Dharma

South Asia Research

Series Editor
Patrick Olivelle
A Publication Series of
The University of Texas South Asia Institute
and
Oxford University Press

INDIAN EPIGRAPHY NECTAR GAZE AND POISON BREATH

A Guide to the Study of Inscriptions in Sanskrit, An Analysis and Translation of the Rajasthani Oral
Prakrit, and the other Indo-Aryan Languages Narrative of Devnarayan
Richard Salomon Aditya Malik

A DICTIONARY OF OLD MARATHI BETWEEN THE EMPIRES

S. G. Tulpule and Anne Feldhaus Society in India 300 BCE to 400 CE
Patrick Olivelle
DONORS, DEVOTEES, AND DAUGHTERS OF GOD
Temple Women in Medieval Tamilnadu MANAGING MONKS
Leslie C. Orr Administrators and Administrative Roles in
Indian Buddhist Monasticism
JIMUTAVAHANA’S DAYABHAGA Jonathan A. Silk
The Hindu Law of Inheritance in Bengal
Edited and Translated with an Introduction and SIVA IN TROUBLE
Notes by Festivals and Rituals at the Pasupatinatha Temple
Ludo Rocher of Deopatan
Axel Michaels
A PORTRAIT OF THE HINDUS
Balthazar Solvyns & the European Image of India A PRIEST’S GUIDE FOR THE
1740–1824 GREAT FESTIVAL
Robert L. Hardgrave Aghorasiva’s Mahotsavavidhi
Richard H. Davis
MANU’S CODE OF LAW
A Critical Edition and Translation of the DHARMA
Manava-Dharmasastra Its Early History in Law, Religion, and Narrative
Patrick Olivelle Alf Hiltebeitel
Dharma
Its Early History in Law, Religion,
and Narrative

alf hiltebeitel

1
1
Oxford University Press, Inc., publishes works that further
Oxford University’s objective of excellence
in research, scholarship, and education.

Oxford New York

Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam

Copyright © 2011 by Oxford University Press, Inc.

Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York 10016
www.oup.com
Oxford is a registered trademark of Oxford University Press
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.

Library of Congress Cataloging-in-Publication Data

Hiltebeitel, Alf.
Dharma: its early history in law, religion, and narrative / Alf Hiltebeitel.
p. cm. — (South Asia research)
Includes bibliographical references and index.
ISBN 978-0-19-539423-8
1. Dharma—History of doctrines. 2. Hinduism—Doctrines—History. 3. Dharma
(Buddhism)—History of doctrines. 4. Buddhism—Doctrines—History. I. Title.
BL1213.52.H55 2010
294.5'2—dc22 2010030676

9 8 7 6 5 4 3 2 1

Printed in the United States of America

on acid-free paper
To Elena Garcés Echavarria
This page intentionally left blank
Contents

Acknowledgments, xi
Abbreviations, xv

1. Introduction, 3
A. Classical Dharma Texts and Their Relative Dating, 5
B. Three Critical Editions, 11
C. Paradigm Shifts on Dharma: The Case of the
Mahābhārata, 20
D. Chapter by Chapter, 29

2. Aśoka Maurya, 35
A. Aśoka’s Inscriptions, 36
B. A Comprehensive Dhaṃma, 45

3. A Vedic History of Dharma, 51

A. Dhárman in Early and Later Ṛgvedic Usages, 52
B. Dhárman as Enigma, 58
C. Dhárman and Ṛgvedic Kingship, 66
D. Mantra Period and Later Saṃhitā Usages, 78
E. The Brāhmaṇas, 84
F. The Upaniṣads, 91

4. Early Buddhism: Three Baskets of Dharma, 103

A. Sūtra Basket Dharma, 107
B. Abhidharma Basket Dharma, 124
C. Vinaya Basket Dharma, 150
viii contents

5. Post-Vedic Brahmanical Dharma, 181

A. Vedic Schools and the Dharmasūtras, 182
B. Toward Consensus in Brahmanical Dharma Texts, 189
C. What’s New with Manu, 196
D. Brahmā in Manu’s Frame Narrative, 208
E. Varṇa (Caste), Āśrama (Life Pattern), the King, Śūdras,
and Women, 215
F. Rājadharma: Establishing a King’s Dharma, 228
G. A Day in a King’s Life, 233

6. Dharma over Time, I: Big Time Dharma, 243

A. Kalpas and Yugas, 244
B. Buddhist and Hindu Kalpas, 246
C. Originary Dharma in the Mahābhārata, 260
D. Kalpas, Manvantaras, and Yugas in Manu and
the Mahābhārata, 266

7. Dharma over Time, II: Prophesies of Disaster, 273

A. The Yuga Purāṇa, 274
B. Variations on the Debacle at Kauśāmbī, 298
C. The Yuga Purāṇa and the Kauśāmbī Myth, 333

8. Women’s Dharma: Śāstric Norms and Epic Narratives, 337

A. Strῑdharma, 338
B. The Law of the Mother, 340
C. Mother Gaṅgā, 345
D. Mother Kālī Satyavatī, 354
E. The Transitional “Three Mothers,” 373
F. Mothers Kuntī and Gāndhārī, 383
G. Kuntī, Mādrī, and Pāṇḍu among the Hundred Peak
Mountain Ṛṣis, 393
H. Settling Mother Kuntī and Her Sons Back at
Hāstinapura, 406
I. Conclusions, 410

9. Two Dharma Biographies? Rāma and Yudhiṣṭhira, 411

A. The Royal Life as Adventure, 412
B. Frames and Frontmatter on Rāma, Yudhiṣṭhira,
and Dharma, 415
C. Sidestories and Subtales, Foregrounding and
Legal Precedent, 420
 contents ix

D. Monstrous Encounters, 429

E. Questionable Killings: Vālin and Droṇa, 453
F. Rāma and Yudhiṣṭhira: Some Comparative Points, 479

10. Draupadī and Sītā: Dharmapatnīs of Two Different Kinds, 481

A. Family Background, Birth, and Childhood, 484
B. Marriage, Divine Plan, Early Signs of Trouble, 488
C. Sītā and Draupadī on Their Svadharma, 495
D. Captivity and Exile, 498

11. Dharma and the Bhagavad Gītā̄, 517

A. Svadharma and Svakarma: Qualities, Merits, and Virtues, 518
B. Who Has Svadharma?, 529
C. Manu and the Bhagavad Gītā: Two Kinds of Karmayoga, 535
D. Where Kṛṣṇa Is There Is Dharma, 542
E. Dharma Rings in the Bhagavad Gītā Proper, 553

12. Dharma and Bhakti, 569

A. Mapping the Divine Plans, 571
B. The Placer and the Ordainer, 585
C. “Avatāra,” 589
D. Friendship, Hospitality, and Separation, 603
E. Ṛṣidharma, 607
F. Rāma and Kṛṣṇa as Guests, Hosts, and Friends, 616

13. Aśvaghoṣa’s Buddhacarita: A Buddhist Reading of the Sanskrit

Epics and Their Treatments of Dharma, 625
A. Aśvaghoṣa’s Buddhacarita, 626
B. The Centrality of Dharma in Aśvaghoṣa’s Buddhacarita, 628
C. Aśvaghoṣa the Brahmin, Buddhist Convert, and Scholar, 633
D. Aśvaghoṣa and Epic Precedents, 635
E. The Buddhacarita and the Rāmāyaṇa, 638
F. The Buddhacarita and the Mahābhārata, 645
G. Postscript on Aśoka, 683

Bibliography, 685
Index, 727
This page intentionally left blank
Acknowledgments

Chapter by chapter this book, like its subject, is what it is thanks to

cumulative fruitful conversations carried out amicably over subjects
of some controversy. I will first mention only those with whom I have
had actual exchanges that bore directly on its genesis and writing,
but, in principle, my engagements with past authors, texts, and
characters in those texts will be hardly less evident to this book’s
readers. I lead off with special thanks to Patrick Olivelle for seeing
this book into the South Asia Research Series, which he serves as
Series Editor. I mention him below in connection with five chapters,
but it has felt to me like I was in conversation with him throughout
the book. I also owe special thanks to Greg Bailey for reading first
drafts of chapters 3, 4, and 5 during his fall semester at George
Washington University in 2005, which was a tremendous help in
getting this project off the ground. Many others also impacted more
than one chapter. For conversations bearing on chapter 1, thanks to
Wendy Doniger, Madeleine Biardeau, Thennilappuram Mahadevan,
Simon Brodbeck, Vishwa Adluri, Joydeep Bagchee, and Jim
Fitzgerald. On chapter 2, thanks to Dan Rudmann for reading it
early, Meera Visvanathan for reading it lately, and Patrick Olivelle for
ongoing input. On chapter 3, thanks to Olivelle for an early reading,
to Jarrod Whitaker for a reading of its Ṛgveda discussion, and to Joel
Brereton, Brian Black, and Steven Lindquist for conversations on the
Upaniṣads. On chapter 4, thanks to Olivelle and John Strong for
reading and commenting on portions and to Strong, Rupert Gethin,
xii acknowledgments

Shayne Clarke, and Black, for supplying helpful bibliography. On chapter 5,

thanks to Olivelle, Adam Bowles, Donald Davis, and Tim Lubin for feedback.
On chapter 6, thanks to Randy Kloetzli for a recent reading, and to Luis
González-Reimann for helpful input. On chapter 7, I am thankful to
Mahadevan and González-Reimann for reading part A and thankful and
much indebted to Jan Nattier for a major reading part B. On chapter 8,
thanks for the invaluable readings it received from Stephanie Jamison,
Laurie Patton, and Uma Chakravarty, and thanks for ongoing feedback
from Perundevi Srinivasan and Elena Garcés. On chapter 9, thanks to Fred
Smith for conversations about possession and to Fitzgerald and Gurcharan
Das for commenting on the heart of it. On chapter 10, thanks to Fernando
Wulff for translating much of it into Spanish, and to Elena for going public
with it at the University of Malaga. Chapter 11 stewed and ripened around
many conversations with Adluri and Bagchee. Chapter 12 benefits from
timely feedback from Brodbeck, Wulff, André Couture, and Christopher
Austin. And chapter 13 was lucky to find Bowles among its early inspira-
tions and Phyllis Granoff, Fitzgerald, and Olivelle among its first readers.
I hope this book will help to make many of these conversations, if they are
not already so, fruitful reciprocally, especially some that have gone long
into the night over disagreements.
Thanks also to the following for memorably pertinent exchanges and sug-
gestions: Aditya Adarkar, Ashok Aklujkar, Marshall Alcorn, Eyal Aviv, Giuliano
Boccali, John and Mary Brockington, Johannes Bronkhorst, Yigal Bronner,
Christopher Chapple, Jonathan Chaves, Ane Kunga Chodron, Tracy Coleman,
David Curley, Waleed El-Ansary, Danielle Feller, Oliver Freiberger, Norman
Girardot, Robert P. Goldman, Sally J. Sutherland Goldman, Barbara Gombach,
Jack Hawley, James Hegarty, John Holt, D. Dennis Hudson, Emily Hudson,
Stanley Insler, Mislav Ježić, Petteri Koskikallio, Meaghan Lisicich, Angelika
Malinar, Greg Mahoney, Tom Michael, Klara Gönk Moačanin, Jason Neelis,
Indira Peterson, Sheldon Pollock, Tamar Reich, Wendy Rodriguez, Henry
Rosemont, T. S. Rukmani, Gregory Schopen, Peter Schreiner, Saswati
Sengupta, David Shulman, Renate Söhnen, Richard Solomon, Bruce Sullivan,
Deepika Tandon, Romila Thapar, Muneo Tokunaga, Georg von Simson, Frits
Staal, Lynn Thomas, Herman Tieken, Yaroslav Vassilkov, and Michael
Witzel.
Thanks to Cynthia Read and Charolotte Steinhardt at Oxford University
Press for their care in seeing this book through production.
Thanks to the George Washington University for a 2005–6 Columbian
College Research Fellowship that allowed me a year’s leave to launch this
research and writing.
 acknowledgments xiii

Thanks to my mother Lucille Hiltebeitel for always remembering to ask

what this book was about. Thanks to my mom, Adam Hiltebeitel, Rachel
Hiltebeitel, Simon Hiltebeitel, Erin Stone and Lucy Stone for the many happy
reminders that other things are more important. Thanks to Elena for shaping
the world in which this book was written.
This page intentionally left blank
Abbreviations

Of Indic Texts and Inscriptions

Ā, ĀpDhS Āpastamba Dharmasūtra

AB Aitareya Brāhmaṇa
AmbS Ambaṭṭha Sutta
AN Aṅguttara Nikaya
AV Atharva Veda
AS Aggañña Sutta
B, BDhS Baudhāyana Dharmasūtra
BĀU Bṛhadāraṇyaka Upaniṣad
BC Buddhacarita
BhavP Bhaviṣya(t) Purāṇa
ChU Chāndogya Upaniṣad
Chinese CS Candragarbha Sūtra
DN Dīgha Nikāya
BhG, Gītā Bhagavad Gītā
G, GDhS Gautama Dharmasūtra
HV Harivaṃśa
KA Kauṭilya Arthaśāstra
KS Kāṭhaka Saṃhitā
KU Kaṭha Upaniṣad
M, Manu Mānava Dharmaśāstra
Mbh, MBh Mahābhārata
MN Majjhima Nikāya
xvi abbreviations

MRE Minor Rock Edict

MS Maitrāyaṇī Saṃhitā
PE Pillar Edict
Rām Rāmāyaṇa
RE Rock Edict
ṚV Ṛgveda
ŚB Śatapatha Brāhmaṇa
ŚBK Śatapatha Brāhmaṇa Kānvīya Recension
ŚBM Śatapatha Brāhmaṇa Mādhyandina Recension
SN Saṃyutta Nikāya
ŚU Śvetāśvatara Upaniṣad
TB Taittirīya Brāhmaṇa
TS Taittirīya Saṃhitā
UMS Umā-Maheśvara Saṃvāda
V, VDhS Vasiṣṭha Dharmasūtra
VS Vājasaneyī Saṃhitā
YK Yājñavalkya-kāṇḍa
YP Yuga Purāṇa
YV Yajur Veda

Other Abbreviations

ABORI Annals of the Bhandarkar Oriental Research Institute

BEFEO Bulletin de l’École Française d’Extrême Orient
CIS Contributions to Indian Sociology
EJVS Electronic Journal of Vedic Studies
HR History of Religions Journal
IIJ Indo-Iranian Journal
IJHS International Journal of Hindu Studies
IT Indologica Taurinensia
JAAR Journal of the American Academy of Religion
JAOS Journal of the American Oriental Society
JAS Journal of Asian Studies
JIABS Journal of the International Association of Buddhist Studies
JIP Journal of Indian Philosophy
JPTS Journal of the Pali Text Society
JRAS Journal of the Royal Asiatic Society
JVS Journal of Vaishnava Studies
KM Kauśambī myth
 abbreviations xvii

MW Monier-Williams’ Sanskrit-English Dictionary

RO Rocznik Orientalistyczny
RoSA Religions of South Asia
WZKS Wiener Zeitschrift für die Kunde Sudasiens
WZKSO Wiener Zeitschrift für die Kunde Sud- und Ostasiens
ZDMG Zeitschrift der Deutschen Morgenländishen Gesellschaft
This page intentionally left blank
Dharma
This page intentionally left blank
1
Introduction

More than for any other project I have undertaken, this one has made
me feel that I should reread virtually everything on India I have ever
read as well as everything I have written. It seems I have been
interested in dharma, however lazily, for a long time. My primary
interests in the Sanskrit epics always dovetailed with dharma, to the
point where my 2001 book, Rethinking the Mahābhārata: The
Education of the Dharma King, put it in the subtitle. Fortunately,
however, I could wait to 2005 to start the writing, for it could not
have taken the form it has, before I was able to digest two works
completed in 2004. The one I read first is Adam Bowles’s
dissertation (2004), now revised as a book (2007), which ranges over
many of the same texts I do as background to his discussion of a
section of the Mahābhārata on “dharma for times of distress.” The
one I read next is the landmark 2004 Journal of Indian Philosophy
volume on dharma, conceptually spearheaded by Patrick Olivelle and
now amplified further as a book titled Dharma (Olivelle 2009). Until
quite recently, scholarly work on dharma has been rather scattered,
and it has been possible to sustain and even promote a nebulous
ahistorical view of the term that many still have today. Thanks to
works of this sort, that has changed.
The findings of this book do, however, differ on some points
from those of Bowles and the authors in Olivelle’s collection.
Researching dharma has proved a rich opportunity to advance new and
unexpected lines of inquiry in areas that have long preoccupied
4 dharma

me: the Sanskrit epics, the historical situations of their composition, and the
ways they come to be received and “read” in Indian culture. But I would like to
mention four findings at this introductory point because they came to me quite
independently of my 2004 readings and each as something of a surprise.
The first finding became part of the plan of a shorter book (Hiltebeitel
2010), one that is now also titled Dharma and was written for undergraduate
readers in a series that allows no footnotes. For that book, one piece of good
fortune carries over to this one. This two-book project on dharma has given me
the opportunity to put some breaks on a career provincialization that had too
easily turned my attentions to Hinduism at the expense of Buddhism. In
teaching the Buddhist poet Aśvaghoṣa’s “Life of the Buddha,” the Buddhacarita,
for the first time in a new Fall 2004 course on South Asian Buddhism, I was
impressed that Aśvaghoṣa’s treatment of dharma was a Buddhist discourse fre-
quently and insistently couched in Brahmanical terms. Aśvaghoṣa deploys
numerous new Brahmanical usages about dharma that cannot be traced to
anything earlier than the Brahmanical dharma texts of our classical period. This
opened the idea that Aśvaghoṣa was not only taking the opportunity to describe
how and why the Buddha searched to discover the “true dharma” but was putting
dharma to use as a term of civil discourse with his Brahmanical counterparts
(both people and texts). This finding remains engrained in this larger book as a
Leitmotif, and not only where it treats Buddhism. It also suggests ways to inter-
pret facets of Brahmanical dharma texts as well, including dharma debates men-
tioned in the dharmasūtras and narrative scenes in the Sanskrit epics.
Of course I wish my short book well, but this was always the real book
I was writing, and the three remaining findings came with it. The second, stim-
ulated by Frederick M. Smith’s monumental study of possession in India
(2006), is that I believe it is now possible to find an answer in our classical
sources, and particularly in the Mahābhārata, to a question I should have asked
more persistently some time ago. In the Tamil cult of the epic’s heroine,
Draupadī, Draupadī’s temples are often called Dharmarāja temples with refer-
ence to her husband, King Yudhiṣṭhira Dharmarāja, who is called “Dharma”
(Tamil Tarumaṉ) for short. How does Dharma personified as a god come to be
a possessing deity? Dharma as a possessing agency can be no more than a brief
focus of this book, but I believe it is one that opens certain unexpected vistas.
The third and fourth findings are ones that shape this book as a whole. As a
newly emergent discourse, dharma could give a “hold” on a world that was chang-
ing. Some months after writing a conference paper titled “Why Itihāsa? New
Possibilities and Limits in Considering the Mahābhārata as History”1 that

1. Hiltebeitel 2009b; the short conference version appears retitled as Hiltebeitel 2010b; the full study with
the original title is chapter 4 in Hiltebeitel 2011-a. The discussion of the Yuga Purāṇa is revised for chapter 7.
 introduction 5

discussed the treatment of yugas or “ages” in the Mahābhārata and the Yuga
Purānạ , it dawned on me that the texts I was treating were engaged not only in a
kind of civil discourse but that such discourse was itself concerned with interpret-
ing dharma over time. My first task in centering this book on that topic was to
write chapters 6 and 7 using that title. This involved comparing a discussion of
Buddhist prophesies of the end of the dharma that I had already written directly
with the Yuga Purānạ ’s Brahmanical prophesies about the decline of the dharma—a
comparison facilitated by the fact that both are ex eventu prophesies of some of the
same historical events. The rest, to make a long story short, is history: a history of
how classical Indian dharma texts envision dharma over time and try to present
it—to borrow a phrase from an American president who at least got to make it a
good campaign slogan—as change you can believe in. Chapters 6 and 7 will now
be this book’s thematic pivot, toward which treating the Aśokan edicts as an his-
torical watershed in chapter 2 will set our initial bearings.
The fourth finding is that by treating certain texts within a concentrated
period and proposing that they be approached within one historical narrative,
dharma would prove to be an historically dispositive term. That is, it would prove
to entail an unfolding set of legal, narrative, and religious projects and strategies
by which it is possible to cut through these texts to show productive ways to envi-
sion their chronology. More than one scholar has described any and every relative
dating of classical Indian texts as “a house of cards.” But the metaphor stacks the
deck in favor of a textual practice that considers every text as itself layered in prin-
ciple with interpolations and strata, such that the deck is virtually infinite and
never far from being toppled by attributions of interpolations or strata in one text
that might reverse its until-then assumed priority or posteriority to others.
I regard this method as having had its field day and to have pushed its ambitious
agenda too indiscriminately. Of the texts that will be central to this book, including
the critical editions of the two Sanskrit epics and The Laws of Manu, the only one
I regard as having demonstrable layers is the Baudhāyana Dharma Śūtra. For the
rest, I regard them as texts that are best considered whole. This means, as I will
outline in a moment, that we will be playing with a deck of fourteen or so cards.
A house of fourteen cards can, I think, be put in some fairly stable order.

A. Classical Dharma Texts and Their Relative Dating

In recognizing that dharma has a history, a book on the topic must thus offer
a textual chronology, however provisional it may be. The term comes to
the foreground in a classical period, a span from between about 300 BCE
and 500 CE. From ca. 1500 to 300 BCE, from the early hymns of the
Ṛgveda through the philosophical speculations of the Upaniṣads, no text is
6 dharma

predominantly about dharma. “Dharma texts” take us into a post-Vedic

“classical” period. The terms in quotes call for some discussion.
The term “classical” is not without its problems. It is, for instance,
conventional to speak of Gupta classicism in association with the notion that
there was a “Hindu renaissance” or “golden age” during the Gupta dynasty (ca.
320–497 CE). But as Romila Thapar cautions, India has just as plausibly had
two other “classical” ages, Mauryan and Mughal, while regional cultures like
the south Indian Cholas have fashioned classical periods as well (Thapar 2002,
280–82). Yet it is not new to speak of the Maurya to the Gupta period as
classical. Indian nationalist historians have done so to foster the notion of an
age centered on the emergence of post-Vedic classical Sanskrit, which probably
takes us back even before the Mauryans to the grammarian Pāṇini (ca. 350
BCE). In any case, my usage, like Thapar’s, is pluralistic, but in a different
sense. It is problematic to think of a single classical period in which classical
texts were being composed in more than one language (Sanskrit, Pāli, Prākrit,
Tamil) and by proponents of contending Brahmanical and Śramaṇic ideolo-
gies. Also, the Mauryan period certainly could be called classical for Sri Lankan
Buddhists, and probably was so, well into the first millennium CE, for Indian
Buddhist communities as well. We will see evidence for this in chapter 13.
My point, however, is not just to pluralize the classical but to use the term to
identify a period in which the term dharma takes on discursive breadth and
density across languages and religious preferences: a period in which this term
provided a forum for the articulation and contestation of new norms. While we
must acknowledge that the Buddha would be preclassical in this sense, since he
is pre-Mauryan, and would almost certainly have said a good deal about dharma,
the actual words attributed to him on the subject in the Pāli canon, which are
conventionally considered to be our earliest window into what he might have
said, are, as we shall see in chapter 4, considered to reflect a state of society not
earlier than around 300 BCE: that is, early Mauryan. This coincides or overlaps
with the third-century BCE dates that have recently gathered some consensus for
the composition of the earliest Brahmanical treatises on dharma, the dharmasūtras.
From this vantage point, one may, and I believe should, view the earliest Buddhist
and Brahmanical dharma texts as framed within a wider discourse about dharma
that we may call classical as it unfolds from the Mauryas through the Guptas. But
it is really between the time of the Mauryans and the Kuṣan ̄ ạ s (the Guptas’ main
immediate precursors) that classical discourse on dharma produced not only
multiple texts, both Brahmanical and Buddhist, but multiple new genres. We
begin to see baroque tendencies in the Kuṣan ̄ ạ period when Mahāyāna Buddhist
authors expand the sūtra format in texts like the Vimalakīrti Sūtra and the Lotus
Sūtra to have the Buddha endorse and utter fantastical novelties about dharma
 introduction 7

that become more classical in China than in India. The Gupta production
of Purānị c texts could then also be called neoclassical or baroque, though in the
different sense of normalizing Brahmanical dharma in a more systematized cos-
mology. Rather than being dharma texts per se, they are texts that move classical
Brahmanical dharma in the direction of an infinitely flexible ideology: a familiar
“sociocosmic” idea of cosmic order, with its tool kit of fully worked out chronom-
etries about vast units of time (kalpas, yugas, manvantaras) and genealogically
based history (solar and lunar dynasties), that can take up the charge of fitting
classical Hindu dharma categories to changing political, geographical, and social
conditions. Meanwhile, when classical usages universalize the term dharma,
regard it as a transcendent value, or project it as the term that defines the truth
claims of a particular religious or civilizational trajectory, they offer new con-
structs that cannot be traced back into prior Vedic texts or the actual word of the
Buddha, much less the teachings of previous Buddhas. Universalistic claims
about an “eternal dharma” have been projected on the past by both Hindus and
Buddhists, but such claims are ways of interpreting the past, and beyond that, the
universe. No preclassical texts use the term dharma in such a fashion.
We must thus identify the texts that participate in this dharma discourse and
offer some idea of their chronology within the formative Mauryan to Kuṣan ̄ ạ
period. In keeping our main focus on texts that can be called “dharma texts,”
including a few that I will call minor classical dharma texts in chapter 7, I will be
foregrounding ones that put dharma front and center as a (if not the) main sub-
ject that concerns them, and that are unimaginable without that subject having
become something of a cause celebre, indeed an intertextual civilizational
discourse, with a lot at stake over its interpretation and implementation, and also
its enjoyment. This definition is meant to include the Sanskrit epics and
Aśvaghoṣa’s Buddhacarita, and to distinguish our dharma texts from some “shas-
tric” or instructional texts on other Brahmanical sciences of the period which,
while they treat dharma as highly important relative to their main topics, do not
treat it as their main topic. For example, the Arthaśāstra looks in on dharma from
the standpoint of the political pursuit of wealth and power; the Kāmasūtra from
the standpoint of erotics; and the Na ṭyaśāstra from that of aesthetics. Likewise,
Patañjali’s commentary on Pāṇini suggests what Ashok Aklujkar (2004) calls a
“grammarian’s dharma”; and the Mīmāṃsāsūtra makes dharma into a transcen-
dent category with which to interpret Vedic ritual. We shall work these approaches
into our discussion at points where they are pertinent, but with the consideration
that their viewpoints arise from engagement—mutual engagement, no doubt—
with what is going on in the dharma texts themselves. They attempt to define the
discourse from carefully refined “infradharmic” or “metadharmic” (notably the
Mīmāṃsā) positions from within or above the fray.
8 dharma

I have thus begun by mentioning major dharma texts and minor ones.
Major ones are those that lend themselves to the cumulative discourse on
dharma that I have been discussing. Minor ones seem to be spin-offs of this
discourse, with more specialized and limited outreach in time and space. The
Yuga Purāṇa seems to have been composed for astronomers. And the Buddhist
prophesies of the end of the dharma had greater impact in China and Tibet
than in India, whose history and geography they reflected and where at least
some of them were composed.
This book will be concerned primarily with twelve major dharma texts or
“text groups” (groupings of collected or thematically related dharma texts) pro-
duced during our classical heydays between the Mauryas and the Kuṣāṇas,
including the Aśokan edicts, which may be called a group of dharma texts. The
following chart lists, in my provisional sequential order, both the twelve major
and two minor dharma texts or text groups that I will be discussing, and breaks
them down into four chronologically defined text clusters. Among the minor
classical dharma texts that will be discussed in chapter 7, I mention here (with
an et cetera) only one—probably the earliest—of the Buddhist texts that will be
compared to the Yuga Purāṇa. This is the Prophesy of Kātyāyana, a text in both

Major classical “Dharma Texts” Minor classical “Dharma Texts”

Cluster 1 (early Maurya):

1. The Aśokan edicts
2. Āpastamba Dharmasūtra
3. Buddhist Nikāyas

Cluster 2 (later Maurya):

4. Buddhist Abhidharma
5. Buddhist Vinaya
6. Gautama Dharmasūtra
7. Baudhāyana Dharmasūtra

Cluster 3 (Śuṅga-Kaṇva or slightly later):

8. Mahābhārata
9. Rāmāyaṇa
10. Manu Smṛti or Mānavadharmaśāstra
(The Laws of Manu)

Cluster 4 (post-Kaṇva to early Kuṣāṇa):

1 1. Vasiṣṭha Dharmasūtra 1. Yuga Purāṇa
12. Aśvaghoṣa’s Buddhacarita 2. Prophesy of Kātyāyana, etc.
 introduction 9

prose and verse versions that were probably composed around the late first to
early second century CE, possibly in Central Asia.
As to text clusters, the notion carries different weight in each usage. Cluster
1 shows only that certain early texts and text collections are relatively contempo-
rary. Since the earliest Buddhist literature may reflect early Mauryan conditions,
since the earliest dharmasūtras cannot be dated precisely with reference to Aśoka,
and since Aśoka’s dharma campaign itself went on for nearly forty years, these
earliest texts can be clustered together without trying to date them further
relative to each other. Cluster 2 is a rather amorphous one at best. It contains
large Buddhist collections that come to be known as “Baskets,” in which some
of the texts included are surely older than the redactions of the collections them-
selves. As mentioned in the last section, the dharmasūtra of Baudhāyana also
appears to be the product of cumulative redactions,2 but this is probably not the
case for that of Gautama, which looks like a single prose composition. Even
more amorphous is Cluster 4, which would appear to have nothing more unit-
ing it than afterthoughts. Only in the case of Cluster 3 do we have a direct rap-
port between the texts themselves. For the two epics and The Laws of Manu
(henceforth Manu), the Śuṅga period of rule by Brahmin kings has long had its
attractions (see Witzel 2006, 482). But while some rapport between these texts
has been widely acknowledged, they have been taken to present intractable inter-
referential problems. I will be urging that we consider a new approach, which I
will get back to in a moment. First, though, since the two epics and Manu will
be most important for us by the criterion of their impact, literary complexity as
poems, and the interpretative challenges they pose, it will be worth briefly sum-
marizing their treatments of dharma along with that of the other great poem we
will be discussing, Aśvaghoṣa’s Buddhacarita.
The vast Mahābhārata is said to have been composed in 100,000 verses
by the sage or Ṛṣi Vyāsa, who frequently enters the story as a grandfather of
the protagonists. It features dharma in three main ways: in didactic sections,
in substories listened to and sometimes told by heroes and heroines, and in
its main story. But Dharma is also a deity who sires one of the heroes, the
eldest of the five Pāṇḍava brothers named Yudhiṣṭhira. From very early on,
even before he becomes king, Yudhiṣṭhira is called Dharmarāja, “the Dharma
King,” which is also an epic epithet for Yama as the god of death and other-
worldly justice. The main story is about a dynastic crisis where two sets of
cousins, the five more noble Pāṇḍavas headed by Yudhiṣṭhira and the more
wicked hundred Kauravas led by the ever-intransigent Duryodhana, go to war
over their divided kingdom after a dicing game in which Yudhiṣṭhira has

2. See Olivelle 1981, 268 n. 23.

10 dharma

gambled away the five brothers’ wife-in-common, Draupadī, and the other
four Pāṇḍavas have sworn to avenge the humiliations and abuse she received
from Duryodhana and some other Kauravas. As in the dice match, so too in
the war, dharma is repeatedly said to be “subtle,” and in the battle, dharmic
and adharmic acts are committed on both sides. Characters are often
delineated through the dilemmas they face in puzzling their way to righteous
yet still ambiguous solutions. The Pāṇḍavas are helped in this, and ultimately
to victory, by both the intervening author and by Kṛṣṇa, who also speaks
authoritatively on dharma throughout, and especially to the middle Pāṇḍava
named Arjuna, Kṛṣṇa’s special friend, in the Bhagavad Gītā, “The Song of the
Lord,” which is considered by many to be this epic’s centerpiece. Both Vyāsa
and Kṛṣṇa are said to be a divine incarnations.
The Rāmāyaṇa, of nearly 20,000 verses, is attributed to the Ṛṣi Vālmīki.
With much less didactic material and fewer substories for prominent charac-
ters to learn from, it focuses on dharma primarily through the adventures of
King Rāma and his wife Sītā, who are presented as paragons of dharma—
though not without episodes that raise questions of their meeting its expecta-
tions and demands. Rāma exemplifies dharma to perfection at least in all his
relations with his father and brothers, if not so certainly elsewhere—most
notably with Sītā and with a monkey named Vālin. It is his perfect allegiance to
the truth of his father’s word that motivates him to undertake fourteen years of
exile to the forest, where Sītā is abducted by the demon king Rāvaṇa. Vālmīki’s
interventions in the story are far less frequent than Vyāsa’s but no less momen-
tous, since he gives refuge to Sītā in his hermitage after Rāma has banished her
even while she is pregnant, helps her raise their twin sons, trains them to sing
the Rāmāyaṇa before Rāma so that he hears his own story, and then, after
Rāma recognizes his sons and asks to see their mother, brings Sītā before him
for her final adieu. Like Kṛṣṇa, Rāma is a divine incarnation. But unlike Kṛṣṇa,
he thinks he is only human.
The Laws of Manu, of 2,675 verses, features only two named characters and
an anonymous host of Ṛṣis. The Ṛṣis ask Manu, who is also known as a primal
sage and king in both epics, to instruct them in dharma, and after Manu tells
them about the creation of the world up to the emergence of humans and their
organization into castes, he then asks his pupil, the sage Bhṛgu, to continue on
his behalf and present Manu’s teachings, which then proceed from the sources
of dharma to all variety of implementations.
Finally, the Adventure of the Buddha has over 2,000 verses, of which only
about the first half survive in Sanskrit. It tells the story of the Buddha’s life
from his conception through his great departure from his father’s royal city, to
his enlightenment, the founding of his order, and his final nirvāṇa, and
 introduction 11

concludes with a homage to Aśoka Maurya. The Buddhacarita is the only one of
these four works to be by a historically identifiable poet. Yet as we shall see in
chapter 13, Aśvaghoṣa names Vyāsa and Vālmīki as precursors, and seems to
know both epics both broadly and well, and he probably also knew Manu.
There is, however, one more text that will come under discussion: the
Harivaṃśa, which is known mainly for narrating supplementary information
about Kṛsṇ ạ that the Mahābhārata did not fully if at all include. It is certainly,
like that epic and the Rāmāyaṇa, a dharma text. I did not list the Harivaṃśa in
our chart. But since, as we shall see, the Mahābhārata calls it and its Bhaviṣya
Parvan (the Harivaṃśa’s concluding unit in its Pune Critical Edition) “appen-
dices” (khilas) of the Mahābhārata, the Harivaṃśa could have been mentioned in
the same cluster as the Mahābhārata, along with the Rāmāyaṇa and Manu.
Indeed, once we begin discussing the Harivaṃśa in chapters 7, 12, and 13, we
will meet new intertextual evidence from two of our classical dharma texts, the
Yuga Purānạ and the Buddhacarita, that could place the Harivaṃśa in that cluster
chronologically: perhaps “late” in it but still somehow within it, and thus earlier
than is usually thought. While such evidence will encourage an hypothesis, it is
best for now to let it arise in context and refrain from prematurely clustering this
“appendix.” The topic is elusive, and more work needs to be done.

B. Three Critical Editions

Aśvaghoṣa, from the first or second century CE, thus strengthens our case for
looking at the Mahābhārata, the Rāmāyaṇa, and Manu as a significant cluster,
since, as will be brought out in chapter 13, he probably did so himself. In this
cluster, however, let me indicate that I view the Mahābhārata as likely a little ear-
lier than the Rāmāyaṇa and Manu, and also as possibly overlapping them during
its slightly longer period of composition (I have proposed that the Mahābhārata
was composed over no more than two generations by a committee working bet-
ween 150 BCE and the turn of the millennium).3 Leaving specifics of such dating
for fuller discussion beginning in chapters 5, this relative chronology does call for
some preliminary clarification of the position I take on the implications of the
critical editions of these three highly prominent dharma texts.
As anyone who knows these three texts at all well can see, I have summa-
rized them in ways that reflect the breadth of their critical editions, and in ways
that go well beyond what most scholars—including some who have taken part

3. Hiltebeitel 2001a, 20–21; 2004a. Although I do not follow him on either dating (see chapter 5), Olivelle
also makes a reasonable argument linking Manu and portions of the Mbh with Kuṣāṇa times (2005a, 24–25).
12 dharma

in editing and translating their critical editions, and others who strongly
advocate the use of those editions—have said when summarizing them or
assessing what they think would have been basic or original to them.4 There
has always been the assumption that the epics should be about conflicts bet-
ween mainly male grown-ups whose adventures lead up to, and end with, a
triumphant good fight;5 and that Manu should be about the law. Anything else
in them has been fair game for the so-called higher criticism and its excavatory
strategies of positing stages and calling passages interpolations. I have cited
elements from the frame stories of all three texts, yet frame stories are among
the first things to be deemed unepical or nonlegal, and thus extraneous.6 As
regards the two epics, I have cited substories and didactic components, which
have likewise come under the axe, as have stories of the heroes’ youths and
postwar aftermaths.7 Divine paternity, mentioned in the case of Yudhiṣṭhira
but extensive in both epics along with other kinds of divine and demonic incar-
nation, has also looked belated and suitable for eviction,8 were it not that it
would strip the main stories.9 Finally, where I mention the divinity of Kṛṣṇa
and Rāma, and moreover that of Vyāsa, I refer to passages typically excised
because they have to do with bhakti, devotion. On bhakti, James Fitzgerald and
John Brockington have done the most recently to reinforce settled opinions.10
Since Fitzgerald’s views are currently the most frequently quoted, we may
take his repeated dismissals of bhakti in the Mahābhārata as late and “Gupta”

4. Compare my summaries of the two epics in Hiltebeitel 2006c and 2006d, which also work from the
complexities of their critical editions, but more expansively than the brief summaries above, with Fitzgerald
2004c on the Mahābhārata and Brockington and Brockington 2006, ix–xxx on the Rāmāyaṇa. More attentive to
the complexities raised by the Rāmāyaṇa’s critical edition is the summary in Goldman and Sutherland (Goldman
2004).
5. See my discussion (1999a, 38–39) of the views of John D. Smith that the “epic world is essentially a
male world” and that goddesses in epics are later developments (1989, 182, 188–89).
6. For the Mbh frame, see Grünendahl 1997, critiqued in Hiltebeitel 2006a; Oberlies 1998, critiqued
in Hiltebeitel 2001a, 93 n. 5. The Rāmāyaṇa frame, introduced in that epic’s upodghāta or “Preamble” and
circled back to in the Uttarakāṇḍa (Book 7), falls victim to the assumption that the text’s allegedly predivinized
“core” is found in Books 2–6. Olivelle 2005a, 88–92 sees most of Manu’s first framing chapter as added in
nine “excurses,” though not the frame story as such involving Manu, Bhṛgu, and the Ṛṣis. On frame excision,
see also Hiltebeitel forthcoming-e.
7. On substories being within the Mahābhārata’s archetypal design, see Hiltebeitel 2005a, forthcoming-d,
and Biardeau 1979, 120 and n. 4. With reservations about the lateness of the “didactic,” see Bowles 2007, 34–35;
2009; Brodbeck 2010a. Cf. Rabault 2004 on opposition to Joseph Dahlmann’s earlier views on the didactic. On
the epic ending shortly after the war, see Tokunaga 2005b and 2009b, as critiqued by Brodbeck 2010a, 159–60.
8. See Hiltebeitel 2001a, 164 n. 118 on Winternitz 1933–34, van Buitenen 1973, and Bigger 1998 viewing
Mbh 1.189 on “The Five Former Indras” as silly, isolated, and thus arguably late even though the critical edition
includes it. For renewed discussion, see chapter 12.
9. Thus Fitzgerald concedes divine paternity in the Mahābhārata’s “initial development as the Pānd ̣ ạ va
epic” (2004c, 54)—a notion he follows up with a theory of the Pānd ̣ ạ vas’ “invention” (2007b); see chapter 7.
10. See Hiltebeitel 2004a, 2005d on Fitzgerald 2001, 2003, and 2004a; Hiltebeitel 1999c and 2001a,
22–23 on Brockington 1998.
 introduction 13

as emblematic. Here is what he says, in an essay meant to introduce the

Mahābhārata, about the congeries of elements that accompany bhakti into the
text after a “main Mahābhārata” had undergone its first written redaction:11 “in
later centuries subsequent redactors wove together meaningfully many of the
new religious ideas (such as elements of Śaivism, the worship of the goddess,
bhakti, the theory of the yugas, and others) that emerged into prominence in
India between the time of that original development and the time that it became
more or less fixed (some time between 300 to 450 CE)” (2004c, 54).
Readers of this book should be prepared to consider evidence that these
“new religious ideas” were already old by Gupta times because the Mahābhārata
had woven them meaningfully into its archetypal text about four or five cen-
turies earlier. The critical editions of the epics give no grounds whatsoever for
any such excisions, and those who want to maintain them have had to continu-
ally reinvent their reasons for doing so with no cumulative results, only
cumulative assumptions.12 I thus advocate a new approach, or better, continue
to advocate one that is no longer so new, which I have been developing since
about 1992, and call my literary turn (see Hiltebeitel 2005c, 81–83; Adluri and
Bagchee 2011). Since I was working at that time on the Mahābhārata, I describe
this turn first as it bears on that epic, but I also consider it to apply in basic ways
to the Rāmāyaṇa and Manu as well. Where the Pune Critical Edition of the
Mahābhārata keeps faithfully to its criteria for inclusion and exclusion,13 I take

11. Fitzgerald 2003, 811 and n. 32. Fitzgerald grants that “main Mahābhārata” is a “vague expression.”
Bronkhorst (in press) is attracted to the idea, but only to provide him with a long time for interpolations, as he
sees them, and without considering the “main Mahābhārata’s” narrative implications. Fitzgerald would seem to
have coined the term, in preference to the often-used “main story,” as one that admits to a text. Fitzgerald agrees
with me that the first (for him, pre-Gupta) redaction, his “main Mahābhārata,” would already have been a written
text (814–15). One tip-off that this construct is unsuccessful and really forced is Fitzgerald’s concession to admit
into this “main Mahābhārata” the “‘black, covert’ characters of the three Kṛṣṇas,” namely, Kṛṣṇa Vāsudeva, Kṛṣṇa
Draupadī, and Kṛṣṇa Dvaipāyana as “ ‘dark,’ ‘obscure,’ or ‘secret’ (kṛṣṇa) holy agents” who are “representative of
the world’s Vedic Brāhmaṇs” (2004c, 56), with Vyāsa further as “the mysterious (kṛṣṇa) agent of Brāhmaṇism”
(61), while rejecting as “late” not only bhakti episodes that link the ‘three Kṛṣnas” together but also the “author”
trope of Vyāsa as developed in the Mahābhārata’s frame narratives (see Fitzgerald 2003, 815–17). On the “three
Kṛṣṇas,” cf. Hiltebeitel [1976] 1990, 60–76.
12. Brodbeck 2010a, 189, n., citing Hiltebeitel 2001a, 163, finds it one of my “most telling criticisms
against the differentiation of textual layers within the Mahābhārata . . . that the results are not cumulative.” The
most ingenious assumption meant to support continued research into prior strata has been offered by Bigger
(1998; 2002) and Fitzgerald (2001, 68), who call the archetype reconstituted by the Pune Critical Edition a “nor-
mative redaction,” implying that once it became the set text, it submerged others that preceded it. Yet whereas
Bigger posits a “fixed oral transmission before the normative redaction was compiled” (24), Fitzgerald holds that
there would probably have been a prior written redaction (2003, 811 and n. 32). I agree with Fitzgerald that the
first “redaction” would have been written and with Bigger that the first written version would be the one approx-
imated by the Critical Edition archetype, but I do not concur with either of them that a secondary “normative
redaction” left prior traces of large scale interpolations to be divined by higher criticism.
13. For balanced discussion, see Sutherland 1992; Brockington 1998, 56–67; Bigger 2002; Mehendale
2009; Brodbeck 2009a, 3–12; 2010a, 154–57. For fuller consideration of the Pune Critical Edition and its critics,
see Hiltebeitel forthcoming-a.
14 dharma

it to be a largely successful reconstitution of the Mahābhārata as a work of writ-

ten literature14 whose literary experiments, symbolic and philosophical com-
plexities, thematic consistencies, and even its jarring juxtapositions and
eye-opening inconsistencies are sufficiently explained by multiple authors
working under a single inspired authorly design. The archetype uncovered by
the Pune Critical Edition does not encourage speculation that it reconstitutes
only a secondary “Gupta redaction” expanded from a less ample and entirely
hypothetical earlier version, whether written or orally “fixed,”15 behind it.
As I have done in various other writings, I will, in this book, sometimes draw
attention to the merits of the text of the Mahābhārata’s critical edition in getting
us back to what I would now call a baseline Mahābhārata from the timespan I
have mentioned.16 I thus do not share the view of many that the Mahābhārata
critical edition is no more than a tool for renewed excavation. There seem to be
three main ideas that run repeatedly through criticisms of the Pune Critical
Edition: that it does not represent the fluidity and orality of the Mahābhārata tra-
dition; that it used Western methods that were inappropriate to Indian texts; and
that it is not a traditional text but is instead “no text.” In an article with the phrase
“no text” in its title, Sally Sutherland (1992) assesses the critical editions of both
epics. Raising these three points, she concludes that “such evidence” does not
warrant that we “turn away from these editions and regret them as dinosaurs of
orientalism” (1992, 88). Yet she offers evidence only on the “no text” charge, and
that only with regard to the Rāmāyaṇa, finding “numerous instances where the
editors . . . have constructed a story” that is not found anywhere outside of the
reconstituted text. She suggests that the Rāmāyaṇa editors may have been
attempting to meld popular variants (86–87). On the other two points, there can
be no evidence for her claims that both epics survived in oral form for over a
thousand years before their earliest “exemplars” (85, 87). Sutherland hints that

14. For the Mahābhārata as written, and with regard to “oral theory,” see Hiltebeitel 1999a, 7–8; 2001a,
2–4, 18–24; 2005c. On orality and writing as they bear on the canonization and transmission of early Buddhist
sources, but weighted to demonstrate orality, see Veidlinger 2006, 17–62. One might hypothesize from his evi-
dence that Buddhists self-consciously weight their written historiography toward orality at “originary” or
“founding moments” (see 24–28, 34–36, 43–54, 61). So, in my view, do the Sanskrit epics in their constructive
fictions about Sanskrit-speaking “bards” (see Hiltebeitel 2001a, 4, 13 n. 31, 96–106; 2005c, 84, 89, 94). There is
no Indian evidence of an epic genre—not to menation oral epic—during the Vedic or pre-Mauryan period.
15. See n. 12 above. Cf. Doniger 2009a, 260, giving the Mahābhārata six centuries to have been composed
“before the Guptas”; van Buitenen (1978, 145–54), regarding its completion as most likely post-Gupta. Bigger also
thinks his “normative redaction” does not require a “famous dynasty” to have produced it (2002, 27). The point
is that we have little reason to suppose Gupta patronage of any Mahābhārata redaction.
16. I arrive at the term “baseline text” in Hiltebeitel forthcoming-d, one of four essays written since com-
pleting this book (see also Hiltebeitel forthcoming-a, -b, and -g) that carry forward my argumentation on these
matters. See also Adluri and Bagchee 2011. Also affirming an important critical edition reading, see Austin
2009, as discussed in chapter 12 § A.
 introduction 15

Vishnu S. Sukthankar’s sense for things Indian may have kept him on track, and
favors his view (as General Editor of the Pune Critical Edition and editor of its
first and third volumes) that “even in its early phase the Mahābhārata tradition
must have been not uniform and simple but multiple and polygenous”
(Sukthankar 1933, lxxix). She prefers this to Franklin Edgerton’s view (as editor of
volume 2) that “every line of the text once had a definite, precise form, even
though we are frequently uncertain about just what that form was. It is not an
indefinite literature we are dealing with but a definite literary composition
(Edgerton 1944, xxxvi–xxxvii). Yet I think Edgerton was more reliable on this
point than Sukthankar (Hiltebeitel 2001a, 24 n. 93).
Unlike Sutherland, Wendy Doniger and Madeleine Biardeau have taken
these three issues as grounds to reject the Pune Critical Edition, each offering
unfortunate broadsides against it. Since I more often than not agree with these
two scholars, and with Georges Dumézil and David Shulman,17 who concur with
them, though more casually, it is worth attempting to clarify my differences with
the four of them on the merits of the Pune Critical Edition. I will show at points
in chapters 9 and 13 how Shulman and Biardeau have led readers down little blind
alleys when preferring to work from Southern or Northern “Vulgate” editions.
It is important to recognize that both Doniger and Biardeau approach the
Mahābhārata from the standpoint of having first worked on the purāṇas. From
this vantage point, we can understand how each brings an avowed preference
for “tradition” and multiple variants over any text.
Doniger’s stance can be traced back to her early work. In The Origins of Evil
(O’Flaherty 1976), which she now speaks of as her “second best book” (Reddy
2009), she submits the Mahābhārata and Hindu mythology more generally to
a “heavy marination in historical periodization” (Hiltebeitel 1979, 270–72).
According to Doniger, the Mahābhārata, as well as other narratives during its
supposedly long period of development, can be traced through “three major
trends: sacrifice, asceticism, and bhakti” (O’Flaherty 1976, 78). Building on
this sense of the Mahābhārata’s longue durée, which can be longer still for
Doniger when she sometimes invokes the Indo-Europeans, she speaks of the

17. See Shulman 2001, 25–26: “the attempt to pin down a precisely delimited text does not appear too
promising, despite the existence today of a so-called ‘critical edition.’ ” Here, Shulman prefers to cite from what
he calls “the Southern Recension (SR) of the Mahābhārata, ed. P. P. S. Sastri (Madras, 1931–33).” Yet as the full
title of Sastri’s first volume—The Mahābhārata Southern Recension, Critically Edited by P. P. S. Sastri (1931)—
shows, Sastri himself aspired to be editing a critical edition: “It is this want of a really representative text of the
Southern Recension that has led the present Editor to undertake the task of a critical edition of the Southern
Recension” (Sastri 1931, xiii). As Sukthankar’s criticism makes clear (1942, xxxiv–xxxv), Sastri’s critical edition of
the Southern Recension was, for many reasons, not as good as the one Sukthankar himself was crafting (in star
passages and appendices) as a provisional byproduct of his work on the full Mahābhārata critical edition. For
example, Sastri used no Malayālam manuscript until the Āraṇyakaparvan (see Hiltebeitel forthcoming-d).
16 dharma

Mahābhārata as having “various recensions . . . even at the earliest stages of its

composition”; and while granting that J. A. B. van Buitenen had good enough
reasons to translate from it, she decries his choice to “approve of” it, since for
her, it is “no text at all” (1978, 22). She wishes that van Buitenen had used his
“excellent judgment” to select good interpolations to fold into his translation,18
and holds that “any structural analysis of the epic would of course demand all
available variants of the text” (23)—as if structural analysis were not also pos-
sible of the text the critical edition gives us. Now, thirty years later, she goes so
far as to say that the Mahābhārata was always a “tradition” and was never a
text (2009a, 263–64, 2009b). In her recent book The Hindus: An Alternate
History, she upgrades an old encyclopedia analogy (critiqued in Hiltebeitel
2001a, 161–63; 2005c, 88) to say the Mahābhārata is “like an ancient Wikipedia,
to which anyone who knew Sanskrit, or who knew someone who knew Sanskrit,
could add a bit here, a bit there” (2009a, 263–64). Then in a book review of
John D. Smith’s abridged translation from the critical edition (2009), she no
longer allows for good reasons to work from it but goes overboard in decrying
Smith’s allegiance to it. After repeating most of the Wikipedia analogy but
dropping its most improbable line about those “who knew someone who knew
Sanskrit,” she sees “a deeply patriarchal metaphor at the root of the whole idea
of tracing text-stemmata.” Finding “no trunk to this textual tree” (despite its
patriarchal root) unless it is a banyan, she claims that “[t]here are several recen-
sions of the Mahabharata, each preserved and cherished by a particular
community” (no evidence for which is mentioned).19 Having reassured us in
her book that the epic is not the “monstrous chaos” once thought (2009a, 264),
she now makes the critical edition the new monster: “The critical edition . . . is
like Frankenstein’s monster, pieced together from various scraps of different
bodies; its only community is that of the Pune scholars, the Frankensteins”
(2009b, 18). It is hard to see this recent assessment as going anywhere, unless
it is backwards. So far Doniger’s only advance is to have recognized that while
she could get by in her book with the bit about those “who knew someone who
knew Sanskrit,” she should drop it from a scholarly book review.
Biardeau, on the other hand, offered her broadside against the Mahābhārata
critical edition toward the beginning of her work on the epic (1968, 1970a,

18. Kulkarni 2009 ends his review of Doniger 2009a: “occasionally, one feels that, rather than dig deeper
into the contradictions of history, she too, like her subjects [i.e., stories] stretching back over thousands of years,
prefers just to tell one more story.”
19. Biardeau also avers that the Northern Vulgate edition of Nı ̄lakaṇṭha “became very popular” and
“received a warmer welcome than the present critical edition” (1968, 121). She does not say where, when, or by
whom it received this “warmer welcome.” My experience in purchasing it in 1974, in the Kinjawadekar 1929–33
edition, was that it was out of print and available in Pune only from local specialists in Sanskrit books.
 introduction 17

1970b), and since then returned to the topic only in asides.20 In her debate over
the merits of the critical edition with V. M. Bedekar, Biardeau also avers that
“such a text never existed” as the one reconstituted (1968, 123). She brings up
Dumézil’s preference for the Calcutta edition over the critical edition (Dumézil
1968, 34) to argue that the current western study of texts is more scientific than
the “outdated” one implemented in the Pune text’s editors and defended by
Bedekar (Biardeau 1970a, 180). Again, it is no coincidence that, as an Indo-
Europeanist, Dumézil takes this similar stance. Yet one hardly needs to read
between the lines to see that Biardeau’s debate was not so much with Bedekar
and the mostly Indian editors of the Pune edition, whom she depicts as strad-
dling an impossible allegiance to traditional Indian punditry and an “outdated”
German philology, but with German philology itself. The third party in this
debate is obviously French Indology. One sees this where Biardeau takes umbrage
at Sukthankar’s dismissal of “my paramaguru” Sylvain Lévi,21 who argued that,
rather than continue reconstructing a critically edited text, Sukthankar would do
better to reprint the northern Vulgate of Nīlakaṇṭha and show all the newly found
variants along with it to let readers decide their own preferred readings (see Lévi
1929, 347; Sukthankar 1933, lxxxiii–lxxxiv; Biardeau 1968, 115–16).22 Biardeau
says that as a “result of the impact of . . . the dominant trend of Western Indian
studies, that is, historical philology,” “modern pandits . . . in the name of
science . . . have introduced the historical dimension into the realm of myth,
where it cannot exist”23 (1968, 122). Bedekar has the appropriate answer: “Even
myths are studied scientifically by anthropologists. But here the matter is quite
otherwise. Are mss [manuscripts] of different recensions and versions myths?”
(1969, 223). The point is not whether one scientific method is right for all texts,
but which methods are best for which texts. Over and over Biardeau lumps
together the pair “epics and purānạ s” on the grounds that both begin as oral, and
thus call for the one and the same type of open textual study. But the problem is
not whether or in what way the Sanskrit epics and purānạ s have oral traditions
behind them, but what impact those oral precursors would have had once those

20. Once she got past these exercises, Biardeau could be an astute critic of the critical edition where it
failed to follow its own principles. See e.g. Biardeau 1979, 118 on Sukthankar’s reconstruction of the name
Duḥṣanta based on the Śatapatha Brāhmaṇa’s usage of Bharata Dauḥṣanti. As Mehendale 2009, 8 notes, the
critical edition sought to avoid “emendations.”
21. Lévi taught on Indian religions when the 5th Section on Sciences Religieuses of the École Pratique des
Hautes Études opened in 1886, which Biardeau was to direct a century later. See R. Rocher 2009, 639, and
635–38 on Lévi’s “barbs” against German scholarly hegemony. Biardeau 1968, 122 implies Indian subordination
to German methods, as cited in my next sentence.
22. I agree with Bedekar’s defense of Sukthankar as to the shortsightedness and really the impossibility of
such an approach, and van Buitenen’s critique of Biardeau for renewing it (see Bedekar 1969, 212–13; van
Buitenen 1978, 152; Biardeau 1968, 119, 122–23). Lévi 1934 softens his critique.
23. Biardeau 1968 32. I reverse the order of the quotes.
18 dharma

texts were written.24 With Biardeau, the problem with this frame of reference is
twofold. First, an epic is not a purānạ .25 As Ludo Rocher demonstrates, there are
excellent reasons to be dubious about critical editions of cross-purānị c textual
pieces like the Purānạ Pañcalakṣaṇa, and likewise of purānạ s themelves.26
But none of these reasons hold for the Mahābhārata or Rāmāyaṇa, both of
which—not to mince words—are, as the amplitude of their critical editions now
demonstrates, epoch-changing monumental literary masterpieces. Second, as
van Buitenen shows, Biardeau’s exemplum to show the superiority of the Vulgate
over the Mahābhārata critical edition by comparing two variants of the story of
King Kārtavīrya Arjuna draws on quasi-purānị c interpolations of the Vulgate to
support her structuralist interpretation.27
Biardeau did, however, refine her views over the years on the question of the
Mahābhārata’s primary orality. Whereas in her 1968 critique of the critical
edition, she consigned oral Mahābhārata to “the bards (usually non-brahmins)”
(1968, 116), in 2002, acknowledging that “Today, certain specialists think that it
is materially impossible to regard [the Mahābhārata] as an oral composition
because of its dimensions,” she weighed in for orality in the name of Brahmins’
memories, the possibility that they could conceive a work on this scale, and their

24. Although Bedekar does not question a single approach to both “epics and purāṇās,” he does point out
Sukthankar’s view that the “problem” of the Mahābhārata was “sui generis” in comparison with Western texts
(Bedekar 1969, 219–20).
25. Shulman raises a seeming problem here, saying the Mahābhārata “is ‘itihāsa’ or, as it calls itself,
‘itihāsapurāṇa’ (see 1.1.16–20)” (2001, 22). This reference is not in the baseline text of the Critical Edition or in
anything in its apparatus around that point. The epic uses the compound only three time, and never to define
itself: first in the dual (itahāsapurāṇābhyām) to say that itihāsa and purāṇa should “support the Veda” (1.1.203;
van Buitenen 1973, 31); next in a plural describing topics studied by Dhṛtarāṣṭra, Pāṇḍu, and Vidura (1.102.18);
and last—the only instance in the singular—describing Bhīṣma as knowing itihāsapurāṇa fully (12.50.34). Thapar
also says the Mahābhārata is itihāsapurāṇa (1993, 136), and equally misleadingly that it is kāvya ([1999] 2002,
5–6). Cf. Pollock 2006, recognizing that the epic’s itihāsa (“narrative of ‘the way it once was’ ”) ”genre identity is
no trivial matter” (17), but still explaining it as kāvya (224) to accommodate it to his binary of culture/power,
kāvya/rājya, poetry/polity (29–30). The epic is not self-identifed as kāvya until the interpolations in which Brahmā
names it so and Gaṇeśa writes it down (Mahābhārata 1, Appendix 1, No. 1; see Hiltebeitel 2011a, 311).
26. L. Rocher 1986, 97–99 objects, with reference to Kirfel’s work (1927) on the Purāṇa Pañcalakṣaṇa or
“Five Features of a Puraṇa,” to the intent to “reconstruct an ideal Ur-text of any purāṇic passage” (97), and the
assumption that “mini-purāṇas are more basic than their conglomerations into large purāṇas in the traditional
sense” (98). “It goes without saying that, once one looks at purāṇas as purely oral tradition, as a tradition carried
forward by individual story-tellers, and which is, therefore, authorless and anonymous, as a tradition only parts
of which have accidentally been committed to writing,—it goes without saying that in that case critical editions
based on the standard rules of textual criticism make little sense” (99). Indeed, the Mahābhārata cites a Vāyu
Purāṇa at 3.189.14, though not any later one known by that name; similarly, Āpastamba Dharmasūtra 2.24.5–6
cites an early Bhaviṣyat Purāna (see L. Rocher 1986, 85–86).
27. Van Buitenen 1978, 145–52. Biardeau’s closing arguments become rather catch-as-catch-can. She
chides Sukthankar for using the Bombay Edition rather than the critical edition for his 1942 lectures, missing the
point that he could not have cited the critical edition because it was not yet complete. Having declared a preference
for “more developed versions,” she brings up the possibility of versions being “shortened (possibly according to
the fees to be expected).” And in suggesting that “rather than reconstruct a single authentic text we had better
published all regional versions” (Biardeau 1970b, 302–3), she raises the red herring of “inauthenticity.”
 introduction 19

habitation of a Vedic “universe of sounds” (Biardeau 2002, 2: 747–49). From

bards to Brahmins is a considerable advance. The notion that the Mahābhārata
goes back to a time when bards eulogized clan warriors has offered up a tribal
conception of the early Mahābhārata (there seems to be no corresponding
insistence on the tribal origins of the Rāmāyaṇa) that the Pune Critical Edition
should help us to put to rest.28 Since about 1947, the bardic component of the
tribal hypothesis has sought help from so-called oral theory. But much as oral
theory has enriched our awareness of performative aspects of Mahābhārata and
Rāmāyaṇa verse (see Hiltebeitel 2001a, 19 and n. 74), it cannot take us back to
pre-Brahmanical origins of these epics. As Biardeau’s 2002 book makes clear in
its title, the Mahābhārata is “a foundational text of Brahmanism.” She and I just
disagreed about whether it was composed orally or in writing.
I thus view the Mahābhārata’s Pune Critical Edition as having a more successful
outcome—indeed, one more successful in uncovering a written archetype—than
its first general editor Sukthankar could perceive. As studies by Thennilappuram
Mahadevan show (2008, 2010), along with studies of mine that build upon his
findings (2006a, forthcoming-a, forthcoming-d), we can posit that an early version
of a Mahābhārata that would lie behind its shortest Northern Recension manu-
script, which is in a Śāradā script,29 would have been brought to south Indian Tamil
country by Brahmins called Pūrvaśikhās (those who wear their tuft of hair in front)
during the Sangam period, where, from its correspondences to the shortest
Southern Recension manuscripts in Malayālam script, we can deduce and indeed
date with good conviction the modifications that produced the earliest Southern
Recension manuscripts. A sedulous and careful makeover by Brahmins who saw
themselves as custodians of the Mahābhārata, yet who nonetheless wanted to
spruce it up in ways they deemed useful and pertinent to their adopted south
Indian milieu,30 would have been accomplished before the Kaḷabhra Interregnum,
ca. 300–500 CE. For at that point the Pūrvaśikhās split into two communities, one
of which—the future Nambudiris in Kerala—would keep their eventual Malayālam
version short; and the other, the Cōḻiya Pūrvaśikhās remaining in the Tamil
country, would see their version in Grantha script overlaid with Northern Recension
additions introduced by the next wave of Brahmins migrating from the north from
the Pallava period on, called Aparaśikhās (those who wear their tuft in back).

28. See Hiltebeitel forthcoming-a on this premise of an originally tribal Mahābhārata, drawing on Adluri and
Bagchee forthcoming particularly for their discussion of the views of the uncle and nephew pair of Adolph
Holtzmanns, but also noting that many scholars outside Germany have also promoted it, notably Hopkins [1901]
1969, 386, 397–98.
29. Mahadevan 2010 calls it the “*Śāradā text,” that is, the “proto-Śāradā text.”
30. On these points, see Hiltebeitel forthcoming-d, a study of the Southern Recension’s major alterations
of the Śakuntalā Upākhyāna, building on the astute insights of Sukthankar 1933, xxxii–xxxvi.
20 dharma

As Mahadevan has noted, his findings for the Mahābhārata critical edition
hold implications for the Baroda Critical Edition of the Rāmāyaṇa, whose early
editors ignored the implications of the similar relative brevity of Malayālam
Rāmāyaṇa manuscripts, making it impossible for the last editors involved in
the project to do any more than acknowledge the seriousness of the oversight
(2008, 99–100 n. 2). Meawhile, Olivelle’s critical edition of Manu simply does
not involve a large apparatus of major passages dropped because they are not
found in all the text’s known manuscripts.31 Indeed, Olivelle too seems to have
taken a literary turn away from the assumptions of predecessors like Georg
Bühler and E. W. Hopkins, who looked behind Manu for nuclear oral precur-
sors. As we shall see in chapter 5, Olivelle, despite still labeling numerous pas-
sages as “excurses,” by which he means interpolations, has discovered a “deep
structure” in Manu that yields a literary text in four main sections (2004b,
xxvii–xxx; cf. 2005a, 7–11). I think what Olivelle calls “excurses” may be thought
of more profitably as segments where Manu resorts to topics that take it beyond
the legal tradition preceding it, and that in such cases Manu may sometimes be
drawing on other knowledge traditions or even sources. In any case, Olivelle
says that Manu’s “deep structure” opens the possibility that, “If not by an
individual, then [Manu] must have been composed by a ‘strong chairman of a
committee’ with the help of research assistants who carried out his plan”
(2005a, 7; cf. 19, 26). One wonders if Doniger still holds to her view that Manu
was “composed in increments over several centuries” (1991, xliv–xlv); “a hotch-
potch” (lv); its “apparent inconsistencies” being “the natural outgrowth of cen-
turies”; and a work whose “śloka verse form” suggests orality (lvii). More
recently, she seems to have acceded to Olivelle’s dating of Manu to about
100 CE (2009a, 202; cf. 26), but she does not reconsider her view of the text.

C. Paradigm Shifts on Dharma: The Case of the Mahābhārata

I cannot expect scholars invested in incrementalist approaches to drop their

investments. I can only urge them to keep an open mind to the possibility that
a literary approach will shed new light on some intertextual problems. Scholarly
orthodoxies are probably harder to challenge after over a century of entrench-
ment and popularization than views on dharma in classical India, which could
admit change over time if not in essence. Every chapter in this book will

31. “All the mss. and commentaries of the MDh contain basically the same text. The additional verses
found in some can easily be detected. There are no major recensions of the MDh giving longer or shorter versions
of the text, as, for example, in the case of the two epics” (Olivelle 2005a, 374).
 introduction 21

wrestle with some aspect of that tension. But since by chapters 5 and 6 it will
have emerged that I regard one text, the Mahābhārata, as having been most
effective in putting dharmic change on the table, and indeed in motivating
change by doing so, I turn now to Bimal K. Matilal’s innovative treatment of
the Mahābhārata as the locus of a paradigm shift on dharma. Matilal’s
discussion occurs in a book of collected essays titled Ethics and Epics (2002)
that was edited by Jonardan Ganeri. As Ganeri says in a luminous study of his
own, citing one of Matilal’s exemplary stories from the Mahābhārata about a
fool named Kauśika, which will be cited below, “The idea that morality has a
history is a difficult one” (Ganeri 2007, 91). This is a nice distillation of
Matilal’s argument. This examination of his insights will serve to preview
some later chapters. I will be introducing Matilal’s approach to Kṛṣṇa along
with a view of the Mahābhārata that can be squared with it: Biardeau’s idea
that the epic’s riposte to Buddhism was achieved by a “bhakti swerve” away
from Brahmanical orthodoxy—a formulation I will open up in chapter 12. And
in relating Matilal’s comparison of the dharmic kings of the Mahābhārata and
Rāmāyaṇa, Yudhiṣṭhira and Rāma respectively, I open a comparison that I
will pursue further in chapter 9.
It has emerged already that, for Vālmīki, Rāma’s perfection in dharma lies
not so much in embodying it, as Yudhiṣṭhira does as the god Dharma’s son, as
in exemplifying its rules; and when rules conflict, in discerning what he con-
siders the higher value, which is always in accord with hierarchical and
patriarchal norms, topped by the truth of his father’s word. According to
Matilal, the ethical virtues extolled in the Rāmāyaṇa are “nothing if not formal-
istic in character,” by which Matilal means that they depend “upon the
fulfillment of a formal promise: of a formal duty of a son to his father, of a hus-
band to his wife, of a friend to his friend” (2002, 85). Rāma thus applies
“rational thinking” to choices that at least he deems to be “only apparent.”
A “higher type of critical thinking,” however, can come into play in “cases of
genuine dilemma” where “the conflict is real” (68). Matilal finds guidance
through such dilemmas from the “devious divinity” Kṛṣṇa whose “non-
omnipotence” he offers in his “defence”—that he did what he could (99–104).
As we shall see, Matilal wants to show that Kṛṣṇa resorts to a special kind of
Archangelic rational argument.
Matilal’s interest in Kṛṣṇa in connection with the “search for a rational
basis for dharma” in India (2002, 74) is surprising, since Kṛṣṇa has more the
reputation of an amoral Machiavellian trickster who is always ready to explain
how his devious means justify ends such as victory that are also inscrutable
divine ends of his own. Moreover, when Matilal says that “God was never cited
as an authority on dharma” in India (idem), Kṛṣṇa looks like an exception,
22 dharma

particularly in the Bhagavad Gītā.32 But I think Matilal’s point can be well taken.
One finds the Gītā in a text with lots of other jostling teachings, where people
often disagree with Kṛṣṇa both directly and indirectly, and where what he says
in the Bhagavad Gītā is never cited as authoritative anywhere else, at least in the
Mahābhārata.33
Let us see how Matilal makes his case with his favorite instance, where
“Kṛṣṇa intervened and became ‘the Archangel’ ” (Matilal 2002, 27; cf. Ganeri
2005, 193–95). This is the convoluted war scene where Yudhiṣṭhira is dis-
traught that Arjuna seems to be dallying once again, just as he did when it
was a question of killing his guru Droṇa (see chapter 9 § E). This time
Yudhiṣṭhira wants Arjuna to finish the job of killing Karṇa (the two are still
yet to know that Karṇa is their eldest brother). Yudhiṣṭhira insults Arjuna’s
bow Gāṇḍīva, whereupon Arjuna, who had vowed to kill anyone who insulted
Gāṇḍīva, rushes at Yudhiṣṭhira sword raised prepared to kill him. Matilal dis-
cusses this case in several contexts34 to make the point that Kṛṣṇa helps
Arjuna resolve a “genuine moral dilemma” (Matilal 2002, 33). Kṛṣṇa says his
own view is that “not to slay living creatures trumps everything,” even truth
(Mbh 8.49.20), and asks Arjuna, “How, like just another uncultivated man,
could you wish to kill your elder brother, a king who knows dharma?” (21). He
says Arjuna’s vow is childish, that he has failed to take dharma’s subtlety into
account (23–24). Granted that “there is nothing higher than truth,” it can be
difficult to know it, and there are circumstances where truth is falsehood and
vice versa (27–29). To illustrate this, he tells Arjuna two stories, catching his
attention in the crosshairs between them to say that keeping to the truth of a
vow can be a foolish consistency.
First, a hunter named Balāka, “always devoted to his svadharma,35 truthful,
and never harboring envy” (35cd), used to hunt to take care of his son, wife,

32. As we shall see in chapter 12 § D, Śiva and Umā also turn to each other as authorities on dharma in
Mbh 13.
33. That Vedāntic schools make it one of their three authoritative texts that call for commentary is a later
matter.
34. See Matilal 2002, 9–11 (detailing a nineteenth-century argument over it by Bankimchandra Chatterjee
and Rabindranath Tagore), 26–32 (the main discussion, which can also be found in Matilal 1989, 9–16), 47 (on
the “situationality” of Kṛṣṇa’s dharma), and 193–94 (differentiating moral and religious truths). The ground-
breaking earlier study of ethical problems in the Mahābhārata is Strauss 1912.
35. On svadharma, see chapters 10 § C and 11 § B. Balāka was probably a Śūdra like the pious hunter in Mbh
3.196–206: an authority on svadharma who gets to instruct another (?) Kauśika who, as Smets notes (2005, 521), was
“befouled” by a she-crane (balākā)! Smets decides they are probably different Kauśikas because their learning dif-
fers: Kṛṣnạ ’s example is “not much learned” (na bahuśruta; Mbh 8.98.49.41b), while the other is explicitly a Veda-
reciter (vedādhyāyī; 3.197.1b) (2005, 521–22, 526–30). But this small difference could be contextual. Kauśika does not
lack learning in Bhīṣma’s brief description of him (12.110.8), and could be Kṛsṇ ạ ’s way of reminding Arjuna that he
(Arjuna) is not as learned as Yudhiṣṭhira. It may be pertinent that Kauśikas are Viśvāmitra Brahmins since
Viśvāmitra starts out a Kṣatriya and has sons of questionable learning (thanks to T. P. Mahadevan on this point).
 introduction 23

blind parents, and others dependent on him. One day, through a shot enabled
by his ability to detect an animal by the noise it made drinking water, he com-
mitted the “very cruel” (sudāruṇa) act of killing a beast such as he had never
seen before called Andha (the “Blind” one). Yet Balāka went to heaven: Andha
had gotten a boon from Brahmā enabling him to annihilate all creatures, and
although Brahmā had blinded Andha, Andha was “resolved” to actualize the
full potential of his boon (31, 33–40). Second, a Brahmin named Kauśika “had
taken a vow of telling the truth always. A group of gangsters, in hot pursuit of
some innocent men, came to ask him which way they had fled. . . . He told the
truth, and the men were chased and killed,” and Kauśika went to a horrible hell.36
As Matilal puts it, Balāka’s story sounds like a “case of moral [we might say
blind or dumb] luck. The goodness of a human life is not always dependent on
the things that the moral agent can control. . . . By contrast, Kauśika had a
project for life. . . . But according to Kṛṣṇa’s ethical system, he acted stupidly,
and chose the alternative that ruined his dream entirely” (2002, 29). Left, after
hearing these two stories, to think it through whether Yudhiṣṭhira should still
be killed (Mbh 8.49.56), Arjuna is remorseful and thankful for being rescued
from disaster. But he still needs Kṛṣṇa’s help in working out a solution, since
his vow must count for something. Always on top of things, Kṛṣṇa arranges for
Arjuna to “kill” Yudhiṣṭhira by addressing his elder brother (guru) in the
familiar, and to atone by “killing” himself with self-praise.37
Matilal calls “genuine” dilemmas like this “action-guide dilemmas,”
“which are not simply products of confusion, and for which, to be sure, there
are no clear-cut solutions.” They arise when “an agent cannot do everything
that is obligatory for him to do in that situation. He feels obliged to do, say, both
X and Y; but it is impossible to do both of them. For the situation is such that
doing X would be undoing Y, and vice versa” (Matilal 2002, 6; cf. 23). On the
present example, Matilal makes a number of rewarding points: that Kṛṣṇa’s
intervention “acknowledges progressive revision” (27),38 demands an “ordering
of priorities” (30), offers reparation for a promise not kept and the remorse for
not keeping it, and salvages what is recoverable from a conflict (32–33). But the
passage holds two more surprises that fit Matilal’s profile of the Archangel:
what it says about poets and about reason.

36. Mbh 8.49.41–47b, for brevity’s sake and because “gangsters” is nice for “cruel dasyus” (43c), quoting
Matilal 2002, 9.
37. See also Hiltebeitel 2001a, 206–7, 269–70; 1984, 24; Reich 1998, 231–45; Biardeau 2002, 2: 348–52,
356–58.
38. Cf. Ganeri 2005, 201, concluding, after discussing Kṛṣṇa’s and others’ roles in the killing of Droṇa,
that, “In the Mahābhārata we see directly a dynamic moral tradition employing its immersed critical principles
in a process of genuine development.”
24 dharma

First, about poets, the passage itself has not forgotten them. Arjuna could
make such a blunder, says Kṛṣṇa, only because he is unacquainted with the
settled opinion (niścayam) of those who have pursued dharma, the poets
(kavayaḥ; 8.49.16)! Nor has Matilal forgotten them in what he says about “two
different types of moral persons as paradigmatic” in their treatment of
dilemmas:

One is the dutiful fulfiller of moral obligations á la Kant. In India we

have Rāma whose moral ideas would fall into this category. The
nature of dharma idealized by Rāma (or Yudhiṣṭhira) seems to have
been very rigid. It seldom bends. The other paradigmatic person we
meet in the moral field can be described as the imaginative poet. He
becomes a perspectivist and understands the contingency of the
human situation. He realizes the necessity for “paradigm shifts”
much like the revolutionary scientists in Thomas Kuhn’s description
of the nature of scientific revolutions. He looks at the particular
situation but also beyond it. He is our Kṛṣṇa. (Matilal 2002, 34)

I believe Matilal’s contrast holds in what it says about Rāma and about linking
Kṛṣṇa with the imaginative poets, for as we shall have numerous occasions to
see, Kṛṣṇa and Vyāsa are Archangelic together. Indeed, we can say there is no
moral distance or tension whatsoever between Kṛṣṇa and Vyāsa such as there
is between Rāma and the Rāmāyaṇa poet Vālmīki. But I believe the parenthesis
on Yudhiṣṭhira is a monkey wrench.39 Rāma gets to fulfill his moral obligations
in the ideal poetic universe crafted for him by Vālmīki to sustain the very image
of Rāma’s perfection, yet also, I urge, to hold it up to question. Yudhiṣṭhira, on
the other hand, must seek to work his way along in a darker world where his
action guides come with intermittent interventions by a variety of Archangelic
agents, including not only Kṛṣṇa and the poet-author (and indeed, more often
than not, the latter) but his father Dharma, other gods and Ṛṣis, and his other
“grandfather” Bhiṣma. As we will see in chapter 9, Yudhiṣṭhira must weigh
what both Kṛṣṇa and Vyāsa have to say in facing his dilemma over truth and
noncruelty in the slaying of Droṇa.
Second, about reason, here is what Kṛṣṇa says Arjuna should draw from
the two stories he has told him:

There will be some indication for you. (Sometimes) supreme

knowledge is difficult to specify. In such cases one resolves (matters)

39. Cf. Ganeri 2005, 192, like Matilal viewing Yudhiṣṭhira as “the inflexible rule follower,” and reading
this dubious understanding into Yudhiṣṭhira’s stance in the killing of Droṇa (see further 2007).
 introduction 25

by reasoning (duṣkaram paramajñānaṃ tarkeṇātra vyavasyati). Surely

many people say, “Śruti is dharma.” I have no trouble with that, but
not everything is enjoined. Promulgation of dharma is done for the
sake of the flourishing (prabhava) of beings. (8.49.47c–49)40

Where Kṛṣṇa says “not everything is enjoined (na hi sarvaṃ vidhīyate),” we can
take him to be saying “not everything has a rule (vidhi).” That is, Kṛṣṇa uses
the verbal root vi-÷ dhā, from which comes the noun vidhi, whose main early
meaning is “rule,” “injunction.” As early as chapter 3, we shall begin to note
this verb and its derivatives, and the associations they often have with the root
÷ dhṛ and its derivatives, including dhárman, dharma, etc. What Kṛṣṇa means
is that where there is no rule, such cases call for reason if dharma is to
flourish.
As far as I can see, leaving out an interpolation cited in chapter 9 § D
where Yudhiṣṭhira is answering “The Yakṣa’s Questions,” no one else speaks
about reason (tarka) in the Sanskrit epics except for the aged Bhīṣma, who, in
most cases, says tarka is to be shunned as a useless science (vidyām . . . nirārthikām;
12.173.45cd, 13.37.12ef) associated with critical and technical “philosophical”
terms like ānvīkṣiki, “investigative science” (12.173.45c; 13.37.12c); hetuvāda,
“causal argumentation” (12.173.46a, 13.37.13ab); nāstikya, “heresy” (12.173.47a);
and tarkaśāstra, “treatises on reason” (12.238.17cd, 12.261.39–40).41 But when
Bhīṣma mentions tarka in rehashing the same two stories (at 12.110.4–15;
see Fitzgerald 2004a, 760), he is, like Kṛṣṇa, talking about moral rather than
philosophical reasoning.42 Kṛṣṇa’s moral reasoning is often challenged, but in
Vyāsa’s Mahābhārata he usually gets the benefit of the doubt.43 And this is so
not only where it concerns conflict between individuals but where it concerns
the welfare of the worlds. Even if Kṛṣṇa advises one “crooked means” (jihmopāya)
after another to keep a bad situation from getting worse, and even if “all that
Kṛṣṇa was able to do was to salvage justice at the end of the battle” (Matilal
2002, 105)—and of course even that is questionable—he and the poet can be
said to have moved things along.

40. Mbh 8.49.48–49 has much in common with 12.110.9–10b. The last lines of these segments about
“reason” (tarka) are identical. Line 8.49.48ab has an alternate initial pāda in the Southern Recension, duṣkaraṃ
pratisaṃkhyānam, which seems to be a partial backreading from 12.110.9cd where Bhı̄ṣma’s shorter account
begins the line similarly with the same ending. Cf. P. L. Vaidya’s paraphrase translation and rephrasing of these
verses’ import in his notes to the Pune Critical Edition (1954, 690).
41. See Halbfass 1988, 273–86, notably 278–79 on Mbh usages, and Rām 2.94.32–33: advice of Rāma in
the same vein.
42. There can, of course, also be dubious moral reasoning, as when Upayāja, a priest, tells King Drupada
that while he won’t perform a homicidal sacrifice for Drupada to kill Droṇa; his unscrupulous and venal older
brother Yāja will agree to do it using the “eye of reason” (tarkacakṣuṣā; 1.155.19b)—background to the sacrifice
that produces Dhṛṣṭadyumna and Draupadī.
43. Duryodhana being an exception worth mentioning.
26 dharma

Matilal thus develops distinctions that are worth pursuing. To Yudhiṣṭhira’s

benefit and in his hearing, Kṛsn ̣ ạ can get Arjuna to appreciate that there is not a
rule for everything. On the other hand, in aspiring to rational consistency in
applying dharma as rules to situations that may seem incompatible, but for
which there must be no dilemma in reality, Rāma is more like Manu, which, as
Donald Davis has argued (2007b), probably introduces its fourth source of
dharma, ātmatuṣṭi, to resolve situations that the other three sources do not clarify
(see chapter 5). Yudhiṣṭhira must make something out of living in Vyāsa and
̣ ạ ’s world that Rāma cannot make out of living in Vālmīki’s. Or, more pre-
Kṛsn
cisely, where Vālmīki leaves us only to imagine what Rāma might actually make,
once he hears it, out of the poetic universe created for him, we can see Yudhiṣṭhira
interacting with both the poet and Kṛsn ̣ ạ . Here are three byplays between
Yudhiṣṭhira and Kṛsn ̣ ạ that show Yudhiṣṭhira’s attunement and participation in
a paradigm shift he thinks he can believe in. The first occurs just before the war
begins, the second in its middle, and the third as it is about to end.
First, there are two “abstract deities,” the Placer (Dhātṛ) and the Ordainer
(Vidhātṛ), whom we will be tracking through several chapters, whose names
are derived from the verbal roots ÷ dhā and vi-÷ dhā, respectively, the latter verb
having been just mentioned. As we shall see by chapter 10, there are some con-
texts in which these appellations may refer implicitly, without naming him
directly, to Kṛṣṇa. Yudhiṣṭhira, however, is the first person in the main narra-
tive to give Kṛṣṇa these names directly, calling him both Dhātṛ and Vidhātṛ,44
and in a decisive context—that of determining who should marshal the Pāṇḍava
armies in the imminent war. After hearing several names proposed by his
brothers, Yudhiṣṭhira says,

When it comes to this whole universe, lads, Keśava, the self of

dharma (dharmātmā), knows all: the relative substance, relative
strength, what has occurred and what is to be sought. Let him be the
marshal of our armies whom Kṛṣṇa Dāśārha45 calls, whether he is
skilled in weapons or not, or if he is old or young. He is our root in
triumph (vijaye mūlam), lads, as well as in reversal (viparyaye).

44. Kṛṣṇa says he is Dhātṛ in the BhG. Cf. 12.60.6, where, after Kṛṣṇa gives Bhīṣma the “divine eye” with
which to see everything pertaining to dharma and artha with a constantly lucid mind (sattvasthaṃ ca mano nityam)
(12.52.15–21), Bhīṣma begins his postwar sermon to Yudhiṣṭhira bowing to “Kṛṣṇa who ordains (namaḥ kṛṣṇāya
vedhase).” Vedhas (probably from vi-÷dhā) is another word for “Ordainer” that is probably cognate with Dhātṛ and
especially Vidhātṛ: “an arranger, disposer, creator (esp. applied to Brahmā, but also to Prajāpati, Viṣṇu, Śiva,
Dharma, the Sun, & co.), Mbh” (MW [1899] 1964, 1018); as a Ṛgvedic usage, also “an august ritual title” (Jamison
1996, 80). See Fitzgerald 2004a, 147–48 and n. 256; 312, 315, proposing that the Rājadharma could originally
have begun here, though, as he admits, it must be read as a resumption.
45. The name means “worthy of respect” (Biardeau 2002, I: 105 n. 23).
 introduction 27

On him rest our lives, kingship, being and nonbeing, happiness and
misery. He is the Placer and Ordainer, and success is founded there
(eṣa dhātā vidhātā ca siddhir atra pratiṣṭhitā). He whom Kṛṣṇa Dāśārha
calls fit shall be the marshal of our armies. Let the best of speakers
speak—the night is running out. (5.149.33–36)

Getting nothing specific in reply from Kṛṣṇa, Yudhiṣṭhira soon takes the
initiative himself and appoints Draupadī’s twin brother Dhṛṣṭadyumna, “who
had been born from the fire to be the death of Droṇa” (154.12cd). I give the flat
translation of dharmātmā as “self of dharma” to keep the focus on what inter-
ests Yudhiṣṭhira, who has earlier recognized Kṛṣṇa as “lord of dharma”
(dharmeśvara) and master of policy (5.28.9; cf. 2.14.9; 18.3). Yudhiṣṭhira makes
a correlation between dharma and Kṛṣṇa’s knowing ability to “set in place” and
“ordain” what will bring success in war. But Kṛṣṇa typically leaves the matter to
Yudhiṣṭhira to decide on his own.
Second, upon the fall of Bhīṣma, Kṛṣṇa congratulates Yudhiṣṭhira, saying
he must have won “by good luck” (diṣṭyā), or perhaps by destiny (daivataiḥ), or
perhaps “having gotten you who can kill with a look (as his foe) he was burnt
by your wrathful eye” (6.115.61–62).

Thus addressed, Dharmarāja answered Janārdana, “Victory is by your

grace, defeat by your wrath. You are surely our refuge, Kṛṣṇa, who
brings fearlessness to bhaktas. Unmarvellous is victory (anāścaryo
jayas) for those whom you, Keśava, always protect in battle and are
always also engaged in their welfare. Having betaken ourselves to you
in every way, there is no marvel (āścaryam). That is my opinion.”
Thus addressed, Janārdana answered, smiling: “Only in you is this
word fit,46 best of kings.” (63–65)

Kṛṣṇa appreciates that, more often than anyone else, Yudhiṣṭhira will see his
marvels as unmarvellous. In this, as we shall see in chapters 12 and 13,
Yudhiṣṭhira as a bhakta knows enough to “see things as they are” in the fashion
of certain Ṛṣis.
Third, upon the fall of the Kaurava king Duryodhana, Yudhiṣṭhira shows
that he can explain himself to Kṛṣṇa in Kṛṣṇa’s own terms. Bhīma, the second
oldest Pāṇḍava born between Yudhiṣṭhira and Arjuna, has downed Duryodhana
by foul means—a “below-the-belt” thigh-smashing strike with his mace or club
that Kṛṣṇa signalled to him to do, through Arjuna. But Bhīma goes further and
crushes Duryodhana’s head with his left foot (9.58.12). This outrages Kṛṣṇa’s

46. Yuktarūpam; proper, suitably formed.

28 dharma

brother Baladeva, who had trained both combatants. But Baladeva protests only
the low blow. Kṛṣṇa makes his own excuses for this foul play that do not con-
vince Baladeva, who leaves the scene in outrage. Nor do they convince the bard
Saṃjaya, who is recounting the story. Saṃjaya calls Kṛṣṇa’s justification a
“fraudulent transgression of dharma (dharmacchalam)” (9.59.22a). The bard’s
comments would seem to set up Kṛṣṇa’s next words, which are to ask
Yudhiṣṭhira how he can excuse Bhīma’s further transgression of trampling his
victim’s head:

Kṛṣṇa said, “To what end, Dharmarāja, do you sanction adharma

when the head of the insensible and fallen Duryodhana, whose
kinsmen are slain, is crushed by Bhīma with his foot? Conversant
with dharma, why do you look on with indifference (upaprekṣasi),
king?” Yudhiṣṭhira said, “It does not please me, Kṛṣṇa, that Wolfbelly
touched a king on the head with his foot out of wrath, and I do not
delight in this destruction of the family. By guile were we always
deceived by the sons of Dhṛtarāṣṭra, who spoke many cruel words.
Indeed, we were exiled to the forest. That grief of Bhīmasena turns
exceedingly in his heart. Reflecting so, Vārṣṇeya, I overlook it.
Therefore, having slain that one bereft of wisdom, covetous, under
the sway of desire (kāmavaśānugam), let the Pāṇḍava gratify his
desire (kāma), be it dharma or adharma!” After Dharmarāja had
spoken, Vāsudeva, that perpetuator of the Yadu family, said with
difficulty (kṛcchrād), “At any rate (or, as desired), let it be so (kāmam
astvevam iti)!” (9.59.29–35)

As we will see, moral choices can be difficult even for Kṛṣṇa. But once again,
the important point is that they can move on. As Kṛṣṇa knows, Yudhiṣṭhira had
decried the crushing of Duryodhana’s head in no uncertain terms. The exchange
changes the mood from gloom at Baladeva’s departure to Bhīma’s wide-eyed
joy at being let off the hook (38). And the adhyāya or “chapter” ends with
Yudhiṣṭhira extolling the “good luck” of having had Kṛṣṇa’s counsel with which
to conquer the earth (83–84).
One would not hear such a relaxation of dharma from Rāma, and indeed,
there was no one in the Rāmāyaṇa like Kṛṣṇa to ask him for one, or to confirm
it. To my mind, these considerations put Yudhiṣṭhira in both camps. Like
Rāma, he is “like one of those philosophers who admit the facticity of moral
dilemmas and insist that our commitment to consistency would require us to
modify the system by reordering priorities or by discarding certain principles”
(Matilal 2002, 27). Indeed, Matilal is citing the philosopher R. M. Hare’s mod-
ification (1986) of the moral precept “do not lie” into “do not lie except to the
 introduction 29

enemy in time of war,” which is precisely what Kṛṣṇa will advise Yudhiṣṭhira to
do between the fellings of Bhīṣma and Duryodhana to bring about the death of
Droṇa “by not very glorious means” (Matilal 2002, 66; cf. 46, 87–88, 95–99).
But Yudhiṣṭhira is also party to Hare’s “higher type of critical thinking,” which,
in Matilal’s terms, still drawing on Hare, would be the side of the intervening
“Archangel” (2002, 26). Indeed, as we will see in chapter 9, when Yudhiṣṭhira
lies to Droṇa at Kṛṣṇa’s prompting, he bends at a point where Rāma’s doing so
would be more or less inconceivable. Both types of moral reasoning can accom-
modate dharma to bhakti and vice versa: in Rāma’s case, as a figure of perfec-
tion worthy not only of the highest “loyalty” but, for those who sense his divinity,
“devotion”; in Yudhiṣṭhira’s, as one who senses his own good fortune in hav-
ing Kṛṣṇa among those who keep appearing to help him see dharma with a
fresh twist.

D. Chapter by Chapter

Matilal’s notion of a paradigm shift around Kṛsn ̣ ạ is, then, one of the ways we
can begin to think about the Mahābhārata as introducing change over time with
regard to dharma. As I have mentioned, it is consonant with another: Biardeau’s
idea of “a bhakti swerve.” From chapters 5 through 12, we shall be opening up
these and other ways through which to appreciate that projecting change in
dharma over time is a large and complex initiative of the Mahābhārata, and also
part of a larger intertextual project that includes initiatives by other texts, but
also, I will argue, involves the Mahābhārata’s response to some of them, and
these other texts’ responses to the Mahābhārata—including, moving on to
chapter 13, the response by the Buddhist poet Aśvaghoṣa in his Buddhacarita.
If, however, the Mahābhārata will be the pivotal text spread over the center
of this book, the Aśokan edicts are pivotal for it in another sense by marking the
historical watershed at which the book begins in chapter 2. We start our
discussion of dharma texts over time by observing that—whatever may have
preceded them in Brahmanical and Buddhist circles—an intertextal history of
Buddhist and Brahmanical dharma texts in their classical period must effec-
tively begin with the Aśokan edicts and what they record of the dhaṃma
campaign that King Aśoka sought to implement through them.47 Yet even as
we let the edicts serve primarily as an historical watershed, they alert us to

47. I thank Dan Rudmann, who, in reading an early draft of Hiltebeitel 2010c as an MA candidate after
auditing the first Freshman Seminar I gave on “Dharma in Buddhism and Hinduism” in 2006, had the idea,
which I adopted, of treating Aśoka as a watershed figure.
30 dharma

some of the ways other texts will thematize change in dharma over time: dysto-
pias and utopias; the role of the king; dharma over dynastic time; dharma over
cosmic time; dharma and meditation; dharma and the biographical; indeed,
quite uniquely, the autobiographical.
Chapters 3 and 4 then take us to the far side of the Aśokan watershed: to
the Brahmanical understandings of dharma that preceded it, in chapter 3; and
to the Buddhist understandings of dhamma and/or dharma that likely preceded,
accompanied, and followed up on it in chapter 4. In chapter 3, we pursue the
implications of dharma, in the form dhárman, having begun in the Ṛgveda as a
new concept; the use of this concept to generate novel enigmas from the earliest
to the latest texts in the Vedic canon; its centrality to changing notions of king-
ship in those texts; and, no less important, to changing notions of the Brahmin.
And in chapter 4, we take up Buddhist understandings of dharma as they were
developed in (or better, into) the three baskets of the early Buddhist canon, tak-
ing their extant collections to have some historical implications themselves for
understanding how Buddhists formulated the Buddha’s teachings over time.
Here, of course, our watershed metaphor will no longer be serving us so well.
Rather than a feature of the South Asian landscape from which water flows in
two separate directions, let us convert the Aśokan watershed into one of those
South Asian imperial projects, an anicut or great dam, toward which numerous
Vedic and Buddhist tributaries must have flowed, gotten dammed up for a
while and forced to intermingle, and from which they then overflowed, each
with new vigor, as the two communities took the potent waters into new flows
and channels of their own further devising.
Gathering force within this current, we then meet in chapter 5 what will be
called a vast Brahmanical society in mutation, codifying and debating its cus-
toms and traditions; providing an all-encompassing cosmological setting for
treating “caste and life-stage” in the context of past, present, and future; giving
explanatory power to dharma over time in relation to the post-cosmogonic mix-
ing of peoples, including Greeks, that follows from creation of the original four
castes; and integrating such subject matter into new poetic genres that we call
epic and dharmaśāstra, which in the case of the Mahābhārata and the
Mānavadharmaśāstra or Laws of Manu include some advice about how a king
should start up a new kingdom and should lead his royal day.
That brings us to what I have called the thematic pivot of this book in
chapters 6 and 7, which take up the theme of dharma over time directly, and
also, in both chapters, comparatively on usages where Buddhist and
Brahmanical texts worked these matters out in significantly different ways.
Chapter 6 will be about how they depicted dharma in relation to the big cos-
mological time units, beginning with the kalpa—in this case, probably in full
 introduction 31

awareness of what each other were doing. Chapter 7 will then be about what I
have called two minor dharma texts, ones that were probably known only
within the two respective traditions, each of which, impacted by much the
same geography and history, prophesied the end result of dharma’s change
over time in a future disaster. These two chapters open new windows on the
two epics, and especially on the Mahābhārata, which remain under discussion
through all remaining chapters.
Chapter 8 will then launch a four-chapter exploration of the epics by
exploring the topic of dharma over time through the three generations of dynastic
instability that precede the main generation in the Mahābhārata. Set loosely in
the armature of a twilight between ages or yugas, this three-generational buildup
makes women’s dharma central to an exploration of the nuances of intersecting
dimensions of time (authorial time, intimations of an overarching divine plan,
maternal time, spousal time, generational time, historical time, time and rivers,
time running out) in which dharmic norms come under repeated challenge and
scrutiny.
Chapter 9 will then look at the question of moral biography in the king’s
dharma as both epics portray their central royal personages of Rāma and
Yudhiṣṭhira: how each epic constructs it differently with regard to instruction
in dharmic precedent, questioning and ambiguity versus moral certainty, and
personal accountability. Building on hints at biography noticed in the Aśokan
edicts (chapter 2) and in portrayals of the Buddha (chapter 4), and also in
Manu’s outline for how a king should lead his day (chapter 5), this chapter
anticipates the critical treatment of the epics in Aśvaghoṣa’s Buddhacarita
(chapter 13) and also builds on the treatment of women’s dharma in chapter 8,
to anticipate further discussion in chapter 10 of such topics as the question of
whether the epics also portray women’s dharma biographically, and the epics’
divine plans in chapters 10 through 12.
With chapter 10, we come back to women’s time: now depicted not so
much through the weave and flow of maternal generations but with respect
to its experience by each epic’s chief heroine. We find that while biographical
time is left more as the province of their royal husbands, the time of the her-
oine queen is episodic. It revolves initially around brief and indirect glimpses
of her passions and wit with respect to scenes of birth, youth, and marriage,
and then zeroes in on her handling of each epic’s central time of crisis, which
involves specifically her violation in a way that affected no woman in the gen-
erations before her. Focusing on Draupadī and Sītā as “legal wives” brings
into relief what each makes of her worst situation, where a woman’s dharma
implies playing for time, biding time while waiting for a husband to get his
act together, and dark thoughts on the fruition of karma. In Sītā’s case, this
32 dharma

includes meditations on the relief she might find in death; in Draupadī’s, it

brings out her flashes of anger and impatience. In both cases, the heroine
intimates a perception of being the pawn and victim of a divine plan beyond
her own devising.
Chapter 11 on dharma in the Bhagavad Gītā then explores how this text,
which from so many angles may be considered to lie at the center of the
Mahābhārata, puts dharma at the center of a vast vision of the workings of time
itself. What, taking an implausible time out in the middle of the battlefield just
before the outbreak of war, did God say to the world’s greatest warrior, who, for
a moment, thought better of being a killer? We approach the Gītā through what
the larger epic shows are its ripple effects in the way the relation between
dharma and time is depicted inward and outward from the Gītā itself. In the
Gītā’s own time, we have a text that takes us from Kṛṣṇa’s revelations about
Time, kalpas, yugas, and the divine plan to his instructions on living dharma
over ordinary time experientally, allowing that Arjuna will need time to digest
what he has to say about fulfilling one’s dharma in a supremely difficult time in
a way that allows one to transcend it.
Chapter 12 then takes us into an attempt to “map” dharma in relation to
bhakti, proposing that such a map must begin with the divine plan (or plans)
one has been hearing about. It becomes a project of mapping the Mahābhārata
in relation to three other texts—not only the Rāmāyaṇa and Manu but also the
Harivaṃśa—as something that relates dharma perhaps first and foremost to
time, and, with it, to notions of yugadharma and avataraṇa. These complex
concepts take in different prophesies about the Kali yuga and themes of divine
“descent,” including the “avatāras” of Viṣṇu, among whom Kalki can end the
Kali yuga, and descents of the goddess Gaṅgā. With Gaṅgā, it also takes in
dynastic descent through generations into a carefully targeted dharmic geog-
raphy. By attending further to how dharma and bhakti interrelate in the lives
of Rāma and Kṛṣṇa on this very terrain, we map dharma and bhakti through
the lifelong interpersonal themes of friendship, hospitality, and ultimate
separation.
Finally, chapter 13 on Aśvaghoṣa’s Buddhacarita brings us back to the
question of dharma in biography. Not only does Aśvaghoṣa construct a biography
of the Buddha from earlier Buddhist sources, he does so taking full cognizance
of the question of precedent raised in the Sanskrit epics. But he is also concerned
to show the uniqueness of the Buddhist dharma and of the Buddha’s discovery of
it. In the context where a Buddha’s greatness lies in rediscovering the dharma
entirely on his own, Aśvaghoṣa portrays how Prince Siddhārtha experiences the
“three signs” of old age, illness, and death, each “for the first time,” as raising for
him the question of an underlying “law” or dharma. Where a Buddha is concerned,
 introduction 33

epic precedent on dharma, subtle, complex, and fascinating as it may be, has no
authority. Aśvaghoṣa undercuts that authority with his marked emphasis that a
younger person in a lineage can outdo an older epigone in achievement. Most
central is the repeated insistence that “there is no wrong time for dharma,” which
provides opportunities for Prince Siddhārtha to trump Brahmanical concerns
for the inherent timeliness of āśramadharma as staggered across a lifetime.
Aśvaghoṣa’s text affords an overview that takes us back to the earlier Buddhist
canon, to Aśoka via a closing tribute, and into the more recent times of the most
high-impact classical Brahmanical dharma texts—the two epics and Manu—all
three of which Aśvaghoṣa seems to know, and in the case of the epics, definitely
savors as poems worthy of engaging for his Buddhist critique.
This page intentionally left blank
2
Aśoka Maurya

To begin with Aśoka Maurya is to begin what might be called the

watershed figure in the history of dharma, and also with a distinctive
set of texts, his rock and pillar edicts, that allow us to begin with what
we might call texts on the ground before we move on to works of
literature. In his “brief edicts,” Aśoka uses the term dhaṃma “about
111 times,” a number that “stands in sharp contrast” to what one
finds “in the much more vast literature of the middle and late vedic
periods” that precede him (Olivelle 2004a, 505).
It is not likely that many could have read the Aśokan inscriptions
even in Aśoka’s own time, since they are among the first-known Indic
texts to use alphabetic writing.1 Moreover, it seems that as other scripts
came into use, Aśoka’s Brāhmī inscriptions were readable for only a
few centuries. By the time Sumudragupta (ruling from ca. 335–376 CE)
had extended the early Gupta empire, his postmortem eulogist seems
to have associated the Aśokan pillar at Allahabad (perhaps originally at
Kausambi) with imperial sway by having it co-inscribed in Sanskrit
with a triumphalist message much opposite the earlier one engraved
on it by Aśoka. This eulogy (praśāsti) describes Samudragupta’s march
of conquest over kings and chieftains across most of north India and
down deep into the south, to Kanchipuram, even though the Guptas
probably never had direct rule of much more than the northern
Ganges plain (see Bloch 1950, 26–27; Thapar 2002, 283–84). The
eulogist would probably not have been able to actually read the Prākrit
of Aśoka’s inscription or its Mauryan Brāhmī script. If the Chinese

1. Nikam and McKeon 1978, 16, say that they were “probably intelligible to the common people
throughout the empire”—though perhaps not directly in writing.
36 dharma

Buddhist pilgrims to India, Fa-hsien (399–414) and Hsüan-tsang (629–45), give

a near-contemporary indication, he would, like them, have probably had access
only to oral pseudo-translations that had no correspondence with what the
Prākrit actually said. The Chinese pilgrims describe “readings” derived from
legends about Aśoka recounted in Buddhist Sanskrit texts like the Aśokāvadāna,
which describes Aśoka’s imperial conquests in a fashion similar to those now
recalled of Samudragupta. By the fourteenth century it was forgotten that the
pillars themselves had anything to do with Aśoka. It was not until soon after the
Brāhmī script was deciphered in 1837 by James Prinsep that the inscriptions
could again be read and Aśoka’s dhaṃma experiment be studied (see Strong
1983, 5–15). Yet it remained conjectural until 1923, when the Maski inscription
was found mentioning his name as “Asoka,” that the edicts were indeed his
(Bloch 1950, 17–18, 145).

A. Aśoka’s Inscriptions

Aśoka was not just a king but an emperor, and his edicts, which he sometimes
called dhaṃmalipi (“dhaṃma texts” or “edicts on dhaṃma”),2 were means to
broadcast an imperial program. He uses the term dhaṃma with some specific
connotations. It is likely that Aśoka was aware that the term carried older
Brahmanical implications of royal authority (see chapter 3). Yet he is perfectly
clear in some inscriptions that he associates it with Buddhism in its sense of
referring to the Buddha’s teachings, which had made him a convert. Even
where the edicts do not refer to Buddhism, it is impossible to overlook the
term’s Buddhist resonances in the name of some secular or Brahmanical read-
ing.3 We also get some hint of the range of meanings that Aśoka imputed to
dhaṃma from a trilingual rock inscription (Kandahar 1), where the Prākrit
dhaṃma is given Greek and Aramaic counterparts: eusebia, “piety, respect for

2. In providing translations from the edicts, I consult mainly Bloch 1950, who, for this compound, gives
“text de la Loi” (90, etc.) and Nikam and McKeon 1978, who give “edict on Dharma” (57, etc.). Bloch notes that
“lipi designates the inscription in its materiality: the edict itself would be śāsana, ‘proclamation’ ” (my transla-
tion), and says lipi is an Indian adaptation, from the root lip, “to smear,” and perhaps likh, “to write,” of Iranian
dipi, which is conserved in edicts from the northwest (90 n. 2). I cite Nikam and McKeon for its accessibility, but
with the caution that they tend to Sanskritize—as can be seen from the form dharma and in many of their locu-
tions and paraphrases. For instance, “to work for the promotion and to prevent the decline of Dharma” (Nikam
and McKeon 1978, 32) can even be said to Hinduize what Bloch translates more literally as “pour qu’un s’attache
au progrès sur ce point et qu’on n’admette aucune défaillance” (Bloch 1950, 101). I soon discuss this passage
from RE 4, and will return to it in chapter 6.
3. I part company from the secularizing slant of Bowles 2007, 130 (“a ‘secular’ reflection of the many
movements which had arisen in the previous couple of centuries”), and to some extent from Thapar 1997, 309
(“His ideas on dhamma borrow from the current debate”). On Brahmanical readings, see, for example, n. 2.
 aśoka maurya 37

gods, kings, and parents,” and qsyt, “truth,” respectively4 (see Olivelle 2004a,
509 n. 26; Falk 2006a, 242–43). Yet these translations are probably most inter-
esting as attempts to find terms that would resonate with speakers of these
languages in the Kandahar area (in today’s Afghanistan), where these languages
were in use, and probably cannot be taken to supply the deepest colorings that
dhaṃma would have had for Aśoka in Prākrit.
The inscriptions are written mainly in two scripts: Brāhmī over most of his
Indian empire; and, in the northwest, Kharoṣṭhī. According to Nikam and
McKeon (1978, 15–16), they convey a “crude” Prākrit that was perhaps close to
Aśoka’s own words. In any case, in rock edict (RE) 8, Aśoka says that, as of his
tenth year, he himself now goes on dhaṃma-tours (dhaṃmayāttā), giving
audience in the provinces (janapadas), preaching on dhaṃma and discussing
questions on dhaṃma (dhaṃmaparipucchā), and that the pleasure he derives
from this is “a second revenue”(Bloch 1950, 112–13). Aśoka begins several of
the rock edicts (RE 5, 6, 9, 11) and all of the pillar edicts (PE) with a kind of oral
formula: “King Priyadassi, beloved of the gods, speaks thus (evaṃ āha).”5 But
then he often ends on the note that the inscription is written (likhitā), in some
cases declaring his intention that it will last a long time (RE 5, PE 7). Clearly
Aśoka felt it was important to set what he had to say about dhaṃma in stone.
But we also can detect a hope that his reign would be credited with implement-
ing change in dhaṃma over time.
Aśoka’s program in edicts can be traced through his reign.6 They have not
only served scholars to compose a biography for Aśoka; they can be read as
autobiography.7 Such matters are nowhere more revealing than in the first
edict he promulgated. Minor Rock Edict 1 (MRE 1) was composed in Aśoka’s
“10th regnal year and inscribed from then onwards” (Falk 2006b, 55; cf.
Anderson 1990). Thanks to the vivid visual documentation and compelling
interpretation of Harry Falk (2006b),8 we are now able to consider it in both a
biographical and geographical context.

4. See Olivelle 2004a, 509 n. 26: “A tantalizing possibility is that Aśoka’s use of dharma may, in fact, have
been influenced by Hellenistic or Persian royal vocabularies.” Cf. Thapar 1997, 281–82.
5. PE 7, a kind of summa, breaks up its long text with ten uses of this oral formula or a variant with āha
(“he said”).
6. I follow Bloch’s (1950, 18–19) and Lamotte’s (1988, 226–27) chronology, for which Nikam and McKeon
(1978, 17–18) give dates correspondingly two years earlier.
7. Cf. Visvanathan 2010, 2: “Here, time sweeps into the records in many ways. It occurs in the references
to his own religious trajectory; in the repeated usages of regnal years; in the appeals to supervisors and teachers
to instruct their pupils in accordance with the old ways (porānāpakiti). . . .”
8. Cakrabarti 2011 demurs at “Falk’s idea that the Samajas of the Asokan time provided occasions for
‘orgies’” (2). He prefers to find “geopolitical aspects of the locations” (10), mainly route alignments, which can be
uncertain (17, 30). He also questions differentiating MRE 1 from MRE 2 in southern edicts (35), but Falk looks solid
on this.
38 dharma

At most of MRE 1’s sixteen known locations it was inscribed in caves or on

hilltops or at other places “generally far removed from habitation sites” but in
“the core of the Mauryan realm” (Falk 2006b, 55). At one hilltop spot in
Karnataka the inscription is atop a nearly inaccessible boulder; at another it is
under a dramatically perched rock overhang (72–73; 86–87). “Some of the
MRE sites on hills and mountains are so impressive by their nature or their
beautiful surroundings or both that their sanctity must go back to a time much
earlier than Aśoka or even the Buddha” (56). Falk even suggests, “Considering
that the text provides the first historical evidence in Brāhmī characters one
wonders if there is a reason for their being kept away from a possible reading
public” (2006b, 55). Falk demonstrates that MRE 1 is “of a completely differ-
ent nature” from the so-called MRE 2, which accompanies and enlarges MRE
1 at only five of its six southern locations, but not at its northern ones, and
never occurs on its own. He outlines twenty rules made by MRE 2 that call on
Aśoka’s dhaṃma bureaucracy to promulgate them (2006b, 57). This bureau-
cracy is announced only in subsequent Rock Edicts (RE), guaranteeing an
interim of several years between the initial promulgation of MRE 1 and the
affixation of MRE 2 to it at southern sites.
As Falk presents it, MRE 1’s “basic text” is as obscure as its locations:

King Devanampiya speaks thus:

For two and a half years I was a (Buddhist) layman and I was not
very zealous. For somewhat more than a year I visited the saṅgha and
became very zealous. The gods have formerly not mingled with men,
but now they are mingled. This is the result of zeal. Now not only
high persons can reach this aim, no, even common people can reach
it if they are zealous. (Falk 2006b, 55)

Note that MRE 1’s basic text does not mention dhaṃma.9 As a Buddhist layman
of increasing zeal, Aśoka would certainly have become familiar with the term.
But he does not make it his key term until the rock edicts. Rather, if MRE 2
follows up MRE 1 by conveying some “rudiments of dhamma” (Thapar 1997,
273), MRE 1 is almost surreptitious in getting that point across. What can it
mean that gods who “have formerly not mingled with men” now do so, and
that “not only high persons can reach this aim, no, even common people can
reach it if they are zealous”? He can hardly be talking about gaining access to
divine realms through Buddhist meditation practices. Just two years later, in

9. There is an exception in the version of MRE 1 at Gurjarrā in Madhya Pradesh, where dhaṃma occurs
in a final sentence not found elsewhere; see Olivelle 2004a, 509 n. 16; Falk 2006b, 77.
 aśoka maurya 39

his twelfth regnal year Aśoka will say that he promulgates dhaṃma by
“summoning people to tableaus of celestial palaces, tableaus of elephants, balls
of fire, and other divine displays” (RE 4). In both texts, Aśoka seems willing to
manipulate images of divinity much as Kauṭilya advises kings to do in the
Arthaśāstra,10 and to have a regal view of the gullibility of his populace.
As Falk remarks, “Although the text comes in plain words it is difficult to
imagine what it is all about. Reduced to its essence it seems to propose that
everyone become a Buddhist layman, develop zeal and thus mingle with the gods.
What sense does it make in a hidden place to read or be told how to mingle with
the gods?” (2006b, 55). Falk argues that “some of its meaning” emerges from
investigating “the places chosen for spreading this message” (2006b, 55).11 The
background for this is Falk’s initial tour of MRE sites in Karnataka with the eth-
nographer Günther-Dietz Sontheimer, on which Sontheimer, four weeks before
passing away, acquainted Falk “with his idea that the places . . . could have been
connected with mother cults in antiquity, very much in the fashion of practices
still current in this area today. . . . These two weeks with him are unforgettably on
my mind. It took years before the accumulated evidence forced me to accept his
interpretation of the sites” (Falk 2006b, 7). Of the 16 known MRE 1 sites, melās or
yātrās occur at nine of them today, and at 3 or 4 others there is evidence of such
festivals having been discontinued. Yet, “If there is truly continuity in practices
and places, then these yātrās were hardly of a Buddhist nature more than 2000
years ago. This . . . must mean that Aśoka had his Buddhist-minded MREs
enscribed at places where they would be seen by large crowds following practices
different from his own” (56). Aśoka called such assemblages samājas, “and tells us
in his first RE that samājas are not to be praised—because animals are killed there.
Such killings no doubt served the needs of a deity at the festival site, just as they

10. See Doniger 2009a, 202–3: Kauṭilya “advises the king to go out in public in the company of several friends
dressed up as gods so that his people will see him hobnobbing with them” (13.1.3–8). Or he might “kill an enemy by
arranging to have the image of a god fall on him” (12.5.1–5), or “have his agents use the blood of animals to cause a
hemorrhage to flow from images of the deities in the territory of the enemy and then have other agents declare defeat
in battle in consequence of the bleeding of the deity” (3.2.27–28).
11. I cannot agree with Thapar 1997, 155 where she associates this reference to gods visiting earth with
Aśoka’s “confidence that his Dhamma had achieved so much good in the country that it was just as it had been in
the righteous days of the Kṛtayuga when the gods in their pleasure visited the earth and associated with the peo-
ple.” As we shall see in this chapter and chapter 6, Aśoka knew and used the term kappa (kalpa) but not yuga,
which was probably introduced in the sense attributed to him only later in Brahmanical texts. Thapar also assumes,
I believe equally anachronistically (see chapter 5 § E), that the Brahmanical concept of varṇaś̄ ramadharma would
have been part of the discussion of what constitutes dharma” in Aśoka’s time (2005e, 436). Bloch 1950, 146 n. 7
rejects a similarly anachronistic view attributed to Sylvain Lévi that Aśoka might simply have had the program of
every Hindu king to implement dharma as his svadharma, refuting this on the basis of Aśoka’s conversion (29–30).
More tellingly, such a concept of svadharma is probably post-Aśokan (see chapters 5 § A; 10 § B) and inconsistent
with Buddhism (see chapter 12).
40 dharma

are common today in temples of Kālī or Śiva” (Falk 2006b, 57). Other “dharma
texts” from the same temporal “cluster” as Aśoka’s edicts (see chapter 1 § A) are
likewise disapproving of samājas. The Pāli canon mentions “samājas taking place
near Rājagṛha” at gatherings on hilltops, and forbids monks and nuns “to visit
them out of a similar disgust” (Falk 2006b, 56). And Āpastamba Dharmasūtra
1.32.19–20 mentions that if a Brahmin should go to a samāja or “fair,” he “should
circumambulate it and go away” (Olivelle 2003, 73; Falk 2006b, 57). Mentioning
further supportive evidence from the orgiastic Vedic Mahāvrata ritual and from
classical Jaina texts, Falk concludes that “folk religion has a long history. Being
rather shapeless with regard to doctrine it nevertheless served basic needs. In fact,
it was and is so vital that it survived all Vedic and many high-caste Hindu develop-
ments with little or no change” (56). Festivals at sites of MRE 1, which also cer-
tainly would have survived Aśoka’s attempted intervention, would thus have
involved blood sacrifices, swinging rites for young people, orgiastic elements, and
practices of trance and possession. Aśoka wanted to tone down popular festivals.
In contrast to the cryptic message and locations of MRE I, the Major Rock
Edicts (RE) were located mainly “in border areas,” where most of them “can be
visited without long walks starting from an ancient city site.” Then last, the Pillar
Edicts (PE) were “chiseled on pillars mainly in UP and the adjoining Nepalese
Terai,” where, “in some way” they “all seem to be connected with places of the
Buddhist saṅgha or were part of an itinerary leading to the birthplace of the Buddha
himself,” which “may explain why some of them are found in rather remote areas”
(Falk 2006b, 55). In RE 13, the famous Kalinga Rock Edict, Aśoka records that in
his eighth year, ca. 260 BCE, he felt remorse over the massive deaths and hard-
ships caused by his conquest of Kalinga, and now “considers conquest by dhaṃma
(dhaṃma-vijaya) the most important conquest” (Bloch 1950, 129–30). From RE 8,
we learn that he became an upāsaka or Buddhist lay disciple, and after about a year
and a half, in ca. 258, following a visit to the Saṅgha that made him more energetic
in his efforts, he either set out for Bodh Gaya or “set out for enlightenment (ayāya
saṃbodhim),” inaugurating a dhaṃma-tour” that took him, for 256 nights, to such
places as Bodh Gaya, scene of the Buddha’s enlightenment, and Lumbini, the
Buddha’s birthplace. The phrase ayāya saṃbodhim is surprising and much debated,
since Aśoka is never more than a layman and a busy king to boot, and also since
the term saṃbodhi most typically describes nothing less than the Buddha’s
complete enlightenment.12 Since no other edict marks Aśoka’s progress in such
terms, it is uncertain how to take the Buddhist intention behind the journey.

12. On the key and often differently interpreted phrase, see Hultzsch 1969, xliii (“a visit to Saṁbodhi, i.e.,
Bodh-Gayā”); Thapar 1997, 37–38 (went to the Bodhi tree); Bloch 1950, 112 n. 6 (possibly implying a quest for
“l’illumination parfaite—après des renaissances,” implying a doctrine not taught formally until the Mahāyāna; or
going on pilgrimage to the Bodhi tree, which, however, does not normally go by the name saṃbodhi (cf. 33). Cf.
Lamotte 1988, 226; cf. Nikam and McKeon 1978, 37, 66.
 aśoka maurya 41

Then, as Aśoka nears the midpoint of his royal career, he promulgates

three rock edicts (RE 3, 4, and 5) that mark his innovations in institutionalizing
a veritable dhaṃma bureaucracy.13 Beginning in his twelfth year (256 BCE),
RE 3 orders that provincial and state officials go about proclaiming dhaṃma in
five-year circuits, endorsing as “good” (sādhu) such things as obedience to
father and mother; liberality to friends, familiars, relatives, Brahmins, and
Samaṇas or ascetics; abstention from killing living creatures; and the minimum
in expenditures and possessions (Bloch 1950, 95–97; Nikam and McKeon
1978, 58). Coming also in his twelfth year, RE 4 is important to this book for
the way Aśoka presents the changes he is seeking to implement in relation to
his ideas about past and future time:

In the past,14 over many centuries, killing, violence done to creatures,

discourtesy to relatives, and disrespect for Brahmins and Samaṇas
have only increased. But now, thanks to the dhaṃma conduct
(dhaṃmacaraṇena) of King Priyadassi, beloved of the gods, the sound of
drums has become the call to dhaṃma. Thanks to king Priyadassi,
beloved of the gods, summoning people to tableaus of celestial palaces,
tableaus of elephants, balls of fire, and other divine displays,
promulgation of dhaṃma has increased that which did not exist over
many centuries: abstention from killing, kindness to creatures, respect
to relatives, respect for Brahmins and Samaṇas, and obedience to
mother, father, and elders. This dhaṃma conduct has increased in
diverse ways, and will increase more thanks to King Priyadassi, beloved
of the gods; the sons, grandsons, and great-grandsons of King
Priyadassi, beloved of the gods, will make this dhaṃma conduct increase
more until the end of the world (āva sa(ṃ)vaṭṭakappā).15 Living according
to dhaṃma and morality (sī lam), they will instruct on dhaṃma. For that
is the best activity, instruction on dhaṃma (dhaṃmānusāsanam).
Moreover, dhaṃma conduct is impossible without morality; but progress
on this point, and the absence of diminution, is good. I had this
inscribed so that one applies himself to progress on this point and so
that one admits no diminution.16 Twelve years after his coronation King
Priyadassi, beloved of the gods, had that engraved.17

13. For extensive discussion, see Thapar 1997, 94–181, 199.

14. Aśoka begins RE 8 with the same phrase (in the Girnār version of this edict), and has a similar phrase
near the beginning of RE 6 (Bloch 1950, 97).
15. See chapter 6 on this usage from the Girnār version of RE 4, and variants.
16. It is here that Nikam and McKeon supply the largely gratuitous translation partially cited in n. 2: “This
edict has been inscribed to inspire my descendants to work for the promotion and to prevent the decline of
Dharma,” paraphrasing as if in echo of the Bhagavad Gītā.
17. Cf. Bloch 1950, 97–101; Nikam and McKeon 1978, 31–32.
42 dharma

Let us note that Aśoka is the first traceable king to speak of change in dharma,
from a worsening if not virtual default of dhaṃma over centuries to generations of
“progress.”18 He distinguishes his reign as a turning point, separating his lineage
from past kings and projecting a grand future not only on a dynastic and imperial
scale but a cosmic one. As Bowles says of such passages where Aśoka speaks of
his heirs, “The overwhelming sense is that Aśoka was heralding, with a consid-
ered rhetoric, what he wanted others to believe was a new era” (Bowles 2004, 97;
2007, 128 modified; cf. Thapar 1997, 155). Yet those are not quite Aśoka’s terms.
Rather, on a point whose significance we shall explore in chapter 6, Aśoka speaks
not of having launched a new era but of having changed the course of dhaṃma in
such a way that its bettering can increase to the end of a kappa or kalpa. For this
chapter, however, it is enough to note that Aśoka uses this terminology only in one
other edict, in RE 5, where, one year later, in his thirteenth year, he indicates that
he is a little less sanguine about his sons and grandsons contributing to such a
cosmic fulfillment. Beginning on the note that good deeds are difficult and that he
and his sons have done many, he says, “If my sons, grandsons, and, after those,
my posterity follow my example until the end of the world (āva saṃvaṭṭakappā),19
they will also do well. But he who neglects even a detail does evil, for sin (pāpam)
is easily done.” Rather than investing much more hope in his offspring, Aśoka
turns immediately to announce a new title for officials who will head a new
department that will take fuller charge of his experiment in social engineering.20
These new officers, also mentioned in MRE 2 (which must thus come after this in
his program), are called dhaṃmamahāmattās: “dhaṃma overseers” or “dhaṃma
superintendents.” Says Aśoka, “In the past there were no dhaṃmamahāmattās.
I have created them. . . .” They are charged to work at instilling dhaṃma among all
sects (savvapāsaṃdesu); among those Greeks, Gandhārans, and other western peo-
ples who are devoted to dhaṃma; and among Brahmins. They are to bring relief to
the poor and aged, and to prisoners. They are assigned everywhere in the kingdom,
from the capital to provincial towns, and even “in all the harems of my brothers
and sisters and other relatives.” Having announced the new portfolio of these
dhaṃma overseers, Aśoka concludes: “This dhaṃma text is inscribed by me to last

18. Where Aśoka mentions “progress” (vaddhī, vaddhi, vaḍḍhī, etc.) and “increase” (vaḍḍhayissati and vari-
ants) in RE 4 (Bloch 1950, 99–100), he uses terms related to Sanskrit Övṛddh, “to increase.” Later, in the pillar
edicts, we find the term dhaṃmavaḍḍhi, “progrès dans le Loi” (Bloch 1950, 167, 168). Cf. Nikam and McKeon
1978, xiii: “increase in morality (Dharma-vṛddhi).”
19. Again from the Girnār version; see chapter 6.
20. Visvanathan adds the revealing point that although Aśoka refers to sons, gransdsons, and further
generations “assuming both the reckoning of descent as well as a familial entitlement to rule,” and “also the
rather frank assessment that the abilities of his descendants might not match up to his hopes,” “there is no
explicit charting of a genealogical line” (Visvanathan 2010, 3). Quite interestingly (see chapter 7), Visvanathan
finds that “the genealogical moment” occurs first among the Śuṅgas (19).
 aśoka maurya 43

a long time, and so that my children conform to it” (Bloch 1950, 101–6; Nikam
and McKeon 1978, 38–39, 58–59). One may read RE 5 as a warning. Aśoka has
realized that his brothers, sisters, and especially his children bear watching, and
that if his descendants are to make dhaṃma increase to the end of the kalpa, it will
take administrative changes.
Finally, Aśoka’s seven pillar edicts all come in a late burst in his twenty-sixth
and twenty-seventh regnal years, ca. 242–241. Aśoka brings an almost retrospec-
tive tinge to the inscriptions on these highly polished lofty pillars, with deepened
reflections on the nature of dhaṃma, the control of sin and passion, regulations
of feast and animal slaughter, and, especially in PE 7, a further filling out of the
dossiers for the dhaṃma overseers and “controllers” or “provincial governors”
(lajjūkas) involved in the widened apparatus of dhaṃma supervision and instruc-
tion.21 The pillar inscriptions suggest that over the years of his reign dhaṃma
“seems to have acquired a far more organized set of rules” (Nikam and McKeon
1978, 8–11; Thapar 1997, 173–79 quoting 173). But at the same time, with what
seems to have been a similarly growing interest, Aśoka came to emphasize that
dhaṃma is more deeply developed by meditation than by moral prescriptions.
There are certainly good reasons to emphasize Aśoka’s political motiva-
tions along with his administrative shrewdness in putting a generalized dhaṃma
to imperial work.22 Drawing an analogy with American “rhetoric centered on
‘family values,’ ” Olivelle says, “For Aśoka and his political operatives, dharma
was the ‘family value’ cliché of the third century BCE. It was a masterly political
move. . . . He claims to have instituted dharma within his bureaucracy, even
appointing dharma spies against his own family” (Olivelle 2005b, 129). Let us
keep this suggestion of dharma as “family values” at play: first to notice that
Aśoka makes his own royal household the scene not only of dharma supervi-
sion, as Olivelle notes, and, at least according to the Aśokāvadāna, of intense
family dramas (see Strong 1983), but also the irony that such an analogy might
hold for those who would see Aśoka espousing a religion that advocated home-
lessness as a higher value than family over and above one that could espouse
the householder life as its highest ideal.23

21. See Bloch 1950, 163, 169 and Nikam and McKeon 1978, 34, 59 translating lajjūka, which the latter give
as “Rājūka.” They were also mentioned in MRE 2. Also among those charged with inculcating dhaṃma were
“superintendents overseeing women, farm superintendants, and other corps of officials” (RE 12: Bloch 1950, 124;
Nikam and McKeon 1978, 52), and others saw to rest houses, wells, and shade trees along the roads for men and
beasts (RE 2; PE 7).
22. See Thapar 1997, 309; Lamotte 1988, 228, 233–36; Nikam and McKeon 1978, xv (McKeon’s Foreword);
Bowles 2007, 131; Olivelle 2004a, 505.
23. See chapters 4 and 5 on Buddhist and Brahmanical views of the householder, and Aśvaghoṣa’s riposte
to Brahmanical views in chapter 13.
44 dharma

While there are also good reasons to emphasize Aśoka’s secular motiva-
tions in implementing such a value, beginning from what we now know about
MRE 1, I see no possibility to doubt his growing seriousness as a Buddhist.24
Among three edicts addressed to the Buddhist Saṃgha (Lamotte 1988, 234–38),
in a Minor Pillar Edict (MPE 1) at Sāñchī, he condemns schism in the order, or
in local saṇghas, both among monks and nuns (Bloch 1950, 152–53; Nikam and
McKeon 1978, 67–68; Tieken 2000). And in the Bairāṭ-Calcutta Edict (formerly
called the Bhābhrā Rock Edict; Falk 2006b, 106), he says, “Whatever the Lord
Buddha has said is of course well said. But it is proper for me to enumerate the
texts which express the true dhaṃma (saddhaṃma) and which may make it
everlasting” or “enduring,”25 whereupon he cites seven such texts (only three of
which can be identified canonically “with anything approaching unanimity”26),
and states it to be his desire that the majority of monks and nuns listen to them
often and reflect on them, and the same for layfolk of both sexes (Bloch 1950,
155). We shall see further evidence of Aśoka’s familiarity with monastic con-
cerns and idioms—as with the powerful and potentially provocative term “true”
or “real dhaṃma” or “good law” (saddhaṃma), just mentioned, which can be
used by Buddhist authors with a certain edge.27 Saddhaṃma occurs just this
once in the edicts (Bowles 2007, 131), but clearly in a telling fashion—indeed as
the only case in the edicts where dhaṃma occurs in a compound as the second
rather than the first member (Olivelle 2005b, 127). It is likely, since the Bairāṭ-
Calcutta Edict is directed to specifically Buddhist audiences, that it reflects
insider language.28

24. See above, n. 2. Cf. Bloch 1950, 29, somewhat overstating the point: “Les édits d’Asoka sont des ordon-
nances d’inspiration bouddhique. . . . elles participent à la fois de la confession et du sermon.” Incidentally, I see
no merit in recent neo-Hindu arguments that the Aśokan edicts were not by Aśoka, but by later Buddhists
reconstructing a gloried past, or even that Aśoka did not exist. Even without Falk’s findings on MRE 1, such an
argument is unconvincing, but MRE 1’s locations make it totally incredible.
25. Nikam and McKeon 1978, 66, with “everlasting”; Bloch 1950, 154 has, more modestly, “propre á ren-
dre la Bonne Loi durable (saddhaṃme cilaṭṭitīke).”
26. Schopen 1997, 24–25. Cf. Bloch 1950, 154–55 n. 6; Lamotte 1988, 234–37. See, however, Thapar 1997,
who supposes Aśoka would have had familiarity with at least “some elements of the Buddhist canon”
(281, 149).
27. See Nattier 1991, 66–68, “The term saddharma (Pāli saddhamma) occurs in some of the earliest layers
of Buddhist literature.” As she indicates, with Buddhists using dharma for the teachings of other leaders, “there
was a need for a term that would refer specifically to ‘the teachings of the Buddha Śākyamuni,’ and it is precisely
in this sense that the term saddharma first begins to appear” (67). Having gained wide circulation, by the first
century BCE it could then be “the basis for the construction of another compound expression,” meaning the
“semblance,” “image,” or “shadow” of the “true dharma,” that could describe transformations even within
Buddhism. See further chapter 7.
28. Thapar 1997, 180 proposes that this edict would have been personal, addressed to two nearby Buddhist
monasteries, and from late in Aśoka’s reign.
 aśoka maurya 45

B. A Comprehensive Dhaṃma

Given everything we have described from the personal and familial to the imperial
and even the cosmic scale, it is fair to say that Aśoka did not shy away from seeing
dhaṃma in a grand picture. Let me indicate five ways that he does this in more
specifically social terms, and then in some further matters of detail.

1. Regarding the “ethical import” of his edicts, Bowles and Lamotte both
mention Aśoka’s repeated injunctions to be respectful and generous to
Brahmins, Samaṇas, parents, teachers, elders, servants, slaves, the
weak, and the poor.29 Without saying what they are, Aśoka also says he
has a respectful concern for the happiness of “all groups” or “classes”
(savvanikāyesu; PE 6: Bloch 1950, 168; Nikam and McKeon 1978, 36).
Although we shall see in chapter 4 that there are parallels and no
doubt precedents in early Buddhist texts, this is the first time we find
dharma intended officially as an ethic that would impact social groups
high, low, and across the board. Yet as Olivelle 2010 notices, Aśoka
never uses the term varṇa or its hierarchical theory, and mentions
Brahmins along with Samaṇas (and as we have seen, sometimes
without them) only as a religious group, and not in a caste hierarchy.
When the Greek ambassador Megasthenes resided at the Mauryan
court in Pāṭaliputra in around 300 BCE during the reign of
Candragupta, Aśoka’s grandfather, he met a sevenfold division of
Indian society that was not reconciled to the Brahmanical varṇa theory
(Karttunen 1997, 82–83; Bronkhorst 2007, 361 n. 22), which Aśoka
also thoroughly ignores.
2. As Nikam and McKeon observe, “Far from restricting Dharma to the
tenets and practices of a single religion, Aśoka asserts . . . that Dharma
is cultivated in all religions and sects, and he seeks to advance Dharma
in all men whatever their religious affiliations; and, true to this
purpose, he instructs Buddhists . . . to pay more attention to Buddhist
texts on Dharma”30—as we have seen, probably holding in reserve
some Buddhist sense of “the true dhaṃma.” Such an intent is also
carried to peoples on and beyond the imperial borders.

29. Bowles 2007, 130; Lamotte 1988, 231. For the most extended lists, see RE 9 and 13. Lubin 2005, 80 n.
7 observes that while Aśoka speaks of giving to both Sramaṇas and Brahmins in the same breath, “when
particular acts of patronage are mentioned, it is almost always the Buddhists who are the beneficiaries.”
30. Nikam and McKeon 1978, 20–21, citing RE 12 and “MRE 3” (the Bairāṭ-Calcutta Edict, just mentioned,
which encourages the study of seven texts), and offering this take on the latter.
46 dharma

3. Such a universal dhaṃma takes on legal ramifications for the king and
the state beyond anything envisioned in the early dharmasūtras. In
PE 4, “controllers” or “provincial governors” are told that uniformity is
desirable in legal procedures and in punishments; they should
therefore give convicted prisoners who are condemned to death a three
days’ respite for appeals, and then, should no appeal be forthcoming,
provision for them to prepare for the other world by giving gifts or
fasting (Bloch 1950, 165; Nikam and McKeon 1978, 60–61). Likewise,
after the Kalinga war, in this still unpacified country,31 in one of his
two additional Kalinga edicts Aśoka reminds his judicial officers in
cities there that no one should be submitted to unjust imprisonment
or torture without reason; when this happens and a prisoner dies,
many suffer (Bloch 1950, 137–39; Nikam and McKeon 1978, 62–63).
As with the suffering from the Kalinga war itself that provoked his
famous “change of heart,” Aśoka remains focused on suffering.
4. Yet, in conquering Kalinga, beyond even the bloodshed and
deportation he caused, Aśoka reflected: “This weighs even more on the
beloved of the gods: Brahmins, Samaṇas, and those of other
communities (paśaṃdā, praṣaṃda), householders (gih[i]tthā, grahatha)
who practice obedience to superiors, obedience to fathers and mothers,
obedience to gurus, perfect courtesy with regard to friends, familiars,
companions, and relatives, with regard to slaves and domestics, and
firmness in the faith, are all victims of the violence, the killing, or of
separation from those who are dear to them. . . .” (RE 13: Bloch 1950,
126–27; Nikam and McKeon 1978, 27–28). Without saying so
explicitly, Aśoka is admitting that he has brought suffering to those
who “do” dharma. Indeed, Aśoka implies that some people are what
could be called “sources of dharma”: the “gift of dhaṃma”
(dhaṃmadānaṃ) is imparted when “Father, son, brother, master,
friend, acquaintance, or even neighbor say, ‘This is good.’ ‘This ought
to be done (sādhu ida kattavyaṃ)’” (RE 11: Bloch 1950, 120; Nikam and
McKeon 1978, 45; cf. RE 9). These ideas do find counterparts in the
early dharmasūtras.32
5. Finally, such matters are tempered by wisdom. As Lamotte remarks,
Aśoka’s usage of dhaṃma recalling past kings is paralleled in Indian

31. See the second additional Kalinga Edict, by which Aśoka wishes to reassure the “unconquerred peoples”
and “border peoples” there (Nikam and McKeon 1978, 53–54).
32. In the late PE 7, Aśoka also says, “Whatever good deeds I have done the people have imitated, and they
have followed them as a model.”
 aśoka maurya 47

“descriptions of lay morality scattered throughout the ancient

anthologies of universal wisdom” and in passages “dispersed
throughout the canonical writings.”33

This sense of the comprehensive applicability of dhaṃma also carries down

to minute detail, as can be seen in a number of areas. One of these is the edicts’
dislike of festivals and critique of useless rites. Having first sought to undercut
popular festivals in MRE 1, in RE 1, Aśoka “banned religious assemblies (samājās)
except for those that propagate his dhaṃma,”34 although he allowed, as we have
seen, for new festivities “summoning people to tableaus of celestial palaces, tab-
leaus of elephants, balls of fire, and other divine displays.” Although we may now
see this as directed primarily, or at least initially, at popular non-Brahmanical
fairs, it would also have impacted Brahmins. As Olivelle indicates, citing RE 1,
“the Aśokan insistence on not killing . . . referred in a special way to sacrificial
killing. . . . This downgrading, if not the abolishing, of animal sacrifice cut at the
very heart of Brahmanical self-definition as the guardians of the
sacrifice, . . . [which] assured rain, prosperity, and social harmony” (Olivelle
2005b, 131–32). Aśoka also claims that ceremonies performed on the occasions
of “sicknesses, marriages of sons and daughters, children’s births, and jour-
neys,” and especially the “many diverse, trivial, and meaningless ceremonies”
performed by women, are of “little use” in this world and none in the next. This
would probably include a generalized reference to Brahmanical domestic or
gṛhya rites. In contrast, “the ceremony of dhaṃma” (dhaṃmamaṅgala) that “con-
sists in proper treatment of slaves and servants, reverence to teachers, self-mastery
in the presence of living beings, and liberality to Samaṇas and Brahmins” does
bear fruit in this world and the next (RE 9: Bloch 1950, 113–15; Nikam and
McKeon 1978, 46–47). Here, as in a number of edicts (RE 6, 11, 13: PE 1, 3, 4;
both additional Kalinga edicts), Aśoka addresses a concern that one also finds in
teachings of the Buddha: that the dhaṃma is pertinent, especially for laymen, to
happiness in both this and the other world. Fruitful ceremonial can take the
form of the “gift of dhaṃma” (dhaṃmadāna) (RE 11) or the reduction or elimina-
tion of killing animals.
Another area implying nuanced applicability of dhaṃma to varied detail
is linguistic. As noticed by Bowles and Olivelle, dhaṃma occurs in the edicts

33. Lamotte 1988, 228; cf. Halbfass 1988, 330, who sees Aśoka as exemplifying the tendency to “ethicize
and universalize” the “golden rule”: something resisted by Kumārila’s Mīmāṃsā, but also found in the
Pañcatantra, the Mbh, etc.
34. Alles 1994, 172, n. 52. Cf. Bloch 1950, 91 n. 10 on the probable range of meaning of samāja, in RE 1,
including ordinary gatherings and religious and festive ones.
48 dharma

with significantly high frequency as the first member of compounds.35 In this

position, “rather than being qualified, therefore, it is in fact the qualifier”
(Bowles 2007, 131); for Aśoka, “everything is dharmic” (Olivelle 2004a, 505).
As Bowles observes, this is “a remarkable contrast” to a predominant usage
of dharma in Brahmanical texts as the second term in compounds to
describe the “laws” or “duties” of social groups and people in particular
social positions, as in kuladharma (“clan-” or “family-dharma”), varṇadharma
(“caste-dharma”), and rāja-dharma (“king’s dharma”). According to Bowles,
in the Aśokan edicts, “dharma is never qualified in this way at all,” and the
only usage of dhaṃma as the “last member of a compound” is the above-
mentioned “saddhaṃma, the true dharma, an obviously different case” (Bowles
2007, 131).
Finally, a third area of especially concentrated application of dhaṃma to
varied and minute detail is that of meditational self-scrutiny. In the addi-
tional Kalinga edict that warns of the deleterious consequences of unjust
imprisonment and torture, Aśoka says, “That is why you must wish to prac-
tice impartiality. But it is not practiced with tendencies like jealousy, anger,
cruelty, haste, stubbornness, laziness, and fatigue. One must wish to escape
these tendencies. The principle of all this is to avoid inconsistency and haste
in the exercise of your functions” (Bloch 1950, 138; Nikam and McKeon
1978, 62). Such an emphasis, which is not absent from other rock edicts,36
increases in the late-life pillar edicts, where it concurs with a concern for
individual merit and an intensified sense of “sin.” In PE 1, we find, “It is dif-
ficult to conquer this world and the next without intense love of dhaṃma
(aggāyā dhaṃmakāmatāyā), intense vigilance, intense docility, intense wari-
ness, and intense energy” (Bloch 1950, 161; Nikam and McKeon 1978, 41).
In PE 2, Aśoka asks, “Dhaṃma is good. But what is dhaṃma? It is the absence
of the causes of sin, abundance of good acts, pity, giving, truthfulness,
and purity ( dayā dāne sacce socaye)” (Bloch 1950 , 162 ; Nikam and McKeon
1978, 41). In PE 3, after remarking that a man is more prone to notice his
worthy deeds than his evil ones, Aśoka says, “This is truly a difficult examina-
tion. But he must consider matters in this way: ‘What one calls access to sin,
that is to know wickedness, hardness, anger, pride, and envy. Now I must not

35. Olivelle counts thirty-one instances (2004a, 509 n. 28), several of which are noticed in this chapter,
including dhaṃmalipi, dhaṃṃayāttā, dhaṃmaparipucchā, dhaṃmavijaya, dhaṃmamahāmātā, dhaṃmamaṅgala,
dhaṃmavaḍdh ̣ i, dhaṃmadāna, dhaṃmaniyama, dhaṃmaguṇa, dhaṃmatthaṃba for Aśoka’s “moral pillars,” and
dhaṃmānusāsana and dhaṃmanusista for “moral instruction.” See also Nikam and McKeon 1978, xiii, 44–45 n. 6.
36. See RE 5 as cited above on the difficulty of doing good deeds and the easiness of bad ones; RE 10 on
Aśoka’s strivings for the other world by escaping bad tendencies, which requires intense effort whether one is
small or great, but is most difficult for the great (Bloch 1950, 118–19; Nikam and McKeon 1978, 48).
 aśoka maurya 49

wish to lose myself for these causes.’ Here is what he should consider above
all: ‘This counts for this world, this counts for the other world.’ ” (Bloch
1950, 163; Nikam and McKeon 1978, 48). In PE 7, his very last, which, like
some of the earlier rock edicts starts off about kings of the past, Aśoka states
that “progress in dhaṃma (dhaṃṃavaḍḍhi) among men has been obtained by
only two means; by dhaṃma rules (dhaṃmaniyamena) and by meditation
(nijjhattiyā). But on this point, rules are of little consequence; meditation is
more important . . . it is by meditation that one obtains the greatest progress
in dhaṃma in view of the conservation of beings and abstention from killing
animals” (Bloch 1950, 172; Nikam and McKeon 1978, 40 [curtailed]). As
Lamotte observes, “He listed the ‘virtues of the Law’ (dharmaguṇa)37 the
practice of which ensures happiness in this world and the next. . . . These
lists of faults and virtues are very similar to those found in the Buddhist writ-
ings, particularly the Abhidharma” (Lamotte 1988, 232). Indeed, we may say
that Aśoka was finding everyday terms with which to encourage his officials
and people to practice the “discrimination of dharmas.” In quasi- or proto-
Abhidharmic terms, Aśoka would be seeking to clarify how such “virtues of
the Law” can be cultivated as “mental events” (dharmas) by “right effort,”
“right mindfulness,” and the “discrimination of dharmas.” We shall return
to this subject in chapter 4.
Aśoka’s interest for the history of dharma is not limited to his own times.
As several scholars have observed from different angles, Brahmins at large
might or might not have ignored the Buddha, but Aśoka’s usage of the term
dhaṃma to define a new imperial non-Brahmanical program was something
that they could not ignore.38 In particular, Brahmins who preserved areas of
Vedic expertise would have found themselves confronted by a campaign that
deemed their rites and knowledge imperially useless, and implicitly treated
them as having the lesser of two moralities and unaccepted views, whatever
they were at this time, of the past, the future, and the legitimate roles of kings.
In reply, they would produce texts that featured Brahmanical dharma in new
post-Vedic genres. As mentioned in chapter 1, such motivations for producing
new texts on dharma might also have spurred the composition of the earliest of

37. Such a linking of dharma and guṇa gives a foretaste of dharma as virtuous quality and merit in the Mbh,
where there is more or less an equation between such compound phrases as sarvaguṇopeta and sarvadharmopeta,
“endowed with every virtue (quality, merit).” See Hiltebeitel [1976] 1990, 192–228.
38. See Olivelle 2004a, 505: Aśoka’s usage of the term dhaṃma to define “a new imperial ideology” of the
Mauryans was a factor that made the Buddhist dharma something that could not be “ignored even by the scho-
lastic Brahmins working within the vedic śākhās”; Bowles 2007, 125: “his ideology,” which “seems heavily influ-
enced by many of the values most vigorously pursued” by the non-Brahmanical Śramaṇa movements, was
“perhaps a key moment in the rise of the concept [dharma] to a central status within the various Indian intellec-
tual and religious traditions.”
50 dharma

these Brahmanical texts slightly before Aśoka, since his two Mauryan predeces-
sors, his grandfather Candragupta (ca. 321–297) and his father Bindusāra
(ca. 297–272), already favored other non-Brahmanical Śramaṇa movements,
Jainism and Ājīvikism, respectively, and Alexander the Great had left Greek
and Iranian ideas of empire in India from 327 BCE on. But it is a good working
hypothesis that even the earliest Brahmanical dharma texts would be no earlier
than these three earliest Mauryans.
3
A Vedic History of Dharma

This chapter will attempt to set our bearings on the early history
of dharma in Brahmanical texts of the Vedic canon, beginning with
the oldest source, the Ṛgveda (henceforth in this chapter ṚV).
Patrick Olivelle’s contribution to the 2004 Journal of Indian
Philosophy volume that he edited, titled “The Semantic History of
Dharma The Middle and Late Vedic Periods” (2004a), was cited in
chapter 2 for its observation that the term dharma is not statisti-
cally, at least, significant in Vedic Brahmanical texts before the
Aśokan edicts. Olivelle offers a new hypothesis on the innovative
character of the early Buddhist usage of dharma toward the end of
the Vedic period. This innovation lay in seizing on a pre-Buddhist
usage having to do with the relationship between kings and their
Vedic divine model, the god Varuṇa, to co-opt this royal term as
chief among a number of royal symbols by which, as leader of an
ascetic movement, the Buddha could lay claim “to a new type of
royal authority.” Shortly after this middle-to-late Vedic period and
the rise of early Buddhism, Aśoka Maurya would have then put the
Buddha’s “royal” transformation of dharma to imperial and more
secular work. In reaction to the Aśokan usage, post-Aśokan
Brahmanical would likewise put the king at the center in such texts
as The Laws of Manu, and the two Sanskrit epics would likewise
put the king at the center, as I will do in this chapter. Olivelle’s
propositions demand careful scrutiny, but I believe most of them
bear up to it.
52 dharma

A. Dhárman in Early and Later Ṛgvedic Usages

As Olivelle recognizes in citing the kingship of Varuṇa, the ṚV, India’s oldest
textual source, becomes the indispensable starting point for a history of
dharma—even though scholarly treatments of dharma usually start well after it
or pay it bare lip service by inadequate summaries and anachronistic backread-
ings. Such distortions are important to clear out of the way. Here are two watch-
words as to what dharma in the ṚV is not.1

1. It is not a Ṛgvedic “cosmic order.” This is the most common

backreading. The classical term dharma is seen as “replacing” a
Ṛgvedic notion of “cosmic order,” allowing one to backread this
understanding of dharma into Vedic dhárman (the precursor formation
in the ṚV). The term said to mean “cosmic order” in the ṚV is ṛta.
Ṛgvedic ṛta, best translated as “truth,” is a cosmic order resonant with
the “truth” of the Ṛgvedic hymns and mantras. The ṚV poets first
discerned it in their inspired compositions, and those who recite their
verses ritually can keep this cosmic order functioning. Since ṛta no
longer means “cosmic order” in classical Hinduism, and dharma sort
of does, it has been convenient to think that dharma not only replaced
ṛta but must always have had some such implication itself.2 But
dhárman did not mean “cosmic order.” Moreover, the ideas of “cosmic
order” to which dharma becomes attached differ from the Ṛgvedic
“cosmic order” denoted by ṛta. Classical usages of dharma put the term
into the service of a sociocosmic order that is more ideology than poetry
or ritual implementation. The Vedic “cosmic order” is something else.
2. Dhárman is not karman. This second type of backreading, also common
but more diverse, interprets dhárman through lenses of karman,
“action.” It is more diverse because karman/karma itself has different
meanings and usages: from its earliest sense of “ritual act” to the later
sense of a “law of karma” that relates actions to reincarnation. While
only the first of these meanings has been read directly into Ṛgvedic

1. I leave aside universalizing neo-Hindu backreadings, for, as Halbfass shows (1988, 334–48), these have
been applied mainly to Upaniṣadic and classical texts.
2. See Horsch 2004, 424–25, 434, 441, tracing the “origins” of dhárman to cosmogonic myth; 427: “It is
the ‘supporting hold- [Halt] giving power’, through which Varuṇa accomplishes his action of making the cosmos
stable.” Cf. Mahoney 1998, 107: “In early Vedic texts dharman refers to an established or proper mode of conduct
that supports or helps maintain the continuing health of the world. According to one such visionary [referring to
the poet of ṚV 6.70.1], for example, it was through Varuṇa’s performance of his dharman that the sky was raised
above the land. . . . The Vedic idea of dharman stands as precedent for the later idea of dharma as responsible,
proper activity that supports the world.”
 a vedic history of dharma 53

usages of dhárman, a fusion of the two has also been smuggled into the
mix. This is the idea that dharma as “duty” and karma as “act” imply
each other if one acts in accord with “one’s own dharma” (svadharma) in
“maintaining the cosmic order.” Such a backreading is found
frequently in popular approaches to the ṚV, where one can read that
each god “performs” his “own dharman” by acting in such a cosmically
supportive fashion.3 Certain classical texts, most notably the Bhagavad Gītā,
do make an equation between “one’s own dharma” and “one’s own
karma,” but that does not justify its wholesale importation back into the
ṚV. Unlike dhárman, both ṛta and karman are indeed governing
concepts4 in the ṚV, and karma remains one through all Indian
traditions. But neither of them ever governs the history of dharma. I am
not persuaded by attempts to reduce dharma to a species of “action.”5

Related to these points is a further misconception. Ṛgvedic dhárman is not

defined by usages in the plural. Here we have the rarity of being misled by
Wilhelm Halbfass, who writes, “Dharman usually appears explicitly in the plural
or possesses at least potentially plural function in meaning” (1988, 314)—the
concluding clause is left unexplained, and clearly begs the question. Of the sixty-
seven instances in the ṚV where one can determine whether dhárman or dharmán
is singular or plural, there are thirty-nine singulars to twenty-eight plurals!
Halbfass builds on this erroneous point to draw a contrast with karman that
entails a backreading from the later dharmasūtra/dharmaśāstra literature: “While
dharma, in its very essence, is subdivided into a countless number of individual
obligations and may, in keeping with the original pluralistic meaning of the
word, be characterized as the very sum or system of such rules, the doctrine of
karma develops one central and universal principle . . .” (322). Anne Monius has
repeated the error while compounding it with the usual reduction of dharma to
“acts”: “In the Vedic world of yajña, ‘sacrifice’ . . . dharma . . . significantly occurs
more often in the plural. Found more than sixty times in the Rig Veda alone,
dharma and its derivatives signal ritual acts . . .” (2005, 331). The mistake has also
been compounded by Rupert Gethin in an attempt to relate Buddhist usages to
prior Brahmanical ones when he proposes “practices” as an “inherited” meaning
of the Buddhist plural usage of dhammas (2004, 530–35). In fact, all three authors

3. See again Mahoney on “performance of his dharman” in the previous note, and further, 1998, 108 on
dharman as “closely associated with” karman, translating dharman as “established rites” in 5.26.6 and “proper
ritual performance” in 8.43.24; cf. Miller 1985, 102: “each [Ṛgvedic] god [is] following the law proper to his own
being, in other words his own dharman” (1985, 102).
4. Cf. Brereton 2004, 449, as cited: compared with ṛtá and vratá, each carrying forward Indo-Iranian
meanings, dhárman was “not a central term in the Ṛgvedic lexicon.”
5. For a different approach on this point, see Fitzgerald 2004a, 106–7.
54 dharma

refer immediately to ṚV 10.90.16, the closing verse of the famous hymn to the
Primal Man (Puruṣasūkta), to exemplify the point about plural usages implying
“acts.” But that is a late ṚV verse—indeed, precisely the one that might first sug-
gest a link between dhárman in the plural and the origins of the caste system;
and, as we shall see, there is probably a better way to translate its plural usage
of dhárman. Indeed, ṚV 10.90 should not be backread either. As one would
expect of any text or language state, the interplay of singulars and plurals can
be illuminating—but not in the sense of pluralized laws or practices that
fractionalize from, or consolidate into, a singular “law” or “order.”
I will proceed from the view that usages of the term dhárman do not begin
from a cosmological/mythical, ritual,6 legal, or religious context, or as an exhibit
of cosmological/mythical, ritual, legal, or religious thinking,7 but as a poetic
conceit, a poetically crafted concept. Evident as this point should be when
studying oral poetry, I am aware of no exception to the rule that scholarly
interpretations of ṚV usages of dhárman do not take this primary consideration
sufficiently into account when delegating the term originally to one or another
such zone. The term dharma, which comes to exemplify Hindu orthopraxy,
“right practice,” takes off from a new idea.8
The ṚV will thus be the initial focus of this chapter. Olivelle’s hypothesis
that royal authority is a major strain of the early meaning of dharma finds
support in Brereton’s essay on “Dhárman in the Rigveda” that precedes Olivelle’s
in the Journal of Indian Philosophy volume. Indeed, Brereton’s study not only
fills in a great need but makes it now possible to continue to explore Ṛgvedic
usages in ways that would not have been possible without his contribution. As
mentioned, we shall treat the relation between dhárman, dharma, and kingship
as the central topic of this chapter, and frame that center with discussions of
dharma as an enigma. Both topics call for some introductory discussion of the
chronological levels of the Ṛgveda.
In opening his study as a “reevaluation of the history of dhárma,” Brereton
begins with several important observations. “Since dhárman is a developing
term in Ṛgveda, its meaning reflects directly its etymology and form. And, hap-
pily, the formation of dhárman is transparent. It is derived from ÷dhṛ ‘uphold,
support, give foundation to’ and a -man suffix. Therefore, it denotes a thing
which upholds or supports, or, more simply, a ‘foundation’. The word dharmán,
a noun of agent, then designates an ‘upholder’ or ‘foundation-giver’” (2004, 450).

6. See Pollock’s view cited below in this paragraph.

7. Brereton 2004, 471: “The uses of the term illustrate the breadth of dhárman and suggest that the
liturgical sense of the term considered earlier is a reflection of the character of the Ṛgveda rather than that of the
word dhárman itself.”
8. Jamison 2010 makes a similar point about the novelty, as it can be traced into its near-emergence in
Ṛgvedic poetry, that a sacrificer’s wife is a good idea for an effective sacrifice.
 a vedic history of dharma 55

Sixty-three usages of dhárman and four of dharmán is “not a small number,”

but “this relatively modest frequency of dhárman nonetheless implies that it
was not a central term in the Ṛgvedic lexicon or in Indian culture of the Ṛgvedic
period” (2004, 449).9 Yet even if not “central,” “dhárman is thoroughly
established in the text, since the word is attested in all its chronological levels,”
with “increasing frequency in the younger layers” (450). Moreover, unlike
Ṛgvedic ṛta (“truth”) and vratá (“commandment”), which are “central,” and,
with their Avestan cognates, point back to “significant roles in the old Indo-
Iranian religious vocabulary,” dhárman, at least in its Ṛgvedic meanings, does
not have such a prehistory. For Brereton, this means that in contrast to the
other two terms, “discussion of dhárman can reasonably begin with the Ṛgveda”
(449). Indeed, setting the old Latin cognate firmus, and extrapolations from it,
aside (see, e.g., Zaehner 1966, 2–3), it would seem that we could also suspect
that dhárman could be a Ṛgvedic coinage, a new term with which to conjure.10
Brereton offers extensive discussion of verses that relate dhárman to gods
who are called king and lord, beginning with Varuṇa, but also opening up the pos-
sibility of a history of this set of associations within the ṚV itself. It will thus be
important to work from the relative chronology of the ṚV’s ten books that Brereton
outlines (2004, 450), and which has gained broad acceptance since it was refined
by Oldenberg in 1888. That chronology recognizes six stages of ṚV composition:

1. Old Family Books: Book 2, associated with the Bhṛgu poets; Book 4,
linked with the Gautama poets; Book 5, with the Atri poets; and Book
6, with the Bharadvāja poets
2. Young Family Books: Book 3, associated with the Viśvāmitra poets; and
Book 7, with the Vasiṣṭha poets
3. Book 1 and most of Book 8 (8.1–46/8, 8.60–103)11
4. Book 9, a collection of hymns to Soma, attributed to various authors,
prominent among them being the Kaśyapa poets
5. Book 10
6. 8.49–59 (the so-called Vālakhilya hymns, treated in the ṚV itself as an
appendix or khila)

9. Olivelle 2005b, 123 and n. 4 also notes fourteen usages of vidharman in the ṚV.
10. My thanks to Joel Brereton, June 17, 2007 e-mail, for further clarification on these points: “When I said
that dhárman was Indic and not Indo-Iranian, I meant that Iranian did not have a corresponding term darman with
anything like the significance of dhárman in Sanskrit. (Middle Persian and modern Persian do have a formally
corresponding darman, but the word means ‘medicine.’) The root dhṛ, however, has a corresponding and very well
attested root dar/d[ə = schwa]r in Iranian.” I had asked him about an instance where Jamison 2007, 37 cites Yasht
44.4 where dərətā . . . yaogət has been translated “Which man has upheld . . . ? . . . Who yoked . . . ?” or “Who
holds . . . ? Who yokes?” Brereton also comments: “I’m not certain whether Latin firmus is etymologically related.
I checked Mayrhofer [1986–2001, I: 779, 780] and he says that a derivation of firmus from IE *dher-mo is unsure.”
11. For a further analysis of the material of these books that will not effect the discussion here, see Witzel
1995, 309–10; 1997a, 262.
56 dharma

Although the Atri hymns mention the Yamunā River to the east and the
Bharadvāja hymns the Gaṅgā still further to the east, the Old Family Books
reflect the earliest polity of the five tribes, centered on the Pūru tribe, which had
settled mainly in the Punjab (Witzel 1995b, 318). The Young Family Books then
center more specifically on the Kurukṣetra area of eastern Punjab and reflect
the ascendancy of the Bharata clan over the Pūrus through the Battle of the Ten
Kings (328–37).
While the situation of ṚV oral poetry is such that each new generation had to
reconstitute its “canon” by memorizing the hymns of earlier generations, new
poets often worked from older models by adapting older formulas, verses, and
themes to new contexts—among them changing political configurations,
including changing conceptions of both earthly and divine kings. Taking the
Family Books together, Varuṇa is certainly the main divine king connected with
dhárman, whereas human rājas, with no mention of their having any special
association with dhárman, can better be called tribal chieftains. The Family Books
describe such chieftains in two contrasting modes: as leaders in yoga, which for
this context means “harnessing” for a battle march in search of booty; and as
leaders during kṣema, “peaceful settlement.” Although nowhere is a chieftain’s
activities mapped with dhárman, some passages suggest that a rāja could take on
both of these leadership roles. According to Whitaker 2006, the celebrated King
Trasadasyu, who seems to have unified the Pūrus and the other four early tribes
by a horse sacrifice or Aśvamedha (Witzel 1995b, 326, 329), has such a dual pro-
file in ṚV 4.38: a hymn to “Indra-Varuṇa” as a deity-pair in which one senses a
possible correlation of Varuṇa with peaceful settlement and Indra with the battle-
trek.12 Indeed, we shall note a verse in one of the Atri hymns that says that Mitra
and Varuṇa, according to their “Foundation” (dhármaṇā), give “peaceful settle-
ments that endure (dhrúvakṣemā)” (ṚV 5.72.2ab). But nothing would allow one to
correlate Mitra or Varuṇa’s dhárman with war. Victorious war will not be com-
pounded with dharma until the classical period.13

12. Cf. 7.89.8, enjoining Varuṇa for both yoga and kṣema; 10.89.10 where Indra is invoked for both. I thank
Whitaker too for his translation of ṚV 5.37.4–5: “This king does not waver/falter, in/by whom Indra drinks the
sharp/intense cow-befriended sóma. With his fighters/true men he drives (cows/wealth) here; he smashes Vṛtra/
Obstacle; he dwells in peace, fostering the settlements, the one whose name is ‘Having Good-Portions.’// He will
thrive/foster in times of settled peace. He will overcome in times of harnessing (for war). He will conquer simul-
taneously both opponents that have come together. He will be dear to the Sun, dear to Fire; the one with pressed
sóma will ritually serve Indra.”
13. The Buddha speaks of Sakka (Indra) winning “a victory by means of righteousness itself” (Saṃyutta
Nikāya 11.6 (Bodhi trans. 2000, 325). Aśoka inscribes dhaṃmavijaya on his Kalinga Edict (RE 13). See Bowles
2007, 128–29 and n. 191, calling attention to occurrences of dharmavijaya in the Rājadharma section of the Mbh,
presumably meaning the concept of “lawful conquest” rather than the actual compound term, which is not
found in either unit that he cites (Mbh 12.59, 12.96). Cf. Mbh 12.97.1 (adharmavijaya or “unlawful conquest”) and
of course BhG 2.31 (“there is nothing more salutary for a Kṣatriya than a war that is lawful [dharmyād hi yuddhāt
śreyo ‘nyat kṣatriyasya na vidyate])”; 2.33 (dharmyaṃ saṃgraham or “lawful war”), on which see chapter 11.
 a vedic history of dharma 57

As to Books 1, 8, and especially 9 (consisting of hymns to Soma), there

would be the possibility that some of their hymns, not to mention themes and
formulae, would be older than their places in this chronology.
Finally, Book 10 marks the completion of the ṚV canon that was appar-
ently undertaken in establishing what Witzel has called India’s first state, that
of the Kurus (1995, 337), now centered in “the modern (eastern) Panjab and
Haryana” (1997a, 266) and expanding Vedic culture into the upper Gaṅgā-
Yamunā valleys (1995, 335). For this Kuru “super-tribe,” we can now call its
rājas, who seem to have elevated themselves from among the Bhāratas or
branched off from them, true “kings” (Witzel 1997a, 264–68; 1997b). These
later ṚV books mark the rise of a more centralized kingship coordinated with
the canonization of the ṚV Collection (Saṃhitā), and also with the need to rede-
fine in relation to usages of dhárman the balance of royal divine power bet-
ween Varuṇa and other gods, but mainly Indra. From the Old Family Books to
Book 10, this canonization process itself and the geographical and political
situations it reflects would seem to run from about the fourteenth to the
twelfth centuries BCE. Whatever the status of Witzel’s construction of the
“Kuru ‘state,’” I believe it has sufficient substance to give it hypothetical status
to present a different picture from one advanced in many publications by
Romila Thapar. According to Thapar, the “lineage based” or “clan based”
society behind Vedic and “early” epic polities remained one “suggestive of
tribal chiefships” (Thapar 2005j, 635) rather than kings, allowing for early
strata of the Sanskrit epics to keep old Vedic legends about “chiefs” rather
than “kings” alive as reflections of the “bardic” side of early Vedic oral culture
(see Thapar 1993, 46–47, 136–41; [1999] 2002, 7–15, 38, 48; 2005i, 622–28;
2005j-k; 2005m, 711). For me, it is important, as I will emphasize, that even
before the “Kuru ‘state,’” Ṛgvedic poetry contrasted human chieftains with an
“imperial” notion of divine kingship seen in Varuṇa. Witzel’s “Kuru ‘state’”
provides a plausible moment from which to recognize that Vedic kings do
become larger scale “kings.” This bears on an idea I began to develop in
chapter 1,14 that the Sanskrit epics are not oral archives of old tribal legends
about “chiefs” but rather written texts that construct an imagined “history” of
“kings”—particularly in the Mahābhārata as the itihāsa (“history”) of the

Bowles also (same note) cites Arthaśāstra [KA] 12.1.10ff. with the comment: “it is curious that while many scholars
assume the KA dates from the time of Candragupta Maurya, and therefore use it to reconstruct Mauryan history,
few if any then consider the consequences of it containing a supposedly Aśokan idea like dharmavijaya.”
Dharmavijayin occurs at KA 12.1.11–12: “The righteous conqueror is satisfied with submission. He [the weak king]
should submit to him . . .” (Kangle [1972] 2003, 2: 460).
14. See further Hiltebeitel 2005c; 2010b; forthcoming-a; in press-a, chapter 4.
58 dharma

Kurus—and that these written texts were composed in the classical period
drawing on post-“Kuru-‘state’” models of kingship.
Now as Brereton remarks, “Interestingly, 7 of the 11 attestations of
dhárman in the oldest Ṛgvedic layer occur within Book 5.” That eleven is the
total in the Old Family Books, and the statistic is still suggestive when one
adds the Young Family Books, whose seven usages (counting one usage of
svádharman at ṚV 3.21.2) make it seven of eighteen for all the Family Books.
According to Brereton, this concentration presents the possibility that the
“increasing occurrence in later books may partly reflect the influence of the
Atri poets” of Book 5 (2004, 450). While there are reasons15 to think they
would not have been the earliest of the early clans of poet-priests to have
their poems collected, I begin both of the following sections with their
poems. The Atri poets were evidently the first to give concentrated thought
to dhárman, and although I know of no evident reason for this, a sequence
in one of their poems, ṚV 5.40.5–9, assembles a number of themes that
might be connected with dhárman in telling how their eponymous ancestor
Atri found the hidden sun by a “fourth formulation” (bráhman), healing it
and freeing it from darkness after it had been punished for an act “contrary
to commandment (apa-vrata),” and established it in heaven in the mid-day
Soma pressing in a rapport with the gods Indra, Varuṇa with a likely allu-
sion to Mitra, and probably Agni as the one who had punished the Sun for
its violation, which was perhaps incest with the Sun’s daughter Uṣas, the
Dawn. While the word dhárman does not appear in these verses or even in
later versions of the myth, we may at least feel our way into the Atri poets’
treatment of dhárman through their ancestor’s manner of restoring a fragile
ritual order to a none too perfect world.16

B. Dhárman as Enigma

To claim an enigmatic meaning for Ṛgvedic dhárman is not to claim much,

since Ṛgvedic poetry simply abounds in enigmatic verses. As we will observe
with variations on the root ÷dhṛ, Ṛgvedic poets like to play with verbal reso-
nances or soundplay and like to structure surprise into their poems by creative

15. Their mention of the Yamunā; their third of four positions in the order of the Early Family Books
(Witzel 1995, 317–18, 326–28, 331–33).
16. See Jamison 1991, 138–42, 188–89, 248–57, 264–66, 271–75, 281–88, 297–303. For Jamison, it is
Agni as “Svar-bhānu (possessing the light of the sun) Āsura [who] punished the sun for this violation of law and
custom” (302)—that of incest with his daughter Uṣas, Dawn.
 a vedic history of dharma 59

compositional devices, including framing, that affect the whole hymn.17 I will
now indicate ways in which specific verses use the word dhárman as a loaded
enigmatic word. Let us begin with the Atri poets of Book 5.
Ṛgveda 5.15 is a hymn to Agni. The first verse calls him “Agni the support
of goods.” The fifth and last verse uses the same term for “support” to describe
how the Atri poet imparts strength to Agni so that he may be helped to great
wealth. This word for “support” is dharúṇa, a derivative of ÷dhṛ like dhárman.
The poet shapes the hymn’s soundplay by surrounding dhárman with these
and other alliterative “supports” derived from ÷dhṛ.18 Moreover, in opening and
closing with verses mentioning dharúṇa, the poet uses a framing design
familiar from many Ṛgvedic hymns of having a poem’s first and last verses
form a “ring” around the composition. Verse 2 is this hymn’s concentration
point regarding dhárman. Here, following one of the verbal derivatives of ÷dhṛ,
dhárman is surrounded twice, this time close at hand, by dharúṇa:

In making powerful the sacrifice in the highest heaven, they

supported (÷dhṛ) the truth (ṛta), itself a support (dharúṇa), by means
of the truth (ṛta)/—they who have reached the men (= the gods) that
have taken their seat upon the Foundation (dhárman) of heaven,
upon the support (dharúṇa); they who, even though they themselves
were born, (have reached) the unborn. (ṚV 5.15.2, slightly modifying
Brereton trans. 2004, 451)

Two groups, each called “they,” have reached heaven. The first refers to ancient
poet-ritualists. By making the sacrifice powerful, they became semi-divine ances-
tors of the poet’s family. The second are “men” but appear to be gods, since two
things said about the first group are not said about them: the first group reached
heaven by sacrifice and were born. In contrast, the apparently “unborn” “men”
who got there first just sit “upon the Foundation, upon the support of heaven.”
It does not seem that dhárman, mentioned only once, and the “supports”
(dharúṇa) that surround it are simply interchangeable. Dhárman is set off by four
usages of dharúṇa—two in this verse and two in the hymn’s first and last verses.

17. See Jamison 2007, 35, 138 on the ṚV poet (kavi) as enigmatist; 79–80, 87, 95, 99 on ring composition;
76, 83–84, 97, 97, 107 on structure, rings, momentum, and movement; 82, 86–87, 104, 114 on “omphalos”
verses; 103 on open-endedness, ambiguity, uncertainty, yet with closure; cf. 71, closed ritual system; 112–14 on
“poetic repair”; cf. 61–64, 85–88.
18. As Brereton notes, T. J. Elizarenkova (1995, 152) describes dharúṇa, “support,” “as the hymn’s ‘magic
word’ ”—one that is “repeated and echoed by other derivatives of √dhṛ throughout the hymn.” Like Warder 1971,
275, who speaks of “a good deal of word play” in verse 2, Elizarenkova says the “phonetic shape of the word
[dharúṇa] has influenced the sound-play of the hymn,” but adds, “it remains unclear what kind of information the
poet intended to convey by this Sprachmalerei” (1995, 152). Geldner also calls the hymn “quite extraordinary”
(“kein ganz gewöhnliches”; 1951, vol. 2, 15) for this run of usages related to this one verbal root.
60 dharma

Dharúṇa comes to evoke a fairly ordinary Ṛgvedic idea of “support” that is

sometimes imaged quite concretely—for instance, where Indra “expanded the
unshakable support that set the atmosphere within the framework of heaven”
(ṚV 1.56.5), or where Soma is called heaven’s “support pillar” (9.2.5; 74.2). The
force of our hymn lies in zeroing in on dhárman not just as another “support” but
as a “Foundation” that is left a little more mysterious than its four surrounding
“supports,” which look something like flying buttresses. Rather than being inter-
changeable with these “supports,” dhárman is their inner “Foundation.” We
could call it the kernel of the poem: a Foundation wrought by poets for “making
truths powerful” and “reaching” heaven. By the truth of their hymns and rites the
poet’s ancestors reached the gods on that well-supported Foundation.
A second illustration comes from one of the Later Family Books. The poet
indicates that he reflects like a “fashioner” (táṣṭā, from ÷takṣ) upon his “inspired
thought” (manīṣaʹm̄ ), and goes on:

And ask about the forceful generations of poets (kavis). Holding the
mind (manodhŕ. taḥ) and performing well (sukŕ. tas), they fashioned
(takṣ) heaven. / And these are your (= Indra’s) leadings forth, which
grow strong and which are won by thought; therefore they go now
upon (that) Foundation (dhárman). (ṚV 3.82.2; slightly modifying
Brereton trans. 2004, 462)

Here dhárman seems to be the “Foundation” of the older hymns by which the
ancient sages “fashioned heaven” (plus “heaven and earth” as the next verse
indicates) by holding (dhṛ) their minds firm or steady and by performing poems
or rituals well, and upon which new generations of poets can still compose
hymns that lead forth Indra to manifest himself in the Rigvedic here and now
(cf. Brereton 2004, 462–63). The correlation of early and later generations of
poets also describes the early ones while answering a seemingly rhetorical
question about them: “And ask about the forceful generations of poets.” Ask
what? And who would be asking whom? One view is that the poet would be
addressing himself. But he could also be addressing the former poets, whom
the previous verse tells us he wishes he could see, or Indra, whom the hymn
and the verse itself invoke.19 We do not know whom the poet is addressing, but
the response lies in something he calls upon himself and probably other poets
to resolve by speaking about dhárman, a Foundation. The answer would thus
be that dhárman lies in poetry or its inspired thought.

19. Cf. Johnson’s notion (1980, 6–12, 53) of Ṛgvedic poetic “symposia” in which poets may have addressed
their sacrificial patrons and “companions” or “friends” (sákhis)—that is, each other—with enigma verses (bráh-
mans, brahmodyas) while competing for recognition and patronage. I note, however, that Houben 2000, 501, puts
the term “symposium” in quotes and says that Johnson provides “few philological data to support his view.”
 a vedic history of dharma 61

I would thus propose that with regard to dhárman the early Ṛgveda leaves
us with two types of deeply wrought enigmas. One, exemplified by our first
example, speaks of a Foundation above, with “supports” around and maybe
below it. Depending on one’s perspective, one could call this type a “highest
Foundation” enigma, or a “turtles all the way down” enigma. In the first case,
I suggest that readers keep in mind a repeatedly asked and resolutely unre-
solved question in the Mahābhārata: What is the highest dharma? And in the
second I refer to an oft-cited “Indian story” about an Englishman who asked
what was below the turtle, who held up the turtle who held up the world, and
was told, “Ah, Sahib, after that it is turtles all the way down.”20 Either way,
although it is couched in circularity and reflexivity, this is a vertical enigma.21
In contrast, circularity and reflexivity are the hallmark of case two where the
poet doubles back on his own inspired thought and finds its dhárman or
Foundation in poetry. We could call this type a “sources of dharma” enigma,
where dharma’s source and unfolding is located in the searching yet regu-
lated minds of the learned: the Vedic poet-Sages or Ṛṣis. As we shall see, this
kind of dharma lives on in many surprising places where ontologies of open-
ness, epistemic strategies of open-endedness,22 and ethics of friendship and
hospitality continue to thrive.
Other enigmatic verses on dhárman focus more on conundrums of ritual:

Being kindled according to the first Foundations (prathamāʹnu

dhármā), he is anointed with unguents—he that fulfills all wishes, /
the flame-haired, ghee-cloaked, purifying Fire, who makes the
sacrifice good—for the sake of sacrifice to the gods.

20. See Geertz’s essay, “Thick Description: Toward an Interpretative Theory of Culture,” in Geertz 1973,
28–29. See Halbfass 1988, 61 and 473 n. 39: the image seems to have gained circulation from John Locke’s An
Essay Concerning Human Understanding II, 13, 19. According to an e-mail circulated on the INDOLOGY list-serve
by Will Sweetman (April 2, 2010), it is “likely that Locke got the image from Samuel Purchas, in whose account
of his pilgrimage (1614, 501), he writes that some Hindus . . . said, that the Earth was borne up by sevcn ele-
phants; and the Elephants’ feet stood on Tortoises, and they were borne by they know not what.” According to
Sweetman, this report may go back “to a letter by the Jesuit Father Emanual de Veiga (1549–1605), written at
Chandagiri on 18th September, 1599.”
21. Brereton reads this verse vertically: the Aṅgirases “have ‘supported,’ or given foundation, to the truth”
(ṛta), which is the sacrifice itself as “the template and ultimate basis for the world. Since it is the basis for the
world, the truth that is the sacrifice is itself a ‘support.’ Moreover, the Aṅgirases supported this truth ‘by means
of the truth,’ that is, by means of the hymns they sang” (2004, 451). Brereton calls attention to ṚV 10.170.2 in
which dhárman and dharúṇa are again both governed by one divas (genitive, “of heaven”) in a context that speaks
of “the support and foundation of heaven in which the sun is ‘embedded’” (452).
22. See Halbfass 1988, 317: “The idea of a primeval opening, separation, holding apart is of extraordinary
importance in Vedic cosmogony,” with later oft-forgotten developments, which, in 1992, 29–32, he traces to a
Ṛgvedic “ontology of openness” that would have priority over an “ontology of substance.” See further Kuiper
1983, 66–89, 130–58; Brereton 2004, 481 on the primal opening between heaven and earth in ṚV 6.70.1; 1999,
258–59 on the openendedness and enigmatic riddle structure of ṚV 10.129 versus the drive to resolve it and bring
it to closure in TB 2.8.9.3–7.
62 dharma

The hotar-priest who is before (hótā pūrvo), o Fire, who performs

sacrifice better, who sits now, as before, and is luck-bringing by
nature—/following his Foundations (tásyāʹnu dhárma), set forth the
sacrifice, o you who are perceptive, and establish the rite for us in our
pursuit of the gods. (ṚV 3.17.1, 5; Brereton trans. 2004, 460–61)

This is one of only two ṚV hymns (the other, 5.63, will be discussed in the next
section) that refer to dhárman in both its first and last verses. Rather than the
highest foundation, one learns here about the first foundations according to
which Agni was first kindled as the prototype for continued sacrifices performed
by “the hotar-priest who is before,” whom Brereton finds among the “puzzling”
things of this verse, since this hotar could be the current one seated before one of
the ritual fires, or Agni himself (2004, 461). Similarly, ṚV 5.51.2 exhorts Agni:

You whose insights are truth (.ŕtadhītaya), come here. You whose
foundations are real (sátyadharmāṇo), (come) to the rite. Drink with
the tongue of fire. (Brereton trans. 2004, 465)

The sobriquet “you whose foundations are real” can relate a grounding in satya,
“the real,” to any number of paradoxes, as is borne out further in later hymns
where it is used not only for Agni (1.12.7) but the unnamed god “Who” (10.121.9),
for the dice in comparison with Savitar (10.34.8) as the god whose “Foundation”
(4.53.3) is that he “compels,” and likewise for Sūrya as he is being identified
with Savitar (10.139.3; see Brereton 2004, 465, 470, 473–74).
Although we will not find that early enigmatic verses make much of a link bet-
ween dhárman and kingship with reference to the Ādityas or Indra, with Agni and
Soma we do meet ritual-based and poetry-inspiring deities who can also be kings.
Agni, for instance, is called the “undeceivable king of the clans and overseer of the
Foundations” (8.43.24), and is praised as a kavi, a “sage” or “poet,” himself:

Praise the sage Fire, whose Foundations are real at the rite (kavím agním
úpa stuhi satyádharmāṇam adhvaré). (RV 1.12.7; Brereton trans. 2004, 465).

According to Brereton, “Here the ‘real Foundations’ are ritual foundations that
Fire creates at the ritual performance” (Ibid.)—indeed, Agni is praised as a
“poet-sage” whose “dhármans are real” in the poetry he inspires. Agni also
causes ‘foundations’ to prosper—typically material goods (466)—beginning
with things here below, but he is also messenger to the gods above:

Being fully kindled, o Fire who conquers thousands, you made the
Foundations (dhármāṇi) thrive, as the praiseworthy messenger of the
gods. (ṚV 5.26.6; Brereton trans. 2004, 466)
 a vedic history of dharma 63

Moreover, in that Agni “assumes the form of the universal fire, the Sun,”
he “becomes the foundation for all things” (Brereton 2004, 452):

The pleasing oblation and drink is poured here in Fire, who finds the
sun and touches heaven. / For him to bear the living world, and yes,
to give it Foundation (dhárman) in conjunction with his own will, the
gods will extend themselves. (ṚV 10.88.1; Ibid.)

Now, as Brereton says regarding this verse, “Like Fire, so also Soma
supports heaven and earth,” citing ṚV 9.86.9. But as 9.86.8–9, which I will get
to shortly, indicates, Soma’s “ritual foundations” are rather different from
Agni’s. For one thing, Agni may once be linked with dhárman as “king of the
clans.” But he is far more frequently linked with dhárman, and typified more
generally, as one or another kind of priest (hotṛ, purohita, etc.; see Macdonnell
[1898] 1974, 96–97). Soma’s connections with dhárman are more typically
linked with kingship. Sometimes this occurs in “contexts involving Varuṇa or
vratá,” which “are suggestive of Varuṇa” (Brereton 2004, 483). Thus in 10.167.3,
“ritual consumption of soma occurs on the ‘foundation’ (dhármaṇi) of king
Soma and Varuṇa” (478); and in 9.35.6, Soma as “the lord of foundation
(dhármaṇas pátiḥ)” possesses royal “command” (vratá) (476; cf. 9.64.1).
Sometimes Indra is the beneficiary, if not explicitly as a king, of the soma drink
as a “foundation.” Thus at 1.55.3, Indra becomes foremost among the gods by
drinking soma, the “Foundation of great manliness”; at 10.44.1, he “dominates”
by the “foundation” and bull-likeness he gets from drinking soma; at 10.50.6,
soma-pressings in specific detail are made thick for his “foundations.” And in
the verses by which Brereton compares Soma with Agni, it is a matter of King
Soma himself:

King (Soma) hides himself in the ocean23 (and) rivers; he follows the
flood of the waters, when placed among the rivers. / (Soma) has
mounted the woollen back (sāʹnu) as he purifies himself, as the
support (dharúṇa) of great heaven on the navel of the earth.
Thundering like the back of heaven (divó ná sāʹnu stanáyan), he
has cried out, by whose Foundations (dhármabhiḥ) heaven and earth
(have foundation). / He purifies himself, rediscovering again and
again his partnership with Indra (índrasya sakhyám). Purifying
himself, Soma sits in the vats. (ṚV 9.86.8–9; Brereton 1981, 124 for
the translation of 8ab, and the rest from Brereton 2004, 452–53).

23. See Brereton 1981, 124 on samudrá here: the “ocean” as all waters, into which soma mixes, and as the
source of rain.
64 dharma

Soma is the support (dharúṇa) of heaven “on the navel of the earth” and the one
“by whose Foundations (dhármans) heaven and earth (have foundation),” who,
in “rediscovering again and again his partnership (sakhyám) with Indra,” is
linked with dhárman not only in line with the foundation of heaven and earth
“buttressed apart” by Varuṇa (6.70.1) but in harmony with the partnership or
friendship of Viṣṇu with Indra that is made “according to the foundations of
Mitra” (8.52.3).24 Indeed, note again the juxtaposition between dharúṇa as a
concrete “support” from “the navel of the earth,” which recurs at 9.72.7a
(Geldner 1951, vol. 3, 79), and the more elliptical25 usage and more ambiguous
sense in the unique mention of dhárman.
It is thus clear that it should not be enough to say that Soma’s or Agni’s
foundation is “in the ritual” or “in the sacrifice,” as Brereton does frequently,
without recognizing that the “ritual foundation” of Agni and especially Soma26
is often itself the subject of well-wrought enigmas. With this in mind, I would
like to close this section by following up the thundering cries that Soma makes
“like the back of heaven” having “mounted the woollen back as he purifies him-
self” on the way to the soma vats. As Brereton says, “The verse describes Soma’s
pressing and purification through the woollen filter, the ‘back of heaven.’27 The
theme of the presence of soma throughout the universe dominates this hymn”
(2004, 452). Such matters are further amplified in three verses of 9.97, the last
two of which have been cited in other connections, but not looked at together:

As he purifies himself, he purifies himself in the direction of the

things dear to him—he the god that fills the gods with his own
juice. / The soma-drop, clothing himself with his Foundations
(dhármāṇi) following the ritual sequence, has enwrapped himself
in the ten fingers on the woollen back (sāʹnu).

When the speech from the thought (mánasas) that is tracking him
fashions (Soma) on the Foundation (dhármaṇi) of the foremost
(thought),28 or in the face of the herd, then the cows, bellowing as
they wished, came to them, their delighted husband, the soma-drop
in the vat.

24. As at 8.52.3 for mitrá, Geldner 1951, vol. 3, 79 has “Freundschaft” for sakhyám at 9.86.9. On Vedic
Viṣṇu, see also Falk 1987.
25. Brereton 2004, 486, n. 5 says that with the “verbal gapping” here, dhárman “suggests a form of ÷dhṛ.”
26. Brereton 2004, 454–59 notes uncertainties, paradox, likely double meanings, and at least triple possi-
bilities where dhárman is mentioned in the Soma hymns of book 9. Cf. Pollock 1997, 402 for a rather too narrow
ritualist interpretation of Vedic usages.
27. Citing Oberlies 1999, 154.
28. See Brereton 2004, 486 n. 11: “the completion of the ellipsis rests principally on the nearby mánasas”
(“from the thought”), and also on his reading of 1.136.1ab.
 a vedic history of dharma 65

The divine giver of drops, sweller of drops, (goes) forth. As the

truth and for the truth, the very wise one purifies himself. / He will
become the Foundation-giver (dharmāʹ), the king of what belongs to
the community. He has been brought forward toward the world by the
ten reins. (ṚV 9.97.12, 22–23; Brereton trans. 2004, 455, 456, 476)

Clothing himself in his foundations, which are the woollen filter that is itself the
back of heaven through which his drops descend into the soma vat, soma is “tracked”
by “speech” that is itself “from thought” on the “Foundation of the foremost
(thought).” Passing through the ten fingers of the priest, soma, as king, thereby
“becomes the Foundation-giver” of what belongs to the community. Again, we see
both verticalities and reflexive circularities, and, more than this, a convergence of
the two types of enigma mentioned earlier. For as Soma descends from the
foundation above to give foundation below, he is tracked by inspired sacred speech
that hangs on the bottomless foundation of thought in the form of mantras.29
Thus, where the “ritual foundations” of Agni, who works mainly from below,
and Soma, who descends mainly from above, are concerned, dhárman relates to
kingship and priesthood not as a matter of force majeur, as it does with the
Ādityas or Indra. What is registered is the toil, fragility, mystery, and sheer poetic
delectability of building a house on Fire and tracking liquid “foundations” from
far above the clouds. As a memorable phrase from the next phases of Vedic liter-
̣ 30
ature will put it, “The gods love what is out of sight (parokṣapriyā devāh).”
Indeed, enigmatic associations of dhárman proliferate in the late books in
connection with the theme of ritual “foundations.” The oft-cited verse at ṚV
10.90.16—beginning, “With the sacrifice the gods sacrificed the sacrifice: these
were the first Foundations, and those, its greatnesses, follow to heaven’s
vault . . .”—is a case in point. As will be noted in the next section, the verse,
enigmatic enough as it stands in the Puruṣasūkta, seems to be borrowed from
the famous Asya vāmasya “riddle hymn.” In that hymn, if Jan Houben is right,
the identical verse ṚV 1.164.50 may be contextualized not only as one of the
hymn’s enigmas in which the “greatnesses” reaching up to heaven’s vault could
refer to the column of steam and flames that rises from a heated pot (the gharma
pot) in the preclassical pravargya ritual but also linked as well to a prior
verse 43 that ends with the phrase “. . . these were the first Foundations
(dhármāṇi).” Here in this riddle hymn, this first reference to the “first dhármans”
seems to enigmatize the preheated pot itself (Houben 2000, 523–25, 536).

29. Cf. Brereton 2004, 456: whereas verse 22c describes the mixing of milk with soma, “The rest of the
verse is about the mixing of a different ingredient in soma’s creation, namely the recitation of the hymn. It is this
thought of the seer which provides the foundation for soma.”
30. Frits Staal’s happy translation (e.g., 1989, 60). See also BĀU 4.2.2.
66 dharma

This concluding line or pāda of 1.164.43 is then repeated in the pādas that con-
clude verse 50 of the same hymn and the identical verse at 10.90.16. In brief,
ṚV 10.90.16 appears to be an enigma built on an enigma.
Still, as we have seen, dhárman also has reference to more empirical “founda-
tions.” One can also track into verses on the “foundations” of King Soma themes
and formulas that are “inherited” from the Ādityas, and above all from Varuṇa.
This notion that deities “inherit” traits and descriptive formulae from other deities
accounts for a recurrent feature of Vedic poetry (see Brereton 2004, 481). On
these matters, Brereton’s down-to-earth conclusions are illuminating. Taking up
the “difficult question” of how dharma inherits the “command” “functions” of
“the close connection between vratá and dhárman,” while “the word vratá itself
becomes circumscribed to a ‘vow,’” he posits that it “may reflect the changing
nature of the state during the Vedic period.”31 Dhárman would thus tie together
the meaning of “royal and foundational authority” with “concreteness and legiti-
macy” to the meaning of “a physical foundation,” pointing toward the “universal
application” of the later dharma, while vratá, “resting on the personal authority of
kings and sovereign gods,” would basically be individualized. Thus “dhárman may
have become the anchor to a broader claim by rulers.” Brereton also discusses
how the “semantic space” of early Vedic ṛtá (truth) is occupied by both dhárma and
satya (real, true), with dhárma becoming linked more with sovereignty than ṛta
was. “Thus a growing authority of the king may have made dhárma a seemingly
more realistic description of the governing principle of the world” (2004, 484). As
these points indicate, Olivelle’s hypothesis that one may begin a “semantic history
of dharma” with Varuṇa thus remains sound.

C. Dhárman and Ṛgvedic Kingship

Let us then continue to trace threads linking dhárman with kingship in the Ṛgveda,
starting again with passages from ṚV Book 5. Brereton cites five verses from
Book 5 hymns that speak of dhárman in connection with Varuṇa or Mitra. These
two sovereign deities are often paired as the foremost among the early class of
Ṛgvedic gods known as the Ādityas. Brereton puts considerable emphasis on
Varuṇa’s associations with the word vratá, which he has long argued (see Brereton
1981) should be translated primarily as “command” or “commandment” in its
early usages, and Mitra’s associations with the word mitrá, meaning first of all
“contract.” These associations, he says, make their “fundamental natures” (as one

31. As Greg Bailey points out, “This begs the question of a ‘state.’ By Witzel’s views [see above] this
meaning is very late” (personal communication, December 2006). I assume Brereton refers to usages from the
late Ṛgveda on, as discussed.
 a vedic history of dharma 67

meaning of dhárman) “transparent” in their names and “give color to the more
neutral dhárman,” so that “dhárman becomes ‘the foundation of authority that
structures society.’ ” This leads him to “an explanation, or at least another nuance,”
of dhárman as it relates to these two gods: “their ‘foundation,’ that is, their nature,
is to represent” the “foundation of authority that structures society” (477). The
poets can thus refer to both their “foundational authority that orders the worlds”
and the “foundational nature” of the Ādityas themselves. Brereton suggests that
these two sides of dhárman are present in the following verses:

Herdsmen of the truth (.ŕtasya gopāv), you two stand upon your
chariot, o you whose Foundations are real (sátyadharmāṇā), in
the furthest heaven (paramé vyòmani).

In accordance with your Foundation (dhármaṇā), o Mitra and

Varuṇa, who perceive inspired words, you two guard your commands
through the craft of a lord (vratāʹ rakṣethe ásurasya māyáyā). / In
accordance with truth (ṛténa), you rule over the whole living world.
You place the sun [Sūrya] here in heaven as your shimmering
chariot. (ṚV 5.63.1ab, 6; Brereton trans. 2004, 477)

By your command (vraténa), you two are those that give peaceful
dwellings that endure, assigning places to the people according to your
Foundation (dhármaṇā). (ṚV 5.72.2ab; Brereton trans. 2004, 478)

Taking up the two nuances just mentioned, Brereton notes that in 5.63.7,
dhármaṇā, “in accordance with your Foundation,” is set in parallel with ṛténa,
“the truth that expresses the right organization of the world,” and draws from
this that dhárman could signify both the “foundation” according to which
“Mitra and Varuṇa guard the commands which keep the world in order” and
“the foundations of Mitra and Varuṇa as the embodiments of the authority to
govern” (477). It may also be noted that, as the first and last verses of this
hymn, 5.63.1 and 7 round out the hymn around images of truth, the real, poetic
inspiration, the solar chariot, and the asuric lordship of the two gods. As to
5.72.2ab, just cited in connection with “peaceful settlements” (kṣema), Brereton
notes that assigning such “places to the people” is more associated with Mitra;
so together with vratá, the verse “again suggests that the dhárman according to
which they act is both their foundation as well as the foundational authority
they apply to the world” (478). We should note that the word that describes
their divine lordship is asura, and that their “assigning places to the people
according to your dhárman” makes no mention of kings of those people having
any resultant or corresponding dhárman.
68 dharma

Let us also note that ṚV 5.69.1 describes Mitra and Varuṇa’s “giving
foundation” to multiple heavens:

The three realms of light, Varuṇa, and the three heavens, the three
airy spaces do you two give foundation [dhṛ], Mitra, / having grown
strong, protecting the emblem of the ruler (amátiṃ kṣatríyasya), in
accordance with his unaging command (vratám . . . ajuryám).
(Brereton trans. 2004, 488 n. 54)

Says Brereton, “Here Mitra, together with Varuṇa, ‘gives foundation’ [dhṛ] to
the three realms of light and the other heavenly spaces”—this, however, in a
verse using the root dhṛ but without the noun dhárman. As I shall note later in
this section, Brereton usually notices this “foundations above” theme, but
seems overall to underplay it. While “giving foundation” to the heavens, the
pair also protects “the emblem of the ruler,” which I take to be the sun as the
emblem of their own heavenly rule (kṣatriya can mean “sovereign” in the ṚV),
and not a standard of human rulers.
Beyond these Atri poems, ṚV 6.70.1–3, a hymn to Heaven and Earth, is
interesting for two usages of dhárman. Verse 1 tells that “Heaven and Earth are
buttressed apart according to the Foundation of Varuṇa (dyāʹvapṛthivīʹ váruṇasya
dhármaṇā vískabhite).” Verse 3 reveals that it is “from the Foundation”
(dhármaṇas pári) of these two worlds that humans propagate, thereby
continuing to imply from verse 2 that the dhárman of Heaven and Earth, by
which these primal parents “rule over this living world,” is the fluids they cir-
culate, among them “the semen . . . which was established in Manu” (trans.
Brereton 2004, 481). According to Brereton, the mention of vratá in verses 2
and 3 means that “Heaven and Earth inherit the characteristic dhárman of
Varuṇa, the authority that here ordains the continuation of the sacrificer’s
line” (481, my italics). Like Heaven and Earth, King Varuṇa presides over the
circulation of waters that bring earthly abundance through good rains (Brereton
1982, 108–11, with several citations in Book 5). We note that in ruling by this
“inherited” dhárman, Heaven and Earth continue a correlation between dhár-
man and kingship that has little if any direct relation to human kings.
Finally, I would like to note verse 7.82.2, from one of the Young Family
Books, which, according to Dumézil, draws “a distinction which has the value
of a theological definition” (1969 62 n. 18): Whereas Indra is svarāʹj, “ruler of
his own (domain),”32 Varuṇa is saṃrāʹj, “universal sovereign.” While neither of
these terms is used to link Indra or Varuṇa with dhárman in the Family

32. Thanks to Jarrod Whitaker (personal communication) for suggesting that I translate svarāʹj in this
fashion, drawing on Schlerath 1960, 132–35, rather than as “king by himself,” which Dumézil prefers.
 a vedic history of dharma 69

Books,33 Brereton calls attention to ṚV 10.65.5 as a key verse for linking the
universal kingship of Mitra and Varuṇa with dhárman:

Strive for the sake of Mitra and of Varuṇa who acts dutifully, for
them, the universal kings (saṃrāʹjā) who, through their thought
(mánasā), are not far away, / whose dominion (dhāʹman) shines aloft
according to their Foundation (dhármaṇā), for whom the two worlds
are twin need and twin course. (Brereton trans. 2004, 478)

Brereton points toward interpreting dhāman, “dominion,” here as being “the

whole heavenly sphere over which Mitra and Varuṇa rule and to which they
give foundation by their dutiful action as gods of alliance and commandment”
(Ibid.). If, however, dhárman is linked with the universal kingship of Mitra and
Varuṇa, it is never connected in the Ṛgveda with Indra’s being svarāj. This sug-
gests the possibility of three developmental stages.
First, although both terms are used in the early ṚV (for Indra as svarāj, see
3.45.5; for his rambunctious warrior bands, the Maruts, see 5.58.1; for Varuṇa
as saṃrāj, see 5.85.1), neither saṃrāj (which is a bit curious) nor svarāj was felt
to have any special rapport with dhárman. Second, verse 7.82.2, from the Young
Family Books, would register an attempt to define a rapport between Varuṇa’s
and Indra’s two types of divine kingship, but still without reference to dhár-
man. Then in the late ṚV, verse 10.65.5 would be a belated attempt to say that
Mitra and Varuṇa’s saṃrāj had been “foundational” all along. With this, now
that Varuṇa’s being saṃrāj is related to dhárman but not in any rapport with
Indra, one may hazard the impression that Indra will not “inherit” whatever
associations he may come to have with dhárman primarily from Varuṇa. So far,
on the basis of what we find in the early ṚV, there appears to be no attempt to
link dhárman with the “independent” and exuberant Indra, which may suggest
that Indra was more the model of the earthly chieftain than Varuṇa, and like
earthly “chiefs,” yet to be reined in as a specific target of this concept.
In this context we may note one more verse from Book 7:

Whatever this deceit that we humans practice against the race of

gods, Varuṇa, / if by inattention we have erased your Foundations
(táva dhármā), do not harm us because of that misdeed, o god. (ṚV
7.80.5; Brereton trans. 2004, 479)

Horsch gives a translation that could suggest that Varuṇa here would be the
projection of a human king: “O Varuṇa, if we have violated your laws out of

33. Indra is also called saṃrāʹj at 4.19.2: “the universal king whose origin is the truth” (saṃrāʹṭ satyáyoniḥ).
See Sohnen 1997, 236.
70 dharma

ignorance, may you not bring us to harm on account of those sins, O god”
(2004, 430 [my italics]). But reading dhárman in the plural as “laws” is at this
point as “anachronistic” as reading it as “ordinances” in the even later verse
10.90.16 (see Brereton 2004, 467). There is no evidence in the early ṚV that
human kings were considered “universal sovereigns” like Mitra and Varuṇa.
Now a good way to transition from the earlier to the later ṚV on this matter
of the relation of dhárman to (primarily divine) kingship is to get our bearings
from the beginning of Brereton’s closing section on “Dhárman as the Foundation
Created by a Sovereign Deity” (475–84), and the beginning of his concluding
summary of that same section. In the first case, he begins: “The last sense in
which dhárman is a ‘foundation’ is the most significant, for it is on this sense
that much of the later development of dhárman and dhárma is established.
A dhárman can be the ‘foundation’ through which a sovereign deity upholds the
life of a community. This foundation can be the material basis for the community,
or it can be prescribed behaviors and social relations which sustain the structure
of the community. In the latter use, it is the sovereign’s ruling ‘authority’ or
‘institute’—and in these ways it may often be translated—upon which the life of
a community depends.” Brereton suggests that the latter usage applies solely to
Varuṇa or Varuṇa together with the Ādityas, and the former solely to Soma, with
whom one finds “some of the clearer instances of the link between dhárman and
a ruler” in which the “foundation” that his “rule provides is likely the material
foundation for the community” (2004, 475). It will suffice to cite two of the four
verses on Soma that Brereton discusses in this connection:

Under whose command (vraté) every people finds foundation

(víśvo . . . jano dādhāʹra), under that of the lord of Foundation
(dhármaṇas páteḥ), who is purifying himself, who brings the foremost
good things. . . . (ṚV 9.35.6; Brereton trans. 2004, 475)

The divine giver of drops, sweller of drops, (goes) forth. As the truth
and for the truth (ṛtám ṛtāʹya), the very wise one purifies himself. /
He will become the Foundation-giver (dharmāʹ bhuvad), the king of
what belongs to the community (vṛjanyàsya rāʹjā). He has been
brought forward toward the world by the ten reins. (ṚV 9.97.23;
Brereton trans. 2004, 476)

In the first of these two verses, Brereton takes “every people” to imply
community that “finds foundation”34 under Soma as “lord of Foundation” in

34. According to Brereton, dādhāʹra is an “isolated use of the active perfect of ÷dhṛ as intransitive” (2004,
488 n. 49).
 a vedic history of dharma 71

the material “good things” that Soma brings (475). In the second, Soma “is the
king who is the Foundation-giver, for he governs what belongs to the community,
that is, its wealth” (476). Brereton notes that the sense of vṛjanyà is uncertain
since it only occurs here in the ṚV, but an earlier verse in the same hymn
“provides an indication of its sense. Here Soma, as lord of the community,
conquers the land and thereby gives the people the space to live”:

He strikes down the demon, and he presses away hostilities on every

side—he who, as king of the community (vṛjánasya rāʹjā), creates
expanse (várivaḥ kṛṇván). (ṚV 9.97.10cd; Brereton trans. 2004, 476)

Says Brereton, “Soma’s kingship is connected to his ability to give his people
the means to raise and pasture their cattle,” which are mentioned just before
the main verse cited here. As has been discussed, some of the Soma hymns
collected in Book 9 may be earlier than their place in the expanding ṚV collec-
tion process would reflect. However, whether these verses are early or not,
while they make a correlation between King Soma’s dhárman and the
community of “every people” (perhaps referring just to every Ārya people), they
make no such correlation explicitly with the kings or chieftains of such peoples.
In Book 8, Agni is still called the “undeceivable king of the clans and overseer
of the Foundations (viśāʹṃ rāʹjānam ádbhutam ádhyakṣaṃ dhármaṇām imám)”
(8.43.24; Brereton trans. 2004, 465).
This brings us to our second transition point from earlier to later ṚV:
Brereton’s concluding summary on the relation between dhárman and divine
kings: “This sense of dhárman as ‘foundational authority’ is a critical source for
the later development of the concept of dhárma, and in considering this aspect
of dhárman, several points relevant to the history of dhárman and dhárma
emerge. First, dhárman implies not just ‘foundational authority’ but ‘royal
authority.’ This facet of its meaning is indicated either by the direct description
of the gods that act in connection with dhárman as kings . . . or by attributing
commanding authority to them” (2004, 482–83). For the first sense, he cites
only one of six of his instances from the earliest ṚV, that of the asuric Mitra and
Varuṇa in 5.63.7. And for the second, three of his six instances come from ear-
lier portions: one where Mitra and Varuṇa assign peaceful enduring dwellings
to the people (5.72.2); a second where Varuṇa buttresses apart Heaven and
Earth (6.70.1–3); and one where men seek atonement for erasing the asura
Varuṇa’s foundations (7.89.5). The first two have been discussed, and none of
them offer us anything further on the “sense of dhárman” as “royal authority”
as “a critical source for the later development of the concept of dhárma.”
That must come from the later ṚV, from which Brereton’s remaining cita-
tions on these meanings derive. These include, in the first meaning, five verses
72 dharma

on “gods that act in connection with dhárman as kings”: Soma (9.97.23), already
described; Mitra and Varuṇa as universal kings (10.65.5), already described;
Soma and Varuṇa (10.167.3); Soma ruling by the power of Mitra and Varuṇa’s
dhárman (9.107.15f.); and asuric Wind assisting Soma (1.134.5). For the second
meaning, three verses attribute “commanding authority” to gods: Soma bringing
̣ giving foundation
goods (9.35.6), already cited; Soma as a bull (9.64.1); and Viṣnu
by his three strides (1.22.18f.). Be it noted that Brereton does not mention ṚV
10.44.1, 10.90.16, or 8.52.3 in this tally, all of them important verses that need to
be discussed as well in the context of changing definitions of the relation between
dhárman and “royal authority.” Since it will not be possible to range over all these
verses, I will look at the three just mentioned from the vantage point of Brereton’s
discussion of them and the other later ṚV passages just mentioned.
Clearly Indra has emerged for us already as a problem for the future history
of dharma. ṚV 10.44 is mainly about him:

As the lord of his own (svápatir), let Indra journey here for his
invigoration (mádāya)—he, the vibrant, who thrusts forward
according to his nature (/‘Foundation’) (dhármaṇā), / who
energetically dominates over all strengths according to his
boundless and great bull-likeness.

Let good things go among us, for I hope for them. Journey here to
the soma-bearer’s stake, which carries his good expectation. / You
[Indra] are master (tvám īśiṣe). Take your seat here on this sacred
grass. Vessels which belong to you are not to be claimed (by another)
according to your [nature/character] Foundation (dhármaṇā). (RV
10.44.1 and 5; Brereton trans. 2004, 472; brackets added as per
discussion below)

Brereton probably does not mention these verses in his closing section on
“Dhárman as the Foundation Created by a Sovereign Deity” (475–84) because
Indra gets his invigoration by Soma and the verses do not call him a king.
Nonetheless, verse 5 invites Indra, “You are master (tvám īśiṣe),” to sit at the
soma sacrifice and to receive vessels that are solely his “according to his dhár-
man (nature, foundation),” on which Brereton says, “It would be reasonable,
then, if the dhárman is that principle according to which the soma cups belong
to Indra and to Indra alone in his foundational nature, his very character as
Indra. Alternately, these vessels may be Indra’s according to their foundation,
that is, according to their place in the ritual.” Moreover, it is by Indra’s own
bull-like dhárman that he dominates and masters, which we already know he
 a vedic history of dharma 73

does as one who is “king of his own (domain)” or svarāj. Brereton notes that
Renou and Geldner translate dhárman frequently (and more often than he) by
its meaning “foundational nature,” and he agrees that this passage is one that
confirms such a meaning. It is thus about Indra as “the lord of his own
(domain)” (svápatir), and “twice speaks of the ‘foundation of Indra,’ the character
that defines his action.” But let us note that although Brereton finds a parallelism
in which dhárman and “bull-likeness” both pertain to Indra and “define who
and what he is” (472), there is no such parallelism between dhármaṇā, the
foundational “nature” by which Indra is vibrant, and his being svápati, “lord of
his own (domain).”35 That is, if it is his bull-like “nature”—his machismo—that
makes him powerful, that enables him to be svápati, his “own-nature” could
not be taken exactly as his sva-dharman, since his invigoration comes from, that
is, has its foundation in, the ritual, the laud, soma, etc.36
In any case, it would seem that ṚV 10.44.1 and 5 deserve mention along
with other verses that are concerned with bringing Indra, however indirectly,
into the sphere connecting dhárman with kingship. Brereton does mention
10.167.3 as among the late verses making such a link, but primarily in connec-
tion with Soma and Varuṇa:

Upon the Foundation (dhármaṇi) of king Soma and Varuṇa, and

under the protection of Bṛhaspati and Anumati, / today, at your
praise, o generous one (Indra), I consumed vats (of soma), o you that
set in place and that set apart (dhāʹtar vídhātaḥ). (ṚV 10.167.3;
Brereton trans. 2004, 478)

Noting that it is rare to see Soma and Varuṇa mentioned together, Brereton
says that their “designation as kings implies that the dhárman of Soma and
Varuṇa is their royal authority. And perhaps, their appearance together reflects
the complementary sides of their dhárman: Soma establishes material
foundation, Varuṇa social foundation” (2004, 479). Here it is interesting that
it is upon the foundation of Kings Soma and Varuṇa that the soma-drinking
ritualist praises Indra as Dhātṛ and Vidhātṛ, nicely translated as “you that set in
place and that set apart”! These two names are used together only one other
time in the ṚV: for the “All-Maker” Viśvakarman at 10.82.2. And where they
are used singly, they occur mostly in the late books (19 of 21 times for dhātṛ;
4 of 6 times for vidhātṛ). This, however, is the only verse in the ṚV that connects
“setting in place and setting apart” with dhárman. It could prove significant that

35. Or “his own master” (Dumézil 1969, 62 n. 18). Dumézil notes that two of the ṚVs three usages of
svápati pertain to Indra (1969, 62 n. 18).
36. The only occurrence of svadharman in the ṚV (see above n. 3), at 3.21.2b, would seem to ask that Agni from
“his own foundation”—no doubt the ritual fire—bestow (÷dhā) what is best “for us” that is acceptable to the gods.
74 dharma

it makes such a connection with regard to a continually active god like Indra,
who probably “inherits” this tag from Viśvakarman, who “sets in place and sets
apart” only in his primal cosmogonic work as the divine architect of the uni-
verse. As we shall see in later chapters, in the Mahābhārata the terms dhātṛ and
vidhātṛ take on a life of their own in association with dharma—and often enough
in relating dharma with bhakti.37
This brings us the Puruṣasūkta, which, as Bailey (1983a, 142) has noted,
uses the verb vi-dhā to describe how the gods divided (vyadadhuḥ) Puruṣa, the
cosmic “Male,” when “they created the physical features of the cosmos and the
four varṇas from the parts of his body” (ṚV 10.90.11–14). It is just after this
“division” or “setting apart” of this victim’s portions that the hymn speaks of
plural dhármans in its last verse (16).38 It can be noticed that verses mentioning
dhárman usually occur early in a hymn, often as the very first stanza. This
occurs thirteen times, and in two of these cases, both from the earlier ṚV, dhár-
man is mentioned in the first and last verses only: in 3.17.1 and 5, discussed in
the last section, and in 5.63.1 and 7, already cited in this section.39 I would sug-
gest that this is because dhárman can itself signal both a poetic “foundation” for
what a hymn says, and something that is to be (or has been) unfolded or unrav-
elled in the hymn. This applies also to four cases where dhárman is mentioned
only in a hymn’s last verse, as in the Puruṣasūkta,40 where its placement would
not be accidental. As noted earlier, the same exact verse also occurs as the fifty-
second of fifty-four in the much more rambling ṚV 1.164 (the riddle-laced Asya
vāmasya hymn), from which 10.90 perhaps lifts it into its much more memo-
rably structured slot.41 The Puruṣa hymn’s earlier verses can thus be reviewed
from the standpoint of the “foundations” proclaimed at its end:

37. See chapter 12. Cf. Halbfass 1988, 317–18, 550–51 n. 35, on other cases of “holding apart”: ṚV 2.13.7;
BĀU 3.8.9 (rather than 2.8.9, as listed). This usage comes in the second of two dialogues in which the bold young
(?) woman Gārgī Vācaknavī finds the right way to get a productive answer from Yājñavalkya after her first ques-
tions about what is woven upon what have misfired (BĀU 3.6); now she comes back to ask him, “The things above
the sky, the things below the earth, and the things between the earth and the sky, as well as all those things people
here refer to as past, present, and future—on what, Yājñavalkya, are all these woven back and forth?” (3.8.7)—
which, it seems, obliges him to finally answer, “the imperishable.” See Hiltebeitel and Kloetzli 2004, 558 for an
interpretation of the two Gārgī passages in sequence. Halbfass notes that this meaning is not forgotten in the
Mbh; see 12.110.11: “the creatures are kept apart, i.e., upheld in their respective identities, by dharma (dharmeṇa
vidhṛtāḥ prajāḥ)—in these cases with ÷dhṛ rather than ÷dhā.” Cf. also Halbfass 1992, 54 on analogous Abhidharma
explanations using dhātu and dharma.
38. The Puruṣasūkta’s uses as a single and whole Ṛgvedic hymn are unusual in later classical Hinduism
(Biardeau 1976, 14–16 and passim; 1991), and must have begun early, to judge from its inclusion in the other
three Vedic Saṃhitās (Witzel 1997a, 283–84).
39. The other eleven that mention dhárman in the first verse are 1.160, 1.187, 3.3, 3.17, 5.63, 6.70, 8.98, 9.7,
9.64, 10.44, and 10.88.
40. The other three instances are 7.89.5, 9.35.6, and 10.175.4. The total of seventeen verses occurring in the
first or last verse is precisely one fourth of the total of verses mentioning dhárman (counting svádharman in 3.21.2).
41. See Bloomfield 1916, 151, considering it secondary in 10.90; Houben 2000, 524–25 and n. 127 in
apparent agreement; Horsch 2004, 444 n. 20.
 a vedic history of dharma 75

With the sacrifice the gods sacrificed the sacrifice (yajñéna yajñám
ayajanta devāʹs): these were the first Foundations (tāʹni dhármāṇi
prathamāʹni āsan), / and those, its greatnesses, follow to heaven’s
vault, where exist the ancient ones who are to be attained, the gods.
(RV 10.90.16; Brereton trans. 2004, 460)

Brereton takes this verse as emblematic of the poets’ concern that dhárman be
a ritual foundation not only for heaven and earth and the gods but “that the
ritual itself have a foundation” (2004, 459–60). Thus, “[t]he ‘first dhármans’
are the model sacrifice instituted by the gods and replicated in human
performance, and as such, they are the ‘foundations’ for the ritual performance”
(460)—itself described in the paradoxical first line that uses three derivatives
of ÷yaj to suggest that Puruṣa is both the sacrificial victim and the means by
which the sacrificial process of the gods who sacrificed him is set in motion
(although Brereton says it might also just mean that the gods “sacrificed again
and again”; Ibid.; cf. Horsch 2004, 428).42 Brereton’s translation “founda-
tions” is especially lucid here, and preferable, as noted, to “ordinances,” “laws,”
or “institutes”; for as he says, in the ṚV “the ritual was varied and fluid” (467).
Yet the hymn also reflects a period when it is getting less so. According to
Witzel, this late hymn provides “the first constitution of India” for the emerg-
ing Kuru state that combined the completion of the ṚV Saṃhitā with the early
collection of the other three Vedas during the post-ṚV “mantra period.” For
among the “first foundations” that the hymn solidifies is “the ‘official’ estab-
lishment of the four classes (varṇa)” of Brahmin or priest from Puruṣa’s
mouth, Rājanya or nobility from his arms, Vaiśya or people from his thighs,
and Śūdra or servant from his feet (verse 12). It thus comes to serve as a charter
for “increasing social stratification,” with cooperation between the armed
nobility and a newly minted class of Brahmins that now cuts across and unifies
the older clans of poet-priests; the “joint power” of these two classes over the
Vaiśya; and further internal varṇa division between Ārya and Śūdra (1997a,
1997b, 267). ṚV 10.90 says nothing direct that would either ground kingship
itself in the “first Foundations,” or the “first dhármans” in kingship. It just
mentions the nobility as a class from which kings would presumably come,
and shows its subordination to the Brahmin class not only by its second posi-
tion but also by saying in verse 13 that Indra and Agni come from the mouth
of Puruṣa just after the Brahmin came from Puruṣa’s mouth in verse 12.43
This priority might suggest that even if Indra and possibly Agni participate as

42. Brereton’s translation of sādhyas as modifying the gods who are “to be attained” may eliminate another
mystery of a class of Sādhya gods prior to the gods.
43. One wants to say the same mouth, but Puruṣa has a thousand heads. . . .
76 dharma

divine kings (which is not clear) in implementing the sacrificial “first

Foundations,” they would do so on the precedent of the sacred speech that the
prior birth of the Brahmin makes possible, and Indra in particular would do
so by association with the purifying ritual Fire. There is nothing in the hymn
about Varuṇa or the “universal sovereignty” (saṃrāj) that links him and Mitra
with dhárman in ṚV 10.65.5. But the elliptical verse 5 says that “Virāj was born
from him (Puruṣa), and Puruṣa from Virāj.” As the feminine principle of this
primal pair, virāj, meaning “Wide Dominion,” may “inherit” the status of
Earth as the foundation (dhárman), along with Heaven, of progeny, which, we
may recall, comes about after Heaven and Earth are “buttressed apart by the
foundation (dhárman) of Varuṇa” in ṚV 6.70.1–3.
Finally, if, with its sixteenth verse, ṚV 10.90 opens a new chapter in the
history of dharma, the one stanza to mention dhárman in the still later and the
apocryphal Vālakhilya hymns may discretely open yet another one:

He (= Indra) who made the solemn words his own, who boldly drank
the soma, / for whom Viṣṇu strode forth his three steps, according to
the Foundations of Mitra (mitrásya dhármabhiḥ). (ṚV 8.52.3; Brereton
trans. 2004, 480)

The hymn celebrates Indra as having drunk the soma with Manu (1a) and as the
beneficiary of Viṣnu ̣ ’s three steps, mentioning that Viṣnu ̣ was acting in accord
“with the Foundations of Mitra.” It further grants Indra the status of “fourth
Āditya” (turīyāditya 7c)—presumably numbering him after Mitra, Varuṇa, and
Aryaman and no doubt imparting a joint sovereign status to him along with the
other three (Brereton 1982, 5) by this association with Mitra. The meaning of
dhárman as “foundations” is again well clarified by Brereton, who explains the
“Foundations of Mitra” in relation to this god’s being the god “contract” (mitrá)
linked with “alliances,” “which refer here to the alliance between Indra and
̣ which is the basis of Viṣnu
Viṣnu ̣ ’s three strides” (480). Yet I believe Brereton
misses two things here: one, in taking Mitra’s “foundations” as the basis for
Viṣnụ ’s three steps rather than just the basis for his alliance with Indra; and
another in taking the three strides themselves as an “image of ascent,” and thus
as “one basis for the occurrence of dhárman here, since it implies the need for a
foundation for that ascent,” even if this is meant “only figuratively” since the
“the real foundation of Viṣnu ̣ ’s ascent is his relationship with Indra” (480). It is
not so evident that Viṣnu ̣ ’s three strides are an ascent or that they need
“foundation.” And it would be more exact to say of Mitra’s “foundations,” which
are in any case in the plural, that, rather than being the “foundation” for the
three strides themselves, they are the basis simply of Viṣnu ̣ ’s contract with
 a vedic history of dharma 77

Indra—or his “friendship” with him,44 this being another and no doubt secondary
meaning of mitrá (Brereton 1982, 15–21, 44–49) but the one where the relation-
ship between Indra and Viṣnu ̣ has, as will now be noted, already been heading.
Fortunately, all this becomes clearer from such earlier hymns as ṚV 4.18.11d
(= 8.100.12a), where Indra asks his “friend Viṣnu ̣ ” (sákhe viṣṇo) to “stride widely”
to open the spaces for the defeat of the demon Vṛtra, and 1.22.18–19ab which
̣ ’s three strides do not require a foundation at all, and especially
shows that Viṣnu
from here below:

Three steps he strode out: he, Viṣṇu, the undeceivable cowherd, /

who gives foundation (dhṛ) to the foundations from there (áto
dhármāṇi dhāráyan) // see the deeds of Viṣṇu!—from where he
watches over his commands (vratāʹni). (ṚV 1.22.18–19ab; Brereton
trans. 2004, 480, slightly modified)

Following Brereton, if “Viṣṇu’s vratāʹni, his ‘commands,’ set this verse within
the context of royal authority and again of Varuṇa” (480), the connection of
Viṣṇu’s three steps with dhárman is again made with one of the two primary
Ādityas, in this case Varuṇa rather than Mitra. Yet although Brereton recog-
nizes that the “‘there’ from which Viṣṇu ‘gives foundation’ is probably heaven,”
he remains caught up with the idea that this is “another passage concerning
Viṣṇu” where “the sense of foundational authority” appears with “the imagery
of . . . physical foundation.” This misses the point that heaven is not exactly a
physical foundation. As Horsch says of this verse, “from the uppermost point,
from the zenith, the ridge of the world-structure, Viṣṇu holds firmly the
dhármāṇi, the fundamental supports, which bestow support and stability on all
the parts of the universe . . . as though the cosmos had been turned upside
down,” calling attention to the inverted tree and other Ṛgvedic and later images
of cosmic inversion (2004, 427–28 and 443 n. 16). Again, Brereton keeps
things down to earth when he says, “In vs. 18, the scene shifts explicitly to
heaven, which is the limit of Viṣṇu’s journey.” True, the prior verse 17 tells that
“Viṣṇu strode out—three times he set down his track—through this (world)
here, / which is drawn together in his dusty (track)” (Brereton trans. 2004,
480). But verse 18, with its shift to heaven, still makes heaven the place from
which Viṣṇu “gives foundation to the foundations” for the three steps that open
up the world for the kingship not of Varuṇa but of his “friend Indra,” with
whom the apocryphal verse 8.52.3 tells us that he “contracts” through the

44. Cf. Geldner 1951, vol. 2, 374, who has “. . . für den Viṣṇu die drei Schritte ausschritt nach den Pflichten
des Freundes” (“. . . for whom Viṣṇu strode out the three steps according to the duties of the friend”) rather than
“. . . according to the foundations of Mitra” for the end of 8.52.3.
78 dharma

“foundations of Mitra.” All this suggests that Viṣṇu’s links with Indra through
dhárman have to do with Mitra and mitrá through the related “friendship” term
sákhya, which defines Viṣṇu’s coming to the aid of Indra with his three steps to
open the space so that Indra, invigorated with Soma, can defeat Vṛtra.
Brereton seems to skip over Viṣnu ̣ , as if he were an unwelcome problem,
perhaps to stick as much as possible with a “Varuṇa context.” But something
more is already going on with Viṣnu ̣ . The terms mitrá, sákhya, and other words
for friendship and alliance open a new semantic arena in which the epics will
work out further implications of dharma where Viṣnu ̣ ’s incarnations or avatāras
concern themselves with matters of alliance, friendship, enmity, self, other, ārya
and non-ārya, and the theme of hospitality which, among other things, will play a
role in the formation of bhakti, which builds on the Vedic theme of honoring a
divine guest (see chapter 12). For now, however, we are also faced with questions:
What could it mean to “give foundation to the foundations?” Or to “give
foundation” or “support” from above? Horsch says the latter idea “seems absurd
to us” and attributes it to “mythical thinking” (2004, 428). I think, however, that
these are among the persistent and well-nurtured enigmas of Ṛgvedic dhárman.

D. Mantra Period and Later Saṃhitā Usages

Moving on into post-Ṛgvedic understandings of dharma, one must posit a

temporal gap (Witzel 1997a, 267) between the ṚV poems and subsequent
“mantra period” texts in which Ṛgvedic mantras, along with new mantras com-
posed in both poetry and prose, were sorted out for expanding liturgical pur-
poses. As Frits Staal puts it, “The Rigveda was created by poets, the other three
by what we would call professors” (2010). These developments were signaled
in the Puruṣasūkta, whose “‘official’ establishment of the four classes
(varṇa) . . . is visible in all Mantra texts” (Witzel 1997a, 267; cf. 274–75). Among
its “first Foundations,” this hymn chartered not only a new class consciousness
of the mouth-born Brahmins but the contents of the three liturgical Vedas,
which, according to its ninth verse, were generated from Puruṣa as well.
In the area of (in modern terms) “the eastern Panjab/Haryana area,” a complex
liturgical system was developed in support of “Kuru dominance and orthopraxy in
the Mantra period” (Ibid., 278–79, 313). The earliest mantra period productions con-
sisted in the systematic application of mainly Ṛgvedic mantras to ritual uses via the
collection of two more “liturgical” Vedas, the Yajur- and Sāma Vedas. A fourth Veda,
the Atharva Veda, more concerned with custom, introduced many new mantras as
healing and sorcery charms and for domestic and some royal rites. With these texts,
we enter into what may be called Brahmanical Hinduism, which we will trace
 a vedic history of dharma 79

further in chapter 5. Each Veda produced specialists whose schools proliferated into
“branches” that could bring Brahmanical ritual into new regions wherever branches
of the three liturgical Vedas could cooperate. This system sets the grounds for what
has been described as Brahmanical “orthopraxy” or “right practice.” The ritual theo-
logians of the four Vedic schools then continue to lay out the performance and
meanings of Vedic sacrifice in prose manuals called Brāhmaṇas. The last Vedic texts
to generate innovative, though not numerous, usages of dharma are then the early
Upaniṣads, which round out the Vedic canon.
For the preclassical history of dharma, this middle-to-late Vedic period, com-
ing between the innovative uses of dhárman in the ṚV and the sudden burst of
interest in dharma in early Buddhism and in both Buddhism and Brahmanism
during the Mauryan period, is surprisingly fallow. What needs to be noticed is that,
aside from a few scattered uses of dhárman or dharma implying custom and rudi-
mentary legal procedure in royal courts, there is cumulative usage in only two
spheres: one, that of the king, where we can see continued development of Ṛgvedic
themes; the other, a new strain concentrating on assertions of Brahmin privilege.
Of post-Ṛgvedic texts, the Atharva Veda [AV] is the earliest to introduce some
distinctive emphases in the area of custom. Being less tied to the ṚV than the
liturgical Sāma- and Yajur Vedas in its hymn-content and to direct use in the
developing śrauta ritual, the AV is informative in suggesting popular usages45
that reflect the developing ritual systematization only obliquely. But its note-
worthy decline in usages from sixty-seven in the ṚV to thirteen allows Olivelle to
introduce his point that, contrary to what one might expect, “dharma was at best
a marginal term and concept within the vocabulary” of middle and late Vedic
texts (2004a, 491). Most revealing is AV 12.1.45, which calls upon the earth that
“bears people in many places, each according to its fixed domicile, with different
languages and various laws (nāʹnādharmāṇam)” (Horsch 2004, 432). Although
Horsch’s translation “laws” is not very secure, we are for the first time in an inter-
esting pluralistic situation where, beside varied modes of residence and speech,
varied dharmas as “customs”46 might be at least a harbinger of a “legal” meaning.
The same would apply where “law” or “custom” is cited in marriage, with “You
are the wife by virtue of dhárman” (14.1.51), and where “a wife is said to lie beside
her dead husband “following an ancient law (or custom) (dhármaṃ purāṇám)”
(18.3.1c; Bowles 2007, 88). One might even detect a recurring theme of dharma
now being articulated in relation to women and the earth. A magical love charm
can still kindle intense burning in a man “according to the foundation of Varuṇa”

45. “Because of its focus on small non-Śrauta rituals, the AV is an irreplaceable source for the material
culture, the customs and beliefs, the desires and sorrows of everyday Vedic life” (Witzel 1997a, 275).
46. Bowles 2007, 87 translates this plural of dharma as “varied . . . custom.”
80 dharma

(váruṇasya dhármaṇā; 6.132).47 We are seeing here usages of both dhárman and
dharma. This departure from the uniform use of dhárman in the ṚV was remarked
on by Horsch (2004, 432, 442), whose main point48 is summarized by Bowles:
“Horsch suggests a semantic reason: while the suffix -man [of dhárman] empha-
sizes the activity of the meaning of the root, the abstract -a ending [of dharma]
accords with the generalized abstract sense of the word as ‘law’” (2007, 88)—or,
by the same token, “custom.” Horsch thus sees the Atharva Veda at the heart of
a change to the more abstract noun dharma, as in the usage of dharma purāṇa,
the “ancient law” just cited in connection with a funerary practice.
Next oldest is the Yajur Veda, the work of Adhvaryu priests who, with their
responsibility of coordinating the use of mantras as yajus or “sacrificial
formulas,” did the most to systematize the śrauta ritual. Their four extant
Yajurveda Collections reflect developments in successive political centers across
northern India, each further to the east,49 and thus bring us back readily to the
subject of the king. Four of their collections are extant: the Kāṭhaka, Maitrāyaṇī,
and Taittirīya Saṃhitās, all of the older so-called Black Yajurveda in which
mantra and commentary are mixed, and the younger Vājasaneyī Saṃhitā in
which mantra and commentary are separated. First, the Kāṭhaka and Maitrāyaṇī
Saṃhitās were composed in the Kurukṣetra area, ca. 1180 to at least 900 BCE,
when the Kuru realm was still at its height in the eastern Punjab/Haryana area,
but now expanding both east and south and beginning to form an alliance with
“the other half of the ‘classical’ Vedic tribal moiety, namely the Pañcāla” (Witzel
1997a, 301). In these first two Yajurveda saṃhitās, the ritual is designed for “the
‘average’ yajamāna” (patron of a sacrifice) as a nobleman, but with attention
also to small chieftains and to the Kuru king, “whose aims are clearly visible in
the development of a ritual for royal consecration, the Rājasūya, which is an
elaboration of the simple abhiṣeka of the Mantra period” (299–300). In the
Taittirīya, however, which begins a frequent use of the compound kurupañcāla
in coordination with the rise of independent Pañcāla kings by about 750 BCE,
one can detect a relocation of the center of orthopraxy eastward toward the
Pañcāla lands—the Gaṅgā-Yamunā doāb (“two river area” or “mesopotamia”)
and surrounding regions of today’s western Uttar Pradesh (301). Here one meets
the emergence to prominence of Pañcāla chiefs,50 new population movements

47. Bowles 2007, 87, but following Brereton’s translation of this phrase in its ṚV usages.
48. Horsch ties this change to a transformation from the “cosmic-ritual level” to the “juridical–ethical
realm” (2004, 432, 442). I agree with Brereton (2004, 485) that Horsch’s overall evolutionary scheme is outdated
and that these terms are not useful to describe the ṚV.
49. See especially Witzel 1989a, 236–49, with maps and charts.
50. According to Witzel 1997a, 302 n. 218, their emergence explains the question asked of Yājñavalkya
about the descendants of the great Kuru King Parikṣit in Bṛhadāraṇyaka Upaniṣad 3.3, “Where are the
Pārikṣitas?”—to which the sage replies, “Indra handed the Pārikṣitas to the wind,” which “carried them to the
place where those who had offered horse sacrifices go” (Olivelle trans. 1998, 83).
 a vedic history of dharma 81

along the rivers and into the interior (but still no indication of large towns or
cities), and Taittirīya subschools extending further east—possibly in relation to
political divisions among the Pañcālas themselves—to the border of the emerg-
ing territories of the Kosalas. There, beginning perhaps from ca. 600 BCE, the
Vājasaneyī project was probably launched among the Kosalas of today’s eastern
Uttar Pradesh, with their kings known as Ikṣvākus, and favored also by the
Videhas, their neighbors to the east in today’s Bihar. These are all names we
shall meet again. By about 500 BCE, these developments would have coincided
with the early phases of India’s second (after the Indus Valley Civilization)
urbanization.
Now where dharma is concerned, the Yajurveda Saṃhitās either illustrate
ways in which older Vedic mantras are applied to specific ritual uses, or offer a
relatively small number of new explications of dharma in prose.51 In these new
usages, one finds three noteworthy developments. First, one meets a new recur-
rent phrase, dhruvéṇa dhármaṇā (“with firm/enduring dhárman”). Olivelle calls
attention to a Maitrayaṇī Saṃhitā passage that offers a rich gloss on the phrase:

Mitra and Varuṇa in accordance with their enduring dharma. Mitra,

indeed, holds it fast (dādhāʹra), and Varuṇa establishes it (kalpayati). For
the upholding (vídhṛtyai)52 of these creatures and for their establishment
(kḷptyai), it is set up. Mitra and Varuṇa are clearly the ones who uphold
dharma among the gods (devāʹnāṃ dhármadhāʹrayau). They have upheld
here the divine dharma (daivám . . . dhármam adīdharatām). (Olivelle
trans. 2004a, 492; MS 3.8.9)

Olivelle notes “that the explanation clearly connects dharma with” ÷dhṛ, and
that such a connection is also evident in Kāṭhaka Saṃhitā 35.7, the first half of
which is from ṚV 10.173.4 and the second half found only in this passage
(Olivelle 2004a, 508 n. 5):

The sky is enduring (dhruvā); the earth is enduring; this whole

universe is enduring. The gods are enduring through dharma, and
the sacrificer (yajamāna) is enduring through the sacrificial animals.
(Olivelle trans. 2004a, 492, slightly modified)

There is an interesting juxtaposition here in that the gods (plural) are to the
sacrificer (single) as dharma (single) is to the sacrificial animals (plural) by
whose sacrifice, understood as dharma, not only would the sacrificer endure
but the gods as well. Such a singular centrality of the sacrificer and dharma

51. Without, according to Olivelle, any noticeable “semantic difference” between dhárman and dharma
(2004, 491). But that would be a matter of degree.
52. Or possibly “the holding apart.”
82 dharma

comes out as well in usage where the phrase dhruvéṇa dhármaṇā is applied to
the planting of the sacrificial post (yūpa) “in accord with the enduring dhárman
of Mitra and Varuṇa.”53 Here, where it is a question of associating dhárman
with a sacrificial post or pillar that is itself an image the (implicitly royal?) sac-
rificer as the measure of the cosmic axis,54 it remains fruitful to see in dhárman
the Ṛgvedic nuance of “foundation.” A nonsacrificial instance of posts being
associated with royal dharma is found in Aśoka’s dhaṃma-ṭhambas or “dhaṃma
pillars.”
Another development, which can be traced from usages in the Maitrāyaṇī
Saṃhitā to ones in the Taittirīya, involves new articulations of rapports between
the king and dharma. According to Maitrāyaṇī Saṃhitā 2.6.8; 4.4.2, Soma,
Indra, Varuṇa, Mitra, and Agni—gods who are all connected with dhárman in
the ṚV—are also invoked as “upholders of dharma” (dharmadhŕ. tas) to make the
new king an upholder of dharma (Olivelle 2004a, 493). Taittirīya Saṃhitā
1.8.16.2 uses a Ṛgvedic epithet of Varuṇa to identify the king explicitly with
him: “You are Varuṇa, whose dharma is true/real (váruṇo ‘si satyadhármā)”
(Idem). Such articulations are also found in the systematization of the two great
royal rituals, the Rājasūya and the horse sacrifice or Aśvamedha.
In the Rājasūya, according to both Maitrāyaṇī Saṃhitā 2.6.6 and Taittirīya
Saṃhitā 1.8.10.1–2, after Mitra is invoked as “lord of truth” (satya) and Varuṇa
as “lord of dharma” (dharmapati), it is announced, “This is your king,
Bhāratas”55—on which Āpastamba Śrautasūtra will make the comment that by
“Bhāratas,” one means “respectively Kurus, Pañcālas, Kuru-Pañcālas, or simply
people” (Heesterman 1957. 70). This “vedic” identification of a people as
“Bhāratas” at their king’s consecration has, of course, a big future. First, the
Śatapatha Brāhmaṇa will give a hint of a story about a king Bharata Dauḥṣanti
who performed an Aśvamedha and had a mother named Śakuntalā who was an
Apsaras (13.5.4.11–14).56 Then the Mahābhārata will foreground its story of the
birth of Bharata from King Duṣyanta and Śakuntalā as its first upākhyāna or
subtale. The Poona Critical Edition establishes that in the Mahābhārata’s base-
line archetypal text, the epic’s first two upākhyānas or substories are those of
Śakuntalā, mother of Bharata, and of Yayāti, father of Pūru, both of which have
as their background the epic’s account of the Aṃśāvataraṇa: the “descent
of the partial incarnations” of gods into human lineages to relieve the burden

53. TS 1.3.1.2; MS 1.2.11; VS 5.27; Olivelle 2004a, 492.

54. For dharma and sacrificial posts, see Biardeau 2004, 99 and passim. Horsch 2004, 446–47 n. 68
observes: “it cannot be surprising that there was an actual Dharma cult during the medieval era . . . in Bengal,”
which, though he does not mention it, makes use of planks cut to the outlines of a stylized yūpa, some with
upraised nails for impalement. See Hiltebeitel 1991, 199 Figure 13 and passim. See chapter 9 § D.
55. Olivelle 2004a, 493. Cf. Olivelle 2004a, 508 n. 8.
56. See Biardeau 1979, 117–19; Thapar (1999) 2002, 10–11; Brodbeck 2009a, 57 n. 47.
 a vedic history of dharma 83

of the earth.57 Whereas the critical edition, following the Northern Recension in
this regard, has the Śakuntalā story, with the birth of Bharata, first, the Southern
Recension reverses that order and puts the Yayāti story, with the birth of Pūru,
first (see Sukthankar 1933, 282). This reversal by the Southern Recension is
likely to be the result of its handlers’ interest in advancing the epic’s own genre
identification as itihāsa, “history” (see Hiltebeitel 2010b; forthcoming-a,
chapter 4; forthcoming-d), since, as the Śakuntalā story itself makes clear
repeatedly, Duṣyanta and Bharata are “Pauravas,” “descendants of Pūru.” But
if the Southern Recension reverses the sequence of the “Vedic” past to make it
more linear, the archetypal Mahā-“Bhārata” would also have had an interest in
anchoring its itihāsa in a refashioned Vedic “Bharata story” that unifies a Vedic
people and a Vedic land. Just as the Mahābhārata serializes the Ṛgvedic names
Pūru, Bharata, and Kuru, the inaugural announcement found in the two
Yajurveda Saṃhitās would presumably have reflected the preeminence of a
Kuru king in the Maitrāyaṇī and his precedence for the Kuru-Pañcālas in the
Taittirīya, while harking back to the younger family books of the ṚV to connect
dharma with this prototypical royal name.58
As to the Aśvamedha, Bowles (2007, 93 n. 47) cites the Taittirīya Saṃhitā
version of a formula that equates the king with dharma: “with my two shins and
my two feet I am dharma (dhármo ‘smi), the king fixed firmly on his people.”59
And Olivelle cites a formula found in all three Black Yajurveda Saṃhitās that
connects dharma with the reins of the king’s chariot (2004a, 493). One may
observe that the king’s association with dharma differs in the Rājasūya and the
Aśvamedha. In the Rājasūya, he is consecrated mainly in the name of Varuṇa
to lawful and truthful rule; in the Aśvamedha to these more martial and even
subduing formulas that connect him with dharma only as a kind of law unto
himself, without mentioning Varuṇa or for that matter any deity. Although the
earliest ṚV knows some form of Aśvamedha,60 its promotion in the Yajurveda
Saṃhitās, according to Witzel, “forms the final, culminating point in the
development of the śrauta ritual”—especially in the White Yajurveda and other
texts developed in the final third center of Vedic canonization in Kosala and
Videha (1997a, 301; 313).

57. See Brodbeck 2009, 28, pointing out that the Yayāti story, including the Uttarayāyata, comprises eighteen
adhyāyas: a signature number of the Mahābhārata that, within it, can signify a kind of epic epitome (see Hiltebeitel
2006a, 231; 2009a, 170).
58. Cf. Brodbeck 2009, 67 on the “prefatorial” treatment of the Śakuntalā story and its “patrilineal import.”
59. Jáṅghābhyāṃ padbhyāʹṃ dhármo ‘smi viśí rāʹjā prátiṣṭitaḥ; see P.-E. Dumont 1948. Bowles indicates that
the formula seems to have been changed from an older version at Maitrāyaṇī Saṃhitā 3.11.8 that “has dhīʹro for
dhármo which suggests a little of the older sense”— that is, the king would say “I am firm” rather than “I am
dharma.”
60. See Witzel 1997a, 313 n. 28: “Found already in ṚV 1.162–63; apparently even earlier at 4.38–42, 4.42.8.”
84 dharma

Finally, apparently reflecting post-Taittirīya Saṃhitā new conceptions in both

the second and third centers, Taittirīya Brāhmaṇa 3.4.1 and Vājasaneyī Saṃhitā
30.6.10 use dharma in a context that Olivelle and Bowles interpret as suggesting an
early association with litigation. In the Puruṣamedha or human sacrifice, which
may be only a theoretical construct working out implications of the Puruṣasūkta,
but in which people are in any case dedicated to various deities, it is prescribed that
a sabhāga be sacrificed to dharma and a deaf man to adharma. Olivelle takes the
sabhāga to be “presumably, a man who participates at the royal audience hall [sabhā]
where judicial proceedings are carried out,” and the deaf man as one excluded
from such debates (2004a, 493–94). Bowles similarly takes the sabhāga to be “a
member of the judicial administration” (2007, 90 n. 34). It would seem that three
novelties have converged here: the earliest suggestion that dharma is connected
with judicial procedure in a royal assembly hall; the first mention of adharma;61
and the first reference to Dharma, not to mention Adharma, as a deity who receives
sacrifice. Needless to say, real or as a simulacrum, the sacrifice of a deaf man and
“a member of the judicial administration” is a pretty rough idea of justice.

E. The Brāhmaṇas

Brāhmaṇas are prose texts in which the ritual theologians of the four Vedic schools
comment on the performance and meanings of the śrauta sacrifice, which they
continue to systematize. I will discuss one passage from the Taittirīya Brāhmaṇa,
which, following from the Taittirīya Saṃhitā, reflects developments in the Pañcāla
area; one from the Aitareya Brāhmaṇa [AB] of the ṚV; and several from the
Śatapatha Brāhmaṇa [ŚB] linked with the Vājasaneyī Saṃhitā. Other than the
Taittirīya passage, all the other passages come from texts or segments of texts (AB
6–8) that were developed east of the Kuru-Pañcāla area in Kosala and Videha,62
“the new political center in eastern North India” (Witzel 1997a, 307). Populations
there were more mixed and varied than in the Kuru-Pañcāla lands, and “only in
part newly stratified.” Kings aspiring to orthodox recognition invited Kuru-Pañcāla
and other Brahmins from the west (rather than from intermediate areas) to attend
their courts and perform their śrauta rituals. Most famous is King Janaka of
Videha, “the main protagonist of Kuru-style Vedic civilization in the late Brāhmaṇa

61. See Horsch 2004, 446 n. 63; Olivelle 2004a, 508 n. 9.

62. “The Kāṇvīya ŚB is from Kosala” [= ŚBK] and develops as “the main Kosala YV text,” the other being the
“closely related ŚBM” [M = Mādhyandina recension] (Witzel 1997a, 313, 314), translated by Eggeling, and probably
from Videha (316). Both it seems go back to a likely lost “original version of the White YV Brāhmaṇa,” and are
“opposed to the strictly western form of the ritual and its texts (such as in MS, KS, TS).” The present ŚB can be
regarded as having 2 main parts: the “Yājñavalkya” section (ŚBM 1–5 = ŚBK 1–7) and the “Śāṇḍilya” section (SBM
6–10 = ŚBK 8–12), with added material to “this combined version” (ŚBM 11–14 = ŚBK 13–17) (Witzel 1997a, 313–15).
 a vedic history of dharma 85

and Upaniṣad periods,”63 who invited “Brahmans from the west to his frequent
brahmodyas and to be his śrauta priests” (311–13 and n. 281). The interest in relocat-
ing orthopraxy can be seen in the way the ŚB and other texts seek “to present a
complete collection, with ‘theological’ discussion, of all the śrauta (and, for the first
time, some gṛhya) materials” (314). This “late Brāhmaṇa period” also “saw the
emergence of a S.E. Koine” covering these areas and including AB 6–8 and the ŚB
among its creations (1989a, 252; 1997a, 318; cf. 331).
As to usages of dharma, there is again nowhere near the frequency of ref-
erences that one would expect if dharma were “a key concept that defined the
Brahmanical view of the world and of human life” (Olivelle 2004a, 494). I turn
first to a Taittirīya passage since it is probably the earliest of those I will discuss.
Taittirīya Brāhmaṇa 3.11.1.20 has a formula that calls the kingdom “abundant
Śrī (Prosperity), the wife of Indra, the wife of dharma (dharmapatnī), who has
arisen over all beings.”64 This could make Śrī the wife of both Indra and Varuṇa,
who has the frequent epithet dharmapati (Olivelle 2004a, 495), or, as I think
more likely,65 she could be the wife of both Indra and Dharma, for as we have
just seen, Dharma is now a deity who receives sacrifices and can probably be
embodied in a sacrificial post. Should it be the latter, we could have an inkling
of the kind of thinking that will allow the Mahābhārata poets to conceive of
their heroine, the Pañcāla princess Draupadī, to have sons of Indra and Dharma
among her five husbands as the incarnation of Śrī.
As to the other usages, since it is not always possible to present matters chro-
nologically, I will do so topically, under three headings: (a) an instance where
dharma continues to be associated with kingship through Varuṇa and again
seems to imply litigation; (b) passages where associations between kingship and
dharma are instead with Indra; and (c) a text that begins to articulate dharma in
terms of the attributes and privileges of Brahmins. The latter, likely from an
added section of the ŚB,66 is probably the most recent of these passages.
Among a few Brāhmaṇa passages67 linking kingship and dharma through
Varuṇa, the most revealing one in the ŚB describes an offering made during
the Rājasūya:

63. He was “a contemporary of [King] Ajātaśatru of Kāśī” (Witzel 1997a, 312).

64. Bowles 2007, 93 n. 48 with the Sanskrit formula.
65. Varuṇa seems presently to be a bachelor.
66. The ŚB includes an “eastern ‘Yājñavalkya’ section” (ŚBK 1–7 = ŚBM 1–5, where “eastern” means from
Videha), a “western ‘Śānḍ ị lya’ section” (ŚBK 8–12 = ŚBM 6–10, where “western” means from Kosala?), and an
addition to “this combined version” (ŚBM 11–14 = ŚBK 13–17 = ŚBM). Our passages come from all three segments.
67. Olivelle 2004a, 495, in TB dharma occurs three of four times “associated with Varuṇa and the overlord
(adhipati).” At TB 3.9.16.2 in connection with the king who is performing the horse sacrifice, commenting on a TS
passage, “The overlord is dharma. He does, indeed, obtain dharma.” At TB 3.11.4.1 and probably 3.11.1.20, Varuṇa is
dharmapati. The fourth seems to be TB 2.5.7.2, on which see Horsch 2004a, 436, just cited regarding “eaters.”
86 dharma

Then to Varuṇa the lord of dharma (dharmapati) he offers a cake

made with barley. Thereby Varuṇa himself, the lord of dharma,
makes him [the king] the lord of dharma. That, surely, is the highest
state (paramátā) when he becomes the lord of dharma. For when
someone attains the highest state, (people) come to him (in matters
relating) to dharma. (ŚB 5.3.3.9; Olivelle trans. 2004a, 495).

According to Olivelle, dharma here “has to do with matters regarding which peo-
ple come to the king and must refer principally to legal disputes. Dharma is thus
placed squarely within the public realm of law and social norms that must be
overseen by the king. We can now understand why the king is the dharmapati, in
the same way as Varuṇa, the sovereign who oversees the cosmic dharma” (Idem,
495–96). Olivelle finds that the ŚB “further supports the meaning of dharma as
social order founded on law” in a “creation story,” which, while not mentioning
kings, links dharma with the waters, and ends: “But where there is no rain, then
the stronger indeed seizes the weaker ones, for the waters are dharma” (ŚB
11.1.6.24). As Olivelle says, “The argument here corresponds to the ‘law of the
fish’ (matsyanyāya) of later Dharmaśāstras,” and also of the Mahābhārata: that
dharma reigns when it rains and anarchy reigns when it does not.68
Meanwhile, the AB passage comes from a late segment of that Brāhmaṇa
(AB 6–8) which “shows great familiarity with the east and the south-east of
Northern India” (Witzel 1997b, 320). The segment “prominently deals with
royal rituals”—our passage comes in a large section that treats “the royal
consecration and the role of the royal priest (purohita)”—and was presumably
composed by Aitareyins who had come east to enhance this involvement. As
Witzel observes, such an emphasis in a ṚV Brāhmaṇa devoted to the Hotṛ priest
may seem surprising, “but this feature agrees well with the efforts of local kings
to enhance their status in the eyes of their more ‘advanced’ western neighbors.”
Indeed, AB 8.14, right in the middle of the passages we shall now discuss, calls
“the great chieftains of the easterners” saṃrāj—a term we have seen linked with
dhárman in the ṚV through Varuṇa—”while those of the central Kuru-Pañcāla
realm retain the rather traditional small state title rājan.” As mentioned, we are
seeing reflected here not only the second urbanization of India, which came
with jungle clearance, increased use of iron tools, and rice cultivation, but the
formation once again of “Vedic” states.69 In what is, according to Olivelle (2004a,
494), the only usage of dharma in the entire AB, the focus of the Rājasūya is
turned from Varuṇa to Indra. In virtually identical passages, once Indra and the

68. Olivelle 2004a, 496; cf. Horsch 2004, 435; Bowles 2007, 90–91 and n. 35–37.
69. All quotes from Witzel 1997a, 321. Witzel even considers the possibility that, while purposefully leav-
ing it unmentioned, AB 6–8 may have been “made under the early Magadha kingdom” (1997a, 320–21 and
n. 333), which soon, at least, came to have heterodox associations.
 a vedic history of dharma 87

king have undergone a “great consecration” (mahābhiṣeka) and are enthroned,

the All-Gods (a generalized divine consensus) announce Indra to the gods while
̣
certain “king-makers” (rājakartāraḥ) present the king to the people ( janāh):

Do ye proclaim him, O people [O gods], as overlord and overlordship

(saṃrājaṃ sāṃrājyam), as paramount ruler and father of paramount
rulers (bhojaṃ bhojapitaram), as self-ruler and self-rule (svarājaṃ
svārājam), as sovereign and sovereignty (virājaṃ vairājyam), as
supreme lord and supreme lordship (parameṣṭhinam pārameṣṭhyam),
as king and father of kings (rājānaṃ rājapitaram). The kṣatra (royal
power) has been born, the kṣatriya has been born, the suzerain of all
creation (viśvasya bhūtasyādhipatir) has been born, the eater of the
commoners (viśām attā) has been born, the slayer of foes has been
born, the guardian of Brahmins (brāhmaṇānāṃ goptā) has been born,
the guardian of dharma (dharmasya goptā) has been born. (AB 8.12,
17; slightly modifying Olivelle 2004a, 494)

We may note that there is no reminiscence in this “great consecration” that the
eastern “peoples” to whom this great king is being presented would any longer be
encouraged to think of themselves as “Bhāratas.” It will be an agenda of the
Mahābhārata to invigorate the idea of bringing eastern Indians in under that rubric
(see chapter 7). But theoretically, it is something of an advance on the Puruṣasūkta
and other late Ṛgvedic hymns that begin to focus on Indra rather than Varuṇa. The
king now has a whole range of grand titles including svarāj, saṃrāj, and even virāj.
He will be a Kṣatriya rooted explicitly in the “royal power” and need come only
implicitly from the rānjanya or nobility; he slays foes; he is proclaimed in the name
of Indra and the absence of Varuṇa;70 and he is positioned vis à vis the castes—here
minus the Śūdras—by his capacity to devour71 commoners and protect Brahmins.
This topic brings us to two ŚB passages that, I believe, carry forward the
Ṛgvedic sense of dharma as enigma. First is a passage that has provided the title
of a book by Charles Malamoud (1996). The intriguing phrase lokapakti,
“cooking the world,” is used at ŚB 11.5.7.1 to hail what is achieved by svādhyāya,
personal vedic recitation or “recitation to oneself.”72 Svādhyāya has by now
become a hallmark, if never quite the monopoly, of the Brahmin class:

70. Cf Bowles 2007, 92–93 on ŚB 5.3.3.9, also concerning the Rājasūya, this time “evoking the special
relation that Varuṇa has with dharma as dharmapati: an oblation made of barley . . . , called a vāruṇá, is offered to
Varuṇa, the lord of the dhárma, thereby he makes the king the lord (páti) of dhárma, and so people come to him
in matters of law (dhárma upayánti). In this case the jurisprudential implications appear obvious.”
71. Cf. Horsch 436: TB 2.5.7.2 makes Agni and Mṛtyu (Death) eaters by Law, that is, where dharma implies
in conformity with their own natures.
72. See chapter 5 § A. It is one of a Brahmin’s five obligatory daily rites, the “sacrifice to brahman.”
According to the commentator Sāyaṇa, “the wise man wards off, by means of the svādhyāya, that suffering of the
ātman that is redeath, punarmṛtyu” (Malamoud 1996, 268 n. 1).
88 dharma

Here is now the praise of the personal recitation (svādhyāya) of the

Veda. The personal recitation and learning are sources of pleasure
for the Brahmin. He acquires presence of mind, becomes
independent (svapati), and acquires wealth day after day. He sleeps
well. He is his own best physician. To him belong mastery of the
senses, the power to find joy in a single object, the development of
intelligence (prajñāvṛddhi), fame, and cooking the world (lokapakti).
The growing intelligence (prájñā várdhamānā) brings to the Brahmin
four dharmas: Brahmanical stature, fitting deportment, fame, and
cooking the world (brāʹhmaṇyaṃ pratirūpacaryāʹm yáśo lokapaktím).
The world, as it is being “cooked,” gratifies (bhunakti) the Brahmin
with four dharmas—with veneration, with gifts, with the condi-
tions of not being oppressed, and of not being subject to capital
punishment (arcáyā ca dāʹnena cājyeyátayā cāvadhyátayā ca).73

By personal recitation, a Brahmin self-generates, as it were, four dharmas by

which the world recognizes him—one of which is that he “cooks the world” to
do just that. And thanks to his having cooked the world in such fashion, the
world gratifies him with four more dharmas that amount to class privileges and
immunities. Scholars have spoken of this usage as “possibly a new connotation
of dharma as either a specific attribute or a right obligation” (Olivelle 2004a,
497), and even as “the most momentous innovation” (Horsch 2004, 436). But
how would one translate dharma here when “cooking the world” is the meta-
phor, and hinge dharma, that takes one from the first four self-generated
dharmas to the second four, which are matters of social recognition? Olivelle
leaves it as “dharmas.” Moreover, in suggesting that the plural usages anticipate
the Buddhist meaning of dharmas as multiple “reals,” he joins Horsch, who
offers, “Here the universal abstract law is therefore split up into a plurality of
concrete factors.” Bowles also seems to accept this Buddhist analogy (2007, 94).
Olivelle also sees this usage as similar to the Śrautasūtras’ “ritual use of sva-
dharma,” which “refers to the fact that a particular rite has its own ritual details
(dharmas) specific and limited to it and not taken over from or extended to
other rites” (2004a, 502). I do not see either of these connotations in “cooking
the world.” The Buddhist theory of multiple “reals” certainly includes no “real”
comparable to “cooking the world”; and ŚB 11.5.7.1 hardly covers specific ritual
details, either of a ritual or a person. Meanwhile, Horsch scales down “cooking
the world” to “maturity in the world (i.e., religious influence in the (social) envi-
ronment),” and, without offering a translation, takes the first four as “the moral

73. First half from Malamoud trans. 1996, 23; second from Olivelle trans. 2004a, 497, both slightly mod-
ified; cf. Horsch 2004, 436; Bowles 2007, 93–94.
 a vedic history of dharma 89

characteristics of the Brahman” and the second four as “rights and privileges”
due to him in return. Malamoud circumlocutes to make do with “duties.”
Bowles says that “Law is becoming duty” here (2007, 94).
None of this quite handles the metaphor “cooking the world.” As the
hinge dharma in the passage, “cooking the world” is how one gets from the
first “four foundations” to the second four. And the first four have their
foundation in a bottomless resource: personal recitation of the Veda. Moreover,
before one gets to the first four, personal recitation is the foundation for such
intermediary supports as learning, presence of mind, independence, wealth,
good sleep, self-medication, mastery of the senses, the power to find joy in a
single object, and—most important in that it is mentioned last and also twice
as another hinge (though not a dharma)—developed and growing intelligence.
As Malamoud demonstrates, the metaphor of “cooking the world” is grounded
in the work and inspiration of the primal sages. Desire, consecration for
sacrifice (dīkṣā), sweat, toil (śrama), creative fervor (tapas), painful overheating,
and exhaustion go into preparing food for gods, who like their food cooked.
These are mystic connections (bandhus) in a “course” or “way” (adhvan) that is
defined by “ceaseless movement . . . from one end of the sacrificial ground to
the other,” one that regulates an “effort” that mimes the “labor of the Vedic
seers, who so strongly desired the Universe, in the beginning, that they
brought it into existence” (Malamoud 1996, 33). “Once consecrated, one pre-
pares a space for himself and one is born into a world one has made by one-
self.” That is, one cooks one’s own world (45). The “metaphorical cooking of
the sacrificer . . . precedes the actual cooking of the offering,” the first being
“the genuine oblation,” the latter a “substitute.” Digestion, milk, sperm, ges-
tation, marriage, cremation, and renunciation are also homologized to cooking.
By cooking himself, the Brahmin can thus perfect a self while also cooking up
the respect others have for him. This world, cooked by the Brahmin, is the
‘created’ world which he creates and organizes around himself in the sacrifice.”
But “the world cooked by sacrificial activity” has no raw natural opposite.
“Everything is already cooked such that all that remains is to re-cook it.
This sacrificial fire fed by the Brahmin does nothing other than redouble the
activity of the sun . . .; ‘That [sun] cooks everything in this world (eṣá vā idáṃ
sárvam pacati), by means of the days and the nights, the fortnights, months,
seasons, and year (saṃvatsareṇa).74 And this [Agni] cooks what has been
cooked by that [sun]: “he is the cooker of that which has been cooked,”
said Bharadvāja of Agni, for he cooks what has been cooked’” (Ibid., 48,

74. The English translation in Malamoud 1996, 48 has, incorrectly, “years” in the plural. The year is a
significant totality in this particular cuisine; see the next passage discussed.
90 dharma

quoting ŚB 10.4.2.19). To paraphrase, everything is cooked all the way up and

all the way down. It is also only a short step to say that “time cooks” (kālaḥ
pacati), a signature saying of the Mahābhārata.75
As regards this innovative usage of dharma on the societal front, one thing
is clear. As Bowles puts it, the passage marks “a clear divergence between the
activities and characteristics of one social group, the brāhmans, and the activ-
ities of other people, ‘those that are being cooked.’ ”76 Indeed, we may detect a
certain correlation with AB 8.12 and 17, where kings and Indra are “guardians
of the Brahmins” and “eaters of the commoners.” On the scale of the sun,
Brahmins may be cooked like everyone else, but at least they cook the world in
a way that avoids their being eaten by kings. I believe the safest thing to say as
regards anticipation of later usages of dharma lies not in new connotations hav-
ing to do with duties, which are not that clear, but in recognizing that while
they have been recording ways in which the king is becoming a law unto him-
self, some Brahmins have begun searching for a way to say similar things about
themselves. Bearing this in mind, however, I think we should also notice that
the second hinge term that has been translated as “developed” and “growing
‘intelligence’” (prajñā) could also be translated as “wisdom.” When the Buddha
and the early Buddhists, soon enough, link dharma with “wisdom” (prajñā)
gained by self-discipline and self-cultivation, they will find Brahmins well
prepared to be among their conversation partners.
Finally, in a passage featuring some of these same cooking themes, ŚB
14.2.2.29 supplies the formula, “A well-supporting (su-dharma) support (dharma)
thou art” (dharmāsi sudharmeti; Eggeling trans. 1963, Part 5, 485) as addressed
to the walking sacrificer—first as the sun, and then as the specially prepared
pravargya pot, said among other things to represent the year. Between the two
formulas, the sun is described as follows: “Truly, that one, who burns yonder, is
the maintenance (dharmā), since he maintains (dhārayati) this universe (or:
everything here), through him this universe is maintained (dhṛtam)” (Horsch
trans. 2004, 433; cf. 446 n. 56). The sun is being compared here with the pra-
vargya pot in which milk steams up in a fiery column. As we saw earlier in this
chapter, both the pot and the fiery column of this ritual may have been enigma-
tized in ṚV 1.164.43 and 50 as providing the “first foundations” (dhármaṇs)
which, in the second of these two verses, reach up to heaven. In the ŚB, the pot

75. This paragraph shortens and modifies a discussion in Hiltebeitel and Kloetzli 2004, 558–59. Recall
that when Agni “assumes the form of the universal fire, the Sun,” he “becomes the foundation [i.e., dhárman]
for all things” (Brereton 2004, 452).
76. Bowles 2007, 94, goes on to say, I think with too much backreading, that “the relationship between
them is reciprocal and complementary” and that “[s]uch a description would not at all be out of place in the
dharma literature.” Varṇa-reciprocity is not the point of this metaphor.
 a vedic history of dharma 91

is itself the sun as the “head of the sacrifice” that is being restored to the
decapitated Viṣnụ as a secret, dangerous, and all-encompassing option within
the soma sacrifice. In this text, the pravargya can be performed within the
Rājasūya as well as in rituals where the sacrificer would be a Brahmin (ŚB
14.2.2.47). But in other texts the pravargya also has strong associations with the
Aśvamedha (Houben 2000, 502, 526). Whatever his status, as he walks the
ritual arena, the sacrificer would be a very enigmatic “support” (dharma) for
such erstwhile “first foundations.” Coming in the last section of the ŚB to have
been composed, the passage may already breathe the air of the Āraṇyakas77 and
Upaniṣads.

F. The Upaniṣads

The next Brahmanical texts to generate innovative, if not numerous, usages of

dharma are the Upaniṣads: texts that “form the (mostly) secret teaching at the
end of Vedic study of each school” (Witzel 1997a, 331). Olivelle finds the term in
“just nine passages” (we shall note one more) in the four oldest Upaniṣads,
among them most notably the Bṛhadāraṇyaka (BĀU) and the Chāndogya (ChU)
(2004a, 498). Twelve or thirteen Upaniṣads are accepted as completing the
Vedic canon, that is, they were produced within one or another of the four Vedic
schools, and are regarded as the “end” or “completion of the Veda” (vedānta).
Most scholars have taken the earlier Upaniṣads to be pre-Buddhist, but there is
also some thought that even the early Upaniṣads may contain post-Buddhist
interpolations, which may include some of our pertinent passages.78 We thus do
well to keep early Buddhism in mind, just as the reverse will be true when we
come to the ways that early Buddhist texts portray the Buddha as speaking about
the Vedas and Brahmins, and also to Brahmins. While the Upaniṣads reflect
intensified interest in increasing social stratification, they are largely silent on
changing conditions that are making northeastern India “a real melting pot”:
the consolidation of older tribes and the settlement of new immigrant ones; the
emergence of large kingdoms and states; and the beginnings of India’s second
urbanization (Witzel 1997a, 328–29). Once the eastern provenance of the BĀU
is recognized as making it a continuation of the ŚB, it is sufficient to recognize
the period as one where “the geographical horizon of the early Upaniṣads
stretches from Gandhāra to Aṅga,” or Afghanistan to Bihar, and that wandering
sages traveled extensively exchanging ideas of widening currency (330).

77. According to Olivelle, there are only three usages in the Ᾱraṇyakas, two rather obscure (2004a, 497).
78. See Bronkhorst 2007, 112–35, especially 119–20, 130; 219–40, most notably 226–30, 236.
92 dharma

The Upaniṣads continue to mention dharma only where Brahmins concern

themselves with matters that do not go explicitly beyond their own class, or in pas-
sages that follow up and continue earlier developments to shore up and define
their duties, options, and privileges as a class. Thus the ChU speaks of the “three
types of persons whose torso is dharma”—“the one who pursues sacrifice, Vedic
recitation, and gift-giving”; “the one who is devoted to austerity”; and “the celibate
student of the Veda living permanently at his teacher’s house” (2.23.1). These
three apparently optional and not yet serialized “modes of life” are pertinent
mainly to Brahmins.79 Such a Brahmin-centered orientation is especially clear in
one of the preliminary cosmogonies in the BĀU. In the beginning, brahman was
alone and “had not fully developed” (na vyabhavat). First, along with certain ruling
gods (Indra, Varuṇa, Soma, Rudra, Yama, Mṛtyu, etc.), it created “the ruling
power” (kṣatra) to which a Brahmin pays homage at a Rājasūya (it is as if Brahmins
have already been created with the brahman). When brahman still “did not become
fully developed,” it created the Vaiśya class with various gods; when it still “did not
become fully developed,” it created the Śūdra class with various gods.80

It still did not become fully developed. So it created dharma, a form

superior to and surpassing itself. And dharma is here the ruling
power of the ruling power (kṣatrasya kṣatram). Hence there is nothing
higher than the dharma (tasmād dharmāt paraṃ nāsti). Therefore, a
weaker man makes demands of a stronger man by appealing to the
Law (dharmeṇa), just as one does by appealing to a king. Now,
dharma is nothing but the truth. Therefore, when a man speaks the
truth, they say (āhur) that he speaks dharma; and when a man speaks
dharma, that he speaks the truth. They are really the same thing.
(Olivelle trans. 1998, 49–51, slightly modified in accordance with
Idem 2005b, 125 and 132; cf. Idem 2004a, 499)

Here for the first time “Law” is an unexceptionable translation of dharma.81

Clearly, it relates to litigation in a king’s court. Granted “the weaker man” could

79. For extensive discussion, see Olivelle 2005b, 53–74; cf. Idem, 1993, 108; 2004a, 500–501; Bowles
2007, 95–97.
80. BĀU 1.4.11–13. The passage completes an adhyāya that begins with an identification of the first being
as puruṣa because he is “shaped like a man” (1.4.1) and describes an initial phase of cosmogony that includes the
creation of male and female from the “self” as “the supercreation of brahman” (Olivelle) or “highest creation of
Brahmā” (Radhakrishnan 1953, 166) (brāhmaṇo ‘tisṛṣṭi; 1.4.6).
81. See Olivelle 2004a, 499 and 508 n. 17; Bowles 2007, 72. Olivelle teases this meaning along a bit, for
example, “By appealing to the Law” and “. . . to the king” are a little free for just the instrumentals dharmeṇa and
rājñā, but the implication is sound. Building up to the summation of his hypothesis of a continuum from King
Varuṇa to the Buddha, Olivelle considers that kṣatrasya kṣatram, “the divine principle that gave legitimacy and
meaning to a worldly ruler, the term associated with the divine sovereign Varuṇa, would be a natural choice to
define the new dispensation, the new truth (satya) that the enlightened one had discovered” (2004a, 504).
 a vedic history of dharma 93

betoken justice for all and that dharma as truth-speaking could mean that he
could count on truthful witnesses. But it is especially Brahmins who would be
true witnesses for themselves.82 The passage and the one on “cooking the world”
are from the same school and milieu. In one, the Brahmin grows the intelli-
gence to cook up dharmas that favor him; in the other, the pre-cosmic brahman
does not fully develop until it has created dharma to keep Brahmins on top.83
As to the dialogical teachings for which the Upaniṣads are best known,
dharma has a few usages that are tangential to expressions of the oneness of the
self (ātman) and brahman. Two occur where the BĀU features its famous sage
Yājñavalkya. Although Yājñavalkya is renowned already in the ŚB, where he
offers novel ideas mainly as an expert on ritual,84 in the BĀU, still the innova-
tive thinker, “he focuses on teachings of the self” (Black 2007a, 75). Yet here we
come up against the question of whether his innovations, particularly about
karma but also about dharma, come up in a pre-Buddhist or post-Buddhist sec-
tion of the BĀU. This Upaniṣad has three sections (kāṇḍas) that were carefully
redacted into an omnibus whole by pulling together oral material from different
Vājasaneyī sources, possibly subschools. Yājñavalkya is known and mentioned
in all three kāṇḍas; but the one in which he makes his most sustained appear-
ances, the Yājñavalkya-kāṇḍa (BĀU 3–4), has late-looking touches: among
them, that whereas elsewhere in this Upaniṣad he is described as a pupil of
Uddālaka Āruṇi (BĀU 6.3.7), the great sage of the ChU, in the Yājñavalkya-
kāṇḍa (henceforth YK) nothing is mentioned of this disciplehood, and, rather
to the contrary, Yājñavalkya finds occasion to “shame” Uddālaka and demon-
strate his own superiority while settling “old scores” (Bronkhorst 2007, 120,
229). Bronkhorst brings this consideration into an argument that the YK is a
late and post-Buddhist piece of “hagiography from beginning to end” (236)
designed to put the origins of the karma-as-reincarnation doctrine into the
mouth of a Brahmin—that is, Yājñavalkya—to relieve Brahmanism of the
embarrassment that this doctrine would otherwise appear to have originated

82. Compare GDhS 21.19 where, again, although “the weaker man” (durbala) who is to be protected from
harm sounds like he could be someone from poor and/or low social circumstances; the context makes it evident
that the rule applies first of all to Brahmins. I believe Olivelle speaks too generally on this matter of “common
law” in “Power of Words,” where he reads the conclusion of the passage just cited (BĀU 1.4.14) anachronistically
to have “confirmed” that an Aśokan sense of dharma for all “must have penetrated the common vocabulary of
the people” (2005b, 132; cf. 125, 129) even before Aśoka. Even if the “they” in āhur (“they say”) “stands for what
is commonly seen among the people” (Ibid.), the people quoted would be primarily Brahmins. Surely the
Upaniṣads do not allow for such generalized ethnography.
83. Cf. Bowles 2007, 72: dharma “is now independent of any other factor, yet clearly associated with the
creation and separation of the four varṇas.” See also Horsch 2004, 437.
84. Black 2007, 68–69, 74–75: “on one occasion Yājñavalkya expresses the view that brahmins are the
most important aspect of performing a sacrifice” not just because they make a place appropriate for sacrifice
anywhere they go (the view of an interlocutor), but because “the individual participants are more essential than
the ritual actions themselves” (75; ŚB 3.1.1.4–5).
94 dharma

from kings85 and gained currency only through the intermediacy of Brahmins—
among others, and above all, of Uddālaka.86 Yet even if Bronkhorst is right
about its implications for tracing the karmic rebirth doctrine to non-Vedic
sources of the “greater Magadha” area, which I believe he makes quite plau-
sible, the YK, which has seemed to many scholars to be the oldest part of the
BĀU, could still contain old and not just archaic features and components.87
Such considerations must apply to what Yājñavalkya has to say about dharma,
which he mentions only twice, and both times in the YK.
The first usage speaks of a sequence of pairs of opposites where the Self
is made of light and the lightless, desire and the desireless, anger and the
angerless, dharma and non-dharma. The list implies an ascendance, with each
pair encompassing the preceding (Olivelle 2004a, 498). What is most inter-
esting is that the passage goes on to offer views on karma that Bronkhorst
takes to exemplify Yājñavalkya’s teaching of “a self that is not affected by
actions” (118), and to illustrate the incongruity of arguing, as some orthogeni-
cist theorists have done, that his karma doctrine could “have arisen in ritualistic
circles” (132–33):

Hence there is this saying: “He’s made of this. He’s made of that.”
What a man turns out to be depends on how he acts (yathākārī ) and
how he conducts himself (yathācārī ). If his actions [karman, singular]
are good, he will turn into something good (puṇyaḥ puṇyena karmaṇā
bhavati). If his actions are bad, he will turn into something bad by bad
action (pāpaḥ pāpena). And so people say: “A person (puruṣa) here
consists simply of desire.” A man resolves in accordance with his
desire (kāma), acts in accordance with his resolve, and turns out to be
in accordance with his action. (BĀU 4.4.5; Olivelle trans. 1998, 121)

According to Bronkhorst, since “there is no such thing as bad ritual activity in

the Veda,” a ritualistic explanation does not hold (2007, 133). “Made of every-
thing” up to and including dharma and adharma, the self remains unaffected,
as the passage goes on to explain: “Now, a man who does not desire—who is

85. “There are reasons to think that the Yājñavalkya-Kānd ̣ ạ . . . was primarily composed to remove the stain
of ignorance from the Vedic tradition”; that it was a “late” plot to unlink the karma teachings from a “non-Brahman-
ical origin” whereby “the doctrine was ‘dressed up’ to look Vedic” (Bronkhorst 2007, 119–20). Of course, all these
passages have echoes of Brahmanical teachings, principally in connection with Soma/the moon; and the king is
represented as a Kṣatriya, and thus presumably ārya, and is not from “greater Magadha,” where the BĀU was likely
composed. If Bronkhorst is right, the redactors of this “hagiography” took Brahmin claims to a new level.
86. On Uddālaka as the chief person through whom the Upaniṣads trace the doctrine of rebirth and
karmic retribution to a king, see BĀU 6.2, ChU 5.3–10, and (more uncertainly as to it being a king) Kauṣītaki
Upaniṣad 1 (Bronkhorst 2007, 112–19; 230–31; cf. Black 2007, 100, 116–19, 124–29). The king in the first two
texts is Jaivali Pravāhana, a Pañcāla who also upstages Brahmins at ChU 1.8.1–8 (Black 2007, 108).
87. For Olivelle, it comes from the BĀU’s “oldest core” (1998, 502–4, 521–23). See Brereton 2006.
 a vedic history of dharma 95

freed from desires, whose desires are fulfilled, whose only desire is his
self—. . . brahman he is, and to brahman he goes” (BĀU 4.5.6; Olivelle 1998,
121). It is thus explained how dharma and its opposite are the highest ground
for understanding how karma works, while the Self is something else: indeed,
the only desire worth really having, the only desire that can be fulfilled. Shortly
before this, Yājñavalkya has compared the state “where all the desires are ful-
filled, where the self is the only desire,” with a man’s embrace of “a woman he
loves” (BĀU 4.3.2). This brings us to a passage that could be said to take us
directly into Yājñavalkya’s lovelife, though probably the end of it.
One might argue that the inclusion of dharma and adharma as merit and
demerit in an ethical explanation of the mechanism of karmic rebirth, an expla-
nation grounded in desire and resolve, would be intelligible in a milieu familiar
with Buddhism, and that the Self as the only desire worth having could be a
riposte to the Buddhist doctrine of no-self (anātman). But the point would have
to be purely comparative, and uncomplicated by any philological or historical
considerations. The second passage—the one that Olivelle omits from his
count of Upaniṣadic usages of dharma—raises different possibilities. It occurs
in the YK version of Yājñavalkya’s famous dialogue with his dear wife Maitreyī
(BĀU 4.5), but not in the other version of that dialogue which occurs in the first
kāṇḍa of the same Upaniṣad at BĀU 2.4. Indeed, each telling has two extant
versions, making four, since both come in a Kāṇva and Mādhyandina recension.
There is a little variation in the two YK passages that mention dharma.
In the YK versions, Maitreyī is acknowledged at the beginning of their
exchange as a brahmavādinī, “a woman who talks about brahman,” or, in Olivelle’s
translation, “a woman who took part in theological discussions” (BĀU 4.5.1).88
Yājñavalkya is telling her about the self as he is making a settlement of his prop-
erty with her and her co-wife Kātyayanī, as he is about to leave them to begin
another “mode of life”89—which probably means that he is about to die.90 After a
stretch in which he tells Maitreyī that what is held dear, beginning with husband
and wife, is dear only out of love or desire for the self,91 he presents a series of less
tangible and more perplexing similes until Maitreyī finally breaks in:

88. Cf. Lindquist 2008, 421: “a talker about brahman”—noting, “This may be a technical term for someone
who participates in public debate (of which we have no direct evidence for Maitreyī), or it may simply mean that she
had a direct interest in discussing religious and philosophical matters.” Lindquist takes the dialogue at BĀU 4.5 to
culminate a “teaching narrative” that gets increasingly private and, in this case, “intimate.”
89. As Olivelle 1998, 112 translates vṛttam here at 4.5.1, attentive to there being no mention of a “life-stage”
or āsʹrama: in particular, no mention of saṃnyāsa as renunciation or retirement. As Olivelle 1981, 266 demon-
strates, samnyāsa in the sense of the final life-stage is not found in the Upaniṣads, and has there only one occur-
rence in Muṇḍaka Upaniṣad 3.2.6, where it “is a discipline undertaken by a yati [ascetic] and, therefore, not
synonymous with a yati’s state of life.”
90. See Brereton 2006, 31, 37, 343 and discussion; Olivelle 1998, 502; Bronkhorst 2007, 233–34.
91. The refrain, repeated twelve times in 4.5.6, is ātmanas tu kāmāya, “but out of love (or desire) for the self.”
96 dharma

“Now, sir, you have utterly confused me! I cannot perceive this at all.”
He replied. “Look—I haven’t said anything confusing. This self,
you see, is imperishable (avināśī); it has an indestructible nature
(anucchitti-dharmā).” (Olivelle trans. 1998, 131)

That is a translation of the Kāṇva recension. The Mādhyandina recension has

an additional phrase at the end:

Then Yājñavalkya said, “Surely, [my dear,] I speak nothing confusing.

Unvanishing (avināśī) is this self here, [my dear]. It’s nature
precludes dissolution (anucchitidharmā), but there is a merging of its
constituents (mātrāsaṃsargas tv asya bhavati).92 (Brereton 2006, 339)

Brereton offers this comment on the Mādhyandina version, which he sees as the
earliest since “it has an older diction”: “Even with Yājñavalkya’s explanation, we
might well remain sympathetic to Maitreyī’s puzzlement. . . . The point he
wishes to make . . . is that it is only perception of objects that disappears at
death. . . . [T]he self does not disappear with the disappearance of sense faculties
and objects of sense and therefore does not disappear at death” (2006, 340). The
parallel version of this dialogue has at this point something entirely different:
Maitreyī is confused because Yājñavalkya has told her “after death there is no
awareness” (BĀU 2.4.13). But Yājñavalkya’s reassurance is similar.93
I will make no attempt to sort out the different views of the textual his-
tories of these variants, on which every possible priority seems to have been
proposed concerning the two versions as a whole, not to mention the recen-
sional differences and individual passages.94 But it seems altogether plausible
that Brereton is right in seeing the YK version of the dialogue as earlier than
the version at BĀU 2.4. It alone seems to retain segments of an “old rhythmic
prose” core that concerned “a conversation about death” (2006, 341), espe-
cially at the beginning and the end, with our passage included in the conclusion.
Yet Brereton’s “reconstruction produces irregular lines” at this point, leaving
him with “no more than possibilities” (339). It may also be that he is right that
the Mādhyandina text is the older of the two YK versions (329, 340), which
were, however, both modified by expansions. But in our present passage, it

92. Cf. Bronkhorst 2007, 233 n. 35 quoting this translation by Slaje (2002, 15): “Look, actually imperish-
able, this [your] central instance [of cognition] (ātman) here bears [indeed] the property of indestructibility.
However, it [re]joins with (saṃsarga) [its causes,] the material components (mātrā).” I cannot imagine this trans-
lation helping Matrieyī, and also must admit to some confusion here on what others say: Bronkhorst, that the two
passages are the same; Olivelle (1999, 522) that the Mādhyandina text “adds” this clause and then follows it with
“a long passage that is identical to” BĀU 4.3.23–30.
93. Olivelle 1998, 69–70; Brereton 1986, 106; 2006, 331; Bronkhorst 2007, 234.
94. See Brereton 2006, 342 n. 48; Bronkhorst 2007, 127; 220 n. 3, 232–35, 239–40, with mostly
dissimilar views.
 a vedic history of dharma 97

would be the Mādhyandina text that expands the Kāṇva text to explain
something further about the self whose “nature precludes dissolution.”95 Thus
Bronkhorst, were he to grant that the YK versions kept elements of an archaic
text, would also have grounds to see its version as a late and post-Buddhist
expansion of that text.
The term dharma (actually dharman) here is usually translated by “nature” or
“property,” and justly so since it follows a long commentarial tradition.96 It can be
related to Ṛgvedic usages where dhárman can mean “foundational nature.” But
note that Halbfass cites the Mādhyandina usage anuchittidharman in this dia-
logue (= ŚB 14.7.3.15) in conjunction with Pāli upādavayadhammin, “subject to
origination and decay” and as an instance where “since ancient times
dharma . . . [can mean] ‘property,’ ‘characteristic attribute,’ ‘essential feature,’ or
more generally . . . ‘defining factor’ or ‘predicate.’” 97 The question remains
whether its usage in the YK is an old part of an early rhythmic prose core, as
Brereton finds possible, or is a novel usage interpolated into the text to make a
clinching point to Maitreyī. If it is the latter, there could be a post-Buddhist com-
plementarity in Yājñavalkya’s two usages of dharma.98 In the first, the self tran-
scends dharma and adharma; in the second, it has an indestructible dharman. It
is interesting that dharman is used in getting as far as one can to the bottom of
the most important things. That is, one could take it that the self “has an inde-
structible ‘foundation.’”
Yājñavalkya, a most interesting character, is the only person in the early
Upaniṣads to have anything to say about dharma in the dialogues for which the
early Upaniṣads are famous. In fact, he says precious little. What is especially
striking is that he uses the term only once—our first YK instance, addressing
King Janaka of Videha—in his extensive conversations at royal courts, where his
hosts and interlocutors include various kings. Yet as Bronkhorst indicates, the
YK makes quite a setting for him to unveil the teaching of karma as reincarna-
tion. Yājñavalkya describes karma as still a secret teaching when he takes another
Brahmin, Jāratkārava Ārtabhaga, privately aside so that it will not be discussed
in public at Janaka’s court, concluding: “A man turns into something good by
good action (puṇyena karmaṇā) and into something bad by bad action.”99 On the

95. Bronkhorst 2007, 239–40 thinks the Kāṇva text is older.

96. Radhakrishnan 1953, 285 does the same, following a gloss by Śankara of dharman as lakṣaṇa, “mark,
trait, feature,” that is, “it has the mark of indestructibility.”
97. Halbfass 1988, 334 and 555 n. 1; cf. 319, 551 n. 44, 555 n. 1. Cf. Gethin 2004, 533, emphasizing the
usage of dharma and dhamma as the second member of a bahuvrīhi compound as having “the sense of a particular
nature or quality possessed by something”; Horsch 2004, 440, 448 n. 86.
98. Brereton says that the seeming shift from a departure denoting death in BĀU 2.4.1 to the YK’s
departure for different “mode of life” (see n. 89 above) could have occurred “to support the emerging dharma
vocabulary and the ideal of renunciation” (2006, 331). This would imply a third usage.
99. BĀU 3.2.13; Olivelle trans. 1998, 81; see Bronkhorst 2007, 122, 232.
98 dharma

other hand, when King Pravāhana Jaivali of Pañcāla teaches the “five fires”/“two
paths” doctrine of transmigration to Uddālaka (ChU 5.3–10), he says it had
before that “never reached Brahmins” so that “[i]n all the worlds . . . government
(praśāsana) has belonged exclusively to royalty (Kṣatriyas).”100 Those “whose
behavior is pleasant” achieve rebirth in the upper three varṇas and “people of
foul behavior can expect to enter a foul womb, like that of a dog, a pig, or an out-
caste woman (caṇḍal̄ ayoni)” (ChU 5.10.7; Olivelle trans. 1998, 235, 237). What
leads to heavenly worlds begins by gift-giving as offering to the gods (ChU
5.10.3); and in the parallel passage of BĀU 6.2.10, the same is achieved “by
offering sacrifices, by giving gifts, and by performing austerities.”101 Similarly,
King Aśvapati Kaikeya instructs five wealthy landlord (mahāśāla) Brahmins
about burning up “all the bad things” (sarve pāpmānaḥ) in the sacrificial fire with
knowledge of “the self common to all men” (ātmānaṃ vaiśvānaram)”—that is, in
the fire of the self that unites all men in the name of Agni. He concludes,

Therefore, even if a man who has this knowledge were to give his
leftovers to an outcaste (caṇḍāla), thereby he would have made an
offering in that self of his which is common to all men. On this,
there is this verse: “As around their mother here, hungry children
gather, So at the fire sacrifice do all the beings gather.” (ChU 5.11–24;
Olivelle trans. 1998, 245)

Both of these learned kings see the big social picture in which sacrificial activity
unfolds: the first divisively, marking off rebirth in the upper varṇas from rebirth as
an outcaste; the second more embracingly. For him, the fire of the self is not only
the great purifier but the mother of the world’s hungry children! He seems to be
admonishing the Brahmins on matters of social privilege when he tells them that
each of them would have suffered a grievous harm had they not come to him. If
so, it is a noteworthy to find such a message treasured in a Brahmin-kept text.
Among later Vedic Upaniṣads the only one to say anything important about
dharma is the Kaṭha Upaniṣad (KU), in its well-known opening dialogue between
the boy Naciketas and Death (Mṛtyu, Yama). Other than Death, it is again a tale
about Brahmins and about death. Naciketas has been sent by his irritated father to
Death because he has belittled the bedraggled cows the father passed off as gifts.
In Death’s house, when Naciketas asks to know whether a man exists after death

100. ChU 5.3.6, Olivelle trans. See Black 2007, 101, 103; Bronkhorst 2007, 113–14, 124, 231. In going over
the question of Kṣatriya/royal “authorship” of various teachings, Black argues that “transmission of Upaniṣadic
knowledge” by Kṣatriyas would be an artifice of Brahmin composers for whom there was “nothing to lose, and a
lot to gain” in portraying “this knowledge as indispensable to the king’s political power” (128–29). But “author-
ship” and “transmission” are not the same as attributing origins.
101. BĀU omits the references to “behavior,” and also the line about “government,” while accentuating
instead Uddalaka’s submission to the king as a pupil (6.2.7–8); cf. Kauṣītaki Upaniṣad 1.
 a vedic history of dharma 99

or not, Death first tells him to ask something else, since “it’s a subtle dharma (aṇur
eṣa dharmaḥ).” But when Naciketas demonstrates his worthiness, Death discloses
“this subtle point of doctrine (dharmyam aṇum),” and in the next stanza one
learns—as Janaka does from Yājñavalkya—that it is “different from dharma and
adharma (anyatra dharmādanyatrādharmād).”102 The KU also has a third usage of
dharma outside the Yama–Naciketas dialogue that speaks of the error of seeing
dharmas as “distinct” (dharmān pṛthak; 4.14). Taking stock of all these usages,
Bowles suggests that one may regard the “subtle dharma” or “doctrine”—which
ultimately refers to the “ātman theory”—as something like a “true dharma” that
“stands over and above” dharma and adharma as a duality, as well as over multiple
dharmas as “various rules and obligations” (2007, 102). Like Horsch, who also
underscores the meaning “doctrine” in the two “subtle dharma” references (2004,
437), Bowles rightly rejects taking the plural as influenced by the Buddhist theory
of dharmas as multiple factors of experience; but in doing so, he adds that, “[o]ddly
enough,” the “subtle dharma” references “are not entirely unlike the Buddha’s use
of the word when he describes the entirety of his teaching as ‘the dhamma.’ ”103
I find Bowles’s discussion illuminating, and all the more so when one sees
how the text buttresses Death’s “subtle dharma” with two words for “foundation”
or “support.” First, one meets the old standard term pratiṣṭhā, used doubly in
2.11 only for it to be superseded:

11. Satisfying desires is the foundation of the universe (jagataḥ

pratiṣṭhām);
Uninterrupted rites bring ultimate security;
Great and widespread praise is the foundation (pratiṣṭhām)—
these you have seen wise Naciketas,
and having seen, firmly rejected.104

Yama commends Naciketas for repudiating these old foundations of Vedic

ritual and mantra. Then, just before he returns in 2.13 to the “subtle dharma”
he had first urged Naciketas to forego, he establishes that it will be about
something much deeper than those old ritual “foundations”:

12. The primeval one who is hard to perceive,

wrapped in mystery, hidden in the cave,
residing within the impenetrable depth—

102. KU 1.21; 2.13–14; Olivelle trans. 1998, 379, 385. Olivelle translates the last reference as “different from
the right doctrine and the wrong,” and suggests that Naciketas is now the speaker, but that it is not certain.
103. Bowles 2007, 102 n. 83. Cf. Olivelle 1993, 69–70, discussing Mbh 14.48.14–17, where, “Bewildered
by the array of expert opinions regarding the true dharma, the seers ask the creator god, Brahmā” about it.
104. This is from Olivelle trans. 1998, 383, except that I prefer “universe” over “world” for jagat.
100 dharma

Regarding him as god, an insight

gained by inner contemplation (adhyātmayogādhimena),
both sorrow and joy the wise abandon.

The “subtle dharma,” just about to be so named, will be about how the mystery
of the “primeval one” hidden in the impenetrable depth of “cave of the heart”105
can be gained by inner yogic contemplation. And this, Yama soon enough
reveals, is to be gained precisely by a newer kind of yogic “support” called
ālambana:106

17. This is the support that’s best (ālambanaṃ śreṣṭham)!

This is the support supreme (ālambanaṃ param)!
And when one knows this support (ālambanam),
he rejoices in brahman’s world (brahmaloke).107

That this “subtle dharma” should be poised between an abandoned “foundation”

of ritual and mantra and a liberating “support” of yoga may remind us of the
way dhárman as a “foundational” enigma was buttressed by usages of dharúṇa—
cosmological “supports”—in ṚV 5.15. On the one hand, this new “subtle
dharma” still has its resonances with both old and new terms meaning
“foundation” and “support.” On the other, it would also seem to have been
party to Brahmanical and probably Buddhist conversations relating dharma to
yoga. On the Brahmanical side, such conversations become commonplace in
the Mahābhārata and Manu, as we shall see in chapter 5. This suggests that
there would have to be a historical dimension to the choice of dharma from
among such older and newer terms for “what holds or supports” to bring these
meanings together.108
The KU is certainly later than the beginnings of Buddhism. More contro-
versially, Bronkhorst makes an interesting case that the same may be said for
the YK passages where Yājñavalkya gives his two ever-so-brief dharma talks.

105. On the “cave of the heart,” see KU 1.14; 3.1; 4.6–7. Although it is perhaps an added verse, just after
this sequence, KU 2.20 says that self hidden in the heart is “finer than the finest (aṇoraṇiyān), larger than the
largest,” confirming by its use again of aṇu that the “subtle doctrine” is about the ātman. Cf. Muṇḍaka Upaniṣad
3.1.7, with the ātman as sūkṣma rather than aṇu: “more minute than the minute (sūkṣmāc ca tat
sūkṣmataram), . . . farther than the farthest . . . yet near at hand . . . hidden within the cave of the heart” Olivelle
1998, 449). Both no doubt hark back to ChU 3.14.2–4: the self in the heart is smaller than a mustard seed,
greater than all the worlds. Cf. BᾹU 4.1.7 on space as the pratiṣṭhā of the heart.
106. Patañjali’s Yogasūtra uses it to mean “support” or “stimulus”; see 1.38, 1.10, 4.11. The support is vedic:
as one leading a student life (brahmacarya), Naciketas’ “support” is to utter the syllable Oṃ which “all the Vedas
disclose” (2.15–16). Cf. Monier-Williams [1899] 1964, 153, mentioning no earlier usages than Pāṇini (probably
fourth century) and a Buddhist usage.
107. Olivelle trans. throughout (1998, 385), except as noted.
108. I thank Greg Bailey for calling my attention to this way of formulating this “historical dimension”
(personal communication, December 2006).
 a vedic history of dharma 101

As with all Upaniṣạdic conversations about dharma, however, kings are mainly
in the background, if they are there at all, and they have nothing to say about
dharma themselves. As to whether Buddhists were in the background, the
Upaniṣads are silent. On the face of it, then, an intriguing fact has emerged.
Remarkably enough, one could say that the Buddha, said to have been born a
prince in a royal family of Ikṣvāku descent, will be the first Kṣatriya to speak
about dharma. And he will do so frequently and most extensively with Brahmins,
while also having numerous conversations with kings.
This page intentionally left blank
4
Early Buddhism
Three Baskets of Dharma

Where to begin with dharma in early Buddhism? Suddenly, we have a

truly vast topic. If we extend the notion of early Buddhism through
our classical period, we have to consider not only the ways dharma is
treated and transmitted in the different collections or “baskets” of
texts that comprise the Pāli canon of Theravāda Buddhism but, once
we begin to unfold that process, some of the variations in the
so-called eighteen schools of “Hīnayāna” Buddhism and some of the
important developments in early Indian (not to mention Chinese)
Mahāyāna. Let me set up three signposts through this maze.
First, I avoid the word “Hīnayāna,” which is a pejorative term
(meaning “inferior vehicle”) coined to advance the innovative
teachings of what was coming to be known as the “great vehicle” or
Mahāyāna. Since “Hīnayāna” does not occur in what seem to be the
earliest Mahāyāna sūtras,1 and since none of its eighteen (really more
than that) schools ever accepts this term, it is better to use a term
that the schools in question might have found unobjectionable.
I adopt the best term I know of: “‘Nikāya’ (sectarian) Buddhism,”
or, for short, “Nikāya Buddhism.”2 Nikāya is used here in its sense
of a “school” or “sect” that defines itself by allegiance to a particular

1. These tend to use the terms śrāvaka and pratyekabuddha to describe non-Mahāyanists
under two headings; see, for example, Nattier [2003] 2005, 84–85; Idem, 174 n. 6 on H īnayāna as a
“back-formation” of Mahāyāna.
2. See Nattier 1991, 9 n. 1 and passim, attributing this usage, without citing a particular work,
to Akira Hirakawa.
104 dharma

transmission of the Buddhist monastic rules, and implies reference to the Vinaya,
the collection or “basket” of texts having to do with monastic discipline.3 It should
not be confused with uses of the same word to describe the five major groupings
of the Buddha’s dialogues or discourses in the Pāli canon: the Dīgha-, Majjhima-,
Saṃyutta-, Aṅguttara-, and Khuddaka-Nikāyas. To keep the distinction visible,
I will put phrases like “Nikāya Buddhism” and “Nikaya schools” in quotes.
The second signpost is already familiar from chapter 2: the Mauryan
period, including the watershed reign of Aśoka Maurya, allows us to look in
two directions. If the Buddha lived ca. 560–480 BCE, or, according to recent
discussions, died around 400 BCE, one has about a century or a little more in
which to imagine (it is always engaging to do so, but we must refrain when it
comes to particular teachings) what the Buddha might have really said based
on what he is said to have said in the earliest texts attributed to him. With
regard to individual teachings, there is no consensus as to which texts are the
earliest, but it is widely agreed that the major edifice of the earliest Buddhist
teachings lies in the first four Nikāya collections and some portions of the fifth,
including the Sutta Nipāta and the Dhammapada. These collections comprise
the bulk of the Sutta Pit·aka. If altogether they reflect social conditions of about
300 BCE,4 the period of the early Mauryas, and were given their current shape
in that period, this means that the Mauryan period marks the end of about a
century or so of undocumentable oral canon formation and launches the
beginning of several centuries of diversified documentable canon textualiza-
tion. But nothing is certain here, as will be noted basket by basket.5
The third signpost, which we see emerging here, is that “Nikāya Buddhism”
can also be defined by the ways the different schools develop in relation to their
notion that the Buddha’s teachings are kept in “three baskets” (pit·akas) of texts.
The organization of a chapter on dharma in early Buddhism around the three bas-
kets may look straightforward. I am told, for instance, that Tibetan Buddhists—
who have a much vaster canon—still speak of the dharma as being like a

3. See Gombrich 1988, 111–12 and passim.

4. See Witzel 1997a, 307–8, 312, dating Buddhist texts in Pāli to “c. 3rd c. B.C.”
5. According to Schmithausen, scholars of early Buddhism have three positions available to them: (a) “the
fundamental homogeneity and substantial authenticity of at least a considerable part of the Nikāyik [i.e., Sutta] mate-
rials . . . yield a fairly coherent picture of the authentic doctrine of the Buddha himself” (1971, 278); (b) “extreme
skepticism” on that point (cf. Williams 2002, 23–30); and (c) that with the tools of higher criticism such as tracing
inconsistencies and supposed contradictions, one can detect a relative chronology that may in some cases take one
back to the Buddha himself. Schmithausen attributes these positions to Gombrich, Schopen, and himself respec-
tively; see Ruegg and Schmithausen 1990, 1–2. I view the first approach as one of engaging but untestable possibil-
ities; the second as one of testable possibilities; and the third one of wishful thinking, especially as regards the Sutta
Pit.aka. Favoring epigraphic evidence, Schopen argues that reconstructions of “higher criticism” rely on cross-school
parallels that are more likely late than early, and that “we cannot know anything definite about the actual doctrinal
content of the nikāya/āgama literature much before the fourth century C.E.” (1997, 23–30).
 early buddhism 105

three-legged stool, one leg in each basket.6 But I have not seen Buddhist dharma
approached elsewhere in this fashion.7 Although Buddhist tradition holds that five
hundred of the Buddha’s disciples recalled and arranged the Buddha’s Sutta and
Vinaya teachings under the names of the Dhamma and Vinaya at the First
Buddhist Council shortly after his death, and that the Abhidhamma is likewise the
original Word of the Buddha, this grouping, probably including its division of
suttas,8 emerges only gradually and not immediately with any reference to bas-
kets. In the dialogues themselves one hears that the “teaching” (dhamma) has
nine “parts” (aṅga). For instance, in Majjhima Nikāya 22, the instructive
Alagaddūpama Sutta or “The Discourse on the Simile of the Water Snake,” the
Buddha says, “a monk learns the teaching (dhammam)—the discourse[s] (suttam),
chants (geyyam), analyses (veyyākaraṇam), verses (gātham), utterances (udānam),
sayings (itivuttakam), birth stories ( jātakam), marvels (adhbhutadhammam), and
dialogues (vedallam).”9 Although I follow Ñānạ moli and Bodhi’s translation of sut-
tam as “discourses,” I provide brackets to register that as a singular noun (like all
the rest), it most likely refers to the Pātimokkha Sutta, the core formulary of the
monastic rule—a point of some importance, since it would indicate a foothold for
the second basket, the Vinaya Pit·aka, in this early dhamma classification.10 Here
the Buddha makes the point that these nine types of teaching are to be examined
“with wisdom” and not “for the sake of criticizing others and for winning in
debates.” According to Lamotte (1988, 149–50), reference to three canonical bas-
kets occurs in inscriptions by the second century BCE, their conventionally listed
order being the Sutta (Sanskrit Sūtra) Pit·aka, Vinaya Pit·aka, and Abhidhamma
(Sanskrit Abhidharma) Pit·aka.
These signposts can help us to keep aware of broad developments in the
canon formation of the “Nikāya schools.” I will organize this chapter around the
Three Baskets—but not in the conventional order just mentioned. Rather, pro-
ceeding by what I sense to be the emerging consensus of the chronology in which
the contents of these baskets filled out into the shape in which we have them,
I will move from the Suttas to the Abhidhamma and save the discussion of

6. Thanks on this point to Ani Kunga Chodron and Lama Kalsang Gyaltsen, personal communication,
September 2007.
7. Though see Bodhi 2000, 31–36: a valuable discussion of the four chief Nikāya collections; their dis-
tinctive features; and their tie-ins with Abhidhamma and Vinaya developments, mentioning throughout their
different presentations of the dhamma.
8. See Manné 1990, 77–81; Collins 1990; Veidlinger 2006, 18–20. Cf. Bodhi 2000, 30: “It is much more
likely that what took place at the First Council was the drafting of a comprehensive scheme for classifying the
suttas (preserved only in the memory banks of the monks) and the appointment of an editorial committee (perhaps
several) to review the material available and cast it into a format conducive to easy memorization and oral trans-
mission” (my italics).
9. Majjhima Nikāya [MN] 22.10 = M I 133 f. See Lamotte 1988, 144–45.
10. See section C. For a different explanation, see Lamotte 1988, 144.
106 dharma

Vinaya for the last, without ever abandoning the sutta texts in either of the latter
two cases.11 To begin with the Sutta Pit·aka is to start with the grouping of texts
that relates to the expertise of “those who maintain the Dhamma”: a “Dhamma”
grouping that refers to the Suttas already in the Mahāparinibbāna Sutta (DN 4.8)
on the Buddha’s last journey and in Vinaya accounts of the two earliest councils,
where, in both contexts, Dhamma and Vinaya are mentioned together as the sum
and standard of the Buddha’s teachings.12 To take up the Abhidhamma between
them is to recognize that, by the time the canon is taking this three-basket design,
Buddhism’s early scholastics were in the middle of things.
The main reason to present matters around the Three Baskets, however, is
to bring out the strikingly different ways that the Buddha’s dharma was devel-
oped in these three areas of specialization. Once one has the three baskets as a
way in which to encompass and transmit an understanding of the totality of the
Buddha’s multifaceted dharma, the “three baskets” become “an ideological con-
cept” (Veidlinger 2006, 20, citing Bizot 1976, 21). Whereas the Sūtras (Pāli,
Suttas) present the dharma’s public face, the other two baskets are in-house spe-
cializations. The sūtras present the Buddha dialoguing, telling engaging narra-
tives to all and sundry, and, even when he addresses only his own monastic
followers, never speaking, as he puts it, “with a closed fist”—that is, esoterically.13
This applies to the five main dialogue collections of the Pāli canon, but especially
to the Dīgha Nikāya, with its preponderance of debates, which “may be regarded
as an exercise of publicity, . . . Something is always at stake. . . . [Debates] are the
records, slanted no doubt in the Buddha’s favour, of public events. They are
entertainments for the purpose of propaganda” (Manné 1990, 73). The notion of
something “at stake” applies especially to what Manné calls “the dramatic debate,”
in which many of the opponents are Brahmins, 32.3 percent being “directed

11 . On “late” dating of the Vinaya, see Schopen 2006, 316–17; Clarke 2009, 4 n. 7 (“around the turn of the
Common Era”). On dating of Vinaya relative to the other baskets, see Bodhi 2000, 35 (later than the Sutta
Basket); Collins 1993, 335 n. 11 (repeated in Collins 1998, 66 n. 86): “recent evidence is tending to suggest that
the version of the [Pāli] Vinaya we have is a later redaction, although it too contains no reference to imperial for-
mations.” He means not only later than the four and the Sutta Nipāta, but apparently “later” than the “always
accepted to be late” Abhidhamma. See further. von Hinüber 1996, 3–14, 18–19, 26: “the cultural environment of
the first four Nikāyas of the Sutta-piṭaka is markedly older than that of the Vinaya-piṭaka”; 64 on the Abhidhamma
as “considerably younger” than both. Gombrich 1988, 91, granting some “strong ground” to “modern skepti-
cism,” despite his preference for early pre-Mauryan dating of the Vinaya; Frauwallner 1956, 154, on the Skandhaka
as “the first great literary work of Buddhism”—though attempting to reconstruct an early prototype of it as the
“original Vinaya” (see Holt 1983, 41–43, with discussion). On the question of pre-Mauryan sutta dating, see the
discussion of the Aggañña Sutta in § C of this chapter.
12. Sutta occurs rather than Dhamma in the Mahāparinibbāna Sutta standing, “so far as we yet know, quite
alone” (see Rhys Davids and Oldenberg [1881] 1968, 1: xxix). See Walshe [1987] 1995, 255.
13. “But Ānanda, what does the order of monks expect of me? I have taught the Dhamma, Ānanda, making
no ‘inner’ and ‘outer’: the Tathāgata has no ‘teacher’s fist’ in respect of doctrines” (Mahāparinibbāna Sutta [DN
16] 2.5; Walshe trans. [1987] 1995, 245).
 early buddhism 107

toward” them (45–57, 79). We shall discuss one of the most dramatic of these in
the next section, the Ambat··tha Sutta (DN 3). The Dīgha Nikāya also has most of
the “magical” (79) and what have been called “mythical” (Gethin 2006, 68 n. 15)
suttas.14 Such public outreach also seems to typify “new” texts included in the
sūtra baskets of the northern “Nikāya schools” (see chapter 7 § B for an example),
and quite clearly applies to Mahāyāna sūtras, which used the sūtra format to
authorize and disseminate new teachings as the Buddha’s Word. In sharp con-
trast, the Abhidharma and Vinaya are each almost totally for those in the
movement. The Abhidharma refines what those really in the know should know.
And the Vinaya defines what monks and nuns should (and should not) do in
common. Yet these two approaches also have their wider public purviews: the
Abhidharma in staking out Buddhism’s place in the erudite but politically impor-
tant settings—sometimes including royal courts—of scholastic debate with other
Indian philosophies; and the Vinaya in regulating the public interface between
Buddhist and non-Buddhist dharmas in practice. Coming to the Vinaya last will
best equip us by the end of this chapter to understand the ways these three bas-
kets bring together varied and overlapping teachings and set the terms for a
discussion of Buddhism’s treatment of the dharma over time in chapter 6. It will
also enable us to consider certain Mahāyāna departures from “Nikāya Buddhism,”
and to take stock of the complexities that distinguish Buddhist and Brahmanical
senses of dharma before beginning to size up the latter in chapter 5.

A. Sūtra Basket Dharma

The most accessible Sūtra Basket comes from the only Nikāya school to have
compiled its canonical scriptures in Pāli rather than Sanskrit: the Theravāda or
School of the Elders, which takes root in Sri Lanka. The Pāli equivalent of
Sanskrit sūtra is sutta.
While the Sutta Piṭaka includes a number of discourses noteworthy for their
basic instructional content, including the Buddha’s famous first sermon known as

14. Gethin mentions eleven of the Dīgha Nikāya’s thirty-four suttas as being more “mythic” than the
others, including the Aggañña Sutta, which will be discussed at the end of this chapter. Manné demonstrates the
DN’s public feature in contrast to the other of the two “‘story tellers’ collections” (1990, 70), the Majjhima
Nikāya, with its higher proportion of sermons and consultations (73–75), in which “60.5% are directed only to
monks,” and in which the “purpose [is] the presentation of the Leader, both as a real person and as an archetupe
(a Tathāgata), and the integration of new monks into the community and the practice. Most of the intimate bio-
graphical suttas appear in this collection” (80–81). Cf. Bodhi 2000, 31; 2005, 11–12, building on Manné’s con-
tribution, and Bodhi 2000, 334 on the Aṅguttara Nikāya, open to “the conventional world of consensual
realities” and presenting people “engaged in a heartfelt quest for happiness and freedom from suffering.” Cf.
Bronkhorst 1985a, 316, on the Saṃyutta Nikāya arranging matters “subject wise” and the Aṅguttara following
“a scheme determined by the number of subdivisions in the items discussed.”
108 dharma

“The Turning of the Wheel Sutta,” it also includes many elegant dialogues famous
as narratives, some of which, as noted, can be called myths.15 Such narratives often
portray the Buddha in interactive settings where his artful teachings, graced with
maieutic ethical and philosophical reasoning and backed by stunning similes, have
their maximum impact. We will open this basket by beginning with suttas of this
type and save discussion of more directly instructional or catechetical teachings for
the next section on Abhidhamma, where we can examine how they served as the
basis for further instructional unfolding. The interactive settings can direct our
initial focus to the likely historical background that the Suttas reflect. We will look
especially at the Buddha’s interactions with kings and others deemed members of
the Khattiya (Sanskrit Kṣatriya) social class; with Brahmins and members of the
lowest social classes; and with householders and his monastic followers.
Keeping in mind Olivelle’s hypothesis linking dharma and royalty,16 let us
turn first to the Ambaṭṭha Sutta (DN 3), in which the Buddha raises the matter
of his royal descent while describing the Sakyans, the clan from which he comes,
as Khattiyas (Kṣatriyas), and saying that “the Sakyans regard King Okkāka as
their ancestor” (DN 3.1.15; Walshe trans. [1987] 1995, 114). Pāli “Okkāka” is
equivalent to Sanskrit “Ikṣvāku.” The Ikṣvākus, it will be recalled, are the royal
clan of Kosala. Along with the kings of Videha (such as Janaka) further to the
east, they had formed “the new political center in eastern North India” (Witzel
1997a, 307) that backed the third and last main phase of Vedic canonization. We
have discussed the late Vedic texts that were composed in this milieu. It is now
in the same area that we find the Buddha claiming that his clan has Kṣatriya
ancestry going back to the eponymous ancestor of the Ikṣvāku line, if not further,
to the first Khattiya, Mahāsammata, to whom the Buddha traces the origin of
kingship in the Aggañña Sutta (DN 3.27.20–21).17 But although it is the same
area, linguistic (Caillat 1997, 26 and passim), social, and political considerations
point to a “gap” of as many as two hundred years18 between the Vedic texts and
the conditions portrayed in the Nikāyas. Both the older tribes of the area and the

15. Among the eleven Gethin 2005, 68 n. 15 mentions in the DN are many that I will mention briefly,
including the Cakkavattisīhanāda and the Mahāsudasana, on which he is writing.
16. Olivelle says dharma was “a natural choice to define the new dispensation” of Buddhism because of its
royal associations: “My hypothesis is that the Buddha borrowed dharma as he did many other symbols to locate
and articulate his new religion. . . . The Buddha’s doctrine is compared to a wheel, a metonym for the war chariot
and conquest; and his first sermon is . . . ‘the Sūtra that set the wheel of dharma rolling.’ The Buddha’s teaching
is śāsana, the counterpart of a royal edict. These are . . . royal symbols used, deliberately I think, to define a new
ascetic group and a new religious ideology” (2004a, 504). As Holt points out too, “justification for viewing the
Buddha’s authority in terms of royal sovereignty” also has to do his being the one who “rules over the religious
life according to the norms of his Dhamma and Vinaya” (1983, 52).
17. According to Walshe 1995, 605 n. 838, Mahāsammata is “the first king of the solar race and ancestor,
among others, of the Sakyan rulers (and hence of Gotama).” Cf. Strong 2010.
18. See above, n. 4. According to Witzel, the Pāli texts “tell us that the area was inhabited not just by the
Kosala-Videhas but by a large number of tribes.” “The important Vedic tribe of the Videha thus is only one and
not a particularly prominent member of the Vajji confederation.”
 early buddhism 109

ones that have recently immigrated there have by now established themselves as
new polities with varied social systems, from monarchic to oligarchic.19 A “fully
developed town civilization” has emerged. And most important, the second
urbanization of India is now in full swing with aggressive metropolitan states
not only in Kosala, ruled by King Pasenadi, but in Kosala’s looming rival
Magadha (barely mentioned in the Vedic texts) under Kings Seniya Bimbisāra
and his son Ajātasattu—all contemporaries of the Buddha.20 The Buddha can
compare the discovering of the dhamma to finding an ancient path to an old and
forgotten city (Nidānasaṃyutta, Saṃyutta Nikāya [SN] 104–7).21
The Ambaṭṭha Sutta can be read with these “melting pot” conditions as
background. Its ambivalent portrayal of King Okkāka and his line are flavored
with contending Brahmanical and Buddhist views on caste and kingship, some
of which would seem to have been simmering for a long time. The Ikṣvākus
“may go back all the way to those once mentioned in the Rigveda (10.60.4), who
may have descended from the royal family of the Pūru tribe,” with some of
them moving eastward after the Pūrus’ defeat in the Battle of the Ten Kings.
But in the Jaiminīya Brāhmaṇa, the Kuru-Pañcālas look down upon the Ikṣvākus
to their east as a “despised and down-trodden” tribe that had eaten Asura food
(Witzel 1997a, 279 n. 93, 318–19). As to the Śakyas, the Śatapatha Brāhmaṇa
includes them among “easterners and others” who have round Asuric graves,
which may be precursors of the Buddhist stūpa (Ibid., 310). Even as an ancestor
of the Sakyas, King Okkāka/Ikṣvāku gets a rather ambivalent portrayal in early
Buddhist sources as the presumed eponym of the Buddha’s own line. In the
Sutta Nipāta, often considered as supplying early sutta material, one learns that
it was under Okkāka’s reign that the degeneration of Brahminhood was brought
about by the institution of animal sacrifice.22 In the Mūlasarvāstivāda Vinaya,
the line going back to the first king Mahāsammata has degenerated and is

19. “The many tribes well known from the Pāli texts, such as the Sakya, Malla, Vajji, Licchavi, Naya,
Kalāma, Buli, Moriya, Vesali, etc. do not (yet) appear in the eastern Vedic texts” (Witzel 1997a, 308).
20. See Witzel 1997a, 308–9 (the second urbanization gets truly underway ca. 450 BCE), 318, 321, 329, 333–34.
21. This passage would seem to be reminiscent of the city Kusāvati, where the Buddha in a previous life
once lived as king Mahāsudassana, dying there in its “Dhamma Palace” (dhamma-pāsāda), which he tells Ānanda
about as he prepares to die at the forlorn town of Kusinārā, where Kusāvati once stood, in the Mahāsudassana
Sutta (DN 17; cf. Mahāparinibbāna Sutta [DN 16] 5.17–18). On this Dhamma Palace as a place for death in
meditation, structured cosmologically as a seven-ringed city horizontally and vertically as the rūpadhātu (world of
form) in which to practice the four Brahmavihāras (the “sublime attitudes” of friendliness, compassion,
sympathetic joy, and equanimity), see Gethin 2005, 74–76, 80, 88–93.
22. See Bodhi 2000, 171–72, 402 n. 214. In this note to Saṃyutta Nikāya 3.9, where the Buddha describes
five “great sacrifices” (mahāyañña), including the Aśvamedha, Rājasūya, and Vājapeya, which “great seers” should
not attend, Bodhi cites the commentary on Sutta Nipāta 299–305, which describes the origin of these rites as
follows: “in the time of the ancient kings the first four sacrifices were actually the four bases of beneficence
(saṅgahavatthu)—giving, pleasant speech, beneficent conduct, and equality of treatment—by means of which
kings conferred benefits on the world. But during the time of king Okkāka the brahmins reinterpreted the bases
of beneficence (which they augmented to five) as bloody sacrifices involving slaughter and violence.” Cf. Tsuchida
1991, 93 n. 57 on Okkāka, and 87 on these five implicitly Vedic sacrifices—on which see further Falk 1988.
110 dharma

about to die out when the son of the last king, falsely accused of murdering a
prostitute, is dying on an impalement stake; then two drops of his semen min-
gled with blood fall from his body and, warmed by the sun, give birth to twin
boys who get the name Ikṣvāku because they take refuge in a sugarcane thicket
(ikṣu-vāta), with the younger twin eventually becoming king under this name
(Strong 2009, 48–49; 2010). In fact, Ikṣvāku does not fare much better in the
Rāmāyaṇa, which celebrates the “house of Ikṣvāku” (ikṣvāku-kula, -vaṃśa) that
puts his name at the source of Rāma’s lineage, but does not say much about
Ikṣvāku himself.23 If Okkāka is the best claim the Sakyas have to royal descent,
then that also could be reflected in the charge made by the leading Brahmins in
the Ambaṭṭha Sutta that their rank among Kṣatriyas is tenuous at best. But sim-
ilar ambiguities pertain to what the Ambaṭṭha Sutta has the Buddha say about
the origins of the regional “Kaṇhāyan” Brahmins.
The Ambaṭṭha Sutta thus gives voice to various challenges that, according
to the Nikāyas, the Buddha is alleged to have put to proponents of Vedic
orthopraxy. It can thus serve to introduce some of the complexities of that
challenge, which is, of course, the challenge of the Buddhist dhamma specif-
ically to Brahmins, who, as Gombrich has noted, are among the Buddha’s
more frequent “interlocutors in the Canon.”24 Indeed, we should note that the
Buddha seems to have expected more from his dialogues with Brahmins than
he did from those at the lower end of the social spectrum. In the Alagaddūpama
Sutta (MN 22), when the Buddha says the nine “parts” of the dhamma are to
be examined “with wisdom” and not to win debating points, he is correcting
an errant “bhikkhu named Ariṭṭha, formerly of the vulture killers,” who has
been caught misrepresenting the teaching with his claim that what the
Buddha has called “obstructions” (which include sensual desires) do not
obstruct the person who engages in them—that is, including sexual
intercourse! The Buddha admonishes Ariṭṭha and the other assembled
monks: “Bhikkhus, that one can engage in sensual pleasures (kāma) without
sensual desires (kāma), without perceptions of sensual desire, without
thoughts of sensual desire—that is impossible.” From this topic he turns to
the simile of the water snake: that when one needs a snake, it makes all the
difference whether one grasps it wrongly or rightly. Without precisely saying
so, the Buddha thus compares the dhamma to a snake (no cigars yet) before
going on to make his famous simile that the dhamma is like a raft—“for the

23. After Manu built Ayodhyā (1.5.6), Ikṣvāku, as Manu’s son (2.102.6), was given the (in itself rather
small; see Pollock 1984, 402–3) land of Kosala (2.43.11), where it is by Ikṣvāku’s “grace” that “all the great kings
of Viśāla are long-lived, mighty, and righteous” (1.46.18). Viśāla, a second city of Kosala founded by Ikṣvāku’s son
Viśāla (1.46.11–12), is Vaisali in Buddhist sources.
24. See Gombrich 1984, 98; Schumann 1989, 187–93; Bailey and Mabbett 2003, 112, 121–29, 134–37, 252, 261.
 early buddhism 111

purpose of crossing over, not for the purpose of grasping”—that you leave
behind when you get to the other shore. Now as Schumann observes, “It is
noteworthy that it was always monks of casteless origin whom the Master had
to reprove for misinterpretations”; “It is noteworthy that in both cases the
Canon mentions the humble origin of the monks so blamed. It seems that
Gotama expected men of no education to have ethical qualities, but not much
power of understanding” (1989, 188; 205). The second reference is to “Sāti,
son of a fisherman,” who also had a “pernicious view” (MN 38). In effect, by
calling attention to their outcaste jātis or “births,” the Buddha was practicing
jāti profiling.25
The Ambaṭṭha Sutta [henceforth AmbS] runs the gamut on these matters,
but it is a dialogue with Brahmins. The best discussion of it I know of, by
Ryūtarō Tsuchida, places it in that context, and brings out two things to keep in
mind. First, as Tsuchida shows, although it is not possible to draw from the
Buddhist canon an “exact and comprehensive categorization” of different types
of Brahmins, “at least two main groups” are repeatedly delineated: “wealthy
Vedic masters living in villages and towns,” the most prominent of whom are
called brāhmaṇamahāsāla, “Brahmins having great halls”; and ascetics with
matted hair called jaṭilā. The AmbS presents one of the exemplary Brahmins of
the great halls, but it also, as we shall see, describes Brahmin ascetics (Tsuchida
1991, 53–54, 60, 90). Secondly, Tsuchida shows that the “longer” suttas of the
Dīgha and Majjhima Nikāyas “do not have so much the character of historical
records as that of literary frameworks, into which doctrinal materials can be
incorporated as the main theme of each particular Sutta.” (1991, 77). They thus
make the task of recognizing Brahmin types easier by deploying “stock-
phrases,” “formulae,” “stereotypical expressions,” and “fixed prose passages.”26
Yet even while they run their characters through such central casting scripts,
these “Sutta authors display their literary skill by giving us lively portraits of
particular Brahmins” (53–54)—ones whose individuality emerges precisely
through and beyond the stereotypes. As we see the Buddha engaging the lead
Brahmin characters of the AmbS, let us compound our attention to the ambi-
guities of caste and kingship with an open question: where is the dhamma in
this text?

25. For a more somewhat more sympathetic reading of these usages, and also citing many others, see
Parasher-Sen 2006, 441.
26. Tsuchida 1991, 77, goes on to mention that the brāhmaṇamahāsāla Jāṇussoṇi is converted in six dif-
ferent suttas, each time by a different sermon of the Buddha! Cf. Manné 1990, 71, 78, carefully speaking “of (oral)
literature” and also of “the strict literary style” (82) of DN and MN suttas, while earlier citing in full the “formulas”
by which to recognize debates, consultations, and sermons in them, many of which we now meet in the AmbS.
For me, granted that there are always questions about prior orality and continued oral usages, the literary
character of the DN and MN suttas, as we have them, is, at least for the ones I discuss, a given.
112 dharma

The Buddha is touring Kosala with some five hundred monks and stays
near a Kosalan Brahmin village. A Brahmin named Pokkharasāti27 lives there
at a populous and well-stocked “handsome estate” (Tsuchida 1991, 56) named
Ukkaṭṭha that has been given to him by King Pasenadi of Kosala as a “royal
gift and with royal powers (rājadāyaṃ brahmadeyam).”28 As Tsuchida indi-
cates, the practice of giving such estates to Brahmins “suggests the close rela-
tionship of Brahmin landownership with royal power.”29 It appears from
Pokkharasāti’s mention in several of the Buddha’s dialogues with Brahmins
that his “authority and reputation . . . were not restricted to his native region
of Kosala but extended throughout the Brahmin society of northeastern
India.”30
Pokkharasāti has heard a “good report” about Gotama: among other things,
that he is “a fully enlightened Buddha” and that he “teaches a Dhamma that is
lovely in its beginning, lovely in its middle, and lovely in its ending.” He sends
his pupil Ambat.t.ha to find out about the good report,31 and to test whether
Gotama is a Great Man by seeing if he has the thirty-two physical marks that,

27. According to Malalasekera [1937] 1983, 2: 247, Sutta commentaries tell of his celestial hue and fea-
tures, and that in the time of the Buddha Kassapa he was a Brahmin “versed in the three Vedas who, having heard
the doctrine and given alms, was reborn in the deva-world. Thereafter, scorning birth in the womb of a woman,
he sprang to life in a lotus” in a Himalayan pond. An ascetic found him and taught him the Vedas, and the king
gave him his estate because of his Vedic learning. The lotus birth gives him his name, which in the Sanskrit
Divyāvadāna is Puṣkarasārin, “Having the Essence of the Lotus.”
28. According to Walshe ([1987] 1995, 548 n. 141), who draws on an earlier translation, this stock phrase
means that the “royal gift” allows him to rule over it “as if he were a king.”
29. See Tsuchida 1991, 57, preceded by a table of seventeen brahmadeyya-s and/or brāhmaṇagāma-s men-
tioned in the Nikāyas, with eleven in Kosala, four in Magadha, one in Aṅga, and one in Malla (56–57). Such estates
seem to have populations consisting mostly of other castes. The prominence of these Brahmans “seems to owe
itself primarily to authority they held within society as orthodox vedic masters and ritual priests” (57). Cf. the con-
cern in Brahmanical dharma texts over the inviolability of Brahman property, to be discussed in chapter 5.
30. Tsuchida 1991, 55. As Tsuchida brings out, he is “the most prominent and authoritative figure”
among the five brāhmaṇamahāsālā in the “Kosala group” of five such persons (54; see Subha Sutta (MN) 99.13;
Tevijja Sutta (DN) 13.2). The five include Caṅk ī and the interesting Jāṇussoṇi, who, at the end of the Subha
Sutta, gets off his fancy chariot in public view to salute—from what he has just learned second-hand of the
Buddha’s teaching—the gain it brings to King Pasenadi that a fully enlightened Tathāgata dwells in his realm
(MN 99.30; cf. MN 27.8). This public avowal has its opposite in the request by the freshly converted Brahmin
Soṇadaṇḍa that the Buddha not take it amiss if, instead of alighting from his chariot to salute him in public,
for which “he will surely incur the reproach of the (Brahmin) assembly,” he merely raise his goad (DN 4.26).
See Tsuchida 1991, 76–77, who notes of this assembly (parisā; Sanskrit pariṣad) that “perhaps it refers to some
kind of assembly presided over by eminent Brahmins” [93 n. 65]). Cf. the five mahāśāla Brahmins who hear
the teachings of King Aśvapati Kaikeya at ChU 5.11.1, mentioned in chapter 3 § F. That mahāśāla Brahmins of
the Kosala group could serve as court chaplains is indicated by the Majjhima Nikāya commentary, which,
according to Ñāṇamoli and Bodhi 2005, 1174 n. 56, says of Jāṇussoṇi that his name “was not a given name but
an honorific title meaning ‘royal chaplain’ (purohita) bestowed on him by the king.”
31. One might, however, also infer from MN 99.10 (Subha Sutta) that, more than just wanting to validate
the “good report,” Pokkharasāti has an axe to grind, for he is represented there as thinking that the claims of
some Brahmins and recluses to reach superhuman states and “distinctions in knowledge and visions worthy of
the noble ones” are “ridiculous.”
 early buddhism 113

“according to the tradition of our mantras,”32 indicate that he would have had
the choice of becoming either “a wheel-turning righteous monarch of the law”
(dhammarāja) or a “fully enlightened Buddha” (1.5). Young Ambat.t.ha is a “stu-
dent of the Vedas, who knows the mantras,” and fully shares his master’s
knowledge. So far, virtually everything has been stock material, paralleled in
other suttas, and so it will continue, leaving us little need for further such
notices except where something distinctive begins to happen, as it does now.
Soon invited into the Buddha’s dwelling, Ambat.t.ha shows discourtesy by
walking and standing rather than sitting like his seated host. Asked would he
behave like this if he were talking to venerable and learned Brahmins, he
says, “No, Reverend Gotama. . . . A Brahmin should walk with a walking
Brahmin, . . . sit with a sitting Brahmin. . . . But as for those shaven little
ascetics, menials, black scourings from Brahmā’s foot, with them it is fitting to
speak just as I do with the Reverend Gotama” (1.10). Ambat.t.ha’s ultimate proof
text is obviously meant to be an implied but also modified Puruṣa Sūkta with
Brahmā in place of Puruṣa,33 such as occurs similarly in Bṛhadāraṇyaka
Upaniṣad 1.4, where, as we have seen (see chapter 3 § F), the four social classes
emerge along with dharma as part of the “full development” of brahman.
It will be left hanging whether young Ambaṭṭha ever changes his views on
these insulting matters, but the Caṅkī Sutta ends with the young scholarly
Brahmin Kāpat.hika drawing the right lesson on the same subject: “Formerly,
Master Gotama, we used to think: ‘Who are these bald-pated recluses, these
swarthy menial offspring of the Kinsman’s [Brahmā’s] feet, that they would
understand the Dhamma?’ But Master Gotama has indeed inspired in me love
for recluses, confidence in recluses, reverence for recluses” (MN 95.34; Ñānạ moli
and Bodhi 2005, 785). The Buddha has engaged Kāpaṭhika even after he has
repeatedly interrupted the Buddha, and been defended by Caṅkī as a very learned
and wise young scholar with whom the Buddha ought still to engage in
discussion, to which the Buddha agrees (MN 95.11–12). For present purposes,
let me just introduce a point that will bear further unfolding. In several of the
suttas in which the Buddha holds dialogues with Brahmins, there comes a point
where one or more senior Brahmins turn matters over to a junior Brahmin

32. See DN 30 and Walshe 610–11 n. 939 positing that although the thirty-two marks “clearly must have
been important in the minds of influential Brahmins in the time of the Buddha, . . . Brahmin tradition has pre-
served very little about them.”
33. See MN 50.13, where Māra Dūsi puts such thoughts into Brahmins’ minds by possessing them to oppose
the monks (bhikkhus) or recluses (samaṇas), and Ñānạ moli and Bodhi 2005, 1251 n. 524, providing the commentarial
explanation that “it was the belief among the brahmins that they were themselves the offspring of Brahmā’s mouth,
the khattiyas of his breast, the vessas of his belly, the suddas of his legs, and the samaṇas of the soles of his feet.” Other
than the Brahmins holding place at the mouth, everything else is centrally (breast rather than arms) or upwardly
(belly rather than thighs, legs rather than feet) displaced to make room for the fifth group at rock bottom.
114 dharma

pupil to match wits with the Buddha.34 From the standpoint of such suttas, it is
as if the senior Brahmins unwittingly or inadvertently put the Buddha in a posi-
tion to test the depth of the young men’s Vedic education, which, of course,
always comes up somewhat short. Indeed, in some of these cases one could
compare him to an initiation master. The point that bears watching, however, is
that even though the Buddhist texts do not offer a precise designation of the
status of these young men, it would appear that they are proximate to a type that
Brahmanical dharma texts call the snātaka or “bath graduate”: the postgraduate
twice-born male, but especially Brahmin, who has undergone the sacred bath
that marks the completion of his Vedic education, after which he may remain in
an intermediate celibate state until marriage, or can even continue to be called a
snātaka after marriage.35 As Tsuchida shows, Buddhist canonical texts have the
equivalent word n(a)hātaka, though it “does not occur very frequently.” But its
usages emphasize “scholarly rather than priestly aspects of the Brahmin” as “a
virtual synonym” for sottiya or “vedic scholar” (Sanskrit śrotriya) and tevijja or
one who has knowledge of the three Vedas (Sanskrit traividya), which the Pāli
texts use more frequently (1991, 70–71).
Ambat.t.ha’s Vedic education is thus being made subject to a Buddhist exam-
ination. And when the Buddha replies to him that his training should have
made him more courteous and Ambat.t.ha hears himself “being called untrained,”
he angrily explains himself by insulting the Sakyans as “menials” no less than
three more times, of which the second is the most interesting. When the Buddha
asks what the Sakyans have done to Ambat.t.ha that he should so insult them,
Ambat.t.ha has a story. Once he went to Kapilavatthu, the Sakyans’ capital,36 on
some business for Pokkharasāti, and while the Sakyans were “laughing and
playing about together” on “the high seats in their meeting hall,” no one offered
him a seat. The Buddha passes this off as “a trifle,” saying, “But Ambat.t.ha, even
the quail, that little bird, can talk as she likes in her own nest. Kapilavatthu is the
Sakyans’ home.” Although Ambat.t.ha never denies that the Sakyans are

34. In Assalāyana Sutta (MN 93), the scholarly sixteen-year-old Assalāyana is promoted by the Brahmins to
dispute the Buddha on “purification for all the four castes.” In Soṇadaṇḍa Sutta (DN 4) 20–22, Soṇadaṇḍa
upholds his young nephew Aṅgaka as an exemplary young scholar and pupil to make a point about wisdom and
morality that the Buddha applauds. And in Brahmāyu Sutta (MN 91) 4–22, the young student Uttara shadows the
Buddha on behalf of his aged preceptor Brahmāyu, the leading householder Brahmin of Mithilā (27). In several
sutta narratives without this explicit frame of senior Brahmin(s) putting forward a junior one, the Buddha also
discourses with Pokkharasāti’s pupils (Tsuchida 1991, 55): see Vāset.t.ha Sutta (MN 98), Tevijja Sutta (DN 13), and
Aggañña Sutta (DN 27), discussed at this chapter’s end, all dialogues with the young Vāset.t.ha and Bhāradvāja;
and Subha Sutta (MN 99) with the young Subba, which, like the Tevijja, treats the four brahmavihāras as the way
to Brahmā. The fatherless “young householder” Sigālaka, who receives the Buddhas instruction on “the right way
to pay homage to the six directions,” is also presumably a Brahmin in Sigālaka Sutta (DN 31).
35. See Tsuchida 1991 70–72. “Bath graduate” is Olivelle’s 1999 and 2005a translation of snātaka.
36. And near the Buddha’s birthplace at Lumbinī.
 early buddhism 115

Khattiyas, the net effect is that he denounces them as crude ones who “do not
pay homage to Brahmins” as menials should (3.1.12–14).
Saying, “This young man goes too far in abusing the Sakyans,” the Buddha
tells him not only that “the Sakyans regard King Okkāka as their ancestor,” but,
turning the tables, that Ambaṭṭha’s own Brahmin clan descends from one of
Okkāka’s slave girls (dāsis). As Bronkhorst has demonstrated, the name
Ambaṣṭha almost certainly references the Ambaṭṭhas, one of the “mixed-caste”
groups enumerated in most of the dharmasūtras and Manu, which agree that
“Ambaṣṭhas were thought of as descendants of a mixed marriage in which the
father belonged to a higher class than the mother, the father most typically being
a Brahmin, the mother a Vaiśya” (2007, 355). In fact, the only dharmasūtra not
to mention Ambaṭṭhas in this way is the Ᾱpastamba Dharmasūtra, for which, as
Olivelle points out, one of the signs that it is earlier than the rest is that it “does
not deal with mixed classes at all, a topic found in all other Dharmasūtras and in
the later Smṛtis” (1999, xxxi; cf. Jha 1970). This reinforces Bronkhorst’s evi-
dence that the naming of Ambaṭṭha as this sutta’s chief interlocutor would be
“late” (by which he seems to mean later than Alexander the Great [2007, 353]),
and that its author would have chosen this name because “cultivated early lis-
teners to the story would know, right from the beginning, that Ambaṭṭha was
not what he claimed to be, viz., a pure-blooded Brahmin. They would know
immediately that he was an empty boaster” (355). Indeed, among Bronkhorst’s
reasons to include the AmbS in a discussion of “late” suttas is its link with the
other two suttas mentioned above about students of Pokkarasāti (353–54). Note
that in contrast to the AmbS, the two others give names of the highest Brahmin
and indeed Vedic pedigree to the two young Brahmins who, unlike Ambaṭṭha,
will become the Buddha’s disciples. Vāseṭṭha and Bhāradvāja would be descen-
dants of the great Vedic poets Vasiṣṭha and Bhāradvāja!37 Here the Buddha is
not so much jāti profiling as he is staying ahead of the Brahmins’ game.
Three times the Buddha presses Ambaṭṭha to admit that he too has heard
about his clan’s mixed origins from learned Brahmins (3.1.15–21), finally backing
his persuasion with a threat that if Ambaṭṭha does not answer the third posing,
his head will split into seven pieces, which is reinforced by the Yakkha Vajirapāni,
whom both of them see holding an iron club in midair and ready to strike.38 But
now, hearing Ambaṭṭha censured by his own companions, the Buddha pulls
back. Not wanting these young men to humiliate Ambaṭṭha further, he allows

37. Bronkhorst also calls attention to the location of the Ambaṣṭhas “among the western people conquered
by Nakula” in the Mbh (2007, 355, citing Mbh 2.29.6 and 19) as evidence “that the author or inventor of this
Buddhist story had heard of” a people that “lived far to the west of the area where the Buddha taught” (355–56).
The AmbS has them come from the south.
38. One might hold this incident in mind for chapter 9’s discussion of “the Yakṣa’s Questions.”
116 dharma

that the slave-girl’s son Kaṇha39 “was a mighty sage” (Isi, Sanskrit Ṛsị ) who “went
to the south country, learnt the mantras of the Brahmins there, and then went to
King Okkāka and asked for his daughter”; when the outraged Okkāka readied his
bow and arrow, the sage was able to put a spell on the weapons until Okkāka,
fearing divine punishment (brahmadaṇdạ ), relented and gave the “mighty sage”
his daughter (22–23). As if invoking Brahmanical Law, the Buddha then takes up
two cases of intercaste marriages and one of banishment from caste to demon-
strate, on highly dubious Brahmanical grounds,40 that Khattiyas are higher than
Brahmins. And in summation, he twice quotes a verse “pronounced by Brahmā
Sanatkumāra”: “The Khattiya’s best among those who value clan; He with
knowledge and conduct is best of gods and men.”41 The two lines of the saying
are clearly contrastive in juxtaposing what is “best” for two different audiences:
those who value clan and those who follow the teachings on “knowledge and
conduct” which characterize those on the Buddhist path.42 The speaker of this
verse, who is also its source in the Aggañña Sutta, is not the “creator” Brahmā but
one of his five sons called “Forever Young.”43 As mentioned above, there too the
Buddha is also talking to Pokkharasāti’s young pupils.
For the first time, Ambat.t.ha is drawn into questioning: “But, Reverend
Gotama, what is this conduct, what is this knowledge?” The Buddha reveals
first what they are not. They are not concerned with reputation based on birth
and clan or the conceit of giving and taking in marriage; rather, they are declared
“from the standpoint of the attainment of unexcelled knowledge-and-conduct”
that comes from “abandoning all such things” (3.2.1). So Ambat.t.ha asks again,
“But, Reverend Gotama, what is this conduct, what is this knowledge?”

39. The name was given because the baby was “black” (3.1.16). Walshe [1987] 1995, 549 n. 152 asks, “Is he
to be identified with Krishna?” There is no narrative basis for an identification with either Kṛṣṇa Vāsudeva or,
more plausibly since he is a great Ṛṣi, Kṛṣṇa Dvaipāyana, author of the Mahābhārata.
40. Should a male Khattiya marry a Brahmin female, or the reverse, Brahmins would give their son a seat,
teach him mantras, and not cover their women before him; but Khattiyas would not give him their Khattiya con-
secration. A banished Brahmin would not be received by Brahmins but a banished Khattiya would (this case
seems to be fallacious, since, although the Buddha asks whether Brahmins would teach a banished Brahmin
mantras, and hears “no,” he does not ask whether they would teach mantras to a banished Khattiya).
41. AmbS 3.1.24–28. In this, the Buddha moves not only from Brahmins to Kṣatriyas as the standard of
orthopraxy, but to dharma as a teaching of what is best for the full range of human society and the world. Beyond
this sutta, Kṣatriyas not only set the standard in the oligarchic polity which the Buddha nostalgically favors but in
the emerging reality of monarchical states.
42. The AmbS is third among the first thirteen suttas in the DN, each of which, according to Gombrich
(1984, 91–133), has a section on sī la or conduct, meaning “decorum.” In the Soṇadaṇḍa Sutta, the fourth and next
in sequence, the Buddha puts sī la on a par with pañña or wisdom, and has also been talking about both together
in the AmbS.
43. See Walshe [1987] 1995, 580 n. 516. According to the Majjhima Nikāya commentary, he “was a youth
who attained jhāna, passed away, and was reborn in the Brahma-world, retaining the same handsome form he
possessed in his existence in the human world” (Ñāṇamoli and Bodhi 2005, 1255 n. 566). This same verse is
quoted and approved by the Buddha at the conclusion of the Aggañña Sutta (DN 27.33). See Collins 1993, 349,
378–79 and discussion later in this chapter.
 early buddhism 117

Ambat.t.ha, a Tathāgata arises in this world an Arahant, fully-enlightened

Buddha, endowed with wisdom and conduct, Well-Farer, Knower of the
worlds, incomparable Trainer of men to be tamed, Teacher of gods and
humans, enlightened and blessed. . . . He preaches the Dhamma which
is lovely in the beginning, lovely in its middle, lovely in its ending, in
the spirit and the letter, and displays the fully perfected and purified
holy life. A disciple goes forth and practices the moralities; he guards the
sense doors, etc.; attains the four jhānas. Thus he develops conduct. He
attains various insights and the cessation of the corruptions. . . .44 And
beyond this there is no further development of knowledge and conduct
that is higher or more perfect. (2.2)

The Buddha adds, however, that there are “four paths of failure”45 along which
ascetics (samaṇas) or Brahmins fail “in pursuit of this unexcelled attainment of
knowledge and conduct.” According to Tsuchida, this grouping of four seems
to sort out ascetics and Brahmins according to the types and locations of their
hermitages or assamas (Sanskrit āśramas), although that word does not occur.
The third category, the ascetic or Brahmin who “builds himself a fire-shrine on
the outskirts of some village or town and dwells there tending his sacred fires”
(2.3; Tsuchida trans. 1991, 85), strikes Tsuchida as a likely approximation, on
the Brahmin side, of the matted hair jat·ila ascetics, who are similarly depicted
in other canonical sources (Ibid.). After hearing of these four in the order of
their decreasing austerity, Ambat.t.ha has to admit that he and his teacher not
only fall short of the highest standard of “unexcelled knowledge and conduct”
but are incapable of undertaking any of the “four paths of failure,” down even
to the easiest! “And yet,” replies the Buddha, “you and your teacher the Brahmin
Pokkharasāti utter these words: ‘These shaven little ascetics, menials, black
scrapings of Brahmā’s foot, what converse can they have with Brahmins learned
in the Three Vedas?’—even though you can’t even manage the duties of one
who has failed. See, Ambat.t.ha, how your teacher has let you down!” (2.5).
Although we were told that Ambat.t.ha “fully shares his master’s knowledge,”
this is the first we hear that the abusive words were his teacher’s as well as his.
Says Tsuchida, although the four hermit types are “pitiable people” from the
standpoint of their following knowledge and practices that the Buddhist path
excels, still, “the attitude of these canonical authors towards Brahmanical her-
mits is not entirely degrading” since “at least they are more highly esteemed
than mahāsāla-Brahmins” such as Pokkharasāti (1991, 85).

44. The italics and ellipsis indicate abridged allusions to the longer teachings in the preceding sutta, the
Sāmaññaphala Sutta (= DN 2). See Walshe [1987] 1995, 549 n. 157.
45. Walshe [1987] 1995, 119–20; literally “outlets of loss” or “leakages” (Ibid., 549 n. 158; Tsuchida 1991, 85).
118 dharma

The Buddha now tells how Pokkharasāti lives opulently “by the grace and
favour of King Pasenadi of Kosala,” yet confers with the king only through a
curtain rather than face to face (3.2.6)—which, the Buddha need not mention,
is the way he converses with King Pasenadi.46 And he gets Ambaṭṭha to agree
that, just as “some workman or workman’s servant” cannot stand in place of
the king and speak for him, so the present-day Brahmins, with their opu-
lence, amusements with women, and fortified towns, are not like “the first
sages of the Brahmins,47 the makers and expounders of the mantras whose
ancient verses are chanted, pronounced and collected48 by the Brahmins of
today” (3.2.6–10). As Tsuchida observes, the AmbS puts this “neutral” stock
unit on the ancient vedic Ṛṣis and their mantras to a rather favorable depic-
tion here of the sages as having none of the ostentation of their current-day
mahāsāla successors, whereas in other suttas the Buddha puts the same Ṛṣis
at the head of “a procession of the blind” (andhaveṇi) that runs from the “first
generation of vain and ignorant Brahmins” down to those with whom he is
speaking.49
Possibly holding back on such a negative evaluation, the Buddha now gives
his dialogue with Ambaṭṭha its final turn. Telling Ambaṭṭha that neither he nor
his teacher is “a sage or one trained in the way of a sage,” he says he will clarify
Ambaṭṭha’s “doubts and perplexities” concerning his person and answer his
questions. Descending from his lodging, he starts to walk with Ambaṭṭha, and,
aware that Ambaṭṭha sees all but two of his thirty-two marks, he “effected by
his psychic power that he could see his sheathed genitals, and then, sticking
out his tongue, he reached out to lick both ears and both nostrils, and then
covered the whole circle of his forehead with his tongue.”50 Finding all the
marks accounted for and considering his mission accomplished, Ambaṭṭha
asks leave and returns to Pokkharasāti, whom he finds “sitting in his park with
a large number of Brahmins, just waiting for Ambaṭṭha” (3.2.10–13). First,
Ambaṭṭha reports that the Buddha has “the thirty-two marks of a Great Man,
all complete, with nothing missing.” Then he recounts their conversation,
upon which Pokkharasāti exclaims, “Well, you’re a fine little scholar,” berates

46. See Aggañña Sutta 3.27.8 (DN 27) on Pasenadi’s homage to the Buddha despite the Sakyans’ homage
to Pasenadi, discussed in Collins 1993, 340–41. Saṃyutta Nikāya 3, the Kosalasaṃyutta, gathers together the
Buddha’s many conversations with King Pasenadi.
47. This would be one of the Buddha’s more positive characterizations of the ancient Vedic Brahmins,
though still couched in ambiguity (see next note).
48. It is, of course, interesting that the Buddha says the Brahmins are still collecting (presumably Vedic)
mantras.
49. Tsuchida 1991, 74; on this simile, see Tevijja Sutta (DN) 13.15; Caṅkī Sutta (MN) 95.13; Subha Sutta
(MN) 99.9.
50. The Buddha pulls this surprise elsewhere: for example, in the Brahmāyu and Sela Suttas (MN 91.6–7,
29–30; 92.13–14). I have not found an illuminating discussion.
 early buddhism 119

him, rues how the Buddha has “brought up more and more things against us”
from hearing such insults, and kicks Ambaṭṭha to the ground (14–15).
It is too late in the evening for Pokkharasāti to do what he would like, which
is to set out at once to see the Buddha, so he leaves early the next morning by
torchlight. Greeted by the Buddha, he exchanges courtesies, sits down, and asks
whether the Buddha recalls a conversation with his student. Hearing “all that
had passed,” Pokkharasāti says, “Reverend Gotama, Ambaṭṭha is a young fool.
May the reverend Gotama pardon him.”51 The Buddha replies, “Brahmin, May
Ambaṭṭha be happy.” The Buddha now also sets Pokkharasāti’s “mind at rest”
as to the thirty-two marks, and Pokkharasāti invites the Buddha and his monks
to accept a meal from him that day, which the Buddha accepts by his silence.52
Pokkharasāti has apparently anticipated this acceptance, for he is able to say, “It
is time, Reverend Gotama, the meal is ready.” And in the early morning, taking
his robe and bowl, the Buddha goes “with his order of monks to Pokkharasāti’s
residence”; Pokkharasāti serves him personally, and the young Brahmin men—
Ambaṭṭha is not mentioned, but who knows?53—serve the monks. And when
the Buddha has finished eating, Pokkharasāti takes a low seat to one side.

And as Pokkharasāti sat there, the Lord delivered a graduated

discourse54 on generosity, on morality and on heaven, showing the
danger, degradation and corruption of sense-desires, and the profit
of renunciation. And when the Lord knew that Pokkharasāti’s mind
was ready, pliable, free from the hindrances, joyful and calm, then he
preached a sermon on Dhamma in brief: on suffering, its origin, its
cessation, and the path. And just as a clean cloth from which all
stains have been removed receives the dye perfectly, so in the
Brahmin Pokkharasāti, as he sat there, there arose the pure and
spotless Dhamma-eye, and he knew: “Whatever things have an origin
must come to cessation.” (3.2.21)

Most of this is stock prose, but in the dialogues with householder Brahmins in the
Dīgha and Majjhima Nikāyas, Pokkharasāti is one of the few who is said to have

51. As Brian Black has noted (American Academy of Religion Annual Meeting, Washington DC, November
2006; 2011), young Ambat.t.ha may remind one in several ways of the “young Śvetaketu,” whom Olivelle 2005b,
13–51 has so richly portrayed as the representative Brahmanical “spoiled brat.” Roughly parallel, Śvetaketu is a
pupil who fully shares his master’s knowledge (BĀU 6.2.4), yet falls short of real knowledge, leaves the master
(his own father, Uddālaka) to seek beyond what he has taught, and gets further learning from a Kṣatriya.
52. 3.2.16–19. On this stock theme, see Bailey and Mabbett 2003, 237–57 on “meal narratives,” especially
with Brahmin landholders; see also Tsuchida 1991, 81–82.
53. Cf. Malalasekera [1937] 1983, 1: 152, agreeing that it is left uncertain whether Ambat.t.ha “became a fol-
lower of the Buddha,” and noting that Buddhaghosa squelches the idea, saying “that the Buddha knew that
Ambat.t.ha would not profit by his discourse in his present life,” and that he preached to Ambat.t.ha “at such
length” so “that it might be repeated to Pokkarasādi.”
54. Cf. Gethin 2004, 517: “by means of a step by step instruction (anupubbī kathā).”
120 dharma

experienced the Dhamma-eye while apparently remaining a lay convert.55 The

sutta then concludes that Pokkharasāti, “having seen, attained, and penetrated the
Dhamma,56 having passed beyond doubt, transcended uncertainty, having gained
perfect confidence in the teacher’s doctrine (dhamma) without relying on others,
said, ‘Excellent, Lord, excellent! It is as if someone were to set up what has been
knocked down, or to point out the way to one who had got lost, or to bring an oil
lamp into a dark place, so those with eyes could see what was there. Just so the
Blessed Lord has expounded the Dhamma in various ways.’ ”57 He announces that
he, his son, wife, ministers, and counselors will take refuge in the Buddha, the
Dhamma, and the Saṅgha, and he asks that Gotama “accept me as a lay follower
who has taken refuge from this day forth as long as life shall last!” Henceforth, he
says, the Buddha will always be welcome in Ukkat.t.ha, and whenever young men
and maidens greet him courteously, it “will be for their welfare and happiness for
a long time.” To which the Buddha replies, “Well said, Brahmin!”
Now if we go back to our observation that the AmbS is noteworthy for engaging
Brahmins on social and political matters, we may ask our accompanying question:
where is the dhamma in this text? My working premise is that we find it in three
places: in the mutual hospitality codes that both sides accept but also put to a test;
in the ways that the text moves the exponents of Buddhist and Brahmanical
dharma from initial incivilities toward a civil and fulfilling discourse; and in the
rhetoric of artful and persuasive teaching that we have just seen emphasized in
closing. These matters would contextualize the suttas historically, since they must
be engaging Brahmanical dharma with such persuasive artistry over some
particular period. In this sutta, at least, the Buddha twists the alleged Brahmanical
law to Buddhist ends, but he is debating matters that are not codified as Brahmanical
dharma before the dharmasūtras: most notably, mixed social classes, which are not
treated systematically in the earliest dharmasūtra of Āpastamba, as well as marriage
law and eligibility for Vedic teaching. But let us work back to the rhetoric by
looking at three ways this sutta and others juxtapose Buddhist and Brahmanical
dharmas: in addressing the training of young men; in responding to householders;
and in treating relations between Kṣatriyas, kings, and Brahmins.

55. The Dhamma-eye (dhamma-cakkhu), which can denote “‘entering the stream’ and thus being set irrevo-
cably on the path” (Walshe [1987] 1995, 547–48 n. 140), also arises like this for the householder Brahmins Brahmāyu
(Brahmāyu Sutta [MN 91] 36) and Kūt.adanta (in the highly stereotyped Kū.tadanta Sutta (DN 5.29; see Tsuchida 1991,
89). As Tsuchida observes, the Sela Sutta (MN 92) is “quite exceptional” in having the householder Brahmin Sela
become a monk and arahant rather than just a lay devotee or upāsaka. But rather than the arising of the Dhamma-eye,
Sela “directly knew” (MN 92.27; Ñānạ moli and Bodhi 2005, 762). But this could be translating the same thing.
56. Walshe’s translation condenses here. Cf. Gethin 2004, 518: having “seen the truth (dit.t.ha-dhammo),
gained the truth (patta-dhammo), known the truth (vidita-dhammo), penetrated the truth (pariyogāḷha-dhammo). . . .”
57. This is a stock passage; cf. the response of the Kalāmas after the Buddha has responded to their “fitting
doubts” in Aṅguttara Nikāya 3.65 (Bodhi 2005, 88–92; 432 n. 8).
 early buddhism 121

1. The training of young men is obviously this sutta’s pivot. It is when

Ambat.t.ha hears himself being called “untrained” for his physical
discourtesies that he insults the Sakyans and the Buddha verbally, not
to mention all the five hundred monks travelling in the Buddha’s
company. Indeed, we could infer that it is they who are well trained by
the evidence that they did not rise to teach young Ambat.t.ha a lesson he
deserved. As the Buddha makes clear, whereas neither Ambat.t.ha nor
Pokkharasāti is “a sage or one trained in the way of a sage,” a Buddha
is an “incomparable Trainer of men to be tamed.” In terms of Vinaya,
the Buddha is contrasting Ambat.t.ha’s rudeness with the training given
to monks and nuns, and particularly the “Training Rules” by which all
monks and nuns govern not only their own communal interactions
but their interactions with householder Brahmins like Pokkharasāti.
2. We see that the Buddha meets Pokkharasāti’s request to host him with
a meal even though the Buddha knows he is behind Ambat.t.ha’s big
insult. Elsewhere in the suttas, the Buddha couches his teaching in
terms meant for the more ordinary householder, a figure of obvious
importance to monks and nuns who beg for their food. In another,
he is asked by a family man from a market town:
Venerable sir, we are laypeople who enjoy sensual pleasures, dwelling
at home in a bed crowded with children, enjoying fine sandalwood,
wearing garlands, scents, and unguents, accepting gold and silver. Let
the Blessed One teach the Dhamma to us in a way that will lead to our
welfare and happiness both in the present life and in the future life as
well. (AN 8.54; Bodhi trans. 2005, 124)

Welfare in the present and future life are two of the three types of benefit that the
Buddha’s dhamma promotes, with the third, unmentioned by this jolly layman,
being liberation, nibbāna (Sanskrit nirvānạ ).58 The Buddha replies that four things
lead to a householder’s welfare and happiness in this life: persistent effort, pro-
tection, good friendship, and balanced living. And four “other things lead to a
family man’s welfare and happiness in the future life”: “Accomplishment in
faith, moral discipline, generosity, and wisdom.” Each gets a pithy summary.
Persistent effort goes into earning a living; protection refers to “wealth earned by
the sweat of his brow, righteous wealth righteously gained”; good friendship is
found in any “village, town, or family” with “householders or their sons, whether

58. See Bodhi 2005, 108–9. Commentators enumerate “three types of benefit” that the “teaching is
intended to promote, graded hierarchically according to their relative merit”: welfare and happiness whose nature
is visible (dit.t.ha-dhamma-hitasukha), that is, in the present life, which is “attained by fulfilling one’s moral and
social responsibilities”; “welfare and happiness pertaining to the next life, attained by engaging in meritorious
deeds”; and “Nibbāna, final release from the cycle of rebirths, attained by developing the Noble Eightfold Path.”
122 dharma

young or old, who are of mature virtue, accomplished in faith, moral discipline,
generosity, and wisdom”; and balanced living calls for a balanced budget. The
four things under “good friendship”—faith, moral discipline, generosity, and
wisdom—are the hinge between the two sets, since they also lead to welfare and
happiness in the future life. The Buddha often treats friendship as the first of
four “Unlimiteds,” also called the four brahmavihāras or “Sublime Attitudes”
(Aronson 1984), which should be cultivated in all circumstances: friendship,
compassion, sympathetic joy, and equanimity. He also says, “This is the entire
holy life, Ānanda, that is, good friendship, good companionship, good comrade-
ship” (SN 45.2; Bodhi trans. 2000, 1524). But here he seems to be linking friend-
ship more with the hospitality shown by “good people,” which is also typical of
the dharmasūtras: the generous family man “dwells at home with a mind devoid
of stinginess, . . . delighting in giving and sharing.” Ultimately, this program for
householders also envisions the third benefit of nirvānạ : the wise family man
“possesses the wisdom that sees into the arising and passing away of phenomena,
that is noble and penetrative and leads to the complete destruction of suffering.”

3. The AmbS mentions several kings in passing, but the Buddha is the
only Kṣatriya to appear in person. Here it is pertinent to recall the
emerging signs we noted in chapter 3 that dharma has begun to imply
litigation at a king’s court. Though not a king, the Buddha speaks
poignantly for the justice of the Sakyan court/sabhā where his father
would be (or have been) king (or, in the “oligarchic-republican”59 polity
ascribed to the Sakyas, seated in good humor on one of those high
seats as primus inter pares). Although royal dharma is not the focus of
the Buddha’s dhamma, he can speak to it. Here he mentions two key
symbols, the king’s cakra and the Brahmin’s daṇḍa, as if in opposition.
Cakra means “wheel” and daṇḍa means “stick,” “rod,” or “rod of
punishment.” On the one hand, the cakra, a symbol for the “wheel-
turning emperor” (cakravartin), is a reminder that the Buddha could
have chosen to be a cakravartin and just king (dharmarāja) himself.60
According to the Cakkavatti-S ı̄hanāda Sutta (DN 26) 1.2 and the
Brahmāyu Sutta (MN 91) 5, a Wheel-turning Monarch “rules without a
rod, without a weapon, by means of the Dhamma.”61 Of course,
cakravartins and just kings are the exception, and the Buddha has his
preference for smaller scale republics, as we have seen in his remark

59. For this terminology, see Collins 1998, 436; cf. 66.
60. I see nothing gained in Biardeau’s view (2002, 1: 97) that the cakravartin is a legendary and late figure
in Buddhist usage.
61. Ñānạ moli and Bodhi trans. 2005, 744; cf. Walshe [1987] 1995, 112, 396: “without stick or sword, by the law.”
 early buddhism 123

about the little female quail. But if there must be those who aspire to
empire, they should measure themselves against the highest standards
of the Buddhist monastic Community or saṅgha:

The Blessed One said: “Monks, even a wheel-turning monarch, a just

and righteous king, does not govern his realm without a co-regent.”
When he had spoken, a certain monk addressed the Blessed One
thus: “But who, venerable sir, is the co-regent of the wheel-turning
monarch, the just and righteous king?” “It is the Dhamma, the law of
righteousness, O monk,” replied the Blessed One. (Aṅguttara Nikāya
[AN] 3: 14; Bodhi trans. 2005, 115)

Just so, in a renowned passage in the Mahāparinibbāna Sutta, before he dies,

the Buddha refuses to name a successor to lead the saṅgha; rather, his own
successor will be the dhamma (DN 16.6.1). On the other hand, the daṇḍa
seems to be a symbol primarily of the Brahmin. In the AmbS, the Buddha
concentrates on the brahmadaṇḍa by which a “great sage” (note, he does not
call him a Brahmin) overwhelms King Okkāka to marry a princess with ill-
gained mantras obtained from Southern Brahmins, surely with implications
that such Brahmins would be less careful about Brahmin pedigree (neither
Kaṇha’s father nor mother is a Brahmin), and that they might not notice if
someone seeking Brahmin knowledge was dark. Brahmanical dharmaśāstra
texts will insist that kings rule with the daṇdạ (see chapter 5), and the
Brahmanical epics will assign both the cakra and daṇdạ to both avatars and kings.
Such an implicit contrast between the cakra and the daṇḍa allows us to
speak to the rhetoric of this and the other discourses we have cited. For if the
dharma as teaching lies in the Buddha’s rhetoric of artful persuasion, one factor
would be that he gives a new, challenging, and critical twist to an old Vedic
term that already carries implications of royal authority. As we have seen, the
Upaniṣads give us brief glimpses of kings who see things (notably including
karma) from the big picture. But those kings are not central speakers in a devel-
oped story who put the big picture together for others, for one and all; and
when they suggest a vision of the big picture, they do not use the term dharma,
which, to date, had been coming more and more to define Brahmin privilege,
and law or justice that privileges mainly Brahmins. This is precisely the
challenge that the dhamma poses in discourses where the Buddha’s primary
interlocutors are Brahmins. As the Buddha says after being asked his opinion
on what Brahmins “prescribe,” “Well, brahmin, has all the world authorized
the brahmins to prescribe. . . ?” (Esukāri Sutta, MN 96.4 and 11). For the first
time, dharma is presented as a tactic of civil discourse for engaging Brahmins,
among others, with the wider implications of what has ostensibly been their own
124 dharma

enigmatic term. Especially in the longer suttas, dhamma is for the first time the
overarching subject of well-rounded narrative (indeed, as we have stressed, of
well-framed and well-rounded intersecting narratives). Throughout, it is there
to be questioned, tested, penetrated, and enjoyed.

B. Abhidharma Basket Dharma

Abhidharma, the “Further” or “Higher Dharma,” is the project of those who

enjoyed taking such questioning, testing, and penetrating to new heights.
Though they fill the Abhidharma Basket with texts said to be the word of the
Buddha, they are clearly later scholastics. Abhidharma “tendencies” can, how-
ever, be traced through usages of the term mātikā (Sanskrit mātrika) for the
“matrix”-type lists or codes one finds in some suttas, and in sutta references to
mātikādharas, “those who maintain the lists,” beside “those who maintain the
Dhamma and the Vinaya.” The Dasuttara Sutta, for instance, lists 550 dhammas
to be cultivated or abandoned. The Saṅgītī Sutta gives an even larger number,
and the Mahāparinibbāna Sutta lists 1,011.62 There is no sure way to identify
such lists as late, but the “dharma theory” that typifies the Abhidharma seems to
take matters into new directions. As we saw in chapter 2, the Aśokan edicts
make it likely that Aśoka was familiar with facets of Abhidharmic dharma theory,
and that it would have been developing by his time—notably among the “Nikāya
schools” whose Abhidharma differences are sometimes traced back to the Third
Council under him (ca. 250 BCE), the Theravādins and Sarvāstivādins. The edi-
fice of an Abhidharma Basket on top of one concerned with Dharma was evi-
dently the work of several further centuries of continuing refinement, intra-school
debate, and interreligious sharpening of ideas. We will limit ourselves to this
dharma theory, which can be only partially traced from the Sūtra Basket.
Abhidharmic dharma theory concerns itself with more technical meanings
of dharma than we have seen so far in the suttas. By now we have met three
basic meanings that are typical of the suttas’ rhetorical purposes: the Buddha’s
“teaching,” “law,” or “doctrine”; the “truth” that the teaching makes accessible
through the Buddhist path; and well-trained “good behavior.” The latter
meaning is also typical of the Vinaya Basket, as is another we have met: that of
“rules,” as in monastic “training rules.”63

62. See Skorupski 1987, 333: “The Dasuttara Sutta enumerates some 550 dharmas to be cultivated or aban-
doned. The Saṅgītī Sutta gives an even larger number of them, and the Mahāparinibbāna Suttanta lists some 1,011.”
Warder 1971, 279–80 argues that the Dasuttara (Dīgha Nikāya 34) and Mahāparinibbāna Suttas are “old” texts.
63. Omitting “rules,” Gethin (2004, 515–16) finds six basic meanings in the Pāli texts: (a) the Buddha’s
“teaching,” (b) “good conduct,” (c) “the ‘truth’ realized by the practice of the Buddhist path,” (d) “any particular ‘nature’
or ‘quality’ that something possesses,” (e) “the underlying ‘natural law or order’ of things,” and ( f ) dharmas plural.
 early buddhism 125

Two other meanings are not specifically Buddhist. One is somewhat rare, at
least in Pāli, where dhamma or dharma can refer to the underlying law or structure
of the cosmos, the truth about it, which both Buddhist and Brahmanical texts can
speak of as an “eternal dharma” that remains unchanged. The other is an old
meaning of dhárman, dharma, and dhamma found in Vedic, classical Sanskrit,
and Pāli, respectively. Here, as the second noun in a type of noun-compound that
we can call “possessive compounds,” the term indicates the particular “nature” or
“quality” that the preceding noun possesses. We have met such a usage, but the
translation disguised it. The stock phrase that described Pokkharasāti’s break-
through, “Whatever things have an origin must come to cessation,” could be
translated “the nature of everything whose nature it is to arise, is to cease” (Gethin
2004, 518, author’s emphasis). That is, whatever has “the nature or quality of aris-
ing (samudaya-dhamma)” has “the nature or quality of cessation (nirodha-
dhamma)”—arising and cessation being the second and third of the four noble
truths. As Gethin observes, this meaning is prominent in Buddhist usage at the
end of bahuvrīhi (Pāli bahubbīhi) compounds. A. K. Warder provides numerous
other such bahuvrīhi usages from the Nikāyas, noting that their meaning “quality”
or “nature” may overlap with some of the other Buddhist meanings mentioned
above.64 As these and other authors have noted,65 the usage has an old Vedic pre-
cedent where dhárman can mean the “‘nature’ or ‘quality’ that something pos-
sesses.” As we saw in chapter 3, Brereton (2004, 472) glosses this meaning as
“foundational nature,” while noting that other translators favor such a rendering
more frequently than he did; and it is deployed when Yājñavalkya tells Maitreyī
that the self “has an indestructible nature” in the Bṛhadāraṇyaka Upaniṣad. As
Gethin shows, there is a route from this fourth meaning to “dharmas plural”: one,
I believe, that offers considerable clarification on this Buddhist plural usage such
as one finds in the aforementioned lists or codes (Skorupski 1987, 333). But to
appreciate the clarification means first considering the problem.

64. Warder 1971, 282–84; cf. Gethin’s paraphrase on watching dhammas as dhammas in the
Mahāsatipat.t.hāna Sutta as revealing “the underlying equivalence between seeing dhammas . . . and seeing the
dhamma or the truth” (2004, 536). One case is the famous raft simile: “Bhikkhus, when you know the Dhamma
to be similar to a raft, you should abandon even the teachings, how much more so the things contrary to the
teachings (dhammā pi pahātabbā vo pageva adhammā).” As Bodhi notes, “The word dhammā is ambiguous here,”
and has been interpreted by one commentary as attachment to “good states”— that is, “dharmas plural”—linked
to serenity and insight meditation that should be abandoned (though not the good states themselves). But Bodhi,
like other recent interpreters, takes it to indicate that attachment to the teachings “can be an obstacle to progress,”
and that adhammā “would include the moral laxity that the bhikkhu Arit.t.ha advocated” (Nāṇamoli and Bodhi
1995, 224–29, 1208–9; cf. Gombrich 1996, 22–23; Williams 2002, 38–39). As we can see, however, it is one of
those usages that is open to overlapping interpretations.
65. Warder 1971, 282–83; Gethin 2004, 532–33; Horsch 2004, 440, 448 n. 86; Halbfass 1988, 319, 334,
551 n. 44, 555 n. 1, citing it in conjunction with Pāli upādavayadhammin, “subject to origination and decay” and as
an instance where “since ancient times dharma . . . [can mean] ‘property,’ ‘characteristic attribute,’ ‘essential fea-
ture,’ or more generally . . . ‘defining factor’ or ‘predicate’ ” (334).
126 dharma

I believe this usage is untraceable to any prior Vedic or Brahmanical plural

meaning, and that it is an important Buddhist innovation in both usage and
meaning. Yet I also believe that scholars have been right to stress that all these
Buddhist meanings of dharma can overlap and be juxtaposed in given contexts,
and that “dharmas plural” gains particular intelligibility as a Buddhist twist on
the old compound meaning of dharma as the “nature,” “quality,” or “property”
that something possesses.
A quick list would include the following among a century’s attempts to
translate or explain this distinctive usage: “truths,” “laws,” “states,” “merits,”
“practices,” “phenomena,” “things,” “elements,” “conditions,” “factors,” “data,”
“qualities,” “forces,” “regularities,” “identifiables,” “noeta,” “irreducibles,”
“mind objects,” and “ideas.” One of the more basic Abhidharma classifications
divides dharmas into physical and mental categories, and some of these empha-
size one dimension or the other, while others suggest attempts to straddle
them. One of the more serviceable straddling translations is “mental events,”
since it is applicable in all Buddhist schools to the review and clarification of
“dharmas plural” in meditation.66 Yet it is awkward as a general definition since
it typifies the idealist “mind-only” doctrine of one of the Mahāyāna schools
called Yogācāra.67 To my mind, “mental events,” “forces,” and “regularities” are
the most fruitful translations of major arcs of the meaning of “dharmas plural.”
But, as we shall see, “qualities” on the one hand and “elements,” “phenomena,”
and even “things” on the other also have their uses in making, respectively,
philological and philosophical points.

B.1. Dharmas Plural in the Nikāyas

If scholars can be found to have come up with such variety in translating

“dharmas plural,” they will have attempted to contextualize their choices. Of
course everyone agrees that the Buddhist usage has to be contextualized some-
where in the teachings of early Buddhism. But from there, agreement breaks
down, resulting in three main approaches. Some of these meanings have been
proposed from the premise that “dharmas plural” is borrowed from prior
Brahmanical usages. But as I have been attempting to show in anticipating this
point, this approach is particularly lacking in evidence. There can be no linking
of the Buddhist usage to an “original pluralistic meaning of the word” dhárman

66. See Warder 1971, 280, commenting on mental/physical distinction in the Mahāsatipat.t.hāna Sutta: “It
could of course be suggested that any dhamma could be regarded as a thought-content in so far as it could be
thought of, as an idea or concept, including physical phenomena.”
67. See Cox 2004, 551 on the Sārvāstivādins’ “fundamental distinction” between material form (rūpa) and
nonmaterial mental events (nāma).
 early buddhism 127

if the word has no such original pluralistic meaning (see chapter 3 § A).
And there is no common transitional meaning, despite some suggestive
candidates,68 that will convincingly get us from Brahmanical lists of privileges,
laws, duties, or ritual details to anything that would explain how the Buddhist
usage would have such a different range of meanings.69 A second approach,
which accompanies the view that the Buddha did not himself teach about
“dharmas plural,” has been to propose that “dharmas plural” is an Abhidharmic
extension of one of the supposedly more basic and earlier Buddhist meanings
of dharma in the singular—that is, the “truth,”70 or the “teaching” or “doctrine.”71
Here there is too much evidence to the contrary. Gethin, for instance, finds it
“undeniable, whether or not one accepts this as something the Buddha himself
taught,” that a “basic understanding” of plural dhammas72 “is firmly established
and imbedded in the Nikāyas. Indeed, I think it is not unreasonable to suggest
that it is the prevalent usage of the word dhamma in the Nikāyas” (2004, 521).
We must come back to the chief contexts in which it is so embedded. A third
approach, exemplified by Halbfass (1992) in his work on the Brahmanical
Vaiśeṣikas and Cox in her work on the Buddhist Sarvāstivādins, has been to
trace the relationship between meanings both in and outside Buddhism
through “the context of early Indian scholasticism”—a “tapestry woven in often
unexpected directions as a result of both internal dynamics and external influ-
ences and events” (Cox 2004, 547). Such an approach, which comes with a fine
note of caution on filling in gaps to tell stories of “development” (554), starts
from texts expressive of the competing and intermingling outlooks of some of
the eighteen schools of “Nikāya Buddhism.” But the method would have rele-
vance to what one finds “even in the early period” (547) in the Nikāya texts as
well, which have their conversational implications and their unexpected leads.
This brings us, then, back to the Nikāyas and to the main contexts in which
they represent the Buddha as teaching about “dharmas plural” within them.
To some extent, interpretations of such dharmas have been affected by the

68. Warder proposes “regularities”; Gethin proposes “practices.”

69. See chapter 3 § E on the views of Horsch, Olivelle, and Bowles as regards mainly the hinge dharma of
“cooking the world” in Śatapatha Brāhmaṇa 11.5.7.1. Horsch proposes that the plural Buddhist usage is “best ren-
dered by ‘legal factors of existence’” (2004, 440).
70. See Horner 1948, 115–16 and Lamotte 1988, 23–26, as cited by Cox 2004, 545, 579. Cf. Gethin 2004, 514,
528–30, on Gombrich 1996, 34–36, and on Geiger and Geiger 1920, for whom, of four meanings, law, teaching,
truth, and thing, the last “far removed” and thus their project to explain: they assume that a prior meaning of
“norms” or “laws” or “truths” becomes pluralized, so that to see them means to see the Law, the Truth.
71. See Bronkhorst 1985a; 1999a, 25; 1999b, 19, for whom, as Cox (2004, 544–45) puts it, “dharma in the
sense of ‘element’ is derived from dharma as ‘teaching’ ”; “efforts were made to distill the most important ideas
and concepts from [the Buddha’s] teaching . . . [which] gave rise to lists of so-called dharmas.” Warder 1971,
278–79, reverses the assumption, suggesting that the meaning “doctrine” would derive from the plural usage.
72. He has just spoken of a “particular understanding” of dhammas as “basic qualities, both mental and
physical, that in some sense constitute experience or reality in its entirety.”
128 dharma

context selected for study. Most discussions have been focused on mindful-
ness. But even though mindfulness is of supreme importance, it tends to be
singled out at the expense of underemphasizing right effort, which has some
more basic implications. For one thing, it has affected Bronkhorst’s view that
the Buddha would not have taught “dharmas plural” himself. Bronkhorst’s
position is facilitated by his almost exclusive attention to the “dhammas ‘on the
side of enlightenment’” (bodhipakkhiyā dhammā, bodhipakṣya-dharmas).73 With
this topic, often grouped but barely named in the Nikāyas (Warder 1971, 279),
Bronkhorst can point to variations in lists that could well have been affected by
the Abhidharmic tendency to increase systematization. But while it is alto-
gether likely that there are earlier and later versions of these lists, it cannot
eliminate the question of whether the Buddha taught through lists, which
leaves the case of the bodhipakṣadharmas rather undecidable (see n. 62 above).
As to the other main contexts, however, they are little affected by the list
criterion.
In all, there are five such main contexts. For whatever expository signifi-
cance it may have, I enumerate and discuss them in the order of their occur-
rence in the Buddha’s life story in relation to the unfolding logic of his
teaching. These contexts are his teachings of: (a) dependent origination
(pat·icca-samuppāda), which is considered to be the culminating insight of his
enlightenment; (b) right effort, first mentioned as the sixth member of the
eightfold path that he declares in his first sermon; (c) right mindfulness, like-
wise first mentioned as the seventh member of the eightfold path; (d) the
“discrimination” or “discernment of dharmas” (dhamma-vicaya, dharma-
pravicaya),74 which occurs in practicing right mindfulness and also as one of
the seven “members of enlightenment” (bojjhaṅga, bodhyaṅga); and (e) the
thirty-seven bodhipakṣyadharmas. I will not address this latter grouping75 other
than to say that it includes not only the seven “members of enlightenment”
but the four foundations of (right) mindfulness and the four right endeavors
of (right) effort, which will be discussed below.
Frequently mentioned in the suttas as a list of wholesome (kusala) dham-
mas, or just dhammas (Bronkhorst 1985a, 305), the Buddha sets them out with
finality in the Mahāparinibbāna Sutta, the sutta on his last days, as “a list of
dhammas he had ascertained and taught for the benefit and happiness of the
people” (Warder 1971, 279).

73. As translated by Warder (1971, 279). Cf. Skorupski 1987, 333: “practice and principles conducive to the
attainment of enlightenment.”
74. Cox 2004, 549. For the Pāli usage, see for example DN 22.16 (Mahāsatipat.t.hāna Sutta).
75. For extensive discussion, see Bronkhorst 1985a; Gethin 1992. Cox 2004, 550 calls it “the most
enduring summary presentation of Buddhist praxis.”
 early buddhism 129

b.1.a. dependent origination (pat. icca-samuppāda, pratītya-samutpāda).

What is often called the “law of dependent origination” (also translated as
“dependent arising,” “conditioned origination,” “conditioned genesis,” “condi-
tioned coproduction”) is regarded as the Buddha’s climactic insight on the
morning of his enlightenment, and as tantamount to that enlightenment. In his
foundational study of the Buddhist dharma theory, Theodore Stcherbatsky offers
a well-considered description:

Although the separate elements (dharmas) are not connected with

one another, either by a pervading stuff in space or by duration in
time, there is, nevertheless, a connexion between them; their
manifestations in time, as well as in space, are subject to definite
laws, the laws of causation. These laws bear the general name of
pratītya-samutpāda. We have seen that the connotation of the word
dharma implies the meaning of elements operating together with
others. The concerted life of the elements (saṃskṛtatva) is but another
name for the laws of causation—the combined origination
(sam-utpāda) of some elements with regard to other elements. Thus it
is that the fundamental idea of Buddhism—the conception of a
plurality of separate elements—includes the idea of the most strict
causality controlling their operation in the world process.
(Stcherbatsky [1922] 1988, 28; author’s italics)

What Stcherbatsky describes as “the concerted life of the elements” means the
“conditionality” of dharmas, and in Theravāda Buddhism all dhammas are
conditioned except nibbāna (Sarvāstvadin Buddhism, as we shall see, identifies
three unconditioned dharmas). One caveat here: “the idea of the most strict cau-
sality” needs softening, and I would suggest—working from some ideas of
Warder—that “the hypothesis of a testable regularity”76 would be a better
starting point. Stcherbatsky’s emphasis on “some elements,” and on “laws”
rather than “law,” is helpful in this regard: one should not rush to confuse
dhammas with “the Dharma.”77

76. See Warder 1971, 289–94, especially 289: “The dhamma hypothesis” seeks “the ‘real’ laws of nature
underlying the surface appearance of things, the ‘real’ forces underlying the personifications of superstition, the
‘real’ way the universe works underlying the imaginary terrors of religion and the imaginary rule of ‘gods.’ ” Also
290: one seeks “fundamental regularities underlying the apparent chaos of data.” Cf. Cox 2004, 547.
77. We may extend here a possible reading of Warder’s “remark” (or aside) that the meaning
“‘doctrine’ . . . appears to derive from the idea of ‘the way things are’ as ascertained by the Buddha through his
experience” (1971, 278). The Buddha experiences the “regularities” on the morning of his enlightenment. Then
he preaches the dhamma. Of course, he is also, earlier, said to have learned and rejected the dhamma of two other
teachers.
130 dharma

According to Gethin, “the question of the relationship between dhamma and

dhammas is perhaps most easily seen with reference to paṭicca-samuppāda. It is
stated in the Nikāyas that he who sees paṭicca-samuppāda sees dhamma and that
he who sees dhamma sees paṭicca-samuppāda. This is in fact a very succinct state-
ment of the principle involved, for what is paṭicca-samuppāda apart from the
interrelatedness of dhammas?”78 Bodhi (2005, 283, 284) translates the statement
as follows: “He who sees dependent arising sees dhamma; he who sees dhamma
sees dependent arising.”79 The statement is attributed to the Buddha by Sāriputra,
his chief disciple on matters of meditation, and, as Gethin observes, “The text
goes on to explain that the five aggregates of attachment have arisen dependently
(paṭicca-samuppanna).”80 Moreover, the MN commentary “glosses the Buddha’s
saying as ‘he who sees causal conditions, sees dependently arisen dhammas.’ ”81
That is, the dhamma seen in seeing dependent origination is taken as the seeing
of “dhammas plural.” Gethin suggests that the commentary makes a “quite delib-
erate play” here “on the meaning of dhamma, a play, moreover, that is entirely
consonant with the Nikāyas.” For in the Nikāyas, seeing how dhammas “arise and
disappear, seeing how they are dependently arisen—one sees the ultimate truth:
he who sees dhammas sees dhamma.” In saying this, Gethin wants to assure
readers that he is not imputing an Abhidhamma meaning to the Nikāyas: he is
“not suggesting that dhamma is used in early Buddhist thought in the sense of an
irreducible element.82 The use of dhamma in the general sense of a mental or
physical quality is quite distinct from the question of the metaphysical and
ontological status of those qualities.” Such ontological and metaphysical matters
are “debated and discussed by the later schools” (2004, 536–37). Gethin con-
tinues to work out this insight around the theme of “watching dhammas as dham-
mas,” which pertains to mindfulness practice.
Now, as Stcherbatsky notes, dependent origination applies, at least first
of all, to “some” dharmas: specifically, the terminology is used to describe the

78. Gethin 1992, 151, as quoted by Cox 2004, 546, who cites also Gethin 1992, 147–54.
79. Yo pat.iccasamuppādaṃ passati so dhammaṃ passati, yo dhammaṃ passati so pat.iccasamuppadaṃ passati;
MN 28.28, 35.
80. The five aggregates (skandhas)—body, feelings, perceptions, motivational forces, and consciousness—
are the constituents of personality and the grounds for thinking there is a unifying “self.” Sāriputra, discussing
the five in relation to craving, continues: “And these five aggregates affected by clinging are dependently arisen.
The desire, indulgence, inclination, and holding based on these five aggregates affected by clinging is the origin
of suffering. The removal of desire and lust, the abandonment of desire and lust for these five aggregates affected
by clinging is the cessation of suffering.” See Ñāṇamoli and Bodhi 2005, 283, 1223 n. 342.
81. Cf. Ñāṇamoli and Bodhi 2005, 1223 n. 341, translating this gloss: “One who sees dependent origination
sees dependently arisen states (pat.icca samuppanne dhamme); one who sees dependently arisen states sees
dependent origination.”
82. Cf. Warder 1971, 278: “The four old Nikāyas are not as clear about dhamma meaning an ‘element’ as
is the Abhidhamma. They seem instead to offer discussions using the word a little more freely, apparently without
defining it, out of which the precise concept of the Abhidhamma might have been extracted.”
 early buddhism 131

(usually) twelve nidānas—“causes” or “sources”—in the so-called formula of

dependent origination. These twelve are old age and death, birth, becoming,
grasping, thirst, sensations, contact, the six sense-fields, name-and-form
(= the five aggregates), consciousness, motivational forces, and ignorance.
Warder calls attention to passages that speak of the idea that these nidānas as
dhammas have “regularity” and serve as “‘stations’ for one another. In other
words the point about dhammas is their regularity, their constant relations
despite the impermanent, transient nature of all of them” (1971, 287). The
key passage occurs in the Paccaya Sutta from the Nidānasaṃyutta83 of the SN
(II 25). In response to his own question, “And what, bhikkhus, is dependent
origination?” the Buddha explains that old age and death have birth as their
condition (and so on back through the other nidānas), whether or not
Tathāgatas arise or do not arise. Then he says, “This dhātu is established
(ṭhitā va sā dhātu), there is a station for dhammas, there is regularity of
dhammas (dhammaṭṭhitatā dhammaniyāmatā), there is specific conditionality.
This a Tathāgata attains enlightenment about.”84 Warder is, of course, mak-
ing a larger point in translating dharmas as “regularities,” and takes the
compound dhammaniyāmatā as “regularity of dhammas.” Niyāmatā could
even mean “reliability”:85 the Buddha discovers regularity amid regularities
that one can rely on. Meanwhile, Warder’s translation “stations” draws on
passages that use the term dhātu to explain how dhammas condition other
dhammas in dependent origination. Although dhātu can mean “element,” as
in the five elements,86 Warder takes it here as “base.” In either case, it would
have to provide an “element” or “base” by which the Buddha elucidates the
causal conditionality of dhammas.87

83. The “Connected Discourses on Causes,” which are all concerned with dependent origination; paccaya
means “conditions.”
84. I slightly modify Warder’s translation (1971, 281–82), and do not follow that of Bodhi (2000, 551, notes
741–42), who translates the two usages of dhamma as “Dhamma.” Bodhi sees support in AN I 286, 8–24, where
dhammat.t.hitatā dhammaniyāmatā is applied to the “three marks” (impermanence, suffering, and non-self), tak-
ing this to mean that it would not apply “specifically” to conditionality in the Saṃyutta passage. But it would apply
specifically in either case. The Saṃyutta text explains the “specific conditionality” of dependent origination.
Indeed, it goes on to apply what is said about conditions to impermanence. Warder translates the Aṅguttara
passage as follows: “this dhātu is established, there is a station for dhammas, there is regularity of dhammas, all
forces are impermanent” (1971, 285).
85. Or as the Pali-English Dictionary puts it, “reliance” (Rhys Davids and Stede [1921–25] 2003, 368).
See Idem on aniyāmena, “without order, aimlessly, at random.”
86. Interestingly, the Pali Dictionary says dhātu is “closely related to dhamma in meaning, only implying a
closer relation to physical substance” (Rhys Davids and Stede [1921–25] 2003, 348). One finds here a continuation
in Pāli of the affinity in Sanskrit between words derived from dhr. (including dharma) and from ÷dhā.
87. See Warder 1971, 281 on the classification of eighteen dhātus, “elements” or “bases,” as basic, distinct
elements yielding stimuli and experiences, including “mind base, dhamma base, consciousness through mind
base,” on which example, “This means that the ‘base’ consisting of thought-contents or mental objects [= dham-
mas] gives rise to contacts, etc., corresponding to it and eventually to searchings and gains of such objects.”
132 dharma

In a conventional Nikāya sense, Sāriputta is praised “for having well

penetrated the dhātu of dhammas [dhammadhātu] after he has understood
conditioned origination.”88 On a larger canvas, for “Buddhists of several
schools afterwards,” dharmadhātu can mean “dhamma base” in the sense of
“the whole physical and mental universe, by reference to its conditioned
nature or to the regularity of the natural laws causing it to evolve. It is thus
equivalent to the ‘base’ or source of all dhammas, of all the elements, the
universe as a quarry of events” (Warder 1971, 282). In either case, as Bodhi
says, “The ultimate purpose of the teaching on dependent origination is
to expose the conditions that sustain the round of rebirths, saṃsāra, so as
to show what must be done to gain release from that round” (2000, 517).
Let us note a provocative statement by Gombrich here: “All that the Buddha
claimed to explain was continued rebirth into this world of suffering” (1988, 8).

b.1.b. setting the wheel in motion. After his enlightenment, as the

Buddha tells it, he reflected that it would be a little hard to communicate “this
abstruse Dhamma which goes against the worldly stream, subtle, deep, and
difficult to see,” and even considered keeping it to himself until the god Brahmā
persuaded him, “There will be those who will understand.”89 As the Buddha
explains in the “Discourse on Setting in Motion the Wheel of the Dhamma,”
although he had reached complete and perfect enlightenment, he could not
claim “to have awakened to it” until he had “thoroughly purified” his knowledge
of the four noble truths to present them in an ordered way. The famous first
sermon gets to the four truths by first equating the fourth truth, that of a “path,”
with a “middle way” between sensual self-gratification and self-mortification,
and declaring it to be an eightfold path. Then he teaches how the first three
truths (suffering, its origin, and its cessation) are to be known, accomplished,
and fully understood, implying that he does the same for the fourth truth. This
means that he has done no more than mention the eightfold path as including
right view, right intention, right speech, right action, right livelihood, right
effort, right mindfulness, and right concentration. The dhamma is then set in
motion with the first conversion when one of his five listeners, Kondañña,
arises with a “dust-free, stainless vision of the Dhamma” and says, “Whatever
is subject to origination is subject to cessation” (SN 5.420–24).
For what follows, it is useful to mention that the eightfold path is conven-
tionally treated under three headings: wisdom (paññā, prajñā), concerned with
right view and right intention; morality, involving right speech, right action,

88. Warder 1971 281, citing Kaḷāra Sutta (Saṃyutta Nikāya II 56).
89. MN 26.19–20; Ñānạ moli and Bodhi [1995] 2005, 260–61; cf. Brahmasaṃyutta 1.1 in the Saṃyutta Nikāya.
 early buddhism 133

and right livelihood; and concentration, requiring right effort, right mindfulness,
and right concentration. The wisdom group calls for “seeing ‘things’ as they
are,” as in a famous passage where the Buddha addresses the Kālāmas, lay folk
and “oligarchs” like his own people, the Sakyas, who want to know how his
dhamma compares with those they have heard from other ascetic teachers and
from Brahmins:

It is fitting for you to be perplexed, O Kālāmas, it is fitting for you to be

in doubt. Doubt has arisen in you about a perplexing matter. Come,
Kālāmas. Do not go by oral tradition, by lineage of teaching, by
hearsay, by a collection of texts, by logic, by inferential reasoning,
by reasoned cogitation, by the acceptance of a view after pondering it,
by the seeming competence of a teacher, or because you think,
“The ascetic is our teacher.” But when you know for yourselves,
“These things are unwholesome; these things are blameable; these
things are censured by the wise; these things, if undertaken and
practiced, lead to harm and suffering then you should abandon them.”
What do you think Kālāmas? When greed, hatred, and delusion
arise in a person, is it for his welfare or harm? . . . (AN 3.65;
Nyaponika Thera and Bodhi trans. 1999, 65)

But it is the concentration group that introduces that terminology of “dharmas

plural” into the path.

b.1.c. right effort. The most basic way that the Buddha teaches right
effort is through “the four right endeavors,” which occur in many places in
stock-phrases:

A monk . . . endeavours so that bad, unwholesome dhammas that have

not arisen, do not arise; . . . he endeavours so that bad, unwholesome
dhammas that have arisen are abandoned; . . . he endeavours so that
wholesome dhammas that have not arisen, arise; he endeavours so that
wholesome dhammas that have arisen, are constant, not lost, increase,
grow, develop, are complete. (DN 33.1.11; Walshe trans. [1987] 1995, 487)

Right effort is thus applied to recognizing dhammas under the headings of

“wholesome” (kusala) and “unwholesome” (akusala) and fostering progress by
abetting or diverting them according to whether they are wholesome or
unwholesome and nascent or current. This practice is basic to what follows as
right mindfulness and right concentration. Right mindfulness provides the
context where seeing dhammas as “wholesome” and “unwholesome” becomes
134 dharma

soteriologically purposeful. And right concentration enables one to enter the

first meditation stage “completely secluded from sense desires and unwhole-
some dhammas” and remain in “joy and happiness” (pītisukham)90 once one has
“perceived the disappearance” of the most pervasive unwholesome dhammas
known as the five hindrances: sensual desire, ill-will, sloth-and-torpor, worry-
and-flurry, and doubt (DN 2.74–75; Walshe trans. [1987] 1995, 593). In a fashion
close to an Abhidharma analysis, one sutta (MN 111) has the Buddha speak
additionally of the wholesome dhammas that are experienced through all four
meditation stages in the “realm of form,” in which right concentration can be
carried out with increasing detachment from the “realm of desire.”91

b.1.d. right mindfulness. For right mindfulness, the key text is the
Mahāsatipat··thāna Sutta, often described as one of the most important in early
Buddhism both because it is so highly treasured and because it treats a practice
that is unparalleled in the yogic meditation practices of Jainism and Hinduism.
The sutta occurs in two similar versions (DN 22; MN 10), both said to have been
taught in the Kuru country: a curious point, since this is outside the area that
the Buddha regularly travelled, and it is not that likely he actually went there.92
In keeping with the location, and because we shall have occasions to return to
it, the topic deserves an aside.

b.i.d.i. the buddha in kuru country. I asked John Strong about this
matter, and he gave it some helpful consideration: “I don’t know if there is an
evolution or change from canonical to commentarial views, but I do get an
impression of ambivalence in the Buddhist materials: Kuru people are untamed,
in-your-face people, in need of being converted to Buddhism; on the other
hand, Kuru people are basically good, wise, and worthy” (Strong 2007 e-mail).
On this more positive side, “The people of Kuru had a reputation for deep
wisdom and good health, and this reputation is mentioned as the reason for the
Buddha having delivered some of his most profound discourses to the Kurus”
(Malalasekera [1937] 1983, 1: 641–42). The Buddha also goes there to a town,

90. See Gethin 2004, 539 n. 22: “a stock description of the attainment of the first jhāna.”
91. See Warder 1971, 280: the eleven presumably wholesome dhammas that carry through all four jhānas
of the realm of form are contact, sensation, perception, volition, thought, will, intentness, energy, mindfulness,
equanimity, and attention, with alertness added in the fourth.
92. On suttas appearing in two collections, see Norman 1983, 31, proposing that up to the second council,
“in early times a large collection of suttas were. . . remembered by heart, and the task of allocating them . . . had
not been finished or the allocation completely agreed, by the time the schools began to separate.” Cf. Manné
1990, 77–78. Although she does not take up this particular case, one could suspect that the Mahāsatipaṭṭhāna’s
Kuru country setting would have been a compromise worked out to accommodate both the DN’s public outreach
feature and the MN’s focus on sermon for monastic training (see above at n. 14).
 early buddhism 135

usually named Kammāsadhamma in the suttas, where he “is supposed to have

resided for a time” (Ibid., 642). There, he delivers not only the Mahāsatipatthāna
Sutta in both versions but, among others, the Mahānidāna (DN 15.1), Māgandiya
(MN 75.1), and Āṇañjasappāya (MN 106.1) Suttas, and two Saṃyutta Nikāya sut-
tas in the Nidānasaṃyutta: 12.60 on causation (known as the little [cūḷa] Nidāna
Sutta; Bodhi 2000, 35) and 12.66 on inner “exploration.” The commentary on
the latter says he went there because “a subtle Dhamma discourse, one stamped
with the three characteristics, had presented itself to him. In this country, it is
said, the people had good roots [supporting conditions for achievement of the
noble Dhamma] and were wise [with the wisdom of a three-rooted rebirth con-
sciousness and pragmatic wisdom]. They were capable of penetrating a deep
Dhamma talk stamped with the three characteristics. Therefore the Buddha
taught here the two Satipat.t.hāna Suttas . . . and other deep suttas.”93 A comple-
mentary image of the Kurus is also found in the Kurudhamma Jātaka (No. 276).
Once, when the Bodhisatta had become king of the Kurus after his father King
Dhanañjaya’s death,94 he observed the five precepts—not taking life, not steal-
ing, not partaking of illicit sex, not lying, and not using intoxicants—under the
name of Kurudhamma. And such was this Kurudhamma while he ruled that it
was also observed by the queen-mother, queen-consort, viceroy, keeper of the
royal granaries, palace porter, and the courtesan of the city! The country thus
prospered and its people were happy. Meanwhile, drought beset the Kalingas,
and their king asked the Kurus for their royal elephant, in hope that it would
bring rain. But when the elephant did not bring rain, the Kalingas deduced that
the Kurus’ prosperity came from the Kurudhamma, and sent messengers to find
out about it. The messengers reported that from the Kuru king on down to the
courtesan, everyone was scrupulous in keeping the Kurudhamma, but also that
each one was attentive to some unwitting violation they had committed—which
only reinforced the rigor that went into observing the Kurudhamma while
emphasizing the “future” Buddhist value of close self-scrutiny. When the
Kalinga king practiced Kurudhamma, rain fell in Kalinga (see Cowell [1895],
2005, 1: 251–60; Malalasekera [1937] 1983. 1: 643).
As to the rougher face of the Kurus, Strong suggests that one may find it
coming out in the stories of two Brahmins named Māgandīya (2007), each of
whom the Buddha converts after he shows some crude misunderstanding:
one, by insisting on offering the Buddha his beautiful daughter; the other, by
“calling the Buddha a rigid repressionist (bhunahu)” (Malalasekera [1937] 1983,

93. Bodhi 2000, 779; subcommentary in brackets; cf Malalasekera [1937] 1983, 1: 529; 642.
94. Dhanañjaya is of course a name for Arjuna in the Mahābhārata, but Pāli literature has too many
people by that name to make much of it. Malalasekera mentions nine, four of them Kuru kings ([1937] 1983, 1:
1130–31).
136 dharma

2: 594–95). Each encounter takes place in the same Kuru country town where
the Buddha delivers his aforementioned deep and subtle sermons.95 But in the
stories that give the town its name, it is not the Kurus who are “in-your-face”
but a cannibal who roves their forests. According to Malalasekera, “Even in
Buddhaghosa’s day the name of the township had two different spellings, and
two etymologies are suggested for the names” ([1937] 1983, 1: 529). As Strong
summarizes, “There seem to be two accepted forms of the name of the town:
Kammāsadamma and Kammāsadhamma. The Dīghanikāya Commentary
(2: 483) offers slightly varying etymological explanations for both, relating each
to the fact that it was the place where (a) the famous ogre ‘Spotted Feet’
(Kammāsapāda) was ‘tamed’ or alternatively (b) the place where he was
disciplined by adherence to Kuruvatta-dhamma.” As Watanabe 1909 demon-
strates, the cannibal king Kammāsapāda-Kalmāṣapāda figures both in Buddhist
stories (notably the Mahāsutasoma Jātaka; no. 537) and in the Mahābhārata
(1.113.21–22; 1.166–73). The epic makes him an Ikṣvāku king of Ayodhyā and
gives him nothing directly to do with Kuru country or the Kurus.96 But in the
Mahāsutasoma Jātaka, he takes to cannibalism as King Brahmadatta of Benares,
and does get entangled with the Kurus when he determines to eat his old friend
and tutor, the Kuru prince Sutasoma (the Bodhisatta), who gets him to renounce
his cannibalism at Kammāsadamma village and restores him to his kingdom.
In this account we have an explanation of the name Kammāsapāda from the
event that his foot is wounded on a stake while he is fleeing some pursuers,97
and the name of the town implies his “restraint” or “taming” (damma) there.98
The alternate spelling of Kammāsadhamma “is explained on the ground that
the people of the Kuru country had a code of honour called the Kuruvattadhamma;
it was there that Kammāsa (already referred to) was converted and made to
accept this code, hence the name of the township” (Malalasekera [1937] 1983,
1: 529).99 Since information on “Kuruvattadhamma” seems to be scarce, I am

95. According to Malalasekera [1937] 1983, 1: 528, the “exact place” the Buddha stayed at this town is
“mentioned only once, namely the fire hut of a brahmin of the Bhāradvāja-gotra, where a grass mat was spread
for him by the brahmin.” It was on this occasion that the second Māgand īya excoriated the Buddha’s Bhāradvāja
host. Cf. Tsuchida 1991, 85, taking this Māgandīya to be an āhitāgni entitled to perform śrauta sacrifices like the
jat.ila ascetics favored in canonical Buddhist texts.
96. His first human meal is, however, of the Ṛṣi Vasiṣṭha’s son Śakti, who had first cursed him to become
a cannibal, and it is thanks to Vasiṣṭha’s interventions that Kalmāṣapāda is relieved of the curse at the same time
that Śakti’s son Parāśara is kept alive in his mother’s womb, making it possible that Parāśara’s son Vyāsa will
compose the Mahābhārata and rescue the Kuru line by siring the Kauravas and Pāṇḍavas’ fathers.
97. As Ensink 1968, 575, 581–82 shows, some purāṇic versions, but not the Mahābhārata ones, account
for the name by a story that he dropped hot water on his feet.
98. See Watanabe 1907, 255–59; Cowell [1895] 2005, 3: 246–79; Malalasekera [1937] 1983, 1: 529–30; 2:
573–74.
99. Matters are in fact still more complicated, as there are said to be two Kammāsadammas (the original
“big” and the other “small”), and two cannibals named Kammāsa, the second converted in the littler
Kammāsadamma in the Jayaddisa Jātaka, no. 513 (Malalasekera [1937] 1983, 529, 943–44).
 early buddhism 137

left to guess that this commenterial term would be another name for, or
refinement on, the “Kurudhamma” mentioned above in the Kurudhamma
Jātaka. If so, what seems to be implied is that the Kurus honored the five pre-
cepts, before the Buddha even taught them, when hosting a cannibal.100
In any case, the place “is described as a nigama [town] of the Kurus, where
the Buddha resided from time to time” (Malalasekera [1937] 1983, 1: 641–42).
Whether or not it was a real place, and irrespective of whether the Buddha actu-
ally went there, it is fascinating that the early Buddhism101 would present the
Buddha going so far out of his way. Whether it was to tame some unrefined
Brahmins of the Kuru country, to tame a cannibal who haunted that country,
or to find the right audience there for his subtle teachings, we find his chief text
on mindfulness in a meaningful interetextual context that is of interregional
interest. It is as if he leaves “greater Magadha” to bring subtle Buddhist dhamma
theory to the rough and tumble attractions of the Vedic heartland.102

b.i.d.ii. establishing mindfulness. The term in the Sutta’s title, sati-

paṭṭhāna (Sanskrit smṛti-upasthāna), is usually taken as “foundation of
mindfulness” (it can also mean “way of establishing mindfulness” or “domain
of mindfulness”). The Buddha says there are four such foundations: “A monk
dwells watching the body as a body, watching feelings as feelings, watching
mind (cittā) as mind, and watching dhammas as dhammas. In each case, he is
to be “ardent, fully aware, and mindful.” Thus if one is distracted from sustained
attention on the body as body to watching feelings as feelings, he should be
aware that he has moved on to feelings (Nyanaponika Thera [1962] 1979,
132 n. 2). In this sense the four “foundations” are thus “domains.”103
The sutta describes twenty-one exercises, fourteen of which focus mindful-
ness on the body, one each on feelings and thoughts, and five on dhammas.

100. The only (seemingly pertinent) information I find by an internet search is a brief mention in
Loomcharoen 2007, 16: “Individual moral development through bhāvanā, for example in Kuruvatta-Dhamma,
begins with the observation of the five precepts as the basis for the practice of satipaṭṭhāna, to finally make that
observation and practice common, routine, a way of life.” At “Kuruvatta-Dhamma” there is a note citing Jātaka
3.368 with commentaries, but I have not been able to follow that up.
101. And not only the Theravāda; Strong 2007 mentions a Divyāvadāna account of the story of the first
Mākandiya where the town is named Kalmāṣyadamya. Cf. Watanabe 1909, 243–70 on the story’s appearance in
Many Sanskrit Buddhist texts; 284 n. 4 on the town’s name as reconstructed from Chinese translations of it.
102. On Kuru country as the central region of Brahmanical dharma, see Macdonell and Keith [1912] 1967,
1: 165–69. On “greater Magadha,” see Bronkhorst 2007. “Some scholars” locate Kammāsadhamma “in the
vicinity of modern Delhi” (Ñāṇamoli and Bodhi 1995, 1188). If Kuruvatta implies Sanskrit Kuruvarta, it could,
like Brahmāvarta or Āryāvarta, suggest a sacred land of dharma (cf. Manu 2, 17–22).
103. I have been following mainly Gethin’s terminology (2004, 520), notably with “watching” for anupas-
sati, along with the translation and notes in Ñāṇamoli and Bodhi [1995] 2005, 145 ff., 1188 ff. Henceforth I mainly
use the latter translation. As Gethin 2004, 529 indicates, anupassati can be also translated as “contemplating” (as
is done by Ñāṇamoli and Bodhi, Nyanaponika Thera [1962] 1979, and others). Beyer 1974, 90 ff. favors
“observing.” The verb has a visual implication and is used in an active sense (thanks to Greg Bailey on the latter
point; personal communication, December 2006).
138 dharma

Each exercise is followed by a discussion of the insight to be developed as one

deepens understanding. Yet there is also a continuing refinement in focus,
since, although dhammas come fourth, they are mentioned in all the intervals
on insight. Keeping the now familiar opening phrasing (it is in these insight
intervals as well as at the beginning), I call attention to the one instance of
dhamma(s) that recurs in all the insight sections by italicizing its translation
“phenomena.” As the monk has dwelt “watching the body per se,” “feelings per
se,” and “thoughts per se,” so now:

Thus he dwells watching dhammas per se [in three ways]: first,

he dwells watching dhammas within himself, or dhammas outside
of himself, or dhammas both inside and outside himself;104 second,
with regard to dhammas105 he dwells watching the phenomena of
origination (samudayadhammā), or passing away, or both origination
and passing away; third, he sustains the awareness “this is a
dhamma”106 in so far as wisdom and recollection allow, and remains
detached, not clinging to anything in the world.107

In the second way of watching, most translators take dhamma in samudaya-

dhammā as referring to “dhammas plural.” But it is possible, given the stock
usage we have noted of the same compound, to take it as a singular and trans-
late it as “having the nature (dhamma) of arising and passing away.”108 This is
plausible. But more likely, it is a case of contextual juxtaposition and overlap-
ping meanings, such as one finds particularly with regard to the two meanings
“nature of” and “dhammas plural” in contexts dealing with or implying (as is
the case here) dependent origination (see Warder 1971, 284, 286, 288). The
usage probably occurs through all the intervals on insight in anticipation of
dhammas being the fourth foundation of mindfulness. For the momentum of
the sutta is to break things down from the most solid, the body, to the most
insubstantial and evanescent, dhammas. Thus commentaries turn to this sutta
to illustrate that dhammas are without essence, lifeless, and empty. In any case,

104. As he did when focused “on his own body, or on the body of another”/ “on his own feelings, or on the
feelings of another”/ “on his own thoughts, or on the thoughts of another.”
105. Likewise, as with the body, etc.
106. Likewise, as he did, “this is the body, etc.”
107. Beyer 1974, 90–99 has for the third way the felicitous, “And he establishes the mindfulness that
‘This is an event’ [‘a body’ etc.] just sufficiently for a bare awareness & bare mindfulness of it; and he dwells in
freedom and does not cling to anything in the world.” But Beyer’s translation reads as if the three ways were
simply cumulative rather than also distinct.
108. As does Bodhi (Ñānạ moli and Bodhi [1995] 2005, 1191 n. 144), who, for the second and third editions
of this translation, notes that “[a] plural sense . . . is not mandatory,” and offers the view that “it is more consistent
with the use of the suffix -dhamma elsewhere to take it to mean ‘subject’ or ‘having the nature of.’ ” In Bodhi’s new
translation, what the monk watches in the body, etc., is thus “its nature of arising” and “its nature of vanishing.”
 early buddhism 139

we find that through all these iterations, this second way of watching, no matter
what one is watching (from breathing to corpses to feelings, thoughts, or dham-
mas), refers to a way of watching dhammas arising and passing away, or in
more common terms, watching “phenomena” or “events” in their “rise and
fall.” In the five exercises that come under the fourth foundation of watching
dhammas as dhammas, all this is practiced with regard to the five hindrances,
five aggregates, six sense-fields (which include mind, and being aware of the
mind as the sense base for dhammas as “thoughts” or “ideas”), seven limbs of
enlightenment (which include the “discrimination of dhammas”), and four
noble truths (which include a reiteration of the four endeavors of right effort).
This closing section on the four noble truths is climactic. As Nyanaponika
Thera says, it unfolds the four truths “in terms of the actual Satipaṭṭhāna prac-
tice,” particularly as regards the reassurance offered that the “act of mindful
noticing will necessarily stop the continued flow of craving; because detached
observation and craving cannot go together” ([1962] 1979, 134 nn. 37 38).
In raising the question of what it means that “a monk dwells watching
dhammas as dhammas,” Gethin remarks that at least for the first four topics,
“[c] learly . . . dhammas are not teachings, practices, truths, or laws,” and that
even with the inclusion of the four noble truths, it is not a matter of “doctrinal
propositions, but realities that have to be understood” (2004, 520–21). That is
well said. But eventually Gethin comes back to “practices” as one of “two basic
meanings of dhamma in early Buddhist texts: the practices recommended by
the Buddha and the basic qualities that constitute reality. The first takes dhamma
as something normative and prescriptive, and the second as something descrip-
tive and factual. Both of these meanings essentially derive from pre-Buddhist
usage but both are adapted to the specifics of Buddhist thought” (534). Gethin
has built this up from his understanding that such meanings are “developed”
from Vedic and Brahmanical ones (531–32). I think he is right in one case but
not in the other, but that, in any case, he offers a misleading dichotomy. The
derivation of “basic qualities” from the bahuvrīhi usage of dharma at the end of
compounds is, as we have seen, unexceptionable and important. But the
Buddha and early Buddhists are just continuing to deploy a feature of Indo-
Aryan linguistic usage rather than deriving a meaning from the past.109 On the
other hand, I do not think one can trace a sense of “practices” back to the ṚV
on the grounds that, “among its earliest uses in the plural,” dhármans “refer to
certain practices—primarily sacrificial rites—as maintaining and supporting

109. Gethin writes, “In fact this usage of dharma in the sense of ‘property’ or ‘characteristic’ would seem to
derive directly from the Vedic usage of dharman to refer to ‘foundational nature of a deity’” (2004, 532). Yet before
this, he says, it is “a common usage in both Pali and Sanskrit and is not a specifically Buddhist usage” (518).
140 dharma

things—the cosmic and social order,” and that “dharmans are therefore
prescribed practices.” Yet Gethin draws from this that Buddhism “inherited” a
“plural usage of dhamma . . . from earlier pre-Buddhist usage, and that for early
Buddhist thought dhammas are in the first place practices, the kinds of behavior
prescribed and recommended on the authority of the Buddha”—indeed, that it
is only “subsequently” that dhamma “comes to refer to the Buddha’s teaching”
(531). This is a case of filling in the gaps with a very imprecise notion—
“practices”—and repeating the error of thinking that there is some kind of
inherent meaning or intrinsic nature to the fact that Ṛgvedic dhárman some-
times occurs in the plural.

b.1.e. discrimination of dharmas. As indicated, mindfulness makes

right effort increasingly purposeful in the soteriology of the eightfold path.
In the skein of the seven members of enlightenment, “discrimination of
dhammas” comes second between (1) mindfulness and (3) effort or energy
(viriya, vīrya), from which one can then develop joy (piti), relaxation
(passaddhi), concentration (samādhi), and equanimity (upekkhā)—four
wholesome states (4–7), often called dhammas, that mark access to the first
and second jhānas.110 One notes that with “discrimination of dhammas,” the
order between mindfulness and effort is now reversed from that in the
eightfold path. Clearly, “effort” now means continued or renewed effort. In
the full sequence of these seven bojjhaṅgas, “discrimination of dhammas”
thus arouses the “tireless energy” (Ñāṇamoli and Bodhi [1995] 2005, 947)
by which to pursue meditation.
According to Cox, “[t]he formulaic description of the discrimination of
dhammas in the Pāli suttas suggests that it functions specifically through insight
(paññā), by which one discriminates, investigates, and reflects thoroughly upon
dhammas” (2004, 550). We find such insight or wisdom (paññā = Sanskrit
prajñā) in a passage where the Buddha unfolds the full bojjhaṅga sequence with
respect to mindfulness in breathing: “Abiding thus mindful, he investigates
and examines that state with wisdom and embarks upon a full inquiry into it.
On whatever occasion, abiding thus mindful, a bhikkhu investigates and exam-
ines that state with wisdom and embarks upon a full inquiry into it” (Ānāpānasati
Sutta (MN 118) 31; Ñāṇamoli and Bodhi trans. [1995] 2005, 946–47). As
Cox shows, however, it is the Sarvāstivādin school that articulates this practice
most assiduously and in effect makes it a linchpin of its more developed
dharma theory.

110. For example, MN 118.29–38. See Rahula 1974, 74–75; Bodhi 2000, 1910. See also n. 91 above
regarding additional dhammas that apply here.
 early buddhism 141

B.2. Dharmas and Abhidharmas

Here we must shift our focus from the Nikāyas to “Nikāya Buddhism.” For with
the Sarvāstivādin Abhidharma and that school’s commentarial discussions, we
meet not only the classical dharma theory among the schools of “Nikāya
Buddhism” but the handling of this topic that was most directly critiqued by the
Mahāyāna (see Williams 2000, 92–95). Sarvāstivādin Abhidharma also makes
the connection between “wisdom” and “discrimination of dharmas,” even offering
the meditator a new “controlling faculty of insight” (prajñendriya).111 As Cox says,
“Abhidharma itself is defined as insight (prajñā), which is then identified with
dharmapravicaya that discriminates dharmas according to their intrinsic nature
[svabhāva], or according to their particular inherent (svalakṣaṇa) and generic char-
acteristics (sāmānyalakṣaṇa).” The activity of dharmapravicaya is “compared to
the skill of a jeweler who can recognize and distinguish any stone,” or to selecting
the right flowers. And as insight, “it becomes the sword that cuts off defilements
preventing them from ever rising again.” The “early Sarvāstivādin texts” thus
clarify how “the discrimination of dharmas acts by separating and clarifying
dharmas of all categories, in particular by distinguishing events that are unvirtu-
ous and result in suffering from those that are virtuous and lead to liberation; the
former are to be abandoned, and the latter are to be cultivated.”112 In this regard,
as Cox shows, the Sarvāstivādins extend “discrimination of dharmas” far beyond
the practice of the seven members of enlightenment to the point of making it the
grand trunk road to right views “as the first stage in a standard series of mental
cultivations that culminate in the faculty of discerning (vipaśyanā)” (550–51).
This, of course, redefines the scope of what is called “insight meditation” (Pāli
vipassanā), by which, however, in either school, one can discover “the way things
really are” (yathābhūtam) (see Gethin 2004, 549; cf. Williams 2000, 81). As
Williams (88, 91–92) demonstrates, “seeing the way things really are” is the
general goal of the Abhidharma and is primarily a goal of insight meditation.
As we saw in chapter 2, there is some suggestion in the Aśokan edicts that
Aśoka was familiar with facets of Abhidharmic dharma theory. Among the

111. Cox 2004, 550, 580 n. 29. This would be beyond the basic six indriyas (the five sense organs plus mind).
Also highly interesting in another list is the inclusion of femininity, masculinity, and the life force among
twenty-two indriyas in a position between the five sense organs and the mind—there, because these “controlling
faculties” “further qualify the final corporal sense of the body” and “determine the affective quality of mental
events” (552, 581 n. 43). Since they are between body and mind but apparently both, they however do not qualify as
“dissociated forces”—dharmas that are neither physical nor mental (on which see § B.2.a). It would be interesting
to know what lies behind this Sarvāstivādin gender theory.
112. Cox 2004, 550. A Mahāyana Perfection of Wisdom view rejects discrimination of dharmas:
“A Bodhisattva should therefore be trained in non-attachment to all dharmas, and in their unreality—in the sense
that he does not construct or discriminate them” (Mahāprajāpāramitā Sutra part 1, Conze trans. 1961, 98 (1, 4, 13),
as cited by Halbfass 1992, 64 n. 27.
142 dharma

schools whose Abhidharma differences are sometimes traced back to the Third
Council under Aśoka are the proto-Theravāda Vibhajyavādins and the
Sarvāstivādins (see Lamotte 1988, 273–74, 277). The sense in which these and
other schools speak of abhidharma on top of dharma certainly marks a refine-
ment within Buddhism that can be only partially traced from the Nikāyas.
Where the latter use the term abhidhamma, “they do not mean to designate any
scriptural code, but simply the ‘Special Dharma’, i.e. the Doctrine pure and
simple, without the intervention of literary developments or the presentation of
individuals. When understood in this sense, abhidhamma is often coupled with
the word abhivinaya” (Lamotte 1988, 180).113 Abhidharmic “tendencies,” on the
other hand, can be better traced through Nikāya usages of the term mātikā
(Sanskrit mātrika) for the “matrix”-type lists of dhammas we have noted in some
suttas. In this fashion, MN 33.9 seems to mention three types of monks: those
“who maintain the Dhamma, the Discipline, and the Codes” (mātikās)
(Ñāṇamoli and Bodhi [1995] 2005, 314). We probably see here an anticipation
of what will be called the “three baskets.”
For our purposes, which are to offer a sense of this new refinement, it
must suffice to look briefly at what the Theravādins and Sarvāstivādins do not
have exactly in common,114 but at what underlies their differences, as a way of
getting at what Abhidharma refinements amount to in historical and everyday
language. I will consider a few topics for which the Sarvāstivādin Abhidharma
is famous from this vantage point.

b.2.a. sā rvā stivā din dharma-ontology. First, let us look at some topics
that have already made their way into our discussion, beginning with the
relation between Abhidharma and insight meditation. According to Williams,
in “seeing dharmas as events (dharmas)” (i.e., “watching dhammas as dhammas”),
all schools were concerned with seeing them “as based perhaps on an event-
ontology, rather than on a substance-ontology. . . . To that extent, one could
argue, the everyday practicalities of insight meditation remain paramount.”
But while the Pāli Abhidhamma left “specific questions of the ontological
nature of dharmas” “relatively unexplored,” such an “interest” was found
“among Sarvāstivādins and their rivals” (2000, 92). This ontological interest
promoted an intensified preoccupation with taxonomy via “analytical matrices
(mātrika) or lists of categories which themselves constitute the dharmas”

113. See, however, Ñāṇamoli and Bodhi [1995] 2005, 1226 n. 362, agreeing with Watanabe 1983, 34–36,
who, in Bodhi’s words, “concludes that the Buddha’s own disciples formed the conception of Abhidhamma as an
elementary philosophical study that attempted to define, analyse, and classify dhammas and to explore their
mutual relations.”
114. According to Williams, the two are “in many respects very similar” (2000, 92).
 early buddhism 143

(Cox 2004, 551). According to Cox, this working up of new matrices of matrices
offered “a taxonomic abstract web of all possible conditions and characteristics
exhibited by actually occurring dharmas” (552). In this vein, the Sarvāstivādins
developed a system of six causes and four conditions, which I shall return
to shortly.
There are, then, lists within lists, but the most representative Sarvāstivādin
list comprises seventy-five dharmas in a “fivefold taxonomy” (pañcavastuka) of
five genera. Therein, one finds eleven dharmas of material form (rūpa), one
dharma of thought (citta), forty-six thought-concomitant dharmas (caitta: whole-
some ones like faith, unwholesome ones like anger, and other related dharmas
that accompany thought), and fourteen “dissociated forces” (dharmas that are
neither physical nor mental).115 These are all conditioned dharmas, the total of
which is seventy-two. Then there are the three unconditioned dharmas. This
taxonomy “attempts to present a complete and systematic listing of all possible
dharmas classified abstractly by distinctive intrinsic nature (svabhāva).” And
quite intriguingly, according to Cox, in doing so it “takes the perspective of the
dharmas themselves” rather than that of the meditator and his or her praxis
(Cox 2004, 553). The seventy-five dharmas that make the cut are considered to
be “primary existents” (dravyasat) or “irreducible simples” by the fact that they
“have svabhāva” (sasvabhāva)—that is, each has its “own” (sva) intrinsic “nature”
(bhāva). They are thereby distinguished from secondary “conceptual existents”
(prajñaptisat) like a table or person, which are real unities for pragmatic pur-
poses only, can be broken down, and are considered to be “without svabhāva
(niḥsvabhāva)” or “empty” (śūnya) of intrinsic nature (Williams 2000, 93–95).

b.2.b. svabhāva. Although Sarvāstivādins are most readily distinguished

for a view of time in which dharmas have real existence not only in the
present but in the past and future, they seem to get most idiosyncratic about
dharmas in three concepts: “intrinsic nature” (svabhāva); one of the fourteen
“dissociated forces” called “possession” (prāpti); and space as one of three
“unconditioned dharmas.” Svabhāva is the hinge on which such differences
turn. According to Cox, Sarvāstivādin commentarial tradition defines
dharmas both etymologically and via svābhāva (or an equivalent term): thus
“dharma means ‘upholding,’ [namely], upholding intrinsic nature (svabhāva)”;
similarly, “Etymologically, ‘dharma’ comes from upholding a particular
inherent characteristic” (nirvacanaṃ tu svalakṣaṇadhāranād dharmaḥ), using

115. See Cox 2004, 554: This third “category of dissociated forces (cittaviprayuktasaṃskāra),” a completely
new one for the Sarvāstivādins, “includes dharmas that were proposed to account for a varied range of experien-
tial or doctrinally necessary events and is, therefore, a miscellany of dharmas not unified by any overall integra-
tive principle other than dissociation from material form and thought.” See Cox 1995.
144 dharma

“svalakṣaṇa in place of svabhāva” (2004, 559, 584 n. 70). With reference to

“dharmas plural,” it is interesting that such definitions derive dharma from
formations of ÷dhṛ, which Cox translates as “upholding.” But let us just note
that nothing would seem to be lost, and something might possibly be gained,
if we broke with this translators’ convention and replaced “upholding” with
“holding,” which is really what ÷dhṛ basically means.116 What Sarvāstivādin
dharmas do “from the perspective of the dharmas themselves” is not really
“uphold their intrinsic natures,” which sounds as if the fivefold taxonomy
was something like a constitution. Being utterly impersonal, such dharmas
simply (although explaining it is not so simple) “hold their intrinsic natures.”
Amid the flux of “conceptual existents,” the meditator can rely on them as
irreducible “holds” precisely because each such dharma holds its particular
svabhāva. As one watches faith or anger rise and fall, one can put a wholesome
hold on the one and let the other go. In everyday English, it would be
something, in the first case, like saying “hold that thought” and reducing it
to “let that mental event take hold”—and in the second, something like
saying “let it let itself go.”
Sanskrit etymologies from ÷dhṛ are interesting in light of what I have sug-
gested is the better part of Gethin’s discussion of “two basic meanings of
dhamma in early Buddhist texts”: his discussion of “qualities”—the meaning
that can be traced to usages of dharma at the end of bahuvrīhi compounds. This
usage also links dharmas with an etymology of dharma by another route, but
since it is the same root, it is one that amounts to the same thing. When
speaking about Nikāya usages at the end of bahuvrīhi compounds, Gethin, as
we have seen, says it “has to mean something like ‘nature’ or ‘characteristic
quality’” (518). The Pāli Abhidhamma commentaries define this meaning more
precisely: “dhamma as the last member of a compound means the natural
condition (pakati) of something, thus to describe someone as jāti-dhamma or
jarā-dhamma means that birth and old age are his ‘natural condition’ (pak-
ati).”117 Whereas Gethin had initially glossed what we are calling “dhammas
plural” as the “a basic mental and physical ‘state’ or ‘thing’” (2004, 516), he
now modifies this to “mental or physical quality.” This would seem to suggest
that the Abhidharmic usage is closer to the possessive compound meaning
“‘nature’ or ‘characteristic quality.’”

116. Note how Fitzgerald renders a similarly etymological Mbh passage: “They say dharma is from ‘holding’
(dhāraṇād dharma ityāhur). Creatures are ‘held apart’ by dharma (dharmeṇa vidhṛtāḥ). So dharma is whatever
involves ‘holding’ (yat syād dhāraṇasaṃyuktaṃ sa dharma); that is the settled conclusion” (2004a, 445,
modified).
117. Gethin 2004, 522; pakati is Sanskrit prakṛti, “nature,” and Gethin adds that “an alternative term used
by the commentaries here is vikāra in the sense of disposition” (2004, 522). Jāti and jarā mean “birth” and “old
age”; they are also causal conditions among the twelve nidānas of dependent origination.
 early buddhism 145

Theravāda Abhidhamma commentaries take this possessive compound

meaning differently as “the natural condition (pakati) of something,” but also
use the Pāli term sabhāva, equivalent to Sanskrit svabhāva, to explain its
meaning. Again, one could say that the bahuvrīhi usage and the sabhāva
meaning are not different in meaning. As Warder says, not uncommonly these
two meanings and usages overlap.118 One could even say that the bahuvrīhi
usage points to the sabhāva meaning. But Pāli sabhāva should not be translated
as “intrinsic” or “inherent nature.”119 Rather, it is meant to explain that dham-
mas are called “particular natures” or “qualities” because they “hold” or “main-
tain” themselves in relation to causal conditions. But whether or not dharmas
are “ontological irreducibles” or just “conditional correlatives,” they have the
similar quality and function of being meditational “holds.”

b.2.c. prā pti. Indeed, one of the Sarvāstivādins’ seventy-five dharmas that
functions most clearly as a kind of “hold” is prāpti, “possession” or “ownership”—
one of the fourteen “dissociated forces” (dharmas) that are neither physical nor
mental. Although it has been compared with other “pseudo-selves” allowed by
different schools to explain varied facets of the personal continuum without a
personal substratum or “soul,”120 as Williams says, it is a “unique Sārvāstivadin
doctrine, and once more a topic of intense debate with others.” According to
Williams, “This prāpti too is an impermanent dharma” (2000, 116–17). A series
of prāptis explains how a bad intention keeps having a new “possession” to
make the bad intention “one’s own.” In effect, everyone lives almost inescapably
in an “ownership society”—and is bound to stay in it by their own karma. For
unenlightened persons, prāptis are “own-its” in series. The advice, “Don’t own
it,” is thus good Buddhist advice, and also ontologically on target, since no
matter what your school, it is never “really” yours to begin with or to end with.
But how do Sarvāstivādins envision getting to that end without ownership?
Fortunately, for an enlightened person, “there is also a different dharma

118. Warder 1971, 282–84, 286: “usages overlap”; “no sharp distinction” between “elements” and
“qualities”/“natures.”
119. Gethin (2004, 533–34) points out that Carter (1978, 61) “has drawn attention to the way in which the
Pali commentaries later come to gloss dhamma at the end of a bahuvrīhi compound both by pakati and svabhāva.”
For Gethin, it follows “that when the commentaries define dhammas as sabhāvas this is not a statement about
their ontological status and that sabhāva should not be translated as ‘inherent existence’, but is merely a gloss
stating that dhammas are ‘particular natures’ or ‘particular qualities.’ Moreover, . . . [to] say that dhammas are
so-called ‘because they maintain (dhārenti) their own particular natures, or because they are maintained
(dhāriyanti) by causal conditions,’ this should be understood, I think, as a direct and deliberate counter to the idea
of dharmas as ‘particular natures’ that are maintained by an underlying substance (dharmin) distinct from them-
selves; it is not intended to define dhammas as ontologically irreducible entities” (2004, 534). This differentiates
a Theravāda position on sabhāva from both Brahmanical (Vaiśeṣika and Nyāya) conceptions of substance and
substratum and the Sarvāstivādin conception of svabhāvas as irreducible reals. See Halbfass 1992, 79, 150–51.
120. See Conze 1962, 132, and on prāpti and aprāpti, 139–41, 161–62.
146 dharma

present, called ‘non-possession’ (aprāpti), which keeps the negative taints from
ever occurring again.” It accounts for how the ownership series can lose its
hold. No other schools held these two dharmas, and the Sautrāntikas especially,
with their intensified emphasis on the “momentariness” of dharmas, rejected
them as “unnecessary” and “absurd.”121 The Theravāda, probably outside this
argument, held that there was an “inactive level of mind” called bhavaṅga “that
is still present when no mental activity is occurring,” and “that makes the link
between a dying person and rebirth” (Williams 2000, 123).

b.2.d. object-support conditions. Finally, if prāpti is meant to explain how

certain things (dharmas) hold course with respect to the right effort of a meditator,
there remain much larger questions as to how all things hold course with respect
to each other. Here we come to space as one of three unconditioned dharmas. In
Sarvāstivādin terms, these are dharmas that are permanent rather than
impermanent, “free from arising and passing away or modification,” and lacking
any “generative cause or dependence upon a collocation of causes and conditions,
as well as any activity that generates its own effect” (Cox 2004, 555–56). Whereas
the Theravāda Abhidhamma allows for only one unconditioned dhamma, nibbāna,
the Sarvāstivādins introduce space (ākāśa) as the third in the company of two types
of nirodha or “cessation” (instead of one nibbāna), which they call “cessation
resulting from consideration” (pratisaṃkhyānirodha) and “cessation not resulting
from consideration” (apratisaṃkhyānirodha).122 Cox raises the pertinent question:
if they are neither conditioned nor condition other dharmas, “what then do
unconditioned dharmas ‘do?’” Here, like some other recent scholars, Cox
emphasizes the “function” of dharmas, how dharmas “function” both specifically
and in general, and how the universe “functions,” given dharmas.123 Cox insists
that unconditioned dharmas “do indeed have a function, and, as in the case of
other impermanent, conditioned dharmas, this function is the basis that determines
their distinctive character and hence both their status as dharmas and their
existence” (Cox 2004, 556). The distinctive function of space, linked to its existence
and knowability, is “not obstructing.” Space can be known by reductio-type
inferential arguments called prasaṅga.124 Among these, the *Mahāvibhāsạ ,̄ a text
from about the second century CE, credits its chief Sarvāstivādin proponent with
saying, “since one observes that there are places without obstruction, one knows

121. Williams 2000, 118. With twenty-four or ten additional dharmas in the “dissociated forces” category,
the Yogācāra kept prāpti but discarded (or perhaps better, abandoned) aprāpti (Takakusu [1947] 1956, 96a).
122. Cox 2004, 556. For an explanation of these two types of nirodha, only the first of which is nirvāna as “due
to the comprehension, by wisdom, of the four holy truths,” see Conze 1962, 162. The Yogācāra Abhidharma adds two
more types of nirodha plus tathatā, “suchness,” to make six unconditioned dharmas (Takakusu [1947] 1956, 96a).
123. See Cox 2004, 555–58, 571–74, 576–78; Gethin 2004, 535: “a functional concept.”; Skorupski 1987, 334.
124. Since the Buddha also speaks of space, it also has scriptural authority.
 early buddhism 147

that space definitely exists as a real entity (dravyatas), because it is precisely space
that has the distinctive function of non-obstruction [anāvṛti] as its characteristic.”125
This argument has an interesting extension when Saṅghabhadra, perhaps two
centuries later, “suggests that space as an unconditioned dharma functions to
provide a place for light; the cause (space) which exists as a real entity in intrinsic
form, is manifest by means of this effect (the appearance of light). . . . If there were
no space, there would be no light. Since there is light, distinctions in material
form are grasped by visual perceptual consciousness. Therefore, space exists
because it is able to provide a place for light, and so forth. It is proven that space
exists as a real entity” (Cox 2004, 586 n. 64). One is reminded of the Vedic
“ontology of openness” that opens spaces (varivas, etc.) for light, worlds, and all
other phenomena by removing “The Obstacle” personified as the demon Vṛtra.
As Halbfass says, although an “ontology of substance” may be more conspicuous
in later Brahmanical thought, the “ontology of openness should by no means be
underestimated” (1992, 32). Indeed, it could be said that with the example of light,
Saṅghabhadra’s additional argument recalls an early Vedic outlook. But that is not
the stated focus. Rather one may note how down-to-earth and practical
Saṅghabhadra’s concern with space and light is. It is really the mental space to
deal with immediate stimuli. The “and so forth” is perhaps the only hint that this
comment could have reference to light-bringing celestial bodies.126 The argument
also positions this Buddhist view over and against most forms of Brahmanical
scholastic thought, which view space or “ether” in a continuum of unfolding
substance that conveys sound rather than light.
And how do conditioned dharmas function in relation to unconditioned
dharmas (including space) in such an open space that includes mental space?
Here we come to the Sarvāstivādins’ system of causes and conditions. Although
Theravādins listed as many as twenty-four conditions (paccaya) and as few as
four, many of which are found by name in the Sarvāstivadin lists of causes and
conditions, the Sarvāstivādins refined matters to six causes (hetu) and four con-
ditions (pratyaya; Pāli paccaya)127—even though, as Lamotte mentions, they
acknowledged “the synonymity of the words” hetu and pratyaya. Lamotte offers
a pithy summary: “The general principle is that all dharmas are causes with

125. Cox 2004, 557, 582–83 n. 59; Williams 2000, 113–14, 289. The asterisk indicates that the work is
known only in Chinese translation. The same applies to the Nyāyānusāraśāstra of Saṅghabhadra, cited next.
126. Cox returns to Saṅghabhadra, for whom “the term dharma does not denote a permanent substance,
but rather a cognitive category, an objective locus identifiable through cognition. The world exists only as cog-
nized, and the regularities of this cognized world that are evident in the given objects (dravya) of our experience
are expressed through its constitutive dharmas” (2004, 578). Once again we come back to Warder’s under-
standing of dharmas as “regularities” in what Halbfass (1988, 317), endorsed by Gethin (2004, 535), calls “a
certain elusive coherence.”
127. Cox 2004, 557 n. 66; cf. Lamotte 1988, 605; Williams 2000, 115–16, 258–59 n. 7; Hirakawa 1990, 179–84.
148 dharma

respect to all the conditioned dharmas, with the exception of themselves,

because no dharma constitutes an obstacle to the arising of dharmas which are
susceptible to arising” (1988, 604). Obviously, it is important that conditioned
dharmas are causes with respect to each other and yet do not get in each others’
way, but Lamotte leaves it to be explained how all dharmas including uncondi-
tioned dharmas function in this manner.
As regards space, one of the six kinds of causes and two of the four kinds of
conditions are revealing on this point. The important cause is called the kāraṇahetu,
which Cox translates as “non-obstructing cause,”128 and the two conditions are
called the “object-support condition” (ālambana-pratyaya) and the “sovereign
condition” (adhipati-pratyaya). According to Cox, “in clarifying precisely what it is
that space ‘does,’” the *Mahāvibhāsạ ̄ states, “‘Space is unconditioned and lacks any
generative activity to produce an effect. However, it does function as the sovereign
condition (adhipati-pratyaya), that is, as a non-obstructing condition with regard to
the various material elements of space.’ The *Mahāvibhāsā thus delineates a series
of sovereign conditions, beginning with this non-obstructing function of
space . . . and ending with thought and thought concomitants” (Cox 2004, 57).129
This progression, allowed by the unconditioned dharma of space, of non-obstructed
sovereign conditions back to thought itself is reminiscent of Ṛgvedic enigmas
concerned with the “sources of dharma,” where dharma’s source and unfolding is
located in the regulated minds of the Ṛsị s. Moreover, with the “object-support
(ālambana) condition,” we find this Buddhist text linking the non-obstruction of
dharmas with a meditational “object support” in a fashion similar—other than
that it is without substance or substratum—to the usage of ālambana along with
Death’s “subtle dharma” of the Self in the Kaṭha Upaniṣad.
An “object-support condition” (ālambana-pratyaya) is also mentioned as
ārammana, “object,” among the up to twenty-four “conditions” (paccaya) enu-
merated in the Theravāda Abhidhamma (Lamotte 1988, 605; Gethin 2004, 527),
where it can also describe the “object” of the consciousness of the nine “transcen-
dent dhammas” (nava lokottara-dhammā, navavidha-lokottara-dhamma) that
include the one unconditioned dhamma, nibbāna.130 And it has some other

128. Cf. Williams 2000, 116, 259–60 n. 7, who translates it as “efficient cause,” which “consists of every
other dharma apart from the dharma that is the effect itself, inasmuch as every dharma either contributes directly
toward bringing about a further dharma (the cause as an ‘empowered’ kāraṇahetu) or does not hinder its produc-
tion (the cause as a ‘powerless’ kāraṇahetu).” Presumably, Cox is addressing how space functions in this second
manner.
129. Cox continues: “The text concludes, ‘if there were no space, this series of causes and conditions pro-
ceeding in this way would not be established. This fault must be avoided. Therefore, space actually exists with
intrinsic form (*svarūpa), and one should not deny it as nonexistent’” (2004, 557).
130. Gethin 2004, 527 discusses Carter, who notes some forty examples in the commentaries that say
dhamma in the Nikāyas should be understood as referring to nine types of transcendent dhammas: the four stages
of the path (from stream-winner to arhat) and their respective fruits topped off by nibbāna. “In the technical
 early buddhism 149

interesting Sarvāstivādin conjunctures. As Cox shows, “[e]lsewhere the

*Mahāvibhāsạ ̄ specifies that unconditioned dharmas function not only as sover-
eign conditions (adhipatya-pratyaya), but also as comprehensive non-obstructive
causes (kāraṇahetu) and as object-support conditions (ālambana-pratyaya). Their
function as comprehensive non-obstructive causes is identical to their function
as sovereign conditions: that is, in the case of all dharmas, unconditioned dharmas
also function not to obstruct the arising of dharmas other than themselves. And
in this function as non-obstructing causes, unconditioned dharmas function as
objects of mental perceptual consciousness, but here also they do not function as
the generative causes for its arising” (Cox 2004, 557–58). Here we see that space
functions as a “comprehensive non-obstructive cause” and as an “object of per-
ceptual consciousness”131 that does not, however, generate its arising, in the same
manner as the other two unconditioned dharmas, which have to do with cessa-
tion and thus nirvānạ . If we can now say, etymologically, that dharmas function
as “holds,” we could say that space functions as a special dharma or irreducible
“real” that holds everything and yet, in allowing everything else to hold that needs
to hold, also functions to not get in the way.
Further, what is said here of unconditioned dharmas pertains to all dharmas.
Even though “a given dharma is marked by a distinctive function in accordance
with its intrinsic nature,” it “can function in any of various ways, any of which
implies its existence. For example, all dharmas function as comprehensive non-
obstructing causes (kāraṇahetu) or as sovereign conditions (adhipati-pratyaya) in
not obstructing the arising of other dharmas. Dharmas also function . . . in the
arising of dharmas of the same type and as the object-support condition (ālambana-
pratyaya) for appropriate varieties of perceptual consciousness. . . . Moreover, for
the Sarvāstivādins, a dharma’s possible modes of functioning are not limited to
the present moment. All dharmas can function as non-obstructing causes, as sov-
ereign conditions, or as object-support conditions [ālambana-pratyaya] at any
given time, that is as past, present, or future” (572). Such usages of alambāna as
a causal condition in meditation are paralleled in the Yogasūtra, where Patañjali
uses ālambana as a technical term for a “support” or “stimulus,” as in: “Or [the
mind-stuff reaches the stable state] by having as the supporting-object a percep-
tion in dream or in sleep (svapna-nidra-jñāna-ālambanaṃvā)” (1.38)—which the
commentary shows could be a dreamed luminous and fragrant image of the

language of the commentaries, this refers to the four kinds of consciousness (citta) that arise as” the eight tran-
scendent dharmas linked to attaining the four paths and fruits “and lastly nibbāna as the ‘unconditioned element’
(asaṃkhata-dhātu), ‘object’ (ārammāṇa) of those classes of consciousness.”
131. See further Cox 2004, 577–78: “for Saṅghabhadra, unconditioned dharmas exhibit capability as non-
obstructing conditions and as object-supports [ālambanas] in the arising of cognition.”
150 dharma

“Exalted Maheśvara” in a forest, from which support the yogin awakens with
“undisturbed calm” (Woods [1927] 2003, 76). Here again, however, whereas a
dream or sleep ālambana is grounded in the same ontology as the deity, a
Buddhist ālambana, like all dharmas, must function in the open.

C. Vinaya Basket Dharma

Not everybody, of course, wants to be a metadharmician. As we have noted,

the Nikāyas mention three types of monks: those who maintain the Dhamma,
the Discipline (Vinaya), and the Codes (mātikās)132—a grouping that would seem
to have evolved further to define specialists in each basket, as well as those rec-
ognized for proficiency in all three.133 In the first part of this section we meet the
vinayadhāras—“experts in law” (von Hinüber 1995, 22)—who specialized in
Vinaya, the monastic Discipline. In the second and third parts we will look at
two narratives that make Vinaya the lifeline of the dharma in time.
Strangely, many Buddhologists and Indologists who have written about
dharma have tended to ignore the Vinaya, as if everything important that was
said about dharma in early Buddhism was said in the other two baskets.134 As
we have seen, one still gets summaries of the Buddhist meanings of dharma
that overlook the fact that it means “rule” with reference to the “offenses” that
comprise the early monastic code.135 Since these were rules for monastic life,
they were necessarily designed to take substance within a larger society. It is
thus fitting to take note of the historical context in which Vinaya has been
studied, which includes its relation to Brahmanical Law. While the need to
explore Vinaya’s relation to dharmaśāstra has been recognized for some time,
most scholars have not been able to say much with any specificity, and indeed
some have been tempted to say much that now looks dubious because they
were working with poor temporo-spatial coordinates.

132. MN 33.9 as cited above; see Ñāṇamoli and Bodhi [1995] 2005, 314; Gombrich 1990, 25–26.
133. One senses that the parallel has emerged with specialists in one or more Veda, as in Buddhist usage
of tevijja.
134. See, for example, Horner [1942] 1997, xi–xiii. For a good commentary on the proclivities behind such
scholarship in Buddhist (and wider studies, see Reynolds 1995, 3: “From the kind of Orientalist perspective that
developed in the early phases of Buddhist studies, true Buddhism was not a religion that had a strong legal com-
ponent. . . . ‘Law,’ when it was used as a translation for Dharma, was used with cosmic, philosophical, and/or
ethical connotations that were never associated—in any intrinsic way—with legal systems as codes.” With
“emphasis on the Buddha . . . and the Dharma . . . rather than the Saṅgha . . . [t]he notion that the vinaya and
vinaya commentaries constituted the textual locus of a specific legal system . . . was never seriously
considered.”
135. See Rhys Davids and Oldenberg [1881] 1968, 1: 3 n. 1, who remark that the translation “offences” for
this usage of Pāli dhammā “is no doubt right in taking the word . . . in a strictly technical, legal sense. ‘Offences’
is however not the right direction in which to limit the general sense. Dhammā must here be ‘Rules.’”
 early buddhism 151

Scholars have been prone to work with too early dates for Vinaya and
dharmaśāstra. As we shall see in the chapter 5, there is a recently emerging con-
sensus around the work of Olivelle that the earliest dharmasūtras entail a response
to Buddhist and other heteropraxies, with none of the dharmasūtras earlier than
the Mauryas. And as regards Vinaya, a consensus seems to be emerging that the
six surviving Vinayas of South Asian origins may be the latest canonical baskets
to be filled, as we have them, reflecting comparatively later processes of compila-
tion than the other two,136 and being less reducible to an archetype than has been
commonly thought.137 Schopen estimates that all six extant Vinayas were com-
piled during the first and second centuries CE (2004, 79–80, 210, 212), the
beginning of the “Middle period” of Indian Buddhism that marks the emergence
of the Mahāyāna (see Schopen 2000, 2006, 321, 325, 345).
Scholars have also been prone to overlook the ways that both Vinaya and
dharmaśāstra reflect locality. Here we face a somewhat counterintuitive situation.
Whereas the dharmasūtras almost certainly come from different locales and
acknowledge custom and tradition as two of the three main sources of dharma
(the third being Veda), and even in some cases take note of regional differences,
they are nonetheless, as Olivelle attests (1999, xxvii–xxviii), very hard to locate
geographically since they each project a pan-regional scope for dharma. Surviving
Vinayas, on the other hand, because they all claim to be the word of the Buddha
for monks and nuns and do not legislate for society, and also because of the pre-
sumption of an original archetype for them, give the appearance of going back
to a translocal common code, but actually reflect clear accommodations to local
custom and tradition that cannot go back to the time of the Buddha. Schopen
observes that “adaptation of monastic rule to local custom can be found in all the
Vinayas,” citing a Mahīśāsaka-vinaya passage that even justifies this:

The Buddha said: “. . . Even if something was authorized by me, if

in another region they don’t consider it pure, no one should follow it.
Even if something was not authorized by me, if in another region
there are people who must necessarily practice it, everyone should
put it into practice.”138

136. See Schopen 1997, 25–29; 2004, 194; 2006. Schopen makes much of the lack of “evidence for
Buddhist monasteries either before or during the Mauryan period” and the prevailing vinaya concern for ordering
life in monasteries, which “did not occur on any scale until well after Aśoka, and probably nearer to the beginning
of the Common Era” (2006, 316). For a tally of conventional views, see Gombrich 91–92, for whom the Vinaya
would be pre-Mauryan, and thus earlier than Aśoka. His premise is that things were “added after the Pali tradi-
tion separated from the rest (which happened during Asoka’s reign)” (91). Cf. Collins 1993, 335 n. 3.
137. See Clarke 2009, 24–26, 31 on questions “higher criticism” might raise regarding a Vinaya rule and
accompanying story about a penance allowed for monks or nuns who have had sex that is found in all six Vinayas
except the Theravāda, which, however, seems to have had such a rule (32–34).
138. From Jaworski 1929–30, 94; see Schopen 2004, 194. Cf. Schopen 2006, 317: “If the compilers of the
various Vinayas considered it ‘highly important’ to regulate the lives of their monks so as to give no cause for
152 dharma

Schopen suggests that local backgrounds can also be hypothesized even in the
presumably earlier Prātimokṣa where the six Prātimokṣas’ greatest variance in
Sekkhiya/Śaikṣa (“training”) rules may reflect their local environments.139
Through close scrutiny of numerous instances of local accommodations in the
Theravāda and Mūlasarvāstivādin Vinayas,140 Schopen has shown that, com-
pared with the Mūlasarvāstivādin Vinaya, which reflects considerable negotia-
tion with dharmaśāstra as an established legal system, the Pāli Vinaya, compiled
in a Sri Lankan legal environment by monks of the ascendant Mahāvihāra sect,
“shows little awareness of the early and elaborate Indian legal system articu-
lated in the Dharmasūtras and Dharmaśāstras” (2004, 210).141
This last point makes it unlikely that one can use dharmasūtra parallels or
precedents to interpret features of a Theravāda Vinaya that reflects little aware-
ness of Brahmanical Law. It is less of a problem to draw parallels and prece-
dents where the Suttas relate Vinaya concerns to dharmaśāstra themes, since
they would likely reflect Vinaya in an earlier stage of development; but still, if
Olivelle is right that the earliest dharmasūtras are no earlier than the Mauryas,
then we can date such parallels no earlier than that. We shall come back to
these matters in closing this chapter on the topic of Vinaya allusions in the
Aggañña Sutta.

complaint to the laity, and if considerations of this sort could only have assumed high importance after Buddhist
groups had permanently settled down, then, since the latter almost certainly did not occur until well after Aśoka
it would be obvious that all the Vinayas we have are late.”
139. Schopen 2004, 357–58 n. 62; cf. Prebish 1996, 263–70 on the Dharmaguptakas added rules on
monks’ behavior at stūpas (270). Regional variation could apply to the Śaikṣa rules regarding indecorous sounds
made while eating alms food: the Mūlasarvāstivādins’ Prātimokṣa has four: the sounds “cuccat,” “suśaśut,”
“thutyut,” and “phuphphuph”; the Mahāsaṅghikas’ has three: “cucu,” “surusuru,” and “śuluśulu” (rules 38–40,
rules 54–57; Prebish 1975, 100–101); the Theravādins’ has two: “capu-capu” or “smacking the lips” and “suru-
suru” or “hissing” (Rhys Davids and Oldenberg [1881] 1968, 1: 64–65; Horner [1942] 1997, 3: 137–38; rules
50–51); and the Sarvāstivādins’ only 1 (rule 73; see Idem 146). Locative variation would also apply to the next most
varied category of Paccitya/Pāyantika rules. For example, the Mūlasarvāstivādins’ rule 33 adds “Brāhmaṇa” and
pluralizes “monks,” where it limits to two or three bowls full of food the amount “many monks” may receive
when they “approach families, if faithful Brāhmaṇa householders invite them.” The corresponding Mahāsaṅghika
rule 38 reads only, “If a family should invite a monk . . .” (Prebish 1975, 79–81); and the much shorter Theravāda
rule 33 is, “Begging straight on from house to house will I eat the alms placed in my bowl” (Rhys Davids and
Oldenberg [1881] 1968, 1: 63).
140. Undercutting presumptions of the relative antiquity of the Pāli Vinaya, see Schopen 1997, 205,
225–26, viewing of it as “in some cases ‘markedly inferior’ to the other Vinayas, and in some cases appear[ing]
decidedly later”; Schopen 2004, 195–203 where mention of royally appointed monastic servants in the Pāli
vinaya “may reflect . . . Śri Lankan practice” in preparing rock-cut cells, and seems to “abridge” what the
Mūlasarvāstivādin-vinaya says; similarly Schopen 2006, 340–45. Cf. Schopen 1997, 86–96, 72–80, 84 n. 22,
205–37; Schopen 2004, 1–18, 94–96, 193–218.
141. See Schopen 2004, 81: “it is becoming ever clearer that the Mūlasarvāstivādin-vinaya may have particu-
larly close ties to brahmanical concerns, and this, in turn may suggest that it was redacted by a community deeply
embedded in the larger Indian, brahmanical world. It may turn out to be the mainstream Indian vinaya. Time will
tell.” See von Hinüber 1995, 43 n. 102 on the “new and quite unexpected light” shed by Schopen’s 1994 success
(=2004, 45–90) “in finding influence of Dharmaśāstra on a Vinaya,” that is, that of the Mūlasarvāstivādins.
 early buddhism 153

C.1. Vinaya and Prātimokṣa

Obviously, use of the name “Vinaya” for a category of the Buddha’s teachings is
early. Here we need two distinctions. First, one must distinguish between the
early monastic code and Vinaya proper, since, although it is integral to Vinaya and
to any discussion of it, the code itself, called the Pātimokkha/Prātimokṣa, stands
textually, as a sutta or sūtra, outside the Vinaya Basket. The Pātimokkha Sutta
would seem to have this separate status because its “liturgical” character gave it a
use outside the “lists of the Piṭaka texts”;142 but one must also not discount the
possibility that it may have been a prior text,143 at least to the Vinaya proper, and
one around which portions of Vinaya textual composition nucleated.144 Second,
one must mark a distinction between developing Vinaya and the developed Vinayas
of “Nikāya Buddhism,” such as have survived from six “Nikāya” schools.
The two main portions of the Vinaya Basket proper, the Sutta Vibhaṅga
(Sūtra Vibhaṅga) and the Khandhaka (Skandhaka), which treat individual and
communal rules, respectively, no doubt recall a process of developing Vinaya
that would have begun early. While the Sutta Vibhaṅga probably nucleated
around the Pātimokkha rules, the Khandhaka may have had similar though
more partial grounding in the kammavācās—“legal formulas” that “have to be
recited to transact legal business of the order,” such as appointing specific
monks to specific tasks and including “admission of new members to the
order.”145 In the Sutta-Vibhaṅga (“The Analysis of the Sutta” [Prebish 1994,
20]), the Buddha explains one-by-one each rule of the Pātimokkha-Sutta and his
subsequent riders (loosenings and tightenings) on that rule as first expounded.146
The “casuistic” nature of the rules and their riders suggests a period beyond the

142. Rhys Davids and Oldenberg [1881] 1968, xiv–xv; “The cause which led the Pātimokkha, and the
Upasampadā-kammavāka, being separately preserved at all, is the same as the cause which led to their exclusion
from the lists of the Pit.aka texts—the fact, that is, of their being liturgical compositions.”
143. Recall that sutta, as listed first of the nine “Parts” or Aṅgas into which the Buddha’s “pre-canonical”
teaching was early divided, may have referred to the Pātimokkha Sutta.
144. This is the standard view, which Schopen seems not to dispute (2004, 133–34); rather, he offers strat-
egies for understanding how Vinayadharas found ways around or sophisticated ways through their schools’
Pratimokṣa rules (Idem, 11–14). Cf. von Hinüber 1995, 14: “The Pātimokkha as we have it today, must have been
formulated at an early date, and not by the Buddha.”
145. See von Hinüber 1995, 19, stressing that “[t]he wording of these formulas is fixed exactly, down to the
correct pronunciation of single sounds; for phonetic mistakes such as pronouncing a labial instead of a nasal in
saṃgha versus saṃghaṃ would result in the invalidity of a legal act” (19–20).
146. See von Hinüber 1996, 13: In the Suttavibhaṅga, “Every single rule is embedded in a text that begins
with an introductory story (vatthu) describing the occasion on which the rule was prescribed by the Buddha. Then
follows the rule as such (paññatti), which may be supplemented by additional conditions (anupaññatti), and
which may be accompanied by word for word explanation (padabhājaniya). Finally, exceptions to the rule are
enumerated (anāpatti ‘no offence’).” Cf. Schopen 2004, 13: such “exceptions” include “exemptions, exclusions,
extenuations.”
154 dharma

lifetime of the Buddha in which case histories,147 some “evidently invented for
the purpose” (Olivelle 1974, 57), with an especially rich imaginary assembled
on sex,148 could have been assembled, and time for vinayadhara casuists to
develop an interpretative stance to pass off their explanations of each rule and
its exceptions as Buddhavacana, the Word of the Buddha.
It is here that we meet the force of Vinaya behind the formation of differ-
ent “Nikāya schools”: “Monks cannot co-operate in a pātimokkha ceremony if
they do not share exactly the same pātimokkha code,” which they (and nuns)
would have ratified each fortnight as a formulary of rules that they were con-
sensually committed to recite on these bimonthly occasions, and from which
each must accept the rule’s discipline for any admitted infraction. If as few as
four monks disagreed with others over the rules, they could form their own
saṅgha with a different Prātimokṣa; and this could eventually be perpetuated in
a nikāya or “sect” with its own ceremony of higher ordination (upasampadā)—a
procedure treated in the Khandhaka. Geographical separation clearly played its
role in such variation, and doctrinal differences could accompany splits, and
may have done so increasingly over time. But a split itself would be formalized
via Vinaya, and “had in any case to be actualized by performing one’s own
pātimokkha ceremony” (Gombrich 1988, 111–12). As mentioned, six Nikāya
schools have left extant Vinayas: one of them, the Theravāda in Pāli; and all the
others in either Chinese (the Mahāsaṅghikas, Sarvāstivādins, Dharmaguptakas,
and Mahiśāsakas) or Tibetan (the Mūlasarvāstivādins) translations from
original Sanskrit compositions, only fragments of which survive.149 It is note-
worthy that the Chinese and Tibetan translations of these “Nikāya Buddhist”
Vinayas came to be used mainly (and in Tibet exclusively) by the Mahāyana.
Gombrich even goes so far as to say the Mahāyana “is not a sect, but a current
of opinion cutting across sects as properly defined. There is no such thing as a
Mahāyāna pātimokkha” (Ibid., 112). Williams goes a step further: “There is no

147. See Beyer 1974, 69; Holt 1983, 47, 52 (“case history after case history”); von Hinüber 1995, 35: evi-
dence at least in the commentaries “that there seems to have been collections of precedents.” Comparatively, see
Olivelle 1974, 57: Vinaya judgments are made “on a particular case or issue at hand” in contrast to “the normal
practice of Hindu law where the codes contain general laws and not casuistry.”
148. Gyatso 2005, 277: “At the very least, the many lusty people and acts depicted in the Suttavibhaṅga
bears witness to a very active sexual imagination. We find monks having sex with fresh corpses, rotting corpses,
dolls, dildos, and a plethora of live partners crossing sex, gender, and species lines in every imaginable way. . . . It
is certainly not impossible that some of the rules were formulated in response to things people actually did. But
. . .” Cf. Clarke 2009, 35–36.
149. See Prebish 1994, 62–63 (a portion of Mahasaṅghika Vinaya); 79–80 (fragments of Sarvāstivādin
Vinaya); 89 (a portion of the Mūlasarvāstivādin Vinayavastu). Prebish 1975 translates the Sanskrit Prātimokṣa Sūtras
of the Mahāsaṅghikas and Mūlasarvāstivādins. According to Prebish and Nattier 1977, 238–39 and Prebish 1996,
261, the earliest split would have occurred as a result of “a reaction on the part of the future Mahāsaṃghikas to unwar-
ranted expansion of the root Vinaya text on the part of the future Sthaviras” (from whom the other five schools
derived; see Prebish 1996, 269).
 early buddhism 155

such thing as a Mahāyāna Vinaya” (2000, 100, author’s italics). Mahāyāna

schools of thought produced new sūtras, and in the case of the Yogacāra, an
Abhidharma, but they used “Nikāya school” Vinayas and are thus, in this sense,
not sects. Rather, as recent research has shown, the early Mahāyāna developed
its new teachings within “Nikāya school” institutions, with all that may imply
about the civil character of their contending discourses on dharma.
Fundamental, then, is a commitment to the idea of Prātimokṣa recitation,
which may indeed go back to a “pre-canonical” sūtra text. In order that the
ongoing discussion remains clear, it is worth outlining the seven basic cate-
gories of Prātimokṣa rules (dhammas, dharmas) while mentioning certain rules
or rule clusters that will be noticed further. The seven categories, which apply
to all six “Nikāya schools,” are listed in their conventional order of recitation,
which moves from the most severe to the most incidental and procedural. Nuns
have considerably more rules and one less category.150 The Pāli spelling is given
first and the Sanskrit second, if they significantly differ:

1. Pārājikā, perhaps meaning “(suffering) defeat” or “downfall” (Prebish

1975, 11): four offenses—sexual intercourse, theft, murder, and flaunt-
ing spiritual attainments—which require expulsion from the Saṅgha.

In the Theravāda Vinaya, sex is most basically defined as “that which is not the
true dhamma” (asaddhamma) and called “village dhamma” and “vile dhamma.”
The Buddha devises a rule that defines penetration, and thus expulsion, by the
“length of the fruit of a sesame plant” (Gyatso 2005, 277–78). In the Aggañña
Sutta, it is straightforwardly condensed among “five impossible things” as “He
cannot have sex” (methunaṃ dhammaṃ pat·iseviṃsu) (Collins 1993, 327, 330).151
This first rule is worth a pause because it will be important. The Sutta-Vibhaṅga
defines it as “the one work” (ekakamma), “one instruction” (ekuddeso), “equal
training” (samasikkhātā): the monk in communion has the same work, rule,
and discipline as everyone in the community. And it is also the place where he
interrogates his own memory and accountability”: “Not only does a monk’s sex
break a law, the first rule of the vinaya. Given the way the rule against sex sig-
nifies the very being of the community, to have sex breaks the basis of law

150. Prebish 1975, 47; the category 3 aniyata dharmas are omitted without gender reciprocity. See
Kalbisingh 1998, 2, listing the numbers of rules by the seven categories and six schools. Nuns rules are most
inflated in category 1 (eight instead of four could require expulsion) and category 5 (rather than 90–92 rules requiring
expiation, nuns had from 141 (Mahāsaṅghikas) to 210 (Mahiśāsakas), with the Theravāda at 166, the Dharmaguptakas
and Sarvāstivādins at 178, and the Mūlasarvāstivādins at 180).
151. On “village dhamma” and “vile dhamma” (visaladhamma) as Suttavibhaṅga glosses on asaddhamma,
see Idem, 278; Gombrich 1988, 61 (gāmadhamma). As we shall see in later chapters, Pāli methunam dhammam
has its equivalent in the Mahābhārata as maithunam dharmam. Both traditions agree that there is a powerful
force called the “law of sex” or “law of copulation.”
156 dharma

itself. . . . The decision to refuse sex means no less than to honor the rule of
monastic law altogether” (Gyatso 2005, 286, 288). Yet the Theravāda is the
only school to make the rule an absolute condition for staying in the monastic
community. In all the other five monastic law codes, there is an “origin tale”
about a monk named Nandika whose seduction by a daughter of Māra, and his
having had no thought in the aftermath of concealing the matter, sets the stage
for the Buddha to establish a rule that allows a repentant monk (or nun) to be
granted a “life-long penance, namely, the status of a śikṣādattaka,” which seems
to mean “one who has been granted the training (or penance)”—“a unique pro-
bationary status within the monastic hierarchy below the most junior of the
monks, but above the novices.”152 This probationary status can apparently also
be granted for the other three pārājikas as well, with there being three possible
outcomes: the monk or nun “either leaves voluntarily, becomes a śikṣādattaka,
or is banished” (Clarke 2009, 22).

2. San· ghādisesa, Sam· ghāvas´es. a: Thirteen offenses calling for a

153

probationary period of temporary exclusion from the order:154 five

deal with sexual transgressions; two with dwelling places; two with
schisms; one with a monk who is difficult to speak to; and one with
monks who corrupt families (Prebish 1975, 11–12).
3. Aniyata: Two offenses where a monk is accused on the testimony of a
trustworthy female lay follower (upāsikā), whose word, as sole witness
to a sexual impropriety, can be evaluated to assess whether it is a
category 1, 2, or 5 offense.155
4. Nissaggiya-Pā cittiya, Nih· sargika-Pā yantika: Thirty offenses requiring
expiation and forfeiture after confession. Ten pertain to robes; ten to
rugs and use of gold and silver; and ten to bowls, medicines (ghee,

152. See Clarke 2009, 8, 16—and correspondingly for penitent nuns, 3 n. 4; 8 n. 3 (I summarize mainly
from the Mūlasarvāstivādin version). On Gyatso’s statement (2005, 276) that sex is “the premiere downfall
(pārājika) that ends a monk’s or nun’s career,” Clarke 2009, 6 n. 19 comments, “Such a conclusion would have
been hardly possible on the basis of the Vinaya preserved in Tibetan.”
153. Prebish 1975, 12 says “No etymological rendering of the term seems to make much sense.”
154. According to Prebish, the probationary period “lasted as many days as the offense was concealed”
(1975, 12) and was designed for offenses that were “concealed for some time” (123 n. 33).
155. von Hinüber 1995, 10–11: “the Suttavibhaṅga adds (and thus at the same time mitigates the rule) that
it is necessary, too, that the monk does not deny having committed the respective offense.” Cf. Prebish 1975, 12:
[showing] “an outstanding and somewhat surprising degree of trust in the female lay follower.” A woman’s tes-
timony can determine whether the monk should be punished for a category 1, 2, or 5 offense requiring expulsion,
probation, or expiation/penance. The two types are distinguished by whether the place where a monk would “sit”
with a woman was secret and convenient for sexual intercourse. In this case all three categories must be consid-
ered. But if in the second case where the spot was only “suitable for speaking to her in lewd words,” the first
category could be ruled out.
 early buddhism 157

butter, oil, honey, and raw sugar) that can be stored only up to seven
days, and further rules for robes.156
5. Pā cittiya, Pā yantika: Ninety to ninety-two offenses, numbered variously
by school, requiring expiation or penance.157 Seventy-four of ninety come
under five categories: (a) moral rules, for example, lying (23); (b) conduct
with nuns and other women (14); (c) food and drink, including rules for
accepting food from lay households (16); (d) application of Dharma,
Vinaya, and Mātṛkā teachings (11); and (e) use of requisites (10). The
remaining sixteen break down further into behavior in the monastery (6);
travel (5); and various types of destruction (5).158
6. Pā.r idesanı̄ya, Pratides´anı̄ya: Four offenses requiring confession: two
concern receiving almsfood in the same place as a nun; one, taking food
from a family in training; and one, taking food in a dangerous place.
7. Sekkhiya, Śaiks.a: The most disparate grouping both in the order listed
and in the total number. Theravādins and Mūlasarvāstivādins,
for instance, have seventy-five and one hundred and eight rules,
respectively.159 Sometimes called training rules, they are concerned
with daily monastic conduct and manners, for which there will be
no sanction or punishment. They include numerous guidelines
for “going amongst the houses” and eating decorously.160
8. Adhikaran. as´amatha: Seven rules providing procedures for resolving
disputes over offenses. These include means of reaching verdicts such
as majority vote by tickets; handling appeals; criteria for assessing
insanity; smoothing things out with other monks and, where an
offense is “connected with householders,” with the laity, after harsh
things were said; and a requirement that a verdict may not be carried
out without the confessing monk’s acknowledgment.161

Suffice it to say that the eighth category shows considerable experience with
litigation, but this can also be detected in some of the prior categories, as will

156. See Prebish 1975, 13: “This is the first class of offenses in the Prātimokṣa in which the numbering system
employed by the various schools became widely divergent.”
157. Lamotte 1988, 166: “faults entailing penance”; Conze [1959] 1977, 76, “offences which, unless
repented and expiated, will be punished by an unfavorable rebirth.” But cf. von Hinüber 1995, 12: “The
consequence of transgressing a Pācittiya are not clear.” He says the term (akin to Sanskrit prāyaścitta, “atone-
ment”) points to a Vedic ritual background as well as to parallel usages in dharmaśāstra.
158. Slightly modifying Prebish’s breakdown (1975, 14–15).
159. Prebish 1975, 15: 66 rules in the Chinese Mahāsaṅghika version; 113 in the Chinese Sarvāstivādin
version; 67 rather than 66 in the Sanskrit Mahāsaṅghika (Idem, 148; 1996, 263).
160. See above n. 139. The Mūlasarvāstivādins add as their Śaikṣa rule 39, “We shall not sit down on a seat
amidst the houses exposing the genitals” (Prebish 1975, 101).
161. Prebish 1975, 16–17; cf. Ñāṇamoli and Bodhi [1995] 2005 852 ff., 1311 ff.
158 dharma

be noted. As Vinaya begins to familiarize us with tensions between Buddhist

and Brahmanical law, it introduces a number of intriguing complexities.

C.2. Vinaya and Small-Scale Early Republics

Gaṇa-saṅgha is a name for the small-scale “oligarchic-republican” polities that

proliferated in northeast India in the early phase of the second urbanization,
where their histories are reflected in narratives about the Buddha, particularly in
the Suttas and the Vinaya. Their currency raises the question of whether the
consensual nature of Vinaya reflects gaṇa-saṅgha laws and polities, such as those
of the Sakyans from whom the Buddha is said to have come, and of whom he
could speak with nostalgia where he defends their conduct (particularly, their
nondeference to Brahmins) at their court at Kapilavatthu by telling young
Ambat.t.ha, “even the quail, that little bird, can talk as she likes in her own nest.”
Among the implications we noted, the gaṇa-saṅghas were endangered during
the Buddha’s lifetime by the ambitions of rising monarchical states,162 and were
also headed for tension with the ways that Brahmanical Law had begun to theo-
rize royal power in the context of varṇa (social class) hierarchy (see Lang 1992).
We see the first of these oppositions and probably also a glimmer of the
second in a famous passage in the Mahāparinibbāna Sutta,163 where the Vajj īs,
another of those ill-fated gaṇa-saṅghas, are in the way of King Ajātasattu of
Magadha. Ajātasattu tells his Brahmin Chief Minister Vassakāra to go and tell
the Buddha, “King Ajātasattu wants to attack the Vajjīs. He said, ‘I will anni-
hilate these splendid and powerful Vajjīs, destroy them, bring them to utter
ruin.’” Ajātasattu orders Vassakāra to listen carefully to what the Buddha says
and report back. Upon hearing this parricide king’s164 precise words, the
Buddha avoids replying to Vassakāra directly.165 Instead he asks his beloved

162. See Collins 1998, 66, 445; Thapar 2002, 137–38, 146–51—though as Thapar points out, gaṇa-saṅghas
have a continued marginalized history at least up to the Gupta period (283–84).
163. See Gombrich 1988, 49: “according to the Buddha (or, strictly speaking, according to words attributed
to him in the Mahāparinibbāna Sutta), he modelled the organization of his Sangha on that of such communities
as his own.” As we shall see, this is at best an inference. Gombrich goes on to find no evidence of caste (varṇa)
or for that matter kingship among the Sakyans, though they did have servants, and to suggest that they may have
made age their “ranking principle” (49–50). See further Idem, 109: “acephalous polities” enjoined to “hold ‘full
and frequent assemblies.’ ” Collins 1998, 436 says the passage appears “in almost exactly the same form” in the
Aṅguttara Nikāya (at IV 16ff.).
164. He had killed his father Bimbisāra, with whom the Buddha had already established the principle of
acceding to kings in matters of Vinaya. When Bimbisāra once requested that the Buddha postpone his bhikkhus’
rain retreat for a month, he accommodated him and told his followers: “I prescribe, O bhikkhus, that you obey kings”
(Book of Discipline 4: 184–85; Holt 1983, 50). Bimbisāra had also asked the Buddha to preach the dhamma on the
Uposatha as other paribbājjakas were doing, whereupon the Buddha said that it would henceforth be the time for
monks to recite the Pātimokkha (Mahāvagga II, 1–5) (Book of the Discipline 4: 130–37; Holt 1983, 38). On treatments
of the future Buddha’s first meeting with Bimbisāra as king of Magadha, see chapter 13.
165. See Collins 1998, 443–44 for a fine discussion of this slight of Vassakāra.
 early buddhism 159

disciple and cousin Ānanda seven times what he has heard about the Vajjīs
with regard to seven aspects of their laws and conduct; and after each answer,
the Buddha replies that so long as the Vajjīs continue that behavior, “one can
expect them to prosper, not to decline.” Finally, he tells Vassakāra that he once
taught the Vajjīs these seven very things himself. Here we see the Buddha as
a virtual (though only partial) law-giver to the republican-oligarchic Vajjīs,
while we can assume that Ajātasattu is getting his main advice from this
Brahmin, who now, as Vassakāra’s probably sly response shows, tells the
Buddha he sees that Ajātasattu will “not be able to conquer the Vajjīs, at least
not (simply) in warfare, without deceit and (fomenting) internal dissension.”
Once Vassakāra leaves with, as he puts it, “a lot to do,”166 the Buddha assem-
bles all the monks dwelling in Rājagaha (the Magadha capital) to commend
seven analogous things that will lead them to prosperity. The passage thus
establishes parallels between the laws and conduct of the Vajjīs and the Vinaya
code for monks, with both said to have the Buddha’s authority behind them:

(1) Just as the Vajjīs were told they should “continue to meet together
in assembly frequently,” so should monks; (2) both should assemble
and conduct their affairs harmoniously; (3) just as the Vajjīs were not
to “establish (any laws) which are not already established” or “rescind
any established (law),” and should “proceed in accordance with the
traditional way of the Vajjīs,” monks should “not establish (any
Monastic Rules) which are not already established” or “rescind any
which are (already) established,” and should “proceed in accordance
with Training Rules as they have been established”;167 (4) both should
venerate and heed their elders; (5) just as the Vajjīs should not “carry
off forcibly women and girls of good family to have as their wives,”
monks should “not fall under the sway of the (kind of) desire which
leads to rebirth, when it arises”;168 (6) just as the Vajjīs should
venerate and maintain their shrines both inside and outside the city,
monks should “look to (secluded) forest dwellings”; and (7) just as
the Vajjīs should “provide proper guard, shelter, and protection for

166. Cf. Collins 1998, 444–45, speaking of Vassakāra’s “sardonic rider” and this final remark as “dripping
with dramatic irony.” Collins says, “it would seem, the Buddha’s remarks have pinpointed just what he needs to
do to bring about what his king Ajātasattu’s violent words have envisaged”; the commentary explains his con-
quest and absorption of the Vajjīs with just such a reading; the Buddha “spoke as he did to Vassakāra out of
compassion for the Vajjīs” as a way to buy them time in the face of the inevitable (445 n. 42).
167. These are the Sekkhiya dhammas “pertaining to training” that form the seventh grouping in the
Pātimokkha. See above.
168. It is suggestive that this desire is compared to marriage by abduction, which Brahmanical law will call
Rākṣasa marriage. As Collins 1998, 443 indicates, it implies “celibacy for monks and thus Pātimokkha rule 1,
requiring expulsion.”
160 dharma

their Arahants,” welcoming future ones and maintaining comfortably

those now in their territory, monks should “continue to establish
mindfulness, so that congenial companions in the celibate life”
might be welcomed to their monastery in the future and maintained
comfortably there now. (see Collins 1998, 437–41, as quoted)

Set toward the beginning of the Buddha’s final journey and “great decease,” the
passage seeks to secure the Vinaya’s future while anticipating the imminent
passing of the laws of the Vajjīs, which, no matter how excellent they may be as
commended by the Buddha himself, are soon to be overwhelmed by a
Brahminically coaxed monarchic order. As Collins says, the passage shows a
“Janus-like ambiguity,” for “the virtues which characterize the monastic life, or
some of them at least, prove incapable of defending a lay community which
adopts them” (1998, 445). The Buddha’s “Janus-like” diversions recall his tacit
but knowing role in bringing about the end of his own royal line, that of the
Śākyas, and with it, the annihilation of their “little republic.”169
Parallels between Vinaya and gaṇa-saṅgha polities have interested scholars
not only in the genetic question of the origins of Vinaya, which must remain a
matter of reasonable speculation, but as evidence of “democratic” trends in
early Indian law. Further evidence for such trends has been sought in guilds,
which may also have shared consensual legal charters and processes with the
early Saṅgha and reflected some of the commercial ties of both types of associ-
ations with the second urbanization. The label “democratic” may serve to
describe a countervailing tendency to monarchism, but it is at best a loose
usage. Any juxtaposition between Vinaya law and secular law must acknowl-
edge that the former, particularly as it requires celibacy, cannot be law for a
whole society. In fact, the Buddha only recommends seven things to the Vajjīs,
not a total legal charter. Commentaries on the Dīgha and Aṅguttara Nikāyas
also speak favorably of “the old laws of the Vajjis,” whose kings made widening
inquiries before punishing a thief (von Hinüber 1995, 33–34). Clearly the two
sets of seven things come to focus not on the transient value of the first set to
the Vajjīs but on the more enduring importance of the second set to the Saṅgha.
And we can see in them some of the things that make Vinaya jurisprudence
distinctive. The Buddha begins with clear references to Pātimokkha rules—in

169. See Strong 2010, notably 13: One way to read “these various episodes in the Buddha’s lifestory . . . is to
see the Buddha as the terminator . . . specifically of the Śākya royal lineage. Indeed, if we are to believe the Pali Vinaya,
the Buddha had a bad reputation as the ender of lines: At that time, [we are told], many distinguished
young . . . noblemen led a religious life under the direction of the Blessed One. The people were annoyed, murmured,
and became angry (saying): ‘The Samaṇa Gotama causes fathers to beget no sons; the Samaṇa Gotama causes wives
to become widows; the Samaṇa Gotama causes families to become extinct.’ ” He cites Rhys Davids 1882, 1: 150.
 early buddhism 161

directives 1 and 2 to the rules concerning schism; in directive 3, he admonishes

not to add or rescind rules and in particular to maintain the Training Rules;
directives 4 and 7 offer strong reminders of Pātimokkha rule number 1, celi-
bacy; and then he goes on to more general Vinaya concerns of the ongoing
Saṅgha—respect for seniority; emphasis on both forest and monastic residence;
and all this for the pursuit of meditation and the comfort of Arhats. Yet albeit
that Vinaya law is distinctive, it remains a royal road into the complexities of
both Buddhist and Brahmanical dharma.

C.3. Vinaya Allusions in the Aggañña Sutta

As we make our way toward Brahmanical dharma, it is worth making one last
pause over a Buddhist text that seems to make allusions to both Dharmasūtra and
Vinaya.170 The Aggañña Sutta (DN 27) has been richly rethought and usefully
retranslated by Steven Collins as “The Discourse on What is Primary.”171 It is an
important text in the comparison of Buddhist and Brahmanical Law for many
reasons, but most famously because it accounts in Buddhist terms for the origins
of kingship. But Collins has made a major advance by detailing this text’s Vinaya
allusions (1993), while Richard Gombrich also calls attention to an intriguing
dharmasūtra parallel in it (1992a, 172–74). Collins sets the Aggañña Sutta [hence-
forth AS] in a pre-Mauryan and pre-imperial setting mainly on the grounds that
it refers to rice surpluses that would have played a part in creating the institution
of kings; that it talks of the first king as a “Great Appointee” (Mahāsammata)
rather than a “Wheel-Turning Emperor” (Cakkavatti, Cakravartin); and that it
makes no reference to imperial formations, as does the Cakkavatti-Sīhanāda
Sutta, which he thus deems to be later (DN 26) (306–11, 316–18, 323–24, 375,
382; cf. 1996, 422–23; 1998, 65, 89). Gombrich too sees the AS as pre-Mauryan.
Yet I do not think either succeeds in making this sutta an exception to the view of
Witzel and others that the suttas reflect conditions of about the early third century
BCE.172 Collins’s key argument is the lack of reference to imperial formations,
which is an argument from silence. But there is also an obvious question: Why
would the first king already be an emperor?
Collins fruitfully approaches the AS as a parable about the fall of Beings
from a “ou-topia, ‘No-place,’ and eu-topia, ‘Good Place’” (1993, 315) into everyday

170. A discussion of the material in the section was presented at the International Association of Buddhist
Studies panel in 2008. It appears in Hiltebeitel 2009c.
171. Collins 1993, 338–49; 1998, 627–34. It was first translated into English by Rhys Davids as “The
Buddhist Book of Genesis,” and often called that since.
172. See Witzel 1997a, 308–9, 312; Gethin 2006, 77, 82–84, discussing a proposed third- or fourth-century
BCE date for the Mahāsudassana Sutta (DN 17).
162 dharma

Indian social conditions, and suggests that the Brahmanical tradition of

dharmaśāstra be studied from a similar “o/eu-topian perspective” (336 n. 27).
There would be something to this, but also the major difference that Brahmanical
dharma was much more based on longstanding cultural specificities (custom,
tradition, Veda) that were not only old and real to Brahmans—as the AS itself
makes clear—but likewise all too real to Indian (if less so to Sri Lankan)
Buddhists. As Collins observes, in contrast to Buddhist texts produced in Sri
Lanka and Southeast Asia, “Indian Buddhist texts were produced in a milieu of
constant endemic ideological plurality: kings, Brahmans, Jains and others all
had their own hierarchical models of social relations” (1993, 325).
If then, as Collins holds, the AS is an especially “context-sensitive” text,
and if the goal, as he holds, is a “reading” that can “be accorded historical
value” (1993, 323), it is my hope that its interface of Vinaya and Dharmaśāstra
will be worth another close look.
Collins thinks the AS “was intended by its earliest composer(s) and redac-
tors to be a humourous parable” (1993, 314). His choice to speak of it as a “par-
able” is adroit, for he wants to show that it is not a “myth” that can be approached
with any profit structurally (312), or as a charter myth of origins.173 He eschews
the chop-and-block methodology of the so-called “higher criticism” that licenses
scholars to dismantle texts by divining their “original”174 components and
detecting later interpolations and additions, among which wrap-around frame
stories are inevitably the easiest things to spot and dismiss.175 From the start,
we should appreciate the gains made by this literary approach, with its

173. Some later Sri Lankan and Southeast Asian Theravāda texts have taken the AS’s “parable,” without the
frame, as “a legitimatory myth-charter” (Collins 1993, 325; cf. Huxley 1996a, 1996b; Collins and Huxley 1996;
Gombrich 1992a, 161–62, 165–66, 175–76, contrasting “myth” and “parody”). Huxley, a legal historian who per-
sists in calling the AS a myth (1996a, 416), belittles Collins’s and Gombrich’s penchant for “romps thorough
literary theory” (412)—on which I applaud them. Collins mentions a Mūlasarvāstivādin-vinaya version without
the frame, but gives no detail (323–24 n. 37). There are three Chinese “versions” called “The Small Sūtra of
Origins,” “The Sūtra of the Four Varṇas,” and “The Sūtra of Origins to the Two Brahmins Vāsiṣṭha and
Bhāradvāja” mentioned by Gombrich 1992a, 165, whose source and study leads him to the opinion that the name
Aggañña Sutta cannot be reconstructed as original. But these look to me like renamings. See also Nattier 1991,
11 n. 3 on these “three translations” and the various “Nikāya schools” behind the likely originals. I know of no explic-
itly Mahāyāna retellings, although Nattier mentions a Mongolian adaptation of its “first king” account (12).
174. See Collins 1993, 312: “the mania—which is what I think it is—for an ‘Ur-text’ is entirely misplaced.”
175. Gombrich 1992a, 165 agrees in principle, but says little about the frame. Both cite and argue with
several scholars on these points. For a good discussion of the issues involved in applying these methods carte
blanche to Indian texts, see Nattier [2003] 2005, 49–50, despite which she recommends textual stratification to
discount the frame of the early Mahāyāna Ugraparipṛcchā Sūtra. Cf. Veidlinger 2006, 47–48, intent on dismiss-
ing the frame of the Traibhūmikathā, a Thai vernacular Buddhist cosmology, because it mentions numerous
textual sources while Veidlinger favors a general oral milieu. The issues will recur for us in discussing Nattier’s
treatment of frame stories in chapter 7, and with regard to the frame stories of the Sanskrit epics and the Mānava
Dharmaśāstra in chapter 5. On the sophistication of the Mahābhārata frames and the cost of excising one of them,
see Hiltebeitel 2006a. Generally, frame-removal arguments pay the cost of yielding preferential readings.
 early buddhism 163

hermeneutic of reading the AS whole, its frame included, even presuming an

initially oral composition.176 For Collins, the AS “as we have it is a coherent and
continuous whole, with lexical, semantic, and thematic elements common to
both the parable of origins and its frame” (324).
Taking aggañña to mean “what is primary,”177 from agga, “first, primary”
(= Sanskrit agra), Collins shows correspondences between these terms and two
other “key-words”: set··tha, “best” (= Sanskrit śreṣ·tha), and the prefix brahma-,
which “allows puns on the name of the Brahmanical god Brahmā” (331–32,
332). Through these “key-words” he focuses in on the question of vaṇṇa, which
(like Sanskrit varṇa) ranges in meaning from social class to color, complexion,
good looks, and “appearance,” with all these senses reinforcing each other
(332–34; cf. Gombrich 1992a, 163). Yet although Collins traces another thread
through the same skein, he does not scrutinize it as closely as I believe it
deserves. Let me call this thread “the housing problem,” taking in representa-
tions of householders, houses, and leaving home. Clearly, in a text where the
Buddha repudiates the reported view of Brahmins that they are the best (set··tha)
social class (vaṇṇa) by saying that ascetics (samaṇas) are “what is primary” in
regard to vaṇṇa, and further that Arhats are “what is primary” among beings,
householders are not “what is primary.” Indeed, across “Nikāya schools,” the
cardinal Prātimokṣa rule about “no sex” would seem to make not having sex
less important than not being a householder.178 Yet what is distinctive about the
AS is not a repudiation of the primacy of Brahmins, whom the Buddha
undermines—often with a certain ambivalence179—right and left in other texts.
Rather, in a narrative thread taken dramatically and pungently from the frame
through the parable, what the Buddha undermines is a view that connects the
“bestness” of Brahmins to the ideal of the householder.
We have met mahāśāla/mahāsāla Brahmins earlier in this chapter; they are
certainly householders, and the term has even been translated that way (Black
2007a, 72). But that is a backreading. In both the Upaniṣads and the Buddhist

176. Collins 1993, 324; cf. 331, 378. Cf. Gethin 2006, 81, 86, crediting Collins on “understanding these
suttas as redacted wholes,” while also arguing that repetition is an oral element in the Mahāsudasana Sutta (93,
99–100). Cf. Manné 1990, 81: “If one accepts the antiquity the category of Debate suttas, then one must accept
that long suttas are not necessarily amalgams of ‘bits’ of the Teaching.”
177. Collins 1993, 331 draws on Gombrich’s discussion. Gombrich 1992a, 169–71 rejects a derivation from
agga-jña, “origin-knowing” (see e.g. Walshe [1987] 1995, 407) and posits “an adjectival suffix -ñña formation”
whereby “aggañña means something like ‘primeval’ and has nothing to do with ‘knowing’ ” (170); it also “appears
to be virtually synonymous” with porāṇa, which it “always immediately follow[s]. Nattier 1991, 11 n. 3 already
translates aggañña as “primeval.”
178. See above at n. 151. Cf. Gyatso 2005, 281 on the first expulsion for sexual intercourse in the Theravāda
story of Sudinna, “to confront the brute fact that he performed householder activity—no matter what the mitigating
circumstances, and no matter what his particular intention or subjective state” (author’s italics).
179. See Vāset.t.ha Sutta, MN 98.27–54, 62–63, identifying the true Brahmin with the qualities of the arah-
ant (Ñāṇamoli and Bodhi [1995] 2005, 1302–3 nn. 904, 907).
164 dharma

suttas, emphasis is on their being landholders, not householders—the Buddhist

usage being not limited to Brahmin landholders (see Bailey and Mabbett 2003,
114–16). The Brahmanical householder ideal takes shape no earlier than the
dharmasūtras,180 and not fully in the earliest of them. It reaches its classical
expression in Manu, which, after proclaiming that the four āśramas “have their
origin in the householder” and “lead a Brahmin who acts in the prescribed
manner to the highest state (paramāṃ gatim),” concludes:

Amongst all of them, however, according to the dictates of Vedic

scripture,181 the householder is said to be the best (gṛhastha ucyate
śresṭhaḥ), for he supports the other three. (M 6.87–89; Olivelle trans.
2005a, 153)

While one can see this famous claim emerging through the dharmasūtras, its
formation would seem to be gradual. First, we may note its incompleteness in what
is likely the earliest dharmasūtra, that of Āpastamba, which says only that, based on
the Veda, rites requiring the wife are superior to what can be achieved by yogic
powers in other āśramas (Ā 2.23.10), and that “immortality consists in offspring”
(24.1). The other early dharmasūtras, Gautama and Baudhāyana, make a different
claim: not that the householder is the “superior” āśrama but that it is the only one,
since from it alone come offspring (G 3.3; B 2.11.27), to which Gautama adds that it
alone is prescribed in the Vedic texts (G 3.36). The Vasiṣ ṭha Dharmasūtra—closer
in time to Manu—then airs a view similar to Manu’s, saying “Of all four āśramas,
the householder is the best” (viśiṣyate).”182 Here again we get a formulation of the
idea that the AS is undermining. Yet it is only in Manu and the Rāmāyaṇa—our
two staunchest Brahmanical dharma texts—that we find the idea expressed pre-
cisely by śreṣ ṭha, as it is in the AS by seṭṭha. In the Rāmāyaṇa, this occurs where
Lakṣmaṇa implores Rāma not to go into exile but rather assume kingship and
govern: “Those who understand dharma, O dharma-knower, say that of the four
āśramas the best āśrama is that of the householder (caturṇam ̄ āśramānạ ṃ
̄ hi
gārhasthyaṃ śreṣṭham āśramam)” (Rām 2.98.58). In the Mahābhārata, Arjuna also
uses the parallel term jyeṣṭha—meaning “elder,” “senior,” “best,” and close in this
sense to “what is primary”—when he tells Yudhiṣṭhira after the war: “Birds, beasts,
and Beings are supported by householders alone, O king; so the householder has
the elder āśrama (tasmād jyeṣ ṭhāśramo gṛhī )” (Mbh 12.23.5). Both epics put these
words into the mouths of junior brothers to pass down the idea of the “bestness” of
the householder from the Brahmin to the royal Kṣatriya householder.

180. Meanwhile, see Olivelle 2004a, 502–3 on the surprisingly limited mention of dharma in the earlier
gṛhyasūtras, none of which exalt the householder.
181. As Olivelle notes (2005a, 292 n.), there is a variant reading, “according the Veda and the smṛtis.”
182. V 8.14. I follow Olivelle’s translation (2005a) here; viśiṣyate could also be translated “is distinguished.”
 early buddhism 165

Leaving discussion of the dates of these texts for chapter 5, I believe we can
take Manu to provide us with a general reference point for Brahmanical views
on householder dharma; and Manu is perhaps, at the earliest, an early Śuṅga
text, according to Witzel (2006, 482), or as late as the Kushanas according to
Olivelle (2005a, 24–25). What is most emphatic in Manu and Vasiṣ·tha, where
it is unquestionably a matter of Brahmins, is that in being “best,” the house-
holder attains what Manu calls “the highest state.” As the Vasiṣ·tha passage
quickly goes on to say, if a householder, on whom “all mendicants (bhikṣukas)183
depend,” follows all his caste rules, a “Brahmin does not fall from the world of
brahman/Brahmā (na brāhmaṇaś cyavate brahmalokāt)” (V 8.15–17). This brings
us very close to the associations Collins notes in the AS between what is “best,”
“what is primary,” and Brahmanical opinion on the primacy of brahman and/
or the god Brahmā. Indeed, Vasiṣ·tha’s opinion is that the “best” are those who,
as Brahmin householders, support “all bhikṣukas.”
Clearly, it is not necessary to argue that the AS is undermining a view dat-
able only to such later classical dharma texts as Manu and Vasiṣ·tha. The Buddha
could be portrayed as countering something like what one finds in nuce in the
earlier dharmasūtras. But if the emerging consensus on dating the dharmasūtras
is on target, the AS would not seem to be countering a pre-Mauryan view of
householder Brahmins. This is one piece of evidence, and I will note others,
that it would not be a pre-Mauryan text.

c.3.a. aggañña sutta frame and story. The AS’s frame sets the Buddha in
conversation with two young Brahmins, Vāset.t.ha and Bhāradvāja. Names like
this make them scions of two of the Ṛgveda’s Family Book clans. Rather exemplary
converts, they are, says Collins, a “good audience” for Brahmanical allusions, and
“as novices they would presumably be becoming familiar with Vinaya rules”
(1993, 319). Early in this chapter, I mentioned this pair as appearing in two other
suttas, the Vāset··tha and Tevijjā Suttas (MN 98; DN 13), as young men in an
intermediate status between student and householder that would correspond to
what Brahmanical dharma texts call the snātaka or “bath graduate”184—which
would in fact qualify them to get married and be householders. As Collins points
out, commentators portray the two as maturing from one sutta to the next: “after
the Vāset··tha Sutta they declared themselves lay followers; after the Tevijjā they did
so again, but thereafter . . . took the Minor Ordination to become Buddhist
novices. At the start of the AS they are aspiring to become monks, hoping to take
the Major Ordination,” and after the AS “they did so and attained liberation”

183. Probably covering non-Brahmanical ascetics; and note the diminutive suffix -ka, which can be dismissive.
184. See § A and n. 35 above. The Vāset.t.ha Sutta is also redacted in the Sutta Nipāta; see Collins 1993, 318;
Siddhatissa [1985] 1998, 70–76.
166 dharma

(Collins 1993, 319). Upon closer scrutiny, one finds in both the Vāset··tha (MN
98.7.1) and Tevijjā Suttas (DN 13.5, 8–9) that Vāset.t.ha’s teacher was Pokkharasāti,
a memorable character from § A of this chapter, which would presumably make
young Vāset.t.ha an erstwhile colleague of young Ambat.t.ha! The beginnings of
the Vāset··tha and Tevijjā Suttas find Vāset.t.ha and Bhāradvāja in areas where
Pokkharasāti and four other “well-known and prosperous” Brahmins are staying,
and they engage the Buddha in conversation only after they have gone strolling
and realized he is camped nearby.185 Near the start of the Vāset··tha Sutta, Vāset.t.ha
tells the Buddha that he and Bhāradvāja “have attained full mastery over all that
the Vedic experts teach; skilled in philology and grammar, we match our teachers
in recitation” (MN 7.2).186 They have thus passed the same kind of curriculum as
Ambat.t.ha.187 But whereas the Ambat··tha Sutta leaves its brash protagonist in a
kind of limbo, these two move along.
In all three of the suttas involving them, Vāset.t.ha and Bhāradvāja hear about
the merits of homelessness over the encumbrances of the householder.188 But in the
AS, now that they are “aspiring to become monks,” they are living with the monks
at Sāvatthi where the Buddha is residing. Here, when they go out strolling they see
the Blessed One doing the same and decide to approach him on the chance of
hearing a Dhamma talk (DN 27.1–2). The first word goes to the Buddha:

Monks, you were (both) born brahmins, in brahmin families,

(but) you have gone forth from home to homelessness, (leaving) your
brahmin family. Surely brahmins (must) revile and abuse you.

Right from the start, the Buddha asks what Brahmins of their own families are
saying, based on their householder ideal, when their sons are transgressing
that ideal as aspirants to the monastic life. This topic at the near side of the AS’s
frame also has an elegant closure when the Buddha ends the sutta invoking the
Ever-Virgin Brahmā—a youthful approved ascetic counterpart to the “Creator”
by that name—as a counter-Vedic model of what is “best” for his two young
interlocutors.189

185. Another is Bhāradvāja’s guru Tārukkha, and the others are Caṅki, Jāṇussoṇi, and Todeyya (MN 98.3;
DN 13.2).
186. Of the two, Vāset.t.ha is always the chief speaker and interlocutor, and in these two other suttas where he
and Bhāradvāja appear in debate, Vāset.t.ha is the one whose view draws the Buddha’s main comment and, in the
first case, his basic agreement.
187. Cf. Collins 1993, 318: “adept in the three vedas . . . philologists, grammarians, like our teachers in
(vedic) recitation.”
188. Fittingly, if we take the commentators’ indications of a continuous development, this theme prog-
resses. Lightly broached in the Vāset.t.ha Sutta (see MN 11.35, 46), it is a major point in the Tevijjā, where “union
with Brahmā” hinges on whether Brahmā is “encumbered” by marriage like Brahmins and unlike Buddhist
monks (DN 13.31–34), Vāset.t.ha says he is unencumbered, not having heard yet of his marrying Sarasvatī.
189. Collins 1993, 377–79. In the Ambat.t.ha Sutta, the same words occur not at the end but fittingly in the
middle of the conversation with Ambat.t.ha, and not preceded by the contrast with the Vedas; see above, n. 41.
 early buddhism 167

Working further into the AS, my purpose will be to trace the interplay in it
between Vinaya allusions and Brahmanical (principally Dharmaśūtra) ones.
I will take it that the allusions are being made to two spheres of dharma that are
under contestation not only with regard to social class and kingship, as others
have recognized, but also with regard to this “housing problem.” I begin on two
points were I agree with both Collins and Gombrich: (a) that the AS’s
conversation about social class is based on what Vāset.t.ha reports to be the con-
temporary Brahmanical understanding (however indirect it may be through
both Buddhist and Brahmanical filters) of the Puruṣa Sūkta’s verse on the ori-
gins of caste (ṚV 10.90.11);190 and (b) that the Buddha is represented as
combining seriousness with humor.191 What Vāset.t.ha reports will in fact be no
surprise to us. It is pretty much what his erstwhile classmate Ambat.t.ha repre-
sented as the views of their teacher Pokkharasāti:192

The brahmin is the best class (vaṇṇa). . . . Brahmins are Brahmā’s

own sons, born from his mouth, born of Brahmā. . . . You have left
the best class and gone (over) to an inferior class, since you have
become wretched shaven-headed (pseudo-)ascetics (samaṇakas),
members of some sect, (no better than) offspring of our Kinsman’s
[i.e. Brahmā’s] feet. It . . . is unseemly, that you have left the best
class. (Collins 1993, 339; AS [DN 27] 3)

The class analysis is familiar from the Ambat··tha Sutta, as also its bearing on
Khattiya and Brahmin rankings;193 but the AS now also addresses the matter
of leaving this “best” of birthrights, and thus raises the matter of leaving
home. Hearing this, immediately the Buddha kicks in with the first signs of
humor:

Surely, monks, the brahmins are not recalling the past when they
say [this]. Brahmin women, (the wives) of brahmins, are seen to
menstruate, become pregnant, give birth and suck; and (so) these
brahmins who say: “the brahmin is the best class . . . brahmins are

190. See Gombrich 1992a, 163–64, 167; Collins 1993, 349–50. Gombrich 1992a, 165 also thinks the
transition from frame to “parable” entails a continuing parody of different Ṛgvedic hymns, from ṚV 10.90 to ṚV
10.129, which I consider less likely, and also holds that the Buddha follows this up with “garbled” parodies of
passages from the Bṛhadāraṇyaka Upaniṣad (169 and nn.). Best just to rely on the ṚV 10.90 allusion which is
virtually undeniable and, as Collins and Gombrich note, is even recognized in a Buddhist subcommentary.
191. Which Gombrich tends to describe as satire, parody, and debunking; see Gombrich 1988, 85: “a
debunking job”; 1992a, 162 and passim; 1992b. Collins, agreeing with Gombrich’s treatment of Brahmanical
references, suggests that the “AS was composed in and for an educated milieu familiar with both styles of
thought, one that could smile at its wit, and appreciate its serious intention” (1993, 318; see further 313–16).
192. Recall Ambat.t.ha Sutta 2.5, as sited above in § A of this chapter.
193. Cf. Walshe [1987] 1995, 604 n. 816. Collins takes the -ka suffix in samaṇaka to give samaṇa/śramaṇa
a “pejorative” sense (1993, 350 n.).
168 dharma

born from Brahmā’s mouth . . . heirs of Brahmā,” are (in fact)

born from vaginas. They are slandering Brahmā, telling lies, and
producing demerit. (Collins 1993, 339; AS [DN 27] 4)

Collins makes the noteworthy point that the Buddha’s remark that Brahmins
misremember the past (porāṇaṃ) “is unexpected,” since Brahmins misre-
member not only their cosmogonic origins but their physical birth from
yonis (which could also be rendered “wombs” or “uteruses”). The “slander”
in question, as a commentarial exegesis recognizes, thus implies that
Brahmins equate Brahmā’s mouth with Brahmin women’s sexual organs.194
The Buddha will continue in the AS, as in other suttas,195 to speak of what
one might call Brahmins’ errant and selective memories (a likely slap at the
single-mindedness of Vedic oral tradition).
Soon, nearing the end of the frame, the Buddha brings these matters back
to the housing problem:

You, monks, are from various castes,196 of various names, from

various clans and various families, and (yet) you have gone from
home to homelessness. When asked who you are, acknowledge that
“we are sons of the Sakyan.” (Collins 1993, 341; AS [DN 27] 9)

The Buddha gives those who go from home to homelessness a new birth and
family,197 if not a home.198 Vāset.t.ha and Bhāradvāja are now staying at a “palatial
monastery . . . outside Sāvatthi” (AS [DN 27] 1.1). This implication of a palatial
monastic life could be another indication that the AS is not pre-Mauryan.199
Moving on from the frame into the parable, the Buddha begins his account
of what has been called “the Buddhist Genesis,” whose implications for relating
Buddhist ideas of dharma and time we shall recall in chapter 6. He starts with

194. See Collins 1993, 350–52, citing Siegel 1987, 207 for his “uninhibited” rendering.
195. See his metaphor of the “file of the blind,” as discussed in § A above.
196. This seems to suggest that Vāset.t.ha and Bhāradvāja have joined the Buddha strolling with an entou-
rage, and indeed, that the Buddha has seized upon their joining the group to make the points he is making.
197. Freiberger 2000, 224. Collins takes the compound Sakyaputtiyā here as “‘Sons of the Sakyan’, in the
sense of the monks of Gotama’s order,” but Freiberger doubts that meaning and says that in being preceded by
Samaṇa,̄ the question should be answered, “We are ascetics who belong to the son of the Sakyas” (“Wir sind
Asketen, die zum Sakyasohn gehören”), that is, who belong to the Buddha. For him, the term is a sociological
identity marker. Yet the passage goes on to use the additional phrase Bhagavato putto, rendered by Collins as “the
Blessed One’s own son,” which according to Freiberger does convey a father–son relationship through the spe-
cial quality of trust in the Buddha and his dharma (225), from which they “are born through his mouth” (225).
Freiberger views the first as an early formulation but thinks the second offers no help in dating (2008 and
personal communication).
198. See Gyatso 2005, 287: “The vinaya rules are the blueprint for the functioning of special kind of
community, a denatured home for the homeles. . . .”
199. Walshe [1987] 1995, 603 reassures us that Visākhā’s “‘mansion’ was a comparatively splendid struc-
ture, though still small to the modern way of thinking.”
 early buddhism 169

what “usually”200 happens when the world contracts and “beings devolve as far
as the Ābhassara world,” where “they remain for a long time, made of mind,”
and further down to where they “die from their Ābhassara bodies” and resume
coming back to this world, where “they remain for a long time,” still “made of
mind” (Collins 1993, 341; AS [DN 27] 10). Following Collins, the AS now begins
to pepper its continued evolution with allusions to Vinaya, especially to the
Pātimokkha. The first such allusions are to a world that would appear to antici-
pate perfect conditions for the sustenance of “what is primary,”201 that is, mind-
made monastic life. While there is still “nothing but water,” an “earth essence”
spreads out on the water that has color and taste like ghee (sappi), cream
(navanīta), and honey (madhu)—three of the five kinds of medicine (bhesajja),
along with oil and molasses, that a Pātimokkha rule (Nissaggiya Pācittiya 23)
allows monks and nuns to store, unlike other foodstuffs, for up to a week but
no longer.202 So far the evolutionary process cannot have yielded medicine,
since beings are still disembodied; rather, the filmy appearance and flavor are
a temptation: “Then, monks, a certain being, greedy by nature, . . . tasted the
earth-essence with his finger,” pleasing himself and producing craving. Other
beings did the same, and then moved on to taking big mouthfuls of it with their
hands. Their mental luminosity disappeared while the sun, moon, and stars
appeared along with the day, night, and seasons (Collins 1993, 342; AS [DN 27] 12).
As Collins shows, tasting food with one’s finger and eating large mouthfuls
with the hands contravene a number of Sekkhiya or Training rules (316;
359–60). This turn in Pātimokkha allusions from primal temptation to training
rules makes narrative sense when it is recalled that two novice monks are the
Buddha’s chief listeners.
This finger-dipping is the first decisive individual act. Through each stage
in this “Fall,” it is always individuals whose acts lead to a general transforma-
tion. The individualism and voluntarism of Vinaya Law have been recognized
as contrastive features over against Dharmaśāstra;203 we may relate them to
other similarly contrastive features of Vinaya that now come into play in the
AS’s account of ongoing social evolution: Vinaya law’s ethics of intention, with
its emphasis on choice, deliberateness, consent/refusal, and consensus (see
Gyatso 2005, 282–88); its recognition that acts have consequences, not only for
the acting individual but social consequences both inside and outside a group

200. Collins 1993, 341, 357 n. finds the words “suggest . . . a studied vagueness about the cosmogony here.”
201. Collins calls the description of ghee, cream, and honey the first of the “verbal reminiscences of the
Monastic Code (Vinaya)” in the AS (1993, 326).
202. Collins 1993, 341–42; AS (DN 27) 11. If they are kept and eaten longer and not discarded, it is a nis-
saggiya pācittiya requiring expiation and forfeiture (326, 329, 358). A monk or nun can eat them medicinally after
noon (368).
203. Gombrich 1988, 11–16 posits a “methodological individualism” lending itself to Vinaya formation.
170 dharma

of actors; and especially that when making rules (not to mention laws) there
will be what Gombrich likes to call “unintended consequences”204 that will lead
to the need for further rules, clarifications, and amendments.
At this point we may ask how the AS brings its account of this continuous
Fall into what Collins calls its “central figure” and “crucial moment.” For
Collins this is the Vinaya-related “motif of ‘making a store.’ ”205 For Gombrich,
we may perhaps locate it at the point that gets his greatest attention: the implic-
itly Brahmanical etymologies by which the Buddha explains the four social
classes. My point will not be that a third moment should trump these two, but
that the housing moment brings the Vinaya allusions and the Brahmanical
allusions into focus together.
There is now a stretch (AS [DN 27] 13–15) where Collins finds no Vinaya
allusions, although their reminiscence surely holds through it. The more earth
essence beings ate, the harder became their bodies. Good and bad looks arose,
and pride and arrogance about them.206 The pride and arrogance made the earth
essence disappear. “When it had disappeared, they came together and lamented,”
uttering aho rasam, “Alas, the essence!”—which people now say when they have
“tasted something good” and mean “Ah, the taste!’ They recall the original, pri-
mary word(s), but they don’t understand what they mean” (13). This is the first
of three etymologies that the Buddha offers to explain Brahmanical loss of
memory, here along with the first collective or concerted activity of lamenting.
Forgetting the meaning could represent the Buddha’s perception of Vedic
memorization,207 and something that must not happen with the Dhamma and
Vinaya, especially in the recitation of the Pātimokkha, which concerns memory
about meaning.208 It is the first of three points where the Buddha explains how
such a loss of meaning occurred. As etymologies, each ties in with those he will
offer in explanation of the names of the four social classes; and all of them go
back to his “unexpected” remark in the frame about Brahmins’ errant mem-
ories when it comes to being born from Brahmā’s mouth.

204. Gombrich 1988 (see Index, p. 237) uses this phrase with cumulative effect, most often with reference
to Vinaya.
205. Collins 1993, 305; cf. 327–28 on Vinaya overtones of the “five ‘impossible things’ which an enlight-
ened monk cannot commit”—one of which is storing food. Collins speaks in 1993 only of “the central figure”
around which Vinaya allusions cluster. But later, he clarifies this as a “crucial moment”: “the text chooses the
moment when one of the beings stores food as a crucial moment in the evolution of society” (1998, 450). Actually,
can there be a “crucial moment” in a continuous Fall? Cf. 329–30: “Each and every event in the degeneration of
beings is in some way related to the monastic order, its ideals and its Code.”
206. Collins relates these vices to “the class-pride of brahmins which began the whole narrative” (1993,
361), and suggests that Vāset.t.ha and Bhāradvāja have “abandoned” them in seeking to become monks (330).
207. See Collins 1993, 362, “the point being that although in general brahmins do not remember the true
account of ‘cosmogony’, when they say this phrase they do recall one small part of it (without understanding).”
208. See Brough 1962. Ānanda does not bother over hearing a verse recited in which the words have
changed but the meaning seems to have remained more or less the same.
 early buddhism 171

When the earth essence disappeared, a “fragrant earth” then appeared like
a mushroom, which again had a color and taste like ghee, cream, and honey.
But after beings had lived on it for a long time, further hardening their bodies
and developing pride and arrogance in their increasing good looks and despising
those who were increasingly ugly, the fragrant earth likewise disappeared and
“a (kind) of creeper appeared” (14). This no longer had the “earth essence” and
the “fragrant earth’s” look and taste, which had motivated the first decisive act
of “finger-dipping” and the first “craving.” Beings started to eat the creeper,
which after a long time disappeared under the same conditions as the earth
essence and the fragrant earth. “When it had disappeared, they came together
and lamented ‘we’ve had it, the creeper has given out on us!’” And again,
though people nowadays use the phrase when “touched by some hardship,”
they forgot the meaning (15; Collins 1993, 342–43).
That is the second of the three etymologies that the Buddha offers to
explain a “primary” loss of memory, and again it comes with the collective or
concerted activity of lamenting. The AS has now been triple-telling in two
converging ways.209 Having come through two instances of concerted activity
and loss of memory, and having perhaps caught the deepening Vinaya and
Brahmanical reminiscences through the three cycles of body- and vaṇṇa-
differentiation now completed, a listener would be within rights to suspect that
the Buddha has been building up to a breaking point. Like Vāset.t.ha and
Bhāradvāja, such a listener could ask: What could possibly happen next after
three rounds of increasing pride and arrogance about the beauty and ugliness
of solidifying bodies?
This brings us to the stretch where Collins and I each find our central
moments, neither of which is the first sex act but both of which follow from it.
With the creeper gone, rice appeared without cultivation; “harvesting was
unknown.” Living off rice for a long time, more good and bad looks came about,
bodies became even solider, and “the female parts appeared in a woman, and
the male parts in a man;210 the woman looked at the man with intense longing,
as did the man at the woman.” An “intense, excessive longing” led to “burning
passion” and “burning on their bodies; because of this burning they had sex.”
For the first time the individual act that changes things takes two individuals.

209. See Gombrich 1992a, 171, observing that the same cycle of increasing body solidifications and vaṇṇa
differentiations has now occurred three times “for no apparent reason”—for which he tries, I think implausibly,
to explain the foods as allusions to Upaniṣadic cosmogony and Vedic ritual (165, 172, 178 nn. 39–41), while dis-
missing the more solid point that, “True, the Buddhist texts tend to say things three times. . . .” (171).
210. According to Buddhaghosa’s rendition in the Visuddhimagga (418 = 12.50–51), the body orifices were
opened by urine and excrement that resulted from digesting rice; see Collins 1993, 365, remarking, “presumably
the previous foodstuffs were too ethereal for such a result.”
172 dharma

For once the Buddha does not say it took a long time. As Collins shows, what is
being described would reach (aspiring) monastic ears as a reminder of Pārājika
and Pātimokkha rule Number 1: no sex.211

When the other beings saw them having sex, some threw earth (at
them), some threw ashes, others cow dung, (saying) “Away with you
and your impurity, away with you and your impurity!” “How could a
being do such a thing to another being?” So nowadays, people in
certain areas, when a bride is being led out, throw dirt, ash or
cow-dung. They recall the original, primary (actions), but they do not
understand what they mean. Monks, what was thought adhamma at
that time is nowadays thought [to be] dhamma. At that time the
beings who took to having sex were prevented from entering either
small or large settlements for a month or two. Accordingly at that
period of time those who indulged excessively in that which is not the
true dhamma (asaddhamme) took to building houses to conceal it.
(Collins 1993, 344, slightly modified)

Then a being who was “lazy by nature” started storing rice for a day for both eve-
ning and morning, and others imitated him; each time someone added another
day of storing until someone did it for eight days, thus contravening the Vinaya
limit of seven. After this, grain no longer regenerated; it required harvesting;
and “rice stood in a clump.”
The first sex act thus leads to houses,212 living in houses leads to food
storage, and food storage leads to the depletion of miracle-grow rice and the
necessity of human labor. As Collins points out, sexual differentiation has
overtaken vaṇṇa pride in defining stages in the Fall (365), but vaṇṇa
differentiation has continued to be reinforced by good and bad looks and will
next be furthered by division of labor, and each vaṇṇa will presumably want
its own houses. To put it simply, making houses accounts for the first house-
holders and explains the bifurcation of the “original saṅgha” into two model
communities. There are those who, as “what is primary,” have now set prece-
dent for the saṅgha that will recall such “primary” ways when a Buddha
appears and rules that sex, storing food for over seven days, and taking meals
after noon are breaches of Vinaya. And there is now a new secondary
community, proto-Brahmanical we could say, anchored in the ways of house-
holders who store food for as long as they want, who eat breakfasts and

211. See Gyatso 2005, 280: “It is much more plausible that what really made sex with a woman worse than
any other kind was its practical upshot: marriage, children, the householders life. . . .”
212. Collins 1993, 326: agārāni . . . kātuṃ; 368: “given the references to houses (plural) here, one must
assume that . . ., again by narrative ellipsis, the practice has spread to others.”
 early buddhism 173

dinners, and who will continue the process of vaṇṇa differentiation that has
been under way long before this sexual revolution in the form of good and
ugly “appearances.”
Collins is attentive to this housing moment: “Making houses . . . contra-
dicts the fundamental symbol of monastic life, ‘going forth from home to
homelessness’ (agarasma anagāryaṃ pabbajā)” (1993, 326). Indeed, his first
book, Selfless Persons, offers a pertinent section on Buddhist “House imagery”
(1982, 165–76), with a subsection on “Leaving home for homelessness”
(167–71) that includes a summary of our sequence:

The connection of ideas is expressed in the comic evolution myth of

the Aggañña Sutta, which satirizes Brahmanical cosmology. When
first the characteristics distinguishing male and female appeared,
sensual passion for each others’ bodies arose in some beings, and
because of this [when?] they began to indulge in sexual intercourse,
other beings, disgusted, stoned them and forced them to build
houses to conceal their immorality. (Collins 1982, 169)

Although “stoned them” is not quite accurate, and also less funny (they only
threw dirt, ash, and cow dung), Collins had already found the AS humorous.
Having not yet discovered its Vinaya allusions, he treated the housing
problem as central. But having since discovered the Vinaya allusions, he
gives the housing moment a Buddhist “symbolic” explanation only, without
noticing that society has evolved the leading institution of dharmaśāstric
Brahmanism.

c.3.b. sly etymologies. Inevitably, the devil is in the details, and I limit
myself to two sets of them: the third etymology connected with concerted
activity and loss of memory, now just behind us; and two of the allegedly
humorous explanations of the naming of Brahmins.
When the Buddha gives this third etymology about forgetting, there is, as
Gombrich and Collins notice, a difference from the other two. In all three he
uses the term akkhara (= Sanskrit akṣara) to recall what was original and is
now forgotten, but this third time he does so to describe actions rather than
words. Thus when people “nowadays . . . throw dirt, ash or cow-dung” at a
bride, we find a parenthesis in Collins’s translation: “They recall the original,
primary (actions), but they do not understand what they mean.” Gombrich
sees this anomaly in the way this third usage etymologizes a no longer under-
stood “custom” rather than a no longer understood sacred “eternal” word
(a possible meaning of Sanskrit akṣara) (1992a, 171). Collins finds Gombrich’s
idea that the Buddha is “deliberately parodying” a Brahmanical view of the
174 dharma

eternality of language “not . . . decidable” (1993, 336 n. 29); but he thinks

Gombrich “must be right” that “akkhara used for actions” would come from
“‘the levelling process typical of oral transmission’, and/or as having ousted
some other word” (367). I agree with Collins that the point about the eternality
of Vedic language is undecidable,213 but I doubt that this third usage of akkhara
is secondary.
In each of the three etymologies connected with loss of memory, there is
both something said and something done. In the first two the Buddha focusses
on the forgotten meaning in what is said, but in each case there is a lamenta-
tion, which I signalled by calling it a “collective or concerted activity.” In the
third case the focus may shift to what is done, but the action whose meaning is
forgotten was accompanied by words: “Away with you and your impurity,214 . . . !”
The important point is the etymology (or etiology) of the “custom” of marriage.
What the Buddha calls “adhamma” has turned into marriage, with the same
action and words that initially disdained sex now continued only as a ritual, and
again, with no one remembering what it was originally about. Marriage is the
first concerted activity unaccompanied by lamentation. Clearly, the rite that
makes householders is founded on something that was originally shunned,
and bifurcates two model communities: one that will see householding as what
is “best” since it produces progeny, and one that will restore “what is primary”
by ruling sex to be its first grounds for expulsion.
The AS continues its account of the evolution of more “bad, unwholesome
things”215 such as “stealing, accusation, lying, punishment and banishment,”
and tells how Mahāsammata became the first king, which alleviates some “bad
things” and compounds others (AS [DN] 27.21–22; Collins 1993, 345–46).
Collins is able to show more references to Pātimokkha rules during this span,
and Gombrich to Brahmanical ideas, particularly in the Buddha’s eight etymol-
ogies (niruktis) for the four social classes. Regarding Collins’s contribution, we
must note his decisive discussion of the very naming of Mahāsammata as a

213. Cf. Veidlinger 2006, 73 on “oral” and “written” vagaries of Pāli “akkhara.”
214. On this repeated nassa asucī, see Collins 1993, 366–67 on a possible Brahmanical nuance (following
a suggestion of Phyllis Granoff) were it to mean “depart impurity” at a wedding where these things “would be
thrown” (no citation is supplied), and a specialized Vinaya usage to mean “ ‘expel’ from the saṅgha.” Cf.
von Hinüber 1995, 36–37 on the Buddha’s use of nāseti in expelling the nun Mettiyā for falsely accusing a monk
of raping her.
215. The “primary” community’s mode of thinking connotes additionally an Abhidhammic sense of
“wholesome” and “unwholesome” as applied to dhammas. In the AS, the beings begin to think repeatedly in the
terms “bad, unwholesome things” (pāpakā akusalā dhammā) right after the rice required harvesting (Collins
1993, 344–45; cf. 331, 345–46). Such an Abhidhammic outlook is also foreshadowed in the AS’s frame, where it
mentions what have been called the “Buddhist Ten Commandments” (Hopkins 1923b; Huxley 1996a, 412) and
speaks of murder, theft, sexual misbehavior, lying, malicious speech, harsh speech, frivolous speech, envy,
malevolence, and holding wrong views as “those things which are unwholesome, blameworthy, not to be fol-
lowed,” in contrast with “those things which are wholesome” (Collins 1993, 339–41).
 early buddhism 175

Vinaya allusion, with “the Great Appointee” evoking Vinaya “monastic organization
and self-government” by “consent,” and with suggestions of a consensual theory
behind kingship (1993, 382–84).
But Collins is framing this discussion around his view that the AS is
pre-Mauryan. Its theory of kingship is also, I believe, a critique of post-Mauryan
Brahmanical notions of (partially) divine kings.216 It may also offer a compro-
mise to the tensions aired in the Mahāparinibbāna Sutta between the Vinaya
overtones of gaṇasaṅgha polities and the imperial monarchic state (see Collins
1998, 436–45), which would mean it reflected the latter’s emergence.
I thus return to the question of dating the AS. While my current argument
has been that the housing moment is pivotal, my underlying argument has been
that this focus must introduce caution in dating the AS to pre-Mauryan times. So
far this caution hinges mainly on how one periodizes the dharmasūtras, whose
dates, I recognize, are not set in stone. I have mentioned some other consider-
ations of this type favoring my position, and will now consider two more. On this,
I turn to Gombrich and Collins’s discussions of the AS’s etymologies of the four
social classes. While leading up to a description without etymology of Samaṇas
“leaving home for homelessness,” the Buddha gives eight etymologies for the four
varṇas, beginning with three for Brahmins, the second and third of which are the
ones of interest. The second includes what Gombrich considers to be an allusion
to the Baudhāyana Dharmasūtra, and the third a seeming reference to books.

The second etymology for Brahmins puns on words linking

fire-tending and meditation, and offers this characterization of what
Gombrich calls “the original good brahmins”: They made leaf-huts
in the forest and meditated in them; without coals or smoke (from
a cooking fire), pestle set down, they went into villages, towns and royal
cities in search of food, in the evening for their evening meal, and in
the morning for their morning meal. (AS [DN] 27.22)217

Gombrich finds an “allusion to brahminical literature” here. Baudhāyana,

he says, “prescribes the way of life of a brahmin ascetic who has renounced
the householder’s life”; as he translates it:

216. See chapters 3 and 5. The AS denies kings any intrinsic authority as divine (e.g., from God’s arms), or
as made so by Brahmins who consecrate them in a Rājasūya as, among other divinities, Dharma. They are
appointed to a function. On “ironical and satirical comment on kings and society” in other suttas (Mahāsudassana,
Cakkavatti Sı̄hanāda) with which I believe the AS is probably contemporary, see Gethin 2005, 85–86, discussing
Collins 1998, 476–96.
217. Collins trans. 1993, 346, for the sake of consistency, but dropping a parenthetic “(and mortar)” after
“pestle” as a distraction, since the pestle (musala) is what counts. Cf. Gombrich 1992a, 173. The AS has vītaṅgara
vitadhūmā paṇṇamusalā, which Gombrich plausibly corrects to sannamusalā. Collins 1993, 372 takes the emen-
dation as one of several possible variants that in the corresponding dharma rule can be read “when the pestle has
been set down.”
176 dharma

A wandering renunciate (parivrājakaḥ) should leave his family and

go forth without possessions according to rule. Going to the forest
with his head shaven except for the topknot, wearing a loincloth,
staying in one place during the rains, with a yellow-stained outer
garment, he should beg food when the pestle has been laid down,
there are no live embers, and the collecting of the plates is over
(sannamusale vyaṅgāre nivṛttaśarāvasampāte bhikṣeta). (B 2.11.16–22;
Gombrich trans. 1992a, 172)

Baudhāyana goes on to speak of nonviolence (ahiṃsā), carrying a water strainer,

and taking a middle course (madhyamaṃ padam) (23–26). As Gombrich notices,
the latter “sounds like an allusion to the Buddhists, even if the passage as a whole
may be giving a more composite picture” (1992a, 173). Both texts thus mention
the laid-down pestle and the cooled embers, and Baudhāyana adds being finished
with “plates.” From other examples, we can posit that each is probably nuancing
a saying such as Gombrich notices “an extra echo” of in Manu:

A renunciant (yati) should always go begging when the pestle has

been laid down (sannamusale), there is no smoke or live embers,
people have finished eating and the plates have been collected.
(M 6.56, Gombrich trans. 1992a, 173)

Manu thus mentions four signs, adding “no smoke.” Not noted by Gombrich,
Vasiṣṭha Dharmasūtra 10.7–8 also has a short variant. The Mahābhārata also
invokes such a saying in its twelfth (12.9.32) and thirteenth Books (13.129.53),
which draw frequently on dharmaśāstra. We shall return to this adage, its vari-
ants, and one of the Mahābhārata’s twists on it in chapter 12. What is striking
for now is that there is no such saying in the earliest dharmasūtras of Āpastamba
or Gautama. This would suggest that the AS shares this formulation with, and
alludes to, texts that are probably no earlier than the second century BCE.218
The Buddha is thus evoking a dharmaśāstra rule for which there are no early
examples. Collins and Gombrich have a looser idea of Baudhāyana’s dating
than I am working with,219 and Gombrich invokes oral tradition to keep the AS
within reach of a pre-Mauryan dating (1992a, 173–74). But the argument looks
strained. We must not allow oral tradition to be a magic wand.
Like the second etymology for Brahmins, the third also raises questions
about orality, but now with a reference to books. Or at least so it seems to me.
Here is the third etymology of Brahmins:

218. Collins 1993, 372 cites two additional passages from the Purāṇas.
219. See Collins 1993, 307 on the “later Vedic texts produced by Brahmins, the dharma-sūtras”; cf.
Gombrich 1988, 55; 1992a, 173–74, pulling the BDhS passage discussed above into a late Vedic time frame.
 early buddhism 177

Some of these beings, monks, were unable to maintain (the life of )

meditation in forest leaf-huts; they went to the outskirts of villages
and towns and lived there making (up) texts (ganthe karontā). (Other)
humans saw them and said “These beings are unable to maintain
(the life of ) meditation . . . they live there making (up) texts. They
do not meditate.” “They do not meditate” (na . . . jjhāyanti), monks
(is what Students (of the Veda) (ajjhāyakā) (means); “ajjhāyaka” was
the third term (for the brahmin class) to appear. (AS [DN] 27.23;
Collins 1993, 346; cf. 373)

As Collins says, this “pun” on ajjhāyaka (= Sanskrit adhyāyaka, “reciter”) is a

“witticism everyone has seen and agrees on” (1993, 374). It is Gombrich’s
prime example of AS humor (1988, 85; 1992a, 163), and, as he observes, it
takes the previous etymology “to set up the joke” (1992a, 173). I do not think,
however, that Collins’s translation of ganthe by “texts” gives the joke its like-
liest target. For one thing, his parenthesis around “(up)” in “making (up) texts”
seems to make up a new joke that is not really there, suggesting fiction. As
Collins notes, Rhys Davids’ translation of ganthe karontā is simply “make
books” (1993, 373). Nothing in the verb suggests that books or texts would
have been made “up.” Collins calls “books” “an anachronism,” and elsewhere
he includes gantha among a number of Pāli words, including pāli itself
(“meaning firstly a line, bridge, or causeway, and thence a ‘text’”), that carry
metaphoric meanings of “text” that are used originally for “oral ‘texts.’”
Gombrich too (1990, 27) criticizes Schopen’s discussion of this passage as
“rare” evidence in Pāli for “books,” and, like Collins (1998, 227), suggests that
Schopen would have done better to speak of a “cult of the text” than a “cult of
the book.” Gombrich takes gantha to mean “knot” here, calling attention to the
phrase manthe ganthetvā in the Sutta Nipāta, Brāhmaṇadhammika Sutta 19,
where it means “‘knot together’ mantras—and the reference is to their
composing Vedic texts” (1990, 27).220 But this verbal usage is easier to take as
“knot together” than the nominal one of gantha with “make” is, which is easier
to take as “book.”221
Now it is widely acknowledged that writing in classical India cannot be
shown to be much if any earlier than the Aśokan edicts. It would not in any case

220. Cf. Collins 1993, 337 n. 36. As Collins summarizes the Brāhmaṇadhammika Sutta, the original
Brahmans “began to covet, inter alia, the wealth and ‘excellent women’ of kings, and so began composing hymns
to acquire them” (1993, 320).
221. Unfortunately, Schopen confines discussion to a footnote, citing this usage only as the first example
of “rare” references “to books” in “canonical Pāli” (1975, 171 n. 46) and not revisiting the matter in the article’s
republication (2005, 59 n. 46).
178 dharma

serve anyone’s argument to say that a text mentioning “books” is pre-Mauryan.

So a lot is at stake in maintaining that the AS is speaking only of “knots.”
Indeed, considerable scholarly energy has gone into keeping the Sutta Pit·aka
oral. Ever since Rhys Davids and Oldenberg, however, it has been recognized
that the Vinaya Pit·aka contains numerous references to writing.222 Vinaya
orality has thus been treated more selectively,223 most notably with reference to
the Pātimokkha. It is of course a leap to suggest that familiarity with writing in
the Vinaya may be pertinent to an uncertain reference to “books” in the AS.
Laced as it is with Vinaya allusions, the AS could still reflect a stage of Vinaya
development older than the Vinaya’s references to writing. Nonetheless,
everyone agrees that the AS is describing the Vedas here, whether as books,
texts, or knots. It is probably a better joke if it is saying that the “Reciters” who
“do not meditate” had started making these worshipful texts into books.
This matter of humor in the AS has drawn some fire. Leaving aside whether
the Buddha himself had the wit that the AS attributes to him, the AS is cer-
tainly humorous, as can be said of some other suttas. The humor is important.
But why? Gethin has been critical of rationalist assumptions behind Gombrich’s
attributions of humor where he finds the narrative implausible (1997, 216;
2006 65–66), while Andrew Huxley remarks on the uncertainties of attrib-
uting humor to a fifth-century BCE text. Yet both are also congenial to the topic
within limits. Crediting Collins’s eye for humor and noting that its role in the
AS was opened up over a century ago by Rhys Davids,224 Gethin remarks that
humor is one of the ways that a Buddhist sutta “has the power to move and
arouse—certainly in its ancient listeners—religious emotions in the manner so
well brought out by Steven Collins in his discussion of the Vessantara Jātaka.”225
We are not far here from acknowledging the place of humor in a palate of
emotions or moods (rasas) such as is theorized in classical Indian aesthetics,
and indeed mentioned in the Rāmāyaṇa (1.4.8), which has so many affinities
with the Vessantara Jātaka, as Gombrich appreciated (1985). Meanwhile,
according to Huxley, “What we really need is an analysis of the N. E. Indian
sense of humour in the 5th century BC graded on a scale from buffoonery to
subtle irony, with appendices showing regional variations among Sakyans,
Magadhans, male adult celibates and enlightened beings. . . .” (1996a, 412).

222. See Rhys Davids and Oldenberg [1881] 1968, xxxii–xxxvi on the “indisputable” evidence for “the
existence of the art of writing at the time when the Vinaya texts were put into their present shape,” “and the use of
written books”—but not for the sacred texts; and with no one owning books, manuscripts, inks, pens, etc. There are
also rules on oral recital and making a quorum for it. For discussion, see Gombrich 1990, 27–28.
223. See however Gombrich 1990, 21, 25.
224. See Rhys Davids 1899, 105–7; Gethin 2006, 65, 85–86, 102–3.
225. Gethin 2006, 102–3, citing Collins 1998, 497–554 on this jātaka. The sutta in question for Gethin is
the Mahāsudassana, in which he finds humor in the poignant portrayal of the king’s wives (2006, 101–2).
 early buddhism 179

I have tried to improve the contextualization by moving three centuries forward

to an intertextual situation that brings out more clearly what is especially
notable about the AS: that it is funny about the Law—by which I mean not only
Vinaya but Dharmaśāstra, and about both in tandem and in juxtaposition.
On such a note, I close with a quote from Huxley, who, along with his con-
genial reservations on AS humor, tells a good joke about it:

I picture the Buddha expounding #18 to #20 [the sections on

transformations in society] absolutely deadpan. Vāset.t.ha and
Bhāradvāja are walking a step behind him nodding their heads and
trying to remember it all. “Wow!” mutters Bhāradvāja, sotto voce,
“the Śākyamuni’s really getting stuck into the human sciences
tonight!” And then the Buddha, equally deadpan, moves on to #21
to #25 [on the origins of kingship and etymologies on caste] and
knocks them dead. As the howls of laughter die down and as the
audience pick themselves off the ground, Vāset.t.ha asks “By the
way, Blessed One, were you serious about that Mahāsammata
stuff?” To which the Buddha replies. . . . (1996a, 412).

For my money, however, the Śākyamuni got stuck into the human sciences
from his first words in the frame, and got his first howls just before Huxley’s
segment when, after saying, “How could a being do such a thing to another
being?” he followed this with “houses to conceal it.”
This page intentionally left blank
5
Post-Vedic Brahmanical
Dharma

Passages from the Vedas have necessarily been sifted by all attempts to
trace the history of dharma. One solid result has been the recognition,
clearest from the studies of Olivelle, that while one can trace early and
interim usages of dharma up to the Middle Vedic period, the concept
does not come to full flower until it is developed in early Buddhism, as
discussed in earlier chapters, and in the Brahmanical texts that will be
introduced in this chapter: the dharmasūtras, The Laws of Manu, and
the two Sanskrit epics: the Mahābhārata with the Bhagavad Gītā, and
the Rāmāyaṇa. We may speak of these texts somewhat loosely as
post-Vedic in that they are composed at a point where the Veda is
undergoing not only canonical closure but reclamation in new
domains and genres, including those represented by these very texts.
While we shall take further note of some other late Vedic1 and early
classical Brahmanical usages, particularly in Sanskrit grammar,
Mīmāṃsā ritual theory, Arthaśāstra political theory, and also in
dramaturgy, aesthetics, erotics, and even (as we shall see in chapter 7)
astronomy that tie in with these developments by making dharma
relevant to their main concerns, these are the main early texts to
make dharma itself their main concern. That will be the
justification for making it the charge of this chapter to take up the
question of the relation between these texts in their treatment of
dharma, for these are the texts that open up the concept for what is

1. For example, the Kaṭha Upaniṣad, which is probably post-Buddhist, treated in chapter 3.
182 dharma

to become known as Hinduism. In doing so, we will keep The Laws of Manu
[henceforth Manu] more or less at the center of this chapter, since with “Manu,”2
we get down to the thicket of it, from which we may hope to emerge onto the
more familiar (at least for this author) terrain of the epics.
In organizing matters in this fashion, we stay with the rough chronology that
Olivelle has continued to sharpen,3 though I must soon question certain points.

A. Vedic Schools and the Dharmasūtras

To probe the provenance of the dharmasūtras and Manu’s relation to them, we

must locate them not only in history but as a genre. At least in the cases of
Āpastamba and Baudhāyana, two of the three oldest dharmasūtras, they come—as
dharmasūtras—third in train, serially and also historically, within a larger genre
called the kalpasūtras, where they follow two earlier types of sūtra texts: the
śrautasūtras—“aphoristic commentaries on the texts of the solemn [śrauta] ritual of
the Brāhmaṇas,” and the gṛhyasūtras—“aphorisms concerning the domestic [gṛhya]
ritual borrowed from the solemn ritual and also the Atharvaveda.”4 The dharmasūtras
of Āpastamba and Baudhāyana are additions to kalpasūtras of Vedic schools associ-
ated with their names, while the other two, Gautama and Vasiṣṭha, are named after
ancient Vedic sages and have more tenuous associations with particular schools
(see Olivelle 2005b, 272). Kalpa derives from the root ÷kḷp, “to form, give form to,”
which also gives us the associated term saṃkalpa, the announcement of a sacrifi-
cial desire for a ritual result or “fruit.” As Biardeau puts it,

On the ritual register, kalpa refers to that which is made and which
takes form and gives a result. It implies a certain order of execution,
a method, under the form of “injunctions”—vidhi—which constitute
precisely the constraining part of the ritual and complete what is set
forth in the Vedic Brāhmaṇa texts. To look at them more closely, the
kalpasūtras are collections of positive and negative injunctions,
injunctions that are directly ritual or in rapport with conduct. They
always imply that something must produce itself given such and
such a ritual action.” (Biardeau 2002, 1: 66–67; my translation)

2. As with other texts named after their reputed authors, I will refer to the text in italics and to the reputed
author without italics, and sometimes in quotation marks, as I do here. Cf. Olivelle 2004b, xxii.
3. For a positive assessments, see Lubin 2005, 80 n. 9 (also on Witzel’s chronology for the earlier Vedic
period), 82–83 n. 12, 92 n. 35; Jamison 2006, 191.
4. Biardeau 2002, 1: 66. Biardeau’s analysis is, of course, rather textual. As we shall see, others emphasize
that the gṛhya- and dharma-sūtras are based on custom. See Jamison 2000 for a study in continuities in such texts
from ritual to legal injunction.
 post-vedic brahmanical dharma 183

That the groundsetting Brahmanical texts on dharma should frame their

discourse under a larger heading of ritual injunction is obviously momentous,
since it makes an implied correlation between dharma and ritual action or
karma that has until now been largely absent.5
The initial phase of kalpasūtra production makes these ritual and, more
precisely, injunctive overtones perfectly clear in usages of dharma. As Olivelle
has demonstrated, even while the śrautasūtras use the term only minimally,
they sharpen it precisely in this direction: “It appears that the expert scholastic
tradition on ritual developed a very special meaning of dharma that . . . refers to
the specific ritual rules or ritual details of a rite.” It could thereby be explained
how certain model rituals or “archetypes” (prakṛtis) provide “dharmas, taken as
ritual details,” that can be extended from the archetypes to other rites (vikṛtis)
that follow their model on these specific ritual details (2004a, 501). As Olivelle
observes, this “specialized and technical meaning,” absent from the canonical
Vedic texts (including the Brāhmaṇas and Upaniṣads), “probably developed
within the expert tradition devoted to ritual that produced both the Śrautasūtras
and the later Mīmāṃsā texts.” Moreover, as he says, “interestingly,” this
meaning comes hand in hand with a distinctive śrautasūtra deployment of the
compound svadharma, “used with a similar meaning” when it refers to instances
where “a particular rite has its own ritual details (dharmas) specific to it and not
taken over from or extended to other rites” (502). The ground is thus set for
svadharma as “own ritual details” to be applied in later Brahmanical texts to the
personal “duties particular to” groups and individuals.6 Although the śrautasūtra
usage of svadharma is probably pre-Buddhist, and we should not in any case
expect either tradition to be reading the other on such fine points, its technical
sense as “own ritual detail” has something of the impersonal quality of the
Buddhist dharma theory, but with an entirely opposite focus on particular unex-
tendable characteristics rather than causal interdependence.
There would seem to be reasons to stress both continuity and discontinuity
from the gṛhyasūtras to the dharmasūtras. On the one hand, Olivelle thinks it
very likely that the same author would have composed both the Āpastamba-
gṛhya- and -dharma-sūtras, which would mean some continuity in what for
Olivelle is the earliest dharmasūtra (2003, 4 n. 5; 2005b, 161, 164 n. 25). On the
other hand, he hypothesizes that the dharmasūtras “constitute a special
category,” “an autonomous genre,” in breaking new ground as the first to the
use the term śāstra self-referentially (2005b, 156), and also, as will be noted, as

5. Except where it has been backread onto older texts by translators and interpreters; see chapter 3 § A.
6. Among the few gṛhyasūtra usages of dharma that Olivelle notes (2004a, 502–3), that of dhārmika for the
Vedic student (Baudhāyana Gṛhyasūtra 3.3.31) could be said to point in this direction. In chapter 9, I will suggest
that “law-abiding” is a useful translation of this usage in the epics.
184 dharma

the first to expound upon its own sources of authority. Olivelle’s point is not
that some of the Vedic schools “did not produce literature on dharma,” which
“they evidently did,” but “that the expert tradition on dharma probably did not
arise as an integral part of the ritual tradition of scholarship” (2005b, 156 n. 6).
I believe this is correct. The distinctive ways in which the śrautasūtras use the
term dharma and svadharma offer significant precedents for the dharmasūtras,
but are not immediate to their chosen tasks.
Moreover, just as texts can disguise the “lived reality of their authors” to
sustain “fictions” of Vedic “timelessness” and eternal continuity (Olivelle
2005b, 160, 170), so can the ways in which they seem discontinuous conceal
underlying continuities. It may well be by design that whereas the gṛhyasūtras
follow the śrautasūtras in tying the kalpasūtras to specific Vedic schools
(śākhās) which can be located to specific regions, such ties become less evident
in the dharmasūtras. As Timothy Lubin puts it, while calling attention to a
“contrast” with the highly developed institutional structures of early
Buddhism, “Brāhmanical institutions were diffuse and intensely localized, at
least prior to the early dharma-texts. . . . The diffusion of textual knowledge
was dependent upon teacher-to-pupil lineages and texts themselves were
treated as belonging to individual descent groups until Mauryan times at
least. Thereafter, a pan-Ārya scholastic and literary tradition began to take
shape (although the core texts continued to be treated as proprietary knowledge
of individual lineages)” (2005, 81–82). That is, while the dharmasūtras retain
more or less formal ties with Vedic śākhās, their new concern lies in creating
a pan-Ārya and thus supra-regional identity around the concept of dharma in
a way that makes it very hard if not impossible to spot their regional prove-
nance.7 Like some others (e.g., Doniger 1991, xxxv–xxxvi; Olivelle 2005b,
165), Lubin calls this a “universal” dharma (92), with Manu completing such
a “universalization” (87), and says this “new Brahmanism, called ‘dharma’
(perhaps in imitation of Buddhist usage of the term), also aimed at establish-
ing an ecumenical set of standards that could serve to coordinate the separate
traditions of the individual Vedic schools” (92). As Biardeau puts it, “For the
first time, one witnesses an attempt to put in systematic order the brahmani-
cal society and its functioning” (2002, 1: 68). But clearly it would be more
correct to call this Ārya dharma a “civilizational” dharma—and a highly flex-
ible and diversified one at that—than a “universal” one. Let me suggest for
now that it is the universal claims of the Buddhist dharma that this supple
civilizational dharma will find repeated reasons and ways to challenge.

7. See Olivelle 1999, xxvii on the question of whether Āpastamba came from the south, introduced by the
comment that “the geographical provenance of these documents is not very clear.”
 post-vedic brahmanical dharma 185

Lubin shows that still more historical information can be culled from the rela-
tion between these three strata of kalpasūtra. Why, he asks, after codifying the “high
cult” of “complex multi-fire rites” in the śrautasūtras, “might the priestly authors
have considered it necessary, in a second stage, to codify the ceremonies outside
the śrauta system as well?” He proposes a sixth- to fourth-century BCE setting
when the Brahmanical system was still poorly adapted to “the dramatic growth of
trading towns and cities” then underway (at least in eastern regions), and still
“dependent upon the stable caste society of the village” (although its textual formu-
lation may have been more stable than its implementations). Lubin discusses a
cluster of “regimens of discipline” (vratas), including Veda study for all three Ārya
classes, Veda recitation (svādhyāya), and studentship (brahmacarya) to be among
the “apparent novelties” that are sacramentalized in the gṛhyasūtras; but he regards
the “five great sacrifices” (mahāyajñas) as foremost among the “hints” or “signs
that the genre came to embody the Zeitgeist of this period” (2005, 83–84).
These five mahāyajñas are frequently cited as having Vedic and śrauta pre-
cedent “in a Brāhmaṇa,”8 but they are especially “rubricated” in the gṛhyasūtras
(Lubin 2005, 86 n. 19), where they are “inserted into the ritual cycle” to be
performed daily as “the most prominent instance of a simplified ritual format”
(88). They include (a) a food offering (bali) on the ground or in the air to Beings
(e.g., crows); (b) food hospitality to guests; (c) at least some wood as a fire-
offering to the gods while saying “Svāhā”; (d) a water offering to ancestors while
saying “Svadhā”; and (e) private Vedic recitation (svādhyāya) as the offering to
Brahman.9 Positing “that all the rewards of a pious śrautin life could be secured
through the regular performance of a few simple ‘super-sacrifices’ (mahāyajñas),”
the gṛhyasūtras mark a “doctrinal shift” (84); this “radical simplification of
Vedic duties,” which includes Vedic recitation itself as one of the five and the
performance of all five as a “marker of Ārya status,” would then be further
generalized as exemplifying Brahmanical dharma in the dharmasūtras.10
Overlapping with this daily domestic routine, a man also has three lifelong
personal debts that more or less correspond to three of the five mahāyajñas. These
are to seers (study), gods (sacrifice), and ancestors (offspring). According to Manu,
a man must fulfill these before he can leave his inheritance to his sons and retire
in equanimity (M 4.257). The three are mentioned at Taittirīya Saṃhitā 6.3.5–10
and Śatapatha Brāhmaṇa 1.7.2.1–6, with the latter adding “the debt of hospitality to

8. Śatapatha Brāhmaṇa 11.5.6.1–10 begins its description of them as “great sacrificial sessions”
(mahāsattrāṇi), implying that they would be equivalent to Soma sacrifices that last several days; see Biardeau
1976, 40–41; Tsuchida 1991, 68. Lubin 2005, 86 n. 19 also cites Taittiriya Ᾱraṇyaka 2.10.14.
9. On svādhyāya, see chapter 3 § E.
10. Ᾱ 1.12.13–1.13.1, with dānam as “giving food” to signify the offering of hospitality to men [Tsuchida 1991, 68];
G 5.3, 8; B 2.11.1–8, emphasizing svādhyāya [6–8]; 3.1.19; V 27.7, as quickly destroying sins, “even grievous sins.”
186 dharma

men, bringing the theology of debts into closer alignment with the five great sacri-
fices” (Olivelle 2005a, 277–78).11 Olivelle considers Manu to have been the first to
use “this theology of debt to defend his position that the orders of life (āśramas) are
to be followed sequentially as an individual grows old and that renunciation is
limited to old age,” though, as he observes, the theology of debt is also “alluded to
in the ‘Mahābhārata’ ” (2008, xxi). Olivelle’s position is that Manu would be earlier
than the Mahābhārata, or at least than this usage in the Mahābhārata. But, more
important on this specific point, he says that Manu would have been the first to use
“the theology of debts (ṛnạ ) to provide theological grounding to his view,” which
was in opposition to that of Baudhāyana, who used the theology of debt “as an
argument against the āśrama system as a whole and against celibate asceticism”
(2008, liii n. 1, citing Olivelle 1993, 86–91). On this matter, Manu states:

Only after he has paid his three debts, should a man set his mind on
release (mokṣa); if he devotes himself to release without paying them,
he will proceed downward. Only after he has studied the Vedas
according to rule, fathered sons in keeping with the Law, and offered
sacrifices according to his ability, should a man set his mind on
release. (Manu 6.35–36; Olivelle 2005a, 600; 2008, xxi)

We will return to debates about the āśrama system later in this chapter, and to
Olivelle’s discussion of Manu’s views on the relation of debt to mokṣa in our
final chapter.
Now if the dates estimated both for this transition and for the society
described in the Pāli suttas (see chapter 4) are within reason, the dharmasūtras
would be beginning their prescriptions for Brahmanical and Ārya culture in con-
ditions chronologically (if not necessarily geographically) more or less at the near
side of the cusp that these Buddhist texts describe. There is a major difference
that will have to be explained: while the dharmasūtras continue to exalt the village
and discourage Brahmins from even going to crowded and noisy places,12 the
Pāli suttas, supposedly reflecting slightly earlier conditions of around 300 BCE,
already present mahāsāla Brahmins—with the emphasis on their being wealthy
landholders, not householders (see chapter 4 § A)—comfortably ensconced in
market towns and capital cities. But otherwise, the Buddhist descriptions high-
light many identical and overlapping Brahmanical “regimens” (studentship with

11. Cf. Olivelle 1993, 46–53; Malamoud 1996, 92–108 (104–5 on the debt to ancestors); and Biardeau
1976, 36, 40–41. The Mahābhārata tells several stories of ancestors appearing to demand that a man have a son,
beginning with the tale of Jaratkāru (1.13, 34–36, 41–44, 48–49, 53).
12. See Ᾱ 1.32.18–19, 21 on avoiding fairs, casinos, and cities; Lubin 2005, 79–80 n. 5 on usages of
nagara, city.
 post-vedic brahmanical dharma 187

celibacy as brahmacariya, Vedic study as ajjhena,13 triple Vedic knowledge as

tevijja, hospitality practices evident in the narratives) and personal types (the
Vedic scholar or sottiya, the “bath-graduate” or n[a]hātaka).14 And they also offer
particularly intriguing treatments of the five mahāyajñas (Pāli mahāyaññas) as
practiced by both householder and ascetic Brahmins.15 Indeed, the Buddhist texts
would seem to delineate the five little great sacrifices (mahāyaññas) described so
far, which the Buddha basically endorses as piety practices wholesome for lay
Brahmin converts, in juxtaposition with five big great sacrifices (mahāyaññas)—
śrauta rites with animal victims that the little great sacrifices would supposedly
have reformed—which he condemns.16 Moreover, the Pāli suttas also seem to
introduce, as a further and still more acceptable modification of both types of
mahāyaññas, a grouping of “five Brāhmaṇadhammā” or “five things that the
brahmins prescribe for the performance of merit, for accomplishing the
wholesome”:17 truthfulness (sacca), austerities (tapa), purity in sexual life (brah-
macariya), erudition (ajjhena), and charity (cāga, Sanskrit tyāga)—each “as old as
Vedic culture, itself” (Tsuchida 1991, 72). In the Subha Sutta (MN 99.18–27), the
Buddha favors this set emphatically by supplementing the five with the addition
of a recommended sixth, “the motive of compassion” (anukampājātika), then
goes on to interpret them in relation to the four “unlimited” Brahmavihāra prac-
tices of friendliness, compassion, sympathetic joy, and equanimity that form “the
path to the company of Brahmā,” and finally declares that what they really are is
“equipment of the mind”—which thrills the great landholder Brahmin Jānu ̣ ssoṇi
when he hears of it (see chapter 4 § A).
The Buddhist texts would no doubt have their own reasons for pushing such
distinctions between different types and modifications of mahāyaññas, and

13. Or ajjhayana, in each case equivalent to Sanskrit adhyayana, studying the Veda.
14. See Tsuchida 1991 on these Pāli terms and their Sanskrit counterparts.
15. What follows is drawn mainly from the rich treatment of these three sets by Tsuchida (1991, 68–90),
but the juxtapositions and usages of “little” and “big” are my own. Cf. Inden 2006, 92: “I am tempted to see these
five Great Sacrifices as dialectically formulated in response to the Great Gifts of the five precepts of Buddhism.”
16. See Tsuchida 1991, 88–89: the five include three known śrauta rites, the Aśvamedha, Puruṣamedha or
human sacrifice, and Vājapeya, and two that are obscure (see Falk 1988). See Bodhi 2000, 171–72, 402 n. 214 on
Kosalasaṃyutta 9, where inclusion of animal sacrifice is said to have been instituted by King Okkāka; and
Kūṭadanta-Sutta (DN 5), where the Buddha condemns mahāyañña with animal sacrifice and tells how in a former
life as a royal chaplain he advised a king to do a beneficial sixteenfold bloodless “great sacrifice” instead (DN
5.9–21; Tsuchida 75–76). Tsuchida 80–83 notes the “semantic twist” (82) put on the usage in MN 92.7–8
(Sela Sutta), where preparations for a visit by the Buddha to the matted-hair ascetic Keṇiya are compared to a
mahāyañña. Should it be one like that planned by the stereotyped landlord Brahmin Kuṭadanta, it would include
“seven hundred bulls, seven hundred bullocks, seven hundred heifers, seven hundred he-goats and seven hundred
rams” (DN 5.1)—as Tsuchida 89 says, “unlikely details” and probably satirical. Cf. Bailey and Mabbett 2003,
249–52, suggesting that the Buddhist texts make an analogy between śrauta rites and “the ‘total’ event of the
Buddha visiting a village.”
17. MN 99.8, Subha Sutta; Ñāṇamoli and Bodhi 2005, 810 translation.
188 dharma

between wealthy landholding Brahmins who could be encouraged to do Buddhist

modifications of the little mahāyaññas, and lean longhaired ascetic Brahmins
called jaṭilas who did bigger mahāyaññas as fire-offering rites in assamas (Sanskrit
āśramas) or hermitages, and who could make impressive converts.18 The
dharmasūtras may be said to unify these matters by leaving big mahāyajñas to the
śrautasūtras that precede them and marking more of a continuum between
Brahmin occupations and lifestyles—notably where Āpastamba defines “the occu-
pations specific to a Brahmin (svakarma brāhmaṇasya)” as “studying, teaching,
sacrificing, officiating at sacrifices, giving gifts, receiving gifts, inheriting, and
gleaning, as well as appropriating things that do not belong to anybody” (4–5)!
Āpastamba thereby begins with what become the standard six “occupations” or
“duties” of Brahmins (see Manu 10.74–75; Mbh 3.189.12c; 13.129.7–8), while with
the last three, idiosyncratic in Āpastamba, before just pushing the envelope at the
end, he comes up with the contrastive pair of “inheriting and gleaning.” This odd
couple, as Biardeau observes, suggests that Āpastamba is putting together two
contradictory lifestyles—the wealthy landed householder and the ascetic—that
Brahmins are newly settling into: “Inheritance presupposes that the Brahmin has
goods to transmit to his descendants, while gleaning, on the contrary, is witness
to the extreme poverty of the Brahmin who lives day to day, of whom some make
an ideal as he remains entirely a householder” (Biardeau 2002, 1: 77). Of course
the pair could correspond fairly well to the mahāsāla and jaṭila Brahmins juxtaposed
in the Pāli suttas. We shall look into this gleaning ideal in later chapters.
It is noteworthy that we have here only reasons to affirm that the Buddhist
suttas describe a slightly earlier state of Brahmanical society than the
dharmasūtras, and we may suspect two reasons why the latter do not yet bring
Brahmins to the market towns and cities: that they continue the conservative
outlook of the gṛhyasūtras, and that the cities were becoming crowded with
nāstikas, including Buddhists who might try to convert them. The epics and
Manu will find ways to bridge this rural/urban divide. But while the epics have
relatively little to say directly about mahāyajñas as a group or a formal term,19

18. Tsuchida 1991, 78, 80, 83–88. The canonical Pāli texts call them jaṭilas with reference to their wearing
matted hair (jaṭā). The Kassapa brothers, for instance, became prominent converts and disciples.
19. Tsuchida 1991, 88 notes that the Mahābhārata calls the Aśvamedha a mahāyajña (12.260.37), but this
does not seem to be a technical usage, as the epics refer frequently to the Aśvamedha as a “great sacrifice,” as also
other śrauta rites (e.g., Mbh 1.13.93, Serpent-sattra; 2.11.6 and 3.241.23, Rājasūya). While the Mbh alludes to the
(little) five frequently, its only explicit references to them that I can find are at 13.129.46, where Śiva ascribes them
to forest-dwelling sages, and 14.15.16 (Anugītā). Nārada probably suggests the small daily routines when he urges
the Pāṇḍavas to offer “great sacrifices” when they have lost the dice match and are about to enter the forest
(2.71.44), whereas Draupadī soon reminds Yudhiṣṭhira that although he now lives in the forest dethroned, “You
have offered up the great sacrifices of the Horse Sacrifice, Rājasūya, Puṇḍarīka, and Gosava with ample stipends
for priests” (3.31.16).
 post-vedic brahmanical dharma 189

they widen them immensely in narrative contexts, where hospitality can be

exchanged also with edifying animals.20 But it is especially Manu who unpacks
the five mahāyajñas, and quite possibly makes them relevant in some ways to
urban and commercial living. At 3.67–89 (etc.), he makes clear for the first
time that one is to use the domestic fire rather than the three śrauta fires for
them, and suggests considerations of nonviolence (ahiṃsā) by introducing the
five by the “inevitable violence” of “the householder’s five slaughter-houses
(pañcasūnā gṛhasthasya)”: the fireplace, grindstone, broom, mortar and pestle,
and water pot, which the great Ṛṣis designed the “great sacrifices” to expiate.21
At 4.21–24, he goes on to say that the five can be performed by interior medita-
tive means: by the sense organs, speech, breath, or knowledge.22 Yet at 6.5–9,
he still enjoins that they be continued by the forest hermit. As Tsuchida
(1991, 68) observes, with Manu, “extolling of the mahāyajña-s has reached the
utmost magnitude,” on which he quotes Manu 3.75: “He should apply himself
here daily to his vedic recitation (svādhyāya) and to making offerings to gods;
for by applying himself to making offerings to gods, he upholds (bibharti) this
world, both the mobile and the immobile.”

B. Toward Consensus in Brahmanical Dharma Texts

Taking Olivelle’s cue that dharma was never a prominent term in early pre-
Buddhist Vedic usages, we have followed his hypothesis on the innovative royal
character of the Buddhist usage of dharma to the point where it is ready to
become more complicated, for Olivelle hypothesizes that it is only once Aśoka
broadcasts the term in his edicts that Brahmanical culture develops texts in
which to articulate dharma as the all-embracing norm of post-Aśokan
Brahmanical culture. As of 1999, of the four earliest dharmasūtras, Olivelle
placed Āpastamba first, in the early third-century BCE, and thus roughly
contemporary with or even prior to Aśoka’s edicts; Gautama second, in
the mid-third-century BCE; Baudhāyana third in the mid-second-century BCE;
and Vasiṣṭha last, bringing us down to the first- or second-century CE (1999,

20. At the sage Upamanyu’s āśrama, for instance, “Mongooses sported with snakes, and tigers with the
deer, like friends” (Mbh 13.14.42ab). It is always an implicit question in animal fables, and sometimes an explicit
one, whether one can learn dharma from animals. The monkey Hanumān says that those who come from animal
wombs, like him, “do not know dharma,” as do men who are “endowed with intellect (buddhi)” (3.146.75); but
then he goes on to speak about it: first as regards its violation by the Pāṇḍavas’ overhunting (77), and then at
length about the dharma of the yugas (148.9–36).
21. He also mentions teaching rather than svādhyāya as the offering to the Veda, before also mentioning
svādhyāya. On these points see the excellent discussion by Biardeau 1976, 41–43.
22. Cf. Mbh 12.12.23: Nakula, recommending the householder life to the reluctant Yudhiṣṭhira after the
war, reminds him that some do the “great sacrifices just with their minds” (mahāyajnān manasāiva vitanvate).
190 dharma

xxviii–xxxiv). In his 2005a critical edition of Manu, holding to the same

sequence, he finds these dates “still . . . reasonable,” but is “inclined now to
place them somewhat later” (2005a, 20–21 n. 32). Olivelle has thus been
revising downward both from others’ earlier datings and his own.23 Most
recently, he writes, “The very creation of a Brahmanical genre of literature ded-
icated to dharma was possibly due to the elevation of this word to the level of
imperial ideology by Aśoka” (2005a, 39; cf. 2005b; cf. 2004a, 506). We have
already found Olivelle’s chronological hypotheses useful in relation to pre- and
post-Mauryan Vinaya datings in chapter 4.24
As Olivelle says, the dharmasūtras are “Dharmaśāstric texts written in the
sūtra mode” (2005b, 165). Other than Āpastamba, the others use the term
dharmaśāstra when referring to themselves (G 11.19; B 1.1.13; V 24.6; 27.19);
and the grammatical commentary of Patañjali, which can be fairly reliably
dated at about 150 BCE, also uses the term dharmaśāstra with reference at least
to dharmasūtra rules and probably to extant dharmasūtra texts (Olivelle 2005b,
161–63). Olivelle distinguishes within the overlap: “In the early literature, . . .
the titles Dharmaśāstra and Dharmasūtra were synonyms, the former referring
to their substance and the latter to their linguistic form” (163). That is, whereas
the term śāstra denotes “instruction” about a topic that can take the form of a
“treatise,” sūtra, meaning “thread,” is used for threaded aphorisms that call for
elucidating commentary. The dharmasūtras are thus included in the explicit
dharma tradition of “treatises on dharma” (dharmaśāstra) along with Manu.
Manu is titled both as Manu Smṛti and Mānava-dharmaśāstra, and with the lat-
ter usage it offers itself as the first text in this dharma tradition (those named
after Yājñavalkya and Nārada will follow) to be called a śāstra rather than a
sūtra, having abandoned the aphoristic sūtra form.
I date Manu and the Rāmāyaṇa a little later than or possibly even overlap-
ping with the completion of the Mahābhārata, which I have urged would have
been composed under the inspired leadership of a main author, whom the epic
itself calls Vyāsa, at the head of some kind of committee or atelier, about which
I have hypothesized two main things: that the interpersonal dynamics of this
group would likely have borne some implied resemblance to the figures and
processes depicted in the epic’s frame stories; and that this team would have

23. Olivelle provides good grounds for revising downward (by roughly a century) from dates he had pro-
posed in Olivelle 1993, 71, 94, 101–3, and for considering Ᾱ prior to G. Cf. Olivelle 1999, xxxi–xxxii on older and
widely cited datings proposed principally by Kane 1962–75, 1: 19–90, 94–112; 3: xvii, who dates G before Ᾱ; and
Biardeau 2002, 1: 68, who considers Kane’s “science sans faille” in placing G first. Biardeau adds, “If the 6th–4th
centuries can represent a plausible point of departure, . . . the blossoming of the sūtra literature corresponds,
come what may, to the birth of Buddhism” (2002, 1: 68).
24. See chapter 4 § C, especially § C.3.
 post-vedic brahmanical dharma 191

accomplished its masterpiece over a period of no more than two generations at

some point between 150 BCE to the turn of the millennium.25 Manu is perhaps,
at the earliest, an early Śuṅga text, according to Witzel (2006, 482), or as late as
the Kushanas according to Olivelle (2005a, 24–25),26 who dates the Mahābhārata
after it (2004b, xxiii; 2005, 23–25). Biardeau dates the Mahābhārata and Manu
as “no doubt nearly contemporary” around 200 BCE, and the Rāmāyaṇa around
100 BCE (1999, xxxiv–xxxiv, li–lii). We shall come back to these probably intrac-
table issues. For the moment, it will suffice to say that Manu gives central
attention to the dharma of kings—Brahmanical kings, Hindu kings-to-be—in
ways that go far beyond anything in the dharmasūtras. The obvious point for
now is that Manu is joined by both epics in making the king’s dharma central.
Having considered Olivelle’s hypothesis so far mainly in terms of king-
ship, we must now begin to address points at which it becomes more complex,
even as he has continued to work on it. In the article “Power of Words” (2005b,
121–35), Olivelle reviews his understanding of what would have happened by
the middle of the third-century BCE to have prompted the creation of the
dharmasūtras, and outlines four factors:

1. The mainly royal term dharma was appropriated by ascetic groups,

including especially Buddhists, and in particular by the charismatic
leaders of such groups, who used the term to define their “law,”
“teaching,” or “view.”
2. They ethicized the term.
3. It became a public term under Aśoka.
4. The kalpasūtra literature “facilitated” a Brahmanical “reappropriation”
of the term by providing “a literary structure already in place” that
had articulated “the ritual strand of the semantics of dharma already
found in the Vedas and the Śrauta-sūtras” (Olivelle 2005b, 132–33).

In a further twist, Olivelle also mentions the topic of the “sources of

dharma” as “one aspect of the emergent brahmanical discourse on dharma that
may have some relationship to” Buddhist and other “ascetic appropriations of

25. Hiltebeitel 2001a, 18–20; 2004a, 215–19; 2005c, 89. For overview and updated discussion, see Adluri
2011. For further refinement as to this span’s historical parameters and discussion of its background events, see
Hiltebeitel 2011a, chapter 4: “Why Itihāsa: New Possibilities and Limits in Considering the Mahābhārata as
History.” Cf. Bodhi 2000, 30 proposing that the arranging of the Buddhist suttas would have been the work of
one or more appointed committees.
26. Olivelle 2005a, 24–25. Cf. 2004b, xxiii; cf. 2005a, 20–25, with mention (21 [cf. 43] of the view of
Jayaswal 1930, 29) that Manu would be “a work of the Śuṅga period during a time of Brahmanical revival after
the Aśokan period,” and thus written probably early “during the last 170 years before the common era.”
Bronkhorst in press considers only Olivelle’s later dating of Manu.
192 dharma

the term” (2005b, 133; cf. 2004a, 506)—a point he then advances in his
“Explorations in the Early History of Dharmaśāstra” (2005b, 155–77; 2006b).
Here, Olivelle considers it additionally “probable” that the dharmasūtra authors
(dharmasūtrakāras) would have been the first in the sūtra tradition to explicitly
address the issue of the “authority” (pramāṇa) of their own “sources” (mūla)
because they were “consciously responding” to “the Buddhist theory of
dharmapramāṇa,” which makes the word of the Buddha, buddhavacana, the
authoritative source of dharma. In contrast to such a reliance on a single char-
ismatic authority,27 the Brahmanical texts diversify the sources of dharma, and
each includes among them some kind of consensus. Āpastamba speaks first of
“accepted customary Laws” (sāmayācārikān dharmān)28 and mentions two kinds
of authority (pramāṇam) on these: “acceptance by those who know dharma”
and the Vedas; Gautama (1.1–2) reverses these priorities, placing Veda first,
and also mentions smṛti—tradition, “textualized memory”29—as a third source
(mūla) between Veda and “practice” (śīla) (1.1–2); Baudhāyana keeps the same
order as Gautama but offers greater detail, mentioning “each Veda,” “what is
given in the tradition” (smārta), and “the conventions of cultured people”
(śiṣṭāgamaḥ)” (1.1–4); and Vasiṣṭha, after prioritizing Veda and tradition (smṛti),
says that “the practice of cultured people (śiṣṭācāraḥ) becomes authoritative
(pramāṇam)” only where these first two “do not address an issue” (1.4–5).
Meanwhile (I think between the first three dharmasūtras and Vasiṣṭha30), Manu
describes “the entire Veda, Tradition, the practice (śīla) of those who know the
Veda” along with “the conduct (ācāra) of good people” (sādhūnām), and “what
is pleasing to oneself” (ātmanas tuṣṭiḥ) as one apparently fourfold31 “root of
Law” (dharmamūlam). The Vedas, tradition or “textualized memory,” and some
kind of custom32 in fact all go into making for these varied kinds of consensus:
first, in Āpastamba and Gautama, the consensus of those who know dharma

27. Buddhavacana can cover not only what the Buddha said, but the texts that recount what he said, and
what can be discerned by trained monks based on what he said or is said to have said. For some discussion, see
Nattier [2003] 2005, 11–14.
28. See Olivelle 1999, 353: “Ᾱpastamba, more than any other author, points to generally accepted custom
as the basis of Law (dharma). He puts the Vedas last, in contrast to the other three who place it first.” But last
would not simply mean least.
29. See Olivelle 2005b, 168: smṛti begins as “a textualized form of memory”; cf. 171.
30. Note that right after discussing the sources and geography of dharma, Vasiṣṭha as it were quotes
“Manu”: “When there are no specific rules in vedic texts, Manu has said that one may follow the Laws of one’s
region, caste, and family” (V 1.17). See Hopkins 1882, 241 on this citation, and more below on his discovery
of the two ways by which Vasiṣṭha cites “Manu.”
31. See Lingat 1973, 6; Olivelle 2005a, 24, n. to M 2.6, on Manu’s use, like Gautama, of śīla, and the com-
ment that “the distinction that commentators seek between ‘practice’ (śīla) and ‘conduct’ (ācāra) may be
misplaced.”
32. Variously called ācāra (Ᾱ 1.1.1; V 1.5); śīla (G 1.2), and āgama (B 1.1.4), with “similar and overlapping
semantic ranges” (Olivelle 2005b, 168). See also Olivelle 2004a, 506.
 post-vedic brahmanical dharma 193

and the Vedas; and then, perhaps more strictly in Baudhāyana, the conventions
of śiṣṭas, the “instructed” cultural elite.33 The dharmasūtra tradition probably
introduces the term śiṣṭa “about the same time” that the grammarian Patañjali
uses it to define those who speak proper grammatical Sanskrit; and both
Baudhāyana and Patañjali use it to introduce a newly conceived terrain called
Āryavarta, the Land of the Āryas, as “the place where śiṣṭas live” (B 1.2.11 and
13)—a combination repeated not only in the Vasiṣṭha (Olivelle 2005b, 133–34)
but also in the Mahābhārata (Hiltebeitel 2001a, 27–28). Biardeau complements
this perspective. On Baudhāyana’s earliest delimitation of the “land of the
Āryas,” she suggests that he is reclaiming the Indo-Gangetic plain for
Brahmanical culture “in an epoch when, by every likelihood, the implantation
of the Buddhists was a secret to no one” (72). As was noted in chapter 4,
Buddhists seem to have had two views of the Kuru area within this region: that
it was a dangerous place where, in a former life, the Bodhisattva once tamed a
cannibal; and that it was a place where people would be distinctly prepared to
hear some of the Buddha’s subtler Dhamma talks.34
Biardeau also points to numerous ways35 in which the earliest dharmasūtras
are likely to have Buddhism in the “background” as their unacknowledged
“ideological enemy”36—a term that is stronger than necessary for many pas-
sages in both traditions, but not for some,37 of which I consider the following
from Āpastamba Dharmasūtra an instructive early case in point:

Let him not follow the Laws (dharmān) for the sake of worldly
benefits, for then the Laws produce no fruits at harvest time. It is like
this. A man plants a mango tree to get fruits, but in addition he

33. Cf. Mbh 13.129.3–5: Śiva tells Umā that for the sake of dharma in the world, Brahmā created three
eternal dharmas (trayo dharmāḥ sanātanāḥ): what is stated in the Veda as highest, what accords with smṛti treatises
as next, and then what is declared that is based in the practices of Śiṣṭas (śiṣṭāciṛṇaḥ paraḥ proktas)—all while
discussing “what is dharma among Brahmins.”
34. See chapter 4 § B.1.d.i. According to Manu 2.17–23, there dharma gets increasingly pure as one moves
inward through four rings that surround a center of the most exemplary conduct through the generations: (a) the
Land of foreigners (Mlecchadeśa) surrounds (b) Ᾱryāvarta; Ᾱryāvarta encompasses (c) Madhyadeśa; Madhyadeśa
surrounds the (d) Land of Brahmin Seers (Brahmarṣideśa) comprised of Kurukṣetra and the lands of the Matsyas,
Pañcālas, and Śūrasenakas (“All the people on earth should learn their respective practices from a Brahmin born
in that land”); and Brahmarṣideśa surrounds (e) the holiest center named Brahmāvarta where the Sarasvatī and
Dṛśadvatī Rivers flow, and where “The conduct handed down from generation to generation among the social
classes and the intermediate classes of that land is called the ‘conduct of the good (sadācāra).’ ” See Olivelle 2005a,
43 on the possible Śuṅga period origins of the concept of Ᾱryāvarta (which could make it later than the early
Buddhist focus on the Kuru country). Cf. Bronkhorst 2007, 1–2 on changing definitions of this geography and his
view of the “enormous divide that existed between Vedic culture and the culture of greater Magadha” (269).
35. For example, the dharmasūtras’ codification and justification of Vedic ritual, including animal sacrifices,
as dharma—something that both the Buddha and Aśoka find useless and delusionary; the increasing attention
given to the king, and the stress on dharma as “merit,” good quality,” or virtue” (Biardeau 2002, 1: 69).
36. Biardeau 2002, I: 70, 82; see 65–83 on the dharmasūtras; also 85–96 on Manu.
37. See Tsuchida 1991, 91 making this point with regard to most but not all early Pāli Buddhist texts.
194 dharma

obtains also shade and fragrance. In like manner, when a man

follows the Law, he obtains, in addition, other benefits. Even if he
does not obtain them, at least no harm is done to the Law (na
dharmahānir bhavati). Let him not become vexed or easily deceived by
the pronouncements of hypocrites, crooks, infidels, and fools.
Dharma and adharma do not go about saying “Here we are!” Nor do
the Gods, Gandharvas, or Ancestors tell us, “This is dharma,” “This is
adharma.”38 An activity that the Āryas praise is dharma, and what they
deplore39 is adharma. He should model his conduct after that which
is unanimously approved in all regions by the Āryas who have been
properly trained, who are elderly and self-possessed, and who are
neither greedy nor deceitful. (ĀpDhS 1.20.1–8)40

In the rogues’ gallery41 just invoked, the “infidel” or nāstika is again a person of
interest in opposition to the āstika, the “yea-sayer,” literally, the “one who says
‘It is.’ ” As Biardeau observes, “in the epic, nāstika is one of the possible
designations of the Buddhists” (2002, 1: 75). Here, Āpastamba mentions nāstikas
among those who, unlike the Gods, Gandharvas, and Ancestors, do, it seems, “go
about saying, ‘This is dharma.’ This is adharma’”!42 Not only should one believe
instead the consensual—indeed, the cross-regional consensual—approval of the
Āryas; to the extend that one shouldn’t believe such talking heads, the passage
suggests a kind of self-validation of dharma similar to what we have seen Manu
introduce as a new and most intriguing source of dharma: ātmanas-tuṣṭi, “what is
pleasing to the self.” Yet as Āpastamba goes on immediately to make clear, at least
as far as he is concerned, not everyone has a self eligible for self-validation, for as
regards the discernment of dharma, only Āryas have such a self.
It is important to distinguish the connotations of the term ātmanas tuṣṭi in
Manu from its later uses in dharmaśāstra, and, still later, its possible implica-
tions for modern Indian jurisprudence. Donald Davis demonstrates that later
dharmaśāstra interprets the term as a default position in resolving legal cases
that cannot be decided by the other three criteria of Veda, custom, or tradition.

38. I follow Fitzgerald 2004b, 671 trans. for this line; the rest, with only slight changes, is from Olivelle 1999.
39. ÷Garh: accuse, charge with, reproach, blame, censure. Censuring Dharma/dharma itself occurs in two
famous scenes in the Mahābhārata. See Hiltebeitel 2001a, 274 on Mbh 18.2.50, where Yudhiṣṭhira does so;
2007b, 46 on 8.66.44, where Karṇa does so.
40. As Olivelle notes (2003, 57), the line may echo Taittirīya Upaniṣad 1.11.4—a passage about consensus
through Brahmins “devoted to the Law,” dharmakāma. The sūtra passage puts more emphasis on “all regions.”
41. The four words used have the following ranges: kuhaka—cheats, rogues, jugglers, impostors; śaṭha—
cheats, rogues, fools; nāstika: naysayers, nihilists; and bālavāda: those who talk like children.
42. See Olivelle 2005b, 130–31: in the Aṭṭakavagga of the Sutta Nipāta, from what seems to be some of the
earliest material in the Pāli canon, Buddhists themselves use dhamma negatively as a term by which other ascetic
schools define something like their distinctive “views.”
 post-vedic brahmanical dharma 195

He opposes Menski’s view that ātmatuṣṭi (for short) is important to a

“self-controlled ordering” that “is the first and foremost method of ‘finding’
dharma.”43 For the “post-traditional” Menski, ātmatuṣṭi is the starting point
for understanding, in current Hindu law, “the Hindu individual’s rights to
participate in the rule-ascertaining process” (126–27). Indeed, writes Menski,
“Self-reflective contemplation should be sufficient to make individuals do the
right thing at any time, which remains the simplest general paraphrase of
dharma” (547). Obviously Manu is not post-traditionalist, and may have more
in mind too than offering a fail-safe position. I believe that where Manu intro-
duces the term, ātmanas tuṣṭi complements the kind of consensus among
Āryas we have just met in Āpastamba. At several points, Manu seems to sur-
round it with Upaniṣadic echoes. For instance, Manu’s main discussion of
avoiding violence is introduced by three verses that seem to entail the notion:

He should carefully avoid all activities that are under someone else’s
control (paravaśaṃ karma), and diligently pursue those that are under
his own control (ātmavaśam). Whatever is under someone else’s
control—that is suffering; whatever is under one’s own control—that
is happiness. He should know that this, in a nutshell, is the
definition of suffering and happiness. He should diligently engage
in those activities that give him inner joy (paritoṣo ‘ntarātmanaḥ)
and avoid those that do not. (4.159–61)

Similarly, “What a man seeks to know with all his heart and is not ashamed to
perform, at which the inner being (ātman) rejoices (tuṣyati)—that is the mark
of the attribute of Goodness (sattvaguṇa)” (12.37).44 In Manu, an implied
consensus would back the one with the attribute of goodness who finds inner
joy in having a self under his own control and no one else’s.
It is thus useful to consider the ways that post-Vedic Brahmanical dharma texts
construe consensus, and build toward it. Here we might expect to find some kind
of differentiation analogous to what we found in chapter 4 among Buddhist texts,
with some—even in the classification of Buddhist suttas—being more for in-house
specialists and others more for public outreach.45 The aphoristic dharmasūtras, like
all sūtras in aphoristic style, are clearly works by and for learned experts to interpret
and apply. The Sanskrit epics are clearly works designed for the widest possible
publics. On this point, Manu lies somewhere quite precisely in between.

43. Davis 2007b, 279, citing Menski (2003) 2005, 126.

44. Cf. Olivelle 2005a, 244 n. to 2.6.
45. Johannes Bronkhorst (2010a) made such a point classifying some texts available to Brahmins during
the classical period: Veda for “internal consumption”; the epics for “external consumption”; and works on state-
craft being “mixed.”
196 dharma

C. What’s New with Manu

According to Olivelle, “Manu introduced two major innovations in comparison

to the previous literature of the legal tradition, the dharmasūtras. First, he com-
posed his text entirely in verse, using the popular simple śloka meter with four
eight-syllable feet. Second, he set his text within a narrative structure that con-
sists of a dialogue between an exalted being in the role of teacher and others
desiring to learn from him” (2004b, xxiii; cf. 2005a, 25). We may speak of these
two innovations as “verse composition” and a “frame story.” Lingat mentions
the same two when addressing the “three main respects” in which not only
Manu but “the dharma-śāstras of . . . Yājñavalkya and Nārada, differ from the
dharma-sūtras” (1973, 73). Along with treating verse composition under the
heading of “form” and the frame story as a new way to formulate authorship and
textual authority, Lingat offers a third difference under the heading of “subject
matter”: Manu’s greater attention to “the duties of the king, . . . including his
judicial functions, and . . . what might be called the ‘legislative element’” (Ibid.).
Olivelle addresses this third difference as a separate “innovation”: “The sections
of Manu dealing with the king, statecraft, and especially judicial procedure, are
either absent or poorly developed in the Dharmasūtras. It was Manu’s innova-
tion to include these discussions in his treatise” (2004b, xix; cf. 2005a, 20). In
keeping major distinctions to two, Olivelle suggests a useful distinction between
form and content: “At the substantive level, the greatest change in the content
of the Dharmaśāstras was the incorporation of matters relating to the king, the
state, and the judiciary (an area I will call artha for the sake of convenience)”
(2005b, 175). Verse composition as form and the frame story as design both
affect the totality of Manu, whereas substantive differences concern only por-
tions, and thus proportions, of Manu’s text. Indeed, Lingat remarks in passing
on another significant difference in content: in contrast to Manu, the dharmasūtras
“contain little or no philosophical speculation” (1973, 74). It is thus possible to
extract four important innovations: two matters of overall form and two having
to do with content. Verse composition and the frame story, both matters of form,
will be the topic of this section. Olivelle’s “artha” nexus and Lingat’s philosophical
overtures, both matters of content, will be treated later and in passing.
On verse composition, Olivelle has written much of importance, but I believe
that he, like all other dharmaśāstra scholars, draws up short on a simple and obvious
point that nonetheless deserves greater consideration: that Manu is a poet. Long
ago, Edward Washburn Hopkins raised the question of whether quotations of
Manu in the dharmasūtras and the Mahābhārata were traceable to Manu, and
found that usually they were not, and, moreover, that similar legal content would
 post-vedic brahmanical dharma 197

sometimes appear in different meters and inverted verse structure, and that
particularly in the area of rājadharma, where Manu expands the content, “the vast
number of verses [in the Mahābhārata] identical with those in the Manu-treatise
[were] not referred to it, or sometimes referred elsewhither” (1882, 250). Based on
such findings, Hopkins submitted “that legal saws and maxims were couched in
such general language and in such plastic swinging verse-form as to be handed
down merely as a whole . . . [and] changed” from text to text: “I fancy they did not
lay much stress on exactness. . . . They did not quote, they paraphrased” (1882,
250). Olivelle rightly criticizes Hopkins for being the first to hypothesize from
such findings that both Manu and the Mahābhārata are compositions collated over
centuries from material that was “floating in the mouths of people and handed
down from generation to generation . . . divorced from authorial intent and agency
and from social, political, and economic context.”46 But Hopkins’s findings on the
plasticity of citations and maxims in these texts are solid enough to cast doubt on
any kind of straightforward linear development from the dharmasūtras to Manu,
and on any likely direct dependence of either Manu on the Mahābhārata or the
reverse.47 While I believe the Mahābhārata is probably earlier than Manu, and will
cite a number of pointers toward that conclusion, I think we must frame the
question without notions of dependency or—with the exception of some verse
maxims, which we should not expect to explain everything—of prior texts from
which each independently borrows.
Taking note that Manu is “older than any of the other metrical
Dharmaśāstras” (2004b, xxii), Olivelle’s way of handling Manu’s innovation
in using verse is also to call attention to the tradition of citing verse maxims
in the early prose Upaniṣads, some early Buddhist texts (2005a, 6), and in
most of the Dharmasūtras. “It appears that during the last few centuries prior
to the common era ślokas had assumed an aura of authority, and proverbial
wisdom was transmitted as memorable verses. The logical outcome of this

46. Olivelle 2005a, 3, citing Hopkins 1882, 268. Cf. Olivelle 2005a, 23 on the similar views of Bühler: that
both the Mahābhārata and Manu “drew on the same stock of ‘floating proverbial wisdom’” (see Bühler ([1886]
1969, lxxiv).
47. Olivelle views this second point differently: “The references and citations collected by Hopkins, I think,
make a compelling case that the author(s) of the epic knew of and drew upon material from [Manu]” (2005a, 23).
He answers Bühler’s point about the Mahābhārata’s lack of exact replication of Manu ([1886] 1969, lxxiv–xcii) by
ascribing differences and confusions to “the team of workers engaged in the epic enterprise,” citing my view on
the epic’s authorship by committee. But then he leans toward Fitzgerald’s view that it underwent a second Gupta
redaction, giving him room to assign Manu to a much later date (2005a, 23–24 n. 41). Bronkhorst still favors the
idea of a Gupta redaction for such reasons (in press; see chapter 1 n. 11), but with no new arguments for it. I think
Hopkins and Bühler were rightly cautious (see Bühler ([1886] 1969, lxxv) about the relation between the epic and
Manu, and I have given many reasons to reject a “second Gupta redaction” of the Mbh (Hiltebeitel 2001a, 25–30;
2004a, 205–6, 213, 220; 2005a, 458–61, 486–93; 2005c, 87; 2005d, 242–46; 2006a, 227–33, 249–53; 2009a,
196–210; 2009b; forthcoming-a; forthcoming-d), which has only fueled stratigraphic fancies and has never had
a convincing Gupta rationale.
198 dharma

tendency was for texts themselves to be composed in verse, lending authority

to the text by its very literary genre.”48 Manu’s “use of verse . . ., therefore,
must have been part of a deliberate plan to lend the kind of authority to his
text that would come only through this literary genre” (2004b, xxiv). We
must, however, separate this approach’s two components: the well-
documented fact that the legal tradition joins with some others in imparting
an aura of authority to verse maxims, and Olivelle’s theory of a “logical out-
come.” I believe that we are left with something a little hard to square in
Olivelle’s theory: that to speak of an “innovation” or even a “deliberate plan”
as a “logical outcome” is to sell it a bit short.49 Indeed, if we follow Olivelle’s
“hypothesis that Gautama had assumed prominence as the chief Dharmasūtra
by the time Manu was composed” (Olivelle 2005b, 273), Manu’s innovation
of composing in verse cannot just be the “logical outcome” of a dharmasūtra
tendency. Rather, as Olivelle amply demonstrates under the title “Manu and
Gautama: A Study of Śāstric Intertextuality” (2005b, 261–74), among the
earlier dharmasūtras, the Gautama Dharmasūtra was Manu’s main model
and the most frequent source for versification of specific sūtras. In composing
in verse, Manu thus makes a deliberate and contrastive departure from his
likely “paradigm”50 among the dharmasūtras, since Gautama is the only one
of them that is composed entirely in prose.
It would thus appear that Manu’s intertextual situation must be wider and
more complex than just the linear legal or Śāstric one, or the certainly less
cogent one of borrowing the aura of Śāstric, early Upaniṣadic, and early
Buddhist aphoristic verse. Olivelle gives a brief next thought to such a wider
setting, but only as an aside to his main and abiding interest, “the legal tradi-
tion”: “We have, of course, the parallel examples of the epics Mahābhārata and
Rāmāyaṇa composed in verse and claiming religious authority. The legal tradi-
tion followed the tradition blazed by Manu; all later Dharmaśāstras are written
in verse.”51 Lingat makes a similar and slightly more bridging nod toward the
epics, noting that the dharmaśāstras are composed “entirely in verse, the meter,
śloka (anuṣṭubh), being used in the two great epics, the Rāmāyaṇa and the
Mahābhārata. The style is less archaic and very close to classical Sanskrit”

48. Olivelle 2005a, 26, which slightly modifies 2004b, xxiv; my emphasis.
49. Bühler creates a similar impression that “it is no more than might be expected” that the stock of metrical
maxims “should have gradually [been] augmented” ([1886] 1969, xci); cf. also Brick 2006, 299–301, following
Olivelle’s explanation to account for changes in the meaning of smṛti from quotable “tradition” to literary category.
50. Olivelle supplies “evidence that Gautama had risen to prominence as the first and perhaps the paradig-
matic Dharmaśāstra” at least by the seventh-century CE (2005b, 272–73).
51. Olivelle 2004b, xxiv; after this same sentence, 2005a, 26 goes on to say that this “move away from
prose . . . continues” in the Purāṇas, while “[t]he artha and kāma traditions continued to produce prose works, as
did the ritual, philosophical, and grammatical traditions.”
 post-vedic brahmanical dharma 199

(1973, 73). But if, as I think we will be able to demonstrate, the epics provide
more than just “parallel examples” and “archaic” stylistic precedents, then the
question of Manu’s “genre” context must be extended to them more seriously.
The Mahābhārata has some interesting things to say about its relation to
other works of (presumably classical) poetry.52 Near the beginning of its first
adhyāya, before the bard Ugraśravas opens his storytelling to the (celestial) Ṛṣis
of the Naimiṣa Forest, he announces:

I will proclaim the thought entire of the infinitely splendid Vyāsa.

Some poets (kavayaḥ) have told it before, others tell it now, and
others too will tell this history (itihāsa) on earth. It is indeed a great
erudition (mahaj jñānam) established in the three worlds that is held
(dhāryate) by the twiceborn in its particulars and totalities. (1.1.23–25)

Then, a little further along, toward the end of the second adhyāya, the bard adds:

From this supreme history (itihāsa) arise the inspirations of poets

(jāyante kavibuddhayaḥ), just as the configurations of the three
worlds53 arise from the five elements. Ancient lore (purāṇa) turns in
the compass of this narrative (ākhyāna), O twiceborns, just as the
four classes of creatures (turn) in the compass of space. Works of
every quality resort to this narrative even as the interacting senses
resort to the manifold workings of the mind. There is no story
(kathā) on earth that hasn’t resorted to this narrative, even as
support for the body comes from resorting to food. This narrative is
lived on by all the best poets even as a lord is born with servants
wishing for promotion.54

Taken together in sequence, these two frame passages—both of which, let us

note (see also Hiltebeitel 2011a, chapter 4), identify the Mahābhārata by its
primary genre identification of itihāsa, “history” (literally, “so indeed it
was”)—first have the bard locate the text as a possession of Brahmins. That
is, it is “held” (dhāryate)—from √dhṛ—by the “twiceborn,” a term by which

52. These two paragraphs reconsider Hiltebeitel 2008a, 201–10 on the two epics in the light also of Manu.
53. 1.2.273d: lokasaṃvidhayas trayaḥ; Van Buitenen (1973, 43) trans. Ganguli [1884–96] 1970, I, 35 has “the
formation of the three worlds.” Nīlakaṇṭha gives a Vedanticizing construal ādhyātmādhibhūtādhidaivānāṃ
samyāgvidhayo racanā (Kinjawadekar, ed. 1929–36, 1: 39) that would suggest “the three dispositions of the world.”
54. Mbh 1.2.237–41:
itihāsottamād asmāj jāyante kavibuddhayaḥ/ pañcabhya iva bhūtebhyo lokasaṃvidhayas trayaḥ//
asyākhyānasya viṣaye purāṇaṃ vartate dvijāḥ/ antarikṣasya viṣaye prajā iva caturvidhāḥ//
kriyāguṇānāṃ sarveṣaṃ̄ idam ākhyānam āśrayaḥ/ indriyaṇ̄āṃ samastānāṃ citrā iva manaḥkriyāḥ//
anāśrityaitad ākhyānaṃ kathā bhuvi na vidyate/ āhāram anapāśritya śarīrasyeva dhāraṇam//
idaṃ sarvaiḥ kavivarair ākhyānam upajīvyate/ udayaprepsubhir bhṛtyair abhijāta ivēśvaraḥ
200 dharma

the Mahābhārata most typically means Brahmins, though we may also detect
an implication that what Brahmins “hold” as dharma would pertain not only
to other “twiceborn” Āryas but to all beings. Then, having reported that he
has heard the Mahābhārata’s debut recital in the world of men at King
Janamejaya’s snake sacrifice, the bard indicates that, now that it is in that
world of men for others to resort to, other poets will be inspired there by it as
they seek to stand on its shoulders. Indeed, the bard seems to sense that other
works of poetry have emerged on the horizon with their authors poised to do
just that.
I believe this possibility should be considered. Just as there is a question of
the relation between the Rāmāyaṇa poet and the sage Vālmīki, who is mentioned
fairly frequently in the Mahābhārata,55 there is the question of the relationship
between the Manu poet and the sage Manu, more frequently mentioned in the
Mahābhārata, and often as a primal lawgiver in connection with epigrammatic
verses on dharma and the promulgation of śāstras.56 Considering the overlap-
ping audiences or reading communities targeted by these works, with all three
envisioning ideal kingdoms with cities,57 and Manu mentioning Brahmins,
moneylenders, merchants, and kings in that order as the four who prosper
while others suffer,58 it is not impossible that Vālmīki and Manu could be noms
de plume taken up from the Mahābhārata by poets familiar with that text.
Indeed, if the Mahābhārata is a work of composite authorship, “Vālmīki” and
“Manu” may even have been apprentice-contributors, or at least persons
familiar with the Mahābhārata project.59 To speak of the temporal priority of
the Mahābhārata over the Rāmāyaṇa and Manu is thus not to rule out the pos-
sibility that the last two might have been started before the Mahābhārata was
finished.

55. See 1.50.14 (Vālmīki praised by Ᾱstīka for his gentle firmness), 2.7.14 (among many famous Ṛṣis in
Indra’s sabhā or hall), 3.83.102 (among select Ṛṣis waiting for the Pāṇḍavas on pilgrimage), 5.81.27 (among illus-
trious Ṛṣis attending Kṛṣṇa’s departure for Hāstinapura to confront the Kauravas), 12.200.4 (among Ṛṣis cited
by Bhīṣma regarding Govinda), and 13.18.7 (addressing Yudhiṣṭhira regarding Śiva, who absolved Vālmīki of a
charge of Brahmanicide and told him, “Your fame shall be foremost (yaśas te ‘gryaṃ bhaviṣyati)” (8f ).
56. See Hopkins 1882, 247–50, 254–55 (see Mbh 12.322.26–42 and 804* on the “treatise embracing
worldly dharma [lokadharma]” of the Citraśikhaṇḍins, whose laws are to be declared by Manu, son of the Self-
existent [41cd]; 12.37.1–6; 12.259.35: “Manu, son of the Self-existent, out of pity for his creatures, declared the law,
that the great fruit thereof might not perish.”), 262, 264. See more widely 251 ff., beginning, “I come now to
Manu himself. God, creator, demi-god, king, and law-maker—these are his roles.”
57. See M 7.115–17; Mbh 3.33.24cd, where, early in their forest exile, Draupadī tells Yudhiṣṭhira that “the
success of houses and cities is caused by man (agāranagarāṇāṃ hi siddhiḥ puruṣahaitukī ).” Cf. Mbh 6.62.40 on
the creation yuga by yuga of seagirt cities ( purīs ).
58. M 8.169; Manu has a penchant for listing things in order (anupūrvaśas, anupūrvyeṇa, kramaśas, yathākramam,
etc.). I would suggest that one could think of the four mentioned here as the chief beneficiaries of Manu.
59. Cf. Hiltebeitel 2001a, 169, attempting to imagine the joint authorship of the Mbh: “The inner core, the
sattrins or committee, would no doubt have had a philosopher and a dharmaśāstra connoisseur among them. . . .”
 post-vedic brahmanical dharma 201

I offer this approach to the relation between these texts not because it can be
supported by conclusive evidence, but because it opens what I believe is a fruitful
perspective on several fronts that bear on how we understand the history of
dharma. First, it suggests that there could be something significant to draw from
dharma’s beginnings in the Ṛgveda not as a legal term but as an (often enigmatic)
coinage of poets. Second, it takes note of a recurring Brahmanical pattern. Manu
and the Rāmāyaṇa can be said to tighten up the more pluralistic, flexible, and
“broad” dharma of the Mahābhārata much in the way that the three other probably
later dharmasūtras tighten up the dharma of Āpastamba (Olivelle 1999, xxxix). For
in tone, at least, the Mahābhārata is closer to Āpastamba, and specifically so where
Āpastamba allows that one may learn aspects of dharma from women and Śūdras,60
and implies that dharma and adharma are too subtle to “go about saying ‘Here we
are!’ ” (Ibid.). Third, it offers a way to map the relation of dharma and bhakti in
these three texts without mapping bhakti out of the epics (see chapter 12). Fourth,
there is more to say about Manu’s place as a poet within what Olivelle calls “the
legal tradition” itself. Yājñavalkya and Nārada, Manu’s successors in writing
Smṛtis, stay in the śloka verse more as legislators than philosophers or poets, at
which they are not Manu’s match. Indeed, their verse sometimes condenses and
strives for precision (Lingat 1973, 98, 102) almost to the point of reverting to the
aphoristic. As Lingat says of Nārada, “The frequent philosophical or moral specu-
lations of Manu have entirely disappeared” (1973, 103); and of Yājñavalkya, “We
are struck, especially if we have just read Manu, by the sober tone, the concise
style, and the strictness with which the topics are arranged. We find none of those
lyrical flights which are, after all, the literary beauty of Manu” (1973, 98). Manu’s
achievement thus stands out—indeed, rather daringly. It is much easier to write
poetry about gods and great heroes and heroines than to write poetry about laws.
Much of our discussion of Manu will revolve around this point, and
I believe many of the passages I quote (and have cited already) support it
without the need to repeatedly call attention to it. Yet it is important not to put
too much of a burden on it. For one thing, Manu is modest. His text makes no
“epic” claims about its poetry being the source of other poetic works, or even,
as the Rāmāyaṇa does, of being the very “first poem” and the source of the śloka
meter (see below). It is also significant that whereas both epics, but especially
the Mahābhārata, punctuate their prevailing śloka narration with what usually

60. Ᾱ 2.15.9; 29.11, 15; Olivelle 2005a, 39–40; compare especially the Mahābhārata’s Pativratā-Upākhyāna
(3.196–206) about a Brahmin who learns dharma from both a woman and a Śūdra. Vyāsa’s son with a Śūdra
woman (Vidura) is major interpreter of dharma, and Vyāsa himself has a fisherwoman (probably Śūdra) mother (see
chapter 8). In contrast, for Manu, see Olivelle 2005a, 39–40 on Śūdra as a “code word” for Buddhists, etc. For
passages that might bear this out, see Manu 4.61; 4.194; 4.218. See 8.20–22 forbidding Śūdras as legal inter-
preters. The Rāmāyaṇa has Rāma kill a Śūdra for performing asceticism (Rām 7.64–67).
202 dharma

seem to be attention-catching verses or strings of them in the more complex

triṣṭubh meter and its variants, and sometimes punctate the strings with even
more dramatically dissonant irregular triṣṭubhs,61 Manu sticks entirely to the
workhorse śloka. Indeed, one of the cases noted by Hopkins in which a śloka of
Manu corresponds to another author’s triṣṭubh is Vasiṣṭha Dharmasūtra 19.37,
which thus “cites” or “quotes” in triṣṭubh what it alleges to have been a “Mānavan
śloka”! Vasiṣṭha has:

With reference to taxes, they also quote this verse of Manu (śulke cāpi
mānavaṃ ślokam udāharanti): “There is no tax when the sum is less
than one Kārṣapaṇa, as also on craftsmen, children, and messengers;
on what is received as alms or what remains after a robbery; and on
vedic scholars, wandering ascetics, and sacrifices.” (V 19.37; Olivelle
1999, 301; 2003, 428–29)

The alleged quote has very little in common with any particular verse in Manu,
which Hopkins attempts to explain by suggesting that Vasiṣṭha would be cit-
ing an earlier form of Manu that would have been “originally written” with
“many” triṣṭubh verses, and that it would now “be impossible to find the same
verse in our treatise,” which, “in shortening the verse to adapt it to its present
metre,” would have left “part of the original . . . omitted.”62 But for most
scholars today, Manu is what it is, and Hopkins’s idea that there would have
been some prior Manu behind Manu has few followers, as can also be said for
the idea that Manu would have versified a prior largely prose and entirely
hypothetical Mānava Dharmasūtra,63 which Hopkins was among the earliest
epic and dharmaśāstra scholars to sensibly reject (1882, 267, 271). As Olivelle
has demonstrated, where Manu versifies, its primary source is Gautama.
But more important, Hopkins’s explanation is not only out of favor and
uneconomical, it undercuts what I consider to be a major but seemingly

61. See Fitzgerald 2005, 138 on “[t]heir psychological functioning, their aesthetic qualities, their interrela-
tionship to other musical and acoustic elements of literary composition”; 146 on “continuous literary aggregation
of stanzas that constitute a ‘passage’ . . . to create literary effects in a literary context,” and the question, “Does the
passage connect with surrounding ślokas or stanzas in fancy meters?” As Fitzgerald mentions, “many of the
epic’s triṣṭubh passages are actually very short” (147). See Söhnen-Thieme 1999, 150 on usages in the dice match
scene of Book 2: “Whereas the triṣṭubh passages provide dialogues and discussions, the anuṣṭubh verses serve
various purposes, the most prominent being narration of action.” I emphasize only specific features of these
authors’ discussions; cf. Hiltebeitel 2001a, 18 n. 70.
62. Hopkins 1882, 243. The Manu verse (M 10.120) that Hopkins cites, while excusing it for its dissimi-
larity on metrical grounds, has only two points in common: a limit of 1 Kārṣapaṇa below which nothing can be
taken from certain groups, though not as taxes but by a conquering king; and the inclusion of craftsmen among
an otherwise altogether different group (Śūdras and artisans) in which, however, instead of being exempt from
taxes they are among those from whom a conquering king can demand services!
63. Bronkhorst in press seems to be a holdout; cf. Bronkhorst 1985b.
 post-vedic brahmanical dharma 203

forgotten discovery that the young Hopkins himself made here: that Vasiṣṭha
cites Manu and Manu in two different ways: “Now it appears to me that there
is an interesting difference in the way in which his quotations are made.”
I notice that when a passage begins with ‘Manu said (abravīt),’ one finds
nothing in Manu exactly corresponding to it”; but “whenever Vāsiṣṭha gives
a quotation which answers exactly to some verse in our present M-treatise,
he always introduces it with the words ‘now they relate on this point a
Manavan verse’ (mānavaṃ cā’tra çlokam udāharanti)” (1882, 241)—or some
close variation, as just cited above in the case where Hopkins most stretches
the point about “answering exactly.” Whereas there is some minimal
dharmasūtra precedent for quoting Manu by the first formula (“he said”),
there is none for quoting Manu using this second formula involving citation
by śloka along with a quotative phrase with udāharanti, “they quote,” pre-
ceded by atha in the dharmasūtras or atra in the Mahābhārata.64 Without
mentioning Manu, however, this formula is used frequently in the
dharmasūtras by both Āpastamba and Baudhāyana (but not by Gautama).
Āpastamba, who is interesting not only for being probably the earliest of the
four surviving dharmasūtrakāras but for its familiarity with something it
calls the Bhaviṣya Purāṇa, begins with ten usages. Two of these refer to what
Olivelle (1999, 30) translates as “two verses from a Purāṇa” (purāṇe ślokau),
one involving a godly quote from Prajāpati and the other a mythological
anecdote (Ā 1.19.13; 2.23.3–5). Others recount illustrative stories (notably
1.22.3–8, an allegory of the eternal being residing in the cave in the heart;
1.32.23, quoting Mṛtyu [Death]; and 2.13.6, pertaining to paternity once one
is in the abode of Yama). Most of the rest are just proverbial sayings (e.g.,
1.19.15; 25.9–10; 35.23–31.1; 2.9.13; 2.17.7–8). Baudhāyana then generalizes
the practice, using the athāpy udāharanti phrasing forty-nine times virtually
anywhere it turns from prose to a quoted verse, one of which is Baudhāyana’s
version of why a woman must be nonindependent through the three phases
of her life (B 2.3.45), whose most quoted version of which is Manu 5.147–48
(see chapter 8). Baudhāyana is proportionally less inclined to relate the
phrase to mythical anecdotes, of which I note only three instances (2.3.31–35,
similar in part to Ā 2.13.6; B 2.4.26, recalling a dialogue between the two
wives of the Mahābhārata’s Yayāti; and 2.11.28 on the origin of the āśramas).
And unlike Āpastamba, it never conjoins the quotative phrase with the term
purāṇa. Gautama then clears the quotative phrase out entirely from its

64. See Tokunaga 2009a, 28: “The particle atha, which matches the style of the treatise, was changed to
atra (‘as to this [point of your question]’) in the dialogue of the Mahābhārata. Just as atrāpi udāharanti corrobo-
rates an instruction in the Mahābhārata, so athāpy udāharanti introduces in the Dharmasūtras a śloka text as an
illustration or corroboration (arthavāda) of the injunction (vidhi) found in the preceding sūtra.”
204 dharma

entirely prose dharmasūtra, yet still cites purāṇa.65 Against this background
where purāṇa overlaps with dharmaśāstra, we cannot expect the Mahābhārata
to be too specific when it mentions either of them.
It is probably revealing that only the two earliest dharmaśāstric texts,
Āpastamba and Gautama, refer to purāṇa as authoritative on dharma.
Meanwhile, whereas Vasiṣṭha is the first and only dharmasūtra to use the
quotative phrase with the source as Manu, or more exactly, a “Mānavan śloka,”
the Mahābhārata is the first and only one of these texts to use it with the
phrase itihāsam purātanam and thus, moreover, to use it in conjunction not
only with the phrase but with the term itihāsa, “history” (see Hiltebeitel 2011a,
chapter 4). None of these other authors link quotations about dharma with the
term itihāsa, even though all the dharmasūtras but Āpastamba use the term
itihāsapurāṇa, albeit in each case only once (G 1.8.6; B 2.9.14; V 27.6).66 The
Mahābhārata uses the atrāpy udāharanti phrasing liberally, especially in the
dharma instructions of its twelfth and thirteenth Books, while Manu and the
Rāmāyaṇa, like Gautama, do not use it at all.67 It would make an interesting
study to look into the ways that four of these classical dharma texts (Āpastamba,
Baudhāyana, Vasiṣṭha, and the Mahābhārata) use the athāpy/atrāpy udāharanti
formula while three (Gautama, Manu, and the Rāmāyaṇa) do not, but one of
the reasons why seems fairly clear. As I believe can be generalized from what
I have tried to show in the case of the Mahābhārata’s use of this formula
(Hiltebeitel 2001b), the intertextual citational interest of the first group bears
a certain resemblance to a scholarly apparatus of footnotes that would be per-
tinent to texts that reflect debates of a scholarly tradition on dharma as legal
precedent, and take some delight in absorbing themselves in a world of varied
and often conflicting views about it. If this is the case, it would suggest that the

65. See Rocher 1986, 85–88, beginning, “The purāṇas are also closely related to the dharmaśāstras. In
fact, the purāṇas are said to be dharmaśāstras. . . . From the time of the early dharmasūtras purāṇas are referred
to among the sources of dharma. Gautama (11.19) requires the king to administer justice in accordance with the
Veda, the Dharmaśāstra, the Aṅgas, the Upavedas, and ‘the Purāṇa.’ ”
66. I avoid discussing the Arthaśāstra, which subordinates dharma to artha. It does cite others’ views (e.g.,
1.4.6), but not by this formula.
67. Hopkins 1882, 259–61 cites what seem to be the only two Mbh passages that link the name Manu (or
a Manu) with ślokas. Neither uses the udāharanti citation, which the Mbh uses widely with other authoritative
sources, although the second passage uses the same verb. 12.56.23 mentions two ślokas “sung (gitāu) by Manu,”
of which 12.56.24 is identical with Manu 9.321 (and Mbh 5.15.32, where, as Hopkins cutely puts it, it is “an
original remark by Agni”), while its sequel verse is untraceable to Manu. Bühler considered the combination of
citing “Manu” with this equivalence in the next verse (about how fire, the Kṣatra, and metal come from water, the
Brahman, and stone respectively, and can be quenched by them) to be one of two pieces of “really indisputable
evidence” that the authors of Mbh Books 12 and 13 “knew a Mānava Dharmaśāstra not identical but closely
connected with our Smṛti” ([1886] 1969, lxxv). But he still has some caution and is relying on lost originals and
textual stratification. Mbh 12.57.43 mentions “two ślokas cited (udāhṛtau) by Manu Prācetasa,” who seems to be a
different Manu (Fitzgerald 2004a, 304 n.) and is in any case the quoter rather than the quoted, with no
correspondence in Manu.
 post-vedic brahmanical dharma 205

second group would be one of texts that claim authority independent of and
above that nexus, no doubt each for its own reasons but in each case in the
name of some kind of streamlining and moral rearmament.68 This point may
now be enriched by a brilliant point made by Muneo Tokunaga: that the
Mahābhārata’s use of this formula treats the informative narrative told as
exemplum as the third member of the classical syllogism to provide illustra-
tion or corroboration of the speaker’s thesis.69 Since, as we have seen, the
Mahābhārata’s primary self-identifying genre term is itihāsa, we may say that
it presents its leading characters not only as “living history” but as interested
in hearing it and citing it in support of their varied views on what we may call
“precedent,” and in some cases “legal precedent,” especially in the Rājadharma
section of Book 12, the Śāntiparvan.70
Yet as regards the matter at hand, what seems to have been missed here,
even by Hopkins, is that Vasiṣṭha consistently knows Manu as a śloka text. To
be sure, Vasiṣṭha is free to attribute things to Manu that one might not find in
Manu and even to cite as “Mānavan” an absent śloka in triṣṭubh. But in the other
cases where he invokes this second formula, Vasiṣṭha takes fewer such liberties
and, as Hopkins shows, answers to Manu more closely.71 Hopkins thus offers
a good reason here, despite himself, to suspect that in being the first
dharmasūtrakāra to recognize Manu as having composed in ślokas, Vasiṣṭha
would be younger than Manu.

68. Cf. Tokunaga 2009, 28–29, observing the same distribution and tracing “the stock phrase
athāpyudāharanti further back to the Gṛhyasūtras,” his explanation for its absence from the Rāmāyaṇa being that
it “stands free from the influence of Vedic exegesis . . . at the beginning of a new era of Sanskrit literature culmi-
nating in classical Kāvya literature.”
69. Tokunaga shows that in such usages, “itihāsas correspond to the third member of the pañcāvayava
syllogism of parārthānumāna in the classical Indian logic. That means, itihāsa plays the role of illustration or
corroboration of the instruction or thesis. . . . It is therefore quite appropriate that an itihāsa is quoted by the word
udāharanti, the noun form of which is udāharaṇa, one of the terms for the third member of the syllogism”
(2009a, 24–25). Tokunaga thinks it “is highly probable that itihāsa was originally the designation of a text not
according to its substance or form, but according to its use in a didactic discussion” (26–27), and shows that Mbh
usages typically occur with instruction either before or after the formula, with the instruction being about factual,
moral, political, philosophical, or religious matters.
70. Since I see the Mahābhārata’s genric and quotative usages going hand in hand, I do not, however, follow
Tokunaga’s idea that the “moral (or dharmic) instructions” found mainly in the Rājadharma mark the point from
which the quotative usage “spread” to later parts of the epic (2009a, 27), or that usages of itihāsa to characterize the
epic in its frames result from “the long history of [its] textual development” as it “gradually changed its nature under
the influence of Vedic exegesis” (29).
71. See Hopkins 1882, 242–43, citing Vasiṣṭha 3.2 [“On this point they cite a verse by Manu: When, without
studying the Veda, a twice-born man strives after other matters, he quickly sinks to the level of a Śūdra in this very
life together with his descendants”] as equivalent to M 2.168; Vasiṣṭha 13.16: [“In this connection, they cite this
verse from Manu: Even after accepting fruits, water, sesame seeds, foodstuffs, or anything else given at an ances-
tral offering, vedic recitation is suspended; a Brahmin’s hand, tradition says, is in his mouth”] as close, with some
variant readings, to M 4.117 (Olivelle 1999, 282 notes that V 13.16 has a variant at B 21.8–10n); and Vasiṣṭha 4.5–7
as having a complex relation to M 5.41 and 48.
206 dharma

This would of course bolster our sense that Manu could be close in time to
the Mahābhārata, for Olivelle dates Vasiṣṭha to the first-century CE (1999,
xxviii–xxxiv) or a little thereafter (2005a, 20–21 n. 32). But it does not support
Olivelle’s dating of Manu as posterior to all the dharmasūtras including Vasiṣṭha,
which he bases mainly on Manu’s “clear advances in thinking on many fronts,
especially in the sections relating to statecraft, royal functions, and judicial
procedure” (2005a, 20).72 Nor does it support his dating of the Mahābhārata to
a time later than Manu, which he finds more likely than the reverse on the
grounds that “a narrative epic would draw on expert śāstras for its discussion of
legal matters than the other way round” (Olivelle 2005a, 23; see n. 47 above).
On the first point, however, one need not assume that a dharmasūtrakāra like
Vasiṣṭha would interest himself in Manu’s areas of expansion when his own
genre would have directed him mainly to the “Mānavan ślokas” concerned with
matters treated by the other dharmasūtrakāras before both of them.73 And as to
dating Manu with the Mahābhārata after it, I believe that Olivelle has pushed
too hard on the lower limit for Manu. But whether Olivelle is right about Manu’s
dates or not, a late date for Manu need not have any bearing on the dating of the
Mahābhārata, which could still be thoroughly or mostly earlier than Manu.74
Olivelle gives Manu a likely date “between the first century BCE and the
second century CE” (2004b, xxiii) or as late even as the “2nd–3rd centuries CE”
(2005a, 25), and, although he considers several factors that could place it early
in this span, he favors such later placement due to the possibility that Manu’s
mention of gold coins could be evidence for a time not before “the earliest native
gold coins discovered in India” from the second-century CE Kushana dynasty.

72. Bronkhorst in press also views Vasiṣṭha as most likely posterior to Manu, contra Olivelle, on the grounds
of Vasiṣṭha’s citing two “almost identical” “mānava ślokas” at V 3.2 (similar to M 2.168, as mentioned in the
previous note), and V 20.18 (similar to M 11.152). But he draws no inference from their being ślokas in Manu. His
main interest in Hopkins’s article is to stratify the Mahābhārata’s citations of “Manu” in accord with Hopkins’s
observation that “[n]ot more than half” of what the Mbh ascribes to “Manu” can be found in Manu (1882, 268).
73. See Olivelle 2005c, 2: “Vasiṣṭha is the only [dharmasūtra] author concerned about meat eating,
indicating that he is living at a time when vegetarianism may have been on the rise”—a topic that interests
Manu (5.48–52, similar to V 4.7a–c, which quotes “Manu” (abravīn manuḥ); cf. M 5.56; 6.05); and one that
also concerns the Mbh (see Mbh 12.323–24; Sutton 2000, 85, 66, 90, 105, 310, 323). Note also the other two
topics of Vasiṣṭha that Olivelle shows to be unparalleled in the other dharmasūtras: one, Vasiṣṭha’s general
comments on women (V 5.1–5) could be taken to outdo Manu’s notorious verse 5.147 on women’s noninde-
pendence, while V 28.7, which seems to finally answer what it takes for women to fall from caste, comes
within a probably late versified addendum to Vasiṣṭha (Olivelle 1999, 397; 2005c, 16, 88); the other, the
king’s control of the economy in the capital, including weights and measures (V 19.13–16; Olivelle 1999,
395; 2005c, 30, 212; cf. M 8.132–37; 403). To make this argument requires only that Vasiṣṭha would know of
Manu as a poetic text and add some points of novel interest on topics already of concern, not that Vasiṣṭha
would have been greatly influenced by or dependent on Manu.
74. Cf. Biardeau 2002, 1: 85 on these matters. Leaving Vasiṣṭha aside (as if not germane to the primary develop-
ments up to Manu?), she says Manu “could be posterior” to the other three, and “could also mark a certain distance in
relation to their content.” It is also “possible” that Manu “already knows the Mahābhārata, indeed the Rāmāyaṇa.”
 post-vedic brahmanical dharma 207

But this may be less decisive than it looks, for as Olivelle admits, “It is, of
course, impossible to say from the archaeological absence that gold coins were
not minted earlier” (2004b, xxiii). For instance, in the so-called “early” Sutta
Nipāta, one hears the story of Bāvarī, a Brahmin who went south, who was one
day asked by another Brahmin for “five hundred gold pieces, but Bāvarī was too
poor to give him such a sum” (Lamotte 1988, 347). Gold pieces may not be
minted coins.75 But even as some other kind of currency they complicate the
use of gold coins for dating texts. The story places Bāvarī in the Deccan
(Dakṣiṇāpatha) on the Godavarī River, which means that wherever they came
from, gold pieces were currency among Brahmins in central India in an “early”
Buddhist sutta. They could have come from Bactria, where, according to Falk,
Graeco-Bactrian kings minted gold coins as early as 170 BCE (2006a, 147). The
Mahābhārata knows this Afghan area well enough for its royal family to have
wooed brides from it twice in the persons of Gāndhārī and Mādrī, who is also
called Bāhlakī, “woman of Balkh” (Mbh 1.116.21a). Such familiarity could con-
textualize a hypothetical familiarity with that area’s currency, especially in the
case of Mādrī who was only available to Pāṇḍu (the Pāṇḍavas’ putative father)
by accepting the Madra custom of paying for the bride (see chapter 8). One
would also need to consider the Buddhists’ tenth precept that monks could not
handle gold or silver, which seems to imply coins (see Gombrich 1988, 103).
In any case, however modest Manu is about being a poet, I think it will be
profitable to follow Vasiṣṭha’s cue in recognizing him as one through his
“verse,” and profitable as well to keep him close to the epics not only by their
common verse composition but as their near contemporary. To begin with,
although it is not one of the fruits of considering Manu as a poet, I consider
Olivelle’s brilliant discovery of the “deep structure” of Manu to be one of the
things that falls within Manu’s poetic range. Olivelle has detected a “signature
of Manu”: the use in “transitional verses” of the expression nibodhata, enjoin-
ing the reader to “learn.” The placement of these transitional verses yields a text
sectioned into four main though uneven parts: a “preamble” (the frame story
on the “Origin of the World”: 1.1–119); an “introduction” (on the “Sources of
Law”: 2.1–24); a long “main section” on the “Dharma of the Four Social Classes”
(with major subsections and sub-subsections also marked off by telltale transi-
tional verses: 2: 25–11.266); and a “postscript” offering a “Determination
Regarding Engagement in Action (karmayoga nirṇayaḥ)” (with two similarly
differentiated subsections dealing with the law of karma and final liberation:
12.3–116) (2004b, xxvii–xxx; cf. 2005a, 7–11). Olivelle regards this “exquisite
structure” to be even deeper than Manu’s organization into twelve chapters

75. As Fitzgerald (2006b) points out to me and Olivelle with reference to the Mbh’s term niṣka.
208 dharma

(adhyāyas), which he sees as “artificial divisions” despite their being “old” and
“followed by all the commentators” (xxvii). But whether or not the four-part
division is older than the twelve chapters, it clearly underlies it, and is, as
Olivelle says, compelling evidence that Manu is not “an edition or version of a
preexisting text, but an original composition written by a single individual. The
kind of deep structure, so subtle yet so clear, makes it impossible to have been
composed either through unconscious accumulation or through a series of edi-
torial interventions spanning long intervals of time. This was conceived and
put together by a single individual with extraordinary ability and a systematic
mind” (Olivelle 2004b, xxii). Let us recall that Lingat also credits this authorial
mind with “lyrical flights” and finds this exquisitely designed text to be one of
“literary beauty.”76 Lately, Olivelle also seems to have become attracted to a
slight modification that could bring him closer to my views on the joint author-
ship of the Mahābhārata: “If not by an individual, then [Manu] must have been
composed by a ‘strong chairman of a committee’ with the help of research
assistants who carried out his plan” (Olivelle 2005a, 7; cf. 19, 26).

D. Brahmā in Manu’s Frame Narrative

To make a thorough break from the kalpasūtras and from the dharmasūtras’ link
with the Vedic śākhās,77 Manu, says Olivelle, “introduced a significant and drastic
innovation: unlike the Dharmasūtras—which were located within the give and
take of an expert tradition, offer glimpses of divergent views within that tradi-
tion, and do not pretend to be anything other than humanly authored works—
the Mānava Dharmaśāstra is presented as a treatise handed down by none other

76. Lingat 1973, 98; cf. Doniger 1991, who views Manu as “composed in increments over several cen-
turies” (xliv–xlv); “a hotchpotch” (lv). It is not clear from this why she concludes her introduction calling it “this
extraordinary text” (lxviii). See chapter 1 § B on this, and her similar view of the Mahābhārata. I also do not follow
Bronkhorst’s attempt (in press) to isolate the often “lyrical” Manu 12 as late because, in his view, it is only there
that Manu discusses “rebirth as determined by karmic retribution” “to explain the superior status of Brahmins.”
The “possible exceptions” (Bronkhorst cites M 6.61–64,, 69–69,, and 10.42) to this “general rule” rather prove
that, as soon as one drops the requirement that the theory would apply only to Brahmins, there is no rule at all.
77. On Manu’s frame and the others to be discussed here, see, briefly, chapter 1 § B. The frame story “sug-
gests” such a break, even though the matter may be more complicated. As Olivelle indicates,“[i]t is evident that
the author of Manu conceived of his śāstra as a charter applicable to all and transcending the narrow boundaries
of vedic śākhās,” and that Kumarila “clearly articulated” a Mīmāṃsā position “[t]hat Manu is not limited to any
śākhā” (2005b, 272, 261 n. 2). Yet Jamison makes a case that Manu turns an adage from earlier ritual texts (“don’t
wake a sleeping fire”) into a law for snātakas (“don’t wake a superior”) from within the Maitrāyaṇīya śākhā of the
Yajurveda (2000, 121–24), challenging Lingat’s view that Manu is “fully detached” from a prior gṛhyasūtra tradi-
tion and that the evidence is lacking for Bühler’s notion of a prior Mānava Dharmasūtra. Cf. Bronkhorst 1985b,
also arguing for a prior Mānava dharma text from the Mānava school of the Maitrāyaṇī Saṃhitā; Biardeau 2002,
1: 66, also viewing such a link as apparently broken (semble rompu).
 post-vedic brahmanical dharma 209

than the creator god Svayaṃbhū. He taught it to his son Manu, who transmitted
it to his disciples, including Bhṛgu, who is made the spokesman and promulgator
within the treatise.”78 Manu leaves traces of its real author’s “pandit mentality.”79
But in presenting its author “as the primeval lawgiver, the Creator himself,” it
leaves no room for “debate, discussion, or scholarly give and take” (2004b, xxv).
Olivelle says the frame “fizzles out” at the end (2004b, xxv), and presents a
“disjunction” between the notion of the Creator Brahmā as the single source of
Law in chapter 1 and Manu’s discussion of four different sources of Law in
chapter 2 (xxvii). But the Creator’s divine authority is established from the
beginning, for the whole. Manu’s author is both a modest poet and a most
immodest “traditional pandit.” We can now see what is involved, at least for
Manu, in doing without an intertextual citational apparatus.
Now in having a direct line to Brahmā, Manu has interesting company in the
persons of the Buddha, Vālmīki, and Vyāsa. The Buddha, as we have seen, not only
tells Brahmins from time to time how to join the company of Brahma; Brahmā
appears soon after the Buddha’s enlightenment to prompt his first sermon by reas-
suring him that, although his enlightenment is indeed profound, “There will be
those who will understand” (see chapter 4 § B.1.b). At the beginning of the
Rāmāyaṇa, Brahmā gives Vālmīki the insight to see all that has happened and is
still yet to happen in Rāma’s life, and promises that Vālmīki’s poem will endure so
long as the rivers and mountains last on earth and that it will all be true (1.2.22–
36).80 Vyāsa also converses with Brahmā—though only belatedly—in a famous
double interpolation. First, in an insertion probably from before the fourth-century
CE, Brahmā appears to Vyāsa to vouch that his epic is a work of poetry (kāvya);
then, in twenty lines inserted probably several centuries later, Brahmā recom-
mends that Vyāsa call upon the elephant-headed god Gaṇeśa to be his scribe.81
The case of the Buddha suggests that Brahmā’s authorization would have to
do with an implied endorsement of the Buddhist dharma by the god of Vedic and
Brahmanical orthodoxy. And the two epics show that, beginning with the Rāmāyaṇa,
Brahmā authorizes and validates poetry. Putting these two points together we have
a suggestive combination for understanding Brahmā’s role in Manu. For Manu,

78. Olivelle 2005b, 272. Cf. Lingat 1973, 74: The dharmasūtras “profess to be nothing more than treatises
written by ordinary mortals and based on traditions of the sages,” but they are also, where deemed appropriate,
intent on differentiating their customs from those of the ancients.
79. See Olivelle 2005a, 29, citing, as seeming slippages, Manu’s (really Brahmā’s) occasional references
to others’ opinions and appeals to other sagely and divine authorities; cf. 311, n. to 8.110 concerning the author’s
lapse in citing Manu’s son Vasiṣṭha’s oath as an “historical episode”; Olivelle 2004b, xxvii.
80. Although the Rāmāyaṇa is called the ādikāvya or “first poem,” the term does not occur in the Baroda
Critical Edition. But it probably should since it occurs in a universally attested sarga where, after Sītā has vanished
into the earth, Brahmā encourages Rāma to hear the rest of this ādikāvya (7, Appendix I, no. 13, lines 31–39).
81. See Lüders 1929, 1144; Sullivan 1990a, 11, 118–19; Hiltebeitel 2008a, 206–9; Adluri 2010b.
210 dharma

Brahmā authorizes poetry as the god who has come to embody Veda or the Vedas,
and more specifically the utterances of Vedic poetry or mantra, including the primal
utterances (vyāhṛtis) by which Brahmā creates the triple world.82 What Manu adds
here is that, even more than authorizing poetry, Brahmā actually composes it and
imparts it to Manu who is now his son: their relation being a fitting one for a text
on law, and particularly on laws of inheritance, for which Manu has been an
authority since the Vedic Mantra Period.83 Thus while Brahmā prompts Manu to
transmit dharma in the form of Brahmā’s own poetry,84 he also prompts Vālmīki to
compose poetry that will be about a paragon of dharma, just as he comes to endorse
Vyāsa’s poetry about dharma in the Mahābhārata.85
These correspondences allow us to call attention to some other similar-
ities in the frames themselves. All three texts frontload their frames to initial
chapters, and although Manu could be said to allow its frame to fizzle out a bit
more than either of the epics does, there are reminders of it both near the
middle (5.1–4) and in the last adhyāya, both at its beginning (12.1–2) and its
end (12.117–26)—reminders that, as with those in the epics, leave it to the
attentive reader or listener to grasp how the frame remains relevant and
threaded into the text, or can even be expanded upon at major points. In each
work, the frame stories tell of the poem’s composition and transmission and
explain its ultimate outreach to a “universal” audience. The Mahābhārata has
three frame narratives, each of which focuses on the transmission of the story
to audiences that are increasingly remote from the main action, ultimately
taking us to the outer frame location of a celestial Naimiṣa Forest where we
overhear the celestial Ṛṣis listening to the Mahābhārata for the first time.86
The Rāmāyaṇa creates just the opposite effect. Instead of reaching universality

82. As Bühler points out, commentators also relate the figure of Manu back to four passages mentioning
“Vedic Mantras which Manu is said to have revealed or seen” ([1886] 1969, xvi; cf. lx).
83. At least since Baudhāyana Dharmasūtra 2.3.2, which quotes Taittirīya Saṃhitā 2.1.9.4, “Manu divided
his estate among his sons,” the dharma tradition has established Manu’s credentials as a lawgiver first of all in
the area of inheritance as a matter of śruti or Veda; see Hopkins 1882, 240–41; Bühler [1886] 1969, lxi. From
Bühler’s survey of Vedic mythemes surrounding Manu (lvii–lxii), others such as “father of mankind” and
“inventor of sacrificial rites,” including especially funeral sacrifices, could lend some support to the prominence
of inheritance in Manu’s pre-legal-tradition profile. Cf. Olivelle 2005a, 18, citing Taitt. Saṃh. 2.2.10.2 for “what
appears to have been a proverbial saying: ‘Whatever Manu has said is medicine.’ ”
84. Olivelle takes up these matters in relation to divine instruction in the Upaniṣads, and the likely
influence of Buddhism in presenting “the words of a single charismatic individual . . . as the sole fountain of
authority in a religious tradition” (2004b, xxv–xxvi).
85. On Vyāsa’s affinities with Brahmā, see Sullivan 1990a. On Brahmā and the Vedas, see Hiltebeitel
2005b, 2005c; on the Rāmāyaṇa inspiration for the Mahābhārata interpolation about Vyāsa and Brahmā, see
Shulman 2001, 33 n. 10.
86. On the Mahābhārata’s three frames and the otherworldliness of the Naimiṣa Forest in its outer frame,
see Hiltebeitel 2001a, 92–161; 2006a. On comparison of the two epics’ frame stories, see Hiltebeitel 2005a,
461–64.
 post-vedic brahmanical dharma 211

through increasing remoteness from the epic’s events, it conjures up a sense

of great immediacy in just one frame by having the poem addressed directly to
Rāma, the hero and universal “perfect man.” Moreover, once Sītā has entered
the earth, Rāma and other earthly listeners are joined, with Brahmā’s permis-
sion, by the celestial Ṛṣis who, in this case, come to an earthly Naimiṣa Forest
to hear the rest of the story (see Hiltebeitel 2001a, 285–86, 317–22; 2005a,
501–2). Meanwhile, Manu is transmitted from Brahmā to Manu to Manu’s
disciples, including Bhṛgu, whom we overhear recounting it to the gathered
celestial Ṛṣis.87 Vālmīki as poet (see Goldman 1976), Śaunaka as chief listener,
and Bhṛgu himself as chief transmitter are all Bhārgavas. In each case, the
poets themselves are present to hear their work first recited: Vālmīki to hear it
recited by Rāma’s twin sons and to vouch with finality for Sītā’s chastity even
as she descends forever into the earth; Vyāsa to hear it recited to Janamejaya
even though he is the latter’s ancestor by six generations; and Manu appar-
ently remains seated to listen to his son Bhṛgu address the assembled Ṛṣis
(M 1.1, 60, 119).
Moreover, if we follow Manu’s frame through its midpoint and final itera-
tions, we do find that it tides itself along through an important theme. In the
frame’s first resurgence, at a major subsection transition, the Ṛṣis, having
heard Bhṛgu describe “the Laws of a bath-graduate” (dharmān snātakasya) (5.1),
chime in with a brief question:

“How, O Lord, does Death (Mṛtyu) prevail over Brahmins, who know
the vedic teachings (vedaśāstravidām) and practice their svadharma
described in this manner?” Bhṛgu, the embodiment of the Law
(dharmātmā) and the son of Manu, said to those great seers: “Listen
to the fault because of which Death seeks to kill Brahmins. Death
seeks to kill Brahmins because of the failure to recite the Vedas, the
dereliction of the rules of proper conduct, laziness, and faults with
respect to food. . . .” (5.2–4)88

The Ṛṣis’ question about the svadharma of Brahmins seems to follow up on

what is said shortly before this, near the end of chapter 4, where dharma alone,
without the prefixed sva-, has been mentioned as the “merit” one should accu-
mulate to be one’s “escort” or companion (sahāya) to the other world:

87. See also M 7.28–39, about the boomerang powers of the personified daṇḍa, Punishment, that the king
should administer only properly, mentioning a midspace cosmological Muni setting that could remind Bhṛgu
and his Ṛṣi/Muni audience of their own setting: “Punishment . . . oppresses the fort, the realm, and the mobile
and immobile world, as well as the sages and gods dwelling in mid-space (antarikṣagatāṃś caiva munīn devāṃś ca
pīḍayet).” Olivelle does not consider this passage to be among his “excurses”; see further Biardeau 2002, 1: 85.
88. Cf. Olivelle trans., 2005a, 138. I thank Greg Bailey for suggestions on this verse.
212 dharma

Gradually and without hurting any creature, he should pile up merit

(dharma) like termites an anthill, so as to secure an escort in the next
world; for in the next world, neither father nor mother stands by him
as his escort; nor does son, wife, or relative. Only merit stands by
him (dharmas tiṣṭhati kevalaḥ).89 Alone a creature is born, and alone it
dies. Alone it enjoys the fruits of its good deeds, alone the fruits of its
evil deeds. While his relatives discard the dead body on earth as if it
were a piece of wood or a clod of earth, and depart with averted faces,
his merit accompanies him (dharmas tam anugacchati). To secure an
escort, therefore, let him gradually pile up merit every day; for with
merit as his escort, he will cross over the darkness that is difficult to
cross. It quickly leads that man, who is devoted to the Law and whose
sins have been erased by ascetic toil, to the next world, glittering with
an ethereal body. (4.238–43, Olivelle 2005a, 136 slightly modified)

Note that Olivelle translates dharma in two ways here: as “merit,” where the met-
aphor of “piling up” more or less requires it, up to the point that “Law” better
describes what a man is devoted to who has secured his escort of merits that
leads him to the other world. Manu is thus returning to these themes, and to its
frame story, when the Ṛsị s ask to learn from Bhṛgu about the contrary situation
where Brahmins practice their “own dharma” yet succumb to death because they
have a countervailing “fault” or “faults” (doṣas). Here it seems that the dharma in
the Brahmin’s svadharma has shaded over into the meaning “merit.”
Manu reiterates this set of themes at several more points, each carefully
nuanced for those in different walks of life, before we meet it one last time at
the end with the final closing-out of the frame. For the wandering ascetic in the
fourth āśrama or “life-stage,” “self” supersedes “merits” as his only “escort” or
“companion”:

Taking delight in what pertains to the self, he should remain seated

without longings or sensual attachments. With himself90 as his only
companion (sahāya), he should walk about here, seeking felicity.
(6.49)

For the king, however:

Justice (dharma) is the only friend (suhṛd) who follows a man in

death; for all else perishes along with the body. (8.17)

89. Cf. Doniger’s misleading rendering: one is unhurtful to other creatures so that they will be one’s
“companions” in the other world, where “religion alone endures” and “follows after him” (1991, 96).
90. Or “the self” (ātmaiva).
 post-vedic brahmanical dharma 213

Note that for a king, the “friend” or “wellwisher” (suhṛd)91 replaces the “escort,”
and dharma is now a matter of “justice,” which the context clearly warrants
since it is a question of the “dharma bull” that the king’s justice can either
impede or uphold (8.16–19). Then, as part of a surprising speech that Manu
puts into the mouth of a judge exhorting truth from witnesses of all four classes,
the judge should say:

Whatever good deeds you have done since birth, dear man, all that
will go to the dogs, if you testify dishonestly. “I am all alone”—should
you think like that about yourself, good man; there dwells always in
your heart this sage (muni), who observes your good and evil deeds.
This god, Yama the son of Vivasvat, dwells in your heart. If you have
no quarrel with him, then you do not have to go to the Ganges or the
Kuru land. (8.90–91)

Here, with Manu bringing together a concatenation of ideas similar to what

one finds in Naciketas’ dialogue with Yama in the Kaṭha Upaniṣad,92 it is now
the Muni Yama as the inner witness in the heart who replaces the inner escort93
and its other surrogates. It continues to say that “if the witness has no quarrel
with Yama, the god of death and the judge of the dead, then there is no need for
him to visit the Ganges or the land of the Kurus to expiate his sin” (Olivelle
2005a, 310, paraphrasing 8.92). Actually, it only says he need not visit “the
Kurus” (kurūn), which may hold some interest in terms of ideas of “Kuru
orthopraxy” that would seem to have taken hold in both Buddhist (see chapter
4 § B) and Brahmanical circles. The Mahābhārata gives some strikingly similar
ideas about the inner witness to the heroes’ dynastic ancestress Śakuntalā when
she demands that King Duṣyanta attest honestly in his royal court, which will
become the court of the Kurus,94 as to his paternity of their son Bharata (Mbh
1.68.25–32).

91. See chapter 12 § D on friendship terms. Before this, suhṛd is used only at 7.32, with respect to the king:
“Within the realm, he should act in accordance with the rules; upon his enemies, he should impose harsh pun-
ishments; toward his suhṛds and loved ones, he should behave without guile; and to Brahmins he should show
compassion.” After 8.17, it is still an affair of the king at 8.335 and 9.294 (the suhṛd as “ally” is one of the seven
“constituents” (prakṛtis) of the realm). Only later at 11.57 and 89 is it used in other contexts.
92. See chapter 3 § F: Naciketas asks Death (Yama, Mṛtyu) for the “subtle dharma” on life after death, and
learns about the mystery of the “primeval one” hidden in the impenetrable depth of “cave of the heart.”
93. Note that just before this at 8.86, along with Yama, Dharma is among the cosmic beings and gods
who knows a witness’s truth; as does the self, a kind of conscience (8.84)—the inner self as witness. Yet this
rebirth cosmology still recalls the Vedic fetters of Varuṇa. Cf. 8.173: “Like Yama, therefore, the king should lay
aside his own likes and dislikes and follow Yama’s pattern of behavior, suppressing anger and mastering his
organs”; 9.307: as Yama holds both friend and foe alike in his grip, he holds his subjects in his grip through his
“Yama-vow.”
94. See Brodbeck 2009a, 133–50; Hiltebeitel forthcoming-d. During Duṣyanta’s reign, it is still yet to
become the “Kuru” capital because Kuru is a later eponym in the Paurava-Bhārata-Kuru/Kaurava dynastic line.
214 dharma

Finally, speaking through Bhṛgu, Manu seems to tie all these themes
together in a quasi-Upaniṣadic closing-out of the frame and the entire text.95
First, we are brought back to the frame:

In this manner, the blessed god, desiring to do what is beneficial for

the people,96 revealed to me in its entirety this highest secret of the Law
(dharmasya paramaṃ guhyam). With a collected mind, a man should
see in the self everything, both the existent and the non-existent; for
when he sees everything in the self, he will not turn his mind to what
is contrary to the Law. All the deities are simply the self, the whole
world abides within the self; for the self gives rise to engagement in
action (karmayoga)97 on the part of these embodied beings. (12.117–19)

Then, after Manu recommends a closing meditational exercise that involves depos-
iting the five elements in aspects of one’s own person and various gods in the mental
and physical organs that orient action (120–22),98 he brings the text to closure on
ways of grasping the supreme Puruṣa, by any one of which a man many attain the
highest—among them, what I would like to call the “subtle inner Manu”:

The ruler of all, more minute than even an atom, resplendent like
gold, and to be grasped by the sleeping mind—he should know him
as the supreme Puruṣa. Some call him Agni, some Manu the
Prajāpati, others Indra, still others Breath, and yet others the eternal
Brahman. This one, pervading all beings by means of the five forms,
makes them go around like a wheel through birth, growth, and death.
When a man thus sees by the self all beings as the self, he becomes
equal towards all and reaches Brahman, the highest state. When a
twice-born recites this Treatise of Manu proclaimed by Bhṛgu, he will
always follow the proper conduct and obtain whatever he desires.
(123–26, slightly modifying Olivelle 2005a, 236)

It turns out that Manu’s frame fizzles out quite lyrically by construing Manu
himself and his Treatise (śāstra) as possibilities for giving definition to one’s
inner escort, inner Muni, inner author, inner text, and inner witness. Manu
pitches his poetry to get dharma under one’s skin.

95. Olivelle, I think quite unnecessarily (if he means by it “interpolation”), regards this as an “excursus”
(2005a, 236). But see chapter 1 § B on my view of such excurses.
96. Or, “for the welfare of the worlds” (lokānāṃ hitakāmyayā; 12.117b).
97. I return to this passage in chapter 11 § C on Manu’s different view of karmayoga from that in the
Bhagavad Gītā.
98. It clearly involves the action orientation of a Brahmin householder, building up to “Viṣṇu in his stride,
Hari [or Hara] in his strength, Fire in his speech, Mitra in his organ of evacuation, and Prajāpati in his organ of
procreation” (12.121).
 post-vedic brahmanical dharma 215

E. Varṇa (Caste), Āśrama (Life Pattern), the King, Śūdras,

and Women

If one sought to inculcate a body of behavioral norms, roles, and rules that
should govern each person’s social life in all its transactions across the spectrum
of social classes and likewise each person’s individual life through all its
options, and further have it calibrated to cover not only individuals in every
social class but every “intermediary” or “mixed” class, one would come up with
something like the theory of varṇāśramadharma, “the Law of caste (varṇa) and
life-stage (āśrama),” as it is usually translated, which textbooks on Hinduism
are prone to attribute to Manu. Since the social classes (varṇa also means “cat-
egory” and “color”), class mixture (varṇasaṃkara), and the life-stages each have
their own calculus, which becomes virtually infinite when the three are interre-
lated, we must continue to treat facets of them in other sections of this chapter
and book. This section will open on a few limiting points, and then explore how
Manu makes the laws of varṇa and āśrama so focal in relation to our other
classical dharma texts.
Regarding the four classes and the mixed classes, Manu and the
Mahābhārata are adamant on one limiting point: there is no fifth varṇa. None
of the dharmasūtras make such a point. The Mahābhārata has the line
“Tradition allows four classes, a fifth is not met (smṛtā varṇāś ca catvāraḥ pañ-
camo nādhigamyate)” (13.47.18), which comes amid discussion of the different
amounts to be inherited by sons of a Brahmin born from wives of the four
classes. Sons born from wives of the three twice-born classes inherit amounts
that decline according to the wife’s rank, even though all three types of sons
would have the Brahmin’s status; but the son of a Śūdra wife does not have
Brahmin status and should receive a small inheritance only optionally, and
only out of “non-cruelty,” “compassion,” or “kindness” (ānṛśaṃsya; 20).99
Obviously, a Brahmin’s inheritance is protected against claims made by sons
of uncategorized wives. This discussion of inheritance sets the stage for one of
mixed classes in the next adhyāya (13.48), in which there are some similarities
to verses at the beginning of Manu 10.100 It is there that Manu takes his stand
that there is “no fifth” while introducing mixed classes as the opening topic of
his tenth chapter on “Rules for Times of Adversity” (Olivelle 2005a, 205):

99. I note several translations of this value in anticipation of discussing it further below and in chapter 9
as a major value in the Mahābhārata, particularly for the Pāṇḍava king Yudhiṣṭhira, who is in fact being addressed
at this point by his “grandfather” Bhīṣma.
100. Dandekar 1996, 1075 notes several of these.
216 dharma

Devoted to their respective activities (svakarmasthā), the three

twice-born classes should study the Veda;101 but it is the Brahmin
who should teach them, not the other two—that is the firm principle.
The Brahmin must know the means of livelihood (vṛttyupāyān) of all
according to rule, and he should teach them to others and follow
them himself. Because of his distinctive qualities, the eminence of
his origin, his holding (dhāraṇāt) of restrictive practices, and the
distinctive nature of his sacrament (saṃskārasya), the Brahmin is the
lord of all the classes. Three classes—Brahmin, Kṣatriya, and
Vaiśya—are twice-born; the fourth, Śūdra, has a single birth. There is
no fifth. (M 10.1–4; Olivelle trans. 2005a, 208, slightly modified)

Presumably, such an insistence follows from an ongoing reading of the

Puruṣasūkta, which mentions only four primal classes, and may be a way of
disallowing claims to the contrary found in some Buddhist texts such as the
Ambaṭṭha Sutta.102 But more directly, it reinforces the explanatory power of the
theory of mixed castes, which explains the origins of all “subcastes” (jātis,103 lit-
erally “births”)—including even Greeks (Yavanas) and outcastes—as the post-
cosmogonic outcome of the different possibilities of varṇa miscegenation.104
Note also that the opening verse concerns svakarma rather than svadharma: the
former, as we shall see in chapter 11, is the greater of the two concerns as
regards mixed classes. As Manu works out the calculus of mixed classes to new
and advanced degrees, it is also interesting that even while he focuses his
geographical attention only on the northern plains,105 he mentions “subcastes”
that extend to the north and south, mentioning as fallen “Vrātya” Kṣatriyas, for
instance, two of the “oligarchic” tribes or gaṇa-saṅgha “republics” of Videha
known to the Buddha106—the Licchivis and Mallas—along with the southern
Draviḍas, all in one verse (10.22). As Olivelle points out, the prior mention of
Yavanas in such a context by Gautama probably helps to date the “mixed caste”

101. The verb adhīyīran (“they should study,” “they should recite”) implies the Veda, but it is only presum-
ably what the Brahmins should teach—dharma being another possibility, as the passage would suggest, or better,
they should teach Veda implying also dharma. Presuming that it refers to teaching Veda, Bühler notes that M
2.241–42 allows that in times of adversity “a Brâhmaṇa may learn the Veda from a non-Brâhmaṇical teacher, and
that hence this rule is not absolute” ([1886] 1969, 401–2). See also 8.390–91, discussed below, which seems to
allow that a king can “teach” (Olivelle 2005a, 188), “establish,” or “settle” (pratipādayet) the svadharma of those in
āśramas (whether “hermitages” or “life stages”).
102. See chapter 4 § A. I am told that some Jātakas also mention a fifth caste.
103. Although as Biardeau 2002, 1: 93 observes, “the term jāti makes only a discrete appearance” at M
10.26, 27.
104. Cf. Mbh 12.285.5–7 differentiating the four varṇas born from Prajāpati from all others born from
miscegenation.
105. See above n. 34. What makes for the “conduct of the good” in Brahmāvarta is that it is “handed down
from generation to generation among the social classes and intermediate classes of that land” (M 2.18).
106. See Bühler [1886] 1969, 406, n. to verse 22. On gaṇa-saṅghas see chapter 4 § C.2.
 post-vedic brahmanical dharma 217

theory, and also to date Gautama relative to Āpastamba, since Āpastamba “does
not deal with mixed classes at all, a topic found in all the other Dharmasūtras
and the later Smṛtis” (1999, xxxi; cf. Jha 1970). As Biardeau puts it, “on the
plane of the imaginary is constructed the complex tissue of a vast society in
mutation” (2002, 1: 93). Her example behind this statement is Manu’s treatment
of the sairandhra, “skilled at adorning and personal attendance” (10.32), as a
mixed caste—fabricated, she suggests, as an explanation for Draupadī’s dis-
guise as a Sairandhrī or hairdresser in the fourth book of the Mahābhārata.
As to the āśramas, the main limiting matters were well underscored in a
groundbreaking study by Olivelle (1993). In brief, the Mahābhārata knows
the āśramas in two ways. It knows them in their “original system” (153–55)
familiar to the dharmasūtras, where they begin very much in flux even by
name and sequence,107 and under some disapproval,108 as four different life-
long choices (vikalpa) to be made before marriage.109 And it knows them in
their “classical system” (148–51) favored (though not exclusively) by Manu
(129), which staggers the four through a male’s life. Where the Mahābhārata
advocates the classical system, however, it does not do so in quite the same
way as Manu. On Manu’s usage, Olivelle offers an important note to Manu
1.114ab, a line in Manu’s table of contents or “synopsis.” This line reads:
strīdharmayogaṃ tāpasyaṃ mokṣaṃ saṃnyāsameva ca; and Olivelle translates it
as follows, while inserting the chapter-and-verse numbers where Manu
addresses these topics:

Law pertaining to women [5.111–45]. Hermit’s life [6.1–32]. Renunciation*

[6.33–85]. Retirement* [6.87–96]. (Olivelle 2005a, 92, 401)

The asterisks after “Renuciation” and “Retirement” direct us to a footnote:

Renuciation (mokṣa), Retirement (saṃnyāsa): the Sanskrit term mokṣa

literally means liberation. Manu, however, attaches a technical

107. Ᾱpastamba treats what is to become the fourth stage (Manu’s vedasaṃnyāsin or “vedic retiree” [M 8.86])
before what is to become the third, calling the two pravrāja, “wanderer” (2.21.7) and vānaprastha, the eventually
conventional name for “forest-dweller” (2.21.18), in that order; G calls the eventual fourth bhikṣu, possibly
reminded of Buddhism, and likewise discusses him before vaikhānasa, “anchorite,” his name for the eventual
third (3.2, 11–35). B (2.11.14–26) and V (9.1, 10.1) both use vānaprastha and parivrājaka in that order. Cf. Biardeau
2002, 1: 70–71, 74, 81, 91–93 on the āśramas in Gautama, Baudhāyana, Ᾱpastamba, and Manu respectively.
108. See Ᾱ 2.23.3–24.14 presenting his view of non-Vedic claims to supernatural powers; G 3.3, 3.36;
B 2.11.27, favoring only the householder (see Olivelle 1999, 374 nn. and 1993, 83–91 on “this early view of
conservative Brahmans”); B 2.11.28, attributing to some the idea that the āśrama divisions were created by
the demon (Asura) Kapila, son of Prahlāda. On these and other references, see Olivelle 1993, 90–99. On
Kapila’s “Greater Magadha” associations, see Bronkhorst 2007, 61–68.
109. V 7.3–5 upholds such choice only on condition that one has completed the student stage (brahma-
carya) and its Veda study without breaking his vow of chastity, while ruling that a student should serve his teacher
until death.
218 dharma

meaning to the term, using it as a synonym of renunciation and the

fourth order of life dedicated exclusively to the search after personal
liberation. . . . Manu makes a clear distinction between this
renunciatory asceticism and the life of a vedic retiree, which he
designates as saṃnyāsa. (Olivelle 2005a, 243)

In an earlier article, Olivelle already touches on Manu 1.114’s differentiation of

mokṣa as “renunciation”110 from saṃnyāsa, or more specifically “the life-style of
the vedasaṃnyāsika that Manu calls saṃnyāsa,” which, at 6.86–96 involves the
abandonment of ritual activity incumbent on a householder (1981 270–71).
Olivelle also shows that in contrast to Manu’s carving out of this technical
“vedic retiree” usage to insist on doing the four āśramas in sequence, the
Mahābhārata is one of just a few texts to introduce what Olivelle calls “the
classical meaning” of saṃnyāsa, in which “saṃnyāsin is commonly used as a
synonym of such terms as parivrājaka, pravrajita, śramaṇa, bhikṣu, and yati”
(265), and in which it is devalued at least for active warriors and kings in both
the Bhagavad Gītā and the Śāntiparvan (268, 272).
I believe Olivelle raises intriguing possibilities in positioning the
Mahābhārata among the earliest texts to have innovated in introducing the gen-
eralized classical usage of saṃnyāsa. But what is ignored in his earlier discussion
is that the Mahābhārata also airs the preclassical system, particularly doing so
in its story of Vyāsa’s firstborn son Śuka, which Olivelle, eleven years later, calls
“the most straightforward presentation of the original [i.e., preclassical āśrama]
system” (1993, 154; cf. 104, 153–55). This is because the Śuka story confirms
that the 25-year-old (Mbh 12.309.62b) Śuka can skip the full sequence of the
four āśramas and seek mokṣa directly from the first āśrama, that is, from brah-
macarya, without marrying, and above all, without waiting for the fourth
āśrama. If the Śuka story presents the preclassical system in conjunction with
questions pro and con about the classical system, this does not encourage the
view that the Mahābhārata’s innovative treatment of the classical system would
itself, in isolation, be late, as Olivelle, at least as of 1982, proposes.111 More
likely, I believe, it just takes a while for the more strictly legal texts to catch up
with the Mahābhārata. I think that in airing both systems, the Mahābhārata
keeps them under debate such as Olivelle himself mentions (1993, 69–70),
taking them up in some harmony with their treatment in the dharmasūtras and

110. Bronkhorst 2010b and in press raises questions as to whether Olivelle’s construal of mokṣa in this
fashion befits Manu. I believe it does, and also befits what the Mahābhārata has to say about mokṣa in the closing
units of the Mokṣadharma Parvan. I will discuss this matter in chapter 13.
111. Olivelle dates the Mahābhārata here later than the Rāmāyaṇa (1982, 267–68, 272 n. 47, 273), and,
on the “classical meaning” as found “especially the Śāntiparvan and the Anugītā, concludes, “We would not be
far wrong in placing this final semantic development of S[aṃnyāsa] around the 3rd–4th century A.D.” (274).
 post-vedic brahmanical dharma 219

probably soon before Manu further codifies them. Unlike the dharmasūtras and
the Mahābhārata, Manu seeks energetically to suppress the pro-choice position
(131–36, 147, 176).
Whatever their views of the other āśramas and the total system, however, all
of these texts, including the Rāmāyaṇa even in its only verse that clearly references
the āśramas,112 agree,113 or at least suggest,114 that the householder life is the best,
mentioning such reasons for this as its being the only one prescribed in the
Vedas, that it supports them all, and that it alone enables socially responsible
human reproduction. As we saw at the end of chapter 4, the Buddha is portrayed
as countering such a view of the householder’s primacy in the Aggañña Sutta.
Coming now to the crediting of varṇāśramadharma to Manu, the Manu
Smṛti never actually mentions the term itself, nor, for that matter, does its first
main successor, the Yājñavalkya Smṛti. As far as these texts are concerned, the
term seems to have emerged among commentators on them,115 who classify
varṇāśramadharma as one of the five or six topics that such Smṛtis cover in
their treatment of dharma: (a) varṇadharma (“injunctions based on varṇa
alone”—i.e., on class or caste), (b) āśramadharma (injunctions concerned
directly with life-stage behaviors), (c) varṇāśramadharma (rules concerned with
their points of intersection), (d) guṇadharma (virtues or duties incident to
personal qualities “such as protection of subjects in the case of a crowned
king”), (e) nimittadharma (secondary or occasional duties “such as expiation on
doing what is forbidden”), and ( f ) sādhāraṇadharma (virtues “common to all
humanity viz., ahiṃsā [non-violence] and other virtues”).116 Such a list of
extracted topics is instructive for its commentarial indication that, prior to the
joint consideration of varṇāśramadharma as topic three, varṇa- and āśrama-
dharma are recognized as each having separate status on their own.
P. V. Kane says, “It will be noticed from the above that all matters (except
sādhāraṇa or sāmānya dharma) have varṇa and āśrama as the pivots round which

112. Rām 2.98.58, in which Bharata uses the adage among his arguments to try to convince his exiled
brother Rāma to return home and rule. Olivelle, who first called attention to the Rāmāyaṇa’s near silence on the
subject (1993, 18 n. 46; 103), suggests the verse would be an interpolation, but his pre-fifth-century BCE date for
the Rāmāyaṇa is too early a rationale for this.
113. G 3.3, 36; B 2.11.27; V 8.14–16; Mbh 12.12.11; 18.27–28; 23.4–5; and 61.15, in which four different
people make this point to Yudhiṣṭḥira early in his postwar miasma; M 6.87–90; Rām 2.98.58 as just noted.
114. Ᾱpastamba says only that, based on the Veda, rites requiring the wife are superior to what can be
achieved by yogic powers achieved in other āśramas (2.23.10), and that “immortality consists in offspring” (24.1).
115. See Kane 1962–75, 1: 2–3 and Vidyarnava and Panshikar 2003, 3–4), from which I take the following
list. According to Kane, Hemādri mentions all six, citing sixteen verses from the Bhaviṣya Purāṇa on them, and
Medhātithi mentions the first five. Vidyarnava translates Vijñāneśvara’s Mitākṣarā, which includes all six in
commenting on Yājñavalkya Smṛti 1.1. Another early usage occurs in inscriptions, where even Buddhist kings are
called “protectors of varṇāśramadharma” in sixth-century inscriptions (Olivelle 1993, 201–4).
116. The parenthesized quotations are from Kane 1962–75, 1: 2–3.
220 dharma

the whole of dharmaśāstra revolves. It is therefore that in ancient smṛtis like

those of Manu (1.2 and 107) and Yājñavalkya (1.1) the sages are represented as
asking the great expounders of these codes to impart to them instruction in the
dharmas of varṇas and āśramas” (1962–75, 1: 3). Yet Kane fudges here. Even
sādhāraṇa or sāmānya dharmas are found to revolve around caste and life-stage.
In the Mahābhārata, which seems to have originated the concept of sādhāraṇa
dharmas, they are mentioned twice in the Mokṣadharma section of the Śāntiparvan
as “Upaniṣadic” virtues linked with the fourth life-stage (12.236.15ab; 262.27cd).
And another Mokṣadharmaparvan unit says they are common to the upper three
varṇas and to Śūdras who follow the “conduct of the good” (12.285.22c–34).
Asked by King Janaka to contrast the dharmas distinctive to each caste
(viśeṣadharmas) and the sāmānya dharmas applicable to all castes, the sage
Parāśara lists an odd assortment of thirteen sādhāraṇa ones117 in the following
order: noncruelty (ānṛśaṃsya), nonviolence (ahiṃsā), vigilance, sharing, doing
ancestral rites, hospitality, truth, non-anger, contentment with one’s own wives
(sic), purity, constant unspitefulness, self-knowledge, and forbearance (23–24).
Kane is also a bit casual in stating that Manu and Yājñavalkya begin with
the sages asking for “instruction in the dharmas of varṇas and āśramas.” It is
true that Yājñavalkya 1.1—in the text’s only verse mentioning varṇa and āśrama
together118—has the sages ask Yājñavalkya, “Tell us completely the Dharmas of
classes (varṇas), of orders (āśramas) and of others.”119 But in the two verses that
Kane cites from Manu, the sages ask only about the classes, first (as in
Yājñavalkya) along with the mixed classes:

Please, Lord, tell us precisely and in the proper order the Laws
(dharmān) of all the social classes (sarvavarṇānām) as well as those
born in between (antaraprabhavānām). (1.2)

In this, the Law (dharma) has been set forth in full—the good and the
bad qualities of actions and the timeless norms of social conduct—
for all four social classes. (1.107)

Unlike the Ṛṣis at the beginning of Yājñavalkya, Manu’s Ṛṣis are yet to fuse the
āśramas into their text-opening preoccupation with straight and mixed classes
(cf. Olivelle 1993, 135). How Manu gets to mentioning varṇa and āśrama
together might thus be revealing.

117. Both terms are used for dharma or dharmas common to all or “held in common.”
118. As determined from the electronic version prepared by Tokunaga 1991.
119. “Others,” according to the Mitākṣarā commentary, refers to mixed classes; Vidyarnava and Panshikar
2003, 3.
 post-vedic brahmanical dharma 221

Manu mentions varṇa and āśrama jointly only twice, but both times
resonantly: the first with some buildup and the second with global finality. This
second usage comes in Manu’s last chapter in a verse claiming Vedic validity
for both schemes in an all-encompassing vast cosmological setting:

The four social classes (varṇas), the three worlds, and the four orders
of life (āśramas), the past, the present, and the future—all are
individually established by the Veda.120

Noticeably absent are the mixed classes. One wonders if this could be a know-
ing acknowledgment of their absence in the Puruṣasūkta, which could account
for everything in this verse except the four āśramas, whose Vedic status is highly
dubious.121
For our purposes, however, the verse where Manu first joins varṇa and
āśrama is even more revealing. As we have just seen, varṇa is a topic from
Manu’s very beginning, and remains so virtually throughout. But āśrama
emerges only intermittently until it is taken up as the main topic of chapter 6—
the only chapter in which varṇa is not mentioned. Their first joint mention
comes only after this, early in chapter 7, amid the launching of “the Laws per-
taining to kings” (7.1; rājadharmān), to which, as Olivelle has put it, Manu will
devote three “innovative” chapters (7–9):

The king was created (sṛṣṭaḥ) as the protector (abhirakṣitā) of people

belonging to all social classes and orders of life (varṇānām
āśramānāṃ ca) who, according to their rank,122 are devoted to the Law
specific to them (sve sve dharme niviṣṭānām). (M 7.35)

Clearly, Manu introduces the joint topic of “caste and life-stage” as belonging to
his expanded coverage of the king; and note that with the verse’s emphatically
doubled sve sve dharme, with which it actually begins, we have a strong indication
that the notion of svadharma is “henceforth” meant to be complicated by
consideration—indeed, calibration—of the countless ways that varṇa and āśrama
would intersect in a kingdom.123 Moreover, as Olivelle notices, with its use of sṛsṭ̣ aḥ
referring back to the king’s having been created (asṛjat) by Brahmā (prabhu) at 7.3,

120. M 12.97. See Olivelle 2005a, 349, noting that some commentators take prasidhyati vedāt to mean
“known from the Veda.”
121. See Olivelle 1996 (= 2005b, 53–74) showing that exegetical reliance on Chāndogya Upaniṣad 2.23.1 to
anchor the āśrama system in a Vedic text is not convincing.
122. Anupūrvaśaḥ, which Olivelle normally translates with “order” rather than “rank”—perhaps in this
case avoiding the double use of “order,” which appears in his translation of āśrama. “Rank” would initially seem
more suitable for varṇa than āśrama, but insofar as varṇa subsumes āśrama, “rank” is a useful rendering here.
123. Curiously, Manu’s other two usages of sve sve, both likewise conveying virtually infinite multiplicity
amid social complexity, are with karmaṇi, that is, karma, “activity”: 8.42, on how men should act far away from
home, and 9.262, on how criminal activities should be publicized and punished.
222 dharma

“this verse concludes the section on the creation of the king to be the protector of
the people” (2005a, 294), for the next verse marks the transition to a new topic, a
day in the life of the king that we will discuss in section G of this chapter.
Our current adhyāya 7 thus begins with a celebrated passage on the creation
of the king by “extracting eternal particles (mātrā nirhṛtya śāśvāvatīḥ)” from
eight deities associated with the cardinal and intermediate directions, while in
his grace (prasāda)124 and anger (krodha) the king is also respectively Padma,
the Lotus goddess of prosperity, and Mṛtyu, Death (7.3–12). The passage has
often brought up the question of whether Manu holds the king to be divine,
which Manu admits to in a qualified way: “it is a great deity [devatā] who stands
here in human form” (10cd). Biardeau suggests good caution before reading
too much into this passage: “scholars who have wanted to see in this creation
of the king from eight divinities an expression of the idea that the king himself
is divine have certainly gone too far. The king is certainly not a god, but one
wishes to render him respectable to all. If there are men who are gods on earth,
it is the Brahmins, and they are so only in a metaphoric sense.”125
It is following this creation of the king that Manu builds up to his first joint
mention of varṇa and āśrama by entailing both of them in a most intense
passage about the birth also of Daṇḍa, Punishment personified. Daṇḍa is
“created” or “issued” (asṛjat) just like the king, and possibly before (pūrvam)
him, which would make him the king’s older brother:

For his [the king’s] sake (tasyārthe), the Lord (Īśvara) formerly created
(asṛjat pūrvam) Punishment, his son—the Law and protector of all
beings—made from the energy (tejas) of Brahman.126 It is the fear of
him that makes all beings, both the mobile and the immobile, accede
to being used and they do not deviate from the Laws proper to them
(svadharmān na calanti ca).127 The king should administer him
appropriately on men who behave improperly, after examining
truthfully the place and the time, as well as their strength and

124. Olivelle 2005a, 154 prefers a more Confucian “benevolence.” Prasāda, translated by Pollock as “grace”
(Rām 3.3.21), occurs in an episode that Pollock describes under the heading of “the liberating power of the king”
(1991, 50–51, 92).
125. Biardeau 2002, 1: 86, citing Kane 1962–75, 3: 23 ff. as regards the metaphoric divinity of Brahmins.
Cf. Pollock 1991, 42–54, 63–67, 71–74; 1986, 21–24, reading “Rāma’s divinity” against a background that would
seem to include this passage from Manu (see Pollock 1991, 64–65; 300 n. to Rām 3.38.12), which, like others he
cites, does not mention Viṣṇu among the deities contributing to the king’s makeup. On the basis of such pas-
sages, Pollock suggests, I think unconvincingly (see Hiltebeitel 2003, 124–28), that Rāma’s association with the
“avatāra theology” would have occurred only gradually (52).
126. Or “of Brahmā” (brahmatejomayam). The “issuing” (÷sṛj) mode of creation is Brahmā’s, for whom the
name Īśvara is nothing uncommon.
127. M 7.15; On this “deviation” (÷cal), cf. Gautama Dharmasūtra 11.9–11. See Olivelle 2005a, 293 on
bhogāya kalpante in this verse, which could also mean “being enjoyed” and “being eaten.” The theme is extended
in the same terms to the whole universe at 7.22–23, elided in this citation.
 post-vedic brahmanical dharma 223

learning (śaktiṃ ca vidyām). Punishment is the king; he is the male

(sa rājā puruṣo daṇḍaḥ);128 he is the leader (netā); he is the ruler
(śāsitā); tradition tells us, he stands as the surety for the Law with
respect to the four orders of life [the four āśramas]. Punishment
disciplines all the subjects. Punishment alone protects them, and
Punishment watches over them as they sleep—Punishment is the
Law, the wise declare. When he is wielded (dhṛ, or held) properly
after careful examination, he gives delight to all the subjects; but
when he is administered without careful examination, he wreaks
total havoc. If the king fails to administer Punishment tirelessly on
those who ought to be punished, the stronger would grill the weak
like fish on a spit (śūle); crows would devour the sacrificial cakes;
dogs would lap up the sacrificial offerings; no one would have any
right of ownership (svāṃyam); and everything would turn topsy-turvy
(adharottaram). . . . All the social classes [the four varṇas] would
become corrupted, all the boundaries129 would be breached; there
would be revolt (prakopa) of all the people as a result of blunders
committed with respect to Punishment. Wherever Punishment,
dark-hued and red-eyed (śyāmo lohitākṣo), prowls about as the slayer
of evil-doers, there the subjects do not go astray—so long as the
administrator130 ascertains correctly. The proper administrator of
Punishment, they say, is a king who speaks the truth, acts after
careful examination, is wise, and has a masterly grasp of Law,
Wealth, and Pleasure.131 . . . For Punishment is immense energy,
and it cannot be wielded (dhṛ) by those with uncultivated selves
(akṛtātmabhiḥ). It assuredly slays a king who deviates from the Law,
along with his relatives; then he [Punishment] oppresses (pīḍayet) the
fort, the realm, and the mobile and immobile world, as well as the
sages and gods dwelling in mid-space. (7.14–21, 24–26, 28–29,
slightly modifying Olivelle trans. 2005a, 154–55)

Two points to begin with. The verse about dark-hued red-eyed Punishment
prowling about (7.25) is identical in all but one word to a passage in the
Mahābhārata (12.15.11), where it is spoken to fortify the clemently inclined
King Yudhiṣṭhira by his harsher younger brother Arjuna. And mid-space

128. Bühler’s parentheses are useful here: “Punishment is (in reality) the king (and) the male” ([1886]
1969, 219). Cf. Biardeau 2002, 1: 89: “the king is no longer king: the true king is his daṇḍa.”
129. Setavaḥ; alternately, “bridges,” “dams.”
130. Although there is no pronoun, Olivelle 2005a, 155 has “its administrator,” noting (294) that netṛ must
refer here to the king as Daṇḍa’s administrator, and that it was translated as “leader” for Daṇḍa himself at 7.17.
131. The three “aims of life” or trivarga (translated by Olivlle as the “triple set”; see § G for discussion), by
which the Punishment-wielding king is said to flourish in the next verse.
224 dharma

(the antarikṣa) would be where the celestial Ṛṣis or Munis have gathered as an
interested party to listen to Bhṛgu recite this text, which, in this passage, also
gets us quickly to the Ṛṣis’ earthly counterparts, the Brahmins, who are
another interested party. For two verses further along, one learns that the
Punishment-wielding king should “have good assistants” (susahāyena, 7.31)—
that is, above all Brahmins. As preamble to the day in the life of the king,
Manu thus creates high drama for its audiences, for whom the destructive
boomerang effects of Punishment extend from here below—beginning with
the fort (durga) that a king should create as the capital of his kingdom—on up
to the gods and Ṛṣis.
As Olivelle observes (2005a, 294), by mentioning “the place and the time”
and urging “careful examination,” this passage also sets the stage for the king
to link Punishment “here below” with the administration of judicial inquiry in
court, which will be among the matters that Manu most richly develops as
follow-up on his king’s day.132 Moreover, we now see that, as prelude to the
verse in which varṇa and āśrama are mentioned together, Manu would have
Punishment correct two problem areas in seemingly different ways, or at least
with what appear to be different motivations. First, those who deviate from the
four āśramas are to be punished after determining their power and learning.
I take this to imply nāstikas and, again, above all Buddhists, who would have
mendicant practices that fall outside the āśramas and lack the right Vedic
learning, but who also just might have some royal backing. Second, irrespec-
tive of learning but with power hanging in the balance, those who breach the
boundaries of the four varṇas from below should be punished lest they give rise
to revolution. As Olivelle brings out regarding verse 7.21, “commentators take
adharottaram to mean that the lower castes would usurp the roles and privi-
leges of upper castes”; and as he says of 7.24, “I take the term prakopa here to
mean revolt or tumult among the populace, a meaning common in the
Arthaśāstra” (2005a, 294).
Manu thus dramatically reduces the opening that one finds in Ᾱpastamba
and the Mahābhārata to learn dharma from Śūdras, which, as we have seen, both
texts mention alongside learning dharma from women. The “mixed” union
of the male Śūdra and the Brahmin woman being Manu’s worst-case scenario

132. Both topics receive attention in connection with the king’s judicial responsibilities: varṇa at 8.172–75,
where it concerns the king’s impartiality to all classes; and āśrama at 8.390–91, concerning the limits of the king’s
intercessory powers when dealing with disputes among “twice-born men living in āśramas.” Olivelle translates
āśrameṣu here as “in hermitages,” but I think “life-stages” is more likely. As Olivelle notes (2005a, 322), “The context
clearly calls for extraordinary individuals,” and some commentators explain, and both Bühler and Doniger translate
it, with reference to disputes about life-stage duties. There are also Mahābhārata instances, to be mentioned, that
would seem to intend both meanings, and that is possible here in Manu as well.
 post-vedic brahmanical dharma 225

(M 10.30), it is probably no coincidence that women and Śūdras are the two
main targets of Manu’s recommendations for spectacular punishments in
public view:

When a woman, arrogant because of the eminence of her relatives

and her own feminine qualities (strīguṇa), becomes unfaithful to her
husband, the king should have her devoured by dogs in a public
square frequented by many. (M 8.371)

If a man of lower class deliberately torments Brahmins, the king

should kill him using graphic modes of execution that strike terror
into men (citrair vadhopāyair udvejanakarair). (M 9.248)

The situations are not exactly parallel. In the first case regarding women, Manu
is only improving on what Gautama has to say:

If a woman has sex with a low-caste man, the king should have her
devoured by dogs and have the man executed, or punish [him?] in the
manner stated above.133 (G 23.14–16)

Manu also improves on the punishment of “the male offender,” who should be
“burnt upon a heated iron bed” on a stack of logs in the same public square
(M 8.372). In the dharmaśāstra, tradition public spectacles for adultery seem to
have preceded those recommended for the execution of criminals to instill
public terror. For the latter, Manu has gone beyond the dharmasūtras to adopt
the terminology of the Arthaśāstra.134 Also, whereas the first type of public
punishment could be local to any offense, the second is called for “in prisons
along the royal highway where people will see the criminals, grieving and muti-
lated” (M 9.288). But there are convergences in the way Manu and the
dharmasūtrakāras treat women and Śūdras. They are mentioned as overlapping
conceptual categories.135 Both were indispensable to the varṇa system for the
“obedience” they should each “like to hear” about and ungrudgingly perform

133. That is, the Śūdra who has sex with an Ᾱrya woman: “his penis should be cut off and his property con-
fiscated” before he is executed (G 12.2–3).
134. On Manu’s use of KA terms to prescribe “colorful” (citra), “clean” (śuddha: probably beheading, “not
found in any other Dharma text”), and “painful” (kleśa) executions, see Olivelle 2005a, 333 n. to 9.279; 334 n. to
9.291. See M 9.288, 9.291–92 (a dishonest goldsmith, “the most wicked of all thorns” who is to be “cut to pieces
with razor knives,” would probably be a Śūdra).
135. For example, as defining levels of suitable punishment (Ᾱ 1.9.24.5: “As with killing Śūdras, so with
killing women”) or exemption from it (G 22.16–17: a non-fertile woman and a Śūdra are equally protected from
murder, and neither as well as a cow or frog). Cf. B 1.10.19.3: “What applies to Śūdras applies to women and cows,
except when it is a Brahmin woman soon after her menstrual period.” Cf. M 11.67: killing a woman or Śūdra are
among “secondary sins causing loss of caste.”
226 dharma

as dharma for Ārya men.136 But neither can hear the Veda or have the initiated
status of upper class identity. Moreover, their mingling defines the menace of
varṇasaṃkara (see G 12.1–7). They are flashpoints of trouble.
In the same vein, where Manu speaks of Punishment protecting the weak
in the context of the right of ownership, it is once again clear that the weak to
be protected are primarily Brahmins, whom Manu (like all our classical
Brahmanical dharma texts) systematically protects from all but the mildest
punishments. Indeed, where Manu and also the Mahābhārata137 say “the
stronger would grill the weak like fish on a spit (śūle),” it is a likely play on the
matsyanyāya, “the maxim of the fish,” according to which a lawless kingdom
is one in which the big fish eat the little fish, implying the impalement138 of
weak Brahmins should a wrong “stronger” party come to power. Manu, how-
ever, is unparalleled by the dharmasūtras or the epics139 in the way it, quite
surprisingly, puts revolution so openly in the air; and we are indebted to
Olivelle for making us aware that suppressing it is an important part of the
politics of this text. Indeed, as I have hinted, Manu’s Daṇḍa personified,
watching us red-eyed even in our sleep, is something like George Orwell’s
Big Brother140—a point that might bear further thought as regards epics in
which the two royal heroes, Rāma and Yudhiṣṭhira, who most represent
dharma, and in Yudhiṣṭhira’s case incarnates it, are precisely elder big
brothers of a more humanly appealing type whose word is nonetheless law—
even and indeed especially when it is harsh in its outcome.
Yet suppressing revolution aside, Manu’s author does have precedent in
one dharmasūtra and a close parallel in the Mahābhārata for the way he brings
varṇa and āśrama together as a unified topic once he has introduced rājadharma,
“the Law of the King” and the “rod of punishment” (daṇḍa) with it.141 Gautama,

136. As with Śūdras’ obedience (suśrūṣa) to the higher classes, strīdharma requires that a wife “should obey
(śuśrūṣeta)” her husband “when he is alive and not be unfaithful to him when he is dead . . . a woman will be
exalted in heaven by the mere fact that she has obediently served (śuśrūṣate) her husband” (M 5.151, 155). Mbh
3.196–206 has twenty-two such usages for the “faithful wife” and the good Śūdra in the Pativratā-Upākhyāna.
Śuśrūṣa, “liking to hear, obedience, service,” is often the stand-alone virtue for Śūdras; cf. Ᾱ 1.1.7; G 28.39; M 1.91,
9.334–35, 10.99; Mbh 3.149.36, 5.70.47, 12.60.28–34, 13.128.56. Cf. BhG 18.44cd, where Kṛṣṇa speaks of
paricaryā, “service,” instead as the “inherent karma of the Śūdra” (ātmakaṃ karma śūdrasyāpi svabhāvajam).
137. The second line of M 7.20, cited above, is identical with the second line of Mbh 12.67.16, with the
latter’s first line being about what happens should there be no king to bear the daṇḍa.
138. Śūla, if single-pronged, implies an impalement stake—a meaning relevant to Śiva’s triśūla or “trident.”
139. As the Marxist scholar Walter Ruben (1968, 111.116–17) makes clear, although “there was no revolu-
tionary class in ancient India” and although most of the “despots” known from texts (including the epics) are
“mythological,” leaving the people to be “consoled with religious stories,” some Buddhist Jātakas envisioned
mass resistance. See Hiltebeitel 2001a, 177.
140. See the use of spies and secret agents in “the eradication of thorns” at M 9.261.
141. There may also be such precedent in the Arthaśāstra: “And after conquering the world, he [the king]
should enjoy it divided into varṇas and āśramas in accordance with his own duty (svadharmeṇa)” (13.4.62).
 post-vedic brahmanical dharma 227

as the only dharmasūtra to mention the two terms together, does so twice
(11.9–11, 11.29) when it gets to the topic of the king in conjunction with an
unpersonified daṇḍa (11.27–32), and a third time in an overarching summary
(19.1) that evidently goes back to those two passages. This will not surprise us,
since Olivelle has demonstrated Manu’s use of Gautama as a source and
model. Indeed, the first of these rules in Gautama has similarities to Manu
7.35, by which, as we have seen, Manu first brings varṇa and āśrama
together:

He [the king] should watch over (abhirakṣet) the social classes and the
orders of life in conformity with their rules (varṇān āśramāṃś ca
nyāyato), and those who stray (calatas) he should guide back to their
respective duties (svadharme sthāpayet), “for the king,” it is stated,
“takes a share of their merits (dharma).” (G 11.9–11)

The king’s role in both passages is to watch over or protect (using forms of
abhi + √rakṣ) the varṇas and the āśramas in the area of his subjects’ svadharma,
which Manu, as we have seen, intensifies by sve sve.142 Meanwhile, the
Mahābhārata mentions varṇa and āśrama together eleven times,143 the first
four occurring in what could be called a cluster in the Rājadharma section of
the Śāntiparvan (Book 12) very near the beginning of Bhīṣma’s postwar
instruction of Yudhiṣṭhira, where they also coincide with Bhīṣma’s introduc-
tion of the daṇḍa (12.59–74).144 Let it suffice to describe this cluster using the
rubrics by which Fitzgerald (2004a, 292) summarizes the segments in which
they appear. The first three occur in a sequence (12.60–66) on “Permitted and
Prohibited Occupations and Life-Patterns and the King’s Responsibility to
Enforce These”; and the fourth comes in the next adhyāya or chapter (12.67)
which treats “The Nature and Character of Kingship.” “Life-pattern” (rather
than “life-stage”) is Fitzgerald’s thoughtful translation of āśrama, and “permitted
occupations” are of course defined by varṇa. It is important that Gautama and
the Mahābhārata are, along with Manu, our only classical dharma texts to make
these correlations: Gautama for its affinity with Manu just mentioned, and the
Mahābhārata, particularly in its aftermath to the war, for the suggestive verbal,
conceptual, and, as we shall see in the next section, narrative affinities that it
alone has with Manu.

142. Cf. also M 7.15 with ÷cal, “to deviate, wander.”

143. At 12.63.11ab, 12.64.24cd, 12.66.37ab, 12.67.1ab, 12.92.7ab, 12.230.014ab, 12.261.44ab, 12.308.177ab,
12.308.180ab, 13.027.0341, and 14.35.27ab.
144. “The Origin of the Daṇḍa” (12.121) comes later, with Daṇḍa again personified as dark (śyāma) but with
many eyes and such additional features as four tusks, eight feet . . . (121.14–16).
228 dharma

F. Rājadharma: Establishing a King’s Dharma

We thus come to the area where, at least with regard to the legal tradition blazed by
the dharmasūtras, both Manu and the Mahābhārata innovate markedly. Manu does
so mainly in terms of subject matter or content: attention to “the king, statecraft,
and especially judicial procedure” (Olivelle 2004b, xix). Olivelle is certainly right to
emphasize Manu’s advances on these interrelated fronts, and that, on statecraft
and judicial procedure, Manu shows familiarity with the Arthaśāstra tradition
(2005b, 275–85). Statecraft as artha and daṇdạ nīti (counsel on coercive authority) is
the subject of an independent and probably prior śāstric “science” on its own (see
Sinha 1991, 369 n. 3), which Manu and the epics treat as if dharma encompassed
that science.145 As regards judicial procedure, Manu works out a well-ordered pre-
sentation of eighteen “grounds for litigation” (often translated “titles of law”) that
seems to be his improvement on a different ordering of sixteen grounds in the
Arthaśāstra (Olivelle 2005a, 13–16). With these eighteen grounds, Manu goes into
an area untouched by the epics. Yet one might leave open the possibility that Manu
gets to the eighteenth ground with the Mahābhārata in view. Why now eighteen
grounds rather than sixteen? Not only is eighteen the Mahābhārata’s signature
number;146 as Olivelle observes, Manu makes “gambling and betting” his eigh-
teenth ground and treats it differently from the other seventeen, for which he pres-
ents rules to establish proper legal procedure. Moreover, unlike Ᾱpastamba, which
presents “rules for the orderly conduct” of gambling itself, Manu considers it a
practice that “should be suppressed rather than regulated” and takes it up along
with theft under the heading of “the eradication of thorns” (kaṇṭakaśodhana)—a
category that in other dharma (and also artha) texts “falls outside the grounds for
litigation” (15). The negative way Manu sets off gambling as the eighteenth ground
is curious. Manu says, “In a former age gambling was seen to create great enmity
(dyūtam etat purā kalpe dṛsṭ̣ aṃ vairakaraṃ mahat); therefore, an intelligent man
should never engage in gambling even for fun” (M 9.227; Olivelle trans. 2005a,
202). As Olivelle indicates, this could refer to the gambling losses of Nala and/or
Yudhiṣṭhira—in either case, something from the Mahābhārata (2005a, 332). But
surely, since Nala actually reconciles with his dicing opponent and no war results,
it is Yudhiṣṭhira’s gambling that would be known for creating “great enmity.”

145. Dharma is the encompassing member of the trivarga that includes dharma, artha, and kāma. See M
2.13, 2.224 (and Olivelle 2005a, 253 on Manu’s six usages of trivarga), and 12.38, a brief ascending correlation of
the trivarga with the three guṇas or “qualities” of matter: kāma with tamas (darkness, torpor), artha with rajas
(vigor, energy), and dharma with sattva (goodness, luminosity). The epics frequently give these “three goals”
personal advocates, with dharma generally trumping.
146. Cf., however, Biardeau 2002, 1: 91: “One does not find oneself surprised to find the ‘affaires courants’
divided into eighteen categories; the number reappears.”
 post-vedic brahmanical dharma 229

What then becomes of interest is that Manu says this occurred “in former kalpa,”
not a former yuga, or “between yugas,” as one might expect from the Mahābhārata.
Noting for the moment that “eon” would have been a useful differential translation
of kalpa, leaving “age” for yuga, I will return to this curiosity in chapter 6.
As to the Mahābhārata itself, Braj Sinha offers an important discussion of
the epic’s treatment of Arthaśāstra and daṇḍanīti in Bhīṣma’s instructions to
Yudhiṣṭhira on rājadharma, which applies equally if not more so to Bhīṣma’s
ongoing treatment of āpaddharma, both in the epic’s Śāntiparvan. According to
Sinha, Bhīṣma bridges the “chasm” (1991, 369) between the Arthaśāstra tradi-
tion of “putting the interests of the King and State above all other interests”
(376) and “the dharma categories of the Dharmasūtras” (370) by “resorting to
the universal principle of dharma as the foundation of the Mahābhārata notion
of rājadharma” (383). Moreover, Bhīṣma does this in parallel with, and with
“echoes” of, the Buddhist “conception of a righteous king and the sociopolitical
and cosmic significance of such a righteous king” (375), and of Buddhist usage
of the term dharmarāja in the Nikāyas and Jātakas (376–83). As will be clear
shortly in this section, however, Sinha’s treatment of the dharmasūtras as
already “concerned with the art of government in a monarchical State” (474)
seems to read Manu, which he does not discuss, back into that tradition.
The “Mahābhārata question” thus raises questions of Manu, and vice versa.
With such questions in mind, I will focus on two features of Manu’s main
chapter 7 on the king: the construction of a fortified capital, and, in the next
section, Manu’s narrative of a king’s day.
After having first mentioned the royal fort (durga) in a passage already
noted, where the boomerang effect of unused Punishment redounds from the
fort and the kingdom up to the gods and sages on high (7.29), Manu speaks of
starting up a royal fort in the following terms:

He should settle (āvaset) in a region that is dry (jāṅgalam . . . deśam),

abounding in grain (sasyasaṃpannam), populated mainly by Āryas,
healthy, beautiful (ramyam), with submissive neighbors, and
providing a comfortable living. A fort secured by a desert, a fort with
an earthen rampart, a fort surrounded by water, a fort protected by a
forest, a fort guarded by soldiers, and a fort protected by a hill
(giridurgam)—finding safety in a fort, he should settle in a fort. He
should try his very best to find safety in a hill fort; for the hill fort,
because of its numerous superior features, is the most excellent of
them. Animals, creatures living in holes, and fish find safety in the
first three of them; and monkeys, humans, and gods
(plavaṃgamanarāmarāḥ) in the last three respectively (kramaśaḥ).
230 dharma

As their enemies do not harm these when they have found safety in a
fort, so his foes do not harm a king who has found safety in a fort. . . .
It should be well supplied with weapons, money, grain, conveyances,
Brahmins, artisans, machines (yantrair), fodder, and water. At its
center he should have a house built (kārayed gṛham) for himself, a
house that is spacious, secure (guptam), suitable for all seasons, and
provided with pools and groves. (7.69–73, 75–76; Olivelle trans.
2005a, 157–58, slightly modified)

Let us note that we are already in a little narrative. Only “after” securing a fort
(which he could presumably conquer as well as build) and made a house for
himself “should” this start-up king marry and appoint a chaplain (purohita)
and other trusted Brahmin officials (7.77–81). Between this minor narrative
and the more foregrounded one that starts him on the morning of his ideal
day (beginning at 7.145 and running through nightfall and the rest of chapter 7),
one hears at some length how the king should conduct matters of war and
conquest (7.87–109) and see to the protection of his subjects (110–44).
In this segment on war and domestic policy that comes between these two
narratives, Manu projects a fairly grand monarch, as indeed it does elsewhere:
not least in the divine infusions that go into the original king’s creation, but also
in the claim that the king is or makes the yuga or “age” (9.301–2), and in a verse
saying, “A king, though a mere child, must never be treated with disrespect,
thinking he is just a human being” (M 7.8), which hints at monarchy on a poten-
tially dynastic scale.147 But in this first minor narrative on starting up a kingdom,
Manu stays within the “little king” idiom typical of the dharmasūtras, where the
Brahmin authors envision close access to the king as if it were open to more than
just themselves.148 It is, however, especially Ᾱpastamba who affects this image,149

147. See von Stietencron 1997, 497–98: “a passage which points to hereditary kingship.” I would not,
however, follow von Stietencron in continuing “. . . at the time of its composition,” as if one could date the
passage differently from others in Manu on this basis of this implication. This type of historicizing is not to be
trusted, particularly now after Olivelle has shown Manu to be a largely unitary text.
148. G 10.36 may signal a small kingdom: one should report something found to the king; similarly V 15.6:
when one adopts a son, inform the king. Cf. G 20.1: “A man should disown a father who assassinates a king.”
Vasiṣṭha, however, supplies royal succession rules, including a stipulation that royal widows may become ascetics
(19.29–34), which sounds possibly epical. Meanwhile based on what little Baudhāyana offers on the king,
Biardeau writes, “the Brahmins of the Baudhāyana school” could not know “a great king. . . . Or possibly these
Brahmins were just more preoccupied with their personal duties” (2002, 1: 73–74).
149. See Biardeau 2002, 1: 79–80: noting that the sūtra style may make economy, “one cannot help but be
struck by the brevity of the description, a brevity that no doubt corresponds to the rusticity of the place. . . . As to
the palace, it is no doubt a simple residence.” See Ᾱ 2.26.2–3: the king who dies attempting to secure wealth
stolen from Brahmins is no different from other heroes (anye śūrāḥ) whose “own body serves as the sacrificial
post (ātmāyūpa) and an unlimited amount is given as the sacrificial fee.” The explanation would befit some village
“hero stones.”
 post-vedic brahmanical dharma 231

and moreover offers Manu’s only dharmasūtra precedent in speaking not only
about starting up a fort but about a fort (puram) at all,150 which he describes with
a rather precise and surprising (see Biardeau 2002, 1: 79–80) ground plan.

He should have a residence and a fort constructed (māpayet), with

their gates facing south. The residence (veśman) is within the fort, and
in front of the residence is the lodge (āvasatha), which is known as the
Audience Hall (āmantraṇam). To the south of the fort is the assembly
hall (sabhā) with doors on both the south and the north sides so that
one can see what goes on within and without.151 In every one of these
buildings fires should be burning continuously, and every day
offerings should be made in these fires in the same manner as in the
domestic ritual. The king should put into the lodge at least those
guests who are vedic scholars. They should be given accommodation,
as well as beds, food, and drink, in accordance with their distinction
(yathāguṇam).152 The king should not live more opulently than his
elders and ministers (gurūn amātyāṃś ca). And in the realm no one
should suffer from hunger, illness, cold, or heat, either through want
or by design. In the middle of the assembly hall he should erect a
gambling table, sprinkle it with water and place there dice—they
should be in pairs, of Vibhītaka seeds, and in adequate numbers.
Āryas who are upright and honest may gamble there. (Ᾱ 2.25.2–13)

As Biardeau observes, both Ᾱpastamba and the Mahābhārata would have a

common rapport here with the śrautasūtras’ precedent-setting descriptions of
“the sabhā and its sabhya fire . . . tied to the ceremony of the solemn consecra-
tion of the king, the rājasūya,” in which the sabhya fire is lit on that occasion for
the Rājasūya-ending dice match that the king must win (2002, 1: 80). Attentive
to the way the Mahābhārata plays these matters out in its dice match sequel to
the Rājasūya, which King Yudhiṣṭhira loses rather than wins, Biardeau finds
the rapprochement “evidently troubling”:

150. See Biardeau 2002, 1: 79: “This capital is probably a fort, if one keeps to the meaning of Vedic pur.”
Ᾱ would seem to echo this Vedic usage also when at 1.22.7 it speaks in Upaniṣadic terms of the self as “the fort
(puram) beyond compare.” G has nothing about a fort, only a sabhā (11.17); B mentions “the bolt of a city gate”
(2.6.13) using puram apparently as “city,” as it also does jointly with nagara (“city”) at 2.6.33; V uses puram in the
compound antaḥpura, “harem,” which as “inner fortress” implies the “fort” meaning of puram. M’s word durga
is not used for “fort” in the dharmasūtras.
151. See Biardeau 2002, 1: 80: “one notes here the absence of any surrounding wall. There is no fortifica-
tion, even if the royal ville would be a protected place. Perhaps the site offers more protection than the complex-
ities of the construction. Let us note the doors that permit seeing arrivals (enemies) from all sides.”
152. Or “according to their merits, qualities”; note the use of guṇa, which comes to be interchangeable in some
contexts with dharma in the sense of “merits, virtues,” to distinguish rank among Vedic scholars or Śrotriyas.
232 dharma

It is not possible to say which text borrows from the other, since both
rely on the same episodes in the solemn ritual (notably of the
Taittirīya school). It is rather the appearance here, in this
dharmasūtra, of a major epic theme that poses a question: one that is
not directly soluble, but which can translate preoccupations that are
“in the air,” which would have to do with the king of the Brahmanical
society. (Biardeau 2002, 1: 180)

I think, however, that Biardeau’s closing explanation suffices without the need
to consider the possibility that the Mahābhārata would borrow from Ᾱpastamba,
much less the reverse.
As to Manu, whatever precedent Ᾱpastamba offers within the legal tradi-
tion for starting up a fort, Manu takes it minimally. Ᾱpastamba makes it explicit
that the start-up king is to have the fort “constructed,” and not only a fort but a
capital complex. On the other hand, Ᾱpastamba gives no hint at any different
types of forts, such as Manu enumerates, that the new king might choose from.
As we shall see, this enumeration seems to serve new purposes in Manu.153
What interests Ᾱpastamba is a ground plan that zeroes in on the gambling hall,
and considering Manu’s antigambling legislation, it is not surprising that
Manu mentions no such multiple constructions and settles for now on building
(kṛ) just the one royal residence with the seemingly indifferent and utterly
generic term “house” (gṛha).
But Manu may know of another precedent for founding a new capital. That
is what the Pānḍ ạ vas do, shortly before Yudhiṣṭhira’s Rājasūya and the disastrous
dice game that follows, when they found their new kingdom at Indraprastha
after burning the Khānd ̣ ạ va Forest. As befits Manu’s prescription, they select a
dry region or savannah (jānġ ala): to be precise, land that is within Kurujāng ̇ ala,
a name for the whole Kuru domain, which according to an earlier passage was
rich in grain (sasya) and delightful and salubrious in the time of the heroes’ par-
ents (Mbh 1.102.1–14) in ways similar to the type of spot that Manu recommends
for a new capital.154 All this is not to say that Manu draws on the Mahābhārata or
vice versa, though, as I have indicated, I think it is the former. Why would
Manu’s king, who has no particular location for his capital, be urged to deploy
soldiers “from the lands of the Kurus, Matsyas, Pañcālas, and Śūrasenas” on his

153. As Biardeau already suggests with regard to the inclusion of the forested fourth type among the best,
“What are the monkeys doing here, if the epic is not at the bottom of the tableau, at least the Mahābhārata which
knows already the story of the Rāmāyaṇa and gives Arjuna the monkey Hanumān as his emblem?” (2002, 1: 90).
̣ ạ vas build their fort there, it would not, of course, have “submissive neighbors,”
154. By the time the Pānd
since their Kaurava cousins and enemies will now occupy the other “half” of the kingdom. The Pānd ̣ ạ vas have their
ceremonial hall (sabhā) constructed there by the Asura Maya, whose name—the Constructor—derives from √mā,
just as Ᾱpastamba uses the optative of that verb in prescribing how the king “should construct” (māpayet) the fort.
 post-vedic brahmanical dharma 233

front lines when he goes to battle (7.193)? One cannot rule out that the peoples
of this Ārya heartland had already become a target for military recruitment.155
But even if it did, the recommendation could still recall that their lands provided
the chief participants in the Mahābhārata war. We can, in any case, say confidently
that Manu knows the epic’s idioms in ways that are not derived from a common
sacrificial nexus such as one catches “in the air” in the epic’s and Ᾱpastamba’s
shared references to the solemn Vedic ritual.

G. A Day in a King’s Life

Manu’s recommendation of the hill fort as the best of the six varieties is not
surprising tactically. But it is intriguing that Manu draws on a poetic
convention of the Mahābhārata to explain its choice. It comes out early at the
beginning of his narrative of a day in the life of the king. The king should rise
in the morning, perform his personal purifications with a collected mind,
make the fire offering, pay respects to Brahmins, and enter “the splendid
assembly hall” (sabhā)—there by now or perhaps there all along, and some-
where near his “house”—to greet the public and then dismiss them; then he
should confer privately with his counselors (7.145–46). In order to do this,

Climbing up to the back of a hill (giripṛṣṭham), or terrace

(prasādaṃ vā), retiring to a solitary spot (rahogataḥ), or withdrawing
to a wild area or a bare tract, he should confer with them unob-
served (avibhāvitaḥ). (7.147; Olivelle 2005a, 162 slightly modified)

Here, I am in agreement with Bühler and Doniger’s “the back of a hill” rather
than Olivelle’s “Climbing up to a hilltop. . . .” The point is to avoid a place
where the king can be easily seen. Although pṛṣṭha as the “back” of an animal
can refer to its “top”156 or spinal “ridge,” it would not mean “top of the moun-
tain” here where the intention is to manage concealment. The king’s conceal-
ment is an important matter that has already been conveyed near the end of the
segment on warfare in two memorable epigrammatic verses (M 7.105–6) that
are all but identical with two verses found in the Ᾱpaddharma section of the
Mahābhārata’s Śāntiparvan (Mbh 12.138.24–25):

He must not let the enemy discover any weakness of his, but discover
any weakness of the enemy; he should hide his limbs like a tortoise

155. See Kolff 1990, 59–64, 160, 171–79; Hiltebeitel 1999a, 299–310; and chapter 7 on these peoples tak-
ing part in a second-century BCE “Northern Midlands Alliance.”
156. There is a variant giriśṛṅgam, the “top” or even “peak” or “horn of a mountain,” which only reveals a
misunderstanding by the would be corrector. Cf. Sternbach 1972 on this place.
234 dharma

and conceal his own weak points. He should ponder over (cintayet)
his affairs like a heron, dart off like a rabbit, snatch like a wolf, and
attack like a lion.157

The hiddenness of “the back of the mountain” is a poetic convention that can
be found in both the Mahābhārata and the Rāmāyaṇa.158 Manu’s usage confirms
this sense of hiddenness and concealment, and that Manu’s author is familiar
with the Mahābhārata poets’ conventions.
The “back of the mountain” has also brought us to the early hours of
Manu’s ideal day in the life of the king. Here, Manu’s precedent may be
Gautama, but Gautama applies this theme not to the king but in a few rules for
the Snātaka or Bath-Graduate in a lengthy ninth chapter on that figure.

He [the Snātaka] should not spend the morning, midday, or after-

noon fruitlessly, but pursue righteousness, wealth, and pleasure
(kuryāt . . . dharmārthakāmebhyaḥ) to the best of his ability, but
among these he should attend chiefly to righteousness (dharma). . . .
Morning and evening, however, he should take his meal, revering the
food, and never disparaging it. (9.46–47, 59)

Gautama’s concern that the Snātaka pursue the triad of dharma, artha, and kāma,
with dharma chief among these, is unique in the dharmasūtras in bringing forth
these three values together, not to mention ranking them. It is, however, typical,
at least as regards the combination dharma and artha, in being addressed to
Brahmins.159 Here we come to some important formulations. In the epics and
Manu, these three values together come to be called the trivarga or “triple-set”;
and when mokṣa is added as a somewhat disjunctive fourth,160 the four are called
puruṣar̄ thas, the four “aims of man” or “goals of human life.”161 In these texts,

157. The only differences arise where Olivelle’s Critical Edition of Manu reverses the second half of each line
on the similes of the rabbit and the lion. See Olivelle 2005a, 298 on these and other similar verses. Belvalkar 1966,
2222 also mentions the parallel. For another on the topic of ascetic aversion to the female body, see M 6.76–77 and
Mbh 12.316.42–43 where it is the celibate Nārada’s message to Vyāsa’s son Śuka (see Hiltebeitel 2001a, 292).
158. On shared Mahābhārata (and to some degree Rāmāyaṇa) poetic conventions, see Hiltebeitel 2004a,
220–26, with discussion of the “back of the mountain” and related usages of pṛsṭ̣ ha in both epics (221–24, especially
223 n. 75), carrying forward from Hiltebeitel 2001a, 290–317 concerning the literary site of Vyāsa’s hermitage.
159. See Ᾱ 1.4.23; 24.23; B 1.4.1; V 7.77 (each on only dharma and artha); while V 1.1 considers dharma the high-
est “goal” (artha) of man. Ᾱ 2.10.14 is the only dharmasūtra passage on both dharma and artha directed to the king.
160. Mokṣa is sometimes called apavarga (frequent in the Mbh, and especially the Mokṣadharmaparvan): the
goal that, in dealing with transcendent matters of salvation, is “away from the classification” of the triple-set. But
there are different “rhetorics,” including a Vedāntic one that treats dharma and mokṣa together as śreyas (“the sov-
ereign good”) and kāma and artha as preyas (“what is pleasant”) (Malamoud 1982, 37).
161. In Manu, the one mention of puruṣārtha, significantly broached in connection with good government
by the king (7.100), refers to four of them, but without any indication of how mokṣa would be pertinent. At 2.224
Manu mentions only the three of the “triple-set.”
 post-vedic brahmanical dharma 235

however, the trivarga is more typically couched in advice given especially to kings,
and also by kings. So it is in Manu, after a general introduction of the trivarga
(2.224)—remember, Manu is legislating as a first king himself:

When a king administers Punishment properly, he flourishes with

respect to the triple-set (trivarga). (7.27ab)

At midday or midnight, when he is not tired or worn out, he should

reflect (cintayet) on these matters either in consultation with his
counselors or alone—on Law, Wealth, and Pleasure, and on how they
may be acquired all together when they are in mutual opposition.
(7.151–52b; see also 7.100)

One may wonder that Manu should transfer a theme from the Snātaka to the
king. But Gautama already brings these two figures quite decisively together:

He [the Snātaka] should approach the king (īśvara) for the sake of
livelihood (yogakṣema), but not anyone else except gods,162 elders, and
righteous people. He should try to live in a place well supplied with
firewood, water, fodder, Kuśa grass, and garland material, served by
many roads, inhabited mainly by Āryas, full of energetic people, and
ruled by a law-abiding man (dhārmikād).163 (9.63–65)

Gautama brings the Snātaka to such a “righteous” king for the sake of yogakṣema,
which Olivelle may undertranslate here, and again later, as “livelihood,” since
Gautama’s other usage certainly has something of the feel of the Ṛgvedic sense
(see chapter 3 § B) of “harnessing” for war (yoga) and “peaceful settlement”
(kṣema) as alternating roles of a king:

He [the king] should also pay heed to what his astrologers and augurs
tell him, for, according to some, welfare (yogakṣema) depends also on
that. In the fire within the assembly hall (śālāgnau), he should
perform rites to secure prosperity (ṛddhi) in connection with a
propitiation (śānti), festive day, military expedition, long life, or
auspiciousness, as well as rites to stir enmity, to subdue or slay his
enemies, or to bring them to their knees. (11.16–17)

162. Olivelle 1999, 376 thinks that this would refer to divine images, suggesting “that a Brahmin in dire
straits may go to a temple to obtain assistance from the temple funds.” More likely, he would approach the deity
in śrauta or gṛhya rites that he could perform remuneratively for others.
163. As, at some other points, for the augmented form dhārmika (Olivelle 1999, 93 has “righteous”).
236 dharma

The other three dharmasūtras do not use this compound term (although see
Ᾱ 1.8.3; 2.5.10).
Manu, however, goes beyond any such a practical rationale for bringing the
Snātaka and the king together. Making one surprising twist, he treats them
both in the same way as role models,164 indeed as characters whom he addresses
by a sort of optative apostrophe using the verbal root Öcint that allows him to
tell the king and the Snātaka—and them alone—not only what they should do
but what they should think.165 These would both appear to be largely “new
fictions” of Manu. As in Manu’s chapter 7 on the king, chapter 4 on the Snātaka
presents these usages in a cluster and as implying something of a sequence,
although with other matters interspersed. At Manu 4.92ab, “He [the Snātaka]
should wake up at the Brahma hour and should think (anucintayet) on dharma
and artha”; at 4.109, where the question of suspending Vedic recitation has
come up, the Snātaka is told of untoward settings and circumstances where “he
should not even review it in his mind” (manasā-api na cintayet); and at 4.258, in
the very last verse of instruction to the Snātaka, who has just been told how to
retire peacefully and leave everything to his son, such advice is capped off with,
“Living alone in a secluded place, he should always reflect (cintayen nityam) on
what is beneficial to himself; for by reflecting alone, he attains supreme bliss.”
Here we see the fully rounded and extended picture Manu wants to give of the
Snātaka, who originally, it seems, is the young Veda scholar between his initi-
ation (upanayana) and marriage who has undergone the rite of samāvartana,
“returning home,” that seems to prepare him to be the prototypical host and
guest of Ārya society (Heesterman 1978, 439–46). Manu’s Snātaka is now not
only the young man who has finished his Vedic study at the home of a guru,
bathed, and returned to his parental home optimally to marry, but a man who will
stay home as the ideal householder until he is ready—still as a Snātaka—to retire.
The Mahābhārata speaks casually of such types when Draupadī boasts that
“Yudhiṣṭhira supported eighty-eight thousand Snātaka householders (snātakāḥ
gṛhamedhinaḥ) with thirty slave girls each” (3.222.41)! As Olivelle has indicated, in
Manu it is something of a maladjustment designed to let the Snātaka’s lifelong
status of “Bath Graduate” serve Manu’s classical prescription that the four āśramas
as life-stages be taken up in sequence and without choice (1993, 137–38).

164. I follow Merton 1957, 303 on “role model” in contrast to “role” as “more restricted in scope, denoting
a more limited identification with an individual in only one or a selected few of his roles.” Cf. Hiltebeitel 2004b,
40. The limited role model features of Manu’s Snātaka and king would have to do with the narrowing ways that
each, but differently, embodies lifelong purity.
165. A similar usage of cintayet is also found eight times in the Arthaśāstra, but without implying a daylong
narrative: on the topic of counsel, after covering the winning of territory (1.15.1); in rules for the king (19.16 and 21);
concerning the superintendant of horses (2.30.3); on secret agents in the disguise of householders, traders, and
ascetics (35.2 and 6); and on rules for the city superintendent (36.1 and 4). Like Manu, Kauṭilya far more frequently
uses vidyāt, “he should know.” As in other matters noted, Kauṭilya would be another of Manu’s likely models.
 post-vedic brahmanical dharma 237

As to the king, this device is something we have now already noticed twice
among the five times Manu uses it for him: the use of the optative third-person
singular of √cint—“he should think” (cintayet)—which, we will recall, the king
should do “like a heron” (M 7.104); and, once his day is underway and he can
find some fresh time for it, either at noon or at midnight, it is something he
should continue to do with his counselors or alone, thinking over the complex-
ities of the trivarga (151–52).166 Considering that this instruction governs the
free hours of the night, these usages cover all phases of the king’s day. Following
the discussion of how the king should position himself in the “maṇḍala of
neighboring kings” (154–59, 207), “He should think constantly (cintayet sadā)
about the sixfold strategy of forging alliances, waging war, marching into battle,
remaining stationary, pursuing a double stratagem, and seeking asylum” (160).
In this vein, there is also a single use of a functionally similar verb: “He should
probe closely (vicārayet) into the current status and the future shape of all his
undertakings, as well as the positives and negatives of all his past undertak-
ings” (178). Then in the afternoon, “After his meal, he should relax in his
private quarters with his women, and after relaxing, once again turn his
attention (cintayet) at the proper time to his affairs” (221). It is a daylong round
of near-constant vigilance that then extends similarly into the evening and
night (223–25), and continues to have implications for the king’s administration
of justice, where, on the matter of sexual crimes against women (the fifteenth
of the grounds for litigation), Manu’s final usage of cintayet serves to warn him,
“There is no greater violation of the Law on earth than killing a Brahmin; there-
fore a king should not even think (manasā-api na cintayet) of killing a
Brahmin”—”even if he has committed every sort of crime” (8.380–81). Manu
would not get this usage from the dharmasūtras, which, although they make
frequent use of the optative, do so primarily regarding only actions, and in any
case not with the root √cint.
Now it should be no surprise to find this usage fairly frequently in the
Mahābhārata, and considering that this epic and Manu share so many idioms,
it is also no surprise that it is used almost exclusively in passages where
instruction is imparted to two kings, Yudhiṣṭhira (twenty-four times) and the
Kauravas’ father Dhṛtarāṣṭra (who is twice told what “he should think” by his
half-brother Vidura [5.33.45, 11.2.17]). The only exceptions occur where Kṛṣṇa
tells Arjuna how and what “he should think” once in the Bhagavad Gītā (6.25;
Mbh 6.28.25) and three times in the Anugītā (14.19.34, 35; 46.40). Sinha’s
discussion of the Mahābhārata’s “metamorphosis” (1991, 382) of the
Arthaśāstra’s treatment of daṇḍanīti is pertinent here:

166. On complexities of the trivarga and Puruṣārthas, see Biardeau 1989a; Malamoud 1982; Fitzgerald
2004b, 672.
238 dharma

Instead of putting the interests of the King and the State above
all other interests, . . . the Mahābhārata makes [its] dharma rāja
subservient to the interests of all. The scope of rājadharma is
expanded to include virtues far beyond the compass of [the]
particularistic varṇa dharma of [the] kṣatriya of casting away life
in battle, protecting the realm, and preventing the intermixture
of varṇa. Rājadharma comprises compassion for all creatures,
knowledge of the ways of the world, and relieving the distressed
and the oppressed. (376)

We shall follow up this point in chapter 9, noting that it has bearing on both
epics in their portrayals of Yudhiṣṭhira and Rāma. For the present, it will suffice
to note that Manu’s king spends his day thinking much more like an Arthaśāstra
king than either of these epic kings.
Nonetheless, Manu and the Mahābhārata have this much in common with
these urgings of the king to think: everything he should consider is a dharma-
topic. No matter how we date the Mahābhārata relative to Manu, we should not
fail to appreciate the latter’s originality here. In the Mahābhārata this usage is a
character-building device, and considering that the epic even tells us that Vyāsa
not only knows his characters’ thoughts but can shape them and, as a character
himself, even prompt what they have to say,167 this is no minor feature. Manu, on
the other hand, uses the same device to give narrative form to a nameless, silent,
and neutral king who has no character at all. I propose to call him a default nar-
rative king. He is a king through whom Manu can fashion a model that could
apply to all kings, little and grand. That is made possible because Manu gives him
both a temporal dynamism through these minor and major narratives (and in
identifying the king with the yuga), and a spatial dynamism with respect to the
“circle of neighboring kings,” about whom he must constantly strategize.168 But
it is also made possible because Manu’s king both acts and thinks. With this
crypto- or proto-narrative device, Manu gets his teachings not only under the
king’s skin but into his head. Though the king is always hedged around by his
ministers, he is also necessarily recognized for having to think for himself.
Now if we look at Manu’s minor and major narrative sequences together, it
appears that the king remains as “small scale” in the second as he was in the first,
which established him on his throne. Manu positions this little king with start-up
capital, not even married before he establishes himself as a king, much as he does

167. See Hiltebeitel 2001a, 33–91, especially 80–82, and chapter 9 below for an example concerning
Yudhiṣṭhira.
168. See Olivelle’s extensive notes on the different types of royal neighbors the king must deal with (2005,
300–303).
 post-vedic brahmanical dharma 239

the Snātaka, each in an in-between world that also includes all kinds of
possibilities—but with marriage for both among the first orders of business. To
enable these portrayals and give direction to their intersecting paths, Manu adds a
chief trait for each: for the Snātaka—Olivelle’s “bath-graduate”—that he has
bathed to last a lifetime; and for the king, that he can wield the daṇdạ with “statutory
purity” or “instant purification,” leaving him unaffected by the impurity of blood-
shed (M 5.93–98; see Olivelle 2005a, 284; von Stietencron 1997). Says Manu:

The taint of impurity (aghadoṣa) does not affect kings, votaries

(vratinām), and those engaged in a sacrificial session (sattriṇām); for
they are seated on the seat of Indra and are ever one with brahman.
(M 5.93)

As Biardeau observes of this and other such passages, “The dharma of the
Kṣatriya king is truly peculiar: his rapport with purity is its essential index. It is
thus that kings are not submitted to impurity, any more than Brahmins are
who are undertaking observances or occupied with sacrificing. . . . One sees
that the violence inherent to royal dharma and to well-regulated exercise of the
daṇḍa poses no particular problem. It is easier for a king to be pure than for a
Brahmin” (Biardeau 2002, i: 91). Moreover, the only dharmasūtra to speak of
the statutory impurity of kings is Vasiṣṭha, who does so while specifically
discussing the king’s role in suppressing crime by punishment (19.38–47).
Since only Vasiṣṭha does this,169 it is again evidence that Vasiṣṭha would be later
than Manu—although in this case, rather than attributing the closely parallel
verse to Manu, Vasiṣṭha attributes it to Yama!

The rule is that kings always become pure immediately after they
carry out capital as well as non-capital punishment: the reason for
this is simply time. In this connection they quote a verse proclaimed
by Yama: “The stain of impurity does not affect kings, as well as
people performing vows and sacrificial sessions, for they are always
seated on the throne of Indra and become one with Brahman.”
(V 19.47–48; Olivelle 2003, 429)

Manu fits his verse into a kind of aside on the king while dealing with the over-
arching topic of bodily purification. Manu does not cite it in his main discussion
of the king, which Vasiṣṭha does, making it likely that it is Vasiṣṭha who has

169. One can see the uniqueness of Vasiṣṭha among the dharmasūtras on this matter from Olivelle 2005c,
24, 152–53, where Olivelle shows dharmasūtra parallels on the treatment of “statutory purity of persons and arti-
cles”; and Olivelle 2005c, 31, 216–26 concerning the king’s role in “crime and punishment.” Olivelle 2003, 687
on V 19.48 merely cites the parallel; Olivelle 2005a 284, on M 5.83, does not.
240 dharma

sharpened the focus of the rule, and is thus, again, the later of the two.
Perhaps his attribution to Yama is Vasiṣṭha’s way of giving the rule greater
weight.
On the other hand, Manu’s singular orchestration of a rapport between the
bath-graduate and the king plays on a theme that was crucial to the way
Brahmanical culture distinguished itself from the heterodoxies: the still wider
rapport between the Brahmin and the king. For classical Brahmanical authors,
such a mode of self-distinction was both a necessity and a matter of choice as
to how to go about theorizing dharma and personalizing it through stories. It
was necessary in that the heterodoxies did not concede superiority to the
Brahmin, and it was a matter of choice where it came to making the king so
pivotal, and not only the king but the urgency of his being a Kṣatriya with Vedic
credentials. The heterodoxies had no qualms about non-Kṣatriya kings and had
nothing at stake in tracing that class back to the Veda.
Manu’s upstart king would find some precedent in Āpastamba. But he
would also be a reflex of the small-scale early Vedic chieftain and the later
Vedic Vrātya, who can lead a band of trekking warriors and can be a Brahmin
as well as a Kṣatriya—figures who anticipate the medieval Rājpūt of “achieved”
rather than “ascribed” status.170 As such, he is a reflection of a real tension at
the heart of Indian kingship that can be traced back at least to the rise of met-
ropolitan states, which swallowed many little kingdoms and would have pro-
voked both resistance and adventurers who founded new ones, and probably
back further to the early Vedic tension between the two types of kingship
exemplified by the emperor Varuṇa and the chieftain Indra. As many have
noted, the Pāṇḍavas themselves can be viewed in a Vrātya or little Rājpūt
mold.171 Once they have established themselves at Kurujāṅgala, they are
unable to assert real sovereignty until the two most martial among them,
Bhīma and Arjuna, along with Kṛṣṇa, disguise themselves as Snātakas; and in
that ambiguous status, which denotes that they have finished their Vedic
studies and puts advantage to the fact that not just Brahmins can be Snātakas,
they go to the great metropolitan capital of Magadha to eliminate the emperor
Jarāsaṃdha (see chapter 13).
Both the Mahābhārata and Manu thus want their start-up kings not to be
upstarts or parvenus, and this requires that they be Kṣatriyas: something that
in the Mahābhārata entails great genealogical contortions—above all by the

170. On the Vrātya, see Hiltebeitel 2001a, 129–40, 148–84 with pertinent bibliography. On the medieval
errant Rājpūt, see Hiltebeitel 1999a, passim, with pertinent bibliography, most notably Kolff 1990.
171. And not just scholars, among them notably C. V. Vaidya (1905, 56–57, 71–75, 164, 176; 1907, 163), but
regional martial oral epic reenplotments of the Mahābhārata from virtually all over India (see Hiltebeitel 1999a,
passim).
 post-vedic brahmanical dharma 241

author, who can intervene to sire the heroes’ fathers as Kṣatriyas even though
he is himself a Brahmin (see chapter 8). All these texts are in fact reinventing
the Kṣatriya. Gautama surely has some kind of “legitimate” Kṣatriya in mind,
but suggests things could be otherwise when he urges that the Snātaka live in
a kingdom “inhabited mainly by Āryas, full of energetic people, and ruled by a
righteous man” (9.65). But Manu admits a more difficult reality, and warns his
Snātaka not to “live in a kingdom ruled by a Śūdra, teeming with unrighteous
people, overrun by people belonging to heretical ascetic sects (pāṣaṇḍagaṇākrānte),
or swamped by lowborn people” (4.61).172 Nor should his Snātaka accept gifts
from an ucchāstravartin king: “one who deviates from the provisions of the
authoritative texts” or “follows a wrong śāstra” (4.87)—“for example, Jain or
Buddhist scriptures” (Olivelle 2005a, 271). As we shall now see, the epics and
Manu would know a “real world” in which “real vedic” Kṣatriyas would have
been hard to find already for centuries.

172. On Śūdras, see above n. 60 and § E above at nn. 133–36. Cf. M 4.218 on Snātaka avoidance of Śūdras,
among others.
This page intentionally left blank
6
Dharma over Time, I
Big Time Dharma

As previous chapters show, while dharma could be spoken of as in

some sense everlasting or eternal, it was also susceptible to change
over time. This chapter and the next will compare ways that both
Buddhist and Brahmanical texts wrestled with this tension. This
chapter will consider the grand schemes of cosmic time in which
familiar classical dharma texts contextualize change. Chapter 7 will
then be concerned with how both traditions produced prophesies of
dharmic doom during virtually the same period focused on much the
same geography: the Brahmanic tradition in the Yuga Purāṇa, and
northern Buddhists in texts that predict what will happen when the
Buddhist dharma ends.
In setting these discussions around the topic of dharma over
time, we thus undertake two surprisingly overdue comparisons.
First, studies of Buddhist and Brahmanical cosmologies have not
noted the contrastive ways they relate change in dharma to their
vastest concepts of time. In this comparison, the likely priority of the
Buddhist kalpa, noted already in two usages in the edicts of Aśoka
Maurya (see chapter 2), will again be important. Second, as to the
prophesies of dharma’s decline that will be taken up in chapter 7,
although they have been the subject of elegant studies, these very
studies were done without considering the Hindu or Buddhist
counterparts, and there has been little follow-up to them. These
comparisons are thus overdue.
244 dharma

A. Kalpas and Yugas

In beginning with large cosmological concepts, I start with some comments in

Johannes Bronkhorst’s 2007 book Greater Magadha: Studies in the Culture of
Early India. Bronkhorst challenges orthogenetic theories that project modern
visions of a unified “India”—grounded in the unfolding of the Veda—back on
India’s post-Vedic classical past. In opening his discussion of the “fundamental
spiritual ideology” that would have distinguished Greater Magadha’s non-Vedic
heterodoxies, Bronkhorst distinguishes four “features” of “what must have
been the culture of Greater Magadha” (55): funerary practices, medicine, the
godlike status of the sage Kapila, and what he calls “cyclical time.” This last
section is, however, brief (69–71) and, I believe, hasty on matters that concern
the two most prominent classical Indian time concepts, kalpas and yugas.
Bronkhorst speaks generally about “a cyclical notion of time, in which kal-
pas, yugas and other time units” impact later Hinduism. Leaving one to think
that kalpas and yugas would have a common background in Greater Magadha,
he does not take note that they are probably concepts with separate early his-
tories. As Luis González-Reimann has observed, the kalpa appears to begin as a
distinctly Buddhist concept with its first documentable usage possibly being by
Aśoka in a couple of his edicts:1 to be precise, the two mid-career edicts that we
discussed in chapter 2. I believe González-Reimann is right to conclude that the
word kalpa was “appropriated” from Buddhist sources into Brahmanical texts.2
Moreover, as Jan Nattier indicates,3 and as González-Reimann also demonstrates,

1. See González-Reimann 2008; cf. 2002, 129, 198 n. 98; cf. 64, 70, 167 n. 19, 169–72 nn. 36–37. Cf.
Pingree 1963, 238; 1990, 275 on the Buddhist kappa being “much earlier” than the yuga, whose appearances in
the Mbh and Manu he dates (I think too tardily) to “about the second century A.D.” I differentiate yuga
chronometry from usages of yuga and even yugānta, which do occur in the Pāli canon; see González-Reimann
2002, 62–64; 70.
2. González-Reimann 2008, 10, adding that during what he takes to be the lengthy history of the
Mahābhārata’s growth, “the cycle of four yugas became principally the cycle of the decay and renewal of dharma,
while the kalpa—by now [i.e., in some Śānti- and Anuśāsana-Parvan passages and in the “late” Bhagavad Gītā]
associated with the day of Brahmā—was reserved exclusively for the period of cosmic world destruction and ren-
ovation” (2008, 10). See Idem, 3–4, discussing Mbh 12.224–25 as one of the epic’s two more “important sections”
on “time periods and/or cosmic cycles,” in which 224.28–40 goes “from a nimeṣa to a day of Brahmā as lasting
1,000–12,000-year periods, with a night of the same duration, with a verse on the length of Brahmā’s day that is
“repeated verbatim” at BhG 8.17, which “establish[es] Kṛṣṇa’s status as the source and end of everything.”
González-Reimann says that the passage “has a parallel” at Manu 1.61–85. I discuss my disagreement with
González-Reimann’s textual stratifications below.
3. See Nattier 1991, 115 n. 130: “The fact that Buddhist texts (from the Gupta period and after) occasionally
refer to the Hindu four-yuga system strongly suggests that the Buddhists had no multi-era system of their
own”—such as the three-age system that Buddhists would introduce in China. Cf. 280–81 n. 3, citing a
Laṅkāvatāra-sūtra passage with reference to the Guptas, with important points about Gupta and post-Gupta
Buddhist usages.
 dharma over time, i 245

yuga calculations are not mentioned by Buddhists until quite late4—nor

does the yuga appear to be a feature of the time-reckoning of the other
heterodoxies.
Moreover, whereas the kalpa can rather reasonably be identified with
calculations of cyclical time, and indeed more so than the competing Jain and
Ā jīvika concepts current in Greater Magadha that have more to do with down-
swings and upswings of time, cyclicity is less clear in the yuga, which lends
itself more readily to accounts of linear and indeed historical time.
Bronkhorst also endorses, if a little tentatively, González-Reimann’s
“conclusion ‘that the yuga theory is a relatively late addition’ ” to the Mahābhārata
(2007, 72, citing González-Reimann 2002, 202). I will not be in agreement
that the yuga theory, with its chronometry of four ages, could be a “late addition”
to the Mahābhārata. But the important point for now is the one of regionality.
Taken together, Bronkhorst and González-Reimann allow us to identify the
kalpa as the distinctively Buddhist entry among the competing notions of the
vastness of time that were advanced in Greater Magadha, and perhaps the only
one that could properly be called cyclical.
This, however, brings us to a major contrast between the Buddhist and
Brahmanical traditions that has so far gone unnoticed. Whereas it can be taken
as axiomatic,5 even though we shall have to reckon with some nuances, that the
Brahmanical tradition dissociates the kalpa from dharma and links dharma
primarily with the yuga, and secondarily with the manvantara (“interval of a
Manu”), early Buddhism, doing without yugas and manvantaras, did link
dharma with the kalpa. Indeed, we saw Aśoka doing precisely this in his two
mid-career edicts. In the first of them, in Rock Edict 4 (RE 4), he expressed his
hope that the changes in dhaṃma conduct that he had initiated would continue
to increase, thanks to his “sons, grandsons, and great-grandsons, . . . until the
end of the world (āva sa(ṃ)vaṭṭakappā).” As we shall see momentarily, the
terms Aśoka uses here are clearly Buddhist, and moreover complex. For in
linking dharma with the kalpa, Buddhism gave the kalpa varieties and complex-
ities that Brahmanical texts ignored—largely, I will be arguing, because the
Brahmanical texts found the yuga theory more congenial to such complexities.
This is a matter of considerable implications.

4. González-Reimann leaves some uncertainty where he says the “yuga theory” “extended beyond Hinduism,
and was early on an important part of Indian Buddhism. Jainism, likewise, . . . although the Jain theory of cosmic
cycles does not include the yugas” (2002, 169–70). The Buddhist (2002, 167) and Jain (188 n. 19) theories he cites
are not early, and it is not clear that they specifically mention yugas. See further González-Reimann 2002, 64 on the
nonuse of “Kali” with yuga in Pāli texts; 70 on the term yugantavatā, “the wind of yugānta,” but only in a Pāli Jātaka
commentary; 129 on a Gupta period Jātaka; 172 on a Kali yuga prophesy by Avalokiteśvara regarding Maheśvara in
perhaps the sixth-century Kāraṇdạ vyūha; 198 n. 98 on a post-Gupta Buddhist adoption of yuga chronology.
5. See Biardeau 1981b, 70, 89–90, 113–14, and the “universe de la bhakti” graphic, 172.
246 dharma

B. Buddhist and Hindu Kalpas

Buddhist cosmology knows kalpas (Pāli kappas) of three overarching types. The
major heading is the “great aeon” (mahākalpa or mahākappa). Great aeons have
four sub-kalpas called “incalculable aeons” or “innumerable aeons” (asankheyya-
kappas, asaṃkhyeya-kalpas). And “incalculable aeons” have sub-kalpas called
“intermediate aeons” or “interval aeons” (antara-kappas, antara-kalpas). If several
Buddhas appear in one “intermediate aeon,” it can be qualified as a “good aeon”
or “blessed aeon” (bhadra-kappa, bhadra-kalpa). In appearing as one of several
Buddhas in the present such kalpa, Śākyamuni and the others make it a “good
aeon” (Strong 2001, 22)—a term found across sectarian boundaries and probably
behind grander Mahāyāna formulations as well (Nattier 1991, 8, 21–22, 25–26).
The complexity thus begins with the four kinds of “incalculable aeons” that
make up a “great aeon.” An incalculable aeon of destruction is followed by an
incalculable aeon of the duration of destruction (when nothing is manifest), an
incalculable aeon of renovation or re-“creation,” and an incalculable aeon of the
duration of renovation (when the world of beings is in full swing until the next
incalculable aeon of destruction). In Pāli, the four incalculable aeons are called
saŋvaṭṭa-, saŋvaṭṭaṭṭhāyi-, vivaṭt ạ -, and vivaṭṭaṭṭhāyi- (AN 2.142.15–28). Clearly, in
aspiring to have turned the corner toward progress in dhaṃma that will last “until
the end of the world (āva sa(ṃ)vaṭṭakappā),” Aśoka’s fourth rock edict supplies a
Prākrit equivalent to the Pāli saŋvaṭṭakappa. Rock Edict 5 then uses the same
phrase and compound in the form āva saṃvaṭṭakappā. Curiously, only the Girnār
versions of these two edicts use this precise terminology. Elsewhere, at four other
sites, each edict has the reduced phrasing āvakappam or a variant—that is, without
using saṃvaṭṭa (Bloch 1950, 100, 102; Hultzsch 1969, 189, 191). This implies only
“the end of a kappa” rather than any specific kind of sub-kappa.6 Odds are that the
two Girnār inscriptions have benefited from a locally knowledgeable, perhaps
“scriptural,” interpolation, perhaps in collaboration with a monk. But the impor-
tant point is that with or without saṃvaṭṭa, it is an Aśokan inscription from Aśoka’s
own time, and shows a contemporary Buddhist transposition into stone of a spe-
cifically Buddhist understanding of “kalpa.” If we concentrate for a moment on
the usage with saṃvatta, matters are not straightforward. The four-term cycle of
incalculable aeons can be abbreviated to saŋvaṭṭa-vivaṭṭakappa (DN 1.32–33), which
Walshe ([1987] 1995, 74) translates as “period[s] of contraction and expansion.”

6. Hultzsch observes the difference, translating āva saṁvaṭa-kapā (or a variant) in the Girnār edicts as
“until the aeon of destruction (of the world) (1969, 7–11), and translating āva kapa[ṁ] in the Kālsī edicts, for in-
stance, as “until the aeon (of the destruction of the world)” (30–34).
 dharma over time, i 247

This seems to reverse the order or the implications of the terms, for according to
Rhys Davids and Stede, “with reference to the Universe and time (kappa),” saŋvaṭṭa
means “rolling on or forward” and refers to the “ascending aeon” as the opposite
of vivaṭṭa, which means “rolling back” with reference to “the descending cycle.”
Thus the pair saŋvaṭṭa-vivaṭṭa refers to “a period in which evolution and dissolu-
tion of the world takes place, a complete world cycle.”7 Rhys Davids and Stede
address the ambiguity that would seem to lie behind Walshe’s translation: “as
‘periods’ of the world they mean practically the same thing and may both be inter-
preted in the sense of a new beginning. . . . We sometimes find vivaṭṭa in the sense
of ‘renewal’ and saŋvaṭṭa in the sense of ‘destruction,’ where we should expect the
opposite for each” ([1921–25] 2003, 637). Most likely, since the Girnār usages
imply a terminus, they would either mean that progress in dhaṃma could con-
tinue “until the end of the ascending cycle,” or, as with Walshe’s translation, “until
the next period of contraction.” What is most interesting for our purposes, how-
ever, is the context in which Aśoka speaks of the “end of the world” with or without
saṃvaṭṭa. As we saw in chapter 2, he positions his reign to have effected a turn-
about in the progress of dhaṃma not only with regard to a kalpic future that will
be secured by his children and descendants but with reference to past “centuries.”
He begins RE 4, “In the past, over many centuries, killing, violence done to crea-
tures, discourtesy to relatives, and disrespect for Brahmins and Samaṇas have
only increased. But now, . . . promulgation of dhaṃma has increased that which
did not exist over many centuries: abstention from killing, kindness to creatures,
respect to relatives, respect for Brahmins and Samaṇas, and obedience to mother,
father, and elders.” Yet all this has fallen flat, a year later, in RE 5, where he has
turned everything over, including implicitly his children and descendants, to his
newly appointed dhaṃma superintendents.
Now if we set aside its mention of disrespect for Samaṇas, RE 4 recalls the
past in terms of the same types of failures in dharma that Brahmanical texts
will soon enough ascribe to the Kali yuga. By RE 5, it is clear that the attempt to
imagine the progress of dhaṃma over the long haul of the kalpa was a bad fit,
and that if Aśoka had been able to conjure with a theory of yugas, he might have
been better served. Not surprisingly, RE 5 is the last we hear from Aśoka about
“the end of the world.”
Indeed, Aśoka’s long progress of dhaṃma is not only hard to square with
Buddhist kalpas in general but with the further complexities one meets in the
four kinds of incalculable kalpas, each of which is said to have twenty intermediate
aeons (antara-kalpas). During an incalculable aeon of destruction, intermediate
aeons are, as Nattier puts it, “rather meaningless” (1991, 16; cf. 2008a, 153),

7. See Rhys Davids and Stede (1921–25) 2003, 656 on saŋvaṭṭa; cf. 187 on kappa; 637 on vivaṭṭa.
248 dharma

since there are no beings affected. But when an incalculable aeon of renovation
begins, “the receptacle world (bhājanaloka) is created” in the first of its twenty
intermediate aeons, “while beings appear in the remaining nineteen.”8 Brahmā
is only the first being caught up in the devolution of beings at the beginning of
an aeon of renovation, and he accordingly gets linked with some false views,
including that of “creation” itself—as in the Brahmajāla Sutta (DN 1), where
Brahmā thinks he is the Creator only because he finds himself the first person
on the scene and desires a creation, and later arrivals subsequently think he cre-
ated them because “he was here first.”9 During this time the duration of life is
said to be infinite. But that begins to change in the first intermediate aeon of an
incalculable aeon of the duration of renovation. And through the last nineteen
intermediate aeons of an incalculable aeon of the duration of renovation, human
life oscillates between eighty-thousand years and ten years (Kloetzli 2007,
153–54). What is most significant about these twenty intermediate aeons of an
incalculable aeon of the duration of renovation is that “they mark the bound-
aries between the periods of decline (when the human life span grows progres-
sively shorter) and those of advance (when the reverse of this process takes
place)” (Nattier 1991, 16). Within this scheme, a Buddha’s dharma has of course
a beginning with his first sermon (see chapter 4 § B), and it will also have an
end. If Buddhas do appear in an intermediate aeon, which is a rarity, it is within
a downswing period of decreasing life span during an aeon of the duration of
renovation (Nattier 2008a, 156). Here there is a difference among Buddhists:
whereas in the Nikāya schools Buddhas cease to appear when human life spans
get shorter than one hundred years, in the Mahāyāna they can continue to be
born even after that in “the period of the Five Corruptions” (Chappell 1980, 141).
The extreme rarity of a Buddha teaching the true dharma thus occurs within a
vast panorama of aeons within aeons.
Like the Brahmajāla Sutta, the Aggañña Sutta (AS), which we discussed at the
end of chapter 4, probably taps into an unschematized and perhaps presystem-
atized version of this cosmology to focus mainly on something like the beginning
of an incalculable aeon of renovation. Both suttas trace the primary evolution of
beings rather wryly, beginning from the beings’ mental existence;10 and the AS
continues to trace their evolution down to gendered and social embodiment and
the “contracting” of the first king. Yet although the intricate Buddhist kappa

8. See Kloetzli 2007, 153–54, drawing this information from Vasubandhu’s Abhidharmakośa.
9. See Brahmajāla Sutta (DN) 1.2.2–5; Walshe [1987] 1995, 75–76, another famous satire of a
Brahmanical view.
10. See chapter 3, n. 200 on Collins’s view of its “studied vagueness about the [ AS’s] cosmogony.”
Up to the point of Brahmā desiring companions and the first finger-dipping, the Brahmajā la and AS pre-
sent situational and even verbatim parallels (see Nattier 1991, 11 n. 4), but then take different turns. The
Brahmajā l a quickly dispenses with anyone taking Brahmā seriously as the Creator ( DN 1.2.6).
 dharma over time, i 249

cosmology is mentioned in the Brahmajāla Sutta and is well established in the Pāli
canon,11 the AS never mentions the term kappa or anything about life spans, and
takes the evolution of beings both deeper back and further forward than an
incalculable aeon of renovation might lead one to expect.12 For present purposes,
however, what is of interest is that whereas the Brahmajāla Sutta does not trace
the course of dharma through time, the AS makes the course of dharma through
time central to its parable. As we saw in chapter 4, before beings start to differen-
tiate “things” (dharmas) that are wholesome from those that are not, they first
come to the point of confusing dhamma with adhamma and engaging in “that
which is not the true dhamma”: householder sexuality. The true dharma (sad-
dhamma, saddharma) must await the coming of a Buddha, something that does
not even happen in a kappa of renovation, but only in a kappa of the duration of
renovation. If the AS includes “what is not the true dhamma” among things that
evolve in a prior incalculable aeon of renovation, or even in its carry-over into an
incalculable aeon of the duration of renovation, it could be telling us what the true
dhamma would default or revert to when a Buddha’s dhamma disappears, which,
on the vast scale of inter-kappa time, happens both too rarely and too quickly.13
The sexualized householder dharma it would default to would, naturally, be
something like Brahmanical householder dharma.
Buddhist chronometry, which never settled on the number of periods the
Buddha’s dharma would pass through before disappearing,14 links periodization
of the Buddhist dharma, from beginning to end, and however many such periods
there may be, to the appearance of a Buddha. When a Buddha delivers his first
sermon and turns the Wheel of the Law, he does so only long after the previous
Buddha’s dharma has disappeared, and only after he himself has rediscovered the
dharma after gaining complete and perfect enlightenment absolutely on his own.15
Buddhas could extend their life span, and thus their dharma, to the end of a kalpa,
but at least in Śākyamuni’s case, which is paradigmatic, they do not.16 And when

11. See SN 2.15.5–8, 10 (Bodhi 2000, 654–57), among the “Connected Discourses on Without Discoverable
Beginning,” five similes: on the length and number of kappas involving a mountain, a city with mustard seeds,
disciples remembering past aeons, the sands of the Gaṅgā (addressing a Brahmin), and a heap of bones; Bodhi
2000, 716, 822 n. 37: Mahāmogallāna could have lived for an aeon; Idem, 1723–24, 1940 n. 249: “the Tathāgata
could live on for the aeon or for the remainder of the aeon”; Idem, 1758: recollecting “A thousand aeons.”
12. See Nattier 1991, 18; Kloetzli 2007, 154, discussing the recapitulation of the AS in the Abhidharmakośa.
The AS also implies a prior incalculable aeon of the duration of destruction as deep background.
13. The Sakkapañha Sutta [DN 21] carries this idea along with an explicit statement as to what can be
known of dharma in a time without a Buddha. Says Sakka to those he “considered to be ascetics and Brahmins”
whom he sought out to teach him, but who rather turned to him for answers: “Then I taught them the Dhamma
as far as I had heard it and practised it” (2.27).
14. See Nattier 1991, 46; 2008, 157–58, showing that there was no standard numerology in Nikāya
Buddhism, and leading into making the same point for the Mahāyāna.
15. For good discussions, see Strong 2001, 15–34; Nattier 2003, 2008.
16. This is addressed implicitly in the Mahāparinibbāna Sutta where the request that the Buddha extend
his life is tied to discussion of the continuation of the dhamma beyond it, though without a timetable. An inverse
250 dharma

a Buddha passes into final nirvāṇa, his dharma, in accord with his fundamental
teaching (dharma) that all conditioned things (dharmas) are impermanent, will
have a limited duration before it disappears like his predecessor’s. These ideas
take many forms, and one cannot be sure that the earliest formulations included
the expectation of a future Buddha. Stories of past and future Buddhas and predic-
tions of the end of the dharma are overlapping ideas, but are usually found sepa-
rately. For instance, in what is probably the earliest end-of-the-dharma prediction,
when the Buddha says the true Dhamma (saddhamma) will last only five hundred
years rather than a thousand because of his decision to admit nuns to the monastic
order, he does not mention the future Buddha Maitreya (Pāli Metteyya). But there
is no reason to think that the idea of future Buddhas is any younger than the
prediction about the effect of nuns on the duration of the dharma.17
Both the duration of a Buddha’s life span and the time his dharma lasts there-
after are said to differ from one Buddha to another. Śākyamuni was able to predict
the duration of his dharma past his lifetime, even if his predictions differ from text
to text (Nattier 1991, 19–20). According to texts that make such calculations, it will
take 5.6 billion or 560 million years between the death of Śākyamuni Buddha and
the coming of the next Buddha, Maitreya. In a single world system like ours, for a
bodhisattva to become a Buddha, he would not only have to wait for the disappear-
ance of his predecessor’s dharma, but to be “the next candidate in line” (Nattier
2003, 183–84; 2008, 159). Even with the innovation of multiple worlds and
Buddhas of the ten directions, he would still have to find a “Buddha-free zone.”18
Like Buddhist kalpas, Hindu kalpas operate cyclically and willy-nilly. In a
technical sense, in each tradition they operate independently of dharma.19 Or by
varied interpretation, they are a dimension of dharma as cosmic law (Hindu), or
of what is true when seen as it really is (Buddhist).20 But whereas Buddhist kalpas
cycle along without divine agency, with Brahmā only deluded in thinking he is the

relation is made explicit in the Chinese translation of the Candragarbha Sūtra, where the Buddha says “he will
renounce one-third of his life span for the benefit of sentient beings, thus prolonging the life span of the Dharma”
(Nattier 1991, 182). On the so-called “infinite life,” according to his name Amitāyus, of Amitabhā, who is mortal
in the earliest translation of the larger Sukhāvatī-vyūha Sūtra, and still “not immortal” in the more celebrated
subsequent translation, see Nattier 2003, 188–92.
17. See Nattier 1991, 30–33: The story blaming nuns is not found in the “surviving literature” of any of the
Mahāsaṅghika schools, but only in “schools on the Sthavira side of the family tree” (most notably the Theravādins,
Sarvāstivādins, and Dharmaguptakas). This could mean that it was not part of the Mahāsaṅghika-Sthavira split
that Nattier dates here to ca. 275 BCE. Nattier offers this as evidence that the story may be no earlier than 200
BCE, but suggests a 340–200 window. For other discussion, see 33 n. 16; cf. Nattier 2008, 156–57.
18. See Nattier 2003, 185, 193; 183–84 on the “structural or cosmological problem” posed in early Buddhism
by its one-Buddha-at-a-time principle. For an articulation of this principle, see Mahāgovinda Sutta [DN 19] 2.14.
19. On the Brahmanical kalpa, see Biardeau 1994, 21–22, noting that on this point it is difficult to con-
found kalpa and yuga.
20. See Nattier 1991, 5 n. 1: “Both Hindus and Buddhists agree that the overall structure of the cosmos
(that is, the rules by which it operates) will remain unchanged.”
 dharma over time, i 251

Creator, for Hindus Brahmā is the Creator. Moreover, in the fully developed form
of what Biardeau calls the epico-purānị c “universe of bhakti,” Brahmā’s divine
agency is subordinated to the still higher agencies of Viṣnụ and Śiva, with whom
his function as Creator is performed cyclically in conjunction with their higher
roles as Preserver and Destroyer, respectively. In these contexts, where the three
deities together can be called the Trimūrti (the “Three Forms” of Brahman or the
“Triune Godhead”), a mahākalpa or “great aeon” becomes the term for a Life of
Brahmā: a hundred years of three-hundred-and-sixty Brahmā Days and Nights
during which the universe formed from the primal Elements as an Egg of Brahmā
(brahmānḍ ạ ) survives until it dissolves back into those Elements. Meanwhile,
kalpas or aeons—with kalpa having become the standard term for a Day of
Brahmā—bubble up, and for each such Day, Brahmā (or Viṣnu ̣ in the form of
Brahmā) (re)-awakens to (re)-create a triple world of earth, atmosphere, and
heaven, which lasts until it is dissolved into “the single ocean,” after which come
Brahmā Nights of equal duration. Thus whether or not Brahmā is really the
Creator, he appears at the beginnings of kalpas.
As in Buddhism, the Hindu kalpa also defines certain possibilities for
salvation that transcend Brahmā. For Buddhists, as noted, it assigns the
interval—a downswing period of decreasing life span within an intermediate
aeon of an incaculable aeon of the duration of renovation—during which a
Buddha may appear, reach nirvāṇa himself, and, at Brahmā’s encouragement,
teach a dharma that brings salvation to arahants. For Hindus, upon the dissolu-
tion of the triple world at the end of a kalpa, saintly beings can resort not to
Brahmā but to Nārāyaṇa, a form of Viṣṇu; and upon the dissolution of the egg
of Brahmā at the end of a Life of Brahmā or mahākalpa, all beings are released
by the dissolution of the Elements thanks to the ultimate divine agency of
Viṣṇu or Śiva as the Supreme Person (Puruṣottama). When a Brahmā dies (or
perhaps better depersonalizes himself and “becomes Brahman”)21 at the com-
pletion of a “great kalpa,” the Hindu dharma becomes irrelevant: matter itself
devolves leaving all beings liberated.22
If, as seems to be the case, these are the primary ways in which the two
traditions initially formulate these soteriological chronometries, it may be said
that epico-purāṇic bhakti Hinduism, or at least later Purāṇic bhakti Hinduism,

21. Cf. Biardeau 1981b, 91, 97–98, 140. Although Brahmā’s “death” seems to be poorly reported, if indeed
all beings include him, it would not be tasteful to exclude him from a universal salvation. I take his transmigra-
tory salvation to be implied by the swan or gander (haṃsa) he rides, which, with its etymological explanation “so
‘ham so ‘ham” (“resolvable to ahaṃ sa, ‘I am that’” [MW 1286]), indicates his oneness with the undying ātman.
22. See Biardeau 1981b, 39, 98, 111–15, 152–57; Hiltebeitel and Kloetzli 2004. It is interesting that
according to the Kevaddha Sutta (DN 11), Brahmā reveals with some embarrassment that he does not know
“where the four elements cease without remainder,” while the Buddha seems to treat the question as if it were
like those that do not lead to edification (see Walshe [1987] 1995, 178–80).
252 dharma

joins the Mahāyāna in converting the kalpa from structuring a drama that con-
cerns the salvation of individuals (once-in-a-blue-moon Buddhas and arahants)
into a drama that envisions universal salvation.23 Indeed, Purāṇic bhakti
Hinduism envisions universal salvation at the end of a mahākalpa (a life of
Brahmā) rather than in an improbably unusual sub-kalpa (intermediate aeon)
of a sub-kalpa (incalculable aeon) of a mahākalpa. We must, however, leave a
gap between epico-purāṇic and Purāṇic bhakti (cf. Biardeau 1994, 16–18), since
the Mahābhārata never explicitly connects the idea of a Life of Brahmā with a
mahākalpa or implies that universal salvation would come with it. Indeed,
González-Reimann points out that the epic uses the term mahākalpa only four
times “without ever defining it, and it doesn’t seem to have a specific meaning.”24
He goes so far as to say that the Purāṇas add the Life of Brahmā as a fourth
cycle beyond the kalpas, yugas, and Manvantaras.25
One passage in a highly devotional Mahābhārata segment called the
Nārāyaṇῑya is especially instructive in confirming such an open meaning of
kalpa in the epic. It occurs in two successive chapters or adhyāyas, and it is
important to look at how kalpa is used in them in context.26 First, in adhyāya
12.326, the Ṛsị Nārada has just returned, rather spectacularly, from White Island
(Śvetadvῑpa) with revelations about Nārāyaṇa, who resides there, and Yudhiṣṭhira,
who is hearing this account from his grandfather Bhīṣma, asks Bhīṣma how it
is that Brahmā seems not to know what Nārada has learned, and whether
Brahmā is really different from Nārāyaṇa (12.326.103). Bhīṣma replies,

Thousands of great aeons and hundreds of great aeons (mahākalpa

sahasrāṇi mahākalpa śatāni ca), and hundreds and thousands of
creations and dissolutions come after one-another. In the beginning
of a creation, the wise lord, the creation-maker Brahmā, is
remembered.27 He knows, moreover, that the best of gods, being the
supreme self (paramātman), the lord, as also source of the self,
is superior to him, O king. (12.326.104–5)

23. The Mahāyāna will promise salvation to all beings once all beings realize that the universe is empty,
but this will still require the appearance of Buddhas and the teaching of their dharma. Similarly, in Hindu
knowledge traditions, in principle, a realization of the Self’s oneness with Brahman shortcuts the skein of time.
24. See González-Reimann 2008, 9, presumably referring to the occurrences he mentions (3 and handout
Table) at Mbh 13.110.71 and in the Nārāyaṇῑya at 12.323.1 and 326.104 (with two).
25. González-Reimann 2002, 4, presumably identifying the mahākalpa (unmentioned here) with the Life
of Brahmā.
26. The contributors to Nārāyaṇīya Studien (Schreiner 1997) all work from the premise that adhyāya 326
ends “Part A” of the Nārāyaṇīya, and adhyāya 327 begins “Part B.” From this standpoint, the transition should be
pivotal, but it goes unexamined because their discussion focuses only on the supposed lateness of Part B.
González-Reimann 2008 discusses usages of kalpa in the two adhyāyas, but not in full context.
27. By Nārāyaṇa, says Ganguli ([1884–96] 1970, 10: 558), I think correctly; alternately, Esnoul 1979, 120:
“Au début d’une création on rapport que Brahmā émet les créatures.”
 dharma over time, i 253

Bhīṣma then recounts how Brahmā revealed this knowledge to other divinities,
and how it was further transmitted to Bhīṣma himself (326.106–11). Bhīṣma
then tells Yudhiṣṭhira that this story is transmitted within a Vāsudeva bhakti
tradition, and that Yudhiṣṭhira should take this opportunity to please Kṛṣṇa-
Janārdana, who is there listening, by worshiping him as Puruṣottama, the
supreme divinity (113–20)—after which “Dharmarāja and his brothers all
became dedicated to Nārāyaṇa (nārāyaṇaparābhavan)” (121). The story seems to
link “remembering” Brahmā with the beginnings of great aeons, and thus,
perhaps implicitly, with Brahmā Days. But nothing is said about such days,
much less their adding up to a Brahmā’s life.
Then in its next chapter, the Nārāyaṇīya comes to a surprising lurch.28 Up
through adhyāya 326, the Nārāyaṇīya has been moving along at the epic’s main
dialogue level. What Bhiṣma and Yudhiṣṭhira have been saying to each other
has been recounted by Vyāsa’s disciple, the Brahmin Vaiśaṃpāyana, to the
Pāṇḍavas’ descendant, King Janamejaya, who is listening to the epic while
performing a vengeful snake sacrifice. It is customary to call this dialogue level
the epic’s inner frame. Now, however, that Vaiśaṃpāyana has finished the story
of White Island, which has involved so many revelations about Nārāyaṇa, sud-
denly the dialogue level shifts to the outer frame in which Śaunaka, the leader
of the Ṛṣis of the Naimiṣa Forest, is listening to the Mahābhārata being
recounted by the bard Ugraśravas, who had come to Naimiṣa Forest to tell the
Ṛṣis there “the same” Mahābhārata that Vaiśaṃpāyana had told Janamejaya.
One has barely heard from this second dialogue level since the Mahābhārata’s
first book, the Ā diparvan, but now the Brahmin Śaunaka, the “anchorman” of
the Naimiṣa Forest sages, chimes in out of the blue,29 and asks Ugraśravas the
first question that has come to mind from hearing the White Island story.
Śaunaka asks about Nārāyaṇa. How, while he is “established in nivṛtti
dharma, enjoying peace, ever the beloved of Bhāgavatas,” do the other gods
come to accept sacrificial shares according to pravṛtti dharmas, while nivṛtti
dharmas are “made for those who have turned aside” (327.2–3)? Here we meet
two broad orientational terms using dharma: pravṛtti-dharma, generally the
normative lifestyle for the householder, and nivṛtti-dharma, the normative
lifestyle for the ascetic. As Bailey (1985) demonstrates, the two include

28. See Hiltebeitel 2006a, 229–49, and, with some corrections, 2011a, chapter 7, discussing this and two
further “dips” to the outer frame erased in the Pune Critical Edition (CE) because the Śāntiparvan editor Belvalkar
(1954–66) relied, I argue mistakenly, on an erasure of the outer frame based on Malayālyam manuscripts. See
Brodbeck 2009a, 236 n. 12, assessing my correction back to the outer frame as “convincingly argued.”
29. Thanks to Wendy Doniger for coming up with the designation “the Mahābhārata’s Anchorman” for
Śaunaka (American Aadamey of Religion Annual Meeting, Washington DC, November 2007). On Naimiṣa
(Winking, Blinking, Twinkling) Forest as a celestial place whose Ṛsị s double as stars, see Hiltebeitel 2001a, 95–161.
254 dharma

ideological implications that are developed, with semantic variations, in both

Brahmanical and Buddhist texts. Literally, pravṛtti means the “act of rolling
onwards” or “turning outwards,” implying modes of engagement with the
world consonant with dharmashastric norms. Nivṛtti, the “act of turning back”
or “returning,” implies liberation from the world through cessation or aban-
donment of activity consonant with dharmashastric norms. A similar term
implying tension with dharmashastric norms is mokṣadharma, which is also
highlighted in adhyāya 327, where it seems to showcase what the whole
Mokṣadharma Parvan is driving toward in using both of these terms.30 While
usages of pravṛtti and nivṛtti are fairly clear by being contrastive, usages of
mokṣadharma seem paradoxical. In a thought-provoking article on the tensions
between sādhāraṇadharma and varṇāśramadharma as both being worldly, and
mokṣadharma, Gerald Larson describes the latter as the dharma that “does not
fit” (1972, 149). Adam Bowles notes that nivṛttidharma overlaps in the
Mahābhārata with mokṣadharma, and remarks that the latter looks at first blush
“like an oxymoron” (2007, 153). I do not think, however, that it was meant not
to fit or to be as oxymoronic as it first looks. We can consider these questions
further, however, in chapter 13.
Now when Ugraśravas answers Śaunaka’s question about pravṛtti and
nivṛtti, he does so by telling him what Vaiśaṃpāyana said when Janamejaya
asked him “the same” question. Janamejaya had asked:

These worlds with Brahmā, men, gods and demons are seen
everywhere to be attached to rites said to assure prosperity.
And mokṣa is said by you, O Brahmin, to be nirvāṇa, the supreme
happiness. . . . Alas, the eternal mokṣadharma is surely difficult to
observe (aho hi duranuṣṭeyo mokṣadharmaḥ sanātanaḥ), abandoning
which all the gods have become enjoyers of rites to gods and ancestors.
How then do Brahmā, Rudra [and other gods . . . ] not know that
dissolution is determined of the self. By that, they are then settled on
a firm, indestructible, undecaying path.31 Those who have remembered
the measure of time are intent upon pravṛtti. Meting out time, that
is the great fault of those given to activity.32 That is my doubt,

30. I take up these matters in Hiltebeitel forthcoming-b, in a collection edited by James Fitzgerald.
31. The close of Janamejaya’s first question in verse 9 is a bit obtuse: pralayaṃ na vijānanti ātmanaḥ pari-
nirmitam/ tatas tenāsthitā mārgaṃ dhruvam akṣayam avyayam (12.327.9). Ganguli ([1884–96] 1970, 10: 561 n. 2
comments: “‘Atmanah parinirmitam pralayam’ means that destruction or cessation of existences brought about
by self-realization. . . .” Esnoul 1979, 125: has Janamejaya asking how the other gods “. . . . ne distinguent-ils pas
que la résorption est déterminée par leur nature même.” The last line is not very clear.
32. 12.327.10: smṛtvā kālaparīmaṇaṃ pravṛttiṃ ye samāshitāḥ/ dosaḥ kālaparīmāṇe mahān eṣa kriyāvatām.
 dharma over time, i 255

O Brahmin, like a vexing thorn in my heart. Cut it by a story of history

(chindhi itihāsakathanāt). My curiosity is surely peaked. Why do those
deities addressed in sacrifices take away (sacrificial) shares? To what
end are the heaven-dwellers worshiped in sacrifice, O Brahmin? Those
who take the share in sacrifices, O best of twiceborns, those sacrificing
with great sacrifices, to whom do they give a share? (12.327.5–13)

That Śaunaka interrupts the main dialogue level or inner frame means that the
epic poets have interrupted Vaiśaṃpāyana’s main narration to take us into
something deep, and, moreover, something that Vaiśaṃpāyana did not tell
Janamejaya in that main narration, but outside it—let us say, during an aside.
Janamejaya is performing his snake sacrifice, which is a pravṛtti rite par excel-
lence that involves “meting out time” for the destruction of all snakes, because
a snake killed his father. His question is as existentially profound as we can
expect of him. Not to disappoint him, Vaiśaṃpāyana answers him by quoting
no one less than Vyāsa himself, the author (who will be revealed a little later in
the Nārāyaṇīya to be a “portion of Nārāyaṇa” himself [12.337.4], and who,
despite his great age, is a seated attendee at Janamejeya’s snake sacrifice33). We
now learn that Vyāsa’s five disciples (with Vaiśaṃpāyana among them) once
had the “very same doubt” while Vyāsa was with them in his hermitage on
Mount Meru (327.16–19), where he composed the Mahābhārata. Again, pre-
sumably during some “aside,” they asked him this question and he gave them
a lengthy reply (327.21–98). Śaunaka’s interruption in adhyāya 327 has thus
taken us not only from the inner frame to the outer frame, but to the originating
outermost frame of the Mahābhārata to get us to the bottom line—the
Mahābhārata author himself—on this “vexing” question.
Vyāsa is reported to have told his five disciples about what Brahmā and the
gods and Ṛsị s once learned when they went to the northern shore of the Milky
Ocean to ask Nārāyaṇa about such matters. There, Nārāyaṇa remarked that he
has consigned the gods to receive offerings until the end of the kalpa (yāvat
kalpakṣayād; 327.60d) according to pravṛtti dharma for the welfare of the world.
To this end, he has assigned seven mindborn Ṛsị s—Marīci, Aṅgiras, Atri,
Pulastya, Pulaha, Kratu, and Vasiṣṭha—to procreation following pravṛtti-dharma
(326.60–62). And he has assigned seven other Ṛsị s—Sana, Sanatsujāta, Sanaka,
Sanandana, Sanatkumāra, Kapila, and Sanātana (“called mental sons of
Brahmā”)—to be “established in nivṛtti dharma” as “the foremost of yoga-
knowers, as also knowers of the Sāṃkhya-dharma,” and “preceptors in mokṣaśāstra
and promulgators of mokṣadharma (mokṣadharmapravartakāh)̣ (64–65). This is a

33. On Vyāsa as attendee, see Mbh 1.48.7ab and Hiltebeitel 2001a, 115 n. 71.
256 dharma

group of perennially liberated Ṛsị s. Finally, Vyāsa tells what happened when all
the other heaven-dwellers but Brahmā had left. When Brahmā remained in
place, Nārāyaṇa appeared to him, “having assumed the great Horse’s Head
(Hayaśiras), reciting the Vedas with their limbs” (327.80–81). The Horse’s Head
now reinforces the distinction between nivṛtti and pravṛtti: he charges Brahmā
to oversee pravṛtti as the “world’s creator” (lokakartā), and promises, before
vanishing, that he himself, Nārāyaṇa, will intervene with various manifestations
(prādurbhāvas)—later, in the Purānạ s, to be called avatāras34—to bear the work
of the gods (surakāryam) whenever things get intolerable (82–86b). Vyāsa then
continues:

So it is that this one of great share, the eternal lotus-naveled one. . . ,

the eternal upholder of sacrifices, has fixed nivṛtti dharma, which is the
destination of those whose teaching is the imperishable. He has (also)
ordained pravṛtti dharmas, having made for the world’s diversity. He is
the beginning, middle, and end of creatures; he is the ordainer and
the ordained, he is the maker and the made. At the end of the yuga he
sleeps after having retracted the worlds; at the beginning of the yuga
he awakens and creates the universe. (12.327.87–89)

Finally, Vyāsa lauds Nārāyaṇa, guarantees that all this is true, and exhorts his
disciples to sing Hari’s praise with Vedic words (327.90–98).35
Now we have just seen Vyāsa use the phrases “end of the yuga” and “beginning
ot the yuga” to describe what happens when Narāyaṇa goes to sleep and awakens,
where he might, were he seeking “Purānic” consistency, have used the term
kalpa. Toward the beginning of Vyāsa’s speech, he has also told what he learned
by going to the northern shore of the milky ocean and performing “the height of
difficult tapas” there. Thanks to Nārāyaṇa’s grace, he obtained the “triple-timed
knowledge” he desired of past, present, and future (12.327.21–23b). With that
knowledge, he now recounts a primary creation down to the great elements
(23c–28), which might suggest a mahākalpa, even though he only uses the term
kalpa for the beginning of it (yathā vṛttaṃ hi kalpādau; 12.327.23e). Finally, as
noted, he also uses the term kalpa when he gets to the end of such a cycle in the
phrase “until the destruction of the kalpa” (yāvat kalpakṣayād; 327.60d). With this

34. For discussion, see chapter 11 § C. Cf. González-Reimann 2008, 6: “we hear that Nārāyaṇa, after declaring
that he created Brahmā at the beginning of every kalpa and that he does so repeatedly in every kalpa, lists his several
appearances on Earth, in what is probably the earliest list of avatāras, although the term used here is prādurbhāva.”
35. With regard to Nārāyaṇa’s taking on this Veda-reciting form “for the sake of pravṛttidharma”
(12.335.68c), the Nārāyaṇīya further discloses that “Nārāyaṇa’s supreme dharma is difficult to reach by those
turning around again. The dharma characterized by pravṛtti also has Nārāyaṇa’s nature (nārāyaṇaparo dharmaḥ
punarāvṛttidurlabhaḥ/ pravṛttilakṣaṇaś caiva dharmo nārāyaṇātmakaḥ)” (12.335.76c–77b).
 dharma over time, i 257

latter phrase as a touchstone, González-Reimann has noticed something very

interesting about these usages of kalpa in the Nārāyaṇīya.
As González-Reimann sees it, adhyāya 12.327 “is of special interest” as a
“discourse on the relative merits of two kinds of dharma, pravṛtti and nivṛtti.”
Therein, Vyāsa explains36 why Viṣṇu/Nārāyaṇa propounds both orientations,
and eventually quotes Viṣṇu himself “that the gods should fulfill their allotted
duties until the end of the kalpa, thus ensuring that the world continues to
operate properly” (2008, 7). Says González-Reimann,

This statement about the need to uphold the brahmanical dharma

until the end of the kalpa is intriguing. It is equivalent to a declara-
tion that is repeated in two of Aśoka’s stone inscriptions. In that
case, King Aśoka declares that his promotion of the Buddhist
dharma has improved life in the world, and that the practice of the
dharma will continue to increase thanks to his children, grandchil-
dren, and great-grandchildren until the end of the kalpa. It appears
possible that the Nārāyaṇīya’s emphasis on the need to continue
the brahmanical dharma until the end of the kalpa is an answer to a
proclamation like that of Aśoka regarding the Buddhist dharma’s
duration during the same period. Besides the thematic similarity,
there is a striking grammatical parallelism as well. The Sanskrit
expression used in the Epic is yāvat kalpakṣayāt, literally meaning
“until the destruction of the kalpa.” The correlative yavāt is followed
by the compound kalpakṣaya in the ablative case, for the purpose of
indicating duration up to a point (the end of the kalpa). Nowhere
else in either of the two epics is the word kalpa used like this.
However, Aśoka’s inscription uses the Prakṛt āva kappa, or kappam,
where āva is the Prakṛt equivalent of Sanskrit yāvat, and kappa is
used adverbially. But to make the parallelism even more striking,
the Girnar versions of both inscriptions has the Prakṛt āva
saṃvaṭṭakappā (Sanskrit yavāt saṃvartakalpāt), where
saṃvaṭṭakappā, “the destruction of the kalpa,” is in the ablative
case,37 as in the Nārāyaṇīya’s Sanskrit version. (González-Reimann
2008, 9)

36. González-Reimann has Vyāsa address Janamejaya directly, but Vaiśaṃpāyana only quotes him here.
Yet Vyāsa does soon address Janamejaya in the Nārāyaṇīya’s third dip to the outer frame, after Śaunaka asks his
“bottom-line question” about Viṣṇu’s Horse’s Head manifestation (Mbh 12.335.21–66; see Hiltebeitel 2006a,
246–49).
37. González-Reimann notes that Bloch 1950, 100 n. 9 indicates that the Girnār inscription “uses what he
describes as an archaic ablative.”
258 dharma

We are, in other words, on some familiar ground here with a surprising new
argument based on the Aśokan usages.
Let us start out by recognizing, with González-Reimann, that the parallelism
is grammatically real, though exceptional in both cases, especially in the appeal
to the singular Girnār edicts for the closest analog on the Aśokan side. Let us
also note that González-Reimann is careful to say that the Nārāyaṇīya’s emphasis
on continuing the brahmanical dharma until the end of the kalpa might not be
a direct answer to the Aśokan proclamation but an answer to “a proclamation
like that of Aśoka regarding the Buddhist dharma’s duration during the same
period” (my italics). That leaves room for all kinds of intermediaries, and for the
Nārāyaṇīya to be as late or early as one might wish. Clearly, it would be difficult
to hold that the Nārāyaṇīya passage was responding directly to the Aśokan edicts,
not to mention the Girnār one in particular, if the Nārāyaṇīya were composed in
the Gupta period, when (as we saw in chapter 2), it seems likely that few if any
could have read the Aśokan edicts any longer. I have produced some arguments
that, even though the Nārāyaṇīya would probably be late within the two-
generation span I assign to the Mahābhārata’s composition, it would still be part
of the baseline archetype exposed by the Critical Edition.38 But I would not see
this parallelism as evidence that the Nārāyaṇīya would be close in time to Aśoka.
Exceptional as the two usages are, they are not so surprising grammatically as to
require a direct relation. This leaves us then with the question of whether the
Nārāyaṇīya might be responding in this usage to Buddhist ideas. Here, although
I think a general answer would be yes, and that González-Reimann may be right
to include this passage among others that are more convincing on this point,
I think it is more significant to recognize that the ideas espoused are likewise
exceptional in both traditions. As pointed out earlier in this chapter, Aśoka’s
idea, described by González-Reimann as saying that the Buddhist dhaṃma will
last to the end of the kappa through his descendants, actually pertains only to
RE 4. By RE 5; he seems to be about to drop it. In relation to more conventional
Buddhist ideas about kalpas, it would seem, in its basic form using only āva
kappa or kappam, to be an unusual extrapolation from them meant to make
what I called a bad fit with Aśoka’s mid-career aspirations for the Mauryan
dynasty. And the modification at Girnār looks like someone’s effort to make
some sense of it in technical Buddhist terms. Meanwhile, on the Brahmanical
side, the Nārāyaṇīya is equally atypical in making a connection—a looser one
than González-Reimann implies, but one nonetheless—between dharma and
the full course of a kalpa. Yes, pravṛtti dharma will continue to the end of the

38. See Hiltebeitel 2006a and 2011a, chapter 7; 2010d, arguing that it would be earlier than Aśvaghoṣa’s
first- or second-century CE Buddhacarita; and forthcoming-a.
 dharma over time, i 259

kalpa, but as something that concerns only the gods who receive sacrifices over
a very long time that has little to do with how human beings lead human lives
through such things as families, dynasties, history, and yugas. The Nārāyaṇīya is
on an entirely different plane here from REs 4 and 5.
The Nārāyaṇīya may thus be said to leave the gap we have been mentioning
between the epics’ open treatment of small-to-vast units of time and the more
systematic treatment of such units that one meets in the Purānạ s. A word,
however, is indispensible here on how the Mahābhārata takes such temporal
matters as only part of a daunting meditation on time that engages the question
of dharma over time in multiple intersecting ways.39 In the Nārāyaṇīya, for
instance, when Śaunaka summarizes what he has learned so far before asking
his third question about the Horse’s Head, he mentions what he has learned
from his second question about the bewildering relation between Nārāyaṇa in
White Island and the Ṛsị s Nara and Nārāyaṇa at their Badarī hermitage in the
Himalayas. There, as he puts it, Nārada had heard of Nārāyaṇa’s “birth in
the house of Dharma in the person of Nara and Nārayaṇa.”40 What is meant by
the house of Dharma? Unfortunately, Śaunaka did not make it a fourth question,
but it is worth asking. The Nārāyaṇīya’s first two mentions of the house of
Dharma connect it with the Pāñcarātra doctrine of the four vyūhas: the four
forms or bodies in which Nārāyaṇa—as Vāsudeva, Saṃkarṣaṇa, Pradyumna, and
Aniruddha—takes part in the process of primary creation. First, Nārada says,
“That object for which you, O self-born being, have taken birth fourfold in the
house of Dharma (dharmagṛhe caturdhā), may it attain its end for the welfare of
the world” (2.322.2a–c). And then Nārāyaṇa says, “These are my best bodies, O
twiceborn, born in the house of Dharma” (mamaitās tanavaḥ śreṣṭhā jātā
dharmagṛhe dvija; 326.13ab), before going on to describe the four vyūhas in their
cosmogonic roles (326.23–43). One gets to Nara and Nārāyaṇa being born in the
house of Dharma only later, in the build up to Śaunaka’s second question. This
occurs when Nara and Nārāyaṇa, speaking unitedly in Badarī, tell Nārada:

We two who have taken birth in the house of Dharma (āvām api ca
dharmasya gṛhe jātau), O best of twiceborns, having resorted to this
spacious delightful retreat, are engaged in fierce tapas. As to the
manifestations who are surely dear to the gods, that will come about
in the three worlds, may it be well for them (bhaviṣyanti trilokasthās
teṣāṃ svasti ity ato). (12.332.19–20)

39. See Vassilkov 1999 on Vyāsa, the author, as kālavādin; Hiltebeitel and Kloetzli 2004 for the epic’s
innovations in “narrative time.”
40. 12.335.1cd: janma dharmagṛhe caiva naranārāyaṇātmakam; see Hiltebeitel 2006a, 245.
260 dharma

It would seem that Nara and Nārāyaṇa’s birth in the house of Dharma puts them
at the same primal plane as the cosmogonic vyūhas, and that their relation is
something more primordial and elevated than the triple world, where the manifes-
tations will come into being (bhaviṣyanti) for its periodic welfare.41 Yet their rapport
also has its backing in the passage where Nārāyaṇa foretells his manifestations:

When all the powerful kings of the earth have come together, Vāsavi
[Arjuna, Vāsava-Indra’s son] will surely be my good companion. So
people will call us the Ṛṣis Nara and Nārāyaṇa; joined together, we
two lords (īśvarau) will burn the warrior class for the sake of the
world’s cause. I will make a terrible dissolution (kariṣye pralayaṃ
ghoram), destroying my own kinsmen. Bearing four forms
(caturmūrtidharo), having done immeasurable acts, I will go to the
worlds of mine that are honored by Brahmins. (12.326.90–93)

It would seem, then, that the house of Dharma brings the primordiality of the
vyūhas together with at least one of the manifestations: that of Kṛsn ̣ ạ . And with
that, we could be reminded that Arjuna might be said to be born into the house
of Dharma in that his older brother Yudhiṣṭhira’s father is precisely the god
Dharma. This is probably not what is meant, but there may be some resonance,
at least for Yudhiṣṭhira as he hears all this. From what we have gathered so far,
I would propose that the house of Dharma is the created universe seen from the
standpoint of Nārāyaṇa’s institution, for it, of both pravṛtti and nivṛtti dharma.
Fortunately, however, we do have one more clue to pursue, in that the
Nārāyaṇīya also knows of Dharma as one of twenty-one Prajāpatis who emerge
more or less coevally into the creation.42 This is the second of two narratives
I will now examine, by which the Mahābhārata describes the origins of dharma
in cosmogonic time. I will then turn to Manu’s treatment of dharma in relation
to the kalpa, yuga, and Manvantara, where it will be also possible to briefly
recapitulate the Mahābhārata’s treatment of these chronometric terms in the
Nārāyaṇīya and elsewhere.

C. Originary Dharma in the Mahābhārata

The Mahābhārata gives its first glimpse of the primordiality of dharma in the
early front matter of the Ā stīkaparvan, which as told from the standpoint of

41. Note that Grünendahl 1997, 232–40 ascribes epic vyūha representations to a “late” textual “layer” intro-
duced by a school of “epic Pāñcarātrins” that would nonetheless be earlier than the Pāñcarātra sect, which may
be from Gupta times.
42. The list (12.321.33–35) begins “Brahmā, Sthāṇu, Manu, Dakṣa, Bhṛgu, Dharma, Tapas, Restraint
(dharmas tapo damaḥ)” (33ab), for which “Dharma and Yama” (dharmas tathā yamaḥ) is a Northern variant.
 dharma over time, i 261

outer frame, sets the background for Janamejaya’s Snake Sacrifice at which the
epic will first be told. Ugraśravas tells Śaunaka how the Nāga (serpent) Śeṣa
separates himself from other Nāgas because they are too filled with enmity.
Śeṣa performs austerities (tapas) until Brahmā notices. Once Brahmā questions
him and is gratified by the explanation, he offers Śeṣa a boon:

I wish to grant you a boon, for I am very pleased with you. It is by

good fortune, O best of snakes, that your mind is set on dharma
(diṣṭyā ca buddhir dharma te niviṣṭā), and on dharma it shall be set
ever firmer. (1.32.15c–16)

Since Śeṣa had never mentioned the term dharma, we may say that Brahmā’s
exegesis on Śeṣa’s tapas as dharma puts the idea into the serpent’s head.43 Śeṣa
adopts it.

This is the boon I desire, Great Grandfather. May my mind rejoice in

dharma, and in tranquility (śama) and austerity, Lord! (32.17)

Brahmā credits Śeṣa that this will be for the well-being of creatures. He
appoints him to encircle the earth with her mountains and oceans and to bear
her on his head, and they work out the details (18–22). Brahmā then con-
cludes, telling him, “You are Śeṣa, the best of snakes, god of dharma
(dharmadeva), for you alone uphold (dhārayase) this earth” (23ab). He thus
offers an etymological link between Śeṣa’s appointment as dharmadeva and
his “upholding” (√dhṛ) of the earth. The etymology is then repeated by
Ugraśravas, who now calls Śeṣa Ananta, making this “god of dharma” “the
endless (ananta) remainder (śeṣa),” and adding that Brahmā gave Ananta the
friendship (sakhāyam) of Garuḍa (24–25).
Śeṣa is identified with the second vyūha Saṃkarṣaṇa in the Pāñcarātra
doctrine unfolded in the Nārāyaṇīya, and when Kṛsn ̣ ạ ’s brother Balarāma-
Saṃkarṣaṇa dies while engaged in yoga, a white snake leaves his mouth and
enters the ocean, where it takes on a mountainous size with a thousand heads
(Mbh 16.5.122–23). From this angle, the initial story about Śeṣa could be located
in the formation of what the Nārāyaṇīya calls the house of Dharma, which would
place Balarāma in that house along with Nara and Nārāyaṇa. But the passages
in Books 1 and 16 leave such interpretation to readers. The Ā stīkaparvan narra-
tive is explicit only about linking Śeṣa with dharma in upholding the earth in
space. But it also implies that he upholds the earth through time, and thus rep-
resents something of the continuity of dharma over time, particularly through

43. Again we may raise here the question “Do animals know dharma?” on which the epics, at least explic-
itly, say that they do not (see chapter 5 n. 20 and chapter 8). Obviously, some animals (given time and the
occasion) appear to be exceptions.
262 dharma

the Days and Nights of Brahmā, or what will be more regularly called kalpas in
the Purānạ s. As the “Endless Remainder,” Śeṣa comes to represent the conti-
nuity of dharma not only through Brahmā Days, when he can uphold the earth,
but through Brahmā Nights, when, with the earth having undergone its periodic
dissolutions into the “single ocean,” he can rise to the surface of that ocean to
form the couch for the sleeping Viṣnu ̣ -Nārāyaṇa, at whose awakening Brahmā
is “remembered.” In this, the Śeṣa narrative is more primal than our next
sequence, but since both Śeṣa and Garuḍa are born into primal genealogies, it is
also in that regard part of our second story as well.
This second narrative calls to mind ways in which the Mahābhārata is in
tune with a cosmogonic sequence that opens the Bṛhadāraṇyaka Upaniṣad
passage (1.4.11–14) in which Brahman is not “fully developed” until it has cre-
ated dharma (chapter 3 § F). The epic also considers dharma something to
account for mythically as a coeval but still not a primary component of the cos-
mogonic process. As noted in the case of the Nārāyaṇīya, the Mahābhārata
knows Dharma as a Prajāpati: a designation that can refer to Brahmā (who is
the first of the Prajāpatis mentioned in the Nārāyaṇīya’s list of twenty-one)—
and also, in contrast to Brahmā’s mindborn sons who remain celibate, “to those
sons of Brahman [Brahmā] who beget children” (Ganguli [1884–96] 1970,
9: 89). The epic tells how such a personified Dharma emerged in Athena-like
fashion—not, however, from the brow:

The blessed Lord Dharma, assuming human form, issued forth

by breaking open the right nipple of Brahmā (stanaṃ tu dakṣiṇam
bhittvā brahmaṇo naravigrahaḥ), bringing happiness to all the
worlds (1.60.30).44

Further, it tells several times how Kaśyapa, Dharma, and Soma (the Moon)
become husbands of the fifty daughters of the Prajāpati Dakṣa. Thirteen of
Dakṣa’s daughters marry Kaśyapa (the primal Ṛṣi who fathers Śeṣa and the
Snakes and Garuḍa and the Birds with the sisters Kadrū and Vinatā), ten marry
Dharma, and twenty-seven wed Soma. Dakṣa’s daughters are putrikās (1.60.11c;
70.7b), which means that he sires them to have sons for his own descent rather
than that of their husbands. Not surprisingly, Manu knows the same story,
giving it two verses (9.128–29) as background to his discussion of putrikās.45
What is pertinent presently is that the wives marrying Dharma and the Moon

44. Thus van Buitenen, trans. 1973, 149, with “nipple.” Cf. Hopkins [1915] 1969, 199, taking it to be Brahmā’s
“heart,” which is somewhat implausible, as it is a matter of the “right” side of his “breast” or “chest” (stanam).
45. For discussion, see Brodbeck 2009a, 48–56, 63–65, contrasting putrikās with pativratās, and 89–90, 102,
on Dakṣa’s daughters, but focusing only on the ones who marry Soma. See also Hopkins [1915] 1969, 190, 199.
 dharma over time, i 263

bring about the conjunction of beings (the generations born of Kaśyapa)46 with
dharma over time. Dharma’s wives and offspring enter the world as qualities or
virtues that coincide with the ordering of time that takes place through the
Moon’s marriage to the twenty-seven daughters of Dakṣa who are the twenty-
seven lunar mansions through which the Moon passes in measuring time
through the month. Dharma’s ten wives are:

Fame (Kīrti), Fortune (Lakṣmī), Resolve (Dhṛti), Intelligence

(Medhā), Growth (Puṣṭi), Faith (Śraddhā), Religious Action (Kriyā; see
MW 321), Intellect (Buddhi), Shame (Lajjā), and Mind (Mati). They
are the doors to Dharma that have been ordained by the Self-
Existent. (1.60.13–14)

Meanwhile, the twenty-seven “faithful wives of the Moon are appointed to the
procession of Time (kālasya nayane yuktā); they are all the fairies of the lunar
mansions (sarvā nakṣatrayoginyo), which regulate life in the world”
(60.15–16).47

Dharma is then said to have . . . three sons, fetching the hearts of all
creatures—Tranquility (Śama), Desire (Kāma), and Joy (Harṣa), who
sustain the world with their glory. Desire’s wife was Pleasure (Rati),
Tranquility’s was Possession (Prāpti), and Joy’s Was Delight (Nandī),
upon whom the worlds are founded (yatra lokāḥ pratiṣṭhitāḥ).
(1.60.31–32)

Desire thus flows from Dharma, and, once Desire is joined with Pleasure,
implying sexuality, Kāma and Rati become a famous mythological couple.
Nonetheless, Dharma’s sons are as abstractly named and disembodied as their
parents (even Kāma becomes “the bodiless God” when Śiva burns him to
ashes). Indeed, with Tranquility marrying Possession (Prāpti), we may be
reminded that the latter has the name of a prominent term among the fourteen
“dissociated forces” (dharmas) of Sarvāstivādin Buddhism that are neither

46. Kaśyapa’s wives give birth to Beings: that is, gods, demons, snakes, birds, fish, animals, Gandharvas,
Apsaras, plants, etc. (1.59.12–50; 12.200.25–28). Men through Manu as a descendant of Kaśyapa are included at
1.70.9–11.
47. Van Buitenen trans. 1973, 148. The point is reinforced just after this, where Dhruva, the Pole Star,
“begot our lord Time, who is the reckoner of the world (dhruvasya putro bhagavān kālo lokaprakālanaḥ)”
(1.60.20cd). Cf. 1.70.8 ff. and 5.106.4–6, with shorter accounts that make these same correlations, the latter
around the theme of the sunrise in the east, “where there are the two eyes of Dharma and there dharma is itself
established” (4cd; Hopkins [1915] 1969, 105 says Dharma’s two eyes are Soma and Agni). 9.34.40–58 is about
Dakṣa demanding that Soma spend equal time with all twenty-seven of Dakṣa’s daughters, rather than preferring
Rohiṇī. 12.200.17–29, explains the time component instead through Viṣṇu. At 12.329.45–46 (in prose), the
Nārāyaṇīya also has the Rohiṇī story, and gives Dakṣa sixty daughters, with the extra ten going to Manu.
264 dharma

physical nor mental,48 and wonder what the outcome or purpose of such an odd
couple might be. Apparently, on the analogy of Desire and Pleasure, Tranquility
comes with Possession or Ownership, which could hardly be a normative
Buddhist view. Indeed, it can be said that what the Mahābhārata personifies as
wives, sons, and daughters-in-law of Dharma are abstract impersonal dharmas
for the Sarvāstivādins. In any case, since Desire, Tranquilty, and Joy have
putrikā mothers, they would be technically sons of Dakṣa—unless Dharma
produces them without his wives, which could be indicated by their being
mentioned separately from his marrying them.49
Adharma, on the other hand, apparently emerges without parents, but still
finds his genealogical niche. After Liquor (Śurā) was born from Varuṇa and his
wife Jyeṣṭhā, Adharma was born from creatures (prajā) who began devouring
each other when they were hungry. Marrying Nirṛti, after whom Rākṣasas are
called Nairṛtas, she (or they?) had three terrible sons: Fear (Bhaya), Panic
(Mahābhāya), and Death (Mṛtyu) (1.60.51–53).
The Mahābhārata can tie such matters in loosely with the yugas. This is
done, for instance, at Mbh 12.220, an adhyāya indicative of an increasingly
devotional momentum in the Mokṣadharma Parvan that culminates in the
Nārāyaṇīya. Once this chapter tells how Dakṣa’s fifty daughters marry Kaśyapa,
Dharma, and the Moon, it bypasses discussion of the wives of Soma and
instead attributes the creation of “day and night, time, the seasons, morning
and evening” to Kṛṣṇa-Viṣṇu-Madhusūdana (200.17–29). Then, the origin of
the four varṇas (a hundred Brahmins, Kṣatriyas, Vaiśyas, and Śūdras come
respectively from Viṣṇu’s mouth, arms, thighs, and feet; 31–32) leads into an
account of the yugas. Sexual union (mathuno dharmaḥ) is traced back only to
the Dvāpara yuga (35–37). And “those without supervision” (niradhiyakṣān),50
that is, outliers in each of the four directions including the Greeks (yauna),
Kāmbojas, Gāndhāras, Kirātas, and Barbaras in the north, are traced back to
the prior Tretā yuga. These were all evil-doers (pāpakṛtas) who “move on this
earth having the nature (dharman) of dogs, crows, ravens, and vultures”
(śvakākabalagṛdhrānāṃ sadharmāno); they did not exist on earth in the Kṛta
yuga, but originated and started multiplying in the Tretā yuga, and contributed

48. Van Buitenen 1973, 149 calls her Attainment. See chapter 4 § B.2.c, translating prāpti as “possession”
or “ownership.” Cf. Hopkins [1915] 1969, 199, translating this Prāpti as “Possession.” The Mbh also gives Prāpti
as the name of a daughter of Jarāsaṃdha—perhaps, as I have said elsewhere, with Buddhist overtones as well (see
Hiltebeitel 1989, 97–98).
49. Elsewhere Dharma’s sons with Dakṣa’s ten daughters are groups of gods: the Vasus, Rudras,
Viśvedevas, Sādhyas, and Maruts (Mbh 12.200.23).
50. Mbh 12.200.38c. The term, which has some variants, is contrasted in the same verse with those who
are svadhyakṣa, “good to be inspected” (MW 1280).
 dharma over time, i 265

to unspecified events51 that occurred at “the terrible twilight time at the end of
that yuga” (tatas tasmin mahāghore saṃdhyākāle yugāntake) (38–43).
But the main account of the marriages of Dakṣa’s daughters that I have been
following is narrated as background to the “divine plan” behind the epic’s main
narrative. At some indeterminate point, but implicitly not too long ago when
demonic forces had already beset the world, the goddess Earth came to Brahmā
to ask help in relieving her of the burden caused by the demons who had taken
birth in human lineages upon her lest she be submerged prematurely in the
waters of time. We are assured that “Earth’s business (bhūmeḥ kṛtyam)52 had long
before been known” to Brahmā, “For how could he, the creator of the universe,
fail to know entirely what is lodged in the minds of the worlds of Gods and
Asuras, Bhārata” (1.58.41–42). So with Nārāyaṇa Vaikuṇtha’s (Viṣnu ̣ ’s) concur-
rence (49–51), Brahmā ordered a “descent of the portions” of the gods into
human lineages.53 Genealogies are then described for the generations of all
beings, with Dharma, as we have seen, among the more primordial gods. And
these genealogies come to focus on the lineage of the heroes of the central
Paurava-Bhārata-Kuru Dynasty into which the Pānd ̣ ạ vas and Kauravas will be born.
Though the divine plan will reach its concentration point in their generation, the
following two verses suggest a long time for it to have ripened: “And so the celes-
tials in succession descended (avateruḥ krameṇa) from heaven to earth for the
destruction of the enemies of the Gods and the well-being of all the worlds: there-
upon they were born in the lineages of Brahmin seers and the dynasties of royal
seers, as desired (yathā kāmam), O tiger among kings” (1.59.3–4). The role of the
gods’ desire (kāma) thus remains instrumental. It is important to recognize that
this account subordinates all genealogy to the divine plan.54
Since the Mahābhārata can be this diffuse and interwoven on topics of dharma
over time, it will be useful to defer our discussion of what it has to say on more
precise chronometry until we have considered their treatment in Manu. In this
regard, I urge consideration of the possibility that Manu may condense and
reorient the Mahābhārata’s take on these matters. Our starting point, however, is
to note that with Manu and the Mahābhārata, early classical Brahmanical texts
already extricate dharma from the rhythm of kalpas and assign the ups and downs

51. Ganguli ([1884–96] 1970, 9: 90 n. 3) says, “It was at this time that that dreadful famine occurred which
compelled the royal sage Viswamitra [Viśvāmitra] to subsist on canine haunch”—referring Mbh 12.139.
52. Also called the “cleansing of the earth” (bhuvaḥ śodhana; 1.58.51a).
53. 1.58.35–51. At 46–47, Brahmā orders the gods to incarnate through bhāgas, “portions” or “shares” of
themselves, as also the Gandharvas and Apsaras. But the more recurrent term is aṃśa, “portion” or “particle,” as
used in aṃśāvataraṇa, “descent of the portions,” summarily at 1.61.99c, and often echoed (see chapter 12).
54. See chapter 12 § A on the Mahābhārata divine plan in relation to divine plans in the Rāmāyaṇa,
Harivaṃśa, and even Manu, and scholarly resistance to it. See Hiltebeitel forthcoming-g. Such resistance has also
long sustained most Rāmāyaṇa studies; see most recently González-Reimann 2006a.
266 dharma

of dharma to the yuga. If we may speak of the Buddhist and Brahmanical tradi-
tions as presenting alternate dramas about the course of dharma over the vastness
of time, the Brahmanical tradition, by moving its more experiential chronometry
from the kalpa to the yuga, would have made this transferal a “wedge issue.” We
may thus ask what it might mean that Brahmanical chronometry detaches dharma
from the kalpa and links it instead with the yuga. To answer this, we should now
consider how the earliest classical dharma texts presented the yuga, kalpa, and
other temporal units before they were lodged into the systematic shape we find
assigned to them in the Purānạ s. I will argue that the Mahābhārata and Manu are
the texts that drove this wedge in its earliest formulations, and that the Yuga
Purānạ takes off from what the Mahābhārata has to say about yugadharma as the
basis for its ex eventu prophesies, which yield information about the historical
conditions under which this larger Brahmanical scheme was generated.

D. Kalpas, Manvantaras, and Yugas in Manu and the Mahābhārata

By Olivelle’s analysis, Manu’s first adhyāya (M 1) begins with thirty-one original

verses followed by two interpolated “excurses” totaling twenty-six verses, and
then, after three more original verses, continues with five more “excurses”
totaling fifty-nine verses (2005a, 53–54). As I indicated in chapter 5 § C, I believe
he underestimates his own findings on the structure of Manu for the text’s
compositional unity, and note that the first use of the signature line at M 2.25
by which Manu indicates divisions of the text can be taken to include the whole
of adhyāya 1, as Olivelle initially does himself (2005a, 7–9, 52). I believe
Olivelle’s willingness to name so many segments by different topics does not
give credit to the significance that cosmology, combined with taxonomy and
pedagogy, probably had for Manu in composing a unifying preamble to his
text; as I suggested in chapter 1, Manu may have drawn on sources from other
knowledge traditions to include some of these topics. I also believe that
Biardeau’s idea that Manu presents elements of primary and secondary cos-
mogony rightly directs us to consider Manu 1, even with its “patchwork”
character, as a whole. Manu contextualizes dharma with respect to units of time
with orderliness and statistical simplicity, but while getting to that, its whole
frame account, as Biardeau says, is a “maladept” “patchwork of more or less
ancient reminiscences” (2002, 1: 93–94), mixing Vedic and Upaniṣadic allu-
sions into an epico-purāṇic sequencing of a primary creation from the Elements
and a secondary creation.55 Moreover, early on, one may see an outcropping of

55. Cf. Bühler’s finding ([1886] 1969, lxxxi–lxxxc) that twenty-three and a half verses from Manu’s cos-
mogony scattered between M 1.18–86 are similar to and in some cases identical with thirty-eight consecutive
 dharma over time, i 267

familiarity with epic bhakti in a verse explaining the name Nārāyaṇa (M 1.10).
Or, as Biardeau, sees it, the verse may reflect a Vedic sense of Nārāyaṇa:

It is in appearance asserted very logically in a place where the waters are

mentioned as appearing, since this is a constant referent for the verse’s
connection with the deluge. The incongruity comes with the name
being one of Viṣnu ̣ when he takes the form of Brahmā to create the
world around the sacrifice in the second creation. But Manu doesn’t
know Viṣnu ̣ , and Brahmā suffices for his Supreme Puruṣa sleeping on
the water. This probably shows Manu’s proximity to the Vedic heritage,
where Nārāyaṇa is known without Viṣnu ̣ . (Biardeau 2002, 1: 96)

In any case, whereas the epic introduces what Biardeau calls a “swerve” in the
Brahmanical tradition toward bhakti, Manu “resists” this swerve “by which the
Brahmanical tradition goes on to renew itself” (2002, 1, 96). Yet Manu seems
to regularize what the epic has to say about dharma over time with regard to
yugas and possibly also about Manvantaras.
As one might expect, where Manu begins organizing its treatment of time
itself, it does so with a chronology of Manus, prioritizing an account of six
Manus in the lineage of Manu Svāyambhuva, and mentioning their Manvantaras
or “Manu Intervals” (M 1.61–63). Manvantaras, which I will say more about
when Manu returns to them, are thus introduced even before units of time are
detailed from a “moment” (nimeṣa) to the two halves of the year, which are
equated to a day and night of the gods (64–67). From the year, a day and night
of Brahmā is calculated in yugas: a four-yuga (caturyugam) sequence constitutes
a yuga or age of the gods (devānāṃ yugam) lasting twelve thousand years, and a
thousand such divine yugas “should be regarded as a single day of Brahmā, and
his night as having the same duration” (68–72).
After this come three intriguing verses on the high value of this knowledge:

Those who know this propitious day of Brahmā having 1,000 Ages
[= divine yugas], as also his night with the same duration—they are the
people who truly know day and night (rātriṃ ca tāvatῑm eva te
̣ At the end of that day-and-night, he awakens from
’horātravido janāh).
his sleep; and when he has woken up, he brings forth the mind, which
is both existent and non-existent (pratibuddhaś ca sṛjati manaḥ
sadasadātmakam). The mind, driven by the desire to create, transmutes

Mbh verses (12.224.11–48 in the Pune CE; 12.232.11–233.17 in the Bombay ed. used by Bühler). The parallels
include a few details from the first cosmogonic stage such as the creation of the great elements (M 1.18ab = Mbh
12.224.43cd), and more from the second, for example, on dharma and the yugas (M 1.81–86 and Mbh
12.224.23–27).
268 dharma

the creation. From the mind is born ether, whose distinctive quality is
said to be sound. (1.73–75; Olivelle trans. 2005a, 90)

Who are these “people’’ ( janāh)̣ who really know? And what do they know?
One thing they know is that the sum of time is counted in terms of Days and
Nights of Brahmā, and not in terms of kalpas. As Biardeau (2002, 1: 94) and
González-Reimann (2008) have noticed, Manu makes no mention of the kalpa
here as an equivalent term for the Day of Brahmā. Yet Manu knows the term,
using it, though only once for a large or at least ancient unit of time, in a
passage noted in chapter 5:

In a former aeon (purā kalpe) gambling was seen to create great

enmity; therefore, an intelligent man should never engage in gam-
bling even for fun. (M 9.227)

Clearly, Manu knows something of the Mahābhārata (see chapter 5 § F).

Manu here is either using kalpa as interchangeable with yuga, which, as we
have already seen, the Mahābhārata also does; or, more likely, it uses kalpa
because its own chronological setting requires a temporal relocation. Since
Manu should be delivering his dharmaśāstra before the Mahābhārata occurs,
it should refer to a Mahābhārata event either in the future, as prophesy, or in
a deeper past than could be suggested by the term yuga. Manu could thus be
implying that what happens in the epic could also have happened at a similar
yuga juncture in another long-ago aeon. This could have a certain coherence
if we remember the primordial standpoint from which Manu speaks as a law-
giver for all times. Not in the habit of prophesy, he has chosen precedent.
In any case, Manu has not yet connected the yugas with dharma. This may be
because their significance to this point is only transitional in accounting for time
from its smallest to largest units. But more likely, it resonates with what “the
people” who know Days and Nights really know: that what belongs to the nature
of mind and is both existent and nonexistent and driven to create by desire is a
cosmogonic impulse that transcends or ontologically precedes dharma. All this
probably fuses a proto-Sāṃkhya cosmogony of the five great Elements from Ether
down to Earth with patchworked Vedic allusions to ṚV 10.129, and to the open-
ing cosmogonies of the Bṛhadāraṇyaka Upaniṣad, including one discussed in
chapter 3 where brahman was alone and “had not fully developed” until it created
dharma (BᾹU 1.4.11–14). Something very un-Buddhist is thus offered in affirm-
ing brahman’s (or Brahmā’s) thinking about creation.
Once the account of primal creation proceeds from mind through the
great Elements (76–78), Manvantaras are calibrated into the day of Brahmā
(79–80). Like creations (sargas) and destructions (saṃhāras), Manvantaras are
 dharma over time, i 269

said by Manu to be “incalculable” (asaṃkhyāni), and all three fall under what
“the Supreme Lord does again and again as a kind of sport.”56 Creations and
destructions imply an equivalence with Days and Nights of Brahmā, which
would be further equivalent to kalpas. We thus find Manu once again not
using the term kalpa where it could, yet evoking a notion of “divine play” that
seems more appropriate to epic descriptions of Viṣṇu or Kṛṣṇa than to
Brahmā. Moreover, creations, destructions, and Manvantaras are given a
descriptor “incalculable” that may echo the four incalculable aeons of
Buddhism, while in very un-Buddhist fashion, they cycle through time by
divine agency. While resisting the epic’s bhakti swerve, Manu thus makes two
moves: Days and Nights of Brahmā are not called kalpas, and Manvantaras
are not explicitly linked with dharma. Given that the text is a “Treatise of the
Laws of Manu,” this might be surprising. But it seems to imply that Manus
teach dharma for all time, and do no more than anticipate its fluctuations
over time.57
Finally, then, dharma is introduced into the flow of time, and solely in
connection with yugas:

In the Kṛta Age, dharma is whole, possessing all four feet; and so is
truth. People never acquire any property through unlawful means.
By acquiring such property, however, dharma is stripped of one foot
in each of the subsequent Ages; through theft, falsehood, and fraud,
dharma disappears a foot at a time. In the Kṛta age, people are free
from sickness, succeed in all their pursuits, and have a life span of
400 years. In the Tretā and each of the subsequent Ages, however,
their life is shortened by a quarter. The life span of mortals given in
the Veda, the benefits of rites, and the power of embodied beings—
they all come to fruition in the world in conformity with each age.
There is one set of dharmas for men in the Kṛta age, another in the
Tretā, still another in the Dvāpara, and a different set in the Kali, in
keeping with the progressive shortening taking place in each Age.
Ascetic toil (tapas), they say, is supreme in the Kṛta Age; knowledge
( jñāna) in Tretā; sacrifice (yajña) in Dvāpara; and gift-giving (dāna)
alone in Kali. (M 1.81–86; Olivelle trans. 2005a, 91, slightly
modified)

56. M 1.80: manvantarāṇyasaṃkhyāni sargaḥ saṃhāra eva ca/ krῑḍannivaitatkurute parameṣṭῑ punaḥ punaḥ.
57. See González-Reimann 2002, 173 on the way Dharmaśāstra commentators “were troubled . . . by the
suggestion that dharma could vary and not be immutable,” mentioning Medhātithi’s view in commenting on
270 dharma

Manu’s last two verses, like some passages in the Mahābhārata,58 thus assert
a set of correlations between the yugas and certain typifying virtues.59
Indeed, here we may reinforce our suggestion that Manu could be con-
densing what the Mahābhārata has to say about dharma over time while resist-
ing its swerve toward bhakti. Manu does not say that Viṣṇu or Kṛṣṇa has
anything to do with the course of dharma over time. In particular, Manu does
not report that Nārāyaṇa changes colors from one yuga to the next, which one
finds reported twice, with the colors for the second and third yugas reversed, in
the Mahābhārata.60 But for the rest, although it does not indulge in anything
like the Mahābhārata’s vast meditation on Time, with its philosophical and
theological implications, it does distill most of what the epic says about basic
temporal units. The Mahābhārata makes brief mention of the Manvantaras in
the Nārāyaṇῑya. And as regards the kalpa, as Biardeau (2002, 1: 94–95) and
González-Reimann (2008, 2–5) show, neither the Mahābhārata nor Manu uses
the term consistently, especially as distinct from the yuga; and neither uses it in
the later Purāṇic fashion as equivalent to a Day of Brahmā—not to mention a
Life of Brahmā. As with Manu’s mention of disastrous dicing in a former kalpa,
the Mahābhārata appears to take advantage of fluctuation in the usages of kalpa
and yuga, which González-Reimann calls “confusion” (2008, 72; 85 n. 150).
But most important, although the Mahābhārata does it so much more col-
orfully and recurrently, both Manu and the epic detail the decline of dharma
through yugas.61 As we saw in chapter 5, Manu begins with its “all-encompassing
vast cosmological setting” in order to treat varṇa and āśrama in a context of
“past, present and future.” It thereby gives explanatory power to dharma over
time while including Greeks, for instance, in its discussion of mixed castes—
and with this, the reinvention of the king as a start-up Kṣatriya who could, “in
reality,” be of any caste. On such matters, Manu would be describing much the
same society as the Mahābhārata, but with an orientation not so much to its

Manu that “dharma itself did not change as the yugas advanced,” but rather “the decreasing capacity of men in
each yuga” in being “able to follow it correctly.”
58. See González-Reimann 2002, 166, 188 n. 13: Mbh 12.224.26–27, closest to Manu’s two verses, lists the
yugadharmas as tapas, jñāna yajña, and dāna; 12.252.8 repeats 12.224.26, preceding it by, “We hear that the teachings
of the Vedas wane according to the yuga”; and at 12.336.28, the Nārāyaṇῑya has Brahmā “create the yugadharmas.”
59. The Purāṇas make such correlations regularly, with variations and elaborations, including sectarian
passages on the good fortune of being born in the Kali yuga (see González-Reimann 2002, 174–78; von
Stietencron 2005, 43), and long lists of practices prohibited in the Kali yuga called kalivarjyas (Lingat 1973,
189–95; González-Reimann 2002, 173).
60. See González-Reimann 2002, 103, 114 n. 61, citing Mbh 3.148.16–33 (Hanumān’s version: white, red,
yellow, black) and Mbh 3.187.31 (Mārkaṇdẹ ya’s version: white, yellow, red, black). The important point, despite
González-Reimann’s wish to say otherwise (105), is that, along with Kṛsṇ ạ , black is being linked with the Kali yuga.
61. See Mbh 12.224.23–27, and what Hanumān has to say on the dharma of the yugas (3.148.9–36, cited
chapter 5 n. 20).
 dharma over time, i 271

past, at the end of a previous yuga62 where, thanks to Kṛṣṇa, it had divine aid
during the transition to this yuga, but to its maintenance and renewal in times
of current and future change. As we have seen, Manu 9.301–2 agrees with, and
probably restates, the Mahābhārata doctrine that the king is, or makes, the
yuga. Indeed, along with the color associations of the yugas with Nārāyaṇa
mentioned above, one of the Mahābhārata’s few other basic yuga ideas to have
a Purāṇic future without Manu replicating it is that the four yugas succeed each
other in the land of Bhārata (Mbh 6.11.3; see González-Reimann 2002, 117 n.
80; 211–14)—which the Purāṇas make something unique to Bhārata.63
Here, however, it becomes necessary to distinguish my approach from
González-Reimann’s, since it will bear on many matters in my discussion of the
Yuga Purānạ in the next chapter. With his view that the Mahābhārata accreted
over many centuries, González-Reimann (2002, 2006b) makes his project the
working out of a long developmental logic of the yuga concept, from “mere met-
aphor” and “simple story” to systematic complexities; but he takes a more
restricted, if still vague, view of “Vaiṣnạ vism,” suggesting that the beginnings of
its possible impact upon the epic would have begun “[a]round the beginning of
the common era” (2006b, 227). As should be clear, my views on these matters
are different. I do not believe the epic accreted over time. And I do not think it
is convincing to stretch out the epic’s treatment of the kalpa and yuga over a long
hypothetical history in order to trace this kind of developmental logic—always
with the assumption that “simple” usages are early and complex ones are “late.”
I believe the epic poets are contributing to the construction of “Vaiṣnạ vism,” not
that some form (or forms) of “Vaiṣnạ vism” impacts the epic only after
“Vaiṣnạ vism” has become a theological movement. The epic’s yuga and kalpa
chronometries are part of that construction and not late additions by a sectarian
system. With these considerations in mind, I would like to move forward to the
Yuga Purānạ by stating two general working premises:

1. Mahābhārata passages that work out the complexities that relate

dharma to time, including passages where this complexity involves

62. See chapter 5 § F at n. 147. Actually, beyond Manu’s agreement that the king is or makes the age
(9.301), what the next verse goes on to say looks like a restatement of what Manu says about the king’s vigilant
day (see chapter 5 § G): “When he is asleep, he is Kali; when he is awake, he is Dvāpara; when he is ready to
undertake operations, he is Tretā; when he is on the march, he is Kṛta” (9.301). For discussion of the more com-
plex and diverse Mahābhārata doctrine, see Biardeau 1994, 47–60; 2002, 1: 609, 635, 1006–8, trying, to my
mind unconvincingly, to link Duryodhana with the Kali yuga in a way that requires that Yudhiṣṭhira inaugurate
a Kṛta yuga; Couture 2006, 75–76, taking the early yugas to represent pre-urban nostalgias; Thomas 2007,
arguing that the epic means to have the king’s making of the age taken literally rather than, as González-Reimann
2002 would have it, as “mere metaphor” (see chapter 7)
63. See Kloetzli 2010, 600, 607–8 on yugas only in Bhārata in the Viṣṇu Purāṇa.
272 dharma

Viṣṇu or Kṛṣṇa, will not be deemed axiomaticallty “late.” These include

especially the Bhagavad Gītā and the Nārāyaṇῑya, and also passages
from the Anuśāsana Parvan (Book 13) and the Forest Book (Book 3).64
For instance, one of González-Reimann’s arguments is that, in a
passage cited above, the Nārāyaṇῑya uses the term kalpa in a way that
suggests the “possibility of a competition with Buddhism,” making
Brahmā subordinate to Viṣṇu as a real creator rather than to the
Buddha as a deluded one. This may be so without its being a
“late” addition in the epic.
2. It is always worth considering that the epic, like Manu, may be using the
terms kalpa and yuga interchangeably not as a result of developmental
“confusion” but strategically to show the interwovenness of these units
over the vastness of time, and possibly as a deconstructive move over
against Buddhist usage, where the term kalpa would stand alone.

64. González-Reimann 2008 remarks that 50 percent of the Mahābhārata’s thirty-six usages of kalpa for
a long time period occur in the Nārāyaṇῑya and nine in the Anuśāsana.
7
Dharma over Time, II
Prophesies of Disaster

This chapter compares Brahmanical and Buddhist prophesies from

virtually the same period in “minor” (as I called them in chapter 1)
classical dharma texts. On the Brahmanical side, we meet a short text
called the Yuga Purāṇa [henceforth YP] embedded in an astronomical
treatise that foretells what will befall dharma toward the end of the
Kali age. On the Buddhist side, we meet a series of Sūtra texts not
found in the Pāli canon, but only in Northern Nikāya schools and in
Mahāyāna variations, which tell of a debacle that will mark the very
end the dharma.
We can now say that these are minor dharma texts in that their
prophesies take place off-stage from the intertextual drama that we
have seen our main classical dharma texts scripting in chapters 4–6.
But they are illuminative of that drama in that they crystallized
Brahmanical and Buddhist understandings of what change entailed
for an imperiled dharma in history as each tradition was conceiving it.
We may thus continue to examine Bronkhorst’s central thesis that
“Greater Magadha” produced a “separate culture” (2007, 1–9), but
with our attention now on the yuga rather than the kalpa. With this
new focus, we will be able to ask about the ways other cultural regions
responded to Greater Magadha’s rise to dominance and fall into
decline, including—if the sources allow it—the ways writers reflecting
the cultures of other regions envisioned, indeed, constructed, their
own alternate histories and different theories of time through which
to frame their chronologies. In these minor dharma texts,
274 dharma

both traditions expound on distinctive concepts of decline against the

background of the same historical turmoil: the incursions of Greeks and Śakas
in the YP; and of Greeks, Śakas, and Pahlavas in the earliest of the Buddhist
texts. The nonmention of Pahlavas in the YP would suggest that it was the
earlier of the two prophesies, and with that in mind, I will treat it first. Although
that evidence is not very definitive, we will find other reasons to suspect that the
Yuga Purāṇa is earlier than the comparable Buddhist prophesies. But
comparison may begin from the premise that both traditions catalyzed their
prophesies against the same historical backdrop.

A. The Yuga Purāṇa

The YP, a short unit of only 115 verses, is introduced by John E. Mitchiner as
having two major interests. It is “the only Indian text” that “refers in any detail
to the presence of Indo-Greeks in India.” And it is “important for its account of
the four Yugas,” being, he thinks, “probably the earliest account of the Yugas
in Indian literature” (1986, vii).1 According to Mitchiner, in terms of its overall
structure, the YP’s “main aim” is “to give an outline account of the principal
peoples and events in each of the four Yugas or Ages, as an illustration of what
came to pass when dharma inevitably declined with the passing of time”
(1986, 50). It is “primarily a Brahmin-oriented text . . . concerned with the
proper maintenance of the four varṇas or castes” (ślokas 15–19).2 “It also decries
the rise of Śūdras, Vṛṣalas (low-born men) and Pāṣaṇḍas (heretics) at the end of
the Kali Yuga” who usurp Brahmin roles (śl 50–55). It opposes changing roles
of women (śl 82–86); is antiascetic, despising low-born Bhikṣukas, “a term
applicable to both Hindu and Buddhist mendicants”(śl 52);3 “and it also con-
demns men for abandoning an active role in favour of taking their ease as
gṛhasthas or house-dwellers” (1986, 47).4

1. This section overlaps with Hiltebeitel 2010b and 2011a, chapter 4. Whereas those essays discuss the
YP’s value in contextualizing the Mahābhārata’s primary genre identification as “history” (itihāsa), on which see
now chapter 1 n. 25, this section is shaped to make this chapter’s comparison with Buddhist narratives. It also
introduces comparison with prophesies by Vyāsa at the end of the Harivaṃśa.
2. Curiously, śl. 19 says that in the Tretā Yuga each class was intent upon its svakarma, but shifts the
terms for the Dvāpara Yuga, in which everyone did their svadharma unquestioningly (śl 27) until the
Mahābhārata war.
3. It also scorns dressing in red (raktavāsas) (Mitchiner 1986, 47, citing śl 86; see 55 n. 112). See Bhattacharya
2008.
4. The criticism of gṛhasthas strikes me as unusual. I am familiar with such a view only in some Buddhist
suttas. Could it air a critical view of wealthy estate-holding mahāsāla Brahmins in the Greater Magadha area?
 dharma over time, ii 275

This so-called “purānạ ” is found in an astronomical/astrological treatise

called the Gārgya-jyotiṣa. Mitchiner argues, I believe convincingly, that it is com-
posed as part of this treatise, although one might consider the possibility that it
was an added component. Mitchiner proposes that the YP was composed in
Ujjain—“itself well-famed as a centre of jyotiṣa-studies” or astronomy and astro-
logical sciences (79–80), probably in Brāhmī (31, cf. 36), around 60–25 BCE,
with the later date the most likely (81; cf. 5, 11, 16). Mitchiner’s main evidence is
that the YP mentions Yavana and Śaka incursions, which continued down to 60
BCE, but not the Pahlava incursions into northwest India that probably did not
occur until early in the first-century CE.5 This is a rather loose terminus. As we
shall see, there could be other reasons why the YP stops its north Indian history
without getting to the Pahlavas or Kuṣan ̄ ạ s. But the selective history it describes
all precedes them. I believe the YP may be a bit later than 25 BCE, perhaps by
half a century, and that it was probably composed during the period of Śaka con-
solidation between ca. 25 BCE6 and the rise of the Kuṣan ̄ ạ s. The Śakas established
themselves as Kṣatrapas (Satraps), forming trade networks and points of power
from the northwest to central north India, and supporting mainly Buddhism.7
The title Gārgya-jyotiṣa implies that the YP’s author would be named Garga
or Gārgya, a sage known in the Mahābhārata. In Mahābhārata Book 12, Garga
is said to have become the “keeper of the year, the almanac maker (sāṃvatsara)”
of the primal—and first favored—king [Pṛthu] Vainya.8 In Book 9, Old (Vṛddha)
Garga, after doing severe tapas at an auspicious tīrtha on the Sarasvatī River, is
said to have obtained the “knowledge of time and of the passing-away (or, evil
effects) of heavenly bodies (jyotiṣāṃ ca vyatikramaḥ),” and of favorable and
unfavorable omens (utpātā dāruṇāś caiva śubhāś ca; 9.36.14–17). The association
of the Sarasvatī with calendrical calculations is reminiscent of yātsattras: jour-
neys up the Sarasvatī to its source in conjunction with the winter solstice, and
then back down. Yātsattras, associated with Vrātyas, are first mentioned in the
Brāhmaṇas and are referenced with surprising frequency in the Mahābhārata. This
citation is a case in point. The listener here is Kṛsn ̣ ạ ’s brother Balarāma, who is

5. See Mitchiner 1986, 81, mentioning Indo-Scythian and Indo-Parthian settlements of ca. 25 CE; cf.
Nattier 1991, 225, citing Frye 1983, 197–204: “the best current estimate is that the Parthians—the last of the three
groups [after the Greeks and Śakas] to arrive on the scene—invaded northwest India sometime during the first
half of the first century CE.” Cf. Nattier 1991, 152 n. 17.
6. See Härtel 2007, 324: “Based on the archaeological data [at Sonkh, near Mathurā], the Mitra period
ends with Level 25 in about ±25 B.C.” Cf. 346.
7. See Neelis 2007, 72–79, making the point that Brahmin texts produced xenologies identifying the Śakas
as adharmic barbarians, even though “Śaka support of Buddhism did not preclude their support of other Indian
religious traditions” (79 n. 100); 2008b, 8–10, adding that “this support did not preclude the patronage of
Brahmins, Jains, and other non-Buddhists.”
8. Mbh 12.59.117cd; Fitzgerald trans. 2004a, 310. He is just called Vainya in this context. Cf. Mitchiner
1986, 42.
276 dharma

making just such a journey (Hiltebeitel 2001a, 120–26, 138–61). Along these
lines, in Book 13, the Anuśāsanaparvan, Gārgya reports that once when Śiva was
“pleased by my mental sacrifice, he gave [me] this great wondrous knowledge
of time contained in sixty-four divisions on the bank of the Sarasvatī, as well as
a thousand sons conversant with brahman like me, and a life-span for me and
my sons of a million years.”9 Note that Śiva extends these life spans without
anyone using the term kalpa. The mention of sixty-four aṅgas is also note-
worthy, since the Gārgya-jyotiṣa says of itself that it has sixty-four aṅgas,
although it includes only sixty-two (Mitchiner 1986, 10). The YP is the Gārgya-
jyotiṣa’s forty-first aṅga, coming in sequence after aṅga 39 on portents—
“Rāṣṭrotpātalakṣaṇa (Signs and Portents of Calamity),” with twelve categories of
signs and portents listed; and aṅga 40 on “Tulākośa (Weighing on a Balance).”
And it is followed by aṅga 42 on “Sarvabhūtaruta (The Cries of all Creatures)”;
aṅga 43 on “Vastracheda (Tears in Clothes)”; and aṅga 44 titled Bṛhaspatipurāṇa,
on Jupiter, one of the other two units called a “purāṇa” (Mitchiner 1986, 14–16,
108–10). This Anuśāsana Parvan passage could be later than the epic’s other
two, since it refers to the same numerology as the treatise, and possibly, as
Mitchiner maintains (1986, 10, 45), to the treatise itself. If the Anuśāsana
Parvan passage shows familiarity with a YP from the late first-century BCE
(Mitchiner’s proposed date), that would still be in time to date it within the time
span I proposed for the Mahābhārata’s composition. On the other hand, the
Gārgya-jyotiṣa’s claim of having sixty-four aṅgas when it has only sixty-two
could suggest that it was striving for a number it never reached on the basis of
Gargya’s epic reputation as an astrologer.10 In any case, I see no reason to
regard any of these passages to be late in the fashion that Mitchiner does in
assigning them to the third-century CE (1986, 5).
Strikingly, in being part of an astrological text, the YP calculates the kalpa
idiosyncratically.11 Yet it innovates in treating the yugas, and thus dharma, from
the standpoint of the predictive sciences. It will be my argument that it does
this building on the Mahābhārata. As Mitchiner recognizes, the events by

9. Mbh 13.18.25–26: catuhṣasṭyaṅgam adadāt kālajñānaṃ mahādbhutam/ sarasvatyās taṭe tusṭo manoya-
jñena pāṇḍava// tulyaṃ mama sahasraṃ tu sutānāṃ brahmavādinām/ āyuś caiva saputrasya saṃvatsaraśatāyutam//
Cf. Mitchiner’s trans. 1986, 102. More fully, see Mitchiner 1986, 5, 7, 10–11, 16, 79, 101–3.
10. As a cosmological and astronomical number, 64 occurs in Archimedes’ “Sand-Reckoner” as the
number 1063 or “one followed by 63 ciphers” by which one counts the grains of sand in the universe (Kloetzli
1983, 16, 115–22).
11. Along with a 10,000,000-year kalpa, and without defining how long each yuga lasts (González-
Reimann 2002, 98), the YP gives the yugas a decimally defined relation in terms of life spans of 100,000 years
in the Kṛta, 10,000 in the Tretā, 1,000 in the Dvāpara, “and (by implication) 100 years in the Kali” (Mitchiner
2002, 43; śls 8, 21, 24, 115). Rather than the epic’s and Manu’s 4–3–2–1 proportion of yuga-durations, the YP may
befit the Gārgya-jyotiṣa’s interest in astronomical calculations. See further Hiltebetel 2011a, chapter 4.
 dharma over time, ii 277

which the YP characterizes the end of the first three yugas are also referred to
in the Mahābhārata,12 though the epic does not refer to the first episode—the
destruction of the demon Tāraka that ends the Kṛta Yuga—in connection with
that yuga. But for the other two, Rāma Jāmadagnya’s destruction of the Kṣatriyas
thrice seven times at the transition from the Tretā to the Dvāpara Yuga, and the
Mahābhārata war with the transition from the Dvāpara to the Kali Yuga, the
Mahābhārata does make the correlations—albeit inconsistently.13 It is important
to register that the Mahābhārata refers to these yuga-junctures while introducing
itself generically as itihāsa, “history,” doing so at its beginning in both cases: for
Rāma Jāmadagnya at the Tretā-to-Dvāpara juncture (Mbh 1.2.3) and for the
Mahābhārata war at the Dvāpara-to-Kali juncture (1.2.9). What is striking is
that the YP shifts from past to future tense right with the Mahābhārata war
(Mitchiner 1986, 35, 50), making that war and the end of the Dvāpara yuga a
prophesy no less than the whole course of the Kali yuga that follows it. Keeping
in mind that Mitchiner brings up these matters under what he calls the YP’s
“main aim,”14 I will now begin to talk about the telos of that text. According to
Mitchiner, the YP is “selective” in two ways.
First, it does not give complete genealogies: that is, it is not focused on an
epic’s dynasty. Mitchiner thinks this would reflect a “fairly early stage in the
formulation of the Epic and Purāṇic genealogical tradition, which was only
subsequently developed into the presentation of complete genealogical lists.
The Yuga-Purāṇa has clearly not derived its account . . . from any other extant
literary sources or from the main Epic and Purāṇic tradition” (1986, 50). Yes
and no. As Mitchiner wants to maintain here, the YP shows unique and original
features in the events and personages it mentions. But his idea that it would
reflect an “early stage” of the epico-purāṇic genealogical tradition is untenable.
It simply lacks a genealogical telos. More than this, von Stietencron shows that
the genealogies introduced in the royal genealogical section of the Purāṇas, the
vaṃśānucarita, were probably first redacted in the period of the Nandas, with
the section on future dynasties, called the Bhaviṣya(t) Purāṇa, then added
during the early Mauryan period (2005, 69–85). Such lists of kings (von
Stietencron calls them “complete genealogical lists,” but that is more doubtful)
would thus be earlier than both the YP and the epics.

12. He says “both in the Epics . . . and throughout the Purānạ s” (1986, 50), implying that these texts would
develop their accounts after the YP. I will, however, be arguing that the YP cannot be earlier than the Mahābhārata.
As to the Rāmāyaṇa, Mitchiner says, interestingly, that the YP ignores “the entire Ramayaṇa tradition” (1986, 45).
13. The Nārāyaṇῑya makes a different correlation for Rāma. When Nārāyaṇa describes his Prādurbhāvas,
he says, “When the twilight of Tretā and Dvāpara has arrived, having become Rāma Dāśarathi, I will become the
lord of the universe (bhaviṣyāmi jagatpati)” (12.326.78). I would attribute the YP’s ignoring of the Rāmāyaṇa (see
previous note) to a lack of interest in its geography and its being kāvya rather than “history.”
14. One would like to know how and whether this aim is related to the aim of the Gārgya-jyotiṣa.
278 dharma

The second area of selectivity gets us to some of the ways the YP shows
unique and original features, including the futurity of the Mahābhārata.
Says Mitchiner, it is neither “a complete” nor “fully consecutive catalogue of
events and characters, but rather presents . . . certain select and prominent
occurrences: it is therefore impossible to reconstruct, from this account
alone, anything more than an outline chronology for the people and events
described” (1986, 50–51). Mitchiner is speaking here mainly about what the
YP presents as late “people and events” of the Kali Yuga, which I will get to
shortly. The only outline chronology it is really interested in, however, is that
of the four yugas, and mainly, beginning with the Mahābhārata, in the future,
as prophesy.

A.1. The Yuga Purāṇa and Yugas in the Mahābhārata

Now Mitchiner maintains that the YP would be “probably the earliest account
of the Yugas in Indian literature” (1986, vii; cf. 35), but on this point his expla-
nation is unconvincing. For Mitchiner, the YP’s “phrasing” of “its accounts of
the Mahābhārata war and the reign of Janamejaya Parīkṣit also in the future
tense . . . suggests” that its “account was composed before such a convention of
the precise dividing-point between past and future time became widely
established” (35). Dividing lines between past and future, however, are more
flexible than the one between the Dvāpara and Kali Yuga, on which the YP
simply follows the Mahābhārata. Indeed, so does the vaṃśānucarita, once the
Bhaviṣya(t) Purāṇa begins the future during the reign of the Pāṇḍava descen-
dant Adhisīmakṛṣṇa (see below). The Mahābhārata probably establishes the
flexibility of this convention when Mārkaṇḍeya shifts to the future tense to
describe the end of the Kali Yuga.15 Moreover, Mitchiner simply overlooks that
YP verses 28–39 are a kind of futuristic Mahābhārata digest, and clearly show
that some kind of Mahābhārata—almost certainly one with yugas—is older
than the YP:16

28. And at the end of that Yuga, the earth will go to destruction; men,
having come under the control of Time, will cook [for] their own
bodies (tasyāpi ca yugasyāṃte medavī̄ kṣayameṣyati/ narāspacam. ti
svāndehān kālasya vaśamāgatāḥ).

15. Mbh 3.188.10 ff. See Hiltebeitel 2005b, 125 n. 41.

16. Cf. González-Reimann 2002, 142–43, maintaining that “a text like the Yuga Purānạ ” could have worked
out the Mahābhārata’s yuga chronology before the epic did so itself, and that the epic “only later incorporated” it.
Yet he acknowledges that the “exact relationship” between the YP and the epic is “difficult to ascertain.”
 dharma over time, ii 279

29. Keśava (Viṣṇu) will arise at the end of the Dvāpara, in order
to destroy horses and elephants, princes, and men (hayānām ca
gajānaṃ ca pārthivānāṃ nṛṇāṃ tathā/ vadhārtham dvāparasyāṃte
samutpatsyati keśavaḥ).
30. [he will be] four-armed, of great valour, bearing the conch,
disc, and mace: [and he will be] called Vāsudeva, the strong one,
dressed in yellow clothes (caturbāhurmahavī̄ryaḥ śaṃkhacakra-
gadādharaḥ/vāsudeva iti khyātaḥ pī̄tāṃbaradharo balī).
31. Then, resembling Kailāsa, wearing a garland of flowers [and]
bearing the plough as weapon, there will arise Yudhiṣṭhira17—the
excellent king of the Pāṇḍavas—for the purpose of slaughter at the
end of the Dvāpara, together with [his] four brothers (tataḥ
kailāsasaṃkāśo vanamali halāyudhaḥ/ pāṇḍavānāṃ varo rājā bhaviṣyati
yudhiṣṭhiraḥ/ vadhārthaṃ dvāparasyāṃte caturbhir bhrātṛbhir saha).
32. [namely] both Bhīmasena the son of Vāyu, and Phalguna
of severe tapas, and the two brothers, Nakula and Sahadeva, born
of the Aśvins (vāyavyo bhīmasenaśca phālgunaśca mahātapāḥ/ nakulaḥ
sahadevaśca bhrātarāv-aśvinātmajau).
33. Also Bhı̄ṣma, Droṇa and others, and the prince
Dhṛṣṭadyumna and Karṇa, the king of Aṅga, together with
Aśvatthāman the invincible (bhīṣmadroṇādayaścaiva dhṛṣṭadyumnaśca
pārthivaḥ/ aṅgarājastathā karṇaḥ sāśvatthāmā ca durjayaḥ).
34. Devaka and Śatadhanvan, and Dāruka the illustrious—they
will arise at the end of the Yuga, in order to protect the world of men
(devakaḥ śatadhanvā ca dārukasya mahāyaśaḥ/ rakṣārthaṃ naralokasya
utpatsyaṃti yugakṣaye).
35. So too Śakuni and Dantavaktra, and Śiśupāla the haughty:
together with Śalya, Rukmi, Jarāsaṃdha, Kṛtavarman [and] Jayadratha
(śakunir-daṃtavaktraśca śiśupālaśca garvitaḥ/ śalyo rukmir-jarāsaṃdhaḥ
kṛtavarmā jayadrathaḥ).
36. The cause [of strife] of these might[y] kings will be Kṛṣṇā,
the daughter of Drupada: [and] the earth will go18 to her destruction
(eteṣāmapi vīrāṇāṃ rājñāṃ heturbhaviṣyati/ drupadasya sutā kṛṣṇā
dehāṃtaragatā mahī).
37. Then, when the destruction of men has occurred and the
circle of kings has ended, there will be the fourth [and] final Yuga

17. As Mitchiner 1986, 46, 90 n. 19 indicates, Yudhisṭhira’s description seems to be borrowed from
Balarāma.
18. Mitchiner 1986, 90 n. 20: lit “went.”
280 dharma

called Kali (tato narakṣaye vṛtte praśāṃte nṛpamaṇḍale/ bhaviṣyati

kalirnāma caturthaṃ paścimaṃ yugaṃ).
38. Then at the start of the Kali Yuga, Janamejaya Pārīkṣit will be
born, illustrious and celebrated on the earth—there is no doubt
(tataḥ kaliyugasyādau pārīkṣij-janamejayaḥ/ pṛthivyāṃ prathitaḥ
śrīmānutpatsyati na saṃśayaḥ).
39. And that king will cause a quarrel with the twice-born:19 his
anger for his insulted wife having come under the power of time
(so ‘pi rājā dvijaiḥ sārdhaṃ virodham upadhāsyati/ dāraviprakṛtāmarṣaḥ
kālasya vaśamāgataḥ). (Mitchiner 1986, 90, xi–xv text)

It will be noted that the YP describes a somewhat rough-edged Mahābhārata

that may reflect popular or “purāṇic” conceptions of it. For instance, the fault-
ing of the war to Draupadī-Kṛṣṇā in śloka 36 could remind one of the way the
Bhaviṣya Purāṇa—perhaps as recently as the nineteenth century—reworked
the medieval Hindi oral epic Ālhā. Reflecting north Indian Rājpūt perceptions,
the Ᾱlhā makes Draupadī the root cause of its “Mahābhārata of the Kali Yuga”
since she took rebirth as its heroine Belā so that she could more fully satisfy her
lust for blood (Hiltebeitel 1999a, 432–33, 494–511). The possibility that proto-
Rājpūt sensibilities might lie behind the YP will be worth holding in mind. Yet
more immediately, ślokas 31, 34–35 and 39 suggest that the YP author knows
something of the Mahābhārata’s “Appendix,” the Harivaṃśa. The description
of Yudhiṣṭhira with traits of Balarāma in verse 31 (see n. 17) suggests a height-
ened familiarity with Balarāma that could derive from that text. The names
cited in verses 34–35 to describe those who arise at the end of the Yuga
concentrate on enemies more of Kṛṣṇa than of the Pāṇḍavas, and Dantavaktra
and Śatadhanvan in particular are far more prominent in the Harivaṃśa than
they are in the Mahābhārata.20 One might think of relating these characteristics

19. See Mitchiner 1986, 51–52 on Janamejaya (śl 37–39): the YP could know a Mahābhārata story of his
killing a Brahmin and regaining favor after an Aśvamedha. See Fitzgerald 2004a, 786 n. to Mbh 12.146.3, joining
attempts to explain away Yudhisṭhira’s hearing from Bhīṣma about their as-yet-unborn descendents Parikṣit and
Janamejaya (called “Bhārata” [147.20d]). Fitzgerald thinks it is “more likely . . . the story originated apart from the
MBh . . . and was inserted . . . relatively late. The entire [Āpaddharma] seems to be significantly later than at least the
first thirty-five chapters of . . . the rājadharma.” Cf. Bowles 2007, 316–18, favoring “narrative recursiveness” here. In
any case, the YP is not referring to this Mbh story but one in the HV, to be discussed below and in chapter 12.
20. The Mahābhārata mentions only Dantavaktra, as an incarnate Krodhavaśa demon (1.61.57–61) and
ally of Jarāsaṃdha (2.13.12).On Dantavaktra in the Harivaṃśa, see HV 80–82 (Dantavaktra, Śatadhanvan,
Śiśupāla, and many others including Duryodhana join Jarāsaṃdha’s attack on Mathurā); 87.26–61 (he
goes with Jarāsaṃdha, Śiśupāla, and Pauṇḍra Vāsudeva to Vidarbha for Rukmiṇī’s wedding with
Śiśupāla, and when Kṛṣṇa abducts the bride, he is among those who attack him); on Śatadhanvan king of the
Kārūṣas, see also HV 18.15; 29.2–19; 98.5. Purāṇic accounts then make Dantavaktra a brother of Ṥiśupāla. In the
Bhāgavata Purāṇa, his slaying by Kṛṣṇa forms a pair with that of Ṥiśupāla. Upon each of their deaths, a ray of light
departs from them and enters Kṛṣṇa (BhāgPur 10.74.45; 78.10). He and Ṥiśupāla thus join the brother-pairs
 dharma over time, ii 281

to what Mitchiner calls the “hybrid” features of the YP’s Sanskrit (1986, 20–36),
but verse 39 also strengthens the case that the YP is in particular referencing
the Harivaṃśa text. The allusion in this last verse to a story about Janamejaya’s
anger over an insult to his wife certainly refers to a scene at the very end of the
Harivaṃśa. This distinctive story, which occurs in conjunction with a prophesy
that Vyāsa makes about near and distant events in the Kali yuga, tells that
Janamejaya blamed his priests for allowing an insult to his wife when she was
fulfilling her duties as queen to simulate sexual union with the slain horse
during Janamejaya’s Aśvamedha sacrifice, and Indra entered the horse’s body
so that he could have intercourse with her.21 In completing three verses that
zero in on the beginning of the Kali Yuga in the reign of the Pāṇḍavas’ descen-
dant Janamejaya, verse 39 gives the impression of rounding off the YP’s digest
of the “epic” by settling us into Kali age as the basis for turning immediately to
other later “future” events within it. I will return to this closing Harivaṃśa
sequence involving Janamejaya’s wife and Vyāsa’s prophesies in chapter 12 § A,
so for the moment let us just note two things. First, we will find in this chapter
that bits of the future history that Vyāsa foresees will help us to contextualize
some historical allusions in the YP. Second, if the YP dates from anywhere
near the late first century BCE, its Harivaṃśa references could be evidence of a
much earlier date for that text than is usually estimated. But leaving these
points for later discussion, our question for now is: why would the YP innovate
by placing the Mahābhārata war (and it’s Harivaṃśa aftermath) in the future?
All we need to know for a reasonable answer this question is that the YP is
part of an astrological text and that its frame story makes the whole unit a narrative
by Śiva to his son Skanda, the general of the Gods. The shift shows the power of
this text to predict the Mahābhārata astronomically, not its priority to the
Mahābhārata. Śiva, as it were, scoops Mārkaṇdẹ ya and Vyāsa (who shows a pro-
phetic side not only in the Harivaṃśa)22 to foretell the Mahābhārata as the entrée

Hiraṇyakaśipu and Hiraṇyākṣa and Rāvaṇa and Kumbhakarṇa in being incarnations of Viṣṇu’s doorkeepers Jaya
and Vijaya, who turned away a visitation of youthful-seeming Ṛṣis (led by Sanaka Ṛṣi) to Viṣṇu’s heavenly palace,
and upon being cursed to take on three lives each as pairs of demon brothers, chose to be reborn as great enemies
who would remember Viṣṇu when he slew them in his avatāras (BhāgPur 3.15.12–36; 7.1–2). HV 24.21–22 knows
Śiśupāla as an incarnation of Hiraṇyakaśipu and mentions Dantavaktra after him genealogically; but the HV does
not know the latter as an incarnation of Hiraṇyākṣa, or either of them as a Krodhavaśa demon.
21. Mitchiner (1986, 51) cites Harivaṃśa 118.11–39. On epic sensitivities concerning such Aśvamedha
scenes, see chapter 8 §§ E and G); Hiltebeitel 2011a, chapter 9; forthcoming-c. In the Mahābhārata, before
Janamejaya is instructed to perform an Aśvamedha, he is given the option of what amounts to an archaic yātsattra
by which he would traverse the Sarasvatī and Dṛṣadvatī Rivers (Mbh 12.148.10–13).
22. See chapter 6 § B on his obtaining knowledge of the three times at White Island, and Mbh 1.119.6–8,
his parting words to his mother, who will go off with the two sisters with whom, at her bidding, he sired the
Kauravas’ and Pāṇḍavas’ fathers: “The times of happiness are past and times of trouble lie ahead. The days grow
worse every new tomorrow, earth herself is aging. A dreadful time is at hand. . .” (van Buitenen 1973, 64).
282 dharma

to the Kali Yuga, which from its beginning thus becomes a somewhat redemptive
tale. Indeed, Śiva does not fail to begin this twelve-verse Mahābhārata-Harivaṃśa
digest by describing Kṛsṇ a in “Vaiṣnạ va” terms (Kṛsn ̣ ạ has four arms, three of his
conventional emblems, and the name Vāsudeva) as the guiding hand behind
23

the Mahābhārata war, thus ruling out any possibility of taking the YP to be assert-
ing a sectarian slant on the epic, and, more important, considering the likely early
date of the YP, demonstrating that Kṛsn ̣ ạ ’s redemptive role in the transition
between the yugas would be known to the YP from the Mahābhārata itself.
From here, the YP shows its teleological hand by moving directly from its
prophesy of the Mahābhārata (cum Harivaṃśa) to more recent persons and
events of the late Kali Yuga. After the twelve verses on the Mahābhārata and
King Janamejaya, the next personage, mentioned immediately, is the Magadha
king Udāyin, of the pre-Nanda Śiśunāgas, and his founding of the new Magadha
capital at Paṭaliputra (śl 40–43)!24 The telos of the YP will thus lie in Śiva’s
predictions connecting the Mahābhārata with certain persons, events, and places
of the late Kali Yuga.
Now as I have already mentioned, Mitchiner argues that the YP was prob-
ably composed not in Magadha but in Ujjain. I believe he is right to argue that
it features a northwestern/north-central Indian perspective rather than what
we can now call a Greater Magadhan northeastern one, which makes it inter-
esting to compare with Vyāsa’s prophesy at the end of the Harivaṃśa, which,
like the YP, is offered from a north-central perspective, and with the Greater
Magadhan outlook of the vamśānucarita and its Bhaviṣya(t) Purāṇa. Actually,
the YP prognosticates about three broad regions, which we may call Greater
Magadha, the Northern Midlands (for what I have just called the northwestern/
north-central area), and the Deccan or South—including the Kāverī River basin
(1986, 75–79 with map facing 78). As already indicated, the YP sketches out its
program highly selectively, mentioning events and personages found nowhere
else in the historical record: among them, its unusual interest among Indian
texts in giving details on the Indo-Greeks in India.25 I limit myself to two fea-
tured themes: (a) the way the YP links its singular account of the Indo-Greeks

23. See śl 30 above. Mitchiner observes the sectarian tone of these two verses on Kṛṣṇa, but correctly
observes that with its overall narration by Śiva, the YP “has no strong sectarian bias” (1986, 46–47).
24. The Bhaviṣya(t) Purāṇa traces Magadha expansionism to Udāyin. See Mitchiner 1986, 52–53: Udāyin,
who bears epithet śiśunāgātmaja, was perhaps the successor of Ajātaśatru, whom Mitchiner dates to ca. 460 BCE;
“for the author of the Yuga-Purāṇa,” the founding “constituted a major event in the Kali Yuga.”
25. Mitchiner 1986, 3, cited above, attempts to tie in these events with Greek presences in central and east-
ern India in inscriptions and with the history of Demetrios and Khāravela (55–58). Fitzgerald 2010 convincingly
dates Khāravela later than Mitchiner and an old consensus, promoting him to “the middle of the first c. BCE,” but
he makes a rather fantastic attempt to imagine this Jaina Cedi king “or a clansman” to have had a “possible role in
sponsoring the Mahābhārata” (109 and n. 14). Khāravela ruled from Kaliṅga (Orissa), on which the Mbh has many
negative things to say (see Salomon 1978; Biardeau 2002, 1: 298; 2: 54, 756–57; Hiltebeitel 2005b, 118–21).
 dharma over time, ii 283

with what I will be calling Puṣyamitra’s Northern Midlands Alliance; and (b) the
YP’s peculiar notion of safe havens in the south. Be it noted that the YP inter-
prets both these themes redemptively: after adharmic rule makes way for the
Greek and Śaka invasions, both incursions, says Śiva, will have redemptive
outcomes, which will include the invaders going back home.

A.2. The Greek Incursion and the Northern Midlands Alliance

The first incursion, then, is by the Greeks into Paṭaliputra. From Udāyin’s
founding of that Magadhan capital, the YP turns immediately to a late Mauryan
king there named Śāliśūka (Mitchiner dates his ascent to about 205 BCE), whose
adharmic rule (śl 44–46; 1986, 91) sets the stage. The YP disparages Śāliśūka as
“an oppressor of his own kingdom” (śl 45), and Mitchiner takes him to display the
non-Brahmanical preferences of the late Mauryas.26 As Mitchiner says, it is not
necessarily a continuous chronology from Śāliśūka to the Yavana incursion, but
since the account goes on to speak of Śuṅga kings, it would appear that the incur-
sion would not be much later than Śāliśūka. We now meet our Northern Midlands
Alliance of Greek, Mathurā, and Pañcāla forces, which pass through Sāketa
(Ayodhyā) on the way to Paṭaliputra, where they tear down the fort and leave the
lands desolate. After this, however, the Yavanas do not remain in “Madhyadeśa,”27
as war will break out in their own realm (1986, 55–58; śl 47–48, 56–57; 91–92).
Such a Yavana incursion in alliance with Pañcāla and Mathurā after the adharmic
rule of Śāliśūka is, according to Mitchiner, most conceivable “around 190 B.C.E.:
a period which saw . . . the secession of Sogdiana in the Indo-Greek realm, which
would doubtless have prompted the return of any Indo-Greek expeditionary force
to their own realm [in Bactria];28 and a period which also witnessed the final down-
fall of the Mauryas and the emergence of the new Śuṅga dynasty under Puṣyamitra
in India” in ca. 187 (1986, 58). Such a course of events may find corroboration in
the recent discovery that a Greek era was dated back to 186/85 BCE.29

26. Purāṇic sources list him as a successor of Aśoka who could have ascended “shortly before 200 BC.”
Cf. Thapar 1997, 183, 191, taking him from purāṇic sources to be the fourth-to-last Maurya, and of a bad reputa-
tion also in the Gārgasaṃhitā. The YP’s last word on him is that “he will, in delusion, cause his oldest brother to
establish a righteous [person] called Vijaya (vijayaṃ nāma dhārmikam.” (śl 46), which Mitchiner says may be a
negative-ironic twist on Aśoka’s dharmavijaya or “conquest through righteousness” (1986, 52–55, 91 and n. 26).
Proposing another interpretation linked with Jainism, see Bhattacharya 2008.
27. See Mitchiner 1986, 56: “ ‘Madhyadeśa’ seems to be intended here in the sense found also in Buddhist
works, denoting the region up to or beyond Kajaṅgala” (in southeast Bihar, “and even Puṇḍra.” Cf. 92 n. 33; Rhys
Davids 1904; Bronkhorst 2007, 1–4.
28. Mitchiner narrows this down from a starting period of ca. 205–190 during the overlapping reigns of
Euthydemos and Demetrios as co-regents in Bactria, and the span from ca. 190–171 when Demetrios became
supreme ruler.
29. I thank Jason Neelis for making this point (personal communication, 2008). On the inscriptional evi-
dence for this Indo-Greek era, see Salomon 2005, 2007, 268; Bracey 2005; Neelis 2007, 70 n. 63.
284 dharma

As we shall see in chapter 12, Vyāsa’s prophesy at the end of the Harivaṃśa
may allude to the overthrow of the last Maurya, Bṛhadratha, by his Brahmin
general Puṣyamitra Śuṅga,. The YP, as we shall see, seems to name Puṣyamitra
by another name, but without reference to these events and only as one of four
kings who rules at Paṭaliputra (Mitchiner 1986, 63). Bronkhorst associates the
YP’s account of the Greeks’ part in this invasion with the way some other
sources represent them as provoking fear, or as a threat to the order of
Brahmanical society (2007, 359). But note Śiva’s emphasis on their retreat. In
any case, Mitchiner shows a way to read these events as background to what
I am calling the text’s redemptive pattern. Says Mitchiner, “the confusion
wrought by the advent of a joint Yavana-Pañcāla-Māthura force may have
provided the opportunity for Puṣyamitra to overthrow the last Maurya king and
establish his own dynasty. . . . [T]o judge from [Puṣyamitra’s] name, it is quite
possible that he himself may have come from the then Maurya domains of
Pañcāla or Mathurā, where many of the subsequent “Mitra” kings are known
to have ruled” (1986, 58–59). Mitchiner goes on to enrich this hypothesis with
a plot:

Seen in this light, therefore, we may suppose that the Pañcālas and
Māthuras—seeking to hasten the fall of a rapidly weakening Maurya
empire and to establish their own independence [which, as shall be
mentioned, they achieved in their independent minting of coins]—
enlisted the help of a Yavana contingent and proceeded to the
Maurya capital to sound the death-knell of that dynasty. If the
“Mitra” Puṣyamitra was indeed of Pañcāla or Māthura origin, the
advent of this force—comprised of many of his compatriots—would
have provided him with an ideal opportunity to dispose of his Maurya
master and to have himself proclaimed as ruler. (Mitchiner 1986, 59;
my insertion)

Mitchiner goes on to propose that the seven kings of Sāketa, whom the YP
mentions next without names or details (śl 58–59), serve not only to take us
down to the conditions that presage the next invasion by the Śakas but might
include descendants of Puṣyamitra as rulers of Kosala, which bordered on
Pañcāla (1986, 59). Whatever we make of Mitchiner’s historical reconstruction,
we are left with the question of why the YP features a Yavana–Pañcāla–Mathurā
coalition in a destruction of the Magadha capital, to which Mitchiner’s solution
offers a cogent answer. Along the same lines, why, as was mentioned in chapter 5,
should Manu 7.193 prescribe that a king, who is given no particular location for
his capital, be urged to deploy soldiers “from the lands of the Kurus, Matsyas,
Pañcālas, and Śūrasenas” on his front lines when he goes to battle? With the
 dharma over time, ii 285

Śūrasenas being from Mathurā, it is much the same population. Manu’s

recommendation would likely reflect that these lands yield loyal soldiers from
the Midlands, as they are said to have done in the Mahābhārata war. And
indeed, before that war, and especially as prelude to it in the Harivaṃśa, it is a
Mathurā–Kuru coalition (displaced, in that Kṛṣṇa has left Mathurā for Dvārakā)
that devises the elimination of King Jarāsaṃdha of Magadha so that Yudhiṣṭhira
Pāṇḍava can undertake a Royal Consecration (Rājasūya) ceremony, as will be
discussed in chapter 13.
Now if the YP does not mention Puṣyamitra Śuṅga in connection with the
overthrow of the Mauryas, it probably does refer to him under the name of
Puṣpaka—identifying him as the cofounder and then sole ruler of the kings
called Agniveśyas (Mitchiner 1986, 3, 75, 93 śl 71–72)—that is, Śuṇga kings30
whom Mitchiner also calls the “the ‘Mitra’ kings” with reference to their coin-
age. Not known by either name from other sources, they are said to have ruled
prosperously at least for a while in Paṭaliputra (a.k.a. Puṣpapura), and also else-
where. Attention thus switches to a land called Bhadrayaka or Bhadrapaka,
otherwise unknown, where Agnimitra, possibly as a Śuṅga viceroy,31 and
Agniveśya, probably another Śuṅga, may have ruled. Agniveśya is said to have
fought the Savaras (today’s Saoras, a tribal population) there, which would sug-
gest the Vindhyas (Mitchiner 1986, 62–63).
Looking at coins found from the Śuṅga period, Mitchiner remarks on the
rise of dynasties across northern India that issued anonymous silver and copper
punch-marked coins first from Paṭaliputra, then briefly at Mathurā, and subse-
quently from mints in Vidiśā, Ujjain, and Eran. In these circumstances, the
central areas of Mathurā, Pañcāla, Kosala, and Kauśāmbī (capital of Vatsa) take
prominence, while “in western India and the Panjab, a number of small states
and tribes asserted their independence and issued coins,” including the Kurus,
Pūrus, Vṛṣṇis, and Yaudheyas. But there were also coins of “further ‘Mitra’
kings—including Indramitra” found at Paṭaliputra (1986, 63–64). With these
so-called Agniveśya or “‘Mitra’ kings,” we thus find a proliferation of “little
kingdoms,” reminding us how that ideal has flourished over the centuries in

30. Śl 61, 70c–81. Says Mitchiner (1986, 62), the “name is apparently given to both the dynasty as a whole
(śloka 62) and also to one of its prominent members, who is nonetheless not its founder (ślokas 79–80).” It is,
however, not clear that the name pertains to the whole line. The uncertainty of the chronology is compounded by
the mention before this (śl 60 and 65–70b) of a low-born ruler named Āmrāṭa, who takes over Paṭaliputra after
an uprising. Mitchiner suggests he may be Khāravela (1986, 60–62), on whom see Fitzgerald 2010 and n 25
above.
31. Agnimitra falls in love with a beauty of the land in Kālidāsa’s Mālavikāgnimitra, where he is Puṣyamitra’s
viceroy in Vidiśa, after being a coruler with him at the beginning of his reign (Mitchiner 1986, 63). Mitchiner
cites an astronomical feature of his description in the YP that may date his Vidiśā viceroyalty’s beginning to 183
BCE (66).
286 dharma

Indian martial culture, and interestingly—given the early names Agnimitra

and Agniveśya—flourished under the name and mythology of Agnikula or
Agnivaṃśa Kṣatriyas or Rājpūts, who are celebrated in the Hindi Ā lhā and sup-
pressed in the Bhaviṣya Purāṇa’s revision of that oral epic story. That is,
Agnikula and Agnivaṃśa are terms used for kings, some of them of Brahmin
extraction, who do not trace their descent in the Lunar or Solar dynasties made
normative in the Mahābhārata and Rāmāyaṇa, respectively (see Hiltebeitel
1999a, 211–363, 439–75; 1999b). This would reinforce our earlier point that
the YP could reflect early “proto-Rājpūt” sensibilities. One feature of the YP’s
redemptive history could thus be called the little kingdomization of Paṭaliputra
as preferable to rule there by the imperial Nandas or Mauryas. In the post-
Maurya period and before the rise of the Śaka Kṣatrapas (Satraps) and Kuṣāṇas,
the Mitras were little kings, the biggest being Puṣyamitra. Says Mitchiner, “We
might even go so far as to suggest that Puṣyamitra was instrumental in estab-
lishing certain viceroys or regents in such regions as Kosala, Pañcāla and
Mathurā who—or whose successors—thereafter asserted their own
independence.” The Śuṅgas quickly lost power in northern India,32 the focus of
power shifting to “Malwa, centered on the three cities of Vidiśā, Eran and
Ujjain”—where Mitchiner posits that the YP would have been written (65).
Meanwhile, in around 110 BCE and close to Vidiśā, something of the Śuṅga
period alliance with the Greeks could still have inspired the Indo-Greek ambas-
sador Heliodorus to inscribe his Vaiṣṇava sentiments on the famous Besnagar
Garuḍa Pillar in a śloka verse similar to one in the Mahābhārata (see Brockington
1998, 134; Witzel 2005, 62, 64, 66).

A.3. The Śaka Invasion and the Southern Safe Havens

It must be admitted that to call this amorcellization of kingship redemptive is to

play into a dead-end historiography. It may be better than life under the Nandas
or Mauryas, but it is going nowhere. In contrast to the YP’s redeptive interpreta-
tion of the foregoing history, Vyāsa’s prophesy at the end of the Harivaṃśa looks
at this same period more somberly: “When the age is declining (yuge kṣīṇe),
there will be great war, great tumult, grain rain, great fear” (HV 117.14a-c). In
these conditions, there will be migrations away from this north central heartland
to all regions but the deep south: “Assailed with fear and hunger, and carrying

32. See Härtel 2007, 346, finding it “amply clear that the Śuṅgas did not inherit the Mauryan empire in its
entirety; various small principalities had cropped up in the various parts of the Mauryan empire along with the
coup d’etat of Puṣyamitra or soon after it.” “The majority of historians agree on dating the beginning of the Local
States of Northern India to the second or later half of the second century B.C., taking for granted the disintegra-
tion of the Śuṅga empire soon after Puṣyamitra.”
 dharma over time, ii 287

their sons on their shoulders, men will cross the Kauśikī River and seek shelter
[east] among Aṅgas, Vaṅgas, Kaliṅgas, [northwest among] Kāśmīras, [in the
Vindhyas among] Mekalās, and among those of Ṛsị kāntagiri [uncertain]. Men
will dwell together with hosts of Mlecchas on whole flanks of the Himalaya, on
the shore of the saltwater ocean, and in forests” (117.28–30). In the YP, however,
the basic redemptive pattern repeats itself. Following an interlude on evil con-
duct of men and women, another bad ruler, King Satuvara (śl 87)—Mitchiner
thinks he is probably Śātavāhana or one of the early Śātavāhana kings (1986,
66–67)—rules for ten years to set the stage for the Śaka invasion, which is men-
tioned in the next verses (śl 88–89), though it is also mentioned earlier (śl 62–64).
This discontinuous chronology would take us down to about 60 BCE. According
to the YP, a Śaka king plunders and destroys until he is killed by a Kaliṅga king
and a group of Sabalas, which for Mitchiner again suggests a Vindhya location,
now close to Ujjain. Thereafter the Śakas, like the Greeks, return to their own city
(1986, 68).
Although the Purāṇas do not mention any Śaka incursion, the story is
related in the cycle that launches the Vikrama era, which was adopted by the
Guptas and Cālukyas.33 The YP’s redemptive predictions now draw to their
climax with a description of what Mitchiner calls “Regions in Which Men Will
Survive and Prosper” (1986, 75–76; śl 98–113). As he indicates, the area
in question lies “between the Vindhyas and the river Krishna (modern western
and central Madhya Pradesh, and Maharashtra), together with the southern
part of the eastern Ghats in Orissa, and the area around the river Kāverī in Tamil
Nadu” (75). As Mitchiner maps the locations (facing p. 78), he must set the
Kāverī area apart, but all the rest are contiguous, and he takes them to imply a
territory enjoying “renewed prosperity” under the Śātavāhana empire (75–76).
Noting the YP’s prognostic emphasis, Mitchiner takes it that its author, who he
says is “well acquainted on the one hand with various regions of northern
India—which are mentioned in the context of the Kali Yuga,” now mentions
these “survival regions for the start of a new (Kṛta) Yuga” (79): “clearly linking
his account of the Yugas to historical events,” this author “believes that, shortly
after the advent of the Śakas, the Kali Yuga had come to an end; and that after
an interval of harsh conditions, a new Kṛta Yuga was beginning to dawn—an
event which he evidently believed was occurring at his own time of writing” (82).34
Mitchiner suggests that others in Ujjain would have shared this view and

33. Mitchiner 1986, 71; cf. 74–75 offering a “summarized reconstruction” of the expulsion of the Śakas
from Ujjain, leading to the founding of the era. On the story cycle, see Brown 1933.
34. Cf. Mitchiner 1990, 321: After the Śakas’ “relatively brief plundering foray,” the YP “describes the end
of the Kali Yuga and the dawn of a new Kṛta Yuga.”
288 dharma

its “optimism” in around 25 BCE, and that this would “explain why the era of
58 B.C.—that was almost certainly founded by [the Indo-Scythian king] Azes35—
came to be called the Kṛta Yuga when it was subsequently adopted by the inhab-
itants of Ujjain and Malwa” (82).
This credulity of the people of Ujjain twenty-five years before the
common era might, however, strike us as rather contrived. The YP never
makes it explicit that its closing events, involving what Mitchiner calls the
“Regions in Which Men Will Survive and Prosper,” have anything to do with
the Kṛta yuga: a point acknowledged by González-Reimann, who nonethe-
less says that “the text implies that at that time, and in those places, a new
Kṛta Yuga will commence” (2002, 99).36 Mitchiner is also misleading when
he says men will “prosper” in these regions due to what he calls “renewed
prosperity” (75–76, 82) under the Śātavāhana empire. They are better named
by González-Reimann, who calls them “safe havens” (2002, 99)—but in
troubled times. Although as we shall see, the safe havens sequence repeat-
edly uses words for “the end of the yuga,” it never uses the term Kṛta Yuga,
and, I will argue, is not describing one as imminent or, much less, as a “real-
ized eschatology.” Here I must disagree not only with Mitchiner but with
González-Reimann, who launches his own discussion of the YP with the
comment that, “At some point in the last centuries B.C.E., and the early
centuries C.E., there must have been real expectations that Kali would end
in the foreseeable future” (2002, 97–98; cf. 144). I believe Mitchiner is
again overstating his case to argue for the priority of the YP’s yuga skein to
that of the Mahābhārata. Mitchiner attempts to make capital of the fact that
the YP does not mention the myth of Kalki, and takes the Mahābhārata’s
prophetic account of that myth by Mārkaṇḍeya—in which Mitchiner finds
similarities to the YP account, and one identical line, half of which we can
basically find as well in Vyāsa’s closing prophesy in the Harivaṃśa37—to be

35. The equation of the Azes and Vikrama eras is disputed in Bracey 2005, based on an inscription dated in
two eras: a Greek and Azes era; he concludes: “The Azes era cannot be dated later than 30 BC or earlier than 80 BC.”
Cf. Cribb 2005. See also González-Reimann 2002, 99; Hiltebeitel 1999a, 263–64: a probably fictional Vikramāditya
is credited with founding an era in 58 or 57 BCE that may have Scytho-Parthian origins in eastern Iran. Śālivāhana,
linked with the Śaka era that seems actually to mark the beginning of Kuṣan ̄ ạ rule in 78 CE, is probably a
personification of the Śātavāhanas (based in Paithan on the Godavari near Aurangabad in Maharashtra, but with
origins in Andhra and north Kanara). He seems to draw his profile especially from the first dynast Simuka-
Śātavāhana (mid-first-century BCE) and Gautamīputra Śātakarṇi (ca. 106–30 CE) who defeated a branch of Śakas.
36. See similarly Bhattacharya 2008: The YP “is the only text to speak of 12 regions (maṇḍalas) that are
peopled after the end of the yuga. It is unique in not terming these survivors as mlecchas.” Bhattacharya misses
the point that these havens would be above all for Brahmins.
37. Mitchiner 1986, 38: YP śl 55ab = Mbh 3.186.33cd: bhovādinas tathā śūdrā brāhmaṇāś cāryavadinaḥ. Van
Buitinen 1975, 587 translates, “The serfs will say ‘Hey you!’, the brahmins will say ‘Pray sir!’ ” It looks to be simply
proverbial. Mitchiner says the Mahābhārata borrows this line (42). But why just a line? HV 116.13cd has (śūdrā
bhovādinaś caiva bhaviṣyanti yugakṣaye).
 dharma over time, ii 289

later.38 But an omission is not a sufficient reason to date one text before the
other. We are, in other words, back to the question of the redemptive telos of
the YP, which obviously—to begin with—simply differs from that in the
Mahābhārata’s Kalki myth, and also from the one in Vyāsa’s Kali yuga
prophesy at the end of the Harivaṃśa, both of which do take us into a new
Kṛta yuga, and, moreover, mention it (in Vyāsa’s prophesy, it will occur when
the Kali yuga’s evils begin to bottom out and dharma’s growth [vṛddhi] begins
to avert its decline; HV 117.40–44). To appreciate the distinctiveness of the
YP’s redemptive telos, we must consider how it uses the notion of “the end
of the yuga” in relation to the account of the safe havens that closes Śiva’s
prophesy to Skanda.
The YP refers to “the end [or destruction] of the yuga” in its run-up to the
description of the safe havens, first to describe the destructive situation: “there
will be an end of the Yuga, the destruction of all living beings (bhaviṣyati
yugasyāntam sarvaprānị vināśanaṃ)” (Mitchiner 1986, 95, śl 91cd). There are then
three successive verses that end in yugakṣaye, “at the end [or destruction] of the
yuga,” for those who will “remain” at that time. Although Mitchiner always trans-
lates yugakṣaya as “end of a yuga,” the more literal translation is “destruction of a
yuga,” which I believe is the preferable translation for YP usages. Notions of yuga
have an entitative aspect as something that has its own dharma, “remnants” (see
below), connective dawns and twilight (in the Mahābhārata), and feet, suggesting
a bull (in Manu 1.81–83; 8.16), and is felt more as an endangered continuum than
as something really about to end (see Koskikallio 1994, 259–60, 267). At such a
time, the good remainder will be persons of “calmness, patience and self-
restraint” (śamakṣamadamās; śl 95c) and “those who maintain firmness”
(dhairyam; śl 97c). But the wicked, those “dear to Kali who ever cause disbelief”
(nāstikyaṃ . . . kalipriyam), will also remain at the destruction of the yuga
(yugakṣaye) (śl 96). Such stock usages of “at the end [or destruction] of the yuga”
for a bad time—what González-Reimann likes to call a “mere metaphor” (2002,
77)—are not sufficient to establish that a Kṛta Yuga follows.
The same compound, yugakṣaye, is then used at the end of the two transitional
verses that describe the “creation” of the twelve safe havens (dvādaśamaṇdạ laḥ):

98. When the world has been afflicted with famine and has become a
terrible fire, [then] from regard for the welfare of living beings, twelve

38. See Mitchiner 1986, 40–42, 44, adducing additionally that the Mahābhārata prefaces its Kalki account
by speaking not of individual kings, as the YP does, but of peoples (3.186.30); the epic’s citation of a Vāyu Purānạ
at 3.189.14 (see chapter 1 n. 25 on named purānạ s known by the Mbh); and that the YP account also omits any
mention of an era of the Seven Ṛsị s, that is, a Manvantara. Bhattacharya 2008 follows Mitchiner on these points.
290 dharma

regions were [i.e., will be] created (durbhikṣābhyāhate loke agnibhūte

sudaruṇe/ avekṣyarthaṃ (var. avakṣyārthaṃ) praṇiṇāṃ sṛṣṭā
dvādaśamaṇḍalaḥ).
99. The remnant in the world who are dear to dharma, those
men who resort to dharma, they will remain at the end of the Yuga,
wearied by hunger and thirst (śeṣā dharmapriyā loke ye narā
dharmasaṃśritāḥ/ kṣutpipāsapariśrāṃtāste sthāsyāṃti yugakṣaye).
(Mitchiner 1986, 96; xxxiv–xxxv)

When it comes now to describing the twelve havens themselves and the condi-
tions under which this remnant of the good will survive, they are said to do so in
four cases “at the end [destruction] of the yuga” (yugakṣaye; śl 102d; 103d; 104d;
113d)—in the last case, which summarizes for the whole, “in that terrible end
[destruction] of the yuga” (ghore tasminyugakṣaye). Clearly none of the remnant
is experiencing anything like a Kṛta Yuga. Indeed, although after mentioning
that those who seek refuge, “longing for a better condition” in the first two
havens, “will attain excellent happiness” (sukhamuttamaṃ) there (śl 101), the
conditions elsewhere seem to be reminiscent of āpaddharma—that is, of topsy-
turvy dharma for times of distress: in riverside, oceanside, mountainous, and
forested regions, “men will live on fish and costly flesh” (102); elsewhere, “. . .
through fish and birds” (107); “. . . on lotus fibres and lotus flowers” (111). And
“on the banks of the Kāverī . . . men will have contentment there, through fish
and boars” (106)! These are safe havens primarily for Brahmins. Indeed, as T. P.
Mahadevan reminds me, supplying the following references, Tamil Sangam
poetry depicts one poet, Kapilar, as a Brahmin who “feasts full on meat” that was
“stewed” or “fried” (Hart 1975, 53; Hart and Heifitz 1999, 248–49).
I believe, then, that we must certainly rule out Mitchiner’s interpretation
that the YP ends with the author and the people of Ujjain enjoying “prosperity”
and happily anticipating or experiencing the Kṛta Yuga. And with the rejection
of that interpretation, we must also dismiss his argument that the account
would be an earlier alternative to the Mahābhārata’s Kalki myth. There is
nothing solid on which to base a claim that it is earlier than the Kalki myth. But
more important, it is not an alternative to it,39 and it is not evidence that the YP
would be authoring “the earliest account of the Yugas in Indian literature”
(1986, vii). But if the YP is not describing a pre-Mahābhārata account of the
Kṛta Yuga, what is it describing and what are its models? I believe there are two
possible answers, both of which would have Mahābhārata prefigurations.

39. Cf. González-Reimann 2002, 99, in agreement with Mitchiner that the Kalki myth is “probably a late
adaptation or reformulation of ideas presented earlier in the Yuga Purāṇa, or some other external source.”
 dharma over time, ii 291

The first explanation would draw on the Mahābhārata doctrine, also

aired in Manu (9.301), that the king makes the yuga. This explanation could
borrow a little from Mitchiner’s view that the Śātavāhana empire had cre-
ated “Regions in Which Men Will Survive and Prosper.” But the explana-
tion has obvious difficulties. Since no particular king is credited, the
doctrine would have to be modified to say that “the empire makes the yuga.”
More than this, it is an account of safe havens in troubled times. Of course,
to say a king creates the yuga is not to limit him, or an empire, to a Kṛta
Yuga. But one does not hear much about kings who create Tretā or Dvāpara
Yugas. For instance, when Kuntī warns her son Yudhiṣṭhira to prepare for
war, she says a good king creates a Kṛta Yuga, lesser kings middling yugas,
and a wicked king goes to hell (Mbh 5.130.14–19). Since in the YP it would
be redundant to create a Kali Yuga, I think this explanation must be put
back on the shelf.
The second explanation, however, is sound, and has a more well-developed
Mahābhārata background, indeed, a myth. Rāma Jāmadagnya empties the
earth of Kṣatriyas twenty-one times. But the job is never complete. Yudhiṣṭhira,
who has heard the story once in the forest, hears it a second time after the war
from Kṛṣṇa, who tells it to deter him from disavowing his hard-won king-
dom.40 In this account, the goddess Earth tells the Brahmin Kaśyapa she has
made safe havens for the eventual regeneration of the Kṣatriya class, and
implores Kaśyapa to reinstate them as kings to protect her. As Kṛṣṇa tells it
(Mbh 12.49.66–75), neo-Kṣatriyas were raised in different forest, mountain,
riverside, and oceanside41 locations by bears, seers, cows, calves, monkeys,
and the ocean (see Hiltebeitel 1999a, 460; Fitzgerald 2004a, 279–80). The
YP, however, would be drawing on this myth to envision safe havens for
Brahmins.42 Similarly in Vyāsa’s prophesy at the end of the Harivaṃśa, those
who migrated out of the northern central midlands with sons on their shoul-
ders and settled in new places, including mountains, oceansides, and forests,
“will subsist on deer, fish, birds, dogs’ feet (śvāpadaiḥ), all insects/worms
(sarvakiṭakaiḥ), honey, vegetables, fruits, and roots,” and will wear tree bark
like Munis (HV 117.32–33).

40. At first, says Kṛṣṇa, Śūdras and Vaiśyas united with Brahmin women to produce a kingless condition
in which the strong ruled the weak (12.49.61–62). The Southern Recension adds that Brahmins abandoned their
dharma and turned to heresies (pāṣaṇḍān; 111* line 4 following 49.62).
41. I assume Ocean did its protecting by the oceanside. Some northern texts add that the ones thus pro-
tected lived among blacksmiths, goldsmiths, and such (vyokāra-hemakārādi; 75 and 114*).
42. Indeed, the Kalki myth would envision something similar, but after the intallation of the Kṛta yuga,
where it describes the return of Brahmins to conquered territories; see chapter 12 § B.
292 dharma

A.4. Yugas, Yavanas, and Current Considerations

So far, I have argued that that the Mahābhārata and Manu are the texts that
made the earliest Brahmanical formulations of yugadharma, doing so in a con-
text where the yuga’s connections with the continuity of dharma over time
would have cut a wedge that distinguished these formulations from Buddhist
ones that traced dharma—or at least “the true dharma”—over time discontinu-
ously through kalpas. As I have mentioned, von Stietencron’s study of geneal-
ogies in the vaṃśānucarita (2005) brings in text-historical and geographical
considerations about the Indian prehistory of this yuga theory that shed
background light on both the Mahābhārata and the YP. For deeper background,
I will suggest that we also think further about the YP’s Yavana–Pañcāla–Mathurā
alliance.
Von Stietencron’s interest lies in the Bhaviṣya(t) Purāṇa (henceforth
BhavP) portion of the vaṃśānucarita, which “consists of a list of kings and sages
from early times up to the fifth or seventh generation after the Mahābhārata
war, written in the past tense; and a second [list] in the form of future [bhaviṣya]
prophesies which continues the line of kings into the Kali age in various redac-
tions until it is completed in the early Gupta era.”43 As noted, the prophesies
occur in the BhavP portion, which von Stietencron distinguishes from what he
calls the “literary framework,” in which the Pāṇḍava descendant Adhisīmakṛṣṇa
performs a sacrifice in the Naimiṣa Forest, where the vaṃśānucarita is recited.
“What is certain,” he says, “is that the redactors belonged to different regions.
The focus of the earlier genealogies is the present-day Uttar Pradesh and the
adjacent areas to the south and west” (2005, 78–79), and “indicative of the area
of the Kurupañcālas” (73). “In contrast, in the BhavP portion the focus of the
first and second redactions is Magadha” (78–79). Von Stietencron considers
the BhavP portion to be the nucleus, that it would have been first redacted
under conditions of “Brahmin restoration” (72), and that the best candidate is
the early reign of Candragupta Maurya before he turned to Jainism (72)—one
reason being its disapproval of the Nandas for overrunning “the old dynasties
of North India” (79–80). In assigning the BhavP to the early reign of the first
Mauryan, Candragupta, von Stietencron brings in what he calls “irrefutable
evidence”—though it is “anything but obvious”—that this account already
existed in Candragupta’s time. Megasthenes reports that Indians listed 153
kings down to Candragupta (whom the Greeks called Sandracottus). The BhavP
lists either 150 or 154, which, says von Stietencron, is too close to Megasthenes’
number “to be a mere coincidence” (2005, 82–83).

43. Von Stietencron 2005, 65; cf. 73 and n. 12: as discussed in chapter 5 § C, ĀpDhS mentions a Bhaviṣya
Purāṇa.
 dharma over time, ii 293

Von Stietencron relates this dating to the yuga concept: “The collapse of the
old Kṣatriya dynasties under the vigorous policies of Mahāpadmananda, who is
explicitly denigrated as the son of a Śūdra (śūdra-yoni), the support extended by
the rulers to non-Vedic religions and the invasions of barbarians—all this is seen
negatively as a sign of Kaliyuga” (2005, 72). The BhavP section would, however,
be narrativizing the past-to-“future” of such defeated royal lines “not for the
purpose of reinstating them as independent regents—since the vast centralized
empire had already proven its advantage—but in order to base the power wrested
from the deposed Nandas on the co-operation of those very people who had suf-
fered due to the former and hated them” (2005, 81). Yet as von Stietencron is
quick to point out, “The yuga doctrine appears to be a relatively late insertion in
the Bhaviṣya(t) Purānạ and certainly postdates Mahāpadma Nanda due to whom
the Kaliyuga grows in vigour (vṛddhiṃ gamiṣyati). If at all conceived by the first
compiler of the Purānạ at the time of this supposed Brahmanical restoration, its
yuga calculations must have been based upon short time cycles which were not
yet in conformity with the later teachings.” He speaks here of “the later yuga
̣ ạ ”—which, as González-
concept” that starts the Kali Yuga “with the death of Kṛsn
Reimann demonstrates (2002, 51–52, 60, 73, 94–97, 105, 115–16), is not
formulated around that precise turning point until the Purānạ s.44 Von Stietencron
would be buttressed on this point by González-Reimann, who documents usages
of yuga in the Ṛgveda, the Brāhmaṇas, and the Jyotiṣa Vedāṇga (which he dates ca.
the fifth-century BCE) for time spans (a generation, a five-year cycle, “an age in
general” [2002, 6–7, 16 n. 10]) that, as we have seen, are far shorter than those
that come to be associated with the yuga in classical, including epic, usages.45
Von Stietencron also refines these chronological considerations about the
two portions of the vaṃśānucarita in relation to geographical ones, calling
attention to their handling of two different groups of kingdoms:

1. Those from Kosala to Avanti, “well documented in the Bhaviṣya(t)

Purāṇa, . . . which were conquered by the Śiśunāgas and the Nandas,
primarily under Mahānandin and Mahāpadma Nanda, and subsumed
under the first great North Indian empire with its centre at Magadha,”
whose expansion would have begun with Udāyin, whom the YP
mentions so prominently.

44. As von Stietencron remarks, “If this concept existed when the Bhaviṣya(t) Purāṇa was compiled, one
would expect that kings from Parikṣit onwards would have all been listed as the future rulers of the Kali Age. This
is not the case. Therefore, there almost certainly existed a Bhaviṣya(t) Purāṇa prior to this doctrine, a text which
started six–eight generations after Parikṣit in the future tense and which remained unmodified in its kernel while
being supplemented at its end. The later yuga doctrine is definitely in existence by the time of the final redaction
in the early Gupta period” (2005, 73).
45. See González-Reimann 2002, 223–25 on use of parvan rather than yuga for dice throws in the Ṣaḍviṃśa
Brāhmaṇa.
294 dharma

2. A list of “perfunctorily mentioned” dynasties originating “in another

context,” including Kurus, Aikṣvākus, Pañcālas, Kāśeyas, Kaliṅgas,
and Śūrasenas, all belonging “to families found in the list of early
dynasties, and whose absence in the Bhaviṣya(t) Purāṇa was bound
to be conspicuous when a complete redaction of the vaṃśānucarita
section was made. At the same time they are families which, in a
favourable or unfavourable sense, are connected to the later Kali Yuga
doctrine. . . . They perish in the Kaliyuga, are known opponents of
Kṛṣṇa, or are supporters of the Bauddhas and Jains.” (Von Stietencron
2005, 79–80)

Von Stietencron is “certain . . . that the redactors belonged to different regions.

The focus of the earlier genealogies is the present-day Uttar Pradesh and the
adjacent areas to the south and west. In contrast, in the Bhaviṣya(t) Purāṇa the
focus of the first and second redactions is Magadha (Bihar); it is the Dekkhan
when it comes to the third and fourth and, with the last redaction, it is once
again Magadha” (2005, 78–79). On the other hand, the “literary framework” is
“indicative of the area of the Kurupañcālas, as seen in the mention of the place
of the prophecy as the Naimiṣa Forest” (73).
Yet perhaps we need not posit separate regions behind the vaṃśānucarita’s
two sections. As the Naimiṣa Forest setting suggests, the Mahābhārata is a
likely model for the belated “literary” framing of the BhavP. We may thus posit,
at some point after the early reign of Candragupta Maurya, that familiarity with
the Mahābhārata had intervened.
Now if, as I have further argued, the YP builds on the Mahābhārata’s
version of yugadharma as the basis for its ex eventu prophesies, and these
prophesies yield information about the historical conditions under which this
chronometry was generated, what are these conditions? Clearly, the Mahābhārata
would have intervened at some time between an early Mauryan redaction of the
BhavP and the late first-century BCE date that is likely for the YP. Although the
YP ends up looking south, it, like the Mahābhārata, Harivaṃśa, and Manu,
places most of its emphasis on northern India, and, like the Mahābhārata more
specifically, as we shall see in chapter 13, on tensions between the northern
midlands and Magadha as the hub of Greater Magadha. Von Stietencron shows
that such tensions would be traceable to the early Mauryan period, and that if
the early Mauryans drew on a yuga concept, which is not necessary to his
argument, it would have designated only short time cycles or generations.
Megasthenes’ report from Candragupta’s Maurya’s time of a genealogy of 153
kings shows that the Mauryans had an early historical record even before Aśoka,
probably without yugas. But for its association of the yuga with dharma over
time, the YP has certainly drawn on the Mahābhārata.
 dharma over time, ii 295

A four-age yuga theory was probably brought into the Mahābhārata as an

historical armature with which to loosely incorporate such genealogical and of
course other highly “coded” information. On that note, before setting our sites
on some comparable Buddhist texts, the following points can summarize
matters for now just on the Mahābhārata and the YP.

1. In comparison with the YP, there is obviously a difference between a

redemptive telos prophesied by Śiva to his son the war god, as if the
two were holding their destructive power in reserve, and a text whose
redemption of dharma prophesies a restoration of the Kṛta Yuga under
Kalki Viṣṇuyaśas, “the fame of Viṣṇu” (Mbh 3.188.89a), and whose
main redemptive narrative—as the YP itself would seem to recognize—
is managed from within by a Kṛṣṇa who takes birth from yuga to yuga
whenever dharma declines.
2. The Mahābhārata provides reassuring models for the continuum of
yugadharma not only in its main story but in the Rāma Jāmadagnya
myth. The YP simply extrapolates this kind of reasurance into Ṥiva’s
long-range prophetic view of recent history. Instead of presenting a
new Kṛṭa yuga as the outcome of the end or destruction of the yuga, its
emphasis is on the reassurance that even in the destruction of a yuga,
meaning even in difficult times, dharma always has a remnant.
3. In comparison with the BhavP’s Magadha-oriented history, which
probably, under the impact of the Mahābhārata’s frame story, only
belatedly prepones an epico-purāṇic Naimiṣa Forest setting for itself,
the YP’s Midlands concentration looks askance at Magadha, but also
looks hopefully to a third region, the South, for the survival of its
dharma remnant. In fact, it looks to two separate regions there: one,
the northern Deccan; the other, the Kāverī River valley, with its
welcoming fish and boars.
4. This look to the south may coincide with another reason why the YP
stops its north Indian history with the Mitra kings and Śakas. Its
redemptive history was proving to be wishful thinking. The Śakas
did not really “go home” like the Greeks. Around 25 BCE, they
established themselves as Kṣatrapa (Satrap) successors to the Mitras,
and remained to stay, forming trade networks and power points at
centers from Taxila in the northwest to Mathurā in the midlands,
and also to Ujjain in the west, where they contributed to the support
and spread of mainly Buddhism.46 There would be no reason for the

46. See Neelis 2007, 72–79; 2008, 8–10; 2009. Eventually, they would be accorded the little king status
of “degenerate (vrātya) Kṣatriyas (Thapar 1992, 153), which would lend itself to the colonial period’s idea of the
“Scythian origins” of the Rājpūts (Hiltebeitel 1999a, 439–40).
296 dharma

YP’s composer to envision safe havens in the south under the

Mitras, but there would have been under the Kṣatrapas.
5. Following the arguments of T. P. Mahadevan (2008), this could
well describe the conditions in which, by around 25 BCE, out-of-
sorts Pūrvaśikhā Brahmins from the Mathurā–Kuru–Pañcāla
northern midlands would have headed south and reached the
Tamil-speaking regions with the Mahābhārata archetype in their
possession.
6. As it does today, that archetype would have included passages
reflecting the theory of four ages and association of the epic’s main
story with the turn from the Dvāpara to the Kali Yuga—albeit, as
González-Reimann has demonstrated, associated loosely, inconsis-
tently, and often metaphorically treated, and sometimes “confusing”
yugas and kalpas (2002, 72, 85 n. 150), since, by the post-Aśokan
time of the epic’s composition, the kalpa, originally a Greater
Magadhan concept of cyclical time, had been taken up in
Brahmanical texts and considerably modified in the Mahābhārata to
round in the yuga.
7. As to the separate origin of the concept of four yugas, a Yavana–
Pañcāla–Mathurā coalition to destroy the Magadha capital in
conjunction with the rise of Puṣyamitra Śuṅga in ca. 187 BCE provides
a point at which we can begin to imagine how the skein of four
declining ages would be among the “connection points” that the
Mahābhārata poets profoundly reworked under the Vedic name of
yugas from what Fernando Wulff calls the Greek repertory with its five
ages, the fourth being an Age of Heroes.47
8. This would have been done by people interested in narrating and
preserving their own history (itihāsa) in relation to a new epic vision of
dharma over time, concentrating on their Kuru–Pañcāla and other pasts,
and promoting in Vedic idioms a view that Vedic dharma perdures,
whatever its pitfalls in history, through all time and generations.48

47. Wulff 2008, 116, 153–56 has reinforced the likelihood that the skein of four declining yugas would
be among the “connection points” that the Mahābhārata poets reworked from “the Greek repertory of the
Mahābhārata” from the Greek scheme of the decline of Justice through five races or ages. On this matter of
longstanding discussion, cf. Beall 2005/6 n. 33, citing Hiltebeitel [1976] 1990, 48–59 for “a summary
treatment” in which I note “that in the Indian case the Bhārata war is an episode between the third and fourth
ages (contra Clay’s [2003], 81 belief that the oriental parallels lack heroes).” For fuller discussion, see
Hiltebeitel 2011a, chapter 4.
48. Visvanathan 2010 makes the intriguing point that “the genealogical moment” when royal genealogies
begin to show on inscriptions is post-Aśokan (see chapter 2), and begins on Śuṅga inscriptions. This Śuṅga royal-
Brahmanical interest could have a contemporary parallel in the Vedic genealogies of the Mahābhārata.
 dharma over time, ii 297

Although this is not history as a scientific sorting out of the past, it implies,
just as any historical fiction would have to, a readership that would know
what a sorted out and sequential history was. If it was done by out-of-sorts
Brahmins in territories ruled by little “Mitra” or “Agniveśya” kings, we do
not have to worry about finding them a big royal patron.49
9. Evidence that the Mahābhārata and these changing times are in
mutual attunement thus includes not only a post-Mauryan framing of
the vaṃśānucarita’s narration in the Naimiṣa Forest and the filling out
of the yuga concept but also the fact that, like the Pañcālas and
Śūrasenas in the YP, the Kauravas have Yavana allies in the
Mahābhārata war (5.196.7, etc.),50 and that next in store, according to
Mārkaṇḍeya’s prophesy, Greeks and Śakas will be among the barbar-
ians who become kings in the Kali yuga (3.186.30).
10. Finally, the essentially linear yuga would have served such poets as a
better armature for history than the kalpa. The Mahābhārata may
“invent” the five Pāṇḍavas, the 100 Kauravas, and the six generations
of which those vying cousins are the fourth.51 But as Witzel shows, its
poets extracted plausible historical data from all parts of the Vedic
canon to trace out those momentous generations’ pasts and futures in
their itihāsa of the Bhāratas. Buddhists, on the other hand, beyond
tracing a discontinuous dharma through a vast cyclical cosmology,
found the kalpa most relevant to Jātakas and remembering past lives,
as in the Buddha’s recollection of a lost city in the Mahāsudassana
Sutta, of which Rupert Gethin can say, “There is nothing here that the
modern mind would be tempted to read as history” (2006, 63). With
the seeming exception of Aśoka, who may have used kalpa with the
connotation of a dynastic era or age, Buddhists had to find other more
immediate ways to trace the decline of dharma in history.

49. See Hiltebeitel 2001a, 19 proposing that the Mahābhārata “was written by out of sorts Brahmans who
may have had some minor king’s or merchant’s patronage, but, probably for personal reasons, show a deep
appreciation of . . . Brahmans reduced to poverty.” Fitzgerald’s turn to Khāravela (2010; see above nn. 25 and 30)
builds from an opposite assumption, which he stated earlier as follows: “The production and promulgation of
this text would have required a major effort and significant expense, so we must imagine the support and backing
of some prince or princes, or direct imperial support” (2001, 69). Cf. Witzel 2005, 48, 60, 62 n. 132, 64, 70.
50. Mbh 5.196.7, etc. It is of interest that the Kauravas’ sister Duḥśalā has Greek cowives who join other
Sindhu, Sauvīra, Gāndhāra, and Kāmboja ones in mourning their husband Jayadratha Saindhava (11.22.10–11).
The Pāṇḍavas also had to subdue “the Greek’s city” (2.28.49) before Yudhisṭhira could do his Rājasūya.
51. Again, a comment on Fitzgerald 2010, who, like many before him (see Hiltebeitel 2001a, 2 n. 10), and
̣ ạ vas. In mentioning the six main genera-
Witzel (2005, 28, 35), wants to limit “invention” in this span to the Pānd
tions, I note that one of the surprises of Wulff’s thesis (2008, 81–92. 102, 205–24) is that the Mahābhārata poets
would have begun reworking a Greek epic repertoire that unfolds from a water goddess who, like Gaṅgā, gives birth
to a mortal son over whom she will eventually grieve. On Gaṅga and the “history of the Bhāratas,” see chapter 8.
298 dharma

B. Variations on the Debacle at Kauśāmbī

This brings us to a set of Buddhist texts that emerged against a background of

incursions not only of the Greeks and Ṥakas but also the Pahlavas—with the
Kuṣan̄ ạ s, Iran, and Rome also appearing in some later versions. As Jan Nattier
has shown, these prophetic texts take us into Buddhism’s historical self-
consciousness (1991, 133–42), which we can now compare with the surprising
historical self-consciousness evident in the YP. In each case, the texts provide a
way to follow a course of predictions as they are tied to changing historical and
social realities, and to ways in which Hindus and Buddhists considered dealings
with their inner and outer others: in the YP, concerning a dharma defined mainly
by population and culture; in the Buddhist texts, a dharma defined principally by
Vinaya. In returning our attention to Buddhist texts, we thus also pick up from
our discussion of Vinaya at the end of chapter 4. There, I sought to move the AS’s
Vinaya argument with householder Brahmanism out of pre-Mauryan times and
into the third- or second-century BCE. I will now move such considerations a few
centuries forward to a point where one can detect more advanced concerns about
the relation between Vinaya and Brahmanical society, and also the emergence of
the Mahāyāna. While keeping in mind the continued unfolding of predictions of
the end of the Buddhist dharma, I will focus on one such prediction that is also a
fascinating narrative, which, for heuristic purposes and in contrast to the AS
“parable,” I will call “the Kauśāmbī myth.”52 Unlike the AS, it is canonical in both
Nikāya-school versions and Mahāyāna versions, and almost certainly reflects his-
torical and geographical conditions in which the spaces shared between these
two great branches of Buddhism would have been part of its background.53 Like
the AS, it is a story about Vinaya rules and the primacy of the arhat.
For nearly a millennium, the Kauśāmbī myth had a surprising vitality,
after which it aroused little interest. I will be discussing how it took on different
contours in different lands during what has been called Buddhism’s Middle
Period, which introduced the Mahāyāna and involved the transmission of
Buddhism to central Asia and China. Although there were Sanskrit or Prakrit
originals behind all the versions I will discuss, none survive. This means that
I will rely on translations and summaries, mainly of Chinese texts, for which

52. For Nattier, “a unique scenario,” “the only canonical tradition that combines a prophesy of complete
extinction of the Dharma with a narrative account of the actions leading up to this event”—compared in this with
the general sense of decline in the Lotus and Sukhāvatῑ-vyūha Sūtras (1991, 131 and n. 27).
53. See Nattier 1991, 127–31 on its distribution; 145 on its “long period of trans-sectarian currency” (145).
Note that I mention “spaces shared” to register that it is not found in the Sri Lankan Pāli canon (on which see
more below).
 dharma over time, ii 299

Nattier 1991 is my main source. My discussion must always be cautions when

Nattier summarizes rather than translates the Chinese texts.54 But I think it is
worth rearranging the deck chairs of this Buddhist Titanic in some new ways.
Nattier discusses thirteen extant versions of the Kauśāmbī myth, most in
Chinese, but also in Tibetan and Khotanese, organizing them into a branching
stemma involving “archetypes”—including an “original” archetype—behind
extant versions. I have five misgivings about this approach.

1. The archetypes are hypothetical and selective.55

2. Texts that synthesize varied versions of the myth are omitted from the
stemma.56
3. The approach encourages thinking about what Nattier calls a “core”
story and treating ways the story is “framed” as secondary.57
4. Based on this “core” Nattier proposes an original historical incident
behind the myth.58
5. Nattier wants to demonstrate that the story is originally nonsectarian,
and remains so. That is, the “core” would be consistent on doctrinal
matters, particularly, as we shall see, in the implications of the
contrastive terms True Dharma and Semblance Dharma.

I cannot go into detail on the first four points beyond what is noted. Suffice
it to say that methodologically, in treating myths or for that matter any narra-
tives, one would naturally prefer to be dealing with actual texts than hypothet-
ical cores and archetypes. It is regarding point 5 that I will be rearranging the
deck chairs. Rather than a branching linear stemma, I will take up versions of
the Kauśāmbī myth in three successive text clusters. These are:

1. an “alpha-cluster,” for what is probably the oldest version (late first-

to second-century CE), linked with the expansion of Buddhism into
Central Asia and reflecting, I believe, early tension between Nikāya
school and Mahāyāna Buddhism;

54. As Nattier reminds me in an e-mail (2008a) after reading a shortened version of this analysis prepared
for a talk, the Chinese summaries she drew on were aided by John McRae (see Nattier 1991, xii). I thank her for her
continued interest. The talk, presented as “The End of the Dharma in India, Central Asia, and China,” was given at
the “Bringing Buddhism to Varied Lands” Colloquium at George Washington University, April 14, 2008.
55. Nattier comments (2008a, 2) that my working with “clusters” is also “selective.” That is to be granted.
56. See Nattier 1991, 214: The stemma principles are unable to resolve three types of complications: deliberate
condensation to fit a stylistic genre, as with the Mahāvibhāsạ ̄ version; use of more than one version in a retelling, also
as with the Mahāvibhāsạ ;̄ and insufficient data to account for lost oral versions, as with Mahāmāyā-Sūtra.
57. For this terminology, see Nattier 1991, 177, 186, 206, 224 on the “core” (the continuity of the core is,
I believe, overstated); 183–85, 207 on the “frame” and “framing.” On textual “frames,” see chapters 4 § C.3 and 5 § C.
58. Nattier asks, “Or might the original story simply have been the narration of an actual historical event which—
despite its manifestly uninspiring content—was simply too vivid a memory to be forgotten?” (219–20). Opening this
thought from the beginning (4), she eventually favors it, but as far as I am concerned it is unconvincing.
300 dharma

2. an “S-cluster,” not much later, linked with Sarvāstivādin and

Mūlasarvāstivādin Buddhism59 as the schools most representative
of entrenched Nikāya school Buddhism in Northern India; and
3. an “M-cluster,” later than the other two by about two centuries, linked
to the spread of the Mahāyāna from India principally to China, and
also Tibet.

Now, according to Nattier, we can say “with certainty” that “the original
Kauśāmbī story was utterly devoid of Mahāyāna content” (1991, 223). Although
Nattier is a shrewd detective on the sectarian provenance and implications of
each text, my suspicion is that even if Nikāya school versions are earlier and do
not mention the Mahāyāna or show its features, each of these clusters, and
most clearly the first and the third, reflect perceptions that Nikāya school–
Mahāyāna tensions were an unstated symptom (others being stated) of what
will bring the dharma to its end. Rather than identifiable content, I believe we
are dealing with knowing allusions. As we have seen already in the three pre-
ceding chapters, Indian dharma texts can be quite coy in alluding to competing
views of dharma. In this myth, the end of the dharma is a topic that engages
Buddhists in envisioning differences mainly among themselves.
Nattier points out that the Kauśāmbī myth is “conspicuous by its absence
from the voluminous Theravādin literature” (1991, 222). While this holds for
the Pāli canon, a late version, apparently minus any reference to Kosambi/
Kauśāmbī, and with other truncations, has made its way into a fifteenth/early
sixteenth-century Pāli chronicle in northern Thailand, the Tamman Mūlasāsanā
(Veidlinger 2006, 22–23). The Buddha prophesies what will happen in 5,000
years. The three baskets will vanish in order: Suttas, Abhidhamma, and Vinaya.
At that point, “‘Vinaya will be corrupted and will vanish.’ And furthermore,
once even the Pātimokkha and Pārājika have disappeared because people who
know them can not be found, the following disturbing scenario will unfold:
‘When a king will desire to hear the Dhamma, he will not be able to find someone
in the kingdom who knows the Dhamma.’” The apparently unnamed king will
put a prize of gold on the back of an elephant and have it paraded around his
kingdom with the announcement that, even if someone knows just “a four-line

59. A plausible approach to the relation between these two schools was presented in the “Early Buddhism”
panel at the 15th Congress of the International Association of Buddhist Studies, June 24, 2008, by Alexander
Wynne, arguing that while they probably did not differ in doctrine or use of extra-Vinaya texts, their different
Vinayas point to their separation over a Vinaya dispute: the Mūlasarvāstivādins, centered around Mathurā,
claimed to be the “source” (mūla) of the Kashmir-based Sarvāstivādins’ Vinaya, that the Sarvāstivādins were an
offshoot, and that their Vinaya was inferior. Cf. Nattier 1991, 43 n. 41, 205–6, 222; Schopen 2004, 25 and 41 n.
35: the name could mean “either ‘the Original Vinaya of the Sarvāstivādins’ or ‘the Vinaya of the Original
Sarvāstivādins.’ ”
 dharma over time, ii 301

verse,” he can claim the prize. But the search will be futile and the prize will be
returned to the king. “Thus at this time the knowledge of the texts and thereby
the doctrine of the Buddha will be lost” (58–60). Missing details of this story
will soon become clear, and the dating of this Theravāda text will not be
surprising should the story have routed itself from China or Tibet.
More immediately, however, the Theravāda Vinaya does tell at length about
the Buddha’s dealings with the fractious monks of Kosambi. The account is fas-
cinating for everyone’s sustained silence on what “that matter” of dispute was,
with the Buddha finally steering the two factions, once they are ready, to a solu-
tion despite having never gotten “to the bottom of ” it, as he himself tells the
Vinaya-master Upāli (Mahāvagga, Khandaka 10 10.6.2).60 According to H. W.
Schumann, “If we are to believe the commentary, it was a lavatory water-jar. . . .
A monk . . . had left the latrine without throwing out the remaining water.”61
Whatever it was, one senses that the Buddha is portrayed as preferring to let the
monks reach their own conclusions than to rule on the specific matter. Other
scholars (Chappell 1980, 128; Lamotte 1988, 198) have thought that they detected
an echo of the first difficulties at Kosambi in the prophetic Kauśāmbī myth, and
I believe this is more plausible than an “original” historical incident. I suggest
three possible thematic continuities. In both, we find a condescending attitude
on the part of the leader of one faction. In both, the contending parties are
unable to perform the prātimokṣa together. And in both, there is a rebuke of the
bickering monks by the laity.62
Let us, however, grant that a myth, not to mention a prophetic one, does
require its historical grounding to be forgotten, or at least misrecognized.
Nattier draws a distinction between an historical arhaticide which she finds “no
persuasive reason to question” and the decline prophesy, which “is another
matter altogether.”63 But we can only progress if we put the arhaticide back in
the myth, leave its historicity as a distraction, and keep our questions to the
historical grounding of the myth itself. Like any good myth, it can be expected
to weave together a number of oppositions, both tacit and explicit; to leave

60. See Rhys Davids and Oldenberg 1982, 2: 285–325; see also the Kosambiya Sutta (MN 48).
61. Schumann 1989, 119, and more fully 117–21. Cf. Ñāṇamoli 1972, 109–19.
62. The only locational note I find is that it would not be surprising “[t]hat a non-Mahāyāna story should
be set at Kauśāmbī, since both Fa-hsien and Hsüan-tsang report it to have been mostly or entirely non-Mahāyāna.
Hsüan-tsang must know the story, since he reports of Kosambī: ‘The law of Śākya becoming extinct, this will be
the very last country in which it will survive’” (Nattier 1991, 223 n. 16).
63. Nattier 1991, 225. This gets quite contrived. Nattier thinks the names of Śiyṣaka, Sūrata, and Aṅgada
meet the test of historical verisimilitude since they are unusual and not “explicitly didactic.” But what could be
more didactic than Śiṣyaka, “He Who Has Disciples”? She asks, would not a different, less “ambigious” kind of
story “have served the purpose better?” “Only later, perhaps, was the story shifted into the future and brought into
conjunction with a prophesy of the end of the Dharma” (220–21).
302 dharma

incongruities from one version to the next; and to be adaptable to different his-
torical, social, and, in the present case, sectarian purposes. Much safer historical
grounding is to be found where Nattier traces the origin of the myth itself—for
her, the tale of the end of the dharma—to a period no earlier than the occur-
rence of actual invasions by Greeks, Śakas, and Parthians, whose three kings
bear these peoples’ names as ethnonyms in the myth.64

B.1. The alpha-cluster and the Prophesy of Kātyāyana

Nattier posits that the oldest extant version of the Kauśāmbī myth is found in
the verse and prose versions of the Prophesy of Kātyāyana. She makes two major
points in favor this priority. The prophesy is made by the Buddha’s disciple
Kātyāyāna and not, as in all other versions, elevated to the sūtra status of
Buddhavacana, “the word of the Buddha”; and the poetic version is the first
textual account of the Kauśāmbī myth to have been translated into Chinese dur-
ing or before the Western Chin dynasty (265–316 CE) (1991, 157–58). The first
point is the hardest to argue around, and looks to me like a solid indicator of
priority. Now, if we look at these two texts as a whole, and ignore the attempt to
discern an original historical incident behind their alleged “core,” we have a
story that may have considerable heuristic value: “an unnamed king ruling at
Kauśāmbī . . . successfully repels a foreign invasion” of Greek, Śaka, and
Parthian kings. As Nattier shows, this places the myth in a period no earlier
than the occurrence of actual invasions by these peoples. To celebrate, the
Kauśāmbī king’s preceptor Śiṣyaka recommends that he invite “the monks
from surrounding regions to Kauśāmbī.”65 There, a confrontation occurs,
which the Prophesy of Kātyāyana in verse recounts as follows:

A group “numbered in the hundreds of thousands” will meet, and on

the 15th day (when the Prātimokṣa is to be recited), a quarrel will
break out. A large group of bhikṣus, thinking themselves superior to
the rest, issue a condescending statement to the effect that they
understand more about Dharma than anyone else. “If there is one
bhikṣu who has attained enlightenment,” they assert, “then he can
preach about the origins and end [of the Dharma], and we will study
his sūtra[s].” Much to their dismay, however, just such a bhikṣu

64. See Nattier 1991, 226, tracing the origins of the myth to “northwest India sometime between the
beginning of the 2nd century CE and the middle of the 3rd” when “throughout most of this period northwest
India was part of the powerful Kushan empire.”
65. The quotes are from Nattier’s description of the original archetype behind the alpha-cluster (1991,
219–20).
 dharma over time, ii 303

happens to be present in the audience. This monk, named Su-lai

(Skt. Sūrata?), rises from his seat and proclaims that he does indeed
keep the precepts, has no doubt concerning the Dharma and Vinaya,
has penetrated the scriptures, and truly understands the meaning
of the path. He rebukes the self-righteous monks for praising
themselves so much, which arouses the anger of a student of the
honored teacher (= Śiṣyaka), named A-ssu («Skt. Aṅgada?), who then
strikes and injures the Arhat. A Dharma-loving Yakṣa66 . . . then
seizes a vajra and kills A-ssu. The earth shakes in the familiar six
ways, various ill omens appear, and the ultimate fate of the Dharma
is described as follows. “The lamp of the Dharma is already out.
The correct scriptures are already destroyed.” (Nattier 1991, 158–59)

Now if Sūrata is killed but not Śiṣyaka, and Śiṣyaka’s crowd get to go on with
their truncated following of the rules, and if “this primitive version” ends with
that, it would mean that the alpha-cluster myth is originally about the end of the
Dharma as something that occurs with the death of the last arhat. To keep
Steven Collins’s phrase alive (1993; see chapter 4 § C.3), the arhat is “what is
primary”—his death alone marks the end of the dharma. Śiṣyaka is not killed as
he is in all other versions than the verse Prophesy of Kātyāyana, including its
prose counterpart. The arhat’s primacy both stands out as exceptional and is
found in what appears to be the earliest version (see Nattier 1991, 210–11).
Indeed, this is an arhat who knows, follows, and dies championing the
Prātimokṣa. With that, we would have a suggestive way to demarcate the two
factions, for the arhat could be “what is primary” only on a Nikāya school side
of the aisle. Although early Mahāyāna practitioners would have followed the
Prātimokṣa and been at least as strict about the rules, the arhat would be pre-
cisely what is not primary for the vehicle of the bodhisattva.67
Indeed, the text summary offers some irresistible clues in this direction.
The “large group” that thinks it knows more about Dharma than anybody else
says that if there is anyone who has attained enlightenment, “they will study his
sūtra[s]”! There is something poignant in an arhat, who is enlightened, sidestep-
ping an invitation to preach “his sūtra[s].” They look to be something other than
the scriptures the arhat has “penetrated” and the “correct scriptures” that are

66. Nattier 1991 reads “demon” at this point. Her e-mail (2008a, 2–3) clarifies this now to a Yakṣa.
67. I thank Nattier 2008a, 3 for her cautionary comments on this point, calling attention to Daniel
Boucher’s forthcoming book on the Rāṣṭrapāla Pāripṛcchā, and maintaining that the Prātimokṣa is “a ‘prerequi-
site for both groups for spiritual advancement, but not the be-all and end-all. . . . Thus the observance of the
Prātimokṣa doesn’t work as a way to distinguish Mahāyāna from non-Mahāyāna practitioners.” My point, how-
ever, is not that the Prātimokṣa distinguishes the two, but the emphasis on the arhat’s allegiance to it.
304 dharma

then said to have been “destroyed” with his passing. It looks like the condescending
faction would be ready to listen if he or someone else were to preach Mahāyāna
teachings. And if it is really a question of “sūtra[s]” and not just his “dharma,”
they could even be participants in what Schopen calls “the cult of the book.”68
A Mahāyāna faction, or better, a mixed monastic group, projected as “all the
monks of Jambudvīpa” (158), would then be left with a Mahāyāna-inclined
Teacher (Śiṣyaka as “He Who Has Disciples” [Nattier 1991, 153; cf. 292]), who
could well be portrayed as insouciant in honoring the Vinaya rules.
Taking this a few steps further, the Prophesy of Kātyāyana in prose mentions
that the dispute is over a prātimokṣa of 250 rules. This opens the possibility that
the myth had this early circulation in the one Nikāya school that had that pre-
cise number, that of the Dharmaguptakas.69 This school had not only the usual
Three Baskets but a fourth, the Bodhisattva Piṭaka.70 With such an “open
canon,”71 and one open in this specific way, the Dharmaguptakas could have
provided a fitting milieu in which to take interest in a myth that involved ten-
sions between the two parties in a still-early phase, such as Buddhist studies
has recently come to recognize,72 of mutual side-by-side communion. It is also

68. Nattier 2008a, 3 sees “no evidence of ‘the cult of the book here,’” saying that what appears above as
“sūtra[s]” could simply be translating “Dharma.” This would not rule out reference to specific sūtra[s], but it would
weaken this part of my argument. So I asked Nattier if she could look again. I thank her for this follow-up
information (2008b), which leaves me some wriggle room: “First of all the term that was translated as ‘sūtra[s]’ . . .
does turn out to be jing (this occurrence is in T2029, 49.11c28) . . . the term generally used to translate ‘sūtra’ in
scripture titles, but [used] more broadly . . . to mean simply ‘classic’ . . . [but also] one of several expressions used
by a number of early translators to translate ‘dharma.’ My own take on this is that for a Chinese audience Buddhism
was represented primarily by its texts, and thus when an Indic-language original talked of the Buddha ‘preaching
the Dharma’ (dharmaṃ deśayati sma, etc.), the Chinese translated it as ‘preaching the classics’ (shuo jing). A differ-
ent explanation is given by Stefano Zacchetti [citing Vetter and Zacchetti 2004, 160], who suggests that this usage
reflects an older meaning of jing, i.e., ‘norm, standard.’ In either case, though, it’s clear that (whatever the ratio-
nale) jing, like jingfa (fa being the standard translation of Dharma in most other contexts), was often used in the
second and third centuries CE to translate ‘dharma.’” Nattier clarifies that it is difficult to identify the underlying
Indic term because the translator is anonymous, but adds that there are “some other items in this same verse
translation that make me think it is extremely likely that he was following this archaic practice of translating
dharma as jing,” citing a reference at 12a5 “to jingjie, ‘classics and precepts’ which looks, in the context, very much
like a rendition of the standard term dharma-vinaya. . . . So, though it’s impossible to prove this 100%, it does seem
most plausible to interpret jing here as Dharma.”
69. Nattier 1991, 166–68, 222–23, though as 222 n. 14 indicates, the reference to 250 rules could have
been added in China.
70. See Nattier 2003, 46 n. 80; 80–83, 129, 274–76; Pagel 1995, 7–36. Curiously, the Bahuśrutīyas also
had a bodhisattvapiṭaka, but in a canon of five baskets (Nattier 2003, 46 n. 80).
71. See Collins 1990 on the important point that among all Buddhist schools and traditions, only the
Theravāda sought to have a closed canon. Another illustration of the openness of the Dharmaguptaka canon is
found in the addition to its Prātimokṣa of twenty-six śaikṣa rules defining “appropriate conduct at a stūpa,” which
Prebish suggests would help to explain “the high status of its Vinaya in the development of Chinese Buddhism”
(1996, 270). This would likely also apply to his sect’s spread into central Asia.
72. See Nattier 2003; Williams 2000, 96–191; 2009, 5–7; Robinson et al. 2005, 43–123, with input from
Nattier (xiv).
 dharma over time, ii 305

of interest that the earliest texts to tell the Kauśāmbī myth in this way show
signs of telling it from a Bactrian perspective. According to Nattier, their likely
use of the Kharoṣṭhi script and the renaming of the three kings in the verse
Prophesy of Kātyāyana to include Rome and Iran suggest that this version of
the myth travelled to Bactria, where the Dharmaguptakas and perhaps the
Mahāsaṅghikas could have used it in their spread through Central Asia (Nattier
1991, 161, 222).
But Rome and Iran aside, the alpha-cluster would still point us to northwest
India for the original.

B.2. The S-cluster of the (Mūla-) Sarvāstivādins

I use the name S-cluster to bring together versions of the Kauśāmbī myth told
with similar contours in four texts, all of Sarvāstivādin (or Mūlasarvāstivādin)
provenance. In these texts, the myth would seem to have been composed and
transmitted in three different genres or formats—sūtra, avadāna,73 and
Abhidharma—within a loose but consistent conceptual framework supplied by
the relatively open canon(s) of these “Sarvāstivādin” schools—relative, that is,,
to the closed Theravāda canon and the still more open Dharmaguptaka can-
on.74 One version is included in the Saṃyukta Ā gama (Saṃyuktāgama), which
means that it has a place in the Sarvāstivādins’ first Basket, its Sūtra (= Pāli
Sutta) Piṭaka. In Northern Nikāya schools, the Saṃyuktāgama is the “equivalent”
of the Saṃyutta Nikāya in the Pāli canon, but the Saṃyutta Nikāya, like the Pāli
canon in toto (as mentioned above), knows nothing of the Kauśāmbī myth.
Another version yields two texts found in Chinese and Tibetan translations of
the Aśokāvadāna, itself probably also originally a Sarvāstivādin text geared to
attracting wider north Indian audiences (Strong 1983, 26–37). The section
telling the Kauśāmbī myth does not, however, appear to be an original feature
of this text, since it is not included in its first Chinese translation in 306 CE or
in the surviving Sanskrit Aśokāvadāna (Strong trans. [1983]; Nattier 1991, 151).
The third text, which seems to be an Abhidharmic digest of several known
versions, is found in the Mahāvibhāṣa,̄ where it seems to be a condensation of

73. On the avadāna genre as stressing reliance on the power of Buddhas in contrast to the jātakas’ emphasis
on self-reliance to become a Buddha, and the relation of both paradigms to early Mahāyāna Bodhisattva sūtras as
“a missing link in the story of the development of Indian Buddhism,” see Robinson et al. 2005, 77–78, 84–86;
cf. Strong 1983, 32–33 relating this aspect of the avadāna genre to bhakti; Rotman 2009, on its devotional
dimension in relation to Buddhist visual culture.
74. Cf. above, n. 71 on closed and open canons. See Strong 1983, 31 on the Sarvāstivādins’ particular will-
ingness “to add more materials to their ‘canon,’” including avadānas, and to write Buddhist texts in Sanskrit. Cf.
Lamotte 1988, 154: As the āgamas “were closed much later, they make room for works of a comparatively recent
date; hence the Saṃyukta . . . contains long extracts from the Aśokāvadāna.”
306 dharma

the myth from within the same school.75 The Kauśāmbī myth undergoes
Chinese translation in these versions first in the Saṃyuktāgama (443 CE), then
in an enlarged Aśokāvadāna (late fifth century), and then in the Mahāvibhāṣā
(656–59 CE) (Lamotte 1988, 201–2; Nattier 1991, 146, 150). This, of course,
cannot supply the order or dates of composition in India. As Nattier notes
regarding the Mahāvibhāṣa,̄ “the bulk of its content is generally assigned to
around the 2nd C. CE” (Nattier 1991, 146, citing Lamotte 1988, 217–18).
My approach to this cluster differs from Nattier’s. She calls the first three
of these texts “the Aśoka-avadāna group,” and leaves the Mahāvibhāṣā version
aside. After noting that it has some affinities with the other three versions,76
she views it as ultimately irrelevant to her stemma because, as an Abhidharma
compendium, it draws from more than one source.77 In selecting the name
“Aśoka-avadāna group,” Nattier would appear to be fortifying her view that an
Aśokāvadāna version would have priority over the Saṃyuktāgama version. But,
as I shall indicate, I think that point is tenuous.
Now the overall impression is that S-cluster versions have two compo-
nents: one reworked and one new. In the first case, the end of the True Dharma
still comes about amid events that include the death of the last arhat. Otherwise,
“the cast of characters is considerably expanded” (Nattier 1991, 152), as is the
plot, in the buildup and the denouement. In recounting the killing of the arhat,
there are new features, but the story discernible in the Prophesy of Kātyāyana
seems to be reduced, and in any case yields no suggestive innuendos that
Nikāya school–Mahāyāna tensions are among the symptoms of the end. Here
is Nattier’s summary of this portion from the Saṃyuktāgama and the two
Aśokāvadana translations:

When the monks gather for the poṣadha (Pāli uposatha) ceremony
[where the prātimokṣa should be recited], Śiṣyaka, the head of the
sangha, wants to recite the monastic rules in abbreviated form.
“If even I, who am the most learned of all, cannot keep the rules in
their entirety, who else can?” he argues. But the arhat Sūrata rises to
object. As in the Mahāvibhāṣā, Sūrata proclaims that he does keep
the monastic rules to the letter, and that Śiṣyaka should therefore

75. See Nattier 1991, 146–50, who I believe is right in underlining the different Abhidharma literary
processes behind it, including likely reference to varied versions that would mean it was not the earliest version,
as suggested by Lamotte (1988, 198, 200).
76. The Mahāvibhāṣā differs in having two invading kings and only one unnamed king of Kauśāmbī, who
could be the father or, I think, more likely the son or an elision of the distinction (Nattier 1991, 152); Nattier rec-
ognizes that the father’s death “is restricted to” the other texts of our S-cluster “and serves as one indicator of the
distinctive character of this textual family” (211).
77. Point for point, see Nattier 1991, 150–52, 185–86; 206, 214–15.
 dharma over time, ii 307

recite the rules in their entirety. A disciple of Śiṣyaka, named

Aṅgada, is outraged at this insult to his teacher, and strikes and kills
the arhat. Aṅgada is in turn killed by a yakṣa, Dadhimukha, and one
or more disciples of Sūrata then kill Śiṣyaka. Then the Buddha-
Dharma, the texts tell us, gradually disappears. (Nattier 1991, 154;
my brackets)

If the alpha-cluster is earlier, as I think remains convincing, it would appear

that any suggestion of Mahāyana allusions has been systematically removed!
What seems to have replaced them is an in-house Sarvāstivādin version cleared
of such implications. Indeed, Lamotte’s translation of the Mahāvibhāṣā’s telling
has a singular twist that would seem to reinforce this conclusion. Śiṣyaka is
now a “Tripiṭaka master,” and when he is asked to recite the Prātimokṣa in
public and wants to recite it in brief, Sūrata says,

“I would like the elder (sthavira) to recite the Tripiṭaka in full.”

Śiṣyaka will answer: “If there is [a monk] in the assembly who is
capable of observing all the precepts in the Prātimokṣa, let him invite
me to recite it in full!” The Arhat will respond: “I myself am capable
of observing the fine details (prāntikoṭi) of the ruling (śīkṣapada)
observed by the bhikṣus when the Buddha was alive; if that is what
you call being capable of observing [the Prātimokṣa] completely, then
I beg you to recite [the Tripiṭaka] in full.” (Lamotte 1988, 200)

Again, the arhat takes his stand on the Vinaya rules,78 but, most interestingly,
he seems to be challenging the “Tripiṭaka master’s” right to his title. In turn,
Śiṣyaka’s dodge seems to suggest that he does not know how to recite anything
in full. In any case, their two types of authority rest on the content of the Three
Baskets and no more than that. Śiṣyaka gives no hint that he would be inter-
ested in hearing some enlightened person’s “sūtra[s],” and is himself supposed
to be an expert in the Three Baskets only. As in the three other S-cluster texts,
the Mahāvibhāṣā knows, as one of three possibilities, “that it was the Arhat’s
disciples who killed the Tripiṭaka master in revenge” (Idem).79 Throughout the
S-cluster, then, the arhat now has his own murderous faction of monks, all pre-
sumably from the same “school” as Śiṣyaka’s, and things are ready to go from
bad to worse.

78. On arhats preserving Vinaya, see Lamotte 1988, 174–75. The introduction to the Mahāsaṅghika
Prātimokṣa speaks of the Prātimokṣa’s preservation as a measure of the duration of the Saddharma, while the
Mūlasarvāstivādin Prātimokṣa’s opening exalts itself as “the ‘compendium of the True Dharma written by the
King of the True Dharma’”! Such openings may be “additions” (Prebish 1975, 119–20 n. 18). The Theravāda
Pātimokkha lacks one.
79. The other two are that the gods, nāgas, and yakṣas did it, or the king did it.
308 dharma

As to the expansions in the run-up and denouement, the new elements in

the S-cluster have two main features: one is new touches that locate the account
in Indian society; and the other is that the narrative now has an epical quality
with some details possibly reminiscent of the Mahābhārata. These features
could of course coincide.
On the first point, one detail clearly links the myth with Brahmanical
society: Śiṣyaka is now the son of a Brahmin named Agnidatta. Other details
appear to be given without caste or religious specificity, which may have been
lost in translation or imply a mixed Brahmanical–Buddhist situation, or both.
Sūrata, for instance, is the son of “an elder named Sudhana,” whatever that
means.80 Five hundred “ministers” (often implying Brahmins) or “heads of
families” serve the king. The monastic community has become “impure and
lazy, . . . due in part to their being accustomed to receive large offerings.” And
a group of 500 laity “censures the monks for their misconduct and urges them
to bring their quarrels to an end.” In the Mahāvibhāṣa,̄ these are 500 “pious
householders” (gṛhapati) who indicate that they themselves have not been short
on largesse, so the bad turn must be the monks’ fault.81 This rebuke by the 500
laity could be a reminiscence of the quarrel between the monks at Kosambi
during the Buddha’s lifetime, for there, what finally drove the two factions to
seek reconciliation was that the “lay-devotees of Kosambī” backed up their cen-
sure by ceasing to show the monks honors or give them food “so that they
would go away, or return to the world, or propitiate the Blessed One” (Vinaya,
Mahāvagga 10.5.1–2; Rhys Davids and Oldenberg [1981] 1968, Part 2: 314). In
any case, the theme is consistent with an overall Vinaya concern that monks
keep a good face for the society that feeds and further maintains them. In the
Kauśāmbī myth, Śiṣyaka is a “leader of the saṅgha” whose shortcomings in
Vinaya represent a faction that has grown too content with royal and public lar-
gesse. Being the son of a Brahmin and leading such a saṅgha places the monks
on a par with what the dharmaśāstras say should be a unique entitlement of
Brahmins: to receive gifts. All this is consonant with Schopen’s picture of the
Mūlasarvāstivādin-vinaya being composed, like other surviving Vinayas as we
have them, in the early centuries of the common era by “profoundly conserva-
tive men” who were “preoccupied—if not obsessed—with avoiding any hint of
social criticism,” and “timid,” particularly in the Mūlasarvāstivādins’ case,
about their “public image” where it came to the revival of early Buddhist forest

80. In the Tibetan version of the Candragarbha Sūtra, Sudhana will be taken as a merchant name (“Good
Wealth”) (Nattier 1991, 246).
81. Nattier 1991, 153–54; Lamotte 1988, 199–200. Of these details, only the rebuke by the 500 laity is men-
tioned in the Mahāvibhāṣā.
 dharma over time, ii 309

and cremation ground practices by, and in the name of, Mahāyāna bodhi-
sattvas. Says Schopen, “This, moreover, is the Buddhism that succeeded in
India, that built and ran Sāñci, the monasteries at Taxila and Pitalkhora. It did
so, it seems, because it had learned how to write a loan contract and how to
‘properly’ behave” (2004, 96; 2006, 316, 345 quoted).
But it is in its epical sweep that the S-cluster is entirely innovative,
with new names and a more elaborate geo-temporal plot. In what is now
Buddhavacana, the Buddha himself prophesies:

One thousand years after the Buddha’s death, the Dharma will be
about to disappear. There will arise three evil kings, “Greek” in the
north, “Saka” in the south, and “Parthian” in the west, and according
to the Saṃyuktāgama, Tuṣāra or “Kushan” in the east.82 These kings
will invade India and persecute Buddhist believers and destroy
Buddhist monuments in their path. King Mahendrasena, ruling in
the east at Kauśāmbī, will have a son named Duṣprasaha, “Hard to
Bear.” That same night sons will also be born to five hundred “great
ministers,”83 and a rain of blood will fall. When Mahendrasena asks a
soothsayer about these omens, he will hear that Duṣprasaha will
grow up to conquer all of Jambudvīpa, killing many people. All this
will come to pass when Duṣprasaha conquers and kills the foreign
invaders. Meanwhile, Śiṣyaka and Surāta will be born in Paṭaliputra:
the one destined to be a great teacher; the other a great arhat. Upon
the death of Mahendrasena, Duṣprasaha, now king, is plunged into
depression and doubt, and asks the Buddhist monks when the
Dharma will be destroyed.84 Learning it will happen in twelve years,
he resolves to hold a great pañcavarṣa feast for that entire period.85
Duṣprasaha now invites all the monks of Jambudvīpa, but the monks
get used to all the royal largesse. In the Mahāvibhāṣā’s description

82. Nattier: a direction normally reserved for the “good king” of Kauśāmbī in other versions (1991, 153
n. 18).
83. This is Nattier’s translation of what the Chinese Saṃyuktāgama has. It finds a corroborating “ministers” in
the Tibetan translation of the Aśokāvadāna, while the Chinese translation of the latter yields “heads of families”; and
in the Tibetan translation of the Candragarbha Sūtra where the Chinese has “elders” (Nattier 1991, 178, 242, 249).
84. See Nattier 1991, 253 and n. 20: the variants raise some uncertainties here.
85. To this point, all events are from Nattier’s summary of the Saṃyuktāgama and Aśokāvadāna versions,
and, with the exception of the invasion by only two barbarian kings and the victorious Kauśāmbī king’s decision
to hold the feast for all the monks of Jambudvīpa, the details so far are not found in the Mahāvibhāṣā. As Nattier
notes, another difference between the latter and the other three accounts is that the Vibhāṣā mentions only one
Kauśāmbī king without naming him (1991, 152). She assumes that because it does not mention the father’s
death, the one king is the father, but that does not seem to be certain.
310 dharma

of their decadence, they gather “to discuss worldly matters

(lokadharma),” make loud confused cries, get “lazy,” sleep late, are
“devoid of reasoning and reflection,” “neglect the true teaching,” and
“no longer follow the practices” (Lamotte 1988, 199). Various
supernaturals are offended and five hundred laity warn the monks to
cease quarreling, all to no avail. When Duṣprasaha hears about the
uposatha-night massacre,86 he flies into a blind rage and begins to kill
more monks and destroy stūpas and monasteries. Finally the Buddha
tells the four World Regents (lokapālas), “The Dharma that you were
assigned to protect has died out.” (see Lamotte 1988, 198–200;
Nattier 1991, 153–54)

Clearly, this end of the Dharma has taken on a more epical coloring. I confine
myself to six features that may resonate particularly with the Mahābhārata.

1. The great turning point on which the end of the dharma hinges is now
a great royal festival, the pañcavarṣa (var. pañcavarṣika) or “quinquennial
assembly,” which is not mentioned in anything connected with the
alpha-cluster. Some things are known about it from the Chinese
pilgrim Hsüan-tsangs’s description of one held by the seventh-century
CE north Indian Buddhist emperor Harṣa of Kanauj (not far from
Kauśāmbī), and from the Sanskrit Aśokāvadāna itself. A pañcavarṣa-
performing king should give away everything accumulated in his
treasury over five previous years in an extravaganza of dāna or giving,
implicitly gambling away his kingdom and himself in expectation that
his wealth and kingdom will be redeemed by further donations made
by his vassals.87 In the Sanskrit Aśokāvadāna, it figures as a turning
point in Aśoka’s tortured rivalry with his son Kunāla (see Strong 1983,
92–96, 264–68), which, as I suggested elsewhere (Hiltebeitel 2001a,
262–63), holds echoes of the Mahābhārata ’s downward turning point
for dharma: the betting at the epic dice match. If the Kauśāmbī story
now has a pañcavarṣa backdrop, it looks disquieting indeed. What
could it mean that King Duṣprasaha intends to stretch such a pageant
to the twelve years predicted for the end of the dharma? Is he gambling
that the True Dharma will somehow not end? That his munificence
toward the monks will somehow help to reverse the prophesy? The
True Dharma appears to be what is at stake, and the dialogue between
Śiṣyaka and Sūrata is as cagey and ominous as that between Aśoka and

86. “That very night, as the uposatha is being celebrated in the monastery, . . .” (Lamotte 1988, 200).
87. On the ritual by Harsha, see Strong 1983, 92–94 with citations, including Lamotte 1988, 60.
 dharma over time, ii 311

Kunāla or the gambling protagonists in the Mahābhārata.88 Are the

500 great ministers born on the same night as Duṣprasaha to a rain
of blood the vassals who should ransom him after all his giving?
2. The formation of a name like Duṣprasaha (“Hard to Bear”) looks suspi-
ciously reminiscent of the way the Mahābhārata names the 100 wicked
Kauravas, brothers of Duryodhana (“Hard to Fight”), many of whom have
names with the same “hard to”/“bad” prefix,89 and all of whom are born
within a month of Duryodhana, who is born to the sound of jackals (Mbh
1.107.24–37). It would help to know more about Duṣprasaha’s melan-
choly and doubt after his father’s death, which comes before he learns
about the twelve years left for the dharma; and about his final paroxysm
of murderous rage that follows the nighttime massacre that brings the
twelve years to their predicted climax. It may correspond to what follows
the final paroxysm of giving that we see in Harsha’s pancavarṣa, where
Harsha dons a worn and common monastic robe and goes “to worship
before the Buddhas of the ten countries” (Strong 1983, 93)—but by
inversion: rather than become a monk, Duṣprasaha joins in their
already-begun slaughter. The suggestion is made with the Mahābhārata
dice match in mind, which, as part of a similar royal ceremony, the Royal
Consecration or Rājasūya, the king should win rather than lose, as he
does in the Mahābhārata story (see van Buitenen 1972).
3. An epic sweep can also be noted in the new chronology. Although
a 1,000-year timetable for the True Dharma seems to enter the
Kauśāmbī myth from its earliest Northern Nikāya tellings on, including
the Prophesy of Kātyāyana in verse, it appears that a twelve-year hinge
period at the end of the 1,000 years is an S-cluster innovation.90

88. After Aśoka announces he will give vast wealth to the saṅgha and bathe the Bodhi Tree, Kunāla ges-
tures that he will double the amount, forcing his father to triple his original offer, which Kunāla quadruples, until
“Aśoka, retaining only the state treasury, makes a total gift to the sangha of his whole kingdom, his harem, his
ministers, his self, and his son Kunāla. . . . Kunāla, finding himself part of his father’s gift, cannot very well outdo
him” (Strong 1983, 95–96). The Mahābhārata parallel envisions a brahmanical king as imperial. Cf. Inden 2006,
91–94, contrasting this imperial mode of Buddhist ceremonial with the preference for Śrauta rites in regional
states, by which regional rulers could “declare their independence by again performing the horse sacrifice and
other Śrauta rites that had been in abeyance,” and gifting land to Brahmins at the end of these rites, “when one
of these imperial kingdoms contracted”—as Puṣyamitra Śuṅga, the Sātavāhanas, and Guptas did (94).
89. In the Mahābhārata, Duṣprasaha is a name for Garuḍa, the king of birds in his role as devourer of
snakes (1.20.11). The forms duṣ- and dur- are equivalent formations of dus- governed by rules of consonant
combination, all meaning “hard to X,” “bad X.”
90. See Nattier 1991, 45 on the Prophesy of Kātyāyana’s distinctive division of the 1,000 years into three 300
years “supplemented (apparently) by a final segment of a hundred years.” This is, of course a summary, but it does not
seem to involve an extra twelve years. On the 1,000-year timetable and the Nikāya schools’ texts more generally, see
62–64, 72. As Nattier argues, the 1,000-year period by itself looks like a reinterpretation of “the original figure of five
hundred years to refer to the duration of the ‘True Dharma’” attendant on the admission of women to the order, and
probably reflects a “crisis of historical self-consciousness” in that the initial 500 years were passing (1991, 62).
312 dharma

The twelve-year hinge suggests a nod to the Brahmanical yuga system.

Although Buddhists probably had not yet begun to refer to this system,
the S-cluster would be familiar with its early usages in the Sanskrit
epics and Manu. It is built on a numerology of 1,000-year units and
multiples thereof, with dawns and twilights one-tenth their length, that
gives the total of 12,000 years, mentioned in Manu,91 for one complete
mahāyuga cycle, the smallest nonfractional division of which is a
twelve-year period. In the Mahābhārata, that is what one finds
stipulated for the Pāṇḍavas’ period of exile once they have lost the dice
match, which itself marks a nadir in the decline of the dharma; on the
Pāṇḍavas’ fulfillment of that twelve years plus a thirteenth year (twelve
months) incognito, the restoration of the dharma hinges. Indeed, as is
widely recognized, the epic’s dice match narrative plays on a correlation
between the names of the yugas and the names of the four throws of the
dice. What we seem to have in the S-cluster narratives of a 1,000-year
period plus (or in some way including) a twelve-year hinge is an allusion
to such ideas without acknowledging them. According to Nattier (1991, 115
n. 130), Indian Buddhists do not begin using the classical Brahmanical
yuga structure until the Gupta period, when they start doing so as a kind
of concession to prevalent Hindu usages (see chapter 6 § A). Without
naming yugas as such, the S-cluster’s 1,012-year chronology already
looks to be intermediary between a Brahmanical view and a more
representative Buddhist one, which would hold, philosophically, that
“the ‘decline of the Dharma’ refers not to the decay of the order of the
universe as such, but primarily to the fading away and eventual disap-
pearance of the teachings of a specific historical figure” (Nattier 1991,
281). While the arhat still accounts for what is primary, the decaying
order of the universe seems to have stepped up its role.
4. The S-cluster’s epical sweep fills out in its geography. But on this topic
I must open with an uncertainty. The term madhyadeśa, the “Middle
Region,” does not appear in available translations or summaries of
S-cluster versions of the Kauśāmbī myth. Yet it is a featured term in both
the earlier alpha-cluster and the later M-cluster.92 I must, however,

91. With computational conciseness, M 1.69–70 reads, “The Kṛta yuga is said to last 4,000 years. It is
preceded by a twilight lasting 400 years and followed by a twilight of the same length. For each of the three
subsequent Ages, as also for the twilights that precede and follow them, the first number of the thousands and
the hundreds is progressively diminished by one.”
92. See Nattier 1991, 158, 160, 166–67 (both versions of the Prophesy of Kātyāyana); 178, 196, 242
(M-cluster, though only in its Tibetan and Khotanese versions). Nattier does not index Madhyadeśa, and does
not include it as one of the fourteen key “plot components” by which she detects differences in texts and
 dharma over time, ii 313

maintain that whether it is named in S-cluster texts or not, Madhyadeśa

is an important idea to keep in mind in discussing them since it is a
famous and indeed prestigious name for the region where Kauśāmbī lies.
In the epics and Manu, Madhyadeśa is a name for the Vedic heartland
of dharma. According to Manu 2.22–21, it encompasses the Land of
Brahmin Seers (Brahmarṣideśa) that is comprised of Kurukṣetra and the
lands of the Matsyas, Pañcālas, and Śūrasenakas, and extends only as far
east as Prayāga at the confluence of the Yamunā and Gaṅgā Rivers.93 Not
surprisingly, Buddhists seem to have had a different idea of the “Middle
Region” as extending well eastward beyond Prayāga. In Pāli texts, at least,
the “Middle Region” or Majjhimadesa is the “country of Central India
which was the birthplace of Buddhism and the region of its early
activities,” and “contained fourteen of the sixteen Mahājanapadas, that is
to say, all but Gandhāra and Kamboja, which belonged to the
Uttarāpatha” (Malalasekera [1937] 1983, 2: 418–19). That is, it included
both Magadha and Aṅga east of Prayāga (484).94 Kauśāmbī, which is a
little upriver west of Prayāga on the northern bank of the Yamunā, is thus
within the Gaṅgā–Yamunā doab, which we may take as the contested real
heartland of both Madhyadeśas. Thus if twelve years is now the dharma’s
hinge of time, Madhyadeśa—especially as so variously conceived—has
for some time now been its pivotal place. Yet if we begin with the
alpha-cluster, we begin with an incongruity, since the alpha-cluster allows
us to envision an original Kauśāmbī myth centered on a “Middle
Region,” with Kauśāmbī in it, that is not in the center but the east, with
invading Greek, Śaka, and Parthian kings having lodged themselves
respectively in the north, south, and west. The S-cluster removes this
incongruity by giving Kauśāmbī an actual “Middle Region” position, even
if it apparently does not give it a Madhyadeśa location by name. Indeed,
it would be locating the births of Śiṣyaka and Sūrata in Paṭaliputra, the
former capital of Aśoka, within the Buddhist Madhyadeśa.

text-groups (210). It is also not mentioned in Lamotte’s translation of the Mahāvibhāṣā account. The S-cluster’s
redactors may have left it unmentioned because they were from that area themselves. If so, my very tentative
hypotheses would be that centers get more interesting when described from afar (as would be the case with the
Bactrian perspective of the α-cluster). See Turner 1973 on “the center out there.”
93. Manu includes it within Āryavarta, which covers north India from ocean to ocean. See chapter 5 § B and
n. 33, and chapter 4 § B.1.d.i on Kuru country in the Pāli canon. Cf. Macdonell and Keith [1912] 1967, 2: 125–27.
94. Majjhimadesa “extended in the east to the town of Kajaṅgala, beyond which was Mahāsāla”
(Malalasekera [1937] 1983, 418); these towns are less easy to locate (see Idem, 1: 481–82 on Kajaṅgala). Cf. T. W.
Rhys Davids 1904; Parasher-Sen 2006, 439–40 on the Pāli Majjhimadesa as ārya versus other regions as milak-
kha (= Skt. mleccha) and “said to abound in ignorant peoples.”
314 dharma

5. Here Nattier has given us the key to understanding the alpha-

cluster’s geographical orientation by alerting us to the likely Bactrian
perspective and “outpost mentality” (1991, 162) of the Prophesy
of Kātyāyana. In contrast, the mentality behind the S-cluster is that
of the predominant conservative Nikāya school(s) of central north India,
that is, of Madhyadeśa itself: the Mūlasarvāstivādins and Sarvāstivādins.
Indeed, one should notice that the one sure sign of Brahmanical
impingement now comes from this Buddhist Madhyadeśa’s further
eastern domain: Śiṣyaka is the Magadha-born son of a Brahmin named
Agnidatta, “Given by Agni,” the Brahmanical god of fire. Whatever one
makes of this likely sign of trouble, Śiṣyaka’s dharma-ending joust with
Sūrata will be between two monks from Magadha.
6. The Saṃyuktāgama mentions Tuṣāra or “Kushan” as a fourth invading
king. In this text and this one alone, “Kushan” will rule in the east,
seemingly overlapping the rule of King Mahendrasena in Kauśāmbī,
also in the east, until Mahendrasena’s son, prince Duṣprasaha, kills
all the invading kings. For Nattier, “The presence of this obvious
interpolation provides one piece of evidence that the Saṃyukta-āgama
version is not the oldest version of the story” (1991, 152 n. 16). Agreed,
a story with the three kings is older. But this text places this detail in
the Buddha’s most prestigious canonical discourses or sūtras.95 I take it
as the tip of an iceberg.

My view is that the Kauśāmbī myth begins in the alpha-cluster as a story

centered on northwest India reflecting the impact of incursions there first by
Greeks, Śakas, and Parthians, all of whom are also mentioned in the Mahābhārata
and Manu. However, since these kings not only invade the northwest but have
also located themselves in the north, south, and west of India, the myth prob-
ably already reflects a geography of Kuṣāṇa rule in the area defined by those
boundaries, and perhaps not yet extending into a Madhyadeśa inclusive of the
east.96 The Saṃyuktāgama version, however, in interpolating the fourth king

95. See Chappell 1980, 128: “probably the oldest version of the story,” on which Nattier comments, “his
reasoning may be that since the āgamas are among the oldest layers of Buddhist canonical literature,” it would
likely “be a relatively ancient version of the tradition” (1991, 150 and n. 10). We cannot say whether the
Saṃyuktāgama version made this alteration before or after the Aśokavādāna versions, which keep things to three
kings; it could still be the oldest text in the S-cluster. In any case, although the avadāna and Mahāvibhāṣā versions
also cast the myth as the word of the Buddha, a version in the Saṃyuktāgama, where it is placed amid the
Buddha’s most prestigious canonical discourses or sūtras, would certainly be the most prestigious version, and
thus call for interpretation on this striking point of difference.
96. As Nattier says, “archetype α—which is the only version I am suggesting was produced under Kushan
rule—clearly contained a list of three kings” (1991, 227 n. 25).
 dharma over time, ii 315

and in placing him at the pivotal east (along with Kausāmbī), must also have a
Kuṣāṇa setting or familiarity with Kuṣāṇa rule. Otherwise, why the addition?
The Kuṣāṇas’ southern capital in Madhyadeśa was at Mathurā. Moreover, as
Nattier says, “the addition of the Kushans . . . in the Saṃyukta-āgama recension
does not in itself require a later date than that of the other texts in this group”
since “recent research” points to the first-century CE for “the earliest Kushan
incursions.”97 Yet one must separate the time when the Kuṣāṇas could be men-
tioned along with the Greeks, Śakas, and Parthians as invaders from the time
they actually ruled from northwest to central India, including Madhyadeśa.
Whereas the alpha-cluster clearly refers only to the earlier invading phase, yet
probably also reflects the second phase of Kuṣāṇa rule as well, the S-cluster
reconfigures the geography in a way that clearly reflects this second phase. The
S-cluster (excepting the Mahāvibhāṣā) shifts the perspective by placing
Kauśāmbī in the center and, in the Saṃyuktāgama, further clarifies that this
area is now central precisely because the Kuṣāṇas have made it so. How are we
to interpret these differences, and how close to each other are these two text
clusters in time? Nattier has set up the problem in an interesting way by pos-
iting that the original archetype α comes before our S-cluster and by interpret-
ing the Kauśāmbī myth against the background of Kuṣāṇa rule. But one or more
S-cluster texts may be closer to her archetype α in time than she suggests98—
indeed, nearly as close to it in time as the earliest alpha-cluster text, the Prophesy
of Kātyāyana in verse. I believe we should see a Kuṣāṇa-rule setting for both the
hypothetical original and the S-cluster. Given the difficulties scholars have had
in dating the Kuṣāṇa emperor Kanishka, this Kuṣāṇa period may be tentatively
assigned from his reign in the “late first or early second century C.E.”99 through
the mid-third century. With these considerations, we may turn to Nattier’s
revealing discussion of the parameters for interpreting what Kuṣāṇa rule
means for the Kauśāmbī myth.
Among the things Nattier finds to be noteworthy in our S-cluster (she
includes the Mahāvibhāṣā in her overview) are that it includes a “lineup of evil,
non-Buddhist kings”—Greeks, Śakas, and Parthians (and the Kushans as well,

97. Nattier 1991, 152 n. 17, disagreeing here with Lamotte 1988, 201–2. Cf. 225: “the best current estimate
is that the Parthians—the last of these three groups to arrive in the scene—invaded northwest India by the first
half of the first century C.E.” In both cases she cites Frye 1983, 177–204.
98. Nattier does not rule anything out. She does not contest Lamotte’s dating of “the bulk” of the
Mahāvibhāṣā to the second-century CE (cited above), and says our S-cluster version must have emerged “during
the period from the 2nd–4th c. CE” (152–53). But her outlook is clear in the way she proportions implied time in
her flow chart. There, the two Prophesy of Kātyāyana texts would be older than the hyperarchetype from which our
S-cluster derives; and the S-cluster texts (with the Aśokāvadāna’s version older than the Saṃyuktāgama’s) would
be contemporary with the hyperarchetype for our M-cluster (1991, 216).
99. See, lately, Robinson et al. 2005, 76.
316 dharma

in the Saṃyukāgama), as ‘enemies of the Dharma’” (154–55). In the Mahāvibhāṣā,

where we are down to two kings without their countries being named, “The two
irreligious kings will be born among foreign slaves (Dasyumleccha); they will be
stupid and will hate and despise the Law of the Buddha.” The “religious king”
of Kauśāmbī “will finally capture the two bad kings” and put them to death
(Lamotte 1988, 198–99). Now as Nattier observes, it is “peculiar . . . from the
point of view of the Buddhist history of the region” that these peoples and
kings, all of whom had exemplars (including Kanishka) who took up Buddhism,
should be portrayed as “enemies of the faith” (155). Asking why this is the case,
she proposes “two possible explanations.” Option one is that the invaders, pre-
sumably as yet with no Buddhists among them, caused devastation to Buddhists
along their routes that would have left “a negative initial impression among
Indian Buddhists” that “persisted long after many” of the invaders’ “descen-
dants had adopted the Buddhist faith.” Option two is that “by the 2nd century
CE (if not before) the trio of Greeks, Śakas, and Parthians had become a stan-
dard topos for ‘non-brahmanic barbarians’ in Indian literature” (155–56)—an
explanation that she has already hinted will be linked with “the experience of
excessive ease and comfort” (117).
Given what has been said above, that even the alpha-cluster reflects not
only the invasion by these peoples but their settling in and around India,
I think that both explanations hold true for the alpha-cluster, and thus as well
for the S-cluster. But whereas the alpha-cluster may harbor more of the first
imprint, the S-cluster shows more of the second, and makes it more inter-
esting. This is not because the second set of factors calls for a later dating, for
as Nattier admits, “the formulation of a tradition referring to these three
powers may well date from as early as the 2nd century CE” (152). It is because
such a dating is well within the Kuṣāṇa period that the alpha- and S-clusters
both seem to reflect. Moreover, the Saṃyuktāgama’s interpolation of a wicked
Kuṣāṇa king means that it extends this “barbarian” topos to the Kuṣāṇas not
only beyond other Buddhist versions of the Kauśāmbī myth but beyond the
Sanskrit epics and Manu, which are liberal in applying it to Greeks, Śakas, and
Parthians but never mention the Kuṣāṇas.100 The S-cluster also seems to be
more aggressive in applying this topos than the alpha-cluster, which, if it
reflects a Bactrian outpost mentality, may not wish to apply it at all. Note also
that it differs too from the more redemptive telos of the Yuga Purāṇa, which,
while mentioning only the Greeks and Sakas and stopping before both the

100. For discussion, see Witzel 2006, 482–84 on Manu and the epics; passim on the Kuṣāṇas. To my
mind, the epics and Manu are all pre-Kuṣāṇa. Cf., however, chapter 5 § C on Olivelle’s later dating.
 dharma over time, ii 317

Parthians and Kuṣāṇas, knows foreigners as mlecchas but sees them as poten-
tial allies in history and does not personalize them through “evil” (or other-
wise) kings.
With option two, we are thus back above all to the (Mūla-) Sarvāstivādins
and the place of the S-cluster within its wider world. And the first thing to
notice is that if this version of the Kauśāmbī myth is especially aggressive in
adopting “a standard topos for ‘non-brahmanic barbarians,’” it does so bow-
ing to a historically and socially identifiable Brahmanical usage. Our task
would be to interpret the S-cluster’s nods to Brahmanical impingement,
traces of which (including epical innovations) we have already noted, within
the wider sociohistorical context of what Nattier calls the pax kushanica
(1991, 227). Nattier comes to the flourishing of Buddhism under the Kushan
rule and the supposed conversion of Kanishka with this question: “What,
then, would have led the Buddhist subjects of such a cosmopolitan realm,
presumably enjoying all the material and spiritual benefits afforded by the
long-lasting pax kushanica, to produce such a prophesy of decline?” (226–27).
Her answer, in accord with option two, is that “the Kauśāmbī prophesy
makes good sense as the product of a Kushan environment . . . in a period of
post-invasion prosperity.” But her evidence is sketchy: the myth resolves
matters to internal sources of decline rather than external ones; “the death
of the dharma” is brought on not by “the upheavals of war” but by “the
munificence of a well-intentioned Buddhist king” (227). This barely accounts
for her hypothetical archetype α, which she suggests is her intention (227 n.
25). Though she need not have done so (the one king of Kauśāmbī in the
alpha-cluster goes unnamed), she mentions Duṣprasaha as the munificent
king in question, who makes his debut in the S-cluster: a pious Buddhist
king to be sure, but also with numerous Brahmanical (including epical) trap-
pings. Most important, Nattier does not look beyond the “standard topos for
‘non-brahmanic barbarians’” to the wider configuration of which it is only a
part: a Brahmanical mixed ethnic, mixed caste, and mixed religious (but not
in any visible way mixed sectarian, in Buddhist terms) packaging that makes
the S-cluster’s versions of the Kauśāmbī myth intelligible. Along with its
reorientation of the myth to a “Middle Region,” this repackaging explains
how the S-cluster, probably close on the heels of the alpha-cluster, makes
full sense—that is, Brahmanical sense—of its Kushan environment by bar-
barizing the Kuṣāṇas and their Greek, Śaka, and Parthian predecessors. In
agreeing here with Brahmans in the naming of foreign “others,” central
north and northwest Indian Nikāya school Buddhists offered an ironic indi-
cation that the myth of the end of the dharma was finding its own prophetic
confirmation.
318 dharma

B.3. The Mahāyāna M-cluster and the Chinese Candragarbha Sūtra

One more cluster beckons. The M-cluster is clearly the main Mahāyāna retelling
of the Kauśāmbī myth. Yet although it is as varied as the other two, there are rea-
sons to focus mainly on one text. The M-cluster’s main destiny is in China and
East Asia, though it also reached Tibet and Khotan. In East Asia its prophesy gets
tied to a system of three periods of the dharma, which Nattier shows is not origi-
nally Indian: those of the True Dharma (saddharma), the Semblance Dharma or
Reflection of the True Dharma (saddharma-pratirūpaka), and the Final Dharma
(Chinese mo-fa; Japanese mappō )—“a term for which no proper Sanskrit equivalent
exists.”101 The Chinese translation just mentioned is of the Candragarbha Sūtra
(henceforth Chinese CS), and I will limit my discussion mainly to it and invoke a
portion of one more text in Chinese translation that can be studied with it, the
“Mahāyāna Mahāparinirvānạ Sūtra” (also called the Nirvāṇa Sūtra).
The M-cluster is later than the other two, and probably comes from the
Gupta period. It is harder to discern the social, institutional, and historical
bases of M-cluster texts in India, as it is for pre-Gupta Mahāyāna Buddhism in
general.102 I will keep to three points that round out our discussion of the other
two clusters: (a) the M-cluster’s timetable and treatment of the Semblance
Dharma; (b) the relation between the core of the Chinese CS and its frame; and
(c) the Mahāyāna rehandling of the Kauśāmbī narrative.

b.3.a. the timetable and the semblance dharma. Taking our point of
departure from its timetable, the Chinese CS is one of several texts that move
the end of the dharma along from 1,000 to one 1,500 years. The Buddha
prophesies that his True Dharma will last 500 years and his Semblance Dharma
will then last 1,000 (Nattier 1991, 48–49). Thousand-year timetables are found
in both Mahāyāna and Nikāya school texts but typify the latter. In contrast,
schedules of 1,500, 2,000, and 2,500 years are found almost exclusively in the
Mahāyāna.103 Yet the True Dharma/Semblance Dharma distinction comes
mainly with 1,000 and 1,500 year timetables.104

101. Nattier 1991, 66, and for fuller discussion, 65–118 for her chapter 4, “The East Asian Tripartite System.”
See also Nattier 2008a, 158; Chappell 1980, 133–35, preferring “Imitation Dharma” for pratirūpaka-dharma.
102. See Schopen 2005, 268–69: “even after its initial appearance in the public domain in the second
century it appears to have remained an extremely limited minority movement . . . that attracted absolutely no
documented public or popular support for at least two more centuries. . . . It suggests that, although there was—as
we know from Chinese translations—a large and early Mahāyāna literature, there was no early organized,
independent, publicly supported movement to which it could have belonged” (author’s italics).
103. A 5,000-year timetable, mentioned above, is found in a fifteenth-century Pāli chronicle from northern
Thailand (Veidlinger 2006, 22–23); and a 2,500-year span is offered by the Theravāda scholiast Buddhaghosa.
104. I say “mainly” because, if I understand Chappell 1980, 139–40 correctly, Nattier seems to overlook an
exception where the distinction is mentioned in a text with a 700-year timetable, the Mahāyāna Mahāparinirvāṇa
Sūtra. See Nattier 1991, 27–29 on this text and 42–58 for wider discussion on these topics.
 dharma over time, ii 319

This distinction itself is also made almost exclusively in Mahāyāna texts,105

and Nattier says “the origins of the term saddharma-pratirūpaka and its pairing
with the term saddharma should be sought in Mahāyāna circles” (1991, 47 n. 55).
But the matter is complicated. I would rather say for now that we can certainly
assume that usage of this distinction would be one of the important elements
in the Mahāyāna transformation of the Kauśāmbī myth in the Chinese CS. Of
Chinese translations that know the Kauśāmbī story, it is the only one to make
this distinction (Nattier 1991, 37–39).106 In this text, the Buddha divides the
1,500 years as follows:

After my nirvāṇa, the True Dharma (saddharma) will remain in the

world for five hundred years. Sentient beings will extinguish the
kleśas, and the bodhisattva with vīrya will achieve fulfillment in the six
pāramitās. Those who practice [the Dharma] will be able to enter
quickly into the peaceful city [of nirvāṇa] which is without āśravas.
And the saddharma-pratirūpaka will remain in the world for a full one
thousand years.107 (Nattier 1991, 48–49; cf. 85)

The “demise of the Dharma” at Kauśāmbī will occur at the end of this second
“considerably less auspicious” period (Idem, 49).
Now if the True Dharma/Semblance Dharma distinction is largely a
Mahāyāna one, what is it? For one thing, it belongs mainly among what
Nattier calls the “‘middle-length’ formulations of 1,000 and 1,500 years,”
which, we may add, belong historically to the “Middle Period” of Indian
Buddhism with its extensions into China—and it would figure to be the
Mahāyāna that is grabbing the extra 500 years.108 Beyond that, as Nattier
richly demonstrates, it is not so easy to say what the distinction is about, espe-
cially since we must dispense with many “commonly held assumptions”
about it, above all that in contrast to the True Dharma, saddharma-pratirūpaka
means a “counterfeit Dharma,” as it has been widely translated (67), implying
“a fake or a forgery” (86). As Nattier illustrates from the Tibetan translation
of the Candragarbha Sūtra, Mahāyāna texts may initially use the two terms as

105. Once one gets to 2,000 and 2,500 years, “this distinction disappears from view” (Nattier 1991, 51).
106. See further Idem, on the three versions of the Chinese Mahāparinirvāṇa Sūtra, with a 700-year
timetable; and 50–51 on the Mahāmāyā Sūtra, with a 1,500-year timetable, but without mentioning the
distinction.
107. Kleśas are “taints,” “hindrances,” or “passions,” the five being sensual desire, ill-will, sloth-and-torpor,
worry-and-flury, and doubt; vīrya is “effort,” one of the six pāramitās or “perfections” of the bodhisattva, and one
of the concentration group within the eightfold path; the āśravas are the four “binding effluents or pollutions of
the mind” (Robinson et al. 2005, 9–10): sensual desire, becoming, views, and ignorance.
108. See Williams 2009, 269–70, n. 12 on Indian Mahāyāna in its early centuries as a Buddhism more
successful for export than for home consumption. See also Schopen 2005, 3–24.
320 dharma

a distinction without a difference. In this instance, the new interlocutor-

bodhisattva of the title, Candragarbha,109 asks the Buddha,

If after the nirvāṇa of the previous Buddha, the reflection of the True
dharma (saddharma-pratirūpaka) [i.e., the Semblance Dharma]
disappeared after [just] seven years had passed, how long will the
True Dharma (saddharma) endure after the Lord Śākyamuni has
attained nirvāṇa?

To which the Buddha replies,

After I have attained nirvāṇa, the reflection of the True Dharma

(saddharma-pratirūpaka) [i.e., the Semblance Dharma] will endure
for two thousand years. (Nattier 1991, 77)

Asked about one thing the Buddha answers about the other.110 Nattier com-
ments, “in this text the two terms are, for all practical purposes, interchange-
able”; rather than being contrasted with a True Dharma, saddharma-pratirūpaka
refers to “the period of time during which the True Dharma will survive.
It . . . is an indication of its very presence” (78). Nattier calls these virtu-
ally undifferentiable usages “inclusive,” and presents evidence that
“exclusive” or contrastive ones with a “two-period system” reflect “a gradual
transition” (82). If so, this would indicate that the Chinese CS could speak
from the earliest Mahāyāna meaning.
As to the Chinese CS and the other texts cited by Nattier that do have contras-
tive usages for two periods of time, Nattier sees them as having “ ‘adjusted’ to con-
form with new developments in the understanding of saddharma-pratirūpaka.”
Yet the chronological evidence is thin here. Based on the Tibetan translation of the
Candragarbha Sūtra where “this twofold division does not appear” (85), Nattier
takes the Chinese CS, translated in 566 CE, to have done the adjusting—even
though the Tibetan translation was done several centuries later.111 I believe Nattier
also tries to improve matters when she goes on to say that, “Even when a clear-cut
distinction between saddharma and saddharma-pratirūpaka begins to emerge, the
latter period is still viewed as positive (if slightly less so than in the preceding
versions)” (86–87). This is the same Semblance Dharma in the Chinese CS that

109. He of course has this role likewise in the Chinese CS; see Nattier 1991, 174–75 and 219 (traced back
to a hyperarchetype). Nattier considers him among “minor Mahāyāna elements” of the frame.
110. Nattier recognizes that the passage could be read as making a consistent distinction between a True
Dharma operative during a Buddha’s lifetime and a reflection of it that survives him. But as she says, such an
“intuitively satisfying” equation has no known textual support (1991, 77 n. 48).
111. Nattier has used the Tibetan version’s longer timetable, which she usually takes to indicate that a text
is later than one with a shorter timetable, to explain the absence of a temporal distinction.
 dharma over time, ii 321

she earlier called a “considerably less auspicious” period with reference to its
preceding the end of the dharma at Kauśāmbī (49, as cited above).
No doubt the Semblance Dharma has had changing and polysemous refer-
ents, among them, as we shall see in the Chinese CS, deteriorating times for
the dharma. But I am not sure that it would have changed meanings in just one
direction.112 I would like to explore a possibility that one of the contrastive
implications of the Semblance Dharma in the Chinese CS would be an intra-
Buddhist sectarian one. The first and most important reason to suspect this is
that it would carry along (and, we would have a right to expect, reverse) the sec-
tarian implications we found in the alpha-cluster. It would thus be not so much
a general meaning as a text-specific one tied to the M-cluster’s rehandling of
the Kauśāmbī myth.
The second reason, though, is a more powerful one. This is that there
seems to be a distant but still pertinent precedent for a sectarian implication in
the one exceptional usage of the equivalent terms saddhamma and saddhamma-
paṭirūpaka in the Pāli canon. This occurs in the Kassapasaṃyutta section of the
Saṃyutta Nikāya,113 where the Buddha tells his disciple Kassapa:

There is no disappearing of the True Dharma (saddhamma), Kassapa,

till a Counterfeit Dharma (saddhamma-paṭirūpaka) arises in the
world; but when a counterfeit doctrine does arise, then there is a
disappearance of the true doctrine. Just as there is no disappearance
of gold so long as there is no counterfeit gold in the world. (slightly
modifying Rhys Davids trans. [1917–30] 1972, 2: 152; see Nattier
1991, 87; cf. Bodhi 2000, 680–81)

As Nattier says, Counterfeit Dharma (Rhys Davids and she have “counter-
feit doctrine”) is clearly a correct translation here; and, as we have seen, she
demonstrates convincingly that it has been misleadingly adopted in translating
Mahāyāna texts. But she drops tangential hints at something more.114 Making

112. As with the “intuitively satisfying” but unattested solution noted above, there would seem to be
something shadowy in pratirūpaka, as borne out in the Tibetan term that translates it, which literally means “bor-
rowed from” and “carries the connotation of ‘reflection,’ ‘image,’ or ‘shadow’” (Nattier 1991, 88).
113. This text opens with Mahākassapa asking the Buddha a loaded question: “Venerable sir, what is the
reason, what is the cause, why formerly were there fewer training rules but more bhikkhus were established in
final knowledge, while now there are more training rules but fewer bhikkhus are established in final knowledge?”
(Bodhi 2000, 680–81). The “now,” of course, begs the historical question of the date of composition, for it pre-
supposes, during the Buddha’s lifetime, an anticipation of the “future” two periods. Nattier 1991, 121–22 also
cites this text’s mention of “five lowering dharmas”/Bodhi: “five detrimental things” (2005, 681) that lead to “the
corruption and disappearance of the True Dharma,” the fourth being when monks, nuns, laymen, and laywomen
are “irreverent and unruly toward the training (sikkhā).” The same question gets a different reply, one more
focused on the training rules, at MN 65 (see Ñāṇamoli and Bodhi 2005, 548).
114. One must gather the following points from footnotes: Nattier 1991, 68–69 n. 9; 87. n. 78; and 88 n. 83.
322 dharma

a suggestive case that such a translation might also be right for “the sole Vinaya
text in which it occurs”—referring to the Sarvāstivādin Vinaya in Chinese trans-
lation (68–69 n. 9), she finds it “intriguing” that “the sense of ‘counterfeit’ is
restricted to Nikāya Buddhist (i.e., non-Mahāyāna) texts . . . in light of the fact
that some we have examined use the term in reference to specific (Mahāyāna)
Buddhist scriptures. From the perspective of the Nikāya Buddhists, of course,
such Mahāyāna scriptures did indeed represent a counterfeit of the True
Dharma.”115 Further, she admits, “It would be interesting to speculate as to
what the authors of this passage understood as to the ‘counterfeit’ Dharma in
question” (124; cf. 120).
That is indeed a deep and fascinating matter, and both the Pāli commen-
tarial tradition and the Theravāda scholarly tradition have offered an answer.
Drawing on Buddhaghosa’s distinction regarding the first period of attainment
when there were numerous arhats versus a second period of learning, each period
is said to have its counterfeit of the True Dharma (saddhamma-paṭirūpaka). In
the first period, the “counterfeit” is the ten corruptions of insight. In the second,
it is “texts other than the authentic Word of the Buddha authorized at the three
Buddhist councils,” with some noted exceptions. Eight such counterfeit texts
are mentioned. Three of these include the descriptive term “Secret” (guḷha) in
their titles; three are extra-canonical “Baskets” (piṭaka); and the other two of are
named after someone’s cloud-like “roar” (gajjita). Most suggestive are the
“Secret Vinaya” (Guḷhavinaya) and the Vedalla Piṭaka, which a subcommentary
connects with the Mahāyāna.116
It is here that matters get complicated with Nattier’s statement that “the
origins of the term saddharma-pratirūpaka and its pairing with the term sadd-
harma should be sought in Mahāyāna circles” (1991, 47 n. 55 as quoted above).
This careful statement peripheralizes this older and clearly earliest Nikāya
school usage,117 and leaves little room to consider the suggestive sectarian and
historical implications that it raises. For Nattier, with the “strong possibility
that the term [saddharma-pratirūpaka] itself was coined in Mahāyāna circles,”
it is “unlikely” to have been “in wide circulation prior to the 1st century BCE”;
as “a Mahāyāna invention . . . it never gained significant currency outside
Mahāyāna circles” (69–70 and n. 17).

115. Nattier 1991, 88 n.83. I do not recall Nattier making this important point about “specific (Mahāyāya)
Buddhist scriptures” with reference to specific texts.
116. Of course all this comes from the Theravāda tradition. It would be more interesting to know what was
made of the corresponding passage in the Saṃyktāgama. See Bodhi 2000, 808–9 n. 312, with citations.
117. See further Nattier 1991, 68, first referring to the usage of saddharma-pratirūpaka in this passage as
the “sole (and extremely atypical) occurrence in the Pāli canon.”
 dharma over time, ii 323

A Nikāya-school sūtra passage does not, however, need to be cross-referenced

or repeated to have achieved wide circulation. Assuming that Indian Mahāyāna
authors are behind the Chinese CS, we do not have to imagine that they had to
know this suggestive usage from the Pāli canon. Unlike the Kauśāmbī myth
itself, which is found in the Saṃyuktāgama but not in the Saṃyutta Nikāya, this
passage occurs in both. As Nattier indicates, in the “two separate translations of
the Saṃyukta-āgama” into Chinese, one finds its conventional “Chinese coun-
terpart hsiang-fa” used either as the equivalent of saddharma-pratirūpaka or as a
component of its equivalent in both translations (86–87 and n. 78).118
Yet the important point so far, which has not (as far as I can see) been
given sufficient scrutiny, would be that in the Chinese CS, it is no longer the
True Dharma that vanishes at Kauśāmbī but the Semblance Dharma. Thus the
Khotanese Candragarbha Sūtra, completely in keeping with the Chinese CS in
maintaining a correlation between the True Dharma/Semblance Dharma dis-
tinction and no more than 1,500-year timetables, ends its version: “and the
reflection of the True Dharma (saddharma-pratirūpaka) in Jambudvīpa will per-
ish completely” (Nattier 1991, 193). In contrast, again in accord with its nondis-
tinction on this matter and 2,000-year timetable, the Tibetan version ends
with, “At that time, on that occasion, the True Dharma (saddharma) in
Jambudvīpa will completely disappear” (250). To invert the usual saying, this is
a non-distinction with a difference.

b.3.b. core and frame. This brings us to our second topic: the frame-core
relation in the Chinese CS.119 In brief, while one can find many clear indications
that the M-cluster builds directly from the S-cluster’s epical innovations,120 it
seems to smooth out the overall story, remove doubt wherever possible over
questionable motivations,121 drop some Brahmanical themes, and omit the

118. Nattier also remarks that Rhys Davids’ translation “counterfeit doctrine” is appropriate not only for the
singular Pāli usage but for “its Chinese parallels” (1991, 89 n. 93). Indeed, whether or not we posit Indian authors
behind the crucial passage in the Chinese CS, we need only assume that its originator—perhaps even its
translator—connected dots from the Saṃyuktāgama. I see no compelling reason to challenge the Indian origins of
passages in the Chinese CS’s telling of the Kauśāmbī myth. For one thing, the discovery of “a small fragment of a
Sanskrit manuscript of the Candragarbha-sūtra” allegedly “corresponding to” the Chinese CS version “demonstrates
that at least a part of that text once existed in Sanskrit” (Nattier 1991, 174 and n. 70).
119. Nattier’s view is helped by the fact that each version frames its core differently (1991, 177). But as the first
frame to have been translated, that of the Chinese CS would be the best candidate to have originated in India.
120. The Tibetan version includes traits of Duṣprasaha that help to explain his prodigious power: an iron
mole in his forehead; an iron body in battle (Nattier 1991, 244). The summary does not mention these traits, so
the Chinese CS may not include them. They are reminiscent of Mahābhārata folklore about Duryodhana,
although one could find such traits in other “iron age” heroes.
121. Śiṣyaka seems to offer his words in good faith. Duṣprasaha is redeemed. Similarly, rather than dying,
King Mahendrasena abdicates the throne (Nattier 1991, 179), in the Tibetan version to renounce the world (224).
324 dharma

Aśokāvadāna setting of the quinquennial assembly. And while it keeps the

basic spatial orientation with selective modifications,122 it changes the temporal
one. In principle, the hinge period is now the 1,000-year Semblance Dharma.
Rather than finding out that he has a twelve-year window123 in which to forestall
the end of the Dharma, Duṣprasaha is concerned for himself, and learns that
he should ask the tripiṭaka-master Śiṣyaka about “the karmic effects of his
military campaigns.” At Śiṣyaka’s recommendation, he “invites all the monks
in Jambudvīpa to Kauśāmbī, but many are killed along the way” (Nattier 1991,
179–80). The Tibetan translation indicates that they come in distinct “saṅghas,”
and that those who die en route do so miserably. A hundred thousand get to the
capital. Considerable emphasis is placed on “all the monks of Jambudvīpa”
being present,124 and in both the Chinese CS and Tibetan versions that
one monk has not arrived: “At this point the king develops a desire to see
an arhat and is informed in a dream about the whereabouts of Sūrata”
(180). Here the Tibetan version makes Sūrata more interesting:125 the divine
voice in Duṣprasaha’s dream tells him that Sūrata is presently on Mount
Gandhamādana—a Himalayan haunt of Ṛṣis in the Mahābhārata—where
presumably the invitation did not reach him. Duṣprasaha should invite Sūrata
and confess to him so his sins will be purified (246). He does not seem to have
heard of Sūrata before this. It takes a dream to find the last arhat and bring him
out of the Himalayas.
Nattier’s summary of what follows is rather brief: “On the day of the
poṣadha ceremony the familiar conflict between Śiṣyaka and Sūrata breaks out”
(1991, 180). As to the usually revealing conversation between them, the Tibetan
version also gives us little to go on:

[some] newly arrived monks will ask the tripiṭaka-knower Śiṣyaka, the
preceptor, to explain the Vinaya. [But] Śiṣyaka will say, “To a man

122. There are three evil kings. Unlike the other two versions, the Chinese CS alters one’s identity, substi-
tuting Persian for Greek; but it is the only one to keep the original Indian locations with Kauśāmbī in the east
(Nattier 1991, 177–78, 187). The M-cluster is apparently divided on mentioning Madhyadeśa (see discussion
above).
123. Three twelve-year periods are folded into the endgame in the Tibetan translation, as if to compensate
for the loss of meaning: Duṣprasaha is twelve when his father yields him the throne (in the Chinese CS he is
seven); he fights the three kings for twelve years; and he learns that if he takes refuge in the three jewels for twelve
years he can purify himself of the killings, which leads him to invite all the monks to the unnamed feast (Nattier
1991, 243–45). In coming from “the country called Paṭaliputra” (Nattier 1991, 244) to “preach the Dharma for the
king,” Śiṣyaka tells him its “‘fifteen hundred years are now fulfilled,’ and that the Dharma is about
to die out” (180).
124. The Tibetan version makes it clear that, as a result of the events at Kauśāmbī, Jambudvīpa—implying
the world—becomes devoid of monks (Nattier 1991, 245).
125. Without indicating his birthplace, as it does with Śiṣyaka, keeping the latter at Paṭaliputra as a son of
the Brahmin Agnidatta, Sūrata is still the son of Sudhana; see discussion above.
 dharma over time, ii 325

who is blind and has no nose or ears, what need is there for a mirror?
What will he be able to see? Even if I explain the Vinaya, you will not
act according to the Vinaya. To those who do not keep the Buddhist
precepts (dharma-śikṣa), what is the use of explaining the Vinaya?”
And when he speaks thus, at that time the arhat will speak with the
voice of a lion: “From the time when I first took the Buddhist
precepts until now, I have not harmed so much as [a blade of] kuśa
grass. Do not say such words. Explain the Vinaya!” (247)

Śiṣyaka may still be dodging. But even if both of them seem to speak in good
faith, the result is the same mayhem.
Arriving on the scene, Duṣprasaha finds “all the monks dead,” “bewails
the fate of the Dharma,” and cremates Śiṣyaka and Sūrata (Nattier 1991, 181).
In the Tibetan version, he is so pious as to say, “The arhat was my mother;
Śiṣyaka was my Dharma treasury” (248). Much ennobled, he does not carry
out the final killings himself or have a sudden rage. But now, instead of being
rebuked by 500 lay disciples, the Chinese CS “suddenly has a contingent of
monks appear out of nowhere” after all the real ones are dead. When
Duṣprasaha asks them to preach the Dharma, he learns they know nothing
about it and realizes that it has “indeed disappeared” (181). The Chinese CS
may condense and revise a story that makes more sense in the Tibetan ver-
sion. There the king’s 500 ministers (the ones who were born to a rain of
blood) now have “pity” for him, and put 500 fake monks to the task of fooling
him that the Dharma has not disappeared. The “impostors,” having no
Buddhist robes to wear, seem to dress up like Śaivite ascetics: “in animal
skins, . . . their hair and beards . . . not shaved” (249; cf. 181). Perhaps the
Tibetan version puts a Śaivite twist on the end of the Dharma. On the other
hand, if a single unreal monk could be the “fourth sign” to Prince Siddhārtha
before any real monks could appear outside his father’s palace, pending his
rediscovery of the Dharma, why should not a contingent of unreal monks
appear out of nowhere as the last sign that his Dharma had vanished?
Clearly, if we are to find redirection out, we might look to the frame. But let
us already note the fighting saṅghas; the general emphasis on Vinaya and pre-
cepts (dharma-śikṣa, the basic training rules?) rather than on the Prātimokṣa;
the mirror offered to all the monks as a reflection of their Vinaya shortcomings;
and the intensified uniqueness of the last arhat who has been out of touch with
the “saṅghas” as a Himalayan recluse. And let us note, along with the mirror,
the dream and the closing emphasis on illusion, which may reflect Mahāyāna
perceptions of reality, that is, of an unfathomable Dharma that reveals itself
mysteriously and would not really be disappearing at all. As one reads in the
326 dharma

Vimalakīrti Sūtra, which must be a fairly early Mahāyāna text (Watson 1997, 1),
“There are lands where similes such as dreams, phantoms, reflections, echoes,
images in a mirror, the moon in the water, or shimmering heat waves are used
to do the Buddha’s work” (123–24).
Turning now to the frame, my difficulty is with the implication that it is
nonessential,126 late, and “probably” added in China. I think it was just
retouched there, and has demonstrably Indian features:127 for one, taking a
page from Indian Pure Land ideas, that by his austerities in previous lives the
Buddha could “make the Dharma grow bright” when he became a Buddha
(Nattier 1991, 182; cf. 2003, 185–93). Granted that the Chinese CS frame was
“added” to a core story, if the frame and reframed core were put together
(“composed”) in India, it might help to explain the Mahāyāna reorientation of
the Kauśāmbī myth.
A bodhisattva asks the frame’s leading question: How can people use skill-
ful means to prolong the existence of the Dharma?—not, as Nattier observes,
How long will the Dharma endure? The Buddha replies “that one hundred years
after his death, the Dharma will gradually [begin to?] decline and disappear, and
delivers a long discourse on the evil actions of future sentient beings,” differen-
tiating the special misconduct of monks, Kṣatriya kings, and Cānd ̣ ạ l̄ a kings
before the core story begins. Then, once the Dharma has disappeared, the
128

frame closes “with a long concluding section.” After describing “the decline to
be suffered by the natural world, human beings themselves, and the Dharma,”
the Buddha says “he will renounce one-third of his life span for the benefit of
sentient beings, thus prolonging the life-span of the Dharma,” and now pro-
nounces the text’s authoritative 1,500-year timetable with its 500-year True
Dharma and 1,000-year Semblance Dharma sequel. He “emphasizes the
importance of giving to the monastic community, even if the monks are
breaking the precepts,” on the grounds that giving to them “is like making

126. Nattier seeks to confine “minor Mahāyāna elements” to the frame while demonstrating that the core
has been little affected by its Mahāyāna handling: “In the Chinese version . . . [t]he main lines of the Kauśāmbī
story are devoid of Mahāyāna content” (1991, 184); “All in all, then, there is little significant innovation. . . .
Despite the peripheral Mahāyāna accretions. . . .” But the frame’s Mahāyāna elements would not look minor if
the core has innovative Mahāyāna resonances.
127. Nattier lists eight indices by which “we may tentatively suggest that the ‘frame’ . . . was added in a
Chinese context.” The most persuasive are two that suggest idiomatic retouches such as are common in Chinese
translations: an accentuation on filial piety, and looking on one’s parents as “field deer.”
128. Nattier 1991, 176–77 (my brackets). The full 1,500-year timetable is predicted in the frame’s closure
(182). Without the brackets, this “hundred” looks inconsistent with the rest of the Chinese CS. Nattier proposes
that it could be a scribal error for “ten hundreds,” and that the text was referring to a 1,000-year timetable like our
S-cluster (176 n. 74). Neither is likely, and Nattier has just noted that the Buddha was not answering about the
duration of the Dharma, which he gets to on the frame’s far side. But note that the argument admits the possi-
bility of an original Indian frame.
 dharma over time, ii 327

offerings to me.” Next, he mentions various austerities he did in previous lives

“in order ‘to make the Dharma grow bright,’” which now becomes a “refrain”
as he offers a dhāraṇī (mantra) “for the purpose of making the Dharma long
endure in the world” (182). With this, “good omens appear, sentient beings of
various types attain new levels of insight, and the Buddha increases the three
‘essential life forces.’”129 He then tells how to divide the merits left over from
the offerings he made to former Buddhas. One part will go to himself.130 One
part will go to “those śrāvakas who are firm in samādhi, leading to liberation,”
which probably implies the 500-year duration of the True Dharma when arhats
will not yet be so scarce. And the last part will go to “śrāvakas who break the
precepts but recite the sūtras and put on monastic robes in the times of the True
and Semblance Dharma.” That is, there will be śrāvakas who break the pre-
cepts, etc., during both periods. The Buddha then closes by asking Maitreya to
protect “these śrāvakas so they will not be poor,” and “to prevent caṇḍāla kings
from harming those śrāvakas who break the precepts.” Finally, all sentient
beings give his sermon a rousing reception (182–83).
Several points call for discussion. First, it is rather easy to show that the
Chinese CS’s frame gives its whole text a Mahāyāna momentum. One can see
this most vividly in the redemption of King Duṣprasaha. Second, as this and
other examples show, the frame is interreferential with the Chinese CS’s core.
For instance, the Buddha’s appeal to give to monks who break the precepts con-
firms Duṣprasaha’s piety in doing so down to the Dharma’s last moment, and
reinforces his redemption. Redemption of a demonic king is typical of Purāṇic
bhakti, which, as we noted in chapter 6, ultimately, like Mahāyāna Buddhism,
envisions a universal salvation open to everyone, “demon devotees” included.
There will be more to say about the Mahāyāna. But it is worth attending before
that to the way the Chinese CS’s frame redescribes Indian social and institu-
tional conditions mentioned in the S-cluster’s main story.
While obviously building on the S-cluster, the Chinese CS appears to clear
out social problems from its core to relocate them in its frame. The frame first
mentions such conditions in its opening where the Buddha predicts the three
types of misconduct that will begin to surface a 100 years after his nirvāṇa,
and then continues in its closing to relate them to later conditions, leaving it

129. Nattier mentions the dhāraṇī and the “three essential life-forces” among the indices of Chinese com-
position, but neither is wholly convincing. She does not make up her mind on the dhāraṇῑ (1991, 219). As to the
three essential life-forces, Nattier says they do “not seem to reflect any term in the Indian Buddhist repertoire”
(1991, 184–85). This term does sound Chinese, but the three are those of the earth, of all living Buddhas, and of
the True Dharma (182 n. 88), which are highly plausible in an Indian context.
130. Which may be enough for the rest of his life, or more likely imply that he has a continued existence
after parinirvāṇa (as we just saw, it seems he can continue to receive offerings).
328 dharma

obvious that they apply to the core story just told. The three types of miscon-
duct predicted are that Buddhist monks will be “more concerned with worldly
success than with religious achievement”; Kṣatriya rulers will cooperate “with
these decadent monks” and interfere “with them by imposing punishment
on those who do not obey the precepts”; and Cāṇḍāla rulers “will conspire
with evil monks to turn out their own version of the Buddhist teachings”
(Nattier 1991, 176–77). The description of worldly monks is now familiar—
among other things, as a Mahāyāna critique of settled life in wealthy Nikāya
school monasteries. Yet these conditions seem to be general: the Buddha asks
Maitreya to protect all wayward śrāvakas “so they will not be poor.”
The matter with Kṣatriya rulers is curious. In contrast with Cāṇḍāla rulers,
they would administer punishment from within the Brahmanical system of
caste and life-stage (see chapter 5), which would leave ambiguities in their deal-
ings with Buddhist monks. As many have noticed, the Prātimokṣa singles out
one rule, the second in the most serious Pārājikā category requiring expulsion,
where a monk is punishable by a king, and that is theft.131 As the Chinese CS
reads here, one could infer that Kṣatriya kings who cooperate with decadent
monks yet interfere with them in imposing punishment on lax ones would do
so by applying the rod of punishment (daṇḍa) to additional Vinaya matters
beyond theft. This would involve a tension between Vinaya and Dharmaśāstra,
which could extend to the fact that Brahmins are ordinarily exempted from
punishments, or if not exempted, given the lightest penalties. Perhaps one may
read between the lines and suspect that Kṣatriya rulers are treating monks
selectively as they do with Brahmins, but not as well. They would treat Buddhists
in ways that Brahmins are immune to. In any case, the differentiation between
the two types of rulers has clearly stepped up what we found in the S-cluster,
where the contrast was between a “good Buddhist king,” presumably a Kṣatriya,
and Mlecchas or “Barbarians,” not Cāṇḍālas or “Untouchables.” Mleccha kings
are a commonplace in the Sanskrit epics, where the same Greeks, Sakas, and
Parthians are mentioned most under that heading just as they are in the
S-cluster. But not Cāṇḍāla kings. I am unaware of such a usage in Brahmanical
literature.132 Whereas the Buddha, on principles that spoke to equality, and

131. See Collins 1993, 370 (on the original rule traced back to King Bimbisāra); von Hinüber 1995, 33–34
and n. 72 (on “legal levels” of persons below the king who may also punish for theft, according to the Sūtra-
Vibhaṅgha).
132. For nuances given these and related terms in Brahmanical, Buddhist, and Jain literatures of our period,
see Parasher-Sen 2006, 431–42. She notes that in using the word mleccha, the Viṣnu ̣ Purānạ introduces a distinction
“between the traditional kṣatriyas and the new ruling elites” that differed from the Brahmin/Kṣatriya distinction of
interest to the epics (433). But the new distinction was still made with the term mleccha, not cānḍ ạ l̄ a.
 dharma over time, ii 329

against Brahmanical ideas of hierarchy based on purity, once exalted the

Khattiya over the Brahmin to defend Buddhist monks from the charges of
Brahmins like Ambaṭṭha and Pokkharasāti that they were Cāṇḍālas scoured
from the soles of Brahmā’s feet, he is now projecting as Buddhist a view that
seems to go beyond Brahmanical texts in naming foreign non-Kṣatriya kings
Cāṇḍālas.133
It is thus the misconduct of Cāṇḍāla rulers that is most intriguing. Unlike
Kṣatriya rulers, they are faulted on both sides of the frame. As we noted, the
Buddha’s final appeal is to ask Maitreya “to prevent caṇḍala kings from harm-
ing those śrāvakas who break the precepts.” Whereas a Kṣatriya king like
Duṣprasaha is now expected to act within the Law, even if it is not Vinaya law,
in punishing monastic laxity, Cāṇḍāla rulers lack such restraints. If we are to
look for “Cāṇḍāla” kings who “harm” monks, we should probably look past the
pax kushanica. The most likely background would be the invasions by the
Hepthalite Huns (“White Huns”), who, led by Toramana and his notoriously
brutal son Mihirakula, devastated monasteries in northwest India and Bactria
and may have provided an impetus for the composition of this version of the
Kauśāmbī myth as well as its transmission in the Chinese CS to northern
China, which also knew Hepthalite invasions and had barbarian kings in the
same period.134 If so, the Chinese CS would have been composed in the late
fifth- or early sixth-century CE, a time of rapid late-Gupta decline, yet still early
enough for its Chinese translation in 566.135
But the Chinese CS also draws a text-historical association with Cāṇḍāla
rulers that could have had nothing to do with persecution: they “will conspire
with evil monks to turn out their own version of the Buddhist teachings.” This
cannot be squared with Hephthalite kings who never converted to Buddhism,

133. As Schopen has shown (2006, 344–47), Buddhists of the Middle Period, both in the Mūlasarvāstivādin-
vinaya and in a Mahāyāna text called the Ratnamegha Sūtra, had also gone beyond Brahmanical usage of calling
others Cāṇḍālas to using this name to differentiate among, and even to identify, themselves. Schopen shows this
with reference to Vinaya rules made to “sanitize” the early Buddhist śmāśānika practice of wearing cloth from
corpses. But as he says, “This same pattern” holds for “the monastic disposal of the dead and monastic inheri-
tance law, and could easily be demonstrated in regard to a very broad range of other, seemingly more mundane
matters, like washing bowls and providing drinking water.”
134. I am condensing a complex matter on which Nattier offers a somewhat different view in debate with
Ryūjō Yamada, who holds that the Hephthalite invasions were a spur to the composition of the Chinese CS (see
Nattier 1991, 110–17, 128, 164). While Nattier is convincing in disputing other views of Yamada’s, her three
points to contest this one are not convincing: that the Kauśāmbī story was already circulating for centuries before
the Hephthalite invasion; that it “exhibits few divergences”; and that it probably had its origins not in persecution
but the experience of excessive comfort (116–17). It is not a question of origins but of a reorientation. Thapar
2002, 286–87 gives a death-date of “about 542” for Mihirakula.
135. Note that Nattier posits three intermediary hyperarchetypes δ, ζ, and θ between our S-cluster and the
Chinese CS (1991, 216–18). I think, however, that the transformations they introduce can be sufficiently
accounted for as reorientations from within the texts that we have.
330 dharma

and does seem to hark back to the pax kushanica. Sure enough, we find the
Kauśāmbī prophesy tied to a negative depiction of the Kuṣāṇas in the most
prestigious version of the S-cluster, the one in the Saṃyuktāgama where
“Kushan” is the fourth invading barbarian king.
Yet if the Chinese CS is recalling the Kuṣāṇas, or more precisely this king
named Kushan, and thus probably Kanishka, in the Cāṇḍāla role of conspiring
“with evil monks to turn out their own version of the Buddhist teachings,” what
teachings would they be? If it is a reference to texts, there would seem to be
only three possibilities: Mahāyāna texts, Nikāya school texts, or both. The first
suspicion would be that we have a reference to new Mahāyāna sūtras written
during or around the Kuṣāṇa period. But that is the least likely solution since
the Chinese CS is unlikely to be casting aspersions on Mahāyāna sutras—
especially in its frame, where its Mahāyāna allegiance is so explicit. The second
possibility is more promising but also more complicated. The reference could
hold a negative memory of Kanishka, whose reputation connects him with a
fourth council where 100,000-verse commentaries were supposedly composed
for each of the three baskets of the Sarvāstivādin tripiṭaka.136 From this angle,
the “evil monks” would be Sarvāstivādins. Also, if one asked which side would
be likelier to garner royal support, this too might favor a Nikāya school candi-
dacy since up to the Gupta period there is so little evidence of institutional
support for the Mahāyāna. If we take this route, the Chinese CS could even be
using the True Dharma/Semblance Dharma distinction to turn the tables on
the main Nikāya school usage of this distinction in the Saṃyuktāgama, where
the “Counterfeit Dharma” would probably have been taken (whatever the plau-
sibility of such a text-historical reading) to refer to the Mahāyāna. Indeed, the
Chinese CS says Cāṇḍāla rulers will conspire in turning out “Buddhist teach-
ings”: not “false teachings,” and not “sūtras.” It could refer to the massive
Sarvāstivādin commentaries or to the voluminous Mūlasarvāstivadin-vinaya!
This looks to me like the deepest level of allusion, but the third possibility must
also be considered. Buddhist literature also viewed Kanishka “as an outstanding
patron of the Dharma” under whom the whole “Buddhist community . . .
flourished” (Nattier 1991, 155, 226). One might also suspect that the Chinese
CS is “Catholic” and prophetic enough to view textual complicity with barbarian
power as a general problem. In this regard, one might recall that both northern
Nikāya schools and the Mahāyāna were developing open canons during the
Kuṣāṇa period, and that these were becoming interpenetrating canons in the

136. Lamotte 1988, 585–86. This is not to assume that this reputation is historically accurate. As Nattier
2008a, 4 cautions, such an assumption, given Kanishka’s varied profile in different texts and by different schools,
“would be very dangerous.”
 dharma over time, ii 331

process of reaching China (and later Tibet). If the Chinese CS is as late and
“Catholic” as I have been suggesting, it could be that its author was content to
bury old squabbles and to veil his allusions to them.

b.3.c. a mahāyāna retelling. This brings us to our last window on the

Kauśāmbī myth in the Chinese CS, on which the frame clearly has more to tell
us about its Mahāyāna rehandling. Let us look to the bodhisattva Moon-Lamp’s
opening question, How can people use skillful means—upāya-kauśalya, the
capstone, and complement to the six perfections of the Bodhisattva path—to
prolong the Dharma’s existence? This is not only a leading question but a
wraparound one—in at least one sense, and I will say two. It is clearly a
wraparound question in the sense that the Buddha answers it directly on the
other side of the frame. After telling the Kauśāmbī story, he uses skillful
means in what Nattier calls his “heroic efforts” to prolong the Dharma by
shaving off a bit of his life span and offering a dhāraṇī (Nattier 1991, 184). But
I think it is also a wraparound to a Mahāyāna reorientation of the Kauśāmbī
myth itself. On this, despite concluding that “the Kauśāmbī story does not
exhibit any major deformations in plot due to changing historical circum-
stances,” Nattier seems to agree: “when the story finally begins to appear in
Mahāyāna contexts the authors seem rather uncomfortable with its negative
conclusions, and show signs of attempting to step back from the finality of the
extinction of the Dharma that it portrays” (223–24). This is a good formulation,
but more needs to be said.
The wraparound question takes us not only to the Buddha’s “heroic efforts”
but prompts the Kauśāmbī prophesy itself, which, in the mode of Mahāyāna
storytelling, is itself an exhibit of using narrative “reality effects” as skillful
means.137 Given the trenchant dialogue in the alpha-cluster and the epical and
demonic innovations of the S-cluster, the Chinese CS is a surprisingly compas-
sionate, or perhaps better, indulgent text. If there is one message it gives over
and over again, both in the frame and core, it is the merit of giving to monks
who break the precepts. As the Buddha says in the frame, giving to them is like
giving to me. Clearly this applies to giving both during the True and Semblance
periods of the Dharma, and implicitly to both Nikāya school and Mahāyāna
monks. But with Moon-Lamp and Candragarbha as the Kauśāmbī myth’s new
bodhisattva interlocutors, it also bears on the Buddha’s austerities in former
lifetimes in the presence of previous Buddhas, since the Chinese CS is set
against the foreground of the Buddha’s soon becoming a previous Buddha

137. See Lopez 1993; the classic dharma narrative “fiction” as skillful means is the Lotus Sūtra’s parable of
the burning house.
332 dharma

himself. Just as the Buddha has merits to spare from his previous lives as a
bodhisattva, the bodhisattvas in his audience could contemplate having the
same in future kalpas or aeons. It cannot be incidental that the story of the last
arhat now features two listening bodhisattvas. The Buddha is telling how
Bodhisattvas using skillful means could prolong the Dharma past the end of
the Dharma. This could be one of the reasons for the audience’s final
applause.138
So far I have offered this interpretation of the Chinese CS’s rehandling of
the Kauśāmbī myth inferentally from the text’s own Mahāyāna logic. I believe
the proposed connections are valid, but that they would look more persuasive if
there were a Mahāyāna version of the Kauśāmbī myth that made these points
less subtly. On the main point, there is. Of three Mahāyāna versions of the
Mahāparinirvāṇa Sūtra translated into Chinese between 410 and 435 CE, one of
them, probably the second translated (between 414 and 421), contains what
Nattier calls “a brief retelling of the Kauśāmbī story” that “is framed by the
statements that ‘the Three Jewels manifest disappearance but this also is not
permanent extinction,’ and ‘My True Dharma is not really extinct! At that time
[Kauśāmbī] will have 120,000 bodhisattvas who will righteously uphold the
teaching’” (Nattier 1991, 223 n. 17). Chappell gives some of the context:

the Nirvāṇa Sūtra depicts the death of the Arhat who teaches the
True Dharma, and the harm which is suffered by 600 monks. But in
response to the lament by the common people that the Nirvāṇa Sūtra
depicts the death of the Arhat who teaches the True Dharma, and the
teaching of the Buddha is destroyed, it reports the Buddha saying:
“My True Dharma really was not destroyed! At that time their
country had 120,000 various bodhisattvas who righteously upheld
my teaching.” Later,139 Kāśyapa explains to the multitude that they
should not be anxious or filled with remorse. “Earthly existence is not
empty. The Tathāgata always exists without changing, and the
Dharma and Sangha as well.” (Chappell 1980, 140)

There are several striking differences here from what one finds in the Chinese
CS. One is that the Chinese CS does not formulate the Dharma’s perpetuation
in relation to the Three Jewels (the Buddha, Dharma, and Saṅgha); rather the

138. Another would be the appeal this text would have in Chinese circles that developed the mo-fa doctrine
of the Latter Day Dharma (Nattier 1991, 90–118). As Nattier says, while “Chinese writers” continued to make
“fairly modest” estimates of the duration of the True and Semblance Dharmas, with the introduction of a mo-fa
lasting a myriad or 10,000 years, “Chinese authors conveyed the distinct impression that the era of the final
dharma was here to stay” (2008a, 157).
139. Does “later” mean after the Kauśāmbī story is told? After the Buddha’s death?
 dharma over time, ii 333

Buddha increases the three “essential life forces” (of the earth, all living
Buddhas, and the True Dharma). But this, a least in theory, is a two-thirds
overlap, and the Chinese CS is surely also hinting at the perpetuation of a
Bodhisattva Saṅgha. Another difference, apparently, is the Buddha’s use of the
past tense to describe Kauśāmbī not so much as a prophesy but as something
that will have happened. Further, it is not the Semblance Dharma that disap-
pears with the last arhat but the True Dharma itself. Indeed, it appears that the
Semblance Dharma continues after Kauśāmbī. According to Chappell, “after
the True Dharma disappears and there is the Imitation Dharma, the Three
Jewels only appear to die.”140 This raises the fascinating possibility that a
Mahāyāna text would correlate its own continuation with the Semblance
Dharma! I would suggest that this is not improbable as one Mahāyāna
interpretation—a positive and appropriative one—of the Semblance Dharma,
which the Chinese CS itself portrays as a dharma that works through dreams,
mirrors, illusions, and skillful fictions.141
Most important, though, is that this Mahāyāna retelling of the Buddha’s last
journey and final conversations seems to give him the same monastic travelling
companions he has in Nikāya school versions of the Mahāparinirvānạ -sūtra (or
Pāli Mahāparinibbāna Sutta). Although there will be 120,000 bodhisattvas to righ-
teously uphold the teachings in Kauśāmbī country after the demise of the True
Dharma, no bodhisattvas seem to have been added to the entourage attending the
Buddha’s own demise—as they are to his audience in the Chinese CS. The con-
clusive point is thus the same. In one way or another, with the passing of the last
arhat and also of Śiṣyaka with his disciples,142 there will be bodhisattvas to con-
tinue the Dharma, which means, of course, the Mahāyāna. Clearly the arhat is no
longer what is primary. He is what was primary. With all due respect to arhats, the
Mahāyāna can get along without them. Their days are numbered, at least in East
Asian Buddhism. Here again we find an ironic indication that the myth of the end
of the dharma would find its own prophetic confirmation.

C. The Yuga Purāṇa and the Kauśāmbī Myth

We are left with the following questions: Does this study of the Kauśāmbī myth
(henceforth KM) reinforce our interpretation of the YP? How do the YP and the

140. Chappell 1980, 140, with “Imitation Dharma” being his translation of saddharma-prātirūpaka.
141. Nattier characterizes this text as offering “a fervent defense of the docetic view of the Buddha” (1991, 39).
142. See Nattier 1991, 215 n. 8: the brief versions of the Kauśāmbī myth in the Mahāyāna Sūtra and the
Mahāparinirvāṇa Sūtra share a detail not found elsewhere: that “the tripiṭaka-master [i.e., Śiṣyaka] is described as
having five hundred followers.” Chappell mentions the harm “suffered by 600 monks” (as quoted above).
334 dharma

KM relate to each other, and what light might they shed mutually and differen-
tially on dharma over time? We have made our central question that of dharma’s
entanglement not only with terminologies of cosmological time but with the
relation of such concepts to change and duration in history.

1. Although the YP and KM each predict events against the background

of Greek and Śaka incusions, no version of the KM would be likely to
know the YP, especially if the earliest alpha-cluster version of the KM
was written from a Bactrian perspective. We may thus rule out a direct
relation between the two narratives.
2. In predicting events, the YP is a once-and-for-all short text, appar-
ently largely forgotten, with no elaborations. As the history it
describes recedes, that history gets left behind in the astrological
treatise that predicted it. In the KM, although the names and number
of the invading kings change, the myth keeps its structure. Thanks to
its distance from the historical events it mythifies, the invading
kings, from the beginning, have had time to become barbarian
ethnonynms. In contrast with the YP, the barbarian kings come to be
stupid enemies of the faith rather than potential allies or contributors
to a divine plan.
3. The YP and KM have overlapping geographies. Each begins from
north-central India: the KM in Kauśāmbī; the YP in areas that would
surround and include Kauśāmbī (which, however, the YP does not
specifically name). Yet they are oriented in totally different directions.
The YP is oriented eastward toward a Magadha transitioning from
power and preeminence to decline, and south into the Deccan and the
Kāverῑ area. It thereby takes in a vision of the whole subcontinent. The
KM, after beginning from a possibly Bactrian perspective, introduces a
central north Indian and eastward Magadha focus only in the S-cluster,
which has its monastic protagonists Śiṣyaka and Sūrata come from
Magadha. This suggests a sense of Magadha as by now a region in
decline, and notably so, for Buddhists. Otherwise the KM’s focus
remains on the northwest, where it positions events to take Buddhism
out of India into China and Tibet.
4. The YP and the KM each appear to be somewhat atypical of their
traditions’ treatments of history: the brahmanical text for being the
more historical; the Buddhist ones for being more mythical in
treating the same events. The YP builds up to the Yavana and Śaka
incursions as real ex eventu prophesy, and describes them with
surprising historical immediacy. In contrast, all versions of the KM
 dharma over time, ii 335

use ex eventu prediction as a trope. Its prophesized history is entirely

fictionalized. Whereas a Pañcāla–Mathurā–Yavana alliance against
Magadha has historical plausibility, a simultaneous alliance of Yavana,
Śaka, and Parthian kings to invade Kauśāmbī is a purely retrospective
imaginary.
5. While not naming Yavana and Śaka kings, the YP describes them as
separately invading peoples in a contentious historical setting,
envisioning possibilities of tactical alliances and strategies of retreat to
southern safe havens. While its strategies of retreat have reminded us
of the Mahābhārata’s Rāma Jāmadagnya myth and Vyāsa’s Kali yuga
prophesy near the end of the Harivaṃśa, its tactical alliances remind
us of these two texts overlapping central story. The KM names only its
foreign kings, and fictitiously. When, from the S-cluster on, the KM
starts naming the Kauśāmbī king, it gives him no strategic possibil-
ities. His sole preoccupation is with Buddhist monks, and in the
M-cluster, he shows some interest in their texts.
6. In contrast to the interest that the Buddhist texts have in kings, the YP,
once one is past its predictions about of the Mahābhārata war and into
its prophesied history, gives no pivotal role to any king (like
Duṣprasaha). This may reflect the relative importance attributed to
royal sponsorship of texts by the two traditions in this period, and
reinforce our explanation of why attempts by modern scholars to find
kings (or at least big kings) as patrons behind early Brahmanical
dharma texts like the YP and the Mahābhārata have been so unproductive.
Big Brahmanical kings were at best still just wishful thinking.
Likewise, if Manu needed a king to sponsor it, why give such a strange
prescription for getting to be one?
7. In each case we have found the intervention of the Mahābhārata. This
is so from the beginning with the YP, for which the Mahābhārata is
certainly foundational. In accord with the Mahābhārata’s own primary
generic self-identification as an itihāsa, the YP views the epic as a
“history” that includes prophesy, and updates the prophesy as such
by turning it over to the predictive sciences and backdating it into the
mouth of Ṥiva. For the KM, the Mahābhārata is of no impact on the
alpha-cluster. But the S-custer fills itself out with likely Mahābhārata
allusions. The S-cluster introduces Mahābhārata themes principally
around its 1,000-year “age” with its twelve-year hinge and its evident
recentering of the KM in a Buddhist “madhyadeśa.” But such yuga-
shaped contouring recedes in the M-cluster. While the Tibetan CS
recalls something like the yuga sequence cast in terms of metals more
336 dharma

like the Greek ages than Brahmanical yugas, it subordinates this

reckoning to its notion of the true and semblance dharma,143 and, like
the M-cluster as a whole, reverts the KM’s time scale to the kalpa to
entail vast predictions about the Buddha’s life. Buddhism never settled
on the number or duration of the ages of the dharma in handling
change over historical time, and it reverts to the kalpa when handling
the course of the true dharma.
8. Finally, the role of prophesy differs. Śiva in the YP is reassuring that
dharma will always have its remnant as history unfolds. Yugadharma is
always in flux, but it is entitative and uninterrupted (at least for a kalpa
or a Life of Brahmā).144 In contrast, Kātyāyana and then the Buddha
describe a downward spiral of events that lead to the disappearance of
the true dharma—with some exceptions for the M-cluster, where the
mirror, the dream, and the closing emphasis on illusion may reflect
perceptions of an unfathomable Dharma that would not really be
disappearing at all. Otherwise, the Buddhist true dharma is discontin-
uous and always interrupted. The true dharma and semblance dharma
may allow for shifting interpretations, but there is no interest in what
I have called the default dharma of the Brahmanical householder,
which is what the Aggañña Sutta rejects in calling it “not the true
dhamma.” The true dharma does not leave remnants. Yet whether it
really ends or not, in the inevitable meanwhiles, Buddhism, unlike
most “world religions,” has the good graces to let the world continue
with or without it.

143. A few verses from the Tibetan Candragarbha Sūtra may remind us of a more Greek than Indian
account of the four ages, and also of the disappearance of the fragrant earth in the Aggañña Sutta: “At that time,
on that occasion, the True Dharma in Jambudvīpa will completely disappear. {And even the letters of the scrip-
tures will become invisible, as if one erased a conch shell.} [Then] gold will become bad silver and stone. As to
silver, it will become brass. Brass will become copper {and iron will become stone.} Jewels and pearls will turn to
horn. Of the six flavors of food, four—sweet, {sour, and salty flavors} and so on—will disappear. [And only] two—
the bitter and the pungent—will remain. (Nattier trans. 1991, 250; {} = not found in every version).
144. Published too recently to have been included in this discussion, but intriguingly pertinent, I believe,
to this entitative aspect of yugadharma, is Kloetzli 2010, 584, 615–18 on the yuga as the “joining” of the days of
men and days of the gods. “The destinies of men seem intertwined and joined with the gods of each age. The
result is four societies defined wholly by number” (608), which he compares to the four societies in Plato’s
Republic: timarchy, oligarchy, democracy, and tyranny (618).
8
Women’s Dharma
Śāstric Norms and Epic Narratives

As mentioned in the last chapter, toward the end of its discussion of

the Yuga Purāṇa, the Mahābhārata’s main story tells of a crisis of six
generations, of which the two sets of vying cousins, the five Pāṇḍavas
and the hundred Kauravas, make up the fourth. As I indicated, if one
wants to speak of “invention,” one must do so with respect to those
six generations in full, beginning with the intervention into the
Mahābhārata’s main dynastic line by the Goddess Gaṅgā. I also
mentioned that the epic poets extract plausible historical data from
the Vedic canon to trace their Bhārata itihāsa or “history of the
Bhāratas” through those six generations into the “future” of the Kali
yuga (see chapter 7 § A.4). As we shall now see, it is specifically
through Gaṅgā that this thread of itihāsa begins to be woven into the
Mahābhārata’s dynastic and generational time.
This book will introduce the topic of women and dharma against
that background. The obvious point is that generations require
women, and the Mahābhārata’s women are great, spirited, and
fantastic.1 We will thus be exploring the topic of dharma over time

1. Although I take a different route through this material in focusing on the women’s agency
with regard to dharma, it is worth quoting Biardeau’s take on the “generations that precede those of
the protagonists”: “It is always by women that the foreign element of disorder in a dynasty that must
remain interrupted is introduced (Devayanī, Śakuntalā, the two Satyavatīs, Ambā, Kuntī . . .), but the
women are also the obligatory mediatrices of the restoration of good order (once again, Devayānī,
Śakuntalā, etc.): the Puruṣa remains immobile facing its Śakti who does not stop manifesting the
phases of her yoga” (1979, 125; my translation). See also Chakravarti 2009, examining much of this
skein through the lens of transition in “gendered social and economic processes” (5).
338 dharma

through the three generations of dynastic instability and crisis that precede the
Pāṇḍavas’ marriage to their own fourth generation’s equally if not more remark-
able woman, Draupadī. As in this fourth generation, when Draupadī becomes
the stake who says dharma itself is at stake at the pivotal epic dice match, the
Mahābhārata women who precede her in the line can be remarkably active and
loquacious about dharma, and especially when it is imperiled.
We keep ourselves largely to the Mahābhārata in this chapter, since the
Rāmāyaṇa does not offer three such dynamic generations before its main story,
does not place itself explicitly in a transition between yugas, does not weave
itself into itihāsa or history, and—most important—does not envision dharma
changing during the lifetime of Rāma, or Rāma ever changing his view of
dharma. We shall come to such matters of dharma over biographical time in
chapter 9.

A. Strῑdharma

Brahmanical norms for women are set forth broadly through the concept of
strīdharma, “Law(s) for women” or “women’s dharma.” While there is no end
of śāstra-type instruction on the “dharma of women” (strī means especially but
not only “wives”), and we looked briefly at some of Manu’s more egregious
ones in chapter 5 § E, it is too easy to suppose that it exhausts the subject.
Nonetheless, Manu encapsulates strīdharma in a famous adage that is
paralleled in many of our texts:

Even in their own homes, a female—whether she is a child, a young

woman, or an old lady—should never carry out any task
independently (na svātantryeṇa).2 As a child, she must remain under
her father’s control; as a young woman, under her husband’s; and
when her husband is dead, under her sons’. She must never seek to
live independently (na bhajet strī svatantratām). She must never want
to separate herself from her father, husband, or sons; for by
separating herself from them, a woman brings disgrace on both
families. (M 5.147–49)

Among the dharmasūtras, the “broader” Āpastamba has no such adage.3 As

Jamison indicates, Āpastamba treats women largely within “the older śrauta
ritual model” of the “household pair,” in which the wife is recognized as one

2. See Olivelle 2005a, 287 on this “cause celebre” verse and on svatantra as a legal term; cf. 9.2–3.
3. Cf. Jamison 2006, 200; on the contrary Ā 2.15.9, 2.29.11 and 15, giving some latitude to women.
 women’s dharma 339

of the “two masters” (svāminau; Ā 2.4.13), along with her husband (2006, 192).
As was mentioned in chapter 5 (§ C at n. 60), Āpastamba also, like the
Mahābhārata, allows that one may learn dharma from women and Śūdras.
Baudhāyana and Vasiṣṭha, however, “quote” similar (unattributed) verses to
Manu’s second of these three, about a woman’s passage through three phases
of life—in Baudhāyana’s case to make the different point that she lacks
strength and cannot inherit property (B 2.3.43–45; cf. V 5.1–3). Gautama,
however, keeping matters to prose, seems to have made the strictures even
tighter than Manu:

A wife (strī) cannot act independently in matters related to the Law

(asvatantrā dharme). She should never go against her husband and
keep her speech, eyes, and actions under strict control. (G 18.1–3)

As recent studies have brought out, these texts do indirectly attribute

mental agency to women (Jamison 2006), and explicitly ascribe economic and
sexual agency to them (Olivelle 2005b, 248–60); but they leave us with what
Jamison calls “something of a paradox—that the more woman is textually
endowed with agency the more her capacity for independence is denied” and
the more she needs to be kept under “guard” (2006, 201). Jamison thinks that
one factor behind the increase of these strictures, and in Manu their misogy-
nist expression, could be the growing prominence of a “new female type, the
independent and religiously unorthodox woman,” the Buddhist and other
heterodox nun (206).
We may expect women’s “non-independence” to have narrative outlets and
subversions. Even Manu, as it contracts women’s worlds, may leave the tiniest
room for women’s dharma to breathe. On “Law Concerning Husband and
Wife,” the sixteenth ground for litigation, Manu begins, “For a husband and
wife who stay on the path (vartmani) pointed out by the Law, I shall declare the
eternal Laws for both when they are together and when they are apart” (9.1).
And, all be it that it is problematic, “even a woman . . . can give testimony.”4 An
epic heroine’s stories could be condensed into, or expanded from, these pro-
nouncements.5 Compared to the dharmasūtrakāras, in fact, Manu does create a
mini-narrative in his three famous verses. Out of Gautama’s strī, which means
“wife” in context, but before that “woman,” and out of the quoted adage about

4. M 8.70: “When there is no one else, even a woman, a child, an old man, a pupil, a relative, a slave, or a
servant may give testimony.” But negatively, see 8.77: “Even one man free from greed may be appointed as a
witness, but never women, even if they are many and honest, because the female mind is unsteady. . . .”
5. Draupadī, Sītā, and Śakuntalā, insofar as all three take their cases to court. Draupadī and Śakuntalā do
so literally; Sītā presents her case to accompany Rāma to the forest during the “court intrigue” of Rām 2, and
calls on divine witness in her two ordeals.
340 dharma

a woman’s three phases of life, Manu gives us a full female life cycle not only
with its wants, but within the “homes” of her successive families.

B. The Law of the Mother

If Āpastamba opens the possibility of learning dharma from women within the
older śrauta model of the household pair, and the epics spin that possibility
out, our questions are: Under what conditions does such spinning out occur?
What do women characters really desire6 . . . to say about dharma that would
also be dharma? Let me start with this last question and propose to explore
such matters chiefly under the name of the mother and call this dharma the
“Law of the Mother.”7 Although a wife may also be a mother, it seems to have
been easier to legislate for woman as wife (thus strīdharma) than as mother
(e.g., mātṛdharma, a term one will not find in any of our classical sources).
The Law of the Mother would, to begin with, be a name for something less
articulated—at least as dharmaśāstra. Similarly, by the end of this chapter, we
will also have discussed something we could call the “Law of the Girl,”
which—as the equally unattested kanyādharma—will have proved equally
resistant to proper codification.
I first used the term “Law of the Mother” in a conference presentation
about two Mahābhārata cults in Tamilnadu (2004c) to address the question of
whose law it is when members of a clan that has temples for Draupadī say they
worship their clan deity Periyantavar, who is none other than a reincarnation of
Duryodhana, because Draupadī gave permission to their ancestors to do so, but
“for only a day since his laws last only a day.” In brief, Duryodhana is Draupadī’s
former enemy who now holds power for a day over spirits (pēy) who are exor-
cised mainly from women who are disrupting clan (kulam) expectations that
they be dutiful wives or daughters. While Duryodhana’s male priests perform
the exorcisms with a clay horse that Duryodhana himself possesses, the women
dispossessed are carefully attended by older women: mothers and others. My
paper asked, “Is it really Duryodhana’s laws that operate or Draupadi’s, who
sets the limit of a day to the laws of Duryodhana? Duryodhana’s festival is in
many ways an inversion of hers. Could one say that at a deeper level,

6. On the semiotics of women’s desire, this chapter is affected by the work of Julia Kristeva.
7. Were this chapter short enough to include discussion of Buddhist sources, as originally conceived, it
would take note especially of the theme of debt to one’s mother that provides one of the main motivations for the
Buddha to permit the founding of an order of nuns (see Ohnuma 2006). Cf. also the practice by both monks and
nuns of making gifts to the sangha in the name of their mother and father, as recorded on early donative inscrip-
tions (see Schopen 1997, 35–42, 56–71).
 women’s dharma 341

Duryodhana’s festival takes place under the Law of the Mother?” In the
discussion that followed, Mark Pizzato of the University of North Carolina,
Charlotte, found the term “suggestive as a Lacanian oxymoron”—which was
gratifying, since I used it bearing in mind that the term “Law of the Father”
encapsulates Jacques Lacan’s view that “It is the name-of-the-father that we
must recognize as the symbolic function which, from the dawn of history, has
identified his person with the figure of the law” (Lacan 1977, 67; Grosz
1990, 71).8 Lacan develops this terminology as a way of reading Freud’s expla-
nation of the origins of patriarchy by the Oedipus myth.
Now the type of exorcisms done in the name of Duryodhana is not done for
him alone, though his are the best coordinated and probably the longest
sustained. Within the same valley north of Dharmapuri, the same type of rites
are also performed, at least recently, by an intermarrying clan in the name of
Periyantavar-Duryodhana’s wife Periyantacci, and a variation has also recently
been incorporated into the festival cycle at one innovative Draupadī temple. In
other words, although these rites have for some time been performed mainly
in the name of a clan father (Duryodhana), they can also be performed in the
name of a clan mother, while beyond that, Draupadī gives permission as a
mother whose festivals make it clear that her husbands’ law, reinstituted
annually by the festival that reenthrones her eldest husband Dharma (Tamil
Tarumar, i.e., Yudhiṣṭhira), is at some profound level also hers (Hiltebeitel
forthcoming-h). Since the Draupadī cult is a distillate of goddess worship with
folk and classical Mahābhāratas, we have an opening here to think back to the
Mahābhārata itself as a discursive world about which we might ask: might the
Law of the Mother be a term worth considering in connection not only with
Draupadī but with other Mahābhārata women whose stories intersect with
hers? For the moment, let us just note that the Mahābhārata is more open to
such an approach than the Rāmāyaṇa, where Rāma’s dharma lies mainly in
upholding the truth of his father’s word and thus standing resolutely for the
Law of the Father. In the Mahābhārata, on the contrary, the fatherless Pāṇḍavas
follow a Law of the Mother when Kuntī, their mother, tells them by an inadver-
tence, thinking they have returned with alms from begging when they are really
bringing home Arjuna’s newly won bride, to “share it all equally” (Mbh 1.182.2),
with the result that all five marry Draupadī (see chapter 10). By the end of this
chapter, and further in chapter 10, I will attempt to put Kuntī’s inadvertence
into a deeper dharma context.

8. I thank Mark Pizzato for his stimulating comments. I was at the time supervising a dissertation on a
Lacanian reading of the Mbh (Custodi 2005). On Lacan, law, and legal theory, see Caudill 1997.
342 dharma

It is beyond my competence to interpret the Mahābhārata in Lacanian

terms, much less revisionist ones. When I have presented this skein to audi-
ences without having time to go into it, I just speak of a “maternal dharma,”
and would expect that for some readers of this chapter that term may be prefer-
able. But it is striking that certain Lacanian phrases seem to be quite illumi-
nating of the Mahābhārata’s unusual treatment of fathers and mothers. Let me
work from a quote that would seem to give a little room for a Law of the Mother
to serve as something more than a suggestive oxymoron:

In fact, the image of the ideal Father is a phantasy of the neurotic.

Beyond the Mother, the real Other of demand, whose desire (that is,
her desire) one wishes she would assuage, there stands out the image
of a father who would close his eyes to desires. The true function of
the Father, which is fundamentally to unite (and not to set in opposi-
tion) a desire and the Law, is even more marked than revealed by this.
The neurotic’s wished-for Father is clearly the dead Father. But
he is also a father who can perfectly master his desire—and the same
can be said of the subject. (Lacan 1977, 321)

If we start with the rival Pāṇḍava and Kaurava cousins and go back no further
than their human if not always “real”9 fathers and grandfathers, it would
appear that these main heroes of the Mahābhārata grow up under a near-total
collapse of the paternal function. Back a generation to the “grandfathers” (hav-
ing that name whether as biological parents or just generationally), one finds
two sonless and dead before their time from martial (Citraṅgada) or libidinal
(Vicitravīrya) excesses, and two still living, both great spokesmen of dharma:
Vyāsa, usually a celibate and apparently unmarried, whose desire for dharma
mixed with desires for the sons he has sired seems to provide his explanation
for composing the whole poem; and Bhīṣma, whose vow of lifelong celibacy to
guarantee a second marriage for his own father makes him a nonbiological
grandfather who has perfectly mastered his desire, whatever the credibility of
that—or the consequences.10 From Vyāsa, then, they have two fathers and
a lower-rank uncle. Oldest is the Kauravas’ blind father Dhṛtarāṣṭra who
cannot really rule and who can never open his eyes to the raging desires of his
sons and nephews or gain wisdom on his own desires. His junior, Pāṇḍu,

9. In most of these cases the real father is human, but the Pāṇḍavas and Karṇa’s fathers are gods. This
complicates any use of Lacan’s distinction between Real, Imaginary, and Symbolic fathers.
10. On Bhīṣma from Freudian perspectives, see Goldman 1978; Fitzgerald 2007. Sax’s critique (2002,
78–92) of Goldman’s psychoanalytic interpretation in favor of a local ethnography of Rajput filial piety, which
Fitzgerald rightly rejects, has in any case little bearing on Bhīṣma’s epic complexities—not least because it has so
little to do with women.
 women’s dharma 343

after a short rule, is clearly the wished-for dead father, who has in fact died
fulfilling his desires during the heroes’ childhood. Once Pāṇḍu has died, the
Pāṇḍavas and Kauravas as parallel paternal cousins should accept paternal
authority from Dhṛtarāṣṭra, since he and Pāṇḍu are of the same rank, having
Kṣatriya mothers and the same Brahmin father, Vyāsa. But his authority is
weak, vacillating, and incurably biased so long as his sons are alive. Meanwhile,
not of the same rank and thus an uncle rather than a father, is Vidura, incar-
nation of the god Dharma, who can speak for dharma profoundly but only
ineffectively since as Vyāsa’s son with a Śūdra woman he cannot rule and has
no authority whatsoever. All this, in Lacan’s terms, would provide an ideal
situation for the Kauravas and Pāṇḍavas to make rival demands for their royal
patrimony.
If our heroes and villains grow up in such a world where their real or at
least human fathers and grandfathers comprise such an ideally dysfunctional
composite of the Symbolic father in whose name they would incarnate the tran-
scendent Law of the Father,11 we might not be surprised to find that it is a world
where a Law of the Mother could take hold, and that it will eventually take five
gods to sire sons in such a family to begin to restore a paternal Law. This would
not be interesting if the name of the father was always “Beyond the Mother,”12
leaving the mother, in situations of default, to be “the agent of the symbolic
father, who enforces the Law-giving and lineage-maintaining power of what
might otherwise be an empty paternal function” (Custodi 2005, 123). There is
some of that. But what the Mahābhārata seems to be wrestling with is that when
women hold the cards in speaking on dharma, and men must learn and listen,
situations may emerge where the Law and indeed the name of the Mother (and
when we get to it, of the Girl) may be vital to the Law’s realization and at least
equally as primordial as the name or Law of the Father.13 This, at least, gives us

11. “The Symbolic father is to be distinguished from the Imaginary father (often surprisingly distant from
the Real father) to whom is related the whole dialectic of aggressivity and identification. In all strictness the
Symbolic father is to be conceived as ‘transcendent,’ as an irreducible given of the signifier. . . . The real father
takes over from the Symbolic father” (Wilder 1968, 271). Cf. S. J. Sutherland 1991, 47: Gāndhārī, Kuntī, and
Mādrī “function almost as if they were the wives of one husband rather than two. After all, both husbands can be
said to be complimentarily dysfunctional.”
12. Cf. Lacan 1977, 311: “The fact that the Father may be regarded as the original representative of this
authority of the Law requires us to specify by what privileged mode of presence he is sustained beyond the sub-
ject who is actually led to occupy the place of the Other, namely, the Mother. The question, therefore, is pushed
still further back.”
13. As with the semiotics of women’s desire (see n. 6 above), this chapter reflects on Kristeva’s departures
from Lacan, with whom she studied, in her treatment of “the law before the Law,” the presymbolic, and the abject
mother (see Oliver 1993, 46–68). My own caution, however, has been to bear in mind that although Kristeva and
the Mahābhārata can be said to be looking for ways to talk about something similar, Kristeva starts with the child–
mother dyad whereas the Mahābhārata starts more with the child-wanting couple and the impressionable girl.
344 dharma

terms with which to consider the distaff side of this family through these three
generations that are rounded off, from this angle, with the five Pāṇḍavas’ legally
extraordinary marriage to Draupadī.
To begin the story of this family’s investment in a primordial Law of
the Mother, we must take ourselves back still one generation further to the
Pāṇḍavas and Kauravas’ two great grandmothers Gaṅgā and Satyavatī. The
best treatments of the skein (a term of choice14) as a whole are by Biardeau
(2002, 1: 212–38) and Brodbeck (2009a, 150–77), but each—for very different
reasons—tends to shortchange the women. Biardeau attends mainly to their
symbolic roles rather than their words, which leaves their relation to dharma
rather schematic, cosmological, and abstract (see above n. 1). And Brodbeck
zeroes in on the patrilineal males for whom wives are more interesting than
mothers, leaving him to admit that in treating female characters, his cate-
gories “cannot do them justice” (261).15 Meanwhile, with closer attention to
the women’s words and moves, claims have been made for two of the women—
Satyavatī (by Ghosh 2000, 33, 42) and Kuntī (by Dhand 2004)—that they are
pivotal for the whole epic, which they certainly are, though neither one more
than the other. Dumézil (1979, 31–45, 66–71), Doniger (1995), Jamison (1996),
and Heesterman (2001, 254–59) have also made important contributions on
the legal ramifications of several of the unusual marriages in the series.
A methodological point here: A. K. Ramanujan had a good impulse when he
criticized me for overemphasizing divine–human connections at the expense
of “the architectonic complexity of the human action of the epic” (1991, 434,
n. 4). But the point is valid only so long as one traces that complexity only in its

14. Along with the cloth (paṭa) (1.3.147, 167) woven by the female weavers Dhātṛ and Vidhātṛ in the epic’s
opening story of Uttaṅka, major Mahābhārata scenes are compared to pictures (citra) drawn on a woven cloth
(paṭa): the outbreak of fighting just after the Bhagavad Gītā (6.42.25); the pause in the battle on the fourteenth
night when the warriors, horses, and elephants sleep (7.159.40); Gāndhārī, Dhṛtarāsṭ̣ ra, and Vidura’s collapse
upon hearing that all hundred Kauravas have been killed (9.1.40); Bhīṣma’s closing silence on his bed of arrows,
having finished his immense dharma talk (13.152.1); and Dhṛtarāsṭ̣ ra, Gāndhārī, and others seeing their slain sons
rise from the Gaṅgā thanks to Vyāsa’s boon of “divine sight” (15.40.20). These usages for such vivid scenes may
anticipate the paṛ (a cloth depicting epic events) used in Rajasthani folk epics (Smith 1991, 5–9, 54–70, 504).
15. This is admitted with reference to “the putrikā/pativratā dichotomy” (Brodbeck 2009a, 261), by which
sons belong lineally to a woman’s father or to her husband (49–56). These categories do not pose overt opposi-
tions in our skein, in which Gaṅgā “is no putrikā” (155) and “her father is absent” (225); the overall intent, even
with the anomalies presented by Gaṅgā and Satyavatī, is to make “good lineal wives” (160) and find dutiful
pativratās (168). Yet Brodbeck does imply that Gaṅgā should have a father; hints at further possibilities in the
sufficiently enigmatic story of King Vasu Uparicara’s paternity of Satyavatī (163); and offers speculation linking
Kiṃdama (the Ṛṣi disguised as a mating deer whom Pāṇḍu killed) with Kuntī and Mādrī as potential putrikās,
which Brodbeck himself seems to reject (175–76). Brodbeck is less inclined to invent lineage connections and
merge characters’ identities in this skein than he is in discussing earlier and later generations—perhaps because
this is more inescapably the “full story.” In any case, he grants that “characters in this plot may seem slightly
independent of those surveyed above” (151), who lived in prior generations.
 women’s dharma 345

own human terms. Our epic’s human action also has cosmological complexity
(see Hiltebeitel 2001b, 267), as does its men’s and women’s words.
With Gaṅgā and Satyavatī on through Kuntī and the other women of her
generation, we find that women’s dharma is woven (sometimes literally so) into
an artful story of increasing crisis among the males in the Kuru dynasty, also
known as the Lunar dynasty due to its descent from the Moon god. In such cir-
cumstances, the men are no less interested in and representative of dharma
than the women, but what the women have to say about it becomes increas-
ingly decisive. Our chief attention will be on what the three generations of
women who marry into this lineage before Draupadī have to say about dharma,
and the moves they make regarding it.

C. Mother Gaṅgā

It all starts with a near trifle in outer space.16 “There was once a king born in the
Ikṣvāku lineage” (Mbh 1.91.1) named Mahābhiṣa who, after countless sacrifices
and a truthful life, became a Rājarṣi or Royal Sage in heaven. The poets are
about to introduce him to the luminous celestial Gaṅgā, her robe the Milky
Way, and their metaphoric range is the night sky, where Ṛṣis, royal and other-
wise, are stars (Mitchiner 1982), and there are mighty winds. The celestial
Gaṅgā is associated with the Parivaha wind that diffuses its waters and carries
them through the sky when “agitated,” affecting the visibility of the sun and the
rising moon (Mbh 12.315.46–48). “Then at some time” (tataḥ kadācit)—note
this cunning narrative convention, which occurs nine times in our skein and
serves as a sort of start-up and then restart mechanism—while the Gods were
doing homage to Brahmā in the Royal Ṛṣis’ company, as “Gaṅgā approached
the Grandfather [Brahmā], her garment, radiant as the moon, was raised by the
wind” (4). Now, as Gaṅgā’s garment lifts,

The host of gods then lowered their faces. But the royal Ṛṣi
Mahābhiṣa looked at the river fearlessly. Mahābhiṣa was disdained by
lord Brahmā, who said, “Born among mortals, you shall again gain
the worlds.” (1.91.5–6)

At this point, in a fairly widespread interpolation remarked on by Biardeau,

Brahmā also says that Gaṅgā will join Mahābhiṣa in this earthly destiny.
According to Biardeau, the two are jointly “condamnés pour une faute légère

16. This section draws some phrasing and insights from Hiltebeitel 2001b, which now reappears slightly
revised as Hiltebeitel 2011a, chapter 13. From another angle, see now also Hiltebeitel forthcoming-g.
346 dharma

dans le ciel” (2002; I: 219; cf. 213). But even with the interpolation, I do not
see this implication. Brahmā addresses only Mahābhiṣa:

Having stayed long among humans, you will obtain the beautiful
worlds, O low-minded one whose mind is seized (hṛtamanas) by
Gaṅgā. She indeed will act disagreeably in the human world. When
you get angry you will then be released from the curse. (1.911*, after
Mbh 1.91.6)

It seems that Brahmā only says that once Mahābhiṣa takes birth as Śaṃtanu,
Gaṅgā will act disagreeably to him; not that she is “condemned” to do so. In
any case, the Pune Critical Edition does well to make these lines superfluous,
for it is immediately apparent that Gaṅgā’s descent is voluntary and amorous:
“The river, best of streams, having seen the king fallen from his firmness
(dhairyāc cyutam), went away musing about him in her heart” (91.8). Brahmā
can be prudish, but he does not hold her accountable for the wind blowing up
her skirt. In fact, he is more often prurient than prudish, and could be punish-
ing Mahābhiṣa for ogling a woman as something Brahmā gets a reputation for
doing himself. If we look back from a Purāṇic perspective, there is an emerging
irony here, since in Purāṇic myth, Brahmā is often the lascivious one disdained
or punished for his gaze, which he directs most famously, but not only, at the
beautiful Pārvatī when she is marrying Śiva (see e.g., Dimmitt and van Buitenen
1978, 34–35, 171; Hiltebeitel 1999c, 68–76).
Mahābhiṣa is told he can choose his father, and picks King Devāpi (91.7).
He gives no reason, but in doing so he shifts from the Ikṣvāku or Solar Dynasty
to the Lunar Dynasty, perhaps to avoid rebirth in his own line which I believe
could be a problem (cf. Brodbeck 2009a, 153–55). Meanwhile Gaṅgā, resuming
her celestial path and still musing on Mahābhiṣa, comes across the eight Vasu
gods smitten with dejection, their figures bedimmed (9), which, they tell her,
results from their having been cursed by Vasiṣṭha, one of the Great Ṛṣis among
the Seven Sages of the Big Dipper, for having come too close to him (10–13)—
all of which seems to continue to have astronomical overtones.17 The Vasus say
Vasiṣṭha cursed them to be born in a womb, and, since they are unwilling to
“enter an inauspicious human-female womb,” they ask Gaṅgā to be their
mother (14–15). When she agrees and says they can choose their father, they
pick “Pratīpa’s son, King Śaṃtanu, renowned as law-abiding (dhārmikaḥ)” (16).
Since Gaṅgā knows that Mahābhiṣa has chosen this same Pratīpa to be his
father, the Vasus’ choice—assuming, as she must, that Mahābhiṣa will be

17. See Hiltebeitel 2001b, 270–71, 274–75. I pass over most of the details, but Vasiṣṭha’s hermitage is on
“a side of Mount Meru” (meroḥ pārśve; 93.6)—the “cosmic mountain” that defines the axis of the Pole Star.
 women’s dharma 347

reborn as Śaṃtanu—is much to her satisfaction: “Such is even my mind, sinless

gods, as you say. I will do his pleasure; that is your desire” (17)—and also hers.
Let us note that we are poised to see for the first but not the last time a woman’s
mind (matam) carry prevailing force in our three-generation skein. When the
Vasus insist that Gaṅgā “must throw his [Mahābhiṣa-Śaṃtanu’s] newborn sons
into the water so that our restoration will not take a long time,” she agrees
again, but with the proviso that Śaṃtanu will keep one son. Each Vasu then
imparts an eighth of his vīrya (energy/manliness/sperm) into a collective
deposit for “the son you and he desire” (19–20): who will be Bhīṣma. But, add
the Vasus, this son “shall not reproduce his lineage among mortals. Thus your
son will be sonless, despite his possessing vīrya” (20–21). With Gaṅgā’s
agreement (samaya) on this further point, the “delighted” Vasus pursue their
course (22). Bhīṣma’s sonlessness is thus stipulated even before his father’s
birth, not to mention his own, by a divine compact or agreement between the
Vasus and his Mother. If Mahābhiṣa cannot reenter the Solar line, he also can
make only a restricted contribution to the Lunar one.
The only usage of dharma, an augmented one, in this first adhyāya of our
skein is the description of the future Śaṃtanu as dhārmika, “law-abiding.” But
samaya is also a legal concept. It is used early by Āpastamba to define the very
“sources of Law,”18 and Gautama says, “The offspring belongs to the man who
fathers it, unless there has been a compact (samaya)” (G 18.9–10)—which, with
a twist, could explain why Śaṃtanu loses his first seven sons.19 At this point at
least, Gaṅgā seems to have brokered this deal to leave one son to her husband,
as both he and she (at least according to the Vasus) desire. And with this, she
now begins to enact dharma in both gestures and words. Appearing to Pratīpa
out of her own waters, she sits on his right thigh and invites him to make love
to her, telling him—and implying a point of dharma20—that “abandonment of
a woman in love is prohibited by the good” (92.3–5). Pratīpa knows his dharma,
but seeks other ways to satisfy it and her. First, he says, “I would not go out of
desire to another’s woman, nor, lovely one, to one not of my own caste. Know
that to be my righteous (dharmya) vow” (6). Of course he wants to know more
about her, so she reassures him that she is “not unapproachable in any way,”

18. Ā 1.1.1–2 uses it to describe “accepted customary laws” (sāmayācārikān dharmān) and “the authority for
their acceptance by those who know the Law” (dharmajñasamayaḥ pramānạ m). See chapter 5 at n. 28; Olivelle 1999,
xl, citing also G 8.11.
19. Cf. Ā 1.13.10; 3.18.13; M 8.218–21; 7.202. The Mahābhārata tells of many samayas, including the
Pāṇḍavas’ agreement to share treasures and their sleeping protocol with Draupadī (1.2.90; 1.204–5). Two of the
first mentioned are the primordial bet on a horse between Kadrū and Vināta, the mothers of snakes
and birds (1.18.5), and the prenuptial agreement between the husband and wife both named Jaratkāru (43.28).
20. This is the point made by the serpent woman Ulūpī to Arjuna, who is married and vowed to some kind
of spurious celibacy: that it is a “subtle” matter of dharma to satisfy a ready woman (1.206.26–33).
348 dharma

and moreover a virgin (kanyā) (7). Given what one soon learns about other
kanyās in this skein, virginity can be renewable, and if Pratīpa’s mention of
“another’s woman” might remind us of anything, it would be that—at least in
post-epic Hinduism—Gaṅgā is the second wife of Śiva.21 Pratīpa, however, con-
tinues to demur: “I abstain from this pleasure to which you press me lest my
violation of dharma would destroy what I vowed” (8). But he can still suggest a
way to solve the dharma-problem her appeal seems to present: he invites her to
become his daughter-in-law because she chose to sit on his right thigh, suitable
for children and daughters-in-law, and avoided his left where a wife or lover
(kāminī) would sit (9–11). Agreeing to what she must have herself foreseen, and
thereby virtually assuring this apparently equally shrewd old king a son they
both desire, Gaṅgā says,

So be it, dharma-knower. May I unite with your son. So by devotion

to you will I love the famous Bhārata lineage (kulam). Whoever are
the kings of the earth, you (plural: your dynasty) are their refuge.
I am unable to speak the qualities (guṇas) that are renowned of your
lineage in even a hundred years; its strictness is peerless (tat
sādhutvam anuttamam). (12c–13)

Now telling Pratīpa he must tell his son that she must never be questioned,
Gaṅgā disappears (14–16).
Even though Pratīpa and his wife are old, he “burns tapas.” And “at a
certain time” Śaṃtanu is born in terms that both recall his recent celestial
identity and give what seems to be a double etymology for his new name:
“Mahābhiṣa became the old couple’s son. He was born the continuity (saṃtāna)
of a peaceful man (śāntasya); therefore he was called Śaṃtanu” (92.18).
Śaṃtanu’s name thus introduces a “continuity” theme that we shall see
unfold,22 and rather ironically, since this heavenly migrant from the Solar
Dynasty’s own continuity in this line, though he does not yet know it, will go
no farther than Bhīṣma. Yet as “Bhīṣma Śāṃtanava,” Bhīṣma will keep his
father’s name as a patronymic while affecting his family’s continuity in nearly
every imaginable way other than by paternity. Indeed, this double etymology
would seem to have been designed less for Śaṃtanu than for Bhīṣma, who will

21. Hopkins [1915] 1969, 6 has two Mahābhārata references to Gaṅgā being Śiva’s wife, but they seem to
come from interpolations. Śiva does break Gaṅgā’s celestial descent at Mbh 3.108.9–14 and 5.109.6, from which
their later (?) marriage results.
22. Cf. Brodbeck 2009a, 157 and n. 15; there is another etymology for Śaṃtanu’s name at 1.90.47b–49.
Within our skein, from this point to the death of Pāṇḍu, there are fourteen of the Mahābhārata’s forty-eight
usages of saṃtāna, seven with kula and one with vaṃśa, and there is no other such cluster. The Rāmāyaṇa uses
saṃtāna only five times, two with kula.
 women’s dharma 349

be preoccupied not only with his line’s continuity but, at peace himself on a
bed of arrows, with the “pacification” (śānti) of Yudhiṣṭhira in his vast postwar
sermon on what we might call the continuity of dharma. Just to keep these
ironies before us, when we come, shortly, to see that his name Bhīṣma will
mean “The Terrible” for his choice not to continue his dynastic paternal line
by siring sons himself, we might take his name Bhīṣma Śāṃtanava to suggest
“The Terrible ‘Continuator.’”
Śaṃtanu now becomes a young man, and Pratīpa describes the beautiful
divine (divyā) woman who may approach him and the conditions under which
she will stay with him, saying they are to be carried out “at my appointment
(man-niyogāt)” (21–23). We must note our skein’s first usage of niyoga, since,
even though it does not mean what it will be all about by its end—that is, the
proxy siring by a man who replaces a deceased or otherwise disabled husband—
the terminology seems to have been adroitly woven into the skien throughout.
But for now, having made Śaṃtanu his heir-apparent, Pratīpa leaves for the
forest. Soon, while hunting “along the Siddha- and Caraṇa-frequented Gaṅgā”
(92.25cd), Śaṃtanu saw:

a superb woman whose figure had an intensive glowing

(jājvalyamānāṃ vapuṣā) that was like the splendor of a lotus, faultless
everywhere, with nice teeth, adorned with divine ornaments, wearing
a subtle cloth, alone, and radiant as the calyx of a lotus. . . . As if
drinking her with his eyes, the king wasn’t satisfied (pibann iva ca
netrābhyāṃ nātṛpyata narādhipaḥ). (92.26–28)

Śaṃtanu resumes the gaze that got him into trouble as Mahābhiṣa—and
indeed, to drink this woman with his eyes and not be satisfied could be a
reminder that he last saw her as a river of stars. Moreover, “having seen the
king of great radiance going about,” she approaches him “dallying” (vilāsinī)
and with “fondness come with affection,” as her dissatisfaction matched his
(29).23 So he asks her to marry him. Remembering now her samaya with the
Vasus, she sets her requirements: the marriage will be “at his will,” but she will
do what she wants, and “I am not to be stopped or harshly addressed; . . . if you
speak obstructively or unkindly I will no doubt leave you” (30–36). For him, it
boils down to this: she is not to be questioned lest she abandon him. For her it
is certainly a kind of women’s law, but not one that many could impose, with a

23. Both “were not [yet] satisfied” (nātṛpyata) by having seen each other. On Lacan’s view of the objecti-
fying disembodied male gaze and the possibilities these two scenes open for seeing in Gaṅgā’s musing and
coyness an exemplum of an embodied female gaze charmed by male vulnerability and “the sense of seeing-
oneself-being-seen,” see Custodi 2005, 165–78.
350 dharma

codicil for a wife’s justifiable abandonment—something opposed by Manu,

who says, as we have seen, “She must never want to separate herself from her
father, husband, or sons” (5.149). As their joys unfold,

by attachment to pleasure, the king, seized [as Brahmā’s interpolated

curse predicted] by the qualities of this foremost woman, was not
aware of the many years, seasons, and months that passed. (41)

In “not so long a time” the eight Vasus are born, and Gaṅgā throws the first
seven into the water, saying “I fulfill your wish.” Śaṃtanu watches, saying
nothing “from fear of abandonment (tyāgād bhītaḥ)” (43–45). Here we may
have the seeds of his eventual longing, once he is abandoned, for another wife,
for clearly Gaṅgā, through the first seven, puts her motherly samaya before any
concern for him.
At last the eighth son is born and Śaṃtanu can stand it no longer. “Yearning
for a son of his own,” he speaks out while she seems to be laughing and says,
“Son-killer, Stop!” to which she replies, “I will not slay your son, but this stay is
now exhausted according to the samaya we made” (46–48). Gaṅgā now reveals
who she is: “I am Gaṅgā, daughter of Jahnu, frequented by the hosts (gaṇas) of
Great Ṛṣis; I have dwelt with you for the sake of success in accomplishing a
purpose in the work of the gods (devakāryārthasiddhyartham)” (49). Although
Gaṅgā emerged from her “Siddha- and Caraṇa-frequented” earthly course to
sit on Pratīpa’s right thigh (25), she now speaks of the heavenly “hosts of Great
Ṛṣis” who frequent her heavenly course. Her “success in accomplishing a
purpose in the work of the gods” is a quite precise and early indication that the
“work of the gods,” which we may call the Mahābhārata’s divine plan, begins to
unfold with a certain complexity. That is, it involves not only the gods and Ṛṣis
but this celestial goddess,24 and it will have to take in more than one generation.
And it could—without it being made explicit—be thought to coincide with the
yugas. The Mahābhārata’s divine plan is thus far more fluid and extensive than
the Rāmāyaṇa’s, which concerns (at least as Vālmīki describes it) mainly just
male gods and Ṛṣis, and does little to trace a course of dharma over genera-
tional time. Outside its yuga references, the Rāmāyaṇa traces dharma over time
mainly through only the one generation of Rāma and his brothers, though it
hints at some laxity of dharma in the long life of Daśaratha,25 and describes a
lengthy course of adharma over time through the long career of Rāvaṇa.

24. Cf. Biardeau 2002, 1: 221: All this would have a “fonction avatārique” since Gaṅgā is doing devakārya,
“ce qui est à faire pour les dieux”—which I translate as “the work of the gods.”
25. See Doniger 2009a, 225, on Daśaratha’s propensity to kāma over artha and dharma, and Rāma’s criti-
cism of this.
 women’s dharma 351

More immediately, we now see that what Gaṅgā has accomplished under
the heading of “work of the gods” has fulfilled a samaya with the Vasus brought
on by a curse by one of the most Vedic of the Great Ṛsị s, Vasiṣṭha. In fact,
although Gaṅgā has just spoken of her prenuptial samaya with Śaṃtanu, there
was no mention of that or any other such legal term during their premarital
understanding. She had shaped that understanding, as she tells him only now,
to be an extension of her samaya with the Vasus: “But this samaya was made
between me and the Vasus—that as soon as they were born, I would release
them from human birth,” leaving it unclear how this affects the one exception
just born other than to say that she has now freed all the Vasus from Vasiṣṭha’s
curse (53–54). Śaṃtanu would probably be too bowled over to catch his wife on
this technicality, but what good would it do him anyway? Instead he asks to
learn more about Vasiṣṭha and the Vasus, and how their contretemps affected
the son he must think he and he alone has just rescued from oblivion (93.1–3).
What Gaṅgā tells him is this: his son is the incarnation of the god Dyaus (Sky),
who was cursed by Vasiṣṭha to take birth in a womb because, as the Vasus’ ring-
leader, he led them, at his wife’s request,26 to abduct Vasiṣṭha’s divine cow (26);
and that, although Vasiṣṭha shortened the terms for the other Vasus, Dyaus was
cursed to “dwell in the human world for a long time by his own karma
(svakarmaṇa)”̄ (36cd; cf. 42). This would imply that Dyaus’s karma will carry
over into this human life, where we might have occasions to wonder whether
traces of Dyaus’s misadventures—foolishness done for a wife, abduction of a
cow and calf—might have been left on Bhīṣma. Gaṅgā then comes to the closing
words of Vasiṣṭha’s curse, some of which sounds good: “. . . He will be a soul of
dharma, conversant with all the scriptures”27 (39ab); while some is bound to be
unsettling to Śaṃtanu: “The high-minded one will not reproduce among
humans . . .” (38cd)—which Gaṅgā repeats not from Vasiṣṭha28 but from her
samaya with the Vasus (91.21–22). Even more troubling, “Devoted to the welfare
and his father’s pleasure, he will forsake the enjoyment of women” (39cd).
This we are hearing for the first time. One cannot avoid the impression that
Gaṅgā is a little free with her sources. Having finished attributing things to
Vasiṣṭha, she offers a brief self-exoneration for throwing the other boys into the
river for the sake of their release (mokṣar̄ tham) from the curse, and upon that,

26. This unnamed wife does not seem to have any trait that would identify her as the Pṛthivī—Earth,
Dyaus’s Vedic wife. Indeed, note that the Vasu who most helped Dyaus steal the cow was Pṛthu, Earth in the
masculine (93.26).
27. The Critical Edition follows the Southern Recension with śāstras (scriptures) rather than śastras
(weapons).
28. Earlier, at 91.12, all Vasiṣṭha said was “Be born in a womb,” and that is all that Gaṅgā has reiterated so
far (92.50; 93.31).
352 dharma

“the Goddess disappeared right there” (43) taking the boy with her. For Gaṅgā to
vanish (antar-adhīyata)—literally, “to put herself within”—is to return to her own
element, whether it be water or space (ākāśa), since she is of course the Ākāśa-
Gaṅgā. In going with her, Bhīṣma’s disappearance is almost like the drowning of
his brothers. But Śaṃtanu knows Gaṅgā has taken him away with the promise of
a long life ahead of him. Śaṃtanu goes “back to his capital afflicted with grief”
(44). Having finally spoken out to keep his eighth son even though he knew it
would mean losing his wife, he has suddenly lost them both.
It is now, with these events behind him, that the narrator Vaiśaṃpāyana
says he will speak of Śaṃtanu’s endless guṇas (93.45d) and “the splendid itihāsa
called Mahābhārata” (93.46cd). The Mahābhārata’s “history” begins with
Gaṅgā’s departure, yet also with her ongoing blessing: for thanks to her “devo-
tion” to Śaṃtanu’s father Pratīpa, she has promised that she will “love the
famous Bhārata lineage” whose guṇas she is unable to recount in “even a hun-
dred years” (92.12c–13). In effect, from a heavenly story moved down to earth,
the Mahābhārata will stay largely on earth. After Śaṃtanu’s “lost time” with
Gaṅgā, time gets condensed into charted time along the epic’s flow, beginning
soon with the return of Bhīṣma. The verse on “itihāsa” ends an adhyāya. In
effect, “history” begins directly after this in the next adhyāya with the story of
Śaṃtanu’s second marriage to Satyavatī, who is already the mother of Vyāsa,
the author.29 How better to begin an “invention” of “history” than by the with-
drawal of the celestial Gaṅgā, whose very intervention has resolved a crisis in
the genealogy that will eventually bring forth—indeed, make possible the births
of—the Kauravas and the Pāṇḍavas?30
During his wifeless rule with his son swept away, Śaṃtanu makes what he
can of dharma, but with a sense of mounting insecurity. A “soul of dharma”
(dharmātmā; 94.1c), he regulated the three puruṣārthas in favor of dharma (3–4);
he performed his Kṣatriya svadharma like none other (5–6b); he supported the
ordered relation of the four classes with Brahmins above and with each class
serving the superior ones with increasing devotion down the scale (8–9; 14ab);
he ruled the earth with knowledge of dharma and attained prosperity through
gifting, dharma, tapas, and truth (10–11; 17); and he kept dharma and the brah-
man foremost in his kingdom and saw to it that no breathing creature was slain

29. With typical recursivity, the Mahābhārata has begun that “prehistory” earlier with the story of Vasu
Uparicara (1.57), which, at Mbh 1.1.50–51, is introduced as one of the three starting points from which some
Brahmins learn the epic—the other two being “from Manu onward” and “from Āstīka onward.”
30. Cf. Brodbeck 2009a, 158 n. 18, resisting such an “astronomical cosmological interpretation” while
attending to genealogical matters, which the Mahābhārata actually subordinates to its divine plan and its cosmo-
logically worked out sense of “history.” Indeed, Gaṅgā’s intervention in the Bhārata genealogy is similar to the
“descent of the Gaṅgā” (gaṅgāvataraṇa) in the Rāmāyaṇa, where she solves a genealogical crisis of the Ikṣvāku
lineage posed by the disappearance of the sons of Sāgara. On these points, see Hiltebeitel forthcoming-g.
 women’s dharma 353

by adharma. Moreover, “rites then began (ārabhyanta tadā kriyāh)̣ for the sake
of sacrifice to Gods, Ṛṣis, and Fathers” (13–15)—as if they had been interrupted.
Indeed, this seeming resumption in fulfilling the “three debts” introduces a
note of unease. What is this king’s lineage? No doubt his rites would be for
Fathers in the Lunar line he has shifted into, but what about this son who, even
if he comes back, will not be able to continue them by having a son himself? All
this time, for thirty-six years, Śaṃtanu “did not obtain pleasure with women”;31
still ruling, he then became a forest rover (18).
Vaiśaṃpāyana now brings the flow of time back to the river: “At a certain
time (kadācit),” hunting while “following the river Gaṅgā, Śaṃtanu saw that
the Bhāgīrathī had little water” (94.21). This use of the river’s name
Bhāgīrathī (omitted from van Buitenen and Ganguli’s translations) evokes a
connection between Śaṃtanu and the Solar-line king Bhagīratha, who first
brought the heavenly Gaṅgā down to earth (Mbh 3.107–8; Rām 1.41–43.6).
Having now also brought Gaṅgā down to earth once to marry him, Śaṃtanu
will meet Bhīṣma arresting her descent. Wondering why “this best of streams
no longer flows swiftly, he saw the occasion (nimittam)”: a stunning youth
using a “divine weapon” to block the entire Gaṅgā with heavenly arrows
(Mbh 94.22–24)!32 Śaṃtanu does not yet remember the son he had seen only
at birth, who has now learned how to bewilder his father like a God or Ṛṣi by
disappearing or “putting himself within” (antaradhīyata) (26–27). Suspecting
this was his son, Śaṃtanu invokes Gaṅgā to appear, as she does: better
dressed this time (“adorned with ornaments, attired in a dustless garment”)
and holding the boy in her right hand (28–31). She tells Śaṃtanu who she is
and what their son has accomplished in his time with her: he has studied the
Vedas and their limbs from Vasiṣṭha; the same śāstras as Uśanas and the son
of Aṅgiras (i.e., Bṛhaspati); all weapons known to Rāma Jāmadagnya; and he
is “skilled in rājadharma and artha” (32–36). Though the poets do not over-
state the matter, Bhīṣma has clearly been brought up by the celestial Gaṅgā
and educated by Gods and celestial Ṛṣis. She has taken him up to the stars,
where he could have learned Veda from a pacified Vasiṣṭha, last seen on the
side of Mount Meru cursing Mahābhiṣa and Dyaus to become our father–
son pair; and other celestial stalwarts have taught him the texts and topics

31. Ratim aprāpnuvan strīṣu; 94.18c. The alpha privative has been overlooked by van Buitenen, 1975, 223
and Ganguli 1: 230.
32. See Brodbeck 2009a, 159 for some interpretive options, among which I stick to mine and Randy
Kloetzli’s, that the eternal river of time is momentarily arrested. I appreciate also Brodbeck’s view that it signifies
Bhīṣma’s “movement back to his father’s house,” but see as counterproductive his notion that Gaṅgā’s “father is
absent” (224). She clearly interrupts the lineage as a goddess, and her father Jahnu has not been shown to be at
all relevant genealogically.
354 dharma

just mentioned, including rājadharma—about which he will expound in his

first postwar anthology on that topic, where he will be more punctilious in
citing his sources than his mother.33 Clearly the “work of the gods” that we
have heard about first from Gaṅgā will have this carry-over on past the end
of the Mahābhārata war. More immediately, though, this training has pre-
pared her son to be the heir-apparent, which Śaṃtanu now makes him (38ef),
even though he has been given the worry that he will be without issue.
Meanwhile, for once there is no mention of Gaṅgā disappearing, leaving an
impression that she has returned for now to being mainly a river. It is the
last one sees her in this skein.

D. Mother Kālī Satyavatī

Four years pass with father and son enjoying each other’s company until, “at a
certain time, going into a forest along the Yamunā River, the king of the earth
scented an indescribably lofty smell” (Mbh 1.94.41). Same restart mechanism,
different river. Śaṃtanu is about to meet a woman introduced earlier in an
account focused on her becoming the mother of Vyāsa. Outside the capital of
King Vasu Uparicara of Cedi, Mount Kolāhala made love to the Śuktimatī River.
King Vasu released the river with a kick, but she had become pregnant with
twins, whom she gave to the king in thanks for freeing her. Vasu made the boy
his marshal and married the girl, Girikā. One day Girikā lovingly told Vasu she
could conceive and did the ritual bath to get a son. But this plan was strangely
interrupted by Vasu’s Fathers who, although they were “pleased”—presumably
at the prospect of a son—told him, “Kill deer!”34 “Unable to transgress the
Fathers’ appointment (niyoga), he went hunting, lovingly musing on Girikā”
(38). In this reverie he ejaculated, and to keep his wife’s readiness from going
to waste, caught the sperm on a leaf, spoke a mantra over it, and, “aware of the
subtleties of Law and Profit,” gave it to a hawk to speed it to her. Another hawk
who thought it was meat attacked the carrier and the sperm fell into the
Yamunā, where an Apsaras named Adrikā, cursed by Brahmā to become a fish,
swallowed it. Fishermen (matsyajīvinaḥ) then caught the fish, pulled out a pair
of twins from it, and told the king. Vasu kept the boy, who became King Matsya,
but since the girl smelled of fish he gave her to the fisherman (dāśāya), saying,

33. Nārada attests to a similar list of Bhīṣma’s sagely sources at 12.38.7–13, when Bhīṣma gets the divine
eye toward the beginning of the Rājadharmaparvan. See Hiltebeitel 2001b.
34. See van Buitenen 1973, 47: “The reason of this demand is not clear: for purposes of offering?” A deer
hunt will figure again in Pāṇḍu’s incapacitation to satisfy his ancestors.
 women’s dharma 355

“Let this one be yours!” Named Satyavatī, she had every virtue (guṇa), but “from
dwelling amid fish-killers (matsyaghāti-), she bore for some time a fishy smell.”
Out of obedience35 to her father, she plied a boat on the water (53–55), and there
one day she caught the eye of the Ṛṣi Parāśara, grandson of Vasiṣṭha, who
seduced her—by arranging for a fog to keep the Ṛṣis on the banks from seeing
them; by promising to restore her virginity so she could return to her father;
and by giving her the boon of smelling good whereby her new scent (gandha)
won her the new names Gandhavatī, “the Fragrant,” and Yojanagandha, “She
whom you could smell a league away” (67). She was “thrilled” with the boon,
and all in all she seems quite the capable negotiator (Ghosh 2000, 35–37). That
same day after Parāśara continued on his way, Vyāsa was born on an island
only to depart that very day himself, his mind set on austerities, with the words,
“Remembered, I will appear when things are to be done” (70cd; 1.57.31–71).
Here Julia Kristeva gives us tools with which to register a useful contrast: If
Gaṅgā is the “sublime mother” whose son Bhīṣma will not be able to separate
from her or relate sexually to other women,36 Satyavatī is the “abject mother”
whose first son separates from her instantly to practice tapas, with its overtones
of gynophobia.37
With this story kept for now as a secret of her past, Satyavatī’s “lofty
smell” enables Śaṃtanu to trace her, and he learns from her, “I am a fisher-
maid; for the sake of dharma I ply a ferry at the appointment (niyogāt) of my
father, the high-souled king of fishermen” (1.94.44). But what is plying a
ferry for the sake of dharma? Biardeau rightly points us ahead to descrip-
tions of Draupadī as the boat that saves the Pāṇḍavas from the dice match
by her question about dharma.38 But this ferryboat is also legal: her “father
who was adjoined to dharma had a ferry” (99.6), and, it seems, appointed
her its captain. And what then is a king of fisherfolk? Let us note that this
adoptive father speaks in a legal manner when he uses terms like niyoga and

35. Śuśrūṣa (obedience, literally “liking to hear”) is the characteristic of women and Śūdras (see chapter 5 § E).
Satyavatī would presumably have been raised as a Śūdra.
36. See Oliver, 1993, 64–65: “Kristeva takes us back to the milky way of the primary dyad. She analyzes
the pulses and jolts of this primary universe made up of only mother and child” (64–65). The heavenly (ākāśa)
Gaṅgā is the Milky Way. Though I know of nothing to tell us that it is milky, it may have some relation to
the Milky Ocean (kṣīroda) that is above Mount Meru and has a northern shore where Brahm nā,
the Gods, and Ṛṣis go, seeking the welfare of the worlds (Mbh 12.322.8; 327.39). See chapter 6 § C on this
Nārāyaṇīya cosmography.
37. See Oliver 1993, 61. I do not push Vyāsa too hard on gynophobia, but a case can be made; see Dhand
2004, 44–45.
38. See Biardeau 2002, 1: 201. Tārī as “ferryboat” means “something that carries across” to the other
shore, “saves.” Cf. Biardeau 219: Satyavatī’s “link with the world of fish places her near the origins and she puts
into the world the essential components of good order.” Biardeau takes her name to mean “She who has within
her all beings.”
356 dharma

now dharmapatnī, “legal wife,” and samaya. Wishing to marry Satyavatī,

Śaṃtanu learns that the fisher king has “a certain desire”:

If you seek her from me as your dharmapatnī, sinless one, then make
a samaya with me by truth, as you are of truthful speech. I will give
this daughter to you by the samaya, king, for there will surely never
be another suitor like you to me. (94.47–49)

His requirement is that her son be king (51). Burning with desire, Śaṃtanu says
he must think about giving him this “boon” (varam; 50, 52, 56), which Bhīṣma
will later call a “bride price” (śulka; 97.14).
Bhīṣma soon sees his father morose and asks why. Śaṃtanu does “not seek
vainly to marry another wife,” but still, he says, “I wish for non-destruction of
continuity” (saṃtānasya–avināśāya; 94.59a–d)—reminding us, with this second
mention of saṃtāna, “continuity” or “succession,” that the first usage built the
term into Śaṃtanu’s own name (92.18). Quoting an adage: “Those conversant
about dharma say, ‘Having one son is to be sonless’” (94.59ef), he fears he will
be left sonless if Bhīṣma dies fighting, as is his wont. But we know he has other
good reasons to worry. What is he to make of a son who will be “devoted to the
welfare and pleasure of his father” and “forsake enjoyment of women”?
“Pondering” his father’s demurral, Bhīṣma learns about the fishermaid from an
“old councilor” of his father’s and goes with some “old Kṣatriyas” to her father.
The meeting “in an assembly of kings” (rājasaṃsadi, 94.68) credits the fisher
king’s royalty and gives further legal force to the proceedings.39 The fisher king
establishes that Satyavatī, no mere ferryboat pick-up girl,40 is “the offspring of an
Ārya who is your equal in qualities, from whose sperm (śukrāt) she appeared”;
moreover, “Your father has been praised by him to me frequently, lad: ‘He
among all kings is worthy of marrying Satyavatī’” (71–72). Bhīṣma would cer-
tainly gather that Satyavatī has a royal father who has delegated this man to
arrange this very match41—and more, that the fisherman has somehow learned
that Vasu was behind her birth from a fish, and has remained in touch with that
king of Cedi, a kingdom located along the southern bank of the Yamunā42 where
Satyavatī washed up with her original fishy smell. If we go back to her birth
story, we can hardly be prepared for these intimations. Nothing was said about

̣
39. Bhīṣma will make his vows “while the earth-protectors were listening (śṛnvatā m bhūmipālānām)” (77c;
cf. 83c, 86c). Cf. Brodbeck 2009a, 162: Satyavatī’s “adoptive father seems to know her true paternity.”
40. Van Buitenen 1973, 447 imagines the father to be “the king of a probably aboriginal fisher tribe” to
square the two stories and make Satyavatī “obviously a fishing and ferrying wench for the taking, [who] is
obviously legitimized as a king’s daughter.”
41. Cf. Biardeau 2002, 1: 214: the fisher king seems to know his daughter’s royal origins.
42. See van Buitenen 1973, 446: “south of the rivers Yamunā and Carmaṇvatī.”
 women’s dharma 357

one of the fishermen being a king, and Vasu showed no interest in the girl’s
marriage prospects, much less her origins from him. Jayatri Ghosh wishes to
make the tale more plausible: “evidently, King Uparicara was attracted to a fish-
erwoman, had children by her, and to spare Queen Girikā’s feelings, the event
was transformed into a fantasy” (2000, 34). But clearly, neither Satyavatī nor her
adoptive father needed to produce credentials at that time. And now these go
unquestioned by Bhīṣma who, as his father’s matchmaker, meets the fisher-
man’s two demands. He renounces kingship to guarantee Satyavatī a royal son
(79); and to see that he will sire no other rival, he vows celibacy even as he says
he will obtain imperishable worlds in heaven (divi).43 All this is said to the
celestial approbation “of the Apsaras, Gods, and the hosts of Ṛsị s (ṛsị gaṇas),”
and earns him the name Bhīṣma, “the Terrible” (86–90).
Speaking on his father’s behalf, Bhīṣma now says, “Ascend the chariot,
mother. Let us go to our homes” (94.91). Under conditions brought about by
the samaya his own mother, Gaṅgā, made with the Vasus, Bhīṣma now makes
the samaya his father left pending in the name of this new mother, Satyavatī.
Śaṃtanu thus enters into both marriages by restrictive samayas made on behalf
of his wives as mothers-to-be; samayas that in each case introduce a super-
vening Law of the Mother that must rise to the occasion of a lineage whose
“continuity” has been disrupted in the person of these two men cursed into
being born into it: Śaṃtanu himself, cursed by Brahmā for gazing fearlessly (in
his former life as Mahābhiṣa) up the skirt of mother Gaṅgā; and Bhīṣma
Śāṃtanava, cursed (in his former life as the Vasu Dyaus) by Vasiṣṭha for steal-
ing his cow. What can be said of this doubled Law of the Mother? Clearly the
two women speak in some fashion together for something primordial that is
deeper than kingship. Their two rivers embrace the “Mesopotamia” (doab) of
“the north of Madhyadeśa,” that exemplary land of dharma.44 They speak for a
primordial dharma sanctioned by the gods and celestial Ṛṣis. One, a goddess,
descends from a celestial river of light; the other, born from a fish, comes from
the terrestrial waters of the always darker Yamunā and is “dark” herself like her
instant-Ṛṣi son, the Dark Island-born Vyāsa (Kṛṣṇa Dvaipāyana Vyāsa). One is
called “child-killer,” but where she kills, it is with waters of liberation; the other
is born bearing the smell of the killing of fish, suggesting something to do with
the Law of the Fish (matsya nyāya) and the ongoing life–death struggle of trans-
migration (saṃsāra), as well as a boat across its waters.45 With Gaṅgā, the first

43. Since the sonless have woeful destinies, this could suggest that Bhīṣma has an intimation of his having
been Dyaus, “Father Sky,” or at least of having been there. Dyaus belongs to the same noun stem (div/dyu) as
div[i]. My thanks to Stephanie Jamison for reinforcing this point (e-mail, June 8, 2007).
44. On madhyadeśa, as last cross-referenced see chapter 5 n. 34, and further chapter 7 §§ A.2, B.2.
45. On these points, cf. Biardeau 2002, 1: 218.
358 dharma

thing this Law of the Mother ameliorated was the punitive character of two
male curses. Gaṅgā was more charmed than bothered by Mahābhiṣa’s imperti-
nence, and, given a situation where she could do no more “in person” than
“love the Kuru dynasty” through two errant generations, perhaps she made up
for the harshness of her first seven maternities by loving Śaṃtanu and by
raising their one surviving son among Ṛṣis to be the line’s highest authority on
dharma. Now Satyavatī, definitely more charmed than bothered by the atten-
tions of the Ṛṣi Parāśara, continues to speak for dharma solutions to the dynas-
ty’s disruptions (cf. Biardeau 2002, 1: 218).
Satyavatī’s marriage to Śaṃtanu is over in four verses (1.95.1–4), barely
enough to mention that they have two sons, and to bring closure with, “the wise
king Śaṃtanu succumbed to the Law of Time (kāladharmam).”46 The next verse
sets the tone and terms for much that follows. It says that when Śaṃtanu went
to heaven and the oldest of the two princes, Citrāng ̇ ada (and note: we have
nothing to tell us whether Citrāṅgada ever married) was enthroned, Bhīṣma was
“settled on the mind (or thought) of Satyavatī” (satyavatyā mate sthitaḥ; 5d). This
half-line becomes a refrain: first, after the extravagantly martial Citrāng ̇ ada’s
untimely death, when it is a question of Bhīṣma governing as regent while 47

Satyavatī’s second son, Vicitravīrya, is still a child (96.1d); and again, immedi-
ately after the extravagantly amorous Vicitravīrya’s untimely demise, when the
continuity of the now thoroughly defunct line will finally call for the extraordi-
nary measure of proxy siring or niyoga (59d). Be it noted that Satyavatī produces
sons whose rules are marked by conquest and desire, but none whose rule is
marked by dharma—a pattern that will be replicated in her three grandsons. As
Ghosh observes, the “reiteration of the phrase satyavatyā mate sthitaḥ underlines
the acknowledgment of her decision-making role” (2000, 40). As we follow the
mind of Satyavatī through these refrains and beyond, two other points bear
watching: first, while the phrase “settled on the mind of Satyavatī” suggests that
Bhīṣma yields to her, it also means that he and she repeatedly put their minds
together; and second, that Vyāsa has told this same mother—his own real
mother—that all she has to do is “remember” him (from ÷smṛ, which often har-
bors feelings of love, longing, and desire) should something need to be done. As
the author, Vyāsa knows his mother’s mind even from afar, and with it her
desires. Perhaps by disappearing at birth with such a promise of return, he
could go for a time—to requote Lacan—“beyond the Mother, the real Other of

46. The phrase kāladharmam upeyivān is formulaic: see 1.70.46d for Yayāti; 12.31.45 for “Excretor of Gold”;
Rām 1.41.9. Kāladharman is also common in the instrumental singular, as with the death of Pāṇḍu (Mbh
1.116.12d).
47. Later, Kṛṣṇa mentions to Yudhiṣṭhira that Bhīṣma told Duryodhana that, when he installed Vicitravīrya
as king, he “became his retainer below him” (bhṛtyo bhūtvā hy adhaś caraḥ; 5.145.21d).
 women’s dharma 359

demand, whose desire (that is, her desire) one wishes she would assuage.” But
of course a mother’s desires may be beginningless and unending.
Just after we hear for the second time that Bhīṣma was “settled on the mind
of Satyavatī” (1.96.1) in his regency during Vicitravīrya’s childhood, we learn
that as soon as the boy matured, Bhīṣma “set his mind (akaron matim) on
Vicitravīrya’s marriage” (2cd). Bhīṣma is minded to abduct three princess-brides
for this ineffectual half-brother of his, and clearly initiates the plan—although as
he sets off armed on a single chariot, it is with his (or their) “mother’s consent”
(anumate mātuḥ; 6). Let us mark this as the first point in a two-way concurrence,
for Satyavatī will similarly ask Bhīṣma’s consent (tava hy anumate; 99.17a) when
she tells him her last idea on how to revive the extinct line once Vicitravīrya is
dead. Etymologically we may say that their two “minds” (mati) lean “toward”
(anu) each other.48 By the time Bhīṣma has completed his “superhuman feat”
(karma-atimānuṣam) of abducting the three princesses and is ready to give them
to Vicitravīrya, we are told that he “made the decision together with Satyavatī”
(96.46). Here we come back to a point made earlier that has now become more
complicated. Bhīṣma “Śāṃtanava,” who carries within his patronym the over-
sight of his family’s “continuity,” began this project by spontaneously making
two vows that complied with something fated for him by his mother Gaṅgā’s
samaya with the Vasus so that he could clinch the samaya by which his father
could marry a second wife. Now this second mother, his father’s widow, has
become his handmaiden in the project of continuity, which has begun to take on
the character of a kind of weave of mutually improvised dharma. Two things
about this dharma: it gets rough and discordant; and Bhīṣma and Satyavatī soon
find themselves co-improvising it with the author.
The abduction of the three princesses sets a rough and discordant tone
from the start. The three are daughters of the king of the Kāśis (kāśipati, kāśirāja)
whose city is Vāraṇāsī (Benares; 1.96.3–4), which is apparently a capital of the
kingdom of Kosala.49 Although the princesses are being given a svayaṃvara (3)
or “self-choice” betrothal ceremony in which they would have some choice in
selecting their husbands from among the assembled kings, Bhīṣma quite liter-
ally subverts the ceremony, at least according to some wordplay by Vaiśaṃpāyana:
“But when the names of the kings were proclaimed by the thousands, the lord
Bhīṣma then chose50 those [maidens] himself (svayaṃ . . . varayām āsa)” (6)!

48. Unlike English “consent,” “permission,” or “concurrence,” anu + mati implies an assenting mind.
Anumati is one of the four Vedic lunar phase goddesses.
49. See Biardeau 2002, 1: 220–21; 225 and n. 2. Kāśi might not yet be a name for Vārāṇasī, and it or its
people may belong to the kingdom of Kosala from the fact that the two younger daughters are both sometimes
called “Kausalyā,” woman of Kosala.
50. Cf. Jamison 1996, 299 n. 38: a “pun” amounting to Bhīṣma’s “self-identifying announcement” that he
“formally wooed” the maidens as Vicitravīrya’s “proxy wooer” (220).
360 dharma

Lifting the women onto his chariot, Bhīṣma is quick to invoke dharma.51 But it
is very selective dharma, and also opaque in the way he reaches the number
eight in describing the legal options for weddings.52 Just before challenging all
the other kings to stop his carrying the princesses away, he recites a blur of
marriage types that are “remembered by the wise” (smṛtam budhaiḥ; 8b)—that
is, probably, sanctioned by “tradition” (smṛti)—building up to the eighth:

Now know the eighth type of marriage remembered by the poets is

the svayaṃvara, which princes praise and observe. But dharma
experts call the bride best who is carried off forcibly. So, kings, I will
carry these off by force! (10c–12b)

It would seem that Bhīṣma is interpreting this “eighth” form as susceptible to the
rules of his sixth, which he has just described simply as “taking leave by force”
(9c), implying the abduction of the bride according to the so-called Rākṣasa mode,
which is implicitly what he turns the Kāśi princesses’ svayaṃvara into. Bhīṣma is
clearly familiar with there being eight (he numbers only the eighth) acceptable
forms of marriage, such as finds its way into some dharmasūtras and Manu.53 But
the svayaṃvara is not among the eight these texts mention, which fits our sense
that Manu speaks from a tradition that avoids role modeling by princely women.54
The Mahābhārata, in fact, accurately recounts the same eight types in the same
order as Manu when King Duṣyanta cites “Manu Svāyambhuva” on their descend-
ing order of preference by varṇa (Mbh 1.67.8c–9b; cf. M 3.21) before he praises
the Gāndharva mode or “love match,” by which he is trying to seduce Śakuntalā.55
It would thus seem that Bhīṣma, like Duṣyanta, knows what he is doing. Taking
advantage of urgency to make his list cursory, which may allow him to be vague
about the Prājāpatya mode, he bends the Law by making double-talk about the

51. As Jamison 1996, 218–26 shows, the chariot-mounting is among the features by which this violent
marriage mode is regulated by sanctified protocols found in the higher marriage modes.
52. I follow van Buitenen’s parsing (1973, 228, 456) and translation of the “extremely elliptical” 1.96.8 as
covering three modes (the brāhma, daiva, and “doubtfully” the prājāpatya) and also his handling of 96.10b as
referring to the svayaṃvara rather than, as per Dumézil, the gāndharva mode. Cf. Brockington 2006, 36; Dumézil
1979, 37, who sees only one mode (the brāhma) in verse 8, and reaches eight by counting a northern interpolation
(1.999*) that mentions the ārṣa mode, even though it is accounted for in 9ab.
53. G 4.6–15, B 1.20.16, and M 20–35 each mention eight. See Dumézil 1979, 35–39 on variations in other
enumerations. As Heesterman 2001, 254 points out, most gṛhyasūtras mention only two types: bride-price (śulka)
or unconditional bride-gift.
54. Jamison (1996, 305 n. 99), noting that “there seems to be no particular evidence for the general
autonomy of Kṣatriya women,” rightly questions Dumézil’s view (1979, 43–44) that the “conflict between the
woman’s right to bestow herself in a Gāndharva marriage and the legal doctrine denying her ‘independence’” is
resolved by an extension to Kṣatriya women of “the ‘autonomy’ characteristic of Kṣatriyas.” But the epics also give
Kṣatriya women a narrative autonomy not found in the legal tradition as wider role models. Cf. Jamison 2001 on
Ṛgvedic formulary evidence for the antiquity of the svayaṃvara.
55. Mbh 1.67.8–14; see Hopkins 1882, 256–57; Jamison 1996, 249–50.
 women’s dharma 361

three types of marriage most suitable for Kṣatriyas, elevating the Rākṣasa mode
while suppressing the Gāndharva mode to make room for the svayaṃvara.56
Among these three, the Gāndharva mode is similar to the svayaṃvara in that in
both the woman participates in choosing: in the Gāndharva mode privately, in
the svayaṃvara publicly.57 And the svayaṃvara and the Rākṣasa modes share feats
of valor that allow Bhīṣma to justify his intrusion as dharma.58 Just to confirm,
Vaiśaṃpāyana calls it kṣatradharma as Bhīṣma fights off the challenge of the last
determined suitor (29).
Nonetheless, although this may all be Bhīṣma’s derring-do, it is important
not to miss that it has Satyavatī’s “consent,” both in the planning and the
result.59 Satyavatī, who would certainly know what it is to be well protected in
marital arrangements, seems to have left these maidens to something close to
the law of the fishes. It is hard to imagine what the three princesses would
think while being carted off on a chariot under attack by thousands of kings: off
on a wild adventure?60 spellbound into inaction?61 traumatized?62 forced to
calculate between wishing for the good or bad aim of their suitors in missing
them? wishing someone would kill Bhīṣma? hoping that Bhīṣma’s arrows pro-
tect them?63 And what would they think, once he has defeated the last chal-
lenger, when he treats them “like daughters-in-law, younger sisters, or
daughters” as they cross the lovely landscape on the way back to Hāstinapura
(96.42–44)? They do not appear to learn this “Terrible” man’s purpose until
this “dharma-knower” has given (pradadau) them to his younger brother and

56. I think van Buitenen is right that “The present series misses the Gāndharva and possibly the Prājāpatya
modes” (1973, 456), whereas Dumézil thinks it only “suppresses the prājāpatya” (1979, 38). Ā 2.5.11; 17–20 and
V 1.28–35 omit the prājāpatya and paiśāca. On the prājāpatya as “the most difficult to distinguish of the four ‘gift’-
marriages,” defined only by a mantra that joins the couple in practicing dharma, see Jamison 1996, 217–18,
saying “probably MBh 1.96” omits it—but without clarifying. Somehow Jamison says Bhīṣma explains “the eight
forms of marriage in wearisome detail” (299 n. 38).
57. See Dumézil 1979, 33 and n. 2 (on the svayaṃvara, “Je suis porté à y voir un dérivé ‘policé’ du mode
gāndharva, propre à une société chevalresque”), 37–38, 54, 67–69, 82.
58. See Jamison 1996, 225 and Heesterman 2001, 256–57 on the vīryaśulka svayaṃvara: in Jamison’s
terms “a pseudo self-choice in which the maiden ‘chooses’ the man who accomplishes a feat of strength or manly
skill set by her father.” But it is unclear how this would apply to the Kāśi princesses, one of whom had preselected
her suitor with her father’s agreement.
59. See Satyavatī’s joyful consent (anumati; Mbh 5.171.4) to the marriage once the three girls arrive, just
before the eldest declines.
60. See Jamison (1996, 224) on the chariot ride to the groom’s home in an ordinary marriage, “now
ensconced on a tiny island of his property.”
61. The eldest of the three, Ambā, later wishes she had jumped off and gone to her fiancé (5.173.4).
62. Regarding the “spectacular violence of the Rākṣasa marriage, Manu speaks of ‘hitting,’ ‘cleaving,’ and
‘breaking’ and of a ‘screaming,’ ‘wailing’ bride” (Heesterman 2001, 256, citing M 3.33; cf. Jamison 1996, 211,
218, 225).
63. In its violence to women, this “superhuman feat” may be comparable to his stopping the Gaṅgā.
362 dharma

begun preparations for their wedding “according to the Law of the good” (satāṃ
dharmeṇa dharmajñaḥ), as he had decided with Satyavatī (96.45–46).
But now the eldest of these still unnamed maidens has a problem.
While Bhīṣma is making wedding preparations, she reveals that she had
prearranged with King Śālva of Saubha to choose (vṛ) him at the svayaṃvara,
having already done so in her heart (manasā), as indeed Śālva had also for-
merly chosen her. Moreover, this was also the desire of her father
(1.96.48–49). In other words, beneath this svayaṃvara there were some ele-
ments of a “love-match,”64 the very thing Bhīṣma suppressed in his
patchwork recital of the eight modes of marriage. Śālva was of course the
last challenger, and the eldest abductee who had chosen him could well
have wished for Bhīṣma’s death—as she will soon do for the rest of her life,
and indeed her next one. Her choices are thus important. On one level, we
may be back in the divine plan, for as Biardeau remarks, rather than taking
on the role of dharmapatnī, a wife in accord with dharma, that Bhīṣma offers
her, she seems to have made the error of choosing for herself an Asura
demon who had incarnated on earth.65 And while that would not tell her
story as she lives it, she is portrayed as defiant. Up to now she seems to be
speaking only to Bhīṣma, but they are actually “in an assembly of Brahmins”
(viprasaṃsadi; 96.50b). There she calls him “dharma-knower” and chal-
lenges him, now that he knows of her prior choice, to conduct himself in
accord with dharma (dharmam ācara) (50)! Bhīṣma, who has subverted the
svayaṃvara into a Rākṣasa-mode abduction and explained it as dharma by
suppressing mention of a Gāndharva match that could be latent within it is
suddenly challenged to rule on the two very things he had ruled out: a wom-
an’s choice and an apparent love match. Moreover, he is challenged to find
a “courtroom” solution to a dharma predicament, as he will be again during
Draupadī’s humiliation at the dice match. But unlike his equivocation on
Draupadī’s question, this time “the dharma-knowing hero” and the
Brahmins decide, with no reported discussion and without consulting
Satyavatī, to give the oldest sister leave to go her own way. In these verses
we learn that her name is Ambā, “Mother,” and in the next verse that her
two sisters, as they marry Vicitravīrya, are called Ambikā and Ambālikā
(52), both diminutives for “Mother.” The isolated “Mother” Ambā gets no

64. The father should be irrelevant to the gāndharva mode (M 3.32; see Dumézil 1979, 34, 43, 53). But see
Mbh 13.44.5 where Bhīṣma says he rejects his own preference in that mode to give the bride to the man she loves
(see Dumézil 1979, 39).
65. Biardeau 2002, 1: 221. Moreover, her father has agreed (Ibid.), which Biardeau suspects may reflect a
connection of Kāśi and Vārāṇasī with Buddhism. Biardeau relies on a widely found northern variant here; but the
CE at 1.61.17 has the demon Ajaka incarnate not in Śālva but in a king named Malla.
 women’s dharma 363

help from these two “little mothers” or her mother-in-law Satyavatī, who
seems to have no voice in this assembly of Brahmins; and her “father’s
former desire” to see her married to Śālva is a matter that has no legal force.
Although she will remain pertinent to this skein, this is the last one hears
of her in it. Bhīṣma will bring her story up to date shortly before the war to
explain why he will not fight a brother of Draupadī named Śikhaṇḍin, who
is Ambā reborn after she had been rejected not only by Śālva but by Bhīṣma
himself and has died vowing Bhīṣma’s death in her next birth.
Neither Ambikā nor Ambālikā are ever ascribed a word. “Tall and dark
(bṛhatī śyāme) and with blue-black curling hair, red-pointed nails, and buxom
breasts and buttocks,” they turn Vicitravīrya from being dharma-minded to
desire-minded (dharmātmā kāmātmā samapadyata) for seven years until he
dies from “consumption” (1.96.53–58). Once again Bhīṣma is “settled on the
mind of Satyavatī” (59d), who is now not only a widow but a grieving childless
mother:

Then Satyavatī, distressed, wretched, eagerly longing for sons, having

done the funeral rites for her son with her daughters-in-law, Bhārata,
a thinking [or proud or eminent] woman (māninī), having reflected
upon dharma and the Fathers’ lineage and the Mothers’ lineage
(dharmaṃ ca pitṛvaṃśaṃ ca mātṛvaṃśaṃ ca), spoke to Gaṅgeya:
“The ancestral offering (piṇḍa), fame, and continuity (saṃtānam)
of the famous Kauravya Śaṃtanu who was always a man of dharma
are established in you.” (97.1–3)

When Satyavatī mentions the Mothers’ lineage along with that of the Fathers,
one might think she is referring to her own birth line (or lines . . . ). But the ances-
tral piṇdạ —a rite where rice balls are offered up on behalf of both paternal and
maternal grandparents in a son’s patriline—tells us otherwise. She is thinking of
all the men (notably her deceased husband and sons) and all the women (dead
and especially living, including herself, Ambikā, and Ambālikā) married into the
one dynastic line whose “continuity” she says she seeks now to secure through
Bhīṣma. I would not rule out that she could also be thinking of her husband’s
first wife Gaṅgā, now back in heaven, but probably not as the recipient of a piṇdạ ,
since Gaṅgā is immortal. Both Gaṅgā and Satyavatī link the dynasty’s “conti-
nuity” through Bhīṣma with dharma: the one through history (itihāsa), the other
through lineage (vaṃśa, kula). With Satyavatī, we recall that the ancestral offering
is one of the three debts and one of the five mahāyajñas that are done for the
benefit of both male and female “fathers” as joint sacrificers.
Clearly, however, even if Satyavatī considers Bhīṣma the “best of the
upholders of dharma” (7b), she can hardly expect that he will do what she
364 dharma

now asks of him, and we know what else she could have in the back of her
mind even if she has apparently not thought of it yet:

I shall enjoin (viniyokṣyāmi) you in a task. When you have heard it,
you can do it. . . . Your brother’s two chief queens (mahiṣyau),
auspicious daughters of the Kāśi king, both lovely and in the bloom
of youth, are desirous of sons, Bhārata. Beget children on them for the
continuity of our line (saṃtānāya kulasya naḥ) by my appointment
(man-niyogāt), fortunate one. You are able to do dharma here. Be
consecrated as king of the realm and rule the Bhāratas. Take a wife by
the Law (dharmeṇa). Do not drown (your) grandfathers. (97.7cd, 9–11)

Satyavatī mentions dharma at every turn (from 97.2–7 she refers to it seven
times), but when it comes to being able to do dharma in two contradictory ways,
Bhīṣma has established his preference for celibacy. No doubt Ghosh has a point
that Satyavatī is taking a courageous risk, as she seems to leave a possibility that
Bhīṣma could have a son who might succeed him as king with a new wife (2000,
41–42), although Satyavatī has certainly made her two widowed daughters-in-law
his first “appointed” priority. But rather than “renouncing her self-interest
irreversibly,”66 I think Satyavatī is urgently throwing different options on the
table more as a hastily calculated risk. Though jumbled, and whether or not she
has her own next step in the back of her mind, she drives straight to the point
that she knows Bhīṣma must reject. At first, she seems to be talking about niyoga
as the appointment to a proxy siring, which perhaps by some wild stretch of the
imagination she might think Bhīṣma would think of a way to do. But in short
compass she asks him to do dharma by breaking both of his vows in order both
to rule and to marry. Let us also note, and hold for further discussion, the incon-
gruity of her mentioning two chief queens (mahiṣīs).
Bhīṣma reminds Satyavatī of both of his vows. Restating them, he says that
even though her words are the highest dharma, nothing will make him forsake
his truth even if “the Dharma King (Yama) should abandon dharma” (97.13–18).
Satyavatī now admits, “I know the truth which you spoke for my sake,” and she
knows that Bhīṣma stands utterly upon it (20–21b); but she persists with one
more appeal:

Look to the Law of Distress. Carry out the ancestral yoke. Act,
scourge of your foes, so the family thread (kulatantu) and dharma will
not be lost and your wellwishers may rejoice. (97.21c–22)

66. Ghosh 2000, 42, tending to read disinterest into Satyavatī’s decisions, I think gratuitously; cf. 2000,
42 on her agreement to seek a Brahmin to undertake a proxy siring.
 women’s dharma 365

According to the narrator Vaiśaṃpāyana, Satyavatī, still “wretched, eagerly

longing for sons,” is now reduced to “babbling and speaking astray from
dharma (dharmād apetam)” (23). It seems that she is left to hold dharma and
what she calls the bare “family thread” or “thread of the line” together all by
herself. As she perceives, there seems to be no dharma solution outside the Law
of Distress, for which Bhīṣma will reveal some uneasiness when he approaches
that topic in the Āpaddharmaparvan of Book 12 (Bowles 2007, 192–94). But let
us note that one imaginable solution goes unimagined: Satyavatī, now certainly
no less a widowed queen than her daughters-in-law, would be available for a
proxy siring herself. Like her two daughters-in-law, whom she will soon describe
as “desirous of sons by the Law” (putrakāme ca dharmataḥ; 99.43), she too, as
just mentioned, is “eagerly longing for sons” (putragṛddhinī). In these epics it
would be no excuse to say she is too old (we have met one comparable figure:
Bhīṣma’s paternal grandmother, the wife of Devāpi; another is Kausalyā in the
Rāmāyaṇa). As we shall see, this unmentionable (certainly Bhīṣma could not
recommend it to his “mother”) has some interest as a structural possibility. But
back to reality, Bhīṣma does have an āpaddharma solution to offer that he can
leave her and himself biologically out of, which he introduces as kṣātra
dharma:

I shall tell you, queen, the eternal Kṣatriya dharma by which even the
continuity of Śaṃtanu (śaṃtanor api saṃtānam) may be imperishable
(akṣayam) on earth. Having heard it, and having looked to the loom
of the world (lokatantram), carry it out with wise household priests
who are skilled in the aims of āpaddharma (97.25–26).

Once she listens, Satyavatī will not be left holding the last thread (tantu) of the
line. For in his tightly packed response that “picks up” twice on the verbal
root÷tan in tantu, “thread,” Bhīṣma offers her as dharma a way that she can
work that thread back into the “continuity” (saṃtāna) of Śaṃtanu, having looked
to the “loom” or “course” (tantra) of the world! It will be Satyavatī’s—and our—
task to unravel this teaser, which punctuates the end of an adhyāya. But first
she must hear more from Bhīṣma, who has yet to know all that lies potent in
his own words.
Without mentioning niyoga (or any derivative), which calls for the interven-
tion of a (preferably older) brother of a deceased or otherwise incapacitated
husband, Bhīṣma recommends—as a kind of alternative, or as Doniger puts it,
as “pseudo-niyoga” (1995, 174)—that a Brahmin do the job, and provides two
precedents. First, placing his proposal under the heading of Kṣatriya dharma,
he invokes the story of Rāma Jāmadagnya: after he killed all the Kṣatriyas
twenty-one times, their widows regenerated that population by uniting with
366 dharma

Brahmins (1.98.1–5).67 Then he works back to that point rather obliquely by

telling of the blind Ṛṣi Dīrghatamas, who was cursed in the womb to (as his
name indicates) “long darkness” (6–19). When an apparently impotent king
Bali engages him to sire sons with his wife, the Ṛṣi does this, but not before the
queen rejects him for his blindness and age and refers him to her Śūdra nurse,
with whom he first sires eleven of “his own” sons (20–33). As several (Doniger
1995, 174–75; Biardeau 2002, 1: 216 n. 21; Dhand 2004, 40, 42) have noted,
there are foreshadowings here—blindness from a curse, rejected proxy, chil-
dren with the servant—of what will soon be replicated in the Kuru line. But in
brief, citing also a “Vedic” adage,68 Bhīṣma thinks all that is necessary is to find
a Brahmin “of stern spirit” (niyātātmā; 4c) and for the women in question to
“keep in mind the Law” (dharmam manasi saṃsthāpya; 4–5). Satyavatī herself
should thus invite a suitable Brahmin of qualities (guṇavān) to prosper the con-
tinuity (saṃtāna) of the lineage “in the fields of Vicitravīrya” (99.1–2).
So it is, with no hint that she has thought of it only suddenly, but rather
because she knows Bhīṣma speaks the truth and because she trusts him, and for
the continuity (saṃtāna) of the line, that Satyavatī now speaks of her premarital
affair “with a faltering voice and as if smiling shyly” (1.99.3–4). She begins with
how lawful everyone is and was: to Bhīṣma, “In our family you are dharma, you are
the truth, you are the ultimate resort”; her father was legally engaged (dharmayukta)
in his ferryboating; and “the supreme Ṛsị Parāśara was the best of dharma’s
upholders” (5–7). She tells how she was seduced by Parāśara’s boons of smelling
beautiful and restored virginity, how once born Vyāsa divided the Veda, etc. And
she concludes that although she can bring Vyāsa there, they must co-invite him:

He surely, when appointed (niyuktaḥ) by me and by you, . . . will

beget beautiful offspring on your brother’s fields. He has said, “You
may remember me when a task need be done.” I shall now recall
him, strong-armed Bhīṣma, if you wish. Surely, Bhīṣma, with your
consent (anumate) the great ascetic will beget sons in the fields of
Vicitravīrya. (15–17)

Note that it is Satyavatī who introduces the terminology of niyoga in this irreg-
ular manner of appointing a Brahmin—but not without regularity, since Vyāsa
(on the mother’s side), like Bhīṣma (on the father’s side), is an older (half-)
brother of Vicitravīrya.

67. See chapter 7 § B.3 on Kṛṣṇa’s version of this story, which is told from the different angle of regenera-
tion in out-of-the-way places.
68. “‘The son is his who took the hand,’ so it is decided in the Vedas” (pāṇigrāhasya tanaya iti vedeṣu
niścitam; 98.5ab). That is, the son belongs to the woman’s husband, dead or alive.
 women’s dharma 367

Bhīṣma folds his hands in homage. Somehow knowing something of

Vyāsa’s reputation, he praises Vyāsa’s insight into the three puruṣārthas, and
approves. Here, for the first and only time in this skein, Satyavatī is given a pro-
vocative name:

Then, when Bhīṣma had given his promise, O scion of the Kurus,
Kālī directed her thought to the Muni Kṛṣṇa Dvaipāyana (21).

As far as I can see, Satyavatī’s name Kālī has not been sufficiently consid-
ered.69 Only Biardeau seems to have grappled with it, but she loses its singu-
larity in her mythographic categories: Satyavatī’s name, “la Noire,” occurs in
our present passage after Satyavatī mentions Vyāsa’s name Kṛṣṇa, “black”
(2002, 1: 216); and when Vaiśaṃpāyana mentions it, he “goes so far as to call
[her] Kālī, ‘la Noire,’ from the name of the goddess under the cruel form of
slayer of the Buffalo asura” (220). Biardeau offers no explanation of what
Satyavatī has to do with slaying the Buffalo Demon, and would seem to over-
particularize70 where Kālī is something more general: the name of a goddess
recognized for being above all, like Śaṃtanu’s other wife Gaṅgā, a very famous
Mother.
This is not an easy point to make, since, as has been recognized for some
time (see Kinsley 1975, 89), there seems to be only one place in the Mahābhārata,
at least in the Critical Edition, which could be used to support the idea that the
epic poets even know this “dread goddess.”71 But it is not a negligible usage,
and despite what I said some years ago, it is not likely to be an “intrusion.”72
“Kālī” is used either as an adjective or a proper name to introduce the goddess
Kālarātrī, the “Night of Time,” also calling her a Kṛtyā, an apparitional goddess
of black magic.73 It describes Kālarātrῑ as she is seen by the warriors about to
be slaughtered by Droṇa’s son Aśvatthāman as he is possessed by Śiva during
the night massacre that brings the Mahābhārata war to its final cataclysm,
and as she was seen by those warriors in dreams all during the war
(10.8.64–67). As Jacques Scheuer cleverly puts it, Kālarātrī “is presented with
the features of a black woman (kālī, the first word of the description, is nearly
a proper name) . . . ” (1982, 316). And indeed, Biardeau takes it as a proper

69. Translating it as “the Dark Woman” (1973, 234), van Buitenen compares it to her supposed name
Kṛṣṇā, “the black one,” as reflecting her “low origins” from an “aboriginal fishing tribe” (447). But he cites no
instance of Satyavatī’s being called Kṛṣṇā, and there is none in Sörensen’s Index of Mbh names ([1904] 1963).
70. Biardeau links epic usages of “Kālī” with the killing of the Buffalo Demon as part of an argument that
it appears in two hymns to Durgā in Books 4 and 6 (2002, 2: 467) that are rejected by the CE as interpolated.
71. Manu already knows of Kālī as Bhadrakālī in connection with the householder’s Bali offerings (M 3.89).
72. Hiltebeitel [1976] 1990, 326. See now Johnson 1998, 115–16 for a better assessment.
73. On Kṛtyā as parallel to Draupadī and later associated with Kālī, see Hiltebeitel 2001a, 190–92;
Türstig 1985.
368 dharma

name.74 In either case, I think it is plausible to think of it as a reference to Kālī

(cf. Johnson 1998, 115). But even if this Kṛtyā Kālarātrī is just another “dark
woman,” we still have the case of Satyavatī, which could stand alone in the epic
as a reference to Kālī. What then can we learn from the uses of this name
for her?
First of all, as regards this one reference to Satyavatī as Kālī in this skein,
I believe Biardeau has the right impulse to note that it occurs after Satyavatī
mentions Vyāsa’s name Kṛṣṇa. Biardeau also links Satyavatī–Kālī with the
“epic constant” of darkness: as with the dark Yamunā from which she is born,
both she and her son are among the dark “personages connected with saving
the dharma of the world of transmigration” (2002, 1: 201). Contextually, how-
ever, when Vaiśaṃpāyana invokes this name, he reminds us, just as Satyavatī
recalls Vyāsa’s birth, indeed, as she brings him back into her world by remem-
bering him, that she was “Kālī” when he was born. This is also what the higher
order narrator, Ugraśravas, has said earlier when he introduced Vyāsa upon his
arrival at Janamejaya’s Snake Sacrifice to hear his epic composition recited for
the first time. Vyāsa is:

he whom Kālī had birthed from Parāśara, son of Śakti, even as a

virgin on a Yamunā island, the grandfather of the Pāṇḍavas who, the
same day he was born from his mother, matured his body by will.
(1.54.2–3b)

The case is also similar in the prose genealogy of the whole dynasty that
Vaiśaṃpāyana delivers just before he spins out our present skein from it. When
he comes to Satyavatī she is called Gandhakālī (90.51), “Kālī of the Smell”;
and although the two names are used together to describe everything she
does, Gandhakālī is used more tightly with reference to her mothering and
premarital birthing of Vyāsa:

Bhīṣma, to do his father a favor, brought him Satyavatī, the mother

whom they called Gandhakālī, whose premarital son (kānīna)
Dvaipāyana was a child from Pārāśara. . . . (1.90.51–52; cf. M 9.160, 172)

Whereas “Kālī of the Smell” would presumably have the smell of dead fish,
which Satyavatī loses while keeping her virginity thanks to Parāśara’s boons,
she is not called Gandhavatī75 or Yojanagandhā (1.57.67) until she has started

74. Biardeau 2002, 2: 464–65: “Kālī, au visage et aux yeux rouges, . . . C’est Kālarātri. . . .” Recall n 38:
Biardeau 2002, 1: 219 explains Satyavatī’s name as “she who has within her all beings.” This is really an etymology
of her name as based on sat, “being,” which could give us the name Satī, Śiva’s first wife, whom the purāṇic Kālī
becomes after she sheds her darkness (Hiltebeitel 1999c).
75. Vyāsa is frequently called the son of Gandhavatī, “the Fragrant,” which could, of course, recall either’s smell.
 women’s dharma 369

to smell like perfume. I think the poets describe Satyavatī as having under-
gone a transformation similar to ones undergone by other goddesses, but
especially by Kālī, from dark to beautiful, usually golden or light (gaurī), to
make her marriageable to Śiva.76
Although it is not the case in our present skein, Satyavatī also carries her
name Kālī into her marriage to Śaṃtanu. There are some striking instances
when Bhīṣma tells Duryodhana the story of Ambā (in the Ambopākhyāna) to
explain why he will not fight Śikhaṇḍin, and again, it is usually a question of
her being a mother.77 For instance, Bhīṣma says that when Śaṃtanu was worried
that having one son was like having none, “Knowing his desire I brought him
the mother Kālī (kālīṃ mātaram āvaham)” (5.145.18cd). But most fascinating
are Bhīṣma’s usages when he describes the dire condition of the heirless realm
after the death of Vicitravīrya:

When Indra no longer rained on the kingless kingdom, the subjects

hastened to me, oppressed by hunger and fear. “All the subjects are
dwindling! Be our king and revive us! Be blessed, drive away the
plagues (ītayo nuda). All your subjects are oppressed by the most
terrible diseases (vyādhibhir) and but few remain, Gāng ̇ eya. You can
rescue us! Dispel the diseases, hero, protect the subjects by dharma
lest the kingdom fall to ruin while you are alive.” The subjects wailed,
yet my mind was not shaken. Remembering the conduct of the good,
I kept to my oath, great king. The cityfolk and my auspicious mother
Kālī (mātākālī ca me śubhā), the retainers, the house priests and
preceptors, and the learned Brahmins, much scorched, told me, “Be
the king!” continuously. . . . At their words, son, I folded my hands,
greatly distressed and unhappy, and told them again and again the
oath I had made out of deference to my father, that for the sake of the
lineage I would keep my seed drawn up and not be king. Then, having
folded my hands, I placated my mother again and again, O king:
“Mother (amba), though I am born from Śaṃtanu and carry the
lineage of Kuru, I cannot belie my oath, withal for your own sake. Do
not appoint (mā . . . niyojaya) the yoke to me. I am your servant and
slave, O mother (amba) who loves her children!” (5.145.24–29, 31–33)

It is fascinating to see Bhīṣma telling Duryodhana that he called Satyavatī ambā

while telling him the story of Ambā—as if all pestilential mothers were one, at

76. The theme occurs in the Devī Māhātmyā and widely in the Purāṇas.
77. Only once do I find the name not referencing Satyavatī as “mother,” when Bhīṣma tells Duryodhana
that he informed “Kālī Satyavatī” and various courtiers that he had given Ambā leave to go to Śālva (5.172.1–2).
370 dharma

least to him.78 The conditions that induce Bhīṣma and Satyavatī to call in Vyāsa,
which as Bhīṣma says was the next thing they did (5.145.34–35), are thus ones
that one could associate with Kālī as a mother of ills and destruction—though
again, this is not to say Satyavatī is that goddess, only that she is being called
Kālī in a way that seems to recall such a goddess. The name deepens Satyavatī’s
primordiality to put it on a par with Gaṅgā’s, with each equally intertwined into
the Kuru dynasty’s lineage problems and solutions. As Bhīṣma tells it, the
public thinks the restoration of dharma is up to him, but he knows that it must
also be up to her. In our skein, these dire conditions will soon be described
much more minimally by Satyavatī to Vyāsa.
But first, even while Vyāsa was propounding the Vedas, he appeared
“instantly, having discerned his mother’s thought” (1.99.22), and without her
having said a word. As she embraces him with tears, she is once again the
fisherwoman (dāśeyī), and he sprinkles his “distressed mother” with water,
greeting her too before saying, “I have come to do what you intend. Command
me, you who know the true nature (tattva) of dharma, and I will do what
pleases you” (25). If there was any doubt when she veered from dharma or
deferred to Bhīṣma about it, here we have the word of the author that she
knows its true nature. Vyāsa even seems eager to act on her command, as if
he equates the Law with whatever pleases his mother. Satyavatī begins with
an assertion of the parity of father and mother in parenting sons: “Sons are
born the common property of the mother and father, O poet (kavi). As the
father is their owner (svāmī), so is the mother, no doubt” (28). Drawing
together the brotherhoods of Vyāsa and Vicitravīrya on the mother’s side and
Bhīṣma on the father’s, and mentioning Bhīṣma’s disinclination to rule or
have children, she now proposes this:

Out of esteem for your brother, for the continuity (saṃtāna) of the
line, at Bhīṣma’s word and by my own appointment (niyogāt),
blameless one, out of compassion (anukrośāt) for beings and for the
protection of everyone, with non-cruelty (ānṛśaṃsyena) you must do
what I am proposing. (32–33)

Note, again, that it is she who must make the appointment. Vyāsa is to beget
sons on the two lovely wives of Vicitravīrya, who are “desirous of sons in accor-
dance with dharma” (34). Vyāsa’s response is to confirm Satyavatī’s profundity
when it comes to dharma:

78. On “ambā” and a suggestive discussion of Ambā as a third “mother” of Bhīṣma along with Gaṅgā and
Satyavatī, see Fitzgerald 2007, 191. Along with the other mother-triads that occur in successive generations, this
set gives Bhīṣma different mothers across at least two generations.
 women’s dharma 371

You know dharma, Satyavatī, both the higher and the lower (paraṃ
cāparam eva ca). And since, knower of dharma, your mind is set on
dharma, therefore by your appointment (niyogāt), having pointed out
what is needed with respect to dharma, I will do what you desire as it
is found to be ancient practice. (36–37)

Indeed, Vyāsa says that Satyavatī speaks both for the higher and lower dharma,
which, in the words of the passage that introduces him at Janamejaya’s Snake
Sacrifice, he can certainly do: “He was a Brahmarṣi who knew the high and the
low (parāvarajñaḥ), a poet, a man of true vows, and pure” (1.54.5cd).
But exactly what would the high and low be in the current situation? Vyāsa
may be preparing Satyavatī for a little problem he has just mentioned: that he
will now indicate what is needed with respect to dharma before he can under-
take “her appointment.” Vyāsa says he will give his brother sons the likes of
Mitra and Varuṇa, but that the two widows must submit to a year-long vow
before he lies with them (1.99.38–39). That may be the higher law, and, if so,
Satyavatī is urgent for a lower one: “In kingless kingdoms there is no rain, no
gods. How, lord, can a kingdom be preserved that has no king!” (99.40c–41b).
In that case, says Vyāsa, to bear a superior child the minimal vow will be that
the widows must bear his ugliness: “my smell (gandha), my looks, my garb, and
my body” (43ab).
Meeting Ambikā in secret, Satyavatī begins, “Kausalyā, what I say to you is
the loom of the law (dharmatantram). Listen to me” (99.45ab). Going on to say
that the appointment is in accord with Bhīṣma’s view (buddhi) and yet still up
to her own view (buddhi), “somehow (kathaṃcid) she persuaded that dharma-
farer [Ambikā] by appealing to dharma” (1.99.46–47, 49). As the “somehow”
suggests, those who are “appointed” are not always so accepting of this “ancient
practice.” Whatever pestilential conditions may have already beset this kingless
kingdom, Satyavatī’s urgency looks like it is the first thing to deprive the Kurus
of kings the likes of the Vedic sovereigns Mitra and Varuṇa. And the second is
that even though Vyāsa has just said that without the year’s vow it will depend
on the deportment of the widows whether they get superior sons, Satyavatī
does not sufficiently warn at least the first of them, Ambikā, and actually seems
to mislead her, telling her to expect a “brother-in-law” (devara), which makes
Ambikā think she will be lying with “Bhīṣma or another of the Kuru bulls”
(100.2–3; see Biardeau 2002, 1: 217; Dhand 2004, 35). The result is that when
Ambikā shuts her eyes at Vyāsa’s ugliness, Vyāsa curses her son to be born
blind “because of his mother’s defect of virtue” (matuḥ . . . vaiguṇyād; 10c).79

79. Cf. 96.45: all three maids were guṇasampannā when abducted. Even though Ambikā is a dharmacāriṇī
(99.49; cf. Gāndhārī at 103.11), she is vaiguṇya—suggesting a disjunction between her dharma and her guṇas.
372 dharma

Satyavatī quickly recognizes that no blind man can be king of the Kurus and
seems to correct what may have been an oversight in her first appointment.
This time she asks Vyāsa to “grant a protector of the lineage of affines (jñāti)
and an increaser of the lineage of Fathers—a second king of the Kuru lineage”
(12). Note her three differentiable usages of “lineage” (vaṃśa) in this one verse.
The lineage of affines would include Satyavatī and Vyāsa, who would belong to
the Kuru vaṃśa on her side as her son, but now also as a father. Vyāsa meets
Satyavatī’s new request, but minimally: because Ambālikā pales, he curses her
son Pāṇḍu to be born pale (14–21).
Finally, Satyavatī gives the first widow a second chance, but when Ambikā
recalls Vyāsa’s smell (gandha) and appearance (obviously, smell stands out in
her recollections) and dresses a beautiful Śūdra servant to replace her in bed,
Vyāsa and the Śūdra woman produce the “half-breed” (karaṇa) Vidura, an incar-
nation of the god Dharma and skilled in both dharma and artha (100.22–28,
101.27–28).80 Whereas the Law of Mother Gaṅgā shortchanged the Kuru
dynasty, giving it kings well-versed in dharma but only for one generation, the
Law of Mother Kālī continues to beset it with too many flawed successors: two
dead, and three now each with a defect as to being king, with only Pāṇḍu’s
blemish leaving him eligible.
Yet if Satyavatī seems to leave nothing but loose ends, she also ties
together the continuity theme with a curious term: what she speaks, she says,
is “the loom of the Law” (dharmatantram). Van Buitenen translates dharma-
tantram as “under the Law” (1973, 235); Ganguli as “It is consistent with
virtue” ([1884–96] 1970, 1: 246). Both are apt but neither captures the verbal
play that makes this usage nodal and justifies such a literal translation once
again. As mentioned, Satyavatī seemed to be left holding the bare “family
thread” or “thread of the line” (kulatantu) alone (97.22) until Bhīṣma offered
his recommendation that she look to the “loom” or “course of the world”
(lokatantram) and consult with household priests conversant with āpaddharma
(26). With the concurrence of Vyāsa, who will serve the Pāṇḍavas as a priest
in their great rituals, though not as their purohita (see Sullivan 1990a, 31–34),

80. See M 9.57–70, 167. Vyāsa would seem to be in violation of Manu’s stipulation that the man should
never approach the same widow twice (60), but then he sleeps not with the widow but her Śūdra maidservant.
Indeed, at M 9.167, niyoga can be a woman’s svadharma! Although the Mahābhārata may imply this when it takes
recourse to dharma as its justification for niyoga, it seems to spare its women characters—and particularly
Kuntī—the argument that it is their svadharma. On Vidura as a karaṇa, see MW 254: a son of mixed class vari-
ously defined; a writer, scribe! In later terms, a Karṇam. One might say that in contrast to Yudhiṣṭhira, about
whom one keeps hearing that it is his very nature or svabhāva to be or embody dharma (see chapters 9 § D.2.b and
10 § C, and Mbh 6.115.63–65 cited in chapter 1 § C), Vidura is a kind of moral accountant—a contrast I owe to a
conversation with Gurcharan Das (October 2010). Das writes of this difference as follows: whereas Yudhiṣṭhira
offers “a view of the world, based on dharma, which he explains as a universal duty of righteousness,” Vidura’s
“moral thinking is based on the consequences of actions rather than duty” (2009, 15–16).
 women’s dharma 373

she now urges as dharmatantra the first of the three unions that will produce
what Bhīṣma will soon enough, after a time of onrolling dharma under his
regency,81 confirm are the needed threads to continue the lineage into the
next generation. As he says to Vyāsa’s three sons, and Vidura in particular,
when he begins to plot their weddings:

Our family, protected by the great-spirited dharma-knowing kings of

old, has never come to ruin here; and through me, Satyavatī, and
the great-spirited Kṛṣṇa [Vyāsa], it has been firmly established in
yourselves, the threads of the line (yusmāsu kulatantuṣu). (103.2–3)

Indeed, well before this in the text (or after it in the sequence of frames), the
higher order narrator Ugraśravas prepared his audience, the Ṛṣis of the Naimiṣa
Forest, to anticipate such “extended” verbal play, saying of Vyāsa when he
comes to Janamejaya’s Snake Sacrifice:

It was he who with holy renown and great fame begot Dhṛtarāṣṭra,
Pāṇḍu, and Vidura, extending (or stringing along) the continuation
(or stretching over) of Śaṃtanu (śaṃtanoḥ saṃtatiṃ tanvan). (1.54.6)

We see then how Mother Kālī Satyavatī has reset the shuttles of the divine
plan along with Bhīṣma and Vyāsa. But let us not forget that Śaṃtanu himself
and above all his first wife, Gaṅgā, set the first threads to this tapestry. Indeed,
still within the higher order narration of the Bard Ugraśravas, I believe we are
taken back to one of the riddles that provide an allegorical backcloth to the
entire Mahābhārata in the story of Uttaṅka (see n. 14). These two Mothers—
the bright woman of the night and the dark woman brought ashore by day—
give genealogical shape to the Uttaṅka story’s unique identification of Dhātṛ
and Vidhātṛ, the Placer and Ordainer, respectively, as goddesses, weavers on
the loom (tantre) of Time whose black and white threads (tantavaḥ) are woven
into nights and days (1.3.167, 172), just as the children and descendants of
these two mothers are threads woven into the nights and days of this text.

E. The Transitional “Three Mothers”

As we move from the generation of Satyavatī’s sons to that of her grandsons,

Bhīṣma once again brings in three women to be queens of the royal line:
81. Most of adhyāya 1.102 describes a time of plenty in which these four are said to have rolled along
(avartata): the Kṛtayuga or Golden Age (5d), the supreme Law (dharmam uttamam; 7d), the Wheel of Dharma
(dharmacakra; 12f) held by Bhīṣma, and a saying that went around in all kingdoms that said, “Of mothers of
heroes (vīrasūnām) the daughters of Kāśi, of countries Kurujāṅgala, of dharma-knowing princes Bhīṣma, and of
cities the Town of the Elephant” (21–22).
374 dharma

Gāndhārī, to marry Dhṛtarāṣṭra; and Kuntī and Mādrī to marry Pāṇḍu (103–5).
He also secures an unnamed bastard daughter of a king82 to be Vidura’s wife,
with whom Vidura has many virtuous sons who are unnamed like their mother
(106.12–14). Presumably Bhīṣma follows custom in finding brides for brothers
in their order of seniority. We must soon bring closure on this headline role of
Bhīṣma, by which Dumézil dubs him “Bhīṣma marieur,” Bhīṣma the match-
maker (1979, 66–71). Dumézil covers Bhīṣma’s part in arranging the mar-
riages of Satyavatī to his father Śaṃtanu; of two of the three sisters he abducted
for his half-brother Vicitravīrya; and now of these three wives for Dhṛtarāṣṭra
and Pāṇḍu (66–67). Indeed, not only can we add to this his arrangement of
Vidura’s marriage, we could also suspect that his matchmaking role was set for
him in his previous life as Dyaus when he was the chief among the eight Vasus
to get the samaya (91.21–22) or boon (varam, 93.40–42) of Gaṅgā by which she
married Mahābhiṣa-Śaṃtanu and became Bhīṣma’s own mother.
In any case, the first thing we may note about “Bhīṣma marieur” is that, once
he has accomplished these three generations of matchmaking, he seems to have
finished all work of this kind. When it comes to the marriage of Draupadī, such
family business has been handed over to his logical counterpart and co-
“grandfather” Vyāsa, who, from Draupadī’s wedding on to the revival of the
̣ ạ vas’ grandson Parikṣit, gets help from Kṛsn
Pānd ̣ ạ in carrying forward the ongoing
wonders of the line’s continuity. From such hindsight, Bhīṣma’s matchmaking
seems to have been doomed to repeat certain types of failures through the two
transitions under his watch—failures that he was hardly the one to correct.
In the first transition, Bhīṣma seems bent on repeating a fateful fascina-
tion with primal Mothers, with the first of whom, Gaṅgā, he in effect—as
Dyaus—arranged his own conception. If I have invoked the primordial
renown of Gaṅgā and Kālī Satyavatī as Mothers, I passed too quickly over an
evident continuity with the names of their three successors: the three Kāśi
princesses Bhīṣma abducts to marry Vicitravīrya. Bhīṣma’s and Satyavatī’s
intention is to bring these three “Mothers” into the lineage before they ever
are such, in this case mothers less by reputation than by their names, which
mean something like Mama (Ambā), Mamita (Ambikā), and Mamacita
(Ambālikā), or, if one prefers, “Mummy, Mummikins, little Mummy”
(Jamison 1996, 67, 69, 79). They are brought in to be mothers like Gaṅgā
and Satyavatī, but unlike them they are brought in violently; and their manner
of becoming mothers will not be to their liking. Moreover, as both Jamison

82. See van Buitenen 1973, 243, with “bastard” for pāraśavī: in Gautama’s possibly earliest account of mixed
classes (Olivelle 1999, xxxi), a child born to a woman three classes below the man (G 4.17). This would describe
Vidura, but be problematic for a king’s daughter if there is “no fifth” class, unless the king were a Brahmin.
 women’s dharma 375

and Biardeau have recognized (without, it seems, anyone else having waded
in on it), these three names in combination cannot be disentangled from
their use in a celebrated—if often called “notorious”—scene in the great royal
horse sacrifice, the Aśvamedha. This usage occurs when the three names are
invoked in a formula special to the Aśvamedha while the chief queen or
mahiṣī is having undercover ritual intercourse with the sacrificed horse. The
formula has what Jamison calls “clean and dirty versions” (79). The first and
most widely mentioned in the texts is used when the mahiṣī is led up to the
slain horse accompanied by her cowives (each of whom may have a hundred
attendants). Here the mantra is:

O Ambā, Ambālī, Ambikā [var. Ambā, Ambikā, Ambālikā]83

No one is leading (nayati) me.
The horsikins is sleeping. (Jamison 1996, 67, trans. TS 7.4.19.1ab)

“Leading,” according to Jamison (67, 274 n. 108), probably introduces a con-

ventional sense of ÷nī to suggest that the mahiṣī is being led, as it were, into
matrimony. Then, once the mahiṣī has settled herself with the horse beneath a
blanket84 and either the Adhvaryu priest or the king (her husband as Sacrificer)
has uttered the most straightforwardly erotic of all the rite’s mantras, focusing
on her sexual pleasure, the mahiṣī, having “manipulate[d] the dead horse into
some sort of copulatory position,” in some texts modifies her mantra to say:

O Ambā, Ambālī, Ambikā85

No one is fucking (yabhati) me.
The horsikins is sleeping. (Jamison 1996, 69, trans. TS 7.4.19.2h)

With or without these words, which “mock” or “scold” the horse, the rite clearly
invites the dead horse to regain its sexual stamina,86 for one of the Aśvamedha’s
purposes is to induce fertility, and specifically to obtain royal offspring (66, 76;
cf. 242). Meanwhile, the cowives and their attendants circulate back and forth

83. The usual opening line ambe ambāly ambike has, as Jamison observes (1996, 274 n. 107; 304 n. 87), a
precise Mahābhārata nomenclature and sequence in VS 23.18: ámbe ámbike ‘mbālike, with the same but for the
initial vocative in MS 3.12.20. According to Jamison (243), “the three vocatives are taken as variants on affec-
tionate terms for “mother,” but together they also add up to the “three Ambikās” of Rudra Tryambaka and its
vṛddhi derivative, the Traiyambakahoma.”
84. Biardeau says, “It seems that the queen complains at being looked at in the sexual act” (2002, 1: 220).
See Jamison 1996, 67: before the Adhvaryu covers the mahiṣī and the horse with a linen blanket, she “lies down
beside the horse and invites it to stretch out its forefeet along with hers.” How that would help achieve a position
for sexual contact is not made clear.
85. VS and MS apparently do not have a “crude version” with the precise Mahābhārata names.
86. See chapters 7 § A.1 and 12 § A on the Yuga Purāṇa’s allusion to the closing episode of the Harivaṃśa
where Indra animates the horse to have intercourse with King Janamejaya’s wife.
376 dharma

around the horse and the mahiṣī slapping their thighs and fanning with their
hems, and then exchange “slangy and crude” riddling mantras with the priests
(65–66, 69–70).
The use of this formula to overcome childlessness gives us, then, an
obvious point in common with what Bhīṣma and Satyavatī initially wish from
the three Kāśi sisters, and continue to wish from the two younger of them even
after they have become widows. But as Jamison and Biardeau recognize, the
connection is anything but obvious, in no small part because there is no
Aśvamedha at this point in the Mahābhārata (indeed, with Bhīṣma as regent,
there is no king to perform one). Jamison, looking back to Vedic sources, pro-
poses that the two usages can be illumined by a third: a “husband-finding”
(pati-vedana) ceremony in its form as an “addendum” to the third of the Four-
monthly (Caturmāsyāni) seasonal rites called the Sākamedha, in which Rudra
is worshiped with a Traiyambaka Homa that includes Aśvamedha features:
especially (along with more back and forth thigh-slapping87) the names of the
“three mothers” formula condensed into “Traiyambaka” as a name for Rudra,
who is worshiped both to remove the young woman’s lack of marital success
and to secure her first unborn descendant (242–43, 245–46). Biardeau, on the
other hand, looks toward Purāṇic usages and finds the “link . . . between the rite
and its epic role” (if I am unpacking her brief discussion correctly) in the prop-
osition that the triple name in the Aśvamedha would evoke one person, the
mahiṣī, “transformed into three women” as the Kāśi sisters, and that the one
woman chiefly in question would thus be Ambikā—whom she calls here “the
mother of Pāṇḍu,” but must mean the mother of Dhṛtarāṣṭra—not only as the
mahiṣī, but in anticipation of her name becoming “one of the most frequent
names of the Goddess” (2002, 1: 220).
I believe that both these authors have found pieces to a puzzle that the
Mahābhārata poets have intentionally left incomplete. And if that is so, then so
it must remain. But I think there are more pieces to be found if we allow for
two things: first, that the Mahābhārata episode alludes knowingly to this
Aśvamedha mantra and its ritual scene without being willing to make it obvious.
For I think what it wants to do is pull off an innuendo that beneath the already
potentially disagreeable submission to proxy siring (niyoga) to which the two
widows are asked to submit, there is the further suggestion that the smelly and
misshapen author Vyāsa is cryptically taking on the role of the revived sacrifi-
cial horse. Second, I would urge that if we dig into the contextual site of this

87. Jamison 1996, 242–44 (it uses a mantra to Tryambaka found in ṚV 7.59.12 that the girl modifies to
request a husband). As she points out, both rites resemble the movements of fire-circling servant maids on the
Mahāvrata day of the Gavāmayana.
 women’s dharma 377

episode while thinking more or less synchronically across the two Sanskrit
epics rather than reaching back to supposedly older versions of the epic story,
which I consider unlikely,88 or resting the case on anticipatory epic outcrop-
pings of postepic goddess mythology, we can see a few more pieces of the
puzzle crop up from the text itself.
The first move, then, is to the Rāmāyaṇa, where a parallel scheme contex-
tualizes this interpretation. It is a case of the Rāmāyaṇa having two
Aśvamedhas where the Mahābhārata seems—until one sees the double
parallel—to have only one. For whereas the Mahābhārata slots this present
story of the rescue of the Kuru line amid (and, I argue, colored by) the other
rescues of the line that come before and after it in the same skein, and has an
actual Aśvamedha in its main narrative only later as a sin-cleansing rite after
the Mahābhārata war (14.70.15–16; Hiltebeitel [1976] 1990, 292), the
Rāmāyaṇa has two Aśvamedhas in its main narrative: the first slotted at
the only parallel point where it is a question of rescuing the Ikṣvāku line by
the birth of Daśaratha’s four sons headed by Rāma,89 and the second, as in the
Mahābhārata, a postwar sin-cleansing90 rite of realm consolidation. Something
of this complexity calls for a fuller exposition elsewhere.91 But let us not miss
the ways that both epics keep the Aśvamedha’s reputation for replenishing
the continuity of royal lineages even in their second Aśvamedhas. In the
Mahābhārata, it is during the postwar Aśvamedha that Kṛṣṇa revives the still-
born Parikṣit, as Vyāsa had promised. And in the Rāmāyaṇa, it is during the
postwar Aśvamedha that Rāma discovers his sons Kuśa and Lava, born to Sītā
during her banishment in Vālmīki’s hermitage, who will now be Rāma’s
recovered heirs.92

88. Jamison 1996, 304 n.94 “assumes that some version of the epic story of Ambā and her sisters already
existed in early Vedic and that these girls were associated with the three Ambikās belonging to Rudra. Despite
the difficulties . . ., I think this is more likely than assuming that a later epic poet simply made up the story of the
abduction and its aftermath and named the female protagonists by plucking some designations out of the ono-
mastic repository of Vedic ritual.” I see the alternative more positively as part of the Mbh’s project of knowing
Vedic allusion (see Hiltebeitel 2001a, 42 and passim).
89. This rite (Rām 1.8–17) is undertaken by the potent Ṛṣi Ṛyaśṛṅga for the “continuity” (saṃtāna; 10.5;
11.2) of the aged and sonless Daśaratha’s line, and is doubled by a putrīyā iṣṭi (or putrakāmeṣṭi) or “son-producing
sacrifice” (14.2–3; 15.24) that yields the porridge that the three queens then share to get pregnant with portions of
Viṣṇu. Cf. Goldman 1984, 60, 74–77.
90. Rāma proposes a Rājasūya sacrifice, but Bharata tells him a horse sacrifice is less destructive and that
the Aśvamedha removes all sins and purifies (Rām 7.75.2). Lakṣmaṇa tells of Indra’s Aśvamedha to recover from
the Brahmanicide of killing Vṛtra, which could allude to Rāma’s Brahmanicide in killing Rāvaṇa, a descendent
of the Brahmin sage Pulastya.
91. See Hiltebeitel 2007a, 126–35 on the Mahābhārata’s postwar Aśvamedha; 2011a, chapter 9, on this
episode along side prewar and postwar Aśvamedhas in both epics.
92. Note that Sītā, the would-be mahiṣī of this piece, is brought to the Aśvamedha not to lie with the horse
but to make the disagreeable attestation to her purity that results in her earthen engulfment, but also confirms
the legitimacy of her two sons.
378 dharma

What then are the additional puzzle pieces that crop up from the site of
Vyāsa’s “appointment” to sire sons on the two Kāśi widows? The first two are
his insistence that what is needed with respect to dharma is that the two
widows undergo a year-long vow before he lies with them, and that should
they do this he will give his deceased brother sons the likes of Mitra and
Varuṇa (1.99.38–39). However general it may appear, the stipulation of a
year-long vow is an exact counterpart to the Aśvamedha requirement that the
king and his queens must remain sexually abstinent during the full year that
the horse wanders free.93 Moreover, Vyāsa’s precision regarding the pair
Mitra and Varuṇa is unusual in the epic and certainly has a Vedic ring.94 It
is important to remember who is speaking here and what he has been doing.
The last we heard of Vyāsa before Satyavatī recalled him as needed was that
he had gone off to the Himalayas to divide the Vedas and impart the
Mahābhārata as the fifth Veda to his five disciples, who were to proclaim it
as the Bhārata (1.57.73–75). These precisions about a year-long vow, night-
visitation, and Vedic-issue sons thus come from a Veda-knowing author who
could be alluding to what we learned in chapter 3 § D: that the Aśvamedha
variously identifies the king with dharma, and that the Rājasūya, the other
great Vedic royal ritual, invokes Mitra as “lord of truth” and Varuṇa as “lord
of dharma” in announcing the newly consecrated Bhārata king (MS 2.6.6;
TS 1.8.10.1–2). Such a momentary fusion of elements from of two great royal
rituals would fit our scene, since Vyāsa would be engendering just such a
king—or two. It is also in this domain of Vedic allusion that I would propose
some explanation of how and why Vyāsa would be encrypting himself as a
revived sacrificial horse.
First of all, the Mahābhārata is certainly in step with the dharmasūtra–
dharmaśāstra literature and the Mīmāṃsā in considering Veda as a source of
dharma. But more than these surrounding literatures, and also more perva-
sively than the Rāmāyaṇa, it is disposed to give new and often enigmatic shape
to Vedic usages and practices in its main and ancillary narratives. What we are
finding in our current skein is that this can be particularly the case where the
Vedic practices themselves—as with this kingless Aśvamedha scene, and also
with niyoga—are already enigmatic and of dark and doubtful dharma from the
standpoint of post-Vedic Brahmanical culture. Their cachet lies in their being

93. See Jamison 1996, 84: “During the year when the horse is journeying, the king lies nightly with his
favorite wife (Vāvātā), but does not have sex.” In ŚB 13.4.1.9, he thinks, “May I, by this austerity (anena tapasā),
reach the end of the year successfully.” Similarly, the horse on its travels is to be kept away from mares, and if it
should mate, an expiation is required.
94. Elsewhere in the Mahābhārata, I can find the two gods occurring alone together only with reference to
Agastya as their son (12.329.38; 13.151.3), and at 14.59.14.
 women’s dharma 379

reminders and remainders of dharma changing over time. It is thus telling that
it would be Vedavyāsa who confirms these practices as dharma. Secondly, Vyāsa
makes a singular appearance in the Nārāyaṇīya, which, I have argued (Hiltebeitel
2006a, 249–50), offers a bhakti decryption of a number of the epic’s Vedic and
purāṇic allusions. As mentoned in chapter 6, the Nārāyaṇīya comes to its deep-
est level of disclosure when Vyāsa tells Janamejaya, who has just been instructed
to perform a Horse Sacrifice on top of his Snake Sacrifice (12.334.8–9), about
the manifestation of Nārāyaṇa as the Horse’s Head (Hayaśiras) while it is
further disclosed that Vyāsa is a “portion” of Nārāyaṇa as well.95 These
Nārāyaṇīya enigmas give a glimpse of Vyāsa’s horseplay that might not be
limited to speaking from the horse’s mouth. The interpretation is in any case
proposed as similar to the one Jamison offers in interpreting Ṛgveda 10.86
about Indra’s monkey-companion Vṛṣākapi “as a veiled Horse Sacrifice” that
describes, among the benefits brought by “Indraṇī’s mating with Vṛṣākapi,”
the restoration of Indra’s worship with bulls and soma, the reaffirmation of his
power, his recovery of good erections, and his attainment of sons (1996, 74–88
quoting 81, 82–83).
Now Vyāsa only says he will sire sons the likes of Mitra and Varuṇa (Mbh
1.99.38b), not kings. But we know that is what is at stake even if we cannot
derive how this would have worked out. Dumézil seems to ignore this verse,
perhaps because it does not help his case that an original set of incarnations
has been effaced. For Dumézil, Pāṇḍu would originally have incarnated Varuṇa;
Dhṛtarāṣṭra and Vidura likewise the two “minor sovereigns” Bhaga (god of des-
tinies) and Aryaman (god of Ārya clans); and Yudhiṣthira would have been
sired by Mitra were it not that a “clumsy retouch” had replaced Mitra by Dharma
(1968, 146–48, 152, 159–60, 170–74). Going well beyond Jamison’s assump-
tion “that some version of the epic story of Ambā and her sisters already existed
in early Vedic” (1996, 304 n. 94; see n. 95 above), Dumézil was convinced that
he could recover a whole Mahābhārata whose “primary form [was] contempo-
rary with the oldest Vedic times, or anterior” (172). That idea seems now to be
a mirage.96 But the real point here is that Vyāsa does not get to sire sons who
would incarnate dharma in the likes and names of Mitra and Varuṇa because

95. Vyāsa’s Horse’s Head story (on which see chapter 6 § B) is called a “purāṇa equal to the Veda (purāṇaṃ
vedasammitam)” (12.335.7b). Recall that the Nārāyaṇīya gets to it by dipping to the outer dialogue frame so that
Śaunaka can ultimately hear what Vyāsa said to Janamejaya about the Horse’s Head.
96. I say this having once been devoted to the idea ([1976] 1990, 57–59; 1982b), and knowing that inter-
esting connections can still be posited on its basis (see Allen 1996, 2005, 2007a, 2007b). But the question of
genre is too important (see Hiltebeitel 2001a, 6–8), and I would agree now with Wulff 2008, 24 that “culture
contact” is usually and over all the better answer. See also chapter 1 n. 28 and Hiltebeitel forthcoming-a on the
improbability that an Indo-European “tribal” epic could lie at the base of the Mahābhārata.
380 dharma

Mother Satyavatī says matters are too pressing to give the widows a year for
such a vow. Instead of being a case of lost Vedic symmetries, it is a matter of
something willfully set ajar.
It is thus not only on male initiative that we find Aśvamedha pieces set
askew. Before Satyavatī gets to Vyāsa, she first tries to enjoin or appoint (vi-ni-
÷yuj; niyoga) Bhīṣma to sire sons with his brother’s “two mahiṣīs.” Can there be
two chief queens (mahiṣīs)? This is the only dual of mahiṣī (mahiṣyau; 1.97.9a) in
either epic. The Rāmāyaṇa’s first Aśvamedha incongruously requires more
than one queen to lie with the horse, but on unequal terms: only Daśaratha’s
aged queen Kausalyā passes the night with the horse out of desire for dharma
(dharmakāmyayā) as his mahiṣī, whereas his other two queens, Kaikeyī and
Sumitrā, cast in the well-recognized Aśvamedha roles of the favorite (vāvātā)
and discarded (parivṛktī) wife, respectively, are said only to have also united
with the horse, apparently more briefly (Rām 1.13.27–28; cf. Jamison 1996, 66,
87, 274 n. 104). Just as the Rāmāyaṇa has intelligible narrative reasons to inno-
vate in giving each of Daśaratha’s queens some time with the horse, the
Mahābhārata would have its reasons to be inventive in having two mahiṣīs.97 Yet
the reasons are immediately curious, since neither Ambikā nor Ambālikā
would be a mahiṣī if Ambā were still around. But of course this is Satyavatī
speaking, not Vyāsa. It is never explicitly two mahiṣīs when Vyāsa lies with
Ambikā and Ambālikā. Satyavatī’s usage at this point in speaking to Bhīṣma
would seem to be a reminder that Ambā’s unavailability as a Mother comple-
ments Bhīṣma’s as a Father, and an intimation that despite the legal incon-
gruity, to speak of Ambikā and Ambālikā as two mahiṣīs gives them both an
equal chance to become mother of the one desired royal heir.
Here some pieces noted by Biardeau and Jamison also fall into new places.
The older of the two widows, Ambikā, should be—now already, but especially
next with Vyāsa—not only (as per Biardeau) the single focus of the three names
of the Aśvamedha mantra; she should be the single mahiṣī. And further (as per
Biardeau), not only does Ambikā become a frequent name for the goddess; as
Jamison shows, hers is the predominant name among the “three Ambikās”

97. Both epics use the term mahiṣī almost exclusively for single chief queens, and where they do other-
wise it is with similar discordant notes for the real mahiṣī. See, for example, Arjuna, already married to
Draupadī, telling Kṛṣṇa his good fortune that Subhadrā will be his mahiṣī (Mbh 1.211.19); Mantharā telling
Kaikeyī that she (rather than Kausalyā) is Daśaratha’s mahiṣī (Rām 2.7.19); Sumantra, as her messenger, calling
her the mahiṣī when he tells Rāma that she and Daśaratha want to see him (14.11); Rāvaṇa inviting Sītā to be
his mahiṣī, with the odd qualifier that she will be his agramahiṣī, “primary chief queen” (3.4.24; 5.18.16)—a
term used nowhere else in the epics’ Critical Editions, although the Southern Recension finally makes
Śakuntalā the agramahiṣī of Duṣyanta (Hiltebeitel forthcoming-d), and Mandodarī would seem to have that
position among the wives of Rāvaṇa. On the other hand, in two plural usages it means mainly “woman” (Mbh
1.187.26; Rām 2.36.7; see 36.1).
 women’s dharma 381

referred to in Rudra’s epithet Tryambaka, which refers to Rudra as “possessing

three Ambikās.”98 Indeed, although the Aśvamedha mantra’s “three mothers”
are enough to call for the question of the epic’s use of these ritual names, it is
worth pressing Jamison’s extension to the discussion of the Traiyambaka
Homa further to ask whether the epic poets would have been alluding to this
rite as well, as Jamison wants to suggest.99 It would seem that the Traiyambaka
Homa’s “‘husband-finding’ spell”—a “somewhat sinister ritual performed
outside consecrated ground in an inauspicious, indeed dangerous place,”
a crossroads—would be an addendum to a year’s-end “Four-monthly” rite for a
girl who is running out of time in finding a mate, and who, with her parents
and no doubt at their urging, is calling on Rudra and his inauspicious sister to
apotropaically remove her unfavorable condition (Jamison 1996, 242–44; 303
n. 80). For Ambikā as Rudra’s sister is linked with autumn, which in some
texts is Rudra’s season of “special murderousness” due to her influence (cf.
241, 245, 304 n. 94).100 All this is fitting for an epic plot in which Ambikā, as a
result of her larger failure in meeting the reduced vow demanded by Vyāsa of
bearing “my smell, my looks, my garb, and my body” (99.43ab), would be dou-
bled by a second mahiṣī, Ambālikā, because Ambikā’s son will be blind and
unfit to rule.
Further, setting aside her assumption (to my mind unconvincing) that this
ritual, like the Aśvamedha, would have complemented an older version of the
epic story, Jamison also relates the Traiyambaka Homa to Bhīṣma’s abduction
of the three Kāśi sisters on two interesting points. First, remarking that “the
connection seems hard to gainsay, especially because the sisters also occur in a
marriage context,” she suggests that “the epic maidens would provide bad role
models for the husband-seeking girl of the Sākamedha” (245). Rather than
assuming an old para-Vedic story, however, I think it preferable to ask what the
epic poets might have made of the model of the Vedic rite. From this angle, the
“husband-finder” of the Traiyambakahoma would be a good role model, given
her bad situation, for something untoward to go further haywire in an epic

98. Jamison 241, 303 n.76. Although “Tryambaka” may—as usually translated—refer in the epic to Śiva’s
having three eyes (see Scheuer 1982, 237–36, 255–56), the matter is uncertain (Hopkins [1915] 1969, 220), and in
the one case where the Mbh gives an etymology (Vyāsa is telling Arjuna how Śiva preceded his chariot in battle), it
refers to his having three goddesses: “And since the Lord of the universe possesses three goddesses—Sky, Waters,
and Earth—he is remembered as Tryambaka” (7.173.89).
99. See Jamison 1996, 243. The Mahābhārata never mentions the pati-vedana or the Traiyambaka Homa.
Generally, the epics seem to overlook the Caturmāsyāni rites. It does mention a Traiyambaka Bali (7.56.1–4, esp.
3d): according to Scheuer 1982, 255–63, it is probably offered nightly throughout the war on Arjuna’s behalf by
his servants, after which Arjuna sleeps on the ground surrounded by weapons. Cf. also Scheuer 258 n. 23 on the
Pāṇḍavas’ offering (upahāra, bali) to Rudra Tryambaka (14.8.23–24; 64.1–8) after a night’s fasting to get hold of
the wealth needed to perform their postwar Aśvamedha.
100. The Mahābhārata does know Śiva as “Ambikā’s husband” (ambikabhartre), yet a brahmacārin (7.57.53).
382 dharma

series of turnabouts, which would include Bhīṣma’s turning of a husband-

finding ritual into a wife-finding one—something even more basic than his
turning a woman’s-choice svayaṃvara into a man’s-choice svayaṃvara, which,
as Jamison observes (1996, 299 n. 38, cited above), he does with verbal preci-
sion. Again it would be a matter of inversion via allusion, although in this case
less explicitly. Second, Jamison goes on to say that while Ambā became a mur-
derous avenger after the abduction causes her to lose her husband, “[e]ven
Ambikā and Ambālikā, though they settled happily enough into their married
life after their unexpected abduction . . . , might not have chosen this particular
method of pati-vedana [husband-finding] if they had their druthers” (245). But
we know they did have their druthers. They were going to have a svayaṃvara. It
would seem that as personifications of “the tryambaka,” the three Kāśi sisters
would be embodiments of a pretext-pati-vedana or “husband-finding” by
svayaṃvara that goes awry, at least for them, because Bhīṣma carts them off
Rākṣasa style to find their rather limited un-chosen husband Vicitravīrya, whom
Ambā in fact rejects.
Of course here we have a way to suggest that if the epic poets recall both Vedic
rites, it is separately: the “husband-finding” ritual would underlie the beginning of
the three sisters’ adventure into the Kuru dynasty. And the Aśvamedha invocation
of Ambā, Ambikā, and Ambālikā would underlie the point where Ambikā and
Ambālikā—and then Ambikā’s Śūdra servant—lie with the smelly misshapen
Vyāsa. Once again, there is something askew with the Śūdra servant in place of the
oldest Mother Ambā, something that carries forward into the unfolding plot and
allows us to move from this one threesome to the next: that is, to the transition
from the three Kāśi sisters to Gāndhārī, Kuntī, and Mādrī.
As we approach our skein’s last segment, let us take note of three con-
tinuities. First, as is apparent by now, in moving from generation one to
two through an intensification of Vedic allusions centered on the Aśvamedha,
we saw no diminution in allusive references to Mother Goddesses or to
Rudra–Śiva. In both cases, we would seem to have cumulative evidence that
the Mahābhārata either knows or anticipates more “purāṇic” mythology
about these deities than it tells us. Is Gaṅgā already married to Śiva when
she marries Śaṃtanu and becomes the mother of Bhīṣma? Is there already
a Kālī behind Kālī–Satyavatī? Do the three Kāśi sisters, probably from
Vārāṇasī, not only recall Vedic affinities with Rudra but already come from
a city associated with Śiva?101 It is unnecessary to push these associations to

101. Biardeau, however, proposes that Kāśi would index Buddhism “where the Buddha preached his first
sermon”; the two sisters’ distaste for Vyāsa would reflect a Buddhist preference for monks, who would not be
drawn into such a scene (2002, 1: 221–22). This seems strained, but, if so, reference to Śiva could cover, as else-
where, for Buddhism.
 women’s dharma 383

a positive answer on each or any count. The important point is that refer-
ences to Śiva and possible allusions to “later” goddesses also remain
prominent in the stories of Gāndhārī and Kuntī. If nothing else, the
Mahābhārata provides a semiotics for later goddesses to find their syntax
with Śiva. Second, we will notice in the movement from generation two to
three a certain consistency, along with a downward vector, that one might
seek to explain as an incomplete fulfillment of the three goals of human life
(the trivarga), perhaps complemented by a Dumézilian analysis of incom-
plete tripartition. While in each generation there are impressive martial
(artha or second function) feats by Citrāṅgada and Pāṇḍu and excessive
amatory (kāma or third function) ones by Vicitravīrya and Dhṛtarāṣṭra, righ-
teous pursuits (dharma or first function) are compromised, first at the top
by Bhīṣma and then at the bottom by Vidura, where dharma is derailed
from actual rule. Third, just as we may assume it would not have been easy
for Bhīṣma and Satyavatī to find a bride for Vicitravīrya, we may assume it
will not be so easy for them to find one for Dhṛtarāṣṭra. As we move from
the three Mothers of generation two to the three Mothers of generation
three, there will be in each case a seniormost mother (Ambā, Gāndhārī)
connected with Śiva who either rejects marriage into the line (Ambā) or
marries into its most unpromising senior branch after the line itself has
split (Gāndhārī).

F. Mothers Kuntī and Gāndhārī

So Bhīṣma once again brings in three women: Gāndhārī (103.9–17), Kuntī, and
Mādrī (105.1–6). As these accounts unfold, two things become apparent. First, it
quickly emerges that our skein makes Kuntī its new rising star,102 for between
Gāndhārī’s marriage and her own, a whole adhyāya (104) is dedicated to Kuntī’s
childhood. She is the only woman marrying into the chief royal line of either epic
whom we get to know well as a girl, and not only now but through her own adult
recollections. Second, this is the first generation in this skein in which there is
rivalry between “cowives.” Although there is rivalry between cowives in the lineage
before our skein (see Dumézil 1973, 16–18; Defourny 1978, 107–37; Brodbeck
2009a, 128–29), the present case takes us back above all—as Gaṅgā and Kālī–
Satyavatī did to the female weavers Dhātṛ and Vidhātṛ—to a prototype from the
early prolegomenal matter in Book 1: in this case, the rivalry between Kadrū and

102. Allowing for overlap and not counting descriptions of their sons at birth, from 103.9 to its end at
119.12 our skein gives 195 verses to Kuntī, eighty-five to Mādrī, and thirty-six to Gāndhārī.
384 dharma

Vinatā, the cowives of the Ṛsị Kaśyapa and the mothers of snakes and birds.103
Gāndhārī and Kuntī compete over who will be the first to give their husbands an
heir, and then Kuntī and Madrī show some rivalry. If one recalls that Kuntī is
̣ ạ ’s paternal aunt, one has one of the reasons why these rivalries revolve around
Kṛsn
her, and why she is pivotal to the change, noticed in the previous section, to Vyāsa
and Kṛsn ̣ ạ in the handling of the family’s dynastic fortunes. Yet even as Kuntī
becomes salient, our skein—right down to its well-marked end—keeps all three
generations of women (Gaṅgā, on this point, excepted) in the living picture.
Bhīṣma prefaces his renewed matchmaking in a passage we have already
noticed. Addressing Vidura, he mentions the continuity (saṃtāna) now
established in the persons of Dhṛtarāṣṭra, Pāṇḍu, and Vidura as “threads of the
line” (103.3). But although Bhīṣma mentions Vidura as one of the “threads,” he
rather tries to draw him into being one of the threaders: “For this lineage to
grow like the ocean, son, I, but especially you, must take care . . .” (103.4).
Bhīṣma speaks favorably of three royal lines he considers to have produced
suitable brides for Dhṛtarāṣṭra and Pāṇḍu, and, saying he thinks these pros-
pects should be wooed, asks Vidura’s opinion. But Vidura demurs: “You are
our father, our mother, and our supreme guru”; Bhīṣma should decide and act
on these matters himself (5–8). Calling Bhīṣma “mother” here could suggest
that Satyavatī, Ambikā, and Ambālikā are not being consulted.
Biardeau offers an incisive key to interpreting Bhīṣma’s choices: Kuntī’s
link with the earth, most notably through her name Pṛthā, is with the earth
totally, even though her Yādava people are not associated with any land, whereas
the names of Gāndhārī (Woman of Gāndhāra) and Mādrī (Woman of Madra)
identify them only by their lands, and with neighboring northwestern lands
ruled by or associated with incarnate demons in an area propagated by
Buddhism (2002, 1: 231–33). Indeed, Madrī is even called Bāhlīkī (Woman of
Bactria)104 once (1.116.21a), when Kuntī says how lucky Mādrī was to have been
the last to see Pāṇḍu’s face making love.

103. See Mbh 1.13.35–14, 18–23, 30.11–35, and 49.3–16. Kadrū and Vinatā’s parallels with Gāndhārī and
Kuntī, widely noted, include rivalry, use of pots for gestation, and the servitude of the one’s sons to the other’s.
Although the parallels crisscross (most notably, it is Vinatā who has a long and aborted gestation), Gāndhārī has
more in common with Kadrū (mother of a thousand snakes) and Kuntī with Vinatā (mother of two birds). But
the key theme of a mother (Kadrū) cursing her sons is unparalleled, unless one reads Gāndhārī’s boon as a
parallel, or perhaps Kuntī’s abandonment of Karṇa.
104. Karṇa mocks the low conduct of Madras at Mbh 8.27–30, once mentioning Gāndhāras, Madras, and
Bāhlīkas (Bactrians) together (27.55cd), particularly deriding Madra women. See Biardeau 2002, 2: 313–15 and n.
11 on the Vāhīka-Bāhlīkas, inclusive of Madras, as (by Karṇa’s etymological tale) those from outside Āryavarta,
descended not from Prajāpati but from two “ghastly” Piśācas who now inhabit the former Vedic heartland of the
Punjab (cf. Hiltebeitel [1976] 1990, 272–73). This Āryavarta corresponds, in fact, to the Buddhist majhimadesa
described in chapter 7 § B.2.
 women’s dharma 385

Dumézil, however, appears to have made the most cogent attempt to inter-
pret these three marriages as an ensemble, taking Bhīṣma to have implemented
them in trifunctional order as “un théorie en action des modes archaïques de
marriage” (1979, 71). Dhṛtarāṣṭra’s marriage with Gāndhārī, who is “given” to
him by her father, mother, and brother (Mbh 1.103.11e, 12c, 15a) with a large
dowry (14c), has earmarks of a Brāhma-mode marriage; Kuntī’s “self-choice”
(svayaṃvara) of Pāṇḍu comes next; and Pāṇḍu’s marriage to Mādrī by pur-
chasing her would be an Āsura-mode wedding (69–71). Dumézil thinks the
operative theory would be older than the classical enumeration of eight types of
marriage in which the Brāhma mode can be deemed suitable for Brahmins; the
svayaṃvara—discounted in the dharmaśāstra lists—special, as we have seen in
§ D above, in the epics, at least, for Kṣatriyas; and the Āsura mode for Vaiśyas
and Śūdras (33). But again, while the theory may be old and parallelled by
marriage sequences in other Indo-European heroic traditions, this does not
give the “entire episode” a special “antiquity” (70). Rather than reflecting an
older account of an archaic trifunctional order translated and updated into
terms of classical dharma, the epic continues to have Bhīṣma arrange mar-
riages in which dharma remains askew. Kuntī’s marriage to Pāṇḍu does not
seem to present any immediate dharma problems, but considering that Kuntī’s
Yādava origins connect her only loosely with a royal lineage,105 it is not clear
why she would be having a svayaṃvara at which to have rather miraculously
“found (avindata)”106 Pāṇḍu “in the midst of a thousand kings” (105.2cd)!
Mādrī’s marriage in the Āsura mode can also be complicated, as Dumézil
acknowledges elsewhere (1968, 75–76), since her brother Śalya is an incarnate
Asura or demon.107 But it is especially in finding a bride for Dhṛtarāṣṭra that
Bhīṣma sets things askew.
Here, where Dumézil sees the most “august” of the Indo-European “pro-
cedures ‘civilisés’” reserved for Dhṛtarāṣṭra, his handling shows the strain of
his compartmentalization. Dumézil presents as lofty a picture as he can: “had
it not been for his infirmity,” Dhṛtarāṣṭra “would have been king,” and his
blindness is a kind of insight (1979, 70). But he fails to appreciate that in
arranging for Dhṛtarāṣṭra to receive a bride as a gift, Bhīṣma enables a violation
of the Kṣatriya ethos of not accepting gifts that he himself had so ably put on

105. The Yādavas descend from Yadu, son of the lunar dynasty ancestor Yayāti, who cursed Yadu and his
descendants to have “no share of royalty” (Mbh 1.79.7). A Northern (N) passage (1.1129*) compounds the story at
this point, having Kuntibhoja arrange the svayaṃvara when no king had asked for Kuntī’s hand!
106. Cf. Brockington 2006, 38. We see again in this formation from ÷vid that a svayaṃvara could be a kind
of “husband-finding” (pati-vedana).
107. A lengthy N passage (Mbh 1, App. I, No. 61) has Śalya, already king of Madra, stipulate that his sister’s
purchase is a matter, whether good (sādhu) or not, of what his family (kule; lines 24–25) has always done.
386 dharma

record when he took the three Kāśi sisters ungiven yet gave them to Vicitravīrya.108
This time, when Bhīṣma goes to extraordinary measures to get someone else to
give a bride to a problem child, it is not in an “august” remnant of a more
civilized time but in the sad light of Dhṛtarāṣṭra’s blindness. Indeed, when
King Subala of Gāndhāra receives Bhīṣma’s messenger, he agrees in mind
(buddhyā) about the groom’s lineage, fame, and conduct, but only after consid-
ering (prasamīkṣya) that Dhṛtarāṣṭra is blind (103.11; cf Brodbeck 2009a, 168).
But it is when Bhīṣma first hears about Gāndhārī herself that we might wonder
how desperate, or perhaps heedless, he is:

Then Bhīṣma heard from Brahmins that Subala’s daughter Gāndhārī

had propitiated the boon-granting god who took the eyes of Bhaga,
Hara (bhaganetraharaṃ haram), and that the auspicious Gāndhārī, it
was said, had obtained the boon of a hundred sons. (103.9)

The associations with Śiva of course continue, but their ominousness is now
rather evident. As we descend into blindness along with this bride who will be
given, Śiva is doubly Hara, the god who “takes” or “removes,” and what he
removes is sight itself from the god of destinies—something that might
complement what we have understood so far: that Dhṛtarāṣṭra’s blindness
resulted from Little Mother Ambikā’s shock upon seeing Vyāsa. If Bhīṣma is
“taking care” in selecting this bride, it would seem to have been bad judgment
on his part, at least knowing of this boon from Śiva, for thinking a hundred
sons was a good idea,109 and on Vidura’s part for giving Bhīṣma carte blanche
in all the matchmaking.
Gāndhārī, a woman who fares in dharma—a dharmacāriṇī (1.103.11ef)—
has already been given by her father when we meet her. Realizing that her
mother and father wish to give her to a blind man (12), she blindfolds her eyes
and says, at least literally, “I would not eat before my husband” (nātyaśnīyāṃ
patim aham; 13e). In saying this, she is twice said to be “vowing utter fidelity to
her husband” (13d; 17b); but taken along with Vaiśaṃpāyana’s adhyāya-closing
statement that she never again mentioned “other men” (17cd), her first curt
and haunting words could suggest that the bandaging of her eyes is not only an
act of spousal consolidarity and deference (not eating before one’s husband is

108. See Hara 1974, 304–5; Jamison 1996, 235; and Heesterman 2001, 255 n. 31, on the Kṣatriya’s
“warrior ethos, which forbids him to accept, let alone ask for gifts,” as an underlying explanation for the
Rākṣasa mode of marriage.
109. A hundred sons is not inherently a bad boon. Sāvitrī gets Yama’s boon of a hundred sons to trick him
into keeping her husband alive, and the same number for her husband’s parents (Mbh 3.381.44–58). Here the
number seems to signal the spread of a martial clan (as it does where the Kauravas are worshiped as clan deities
in the Tamil cult mentioned in § B above).
 women’s dharma 387

just ordinary good behavior) but a curbing of her desire—about which there is
more that can be said. Mahābhārata folklore of northwest Tamilnadu knows an
irresistibly pertinent story. Gāndhārī once wanted to make Duryodhana invul-
nerable by the power amassed from bandaging her eyes. She would lift her
bandage to see him, and told him to come before her naked. But Kṛṣṇa got
Duryodhana’s sister Duḥśalā to appear, and Duryodhana, feeling shy, covered
his genitals and thighs with a banana leaf, leaving that area vulnerable when
Gāndhārī looked. The subtext of desire is evident: Gāndhārī wanted to see her
son naked. And the banana leaf could suggest that Duryodhana sensed that she
wanted not only to see but eat, which could recall what she says in Sanskrit
about not eating before her husband.110 For now, however, we must appreciate
that Dhṛtarāṣṭra’s Little Mother Ambikā and his fiancée Gāndhārī have both
closed their eyes around their desires. But whereas Ambikā did so for only a
fairly short but fateful time—maybe just upon first seeing Vyāsa, probably for
the duration of their unpleasant encounter—Gāndhārī has vowed to close them
for a lifetime.
Before the Critical Edition gets to Vaiśaṃpāyana’s briefer descriptions of
Bhīṣma’s arrangements for Kuntī and Mādrī to marry Pāṇḍu, it now pauses for
an adhyāya, as noted, over Kuntī’s childhood to tell how she became a mother
even while remaining a kanyā, and to explain her secrecy about Karṇa, the son
she bore and abandoned “to hide her misconduct and out of fear of her rela-
tions” (1.104.13cd). Like Gaṅgā, Kuntī puts her son in a river,111 and like Satyavatī
she gets to remain a kanyā and hides the youthful affair from her adoptive rela-
tives.112 But in Kuntī’s case, compared with Satyavatī’s, being a kanyā has more
the surcharge of her being “just a girl.”113 Unlike Satyavatī, Kuntī does not
bargain to remain a virgin, she just remains a girl: indeed, in a fuller version of
her story, “a free female” (svatantrā). In that elaboration, Sūrya offers gratis that
she will remain a kanyā even after their union, clarifying with a contrived but
surprising etymology: “O sweet-smiler, neither your father, mother, nor elders
prevail; hail to you of choice hips. Hear my word. A free female (svatantrā),

110. Perundevi Srinivasan is gathering variants of this story. I thank her for her thoughts on the banana leaf.
111. Of course, whereas Gaṅgā “immersed” (amajjayat; 92.44d) her first seven sons as they wished so that
they could return to heaven, Kuntī “abandoned,” “cast,” or “poured out” (utsasarja; 104.13c) Karṇa to a life of
earthly resentments.
112. Their adoptions are curious. Both have their premarital affairs apart from their “real” parents
(Satyavatī’s mother being an Apsaras-turned-fish). They also attenuate their already loose ties to royal lines. And
Kuntī’s foreshadows that of Karṇa.
113. Van Buitenen captures this sense in translating what Kuntī says when she tells Karṇa shortly before
the war how she conceived him: “Then, out of curiosity and childishness, . . . I, being just a girl (kanyā satī), made
the Sun god come to me” (5.142.21, 23–25; van Buitenen 1978, 452). Van Buitenen 1973, 241 missed the oppor-
tunity to translate kanyā satī in this fashion at 1.104.8cd. Cf. Oliver 1993, 54 on Kristeva’s view of “virginity.”
388 dharma

since she desires all (sarvān kāmayate yasmāt), is called kanyā from the root
kan” (3.291.12–13). Sūrya says, I believe, that as a free female, a girl is free to
desire the world (or all things) and to make the world (or all things) desirable.114
Known from birth as Pṛthā—“the Wide,” evoking the Broad Earth Pṛthivī and
her “girlish wide eyes” (kanyām . . . pṛthām pṛthulalocanām; 3.287.12cd)—she is
the eldest child of the Yādava chief Śūra and older sister of Vasudeva (who is
perhaps still yet to become Kṛṣṇa’s father), and she gets her name Kuntī from
her adoptive father Kuntibhoja, the otherwise childless son of Śūra’s father’s
sister, because Śūra had promised and given Kuntibhoja his firstborn child as
“a friend to a great-souled friend” (sakhā sakhye mahātmane; 1.104.1–3). This
exchange is not one that Kuntī will later recall happily. As she tells Kṛṣṇa dur-
ing his prewar embassy to the Kauravas,

I censure not myself nor Suyodhana,115 but my father by whom I was

transferred to Kuntibhoja as wealth is by rogues. I was a child playing
with a ball in my hand when your grandfather gave me to Kuntibhoja
as a friend to a great-souled friend! I was humiliated (nikṛtā) by my
father and maternal uncles,116 Foe-scorcher. (5.88.61–63b)

Yet from this disquieting friendship unfolds the friendship of Kṛṣṇa and Arjuna
and a hidden friendship between Kṛṣṇa and Karṇa that we will return to in
chapter 12 (see Hiltebeitel 2007b, 30–31). For now, however, the focus is on
how Kuntī got the mantra that made her a special kind of mother:

In her father’s [Kuntibhoja’s] house she was appointed (niyuktā) to

the honoring of Gods and guests; and so she came once to serve that
fierce and terrible Brahmin of strict vows whose design in dharma is
hidden (nigūdh ̣ aniścayaṃ dharme), whom they know as Durvāsas. This
fierce man of honed spirit she satisfied with all her efforts, and the
Muni, with foresight into the Law of Distress (āpaddharma-anvavekṣayā),
gave her a mantra combined with sorcery (mantram . . . abhicāra-
abhisaṃyuktam) and said to her: “Whichever God you call up with
this mantra, by this or that one’s grace there will be a son for you.”
When that Brahmin had said this, from curiosity, and being just a
girl (kanyā satī), the famous one then invoked the Sun God. . . .
(1.104.4–8)

114. The causative kāmayate (÷kam) reads either way; ÷kan means “to be satisfied or pleased; to agree to, to
accept with satisfaction; to please, be liked or wished for.”
115. A name for Duryodhana: Biardeau (2002, 1: 886 n. 4, 900) suggests, “Easy to combat.”
116. Śvaśura (plural): usually “fathers-in-law” (see van Buitenen 1978, 370), but “also applied to a maternal
uncle and any venerable person” (MW 1105). This is surely Kuntī’s meaning. She is complaining about her
father’s deal with a maternal uncle, and had no fathers-in-law yet (she calls Vyāsa her śvaśura at 15.38.1a).
 women’s dharma 389

From this momentous passage, whose essentials Kuntī herself will soon
rephrase,117 we need only know that Durvāsas is an irascible and inveterately
hungry Ṛṣi who incarnates Śiva, to sense all that would be at work when,
with “foresight in āpaddharma” and as “one whose design in dharma
is hidden,” he gives this sorcerous Mother-making mantra to this
wide-eyed girl.
In the fullest version of this story, Kuntibhoja charges Kuntī to give to
Durvāsas ungrudgingly or disinterestedly (amatsarāt; 3.287.15d), to which she
replies, “It is by my own nature (svabhāva) that I would honor the twiceborn”
(288.2ab)—powerful words from a girl. Here no sorcery is mentioned, but she
seems to paralyze the god (289.17–18), and Sūrya calls Kuntī mattakāśinī
(291.21), which could be “bewitching one.” Much later, when she asks Vyāsa to
raise the slain Karṇa from the Gaṅgā so she can see him while Dhṛtarāṣṭra and
Gāndhārī also see the slain Kauravas, she has more rueful memories: that
Durvāsas and Sūrya both threatened to curse her, and that she again became a
kanyā by the god’s grace as Durvāsas had foretold (15.38.1–17). Here she also
recalls that Durvāsas told her she would become the mother of Dharma (dhar-
masya jananī; 6a)! In any case, as we shall soon see, not only the gods but all the
Ṛṣis must know that Kuntī keeps her secret sorcerous knowledge into her
marriage to Pāṇḍu as part of her “innate” service to the twiceborn, and thus to
the Ṛṣis themselves. This makes it part of the divine plan. But more than this,
again something we could call primordial is drawn from the distaff side into
the Kuru line: this time in a secret that the gods and Ṛṣis keep with a “free
female” girl made old (as a mother) before her time. I would suggest that some
such sense lies behind an understanding of Pṛthā–Kuntī in the north Indian
Mahābhārata folklore of sub-Himalayan Garhwal, where she is renowned for
her great old age in the Pāṇḍav Līlās (dance-dramas about the Pānd ̣ ̣avas and
company) and known as the elder sister of Bhūmī, Mother Earth (Sax 2002,
71–74)! In one story, Arjuna follows her and Draupadī—his mother and his
wife—only to find out who they really are: the chief “hags” among the Sixty-
Four Yoginīs, with Kuntī the eldest among them, who appear before him as
vultures to forecast the Mahābhārata war (Sax 2002, 153–55; cf. 144). “Hags” is
Sax’s translation of Garhwali pañcāli, “bird” (Hindi pakṣī), evoking Draupadī’s
name Pāñcālī. Note that Kuntī and Draupadī are birds, of whom one is soon
reminded—along with the archetypal story of Kadrū and Vinatā—by the story
of Gāndhārī’s pregnancy.

117. She repeats much of 1.104.4–6 at 1.113.31–33, condensing “gods and guests” to the essential “guests,” and
speaks of her “sacrifices” (yajñair) rather then “efforts” (yatnair).
390 dharma

As Gāndhārī’s story is told, it mirrors Kuntī’s in ways that deepen their

differences around themes we have been discussing.

Gāndhārī once satisfied Dvaipāyana [Vyāsa] when he had arrived

exhausted with hunger and fatigue. Vyāsa granted her a boon.
(Mbh 1.107.7)

Vyāsa’s hunger, elsewhere unheard of, is here to match that of Durvāsas,

which in Kuntī’s story is only implicit yet there by his ornery reputation. Each
Ṛṣi is “satisfied” using the same verbal root tuṣ: just as Kuntī “satisfied”
(atoṣayat; 104.5c) Durvāsas, an incarnation of Śiva, and Gāndhārī “satisfied”
(toṣayām āsa; 107.7c) Vyāsa, an incarnation of Viṣṇu–Nārāyaṇa. Their involve-
ments thus reflect the intertwined workings of the two major gods, and not
just on the sides where one would expect them. Yet if these two Ṛṣis are simi-
larly satisfied, they reward their hostesses differently. Durvāsas gives an unso-
licited mantra to a curious unmarried wide-eyed girl that opens the world
ahead of her. Vyāsa gives a boon or choice (varam) to a married woman who
has closed her eyes to the world, who now chooses—and there is no saying
what else she could have chosen—only to go deeper into her own darkness by
fine-tuning the troublesome boon that she has already gotten from Śiva: “She
chose that her hundred sons would resemble her husband (sadṛśam bhartuḥ)”
(107.8ab)—a man whom she had never seen, and how they should resemble
him she does not say (it is tempting to take sadṛśam, “resembling, similar to,
with the appearance or look of,” as a pun on Dhṛtarāṣṭra’s blindness).118 Indeed,
Vyāsa had already predicted at Dhṛtarāṣṭra’s conception that this son of his
would have a hundred powerful (mahābalāḥ) sons (100.10ab), so if Gāndhārī
adds anything, it is not to their strength. And while it is not clear whether
Vyāsa adds anything at this point either, the outcome is that Gāndhārī will
have all her sons, plus an unasked-for daughter gratis,119 the aforementioned
Duḥśalā, all in one horrendous two-year pregnancy that she interrupts only
to abort it when she hears that Kuntī has had a son “of splendor like the

118. Cf. Biardeau 202, 1: 230: to wish for sons like their father is “the wish of every femme dharmique, but
is it wise in this case?” As with Gāndhārī’s utterance when she puts on her blindfold, her everyday wifely senti-
ments sound foreboding. Brodbeck 2009a, 168–69, however, implies that Śiva’s and Vyāsa’s boons to Gāndhārī
are the same.
119. In a widely found adhyāya (omitted only in four N mss., including the important Śāradā one [on which
see Sukthankar 1933, lxv]), Gāndhārī interrupts Vyāsa to ask for a daughter (1 Appendix I, no. 63, after 1.107;
Ganguli [1884–96] 1970, 1: 260–61). I stress this daughter’s gratis appearance because it parallels Draupadī’s
unasked-for birth (see Hiltebeitel 2001a, 187), which—in Draupadī’s case—makes it part of the divine plan, for
Draupadī is born specifically to do surakārya, “the work of the gods” (Mbh 1.155.45; see chapters 10, 12). I would
not agree that Duḥśālā “is apparently no part of Vyāsa’s plans for the fleshball” (Brodbeck 2009a, 170).
 women’s dharma 391

morning sun (bālārkasamatejasam)” (107.10ab)—presumably Yudhiṣṭhira but

sounding (no doubt intentionally) more like Karṇa, as is also the case when
Gāndhārī tells Vyāsa just after this that she aborted her belly because she had
“heard that the eldest son of Kuntī (jyeṣṭhaṃ kuntīsutam) was born of splendor
like the sun (ravisamaprabham)” (107.15).120 It would seem that we are entitled
to fantasize. Although jyeṣṭha could mean “most excellent” rather than “eldest,”
the latter very common meaning, when referring to siblings, would suggest
that Gāndhārī knows that Kuntī now has more than one son! Yet it would still
take a lot to explain why Gāndhārī would be jealous about a déclassé son born
years ago out of wedlock, or why she would be referring to Karṇa while really
worrying about Yudhiṣṭhira. The ambiguity would seem to foreshadow the
emerging affinities, despite their opposition, between Karṇa and Yudhiṣṭhira
as the two most legitimate heirs to the throne. In any case, Gāndhārī’s miscar-
riage requires Vyāsa himself to save the day, as it were, by dividing the mass
of flesh into a hundred and one pieces while giving directions on how to let
them continue to gestate in pots.121
Even though we cannot be sure whose birth Gāndhārī is talking about,
much less how she could have heard about the birth of either Karṇa or
Yudhiṣṭhira at this point, her hearing about Kuntī’s sunlike firstborn son makes
their rivalry the turning point in the Kauravas’ birthing. This simultaneity was
important enough to have required this seeming mention of Yudhiṣṭhira’s
birth before our skein actually gets to it, and to interrupt the narrative of Pāṇḍu’s
strange reign—at least in the Critical Edition, which again, I believe, shows the
soundness of defaulting to the Northern Recension where the Southern
Recension gives the story differently.122 In the Critical Edition, Pāṇḍu’s rule is
narrated in two stages that cover three phases of his career, with the middle
phase interrupted by the story of Gāndhārī’s pregnancy. These three phases

120. As Brodbeck has noted (personal communication), and, though I take the possibility that these allu-
sions are to Karṇa from him, he credits them (Brodbeck 2009a, 170 n. 12) to Bowles 2008, 1: 44 n. 5.
121. Mbh 1.107.13–23. Biardeau 2002, 1: 233 finds it tempting to link Gāndhārī’s name Saubaleyī here with
the bali offerings (one of the five mahāyajñas) made to inferior divinities and thrown on the ground.
122. As Sukthankar 1933, 474–75 shows, the Southern Recension’s “entirely different” handling of
1.106.11–114.15ff. makes the Pāṇḍava and Kaurava princes’ births follow one temporal line. Beginning with an
excision of 109.1–4 where Janamajaya asks Vaiśaṃpāyana to return to the story of the Pāṇḍavas’ partial divine
incarnations, it does not divide Pāṇḍu’s story by Gāndhārī’s, and continues with Pāṇḍu’s from 109.5 through
Yudhiṣṭhira’s birth before it gets to Gāndhārī’s pregnancy. Meanwhile, as S joins 107.8 to 111.12 to make Pāṇḍu’s
story continuous, it drops the first reference to Gāndhārī’s learning of Kuntī’s sunlike son (107.10), and once it
has recounted her pregnancy, supplies six lines (1135*) to take one back to Pāṇḍu’s plight. Gāndhārī’s remaining
mention of Kuntī’s sunlike child (107.15) is thus deferred to follow Yudhiṣṭhira’s birth (114.1–7), eliminating any
ambiguity as to who he is and diminishing her rivalry with Kuntī. The extensive revision is another example of
S’s replacement of ambiguity and literary experimentation with a flatter and more easily communicated orderli-
ness. See Hiltebeitel 2006a, 252–53; forthcoming-a; forthcoming-d; Mahadevan 2008.
392 dharma

clearly follow the one-sided role models set for him by his three or four
“fathers.”123 The first (105.6–106.5) begins when he has married Mādrī:

When he had wed her, Pāṇḍu, who was endowed with both strength
and enterprise, desired to conquer the earth and went at his enemies
in all their multitudes. (105.7)

In this expansionist124 martial phase, he is like Citrāṅgada and/or Bhīṣma, but

his conquests—most notably in Magadha as part of a drive to the east (105.
10–12)—now take place in a “real” political geography that will also challenge
his Pāṇḍava sons (see Biardeau 2002, 1: 225). Note that when Pāṇḍu returns
and greets the feet of his “father” (105.25), the narrator is referring to “father”
Bhīṣma just as Pāṇḍu is about to deliver the war booty to him, Satyavatī, and
Ambālikā, with something too for Vidura (106.1–2). Second is a phase of
hunting life, in two parts. First, before we hear about Gāndhārī’s pregnancy, we
learn that Pāṇḍu was provisioned in the forests by his brother Dhṛtarāṣṭra and
joined there by his two wives in an amorous setting.125 Then after the narration
of Gāndhārī’s boon and the births of the hundred Kauravas (107–8), Pāṇḍu
shoots a Ṛṣi named Kiṃdama,126 who is mating while disguised as a deer
(109.5–31). According to the dying Kiṃdama, Pāṇḍu is now tinged with lust
and greed (109.11cd), which Pāṇḍu then links back to the lustfulness of
Vicitravīrya as a strong vice he must get rid of (110.2–6). Note that Kiṃdama
curses Pāṇḍu because of his cruelty in shooting a mating creature, which was
“most unrighteous” (adharmiṣṭham; 21d) because it “frustrated a cherished
fruit of the puruṣārthas” (19e; cf. 23d)—that is, the trivarga—which, as a king,
Pāṇḍu was supposed to protect. Finally, Pāṇḍu’s third phase (110.3–116.12)
occurs under Kiṃdama’s curse that he too will die should he engage in the act
of love. Even though he manages to have sons before he dies under these con-
ditions, “his last years pass without his exercising his royal duties in a normal
manner” (Biardeau 2002, 1: 237). Here Pāṇḍu’s turn to sexual continence puts
him again in the same league as Bhīṣma. But more than this, what motivates
him to practice tapas is the example of his “real” father Vyāsa (110.3–6, 29); and
if Vyāsa is his model, it could also motivate Pāṇḍu’s eventual incontinence.127

123. Cf. Brodbeck 2009a, 175, making do with having Pāṇḍu “imitate aspects” of only two fathers, Vyāsa
and Vicitravīrya. The additional two would be Bhīṣma and Citrāṅgada.
124. See 102.12; 105.19–23 and 26: he reclaims kingdoms (rāṣṭrāṇi) the Kurus had lost, and takes captives
from them.
125. Mbh 1.106.6–11. Biardeau 2002, 1: 235 says Pāṇḍu is now living an irresponsible life, having given up
his responsibilities in the capital.
126. See Biardeau 2002, 1: 226 n. 3: This Ṛṣi “is not a model of asceticism,” his name meaning “Who has
some self-mastery,” “More or less master of himself.”
127. Vyāsa sires his first son Śuka by ejaculating after seeing a nymph transformed into a female parrot
(12.311.1–10; see Hiltebeitel 2001a, 286–87), and also interrupts his tapas to sire Dhṛtarāsṭ ra, Pānd
̣ u
̣ , and Vidura.
 women’s dharma 393

Throughout all this, poor Pāṇḍu seems to have only increasing difficulties with
this fractured paternal imago, and gets very little from his silent mothers. What
good fortune he has must come from his wives, who love him.
Still following the Critical Edition and the Northern Recension, we now
come to the turning point in the life of Kuntī, Mādrī, and Pāṇḍu. In four
verses that the Southern Recension discards (109.1–4; see n. 122 above), King
Janamejaya, having heard about Gāndhārī’s pregnancy and the news she gets
of Kuntī’s sunlike child, now asks Vaiśaṃpāyana to tell him in full about the
Pāṇḍavas’ births as partial incarnations of gods. It is these verses that prompt
Vaiśaṃpāyana to relate the story of Pāṇḍu’s deer hunt and Kiṃdama’s curse,
which occurs, of course, in both recensions: Pāṇḍu will die “when you are
lying with a woman you love, overcome by love, blinded by your passion,”
moving in that happiness not only himself but bringing about the death of
the beloved he lies with, who will follow him out of devotion (bhakti), coming
under the sway of the King of Ghosts.128 It could be either wife, and it is
clearly bad news for both of them even if it is not clear what they learn about
the curse. There is nothing to indicate that they heard it uttered, or what
Pāṇḍu told them about it. All we know is this: when the deer-Ṛṣi died after
speaking, Pāṇḍu was overwhelmed with grief (śokārtaḥ); then together with
his wives he was stricken with grief and sorrow (śokaduḥkhārtaḥ) as they
lamented the deer like a kinsman (109.31–110.1)! As Pāṇḍu turns his mind to
asceticism, he speaks only of how all this affects him. And with that we
embark on the third and last phase of his career—modeled, as he says, after
Vyāsa—and the point toward which this chapter (and indeed, our whole
skein) has been driving: his dialogue with Kuntī. But we cannot get there
before another intervention by the Ṛṣis.

G. Kuntī, Mādrī, and Pāṇḍu among the Hundred Peak

Mountain Ṛṣis

Pānḍ u
̣ now utters one of the Mahābhārata’s moving evocations of the renunciant
life, holding it up as a sad but also beautiful ideal that the epic, however, will not
allow its householder kings to live. Indeed, it will be like father like son all over
̣ u
again. Just after Pānd ̣ says he will emulate his father Vyāsa by yoking himself to
austerities (110.6), he expresses himself in a train of thought (110.7–18) that his
son Yudhiṣṭhira will also follow, quite precisely from point to point and using

128. Mbh 1.109.28–30. The description suggests that this double Liebestod under Yama’s sway could
evoke Yama’s connections with possession; see chapter 8 § D.2.b.
394 dharma

many of the exact same words, phrases, and lines,129 when he says he wants to
renounce the world after the war upon Kuntī’s revelation that Karṇa was his elder
brother. No matter how poignantly such a royal householder expresses this ideal,
it moves his household to speak against it, and Pānd ̣ u
̣ ’s entire household at this
time and place is Kuntī and Mādrī, who now protest when he finishes this speech,
which seems to ask them, while he remains behind to begin a life of solitude, to
accompany Pānd ̣ u
̣ ’s retinue (there are unnamed Brahmins, followers, and
servants with them [110.37, 39]) to Hāstinapura, the City of the Elephant, to bring
the court there word of his decision (110.22–24). Saying the same words jointly
(25), Kuntī and Mādrī reply,

There are other stages of life (anye . . . āśramāḥ) that you can
undertake together with us, your wives by the Law (dharmapatnīs),
O Bharata Bull, and still do great austerities. And you surely and
without fail will find heaven too. We shall abjure all our senses and,
devoted to our husband’s world and forsaking the pleasures of love,
we too shall undertake severe austerities. If you desert us, wise king
who are the lord of your people, then of a certainty we shall give up
our lives this very day. (110.26–28)

Kuntī and Mādrī thus speak jointly as dharmapatnīs (not as co-mahiṣīs) to

redirect Pāṇḍu toward an āśrama or “life-stage” that allows their participa-
tion.130 If that is their will, he says, he will be a vānaprastha with them and
follow the “harsh and ever harsher rule (vidhi) of the Forest Treatises”
(110.34–35). Although only Pāṇḍu mentions texts, his two dharma-wives
would seem to know the gist of what is in them. The three now remove their
royal paraphernalia for their retinue to take back to Hāstinapura along with
word of his decision (36–40); and when Dhṛtarāṣṭra hears “from them as to
all that had happened in the great forest, he mourned after Pāṇḍu” (41).
Although Vaiśaṃpāyana has by now (in the Critical Edition) told Janamejaya
about Dhṛtarāṣṭra’s hundred sons, Gāndhārī has not yet had them. But the
race is on and no one could now expect Kuntī to become a mother.

129. Mbh 12.9.12–28; see Klaes 1975, 113, 129 n. 10; Fitzgerald 2002, 670 n. to Mbh 12.308.36. Yudhiṣṭhira
first projects himself as a solitary vānaprastha (12.9.1–11; Fitzgerald 2004a, 685); but then, as his begging shows,
he sees himself as a renunciant. The parallels are as follows: 1.110.7–9 ≈ 12.9.12–14; 1.110.10 ≈ 12.9.17; 1.110.11 ≈
12.9.16 (with a verse-order reversal); 1.110.12ab ≈ 12.9.23ab; 1.110.14 = 12.9.25 excepting a verb; 1.110.15 = 12.9.24
except for a word-reversal; and 1.110.16–18 ≈ 12.9.26–28, with 16ab = 26ab, 17b = 27b, and 18cd = 28cd about
seeking “the Law (dharman) of the wind.” Most likely Pāṇḍu’s is the trunk version while Yudhiṣṭhira is given
some intermittent fancies (e.g., he will be the last after-hours guest on his begging rounds [12.9.22]) before each
concludes on a distinctive note: doggish lust in Pāṇḍu’s case, perhaps his way of recalling the mating deer he just
shot; and massive guilt in Yudhiṣṭhira’s.
130. Compare the situation with Yudhiṣṭhira mentioned in the previous note. He starts out envisioning
the vānaprastha mode and then, perhaps aware that his wife and brothers have had their fill of the forest, raises
the stakes to a renunciatory mode.
 women’s dharma 395

The three now start off into the Himalayas, crossing over certain
mountains—Himavat, Gandhamādana (220.42–43)—that their sons and
Draupadī’s travels will make more familiar.131 But their further journey, on
which they begin to be “protected by Great Beings (mahābhūtas, possibly the
five elements), Perfected Ones (Siddhas), and the Supreme Ṛṣis (Paramarṣis),”
takes them to Mount Śataśṛṅga, the Hundred-Peak Mountain (110.44–45), a
destination that is otherwise, it seems, virtually uncharted.132 But I believe it is
invoked by Damayantī when she is estranged from her husband Nala and
addresses a mountain somewhere in central India:

O best of mountains, have you with your hundred peaks (śṛṅgaśatair)

that scrape the sky perchance seen king Nala in this terrible forest?
(3.61.50)

What is most pertinent about this turning point in the famous love story of
Nala and Damayantī, a “mirror story” to the Pāṇḍavas and Draupadī’s own
forest troubles,133 is that just after Damayantī appeals to this mountain, she
turns north and walks for three days and nights (56) only to come upon a mirac-
ulous “circle of hermitages” (93d) “looking like a heavenly park” (57) populated
by Ṛṣi-Muni ascetics (tāpasaḥ) “the likes of Vasiṣṭha, Bhṛgu, and Atri,” who
“lived on water or off the wind, or fed on leaves” (58–59). Just as the Ṛṣis are
now protecting Pāṇḍu, Kuntī, and Madrī—but especially, as we shall see, the
two women—these Ṛṣis (really, “the same” Ṛṣis) have apparently heeded
Damayantī’s call to the hundred-peaked mountain and have arranged for her to
find them so that they can reassure her that she will find Nala, “the best of
dharma’s upholders,” restored to her and to his kingdom (88)—only to vanish,
along with their hermitages, leaving her amazed and wondering whether she
had seen them and their hermitages only in a dream (93).
But whereas Damayantī brings the Ṛṣis of the north to central India, Kuntī,
Mādrī, and Pāṇḍu scale the Ṛṣis’ Hundred-Peak Mountain themselves. For four
verses, the focus is entirely on Pāṇḍu as he “became dear to the sight of the
hosts of Siddhas and Cāraṇas” (111.1):

To some he became a brother, to some he became a friend (sakhā),

and other Ṛṣis watched over him like a son. (3)

131. On Gandhamādana, prominent in many Mbh episodes and in the Nārāyaṇīya, see Hiltebeitel forthcoming-a;
it will also be the retreat of the last Arhat in the Tibetan Candragarbha Sūtra (chapter 7 § B.3.b).
132. Its only other mention occurs as Arjuna and Kṛsn ̣ ạ pass over it—along with the Śaryāti Forest and the
holy places of the Horse’s Head and Ātharvaṇa (7.57.28)—on the way to asking Śiva for the Pāśupata weapon. The
Rāmāyaṇa seems to have relocated Śataśṛng ̇ a to the west, “where the Sindhu river meets the ocean” (4.41.12)!
133. See Biardeau 2002, 1: 412–13 (mentioning it in this category along with the Rāma- and Sāvitrī-
upākhyānas). On upākhyānas in the Mahābhārata, see chapter 9 § B; Hiltebeitel 2005a.
396 dharma

Let us note these “hosts” (saṅghas) of Ṛṣis, whose likes we shall meet again in
chapter 12. In their midst, Pāṇḍu was “intent upon going to heaven by his own
power (svargaṃ gantuṃ parākrāntaḥ svena vīryeṇa)” (2), and “after a long time
he reached such stainless austerity that he became like a Brahma-Ṛṣi
(brahmarṣisadṛśaḥ)” (4)—that is, he “looked like” or “resembled” one, yet did
not become one, even if he seems to have momentarily convinced himself oth-
erwise. For at this point, “wishing to cross to the other shore of heaven, head-
ing north from Śataśṛṅga,134 he set forth with his two wives” (5). His wishing to
make this crossing, even with Kuntī and Mādrī, is motivated before this last
verse by a fairly widely found eight-line mainly Southern interpolation.135
Hearing from the Great Ṛṣis that they are starting out on a new moon-night’s
journey to visit Brahmā for a great gathering of the great-souled Gods, Ṛṣis,
and Fathers, who are desirous of seeing the Self-Existent in Brahmaloka, Pāṇḍu
suddenly gets up wishing to accompany them! The passage is cosmologically
interesting for its new moon setting, its similarity to Buddhist instructions on
the path to the company of Brahmā (see chapter 4 § A), and its suggestion that
what motivates Pāṇḍu is to shortcircuit the system by joining the company of
the heavenly Fathers before he is one. But it is clearly an interpolation, occur-
ring throughout the Southern Recension, from which it seems to have entered
all the Northern Devanāgari manuscripts but one, yet none of the others in the
Northern Recension. It is another example of Southern ingenuity, and with or
without it, the ascetics now tell Pāṇḍu enough is enough.
Without making it quite explicit, the ascetics describe the journey beyond
Śataśṛṅga as reserved for immortals (as the interpolation spells out). And while
saying nothing about Pāṇḍu’s being childless, they state their concern for his
wives:

Going higher and higher facing northward up the king of mountains,

we have seen the peak’s many inaccessible regions. . . . There are
regions of perpetual snow where no tree grows, no deer or birds live;
there are some great continents (kecin mahāvarṣā), some inaccessible
passes. No bird could cross them, much less animals. Only the wind
has gone beyond, and the Siddhas, the Supreme Ṛṣis. Not deserving
the misery, how could these two princesses not sink on this king of
mountains? Don’t go, Bharata bull! (111.5e–6b, 8–10)

134. Damayantī’s three-day journey north from a hundred-peaked mountain before the Ṛṣis appear to her
may echo this route.
135. Mbh 1.1171*. See Biardeau 2002, 1: 227; Ganguli [1884–96] 1970, 1: 267. This is similar to the one at
the beginning of our skein where King Mahābhiṣa offends Brahmā by gazing at Gaṅgā. Note the interpolation
there as well.
 women’s dharma 397

The ascetics are of course politely telling Pāṇḍu, without mentioning his defi-
ciencies, that his wives are not ready to ascend bodily to heaven. But Pāṇḍu also
understands (again, as the interpolation has hinted) that they are saying he also
is not ready because he is childless (11). In a brief Southern interpolation, he
continues to defy the system and would have all three of them take heaven by
storm:

By terrible tapas, together with my wives, my life abandoned, even

without offspring, I will find heaven by terrible action (karma).
(1.1177*)

But in the Critical Edition—in lines that the Southern Recension preponed
when it transitioned from Gāndhārī’s gestation to an earlier moment in
Pāṇḍu’s plight (see n. 122 above), leaving Pāṇḍu here with just these bold
but hopeless words—it is a deflated Pāṇḍu who concedes his childlessness.
He now admits that he has fulfilled three of his four debts: those to Gods,
Ṛṣis, and Men, but not the one to his Fathers.136 Troubled that his Fathers
will perish when he does, he asks the ascetics how he might have “offspring
in my field,” meaning his wives (111.11–17). Surprisingly (at least, we would
assume, to him), the ascetics are encouraging. By their divine eye
(divyacakṣuṣā), they foresee him having “godlike, beautiful, flawless off-
spring” and urge him to apply his intellect and make the effort (18–20).
Minimally, they must know that Kuntī has her boon. But note how they set
Pāṇḍu up to think the solution could be at hand in some other way, perhaps
with one of them—as if with a wink and a nod he should recall how he him-
self was conceived by the Brahmin Ṛṣi Vyāsa.137 Indeed, before Pāṇḍu speaks
to Kuntī, the Southern Recension now has him bring both his wives near to
explain how

lower persons in distress (āpadi) desire a son from the higher, and
the straight (sādhavaḥ) desire offspring, the fruit of dharma, from
the best—(1.1181*, lines 3–4)138

136. Pāṇḍu improvises to combine the three debts and the five mahāyajñas (see chapter 5 § A), leaving out
the offering to Bhūtas and introducing the novel note that one performs the offering or debt to men by noncruelty
(ānṛśaṃsyena mānavān; 111.14d), which is what Kiṃdama said Pāṇḍu lacked when he shot a mating deer (109.18d),
as Pāṇḍu soon remembers (111.26). On this virtue, see further chapter 9.
137. Dhand 2004, 41 misses that the ascetics foresee the outcome and must know Kuntī’s boon, saying
that these verses “propose the timely solution of niyoga.” But yes, Pāṇḍu “seizes on” this implication “with
enthusiasm.”
138. The Southern Recension, both flat and arch here, cribs the gist of these lines from what Pāṇḍu soon
tells Kuntī at 111.30c–31b.
398 dharma

whereupon, as “he thought about a qualified (guṇavantam) Brahmin, the

dharma-knowing” Pāṇḍu “brought” Kuntī and Mādrī “right near a conclave of
great Brahmins” (lines 5–6)! The Critical Edition (including the Southern
Recension when it returns from this interpolation) allows Pāṇḍu to be a little
(if not much) more subtle as he has his conversation with “his famous
dharmapatnī Kuntī in private (vijane)” or off to the side (111.22ab).
So now we come to the point of having tracked the term niyoga and its
verbal variations through so many vicissitudes. If other women have been
appointed or involved in appointing themselves or other women to this or that
contribution to the continuity of the Kuru line, none since Satyavatī got to voice
an opinion on the practice of niyoga itself, not to mention its variations, over-
tones, and ramifications. That is saved for Kuntī, now a woman who knows to
speak her mind on dharma, but one who also knew to do so when she was just
a “free female” girl.139 Arti Dhand has done a good job bringing out the main
import in Pāṇḍu and Kuntī’s niyoga dialogue (2004, 40–43), but there are still
things yet to notice. As Dhand says, usefully summarizing, “Pāṇḍu goes to
some contrivance to convince Kuntī that what he is proposing is a meritorious
act,” first “telling her about the six types of heirs that are possible,” then quot-
ing Manu “that any child of hers would legally be considered a child of his,”
before he finally, “to fully legitimize his proposal,” offers her precedent by a
story (41).
It is in their dialogue through stories and counterstories that the dharma
issues come alive, and Pāṇḍu’s opener can justly be called a Lulu:

Listen, Kuntī, to this story (kathā) about Śāradaṇḍāyanī, that hero’s

wife who was appointed (niyuktā) by the elders to bear a child. Pious
and bathed, at night, Kuntī, at a crossroads, having chosen an
accomplished Brahmin with a flower, having offered into Fire in the
rite to conceive a son, she lived with him when that rite was finished.
She gave birth to three warlike sons, Durjaya and so on, and so you
too, beautiful wife, must by my appointment (niyogāt) quickly rise to
conceive a son from a Brahmin of superior austerities. (111.33–36)

In the name of niyoga, Pāṇḍu is pressing Kuntī to do something similar to

what the young unmarried daughter would do who is driven to perform the
“husband-finding spell” of the “Three Ambikā Homa”: she should go stand

139. In the fullest account of her impregnation by Sūrya, Kuntī asks him, “But how can I make a gift of
myself that is surely not to be made?’ . . . But if you think that this is dharma, best of burners, I will do your wish
without being given away by relatives. Having made you the gift of myself (ātmapradānam), I shall remain vir-
tuous (satī)” (3.291.5cd, 10–11ab).
 women’s dharma 399

at night at a crossroads seeking a way to get pregnant, and without parental

or, it seems, spousal supervision! I do not press the matter of the crossroads
being dangerous140 only because the Śataśṛṅga ascetics all seem to be of
the holiest sort. Kuntī has her chance to say something important, and
what she starts off with is, in my opinion, one of the highlights of the
Mahābhārata:

On no account, dharma-knower, can you speak to me like this (na

mām arhasi dharmajña vaktum evaṃ kathaṃcana), your devoted
dharma-wife, O lotus-eyed one. (112.2)

As she makes up her mind whether and when to tell Pāṇḍu about her secret
mantra, she is ready to stand her ground in a discussion that will turn on
dharma, her affection for her husband, and her unreceptiveness to niyoga.
She is, to begin with, just as encouraging as the Hundred Peak Mountain
ascetics:

You yourself, strong-armed Bhārata hero, will give birth to heroic

offspring in accord with dharma. I shall go to heaven with you, tiger
among men. And for offspring, you only come to me, joy of the
Kurus! Surely I will not go even in my mind to any man but you.
(112.3–5b)

He has come to her, but not on terms she can yet accept. We see that she holds
back yet promises all. Kuntī, as Dhand says, is “quite as deft in debate as Pāṇḍu
himself,” and she now cites a counterstory to match his (2004, 41). She calls
her story (kathā) “purāṇic” (112.6ab, cf. 13b), and it is more ludicrous than his.
As Dhand states briefly, “Bhadrā Kākṣavatī, . . . deeply aggrieved at the death of
her husband, through fantastic resolve succeeds in having her husband return
periodically to life to sire seven worthy sons on her” (41). Minimally, Kuntī tells
a story of successful necrophilia. But in the last verse 34, she does not, as Dhand
states, propose a “similar code of conduct for herself”; rather, if she sets any-
one’s course by this story, it is Pāṇḍu’s, who, she explains, need not periodically
rise from the dead:

And so you too, Bharata bull, are able with just your mind (eva manasā)
to beget sons on me by the lasting yoga-power of your tapas. (112.34)

She is really saying that Pāṇḍu will be able to contribute mentally to the use of
her mantra, but that still lies beyond his grasp. But he does grasp, even as he is

140. See 5.37.26, where, Vidura tells Dhṛtarāṣṭra: “A sensible man will never feel free to enter a stranger’s
house at the wrong hour, nor stand at night concealed at a crossroads, nor solicit a woman of baronial rank.”
400 dharma

about to tell the next counterstory, that her lesson has been about the resusci-
tated husband and would somehow have to be for him:

Yes, so did (cakāra) Vyuṣitāśva of old, beautiful Kuntī, just as you have
told. He was surely like an immortal (sa hyāsīd amaropamaḥ). (113.2)

I believe that necrophilia is a key to Kuntī’s story, but that it is not only about a
king who impregnates his queen from his corpse (112.29–33). Kuntī, I will now
contend, answers Pāṇḍu with reminders of the circumstances of his own con-
ception, and of dimensions of it that had less to do with niyoga, which she is
refusing, than with its having been a cryptic and, at least to the “two mahiṣīs,”
repulsive Aśvamedha. If so, all this would help to explain how Kuntī’s answer
to Pāṇḍu is a rejection of his explicit mention of niyoga without her ever
mentioning it herself. Here are the clues to such a reading.
Like Vicitravīrya with his two wives, King Vyuṣitāśva died of the extremely
rare (in the epic) disease of “consumption” (yakṣmāṇam)141 after lusting madly
(kāmasammataḥ) for Bhadrā (112.16). Also like Vicitravīrya, Vyuṣitāśva dies
sonless. But unlike Vicitravīrya, this “most Law-minded” and “dharma-spirited”
scion of the Pūru line (this would make him one of Pāṇḍu’s ancestors) made
his fame as a yajamāna or sacrificer (8) in numerous Soma sacrifices (9, 14).
And the big event of his reign, by which he subdued the kings in each direction
(11), was a horse sacrifice, which Kuntī sonorously links with his name and
with his reputation for extraordinary strength:

At his Aśvamedha great-sacrifice, Vyuṣitāśva the majestic (aśvamedhe

mahāyajñe vyuṣitāśvaḥ pratāpavān) surely became Indra among kings,
endowed with the strength of ten elephants. (12)

The rite even occasions a song telling that he protected all the social classes
like a father his sons (13)—perhaps a reminder that he probably had none.
But we only know for sure that he was sonless with Bhadrā, who may or may
not have been—but probably was—his mahiṣī. When he dies Bhadrā is not
only his grieving wife (bhāryā; 112.17d); lamenting his demise and wanting to
join him in the other world, she says, “Faithful as a shadow, king, I shall ever
do your will, always loving to please you (nityam priyahite ratā)” (23). A mahiṣī
is formulaically “dear” (priya) to her royal husband.142 When his corpse
141. The epics use it only for Vyuṣitāśva and Vicitravīrya (1.96.57d; 5.145.23f)—with the same verb
samapadyata.
142. Draupadı̄̄ is the “dear” mahiṣī of Yudhiṣṭhira (Mbh 4.15.31, 16.12; 10.11.17) and of all five Pāṇḍavas
(4.20.19, 5.80.22); so is Sı̄tā to Rāma (Rām 4.48.18, 5.12.43–44, 13.46; Mbh 4.20.10 according to Draupadī);
Śakuntalā finally to Duṣyanta (1.69.43); Indrāṇī to Indra (5.11.13; 12.22, while coveted by Nahuṣa); and Tārā was
dear to Vālin according to their son Aṅgada (Rām 4.54.3). Both Draupadī (Mbh 4.19.10; 20.10, 19; 5.80.9) and
Sı̄tā (Rām 6.38.3) speak of the title with pride, and Vālmīkī uses it even when he welcomes the banished Sītā to
his ashram (7.48.8).
 women’s dharma 401

impregnates her she is also a pativratā (32b). In any case, while there are no
details on the rite, his name with -aśva means “the Daybreak Horse,” and pos-
sibly also “the One Who is Possessed by the Horse.”143 The first meaning
clearly resonates with his story when Kuntī identifies him with a rising sun
before his Aśvamedha (10) and with a setting one after it (17b), dying of lust-
driven consumption from his lovemaking with Bhadrā, during which, if she
was his mahiṣī, they would have been making up for a year’s lost time after
her cohabitation with the horse (recall that the king may utter the most erotic
mantras to the mahiṣī while she is lying with the horse). The meaning
“Daybreak Horse” would suffice to carry along an Aśvamedha innuendo,
which calls only for something beside his “inner voice” (vāk . . . antarhitā) to
arise from his corpse. But “possession” may also lie latent in the story since
something beside or above the strength of ten elephants enables him to speak
from his corpse and impregnate Bhadrā, who would already have had some
familiarity with his Aśvamedha and its horse if she was his mahiṣī. A posses-
sion scenario may also be at play from her side, for before she “embraced the
corpse” (taṃ śavaṃ sampariṣvajya) and heard its inner voice’s directions on
her timing and a bath (29–31), she uttered a long lament in which she vowed
to lie from that day forward on a bed of kuśa grass “possessed (āviṣṭā) by sor-
row and intent on seeing you (tvaddarśanaparāyaṇā)” (27). It thus seems that
just as Pāṇḍu’s story outrageously enjoins Kuntī to perform niyoga by recall-
ing something similar to the “husband-finding spell” of the “Three Ambikā
Homa,” Kuntī’s story replies in kind, and equally outrageously, by getting
Pāṇḍu to consider some of the deeper elements of what niyoga—which Kuntī
will have no part of—meant for his own mother Ambālikā and her older sister
Ambikā. Kuntī thus answers not in the name of niyoga itself but in the name
of its Aśvamedha reverberations, which, it now appears, Pāṇḍu’s conception
has only doubled within his own dynastic line. Yet the matter is curious.
Vyuṣitāśva is omitted from the Pūru–Bhārata–Kuru line’s genealogies (see
Brodbeck 2009a, 24–27). Could Kuntī be a better chronicler than
Vaiśaṃpāyana? Could she be making this up as genealogy by invention?
Brodbeck mentions her story (2009a, 176), but is silent on this question.
Kuntī has in any case mastered the art of deflection. But Pāṇḍu, while
admitting that Vyuṣitāśva “was surely like an immortal,”144 still wants to return
to the subject of niyoga, and takes a deep route back to it through the sources of
Law. What he says Kuntī should now listen to is not precisely an old story but

143. See MW 1040–41. As a neuter noun, vyuṣita, “daybreak,” derives from vi-2.÷vas, “to shine forth” (from
vi-uṣ, “to dawn”). But as an adjective from vi-5.÷vas, “to abide, dwell, live,” it can mean “inhabited by” in
compounds.
144. 1.113.2 as cited. I see nothing to suggest that Pāṇḍu “dismisses Kuntī’s narrative as naïve” (Dhand
2004, 42).
402 dharma

what he calls, in impressive Vedicizing tones, an “ancient Law (dharmam . . .

purāṇam) seen (dṛṣṭam) by the dharma-knowing great-spirited Ṛṣis” (113.3; cf.
6ef) that was repealed by none other than that Upaniṣadic enfant terrible and
“spoiled brat” Śvetaketu.145 In former times women went around in the open
(anavṛtāḥ), were free (svatantrāḥ), and took pleasure as it pleased them, but all
that ended after young Śvetaketu, outraged at seeing a Brahmin making off
with his mother, heard his father, the great Ṛṣi Uddālaka, defend the law: “This
is eternal dharma” (eṣa dharmaḥ sanātanaḥ; 13d)! Śvetaketu did not concur, and
laid down a new rule or limit (maryādā) that, once past the main point and its
obverse, comes to a rider that is suspiciously convenient to Pāṇḍu’s current
cause. Śvetaketu ruled that women’s faithlessness to a husband will from now
on be a grievous sin of aborticide (or brahmanicide), as will seducing someone
else’s chaste wife, and that “a wife who is appointed (niyuktā) by her husband
to conceive a child and refuses shall incur the same evil!” (Dhand calls this last
point “a somewhat sinister twist” [2004, 41].) Although Pāṇḍu is a Kṣatriya, he
is quoting a Brahmin—Śvetaketu—for whom bhrūṇahatyā (1.113.17), “killing an
embryo,” can mean Brahmanicide as well.146 Brahmanical dharma texts are
written from a Brahmin’s viewpoint, and any interference to a Brahmin’s seed
is a murder of Brahmins. Pāṇḍu thinks Kuntī should understand her rejection
of niyoga as a Brahmanicide because she won’t perform it with a Brahmin! The
account is interesting for its way of imputing change to dharma while abro-
gating even the “eternal dharma.” But Pāṇḍu is clearly less interested in
legislative history than in suggesting that Kuntī is acting like she thinks the old
law still applies to her. And in this, even if Pāṇḍu may sound like he wants to
raise the stakes by “speaking Law” rather than just a story, and however conve-
nient, and I think desperate, his argument now sounds, he has deepened the
dialogue with an ironic glimmer of insight into his wife, who, ever since she
was a “free” girl, has kept the secret of a mantra given her by a Ṛṣi “whose
design in dharma was [as she knows] hidden”—a secret that the Ṛṣis of Mount
Śataśṛṅga are now protecting along with her.
Indeed, imagine what Kuntī heard when Pāṇḍu began this narrative about
ancient and eternal Law: “From youth on they were faithless to their husbands,
but yet not lawless, for such was the Law in the olden days” (113.5). Kuntī will
still keep that secret: that she was faithless to Pāṇḍu, her future husband, when
she became a child-mother.

145. See chapter 4, n. 51; Olivelle 2005b, 13–51. Olivelle does not discuss this Mbh story (which makes his
“spoiled brat” additionally into a mommy’s boy) in that piece, but does in connection with Manu’s injunction (M
9.5–12) that women should be guarded (2005b, 257 and n. 32)—though mistaking the free partnering of
Śvetaketu’s mother for the rape of his wife. See now Black 2011, 137, 146–47.
146. See Olivelle 1999, 363, n. to Ā 1.19.15; Fitzgerald 2004a, 695, n. 50 to Mbh 12.15.55.
 women’s dharma 403

The point about this secret that is now emerging, and the point of giving
Pāṇḍu and Kuntī such a mutually probing exchange, is that she is still free to do
with her mantra what she will, and to that extent at least, the ancient Ṛṣis’
eternal dharma that women are svatantra lives on in her.147 Once the kanyā who
could “desire all,” she is now the woman who will tell her sons—speaking
about another kanyā, Draupadī—to “share it all equally” and have it all turn out
dharma, as it will now if Pāṇḍu will only calm down.
Pānḍ u
̣ is not finished, but with this legal tale he seems to have run out of good
ideas. He cites two more brief precedents including his own (113.21–23); and he
makes a few more strident statements about dharma to continue pressing “the
entire patriarchal establishment of ancient India into the service of his argument,
alternately cajoling and bullying Kuntī” (Dhand 2004, 41). Finally, when he says,

Dharma-knowers also know, princess, that whatever a husband tells

his wife, whether dharma or even adharma, that is to be done—(27)

he might seem to have pushed matters beyond the limit. But that is also his
segue to saying a wife should “especially” do her husband’s bidding if he is
longing for sons and without the power to engender them,

as I am, flawless Kuntī, longing to set eyes on a son. Thus with folded
hands, cupped like lotus petals with red fingers, this añjali is raised,
lovely one, to my head for the sake of your grace (prasādārtham). (113.
28cd–29)

With his hands still, we may assume, so beseeching, he asks one last time
that she go “by my appointment” (man-niyogāt) to a highly qualified Brahmin
so that “on your doing, broad-hipped lady, I may go the way of those who have
sons” (30). As Dhand says, it is this “truly extraordinary gesture, the parallel
of which is hardly to be found in any of the Indian epics,” and Pāṇḍu’s “finally
actually pleading” (2004, 42)—and, we might add, his finally acknowledging
that it is totally up to her—that convinces Kuntī to reveal the secret of her
mantra.
For the rest, we can limit ourselves, while keeping in mind the women’s
rivalries and the question of succession, to three points: the mantra; some dif-
ferent features in the conceptions and births of Yudhiṣṭhira and Arjuna; and
the initiative of the Ṛṣis in settling the Pāṇḍavas and Kuntī at Hāstinapura.

147. To be sure, she is a special case: “Kuntī was fortunate: unlike Ambikā and Ambālikā, she had a mag-
ical formula. . . . Ordinary women did not possess special mantras to call on beautiful, well-scented, well-behaved
gods” (Dhand 2004, 42–43).
404 dharma

One first hears about the mantra in the Mahābhārata’s prologue, where
one verse describes the conceptions of all five brothers:

Their two mothers were impregnated, in accordance with a secret

Law (dharma-upaniṣadam), from Dharma, the Wind God, Śakra, and
the twin Gods the Aśvins. (1.1.69)

“Secret law”—van Buitenen’s translation—seems to call upon the esoteric

character of the Upaniṣads. Upaniṣad can mean a secret teaching or spell, but it
is in any case something “Vedic” and would have the authority of the Vedic
Ṛsị s—like Pānḍ u
̣ ’s “ancient law” endorsed by the Upaniṣadic sage Uddālaka—
even if Kuntī gets her mantra only from Durvāsas and not his alter ego Vedavyāsa.
A dharma-upaniṣad, however, is something new,148 for which “secret law” or per-
haps “legal spell” may have to do while keeping such Vedic overtones in mind.149
As we have seen, Durvāsas gave this mantra to Kuntī as “one whose design in
dharma is hidden (nigūdh ̣ aniścayaṃ dharme)” (1.104.5a), a phrase (or verbal for-
mula) that Kuntī knows well enough to repeat it exactly now to Pānd ̣ u
̣ (113.33a).
She also repeats that the mantra is “accompanied by sorcery” (abhicārasaṃyuktam;
34a), and further calls it a mantragrāmam (34c) or “canon of spells.” When that
term is used in the longer version of Karṇa’s birth, it is said to be something
“heard in the Atharvaśiras” (atharvaśirasi śrutam; 3.289.20).150 Its hidden dharma
is thus not only Upaniṣadic but Atharvanic, and sorcerous in that connection.
Yet we could not find much of sorcery in Kuntī’s first tryout of the mantra
with Suryā. Abhicāra comes explicitly into question only in the conception of
Yudhiṣṭhira, and at Pānd ̣ ụ ’s insistence. Making his mind up immediately, Pānḍ u
̣
tells Kuntī, “Right now call Dharma” (39), and after explaining his choice, con-
cludes, “With service and sorcery (upacāra-abhicārābhyām) propitiate Dharma”
(42cd). Why the link between urgency and sorcery? For all his talk about Fathers,
Pānd ̣ u
̣ also wants a son quickly who will be in line for the throne:

Call Dharma, lovely woman, for he among the gods partakes of merit
(puṇyabhāk). For Dharma could not join yoke with us if it were
adharma and people will now think that this is the Law (dharmo
ayam). And of a certainty this son shall become the standard of Law
(dhārmikaḥ) for the Kurus. (39c–41b)

148. It occurs only here. See Scheuer 1982, 74–75 with discussion and bibliography, notably Katre 1943, 122,
who, drawing on the commentaries of Nīlakaṇṭha and Devabodha, says that “Dharma stands for Āpaddharma and
upaniṣad for mantragrāma,” and takes mantra-upaniṣad as a parallel, which he translates as “secret mantra.”
149. To conceive Yudhiṣṭhira from Dharma, Kuntī “muttered what was to be muttered according to rule
(jajāpa japyaṃ vidhivat)” (114.2cd), thus intoning it as a softly muttered Vedic prayer (japa).
150. For other usages, see 1.53.4; 3.290.1 and 2 (also in the account of Karṇa’s birth); cf. Rām 1.21.10c;
1.26.21d.
 women’s dharma 405

If succession will hinge on this son’s legitimacy, Dharma is the best he can do.
As Kuntī now invokes Dharma, Vaiśaṃpāyana interjects that Gāndhārī has
already been pregnant for a year (114.1); and soon enough (that is, before the
birth of Duryodhana), Dharma’s son is born to a brief announcement from a
disembodied voice that this “firstborn son of Pāṇḍu called Yudhiṣṭhira” will be
“the best of dharma’s upholders” (5–6). Arjuna’s conception and birth are quite
different. Following another update that Bhīma was born on the same day as
Duryodhana (114.14cd)—which would suggest that Gāndhārī’s pregnancy
lasted a long two years—the race is over and the couple can take their time for
a son from Indra who will be Pāṇḍu’s “choicest” (18). “Having consulted with
the Great Ṛṣis,” Pāṇḍu “directed Kuntī to do an auspicious yearlong vow”
(19cd)—exactly what Ambikā and Ambālikā did not have time for (cf. Brodbeck
2009a, 183)—and undertakes his own arduous tapas, until, after a long time,
Indra announces his compliance (20–22) and foretells the grandeurs of his
son-to-be, as does the disembodied voice, now that of the wind in space (vāyur
ākāśe), once Arjuna is born (23–36). This divine wind is heard not only by
Pāṇḍu and Kuntī but by the Hundred Peak Mountain ascetics, filling them
with joy, and setting off such a vast clamorous celebration of the Gods, Divine
Ṛṣis—a triṣṭubh verse is set off for the Seven Maharṣis of the Big Dipper (41)—
and all manner of other celestials and even infernals (the snake sons of Kadrū
appear along with the bird sons of Vinatā; 40a, 60–62) that those “best of
Munis” (63b), who by now seem to include both the Hundred Peak ascetics
and the celestial Ṛṣis, are astounded and all the more exultant over the Pāṇḍavas
(37–63).
Pāṇḍu, however, is greedy for even more sons, and when Kuntī demurs,
Vaiśaṃpāyana gives one last update that Gāndhārī has also now had her
numerous sons (1.115.1b). Mādrī now coaxes Pāṇḍu to see if Kuntī will give her
a chance with the mantra (1.115.1–18), thus activating the latent rivalry between
these two that plays out from this point on.151 Kuntī’s demurral comes with an
interesting dharma adage, and she is surprised that Pāṇḍu is not respecting it:

They do not speak of a fourth son even in a time of distress. After

three she would be a loose woman,152 after four a whore. Knowing
this Law, which stands to reason (buddhigaṃyam), how do you
transgress it and, as though besottedly, speak about offspring!
(114.65c–66)

151. See 1.115.23, where Kuntī refuses Pāṇḍu’s request to let Mādrī have another use of her mantra: “I said
to her,’ For this once,’ and she got two! I was deceived. I fear that she will best me. That is the way with
women!”
152. Paraṃcāriṇī, one who would now have a reputation for “moving with others.”
406 dharma

Kuntī cites three sons as a dharma limit that she has already stretched. Manu
makes one son the limit for niyoga, but allows that some say two (M 9.60–61).
But we know that Kuntī now has four sons, counting her secret son Karṇa by
Sūrya, and moreover that she has now had five “men,” counting Pāṇḍu. This
will be the number of Draupadī’s husbands when Karṇa uses the same term
for “whore” (bandhakī) to revile her at the dice match for “submitting to many
men” (Mbh 2.61.35). Again, Kuntī knows the Law, threads her way carefully,
and is as free with her mantra to stop as she was to start.

H. Settling Mother Kuntī and Her Sons Back at Hāstinapura

Now that all the children are in place, the Pāṇḍava boys and their parents have
a little time with the Śataśṛṅga ascetics. If Pāṇḍu and his wives hiked to
Śataśṛṅga, they might have considered making the return trip, even if they
would have been slowed down by the children. They did not. Maybe Pāṇḍu was
embarrassed about his curse, or knew his story would be hard to explain. The
journey to the plains is shrouded in mystery, and some have seized upon this
to infer that the Pāṇḍavas are “inventions,” while others have made them out
to be impostors, possibly Himalayan tribals from a polyandrous culture, and in
any case not really the sons of Pāṇḍu, which is of course perfectly true.
Compounding the mystery is the conundrum that Mādrī and Pāṇḍu seem to be
cremated twice: first in the mountains and then outside Hāstinapura. I would
just say that I do not think there is an ethnographic key to the Pāṇḍavas’ origins
or their polyandry, but it would be nice to know more about royal funerals of
the time or times of the epic’s composition. The point we must satisfy our-
selves with is that the Mahābhārata refers all these problems to the Śataśṛṅga
ascetics.
These characters now reenter the narrative immediately after the births of
Mādrī’s twins:

Then those who dwelt on Śataśṛṅga gave [the five] their names, with
devotion (bhakti) and ritual and benedictions, O king. . . . And as they
grew up there on the holy Himalayan mountain, they brought
wonder to the Great Ṛṣis who had foregathered there. (1.115.19, 27)

From here on, although both are repeatedly mentioned, the resident mountain-
dwelling ascetics and the celestial Great Ṛṣis have merged and cannot be kept
distinct. Love now springs its last time in Pāṇḍu’s heart (116.4) as he succumbs
to the “law of copulation” (maithunadharma; 9c) according to the “law of time”
 women’s dharma 407

(kāladharma; 12c), and Mādrī is the last to see desire on his face. As we saw in
chapter 4, Pāli texts associate the powerful “law of sex” with village law and
building houses to conceal it. The Mahābhārata now associates it with “law of
time!”153 Kuntī and Mādrī, rivals nearly to the end, grieve and scream before
working out which of them will lie with Pāṇḍu on the pyre and which one will
take care of the children (13–31). Their last eloquent, noble, and sad exchange
gives Mādrī the last word:

As he was lying with me the best of Bhāratas was cheated of his

love. So how should I deprive him of his love in Yama’s seat? Nor
will I go on living treating your children the same as mine, noble
lady, for evil would touch me that way! Therefore you, Kuntī, must
treat my twins as your own sons. The king went to his death
making love to me—let this carcass of mine be burnt with the body
of the king that covers mine so well. Do this as a favor, noble lady!
Watch over our children and think kindly of me—there is nothing
else I could charge you with. (116.26–30b)

They do not mention (and must not know) that Mādrī’s lot of joining Pāṇḍu in
death was foreseen in Pāṇḍu’s curse. It can be noticed that whereas the
Kaurava–Pāṇḍava men’s rivalries can be resolved only in heaven, their wives’
differences can be resolved on earth.
And now the text is where it wants to be, with the Great Ṛṣis and the
ascetics of Śataśṛṅga helpmates together in securing the “work of the gods” and
the “welfare of the world”:

After they performed Pāṇḍu’s final bath, the god-like Great Ṛṣis then
took counsel, those ascetics having assembled. (117.1)

Although it is hard to tell with the honorifics, it seems like the Great Ṛṣis are
addressing the ascetics:

King Pāṇḍu has left his barely born children and wives154 in trust
with your worships (bhavatām) here as he went to heaven. (117.3)

In any case,

Having thus taken counsel with one-another, those delighters in the

welfare of all beings (sarvabhūtahiteratāḥ) put Pāṇḍu’s children

153. See chapters 4 § C.1, n. 151; C.3.a and 6 § B on Mbh 12.200.35–37, where maithunadharma begins in
the Dvapāra yuga.
154. Dārān, masculine plural, means both wives here and below, although only Kuntī is now alive.
408 dharma

before them and set off to the Elephant City. And the lofty minded
Siddhas set their minds on journeying there to give the Pāṇḍavas to
Bhīṣma and Dhṛtaraṣṭra. At that very instant, having taken them, all
the ascetics set out with Pāṇḍu’s wives, sons, and body. (117.4–6)

The arrival of thousands of Cāraṇas and Munis astonishes the people of the
capital, and as the sun rises throngs of men and women from all four social
classes come to see the ascetics, showing no trace of jealousy and becoming
minded of dharma (dharmabuddhayaḥ; 12d), while members of the royal
household, including Bhīṣma, Dhṛtarāṣṭra, Vidura, Satyavatī, Kausalyā (pre-
sumably Ambālikā), Gāndhārī, and Duryodhana and all his brothers come out
to greet “all those hosts of Great Ṛṣis (maharṣigaṇān) with bowed heads.”
Bhīṣma then, once “the mass of people all around had fallen silent, offered
kingship and the kingdom to the Great Ṛṣis” (8–18)—a pro forma gesture that
Mahābhārata kings often make to such visitors that requires no reply but
acknowledges their higher authority (i.e., “real” rule).

Then the oldest of them, a Great Ṛṣi wearing the braid and deerskin,
arose and knowing the Great Ṛṣis’ mind (maharṣimatam) spoke.
(117.19)

Accountable for the collective “mind of the Maharṣis,” this unnamed senior
Maharṣi155 now makes a fairly lengthy (117.20–31) speech, telling how Pāṇḍu
arrived at Śataśrṅga “having renounced love and pleasure,” how the Pāṇḍavas
were born despite that, how Pāṇḍu enjoyed his sons’ childhood and never
strayed from the path of the good until he went to the world of the Fathers sev-
enteen days ago, and how Mādrī joined him on the pyre faithful to the world of
her husband. This senior Maharṣi speaks on three practical matters. First,
saying “these are their two bodies,” the Kauravas should see to the “remaining
rites” for Pāṇḍu and Mādrī (29–30). Second, while not quite saying it, he
strongly hints that the preferred heir should come through Pāṇḍu:

This ancestral lineage was again uplifted by Pāṇḍu while the famous
one dwelt in the forest always abiding by dharma. . . . And when the
rite for the departed has been done, let famous Pāṇḍu, who knew
every dharma, the bearer of the dynasty of the Kurus, receive the
ancestral offering. (25, 31)

155. One is reminded of Śaunaka, “the Mahābhārata’s anchorman” (see chapter 6 n. 29). Also, if the
compound maharṣimatam is read as “the great Ṛṣi’s mind,” one could take it as echoing passages that call the
Mahābhārata Vyāsa’s “thought entire (kṛtsnaṃ matam)” (1.1.23; 55.2; Hiltebeitel 2001a, 12; cf. 54), in which case
this Ṛṣi who might remind us of Śaunaka could be cryptically tuned into the thought of the author.
 women’s dharma 409

Finally and accordingly, when it comes to describing the births of the

Pāṇḍavas, it is Pāṇḍu’s firstborn who really counts:

While he [Pāṇḍu] himself lived keeping a vow of celibacy, by divine

instrument (divyena hetunā) this son Yudhiṣṭhira was born to him,
begotten by Dharma himself. (21)

Enough said, and really, who would believe anyone else!156 I believe this is the
first inkling of the divine plan made public in the course of the story, and it is
the Ṛṣis’ role to make it so. And with that, as they did after imparting their mes-
sage to Damayantī,

even as the Kurus were looking the Cāraṇas and Secret Ones
(Guhyakas) all disappeared that instant. And upon seeing the host of
Ṛṣis and Siddhas (ṛṣisiddhagaṇa) vanish there like a castle in the
sky,157 they attained the highest wonder. (33–34)

It helps, however, to have Vidura—who incarnates the same god Dharma—

there to be the first to speak, ordering that the rite for the departed be properly
begun and confirming, while somewhat understating the matter, that Pāṇḍu’s
“five heroic sons were born like sons of Gods (surasutopamāḥ)” (118.4c).
When this ceremony has settled Pāṇḍu and Mādrī among the line’s hon-
ored forebears, Vyāsa then brings our skein to its end:

When the śrāddha had been completed, Vyāsa looked upon the
grieving people and said to his mother Satyavatī, who was blinded by
pains of sorrow, “The times of happiness are past and times of
trouble lie ahead. The days grow worse every new tomorrow. Earth’s
youth is gone. A dreadful time is at hand, confounded by much
trickery (māyā), beset by many vices, when all conduct and acts of
dharma shall be soiled. Go now, leave it all. Yoke yourself and live in
the wilderness of austerities, lest you must witness the ruination of
this your own family.” (118.5–8)

Satyavatī consents, tells Ambikā she has heard that her son’s “bad policy” will
bring about the destruction of the Bhāratas, and Ambikā consents to leave as
well. Bidding adieu to Bhīṣma, Satyavatī and Ambikā then set off with Ambālikā,

156. The short account of this scene in the epic’s prologue has the Ṛṣis speak together, before disappear-
ing, telling only that the boys are the sons of Pāṇḍu. But after this, “Some said, ‘They are not his.’ Others, ‘They
are his.’ Others again, ‘How can they be his when Pāṇḍu has been dead long since?’” But all agreed, “They must
be bid welcome!” (Mbh 1.1.72–74). See Brodbeck 2009a, 172, 175.
157. Van Buitenen 1973, 262 for gandharvanagara, “a city of the Gandharvas,” the heavenly musicians.
410 dharma

who is still distraught over the death of Pāṇḍu, for the forest. There, after the
fiercest austerities, their lives end (9–12). Kuntī and Gāndhārī are now the
line’s only living mothers.

I. Conclusions

Looking back over this chapter, we see that dharma is spun out from all variety of
Brahmanical sources: regional and family customs, anecdotes and proverbs,
Veda—especially in relaying it through the great Vedic Ṛsị s, and in some sense
what is pleasing to the self as regards both men and women. As we have sug-
gested, it may be that the Mahābhārata picks up on an implication of the
dharmasūtras, which, like our skein, use the enigmatic term dharmatantra, “loom
of dharma.” But unlike the dharmasūtras, this skein presents women who take
initiative in the threading and are indispensable to the texture. They are
back-clothed by female prototypes like the weavers of night and day and the snake
and bird mothers Kadrū and Vinatā. Great things could also be said individually
about other women’s interventions in briefer crises of the Paurava–Bhārata–
Kuru line (Devayānī and Śarmiṣṭhā, the two rivaling wives of Yayatī; Śakuntalā;
Tapatī158) or about princesses married into other lines (Damayantī, Sītā, Sāvitrī)
whom the Pānd ̣ ạ vas and Draupadī hear about in the forest. But nowhere else in
either epic is there a skein that gives women such repeated prominence not only
in the succession of their lives but in the ramifications of their lives for each
other’s lives, or does so with such recurring focus on a primordial Law of the
Mother. Each of these women plays her part in this textualization of dharma,
bringing home its nuances—whether in questioning it, interpreting it, raising
questions by her silences, or even by a slip of the tongue. Though one sees it most
artfully in Kuntī, each one, pace Manu, raises the question of her svatantra,
“whether she is a child, a young woman, or an old lady.”

158. On Śakuntalā, see Biardeau 1979; Hiltebeitel forthcoming-d. On Tapatī, note that Jamison 1996, 305
n. 98 remarks on how she “dutifully cites man’s views on women’s lack of independence” at Mbh 1.161.14.
9
Two Dharma Biographies?
Rāma and Yudhiṣṭhira

The question mark in the title of this chapter looks ahead to this
book’s last chapter. There, I will argue that the Buddhacarita, “The
Adventure of the Buddha,” offers the first known close and critical
reading of the Sanskrit epics. Its poet Aśvaghoṣa, of the first or
second century CE, was familiar with both the Rāmāyaṇa and the
Mahābhārata. When he portrays the Buddha-to-be as saying, “there is
no such thing as a wrong time for dharma” (BC 6.21), both Rāma and
Yudhiṣṭhira would fail to realize this because they are caught up in
the particulars of Brahmanical dharma. As we know, the Buddha
discovers “the true dharma.” Do characters in the Sanskrit epics learn
or make discoveries about dharma? With this question we extend our
inquiry into biography as a representation of living dharma over time.
We have already noticed this dimension in the Aśokan edicts and in
the Pāli suttas—among the latter, especially in the Majjhima Nikāya,
and above all in the Mahāparinibbāna Sutta, which is in the Dīgha
Nikāya.1 We may now consider whether the two classics build on
these precedents to imagine what dealing with dharma over time
means for a leading epic hero.
This matter of precedent is, of course, not to be taken without
anticipating disagreement. Some, for instance, have introduced the

1. See chapter 2 on Aśoka; chapter 4 n. 14 on Manné’s discussion (1990, 79–81) of the “intimate
biographical suttas” of the MN (n. 99). Manné seems to suggest that the Mahāparinibbāna Sutta may
have “found its way into” the DN as an exception because it arouses interest to “inspire conversion,”
and contains “accounts of converts and supporters from many different areas of society” (79).
412 dharma

possibility that Yudhiṣṭhira’s portrayal may owe something to Aśoka’s, and I

have suggested that Vālmīki’s portrayal of Rāma takes inspiration from the
Mahābhārata.2 Obviously Aśvaghoṣa’s portrayal of the Buddha builds from the
Sūtra Piṭaka.

A. The Royal Life as Adventure

A preliminary picture of Rāma and Yudhiṣṭhira’s dharma biographies can be

given in outline in relation to the similar ways their royal lives are structured
through each epic’s “archetypal” organization into “Books.” For the sake of
ongoing reference, I present a template shared by the Rāmāyaṇa’s seven Books
with the eighteen Books of the Mahābhārata in their Critical Editions.3

Book 1 introduces frame stories involving the composition and trans-

mission of the poem, dynastic origins connected with Vedic names, and
the youthful lives and marriages of the heroes. Rāma marries Sītā.
Yudhiṣṭhira and his four brothers, the Pānd ̣ ạ vas, marry Draupadī.
Each Book 2 then describes a pivotal “court intrigue” highlighting
the truthfulness of the king. In Rāma’s case, his truth is exalted in
that he upholds not only his own truth but his father’s, who had
promised Rāma’s stepmother to make their son Bharata, Rāma’s
junior, king instead of Rāma, and to send Rāma into exile. In
Yudhiṣṭhira’s case, his truth is questioned at the dice match with his
cousins, the Kauravas, and left ambiguous when he leaves Draupadī’s
question—whether, if he wagered himself first, he had a right to bet
her—unanswered.4
Book 3 is in each case about a forest exile. The kings and their
companions enter and exit the forests with monstrous encounters, and
while they are there, they receive instructive guidance from great sages
or Ṛsị s. After abduction attempts on their wives, in Sītā’s case success-
ful, the second monstrous encounters are transformative, ending their

2. It has been proposed that specific Mahābhārata characters are modeled after Aśoka: Yudhiṣṭhira
according to Fitzgerald 2001, 2004a, 103, 114–23, 129, and Sutton 1997; Arjuna according to Selvanayagan 1992.
For discussion see Hiltebeitel 2005b; Bowles 2007, 126–27; Das 2009, 243–46. Meanwhile, Hiltebeitel 1989
and Biardeau 2002 see the Mahābhārata as a riposte to Buddhism without such singular biographical modeling.
For arguments implying that Vālmīki’s portrayal of Rāma draws from the Mahābhārata, including the
Rāmopākyana, see Hiltebeitel 2009a., and chapters 5 §§ C and D of this book.
3. As my first attempt to outline such a template indicated (2005a, 460–61), Pollock 1986, 38–42 and
Biardeau 1997a, 77–119 anticipate such an approach through the first two Books.
4. On this episode, see Hiltebeitel, 2001a, 240–77; Das 2006, 99–133.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 413

forest trials so that they can return to society—in Rāma’s case to a

society of monkeys. Rāma encounters a headless monster who tells him
to befriend the exiled monkey king Sugrīva to learn about Sītā’s abduc-
tion. Yudhiṣṭhira meets up with his father Dharma disguised as a
seemingly murderous Yakṣa, whose second of three boons, once
Yudhiṣṭhira answers his questions, is that the five Pānd
̣ ạ vas and
Draupadī will be able to pass their final thirteenth year of exile incognito
in a place of their choosing. Note that the Buddha ends his forest life to
return to society after his encounter with Māra, the devilish god of death.
Book 4 is then about inversions. Rāma gets involved with the topsy
turvy world of the monkeys’ capital, in which the royal monkey brothers
Vālin and Sugrīva play out a reverse image of Rāma’s own story of exile,
wife-abduction, and fraternal rivalry for the throne. The Pānd ̣ ạ vas
assume topsy turvy disguises in the kingdom of Matsya, “Fish.”
Book 5 is then about “Efforts” made in Preparation for War: by
both sides in the Mahābhārata; by all the monkeys in the Rāmāyaṇa
(5.10.24; 33.36, using the Mahābhārata’s term for such efforts,
udyoga). In each epic a divine messenger goes into the enemy camp
where he reveals an overpowering nature while upstaging attempts to
hold him captive. In the Rāmāyaṇa this messenger is Rāma’s devoted
monkey Hanumān; in the Mahābhārata it is Kṛṣṇa.
Books 6 in the Rāmāyaṇa and 6–11 in the Mahābhārata are then
War Books.
And finally there are the denouements and returns to the frames
in Rāmāyaṇa 7 and Mahābhārata 12–18. Here the two kings’ dharma
biographies are handled very differently. Once their main exploits are
over, Rāma becomes the primary listener to his own adventure,
whereas Yudhiṣṭhira goes on learning more and more about being a
Dharma King, and is tested two last times by his father Dharma.

As the differences show, the Rāmāyaṇa works this common blueprint

along most carefully through Books 1–5. I favor the priority of the Mahābhārata
and present matters from that standpoint, with the corollaries that Rāmāyaṇa
Books 1 and 7 are integral to its earliest design and that the Rāmāyaṇa poet is
familiar with the Mahābhārata’s archetypal design and intent upon refining it.
In making such a comparison, one must also bear in mind that it is com-
plicated by Yudhiṣṭhira’s relation as the son of Dharma5 to Kṛṣṇa, the avatar or

5. Yudhiṣṭhira’s counterpart in the Vālmīki Rāmāyaṇa as a son of Dharma is the monkey Suṣeṇa (6.21.22;
33.14), father of Vālin’s wife Tārā, and a healer. The Rāmopkhyāna knows Suṣeṇa but not as the son of Dharma.
414 dharma

incarnation of the supreme deity Viṣṇu. In the Rāmāyaṇa, Rāma is likewise

Viṣṇu incarnate. That is one source of his perfection and truth, however unbe-
knownst it may be to him. Whereas Kṛṣṇa knows and indeed occasionally con-
firms his divinity, Rāma must consider himself human until he has killed
Rāvaṇa because Rāvaṇa has a boon that he can be killed only by a man.6 Indeed,
uncomfortable though it has always been to modern readers, one of the impli-
cations of Rāma’s not knowing that he is really divine until Brahmā tells him
so twice is that Vālmīkī gives Rāma a split personality. His dharma biography
as a human is ruptured at these two points, both of which occur after he has
killed Rāvaṇa and submitted Sītā to ordeals because he or others doubted that
she would have been chaste with Rāvaṇa. Even after Sītā leaves the world for-
ever rather than submit to the second ordeal, Rāma’s human life is depicted as
going on without ostensible reflection on his divinity. Yudhiṣṭhira, on the other
hand, must puzzle his way to truths in relation to a rather unusual father and
an avatar whose advice is not only subtle but shady and often withheld from
him, and who often, I dare say, simply cannot be believed.7
It would be possible to present both biographies in fuller narrative than
space allows.8 I will concentrate in this chapter on four points of comparison:
the ways Rāma and Yudhiṣṭhira are introduced in their epics’ frames; their
position in relation to substories; the monstrous encounters that bring their
forest adventures to a close; and the ways they handle killing foes in the two
scenes where the morality of their doing so comes most into question—Rāma’s
killing of the monkey king Vālin from ambush, and Yudhiṣṭhira’s lie to enable
the killing of the Brahmin Droṇa. A fuller treatment of Rāma and Yudhiṣṭhira’s
dharma biographies would want to take up at least five wider areas in which
they are delineated: their treatment of women, especially their wives; their filial
obligations to parents and to others within their families; their instruction by
the wise; their understanding and handling of royal obligations; and what they
learn about it all by the end. This chapter will touch on some of these matters
while opening others for further discussion in subsequent chapters.

6. See Pollock 1984, with support from Goldman 1995, 74, 80; Goldman and Sutherland Goldman 1996,
30; 2009, 1,445 (n. to 6.105.10). Pollock has demonstrated that Rāma’s seeming humanity is a structural piece of
the story threaded into the poem along with a boon obtained by Rāvaṇa from Brahmā: that of invulnerability to
death from all different classes of beings other than humans, whom Rāvaṇa omitted because he disdained them
(7.10.13–20). Rāma must thus be a man to slay Rāvaṇa and must think he is one until he accomplishes this goal.
In agreement, see Hiltebeitel 2003. Ignoring such recent discussion, González-Reimann 2006a draws an
increasingly tenuous distinction between textual and literary analysis (204) while making mostly literary points
himself in a misfired attempt to reclaim an undivinized Rāma for “the general tone of the narrative” (213), and to
paint Pollock as an unwitting Vaiṣṇava hermeneut.
7. For preliminaries on these Mahābhārata matters, see chapter 1 § C. For their further unfolding, see
chapter 12.
8. For Yudhiṣṭhira, see Klaes 1975, especially 24, 29–30, 32–34, 43, 49, 112; Bose 1986.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 415

B. Frames and Frontmatter on Rāma, Yudhiṣṭhira, and Dharma

In chapter 5 (§§ C and D), we noted some affinities between the use of frame
stories in Manu and the two epics, and remarked on one way that Manu and the
Rāmāyaṇa are closest to each other: their author-poets draw direct (uninterpo-
lated) inspiration from Brahmā, denoting unmediated “Vedic” authority. We
also noted a singularity of the Rāmāyaṇa frame. Whereas the Mahābhārata’s
and Manu’s frames are addressed only to remote, universal listeners who, in
the Mahābhārata’s case, are several generations, if not galaxies, removed from
Yudhiṣṭhira and other heroes whose lives are over, the Rāmāyaṇa’s first
universal listener is none other than Rāma himself who will hear the epic, his
own story, as the “perfect man.” In its first four-chapter (sarga) Upodghāta or
“Preamble,” the Rāmāyaṇa’s frame presents two progressive unfoldings of the
story—the first told by the Ṛṣi Nārada to the Ṛṣi and poet-to-be Vālmīki, the
second envisioned by Vālmīki himself, now a poet—that trace their way into
the third full unfolding of the Vālmīki Rāmāyaṇa itself.9
The first unfolding comes in Nārada’s response to Vālmīki’s text-opening
question—or overarching question—as to whether there is an ideal man in the
world today:

Is there a man in the world today who is truly virtuous (guṇavān) and
energetic and yet knows both dharma and how to act upon it
(dharmajñaś ca kṛtajñaś ca)? Who always speaks the truth and holds
firmly to his vows? Who exemplifies proper conduct and is benevolent
to all creatures? Who is learned, capable, and a pleasure to behold?
Who is self-controlled, having subdued his anger? Who is both
judicious and free from envy? Who, when his fury is aroused in battle,
is feared even by the gods? This is what I want to hear, for my desire to
know is very strong. Great Ṛsị , you must know of such a man. (1.1.2–5)

The first verses pose the question in terms of paired dharma traits combined
with their implementation.10 From there on, Vālmīki asks after particular

9. For discussions of the Upodghāta from different perspectives, see Bhandare 1920, 1–43 and critical
notes; Goldman 1984, 67–73, 273–88; Brockington 1998, 2–3, 380, 395. Some aspects of this treatment of the
Upodghāta were discussed earlier in Hiltebeitel 2005a.
10. Goldman trans. (1984, 121), excepting this book’s usual insertion of “dharma.” But let us note that
kṛtajña can quite ordinarily mean “grateful.” Vālmīki’s first doublet of a virtue and its implementation could thus
be translated “who knows dharma and is grateful” rather than “. . . and how to act upon it.” This would change the
implications of Vālmīki’s question in two ways: first, with respect to what it means to implement dharma, that is,
why gratitude would implement dharma just as holding firm to vows would implement truth; and second, partic-
ularly with respect to Sītā, who will hold this quality of Rāma up to question, as we shall see in chapter 10 § D.
416 dharma

virtues, on the way to addressing the matter of how this person would handle
anger. By the last verse, it is clearly a leading question, for Vālmīki certainly
suspects or even knows that Nārada has an answer. The Rāmāyaṇa frame thus
opens from this dharma question, and leaves the question open for a moment
before Nārada replies.
Nārada meets the requirements of the question with an entirely laudatory
brief account of Rāma’s virtues and adult life, ostensibly to date (1.1.7–76). He
responds as the question requires, first admitting that the “celebrated virtues”
asked for are hard to find in one person, but that upon consideration, he has
the answer: Rāma (7–8); and then reciting in detail the qualities Rāma pos-
sesses, including certain godlike ones (9–18) that might remind us that Manu’s
king is born from the particles of various gods (M 7.5–7; see chapter 5 § E).
Then he embarks upon Rāma’s story (19–79). This section provides the
so-called Saṃkṣipta (“condensed”) Rāmāyaṇa, which, because it does not
include anything from the first or last Books, has been taken to predate “even
the oldest stratum” of Book 1 while postdating “the central body of narrative
that is summarized in this section” (Goldman 2004, 68)—a rather acrobatic
proposition that would require all the iterations of Rāma’s story in the
Upodghāta to reflect different stages in the Rāmāyaṇa’s textual formation
(cf. Brockington 1998, 381). I propose, on the contrary, that Nārada’s account
is purposefully “condensed” for the further iterations to unfold from it in the
mind of the poet (and, of course, the first listener, and other listeners and
readers).
I will note only a few points of interest in Nārada’s condensation. Saying
the minimum about Rāma’s killing of Vālin (49, 55), he hardly hints at
anything problematic in Rāma’s life and omits Sītā’s fire ordeal while con-
cluding with Rāma and Sītā’s return to Ayodhyā to recover his kingdom (70).
In bringing Rāma to mid-career, Nārada speaks of his rule as already a kind
of golden age (“like the Kṛta Yuga,” 73) extending into the future in which
Rāma “is performing hundreds of horse sacrifices involving vast quantities
of gold,” “bringing about the establishment of royal lines of a hundred qual-
ities, and appointing the four social classes each to its own dharma in this
world (cāturvarṇyaṃ ca loke ‘smin sve sve dharme niyokṣyati)”—again
sounding like Manu (e.g., 7.35)—before he will go to heaven after ruling for
eleven thousand years (74–76).11 In mentioning the “hundreds of horse
sacrifices” that spangle Rāma’s ongoing future present, Nārada leaves

11. This comes right after the Saṃkṣipta section, and seems to refer to the statement made twice at the end
of the Yuddhakāṇḍa that Rāma ruled for ten thousand years (6.112.82, 90). But as the conclusion of one contin-
uous iteration, it also implies Rāma’s later years that are described in Book 7, where that figure is also mentioned
(7.94.12; cf. 92.16: he spent ten thousand years seeing to state affairs).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 417

unmentioned the one horse sacrifice narrated in Book 7, undoubtedly the first,
where the banished Sītā refuses to return to Rāma and enters the earth.12
The Rāmāyaṇa frame thus opens onto a king who is still alive, about whom
Vālmīki will be able to compose his Rāmacarita/Rāmāyaṇa13 to help readers, and
perhaps himself, decide whether Nārada has truly answered his question. The
contrasts with the Mahābhārata are striking. Yudhiṣṭhira is also mentioned from
time to time in the Mahābhārata’s frontmatter in ways that highlight associations
with dharma. First, a two-verse allegory compares “Duryodhana made of wrath
(manyumayo)” and “Yudhiṣṭhira made of dharma (dharmamayo)” as each a “great
tree” with others on each side as the tree’s crotch, branches, flowers, and roots
(1.1.65–66).14 And soon thereafter, Yudhiṣṭhira begins to be referred to frequently
as Dharmarāja, the Dharma King:15 first, as the story is initially digested in
Dhṛtarāsṭ̣ ra’s plaint as to all he should have done had he known better (1.1.108b,
113b, 148b; see E. Hudson 2007, 36–42), which is part of an adhyāya-long con-
spectus called the Anukramaṇī; and then again in the Parvasaṃgraha (2.62d,
175b, 196d, 217b, 228c), the epic’s table of contents (1.2.33–233). But the
Mahābhārata’s frontmatter and outer frames can hardly be said to open directly
onto Yudhiṣṭhira, who is not made a focus of narrative until the Pānd ̣ ạ vas’ great
grandson Janamejaya asks his next-to-last opening question to launch the main
narrator Vaiśaṃpāyana on his inner frame recitation of the great story of
Janamejaya’s deceased ancestors: “How could the best of dharma’s upholders,
the son of Dharma, who knew dharma, endure such oppression, of which he was
undeserving?” (1.56.9).16 Yudhiṣṭhira’s forte is asking questions17—something

12. The Mahābhārata’s Rāmopākhyāna also ends with the return to Ayodhyā (Mbh 3.275.66) before briefly
mentioning some events that follow Rāma’s enthronement, the last of which is curious: “Then, assisted by
gods and Ṛṣis (devarṣisahitaḥ), he offered, along the bank of the River Gomatī, ten unimpeded horse sacrifices”
(275.69). Devarṣisahitaḥ, which could be translated “together with the divine Ṛṣis” [one or many]), would leave
the Rāmāyaṇa poet a place to insert himself (Vālmīki is unmentioned in the Rāmopākhyāna) among the Ṛṣis who
might have come to one such Aśvamedha, but hardly a place for Sītā to have come there with him.
13. Nārada concludes with a phalaśruti that either counts his own description as a “Rāmacarita” and
“Rāmāyaṇa” already, or, more likely, calls for a text by those names that is yet to be composed (1.1.77–79).
14. Here Dhṛtarāṣṭra is the Kaurava tree’s root, and Kṛṣṇa, brahman, and Brahmins the Pāṇḍava tree’s
root. Cf. 7.157.22–24: “Kṛṣṇa is the root of the Pāṇḍus and Pārtha is like the risen trunk; the other Pārthas are
branches, and the Pāñcālas like leaves. The Pāṇḍavas have Kṛṣṇa as refuge, Kṛṣṇa as strength, Kṛṣṇa as lord—
Kṛṣṇa is their central support even as the moon among stars. Therefore, O Sūta’s son, avoiding the branches and
trunk, bring down Kṛṣṇa (kṛṣṇaṃ nikṛndhi) who is always everywhere the Pāṇḍus’ root.” Dhṛtarāṣṭra has asked
why a one-use sure-shot weapon was wasted on the Rākṣasa Ghaṭotkaca in the night fighting after the fourteenth
day of battle after Saṃjaya has told him that every night Duryodhana, Śakuni, Duḥśāsana, and Saṃjaya had told
Karṇa to kill Arjuna or else Kṛṣṇa with that special spear, See Hiltebeitel 2007, 35.
15. The Rāmāyaṇa also knows Yama as Dharmarāja (2.58.23d, 33c; 7.18.5, 23; 7.22.26), and even gives two
prominent kings that title, Kaikeyī’s father Aśvapati and Daśaratha (2.68.9; 75.6); but never, as is commonly
thought, Rāma.
16. See Klaes 1975, 24 on this question and its pervasiveness in opening the narration of Yudhiṣṭhira’s life.
17. Before Book 3, Yudhiṣṭhira is the first one with a ready answer when King Drupada doubts the dharma
of his daughter’s marriage to five men (Book 1), and it is Yudhiṣṭhira’s not answering Draupadī’s dharma question
after he has wagered her in the dice match that drives Book 2.
418 dharma

more rarely seen in Rāma—rather than being the answer to one. Both texts, how-
ever, are ingenious in leaving these early dharma questions posed by and about
their royal protagonists hanging in the air.
Once Nārada has taken leave for the sky (1.2.2), in the Rāmāyaṇa’s second
sarga, Vālmīki witnesses near his āśrama the grieving cries of a female Krauñca
bird over the slaying of her mate by a “cruel hunter.” This “brought forth the
compassion of that Ṛṣi who was the soul of dharma” and provoked him to say,
“This is adharma” (adharmo ‘yam), or as Goldman nicely renders it, “This is
wrong!” (1984, 127). Vālmīki then breaks into the utterance that creates “verse”
(and thus poetry) out of “grief” (śloka out of śoka; 1.2.9–17). Brahmā now
appears (22–36) to claim that the inspired verse came from him, and to urge
that Vālmīki should now compose in its entirety the adventure (carita) of the
“dharma-souled, virtuous (guṇavat), wise, and steadfast Rāma” for the world to
hear, “just as you heard it from Nārada” (30–31); and he gives Vālmīki the
insight to see what he did not know and what is still yet to happen in Rāma’s
life, with his word that his poem will endure so long as the rivers and moun-
tains last on earth and that it will all be true (33–35). Brahmā thus assures
Vālmīki that he will know things left out in Nārada’s encomium. Upon
Brahmā’s vanishing, Vālmīki now adopts the idea of composing “the entire
Rāmāyaṇa kāvya in verses such as these” (1.2.40cd)—that is, such as the śloka
he has just uttered.
With that, the third sarga tells how the righteous (dharmātmā) Vālmīki
righteously (dharmeṇa)18 enters into this story so exemplary of righteousness
(dharmasaṃhitam) for the first time by a sort of meditative preview (1.3.1–2), to
which a widely attested and well-known interpolation (1.154*) gives “additional
details on the sage’s vision,” including a famous verse that says, “Then, deep in
his yogic trance, that dharma-knower saw all that had taken place before him as
clearly as an āmalaka fruit placed in one’s hand” (lines 7–8; Goldman 1984,
283). Now Vālmīki offers a kind of first unfolding of what his poem will contain
(3–28). Rāma will hear Viśvāmitra’s “various other marvellous stories (nānā
citrāḥ kathāś canyāh)”
̣ (4); he will square off with Rāma Jāmadagnya (5); and, as
Nārada already indicated, he will meet Bharadvāja (8). Most important, while
adding nothing problematic on Rāma’s slaying of Vālin (15–16) and still without
mentioning Sītā’s fire ordeal, Vālmīki closes with Sītā’s banishment (28),
taking the story for the first time into Book 7.19 We still remain in the narrative
middle of Rāma’s long life, but a bit further along and into it—and implicitly

18. Goldman 1984, 283 notes of this dharmeṇa: “The term is quite difficult here.” He follows a commen-
tarial reading that seems to seek a contextual meaning, translating it as “through profound meditation.”
19. See Goldman 1984, 285: “Note that this is the only reference to events that occur in the Uttarakāṇḍa”
(7.46). Yet see n. 11 above on Rāma’s future.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 419

into the question of the relation between Rāma’s life and Vālmīki’s, for when
Sītā is banished, Rāma’s brother Lakṣmaṇa takes her secretly—and she is
pregnant—to Vālmīki’s hermitage, where Rāma and Sītā’s twins are born.
Is this still yet to happen, or are Sītā and the twins already there? That Vālmīki’s
poetic inspiration was just before this triggered by his compassion for the
painful cries of a female bird suggests at the very least a foreshadowing of Sītā’s
estrangement from Rāma, of which Vālmīki now has been given sure knowledge
by Brahmā.
Then, looking back upon the poem’s completion, the fourth sarga, con-
cluding the Upodghāta, tells how Vālmīki taught the dharma-knowing princes
Kuśa and Lava (4), Sītā’s and Rāma’s sons, the “unsurpassed tale exemplary of
dharma” (11ab) that made “the dharma-loving Munis glad at heart” (15ab). Once
it is implied that Vālmīki has now composed his poem, he calls “the whole
Rāmāyaṇa poem (kāvya) the great adventure of Sītā (sītāyāś caritam mahat) and
the slaying of Paulastya [Rāvaṇa]” (1.4.6), thereby suggesting that although
Rāma’s carita or adventure (2.30–31) has been his starting point, his complete
poem is about Sītā’s adventure, or these two adventures that are about to unfold
as one: the “profound adventure” (mahānubhāvaṃ caritam) that Rāma prepares
himself to hear at the end of the Upodghāta (4.26). For here, amid what I think
are restrained hints at the setting20 in which Vālmīki’s poem will finally be
recited (21–27), we have the captivating verse in which Rāma invites the two
youths to begin singing it before him, its own hero: “Moreover, it is said that
the profound adventure they tell is highly beneficial even for me. Listen to it”
(1.4.26cd). We are not told who said this to Rāma, or whom he is addressing.
But that would now include “we, the readers.” We are thus invited to figure out
how this story will be of benefit even for Rāma, which requires that we wait to
learn more about the scene of this recital, which will take us further beyond
Sītā’s banishment into the seventh book. By this time, the point seems to be
that Vālmīki’s poem will be about how the perfection of Rāma could relate to
the compassion and grief that Vālmīki has felt over the banishment of Sītā.
The Rāmāyaṇa’s frame is then picked up in Book 7, when Vālmīki says his
first words within the story itself, welcoming the banished Sītā to his ashram
with hints at the preamble when he tells her he knows everything by his
“concentration on dharma” (dharmasamādhi), and above all that she is “without
sin” (apāpām; 10a) and of “pure character” (viśuddhabhāva; 7.48.9–10). Kuśa
and Lava then pick up from the preamble directly by singing their parents’
“profound adventure” to their father in person during intervals of Rāma’s horse

20. The Aśvamedha is not mentioned and Rāma is on a throne and amid an assembly (pariṣad). As Goldman
indicates (1984, 288, notes to verses 13, 21, and 27), it is possible to take this as implying Rāma’s court at Ayodhyā,
but commentators and interpolations have set it at the Aśvamedha, which I think is probably implied.
420 dharma

sacrifice, where Vālmīki’s dramatic entry presents the occasion to reveal, or

better suggest, the poetic heart of the whole poem through its effects on its
hero and its heroine (see Shulman 2001, 255–92).
Information on the Mahābhārata’s frame is also resumed with further rev-
elations in Book 12, but, as I have indicated, the Mahābhārata’s frame narra-
tions are generations removed and cosmologically remote from the central
story. Nonetheless, in both epics there is a moment where the author emerges
from the frames to speak directly to the epic’s main listener. In the Rāmāyaṇa,
this occurs at this climactic moment when Vālmīki addresses Rāma and con-
firms Sītā’s dharma before she enters the earth (7.87.15–88.10; Hiltebeitel
2001a, 321). In the Mahābhārata, Vyāsa cannot, of course, address Yudhiṣṭhira
in the frames, although he does so, as we shall see, on several occasions in the
main story while Yudhiṣṭhira is alive. But there is one occasion where Vyāsa
addresses Yudhiṣṭhira’s successor Janamejaya directly in answer to a culmi-
nating question of the Nārāyaṇῑya. This is where Vyāsa narrates the subtale of
the Horse’s Head (see chapters 6 § B, 8 § E). For Yudhiṣṭhira to hear this is, of
course, incongruous. But why not? If Rāma can hear about his own future,
Yudhiṣṭhira could hear about that of the first and future listener to the
Mahābhārata.
These authorial frames thus have much to tell us about how the two epics
position their poet-authors in relation to questions about their main heroes’
dharma. The Rāmāyaṇa frame is shorter, more concentrated on the poet’s
relations first to the heroine and then to the hero, and more poetically account-
able in allowing one to trace questions of dharma into the main narrative and
the whole poem. It opens like a flower, petal by petal from a “perfect” bud. At
no point in its unfolding does it include a retrospective digest like the
Mahābhārata’s Anukramaṇī and Parvasaṃgraha. Its unfolding table of con-
tents never widens to the whole poem, but only leads into it. In contrast, once
the Mahābhārata opens a clearing on its two great trees of dharma and
adharma, it soon surrounds them with the forest.

C. Sidestories and Subtales, Foregrounding

and Legal Precedent

Both epics raise the matter of ancillary stories within their introductory material
and frames, but in different ways—in each case, with important bearing on
their differing treatments of dharma. We can approach these differences with an
initial pair of hypotheses. First, regarding what the two epics have in common,
in each epic, one can usefully differentiate early ancillary narratives that provide
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 421

background for the development of the central story from later ones that leading
characters, now adults in the midst of life’s complexities, hear (and sometimes
tell) themselves for their own edification or entertainment. Second, regarding a
point of difference, whereas the Rāmāyaṇa’s early ancillary stories are concerned
to foreground the extraordinary dharmic character of Rāma that is emerging
along with their telling, the Mahābhārata’s early ones are concerned with legal
precedent in gray areas of dharma that these stories expose while arriving at
legally binding solutions that will have their bearing on the more ambiguous
world in which the chief characters are still yet to emerge.
The Rāmāyaṇa’s frame offers no complete table of contents, but in men-
tioning Viśvāmitra’s “various other marvellous stories,” it does not fail to indi-
cate that it will include ancillary stories. According to Goldman, Viśvāmitra’s
stories “are not directly part of the Rāmacaritam, such as the episodes of the
origin of the Ganges (1.34), the birth of Kumāra (1.36), the churning of the
ocean (1.44), the penances of Diti (1.46), and so on”; he goes on to generalize
that such tales are, “for the most part, recounted in the first and last books of
the Rāmāyaṇa” (1984, 283).21 Yet he stops short of mentioning that Viśvāmitra
concludes his run with the story of the Ṛṣi Gautama and his wife Ahalyā
(1.47–48), which engages Rāma directly when Rāma’s presence redeems
Ahalyā of Gautama’s curse for her having slept with Indra—a cautionary tale
about marriage and sexuality before Rāma’s marriage to Sītā (Sutherland
Goldman 2004, 72) that comes early within the sequence and involves one of
the great Vedic Ṛṣis who will, in effect, oversee Rāma’s career path.22 Indeed it
is also possible that Vālmikī counts the narratives that Rāma, Lakṣmaṇa, and
King Janaka (Sītā’s father) next hear from Gautama’s eldest son Śatānanda
(Janaka’s domestic priest) under the rubric of Viśvāmitra’s marvellous stories,
since they are mainly about Viśvāmitra and include his rivalry with Vasiṣṭha,
another of those great Vedic Ṛṣis who was also already very much part of
Rāma’s life.
Now a curious feature of this sequence of stories told by (1.22–24, 28,
31–48) and about (50–64) Viśvāmitra is the frequency with which the augmented
form dhārmik- is used, sometimes even hyperaugmented by incremental pre-
fixes like su-, “very”; ati-, “excessively”; parama-, “supremely”; and bhṛśa-,
“severely.” The Princeton Rāmāyaṇa translators do not nuance these terms
much and tend to render them, like much else, simply as “righteous.” But I
believe van Buitenen has the right impulse when he translates dhārmik- as
“law-abiding,” which I have on occasion been adopting in the sense of one in

21. Cf. Brockington 1998, 382: “The entire Viśvāmitra episode (1.31–64) lacks direct relevance to the main
story.”
22. See Hiltebeitel 2009a and chapter 12.
422 dharma

whom dharma not only abides but does so intensively.23 Mahābhārata usages
are proportionally much fewer and more scattered. But in the Rāmāyaṇa we
find them frontloaded in Book 1 mainly through the stories told by and about
Viśvāmitra. These Book 1 stories groom young Rāma and Lakṣmaṇa with tales
about intensely law-abiding Vedic Ṛṣis who are peers of Viśvāmitra, while both
Books 1 and 2 regale them further with references to intensely law-abiding
kings in the Kauśika line of Viśvāmitra before he becomes a Brahmin; in the
line of Janaka that Rāma will marry into; and of course in Rāma’s own Ikṣvāku
dynasty.24 All these tales told by intensely law-abiding interlocutors about
intensely law-abiding Ṛṣis and dynastic scions, all converging in Rāma’s own
solar line,25 are, of course, precedent for Rāma’s becoming intensely law-abiding
in Book 2 up through his departure for the forest and his meeting there with
Bharata.
Here are some of the more illustrative instances in Book 2, where Rāma
has by now emerged as the exemplar of all this dharmic intensity. Daśaratha
says to Kaikeyī,

I am bound by dharma’s bond (dharmabandhena baddho ‘smi). My

mind is failing me! I want to see law-abiding Rāma, my beloved
eldest son. (2.12.16)

23. Of van Buitenen’s translations of the usage, see 1.5.7, 1.46.22, 1.57.51, 5.39.37, 5.88.34, and notably,
from the sage hunter, 3.198.28: “If our king Janaka had a bad son who was a jailbird he would throw him in the
dungeon; but he does not bother a law-abiding man” (1978, 620). It would also befit Arjuna’s caution to Bhīma,
when Bhīma is ready to burn Yudhiṣṭhira’s arms for allowing Draupadī’s disrobing, which he translates differ-
ently: “no one may overreach a law-abiding elder brother (bhrātaraṃ dhārmikaṃ jyeṣṭham)” (2.61.8).
24. Book 1 has thirty-two of the Rāmāyaṇa’s sixty-seven usages of dhārmik-, with twenty-two of these in the
stories by and about Viśvāmitra. Of these, nine are about Viśvāmitra or members of his prior royal family, the
Kauśikas, with three for King Kuśanābha (1.32.5, 20; 50.18); six describe Ikṣvākus (Aṃśumant, 41.1; Bhagīratha,
41.7, 11; Viśāla, 46.11; Kuśāśva, 46.15; and Triśaṅku, 58.2); three reference Janaka (69.7) and others in his line
(69.2, 70.8); three are about Viśvāmitra as a Ṛṣi (51.7; 57.23, 64.22); one each describe Bhṛgu (37.11) and Vasiṣṭha
(53.4); one references Aditi’s sons, that is, the gods (44.14); and one is about Rāma and Lakṣmaṇa as they rever-
ence the Sarayū River (1.23.10). Book 2 has twenty-one of the remaining thirty-five usages, two of which describe
the Ikṣvāku ancestors Sagara (32.20) and Nābhāga (102.27); four each reference Daśaratha and Bharata; two ref-
erence Rāma; and one describes the boy whom Daśaratha killed by mistake (58.26). Books 1 and 2, with 5262.5
(28 percent) of the Rāmāyaṇa’s 19,100 verses (Brockington 1998, 65), thus has fifty-three (79 percent) of its sixty-
seven usages.
25. Vālmīki seems to protest too much in making the solar line so much more “law-abiding” than the
lunar (it includes, among other flawed monarchs, the cannibal Kalmāṣapāda [see chapter 4 § B.1.d.i], though
not, in Vālmīki’s genealogies, the wayward Mahābhiṣa [see chapter 8 § B and Brodbeck 2010d]). Kālidāsa picks
up on the utterly normative character of solar line kings, “who, in childhood, studied all good arts, and next in
youth sought each worldly joy; who in age lived hermit’s lives; and cast away their bodies by devotion’s power at
last” (Raghuvaṃśa 1.8; Devadhar [1985] 2005, 2; reference thanks to Vishwa Adluri). Thapar 2005m, 723–28
contrasts the two lines in terms of solar primogeniture versus lunar segmentation, geographical spread, and
openness to non-normative kinship patterns: “rulers of Kosala and Videha are . . . seen as belonging to co-lateral
lines” and “Ikṣvāku descendants seem hardly able to move away from the middle Ganga plain” (724–25).
Ikṣvāku dynasty constructions, which include Ikṣvāku connections for the Buddha (see chapter 4 § A), merit
further study.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 423

As Rāma leaves Ayodhyā, Daśaratha tries to hold him in view:

As long as the king could see his beloved law-abiding son, he seemed
to stand firm on the ground just to have him in sight. (2.37.2)

Soon Rāma’s charioteer Sumantra returns to the desolate city, saying,

We shall never again show ourselves at feasts or sacrifices, at weddings

or great assemblies, since law-abiding Rama will not be there. (2.51.11)

Finally, as Book 2 rounds to its close, Bharata, finding Daśaratha dead upon his
return to Ayodhyā, tells his mother Kaikeyī it is she who should be banished,
and snaps:

What possible wrong could severely law-abiding (bhṛśadhārmikaḥ)

Rāma have done you, that because of you they [Daśaratha and Rāma]
should have found death and banishment both at once? (2.68.3).

After Book 2, one meets such terms far less often, with only two more usages for
Rāma—the first couched in the negative when Vālin rebukes him for shooting him
from ambush and says he did not know Rāma was “an un-law-abiding hypocrite
(dharmadhvajam adhārmikam), like a well overgrown with grass” (4.17.18).26
If one looks further into the Rāmāyaṇa’s more edifying and entertaining
ancillary tales, while it is true that most of them are in its first or last books, not
all are, and some are told by Rāma and Sītā themselves.27 As in the Mahābhārata,
which knows many of the same ancillary narratives, they are told to shed light
on evolving situations in the main story, and are directly or indirectly illustra-
tive of dharma. But the differences with the Mahābhārata are equally striking.
First, Vālmīki has no other generic term for them than “marvellous” or
“colorful” stories, kathā, and thus does not set them off generically.28 In this

26. The last usage for Rāma comes at 6.11.27 where Rāma’s monkey counsellors praise him for acting
deliberately in receiving Rāvaṇa’s brother Vibhīṣaṇa.
27. Those in Books 2–7 include Rāma’s tale of the Ṛṣi Agastya’s destruction of the Asuras Vātāpi and Ilvala
and his arresting of the growth of Mount Vindhya (3.10.52–64, 77–83; cf. Mbh 3.94–108); and the story of the
bi-gendered Ila/Ilā (7.78–79; on which see chapter 12 § C), followed by that of Purūravas (Rām 7.80–81). Sītā’s
include her story to Anasūyā of her own marvellous birth (2.110–11), on which see chapter 10; and her story to
Rāma about the ascetic seduced by Indra’s sword who turned violent and went to hell (3.8). Others include
Bharata’s story to Kaikeyī about the cow Surabhī (2.68); Vasiṣṭha’s to Rāma of the origin of the world and the
Ikṣvākus’ genealogy and tradition of primogeniture (2.102); Sugrīva’s to Rāma about the buffalo demon
Dundubhi (4.11); and the stories of the sage Niśākara-Muni (4.49) and the Puṣpaka chariot (6.109). Cf. Hiltebeitel
2005a, 470 n. 31. A thorough inventory would be welcome.
28. The Rāmāyaṇa makes very selective use of the terms by which it defines itself in course, tying them
from the beginning into its poetic inspiration and using them only rarely to define other stories told in passing
(e.g., ākhyāna once for the “Descent of the Gaṅgā”; purāṇa once for old prophetic lore on Rāma’s future life)—
never upākhyāna. Uses of kathā can be more poignant, as when Anasūyā asks Sītā about her birth and
“svayaṃvara” as a story she has heard (Rām 2.110.23–24); again, see chapter 10.
424 dharma

fashion, and in contrast to the Mahābhārata, the terms are used strategically
rather than definitionally, and they are not used to emphasize the interplay bet-
ween the Rāmāyaṇa’s parts and its whole. Emerging from and flowing back
into the passages that frame the Rāmāyaṇa (the upodghāta and the Aśvamedha
recital scene), such stories fall within a single poetic narrative that is portrayed
as being addressed uninterruptedly to Rāma (with the one exception at the
recital scene where Rāma tells someone else or some others to “listen”). They
are not conveyed to multiple audiences within the main story or ultimately to
further audiences through a thrice-told stacking of dialogical frames (see
Shulman 2001, 28–33; Hiltebeitel 2005a, 464).
Second, the Rāmāyaṇa’s single unfolding frame leaves the rest of the
Rāmāyaṇa to be all in verse without unversified phrases like “So and so
spoke (uvāca)” to set off units by speakers, and without dialogical shifts
from frame to frame or from lead narrators to subnarrators, including nar-
rators of ancillary narratives. Thus third, no matter who tells such ancillary
tales to whom, and even if the primary speaker or listener is Rāma, they are
all told ultimately to Rāma within the “profound adventure” he hears from
his sons. Indeed, Rāma hears his boys start with the first twenty sargas
beginning “from the sight of Nārada” (nārada-darśanāt) (7.84.11)—that is,
from the beginning of the upodghāta on. Rāma would have heard all three
of Vālmīki’s inspirations and all about how extraordinarily law-abiding he
and so many others were who all behaved so well! Rather than having
dialogical frames, the Rāmāyaṇa frame makes the whole poem an apostrophe
to Rāma.
The Mahābhārata differs. It has unversified inset phrases to indicate
speaker shifts. It has multiple “chief listeners” and no one character within the
main story who listens to the whole. And it sets off its ancillary tales both in its
frontmatter and its framing with specific generic terms. Having proceeded so
far with the neutral term “ancillary stories” (see Gombach 2000; cf. Nanavati
1982), we may now call the Rāmāyāṇa’s ancillary stories sidestories in contrast
to those in the Mahābhārata, which can be called subtales—one reason for this
being that they are all told by subnarrators. Be it noted that while I work with a
list of sixty-seven stories defined as subtales in one way or another within the
Mahābhārata, this is not a closed group, and many more of this epic’s ancillary
tales conform to a “subtale type.”29
In having numerous ancillary stories cited in its table of contents, the
Parvasaṃgraha’s chief terms for them are akhyāna, “tale,” and upākhyāna,

29. For my count of sixty-seven Mahābhārata upākhyānas, along with the point that “it is not a boundaried
group,” see Hiltebeitel 2005a, 467–69. For a looser list of thirty upākhyānas (including three in Appendices) in
just the Śāntiparvan (where I list only fourteen), see Belvalkar 1954–66, clxiii (the list is by V. M. Bedekar).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 425

“subtale.”30 Both are also used for the Mahābhārata as a whole: ākhyāna
frequently,31 leaving a rather bizarre impression at the one such usage of
upākhyāna (1.2.236a), which, right after the Parvasaṃgraha, seems to suggest
that the Mahābhārata is a subtale of something else!32 The Mahābhārata also
reserves the term upākhyāna for tales told within or pending from the first dia-
logue level of Vaiśaṃpāyana and Janamejaya. It does not use it for stories in the
Pauṣya (1.3), Pauloma- (1.4–12), and Āstīka-Parvans (1.13–53), which, as extended
frontmatter, are narrated to Śaunaka and the other Naimiṣa Forest Ṛṣis by the
bard Ugraśravas, and include the stories of Kadrū and Vinatā, Garuḍa, the
Churning of the Ocean, and so on. The Parvasaṃgraha reflects this, mentioning
as upākhyānas only substories told within the main narrative, which makes
them tales told to members of the larger Kaurava–Pāṇḍava household, including
Janamejaya, to enable those listeners to consider what varied narrators deem to
be pertinent to the listeners’ questions and situations. This distinctive use of
upākhyāna only for stories subordinate (upa-) to the main story (often called an
ākhyāna) suggests what may be the primary significance of the term upākhyāna,
which seems to originate with the Mahābhārata.33 In contrast, it is possible to
say that the number of upākhyānas in the Rāmāyaṇa is zero.34
Mahābhārata substories are important to this chapter because in the next
section I will involve them in my interpretation of the episode of “The Yakṣa’s
Questions.” One surprise about that episode is that the Parvasaṃgraha calls it
an upākhyāna, describing it as “The Firesticks Subtale where Dharma Instructs
His Son (āraṇeyam upākhyānaṃ yatra dharmo ‘nvaśāt sutam)” (1.2.127ab). To
call it a subtale looks incongruous, for not only is it part of the epic’s main
story,35 it is in fact unique because its listener hears questions rather than a

30. For the seven upākhyānas mentioned in the Parvasaṃgraha, see Hiltebeitel 2005a, 469 n. 29. Ākhyāna
is used there ten times to describe thirteen tales, all but two of which are elsewhere also called upākhyānas
(it seems that metrical fit may sometimes decide which term is used).
31. It is the most frequently used term for the whole, used even more often than itihāsa (Hiltebeitel 2005a,
465).
32. I have suggested Veda (Hiltebeitel 2005a, 471), not entirely whimsically since the Mahābhārata is the
“fifth Veda.” On these matters, see Idem, 465, 470–71. The passage (1.2.236–41) is rich in genre terms and in
referring to the Mahābhārata as a source that other poets will live off of; see chapter 5 § C; Hiltebeitel 2009a,
202–5.
33. I propose this as an improvement over my suggestion in Hiltebeitel 2005a, 270–72 that whereas
ākhyānas can be overarching tales containing subnarratives, as with the multistoried Ᾱstīkaparvan (Mbh 1.13–53)
billed as an ākhyāna at 23.5cd and 13.4a), upākhyānas tend to be uninterrupted single or closely intertwined
tales.
34. See Hiltebeitel 2005a, 470, 476. There is an interpolated verse in the Rāmāyaṇa’s Aśvamedha recital
scene where the twins tell Rāma that the Rāmāyaṇa has 24,000 verses and a 100 upākhyānas (7.1328*, following
7.85.20), but this is only suggestive of Mahābhārata influence.
35. Only three others—the Saubhavadha- (3.15–23), Ambā- (5.170–93), and Nakula-Upākhyānas
(14.92–96)—involve main characters in their current lives. I regret saying that the Yakṣa episode is the only
such case (2005a, 484).
426 dharma

sub-story. In effect, Yudhiṣṭhira lives a substory, and draws, I will argue, on

information learned from hearing other substories in his answers. I will thus
interpret the episode as a “substory clearing house”: one in which Yudhiṣṭhira
is tested on what he has learned so far in life, for which the subtales he has
heard in the forest provide a fair index. This offers us a way to approach the
Mahābhārata’s subtales by thinking of them in four broad groups. First come
ten in Book 1 that Yudhiṣṭhira has not heard. Second are the twenty-one, all but
one of them recounted in Book 3, that he hears by the time he meets his father
Dharma disguised as a Yakṣa. Eight subtales are then told between Books 5
and 9 that relate mainly to themes of war. Of these, Yudhiṣṭhira hears only the
first, which recounts a set of stories about Indra through which the narrator
predicts Yudhiṣṭhira’s victory (Mbh 5.9–18). Finally, there are twenty-seven
postwar subtales, of which Yudhiṣṭhira hears twenty-six, all but one of them in
Books 12 and 13.36 As hypothesized at the beginning this section, unlike the
Rāmāyaṇa’s early sidestories which foreground the dharmic excellence of the
hero, some of the Mahābhārata’s earliest subtales (I will now mention the first
five) introduce matters of legal precedent.
We met one of the clearest examples of this in chapter 8 § G, where Pānd ̣ u
̣
tells Kuntī the story of young Śvetaketu’s repeal of the “eternal dharma” of wom-
en’s sexual freedom by ruling that henceforth, women’s unfaithfulness will con-
stitute aborticide or brahmanicide, as will seducing someone else’s chaste wife,
and that a wife who refuses a husband who tells her to perform niyoga “shall
incur the same evil.” As noted, Pānd ̣ u
̣ introduces this story as “ancient Law
(dharmam . . . purānạ m)” (1.113.3), which means that it is not explicitly an
upākhyāna. But it is his direct answer to Kuntī’s necrophilia story (kathā) about
King Vyuṣitāśva, which is the Mahābhārata’s fifth subtale called the Vyuśitāśva-
Upākhyāna—the only one told by a woman, though, as we shall see in chapter 10,
Draupadī also tells a story of the “subtale type.” Among the early subtales, how-
ever, the fourth is also about legal precedent and a change in prior law. This is
the Aṇīmānḍ ạ vya-Upākhyāna (1.101), which introduces a change in law when
the sage Mānd ̣ ạ vya learns from Dharma—who is functionally tantamount here
to Yama Dharmarāja, the god of death who oversees karmic retribution—that
the reason he ended his life with an impalement stake up his rectum is that he
mistreated insects similarly in his childhood. Mānd ̣ ạ vya establishes that there
will henceforth be a “limit (maryādā) on the fruition of dharma”: that sins com-
mitted before the age of fourteen will not be counted an offense; and he curses
Dharma to take birth as Vidura (see Hiltebeitel 2001a, 192–95).

36. This classification differs from those in Hiltebeitel 2005a (where, for some reason, I have Yudhiṣṭhira
hear forty-nine subtales [472] rather than the forty-eight reached here).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 427

Equally significant, in a different way, are the first three subtales. These intro-
duce legal precedent not so much as “supreme court” changes in the law itself, as
in setting the legal precedents for handling crisis situations in the lunar dynasty.
As the epic’s first subtale, the Śakuntalā-Upākhyāna (1.62–69) encodes caste pre-
rogatives37 and gender roles.38 King Duṣyanta faces Śakuntalā’s paternity suit as a
legal court case, in which she invokes the agreement (samaya; 69.25) she secured
before agreeing to make love to him; and the court case concerns the royal inher-
itance of their son, the eponym Bharata. In the second subtale, the Yayāti-
Upākhyāna (1.70–80), Yayāti mishandles matters involving his rival wives—the
first a Brahmin and the second an asuric Kṣatriya—with his inheritance going to
his youngest son Pūru, born to the junior wife rather than to his “disqualified
eldest” Yadu, son of the Brāhmaṇī.39 The third upākhyāna is then one we have
met. In the Mahābhīṣa-Upākhyāna (1.91), an adhyāya early in the story of Mother
Gaṅgā’s intervention in this same dynastic line, one learns of the agreements
made by Gaṅgā and the Vasus after King Mahābhiṣa’s heavenly indiscretion,
which all take on binding legal force in the lives of King Śaṃtanu and his son
Bhīṣma (see chapter 8 § B). In effect, these lessons in legal precedent present just
the opposite of what we have seen in the foregrounding that Rāma’s ever so
law-abiding ancestors provide for him. It is not that lunar dynasty kings are less
law-abiding. It is just that they bring the law into play because they like to mess
around, which results in legal tangles that carry forward into the main story.
With this in mind, and also taking into consideration our four-phase catego-
rization of Mahābhārata subtales, it is useful to reflect on Masaji Chiba’s compar-
ative analysis of a “three-level structure of law” as involving official law, unofficial
or informal law, and basic legal postulates (1986, 5–7).40 The first five upākhyānas
(the fifth, i.e., along with its counterstory told by Pānḍ u
̣ ), present official law: in
the first four cases direct from Vaiśaṃpāyana and thus straight from Vyāsa; in the

37. See Biardeau 1979, 118: “the apparently secondary accounts with which the epic is stuffed, far from
being what one lately calls interpolations, are the reprise under a symbolic form of the dominant message of the
principal account, which they thus aide to decipher while contributing to the progression of the intrigue.
Śakuntalā’s birth is no exception to the rule: it is clear right away that the theme of varṇa mixing takes charge of
the story.” Her father Viśvāmitra wants to be a Brahmin but is still a Kṣatriya when he sires her with the Apsaras
Menakā, which satisfies Duṣyanta that she is a Kṣatriya so that he can propose their instant Gandharva union,
even though later, in rejecting her in court, he calls Viśvāmitra “this son of a Kṣatriya avidly desirous of
Brahminhood (brāhmaṇatve lubdhaḥ)” (68.74). Śakuntalā has a different idea of her varṇa status at least on her
mother’s side, which is that Menakā is brahmayoni, “born of Brahmā,” or “of Brahmanic birth” (68.68), making
her the projection of Viśvāmitra’s desire to become a Brahmin, and giving Śakuntalā a superior status to Duṣyanta
(Biardeau 118, 120).
38. Śakuntalā puts some emphasis on her preparation to be a devoted wife or pativratā, which the Southern
Recension intensifies, tying this reassurance into her becoming Duṣyanta’s agramahiṣī or “primary chief queen”
(Hiltebeitel forthcoming-d).
39. See Dumézil 1973, 15–27; Defourny 1978, 20–37, 57–157; Biardeau 1979, 115, 125.
40. Chiba 1986, 5–7. See Menski (2003) 2005, 71 and passim with reference to “Hindu law.”
428 dharma

fifth case from a more dubious source, Pānd ̣ u

̣ . The subtales Yudhiṣṭhira then
listens to up to “The Yakṣa’s Questions” are more for his and the Pānd ̣ ạ vas’
entertainment and edification. Some of them present “mirror stories” held up to
the predicaments the Pānd ̣ ạ vas and Draupadī find themselves in,41 but all of
them may be said, generally, to present narratives from which matters of prece-
dent in unofficial or informal law can be inferred by their careful first listeners
headed by Yudhiṣṭhira. The third group surrounding the war is then for the most
part less interesting regarding law and its impact on Yudhiṣṭhira or the dynastic
line, with an exception for the Ambā-Upākhyāna (5.170–93), which takes us back
into the miasmas brought about by Bhīṣma’s abduction of the “three Ambikās”
of Kāśi and explains his singular vow not to fight with anyone who has been born
a woman (see chapter 8 §§ D and E). Finally, the fourth group of postwar subtales
can be said to engage a now more mature and wisened Yudhiṣṭhira in thinking
through basic legal postulates as he listens with his wife and brothers.42 Chiba
defines these as follows: “A legal postulate is a value principle or value system spe-
cifically connected with a particular official or unofficial law, which acts to found,
justify, or orient the latter” (1986, 6; author’s italics). This fits quite nicely the
broad headings of Rājadharma, Āpaddharma, Mokṣadharma,43 and Dānadharma
through which nearly all the postwar upākhyanas—including the White Island
“narrative,” said to be the “essence” of all the upākhyānas,44 and the “Horse’s
Head Upākhyāna” (12.334) that it leads Śaunaka to ask about—are threaded into
Yudhiṣṭhira’s dialogue with Bhīṣma. Moreover, it is mainly in this skein up to
and including the last subtale that Yudhiṣthira hears a number of what I have
called “puzzle piece” upākhyānas in which Dharma either appears himself in the
story, sometimes in disguise, as he did to ask Yudhiṣṭhira “The Yakṣa’s
Questions,” or in which dharma/dharman is disguised in puzzling characters
who have that term in their names.45
This four-phase analysis of subtales can thus help us to contextualize the
episode of “The Yakṣa’s Questions.” Yudhiṣṭhira’s forte of repeatedly asking

41. Biardeau 2002, 1: 412–13 uses this term for the Nala-Upākhyāna (3.50–78), Rāma-Upākhyāna (3.257–76),
and Sāvitrī-Upākhyāna (3.277–83)—three of the best known, and all from the Forest Book 3, to which I have suggested
adding the Sunda-Upasunda-Upākhyāna (1.204) which, being about two demon brothers who fight over one wife,
̣ ạ vas’ polyandry (Hiltebeitel 2005a, 482; cf. 476, 483).
holds up an “inverse mirror” to the Pānd
42. See Hiltebeitel 2005a, 472, and 490 on Mbh 13.57.42–44, the one mention of Draupadī listening to
Bhīṣma, which she has probably been doing all along.
43. See Fitzgerald 1980, 231: “So the majority of texts collected in the MDh focus directly on mokṣadharma-s,
that is, behavioral or attitudinal norms (dharma-s) leading to mokṣa, ultimate personal transcendence of the
limits, pain, and misery common to the situation of all living beings.”
44. See chapter 6 § B: in the Nārāyaṇīya, when Bhīṣma tells Yudhiṣṭhira about Nārada’s journey to White
Island, he indicates that this “narrative” (ākhyāna) embodies the “essence” of all the other dharmya subtales he
has transmitted (12.326.114–15). See Hiltebeitel 2005a, 506–507; cf. also Mbh 12.238.13–15.
45. For preliminary discussion, see Hiltebeitel 2005a, 487–88, 491–92.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 429

and sometimes answering questions about dharma governs much of the

Mahābhārata, particularly, as we have outlined, in the Forest Book 3, and in the
oft-called “didactic” Books 12 and 13.

D. Monstrous Encounters

At the beginning and end of each epic’s third Book, as Rāma and Yudhiṣṭhira
enter and leave the forests, they encounter monsters who provide the Forest
Books with bookends. The first monsters, Kirmīra in the Mahābhārata (3.11–12)
and Virādha in the Rāmāyaṇa (3.3), merely guard access to the forests. They have
little immediately to do with dharma. But two things are interesting about the
Rāmāyaṇa’s Virādha episode. One is that it is not mentioned in the Rāmopākhyana,
the Mahābhārata’s subtale on Rāma and its main telling of the Rāma story.46 This
means that although the bookend symmetry one finds in the Rāmāyaṇa could be
modeled on the Mahābhārata, it cannot be modeled on the Rāmopākhyāna. The
second is that the Rāmāyaṇa’s forest-entry and forest-exit scenes have far greater
symmetry between them than the Mahābhārata’s. Whereas the Mahābhārata’s
Kirmīra and Yakṣa encounters have only their benchmark positions in common,
the two Rāmāyaṇa encounters are virtual duplicates. A brief summary of the
beginning of Rāmāyaṇa Book 3 will allow me to make this point.
This Forest Book’s very first line finds Rāma, Lakṣmaṇa, and Sītā entering the
vast Daṇdạ ka Forest: the southern destination required of Rāma by Kaikeyī (Rām
2.10.28) and the forest into which the great Ṛsị s, who gather fruits there, have both
collectively and individually set forth Rāma’s path (2.111.19). The trio first sees a
circle of ashrams whose unnamed sages welcome Rāma and request his protec-
tion, since they regard him as their king whether he is in the city or the forest. But
soon Virādha (“One Who Thwarts”) looms before them, seizes Sītā, and chal-
lenges the brothers for entering this forest with her. Pained by seeing anyone else
touch Sītā, Rāma fills the huge Rākṣasa with arrows and he and Lakṣmaṇa then
each break off an arm to release her. Asking Rāma and Lakṣmaṇa who they are
(3.3.1), Virādha realizes he has been slain by Rāma: a long-awaited blessing that
relieves him from a curse. He can now go to heaven, but before that he tells Rāma
to go next to the Ṛsị Śarabhaṅga who “will see to your welfare” (3.22–24).
From there, Rāma is relayed by Ṛsị s from one forest site to another until
Sītā is abducted, whereupon Rāma and Lakṣmaṇa start looking for her and
happen upon a second Monster, Kabandha. As we turn to that story, which does

46. See Mbh 3.261.39–40, where it would be slotted: following Bharata’s return to Ayodhyā, Rāma, “fearing
that the townsfolk and countrymen would return,” entered the Daṇdạ ka Forest “by the hermitage of Śarabhaṅga.”
430 dharma

have a slot in the Rāmopākhyana, the obvious difference is that this second mon-
ster cannot grab Sītā, who is now abducted. I believe that Vālmīki would have
had three reasons to invent the Virādha encounter. As preparation for Sītā’s
abduction by Rāvaṇa in this Book, he provides a glimpse of how Rāma reacts to
Sītā being touched. It supplies a bookend on the model of the two monstrous
encounters in the Mahābhārata. And it is a rehearsal for the Kabandha encounter,
which the Rāmopākhyāna does include, quite indispensably.47

D.1. Rāma and Kabandha

When Rāma and Lakṣmaṇa start looking for Sītā, having learned from the dying
vulture Jaṭāyus only that she has been abducted by Rāvaṇa, they pass into the
Krauñca Forest, still hoping to find her. There is no Krauñca Forest in the
Rāmopākhyana. It could be another invention of Vālmīkī to remind readers that
the suffering of a female krauñca bird underlines the sentiments of grief in sep-
aration from one’s beloved in Vālmīki’s poem. Here the brothers encounter
Kabandha, whose name means “Headless trunk” and can also denote a sacrificial
post. Kabandha guards the way past him as Virādha did; the encounter is sim-
ilar, but with one more exception. Kabandha will tell how to find the abducted
Sītā, pointing the heroes out of the forests and into the impolitic rivalries of the
monkey capital of Kiṣkindhā. Kabandha has neither a neck nor head but a
single-eyed face in his stomach, a huge mouth to devour animals, and long arms
to grab them that suddenly seize the brothers (3.65.15–20). Asking Rāma and
Lakṣmaṇa who they are, and to state their purpose or be devoured (24–26), the
heroes muster their courage and each sever one arm (66.6). Hearing from
Lakṣmaṇa that his arms have been severed by Rāma, “who has the power of a
god” (asya devaprabhāvasya; 11a), Kabandha realizes that this brings an end to a
long curse which had restricted him to this hideous form until Rāma should cut
off his arms and cremate him (67.6, 15). When Kabandha rises lustrously from
his pyre, he says that Rāvaṇa’s abode may be found if Rāma allies with Sugrīva:
“Quickly make him a comrade (vayasya), having gone there now, Rāghava,
sealing your compact in the presence of blazing fire to shun all trickery.”48

47. A suggestive piece of evidence that Vālmīki has imported features of the Kabandha story back into his
Virādha story is that Virādha is a former Gandharva named Tumburu (Rām 3.3.8), just as Kabandha is a former
Gandharva named Viśvāvasu in the Rāmopākhyana (Mbh 3.63,38), whereas in the Rāmāyaṇa Vālmīki upgrades
Kabandha into a former unnamed Dānava (Rām 3.66.8).
48. Rām 3.68.13. In the Rāmopākhyāna, the killing of Kabandha sets free a puruṣa that not only reveals
Kabandha’s prior identity to have been that of the Gandharva Viśvāvasu cursed by Brahmā to pass through a
Rākṣasa womb (34–38), but the partly erroneous information that Sītā was abducted by Rāvaṇa, who dwells in
Laṅkā, and that Rāma should seek out the help of Sugrīva, who “surely knows Rāvaṇa’s seat” (Mbh 3.263.34–42;
van Buitenen 1978, 737). See Hiltebeitel 2009a, 194–95.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 431

The fairly unusual term vayasya, which has no corresponding usage in the
Rāmopākhyāna, literally means “contemporary,” but is used in the sense of
“commiserator” or “sympathizer.” Vālmīki will continue to use it to define Rāma
and Sugrīva’s special kind of friendship. Meanwhile, the caution to avoid trickery
makes Kabandha’s advice moral, at least as it bears on Rāma’s arrangements
with Sugrīva, if not with Sugrīva’s brother Vālin, Sugrīva’s rival for the throne—a
matter we will take up in section D.
In both epics, the second monstrous encounter thus has the role of guid-
ing the king beyond the forests as he begins his return to society—in Rāma’s
case, entering the no-man’s land of monkeys which inversely mirrors the
politics of Ayodhyā; in Yudhiṣṭhira’s a place of hiding—before the eventual
demands of war. The king’s dharma is changing along with his circumstances,
and these second monster encounters are pivotal to the changes he needs to
make. Rāma and Yudhiṣṭhira are no longer honorable exiles who have the
leisure to model their royal dharma and their meritorious words and actions on
the sagely instructions they receive along their forest route. They will have to
face the calamities their dharma biographies have been preparing them for.
Kabandha makes this point only implicitly, never mentioning dharma directly.
In the Mahābhārata, this second encounter places Yudhiṣṭhira’s dharma under
an explicit and thorough review.

D.2. The Yakṣa’s Questions

It would seem straightforward to examine the Mahābhārata episode of “The

Yakṣa’s Questions” (Mbh 3.295–99) from the standpoint of dharma, since the
son whom Dharma desires to know through questions, Yudhiṣṭhira, is called
Dharmarāja and Dharmaputra (son of Dharma). But a difficulty lies in the
fact that readings of the episode have done little to explore the relationship
between dharma (lower case) and Dharma (the deity),49 or have sought to
diminish its significance by maintaining that the episode is really about
something else. The two sustained readings that I am aware of, both pub-
lished in 1991, are of this second type. Gail Hinich Sutherland interprets the
story as one in a “cycle” of four about aquatic spirits, taking Varuṇa rather
than Dharma as the Yakṣa’s prototype, since he is the lord of waters, and
approaching D/dharma as embodying an opposition between natural and

49. I set aside Biardeau 2002, 1: 755–60, which builds on the promising point that “Tout cet épisode est un
construction du dieu Dharma” (759), but does not pursue it very far, and not at all into the details of the Yakṣa’s
questions. See also Hiltebeitel [1976] 1990, 186–91, attempting to locate the episode among Indo-European
parallels; Laine 1991, 280–81, taking the episode as initiatory, with Dharma “understood . . . in terms of universal,
saṃnyāsic virtues, the anti-structural values of communitas”; Klaes 1975, 78–81, with some points to be noted.
432 dharma

social Law that the waters and Varuṇa somehow mediate (1991, 85–103).
Sutherland is not much concerned with the episode’s Mahābhārata context,
though in bringing in the Buddhist Devadhamma Jātaka ( Jātaka 6, Sutherland
93–96; Cowell [1895] 2005, 1: 21–23) as the only strong parallel, she can help
one to formulate hypotheses about how the Mahābhārata’s allomorph is
deployed in the epic.
The episode’s most stimulating reading, David Shulman’s, finds it to be
about a “preoccupation with language,” and that it encourages a “dissection of
language levels” ([1991] 2001, 42, 51). Shulman lodges dharma at an
intermediate level of meaning between a “simple truth” or “news” (vartika)
level, and a “context-dependent” daiva or destined level that opens out on the
“larger epic frame” through the “eyes of the narrator-poets” (53–55). But
whereas the first level interests him for its direct formulation of obvious truth
and the third “probably reflects a latent theory of knowledge in relation to lan-
guage” in opening the episode to Yudhiṣṭhira’s destiny (55–56), the dharma
level concerns ethical language that has lesser interest from this perspective.
Shulman sees it as the one whose “truth level” is most open to “potential
disjunction” and “gaps” (51), since Yudhiṣṭhira’s “chosen ideal” (57) of
ānṛśaṃsya, which he translates mainly as “non-injury” (47; cf. 50, 58), is a
dharma value that Yudhiṣṭhira will be unable to uphold, so that by the end the
best he can do is to “curse” Dharma by linguistic means of negation (57–59).
I have urged that “non-injury” should be avoided in translating ānṛśaṃsya,
since it does better to translate what falls under the domain of ahiṃsā, “non-
violence” or “non-harming.” I prefer to translate it more or less literally as
“noncruelty.”50 Also, Yudhiṣṭhira is not unable to uphold this ideal, but rather
keeps it before him throughout his life51—including, as we shall soon see, even
in the Mahābhārata war.
Now, while an episode based on riddles is certainly enriched by analysis of
its language, I feel that it is unlikely, given the long-standing Vedic association
of dhárman with enigmas (see chapter 3 § C) and the puzzle-packed character
of the Mahābhārata as a whole, that the episode has an unusual linguistic
preoccupation. In fact, in three cases where Shulman assembles evidence to

50. From anṛśaṃsa, “uncruel”; “ānṛśaṃsya begins from a feeling of the ‘absence of injuring men’ (nṛ), . . .
as if that were a good and realistic starting point for a species trying to imagine a way out of its own cycles of
violence” (Hiltebeitel 2001a, 211). James Fitzgerald (personal comunication, 2008) says he likes to translate it as
“kindness,” which I too like—for reasons I will bring out below—for its resonance with kin, suggesting the coin-
age “kindredness.” Others have done with “compassion” or “benevolence.” See chapter 5 § D at n. 117, as the first
listed among thirteen sādhāraṇa darmas (Mbh 12.285.23).
51. See Hiltebeitel 2001a, 202–14, 260, 268–70, 275 (he also hears about it in stories told to him: 230–31
299, 303). The point is nicely traced by Klaes (1975, 71–73, 76–81, 89, 94, 96, 112, 116, 124–28, 133–36) and Lath
(1990).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 433

that effect, he puts a twist on usages that are not accepted by the Pune Critical
Edition.52 Nor does he discuss one of the question-and-answer sets that explic-
itly links language with dharma:

Yakṣa: What in a word has to do with dharma (kiṃ svid ekapadaṃ

dharmyam)? What in a word is fame? What in a word has to do with
heaven? What in a word is happiness?
Yudhiṣṭhira: In a word, skill has to do with dharma. In a word,
giving is fame. In a word, truth has to do with heaven. In a word,
morality (śīla) is happiness. (13.297.48–49)

Yudhiṣṭhira and Dharma seem to agree that dharma has to do with skill in find-
ing the right encapsulating word. Moreover, while dissection into levels can be
a useful heuristic, it is a scholarly contrivance, and in this case a problematic
one from the start, since Shulman draws the terms for his first “news” level
from an interpolation that was added to another interpolation: a radically nega-
tive statement about reasoning (tarka) in relation to śruti and dharma that was
only belatedly put into Yudhiṣṭhira’s mouth:

Reasoning is without foundation (tarko pratiṣṭhaḥ); the śrutis contra-

dict one-another (śrutayo vibhinnā); there is not even one Ṛṣi whose
opinion is authoritative (yasya matam pramāṇam). The truth about
dharma (dharmasya tattvam) is hidden in a cave. (3, App. 32, lines
65–68 = Kinjawadekar 3.313.117)

This lovely verse,53 which may recall the Kaṭha Upaniṣad’s identification of its
“subtle dharma” with the ātman hidden in the cave of the heart (KU 2.11–12),
is an obvious interpolation added onto an equally obvious prior interpolation.
The verse occurs in only one manuscript, the so-called “Vulgate” edited by
Nīlakaṇṭha;54 and it comes amid fourteen lines added to a prior interpolation

52. See Shulman 2001, 41–42 on Vulgate [= Vlg] 3.312.1c with artham rather than CE 3.296.1c with atra,
taking the former as “meaning” where, whatever it means (Johnson 2005, 281 ignores it), the CE does without it;
Idem 44–45, 59, citing and translating vṛnute (Vlg 313.53c) rather than CE vṛścate (3.297.35c), taking the former
(÷vṛ) to indicate that the Vedic verse (ṛc; Vlg) or speech (vāk; CE) “alone veils (or envelopes) the sacrifice which does
not extend beyond it,” whereas the CE’s ÷vraśc with vāk suggests the less cryptic but more mantrically interesting
“speech alone cuts up (or cuts off ) the sacrifice which does not extend beyond it”; Idem 49 on Vlg. 313.120b with
śabdaḥ puṇyena karmaṇā rather than CE 3.297.63b with śabdaḥ puṇyasya karmaṇaḥ, translating the former, “The
word [or sound] touches heaven and earth together with [in association with, through] a good deed; as long as that
word exists, one may be called a man” (Shulman’s brackets, my italics) and taking this to suggest “that a human
being is someone who, through language, connects disjointed domains,” whereas it might suffice to say (with
van Buitenen 1975, 803) that “word of a good deed” or “the repute of a good deed touches heaven and earth.”
53. On the verse’s renown, see Matilal 2002, 67; Shulman 2001, 54.
54. As Sukthankar 1942, 1,032 notes, “Dn alone cont. (!)” Does the exclamation point reflect a suspicion
that the commentator could be the author of this verse? I believe that it would be a legitimate suspicion. Cf.
Austin 2009, 609–15 on another intervention by Nīlakaṇṭha.
434 dharma

found in only three Northern manuscripts, the Vulgate included, “which,”

says the Critical Edition editor, “from the documentary viewpoint is a pal-
pable interpolation, as is also evident from intrinsic considerations”
(Sukthankar 1942, 1,032). Having just left Draupadī at the hermitage of
Mārkaṇḍeya, who has by this time prophesied the Kali yuga (see chapter 7)
and told many tales featuring dharma, it is unlikely to be Yudhiṣṭhira’s view
of Ṛṣis at the end of his twelve years in the forest that no Ṛṣi’s opinion on
dharma is authoritative. Yet Shulman sees Yudhiṣṭhira’s “exposition” as
“wholly appropriate to the epic world. Logic is no use, even the Veda and the
sages offer mutually contradictory opinions . . . ; the truth of dharma is hidden
and enigmatic perhaps ultimately beyond recovery; the world’s creatures die
before our eyes day after day and we still pursue flimsy illusions of security”
(2001, 54). We also saw in chapter 1 that there is no reason to tag Yudhiṣṭhira
with a rejection of logic. The fourteen-line interpolation is a beautiful
passage,55 and we cannot always be this insistent in favoring the Critical
Edition. But in this episode, the Critical Edition yields a highly credible text
in which the Yakṣa has already ended his questions, and Yudhiṣṭhira his
answers, in a provocatively satisfying way without these addenda. According
to the Critical Edition, Yudhiṣṭhira turns the last question—the eighteenth—
back on his interrogator:

Yakṣa: What is direction? What is proclaimed water? What is food,

Pārtha, and what poison? Tell the time of a ritual for the dead
(a śrāddha) and then drink and carry.
Yudhiṣṭhira: The good are direction, space is water, the cow is
food, a request56 is poison; a Brahmin is the time for a ritual for the
dead—or what do you think, Yakṣa? (3.297.60–61)

Indeed, the Yakṣa has invited this turnabout and implied the end of this skein
of questions by telling Yudhiṣṭhira he can finally drink if he answers this one.
As we shall see, Shulman shows nicely how the questions and answers found
in the Critical Edition and the Southern Recension build only to this point. But
he leaves an impression of parity between versions, remarking that “the Vulgate
continues with several further questions before reaching its own interesting

55. It begins with the Yakṣa asking, “Who is happy? What is amazing? What is the path, and what the simple
truth (vārtika; cf. Johnson 2005, 319: “the news”)? Answer these four questions and let your dead kinsmen live
̣ (CE 3, App. 32, lines 59–61; Kinjawadekar 3.313.114). In brief, Yudhiṣṭhira’s four answers
(mṛtā jīvantu bāndavāh)”
are that the happy find happiness at home; the wonder is that people look for stability in a world of daily mortality;
the path is “the way the great have gone” despite the hiddenness of dharma and the lack of any Ṛsị who could
authorize it; and the “news” or “simple truth” is that “time cooks all beings” (lines 62–72; Vulgate 3.313.115–18).
56. Perhaps “begging,” for prārthanā. See Johnson 2005, 311.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 435

closure” (51–52). The Critical Edition cannot be neglected when it uncovers

eighteen original questions, that being the Mahābhārata’s signature number
for a totality.57
I am thus not convinced that a dharma level is sandwiched between two
more interesting language levels, nor am I clear how a daiva level might relate to
a latent theory of language. I will open three lines of inquiry. One, already pre-
viewed, is that the episode is an upākhyāna clearing house. Second, I will propose
that Dharma in the form of a Yakṣa signals not only his relation to death and
Yama but a typical “possessing spirit.” The Yakṣa does not really kill the four
brothers; they only “fall down” or “collapse” (ni-pat), and Dharmarāja becomes
something of an exorcist in getting the possessing spirit to reveal his true iden-
tity. The case for possession is easier to make in the Devadhamma Jātaka, where
the Yakkha takes the two younger brothers to a cave and binds them, and is then,
like so many of his ilk in Buddhist narratives, converted to the “true” or “good
dhamma,” which is what this Yakkha’s single question was about, which only the
Bodhisatta could answer without even needing to be asked.58 Third, there is more
to say about Yudhiṣṭhira’s chosen value of noncruelty. Along the way I will also
continue to emphasize the dispositive character of the Critical Edition.

d.2.A. puzzle pieces and substories. As we have seen, the Parvasaṃgraha

calls this episode “The Firesticks Subtale where Dharma Instructs His Son”
(1.2.127ab). The name partly resembles one given near the end of the unit itself
to describe the benefits of reciting it: “The Great Rising Up and Meeting of the
Father and the Son (idaṃ samutthāna-samāgamaṃ mahat pituś ca putrasya ca)”
(3.298.27ab). I will get back to this usage of samutthāna, which I have translated
as “Rising Up.” But first let us consider the similarity with the Nakula-Upākhyāna
(14.92–96). Being the only other upākhyāna to be both part of the Mahābhārata’s
main story and have Yudhiṣṭhira as its interlocutor, it may give an index as to
why the Mahābhārata’s “table of contents” strains the upākhyāna category to
include “The Yakṣa’s Questions.” Like the latter, the Nakula-Upākhyāna or “The
Mongoose Subtale” ends one of the Mahābhārata’s major parvans (Book 14)

57. Recognizing this number and its implications, see McGrath 2004, 198. Cf. Stein 1936, 1937. While the
Southern Recension reaches the same turning point, it adds a few questions and thus has a different number of
them. Cf. Shulman 2001, 61–62 on the decisive thirteenth and last riddle in the Aṣṭavakrīya-Upākhyāna, and the
significance of that “totality” number with respect to the Pāṇḍavas’ year in hiding. If one totality number hints at
the completion of the Pāṇḍavas’ exile, the other could suggest their completion of the Mahābhārata.
58. Hinich Sutherland 1991, 93–94. For a Buddhist story from the Mūlasarvāstivādin Vinaya involving
Yakṣas and suggestive of exorcism, see Schopen 2006, 340–43 and n. 47. It would seem that the deceased boy’s
preta or “ghost” is mentioned after gods, Nāgas, and Yakṣas as a lower spirit that cannot get in the monastery door
because, unlike the others, it has not become devoted to the Buddha. Gods, Nāgas, and Yakṣas are possessing
spirits that can be brought under the Buddha’s warrant; see Obeyesekere 1984, 56–106.
436 dharma

with an appearance of Dharma in disguise, this time as a mongoose doubled by

a further story about Dharma where he dons another disguise. This is the last
subtale in the Mahābhārata, and its placement brings closure to the Aśvamedha
sacrifice that is supposed to have cleared Yudhiṣṭhira of his postwar self-
recriminations, but now leaves him with this mongoose to puzzle over, who
tells him his sacrifice was not worth beans, or, more exactly, worth the grain a
gleaner gets from picking over fields like a pigeon to feed his guests before he
feeds his family or himself. The “Mongoose Subtale” is also the last of the string
of subtales that I call puzzle pieces, most of which come at the ends of major
units—parvans or sub-parvans.59 All the puzzle piece subtales, like “The Yakṣa’s
Questions,”60 mark transitions in the lifelong education of Yudhiṣṭhira.
Coming as the first unit of this puzzle-piece type to engage Yudhiṣṭhira, but
also, as we saw in section B, coming after his hearing a whole skein of enter-
taining and edifying forest tales, “The Yakṣa’s Questions” is thus intriguing
because so many upākhyāna subcurrents run through it. What are Yudhiṣṭhira’s
priorities to go charging off after a Brahmin’s firesticks, dragging his brothers
with him and leaving Draupadī behind so soon after they have all heard
Mārkaṇdẹ ya tell them the Rāma-Upākhyāna and the Sāvitrī-Upākhyāna, the first
to relieve Yudhiṣṭhira’s torments over Draupadī’s abduction after she was also
left behind, and the second to answer his question whether there was ever
another woman as devoted to her husband(s)? The entire Yakṣa episode begins
with Janamejaya asking Vaiśaṃpāyana what the Pānd ̣ ạ vas did next after Draupadī
was abducted and they had gained her back (3.295.1–2), and immediately shows
Yudhiṣṭhira to be as headstrong about a Brahmin’s firesticks as he will shortly
be about playing dice in disguise, which he learned how to do after listening to
the Nala-Upākhyāna.61 Such subtale repercussions only continue to widen.62

59. See above at n. 45. Bhīṣma recounts such subtales near or at the very end of the first three and at the
beginning of the fourth subparvans of Books 12 and 13. Fitzgerald 2004a, 152, supposes that their positions indi-
cate lateness, but their recurrence in such slots raises the stakes for such an argument. I prefer an argument by
design. The Śārṅgaka-Upākhyāna at the end of Book 1 (1.220–25) could also be called a puzzle piece, but it does
not involve Yudhiṣṭhira (see Hiltebeitel 2007a). Books 2, 5, and 10 also end with riddle-like disclosures that
explain the prior action or reveal mysteries through stories that explain both past and future action. On the con-
trary, the war books begin with disclosures of what will happen at their ends.
60. On “The Yakṣa’s Questions” as “a major point of transition,” with the final test occurring “again at a
moment of closure,” see Shulman 2001, 41, 56.
61. I was mistaken to say that Yudhiṣṭhira gets this knowledge as one of his boons from answering the
Yakṣa’s questions (Hiltebeitel 2005a, 484), although Dharma implicitly confirms it by guaranteeing the Pānd ̣ ạ vas’
success in concealing themselves. It comes from Bṛhadaśva (3.78.14–17), the narrator of the Nala-Upākhyāna.
62. This is not the first time that chasing after a Brahmin’s stolen goods was an all-consuming priority.
Earlier, Arjuna reasoned it would be an overruling dharma to recover a Brahmin’s stolen cows even though it
meant interrupting Yudhiṣṭhira and Draupadī in the bedroom, where Arjuna had left his weapons, and a breach
of the agreement the Pāṇḍavas had accepted after hearing the Sunda-Upasunda-Upākhyāna from the busybody
Ṛṣi Nārada. See Biardeau 2002, 1: 760 on the two incidents; Hiltebeitel 2001a, 264–68, 270 on the earlier epi-
sode and subtale and their repercussions, which involve Arjuna finding other wives, including Subhadrā.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 437

I thus suggest that the Yakṣa’s test challenges Yudhiṣṭhira to answer what
he has learned by breaking down the narrative information he has received to
its most irreducible units. “What in a word has to do with dharma?” (297.48a),
quoted above, was the Yakṣa’s twelfth question. The thirteenth is then espe-
cially suggestive of lessons learned from Yudhiṣṭhira’s own life and from
hearing about others’ stories in subtales:

Yakṣa: What is the self of a man, what is the friend made by fate,
what is the support of his life, what is his highest resort?
Yudhiṣṭhira: A son is the self of a man, a wife the friend made by
fate, rain supports his life, giving is his highest resort.63

Each question and answer can be read with such encapsulation in mind. Note
that Shulman shows how the Yakṣa’s initial skein of (eighteen) questions
suggests “a rough typology” with a narrative structure, beginning with ulti-
macy and cosmology, “including social cosmology”; moving on to cosmic
linkages through sacrifice; boundaries between “human/non-human” and
“living/dead”; alternation between the “dehumanizing” or “disanimating”
and “the human, but still with cosmological coloring”; “‘monk’s riddles’”
with “moral or ethical culmination”; “identification puzzles”; and finally rid-
dles about death (2001, 44–47, 52–53). Death and the overlapping identities
of Dharma and Yama are recurrent themes in all four phases of subtales.64
Indeed, the last subtale Yudhiṣṭhira has heard in the forest before “The
Yakṣa’s Questions” is the Sāvitrī-Upākhyāna in which Sāvitrī brings her hus-
band Satyavan back from death through her dialogue with Yama. More on
that subtale in a moment.
Unlike most or all of the other forty-eight subtales addressed to all or most
of the Pāṇḍavas, and usually Draupadī, “The Yakṣa’s Questions” is addressed
almost solely to Yudhiṣṭhira. Dharma first addresses Yudhiṣṭhira’s brothers
only separately as a disembodied voice to tell them they must answer his

63. Mbh 3.297.50–51: Yakṣa: kiṃ svid ātmā manuṣyasya kiṃ svid daivakṛtaḥ sakhā/upajīvanaṃ kiṃ svid asya
kiṃ svid asya parāyaṇam. Yudhiṣṭhira: putra ātmā manuṣyasya bhāryā daivakṛtaḥ sakhā/upajīvanaṃ ca parjanyo
dānam asya parāyaṇam (van Buitenen 1975, 802 trans.). Note that Yudhiṣṭhira overlooks Dharma’s gender cue to
name a friend in the masculine, and that he singles out dānam, giving, anticipating the yugadharma of the Kali
yuga (see chapter 6 at n. 58). The hypermetric 51c has a metrically correct variant in S with upajivyas for
upajīvanam, but that is not the case in 50c; see Shulman 2001, 44: “Wisdom, or survival” requires metric answers;
silence “would be fatal.”
64. See Hiltebeitel 2005a, 475 n. 46, citing eight subtales involving Dharma and four involving Yama,
having missed an encounter in hell with Dharma-Dharmarāja (Yama) in the Jantu-Upākhyāna (3.128.10–16).
Among these, by the end of Book 3, Yudhiṣṭhira has heard the Jantu- and the Sāvitrī-Upākhyānas. Readers also
know the Aṇīmāṇḍavya- and Pañcendra-Upākhyāna (1.189) in which Vyāsa reveals a connection between a
sacrifice of the gods in which Yama interrupts his power over death, the births of the Pāṇḍavas and Draupadī,
and the surfeit of death that will occur for the kingship of Yama-Dharma’s son in the Mahābhārata war (Hiltebeitel
2001a, 119–20; 2005a, 481–82).
438 dharma

questions before they drink from a lake, and “kills” them when they do not pay
heed. Then even after he revives them, he still addresses only Yudhiṣṭhira,
though the brothers are now listening. Later in Books 17 and 18 when Dharma
gives Yudhiṣṭhira two final “tests” (jijñāsās) that recall this one, he addresses
him alone again, since the rest have just “really” died, and at last reveals, “This
was the third test I made for you, king (eṣā tṛtīyā jijñāsā tava rājan kṛtā mayā)”
(18.3.30ab). The last two tests come when Dharma sheds his last disguise as
Yudhiṣṭhira’s sole remaining companion, a dog, and when he becomes a divine
messenger to show Yudhiṣṭhira his loved ones in hell. These two final
appearances of Dharma are parvan-closing puzzle pieces as well (see Hiltebeitel
2001a, 271–77). As tests, each jijñāsā is literally a “desire to know”: Dharma
wants to know his son, and whether his son knows D/dharma.

d.2.b. possession, exorcism, and detection. Let us now consider what it

might mean that Dharma in the form of a Yakṣa signals not only a relation to
Yama but a typical “possessing spirit” or “deity,” and that the Yakṣa’s
interrogation of Yudhiṣṭhira is one by which Yudhiṣṭhira is able to get him to
reveal his true identity as Dharma by asking a final question himself, as would
be done in an exorcism.
Frederick Smith’s recent book on possession offers a provocative contrast
regarding the Sanskrit epics. Whereas the Rāmāyaṇa seems to have scant refer-
ence to possession,65 the Mahābhārata is “the single text in South Asian literary
history with the greatest concentration of possession” (2006, 250). Smith is rea-
sonably representative in making Mahābhārata citations, but I will suggest that
he might have organized his discussion differently. Taking a lexical approach,
he finds that the Mahābhārata presents numerous instances of the three main
terms for possession in Sanskrit and Indian vernaculars: āveśa, the main term,
denoting rough and ready possession by forces (gods, spirits, demons) or emo-
tions (like wrath) and involving “a shift in personality”; graha/grahaṇa, implying
“seizure”; and praveśa, a more benign and “hazy” experience of “pervasion”
(580–81). But Smith tends to see the Mahābhārata as a palimpsest with no dis-
cernible form in which there would be no need to differentiate possessions that
carry along the main story from ones told of in subtales. First, of possessions

65. Smith says the Rāmāyaṇa seems to have largely “expunged or censored . . . popular religious praxis,”
and that the only instance of āveśa there attributes notions of possession only to “unnamed urban citizens, the hoi
polloi, . . . and at that only rumored” (2006, 278; citing Rām 2.30.10). But there are other scenes that suggest pos-
session: Rāma’s sighting of his mother as he is leaving Ayodhyā, “who almost seemed to be dancing” (2.35.32ef);
Rāma’s madness upon the loss of Sītā (3.58–62; see Pollock 1991, 55–67); Sītā’s seeming madness in captivity
(5.24.2, on which see chapter 10 § D); and the monkeys’ antics in the Honey Forest as they celebrate before telling
Rāma they have found Sītā (5.59–62). But the point can be well taken that where Rāma, his brothers, Sītā, and
other main characters are concerned, such suggestions of possession, if they are such, are not explicit.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 439

found in the main story, Smith profitably discusses the prewar possessions
(āveśa) by demons of Karṇa, Droṇa, Bhīṣma, etc., mentioned toward the end of
Book 3 (267–68; 3.240.10–34); the war-ending possession of Aśvatthāman in
Book 10 as an apocalyptic grahaṇa (271; 10.6–7); and Vidura’s life-ending pos-
session by pervasion (pra-viś) of Yudhiṣṭhira in Book 15, with both characters
being personifications of Dharma so that Dharma henceforth more fully per-
vades just his son Yudhiṣṭhira (259; 15.33.25–28). But Smith does not mention
that after the death of Irāvat (Arjuna’s son with the serpent woman Ulūpī) on
the eighth day of battle, the warriors on both sides fought on with heightened
intensity, “possessed (āviṣṭāh)̣ by Rākṣasas and Bhūtas” (Mbh 6.86.85). Second,
of possessions found in subtales, Smith discusses several fantastic cases,66 and
notably that of Nala in the subtale of Nala and Damayantī (251–55), which,
although he does not bring this out, opens retrospectives on possession in the
epic dicing scene (see Hiltebeitel 2001a, 220–46, 253, 261–62, 361). While the
episodic instances are varied and entertaining, the increasingly intensified main
sequence, which one could say is “resolved” in D/dharma’s “pervasion” of
Yudhiṣṭhira by Vidura, along with the retrospective on dicing and madness
offered by the Nalopākhyāna, now suggests to me the possibility that the
Mahābhārata could be read, given certain milieus, as a possession script, with
Dharma personified as the ultimate therapeutic agent whose tests and ques-
tions bring Yudhiṣṭhira through a transference to be able to take on that
therapeutic role of embodying dharma in the main story. Let me make two
points in favor of this interpretation before proceeding to the case in point.
First, to think of Dharma as a possessing deity may seem strange, and in
any case a departure from the way dharma is represented in the legal literature.
Yet Sudhir Kakar brings out opportunities to study the ways possession rituals
are scripted for the god Dharmarāja in Rajasthan, where Dharmarāja is one of
the three main deities, along with Bālājī-Hanumān as chief magistrate and
Bhairava, who oversee exorcisms at the Bālājī temple in Mehndipur.67 What
does Dharma have to do with possession? Dharma here is Pretarāja, King of
Ghosts, and interchangeable with Yama; the possessions seek relief in Bālājī’s
court (Urdu adālat or Hindi darbar), in which both the client-petitioner and the
possessing agent submit to legal procedures while in a trance called peśī—in
Urdu, “literally a ‘hearing’ or ‘appearance’ in court” (Smith 2006, 115). Although

66. The possession of the cannibal King Kalmāṣapāda (265–67), on whom see chapter 4 § B.1.d.i; a yogic
possession mentioned below; the story of Skanda’s link with eighteen grahas, mainly child-“seizers,” which
Smith handles superbly for its connections with Āyurvedic treatments of the same (272–75).
67. See Kakar 1982, 53–88. See Smith’s discussion, 2006, 114–19, calling this “model of possession,
exorcism, and healing . . . exceptional in South Asia” (117); 115, 119, and 160 n. 36 on his visit there and observa-
tions. Cf. the bibliography cited by Smith 2006, 628, 650. In describing the scene there, Smith discusses only
recent fieldwork and his own visit, and barely mentions Kakar’s prior discussion.
440 dharma

recent fieldwork mentions that Dharmarāja-Pretarāj, like Bhairava, is called

Bālājī’s “messenger” (dūt), Kakar does not mention such terms of subordina-
tion, and it seems, if we compare Kakar’s description (1982, 64–68, 86) to
more recent ones, that the court atmosphere now defined primarily around
Bālājī-Hanumān has somewhat shifted to him from Dharma-Pretarāj, who in
any case still holds “court” on his own (64). The association of possession with
Dharma/Dharmarāja is not something that occurs only in Rajasthan, but at
Dharmagajans in Bengal68 and at Draupadī festivals, which often take place at
temples named not after Draupadī herself but after Yudhiṣṭhira as Dharmarāja,
or just Tarumaṉ (i.e., Dharma) for short. Smith speaks of Hanumān and
Bhairava as deities he has found linked with possession in Sanskrit texts (2006,
119), but he does not say this of Dharmarāja-Yama. I will be proposing, how-
ever, that such a link can also be traced into “The Yakṣa’s Questions.”
Second, to read the Mahābhārata this way has been an Indian cultural
practice. Three Tamil cults—the Draupadī cult,69 the cult of Aravāṉ-Kūttāṇṭavar
(= Irāvat),70 and the cult of Duryodhana as the clan deity Periyantavar (see
chapter 8 § B)—have taken the Mahābhārata story itself, each as they under-
stand it, as festival possession scripts. At Draupadī/Dharmarāja festivals espe-
cially, where possession scenes are recurrent and cumulatively intensified as
the festival progresses, the arrangements for actors impersonating epic
characters to enact possession in nightlong terukkūttu dramas that spill over
into public rituals of mass possession follow the main epic narrative closely.71
Such a correlation between the main Mahābhārata story and enacted posses-
sion differs from the emphasis on oracular possession in the Pāṇḍav Līlās of
Garhwal, described by William Sax.72 In other words, these north and south
Indian cults script possession differently and script different types of possession,

68. See Korom 2004; Curley 2005; Hiltebeitel 1991, 182–207 on these three Dharmarāja cults in different
corners of India.
69. The scripting of Draupadī cult rituals highlights the figure of Pōttu Rāja, whom Smith does not
mention, despite having reviewed Biardeau 1989b (= 2004) and Hiltebeitel 1988 and 1991, which discuss him
intensively. As Biardeau has shown, one finds in Pōtu Rāja/Pōttu Rāja a link between the Vedic sacrificial post and
painful south Indian possession ordeals linked with the goddess. Smith 1994 reviewed these books.
70. See Hiltebeitel 1995; 1998a; 1999b. These three articles are overlooked by Smith, as is Hiltebeitel
1999a, which also discusses the Aravāṉ-Kūttāṇṭavar cult and wider Indian variants in its chapter 12. Aravāṉ,
whose painful self-sacrifice to Kālī involves mutilating his body in thirty-two places, is known to do this in the
Tamil Mahābhārata tradition as early as the ninth century, in Peruntevaṉar’s Pārata Veṇpā. As a prominent ritual
enacted at both Draupadī and Aravāṉ festivals, it seems to have spun off into a cult of his own where, under the
name Kūttāṇṭavar, Aravāṉ’s battlefield sacrifice or kaḷappali has local variations that include worship by Alis,
transvestites, and “eunuchs,” including castrati, who relate their sufferings, for some including their own bodily
mutilations, to his, and script their possession rituals to the festival enactment of Aravāṉ’s kaḷappali. See now
Hiltebeitel 2011b, chapters 11-14, revising the three articles just cited to tie in with a new chapter 14.
71. This point was underscored by Frasca 1990, on which see Smith 2006, 76.
72. See Sax 2002; Smith 2006, 76–77, 412 n. 77, mentioning the two cults. I believe, however, that it
would be worth asking whether a sense of dharma affects Pāṇḍāv Līlā possessions.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 441

yet all of them do so as readings of the Mahābhārata. I believe we have support

here for the cultural reading practice I am proposing, but what kind of reading
is it? Let me recall a summarizing statement I made on the way Draupadī cult
terukkūttu dramas focus in on epic characters who are in the very thick of the
epic’s kin-defined main family feud:

. . . that is what the Mahābhārata has become in village eyes that view
the Terukkūttu: a royal family feud among paṅkāḷis over the rights of
inheritance. It dwells on the same issues—a paṅkāḷi kāccal, or “fever
besetting a group of male descendants”—that Brenda Beck (1982,
174) has traced through another creation of the Tamil “folk epic”
tradition, the Elder Brothers Story. (Hiltebeitel 1988, 398)

Indeed, the Elder Brothers Story involves an oral epic reenplotment of the
Mahābhārata.73
Let us mark this family orientation. Significantly for Tamil epic heroines,
paṅkāḷis are “share-holders” of a “family’s” holdings: the men in the family a
woman has married into, her male in-laws (Hiltebeitel 1988, 9, 306). Smith
twice mentions the exorcistic practice of nailing a possessed client’s (usually a
woman’s) hair to a tree to fix the spirit there, one case in Varanasi, the other in
Kerala (2006, 528, 546). This also occurs in exorcisms detailed by Isabel
Nabokov in northern Tamilnadu (2000)—as it does in the Duryodhana-
Periyantavar cult, where most of the exorcisms by “Duryodhana” are done for
women deemed possessed by others in their families, typically for their irreg-
ular behavior as daughters or wives. Smith is cautious on Nabokov’s stance
(2000, 15) that such rituals do not reintegrate participants but rather fragment
them “to the point of splitting them apart” (quoted in Smith 2006, 75). “The
Yakṣa’s Questions” qualifies on both grounds. Dharma both fractures his son’s
world and enables him to reintegrate it—at least as regards his brothers,
Draupadī’s most immediate paṅkāḷis.
Now as we begin to consider “The Yakṣa’s Questions,” I believe we can
look at Yudhiṣṭhira not only as a kind of exorcist in raising questions of the
Yakṣa but, like any exorcist, as a kind of detective in getting the possessing
spirit to reveal his true identity. In doing so, we must question Shulman’s view
that the Dharma who imposes these tests is bringing about “evil” (Shulman
2001, 41–42) or a “blatant injustice evident in the reality” (58)—which always
turns out to be illusory. We will also have to reconsider his portrait of Yudhiṣṭhira
in this episode as unduly morbid and unknowing, caught up in a world “already
coming to an end” though “he does not know this,” with an identity “still

73. See Hiltebeitel 1999a, 45–46 and passim on reenplotments of various types of intensified family dramas.
442 dharma

hidden from him” (53), and with a “conceptual plan” that will eventually have
“entirely collapsed” (57) by the time he encounters his father again those two
final times. For one thing, Yudhiṣṭhira’s identity as Dharma’s son is hardly
hidden from him, and has indeed been public knowledge since he was a boy
when the great Ṛṣis came down from the Hundred Peak Mountain to announce
him at Hāstinapura (see chapter 8 § H). For another, if, as Shulman seems to
suggest, Yudhiṣṭhira’s skill in making “countermoves” to the Yakṣa’s ques-
tions is reminiscent of the epic’s dice game (44), it would be pertinent that he
has learned the “heart of the dice” before this encounter. Granted, however,
that “The Yakṣa’s Questions” engage Yudhiṣṭhira in “a cunning and potentially
deadly game” (44); that he is often enough “a tormented and embattled figure”
(58); that his three confrontations with his father Dharma “disguise an identity
carrying a powerful destructive charge, which the self-identifying subject
resists” (62); and that “[i]n such a world, one mostly fights for time” (40).

d.2.c. detecting dharma. So Yudhiṣṭhira is finally left to his own resources

in answering the Yakṣa’s questions. His four younger brothers have succumbed
to some enchantment at a lake, and Yudhiṣṭhira arrives to face “a strange, one-
eyed, fiery creature standing on a log beside the water” (Shulman 2001, 41–42).
Actually, it is more likely that the Yakṣa’s eyes are in some way “asymmetrical”
(virūpākṣa; 3.297.20a), so it may not be “one-eyed” like Kabandha.74 Also, rather
suitably, considering the episode’s complex boundary symbolism, the Critical
Edition has the Yakṣa standing on a “dam” (setu; 3.97.2a) rather than the “log”
or “tree” (vṛkṣa) that Shulman finds in the Vulgate.75 More important, however,
a good deal has gone on before this that bears on how Yudhiṣṭhira approaches
this unusual father.
I begin from Janamejeya’s leading question, already noticed, as to what the
Pānḍ ạ vas did next after they got Draupadī back from being abducted. The
assumption cannot be tested, but it is probably a fair guess that Janamejaya
already knows the basic plot when he asks this question.76 Vaiśaṃpāyana
supplies narrative voiceover frequently in “The Yakṣa’s Questions,” and even
though Janamejaya never asks him anything further during the episode, these
voiceovers are a constant reminder not only that the story is addressed to him,
but a sustained set of cues as to how he might figure out, if he is listening care-
fully, what is really going on. Janamejaya, and other careful listeners, will thus

74. Cf. Biardeau 2002, 1: 251 and n. 33, where this name describes the three eyes of Śiva. Cf. Hinich
Sutherland 1991, 90: “The Yakṣa had an enormous body like the elephant Virupakṣa who holds up the surface
of the earth.”
75. Toward the end of the episode, Yudhiṣṭhira asks who the Yakṣa really is “standing on one foot in the
water (sarasy ekena pādena tiṣṭantam . . .)” (3.298.2ab).
76. See Janamejaya’s early question in the epic, cited above at n. 16.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 443

̣ ạ vas, including Yudhiṣṭhira.

always be at least a step or two in front of the Pānd
For instance, Vaiśaṃpāyana uses the phrase Yakṣa uvāca, “The Yakṣa said,” sev-
eral times (3.296.30–37) before he tells Janamejaya point blank that he has been
describing a Yakṣa giving warnings to Arjuna and Bhīma (38), and all this comes
well before Yudhiṣṭhira learns that he is talking to a Yakṣa on a dam (297.18).
But the important matter is that Yudhiṣṭhira is never far behind Janamejaya
(and other careful listeners) in putting cues and clues together. As with
Janamejaya from the beginning, there is the possibility by the end that
Yudhiṣṭhira has all along had an inkling of what is going on. Let us see how this
comes together.
While our first listener Janamejaya might wonder from the start why
Yudhiṣṭhira should tell Nakula to look in all ten directions for signs of water,77
more reliable cues start cropping up in descriptions of the brothers’ “deaths”
after they have drunk from the lake without heeding the Yakṣa’s warnings to
first answer his questions. When Vaiśaṃpāyana describes these scenes, it is
always with the perfect of ni-÷pat, “he fell down” (nipāpata; 296.14, 20, 31, 38),
which suggests that the brothers have “fallen down dead,”78 but actually says no
more than that they have “fallen down” or “collapsed.” On the contrary, when
the brothers arrive one after another to find the ones who preceded them on the
ground, it becomes a matter of what they perceive, and only there does one
hear that they see the others as “slain” (÷han). In the first two cases, Sahadeva
and Arjuna “saw” their brothers “dead” or “slain” (dadarśa . . . hatam/au;
296.17, 22). With Bhīma, Vaiśaṃpāyana replays the note of uncertainty,
describing not what Bhīma saw but that “he went to the same spot to where his
brothers had been felled” or “made to collapse” (nipātitāh;̣ 34). Finally with
Yudhiṣṭhira it is most intense, when, upon first arrival, “he saw his brothers
slain (dadarśa hatān), as weighty as Śakra, fallen like the World Guardians at
the end of a yuga” (yugānte; 3.297.1). But that is only his first impression.79
Moreover, if he will be the only one to approach the scene as an exorcist, he is
not the first to approach it as a detective. Arjuna and Bhīma also did some pre-
liminary sleuthing before succumbing to their thirst. Arjuna was the only
Pāṇḍava to bring weapons—in his case, his bow and arrows—to the lake, and

77. Mbh 3.296.5. Hinich Sutherland raises a good point about the Devadhamma Jātaka that can be consid-
ered in parallel. Though the senior brother, the Bodhisatta Prince, “tells the yakkha that the Sun Prince’s welfare
is his primary concern, the fact is that he has sent him (even before his blood brother) into danger” (1991, 96). The
Sun Prince is the Nakula of the piece: the younger half-brother of another mother whose life he will choose for the
Yakkha to spare rather than his blood brother, the Moon Prince. One has to wonder whether these names reflect
the Solar and Lunar Dynasties.
78. As Shulman says; cf. Hinich Sutherland 1991, 89: “drops dead.”
79. Yudhiṣṭhira’s reference to a yugānta here could be not only intentionally “mere” (see González-Reimann
2002) as a metaphor but also condense a correspondence between the four fallen brothers and the four yugas, and
thus further, with respect to Yudhiṣṭhira himself, a confirmation that the king will have to make a new yuga.
444 dharma

at least learned, before drinking, that he could not get the disembodied voice to
materialize by challenging it to do so, or hit it by filling the air with arrows or
shooting at the sound (296.21–31). And Bhīma actually paused before drinking
with the thought that his brothers’ fall was “the work of Yakṣas or Rākṣasas”
and that drinking the water would help him put up a good fight (35–36). All this
lets Janamejeya know that these two brothers were not entirely clueless, and
that Yudhiṣṭhira will at least not make the same mistake as Arjuna.80
It is now Yudhiṣṭhira’s turn, and our time to see how he gets past his first
impressions. But let us first appreciate the powerful way that Vaiśaṃpāyana
describes how he sets out after his missing brothers:

Thereupon Kuntī’s son the king, a bull among men, began

pondering (vicintya), and the strong-armed man rose up
(samutthāya) with his mind on fire. (3.296.39)

Here we meet another formation from the same verb sam-ut+÷sthā that occurs
in the verse describing the blessings of reciting the episode, calling it “The
Great Rising Up (samutthāna) and Meeting of Father and Son” (298.27).
It would seem that this titular usage would refer back, above all, to this very
moment of Yudhiṣṭhira’s “getting up” to meet his father rather than the
“revival” of his brothers, as others have taken it,81 for which no formation of the
verb sam-ut+÷sthā is ever used, although by that point all these arisings could
also be referred to together. This verb has at least four other very charged usages
for Yudhiṣṭhira’s getting up as a king: he “gets up” to go tell Draupadī she has
been won by the Kauravas in the dicing (2.60.19); Draupadī uses this term to
challenge him to “exertion” during their first exchange in the forest (3.33.6–7,
53; see chapter 10 § D); he has not “gotten up” but has been “lying on Draupadī’s
bed” (8.49.83) wounded while everyone else is off fighting when he insults
Arjuna for taking so long to kill Karṇa (8.50.29); but he does “get up,” expect-
ing to hear that Karṇa is finally slain, when he sees Arjuna and Kṛṣṇa approach-
ing him again (8.69.11).82 As we saw in chapter 5, it is also a repeated trope in
the postwar anthologies and even in the Mānavadharmaśāstra that much
depends on a king’s ability to ponder or think for himself (cint, vi+÷cint,
sam+÷cint) and to do so on his feet.

80. In coming armed, Arjuna is like the Bodhisatta prince Mahiṃsāsa in the Devadhamma Jātaka (Cowell
[1895] 2005, 26), who, unlike his two younger brothers, brings his sword and bow. He is thus like Arjuna in
coming armed with weapons and like Yudhiṣṭhira in coming with the wisdom to answer the Yakkha’s one
persistent question.
81. Cf. van Buitenen, 1975, 804 and Shulman 2001, 48 n. 6, both with “revival”; Johnson 2005, 329 with
the even more unlikely “recovery.”
82. See Hiltebeitel 2007b, 57, 73 n. 216. The term is also used to describe Draupadī when she rises up
from the earthen altar (vedi) at her birth (1.55.41; 5.80.21).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 445

Immediately after shedding tears of grief at seeing his brothers “dead”

(297.1–2), Yudhiṣṭhira starts to wonder or ponder (vi+÷cint).83 Although using
the verbal root han, he is now the first Pāṇḍava to use the verb ni+÷pat by which
Vaiśaṃpāyana has been cuing Janamejaya that the four younger brothers are
not really dead but “collapsed” or “fallen”:

With his intellect he pondered (buddhyā vicintayām āsa), “By whom

have the heroes been made to fall (nipātitāḥ)? There is no mark of a
weapon on them, no trace (or footprint, padam) of anyone here.
I think this is a great being (bhūtaṃ mahat) by whom my brothers
are slain (hatāḥ).” (3.297.3c–4d)

Unlike the four others, each overwhelmed by thirst, Yudhiṣṭhira starts using his
“wits” or “intellect” (buddhi) like a sleuth who has arrived at a possible crime scene,
but one in which he is beginning to see the work not of just of a Yakṣa but a “great
Bhūta,” an even more familiar term for a possessing spirit (although it could also
refer to the water as one of the five “great elements”). Yudhiṣṭhira now focuses his
pondering buddhi “attentively” or “one-pointedly” (ekāgram), using a popular yoga
terminology with which both epics are familiar,84 on the matter at hand:

I will ponder this attentively (ekāgraṃ cintayiṣyāmi)—or, having

drunk the water, I will know (pitvā vetsyāmi vā jalam). (297.4ef)

Note the carefully constructed disjunctive thought by which he delays the

thirst-driven alternative. Vaiśāṃpāyana now tells how Yudhiṣṭhira first “con-
sidered various possibilities (bahudhā samacintayat)”—that the “crooked-
minded” (jihmabuddhi) Duryodhana could have concocted all this with Śakuni
or carried it out by hidden henchmen (5–7)—and then again returns to
Yudhiṣṭhira’s direct thoughts just before Yudhiṣṭhira wades into the water:

“This, its (the lake’s) water, was not fouled by poison since my
brothers’ facial color is clear,” he pondered (acintayat). “Who else but
Yama, the Finisher of Time, could match, one by one, these best of
men who had the power of a flood?” With this apprehension
(adhyavasāyena), he plunged into the water, and as he was immersing
himself he heard (a voice) from the sky. (297.8–10)

83. Hinich Sutherland 1991, 89 notices Yudhiṣṭhira’s “wonder” at his brothers “bloodless demise,” but
makes nothing of it.
84. This is the only epic verse to combine eka-agra with Öcint, not to mention buddhi. For “proto”-systematic
Yoga usages of eka-agra, see Mbh 12.228.36, 272.33; 14.19.34 and 50; 14.30.28; and, among many narrative usages
implying this sense, 11.12.13; Rām 4.51.1. Ekāgratā, “one-pointed concentration,” like buddhi, becomes a technical
term in the classical Yoga system. Also carrying along this yogic interlude is the usage of adhyavasāya at 3.297.10a
to describe how Yudhiṣṭhira rounds off his thought at this point.
446 dharma

I translate the unusual usage of adhyavasāya here as “apprehension” primarily in

a philosophical sense. That is its only sense elsewhere in the epic, where it appears
in the Mokṣadharma or the Anugītā, and always describes what the buddhi can
“apprehend,” “ascertain,” or “decide” beyond the agitation of the mind.85 But it
also works doubly to convey Yudhiṣṭhira’s apprehensive understanding of his
situation. As he steps into the water, he is, to be sure, still thinking his brothers
look like they are “slain,” but now he has the more specific “apprehension” that
they may have been overcome by “Yama, the Finisher of Time.”86 The Yama whom
Yudhiṣṭhira apprehends here should be the one he has most recently heard about
in the Sāvitrī-Upākhyāna. That Yama is one who not only doubles there for
Yudhiṣṭhira’s father Dharma but returns Sāvitrī’s husband Satyavan from a not
quite complete death. Indeed, Yudhiṣṭhira has just heard how Yama released
Satyavan because of Sāvitrī’s extraordinary fidelity, and also because she answered
Yama’s questions. Yama had taken away Satyavan’s soul. That too was a posses-
sion, but not an exorcism because Sāvitrī knew whom she was talking to.
The celestial voice now says it belongs to a crane (baka)—perhaps another
lexeme of possession. As Dan Rudmann reminded me,87 one of the ten Tantric
Mahāvidyās is the “Crane-faced” Bagalāmukhī, who has the ability to paralyze
(stambhana), particularly in arresting the power of speech. With the crane’s
yogī-like concentration, it can use this trait to immobilize fish.88 Yudhiṣṭhira
will take the name Kaṅka, “Heron,” for the disguise he will assume in the
“Fish”-kingdom of Matsya. This episode provides the immediate antecedent of
his choice of that name.
Now it seems Yudhiṣṭhira never sees this crane which Vaiśaṃpāyana has
already been calling a Yakṣa. Indeed, it is interesting that Yudhiṣṭhira may not see
the crane, whose invisibility would be an added pretext for his choosing the name
“Heron” to disguise himself. On the other hand, if Janamejaya’s memory is acute,
he could recall that near the beginning of the Pānd ̣ ạ v̄ as’ forest exile, in a scene we
shall examine in chapter 10, Draupadī emphasized the visibility of a crane to make

85. Among these other usages, Yudhiṣṭhira hears what Vyāsa told Śuka: “The mind sets loose mental expe-
rience, the intellect is the Apprehender (manas prasṛjate bhāvaṃ buddhir adhyavasāyinī)” (12.240.1); Bhīṣma also
tells Yudhiṣṭhira, “The eye is for perceiving; so the mind produces doubt. The intellect is for apprehending (bud-
dhir adhyavasāyāya); the field-knower exists as witness” (12.187.12); cf. 14.43.33, beginning the same way.
86. He does not at this point say they are “dead” (mṛta-), as does a line interpolated here that spoils the
effects: “Even though they are dead it did not cause any disfigurement (mṛtānām api caitāṣāṃ vikṛtaṃ naiva jāyate)”
(Vlg 3.313.26ab). Like another usage of mṛta- (cited above) where the Yakṣa asks Yudhiṣṭhira his four superfluous
closing questions, it shows the inattentive hand of this particular interpolator.
87. Presently a doctoral candidate in South Asian studies at Texas-Austin, Rudmann was then my student.
88. Kinsley 1997, 47, 197, and 272 n. 13. Bagalāmukhī (bagala = baka) may also have a crane’s head or
beak, or ride one as her vehicle (193–208). She can be depicted as sitting on a corpse or preta, probably evoking a
śava-sādhanā (201–206), wherein at some “points it seems as if one is trying to control the corpse or the spirit
that may inhabit it,” which “reminds one of possession cults,” with “the possibility of the corpse reviving or
becoming aggressive (204–205). In one myth, she “sported in” a “pond of turmeric” (193).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 447

the point to Yudhiṣṭhira that all creatures live off their own “total effort,” using the
same term, samutthānam (3.33.6–7) that, as just noted, comes at the end of this
episode to describe Yudhiṣṭhira’s great “rising up” to encounter his father. But
whether Yudhiṣṭhira sees this crane now or ever, he does not believe the voice
comes from a crane. Here we see Yudhiṣṭhira not only processing information
but beginning to raise questions, which we will have to examine. And we see the
possessing “Bhūta” doing a typical trick of not admitting his real identity. And
shortly, after Yudhiṣṭhira has voiced doubt and raised these questions, the voice
changes its tune to identify itself as belonging to a Yakṣa, which Yudhiṣṭhira is
finally able to confirm “having approached and stood while the Yakṣa was
speaking” (yakṣasya bruvato . . . upakramya tadā sthitaḥ; 3.297.19cd). This prodigy
now looms before him odd-eyed, fiery and tall as a palm tree. One would think
that Yudhiṣṭhira has gotten out of the water, but he could also be making an
aquatic approach, standing like a crane or heron himself ! In any case, the Yakṣa’s
pseudo-crane voice has led him on to its Yakṣa apparition. What has happened in
this interval that Yudhiṣṭhira should have gotten this revised information?
It seems that the Yakṣa first poses his apparently invisible and in any case
unreal “crane” identity, which he has not mentioned to any of the other brothers,
as a kind of opening riddle intended solely for Yudhiṣṭhira, and linked to what
he says along with it:

I am a crane (baka) living on duckweed and fish. Your younger

brothers have been brought by me under the sway of the departed
(nītāḥ pretavaśam). You will be the fifth, son of a king (rājaputra),
if you do not reply to the questions I ask. Commit no violence, lad,
I have a prior claim. But answer my questions, Kaunteya, then you
can drink and carry. (3.297.11–12)

If Yudhiṣṭhira has been picking up clues, there are more here to consider in the
voice’s somewhat ambiguous description of the condition of his brothers and
the term by which it first addresses him. If the brothers have been “brought
under the sway of the departed,” they would have been led to the realm of Yama
Pretarāja, “King of the departed” (preta means “ghosts,” “the dead”); and if the
prodigy calls Yudhiṣṭhira “son of a king,” it could refer to his being the son of
King Pāṇḍu or that of King Yama Dharmarāja, that is, Yama as he doubles for
Dharma (or vice versa). Otherwise, except for variations in the ways the voice
addresses each brother after calling him “lad” (tāta), it ends with the very same
words to Yudhiṣṭhira that it spoke to all but Arjuna,89 posing another apparent

89. Whose attempt at a frontal assault led to different phrasing; he says to Arjuna, “Why did you come
near? You cannot take this water by force. If you answer my questions, Kaunteya, then you shall drink and fetch
the water, Bhārata!” (3.297.25c–26).
448 dharma

riddle about “my prior claim” (mama pūrvaparigrahaḥ). When Yudhiṣṭhira

agrees to answer the Yakṣa’s questions before drinking, he will say explicitly
that he does not covet the Yakṣa’s “prior claim” (297.24).90
Let us note that Yudhiṣṭhira also does not challenge that this being holds
his brothers under the sway of death.91 Rather, in challenging its identity as a
crane, Yudhiṣṭhira makes several of the clues he has so far pondered relevant
to his continuing questions:

I ask you, who are you, a God (pṛcchāmi ko bhavān devo)? This was
not done by a śakuni! (3.297.13cd)

Yudhiṣṭhira is also riddling.92 He seems to expect the so-called “crane” to appre-

ciate that he is ruling out two possibilities: that a śakuni or “bird” (typically a
bird of ill omen) could be behind this disaster; and that his nemesis Śakuni
could any longer be considered a suspect now that this “crane’s” voice has
spoken. Also, having already noted that his brothers still have a good com-
plexion, he does not grant that they are dead but only credits that the perpe-
trator must be great and strong to have “made the four mountain-like brothers
fall on the earth (pātitā bhuvi)” (297.14–15). Finally, it seems that he has begun
to see the form of this great being when he builds up to one further question
that gets the voice to reveal that it belongs not to a crane but to a Yakṣa standing
on the dam before him:

I do not know what you are doing. I do not fathom your intention
(kāṅkṣitam). Great curiosity is aroused and my alarm (sādhvasam)
comes. By what is my heart anguished and fever brought to my head?
Therefore I ask (pṛcchāmi), lord: Who are you, standing here?
(3.297.16–17)

The desiderative kānk̇ ṣitam, “intention,” sets a precedent for Dharma to answer
later that he “desired to see” Yudhiṣṭhira, and further that he came out of a “desire
to know” or “test” him (298.6, 10 and 13, with didṛkṣur and jijñāsur/jijñāsā).

90. Shulman’s translation. There is an illuminating prior usage much earlier in the Forest Book in a sim-
ilar initiatory scene when a disguised Śiva lays prior claim to a boar he and Arjuna had shot simultaneously.
According to Arjuna’s description of his quest for the Pāśupata weapon from Śiva, when he and Śiva, who had
disguised himself as a low caste forest hunter (kirāta), had simultaneously shot a boar, the hunter used this exact
phrase, “my prior claim” (mama pūrvaparigrahaḥ; 3.163.22; cf. 40.22), to argue that the boar was his since he was
the first to have aimed at it. Arjuna then challenges the hunter’s claim and the two wrestle, before Śiva reveals his
divinity and bestows the boon Arjuna came for.
91. The Yakṣa uses strong but still ambiguous terms such as nihatāḥ, “struck down,” “slain” (3.297.18d),
sūditāḥ, “finished off” (297.22d, for which Vlg. 3.313.41d has mṛditāḥ, “crushed”).
92. See Hiltebeitel [1976] 1990, 247–48 on Yudhiṣṭhira’s ability to discern riddles coming from such
other dharma authorities as Bhīṣma and Vidura.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 449

Yudhiṣṭhira has thus gotten Dharma to reveal the penultimate of his serial
self-disclosures, that of a Yakṣa about to test him with questions, by question-
ing Dharma himself, as he will do twice more: first, as noted, after he has
answered Dharma’s eighteenth question, the last in the Yakṣa’s main series;
and finally, once again after the Yakṣa has revived his brothers. For now, the
Yakṣa lets Yudhiṣṭhira know, “You have answered my questions in accordance
with reality (yāthātathyam)” (297.62a; Shulman 2001, 54), and asks one more
question himself, “But now tell me, who is a man, and what man owns all
wealth?” (62cd), to which Yudhiṣṭḥira replies,

Word (śabdaḥ; sound, repute) of a good deed touches heaven and

earth; one is called a man as long as that word lasts. And the man for
whom the dear and the undear are the same, as also happiness and
suffering, and both past and future, owns all wealth. (63–64)

This question is marked off from the earlier eighteen not only by the preceding
turnabout. It differs in the one-line and two-topic brevity with which it is asked,
in the relaxed two-verse reply by which Yudhiṣṭhira answers it, and in the
further conversational tone that follows. But above all, as the Yakṣa’s follow-up
makes clear, this briefer question differs in having been a leading question:

You have told of man, king, and of the man who has all wealth.
Therefore, let one of your brothers live, he whom you choose. (297.65)

When Yudhiṣṭhira chooses Nakula, the Yakṣa seems surprised and asks at
length (66–70) why he did not choose Bhīma or Arjuna, on whom, he points
out, so much depends; also Nakula, as he says twice, is only a half-brother, a
sāpatna: literally, “one born to a rival” or “cowife.” That is, Yudhiṣṭhira has
chosen a son of Mādrī who was a rival of his own (and Bhīma and Arjuna’s)
mother Kuntī. This looks like another leading question since it anticipates
Yudhiṣṭhira’s reply, which is:

Noncruelty is the highest dharma (ānṛśaṃsya parodharmaḥ), and

to my mind higher than the final goal (paramārthāc ca me matam).
I seek to do a noncruelty (ānṛśaṃsyaṃ cikīrṣāmi). Let Nakula live,
Yakṣa. “The king always has the character of dharma (dharmaśilaḥ
sadā rājā iti),” so people know of me. I will not stray from my own
particular dharmas (svadharmān na caliṣyāmi). Let Nakula live,
Yakṣa. As is Kuntī so is Mādrī: for me there is no distinction
between them. For the two mothers I want the same. Let Nakula
live, Yakṣa. (297.71–73)

This gets us to this king’s “chosen ideal.”

450 dharma

d.2.d. noncruelty. This climactic question gives Yudhiṣṭhira the opportunity

to answer for once what he considers to be a king’s svadharmas, plural.
He asserts that for a king known for having “the character of dharma,” the
highest dharma is noncruelty, a virtue we have met before (see n. 50 above) and
will meet again in this chapter. He now exemplifies it by making no distinction
between his two mothers. Minimally, he sees that a king’s dharma begins with
what he does at home. But he could also be amplifying on something he has
said in reply to the Yakṣa’s fifteenth question, which could anticipate the last
two questions that concern rites for the dead:

Yakṣa: What is the highest dharma in the world, what dharma always
bears fruit, what when restrained does not bring grief, with whom
[plural] does the bond never wear away?
Yudhiṣṭhira: Noncruelty is the highest dharma, Vedic dharma
(trayīdharmaḥ) always bears fruit, the mind does not bring grief when
restrained, the bond with the good [plural] is never worn away
(sadbhiḥ saṃdhir na jīryate).

Indeed, this answer reinforces our sense that “The Yakṣa’s Questions” are a
subtale clearing house, for Yudhiṣṭhira has heard this virtue exalted as the
highest dharma quite memorably in two of the most prominent of Forest Book
subtales, the Nala-Upākhyāna (3.67.15) and the Pativratā-Upākhyāna (203.41;
cf. 206.3). And he has also heard it recommended in Book 3 by his ancestor
Nahuṣa (177.18), who was cursed to become a boa constrictor, and got to
question Yudhiṣṭhira, somewhat like Dharma, before he would release the
nearly strangled Bhīma.93 As Shulman acknowledges, this is “the most mean-
ingful set” of questions and answers among questions ten to fifteen that
opened a “moral direction.” But in translating ānṛśaṃsya as “non-injury” and
seeing Yudhiṣṭhira’s “chosen ideal” as undercut, Shulman downplays the
dharma “level” of the riddle exchange: “Once the moral direction has opened
up, there is no stopping it from this point.”94 I believe, however, that this is not
the time for dharma fatigue. Question-and-answer fifteen point directly to rites
for the dead in the connection it suggests between noncruelty and “the bond
with the good that is never worn away”—a connection that Yudhiṣṭhira’s

93. See Hiltebeitel 2001a, 209 on these passages in relation to “The Yakṣa’s Questions,” along with 205–6
and 230–31.
94. Shulman 2001, 47. He continues: the best that can be said is that he “is summing up his life’s wisdom”
with a response he “will repeat . . . later,” and that the Yakṣa is “satisfied.” Shulman is referring to Dharma’s final
tests of Yudhiṣṭhira, where this chosen ideal is again tested that one last time, again in confronting death.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 451

choice of Nakula then puts into practice. His choice thus not only makes no
distinction between the two mothers but between past and future and life and
death, since, unlike his own mother, Nakula’s mother is dead. Like Pāṇḍu,
and for the moment somewhat like Nakula, she is a preta in the realm of Yama.
The Yakṣa now not only endorses Yudhiṣṭhira’s answer but, also mentioning
noncruelty himself, exemplifies it as well by bringing all four of his brothers
back to life (297.74). This exorcistic conversation has rectified a seeming
rupture between past and future and the living and the dead.
Yudhiṣṭhira now presses on with his final questions to the Yakṣa, quickly
including the same quarter-verse question with which he stated his doubt about
the “bird” to state the same doubt that he is dealing with a Yakṣa:

I ask you, who are you, a God (pṛcchāmi ko bhavān devo)? I do not
think you are a Yakṣa. . . . Surely these brothers of mine are fighters
of hundreds of thousands. I do not perceive the means (yogam) by
which they could be brought down (vinipātitam).95 Now that they are
sweetly awakened, I notice their sense faculties (have returned). Are
you our wellwisher, or are you even our father (sa bhavān suhṛd
asmākam atha vā naḥ pitā bhavān)? (3.298.2cd, 4–5)

Note the plural. Dharma is part of the whole family. Yudhiṣṭhira has finally put
the question to him that we may suspect he had in mind all along. Indeed, the
Yakṣa sequence itself began with Nakula thirsty and impatient, asking why
they suffer in this disaster, and Yudhiṣṭhira answering, “There is no limit to
misfortune, and neither its reason nor its cause can be ascertained. It is Dharma
who here distributes the fruits of both virtue and vice” (3.296.1)!96 Yudhiṣṭhira’s
question now brings the satisfying answer:

I am Dharma, lad, your begetter, O man of mild prowess. Know, bull

of the Bhāratas, that I have come out of a desire to see you (tvāṃ
didṛkṣur). (298.6)

Indeed, the framing of the questions by which Yudhiṣṭhira smokes out Dharma
suggests that he knows only all too well the darker side of this divine father,
who would be the very god to have the means (yoga) to bring his brothers under
the sway of death.97 Yet the extraordinary thing by this time is that, with his
brothers now restored, Yudhiṣṭhira can now ask, “Are you our wellwisher,”

95. The Critical Edition brings out a much more interesting line here than the Vulgate’s, which has: “I do
not perceive the warrior by whom they were all made to collapse (taṃ yodhaṃ na prapaśyāmi yena sarve nipātitāḥ)”
(Vlg. 3.313.4cd).
96. See Klaes 1975, 79. Van Buitenen 1975, 797 has “the Law distributes. . . .”
97. When Dharma sired Yudhiṣṭhira, he took a yogamūrti (1.114.3); see Hiltebeitel 2001a, 188.
452 dharma

“our friend” (suhṛd)? Yudhiṣṭhira and Dharma both play at the edge of death
but there is no crime. The perpetrator not only lays the clues that will lead to his
detection; he has come to be found out.
Once Dharma has given the boons that the Brahmin can have his firesticks
back (which the Pāṇḍavas will have to return to him) and that the Pāṇḍavas will
be able to pass their thirteenth year incognito, he presses Yudhiṣṭhira to choose
a third boon and Yudhiṣṭhira asks, “May I ever conquer greed, delusion, and
anger, and may my mind be always on giving, austerity, and truth” (3.298.23).
To this, Dharma replies,

Endowed with every virtue, Pāṇḍava, you your honor are Dharma by
your own nature, and again it will be as you have said (upapanno
guṇaiḥ sarvaiḥ svabhāvenāsi pāṇḍava/bhavān dharmaḥ punaś caiva
yathoktaṃ te bhaviṣyati). (298.24)

Dharma takes advantage of the Upaniṣadic notion that the son replicates the
father to identify Yudhiṣṭhira by his own name, Dharma.98 Having done so,
he vanishes in the next verse, leaving Yudhiṣṭhira not only with a confirma-
tion of their “inherent” (svabhāva) identity but with the promise that “it will
be” his son’s future to work out what it means to “be Dharma.” For the pre-
sent, Yudhiṣṭhira chooses a complement of virtues that is inherently that of
Dharma, encapsulating the “chosen ideal” of noncruelty they seem to share
in the triple set of giving, austerity, and truth. To conclude with the obvious:
if Dharma and Yudhiṣṭhira both embody the ideal of noncruelty, it would
have a bearing on how Dharma made the younger Pāṇḍavas all “fall down.”
What is striking in this set of values, which we shall be tracking further,
is that they are validated in a royal family: one which, as it hurtles toward a
cataclysmic war, has been all but dysfunctional, and increasingly so, for four
generations (see chapter 8). On this point we may offer a hypothesis that
connects the main topics of this section. If the Mahābhārata bears out
cultural readings as a possession script with D/dharma as the ultimate
therapeutic agent, we may propose that in this transformative episode,
the father–son questions and answers refract narrated life experiences to
bring Yudhiṣṭhira through a transference that will enable him to conduct
the therapeutic role of embodying dharma through the traumas to come.
It has tested this king not only so that he can reach a new self-understanding
but prepared him to keep dharma at the center of an ongoing family

98. As Tamil Tarumar, this is Yudhiṣṭhira’s ordinary name at Draupadī and Dharmarāja temples and
festivals in Tamilnadu. This is more than just confirming that the boon is “tautological or tautidentical” because
“Yudhiṣṭhira character is in any case exactly as he asks that it be” (Shulman 2001, 48).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 453

crisis. Considering the likely post-Aśokan ambitions of this text, we would

have here the family therapy of a world, civilization, or nation99 (see chapter
1 § B).
Finally, as his memorial to Mādrī shows, Yudhiṣṭhira still keeps to a law of
the mothers, whereas being “law-abiding” in the Rāmāyaṇa has to do only with
Rāma’s fathers.

E. Questionable Killings: Vālin and Droṇa

Once Rāma and Yudhiṣṭhira have exited the forests, they begin to take a more
“pragmatic” view of the world (as Das 2009, 78 says this of Yudhiṣṭhira).
Rāma must make compromises to get his wife back, and Yudhiṣṭhira must
adjust his predilections for noncruelty and peace to take charge of a war effort
that will require his trust in the inscrutable counsel of Kṛṣṇa.100 If their moral
biographies prepare them to face their lives’ calamities, and prepare readers
to understand their responses as exemplifying dharma with the possibility
also of violating it, each does so with some complexity, and with differences
in accent. One way into this complexity is to examine the way they handle
situations where their dharma is questioned. A difference to note immedi-
ately is that whereas Yudhiṣṭhira is among those to question his own dharma,
even to the extent of confessing sins, Rāma never does this. Yudhiṣṭhira is
prone to moral self-scrutiny of a type that we can imagine of Rāma—if his
split personality allowed it—only after he starts hearing his adventure sung
by his sons. Rather than self-exoneration, Yudhiṣṭhira confesses his part in
killing Droṇa as one of four sins that most bother him after the Mahābhārata
war (Mbh 12.27.4–22).
After lamenting his part in bringing about the fall of Bhīṣma, Yudhiṣṭhira
admits that, when Droṇa expected the truth from him as to whether his
son Aśvatthāman was slain, “I acted falsely by saying ‘elephant’ under my
breath,” and was “a kingship-coveting sinful guru-slayer” who “resorted to the
armor-skin of truth (satyakañcukam āsthāya)” (15–17).101 He expects karmic

99. Regarding the proto-national scope of the Mahābhārata, see Hiltebeitel forthcoming-d, on its naming
of India (Bhāratavarṣa) after Śakuntalā’s son Bharata. Contrary to van der Veer 1999, the editors of both epics’
critical editions had fair reason to vie over which was India’s “national epic.”
100. See Das 2008, 78–79, citing Mbh 5.28 as marking this pragmatic turn but leaving Kṛṣṇa out of the
equation, which Yudhiṣṭhira does not do (see verses 9–14).
101. For reasons of continuity with the discussion in text, I use Ganeri’s translation “armor-skin of truth”
for satyakañcukam (Ganeri 2007, 72). Fitzgerald 2004a, 223 very nicely has Yudhhiṣṭhira say, “I put a little jacket
on the truth” (2004a, 221 and 702, with explanation).
454 dharma

consequences, asking what worlds he will now obtain due to that “terrible act”
(karma daruṇam; 17ef)—for which Dharma will show him a karmic hell at the
epic’s closing turnabout (18.3.4).
Then he gives one verse to Karṇa: “And having caused Karṇa, unretreating
in battle, my eldest brother, exceedingly fierce, to be slain, who is there that is
more a sinner (pāpakṛ tamaḥ) than I?” (12.27.18). Note that while Yudhiṣṭhira
admits deception in the case of Droṇa and agitates over its consequences, he
does not do so in regretting his “most sinful” part in the death of Karṇa. Yet
there is deception here too, and it is more unsettling. As Fitzgerald vividly puts
it, before the war, Yudhiṣṭhira had “craftily suborned” Mādrī’s brother Śalya, in
effect his maternal uncle, to undermine Karṇa’s energy when he would serve
as Karṇa’s charioteer against Arjuna (5.8.25–32), and then eventually killed this
“traitorous coconspirator” in his only major triumph in man-to-man combat
on the war’s final day (2004a, 88, 89; cf. Hiltebeitel 1976 [1990], 271). When
Yudhiṣṭhira laments his part in Karṇa’s killing, he admits neither this decep-
tion nor its seeming cover-up, but only his guilt over killing an older brother.
Yudhiṣṭhira had developed a grand obsession to kill Karṇa, who, he learned,
had sworn to Duryodhana in the Kaurava court, “I shall not wash my feet until
Arjuna is dead” (3.243.15–20). Even before the war starts, Yudhiṣṭhira has such
fear of Karṇa that his killing is the final matter he seeks to contrive just before
the fighting breaks out. What I wish to explore is that the collusion with Śalya
is a surface reflection of a secret and deeper counterdeception that Yudhiṣṭhira
was involved in all along, with Karṇa, Kuntī, and Kṛṣṇa: that of deceiving him-
self. When Karṇa lies dead, Yudhiṣṭhira reveals that ever since the dice match,
he should have recognized Karṇa as his brother since his feet were like Kuntī’s
(12.1.41–42; see Fitzgerald 2004a, 78 n. on other references). That is, I believe,
he should have bowed to Karṇa’s feet as he does to his mother’s.
Last, Yudhiṣṭhira blames himself for the deaths of Draupadī’s five sons
and Abhimanyu. Having played his part in bringing about the killings of his
family’s legitimate heirs, he calls himself “a destroyer of the Earth.” Note that
he does not mention other Pāṇḍava sons slain in the war: in particular Bhīma’s
son Ghaṭotkaca, whose death he would have a hard time faulting himself for,
as we shall see.102 At this point Yudhiṣṭhira determines to fast to death until
Vyāsa—playing his familiar role of Yudhiṣṭhira’s prompter—tells him to get
on with the life he was created for (23–32)!
One might seek to interpret these sins, as I did over thirty years ago, as
falling into trifunctional zones of the type analyzed in Dumézil’s comparative

102. Tamil traditions do give him some accountability for the death of Irāvat-Aravāṉ (see Hiltebeitel 1988,
321–22; 2011b, chapter 14).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 455

handling of the “three sins of the warrior” (Dumézil 1969): killing Bhīṣma and
Droṇa as violations against “first function” values of vows and truth endorsed by
Brahmins; slaying Karṇa as a “second function” violation of warrior codes upheld
by Kṣatriyas; and the deaths of the heirs as a violation of “third function” values
linked with women’s fertility and the welfare of the Earth. But I would no longer
argue that the sins of the warrior can be transposed to a king, or that Yudhiṣṭhira’s
“rehabilitation” of dharma “is one of the three functions itself, over and against
the forces of impurity, adharma, disorder, chaos” (Hiltebeitel 1976 [1990], 282).
Daniel Dubuisson (1979) attempted, with I think even more forced results, to
relate this theme to Rāma, who admits to doing nothing wrong: the killing of
Rāvaṇa as his “first function” sin because Rāvaṇa is a Brahmin; the killing of
Vālin as his “second function” sin because Rāma violates the warrior code; and
the repudiation of Sītā as a violation of the “third function” sanctity of femininity
and the Earth.103 The best one can say is that these arguments give a sense of
three “legitimate” areas where these two kings’ culpability has been questioned;
and that they underscore affinities in the killings of Vālin and Karṇa.
Yet the killings that stand out for lowering Rāma and Yudhiṣṭhira’s repu-
tations are not those of Vālin and Karṇa but those of Vālin and Droṇa. There is
a tradition of aligning these two episodes. Some Northern Recension manu-
scripts interpolate a comparison into the Mahābhārata itself in the aftermath of
Yudhiṣṭhira’s lie and Droṇa’s death. Arjuna tells Yudhiṣṭhira, “The ill fame will
stay long in this triple world with its animate and inanimate things—as in
Rāma from the killing of Vālin, so in bringing about the fall of Droṇa” (Mbh
7.1375* lines 1–2). This probably reflects a tradition of focusing on the episodes
together because of the similar terms used to rebuke Rāma and Yudhiṣṭhira by
their detractors. But the point is not just to align these asymmetrical scenes:
one a martial act with an arrow shot to kill a monkey king, the other a verbal act
of lying to abet killing a Brahmin; one done before the Rāmāyaṇa war, the other
during the Mahābhārata war. Rather it is to see how these two dharma biogra-
phies (to the extent that they are such) are constructed with such famous sins
made so prominent in the telling.

E.1. Rāma and the Ambush of Vālin

I must again be brief with Rāma and save comparison for the end. At Kabandha’s
advice, Rāma befriends the younger of the two royal monkey-brothers, Sugrīva.
An exiled king like Rāma himself, Sugrīva wants to kill his older brother Vālin.

103. See Dubuisson 1979. I say misleading results, for one thing, because with the “second function” sin
coming first and the “first function” sin second, their sequence does not follow a trifunctional order. This requires
Dubuisson to ignore one of Dumézil’s self-imposed restrains on cherry-picking.
456 dharma

But from Vālin’s standpoint, Sugrīva had come to the throne only as a usurper.
Without questioning either brother, Rāma hears Sugrīva’s dubious stories
about how Vālin recovered his throne from him and forced him into exile.
Rāma then tells Sugrīva to challenge Vālin to single combat, and shoots Vālin
from ambush. Against the dying Vālin’s complaints that Rāma bears the ban-
ner of dharma unrighteously,104 Rāma has only dubious replies, even though
the text and its commentators certainly honor them. Among them are the fol-
lowing: Rāma acts as his younger brother Bharata’s proxy (4.18.7–11, 23–25) in
that princes go about the world guarding dharma (9), which is subtle (15). Vālin
is only a monkey and cannot understand dharma (5, 16, 39), yet deserves this
punishment for the sin of taking his brother’s wife (18–21)! Having “made
Bharata’s command our sacred law,” Vālin was punished for “transgressing
the proper limits” (25: maryādān). Moreover, Rāma had promised Sugrīva to
kill Vālin and his Truth is unexceptionable (27). As Rosalind Lefeber says, this
is the “excuse that dominates” (1994, 45), not only since Rāma’s word must
always be “true” but since everyone knows it has been promised beforehand.
Yet as if that were not enough, “Men seeking meat shoot animals that are atten-
tive or inattentive or even facing the other way, and there is nothing wrong with
this” (35). Finally, to top it all off, kings should not be harmed or censured as
“they are gods in human form going about on earth” (38). We do not know
whether it is Rāma’s divine or human side that is speaking, and neither does
he. And perhaps that uncertainty leaves him less subject to criticism. But every
argument is opportunistic, as was believing Sugrīva’s story in the first place.
Rāma is above reflecting on them himself, and the only “personage” to have
challenged him lies beaten down to submission (40–44). My impression is that
this grab bag of arguments, which are strict to the extreme in their rationaliza-
tions of “punishment” in the name of dharma, are implicitly left open to further
questioning by listeners and readers who are not monkeys.105
That Rāma can invoke his Truth in the killing of Vālin even in support of
an ethically dubious and self-serving promise is testimony to his reputation for
this virtue, on which his exemplification of dharma unquestionably stands.
Rāma’s dharma biography revolves around upholding not only his own truth
but his father’s.

104. Rām 4.17.18; see above at n. 26.

105. What of the question, “Do animals, and particularly monkeys, know dharma?” Hanumān has the
same view when he meets Bhīma in the Mahābhārata: “We who come from animal wombs do not know the Law”
(vayaṃ dharmaṃ na jānīmas tiryakyoniṃ samāśritāh;̣ 3.146.75cd). But Hanumān is clearly being ironic, since he
goes on to chide his younger brother Bhīma for sullying the forest in ways that show that it is Bhīma who “does
not know the Law” (76–77), and stops him out of compassion and goodheartedness lest Bhīma menace the high-
mountain Siddhas (76–80). Clearly Indian animals do know dharma in fable literatures like the Jātakas and
Pañcatantra. Cf. Hiltebeitel 2001a, 195–202 (“Talking with Animals”).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 457

E.2. Yudhiṣṭhira’s Lie to Droṇa

In the Rāmāyaṇa, no virtue is higher than truth. It is a different matter in the

Mahābhārata, which resists universalizing any virtue and always has a contextual
answer when someone asks, “What is the highest dharma?” There are, however,
three leading answers to this question: noncruelty, truth, and nonviolence.106
Truth is obviously important in Droṇa’s killing, and Jonardon Ganeri brings this
out richly in two studies—the two most important to expand our understanding
of the episode.107 But one of my three main criticisms will be that he is fixated on
only this one virtue. The story of Droṇa’s killing brings all three of these virtues
into play. I shall explain my two other differences with Ganeri shortly.

e.2.a. good men are hard to find. Karṇa and Droṇa are something of a
pair, particularly in the scenes building up to the killing of Droṇa, where we
will find several intensified usages of the names “Droṇa and Karṇa”
(droṇakarṇau; 7.158.32–34; 44–45) in the dual.108 As we shall see, the pairing
holds the implication that the two together are insuperable. Karṇa has been
called the Mahābhārata’s tragic hero, since he bears up to Aristotle’s type as
a flawed good man.109 As Kṛṣna puts it after Karṇa’s death: “He who announced
Draupadī won by dice” was “the vilest of good men (satpuruṣas)” (Mbh
8.69.17)! To be sure, the Mahābhārata finds many ways to say that no one is
purely good or wicked, and that some goodness can be found in every man or
woman. But good men called satpuruṣas are hard to find.110 Yudhiṣṭhira is
never called a satpuruṣa, although before the dice match he was naive enough
to chide Śakuni about the deceitfulness of gambling and say, “the vow of a

106. See Hiltebeitel 2001a, 209–14 on these three as being statistically the most often cited among the
so-called “highest dharmas.” On non-violence, see Proudfoot 1982.
107. Ganeri 2005, 2007. These two studies contain much overlap, as will be occasionally noted, but I will
cite mainly the second since it introduces new arguments.
108. See Von Simson 1968, 41, noticing this and suggesting an analogy with the single name Kumbhakarṇa
in the Rāmāyaṇa, which means the same thing—“pot-ear”—as the compound Droṇa–Karṇa.
109. See Shulman 1985, 35, 44–46, and 380–400 on the “Tamil version of the tale of Karṇa, which even
Aristotle would have deemed tragic” (45). Cf. Adarkar 2001; Redfield 1994 on “the tragedy of Hector.”
110. Usages are either general (2.53.8, cited above; 3.241.32; 5.33.94), exhortative (3.34.2; 281.29; 6.20.4;
7.2.20; 12.71.6; 120.47), cautionary (4.13.15), or simply scattered (6.50.85; 56.28; 75.17; 7.118.9; 133.19; 11.19.16;
13.11.16). The term is unused in the dharmasūtras or Manu. It has a few uses in the Rāmāyaṇa. It seems more
common in Buddhist usages. See Rhys Davids and Stede [1921–25] 2003, 87 on asappurisa, “a low, bad, or
unworthy man,” citing various texts; Nattier [2003] 2005, 223 n. 93; Schopen 1997, 178, 187, 189. Although there
are numerous terms that can define a good woman, including pativratā, satī, and sādhvī, the corresponding term
satstrī is seldom used. Its main use is for Draupadī. Just as she is heading into exile, Kuntī reminds her that “good
women are not timorous” (satstrīṇāṃ vaiklavyaṃ nopajāyate; 2.70.7b); and in opening and closing her colloquy
with Satyabhāmā, Draupadī says of herself that she avoids the ways of “women who are not good” (asatstrīṇām;
3.222.9 and 57). Otherwise, cf. 3.72.26 on Damayantī, with Draupadī listening, and the exemplary Śāṇḍilī of the
Śāṇḍilī-Sumanā-Saṃvāda (13.124; see v. 1), whom she probably hears about too.
458 dharma

good man” is to undertake “honest, uncrooked war” (2.53.6–8). Droṇa is

never, and I believe never could be, called a satpuruṣa, and neither is Arjuna.
To put Kṛsn ̣ ạ ’s words in context, one must go back to the dice match. As
Biardeau remarks, when Draupadī raised her insistent dharma question after
being gambled, Bhīṣma at least spoke out, but Droṇa said nothing111—a fact that
Yudhiṣṭhira will recall with grim finality after Droṇa’s death (Mbh 7.170.32).
Biardeau also regards it as an innuendo (sous-endendu) that “the death of Droṇa
after that of Bhīṣma brings to a close the elimination of the enlightened (éclairé)
partisans of dharma from the camp where they have a role opposed to their
deeper nature” (2002, 2: 254). But she underestimates Karṇa. Karṇa also spoke
out after the dicing, extremely harshly, calling Draupadī a whore for having five
men before ordering her disrobing (2.61.35). But remarkably, he comes to regret
this. When Kṛsn ̣ ạ secretly invites Karṇa to join the Pānd ̣ ạ va side, Karṇa admits
that he now “burns from the karma” of his harsh words to the Pānd ̣ ạ vas that he
uttered to please Duryodhana (5.139.45), which must include his insults to
Draupadī. And on the night before Karṇa takes command of the Kaurava army
he even shares regrets with Duryodhana, “recalling the pain they had caused
Kṛsṇ ạ ̄ [Draupadī] at the dicing” (8.1.7)! Thus when Kṛsn ̣ ạ says of the freshly slain
Karṇa, “He who announced Kṛsn ̣ ạ ̄ won by dice, the vilest of good men
(satpuruṣad̄ hamaḥ)—today the earth drinks that Sūta’s son’s blood” (8.69.17),
it is with terrible irony. But Kṛsn ̣ ạ also overstates the matter, for the terms under
which Kṛsn ̣ ạ knows Karṇa to be a “good man” (satpuruṣa) are hardly vile. Just
before Kṛsn ̣ ạ says this, Vaiśaṃpāyana lauds the fallen Karṇa as a satpuruṣa without
deprecation: “the one who also said ‘I give,’ and not ‘It is not so,’ when sought by
seekers, a good man always with the good (sadbhiḥ sadā satpuruṣaḥ)” (8.68.44).
When Kuntī, the mother who had abandoned Karṇa at birth, approached him as
just such a favor-seeker before the war, “he turned around, and seeing Kuntī,
saluted her with joined palms, as was proper, this proud man of great splendor
(tejas), the best of dharma’s upholders” (5.142.30). He tells her he will not do her
bidding; he will fight her sons with all his strength. But he will spare her other
four sons in battle and only seek to kill Arjuna. He or Arjuna will die, but five will
survive. And he will do this—in his own words—“while trying to persevere in the
conduct of noncruelty that befits a good man (ānṛśaṃsyam atho vṛttaṃ rakṣan
satpuruṣa-ucitam)!112 Kuntī leaves knowing she could not have asked for more.

111. See Biardeau 2002, 1: 246, 268. After telling Draupadī “the matter is subtle,” Bhīṣma says, “Droṇa and
the other elders who are wise in the Law sit bent over as though spiritless with empty bodies (ete droṇad̄ ayaś caiva
vṛddhā dharmavido janāh/̣ śūnyaiḥ śarīrais tiṣṭhanti gatāsava ivānatāh)̣ . But Yudhiṣṭhira, I think, is the authority in
this question; let he himself speak out and say whether you have been won or not” (2.62.16, 20–21).
112. Mbh 5.144.19ab. Shulman 1985, 384, notes this passage, and van Buitenen’s “insightful translation”:
“His goal is ‘to persevere in the humane conduct that becomes a decent man.’”
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 459

This matter of Karṇa being a satpuruṣa is also invoked indirectly, and with
deepening ironies, by Śalya in two upākhyānas. First, near the beginning of the
war preparations in Book 5, Śalya, while agreeing to collude in the destruction
of Karṇa’s tejas, tells Yudhiṣṭhira the Indravijaya-Upākhyāna (5.9–18). Śalya
recalls how the Ṛṣis cajoled the Asura Vṛtra to make a fateful pact with Indra by
answering his question, “How can there be friendship (sakhyam) between our
two splendors (tejasor hi dvayor)?” (10.22cd)—that is, between Vṛtra’s tejas and
Indra’s—to which the Ṛṣis answered, “One does not transgress a pact with a
good man” (10.23c)! Whichever is the “good man,” Indra will break this pact
with the compliant Vṛtra, and the killing of Karṇa will be mysteriously com-
pared by Kṛṣṇa—as always in harmony with the Ṛṣis—to Indra’s killing of
Vṛtra.113 Later, in the war book on Karṇa’s killing, Śalya, now Karṇa’s chari-
oteer, tells Karṇa the Haṃsa-Kākīya-Upākhyāna or “Subtale of the Swan
and Crow” (8.28) to undermine Karṇa’s tejas just before his duel with Arjuna.
This story tells how the crow (being compared to Karṇa) falters, trying to match
the swan’s flight over the ocean, until the swan (compared to Arjuna) sees the
crow “about to sink and, remembering the vow of a good man” (28.44), accepts
the crow’s pleas for protection, but not before first reminding him of all his idle
boasting before he got out over the water. Like his father Indra against Vṛtra,
Arjuna will be a “good man” only in the breach in his duel with Karṇa. And
unlike the swan with which he has been compared, Arjuna will recall no “vow
of a good man” when Karṇa’s chariot wheel gets stuck in the earth. Rather than
letting Karṇa extract it so that they can resume fighting fairly, Arjuna kills him
at Kṛṣṇa’s prodding (9.66.59–67.24). Indeed, when Arjuna is about to kill
Karṇa, Karṇa says to him, “Having seen this wheel of mine swallowed by fate,
Arjuna, abandon the intention practiced by a lowly man” (8.66.61). When
Arjuna then kills the disadvantaged Karṇa, it is, at least in Karṇa’s words, as a
kāpuruṣa, the very opposite of a satpuruṣa.
Satpuruṣa seems to have this special force for Karṇa because he is good
without the benefit of high social station. Considering Karṇa’s reputation for
friendship and giving, especially to beggars (not only Indra comes to him in this
guise; so really does Kuntī), his being a satpuruṣa has a strikingly analogous use
in the Ugraparipṛcchā Sūtra. There, the Buddha urges that the householder pur-
suing the bodhisattva path be a satpuruṣa who first and foremost cultivates the
perfection of giving (dānapāramīta) by giving to beggars who are themselves to
be addressed as “satpuruṣa” when the householder cannot give all they ask for,
and who are to be thought of as one’s kalyānạ mitra—a friend who appears
by good fortune who enables one to practice maitrī, “friendship” or “loving

113. On Mbh 8.69.1–3 especially, after several preceding verses, see Hiltebeitel 2007b, 56–57 and 72–73 n. 211.
460 dharma

kindness,” the first of the four sublime abodes (see chapter 4) and the one
singled out for universal application, rather than karuṇā (“compassion”), in this
early Mahāyāna sutra. 114 That the Mahābhārata invests such energy in an exem-
plary good man of low station means that there may be others of his kind whose
mistreatment and killing might trouble the conscience of a fairly good king.

e.2.b. a hinge passage. These reflections bring us to the killing of just such a
person: Ghaṭotkaca, a son of Bhīma. Before marrying Draupadī, Bhīma sired
Ghaṭotkoca with a lady of the wilds, a Rākṣasī or demoness named Hiḍimbā.
While Ghaṭotkaca’s death is episodic to our concern with the killing of Droṇa,
Bhīma’s part becomes important. Ghaṭotkaca’s killing occurs at a hinge in the
text toward the close of the Mahābhārata’s seventh Book, the “Book of Droṇa,”
which covers the five days that Droṇa marshals the Kaurava army. While the
epic has eighteen major parvans, it also has a hundred “sub-parvans” or
upaparvans, and the three sub-books that end Book 7 are titled “The Death of
Ghaṭotkaca” (7.122–54), “The Death of Droṇa” (7.155–65), and a closing sub-
book on recriminations and further fighting called the “The Release of the
Nārāyaṇa Weapon” (7.166–73). As established by the Critical Edition based on
good manuscript evidence, the hinge occurs at the beginning of the “Death of
Droṇa Sub-book” (7.155–58), but its character as a swing segment can be
underscored by noting that numerous manuscripts consider these four chapters
to form the end of the “Death of Ghaṭotkaca Sub-book” instead.115
The “Death of Ghaṭotkaca Sub-book” recounts a terrifying night battle that
is resumed even in its sequel (Mbh 7.129–61). Ghaṭotkaca’s demonic powers
grow so great at night that Duryodhana presses Karṇa to use up a sure-shot
one-use weapon he was saving for Arjuna. As the episode is rehashed in the
hinge segment, Kṛṣṇa tells Arjuna why he just did a little dance over Ghaṭotkaca’s
death.116 Kṛṣṇa’s tactics and evasions had succeeded in depriving Karṇa of the
one weapon by which he could have killed Arjuna. Arjuna will now be able to
kill Karṇa when his chariot wheel gets stuck—to which Kṛṣṇa adds that several
other foes, including Ekalavya, “were one after another all slain by diverse
means for your sake by me.”117 Kṛṣṇa goes on to say that he killed these foes
because, had they sided with Duryodhana, he could have conquered the earth

114. See Nattier 2002, 146, 223 and n. 93; 228; 241–45; 257–59 and n. 329). Cf. Schopen 1997, 186–87, on
uses in inscriptions to describe locally appreciated deceased monks, not “great saints,” recognized for being
“little more than ‘a good and worthy man.’” See the Florian Henckel von Donnersmarck’s 2007 film “The Lives
of Others” for a stunningly similar character and narrative.
115. De 1958, xviii–xix, 895, 916. The CE decision is based mainly on what can be considered early Śāradā,
Kāśmīri, and Malāyalam manuscripts.
116. On this dance (Mbh 7.155.2–3) and a fuller discussion of the sequence, see Hiltebeitel 2007b, 33–36.
117. Mbh 7.155.11–29. The others were Śiśupāla and Jarāsaṃdha, on whom see chapter 13 § F.1. On Ekalavya
in current discussions see Shankar 1994.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 461

(156.2–5). Since Arjuna knows how the others were slain from being present at
their killings, Kṛṣṇa must be directing his curiosity to Ekalavya. Arjuna does
know something about Ekalavya’s killing, since he recounted it among Kṛṣṇa’s
feats in a prewar meeting (5.47.71). What would appear news to him is why
Kṛṣṇa did it. Says Kṛṣṇa,

For your sake the Niṣāda’s son, whose prowess was true, who was
incapable of being baffled, was deprived of his thumb by Droṇa,
assuming the position of his preceptor, by an act of guile. . . . With
his thumb, Ekalavya was incapable of being vanquished. . . . For your
sake he was slain by me at the van of battle. (7.156.17–21)

Ekalavya’s disablement was really prompted by Arjuna, who, in his youthful

training in arms, felt threatened that Ekalavya would replace him as Droṇa’s best
disciple (1.123.10–39). This reminiscence complicates any view of Arjuna’s motiva-
tions in the Droṇa-killing episode. I take it as a warning against overindulging in
admiration for either Droṇa or Arjuna, but especially for Arjuna, in what follows.
Kṛṣṇa also adds that Ghaṭotkaca hated Brahmins and that Kṛṣṇa would
have killed him too for the sake of upholding dharma had no one else done so
(7.156.23–29). Kṛṣṇa is making all this up about Ghaṭokaca’s adharma and
hatred of Brahmins, or at least contradicting Vaiśaṃpāyana’s main narrative.
When the Pāṇḍavas were in exile and Bhīma called on Ghaṭotkaca to carry
Draupadī on his back so that he could fly her over a mountain, Ghaṭotkaca’s
dharma was underscored and he even carried the Pāṇḍavas’ attending Brahmins
on his back along with her (3.144.25–145.9)!
Here I begin to make two arguments. First, mutual complicities in the
death of Droṇa have not been sufficiently understood because no one has taken
account of this preceding sequence. Second, Yudhiṣṭhira’s particular part in
killing Droṇa now begins to be intertwined with his desire to kill Karṇa because
of what seems to him to be the dilatory behavior of Arjuna in fighting either of
them. These arguments bring me to my second and third overriding differ-
ences with Ganeri. My second is that whereas I approach matters from
Yudhiṣṭhira’s—that is, the liar’s own—perspective, Ganeri wants to see mat-
ters mainly from Arjuna’s.118 This is the case in both of his studies, but it is

118. While vaunting Arjuna, Ganeri tends to caricature Yudhiṣṭhira, attributing generalities to him without
specific attribution. Yudhiṣṭhira is thus “rule bound” (2007, 74, which Ganeri gets from Matilal; see chapter 1);
supposedly, “it is . . . his particular duty (svadharma) to speak only the truth” (84). Ganeri tends to overstatement
in faulting Yudhiṣṭhira: After his chariot “crashes to the ground! Yudhiṣṭhira will never again be taken at his
word” (Ganeri 2007, 67). No “crash” is evident in the text, and I know of nothing to support this alleged abrupt
change in Yudhiṣṭḥira’s credibility. Twice Ganeri makes Yudhiṣṭhira accountable for “murder” (71, 72). Most
important, I do not think Yudhiṣṭhira believes he is being truthful when he is lying to Droṇa (79, 82), and cer-
tainly not in the aftermath, where, as we shall see, only Dhṛṣṭadyumna makes such a claim.
462 dharma

most pointedly so in the second, whose title, “A Cloak of Clever Words: The
Deconstruction of Deceit in the Mahābhārata,” rebuilds the argument around
Arjuna’s denunciation of Yudhiṣṭhira’s lie: “You told our teacher that ‘the
kuñjara is slain’; this being but a falsehood wearing the truth as an armour-
skin” (satyakañcukaṃ nāma praviṣṭena tato ‘nṛtam; 7.167.35; Ganeri 2007, 72).
In offering a new twist on what Yudhiṣṭhira’s might mean by kuñjara, a word
for “elephant,” Ganeri will conclude that he “covers truth in a very fine cloak”
(84), which matches his interpretation with Arjuna’s. As we have seen
(§ E above), Yudhiṣṭhira owns up to this charge of Arjuna in his postwar mea
culpa; so it is not that Arjuna is wrong. But there will be more to say about
Yudhiṣṭhira’s own entanglements.
My third difference with Ganeri is that in favoring Arjuna over Yudhiṣṭhira,
he shapes his interpretation around personalistic and individualistic ethical
and philosophical questions that arise from the Bhagavad Gītā,119 but do not
take into account the context of Yudhiṣṭhira’s lie, or his wider responsibilities
as a king. To move ahead, I must thus offer some caveats and say a few words
in advance on the Bhagavad Gītā. First, it can be misleading to think, as Ganeri
does, that Kṛṣṇa’s advice should always be squared with what he says in the
Gītā.120 Kṛṣṇa is quite capable of talking out of different sides of his mouth to
different people. What he has to say to Arjuna in the Gītā is for a particular time
and place, and is neither his once-and-for-all pronouncement to Arjuna nor his
one-cut-for-all-sizes message to everyone else in the Mahābhārata. Thus, sec-
ond, even if one grants that Arjuna gets Kṛṣṇa’s most carefully considered
teachings in the Gītā, those teachings do not let Arjuna off the hook for his
actions and especially his inactions in scenes of battle. Many scholars give
Arjuna a free pass for having heard the Gītā and continued to act on it to the
minimal extent that he always heeds the battle advice of Kṛṣṇa, his divine
charioteer. But as Arjuna himself will admit, even after charging Yudhiṣṭhira
with the unrighteous killing of Droṇa, “Droṇa was neglected by me even though
he was being killed because of my desire for the kingdom” (7.167.49). Arjuna
admits here to inaction out of the desire for fruits, the very inverse of one of
Kṛṣṇa’s main Gītā teachings. Yudhiṣṭhira, on the other hand, is usually left to
his own resources in the Mahābhārata war, except when he is fortunate enough

119. Ganeri’s focus on truth, drawing on the philosopher Bernard Williams, emphasizes virtues of self-
preservation, integrity, calmness, and steadiness of mind (2005, 186; 2007, 86, 235)—all virtues that the Bhagavad
Gītā emphasizes as pertinent to a warrior facing battle, as Ganeri shows—but they are not advice to a king.
120. See Ganeri 2005, 188–86, 189–90, 195; 2007, 67–68. The determination to measure all by the
Bhagavad Gītā is most unfortunate at the top of Ganeri 2005, 190: “If Yudhiṣṭhira embodies one strand in the
closely woven ethics of the Bhagavdagītā, and Arjuna another, then Bhīma manifests a third, the voice of caste,
hierarchy, and social order.” These are far more Arjuna’s worries (see chapter 11) than Bhīma’s; and neither
Bhīma nor Yudhiṣṭhira had the benefit of hearing the Bhagavad Gītā.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 463

to hear from Kṛṣṇa or from that other unimpeachable authority, his own special
prompter, the author Vyāsa. Vyāsa now makes just such an appearance, still
within our hinge passage, before Yudhiṣṭhira.
It is still in the dead of night. The fighting is yet to be resumed and
Yudhiṣṭhira is not party to any of the information about the killing of Ghaṭotkaca
that Kṛṣṇa has imparted, perhaps secretly, to Arjuna. Unlike Kṛṣṇa and Arjuna,
Yudhiṣṭhira begins with feelings of grief over his nephew Ghaṭotkaca, which
he expresses briefly for now to Ghaṭotkaca’s father Bhīma, urging him against
the Kaurava host:

“O strong-armed one, resist the Dhārtarāṣṭra’s army. With the attack

on Hiḍimbā’s son, a great bewilderment possesses me (moho mām
āviśan mahān).” (7.158.21c–d)

Meanwhile, if Bhīma feels anything over the death of this son, it is untold,
though I think to be assumed, and presumably Yudhiṣṭhira senses this as he
sends him into the fray, principally against Droṇa and in protection of the
Pāñcālas whom Droṇa is destroying.
Then, “Having assigned Bhīma in this fashion, he [Yudhiṣṭhira] got into his
own chariot. His face full of tears, sighing again and again, the king became ter-
ribly dejected (kaśmalaṃ prāviśad ghoram)121 at the sight of Karṇa’s prowess. Seeing
him so agitated (vyathitam), Kṛsn ̣ ạ said a word” (158.22–23b). Yet what Kṛsn
̣ ạ says
is once again less interesting than what he does not say. That is, he says nothing
about why he engineered the death of Ghaṭotkaca, much less his little dance over
it. He just tells Yudhiṣṭhira to get over his agitation (vyatham) and feebleness (vai-
klavyam) and to “rise up” (uttiṣṭha) and “bear the heavy yoke” (vaha gurvīṃ dhuram),
lest victory become uncertain (23c–24). Wiping his eyes, Yudhiṣṭhira tells Kṛsn ̣ ạ ,

The excellent way of the Laws (dharmāṇāṃ paramā gatiḥ) is known to

you, strong-armed one: the fruit of Brahmanicide is his who does not
become aware of what has been done [for him]. (26)

It seems something of a rebuke at Kṛṣṇa’s callousness, but also a begrudg-

ing admission that Yudhiṣṭhira will not overlook the gratitude he owes to
others, and particularly to Kṛṣṇa and Arjuna. But the tone also registers that
he knows gratitude to this pair will come with its costs, and that his gratitude
to others—in particular Bhīma—must also be considered.
Moreover, if Yudhiṣṭhira’s words are an anticipation of what will be at
stake in settling the blame for the Brahmanicide of Droṇa, they also raise

121. Ganguli [1884–96] 1970, 6: 427: “became exceedingly cheerless”; kaśmalam can mean timid, pusil-
lanimous; dejected; distressed. On pra-÷viś/praveśa, see F. Smith’s discussion of possession in § D.2.b above.
464 dharma

the question, “Who among those responsible for Droṇa’s death will not
become aware of what has been done for him?” Yudhiṣṭhira, who was pre-
sent when Bhīma called on Ghaṭotkaca to carry Draupadī and who even
urged him to fly low so she would not be upset (3.145.6), does recall what
was done for him and the Pāṇḍavas in the forest. He reminds Kṛṣṇa that
while Arjuna was away seeking divine weapons, Ghaṭotkaca carried the
exhausted Pāñcālī on his back as they crossed Mount Gandhamādana, and
achieved other difficult feats “for my sake” (7.158.27–33). If Yudhiṣṭhira was
brief in expressing his grief over Ghaṭotkoca to Bhīma, he now makes clear
a genuine affection for him.122

My affection for Ghaṭotkaca, that Indra among Rākṣasas, is twice the

affection I bear naturally (svabhāvād) for Sahadeva. That strong-armed
one was devoted to me (bhaktaś ca me), and I was dear to him and he
was dear to me. Scorched by grief, that is why I become dejected,
Vārṣṇeya. See our troops, Vārṣṇeya, routed by the Kauravas. And
Droṇa and Karṇa contending together (droṇakarṇau ca saṃyattau) in
battle, see the two great chariot warriors. See the Pāṇḍava army
crushed at night like a great reed forest by two maddened elephants.
(158.31–34)

Here, we find the first of several usages within our swing passage of Droṇa and
Karṇa in the dual, and coming with this triple usage of paśya, “see!” Yudhiṣṭhira
remarks that without respecting Bhīma or Arjuna, Droṇa, Karṇa, and
Duryodhana now roar in battle, having slain Ghaṭotkaca (35–36), and asks,

How, O Janārdana, when we are alive, and you too, could Hiḍimbā’s
son have died when engaged with the Sūta’s son? . . . in the sight of
Savyasācin [Arjuna] . . . ? (37–38b)

Yudhiṣṭhira’s gratitude to Kṛṣṇa and Arjuna exasperates him, and he continues

to sharpen this point. Where was Arjuna when Abhimanyu was killed? Why all
the fuss to punish Jayadratha for that when his contribution was small, and
when it was really brought about, as Yudhiṣṭhira sees it, mainly by Droṇa and
Karṇa (39–43):

If in killing enemies it should be done right (nyāyyo bhavet kartum)

by the Pāṇḍavas, then Droṇa and Karṇa (droṇakarṇau) should have

122. Yudhiṣṭhira’s affection for Ghaṭotkaca may carry down to Janamejaya, who, in his last profound
question at the end of the epic, mentions Ghaṭotkaca as the last person he wonders about as to what destination
all the divine and demonically incarnated heroes and heroines reached “at the end of their karma” (Mbh 18.5.3–5;
see Austin 2009, 600–601).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 465

been slain before this in battle, that is what I think. These two surely
are the root of our woes (mūlaṃ hi duḥkhānām), bull among men.
(44–45b)

Again, he repeats, Arjuna slew Jayadratha when he should have killed Droṇa
and Karṇa (46–47b). And clearly implying that Kṛṣṇa and Arjuna have been
lackluster where it counts most, he says, “I will go myself with the desire of
slaying Karṇa (karṇajighāṃsayā)” while Bhī ma opposes Droṇa (47c–f). Note
that Yudhiṣṭhira never gets an answer from Kṛṣṇa on any of this, and seems to
know he is asking him rhetorical questions, and presumably all in Arjuna’s
hearing. His desire to fight Karṇa seems genuine, but postured. Yudhiṣthira
sets off, “holding his formidable bow and blowing his conch fiercely (bhaira-
vam)” (48cd), followed by a large but unpromising force123 and observed by
Kṛṣṇa, who tells Arjuna,

This Yudhiṣṭhira possessed by wrath (krodhāviṣṭaḥ) goes forth hastily

desirous of killing the Sūta’s son. His neglect is not proper
(tasyopekṣā na yujyate) (158.51c–f)

̣ ạ maneuvers himself and Arjuna to follow Yudhiṣṭhira from a distance (52).

Kṛsn
Clearly it has been a testy exchange. Yudhiṣṭhira has every right to think he
is being left out of some master plan. Arjuna could be thinking he would like
to let Yudhiṣṭhira stew in his own juices after all the innuendos he has just
heard about his lackadaisical fighting. Meanwhile, Kṛṣṇa can keep tabs but it is
not his usual role to tell secrets or talk sense to Yudhiṣṭhira. It is time for a visit
from Vyāsa, who sees Yudhiṣṭhira speeding toward Karṇa and tells him some,
but not all, of what Kṛṣṇa has not told him. Vyāsa tells Yudhiṣṭhira it is by good
luck (diṣṭyā) that Arjuna is still alive, having encountered Karṇa at all while he
had the Unfailing Weapon; by good luck, Arjuna did not engage Karṇa in a
chariot duel then, which would have led to Karṇa killing him; by good luck
Ghaṭotkaca was slain, indeed on Yudhiṣṭhira’s behalf. Yudhiṣṭhira should put
aside anger and grief and unite with all his brothers and allies in fighting the
Kauravas:

“In five days the Earth will be yours! And always, O tiger among
men, keep just considering dharma (dharmam eva vicintaya)! Highly
pleased, let yourself resort to noncruelty, austerity, generosity,
forbearance, and truth (ānṛśāṃsyaṃ tapo dānaṃ kṣamāṃ satyam . . .
̣ Pāṇḍava. Where dharma is there is victory.” Having so
sevathāh),
spoken to the Pāṇḍava, Vyasa disappeared right there. (158.60e–62)

123. It is headed by Śikhaṇḍin, the brother of Draupadī whose main martial trait is his effeminity.
466 dharma

Noncruelty is the first virtue mentioned, truth the last. Coming from the author,
I believe we are within reason to take this as a set of virtues concerned with
what the philosopher Bernard Williams calls “the values” and “virtues of
truth.”124 I take it as an invitation to make some sense of Yudhiṣṭhira’s choices
and actions which this hinge passage has served to frame, and which the “Death
of Droṇa Sub-book” will now go on to describe.
What is striking in this set of values is that Vyāsa brings Yudhiṣṭhira back
to the virtue of noncruelty. As we saw at the end of “The Yakṣa’s Questions,” this
was the virtue behind Yudhiṣṭhira’s choice to revive Nakula rather than Bhīma
or Arjuna so that Kuntī’s cowife Mādrī would have one son like her rival, even
though Mādrī was long dead (3.297.71); it was the virtue by which Karṇa wished
to be a “good man” in telling Kuntī she would still have five sons; and now it is
the top virtue that the author, Pānd ̣ u
̣ ’s father and Yudhiṣṭhira’s grandfather,
mentions to keep Yudhiṣṭhira from facing Karṇa, which could endanger Karṇa’s
promise to Kuntī. Indeed, not only does Yudhiṣṭhira’s grandfather speak for this
virtue, so does his “real” father Dharma: as the Yakṣa who endorsed it when
Yudhiṣṭhira chose the revival of Nakula. Before any of this, it lay at the root of
the curse that led to the death of the Pānd ̣ ạ vas’ lineal father Pānd
̣ u
̣ when he shot
a pair of deer unaware that the buck was a Ṛsị uniting with an apparently real
doe (Brodbeck 209a, 174 n. 23), and earned the Ṛsị ’s curse—for lacking the
noncruelty to wait until they were finished making love (1.109.5–31). In relating
noncruelty as the highest dharma to a primary context in this royal family, there
is a thread here through which the Mahābhārata is saying something distinctive
about the dharma of kings. For one thing, a king’s family drama is everybody’s
family drama. Moreover, for a king, noncruelty is a family value in the largest
sense. It is a “creature-feeling” that extends “across the great divides”:125 of
humans and animals (as it begins in the breach when Pānd ̣ u
̣ shoots the mating
deer); of the living and the dead (as with keeping a son for the dead Mādrī); of
those of high and low standing (in Karṇa’s choice to remain a man of low station
while honoring the mother who abandoned him, and in Yudhiṣṭhira’s feeling
for Nakula, not to mention Ghaṭotkaca); and ultimately, in the person of Dharma,
of a hidden force that governs the universal rhythms of life and death.
As with Kṛṣṇa’s reminder of Ekalavya, I take Vyāsa’s words about
“noncruelty” to hang over Droṇa’s killing. If Arjuna keeps listening to Kṛṣṇa,

124. See Williams 2002, 59–61; cf. Ganeri 2005, 196; 2007, 81, 231–36, discussing Bhīṣma’s list to
Yudhiṣṭhira, after the war, of thirteen truth-related virtues at Mbh 12.156—noncruelty not being among them,
possibly because Bhīṣma is past speaking about rājadharma and has taken up truth in relation to times of distress
(āpaddharma). Recall that Yudhiṣṭhira asks for a reduced triple set of truth-related virtues from Dharma at the
end the Forest Exile and “The Yakṣa’s Questions”: giving, austerity, and truth, which I have taken as encapsu-
lating his chosen ideal of noncruelty at that point.
125. See Hiltebeitel 2001a, 213. Cf. Klaes 1975, 81: “Ānṛśaṃsya demanded of him understanding, honest
judgment and fellow-feeling.”
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 467

he will keep wanting—despite the cruelty it brings upon others—to spare his
benefactor Droṇa. The answer to the question of who among those responsible
for Droṇa’s death will not become aware of what has been done for him will be
Arjuna. Kṛṣṇa is protecting Arjuna in his complicity with Droṇa, and although
Arjuna does not want to see it, everyone else knows about it, as we shall soon
see. Meanwhile, if Yudhiṣṭhira has listened to the author, he might find grounds
to kill Droṇa by putting noncruelty before truth.

e.2.c. the death of droṇa. When the moon rises after both sides had agreed
to catch some sleep, the “Death of Droṇa” sub-book now continues with the
resumption of the night battle. Droṇa opposes the Pāṇḍavas and Pāñcālas. The
Pāñcālas are Draupadī’s people. Through her marriage to the five Pāṇḍavas,
they are the Pāṇḍavas’ closest allies. Their leading fighters are Draupadī’s twin
brother Dhṛṣṭadyumna and their king, Draupadī’s father Drupada. Droṇa has
feuded with Drupada since they were young men.
Kṛṣṇa now tells Arjuna to position himself to fight Droṇa and Karṇa.
Bhīma seconds this, calling Arjuna Bībhatsu: a name of his that can mean “the
Repugnant one,” “the one who is averse to something, who loathes,” which can
be suggestive when Arjuna is averse to fighting Droṇa and Karṇa.126 Bhīma
says that if Arjuna dawdles, it would be a noncruelty and an untruth:

O Arjuna, Arjuna Bībhatsu, listen to my word truly. The time has

come for what a Kṣatriya woman bears a son. If at this time you do
not strive to win the good, . . . you will do a noncruelty. Pay with vigor
the debt you owe to truth, prosperity, dharma, and fame. (7.161.7–9)

Bhīma’s call for Arjuna to override noncruelty in the name of truth is a reminder
(but of course only to listeners) and reversal of Vyāsa’s words to Yudhiṣṭhira.
Yudhiṣṭhira should heed five virtues with noncruelty first; Arjuna should dis-
miss noncruelty and heed four other values headed by the truth of his martial
vows. If Arjuna is disposed toward noncruelty now, it can only be toward Droṇa.
And Bhīma is saying that this is a cruelty toward their mother and brothers, not
to mention the Pāñcālas, that would amount to a lie should he continue in it.
Here we see a deeper level of Arjuna’s complicity with Droṇa over Ekalavya.
Droṇa had Ekalavya cut off his thumb not only so that Arjuna would be Droṇa’s
best disciple but so that Arjuna would help Droṇa avenge himself against

126. See Biardeau 2002, 1: 843, 851–52; 2: 358, 514: Biardeau treats the name Bībhatsu as having this sense
especially in interactions between Karṇa and Arjuna, noting its uses also in describing Arjuna’s repugnance at
killing Bhīṣma (2: 110), Jayadratha (2: 199), and Droṇa (2: 255), in whose case we will note several striking exam-
ples. The sub-book on the killing of Droṇa actually begins on this note. As the warriors on both sides fight on
“blind with sleep,” Bībhatsu says all should take time out, and both sides agree, as do the Gods and Ṛṣis, who
applaud, and even the Kurus bless Arjuna with thanks (7.159.16–34).
468 dharma

Drupada by conquering the Pāñcālas, enabling Droṇa to claim for himself the
northern half of the Pañcāla kingdom as his (Mbh 1.128.5–15).
Still at night, Droṇa moves away from a brief salvo with Arjuna to station
himself to the north like a smokeless fire (7.161.21), and from there he kills two
major Pāñcālas—a son of Dhṛṣṭadyumna and King Drupada himself—and
also King Virāṭa of the Matsyas. When the sun rises, Dhṛṣṭadyumna swears he
will kill Droṇa today (7.161.21–51). Arjuna’s noncruelty to Droṇa has just
allowed Droṇa to bring his feud with Drupada to its end, just as it will continue
to let Droṇa bring massive cruelty to the Pāñcālas.
From here, let us go straight to the two lengthy chapters (7.164–65) on
Yudhiṣṭhira’s lie and Droṇa’s death. For a bit, the two youngest Pāṇḍavas, the
twins Nakula and Sahadeva, fighting along side Dhṛṣṭadyumna, oppose
Droṇa and there is a “fair fight” (7.164.7–18). Yudhiṣṭhira now invokes the
dharma of Kṣatriyas and orders the Pāñcālas to counter Droṇa, implicitly
rallying them after they have lost their elderly king. Here there is an inter-
esting verse, considering what is said before and after it about Arjuna and
Yudhiṣṭhira:

Among the sons of Pānd ̣ u

̣ there were three non-crooked chariot
warriors, the twins and Bhīma, and they cried out to Dhanaṃjaya. (55)

As Droṇa’s destruction of the Pāñcālas increases, the three “uncrooked war-

riors” are fearful—as are Yudhiṣṭhira and the Pāñcālas—that Arjuna lacks
commitment to fight Droṇa (63–65). Throughout the war, it is the bard Saṃjaya
speaking, to whom Vyāsa has given a “divine eye” so that he can know every-
thing on the battlefield and chronicle the war for the blind Dhṛtarāṣṭra, father
of the Kauravas and their nominal king. Saṃjaya thus sets the terms by which
Dhṛtarāṣṭra, as first listener, would begin to weigh the actions of Arjuna and
Yudhiṣṭhira in the killing of Droṇa, which Dhṛtarāṣtra knows has already
happened.127
At this impasse, Kṛṣṇa now makes what Ganeri calls an “astonishingly
base” pitch (2005, 183; 2007, 66):

Beholding the sons of Kuntī afflicted with the shafts of Droṇa and
inspired with fear, Keśava, endowed with intelligence and devoted to
their welfare, addressed Arjuna and said, “This herdsman of chariot-
herdsmen can in no way be conquered in battle by warfare, even by
the slayer of Vṛtra [Indra] in war. (7.164.66–67)

127. As mentioned above in n. 59, each battle book begins with Saṃjaya’s report to Dhṛtarāṣṭra on the
death of the Kaurava marshal after whom the book is named.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 469

With his bow he is the best of bowmen even among the gods
with Vāsava [Indra], but with weapons cast down, he can be killed in
battle by men (nṛbhiḥ). (1305*)128
Casting aside dharma, Pāṇḍava, adopt a means to victory
(āsthīyatāṃ jaye yogo) so that Droṇa of the golden car may not slay us
all in battle. Upon the death of Aśvatthāman, he will cease to fight,
I think. Let some man tell him (kaścid asmai śaṃsatu mānavaḥ) that
that one is slain in battle.” (68–69)

Ganeri argues, based on the interpolation recognized by the Critical Edition (the
“star passage” inset above), that it is possible without the insert—that is, from the
CE’s reconstituted text—to take Kṛsn ̣ ạ as recommending “only that Droṇa be dis-
armed and not that he be killed” (2007, 66). This is true. But who would take it
that way? Not Bhīma, at least. When Bhīma reports to Yudhiṣṭhira that he has
killed an elephant with the same name as Droṇa’s son, he begins, “As soon,
O king, as I heard of the means (upāya) by which the high-souled Droṇa might
be slain, I immediately slew a mighty elephant . . . named Aśvattthāman”
(7.164.100–101a). Bhīma understands the disarming of Droṇa to be Kṛsn ̣ ạ ’s indi-
cation of the “means” to kill him, and Yudhiṣṭhira says nothing to correct Bhīma.
Arjuna, however, could—rather wishfully and, if so, I think mistakenly—take
Kṛsṇ ạ to be implying that Droṇa might be spared. If so, it might help to explain
a delayed reaction we will note on his part, which comes with his wish to see
Droṇa spared at his very end. The Critical Edition thus allows us to see that Kṛsn ̣ ạ
could very well be couching his pitch in ambiguous words that Arjuna could find
acceptable. But Ganeri ignores an important point: Kṛsn ̣ ạ begins this base sug-
gestion by addressing Arjuna. In effect, by making it look like Arjuna will be
involved, he is rescuing him from the charge of not pitching in. He addresses
Arjuna throughout,129 but cleverly widens his appeal to “let some man tell him”
at the end, since it is easy to anticipate that Arjuna will not agree to any of this.
If Saṃjaya has just implied that Arjuna and Yudhiṣṭhira are crooked warriors,
̣ ạ would be urging someone else than Arjuna to be that man.
Kṛsn
Arjuna does not approve Kṛṣṇa’s counsel, the rest approve, Yudhiṣṭhira
with difficulty, whereupon Bhīma goes off and kills the elephant named
Aśvatthāman (7.164.70–71). It would seem that the others would know what
Bhīma does, but apparently they do not! Bhīma kills the elephant, it seems, to
give himself an excuse to tell Droṇa that “Aśvatthāman” is dead, unless it is in

128. The record on this verse is not as good as Ganeri 2007, 66 indicates, where he says it appears in “all except
the Kaśmīrī.” By no means do all manuscripts in the other scripts have this verse. “All” cannot mean “all mss.”
129. At line 68, he is still addressing Arjuna as “Pānd ̣ ạ va,” although some northern manuscripts make it
plural.
470 dharma

anticipation of presenting Yudhiṣṭhira with a way to remove his “difficulty.”

But we probably cannot have Bhīma seeing that far ahead or being that glib
about his older brother’s veracity.
Bhīma says to Droṇa, apparently without others in earshot,130 and “with some
bashfulness,” “Aśvatthāman has been slain”—on which Saṃjaya comments:

That elephant named Aśvatthāman having been thus slain, Bhīma

spoke of Aśvatthāman’s slaughter. Keeping the true fact within his
mind, Bhīma was then uttering a falsehood. (72–73)

Saṃjaya does make Bhīma sound a little crooked, but what Bhīma says to Droṇa
is unimpeachable. Droṇa, however, does not believe Bhīma’s “highly disagree-
able” words and returns to fighting the Pāñcālas.
Seeing Droṇa acting “for the non-existence of the Kṣatriyas,” the Seven
Sages (in one aspect, the seven stars of the Big Dipper), Agni, and other celes-
tials, desirous of leading Droṇa to the world of Brahmā, now say he is acting
contrary to dharma; he should not continue such highly cruel karma since he
is a Brahmin enjoined by the dharma of truth. He should lay down his weapons.
Droṇa becomes downcast and asks Yudhiṣṭhira whether his son is “not slain
or slain,” as he had always hoped for truth from Yudhiṣṭhira since he was a
boy (7.164.72–96). Not mincing words,131 he puts Yudhiṣṭhira on the spot not
only to tell the truth and to tell it now but to say that Bhīma is not a liar.
Attentive listeners like Dhṛtarāṣṭra could now think back and recall a prior
scene where Yudhiṣṭhira’s truthfulness first came under question. At the
climax of the gambling match in Book 2, when Draupadī asked whether
Yudhiṣṭhira had bet himself before he bet her, which she knew he had,
“Yudhiṣṭhira made no reply, whether good or ill” (2.60.9). When Duryodhana
called for his answer, knowing that a “no” would be a lie and a “yes” would
imply one, Bhīma and Arjuna took opposite positions. Bhīma, showing
deference to his elder brother, a “guru,” stood up for Yudhiṣṭhira’s truth, saying
that as the “lord” and “owner” of all the Pāṇḍavas’ merits and austerities,
Yudhiṣṭhira was within rights to wager Draupadī even after he had wagered
himself (2.62.32–33). Arjuna, however, stirred up the howls of jackals, which
finally drowned out the tense debate, by suggesting that Yudhiṣṭhira would no

130. Ganeri 2007, 66 mistakenly has it that Bhīma kills the elephant and “then announces to Yudhiṣṭhira
that Aśvatthāman has been slain.” As verse 72 indicates, he says this only to Droṇa. Presumably this is a slip on
Ganeri’s part, but indicative of his attempted widening of the case against Yudhiṣṭhira.
131. One of the fine insights in Ganeri’s studies is his recognition that Droṇa’s insistence is inquisitional, and
that this raises the question of whether Droṇa has a right to the truth he asks for (2005, 194–95; 2007, 84, 91): “he
is ‘trading upon’ Yudhiṣṭhira’s virtue . . . Droṇa, we might say, is trying to be a ‘free-rider’ on Yudhiṣṭhira’s goodness.
Yudhiṣṭhira simply beats him to his own game (for, as Bhīṣma remarks at 12.110.26, ‘One who uses illusion (māyā)
should be met with illusion, one who is good should be met with goodness’)” (2005, 195; 2007, 84).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 471

longer have been master of anyone once he had bet and lost himself (63.21).132
Now, facing Droṇa’s question, Yudhiṣṭhira is once again well aware of the
demands of truth, but this time he cannot rescue anything by keeping silent.
But if he listens to the author and puts noncruelty before truth, he does have
the choice of the lesser of two cruelties. It is better to slay Droṇa than to make
Bhīma a liar. Each side of this equation must be weighed. Droṇa is on a terrible
rampage killing Draupadī’s kin, the Pāñcālas, and is doing so with a somewhat
illicit Brahma-weapon. Meanwhile, Bhīma is not only Yudhiṣṭhira’s most loyal
brother; he has just lost his son.
It is here that my second (Ganeri takes Arjuna’s perspective) and third (he
ignores Yudhiṣṭhira’s distinction as a king) overriding differences with Ganeri
come back to the first (he fixates only on truth). Ganeri maintains that
Yudhiṣṭhira

lacked the “reflective understanding” of the practice of truth that

would have permitted him either to endorse the lie as Kṛṣṇa did, or to
follow Arjuna and reject it. He knew that it was good to be truthful,
but he had no insight into the framework of correlative virtues that
go to makes [sic] sense of truthfulness as a good in itself. Kṛṣṇa
managed to impart some of that understanding to Arjuna, but had
less success with Yudhiṣṭhira. (2005, 200; cf. 2007, 74)

It is quite the opposite. Yudhiṣṭhira would not “follow” Arjuna, and it is not a
question of truthfulness as “a good in itself.” Yudhiṣṭhira gets his insight into
a framework of correlative virtues that subordinate truth to noncruelty, which
is successfully communicated to him not by Kṛṣṇa but by the author, who tells
Yudhiṣṭhira he should “keep considering” (vicintaya) them as his recourse to
dharma.
Yet Kṛṣṇa now tells Yudhiṣṭhira “truly” that if Droṇa fights for half a day
Yudhiṣṭhira’s army will be annihilated:

Save us, honored one, from Droṇa. Falsehood would be better than
truth. By telling an untruth for saving a life, one is not touched by
sin. (7.164.99)

Kṛṣṇa is not too credible when he says that Droṇa can now do in half a day what
he had put off for fourteen, but his subordination of truth to saving life is
consistent with other passages in the Mahābhārata,133 and can be squared here

132. On their positions in this debate, see Hiltebeitel 2001a, 257–59; also Klaes 1975, 45.
133. See especially 8.49.20, where Kṛṣṇa tells Arjuna his own view is that “not to slay living creatures
trumps everything,” even truth (8.49.20). See Matilal 2002, 27 and passim; Hiltebeitel 2001a, 206–7; Ganeri
2005, 193–95; and chapter 1 § C on the scene where Kṛṣṇa says this and tells Arjuna the story of Bālaka.
472 dharma

with Yudhiṣṭhira’s chosen value of noncruelty, just highlighted by Vyāsa.

Yudhiṣṭhira does not act directly on Kṛṣṇa’s prompting but only after Bhīma
fills in what he has done. As soon as Bhīma heard “the means,” he killed an
elephant named Aśvatthāman: “I then went to Droṇa and told him, ‘Aśvatthāman
has been slain, Brahmin! So stop fighting’” (102). We get the impression that
this is news to both Kṛṣṇa and Yudhiṣṭhira. Bhīma then tells how Droṇa did
not believe him and urges Yudhiṣṭhira to tell him the same:

As you are desirous of victory, accept the words of Govinda. . . . In

this world of men you are reputed to be truthful. (103–104)

Kṛṣṇa is still asking, ostensibly, that Yudhiṣṭhira tell a straight lie, even if it is
done “with difficulty.” It is what Bhīma has done that gives Yudhiṣṭhira the
added urgency to do it “with difficulty” now. While Bhīma is simply telling the
truth as to what he did, he encumbers Yudhiṣṭhira with a further complication:
What to say about the elephant!
At this turning point, according to Saṃjaya, Yudhiṣṭhira’s lie will have been
triply prompted—by the words of Bhīma and Kṛsn ̣ ạ , and because of “destiny”
(7.164.105). The crucial next verse includes what Yudhiṣṭhira says to Droṇa:

Sinking in fear of untruth but addicted to victory, Yudhiṣṭhira

indistinctly said, “King [he] is slain, the elephant” (tam atathyabhaye
magno jaye sakto yudhiṣṭhiraḥ/avyaktam abravīd rājan hataḥ kuñjara
ityuta; 106)134

Whereupon Yudhiṣṭhira’s chariot touched the ground (107).

The translated verse is elliptical and uncertain. It would be interesting if
Yudhiṣṭhira really calls Droṇa “King” (rājan) here135 (alternately, the voctive
rājan could be addressed to Dhṛtarāṣṭra, which is probably more metrically
apposite, but then even more minimalist in what Yudhiṣṭhira is left to say136).
If so, Yudhiṣṭhira would have to be reminding Droṇa of two disagreeable truths
he is not asking about: that long ago Droṇa displaced Drupada as king of half

134. Ganeri 2005 184; 2007, 67, 82 has avaktavyam (“not to be said”) instead of avyaktam (“indistinctly”)!
I cannot trace this hypermetrical substitution to either the Vulgate or the Critical Edition apparatus. It enables him to
translate, “Yudhiṣṭhira, equivocating, spoke out of turn and said ‘Lord, [he] is slain, the elephant.” I take it that he
translates avaktavyam as “spoke out of turn” while adding “equivocating.” Yet apart from this, Ganeri still registers
that it may be “a sort of mental restriction” that “has Yudhiṣṭhira add indistinctly ‘the elephant’. . .in a voice calculated
not to carry” (2007, 82).
135. Ganeri takes the vocative rājan as “Lord” rather than “king” (see previous note).
136. See Smith 2009, 447: “Deeply fearful of lying, but longing for victory, O king, he spoke thus to Droṇa.
‘Aśvatthāman is slain,’ he said; then, in an undertone, ‘the elephant.’” Translating the Vulgate (7.190.55), which is
identical, Ganguli [1884–96] 1970, 6: 448 ignores the vocative: “Fearing to utter an untruth, but earnestly desirous
of victory, Yudhishthira said that Aswatthaman was dead, adding indistinctly the word elephant (after the name)”
(translator’s emphasis); so too do Biardeau and Péterfalvi (1986, 2: 141).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 473

of Pañcāla, and has now killed him, while continuing to slaughter the Pāñcāla
warriors. But it is the more disagreeable word that implies the death of his son
that Droṇa is asking about. Here Ganeri’s second approach to this episode has
widened the possibilities of interpretation. Noting that Yudhiṣṭhira does not
mention the name “Aśvatthāman” here, which some—including Ganeri and
myself137—have left the impression that he did, Ganeri proposes a quite inge-
nious explanation that Yudhiṣṭhira’s word for “elephant”—kuñjara, which can
be a name for the aśvattha or fig tree, ficus religiosa—could be an “affectionate”
or “playful” name for Aśvatthāman (2007, 83):

On this interpretation, Yudhiṣṭhira wants it to be the case that what

he actually says is that the elephant is slain, something that is true.
He wants Droṇa to think that he has referred to Aśvatthāman using a
word that is synonymous with one that sounds like the name of
Droṇa’s son. This second possibility would have the advantage of
distancing him from Bhīma’s much cruder deception. Yudhiṣṭhira’s
utterance now resembles the words of an oracle in the way they leave
the final responsibility for their interpretation with the listener.
Yudhiṣṭhira intends Droṇa to think that he is engaging in a rather
subtle wordplay. . . . The utterance is deceitful, but it is much less
clear that it is also untrue; the deceit rests instead in the surreptitious
use of conventional implicature. (Ganeri 2007, 83)

All this is very clever, and not to be entirely ruled out. But there are too many
problems to make it more than a long shot. First, I know of no evidence that
Aśvatthāman is ever called “kuñjara.”138 Second, it is not likely that Yudhiṣṭhira
would be affectionately playful or that he would engage in subtle wordplay
about the death of someone’s son. If, as we have just seen, he is calling Droṇa
“king,” he is being quite straightforwardly brusque and unpleasant. Further,
Yudhiṣṭhira does not want to distance himself from Bhīma’s assertion; he
wants to save Bhīma’s reputation. But most fatally, unless it could be substan-
tiated otherwise,139 Droṇa would not have heard the “indistinct” word kuñjara,
and thus could not have understood what Yudhiṣṭhira supposedly intended by it.

137. See Ganeri 2007, 82 and n. 16, citing Hiltebeitel [1976] 1990, 251 n. 16, taking it that such a reading
would imply the “possibility . . . that the reader is meant to provide it.” Ganeri 2005, 190: “His infamous muttered
utterance of the word ‘elephant’ in apposition with ‘Aśvatthāman is dead’”; Cf. Fitzgerald 2004a, 702: “saying the
word ‘elephant’ under his breath after the name Aśvatthāman”; Smith 2009, 477, as cited above, n. 136.
138. See Sörensen [1904] 1963, 435: in the Mahābhārata, it is a name only of a serpent and a minor prince.
139. Although Ganeri supplies an aberrant and unattestable text for Yudhiṣṭhira’s words that might make
this interpretation possible (see n. 134 above), he does not make the connection between this text’s translation
and the question of what Yudhiṣṭhira would have left audible. Without some such connection, the kuñjara-
āśvattha explanation is impossible.
474 dharma

At least on his own later testimony, Yudhiṣṭhira says, “I acted falsely by saying
‘elephant’ under my breath” (kuñjaraṃ cāntaraṃ kṛtvā mithyopacaritaṃ mayā;
Mbh 12.27.16ab; see above § E).
Whatever Yudhiṣṭhira means and however Droṇa takes it, Droṇa now
hears Yudhiṣṭhira’s disagreeable word seconded by Bhīma, who also tells him
that Brahmins should stick to their own jobs (svakarma) rather than engage in
fighting; reminds him that as a man learned in brahman, he should know that
“nonviolence (ahiṃsā) to all creatures” is the highest dharma; and says he
should not doubt Yudhiṣṭhira’s word (165.28–32). Becoming dejected, Droṇa
restrains his celestial weapons, and Dhṛṣṭadyumna rushes forth to behead him
(46–49). At this point Saṃjaya tells Dhṛtarāṣtra that Arjuna was saying
(uktavān)140 to Dhṛṣṭadyumna,

“bring the preceptor alive, don’t slay him. He is not to be slain, not to
be slain,” and so too your [Kaurava] troops [said the same]. Calling
out with compassion Arjuna ran toward him, and when Arjuna cried
out, so did the kings on all sides. (7.165.50–52)

Soon, telling Aśvatthāman what happened, his maternal uncle Kṛpa confirms
this appeal by Arjuna, and that he was joined in it by the Kauravas (165.122–23).
And when Arjuna recounts the scene amid his regrets and recriminations, he
says, “When I was crying out mightily as one eager for his preceptor
(ācāryagṛddhini), the disciple [Dhṛṣṭadyumna] killed the guru not heeding his
svadharma” (167.41). In all these descriptions, Arjuna’s reaction is a delayed
one that comes just after Droṇa’s beheading. But it reinforces the impression,
mentioned earlier, that Arjuna, unlike Bhīma, might have taken Kṛṣṇa to be
saying Droṇa could be spared.
There follows, as just indicated, a recounting of Yudhiṣṭhira’s lie to
Aśvatthāman, who had been elsewhere during his father’s beheading. Kṛpa
tells him things that he would have had to have got wind of from news crossing
the battle lines. He first tells Aśvatthāman what Kṛṣṇa said:

“Dispensing with dharma, O Pāṇḍavas, protect victory so that

Droṇa’s golden chariot may not slay all of you in battle. When
Aśvatthāman is slain, this one would not fight. So I think. Let some
man falsely tell him that he is slain in battle.” Hearing these words,
Kuntī’s son Dhanaṃjaya approved them not, but all others approved,
Yudhiṣṭhira with difficulty. (165.110–12)

140. The past active participle could register this as something Arjuna was saying while it was happening
or had already happened. At best, it seems, he would have said and done all this too late.
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 475

Other than that it leaves no doubt at all that Kṛṣṇa intends Droṇa’s death, it is
much as in Saṃjaya’s account. Next Kṛpa goes straight to what Droṇa heard
from Bhīma: With some shame, Bhīma told your father, “Aśvatthāman has
been slain!” But Droṇa doubted it and asked Dharmarāja whether you were
“slain or not slain in combat” (113–14)—again, not mincing words. Now Kṛpa
gets to the elephant, which it seems that he was able to hear Yudhiṣṭhira
mention even though Droṇa presumably did not, unless the story has gotten
around:

Sunk in the fear of a lie, addicted to victory, Yudhiṣṭhira said of

Aśvatthāman, “This one is slain,” saying also “Elephant.” . . . Then,
having approached Droṇa, he said this loudly: “He for whom you
bear weapons, looking upon whom you live, your ever beloved son,
he, Aśvatthāman, has been felled.” (115–16)

Having heard that disagreeable word, Droṇa became dejected, restrained his
celestial weapons, and did not fight as before. Of cruel karma, Dhṛṣṭadyumna
then rushed and killed him (117–18).
In closing, Kṛpa seems to embellish on the bare minimum of what
Yudhiṣṭhira is said to have first said, closing off any subtleties and reinforcing
the cruelty that threads through the whole scene: “Though thus forbidden by
the Kauravas and Arjuna, your father was slain with cruelty” (165.123). He
excuses Arjuna of cruelty while blaming mostly Dhṛṣṭadyumna. As to
Yudhiṣṭhira, he says he mentioned the word “elephant” as if he thought men-
tioning it lessened the lie. The added words Kṛpa gives Yudhiṣṭhira about
Droṇa’s “beloved son” being “felled” look like a lie and a cruel one at that.
Kṛpa’s report has fuzzy features, and is aired among Kaurava partisans.
After Droṇa is killed, Dhṛtarāṣṭra raises the question of blame and Saṃjaya
reports various incriminating speeches. My position is that we should not over-
esteem the words of Arjuna, who is seconded by Sātyaki, a kinsman of Kṛṣṇa,
or underrate those of Bhīma, seconded by Dhṛṣṭadyumna. As to Yudhiṣṭhira,
we should once again not underestimate his silence. First, Aśvātthāman and
Dhṛṣṭadyumna, speaking in the opposed camps, frame the discussion of cru-
elty, truth, and blame. Aśvatthāman, based on what he has learned from Kṛpa,
calls Yudhiṣṭhira a cruel liar: “Having heard about the ignobility of the very
cruel son of Dharma,” he calls Yudhiṣṭhira a “flag-waver of dharma” or hypo-
crite (7.166.19); this “‘Dharma king’ made the teacher surrender his weapons
by fraud” (25–27). Later, he says, “Yudhiṣṭhira, resorting to the garb of dharma,
caused the preceptor who was fighting to release his weapons” (170.5–6). On
the other hand, Dhṛṣṭadyumna fully exonerates Yudhiṣṭhira. Asking which of
the six jobs of Brahmins Droṇa ever did (7.168.22), he says,
476 dharma

Departing from his svadharma and resorting to kṣatradharma, this

doer of low deeds kills us by an inhuman weapon (amānuṣeṇa . . .
astreṇa). Calling himself a Brahmin, he summoned an illusion
(māyā) of an unendurable kind, and by an illusion has he today been
killed, Arjuna, what is improper in this? (24–25).

Yudhiṣṭhira just responded in kind with a counteractive illusion. The second

illusion must be Yudhiṣṭhira’s lie, which he uttered fully cognizant of the moral
difficulty in following through on the advice of the master illusionist Kṛṣṇa.
Yudhiṣṭhira’s lie produced an illusion to Droṇa that upheld the truth of Bhīma
and slowed down the unendurable slaughter of the Pāñcālas.141 Indeed, for
Dhṛṣṭadyumna, Yudhiṣṭhira did not lie, and to have killed such a warmon-
gering Brahmin was virtuous:

The eldest Pāṇḍava is not untruthful (nānṛtaḥ pāṇḍavo jyeṣṭho)! I am

not unlawabiding, Arjuna. The sinner slain was a disciple-hater
(śiṣyadhruṅnihataḥ pāpo)! Fight, let victory be yours! (39)

The charge of disciple-hater—emphatically set in the last line of an adhyāya—

could lead Arjuna’s thoughts, or at least those of other listeners, back to
Ekalavya. Meanwhile, it begins to look like Dhṛṣṭadyumna appreciates
Yudhiṣṭhira’s linguistic improvisation as something that would fit Vyāsa’s
advice to put noncruelty before truth.
In this episode, then, of the Mahābhārata’s three “highest dharmas,”
noncruelty remains the noblest ideal for a king, even trumping truth in
Yudhiṣṭhira’s current dilemma. Nonviolence is mentioned only once, when
Bhīma wants Droṇa to lay down his weapons and reminds him that he
should consider it “the highest dharma”—for Brahmins. But in the same
context, the Seven Sages had told Droṇa that a Brahmin should be respon-
sive to the dharma of truth. The Mahābhārata refuses to absolutize any of
these virtues. At best, nonviolence is a noble ideal that can be implemented
only selectively, as when a Brahmin may be urged to relinquish his weap-
ons.142 Vyāsa could not have recommended it to Yudhiṣṭhira in the heat of
battle as he did noncruelty first and truth fifth. Yet truth should override

141. Another strength of Ganeri’s two studies is his demonstration that the “clash of illusions” (2007, 63)
has been a developing theme throughout the Droṇaparvan. See Ganeri 2005, 181–82, 188–89, 195.
142. The Mahābhārata puts its most memorable twist on ahiṃsā into the mouth of the “dharmic hunter”
who has become a virtuous meat-salesman in the Pativratā-Upākhyāna, whose assessment Yudhiṣṭhira hears in
the forest: “Surely what was said by those astonished men of old was, ‘Nonviolence!’ Who in this world does not
harm living beings? . . . no one in the world. . . . Even ascetics devoted to nonviolence surely do violence, although
by their effort it may be lessened” (3.199.28–29). After the war, Arjuna agrees with this, telling Yudhiṣṭhira that
nonviolence is not only impossible but delusory (12.15.20–28).
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 477

noncruelty for the pure Kṣatriya Arjuna where noncruelty allows him to
neglect his vows and evade battle out of his “repugnance” to fight his guru.
The tensions between these values continue to unfold as each of the three
oldest Pāṇḍavas has his final say about the killing of Droṇa. Arjuna leads off,
charging Yudhiṣṭhira, “The teacher was told a deceit by you, honored one, as a
means of gaining the kingdom. Done by one who knows dharma, this is a very
great adharma” (7.167.33). The interpolated verse cited earlier in which Arjuna
compares Yudhiṣṭhira’s lie to Rāma’s killing of Vālin comes here, after which
Arjuna continues with the charge that Ganeri features in his second article:

You told our teacher that “the kuñjara is slain”; this being but a
falsehood wearing the truth as an armour-skin. . . . It was even a
disciple who, casting off the eternal dharma, slew his own preceptor
who was filled with affection for him, while indeed the preceptor was
possessed by grief and unwilling to fight. . . . When I was crying out
mightily as one eager for his preceptor, the disciple killed the guru
not heeding his svadharma. When most of our lives has gone by and
the remainder is shorter, this great adharma done is now its disfig-
urement. Always like a father in goodheartedness, he was surely like
a father according to dharma. (41–43)

Arjuna has a kind of tunnel vision: Droṇa’s goodheartedness may extend to the
Pāṇḍavas but is hardly to be credited toward the Pāñcālas, not to mention
Ekalavya, whom Arjuna was not long ago reminded of. Despite going on to
note the great wealth Droṇa attained while serving the Kauravas (45), Arjuna
says he laid down his weapons like a Muni or silent sage (50).
When no one said a word, unpleasant or pleasant, Bhīma answered Arjuna
“as if in derision (utsmayann iva),”143 debunking him for speaking about dharma
“like a Muni” himself, “like a Brahmin of rigid vows whose weapons are laid
down” (168.1–3). Bhῑma implies that Arjuna is identifying too much with
Droṇa, who was not a forest Muni but an urban hanger-on144 and was compro-
mised when he was in the forest training Arjuna—probably another reminder
of Ekalavya. This impression grows as Bhīma further reminds Arjuna that, as
a Kṣatriya, he should rescue others from wounds rather than dwell on
noncruelty, and show dharma, fame, and prosperity to those who are good (4):

What good luck, unfallen one, that your intellect is constantly on

noncruelty! But while you were conducting yourself according to

143. Bhīma is “angry” (kruddaḥ) here; this is more than just a “smile.” See MW 183 for ut-smi and derivatives.
144. See Biardeau 1981a, 83–87 on Droṇa’s greed and “search for riches” first in the Pañcāla capital, then
with the Kauravas.
478 dharma

dharma, the kingdom was seized in accord with adharma, and

Draupadī, roughly handled, was led into the assembly by enemies,
and we were exiled to the woods. . . . Formerly you said, “We will go
to battle to the utmost and strive according to our power.” But today
you reproach us. You do not desire to know your svadharma or even
your own deceitful speech. . . . A law-abiding man, you profoundly
misunderstand adharma. (5–16)

Bhīma accuses Arjuna of varṇa-confusion like Droṇa’s. If he is the impeccable

Kṣatriya he is supposed to be, he is, as before, not one to be putting noncruelty
before the truth of his word in battle. But now the contrast is not with Yudhiṣṭhira
but with Droṇa as a mock-Muni.
Arjuna is rarely mocked like this. All became silent while he cast sidelong
glances and sighed tearfully (7.169.7). “Likewise, Yudhiṣṭhira, the twins, Kṛṣṇa
and others were very ashamed” (8).
There seems to be a fine line separating what Saṃjaya says on his own
because he is speaking in the pro-Kaurava orbit of Dhṛtarāṣṭra—for example,
that Dhṛṣṭadyumna’s words and deeds are cruel and crooked—and what he
reports as “fact” accessible through his divine eye, which would include the
spoken views of others. After Dhṛṣṭadyumna and Sātyaki nearly come to blows,
Yudhiṣṭhira finally speaks for the first time since whispering “kuñjara.” As
Aśvatthāman releases the terrible fiery Nārāyaṇa weapon, Yudhiṣṭhira airs his
sense of the recent recriminations. Seeing his fighters listless, in despairing
words that brim with stinging rebukes for what he has had to hear particularly
from Arjuna, and also with seeming sarcasm for Kṛṣṇa who has likewise said
nothing since he urged the lie:

Dharma’s son said, “Fly away with the Pāñcāla army, Dhṛṣṭadyumna.
Sātyaki, you too go to your homes, surrounded by the Vṛṣṇis and
Andhakas. Dharma-souled Vāsudeva will also make himself safe. He
is competent to instruct the world, what more himself! . . . Let the
desire (kāma) of this Bībhatsu succeed quickly with regard to me.
The preceptor of fine conduct has been felled by me in battle; . . . he
who, with his knowledge of the Brahma-weapon, scrupled not to fell
the Pāñcālas and their roots, who had exerted themselves for my
victory; . . . he who did supreme high goodheartedness to us—(now
that he is) slain, for his sake I will die with my kinsmen.”
(170.26–36)

Yudhiṣṭhira slants Arjuna for his unique accessibility to Kṛṣṇa, who leaves
Yudhiṣṭhira so often in the dark. He accepts fault in bringing about the death
 two dharma biographies? rā m a and yudhiṣ t ̣ h ira 479

of this guru, but paints Droṇa’s “goodheartedness” to the Pāṇḍavas, which

Arjuna had so touted, as a fraud. All the while pointedly saying that the current
impasse is in some way an outcome of Arjuna’s desire, he recalls matters that
would sting Arjuna most: Droṇa’s parts in killing Arjuna’s son Abhimanyu
and in suiting up Duryodhana in invulnerable armor so that he could protect
Jayadratha, the ally blamed for Abhimanyu’s killing, whom Arjuna had vowed
to kill in a day or enter fire if he failed (170.31–33). In mentioning Droṇa’s cru-
elties, however, Yudhiṣṭhira ties them back only to Book 2, recalling, as noted,
that Droṇa said nothing as the Kauravas were about to lead Draupadī into
slavery (32). In not recalling Droṇa’s and Arjuna’s cruelty to Ekalavya, which
occurred in Book 1, Yudhiṣṭhira could be subordinating truth to noncruelty
even in what he does not say to his harsh younger brother.
Finally, as Arjuna duels with Aśvatthāman, it is Arjuna’s turn to say sur-
prising things, insulting Aśvatthāman while vaunting Dhṛṣṭadyumna. Hearing
about this, Dhṛtarāṣṭra recalls that Arjuna and Aśvatthāman have great respect
and affection for each other, and asks, “Why then these never before said
harsh words from Bībhatsu?” (7.172.8ab). Saṃjaya’s main answer is that
Arjuna was

pierced to the vitals by the words of Yudhiṣṭhira, and when an inner

breaking was produced, having been reminded of that grief, lord,
Bībhatsu’s wrath arose from that grief as had never been before.
Therefore, he addressed Droṇa’s son unworthily, coarsely, disagree-
ably, the preceptor’s son who was worthy of honor, harshly like a
lowly man. (10–12).

Arjuna is brought to behave as a lowly man, a kāpuruṣa, just as we have seen that
he will do again two days later, at least according to Karṇa, when he shoots Karṇa
from advantage. We would seem to need our chain of displacements going back
to Ekalavya to explain how it is that Yudhiṣṭhira’s present words bring such lowly
words from Arjuna, who was earlier so eager to denounce him.

F. Rāma and Yudhiṣṭhira: Some Comparative Points

In being reminded of Karṇa’s charge that Arjuna would be a “lowly man” in

shooting him at a disadvantage, the thought might have recrossed some readers’
minds that we should go back to comparing the sins of Rāma and Arjuna rather
than those of Rāma and Yudhiṣṭhira. The killings of Vālin and Karṇa are
certainly an iconic and structural pair. Both are deemed lowly when shot at a
480 dharma

disadvantage. The opponents in each case have the same divine parents, but
crisscrossed: Arjuna and Vālin are sons of Indra; Karṇa and Sugrīva are sons of
Sūrya. An incarnation of Viṣṇu in each case provides the justifications.
Whatever else that comparison might lead us to by itself, it is illuminating for
the mirror it holds up to the comparison of Rāma with Yudhiṣṭhira.
One thing it tells us is that both epics take pains not to portray either king
as a lowly man. Another is that the reasons for sensing affinities between these
acts of Rāma and Yudhiṣṭhira lie not in the acts themselves but in the language
used in accusing them. According to their detractors, each is a dharmadhvajin,
one who bears the banner of dharma unrighteously (Rām 4.17.18; Mbh 7.166.19).
A third is that unlike either Yudhiṣṭhira or Arjuna, only Rāma feels called upon
to defend himself. Fourth, only Yudhiṣṭhira admits guilt, that is, admits him-
self accountable as a sinner. To do something with difficulty is to do so having
considered the consequences and been willing to suffer them. Whether the
word “elephant” came out impulsively because of an inner inclination to truth,
or because he improvised an illusion of truth that satisfied some verbal scruple,
he can be said to have recognized the difficulty and put noncruelty before truth,
as the author had told him to do.
Rāma, of course, does not put noncruelty above truth or have his words
ever challenged. He also gets away with numerous cruelties unchallenged,
unless it is by Vālmīki and Sītā, or vainly by Vālin; and he is never called upon
to weigh truth against these cruelties. It just trumps them. We hear about him
as he hears about himself, only from Vālmīki. With Yudhiṣṭhira, his cruelties
and noncruelties are for all to weigh, including himself. If extenuating circum-
stances should lighten Yudhiṣṭhira’s sentence, the loss of his chariot’s air
cushion and a brief agonizing look at Hell before he finally reaches Heaven at
life’s end seem just about right. I assume that these are effects of an “author
function” that can be distinguished from a “bardic function.”
With Yudhiṣṭhira we have a rare kind of man in the Indian epics. The pre-
vailing idea, which Rāma and Arjuna, or better, Rāma, Arjuna, and Kṛṣṇa,
exemplify to the hilt, is that the knowing Self knows itself not to be accountable
for staining acts. The compound “of stainless acts,” reserved for men, is used
mainly for these three characters, and for Rāma in both epics. What is striking
about Yudhiṣṭhira is that even though he hears of this no-fault clause in stories
told to him while he is in the forest,145 when it comes to thinking through mat-
ters in terms of dharma, as is his wont, he seems to have no use for it.
Yudhiṣṭhira’s dharma biography comes from within. Unlike Rāma’s and like
the Buddha’s, it is one of ongoing reflection.

145. Once again, it is in the Nala- (3.74.16) and Pativratā- (3.198.51) Upākhyānas. See Hiltebeitel 2001a, 235.
10
Draupadī and Sītā
Dharmapatnīs of Two Different Kinds

This chapter will compare Draupadī and Sītā as dharmapatnīs, “lawful

wives.” As we saw in chapter 8, the term occurs where Satyavatī’s
“f isher-king” father secures the fateful agreement that allows her
marriage with Śamtanu
· and def ines the “terrible” double
renunciation of Bhīṣma. And it was used several times during the
frantic last exchanges between Pāṇḍu and his two dharmapatnīs
Kuntī and Mādrī. As in these cases, “lawful wife” is used for
Draupadī and Sītā in situations of duress. Indeed, so are other terms
that might be thought to fall naturally under the heading of
strīdharma, but are likewise situational to dharma in times of distress,
as we shall see with the epics’ rare references to women’s
svadharma.1 More than any other term, though, Draupadī and Sītā’s
status as dharmapatnīs throws into relief how they are jeopardized by
what we are getting to know as these epics’ divine plans. In
Draupadī’s case, her polyandry makes her status as a “lawful wife”
uniquely problematic, and the Ṛṣi Nārada is the first to call her
a dharmapatnī (1.200.17b) when he comes to tell the Pānd ̣ ạ vas they
need a rule of rotation to keep them from fighting over her—which
leads him to tell them the Sunda-Upasunda Subtale about the two
demon brothers who killed each other over a woman fashioned by the

1. See § C. The same may be said of usages of pativratā (a woman “avowed to her husband,” a
faithful wife), which tends to be used when someone questions a woman’s fidelity, seeks to regulate
it, or offers reassurances that it is secure—a point I owe to Vishwa Adluri (personal communication),
but on which see also Dhand 2008, 31–32, 125, 160–80, 190–95; Hiltebeitel forthcoming-d.
482 dharma

gods to divide them.2 Next, Dhṛtarāsṭ̣ ra uses the term to fault Duryodhana just
after Draupadī has been abused at the dice match (2.63.5), and Draupadī, soon
finding herself in the forest, tells Kṛsn ̣ ạ , “I detest the Pāṇdạ vas, those great
strongmen in war, who looked on while their illustrious dharmapatnī̄ was
molested” (3.13.58)! Soon thereafter, Saṃjaya tells Dhṛtarāsṭ̣ ra, “The Paṇdạ vas of
boundless luster have been possessed by fury ever since they saw Draupadī, their
illustrious dharmapatnī, brought into the assembly hall” (3.46.20). As to Sītā, she
is first called “Rāma’s illustrious dharmapatnī̄ ” by the jealous and vengeful
Rākṣasī Śūrpaṇakhā, who reports on Sītā’s beauty to spur Rāvaṇa to desire her
(3.32.14). Sītā then speaks of herself in this manner while Rāvaṇa is abducting
her (3.47.28), as do the vulture Jaṭāyus when he tries to defend her (3.48.5) and
Hanumān when he is trying to find her in Laṅkā (5.7.68), and again when
Hanumān goes to bring Sītā before her unexpectedly surly husband after Rāvaṇa
and Laṅkā have fallen (6.101.39). These instances, which virtually exhaust the
epics’ usage of this term for Draupadī and Sītā, surround passages we shall com-
pare at the end of this chapter, where each heroine speaks for herself about the
revolution that has changed her life. As we shall see from the stage-setting stories
of their births and marriages, they become “lawful wives” in very different ways.
I will maintain that these contrasts are intentional. In the Mahābhārata,
Draupadī is contrasted to Sītā directly when she and the Pānd ̣ ạ vas sit down to hear
the Rāmopākhyāna. As mentioned in chapter 9, I regard the Rāmopākhyāna as one
of Vālmīki’s sources.3 This chapter will be able to put some new evidence behind
that idea, for we will be able to see some of Vālmīki’s innovations in constructing a
purposively different kind of heroine. My springboard to thinking about this con-
trast has been a title given to a forthcoming collection of my essays centered on
Draupadī in the epic and in her cult in Tamilnadu: When the Goddess Was a Woman.4
That title can alert us to one of the first things that distinguish the two heroines.
Much as the Mahābhārata poet Vyāsa wants us to understand that Draupadī lives
her life as a woman, her power and the troubles she and others experience around
that power can be said to relate to divine mysteries that work through her human
body arising from the fact that she is the incarnation of the goddess Śrī, “Prosperity.”
As far as the two epics are concerned, the Goddess becomes a woman prominently
only in the Mahābhārata. As we saw in chapter 8 with Gaṅgā, Draupadī was not the
first goddess to become a woman in that epic. Indeed, one might even say it runs
in the family. But she is the one at the heart of the story.
Sītā is a different kind of woman. Brahmā does finally tell Rāma that Sītā is
Lakṣmī when he discloses Rāma’s divine identity to him after Sītā has gone

2. See chapter 9 nn. 41 and 62.

3. See chapter 9 n. 2 and Hiltebeitel 2009a.
4. I thank Vishwa Adluri, one of the book’s two editors, for conceiving of this title (see Hiltebeitel 2011b).
 draupadī and sī tā 483

through her fire ordeal (Rām 6.105.25). But he leaves any implications unstated.5
In the story of Rāma and his brothers being incarnations of Viṣnu ̣ , there is
nothing about Sītā (Rām 1.14–17). The Rāmāyaṇa offers no myth linking the
births of Sītā and Rāma such as Vyāsa himself narrates in the Pañcendra-
Upākhyāna to link the divine births of Draupadī and the Pānd ̣ ạ vas with Viṣnu
̣ ’s
double incarnation as Kṛsn ̣ ạ and Balarāma (Mbh 1.189). Sītā is usually just said
to be “like Śrī,”6 and no more than that. Vālmīki wants us to understand that
even though Sītā’s life is equally caught up in divine mysteries, and that these
might even suggest that she is a goddess, she is first and foremost a very lovely
woman. In comparing the two heroines, I sometimes tell students that whereas
Draupadī is a force, “a real man-eater” as a 1980s song puts it,7 Sītā is the kind
of woman I would like to talk to.8 And in fact, that can get us to what is most
distinctive about Sītā. It is not that divine mysteries inhere in her birth but that
poetic mysteries arise from her relation to the poet. For whereas Vyāsa never
(at least as far as I can recall) says a direct word to Draupadī, Vālmīki composed
his poem when Sītā was living in his hermitage after Rāma banished her. There
she gave birth to Rāma’s twin sons Kuśa and Lava. I dare say the Vālmīki
Rāmāyaṇa is unimaginable without our imagining these unreported conversa-
tions between the heroine and the poet, who was not only her refuge during her
banishment but, one could say, the midwife of the poem and of her children.9
As in chapter 9, we will be able to see these differences best through being
aware of what the two epics have in common, which emerges from their sim-
ilar archetypal structuring. But were we to tour the frames and substories or
take the same Book-by-Book approach as we did in raising the question of biog-
raphy in chapter 9, we would find that these structuring features are not as
conducive to thinking about the two heroines biographically as they are for
their two royal husbands. Vālmīki gives unparalleled prominence to Sītā in the
Rāmāyaṇa’s frame story. Let us recall that he puts “the great adventure of Sītā”

5. Brahmā could connect it here with his questioning Rāma’s treatment of Sītā in putting her through her
fire ordeal, but he does not. See Sutherland Goldman 2001, 394 n. 30 on other intimations in the Rāmāyaṇa that
Sītā is divine. But her divinity is not foregrounded there as it is, for example, in the Adhyātma Rāmāyaṇa or
Tulsidas’s Rāmcaritmanas.
6. Sītā is compared with Śrī for instance, at Rām 1.76.17; 2.39.12; 5.14.6; 6.5.12;101.43.
7. “Maneater,” words by Daryl Hall, John Oates, and Sara Allen, music and 1982 recording by Hall and Oates.
Its refrain: “(Oh-oh, here she comes) Watch out boy she’ll chew you up, (Oh-oh, here she comes) She’s a maneater.”
Vaiśaṃpāyana compares Draupadī to a tigress (vyāghravadhūm iva; Mbh 3.248.17d) being approached by a jackal when
Jayadratha sends his henchman to find out who she is before he abducts her in the forest; see Hiltebeitel 1991b, 508.
8. What I am saying here is meant to be a different way of saying something that emerges from Richman
2001 by focusing on the tradition of “Questioning Rāmāyaṇas,” particularly the article by David Shulman on
Bhavabhūti’s Uttararāmacarita, which imagines Rāma being able to talk to Sītā even after her banishment and
last ordeal (see also Shulman 2001, 255–92). See also Hess 2001 on Hindi praśna pamphlets which collect
“questions” people ask of the Tulsi Rāmāyaṇ, and Erndl 1991 imagining Sītā conversing with Śūrpaṇakhā.
9. See Hiltebeitel 2001a, 320: “Vālmīki helped her raise these boys, who are called his children, ‘the
children of the Muni’” (munidārakau; Rām 7.84.9d, 17d, and 19d).
484 dharma

on a par with that of Rāma, and that the Rāmāyaṇa’s Preamble closes with
Rāma listening to his sons sing their story. Nothing makes Draupadī central to
the Mahābhārata frontmatter in any way comparable to Yudhiṣṭhira being
described as the trunk of the great tree of dharma. Yet conversely, although Sītā
gets to tell a few sidestories, the Rāmāyaṇa presents nothing like Draupadī’s
engagement with substories. Draupadī gets to listen to almost as many as
Yudhiṣṭhira does, but only a few of them—the “mirror stories,” maybe the
Bhaṅgāśvana-Upākhyāna, where Yudhiṣṭhira asks Bhīṣma who gets more
sexual pleasure, men or women (13.12.1), and Bhīṣma says women10—benefit
her directly, and she never has to answer a Yakṣa’s questions about them.
Insofar as the two epics share a common skein, it clearly unfolds from the
men’s adventures. The two female leads get more staggered attention at differ-
ent points and with more contrasting accents. Draupadī has children early; Sītā
very late. The defining outrage against Draupadī occurs in Book 2, against Sītā
in Book 3. From the end of war to the ends of their lives, the differences increase.
Though the presentation of their lives is more consecutive than Karṇa’s, like
him, each heroine is “the subject of a fragmented countertext . . . that the poets
leave readers to piece together from segments where she is part of the main story
and patches where she is the subject of selected memories—not only others’
memories but her own.”11 The two epics tell us enough about Karṇa or their chief
heroines to allow a scholar, novelist, or filmmaker to reconstruct their biogra-
phies.12 But that is not their purpose. Taking another tack, I will thus turn to a four-
part comparison between Draupaḍī and Ṣītā involving first, their births; second,
their marriages; third, what they have to say, each at a rare moment, about their
svadharma; and fourth, what they make of their worst situations.

A. Family Background, Birth, and Childhood

Draupadī and Sītā begin their lives as extraordinary females who are “ayonijā,”
“not born from a womb,” with both, it seems, being born from the earth.
Draupadī takes birth from a Vedic earthen altar called a vedi; Sītā from a “furrow”
(sītā̄), from which she gets her name.
Regarding Draupadī’s birth, there is a widespread popular conception
that she was born from fire, not the earth; but that is apparently not so in the

10. On the “mirror stories” in Book 3 being held up to the forest predicaments of Draupadī and the Pānd ̣ ạ vas,
see chapter 9 at n. 41. I speculate on Draupadī as listener to the Bhaṅgāśvana-Upākhyāna in Hiltebeitel 2005a, 49.
11. I switch the pronouns in this otherwise quoted passage from Hiltebeitel 2007b, 27. Cf. Sutherland
Goldman’s fine discussions of “gendered spaces” in the text (2001, 2004, 2009).
12. Ray 1995 and Divakaruni 2008 are two novels about Draupadī; Nina Paley’s “Sītā Sings the Blues” is
a 2009 film about Sītā. “Karnan” with Shivaji Ganesan is a celebrated 1963 Tamil film about Karṇa.
 draupadī and sī tā 485

Sanskrit Mahābhārata. Although vedi is a general term for an “altar” that could
have a fire on it, the kind of vedi Draupadī is born from is an earthen altar with
an inner curvature said to resemble a woman’s waist. Sacrificial implements
can be placed on this kind of vedi while they are not in use on top of a strew of
sacred grass, but it would not be a fire altar. When Draupadī is born from a
vedi she is compared to one of this kind, which has features of a woman’s
torso, tapering at the middle between wider “shoulders” and “hips”:

. . . a Pāñcālī girl arose from the middle of the vedi,

well-apportioned with limbs one ought to see, having a vedi’s waist,
a delight to the mind. . . . (Mbh 1.155.41)

The Mahābhārata poets are, however, rather indirect in telling us about the rite
that produces Draupadī (see Hiltebeitel 2001a, 186–89). The few things we
know about it include that her father, king Drupada of Pañcāla, wants a son who
will avenge him against his Brahmin enemy Droṇa, who took half his kingdom
(1.128.1–4; 154.20–22; see chapter 9). His intention is to kill Droṇa, so this
would require a nefarious rite since Brahminicide is the worst of sins. Drupada
must thus go to considerable trouble to find priests willing to perform it
(155.1–30). The rite begins when Drupada pronounces his lethal purpose; and
“at the end of the offering” (havanasyānte) his queen is summoned but told the
rite is efficacious no matter what she does. Immediately, a fire-hued son, incar-
nation of the fire god Agni, rises armed from a sacrificial fire and rides forth
on a chariot. As the thrilled Pāñcālas roar approval, a heavenly voice announces
that the boy, Dhṛsṭ̣ adyumna, will kill Droṇa (33–40). The rite’s purpose has thus
been fulfilled. But immediately, continuing from the verse mentioned above
that begins, “And also a Pāñcālī girl arose from the middle of the vedi,” Drupada,
beyond the stated end (purpose, completion) of his rite, also gets a daughter:

Dark, her eyes like lotus petals, hair dark-bluish and curling, having
taken human form clearly possessing the hue of an immortal, her
fragrance, like that of the blue lotus, wafted for a league. . . . And just
as that full-hipped one was born, a disembodied voice said: “Best
among all women, Kṛṣṇā will lead the warrior class to destruction.
The fair-waisted one will in time accomplish the work of the gods
(surakāryam).” (155.42–45)

The poet goes on to say that she has a fire-like radiance, but never that she
was born from fire. Draupadī comes gratis.13 In being born to “accomplish the

13. Brodbeck 2009b, 153–54 seems to agree with Hiltebeitel 2001a, 181–92, on the basis of a parallel
Mahābhārata narrative, that Draupadī’s birth is “superfluous.” But Brodbeck (2009a, 65–66) implies that a more
486 dharma

work of the gods,” the purpose behind her human birth is not human but
divine. As with her brother’s birth, it is an outcropping of the Mahābhārata’s
divine plan.
Just as it took a sacrificial rite to produce Draupadī, commentators on the
Rāmāyaṇa suggest that Sītā’s birth also occurred as an outcome of sacrificial
activity. Where Sītā’s father Janaka first describes her birth, the word translated as
“clearing” (śodhayatā) literally means “cleaning or purifying” and has been taken
to refer to a plowing done “for the laying of the fire of the sacrifice.” Says Janaka:

Now one time, as I was plowing a field, a girl sprang up behind my

plow. I found her as I was clearing the field, and she is thus known
by the name of Sītā, furrow. (Rām 1.65.14; see Goldman 1984, 385)

Vālmīki, however, does not link Sītā’s birth with any Vedic sacrificial narrative
such as one finds in Draupadī’s birth. Indeed, unlike Draupadī’s birth, which
is described by the Mahābhārata’s main narrator Vaiśaṃpāyana, Sītā’s birth is
never directly told in Vālmīki’s narration, but only in four characters’ recollec-
tions. First, as just quoted, Janaka recalls it to Rāma. Second, it is a question of
a “story” that Sītā is asked to recount by Anasūyā, an ascetic woman she meets
early in the forest toward the end of Book 2, who had heard the “story” but
wants to hear it from its lovely subject herself (2.110.22; 23; 111.1). Third,
Hanumān recalls it when he is telling Sītā her own story to reassure her that
she can trust him when he finds her in her captivity (5.14.16). A fourth and last
reference, however, complicates our picture.
After Rāma has slain Rāvaṇa, the Ṛṣi Agastya can finally tell Rāma who
Sītā was in her previous life. She was an ascetic woman named Vedavatī,
daughter of a law-abiding Brahmin Ṛṣi, himself the son of the chaplain of the
gods, Bṛhaspati. Vedavatī gets her name from being born from her father’s
constant Vedic recitation (Rām 7.17.8). She was much courted by gods, demons,
and other celestials, but her father had chosen her for Viṣṇu. Because of this,
a demon king killed her father while he slept, whereupon her mother joined
her father on the pyre. Making her father’s will to marry Viṣṇu her own, she did
austerities in a Himalayan forest to win Viṣṇu as her husband (7.17.1–17).
Vedavatī told all this to Rāvaṇa, who found her in her retreat. When Rāvaṇa
grabbed her by the hair, she avoided him by cutting it off. Her dying words as

normal kind of “close encounter” would have been behind Draupadī and her brother’s births. Such thought exer-
cises, however, diminish our understanding of epic genealogies with their assumption that the unions behind
them would have to have been ordinary. Cf. chapter 8, nn. 15, 32 (Brodbeck on Gaṅgā’s father); 40 (van Buitenen
on Satyavatī having been a fishing wench); Ghosh 2000, 34 on Satyavatī’s parenting, cited after n. 42; Chakravarti
2006, 260, speaking of Śaṃtanu’s “marriage to the apsara Ganga.”
 draupadī and sī tā 487

she entered fire were an act of truth by which she sought to be reborn as
“a female not born from a womb (ayonijā), a good woman (sādhvī̄), the daughter of
a virtuous man” (27). It is by this means, says Agastya to Rāma, that Vedavatī was
able to bring about Rāvaṇa’s death by appealing to “your inhuman manliness”
(vīryam amānuṣam; Rām 7.17.29) when she was “reborn among mortals on
a field that was turned by the blade of a plow, like a crest of fire on a vedi” (30).
That is, without it quite being Vedavatī’s stated intent to kill Rāvaṇa in her next
birth, in being reborn as Sītā she was able to accomplish just that by inspiring
Rāma, who, now that Rāvaṇa is slain, has the tools to understand what Agastya
means by Rāma’s “inhuman manliness.” “Inhuman” here means not just
“extraordinary”; it implies “divine.”14 Until Rāma slew Rāvaṇa, he had thought
he was human. But he learned thereafter that he was Viṣṇu, whose birth as
Rāma enabled him to kill Rāvaṇa when that demon had disdained men and
excluded them from the boon that protected him from being killed by creatures
he thought more powerful (Pollock 1984).15 Sītā is thus also born from a rite:
one of self-immolation. But in her case the stated intent is hers, and ostensibly
purely noble. Although killing a foe, and indeed a Brahmin (for Rāvaṇa is a
Brahmin), results from both rites, in neither case is killing a stated intent of the
heroine herself. Vedavatī could not have known that Viṣṇu would be born as a
man to kill Rāvaṇa. (For the intentions behind Draupadī’s birth, we must wait
to discuss her wedding.)
It is interesting that according to Agastya, when Sītā was born on a plowed
field, she appeared “like a crest of fire on a vedi.” Vālmīki would seem to have
magnified the vedi in question into a fire altar (in “Vedic” terms, either a mahāvedi
or uttaravedi). He does not get this image from the Mahābhārata’s main version
of the Rāma story, the Rāmopākhyāna, which knows nothing of Sītā being
ayonijā or of a story of her birth in a furrow. The Rāmopākhyāna accounts for
Sītā’s birth and marriage to Rāma in one and a half verses, with the meager
information that the “Artisan” god Tvaṣṭṛ, elsewhere in the Ramopākhyāna
called the “All-Maker” Viśvakarman (Mbh 3.267.41), “made” (cakāra) Sītā to be
Rāma’s beloved wife (Mbh 3.258.9–10). This suggests that it is Vālmīki who
makes Sītā resemble Draupaḍī. But whereas Draupaḍī’s divine origin makes
her complicit in the Mahābhārata’s divine plan, Sītā is born innocent of the
divine plan of the Rāmāyaṇa.

14. I cited one of many such usages in chapter 9 § E.2.c, where Dhṛṣṭadyumna, referring to Droṇa’s use
of the Brahma-weapon, says “this doer of low deeds kills us by an inhuman weapon (amānuṣeṇa . . . astreṇa; Mbh
7.168.24). See chapter 12 for further discussion.
15. Pollock 1984; see chapter 9 n. 6. Rāma has learned of his divinity by now since hearing it from Brahmā
(Rām 6.105.9–29) after his abusive treatment of Sītā at her f ire ordeal, with the information included that he was
born to kill Rāvaṇa (25–26).
488 dharma

One other feature of the births of Draupadī and Sītā opens onto matters
that relate to the agonistic oppositions that underlie each epic’s narrative.
Draupadī’s birth demonstrates the poets’ determination to identify her with a
nefarious darkness that arises from the agonistic dimensions of her birth and
is resonant in three of her names: Kṛṣṇā, “Black”; Pāñcālī, invoked at her birth;
and another name we shall come to, Yājñasenī. Pāñcālī has powerful overtones:
among them an evocation of the number five, pañca, which may predispose her
to marrying five men; and an extended meaning of “puppet” discernible in the
word pañcālika, a little doll.16 The Pāñcāla family cycle is deeply embedded in
the Mahābhārata’s central narrative, where it recalls a period in middle Vedic
culture when Kuru and Pañcāla kings ruled the “mesopotamia” where the
Ganges and Yamunā Rivers converge, and represented a ritual complemen-
tarity for the codifiers of the Vedic sacrifice.17 In contrast, Sītā’s family, the royal
Videhas of Mithilā, are of no wider interest to Vālmīki than for being a collat-
eral line to the Ikṣvākus who can supply brides to Rāma and his brothers in
a joint nuptials (Rām 1.72.14–23). As far as Sītā is concerned, Vaidehī and Maithilī
are just affectionate names for her. Putting these matters together, one can say
that Vālmīki draws on, or perhaps constructs, a folkloric birth for Sītā, and leaves
the Videhas out of the complexities of his narration.18

B. Marriage, Divine Plan, Early Signs of Trouble

It is clearly, however, their marriages toward which the stories of Draupadī

and Sītā are driving, each as part of a divine plan.
Draupadī is born to do “the work of the gods.” When Yudhiṣṭhira sees
himself and his brothers overwhelmed by her beauty and decides all five will
marry her jointly, he sees that “Pāñcālī’s winsome beauty, ordained by the
Ordainer (Vidhātṛ) himself, surpassed all other women and beguiled all creatures”
(Mbh 1.182.13). And at the end of his own life, Indra tells him:

O Yudhiṣṭhira, she is Śrī, who took the form of Draupadī for your
sake, becoming human though not born of a womb, beloved of the
world, she who smells good, born into the line of Drupada and
supported by you, fashioned by the Trident-Bearer [Śiva] for the sake
of your pleasure. (18.4.9–10)

16. See, for her most recent discussion, Biardeau 2002, 2: 358 n. 37.
17. See chapters 3 § B, 4 § B.1.d.1, 5 § B, and 7 § B.2.
18. Those complexities are known in the Mahābhārata, which makes plays on the meaning “bodiless” for
“Videha.”
 draupadī and sī tā 489

Now, from its first mention within the Mahābhārata’s frame, Draupadī’s
marriage is called a svayaṃvara (Mbh 1.2.36). As we saw in chapter 8, this is the
celebrated heroic mode of marrying in which a princess “chooses” a husband
in a ceremony that also calls on suitors to distinguish themselves. Draupadī’s
vivid marriage tale emerges from the agonistic “work of the gods” that she is
born to enable.
The five Pānd ̣ ạ va brothers are disguised as young Veda students so that
Duryodhana will not know that they and their mother Kuntī have survived his
plot to kill them in a burning house. Having settled in a town for a while to study
Veda, they begin to hear about the svayaṃvara being planned for Draupadī, who
was born not from a womb but from the middle of an altar (Mbh 1.153.1–10).
When Kuntī sees that her sons are unsettled by this story “as if they were struck
by spears,” she says it is a good time to leave for Pañcāla, where “the Pāñcālas
are heard to be generous with alms” (1.156.1–7). Since they are living off the
alms they gather, Kuntī is showing a maternal interest in her sons’ dinner. But
her usage of the term “alms” after seeing them so unsettled at hearing about
Draupadī suggests she may already be beginning to give the term a double
meaning. On the way to Pañcāla, Vyāsa drops in on them and tells them a story:
An unnamed maiden, “daughter of a great-spirited seer, with a narrow waist,
full hips, and a lovely brow—a girl favored with all virtues,” once did mortifica-
tions to ask Śiva for a husband, since, because of unnamed acts she had done,
she was “ill-fortuned” and had not found one even though she was beautiful and
good. When Śiva granted her request, she repeated it five times, upon which he
favored her to have five husbands at once. She is now the daughter of Drupada
whom they have been hearing so much about, “the blameless Kṛsn ̣ ạ . . . destined
to be your wife!” (1.156.11–157.14). Vyāsa urges them on and departs.
Once the Pāṇḍavas arrive in the Pañcāla capital, they continue their begging
rounds and take lodging with a Brahmin who maintains a potter’s workshop
(Mbh 1.176.6).19 Here we learn King Drupada’s motivation in holding a
svayaṃvara.

It had always been Yajñasena’s [that is, Drupada’s] wish to give Kṛṣṇā
to the diademed Arjuna, but he did not divulge it. Since
he hoped to search out the Pāṇḍavas, [he] had a very hard bow made,
well-nigh impossible to bend. (Mbh 1.176.8–9)

Drupada knows that Arjuna helped Droṇa deprive him of half his kingdom.
Presumably he wants his daughter to forge a Pāṇḍava–Pāñcāla alliance that

19. The potter-Brahmin is a Bhārgava—all this an unusual combination that may suggest a martial
inclination (Biardeau 1967–68).
490 dharma

would keep Arjuna from siding with Droṇa again. The name Yajñasena—
meaning “He whose army is the sacrifice”—has several uses in this sequence
describing Drupada’s paternity of Draupadī.20 As a patronymic, it gives her the
name Yājñasenī, “She whose army is sacrificial.” Both share the name’s
etymology, but she is the one for whom it has the richest implications. For her,
it first occurs when the Pāṇḍavas first hear about her:

And there, O Janamejaya, at the end of a story, the Brahmin told

of the wondrous svayaṃvara of Yājñasenī among the Pāñcālas,
of the origin of Dhṛṣṭadyumna [and]21 of Kṛṣṇā’s wombless birth at
Drupada’s great sacrifice. (Mbh 1.153.7–8)

Since the divine plan will require a great “sacrifice of battle” to be fought to
avenge Draupadī, her name Yājñasenī may spill over from the vengeful purpose
of her father’s rite to refer to the extra “work of the gods”: the destruction of the
warrior class for which she is born. For Draupadī Yājñasenī, the two armies
will be the sacrifice of battle as both sacrificers and victims.22
At Draupadī’s svayaṃvara, the feat called for is double. Contestants must
first bend and string the very hard bow Drupada has prepared, and hit a difficult
target. Kings come from all over. All find their way to an arena, and there, while
the Pānḍ ạ vas sit with the Brahmins and the crowd swells, Draupadī descends23
on the sixteenth day (Mbh 1.176.9–30). Dhṛsṭ̣ adyumna announces the challenge
and names the Kṣatriyas who have come, ostensibly24 as contestants (177).
Finally, after the Kṣatriyas have all exhibited their futility, Arjuna rises from
among the Brahmins and in the twinkling of an eye strings the bow and hits the
target. Drupada is pleased, and although he is yet to know who has won his
daughter, he stands ready to aid him with his army. As an uproar mounts from
the disgruntled Kṣatriyas, Yudhiṣṭhira and the Pānd ̣ ạ va twins beat a hasty retreat
back to the potter’s house. Draupadī, smiling, garlands Arjuna, completing the
rite itself, and begins following him as his wife (179.16–23).
Now Kṛṣṇa, Kuntī, and Vyāsa intervene in ways that complicate this happy
outcome to bring about what Vyāsa has already seeded in the Pāṇḍavas and

20. At Mbh 1.156.7; 175.7; 187.18; 190.6.

21. Śikhaṇḍin is also mentioned here: another brother who was, however, born a girl, having been Ambā
in his previous life. A sex change will allow him to avenge Ambā against Bhīṣma.
22. References to Draupadī by this name in Tamil as Yākaceṉi suggest that the Sanskrit precursor
Yājñasenī was not understood casually; see Hiltebeitel 1988, 194, 338, 392.
23. The description is avatīrṇā tato raṅgaṃ draupadī (1.176.30bc): “Draupadī then descended into the
arena”—as Couture 2001, 320 points out, the raṅga is a kind of public theatrical “stage.” His article makes an
important contribution to the background of the avatāra concept by tracing all usages of the verbal root ava-√t r̄
·
in the Mahābhārata (see chapter 12).
24. “Ostensibly,” because he mentions Kṛṣṇa. As we will see, Kṛṣṇa is not there as a contestant.
 draupadī and sī tā 491

their mother’s minds. Kṛṣṇa makes his first appearance in the Mahābhārata
plot: a modest one that it is tempting to call a cameo. He has two roles, maybe
three. One is to be the first to recognize the Pāṇḍavas (180.17–22). We are yet
to learn that he and Draupadī have some kind of unique and special friend-
ship, but since they do, he is probably there also to see that she falls into
the right hands—or better, that she does not fall into the wrong ones.25 But
most important, he gets the Kṣatriyas to stop fighting by announcing that
Draupadī has been “won according to dharma” (181.32). The fact that the
fighting stops instantly shows that Kṛṣṇa speaks, at this point in the text, with
unexpected authority. Saying “she is won according to dharma” (dharmeṇa
labdhā) makes his first word in the Mahābhārata “dharma,” and lets us
know from the first that dharma is something he speaks with authority about.
Kuntī’s intervention takes us back to the potter’s shop.26 One does not know
whether Yudhiṣṭhira and the twins are back, and have brought her any news from
the svayaṃvara. But her anxiety over Arjuna and Bhīma builds “when the time for
begging passed.” Again she has that maternal interest in the family’s dinner, but
that cannot be all she is thinking about, for she asks herself, “But still, could a
thought born from the great-spirited Vyāsa be overturned?” (Mbh 1.181.37–39).
She is recalling Vyāsa’s prediction that her five sons will marry Draupadī. While
she is preoccupied with such uncertainties, Arjuna and Bhīma arrive and,

highly pleased, announced [Draupadī], saying “Alms” (bhikṣā)!

[Kuntī] had gone inside the hut then without seeing her sons, and
said, “Enjoy it all equally.” But afterward, setting eyes on the girl,
Kuntī said, “Woe! What have I said?” (1.182.1–2)

The announcement of Draupadī as “alms” suggests either that Arjuna and Bhīma
are complicit or have a chancy sense of humor. And we have just seen that Kuntī
has Draupadī on her mind as well as food. A desiderative from the verbal root
√bhaj, “to share or partake,” bhikṣa,̄ “alms,” is literally what one “desires to share
or partake.”27 If Freud can be given a moment here, they are all complicit.

25. On the epic scenes of their friendship, see Hiltebeitel 2007a. Insofar as Draupadī is Śrī incarnate, the
Mahābhārata wants to be subtle in putting Viṣṇu’s perennial divine partnership with Śrī on deep background.
26. See chapter 8 § B, taking Kuntī’s role in this episode, and the Pāṇḍavas’ heeding her, to exemplify the
law of the mother.
27. A desiderative from √bhaj, “to share or partake,” bhikṣā is literally “desire to partake.” Tamil adapta-
tions in the medieval Villipāratam and Draupadī cult dramas revise the sexual innuendo. Arjuna says, “Look at
the girl (kanni) I’ve brought,” to which Kuntī, allegedly hard of hearing, says, “If it is a kani (fruit), all five must
share it equally” (Hiltebeitel 1988, 200; cf. 279–90 on Draupadī as peeled “fruit”). A bit further along in the Mbh,
Yudhiṣṭhira changes the metaphor to explain matters to Draupadī’s father: she is a “treasure” (ratnabhūtā), and
the Pāṇḍavas have an agreement (samaya) to share every treasure (Mbh 1.187.23–24). Leaving Kuntī’s inadver-
tency unmentioned, he still says his mind (heart, desire) follows her word: “So mother says, and so goes my mind
(evaṃ caiva vadaty ambā mama caiva manogatam). This is firm dharma, king. Carry it out unhesitantly” (29).
492 dharma

As we have seen in chapter 8, Kuntī knows the Law, and that is what
concerns her now: “Afraid of adharma and ashamed,” she takes Draupadī by
the hand and goes to ask Yudhiṣṭhira how her word is not to be made untrue,
and how Draupadī too will not incur unprecedented adharma. Yudhiṣṭhira rec-
ommends Arjuna as groom, and Arjuna recommends Yudhiṣṭhira since the
oldest brother should marry first. But he leaves it up to Yudhiṣṭhira to decide
what will best meet the needs of dharma and please the Pañcāla king. Yudhiṣṭhira
then sees how the beautiful Draupadī churns the hearts of all five, and, “remem-
bering the entire word” of Vyāsa, he said, “The lovely Draupadī shall be the wife
of all of us!” (1.182.8–15).
Drupada, for his part, is dubious until Vyāsa “by chance arrived” (Mbh
1.187.20–32). Vyāsa takes Drupada aside to tell him the aforementioned
Pañcendra-Upākhyāna, by which he “authorizes” the polyandry fully and
squares the Pāṇḍavas’ plan with the divine one. Formerly, Vyāsa begins, some
gods sat at a joint sacrifice. While detained as a participant, Yama, god of Death,
stops killing creatures. Some gods grow anxious about the proliferation of
humans and appeal to Brahmā that there is no longer a distinction between
mortals and immortals. Brahmā says not to worry, when Yama has finished, he
will have power over men. The reassured gods then go on sacrificing until
Indra sees golden lotuses floating down the Gaṅgā. These he traces upriver to
the tears of Śrī, who is weeping at the river’s source over the fall of four former
Indras. Drupada now hears that the Pāndavas and Draupadī are the five
Indras and Śrī, ordained by Śiva to become mortals and marry. Only by
“unbearable” and lethal karma will they be able to regain the world of Indra
(Mbh 1.189.1–26).
This subtale and the main story tie together: Yama will not be alone in
bringing death to the human world. The conclusion of the gods’ rite will take
place in the Mahābhārata war, which will be fought for the sovereignty of
Yudhiṣṭhira who, like Yama, is called Dharmarāja, and who is really Yudhiṣṭhira’s
father Dharma—Death, as it were, warmed over as Justice.28 Vyāsa goes on to
complete this window into the divine plan by accounting for the birth of Kṛsn ̣ ạ ,
and gives Drupada the “divine eye” (189.35–36) to see the truth of it all. He then
reinforces Śiva’s part in what is ordained for Draupadī by retelling to Drupada
the story of his daughter’s previous birth and boon of five husbands, by which
Vyāsa had earlier put the polyandry idea into Yudhiṣṭhira’s head.
The Mahābhārata’s divine plan has certainly advanced and become clearer,
if also more complicated. We note that Gaṅgā keeps a residual role. The tears of

28. See Hiltebeitel 2001a, 119–21. As discussed in chapter 9 § C, “The Yakṣa’s Questions” marks
Yudhiṣṭhira’s integration of these deadly roles and themes into his identity, confirmed by his father Dharma, as
Dharma.
 draupadī and sī tā 493

Śrī descend from the source of her earthly waters, to be traced back to Śrī by Indra
after the rite of Yama has finished. Once Śiva ordains the births of the Pānd ̣ ạ vas
and Draupadī, the gods then go to Viṣnu ̣ , who confirms all of Śiva’s arrange-
ments. Then, plucking a white and black hair from his head, Viṣnu ̣ ordains his
own incarnations as Baladeva and Kṛsn ̣ ạ (Mbh 1.189.31). Unlike the Pānd
̣ ạ vas and
Draupadī who are ordained by Śiva to lives of lethal karma, Baladeva and Kṛsn ̣ ạ
are ordained by Viṣnu ̣ to in a certain manner direct and contain their violence—as
̣ ạ has just done at Draupadī’s svayaṃvara by saying his first word, dharma.
Kṛsn
Sītā’s marriage “story” has a reputation for being similar, since both are
known as svayaṃvaras. But just as with her birth, we see evidence that Vālmīki
is innovating. As already noted, the Rāmopākhyāna takes care of Sītā’s birth and
marriage cursorily without mentioning either her birth from the earth or her
svayaṃvara. It seems that Vālmīki enhances Sītā’s story by making it more like
Draupadī’s in these early episodes, even while making Sītā herself unlike
Draupadī in ways that have to do with her relatively undivine womanhood and
the purity and monogamous morality of her marriage. Marrying Rāma certainly
draws Sītā into the text’s divine plan. But unlike Draupadī, she is not part of the
gods’ formulation of that plan, which, in the Rāmāyaṇa, accounts only for incar-
nations of male divinities. Otherwise, the Rāmopākhyāna and the Rāmāyaṇa
agree that Viṣnu ̣ descends as Rāma and that the gods led by Indra take birth
among monkeys and bears to assist in the divine plan (Mbh 3.260.4–15).
In the Rāmāyaṇa, Sītā’s father Janaka never mentions a svayaṃvara. He
says his daughter may be won by a vī̄ryaśulkā, a “bride-price of heroism” (Rām
1.65.15, 17).29 Surprisingly, just as Sītā’s birth from the earth is never directly
told in Vālmīki’s narration, but only as a “story” that certain characters recall,
the same applies to her marriage being a svayaṃvara. But whereas the birth
story is recalled on four occasions, a “svayaṃvara story” is recalled only once,
along with the birth story, during Sītā’s exchange with the wife of one of the
great Vedic Ṛṣis whom she and Rāma meet in the forest: Anasūyā, the wife
of Atri. When Sītā tells Anasūyā the story of her birth, Anasūyā’s question is
really about her having heard that Sītā had a svayaṃvara, whereupon the two
of them become the only persons to use this term for Sītā’s marriage in
Vālmīki’s whole text:

Anasūyā put a question to her about a certain story she was fond of.
“It was at a svayaṃvara, they say, that glorious Rāghava obtained you,

29. This phrase seems to denote a certain kind of svayaṃvara (Jamison 1996, 225; see chapter 8, n. 59), and
a Mahābhārata line uses its two components separately to describe the “price” (śulka) Drupada set by which a hero
(vīra) won Draupadī (1.185.23). Yet the basic term śulka, “price,” is used in marriage laws to describe the purchase
of a bride through the marriage form known as “demonic,” which is said to be worthy of Vaiśyas, not of Kṣatriyas,
since the usual bride-price is money, dhanaśulkā.
494 dharma

Sītā. This is at least the story that has reached my ears. I should like
to hear that story in full, Maithilī, exactly as it happened, in its
entirety. Would you tell it to me, please?” (Rām 2.110.23–24)

Vālmīki would seem to be drawing on a folklore for Sītā, or perhaps constructing

one—and in any case reconstructing Sītā’s story to parallel and rival Draupadī’s.
Sītā answers Anasūyā that it was her not being born from a womb that led
her father, after much worry, to think of holding a svayaṃvara for her. The
point for both Draupadī and Sītā seems to be that, even if such an abnormal
birth makes it hard to find a good match, lords of the land would want to vie for
such an earth-born bride.30 Sītā says Janaka had received an immense heavy
bow and two quivers of divine origin, and invited the kings to win his daughter
by raising and stringing it. But the kings only looked and left, unable to lift it.31
After a long time, one day Rāma and Lakṣmaṇa came, eager to see the bow; and
again Janaka brought it out. Like Arjuna, Rāma strung and drew it “in the twin-
kling of an eye,” but more than this (or less, since he did not have to shoot at
anything), he broke it (2.110.36–47). “Thereupon,” says Sītā,

. . . my father, true to his agreement, raised up a splendid water vessel,

ready to bestow me on Rāma. But ready though my father was to
bestow me, Rāma would not accept me right away, for he did not know
the will of his father. . . . So my father invited my father-in-law, . . . and
afterwards bestowed me on the celebrated Rāma. . . . And that is how I
was bestowed on Rāma, there at the svayaṃvara, and justly (dharmeṇa)
I love my husband, the best of men. (2.110.48–52)

Sītā’s story is known for a certain “simplicity,”32 but one can also feel the strains
in it. If she had a svayaṃvara, it was an interrupted, disappointing, and even
failed one between the suitors’ departure and Rāma’s arrival long after, with no
rivals remaining. Nothing is left of the bride’s “self-choice.” Unlike Draupadī,
who gets to garland Arjuna with a smile, it is not Sītā but her father who acts
for her by lifting a vessel. Unlike Draupadī, who immediately sets off with
Arjuna, Sīta has to wait until approval comes from Rāma’s father. Unlike the
Pāṇḍavas, who fit their marriage of Draupadī to the word and law of their
mother, Rāma upholds the word and law of his father. Sītā does get to say,
however belatedly, that it all came out “justly” (dharmeṇa). In the Mahābhārata,
Krishna makes that pronouncement on the occasion itself.

30. In his earlier account, Janaka links Sītā’s ayonijā birth with his setting of the vīryaśulkā as the “price”
of winning her (Rām 1.65.15).
31. In Janaka’s version, they offered a long siege (1.65.21–25).
32. See Pollock 1986, 525.
 draupadī and sī tā 495

As far as I am aware, the Mahābhārata keeps Draupadī quiet on the dharma

of her polyandry, but folklores and modern fictions give her plenty of complaints.33
Although Draupadī has not yet said a thing, we can see how the questionable
dharma of her marriage implies her complicity in its outcome no less than her
̣ ạ ’s—even though all Kṛsn
mother-in-law’s, her husbands’, the author’s, and Kṛsn ̣ ạ
actually spoke to was how she was won. One can only imagine what she was
thinking. As to Sītā, we do have her afterthoughts, and see her rather wistfully
reconciling love to dharma while rationalizing Rāma’s delay in marrying her.

C. Sītā and Draupadī on Their Svadharma

Generally speaking, there are remarkably few places in the epics where the
term svadharma is used with reference to women. The point is worth taking up
in anticipation of chapter 11 on the Bhagavad Gītā, where the term has earned
a certain renown. In the only occurrence I have found in the Rāmāyaṇa, Rāma
tells Sītā it would be her svadharma to stay behind and not do what she wants
(Rām 2.25.2), which is to accompany him during his forest exile. Sītā then gets
her way by quoting a supposed “śruti” verse on the sanctity of marriage:

When in this world a woman’s father gives her to a man by means

of the ritual waters and in accord with [his?] svadharma (adbhir dattā
svadharmeṇa), she remains his even in death. (Rām 2.26.16)

This verse is not clear whether the svadharma in question is her father’s or
hers, but before this, where it is clearly hers, the most interesting thing about
Sītā’s svadharma is that she does not accept it. Moreover, Rāma accepts her
nonacceptance (something inconceivable when it comes to his own svadharma).
Rāma seems to be making her svadharma up, or, perhaps more fairly, he is in
the position of a man having little credibility when it comes to telling a woman
what is in her own best interest (to suggest a parallel idiom).
As far as I can see, the Mahābhārata also gives Draupadī just one scene in
which to use the term. At a point when she is alone in the forest because the
Pāṇḍavas have gone hunting, she is confronted by a king who Vaiśaṃpāyana
says has approached her like a jackal approaching a tigress (3.248.17; see above
n. 7). Her visitor has come to find out who she is on behalf of King Jayadratha
Saindhava (the husband of the hundred Kauravas’ sister Duḥśālā), who has
spotted the beautiful Draupadī alone and is scheming to abduct her. Sensing
danger, Draupadī responds,

33. See Hiltebeitel 1999a; Ray, 1995; Divakaruni 2008.

496 dharma

For I am alone in the forest, ruled by my svadharma (niyatā

svadharme; i.e., as an orthodox wife), so how may I converse with
you when you are also here alone? (3.250.3; Leslie 1989, 173–74
with gloss and translation slightly modified)

Having told the messenger how guests are dear to Yudhiṣṭhira (priyātithir
dharmasutaḥ) and that Jayadratha must wait for her husbands’ return (8),
Draupadī continues to make svadharma a pretext to keep Jayadratha at bay:

Thinking (only) of [her] svadharma pertaining to guests

(atithisvadharmam), she entered that excellent hut made of leaves. (9)

Here, as Leslie’s gloss makes clear, Draupadī defines her svadharma by her
joint marital obligations to greet guests, and in the second usage it is again not
that clear whether the svadharma is hers or Yudhiṣṭhira’s. Ganguli takes it to be
his: “remembering well her husband’s character for hospitality” ([1884–96]
1970, 3: 570); van Buitenen (1978, 710) and Leslie to be hers, as I have trans-
lated it here. The eighteenth-century Tryambakayajvan, writing on women’s
dharma, takes this passage as exemplary for a woman who must speak to an
unrelated man (parapuruṣa) in her husband’s absence (Leslie 1989, 173–74).
It is worth asking after the principles behind such usages. Manu speaks of
female svadharma only once, but is certainly revealing as to one way that it may
be used. Even though Manu scorns the practice of levirate, he allows that a
woman may be “appointed in accordance with the Law specific to her
(svadharmeṇa niyuktāyām)”—that is, “in accord with her svadharma,” if her
husband “is dead, impotent, or sick” (9.64–66, 167). As formulated, Manu
completely ignores a woman’s desire or will; and, as we saw in chapter 8, several
Mahābhārata women, without hearing niyoga called a “law of their own,” find
the practice disagreeable. The rule does, however, serve to point out that Manu’s
one reference to a woman’s svadharma confines it to a context of marriage, albeit
marriage under duress. Not surprisingly, Rāma’s first usage is a little like Manu’s
in defining a woman’s svadharma by a supposed marital requirement. Though
our sample is small, it seems possible to generalize from a third example that
when men are speaking, that would be the result to expect.
Here it will be worth keeping in mind that, wherever Kṣatriya and other
non-Brahmin women are the ones concerned, they are supposed to fit their
expectations into an assortment of legally recognized forms of marriage that
the Law books define primarily for Brahmins: a constraint we met in the story
of Ambā. Her story includes another of the very few instances I have found
where the epics use svadharma with reference to women, and offers a useful
illustration of the principle just mentioned. When the powerful sage Rāma
 draupadī and sī tā 497

Jāmadagnya arrives to take up Ambā’s cause and challenge Bhīṣma on her

behalf for his part in ruining her marriage to King Śālva, he says,

Bhīṣma, what thoughts prompted you first to abduct the Kāśi king’s
daughter against her will, and again to let her go? . . . Now, because
you abducted her she has been rejected by Śālva. Therefore take her
back by my appointment (man niyogāt), Bhārata. Let the princess
regain her own Law (svadharmam . . . labhatu), tiger among men.
(Mbh 5.178.5, 7–8)

In Ambā’s case, the principle is clear: her svadharma hinges entirely on marriage
as something arranged and “appointed” by men. Whatever she may have done
out of will or love to regain her betrothed, she lost her chance for such a sva-
dharma when he rejected her—and also because Bhīṣma let her go, which, had
he not done so, would have given her an alternate svadharma as a lawful wife to
one of his nephews, for whom he had abducted her and her two sisters. The
proposed resolution that the sage re-“appoint” her to Bhīṣma is no solution at
all, since Bhīṣma has made a lifelong vow of celibacy.
Yet we have seen a counterprinciple at work when Sītā and Draupadī speak
about their svadharma themselves. In Sītā’s case it enables her to negotiate and
even change Rāma’s mind. And Draupadī uses it to maneuver and play for
time. Two more examples where women speak for themselves confirm that
there is a gender-specific counterprinciple at work. In one, the Rākṣasī Hiḍimbā
tells Kuntī that she is abandoning her friends, kin, and svadharma to marry
Kuntī’s son Bhīma (Mbh 1.143.7). It not clear what she means by her svadharma,
but, like Sītā, she is free to abandon it and it seems generous of her to do so.
Since mention of her kin suggests a reference to Rākṣasa habits, it could be that
in making a love match she renounces the exciting Rākṣasa mode of marriage
that would call for her abduction, or even that she foregoes eating Bhīma
(Wendy Doniger, conversation, April 2011). So far, though, Sītā, Draupadī, and
Hiḍimbā are still, like Rāma, Manu, and Rāma Jāmadagnya, relating their
usages of women’s svadharma to marriage. Fortunately, we have one last in-
stance on the distaff side where this is not a requirement. This is a usage by a
Yoga-specialist named Sulabhā, who tells the philosopher king Janaka,

Firmly devoted to my svadharma, I am not one who makes confusion

of dharma (na dharmasaṃkarakarī svadharme ‘smi dhṛtavratā).
(12.308.185cd)34

34. Cf. Fitzgerald 2002, 667–68: “I am not creating a mixture of Lawful Norms. I hold firm to the practices
of my own Lawful Duty.” Fitzgerald’s study is rich on this Mokṣadharmaparvan unit and Sulabhā’s background
498 dharma

Sulabhā speaks as a nun or ascetic (bhῑkṣukῑ, 308.7) and as a gender philosopher

from a nondualist Sāṃkhya position for whom all beings, whether male or
female, have bodies and components made equally of the same matter (prakṛti;
113–15). It is thus important that she speaks as a woman (see Sutton 2000,
440–43). But where she mentions her svadharma, it is to claim that she is free
(svatantrā; 138–40) from the constraints and confusions of dharma that Janaka
would like to say she has violated by possessing his body to see if he is as free
of worldly attachments as he claims. As Sutton says, Sulabhā’s response can be
recognized as “representing the critique of svadharma typical of nivṛtti” (2000,
441), by which he means a critique from the standpoint of renunciation that
can imply a critique of svadharma. Janaka does not mention svadharma, how-
ever, and accuses Sulabhā mainly of varṇasaṃkara, āśramasaṃkara,
gotrasaṃkara, and finally dharmasaṃkara (Mbh 12.308.59–62)—that is, he
charges her with “confusion” or “mixing” of caste, life-stage, lineage, and of
dharma itself. He charges her with the whole gamut, but she proves him wrong
on each count. When she finally mentions her firmness in svadharma along
with her not committing the pinnacle of these confusions, she formulates her
svadharma as a rather bold, unusual, and probably ironic expression of her
socially untrammeled philosophical standpoint.
There seems to be a considerable gap between what men prescribe as
women’s svadharma and what women can make of it as a law of their own.

D. Captivity and Exile

As we saw in section B with Anasūyā, and just now even in her brief exchange
with Rāma, Rāmāyaṇa characters also think of Sītā as good to talk to. But that
is not to disparage Draupadī. It is now time to concentrate on what they have to
say under real duress. I turn to passages where their voices are strongest—not
only in what their words recall but in what they anticipate in the fuller narra-
tives. Regarding what they have in common, I choose episodes where they first
speak out after their identity as dharmapatnīs has been brutally violated. At this
point our tracing of these epics’ divine plans moves into a discussion of dharma
and bhakti that will be a consideration through the remaining chapters. As we
know from Fred Hardy (1983, 5–9, 331–429, 527–34) and Karen Prentiss

and philosophical outlook, but he does not comment on this verse. As he points out, she is called a “follower of
yogadharma” whose positions “resemble aspects of ‘Sāṃkhya’ doctrine and Buddhist themes” (642). Her name
“means ‘easy to get,’ which has an obvious ironic pertinence to their encounter” (668). Their dialogue takes place
in Janaka’s court, where he makes a “public fuss” (654). See also Vanita 2003.
 draupadī and sī tā 499

(1999, 53), later vernacular bhakti traditions make the woman’s voice a vehicle
for devotional sentiments in the mood of love in separation. This section will
suggest that the epics lay some early ground for this.35 The two episodes are,
however, asymmetrical. The first difference is that whereas Draupadī will talk
about divinity to a husband who is with her, Sītā addresses her thoughts on
such matters to a husband from whom she is painfully apart.
Our text for Sītā finds her being held captive in a grove guarded by Rākṣasīs
and hounded by Rāvaṇa, who has given her two months to live, threatening
that if she does not come to love him the Rākṣasīs will prepare her for his
breakfast or eat her themselves. Yet Hanumān has found her. While he hides
in a tree, deciding how best to speak to her, she makes three speeches, osten-
sibly to herself (Rām 5.23–26). As Hanumān sees her make these three speeches
amid exchanges with the Rākṣasīs, it is not always clear what he actually hears,
since her words seem at times to become soliloquies, particularly as she more
and more addresses herself to Rāma. Through all these speeches, she can be
said to be “brooding,” in each with variations on √cint, “to think.” Going beyond
Manu, she is a woman thinking for herself.
One can see Sītā’s deepening despair through her exclamations and interjec-
tions. She makes five appeals using words whose overlapping range can be trans-
lated “oh” or “alas.”36 In the first speech, she calls out by name to Rāma, Lakṣmaṇa,
and their two mothers: “‘O Rāma!’ and then again ‘Oh Lakṣmaṇa!’ ‘Oh Kausalyā,
my mother-in-law!’ ‘Oh Sumitrā!’” (Rām 5.23.11cd). The rest all occur in her third
speech, which she begins, “Alas, virtuous folk in the world have a popular saying
that there is no such thing as an untimely death. Alas, it must be true if I, who
lack all merit, have managed to survive even for a moment under such abuse”
(26.3). Next, she says, “Alas, the two months allotted me . . . will soon elapse,”
and, in a passage we shall revisit, again calls on the four mothers plus her own
mother as well (26.7–8). Finally, she saves her last call for Rāma only, and with
rather clear bhakti overtones—“Alas, Rāma of the long arms, true to your vows!
Alas, you whose face rivals the full moon! Alas, you benefactor and beloved of the
living world! You do not realize that I am to be slain by Rākṣasas” (26.11)—from
which point she launches her closing apostrophe to him. She also makes three
“curses,” one in each brooding: first, cursing the human state (23.20); second,
cursing herself as “ignoble and unchaste” to have survived even a moment
without Rāma (24.6); and then cursing life itself, which she would abandon had
she poison or a weapon to do so (26.15–16).

35. For fuller “mapping” of bhakti in the passages treated in this section, see Hiltebeitel 2011a, chapter 11.
36. Hā, “an exclamation expressive of pain, anger, astonishment, satisfaction (= ah! alas! oh! ha!)” (MW
1,296); bata, “an interjection expressing astonishment or regret, generally = ah! oh! alas!” (MW 718).
500 dharma

Let us look at these three speeches for the points where Sītā addresses Rāma
directly, even though he is not there. One translating team, inclined toward bhakti
excisions, has cut Sīta short, removing the end of the second speech and all of the
third (Brockington and Brockington 2006, 208–10). But the devotional overtones
have direction, and run throughout. Be it noted that while translators, quite sen-
sibly, have wanted to keep Sītā talking on an intelligible human and wifely plane,
it has meant undertranslating certain loaded words: most notably the impossible-
to-translate ātman, “self,” but also words that I will translate with reference to
gratitude,37 pity, compassion, lordship, power, abandonment, and belovedness38
so as to bring out her speeches’ bhakti overtones. I am not saying this is the “right”
way to translate these terms, just that we may trace a devotional thread through
them that is intertwined with the ruptured domestic thread of Sītā and Rāma’s
marriage, and also her thoughts on Rāma simply as a human king.
As Sītā’s first speech (23.11–20) opens, she recalls a popular maxim quoted
by paṇḍits on how death comes only at its appointed time (12–13), and grieves
that her death will come now separated from Rāma, whom she then dwells on
until this train of thought ends:

This pitiable woman, whose merit must be small (alpapuṇyā kṛpaṇā),

like a woman without a lord (anāthavat), must surely perish, like a
laden vessel struck by strong winds in the midst of the ocean. Unable
to see my husband and come under the Rākṣasīs’ power (vaśam),
I am collapsing under my grief like a riverbank undercut by water.
How fortunate are those who are able to see my lord (nātham)—his
eyes like the inner petals of a lotus—who walks with the valorous gait
of a lion and is yet grateful (kṛtajñam), a speaker of what is beloved
(priyavādinam). Separated from Rāma who knows himself (rāmeṇa
viditātmanā),39 there is no way that I can survive any more than if
I had consumed virulent poison. What kind of sin did I commit in
a former body that I obtain such cruel and terrible suffering?40

37. The term in question will be kṛtajña (5.23.16d; 24.12a; 26.12d). See MW: “knowing what is right, correct
in conduct, Mbh xii.104.6; acknowledging past services or benefits, mindful of former aid or favours, grateful,
Mn, Yājñ, etc.” Goldman and Sutherland Goldman prefer “accomplished.” See chapter 9 n. 10.
38. The little word priyam, “dear,” which I will continue to translate in that fashion, carries a big load in
these speeches. In addition to instances cited below, see 5.24.7: “What desire can I have for life or happiness
without my beloved (priyaṃ vinā), that lord/husband (bhartāram) of all the sea-bounded earth whose word is
beloved (priyaṃvadam)? I shall cast off my body; let them cut it up or eat it. For without my beloved (priyavarjita),
I cannot long endure this suffering.”
39. Goldman and Sutherland Goldman 1996, 180, “celebrated”; Brockington and Brockington 2006, 207,
“sagacious.” Viditātman recurs at the end of Sītā’s second speech at 5.24.49.
40. On ruminations about hardships following from former karma being, in the Mahābhārata, “a trait
especially of women characters,” see Hiltebeitel 2007a, 112–14.
 draupadī and sī tā 501

Engulfed by this great grief, I want to abandon life. Guarded by these

Rākṣasīs, I will never see Rāma again. A curse on this human state!
A curse on being under another’s power (paravaśyatām). Although
I wish to, I cannot end my life. (23.14–20)

Sītā imagines the good fortune of others who might see Rama, without yet
saying who they might be. She builds up a shaky image of him. She thinks he
“knows himself,” but he cannot really know his divine nature until he has killed
Rāvaṇa. She grieves at being under “another’s power,” which in Manu defines
lack of freedom.41 Ostensibly she is held captive by the Rākṣasīs and Rāvaṇa.
But she intimates that she is under some still higher power: her own karma? a
lord’s who should be grateful, who should know himself?42 As her own imagin-
ings continue, Rāma will not be so perfect.
Sītā’s second and longest speech (5.24.3–49) follows a vivid depiction of
her: “Grieving like a woman possessed, or a madwoman, or a woman in a state
of utter confusion (unmatteva pramatteva bhrāntacitteva śocatī), she rolled on
the ground like a filly” (5.24.2). She “broods” further on her captivity under the
“Rākṣasa women’s power” and on her separation from Rāma (3–5), curses
herself (6), addresses the Rākṣasīs and Rākṣasas including Rāvaṇa while predict-
ing their downfall and Laṅkā’s destruction (11–25),43 comes back to herself (35),
and ends on the theme of being under “Rāvaṇa’s power” (49). Here, she first
centers her attention on Rāma wondering why he does not come for her:

Rāghava is renowned, wise, grateful, and compassionate (kṛtajñaś ca

sānukrośaḥ). Therefore I think it must be the exhaustion of my good
fortune that has made this man of good conduct uncompassionate

41. M 4.159, as cited in chapter 5 § B in the context of Manu’s discussion of ātmatuṣṭi.

42. For a powerful study of this whole sequence, see Sutherland Goldman 2001, 223; 232–38, for whom
“the passage represents the lowest emotional ebb of the epic” (237). Sutherland Goldman makes this case from
the standpoint of Sītā’s words at Rām 5.23.10: “how wretched to be under the power of another (dhig astu
paravaśyatām).” As Sutherland Goldman points out, the words are “ambiguous, since the word ‘para’ could refer
equally to Rāma, Rāvaṇa, or to the rākṣasī guards” (233). Her favored reading, noting that the commentator
Govindarāja says the same, is that Sītā is referring to Rāma; but Sutherland Goldman takes this only in
Govindarāja’s sense that it refers to a wife’s societally structured inability to make decisions, such as killing her-
self, without her husband’s permission (233–34; 237; 394 n. 36). I doubt that Vālmīki’s Sītā is as straight-jacketed
as this commentator’s. Sutherland-Goldman recalls that “Pollock has argued convincingly” that Rāma’s divinity
“is fundamental to the epic narrative” (335 n. 30)—referring to two of Pollock’s discussions of this matter (1984;
1991, 51–52), but not his exploration of the importance of the secret of Rāma’s divinity as a sustained narrative
theme (1984). I will be arguing that Vālmīki presents Sītā’s “lowest emotional ebb” (including these words)
within the tensions sustained by this secret, and, moreover, that the whole passage repeatedly brings
this tension into play and alludes to its eventual unfolding.
43. Brockington and Brockington, as usual inclined toward bhakti excisions, think Sītā originally would
break off precisely here, and that the poem does not resume until 5.25.1, with the Rākṣasīs’ outrage at Sītā’s resis-
tance. They also remove 5.26, containing Sītā’s third speech entirely, along with 5.27 (2006, 208–10). This means
that of the four segments we are discussing, they leave only the first two.
502 dharma

(sadvṛtto niranukrośaḥ).44 For why has he who singlehandedly

annihilated fourteen thousand Rākṣasas in Janasthāna not come for
me? This Rāvaṇa, who holds me captive, surely has very little
strength. Surely my husband is capable of killing him in battle. Why
then has Rāma, who slew in battle that bull among Rākṣasas Virādha
in the Daṇḍaka Forest, not come for me? Granted, it is difficult to
assault Laṅkā, which is situated in the middle of the ocean. Still,
there is nothing in the world that can stop the flight of Rāghava’s
arrows. Why has Rāma, so firm in his valor, not come to rescue his
cherished wife, who is carried off by a Rākṣasa? I think Lakṣmaṇa’s
older brother must not know that I am here. For if that mighty man
knew it, would he then endure this outrage? (24.12–18)

Her key verse here is the first one, bringing Rāma’s compassion into question
around the thought that he might become uncompassionate now that she has
been abducted—as he will in fact be in imposing her two ordeals on her. Sītā
then multiplies these anxieties in this second speech’s closing:

How, in my great suffering, shall I do without him—without seeing

my beloved Rāma, the corners of his eyes bloodred? . . . Rāma must
not know that I am alive. For if he and Lakṣmaṇa knew, it is impossible
that the two of them would not scour the earth for me. Surely
Lakṣmaṇa’s heroic elder brother has gone—out of grief for me—from
here to the world of the gods, having abandoned his body on earth.
Fortunate are the gods, Gandharvas, Siddhas, and supreme Ṛṣis who
can now see Rāma, my lotus-eyed lord. Or perhaps this wise royal Ṛṣi
Rāma who loves dharma and is the Supreme Self (paramātman) has
no use for me as his wife. There would be love for the one that is
seen; there is no affection on the part of one who does not see.
Ingrates (kṛtaghnāḥ) destroy; Rāma will not destroy.45 Is it that I am
completely devoid of qualities, or is it just the exhaustion of my good
fortune, that I, Sītā, should be without Rāma, who is deserving of the
best? It would be better for me to die. . . . Or perhaps the two
brothers, best of men, have laid down their weapons and are wandering
in the forest as forest dwellers, subsisting on roots and fruits.
Or perhaps Rāvaṇa . . . has slain the heroic brothers Rāma and

44. Cf. Brockington and Brockington 2006, 208: “The Rāghava is famed for his wisdom, reason, and
tenderness, but I suspect that this disastrous misfortune of mine has turned his righteousness to ruthlessness.”
Goldman and Goldman 1996, 420 n., however, recognize that the verse is about “compassion.”
45. Basically Goldman and Sutherland Goldman’s proposed “literal translation” of this “highly elliptical”
verse 41 (1996, 421).
 draupadī and sī tā 503

Lakṣmaṇa by means of some trick. At such a time as this I can wish

only to die. . . . Fortunate, indeed, are those great-souled,
great-fortuned Munis who are revered for their truth, their selves
conquered, for whom there is neither beloved nor unbeloved.
Homage to those great-souled ones who detach themselves from
both! Abandoned here by my beloved Rāma whose self is known
(rāmeṇa viditātmanā), and fallen under the power of the wicked
Rāvaṇa, I shall end my life. (24.35–49)

The passage widens the scope of Sītā’s wavering. While imagining that Rāma
may have gone to heaven out of grief for her, or that he and Lakṣmaṇa have
relinquished their weapons and become forest wanderers, or that Rāvaṇa could
have killed them by some trick, she broods twice on the “fortunate” celestial
Ṛṣis or Munis in tellingly contrastive terms. First she imagines them greeting
Rāma in heaven: “Fortunate are the gods, Gandharvas, Siddhas, and supreme
Ṛṣis who can now see Rāma, my lotus-eyed lord” (39). But then she brings
them down to the circumstances of her own despair: “Fortunate, indeed, are
those great-souled, great-fortuned Munis who are revered for their truth
(satyasammatāh),̣ their selves conquered, for whom there is neither beloved nor
unbeloved. Homage to those great-souled ones who detach themselves from
both!” (47–48). According to commentators, the term satyasammatāḥ “suggests
the sages’ capacity to transcend saṃsāra” (Goldman and Sutherland Goldman
1996, 422). Sītā cannot aspire to this kind of detachment. Moreover, in the
midst of all this, she has had the disturbing yet also penetrating thought that
“perhaps this wise royal Ṛṣi Rāma who loves dharma and is the Supreme Self
has no use for me as his wife” (40). Commentators have not missed this verse,
one view being “that Rāma, as the paramātmā and as one whose mind consists
of dharma, does not require a wife [to assist him in practicing dharma]” (Goldman
and Sutherland Goldman 1996, 421; cf. Sutherland Goldman 2001, 235). Sītā brings
her thoughts on this anxiety to some complex and subtle questions. If, as we
have noted, the fruits of dharma can often be unseen, what of Rāma and Sītā’s
love when she is unseen? Raising the image of Rāma as a potential ingrate,
who, to the extent that he does “know himself,” would “love dharma” more
than her, she reassures herself, perhaps wishfully, that Rāma “will not destroy,”
yet asks whether her own “qualities” (guṇas) are such as to have left her now
without him. Indeed, Rāma will not involve Sītā in his ritual duties after the
war. Her wifely dharma will be cast aside because he feels he must acknowl-
edge the public’s doubts about her time with Rāvaṇa.
Between the second and third speeches a good Rākṣasī named Trijaṭā
describes a dream that augurs well for Sītā. Among the auspicious tidings,
504 dharma

she says, “I saw that lotus-eyed woman rise from her husband’s lap to stroke the
sun and moon with her hands” (25.15); and, while identifying other good omens
that show Sītā will soon hear welcome news, she says, “This lady (devī) does not
deserve to suffer, and she is the one I saw standing in the sky” (33). The celestial
vision certainly carries an intimation that this devī is a more than ordinary lady.
Sītā does not hear Trijaṭā describe her dream, and her third speech carries
on her lonely soliloquy with mixed apostrophe to Rāma (5.26.3–16), ending at
the point where even she senses the favorable omens that presage the presence
of Hanumān. Again, there is a vivid transitional description: “Surrounded by
the Rākṣasī women and cruelly menaced by the words of Rāvaṇa, timorous Sītā
lamented, like a little girl abandoned in the midst of a desolate wilderness” (2).
Her soliloquy now shifts from ślokas to doubled triṣṭubh verses, intensifying her
grieving tone.46 As in the first speech, she begins with a popular saying: “Alas,
the good speak this truth in the world: that there is no such thing as untimely
death. Alas it must be true if I, who lack all merit, have managed to survive even
for a moment under such abuse” (3). Surely suicide would not be a fault when
“I am about to be killed by this one unbeloved in my sight, to whom I could not
give my affection any more than a twice-born could teach mantra to a non-
twiceborn” (5).47 For the rest (6–16), it is her concluding apostrophe to Rāma:

Surely if Rāma, the lord of the world (lokanātha), does not come,
the king of Rākṣasas will soon dismember me with sharp knives, as
a surgeon might a fetus in the womb.48 Alas, the two months
allotted me, who have already suffered so long, will soon elapse!
Then it will be for me, just as it would be for a thief, imprisoned
and condemned to death for a crime against the crown, on the
morning of his execution. Oh Rāma! Oh Lakṣmaṇa! Oh Sumitra!

46. See Sutherland Goldman 2001, 236 and 394 n. 41; J. Brockington 1998, 275, “containing Sītā’s
renewed lament”; 390, “5.26–27 are again entirely in longer meters and show a high proportion of long com-
pounds and similes—obvious signs of expansion in the second stage of growth.” Brockington does not address
5.24.38–49, where the bhakti is clearest, and never really correlates bhakti with any specific stage, but rather typ-
ifies it by varying degrees of being “late,” while also perhaps covering it with such overviews as, for stage 5, “after
the [Mahābhārata] epic was committed to writing, the number of manuscripts needed for the purpose seems to
have become a virtual library of Indian tradition” open to “new material of all sorts” (20–21; cf. 159). Reflecting
similarly on the Rāmāyaṇa (379 ff.), he sees both epics as having their “particular character . . . in part due to their
position at a time of transition . . . brought to an end by the transition from an oral to a written tradition” (27).
47. Sītā compares herself here to the Veda, playing on a recurrent theme brought up shortly before her
three broodings in a famous passage where Hanumān looks down upon her from his tree, sees her emaciated,
and reflects “with uncertainty: for she seemed barely discernible, like some Vedic text once learned by heart but
now nearly lost for lack or recitation” (5.13.36). Indeed, this could be an “unconscious” theme on her part, since
Rāma eventually learns that Sītā, in her previous life as Vedavatī, got that name from being born from her father’s
constant Vedic recitation (7.17.8). See § A.
48. Cf. Goldman and Sutherland Goldman 1996, 58, citing Rām 5.17.1–3 where Sītā, at the first sight
of Rāvaṇa coming to see her in the grove, “curls up in a fetal position, trembling like a plantain tree in a gale.”
 draupadī and sī tā 505

O mother of Rāma and my own mother as well! I, this luckless

woman, will perish like a ship foundering in a storm at sea. Those
vigorous sons of the lord of men must have been killed on my
account through the strength of that creature in the form of a
deer. . . . It must have been Time itself in the guise of a deer who
deluded hapless me at that time when, fool that I am, I sent away
my husband [and his] younger brother. . . . Alas, Rāma of the long
arms, true to your vows! Alas, you whose face rivals the full moon!
Alas, you benefactor and beloved of the living world (hā jīvalokasya
hitaḥ priyaś ca)! You do not realize that I am to be slain by Rākṣasas.
My taking you for my sole divinity (ananyadevatvam), my long
suffering, my sleeping on the ground, and my rigorous adherence
to dharma—this devotion to my husband has been fruitless
(viphalam), like the favors men do for ingrates. Surely this dharma
adventure of mine (dharmaś carito mama) has been vain and my
exclusive devotion to my husband useless. For, pale and emaciated,
I cannot see you; I am cut off from you without hope for our
reunion. Once you have carried out your father’s orders to the letter
and have returned from the forest with your vow accomplished, you
will, I think, make love with wide-eyed women, carefree, your
purpose accomplished. But as for hapless me, Rāma, after having
loved you so long, given you all my heart—to my own undoing—
and practiced my vows and penances in vain, I shall abandon my
accursed life. (26.6–15)

Those who see Rāma, always potentially a bhakti idiom, are no longer the
celestial denizens but wide-eyed lovers. We see how Sītā keeps dharma and
bhakti at play along with all the strands of her predicament before finally
letting them unravel in her version of a wife’s worst-case imaginings, well in
tune with Rāma’s overriding concern for his father’s truth, if not when imag-
ining his infidelities. Sītā also deepens the implications of her first invocation
of Rāma, Lakṣmaṇa, and their mothers, this time mentioning all four again
but adding “and my own mother.” For this, she uses an appropriately different
term for “mother,” jananī, “begetter,” or “genetrix,” hinting that she is speaking
of the Earth, to whom she will finally appeal at her last ordeal when she will
really end her life.49 But for now, as she determines to hang herself from her

49. Note how Sītā’s next words, coming right after the invocation of her mother, may compare herself to
the earth as a ship tossed at sea: “I, this luckless woman, will perish like a ship foundering in a storm at sea.” This
would parallel a strong image of Draupadī after her gambling, who, according to Karṇa, “became the salvation
(śānti) of the Pāṇḍavas! When they were sinking, boatless and drowning, in the plumbless ocean, the Pāñcālī
506 dharma

ponytail,50 her thoughts turn more fondly to Rāma as she too becomes aware
of the good omens noticed by Trijaṭā and takes hold of the branch of a great
flowering tree (18–20).
I now turn to a scene early in Draupadī and the Pāṇḍavas’ forest exile. I will
maintain that Draupadī is portrayed, during a single conversation where it is
particularly apt, somewhat surprisingly as a materialist. By this I mean that we
can read behind a story she tells and some positions she takes that the epic
poets have cast her, during this exchange, as expressing herself in terms that
call to mind, though of course rather subtly,51 associations with the Cārvāka or
Lokāyata philosophy that was considered a heresy by both Brahmanical and
Buddhist authors during the classical and medieval periods. Some of her views
could also be called “empiricist,” but “materialist” is the key term in the history
of Indian philosophy to allow us to keep her story and her views in focus and in
context. I do not propose that we are meant to think of Draupadī as a materialist
beyond this episode. But once we see how she is portrayed in it, we will be able
to see that the episode has built upon traits of hers found elsewhere in the
Mahābhārata that lend depth to her materialist turn of thought while she airs
it. As will be seen, it will at times be appropriate to qualify her views as “quasi”-
materialist, but that will not, for my argument, undercut their materialist base.
Insofar as she might be a materialist in this episode, we might also ask whether,
along with that outlook, she is taking a kind of time out from bhakti.
Let us begin with a brief account of Johannes Bronkhorst’s findings on the
early Materialists. Bronkhorst argues that although a doctrine is attributed to
them which holds that there is no consciousness beyond the four elements of
earth, water, fire, and air, the main thrust of what singled them out was an
opposition to the new doctrines of karma, rebirth, retribution, and “another
world.” For some time they joined early Mīmāṃsakas in opposition to all but
the last of these doctrines, and before the two came to be depicted as oppo-
nents, both would have been defenders of Vedic Brahmanism against these

became the Pāṇḍavas’ boat to set them ashore (Mbh 2.64.2d–3). Both are likely evocations of the image of the
heroine as boat and as the imperiled Earth sinking under her burden, as in the Mahābhārata’s avataraṇa myth
(see Biardeau 2002, 1: 385).
50. Rām 5.26.17. Sītā wears her hair in an ekaveṇī (5.18.8; 55.7; 57.12, etc.), a term that means literally a
“single-plait braid” but, in contrast to braids of three or more plaits, actually means an unplaited “braid” or a
(long) ponytail such as virahiṇīs or women separated from their husbands conventionally wear in classical
Sanskrit literatures. See Hiltebeitel 1980–81, 198. Goldman and Sutherand Goldman offer “woven braid” (1996,
189; Sutherland Goldman 2001, 237).
51. In this, it differs from the portrayal of the Rākṣasa demon named Cārvāka who, in Brahmin guise,
denounces Yudhiṣṭhira after the war and is summarily executed by the syllable “Hum!” (Mbh 12.39).Yet
Draupadī’s counsel does bear some affinity, positionally, with that of Jābāli in Rām 2.100, who presents a Cārvāka
or Lokāyata view in advising Rāma not to go to the forest at all—that is, he should not renounce the throne. See
Pollock 1986, 511–13.
 draupadī and sī tā 507

late-Vedic “new thought” trends. Early Cārvākas would also have been urban
Brahmins drawn to royal capitals of the type reflected in the Arthaśāstra and
the Kāmasūtra—and, we might add, in the Rāmāyaṇa’s depiction of Jābāli
(see n. 51). And at some indeterminate but probably early point, they had a sūtra
attributed to Bṛhaspati,52 chaplain of the Vedic gods (see Bronkhorst 2007,
150–59, 161–62, 172, 309–10, 363–66).
Now the episode where Draupadī sounds like a materialist occurs thirteen
months into the forest exile (Mbh 3.36.31–32) that she and the Pāṇḍavas must
undertake as the outcome of the epic dice match. It presents the first conversation
she has with her five husbands, and particularly with Yudhiṣṭhira, since that
scene of humiliation and outrage for all of them, but especially for her.53At the
beginning of this exchange, she is introduced as a paṇḍitā, a “scholar,” a learned
or wise lady, or “lady pandit.”54 This is, I believe, the only time the epic describes
her in this way, which suggests that, even though it will not endorse her views,
it wants to present them as the result of some considered opinion.
Initially, Draupadī berates Yudhiṣṭhira for his lack of kingly authority and
manyu, manly wrath (3.28.19–34), and tells him a story to chide him for
exalting kṣama, patience, or forbearance (3.29).55 Yudhiṣṭḥira later returns to
the topic of manly wrath to claim a shred of it and at the same time show
its dangers, claiming that the reason he did not stop the dicing is that he lost
his head to manly wrath upon seeing how his opponent Śakuni was cheating!56
But for now, with her exasperation mounting, Draupadī introduces a new turn
with the exclamation, “Glory be to the Placer and Ordainer (namo dhātre
vidhātre ca) who have befuddled you!”57 Draupadī then tells a second story of

52. See Bronkhorst 2008, 150–53. It is not clear whether one can date the attribution earlier than the
Mahābhārata, but that does not rule out Draupadī’s making an early reference to such a tradition. See Bronkhorst
2009, 157 n. 56, translating Arthaśāstra 1.2.4–5: “‘The science of material welfare and the science of government
and politics [are the only sciences],’ say the followers of Bṛhaspati. For the Vedic lore is only a cloak for one con-
versant with the ways of the world.” Cf. Chattopadhyaya [1959] 1973, 6, 127; cf. 128, 577–78 for Vedic descriptions
of Bṛhaspati–Brahmaṇaspati that could have been taken in a materialist direction.
53. The exchange, quite lengthy, has been treated in some detail by Bailey (1983a, 150–57), Hill (2001,
168–76), Biardeau (2002, 1: 423–26, 437–44), and Malinar (2007b).
54. This, according to the narrative voice of Vaiśaṃpāyana, which introduces her as “dear and beautiful, a
scholar (paṇḍitā) and devoted wife (pativratā)” (Mbh 3.28.2)—for which van Buitenen 1975, 274 reads “beloved
and lovely, wise and faithful.” Cf. Hiltebeitel 2001a, 268; Malinar 2007b, 90; Karve 1974, 90: “hardly
a complimentary epithet in the eyes of the Kshatriyas of the Mahābhārata”—forgetting that it is the Brahmin
Vaiśaṃpāyana speaking.
55. See Malinar 2007b, 81 on the distinction early in their conversation between manyu and kṣama.
56. As Biardeau (2002, 1: 426, 440) observes, Yudhiṣṭhira soon offers what seems a new explanation of
why he did not stop the dice match: that he lost his head to this kind of anger upon seeing how his opponent was
cheating (Mbh 3.35.4–5).
57. Mbh 3.31.1 ab; thus van Buitenen 1975, 279, capturing Draupadī’s exasperation. Cf. van Buitenen 1975,
297: “Homage to the Placer and Disposer (namo dhātre vidhātre ca), go a safe and healthy path” (3.38.25ab), where she
uses the same words in her adieu to her beloved Arjuna when he sets off for divine weapons in the Himalayas.
508 dharma

sorts (it could actually be called more of an allegory), referring to it as an

“ancient tradition” (itihāsam purātanam), and citing it with an abbreviation of
the conventional formulaic atrāpy udāharanti (“as to this, they quote”) mode of
attribution that one finds also in some of the early dharmasūtras (with atha
there, rather than atra) and widely in the Mahābhārata’s postwar didactic
books.58 Curiously, given that such formulaic usages typically mention a
source for their account, Draupadī does not give one at this point for the
“ancient tradition” she cites here (3.31.20–21)—apparently leaving that for
later, as I will mention in a moment. This second story or allegory is about
how the aforementioned Placer is behind everything: “As wooden puppets are
assembled, so are these creatures, king; he makes the body and limbs move”
(31.20–22). Having heard this account, Yudhiṣṭhira then charges Draupadī
with heresy (nāstikyam; 3.32.1, 5), characterizing her view as an excessive doubt
in dharma that can lead one to be reborn as an animal (6).59 Draupadī will then
conclude her dialogue with Yudhiṣṭhira by naming her source somewhat indi-
rectly. Once, while she was doing errands and sitting on her father’s lap, she
overheard a learned Brahmin who had come to her father’s house and spoken
to her brothers on subjects that had first been propounded by Bṛhaspati
(33.56–58).
Here we have some signposts by which to recognize where Draupadī’s
views have been nurtured. The source for her “puppet speech” (see Hiltebeitel
2001a, 214 n. 106; 269) is a Brahmin who visited her father, King Drupada’s
capital, when she was a girl. Since this Brahmin comes to the Pañcāla capital,
he is certainly being depicted as an urban Brahmin, whose own source for his
“story” is ultimately Bṛhaspati. From what we have learned from Bronkhorst’s
depiction of the early Cārvākas, this is enough circumstantial evidence to look
further into what Draupadī has to say, and what Yudhiṣṭhira has to say about
what she says. I limit discussion to four things. First, I will delineate what
I believe are two dimensions of Draupadī’s alleged heresy that emerge from
this spousal dialogue. Second, to the extent that this heresy would appear to be
about a divinity (or two) whom neither speaker decisively names, other than as
the Placer and the Ordainer, we have the question of who they appear to be
talking about. Third is the question of the nature of this heresy60—that is,

58. See Tokunaga 2009a, 28, and for discussion, chapter 5 § C.

59. See Malinar 2007b, 89, noting Yudhiṣṭhira’s use of the prefix ati-, signifying what is “excessive,” “too
much,” or (in his eyes) “deviant,” as in this verse qualifying her “doubts.”
60. Malinar’s discussion of “Draupadī the heretic” (2007b, 86–88) is, I believe, too concessive to
Yudhiṣṭhira, and does not find that Draupadī has a consistent and respectable point of view. I thus do not think
that the debate “combines a symmetry of [gender] representation with a symmetry of female subordination” (91)
by, among other things, showing that “[a] nāstika queen on the throne is as scandalous as a king turned into a
deplorable weakling” (88).
 draupadī and sī tā 509

whether it is, or has at least strong shades of, materialism. And fourth, I close
with the question of whether such a materialist, or at least quasi-materialist,
outlook relates to other things the Mahābhārata tells us about Draupadī.
1. The first level of Draupadī’s alleged heresy comes out in what she has to
say in her “puppet speech” about the Placer: “As wooden puppets are assem-
bled, so are these creatures, king; he makes the body and limbs move” (22).
Carrying this idea along with several “string” similes (“like a pearl strung on a
string” among them), and remarking that creatures can be pushed along
(preritāḥ) by the Lord to heaven or hell (24–25, 27), she builds up to this:

So all beings come under the Placer’s sway. Adjoined to noble or

again wicked acts, the Lord, penetrating beings, moves them, and he
is not perceived. This body they call the “field” is the Placer’s mere
instrument (hetumātram) by which the Lord causes action that has
good or bad fruit. Behold this magical prowess as it is done by a Lord
who kills beings with beings, having bewildered (mohayitvā) (them)
with his own magic. . . . Having made a disguise, Yudhiṣṭhira, the
god Bhagavān, the Self-Existent Great Grandfather hurts creatures
with creatures. Joining them together and disjoining them, doing as
he will, the Lord Bhagavān plays (krīḍate) with beings like a child with
playthings (krīḍanakair). The Placer does not act toward beings like
a father or mother. He seems to act out of fury (roṣād). He is like
another person (yathā ayam itaro janaḥ). (3.31.28–37)

Janaḥ in the singular could also be taken collectively: “This one is like other
people,” “folk.” Let us note that even after Yudhiṣṭhira has mentioned heresy,
Draupadī does not change her tune that the body is an instrument (kāraṇam)
by which the Placer, now as the Great Lord (Maheśvara), makes powerless
beings move along (prerayati) enjoined to this or that task (33.21–22).
Of the two dimensions of Draupadī’s heresy, Yudhiṣṭhira seems least
interested in this one, probably getting to it only toward the end of his response
when he reminds Draupadī of her birth, ostensibly to illustrate the principle
that acts bear fruit:

So it is in you: recall your birth, Kṛṣṇā, as it is heard (yathā śrutam);

and you know also how the ardent Dhṛṣṭadyumna [her brother] was
born. This is a sufficient analogy, sweet smiling woman. [Knowing
that] “Of action, there is fruit (karmaṇām phalam astīti),” the wise
man is content with even a little. . . . Fruition of both meritorious and
wicked acts and their origin and disappearance are mysteries of the
gods (devaguhyāni), beautiful woman. Nobody knows them, these
510 dharma

creatures are bewildered about them. They are guarded by the gods;
surely the gods’ magic is hidden (rakṣāṇyetāni devānāṃ gūḍhamāyā hi
devatā). (32.30–31, 33–34)

Yudhiṣṭhira thus grants that beings are bewildered, but not as puppets of a
puppeteer. He rather suggests that the best karmic option for beings is to make
up a bewildered audience to the mysteries of the gods. If he is bothered by the
puppeteer heresy, it is not much. Rather, he concludes with a ringing endorse-
ment not only of the ontology of acts but of dharma and of everything the Placer
does as the “supreme deity”:

Resolving that “Everything is” (sarvam astīti), set free your heretical
heart (nāstikyaṃ bhāvam utsṛja)! Do not revile the Placer, the Lord of
the beings. Learn of him, bow to him. Let not your buddhi be like
this. Do not, O Kṛṣṇā, censure in any way the supreme deity
(uttamaṃ daivatam) by whose grace a devoted (bhakta) mortal attains
immortality! (32.38c–40)

Draupadī’s birth from an earthen altar, just after her brother Dhṛṣṭadyumna’s
birth from fire, is frequently cited as something known to the principal epic
characters. So Yudhiṣṭhira can remind her of it as one of the “mysteries of the
gods” while turning the conversation also to the workings of karma.61 We can
understand why Draupadī might have a sense that she is buffeted around like
a puppet. There would be a certain experiential quality to Draupadī’s heretical
heart or disposition (bhāva) on this point, since one of her birth-given names,
Pāñcālī, means not only a daughter of Pañcāla but evokes a word for
“marionette.”62
But if the puppeteer heresy is more Draupadī’s problem and does not
much bother Yudhiṣṭhira, it is the other way around with her second heresy,
which gets to the core of Yudhiṣṭhira’s self-understanding. Virtually the first
thing he says in response is, “My mind is beholden to dharma by its own
̣ ạ ”̄ (3.32.4). We have seen the second dimension of
nature (svabhāvāt), Kṛsn
Draupadī’s heresy taking shape in their conversation about the fruition of karma.

61. On this precise double frame of reference as the basis for Janamejaya’s last question in the Mahābhārata,
see Austin 2009 and discussion in chapter 12.
62. See Biardeau 2002, 1: 438. We also know that her birth was coordained (vi-√dhā) by Śiva and Nārāyaṇa.
This was supposed to be a “secret of the gods” (Mbh 1.189) communicated by Vyāsa to her father, but Yudhiṣṭhira
seems to be intimating the news had gotten around. He learns at the epic’s end that Draupadī was “fashioned
(nirmitā) for your [his] pleasure’s sake (ratyartham) by the Trident-Bearer” (18.4.10cd). Draupadī’s creation out of
a sacrifice by the gods has a counterpart in Tilottamā (Mbh 1.201–4), whom the gods fashion from beautiful bits
of this and that to cause a rivalry between two demon brothers so they will regulate their polyandrous marriage
with Draupadī. That the gods can fashion females as “mere instruments” to their higher ends could also be a
reading of Durgā’s birth in the Devī Māhātmyā.
 draupadī and sī tā 511

What exercises Yudhiṣṭhira is what Draupadī says last, condemning the

Placer (39)—whom she has already said “is like another person” or “other
folk” (3.31.37)—before asking a question with two possible conclusions:

Having given prosperity to the Dhārtarāṣṭra [Duryodhana] who

transgresses the noble treatises, is cruel, greedy, and unrespectful of
dharma, what fruit does the Placer eat (dhātā kiṃ phalam aśnute)?
If karma done pursues its doer and not another, then surely the Lord
is stained (lipyate) by the wicked karma he has done. Or if the wicked
karma done does not pursue its doer, then mere power is the cause
here, and I grieve for weak people. (3.31.40–42)

It is true that one would not always want to translate the idiom “phalam √aś” as
“eat the fruit” every time it is used with karma.63 But Draupadī is definitely
being literal, for even after Yudhiṣṭhira has mentioned heresy, she comes back
to this image with unmistakable clarity in her only reference to both the Placer
and Ordainer since she opened this topic saying “Glory be” to them. Having
first mentioned the example of babies sucking their mother’s breast (33.4) to
show that all beings obtain livelihood from what they do, she says:

All beings know exertion, Bhārata, and visibly, having the world as
witness (lokasākṣikam), they eat the fruit of their actions (phalam
aśnanti karmaṇām). I see that creatures live off their own total effort
(svaṃ samutthānam), even the Placer and Ordainer, as does this crane
in the water. (33.6–7)64

Draupadī’s words here show a materialist bent, for in philosophical terms, she
is, like a materialist, stating that “visible evidence” or “perception” (pratyakṣa)
is implicitly her standard (pramāṇam) for knowledge. Nor would she be contra-
dicting herself by mentioning deities, since Indian materialism does not require
atheism.65 Moreover, Draupadī may also be extending or widening her materi-
alism to include the Jain notion that one is stained by one’s karma, which,
unlike Brahmanical notions, involves a material concept of karma. She has
stood her ground against Yudhiṣṭhira’s claim that “having the world as wit-
ness” (lokasākṣikam) is for the fool who pleasures his senses and is confused
about everything else (32.16), and she has not given in to his interpretation of
karma’s fruition through hidden divine forces. Not only does she want effort.

63. Van Buitenen has “What does the Placer gain?” (1975, 281); Hill has “What advantage does the Placer gain?”
(2001, 171). Cf. “they eat the fruit of their actions” for the second usage quoted below (van Buitenen 1975, 283).
64. Draupadī’s mention of the crane here was cited in chapter 9 § D.2.c with reference to Dharma’s
pre-Yakṣa crane disguise.
65. The Vedic god Bṛhaspati, said to have authored their sūtra, would have to be “material.”
512 dharma

She wants to see results! Note that later in this exchange, Draupadī fits these
ideas to counsel that we could expect from a materialist urban Brahmin, telling
Yudhiṣṭhira that “the success of houses and cities is caused by man
(agāranagarāṇāṃ hi siddhiḥ puruṣahaitukī)” (3.33.24cd).
Such materialism would be heresy enough, and that is what appears to
bother Yudhiṣṭhira when he defends his innate bent toward dharma66 and turns
the conversation from one about the unseen fruits of karma, action, about
which he seems not to have convinced her, to one about the unseen “fruits of
dharma” (dharmaphala; 3.32.2, 4–5; cf. 29, 36–37). He defends the Placer and
dharma together (dhātāraṃ dharmam eva ca; 32.14b) for establishing the “Ṛṣis’
standard (or authority)” (ārṣaṃ pramāṇam; 20a), without which “the universe
would sink into foundationless darkness (apratiṣṭhe tamasi)” (23cd). Clearly the
text is allowing the Ṛṣis’ standard established by the Placer and dharma to
trump Draupadī’s standard of perception, but both are evoked as respectable
philosophical positions. Yudhiṣṭhira has been speaking like a good theist
Mīmāṃsaka67 in seeing divinity behind the spiritual fruits of dharma. But in
rising to such a defense of dharma, which is also an expedient self-defense,
Yudhiṣṭhira is again indirect at best in addressing Draupadī’s heresy. No doubt
he wants to assure her that since his acts flow from his innate bent toward
dharma, they would not stain him. But Draupadī has been talking about a stain
on the Placer. Who is he—at least for now, to these two speakers?
2. In mentioning mainly the Placer, Draupadī and Yudhiṣṭhira seem to
be talking about Brahmā, at least in using some of his epithets. But Draupaḍī
also uses the names Bhagavān, Īśvara,68 and Maheśvara, which have wider
and more devotional ambience;69 and Yudhiṣṭhira thinks of the Placer as

66. Kṛṣṇa confirms this svabhāva of Yudhiṣṭhira at Mbh 3.180.18, and, as we saw in chapter 9, his father
Dharma confirms that Yudhiṣṭhira is Dharma.
67. See Bailey 1983a, 155 n. 60: Yudhiṣṭhira’s view “corresponds in essence to the view maintained in the
Brāhmaṇas and amongst the Mīmāṃsākas (as the doctrine of apūrva) that the result of a sacrifice is invisible, that
it only occurs after death with the attainment of heaven.”
68. Malinar speaks of the Placer and Ordainer among “the īśvaras, the lords who rule over the fruits of
one’s efforts” (2007b, 87), going on to speak of them as the “hidden lords” mentioned where Yudhiṣṭhira says,
“the gods’ magic is hidden” (89–90, citing 3.32.33–34 [see above]). But while there are thirteen usages in the dia-
logue of īśvara in the singular, several of them explicitly referring to Dhātṛ (31.20, 24–25, 27–28; 32.27–28, 39;
33.19), the only usage of “lords” in the plural occurs where Yudhiṣṭhira is speaking about “the Ṛṣis, gods,
Gandhavas, Asuras, and Rākṣasas” as “lordly beings (īśvarāh)”̣ who follow dharma knowing that the Placer is the
“giver of [its] fruits” (phaladam . . . dhātāram) (32.27–28).
69. Bailey, Hill, and Biardeau agree that Draupadī and Yudhiṣṭhira seem to be talking about Brahmā. For
Bailey, Dhātṛ “is really Brahmā in his ordaining role”—a “conclusion strengthened by the fact that two of
Brahmā’s most common epithets—Svayambhū and Prapitāmaha—are named here” (1983a, 151 and n. 47). Cf.
Hill 2001, 176. Yet that does not account for names like Bhagavān, Īśvara, and Maheśvara. More solid is Biardeau’s
local contextualization. She thinks Draupadī and Yudhiṣṭhira may be talking at cross-purposes: she, about
Brahmā with reference to early Upaniṣadic ideas of brahman in relation to the abandonment of rites, “even if
Yudhiṣṭhira remains faithful precisely to rites” (2002, 1: 439). For Draupadī, Dhātṛ is “no doubt Brahmā,
 draupadī and sī tā 513

“the supreme deity (uttamam · daivatam) by whose grace a devoted mortal

attains immortality.” It has been suggested that when the Placer and Ordainer
are named in situations of misfortune, it is as if the speakers are reluctant to
blame their personally chosen grace-bestowing highest deity by name.70 In
such a world two characters could be talking about the same or different deities
in a way that each might or might not understand. Draupadī’s story about the
Placer comes from Bṛhaspati, who, as we have seen, gets a reputation as chap-
lain of the Vedic gods for having composed a materialist sūtra. There are also
Jain and Buddhist texts that mock the idea of a creator god.71 The Mahābhārata
does not openly cite heterodox texts (real or imaginary), but here it seems to do
so covertly, under a Vedic cover and in a woman’s voice! Draupadī is particu-
larly out of sorts at this point, and is openly fed up with Yudhiṣṭhira. But she
would also have reasons to be fed up with a divine plan that has put her
through her ordeal at the dicing and now “placed” her in the forest where she
is the most discomfited, being a woman, of all those beginning their years of
exile (Biardeau 2002, 1: 437). If she would have a complaint about the Placer,
then what about a god whom she has asked, shortly before this on Kṛṣṇa’s first
visit to the exiled Pāṇḍavas: “You are the lord of all beings, both divine and
human. Then how was it that a woman like me, wife to the Pārthas, sister of
Dhrṣṭadyumna, your friend (tava . . . sakhī), Kṛṣṇa, came to be dragged into the
hall?”72
It could thus be that Yudhiṣṭhira is heading Draupadī off, implicitly, from
speaking ill of her friend Kṛṣṇa, that grand illusionist who will tell Arjuna in
the Bhagavad Gītā, “The Lord of all beings resides in the region of the heart,
Arjuna, making all beings reel, mounted to a device (yantra) by his power of
illusion” (BhG 18.61)—which has reminded the philosopher Śankara of a

who personifies the brahman that her husband Dharmarāja has invoked here” (437–38)—referring to
Yudhiṣṭhira’s prior statement that “Patience is brahman (kṣanti brahma)” (3.30.37–41), which could well have
triggered Draupadī’s impatience. But Biardeau limits her contextualization to just this passage: “We are still not
in the problematic of the Bhagavad Gītā here . . .” (424, n. 6). Yes, the Bhagavad Gītā is yet to happen, but the point
is unrealistically preemptive as regards readers.
70. I find Hill suggestive on usages of the names Dhātṛ and Vidhātṛ “where the context is one of misfor-
tune, adversity or grief. Where these abstractions are apparent epithets for the great Gods, it is as if devotees,
reluctant to directly blame their chosen conception of the divine, preferred to have recourse to a barely disguised
epithet” (2001, 366). Cf. Biardeau (as cited in the previous note); Malinar 2007, 86: Yudhiṣṭhira “does not deal
directly with her line of thought.”
71. See Chattopadhyaya [1959] 1973, 496–97; Embree 1988, 80–82 (the ninth-century CE Mahāpurāṇa, a
Jaina text). Cf. chapter 6 at n. 9 on the Buddhist Brahmajāla Sutta.
72. Mbh 3.13.52c–53. In an interpolated verse just before this after 13.46, she says, “With Brahmā, Śaṃkara,
Śakra and so forth, with the hosts of gods again and again, you sport like a child with playthings (krīḍāse tvaṃ . . .
bālaḥ krīḍanakair iva)” (3.55*). With the same last pāda (in boldface) as 3.31.36d, the insertion (found only in the
Vulgate and one Kāśmīri ms.) looks like a backreading from 3.31.36, taking the latter to imply Kṛṣṇa. 3.55* could
be taken to say that even the gods are among the playthings, which could be implied by 3.31.36 as well.
514 dharma

puppet play.73 Yudhiṣṭhira could have reasons to hear Draupadī’s words that
way. When she was summoned to the gambling hall, menstruating and in a
single garment, he would have heard her say,

So now the All-Ordainer disposed (vyadadhāt saṃvidhātā), touching

both who are touched, the wise and the fool. He said, “In this world
dharma is alone supreme.” Protecting, he will dispose peace.
(Mbh 2.60.13)

Her words are the epic’s single mention of the All-Ordainer, who seems to
cover for both the Placer and the Ordainer, and to leave open the possibility that
she is talking about Kṛṣṇa.
In fact, there are vivid echoes between Draupadī’s “heresy” and Kṛsn ̣ ạ ’s
words with his other special friend Arjuna in the Bhagavad Gītā: the subject of
our next chapter. Like Draupadī, who protests the point, Arjuna hears that he
should consider himself a “mere instrument” (BhG 11.33) of a god on whom
everything is “strung like heaps of pearls on a string” (7.7). Indeed, she says
things that would rather defy the Gītā. Whereas Draupadī says, “The Placer does
not act toward beings like a father or mother. He seems to act out of fury. He is
like another person,” Kṛsn ̣ ạ says, “I am the father of this universe, the mother,
the Placer, the grandfather” (9.17). When Draupadī asks, “What fruit does the
Placer eat?” that might either stain him or leave people powerless before mere
power, Kṛsṇ ạ says, “I eat” (aśnāmi) whatever is offered to me with bhakti—“a leaf,
a flower, a fruit (phala), or water”—so as to free devotees from the good or bad
fruits (phala) of binding karma (9.26–28), while Arjuna sees him with crushed
heads stuck between his teeth (11.27). Kṛsn ̣ ạ would not confirm Draupadī’s prop-
osition that “the Lord is stained by the wicked karma he has done,” since he says,
“The four-varṇa-system was created by me with distinctions as to qualities and
acts (karmas). Even though I am its doer, know me as the unchanging non-doer.
Acts do not stain me (na māṃ karmānị limpanti); I have no yearning for the fruits
of acts. Whoever comprehends me thus is not bound by acts” (4.13–14).
3. As we have seen, Bronkhorst considers the materialists to have been
motivated above all by their opposition to doctrines of karmic retribution and

73. Edgerton notes, “As puppets in a puppet play, according to Śankara’s plausible suggestion” (1952,
190). Draupadī speaks of “wooden puppets” (dārumayī yoṣā; 3.31.22a); Kṛṣṇa of what is “mounted to a yantra
by māyā” (yantrārūḍhāni māyayā). Yantra itself seems to imply “puppet” where Saṃjaya tells Dhṛtarāṣṭra that
man is not independent in doing good or bad acts but “is caused to act like a wooden device (kāryate
dāruyantravat)” (5.156.14; cf. van Buitenen 1978, 475: “helplessly manipulated like a wooden puppet”). Before
this, well aware that Kṛṣṇa is the decisive ally to have on one’s side, Dhṛtarāṣṭra says, with I think suggestive
irony, “Man is lordless/powerless (anīśvara) in prosperity and adversity, like a wooden puppet strung on a
string (sūtraprotā dārumayīva yoṣā); he is (put) under the sway of fate by the Placer. . . . I cannot abandon my
son. Where dharma is there is victory” (5.39.1a–c, 7cd).
 draupadī and sī tā 515

to the idea of another world, heaven. In looking at how Draupadī handles these
issues, we see that the two dimensions of her heresy—her “puppet story” and
her inference that the Placer is stained by his own karma—are complementary.
Draupadī does have ideas about karma, but she hedges on them. In saying,
“This body they call the ‘field’ is the Placer’s mere instrument (hetumātram) by
which the Lord causes action that has good or bad fruit” (3.31.30), she draws no
connection between karma and any kind of “soul.” Then she puts her ideas on
karma in the conditional and frames them in Jain-sounding material terms:
“If karma done pursues its doer and not another, then surely the Lord is stained
(lipyate) by the wicked karma he has done. Or if the wicked karma done does
not pursue its doer, then mere power is the cause here, and I grieve for weak
people” (3.31.40–42). On karma, then, it seems that we could legitimately say
that Draupadī has some materialist reservations, which she airs in these
particular terms because she has been drawn into a theistic “proto-Sāṃkhya”
frame of reference typical of the Mahābhārata—including, of course, the
Bhagavad Gῑtā. Regarding heaven, again, she does accommodate herself to
notions of another world, saying that the Lord pushes people to heaven or hell
(31.24–25, 27). But here too she uses the very mechanical, and plausibly mate-
rialistic, imagery of the Placer as a puppeteer moving his playthings around
with strings attached. She concedes to Yudhiṣṭhira that one can talk about these
things, but he is the only one to have positive and affirmative things to say
about them. But where she concedes nothing and is consistent throughout,
and most in opposition to Yudhiṣṭhira, is that “perception” is her standard.
4. We come, then, to the question of whether this materialist side to
Draupadī rings true of other things the epic tells us about her. On the most
solid side of what Brahmanical philosophies call matter (prakṛti), she is born
from an earthen altar and given a name that implies “Puppet.”74 At the other
end of the spectrum of prakṛti, we find her associated at times with matter’s
most subtle evolute, the buddhi or “intellect.” Instances of this association
are especially pronounced in the dialogue with Yudhiṣṭhira. As has been
cited, Yudhiṣṭhira tells her, “Let not your buddhi be like this. Do not, O Kṛṣṇā,
censure in any way the supreme deity (uttamaṃ daivatam) by whose grace a
devoted (bhakta) mortal attains immortality” (32.40). From the beginning of
their dialogue, they set out their differences as a matter of two intellects, and
repeatedly use the term buddhi with nuances that range from “intellect” to
“mind” to “mental attitude,” implying the possibility of differentiating their
philosophical positions. Yudhiṣṭhira begins to meet Draupadī’s challenge by
linking her with “those of slow intellect” (mandabuddhi; 3.32.9). But soon he

74. This paragraph is indebted to, and, I would suggest, can be read beside, Butler 1993.
516 dharma

wants to change her mind: “Let not your buddhi be like this” (32.39), sig-
naling that he intends a reasoned debate, one in which not only should she
consider his words but he must be willing to weigh her charge that his mind
was overtaken by the “spirit (buddhi) of gambling” (31.18cd). He is thus intel-
lectually engaged when Draupadī moves on from his riposte to challenge
him to “rising up” or “exertion” (utthānam, samutthānam; 33.6–7, 53)75 by
arguing that “it is the mind (buddhi) to act that is extolled”—from which she
goes on to speak of the fatalist as sudurbuddhi, of very bad intellect, and the
“believer in chance” (haṭhabuddhir) as one who is disposed to just getting by
(33.11–13). Eventually she builds to her last word on the subject: she wants
him to be “resolute on his own with his intellect in front (buddhipūrvaṃ
svayaṃdhīraḥ)” (33.23c). Their differences thus include an airing of contend-
ing philosophical outlooks—fatalism and belief in chance having now been
mentioned. As a materialist, however, Draupadī can thus speak for the bud-
dhi as primal matter’s subtlest aspect. If so, this would be in accord with
Yudhiṣṭhira’s final tribute to her after she has passed away: that she was “the
great dark one endowed with intellectual substance” (bṛhatī śyāmā
buddhisattvaguṇānvitā).76 Aware of the strings attached, from birth to death,
not consecutively but when she comes on stage in bits and pieces, Draupadī
performs a biography of matter.
Finally, let us tie some things from these last two chapters together.
When Yudhiṣṭhira was leaving this world, his very last thought was that he
wanted to ask Draupadī a question, only to learn that she had returned to her
divine nature as the goddess Śrī.77 Whatever Yudhiṣṭhira wanted to ask
Draupaḍī, the poet leaves it for us to imagine. But it is probably fair to say that
he wanted to round off some old debate with that “lady pandit” who could
engage him like none other. In Rāma’s case we have a far vaster question:
Someone told him it would be beneficial for him to hear two boys recite a
poem, which turns out to be about him. Then shortly after the boys have
begun, Sītā enters the earth, rejecting the opportunity Rāma has offered for
their reunion. Would what was beneficial for Rāma have included his listening
to the fading voice of Sītā? And if so, would he have listened as one who
finally understood his own divinity, or—more interestingly—as a just and
perfect man?

75. As discussed in chapter 9 § D.2.c, this is one of five instances where Yudhiṣṭhira’s “rising up” is made
prominent at highly dramatic moments. The same verb is also used to describe Draupadī when she rises up from
the earthen altar (vedi) of her birth (1.55.41; 5.80.21).
76. See Hiltebeitel 2007a, 110, 133–35, with background for this interpretation in the “three unborns” of
Śvetāśvatara Upaniṣad 4.6–7; cf. 2001a, 272–73, on other possible translations, beginning with “the great dark
one rich in spirit, character, and virtue.”
77. Mbh 18.4.8–10; see Hiltebeitel 2001a, 276.
11
Dharma and the Bhagavad Gı̄tā

This chapter will work toward understanding the way dharma is

presented in the Bhagavad Gıtā ̄ . In part because the Gıtā
̄ allows us to
proceed from the relatively familiar to the relatively less so, it will
allow us to bring greater clarity to the Mahābhārata’s new formula-
tions of temporally oriented dharma that we opened discussions on in
chapters 6 and 7 in contrasting the epic’s and Manu’s treatments of
yugas and kalpas. We may keep these contrasting treatments in mind,
as we will be returning to them toward this chapter’s end. But I begin
by contrasting some of the Bhagavad Gıtā ̄ ’s
̄ more basic formulations
with Manu’s treatment of the same terms: the limits and specificities
of the concepts of svadharma, svakarma, svabhāva taken as a set,1 and
also karmayoga. This will bring out how the epic, and especially the
Bhagavad Gıtā̄ , figure the Kṣatriya most pivotally in these formula-
tions. Then, by exploring how the Bhagavad Gıtā ̄ is contextualized in
the wider Mahābhārata and tracing in particular one of its “ripple
effects” through a formula that relates dharma and victory to Kṛsn ̣ ạ ,
I will demonstrate how the Gıtā ̄ itself presents dharma through a
ring structure. We will thus be able to look more closely at what the
Bhagavad Gıtā ̄ (henceforth, in this chapter, BhG or Gıtā
̄ ) brings to
the topic of dharma over time for the warrior as distinct from the
king. And we will set some terms for the exploration of the wider
Mahābhārata’s treatment of dharma and bhakti in chapter 12.

1. Regarding usages of svabhāva, it would seem a plausible hypothesis that, in developing their
position of “realism,” the Sarvāstivādins, a predominantly north central and northwest Indian school,
̄ , and other roughly contem-
could have had overlapping concerns with the Mahābhārata, including the Gıtā
porary Brahmanical texts in shoring up what they meant by dharma and dharmas. Cf. chapter 4 § B.2.b.
518 dharma

A. Svadharma and Svakarma: Qualities, Merits, and Virtues

As brought out in chapter 5, Manu’s crypto-narrative orchestration of the rela-

tion between the Snātaka and the king is one of many attempts to address a
larger problem that was crucial to Brahmanical culture in the way it distin-
guished itself from the heterodoxies: the relation between the Brahmin and the
king, who should be a Kṣatriya. Theorizing the relation between the Brahmin
and the Kṣatriya, with the king as a kind of linchpin, was the key to offering
dharma as an all-encompassing Brahmanical social order. And the key to the
theory itself was the concept of svadharma.
In his article “On the Rhetoric and Semantics of the Puruṣarthas,” Charles
Malamoud provides an elegant extraction of this theory at the beginning of an
appendix titled “On the Correspondence between the Puruṣārthas the Varṇas
and the Ā śramas”:

What provides the foundation for Brahman superiority is the fact that
their svadharma is of the same nature as dharma in general. Their
specialty in the code is to hold the keys to the code; they watch over and
judge the whole of the svadharma. This peculiar affinity that the
Brahmans have for dharma is derived from their alone being qualified to
teach vedic texts which are the ultimate source of dharma; they are also
alone in being allowed to officiate in śrauta ceremonies; their actual
presence and direct influence are indispensable for what is the essential
part of dharma—sacrifice. (Malamoud 1982, 49)

This formulation is impeccable in placing svadharma at the foundation of

Brahmins’ theoretical primacy, and with it the matter of Brahmins’
“qualifications”—which, let us begin to note, imply special “qualities”—for jobs
open only to themselves as Veda specialists who would indeed have motivation
to make (and keep) sacrifice “the essential part of dharma.” It is, however, a syn-
chronic distillation from many texts, and when we look at the texts, we see that
even there it is something of an illusion, especially where it comes to such a
notion as “the whole of the svadharma.” The texts’ authors may have had such a
synthesis in the back of their minds, but the texts themselves are far more com-
plex, varied, incomplete, and revealing in making the points that might at most
only suggest such an orderly statement. This is particularly the case in two areas
where such a synchronic view of svadharma has given rise to unwarranted gen-
eralizations and extensions: that svadharma applies in principle to all humans
individually and equally in every walk of life; and that it pertains to all humans
individually and equally as well as to suprahuman entities.
 dharma and the bhagavad gı̄ t ā 519

The first generalization breaks down when one begins to notice how the
notion of svadharma is developed in association with Brahmin jobs, which, as
we shall see, are usually discussed not under the heading of dharma but rather
under that of karma as “suitable occupations” (svakarma), and with new config-
urations of “qualities,” “merits,” and “virtues.” It is within this novel discourse,
which does not extend back to older meanings of dharma, that Fitzgerald can
generalize about dharma in the Mahābhārata as “being a species of karman”
(2004b, 676; cf. 679, 682, 684 n. 12) based on its having three “layers” of
meaning: normative and especially sacrificial action (karman) (with Fitzgerald
emphasizing that such action is beneficial to a person after death); a more
abstract ethical “‘quality of the correctness, rightness, goodness, or justice’ of
some action”; and “the inner attributes (guṇas) of a person” (674–75).
The second generalization only builds out from the tenuous premises that
allow the first one. As was noted in chapter 3, it has encouraged an unwar-
ranted backreading of the way “each [Ṛgvedic] god follow[s] the law proper to
his own being, in other words his own dharman” (Miller 1985, 102). As we shall
see in the next section, it has also been pushed forward into classical Hinduism
as a cosmic synchronic principle that can be applied most anywhere, often with
little or no warrant from the texts.
Yet svadharma has somehow “become what it is.”2 As far as I can see, the
term gets its first workout, together with svakarma, in the dharmasūtras, where
a kind of semantic drift between the two terms is set in motion; and the two are
then further developed as governing paradigms in Manu and the Mahābhārata,
where they are harnessed, for the Mahābhārata at least in the BhG,3 to what I
consider to be two different notions of karmayoga.4 The Rāmāyaṇa knows these
terms, but uses them less intensively. In the case of svadharma, this is probably
because it projects an earlier world where the social classes are harmonious.
There is less need to shore up matters of svadharma (which it rarely mentions)
since class-mixture does not really exist. The only two usages of the term for
“mixture” or “confusion” (saṃkara) are preemptive: there was no class-mixture
when Rāma’s father ruled (Rām 1.6.12), and Rāma goes into exile holding to his
father’s truth to prevent the world from falling “into confusion” (2.101.6). His
killing of the upstart Śūdra Śambūka is likewise a preemptive strike against a
form of class-mixture: according to the Ṛṣi Nārada, who puts Rāma to this task,

2. Recall that the compound svadharman has one occurrence in the Ṛgveda at ṚV 3.21.2b, where it seems
to ask Agni as the ritual f ire to bestow what is best “for us” from “his own foundation”—a usage that seems to
have no follow-up in the Vedic canon itself, and to be unrelated to a distinctive usage in the Śrautasūtras
(see chapter 3 § E). According to Gombrich (1985, 436; 1988, 67), this “Hindu notion” seems to have no equivalent
in the Pāli canon.
3. See Fitzgerald 2002, 641–42, 47, 650 n. 5.
4. See § C for some recent debate on this issue.
520 dharma

a Śūdra should only be able perform tapas in the Kali yuga, which is yet to come
(7.65.22–23). As to karmayoga, there is just one curious instance at 5.45.30
describing Hanumān as svakarmayoga ca vidhāya, which Goldman and
Sutherland Goldman translate “settling into his own plan of action.”
With svadharma, we may thus begin to notice a pattern developed in the
dharmasūtras that Manu and the Gıtā ̄ only reinforce. The dharmasūtras use the
term svadharma mainly to prescribe or “legislate,” as it were, what is generally
appropriate and to be enforced (by the king or, in the BhG and in at least
the twelfth chapter of Manu,5 by the karmic mechanism of reincarnation) for all
the varṇas and āśramas.6 But their particular target in prescribing svadharma is
the classes below the Brahmin (Ā 1.18.3)—and especially the Kṣatriya, and still
more singularly the Kṣatriya king. Yet when the same Brahmin authors speak
specifically of themselves and the privileges and occupations reserved for them,
they use instead the term svakarma, thereby speaking of their “own actions” or
“activities” or “own occupations” rather than their “own dharma.” And when they
speak of other classes’ “appropriate actions” or “occupations” (svakarmas) instead
of their svadharma, they tend to do so where the activities of such others would
impact directly upon Brahmins: as with what Brahmins can eat that others might
offer (G 17.1; B 1.3.17–18), or where occupations reserved for Brahmins set the
paradigm for what other classes may and may not do. As we saw in chapter 5,
Ᾱpastamba’s treatment of this matter, with three idiosyncratic additions to the
basic set of six, sets the tone. Before he lists the occupations of the two Ārya castes
below the Brahmin (2.10.6–7), he defines those reserved for the Brahmin first, as
follows: “The occupations specific to a Brahmin (svakarma brāhmaṇasya) are
studying, teaching, sacrificing, officiating at sacrifices, giving gifts, receiving
gifts, inheriting, and gleaning, as well as appropriating things that do not belong
to anybody” (4–5). The first six of these nine “occupations specific to a Brahmin”
become standard (see M 10.74–75, cited below; Mbh 5.29.26;7 7.168.22–23;8
12.297.15; 13.129.7–8). The overall implication, as I see it, is that the svakarma of
Brahmins defines the “archetype” or default position of dharma implicitly,
without mentioning the term dharma, for all Āryas, with only some of their “own
actions” or “occupations”—especially giving but not receiving gifts—being

5. See Bronkhorst’s notion (in press), cited in chapter 5 n. 76, that Manu 12 is late because it is only there
that Manu discusses “rebirth as determined by karmic retribution” “to explain the superior status of
Brahmins.”
6. See Ā 2.2.2 (mentioning only varṇas); G 11.29 (mentioning both). In a passage that Olivelle regards as
late, B 2.17.4 allows that a man may undertake renunciation (saṃnyāsa) if he is a widower or after he has “settled
his children in their respective duties.”
7. Kṛṣṇa, speaking to Saṃjaya here, mentions additionally that Brahmins should visit tīrthas.
8. Dhṛṣṭadyumna, defending his killing of Droṇa, asks which of the six actions (karmāṇi) Droṇa fulfilled
(see chapter 9).
 dharma and the bhagavad gı̄ t ā 521

appropriate for Kṣatriyas and Vaiśyas.9 Meanwhile, in this passage from Āpastamba
“the Śūdra is purely and simply eliminated, or perhaps there is nothing new to
say as regards varṇa concerning him.” Biardeau says the passage concerns itself
with “a Brahmin’s dharma” (2002, 1: 77), but surprisingly, that term is not even
used with regard to Brahmins. Yet the Brahmins’ svakarma does of course pro-
vide those below them with an archetype that models the svadharma of Kṣatriyas
above all, and others below them, on sacrificial ritual.
Now as we saw in chapter 5, the pivotal figure for and through whom
dharmasūtra authors begin to legislate dharma for the larger society is the king.
Gautama has a fine passage on this subject, part of which, now italicized, was
quoted in that chapter’s discussion of varṇa and āśrama.

The king rules over all except Brahmins. He should be correct in his
actions and speech, and trained in the triple Veda and logic
(ānvīkṣikī). Let him be upright,10 keep his senses under control,
surround himself with men of quality [have companions who possess
guṇas: guṇavatsahāya-], and adopt sound policies. He should be
impartial to all his subjects and work for their welfare. As he sits on a
high seat, all except Brahmins should pay him homage seated at a
lower level, and even Brahmins should honor him. He should watch
over the social classes and the orders of life in conformity with their rules,
and those who stray he should guide back to their respective duties [that is,
to their svadharma, singular], “for the king,” it is stated, “takes a share of
their merits (dharmas).” He should appoint as his personal priest a
Brahmin who is learned, born in a good family, eloquent, handsome,
mature, and virtuous (śilasaṃpannam); who lives according to the
rules (nyāyavṛttam); and who is austere. He should undertake rites
(karmāṇi) only with his support, “for a Kṣatriya, when he is supported
by a Brahmin,” it is said, “prospers and never falters.” (G 11.1–14)

It is, of course, implied that all the classes and āśramas have their svadharma, but the
focus is on the king’s reaping that “cumulative svadharma” as merit (dharma) for
himself. His own svadharma is not directly clarified, other than that one may infer it
from his being a Kṣatriya. It is not to be confused with the merit (dharma) he reaps.
In reaping the “merits” (dharma) that come from seeing to it that the social classes

9. Early in the Rājadharma section of Mahābhārata Book 12, where Bhīṣma first treats the four varṇas at
length (12.60.6–35), he mentions the six in terms of the three each that Kṣatriyas can and cannot do (13c–14a),
having stated that the dharma or “virtue” solely ordained for Brahmins is dama or “restraint,” which goes along
with their teaching and recitation of the Vedas (8c–9).
10. Śūci, which Olivelle also sometimes translates as “pure,” “clean”; for extended discussion, see Olivelle
2005b, 220–29.
522 dharma

and orders of life do not stray from their svadharma,11 the king should keep his own
senses under control, implying the mastery of ethical values that have come by the
time of the later Upaniṣads and the emergence of the heterodoxies to be associated
with renunciation, yoga, and the reduction of violence. Under the impact of the new
renunciatory and yogic ethics, a new concatenation of merits, virtues, and qualities
is thus coming to be indexed by increasingly interchangeable usages of the terms
dharma and guṇa (“quality” or “attribute”:),12 along with which we must also mention
puṇya (“merit”), to characterize people who are virtuous or moral (śilasaṃpannam).
In their overlap, all three terms can carry karmic momentum and transferability.
But though one may generalize, they also operate in somewhat different contexts.
With dharma, Manu’s sixth chapter on the āśramas uses the term to define
the set of virtues that makes up the “Ten-Point Law” that achieves the highest
aim for the four life-stages, the last two of which, above all, would call for the
mainly yogic and renunciatory virtues now mentioned while discussing those
two most life-fulfilling āśramas:

Twice-born men belonging to these four āśramas must always observe

the ten-point Law diligently. Resolve, forbearance, self-control (dama),
refraining from theft, performing purifications, mastering the organs,
understanding, learning, truthfulness, and suppressing
anger. . . . Those Brahmins who learn the ten points of Law, and, after
learning, follow them, attain the highest state. (M 6.91–93)

With the addition of ahiṃsā, this list is boiled down with regard to the four
varṇas, where perhaps there is a need to be more basic:

Abstention from injuring (ahiṃsā), truthfulness, refraining from anger

(variant: not stealing), purification, and mastering the organs—this,
Manu has declared, is the gist of the Law for the four classes. (M 10.63)

Lists of virtues like this can be rather ad hoc.13 The Mahābhārata’s instruc-
tions on Rājadharma offer a half-equivalent list of ten dharmas, mentioning

11. His role of dealing with those who “stray” (Öcal) is presented even more starkly in Manu, where it is a
question specifically of the king’s need to apply punishment (daṇḍa) (which Gautama deals with subjacently
[11.28ff.]): “It is the fear of him that makes all beings, both the mobile and the immobile, accede to being used
(bhogāya kalpante: literally, being enjoyed, eaten) and to not deviate (Öcal) from the Law proper to them
(svadharmāt: from their svadharma)” (Manu 7.15).
12. The Mahābhārata orchestrates interchangeable usages of the compounds sarvaguṇopeta (“endowed
with every quality”) and sarvadharmopeta (“endowed with every virtue” or “merit”) in many narratives that
describe combinations of yogic and royal virtues of kings and queens (Hiltebeitel [1976] 1990, 192–228).
13. See Ganeri 2007, 231–36 on Mbh 12.156.22–26 as Bhīṣma’s teaching to Yudhiṣṭhira of thirteen virtues of
truth, and BhG 16.1–4, where Kṛsn ̣ ạ mentions truth (satya) among twenty-six virtues that “comprise the divine
complement of virtues of him who is born to it.” The thirteen virtues of truth differ from the thirteen sādhāraṇa
dharmas soon thereafter mentioned by Bhīṣma, which are headed by noncruelty (12.285.23–24; see chapter 5 § E).
 dharma and the bhagavad gı̄ t ā 523

nine (including sharing, begetting offspring on one’s own wife, and supporting
one’s dependents) as pertinent to all varṇas, but reserving self-control or
restraint (dama) as the dharma exclusive to Brahmins (Mbh 12.60.7–8; see n. 9
above). As we saw in chapter 6 § C, the god Dharma himself has ten wives
named after more or less feminine virtues (1.60.13–14); and, after Dharma dis-
guises himself as a Yakṣa to test his son Yudhiṣṭhira, he tells him of ten virtues
that are dharma’s “bodies” (3.298.7). These are different lists. Only one of
Dharma’s wives (Dhṛti or Resolve) and three of his bodies (truth, purity, and
restraint) coincide precisely with Manu’s ten.
Puṇya too is meritorious action especially as regards reaping the fruits of
karma,14 usually in terms of afterworlds, including heaven, that fall short of
“the highest state” attainable by Manu’s “ten-point Law.”15
Guṇa, on the other hand, introduces the differential calculus of the three
“qualities” (guṇas) of matter, the chief ingredients of the soup of life that
enable one to distinguish those of “good quality” (sattvaguṇa) from the rest.
Kṛṣṇa makes the three guṇas highly pertinent in his closing instruction to
Arjuna in the BhG:

That intelligence (buddhi) proceeds from sattva, Pārtha, which

understands when to act and when not (pravṛttiṃ ca nivṛttiṃ ca), what
is a task and what not, what is a cause of fear and what not, what is
bondage and what deliverance. To rajas goes the intelligence with
which one incorrectly perceives dharma and adharma, task and no-task.
An intelligence is inspired by tamas, Pārtha, when, obscured by
darkness, it mistakes adharma for dharma and perceives all matters
topsy-turvy. (BhG 18.30–32; van Buitenen 1981, 141 slightly modified)

As one can see, sattva guṇa provides the intellect (buddhi) with the particular
aptitude to differentiate between what the Nārāyaṇīya calls pravṛtti- and
nivṛtti-dharmas. Also clarifying that guṇas can lead to high soteriological ends,
Manu exerts considerable energy in showing through six widening iterations
how good guṇas correlate with favorable reincarnation and the “fruits of
action” (12.24–50).
Manu is more interested, however, in demonstrating the social benefits of
good guṇas. It is to be assumed that the Brahmins whom the king should surround
himself with as advisors would have the innately defined “sattvic qualities” or virtuous

14. See M 5.53, using puṇyaphalam, “the reward for meritorious acts,” in this sense afterlife merits. See
the similar uses of puṇya at 8.90–91 with just puṇya). On puṇyaphala, note Ā 2.14.18, pertaining to a husband
and wife; 2.24.13, pertaining to “the bodies of those seers who have done meritorious deeds shining brilliantly
far above”—probably referring to the Seven Ṛṣis of the Big Dipper (Olivelle 1999, 372).
15. See the Uttara-Yāyata’s story of King Yayāti (Mbh 1.81–88) discussed in § C.3.
524 dharma

guṇas that entitle them to their svakarma, and to decide others’ dharma. As Malamoud
suggests, in so far as Brahmins have an implied or unstated svadharma, it is to be
“qualified” innately to “hold the keys to the code.” Manu’s first chapter moves for-
ward its whole program by relating class and quality (varṇa and guṇa) to proper
conduct (ācāra) as “the highest dharma” among the “roots of Law”:

In this, the Law has been set forth in full—the good and the bad
qualities of actions (guṇadoṣau ca karmaṇām) and the timeless norms
of social conduct (ācāra)—for all four social classes. Proper conduct
is the highest Law (ācāraḥ paramo dharmaḥ), as well as what is
declared by the Veda and given in traditional texts. (1.107–8b, cited
above, chapter 5 § D)

Illustrative of the guṇas’ force as the operational virtues of twice-born men is

a fascinating pair of brief stories that Manu alludes to together, the second of
which would seem to come from the Mahābhārata. Manu’s purpose is to
describe how low-born women may improve their guṇas just by their being
married to Brahmin husbands. The two stories are about Arundhatī (called
Akṣamālā by Manu), the wife of the great Ṛṣi Vasiṣṭha, whose story Manu
could find elsewhere, though I do not know where he would find the twist
that Arundhatī is of lowly birth; and Śārṅgī, the wife of Mandapāla (9.22–24).
Unless Manu made this second story up or knew it from some lost variant, he
probably could have found it only in the Mahābhārata, where it is a subtale
told in the aftermath to the Burning of the Khāṇḍava Forest. The epic
calls Mandapāla’s wife Jaritā, but Manu probably knows her from a version of
the epic account or one similar to it, since his name for her, Śārṅgī, refers to
the fact that she is a Śārṅgaka bird whom Mandapāla became a bird himself
to marry so that he could quickly have children to satisfy his debt to his ances-
tors.16 What is fascinating here is that “Manu” puts a totally different twist on
the story to make his point that a husband’s guṇas can elevate those of a wife.
Not only does he omit to mention that Mandapāla became a bird himself,17
and thus at least for the time being was as lowly as Śārṅgī; he cheats Jaritā of
being the heroine of the story.18 Manu’s concern that a husband’s guṇas have
such uplifting power is well explained with reference to offspring:

16. On the three debts, see chapter 5 § A in connection with the five great sacrifices; chapter 8 § D and § G
in connection with Satyavatī and Pāṇḍu, the latter mentioning four debts.
17. The story is facilitated by the fact that the term dvija, “twiceborn,” refers to both birds and Brahmins.
18. See Hiltebeitel 2007a, 118–23 on the Śārṅgaka-Upākhyāna. The Vaṭṭaka-Jātaka is a Buddhist parallel
(Cowell [1895] 2005, 88–90; see Söhnen-Thieme 2005), but with different names and nothing that could have
affected Manu.
 dharma and the bhagavad gı̄ t ā 525

If it be asked: who is superior? A child born accidently (yadṛcchayā) to

a Brahmin man by a non-Ārya woman or a child of a non-Ārya man
by a Brahmin woman? This is the resolution: a child born to an Ārya
man by a non-Ārya woman becomes an Ārya by reason of his guṇas,
while a child born to a non-Ārya man is a non-Ārya. Neither of these
should be able to receive vedic initiation—that is the settled Law; the
former because of the inferiority (vaiguṇyād) of his birth and the
latter because he was born in the inverse order of class (pratilomataḥ).
(M 10.66–68; Olivelle 2005a, 211 slightly modified)

The Brahmin father’s son no doubt inherits his father’s guṇas but is still “without
guṇas” (vaiguṇya) because he has inherited nothing of that sort from his mother.
As Manu goes on to explain, the father is seed, the mother field (70–72)—a
recurrent metaphor that we noticed in chapter 8 with regard to women in the
Kuru line (Ambikā and Ambālikā, Kuntī and Mādrī) and shall notice again.
I cannot treat here the considerable number of passages where Manu follows
suit in speaking mainly of Brahmin karma and svakarma around the issue of
jobs, and of dharma and svadharma as legislated mainly for others, and particu-
larly Kṣatriyas and Kings.19 It must suffice to treat the main passage where Manu
deploys much the same distinctions between svakarma and svadharma that we
have seen in the dharmasūtras, but by sharpening them and widening their appli-
cations, and then consider his treatment of the svadharma of Kṣatriyas. Here are
his refinements on “the occupations appropriate to” Brahmins:

Brahmins who are established in that whose source is the Veda and
are devoted to the activity specific to them (svakarmaṇi; singular,
whereas Olivelle pluralizes it) should duly live by the six occupations
(ṣaṭ karmāṇi) in their proper order: teaching and studying, offering
sacrifices and officiating at sacrifices, giving and accepting gifts are
the six occupations of a highest-born person. Of these six activities
(karmas), however, three provide him with a livelihood: officiating at
sacrifices, teaching, and accepting gifts from a completely pure
person. From the Brahmin, [these very] three Laws (trayo dharmā)
are suspended with respect to the Kṣatriya: teaching, officiating at
sacrifices, and the third, accepting gifts; the same are suspended
also with respect to the Vaiśya—that is the settled rule; for Manu,
the Prajāpati, has not prescribed these Laws (tān dharmān manur āha)
with respect to these two. Use of arms and weapons has been

19. See M 1.53; 1.107, 2.183; 4.03; 4.155 ff., 10.1–3; cf. 2.8, 3.3, 3.235, 4.3, 4.155; 5.2, 6.91–93, 6.97, 7.36,
8.41–42, 8.390–91, 9.251, 10.95–97, 11.84.
526 dharma

prescribed20 for the Kṣatriya, and trade, animal husbandry, and

agriculture for the Vaiśya. Their Law (dharma), however, is giving
gifts, studying, and offering sacrifices. Among the activities specific
to each (svakarmasu), the most admirable are: studying the Veda for a
Brahmin, protecting the people for a Kṣatriya, and trade for a Vaiśya.
When a Brahmin is unable to earn a living by means of the activity
[again, singular] specific to him (svena karmaṇā) given above, he may
live by means of the Kṣatriya Law (kṣatriyadharmeṇa). (M 10.74–81,
Olivelle 2005a, 211–12 slightly modified)

The passage follows Manu’s presentation on mixed castes, and is clearly meant
to define the order that mixed castes must not be allowed to adulterate. Again,
it is primarily Brahmins who have svakarma. Their svakarma is glossed as
dharma (not svadharma) in two contexts: when it prescribes distinctions between
the three things Brahmins alone can do and the three that Kṣatriyas and Vaiśyas
can do along with them; and when, in times of distress, it is “dharma” for a
Brahmin to adopt Kṣatriyadharma, taking up arms to protect people.
Indeed, Manu has by this time defined Kṣatriyadharma fairly carefully:

When a man is killed in battle with upraised weapons according to

the Kṣatriya law (kṣatradharma), the settled rule is that for him both a
sacrifice (yajña) and a purification (āśaucam) are accomplished
instantly (sadyaḥ). (5.98)

This follows from Manu’s discussion of the instant purification of kings that
we noticed in chapter 5. Moreover, the sacrificial sanction is sustained by a self-
sacrificial warrior code:

When challenged by rivals—whether they are stronger, weaker, or of

equal strength—as he protects his subjects, a king must never back
away from battle, recalling the Law of Kṣatriyas (kṣatraṃ dharmam
anusmaran). Refusal to turn back in battle, protecting the subjects,
and obedient service to Brahmins—for kings, these are the best
means of securing happiness. When kings fight each other in battle,
with all their strength, seeking to kill each other and refusing to turn
back, they go to heaven (svargam). (7.87–89)

And, rounding off this section:

I have set forth above the eternal Law of warriors (yodhadharmaḥ

sanātanaḥ) without elaboration. A Kṣatriya must never deviate from
this Law (asmād dharmāt), as he kills his enemies in battle. (7.98)
20. Olivelle inserts “as the livelihood,” but this is only implied.
 dharma and the bhagavad gı̄ t ā 527

Now as we saw in chapter 5, where “Manu” establishes the principle that

there is “no fifth” class, he does so by emphasizing that teaching is one
of the three jobs reserved for Brahmins. He makes this the pretext for posi-
tioning the Brahmin as the one who, in times of adversity, defines caste duties
across the board:

The Brahmin must know the means of livelihood of all according to

rule, and he should teach them to others and follow them himself.
Because of his distinctive qualities, the eminence of his origin, his
holding of restrictive practices, and the distinctive nature of his
sacrament, the Brahmin is the lord of all the classes. (10.2–3)

This teaching on teaching is, in effect, Manu’s answer to the Buddha’s question
in the Esukāri Sutta: “Well, Brahmin, has all the world authorized the Brahmins
to prescribe? (MN 96.4 and 11; see chapter 4 § A). Manu’s insistence that a
Brahmin can adopt Kṣatriyadharma is thus especially weighted, suggesting
one of the values of “legislating” that prerogative. When Manu then discusses
the possibility that a Kṣatriya might “dip” to the occupations of a Śūdra, it is
with a verse that parallels a famous verse from the BhG,21 where it concerns
Kṛṣṇa’s guidance of the exemplary Kṣatriya Arjuna:

Far better to carry out one’s own Law (svadharma) imperfectly than
that of someone else’s perfectly; for a man who lives according to
someone else’s Law (paradharmeṇa) falls immediately from his caste.
(Manu 10.97)

But when it comes to Vaiśyas and Śūdras, although Brahmins may likewise
take up their “occupations” in times of distress, there is little interest in calling
them Vaiśya-dharma and none at all in speaking of Śūdra-dharma. Manu makes
only one mention of Vaiśyadharma (10.98) and none of Śūdradharma, con-
ceding only that Śūdras can and should follow dharma or Law, and implying
that, were there such a thing as Śūdradharma, it would be only by upward imi-
tation (10.127–28) and not any kind of svadharma (cf. 10.50).
Likewise, in the discussion of mixed castes that precedes the main passage
on Brahmin occupations, Manu—as is well known and for many notorious—
also defines mixed castes by their vast array of occupations (10.1–3 and ff.), and
very little in terms of dharma, much less svadharma. In the section on mixed
classes (M 10.5–73), such groups are ascribed only svakarma (10.40, 50).
Svadharma is never used, and dharma is used only five times. Three of these
instances state general rules: Manu’s rules on mixed class offspring are

21. BhG 3.35; see Doniger 1991, 246; also 18.47. But Doniger also mentions M 6.66, which is a misleading
parallel since it is about āśramadharma.
528 dharma

“righteous” (dharmya) (10.7); the “delinquent-born” and Śūdras are sadharmāṇaḥ

(“have the same nature” or “characteristics”; 10.41); the “gist of the Law” for all
four classes lies in certain virtues (10.63). And the rest define a Brahmin per-
spective from the standpoint of their separation: those “who follow dharma”
must have no contact with two classes of outcastes (10.53); offspring of only one
Brahmin parent are denied Vedic initiation (asaṃskāryau), whether the father
or mother is the Brahmin (10.68). Indeed, it would seem that this dodge has
been operating from the beginning of Manu, where it can be traced to Manu’s
frame story. The first usage of “dharma” in Manu comes when the Ṛṣis ask
about “the dharmas of those born in between” (1.2). The second, and the first by
“Manu” himself, occurs with the assertion that dharma and adharma were
introduced to establish “distinctions among karmas (karmaṇāṃ ca vivekārtham)”
(1.26), implying not only “activities” but a system of job reservations for all
above the Śūdra, but especially for Brahmins.
As to the Mahābhārata, it must suffice to mention only a few statistics as
the basis for further analysis, while citing what Bhīṣma has to say on the sub-
ject: first as he introduces varṇadharma toward the beginning of his postwar
teachings on the dharma of kings in Book 12, and then, much later, as he quotes
Śiva toward the end of Book 13. In the initial passage, Bhīṣma first treats
Brahmins rather cursorily. Having mentioned dharma only in the sense of the
nine “virtues” pertinent to all varṇas, with “restraint” reserved to Brahmins,
and having begun with Brahmins’ activities as “things to be done” or “duties”
(kāryas) (12.60.12a), he turns to Kṣatriya dharma (13) on the way to making
these pronouncements: “They say the dharma for those of Kṣatriya connection
(kṣatrabandhūnāṃ dharmam) is primarily killing; there is nothing more impor-
tant to do than the destruction of robbers (dasyus)” (17); and “for the king in
particular who wishes to obtain dharma, it is to make war (rājñā viśeṣeṇa
yoddhavyaṃ dharmam īpṣatā)” (18cd). As one would expect in an epic mostly
about Kṣatriyas, kṣatriyadharma, which, along with the far more common
kṣatradharma, occurs by a rough count about 175 times in the Pune Critical
Edition (and in every Book except Books 16 and 17). (That compares with
12 times in the Rāmāyaṇa, where the term “warrior dharma” is curiously absent
from Book 6, that epic’s War Book.) Vaiśyadharma and near equivalents occur
about 9 times; śūdradharma and near equivalents about 6 times (3.149.36,
5.29.24; 12.60.27, 13.128.56 and 58, 13.129.15); and brāhmaṇadharma a mere
once, at 13.131.8a, where, like the last three references to Śūdradharma, it occurs
in a “Dialogue Between Umā and Maheśvara” (Umā-Maheśvara-Saṃvāda).
Among the many intriguing things the Goddess wants to hear about from Śiva,
one topic is the dharma of Brahmins (13.128.29a). Śiva focuses his reply
on householder Brahmins as “gods on earth” (bhūmidevāh), ̣ for whom he
 dharma and the bhagavad gı̄ t ā 529

prescribes fasting, honoring gurus and gods, hearing of mysteries and the
observance of Vedic vows, mendicancy, wearing the sacred thread, Veda recita-
tion, brahmacarya, returning home from their preceptor’s house to marry,
avoiding the food of Śūdras, welcoming guests and eating only after all others
are served, performing both vegetarian and animal sacrifices while practicing
ahiṃsā, and generally—in his favorite phrase (used eleven times in this
dialogue)—following the “path of the good” (satpatha) (13.128.30–45).
Meanwhile, what Śiva prescribes for Kṣatriyas, and in particular kings, is to
protect subjects, be firm in litigation (vyavahārasthitir), and “die in battle, hav-
ing gone forth on behalf of cows and Brahmins” (13.128.46–52). In this long
dialogue, Śiva uses the term svadharma only twice: once, in another exception
to our rule, to say a Brahmin who falls from his svadharma may attain Śūdrahood
and go to hell (13.131.12), and just after this to say more or less the same thing
about those of all three Ārya varṇas. Umā never uses it in answering Śiva’s fol-
low-up question about women’s dharma (13.134.11–29).

B. Who Has Svadharma?

At least in the case of kṣatriya- or kṣatra-dharma, compound uses of varṇa +

dharma do correspond to a varṇa’s svadharma, and we may infer the same for
the other three social classes. But if, in Malamoud’s terms, the Brahmins
“watch over and judge the whole of the svadharma,” they have set a limit on
their vigilance. The epic, like Manu, makes no such correlation between
dharma and svadharma for further classes of beings. Svadharma is restricted
to the four varṇas only. This comes across quite vividly in a passage that Wendy
Doniger O’Flaherty cites on the subject of demons’ supposed svadharma—
one that actually mentions the term. O’Flaherty says it concerns “demons,
Rākṣasas, Piśācas, and others” who have “abandoned their svadharma” (1976,
68), but it would rather appear that they never had any svadharmas to abandon.
The passage, from the Mokṣadharmaparvan of the epic’s twelfth Book,
recounts a taut and erudite conversation between two of the great Vedic
Brahmins, Bhṛgu and Bharadvāja, in which the origins of svadharma would
seem to have been woven into a reminiscence of the cosmogony in
Bṛhadāraṇyaka Upaniṣad 1.4.11–14 (see chapters 3 § F and 6 § C). As will be
recalled, the Upaniṣad tells how brahman was at first alone and “had not fully
developed” until it created the “ruling power,” Vaiśyas, Śūdras, and the Law
(dharma) as truth, which made it possible to coordinate the three sub-Brahmin
classes to make the Brahmin “the power behind the throne.” The passage
would seem not only to expand on these notions but also to dissolve them
530 dharma

with typical epic irony in its play on varṇa as both “color” and “class,” before
reconsolidating and setting limits:

Bhṛgu said,
Brahmā the Lord of Creatures first created Brahmins, who were
produced from his own radiance (tejas), their splendor like the Sun
and Agni. The Lord then ordained (vidadhe) truth, dharma, austerity
(tapas), the eternal Veda, and pure conduct (ācāraṃ caiva śaucam)
for attaining heaven. Gods, Demons (Dānavas, Daityas, and
Asuras), Gandharvas, Great Snakes, Yakṣas, Rākṣasas, Nāgas,
Piśacas, and also Men—Brahmins, Kṣatriyas, Vaiśyas, and Śūdras—
and whatever other groups there are among Beings, O best of the
twiceborn, these too he meted out (nirmame). The Brahmins’ varṇa
was white, the Kṣatriyas’ was red, the Vaiśyas’ varṇa was yellow, and
the Śūdras’ black.
Bharadvāja said,
If varṇa is distinguished by the varṇa of what applies to the four
varṇas, then mixing of varṇa (varṇasaṃkaraḥ) is certainly seen
among all the varṇas! Desire, anger, fear, greed, grief, anxiety,
hunger, and toil prevail over us all. How can varṇa be distinguished?
Sweat, urine, feces, phlegm, bile, and blood flow from all bodies.
How can varṇa be distinguished? The kinds of birth (jātis) of mobile
and immobile beings are innumerable. How can there be decisive-
ness on varṇa (varṇaviniścayaḥ) when these varṇas are so various?
Bhṛgu said,
There is no distinction of varṇas. This whole universe is brahman.
It was created formerly by Brahmā, surely, and came to be
classified by acts (karmabhir varṇatāṃ gatam). Twiceborns (dvijas)
for whom desire and pleasures were dear, harsh, inclined to anger,
given to boldness, and red-limbed, once they abandoned their
svadharma they attained Kṣatriyahood. Twiceborns who had taken
to cattle-rearing, yellowish, living off the plough, who did not
follow their svadharma, they attained Vaiśyahood. Twiceborns who
were fond of violence and lying, covetous, living off work of all
kinds, darkish, fallen from purity, they attained Śūdrahood. Thus
separated by these occupations (ityetair karmabhir vyastā), twice-
borns attained difference of varṇa. Dharma, sacrifice, and rites are
never prohibited for these—surely, these are the four varṇas about
whom vedic speech is holy. Brahmins are established on the loom
of dharma (brāhmaṇā dharmatantrasthās). Their austerity is not
 dharma and the bhagavad gı̄ t ā 531

destroyed. Their vows and restraints are always engaged in

upholding [root dhṛ] Brahmā. And those who do not know about
what Brahmā formerly created, not knowing that, for these there
are many other varied kinds of births (jātis) here and there.
Piśācas, Rākṣasas, Pretas, and the various kinds of Mlecchas are
motivated by unruly conduct (svacchanda-ācara-ceṣṭitāḥ), their
knowledge and understanding destroyed. Offspring having the
sacraments (saṃskāras) of Brahmins, resolved on doing their
svadharma, were created by the Ṛṣis one after another by their own
austerity. Primal, arisen from the god,22 rooted in brahman,
imperishable and unchanging—that creation is called mental, the
ultimate recourse (or supreme end) of the loom of dharma
(dharmatantraparāyaṇā). (Mbh 12.181)

The term dharmatantra appears several times in the dharmasūtras, and also at
a few other points in the Mahābhārata. As in chapter 8, I take a little liberty
by translating tantra here by its main meaning “loom.” But I believe that it
feels right for this double usage in a text that is not averse to bringing out the
metaphoric power of practical and also sacrificial technologies when speaking
of cosmic processes—most typically “churning,” but also weaving, for which
the term tantra makes other suggestive epic appearances, most notably in
suggesting a comparison between the Mahābhārata’s own self-declared
length of a hundred thousand couplets (1.56.13) and a lost prototype attrib-
uted to a group of ancient Ṛṣis known as the Citraśikhaṇḍins, who, “having
become of a single thought, promulgated23 a supreme treatise (tair ekamati-
bhir bhūtvā yat proktaṃ śāstram uttamam) . . . consisting of a hundred thou-
sand verses, from which proceeds dharma for the entire loom of the worlds
(kṛtaṃ śatasahasraṃ hi ślokānām idam uttamam/ lokatantrasya kṛtsnasya
yasmād dharmaḥ pravartate).”24
In any case, Bharadvāja’s questions make it clear what the passage is
driving at and what Bhṛgu answers to with this metaphor. Toward the end of
his queries Bharadvāja mentions jātis (“kinds of birth,” including demons and

22. I believe Brahmā is implied, but not in a capital letter monotheistic sense.
23. For proktam, see Minkowski 1989, 402, 411–12 concerning pra + vac/proktaḥ as having Vedic over-
tones, with “the sense of an original utterance”; cf. Hiltebeitel 2001a, 98–99.
24. Mbh 12.322.26d and 36. See Hiltebeitel 2005a, 455–56). Other suggestive Mahābhārata usages are
discussed in chapter 8 § D. See also 3.181.11c (plural); 5.29.19a and c. “Loom, fabric, or course of the Law” may
also be pertinent in some dharmasūtra usages; see especially G 13.11 and 18.32, in which it is governed by the root
pīḍ, “to harm” or “oppress,” and B 1.10.8 with the similar upa + Örudh, “to thwart.” Cf. Olivelle 1999, 327 with
dharmasūtra citations and comment: “the exact meaning of the expression dharmatantra is unclear. . . . I think
tantra here means something like ‘the working’ or ‘the execution.’” Manu does not use the term.
532 dharma

of course “subcastes”), interchangeably with varṇas, but what he wants of

Bhṛgu is to hear something decisive on varṇa (varṇaviniścayaḥ) that differenti-
ates the four social classes from the plethora of jātis. Bhṛgu’s decisiveness lies
in making the four varṇas the cut-off line for svadharma. The only real, that is,
original, svadharma is for once that of the Brahmins themselves. While they
provide the archetype, the middle two classes originated from not following
that original svadharma, and the same holds for Śūdras even though in their
case svadharma is not even mentioned, only their fall from purity. As with the
“full development” of Brahmin in Bṛhadāraṇyaka Upaniṣad 1.3.11–14, the
current svadharmas of these classes would, by implication, have to have been
further “ordained” for them once they became distinguishable “by their acts,”
in what the eternal Veda has to say with regard to truth, dharma, austerity, and
pure conduct. These practices are not prohibited for the lower three varṇas, but
it is doubtful they pertain to the jātis—among whom “Piśācas, Rākṣasas, Pretas,
and the various kinds of Mlecchas are motivated by unruly conduct.” Heaven is
open to all four varṇas, and, it would seem, to them alone. These four classes
are, after all, mentioned in the Puruṣa Sūkta. Meanwhile, the other “innumer-
able” jātis, unclassified in the Puruṣa Sūkta, need not apply.
Such a cutoff on svadharma has further implications for other usages of
dharma where we might think it applies, but where precious little if any
guidance is offered as to what it might precisely be. Āśramadharma can also be
svadharma, at least for the first three āśramas. But āśramas, at least in our
classical dharma texts, are open only to the three upper varṇas and designed
pretty much around males. As regards groups, we can quickly grasp that if
jātidharma is not svadharma, then neither is kuladharma or the dharma of those
who live in a region (deśadharma) or village (grāmyadharma). Nor is it possible
to say that epic and Purāṇic gods and demons have svadharma. Most influential
to the contrary has been O’Flaherty, who introduces “the particular duty [i.e.,
svadharma] of demons to be evil” as something that should be appreciated by
“the gratified structuralist” (1976, 64; cf. 149), along with their capacity to vio-
late this svadharma by being good (130). This is virtually a pure hypothetical, for
I am aware of only one context that ties the notion of svadharma to any demons,
and that occurs with reference to the legal category of Rākṣasa marriage. In the
Rāmāyaṇa, Rāvaṇa tells Sītā that “making love to other men’s wives and even
carrying them off by force is the svadharma of Rākṣasas” (Rām 5.18.5). Similarly,
the Rākṣasī Hiḍimbā tells Kuntī that she is abandoning her svadharma (i.e.,
foregoing abduction) to marry Bhīma (Mbh 1.143.7; see chapter 10 § C). Rākṣasa
svadharma would seem to crop up as an ironic exception precisely for its overlap
with Kṣatriya svadharma, since it can help to explain how even noble Kṣatriyas
like Arjuna and Bhīṣma can engage in the Rākṣasa mode of marriage by
 dharma and the bhagavad gı̄ t ā 533

abduction—in their cases, of unmarried women. Otherwise, the notion that

svadharma applies to demons—especially Asuras—being “evil” cannot be con-
firmed in the epics,25 and it would seem to be hard to find even in the Purāṇas.26
Yet O’Flaherty goes on to make further such extrapolations: that Death has a
svadharma to kill (221, 229–31, 235); that Viṣṇu, Śiva, and Brahmā have svadhar-
mas to create, destroy, etc. (225–26; cf. 1973, 139); and that “svadharma can be
transcended by bhakti” (1976, 227; cf. 81–83, 131, 237, 377–78). On the first two
points, the Mahābhārata passages in question again make no mention of such
global divine svadharmas.27 And the last point is bound up with an attempt to
confine Max Weber’s understanding of India’s coordination of caste dharma
with the law of karma—for Weber, “the most consistent theodicy ever pro-
duced” (1958, 121)—to an “orthodox period” (81) between earlier “ascetic” and
later bhakti periods: something that is untenable in general, and more specifi-
cally so for the Mahābhārata28 and the BhG. We must look beyond such grati-
fying syntheses—be they structuralist, historical, or ethical—to ask what it is
that the texts actually say themselves about svadharma, when they begin to say
it, what they say about it, and what they do not say about it.
Most interesting are the cases where the absence of svadharma has to do
not only with classes but with individuals and their predicaments. As we saw in
chapter 10 § C, Hiḍimbā’s comment is one of only five instances where the
epics mention svadharma in connection with women, for whom—as distinct

25. I have checked the Mbh passages O’Flaherty cites in support of this formulation and found that either
they do not use the actual term svadharma., or use it but not in the sense ascribed, In Mbh 3.92.3–12, the Ṛsị
Lomaśa tells Yudhiṣṭhira that demons fall because they lapse into adharma (among other things by not performing
rites), and not attributing this to any svadharma of theirs. 12.221.28 does have Śrī describe how the demons origi-
nally did their svadharma, but by being good in all the right ways rather than by being evil: they fell to bad ways (48)
with their wives and children (59–61), began to mix castes (64), to enjoy sports that involved cross-dressing (67),
and to allow daughters-in-law to lecture and rebuke their husbands in front of in-laws, the husband’s parents (75),
whereupon Śrī then came over to the Gods and Indra, along with eight other goddesses (81–82). Their deteriora-
tion, which would seem to have perfectly human models, does not result from any asuric svadharma to be evil.
26. I cannot check all of O’Flaherty’s numerous Purānị c references, in some of which her synthesis may
apply. But where she refers to the svadharma of demons in Vāmana Purānạ 11–16 (1976, 129–31), the passage’s two
sole usages of svadharma refer it only to the fourfold human social order and the āśramas suitable for specific varṇas
(VāmP 15.64–66), and not to demons and other nonhuman beings, who have only dharmas (11.15–27), not svadhar-
mas. Likewise, in this purānạ ’s lengthy account of the rise and fall of the Asura Bali, although a demon may be said
to have a wicked or cruel “nature” individually as his svabhāva (VāmP 59.26, 31), the only references to anyone’s
svadharma again pertain to those in the four varṇas (48.14; 49.13)—this, despite the fact that Deborah Soifer draws
this passage into a discussion of Bali’s alleged demonic svadharma that relies on O’Flaherty’s categories and period-
ization (1991115–30). Soifer seems to hint at the difficulty where she mentions that the gods find Bali’s rule unten-
able because he is “devoted to his asuric svadharma (which is adharma) or . . . more frequently devoted to sanātana
dharma” (119). Soifer’s translations supply numerous instances of Bali’s problematic allegiance to sanātanadharma,
but the only instance I see of svadharma again refers it to the humans over whom he rules (Soifer 1991, 243).
27. Hill 2001 has also worked from these premises to argue that demons’ svadharma is adharma (104–14),
and to speak of “Skanda’s svadharma” (122) in a Mbh passage (3.207–21) that never mentions it. See also Leslie
1989, 265, comparing (otherwise astutely) the dharmas and svabhāvas of demons and women.
28. See my discussion in chapter 1 § B of O’Flaherty 1976’s “heavy marination in historical periodization.”
534 dharma

from Manu’s one usage—it seems to leave open some ironic spaces for negoti-
ation. Indeed, Bhīṣma seems to treat svadharma ironically especially in the
Mokṣadharma anthology. As in his presentation of Bhṛgu’s discussion of varṇa,
he gives the topic of women’s svadharma a special ironic twist in the only case
mentioned by him, that of Sulabhā, who overturns King Janaka’s varṇa, āśrama,
gotra, and gender fixations from the standpoint of her nivṛtti argument. But she
still must make her argument in those terms. All the usages of women’s sva-
dharma, including that of Hiḍimbā, are voiced in the context of marital concerns
that answer to categories of varṇa.
Meanwhile, although kings get the merits (dharma) of others’ svadharma
in various omnibus ways,29 it is a real challenge, as Yudhiṣṭira’s predicament
constantly demonstrates, to figure out what that means when one is constantly
reminded that his own real svadharma is his kṣatriyadharma to give, study,
sacrifice, fight, and kill or be killed. Here is what Kṛṣṇa has to tell Yudhiṣṭḥira
on this score in the very first words of the visit he makes with Satyabhāmā to
see the Pāṇḍavas and Draupadī in the forest:

Dharma is higher than obtaining a kingdom; they say austerity is its

goal, king. While you did your svadharma (caratā svadharmam)
uprightly and truly, you have won this world and the one beyond.
At first you studied, practiced vows, having acquired the entire lore of
the bow (dhanurveda); Having acquired wealth by kṣātradharma,
surely all the ancient sacrifices were obtained. You did not find
pleasure in village laws (grāmyadharmeṣu), Indra among men, pursue
desires (kāmān), or abandon dharma out of greed for gain (artha).
And thus by nature (svabhāvād) you are King Dharma. (3.180.16–18;
see van Buitenen 1975, 571 for the last sentence)

More bellicose is what Kuntī says Kṛṣṇa should tell Yudhiṣṭhira about these
matters when he returns to the Pāṇḍavas from his prewar “peace” embassy to
the Kaurava court. Telling him to warn Yudhiṣṭhira, “Your dharma is greatly
declined” (5.130.5c), she finds her son’s intellect to be stricken by rote learning
from seeking only one dharma (6cd); Kṣatriyas were created to protect subjects
and a king gets a fourth of their dharma from doing so (11); a daṇdạ -wielding
king is worthy of divinity and his svadharma should be the application of the

29. See Mbh 12.65–66 where Bhīṣma, trying to deter Yudhiṣṭhira from abandoning the householder
āśrama for a renunciant one, says that rājadharma is itself a kind of omnibus āśrama—which curiously pertains
to the confusion a king must address in a kingdom where the Ārya varṇas mix with varied Mlecchas (Greeks,
Chinese, Śakas, etc.), and in which dasyus (robbers, local antisocial elements, including what the British called
“Criminal Tribes”) are “even following the four Life-Patterns, though under different outer manifestations
(liṅgāntare)” (65.23). See Fitzgerald 2004a, 732–33 for a note on this terminology.
 dharma and the bhagavad gı̄ t ā 535

rod of force (daṇḍanītī) that regulates the four varṇas (13); such a king creates a
Kṛtayuga, lesser kings middling yugas, and a wicked king goes to hell (14–19;
see chapter 7 § A.3). She does not approve this son’s noncruelty (ānṛśaṃsya),
and neither she, Pāṇḍu, nor “Grandfather” (presumably Bhīṣma, since Vyāsa
does endorse this virtue, as we saw in chapter 9) ever prayed that he be blessed
by the wisdom he lives by (20–21)! Disappointed in having given birth to a son
who leaves her on the dole living off the Kauravas (31), she concludes, “Fight by
rājadharma, don’t drown your grandfathers. Do not, with your merit (puṇya)
exhausted, take an evil turn with your brothers” (32). The rationale for the
king’s svadharma seems quite convoluted. Perhaps this merit is the cumulative
dharma he is neglecting in not overseeing others’ svadharma, which should be
his own svadharma, which is still nothing but his Kṣatriya dharma.
Fortunately or not, then, most guidance on a king’s svadharma is couched
as kṣatriyadharma. Yudhiṣṭhira is inundated with the latter, and contemplates
its implications with some aversion as he speaks to Kṛṣṇa preparing for the
virtual inevitability of war: “Kṣatriya kills Kṣatriya, fish lives on fish, dog kills
dog—see dharma as it comes down (paśya dharmo yathāgataḥ!)” (5.70.48).
Ganguli translates this misleadingly, implying that dogs and fish have svad-
harma: “Behold . . . how each of these followeth his peculiar virtue” (162). But
these are proverbial references to what kṣatriyadharma is like as “custom”
(āgama), not cases of animal svadharma, which is even harder to find than
demon svadharma and the svadharma of women and kings. More attuned to
such a dog eats dog world, and more receptive to Kṛṣṇa’s guidance on Kṣatriya
svadharma, is Yudhiṣṭhira’s dashing younger brother Arjuna, who becomes
especially receptive in the Bhagavad Gıtā ̄ .

C. Manu and the Bhagavad Gıtā

̄ : Two Kinds of Karmayoga

The term karmayoga makes only one appearance in the dharmasūtras, where, in
Olivelle’s translation, it means “ritual use” when Āpastamba says, “The suspension
of vedic recitation laid down in the vedic texts refers to vedic recitation and not
the ritual use of vedic formulas (na karmayoge mantrānạ m ̄ )” (Ā 1.12.9).30 Now, just
as the BhG offers a distinctive presentation of svadharma as a concept found
prominently in both Manu and the Mahābhārata, something similar may be said
of karmayoga, but this second term has more restrictive usage in each text. In this
case, while Manu and the Gıtā ̄ each offer distinctive interpretations of karmayoga,
elsewhere in the Mahābhārata the term seems barely adrift. It “occurs five times

30. Olivelle (1999, 22) cites Pūrvamīmāṃsāsūtra 12.3.19 on this, suggesting it corresponds to a Mīmāṃsā
usage.
536 dharma

in the Gıtā̄ (at 3.3 and 7, twice in 5.2, and at 13.24) and only four other times
throughout the MBh” (Fitzgerald 2002, 650 n. 5), none of which gives the impres-
sion of being affected by its BhG usages or even of developing the term in any
direction.31 At Mbh 3.33.50, Draupadī tells Yudhiṣṭhira that “in the ‘disciplines of
action’ (plural) valor is the principal force” (bhūyiṣṭhaṃ karmayogeṣu sarva eva
parākramaḥ). As noted in chapter 10 § D, she seems in this speech to anticipate
things that Kṛsn ̣ ạ will say in the BhG, but not in this case. The other three occur-
rences come in Bhīṣma’s Mokṣadharma teachings in the Śānti Parvan. At
12.194.11, where Bhīṣma tells how Manu responded to Bṛhaspati’s request to dis-
tinguish the fruits of knowledge from those of action (9), Manu says something
like “‘enjoinments to action’ (plural) in(volving) Vedic chanting have the nature
of desire” (kāmātmakāśchandasi karmayogā). As we will see, the quoted Manu
sounds like the text Manu on the point about desire. The usage, if not the point,
is also similar to Āpastamba’s mention, just cited, of “the ‘ritual use’ of vedic
formulas.” At 12.206.11, Bhīṣma himself, in a rant against women, describes
how an intelligent person should disdain children as vermin because they are
born “from a natural ‘enjoinment to activity’” (svabhāvātkarmayogād vā tānupekṣeta
buddhimān). Here, as in the Gıtā ̄ , there is a connection between karmayoga and
svabhāva (“natural” here), but Bhīṣma is hardly making the same point. And at
12.286.17, Bhīṣma quotes Parāśara on the topic of reincarnation to have said,
“Caused by ‘union with acts,’ one is born here and there” (bhāvitaṃ karmayogena
jāyate tatra tatra ha). Recall too that the sole usage in the Rāmāyaṇa describes
Hanumān as “settling into his own plan of action” (svakarmayoga ca vidhāya;
5.40.30). The originality of the Gıtā ̄ stands out against this background, and calls
for restraint, as I pointed out in chapter 9,32 when it comes to reading the Gıtā ̄ ’s
̄
karmayoga teaching into the rest of the text. More so even than svadharma, it
has special import to Arjuna in coming just before the Mahābhārata war. Meanwhile,
Manu’s formulations on karmayoga are concentrated in its final and most
“philosophical” chapter 12 on the fruits of action and the process of reincarnation.
To read translations of these texts, one would never gather that they are
rendering the same term. Although Manu has a distinctive take on karmayoga,

31. See, however, Fitzgerald 2002, 641, 647, who positions the karmayoga doctrine to be developing toward
̄ .
its “late” and apparently “Gupta” articulation in the Gıtā
̄ . Dhand 2004
32. See chapter 9 § E.2.b on Arjuna’s “free pass” through the war from having heard the Gıtā
provides a questionable extension in her discussion of the Śūdra woman’s easy manner in having a substitute sexual
union with the Mahābhārata’s smelly ascetic author Vyāsa (see chapter 8 § D), and even more implausibly makes
such a case for Rāma in his banishment of Sītā (2002). Similarly, I am not persuaded by Brodbeck that the impulse
of epic males to derail their marriages is a “gendered soteriology” that allows “continued performance of social
dharmas in a spirit of karmayoga” (2007a, 145, 159, 164), citing in the same vein Dhand 2007—both extending the
principle by analogy to other figures. See likewise Brodbeck 2009c, 47–49, generalizing it to a warrior ideology,
lokasaṃgraha, and the deva-asura conflict. Whatever one may read into them, the epics do not make these moves.
 dharma and the bhagavad gı̄ t ā 537

it almost seems that Manu’s translators have rendered it in ways that would
avoid suspicion that they were contaminating Manu by a Gıtā ̄ reading. But
clearly it is the same term, and I think the two texts’ usages probably have
some kind of relation to each other in the pivotal ways they each position the
concept. In both texts one can easily tie in the usages of karmayoga with those
of svadharma and svakarma as they relate to theories of ritual and ideologies
of sacrifice; but the theory and ideology differ in each text. In each case
karmayoga is clustered with a different range of concepts.
In Manu’s very first usage of karmayoga, the motivation for action entails
desire in a way that is difficult to square with the Gıtā̄ ̄’s ideal of desireless
action:33

To be motivated by desire is not commended, but it is impossible here

to be free from desire (akāmatā); for it is desire that prompts vedic
study and the performance of vedic rites (karmayogaś ca vaidikaḥ).
Intention (saṃkalpa) is the root of desire; intention is the wellspring of
sacrifice and intention triggers every religious observance and every
rule of restraint—so the tradition declares. Nowhere in this world do
we see any activity done by a man free from desire (akāmasya), for
whatever at all that a man may do, it is the work of someone who
desired it. By engaging in them properly, a man obtains the world of
the immortals and, in this world, obtains all his desires just as he
intended. (2.2–5; Olivelle trans. 94, dubbing this an “excursus”)

Christopher Framarin (2006) argues that Manu’s views in these verses are
̄ ’s
̄ than has been thought, but subsequently acknowledges that
closer to the Gıtā
there are also significant differences (Framarin 2009, 88–91, 93). In brief,
“Unlike the Manusmṛti, however, the Gıtā̄ does not limit the advice to act without
̄ ’s advice to act without desire
desire to a narrow class of ritual actions. The Gıtā
for ends extends to all actions” (93). Framarin’s argument that Manu calls for
action without desire relies on a reading of saṃkalpa, in the second verse above,
by the eighth- or ninth-century commentator Medhātithi. It would seem that
Medhātithi seeks to square Manu 2.2–5 with the Gıtā ̄ ’s desireless action by
interpreting saṃkalpa “not [as] an intention, but a belief or cognition” (Framarin
2006, 402; 2009, 79), and taking Manu to refer to nityakarmas (“ritual actions

33. The BhG associates karmayoga with “unattached (asakta) action or karma” (3.19) that is “devoid of the
intention (saṃkalpa)” to achieve a ritually defined desire (4.19), and, moreover, “renouncing without exception all
objects of desire that are rooted in intentions (saṃkalpaprabhavān kāmāṃs tyaktvā sarvān)” (6.24). This is else-
where called niṣkāma-karma or “action without the desire for its fruits.” As Bagchee (in press) points out, the BhG
itself does not use this term, which seems to have been coined for it by commentators—although we will see it
at Manu 12.90.
538 dharma

for which no result is mentioned in the Vedas”), in which “the desire that is
prohibited is the desire for the fruit,” as opposed to kāmyakarmas (“ritual actions
performed for the sake of some phala”) (Framarin 2006, 404–5; 2009, 83–84
for the definitions). Framarin then interprets saṃkalpa here as a “belief” that
motivates desire “for the means to one’s end, since it is the result of a means-
end belief,” the means being “the most important thing” (2006, 403, 405).
Medhātithi’s interpretation puts Manu in the position of “prohibiting” certain
desires, which Manu does not do, and “the most important thing” would, I think,
still be a “desire for dharma” (dharmakāma) as a means. Note that Framarin follows
up this comparison of the BhG and Manu with a “reconsideration” of Yogasūtra 1.25,
where, “if Īśvara has a desire at all—and remember, the Yogasūtra does not say
this—it must be a desire for the means of teaching knowledge and dharma”
(Framarin 2009, 92, referencing 38–39). The phrase dharmakāma is not found
in Manu, but is common enough in the Mahābhārata, where, for instance,
Draupadī blames the Pānd ̣ ạ vas’ troubles on Yudhiṣṭhira’s “desire for dharma” at
3.34.8, and Mārkaṇdẹ ya prophesies that Brahmins will be “desirous of dharma”
after Kalki has brought back the Kṛta yuga (3.189.11). In terms of telling us
what Manu actually says, Brodbeck finds Framarin’s “recourse here to
Medhātithi . . . unconvincing” (2009b, 137), as I do, whereas Bagchee (in press)
appreciates Framarin’s vigorous articulation across texts of a recurrent Indian
epistemology of desireless action. Since I am not concerned with the literal via-
bility of desireless action in the Gıtā ̄ , which is Framarin’s main concern in
critiquing “non-literal” readings of it by Brodbeck and others,34 the matter of
whether Manu might be consonant with it is not pressing to this discussion.
Brodbeck concedes that Framarin “convincingly defends the notion of literally
desireless action against clear and present lines of attack” (2009b, 137).
Now as Brodbeck observes, in Manu 2.2–5, “performance of vedic rites,” or,
as he translates it, “‘engagement in Vedic action,’ may also be translated . . . ‘Vedic
karmayoga’” (2004, 85).35 As noted, Manu compounds karmayoga primarily with
ritual rules (and thus implicitly with both svadharma and svakarma), but never
relates these concepts to any usage of svabhāva, which it keeps to two topics: the
“natural range” of the black buck (2.23), and the “very nature” of women, which is
“to corrupt men.”36 Manu’s other uses of karmayoga are enough to indicate that

34. See Framarin 2009, 8–14, critiquing Brodbeck 2004, which will be discussed.
35. Brodbeck 2004 marks the contrast with the BhG but without, in that article, addressing the Gıtā ̄ ’s
usages of karmayoga directly.
36. See M 2.213 (Olivelle 2005a, 105). See also M 9.15–20, beginning, “Lechery, fickleness of mind, and
hardheartedness are innate (svabhāvataḥ) in them; recognizing thus the nature (svabhāvam) produced in them at
creation by Prajāpati, a man should make the utmost effort at guarding them.” These usages of svabhāva parallel
that of svālakṣaṇya to similarly describe women’s “true character” as promiscuous at the end of this passage.
 dharma and the bhagavad gı̄ t ā 539

they are not late afterthoughts to the text, and Olivelle includes only the first usage
among what he calls “excurses.” At 2.68 Olivelle translates the term karmayoga as
“the activities connected with” the upanayana or sacred thread ceremony, and this
is in a transitional verse with what he regards as Manu’s signature transition-
marker, nibodhata, “Listen”; at 6.86, in another transitional verse with nibodhata,
Olivelle translates karmayoga as “ritual discipline” in “the ritual discipline of vedic
retirees” (vedasaṃnyāsins); at 10.115, among the seven means of acquiring wealth,
it surely means more than just “work”;37 and at 12.2, in something like the transi-
tional verses, there is a return to the frame where the great Ṛsị s are told, “Listen to
the determination with respect to the engagement in action (karmayoga).” Finally,
in chapter 12, where Manu unfolds “Vedic karmayoga” most fully, he does so in
relation to ideas of pravṛtti and nivṛtti (important also in the BhG as well as in the
Nārāyaṇīya) that bear especially on reincarnation and mokṣa:

One should understand that acts prescribed by the Veda (karma

vaidakam) are always a more effective means of securing the highest
good both here and in the hereafter than the above six activities.38 All
these activities without exception are included within the scheme of
the acts prescribed by the Veda (vaidike karmayoge), each in the proper
order within the rules of a corresponding act (kriyāvidhau). Acts
prescribed by the Veda are of two kinds: advancing (pravṛttam), which
procures the enhancement of happiness; and arresting (nivṛttam),
which procures the supreme good. An action performed to obtain a
desire here or in the hereafter is called an “advancing act” (pravṛttam
karma), whereas an action performed without desire (niṣkāmam) and
prompted by knowledge is said to be an “arresting act” (nivṛttam). By
engaging in advancing acts, a man attains equality with the gods; by
engaging in arresting acts, on the other hand, he transcends the five
elements. (M 12.86–90; Olivelle trans. 2005a, 234)

Again, “the scheme of the acts prescribed by the Veda (vaidike karmayoge)”
could be translated “Vedic karmayoga.” It is not clear how one is to reconcile
Manu’s earlier statements that no action was ever done “without desire”
(akāma) with this allowance for nivṛtta actions to be niṣkāma. In context, it
appears to be an allowance for what the Nārāyaṇīya calls nivṛtti-dharma (or
what the BhG calls jñānayoga), differentiating it from “Vedic karmayoga” rather

37. Olivelle’s translation; Doniger (1991) offers “working,” and Bühler [1886] 1969 “the performance of
work.”
38. M 12.83: the six are Vedic recitation, tapas, knowledge, controlling the senses, noninjury, and service
of the teacher.
540 dharma

than correlating the two.39 But one thing is definite: Manu does not subordinate
“Vedic karmayoga” to bhaktiyoga or to any ideas about “inherent natures” that
underlie dharmic actions.
̄ , as is well known,40 karmayoga is taught by a Kṣatriya incar-
In the G ıtā
nation of God, himself the ultimate karmayogin, to a Kṣatriya prince. Once
Kṛṣṇa has introduced the term at BhG 3.2–7, he relates it to sacrifice in its
“cosmogonic function as being an integral part of creation,” leading up to
his indication of how he himself acts in the three worlds with nothing to
accomplish (3.9–24). Then, after Kṛṣṇa mentions it again as “higher than
the renunciation of acts” (karmasaṃnyāsād) (5.2), he teaches in chapters 5
and 6 “why karmayoga is possible” as “one of the most important features of
the cosmic cause of all activity,” leading up the invitation to give up one’s
attachment to action by offering it up to Kṛṣṇa “as the only agent and lord.”41
By this point, karmayoga is arrayed with the two other yogas or “disciplines”
of jñānayoga and bhaktiyoga. Like Īśvara in the Yogasūtra, Kṛṣṇa may have no
desire except to teach dharma, as Framarin makes plausible, pointing out
that BhG 3.25 is “one of the only passages in the G ıtā̄ in which Kṛṣṇa seems
to endorse desire. The passage reads: ‘As the ignorant act attached in action,
O Arjuna, so the wise should act without attachment, desiring (cikīrṣur)
lokasaṃgraha’” (2009, 90, citing Sadhale trans. 1985, 25–26)—that is, “the
holding together of the world.” As Bagchee (in press) comments on
Framarin’s discussion of this desiderative usage of Ökṛ, “to do,” “The wise
person’s desire is not a desire for a specific end (which would dispose him to
joy or disappointment depending upon the outcome): it is a ‘desir[e] to do.’ ”42
Indeed, at 7.11cd, Kṛṣṇa says, “I am the desire in beings that does not run
counter to dharma (dharmāviruddho bhuteṣu kāmo ‘smi bharatarṣabha)” (cf.
van Buitenen 1981, 99). In any case, Kṛṣṇa explains svadharma and svakarma
against this background, but compounded by the additional concept of
svabhāva, “inherent nature.” That is, according to Kṛṣṇa, doing one’s duty
and occupations properly springs ultimately from one’s “intrinsic,” “innate,”
or “inherent nature.” Of all the places in the Mahābhārata that present exten-
sive discussions relating svadharma and svakarma,43 the G ıtā ̄ presents the

39. As Brodbeck 2009b, 138 points out in questioning Framarin’s view that Manu 2.2–4 teaches desireless
action in a way consonant with the BhG, “Manusmṛti 12.88–90, which Framarin does not mention, presents a
more compatible view.”
40. See van Buitenen 1981, 12, 18–20; Brodbeck 2004; Hill 2001, 331–34, 342, 351; Woods 2001, 71–76,
143, 172, 182; Sutton 2000, 65, 126, 137, 330.
41. I paraphrase and quote from Malinar 2007a, 81–82, 128–29, 155.
42. See BhG 18.30, describing the sattvic temperament that can discriminate pravṛtti from nivṛtti, what is
to be done (kārya) from what is not to be done (akārya), and bondage from mokṣa.
43. This occurs for instance in Hanumān’s encounter with Bhīma (3.148.17; 149.25–50); in the “Colloquy
of the Brahmin and the Hunter” (3.198.25–38; 199.14–15, 34); and in the “Instruction of Śuka” (12.309.46–90;
cf. 2.50.6–7; 12.67.30–31; 12.107.14–16; 12.110.29–24l).
 dharma and the bhagavad gı̄ t ā 541

only case where these terms are compounded by this additional “inherent”
grounding, which is ultimately a grounding in Kṛṣṇa’s lower nature or
prakṛti.
Now as Brodbeck points out, if we are to understand the Gıtā ̄ as having any
impact on Arjuna, “the notion of svabhāva used here must logically be specific
to individual people rather than to individual varṇas. We would even want to go
further and describe svabhāva as variable within one lifetime” (2004, 90); it
must be “a continuously varying quality” (99) if it is to have any bearing on the
change Arjuna undergoes from the Gīta’s beginning to its end. Moreover, this
manner of acting that Kṛṣṇa teaches Arjuna is presented as having “universal
applicability” that extends rhetorically to “the text’s audience” (81–82). Although
Brodbeck now seems to be persuaded by Framarin that it can be taken literally,
he argues—I think plausibly enough in context—that “the availability of non-
attachment in action functions as a narrative fiction to explain, on the conven-
tional level, how Arjuna can satisfactorily be persuaded to fight”; “the universal
applicability of Kṛṣṇa’s technique is a conceit of the way in which the text
reports Arjuna’s changing his mind” (2004, 100). There is in fact nothing to
directly indicate how or whether Arjuna actually understood or benefitted from
Kṛṣṇa’s karmayoga teaching, and some later evidence that it did not stick:
Arjuna tells Kṛṣṇa that what he said in the BhG is “lost” (naṣṭam) to him and
that he is curious to hear it again (Mbh 14.16.6–7)! To which Kṛṣṇa, not at all
pleased (10), replies with the Anugītā in Book 14, never getting back to the term
karmayoga, which we could say—the Mahābhārata war now being over—is lost
for good. My sense of Kṛṣṇa’s teaching of karmayoga in the BhG is that it is
something like Gutei’s (Chü Chih’s) “one finger Zen.” Whenever Gutei was
asked a question,

all he did was raise one finger. Once a disciple of his was asked by
a visitor: What is the main point of the Law your master teaches?
So the disciple raised his finger. . . . (Beyer 1974, 263)

For Gutei himself, like Kṛṣṇa, it may have been “more than I could use up in
an entire life time,” and it brought some sudden enlightenment to the “little
disciple” when he got his finger cut off. But it was not something that Gutei or
anyone else could ever teach in the same way again.
Moreover, we may note that when Kṛṣṇa first speaks to Arjuna about
karmayoga at BhG 3.3, it is in answer to a question of Arjuna’s that he answers,
if at all, only by deflection or perhaps a long delay: “If more important than
action The mental attitude is held of Thee, Janārdana, Then why to violent
action (karmaṇi ghore) Dost Thou enjoin me, Keśava?” (BhG 3.1; Edgerton trans.
1952 33). As Brodbeck says, “There is no getting around it: the extent of Kṛṣṇa’s
‘rational assessment of the situation’, at least as far as ethics is concerned,
542 dharma

is that Arjuna is a kṣatriya and so must—and will—fight” (2004, 98). That is,
Kṛṣṇa’s teaching of karmayoga is tailor-made for the consummate warrior
Kṣatriya, and not for the king (or at least not for Yudhiṣṭhira).
One of the questions driving this chapter so far is by now no doubt clear
enough. Out of the nexus of Kṛsn ̣ ạ ’s teachings, many wondrous things contribute
to making Arjuna lose track of his lingering question, “Then why to violent
action?” One of these is the idea, on which the Mahābhārata and Manu 7.87 agree,
that warriors who die in battle go to heaven, which Kṛsn ̣ ạ has already delivered as
part of his shock treatment (BhG 2.33–37) before he gets to any yogas or “disci-
plines.” Manu also declares that one should never kill an animal out of desire, that
“killing in sacrifice is not killing,” that when plants and animals die in sacrifice
they earn superior births, and that “when killing is sanctioned by the Veda it
should definitely be regarded as non-killing” (5.37–44)—all of which Brian
K. Smith assigns to the “fog machine” (Doniger and Smith 1991, xlii). In February
2006, I read a dissertation proposal on the “inner jihad,”44 and remarked to its
author, Waleed El-Ansary, that it sounded like karmayoga, by which of course
I meant the BhG’s karmayoga, with which he was prepared to agree. Like “inner
jihad,” Arjuna’s inner struggle is one thing, and deserves respect, even if he seems
to be forgetful; but arguments to kill others because God says so are something
else, and, as I went on to say, an agony of our times, and obviously not of our times
alone. Of the cluster of ideas that Kṛsn ̣ ạ assembles in the BhG to get Arjuna to
fight, it is not svadharma alone that must be singled out. Kṛsn ̣ ạ ’s teaching to kill
with indifference runs the risk of winning the warrior who has not quite mastered
it the same prize—heaven being just a favorable rebirth—as the sacrificial goat.

D. Where Kṛṣṇa Is There Is Dharma

We have now begun to notice that the BhG has ripples in the wider Mahābhārata
text. Would that this could bring pause to the long debate over whether the BhG
is an interpolation entirely, or is one all but for a few old verses near its beginning
(possibly its middle) and its end that belong to some kind of original tribal war
narrative—in either case with the shared premise that what is interpolated
would itself be a patchwork of layered inserts. Among those who have enlivened
this debate, there is no serious consideration that the BhG could be an integral
unit of a larger whole, and no questioning whether the tools of the so-called
“higher criticism” have been fruitfully applied in the Gıtā ̄ ’s case. Suffice it to say
that the few cumulative results of this approach have been based on unsound

44. See now El-Ansary 2006, 209–69 for a hermeneutic also of the “outer jihād.”
 dharma and the bhagavad gı̄ t ā 543

assumptions.45 Yet some scholars taking this stratifying approach have recog-
nized that the BhG does resonate with other units, passages, and themes in the
Mahābhārata. They have thus added to cumulative results on this front—though
with the difference that, for them, such correspondences show either layers of
interpolation that coincide with those in the BhG, or the BhG’s influence on the
larger text, which would make these other units, passages, or themes later even
̄ (or of the corresponding layers in the Gıtā
than the Gıtā ̄ ).46 As one of those who
views the BhG as an integral unit of the larger epic as conceived, I take the
opposite view on this more central matter that still eludes real debate. But the
important point for now is to register that resonances with the BhG are not
limited to ones that foreshadow or build up to the beginning of the war narra-
tive, such as Arjuna’s charioteering for the fainthearted Matsya prince in their
duel with the Kauravas that ends the Pānd ̣ ạ vas’ thirteenth year of exile in dis-
guise “as a precursor and, in a way, a parody” of the Gıtā ̄ and of Kṛsn
̣ ạ ’s chario-
teering for Arjuna (Goldman 1995, 94), or in exchanges during the prewar
Udogaparvan.47 They also include philosophical points made in debates with
Yudhiṣṭhira by Arjuna and Vyāsa after the war is over,48 and even by Draupadī
in the forest well before the war has begun, as we saw in chapter 10 § D. Of
course this is not to say that these dialogical interludes are not part of the war
narrative. They are. But why should not a war narrative be interrupted by
philosophical, ethical, and theological exchanges in the Mahābhārata? Or indeed,
further supplemented by such exchanges in the Nārāyaṇīya or the Anugītā?
These questions position us to begin our discussion, announced at the
beginning of this chapter, of how the BhG presents dharma through a ring struc-
ture. From ripples to rings is of course (for a stone’s throw) just a movement

45. See Adluri and Bagchee forthcoming for a discussion of German BhG scholarship, all of which (except
Joseph Dahlmann) has proceeded from the shared assumption that a war narrative would not be interrupted by
a philosophical discussion! See Hiltebeitel forthcoming-a and chapter 1 § B on the “tribal” premise backing this
extraordinary assumption.
46. See Grünendahl 1997; Malinar 1997, both on correspondence between the BhG and the Nārāyaṇīya,
deemed on the whole later than the BhG, with Grünendahl arguing that such correspondences derive from a very
late coating that would include a unit called the Viśvopākhyāna, discussed above in the text; cf. chapter 6 n. 42.
47. This is the starting point for Malinar’s discussion of the background of the Gıtā ̄ in debates over war
and peace (2007a, 35–53). Cf. Hiltebeitel [1976] 1990, 14–40 on Udyogaparvan-BhG continuities in the two the-
ophanies there of Kṛṣṇa.
48. See Hiltebeitel 2005d, 251–58, “The Bhagavad Gıtā ̄ and the Cheering Up of Yudhiṣṭhira,” especially 256
on Mbh 12.32.11–24. After the war, Vyāsa speaks implicitly to the pertinence of the Gıtā ̄ ’s teachings to Yudhiṣṭhira,
offering four perspectives on what accounts for action: the Lord, man, chance, and karma, saying this about the first:
“When men who have been commanded by the Lord (īśvareṇa niyuktāh)̣ do a good or bad deed, the consequences
of that deed go to the Lord. For obviously if a man were to chop down a tree in the forest with an axe, the evil would
belong to the man doing the chopping and not at all to the axe” (Fitzgerald 2004a, 241–42). As in BhG 11.33d, a king
using the daṇdạ , like Arjuna taking up his weapons, would, like the axe, be the “mere instrument” (nimittamātram)
of the Lord. Vyāsa concludes that from this standpoint, “It would not be right that one should acquire consequences
effected by another. Therefore assign it to the Lord (tasmāc ca īśvare tan niveśaya)” (cf. BhG 12.6–8).
544 dharma

inward. In this section, I still discuss ripples outward from the BhG, beginning
from adhyāyas on the edges of the BhG proper (Mbh 6.23–40) in the “Bhagavad
̄ Sub-book” or upaparvan called the Bhagavadgītāparvan (Mbh 6.14–41).
Gıtā
There, on both sides of the eighteen-adhyāya BhG, we find a formulaic verse
̣ ạ and dharma that ripples out into the wider text. After tracing how
relating Kṛsn
this theme ripples out from the BhG proper into the wider Mahābhārata, in the
final section of this chapter I will zero in on rings of dharma in the BhG itself.
We thus turn to a well-known set of signature Mahābhārata formulae.
Setting them initially in boldface, let us begin with usages of these formulae
closest to the BhG proper that occur in verses within the Bhagavadgītāparvan.
We will then look into their recurrence, contextualizations, and wider reverber-
ations. On the near side of the Gıtā̄ , we have Arjuna tell Yudhiṣṭhira:

tyaktvādharmaṃ ca lobhaṃ ca mohaṃ codyamam āsthitāḥ/

Abandon adharma, greed, and delusion, be enterprising,
yudhyadhvam anahaṃkārā yato dharmas tato jayaḥ//
and then fight without thought of ego. Where dharma is there is victory.
evaṃ rājan vijānīhi dhruvo ‘smākaṃ raṇe jayaḥ/
So know, king, that our victory in battle is assured,
yathā me nāradaḥ prāha yataḥ kṛṣṇas tato jayaḥ//
for as Nārada has told me, “Where Kṛṣṇa there is victory.”

anantatejā govindaḥ śatrupūgeṣu nirvyathaḥ/

Govinda of infinite splendor strides unconcernedly among the
multitude of his enemies,
puruṣaḥ sanātanatamo yataḥ kṛṣṇas tato jayaḥ//
the most everlasting Puruṣa. Where Kṛsn ̣ ạ is there is victory.
(6.21.11–12, 14)

̄ ’s far side, Droṇa tells Yudhiṣṭhira:

And on the Gıtā

yato dharmas tataḥ kṛṣṇo yataḥ kṛṣṇas tato jayaḥ/

Where dharma is there is Kṛṣṇa, where Kṛṣṇa is there is victory.
yudhyasva gaccha kaunteya pṛccha māṃ kim bravīmi te//
Go and fight, Kaunteya. What can I tell you? Ask me. (6.41.55)49

Within these verses, we find second half-lines (pādas) telling us, “Where dharma
is, there is victory” (6.20.11d), or “Where Kṛṣṇa is, there is victory” (6.20.12d;
6.20.14d; 6.41.55b). Preceding the last of these pādas, we also find an initial

49. See van Buitenen 1981, 64–65 and 150–51 for the text and his translation, which I largely follow here,
of these first two in-close citations. One of the attractions of his translation of the full Bhagavadgītāparvan is that
it showcases this initial framing.
 dharma and the bhagavad gı̄ t ā 545

first pāda telling us, “Where dharma is, there is Kṛṣṇa” (6.41.55a), giving us a
complete combining of these two formulas. This line’s indication that dharma
and Kṛṣṇa go together also occurs in reverse form in an initial pāda followed by
the other of our two closing pādas. Indeed, soon after the Bhagavadgītāparvan,
we hear Bhīṣma tell Duryodhana:

rājan sattvamayo hyeṣa tamorāgavivarjitaḥ/

O king, made of goodness, this one is divested of darkness and passion.
yataḥ kṛṣṇas tato dharmo yato dharmas tato jaya//
Where Kṛsn ̣ ạ is there is dharma, where dharma is there is victory.(6.62.34)

Let us break these formulas down and contextualize their usages.

The first thing to note is that Kṛsn ̣ ạ and dharma go together: either by the
phrase “Where dharma is there is Kṛsn ̣ ạ (yato dharmas tataḥ kṛsn ̣ ọ )” (6.41.55a, just
after the Gıtā ̄ proper; 9.61.30c), or its reverse, “Where Kṛsn ̣ ạ is there is dharma
(yataḥ kṛsṇ ạ s tato dharmo)” (6.62.34c; 13.153.39c).50 In either case, all these state-
ments occur at the beginning of a line or hemistich in which they serve as the
major premise to a subordinate but emphatically punctuated sequel pāda in which
the second term mentioned in the initial pāda becomes the f irst in the second.
Thus in the first two cases, where Kṛsn ̣ ạ is mentioned after dharma, it is “Where
̣ ạ is there is victory (yataḥ/yathā kṛsn
Kṛsn ̣ ạ s tato jayaḥ)” (6.41.55b; see also 9.61.30d);
and in the second pair where dharma is mentioned after Kṛsn ̣ ạ , we find “Where
dharma is there is victory (yato dharmas tato jayaḥ)” (6.62.34d; see also 13.153.39d).
Of these two subordinate claims, the first—“Where Kṛsn ̣ ạ is there is victory”—
occurs four times independently (1.197.25d; 5.66.9d; 6.21.12d and 14d); and the
second—“Where dharma is there is victory”—occurs eleven times independently
(5.39.7d; 5.141.33d, 5.146.16d, 6.2.14d, 6.21.11d, 6.61.16d, 6.117.33d, 7.158.62b,
9.62.58d, 11.13.9d, 11.17.6d), with all fifteen of these more curtailed usages kept to
second pādas where they follow from other initial premises or lead-ins. There is
obviously a circularity between the three terms dharma, Kṛsn ̣ ạ , and victory (jaya),
but with only the latter mentioned in all these citations, and solely in concluding
pādas. My shorthand for such usages will be “the yatas tatas formula.”
These phrases are well known since they were discussed in a 1918–20
article by Sylvain Lévi, but not much seems to have been said about them
since.51 Lévi’s article is about jaya, “victory,” and has the merit of demonstrating

50. Brockington 1998, 146–47 sees the first two usages set in late units, but allows as to “the account of the
actual death of Bhīṣma” that it “may well belong to an appreciably earlier stage of the epic’s growth” (152) than the
later portions of Books 12 and 13.
51. See Matilal 2002, 99, noting that the differentiated phrases imply that “Kṛṣṇa bears the entire respon-
sibility” for the tenuous link between dharma and the outcome of victory, but without the “story-teller” making
him omnipotent.
546 dharma

that these tags tie in with a fan of other formulaic and quasi-formulaic usages:
that jaya is mentioned along with Nārāyaṇa and Nara in the benedictory verse
that prefaces each book of the Mahābhārata; that Nārāyaṇa and Nara (as one is
also intermittently reminded) are ancient identities of Kṛṣṇa and Arjuna as
former Ṛṣis; that Jaya is a name for the Mahābhārata itself (13–15); and, as Lévi
could have added, that Jaya and Vijaya, both meaning “victory,” are names for
Arjuna, so that all the lines that mention jaya can suggest Arjuna’s companion-
ship with Kṛṣṇa and his fidelity to Kṣatriyadharma. But I believe that Lévi’s
emphasis on jaya also led him to misinterpret the pādas that relate Kṛṣṇa and
dharma. According to Lévi, the maxim “Where Kṛṣṇa is there is dharma”

seems to proclaim a lesson of transcendent morality. Right gives the

victory. But it would be reading the Mahābhārata false. . . . It is, like
all the creations of the Hindu genius, a work of a caste and a sect . . .;
it preaches to the kṣatriyas the cult of Kṛṣṇa as a guarantee of success
and welfare. For the kṣatriya success means victory, jaya; and the
safety (and salvation) of the kṣatriya is none else but Kṛṣṇa the God
of the kṣatriyas . . . ; where there is Kṛṣṇa there is law (dharma), the
law proper to the kṣatriya. . . . The Mahābhārata in its entirety is the
amplification of these principles, which come to a focus and stand
out very clearly in the Bhagavadgītā. This incomparable dialogue,
often considered a sublime hors d’oeuvre, is quite, on the contrary,
the very heart and kernel of the work. (Lévi 1918–20, 15–16)

I believe Lévi is right to speak of the BhG’s centrality in relation to Kṣatriyadharma,

but I would demur on the Mahābhārata’s deriving from some kind of Kṣatriya
“sect,” and on its being an amplification of Kṣatriya principles (it is a palate of
Brahmanical principles for Kṣatriyas). Lévi is following a fashion that took Jaya
to imply an original Mahābhārata. In any case, one needs to be clearer that this
centrality is one of design. These lines about Kṛṣṇa, dharma, and victory must
be read in their epic contexts rather than abstracted into principles. The first
thing we have noticed is that a significant cluster of these citations (6.21.11–12
and 14, 6.41.55) occur in segments of the Bhagavadgītāparvan that frame the
BhG. Indeed, two of the only four instances that mention Kṛṣṇa along with
dharma occur right after the Gıtā ̄ ’s dharma revelations. The Gıtā ̄ ’s ring struc-
ture of dharma thus has rings of dharma around it. Equally interesting: all but
one52 of these passages are addressed to one or another of the epic’s three most

52. I leave aside this one passage: Karṇa’s prophetic dream, described to Kṛṣṇa in their secret meeting in
Book 5, which is about Yudhiṣṭhira but does not reach his ears. Having dreamed of Yudhiṣṭhira victorious on a
mountain of bones and Bhīma and Arjuna in stances of victory, Karṇa says, “I know, Hṛśīkeśa, where dharma is
there is victory” (5.141.33cd).
 dharma and the bhagavad gı̄ t ā 547

implicated kings: Dhṛtarāṣṭra, Duryodhana, and Yudhiṣṭhira. They are not

addressed to Arjuna-Jaya, although one of them is addressed by him to
Yudhiṣṭhira. Given these considerations, we can expect to find that in building
up to the BhG and the threshold of the Mahābhārata war, these tags would be
telling these kings things about Kṛṣṇa’s relation to dharma and victory that dif-
fer from what Kṛṣṇa has to say about these matters to Arjuna. Let us look first
at what Dhṛtarāṣṭra and Duryodhana hear, and then what Yudhiṣṭhira hears.
Of all these figures, Dhṛtarāsṭ̣ ra is by far the very f irst to catch wind of a yatas
tatas formula when, in Book 1, his half-brother Vidura—who incarnates the god
Dharma—tells him, soon after the Pānd ̣ ạ vas have married Draupadī and emerged
from hiding as Brahmins, that the Kauravas should give them half the kingdom.
Vidura speaks as if good relations with the Pānd ̣ ạ vas will make their allies also
allies of the Kauravas, and says to Dhṛtarāsṭ̣ ra, “The Dāśārhas [Kṛsn ̣ ạ ’s people] are
mighty and numerous, king; where Kṛsn ̣ ạ is there they would be; where Kṛsn ̣ ạ is
there is victory (yataḥ kṛsn ̣ ạ s tatas te syur yataḥ kṛsn ̣ ạ s tato jayaḥ)” (1.197.25). As
noted, none of these passages before the Gıtā ̄ mentions Kṛsn ̣ ạ ’s connection with
dharma, but the intelligent reader (if not yet Dhṛtarāsṭ̣ ra) can piece that in as
background, since shortly before this Kṛsn ̣ ạ ’s very first word in the epic is “dharma.”
As we noted in chapter 10 § B, he utters it after he has recognized Arjuna in his
Brahmin guise and endorses the outcome of Draupadī’s svayaṃvara.
We next find Dhṛtarāsṭ̣ ra, during a sleepless vigil as the war looms, telling
Vidura that despite all his good advice, “I cannot abandon my son—where
dharma is there is victory (yato dharmas tato jayaḥ)” (5.39.7cd). In this first prewar
usage of this phrase by one of the Kauravas, the curtailed adage may suggest the
wistful hope that dharma and victory together might suffice without Kṛsn ̣ ạ even
though, if Dhṛtarāsṭ̣ ra remembers what Vidura told him in Book 1, he would at
least know that Kṛsn ̣ ạ and victory go together.53 Then, just after Vyāsa has given
Saṃjaya the divine eye by which he will be able to narrate the war, beginning
with the BhG, Vyāsa utters the same shorthand formula, narrowing any room
for Dhṛtarāsṭ̣ ra to take it as a seed of hope: “O Bharata bull, I will spread the
fame of all these Kurus and Pānd ̣ ạ vas. Do not grieve. This is formerly fated
(diṣṭam etat purā), and you are not to grieve about it. It cannot be checked. Where
dharma is there is victory (yato dharmas tato jayaḥ)” (6.2.13–14).
Once the fighting is underway, Dhṛtarāsṭ̣ ra next hears about such matters
when he asks Saṃjaya why the fifth day of battle has gone badly for his sons.
Here, in a highly devotional unit known as the Viśvopākhyāna, Saṃjaya replies
that the Pānd ̣ ạ vas’ success owes nothing to magic (māyā), incantations (mantras),

53. At 5.66.9d, Saṃjaya tells Dhṛtarāṣṭra, Vyāsa, and Gāndhārī, “Where truth is there is dharma, . . . where
Kṛṣṇa is there is victory” before calling Kṛṣṇa the turner of the wheels of time and space (kālacakra, jagaccakra,
yugacakra; 66.12).
548 dharma

or terrors; they fight righteously (dharmeṇa); “they do not turn away from battle,
are endowed with virtues (dharmopetāh), ̣ very powerful, and are joined to the
highest prosperity: where dharma is there is victory” (śriyā paramayā yuktā yato
dharmas tato jayaḥ) (6.61.14–16). Again, instead of hearing about Kṛsn ̣ ạ and vic-
tory, Dhṛtarāsṭ̣ ra hears about dharma and victory, and is still yet to hear about
̣ ạ and dharma. That is soon remedied, however. In a reply that includes the
Kṛsn
third passage cited at the beginning of this section, Saṃjaya now tells Dhṛtarāsṭ̣ ra
that Duryodhana asked Bhīṣma that same question the night after the fifth day’s
rout, with the result that Bhīṣma’s reply reaches the ears of both royal Kauravas:

You were formerly forbidden, son, by the Veda-conversant Munis,

“Do not go to war with the intelligent Vāsudeva, as also with the
Pāṇḍavas.” From folly you did not understand this. I think you are
a cruel Rākṣasa, and that you are enveloped in darkness (tamas). That
is why you hate Govinda and the Pāṇḍava Dhanaṃjaya, for would
another man hate the godly pair Nara and Nārāyaṇa? Therefore, king,
I tell you, this one is indeed everlasting, unfading, always containing
the entire world, the Ruler, Placer (Dhātṛ), Upholder (Dhara), the
fixed one who holds up (Ödhṛ) the triple world, the Lord, master of
the mobile and immobile, Warrior, Victory (jaya), Victor (jetṛ), the
nature of all, the Lord, O king, made of goodness (sattva), this one is
divested of darkness and passion (rāga). Where Kṛṣṇa is there is
dharma, where dharma is there is victory (yataḥ kṛṣṇas tato dharmo
yato dharmas tato jayaḥ). By this one’s yoga of greatness and also the
yoga of his self,54 the sons of Pāṇḍu are upheld (dhṛtāḥ), king, and
victory (jaya) will be theirs. It is he who always grants the Pāṇḍavas
understanding enjoined to the good, and so too strength always in
battle, and he protects them too from fears. He is the everlasting god,
all made of mystery, blessed (śiva). The one whom you ask me about
is known as Vāsudeva. He is served and worshiped by noteworthy
Brahmins, Kṣatriyas, Vaiśyas, and Śūdras who are ever joined to their
own distinctive jobs (svakarma). At the end of the Dvāpara Yuga and
the beginning of the Kali Yuga, it is he who is sung of with
Saṃkarṣaṇa following the Sātvata rule (sātvataṃ vidhim āsthāya).55

54. Mbh 6.62.35ab: tasya māhātmyayogena yogenātmana eva ca. Kṛṣṇa’s divine yoga is sometimes hard to
translate, and van Buitenen’s “supernal yoga” or “divine wizardry” (for yogam aiśvaram at BhG 9.5b; 1981, 105
and 166 n.3), while recognizing this, is not bad. Cf. his “sovereign yoga” for the same at 11.8d.
55. See González-Reimann 2002, 86, 89–90, on this passage among the “only nine” in the Mahābhārata to
tie events to this yuga juncture, finding all such passages for one reason or another to be “late.” On the phrase “Sātvata
rule,” see Mbh 12.322.19a, 23cd in the Nārāyaṇīya, with which the Viśvopākhyāna is conceptually connected.
 dharma and the bhagavad gı̄ t ā 549

Again and again this Vāsudeva creates the world of mortals and all
the Asuras, the cities girt by the sea, and from yuga to yuga the
human habitation. (6.62.30–40)

Duryodhana, of course, refuses to be impressed.

Finally, Bhīṣma brings such terms up one last time when, at the end of his
long postwar oration, he asks Kṛṣṇa and the others present for leave to part this
world. Amid praising Yudhiṣṭhira for his dedication to noncruelty (ānṛśaṃsya)
and devotion to his guru, and telling Dhṛtarāṣṭra he should protect the Pāṇḍavas,
he tells Dhṛtarāṣṭra that his sons were wicked and should not be grieved, and
that Duryodhana ignored Bhīṣma’s lesson again and again (and presumably
not only in the Viśvopākhyāna):

Formerly I told the slow wicked-minded Duryodhana, “Where Kṛṣṇa

is there is victory, where dharma is there is victory (yataḥ kṛṣṇas tato
dharmo yato dharmas tato jayaḥ). With Vāsudeva as refuge (tīrtha),
son, having soothed the Pāṇḍavas, it is beyond your time for healing.
So I said again and again. But that very slow-witted fool did not heed
my word. Having right here brought about the marring of the earth,
he then died.” (13.153.39–41)

Note how the specificities of the earlier citations now fan out into a larger but
more imprecise impression of the text. Although we cannot pin down when
Bhīṣma kept saying this, Duryodhana heard it from him again and again,
including (it seems) “right here” at Kurukṣetra.
Indeed, we learn that Duryodhana also got the more truncated “where
dharma is there is victory” message repeatedly, and in similar untraceable
circumstances, on at least three further occasions from Droṇa and his mother,
Gāndhārī. As Karṇa tells Kṛsn
̣ ạ in their secret prewar meeting, Droṇa once told
Duryodhana, “The man of the white horses [Arjuna] means as much to me as
Aśvatthāman [Droṇa’s own son]. Why use many words? Where dharma is there
is victory (yato dharmas tato jayaḥ).”56 With Gāndhārī, it is a matter of reminders
of her own utterances to calm her when her sons are no more. When Kṛsn ̣ ạ visits
her right after Duryodhana’s death, he recalls what she told Duryodhana in the
Kaurava assembly during Kṛsn ̣ ạ ’s prewar embassy there in Book 5: “Desirous of
victory, Duryodhana was addressed this harsh word by you: ‘Fool (mūḍha), listen
to my word: Where dharma is there is victory (yato dharmas tato jayaḥ)’” (9.62.58).
Soon after, as she is nearing the battlefield with Dhṛtarāsṭ̣ ra, Vyāsa reads her

56. Mbh 5.146.16. Keep in mind that Droṇa’s complicity with Arjuna extends into the scenes of his killing;
see chapter 9 § E.2.b.
550 dharma

thoughts and makes one of his sudden appearances to deflect her with a reminder
of what she said eighteen days earlier when Duryodhana asked her for her prewar
blessing: “Desirous of victory, he asked you this time and time again, Gāndhārī,
and you told him, ‘Where dharma is there is victory (yato dharmas tato jayaḥ)’’’
̣ ạ , who is with her as she finds Duryodhana’s
(11.13.9). Finally, she tells Kṛsn
corpse and embraces it, that she had resigned herself to his doom: “O Vārṣnẹ ya,
when this war that would annihilate kinsmen stood at hand, this one, this most
excellent of kings, his hands folded in respect, said to me, ‘My mother must wish
me victory in this war between kinsmen.’ Realizing the whole imminent disaster,
I said to him, O tiger among men, ‘Where dharma is there is victory (yato dharmas
tato jayaḥ)’ ” (11.17.5–6). Perhaps, as with Dhṛtarāsṭ̣ ra through his wistfulness,
Duryodhana may have heard something hopeful in his mother’s words, through
his intrepidity and intransigence. It would appear, at least from their last embrace,
that they were said with some affection.
Turning now to Yudhiṣṭhira, the lesson comes in parallel but opposite
terms in that it falls on open ears, is heeded, and indeed remembered. With
this, two things are to be noted. First, Yudhiṣṭhira—unlike Duryodhana but
also unlike Arjuna—needs no reminders. Second, he first learns what he
needs to know about victory, dharma, and Kṛṣṇa in two passages of the
Bhagavadgītāparvan that surround the BhG’s own revelations about dharma,
and then, at the end of the war, tells Kṛṣṇa that he has heard it even before this
from Vyāsa at Upaplavya, presumably somewhere in Book 4 or 5.
As the Bhagavadgītāparvan builds up to “the BhG proper,” Yudhiṣṭhira
tells Arjuna that he is dispirited (viṣaṇṇam) at seeing the Kaurava army vaster
than theirs. In a passage that includes the first three verses cited at the beginning
of this section, Arjuna replies with three insistent assurances that victory comes
with either dharma (6.21.11d) or Kṛṣṇa (6.21.12d and 14d):
Listen, king, how fewer men triumph over many talented and shrewd
warriors. I shall tell you how, and do not gainsay me. The Ṛsị Nārada
̣ ạ va, and so do Bhīṣma and Droṇa.58 It is told that the
knows it,57 Pānd
Grandfather59 once said to great Indra and the other celestials
concerning this very issue at the battle of the Gods and Asuras, “Those
who seek victory win not so much by strength and might as by truth
and noncruelty (ānṛśaṃsya) and by dharma and enterprise. Abandon
adharma, greed, and delusion,60 be enterprising, and then fight without
thought of ego. Where dharma is there is victory (yato dharmas tato

57. Arjuna is implying that Nārada would be his source, as he clarifies further along at verse 12.
58. As will be pertinent to their consent to Yudhiṣṭhira, discussed above in the text.
59. That is, Brahmā, but note that at 6.21.3 Yudhiṣṭhira has just been thinking here about “Grandfather”
Bhīṣma’s siding with the Kauravas, whom Arjuna has just mentioned.
60. Greed (lobha) and delusion (moha) seem to be paired opposite to noncruelty and truth, respectively.
 dharma and the bhagavad gı̄ t ā 551

jayaḥ).” So know, king, that our victory in battle is assured,61 for as

Nārada has told me, “Where Kṛsn ̣ ạ s tato
̣ ạ is there is victory (yataḥ kṛsn
jayaḥ).” Victory is a talent with Kṛsn ̣ ạ ,62 and it follows closely behind the
Mādhava. And no less than victory (vijaya) is modesty a virtue (guṇa)
with him. Govinda of infinite splendor strides unconcernedly among
the multitude of his enemies, the most everlasting Puruṣa. Where
̣ ạ s tato jayaḥ). Once, having become
̣ ạ is there is victory (yataḥ kṛsn
Kṛsn
Hari Vaikuṇṭha of the ever-sharp arrows, he said thunderously to the
Gods and Asuras, “Who shall win?” And they won who said, “With
Kṛsṇ ạ we shall win!” Indeed, it was by his grace that Śakra and the Gods
won the Three Worlds. I see no cause for concern at all, Bhārata, when
he, the ruler of the world and the lord of the Thirty, wishes your victory.
(6.21.7–17, modifying van Buitenen trans. 1981, 65)
Again, as with the Kaurava kings, we await the conclusion of the Gıtā ̄ for Yudhiṣṭhira
to hear the completion of the equation. This occurs in the first adhyāya just after the
BhG proper that closes the Bhagavadgītaparvan. Dismounting his chariot with an
uncanny purposefulness that reminds us, by contrast, with Arjuna’s lack of purpose
in dismounting his chariot at the beginning of the Gıtā ̄ , Yudhiṣṭhira walks across
the divide between the armies to ask his elders (gurus) on the opposing side not
̣ ạ explains, “as if smiling,” brings
only for their consent to fight them (which, as Kṛsn
victory; 16–19) but also how he might kill them. Coming to Droṇa second after
Bhīṣma, he says, “Wish me victory, Brahmin, and counsel me for my good. Fight
for the Kauravas, that is the boon I ask” (6.41.53); to which Droṇa replies in a
passage that includes the second of the citations that opened this section:
Your victory is assured (cf. 6.21.12b), king, for you have Hari as your
counselor (mantrin). I recognize you: you shall defeat your foes in
battle. Where dharma is there is Kṛṣṇa, where Kṛṣṇa is there is
victory (yato dharmas tataḥ kṛṣṇo yataḥ kṛṣṇas tato jayaḥ).
Go and fight, Kaunteya. What can I tell you? Ask me. (6.41.54–55)

As we are seeing, the yatas tatas revelations reach not only the Pāṇḍavas but the
Kauravas, and other ears as well as they fan out into the text. As we shall
maintain in the next section, these ripples ring out from BhG itself.
Next, then, on the fourteenth day of battle, Vyāsa stops Yudhiṣṭhira from
setting off after Karṇa, for whom Yudhiṣṭhira is no match and in any case
is Arjuna’s sworn foe. In a passage treated prominently in chapter 9, Vyāsa
promises Yudhiṣṭhira that in five days the earth will be his and tells him how
to keep himself healthy: “Thinking always of dharma, O tiger among men, may

61. 6.21.12ab: dhruvo ‘smākaṃ raṇe jayaḥ (cf. 6.41.54a).

62. Following van Buitenen for guṇabhūto jayaḥ kṛṣṇe (6.21.13a).
552 dharma

you be pleased to practice noncruelty (ānṛśaṃsya), penance, gift, forbearance,

and truth. Where dharma is there is victory (yato dharmas tato jayaḥ)”
(7.158.61–62). Here we have the author intervene to indicate that, for dharma
to triumph, his protege Yudhiṣṭhira Dharmarāja had better stay alive, and that
the conjunction of dharma and victory needs a little more time to ripen.
Finally, when the war is over and just before Kṛṣṇa says it would not be
good to spend the night with the surviving Pāṇḍava troops, who will be massa-
cred that night in their sleep, Yudhiṣṭhira asks Kṛṣṇa what he is to make of the
sudden combustion of Arjuna’s chariot outside the tent of the defeated Kauravas
just after Kṛṣṇa has ordered Arjuna to dismount from it for the last time. Kṛṣṇa
embraces Yudhiṣṭhira and says:

By good luck (diṣṭyā) you conquer, Kaunteya. By good luck your

enemies are conquered. By good luck the wielder of the Gāṇḍīva
bow [Arjuna] and Bhīmasena Pāṇḍava have escaped from this
hero-destroying war, their foes struck down. Quickly do the activities,
Bhārata, for the time at hand. When I arrived at Upaplavya, having
brought me the honey mixture [the madhuparka that is offered to a
guest] with the Gāṇḍīva bowman, you formerly told me, “Kṛṣṇa, this
Dhanaṃjaya [Arjuna] is your brother and friend (bhrātā sakhā caiva
tava). He is to be protected in every distress (sarvāsv-āpatsu), great
armed lord.” When you said this, I said, “Yes.” This Savyasācin was
guarded. Victory is yours, lord of men. With his brothers, king of
kings, that champion (śūra) of true prowess has escaped from this
hair-raising hero-destroying war. Thus addressed by Kṛsn ̣ ̣a, O king,
Dharmarāja Yudhiṣṭhira, his hair bristling, answered Janārdana,
“Foe-tamer, when the Brahma-weapon was released by Droṇa and
Karṇa, who else than you could prevail in person, even including the
thunder-wielding Puraṃdara [Indra]? By your grace the many were
conquered in battle, and Pārtha entering the great battle was not one
who turned his back. So too, great-armed one, I have obtained an
auspicious way of splendor, following diverse activities with many
proceedings.63 At Upaplavya,64 the great Ṛṣi Kṛṣṇa-Dvaipāyana told
me, ‘Where dharma is there is Kṛṣṇa, where Kṛṣṇa is there is victory
(yato dharmas tataḥ kṛṣṇo yathā kṛṣṇas tato jayaḥ).’” (9.61.21–30)

63. Mbh 9.61.29: tathaiva ca mahābāho paryāyair bahubhir mayā/karmaṇām anusaṃtānaṃ tejasaś ca gatiḥ
śubhā. My translation of this verse is uncertain.
64. Upaplavye; or possibly, “During the disaster,” since upaplavya could have that meaning (see Matilal’s
translation of this verse [2002, 99]). But since the passage is already talking about Upaplavya as a place, this
seems unlikely.
 dharma and the bhagavad gı̄ t ā 553

We see how well constructed this set of passages is in continuing to recall

Yudhiṣṭhira’s initial anxiety over the greater numbers on the Kaurava side, and
in introducing matters of friendship and hospitality in conjunction with these
formulas about dharma and Kṛsn ̣ ạ . As with the concluding passage in the
sequence concerning Dhṛtarāsṭ̣ ra and Duryodhana, the last one in this sequence
also opens wider circles of information that leave questions about the main nar-
rative. The description of Kṛsṇ ạ ’s welcoming at Upaplavya, the Pānd ̣ ạ vas’ prewar
camp at the end of Book 4 and through most of Book 5, is not drawn from
descriptions of his arrival or return there. Nor can we locate the time at Upaplavya
when Vyāsa would have told Yudhiṣṭhira that dharma, Kṛsn ̣ ạ , and victory work
together since there is no point in this stretch where Vyāsa is said to have visited
the Pānḍ ạ vas there at all.
Indeed, several of these passages make it a point to attribute the disclo-
sures of this type to one or more of the great Ṛṣis: Nārada, as we have seen, told
it to Arjuna. Bhīṣma65 and Droṇa knew it at 6.21.12. And Rāma Jāmadagnya,
Mārkaṇḍeya, Vyāsa, and Nārada are all sources for what Bhīṣma tells
Duryodhana in the Viśvopākhyāna (at 6.62.27). We shall be discussing these
networks of Ṛṣis further in chapter 12. There are also suggestive points
regarding Yudhiṣṭhira and Arjuna. If Yudhiṣṭhira got this message from Vyāsa
at Upaplavya, he would have heard it before the Bhagavadgītaparvan. And if
Arjuna downloads it to Yudhiṣṭhira at the near side of the Bhagavadgītaparvan,
he would have to have heard it before the BhG.

E. Dharma Rings in the Bhagavad Gıtā

̄ Proper

As we have seen earlier in this chapter, the Mahābhārata introduces a number

of new conceptualizations of dharma, among them the BhG’s unique theological
grounding of svadharma and karmayoga that would not have appealed to Manu.
The BhG puts the concepts of svadharma and karmayoga to work as teachings
by which an incarnate deity convinces a consummate but irresolute warrior,
who is not a king, to fight. Though Arjuna hears several times from Kṛṣṇa
about winning the kingdom, it would not be his to rule.66 Indeed, before we go
further, it is important to mention that Kṛṣṇa is not a king either.67

65. Bhīṣma has a pipeline to the great Ṛsị s; see Hiltebeitel 2001b on his celestial sources as the son of Gaṅgā.
66. The point sets me in opposition to some who have argued otherwise, particularly with regard to the
BhG: see Biardeau 1981a, 93–94, offering a brilliant depiction of Arjuna’s svadharma as the “ideal king” that,
however, sidesteps Yudhiṣṭhira to impose the BhG’s svadharma on the king’s dharma (cf. Biardeau 1997, 90;
2002, 1: 155–58); Malinar 2007a, 11 calls Arjuna “a potential king” to consider him as a king subordinate to Kṛṣṇa
as an īśvara, implying cosmic kingship in a “cosmological monotheism” (4–13, 146–48, 233–41).
67. This is a complicating matter with many resonances, beginning with the curse of Yayāti that whereas
the descendants of his youngest son Pūru—that is, the Paurava line that will include the Kauravas and
554 dharma

Now although Kṛṣṇa makes the BhG the occasion to range over a number
of soteriological, philosophical, and practical options that have interested com-
mentarial and scholarly discussion more than what he says about dharma, no
one would deny he also deserves a reputation in the BhG for having things to
say about dharma. Yet if one puts a few mostly minor passages aside,68 he really
says only a few things about dharma per se. Most of his prominent references
to dharma occur in what I will be calling an informal ring structure: not a formal
ring of the type appreciated by folklorists where a text exhibits a self-conscious
geometry of units and themes converging on a central nugget,69 but one that
allows the BhG to be also about other things to which dharma is kept pertinent
through deepening reminders of its relevance. There may also be a further ring
that is off-center or deferred. We shall examine what these patterns tell us about
dharma as these rings close in on the text’s important teachings.
Let us peel inward starting from the outer layers. In reencountering three
familiar terms, I will continue to leave svadharma to speak for itself and trans-
late svabhāva and svakarma as “inherent nature” and “own jobs,” respectively.

Ring 1:

1. In the BhG’s very first words, Dhṛtarāṣṭra asks Saṃjaya what happened
“on the field of dharma (dharmakṣetre), the Kuru field” (BhG 1.1),
between “my sons,” the Kauravas, and their foes.
̄ ends, Saṃjaya tells Dhṛtarāṣṭra that the last thing Kṛṣṇa
2. As the Gıtā
told Arjuna, before asking him if he understood, was that theirs was
a “righteous dialogue”—that is, it was dharmya, about dharma—and that
̣ ạ , and
whoever learns it will offer it up as a sacrifice of knowledge to Kṛsn
whoever listens to it will be released to blessed worlds (18.70–72).

̣ ạ –Arjuna dialogue
To grasp this first ring one must know that the BhG’s Kṛsn
is framed by a dialogue between Dhṛtarāsṭ̣ ra, the blind father of the doomed
Kauravas, and the bard Saṃjaya. Thanks to a recent gift of the “divine eye” from

Pāṇḍavas—will inherit his kingship, the descendants of his oldest son Yadu will be “without a share of king-
ship” (Mbh 1.79.7; see Defourny 1976, 134–35). Kṛṣṇa and his Yādava people can thus be called lowly cowherds
in Śiśupāla’s tirade against Kṛṣṇa in Book 2.
68. BhG 9.21, a reference to the trayīdharma, “the Law of the three (Vedas),” as applying to temporary des-
tinations in the triple world obtained by desires; 9.31, the higher reward for the bhakta who is dharmātmā, “of
righteous self”; 18.34, a reference to dharma along with kāma and artha. Not minor, however, is 7.11 on Kṛṣṇa’s
being “the desire in beings that does not run counter to dharma,” as mentioned in § C.
69. See Tubb 2002; Brodbeck 2006. Cf. D. Hudson 1994, 2001, more pertinent to the approach here and
discussed above in the text.
 dharma and the bhagavad gı̄ t ā 555

the epic’s putative “author” Vyāsa, Saṃjaya can report the entire war account.
The BhG will end with Saṃjaya telling Dhṛtarāsṭ̣ ra that its supreme secret comes
to them by “Vyāsa’s grace” (18.75).
This framing dialogue makes dharma the BhG’s very first word, while the
last mention of dharma, by Kṛṣṇa, confirms the “righteous” nature of the whole
exchange. Arjuna then replies that he will stand firm and do as Kṛṣṇa bids with
his confusion gone and his memory restored (18.73), after which Saṃjaya again
speaks from the frame to wrap things up (74–78). These opening and closing
usages are more subtle than they look. Not only is Arjuna’s “memory” restored.
The ending recalls that when Saṃjaya received the divine eye from Vyāsa (Mbh
6.2.9–13; 16.5–10), it was not only so that the blind old king could hear what he
would otherwise be missing, but the same for all audiences of the text. Note how
Dhṛtarāsṭ̣ ra responded to this transaction with the following appeal to Vyāsa, his
own father, whom he recognizes as having the real “control” over what will now
go into the “textualized memory” (Olivelle 2005b, 168, for smṛti) of the war:
“I beseech you. You are of immeasurable power. You show the way and are firm.
They [i.e., my Kaurava sons] are not even under my control, Maharṣi. You can
enable me not to commit sin here. Surely you are dharma, the purifier, fame,
glory, bearing, and memory (smṛtiḥ)” (6.4.12–13; see Hiltebeitel 2001a, 55–59).
Arjuna’s memory is fused into Dhṛtarāsṭ̣ ra’s, Dhṛtarāsṭ̣ ra’s into ours.
The BhG’s opening has further reverberations. The reference to a “field of
dharma” recalls that the rules of fair fighting were agreed upon earlier that day,
when both sides “established the dharmas (laws, rules of engagement) of
battle.”70 Most of these rules will be broken on this very field, and not infre-
quently at the advice of Kṛṣṇa. That this dharma-field is “the field of Kuru” also
recalls an ancestor of the Kuru line, Kuru, who “made Kurukṣetra meritorious
(puṇyam) by his austerities” (1.89.44). During the war narrative, we learn that
King Kuru did such austerities at Kurukṣetra that he got the boon from Indra
that ascetics and warriors who died there in battle would both go straight to
heaven cleansed of their wicked acts by its very dust (9.52.13, 18). These uses
converge when Bhīṣma is about to give his lengthy dharma oration after the
war. As Kṛṣṇa and the Pāṇḍavas come to the spot where Bhīṣma lies on his
hero’s bed of arrows, Kurukṣetra is called “the field of the whole of the Law”
(12.53.23), expanding on the f irst words of the BhG and projecting them toward
some kind of completion in the dharma instructions of Bhīṣma, which the
war’s survivors are arriving to hear before he dies and goes to heaven. Further,

70. Mbh 6.1.26–33, beginning, “Then the Kurus, Pāṇḍavas, and Somakas made a covenant (samayaṃ
cakruḥ), and they established the dharmas (rules or regulations) of battle (dharmāṃś ca sthāpayām āsur yuddhānām
bharatarṣabha).” See Matilal 2002, 94–95; Mehendale 1995.
556 dharma

the opening “field” references resonate with what Kṛṣṇa has to say within the
̄ about the “field-knower.”71 In the BhG’s philosophical context, the “field”
Gıtā
of “Nature” is not only the body and mind but the world of cosmic evolutes
made up of three ever-entangling “Qualities” (guṇas) known as Goodness (sat-
tva), Passion (rajas), and Darkness (tamas), which the “knowing” or “witness-
ing” Self can level from its transcendent standpoint. This brings out another
sense of the “field of dharma” as a field where seeds of merit (dharma) and
demerit (adharma) are sown and their fruits reaped.
The BhG’s usage of “field-knower” can be illumined by the Uttara-Yāyāta,
a subtale found earlier in the epic (1.81–88), where Yayāti, another ancestor of
the Kauravas and Pāṇḍavas more ancient in the lineage than Kuru, serves as a
kind of Upaniṣadic guide to the afterworld during an interval after he has been
bounced out of heaven for a prideful utterance and is headed for hell (see van
Buitenen 1967–68; Dumézil 1973, 28–37). Rather exceptionally, he is able to
steer himself along the path of his fall, and heads for some smoke he sees
rising from the hallowed Naimiṣa Forest, the Twinkling, Blinking, or Winking
Forest—a place where epic tales can take strange turns since it marks a point
of intersection between heavenly, earthly, and infernal realms.72 Yayāti is aware
that where there’s smoke like this, there’s a Vedic sacrifice, and thus people
who are “good.” The smoke turns out to be rising from a Vājapeya sacrifice
being performed by his four grandsons, the first of whom, Aṣṭaka, is intro-
duced as “a protector of the rules of the true dharma (saddharmavidhāna-
goptā)” (1.83.6d). When Yayāti tells the four who he is and they see him hanging
in midair over that hallowed terrain, impressed that such an interloper could be
an expert on the “laws” of the “field”73 of retribution, they each address him in
turn using the same quasi-formulaic half line: “You are I think a knower of the
field of this dharma (kṣetrajñaṃ tvāṃ tasya dharmasya manye)” (1.87.8d; 1.87.13d;
1.88.1d; 1.88.6d). Their usage suggests that “this dharma” would have the sense
of a “doctrine,” “law,” or “teaching” that the fallen one might impart to them,
much as with Yama’s dharma in the Kaṭha Upaniṣad (see chapter 4 § F), and, of
course, like the Buddha’s dharma. The story uses the phrase “knowledge of the
field . . . of dharma” with reference to its “doctrine” of retribution, which turns

71. In Matilal’s terms, kṣetrajña brings out another meaning of the BhG’s opening words: “Dharmakṣetra—
the field where the seeds of moral merit/demerit are sown in order to bring forth the harvest of karma or just
desert” (2002, 93).
72. I thank Danielle Feller (personal communication, August 2008) for bringing out this point with refer-
ence to Sītā’s descent from there into the underworld at Rām 7.82.13. Cf. Hiltebeitel 1999a, 266, 285–93; 2001a,
124, 319–20.
73. “As one who possesses knowledge of the field you surely speak the laws” (kṣetrajñavad bhāṣase tvaṃ hi
dharmān)” (1.84,12d).
 dharma and the bhagavad gı̄ t ā 557

out to concern a notion of “merit” (puṇya) by which Yayāti’s grandsons can

transfer their own merits to him to reverse his downward course and enable all
five of them to ascend to heaven together. From this we might extrapolate that
where Kṛṣṇa concludes the dialogue with Arjuna saying that it “has to do with
dharma,” his portion of that dialogue could also be taken as his “teaching.”
There is one place in the BhG where he uses dharma in that way. As Kṛṣṇa pre-
pares Arjuna to see his cosmic divine form, he bills the revelation as a “royal
wisdom, royal mystery, and ultimate purifier”74 that is “lawful” or “about
dharma” (dharmya), and says, “Men who lack faith in this dharma, enemy-
burner, having failed to reach me, they return to the runaround of deaths”
(9.2–3). But for the most part, Kṛṣṇa’s teaching is presented not as a single
dharma but as a variety of “disciplines” or yogas.
When Kṛṣṇa says at the far side of Ring 1 that his dialogue with Arjuna has
been “righteous” or “about dharma,” this usage holds some further interests.
Saṃjaya also qualifies “this dialogue” twice as “wondrous” and once as
“hair-raising” or “enrapturing” (18.74, 76). And Kṛṣṇa speaks of it indirectly as
dharmya toward the Gıtā̄ ’s middle:

But those who revere this righteous (dharmya) elixir as it has been
uttered, having faith (in it?) and intent on me, these devotees are
exceedingly dear to me. (12.20)

Moreover, in a Mahābhārata that is not only full of dialogues (saṃvādas) but set
in and sustained by dialogical frame stories, “this righteous dialogue between
the two of them (imaṃ dharmyaṃ saṃvādam āvayoḥ)” (BhG 18.70ab)—that is,
between Kṛṣṇa and Arjuna—presents the only instance where a saṃvāda is
described as dharmya.
To summarize, Ring 1 first mentions dharma in a way that resonates with
geographical, political, genealogical, soteriological, ethical, and philosophical
ideas, and closes with a confirmation that dharma has been what the whole
Kṛṣṇa–Arjuna dialogue has been about. When the BhG mentions the Kuru
Field as a Dharma Field in its very first words and closes on its being “about
dharma,” it encircles Kṛṣṇa’s argument for “just (dharmya) war” (see Ring 3),
placing its dialogue in an especially politicized “field of merit” that gives war-
riors slain there a ticket to heaven and probably still evokes what is left of the

74. As McComas Taylor pointed out at the fifth meeting of the Dubrovnik International Conference on the
Sanskrit Epics and Purāṇas (August 2008), most translators and commentators take rājavidyā as a karmadhāraya
compound, thus “royal wisdom” or “royal science,” and not as a tatpuruṣa compound, by which Malinar takes it
to mean “the knowledge of kings” (2007a, 12; 144–50, 225, 232; cf. 180). See further Adluri 2010d, 103–5 and
above n. 66 on the matter of Arjuna’s not being a king.
558 dharma

prestige of the late Ṛgvedic “Kuru state.” The compound Kuru dharma, “law
of the Kurus,” has six usages in the Mahābhārata, the most interesting being
where the Kauravas are said to have (b)reached the “limit of the Kuru dharma
(kurudharmavelām)” (2.60.33) during the disrobing of Draupadī.75 Manu also
hallows the place of “the Kurus” as a holy site (8.91) that would lie within his
heartland of dharma (see chapter 5). And surprisingly, even Buddhist texts
acknowledge the region as one where a Kurudhamma prepares its public for
special teachings that the Buddha is alleged to have imparted there (see
chapter 4 § B.1.d.i).

Ring 2:

1. The BhG’s principal dialogue first makes dharma a focus when Arjuna
relates his despondency over fighting kinsmen to his angst about
“clan” or “family dharma,” “the dharma of social class by birth,” and
“class-mixture” (kuladharma, jātidharma, and varṇasaṃkara), and his
fear of hell (1.40–44).
2. Kṛṣṇa brings these matters to resolution near the end when he tells
Arjuna he should abandon all dharmas since Kṛṣṇa will release him
from every sin (18.66).76

Arjuna explains his famous despondency, which brought him to lay down
his weapons, as arising from the “the taint caused by destruction of the clan” or
“family” (kula; 1.38–39). But the heart of it is a horror of class-mixture, really
miscegenation. His argument is that if he kills his relatives, adharma will result
for his clan and social class because the women of the clan will become corrupt
and engage in mixed unions. If that occurs, the clan’s ancestors will go to hell
because no one will perform their rites, and he himself will go to hell for bring-
ing all this about (40–44). There have been attempts to widen the scope of
“Arjuna’s dilemma.” One is to interpret it as an aversion, however fleeting, to
war and killing. But the word for what makes Arjuna despondent, kṛpa, by
which he is twice said to be “possessed” (1.28; 2.1), means “pity” before “com-
passion,” and is the first thing Arjuna asks for clarity on, admitting that he is
“afflicted by the taint” of it (2.7). A paragon warrior, Arjuna is pitying those he
has been so well trained to kill—especially, we may guess, for his guru Droṇa

75. Draupadī reincarnates as Belā in the Hindi oral epic, Ālhā, which equals Velā, “the Limit,” in the
Bhaviṣya Purāṇa; see Hiltebeitel 1999a, 125–37 and passim. See also 1.97.11c; 3.131.12d; 8.23.46c and 14.77.37c
(kurudharmajña), and 12.192.96c (kurudharmam adharmaṃ vā).
76. Note, this all parallels Yudhiṣṭhira’s fear of falling into hell, but the larger problems (social class for
Arjuna, life-stages for Yudhiṣṭhira) and the solutions differ.
 dharma and the bhagavad gı̄ t ā 559

who did that training of Arjuna, as we saw in chapter 9. Another is to consider

his dilemma as a conflict of duties within the framework of the Law of Class
and Life-stage. Along with being a Kṣatriya, Arjuna is married, and has been
imagined to be pinioning his domestic duties over against an inclination toward
renunciation and nonviolence. Cited in this regard is a verse where he says it
would be “better to enjoy almsfood” than “enjoyments smeared with blood”
(2.5). But Arjuna remains focussed entirely on the clan and class issues of
killing his elders and does not, as Yudhiṣṭhira does after the war, consider the
beggar option a real one.77 Indeed, the BhG never mentions “life-stage” consid-
erations at all.
We may thus say that when Kṛṣṇa tells Arjuna at the far side of Ring 2 that
he should abandon all dharmas, which may, as Śankara has maintained, refer
as well to adharmas (Sharma 1986, 89), he would be referring above all to
those that have reduced Arjuna to this temporary inaction and pity. Though
the verse uses the root Ömuc, “to release,” when Kṛṣṇa says “I will free you
from all sins,” he is not talking about mokṣa as final liberation. As Ring 3 now
makes evident, for Kṛṣṇa to release Arjuna from every sin will still hold him to
doing his Kṣatriya duties, but in a new spirit of abandoning desire for fruits or
results.

Ring 3:

1. Kṛṣṇa gets Arjuna to concentrate his multiple dharma anxieties on just

one matter: Kṣatriya svadharma. When Arjuna says his inherent nature
(svabhāva) is afflicted by the taint of what comes from pity and asks
Kṛṣṇa to relieve his confusion about dharma (2.7), Kṛṣṇa answers that
a Kṣatriya can find nothing better than a “lawful” or “just war” in
which “either you are killed and go to heaven” or “win and enjoy the
earth” (2.31–37). Kṛṣṇa then rounds off the point with two adages: first,
that a little dharma goes a long way (2.40); and second, that doing
“one’s svadharma” is better than doing another’s (3.55) because “the
best” is the man who acts disinterestedly for the “holding together of
the world” (lokasaṃgraha), just as Kṛṣṇa does himself (3.19–26).
2. This round of topics is brought to resolution, still near the end, when
Kṛṣṇa explains svadharma in terms of each of the four social classes’
“own jobs” and inherent natures (18.41–47).

77. Contrary to Matilal 2002, 74 (see 43–44, 73–74), Arjuna is not thinking of becoming “a sanyāsin or a
wandering monk rather than a king” (sic) at BhG 2.5, where he mentions his preference for “eating almsfood”
(bhaikṣyam). Although Arjuna’s dilemma goes deeper, in this verse he is only making a spur-of-the-moment
judgment; and he is not a king. Indeed, Matilal himself recognizes that the beggar option is not real (2002, 8).
560 dharma

Ring 3 concerns the duties specific to a warrior. That is what Arjuna truly
is, and not a king. Modeled on the warrior, each social class will work for the
“holding together of the world” by doing its svadharma in the “own jobs” that
are inherent to them.
The correspondences between the near and far sides of ring 3 are the clear-
est we encounter, as is the movement toward closure. The central passage on
the near side reads:

Look to your svadharma and do not waver, for a Kṣatriya can find
nothing better than a lawful war. It is an open door to heaven, happily
happened upon; and blessed are the warriors, Pārtha, who find a war
like that! Or suppose you will not engage in this lawful war: then you
give up your svadharma and honor and incur sin. (2.31–33)

The idea of “lawful” or “just” (dharmya) war is challenging. It cannot refer to the
compact both sides make to fight fairly, which Kṛsn ̣ ạ himself will ignore. But it
would have behind it the justice of the Pānd ̣ ạ vas’ cause, and, coming from Kṛsn ̣ ạ ,
the fact that he was the last to make efforts to negotiate a peaceful settlement. In
legally representing the Pānd ̣ ạ va cause, however, Kṛsn ̣ ạ was speaking primarily
for Yudhiṣṭhira, the Dharma King. Here he is speaking to Arjuna, for whom he
has reduced what is “lawful” entirely to Kṣatriya dharma—indeed, to Kṣatriya
svadharma, which promises heaven for those who die in battle. Although Manu
also says that slain warriors go to heaven (7.89), it is worth mentioning the
Buddha’s nonconcurrence on this point. When pressed by martial types of “head-
men” (gāmaṇi) who are clearly dubious about such guarantees, he revealed with
great reluctance that a soldier who dies in battle does not go to heaven but to the
“Battle-Slain Hell,” since he dies with “his mind already low, depraved” and
“misdirected” toward killing others (SN, Gāmaṇisaṃyutta 3–5).78 Yet Buddhism
influenced another usage that could point to a deeper sense. This is the idea
broached by Aśoka after the terrible Kalinga war, that henceforth he “considers
conquest by dhaṃma the most important conquest” (see chapter 2). The real
“just war” would be the one fought within. Gandhi brought out this interpreta-
tion of Arjuna’s true battle in the Bhagavad Gıtā ̄ , and with it the idea that
svadharma means something like “conscience.”79

78. See Schmithausen 1999, 48 citing three “almost identical sermons” in the Saṃyutta Nikāya [SN
4.308–11]; Bodhi 2000, 1334–36 translates only the first one. Though the Buddha is reluctant to say this because he
expects the soldier to take it poorly, the soldier replies gratefully upon realizing that he has “been tricked, cheated,
and deceived for a long time by those mercenaries of old” (Bodhi 2000, 1335). See Sinha 1991, 374–82 more widely
on the Nikāyas’ “scathing criticism” of kshattavijja (“kṣatriya science”)—a term perhaps that reduces it to something
more ordinary than dharma.
79. I have long believed I owe this attribution to Rudolph and Rudolph 1967, but I cannot now locate the
term in question there.
 dharma and the bhagavad gı̄ t ā 561

̣ ạ is not encouraging of such an interpretation.

On the far side of Ring 3, Kṛsn
Backed up by his final words on the three Qualities of Matter, he returns to the topic
of svadharma after explaining the proper functioning of the four social classes:

The jobs of Brahmins, Kṣatriyas, Vaiśyas, and Śūdras, enemy-burner,

are distinguished according to the Qualities that spring from their
inherent nature. Tranquility, restraint, austerity, purity, patience,
uprightness, knowledge, discernment, and orthodoxy are the
Brahmin’s job born from his inherent nature. Championing, energy,
bearing, skill, not fleeing in battle, the gift, and lordly nature are the
Kṣatriya’s job born from his inherent nature. Agriculture, herding,
and trade are the Vaiśya’s job, born from his inherent nature, while
the inherent nature born to the Śūdra has the character of service.
Contented each in his own job, a man attains complete fulfillment.
Engaged in his own job, hear how he finds that perfection. A man
finds perfection by his own job having worshiped him by whom all
this is strung, whence beings are motivated to activity. Better one’s
svadharma imperfectly performed than another’s dharma done
perfectly; doing the job regulated by his inherent nature, he does not
incur fault. (18.40–47)

However, one translates karman here, which I have rendered as “jobs,” one
should not obscure the distinction between the last “job” reference and the one
closing instance of svadharma, as some have done by fudging karman there as
“duties.” Kṛṣṇa is fine-tuning a well-known dharmaśāstra job scheme.
Clearly, there has been movement here from one side of Ring 3 to the
other. While the last verse, beginning “Better one’s svadharma,” has the same
famous first line as a verse on the near side of this ring, their second lines dif-
fer. In the earlier verse, the second line simply reinforces Kṣatriya svadharma:
“Better death in one’s svadharma; another’s dharma brings danger” (3.35). In
the later verse, the second line refers svadharma back to the two terms that
govern the passage, karma and svabhāva: “doing the job regulated by his
inherent nature, he does not incur fault.” No longer needing to convince Arjuna
to do his Kṣatriya svadharma, Kṛṣṇa now uses the Kṣatriya as the role model to
talk about the jobs of all four social classes. More than that, he ontologizes each
job in its respective inherent nature. Here Kṛṣṇa finally straightens out the
issue of “class-mixture” that defined adharma for Arjuna (1.38–44) and para-
lyzed him down to his inherent nature (2.7).
On the near side of Ring 3, when Kṛṣṇa tells Arjuna how “the best” act dis-
interestedly for the “holding together of the world” (3.20, 25), he says that if he
himself did not do this “untiringly, at all times, people all around would follow
562 dharma

my lead. These people would collapse if I did not act: I would be the author of
mixing” (3.23–24)—which certainly includes “mixing of classes.” Now at the
far side of Ring 3, men of each social class can find perfection through content-
ment in their own jobs, having worshiped Kṛṣṇa as the lord who motivates all
beings who are “strung like strands of pearls upon a string” (7.7), and a string
that is “unmanifest” (9.4). Every “inherent nature” comes from “the over-soul,”
meaning Kṛṣṇa, in the first place (3.30, 7.29, 8.3).
Warrior svadharma thus gives a certain patina to everyone else’s “own job.”
It is this kind of scheme that Kṛṣṇa is reformulating at the far side of Ring 3:
one whose basics are found in all our classical Brahmanical dharma texts. This
returns us to the question, “Who really has svadharma?” Despite what has been
written on the premise that svadharma provides a kind of cosmic moral matrix
for every individual human and other being, not to mention every individual
action, we have seen that there is not much evidence that the concept has such
a global reach. On the contrary, it is the svakarma of the Brahmin that provides
the paradigm that models the activities of other classes on prerogatives
grounded in sacrificial ritual. Kṣatriya svadharma, on the other hand, is a role
model for all classes to fulfill duties that uphold the Brahmanical order.80

Ring 4:

1. Kṛṣṇa claims that he himself is the restorer of dharma (4.7–8).

2. Arjuna, during his exhilarated description of Kṛṣṇa’s revelation of his
divine form, accepts this, seeing that Kṛṣṇa is “the unchanging
protector of the everlasting dharma” (śāśvatadharmagoptā; 11.18).

So far, most of our dharma citations have come from the BhG’s edges: on
the near side, from chapters 1 to 3, and on the far side all from chapter 18.
̄ ’s center, things get more dif-
When it comes to Kṛṣṇa’s revelations at the Gıtā
fuse. But it is simple enough to appreciate that the two passages just cited
define a fourth ring. On one side, we have come to the famous passage where
Kṛṣṇa reveals how he provides divine intervention whenever dharma is nega-
tively affected in the course of time:

Whenever there is a waning of dharma and a surge of adharma,

O Bhārata, then I create myself; for the complete rescue of the good
and for the destruction of the wicked, for the sake of the establishment
of dharma I come into being from yuga to yuga. (4.7–8)

80. I use role model in the sense of Robert Merton, the term’s coiner, for whom it meant someone who
sets a model for a narrowly defined behavior. See Hiltebeitel 2004b.
 dharma and the bhagavad gı̄ t ā 563

Here we have an idea with a long future, for the theme of Viṣṇu as Preserver of
dharma comes to be associated with the avatāra doctrine of “incarnation” or,
preferably, divine “descent” (see chapters 6 and 12). Even though the prevailing
intention of Western “higher criticism” has been to explain away as late “devel-
opments” every passage in both epics that could relate to this doctrine, there
are much better reasons to think that it is under construction in them from
their very conception—and I do not think that these verses would be an
exception. Meanwhile, on the farther side of Ring 4, as Arjuna stands in awe
before Kṛṣṇa’s Universal Form, we hear him say,

You are the highest syllable to be known, you are the supreme
resting place of this all, you are the unchanging protector of the
everlasting dharma, I hold you to be the eternal Puruṣa. (11.18)

Here, where Kṛṣṇa will soon reveal himself to be “Time grown old for the
destruction of the worlds” and urge Arjuna to be Time’s “mere instrument”
(11.33), we again see movement. First, Kṛṣṇa tells Arjuna that he rescues dharma
from yuga to yuga; next, he moves on to explain his vaster role in bringing
about creation and dissolution through the rhythm of kalpas (9.7–8); and only
after that will Arjuna recognize him to be “the unchanging protector of the
everlasting dharma.” Moreover, between this recognition and learning that
Kṛṣṇa is “Time grown old,” Arjuna twice cries out “O Viṣṇu!” (11.24, 30).
This ring’s verses bearing on dharma over divinely ordered time thus form
a powerful overarching statement. Yet what they overarch is Kṛṣṇa’s instruc-
tions on living dharma over ordinary time experientially. With patience and
affection for Arjuna, he fine-tunes this instruction all over the BhG text not
only for Arjuna but, by the grace of Vyāsa, for the inner warrior in anyone.

A Ring Off-Center?

If, nonetheless, we ask whether the BhG offers a central focus on dharma, we
should know that it will remain a question. I believe there are four choices. The
first and simplest is to take Arjuna’s euphoric description of Kṛṣṇa as “the pro-
tector of everlasting dharma” to be that center. It has the merit of being the one
mention of dharma in the famous eleventh chapter, which some would take as
the acme of the BhG in that it discloses its theophanic structure. Another would
̄ to have saved its deepest disclosure for the verse where Kṛṣṇa
be for the Gıtā
tells Arjuna he should abandon all dharmas since Kṛṣṇa will release him from
every sin (18.66). For Śrī Vaiṣṇava sectarians of Viṣṇu, this verse is called the
̄ ’s carama śloka or “final, summarizing verse.” Yet if our analysis has any
Gıtā
merit, either would be a disappointing conclusion, since both occur on the far
564 dharma

sides of rings rather than at any real center. An interesting candidate nested
within all four rings has been mentioned. This is the one instance where Kṛṣṇa
speaks of dharma in the sense of his “teaching” or “doctrine”: “Men who lack
faith in this dharma, enemy-burner, having failed to reach me, they return to
the runaround of deaths” (9.2–3). This is a fairly powerful verse, and it has
clear soteriological implications, but of a negative sort. Arjuna would have
a right to expect something more positive from Kṛṣṇa, who has moved him
beyond his initial fear of going to hell.
This brings us to our final candidate, which takes a little explaining as to
how it could be at a heart of things. In introducing the subject of dharma rings
circling around and within the BhG, I have referred to the possibility that the
̄ ’s deepest message on dharma may be lodged in a center that seems off-
Gıtā
center or deferred to its twelfth to sixteenth chapters. Dennis Hudson has pro-
posed that these five chapters have what has been called a “barleycorn” structure,
with chapter 14 as the kernel. To simplify, having resolved Arjuna’s familial
and class anxieties in chapters 1 to 6, and built up to his awesome theophany
from chapters 7 to 11, Kṛṣṇa now pauses to tell Arjuna some of the BhG’s deep
implications, couching them as secrets and mysteries. Given that chapters 12
and 16 and 13 and 15 can be read as continuous discussions, chapter 14 would
lie at the center of this pattern, possibly on the analogy of the fourteenth night
being that of the full moon, and marking a transition from the increasingly
luminous to the gathering darkness. Looked at in this fashion, after closing
chapter 16 on the topic of “demonic people,” Kṛṣṇa would devote chapter 17 to
people of different faiths before offering the encouraging closures of chapter 18
(cf. D. Hudson 2001).
There are some problems with this theory (see Hiltebeitel 2002), but they
diminish as one works inward. Chapters 12 and 16 can be read continuously
only as a discussion of the virtues and vices held by people of different natures.
Kṛṣṇa mentions four types of devotees who are each dear to him in chapter 12,
and briefly takes up “divine people” (who may be the same as those mentioned
in chapter 12) in chapter 16 before getting to the “demonic people” just men-
tioned. Chapters 13 and 15, however, feel more thematically continuous: the
former introduces Kṛṣṇa as the field-knower who in the latter plants the first
seed that yields the cosmic upside-down fig tree, which Kṛṣṇa invites Arjuna to
fell at its roots with the axe of detachment. This seeding theme readily relates
to the kernel verses of chapter 14. This is how Kṛṣṇa begins that chapter:

Further I shall declare the supreme knowledge of knowledges

knowing which all the Munis have gone from this world to supreme
success. Having resorted to this knowledge, they came to have the
 dharma and the bhagavad gı̄ t ā 565

same nature as me (mama sādharmyam). Even at the Creation they

do not take birth, and they are not disturbed at the Dissolution.
My womb is the great brahman. . . . (14.1–3)

Clearly, these words meet our criteria for being a kernel about dharma. Their mes-
sage is positive. They seem to be centered in a ring of five chapters: albeit off-
center from our other rings, yet in a ring-pattern plausibly designed for them. And
their first verse is obvious about declaring its centrality as “the supreme knowledge
of knowledges.” But what is dharma here, and what is the positive message?
Most translators have found ways to translate sādharmyam without making
obvious reference to any of the usual meanings of dharma. But there is no good
reason to be obscure. The term derives from sa-dharma, which can mean either
“having the same nature” or “subject to the same laws or duties.” Clearly, the
former is preferable,81 as it is when Manu says the “delinquent-born” and
Śūdras “have the same natures” (M 10.41). Much as when Kṛṣṇa describes him-
self as “Time grown old for the destruction of the worlds” (11.33), his “nature”
survives the Creation and Dissolution of the universe. This befits a god credited
not only with preserving the universe but the eternal dharma. Yet the good
news is more immediate. The Munis who “have gone from this world to
supreme success” have the “same nature as me,” being neither reborn nor
disturbed. It is on that note that Kṛṣṇa continues to be reassuring:

My womb is the great brahman. In it I place the germ and the origin
of all beings comes about, Bhārata. In all wombs, Kaunteya, whatever
forms come into being, the great brahman is their womb. I am the
father who bestows the seed. (14.3–4)

From this point Kṛṣṇa begins talking about the three Qualities of Nature and
the types of bondage each incurs, yet which can be transcended by attaining
Kṛṣṇa’s “being” (bhāva; 19).
We may find it surprising after all the talk about Kṣatriya dharma, and in
particular Kṣatriya svadharma, to find Kṛṣṇa telling Arjuna that the model for
attaining “my being” (mad-bhāvam)—which would have to be Kṛṣṇa’s
svabhāva—lies in the Munis who have “the same dharma as me,” in that they
survive Creations and Dissolutions of the universe unaffected and undisturbed.
Yet Arjuna has been prepared to understand this in the BhG itself. Earlier, he
has been told how the Munis and Ṛṣis are yogins who attain “the felicity of
Brahman” (brahma-nirvāṇa) and “become brahman (brahma-bhūta)” (5.24):

81. And usually preferred, although see Thompson 2008, 67, who, in his usually careful translation, takes,
I believe, the less likely alternative and, without explanation, uses two words to translate dharma: “come to have
the same virtues and duties that I have.”
566 dharma

The Ṛṣis obtain the felicity of brahman, their sins destroyed, their
doubts cleft, their selves restrained, delighted in the welfare of all
beings. To ascetics detached from desire and anger, their minds
tamed, who know themselves, the felicity of brahman lies near.
Keeping outside contacts out, centering the eye between the eyebrows,
evening out inhalation and exhalation within the nostrils, controlling
the senses, mind, and spirit, the Muni intent upon mokṣa, whose
desire, fear, and anger are gone, is released forever. Knowing that
I am the recipient of sacrifices and austerities, the great lord of all the
world, the friend of all beings, he attains peace. (5.25–29)

Unlike such Sages, however, Arjuna, when he was given the divine eye to
witness Kṛṣṇa’s Universal Form as Creator and Destroyer, was affected, indeed
overwhelmed, left stammering and bowing in adoration, imploring Kṛṣṇa to
show his grace “as a father to a son, as a friend to a friend, as a beloved to
a beloved” (11.44).
̣ ạ ’s theophany may make his case for Kṣatriya svadharma overwhelming,
Kṛsn
but he is not rushing his friend Arjuna on the deeper matters. Earlier, he has
told him it takes a while to take in what he is saying:

For there is no purifier here the like of knowledge; in time, one who
is perfected by yoga finds that in himself. The one who has faith
obtains knowledge, intent upon it, his senses controlled; having
obtained knowledge, in not a long time he finds the highest peace.
(4.38–39; cf. 5.6)

Mindful of Kṛṣṇa’s relaxed approach, let us now look at how chapter 14, pro-
posed as a deferred center, closes after Arjuna has learned that attaining “the
supreme knowledge of knowledges” has to do not only with being, like the
Munis, of the “same nature” as Kṛṣṇa, but with transcending the three Qualities
of Nature to attain Kṛṣṇa’s own “being.”
Arjuna wants to know the traits and conduct of one who transcends the
three Qualities, and how one does it. Kṛṣṇa replies:

He does not hate illumination, activity, and even delusion when they
arise, Pāṇḍava, nor wish for them when they have ceased. Sitting as
one who is sitting apart, who is not agitated by the Qualities, thinking
only, “The Qualities are at work,” who remains firm and is not
stirred; who is the same in happiness and unhappiness, self-abiding,
for whom clods, stones, and gold are the same, alike to those dear
and undear, steady, alike to blame and self-praise, alike to honor and
dishonor, alike to the sides of friend and foe, who abandons all
 dharma and the bhagavad gı̄ t ā 567

undertakings, he is said to have transcended the Qualities. And he

who serves me with unswerving bhaktiyoga, having transcended
these Qualities, is fit for becoming brahman. For I am the foundation
of brahman, of the immortal and the unchanging, of the everlasting
dharma, and of the absolute happiness. (14.21–27)

Kṛṣṇa not only allows that Arjuna will need time to digest what he has to say;
he says, do not hate what arises from the three Qualities—illumination from
Goodness, activity from Passion, or even delusion or bewilderment (moha)
from Darkness, or wish for whatever of them has ceased. This could describe
how Kṛṣṇa or the Sages experience the Qualities, or it could be preparing
Arjuna for a long, passionate, and bewildering war.82 The three Qualities
function here something like the Buddhist dharma theory: one should not be
attached to them as they rise and fall, they are not what one really is. But to
know the “knowledge of knowledges” is to know that they arise from Matter
that is seeded by Kṛṣṇa as the “womb of the great Brahman” in which all selves
find their origin, and their absolute happiness and highest felicity. For that
knowledge to make one “fit for becoming brahman,” it would also recall the
earlier passage just cited (5.24–29), where “becoming Brahman” was likewise
used to describe the Ṛṣis and Munis’ attainment of brahma-nirvāṇa, which
I have translated as “the felicity of brahman.” Although it is controversial, the
fact that the BhG makes such a strong use of the term brahma-nirvāṇa in
conjunction with “becoming brahman,” which is also used in early Buddhist
texts to describe nibbāna and even the Buddha’s attainment of it (see SN iv.
94–95), is probably an indication, one of many, that the Mahābhārata wants
Kṛṣṇa to be saying something different from the Buddha.
The BhG does not tell us what Arjuna makes of this “knowledge of knowl-
edges” that reveals the endgame of bhaktiyoga, or how he factors these inter-
ludes about the Ṛṣis into his more pressing concerns with Kṣatriya svadharma.
But these passages are suggestive for getting at what the Mahābhārata, and
probably also the Rāmāyaṇa, has to say about the relation between dharma and
bhakti, the topic of our next chapter. Let us anticipate a finding we shall meet
there, that the Mahābhārata speaks of a lofty concept it calls both “the dharma
of the Ṛṣis” and “the dharma of the Munis,” and ask, on the hypothesis that

82. See also BhG 18.30–32: whereas the sattvic temperament has the intellect (buddhi) with which to dis-
criminate pravṛtti (worldly activity) and nivṛtti (cessation from activity), things to be done (kārya) and not to be
done (akārya), and bondage from mokṣa, the rajasic and tamasic temperaments have, respectively, the buddhi
only to incorrectly discriminate between dharma and adharma and to confuse them. These latter two tempera-
ments may thus have the capacity to know right from wrong, but are prone to mistake or muddle them and need
the sattvic temperament to tell them what is “to be done.” This of course restates the svakarma/svadharma oppo-
sition in terms of the guṇas.
568 dharma

Kṛṣṇa could be talking about it, what the BhG would have told us about this
unusual dharma. From these passages, we can start out minimally with this. If
Kṛṣṇa and the liberated Ṛṣis know themselves to have “the same nature,” and
if Arjuna could know this too, given time (and Time), it would relate to a
delight in the “welfare of all beings” and a friendship extended to “friend and
foe alike.”
̄ has to say
Finally, if BhG 14 gives us the deferred heart of what the Gıtā
about dharma, it is not so much a center as another ring, beginning with
dharma in the sense of Kṛṣṇa’s ultimate salvif ic “nature” and ending with the
“everlasting dharma” that has its foundation in Kṛṣṇa like the brahman one
can become through knowledge and unswerving bhaktiyoga. I suggest we
think of this centerless ring as centered in “becoming brahman” like the
perfected Ṛṣis who attain brahman’s felicity, coming “to have the same nature
as me”: a golden ring to catch while the merry-go-round goes round.
12
Dharma and Bhakti

A felicitous contrast made by Madeleine Biardeau offers us a point of

departure for this chapter. Whereas the Mahābhārata marks a bhakti
“swerve” (écart) in the Brahmanical tradition, Manu, she says, “‘budges’
as little as possible” in its allegiance to the Veda (2002, 1: 85, 87, 96).
My working hypothesis for this chapter is that Manu and the Rāmāyaṇa,
both probably from around the same time and a little younger than the
Mahābhārata, could have shared a perception that the Mahābhārata’s
treatment of dharma was ambiguous and its presentation of bhakti
amorphous—an indecisive and potentially exasperating combination
that each would seek to tighten up, but in different ways: Manu, by
screening out bhakti and getting orthodox about dharma; the Rāmāyaṇa,
by streamlining and straightening out both dharma and bhakti around
the figure of a royal perfect man. For Manu, bhakti was not germane to
its programs of making dharma a primordial civilizational value and
averting upheaval by reinforcing orthopraxy. For Vālmīki, who may have
felt less agitation about the heterodoxies than these other two texts,1 but
who would seem to have shared Manu’s worries about uprising Śūdras,2

1. See Biardeau 1999, xxxiv–xxxv, seeing both epics as “ripostes” to “imperial Buddhism,” but
with “le menace” less present in the Rāmāyaṇa, where the incarnation of Viṣṇu (Rāma) is more
“ancient” than Kṛṣṇa and does not, like Kṛṣṇa, yet have a cult on the ground in northern India; and
where the “présence d’un danger bouddhique” is displaced on to Rākṣasas who have their southern
base in distant Laṅkā, “an island that has just been converted.”
2. On Manu and Śūdras, see chapter 5 § E. The Rāmāyaṇa takes this stance in its story of
Rāma’s killing of the upstart Śūdra Śambūka (7.65), mentioned briefly in chapter 11 § A. A reference
to this story occurs at Mbh 12.149.61–63; see Brockington 1998, 427; Fitzgerald 2004a, 770.
570 dharma

divinity could be useful in portraying an ancient king. On such a figure, he could

focus a few of the bhakti themes developed in the Mahābhārata, but often shading
the term bhakti to mean “loyalty” rather than “devotion”3—as more befitting a
divine king (such as one meets also in Manu) than a divine friend like Kṛsn ̣ ạ .
Keeping Manu in mind as what might be called an interested but bhakti-
allergic onlooker, a book on dharma offers an opportunity to think about how bhakti
and dharma are related in the two epics. Rather than looking at them for ways to
separate out bhakti and dharma into isolable thematic strands4 or strata,5 it could be
fruitful to map dharma and bhakti in both epics together. But what kind of cartog-
raphy best suits this two-text terrain? As I have maintained at various points in
chapters 9 and 11, it would not be a matter of making the centrality of the Bhagavad
Gītā a pretext to read karmayoga, with or without its bhaktiyoga-overtones, as a
cross-the-board hermeneutic. Yet the Gītā does point out a useful direction. For
one thing, in being—at least as we have it—a didactic bhakti text, it reminds us that
the vast scholarly operation of separating out bhakti from dharma (as the “didactic,”
along with other strands, like “epic philosophy”) has always required a certain art-
istry in keeping such things separate.6 More constructively, though, the Gītā is also
a text that displays the two features that I have found most useful in mapping the
relation between dharma and bhakti across both epics. One is friendship, which has
provided the primary contour lines in three essays I have written on “mapping
bhakti in the Sanskrit epics”: the first, mapping bhakti with just friendship; the
second, mapping it with hospitality and friendship; and the third, mapping it with
friendship, hospitality, and separation.7 The other feature is the notion of a divine
plan. In this case, writing this book about dharma has led me for the first time to
try to seriously address this topic. I broached the relation been dharma and the

3. But see Rām 2.40.27 (as Rāma leaves Ayodhyā all moving and unmoving things have devotion to him
and ask his devotion to them); 6.105.28 (men of devotion will praise Rāma as God now that he has slain Rāvaṇa).
More as “loyalty,” see 2.46.30 (Rāma’s charioteer Sumantra asks him out of loyalty to return to Ayodhyā); 3.15.25
(Bharata remains loyal to Rāma).
4. Fitzgerald, who is fond of the metaphor of “threads” running through the Mahābhārata, speaks to “the
potential value of a careful mapping of the thematic threads of the vast epic” (2006a, 272), and has offered ways
to reorient discussion principally around the politics of that text. He makes a good point that the Mahābhārata
poets have a political agenda centered on the vigorous promotion of varṇadharma, svadharma, a king empowered
with the rod of punishment (daṇḍa), and the abhorrence of varṇasaṃkara (mixture of social classes), and that this
agenda is intelligible in a post-Mauryan setting (2006a, 275–77). Regarding bhakti, however, his practice is to
map it out as “late.”
5. Brockington 1998 has done the most to isolate dharma and bhakti in the epics into separable textual
units, from as small as the verse to as large as whole books, so as to date additions to the supposedly original
bardic heroic core.
6. The problem comes up most interestingly in the Mokṣadharma Parvan, where a prominent bhakti unit,
the Nārāyaṇīya, comes near the end. In Hiltebeitel 2005d, 259–61, I mentioned Zaehner 1963 and Brockington
2000b as shedding some light on this question, which I will return to in chapter 13.
7. See Hiltebeitel 2010a; in press, and 2011a, chapter 11 respectively. See also Hiltebeitel 2007a, and for
initial discussion of separation, chapter 10 § D.
 dharma and bhakti 571

epics’ divine plans in chapter 6, and have used the term repeatedly since then—
except in chapter 11, where the Bhagavad Gītā makes mentioning it superfluous.
Mapping dharma and bhakti in the Mahābhārata’s divine plan now also provides
the occasion to develop a point noticed briefly in chapters 1 and 7: that the
Mahābhārata represents itself as including the Harivaṃśa as its “Appendix.”

A. Mapping the Divine Plans

A few words are in order about where we are already with the project of mapping
these epical divine plans, how we can proceed further, and how mapping divine
plans will tie in with mapping friendship, hospitality, and separation. First, we
may distinguish the divine plans of the epics from Brahmā’s undertaking of insti-
tuting dharma for all times (including yugas and manvantaras) through Manus,
since the latter includes no supernatural incarnations. Second, here in three steps
is what we have mapped so far of the divine plans of the epics themselves.8

1. We have reached a point where we are able to discern something about

the beginning and duration of the divine plans in both epics. In the
Mahābhārata, from one angle, the divine plan seems to kick in with
Gaṅgā’s intervention in the Pūru–Bhārata–Kuru lineage; and, from
another, to be fulfilled at some indeterminate or variously described
point in the transition between the Dvāpara and Kali yugas.9 The
Rāmāyaṇa’s divine plan follows from the loophole in the boon Rāvaṇa
obtains from Brahmā: the gods can take advantage of the fact that,
because Rāvaṇ̣a disdained men, he did not request invulnerability
from them.10 Each epic also has singular episodes that offer
confirmations that things are moving along according to plan.
Some are abrupt and unsettling, like Kṛ̣ṣṇa’s little dance at the
death of Ghaṭotkaca,11 or the Ṛ̣ṣis’ “thrill” at the abduction of Sītā

8. Some of this section is distilled in Hiltebeitel forthcoming-g.

9. On Gaṅgā’s intervention, see chapter 8 § C. For some discussion of the “twilight” transition between the
Dvāpara and Kali yugas, see chapter 6 §§ C and D and chapter 7 § A. The most arresting of the Mahābhārata’s nine
passages (see González-Reimann 2002, 86) to describe the Dvāpara–Kali transition pinpoints the Kali yuga’s arrival,
̣ ạ ’s words, to the fall of Duryodhana after he has been dealt a low blow to the thigh by Bhīma (Mbh 9.59.21). It
in Kṛsn
is the only one of the nine for which González-Reimann can do nothing stratigraphic to say it is “late,” since it is part
of the main story and, moreover, spoken by Kṛsn ̣ ạ . So he resorts to the most unconvincing types of higher critical
explanations: it “would have been inserted” to explain “the blatant breach of dharma this incident entails,” and its
“probable lateness . . . can also be gleaned at from the fact that it is put into Kṛsn ̣ ạ ’s mouth” (101–2). For different
interpretations of this episode, see Biardeau 1994, 45–47; 1997a, 111; Hiltebeitel 2001a, 152–53 n. 92.
10. See Pollock 1984, cited in chapter 9 n. 6 and chapter 10 n. 15.
11. See chapter 9 § E.2.b, and see Couture 2001, 322 on Kṛṣṇa’s dance maneuvers as avataric stagecraft in
the killing of Kaṃsa.
572 dharma

(Rām 3.50.10–11), cited below. Others are long awaited and more
affirmative, as when the Great Ṛ̣ṣis on the Hundred Peak Mountain
give names to the newborn Pāṇḍavas “with bhakti” (chapter 8 § H), or
when the three eldest Pāṇḍavas confirm support from their divine
paternal connections, and in Arjuna’s case from Śiva, in Mahābhārata
Book 3;12 or when Rāma challenges the Ocean to allow him passage to
Laṅkā (Rām 6.14). Many of these scenes can be called bhakti tableaux,13
and once one is accustomed to recognizing them, they are not hard to
find or easy to ignore. One can also track the divine plans in the epics’
great royal rituals, the Rājasūya and Aśvamedha, which are, as sacri-
fices, already arenas in which the gods, in principle, can touch base on
earthly matters that concern them.
2. We have identified two terms that index the divine plans dispositively
in both epics. Most significant so far, because of the way it has allowed
us to detect outcroppings of the divine plan in the Mahābhārata, is “the
work of the gods” (devakārya, surakārya).14 The Rāmāyaṇ̣a also uses the
phrase, but with a noteworthy difference. Whereas the Mahābhārata
usually uses it in disclosures about (and sometimes to) characters of
the main story,15 the Rāmāyaṇ̣a has Rāma hear the phrase mostly from
Ṛ̣ṣis who are telling him about how the work of the gods worked long
ago.16 The second term is “the secret of the gods”: paradoxically,

12. Arjuna with Śiva and Indra (Mbh 3.38–45); Bhīma with Vāyu through meeting Hanumān, both being
Vāyu’s sons (3.147–50); Yudhiṣṭhira with Dharma (see chapter 9 § D). Bhīma gets hints about the divine plan
from Hanumān at 3.150.5–9.
13. See Hiltebeitel 1984, 2 for this term, by which I meant to invoke iconic visual representations that can
also be represented in ritualized drama.
14. We have seen it used with reference to Gaṅgā (1.92.49; see chapter 8 § C) and Draupadī (1.155.45; see
chapter 10 § D). Hardly irrelevantly, devakārya also often means “worship,” “what must be done for the gods”
(Biardeau 1981b, 80; 2002, I: 221).
15. See Mbh 3.41.37: Indra commissions Arjuna to do “very great devakārya” by fighting the gods’ enemies
in heaven; 3.89.17: reporting Arjuna to have done this “great surakārya”; 3.164.16: the World Guardians tell
Arjuna he has seen Śiva to accomplish their surakārya; 3.181.39: Mārkaṇḍeya tells Yudhiṣṭhira the Pāṇḍavas have
come to earth from the world beyond for the sake of the surakārya; and most representatively, at 15.39.5–7 Vyāsa,
preparing the blindfolded Gāndhārī to see her slain sons and others rise from the Gaṅgā, tells her “O faultless
one, the surakārya could not but be accomplished. All these descended [avaterur] to the surface of the earth with
their divine portions, Gandharvas, Apsarases, Piśācas, Guhya[ka]s, and Rākṣasas, as also even meritorious folk—
Siddhas and Devarṣis, gods and Dānavas and taintless Brahmarṣis; they met death on the Kurukṣetra battlefield”
(see Ganguli [1884–96] 1970, 12: 238). In an exception to such disclosures being made usually to human charac-
ters, Śiva tells Umā that, in being half his body, she “does the surakārya and undertakes the continuity of the
worlds (surakāryakarī ca tvaṃ lokasaṃtānakāriṇī)” (13.134.9–10; cf. Ganguli 11: 315).
16. See Rām 1.63.11: Viśvāmitra tells Rāma how the gods got Agni to do their surakārya in contributing to
the birth of Kārttikeya; 2.109.12: Atri tells Rāma how his wife Anasūyā once occasioned the devakārya, ending a
drought by using her ascetic power to get the Gaṅgā to flow; 3.10.15: the Muni Dharmabhṛt tells Rāma why mirac-
ulous sounds come from a lake: f ive Apsarases interrupted another sage’s tapas to accomplish the kārya of the
gods (surāṇām); the sage then built an underwater house where they make love and beautiful music. Similarly,
but without using these terms, Viśvāmitra tells Rāma that the digging up of the earth by Rāma’s ancestors, the
 dharma and bhakti 573

a more obvious term, since, at least in the Mahābhārata, one meets

it in discrete scenes where sagely narrators disclose the Mahābhārata’s
godly subplot,17 much to the distaste of certain scholars who share van
Buitenen’s view that these are “late” scenes of “inept mythification.”18
Yet we have also noticed the term cropping up half-knowingly in
Yudhiṣ̣ṭhira’s contretemps with Draupadī in Mahābhārata Book 3
(see chapter 10 § D). Here, I must regretfully disagree with Brodbeck’s
approach to the “secret of the gods” in an article that is favorable to the
epic’s short-term composition (2009c, 37–38 and n. 17). Although it is
true that most of the Mahābhārata’s Kṣ̣atriya women, whose grief over
their slain husbands the article is about, are uninformed about the
“secret of the gods” and are victims of its “Kṣ̣atriya ideology” (47), it is
overstated and an oversight to say, “The MBh’s central characters are
ignorant” of that secret (see nn. 15 and 17 above). Brodbeck speaks of
a “lower level” of human action that “remains primary” over a “higher
level” (40) one, whose “cosmic interpretation of the war is used
sparingly” (50). Dividing matters so that the lower level one is for epic
characters and the higher level one “primarily to remind the audience”
(39–40, cf. 51) is to overlook how the characters, often as audience
themselves, are drawn into the same story, and how they are often only
a step “behind” Janamejaya and other audiences in figuring things out
(see chapter 9 on Yudhiṣṭhira and “The Yakṣ̣a’s Questions”).19 It is

sons of Sagara, was something Brahmā reassured the gods about (1.38.23–39.4) since it was foreseen as prelude
to the descent of the Gaṅgā. See, however, 6.105.26 where Brahmā, just after revealing Rāma and Sītā’s divine
identities, tells Rāma, “Thus have you accomplished our purpose (tad idaṃ naḥ kṛtaṃ kāryaṃ tvayā). . . . Rāvaṇa
has been slain. Now, Rāma, in your delight, please return to heaven.” Which does not happen yet (see Goldman
and Sutherland Goldman 2009, 1,454), as Agni now returns Sītā to Rāma (106.1–9), and Śiva tells Rāma to wait
until he has established his lineage in Ayodhyā (107.5–6).
17. Vyāsa to Drupada (1.189); Vyāsa to Dhṛtarāsṭ ra (11.8.20–26 and 15.35.11–22); Vyāsa to Gāndhārī
(15.39.5–16); Nārada musing to himself as reported by Vaiśaṃpāyana to Janamejaya (2.33.11–20); and Vaiśaṃpāyana
directly to Janamejaya (1.61; 1.109.1–4; 18.5); in the latter case by a simple allusion (18.5) since Janamejaya has
heard all these other iterations by that point at the end. Assembling most of these passages and discussing espe-
cially the latter, see Austin 2009, 601–3, 606–7, 610, 619–23. Mentioning most of these and others as well,
including Duryodhana’s learning of the secret of the Asuras (Mbh 3.240.103,0), see Brodbeck 2009c, 33–41.
18. See Hiltebeitel 2001a, 164 n. 118 on van Buitenen’s view (1973, xix–xx), shared by others including
Bigger (1998, 100) and going back at least to Winternitz 1933–34, 174, that the story Vyāsa tells Drupada to justify
Draupadī’s polyandry is “silly” and “inept mythification.” Cf. Malinar 2007a, 3, finding no “overarching frame-
work” or “red thread” such as in the Iliad; E. Hudson 2006, 148–59, finding Vyāsa’s explanation to Dhṛtarāṣṭra
that he has lost his sons by a “divine design” to unburden the earth (11.8) less persuasive than consolations by
Saṃjaya and Vidura. See also Austin 2009, 620–22, critiquing additionally, among others, Mangels 1994,
55–59, for positing “a one-time act of mythologization” of an originally bardic core.
19. Finding “depart[ure] in details” and “discrepancies” (2009c, 36) in the ways the divine secret (whose
first disclosure, involving the unburdening of the earth, is misleadingly called “the frame story”) is told in dif-
ferent situations, Brodbeck gives van Buitenen’s “inept mythologization” a “whether . . . or not” free pass (37)
and even what seems to be a final note of sympathy (54). Similarly, in his book on Mahābhārata genealogy, he
writes “we see how cosmic stories might retrospectively arise” (2009a, 262).
574 dharma

really a strain to imagine Draupadī or Sītā, and others in their stories,

not being intrigued by their miraculous births.20 Again, the Rāmāyaṇ̣a
brings a different twist. Rather than the secret of the gods being an
anterior story, the secret of the gods is Rāma himself. As Daśaratha’s
ghost says to Lakṣ̣maṇ̣a, among other revelations of the divine plan
once Rāvaṇ̣a is slain: the gods say that Rāma is “Brahman, the unman-
ifest and imperishable Supreme Spirit, the secret heart of the gods
(devānāṃ hṛ̣dayam . . . guhyam)” (6.107.31; Goldman and Sutherland
Goldman 2009, 465). Also, in the Mahābhārata only, both the
“work of the gods” and the “secret of the gods” refer to what is called
“Earth’s business” and “the purification of the earth.”21
3. These dispositive terms have allowed us to track other more
ambiguous verbal cues found in both epics that can be read as
referring to their divine plans. One, which I follow up in the next
section, has been noted in derivatives of Ödhā and vi-Ödhā to indicate
what is “ordained.” Another is the use of amānuṣ̣a, “inhuman,” to
describe Rāma’s “inhuman manliness.”22 This is actually a rich vein
of half-realized truths and cross-references among the themes so far
mentioned, from which I cull just a few. The Mahābhārata’s
Parvasaṃgraha says the epic will tell about Draupadī’s “superhuman
marriage as ordained by the gods” (devavihito vivāhaś cāpy amānuṣ̣aḥ̣;
Mbh 1.2.88). Ghaṭotkaca was “inhuman though born from a human”
(amānuṣ̣aṃ mānuṣ̣ajaṃ; 1.143.30). Draupadī in disguise hears that she
will steal everyone’s love with her “superhuman body (vapus
amānuṣ̣am)” (4.8.23). Dhṛ̣taraṣṭra tells Saṃjaya how he fears each
Pāṇḍava: “Those Indras among men have cast out their superhuman
net (amānuṣ̣aṃ manuṣ̣yendrair jalaṃ vitatam) in the middle of my
army and will kill it off” (5.52.7).23 Finally, having told Dhṛtarāṣṭra
some of the divine secret as it bears on the god Dharma’s double
incarnation in Vidura and Yudhiṣ̣ṭhira, and leading up to his miracle
of making the slain warriors, including the blind old king’s sons,
appear for one night rising out of the Gaṅgā, Vyāsa asks him, “What
superhuman (amānuṣ̣am) [feat] do you wish to obtain from me,

20. See chapter 10 § A, and n. 13 of that chapter for Brodbeck’s handling of Draupadī’s birth.
21. On “Earth’s business” (bhūmeḥ kṛtyam; Mbh 1.58.41) and “the purification of the earth” (bhuvaḥ
śodhana; 1.58.51), see chapter 6 § C at n. 52.
22. On Rāma’s vīryam amānuṣam, see Rām 7.17.29 and chapter 10 § A, noting a Mahābhārata usage cited
in chapter 9.
23. Note that Bigger’s dismissal of the story of the five former Indras (the Pañcendra-Upākhyāna) on the
grounds that “it stands isolated in the Mahābhārata,” and that one “can take out these verses” (1998, 159, 161; cf.
100), does not stand up to such half-knowing references to the Pāṇḍavas.
 dharma and bhakti 575

O king? To see, to touch, or to hear? Speak, I will do just that”

(15.35.25). Meanwhile, in the Rāmāyaṇ̣a, when Bharata tries to
convince Rāma to return with him to Ayodhyā, he says, “Some say a
king is mortal; I esteem him a god, whose conduct in matters of
dharma and artha they say is inhuman (yasya . . . vṛ̣ttam āhur
amānuṣ̣am)” (Rām 2.95.4). I also flagged a third such verbal cue
where Draupadī “descends” into the arena for her svayaṃvara (Mbh
1.176.9–30). This is one of numerous instances in both epics where
derivatives of the prefixed verbal root ava-√tr.¯ hint at what lies ahead in
the future Purāṇic development of the term avatāra.24 The epics’ avant
la lettre unfolding of what is to become of this concept will be the
topic of the third section of this chapter.

These are all matters one can build from, and, in proceeding further, we are
fortunate to be able to draw on three scholars who, in the last decade, have shown
that the dismissal of the Mahābhārata’s divine plan should not be so easy. Fernando
Wulff’s long chapter on “El Plan Divino” (2008, 81–146) serves to introduce all
the other “connection points” that he believes the Mahābhārata poets reworked
from a “Greek repertory.”25 I can sum up what is pertinent to our discussion in
four points: (a) The “plan of Zeus,” which is also a secret, is undertaken to bring
about the unburdening of the earth after Zeus’s agreement to do so with the earth
goddess Gaia, and his deliberation with Themis, the “divine embodiment of the
natural order,” in whom Wulff sees a similarity to Dharma (111–12). (b) These
stories on the background of the Trojan war are found in texts closest in time to
the Iliad, most notably the Cypria, known through a digest by Proclus and some
fragments, but also echoed in the Odyssey and Hesiod (110–14), that interpret the
Trojan war as “the story of an announced annihilation not only of Troy but of an
entire generation of heroes” (81). (c) That interpretation would be the window
through which post-Alexander the Great Indian poets could have come to know
the Greek epic. (d) Yet classical Greek scholarship in the last two centuries has
renewed a resistance to this interpretation that goes back to Socrates, which is

24. See chapter 10 n. 23. As noted, Couture 2001 cites this passage in an article that traces all instances of
the verbal root ava-√tṝ in the Mahābhārata. I too was tracking such usages in Hiltebeitel 2001a, 70 n. 135 (the
Pānḍ ạ vas and Kṛsn
̣ ạ “descending upon Kurukṣetra” to hear Bhīṣma [1.48.1–3]), 146 (Balarāma “descending from
Plakṣa Prasravaṇa” to Kurukṣetra [9.53.33]), 232 (Nala as the dwarf Bāhuka and his nondriving charioteer Vārṣnẹ ya
“descending the superb chariot” [3.71.18]), 295 (Vyāsa’s four disciples “descending to earth” [12.315.7–8])—the last
two of which Couture does not discuss. In Hiltebeitel 2004a, 224–26, not yet having found Couture’s article,
I discussed these and other passages while developing ideas about a “politics of bhakti” and “an ava-√tṝ convention”
of using derivative forms avant la lettre of actual usages of the term avatāra. One can now benefit greatly from
Couture’s advancement of the discussion around the idea of theatrical usages.
25. Wulff’s book is being translated into English, and I thank him for letting me read drafts of chapters 1
and 2, from which the following quotes are taken. Cf. chapter 7 § A.4 on Wulff’s treatments of the yugas and the
Greek ages, and of Bhīṣma and Achilles as mourned-for sons of water goddesses.
576 dharma

“to deny mythology” (“negar la mitologia,” 109). Clearly the last two centuries of
classical scholarship on Homer has, on these matters, been echoed in Mahābhārata
scholarship like that of van Buitenen. Wulff’s main thesis on the Mahābhārata’s
genesis has the attraction of offering a solution to the question raised by epic as a
new Indian genre (see Hiltebeitel 2001a, 5–7; 2005a, 87 and n. 18).
The other two contributions are those of André Couture (2001), who dis-
cusses theatrical overtones of the root ava-Ötṝ, and Christopher Austin (2009),
who builds on Couture’s point to reinforce the profundity of “Janamejaya’s
Last Question” as to what happened to the Mahābhārata’s supernaturally incar-
nated heroes and heroines once they came to “the end of their karma” in
heaven. I am persuaded by major conclusions of both articles. Couture is con-
vincing that a classical theatrical usage of derivatives of ava-√tṝ yields “a precise
technical term used to describe that movement performed by actors who move
from the stage wings onto the stage itself ” (2001, 319, 324 n. 8), and that this
theatrical usage is pertinent to many usages in the Mahābhārata and Harivaṃśa
that make the world a stage for the play of the gods. As we shall see, the same
applies to the Rāma story in both the Rāmopākhyāna and the Rāmāyaṇa. And
Austin is persuasive about the value of what the Mahābhārata Critical Edition
has Vaiśaṃpāyana ever-so-briefly say in response to Janamejaya’s last question.
Contrary to Nīlakaṇṭha’s forced reading (and redaction) of Vaiśaṃpāyana’s
answer to Janamejaya, in which Nīlakaṇṭha goes to great lengths to argue, in
accord with his own contemporary understanding of karma,26 that some of those
who had supernatural incarnations would have had subsequent karmic des-
tinies and others would not have, Vaiśaṃpāyana’s important real answer is that
everyone dissolved back into their supernatural natures. As Austin astutely
sees, the implication is that they were not “‘like us.’. . . Rather, their post-death
fate is unique to them, and it takes place without further reference to worldly
births, saṃsāric existence, or mokṣa” (2009, 619; my italics). This insight allows
us to deduce that this ending of the Mahābhārata tells us the precise end of
its divine plan. As far as the Mahābhārata is concerned, these heroes and
heroines will not reincarnate,27 nor is there any question of their achieving

26. As Austin points out (2009, 13), the Mahābhārata does not know the later Vedāntic distinction bet-
ween prārabdha karmas, actions whose residues “have determined the present life form and actively ripen (vipāka)
in the present lifetime; their exhaustion is simultaneous with the termination of that lifespan,” and two other
types of karma that carry karmic residues along from past to future lifetimes—a typology that would have inspired
Nīlakaṇṭha to introduce his distinction between heroes ready for liberation and others who are not.
27. This is a fascinating matter, on which Indian martial oral folk epics generated counteropinions in the
medieval period well before Nīlakaṇṭha’s seventeenth-century times, and it is indeed worth considering that
Nīlakaṇṭha may have been motivated to make this argument not only as a Vedāntin but as one who could have been
̄ , popular and well rooted at least today in the Vārānạ sī area where Nīlakaṇṭha
familiar with the Hindi oral epic Alhā
lived, in which Draupadī and many male heroes are reincarnated in this “Mahābhārata of the Kali yuga” to carry out
their “unfinished business” from the Mahābhārata (see chapter 7 § A.1; Hiltebeitel 1999a, 121–296).
 dharma and bhakti 577

mokṣa.28 Kṛṣṇa’s miraculous postwar revival of the stillborn Parikṣit, the

grandson of Arjuna and son of Abhimanyu who incarnated the Moon’s
splendor, was of course part of the divine plan through which Parikṣit could
revive the lunar dynasty in the setting-in of the Kaliyuga.29 But Parikṣit himself
incarnated no one, and neither did his son Janamejaya, who got to live in “real
Kali yuga time” like us. Rather than living in and as part of a divine plan, he,
like us, only gets to hear and ask about a divine fait accompli.
Of course one cannot expect the Mahābhārata to give us total closure on its
divine plan. No doubt it has other as yet unplotted moves, one of which was to
number the “Harivaṃśa Purānạ called an Appendix (khila)” and the “Bhaviṣyat
(‘Future’) called a great wonder among the Appendices (khileṣu)” as the ninety-ninth
and last of its hundred upaparvans or major units in the Parvasaṃgraha (Mbh
1.2.69). Couture addresses what the Mahābhārata might mean by an “appendix”
with “appendices,” noting that “the Indian tradition would rather consider the
HV as a set of addenda completing the MBh,” and that “the HV must be con-
sidered globally to be a set of khilas to the MBh, i.e., as supplements necessary
to a full understanding of the MBh” (1996, 132, 135). Although the Harivaṃśa
actually contains three parvans rather than two, the two that the Mahābhārata
refers to as upaparvans may be meant to encompass all three (Brodbeck 2010e).
In any case, they would “append” further information about the divine plan
that would refocus what has been said so far.
As Couture says, “The HV presents Viṣnu ̣ ’s incarnation as Kṛsn
̣ ạ with more
precise wording, using the same [ava-√tr̄. ] vocabulary but adding some very impor-
tant paraphrases” (2001, 315). Among these, we learn that Viṣnu ̣ receives the news
that Earth is not only being trampled but urbanized30 when he rises from his
cosmic sleep, “just as the Dvāpara yuga is coming to a close.” The gods “ask
Brahmā to tell them what each of them should do (yat . . . kāryam)”31 to relieve
Earth’s pain. For their aṃśāvataraṇa or “descent in portions,” they say, “Let us also
create bodies without passing though an uterus (ayonijāś cāpi tanūḥ sṛjāma).”
Brahmā tells them they will form opposite factions and fight against one another.
Nārada prods Viṣnu ̣ to first dispose of Kaṃsa at Mathurā; and so on, focusing in on

28. This renders rather moot any notion that Kṛṣṇa might be promising Arjuna actual release from
saṃsāra at BhG 18.66 at the far side of the Gītā’s third ring of dharma (see chapter 11 § D). Cf. de Bruin 1998 on
the Tamil Mahābhārata folk theater’s understanding of Karṇa’s “mokṣa” as his death.
29. On this episode (Mahābhārata 14.65–69.11, especially 68.18–24), its lunar symbolism and narrative
ramifications, see Hiltebeitel [1976] 1990, 337–38, 349–50; 2001a, 75–76.
30. The mention of cities burdening the earth is specific only in the Harivaṃśa; cf. Couture 2006, 75 on “violent
cities” in contrast to “the pastoral ambiance of Vedic lore” that he finds behind the image of the dharma bull.
31. Note that Couture highlights the usage of kāryam at HV 43.3. Cf. 43.11: “Having this great resolution of
the gods assembled for one common cause (surāṇām ekakāryāṇām), the grandfather of all, eulogized by the celes-
tials, said to them” (Dutt 1897, 229). See also HV 13.69 (devakāryād); 62.17 (devānāṃ . . . kāryam avyayam); 65.44
(kāryaṃ surāṇām); 81.2 (kāryartho devatānām); 87.41 (devatākārye); 93.6 (surakāryeṣu sarveṣu).
578 dharma

Krṣnạ ’s parents and wives (Couture 2001, 315–16, summarizing HV 43.6–45.80).

We would not have known from the Mahābhārata that Earth’s plight included cit-
̣ awakened to it just before the Dvāpara yuga.32 Nor had we heard
ies, or that Viṣnu
about the gods’ determination to have ayonija bodies. We did see (in chapter 8 § C)
that when the Vasus were cursed to born in a womb, they got Gaṅgā to be their
mother so they would not have to enter “an inauspicious human-female womb”
(na mānuṣīṇaṃ̄ jaṭharaṃ . . . aśubham; Mbh 1.91.14cd). This could be in the
Harivaṃśa’s time frame of the near end of the Dvāpara yuga, so it is probably what
the Harivaṃśa has in mind. Yet the only Mahābhārata characters among those
who take part in the aṃśāvataraṇa to be explicitly ayonija are Draupadī (1.153.8) and
the pot-born Droṇa (1.61.63). While her brother Dhṛsṭ̣ adyumna’s birth from fire
would qualify him as well, most other key figures seem to take birth from human
mothers. This is explicitly so for Duryodhana, who “was a piece of Kali (kaler
aṃśaḥ) born in Gāndhārī’s belly (jaṭhare) to effect the destruction of the worlds”
(11.8.27cd; Fitzgerald trans. 2004a, 41). And Vidura was born in the womb of a
Śūdra (śūdrayonau; 1.101.25c, 27d) because Dharma was cursed by Aṇīmāṇdạ vya.
Interestingly, whereas Yudhiṣṭhira is a “portion” (aṃśa) of Dharma (1.61.84), this
is not, as far as I can see, said of Vidura, so Dharma presumably becomes “fully”
Vidura,33 while in his all-pervading aspect (15.35.19–20) he is also able to subse-
quently sire Yudhiṣṭhira and have his three paternal moments to test him as a
Yakṣa, dog, and afterworld psychopomp—all of which reinforces the way dharma
is so especially and distinctly imbricated in the Mahābhārata’s divine plan.
But most informatively, the Bhaviṣya Parvan (HV 114–18) ends the Pune
Critical edition of the Harivaṃśa, which completed the Pune Critical Edition of
the Mahābhārata. As was mentioned in chapter 7 (§ A.1), the Yuga Purāṇa
rounds off its prophetic twelve-verse digest of Mahābhārata events leading into
the Kali yuga with an allusion to a Bhaviṣya Parvan episode we shall soon now
be discussing. But shortly before that episode itself, the Bhaviṣya Parvan gives
Janamejaya an opportunity to ask more last questions, and the last one he asks
that reflects on the secret of the gods (without calling it that) is nearly as inter-
esting as his last question in Mahābhārata Book 18. It now too plumbs not only
the Mahābhārata, but the depths of the combined Mahābhārata-Harivaṃśa that
he has now just heard almost to the end.
A taut and tightly packed unit, the Bhaviṣya Parvan begins with Śaunaka
asking some questions of Sauti (the bard) about the sons Janamejaya had with his

32. Though Kṛsn ̣ ạ does something transparently similar at the beginning of the Udyogaparvan, waking up at
Dvārakā to see that Duryodhana and Arjuna have come to his bedside seeking his favor in the upcoming war (see
Hiltebeitel [1976] 1990, 103–7).
33. One repeated phrase is dharmo vidurarūpeṇa, “Dharma in the form of Vidura” (1.100.28; 101.27). Cf.
15.35.12: dharmo viduratāṃ gataḥ, “Dharma attained the state or condition of being Vidura” (see MW 963).
 dharma and bhakti 579

wife, initially just called Kāśyā (114.2), who will reappear shortly with a fuller name.
He thus learns about descendants of Janamejaya through two generations down
to a foundling named “Goat-Sides” (Ajapārśva), who is called “the founder of the
Pānḍ ạ va vaṃśa” (114.16). Then, after Śaunaka congratulates Sauti on how much
hearing the “entire Harivaṃśa pleases us” (HV 115.1–2), Sauti tells him what
Janamejaya did next. Having finished his snake sacrifice in Takṣaśilā, he started to
collect materials for a horse sacrifice, which would be completed in Hāstinapura.
Still, however, in Takṣaśilā, Janamejaya summoned his priests and ordered them
to release the horse (115.4–6 ). Learning about it, Vyāsa, “the foremost of the omni-
scient (sarvaparāvarajñaḥ), came suddenly (sahasājagāma) to see” (7). Janamejaya
received him with customary rites of hospitality, and when both had joined the
seated attendees they engaged in “diverse and variegated stories that were
connected with Veda (kathā bahuvidhāś citrāś cakrate vedasaṃhitāh).” ̣ And “at the
end of a story (kathānte),” Janamejaya addressed Vyāsa, “the Muni who was the
Pānd ̣ ạ vas’ grandfather and his own fore-grandfather (prapitāmahaḥ; actually his
great great grandfather)” (8–10). There is a medley here of Mahābhārata echoes.
As elsewhere, the epic mentions variegated stories on Vedic themes exchanged at
Naimṣa Forest occasions (Hiltebeitel 2001a, 99–100, 123). As in the Nārāyaṇīya,
there is a “dip” from the outer frame dialogue between Śaunaka and Sauti through
the inner frame one between Janamejaya and Vaiśaṃpāyana to the ultimate
authority of Vyāsa (Hiltebeitel 2006a; 2011a, chapter 7), who sits here recalling
what I have called the outermost frame in which he first told the Mahābhārata to
Vaiśaṃpāyana and four other disciples before Vaiśaṃpayana narrated it to
Janamejaya (2001a, 92; 2011a, chapter 7). Vyāsa was a sadasya or seated attendee
during the recitation of the Mahābhārata at Janamejaya’s snake sacrifice, and has
now returned as an attendee at the launching of Janamejaya’s horse sacrifice,
which has apparently begun with the release of the horse. He has come suddenly
and will stay only briefly, taking leave after foretelling some matters of the near
and distant future.
Once Vyāsa had been welcomed, then, Janamejaya said,

The Mahābhārata narrative has many meanings and great extent;

by being agreeable to hear, it is like it has gone by for me in a
moment (mahābhāratam ākhyānaṃ bahvarthaṃ bahuvistaram/
nimeṣamātram iva me sukhaśravyatayā gatam. (HV 115.11)

Another nice Mahābhārata touch here: in quoting Janamejaya’s fleeting

description of the Mahābhārata going by in a moment, twinkle, or blink (nimeṣa)
to Śaunaka among the heavenly Ṛsis of the Naimiṣa (Twinkling) Forest, Sauti
is collapsing the inner and outer frames while Janamejaya is collapsing the
Harivaṃśa into the Mahābhārata. Yet Janamejaya, having lauded the author in
580 dharma

this fashion, says he still is like a man dissatisfied with nectar or heaven, and
does not find satisfaction (tṛptim) in the Mahābhārata story (13). Before getting
to the main question he has in mind, he asks the “omniscient” Vyāsa whether
Yudhiṣṭhira’s Rājasūya was the cause of the destruction of the Kurus (13–14).34
As A. Harindranath (2010) puts it, “Janamejaya . . . correctly deduces that
Yudhiṣṭhira’s Rājasūya was the root cause.” Harindranath’s study expands
from the moment in the Mahābhārata on the day of Yudhiṣṭhira’s royal conse-
cration, just before Kṛṣṇa’s killing of Śiśupāla, when “Nārada watches the vast
assemblage of kings and recalls a tale he had heard long ago in the dwelling of
Brahmā at the time of the aṃśāvataraṇam, [and] it begins to dawn on him that
this human assembly was in fact an assembly of gods” all doomed to destruc-
tion (Mbh 2.33.11–20; cf. Couture 2001, 321). Janamejaya recalls this, and
having also “listened carefully to the entire (disastrous) history of his great
grandfathers, having learned about the Rājasūyas of Varuṇa and Soma and the
accompanying great wars,” he “now confronts Vyāsa” (Harindranath 2010):

You are even the grandfather of our ancestors, knowledgeable of past

and future, and our first lord. How, with you as their guide, did those
intelligent ones, as if having none to govern them, and deviating
from the paths of morality, commit sin?35

Vyāsa replies:

Being seized by time, child, your grandfathers did not ask me about
the future. And unasked, I did not speak. Besides, I see the
announcement of the future as unsuitable; I am surely not able to
counteract a course that is conditioned by time.36

Vyāsa then characteristically takes what he wants from this exchange to direct
Janamejaya’s interest to two futures he will now tell him about (once he puts the
questions into Janamejaya’s head). One is the future of the Kali yuga, in which
Janamejaya recognizes himself to be now living (HV 116.2). As we have noted in
chapter 7, bits of this future history that Vyāsa foresees were helpful to us in
contextualizing some of the historical allusions in the Yuga Purānạ . The other
concerns the near-future outcome of Janamejaya’s horse sacrifice, which Vyāsa
tells Janamejaya he will be unable to counteract (115.26). In describing the Kali

34. It would seem that Janamejaya has set his mind on an Aśvamedha mindful of the dangers of a
Rājāsūya. Cf. Lakṣmaṇa’s similar advice to Rāma on this comparative topic (Rām 7.75).
35. HV 115.22–23: bhavān api ca sarveṣāṃ pūrveṣāṃ naḥ pitāmahaḥ/ atītānāgatajñaś ca nāthaś cādikaraś ca
naḥ// te kathaṃ bhagavan netrā buddhimantaś cyutā nayāt/ anātha hy aparādhyante kunetaraś ca mānavāḥ (Dutt
trans. 1897, 820, modified).
36. HV 115.24–25: kālenādya parītās te tava vatsa pitāmahāh/̣ na māṃ bhaviṣyaṃ papracchur na cāpṛṣṭo bravīmy
aham// niḥsāmarthyaṃ ca paśyāmi bhaviṣyasya nivedanam/ parihartuṃ na śaksyāmi kālaniṣṭhāṃ hi tām gatiṃ.
 dharma and bhakti 581

yuga over most of two adhyāyas, Vyāsa centers his two prophesies on a point
where they converge. The outcome of Janemejaya’s Aśvamedha will affect future
Aśvamedhas in the Kali yuga—a matter we must return to shortly. After saying
all this, Vyāsa then coyly takes leave with the words, “We will see (each other?)
again” (punar drakṣyāma ity uktvā; HV 1.118.5). Once Vyāsa leaves, kings and
Ṛsị s, including Āstīka (7), then depart after him. Āstīka’s departure makes it
clear they are leaving Takṣaśilā, where Āstīka interceded at Janamejaya’s snake
sacrifice to rescue the snakes. In the next verse, having cast off his wrath (roṣam
utsṛjya) at the dreadful snakes, Janamejaya goes back to Hāstinapura and rules
well (8), whereupon, “a short time later (kasyacit atha kālasya),” he was “cons-
crated for the horse sacrifice (dīkṣito vājimedhāya)” (11). The horse, it seems,
would in the meantime have returned to Hāstinapura, the Kuru capital.
This brings us now to the Bhaviṣya Parvan episode that—in the Pune
Critical Edition—closes both the Harivaṃśa and, in one of its most capacious
self-definitions, the Mahābhārata. The spotlight turns immediately to one
scene, and the main event all happens rather quickly:

When that one’s horse was suffocated there (saṃjñaptam aśvaṃ tatrāsya),
the queen Kāśyā Vapuṣṭamā, having approached, then lay down
according to the rite prescribed by rule (saṃviveśopagamyātha vidhidṛṣṭena
karmaṇa).̄ But Vāsava then desired (cakame) that lady of faultless limbs.
Having possessed the suffocated horse (saṃjñaptam aśvam āviśya),37 he
became mingled with her (tayā miśrībabhūva saḥ). When that transforma-
tion (vikāra) occurred, he [Janamejaya], having come to know the truth
about it, said to the Adhvaryu, “This horse of yours is unsuffocated.
Perish!38 The Adhvaryu, knowledgable as to Indra’s conduct, told the
royal sage and cursed Puraṃdara.” (HV 118.12–15)

We do not learn what the Adhvaryu’s curse of Indra entailed, since Sauti
focuses only on Janamejaya’s reaction, which shows that when he described
this king’s “casting off his wrath” at the end of his snake sacrifice, it was
premature. Uttering an act of truth based on the merits gained from his sacri-
fices, penances, and protection of his people, Janamejaya says, “Hear this
(śrūyatām idam). From this day on, Kṣatriyas will no longer offer the Aśvamedha
sacrifice to the unstable Indra, who has not conquered his senses” (16–17).
Tapping then more deeply into his anger, he faults the Adhvaryu for allowing

37. Cf. HV 118.34: aiśvaryeṇāśvam āviśya, “by his lordly power he possessed the horse.” Cf. chapter 8 § G
on the dead King Vyuṣitāśva’s possession of the horse that impregnates his widow, in a story told by Kuntī.
38. Or “fall,” for dhvaṃsa (HV 118.14d). For similar uses, see Mbh 3.178.37d; 5.17.15a, both addressed to the
Pāṇḍavas ancestor Nahuṣa who will perish and fall into becoming as a snake. Janamejaya is addressing the
Adhvaryu, not the horse (Dutt 1897, 831 has “Kill him at once!”).
582 dharma

the spoiling of the sacrifice, and tells all the priests (ṛtvijas) to leave his territory.
And getting still angrier, he goes to his women’s quarters and tells his other
wives to drive away the unchaste (asatīm) Vapuṣṭamā (18–22).
The Gandharva Viśvāvasu—a curious incercessor who perhaps knows more
than most about horses and, in any case, about heavenly nymphs or Apsarases39—
then calms Janamejaya down by telling him what was really behind all this. The
Aśvamedha was Janamejaya’s three-hundredth sacrifice, and Indra, thinking
that Janamejaya could surpass him, sought a loophole and put this obstacle in
the way. Moreover, Vapuṣṭamā was formerly the Apsaras Rambhā in a previous
life, and Indra took pleasure only with her, not with Vapuṣṭamā! Less implausi-
bly40 but perhaps more reassuringly, Viśvāvasu adds that since Indra has great
splendor and is desirous of victory, he would not violate the wives of a descen-
dant (Janamejaya is a lineal descendant of Indra’s son Arjuna). In any case,
consoling Janamejaya, as Vyāsa had earlier, that destiny is hard to withstand, he
tells him he should not blame Indra, his guru, Vapuṣṭamā, or himself;
Vapuṣṭamā is sinless (apāpām) and should be honored, and Janamejaya should
continue to enjoy her as the jewel among women that she is (HV 118.24–38).
From here, before the Bhaviṣya Parvan ends with a brief phalaśruti and a final
verse inviting further questions,41 Sauti says that Janamejaya got his mind off
his anxieties by performing a supreme pacification rite that was devoted to
dharma (śāntiṃ parām . . . dharmajuṣṭam; 39), and that with his mind on dharma
and directed at pleasing Vapuṣṭamā, he did not stop honoring Brahmins; did
not stop frequently performing sacrifices42 or looking after his realm; and did
not scold (na parigarhati) Vapuṣṭamā (39–41). But no more horse sacrifices.
Janamejaya has thus pronounced the discontinuation of the Aśvamedha by
Kṣatriyas. Yet his pronouncement would have behind it what Vyāsa told him
about this very matter during his brief prophetic visit. Indeed, Vyāsa foresaw
precisely this, and communicated it partially to Janamejaya, virtually putting the
idea of the ban into his head. Vyāsa had begun, “It is known that the Aśvamedha
is the best sacrifice for Kṣatriyas. That being its nature, Vāsava will treat your
sacrifice with indignity” (tena bhāvena te yajñaṃ vāsavo dharṣayiṣyati)” (HV
115.28). Vyāsa warned Janamejaya that even if he could withstand fate by manli-
ness, he should not perform the rite; nonetheless, neither Indra “nor your

39. On Viśvāvasu as possibly the original single Ṛgvedic Gandharva from whom the class of them derives,
and on their possibly old connections with horses and continuing ones with Apsarases, see Oberlies 2009.
40. When we were told just before this that Indra desired that lady of faultless limbs (HV 118.13), it was with
reference only to Vapuṣṭamā. There was no mention of Rambhā.
41. Sauti asks Śaunaka, “What else do you want? What do I tell you?” (kim aparam icchasi kiṃ bravīmi te;
HV 118.51).
42. Given that Janamejaya will go on performing other sacrifices in the Kali yuga, it would not have been
just the number of sacrifices that incited Indra, but the three-hundredth being an Aśvamedha.
 dharma and bhakti 583

sacrificing priest”43 will commit sin (29–32). Janamejaya asks what the sign or
occasion (nimittam) will be that his Aśvamedha is going off track, and says he
can call it off (33). Vyāsa replies that the occasion will be what Janamejaya does
to a Brahmin out of anger; he would do well to avoid it. But if Janamejaya holds
this Aśvamedha, Kṣatriyas will not offer it as long as the earth shall last (34–35).
This is a rather emphatic point, and brings out that the horse sacrifice stands
paramount among animal sacrifices, which come under critique in the
Mahābhārata, notably in the Nārāyaṇīya’s story of King Vasu Uparicara (Mbh
12.322–24), and further along when Yudhiṣṭhira’s postwar Aśvamedha is
exposed to have been worth little (see chapter 9 § D.2.a). Indeed, Yudhiṣṭhira’s
Aśvamedha now stands out as both the last one completed by a Kṣatriya and the
last one done before the onset of the Kali yuga,44 in which Janamejaya failed to
replicate it. Realizing that his fiery curse of a Brahmin will end his Aśvamedha
and that he will be the occasion, Janamejaya is deeply fearful and asks how one
like himself, devoted to good deeds, can strive for higher worlds if he is like a
bird in a noose striving to reach the sky. He asks Vyāsa to console him that there
will be “a renewal of the sacrifice (yajñasya punarāvṛtti)” (HV 115. 36–38). Vyāsa
begins his reply by predicting something obscurely technical about the mys-
teries of transmission: “A received sacrifice (upāttayajño) will remain among
gods and Brahmins—an energy handed over by an energy, it will abide in just
an energy (tejasābhyāhṛtaṃ tejas tejasyevāvatiṣṭhate)” (39). But he then predicts
something quite concrete: “Springing forth, there will be a certain army-leader
(senānī), a Kāśyapa Brahmin who will again restore the Aśvamedha in the Kali
yuga” (40), and one of his successors “like a white planet” will offer a disastrous
Rājasūya (41). It is not clear what the following verse describes, but Vyāsa seems
to be referring to the horse: “In accord with its strength, it will bestow fruits on
men performing it, and it will range about surrounded by Great Ṛsị s at the gate
of the end of the yuga” (42). Vyāsa then concludes on another obscure note that
is intelligible, however minimally, for linking the Kali yuga with its conventional
yugadharma of giving, dāna, which here as elsewhere probably denotes bhakti
religiosity: “Then a subtle dharma of great consequence, invincible, lax about the
four life-stages, will advance. Then men will attain perfection with small aus-
terity; fortunate indeed, they will do dharma at the yuga’s end, O Janamejaya.”45

43. I am uncertain here. Tava yajamānasya, “of your yajamāna,” seems to imply someone other than Janamejaya,
who would presumably be his own sacrifice’s yajamāna or patron. Dutt 1997, 821 has “presiding priests.”
44. See Koskikallio 1994, 264, 267, indicating that these points emerge also from the Jaiminibhārata—a
much later text “that most probably dates from between AD 1050 and 1250” (Koskikallio 1999, 227 n. 1).
45. HV 115.44–45: tadā sūkṣmo mahodarko dustaro dānamūlavān/ cāturāśramyaśithalo dharmaḥ
pravicaliṣyati// tadā hy alpena tapasā siddhiṃ yāsyanti mānavāh/ dhanyā dharmaṃ cariṣyanti yugāmte janamejaya.
On the yet to be fully explored relationship between dāna and bhakti in the Kali yuga, see Koskikallio 1994,
254–55, 265–66; von Steitencron 2005, 39–47.
584 dharma

If we now return to the problem of dating this Bhaviṣya Parvan prophesy

through the Yuga Purāṇa’s reference to it in rounding off its equally prophetic
Mahābhārata-Harivaṃśa digest, we cannot ignore the explanations many
would likely give for these passages. The Yuga Purāṇa’s allusioins to the
Harivaṃśa could be free floating oral folklore not yet attached to a composed
Harivaṃśa text. Or the Yuga Purāṇa verse or verses that allude to Harivaṃśa
material could be interpolated. Or the Yuga Purāṇa would have to be much
later than has been thought. But the Yuga Purāṇa’s dating is reasonable and
has looked sound to all who have examined it. As to the first two two argu-
ments, they might be plausible for references to Dantavaktra and Śatadhanvan
in YP 34–35 (see chapter 7 § A.1). But they would clearly look contrived were
they applied to the artful way verse 39 references the story of Janamejaya’s
Aśvamedha and Vapuṣṭamā.46 It is beginning to look like we must consider
the Harivaṃśa as something not as far removed in time from the Mahābhārata
as has been for a long time thought,47 and as part of the Mahābhārata project
and plan from at least the time it was reaching completion. Moreover, as
R. Tsuchida (2009) demonstrated in a presentation at the 14th World Sanskrit
Conference in Kyoto, it is quite plausible that Vyāsa’s prophesy about an army-
leader or general who becomes an Aśvamedha-performing Brahmin would
refer to Puṣyamitra Śuṅga,48 who, after overthrowing the Mauryan dynasty as
a Brahmin general in about 185 BCE, is known for renewing the Aśvamedha
by perfoming it twice (Hiltebeitel 2001a, 16; Falk 2006a, 149). That Vyāsa
should prophesy these events as coinciding with the emergence of a “subtle
dharma” of bhakti would then involve a mapping of dharma and bhakti tempo-
rally down to the very historical period and social milieu in which we have
argued that the Kali yuga concept took hold.
This brings us to our own rounding-off question for this section, which is
how mapping divine plans will tie in with mapping friendship, hospitality, and
separation. Friendship takes us to the kinds of intimacies the epics allow their
readers to have with gods who once walked the earth humanly for the “welfare

46. Mitchiner (1986, 51 and n. 98) cites Arthaśāstra 1.6.6 as referring to Janamejaya’s perishing after a
“quarrel with the Brahmans,” but the verse also mentions his violence against them and makes no reference to
the setting, or to Janameyaya’s wife. It is thus not the same story, and in any case is mentioned along with a series
of incidents from both epics (Kangle [1972] 2003, 12). But as Brodbeck mentions, the Mahābhārata knows
Vapuṣṭamā as Janamejaya’s attractive (her name means “Most Beautiful”) wife, and also several Janamejayas
(2009a, 27, 234–48). It could thus allude to this Harivaṃśa story.
47. For considerations pointing in that direction, see importantly Bhattacharyya 1956, 155–56, 161; Couture
1996, 135–36 and n. 29; 2001; Mahadevan 2010; Brodbeck 2010d.
48. Tsuchida takes up Puṣyamitra’s gotra in his article, “Some Reflections on the Chronological Problems of
the Mahābhārata,” Studies in Indian Philosophy and Buddhism 16 (2009): 1–24, which I did not find until too late to
discuss it in this work. S. Bhattacharyya 1956, 160 says the Northern Vulgate variant audbhijja, which describes this
horse sacrifice’s reviver, could be cryptically applied to the Śuṅgas, both meaning “sprout.” The Critical Edition
favors a Southern reading audbhido (115.40a), which could imply the same.
 dharma and bhakti 585

of the world.” With hospitality, we can anticipate, in Paul Younger’s terms, that
the bhakti idiom of “playing host to deity” would apply not only to temples and
festivals where deities are both guests and hosts at once (2002, 13–14) but to
texts. Into these texts, as we can now say, the supernatural “descended,” and as
its “portions” leave, their departures can be experienced in what becomes the
bhakti idiom of love in separation (viraha). Mapping bhakti and dharma in
the Sanskrit epics thus calls for a three-dimensional map that can plot vertical
and horizontal movements, temporo-spatial coordinates, and textual and geo-
graphical terrains.
I will begin with two topics that follow from earlier chapters, but require
fuller discussion. First, we need a fuller picture of the ways the two epics, and
especially the Mahābhārata, invoke the deities Dhātṛ and Vidhātṛ where uncer-
tainties arise about the relation between dharma and the divine plan. This will
lead us into a discussion of the notion that Rāma and Kṛsn ̣ ạ are incarnations of
̣ , and, with it, the Mahābhārata’s treatment of other such incarnations and
Viṣnu
their appearances in time. Finally, mapping the interrelated practices or dis-
courses of friendship, hospitality, and separation will take us into the “middle
land” of what dharma and bhakti are about in both epics: the dharma of the Ṛsị s.

B. The Placer and the Ordainer

Even though Manu resists the Mahābhārata’s bhakti swerve and the Rāmāyaṇa
streamlines it, they all agree that rules (vidhi) are divinely ordained, and that a
conventional way of saying this is to draw on the etymological link between vidhi,
“rules,” especially, but not only, Vedic injunctions drawn from sacrificial rules,49
and the two ancient Ṛgvedic abstract divinities named Dhātṛ and Vidhātṛ, the
Placer and the Ordainer.50 These two deities are singularly pertinent to the project
of mapping bhakti together with dharma, for they can be signposts of the work-
ings of dharma over time—especially at moments where there may be questions
raised and sometimes answered about a text’s divine plan. For instance, we have
seen in chapter 8 that they are introduced into the Mahābhārata as the female
personifications of the weave of Days and Nights. And as we observed in chapter
11, Draupadī and Yudhiṣṭhira’s dialogue early in their forest exile figures Dhātṛ

49. Bailey 1983, 142–43 and n. 20 notes that this “older view of vidhi as an ordinance which determines the
procedure of a rite” has numerous Mahābhārata usages “in the phrase vidhidṛṣṭena karmaṇa,̄ ‘[performed] with a rite
existing in the rules.’”
50. As Brereton’s translation of ṚV 10.167.3 suggests, the names imply an opposition: Dhātṛ as “one who
sets in place” and Vidhātṛ as “one who sets apart” (see chapter 3 § B). Translators of classical texts, however, have
favored such renderings as the Placer (or Creator) for Dhātṛ and the Disposer (or Ordainer) for Vidhātṛ.
I will continue to call them the Placer and Ordainer.
586 dharma

and Vidhātṛ as abstract deities through whom Draupadī questions the effectiveness
of dharma in the world, while both seem to be speaking about more familiar
deities under these two names to either denounce or extol them. And when
Yudhiṣṭhira refers to Draupadī’s birth as among the mysteries of the gods that is
yet to take fruit, he intimates that the Placer and Ordainer are behind a divine
plan. The pair Dhātṛ and Vidhātṛ are, however, abstractions for what would lie
before the divine plan chronologically, yet figures who can be named when char-
acters allude to its effects upon them and express their views about other forces
that are at work, like karma and fate, in producing what seems their sorry plight.
The Placer and Ordainer’s contributions are generally achronological. In speaking
at this abstract level, both Draupadī, Yudhiṣṭhira, as well as other characters,
typically stay away from any reference to more concrete manifestations of divinity
with whom they are familiar, for instance, Kṛsn ̣ ạ himself. Let us look at Dhātṛ
and Vidhātṛ more closely.
While these two deities often have what seems to be overlapping spheres
of governance, it is possible even in Manu’s three references to them to see
some latent distinctions. While Dhātṛ makes rules (vidhis) on meat in sacrifice
(M 5.30–31) and determines that Ā rya and non-Ārya are neither equal nor non-
equal (10.73), Brahmins are Vidhātṛ (11.35) for their power to regulate the varṇas.
Indeed, in the Mahābhārata, one of Bhīṣma’s postwar teachings on dharma for
times of distress (āpaddharma)—with no other mention of either Dhātṛ or
Vidhātṛ in the Āpaddharmaparvan—is that Brahmins are Vidhātṛ specifically
for just such times (12.159.18). As with mixed classes, it is as if nothing was
“ordained,” much less “set in place,” for times of distress (āpad) other than that
Brahmins are to decide what to do about them. The word vidhi as “rule” is
everywhere in Manu. For instance, the last word on the king is that he
has “eternal rules of action (karmavidhir . . . sanātanaḥ)” (M 9.325). Manu is
consonant here with the Mīmāṃsā school’s philosophy of “reflection” (mīmāṃsā)
on Vedic ritual, which maintains that “vidhis properly interpreted are the main
source of dharma. . . . [and that since] dharma can only be acquired by following
the injunctions of the Vedas they should all be interpreted as giving us
injunctions.”51
As Greg Bailey has shown, both epics have put such a nexus of rules in play;
he takes them to be similar in doing so. While relating Dhātṛ and Vidhātṛ pri-
marily to Brahmā in both epics, Bailey does not differentiate the two deities, and
stresses Vidhātṛ for his etymological link with vidhi (1983, 141–42). As he indi-
cates, in the epics vidhi, “rule,” can also be a word for fate equivalent to daiva as
well as a name for Brahmā. Although such a dharma nexus comes into being

51. See Bailey 1983, 152, who is quoting Dasgupta [1922] 1975, 5: 1183–84 here. See also Bailey 1983, 141–43.
 dharma and bhakti 587

with the Veda, it is not exactly created, and is not to be confused with the
personified god Dharma whose early cosmogonic birth the Mahābhārata
accounts for genealogically (see chapter 6 § A.3). As Bailey observes, the epics
and Purānạ s say relatively little about the creation of dharma or its equivalent
here, the Veda, in their primary and secondary cosmogonies (dharma is, after
all, by their understanding “eternal”). But a creation, or better recreation or
“reissuing,” of dharma does occur in so far as Brahmā (or his sometimes surro-
gates Vidhātṛ and Dhātṛ) fits the eternal dharma’s supposedly Vedic vidhis to
activities (karma) that should regulate the “spatio-temporal world of dharma and
adharma,” that is, the triple world (see Bailey 1983, 139–40). For humans, such
rules, which have their model in sacrifice, apply primarily to the four varṇa cat-
egories that were created when the gods divided (vyadadhuḥ) Puruṣa and laid the
first dhármans or “foundations” in ṚV 10.90.11–16 (see chapter 3 § B).
In the Rāmāyaṇa, where Bailey f inds Rāma to be totally caught up in
sorting out “rules,” Bailey thus shows that the few references to Dhātṛ and
Vidhātṛ probably refer to Brahmā,52 as is also the case in Manu where the
pattern of instituting Brahmā’s vidhis is obvious. But Brahmā does more
“ordaining” in the Rāmāyaṇa than in the Mahābhārata. As Bailey indicates
(1983, 143–49), Brahmā plots much of Rāma’s story.53 Yet it is not just
Brahmā who shapes what is ordained; Brahmā puts the story into Vālmīki’s
hands to make it a poem (kāvya; Rām 1.2.22–40; see chapter 5). Only the
Vālmīki Rāmāyaṇa tells us that Sītā’s banishment was “appointed by Dhātṛ”
(7.47.33)—as if Vākmīki leaves a space between Brahmā (taking Dhātṛ as
Brahmā) and the poet.
In the Mahābhārata, however, where Dhātṛ and Vidhātṛ appear in 140
usages,54 Bailey is less successful in posing a consistent link between Brahmā
and the pair. Rather, with this epic’s bhakti swerve, who these two deities are
and what they “place” and “ordain” is more a matter to ponder—like dharma
itself, with its rules or injunctions and consequent dilemmas. What distinc-
tions there are between Dhātṛ and Vidhātṛ are subordinate to their comple-
mentarity and the sense that their rapport can be processual, with the Placer

52. The Rāmāyaṇa has five occurrences of Dhātṛ alone; zero of Vidhātṛ alone; and three of them together.
53. I agree with Bailey’s overall analysis, but believe his attribution of “absolute determinism” to Brahmā’s
ordinances is not well considered. He overextends the implications of vidhi as “rule” to include ways that Brahmā
“guides the action by means of timely ordinances, curses, and the giving of boons” (147). And where the
Rāmopākhyāna mentions vidhi at the point of Sītā’s fascination with the golden deer (Mbh 3.262.17) that Rāma
will chase for her, enabling her abduction, Bailey treats this Mbh subtale as if it were generalized Rāmāyaṇa. This
overlooks the ways that Vālmīki introduces different (and I believe new) nuances and complications by the play
he gives the poet as an intervening character. See Hiltebeitel 2009a.
54. Dhātṛ is mentioned alone seventy-seven times; Vidhātṛ twenty-seven times; and seventeen passages
mention them together. Counting a powerful single usage of Saṃvidhātṛ and a usage that includes Saṃdhātṛ
among the thousand names of Viṣṇu (13.135.35b), the Mbh would have a total of 140 for both names.
588 dharma

always having priority.55 Thus in one of the few Mahābhārata passages that
delineates their roles one after another,56 it is by honoring the ancient rule
(vidhi) “ordained by the Placer (dhātrā . . . vihitaḥ)” that the Seven Seers
shine in the sky and elephants stand tall as mountains, whereas all beings do
what has been “ordained by the Ordainer (vihitaṃ vidhātrā) according to
their own kind.”57 Otherwise, since most passages that mention the pair
together suggest a joint operation, one must look to separate depictions to
follow up suggestions of contrast. Cosmologically, certain passages give the
impression that Dhātṛ takes care of large-scale work. For instance, once
beings are linked with their dharma or adharma, injuriousness or noninjuri-
ousness, etc., at each new creation, Dhātṛ thinks (manyate) variety into the
great elements, sense-objects, and their forms (mūrtiṣu), and ordains
(vidadhāti) the apportionment or distribution (viniyogam) of beings (Mbh
12.224.48–49).58 Other passages then imply that Vidhātṛ sees to details over
time. Thus everything “ordained by the Ordainer for man,” notably goals or
riches (arthān), “is acquired in Time” (Mbh 12.26.25–26). On this basis,
Arjuna can tell Yudhiṣṭhira that the daṇḍa was “ordained by the Ordainer”
to prevent confusion among the four classes and to preserve dharma and
artha; if, for example, the Rod did not rule, a girl would not marry . . .
(121,535–37)! The Mahābhārata also suggests contrasting usages for the
individual: whereas Dhātṛ seems to have the most to do with setting things
(seed, svakarma, svadharma, fate) in place for the beginning, end, and thus
the whole of a lifetime, Vidhātṛ may have more to do with the specific
working out of fate and karma at junctures within a lifetime. Thus, whereas
Dhṛtarāṣṭra can lament that he must have done something wrong in earlier
births since Dhātṛ has joined him to such wretched deeds in this one

55. See Monier-Williams [1899] 1964, 552 on Niyati (Destiny) and Āyati (the Long Run, Posterity) as
daughters of Meru and Dhātṛ and Vidhātṛ’s Purāṇic wives.
56. See also Mbh 12.251.25: after Dhātṛ ordained wealth to be used not only for oneself but for others,
Vidhātṛ anciently ordained “holding the world together” (lokasaṃgraha; cf. BhG 3.20), apparently implying each
class’s contribution to the social good. Also potentially contrastive are separate mentions in the Viṣṇusahasranāma
(The Thousand Names of Viṣṇu): Dhātṛ among names suggestive of Viṣṇu’s creative role in relation to Brahmā:
“The Fundamental Sustainer, Dhātṛ, He Who Smiles (or Shines) Like an Opening Flower [at the Creation?], Ever-
Awakened” (ādhāranilayo dhātā puṣpahāsaḥ prajāgaraḥ; 13.135.114); Vidhātṛ among names suggestive of the sun’s
rays and the accomplishment of activities: “Thousand-Rayed, Vidhātṛ, the One Marked with Accomplishments
(sahasraṃśur vidhātā kṛtalakṣaṇaḥ; 13.135.64); cf. Chinmayananda 1993, 238, 133–34.
57. Mbh 3.26.13–15. I flag the double use of vihita, from vi-√dhā, with the instrumentals of both names. It
phonetically accompanies Vidhātṛ, with whom it occurs in thirteen of the Mahābhārata’s twenty-seven mentions
of him, more naturally than Dhātṛ, with whom it occurs in seven of seventy-seven mentions. Parallel construc-
tions with the more versatile Dhātṛ use sṛṣṭa (created or issued, 9.30.34; 12.27.32, 66.20; 13.14.58; 129.2); diṣṭa
(appointed or fated, 2.51.25, 52.14; 5.71.4); ādiṣṭa (determined or commanded, 9.58.9; 12.20.10); nirdiṣṭa (assigned
or appointed, 9.64.22); saṃkalpita (determined or intended, 3.20.24); codita (impelled, 5.50.27); and kṛta (made,
14.79.15). Cf. also nirmita (fashioned, meted out) at Rām 1.14.18.
58. Cf. Bailey 1983, 140 on Dhātṛ’s heavy-work in Mārkaṇḍeya Puraṇā 48.39–44.
 dharma and bhakti 589

(11.1.18),59 Saṃjaya can tell him, “Who can ward off fate (daivam)? No one
steps beyond the path ordained by the Ordainer (vidhātṛvihitam mārgam)”
(1.1.186c–187b).60 From these mostly latent contrasts, many of the ideas that
might be communicated by epic characters mentioning Dhātṛ and Vidhātṛ
have emerged. I believe they can be broken down into five clearly interre-
lated clusters over which we must not linger. These five are seed-placing,
which we might in some instances call genetic engineering;61 character
formation, sometimes implying the work of the author;62 karma, sometimes
in conjunction with dharma and svadharma;63 fate;64 and the food chain.65 It
is easy to see how dharma can intertwine all of these clusters with rules that
can provoke dilemmas.

C. “Avatāra”

The Bhagavad Gītā̄ does not use the term avatāra, and the Mahābhārata uses it
quite inconspicuously only once.66 Yet it provides early accounts of all but one

59. Cf. Mbh 3.199.14–16, where Dhātṛ’s “rule” (vidhi) covers the lifelong svakarma and svadharma of a
pious Śūdra meatseller; 5.173.5–6, where Ambā lists the Placer last (after Bhīṣma, her father, herself, and her
once-betrothed Śālva) among those by whose folly she came to her plight; 6.108.18, where according to Droṇa,
Ambā’s reincarnation as Śikhaṇḍin was “ordained (vihita) by the Placer” to have been that of a woman, whereas
fate (daiva) made him a man.
60. Cf. 1.99.29, where Satyavatī tells Vyāsa he is her “firstborn son, ordained by the Ordainer” (1.99.29).
Also “ordained by the Ordainer” are Draupadī’s winsome beauty (1.182.13) and rules of penance for Brahmins
who wish to right what is done wrong (5.28.5). As noted, however, Sītā’s banishment was appointed (pradṛśyate)
by Dhātṛ (Rām 7.47.33).
61. See Mbh 3.31.21 and 37 (Draupadī says the Placer sets the course even before the ejaculation of seed;
12.219.112 (one repeatedly must reside in wombs where one is placed by Dhātṛ). Or, a disembodied heavenly
voice can settle who fathered Śakuntalā’s son Bharata by saying to King Duṣyanta, “You are the Placer (dhātṛ) of
this embryo, Śakuntalā has spoken the truth” (1.69.30 = 1.90.32). For related verbal uses of √dhā with seeding the
womb, see 1.85.12ab; 1.99.4; and 1.99.41.
62. See 12.27.32: tiring of Yudhiṣṭhira’s grief, Vyāsa says, “As the Placer has created you for deeds, do
them. Your perfection will come from that. You are not your own master all by yourself, king.” The passage
(26–32) hints at Vyāsa’s authorship. Several characters are said to have been created by the Placer, notably
Śikhaṇḍin: says Droṇa, Bhīṣma will not fight him because he was “ordained a woman by the Placer, and by fate
(daiva) again became a man” (6.108.18)—a distinction already mentioned suggesting that Bhīṣma holds to the
created order. Cf. 1.99.29 on Vyāsa; 1.182.13 on Draupadī; 9.5.12 on Aśvatthāman.
63. See chapter 10 § D for Draupadī’s closing statement on Dhātṛ and karma to Yudhiṣṭhira (3.33.3–35), after
which Bhīma intervenes mentioning Dhātṛ in connection with karma, dharma, and svadharma (3.34.52–54).
64. According to Dhṛtarāṣṭra (2.51.25) and Yudhiṣṭhira (2.52.14), the dice match is fated or appointed
(diṣṭa) by the Placer. Yudhiṣṭhira says Dhṛtarāṣṭra cannot abandon Duryodhana whom “Dhātṛ has made subject
to fate” (5.39.1). And Dhṛtarāṣṭra says his sons may as well be fighting the Wind, “fatally (niyatam) impelled by
Dhātṛ, as game by a lion” (5.50.27).
65. Dhātṛ ordained the food for all creatures on earth to live on, protected by their own acts (svakarmabhiḥ
Mbh 12.277.18). Snakes are the food ordained for Garuḍa by Vidhātṛ (1.14.23) or by Dhātṛ (5.103.4).
66. I thank Couture for correcting my impression that it does not occur at all. The instance refers to the
ridges of Mount Gandhamādana, and “means ‘a new descent of beauty, or a new way of making beauty visible on
earth’” (Couture 2001, 314, citing Mbh 3.146.33).
590 dharma

of the figures later named in Purāṇic lists of the ten avatars—the omitted one,
not surprisingly, being the Buddha. In telling these stories, the Mahābhārata
does not typically speak of some of these figures as incarnations of Viṣṇu: most
notably Rāma Jāmadagnya, also called Bhārgava Rāma (and later, as an avatar,
Paraśurāma), and the Tortoise and Fish (the latter is still identified, as in the
Śatapatha Brāhmaṇa, with Prajāpati; Mbh 3.185.48).67 But when the Pāṇḍavas
meet the two Ṛsị s Lomaśa and Mārkaṇdẹ ya during their forest sojourn of Book 3,
these sages tell of five of these figures in varied incarnational terms: the Boar,
Man-Lion, and Dwarf as “forms” (rūpa) or “bodies” (vapus) of Viṣṇu
(3.100.19–21); Rāma as “the Vaiṣṇava hero” (taṃ vaiṣṇavaṃ śūram; 3.275.65) in
the Rāmopākhyāna;68 and Kalki as the “fame (yaśas) of Viṣṇu,” or Viṣṇuyaśas.69
With Kṛṣṇa making six, this group seems to have been introduced as a sort of
incarnational nucleus. And Rāma Jāmadagnya is mentioned as one of a cluster
of “manifestations” (prādhurbhāvas) of Viṣṇu-Nārāyaṇa in the Nārāyaṇīya
(12.326.77).70 There Nārāyaṇa tells Nārada about the manifestations he promised
Brahmā he would make down through the yugas on behalf of the earth, the
triple world, and the gods (including above all Brahmā himself, the creator) in
this order: Boar, Man-Lion, Dwarf (not named but narratively indexed), Rāma
Jāmadagnya, Rāma Dāśarathi (Rāma of the Rāmāyaṇa), and Kṛṣṇa.71 The avatar
mythology is thus more than just nascent in the Mahābhārata,72 though it is yet
to be shaped into the later formulations that require for some avatars, such as
the Fish, Tortoise, and Boar, that new versions of their myths provide demons
so that they can have someone wicked to conquer (see Biardeau 1989a, 103).
The wicked are only generalized at Bhagavad Gītā 4.7–8. But the epic tunes this
theme more finely in Mārkaṇḍeya’s highly featured rendition of the Kalki myth,
which may be the Mahābhārata’s most explicit formulation of the “just war”
theology one finds in the Bhagavad Gītā̄—and for Yudhiṣṭhira’s ears, primarily,
rather than Arjuna’s. As was discussed in chapter 7, the Kalki myth in the

67. See González-Reimann 2006b; Soifer 1991, 32 on the identification of Prajāpati also as boar and tor-
toise in the Brāhmaṇas.
68. In the Rāmopākhyāna; see also 3.260.5–7: earlier in this subtale, Brahmā says, “For that purpose the
four-armed Viṣṇu has descended (avatīrṇo. . viṣṇuḥ) at my command,” as discussed above in the text.
69. van Buitenen 1975, 597 takes viṣṇuyaśas (Mbh 3.188.89) as another name of Kalki; Ganguli [1884–96]
1970, 3: 412 as “he will glorify Vishnu.”
70. See chapter 6 § B. As Hacker (1960, 50–52) and Couture (2001, 313) observe, this term seems to pre-
cede avatāra in referring to Viṣṇu’s incarnations.
71. Mbh 12.326.61, 72–95, 327.85. Also among these “manifestations” are the Gander (Haṃsa) and Horse’s
head (Hayaśiras). In an interpolation found in mainly Telugu manuscripts (12.835* lines 3–5), after 12.326.76,
Nārāyaṇa also tells Nārada a classical Purāṇic list of ten prādurbhāvas: Fish, Tortoise, Boar, Man-Lion, Dwarf,
Rāma, Rāma, Rāma, Kṛṣṇa, and Kalki. Note that with the cutely named three “Rāmas” [Jāmadagnya, Dāśarathi,
and Balarāma], the list continues to have no place for the Buddha (see Oberlies 1997, 130).
72. See Pollock 1991, 38 n. 71 on “paradigmatic reference in the old battle books” of the Mahābhārata to
the Man-Lion myth, citing 7.164.146, 168.21; cf. 3.100.20.
 dharma and bhakti 591

Mahābhārata offers a different outcome from Vyāsa’s prophesy at the end of

the Harivaṃśa, where the turn to the Kṛta yuga will occur without Viṣṇu mak-
ing an intervention. Although the Harivaṃśa does know Kalki Viṣṇuyaśas for
brief (at least in the Critical Edition) mention among Viṣṇu’s manifestations
(HV 31.148), its incarnational focus is more on Kṛṣṇa.
Once this Kalki myth is seen in its Mahābhārata context, it is clear that it
carries along the epic’s incarnational scheme. While the Pāṇḍavas are in the
forest with Kṛṣṇa visiting, and listening to Mārkaṇḍeya’s tales, this antediluvian
Ṛṣi recounts that while he was surviving the dissolution of the universe
swimming alone on the endless cosmic waters, he saw a child sleeping on a leaf
of a banyan tree. Mārkaṇdẹ ya learns from this babe that, after he had long wan-
dered in the worldly byways of the child’s cosmic body, he had issued from its
mouth into the waters. The child is none other than Nārāyaṇa, now awakened
from the form in which he sleeps through the pralaya. The child tells Mārkaṇdẹ ya
of his greatness as Nārāyaṇa, starting out (all but the vocative) with the same
words as Bhagavad Gītā 4.7, and for the only other time in the epic:

Whenever, sage, the Law languishes and Unlaw rears up (yadā yadā
dharmasya glānir . . . abhyutthānam adharmasya), I create myself
(ātmānaṃ sṛjāmi). When Daityas bent on harm spring up invincible
to the chiefs of the Gods, and terrifying Rākṣasas, then I take on birth
in the dwellings of the virtuous and, entering a human body (praviṣṭo
mānuṣaṃ deham), I appease (praśamyāmi) it all. (3.187.26–28; van
Buitenen 1975, 592)

Mārkaṇdẹ ya then further reveals that this child is none other Kṛsn ̣ ạ in the
Pānḍ ạ vas and Draupadī’s very company, “your ally Janārdana . . . who sits here
as though at play, . . . the Placer (Dhātṛ) and Ordainer (Vidhātṛ) and Destroyer
(Saṃhartṛ), this Govinda, . . . the unborn God of the beginning, Viṣnu ̣ the Person
of the yellow robe,” and that the five brothers and Draupadī73 should “go to him
for refuge, he will grant it” (187.50–53). It is just after the Pānḍ ạ vas receive com-
fort from Kṛsn ̣ ạ (188.2) that Yudhiṣṭhira is inspired to ask Mārkaṇdẹ ya what will
happen “at the destruction of the yuga (yugakṣaye)” (6).74
As can be recalled from chapter 7, once Mārkaṇḍeya dispenses with the
first three yugas using imperfect or present tense verbs (3.188.10 ff.), he
describes the Kali yuga prophetically using future tense ones (15 ff.) to tell how
Kalki, “the Fame of Viṣṇu, a Brahmin prodded by Time” (89ab), will come

73. She bows with the rest to Kṛṣṇa in the next verse (3.188.1).
74. On these passages, see further Hiltebeitel 2005b, 124–27; on similar usages of yugakṣaye in the Yuga
Purāṇa, see chapter 7 §§ A.1 and A.3, with the difference that that text does not lead on to a next yuga.
592 dharma

under conditions of faltering dharma. As with the Bhagavad Gītā, these

conditions begin with the debasement of class roles and mixed marriages
(14–18); but they continue rather differently:

The Brahmins shall find fault with the Veda and abandon their vows;
seduced75 by argumentation/logic (hetuvādavilobhitāḥ),76 they will
neither offer worship nor sacrifice. . . . The entire world will be
barbarized (mlecchabhūtam),77 without rites and sacrifices, without
joy, and also without festivals. (3.188.26, 29)

This may recall the “repressive” centralization that marked imperial policies of
the Mauryas, among whom Aśoka explicitly sought to neutralize popular
religious assemblies.78 Further along, we find the land dotted with eḍūkas,
which I believe Hopkins and Biardeau79 are right to have identified, at least in
this context, as referencing Buddhist reliquaries or stūpas:

The world will be totally upside down: people will abandon the
gods and offer pūjā to eḍūkas; Śūdras will refuse to serve the
twiceborn at the collapse of the yuga. In the hermitages of the great
Ṛṣis, in the settlements of Brahmins, at the gods’ temples
(devasthāneṣu), in the Caitya sanctuaries, and in the abodes of
Nāgas, the Earth will be marked by eḍūkas and not adorned by
houses of the gods (eḍūkacihnā pṛthivī na devagṛhabhūṣitā). At the
expiration of the yuga, that will be the mark of the yuga’s end.
When men become ever-gruesome dharma-lacking meat-eaters and
liquor-drinkers, the yuga will collapse. . . . Then the earth will soon
be overrun by barbarians (mlecchas) while Brahmins, out of fear of
the tax burden (karabhārabhayād), will flee to the ten directions.
(64–67, 70)

Still more horrors are to follow (including the appearance of six suns; 188.7580),
until, says Mārkaṇḍeya, the Kṛta yuga will begin anew (87):

75. Where the Critical Edition has vilobhitāḥ here, the Vulgate has vimohitāḥ. “deceived.”
76. Biardeau glosses this as “led into error by discussions that pretend to be logical,” implying conversa-
tions especially with Buddhists that would take place outside the ritual arena (2002, 2: 760; cf. 778).
77. Or “mlecchified” (Ganguli [1884–96] 1970, 3: 409). Mārkaṇḍeya repeats this phrase; see 3.188.37a
(variant) and 45a.
78. See chapter 2 on Minor Rock Edict 1 and Aśoka’s prohibition of samājas.
79. See Hopkins ([1901] 1969, 399; cf. 391, 475); Biardeau (2002, 2: 759–60; see also Bronkhorst 2007, 5–6,
359–60: “The term eḍuka (Buddhist Sanskrit eluka) refers no doubt to stūpas, but our passage does not tell us whether
Buddhist, Jaina, or Ajīvika stūpas are meant” (5 n. 14). As mentioned in chapter 6, Bronkorst seems to defer to
González-Reimann 2002, 95 ff. on a late dating for the passage, which I see no reason to accept.
80. As Biardeau 1994, 23 observes, it is clearly the end of a Kali yuga preceding a Kṛta yuga, but the sudden
appearance of the six suns is clearly a “borrowing” (“emprunt”) from a kalpa-ending pralaya.
 dharma and bhakti 593

A Brahmin by the name of Kalki Viṣnu ̣ yaśas will arise, prodded by

Time (kālapracoditaḥ), of great prowess, wisdom, and might. He will
be born in the village of Sambhala in an auspicious Brahmin
dwelling, and at his mere thought all vehicles, weapons, warriors,
arms, and coats of mail will wait on him. He will be king, a Turner of
the Wheel (cakravartī), victorious by dharma (dharmavijayī), and he
will bring this turbulent world to tranquility. That rising Brahmin,
blazing, ending the destruction, noble minded, will be the destruction
of all and the one who makes the yuga turn. Surrounded by Brahmins,
that Brahmin will extirpate the lowly barbarian hosts (kṣudrān . . .
mlecchagaṇan ̄ ) wherever they are.81 After destroying the robbers he
will ritually make over (kalpayiṣyati) this earth to the twiceborn at a
great sacrifice (mahāyajñe), the horse sacrifice. He will reestablish the
auspicious limits that the Self-Existent has ordained. And when he
has grown old in works of holy fame, he will retire to the forest.
People who live in the world will follow his morality (śīlam). And with
the thieves (cora) destroyed by the Brahmins, safety will prevail.
Establishing black antelope skins, spears, tridents, and emblematic
arms in the conquered territories (deśeṣu vijiteṣu), that tiger-like
Brahmin Kalki, praised by the chief Brahmins and honoring their
leaders, shall walk the earth forever bent upon slaughter of the
robbers (dasyus). The robbers will wail piteously, “Ah father, Ah son!”
as he leads them to destruction. Adharma will decline and dharma
increase, Bhārata, and the people will observe the rites when the Kṛta
age arrives. Ārāmas (resting-places), caityas (sanctuaries), temple
tanks, wells, and the many ceremonies (kriyāś ca vividhā) will reappear
in the Kṛta yuga. Brahmins will be strict (sādhavaḥ), Munis will do
tapas, hermitages with heretics (āśramāḥ sahapāṣaṇḍāh)̣ will be firm in
truth; people will be subjects. . . . (3.188.89–189.9)

Epic rarities like Buddhist shrines and heretical hermitages,82 a world restored
to the great Ṛṣis and Brahmanical temples,83 Brahmins fleeing tax burdens
imposed by barbarian kings, Brahmin fighters regaining “conquered terri-
tories” apparently without Kṣatriya help and redharmifying them with antelope
skins: this heady “futurist” mix is adroit in admitting conditions that are not

81. An adhyāya break occurs here, resumed with “Vaiśaṃpāyana said” as the next adhyāya begins.
82. It would seem that the āśramas that give residence to heretics will cease to do so. It is interesting that this
Brahmanical term may be used to describe post-Aśokan Buddhist vihāras (on which see Schopen 2004, 76–77).
83. This prophesy provides not only the sole instances of eḍūka in the Mahābhārata but those as well of
devasthāna and devagṛha in the sense of “temple” (to which the Rāmāyaṇa adds a devasthāna only at 2.94.3).
594 dharma

otherwise given narration in the archaicized “epic ages” of either epic. It can be
taken as ex eventu prophesy based on conditions familiar from the time of
the passage’s author, and still familiar as well to the authors of Manu84 and the
Yuga Purāṇa. But even without Kṣatriyas being mentioned to this point, the
myth goes on to charter the reestablishment of the same societal roles and
classes as the Bhagavad Gītā: Brahmins will be devoted to soft recitation of the
Veda (japa) and sacrifices; they will love dharma (dharmakāmāḥ) and be dedi-
cated to their six occupations (ṣaṣkarmaniratā viprāḥ). Kings will rule this earth
by the Law and delight in protection. Vaiśyas will be dedicated to practical
affairs (vyavahāraratāḥ). Śūdras will take delight in what they hear from
(i.e., they will be obedient to) the three classes (śuśrūṣāyāṃ ratāḥ śūdrās tathā
varṇatrayasya ca). The Cakravartin Kalki has been “victorious by dharma”; “this
dharma” will last through the first three yugas (11–13), with its complete carry-
over into a Kṛta Yuga and beyond to a point in the next mahāyuga cycle. The
Kalki myth does for the ideal Brahmin what the Bhagavad Gītā does for the
ideal Kṣatriya, and in a future that confirms the standards of “just war” for both
classes: indeed for all ages and all varṇas. It translates the Gītā’s concept of “just
war” from kṣatriya svadharma to a version of Rājadharma regenerated by
Brahmins,85 and has the ears of the Brahmin-oriented Yudhiṣṭhira with the
battle of Kurukṣetra still to come.
Indeed, rather than it being an impediment to dating the composition of
the Mahābhārata to the Śuṅga period or a time shortly after it because the epic
would oppose the Śuṅgas for being Brahmins rather than Kṣatriyas, the Kalki
myth would legitimize Śuṅga rule—in theory—as a desirable interim mea-
sure.86 It is important to recognize that, in theory, varṇa is nongenealogical.
Beginning right from the Puruṣa Sūkta, the four orders are born, with the
Brahmin first, not as lineages but as categories. As a Brahmin, the Cakravartin
Kalki reestablishes the four orders. Indeed, so does Kaśyapa after the decima-
tion of the Kṣatriyas by Rāma Jāmadagnya (see chapter 7 § A.3).
But we are not done with the latency of the avatar doctrine in the
Mahābhārata. This epic’s uses of the verb ava-√tṝ and derivatives also point
toward incipient developments of the concept itself. Here we begin to

84. Both the Mahābhārata’s postwar anthologies and Manu agree that Brahmin Veda scholars (śrotriyas)
should be exempt from taxation (Mbh 12.77.7–9; cf. 97.21; M 7.133; 8.394) and that Brahmins have the right to
take up kṣatradharma in times of distress (Mbh 12.79.1; M 10.81).
85. So too the Rāma Jāmadagnya story, in its repeated killings of the Kṣatriyas at Kurukṣetra, presents a
Brahmin who can be treated as a king. See Mbh 7 Appendix I, No. 8, listing him among the “Sixteen Kings” men-
tioned (at line 853, he brought all the eighteen islands under subjugation). The CE considers this unit an interpo-
lation because Śāradā and Kaśmīrī manuscripts omit it, accepting the version at 12.29 which has fourteen of the
same kings, but not Rāma Jāmadagnya. See Hiltebeitel [1976] 1990, 346 n. 28.
86. This departs from the position I took in Hiltebeitel 2001a, 17, which is also taken by Fitzgerald 2004a, 122.
 dharma and bhakti 595

explore how the metaphor of “descent,” which, as we have noted, Couture

has developed in relation to the descent of actors onto a stage, does mix
with the theme of genealogy. Here the Mahābhārata provides an instance
that links the two epics. In the Mahābhārata’s Rāmopākhyāna, at the
moment of Rāma’s conception, Brahmā tells the gods and Ṛṣis how Rāvaṇa
will be killed: “For that purpose the four-armed Viṣṇu has descended
(avatīrṇo . . . viṣṇuḥ) at my command” (Mbh 3.260.5)—whereupon Brahmā
goes on to direct the hosts of gods to take birth on earth as “Viṣṇu’s com-
panions (viṣṇoḥ sahāyān)” (3.260.6–7), that is, as the monkeys and bears
who will be his allies.87
Among terms derived from avaÖtṝ, however, the one with the greatest
genealogical range in the epics and the Harivaṃśa is avataraṇa. Couture cen-
ters discussion on two usages. He observes that both the Mahābhārata and the
Harivaṃśa use avataraṇa prominently with reference to the aṃśāvataraṇa or
“descent in portions” of deities and other supernaturals, and remarks that in
the Harivaṃśa, examples of Viṣṇu’s avataraṇa are also “easily found.”88 Viṣṇu
Purāṇa 5.1.2 then “replaces” aṃśa-avataraṇa with aṃśa-avatāra with respect to
Viṣṇu’s birth as Kṛṣṇa in the family of Yadu (2001, 314).
Couture also touches on another usage of avataraṇa in the Mahābhārata,
that of Gaṅgā (gaṅgāvataraṇa), this time without any such usage in the
Harivaṃśa’s Critical Edition, although there is an interpolated one in the
Harivaṃśa Vulgate.89 Indeed, why should the Harivaṃśa have a gaṅgāvataraṇa.
Mathurā is, after all, on the Yamunā River, not the Gaṅgā. In this case, how-
ever, the term gaṅgāvataraṇa is used most prominently in the Rāmāyaṇa
(Rām 1.41.6, 43.13). There, Rāma hears from Viśvāmitra how his Ikṣvāku
ancestor Bhagīratha, by his tapas, brought down the celestial Gaṅgā from
the sky (gaganāt; 42.7), her “mighty fall” (gaṅgāpatanamuttamam; 42.10)
broken by passing through the coils of Śiva’s matted hair, so that her waters
would cleanse the ashes of “all his grandfathers” (43.7), the 60,000 sons of
Sagara, who had dug up the earth looking for their father’s stolen sacrificial
horse and been reduced to ashes by the sage Kapila when they finally found
it. The Gaṅgā will descend both to cleanse the ashes of Sagara’s sons and to
fill the great excavations they made in the earth—the latter, as foreseen by
Brahmā, is all part of a former divine plan, about which he reassures the

87. The same term (sahāya) is used by Vaiśaṃpāyana in the Rāmopākhyāna’s “frame” to describe the monkeys,
bears, Pānḍ ạ vas, and allies of Indra (276.5–10—four times).
88. Couture 2001, 314, 322 cites HV 44.82, addressed by Nārada to Viṣṇu, as one instance: “At the
moment of your descent (or entrance) [on the stage at Mathurā), Kaṃsa will perish (tavāvataraṇe . . . kaṃsaḥ sa
vinaśiṣyati).”
89. HV Vlg. 2.93.24–27; Couture 2001, 324 n. 8.
596 dharma

gods (1.38.23–39.4). Yet there is a third and still more basic reason for this
descent of the Gaṅgā, which Viśvāmitra suggests at the very beginning of his story:

Gently smiling, Viśvāmitra once more addressed Kākutstha, “Rāma,

you shall now hear the whole story of great Sagara. Himavat,
greatest of mountains and father-in-law of Śaṃkara, stands con-
fronting the Vindhya range. The two in fact directly face each other.
Best of men, tiger among men, the sacrifice took place between
these two ranges, for it is said that this is the best region (deśa) for
sacrifices.” (Rām 1.38.3–5; Goldman 1984, 197–98 trans.)

That is an expansive version of the “region best for sacrifices,” which is obviously
meant to include Kosala, where the Gaṅgā flows by the kingdom of the Ikṣvākus.
King Sagara’s sacrifice began in that land before it was sanctified by the river.
Couture takes brief note that the Mahābhārata knows the term gaṅgāvataraṇa
(3.106.38; 3.108.13) along with this basic story, but the ways it contrasts with the
Rāmāyaṇa are informative. First, it fits it into the story of Agastya (3.104–8). In
the Agastya Upākhyāna, the sons of Sagara do not go looking for the horse in
their massive diggings into the earth; they go to the ocean’s floor after Agastya
has drunk up the ocean. The Mahābhārata thus provides a different background
for Gaṅgā’s descent. Agastya does not figure in the Rāmāyaṇa’s account; the sons
of Sagara do not dig up the earth in the Mahābhārata’s. In the same note, Couture
also cites another Mahābhārata usage, actually of gaṅgāvatīrṇa (13.27.87), which
occurs in a long recitation of Gaṅgā’s merits by her son Bhīṣma after Yudhiṣṭhira
has asked him where one finds the most meritorious regions (deśas), countries
(janapadas), hermitages, mountains, and rivers—to which Bhīṣma replies at
length on the misfortune of those who live in countries not watered by the Gaṅgā
(13.27.18, 25–105). From the standpoint of both location and genealogy, then, the
gaṅgāvataraṇa is more the Rāmāyaṇa’s story than the Mahābhārata’s. But the
person of Bhīṣma Gānġ eya reminds us that the Mahābhārata has done something
similar to the Rāmāyaṇa in bringing Mother Gaṅgā down to earth. It has done it
upriver in the hallowed land of the Kurus. As we have seen in chapter 8, which is
where the divine plan seems to enter the itihāsa of the Bhāratas after Gaṅgā tells
King Pratīpa,

So be it, dharma-knower. May I unite with your son. So by devotion to

you will I love the famous Bhārata lineage. Whoever are the kings of the
earth, you[r dynasty is] their refuge. I am unable to speak the qualities
that are renowned of your lineage in even a hundred years. (1.92.12–13)

Although this descent by Gaṅgā is not explicitly referred to as gaṅgāvataraṇa, it

is, to put it simply, recounted within, and as a defining moment in, the
Mahābhārata’s aṃśāvataraṇa.
 dharma and bhakti 597

We can thus see that in both epics, Gaṅgā intervenes in their dynastic
genealogies and descends into their central lands. The Rāmāyaṇa tells of her
origins as an earthly river; the Mahābhārata about her additional descent in
person. In the Mahābhārata, however, her descent into Kuru country, and into
dramas of the Kuru court, capital, and lands, provides just one of the arresting
usages of the term avataraṇa. Here, I believe that Couture’s attention to the
theatrical resonances of the term have brought him up short of seeing a major
thread in the Mahābhārata’s usages. But I must add that it is my own fieldwork
on Tamil Mahābhārata dramas performed at Draupadī and Dharmarāja tem-
ples and festivals that brings me to go this extra mile. Draupadī cult terukkūttus
not only stage the Mahābhārata. They stage it, even though the actors speak
Tamil in Tamil villages, at such places as Vraja if there is an opening play about
Kṛṣṇa; at the Pañcāla capital for Draupadī’s svayaṃvara; from there on, once
the Pāṇḍavas bring her home, at the Pāṇḍava and Kuru courts at Indraprastha
and Hāstinapura; and finally at Kurukṣetra.90 With that in mind, I think we
should read another Mahābhārata instance of avataraṇa both theatrically, as
Couture does,91 and from this additional geographical standpoint.
The term makes a striking double appearance when Kṛṣṇa, having decided
to make one last effort to avoid war, determines to go as the Pāṇḍavas last
ambassador to the Kuru court. Both Yudhiṣṭhira and Vidura, the son and incar-
nation of Dharma, respectively, express reservations as to whether Kṛṣṇa should
“descend” into what, genealogically, amounts to the epic’s camp of demons.
First, before Kṛṣṇa departs, he hears from Yudhiṣṭhira,

I do not agree, Kṛṣṇa, that you should go to the Kurus. Suyodhana

will not accept your advice. The earth’s assembled warrior class
comes under the sway of Suyodhana. Your descent in their midst
does not please [me], Kṛṣṇa (teṣām madhyāvataraṇaṃ tava kṛṣṇa na
rocaye). Surely, not a thing would comfort us, not divinity—let alone
happiness—not the overlordship of all Immortals, with your
obstruction, Mādhava. (5.70.82–84)

Then, once Kṛṣṇa has arrived at the Kuru court, he hears the same, with only
a tweaking of the pronouns and the syntax, from Vidura:

“When all those villains are huddled together, your descent in [their]
midst does not please me, Kṛṣṇa (tava madhyāvataraṇaṃ mama kṛṣṇa
na rocaye).” (5.90.15)

90. See Hiltebeitel 1988, 131–435. The somewhat optional play for Kṛṣṇa is called Kaṇṇaṉ Jalakkiriṭai,
“Kṛṣṇa’s Water Sports,” in which Vraja is āyarpāṭi, “the village of the cowherders” (185–86).
91. Couture interprets the two passages mentioned in the text with the comment that avataraṇa can sug-
gest “the appearance of an outsider in the middle of a place, a group, or even a region” (2001, 325 n. 12).
598 dharma

This double usage of avataraṇa, which has to do with the hospitality Kṛsn ̣ ạ will
receive at the Kuru court, recalls Nārada’s vision, at the Pānd ̣ ạ va court in
Indraprastha, of the impending doom that awaits all the gathered Kṣatriyas at
Yudhiṣṭhira’s Rājasūya. “Descent” is a major meaning of avataraṇa in both
epics and the Harivaṃśa, and although there is another meaning that it is linked
with, “the taking down of a load” that is used in the Mahābhārata for the “unbur-
dening of the Earth,” it is hard to see that the latter would be the more basic or
older meaning.92 I would agree with Couture that both meanings would be at
play when “the term avatṝ alerts the audience listening to the great Epics to the
extended śleṣa (or double entendre) that runs throughout the entire text” (323).
These instances support considering another intriguing usage in the same
vein. The verb ava-tṝ occurs in six places where divinely incarnated heroes
“descend to Kurukṣetra,” which is, after all, not only the “dharma field” where
they fight the Mahābhārata war but the “high altar (uttaravedi) of Prajāpati”
(Mbh 9.52.20), to which the Vedic gods descend in sacrifice, and to which Indra
in particular descended to give King Kuru the boon that warriors who die there
will go straight to heaven. Some of the most interesting of these usages occur
in close proximity. First, after the war, the Pāṇḍavas and Kṛṣṇa return to the
battlefield and reach Kurukṣetra (avatīrya kurukṣetram; 12.48.3) now mired with
the remains of the slain warriors, “strewn with what seemed to be billions of
human skulls” (Fitzgerald 2004a, 275). And shortly after, with two usages, they
cross to where Bhīṣma will soon begin his lengthy dharma oration:

Seeming to gulp down the sky (grasanta ivākāśam), these [Kṛṣṇa’s]

swift, powerful horses crossed [down to] the Field of Kuru
(kurukṣetram avātaran), the field of the whole of the Law (kṣetraṃ
dharmasya kṛtsnasya). They went where the lord Bhīma rested upon
the bed of arrows in the midst of Brahmarṣis, like Brahmā amidst the
crowd of the Gods. Govinda then descended from his chariot
(avatīrya govindo rathāt), and so did Yudhiṣṭhira, Bhīma, the Gāṇḍīva
bowman [Arjuna], the twins, and Sātyaki.93

As noted in chapter 11, “the field of the whole of the Law” projects the first
words of the Bhagavad Gītā—kurukṣetre dharmakṣetre—toward some kind

92. Fitzgerald 2004b, 55 says avataraṇa’s first meaning is “a ‘taking down,’ a relieving of the burden that
oppressed the earth,” and “[o]nly later . . . came to signify the ‘descent’ of a deity for such a rescue mission.” See
Hacker 1960, 58–60 with this view, but also mentioning that the term takes on a doubled meaning in the text.
Couture finds that the Mahābhārata uses the term avataraṇa four times in the sense of “taking down a load” (319).
93. Mbh 12.53.23–25; Fitzgerald 2004a, 286 trans., slightly modified, with my brackets. Sātyaki is a kins-
man of Kṛṣṇa, and these seven are the sole major survivors of the battle on the Pāṇḍava side.
 dharma and bhakti 599

of completion in the dharma anthologies of Bhīṣma, which these seven war-

survivors are arriving to hear. As Fitzgerald’s translation of the three usages
of ava-tṝ indicate, it is possible to render some of these usages as “reach” or
“cross,” but why obscure the nuance of “descent” when the crossing is com-
pared to the gulping down of the sky—indeed, of space (ākāśa)? In fact, the
descending subjects in all six passages are linked only with the Pāṇḍava
side and/or the entourage of Kṛṣṇa.94 This suggests that they are describing
what Fitzgerald has called the “divine raiding party of the gods” (2004c), by
which he is referring to the divine plan of the aṃśāvataraṇa by which the
gods will relieve the burden of the earth. I would thus submit that we have
cumulative evidence for what I would call a Mahābhārata “descent
convention”95 that uses derivatives of the verb ava-tṝ before the noun avatāra
becomes a favored Purāṇic term. I believe Couture is entirely right to call
attention to theatrical resonances in many of the most vivid usages, but I
believe the ultimate stages or arenas in question, in the context of the
Mahābhārata, have to do with what the epic calls the Kuru dharma. One is
the arena of Vedic sacrifice, especially at Kurukṣetra as the “altar of the
gods.”96 After all, where did Śrī and Agni first come forth as Draupadī and
Dhṛṣṭadyumna to do the “work of the gods” but at the sacrifice of King
Yajñasena? And the others are the Kuru courts of Law at Indraprastha and
Hāstinapura.
We may thus consider the Mahābhārata to have launched both an incarna-
tional (and thus genealogical) theology centered on Viṣnu ̣ and a vocabulary of
divine descent centered on Kṛsn ̣ ạ (and also used for Rāma), but without yet
coordinating the two by the actual term avatāra. For the Rāmāyaṇa itself, we
̣ , but without
have only the former. Its divine plan is likewise centered on Viṣnu
the verbal play of a “descent convention.” The aged and sonless King Daśaratha’s
three wives bear four sons (Rām 1.14–17.9), all partial incarnations of Viṣnu ̣ ,
who has chosen Daśaratha to be his father (15.7) so that he and other gods, who
get their directions from Brahmā to take birth as monkeys (16.1–6), can dispatch

94. The other usages of ava-tṝ with Kurukṣetra describe the Pāṇḍavas and Somakas (6.1.3); Baladeva
(9.53.33); Yudhiṣṭhira (15.30.16); and Arjuna accompanying the remaining Vṛṣṇi women after the death of their
kinsman Kṛṣṇa (16.8.65).
95. On poetic conventions in the epics, see Hiltebeitel 2004a, 220–26, in a discussion that addresses
some of the passages mentioned here. Against this background, one may also recall the instance of Vārṣṇeya,
Nala’s charioteer in the Nalopākhyāna with one of the easily recognizable names of Kṛṣṇa, who is found descend-
ing (avatīrya vārṣṇeyo) from a chariot that Nala is actually driving. As such, he doubles in the Nala story, as a
prefiguration of Kṛṣṇa who is called Vārṣṇeya, as a member of the Vṛṣṇi clan (Hiltebeitel 2001a, 219, 225–26).
96. Following up on the implications of Nārada’s vision at the Rājasūya, Couture does say, “The Kurukṣetra
itself, the battlefield on which the war takes place, looks like a mythic raṅgabhūmi where Devas as well as Asuras
come to fight” (2001, 321). Of course the Nāṭyaśāstra itself interprets the origins of theater in sacrifice.
600 dharma

Rāvaṇa. The firstborn Rāma gets the biggest share (bhāga) of Viṣnụ ’s endowment;
Bharata the next largest; and Lakṣmaṇa and Śatrughna smaller portions
(17.6–9).
Vālmīki, as we shall see, for the most part leaves Rāma’s incarnation of
̣ to hints (reminders, foreshadowings, similes, allusions), but sometimes
Viṣnu
makes it obvious. Let me just note one instance that goes beyond what is said in
the Rāmopākhyāna, yet is structurally paralleled in the Mahābhārata. Amid the
prewar dialogues that take place on both sides, Rāvaṇa’s wise maternal
grandfather Mālyavān counsels him to make peace with Rāma and return Sītā.
Explaining that the gods and Ṛsị s desire Rāma’s victory, Mālyavān differentiates
dharma and adharma as divine and demonic alternatives, alludes to the
(Mahābhārata) idea that the king defines the yuga, and says that throughout the
regions, the Ṛsị s, equal to Agni, are performing fiery Vedic rites and austerities
that are damaging the Rākṣasas. Foreseeing their destruction, he notes the sin-
ister omens surrounding Laṅkā, and concludes, “I think Rāma is Viṣnu ̣ abiding
in a human body. Surely this Rāghava is no mere man by whom this bridge, the
highest wonder, is built over the ocean. Make peace with Rāma” (6.26.31–32).
Mālyavān not only sees things rightly but, with the inversion of seeing them
from the hopeless side of the demons, provides analogs to certain features of the
Bhagavad Gī̄tā: a theology for the war about to happen and a disclosure of the
hidden divinity behind it—in this case, hidden so far even from himself.97
Rāvaṇa is, of course, no more impressed by such warnings than Duryodhana.
With his focus mainly on Ayodhyā, Laṅkā, and the southward-reaching for-
ests of Rāma’s exile, Vālmīki, as we have seen, gives the gaṅgāvataraṇa an expan-
sive eastward orientation. But the Rāmāyaṇa does mention madhyadeśa once,
and in a surprising context that shows Rāma to be perfectly aware of what lies
upriver in the lands of the lunar dynasty, and his brother Bharata to have a likely
intimation of the Mahābhārata itself. Rāma tells Bharata and Lakṣmaṇa that now
that he has “truly done the excellent work of the twiceborn (kṛtaṃ mayā yathā
tathyaṃ dvijakāryam anuttamam)” (7.74.3ab), he would like to perform a Rājasūya
sacrifice. By this time, Rāma can speak of dvijakāryam, “the work of the twice-
born,” as equivalent to devakāryam, “the work of the gods.” Presumably he regards
a Rājasūya as appropriate to his being consecrated at last as king. He says he is
motivated to restore the “bridge of dharma” (dharmasetu), and regards the
Rājasūya as “eternal dharma” (3–4). But Bharata, in terms that certainly remind
one more of Yudhiṣṭhira’s rājasūya than any of the divine ones Rāma mentions

97. Cf. Goldman and Sutherland Goldman 2009, 33 (I leave above what I wrote before I could see what
they write): “Indeed [Mālyavān’s] reference to the eternal conflict between dharma and adharma is reminiscent of
Vaiṣṇava framings of the theology of the avatāra as put forth in such texts as the Bhagavad Gītā.”
 dharma and bhakti 601

as precedent,98 asks Rāma how he could perform a sacrifice “where one sees the
destruction of the lineages of the earth’s kings (pṛthivyām rājavaṃśānāṃ vināśo
yatra dṛśyate)” (74.12). After Rāma gratefully accepts Bharata’s counsel and
decides against a Rājasūya, Lakṣmaṇa chimes in to recommend a horse sacrifice
instead, since it frees from sins, and tells how Indra performed an Aśvamedha
that rid him of the sin of killing the Brahmin Vṛtra (7.75–77)—all of which
reminds one, without saying so, that Rāvaṇa was a Brahmin. Rāma then agrees
with his brothers and tells another story about the powers of the Aśvamedha: one
that goes back into the antiquities of the lunar dynasty.
King Ila of Balkh (Bactria) once went on a terrible hunting—better,
slaughtering—expedition that led him from his Afghan capital to some fairly
high reaches of the Gaṅgā. There, amusing Umā, Śiva had transformed him-
self and all creatures around him into females. Ila and all his army became
women. Umā softened matters a bit by allowing Ila to change sex on a monthly
basis, as Ilā and Ila, stipulating that neither would remember anything of the
alternations (78.26–28). Soon thereafter, Ilā caught the eye of Budha, the
planet Mercury and son of the Moon, who was doing tapas in that forested
region, and after a brief courtship,99 she became pregnant. By the time Budha
saw her first change into a man, he had a sense of what was going on, and got
Ila to agree to stay with him for a year, during which, in the ninth month,
after alternating sexes in the interim, Ilā gave birth to Budha’s son Purūravas
(79–80). Budha and a conclave of Ṛṣis then consulted as to what to do next,
when Ila’s father Kardama, presumably in retirement, happened by from
Balkh with some other sages.100 Kardama pronounced that the only remedy
he could see to get his son’s masculinity back was for the sages to perform an
Aśvamedha near Budha’s ashram, honoring the bull-bannered Śiva (81.12–15).
When the horse sacrifice was over, Śiva told the priests he was pleased with
their bhakti and asked, “What shall I do for the king of the Bāhlis?” He met
their request for the return of Ila’s manhood, vanished, and all the sages
returned whence they came. King Ila renounced his capital in Balkh, leaving
it to his son Śaśabindu, and “founded Pratiṣṭhāna city, unsurpassed and cel-
ebrated, in madhyadeśa (madhyadeśe hyanuttamam/niveśayāmāsa puraṃ
pratiṣṭhānaṃ yaśaskaram)” (78.21b–d). There King Purūravas, Ila’s son,

98. Rāma mentions one by which Mitra obtained Varuṇahood, which sounds perhaps benign, and also
mentions one done “in accord with dharma by the dharma-knowing Soma,” which the Mahābhārata tells us led
to no good. See Harindranath 2010.
99. This involves Budha learning from Ila’s feminized armed forces that Ilā has no husband, and turning
them into Kiṃpuruṣīs, “What-Woman” who will henceforth roam the slopes with Kiṃpuruṣas, the “What-Men”
who often go about with the Gandharvas (1.79.20–24).
100. Kardama means “mud, slime, mire, clay, dirt, filth” (MW 258), and his name presumably reflects the
typical aspersions on the people of Bactria and Madra one finds in the Mahābhārata. See chapter 8 § F and n. 104.
602 dharma

succeeded him (22). Note that Purūravas is called the son of the male rather
than the female Ila/Ilā, and is not called the son of Budha. Rāma is perhaps
nodding to some preferred notion of primogeniture.101 His story stops here,
but in the Mahābhārata the lunar line can be traced from Purūravas through
Āyus, Nahuṣa, and Yayāti to Pūru, the youngest of Yayāti’s five sons who
inherits madhyadeśa when Yayāti distributes the earth among them.102
Rāma’s tale sounds at times like a spoof on the lunar dynasty’s early gene-
alogy, though we should pause over the point that, having now banished Sītā,
he will be performing his Aśvamedha with a golden statue as her substitute,
which he might be making light of as well. The Aśvamedha that restores Ila’s
masculinity does not say how that exactly worked out, or what if anything
he/she did in it as Ilā. Pratiṣṭhāna, however, gets a striking description in the
Mahābhārata that makes a fitting close to this section on “descent” and the
geography of madhyadeśa. Says Nārada to the pilgrim Pāṇḍavas:

The land between the Ganges and the Yamunā is known as the vagina
of the earth (pṛthivyā jaghanaṃ smṛtam); and Prayāga and Pratiṣṭhāna,
thus the seers know, form the end of the vagina, the vulva (prayāgaṃ
jaghanasyāntam upasthamṛsạ yo viduḥ). Prayāga, Kambala, Aśvatara,
and the ford Bhogavatī are declared to be the altars of Prajāpati (vedī
proktā prajāpateḥ). There the Vedas and sacrifices take on bodily form,
Yudhiṣṭhira, and with the Seers of great vows wait upon Prajāpati.
(3.83.71–73; van Buitenen trans. 1975, 396)

The Mahābhārata Critical Edition does not seem to have much more to tell us
about Pratiṣṭhāna other than that it still remained the Lunar Dynasty capital
when Garuḍa told the indebted disciple Gālava, “My friend (sakhā) in the Lunar
lineage . . . Yayāti, son of Nahuṣa” lived there, and flew Gālava to Pratiṣṭhāna to see
if Yayāti could help him pay off the debt (5.112–6.9). By the time of King Duṣyanta,
the dynasty had settled at Hāstinapura. But the Southern Recension interpolates
a long passage in which the dynastic progenitrix Śakuntalā and her son Bharata
approach Hāstinapura by way of Pratiṣṭhāna, which is portrayed as still thriving as
a kind of idyllic place (Mbh 1, App. I, Nos. 47 and 48). According to Wilson,
“Pratishṭhána was situated on the eastern side of the confluence of the Ganges
and the Jumna” ([1840] 1972, 280 n. 7). This would put it opposite to Prayāga,
which would be inside the confluence, and might help to explain how the two sites

101. See Brodbeck 2009a, 101, citing Mbh 1.70.16: “Purūravas the sapient was born from Ilā; indeed, she
was his mother and his father too—or that’s what we’ve heard.” Brodbeck interprets Ilā as a putrikā, and as pro-
ducing “heirs for two families” (102, 144). See further Brodbeck 2010d.
102. Mbh 1.79. See Dumézil 1973, 15–20; Defourny 1978, 134–37.
 dharma and bhakti 603

together form the “end of the vagina.” It would be nice to know more about the
Lunar Dynasty’s capitals in Madhyadeśa. In the Rāmāyaṇa account, Ila seems to
have come downriver a considerable way to establish Pratiṣṭhāna for Purūravas at
the outer tip of the Doab. Perhaps it is a way of saying that when Pūru inherited
Madhyadeśa, he inherited it right at the center.103

D. Friendship, Hospitality, and Separation

So we have now mapped dharma and bhakti down to some places on earth. We
turn to friendship, hospitality, and separation as interrelated themes or dis-
courses, each involving norms and practices, in which dharma and bhakti coin-
cide.104 With hospitality, if we simply ask who hosts Kṛṣṇa and Rāma in these
epics? who do they host in turn? how and where do they go about it? what are
the issues involved? what is the tone or mood created? we get into revealing
material. And similarly with friendship. We have seen in chapter 10 how Sītā
suffers in separation from Rāma. A wrenching devotional tone or mood can
also be noticed in circumstances where Rāma or Kṛṣṇa are parting and people
accompany them as far as possible and then hold them dearly in mind.
The epics are sometimes very attentive to such departures, and to the
demeanor and moods of those who follow Rāma and Kṛṣṇa as far as they can,
in ways that are not the case for the comings and goings of other “characters.”
Rāma’s prolonged departure from Ayodhyā is perhaps well enough known to
be recalled only in a few details below. The Mahābhārata describes four such
departures of Kṛṣṇa, plus his final departure from the world.105 Most represen-
tative is his poignant last farewell, which comes after Kṛṣṇa, out of goodheart-
edness (sauhṛdāt; 14.16.6), has lingered longer than he wished to help Arjuna
to remember the Bhagavad Gītā by telling him the Anugītā:

Hari . . . then set out from the City of the Elephant [Hāstinapura] on his
divine chariot drawn by four horses. Having with Yudhiṣṭhira’s
permission caused his sister Subhadrā to mount and also his paternal

103. On madhyadeśa, see (as last cross-referenced) chapter 8 n. 44. On Pratiṣṭhāna, Dubey 2001, 14–39, takes its
likely location to be the unexcavated Jhūsī mound across the Ganges from Allahabad, occupied into medieval times.
104. For further discussion of these themes, see the four studies cited at the beginning of § A above.
105. The first two come after the founding of the Pāṇḍavas’ first capital at Indraprastha (2.2.1–23) and the
stated completion of Yudhiṣṭhira’s Rājasūya (2.42.45–59); the third, when Kṛṣṇa leaves the Pāṇḍavas at Upaplavya
(their prewar camp) to go to the Kurus (5.81.6–57); the fourth, discussed in the text, before Yudhiṣṭhira’s
Aśvamedha in Book 14. Kṛṣṇa’s final departure comes in Book 16 (see two notes down). Taken all together, it may
be noticed that the focus shifts from leaving Yudhiṣṭhira in the first two to leaving Arjuna in the last two, while
in the third Kṛṣṇa leaves both of them. Only with Yudhiṣṭhira does Kṛṣṇa affirm a saṃvid (covenant, engagement;
2.2.19c, 42.59a) in such a passage.
604 dharma

aunt (Kuntī), the great-armed Janārdana set out surrounded by the

cityfolk. The one (Arjuna) whose banner bore the best of monkeys,
Sātyaki, the twin sons of Mādrī, Vidura of unfathomable intellect
(agādhabuddhir),106 and Bhīma himself who had an elephant king’s stride
followed Mādhava. Then, having made Vidura and all the heroic
increasers of the Kuru kingdom return, Janārdana said to (his charioteer)
Dāruka, “Urge the steeds to speed.” . . . While Vārṣnẹ ya was proceeding
to Dvāraka, O Bharata bull, those foe-scorchers with their retinue, having
embraced [him], turned back. Again and again Phalguna [Arjuna]
embraced Vārṣnẹ ya and as long as he was in eyes’ range, he saw him
again and again. And even so, Pārtha withdrew that sight fixed on
Govinda with difficulty, and the unvanquished Kṛsn ̣ ạ [withdrew his sight
with difficulty from Arjuna] as well. (14.51.52c–55, 52.1–3)

In that Arjuna and Kṛṣṇa will have no later farewells,107 the scene anticipates
Arjuna’s forlorn grief for Kṛṣṇa when he finds he has left the world.108 As they
would be in later texts, such scenes play on the bhakti trope of viraha or “love in
separation,” and it is worth pointing out that Aśvaghoṣa, who knew both epics,
already develops this idiom in his portrayal of the Buddha’s Great Departure,
as we shall see in chapter 13.
Hospitality and friendship are not the first things one thinks of regarding
epic treatments of dharma and bhakti. But if a post-Mauryan Brahmanical
ideology puts kingship front and center in both epics and Manu, and may even
be said to be what their arguments are ostensibly about, this is not, in any of
these texts, the grounds on which the argument was capable of being won. How
one wishes to order society is one thing. How people get along is another. That
was encouraged by invoking hospitality and friendship among the more open
and flexible civilizational discourses and practices (another is the gift) familiar
as custom throughout South Asian Arya ̄ culture under endless local and regional
variations. Indeed, these are discourses and practices that the Buddha himself

106. An epithet of Vidura also at 3.5.1, 5.30.29, and 5.81.48, and otherwise only of Vyāsa at 18.5.28.
107. There is no description of Kṛṣṇa’s departure after returning for Yudhiṣṭhira’s Aśvamedha; rather, the
focus is on his description of Arjuna’s return, looking gaunt, after a year of guarding the sacrificial horse. Kṛṣṇa
is then among those welcoming Arjuna’s return (14.89). See Hiltebeitel 2007a, 129–30.
108. See Hiltebeitel 2001a, 85–91. Thirty-six years after the war, when, in accord with Gāndhārī’s curse,
all of Kṛṣṇa’s clansmen slaughter each other in a brawl, before Kṛṣṇa ascends to heaven, he sends for Arjuna to
protect the Yādava women. Arjuna arrives to see Dvārakā bereft and hears the wails of Kṛṣṇa’s 16,000 wives
(16.6.4–6). Kṛṣṇa’s agonized father Vasudeva wonders how his son, the lord of the universe (jagataḥ prabhuḥ)
who slew Kaṃsa, Ekalavya, Śiśupāla, and others, could have overlooked the slaughter of his own kinsmen, and
tells Arjuna Kṛṣṇa’s prediction that when Arjuna leaves Dvārakā, the ocean will deluge the city (7.9–17). Arjuna
says he cannot look at the earth without Kṛṣṇa, and that his brothers and Draupadī feel the same. He also knows
it is time for them all to move on (saṃkramaṇa), which can mean their passage together into the other world
(8.1–4). With Kṛṣṇa gone, Arjuna then finds he has little power to protect the women.
 dharma and bhakti 605

relied on rather than criticized (as he did class), and could do no better than
attempt to refine, as we saw in chapter 4. In reformulating hospitality and friend-
ship as dharma, the dharmasūtras and Manu all sought to harmonize custom
with Vedic practices and discourses, and so did the epics. But the epics could
give this amalgam far more complex treatment by telling stories, among others,
about hallowed ancient Ṛsị s and a god among men. In making this “swerve,”
they could develop it in narratives that were far more nuanced than incessant
top-heavy reminders that the four social classes were created from Puruṣa.
For the most part, the basic vocabularies on friendship and hospitality are
shared by the epics and the dharma literature. But the epics also innovate and
archaize. Several Sanskrit words are often translated by “friend,” three of which,
for our purposes, are important to differentiate. Mitra, a Vedic term, has a pri-
mary meaning of “ally,” but can also sometimes be translated more usefully as
“friend.” Suhṛd, which seems to be an epic coinage, literally means “good-
hearted,” “one with a good heart,” and will be translated as “wellwisher,” with
all the ambiguity that the term can suggest, while its derivatives sauhṛda/
sauharda/sauhārda can be translated “goodhearted.” Most important, already
mentioned in connection with Kṛṣṇa’s special friendships with Draupadī and
Arjuna and the only term I will translate as “friend,” is sakhi, with its derivative
sakhya, “friendship.” It is important to note that Mahābhārata usages, espe-
cially, draw on Vedic precedents in giving this term two senses, which I call
“pact friendship,” as when Indra makes pacts with middle-Vedic demons such
as Vṛtra and Namuci,109 and “intimate friendship,” as when Indra is helped to
defeat Vṛtra by his “intimate friend” Viṣṇu (índrasya yújya sákhā) in Ṛgveda
1.22.19. Elsewhere (Hiltebeitel 2007a, 131), I have observed that Kṛṣṇa’s friend-
ship with Draupadī comes explicitly into play only in two scenes where her
difficulties are colored by Vedic ritual injunctions: one, where her disrobing
occurs in an inversion of the outcome of the ritually required dice match that
should have ended the Rājasūya with Yudhiṣṭhira winning;110 the other, in her
underplayed fulfillment of the rule that the mahiṣī expose herself sexually to the
sacrificed horse in the Aśvamedha. In each case, Kṛṣṇa intervenes in ways that
lighten Draupadī’s sexual humiliation. As the epic wife takes on roles as victim111
within the arena of the great Vedic royal rites, she has this special sakhi

109. Garuḍa’s friendship with Yayāti, mentioned in the previous section, is presumably of this kind, as are
most interspecies friendships.
110. See van Buitenen 1972. Draupadī mentions her friendship with Kṛṣṇa thrice in this vein: at 2.62.9–11,
3.13.52–53, and 5.80.21–26. See Hiltebeitel 2007a, 127–28.
111. See Jamison 1996, 256 on the “sacrificed” sacrificer’s wife playing the role of mediator between men
and gods. I would suggest that this is the Vedic ground from which this aspect of Draupadī’s relation to Kṛṣṇa is
developed by the epic poets. Note that Vālmīki spares Sītā this indignity at Rāma’s Aśvamedha by supplying
Rāma with a golden statue of her in her stead.
606 dharma

to soften the complications she faces in what the text requires of her as it
transposes these rituals from rules into narratives. Similarly, Kṛṣṇa’s sakhya
with Arjuna comes most explicitly into play in the Bhagavad Gītā as Arjuna
faces ahead to the dilemmas of the “sacrifice of battle.” In these moves from
rite to narrative, it is as if the late Vedic slogan that “Viṣṇu is the sacrifice”112
were reassuringly reenplotted as “where Kṛṣṇa is, there is dharma.”
Meanwhile, as to hospitality, atithi is the main old word for “guest,” and
ātithyam for “hospitality,” while there is no consistent term for host, that concept
being more contextual. A host may be found in a house, a sacrifice, a performance,
perhaps on a chariot, in a heart, etc. A few preepic usages are also interesting
here. I believe there is something to be made of the fact that Viṣnu ̣ is “the ‘guest’
par excellence” in the ātitheṣṭi, the ātithya or “guest offering rites” rites men-
tioned in the Śatapatha Brāhmaṇa (3.4.1.1) and Āpastamba Śrautasūtra (10.30.1–31,
31.6–7), which “call the gods to mind,” and to which Viṣnu ̣ is “invited” by the
acchāvāka or “inviter” priest, a deputy of the Hotṛ, by reciting Ṛgveda 1.154.1–3,
praising Viṣnụ for “traversing three times,” he whose “power . . . is like a terri-
fying, hungry, wild animal who dwells in the mountains (or in speech), the one
of many hymns” in whose “three steps, all worlds abide” (Patton trans. 2005, 151).
Used in simple iṣṭi or cake-offering rites, the ātithyeṣṭi can also be used “for the
reception of Soma stalks, which are considered kingly, or a royal guest.”
According to Laurie Patton, the verses, according to the Śaṅkhyāyana Śrautasūtra,
are also used during the cutting of the victim for Viṣnu ̣ in the animal sacrifice
(paśuyajña), while according to the Ṛg Vidhāna (1.136–37) they are to be recited
for the adept “to attain dharma, intelligence (jñānam), sons, the increase of
Brahmā, and the highest abode of eternal light” (2005, 9, 171–73). Also on
guests, and suggestive in the same vein, is Āpastamba Dharmasūtra 2.7.5–10:

“Whether you hold them dear or not,” it is stated, “guests lead you to
heaven.” When a man gives food in the morning, at noon, and in the
evening, they constitute the three pressings of Soma; when he rises
as his guest gets up to leave, it constitutes the final rite of the Soma
sacrifice; when he addresses the guest with kind words, it constitutes
the praise of the priestly fee; when he follows the guest as he leaves,
it constitutes the Viṣṇu steps; and when he returns, it constitutes the
final bath. (ĀpDhS 2.7.5–10)

It is interesting, in the light of passages on Kṛṣṇa’s and Rāma’s departures, that

following the guest until he leaves should “constitute the Viṣṇu steps.”
Āpastamba is vivid in exemplifying dharma as hospitality:

̣ ḥ occurs frequently in the Brāhmaṇas; see Hiltebeitel [1976] 1990, 105, 293–94, 333, 356.
112. Yajño vai viṣnu
 dharma and bhakti 607

A guest comes like a blazing fire. When someone has studied one
branch from each of the Vedas in accordance with dharma, he is called
a “Vedic scholar.” When such a man comes to the home of a house-
holder devoted to his svadharma—and he comes for no other purpose
than to put dharma first—then he is called a “guest.” By paying him
homage, the householder obtains peace and heaven. (2.6.3–6)

The “blazing fire” like which a guest comes is ”agni”: it could be the god Agni
visiting, or the opportunity to offer one of the five daily “great sacrifices” to feed
an almsman. This brings us to a sort of touchstone-text: a surprising unit from
the Mahābhārata’s 13th Book that closes King Yudhiṣṭhira’s postwar instruction
on the topic of “the Law of Giving” after the learned Bhīṣma has covered three
prior dharma topics in Book 12—a curriculum that we shall revisit in chapter 13.

E. Ṛṣidharma

This touchstone is the “Dialogue Between Umā and Maheśvara” or Umā-

Maheśvara-Saṃvāda (13.126–34, henceforth UMS). The expansive unit comes
at one of several points of increasing “bonhomie” (as Bowles [2007, 387, 390]
has put it) that can be traced through the entire arc of Yudhiṣṭhira’s postwar
instruction, and also one that could have something to do with a staggered but
cumulatively increasing emphasis on bhakti through that four-part dharma
instruction where Bhīṣma coordinates his lawgiving with bhakti.113 It begins
with Yudhiṣṭhira wanting to know more about Kṛṣṇa who has up to now been
standing by, once again the listening guest, ever since Bhīṣma began his long
postwar dharma talk on the subject of Rājadharma. Says Yudhiṣṭhira:

If I am favored by you along with my brothers, O sinless one, you can

speak to our question that I ask you, O king. This Nārāyaṇa, possessed
of prosperity (śrī), is esteemed by all kings, and even he waits upon
you with much respect and deference (bahumānena praśrayeṇa ca). In
his presence, you can speak out of affection for me and for the benefit
of all these lords of the earth, my brothers. (13.126.4–6)

̣ ạ that
Bhīṣma is overjoyed to speak on this subject, and tells one story about Kṛsn
leads him to a second in which he quotes Nārada to narrate the actual UMS.

113. A possibility Zaehner (1963) and Brockington (2000b) raise in the Mokṣadharma, with its Nārāyaṇīya
toward the end of that third Śāntiparvan anthology. Indeed, soon after the UMS, Bhīṣma, claiming that he is los-
ing strength, directs Yudhiṣṭhira to ask a question about Brahmins to Kṛṣṇa himself, which turns into further
questions about the irascible Ṛṣi Durvāsas that lead Kṛṣṇa to speak also about the greatness of Śiva (13.143–46)—
before the dialogue returns for its windup with Bhīṣma. See Hiltebeitel 2005d, 260.
608 dharma

As one would expect, there are threads that link the two stories. The first tells
how Kṛsn ̣ ạ once undertook a twelve-year vow at some mountain to obtain a son
who would be his equal in vigor and would have half the energy of Śiva.114
Upon completing the consecration (dīkṣa), ̄ Kṛsn
̣ ạ was visited by Nārada,
Vyāsa, and other important sages with their disciples—a veritable Ṛsị saṅgha
(13.126.49a), in fact—and received them with rites of hospitality (atithisatkāram)
like those offered to the gods (13). The term Ṛsị saṅgha is worth tracking along
with its equivalent, the “host of Munis” or Munigaṇa. Once Kṛsn ̣ ạ had seated
them all on gorgeous seats, they chatted amiably on matters of dharma until
Kṛsṇ ạ released a fire from his mouth that burnt the entire mountain with its
living creatures and came back to him like a docile disciple, after which he brought
everything back to life. Noticing that his well-travelled visitors were surprised
(vismito) and given to dismay (vismaya) at this inconceivable wonder (adbhutam
acintyam), Kṛsn ̣ ạ explained that the fire contained his wrath and was his Vaiṣnạ va
tejas or “energy of Viṣnu ̣ .”115 Then he asked them to tell him something highly
wonderful (āścaryaṃ paramaṃ kiṃcit) that they had seen or heard on earth or in
heaven, even while admitting that every such wonder is known to him, since
what is “told by the pious and heard by the good is surely beneficial and lasts long
on earth, like writing engraved on a mountain” (43)! Once the Ṛsị saṅgha is
f inished glorifying (vardhayantas) and worshiping (pūjayantas) Madhusūdana
with hymns of praise (46), the sages appoint Nārada to tell the sequel.
Nārada, now called “the wellwisher of Nārāyaṇa” (nārāyaṇasuhṛd; 13.127.1a),
then leads us into an idyll of Śiva and Umā on Mount Himavat (1–39) that is
interrupted when Umā covers Mahādeva’s eyes in jest (26), leaving the world
in darkness and distress until Śiva’s third eye (30) opens and burns the
mountain—which Umā then gets him to restore to its natural beauty. This
inspires Umā to ask a series of questions on topics that intrigue her: first Śiva’s
third eye,116 then his four faces, his bull, and his preference for the cremato-
rium, which he explains by the delight he takes there with his ghostly host
(bhūtasaṅgha, bhūtagaṇa; 128.18)—that leads Umā to ask next, for the benefit of

114. Mbh 13.126.33 and 45. It is not immediately clear who this son would be. As one learns in the
Upamanyu-Upākhyāna (13.14–18), Jāmbavatī once reminded Kṛṣṇa, after twelve years, that she wished him to
obtain an heroic son for her “like himself” (ātmatulya) by doing something similar to the twelve-year vow to Śiva
he had done to obtain the eight sons of Rukmiṇī (13.14.12–16). Jāmbavatī, a relative of Jāmbavat, the king of bears
in the Rām, is here “the daughter of the Indra of monkeys” (kapīndraputrī, 24b), and their son will be Sāmba, who
would seem to be the likeliest candidate. Rukmiṇī’s eight sons with Kṛṣṇa are listed here at verse 17. On Samba,
see von Simson 2007.
115. According to Biardeau 1994, 63–88, the avatāra concept conceives Viṣṇu’s incarnations to have not
just a dharma-preserving aspect but, simultaneously, a raudra (terrifying) aspect that preserves dharma through
Time’s destruction and recreation (see Biardeau 1994, 63–88).
116. See O’Flaherty 1973, 248–51 on this myth and what seem to all be later variants (also 31–32, 190–91).
 dharma and bhakti 609

the hosts of Ṛṣis and Munis (ṛṣisaṅgha, munigaṇa) who attend them on Mount
Himavat, what dharma is and how men learn of it (128.21–23), and then to ask
about the dharma of the four social classes, and particularly that of Brahmins
(13.128.29a). Śiva, who has just stated his rather capacious view of dharma’s
many branches (dharmo bahuśākhaḥ; 128.27), focuses his reply on householder
Brahmins as “gods on earth” (bhūmidevāh), ̣ for whom several of his prescrip-
tions involve hospitality practices—honoring one’s gurus and gods, and keep-
ing houses properly rubbed (with cowdung and water) and fumigated with the
smoke of clarified butter (128.30–34, 35–45)—related to the five daily mahāyajñas
underscored in the Gṛhyasūtras and further developed especially as dharma in
the Dharmasūtras and Manu (see chapter 5).
After hearing more about householders, however, Umā wants to know
about the dharma of Ṛṣis or Munis themselves (using the compounds ṛṣidharma
and munidharma at 13.129.32 and 34). Śiva, who delights not only in his hosts
of ghosts but praises outlandish ascetics, now uses the verbal root √uñch, “to
glean,” several times to describe different types of “gleaners”: Phenapas,
“Foam-drinkers,” who as it were “glean” or perhaps better “skim” their nour-
ishment from the foam of waters left over from sacrifices (36–38b); Cakracaras
who practice the dharma of compassion (dayādharma), rove in the world of the
gods (devaloka), and “glean” their food from moonbeams (43–44, 48); others
who live with their wives, practice the five mahāyajñas, and “glean” their nour-
ishment from the flames of their fires; and thumb-sized Vālakhilyas who live
in the solar disc, wear deer skins or tree bark, follow the actual uñcha or
“gleaning” mode of life by subsisting like birds on grains left in the fields
(uñcham uñchanti dharmajñāḥ śākunīṃ vṛttim āsthitāḥ), and “attain equity with
the gods in accomplishing the purpose of the gods’ work (te suraiḥ samatāṃ
yānti surakāryārthasiddhaye)” (129.39c–42b). By now we know enough about
the phrase surakārya, “the work of the gods,” to know why these gleaners are on
a plane of “equity with the gods.”
It is with reference to this varied “gleaning” ideal that Śiva then summarizes
the dharma of Ṛsị s and Munis in the passage that should interest us most:

In all the Ṛṣidharmas, selves are to be conquered, sense faculties are

conquered. Afterwards, desire and wrath are then to be conquered, so
I think. Maintaining the Agnihotra, sitting for the Dharmarātri,117
Soma sacrifice, assenting to approval, and [offering] honoraria at
sacrifices as the fifth, [plus] the daily dharma of sacrificial acts:

117. I do not know what this is. Ganguli [1884–96] 1970, 11: 296 has: “occupying a fixed seat employing
oneself the while in the sacrifice called Dharmaratri.” A nightlong vow is probably suggested. Though a Bengali,
he seems not to have been tempted to relate it to a Dharmagajan. On this Dharma cult, see chapter 9 § D.2.b.
610 dharma

pleasure in the worship of Pitṛs and gods and hospitality to all—

[these] are to be done with food acquired by gleaning. Turning away
from food prepared with cows’ milk and taking pleasure in lying on
the bare ground, yoga, enjoying vegetables and leaves, eating fruits
and roots, partaking of wind, water, and duck weed—these are some
observances of the Ṛṣis by which they conquer the way of the
unsubjugated. When there is no more smoke, when the pestle is set
down, when there are no more coals, when the people have eaten
their meal, when the handing around of vessels is over, when the
time for asking alms has passed by, surely [it is then, still] longing for
a guest, [that] one eats the food left over. Delighted by the dharma of
truth, patient, he is yoked to the Munidharma. Not arrogant or proud,
the one who is neither heedless nor surprised (vismita), a friend alike
to friend and foe, he is the foremost knower of dharma.118

As mentioned in chapter 4, the basic hospitality practice of waiting before one

begs for food until the four (or fewer) signals of the settling or setting down of
the pestle, smoke, embers, and plates occurs as a dharmaśāstra adage in
Baudhāyana (BDhS 2.11.22), Vasiṣṭha (VDhS 10.7–8), Manu (6.56), and in the
Mahābhārata (at 12.9.22), where Yudhiṣṭhira tells how he would like to be a
carefree mendicant just after the war. It is also mentioned in the Aggañña Sutta,
where the Buddha, offering a sly etymology for Brahmins by punning on the
words linking fire-tending and meditation, speaks of the prelapsarian practices
of the original good Brahmins:

They made leaf-huts in the forest and meditated in them; without

coals or smoke (from a cooking fire), pestle set down, they went into
villages, towns and royal cities in search of food, in the evening for
their evening meal, and in the morning for their morning meal.
(DN 27.22)

118. Mbh 13.129.48–55: sarveṣv evarṣidharmeṣu jeyātmā jitendriyaḥ/ kāmakrodhau tataḥ paścāj jetavyāv iti me
matiḥ// agnihotraparispando dharmarātrisamāsanam/ somayajñābhyanujñānam pañcamī yajnadakṣiṇā// nityaṃ
yajñakriyādharmaḥ pitṛdevārcane ratiḥ/ sarvātithyaṃ ca kartavyam annenoñchārjitena vai// nivṛttir upabhogasya
gorasānāṃ ca vai ratiḥ/ sthaṇḍile śayanaṃ yogaḥ śākaparṇaniṣevaṇam// phalamūlāśanaṃ vāyur āpaḥ śaivalabhakṣaṇam/
ṛṣīṇāṃ niyamā hyete yair jayantyajitāṃ gatim// vidhūme nyastamusale vyaṅgāre bhuktavajjane/ atītapātrasaṃcāre kāle
vigatabhaikṣake// atithiṃ kāṅkṣamāṇo vai śēṣānnakṛtabhojanaḥ/ satyadharmaratiḥ kṣanto munidharmeṇa yujyate//
na stambhī na ca mānī yo na pramatto na vismitaḥ/ mitrāmitrasamo maitro yaḥ sa dharmaviduttamaḥ. For the most
evocative lines in verses 53–54b, Ganguli ([1884–96] 1970, 11: 297) has this lovely evocation of a life he surely
knew better than I: “When the smoke has ceased to curl upwards from a house, when the husking machine has
ceased to ply, when all the inmates have taken their food, when dishes are no longer carried from room to room,
when mendicants have ceased to walk the streets, it is then that a man desiring to have a guest (but finding his
desire ungratified), should eat what remnant of food may still occur in his house.”
 dharma and bhakti 611

Here it seems to be a question of putting these things aside oneself before

going begging. Except for Śiva’s usage, everyone else describes when a begging
guest gets to eat, not when the host waiting for a guest gets to eat. Achieving a
startling effect, the Mahābhārata has transformed the adage to describe
Ṛṣidharma as what householder hosts do themselves, longing for a guest to
honor with their meager fare.
Now it is not surprising that Umā and Śiva, surrounded in their Himalayan
retreat by hosts of ghosts and eccentric sages, would share an interest in
unusual and surprising dharmas. But the term Ṛṣidharma was a surprise to me,
as I had never even considered that there was such a thing.119 One does not hear
about Ṛṣidharma or Munidharma from Manu, even though the celestial sages
form the first audience of the Mānava Dharmaśāstra. Nor, even though we may
have found Kṛṣṇa referring to it indirectly when he tells Arjuna about Munis
who “came to have the same nature (dharma) as me” and who “even at the
Creation do not take birth, and are not disturbed at the Dissolution” (BhG 14,2),
it is not the sort of thing one would have expected in the Bhagavad Gītā.120
It shows the value of having more than one god’s opinion about dharma.
Indeed, Śiva will round off his dialogue with Umā by getting her reply to his
question about the dharma of women (134.1–29).
The terms Ṛṣidharma and Munidharma occur five and four times respec-
tively in the Mahābhārata, all but two of them in the UMS, and all of them in
Bhīṣma’s postwar instructions.121 Clearly, Mārkaṇḍeya and Nārada exemplify
this “dharma of the Sages”: the first by surviving and witnessing the Dissolution;
the latter by narrating Śiva and Pārvatī’s dialogue in the UMS, where the terms
are most fully developed. Neither term is found in the Rāmāyaṇa. Yet Ṛṣidharma
is all over the place in both epics, and has much to do in them not only with
dharma but also with bhakti. As the quote from Śiva would suggest, it has to do
with hospitality and friendship; and if gleaner Ṛṣis “attain equity with the gods
in accomplishing the purpose of the gods’ work” (13.129.39c–42b), we can see
that Ṛṣidharma will be consonant with the divine plan. One finds it above all in
the Ṛṣis the heroes meet. But let us begin by noting some of the ways it can be
appreciated straight from this passage. Most crucial is the very idea that there
is a Ṛṣidharma or Munidharma that is of interest to such a grand audience as a

119. Bronkhorst 2007, 81 cites a Mokṣadharma passage that mentions “the Dharma of Ṛṣis,” but does not
pursue the term. The passage (12.185.1.1–2), in prose, describes the ascetic regimes of “forest-dwellers [who]
follow the Ṛṣidharma” (vānaprasthaḥ khalv ṛṣidharmamanusarantaḥ; 1.1a).
120. Yudhiṣṭhira seems unfazed by it, and formulates his next question on the One God, whom Bhīṣma
identifies as Vāsudeva, and the highest dharma (13.135.2–3), which turns out to be the recitation of Viṣṇu’s
thousand names.
121. All but two are in our current dialogue—those at 12.61.10, where Bhīṣma speaks of “the very difficult
householder dharma seen as Munidharma” and at 12.185.1–2, as mentioned two notes above.
612 dharma

company of Ṛṣis or a host of Munis. I will first deal with practitioners of

Ṛṣidharma—at least as Śiva describes them—and then turn to the topic of Ṛṣi
companies as audiences.
It seems that despite their elemental modes of subsistence and access to
higher worlds, the gleaners of the various types who practice Ṛṣi- or
Muni-dharma still lead earthly lives, and would not be among those in the Ṛṣi
companies or Muni hosts who attend Umā and Śiva’s conversation or come
in conclave to honor Kṛṣṇa. As one can gather from Baudhāyana and Manu,
the gleaner’s mode of life is an ideal for real people that in some sense sur-
passes the categories of Class and Life-stage Dharma, and has nothing to do
with class-mixing.122 To correct some misconceptions, gleaners are not neces-
sarily Brahmins, and do not belong in a statement about Brahmins “forced”
by varṇasaṃkara “to subsist by performing military roles, shopkeeping or
even engaging in agriculture—or they are reduced to ‘gleaning’ and often
starve. The king who allows such degradations to occur is said to be seriously
deficient” (Fitzgerald 2006a, 275). As Manu says, although a king “may eke
out a living by gleaning (śīloñchena), his fame spreads in the world like a drop
of oil on water” (7.33). But as Manu indicates,123 it is mainly Brahmins who
would have this high ideal: that of a sort of liminal householder life some-
where between the second and third āśrama, enjoined to perform the five
mahāyajñas but with what seems the absolute minimum that its practitioners
might offer to their own families, not to mention guests. Among those who
are mentioned by Śiva, the foam-, moonbeam-, and fire-gleaners could be
forest-dwellers, but the actual grain-gleaners would have to live near agricul-
tural fields even if they are dressed in leaves and tree bark. It is indeed a high
ideal in the Mahābhārata, where it is exalted in such terms as these to
Yudhiṣṭhira in several upākhyānas, where a god in disguise (his father
Dharma [see chapter 9]), or a saint (that irascible clone of Śiva, Durvāsas),
might come as a hungry guest to test its practitioner. Yet one notices that
Yudhiṣṭhira and his companions only hear about such gleaners and never
meet one, especially in the forest. Clearly the ideal is precisely this: to live out
daily, as Śiva says, the challenges of hospitality and friendship to all with
patience in a world of marvels, indeed, where one is oneself one of the mar-
vels, but as one who is surprised by nothing. It remains a living ideal in India,

122. After telling Yudhiṣṭhira the Uñchvṛttyupākhyāna, in which the snake Padmanābha praises a perfected
Muni who had gone to heaven by undertaking this practice, Bhīṣma says it is “best dharma for those in the life-
stages” (dharmamāśramiṇaṃ ̄ śreṣṭham; 12.340.1), possibly implying that it can be practiced in any of the four.
123. See M 4.7 (“the length of time one stores up grain in advance shows the superiority of those who store
the least”); cf. 3.100; 4.4–10 (limiting the daily sacrifices, which, according to 6.5, the forest hermit should
perform). Cf. chapter 4 §§ C.1 and C.3 on the Buddhist monastic rule of limiting food storage to seven days.
 dharma and bhakti 613

as I learned from my friend T. P. Mahadevan, who grew up in Kerala. Whether

his mother had ever seen one or not, she thought there was no one finer than
an uñchavṛtti Brahmin who (quoting Thennilapuram Mahadevan, e-mail,
July 7, 2006) lived a life beyond “the taint of the dakṣiṇā that regular Vaidika
Brahmins had to endure” and could rise “above it all, a liberated soul squar-
ing the circle, as it were, in the world but out of it.”
As to the audiences of more celestial Ṛṣis and Munis such as the two
Ṛṣisaṅghas that attend to Kṛṣṇa and to Umā and Śiva, here we meet variants
of the widest audiences of the two epics and Manu: in the Mahābhārata, the
stellar Naimiṣa Forest Ṛṣis who hear the epic from the bard Ugraśravas in its
outer frame; the celestial Ṛṣis who come to the earthly Naimiṣa Forest from
Brahmaloka with Brahmā’s permission to hear the rest of the Rāmāyaṇa
from Kuśa and Lava once Sītā has descended into the earth;124 and the celestial
Seers who listen to Manu’s son Bhṛgu recite the Manusmṛti. We also find
another variation in the last verses of the Śvetāśvatara Upaniṣad, a text whose
chronological relation to the epics and Manu is perhaps too close to call:125

By the power of his austerities and by the grace of God

(devaprasādād), the wise Śvetāśvatara first came to know brahman
and then proclaimed it to those who had passed beyond their order[s]
of life (atyāśramibhyaḥ) as the highest means of purification that
brings delight (juṣṭam) to the company of seers (ṛṣisaṅgha).
This supreme secret was proclaimed during a former age
(purākalpe) in the Vedānta. One should not disclose it to a person
who is not of a tranquil disposition, or who is not one’s son or pupil.
Only in a man who has the deepest love for god (yasya deve parā
bhaktir), and who shows the same love for his teacher as towards
God, do these points (arthāḥ) declared by the Noble One shine forth
(prakaśante mahātmanaḥ, which is repeated). (ŚU 6.21–23, Olivelle
trans. 1998, 422–23)

Here we meet the term Ṛṣisaṅgha explicitly, and again with some differentiation
made between this “company of seers” and those in whom they delight who
have “passed beyond their order(s) of life as the highest means of purification.”
It is not certain whether the expression translated here as “beyond their order(s)

124. Rām 7, Appendix I, No. 13, lines 21–49. Although the Baroda CE rejects this sarga, it does so only on
the grounds that without it “the continuity of the narration . . . is not hampered and appears in a better order”
(Shah 1975, 29).
125. I would not go as far as Oberlies, who argues that the Śvetāśvatara is from around 0–200 A.D. (1997,
86; 1988, 57–59), though he may be right (even if I doubt it) that it is younger than the BhG. Cf. Olivelle 1996,
252: “Its thought and vocabulary are close to that other theistic document, the Bhagavad Gītā.”
614 dharma

of life” refers to all four life-stages or just the householder stage, but, given the
liminal state of certain highly purified gleaners, I think most translators are
right to think of the former.126
So what are these high sagely audiences that take such delights: in this
Upaniṣad, a delight in those who have “passed beyond their order(s) of life as the
highest means of purification” from receiving a “supreme secret proclaimed dur-
ing a former aeon”—probably one of the earliest usages of kalpa in a Brahmanical
text127—that now “shines forth” to them after it was disclosed as brahman to
Śvetāśvatara by the grace of God; in the Gītā, in obtaining from Kṛsn ̣ ạ , “the
supreme knowledge of knowledges knowing which all the Munis have gone
from this world to supreme success,” having attained the “same nature as me”
such that they are unaffected through Dissolutions and Creations; in our current
example from the Mahābhārata, a delight in pralayic stories that provoke wonder
and surprise? What is interesting is how these bhakti texts do something bhakti
is famous for: constructing community, and with it audience. I think Manu
shows familiarity with this novel textual practice, without mentioning the term
saṅgha, to give his unbudging orthodoxy some similar kind of constituency. But
the term has this cachet elsewhere. The use of the term Ṛsị saṅgha would also
have counterparts in the wider sense of the Buddhist second refuge “in the
Sangha,” which is also and from the first an audience for texts, and in the use of
the same term for the first audiences of Tamil literature.
These universal listeners are not only the greatest of sages, enough of them
Vedic to make that part of their aura, but a model audience that listens for and
with a carefully constructed “us.” Indeed, in very broad terms, we can say that
they are the hosts of the texts, and in the Mahābhārata, sometimes the hosts of
subunits such as the UMS, and of course not just hosts of the texts but of their
readers since they set the model for receiving the texts into human lives and
homes and open questions of interpretation just by their overhearing presence,
not to mention specific interventions where they “do the gods’ work.” In the
UMS, the first thing to notice is that first Kṛṣṇa and then Śiva and Umā are
hosts to two successive Ṛṣisaṅghas. In the first case, Kṛṣṇa registers the

126. Given the liminal state of certain highly purified gleaners, I am not sure that Olivelle is right to dis-
count the assumption of “many scholars . . . that the expression atyāśramibhyaj . . . refers to those who had passed
beyond the four stages of life known as the āśramas,” maintaining, on the contrary, that “āśrama here means just
the householder life and the expression refers to ascetics who have moved beyond the householder life” (1998,
628 n.). As Olivelle indicates, there are later instances where term atyāśramin does hold “the possibility of living
beyond the āśramas.” In any case, the closing ŚU passage makes the differentiation mentioned, and might pro-
vide an early instance of this later āśrama-transcending meaning by its bhakti context, which is similar to that in
the UMS—for one thing, in stressing “the highest means of purification.”
127. Cf. two notes above: the ŚU may be no earlier than the Bhagavad Gītā. Olivelle, as quoted, translates it as
“age” rather than “aeon,” and, as often in the Mahābhārata, it is indeed hard to know what is meant. See chapter 6.
 dharma and bhakti 615

Ṛṣisaṅgha’s surprise at the miracle of the fire from his mouth, and encourages
their telling him a wondrous story that will last “long on earth, like writing
engraved on a mountain” (śaile lekhyam ivārpitam; 43d)—an interesting juxta-
position of oral and written modes of communication, and reminiscent of
Hanumān’s parting words to Rāma, stressing solely orality: “As long as I hear
Rāma-kathā on the face of the earth, so long will my breaths reside in my body”
(Rām 7.39.16).128 Like Hanumān, Kṛṣṇa and Rāma are both hosts to stories: in
Kṛṣṇa’s case, one about Umā and Śiva; in Rāma’s, his own. But Kṛṣṇa, as guest
and host, listens along with the Pāṇḍavas to countless other stories, including
others too about himself, as we noticed him doing earlier in this chapter
(see § C) when Mārkaṇḍeya sprang the revelation that the child he had just
described surviving the cosmic dissolution was none other than this Kṛṣṇa in
the Pāṇḍavas and Draupadī’s very company, “your ally Janārdana, . . . who sits
here as though at play,” and told the Pāṇḍavas and Draupadī they should “go to
him for refuge” (3.187.50–53). From the standpoint of the UMS, the Pāṇḍavas
and Draupadī have now f inally already taken refuge in Kṛṣṇa by becoming
“dedicated to Nārāyaṇa” after hearing the White Island story in the Nārāyaṇīya,129
again, with Kṛṣṇa-Janārdana present and listening. Meanwhile, in the case of
Umā and Śiva, it is their solicitation of their own attendant Ṛṣisaṅgha that
spurs Umā, once she has satisfied her curiosity about Śiva’s third eye, four
faces, Nandin, and Bhūtas, to ask what dharma is and how men learn of it (Mbh
13.128.21–23), which leads on to her question about Ṛṣidharma.
These highly reflexive and nearly metatextual considerations may seem
only to be connecting the dots between passages that have long been consid-
ered late, and moreover, pointing us away from the two Sanskrit epics’ main
stories. But that is only if we grant those isolative premises. As I have said,
Ṛṣidharma is all over the place in both texts, where it has very much to do not
only with hospitality and friendship but with a kind of intimacy and complicity
in these matters that is fostered by the texts themselves, and indeed by their
poets under the noms de plume of their Ṛṣi authors. Vālmīki composes the
Rāmāyaṇa and imparts it to Kuśa and Lava at his āśrama as nothing less than
the host of Sītā, and he comes with the twins to have them sing the Rāmāyaṇa
at Rāma’s Aśvamedha as a guest of Rāma, and then arrives there again with
Sītā for her final ordeal. Vyāsa is time and again a surprise visitor and occa-
sional host, and a nearly but not entirely silent and definitely honored guest as

128. On orality and writing, see Hiltebeitel 2005c. “Writing on a mountain” does, of course, suggest a
post-Aśokan verse, but that dating would surprise no one.
129. “Having heard this best of narratives, O Janamejaya, King Dharma and all his brothers became
dedicated to Nārāyaṇa” (12.326.121). See chapter 6 § B.
616 dharma

a sadasya at Janamejaya’s snake sacrifice where he is present as a great-great-great

grandfather of that host for the inner frame recital of his own composition.130 If
Vālmīki is intimate with his characters not only through his proximity to Sītā
but through a boon from Brahmā, Vyāsa is still more intimate—literally so
with two Kuru queens and one of their maidservants, and also with his charac-
ters’ thoughts, beginning with those of his mother, and including those of
Kṛṣṇa (Hiltebeitel 2001a, 32–91). Through these author-sages we are party to
thoughts of Rāma and Kṛṣṇa and the generations of heroes and heroines who
share their time on earth. Yet paradoxically, this intimacy and complicity gifted
by the poets to readers is nurtured in our separation, in that these gods, heroes,
and heroines lived in another time, and they are “not like us.”

F. Rāma and Kṛṣṇa as Guests, Hosts, and Friends

Let us then take a closer look at a few scenes that show the work of the poets
in creating this intimate camaraderie around hospitality and friendship.
I focus on examples of each from each epic, noting that hospitality and friend-
ship also often imply each other, and that both always imply separation. Each
also implies a politics of bhakti that maps differently in each epic: as a
master–servant dialectic in the Rāmāyaṇa centered on an exaltation of the
figure of the divine king; and, in the Mahābhārata, a politics of friend and foe.
Again, while never forgetting varṇāśramadharma and indeed reinforcing it,
these instances also open on to matters that play on its interstices and extend
well beyond its limits.
If one looks at the Rāmāyaṇa in terms of hospitality for and by Rāma, there
are two broad phases. In the first five Books one meets Rāma mostly as a guest
of forest sages, and as one who refuses to be hosted in the monkey capital of
Kiṣkindhā. But the tables begin to turn in Book 6, where he gives refuge to the
Rākṣasa Vibhīṣaṇa, a “good demon” who is a younger brother of Rāvaṇa. And
they are fully turned in Book 7 where, as king, Rāma finally gets to host just
about everybody—Rākṣasas, Monkeys, Ṛṣis, and kings—more than once,
including, as noted, Vālmīki, and ultimately even that intemperate hungry sage
Durvāsas whose demand for food signals Rāma and his brothers’ last days.
The first phase is interesting because it brings a neo-Vedic variety of
Ṛṣidharma right into the heart of the story to mark Rāma’s path and to provide

130. As we saw in § A of this chapter, the Harivaṃśa picks up on this presence to give Janamejaya the
opportunity to ask his last questions directly to Vyāsa. On his presence there in the Mahābhārata, see chapter
6 § B (n. 33); Hiltebeitel 2006a, 245–49.
 dharma and bhakti 617

him and his brothers,131 and soon enough Sītā as well, with mentoring or
guidance. This is Vālmīki’s idea, for only one of the great Vedic Ṛṣis Rāma
encounters, Vasiṣṭha, appears in the Mahābhārata’s Rāmopākhyāna. Moreover,
one of them, Agastya, is added to complete a relay system that overlaps with
three Ṛṣi groupings that Vālmīki would have known: the set of Seven Ṛṣis,
whom the Mahābhārata knows by now as the stars of the Big Dipper,132 plus
Agastya as the southern star Canopus; the same set as the eight eponymous
ancestors to whom all Brahmins trace their primary lineages or gotras;133 and
the group of eponymous Ṛṣi-poets of the oldest Family Books of the Ṛgveda.134
Vālmīki thereby presents himself and Rāma not only as contemporaries of the
most hallowed Vedic Ṛṣis; he maps these Ṛṣis from Ayodhyā deep into the for-
ests of India to guide Rāma from conception to the abduction of Sītā. For
Agastya, after hosting them, directs Rāma, Lakṣmaṇa, and Sītā further south to
the place where Sītā will be abducted.135 Four of these Ṛṣis host Rāma in their
hermitages: Viśvāmitra, Bharadvāja, Atri, and Agastya; and he goes with
Viśvāmitra to the former hermitage of Gautama to hear the story of Gautama’s
wife Ahalyā, who provides Rāma and company hospitality (ātithyam; 1.47.18ab)
before he reunites the couple after a long estrangement. Bharadvāja (2.48.13)
and Agastya explicitly provide Rāma hospitality, each calling him their “beloved
guest” (priya-atithi), as does another forest Ṛsị named Śarabhaṅga (3.4.25d).136 The
most telling instance is the last, where Agastya says to Rāma with joined hands:

King of the whole world, one who fares in dharma, a great chariot
warrior, a man offered reverence and esteem, you have come as my
beloved guest. (3.11.27)

131. Lakṣmaṇa is usually with Rāma when Rāma meets them; Bharata—once Rāma has gone into exile
with Lakṣmaṇa—meets Bharadvāja (ca. 2.66–80) independently, and so too deals independently with Vasiṣṭha
(2.107.6a–c; see also 2.96.24–29). Śatrughna also meets with Vālmīki in a passage that the Critical Edition
rejects even though it appears in all the manuscripts collated, when he stops at Vālmīki’s hermitage, hears the
twins’ elegant recitals, and promises that he and his army will keep their birth secret from Rāma (7, Appendix 1,
No. 9; Shah 1975, 26–27).
132. The Seven Ṛṣis as Big Dipper are Viśvāmitra, Gautama, Atri, Bharadvāja, Vasiṣṭha, Jamadagni (Rāma
Jāmadagnya’s deceased father), and Kaśyapa, ancestor of Ṛśyaśṛṅga who oversees Rāma’s conception. The seven
are also listed as the composers of Ṛgveda 9.67 and 10.137.
133. The same seven as in the previous note, plus Agastya. See Hiltebeitel 1977 for discussion of all these sets.
134. See chapter 3 § B: Viśvāmitra of Book 3; Gautama of Book 4; Atri of Book 5; Bharadvāja of Book 6; and
Vasiṣṭha of Book 7; plus Rāma Jāmadagnya, descended from Bhṛgu and the Bhṛgu poets of Book 2.
135. See Hiltebeitel 2009a, and for some discussion of these scenes and Rāmāyaṇa hospitality practices,
Brockington and Brockington 2006, xvi, 381–83.
136. The phrase has ordinary uses: Daśaratha welcomes Bharata’s maternal uncle Yudhājit with it (1.72.6),
and uses it to describe how the boy he killed by accident would come “like a welcome guest” (priyam ivātithim) to
feed his parents (2.58.28). But this second usage also reflects back on how Daśaratha yearns for the departed Rāma.
As we shall see, priyātithi is also used when the residents of Ayodhyā return to the city and lament the “beloved
guest” (!) who has left for exile (2.42.34).
618 dharma

As with the earthly gleaners described by Śiva, these Rāmāyaṇa sages do the
gods’ work and would seem to have more than an inkling of what it is. Once
Agastya recognizes his “beloved guest” even in exile as the king of the world,
he gives Rāma a bow of Viṣnụ (3.11.29–32). When Sītā is soon abducted, Brahmā
speaks of it as something that had to be done and the Daṇdạ ka Forest Ṛsị s are
“thrilled and agitated” (prahṛṣṭā vyathitāś ca) at the sight (3.50.10–11).137
As to friendship in the Rāmāyaṇa, it is the second book that brings the sub-
ject to the fore. Here we meet the first articulation of sakhi as a “friend to
master” (sakhi to bhartṛ) relation specifically focused on Rāma in his dealings
with Guha, a low caste or “tribal” Niṣāda.138 Otherwise, as Rāma emerges into
view in his court and palace life as he is about to be made heir-apparent, sakhi
is never used for any relation he has with anyone else in this Book. In departing
Ayodhyā, Rāma leaves no friends behind—that is, sakhi friends. What he leaves
behind are wellwishers, suhṛds. While suhṛd is used fairly frequently, what is
noteworthy is that although Rāma has suhṛds, when they are around him they
are impersonal, never named.139 For example, “As for Rāma’s wellwishers, they
were all bewildered: crushed by the weight of their grief, they could not rise
from where they had fallen” (2.36.16). The wellwishers are last in the run of
those bidding Rāma adieu, mentioned just before the city, “Ayodhyā, with all
its hosts of soldiers and herds and horses and elephants . . .”—after which Rāma
disappears from the sight of those remaining in the capital (36.17–37.2). After
following Rāma’s departure as far as they could, they would be among the res-
idents who return to the city and say, “Every hill and grove Rāma visits will treat
him as a beloved guest (priyātithi) and not fail to accord him hospitality”
(2.42.34). Only when Sītā is abducted does Rāma begin to make sakhi friends
among animals, most notably with Sugrīva,140 through whom he also inspires

137. As mentioned above in § A. I believe this is right, although Pollock 1991, 196 has “shuddered and
trembled to see” here. Although Pollock 1984 worked out the implications of the boon to Rāvaṇa that requires
Rāma not to know his divinity until he has killed him, he does not see them in passages like this.
138. See 2.44.9b and 14a; 2.78.5ab and 11cd; 79.2b, 80.7a, 81.16b: Guha is friend and Rāma specifically
master in the second and third of these citations. Ekalavya is also a Niṣāda; see chapter 9.
139. See Rām 2.3.29ab; 2.5.12ab; 2.15.1, 4ab and 9ab; 2.16.55 and 61; 2.31.4 (he has taken leave here of “all
his suhṛds”). Lakṣmaṇa likewise has his suhṛjjana, a depersonalized throng of wellwishers (28.11cd and 15ab),
whom Rāma bids him say adieu to once Lakṣmaṇa has pleased Rāma with his detailed promise of services to him
and Sītā in the forest. Bharata, however, at least in Rājagha among his mother’s people of Kekaya, has sakhis who
can speak to him as his suhṛds (2.63.5–7). This is no doubt to nuance the fact that these “friends of Bharata” are
still, as far as the narrative has developed, potential rivals of Rāma, for Bharata’s own loyalty to (and saubhrātra
with) Rāma is being at this point still held up to question.
140. Beginning with the vulture Jaṭāyus and later including Jaṭāyus’s brother Saṃpāti, both former sakhis
of Daśaratha (3.13.41; 4.55.20; 56 9 and 12). As with Sugrīva, with whom Rāma “makes friends” on the advice of
Kabandha (see 3.67.30; 4.5.7; 4.8.2d and 39cd; cf. the Rāmopākhyana’s usages at Mbh 3.263.39d; 264.11c, 57a),
sakhi at first means “pact” friend, but in Sugrīva’s case it becomes also “intimate.” On other terms (notably vayasya,
used artfully by Vālmīki for this budding friendship with Sugrīva, but not used in the Rāmopākhyāna), see
Hiltebeitel 2009a and chapter 9 § D.1.
 dharma and bhakti 619

the devotion of Hanumān. As befits the Rāmāyaṇa’s master–servant politics, it

restricts the role of friend to low-status subordinates, and separates the roles of
friend and devotee, leaving the one no less a servant than the other.
As to the Mahābhārata, scenes of friendship and hospitality are more
numerous and also harder to separate from each other. Things are also more
complex in other ways. As with Rāma, one would want to track the hosting and
guesting relation between Kṛsn ̣ ạ and the Ṛsị s and the semantics of Kṛsn ̣ ạ ’s
“friendship” throughout the text as they relate to movements within and away
from the central kingdom and its capital, or in the Mahābhārata’s case, its two
capitals. Indeed, even before Kṛsn ̣ ạ is mentioned in the action, we saw how the
Mount Śataśṛnġ a Ṛsị s, all cued to the divine plan, and some of them, at least,
celestial, brought the Pānd ̣ ạ va boys to Hāstinapura (see chapter 8 §§ E–H). With
̣ ạ then coming into the action, there are many scenes with bhakti overtones
Kṛsn
that clearly relate hospitality and friendship to dharma: some outside Kuru land,
some in it, but all implying Kṛsn ̣ ạ ’s involvement with the Kurus. Kṛsn ̣ ạ wel-
comes Arjuna at Dvārakā by helping him abduct his sister Subhadrā (Mbh
1.211–12), who will be the mother of Abhimanyu through whom the lunar
dynasty will have continuity into the Kali yuga. With his plan to recover the “mid-
dle country” (avanīṃ madhyamām) from Jarāsaṃdha’s universal sovereignty
(sāṃrājyam) (2.13.7–8), Kṛsn ̣ ạ breaches Jarāsaṃdha’s hospitality at Magadha (see
chapter 13). Bhīṣma selects Kṛsn ̣ ạ to receive the “guest-gift” (arghya) at
Yudhiṣṭhira’s Rājasūya because he is the Supreme Person (puruṣottama;
2.42.24d). When Nārada leads the Ṛsị s there in time for the guest-gift (2.33.1–21),
he perceives there, in the very midst of the Pānd ̣ ạ va–Kuru hall at Indraprastha,
that Hari Nārāyaṇa who has “become man” and is the lord praised by sacrifices
(20) will lead those assembled to their doom (2.34–42). Kṛsn ̣ ạ is then absent
at the dice match in the Kuru court at Hāstinapura, where Draupadī is staked
and miraculously saved from the attempt to disrobe her.141 As events lead
on to war, in a scene that sets the stage for Kurukṣetra itself, Kṛsn ̣ ạ awakens
at Dvārakā to find Arjuna and Duryodhana seeking boons from him. Duryodhana
claims that Kṛsṇ ạ has an equal friendship (samam . . . sakhyam) and relationship
(sambandhakam) to both himself and Arjuna (5.7.10), which Kṛsn ̣ ạ (that “friend
alike to friend and foe”) does not deny, but rather devises the order of choosing
so that Arjuna gets the first choice between Kṛsn ̣ ạ as a noncombatant charioteer

141. For what must remain for now my best attempts to understand that scene, see Hiltebeitel [1976]
1990, 86–101 on Kṛṣṇa’s absence from the dice match as structured by “the epic scenarists” (93) to carry forward
the divine plan, and 2001a, 241–59, leaving open the question of whether Draupadī at least called on Kṛṣṇa in his
absence, during her scene of distress, since both of them refer to her doing so later (5.58.21; 80.23–26). See now
Bhattacharya 2009: despite recognizing evidence to the contrary outside the episode at Mbh 9.58.10, which he
has no good explanation for (96), he wishes to argue that Draupadī was never disrobed at all.
620 dharma

(which Arjuna chooses) and a whole army division of Kṛsn ̣ ạ ’s Nārāyaṇa Gopā
warriors (which seems to content Duryodhana). 142

Unlike Rāma, Kṛṣṇa is never a king with a need for subordinates or gran-
diose hosting obligations;143 and he tends not to play things straight. Noting
again that many of these scenes, and others, do not allow us to forget that Kṛṣṇa
is the special friend of Arjuna and Draupadī, I will choose examples from
Kṛṣṇa’s embassy to the Kuru court in Book 5. For this, we will need to remember
that Karṇa is the Pāṇḍavas’ arch foe, and recall that Vidura incarnates Dharma
and often speaks for dharma in the Kuru court.
Here, for once, as noted above where Yudhiṣṭhira and Vidura both warn
̣ ạ about “descending” into the enemy camp, Kṛsn
Kṛsn ̣ ạ is being hosted by the
Kauravas. The lengthy sub-book on Kṛsn ̣ ạ ’s embassy—“The Coming of the
Lord” (Bhagavatyānaparvan)—is filled with moments that could illustrate our
points, including Dhṛtarāsṭ̣ ra’s raptures at the thought of giving Kṛsn ̣ ạ fabulous
gifts and displays of welcome (5.84); Vidura’s advice that it would be enough for
Dhṛtarāsṭ̣ ra to give Kṛsṇ ạ “a beloved’s hospitality” (5.85.14) as he deserves—just
a jar of water to drink, water to wash his feet, an inquiry into his health, and
what he (supposedly) really wants, peace (85.13–16); and Duryodhana’s plans to
capture Kṛsn ̣ ạ (86.13). But let us focus in on the Ṛsị s. As Kṛsn ̣ ạ leaves the
Pānḍ ạ vas’ camp, various birds circle auspiciously above him, and then some of
the great Brahma-Ṛsị s and Divine Ṛsị s (brahmadevarṣayaḥ)—Vasiṣṭha, Nārada,
Vālmīkā (sic), and Bhṛgu among them—circumambulate him and perform
smokeless rites with mantras on his behalf (5.81.26–29). We do not hear where
they have come from until he has set out:

Along the path (adhvani) Keśava of strong arms saw the Ṛṣis blazing
with brahmic luster, who stood on both sides of the road (sthitān
ubhayataḥ pathi). Quickly descending from the chariot (avatīrya
rathāt), Janārdana, having bowed, honoring all these Ṛṣis duly, said,
“Is good health in the worlds, is dharma observed, do the other three
varṇas abide under the Brahmins’ rule (śāsane)?” Paying them
homage (pūjā), Madhusūdana proclaimed, “Where have you reached
perfection, your worships (bhagavantaḥ)? What path has brought you

142. All such scenes would in principle be rejected from Fitzgerald’s “main Mahābhārata” (2003, 811), on
which see chapter 1 § B. According to Fitzgerald, bhakti runs were “probably added some time between 50 CE
and 350 CE,” including “the essential portions of the Bhagavad Gītā, all episodes that elaborate some theme of
devotion to Viṣṇu, Śiva, or Kṛṣṇa,” including these two and “several highly polished expressions of Kṛṣṇa bhakti
in the narrative wake of Yudhiṣṭhira’s abhiṣeka” between 12.40 and 12.56 (2006, 272–73).
143. At Mbh 13.16.8cd, however, Kṛṣṇa does get an eighth boon from Umā that enables him to feed 7,000
guests daily at his palace.
 dharma and bhakti 621

here? What is to be done for your worships? What can I do for you?
For what purpose have your worships come down to earth?
(kenārthenopasaṃprāptā bhagavanto mahītalam).” (5.81.61–64)

This show of roadside hospitality to the celestial Ṛṣis now turns into a remark-
able scene of friendship: Rāma Jāmadagnya, “embracing Govinda as an old
friend in good conduct (purā sucarite sakhā)” (65cd), speaks for them. Having
witnessed the old battles of the Gods and the Demons, the Ṛṣis are “everywhere
(sarvatas) wishing to see the gathering of the royal Kṣatra, the kings sitting in
the hall, and yourself speaking the truth, Janārdana. We are coming to watch
this grand spectacle, Keśava” (67–68). Jāmadagnya in particular wishes to hear
what Kṛṣṇa, Bhīṣma, Droṇa, Vidura and others will say, and closes, “Go unhin-
dered, hero, we shall see you in the hall” (72cd). Note the depth of precedent
behind this exchange: not only have the Ṛṣis witnessed the old Deva–Asura
wars; there is the veiled reference to an old sakhi “friendship” between Kṛṣṇa
and Rāma Jāmadagnya. This is probably an affinity in their tasks, since this
Rāma brought about the “destruction of the Kṣatriyas,” as Kṛṣṇa will now do as
well. Or did they meet before when Kṛṣṇa was Rāma?
As we have noted, Kṛṣṇa’s welcome by the Kurus is no simple matter: he
rejects Duryodhana’s hospitality and spends the night at Vidura’s. But finally
Kṛṣṇa enters the hall—in the same capital at which Draupadī was disrobed,
if not the same hall:

There standing amidst the kings, the foe-scorcher Dāśārha, conqueror

of enemy cities, saw the Ṛsị s hovering in the sky (apaśyad
antarikṣasthān ṛsị̄ n). While watching them, headed by Nārada, Dāśārha
said softly to Bhīṣma, “King, the Ṛsị s have come to watch the earthly
assembly, and should be invited and honored with seats and full
hospitality (satkāreṇa ca bhūyasā). No one can sit before they are seated.
Let homage (pūjā) be paid to these Munis whose souls have been
perfected.” Śāṃtanava, seeing the Ṛsị s arrived at the gate of the hall,
hurriedly ordered the servants: “Seats!” and they brought large and
wide smooth seats that sparkled with gold and gems. (5.92.40–45)

Clearly, the poets leave questions that an attentive reader or Ṛṣisaṅgha might
ask: Was the hall open to the sky? Did Bhīṣma only see the Ṛṣis when they got
to the gate? How many seats did the servants provide? How many angels can
dance on the head of a pin? But this is serious business.144 Once everyone falls

144. Cf. the fantastic provision of theologically suitable seats for vast celestial and earthly audiences to hear
the dharma in the Vimalakīrti Sūtra (Watson 1997, 75–78, 115, 121–22),
622 dharma

silent looking at the elixir of the dark yellow-robed Lord (51–53) in the middle of
the court, Kṛṣṇa makes his first speech, beginning, “Let there be peace . . .”
(93.3ab).
The Ṛsị s are not silent in this assembly. Rāma Jāmadagnya warns Duryodhana
with the story of the ancient king Dambhodbhava who foolishly challenged the
Ṛsị s Nara and Nārāyaṇa at their Badarī hermitage (5.94). Other Ṛsị s follow with
further stories. Then Kṛsn ̣ ạ addresses Duryodhana directly, as do others, including
his mother, all to no avail, before Duryodhana tries to capture Kṛsn ̣ ạ . Kṛsn
̣ ạ then
gives the Ṛsị s, along with only Droṇa, Bhīṣma, Vidura, and Saṃjaya, the divine
eye to see his theophany (5.129.13–14), and quits the court. Having met with the
Pānd ̣ ạ vas’ mother Kuntī, who sends words to her sons through him, his last
business before going back to the Pānd ̣ ạ vas is with Karṇa.
This scene at the end of Kṛṣṇa’s embassy illustrates our whole nexus.
Dhṛtarāṣṭra said, “Before he rode out amidst princes and councilors, O Saṃjaya,
Madhusūdana had Karṇa mount his chariot”—a kind of hospitality, I believe,
for two men away from home. Dhṛtarāṣṭra asks, “What consolations (kāni
sāntvāni)”145 did Kṛṣṇa-Govinda offer with his “voice roaring like a flood or a
cloud” to the Sūta’s son, “whether gently or sharply” (mṛdu vā yadi vā tīkṣṇam)
(5.138.1–3). Saṃjaya replies, “Hear from me” what the two said “in the course
of their conversing, in words that were smooth (ślakṣṇāni) and gentle (mṛdūni),
dear (priyāṇi), joined with dharma (dharmayuktāni), truthful (satyāni), helpful
(hitāni), and to be cherished in the heart (hṛdayagrahaṇiyāni)” (4–5). Recounting
Karṇa’s true origins, Kṛṣṇa tells him he is legally a Pāṇḍava—on his father’s
side related to the Pāṇḍavas, on his mother’s to the Andhakas and Vṛṣṇis—and
offers him the kingship. The Pāṇḍavas and the Andhakas and Vṛṣṇis will clasp
his feet. Draupadī will make him her sixth. Hospitality indeed! Kṛṣṇa himself
will consecrate him. As Karṇa rides the royal chariot, Yudhiṣṭhira, as his heir-
apparent, will fan him. Arjuna will drive the chariot. Karṇa will have a whole
new life, on which Kṛṣṇa concludes: “Your allies shall shudder with joy, your
enemies with fear. Today let there be good brotherhood (saubhrātram) between
you and your Pāṇḍava brothers!” (138.9–28). Both Kṛṣṇa and Karṇa would
know how these changes for Karṇa would also change things entirely for
Karṇa’s real friend (sakhi), Duryodhana, and for the larger Pāṇḍava–Kaurava
saubhrātra-“brotherhood,” of whom Kṛṣṇa says nothing here.146 Karṇa does not
doubt that Kṛṣṇa is a wellwisher speaking from love or affection, and further,
that he speaks to Karṇa’s best interests “out of friendship” (139.1).

145. Conciliations, mild words? Van Buitenen has “blandishments” (1978, 444).
146. Kuntī too says to Karṇa, just after this: “Let the Kurus today witness the meeting of Karṇa and Arjuna
in a spirit of brotherhood (saubhrātreṇa)” (5.143.9).
 dharma and bhakti 623

The theme of these words being heartfelt is then carried through to the
parting words of their meeting, where Kṛṣṇa responds to Karṇa’s explanation
of why he must refuse. “Kṛṣṇa said, ‘Of a certainty the destruction of the earth
is now near, for my words do not touch your heart, Karṇa (tathā hi me vacaḥ
karṇa na upaiti hṛdayaṃ tava). When the destruction of all creatures is at hand,
bad policy disguised as good does not crawl off from the heart (anayo
nayasaṃkāśo hṛdayān nāpasarpati)” (141.43–44). Kṛṣṇa commends Karṇa for
holding to good policy from his heart, despite rejecting Kṛṣṇa’s offer. As they
part, Karṇa “clasped Mādhava tightly (pariṣvajya ca pīḍitam)” (47b).147 As befits
the Mahābhārata’s bhakti politics of friend and foe, Kṛṣṇa and Karṇa’s embrace
extends the circle of Kṛṣṇa’s men and women friends to include even the sworn
enemy, and, as with Arjuna, combines the roles of friend and devotee.
Now as Saṃjaya told the blind old Kaurava king, Kṛṣṇa’s words in this dia-
logue were “smooth and gentle, dear, joined with dharma, truthful, helpful,
and to be cherished in the heart.” In whose heart? Karṇa’s, whose words are
equally heartfelt? Dhṛtarāṣṭra’s, as first listener? The hearts of the seated
attendees listening to the Mahābhārata’s first recital at the Snake Sacrifice of
Pāṇḍavas’ great grandson? The hearts of the great Ṛṣis listening to its retelling
in the Naimiṣa Forest? All listeners and readers? The phrase puts big asterisks
to these heartfelt words, which provide one of those points where we can feel
that the notion of the “wellwisher” ultimately extends to the bhakti community
of readers that the Mahābhārata seeks to create as its audience of aficionados:
that is, what classical Indian aesthetic theory calls its sahṛdayas, those who
appreciate a work of art because they are “with it at heart.”
In conclusion, we come back to the necessities and niceties of textual dis-
cretion in portraying the hiddenness of gods among men, which constrains
even Kṛṣṇa to operate within human limits, and in Rāma’s case is tacitly struc-
tured into the Rāmāyaṇa around the boon that Rāvaṇa can be slain only by a
man. Bhakti and dharma are in the sinews of these texts as we have them.
Positing political ideologies or class interests as prior to their bhakti can only
dim what they are, and positing pre-Mauryan oral precursors with no evidence
that their formulations of dharma could be that early is to beg endless questions
in a vacuum.148 Rather than a “rage” at the loss of Brahmin privilege under

147. This embrace may recall the parallels between Sugrīva and Karṇa in the killings of Vālin and Droṇa
discussed in chapter 9, for Rāma and Sugrīva also seal their friendship with such a close embrace (paryaṣvajata
pīḍitam; Rām 4.5.13)—in each epic, an embrace between Viṣnu ̣ ’s incarnation and a son of the Sun god. As with
Rāma and Sugrīva, the embrace of Kṛsn ̣ ạ and Karṇa sets the conditions under which each will do the other’s will.
148. As evident in earlier chapters, I am in working concurrence with Olivelle’s post-Aśokan dating of the
dharmasūtras (2006b), and do not consider the epics’ understanding of dharma to be older than the dharmasūtras. The
problem arises, for instance, where Fitzgerald says that “Aśoka preempted the brahmin monopoly on the teaching of
dharma” (2006a, 276). As Olivelle’s careful reading (2004a) of pre-dharmasūtra Vedic and sūtra texts demonstrates,
624 dharma

Aśoka and the Mauryas,149 I believe that the rapport the Mahābhārata sustains
between bhakti and dharma reflects a later, post-Mauryan, sly and confident
sense of taking over the game. For the Rāmāyaṇa, it would appear to be the
same game disambiguated by adding live Vedic Brahmins to exemplify dharma
as hosts and guests through the hero’s long career. In both epics, however,
bhakti is a trump card played discretely and not that often (though certainly
more often in the Mahābhārata). And it is played with a deck stacked with Ṛṣis
who, by “doing the gods’ work,” map a new Brahmanical dharma over time and
across the land, making it familiar in every sense. It is this combination that
could give life to a king who has listened well enough to Kṛṣṇa and even better
to the Ṛṣis, as is the case with Yudhiṣṭhira the son of Dharma; or to a king who,
as a hidden god himself, could be the Vedic Ṛṣis’ “beloved guest.” As we shall
now see in chapter 13, Aśvaghoṣa could respond to this game from a familiarity
with both epics, to map bhakti into the Buddha’s quest for the “true dharma.”

there is no evidence of Brahmins “teaching dharma” or monopolizing that topic in them. Moreover, even in the
dharmasūtras the monopoly that Brahmins assert is not in teaching dharma but, as we have seen repeatedly since
chapter 5, in the ritual sphere of teaching Vedic recitation (adhyāpanam), performing sacrifices, and receiving gifts—
the three “jobs” (karmas) reserved for them.
149. Fitzgerald posits a post-Aśokan “deep and bitter political rage at the center of the Māhabhārata” (2001, 85).
13
Aśvaghoṣa’s Buddhacarita
A Buddhist Reading of the Sanskrit Epics
and Their Treatments of Dharma

One finds mounting evidence that classical Buddhist and

Brahmanical dharma texts use the term dharma knowingly as
regards each others’ usages.1 Yet once we are past the Nikāyas,
both traditions were sparing in direct references to each other.
The Sanskrit epics, as we have seen, adopted what can be called,
at best, a civil silence toward Buddhism, and for the most part
Buddhist texts responded in kind.2 This chapter will deal with a
major exception: the “Life of the Buddha” by the Sanskrit poet
Aśvaghoṣa. By the first or second-century CE, Aśvaghoṣa’s time,
a Brahmanical–Buddhist interface had gone on for roughly half a
millennium without any known text being direct about it since the
Nikāyas. No doubt that is where Aśvaghoṣa saw his inspiration to
reengage. Yet he did not do so under the same conditions as the
Buddha or the Nikāyas. Biardeau proposes a good angle from which
to catch up quickly on the new conditions. By Aśvaghoṣa’s time,
two forms of bhakti, Brahmanical and Buddhist, were developing

1. See chapters 4, 6, and 7 on Buddhist texts; 5, 6, and 12 § C Brahmanical ones. This chapter
revises Hiltebeitel 2006b, which, as the first chapter drafted for this book, was written to be its last one.
2. Though with exceptions, most notably in the Jātakas which have parallel material that is
difficult to date relative to the Sanskrit epics, but which often looks parodic or displaced; see
Hiltebeitel 1976 [1990], 64–67; Gombrich 1985 as mentioned in the text. See also Hiltebeitel 2005a,
459 n. 15 on the Kuṣan ̄ ạ period “Spitzer manuscript” found in east Turkestan and most recently
presented by Franco 2004.
626 dharma

beside each other, in the latter case among Buddhists who were “for the most
part of Indian origins and inserted in the society of castes,” fully “at home”
there, with no one desiring their departure, despite the “Brahmanical mani-
festo” provided by the two epics (Biardeau 2002, 2: 776). As we shall see, the
description fits Aśvaghoṣa quite well.

A. Aśvaghoṣa’s Buddhacarita

Aśvaghoṣa’s Buddhacarita is worth looking at not only for what it says directly
about Buddhism and the Sanskrit epics but about what his treatment of them
might be able to tell us about how dharma, and particularly royal dharma,
remained a central topic of this intertextual and interreligious game. In recog-
nizing that Aśvaghoṣa focuses his Buddhist narrative on dharma, and positing
that one of the main things that would have interested him in the Brahmanical
epics would have been their treatment of dharma, we might also be able to
improve upon earlier treatments of the question of what kind of Mahābhārata
and Rāmāyaṇa Aśvaghoṣa would have been responding to. This means that we
first need to look more closely into Aśvaghoṣa’s likely dates.
Étienne Lamotte upholds Chinese traditions that Aśvaghoṣa was “contempo-
rary with Kaniṣka” whom Lamotte dates at “ca. 128–51” CE ([1958] 1988, 591, 655).
However, as Lamotte puts it, this association of Aśvaghoṣa with Kaniṣka comes
from fourth and fifth century “Chinese documentation on Indian origins of poor
quality and without historical interest” ([1958] 1988, 698), so it is not clear why he
upholds these sources on the connection of Aśvaghoṣa with Kaniṣka in opposition
to others’ skepticism about it. E. H. Johnston (2004), who after nearly seventy
years still holds place as the best introduction to Aśvaghoṣa,3 prefers a pre-Kaniṣka
date for him, noting that Chinese tradition made Aśvaghoṣa into an exorcist saint
(2004, xv, xxxv). Taking Kaniṣka’s likely date to be ca. 75–125 CE, Johnston places
Aśvaghoṣa “between 50 B.C. and 100 A.D., with a preference for the first half of
the first century A.D.” (xvii).4 In offering one of the first recent discussions of the

3. Johnston’s monumental ten-year study of the Buddhacarita provides a critically edited text through most of
the first fourteen cantos (Part 1); translation of those cantos with lengthy Introduction plus extensive notes on the text
and the translations (Part 2); and translation of the last fourteen cantos mainly from the Tibetan, with an attempted
rough reconstruction of the Sanskrit from both fifth-century Chinese and later Tibetan translations (Part 3). Reference
to “Parts” will be made only to Part 1. Although Olivelle 2008 offers an important new translation of the extant Sanskrit
portions and an excellent discussion of the Buddhacarita’s interest in dharma generously citing my treatment in
Hiltebeitel 2006a, I will continue to cite or work mainly from Johnston’s translation in this chapter, turning to
Olivelle’s translation and introduction only where new insights or clarity arise—as will be the case especially in § F.2
on “Buddhist Mokṣadharma.”
4. Johnston throughout speaks of Aśvaghoṣa as a first-century CE poet; see 2004, xiii–xvii; xxxviii, xl. He
had changed his view since translating the Saundarananda; see Johnston 1928, vi: “generally agreed to have
flourished early in the second century A.D.”
 aśvaghoṣ a ’s buddhacarita 627

Buddhacarita, the 2005 fifth edition of Buddhist Religions: A Historical Introduction

by Richard Robinson, Willard Johnson, and Thanissaro Bhikkhu (a.k.a. Geoffrey
DeGraff) likewise gives Kaniṣka’s date as “late first or early second century C.E.”
(76), and treats Aśvaghoṣa as preceding him in “approximately the first century
C.E.” (5). This edition, which both refines and considerably extends (8–11) what
the fourth edition of 1997 had to say about Aśvaghoṣa and the Buddhacarita, con-
textualizes Aśvaghoṣa as a contributor to a first-century turn to writing affecting
both Theravāda and Sanskrit Buddhist texts, a turn that further “parallels a con-
temporary development in Indian fine literature” in which “some of the greatest
poets and prose stylists of this period—Aśvaghoṣa, Mātṛceta, and Ārya Śūra—
[were] Buddhist monks” (77). Richard Salomon points to “inscriptional specimens
of kāvya . . . now available as early as the beginning of the first century,” which are
“consistent with the evidence of literary sources themselves, notably the works of
Aśvaghoṣa which point toward a flourishing kāvya in the first century A.D.”5 Most
intriguing to me has been Giuliano Boccali’s observation that a totally new kāvya
sensibility can be noticed when both Aśvaghoṣa (see Buddhacarita [BC] 4.30) and
Hāla in the Sattasaī (the oldest anthology of Prakrit poems), both around the same
time—which for Boccali is the first-century CE6—introduce women pretending to
stumble to attract the hero’s attention: something, Boccali noted, that we would
not imagine in any prior literature, including the Sanskrit epics, which are “totally
lacking in such stereotypes of love.”7
In brief, although there are those who lean toward a second-century dating,8
there is a good weight of varied scholarly considerations favoring the first
century. Moreover, Johnston shows that the Tibetan and especially the fifth-
century CE Chinese translator must have had a Buddhacarita that does not
differ much from the oldest surviving Sanskrit manuscript, which he dates to
1300 +/−50 (vii). Lamotte ([1958] 1988, 656) and Beal (1968) date this Chinese
translation by Dharmakṣema or Dharmarakṣa to around 420, establishing that
a quite stable Buddhacarita, like the one we have, had come to China at least by
the early fifth century. This guarantees that virtually all the verses of the oldest
Sanskrit manuscript (and the three others used by Cowell ([1894] 1968) that,

5. Salomon 1998, 233. See similarly Dimock, Gerow, Naim, Ramanujan, Roadarmel, and van Buitenen 1974,
119, connecting Aśvaghoṣa with first-century CE praśasti inscriptions and developments in kāvya (the author of this
segment is Edwin Gerow).
6. Martha Selby (2003, xxvi) dates Hāla’s reign at Pratiṣṭhana/Paithan to 20–24 CE. This Śatavāhana
capital is not to be confused with the Pratiṣṭhāna mentioned as the first lunar dynasty capital in chapter 12 § A.
7. Guliano Boccali, “Introduction” to concluding roundtable discussion on “Origins of Mahākāvya:
Problems and Perspectives,” Origins of Mahākāvya: International Seminar, Università degli Studi de Milano,
Milan, June 4–5, 2004.
8. Olivelle 1993, 121, having first accepted Johnston’s first-century CE date, more recently says Aśvaghoṣa is
“generally assigned to the 1st–2nd centuries C.E.” (2005, 24; see also Olivelle 2008, xix–xxv); Strong offers “second
century A.D.?” (1983, 31).
628 dharma

according to Johnston, derive from it) would “be either part of the original or
old interpolations” (2004, Part 1, viii). This does not deter Johnston from
devoting a page to “almost certain” and “doubtful” interpolations (Part 1, xvii–xviii),
but these are neither numerous nor extensive.

B. The Centrality of Dharma in Aśvaghoṣa’s Buddhacarita

It is a surprising point to have to make that Aśvaghoṣa would be centrally

concerned with dharma, but others, with the exception recently of Olivelle
(2008), seem to have missed it. According to Robinson et al., “Aśvaghoṣa’s main
concern in portraying the Buddha’s teaching career is to refute the various
Brahmanical positions extant in his day. Thus he emphasizes the philosophical
side of the Buddha’s teaching role almost—albeit not entirely—to the exclusion
of the religious side” (2005, 23). It is important to their presentation that “[t]he
Buddhacarita is among the earliest extant texts to explicitly state that there is no
self.”9 According to Lamotte, Aśvaghoṣa’s “Buddhacarita and Saundarananda
are on a level with the classical mahākāvya. The scholastic parts remain faithful
to the traditional vocabulary and phraseology; the narrative and descriptive parts
abound in brilliant images, figures of style, complicated metres, and learned
grammatical forms. The author seems to have wanted to dazzle his less
knowledgeable colleagues by fully deploying his brahmanical virtuosity. His
search for effect and his conciseness, taken almost to the point of unintelligibility,
give the impression of a decadent art” (Lamotte [1958] 1988, 591–92).
Johnston acknowledges Aśvaghoṣa’s interest in refuting Brahmanical traditions,
especially with regard to the proto-Sāṃkhya that Aśvaghoṣa puts into the mouth
of Arādạ Kālāma (2004, lvi–lxii), and he discusses at length Aśvaghoṣa’s standing
as a kāvya poet (lxxix ff.). But Johnston’s Aśvaghoṣa is more multifaceted.
Johnston underscores how “the breath of bhakti” (xxvi) animates certain passages
emphasizing śraddhā or “faith,”10 but with a restraint toward the miraculous11—
“more by devotion to the Buddha and a respect for scripture than a love for the

9. Robinson et al. 2005, 91; see B 14.84; 15.80–86 (teaching to Śreṇya-Bimbisāra), 26.18.
10. Most of these are in the Saundarananda, but he also cites Canto 27 in the Buddhacarita (Johnston
2004, xxv–xxvii). See also xxxiv, xxxvii, xcvi, and Aśvaghoṣa’s interest in pari-pratyaya, “reliance on others”
(xxxiv–xxxv), which Johnston relates to Mahādeva’s five points about the arhat, and to the Mahāsaṅghikas
(Johnston 2004, xxvii–xxxi), a sect that revered Mahā-Kaśyapa (xxvii, xxviii), to whom Aśvaghoṣa gives major
billing.
11. See Johnston 2004, xxxix and Buddhacarita 1.11, where, rather than mention the Buddha’s descent
from the Tuṣita heaven, one reads cyutaḥ khād iva, “as if he came from the sky.” Cf. Saundarananda 2.48–50,
where such birth miracles are mentioned.
 aśvaghoṣ a ’s buddhacarita 629

marvelous.”12 Here too Johnston alludes to Aśvaghoṣa’s knowledge of texts,

a point I will turn to in the next section. But Johnston never once mentions a
concern with dharma, coming close only once with a statement that Aśvaghoṣa’s
“standpoint remains entirely moral, free from any attempt at metaphysical spec-
ulation” (2004, xli; my italics). Scholarly work on dharma by Johnston’s time
seems to have been rather scattered, and he might have had a somewhat nebu-
lous ahistorical view of the term that many still have today.
Since Aśvaghoṣa’s interest in the topic of dharma will remain central to
this chapter, I will limit discussion for now to two points that will demonstrate,
hopefully sufficiently, that the unfolding of dharma from a Buddhist perspec-
tive is probably Aśvaghoṣa’s most central concern. For the first of these, let me
just say quickly that Aśvaghoṣa clearly makes it his task to attempt a virtuoso
rehearsal and contextualization of all the varied Buddhist and Brahmanical uses
and meanings of dharma likely to have been known to him. Thus on the
Buddhist side he treats many of the basic Buddhist meanings of dharma that
were noted in chapter 4, offering precise moments for dharmas (plural) as “ele-
ments of existence”13 and for dharma as “inherent quality,”14 as well as for such
staples as the saddharma or “true dharma,”15 the dharmacakrapravartana or
“Turning of the Wheel of the Dharma” (BC 15.54–44), and even the dharmakāya
(24.10). And on the Brahmanical side, while giving direct reference to varṇa

12. Johnston 2004, xxxix–xl; see, for example, BC 6.68, describing the groom Chandaka’s return:
“Sometimes he brooded and sometimes he lamented, sometimes he stumbled and sometimes he fell. So jour-
neying in grief under the force of his devotion (bhaktivaśena), he performed many actions along the road in complete
abandon.” The passage combines kāvya style, used earlier with the smitten women, with viraha bhakti, with a
result that Chandaka acts much like a Gopī. The opening of the same Canto at 6.5–8 combines with this end to
make it a bhakti set piece. That Aśvaghoṣa recognizes such conventions is an indication that they are established
by the time of his composition. See similarly 9.8 and 9.80–82 (a set piece on rājabhakti as inadvertent buddha-
bhakti by two Brahmins). On the “double sense” of bhakti in 4.32, see Johnston 49 n. 32, in agreement with
Gawronski 1914–15, 26.
13. See especially BC 12.106, in which the Buddha is reflecting just before his five companions leave him
and he goes to sit under the bodhi tree: “By the practice of trance those dharmas are obtained through which is
won the highest, peaceful stage, so hard to reach, which is ageless and deathless (dhyānapravartanād dharmāḥ
prāpyante yair āpyate/ durlabhaṃ śāntam ajaram paraṃ tad amṛtaṃ padam).” As Johnston indicates “The refer-
ence is to the bodhipakṣika dharmas” (2004, 184, n. 106). This is I believe the first usage in the text of the technical
sense of dharmas in the plural. Johnston also reconstructs this plural usage from the Tibetan and Chinese trans-
lations also at 17.18 and 24.27.
14. See BC 12.70, where the prince thanks Arāḍa Kālāma but ponders, expresses reservations, and moves
on: “For I am of the opinion that the field-knower, although liberated from the primary (prakṛti) and secondary
(vikāra) constituents, still possesses the quality (dharman) of giving birth and also [the quality (dharman)] of being
seed: vikāraprakṛtibhyo hi kṣetrajñam muktam apy aham/ manye prasavadharmāṇam bījadharmāṇam eva ca.”
15. See BC 13.1 (Māra as saddharmaripus, “enemy of the true dharma”); 13.31 (the divine sages in their pure
abodes are “devoted to the good law”; continuing: they are dharmātmā, “given to dharma” (Johnston), whereas
Māra’s hosts are hiṃsātmā, “cruel” (Johnston), or “given to violence” [13.32]). The term also occurs when Asita
comes “thirsting for the holy Law” (1.49) and predicts that the Buddha will deliver it (1.74), and in the ironic
words of Chandaka at 6.31.
630 dharma

(caste) only in passing (4.18) and spinning out debates about āśramadharma
without ever precisely calling it that,16 he provides special moments for dharma
in the trivarga (10.28–38, 11.58), kuladharma (10.39), and the three debts a man
owes to his ancestors, the seers, and the gods (9.65).17 I will return to all these
matters.18
Second, I would like to illustrate as a prime example of the salience of this
concern, and for its foundational importance for all that follows, how Aśvaghoṣa
presents the famous story of how the princely young Buddha-to-be encounters
the four signs. For the first outing (BC 3.26–38), the Śuddhādhivāsa gods create
the “illusion of an old man” (26). The prince19 asks his charioteer about it:
“Is this some transformation in him, or his original state, or mere chance
(yadṛcchā)?” Thanks to the gods’ confusing the charioteer into spilling the
beans about old age, the prince, having learned the truth, “started a little (calitaḥ
ca kiṃcid)” and offered this first response: “Will this evil come upon me also?
(kim eṣa doṣo bhavitā mamāpi)” (32)—a rather shallow response compared to
what he says when next confronted with signs two and three. For now, he asks
to be taken back to the city; he cannot take pleasure “when the fear of old age
rules in my mind (jarābhaye cetasi vartamāne)” (37d). For the second outing
(3.39–53), the same gods fashion a diseased man. The prince’s first thoughts on
this are more reflective: “Thereupon the king’s son looked at the man compas-
sionately (sānukampyo) and spoke: ‘Is this evil (doṣa) peculiar to him, or is the
danger of disease (rogabhayam) common to all men?’ ” (43). Made aware of the
realities, he observes the “vast ignorance (vistīrṇam ajñānam)” of men “who
sport under the very shadow of disease” (46). When he has returned to the
palace, his father, sensing the prince had “already abandoned” him (49), scolds
“the officer in charge of clearing the roads,” but with no severe punishment,
and prepares another outing hoping to change the prince’s mood. For the third
outing (3.54–65), the same gods now fashion a lifeless man, arranging it so that
only the prince and charioteer see it (54)! Now the prince’s question is still
more sophisticated: “Is this law of being (dharmaḥ) peculiar to this man, or is

16. As Olivelle 1993, 120–21 on BC 5.30–33 shows, Brahmanical āśrama (“life-pattern” or “life-stage”) for-
mulations were still in flux. Curiously, the Ugraparipṛcchā, an early Mahāyāna text probably from around the
first-century CE (Nattier [2003] 2005, 41–45, 193), and thus around or shortly before Aśvaghoṣa’s dates, shows
similar early variation in formulations on the “stages” (bhūmis) of the Bodhisattva career (151–52), along with
intense recommendation that not only monks but lay householders take up this arduous path to Buddhahood.
17. On Aśvaghoṣa’s treatment of these, possibly referring to Manu, and its understanding of the term
mokṣa, see Olivelle 2008, xx–xxii and § F.2.
18. See also 4.83, 7.14, and especially 9.76 and 13.49 for criticism of the uncertainties and wavering of
traditional āgamic authorities.
19. As I will usually call him, except where Aśvaghoṣa uses other terms for him, notably Bodhisattva,
which, since it occurs for the living prince for the first time at BC 9.30, I will not use to describe him before that
point in the text.
 aśvaghoṣ a ’s buddhacarita 631

such the end of all creatures? (kiṃ kevalo ‘syaiva janasya dharmaḥ sarvaprajānām
ayam īdṛśo ‘ntaḥ)?” (58cd). To which the charioteer replies, “This is the last act
for all creatures (sarvaprajānām idam antakarma). Destruction is inevitable for
all in the world, be he of low or middle or high degree” (59). In short, from first
asking about only himself with regard to old age to asking about whether dis-
ease is unique to one or common to all, he is now, when it comes to the dead
man, still framing the question in the same way as for the diseased man, but
not only asking whether death applies to one or to all but asking after the under-
lying “law” (dharma) that results in death. But whereas the prince asks about a
“law,” the charioteer answers him only in terms of “acts,” very nicely translated
as “the last act.” So the discovery of such a law will remain the prince’s prob-
lem.20 He is not handed such a law by a charioteer—I am, of course, alluding to
the Bhagavad Gītā—or anyone else. Instead of dharma being revealed, it is
approached through developing insight.21 As elsewhere, there is a convergence
point between dharma and mṛtyu,22 and perhaps of the two with ignorance
(here ajñāna). Now the prince suddenly becomes “faint on hearing of death”
(śrutvaiva mṛtyum; 60), grabs the chariot rail, and then reflects “in a melodious
voice” (61) that “this is the end appointed for all creatures (iyaṃ ca niṣṭhā niyatā
prajānām)” and how, to appear happy, men must harden their hearts for them
to be in good cheer as they fare along the road (adhvan; 61). He asks to return
to the city as it is no time for pleasure resorts (62), but the driver goes at the
king’s behest to a grove prepared in advance, a park filled with birds and
beautiful women, which the prince experiences as if he were a Muni carried
there by force to a place presenting “obstacles” (65). This sylvan pause gives
Aśvaghoṣa the opportunity to devote the next canto of lacy kāvya to the wiles of
women (one of whom, as noted, even pretends to stumble), and the prince’s
newfound indifference to them, before he is visited by the fourth sign (5.1–15).
The prince now heads out, again with his father’s permission, to see the
forests, taking a retinue of companions (sakhibhis; BC 5.2) who are the sons
of ministers. He rides Kanthaka, but the charioteer is not with him. Going to

20. It is interesting to compare Aśvaghoṣa’s innovations with the Mahāpadāna Sutta scene (DN 14.2),
where the Buddha describes how the former Buddha Vipassī responded to his charioteer upon learning of the four
signs. Vipassī has a similar shallow response to the first sign: “But am I liable to become old, and not exempt from
old age?” (Walshe [1987] 1995, 208). But Vipassī’s responses remain at this level through signs two and three, with
no one mentioning dhamma until the shaven-headed wanderer (pabbajita), as the fourth sign, seems to introduce
the term to him: “Prince, by one who has gone forth we mean one who truly follows Dhamma” (Ibid., 210).
21. At BC 7.46, just after the great departure, he tells the first anchorites he meets that he is still “a novice
at dharma (me dharmanavagrahasya).” Cf. Gawronski 1914–15, 33, taking this as “(of me) who have newly taken to
the dharma i.e. who am a neophyte regarding it,” and citing 11.7 as a further unfolding of this theme.
22. See chapter 3 § F on the opening of the Kaṭha Upaniṣad; chapter 9 § C on the epics and their
substories.
632 dharma

distant jungle-land (presumably “savannah”) he sees the soil being ploughed,

and, seeing insects cut up, he mourns for them as for his own kindred (5).
Seeking clearness of mind, he stops his wellwishers (suhṛdas, 7) and goes to sit
beneath a Jambū tree.23 There, “reflecting on the origin and destruction of
creation (jagataḥ prabhavavyayau vicinvan)” and taking “the path of mental still-
ness” (9), he enters the “first trance of calmness” (10) and attains “concentration
of mind (manaḥ samādhim)” (11). And, having rightly perceived it, he meditates
on the “course of the world (lokagatim).” This meditation soon carries forward
from what was brought into focus around the term dharma as he encountered
the third sign:

A wretched thing it is indeed (kṛpaṇam bata) that a man, who is

himself helpless and subject to the law of old age, disease, and
destruction (vyādhijarāvināśa-dharmā), should in his ignorance and
the blindness of his conceit, pay no heed to another (param ajño) who
is the victim of old age, disease, or death [my italics]. For if I, who am
myself such, should pay no heed to another whose nature is equally
such, it would not be right or fitting in me, who have knowledge of
this, the ultimate law (paramaṃ dharmam imaṃ vijanato me).24

He is realizing that this “law” involves a recognition of “the other” with whom
all are in this together, which carries forward from the progression through the
first three signs. And after verses 14–15 describe this insight further and its
neutralizing of the passions in the prince, it is now the moment for the arrival
of the fourth sign (5.16–21), which, rather than provoking these reflections,
comes in response to them. Not fabricated by the gods like the other three
signs, a śramaṇa appears as a bhikṣu or mendicant (5.16), and says, “In fear of
birth and death [I] have left the home life for the sake of mokṣa (pravrajato ‘smi
mokṣahetoḥ)” (17). He is a homeless wanderer-seeker, “accepting any alms
I may receive (yathopopannabhaikṣaḥ)” (19), and, moreover, a “heavenly being
who in that form had seen other Buddhas, and has encountered the prince to
rouse his attention (smṛti)” (20), which he gets. For, “When that being went like
a bird to heaven, the best of men was thrilled and amazed. And he gained
awareness of dharma (upalabhya tataś ca dharmasaṃjñām) and set his mind on
the way to leave his home.”25 When he returns to the palace, it is “with yearning

23. In the Nidānakathā, this episode occurs when he is a mere child with nurses (see Warren 1998, 53–55).
24. BC 5.12–13: kṛpaṇam bata yaj janaḥ svayaṃ sann/ avaśo vyādhijarāvināśadharmā//jarayārditam āturam
mṛtam vā/ param ajño vijugupsate madāndhaḥ//iha ced aham īdṛśaḥ svayaṃ san/ vijugupseya paraṃ tathā
svabhāvam//na bhavet sadṛśaṃ hi tat kṣamaṃ vā/ paramaṃ dharmam imaṃ vijānato me.
25. BC 5.21: gaganaṃ khagavad gate ca tasmin/ nṛvara saṃjahṛṣe visismiye ca//upalabhya tataś ca
dharmasaṃjñām/ abhiniryāṇa vidhau matiṃ cakāra.
 aśvaghoṣ a ’s buddhacarita 633

aroused for the imperishable dharma (akṣayadharmajātarāgaḥ)” (25–26). One

might wonder whether Aśvaghoṣa draws a contrast with the term
sanātanadharma, “eternal dharma,” which he would have had opportunity to
know from both Sanskrit epics. An eternal dharma invokes the eternal Veda
and a dharma that, while beyond appearances, is always subtly present, whereas
an imperishable dharma could avoid these implications and evoke something
that neither perishes nor originates but can always be rediscovered.26 In any
case, this birdlike divine creature sets the prince to the task of unfolding this
new awareness of dharma he has already begun discovering on his own by
setting his mind on departure from home—which is clearly not the locus of this
dharma, although it will not fail to bear upon it.

C. Aśvaghoṣa the Brahmin, Buddhist Convert, and Scholar

On one matter, all agree, even though it is again only Chinese sources that
actually state it: that Aśvaghoṣa was a Brahmin convert to Buddhism. Johnston
gives numerous reasons to accept the Chinese tradition on this one point
(2004, xviii), and actually hazards to speak of “the zeal of the convert” (xcvi).
But Johnston’s first claim for Aśvaghoṣa under the rubric of converted Brahmin
is that “he had an acquaintance, so wide that no parallel can be found to it
among other Buddhist writers, with all departments of Brahmanical learning”
(lviii)—a topic to which Johnston devotes a whole section under the heading of
“The Scholar” (xlvii–lxxix). He thus credits Aśvaghoṣa with Ṛgvedic knowledge;27
familiarity with Brahmanical ritual texts (xlv, lxxviii, lxxiii–lxxxiv); the Upaniṣads
(xlv–vi), early nīti (l–li), medical, astronomical/astrological, and śilpa (lii–liii)
texts; early Sāṃkhya, Yoga, and possibly Vaiśeṣika texts (lvi–lxii); contemporary
developments in kāvya (lxii–lxiii); and of course the two epics in some form
(our next topic). But in a fascinating oversight or omission, he makes no attempt
to relate Aśvaghoṣa’s knowledge of Brahmanical dharma to any dharma litera-
ture. Perhaps he assumed that the epics were sufficient to cover what Aśvaghoṣa

See Johnston 2004, 65 n. 21 on dharmasaṃjñā with upa-labh, in the “technical sense of the action of the
mind in forming ideas or conceptions, based on the perceptions presented to it by the senses.”
26. Horsch 2004, 439 mentions, without citation, early Buddhist usage of sanātana (“eternal”) and akālika
(“timeless”) for dhamma as “correspond[ing] to the sanātano dharmaḥ of the Hindu philosophers,” but that the
dhamma is “fixed” whether Tathāgatas rise up or do not (citing Saṃyutta Nikāya 2, p. 24, W. Geiger trans.). Cf.
Nattier 2003, 142: a Bodhisattva must “be born in his final life into a world devoid of Buddhism, where he will
rediscover its truths for himself.” See chapter 6 above.
27. See Johnston 2005, xlv and 124–25, note to BC 14.9: “The legend of Vasiṣṭha’s descent from Urvaśī is
alluded to in the RigVeda,” which the verse refers to, although it “had already been lost sight of by the time of the
epics.”
634 dharma

knew of Brahmanical dharma, but that, I believe, would be a very risky assump-
tion. The prince’s friend Udāyin does cite epic precedents as to the duty to ful-
fill women’s desire at BC 4.66–67 (though not the most obvious such case:
Arjuna’s accommodation of the Nāgī or serpent woman Ulūpī, which hinges
on her interpretation of this “highest dharma” [Mbh 1.206.23–33])—counsel
which Aśvaghoṣa describes as “specious words, supported by scriptural tradi-
tion (āgama)” (BC 4.83) that the prince deafeningly rejects (84–99). But there
is probably more than epic precedent when, soon after King Śuddhodana’s rule
is compared to that of “Manu, son of the sun” (2.16), Aśvaghoṣa describes the
young prince growing up in a kingdom where his father not only practiced all
the virtues of self-restraint, offered large fire ceremonies (36), and drank soma
as enjoined by the Vedas (37), but judged petitions impartially “and observed
purity of justice (vyavahāra-śuddham) as being holy (śivam)” (39), did not exe-
cute the guilty but imposed mild punishments (42), and taxed fairly (44)—all
this while the king “pondered on the Śāstra” (vimamarśa śāstram, 52).28
In any case, Johnston makes several astute assessments on Aśvaghoṣa’s
erudition that are worth quoting. First, he says that “Aśvaghoṣa writes for a
circle in which Brahmanical learning and ideas are supreme; his references to
Brahmins personally and to their institutions are always worded with the
greatest respect, and his many mythological parallels are all drawn from
Brahmanical sources.”29 Second, Johnston says that Aśvaghoṣa’s accuracy and
even pedantry bind us “to assume that his learned references are strictly in
accordance with the authorities he used,” even though “these authorities are
for the most part no longer extant” (xliv). Third, and most important, he
observes that Aśvaghoṣa “seems at times to delight in expressing Buddhist
views in a way that would remind Hindu readers of their own authorities” (lv).
If so, for the long run, at least, this was probably wishful thinking, as his verse
was little cited after Kālidāsa (lxxix–lxxxii) and only half-survived in four Sanskrit
manuscripts until modern interests somewhat resurrected him. Johnston also
remarks that by “introducing so much Hindu learning [Aśvaghoṣa] offended
against the puritan moment in Buddhism” (xxxvii), which likewise did little to
later acclaim him—at least in subsequent Indian Buddhist texts, although the
Chinese Pilgrim I-Tsing found the Buddhacarita popular in India in the seventh

28. Vyāvahāra is “justice” in “judicial procedure,” that is, jurisprudence, as the examples make evident—an
important innovative emphasis in Manu (see chapter 5). By the Buddha’s account, his predecessor Vipassī (see
n. 20 above) was also, even as a boy, instructed by his father when trying legal cases (Mahāpadāna Sutta
[DN 14] 1.37).
29. Johnston 2004, xv–xvi. See especially BC 7.45, where the prince shows respect toward the tapasvins—
“the upright-souled sages, the supporters of religion (dharmabhṛtām)—of the penance grove.” Johnston 2004,
xvi, n. 1 notes two exceptions, whose genuineness he doubts, in the Saundarananda. In any case, the point applies
to the Buddhacarita.
 aśvaghoṣ a ’s buddhacarita 635

century (Johnston 2004, xxxv–xxxvi) when he travelled in northeastern India

around 672. Indeed, that may have been the level at which this “Buddhist epic”
would have had its longest run in India. Recalling Johnston’s emphasis on
faith and bhakti in the Buddhacarita, it is not uninteresting that a text composed
by a Brahmin convert to Buddhism who knew the epics and was attentive to
Buddhist bhakti would find its longest response to be a “popular” one.

D. Aśvaghoṣa and Epic Precedents

Aśvaghoṣa thus presents us with the opportunity to study a “close reading”

of both Brahmanical epics: close both in time, for I do not think it very likely
that written versions of either epic can be more than three centuries earlier
than Aśvaghoṣa, and more likely only preceded him by about a century or at the
most two; and in relation to the question Johnston raises by insisting that
Aśvaghoṣa is scrupulous in citing his authorities. With these points in mind, it
is worth making a few observations about how Aśvaghoṣa treats both epics
together before looking at the ways he treats each distinctly.
First, it seems there are recurrent points where he alludes to the two epics
either together or alternately. Most striking is the first such instance when King
Śuddhodana’s court Brahmins interpret the baby prince’s birth signs and refer
to various texts, their authors, and then other heroic figures before the sage
Asita arrives to read the signs definitively. To make the point that “Anyone may
attain pre-eminence anywhere in the world, for in the case of the kings and
seers the sons accomplished the various deeds their ancestors failed to do” (BC
1.46), these court Brahmins mention the following instances (I paraphrase
from 1.41–45):

41. Although Bhṛgu and Aṅgiras were the founders of families, it was
not they who created (cakratus) the “science of royal policy”
(rājaśāstra), but their sons Śukra and Bṛhaspati.
42. The son of Sarasvatī promulgated again the lost Veda ( jagāda naṣṭaṃ
vedam) and Vyāsa divided it into many sections, which Vasiṣṭha
(his great grandfather) had not done.
43. And Vālmīki was the first to create poetry (vālmīkir adau ca sasarja
padyam), which Cyavana30 did not do; and Ātreya31 proclaimed the
science of healing which Atri did not discover.

30. Another Bhārgava; but see Johnston 2004, 10 n. 43.

31. Perhaps alluding to Caraka; see Johnston 2004, 70, n. 43.
636 dharma

44. Viśvāmitra won Brahminhood (dvijatvam) which Kuśika (his

grandfather) did not, and Sagara set a limit for the ocean which his
Ikṣvāku predecessors did not achieve.
45. Janaka gained preeminence in instructing the twiceborn in yoga, and
Śūra (Kṛṣṇa’s father) and his kin were incapable of the celebrated
deeds (khyātāni karmāṇi) of Śauri (i.e., Kṛṣṇa).

Verses 42–43 establish a clear Mahābhārata–Rāmāyaṇa alternation (Vyāsa and

Vālmīki), whereas the rest refer to sages and kings known in both epics. This
alternance and fusion, which occurs repeatedly, suggests a kind of śleṣa inten-
tion toward the two epics.32 Moreover, it would be hard to explain how Aśvaghoṣa
would know what he knows about the two poets, Vyāsa and Vālmīki, unless he
were familiar with material from the twelfth book of the Mahābhārata (if not
also the first) and from the first book of the Rāmāyaṇa (if not also the seventh).
As Johnston remarks, one may infer from a verse in Aśvaghoṣa’s earlier work,
the Saundarananda, “that the story of Vālmīki’s having taught the poem to
Kuśa and Lava was familiar to him (2004, xlix). In fact, the verse credits Vālmīki
with having performed the twins’ childhood rites, and both Vālmīki and the
boys with being “inspired” (dhīmat).33
But Aśvaghoṣa also has a point in making epic and other Brahmanical
mythological allusions, though some of them are certainly obscure.34 It is to
bring across a realization that, no matter how illuminating heroic, sagely, and
divine precedents may be as parallels, they are ultimately irrelevant to the
achievement of the Buddha. Moreover, he makes this thoroughly intelligible
Buddhist point in a manner that undercuts Brahmanical practice at one of its
nerve centers. Aśvaghoṣa’s manner of citation sets him apart from the practice
of citing heroic and divine precedents that is found in both epics, but especially
the Mahābhārata with its use of the athāpy udāharanti (“now they also quote”)
citational formula that is also used in most of the dharmasūtras (see chapter 5 § C).
Indeed, the Mahābhārata cites precedents not only in substories featuring the
great Vedic Ṛsị s (see chapter 9 § B) but in the words of many leading characters,

32. Johnston 2004, xciii–xcvi observes something analogous in Aśvaghoṣa’s allowance of double
Brahmanical and Buddhist meanings in saṃdhi passages with “a negative disappearing” (BC 3.25; 12.82 [he prob-
ably means 12.81]).
33. Saundarananda 1.26cd: vālmīkiriva dhīmāṃs ca dhīmator maithileyayoḥ; see Johnston 1928, 3 n. 26:
“inspired” for dhīmat, referring “to Vālmīki’s poetic inspiration in composing the Rāmāyaṇa and to Kuśa and
Lava’s artistic skill in repeating it.” Yet Johnston 2004, xlix says, “As regards the Uttarakāṇḍa, I can find no
reason to suppose that the poet knew any portion of it.”
34. For unknown and uncertain references and usually Johnston’s discussion thereof, see BC 41.16–18;
4.72–75; 4.80 (? Karālajanaka); 9.20; 9.69–70; 11.15, 11.18; 11.31 (Mekhala-Daṇḍakas); 13.11 (Śūrpaka, the fishes’
foe); 28.32 (Eli and Paka).
 aśvaghoṣ a ’s buddhacarita 637

and above all those of the author and the deity, that is, of Vyāsa and Kṛsṇ ạ .35 On
these matters, Aśvaghoṣa’s stance is most vivid on the crucial point of the
decision not to return home after the Buddha’s “great departure.” Here, the
prince dispenses with royal precedents for returning home from the forest,
including the precedent of Rāma, by saying to one of his father’s emissaries,
“And as for your quoting the instances of Rāma and the others to justify my
return [home], they do not prove your case; for those who have broken their
vows are not competent authorities in deciding matters of dharma (na te
pramānạ ṃ na hi dharmaniścayeṣv alaṃ pramānạ ȳ a parikṣaya vratā)” (BC 9.77).
Rāma may offer precedent but not an “authority” (pramānạ m)! Moreover, we are
left with the tantalizing question of what vow Rāma might have broken,36 for it
is almost certainly king Rāma, son of Daśaratha, who is being kept in focus here,
even though Aśvaghoṣa can also refer to Rāma Dāśarathī and Rāma Jāmadagnya
in one and the same breath.37 Further along, one hears similarly how “Vasiṣṭha,
Atri, and others came under the dominion of time”; so too Yayāti, etc., and hun-
dreds of Indras, whereas Sambuddhas entered nirvānạ (24.38–42). Finally, in
the last canto, when seven kings are ready to go to war over the Buddha’s bones
and cite as heroic precedents for doing so Śiśupāla’s stand against Kṛsn ̣ ạ , the
end of Vṛsṇ ị s and Andhakas over a woman, Bhārgava Rāma’s decimation of the
Kṣatriyas, and Rāvaṇa’s infatuation with Sītā (28.28–31), the point could not
be clearer that heroic precedents from the Brahmanical epics are dangerous. Or,
as John Brockington puts it in the case of the destruction of the Vṛsn ̣ ị s and
Andhakas, the story “figures as a moral warning” (1998, 484). Let us also note
that Aśvaghoṣa does not miss the opportunity to lace his epic precedents with
allusions to many of the great Vedic Ṛsị s and their families—in which descen-
dants may surpass their eponymous ancestors. Indeed, without quite saying so,
he restates the Nikāya view that tracing precedent to the great Vedic Ṛsị s is to
follow a “procession of the blind” (see chapter 4 § A).
Yet we will also have occasions to note that Aśvaghoṣa, probably both as a
kāvya poet and a Buddhist convert, could have his reasons for treating epic allu-
sions with a little play. At Buddhacarita 4.16, for instance, Udāyin begins urging
the women to show some gumption in seducing the prince: “Of old time,
for instance, the great seer, Vyāsa, whom even the gods could hardly contend

35. For some discussion, see Hiltebeitel [1976] 1990, 261–66 (Kṛsn ̣ ạ reveals divine precedents for Arjuna’s
̣ ạ reveal divine precedents for Yudhiṣṭhira’s Aśvamedha); 2001a, 73 (Idem),
killing of Karṇa); 289–96 (Vyāsa and Kṛsn
49, and 118–20 (Vyāsa reveals precedents for the polyandric marriage of Draupadī, on which see chapter 10 § B).
36. Neither Johnston nor Olivelle offers a suggestion.
37. See in the same canto BC 9.25, where the prince hears about both Rāmas and Bhīṣma as exemplars of
doing deeds to please their fathers. See also 9.69, where he hears, “So too Rāma left the penance grove and
protected the earth, when it was oppressed by the infidel (anāryais)”—on which Johnston is no doubt right that
this probably refers to Bhārgava Rāma (2004, 137, n. 69).
638 dharma

with, was kicked with her foot by the harlot (veśavadhvā), Kāśisundarī.” Johnston
says, “The story is unidentified and it is uncertain if Kāśisundarī is a proper
name or not” (2004, 46 n. 16). But most likely it unfolds, a bit bawdily, from
the night Vyāsa spends happily siring Vidura with the Śūdra servant-woman of
the Kāśi princess Ambikā, whom Ambikā adorns with her own jewels so that
she looks “like an Apsaras” (svair bhūṣaṇair dāsīm bhūṣayitvā apsaropamām;
Mbh 1.100.23)—that is, a beautiful heavenly courtesan—and sends to Vyāsa in
her own stead, apparently to try to fool him (100.23–101.1; see chapter 8, §§ D
and E). No doubt this maid would also be from Kāśi, and thus either named
Kāśisundarī or described as “the beautiful Kāśi woman.”38 In effect, Udāyin
would be saying, If nothing else works, give the prince a kick.
As we now proceed to Aśvaghoṣa’s close reading of the epics themselves,
I think we can thus allow ourselves a caveat with regard to Johnston’s insis-
tence that Aśvaghoṣa is scrupulous in citing authorities. I certainly believe that
Aśvaghoṣa wants to be understood by those who know the epic texts. But it is
unlikely that he or they knew them only as written texts, since by his time they
no doubt already served as the basis for oral adumbrations in both Brahmanical
and Buddhist circles in which either and indeed both together could have some
fun with the text. This point is worth keeping in mind as we now address the
more serious matters that interest Aśvaghoṣa in juxtaposing the life of the
Buddha to scenes in both epics, not only separately but together, where they
exemplify their different but also complementary guidelines on a basic problem
raised by the Brahmanical dharma of householder kings.

E. The Buddhacarita and the Rāmāyaṇa

As Johnston points out, Aśvaghoṣa’s treatment of the Rāmāyaṇa is more direct

than that of the Mahābhārata, since, as we have already begun to notice, he
makes frequent reference to the life of the Rāmāyaṇa’s main hero. Johnston
picks up on the Buddhacarita’s closing colophon, where the poet writes of
himself as “Aśvaghoṣa of Sāketa” [i.e., Ayodhyā],39 for a likely explanation:

The case is entirely different with the Rāmāyaṇa, for which an

inhabitant of Sāketa, the scene of its most poignant episodes and the

38. Sullivan, who discusses this and a similar verse in Aśvaghoṣa’s Saundarananda (7.30), considers
Kāśisundarī to have been Ambikā herself, but this is a more unlikely solution since Ambikā would have had to
confront Vyāsa directly to have (in Sullivan’s words) so “decisively rejected” him (1990b, 291), and since the verse
is intended as inspiration in the arts of seduction.
39. Johnston 2004, Part 3, 124. Cf. Lamotte 656: “a native of Sāketa who had converted to Buddhism.”
 aśvaghoṣ a ’s buddhacarita 639

capital of its dynasty, could not but keep a warm place in his heart,
however his religious beliefs had changed. Aśvaghoṣa never tires of
reminding us that the Buddha belonged to the dynasty of his home
and strikes this note in the very first verse of the Buddhacarita.40

From this no doubt important point, Johnston turns to “enquire to what extent
he [Aśvaghoṣa] knew the poem in its present form” (2004, xlviii), favoring the
view of Andrzej Gawronski, who, he says, has

. . . proved conclusively, as I hold, that Aśvaghoṣa knew certain

portions of the second book, the Ayodhyakāṇḍa, in very much the
condition that we have them in today and that he took pleasure in
drawing a comparison between the Buddha quitting his home and
Rāma leaving for the forest. That he knew the continuation of the
story is proved from a reference in B., xxciii. 31 [concerning the bad
precedent, just cited, of Rāvaṇa’s doomed infatuation with Sītā], but
whether in the present form or not is not clear from the wording.
It certainly does seem that there are many future passages in the later
books likely to have influenced the Buddhist poet. . . . The question
really turns on whether Aśvaghoṣa knew some or all of the passages
in the Rāmāyaṇa, describing how Hanumān visited Rāvaṇa’s palace
and saw the women asleep. (Johnston 2004, xlviii)

In fact, Gawronski limited his discussion to Rāmāyaṇa Book 2 because he

found the parallels more direct there and a larger comparison too unwieldy
(1919, 27–28); he felt enabled “to conclude with a sufficient amount of cer-
tainty that at the time of Aśvaghoṣa there existed at least Book II of the Rāmāyaṇa
(but most probably the remaining genuine books also) in much the same form
as is known to us today” (40). Gawronski flagged most of the Book 2 passages
that I will discuss. As to the well-known kāvya question of the similarities bet-
ween Hanumān’s viewing the sleeping women in Rāvaṇa’s palace and the
Buddhacarita’s sleeping harem scene on the night of the Buddha’s great
departure, Johnston says he “will refrain from giving a definite answer” until
there is a Rāmāyaṇa critical edition (2004, xlvii). On this matter, Brockington
takes a favorable view, as do I, of V. Raghavan’s demonstration (1956) that
Aśvaghoṣa borrows the harem scene from Sundarakāṇḍa 5.7–9, “including
parallels in wording” (Brockington 1998, 485).
Beyond these probably unnecessary cautions, Johnston makes some inter-
esting observations about intratextual intricacies: that there is a problem with

40. Johnston 2004, xlvii. See also BC 10.23; 13.1 (implied); 14.92; 17.6.
640 dharma

whether Viśvāmitra is seduced by Ghṛtācī, as Aśvaghoṣa has it along with a verse

in Rāmāyaṇa Book 4, or Menakā, who is the seductress in the story in Rāmāyaṇa
Book 1;41 and that Aśvaghoṣa would seem to have needed the Rāmopākhyāna to
explain why he has Vāmadeva and Vasiṣṭha visit Rama in the forest (Johnston
2004, xlix–l). But these cautions and conundrums have to do not with the heart
of Aśvaghoṣa’s interests in the Rāmāyaṇa, but with his selective pattern of mak-
ing allusions as ultimately negative precedents, which I have already discussed.
The heart of the matter is, as Johnston puts it, that Aśvaghoṣa “took pleasure in
drawing a comparison between the Buddha quitting his home and Rāma leav-
ing for the forest” (xlviii). Indeed, the Buddhacarita has this much in common
with the Pāli Vessantara Jāṭaka, which, as Gombrich (1985) shows, involves
detailed but more indirect correspondences not between Rāma and the Buddha,
but between Rāma and the Buddha in has very last life as Prince Vessantara.42
For Aśvaghoṣa, however, it is not just a matter of poetic pleasure (such as
might be the case in drawing from the Rāmāyaṇa’s sleeping harem scene), as
Johnston seems to imply. What interests Aśvaghoṣa is the opportunity Rāma’s
departure offers to draw a contrast between Brahmanical dharma and Buddhist
dharma. Taking into account only the first fourteen cantos of the Buddhacarita
(those for which we have extant Sanksrit texts), the prince, up to his enlighten-
ment, has no less than thirteen interlocutors with whom he hones his views on
dharma: (a) his charioteer, through the first three signs (BC 3.26–65); (b) Udāyin
(4.9–23, 56–99); (c) the Śramaṇa who appears as the fourth sign (5.9–21); (d) a
“nobleman’s daughter” (rājakanyā), elsewhere43 known as Kisā Gotamī, whose
words of praise upon seeing his return from the fourth sign crystallize his
silent resolve to pursue “the means to final nirvāṇa (parinirvāṇavidhau matiṃ
cakāra)” and “the imperishable dharma” (5.23–26); (e) his father (5.27–46, this
being the only point where he addresses his son directly); ( f ) the horse
Kanthaka44 (5.68–72, a one-way conversation in which the prince voices his
readiness for the great departure after the Akaniṣṭha deities have arranged the
sleeping harem scene); (g) his groom Chandaka (6.1–52, when the prince sends

41. For the first, Johnston 2004, xlix gives Rām 4.35.7, which is 4.34.7 in the Baroda Critical Edition; the
second is CE 1.62.4–13. As Lefeber 1994, 289 notes, some commentators identify the two Apsarases as one and
the same.
42. Gombrich astutely suggests that this deflection to a previous life “reflects the hostility of Theravāda
Buddhism (though the VJ story was not confined to the Theravāda) to the values embodied in the Rām,” and
agrees with Heinz Bechert (1979, 28) that this would further have to do with the Rāmāyaṇa’s being “unaccept-
able to the Sinhalese because it contradicts their view of the island’s history”—especially in Rāmāyaṇa Book 6.
Aśvaghoṣa would not have this Laṅkan problem with Vālmīki.
43. And in different circumstances; see the Nidānakathā version in Warren 1998, 59.
44. Aśvaghoṣa speaks of it as King Śuddhodana’s horse, which he has ridden in battle (BC 5.75). It is not
born, along with the groom, at the same time as the Buddha, as in the Nidānakathā (see Warren 1998, 48).
 aśvaghoṣ a ’s buddhacarita 641

him home after making the great departure);45 (h) the anchorites of a Bhārgava
penance grove, there with their wives (7.1–58); (i) the Purohita or chaplain
(9.81–51); and ( j) the Minister,46 jointly sent by the king to the penance grove to
speak for him and the Ikṣvāku line (9.4); (k) Śreṇya-Bimbisara, king of Magadha
(10.22–11.71); (l) Arāḍa Kālāma (12.1–83); and (m) Māra (13.1–69).
In at least four of these cases, Aśvaghoṣa relates the prince’s departure
directly to the Rāmāyaṇa (or in the fourth case possibly to the Rāmopākhyāna).
First and foremost, King Śuddhodana compares his grief to that of “Daśaratha
friend of Indra,” and envies Daśaratha for going to heaven when Rāma did not
return (BC 8.79–91): “Thus the king grieved over the separation from his son
and lost his steadfastness, though it was innate like the solidity of the earth; and
as if in delirium, he uttered many laments, like Daśaratha overwhelmed by
grief for Rāma” (BC 8.81). Grief (śoka) is of course the Rāmāyaṇa’s underlying
sthāyibhāva or “stable aesthetic emotion” in relation to karuṇa,̄ “pity” as its pre-
dominant aesthetic flavor (aṅgī rasa),47 and it characterizes King Śuddhodana’s
feelings for his son throughout the Buddhacarita.48 Second, the groom Chandaka
says, “I cannot abandon you as Sumantra did Rāghava” (6.36). Third, when he
and the riderless horse return, the townsfolk “shed tears in the road, as
happened of old when the chariot of Daśaratha’s son returned” (8.8). Fourth, as
already noted, the chaplain (purohita) and minister are compared, as emissaries,
to Vāmadeva and Vasiṣṭha visiting Rāma in the forest (9.9).
But there are also indirect allusions to the Rāma story. Indeed, if the two
emissaries seem to step into their roles with Rāmāyaṇa echoes,49 the same
can be said of the prince’s encounter with the many Ṛṣis who dwell in a pen-
ance grove together with their wives (BC 7.3). I would propose that Aśvaghoṣa
builds up this scene to represent the vānaprastha mode of life idealized in
the forest books of both epics,50 but especially in the Rāmāyaṇa, where Rāma
is relayed between Vedic sages, one of whom, Atri, is explicitly ensconced in
the forest with his wife Anasūyā.51 In any case, the prince’s descent in the

45. Assuming that Chandaka is different from the unnamed charioteer.

46. BC 9.52–79, referred to, when the prince dismisses them, as tau havyamantrākṛtau, “the officers who
were in charge of the king’s sacrifices and his counsel chamber” (BC 10.1).
47. See my discussion in Hiltebeitel 2008a.
48. See BC 1.76 (Śuddhodana warned not to grieve over his son’s inevitable enlightenment); 6.19–20
(with likely Rām echoes in the prince’s reference to the road of his [Ikṣvāku] ancestors) and 6.30–31 (Chandaka’s
response); and especially 9.13–15, 9.29 (as aired by the Purohita, whom the prince answers on this point at
9.33–35). Meanwhile others also grieve throughout Canto 8 (śoka is mentioned twelve times there) when it is
realized that the prince has not returned with Chandaka and Kanthaka.
49. They are not found in the Nidānakathā.
50. See Biardeau 2002, vol. 2, 70–71, 75–76, 82 on these often married forest hermits, their hospitality to
epic princes, and their probably prior portrayal as well in the dharmasūtras.
51. See chapter 10, §§ A and B, chapter 12 § F.
642 dharma

Rāmāyaṇa’s dynastic lineage is certainly invoked when the teats of the

ashram cows in this “workshop as it were of dharma”52 flow upon first seeing
the prince as “the lamp of the Ikṣvāku race” (7.6)! Further, while each of
these thirteen interlocutors voices or hears words in the prince’s presence,
his abandoned wife Yaśodharā’s words in his absence are, I think, also
spoken in evocation of Sītā’s lonely soliloquies in the absence of Rāma (see
chapter 10 § D):

If he wishes to carry out dharma and yet casts me off, his lawful
partner in the duties of religion and now husbandless, in what
respect is there dharma for him who wishes to follow austerities
separated from his lawful partner?
Surely he has not heard of our ancestors, Mahasudarśa and the
other kings of old, who took their wives with them to the forest, since
he thus intends to carry out dharma without me.53

Whoever Mahāsudarśa may be,54 Yaśodharā would count Rāma among her
husband’s Ikṣvāku ancestors. This thread of direct and indirect Rāmāyaṇa evo-
cations comes to a decisive climax, in a passage cited earlier, when the prince
tells his father’s Purohita and Minister emissaries that Rāma is not an authority
(pramāṇam) on dharma (9.77).
For now, it must suffice to note that Aśvaghoṣa finds seven of the thirteen
champions of Brahmanical dharma—the father, the groom, the riderless horse
(rather than the empty chariot), the two emissaries, the wife, and the anchorites
(with their cows)—suitable, even if at a stretch, for evocations of the Rāma
story. It would take more space than it merits to demonstrate that, even beyond
these seven, all thirteen speak for one or another form Brahmanical dharma—
including, as we shall see, Māra. Suffice it to say that through the run of
Rāmāyaṇa precedents that ends with the prince dismissing them, it is, from
early on, the “variegated dharma (dharmaṃ vividham)” (BC 2.54) performed by

52. B 7.33; see Johnston 2004, 98 n. 33, crediting Garwonski 1919, 14–15 on this reading, but I think a little
too quickly dismissing his extension of the image to mean “forge, smithy,” making the penance grove “like a
forge of dharma in full activity (dharmasya karmāntam iva pravṛttam).”
53. BC 8.61: sa mām anāthāṃ sahadharmacāriṇīm/ apāsya dharmaṃ yadi kartum icchati/kuto ‘sya dharmaḥ
sahadharmacāriṇīṃ/ vinā tapo yaḥ paribhoktum icchati// 62. śṛnọ ti nūnaṃ sa na pūrvapārthivān/ mahāsudarśaprabhṛtīn
pitāmahān/vanāni patnīsahitān upeyuṣas/ tathā hi dharmam madṛte cikīrṣati.
Gawronski 1919, 35–36 remarks that the previous lines 8.55–58 of Yaśodharā’s lament and her contrast of
“the easy life he has enjoyed thus far and the drawbacks of dwelling in a hermitage” have another Rāmāyaṇa
parallel, but the words there are Daśaratha’s, the verses occur in a long interpolation (Rām 2, Appendix 1, No. 9,
lines 180–87), and the theme is perhaps rather a cliché.
54. Johnston 2004, 117 says he “is presumably the Mahāsudassana of the genealogies of the Dīpavaṃsa
and Mahāvaṃsa.” I am not sure one can rule out the subject of the Mahāsudassana Sutta (DN 17), whom the
Buddha recalls as one of this former lives (see Gethin 2006).
 aśvaghoṣ a ’s buddhacarita 643

King Śuddhodana as a śāstra-pondering king—one who, among other duties,

has just secured the continuance of his royal line through the birth of his son
(2.52–53)—that anchors all these Brahmanical concerns. The ultimate irony of
this portrayal of Brahmanical royal dharma by a Buddhist poet comes across
when Chandaka appeals that the prince should not abandon his loving father
“like a heretic abandoning the true dharma” (saddharmam iva nāstikaḥ; 6.31d).
Chandaka’s words rather yield the inadvertent message that, if the prince were
to rejoin his father, he would be a “nay-sayer” to “the true dharma” by abandon-
ing it before finding it.
This Rāmāyaṇa-related nexus runs mainly through the Buddhacarita’s
first nine cantos. Indeed, the only continuation I can see in later cantos comes
after the Buddha’s enlightenment, when “The seers of the Ikṣvāku race who
had been rulers of men, the royal seers and the great seers, filled with wonder
and joy at his achievement, stood in their mansions in the heavens reverencing
him” (BC 14.92). What a lovely twist to leave us wondering whether Rāma is
among them! Within this Rāmāyaṇa skein, there seem to be two sets of con-
cerns, each with numerous subsidiary considerations: gṛhasthadharma, or the
duties of a householder; and priorities regarding the second and third stages
of life as they bear upon kings in the scheme of āśramadharma, the ideal
sequence of the four stages of life—a term not used in the first half of the
Buddhacarita, but one whose currency is certainly implied, as when King
Śuddhodana tells the prince not to violate their “proper order.”55 This of course
means that the two concerns intersect, since according to the classical formu-
lation of the āśrama system (Olivelle 1993, 27, 30), the householder mode is
the second life-stage.
We see this intersection from the Buddhacarita’s first mention of
gṛhasthadharma, which, fittingly, comes right when King Śuddhodana first
faces his son’s determination to abandon both home and his succession to the
throne, and thus frames the issue as one of royal dharma. Says the father to
the son:

But, O lover of dharma, it is now my time for dharma, after I have

devolved the sovereignty onto you, the cynosure of all eyes; but if you
were to forcibly quit your father (gurum), O firmly courageous one,
your dharma would become adharma.

55. BC 5.32. See Olivelle 1993, 121 and n. 30, so translating vikrama at 5.32c, and commenting that
Johnston’s translation “misses the point.” Cf. Olivelle 2008, 137: “the right order.” See also 10.33, discussed in
the text.
644 dharma

Therefore give up this your resolve. Devote yourself for the

present to householder dharma (bhava tāvan nirato gṛhasthadharme).
For entry to the penance grove is agreeable to a man, after he has
enjoyed the delights of youth.56

Note that “entry to the penance grove” (tapovanapraveśa) is also used for the
forest-dwelling anchorites when they return to their “dharma workshop” at
7.58. This suggests that the term characterizes the third life-stage of the “for-
est-dweller” or vānaprastha (even though the text does not mention the term),
and that King Śuddhodana, at least, conceives the tension between him and
his son as one to be worked out between the “dharmas” of the second and third
life-stages, and not the second and the fourth. This is so even after the prince
hears the “nobleman’s daughter” utter the ambiguous word nirvṛtā—by which
she is describing the woman who would be “blessed” (Johnston 2004, 66) or
“happy” to have such a husband as he, but which fills him with the “supreme
calm (śamaṃ param)” that inspires him to win parinirvāṇa (5.24–25)—and
tells his father that he has decided to seek mokṣa (5.28), preferring that to the
word nirvāṇa, which is not used elsewhere in the first fourteen cantos to
describe the prince’s quest for it. Almost perversely, King Śuddhodana avoids
talking in such terms and, in the passage just cited, immediately rephrases his
son’s resolve into a premature decision for the penance grove and the implied
vānaprastha-dharma. Indeed, King Śuddhodana carries his seemingly delib-
erate misunderstanding to an offer to go to the forest rather than his son.57
This matter of untimely dharma being adharma percolates along through the
prince’s interactions with Chandaka (6.21), the king’s two Brahmin emissaries
(9.14–17; 9.53), and even Śreṇya Bimbisāra (10.33), and gives the prince sev-
eral opportunities to trump these Brahmanical concerns for the inherent time-
liness of āśramadharma with Buddhist rejoinders that “there is no such thing
as a wrong time for dharma” (6.21; cf. 9.37–38, 11.62–63). On the whole, such
concerns parallel the situation in the Rāmāyaṇa, which does not concern its
hero with any inclination toward mokṣa or the fourth life-stage of renunciation
(saṃnyāsa).58

56. BC 5.32. mama tu priya dharma dharmakālas/ tvayi lakṣmīm avasṛjya lakṣmabhūte/sthiravikramavikrameṇa
dharmas/ tava hitvā tu guruṃ bhaved adharmaḥ// 33. tad imaṃ vyavasāyam utsṛja/ tvaṃ bhava tāvan nirato
gṛhasthadharme/puruṣasya vahaḥsukhāni bhuktvā/ ramaṇīyo tapovanapraveśaḥ.
57. BC 5.32. On abdication by kings in favor of their sons, see Olivelle 1993, 116: “The epics contain
numerous accounts of famous kings who followed this custom” (with citations, n. 15).
58. This would be one reason why the Rāmāyaṇa has little to say about the āśrama system. Finding only
one reference (Rām 2.98.58), which leads him to think it would be an interpolation, Olivelle 1993, 103 supposes
that the Rāmāyaṇa would be older than this system, but his dates (pre-fifth-century BCE) for the Rāmāyaṇa were
at that point far too early.
 aśvaghoṣ a ’s buddhacarita 645

Yet the prince begins to break past this Rāmāyaṇa scenario in the penance
grove when he tells the anchorites that one of the reasons he does not stay with
them is that their practice of tapas yields merely “Paradise.”59 Unlike King
Śuddhodana, the anchorites know what he is talking about and tell him that if
he prefers release (which they call both mokṣa and apavarga, what is beyond the
trivarga) over Paradise (7.52–53), he should seek out Arāḍa Kālāma. They clearly
know of a fourth stage of life. With this, we put the Rāmāyaṇa behind us and
turn to Aśvaghoṣa’s treatment of the Mahābhārata, in which all four life-patterns
are a matter of major scrutiny and debate.
This returns us to the matter of dating the Mahābhārata relative to the
dharmasūtras and Manu. As Olivelle shows, the Mahābhārata knows the āśramas
both in their “original system” (1993, 153–55) known to the dharmasūtras, where
they are four different lifelong choices (vikalpa) to be made before marriage, and
in their “classical system” (148–51) favored (though not exclusively) by Manu
(129), which staggers the four through a male’s life (see chapter 5 § E). I think
that in airing both systems, the Mahābhārata brings them under debate such as
Olivelle himself mentions (69–70), taking them up au courant with their prior
treatment in the earlier dharmasūtras some time between 150 BCE and the turn
of the millennium, and probably soon before Manu further codifies them. The
Buddhacarita’s view that there is no wrong time for dharma then looks to be a
typically Buddhist expression of the pro-choice position that Manu, unlike the
dharmasūtras and Mahābhārata, seeks so energetically to suppress (Olivelle 1993,
131–36, 147, 176). Indeed, while making a negative evaluation of this position, the
Mahābhārata includes a prophesy to King Māndhātar by Viṣnu ̣ in the guise of
Indra that would seem to link free choice of āśramas (āśramānạ ṃ ̄ vikalpāh)̣ with
the proliferation of Buddhists (bhikṣavo liṅginas tathā) after the passing of the
Kṛta yuga (Mbh 12.65.25). In any case, I do not share Olivelle’s 1993 acceptance of
a period of eight centuries of Mahābhārata composition (148), or, as will become
clear in what now follows, his view of the “admittedly late didactic sections of the
Mahābhārata” (161). And indeed, Olivelle seems recently to have been rethinking
these very matters (2005a, 5–6, 23–24, 37–38).

F. The Buddhacarita and the Mahābhārata

Johnston sees the Mahābhārata as posing different problems from the

Rāmāyaṇa, proposing that Aśvaghoṣa might know it in a form no longer

59. Divam, svarga; BC 7.18–26, 48–53. That svarga is a this-worldly condition is emphasized from the beginning
when we learn that King Śuddhodana’s kingdom was like svarga to his subjects upon his son’s birth (BC 2.12–13).
646 dharma

available to us (2004, xlvi), perhaps even in an early “kāvya form, which is now
irretrievably lost to us,”60 and noting that, “As for proper names, allusions to
the main characters are very thin” (xlvi–xlvii). Johnston is certainly right that
the Buddhacarita is nearly silent on the Mahābhārata’s main story. The text
does not mention Arjuna, Yudhiṣṭhira, Draupadī, Duryodhana, Karṇa, and so
on. And given that fact, we can go even a little beyond Johnston and say that
Aśvaghoṣa is not really interested in touching base with any of this epic’s high
dramas, as he is with Rāma’s departure from Ayodhyā. Yet Aśvaghoṣa does
refer to “the entire destruction of the Kurus” at BC 11.31; “to Bhīṣma for a story
known to the Harivaṃśa but not to the epic” at 11.18, once again suggesting a
surprisingly early awareness of the Harivaṃśa;61 to Pāṇḍu and Mādrī at 4.79;
and to “many legends . . . found in the MBh, but not always in quite the same
form” (xlvii). Curiously, Johnston neglects to mention references to Vyāsa and
Kṛṣṇa, most of which I have noted, and which have one point of interest in that
several of them come, combined with similar Rāmāyaṇa references, near the
Buddhacarita’s beginning and end,62 where we might consider them as points
for his readers’ entry and departure, or as figures whom Aśvaghoṣa references
to frame his text.
Yet once we look past the allusions and negative precedents, we find that
Aśvaghoṣa engages the Mahābhārata for much the same reason as the
Rāmāyaṇa: his interest in the relation between Buddhist and Brahmanical
dharma in connection with questions that bear on the prince’s great departure.
But now the discourse is taken to a higher register: from the constraints of the
prince’s tussle with his father over the royal protocols for gṛhasthadharma and
the ascetic regimes of the forest-dweller, we move on to the search for “the true
dharma.” From the time that the anchorites in the penance grove tell the prince
to seek out Arāḍa Kālāma through his meetings en route with his father’s two
emissaries and King Śreṇya-Bimbisāra of Magadha, and finally, after his meet-
ing with Arāḍa and the period the prince performs penances, the challenge of
Māra, the prince’s quest for mokṣa takes hold. And with it, we find what I would
propose are two kinds of close but indirect readings of the Mahābhārata: one
concerning some of its “didactic” teachings mainly about mokṣa,63 and one

60. Johnston 2004, xlvii; I would just remark that this type of explanation leads us nowhere.
61. See Johnston 2004, xlvii, 152. Bhīṣma himself tells of slaying Ugrāyudha, who killed many Pañcālas
and demanded Bhīṣma’s “mother” Satyavatī, called Gandhakālī here, after Bhīṣma’s father Śaṃtanu died (HV
15.28–68). Again, a distinctive story suggests a reference to the Harivaṃśa as a text. Cf. chapters 7 § A and 12 §§ A
and C on what could be still earlier references to the Harivaṃśa in the Yuga Purāṇa.
62. On Vyāsa, see not only BC 1.42 but also 4.16 (discussed above) and 4.76 (implied); on Kṛṣṇa see 1.45,
28.28–29.
63. The term nirvāṇa is barely used in the first half of the Buddhacarita.
 aśvaghoṣ a ’s buddhacarita 647

referencing an early Mahābhārata episode that I have already mentioned, the

killing of Jarāsaṃdha, king of Magadha.64 Let us look first at the latter.

F.1. Entering Magadha

Despite the anchorites’ admonition that the prince should head north to pursue
the highest dharma, and take not a step toward the south (BC 7.41), he proceeds
south into the Magadha capital of Rājagṛha, ruled by Bimbisāra, on his way
toward Arāḍa Kālāma’s hermitage in the Vindhyas.65 Certain verses describing
his approach are interesting:

6. On seeing him, the gaudily-dressed felt ashamed and the

chatterers on the roadside fell silent; as in the presence of Dharma
incarnate none think thoughts not directed to the way of salvation, so
no one indulged in improper thoughts. (“No one had an improper
thought,66 as if they were in the presence of dharma in physical
form.” Olivelle 2008, 181)

9. And Rājagṛha’s Goddess of Fortune was perturbed on seeing him,

who was worthy of ruling the earth and was yet in a bhikṣu’s robe,
with the circle of hair between his brows, with the long eyes, radiant
body and hands that were beautifully webbed.67

For the very first time, Aśvaghoṣa describes the prince as “dressed,” or
“disguised,” as a bhikṣu (bhikṣuveṣam), just like the śramaṇa who appeared
before him in that guise as the fourth sign. Indeed, that it was a guise for the
śramaṇa is emphasized in the Nidānakathā, which remarks that it was a sign
of things to come sent from the gods, since there were no bhikkhus at the time
of the fourth sign’s appearance (Warren 1998, 57). Along his way, the prince

64. See briefly chapters 5 § G, 7 § A.1, 12 § F. Jarāsaṃdha must be eliminated before Yudhiṣṭhira’s
Rājasūya, since he stands in the way of that ritual’s objective of establishing universal sovereignty over other
kings.
65. BC 7.57; see 7.58: leaving the penance grove, he “proceeded on his way,” presumably, as pointed out to
him, toward Arāḍa’s hermitage at Vindhyakoṣṭha (7.54), which Johnston locates in the Vindhyas, noting evidence
that the Vindhyas may have been the site of a Sāṃkhya school associated with the name Vindhyavāsin (2004,
102, n. 54), whom Larson and Bhattacharya date to ca. 300–400 CE (1987, 15, 143). Arāḍa/Ārāḍa never seems that
far south in other sources.
66. Johnston’s “not directed to the way of salvation” is just paraphrase.
67. BC 10.6. taṃ jihriyuḥ prekṣya vicitraveṣah ̄ /̣ prakīrṇavacaḥ pathi maunam īyuḥ/dharmasya sākṣad̄ iva
saṃnikarṣe/ na kaścid anyāyamatir babhūva//. . . 9. dṛṣṭvā va sorṇabhruvam āyatākṣaṃ/ jvalacharīraṃ śubhajālahastam/
tam bhikṣuveṣaṃ kṣitipālanārhaṃ/ saṃcuksubhe rājagṛhasya lakṣmīḥ.
648 dharma

stills the improper thoughts of those who see him appear “like Dharma incar-
nate” or “like dharma in physical form.” Since we have established that
Aśvaghoṣa is writing for both Brahmanical and Buddhist audiences, I feel no
need to arbitrate here between a mythological Dharma and an impersonal
dharma. If the former implies some hint of a divine plan, that would be
consonant with the chastened thoughts being those not only of the city’s bon
vivants but of Rājagṛha’s Goddess of Fortune (lakṣmī), who understands that,
despite the bhikṣu dress or guise, the Bodhisattva is fit to rule the earth. When
Bimbisāra, who might thus have reasons for concern, sees him too from a
palace balcony, he orders an officer to report on the prince’s movements. The
prince moves calmly, now begging for food apparently for the first time—that
it is the first time is suggested in the Nidānakathā, where he has to force
down some almsfood that is disgusting (see Nakamura 2000, 124–25)—
accepting what comes to him without distinction. Taking his meal at a lonely
rivulet (Aśvaghoṣa does not, like the Nidānakathā, have him nearly vomit),
from there he climbs Mount Pāṇḍava (BC 10.13–14). Hearing of this destina-
tion, Bimbisāra, who is now described as pāṇḍavatulyavīryaḥ—which
Johnston translates, “in heroism the peer of Pāṇḍu’s son,” but which can
better be simplified to “in heroism equal to a Pāṇḍava” or “equal in valor to
the Pāṇḍavas” (Olivelle 2008, 85)—then ascends the same Pāṇḍava Mountain
(17), where he sees the Bodhisattva (18) sitting “in majestic beauty and tran-
quility like some being magically projected by Dharma” (taṃ rūpalakṣmyā ca
śamena cāiva dharmasya nirmaṇam ivopaviṣṭam; 19). Or, quite sufficiently
put: “He sat with calm and resplendent beauty, like the image of dharma”
(Olivelle 2008, 185).
Although I have found no one who has given it a moment’s notice, Buddhist
tradition itself thus makes one of the five peaks of Rājagṛha, or at least part of
one of them, the Pāṇḍava Mountain. This is the case already in the Suttanipāta,
from the Khuddaka Nikāya, which, usually accepted as part of the Pāli canon, is
certainly older than Aśvaghoṣa, and seems to be a basis for the more developed
account mentioning the same mountain in the Nidānakathā.68 The latter is
ascribed to the fifth-century AD, although Johnston thinks that Aśvaghoṣa
“may be presumed to have used an earlier version [of it], no longer in existence”
(2004, xl), as one of his sources. In the Mahābhārata, not surprisingly, there is
no mountain by that name. Rather, when Kṛṣṇa, Arjuna, and Bhīma approach
Magadha to kill Jarāsaṃdha and reach a certain Mount Goratha, they set eyes

68. See Nakamura 2000, 122, 124; Thomas [1927] 2000, 68. Mount Pāṇḍava is also a stable fixture in The
Gilgit Manuscript of the Saṃghabhedavāstu, being the seventeenth and Last Section of the Vinaya of the Mūlasarvāstivādins
(Strong 2001, 14) and the Lalitavistara (to judge from Poppe 1967, 134).
 aśvaghoṣ a ’s buddhacarita 649

on “Magadha city” (Mbh 2.18.30), which Kṛṣṇa describes as having “five

beautiful mountains: the wide Vaihāra, Varāha, Vṛṣabha, Ṛṣigiri, and Caityaka”
that “stand guard over Girivraja.”69 We note the Mahābhārata’s name for the
city is Girivraja, not Rājagṛha.70 The Mahābhārata means by this name not just
the Magadha capital but the “mountain corral” (giri-vraja) where Jarāsaṃdha
keeps eighty-six of the world’s hundred kings imprisoned (see Biardeau 2002,
vol. 1, 327).
Buddhist tradition thus references the Pāṇḍavas, and one may assume the
Mahābhārata, and in all likelihood the Jarāsaṃdha episode, when it has the
prince cross the Pāṇḍavas’ tracks on Pāṇḍava Mountain.71 From this, the most
straightforward assumption would be that the Buddhists have named as
“Pāṇḍava Mountain” the mountain, or at least part of the mountain, that
Kṛṣṇa, Arjuna, and Bhīma ascend: the Caityaka Peak, as seems to be borne out
by details supplied by H. W. Schumann.72 Yet Aśvaghoṣa goes beyond other
Buddhist sources in describing the Bodhisattva’s trek here. And though it is
not a matter one can demonstrate with a perfect parallel fit, it seems from
some of Aśvaghoṣa’s new themes, similes, and points of emphasis that he
does so not only out of a residue of folklore but with a Mahābhārata “textually”
in view.

69. Mbh 2.19.2–3. As noted by Brockington 2002, 79, a five-verse Southern Recension insertion
amplifies the description of the mountains (2.206*, after 2.19.10), but adds nothing noteworthy for our pur-
poses unless perhaps that Caityaka is giriśreṣṭha, “the best of peaks” (line 2), and that the five are now num-
bered as Pāṇḍara (presumably Vṛṣabha, unless, perhaps under Buddhist influence, this interpolation is
trying to find an alternate place for an intentionally disguised or just garbled “Pāṇḍava” mountain), Vipula,
Vārāha, Caityaka, and Mātaṅga (Ṛṣigiri); the latter reminding us perhaps of the “untouchable” Ṛṣi Mataṅga
of a forest hermitage near Kiṣkindhā in the Rāmāyaṇa, who has a splendid mountain named after him at
Vijayanagar.
70. Biardeau 2002, 1: 330 introduces a little uncertainty as to whether Girivraja and Rājagṛha are the same,
but that they are early and later names for at least parts of the same city seems well enough established. See van
Buitenen 1975, 15–16; Lamotte 1988, 17–18; Schumann 1989, 90. The Buddhacarita uses both Rājagṛha (10.1 and
9) and Girivraja (11.73). For the Mahābhārata to use only Girivraja is probably an archaism.
71. Indeed, if one assumes that the Buddhist tradition works from oral Magadha stories before the epic’s
written text—one would presumably have to presume a proto-Jarāsaṃdhavadha—then the Suttanipāta account
may be older than the Mahābhārata, since the Suttanipāta is thought to present some of the earliest sources on the
Buddha legend (Lamotte 1988, 660; Nakamura 2000, 19, 123–24, 131–34; Thomas [1927] 2000, 273). From this
standpoint, the Mahābhārata would still remain within its game plan if it concealed the name “Pānd ̣ ạ va
Mountain.”
72. That is, by correlating the map in Schumann 1989, 90 with what he says on p. 46: “the Pāṇḍava
hill, the north-easterly of the five hills surrounding Rājagaha.” The map names six mountains around “Old
Rājagaha or Giribbaja”: Vaibhara to the west, Vipula north, Rama northeast, Chattha with the Vulture Peak
to the east, Udaya southeast, and Sona southwest. Chattha Mountain would thus be in the right position to
be both the likely alternate for Caityaka and another name for Pāṇḍava, although the map does not show this
latter name. Note that Vaibhara is the only other mountain with a similar name in both texts. Rājagṛha
became the site of “eighteen vast monasteries” (Lamotte 1988, 17–18 (19)—presumably vihāras, from which
comes also the name Bihar). Lodhra trees cover the Pāṇḍava Mountain (BC 10.15), or all five peaks
(Mbh 2.19.4).
650 dharma

Rather than go over the Jarāsaṃdha episode in detail, as several have

done,73 I present the following chart of parallels and oppositions, which should
suffice to give a basic idea of why a journey of two Pāṇḍavas and Kṛṣṇa to
Magadha would have interested Buddhists before Aśvaghoṣa. Further, by
accenting what appear to be Aśvaghoṣa’s most important innovations in bold
face,74 it should afford a basic idea of what interested him in the Mahābhārata’s
Jarāsaṃdhavadha episode in particular. I suggest that one first read the unac-
cented sequence that reflects the prior Buddhist story (items 1, 2, 4, 6, 8, 11,
and 14), and then the whole alignment (charted on the next two pages) to see
what Aśvaghoṣa seems to have made of it.
The first thing worth noting is that Aśvaghoṣa introduces an epic tone.
Magadha’s goddess of Fortune or Lakṣmī shows her favor on the prince (item
BC 5), and Bimbisāra challenges the prince to take up arms against his foes with
Bimbisāra as an ally (item BC 12). The challenge is particularly gratuitous, and
when it is noticed that Bimbisāra makes it upon seeing the prince in the guise
of a bhikṣu, one gets a good index that Aśvaghoṣa is taking the Pānḍ ạ vas’ bath-
graduate guises as his epic touchstone (see items Mbh 2, 5, 10, 12).75 In each case
it is a matter of responding to a challenge posed by thinly disguised Kṣatriyas: in
one case three Kṣatriyas disguised as snātakas (admittedly Kṣatriya ones but
adopting a predominantly Brahmin role);76 in the other a prince in a mendicant
garb that some texts, including some passages in the Mahābhārata, say should

73. See Biardeau 2002, vol. 1, 324–54 and vol. 2, 755–58, for her most recent discussion; Brockington
2002; van Buitenen 1975, 11–18; Hiltebeitel 1989, 2005b. I am not persuaded by Brockington’s method of dating
the whole episode as “late” and “added”: he seems to accept the criterion of “grounds of content” [73], and
includes among his own criteria, “starting from the premise that [it] . . . is anomalous” [74], that it is “extraneous
to the plot of the MBh” [80], and, as is most symptomatic of this method, observing that it “reflects relatively late
Vaiṣṇava–Śaiva opposition” [82]). Nonetheless, he proposes for its composition an “immediately post Mauryan”
Śuṅga date (2002, 84–85) of “the later part of the 2nd century or, perhaps most probably, the first century B.C.”
(86). Such a date for me is not, however, late; rather, it is attractive for the larger Mahābhārata archetype, parts
and whole, which, as Brockington mentions (79), includes the episode (see Hiltebeitel 2001a, 20–31; 2011a,
chapters 4 and 5).
74. In these determinations, I have consulted the treatment in other versions in Nakamura 2000, 120–24;
Thomas [1927] 2000, 68–70; Strong 2001, 10–18; and Poppe 1967, 133–42. On the other hand, it is fascinating
to see how all of Aśvaghoṣa’s clearly “Indian” nuances are lost in the Sanskrit-to-Chinese translation; see Beal
1968, 111–19.
75. This garb could also have more indirect Rāmāyaṇa echoes, since both Rāvaṇa 3.44.8; 47.6) and
Hanumān (4.3.8; 3.21; 5.14) make rather famous turning-point appearances “in the form of a bhikṣu
(bhikṣurūpa).”
76. See chapter 5 § G on snātakas in connection with this episode. Cf. Olivelle 1993, 220–21: A snātaka
“is considered so sacred and his status so eminent, that many authorities give him precedence over even a
king: if a king meets him on the road it is the king who should salute the latter with respect” (with citations).
Van Buitenen actually wonders whether “the meaning of snātaka might be extended to anyone under a stu-
dious vow of life, and to include the new mendicants who followed the Buddha or Jīna, but that cannot be
made out” (1975, 17).
 aśvaghoṣ a ’s buddhacarita 651

Buddhacarita [BC] Mahābhārata [Mbh]

1. The prince enters the city of the five hills 1. Kṛṣṇa and the two Pāṇḍavas, Bhīma and
(10.2). Arjuna, approach the city of five hills.
2. The prince makes his first appearance 2. Following Kṛṣṇa’s counsel, the three are
dressed (or disguised) as a bhikṣu (9). disguised as snātaka Brahmins (2.18.21).
3. He seems to onlookers like Dharma 3. In Kṛṣṇa is prudent policy (naya, nīti), in
incarnate/like dharma in physical form (6). Bhīma strength, in Arjuna victory (14.9,
18.3). Prudent policy turns out to have been
tricky dharma (see item 14 below).
4. First amazed (visismiye), onlookers then 4. Onlookers “fell to wondering” (vismayaḥ
fall still and silent and have no unruly samajāyata) (19.27) and are at first baffled.
thoughts (anyāyamatir) (2–6).
5. The city’s Lakṣmī shows favor on the 5. Kṛṣṇa soon reveals that Śrī favors Kṣatriya
prince (9). snātakas who wear garlands (19.46).
6. The prince climbs Pāṇḍava Mountain 6. The two Pāṇḍavas and Kṛṣṇa climb
(14). After receiving a report of his ascent, Caityaka Mountain.
so does King Bimbisāra, “in heroism equal
to a Pāṇḍava” (17).
7. There Bimbisāra sees the tranquil 7. There they destroy the “horn” (śṛṅgam) of
cross-legged Bodhisattva “being as it were a Caityaka Mountain (19.18).
horn (s´ṛṅgabhūtam) of the mountain” (18).
8. Bimbisāra thus shows deference and 8. The two Pāṇḍavas and Kṛṣṇa come to
hospitality by coming to the mountain. King Jarāsaṃdha’s palace, where they reject
his hospitality (19.34).
9. The prince looks to the king “like some 9. Jarāsaṃdha asks about the dharma of this
being magically projected by Dharma”/“like rejection, and about the trio’s disguises.
the image of dharma” (19).
10. The king and prince debate about 10. Kṛṣṇa and Jarāsaṃdha air opposing
dharma: Bimbisāra aligns the trivarga with views of dharma: Kṛṣṇa reveals they are
age: pleasure for youth, wealth for middle Kṣatriya snātakas, and hold him as enemy;
years, dharma for old age (33–37); but the asking why, Jarāsaṃdha protests himself a
prince, seeing danger in old age and death, ruler by dharma (20.3–5). Kṛṣṇa says the trio
“resorts to this dharma out of longing for follows dharma in opposing Jarāsaṃdha’s
release (mumukṣayā)” (11.7). He should do plan to sacrifice 100 kings to Rudra (20.9),
kuladharma and offer sacrifices (10.39–40); but Jarāsaṃdha sees it as Kṣatriya dharma to
but he does “not approve of sacrifices” or of treat captives as one pleases (20.26).
“happiness sought at the price of another’s
suffering” (11.64–67), etc.
11. Bimbisāra offers the prince half his 11. Kṛṣṇa’s plan will eliminate Magadha’s
Magadha kingdom (10.25–26), which the sovereignty so that Dharmarāja Yudhiṣṭhira
prince explicitly rejects (11.49–56). can be universal monarch by performing a
Rājasūya sacrifice.
12. Bimbisāra also challenges the prince to 12. Kṛṣṇa challenges Jarāsaṃdha to fight one
fight his foes; moved as he is by compassion of the trio in the guise and garb of bath-
at seeing him, a Kṣatriya, in the garb or guise graduates, now revealing who they are
of a mendicant (bhikṣu), Bimbisāra would be (20.23–24).
the prince’s ally (10.27–32).
13. The prince implicitly rejects such a fight. 13. Jarāsanṃdha chooses to fight Bhīma
(21.3), as Kṛṣṇa had devised (20.32–34).
(continued)
652 dharma

(continued)

Buddhacarita [BC] Mahābhārata [Mbh]

14. The prince promises to come back as a 14. The freed kings imprisoned in Girivraja,
Buddha (11.72–73), at which point he will recognize that Kṛṣṇa protects dharma, and
preach the dharma that converts Bimbisāra that he is Viṣṇu (22.31–32).
and many other Magadhans.

be restricted to Brahmins.77 Moreover, Bimbisāra challenges the Bodhisattva

precisely while calling attention to his appearance, calling him bhikṣaś̄ ramakāma,
“lover of the mendicant stage of life” (10.33), thereby providing the one instance
in the text where āśrama clearly means “mode” or “stage of life” rather than
“hermitage.”78 Just as King Śuddhodana tells the prince not to go against the
“proper order” of the implied āśramas, so now Bimbisāra seconds the point with
additional unusual arguments correlating the triple set with the life-stages (item
B 10),79 and with this specific challenge to the Bodhisattva’s appearance as a
bhikṣu. Both kings are making a “legitimate” point, for they would be speaking
as “protectors of varṇśramadharma,” a role that even Buddhist kings come to
play in sixth-century inscriptions (Olivelle 1993, 201–4).
Next, as one would expect of an accomplished kāvya poet, Aśvaghoṣa tips his
hand further with his similes. To begin with, when the prince has climbed Mount
̣ ạ va, “On that mountain (avau),80 . . . he, the sun of mankind (nṛsūrya),
Pānd
appeared in his ochre-colored robe like the sun in the early morning (bālasūrya)
above the eastern mountain.”81 As Gawronski puts it in the only scholarly note
I have found on these matters, “the future Buddha standing on [or, better,
ascending] the Pāṇḍava mountain, clad as he is in his red garment, is compared
to the rising sun touching the verge of the eastern mountain” (1914–15, 37).

77. See Olivelle 1993, 195 and n. 40, noting that “there are numerous texts in the Mahābhārata that declare
religious mendicancy to be the special dharma of Brahmins: 3.34.49–50; 5.71.3” [both addressed to Yudhiṣṭhira],
and pointing to Mbh 12.10–25 where this point is made to Yudhiṣṭhira at the beginning of the Śāntiparvan.
Although never using the compound bhikṣāśrama, the Mahābhārata sometimes uses bhikṣu or bhikṣuka to cover
the fourth life-stage (12.14.12; 12.37.28; 14.45.13).
78. Aśvaghoṣa could have the precedent of Gautama Dharmasūtra 3.1 and 3.11–25 in identifying a
Brahmanical “life-stage” with the term bhikṣu.
79. Bimbisāra’s correlation of three periods of life with the trivarga (item 10) is interesting as being not
reducible to the āśrama system, and as having a counterpart in Kāmasūtra 1.2.1–6—but there with different cor-
relations: youth should be devoted to aims (artha) such as learning, prime years to kāma, and old age to dharma
and mokṣa (see Olivelle 1993, 30–31 n. 85, 133, 218).
80. Gawronski 1914–15, 37 had noted that some word for “mountain” was necessary, and proposed girau
rather than vane, “in the forest,” having read the latter in Cowell’s edition and translation. See Cowell 1968, 106
and Johnston 2004, 143 n. 15 confirming avi as “a certain reading” based on his primary manuscript and the
Tibetan translation. Olivelle 2008, 283 reads avi as “hill.”
81. BC 10.15: tasminnavau lodhravanopagūḍhe/ mayūranādapratipūrṇakuñje//kāṣāyavāsāḥ sa babhau
nṛsūryo/ yathodayasyopari bālasūryaḥ.
 aśvaghoṣ a ’s buddhacarita 653

That is how Aśvaghoṣa describes what King Bimbisāra’s officer sees (10.16), and
perhaps what the officer reported back to Bimbisāra. But now, when Bimbisāra
himself ascends this same mountain with the heroism of a Pāṇḍava, he sees the
tranquil cross-legged Bodhisattva “being as it were a horn (śṛnġ abhūtam) of the
mountain” (item BC 7). That is, the rising sun of mankind has become the “horn”
of the very Pānd ̣ ạ va Mountain he and Bimbisāra have just climbed. I cannot ima-
gine that Aśvaghoṣa has any other first pretext for introducing82 this singular,
surprising, and somewhat strained simile than a reference to the Mahābhārata’s
double use of śṛnġ am to describe what it is on Caityaka Mountain that the two
Pānḍ ạ vas and Kṛsn ̣ ạ destroy (items Mbh 6–7).83 Van Buitenen takes both usages
as “tower” (1975, 69, 70), which is certainly a guess. Biardeau (2002, 1: 351) also
calls attention to a verse (Mbh 2.208*) that has the trio break three drums (bherī)
and the wall of a caitya (caityaprakāram) on the peak, but this verse is found only
in four manuscripts, including the Vulgate (which Biardeau favors), and is clearly
an interpolation. The three drums made by Jarāsaṃdha’s father are mentioned
just before the insertion (2.19.15–16), but without the interpolated verse that
follows; there is nothing to say they were destroyed, and nothing about a Caitya
wall, which is clearly a belated explanation built on the mountain’s name.84 On
the contrary, Kṛsn ̣ ạ establishes the first meaning of śṛnġ a for the whole passage
when he describes Girivraja’s five mountains as all having “great horns and cool
trees” (mahāśṛnġ āḥ parvatāḥ śitaladrumāh;̣ 19.3).
Aśvaghoṣa could also have a second pretext for using the word śṛnġ a to describe
the tranquilly seated prince: the word’s symbolic significance is brought out in a
Harivaṃśa passage that asks a question about the same Jarāsaṃdha cycle:

To what end did the slayer of Madhu (Kṛsn ̣ ạ ) abandon Mathurā, that
(zebu)’s hump of the Middle Country, the sole abode of Lakṣmī, easily
perceived as the horn of the earth (śṛnġ a pṛthivyāh),
̣ rich in money and
grain, abounding in water, rich in Āryas, the choicest of residences?85

82. Note that the Suttanipāta uses different images when the king’s messengers report back and say,
“Great king, the bhikkhu sits in a mountain cave on the front side of Mount Pāṇḍava, like a tiger or a bull or a
lion” (Nakamura 2000, 122)—a scene that could also evoke girivraja as the “mountain corral” in which Jarāsaṃdha
imprisons the eighty-six kings. See nn. 69, 70 above on the name Girivraja and the possibility that the Suttanipāta
could precede the epic text.
83. At Mbh 2.19.18, this śṛṅga is described as garlanded, and at 19.41 Jarāsaṃdha mentions it again when
he asks how the trio broke it (caityakaṃ ca gireḥ śṛṅgaṃ bhittvā kim; 19.41).
84. Kosambi must pick up on some such tradition when he writes, “But the senseless desecration of the
holy antique caitya at Rajgir (presumably the Pāsāṇaka Cetiya where the Buddha rested so often) by Bhīma and
Kṛṣṇa seems wanton sacrilege (Mbh 2.19.19), unsupported by any other record” (1964, 36–37; 1975, 126), on
which Brockington comments, “Why he should see the reference to the monument as being a Buddhist caitya is
equally unclear” (2002, 79–80).
85. HV 1.57.2–3: kim arthaṃ ca parityajya mathurāṃ madhusūdanaḥ/ madhyadeśasya kakudaṃ dhāma lakṣmyāś
̣ yajalabhūyiṣṭam adhiṣṭhānavarottamam.
ca kevalam//śṛnġ a pṛthivyāḥ svālakṣyaṃ prabhūtadhānadhanyavat/ āryādh
654 dharma

This “horn of the earth,” along side the zebu’s hump as the sole abode of
Lakṣmī, evokes associations of Kṛṣṇa with the horn in contested situations
where he uses his Śārṅga bow in battles, and, even more particularly, associa-
tions with Viṣṇu’s Fish and Boar avatāras where he uses the “single horn” or
“single tusk” (in either case, ekaśṛṅga) to rescue Manu’s ark and the earth.86 In
other words, in the Jarāsaṃdha cycle, the horn is a symbol of unique sover-
eignty in contested circumstances, which makes it fitting that Kṛṣṇa and the
two Pāṇḍavas break the horn of Magadha’s Caityaka Mountain—no matter
how difficult it is to imagine—with their bare arms.87 For they are intent, in the
Mahābhārata’s terms, upon eliminating Jarāsaṃdha’s rivalry of Yudhiṣṭhira
for the title of universal sovereign (saṃrāj), and, in the Harivaṃśa’s terms, upon
restoring the unique centrality of Mathurā to the Middle Country (madhyadeśa),
even in Kṛṣṇa’s absence from it.
At one level, what is being contested in the Buddhacarita is thus, of course,
royal sovereignty, Lakṣmī, who favors the prince even though he declines royal
sovereignty when Bimbisāra offers it. But as Aśvaghoṣa registers in further
similes, in fact by doubling one simile, what is really contested is the dharma:
the prince seems to onlookers “like Dharma incarnate” or “like dharma in
physical form” (item BC 3), and to Bimbisāra he looks “like some being magi-
cally projected by Dharma” or “like the image of dharma” (item BC 9).88 This is
the force of the way Aśvaghoṣa unfolds this matter as one that has to do not
with a debate about the Śaiva–Vaiṣṇava overtones of Kṣatriya dharma, such as
occurs between Jarāsaṃdha and Kṛṣṇa in the Mahābhārata (items Mbh 10, 14),
but one that has to do quite explicitly with oppositions between Brahmanical
royal dharma and Buddhist dharma—the latter as it is, so to speak, taking shape
in the Bodhisattva-prince’s mind and the minds of those who see him. But the
force of the Mahābhārata story, given the no doubt intended ambiguity of the
term caitya, which can have both Brahmanical and Buddhist meanings,89 and
given as well the results of over a century of scholarship that has sensed this
ambiguity,90 is that it can be taken not only as a story reflecting Śaiva–Vaiṣṇava
opposition but Brahmanical–Buddhist opposition as well.

See Hiltebeitel 1989 on this verse and the answer to the question raised in it, which I now bring to bear on
the discussion here.
86. See Hiltebeitel 1989, 96, citing Defourny 1976, 17–23.
87. That is indeed how Ganguli translates the passage ([1884–96] 1970, vol. 2, Sabha Parva, 52).
88. Cf. Saundarananda 2.56cd: the Buddha at birth “shone with the majesty of holy calm like the Law of
Righteousness in bodily form (babhrāje śāntayā lakṣmyā dharmo vigrahavān iva).”
89. Biardeau richly develops this point; see now 2002, 1: 322 n. 2; 344 (with Gṛhya Sūtra references);
330–31; 350.
90. See Brockington 2002, 79 and Hiltebeitel 2005b, tracing this impulse to (the younger) Adolf
Holtzmann 1892–95, and above, n. 84, on Kosambi.
 aśvaghoṣ a ’s buddhacarita 655

That brings us to a third pretext for Aśvaghoṣa’s surprising “horn” simile.

For when one takes the force of the “horn” and “Dharma/dharma” similes in
conjunction with the fact that it is Bimbisāra, not the prince, who is made
“equal to a Pāṇḍava in heroism” and who sees the Bodhisattva as if he had
become the horn of the mountain, one could take it that Bimbisāra sees not
only Dharma incarnate or dharma in physical form but a cross-legged
Bodhisattva appearing as the restored horn of the mountain that the Pāṇḍavas
and Kṛṣṇa broke down.
As I attempted to show in Rethinking India’s Oral and Classical Epics, north
Indian Ālhā traditions, both in this Hindi oral epic and in the Bhaviṣya Purāṇa’s
retelling of Ālhā in Sanskrit, draw on Ismaili traditions to transpose the
Jarāsaṃdhavadha into a Rājpūt rivalry that was also read in terms of opposi-
tion over empire, in this case between Hindu and Muslim rule.91 The
Mahābhārata episode has been open to such readings because it has to do with
religious overtones of rivalry over empire, which itself is one of the reasons it
cannot be persuasive, no matter how many stylistic or “late devotional” criteria
one enlists, to argue that the Jarāsaṃdhavadha is extraneous to the
Mahābhārata.92 This is the real hinge upon which Aśvaghoṣa opens his close
reading of this episode. For although it may look like a weak point to align the
Bodhisattva with Kṛṣṇa on the matter of the Bodhisattva’s double appearance
as Dharma incarnate or dharma in physical form, we are at the deepest level at
which Aśvaghoṣa engages this Mahābhārata scene: the level of Brahmanical
versus Hindu bhakti, which we have seen underscored by Biardeau. The posi-
tion of Kṛṣṇa in representing Brahmanical dharma in the Jarāsaṃdha episode
is decisive. For the first thing to strike one is that the Mahābhārata’s actual
“Dharma incarnate” and embodied exemplar of dharma, Dharmarāja
Dharmaputra Yudhiṣṭhira, is precisely not among the trio assaulting Magadha,

91. Hiltebeitel 1999a, 150–64 and 344–51 on Ismaili gināns about the Buddha and “Kalinga” (an allomorph
of Jarāsaṃdha), though the stories do not relate the two directly. Cf. Khan 2005: although the gināns do not
mention the Buddha’s preenlightenment entry into Magadha, they bring him in to address Yudhiṣṭhira’s postwar
consternation (= Mbh Book 12, etc.); and when he comes before the Pāṇḍavas he “has a very strange appearance:
apart from posing as a religious mendicant, he looks like a warrior, donning Muslim dress. . . . Besides he is a
caṇḍāla . . . and a leper, from whose body emanates an unbearable odour” (2005, 328; cf. 330, 333, 340). After he
challenges Bhīma at the Pāṇḍavas’ gate, his Satpanth Ismaili teachings are rich in overtones of bhakti and are
presented as dharma (329, 333). Undercutting the Brahmins who are performing a “huge sacrifice” on
Yudhiṣṭhira’s behalf, he says “their sacrifice is useless” (as does the half-golden mongoose at the end of Mbh
Book 14); yet before he retires to the Himalayas he convinces the Pāṇḍavas to sacrifice a cow (none other than the
Kāmadhenu or “Cow of Wishes”) for a final shared meal that will make possible their liberation (128–31). As
Khan says, the gināns may draw not only on Hindu sources but Buddhist ones (326, 337–41)—one wonders, with
what ironies.
92. See Hiltebeitel 2001a, 8, noting that “this sequence provides in a flurry most of the Mahābhārata’s
usages of the terms saṃrāj, ‘emperor,’ and sāṃrājya, ‘empire,’” and mentioning some of the scholars who con-
tinue to hold the view that it is late and extraneous, which has now been restated by Brockington 2002.
656 dharma

among whom, as Bhīma says first and Kṛṣṇa then confirms, Kṛṣṇa represents
prudent policy (naya, nīti), Bhīma strength, and Arjuna victory (item Mbh 3;
Mbh 2.14.9; 18.3). Yet what Yudhiṣṭhira says before the trio departs is perti-
nent to this train of associations. Fearing Jarāsaṃdha’s might and ready to
change his mind about performing a Rājasūya, Yudhiṣṭhira says, “Bhīma and
Arjuna are my two eyes, Janārdana I deem my mind (manas); what kind of life
shall be left for me without mind or eyes (manaś cakṣur vihīnasya)?” (Mbh
2.15.2).93 Kṛṣṇa supplies “policy” (naya, nīti) that will turn out to be tricky
dharma, or more precisely upāyadharma (item Mbh 14)—a “dharma of strategy”
or “means,” such as Kṛṣṇa often deploys.94 But it will be done fully in accord
with the mind of King Dharma. Indeed, the two verses that identify Kṛṣṇa
with policy and Arjuna with victory resonate with the famous tag line that first
occurs right after the Bhagavad Gītā when Droṇa tells Yudhiṣṭhira, as if he
needed to know it, “Where dharma is there is Kṛṣṇa; where Kṛṣṇa is there is
victory.”95 In short, Aśvaghoṣa’s reading of the Jarāsaṃdha episode could be
summed up as follows: where Kṛṣṇa was, there now is the dharma looking
personally like the horn of a mountain.

F.2. Buddhist Mokṣadharma

It may be no mere coincidence that Aśvaghoṣa focuses on pivotal matters

bearing on Rāma and Kṛṣṇa in the second books of each epic, reading each in
part through a contrast of Brahmanical and Buddhist modes of bhakti. Yet if
structuralism and symbolism are not the most reliable indicators of intertex-
tual history,96 we have another marker of Aśvaghoṣa’s reading of the Mahābhārata

93. See Biardeau 2002, 1: 328 on this passage.

94. On nīti as upāyadharma in the Mahābhārata, see Bowles 2007, 199–204. See especially 199 and n. 33,
citing Mbh 12.101.2 and 128.13, both from the Rājadharma Parvan, but the latter from an adhyāya transitional to
the Āpaddharmaparvan. Bowles comments: “The idea of a dharma of ‘strategy,’ a ‘strategic dharma,’ or ‘an expe-
dient abundant in dharma,’ is, in many ways, collateral with the idea of a proper form of conduct (dharma) for a
king in times of distress, since a king must employ some form of strategy or policy to overcome difficulties that
might arise for his kingdom. Indeed, in a nīti context, upāyadharma could almost be considered a synonym for
āpaddharma.” Although he explains it only as upāya, Kṛṣṇa is of course the master of upāyadharma and of “many
crooked means” (jihmair upāyair bahubhir; see 9.60.29) throughout the Mahābhārata war.
95. Mbh 6.41.55. As mentioned in chapter 11 § D, the line is repeated at 9.61.30, and has the variant,
“Where Kṛṣṇa is, there is dharma; where dharma is, there is victory” at 6.62.34 and 13.153.39.
96. Aśvaghoṣa provides one more piece of possible evidence of familiarity with the Jarāsaṃdhavadha: a
curious pair of verses, one about a certain Kakṣīvat (BC 1.10), of whom Johnston 2004, 3 n. 10 says “nothing
is known”; the other about a certain Manthala Gautama, likewise untraced, who carried corpses to please a
courtesan named Jaṅghā (4.17). These verses may recall some equally obscure verses in the Jarāsaṃdha story
where a Kākṣīvat is fathered on a Śūdra woman by a Ṛṣi Gautama who dwelt at Magadha because he favored
the Magadha vaṃśa, and was also sought out by the Aṅgas and Vaṅgas, also northeastern peoples (Mbh
2.19.5–7).
 aśvaghoṣ a ’s buddhacarita 657

that may provide a more reliable gauge, even though it points to the same his-
torical conclusions. This is Aśvaghoṣa’s interest in the Mahābhārata’s didactic
corpus on mokṣa.
Here we come to a point that several have noticed: Aśvaghoṣa seems to
know the Mahābhārata’s Mokṣadharma Parvan, or at least material in it.97
Johnston cites, without ever making it clear if he ever discusses it, a discussion
by T. Byōdō (n.d.), which has a title to this effect. I have not been able to locate
it, but Muneo Tokunaga (2005a) spoke on this subject at a London conference
a week before I met him at the Dubrovnik conference where I presented a draft
of this chapter, and kindly made his paper available to me when I learned of it.
Tokunaga begins with an acknowledgment of a 1930 book by Tsusho Byodo,
of which Johnston had apparently read an English appendix. Tokunaga
summarizes Byodo’s work as being interested “mainly in philosophical
matters,” with Byodo’s comparison with the Mahābhārata “centered in the
Mokṣadharmaparvan” (2005a, 1). According to Tokunaga,

Results of his comparison of the texts are summarized under five

heads: (1) myths, (2) Sāṃkhya teachers, (3) the topic “a younger one
sometimes supersedes an older in achievement,”98 (4) thought-
historical, rhetorical, linguistic correspondence, and (5) the relation-
ship between the Buddhacarita and the Mokṣadharmaparvan
(pp. 543–64). In conclusion, he says that Aśvaghoṣa was influenced
by the Mokṣadharma in his composition of the Buddhacarita (p. 560).
(Tokunaga 2005a, 1)

For Tokunaga, “this assumption is not impossible,” but he moves on to some

views of Johnston’s: that “it is more natural to suppose that the common matter
goes back to a single original,”99 even though Tokunaga finds Johnston going

97. See Hopkins 1901, 387–88; Larson and Bhattacharya 1987, 7, 14–15, 110–22, 129–40, assuming,
I think wrongly, that Sāṃkhya references in the Mokṣadharma would be from the first to third- or fourth-centuries
CE (113), and thus mostly later than Aśvaghoṣa, even though they date the Mokṣadharma itself to 200 BCE–200
CE (14). Backed by arguments in Hiltebeitel 2005d and 2006a, I will be arguing that even the so-called late
portions of the Mokṣadharma Parvan, including the Nārāyaṇīya, would have probably preceded Aśvaghoṣa.
98. See BC 1.41–45 as cited above toward the beginning of § D of this chapter.
99. Johnston 2004, xlvi, noting that “much of Arāḍa’s exposition of the Sāṃkhya system has close paral-
lels in the Mokṣadharma, the connection in one case extending over several verses of the same passage,” and
suggesting that “the common matter goes back to . . . possibly a textbook of the Vārṣagaṇya school.” As Larson
and Bhattacharya 1987, 131 observe, “Varṣagaṇya” at Mbh 12.306.57 occurs in a list of “many older teachers of
Sāṃkhya and Yoga.” Assuming this list would be from “the first centuries of the Common Era” (Ibid.), they do
not relate Aśvaghoṣa’s portrayal to such a context, but they do note (136, 138), as does Johnston (2004, lvi, 172 n.
33), that at BC 12.33 Aśvaghoṣa may be quoting the aphorism pañcaparvā avidyā, “there are five kinds of
ignorance”—from Vārṣagaṇya, since it is elsewhere attributed to him. They date Vārṣagaṇya to ca. 100–300 CE.
The Mokṣadharma reference should, I think, support the priority of Vārṣagaṇya, if not necessarily their datings.
658 dharma

too far when he states that “despite the many parallels we cannot establish that
Aśvaghoṣa knew any portion of the epic in the form in which we now have it”
(Ibid., xlvii; Tokunaga 2005, 1). I am encouraged by Tokunaga on this point, on
which John Brockington is both more succinct and more extensive: Aśvaghoṣa
“definitely draws on the Śāntiparvan” (1998, 483). I agree with both Tokunaga
and Brockington. I also find very attractive Tokunaga’s demonstration that
Cantos 9 and 10 of the Buddhacarita involve a reading of (Tokunaga says “are
based on”) the first “forty-five or so chapters in narrative form of the extant
Śāntiparvan” (Ibid.). For reasons that will become clear, if he is right, his dem-
onstration reinforces my hypotheses, and I will refer to it as a supportive
argument. It would seem likely to be a question not only of elements of the
Mokṣadharma and the Buddhacarita drawing on some common sources but of
Aśvaghoṣa having read the Śāntiparvan in some state of “extant” totality
involving both its beginning and its last four major units.
One trace of the range of Aśvaghoṣa’s familiarity with the Śāntiparvan
could be his reference at Buddhacarita 8.77 to the story of a Suvarṇaṣṭhīvin,
“Excretor of Gold” (Fitzgerald 2004a, 236–37), who figures in an upākhyāna
told by Kṛṣṇa and Nārada toward the beginning of Book 12.100 This comes not
within the segment of the Buddhacarita that Tokunaga discusses, but in the
canto just before the two in which he finds parallels in the first part of the
Śāntiparvan. I have also discussed the first forty or so chapters of the Śāntiparvan
from another angle (Hiltebeitel 2005d, 249–58): that they present Yudhiṣṭhira
with arguments from the Bhagavad Gītā for him to reject as inadequate in his
postwar situation, while at the same time foreshadowing the need for instruction
that will prove acceptable to him. That will be the instruction that Yudhiṣṭhira
receives from Bhīṣma in the four subparvans that proceed from the Śāntiparvan’s
opening: the Rāja-, Āpad-, Mokṣa-, and Dāna-dharma Parvans which, together,
comprise nearly all of the Śānti and Anuśāsana Parvans. As to the Buddhacarita,
though, Tokunaga is certainly right in turning our main attention to its ninth
and tenth cantos.
Canto 9 is King Śuddhodana’s “Deputation” of his Purohita and Minister to
find his son the prince, and chapter 10 is “Bimbisāra’s Visit,” which we have just
been looking at from another angle. I propose that Canto 9 is a hinge chapter for
Aśvaghoṣa that allows him to transition from a Rāmāyaṇa reading to a
Mahābhārata one. This means that the Purohita and the Minister get to double

100. It is Suvarṇaniṣṭhīvin in Aśvaghoṣa’s spelling. I am not persuaded by Johnston’s point (2004,

120 n. 77) that Aśvaghoṣa’s silence on the son’s coming back to life “suggests that the poet knew only a version
in which the happy ending had not been added.” Aśvaghoṣa is not trying to tell the whole story in one verse but
making what he wants of the story in what is contextually a perfectly intelligible allusion.
 aśvaghoṣ a ’s buddhacarita 659

not only for Rāma’s two Brahmin visitors in the forest but for the postwar com-
forters of Yudhiṣṭhira: the first explicitly, the second only implicitly. Yet this
Mahābhārata reading would not be limited to Cantos 9 and 10 but carry over
from Canto 10 into Canto 11 where it is anchored in the full meeting with
Bimbisāra as an evocation of the Jarāsaṃdhavadha. Following Canto 12, in which,
as now noted, several have long seen parallels between Arādạ Kālāma’s proto-
Sāṃkhya and certain teachings of the Mokṣadharma Parvan, this Mahābhārata
reading would then be concluded in the encounter with Māra in Canto 13.
To understand how Aśvaghoṣa makes Canto 9 a hinge to these unfold-
ings, we must note two matters. First, such a Brahmanical deputation of
a Purohita and Minister to find the prince in the forest seems to be an inven-
tion by Aśvaghoṣa. In fact, it is new to his Buddhacarita. In his earlier
Saundarananda, the events from the prince’s great departure to his encounter
with Māra take only eight verses (3.2–9) without mentioning either the dep-
utation or the first meeting with King Bimbisāra. Second, we must look back
to a line near the end of Canto 8 where the Purohita and the Minister define
their mission to King Śuddhodana: “Just let there be a war of many kinds
between your son and the various prescriptions of scripture.”101 These two
speakers will come to this war armed with with Brahmanical scriptures,
which the prince will handle rather easily; but, more than this, it sets the
terms for the Bodhisattva’s inner struggle102 that carries through all these
cantos to his ultimate contest with Māra.
What I would like to emphasize, however, is that, important as it is that
Aśvaghoṣa knows something of the Mahābhārata’s Śāntiparvan as a whole,
including its Mokṣadharma Parvan, it is even more interesting that he knows
and probably takes from it the term mokṣadharma. Our chief question is, “Why
would teachings on mokṣadharma have interested him?” Let us begin with an
observation that although Buddhist and Brahmanical texts do sometimes have
differences they want to stress when one uses the term nirvāṇa and the other
the term mokṣa, neither has any internal ban on using the other’s term, and
both would know them as potentially alternates.103 Like the Buddha, Yudhiṣṭhira,
near the beginning of Book 12, wishes to take up the mendicant life and praises

101. BC 8.85cd: bahuvidham iha yuddham astu tāvat/ tava tanayasya vidheś ca tasya tasya. For yuddham,
Johnston has “struggle” (2004, 122) rather than “war.” Olivelle 2008, 241 translates “Let a battle be waged on
many fronts between your son and the diverse rules of scripture.”
102. Note that Fitzgerald speaks of “[t]he inner battle that . . . takes place within Yudhiṣṭhira” (2004a, 179),
occurring (better beginning) at Mbh 12.17, while Arjuna briefly refers to this process as still lying ahead: “Now
conquer yourself (vijitātmā . . . bhava)” (22.10cd). Yudhiṣṭhira’s “inner battle” continues through Books 12 and 13,
and indeed beyond.
103. See, for example, BhG 5.24–29 as cited in chapter 11 § E (conclusion).
660 dharma

the pursuit of mokṣa;104 but unlike the Buddha, he will not do so. Instead, he
agrees to listen to Bhīṣma, knowing that he will be dissuaded from pursuing it.
Thus a point of Olivelle’s is worth reconsidering here: that the Buddhacarita
speaks of mokṣa “in the technical meaning given to it by Manu, namely,
renunciatory asceticism of a wandering mendicant” (2008, xxi–xxii; see
chapter 5 § E). This may also be true of the Buddhacarita’s usages of mokṣadharma.
To explore this possibility, we now have Aśvaghoṣa responding to the treatment
of mokṣa in two texts, Manu and the Mahābhārata. For where he uses the term
mokṣadharma, he would be responding not to Manu, which never uses it, but to
the Mahābhārata. As we shall see, Aśvaghoṣa’s response to Manu parallels its
dharmaśāstric formulation on debt in just two verses (see above n. 17). More
complex, his response to the Mahābhārata takes in a full narrative parallel bet-
ween the Bodhisattva-prince and a king who is disposed, like him, toward
renunciatory asceticism and “mokṣa,” but who, unlike the Bodhisattva,
renounces them! Before looking at how Aśvaghoṣa uses the term mokṣadharma,
we must thus see how it is used in the Mahābhārata.

mokṣa in manu and mokṣadharma in the mahābhārata. Olivelle’s

F .2. A .

discussion of the parallel with Manu on debt at Buddhacarita 9.65–66 must thus
be our starting point. While making a “case . . . that Aśvaghoṣa knew Manu’s work
on dharma” (xix), Olivelle acknowledges Johnston’s recognition that Aśvaghoṣa
also “knew the ‘Rāmāyaṇa’ and presents the Buddha as the new Rāma” (2008,
xxii). But he is silent on Aśvaghoṣa’s relation to the Mahābhārata.
Olivelle gets to Buddhacarita 9.65–66 having begun a discussion of
Aśvaghoṣa’s treatment of what he calls “the theology of debt,” and introduces
the two verses by noting that “[t]hese words are put into the mouth of the coun-
selor of the Buddha’s father.” This counselor or mantrin is the “minister” who
has accompanied King Śuddhodana’s chaplain or Purohita to find prince
Siddhārtha in the forest. In Olivelle’s translation, the Minister says:

A man is released from his debts

to his ancestors through offspring,
to seers through studying the Vedas,
and to the gods through sacrifices;

104. The Śāntiparvan is actually rather slow in getting to the term mokṣa. Yudhiṣṭhira does not mention it
when he first envisions taking to the road as a renunciant (Mbh 12.9), though he worries about the effects of bad
karma brought on by his royal acts, and their possible effects on future existences (30–36). The first reference to
mokṣa occurs when Yudhiṣṭhira—in what Fitzgerald calls his “inner battle” (see n. 102)—recalls a verse sung by King
Janaka of Mithilā, “who was beyond the pairs of opposites, who had gained Absolute Freedom (mokṣaṃ samanupaśyatā),
and who had Absolute Freedom in full view (vimuktena)” (Mbh 12.17.17). As Fitzgerald notes (2004a, 685), “the
current context marks the first serious appearance of the theme of mokṣa in The Book of Peace.” As we shall see, King
Janaka is an ambiguous figure in the Śāntiparvan’s portrayal of mokṣa, and specifically of mokṣadharma.
 aśvaghoṣ a ’s buddhacarita 661

A man is born with these three debts,

whoever is released from these,
for him alone, they say, is release (yasyāsti mokṣaḥ kila
tasya mokṣaḥ).

Release is open to one, experts say,

who strives following this sequence of rules (ity evam etena

vidhikrameṇa/ mokṣaṃ sayatnasya vadanti taj jñāḥ);
Those who desire release violating that sequence (vikrameṇa
mumukṣavaḥ), only get fatigued though they expend
much effort. (Olivelle 2008, xx–xxi; 266–67; 457)

As we saw in chapter 5 § A, Olivelle considers Manu to have been the first to

use this theology of debt “to defend his position that the orders of life (āśramas)
are to be followed sequentially as an individual grows old and that renunciation
is limited to old age” (2008, xxi). Olivelle buttresses this point with the obser-
vation that “these two verses of Aśvaghoṣa parallel” two verses in Manu’s sixth
chapter on the āśramas, which read:

Only after he has paid his three debts, should a man set his mind on
release (mano mokṣe niveśayet); if he devotes himself to release
without paying them (anapākṛtya mokṣaṃ tu), he will proceed
downward. Only after he has studied the Vedas according to rule,
fathered sons in keeping with the Law, and offered sacrifices
according to his ability, should a man set his mind on release (mokṣe
niveśayet). (Manu 6.35–36; Olivelle 2005[a], 600; 2008, xxi)105

One can see that Aśvaghoṣa’s two verses are quite close conceptually to Manu’s
two verses.
Now it is in annotating these two verses in Manu that Olivelle makes the
observation I have been quoting about the “technical meaning” Manu gives to
mokṣa. In a parenthesis in that note which I mention only now, he urges that
readers should “see Olivelle 2005[a], 243.” Readers who follow up this recom-
mendation are referred there to his 1981 article titled “Contributions to the
Semantic History of Saṃnyāsa.” Fortunately, the 2005a distillation is almost
sufficient for our present concerns. It occurs in a note to Manu 1.114ab, a line in
Manu’s table of contents or “synopsis,” where, as cited in chapter 5 § E, Olivelle uses
asterisks to call attention to Manu’s treatment of mokṣa and saṃnyāsa as “Renunciation*
[6.33–85]” and “Retirement* [6.87–96]” respectively (2005a, 92, 401). As mentioned

105. As cited in chapter 5 § A. Cf. Olivelle 2005a, 150, translating the three usages of mokṣa in this passage
by “renunciation” instead of “release,” and with reference to his note to Manu 1.114, on which see the text.
662 dharma

in chapter 5, the asterisks direct us to the footnote in question, which I quoted in

part, leaving an ellipsis that is now pertinent. After stating that Manu attaches
its particular “technical meaning” to mokṣa making it “a synonym of renuncia-
tion and the fourth order of life dedicated exclusively to the search after personal
liberation,” Olivelle continues with what I elided: “The term has the same
meaning when used in the common compound mokṣadharma, which is a sec-
tion of the Mahābhārata and a distinct topic in medieval legal digests (nibandha)”
(2005a, 243). Further, after stating that Manu distinguishes clearly “between
this renunciatory asceticism and the life of a vedic retiree,” which Manu “desig-
nates as saṃnyāsa,” Olivelle goes on to say that other translators “ignore the
technical use of the two terms here,” and references his aforementioned 1981
article for “a more detailed study.” Two things were noted from that article in
chapter 5. First, Olivelle underscores that the differentiation of mokṣa as “renun-
ciation” from saṃnyāsa as “retirement” at Manu 1.114 involves for the latter the
abandonment of ritual activity incumbent on a householder (1981, 270–71; cit-
ing M 6.86–96). Second, Olivelle shows that, in contrast to Manu’s carving out
of its technical “vedic retiree” usage to insist on doing the four āśramas in
sequence, the Mahābhārata is one of just a few texts to introduce what Olivelle
calls “the classical meaning” of saṃnyāsa. That is, the Mahābhārata treats
saṃnyāsa, often with some disfavor, as what comes to be thought of as its ordi-
nary meaning, different from Manu’s, in which it is “commonly used as a syn-
onym of such terms as parivrājaka, pravrajita, śramaṇa, bhikṣu, and yati” (265).
Moreover, Olivelle shows that while the Bhagavad Gītā offers a “diatribe against
renunciation” (268), it also introduces the positive twist that what is renounced
with real saṃnyāsa is not just karma (ritual or otherwise) but the attachment
(saṅga) to karma and its fruits (karmaphala) (269–70, 272). This, of course,
bears on the distinction between Manu and the Gītā’s different understandings
of karmayoga (see chapter 11 § C).
We thus find Olivelle relating Manu’s technical usage of mokṣa directly to
the Mahābhārata’s usage of mokṣadharma in the Mokṣadharma Parvan. As
Olivelle’s comment seems to reflect, the term mokṣadharma is not found in
either the Rāmāyaṇa or Manu. Aśvaghoṣa would relate his usages of
mokṣadharma solely to the Mahābhārata, and particularly so in the section of
the Buddhacarita that Olivelle cites on the “theology of debt.” For as we shall
see in the next section, it is not long before King Śuddhodhana’s Minister tries
the argument that seeking mokṣa without fulfilling one’s three debts is likely to
be a failure (9.65–66), that his companion, the Purohita, is the first to speak of
mokṣadharma in terms that the Buddha-to-be will reject: that kings can win or
obtain mokṣadharma while remaining householders in the lap of luxury
(9.17–19). It is not hard to see how Aśvaghoṣa would represent two Brahmin
 aśvaghoṣ a ’s buddhacarita 663

counselors as arguing that mokṣa must be deferred by an indebted royal house-

holder. Yudhiṣṭhira’s brother Bhīma makes a similar link in a sarcastic response
to Yudhiṣṭhira’s desire for saṃnyāsa near the beginning of the Śāntiparvan:

Renunciation should be made at a time of great distress (āpat kāle hi

saṃnyāsaḥ kartavya), by one who is overcome by old age, or by one who
has been cheated by his enemies; so it is decreed. . . . If one resorts to
this baldness, this sham-Law (mauṇdỵ am . . . dharmacchadmasamsthāya),
and supports only himself, it is possible for him to subsist, but not to
live. But then again, a person can live, and live comfortably, by himself
in the forests, by not supporting any sons or grandsons, Gods, seers,
guests, or ancestors (abibhratā putrapautrāndevaṛsị̄ natithīnpitṝn). The
small animals of the wild do not win heaven in this fashion, nor do the
wild boars, nor do the birds; and people do not say it is a holy deed for
them. King, if anyone could attain perfection from mere renunciation
(saṃnyāsataḥ), then the mountains and trees would quickly obtain
perfection. (Mbh 12.10.17, 21–24; Fitzgerald trans. 2004a, 187–88)

Along with linking renunciation with old age, as King Bimbisāra will do at
Buddhacarita 10.33–37 (see the chart in § E above, item BC 10), Bhīma debunks
the renunciant ideal as a kind of wishful thinking if it implies being free of the
three debts to gods, seers and ancestors, which are included among the obliga-
tions he mentions. And this comes before Yudhiṣṭhira speaks about mokṣa (see
n. 103 above) in addition to his wish to undertake saṃnyāsa.
Let me then propose, tentatively at this point, since it is before looking at
the Buddhacarita’s own usages of mokṣadharma, that Aśvaghoṣa, in his “critical
reading” of both epics, first uses the term to talk about nirvāṇa, but in a way
that is meant to address Brahmanical usage of the term mokṣadharma in the
Mahābhārata as coming up short, from a Buddhist perspective, on the very
question at hand: the idea that mokṣa (i.e., nirvāṇa) would be formulated in rela-
tion to the ambiguities of the renunciatory asceticism of a wandering mendi-
cant, the ideals that engage Yudhiṣṭhira at the beginning of the Śāntiparvan
until he gives them up, once Bhīṣma has turned his attention from mokṣadharma
to dānadharma, by the Śāntiparvan’s end. In these circumstances, it would be
significant to know more about how the term mokṣadharma is in fact used in
the Mokṣadharma Parvan and elsewhere in the Mahābhārata.
Although there are a few usages of mokṣadharma outside the Mokṣadharma
Parvan, I think that Aśvaghoṣa would be referring mainly to the Mokṣadharma
Parvan, where the weight and dramatic centrality of the Mahābhārata’s teach-
ings on the topic certainly apply. This, as we have seen, is also the view of
Byodo [1930] 1969 and Tokunaga (2005). But I would propose additionally,
664 dharma

although it cannot be proven because Aśvaghoṣa makes no reference to any

specific Mokṣadharma Parvan units, that it would be rather unsuccessful to
argue that the Mokṣadharma Parvan’s last four units would not have been
included in the Mahābhārata that Aśvaghoṣa was critiquing, because they are
precisely the units where his argument most directly applies. These four are:
(a) “The Dialogue between King Janaka and the Woman Renouncer Sulabhā”
(12.308);106 (b) the story of Śuka (12.309–20);107 (c) the Nārāyaṇīya (12.321–39);
and (d) “The Substory about the Gleaner” (12.340–53). A few words are thus
called for on these four closing units.
First, the Sulabhā story is important mainly for setting up King Janaka to
Sulabhā’s exposé that his much vaunted reputation for living in a state of mokṣa
while enjoying all the trappings of royalty is not to be believed. Her verdict,
which the text endorses (Fitzgerald 2002, 651 n. 10), is,

You have fallen away from the householder pattern of life without
having reached mokṣa that is so hard to understand; you exist
between these two, babbling about mokṣa. (12.308.175; Fitzgerald
trans. 2002, 667, modified)

As noted in chapter 10 § C, when Sulabhā enters yogically into Janaka’s being to

test this claim, he accuses her of a series of violatory “mixings” of varṇa, āśrama,
gotra, and dharma (Mbh 12.308.59–62). Their dialogue includes three usages of
mokṣadharma, all implying rules of some kind. Both Janaka and Sulabhā speak
of the mokṣadharmas they follow, apparently as “rules” of their yogic disciplines.
First, Janaka speaks of the “the threefold mokṣadharma” (trividhe mokṣadharme;
25c) or “threefold mokṣa” (trividhaṃ mokṣam; 27c), explaining, as to the former:

On knowledge of Sāṃkhya, on yoga, and on the rule of protecting the

earth, on this threefold mokṣadharma I set out, my doubt severed.108

Fitzgerald translates mokṣadharma here as “Rule for Absolute Freedom,”109 but

it is hard to make much of this fractioned mokṣadharma/mokṣa unless its being
fractioned is one of its implied flaws. The second usage comes when Janaka

106. As Fitzgerald 2002, 641 titles it in translating Sulabhājanakasaṃvāda.

107. This shorthand applies to the Śukānuśāsanam (“Instruction of Śuka”; 12.309), Śukotpatti (“Origin of
Śuka; 310–15), and what is most generally referred to as the Śukābhipatanam (“The Flying about of Śuka”;
316–20).
108. Mbh 12.308.25: sāṃkhyajñāne tathā yoge mahīpālavidhau tathā/ trividhe mokṣadharme ‘smin gatādhvā
chinnasaṃśayaḥ. Note that “the rule for protecting the earth” could just be “the rule of a king.”
109. Fitzgerald 2002, 656 translates: “With my doubts completely dispelled, I set out on the threefold Rule for
Absolute Freedom: The Sāmḳ hya Knowledge of Discrimination, the Discipline of Yoga Meditation, and the Rule for
protecting the earth.”
 aśvaghoṣ a ’s buddhacarita 665

questions Sulabhā’s propriety in bringing about the various kinds of “mixtures”

(saṃkaras) he imagines her to be guilty of by entering him. Regarding our term,
he focuses on an implied mixing of the āśramas:

You live in the mokṣadharmas (vartase mokṣadharmeṣu) but I am in

the householder āśrama. This second mixing, that of the āśramas, is a
very grievous offense of yours. (12.308.60)

Here mokṣadharmas would presumably refer to the fourth life-pattern, and

again imply “rules” of discipline. Finally, toward the end of Sulabhā’s disproof
of Janaka’s charges, she says, “Trained in the mokṣadharmas, I fare alone
according to the vow of a Muni.”110 These rule-implying usages are important,
but, as we shall see, idiosyncratic to this unit. Narratively and developmentally,
the dialogue’s main purpose, coming near the end of the Śāntiparvan, may
include a final jolting of Yudhiṣṭhira as to the improbable outcome of a king’s
practicing “rules of mokṣa” as a royal householder. But it could also remind
him that Arjuna had debunked the same Janaka in less decisive but far more
vivid and caustic terms near the beginning of the Śāntiparvan in his outrage at
Yudhiṣṭhira’s first musings about mokṣa (see n. 104 above) after disclosing that
he wanted to renounce the kingdom and take off as a wandering almsman.
There, Arjuna quotes what Janaka’s wife said to him:

Having given up brilliant Royal Splendor, you look like a dog! Your
mother has no son now, and I, the princess of Kosala, have no
husband at all because of you. Eighty Kṣatriya women desiring
dharma attended you (dharmakāmās tvāṃ kṣatriyāḥ paryupāsate),
waiting for your commands—pitiable women motivated by their
desire for the fruits of their actions. Having deprived them of their
fruits, what heavenly worlds will you go to now, king, given that
mokṣa is quite uncertain for souls that others depend upon
(samśayite mokṣe paratantreṣu dehiṣu). Your deeds are wicked, and
you do not have either the higher world or the lower one since you
want to live after abandoning your lawful wives (dharmyān dārān
parityajya). (12.18.12–15; Fitzgerald trans. 2004a, 202, slightly
modified)

We can see how well rounded the Śāntiparvan is here, and how rich it would
have been for Aśvaghoṣa if he were to critique both its beginning and its end.
Indeed, Sulabhā also reminds Janaka of his “wife” and other “wives”:

110. Mbh 12.308.184cd: vinītā mokṣadharmeṣu carāmyekā munivratam.

666 dharma

The king who rules this whole earth under a single parasol dwells in
just one city and in this city he stays in just one palace, where he lies
upon his bed at night, and when he rests in that bed, half of it
belongs to his wife; thus is he joined to the consequences of actions
by these attachments. . . . He is never in complete control, even when
spending time in play with his wives. (12.308.135–36, 139ab;
Fitzgerald trans. 2002, 664)

Sulabhā, Arjuna, and Janaka’s wife agree that Janaka’s affectations to mokṣa are
quite improbable.
It is, however, in the Mokṣadharma Parvan’s final three units that the
term mokṣadharma is used in a way that marks off the full closure of the
Śāntiparvan around Yudhiṣṭhira’s turn away from this topic. Indeed, it would
be unpromising to argue that the Nārāyaṇīya would have been so much later
than the two units surrounding it that it could have been interpolated bet-
ween them, after Aśvaghoṣa, in Gupta times (as is usually assumed). These
two units frame the Nārāyaṇīya by using that term each only once—in the
Śuka story only in its very last verse, just before the Nārāyaṇīya, where
Yudhiṣṭhira hears,

Whoever, devoted to tranquility, would recall this meritorious history

that pertains to matters of mokṣadharma, he attains the supreme way.111

and, in “The Substory about the Gleaner,” only in its very first verse, where, just
after the Nārāyaṇīya, Yudhiṣṭhira asks,

Now, grandfather, that you have addressed the auspicious dharmas

that have to do with mokṣadharma, you can tell me, lord, about the
best dharma for those who pursue the āśramas.112

The Śuka story and “The Substory about the Gleaner” thus frame what the
Nārāyaṇīya has to say about mokṣadharma. Let us see how.
When Yudhiṣṭhira asks to know more about Śuka, he is asking his
grandfather Bhīṣma about the firstborn son of his other grandfather, indeed
his real grandfather genetically, Vyāsa. Śuka would be his father Pāṇḍu’s eldest
brother.113 The Śuka story is obviously a family matter, and comes at a point
where Yudhiṣṭhira is marking a turn toward adjusting to his familial and

111. Mbh 12.320.41: itihāsam imam puṇyam mokṣadharmārthasaṃhitam/ dhārayed yaḥ śamaparaḥ sa gacchet
paramāṃ gatim.
112. Mbh 12.340.1: dharmāḥ pitāmahenoktā mokṣadharmāśritāḥ śubhāh/̣ dharmam āśramināṃ śreṣṭhaṃ vaktum
arhati me bhavān.
113. For fuller discussion, see Hiltebeitel 2001a, 279–80. For Yudhiṣṭhira’s opening questions, see Mbh
12.310.1–5.
 aśvaghoṣ a ’s buddhacarita 667

dynastic responsibilities, which involve ruling the Kuru kingdom. Moreover, as

Olivelle mentions, along with the Mahābhārata’s bringing early prominence to
the “classical” āśrama system, the Śuka story is its “most straightforward pre-
sentation of the original [preclassical āśrama] system” (1993, 154). This is
because it confirms that Śuka can skip the full sequence of the four āśramas
and seek release directly as a continuation from the first, that is, from brahma-
carya, without marrying, and above all, without waiting for the fourth. The Śuka
story that Bhīṣma tells is about how Śuka obtained mokṣa, having received his
father Vyāsa’s and also King Janaka’s instruction, and also due to a somewhat
inherent disposition toward it. Most scholars, and perhaps Yudhiṣṭhira, take
Śuka’s mokṣa to be his exit from the world of saṃsāra.114 I say this might be
Yudhiṣṭhira’s impression, since the Pāṇḍavas are told in Book 3 to visit a tīrtha
named Vyāsasthalī where Vyāsa was consumed with grief over his son, pre-
sumably Śuka, and was resolved to give up the body until he was “made to get
up again by the gods.”115 But whether Yudhiṣṭhira knows it or not, we know that
Śuka has not left the world of saṃsāra, since three generations after Yudhiṣṭḥira,
he joins his father Vyāsa as an attendee at Janamejaya’s snake sacrifice to hear
the Mahābhārata told for the first time in the human world by Vaiśaṃpāyana.116
Indeed Śuka and Vyāsa are among the attendees who decide the fate of the
snakes. Yudhiṣṭhira might also pick up a hint of how Śuka might be living on
after obtaining mokṣa from what Bhīṣma tells him just before he begins with
the Śuka story proper:

Approach life’s journey by [eating] the remains of gods and guests

(devatātithiśeṣeṇa yātrām prāṇasya saṃśraya; 12.309.5cd).

The bird-like Śuka (his name means “Parrot”) is encouraged to do something

like gleaning (uñchavṛtti), which could tie in with the titular subject of the
Mokṣadharma’s third and last unit. In any case, Śuka’s manner would fit what
Olivelle calls the “technical meaning” that Manu gives to mokṣa, “namely,
renunciatory asceticism of a wandering mendicant, . . . rather than simply
liberation from the cycle of rebirth.” Moreover, as we have noted, the last verse
of his story mentions the term mokṣadharma to open up that subject for its
most sustained treatment in the Nārāyaṇīya.
As to the third and final unit of the Mokṣadharma Parvan, “The Subtale
about the Gleaner” tells of a Brahmin named Dharmāraṇya who hears that the

114. See, typically, Sörensen [1904] 1963, 216: “Ç. Obtained liberation, Vyāsa lamented his death.” Cf.
Hiltebeitel 2001a, 282–84, 317.
115. Mbh 3.81.81–82, which concludes: kṛto devaiś ca rājendra punar utthāpitas tadā; cf. Hiltebeitel 2001a,
43, 282.
116. Mbh 1.48.7ab; see Hiltebeitel 2001a, 115 and n. 71.
668 dharma

“highest wonder” a snake-king named Padmanābha saw while pulling the Sun’s
chariot was a refulgent being attaining liberation by entering the “solar disc” in
a moment. That could remind Yudhiṣṭhira of Śuka, and who knows, maybe it
was him. As Belvalkar puts it in describing this story’s opening adhyāya 340:

Yudhiṣṭhira says to Bhīṣma that, though he has listened to his

discourses on the Mokṣa-dharma, he still desires to hear from him
the highest Dharma which is to be practised by persons performing
the duties of the four āśramas. Thereupon Bhīṣma tells him there
are many ways of practising the highest Dharma. (Belvalkar
1954–66, ccxxxi)

Clearly, Bhīṣma is transitioning Yudhiṣṭhira away from mokṣadharma, about

which he has more or less heard enough, to the topic of the āśramas, which
implies his remaining in the householder stage as a royal householder—the
very thing that king Śuddhodana’s minister holds up for the Bodhisattva to con-
sider doing himself. And indeed, “The Subtale of the Gleaner” will tell about a
householder reaching the highest goal, albeit not as a king but a gleaner, and
without further mentioning mokṣa. A fuller discussion of this point would want
to take in further complexities of the Śuka and Gleaner stories,117 but for pre-
sent purposes it will suffice to say that Yudhiṣṭhira has turned his corner pre-
cisely by hearing the unit between them, the Nārāyaṇīya. For our discussion of
the Buddhacarita and also for our wider treatment of the Nārāyaṇīya in this
book, it is the Nārāyaṇīya that has the most importance.
The Nārāyaṇīya is too complex a text, and contains too many plots and sub-
plots, to be usefully summarized. Suffice it to say that can be useful to divide it
into a Part A, which tells how the Ṛṣi Nārada journeyed to and back from
Śvetadvīpa (“White Island”), where he got a special vision (darśana) of Nārāyaṇa;
and a Part B, which includes three dips to the Mahābhārata’s outer frame where
Śaunaka gets to ask questions to the bard Sauti or Ugraśravas. To appreciate the
significance of these dips and the questions that emerge through them, one
must recognize that S. K. Belvalkar (1954–66), basing his decision on obvious
changes that were introduced into what have become manuscripts in the
Malayālam script, erred in editing the Pune Critical Edition’s Śāntiparvan when
he reverted the conversation between Śaunaka and Ugraśravas to an inner
frame one between Janamajeya and Vaiśaṃ pāyana.118 Yudhiṣṭhira’s turning

117. Hiltebeitel forthcoming-b attempts this. There are plans to bring out a conference volume from the
Brown University conference at which it was presented.
118. See Hiltebeitel 2006a: I agree with Oberlies 1997 and Grünendahl 1997’s assessments of Belvalkar’s
error in editing, but not with their interpretations of Part B as later than Part A, or with Grünendahl’s view that
Belvalker’s mistake is ultimately insignificant in matters of interpretation.
 aśvaghoṣ a ’s buddhacarita 669

point occurs toward the end of Part A, and is clearly a moment of family
bonding. Having heard the White Island story, he and his brothers become
devoted to Nārāyaṇa, with Kṛṣṇa also listening in and standing by.119 The next
adhyāya (12.327), which begins Part B, is then the Nārāyaṇīya’s showcase for
the term mokṣadharma, being the only adhyāya in the Nārāyaṇīya to mention
the term, which it does three times. The term mokṣadharma does not occur
again until Yudhiṣṭḥira credits Bhīṣma with teaching him about it in the first
verse of “The Subtale about the Gleaner,” as noted. And thereafter, Bhīṣma
only mentions mokṣadharma one more time in a stray line120 more than halfway
through the Dānadharma Parvan. Yet it would be a mistake to think that the
Nārāyaṇīya leaves the concept behind after adhyāya 327, because it is intro-
duced there in conjunction with the somewhat overlapping term nivṛtti-dharma,
which can be said to thread the purport of mokṣadharma into further reaches of
the Nārāyaṇīya. Nivṛtti, either in the compound nivṛttidharma, or with that
meaning, has five usages along with the three of mokṣadharma in adhyāya 327.
More than this, in Part B, in the Nārāyaṇīya’s second dip to the outer frame
(see Hiltebeitel 2006a, 239–43), the verb ni-vṛt is used twice to describe
Nārada’s running “return” (12.331.16a, 20c) from seeing Nārāyaṇa on White
Island to see Nara and Nārāyaṇa at their Badarī āśrama (basically, to see what
the two Nārāyaṇas and Nara all have in common). This is one of the anomalies
that so intrigues Śaunaka that he asks his second leading question to Sauti
about it. Clearly, as we could also show with Śuka, it has to do with returning
(ni-vṛt) “here” to this world.121
Now, once we correct Belvalkar’s attempt to restore the inner frame dia-
logue between Janamejaya and Vaiśaṃ pāyana to the outer frame one between
Śaunaka and Ugraśravas, we find that adhyāya 12.327, at the beginning of
Part B, introduces the topic of mokṣadharma while making the Nārāyaṇīya’s
first dip to the outer frame. Basically, Śaunaka asks Sauti the first question that
has come to mind from hearing the White Island story, and Sauti answers by
telling him what Vaiśaṃ pāyana said when asked “the same” question by
Janamejaya, which was to tell him what Vyāsa once told his five disciples,

119. Mbh 12.326.121: “Having heard this best of Narratives, O Janamejaya, King Dharma and all his
brothers became devoted to Nārāyaṇa.” See chapter 12 §§ C and E on this and similar scenes at 3.187.50–53 and
13.126.4–6.
120. It occurs in a unit called Śrāddha-Kalpa, “Procedures for Ancestral Rites” (13.87–92), in an adhyāya
where Bhīṣma distinguishes Brahmins who are unsuitable to hire for śrāddhas from those who are suitable, men-
tioning among the latter “Yatis conversant with mokṣadharma” (yatayo mokṣadharmajñā; 3.90.25c).
121. See the repeated uses of iha, “here,” in this second dip to describe Nārada’s arrival at Badarī (331.21d;
38d; 51e). On the Śuka story, cf. Hiltebeitel 2001a, 286–94, especially with reference to 12.314.33–36, where
Vyāsa’s disciples, including Śuka, ask his favor that the Vedas should “abide here,” probably including the
Mahābhārata as “this (ayam) Veda.”
670 dharma

including Vaiśaṃ pāyana and Śuka (see Hiltebeitel 2006a, 233–39). For pre-
sent purposes, it must suffice to give the contextual flavor of the three usages
of mokṣadharma.
Śaunaka opens things up in Part B by asking about Nārāyaṇa: how, while
he is “established in nivṛtti dharma, enjoying peace, ever the beloved of
Bhagavatas,” do the other gods come to accept shares according to pravṛtti
dharmas, while nivṛtti dharmas are “made for those who have turned aside”
(327.2–3). The first use of mokṣadharma now addresses this tension. Sauti
recalls the purportedly similar question that Janamejaya asked Vaiśaṃ pāyana,
from which I cull only the verses with which he begins:

[Janamejaya said,] These worlds with Brahmā, men, gods and demons
are seen everywhere to be attached to rites said to assure prosperity.
And mokṣa is said by you, O Brahmin, to be nirvāṇa, the supreme
happiness. And those who are released are beyond merit and sin; we
hear they enter the god of a thousand rays. Alas, the eternal
mokṣadharma is surely difficult to observe (aho hi duranuṣṭeyo
mokṣadharmaḥ sanātanaḥ), abandoning which all the gods have
become enjoyers of rites to gods and ancestors (havya-kavya).
(12.327.5–7)

Imagine Aśvaghoṣa, if he read this, raising his eyebrows at the comparison bet-
ween mokṣa and nirvāṇa—with both as something that all the gods “abandon”
despite its being “eternal”!122 So far one might suspect that in being placed
beside mokṣa as compared with nirvāṇa, “the eternal mokṣadharma” would have
to do more here with liberation from saṃsāra than with renunciatory asceti-
cism. But with it being the gods who are in question, we can already see that
they do not need either. The next usage comes where Vaiśaṃ pāyana is quoting
what Vyāsa told him and his other four disciples, including Śuka, about what
Brahmā and the gods and Ṛṣis once learned when they went to ask Nārāyaṇa
about such matters on the northern shore of the Milky Ocean, where they
found him. There, Nārāyaṇa remarked that, while he has consigned the gods to
receive offerings until the end of the kalpa according to pravṛtti dharma for the
welfare of the world, and has assigned seven mindborn Ṛṣis—Marīci, Aṅgiras,
Atri, Pulastya, Pulaha, Kratu, and Vasiṣṭha—to procreation following pravṛtti-
dharma (326.60–62), he has also assigned seven other Ṛṣis—Sana, Sanatsujāta,
Sanaka, Sanandana, Sanatkumāra, Kapila, and Sanātana, “called mental sons
of Brahmā” (64–65)—to do the following:

122. Cf. Mbh 12.326.63ab: “The highest nivṛtti is known as the extinction all dharmas” (nirvāṇaṃ sarva
dharmāṇāṃ nivṛttiḥ paramā smṛtā).
 aśvaghoṣ a ’s buddhacarita 671

With knowledge that comes of itself, they are established in nivṛtti

dharma. They are the foremost of yoga-knowers, as also knowers of
the Sāṃkhya-dharma. They are preceptors in mokṣaśāstra and
promulgators of mokṣadharma (mokṣadharmapravartakāḥ).123

Clearly, we know this group, some of them from the Mahābhārata itself, as
perennial Ṛṣis of the type whose mokṣa entails their returning occasionally to
this world to tell us about it. Indeed, we have met an unaging Brahmā
Sanatkumāra in the Ambaṭṭha and Aggañña Suttas (see chapter 4 § A). Finally,
the third usage comes when Vyāsa tells what happened when all the other
heaven-dwellers but Brahmā had gone. When Brahmā remained in place,
“desiring to see the blessed lord who takes on the body of Aniruddha, the god,
having assumed the great Horse’s Head (Hayaśiras), appeared to him, reciting
the Vedas with their limbs. . . .” (327.80–81). The Horse’s Head now reinforces
the distinctions between nivṛtti and pravṛtti with special attention to Brahmā’s
charge to oversee pravṛtti as the “world’s creator” (lokakartā), and promises,
before vanishing, that he (the Horse’s Head is of course Nārāyaṇa) will inter-
vene with various manifestations (prādurbhāvas) to bear the work of the gods
(surakāryam) whenever things get intolerable (82–86b). Vyāsa then continues:

So it is that this one of great share, the eternal lotus-naveled one124 . . .,

the eternal upholder of sacrifices, has fixed nivṛtti dharma, which is
the destination of those whose teaching is the imperishable. He has
(also) ordained pravṛtti dharmas, having made for the world’s diversity.
He is the beginning, middle, and end of creatures; he is the ordainer
and the ordained, he is the maker and the made. At the end of the
yuga he sleeps after having retracted the worlds; at the beginning of
the yuga he awakens and creates the universe. (12.327.87–89)

This is all buildup to the Nārāyaṇīya’s final usage of mokṣadharma, for Vyāsa
now starts a laud of Nārāyaṇa (327.90–96) that includes this verse:

O you who always dwells on the ocean, O Hari, you whose hair is like
muñja grass,
O you who are the peace of all beings, you who imparts mokṣa-
dharma (mokṣadharmānubhāṣine). . . . 125

123. Mbh 12.327.65c–66: svayamāgatavijñānā nivṛttaṃ dharmam āsthitāḥ// ete yogavido mukhyāḥ
sāṃkhyadharmavidas tathā/ ācāryā mokṣaśāstre ca mokṣadharmapravartakāḥ.
124. He refers to Nārāyaṇa as “the eternal Padmanābha” (padmanābhaḥ sanātanaḥ)—a name we meet in
the next unit, “The Subtale about the Gleaner,” as the name of the snake who drives the sun’s chariot.
125. Mbh 12.327.93: samudravāsine nityaṃ haraye munjakeśine/ śāntaye sarvabhūtānāṃ
mokṣadharmānubhāṣine.
672 dharma

Vyāsa then concludes his laud with a guarantee to his disciples that all this is
true, and exhorts them to sing Hari’s praise with Vedic words (327.97–98),
whereupon Vaiśaṃ pāyana winds up this quotation from his guru by telling
Janamejaya that “all of Veda-Vyāsa’s disciples and his son Śuka, the foremost
knower of dharma,” did as he said (327.99).
In his second, central, and transitional moment in presenting mokṣadharma,
Bhīṣma thus flavors it as something well beyond Yudhiṣṭhira’s current intent:
it is a “knowledge that comes of itself” to those who “are established in nivṛtti
dharma. They are the foremost of yoga-knowers, as also knowers of the
Sāṃ khya-dharma. They are preceptors in mokṣaśāstra and promulgators of
mokṣadharma.” Indeed, considering what we learned about how Yudhiṣṭhira
and other supernaturally incarnated characters in the Mahābhārata will come
to “the end of their karma” in a way that makes mokṣa moot (see chapter 12
§ A), we might even say that it is beyond his capacity. In any case, Bhīṣma’s
overall order of business has been to secure Yudhiṣṭhira’s commitment to the
dharma of kings, that is, to rājadharma, the first of the four topics that comprise
his curriculum for the war-torn king. When Bhīṣma gets to mokṣadharma,
it has become an interesting theoretical matter that Yudhiṣṭhira has more or
less agreed to put behind him. For Aśvaghoṣa, it is this tension with rājadharma
that would make Yudhiṣṭhira’s shuttling of mokṣadharma interesting, and we
will find evidence for this in ways that he juxtaposes the two terms.
Curiously, the Mahābhārata’s use, and it seems, possible invention, of the
term mokṣadharma has an anomalous, even uncomfortable feel to it. Wasn’t it
one of the high points of the Bhagavad Gītā that Kṛṣṇa told Arjuna he should
abandon all dharmas since Kṛṣṇa will “release” him from every sin (18.66)?
How can there be “rules” or “laws on salvation or release,” or even one rule or
law on it, if one is to be released from all laws? In a thought-provoking article
on the tensions between sādhāraṇadharma and varṇāśramadharma as worldly,
and mokṣadharma, Gerald Larson describes the latter as the dharma that “does
not fit” (1972, 149). Adam Bowles notes that nivṛttidharma overlaps in the
Mahābhārata with mokṣadharma, and remarks that the latter looks at first blush
“like an oxymoron.”126 I do not think, however, that it was meant not to fit or to
be as oxymoronic as it first looks. But translating the dharma in it is certainly
less straightforward than it is in the titles for the other three of Bhīṣma’s anthol-
ogies. This is how James Fitzgerald broached it in his 1980 dissertation:

So the majority of texts collected in the MDh focus directly on

mokṣadharma-s, that is, behavioral or attitudinal norms (dharma-s)

126. See Bowles 2007, 153, citing Biardeau 1981a, 94; Olivelle 1992, 53: “a term that may have been viewed
as a contradiction by early Brāhmaṇical theologians, becomes a respected branch of the science of dharma.”
 aśvaghoṣ a ’s buddhacarita 673

leading to mokṣa, ultimate personal transcendence of the limits, pain,

and misery common to the situation of all living beings. From the
doctrinal, or thematic, perspective, the collection is best understood
in terms of a general distinction between 1) texts which address
directly some mokṣa theme and 2) texts which address mokṣa related
themes more indirectly, by way of working through problems posed
in terms of traditional dharmic categories. The texts of this latter type
confront the practical dharmic implications as well as the theoretical
arguments of mokṣa oriented themes. (Fitzgerald 1980, 231)

I like this statement for its attention to the tension between both dharma and
mokṣa in the term mokṣadharma, and for its this-textly and this-worldly orienta-
tion, which clearly applies to the Nārāyaṇīya and the two units that surround it.
Equally clearly, that is not the perspective of Aśvaghoṣa.

f.2.b. mokṣadharma in the buddhacarita. I present Aśvaghoṣa’s reading

of a full Śāntiparvan as a working hypothesis worth considering, well aware
that it will be met skeptically by some who continue to hold the Nārāyaṇīya,
in particular, to be late, and aware too that it requires untestable assumptions
about Aśvaghoṣa’s reading habits. Coming now to what Aśvaghoṣa actually
wrote, there are three usages of mokṣadharma in the surviving Sanskrit
portions of the first half of the Buddhacarita, all in the most Mahābhārata-
related segment described above. Before turning to them, however, two things
are worth noting.
First, reinforcing the possibility that Aśvaghoṣa could be recalling the
Nārāyaṇīya is the fact that, like that text, he uses mokṣadharma in ways that
overlap with nivṛtti-dharma in contrast with pravṛtti. For instance, within our
Mahābhārata sector, the prince winds up his exchange with King Bimbisāra:

So, whether one is young, old, or even a child,

one should quickly act in such a way here
That, endowed with dharma, with a perfected self,
one will win the continued life one seeks (pravṛttir iṣṭā),
or the total cessation of such life (vinivṛttir eva vā).
(BC 11.63; Olivelle trans. 2008, 320)

And shortly before that segment, he tells the anchorites, “the dharma of cessa-
tion from activity (nivṛttidharma) is apart from the continuance of active being
(pravṛttyā)” (7.48). The fact that Aśvaghoṣa uses both mokṣa and nivṛtti (along
with pravṛtti) in his earlier Saundarananda, but not in compounds with “dharma,”
could suggest that he did his close reading of the Mokṣadharma Parvan between
674 dharma

writing these two kāvyas. For his earlier work, he coins the decisive compound
mokṣamārga (1.1; cf. 17.13), which does not occur in either epic.
Second, assuming that Johnston was consistent in choosing the phrase “law
of salvation” in his attempt to reconstruct the Sanskrit from its Tibetan and
Chinese translations, it is worth noting the tenor of two likely additional usages of
the term mokṣadharma itself after the surviving Sanskrit portions of the
Buddhacarita. First, resting after his enlightenment and preparing to preach, the
Buddha saw that “the law of salvation was exceeding subtle” (BC 14.96). And
second, just after turning the “Wheel of the Law” with his first sermon and
converting his first five disciples, “the Omniscient established the Law of Salvation”
with further preaching and more conversions (16.1). “Law of salvation” (that is,
probably mokṣadharma) would seem to reach its full impact as one of Aśvaghoṣa’s
terms for the true dharma itself as the “Law” and “Teaching” of the newly enlight-
ened Buddha. If Aśvaghoṣa is critiquing the usages of mokṣadharma in the
Nārāyaṇīya, that would be what the gods and Yudhiṣthira abandon.
Turning now to the three verifiable usages, the first two occur in the exchange
between the prince and the Purohita. When the Purohita and the Minister arrive,
they find the prince sitting below a tree (BC 9.8). In being the first to convey the
message of the prince’s father, the Purohita seems to mix the king’s sentiments with
some new words of his own. Now acknowledging that the prince’s “fixed resolve
with regard to dharma” will be realized as his “future goal,” but invoking once again
the father’s massive grief that the prince is doing this “at the wrong time” (9.14–16),
the Purohita continues to report the words of the prince’s father as follows:

17. Therefore enjoy lordship for the present over the earth and you shall
go to the forest at the time approved by the scriptures (śāstradṛṣṭe).
Have regard for me, your unlucky father, for dharma consists in
compassion for all creatures.
18. Nor is it only in the forest that this dharma is achieved; its achievement
is certain for the self-controlled in a city too. Purpose and effort are
the means in this matter; for the forest and the badges of mendicancy
are the mark of the faint-hearted.
19. The dharma of salvation (mokṣadharma) has been obtained by kings
even though they remained at home, wearing the royal tiara, with
strings of pearls hanging over their shoulders and their arms fortified
by rings, as they lay cradled in the lap of imperial Fortune (lakṣmī).127

127. BC 9.17. tad bhuṅkṣva tāvad vasudhādhipatyaṃ/ kāle vanaṃ yāsyasi śāstradṛṣṭe//aniṣṭabandhau kuru
mayy apekṣam/ sarveṣu bhūteṣu dayā hi dharmaḥ//
18. na caiṣa dharmo vana eva siddhaḥ/ pure ‘pi siddhir niyatā yatīnām// buddhiś ca yatnaś ca nimittam atra/
vanaṃ ca liṅgaṃ ca hi bhīrucihnam//
19. maulīdharair aṃsaviṣaktahāraiḥ/ keyūraviṣṭabdhabhujair narendraiḥ// lakṣmyaṅkamadhye parivartamānaiḥ/
prāpto gṛhasthair mokṣadharmaḥ.
 aśvaghoṣ a ’s buddhacarita 675

Here is Olivelle’s translation of the last of these three verses, which catches
how Aśvaghoṣa makes mokṣadharma the verse’s very last word:

Kings, even while remaining householders

cradled in the lap of royal fortune
crowns upon their heads,
pearl strings on shoulders,
arms bound with bracelets,
have won the dharma of release.
(BC 9.19; Olivelle 2009, 251)

In being the first to mention mokṣadharma, the Purohita prompts the

Bodhisattva’s doubt that release can be won in the lap of royal luxury, where-
upon the Bodhisattva states his firm resolve not to seek it there himself whether
it is possible are not. The Purohita goes on, purportedly in the father’s words,
to mention Bali and Janaka of Videha among several otherwise obscure house-
holder kings128 who “were well trained in dharma rules leading to highest bliss”
(naiḥśreyase dharmavidhau vinītān)129 even while they remained gṛhasthas
(20–21). In invoking King Janaka here, Aśvaghoṣa could be recalling Arjuna’s
quotation from Janaka’s wife and/or the Sulabhā story, which mentions the
term mokṣadharma in Sulabhā’s exposé of the hypocrisy of Janaka’s claims to
being liberated. As further evidence of Aśvaghoṣa’s Mahābhārata frame of ref-
erence here, it is at this point that the Purohita recalls Bhīṣma, though not in
connection with his postwar oration.130 In any case, when the prince replies
“after a moment’s meditation (dhyātvā muhūrtam)” (BC 9.30), he says that fear
of the three signs has left him no choice but to leave, even knowing the fatherly
affections involved (31); in a world of wayfarers, why cherish grief ?131 However

128. Johnston cannot trace some of these (2004, 126–27, n. 20).

129. Olivelle’s more informative translation (2008, 251) suggests something in which Buddhist readers
could hear an ironic reference to monastic training. Johnston (2004, 127) has, “versed in the method of practising
the dharma that leads to final beatitude.”
130. He recalls “the deeds done by Bhīṣma, who sprang from the womb of Gaṅgā, Rāma, and Rāma
Jāmadagnya, to please their fathers” (25), and concludes by again recalling the grief caused to others whom the
prince has left behind (26–29). Nothing in the Mahābhārata suggests that Bhīṣma obtains mokṣa. When he sends
his concentrated breaths through the crown of his head up to heaven, Kṛṣṇa tells the weeping Gaṅgā, who has
emerged from her waters, that he has rejoined the Vasus (Mbh 13.154.2–7, 18, 28, 32). See however chapter 12 §
A; Austin 2009, 601, 614–15: Bhīṣṃa would be among those whom Nīlakaṇtha thinks would have obtained
mokṣa (but then so would Yudhiṣṭhira).
131. BC 9.35. These responses may recall the Śuka story near the end of the Mokṣadharma Parvan, in which
Janaka of Videha is cast, even in his own palace, as an expert on renunciation, and in which Vyāsa confronts his
fatherly affections for his ultimately affectless son Śuka as the latter makes his mokṣa-departure. See Hiltebeitel
2001, 278–322. On Janaka in other such contexts, see Olivelle 1993, 238–40.
676 dharma

noble it is that his father wishes to hand over the kingdom to him, he rejects
kingship as an “abode of delusion in which are to be found fearfulness, the
intoxication of pride, weariness and the loss [or oppression, or “squeezing”] of
dharma by the mishandling of others (parāpacāreṇa ca dharmapīḍā).”132 It may
be “praiseworthy for kings to leave their kingdoms and enter the forest in the
desire for dharma (dharmābhilāṣeṇa), but it is not fitting to break one’s vow and
forsaking the forest to go to one’s home”; for a man of resolution who has gone
to the forest out of desire for dharma, return to the city would be like eating
one’s own vomit, like reentering a burning house (44–47). And now, coming
to the Buddhacarita’s second usage of mokṣadharma, he says with precise and
loaded words on our central point:

48. As for the revelation (śruti!) that kings obtained final emancipation
(mokṣa) while remaining as householders (nṛpā gṛhasthā),133 this
is not the case. How can the dharma of salvation (mokṣadharma) in
which quietude (śama) predominates be reconciled with the dharma of
kings (rājadharma) in which severity of action (daṇdạ ) predominates?134

Going on to argue that “quietude and severity are incompatible (śamaś ca

taikṣṇyaṃ ca hi nopapannam)” for a king (49), he even subjects the Purohita’s
affirmative proposition to a dialectical critique:

50. Either therefore those lords of the earth resolutely cast aside their
kingdoms and obtained quietude, or, stained by kingship, they
claimed to have attained liberation on the ground that their senses
were under control, but in fact only reached a state that was not final.
51. Or let it be conceded they attained quietude while holding kingship,
still I have not gone to the forest with an undecided mind; for having
cut through the net known as home and kindred I am freed and have
no intention of re-entering that net.135

132. BC 9.39–40. For “squeezing,” see Bowles 2004, 199 n. 33, on dharmam prapīḍya at Mbh 12.101.2.
Johnston 2004, 131 n. 40, also notes a usage of dharmapīḍa at what is now Critical Edition 13.96.10, which is a
verse in which Agastya tells that he has heard, “Time harms (kills, saps) the energy of dharma (kālo hiṃsate
dharmavīryam),” coming in a series of stories about when it is dharma not to accept gifts (13.94–96).
133. Johnston 2004, translates, “As for the tradition that kings obtained final emancipation while remain-
ing in their homes. . .”—which I change for the obvious points of emphasis. Cf. Olivelle 2008, 261: “As for the
scripture that householder kings have attained release.”
134. BC 9.48. yā ca śrutir mokṣam avāptavanto/ nṛpā gṛhasthā iti naitad asti//śamapradhānaḥ kva ca
mokṣadharmo/ daṇḍapradhānaḥ kva ca rājadharmaḥ.
Cf. Olivelle 2008, 261 for the second line: “The dharma of release, where calm prevails, and the dharma of
kings, where force prevails—how far apart are these!”
135. BC 9.50. tan niścayād vā vasudhādhipās te/ rājyāni muktvā śamam āptavantaḥ//rājyāṅgitā vā
nibhṛtendriyatvād/ anaiṣṭike mokṣakṛtābhimānāḥ//
51. teṣāṃ ca rājye ‘stu śamo yathāvat/ prāpto vanaṃ nāham aniścayena//chittvā hi pāśaṃ gṛhabandhusaṃjñam/
muktaḥ punar na pravivikṣur asmi.
 aśvaghoṣ a ’s buddhacarita 677

What a crystal-clear Buddhist critique of the ambiguities of the Brahmanical

position! And, I think implicitly, what a subtle response to the nearly interminable
indecisiveness and ultimate resignation to rājadharma and gṛhasthadharma,
while putting aside mokṣadharma, of Yudhiṣṭhira Dharmarāja!
The third usage comes early from Māra, fingering an arrow (BC 13.8) as he
first verbally challenges the Bodhisattva’s right to sit beneath the bodhi tree:

9. Up, up, Sir Kṣatriya, afraid of death. Follow your own dharma (cara
svadharmam), give up the dharma of liberation (tyaja mokṣadharmam).
Subdue the world with both arrows and sacrifices, and from the
world obtain the world of Vāsava.136

This is the first and only usage of svadharma in the first fourteen cantos of the
Buddhacarita, and, as far as I can see, the only one likely in the entire text.
Note that whereas in the first usage of mokṣadharma the Purohita says it is
possible to combine mokṣadharma with gṛhasthadharma, and in the second the
prince contrasts mokṣadharma with rājadharma,137 Māra now contrasts it with
svadharma.
Indeed, as we now see, up until the likely usages of mokṣadharma to refer to
the true dharma after the Buddha’s enlightenment, Aśvaghoṣa uses it contras-
tively with a definite Mahābhārata cachet. In each case his usages might be
intended to prickle Brahmanical ears with references to the postwar predica-
ment of Yudhiṣṭhira, who of course wants to do something like what the Buddha
does and is persuaded not to. But this third usage more pointedly echoes the
prewar dilemma of Arjuna, which can at least be said to anticipate Yudhiṣṭhira’s
postwar one.138 From the first word uttiṣṭha, the imperative “Up, up” or “Arise,”

That householders can obtain liberating knowledge could be seen as the Mīmāṃsā position; see Olivelle
1993, 238–40.
136. BC 13.9: uttiṣṭha bhoḥ kṣatriya mṛtyubhīta/ cara svadharmaṃ tyaja mokṣadharmam//bāṇaiś ca yajñaiś ca
vinīya lokam/ lokāt padam prāpnuhi vāsavasya.
Schreiner 1990 brings out that there is a variant varasva dharmam, “choose dharma,” for cara svadharmam.
Weaker and noncontrastive, I think we can treat it as secondary. Cf. Olivelle 208, 375:”Rise up, O Warrior, afraid
of death! Follow the dharma that’s your own.”
137. As to such a contrast, a further likely usage of rājadharma occurs when the Buddha goes to Kosala to
meet King Prasenajit, and hears from him, “O Lord, I have suffered and been harassed by passion (rāga) and the
kingly profession (rājadharma)” (20.10), to which the Buddha replies at length (12–51) as to how kings can benefit
from the Buddha’s teaching or law (14–17), earlier called his (mokṣa-)dharma.
138. See chapter 10 § E. Olivelle 1993, 103–6, 150 also sees their dilemmas in parallel and brings out that, in
contrast to Arjuna who never hears about āśramas in the Gītā, Yudhiṣṭhira wants to hear about them at length.
Olivelle’s treatment of the Bhagavad Gītā’s emphasis on svadharma and varṇa (caste) rather than āśrama is full of
implications for understanding these two brothers’ differences (105–6, 197), but it is not “likely that the author [of the
BhG] would not have known the classical [āśrama] system” (Olivelle 1993, 105) such as it was known to the author of
the beginning of the Śāntiparvan.
678 dharma

Aśvaghoṣa puts Māra’s insulting challenge in the simplest language of the

Bhagavad Gītā,139 recalling especially Bhagavad Gītā 2.31–37 where this command
at 2.37 is preceded by urgings that Arjuna do his Kṣatriya svadharma in some of
̣ ạ ’s most insulting prods, goading him, just as Māra does the Bodhisattva,
Kṛsn
to stop looking like he is abstaining from battle through fear (2.35), and prom-
ising him a warrior’s heaven of no ultimate value (2.32, 37). The upshot for
Aśvaghoṣa is that Māra’s challenge to fight and perform Kṣatriya svadharma
rather than pursue mokṣadharma not only recalls the Gītā but puts Kṛsn ̣ ạ ’s
words into the mouth of the devil. Here Aśvaghoṣa avails himself of a “Hindu
notion” that, according to Gombrich, finds no place in the Pāli canon.140
Yet the Śāntiparvan remains Aśvaghoṣa’s deeper frame of reference. As
noted in chapter 11 § B, Yudhiṣṭhira has more trouble than Arjuna with this
svadharma concept, and early in his postwar predicament he rejects Arjuna’s
own advocacy of it.141 Moreover, as we have noted in this chapter, throughout
the Śāntiparvan he is guided through a tension that Arjuna scorns from the
start: that between rājadharma and mokṣadharma. As we have seen, these
terms furnish the titles of the first and third subparvans of the Śāntiparvan
where they provide the scansion and arc142 of Yudhiṣṭhira’s postwar instruction
by Bhīṣma that runs through what Fitzgerald (2004a) calls the four anthol-
ogies of the Śānti and Anuśāsana Parvans. This arc of teaching, levelled at
Yudhiṣṭhira but overheard by all the Pāṇḍavas and Draupadī as well
(13.57.42–44), goes through four dharma topics: rājadharma, āpaddharma,
mokṣadharma, and (in the Anuśāsanaparvan) dānadharma. It is this arc or

139. Kṛṣṇa tells Arjuna “Arise!” four times: BhG 2.3; 2.37, 4.42, and finally more or less decisively at 11.33.
Māra uses the verb three times in his short speech (13.9–13), twice in the imperative.
140. Gombrich 1985, 436 says of this “Hindu notion,” “Buddhists do not even have the term svadharma
(Pali *sadhamma). . . .” As we have seen, Aśvaghoṣa can be a bit arch at times when he symbolically juxtaposes
Kṛṣṇa and the Buddha. Unlike Bimbisāra, who also—if only in the Buddhacarita—challenges the Bodhisattva to
take up arms, Māra must be overcome, and, with him, so too must such (from the Buddhist perspective) conve-
nient and self-serving ideas as the svadharma of princes.
141. Just after Yudhiṣṭhira says he is renouncing the kingdom and going to the forest, Arjuna invokes
Yudhiṣṭhira’s svadharma amid insulting and mocking words (Mbh 12.8.3–5) reminiscent of Kṛṣṇa’s in the Gītā.
Vyāsa summarizes some BhG theology as well (12.26.14–16; 32.11–15; 34.4–7)—all to no avail. Yudhiṣṭhira finds
these arguments inadequate, eventually requiring Vyāsa to come up with a ritual solution (the Aśvamedha
sacrifice of Book 14), which Vyāsa already anticipates in this early Śāntiparvan sequence (12.32.20–24). See
Hiltebeitel 2005d, 251–58.
142. Cf. Bowles 2007, 297 on the sequence of the three Śāntiparvan anthologies: “A logic of action informs
this structure, a logic that models the proper duties of the royal life. A king’s desire for salvation must follow the
proper completion of his royal duty, or, rather, it follows from the proper completion of his royal duty. The syntactic
order of the ŚP text . . . mirrors, therefore, the proper syntactic order of the royal life and the proper order of the
king’s concerns” (2004, 297; 2007, 391). After quoting this passage as it first appeared in Bowles 2004, I wrote: “I
believe Bowles has found the right terms here for us to deepen our investigation of the fourth anthology: Would not
dānadharma follow mokṣadharma in ‘the proper syntactic order of the royal life’?” (2005d, 261). See also Hiltebeitel
2005a, 486–91 on this transition from Book 12 to Book 13.
 aśvaghoṣ a ’s buddhacarita 679

sequence through which Yudhiṣṭhira must not only learn about kingship and
its distresses but renounce his inclination to pursue mokṣa, and finally, in the
dānadharma, abandon his wish to retreat to an ashram (Ibid.) in order to
become a giving king. As far as I am able to discern, this fourfold sequence is
unique in Indian dharma literature to the Mahābhārata, and may, I believe, be
called one of its signature formulations about dharma.143 It presents an out-
come that the Buddha must, at least for himself, reject, but not one that he
would necessarily reject for all. Indeed, Aśvaghoṣa has found it worth engaging,
for I believe that his juxtaposition of rājadharma and mokṣadharma, along with
his demonstrations of textual familiarity with both the Rājadharma- and
Mokṣadharma-Parvans of the Śāntiparvan, show that he has the first and third
units of this arc firmly in view. But what about āpad and dāna?
With āpad the evidence is not very strong, but still worth considering.
Āpad comes up only once in the first fourteen cantos, and not in the segment
where Aśvaghoṣa undertakes what I have called a Mahābhārata reading. When
the prince addresses the horse Kanthaka in preparation for his great departure,
he says:

5.76. Easy it is to find companions for battle, for the pleasure of

acquiring the objects of sense and for the accumulation of wealth;
but hard it is for a man to find companions when he has fallen into
distress (āpadi) or attaches himself to dharma.144

It is emphasized in this speech that the prince speaks to Kanthaka as a

companion (sahāya) and friend (suhṛd; 5.79), but while foreshadowing that the
prince will have to make his battle alone, without friends. Thus the interesting
juxtaposition: companions are hard to find “for a man who has fallen into āpad
or attaches himself to dharma.” Since this is the only usage of āpad in the
Buddhacarita, it is hard to say whether the prince uses the term to define his

143. It is, however, worth noting an intriguing parallel, though not a likely influence one way or another,
in the addition of a Bodhisattvapiṭaka as a fourth canonical “basket” (piṭaka) by the Dharmaguptakas (see Pagel
1995, 7–36; Nattier 2003, 46 n. 80; 80–83, 129, 274–76). With four “baskets” (which denote collections of man-
uscripts) we have an analogy with Fitzgerald’s notion of four “anthologies.” And, putting aside the obvious reser-
vation that one collection is for monks and the other for an epic king, there would also be some minimal
correspondence in the last two pairings between the two sets of four in sequence: (a) dharmapiṭaka: rājadharma;
(b) vinayapiṭaka: āpaddharma; (c) abhidharmapiṭaka: mokṣadharma; and (d) bodhisattvapiṭaka: dānadharma—with
the Bodhisattva basket stressing the practice and teaching of the six pāramītas that begin with dāna (Nattier 2003,
154 n. 38; 186). Curiously, the Bahuśrutīyas, with whom Johnston attempts to link Aśvaghoṣa (2004, xxx–xxxv),
also had a bodhisattvapiṭaka, but in a canon of five baskets (Nattier 2003, 46 n. 80). I believe that Nattier’s study
of Ugraparipṛcchā could open new considerations on the sectarian and intertextual placement of Aśvaghoṣa (see
n. 16 above).
144. BC 5.76: sulabhāḥ khalu saṃyuge sahāyā/ viṣayāv āptasukhe dhanārjane vā puruṣasya tu durlabhāḥ
sahāyāḥ/ patitasyāpadi dharmasaṃśraye vā
680 dharma

present situation or is speaking disjunctively and implying that, rather than

being in distress, he is only attaching himself to dharma. One is perhaps helped
by a verse in which Udāyin says he speaks out of friendship offered in adver-
sity, using āpad’s near-synonym vyasana (BC 4.64) (see Bowles 2007, 51–68),
when he counsels the prince to gratify the women who are trying to seduce him
between the third and fourth signs. This suggests that the prince’s situation is
adversity (āpad, vyasana) as others see it, but as he is beginning to see it himself
when he speaks to Kanthaka, it is not adversity once he has begun resorting to
dharma. Also interesting, and within the Buddhacarita’s Mahābhārata sector, is
a verse using vyasana, where the prince responds to the Purohita:

9.41. For kingship is at the same time full of delights and the vehicle
of calamity (vyasanāśrayam), like a golden palace all on fire, like
dainty food mixed with poison, or like a lotus pond infested with
crocodiles.145

These, however, are no more than reminders of a general theme. In any case,
the disjunctive use of āpad and dharma makes it clear that there is no question
of a compound āpaddharma. The best we get is a negative explanation as to why
āpad would not be used in the first half of the Buddhacarita in the sense of
āpaddharma. Unless perhaps one thinks of Māra, there are no princes or kings
in distress over the possibility of losing their kingdoms in the text’s first four-
teen cantos.146
As to dāna, quite surprisingly, there is no use of the term in the first half
of the Buddhacarita. But giving is made an important matter in Canto 18 where,
not surprisingly, the Buddha is addressing not a king but one of those wealthy
merchants, gahapatis or gṛhapatis, so important to both Theravāda and early
Mahāyāna texts147 for the economic support of early Indian Buddhism.
A wealthy merchant of Kosala named Sudatta, “who was in the habit of giving
wealth to the destitute,” came “from the north” at night (BC 18.1–2) to see the
Buddha in Rājagṛha. Having welcomed him, the Buddha turns quickly to “the
fame in this world and the reward in the hereafter [that] arise from giving,” and
urges that “at the proper time” Sudatta should “give the treasure that is won
through the Law” (5). After hearing an initial sermon mainly on imperma-
nence, Sudatta “obtained the first fruit of practice of the Law; and . . . only one

145. BC 9.41: jāmbūnadaṃ harmyam iva pradīptam/ viṣeṇa saṃyuktam ivottamānnam/grāhakulaṃ cāmbv iva
saravindam/ rājyaṃ hi ramyaṃ vyasanāśrayaṃ ca.
146. It is a different matter in the second half with King Prasenajit of Kosala (see n. 137 above), and of
course with Bimbisāra too, who will be murdered by his son Ajātaśatru.
147. See Bailey and Mabbett 2003, 43–53 and passim; Nattier [2003] 2005, 23–31 and passim.
 aśvaghoṣ a ’s buddhacarita 681

drop remained over from the great ocean of suffering for him. Though living
in the house, he realized by insight the highest good” (15–16). As with
Yudhiṣṭhira, whom the Buddhacarita never, of course, criticizes, somebody has
to do this job of giving, and must be educated to do it in the right spirit. After a
lengthy interval in which the poet describes Sudatta’s insight in terms of the
Brahmanical views he now gives up, including those about a deity (18–29), we
return to Sudatta as he is offering to donate a monastery at Śrāvastī (57). Here
the Buddha praises giving at length (61–80), mentioning that it is “one of the
elements of salvation” (74); expounding on the varied virtues of giving wealth,
food, clothes, abodes, vehicles, and lamps (76–78); and concluding that
Sudatta’s gift is of the best kind since it “has no ulterior motive” (79). Sudatta’s
gift will be land: the Jeta grove for the Jetuvana vihāra (81–85). The verses on
the varied merits of giving different things could be called a capsule Dānadharma,
since they are reminiscent of the middle third of the Dānadharma Parvan in
which Bhīṣma regales Yudhiṣṭhira on the merits of giving all the same things,
though above all, giving food and land to Brahmins. As with āpaddharma, we
must again pose a negative explanation, this time as to why dāna is not used in
the first half of the Buddhacarita, but in this case expounded upon in the sec-
ond. It is not a matter of import until the Buddha must develop a postenlight-
enment theory of the gift148—albeit without any evidence that it would have
been called dānadharma.149
Nonetheless, as I have shown elsewhere, giving food is among the topics
brought up toward the end of Yudhiṣṭhira’s discussions with Vyāsa in the
first forty adhyāyas of the Śāntiparvan, where the topics of all four dharma
anthologies are in fact anticipated.150 So if Aśvaghoṣa is familiar with that seg-
ment, he would be familiar at least with these topics, if not with the plan and
contents of the four subparvans themselves that describe the full arc of
Yudhiṣṭhira’s postwar education on dharma. Actually, however, all four terms
are also developed earlier in the Mahābhārata. Not counting the epic’s
Parvasaṃgraha or table of contents, where dānadharma is the only one not
mentioned, there are, prior to the Śāntiparvan, fourteen usages of rājadharma,
nine of āpaddharma, four of mokṣadharma, and six of dānadharma, each with
both singular and plural (-dharmas) instances. Insofar as Yudhiṣṭhira is

148. See the succinct and elegant essay on this subject, said to be “highly theorized in Indian Buddhist
textual discourse,” by Reiko Ohnuma (2005, quoting from p. 102).
149. Note that Aśoka uses the reverse compound dharmadāna (Nikam and McKeon 1978, xii, 44–45 [Rock
Edict 11]; Olivelle 2004a, 509 n. 28). Typically, his mainly Buddhist usages stresses the applicability of dharma as
a universal term to dāna, whereas the Mbh usage makes dāna part of a structured curriculum of separate but
integrated “laws.”
150. See Hiltebeitel 2005d, 259, citing Mbh 12.37.1–2 and 43 on food and giving.
682 dharma

addressed about each of these dharmas, they pace him toward their grand
unfolding to him in Books 12 and 13. Further, once past Book 13, there is also
follow-up in Book 14 on both mokṣadharma (Mbh 14.2.17; 16.16; 19.63 and
49) and dānadharma (14.2.19; 4.7; 94.34)—the two that would still be ringing
most insistently in Yudhiṣṭhira’s (and readers’) ears. Early uses of rājadharma
are basic and not surprising. It is the only one of the four mentioned in the
Rāmāyaṇa (seven times), and, as treated in the dharmasūtras and Manu, even
Bhīma can remind Yudhiṣṭhira that Manu spoke on rājadharma (Mbh
3.36.20). Yudhiṣṭhira also hears about āpaddharma as something basic and
au courant from Vidura after the return dice match, with a warning to pro-
ceed carefully as the Pāṇḍavas prepare for exile (2.69.19). But mokṣadharma
and dānadharma are novel enough matters to be the subjects of upākhyānas
or subtales told to Yudhiṣṭhira and company in the Forest Book. From
Mārkaṇḍeya, Yudhiṣṭhira learns that he has just heard “the entire mokṣadharma
(kṛtsne mokṣadharme)” (3.204.1) in a speech by a meatseller known as the
pious hunter (dharmavyādha)—even though he has given up hunting and
just markets meat that others supply—to a Brahmin in the Pativratā-
Upākhyāna (3.196–206; van Buitenen 1975, 617–38). And dānadharma is a
topic Yudhiṣṭhira wants to know about enough to ask the author himself,
Vyāsa, which weighs more in the afterworld, dānadharma or tapas (3.245.26).
Vyāsa favors dānadharma so long as one gives rightfully obtained wealth
(245.32), which leads him to recount the Mudgala-Upākhyāna about the Ṛṣi
Mudgala who gave unstintingly to guests what little he had garnered from
living righteously off what he gleaned from harvested fields.151 When an
envoy of the gods tries to interest Mudgala in ascending with him to heaven,
he tells Mudgala he will find there “the Law-minded, the masters of self, the
serene and controlled and unenvious, those accustomed to the Law of giving
(dānadharmaratāḥ), and champions with the scars showing.”152 But Mudgala
rejects heaven in favor of “the eternal and supreme perfection that is marked
by Extinction (śāśvatīṃ siddhiṃ parāṃ nirvāṇa-lakṣaṇām)” (247.43; van
Buitenen 705). As this upākhyāna shows, the unfolding of dānadharma
involves weighting it favorably over tapas: a matter that is returned to repeat-
edly in the Dānadharma Parvan,153 and one that deserves further study

151. He is one of the uñchavṛtti/”gleaner” Brahmins and exemplars of Ṛsị dharma alluded to in chapter 12 § E.
152. Mbh 3.247.4; van Buitenen 1975, 703. It is interesting to see van Buitenen translate dānadharma this
way for the first time, having seemingly struggled with it before this: translating it as a dvandva (1.94.11 and 17),
omitting its translation (3.155.10), and trying out “the merits of gifts” and just “giving” earlier in the Mudgala-
Upākhyāna.
153. From Mbh 13.57 on, see 13.93–94, 106, and 109–10. In some passages contrasting the two, tapas is
associated with sacrifice and fasting, as it is in the description of the anchorites in BC Canto 7.
 aśvaghoṣ a ’s buddhacarita 683

(see Olivelle 1993, 162–70). Indeed, a preference for dānadharma over tapas
would probably win Aśvaghoṣa’s and the Buddha’s agreement, as would
Mudgala’s spurning of heaven for nirvāṇa.
Within the skein of Books 12 and 13, however, it is clear that what counts
most for Aśvaghoṣa is mokṣadharma, which he seems to have introduced into
Buddhist literature as a way to translate nirvāṇa that would clarify in both
Buddhist and Brahmanical circles what is comparable and what is distinctive
about Buddhist and Brahmanical dharmas. I remain under the impression that
neither mokṣadharma nor a would-be Pāli equivalent has appeared in Buddhist
texts before Aśvaghoṣa.

G. Postscript on Aśoka

Aśvaghoṣa gives three verses to Aśoka toward the very end of the Buddhacarita,
telling how he put the Buddha’s dharma on the map:

63. In the course of time king Aśoka was born, who was devoted to the
faith; he caused grief to proud enemies and removed the grief of
people in suffering, being pleasant to look on as an aśoka tree, laden
with blossoms and fruit.
64. The noble glory of the Maurya race, he set to work for the good of his
subjects to provide the whole earth with stūpas, and so he who has
been called Caṇḍāśoka became Aśoka Dharmarāja.
65. The Maurya took the relics of the Seer from the seven stūpas in which
they had been deposited, and distributed them in due course in a
single day over eighty thousand majestic stūpas, which shone with the
brilliancy of autumn clouds. (BC 28. 63–65)

Now one touchstone in marking a slightly less than civil recognition of

Buddhism in the Mahābhārata was noted in chapter 12 § C: Mārkaṇḍeya’s
prophesy about a Kali yuga overrun with eḍūkas (Mbh 3.188.64–67, 70), eḍūka
being the oldest term for Buddhist reliquaries, to begin with those for the
bones of the Buddha after his cremation, and found in both Sanskrit and Pāli
as a term for stūpas. A Buddhist counterprophesy can be found in the
Mañjuśrīmūlakalpa, where the Buddha tells King Ajataśātru of Magadha (who
had by now killed his father Bimbisāra):

After my decease, the masters of the world will kill each other from
father to son; the bhikṣus will be engrossed in business affairs and the
people, victims of greed. The laity will lose their faith, will kill and
684 dharma

spy on one another. The land will be invaded by Devas and Tīrthikas,
and the population will place its faith in the Brāhmins; men will take
pleasure in killing living beings and will lead a loose life.154

Devas and Tīrthakas would seem to be Brahmanical temples and other holy
places served by Brahmins. Note that the Mahābhārata passage also makes a
rare predictive reference to Brahmanical temples (devasthānas; 3.188.65c), along
with Brahmin settlements and hermitages of the great Ṛṣis, as being sup-
planted by the Buddhist eḍūkas. Clearly the contrast is instructive: eḍūka may
cover a variety of funereal or other mounds, but the ones Mārkaṇḍeya men-
tions belong to a non-Brahmanical proliferation of the “future.”
As John Strong shows, in the Aśokāvadāna, it is Aśoka’s proliferation of
stūpas, which that text actually calls dharmarājikās, that marks the transition
from his being called Caṇḍāśoka, “Aśoka the Fierce,” to Dharmāśoka,
“Aśoka the Righteous”—a term for which Aśvaghoṣa lets “Aśoka
Dharmarāja” stand alone, and that gives Aśoka the name Dharmarāja,
which is also an epithet for the Buddha, in part because the building of
stūpas represents “the reconstruction of the Buddha’s body” (1983, 117–18).
For Aśvaghoṣa, as for his Aśoka, dharma seems to be intended more as a
universal value than a civilizational one—something the Buddhist dharma
claims to make possible for everyone on levelled terms. If indeed the
Sanskrit epics and Manu present civilizationally attuned ripostes to a never
quite mentioned Buddhist dharma, Aśvaghoṣa would seem to provide both
the Buddha and Aśoka as, in his mind, still potent challenges to the under-
standings of dharma in those texts, and particularly to the Mahābhārata’s
never quite mentioned main story.

154. See Lamotte 1998, 94 [103], translating Mañjuśrīmūlakalpa verses 236–48.

Bibliography

Abé, Ryūichi. 2005. Word. In Donald S. Lopez, Jr., ed. Critical Terms for the
Study of Buddhism, 291–310. Chicago: University of Chicago Press.
Adarkar, Aditya. 2001. Karṇa in the Mahābhārata. Ph.D. Dissertation,
University of Chicago.
——— . 2005. The Untested Dharma is Not Worth Living. IJHS 9/1–3: 117–30.
Adluri, Vishwa, ed. 2010a. Ways and Reasons for Thinking about the
Mahābhārata as a Whole. Pune: Bhandarkar Oriental Research Institute.
——— . 2010b. Hermeneutics and Narrative Architecture in the
Mahābhārata. In Adluri, ed. 2010a, 31–57.
——— . 2010c. The Perils of Textual Transmission: Decapitation and
Recapitulation. Seminar, no. 608, special issue titled The Enduring Epic:
A Symposium on Some Concerns Raised in the Mahābhārata. 608:
48–54. http://www.india.seminar.com/2010/608_vishwa_adluri.htm.
——— . 2010d. Review of Malinar 2007a. HR 50/1: 102–07.
——— and Joydeep Bagchee. 2011. Introduction to Hiltebeitel 2011a,
——— . Forthcoming. The Nay Science: German Orientalism and the Reception
of the Bhagavad Gītā. Under press consideration.
Aklujkar, Ashok. 2004. Can the Grammarians Dharma be a Dharma For All?
JIP 32/5–6: 687–732.
Aklujkar, Vidyut. 1991. Savitrī: Old and New. In Arvind Sharma, ed. Essays in
the Mahābhārata, 324–33. Leiden: Brill.
Allchin, F. R. and K. R. Norman. 1985. Guide to the Asokan Inscriptions. South
Asian Studies I, 43–50.
Allen, Nicholas. 1996. The Hero’s Five Relationships: A Proto-Indo-
European Story. In Julia Leslie, ed. Myth and Mythmaking: Continuous
Evolution in Indian Tradition, 1–20. Richmond: Curzon.
686 bibliography

——— . 2005. Romulus et Bhîshma: Structures Entrecroisées. Anthropologie et Sociétés

29/2: 21–44.
——— . 2007a. Bhīṣma as Matchmaker. In Brodbeck and Black, eds. 2007, 176–88.
——— . 2007b. The Pāṇḍavas’ Five Journeys and the Structure of the Mahābhārata.
Religions of South Asia 1/2: 165–81.
Alles, Gregory D. 1994. The Iliad, the Rāmāyaṇa, and the Work of Religion: Failed
Persuasion and Religious Mystification. University Park, PA: Pennsylvania State
University Press.
Anderson, Paul Kent. 1990. Studies in the Minor Rock Edicts of Aśoka. Critical Editon.
Freiburg: Hedwig Falk.
Apte, V. S. 1957–1959. Sanskrit-English Dictionary, 3 vols. Poona: Prasad Prakashan.
Aronson, Harvey. 1984. Buddhist and Non-Buddhist Approaches to the Sublime
Attitudes (Brahma-vihāra). In Gatare Dhammapadā (?), R. Gombrich, and K. R.
Norman, eds. Buddhist Studies in Honour of Hammalava Sadhātissa, 16–24.
Nugegoda, Sri Lanka: University of Sri Jayawardanapura.
Austin, Christopher R. 2009. Janamejaya’s Last Question. JIP 37/6: 597–625.
Bagchee, Joydeep. In press. Review of Framarin 2009.
Bailey, Greg. 1983a. The Mythology of Brahmā. Delhi: Oxford University Press.
——— . 1983b. Suffering in the Mahābhārata: Draupadī and Yudhiṣṭhira. Puruṣārtha
7: 109–29.
——— . 1985. Materials for the Study of Ancient Indian Ideologies: Pravṛṛtti and Nivṛtti.
Publicazioni di “Indiologica Taurinensia.” Collana di Lettura diretta da Oscar
Botto. Turin.
——— . 1988. The Semantics of Bhakti in the Vāmana Purāṇa. Rivista degli Studi
Orientali 62: 25–57.
——— . 2003. The Mahābhārata as Counterpoint to the Pāli Canon. Paper presented
at the Conference on Religions in the Indic Civilization, Center for Developing
Societies, New Delhi, December 18–21. Courtesy of the author.
——— . 2008. On the Significance of the Mahābhārata as a Cultural Artefact in the
Early History of India (400 BCE–400 CE). Indologica Taurinensia 34: 13–37.
——— . N.d.-a. Contrasting Ideologies in the Nārāyaṇīyaparvan: Chapter 327 and the
Definitions of Pravṛtti and Nivṛtti. Courtesy of the author.
——— . N.d.-b. Preliminary Notes on Pāli Vatt and Sanskrit Vṛt in the Mahābhārata.
La Trobe University Pravṛtti/Nivṛtti Project. Courtesy of the author.
——— and Ian Mabbett. 2003. The Sociology of Early Buddhism. Cambridge:
Cambridge University Press.
Bareau, André. 1974. La jeunesse du Buddha dans les sūtrapiṭaka et les vinayapiṭaka
anciens. BEFEO 61: 194–274.
——— . 1993. Le Buddha et les rois. BEFEO 80/1: 15–39.
Beal, Samuel. 1968. The Fo-Sho-Hing-Tsan-King: A Life of the Buddha by Aśvaghosha
Bodhisattva Translated from Sanskrit into Chinese by Dharmaraksha, A.D. 420.
First published 1883. Sacred Books of the East, vol. 19. Delhi: Motilal
Banarsidass.
Beall, E. F. 2005/2006. Hesiod’s Treatise on Justice: “Works and Days” 109–380. The
Classical Journal 101/2: 161–82.
 bibliography 687

Bechert, Heinz. 1979. The Beginnings of Buddhist Historiography in Ceylon. Ancient

Ceylon 3: 21–28.
——— , ed. 1991–92. The Dating of the Historical Buddha/ Die Datierung des histo-
rischen Buddha. Abhandlungen d. Akademie d. Wissenschaften in Göttingen. 2
parts. Göttingen: Vandenhoeck and Ruprecht.
——— . 1982. The Date of the Buddha Reconsidered. IT 10: 29–36.
Beck, Brenda E. F. 1982. The Three Twins: The Telling of a South Indian Folk Epic.
Bloomington: Indiana University Press.
Bedekar, V. M. 1969. Principles of Mahābhārata Textual Criticsm: The Need for
Restatement. Purāṇa 11: 210–28.
Belvalkar, Shripad Krishna, ed. 1954–66. Śāntiparvan, with Appendix I and Critical
Notes. In V. S. Sukthankar et al., eds. Mahābhārata: Critical Edition, 24 vols. with
Harivaṃśa. Poona: Bhandarkar Oriental Research Institute.
Beyer, Stephen. 1974. The Buddhist Experience: Sources and Interpretations. Belmont,
CA: Wadsworth.
Bhandare, M. S. 1920. The Bālakāṇḍa of the Vālmīki Rāmāyaṇa. Bombay: Standard
Publishing Co.
Bhattacharya, Pradeep. 2008. Bridging a Gap in Puranic History. “The Sampradaya
Sun. Independent Vaisnava News. Feature Stories.” http://www.harekrsna.com/
sun/features/08-08/features1119.htm
——— . 2009. Was Draupadī Ever Sought To Be Disrobed? In Chakravarty ed., 2009,
89–99.
Bhattacharya, S. 1983–84. Political Authority and Brāhmaṇa-Kṣatriya Relationship in
Early India— An Aspect of the Power Elite Configuration. Indian Historical
Review 10/1–2: 1–20.
Bhattacharyya, Sivaprasad. 1956. The Approximate Date of the Harivaṃśa. Journal of
the Asiatic Society: Letters 22, 2: 155–61.
Biardeau, Madeleine. 1968. Some More Considerations about Textual Criticism.
Purāṇa 10/2: 115–23.
——— . 1970a. Letter to the Editor. Purāṇa 12/1: 180–1.
——— . 1970b. The Story of Arjuna Kārtavīrya without Reconstruction. Purāṇa 12/2:
286–303.
——— . 1976. Le Sacrifice dans L’Hindouisme. In Biardeau and Charles Malamoud,
eds. Le sacrifice dans l’Inde ancienne, 7–154. Paris: Presses Universitaires de
France.
——— . 1976. Études de mythologie hindoue: 4. Bhakti et avatāra. BEFEO 63: 87–237.
——— . 1978. Études de mythologie hindoue: 5. Bhakti et avatāra. BEFEO 65: 111–263.
——— . 1979. Śakuntalā dans l’épopée. IT 7: 115–25.
——— . 1981a. The Salvation of the King in the Mahābhārata. CIS n.s. 15/1 & 2: 75–97.
——— . 1981b. Études de mythologie hindoue, 1: Cosmogonies purānị ques. Publication de
l’École Française d’Extrême Orient, vol. 128: Paris: École Française d’Extrême Orient.
——— . 1984. Nala et Damayantī. Héros épiques. Part 1. IIJ 27: 247–74.
——— . 1985. Nala et Damayantī. Héros épiques. Part 2. IIJ 28: 1–34.
——— . 1989a. Hinduism: The Anthropology of a Civilization. Trans. Richard Nice.
Oxford: Oxford University Press.
688 bibliography

——— . 1989b. Histoires de poteaux: Variations védiques autour de la Déesse hindoue.

Paris: École française d’Extrême-Orient.
——— . 1991. Vedic Cosmogony. David White, trans. In Yves Bonnefoy, compiler.
Mythologies. Translation Directed by Wendy Doniger, 2 vols. Vol. 2: 806–7.
Chicago: University of Chicago Press.
——— . 1994. Études de mythologie hindoue, 2: Bhakti et avatāra. Publication de l’École
Française d’Extrême Orient, vol. 171. Pondicherry: École Française d’Extrême
Orient.
——— . 1997a. Some Remarks on the Links between the Epics, the Purāṇas and their
Vedic Sources. In Gerhard Oberhammer, ed. Studies in Hinduism: Vedism and
Hinduism, 77–119. Vienna: Verlag der Österreichischen Akademie der
Wissenscahaften.
——— . 1997b. Un certain Kīcaka. In Siegfried Lienhard and Irma Piovano, eds. Lex et
Litterae: Studies in Honour of Professor Oscar Botto. Alessandria: Editioni dell’Orso.
——— . 1999. Le Rāmāyaṇa de Vālmīki. Paris: Gallimard.
——— . 2002. Le Mahābhārata: Un récit fondateur du brahmanisme et son interprétation,
2 vols. Paris: Seuil.
——— . 2004. Stories about Posts: Vedic Variations around the Hindu Goddess. Alf
Hiltebeitel, Marie-Louise Reiniche, and James Walker, trans. Chicago: University
of Chicago Press.
——— and Jean-Michel Péterfalvi, trans. 1985, 1986. Le Mahābhārata. Vol. 1, Books
1–5; Vol. 2, Books 6–18. Paris: Flammarion.
Bigger, Andreas. 1998. Balarāma im Mahābhārata. Beiträge zur Indologie 30.
Wiesbaden: Otto Harassowitz.
——— . 2002. The Normative Redaction of the Mahābhārata: Possibilities and
Limitations of a Working Hypothesis. In Mary Brockington ed., 2002, 17–31.
Bizot, François. 1976. Le figuier a cinq branches. Paris: École Française d’Extrême
Orient.
Black, Brian. 2006. Paper on Ambaṭṭha and Śvetaketu. AAR Annual Meeting,
Washington, DC.
——— . 2007a. The Character of the Self in Ancient India: Priests, Kings, and Women in
the Early Upaniṣads. Albany: SUNY Press.
——— . 2007b. Eavesdropping on the Epic: Female Listeners in the Mahābhārata. In
Brodbeck and Black, eds. 2007, 53–78.
——— . 2011. Ambaṭṭha and Śvetaketu: Literary Connections between the Upaniṣads
and Early Buddhist Narratives. JAAR 79, 1: 136–61.
Bloch, Jules. 1950. Les inscriptions d’Aśoka. Paris: Société d’Édition « Les Belles
Lettres.»
Bloomfield, Maurice. 1916. Rig-Veda Repetitions: The Repeated Verses and Distichs
and Stanzas of the Rig-Veda in Systematic Presentation and with Critical
Discussion. Harvard Oriental Series, 20 and 24. Cambridge: Cambridge
University Press.
Boccali, Giuliano. 2004. Introduction. Concluding roundtable discussion on “Origins
of Mahākāvya: Problems and Perspectives”. Origins of Mahākāvya: International
Seminar, June 4–5. Milan: Università degli Studi di Milano.
 bibliography 689

Bodhi, Bhikkhu, trans. 2000. The Connected Discourses of the Buddha: A Translation of
the Saṃyutta Nikāya. The Teachings of the Buddha. Boston: Wisdom Publications.
——— , ed. 2005. In the Buddha’s Words: An Anthology of Discourses from the Pāli
Canon. Teachings of the Buddha. Somerville, MA: Wisdom Books.
Bose, Mandakranta, ed. 2004. The Rāmāyaṇa Revisited. New York: Oxford.
Bose, Buddhadeva. 1986. The Book of Yudhisthir: A Study of the Mahabharat of Vyas.
Sujit Mukherji trans., from the Bengali Mahabharater Katha (1974). Hyderabad:
Sangam Books.
Bowles, Adam. 2004. Dharma, Disorder and the Political in Ancient India: The
Āpaddharmaparvan of the Mahābhārata. Ph.D. dissertation, La Trobe University.
——— . 2007. Dharma, Disorder and the Political in Ancient India: The
Ā paddharmaparvan of the Mahābhārata. Leiden: Brill.
——— . trans. 2008. Mahabharata Book 8. Karna, 2 vols. Clay Sanskrit Library.
New York: New York University Press and JJC Foundation.
——— . 2009. Framing Bhīṣma’s Royal Instructions: The Mahābhārata and the
Problem of its ‘Design’. In Koskikallio, ed. 2009, 121–35.
Bracey, Robert. 2005. The Azes Era: A New Discovery and a New Problem. http://
www.kushan.org/essays/chronology/azesvikrama.htm
Brereton, Joel P. 1981. The Ṛgvedic Ā dityas. American Oriental Series, vol. 63.
New Haven: American Oriental Society.
——— . 1986. ‘Tat Tvam Asi’ in Context. ZDMG 136: 98–109.
——— . 1997. “Why is a Sleeping Dog Like the Vedic Sacrifice”: The Structure of an
Upaniṣadic Brahmodya. In Witzel ed., 1997c, 1–14.
——— . 1999. Edifying Puzzlement: Ṛgveda 10.129 and the Uses of Enigma. JAOS
119/2: 248–60.
——— . 2002. The Race of Mudgala and Mudgalānī. JAOS 122: 224–34.
——— . 2004. Dhárman in the Ṛgveda. JIP 32/5–6: 449–89.
——— . 2006. The Composition of the Maitreyī Dialogue in the Bṛhadāraṇyaka
Upaniṣad. JAOS 126/3: 323–46.
Brick, D. 2006. Transforming Tradition into Texts: The Early Development of Smṛti.
IIJ 34: 287–302.
Brockington, John. 1998. The Sanskrit Epics. In J. Bronkhorst, ed. Handbuch der
Orientalistik, Zweite Abteilung, Indien, vol. 12. Leiden: Brill.
——— . 2000a. Epic Threads: John Brockington on the Sanskrit Epics. In Greg Bailey and
Mary Brockington, eds. New Delhi: Oxford University Press.
——— . 2000b. The Structure of the Mokṣa-dharma-parvan of the Mahābhārata. In
Piotr Balcerowicz and Marek Mejor, eds. On the Understanding of Other Cultures.
Proceedings of the International Conference on Sanskrit and Related Studies to
Commemorate the Centenary of the Birth of Stanislaw Schayer (1899–1941),
pp. [incl. 72]. Warsaw: Oriental Institute, Warsaw University.
——— . 2002. Jarāsaṃdha of Magadha (MBh 2,15–22). In M. Brockington ed. 2002:
73–88.
——— . 2005. The Epics in the Bhakti Tradition. In Anna S. King and John
Brockington, eds. The Intimate Other: Love Divine in Indic Religions, 31–53.
New Delhi: Orient Longman.
690 bibliography

——— . 2006. Epic Svayaṃvaras. In Raghunath Panda and Madhusudan Mishra, eds.
Voice of the Orient (A Tribute to Upendranath Dahl), 35–42. Delhi: Eastern Book
Linkers.
——— and Mary Brockington. 2006. Rāma the Steadfast: An Early Form of the
Rāmāyaṇa. London: Penguin.
Brockington, Mary. 2001. Husband or King? Yudhiṣṭhira’s Dilemma in the
Mahābhārata. IIJ 4/(2001): 253–63.
——— , ed. 2002. Stages and Transitions: Temporal and Historical Frameworks in Epic
and Purāṇic Literature. Proceedings of the Second Dubrovnik International
Conference on the Sanskrit Epics and Purāṇas, August 1999. Zagreb: Croatian
Academy of Sciences and Arts.
——— . Forthcoming. Husband or Slave? Interpreting the Hero of the Mahābhārata.
Brockington, Mary and Peter Schreiner, eds. 1999. Composing a Tradition: Concepts,
Techniques and Relationships. Proceedings of the First Dubrovnik International
Conference on the Sanskrit Epics and Purāṇas, August 1997. Zagreb: Croatian
Academy of Sciences and Arts.
Brodbeck, Simon Pearse. 2003. Kṛṣṇa’s Action as the Paradigm of asakta karman in
the Bhagavadgītā. In Renata Czekalska and Halina Marlewicz, eds. Second
International Conference on Indian Studies: Proceedings. (Cracow Indological
Studies, 4–5), 85–112. Cracow: Ksiegarnia Akademicka.
——— . 2004. Calling Kṛṣṇa’s Bluff: Non-attached Action in the Bhagavadgītā. JIP 32:
81–103.
——— . 2006. Ekalavya and Mahābhārata 1.121–28. IJHS 10: 1–34.
——— . 2007a. Gendered Soteriology: Marriage and the Karmayoga. In Brodbeck and
Black, eds. 2007, 144–75.
——— . 2007b. Review of M. Brockington ed. 2002 and P. Koskikallio ed. 2005. JRAS
(third series) 17/1: 81–6.
——— . 2009a. The Mahābhārata Patriline: Gender, Culture, and the Royal Hereditary.
Farnham: Ashgate.
——— . 2009b. The Bhāradvāja Pattern in the Mahābhārata. In Koskikallio, ed.
2009, 137–79.
——— . 2009c. Husbands of Earth: Kṣatriyas, Females, and Female Kṣatriyas in the
Strīparvan of the Mahābhārata. In Goldman and Tokunaga, eds. 2009, 33–63.
——— . 2010a. Introduction: Hermeneutics and Criticism. In Adluri, ed. 2010a, 7–30.
——— . 2010b. Review of Framarin 2009. In Religious Studies, Vol. 46, 135–40.
——— . 2010c. Refuge and Reform: Snakes, Gleaners, and Niṣad ̄ as in the Sanskrit Epics.
Jawaharlal Nehru University History Seminar, 13 January. Courtesy of the author.
——— . 2010d. Ilā and the Yādavas: Somavaṃśa and Putrikā in the Sanskrit Epics.
Cardiff University Workshop: Genealogy in South Asia, May 26–28.
——— . 2010e. Personal communication in conversation, Madison Wisconsin,
October.
Brodbeck, Simon and Brian Black, eds. 2007. Gender and Narrative in the
Mahābhārata. London: Routledge.
Bronkhorst, Johannes. 1981. The Orthoepic Diaskeuasis of the Ṛgveda and the Date
of Pāṇini. IIJ 23: 83–95.
 bibliography 691

——— . 1985a. Dharma and Abhidharma. Bulletin of the School of Oriental and African
Studies 48: 305–20.
——— . 1985b. The Origin of an Indian Dietary Rule: Evidence for a Lost Mānava
Work on Dharma. Aligarh Journal of Oriental Studies 2.1–2 (Ram Suresh Tripathi
Commemoration Volume): 123–32.
——— . 1993. The Two Sources of Indian Asceticism. Bern: Peter Lang.
——— . 1998. Did the Buddha Believe in Karma and Rebirth? JIABS 21/1: 1–19.
——— . 1999a. Language et réalité: sur un episode de la pensée indienne. Bibliothèque de
l’Ecole des Hautes Etudes, Sciences Religieuses 105. Brepols: Turnhout.
——— . 1999b. Why Is There Philosophy in India? Amsterdam: Royal Netherlands
Academy of Arts and Sciences.
——— . 2000. Karma and Teleology: A Problem and It’s Solutions in Indian Philosophy.
Studia Philologica Buddhica Monograph Series, 15. Tokyo: International Institute
for Buddhist Studies.
——— . 2005a. The Background of Early Buddhism. Paper presented at the 4th
Dubrovnik International Conference on the Sanskrit Epics and Purāṇas.
Dubrovnik. September.
——— . 2005b. Renunciation in the Rājadharmaparvan (draft). Paper presented at the
4th Dubrovnik International Conference on the Sanskrit Epics and Purāṇas,
Dubrovnik. September.
——— . 2007. Greater Magadha. Studies in the Culture of Early India. Handbook of
Oriental Studies, Section 2: India, Vol. 19. Leiden: E. J. Brill.
——— . 2008. Aux origines de la philisophie Indienne. Collection Le Maitre et le
disciple. Infolio editions.
——— . 2010a. The Mahābhārata and the Revival of Brahmanism. Presentation at The
Brown Conference on Early Indian Philosophy in the Mahābhārata. Providence,
RI, April 9.
——— . 2010b. Who Is Liberated? The Notion of Liberation While Alive in Some Indian
Texts. Asiatische Studien/Études Asiatiques. Preveiw copy courtesy of the author.
——— . In press. Manu and the Mahābhārata. In L. Kulikov and M. Rusanov, eds. T.
Ya. Elizarenkova Memorial Volume, Book 2. Moscow: Russian State University for
the Humanities. Courtesy of the author.
Brough, John. 1962. The Gandhārī Dharmapāda: Edited with an Introduction and
Commentary. London Oriental Series, vol. 7. London: Oxford University Press.
Brown, W. Norman. 1972. Duty as Truth in Ancient India. Proceedings of the American
Philosophical Society 116/3: 252–68.
——— . 1933. The Story of Kālaka. Oriental Series 1. Washington, DC: Smithsonian
Institute, Freer Gallery of Art.
Bühler, Georg, trans. [1886] 1969. The Laws of Manu. New York: Dover.
Burnell, A. and E. W. Hopkins. [1891] 1971. Mānavadharmaśāstra. London: K. Paul,
Trench, Trübner & Co. Repr. New Delhi: Orinetal Books Reprint Corp.
Butler, Judith. 1993. Bodies That Matter: On the Discursive Limits of Sex. London:
Routledge.
Byodo, Tsyūsho. [1930] 1969. Aśvaghoṣa’s Acquaintance with the Mokshadharma of
the Mahābhārata. Resume of A Study of Indian Buddhist Literature, No. 1: A Study
692 bibliography

of the Sanskrit Buddhacarita (Indo-Bukkyō-Bungaku-no Kenkyū: Bonbun Bussho-

gōsan-no kenkyū), 565–8. Tokyo: Institute of Indological Studies. Kindly provided
by Muneo Tokunaga.
Caillat, Colette. 1997. Vedic and Early Middle Indo-Aryan. In Witzel ed., 1997c, 15–32.
Carter, John Ross. 1976. Traditional Definitions of the Term dhamma. Philosophy East
and West 26: 329–37.
——— . 1978. Dhamma: Western Academic and Sinhalese Buddhist Interpretations:
A Study of a Religious Concept. Tokyo: Hokusuido Press.
Caudill, David S. 1997. Lacan and the Subject of Law: Toward a Psychoanalytic Critical
Legal Theory. Humanity Books. ISBN 1573923141.
Chakrabarti, Dilip K. 2011. Royal Messages by the Wayside: Historical Geography of the
Asokan Edicts. New Delhi: Aryan Books International.
Chakravarti, Uma. 1987. The Social Dimensions of Early Buddhism. Delhi: Oxford
University Press.
——— . 2006. Everyday Lives, Everyday Histories: Beyond the Kings and Brahmanas of
‘Ancient’ India. New Delhi: Tulika Books.
——— . 2009. Of Meta-narratives and ‘Master’ Paradigms: Sexuality and the Reification
of Women in Early India. Monograph. New Delhi: Centre for Women’s
Development Studies.
Chakravarty, Kalyan Kumar, ed. 2009. Text and Variations of the Mahābhārata:
Contextual, Regional, and Performative Traditions. National Mission for
Manuscripts. New Delhi: Indira Gandhi National Centre for the Arts, Munshiram
Manoharlal Publishers.
Chappell, David W. 1980. Early Forebodings of the Death of Buddhism. Numen 27:
122–53.
Chattopadhyaya, Debiprasad. [1959] 1973. Lokāyata: A Study in Ancient Indian
Materialism. New Delhi: People’s Publishing House.
Chiba, Masaji. 1986. Asian Indigenous Law: In Interaction with Received Law. London:
KPI Limited.
Chinmayananda, Swami. 1993. Thousand Ways to the Transcendental: Vishnu
Sahasranama. Bombay: Central Chinmaya Mission Trust.
Clarke, Shayne. 2009. Monks Who Have Sex: Pārājika Penance in Indian Buddhist
Monasticisms. JIP 37: 1–43.
Clay, Jenny Strauss. 2003. Hesiod’s Cosmos. Cambridge: Cambridge University Press.
Clooney, Francis X. 1985. The Concept of Dharma in the Mīmāṃsā Sutras of Jaimini.
In Professor Kuppaswami Sastri Birth Centenary Commemoration Volume, Part 2,
175–87. Madras: Kuppaswami Sastri Research Institute.
——— . 1990. Thinking Ritually: Rediscovering the Pūrva Mīmāṃsā of Jaimini. Vienna:
Publications of the De Nobili Research Library.
Collins, Steven. 1982. Selfless Persons: Imagery and Thought in Theravāda Buddhism.
Cambridge: Cambridge University Press.
——— . 1988. Monasticism, Utopias and Comparative Social Theory. Religion 18: 101–35.
——— . 1990. On the Very Idea of the Pali Canon. JPTS 15: 89–126.
——— . 1992a. Nirvāṇa, Time and Narrative. HR 31: 215–46.
——— . 1992b. Oral Aspects of Pāli Literature. IIJ 35: 121–35.
 bibliography 693

——— . 1993. The Discourse on What is Primary (Aggañña Sutta). An Annotated

Translation. JIP 21: 301–93.
——— . 1996. The Lion’s Roar on the Wheel-Turning King: A Response to Andrew
Huxley’s ‘The Buddha and the Social Contract’. JIP 24: 421–46.
——— . 1998. Nirvāṇa and Other Buddhist Felicities. Utopias of the Pali Imaginaire.
New York: Cambridge.
——— and Andrew Huxley. 1996. The Post-Canonical Adventures of Mahāsammata.
JIP 24: 623–48.
Conze, Edward, trans. 1961. The Large Sūtra on Perfect Wisdom. London: Luzac.
——— . 1962. Buddhist Thought In India. London: Allen and Unwin.
——— . 1967. Materials for a Dictionary of the Prajñāpāramitā Literature. Tokyo: Suzuki
Research Foundation.
——— . [1959] 1977. Buddhist Scriptures. Harmondsworth: Penguin.
Cousins, Lance. 1995. Abhidhamma. In J. R. Hinnells, ed. A New Dictionary of
Religions. Oxford and Cambridge MA: Blackwell.
——— . 1996. The Dating of the Historical Buddha: A Review Article. JRAS, Series 3,
6/1: 57–63. Posted online on the Indology website at: http://www.ucl.ac.
uk/~ucgadkw/indology.html
Couture, André. 1996. The Harivaṃśa: A Supplement to the Mahābhārata. JVS 4/3:
127–38.
——— . 2001: From Viṣṇu’s Deeds to Viṣṇu’s Play, or Observations on the Word
Avatara as a Designation for the Manifestations of Viṣṇu. JIP 29/3: 313–26.
——— . 2006. Dharma as a Four-Legged Bull: A Note on an Epic and Purânic Theme.
In Raghunath Panda and Madhusudan Mishra, eds. Voice of the Orient A Tribute
to Prof. Upendranath Dhal, 69–76. Delhi: Eastern Book Linkers.
Cowell, E. B. [1894] 1968. The Buddha-karita of Asvaghosha. Sacred Books of the East,
vol. 49. Delhi: Motilal Banarsidass.
——— . [1895] 2005. The Jātaka or Stories of the Buddha’s Former Births, Vols. 1–2.
Delhi: Motilal Banarsidass.
Cox, Collett. 1992. Attainment through Abndonment: The Sarvāstivādin Path of
Removing Defilements. In R. E. Buswell, Jr. and R. Gimello, eds. Paths to
Liberation: The Mārga and its Transformations in Buddhist Thought. Kuroda
Institute. Studies in East Asian Buddhism 7, 63–105. Hawaii: University of
Hawaii Press.
——— . 1995. Disputed Dharmas: Early Buddhist Theories on Existence.
Studia Philologica Budhica 9. Tokyo: International Institute for Buddhist
Studies.
——— . 2004. From Category to Ontology: The Changing Role of Dharma in
Sārvāstivāda Abhidharma. JIP 32: 543–97.
Cribb, J. 2005. The Greek Kingdom in Bactria, Its Coinage and Collapse. In Osmund
Bopearachchi and Marie-Françoise Boussac, eds. Afghanistan, Ancien Carrefour
entre l’est et l’ouest. Indicopleustoi: Archaeologies of the Indian Ocean, Vol. 3, 207–25.
Turnhout: Brepols.
Curley, David. 2005. Battle and Self-Sacrifice in a Bengali Warrior’s Epic: Lāusen’s
Quest to Be a Rājā in Dharmamaṅgal. Manuscript courtesy of the author.
694 bibliography

Custodi, Andrea. 2005. Dharma and Desire: Lacan and the Left Half of the
Mahābhārata. Ph.D. Dissertation, George Washington University,
Washington, DC.
Dandekar, R. N. 1966. Anuśāsanaparvan. Introduction and Apparatus. Vol. 17 of V. S.
Sukthankar et al., eds. 1933–70. Mahābhārata: Critical Edition, 24 vols. with
Harivaṃśa. Poona: Bhandarkar Oriental Research Institute.
Das, Gurcharan. 2006. On the Difficulty of Being Good; Draupadi’s Question:
Lessons for Public and Corporate Governance. In Pradeep Trikha, ed. Textuality
and Intertextuality in the Mahābhārata, 5–31, 99–133. New Delhi: Sarup & Sons.
——— . 2008. Foreword to Kathleen Garbutt trans. Mahābhārata Book 5, Preparations for
War, Vol. 1, xv–xxiii. Clay Sanskrit Series. New York: New York University Press.
——— . 2009. The Difficulty of Being Good: On the Subtle Art of Dharma. New Delhi:
Allen Lane, an Imprint of Penguin.
Davis, Donald R. 2007a. Hinduism as a Legal Tradition. JAAR 75/2: 241–67.
——— . 2007b. On Ātmatuṣṭi as a Source of Dharma. JAOS 127/3: 279–96.
De, Sushil Kumar, ed. 1958. The Droṇaparvan. In Vishnu S. Sukthankar and S. K.
Belvalkar, general eds. The Mahābhārata for the First Time Critically Edited, Vol. 8.
Poona: Bhandarkar Oriental Institute.
De Bruin, Hanne M., trans. 1998. Karṇamōksam: A Play by Pukalentippulavar.
Pondicherry: Institut Française de Pondichéry: École Française d’Extrême Orient
and International Institute for Aaian Studies.
Defourny, Michel. 1976. Note sur le symbolisme de la corne dans le Mahābhārata et la
mythologie classique. IIJ 18: 17–23.
——— . 1978. Le mythe de Yayāti dans la littérature épique et purāṇique. Étude de
mythologie hindoue. Paris: Société d’Edition “Les Belles Lettres”.
Derrett, J. D. M. 1965. A Newly-discovered Contact between Arthaśāstra and
Dharmaśāstra. ZDMG 115: 134–52.
——— . 1975. Manuśāstravivaraṇa. Translation of text and Bhāruci’s commentary on
Books, 2 vols, 6–12. Wiesbaden: Franz Steiner Verlag.
——— . 1976. Rājadharma. JAS 35: 597–609.
Devadhar, C. R., ed. and trans. [1985] 2005. Raguvaṃśa of Kālidāsa: Edited with Critical
Introduction, English Translation and Notes. Delhi: Motilal Banarsidass.
Dhand, Arti. 2002. The Dharma of Ethics, the Ethics of Dharma. Journal of Religious
Ethics 30/3: 347–72.
——— . 2004. The Subversive Nature of Virtue in the Mahābhārata: A Tale about
Women, Smelly Ascetics and God. JAAR 72/1: 33–58.
——— . 2007. Paradigms of the Good in the Mahābhārata: Śuka and Sulabhā in the
Quagmires of Ethics. In Brodbeck and Black, eds. 2007, 258–78.
——— . 2008. Woman as Fire, Woman as Sage: Sexual Ideology in the Mahābhārata.
Albany: SUNY Press.
Dimock, Edward C., Jr., Edwin Gerow, C. M. Naim, A. K. Ramanujan, Gordon
Roadarmel, and J. A. B. van Buitenen. 1974. The Literatures of India: An
Introduction. Chicago: University of Chicago Press.
Dimmitt, Cornelia and J. A. B. van Buitenen, trans. 1978. Classical Hindu Mythology:
A Reader in the Sanskrit Purāṇas. Philadelphia: Temple University Press.
 bibliography 695

Divakaruni, Chitra Bannerjee. 2008. The Palace of Illusions. New York: Doubleday.
Doniger, Wendy, trans., with Brian K. Smith. 1991. The Laws of Manu. London: Penguin.
——— 1995. Begetting on the Margin: Adultery and Surrogate Pseudomarriage in
Hinduism. In Lindsey Harlan and Paul B. Courtright, eds. From the Margins of
Hindu Marriage: Essays on Gender, Religion, and Culture, 160–83. Oxford: Oxford
University Press.
——— 1999. The Rig Veda. London: Penguin.
——— 2009a. The Hindus: An Alternate History. New York: Penguin Press.
——— 2009b. How to Escape from the Curse. Review of John Smith 2009. London
Review of Books 32/19: 17–18.
Dubey, D. P. 2001. Prayāga: The Site of the Kumbha Mela. New Delhi: Aryan Books
International.
Dubuisson, D. 1979. Trois théses sur le Rāmāyaṇa. Annales Économies Sociétés
Civilisations 3: 464–89.
Dumézil, Georges. 1968. Mythe et épopée, 1: L’Idéologie des trois fonctions dans les épopées
des peuples indo-européens. Paris: Gallimard.
——— . 1969. The Destiny of the Warrior. Alf Hiltebeitel trans. Chicago: University
of Chicago Press.
——— . 1973. The Destiny of a King. Alf Hiltebeitel, trans. Chicago: University of
Chicago Press.
——— . 1979. Mariages indo-européens suivi de quinze questions romaines. Paris: Payot.
Dumont, P.-E. 1948. The Horse-Sacrifice in the Taittirīya-Brāhmaṇa. The Eighth and
Ninth Prapāṭhakas of the Third Kāṇḍa of the Taittirīya-Brāhmaṇa with
Translation. Proceedings of the American Philosophical Foundation 92: 492f.
Dutt, M. N. trans. 1897. A Prose English Translation of the Harivaṃśa. Calcutta:
Elusium Press.
Dutt, Nalinaksha. 1960. Early Buddhist Monachism. Calcutta: Calcutta Oriental Book
Agency.
Edgerton, Franklin, trans. 1952. The Bhagavad Gītā. Part 1. Harvard Oriental Series 38.
Cambridge: Harvard University Press.
Eggeling, Julius, trans. 1963. The Śatapatha-Brāhmaṇa According to the Text of the
Mādhyandina School, Part 5: Books 11–14. Repr. Delhi: Motilal Banarsidass.
El-Ansary, Waleed. 2006. The Spiritual Significance of Jihād in the Islamic Approach
to Markets and the Environment. Doctoral Dissertation in Human Sciences, The
George Washington University.
Elizarenkova, T. J. 1995. Language and Style of the Vedic Ṛṣis. Albany: SUNY Press.
Embree, Ainslie T. 1988. Sources of Indian Tradition, vol. 1: From the Beginning to 1800,
2nd ed. New York: Columbia University Press.
Ensink, J. 1968. Mitrasaha, Sudāsa’s Son, with the Spotted Feet. In J. C. Heesterman,
G. S. Schokker, and V. I. Subramoniam, eds. Pratidānam: Indian, Iranian, and
Indo-European Studies Presented to Franciscus Bernardus Jacobus Kuiper on His
Sixtieth Birthday, 573–84. Janua Linguarum, Series Major 34. The Hague:
Mouton.
Erdosy, G. 1985. Origins of Cities in the Ganges Valley. Journal of the Economic and
Social History of the Orient 28: 81–109.
696 bibliography

——— . 1987. Early Historic Cities of Northern India. South Asian Studies 3: 1–23.
——— . 1995. City States of North India and Pakistan in the Time of the Buddha. In
F. R. Allchin, ed. The Archaeology of Early Historic South Asia, 99–122. Cambridge:
Cambridge University Press.
Erndl, Kathleen. 1991. The Mutilation of Śurpaṇakhā, In Richman 1991: 67–88.
Esnoul, Ann-Marie. 1979. Nārāyaṇīya Parvan du Mahābhārata. Paris: Société
d’Édition.
Falk, Harry. 1987. Viṣṇu im Veda. In Harry Falk, ed. Hinduismus und Buddhismus:
Festschrift für Ulrich Schneider, 112–33. Freiberg: Hedwig Falk.
——— . 1988. Vedische Opfer im Pāli-Kanon. Bulletin d’Études Indiennes 6: 225–54.
——— . 2006a. The Tidal Waves of Indian History: Between the Empires and Beyond.
In Olivelle, ed. 2006, 147–66.
——— . 2006b. Aśokan Sites and Artefacts: A Source-Book with Bibliography. Mainz am
Rhein: Verlag Philipp von Zabern.
Fitzgerald, James L. 1980. The Mokṣa Anthology of the Great Bhārata: An Initial
Survey of Structural Issues, Themes, and Rhetorical Strategies. Ph.D.
Dissertation, University of Chicago, Chicago.
——— . 2001. Making Yudhiṣṭhira the King: The Dialectics and the Politics of
Violence in the Mahābhārata. RO 54/1: 63–92.
——— . 2002. Nun Befuddles King, Shows Karmayoga Does Not Work: Sulabhā’s
Refutation of King Janaka at MBh 12.308. JIP 30/6: 641–77.
——— . 2003. The Many Voices of the Mahābhārata. Review of Hiltebeitel 2001.
JAOS 123/4: 803–18.
——— . 2004a. trans. and ed. The Mahābhārata, vol. 7: 11, The Book of the Women; 12,
The Book of Peace, Part One. Chicago: University of Chicago Press.
——— . 2004b. Dharma and its Translation in the Mahābhārata. JIP 32/5–6: 671–85.
——— . 2004c. Mahābhārata. In Mittal, ed., 2004, 52–74.
——— . 2005. Toward a Database of the non-Anuṣṭubh Verses of the Mahābhārata. In
Koskikallio, ed. 2005, 137–48.
——— . 2006a. Negotiating the Shape of “Scripture”: New Perspectives on the
Development and Growth of the Mahābhārata between the Empires. In Olivelle,
ed. 2006, 257–86.
——— . 2006b. Niṣka. Jewelry vs. Coins in MBh. E-mail to Patrick Olivelle and
Hiltebeitel, December 19.
——— . 2007. Bhīṣma Beyond Freud, The Fall of the Sky: Bhīṣma in the
Mahābhārata. In Brodbeck and Black eds. 2007, 189–207.
——— . 2010. No Contest between Memory and Invention: The Invention of the
Pāṇḍava Heroes of the Mahābhārata. In David Konstan and Kurt A. Raaflaub eds.
Epic and History. 103–21. Malden, MA: Wiley-Blackwell.
Framarin, Christopher G. 2006. Motivation in the Manusmṛti. JIP 34: 397–413.
——— . 2009. Desire and Motivation in Indian Philsophy. Hindu Studies Series.
London and New York: Routledge.
Francavilla, Domenico. 2006. The Roots of Hindu Jurisprudence: Sources of Dharma and
Interpretation in Mīmāṃssā and Dharmaśāstra. Corpus Iuris Sanscriticum. Vol. 7.
Torino: CESMEO.
 bibliography 697

Franco, Eli. 2004. The Spitzer Manuscript: The Earliest Philosophical Manuscript in
Sanskrit, 2 vols. Vienna: Verlag der Österreichischen Akademie der
Wissenschaften.
Frasca, Richard. 1990. The Theatre of the Mahābhārata: Terukkūttu Performances in
South India. Honolulu: University of Hawaii Press.
Frauwallner, Erich. 1956. The Earliest Vinaya and the Beginnings of Buddhist Literature.
Rome: Istituto Itialiano per il Medio ed Estremo Oriente.
——— . 1995. Studies in Abhidharma Literature and the Origins of Buddhist Philosophical
Systems. Sophie Francis Kidd trans. Albany: SUNY Press.
Freiberger, Oliver. 2000. Der Orden in der Lehre: Zur religiösen Deutung des Saṅgha im
frühen Buddhismus. Wiesbaden: Harrassowitz Verlag.
——— . 2008. July 28 e-mail.
Frye, Richard N. 1983. History of Ancient Iran. Munich: Beck.
Gallop, Jane. 1985. Reading Lacan. Ithaca: Cornell University Press.
Ganeri, Jonardon. 2005. A Dynamic Tradition of Truth-Telling: Moral Innovation in
the Mahābhārata, In Federico Squarcini, ed. Boundary Dynamics and Construction
of Traditions in South Asia, 175–201. Florence: University of Florence Press and
Munshiram Manoharlal.
——— . 2007. The Concealed Art of the Soul. New York: Oxford University Press.
Ganguli, Kisari Mohan, trans., and Pratap Chandra Roy, publisher. [1884–96] 1970.
The Mahabharata, 12 vols. New Delhi: Munshiram Manoharlal.
Gawronski, Andrzej. 1914–15. Gleanings from Aśvaghoṣa’s Buddhacarita. RO 1–42.
——— . 1919. Studies about the Sanskrit Buddhist Literature. Cracow: Nakladem
Akademji Umiejetnosci.
Geertz, Clifford. 1973. The Interpretation of Cultures: Selected Essays. New York: Basic Books.
Geiger, Magdelene and Wilhelm Geiger. 1920. Pāli Dhamma vornehmlich in der
kanonischen Literatur. Munich: Bayerischen Akademie der Wissenschaften.
Reprinted in W. Geiger, 1973, 102–228.
Geiger, Wilhelm. 1973. Kleine Schriften zur Indologie und Buddhismuskunde.
Wiesbaden: Franz Steiner.
Geldner, Karl Friedrich. 1951. Der Rigveda. Aus dem Sanskrit ins Deutsch übersetzt und
mit einem laufenden Kommentar versehen, 3 vols. Harvard Oriental Series, vols.
33–35. Cambridge, MA: Harvard University Press.
Gethin, Rupert. 1986. The Five Khandas: Their Treatment in the Nikāyas and early
Abhidhamma. JIP 13: 35–53.
——— . 1992. The Buddhist Path to Awakening: A Study of the Bodhi-Pakkhiyā
Dhammā. Brill’s Indological Library 7. Leiden: Brill.
——— . 1997. Cosmology and Meditation: From the Aggañña Sutta to the Mahāyāna.
HR 36/3: 183–217.
——— . 2004. He Who Sees Dhamma Sees Dhammas. JIP 32/5–6: 513–42.
——— . 2006. Mythology as Meditation: From the Mahāsudassana Sutta to the
Sukhāvatīvyūha Sūtra. JPTS 28: 63–112.
Ghosh, Jayatri. 2000. Satyavatī: Matriarch of the Mahābhārata. In Mandakranta Bose,
ed. Faces of the Feminine in Ancient, Medieval, and Modern India. New York: Oxford
University Press.
698 bibliography

Glucklich, Ariel. 1988. The Royal Scepter (Daṇḍa) as Legal Punishment and Sacred
Symbol. HR 28: 97–122.
——— . 1994. The Sense of Adharma. New York: Oxford Univesity Press.
Gokhale, B. G. 1965. The Early Buddhist Elite. Journal of Indian History 42/2:
391–402.
——— . 1980. Bhakti in Early Buddhism. Journal of Asian and African Studies 15:
16–28.
Goldman, Robert P. 1976. Vālmīki and the Bhṛgu Connection. JAOS 96/1: 97–101.
——— . 1978. Fathers, Sons, and Gurus: Oedipal Conflict in the Sanskrit Epics. JIP
6: 325–93.
——— . trans. 1984. The Rāmāyaṇa of Vālmīki, vol. 1: Bālakāṇḍa. Princeton: Princeton
University Press.
——— . 1995. Gods in Hiding: The Mahābhārata’s Virāṭa Parvan and the Divinity of
the Indian Epic Hero. In Narang 1995, ed., 73–100.
——— . 1996. Vālmīki and Vyāsa: Their Contribution to India’s Discourse on
Ethnicity. Journal of the Oriental Institute of Baroda 46/1–2: 1–14.
——— . 2004. Resisting Rāma: Dharmic Debates on Gender and Hierarchy and the
Work of the Vālmīki Rāmāyaṇa. In Bose, ed. 2004: 19–46.
——— and Muneo Tokunaga, eds. 2009. Epic Undertakings: Papers of the 12th World
Sanskrit Conference, Vol. 2. Delhi: Motilal Banarsidass.
——— and Sally J. Sutherland Goldman, trans. 1996. The Rāmāyaṇa of Vālmīki,
Vol. 5: Sundarakāṇḍa. Princeton: Princeton University Press.
——— . 2004. Rāmāyaṇā. In Mittal and Thursby, eds., 75–96.
——— and Barend A. van Nooten, trans. 2009. The Rāmāyaṇa of Vālmīki,
Vol. 6: Yuddhakāṇḍa. Princeton: Princeton University Press.
Gombach, Barbara. 2000. Ancillary Stories in the Sanskrit Mahābhārata. Ph.D.
Dissertation, 2 parts, Columbia University.
Gombrich, Richard F. 1984. Notes on the Brahmanical Background to Buddhist
Ethics. In G. Dhammapala et al. eds. Buddhist Studies in Honour of Hammalava
Saddhātissa, 88–102. Nugegoda, Sri Lanka: Hammalava Saddhātissa Felicitation
Volume Committee.
——— . 1985. The Vessantara Jātaka, the Rāmāyaṇa and the Dasaratha Jātaka. Ernest
Bender, ed. Special Issue. Indological Studies Dedicated to Daniel H. H. Ingalls.
JAOS 105/3: 427–37..
——— . 1988. Theravāda Buddhism: A Social History from Ancient Benares to Modern
Colombo. London: Routledge and Kegan Paul.
——— . 1990. Recovering the Buddha’s Message. In Tadeusz Skorupski, ed.
The Buddhist Forum, Vol. 1. London.
——— . 1992a. The Buddha’s Book of Genesis? IIJ 35: 179–91.
——— . 1992b. Why is a Khattiya Called a Khattiya? The Aggañña Sutta Revisited.
JPTS 17: 213–14.
——— . 1996. How Buddhism Began: The Conditioned Genesis of the Early Teachings.
London: Athlone.
Gonda, Jan. 1975. Pratiṣṭhā. In Selected Studies. Vol. 2. Leiden, pp. 338–74. In Jan C.
Heesterman, ed., Pratidānam. Indian, Iranian and Indo-European Studies Presented
 bibliography 699

to F. B. J. Kuiper on his sixtieth birthday, 563–72. Janua Linguarum, Series Major

34. The Hague: Mouton.
González-Reimann, Luis. 2002. The Mahābhārata and the Yugas: India’s Great Epic
Poem and the Hindu System of World Ages. New York: Peter Lang.
——— . 2003. “Time” in the Mahābhārata, and the Time of the Mahābhārata. Panel
on “Rethinking the Mahābhārata,” 32nd Annual Conference on South Asia,
Madison, WI. Courtesy of the author. See now González-Reimann 2010.
——— . 2006a. The Divinity of Rāma in the Rāmāyaṇa of Vālmīki. JIP 34: 203–20.
——— . 2006b. Viṣṇu as Fish: The Growth of a Story, from the Brāhmaṇas to the
Purāṇas. JVS 5/1: 221–38.
——— . 2008. Kalpas in the Mahābhārata: The Days and Nights of Brahmā.
Presentation at the American Oriental Society Annual meeting, Chicago.
Courtesy of the author.
——— . 2010. “Time” in the Mahābhārata, and the Time of the Mahābhārata. In Sheldon
I. Pollock, ed. Epic and Argument in Sanskrit Literary History, 61–73. Delhi: Manohar.
——— . Forthcoming. Cosmic Cycles, Cosmology and Cosmography. Brill’s
Encyclopedia of Hinduism.
Grosz, Elizabeth. 1990. Jacques Lacan: A Feminist Introduction. London: Routledge.
Grünendahl, Reinhold. 1997. Zur Textkritik des Nārāyaṇīya, chapter 2; Zur Stelling der
Nārāyaṇīya im Mahābhārata, chapter 7. In Schreiner, ed. 1997, 30–73. 197–240.
Gupta, Anand Swarup, ed. and Satyamsu Mohan Mukhopadhyaya, trans. 1968. The
Vāmana Purāṇa. Varanasi: All India Kashiraj Trust.
Gyatso, Janet. 2005. Sex. In Donald S. Lopez, Jr., ed. Critical Terms for the Study of
Buddhism, 271–91. Chicago: University of Chicago Press.
Hacker, Paul. 1960. Zur Entwicklung der Avatāralehre. Wiener Zeitschrift für die Kunde
Süd- ind Ostasiens 4: 47–70 (= Kleine Schriften, 404–27).
——— . 1965. Dharma im Hinduismus. Zeitschrift für Missionwissenschaft und
Religionswissenschaft 49: 96–100. Also in L. Schmithausen, ed., Kleine Schriften,
pp. 496–509. Wiesbaden: Franz Steiner, 1978. See Hacker 2006, 2009.
——— . [2006] 2009. Dharma in Hinduism, translated by Donald R. Davis, Jr. JIP
34/5: 479–96. Reprinted in Olivelle 2009, 475–92.
Halbfass, Wilhelm. 1988. India and Europe: An Essay in Understanding. Albany:
SUNY Press.
——— . 1992. On Being and What there Is: Classical Vaiśeṣika and the History of Indian
Ontology. Albany: SUNY Press.
Hara, Minoru. 1974. A Note on the Rākṣasa Form of Marriage. JAOS 94: 296–306.
——— . 2001. The Hindu Concept of Friendship: A Note on Sanskrit praṇaya. Revista
degli Studi Orientali 75/1–4: 157–87.
Hardy, Friedhelm. 1983. Viraha-Bhakti: The Early History of Kṛṣṇa Devotion in South
India. Delhi: Oxford University Press.
Hare, E. M. et al., eds. 1952–. Pāli Tripiṭakam Concordance. London: Pāli Text Society
and Luzac.
Hare, R. M. 1986. The Language of Morals. Oxford: Clarendon Press.
Harindranath, A. 2010. On the Killing of Sisupala. http://.boloji.com.hinduism/177.
html, May 16.
700 bibliography

Hart, George L. 1975. Poems of Ancient Tamil, Their Milieu and their Sanskrit
Counterparts. Berkeley: University of California Press.
——— and Hank Heifitz, trans. and eds. 1999. Four Hundred Songs of War and
Wisdom: An Anthology of Poems from Classical Tamil: The Purananuru. New York:
Columbia University Press.
Härtel, Herbert. 2007. The Pre-Kuṣāṇa and Early Kuṣāṇa Levels at Sonkh. In
Srinivasan, ed. 2007: 319–50.
Heesterman, J. C. 1957. The Ancient Indian Royal Consecration: The Rājasūya Described
According to the Yajus Texts and Annotated. ‘s-Gravenhage: Mouton & Co.
——— . 1964. Brahmin, Ritual and Renouncer. WZKSO 8: 1–31.
——— . 1978. The Return of the Veda Scholar. In Heesterman, ed., Pratidānam.
Indian, Iranian and Indo-European Studies Presented to F. B. J. Kuiper on His
Sixtieth Birthday. Janua Linguarum, Series Major 34, 436–47. The Hague:
Mouton.
——— . 1978. Veda and Dharma. In O’Flaherty and Derrett, eds. 1978, 80–95.
——— . 2001. Gift, Marriage, and the Denial of Reciprocity. In Karttunen and
Koskikallio, eds. 2001, 243–59.
Hess, Linda, 2001. Lovers’ Doubts: Questioning the Tulsi Rāmāyaṇ. In Richman, ed.
2001, 25–48.
Hill, Peter. 2001. Fate, Predestination and Human Action in the Mahābhārata: A Study
in the History of Ideas. New Delhi: Munshiram Manoharlal.
Hiltebeitel, Alf. [1976] 1990. The Ritual of Battle. Krishna in the Mahābhārata. Albany:
SUNY Press.
——— . 1977. Nahuṣa in the Skies: A Human King of Heaven. HR 16/4: 329–50.
——— . 1979. Hindu Mythology and Its Evils. Review of O’Flaherty 1976. HR 18/3:
269–75.
——— . 1980a. Draupadī’s Garments. IIJ 22: 97–112.
——— . 1980b. Śiva, the Goddess, and the Disguises of the Pāṇḍavas and Draupadī.
HR 20: 147–74.
——— . 1980–1981. Sītā Vibhūṣitā: The Jewels for Her Journey. Ludwik Sternbach
Commemoration Volume. IT 8–9: 193–200.
——— . 1981. Draupadī’s Hair. In Madeleine Biardeau, ed., Autour de la déesse hindoue.
Puruṣārtha, Vol. 5, 179–214.
——— . 1982a. Sexuality and Sacrifice: Convergent Subcurrents in the Firewalking
Cult of Draupadī. In Fred W. Clothey, ed., Images of Man: Religion and Historical
Process in South Asia, 72–111. Madras: New Era Publications.
——— . 1982b. Brothers, Friends, and Charioteers: Parallel Episodes in the Irish and
Indian Epics. In Edgar C. Polomé, ed., Homage to Georges Dumézil, Journal of
Indo-European Studies. Monograph, No. 3, 85–112.
——— . 1984. The Two Kṛṣṇas on One Chariot: Upaniṣadic Imagery and Epic
Mythology. HR 24: 1–26.
——— . 1985. Purity and Auspiciousness in the Sanskrit Epics. In Frédérique Appfel
Marglin and John Carman, eds., Essays on Purity and Auspiciousness. Journal of
Developing Societies 1: 41–54.
 bibliography 701

——— . 1988. The Cult of Draupadī, 1: Mythologies, From Gingee to Kurukṣetra. Chicago:
University of Chicago Press.
——— . 1989. Kṛṣṇa at Mathurā. In Doris Srinivasan, gen. ed. Mathurā: A Cultural
Heritage, 93–102. New Delhi: Manohar and American Institute of Indian
Studies.
——— . 1991a. The folklore of Draupadī: Sarees and Hair. In Arjuna Appadurai,
Frank Korom, and Margaret Mills, eds., Gender, Genre, and Power in South Asian
Expressive Traditions, 395–427. Philadelphia: University of Pennsylvania Press.
——— . 1991b. The Cult of Draupadī, 2: On Hindu Ritual and the Goddess. Chicago:
University of Chicago Press.
——— . 1995. Dying Before the Mahābhārata War: Martial and Transsexual
Bodybuilding for Aravāṉ. JAS 54/2: 447–73.
——— . 1998a. Hair Like Snakes and Moustached Brides: Crossed Gender in an
Indian Folk Cult. In Hiltebeitel and Barbara D. Miller, eds., Hair: Its Meaning and
Power in Asian Cultures, 143–76. Albany: SUNY Press.
——— . 1998b. Empire, Invasion, and India’s National Epics. IJHS 2/3: 387–421.
——— . 1998c. Review of Brockington 1998. IIJ 43/(2000): 161–9.
——— . 1999a. Rethinking India’s Oral and Classical Epics: Draupadī among Muslims,
Rajputs, and Dalits. Chicago: University of Chicago Press.
——— . 1999b. Kuttāṇṭavar: The Divine Lives of a Severed Head. In Elizabeth
Schömbucher and Claus Peter Zoller, eds. Ways of Dying: Death and its Meaning
in South Asia, 276–310. Delhi: Manohar.
——— . 1999c. Fathers of the Bride, Fathers of Satī: Myths, Rites, and Scholarly
Practices. Thamyris: Mythmaking from Past to Present 6/1: 65–94.
——— . 2000. Draupadī’s Question. In Kathleen Erndl and Alf Hiltebeitel, eds., Is the
Goddess a Feminist? The Politics of South Asian Goddesses, 113–22. New York and
Sheffield, England: New York University Press and Sheffield Academic Press.
——— . 2001a. Rethinking the Mahābhārata: A Reader’s Guide to the Education of the
Dharma King. Chicago: University of Chicago Press.
——— . 2001b. Bhīṣma’s Sources. In Karttunen and Koskikallio, eds. 2001, 261–78.
——— . 2002. Allusion and Explication in the Gītā’s Epic World. In John Stratton
Hawley, ed. Essays from the 2002 Seminar, “A Temple and a Text: Two Monuments
in the Intellectul Life of Dennis Hudson”. JVS 11/1: 131–40.
——— . 2003. India’s Epics: Writing, Orality, and Divinity. In Arvind Sharma, ed. The
Study of Hinduism, 114–38. Columbia, SC: University of South Carolina Press.
——— . 2004a. More Rethinking the Mahābhārata: Toward a Politics of Bhakti. IIJ
47: 203–27.
——— . 2004b. Role, Role Model, and Function: The Sanskrit Epic Warrior in
Comparison and Theory. In Jacqueline Suthren Hirst and Lynn Thomas, eds.
Playing For Real: Hindu Role Models, Religion, and Gender, 27–50. Delhi: Oxford.
——— . 2004c. The Good Fortune of Getting Stepped On: Some South Indian
Exorcisms. Conference on Identity, Group Psychology and Violence: On the
Emergence of Love and Rage in Human Networks. May 21–23. Co-sponsored by
The Washington School of Psychiatry, George Washington University Human
702 bibliography

Sciences Program, and the journal Psychoanalysis, Culture, and Society. “Women,
Violence, Ritual” panel.
——— . 2005a. Not Without Subtales: Telling Laws and Truths in the Sanskrit Epics.
JIP 33: 455–511.
——— . 2005b. Buddhism and the Mahābhārata: Boundary Dynamics and Textual
Practice. In Federico Squarcini, ed. Boundary Dynamics and Construction of
Traditions in South Asia, 107–31. Florence: University of Florence Press and
Munshiram Manoharlal.
——— . 2005c. Weighting Orality and Writing in the Sanskrit Epics. In Koskikallio,
ed. 2005, 81–111.
——— . 2005d. On Reading Fitzgerald’s Vyāsa. JAOS 125/2: 241–61.
——— . 2006a. The Nārāyaṇīya and the Early Reading Communities of the
Mahābhārata. In Olivelle, ed. 2006, 227–55.
——— . 2006b. Aśvaghoṣa’s Buddhacarita: The First Known Close and Critical
Reading of the Brahmanical Sanskrit Epics. JIP 34: 229–86.
——— . 2006c. Mahābhārata. In Stanley Wolpert, ed. Encyclopedia of India, 4 vols.
Thompson Gale Corporation. Vol. 3, 82–93.
——— . 2006d. Rāmāyaṇa. In Stanley Wolpert, ed. Encyclopedia of India, 4 vols.
Thompson Gale Corporation. Vol. 3, 390–9.
——— . 2007a. Among Friends: Marriage, Women, and Some Little Birds. In
Brodbeck and Black, eds. 2007, 110–43.
——— . 2007b. Krishna and the Death of Karna. In Edwin Bryant, ed. Krishna:
A Sourcebook, 23–76. New York: Oxford University Press.
——— . 2009a. Authorial Paths through the Two Sanskrit Epics, Via the
Rāmopākhyāna. In Goldman and Tokunaga, eds. 2009, 169–214.
——— . 2009b. Why Itihāsa? New Possibilities and Limits in Considering the
Mahābhārata as History. Presentation at Kyoto, 14th World Sanskrit Conference.
To appear in Hiltebeitel 2011a.
——— . 2009c. Recontextualizing Satire of Dharmaśāstra in the Aggañña Sutta. RoSA
3/1: 77–92.
——— . 2010a. Mapping Bhakti through Friendship in the Sanskrit Epics. In Sheldon
I. Pollock, ed. Epic and Argument in Sanskrit Literary History, 91–116. Delhi:
Manohar.
——— . 2010b. Why Itihāsa? New Possibilities in Considering the Mahābhārata’s
Intention as History. In Adluri, ed. 2010a, 59–86.
——— . 2010c. Dharma. South Asian Spirituality Series, Henry Rosemont, ed.
Honolulu: University of Hawai’i Press.
——— . 2011a. Reading the Fifth Veda. Selected Essays, Vol. 1. In Vishwa Adluri and
Joydeep Bagchee, eds. Leiden: E. J. Brill.
——— . 2011b. When the Goddess Was a Woman. Selected Essays, Vol. 2. In Vishwa
Adluri and Joydeep Bagchee, eds. Leiden: E. J. Brill.
——— . In press. Mapping Bhakti through Hospitality and Friendship in the Sanskrit
Epics. In John Brockington, ed. Battles, Bards, Brāhmans. Papers from the Epics
Section of the 13th World Sanskrit Conference, Edinburgh, July 2006, Vol. 2.
Petteri Koskikallio and Asko Parpola, gen. eds, 113–38. Delhi: Motilal Banarsidass.
 bibliography 703

——— . Forthcoming-a. On Sukthankar’s “S,” the Śākuntalā-Upākhyāna, and the

Shortsightedness of Some Well-Known Criticisms of the Pune Critical Edition. In
Vishwa Adluri ed. Papers from 2010 University of Wisconsin South Asia
Conference double panel on the Critical Pune Edition. In JVS.
——— . Forthcoming-b. Mokṣa and Dharma in the Mokṣadharma. In James L.
Fitzgerald ed. Papers from 2010 Brown University Conference on Philosophy in
the Mokṣadharma Parvan. In JIP.
——— . Forthcoming-c. Epic Aśvamedhas. In Joel Brereton, ed. Vedic Workshop
Proceedings. Fourth International Vedic Workshop: The Vedas in Culture and
History, Austin, TX, May 2007. Delhi: Motilal Banarsidass.
——— . Forthcoming-d. The Southern Recension Reading of the Śakuntalā Story as
Its First Reading: What It Can Tell Us about the Original and the Second Reading
by Kālidāsa. In Deepika Tandon and Saswati Sengupta eds. Revisiting Kālidāsa’s
Abhijñāna-śākuntalam: Land, Love, Languages: Forms of Exchange in Ancient India.
New Delhi: Orient Blackswan.
——— . Forthcoming-e. The Archetypal Design of the Two Sanskrit Epics. In prepara-
tion for Dubrovnik International Conference on the Sanskrit Epics and Purāṇas
5th volume of conference papers.
——— . Forthcoming-f. Ritual as Dharma: The Narrowing and Widening of a Key
Term. In Jason D. Fuller, ed. Essays on Ritual Honoring Fred Clothey (title and
publisher uncertain).
——— . Forthcoming-g. Between History and Divine Plan: The Mahābhārata’s Royal
Patriline in Context. In Simon Brodbeck and James Hegarty eds. Papers from
2010 Cardiff University Conference on Genealogy in South Asia. In RoSA.
——— . Forthcoming-h. Just My Imagination? Puzzling thorugh a Duryodhana
Festival Near Dharmapuri, Tamilnadu. In Hilary Rodriguez ed. Hinduism in
Practice. London: Routledge.
Hiltebeitel, Alf and Randy Kloetzli. 2004. Kāla. In Mittal and Thursby, eds. 2004, 55–86.
Hinüber, Oskar von. 1992. The Arising of an Offence: Āpattisumuṭṭhāna: A Note on
the Structure and History of the Theravāda-Vinaya. JPTS 16: 68 n. 13.
——— . 1994. Die neun Aṅgas: Ein früher Versuch Einteilung Buddhisticher Texte.
Wiener Zeitschrift für die Kunde Südaiens 38: 121–36.
——— . 1995. Buddhist Law According to the Theravāda-Vinaya: A Survey of Theory
and Practice. JIABS, issue on Buddhism and Law, 18/1: 7–45.
——— . 1996. A Handbook of Pāli Literature. Indian Philology and South Asian
Studies, Vol. 2. Berlin: Walter de Gruyter.
Hirakawa, A. 1990. A History of Indian Buddhism: From Śākyamuni to Early Mahāyana.
Paul Groner, trans. and ed. Honolulu: University of Hawaii Press.
Hock, Hans Henrich. 2002. The Yājñavalkya Cycle in the Bṛhad Āraṇyaka Upaniṣad.
JAOS 122/2: 278–86.
Holt, John C. 1983. Discipline: The Canonical Buddhism of the Vinayapiṭaka. Delhi:
Motilal Banarsidass.
Holtzmann, Adolf. 1892–95. Das Mahābhārata und seine Theile, 4 vols. Kiel: C. F. Haesler.
Hopkins, E. W. 1882. On the Professed Quotations from Manu Found in the
Mahābhārata. JAOS 11: 239–75.
704 bibliography

——— . 1898a. The Bhārata and the Great Bhārata. American Journal of Philology 19: 1–24.
——— . 1898b. Parallel Features in the Two Sanskrit Epics. American Journal of
Philology 19: 138–51.
——— . 1901. Notes on the Çvetāçvatara, the Buddhacarita, etc. JAOS 22: 380–93.
——— . [1901] 1969. The Great Epic of India: Its Character and Origin. Calcutta:
Punthi Pustak.
——— . [1915] 1969. Epic Mythology. New York: Biblo and Tannen.
——— . 1923a. Review of Hermann Oldenberg, Das Mahābhārata, Seine Entstehung,
sein Inhalt, seine Form. JAOS 43: 56–50.
——— . 1923b. A Buddhistic Passage in Manu. JAOS 43: 244–46.
Horner, I. B. 1948. Early Buddhist Dhamma. Artibus Asiae 11: 115–23.
——— . trans. 1954. The Middle Length Sayings. London: Pali Text Society.
——— . [1942] 1997. The Book of the Discipline (Vinaya-Pitaka). Vol. 3. Suttavibhanga.
Oxford: Pali Text Society.
Horsch, Paul. 1957. The Wheel: An Indian Pattern of World-Interpretation. Sino-
Indian Studies, 5/3–4: 69–71.
——— . 2004. From Creation Myth to World Law: The Early History of Dharma.
Jarrod L. Whitaker, trans. JIP 32/5–6: 423–48.
Houben, Jan. 1999. To Kill or Not to Kill the Sacrificial Animal (yajña-paśu)?
Arguments and Perspectives in Brahmanical Ethical Philosophy. In J. Houben
and R. van Kooij, eds. Violence Denied: Violence, Non-Violence and the
Rationaliation of Violence in South Asian Cultural History, 105–84. Leiden: Brill.
——— . 2000. The Ritual Pragmatics of a Vedic Hymn: The “Riddle Hymn” and the
Pravargya Ritual. JAOS 120/4: 499–536.
Hudson, D. Dennis. 1994. Kṛṣṇa’s Maṇḍala. Draft 7/6/94. Personal communication
from the author.
——— . 2001. The ‘Barley-Corn’ Pattern of Bhagavad-Gītā 12–16. JVS 9/2: 181–95.
——— . 2010. Krishna’s Mandala: Bhagavata Religion and Beyond. In John Stratton
Hawley, ed. New Delhi: Oxford Univerity Press.
Hudson, Emily. 2006. Disorienting Dharma: Ethics and the Poetics of Suffering in
the Mahābhārata. Ph.D. Dissertation, Emory University.
——— . 2007. Listen But Do Not Grieve: Grief, Paternity, and Time in the Laments of
Dhṛtarāṣṭra. In Brodbeck and Black, eds. 2007, 35–52.
Hultzsch, E. 1969. Inscriptions of Asoka. New Edition. Corpus Inscriptionum
Indicarum, Vol. 1. Delhi: Indological Book House.
Huxley, Andrew. 1990. How Buddhist Is Theravāda Buddhist Law? In T. Skorupski,
ed. The Buddhist Forum, Vol. 1, London, 41–85.
——— . 1995. Buddhism and Law—The View from Mandalay. JIABS, issue on
Buddhism and Law, 18/1: 47–96.
——— . 1996a. The Buddha and the Social Contract. JIP 24: 407–20.
——— . 1996b. When Manu Met Mahāsammata. JIP 24: 593–621.
Inden, Ronald. 2006. Text and Practice: Essays on South Asian History. Introd. by Daud
Ali. New Delhi: Oxford.
Jamison, Stephanie. 1991. The Ravenous Hyenas and the Wounded Sun: Myth and Ritual
in Ancient India. Myth and Poetics Series. Ithaca: Cornell University Press.
 bibliography 705

——— . 1996. Sacrificed Wife Sacrificer’s Wife: Women, Ritual, and Hospitality in
Ancient India. Oxford: Oxford University Press.
——— . 1998. Rhinoceros Toes, Manu V.17–18, and the Development of the Dharma
System. JAOS 118/2: 249–56.
——— . 2000. A Sanskrit Maxim and Its Ritual and Legal Applications. In A. Hintze
and E. Tichy, eds. Anusantatyai: Festschrift für Johanna Narten zum 70 Geburtstag,
111–25. Beiheft: Münchener Studien zur Sprachwissenschaft, 19.
——— . 2001. The Rigvedic Svayaṃvara? Formulaic Evidence. In Karttunen and
Koskikallio, eds. 2001, 303–15.
——— . 2006. Women “Between the Empires” and “Between the Lines”. In Olivelle,
ed. 2006, 191–214.
——— . 2007. The Rig Veda Between Two Worlds. Collège de France, Publications de
l’Institut de Civilisation Indienne, Fascicule 74. Paris: Diffusion de Boccard.
——— . 2010. The Secret Life of Texts. Presidential Address, American Oriental Society.
St. Louis, MO (to appear in a forthcoming issue of JAOS).
Jatavallabhula, D. F. 1999. Raṇayajña: The Mahābhārata War as a Sacirfice. In
J. Houben and R. van Kooij, eds. Violence Denied: Violence, Non-Violence and the
Rationaliation of Violence in South Asian Cultural History, 69–103. Leiden: Brill.
Jayaswal, K. P. 1930. Manu and Yājñavalkya: A Comparison and a Contrast, a Treatise on
the Basic Hindu Law. Calcuta: Butterworth.
Jaworski, J. 1929–30. Le section de la nourriture dans le vinaya des mahīśāsaka. RO 7.
Jha, V. N. 1970. Varṇasaṃkara in the Dharma Sūtras: Theory and Practice. Journal of
the Economic and Social History of the Orient 13: 273–88.
——— . 1974. From Tribe to Untouchable: The Case of the Niṣādas. In R. S. Sharma
and V. N. Jha, eds. Indian Society: Historical Probings in Memory of D. D. Kosambi,
67–84. New Delhi: People’s Publishing House.
Johnson, William J., trans. 1998. The Sauptikaparvan of the Mahābhārata: The Massacre
at Night. Oxford: Oxford University Press.
——— , trans. 2005. Mahābhārata Book Three: The Forest. Volume Four. Clay Sanskrit
Library. New York: New York University Press and JJC Foundation.
Johnson, Willard L. 1980. Poetry and Speculation of the Ṛg Veda. Berkeley: University
of California Press.
Johnston, E. H. 1928. The Saundarananda of Aśvaghoṣa (Sanskrit text), The
Saundarananda, or Nanda the Fair (translation). London: Humphrey Milford,
Oxford University Press, 1932.
——— . 2004. Aśvaghoṣa’s Buddhacarita or Acts of the Buddha. Part 1, Sanskrit Text,
Sargas 1–14; Part 2, Introduction and Translation; Part 3, Translation of Cantos
15–28 from Tibetan and Chinese versions. First published Lahore 1936. Delhi:
Motilal Banarsidass.
Kakar, Sudhir. 1982. Shamans, Mystics, and Doctors: A Psychoanalytic Inquiry into India
and its Healing Traditions. Chicago: University of Chicago Press.
Kalbisingh, Chatsumara. 1998. The Bhikkunī Pātimokkha of the Six Schools Translated
into English. Delhi: Sri Satguru Publishers, Indian Books Centre.
Kane, Pandurang Vaman. 1962–75. History of Dharmaśāstra (Ancient and Medieval Religious
and Civil Law in India), 5 vols. Poona: Bhandarkar Oriental Research institute.
706 bibliography

Kangle, R. P. 1960–61. Bhāradvāja: An Ancient Teacher of Political Science. Bharatiya

Vidya 20–21: 333–9.
——— . 1964. Manu and Kauilya. Indian Antiquary, 3rd ser./1: 48–54.
——— . [1972] 2003. Kauṭilya’s Arthaśāstra. Part 2: English Translation with Explanatory
Notes. Delhi: Motilal Banarsidass.
Karttunen, Klaus. 1997. India and the Hellenistic World. Studia Orentalia 83. Helsinki:
Finnish Oriental Society.
——— and Petteri Koskikallio, eds. 2001. Vidyārṇavavandanam: Essays in Honour
of Asko Parpola. Studia Orientalia 94. Helsinki: Finnish Oriental Society.
Karunadasa, Y. 1996. The Dhamma Theory: Philosophical Cornerstone of the
Abhidhamma. The Wheel Publication 412/413. Kandy: Buddhist Publication
Society.
Karve, Iravati. 1974. Yuganta: The End of an Epoch. New Delhi: Sangam Press.
Katre, S. L. 1943. Dharmopaniṣad in Mahābhārata. Journal of the Ganganath Jha
Research Institute 1: 182–22.
Khan, Dominique-Sila. 2005. Reimagining the Buddha. JIP 33: 321–42.
Kinjawadekar, Ramachandrashāstri, ed. 1929–1936. Mahābhāratam with the
Commentary of Nīlakaṇṭḥa. 7 Vols. Vols. 1–6, Mahābhārata; Vol. 7,
Harivanshaparvan. Poona: Chitrashala Press.
Kinsley, David. 1975. The Sword and the Flute: Kālī and Kṛṣṇa, Dark VIsions of the Terrible
and the Sublime in Hindu Mythology. Berkeley: University of California Press.
——— . 1997. Tantric Visions of the Divine Feminine: The Ten Mahāvidyās. Berkeley:
University of California Press.
Kirfel, Willibald. 1927. Das Purāṇa Pañcalakṣaṇa: Versuch einer Textgeschicihte. Bonn:
Schroeder.
Klaes, Norbert. 1975. Conscience and Consciousness: Ethical Problems of the Mahābhārata.
Bangalore: Dharmaram College.
Kloetzli, W. Randolph. 1983. Buddhist Cosmology (From Single World System to Pure Land:
Science and Theology in the Images of Motion and Light). Delhi: Motilal Banarsidass.
——— . 2007. Nous and Nirvāṇa: Conversations with Plotinus—An Essay in Buddhist
Cosmology. Philosophy East and West 57/2: 140–77.
——— . 2010. Ptolemy and Purāṇa: Gods Born as Men. JIP 38: 583–623.
Kolff, Dirk H. A. 1990. Naukar, Rajput, and Sepoy: The Ethnohistory of the Military
Labour Market in Hindustan, 1450–1850. Cambridge: Cambridge University Press.
Korom, Frank J. 2004. The Bengali Dharmarāj in Text and Context: Some Parallels. In
Olivelle, ed. 2004, 843–70.
Kosambi, D. D. 1946. Early Stages of the Caste System in Northern India. Journal of
the Bombay Branch of the Royal Asiatic Society 22: 33–48.
——— . 1952. Ancient Kosala and Magadha. Journal of the Bombay Branch of the Royal
Asiatic Society 27: 183.
——— . 1956. An Introduction to the Study of Ancient Indian History. Bombay: Popular
Book Depot.
——— . 1964. The Autochthonous Element in the Mahābhārata. JAOS 84: 31–44.
Koskikallio, Petteri. 1994. When Time Turns: Yugas, Ideologies, Sacrifices. Studia
Orientalia 73: 253–71.
 bibliography 707

——— . 1999. The Horse Sacrifice in the Pātālakhaṇḍa of the Padmapurāṇa. In

Brockington and Schriener eds. 1999, 227–43.
——— , ed. 2005. Epics, Khilas, and Purāṇas: Continuities and Ruptures. Proceedings of
the Third Dubrovnik International Conference on the Sanskrit Epics and
Purāṇas, September 2002. Zagreb: Croatian Academy of Sciences and Arts.
——— , ed. 2009. Parallels and Comparisons. Proceedings of the Fourth Dubrovnik
International Conference on the Sanskrit Epics and Purāṇas, September 2005.
Zagreb: Croatian Academy of Sciences and Arts.
Krishnan, Y. 1989. The Meaning of the Puruṣārthas in the Mahābhārata. In Matilal,
ed. 1989, 53–68.
Kuiper, F. B. J. 1983. Ancient Indian Cosmogony. John Irwin, ed. New Delhi: Vikas
Publishing House.
Kulkarni, Sunil. 2009. When Śloka Speaks to Śloka: A Text-Centric Conspectus
Studies Hinduism’s Pluralistic Ways, Ignores Living Reality. Review of Doniger
2009a. Outlook India. http://www.outlookindia.com/article.aspx?262452
Kulke, Hermann. 1992. The Rājasūya: A Paradigm of Early State Formation. In A. W.
van den Hoek, D. H. A. Kolff, and M. S. Oort, eds. Ritual, State and History in
South Asia, 188–98. Leiden: Brill.
Lacan, Jacques. 1977. Écrits: A Selection. Alan Sheridan, trans. New York: W. W.
Norton & Company.
Laine, J. W. 1981. The Creation Account of Manusmṛti. ABORI 62: 157–68.
——— . 1991. Out of Character: Marginal Voices and Role-Transcendence in the
Mahābhārata’s Book of the Forest. JIP 19: 273–96.
Lamotte, Étienne. 1988. History of Indian Buddhism: From the Origins to the Śaka Era.
Sara Webb-Boin, trans., under the supervision of Jean Dantinne. Publications de
l’Institut Orientaliste de Louvain, 36. Louvain-la-neuve: Institut Orientaliste.
Lang, K. 1992. Āryadeva and Candrakīrti on the Dharma of Kings. Asiatische Studien
46/1: 232–43.
Lariviere, Richard W. ed. and trans. 2003. The Nāradasmṛti, 2nd ed. Delhi: Motilal
Banarsidass.
——— . 2004. Dharmaśāstra, Custom, ‘Real Law’ and ‘Aprocryphal Smṛtis’. JIP
32/5–6: 611–27.
Larson, Gerald James. 1972. The Trimūrti of Dharma in Indian Thought: Paradox and
Contradiction. Philosophy East and West 22/2: 145–53.
——— . 1995. India’s Agony over Religion. Albany: SUNY Press.
——— , ed. 2001. Religion and Personal Law in Secular India: A Call to Judgment.
Bloomington: Indiana University Press.
——— and Ram Shankar Bhattacharya, eds. 1987. Sāṃkhya: A Dualist Tradition in
Indian Philosophy. Encyclopedia of Indian Philosophies. Princeton: Princeton
University Press.
Lath, Mukund. 1990. The Concept of Ānṛśaṃsya in the Mahābhārata. In R. N.
Dandekar, ed. The Mahābhārata Revisited, 113–19. New Delhi: Sahitya
Akademi.
Lefeber, Roslalind, trans. 1994. The Rāmāyaṇa of Vālmīki, Vol. 4: Kiṣkindhā-kāṇḍa.
Princeton: Princeton University Press.
708 bibliography

Leslie, Julia, ed. 1989. The Perfect Wife: The Orthodox Hindu Woman according to the
Strīdharmapaddhati of Tryambakayajvan. Oxford University South Asian Studies
Series. Delhi: Oxford University Press.
——— . 1991. Problems of Dharma: Rules and Remedies in Classical Indian Law. Panels
of the 7th World Sanskrit Conference, Johannes Bronkhorst, ed., Vol. 9. Leiden: Brill.
Lévi, Sylvain. 1918–1920. Tato Jayam Udirayet. L. G. Khare, trans. ABORI 1: 13–20.
——— . 1929. The Mahabharata for the first time critically edited by Vishnu S.
Sukthankar. Journal Asiatique 215: 345–48.
——— . 1934. The Mahabharata for the first time critically edited by Vishnu S.
Sukthankar. Journal Asiatique 225: 281–83.
Lindquist, Steven. 2008. Gender at Janaka’s Court: Women in the Bṛhadāraṇyaka
Upaniṣad Reconsidered. JIP 36: 405–26.
Lingat, R. 1962. Time and Dharma (on Manu I, 85–86). CIS 6: 7–16.
——— . 1967. Encore “ayāya saṃbodhiṃ.” Journal Asiatique 255: 195–8.
——— . 1973. The Classical Law of India. J. Duncan M. Derrett, trans. Berkeley:
University of California Press.
Loomcharoen, Nareerat. 2007. An Analytical Study of Phra Dhammasingaburācariya’s
(CaranṬhitadhammo) Method of Teaching the Law of Karma. MA thesis,
Mahachulalongkornvidyalaya University, Bangkok.
Lopez, Donald S., Jr. 1993. The Institution of Fiction in Mahāyāna Buddhism. In
Schlomo Biderman and Ben-Ami Scharfstein, eds., Myths and Fictions, 355–88.
Leiden: E. J. Brill.
Lubin, Timothy. 2001. Vratá Divine and Human in the Early Veda. JAOS 121/4: 565–79.
——— . 2005. The Transmission, Patronage, and Prestige of Brahmanical Piety from
the Mauryas to the Guptas. In Federico Squarcini, ed. Boundary Dynamics and
Construction of Traditions in South Asia, 77–103. Florence: University of Florence
Press and Munshiram Manoharlal.
Lubotsky, Alexander. 1997. A Ṛgvedic Word Concordance, 2 vols. New Haven: American
Oriental Society.
Lüders, Heinrich. 1929. Review of the Mahābhārata’s Critical Edition, vol. 1. Deutsche
Literaturzeitung 24: 1138–46.
Macdonell, Arthur Anthony. [1898] 1974. Vedic Mythology. Delhi: Motilal Banarsidass.
——— and Arthur Berriedale Keith. [1912] 1967. Vedic Index of Names and Subjects, 2
vols. Delhi: Motilal Banarsidass.
Madan, T. N. 1982. Way of Life: King, Householder, Renouncer. Essays in Honour of Louis
Dumont. New Delhi: Vikas Publishing House.
Mahadevan, Thennilapuram P. 2008. On the Southern Reccension of the
Mahābhārata, Brahman Migrations, and Brāhmī Paleography. EJVS 15/2: 1–143.
——— 2010. The Southern Recension of the Mahābhārata: The Harivaṃśa and Āḻvār
Vaiṣṇavism. In Adluri, ed. 2010a, 87–128.
——— 2009. The Critical Edition of the Mahābhārata—Its Constitution,
Achievements, and Limitations. In Chakravarty, ed. 2009, 3–24.
Mahoney, William. 1998. The Artful Universe. Albany: SUNY Press.
Malalasekera, G. P. (1937) 1983. Dictionary of Pāli Proper Names, 2 vols. Oriental
Reprint. New Delhi: Motilal Banarsidass.
 bibliography 709

Malamoud, Charles. 1982. On the Rhetoric and Semantics of the Puruṣārthas. In

Madan, ed. 1982, 32–54.
——— . 1996. Cooking the World: Ritual and Thought in Ancient India. Trans. David
White. Delhi: Oxford University Press (= Cuire le monde. Rite et pensée dans l’Inde
ancienne. Paris: Éditions la découverte, 1989).
Malinar, Angelika. 1997. Nārāyaṇa und Kṛṣṇa: Aspekte der Gotteslehre des
Nārāyaṇīya im Vergleich zur Bhagavadgītā. In Screiner, ed. 1997, 241–95.
——— . 2007a. The Bhagavadgītā: Doctrines and Contexts. Cambridge: Cambridge
University Press.
——— . 2007b. Arguments of a Queen: Draupadī’s Views on Kingship. In Brodbeck
and Black, eds. 2007, 79–96.
Mangels, Annette. 1994. Zur Erzähltechnik im Mahābhārata. Hamburg: Verlag
Dr. Kovač.
Manné, Joy. 1990. Categories of Sutta in the Pāli Nikāyas and Their Implications for
Our Appreciation of the Buddhist Teaching and Literature. JPTS 15: 29–87.
Masson, J. L. 1975. Fratricide among the Monkeys: Psychoanalytic Observations on an
Episode in the Vālmīkirāmāyaṇa. JAOS 95: 672–78.
Matilal, Bimal Krishna. 1989. Moral Dilemmas: Insights from the Indian Epics. 1989.
In Matilal, ed. Moral Dilemmas in the Mahābhārata, 1–19. Delhi: Motilal
Banarsidass.
——— . 2002. Ethics and Epics: The Collected Essays of Bimal Krishna Matilal. Jonardon
Ganeri, ed. New Delhi: Oxford University Press.
Matthews, Bruce. 1983. Craving and Salvation: A Study in Buddhist Soteriology.
Waterloo, Ontario, Canada: Wilfrid Laurier Press.
McGrath, Kevin. 2004. The Sanskrit Hero: Karṇa in Epic Mahābhārata. Leiden:
E. J. Brill.
Mehendale, M. A. 1995. Reflections on the Mahābhārata War. Shimla: Institute of
Advanced Study.
Meiland, Justin. 2005. Mahābhārata IX: Śalya (vol. 1). Clay Sanskrit Library. New York:
New York University Press.
Menski, Werner F. [2003] 2005. Hindu Law: Beyond Tradition and Modernity. Oxford:
Oxford University Press.
Merton, Robert K. 1957. Social Theory and Social Structure. Glencoe: Free Press.
Miller, Jeannine. 1985. The Vision of the Cosmic Order in the Vedas. London: Routledge
and Kegan Paul.
Minkowski, Chirstopher Z. 1989. Janamejaya’s Sattra and Ritual Structure. JAOS
109/3: 401–20.
Mitchiner, John E. 1982. Traditions of the Seven Rishis. Delhi: Motilal Banarsidass.
——— . 1986. The Yuga Purāṇa: Critically Edited with an English Translation and a
Detailed Introduction. Bibliotheca Indica, Calcutta: The Asiatic Society.
——— . 1990. The Yuga Purāṇa: A Footnote. Bulletin of the School of Oriental and
African Studies 53/2: 320–2.
——— . 2002. The Yuga Purāṇa: Critically Edited with an English Translation and a
Detailed Introduction. Revised ed. Kolkata: The Asiatic Society.
Mittal, Sushil and Gene Thursby, eds. 2004. The Hindu World. London: Routledge.
710 bibliography

Monier-Williams, Monier. [1899] 1964. A Sanskrit-English Dictionary. Oxford:

Clarendon Press.
Monius, Anne E. 2005. Origins of Hindu Ethics. In William Schweiker, ed. The
Blackwell Companion to Religious Ethics, 330–40. Malden, MA: Blackwell.
Nabokov, Isabel. 2000. Religion against the Self: An Anthology of Tamil Rituals. Oxford:
OUP.
Nakamura, Hajime. 2000. Gotama Buddha: A Biography Based on the Most Reliable
Texts. Gaynor Sekimori trans. Tokyo: Kosei Publishing Co.
Ñāṇamoli, Bhikkhu. 1972. The Life of the Buddha According to the Pali Canon. Kandy,
Sri Lanka: Buddhist Publication Society.
——— and Bhikkhu Bodhi, trans. 2005. The Middle Length Discourses of the Buddha: A
Translation of the Majjhima Nikāya. The Teachings of the Buddha. 3rd. ed. Boston:
Wisdom Publications.
Nanavati, Rajendra I. 1982. Secondary Tales of the Two Great Epics. L.D. Series 88.
Ahmedabad: L.D. Institute of Indology.
Narang, Satya Pal, ed. 1995. Modern Evaluation of the Mahābhārata. Prof. R. K. Sharma
Felicitation Volume. Delhi: Nag Publishers.
Nattier, Jan. 1991. Once Upon a Future Time: Studies in a Buddhist Prophecy of Decline.
Berkeley: Asian Humanities Press.
——— . 2000. The Realm of Akṣobhya: A Missing Piece in the History of Pure Land
Buddhism. JIABS 23/1: 71–102.
——— . 2003. The Indian Roots of Pure Land Buddhism: Insights from the Oldest
Chinese Version of the Larger Sukhāvatīvyūha. Pacific World. Journal of the
Institute of Buddhist Studies. Third Series, No. 5, 175–201.
——— . [2003] 2005. A Few Good Men: The Bodhisattva Path according to The Inquiry of
Ugra (Ugraparipṛcchā). Honolulu: University of Hawaii Press.
——— . 2008a. Buddhist Eschatology. In Jerry L. Walls, ed. The Oxford Handbook of
Eschatology, 151–69. Oxford: Oxford University Press.
——— 2008b. E-mail April 26.
——— . 2008c. E-mail July 2.
Neelis, Jason. 2002. La Vielle Route Reconsidered: Alternative Paths for Early
Transmission of Buddhism Beyond the Borderlands of South Asia. Bulletin of the
Asia Institute n.s. 16: 143–64.
——— . 2007. Passages to India: Śaka and Kuṣāṇa Migrations in Historical Context.
In Srinivasan, ed. 2007, 55–94.
——— . 2008a. Review of Bronkhorst 2007. JRAS Series 3, 18/3: 381–3.
——— . 2008b. As the Dharmacakra Turns: Historical Memory in Buddhist
Narratives of the Past. “Thinking, Recording, and Writing History in the Ancient
World”. Ancient Studies conference in Comparative History, Brown University,
December 12–14. Courtesy of the author.
——— . 2009. Saka Religious and Literary Patronage Networks in Early South Asian
History. Presentation draft for American Oriental Society meeting, Albuquerque,
NM, 3/16/09. Courtesy of the Author.
Nikam, N. A. and Richard McKeon, ed. and trans. [1959] 1978. The Edicts of Asoka.
Chicaigo: University of Chicago Press.
 bibliography 711

Norman, K. R. 1958. Aśokan silā-thaṃbha-s and dhaṃma-thaṃbha-s. In Samaresh

Bandopadhyay, ed., Ācarya-vandanā: D. R. Bandharkar Birth Centenary Volume,
311–18. Calcutta: University of Calcutta. (Collected Papers II, 224–32).
——— . 1975. Aśoka and Capital Punishment. JRAS: 16–24. (Collected Papers II,
200–13).
——— . 1980. Four Etymologies from the Sabhiya-Sutta. In: Buddhist Studies in
Honour of Walpola Rahula. London: Mimamsa.
——— . 1981. A Note on Attā in the Alagaddupama Sutta. In: Studies in Indian
Philosophy: A Memorial Volume in Honour of Pt Sukhlaji Sanghvi, 19–29.
Ahmedabad: LD Institute of Indology.
——— . 1983. Pāli Literature, Including the Canonical Literature in Prakrit and Sanskrit
of all the Hīnayāna Schools of Buddhism. Wiesbaden: Harrassowitz.
——— . 1992. The Group of Discourses (Sutta Nipāta): Volume 2, Revised Translation.
Oxford: PTS.
Nyanaponika Thera. [1962] 1979. The Heart of Buddhist Meditation. New York: Samuel
Weiser.
——— . 1998. Abhidhamma Studies: Buddhist Explorations of Consciousness and Time,
4th ed., revised and enlarged. Boston: Wisdom.
——— and Bhikkhu Bodhi, trans. and eds. 1999. Numerical Discourses of the Buddha:
An Anthology of Suttas from the Aṅguttara Nikāya. Walnut Creek, CA: Altamira
Press, 1999.
Nyanatiloka Bhikkhu, ed. 1972. Buddhist Dictionary: Manual of Buddhist Terms and
Doctrines, 3rd ed. Columbo: Frewin.
Oberlies, Thomas. 1997. Die Textgeschichte der Śvetadvīpa-Episode des Nārāyaṇīya
(Mahābhārata 12.321–26), chapter 3; Die Textgeschichte des Adhyāya 326, chapter
4; Gottesnamen, Opferkonzeption und Zeithorizonte im Teil A des Nārāyaṇīya,
chapter 5. In Schriener ed., 1997, 75–118; 119–38; 139–57.
——— . [1988] 1999. Die Religion des Ṛgveda. 2 vols. Komposisitionsanalyse der Soma-
Hymnen des Ṛgveda. Vienna: De Nobili.
——— . 1998. Die Ratschläge der Sehers Nārada: Ritual an und unter der Oberfläche
des Mahābhārata. In Hildegard L. C. Tristram, ed. New Methods in the Research of
Epic = Neue Methoden der Epenforschung, 125–41. Tübingen: Gunter Narr Verlag.
——— . 2009. Gandharvas and Apsarases. In Brill’s Encyclopedia of Hinduism, Vol. 1.
565–70. Leiden: Brill.
Obeyesekere, Gananath. 1984. The Cult of the Goddess Pattini. Chicago: University of
Chicago Press.
——— . 1990. The Work of Culture: Symbolic Transformation in Psychoanalysis and
Anthropology. Chicago: University of Chicago Press.
O’Flaherty, Wendy Doniger. 1973. Asceticism and Eroticism in the Mythology of Śiva.
London: School of Oriental and African Studies.
——— . 1976. The Origins of Evil in Hindu Mythology. Berkeley: University of
California Press.
——— , trans. 1981. The Rig Veda. London: Penguin.
——— and J. Duncan M. Derrett, eds., 1978. The Concept of Duty in South Asia.
New Delhi: Vikas.
712 bibliography

Oguibenine, B. 1983. From a Vedic Ritual to the Buddhist Practice of Initiation into
the Doctrine. In P. Denwood and A. Piatagorsky, eds. Buddhist Studies Ancient and
Modern, 107–23. London: Curzon Press.
Ohnuma, Reiko. 2005. Gift. In Donald S. Lopez, Jr., ed. Critical Terms for the Study of
Buddhism, 103–23. Chicago: University of Chicago Press.
——— . 2006. Debt to the Mother: A Neglected Aspect of the Founding of the
Buddhist Nuns’ Order. JAAR 74/4: 861–901.
Oldenberg, Hermann. 1882. Buddha, His Life, His Doctrine, His Order. Trans. W. Hoey.
Calcutta: Book Company.
Olivelle, Patrick. 1974. The Origin and Early Development of Buddhist Monachism.
Colombo: Gunasena.
——— . 1981. Contributions to the Semantic History of Saṃnyāsa. JAOS 101: 265–74.
——— . 1984. Renouncer and Renunciation in the Dharmaśāstras. In Richard W.
Lariviere, ed. Studies in Dharmaśāstra, 81–152. Calcutta: Firma KLM.
——— . 1987. King and Ascetic: State Control of Asceticism in the Arthaśāstra. In
Festschrift for Ludo Rocher. Adyar Library Bulletin 51: 39–59.
——— . 1991. From Feast to Fast: Food and the Indian Ascetic. In Julia Leslie, ed.
Problems of Dharma: Rules and Remedies in Classical Indian Law, 17–36. Panels
of the 7th World Sanskrit Conference, ed. Johannes Bronkhorst, vol. 9.
Leiden: Brill.
——— . 1992. Saṃnyāsa Upaniṣads: Hindu Scriptures on Asceticism and Renunciation.
New York: Oxford University Press.
——— . 1993. The Āśrama System: The History and Hermeneutics of a Religious
Institution. New York: Oxford University Press.
——— . 1996a. Upaniṣads. World Classics. Oxford: Oxford University Press.
——— . 1996b. Dharmaskandhāḥ and Brahmasaṃsthaḥ: A Study of Chāndogya
Upaniṣad 2.23.1. JAOS 116: 205–19.
——— . 1998. The Early Upaniṣads: Annotated Text and Translation. New York: Oxford
University Press.
——— . trans. 1999. Dharmasūtras. The Law Codes of Ancient India. Oxford: Oxford
University Press.
——— . trans. 2003. Dharmasūtras: The Law Codes of Āpastamba, Gautama,
Baudhāyana, and Vasiṣṭha. Annotated Text and Translation. Delhi: Motilal
Banarsidass.
——— . 2004a. The Semantic History of Dharma The Middle and Late Vedic Periods.
JIP 32/5–6: 491–511.
——— . trans. 2004b. The Law Code of Manu. Oxford: Oxford University Press.
——— . trans. and ed. 2005a. Manu’s Code of Law: A Critical Edition and Translation of
the Mānava-Dharmaśāstra. Oxford: Oxford University Press.
——— . 2005b. Language, Texts, and Society: Explorations in Ancient Indian Culture and
Religion. Florence: University of Florence Press and Munshiram Manoharlal.
——— . ed. and trans. 2005c. Dharmasūtra Parallels: Containing the Dharmasūtras of
Ᾱpastamba, Gautama, Baudhāyana, and Vasiṣṭha. Delhi: Motilal Banarsidass.
——— . ed. 2006a. Between the Empires: Society in India 300 BCE to 400 CE. New York:
Oxford.
 bibliography 713

——— . 2006b. Explorations in the Early History of the Dharmaśāstra. In Olivelle, ed.
2006a, 169–90.
——— . trans. 2008. Life of the Buddha by Aśvaghoṣa. Clay Sanskrit Library. New York:
NYU Press & JJC Foundation.
——— , ed. 2009. Dharma: Studies in its Semantic, Cultural and Religious History. First
Enlarged Indian Edition. Reprinted from JIP 32, 5–6. Delhi: Motilal Banarsidass.
——— . 2010. The Brāhmaṇa and the Śūdra: Brahmanical Exceptionalism and
Resistance in the Legal Texts. American Academy of Religion Annual Meeting,
Atlanta. Courtesy of the author.
Oliver, Kelly. 1993. Reading Kristeva: Unraveling the Double-bind. Bloomington:
University of Indiana Press.
Pagel, Ulrich. 1995. The Bodhisattvapitaka: Its Doctrines, Practices, and their Position in
the Mahayana Literature. Tring, England: Institute of Buddhist Studies.
Pande, G. C. 1957. Studies in the Origins of Buddhism. Allahabad: University of Allahabad.
——— . 1978. Śramaṇa Tradition: Its History and Contribution to Indian Culture.
Ahmedabad: L. D. Institute of Indology.
Parasher, A. 1991. Mlecchas in Early India: A Study of Attitudes toward Outsiders up to
AD 600. Delhi: Munshiram Manoharlal.
Parasher-Sen, Aloka. 2006. Naming and Social Exclusion: The Outcaste and the
Outsider. In Olivelle, ed. 2006, 415–55.
Pathak, Shubha. 2006. Why Do Displaced Kings Become Poets in the Sanskrit Epics?
Modelling Dharma in the Affirmative Rāmāyaṇa and the Interrogative
Mahābhārata. IJHS 10: 127–49.
Patton, Laurie L. 2005a. Bringing the Gods to Mind: Mantra and Ritual in Early Indian
Sacrifice. Berkeley: University of California Press.
——— . 2005b. Trita’s Tumble and Agastya’s Ancestors: On the Narrative
Construction of Dharma. In Federico Squarcini, ed. Boundary Dynamics and
Construction of Traditions in South Asia, 133–57. Florence: University of Florence
Press and Munshiram Manoharlal.
——— . 2007. How Do You Conduct Yourself ? Conduct and the Construction of a
Dialogical Self in the Mahābhārata. In Brodbeck and Black, eds. 2007, 97–109.
Pingree, David. 1963. Astronomy and Astrology in India and Iran. Isis 54/2: 229–46.
——— . ed., trans, and commentator. 1978. The Yavanajātaka of Sphujidhvaja, 2 vols.
Harvard Oriental Series 48. Cambridge, MA: Harvard University Press.
——— . 1990. The Purāṇas and Jyotiḥśāstra: Astronomy. JAOS 110/2: 274–80.
Pollock, Sheldon. 1984. Ātmānam mānuṣam manye: Dharmākūtam on the Divinity of
Rāma. Journal of the Oriental Institute of Baroda 33: 505–28.
——— , trans. 1986. The Rāmāyaṇa of Vālmīki: An Epic of Ancient India, vol. 2:
Ayodhyākāṇḍa. Princeton: Princeton University Press.
——— . 1989a. The Idea of Śāstra in Traditional India. In A. L. Dallapicola and
S. Zingel-Avé Lallemant, eds. The Sastric Tradition in the Indian Arts. Beiträge zur
Südasienforschung, 17–26. Wiesbaden: Steiner.
——— . 1989b. Playing by the Rules: Śāstra and Sanskrit Literature. In A. L.
Dallapicola and S. Zingel-Avé Lallemant, eds. The Sastric Tradition in the Indian
Arts. Beiträge zur Südasienforschung, 301–12. Wiesbaden: Steiner.
714 bibliography

——— . 1990. From Discourse of Ritual to Discourse of Power in Sanskrit Culture.

Journal of Ritual Studies 4/2: 315–45.
——— , trans. 1991. The Rāmāyaṇa of Vālmīki: An Epic of Ancient India, vol. 3:
Araṇyakāṇḍa. Princeton: Princeton University Press.
——— . 1997. The ‘Revelation’ of ‘Tradition’: Śruti, Smṛti and the Sanskrit Discourse
of Power. In Siegfried Lienhard and Irma Piovano, eds. Lex et Litterae: Studies in
Honour of Professor Oscar Botto. Turin: Edizioni dell’Orso.
——— . 2006. The Language of the Gods in the World of Men: Sanskrit, Culture, and
Power in Premodern India. New Delhi: Permanent Black.
Poppe, Nicholas. 1967. The Twelve Deeds of Buddha: A Mongolian Version of the
Lalitavistara. Seattle: University of Washington Press.
Prebish, Charles S. 1975. Buddhist Monastic Discipline: The Sanskrit Prātimokṣa Sūtras
of the Mahāsaṃghikas and Mūlasarvāstivādins. University Park, PA: Pennsylvania
State University Press.
——— . 1994. A Survey of Vinaya Literature. Taipei: Jade Scepter Imprint, Jin Luen
Publishing House.
——— . 1996. Śaikṣa Dharmas Revisited: Further Considerations of Mahāsaṅghika
Origins. HR 35/3: 258–70.
——— and Janice J. Nattier. 1977. Mahāsaṅghika Origins: The Beginnings of Buddhist
Sectarianism. HR 16: 237–72.
Prentiss, Karen Pechilis. 1999. The Embodiment of Bhakti. New York: Oxford
University Press.
Proudfoot, Ian. 1987. Ahiṃsā and a Mahābhārata Story: The Development of the Story of
Tulādhāra in the Mahābhārata in Connection with Non-Violence, Cow-Protection,
and Sacrifice. Canberra: Faculty of Asian Studies, Australian National University.
Purchas, Samuel. 1614. Purchas His Pilgrimage, Or, Relations of the VVorld and the Religions
Obsuerued in All Ages and Places Discouered, from the Creation unto This Present: In
Four Parts, This First Containeth a Theological and Geographical History of Asia,
Africaa, and America, with the Ijlands Adjacent…. 2nd ed., Much Enlarged. London:
William Stansbie. G159.P976 1614 Rare Books Special Collections Reading Room.
Rabault, Pascal. 2004. Le Mahābhārata dans l’Indianisme allemande: Genèse d’un
objet scientifique. In Marc Cluet, ed. La Fascination de l’Inde en Allemagne
1800–1933, 63–90. Rennes: Presses Universitaires de Rennes.
Radhakrishnan, Sarvepalli. 1962. Indian Philosophy. London: Allen and Unwin.
——— . ed. and trans. 1953. The Principal Upaniṣads. The Muirhead Library of
Philosophy. London: George Allen & Unwin.
Raghavan, V. 1956. Buddhological Texts and the Epics. Adyar Library Bulletin 20:
349–59.
Rahula, Walpole. 1974. What the Buddha Taught. Expanded second edition. New York:
Grove Press.
Ramanujan, A. K. 1989. Is there an Indian Way of Thinking? An Informal Essay. CIS
(n.s.) 23/1: 41–58.
Ray, Anita C. 1998. An Analysis of the Sāvitrī Legend in Ancient Indian Literature and
Culture. Ph.D. Dissertation, La Trobe University.
 bibliography 715

Ray, H. P. 1986. Monastery and Guild. Oxford: Oxford University Press.

——— . 1988. The Yavana Presence in Ancient India. Journal of the Economic and
Social History of the Orient 31: 311–25.
Ray, Pratibha. 1995. Yajnaseni: The story of Draupadī. Trans. Pradip Bhattacharya.
Calcutta: Rupa and Co.
Reddy, Sheela. 2009. Ram Was Happy with Sita . . . Indulging Her in Every Way . . . And
Then He Threw Her Out. Sheela Reddy Interviews Wendy Doniger. Outlook India,
October 26. http://www.outlookindia.com/article.aspx?262346.
Redfield, James M. 1994. Nature and Culture in the Iliad: The Tragedy of Hector.
Durham: Duke University Press.
Reich, Tamar. 1998. A Battlefield of a Text: Inner Textual Interpretataion in the
Sanksrit Mahābhārata. Ph.D. Dissertation. Chicago: University of Chicago.
Renou, Louis. 1964. Sur deux mots du Rgveda. Journal Asiatique 252: 159–67.
Reynolds, Frank. 1972. The Two Wheels of Dhamma. In B. I. Smith, ed. The Two
Wheels of Dhamma. Chambersburg: American Academy of Religion.
——— . 1995. Buddhism and Law—Preface. JIABS, issue on Buddhism and Law,
18/1: 1–6.
Rhys Davids, T. W. 1982. Vinaya Texts, 3 vols. Oxford: Oxford University Press.
——— . 1903. Buddhist India. New York: G. P. Putnam’s Sons.
——— . 1904. On the Buddhist “Middle Country.” JRAS 194, 83 sq., with Fleet’s
corrections, ibid., 1907, 657.
——— , trans. [1917–30] 1972. The Book of Kindred Sayings (Saṃyutta Nikāya), 5 Parts.
London: Pali Text Society.
——— and Hermann Oldenberg. [1881] 1968. Vinaya Texts, 3 parts. Sacred Books of
the East, vol. 13, 17, 20. Delhi: Motilal Banarsidass.
——— and William Stede. [1921–25] 2003. Pali-English Dictionary. Delhi: Motilal
Banarsidass.
Richman, Paula, ed. 1991. Many Rāmāyaṇas: The Diversity of a Narrative Tradition in
South Asia. Berkeley: University of California Press.
——— , ed. 2001. Questioning Rāmāyaṇas: A South Asian Narrative Tradition. Berkeley:
University of California Press.
Robinson, Richard, Willard Johnson, and Thanissaro Bhikkhu. 2005. Buddhist
Religions: A Historical Introduction, 5th ed. Belmont, CA: Wadsworth.
Rocher, Ludo. 1964. The Theory of Proof in Ancient Hindu Law. Recuils de la Societé
Jean Bodin, Vol. 18, 325–71. Brussels: Les Éditions de la Librarie Encyclopedique.
——— . 1980. Karma and Rebirth in the Dharmaśāstras. In Wendy Doniger
O’Flaherty, ed. Karma and Rebirth in Classical Indian Tradition, 61–89. Berkeley:
University of California Press.
——— . 1986. The Purāṇas. A History of Indian Literature, vol. 2. Epics and Sanskrit
Religious Literature. Wiesbaden: Otto Harrassowitz.
Rocher, Rosanne. 2009. New Perspectives on the History of Indian Studies in
Continental Europe. JAOS 129,4: 635–43.
Rotman, Andy. 2009. Thus Have I Seen: Visualizing Faith in Early Buddhist Literature.
New York: Oxford University Press.
716 bibliography

Ruben, Walter. 1968. Fighting against Despots in Old Indian Literature. ABORI
48–49: 111–18.
Rudolph, Lloyd I. and Susanne Hoeber Rudolph. 1967. The Modernity of Tradition:
Political Development in India. Chicago: University of Chicago Press.
Ruegg, D. S. and L. Schmithausen, eds. 1990. Earliest Buddhism and Madhyamaka:
Panels of the VIIth World Sanskrit Conference, Kern Institute, Leiden, August 23–29,
1987. Copenhagen: Brill.
Rukmani, T. S. 2005. The Mahābhārata: What Is Not Here Is Nowhere Else. New Delhi:
Munshiram Manoharlal.
Sachau, E. C. 1964. Alberuni’s India, 2 vols in 1. Delhi: S. Chand & Co.
Sadhale, G. S., ed. 1985. The Bhagavad Gītā with Eleven Commentaries, Vol. 1. Delhi:
Parimal Publications.
Salomon, Richard. 1978. The Three Cursed Rivers of the East, and Their Significance
for the Historical Geography of Ancient India. Adyar Librrary Bulletin 42: 31–60.
——— . 1998. Indian Epigraphy: A Guide to the Study of Inscriptions in Sanskrit, Prakrit,
and Other Indo-Aryan Languages. Oxford: Oxford University Press.
——— . 2005. The Indo-Greek Era of 186/5 B.C. in a Buddhist Reliquary Inscription.
In Osmund Bopearachchi and Marie-Françoise Boussac, eds. Afghanistan, Ancien
Carrefour entre l’est et l’ouest. Indicopleustoi: Archaeologies of the Indian Ocean,
Vol. 3, 378–83. Turnhout: Brepols.
——— . 2007. Dynastic and Institutional Connections in the Pre- and Early
Kuṣāṇa Period: New Manuscript and Literary Evidence. In Srinivasan, ed.
2007, 267–86.
Sastri, P. P. S. 1931. The Mahābhārata Southern Recension, Citically Edited by P. P. S.
Sastri. Vol. 1—Ādi Parvan—Part 1. Madras: V. Ramaswami Sastralu & Sons.
Sax, William. 2002. Dancing the Self: Personhood and Performance in the Pāṇḍav Līlā of
Garhwal. New York: OUP.
Scharfe, Helmut. 1993. Investigations in Kauṭalya’s Manual of Political Science.
Wiesbaden: Harrassowitz.
Scheuer, Jacques. 1982. Śiva dans le Mahābhārata. Bibliothèque de l’École des Haute
Études, Sciences Religieuses, vol. 56. Paris: Presses Universitaires de France.
Schlerath, Bernfried. 1960. Das Königtum im Rig- und Atharvaveda: Ein Beitrag zur
indogermanishen Kulturgeschichte, Wiesbaden: Harrassowitz.
Schmithausen, Lambert. 1990. Preface to Ruegg and Schmithausen 1990.
——— . 1999. Aspects of the Buddhist Attitude towards War. In J. Houben and R. van
Kooij, eds. Violence Denied: Violence, Non-Violence and the Rationaliation of Violence
in South Asian Cultural History, 45–67. Leiden: Brill.
Schopen, Gregory. 1975. The Phrase ‘sa pṛthivīpradeśaś caityabhīto bhavet’ in the
Vajrachedikā: Notes on the Cult of the Book in Mahāyāna. IIJ 27: 147–81.
——— . 1997. Bones, Stones, and Buddhist Monks: Collected Papers on the Archaeology,
Epigraphy, and Texts of Monastic Buddhism in India. Honolulu: University of
Hawaii Press.
——— . 2000. The Mahāyāna and the Middle Period in Indian Buddhism:
Through a Chinese Looking Glass. Eastern Buddhist n.s. 32/2: 1–25
(= Schopen 2005, x–xx).
 bibliography 717

——— . 2004. Buddhist Monks and Business Matters: Still More Papers on Monastic
Buddhism in India. Honolulu: University of Hawaii Press.
——— . 2005. Figments and Fragments of Mahāyāna Buddhism in India: More Collected
Papers. Honolulu: University of Hawaii Press.
——— . 2006. A Well-Sanitized Shroud: Asceticism and Institutional Values in the
Middle Period of Buddhist Monasticism. In Olivelle, ed. 2006, 315–47.
Schreiner, Peter. 1990. A/svagho.sa’s Buddhacarita: A Machine Readable
Transliteration. “Gretil” website.
——— , ed. 1997. Nārāyaṇīya-Studien. Purāṇa Research Publications—Tubingen,
Vol. 6, Wiesbaden: Harrassowitz Verlag.
——— , ed. 2005. Electronic Transliteration of the Harivaṃśa. From the Pune Critical
Edition, P L. Vaidya, ed. 1969–71. A Collaboration of Scholars from the Dubrovnik
International Conference on the Epics and Purāṇas. Department of Indology,
University of Zurich website.
Schumann, H. W. 1989. The Historical Buddha: The Times, Life and Teaching of the
Founder of Buddhism. M. O’C. Walshe, trans. London: Arkana.
Selby, Martha Ann. 2003. The Circle of the Six Seasons: A Selection from Old Tamil,
Prākrit and Sanskrit Poetry. Delhi: Penguin Books India.
Selvanayagam, Israel. 1992. Aśoka and Arjuna as Counterfigures Standing on the
Field of Dharma: A Historical-Hermeneutical Perspective. HR 32/1: 59–75.
Shah, Shalini. 2002. Sexual Politics and the Cult of the Mother Goddess in the
Mahābhārata. In Nilima Chitgopekar, ed. Invoking Goddesses: Gender Politics in
Indian Religion, 62–81. New Delhi: Har-Anand.
Shankar, S. 1994. The Thumb of Ekalavya: Postcolonial Studies and the “Third World”
Scholar in a Neocolonial World. World Literature Today 68/3: 479–87.
Sharma, Arvind. 1986. The Hindu Gītā: Ancient and Classical Interpretations of the
Bhagavad Gītā. La Salle, IL: Open Court.
——— , ed. 1991. Essays in the Mahābhārata. Leiden: Brill.
——— . 2003. Hinduism and Its Sense of History. New Delhi: Oxford
University Press.
Shulman, David Dean. 1985. The King and the Clown in South Indian Myth and Poetry.
Princeton: Princeton University Press.
——— . 1991a. Fire and Flood: The Testing of Sītā in Kampaṉ’s Irāmāvatāram. In
Richman 1991: 89–113.
——— . 1991b. Toward a Historical Poetics of the Sanskrit Epics. International Folklore
Review (1991) 11: 9–17. Also in Shulman 2001.
——— . 1996. The Yakṣa’s Questions. In Galit Hasan-Rokem and Shulman, eds.
Untying the Knot: On Riddles and their Enigmatic Modes, 151–68. New York: Oxford
University Press. Also in Shulman 2001.
——— . 2001. The Wisdom of Poets: Studies in Tamil, Telugu, and Sanskrit. Delhi:
Oxford University Press.
——— . 2009. A Passion for Hindu Myths. Review of Doniger 2009a. New York
Review of Books 56: 18 (November 19).
Siddhatissa, H. trans. [1985] 1998. The Sutta-Nipāta. London: Routledge-Curzon.
Siegel, Lee. 1987. Laughing Matters. Chicago: University of Chicago Press.
718 bibliography

Singh, Avadeesh Kumar. 2006. Ramayaṇa in the Mahabharata: A Study of

“Ramopakhyana” in the Mahābhārata. In Pradeep Trikha, ed. Textuality and
Intertextuality in the Mahābhārata, 32–8. New Delhi: Sarup & Sons.
Singh, Balbir. 1971. Dharma: Man, Religion, and Society. New York: Humanities Press.
Sinha, Braj. 1991. Arthaśāstra Categories in the Mahābhārata: From Daṇḍanīti to
Rājadharma. In Arvind Sharma, ed. Essays on the Mahābhārata, 369–83. Leiden:
E. J. Brill.
Skorupski, Tadeusz. 1987. Dharmas. In Mircea Eliade, gen. ed. Encyclopedia of
Religion. Vol. X, 332–8. New York: Macmillan Free Press.
Slaje, Walter. 2002. Water and Salt (III): An Analysis and New Translation of the
Yājñavalkya-Maitreyī Dialogue. IIJ 45: 205–20.
Smets, Sandra. 2005. The Story of Kauśika (Mbh 3,196–206): Parallel Topics with the
Jaiminīyasaṃhitā of the Brahmāṇḍapurāṇa. In Koskikallio, ed. 2005, 516–33.
Smith, Frederick M. 1994. Review of Biardeau 1989b and Hiltebeitel 1991b. IIJ 37:
253–62.
——— . 2006. The Self Possessed: Diety and Spirit Possession in South Asian Literature
and Civilization. New York: Columbia University Press.
Smith, John D. 1991. The Epic of Pābūjī: A Study, Transcription and Translation.
Cambridge: Cambridge University Press.
——— . 2009. The Mahābhārata: An Abridged Translation. London: Penguin Classics.
Smith, R. S. 1969. On the Original Meaning of Taittirīya Upaniṣad 1.11.1–4. In Dennis
Sinor, ed. American Oriental Society Middle West Branch Simi-Centennial Volume.
211.16. Bloomington: Indiana University Press.
Söhnen, Renate. [1997] 1995. On the Concept and Function of satya (“truth”) in
Ancient Indian Literature. In C Galewicz et al., eds. International Conference on
Sanskrit and Related Studies, September 23–26, 1993: Proceedings, 235–44. Cracow:
Enigma Press.
——— . 1997. Rise and Decline of the Indra Religion in the Veda. In Witzel, ed.,
1997c, 235–43.
——— . 1999. On the Composition of the Dyūtaparvan in the Mahābhārata. In
Brockington and Schreiner, eds. 1999, 139–54.
——— . 2005. Buddhist Tales in the Mahābhārata. “Epic Constructions: Gender,
Myth, and Society in the Mahābhārata” Conference, July 7–9, School of Oriental
and African Studies, London.
Soifer, Deborah A. 1991. The Myths of Narasiṃha and Vāmana: Two Avatars in
Cosmological Perspective. Albany: SUNY Press.
Sörensen, S. [1904] 1963. An Index to the Names in the Mahābhārata. Delhi: Motilal
Banarsidass.
Srinivasan, Doris Meth, ed. 2007. On the Cusp of an Era: Art in the Pre-Kuṣāṇa World.
Brill’s Inneer Asian Library 18. Leiden: Brill.
Staal, J. F. 2010. “Nambudiri Veda Recitation” and the Gonda Series. Introductory
Essay Prepared for a Forthcoming Reissue of Staal, Nambudiri Veda Recitation.
S’Gravenhage: Mouton & Co., 1961.
Stcherbatsky, Th. [1922] 1988. The Central Conception of Buddhism and the Meaning
of the Word ‘dharma’. Delhi: Motilal Banarsidass.
 bibliography 719

Stein, Otto. 1936. The Numeral 18. The Poona Orientalist 1/3: 1–37.
——— . 1937. Additional Notes on the Numeral 18. The Poona Orientalist 2/3: 164–5.
Sternbach, L. 1972. Place and Time of King’s Counsel with Ministers according to the
Dharmmaśāstra-s, Arthaśāstra-s and Subhāṣita-Saṃgraha-s. In J. Ersink and
P. Gaeffke, eds. India Maior: Congratulatory Volume Presented to Jan Gonda,
200–7. Leiden: Brill.
——— . 1974. The Mānava Dharmaśāstra I–III and the Bhaviṣya Purāṇa. Varanasi:
All-India Kashiraj Trust.
Stietencron, H. von. 1997. The Non-Existence of Impurity and the Legitimation of
Kings. In S. Lienhard and I. Piovano, eds. Lex et Litterae: Studies in Honour of
Professor Oscar Botto, 487–508. Turin: Edizioni dell’Orso.
Strauss, Otto. 1912. Ethische Probleme aus dem Mahābhārata. Firenze: Tipografia
Galileiana.
Strong, John S. 1983. The Legend of King Aśoka: A Study and Translation of the
Aśokāvadāna. Princeton: Princeton University Press.
——— . 2001. The Buddha: A Short Biography. Oxford: Oneworld.
——— . 2002. The Experience of Buddhism: Sources and Interpretations, 2nd ed.
Belmont, CA: Wadsworth.
——— . 2007. E-mail exchange discussing the Buddha’s teachings in Kuru country.
——— . 2009. The Buddha: A Beginner’s Guide. Oxford: OneWorld Publishers.
——— . 2010. The Many Lineages of the Buddha. Cardiff University Workshop:
Genealogy in South Asia, May 26–28.
Sukthankar, Vishnu S., ed. 1933. Ādiparvan. Introduction and Apparatus. In
Sukthankar et al., eds. Mahābhārata: Critical Edition, 24 vols. with Harivaṃśa.
Poona: Bhandarkar Oriental Research Institute. 1933–70, vol. 1, with
Prolegomena, i–cx.
——— , ed. 1942. Āraṇyakaparvan. Introduction. In Sukthankar et al., eds.
Mahābhārata: Critical Edition, 24 vols. with Harivaṃśa. Poona: Bhandarkar
Oriental Research Institute. 1933–70, vol. 3, i–xxviii.
Sullivan, Bruce M. 1990a. Kṛṣṇa Dvaipāyana Vyāsa and the Mahābhārata: A New
Interpretation. Leiden: Brill.
——— . 1990b. A Note on Kāśisundarī and Kṛṣṇa Dvaipāyana Vyāsa. ABORI 71:
290–2.
——— . 1995. Author and Authority in the Epic. In Narang, ed. 1995, 19–23.
Sutherland, Gail Hinich. 1991. The Disguises of the Demon. Albany: SUNY Press.
Sutherland, Sally J. 1991. The Bad Seed: Senior Wives and Elder Sons. In S. J.
Sutherland, ed. Bridging Worlds: Studies on Women in South Asia. 24–52.
Berkeley: Centers for South and Southeast Asia Studies, International and Asia
Studies.
——— . 1992. The Text Which is No Text: Critical Edition as Text. In Cornelia N.
Moore and Lucy Lower, eds. Translation East and West: A Cross-Cultural Approach.
82–92. Honolulu: University of Hawaii and the East-West Center.
Sutherland Goldman, Sally J. 2001. The Voice of Sītā in Vālmīki’s Sundarakāṇḍa. In
Paula Richman, ed. Questioning Rāmāyaṇas: A South Asian Tradition, 223–38,
390–5. Berkeley: University of California Press.
720 bibliography

——— . 2004. Gendered Narratives: Gender, Space, and Narrative Structures in

Vālmīki’s Bālakāṇḍa. In Bose, ed. 2004, 47–86.
——— . 2009. Sītā’s War: Gender and Narrative in the Yuddhakāṇḍa of Vālmīki’s
Rāmāyaṇa. In Goldman and Tokunaga 2009: 139–68.
Sutton, Nicholas. 1997. Aśoka and Yudhiṣṭhira: A Historical Setting for the Ideological
Tensions of the Mahābhārata. Religion 27: 331–41.
——— . 2000. Religious Doctrines in the Mahābhārata. Delhi: Motilal Banarsidass.
——— . 2005. A Note on the Development of Emotionaal Bhakti: Epic Shaivism in the
Mahābhārata. ABORI 86: 153–66.
——— . 2006. Text in Context in Understanding the Mahābhārata. JVS 14/2: 81–92.
Takakusu, Junjirō. [1947] 1956. The Essentials of Buddhist Philosophy. Bombay: Asia
Publishing House.
Tambiah, Stanley J. 1976. World Conqueror and World Renouncer: A Study of Buddhism
and Polity in Thailand against a Historical Background. Cambridge: CUP.
——— . 1989. King Mahāsammata: The First King in the Buddhist Story of Creation and
its Persisting Relevance. Journal of the Anthropological Society of Oxford 20: 101–22.
Thapar, Romila. 1982. (= 2005n). Householders and Renouncers in the Brahmanical
and Buddhist Traditions. In Madan, ed. 1982, 273–98.
——— . 1993. From Lineage to State: Social Formations in the Mid-First Millennium in
the Ganga Valley. Delhi: Oxford University Press.
——— . 1997. Aśoka and the Decline of the Mauryas, Rev. ed. Delhi: Oxford University
Press.
——— . (1999) 2002. Śakuntalā: Texts, Readings, Histories. London: Anthem Press.
——— . 2002. Early India: From the Origins to 1200. London: Penguin.
——— . (2000) 2005. Cultural Pasts: Essays in Early Indian History. 4th printing.
New Delhi: Oxford University Press.
——— . 2005a. Society and Historical Consciousness: The Itihāsa-purāṇa Tradition.
In Thapar, ed. (2000) 2005, 123–54.
——— . 2005b. The Oral and the Written in Early India. In Thapar, ed. (2000) 2005,
195–212.
——— . 2005c. The Image of the Barbarian in Early India. In Thapar, ed. (2000)
2005, 235–70.
——— . 2005d. The Early History of Mathurā: Up to and Including the Mauryan
Period. In Thapar (2000) 2005, 396–410.
——— . 2005e. Aśoka and Buddhism as Reflected in the Aśokan Edicts. In Thapar, ed.
(2000) 2005, 422–238
——— . 2005f. Literacy and Communication: Some Thoughts on the Inscriptions of
Aśoka. In Thapar (2000) 2005, 439–52.
——— . 2005g. Epigraphic Evidence and Some Indo-Hellenistic Contacts during the
Mauryan Period. In Thapar, ed. (2000) 2005, 452–61.
——— . 2005h. The Mauryas Revisited. In Thapar, ed. (2000) 2005, 462–517.
——— . 2005i. The Historian and the Epic. In Thapar, ed. (2000) 2005, 613–29.
——— . 2005j. Some Aspects of the Economic Data in the Mahābhārata. In Thapar,
ed. (2000) 2005, 630–46.
 bibliography 721

——— . 2005k. The Rāmāyaṇa: Theme and Variation. In Thapar, ed. (2000) 2005,
647–79.
——— . 2005l. As Long as the Moon and the Sun Endure. In Thapar, ed. (2000) 2005,
696–706.
——— . 2005m. Genealogical Patterns as Perceptions of the Past. In Thapar, ed.
(2000) 2005, 709–53.
——— . 2005n. The Householder and the Renouncer in the Brahmanical and
Buddhist Traditions. In Thapar, ed. (2000) 2005, 914–45.
Thomas, Edward J. [1927] 2000. The Life of Buddha as Legend and History. Mineola,
NY: Dover.
Thomas, Lynn. 1994. The Identity of the Destroyer in the Mahābhārata. Numen 41:
255–72.
——— . 2007. Does the Age Make the King or the King Make the Age? Exploring the
Relationship between the King and the Yugas in the Mahābhārata. RoSA 1/2:
183–201.
Thompson, George. 2008. The Bhagavad Gītā: A New Translation. New York: North
Point Press.
Tieken, Herman. 2000. Aśoka and the Buddhist Saṃgha: A Study of Aśoka’s
Schism Edict and Minor Rock Edict I. Bulletin of the School of Oriental and
African Studies 63: 1–30.
——— . 2004. The Mahābhārata after the Great Battle. Wiener Zeitschrift für die Kunde
Südasiens 48: 5–46.
——— . 2008. A Propos Three Recent Publications on the Question of the Dating of
Old Tamil Caṅkam Poetry. Asiatische Studien/Études Asiatiques 62/2: 575–605.
Tokunaga, Muneo. 1991. Machine-readable text of the Yaajnavalkya-smRti of
Yogiizvara Yaajnavalkya. With the commentary MitaakSaraa of Vijnaanezvara,
Notes, Variant readings, etc. Checked by Yasuke Ikari. Completed on March 12,
1991. Ed. by Narayan Ram Acharya. Nirnayasagara Press, Bombay, 1949.
——— . [1991] 1994. Machine-readable text of the Mahaabhaarata based on the Poona
Critical Edition. First revised version. Kyoto.
——— . 2005a. Buddhacarita and Mahābhārata: A New Perspective. 14th Conference
of the International Association of Buddhist Studies. Courtesy of the Author.
——— . 2005b. Udakakriyā and the Śāntiparvan. In Koskikallio, ed. 2005, 169–81.
——— . 2009a. Vedic Exegesis and Epic Poetry: A Note on atrāpyudāharanti. In
Koskikallio, ed. 2009, 21–30.
——— . 2009b. Bhīṣma’s Discourse as a śokāpanodana. In Robert P. Goldman and
Tokunaga, eds. 2009, 371–82.
Trenckner, V. 1924–. Critical Pāli Dictionary. Copenhagen: Munksgard.
Tsuchida, R. 1991. Two Categories of Brahmins in the Early Buddhist Period. The
Memoirs of the Toyo Bunko 49: 51–95.
——— . 1996. Versuch einer Interpretation von Chāndogya-Upaniṣad 2, 23. Studien
zur Indologie und Iranistik 20: 453–84.
——— . 2009. Janamejaya and Puṣyamitra. Presentation at the 14th World Sanskrit
Conference, Kyoto, September 3.
722 bibliography

Tubb, Gary A. 1991. Śāntarasa in the Mahābhārata. In Arvind Sharma, ed. Essays on the
Mahābhārata, 171–204. Leiden: Brill.
——— . 2002. Numbers that Matter. JVS 1/1: 147–52.
Turner, Victor. 1973. The Center out There: Pilgrim’s Goal. HR 12/3: 191–230.
Türstig, Hans-Georg. 1985. The Indian Sorcery Called abhicāra. WZKS 29: 69–117.
Vaidya, Chintaman Vinayak. 1905. The Mahābhārata: A Criticism. Bombay: A. J.
Cambridge.
——— . 1907. Epic India, or India as Described in the Mahabharata and the Ramayana.
Bombay: Mrs. Radhabai Atmaram Sagoon.
Vaidya, Parashuram Lakshman, ed. 1954. The Karṇaparvan. In Vishnu S. Sukthankar
and S. K. Belvalkar, general eds. The Mahābhārata for the First Time Critically
Edited, Vol. 10. Poona: Bhandarkar Oriental Institute.
——— . 1969–71. Harivaṃśa. Vishnu S. Sukthankar and S. K. Belvalkar, general eds.
The Mahābhārata for the First Time Critically Edited, 2 vols. Poona: Bhandarkar
Oriental Institute.
van Buitenen, J. A. B. 1967–68. Some Notes on the Uttara-yāyāta. Festschrift for
V. Raghavan. Adyar Library Bulletin 31–32: 617–35.
——— . 1972. On the Structure of the Sabhāparvan of the Mahābhārata. In J. Ensink
and P. Gaeffke, eds., India Maior: Congratulatory Volume Presented to J. Gonda,
68–84. Leiden: E. J. Brill.
——— . 1973. The Mahābhārata, Vol. I. The Book of the Beginning. Chicago: University
of Chicago Press.
——— . 1975. The Mahābhārata, vol. II: 2. The Book of the Assembly Hall; 3. The Book of
the Forest. Chicago: University of Chicago Press.
——— . 1978. The Mahābhārata, vol. III: 4. The Book ofVirāṭa; 5. The Book of the Effort.
Chicago: University of Chicago Press.
——— . 1981. The Bhagavadgītā in the Mahābhārata: A Bilingual Translation. Chicago:
University of Chicago Press.
Van der Veer, Peter. 1999. Monumental Texts: The Critical Edition of India’s National
Heritage. In Jackie Assayag, ed. The Resources of History: Tradition, Narration, and
Nation in South Asia. 113–24. Pondicherry: École Française d’Extrême-Orient.
Vanita, Ruth. 2003. The Self is Not Gendered. Sulabha’s Debate with King Janaka.
National Women’s Studies Association Journal 15/2: 76–93.
Vassilkov, Yaroslav. 1999. Kālavāda (The Doctrine of Cyclical Time) in the
Mahābhārata and
——— . 2002. Indian Practice of Pilgrimage and the Growth of the Mahābhārata in
the Light of New Epigraphical Sources. In Mary Brockington ed., 2002, 133–56.
Veidlinger, Daniel M. 2006. Spreading the Dhamma: Writing, Orality, and Textual
Transmission in Buddhist Northern Thailand. Honolulu: University of Hawai’i Press.
Vetter, Tillman. 1991. Some Remarks on Older Parts of the Suttanipāta. In D. S.
Ruegg and L. Schmithausen, eds. Earliest Buddhism and Madhyamaka, 36–56.
Leiden: Brill.
——— and Stefano Zacchetti. 2004. On Jingfa @# in Early Buddhist Translations.
Annual Report of the International Researh Institute for Advanced Buddhology at Soka
University for the Year 2003, Vol. 2, 159–66.
 bibliography 723

Vidyarnava, Rai Bahadur Srisa Chandra trans. and Wasudev Laxman Shastri
Panshikar, ed. 2003. [reprint]. Yājñavalkya Smṛti with the Commentary of
Vijñāneśvara Called The Mitākṣarā. Chowkhamba Sanskrit Studies 123. Varanasi:
Chowkhamba Sanskrit Series Office.
Visvanathan, Meera. 2010. Before Genealogy: Marking Descent in the Insctiptions
of Early Historic India. Cardiff University Workshop: Genealogy in South Asia,
May 26–28.
von Simson, Georg. 1968. The Mythic Origin of Droṇa and Karṇa. Journal of the Bihar
Research Society 14: 40–4.
——— . 1984. The Mythic Background of the Mahābhārata. IT 12: 191–223.
——— . 2007. Kṛṣṇa ad Samba: Faked Gender and Other Ambiguities on the
Background of Lunar and Solar Myth. In Brodbeck and Black, eds. 2007, 230–57.
von Stietencron, Heinrich. 2005. Hindu Myth, Hindu History: Religion, Art, and
Politics. New Delhi: Permanent Black.
Walshe, Maurice, trans. [1987] 1995. The Long Discourses of the Buddha: A Translation of
the Dīgha Nikāya. The Teachings of the Buddha. Boston: Wisdom Publications.
Warder, A. K. 1971. Dharmas and Data. JIP 1: 272–95.
Warren, Henry Clarke. 1998. Buddhism in Translations. Delhi: Motilal Banarsidass.
Watanabe, K. 1909. The Story of Kalmāṣapāda and its Evolution in Indian Literature
(A Study in the Mahābhārata and the Jātaka). JPTS, 236–310.
Watkins, Calvert. 1994a. Studies in Indo-European Legal Language, Institutions, and
Mythology. In Selected Writings, Vol. 2, 422–55. Innsbuck: Innsbrucker Beitrage
zur Sprachwissenschaft.
——— . 1994b. In the Interstices of Procedure. Ibid., 718–27.
Watson, Burton, trans. 1997. The Vimalakirti Sutra. Translations from the Asian
Classics. New York: Columbia University Press.
Weber, Max. 1958. The Religion of India: The Sociology of Hinduism and Buddhism.
Hans H. Gerth and Don Martindale trans. New York: The Free Press.
West, Emily Blanchard. 2005–6. An Indic Reflex of the Homeric Cyclopeia. Classical
Journal 101/2: 125–60.
Wezler, A. 1982. Manu’s Omniscience: On the Interpretation of Manusmṛti II.7. In
G. D. Sontheimer and P. K. Aithal, eds. Indology and Law: Studies in Honour of
Professor J. Duncan M. Derrett, 79–105. Wiesbaden: Steiner.
——— . 2001. Some Remarks on Nirukta 1.20 sākṣātkṛtadharmāṇa ṛṣayo. In
A. Michaels, ed. The Pandit: Traditional Scholarship in India. Delhi: Manohar.
Whitaker, Jarrod L. 2001. Divine Weapons and tejas in the Two Indian Epics. IIJ
43/2: 87–113.
——— . 2006. E-mail, February 6.
Wilder, Anthony. 1968. Jacques Lacan. The Language of the Self. The Function of
Language in Psychoanalysis. Baltimore: Johns Hopkins University Press.
Williams, Bernard Arthur Owen. 2002. Truth and Truthfulness: An Essay in Genealogy.
Princeton: Princeton University Press.
Williams, Paul. 1981. On the Abhidharma Ontology. JIP 9: 227–57.
——— , with Anthony Tribe. 2000. Buddhist Thought: A Complete Introduction to the
Indian Tradition. London: Routledge.
724 bibliography

——— . 2009. Mahāyāna Buddhism: The Doctrinal Foundations. 2nd ed. New York:
Routledge.
Willman-Grabowska, H. de. 1934. Evolution Semantique du mot ‘Dharma.’ RO 10:
38–50.
Wilson, H. H., trans. [1840] 1972. The Vishnu Purana: A System of Mythology and
Tradition. Calcutta: Punthi Pustak.
Winternitz, Moriz. 1933–34. The Critical Edition of the Mahābhārata: Ādiparvan.
ABORI 1933–34: 149–75.
Witzel, Michael. 1987a. The Case of the Shattered Head. Studien zur Indologie und
Iranistik 13/14: 363–416.
——— . 1987b. On the Localization of Vedic Texts and Schools (Materials on Vedic
Śakhas, 7). In Gilbert Pollet, ed., India and the Ancient World: History, Trade and
Culture Before A.D. 650, 173–213. Leuven: Departement Oriëntalistiek.
——— . 1987c. On the Origin of the Literary Device of the “Frame Story” in Old
Indian Literature. In H. Falk, ed., Hinduismus und Buddhismus: Festschrift für
Ulrich Schneider, 380–414. Freiburg: Hedwig Falk.
——— . 1989a. Tracing the Vedic dialects. In Collette Caillat, ed., Dialectes dans les
littératures indo-aryennes. Acts of the Colloque International, Paris, Sept 16–18,
1986. Publications de L’Institut de Civilisation Indienne, Series 8, Fasc. 55,
97–265. Paris: Edition-Diffusion de Bocard.
——— . 1989b. Ṛgvedic History: Poets, Chieftains and Polities. In Collette Caillat, ed.,
Dialectes dans les littératures indo-aryennes, 307–52. Paris: Édition-Diffusion de Bocard.
——— . 1995a. Early Indian History: Linguistic and Textual Parametres. In George
Erdosy, ed. The Indo-Aryans of Ancient South Asia: Language, Material Culture and
Ethnicity, 85–125. New Delhi: Munshiram Manoharlal.
——— . 1995b. Ṛgvedic History: Poets, Chieftains and Polities. In George Erdosy, ed.
The Indo-Aryans of Ancient South Asia: Language, Material Culture and Ethnicity,
293–352. New Delhi: Munshiram Manoharlal.
——— . 1997a. The Development of the Vedic Canon and its Schools: The Social and
Political Milieu. In Witzel ed., 1997c, 257–345.
——— . 1997b. Early Sanskritization. Origins and Development of the Kuru state. In
B. Kölver, ed. Recht, Staat und Verwaltung im klassischen Indien. The State, the Law,
and Administration in Classical India. 27.52. Munich: R. Oldenbourg. = 1995a
(Preprint from the author via Madeleine Biardeau).
——— , ed. 1997c. Inside the Texts—Beyond the Texts: New Approaches to the Study of the
Vedas, Harvard Oriental Series Opera Minora, Vol. 2. Cambridge: Department of
Sanskrit and Indian Studies, Harvard University.
——— . 2005. The Vedas and the Epics: Some Comparative Notes on Persons,
Lineages, Geography, and Grammar. In Koskikallio, ed. 2005, 21–80.
——— . 2006. Brahmanical Reactions to Foreign Influences and to Social and
Religious Change. In Olivelle, ed. 2006, 457–99.
Woods, James Hampton. [1927] 2003. The Yugasūtras of Patañjali. Mineola, NY: Dover.
Woods, Julian F. 2001. Destiny and Human Initiative in the Mahābhārata. Albany:
SUNY Press.
 bibliography 725

Wulff Alonso, Fernando. 2008. Grecia en la India: El reportorio griego del Mahābhārata.
Madrid: Ediciones Akal.
Younger, Paul. 2002. Playing Host to Deity: Festival Religion in the South Indian
Tradition. Oxford: Oxford University Press.
Zaehner, Robert C. 1963. Utopia and Beyond: Some Indian Views. Eranos-Jahrbuch
32: 281–309.
——— . 1966. Hinduism. New York: Oxford University Press.
——— . 1970. Concordant Discordant. Oxford: OUP.
——— . [1969] 1973. The Bhagavad-Gītā with a Commentary Based on the Original
Sources. London: Oxford University Press.
This page intentionally left blank
Index

Abhidharma, Abhidhamma, 8, 49, 379–83, 428, 496–97, 588 n. 59;

106–8, 124–50, 155, 147 n. 215, and Ambikā and Ambālikā in
300, 305–6, 679 n. 153 Aśvamedha, 375; reborn as
adharma, 99, 194, 418, 420, 478, Śikhaṇdị n, 490 n. 21. See also
492, 558 Tryambaka
Adharma, a deity, 84 Ambikā, 362–63, 374–75, 380–82,
Adluri, Vishwa, 14 n. 16, 19 n. 28, 191 384, 401, 4905, 409, 418, 525
n. 25, 209 n. 81, 422 n. 25, 481 Ambālikā, 362–63, 374–75, 377,
n. 1, 482 n. 4, 543 n. 45, 557 n. 74 380–82, 384, 392, 401, 405,
Agastya, 378 n. 94, 423 n. 27, 408–9, 525
486–87, 596, 617–18, 676 n. 132 Ambaṭṭha, 112–21, 158, 166–67, 329
Aggañña Sutta, 107 n. 14, 108, 116, Ambaṭṭha Sutta, 107–21, 166–67,
152, 155, 161–79, 219, 248, 336, 216, 671
610, 671 aṃśāvataraṇa, 82–83, 265 n. 53, 526
Agni, 58–59, 62–65, 71, 82, 87 n. 71, n. 49, 577–78, 580, 595
89, 98, 204 n. 67, 214, 263 Aṇīmāṇḍavya, 326, 437, 578
n. 47, 314, 470, 485, 530, 572 ānṛśaṃsya, noncruelty, 24, 215, 220,
n. 16, 599–600, 607 397 n. 136, 432, 435, 449–53,
Agnimitra, 285–86 457–58, 465–68, 471–72,
Agnivaṃśa, Agnikula, 286 476–80, 522 n. 13, 535,
Agniveśya kings, 285–86, 297 549–50, 552
ahiṃsā, nonviolence, 176, 189, Ā pastamba Dharmasūtra, 8, 40, 164,
219–20, 432, 457, 474, 476, 559 176, 182–83, 188–90, 192–95,
Ajātasattu, Ajātaśatru, 85 n. 63, 109, 201, 203–4, 207, 217 n. 107, 219
158–59, 282 n. 24, 680 n. 14 n. 114, 224, 228, 230–33, 240,
Ambā, 337 n. 1, 361 n. 61; 362–63, 520–21, 535–36, 608–9; and
369–70, 374–75, 377 n. 88, women’s dharma, 338–40, 347
728 index

arhat, arahant, 117, 148 n. 30, 160–61, Aśoka Maurya, 5, 8–9, 11, 30–31, 33, 35–51,
163, 251–52, 303, 306–7, 309, 322, 56–57 n. 13, 82, 93 n. 80, 104, 124,
324–25, 327, 322–33, 395 n. 131, 628 141–42, 151 n. 136; 152 n. 138, 177,
n. 10; arhaticide, 301, 306–7; as 189–91, 223, 244–47, 257–58, 283
“what is primary,” 161, 163–65, 169, n. 26, 294, 296–97, 313, 412, 592,
172, 174, 298, 303, 312, 333 623–24, 683–84; Buddhist affiliation
Arjuna (Pāṇḍava), 164, 237, 347 n. 20, of, 36, 38–40, 44–45, 49, 228, 124,
380 n. 97, 381 nn. 98 and 99, 439, 141; critique of useless rites, 47, 49,
443–44, 447, 449, 458, 464, 466, 193 n. 35; in Buddhacarita, 683–84;
475–76, 532, 543, 549, 577, 582, Kalinga war of, 40, 46–48, 56
598, 634, 646, 648–49, 651, 656, n. 13, 580; as watershed figure, 31,
655–66, 675; in Bhagavad Gītā, 10, 296, 453, 623
12, 237, 462, 513–14, 523, 527, Aśokan edicts, 5, 8, 177, 223, 411, 615
536–37, 540–42, 547, 551–68, 590, n. 128, 681 n. 145; as autobiography,
611, 672, 377–78; and Draupadī, 30–31, 37; Minor Rock Edicts (MRE)
341, 389, 489–92, 494, 507 n. 57, 1 and 2, 37–44, 47; Rock Edicts (RE) 4
547; and Droṇa, 22, 460–61, and 5, 244–47, 257–58
465–67, 467–70, 474–77, 489–90; Aśokāvadāna, 306, 305–6, 309 nn. 83
and Ekalavya, 461, 466–67; and and 85, 310–11, 314 n. 95, 315
Karṇa, 22, 444, 454, 458–59, n. 98, 324
460–61, 463–65, 467, 622–23; and āśramas, four life-patterns or life-stages,
Kṛṣṇa, 10, 22–24, 26–27, 260, 388, 164, 215–24, 236, 522, 558 n. 76 584,
417 n. 14, 444, 460–61, 463–65, 614, 644, 652, 667–68; pro-choice
467–72, 478, 514, 547, 552, 603–6, position, 219, 645
619–20, 622–23; and Śiva, 448 Aśvaghoṣa, 4, 7–9, 11, 29, 31–33, 258
n. 90, 572; as Bībhatsu, 467, 477; n. 38, 411–12, 604, 624–29, 633–41,
as karmayogin, 536–37, 540–42; as 645–50, 652–66, 670, 672–75,
a lowly man, 459, 479; as formerly 677–79, 81, 683–84; knows
the Ṛṣi Nara, 259–61, 546, 548; Harivaṃśa, 646
hears Anugītā after forgetting Aśvamedha, horse sacrifice, 58, 80, 183,
Bhagavad Gītā, 541, 550; 546, 603; 109 n. 22, 187 n. 16, 188 n. 19, 191,
not a king, 553; prewar dilemma 572; in Rāmāyaṇa, 417 n. 12, 424,
of, 558–59, 677; rebuked by 425 n. 34, 601–2, 605 n. 11, 615;
Bhīma, 467, 477–78; rebukes Janamejaya’s, 28, 580–84; role
Aśvatthāman, 479 of mahiṣī or chief queen in,
Arthaśāstra, 7, 39, 57 n. 13, 181, 224–26, 375–82, 400, 605; Vyuṣitāśva’s,
228–29, 236 n. 165, 237–38, 507 400–1; Yudhiṣṭhira’s, 436, 603
n. 52, 584 n. 46 n. 105, 604 n. 107, 637 n. 35, 678
Ā rya, 78, 94 n. 75, 184–86, 193–95, 200, n. 141
226, 229, 231, 233, 235–36, 241, 379, Aśvatthāman, 279, 367, 439, 453,
604; anārya, 637 n. 37; Ā ryāvarta, 469–75, 478–79; elephant named
193, 313 nn. 93 ad 94; and Śūdra, 75; Aśvatthāman, 453, 462, 469–70,
Indo-Aryan languages, 139 472–75, 480
Aryaman, 76, 379 Atharva Veda, 78–80
 index 729

ātman, self, 87 n. 72, 93, 96 n. 92, 585, 587, 601, 604, 607, 613–14,
98–99, 100 n. 105, 433; no self, 95, 620 n. 142, 655; bhakti “swerve”
628; Paramātman, 252, 502 of Mahābhārata, 21, 29, 270, 569;
ātmanas ṭuṣṭi, ātmatuṣti, what is pleasing bhaktiyoga, 540, 567–68; in
to the self, 192, 194–95, 410 Buddhacarita, 624–25, 628–29 and
Atri poets of Ṛgveda, 55, 58–59, 68; Atri, n. 12, 635, 655–56; dharma and, 32,
617, 635, 637, 641; Atri and 74, 201, 517, 533, 569–624; mapping
Anasūyā, 423 n. 27, 486, 493–94, dharma and bhakti, 32, 570–71,
496, 572 n. 16, 641 584–85; politics of bhakti, 616, 623
Austin, Christopher, 14 n. 16, 433 n. 54, Bhāratavarṣa, land of Bhārata, 271, 453
464 n. 122, 510 n. 61, 573 n. 17, 576, n. 99
675 n. 130 Bhārgava Rāma, Rāma Jāmadagnya, 277,
avatāra, avatar, 32, 78, 123, 256, 350 n. 24, 295, 335, 353, 418, 496–97, 553, 585
490 n. 23, 575–77, 589–607, 608 n. 15; n. 48, 594, 617 nn. 132 and 134,
of Viṣṇu, 32, 78, 281 n. 20, 564, 577. 621–22, 637, 675 n. 130; as avatar,
See also Buddha, Bhārgava Rāma, 590; slaughtered Kṣatriyas
Boar, Fish, Kalki, Kṛṣṇa, Man-lion, twenty-one times, 291, 365
prādurbhāva, Rāma, Tortoise, Vāmana Bhaviṣya(t) Purāṇa, 18 n. 26, 203, 219
avataraṇa, 32, 595, 597–99. See also n. 115, 277–78, 282, 292–94; extant
aṃśāvataraṇa, gaṅgāvataraṇa Bhaviṣya Purāṇa, 280, 286, 655
Bhaviṣya Parvan of Harivaṃśa, 11, 577–78,
Bagchee, Joydeep, 14 n. 16, 19 n. 28, 537 581–82, 584
n. 43, 538, 540, 543 n. 45 Bhīṣma, 279, 342, 357 n. 43, 548–49, 619,
Bailey, Greg, 66 n. 31, 74, 100 n. 108, 137 621–22; and Ambā, 362, 490 n. 21;
n. 103, 211 n. 88, 253–54, 507 n. 53, 497; as “grandfather,” 24, 215 n. 99; as
512 nn. 67 and 69, 586–87 matchmaker, 356, 359–60, 374,
Baudhāyana Dharmasūtra, 5, 8–9, 164, 385–86; celestial education of, 353, 553
175–76, 182, 186, 189, 192–93, n. 65; celibacy of, 340, 342, 356–57,
203–4, 210 n. 83, 217 n. 107, 230 481; in Harivaṃśa, 646; postwar
n. 147, 339, 610, 612 dharma oration of, 222, 229, 253, 280
Besnagar Garuḍa pillar of Heliodorus, 286 n. 19, 428, 466 n. 124, 481, 521 n. 9,
Bhagavad Gītā, 10, 22, 32, 41 n. 16, 53, 534 n. 29, 536, 555, 575 n. 24, 586, 596,
181, 218, 237, 244 n. 2, 272, 344 607, 611; his name Śāṃtanava,
n. 14, 4622, 495, 513–15, 517–68, 348–49, 357, 359; son of Gaṅgā, 347–54
570–71, 589–92, 594, 598, 600, Biardeau, Madeleine, 12 n. 7, 82 n. 54, 122
603, 606, 611, 620 n. 142, 631, 656, n. 60, 271 n. 62, 440 n. 69, 590, 608
658, 662, 672, 677–78; barleycorn n. 115; on bhakti, 29, 245 n. 5, 251–52,
structure in, 564; lawful or just war 569; on dharmaśāstra, 182, 184, 188,
in, 56 n. 13, 560, 590; ring structure 189 n. 21, 190 n. 23, 191, 193–94, 206
of, 517, 543, 546, 554, 559, 594. See n. 74, 208 n. 76, 216 n. 103, 217, 222,
also Mahābhārata 223 n. 128, 228 n. 146, 230 nn. 148
bhakti, 12–13, 15, 29, 78, 245 n. 5, 251–53, and 149, 231–32, 239, 266–67; on
267, 269, 327, 379, 393, 406, kalpa and yuga, 250 n. 19, 268, 70; on
498–501, 504 n. 46, 505–6, 514, 533, Mahābhārata as riposte to Buddhism,
730 index

Biardeau, Madeleine (continued) of, 251; in Nārāyaṇīya, 251–56, 260

21, 362 n. 65, 412 n. 2, 592, 625–26, n. 42, 262, 270 n. 58; life of, 251–53,
654 n. 89, 655; on Mahābhārata, 337 270, 336; mental sons of, 255, 262,
n. 1, 344–45, 355–59, 362, 366–69, 371, 670–71; persuades Buddha to teach,
375–76, 380, 382 n. 101, 384, 391 n. 121, 132, 209; path to the company of,
392, 427 n. 37, 428 n. 41, 431 n. 49, 436 187, 209, 396; recommends Gaṇeśa
n. 62, 458, 467 n. 126, 477 n. 244, 506 as Vyāsa’s scribe, 18 n. 25, 209;
n. 49, 507 n. 56, 512 n. 69, 513, 521, 553 weapon of, 471, 478, 487, 552
n. 66, 549–50; on Mahābhārata Critical Brahmā Sanatkumāra, “Forever Young,”
Edition and Vulgate, 15–19, 653; on 116, 671; Sanatkumāra as mental son
reign of Yudhiṣṭhira launching a Kṛta of Brahmā, 255–56
yuga, 271 n. 62 brahmacarya, 100 n. 106, 185, 187, 217
Bimbisāra, also called Seniya and Śreṇya, n. 209, 218, 381 n. 100, 667
King of Magadha, 109, 158 n. 64, 328 brahmadeya, 112
n. 131, 628 n. 9, 641, 644, 646–55, Brahmadhammika Sutta, 117
658, 663, 673, 678 n. 140, 683 Brahmajāla Sutta, 248–50, 513 n. 71
Boar, 654 Brahmaloka, 100, 165, 396, 407, 613
bodhisattva, 141 n. 112, 193, 250, 303–5, Brahman, brahman, 188, 214, 222, 239,
309, 319, 326, 331–33, 459, 630 251, 252 n. 23, 262, 268, 276, 352,
n. 16; Candragarbha bodhisattva, 417 n. 14, 474, 513 n. 69, 529–31,
320, 331; pre-enlightenment Buddha 565–68, 574, 613–14
as, 630 n. 19, 633 n. 26, 648–49, Brāhmaṇas (texts), 58, 60 n. 19, 92–95,
651–55, 659–60, 668 100, 165, 182–83, 185–86, 275, 293,
Bodhisattva Piṭaka, 304, 679 n. 143 590, 606
bojjhaṅga, bodhyaṅga, 127, 140 Brahmanicide, 200 n. 55, 377 n. 90, 402,
Bowles, Adam, 3, 12 n. 7, 36 n. 3, 42, 44–49, 426, 463
56 n. 13, 79–80, 83–90, 93 n. 83, 99, Brahmarṣideśa, 193 n. 34, 313
254, 280 n. 19, 365, 391 n. 120, 607, Brahmāvarta, 137 n. 102, 193 n. 34,
656 n. 94, 672, 676 n. 132, 678 n. 142, 216 n. 105
680 Brahmins, 18–19; under Aśoka Maurya,
Brahmā, 26 n. 44, 99 n. 103, 163, 165–66, 41, 45–47, 49; mahāsāla wealthy
193 n. 33, 208–11, 221–22, 244 n. 3, landlord Brahmins, 98, 111–12,
248–49, 262, 265, 267, 272, 345–46, 117–18, 186, 188, 274 n. 33;
350, 354, 357, 396, 414, 427 n. 37, privileges of, 79, 85, 88–89, 92, 98,
430 n. 48, 482, 492, 512–13, 530–31, 123, 127, 224, 510, 623; Pūrvaśikhās
533, 550, 571, 573 n. 16, 577, 577, 580, and Aparaśikhās, 19; Vedic oral
568, 587–88, 590, 595–96, 598–99, tradition of, 18, 54–56. See also
606, 613, 616, 618; and poetry, 18 Buddha and Brahmins, six jobs of
n. 25, 209–10, 415, 418–19; and Brahmins, svadharma, svadharma,
Śeṣa, 261; boons given by, 23, 414 svakarma
n. 6, 487 n. 15; Brahmins born from Brereton, Joel, 54, 56, 58–73, 75–78, 80
his mouth, Samaṇas from his feet, n. 84, 96–97, 125
113, 117, 167–68, 329; days and Bṛhaspati, Brahmanaspati, 73. 276, 353,
nights of, 251, 253, 262, 267–70; egg 486, 507–8, 511 n. 65, 513, 536
 index 731

Brodbeck, Simon, 12 n. 7, 13 n. 12, 83 654; centrality of dharma in, 4, 625–33;

nn. 57 and 58, 253 n. 28, 262 n. 45, mokṣadharma in, 656–63, 673–82
344, 346, 352 n. 30, 353 n. 32, 356 Buddhavacana, word of the Buddha, 154,
n. 39, 383, 386, 390 nn. 118 and 119, 192, 302, 309
91 n. 120, 392 n. 123, 401, 405, 466,
485 n. 13, 536 n. 32, 538, 540 n. 39, Cakkavatti Sīhanāda Sutta, 108 n. 15, 122,
541, 573, 579, 584 nn. 46 and 47, 161, 175 n. 216
602 n. 101 cakra, wheel, discus, 122–23, 279, 373
Brockington, John, 12, 286, 416, 421 n. 81, 547 n. 53, 629
n. 21, 545 n. 50, 570 n. 5, 607 n. 13, cakravartin, wheel-turning monarch, 122,
637, 639, 649 n. 69, 650 n. 73, 653 161, 373 n. 81, 593–94
n. 84, 655 n. 92, 658; and Mary Caṇḍāla, Cāṇḍāla, 98, 326–30, 655 n. 91
Brockington, 500–4 Candragarbha Sūtra, Chinese, 250 n. 16,
Bronkhorst, Johannes, 13 n. 11, 45, 91 309 n. 83, 318–36; question
n. 78, 93–97, 100, 107 n. 14, 115, 127 of Indian origins, 323–24, 326–27,
n. 71, 137 n. 107, 191 n. 26, 193 n. 34, 329; Tibetan, 308 n. 80, 309 n. 83,
195 n. 45, 197 n. 47, 202 n. 63, 206 319–21, 323–26, 335–36, 395 n. 31
n. 72, 208 nn. 76 and 77, 218 n. 110, Candragupta Maurya, 45, 56, 57 n. 13,
244–45, 273, 283–84, 506, 506–8, 292, 294
514, 520 n. 5, 611 n. 119 caste, class, social classes, 10, 30, 40, 45,
Buddha, 4, 6–7, 10–11, 31–32, 36, 38, 48, 54, 87, 98, 108–9, 111–16, 158,
40, 44, 47, 49, 51, 56 n. 13, 92 165, 167–70, 174–75, 179, 185, 193
n. 81, 567, 610; and Buddhas in n. 34, 206 n. 73, 215–24, 227, 270,
Buddhist texts, 30, 103–79, 246–52, 347, 400, 408, 427, 448 n. 90, 462
272, 297, 300–2, 305, 307–11, n. 120, 498, 519–28, 532–33, 546,
314–19, 322, 326–33, 336, 558, 560, 559–62, 570 n. 4, 605, 609, 618,
604–5, 625–56, 659–62, 674–84; 626, 630, 677 n. 138; mixed castes,
and Brahmins, 30, 101, 106–23, 133, miscegenation, 115, 120, 215–17,
135–37, 158–60, 162 n. 173, 163–79, 220–21, 224, 270, 317, 372 n. 80, 374
527; as avatar, 590; enlightenment n. 82, 526–27, 555, 586, 592; no fifth
of, 10, 40, 128–29, 131–32, 209, 249, caste, 215, 527; untouchability, 328,
302, 640, 641 n. 48, 655 n. 91, 674, 649 n. 69; varṇadharma, 48, 207,
677, 681; 161; first sermon of, 107–8, 219, 416, 521–22, 528–29, 533, 558,
128, 132–33; four signs episode, 570 n. 4. See also Cāṇḍālas,
630–33; great departure of, 10, 631 varṇāśramadharma
n. 21, 637, 639–41, 646, 659, 679; celibacy, 92, 114, 159–61, 178, 186–87,
in Kuru country, 134–37, 558; jāti 234 n. 157, 262, 342, 397 n. 20, 357,
profiling by, 111, 115; on whether 364, 409, 497. See also Bhīṣma,
slain soldiers attain heaven, 560. See brahmacarya
also Buddhacarita; Buddhavacana China, 7–8, 244 n. 3, 298–301, 304 nn. 68
Buddhacarita, Adventure of the Buddha, 4, and 69, 318–19, 326, 329–36, 627.
7–9, 11, 29, 31–32, 258 n. 38, 411, See also Candragarbha Sūtra, Chinese
625–84; bodhisattva appears like Classical dharma texts, 5–11; minor
Dharma incarnate in, 647–48, 651, classical dharma texts, 7–8, 31, 273
732 index

Collins, Steven, 160–78, 303, 304 n. 71 dharma, 3–4, 29–33; as civil discourse,
cooking the world, 87–90, 193, 127 n. 69 4–5, 123–24; as enigma, 30, 54,
courtesy, discourtesy, 41, 46, 113–14, 58–66, 78, 87, 90–91, 100, 124, 148,
120, 241 201, 378–89, 410, 432, 434; as
Couture, André, 271 n. 62, 490 n. 23, 571 subtle, 10, 22, 33, 99–100, 132,
n. 11, 575 n. 24, 578–80, 584 n. 47, 135–37, 178, 193, 201, 213–14, 347
589 n. 66, 590 n. 70, 595–99 n. 20, 354, 414, 455, 458 n. 111, 473,
critical editions, 5, 11–20. See also 455, 583–84, 674; decline of, 5, 36
Mahābhārata, Rāmāyaṇa, n. 2, 41 n. 6, 270, 296–97, 312;
Mānavadharmaśāstra desire for, 342, 380, 538, 676; end of
custom, 30, 58 n. 16, 78–80, 151, 162, in Buddhism, 5, 8, 250, 298–304,
173–74, 182 n. 4, 192, 194–95, 207, 306, 310, 317–18, 321, 324–25,
209 n. 78, 347 n. 18, 374, 410, 535, 332–33; eternal or everlasting, 7, 125,
579, 604–5, 644 n. 57 195 n. 33, 243, 254, 339, 365, 402–3,
426, 477, 526, 565, 587, 600, 633,
dāna, gift, giving, 185 n. 10, 270 n. 58, 670; in Brahmanical legal tradition,
320, 437, 635, 680–81; as 20, 196–201, 210 n. 83, 228–32, 360
yugadharma of Kali yuga, 269, 583; n. 54; in history, 4–5; increase of, 39,
dānadharma, 428, 658, 663, 669, 42–430; Kṛsn ̣ ạ ’s first word in
678–79, 681–83; dānapāramitā, 459; Mahābhārata, 491; sources of, 10,
dānadhaṃmam according to Aśoka, 46, 61, 148, 191–92, 204 n. 65, 207,
46–48 209, 347, 477
daṇḍa, rod of punishment, 122–23, 227, Dharma, as deity, 9, 85, 587; father of
239, 328, 522 n. 4, 534–35, 543 n. 48, Yudhiṣṭhira, 21, 24; 260, 343, 574;
570 n. 4, 588, 676; brahmadaṇḍa, house of, 259–60; incarnates in
116, 121, 231; daṇḍanīti, 228–29, 237 Vidura, 372, 409, 547, 574; ten wives
Daṇḍa personified, 211 n. 87, 222–23, of, 262–64; three sons of, 263–64.
226–27 See also Vidura, Yudhiṣṭhira
Dantavaktra, 279–81 Dharma bull, 213, 289, 577 n. 30
Davis, Donald, 26, 194–95 dharma, etymologies and meanings of,
debt, 416, 524, 602, 660; three 36, 54–55, 66–67, 71–72, 100,
debts; 185–86, 353, 363, 397, 524 124–27, 138–39, 144–45, 150, 519,
n. 16, 630, 660–63; four debts, 397, 565. See also dhárman
524 n. 16; “theology of debt,” 186, dharmakṣetra, field of dharma, 556–67
660–62 Dharmaguptakas, 152 n. 139, 154–55, 250
dependent origination, law of dependent n. 17, 304–5, 670 n. 143
arising, 128–32, 138, 144 n. 117 dhárman, 25, 30, 52–59; as foundation,
deśadharma, regional dharma, 151–52, 54–79; 82; 90–91, 97; as nature or
184, 532 quality, 66–67, 97, 125, 629 n. 14;
devakārya, surakārya, “work of the and peaceful settlement, 56, 67, 235;
gods,” 256, 350, 390 n. 16, 485, 572, dharmán as foundation-giver, 54
577 n. 51, 600, 609, 671. See also dharmapatnī, legal wife, 356, 362, 394,
divine plan 398, 481–82, 498; 85, Śrī as wife
dhamma-eye, 119–20 of Dharma
 index 733

dharmapramāṇa, Buddhist theory of, Doniger, Wendy (a.k.a. Wendy Doniger

192 O’Flaherty), 14 n. 15, 15–16, 39 n. ‘8,
dharmarāja, epithet of Buddha, 614; ‘84, 108 n. 76, 212 n. 89, 233, 253
name for Aśoka, 683–84; name n. 29, 350 n. 25, 355, 366, 497, 527
of Yama, 417 n. 15, 426, 437 n. 64, n. 21, 529, 532–33, 608 n. 116
440, 447, 492; name of Yama as Draupadī, 10, 27, 31–32, 188, 217, 236,
Pretarāja in Rajasthan, 439–40. 280–81, 338, 341, 345, 355, 363, 389,
See also Yudhiṣṭhira 395, 400 n. 142, 406, 412–13, 326,
Dharmarāja and Draupadī temples in 428, 444, 446, 454, 457, 561, 464,
Tamilnadu, 4, 440, 452 n. 98, 597 467, 470–71, 478–82, 499, 534, 536,
dharmas plural, Buddhist usage, 124–34, 538, 586, 589, 591, 615, 678; and
137–44; bodhipakṣya dharmas, Duryodhana-Periyantavar cult,
bodhipakkiyā dhammā, 128; lists of 340–41; as Kṛsn ̣ ạ ,̄ 413 n. 11; 280, 458;
phenomena, 124, 128; seventy-five birth of, 25 n. 42, 390 n. 19, 482–86,
of Sarvāstivādins, 145; wholesome 575–76; cult and temples of, 4,
and unwholesome, 128, 133–34, 130, 440–41, 452 n. 98, 597; disrobing of,
143–44, 174, 249 362, 422 n. 23, 558, 621; friendship
dharmasūtras, 6, 9, 46, 15, 120, 122, with Kṛsṇ ạ , 491, 605–6, 615, 619 n. 41,
151–52, 164–65, 175, 182–98, 201–8, 620, 622; in exile, 201 n. 57, 395, 410,
215–19, 225–26, 234–37, 339, 360, 434, 436, 506–16, 543; in court, 339
410, 451 n. 110, 508, 516, 525, 531, n. 5; lady pundit, 507; materialism of,
535, 605, 609, 623 n. 148, 636, 638, 506–7, 509, 511, 513–16; question of,
645, 682; and dharmaśāstra, 115, 152, 362, 412, 417 n. 17, 458, 470;
178, 181, 190–98; and king, 228–31 svayaṃvara and polyandry of, 5, 341,
“dharmic hunter,” 201 n. 60, 486 n 142, 344, 347 n. 9, 374, 403, 488 n. 93,
589 n. 59, 682 495–96, 547, 573–75, 637 n. 35
Dhātṛ and Vidhātṛ, 26, 73–74, 488, 507, Droṇa, 22–25, 27, 29, 279, 414, 439,
512–13, 548, 585–89, 591; as female 453–55, 457–79, 485, 487 n. 14,
weavers of days and nights, 334 489–90, 520 n. 8, 544, 549–53, 558,
n. 14, 373, 384 578, 621–22, 656
Dhṛtarāṣṭra, 237, 390, 392 n. 27, 468, Dumézil, 15, 17, 68, 344, 360–62, 374,
474, 482, 514 n. 73, 573 nn. 17 and 379, 383, 385, 427 n. 29, 455
18, 574, 588, 620, 622–23 Duryodhana, 9–10,27–29, 271 n. 62, 280
Dhṛṣṭadyumna, 279, 485, 487 n. 14, 490, n. 20, 311, 369, 387–88, 405, 408, 4417,
509, 513, 578, 599; incarnation 445, 454, 458, 460, 464, 479, 482,
of Agni, 485, 510 489, 511, 545–50, 553, 571 N. 9, 578
Discrimination of dharmas, 49, 139–42 N. 32, 589 n. 64, 600; as Periyantavar,
Divine plan, 31; in Mahābhārata, 31–32, 340–41, 440–41; piece of Kali, 578
265, 334, 350, 352 n. 30, 362, 373, Duṣprasaha, 309–11, 323–25, 327, 329
389, 390 n. 19, 409, 481, 486–95,
498, 513, 570–78, 584–86, 596, earth, 39 n. 11, 63, 64, 68, 71, 76, 79, 81,
599, 611, 619, 648; in Rāmāyaṇa, 83, 199, 256, 261–62, 264, 303,
350, 481, 487–95, 498, 511, 575, 595 351–53, 384, 455–56, 456, 465,
734 index

earth (continued) five brāhmaṇadhammā, 187

551, 584, 590, 596, 621, 648, five mahāyajñas, great sacrifices, 185,
653–54, 683; aging of, 281 n. 22, 187–89, 263, 391 n. 121, 397 n. 136,
409; and Draupadī, 484–85, 510, 400, 593, 609, 612
515; and Karṇa, 458–59; and Sītā, five hindrances, 119, 134, 139
210–11, 377 n. 92, 417, 420, 493, four foundations of mindfulness, 128,
505, 516, 613; Brahmins as gods on, 137–39
222, 528, 609; destruction of, 278, four life-patterns or life-stages.
454, 623; distributed by Yayāti to his See āśramas
sons, 602; dug up by sons of Sagara, four noble truths, 125, 132, 139, 146 n. 22
572 n. 16, 595–96; earth’s business, four paths of failure, 117
265, 362, 574; earth essence, 169–71; four puruṣārthas, 234, 237 n. 166, 352,
fragrant earth, 171, 336 n. 143; held 367, 392, 518
up by elephants, 61, 442 n. 79; Kuntī four right endeavors, 128, 133, 139
as Pṛthā (Pṛthivī), 388–891; in Kalki four sublime attitudes or Brahmavihāras,
myth, 592–95; Pratiṣṭhāna as vagina 114 n. 34, 122, 187
of, 602; relieving the burden of, 265, Framarin, Christopher, 537–48, 540–41
517 n. 15, 573 nn. 18 and 19, 575 friendship, 32, 61, 121–22, 459, 491 n. 25,
(Greek), 577–78, 598–99, 601; 533, 570–71, 584–85, 603–5, 611–12,
urbanization of in Harivaṃśa, 577 615–16; in Rāmāyaṇa, 431, 618–19;
eḍūkas, sepulchral mounds in Kali age, Kṛṣṇa’s friendships in Mahābhārata,
592–93, 683–84 388, 491, 508, 619–24
Ekalavya, 460–61, 466–67, 476–77, 479,
604 n. 108, 618 n. 138 gaṇa-saṅghas, 158, 160
Gāndhārī, 207, 343 n. 11, 344 n. 14,
Falk, Harry, 37–40, 44, 64 n. 24, 371 n. 79, 374, 382–94, 397, 405,
109 n. 22, 187 n. 16, 207 408, 510, 447 n. 53, 549–50,
Fish avatar, 590, 654. See also Satyavatī 572 n. 15, 573 n. 17, 578, 604 n. 108
Fitzgerald, James L., 194 n. 34, 202 n. 61, Gandhi, Mohandas K., 560
342 n. 10, 370 n. 78, 412 n. 2, 432, 436, Ganeri, Jonardon, 21–24, 353 n. 101,
454, 519, 536 n. 32, 570 n. 4, 594 n. 86, 461–62, 466 n. 124, 468–73,
598–99, 612, 623–24, 659 n. 102; on 476–77, 522 n. 13
bhakti, 12, 620 n. 42; on features of Gaṅgā, 573, 80, 90, 249 n. 14, 595–96, 601;
Śāntiparvan, 26 n. 44, 144 n. 116, 116, goddess, 32, 56, 297 n. 51, 337,
204 n. 67, 227, 275, 280 n. 19, 394 344–55, 357–59, 363, 597, 675 n. 130;
n. 129, 428 n. 43, 453 n. 101, 447 n. 34, gaṅgāvataraṇa, 352 n. 30, 423 n. 28,
543 n. 48; on four anthologies, 679; 595–96, 595–97
on Guptas, 12–13, 197 n. 47; on Gautama Dharmasūtra, 8–9, 164, 176,
“invention” of the Pānḍ ạ vas, 12 n. 9, 182, 189, 192, 198, 202–4, 218–20,
298 n. 51; on Khāravela, 282 n. 25, 225–27, 234, 241, 339, 347,
285 n. 30, 298 n. 49 374 n. 82, 521–22, 652 n. 78
five former Indras, 12 n. 8, 437 n. 64, Gautama and Ahalyā, 421, 617
483, 574 n. 23 Gethin, Rupert, 53, 97 n. 97, 107–9,
five skandhas, aggregates, 130–31, 139 124 n. 63, 125, 127, 130, 137 n. 103,
 index 735

139–441, 144–48, 163 n. 176, Hanumān, 189 n. 20, 232 n. 153, 270
178, 297 nn. 60 and 61, 413, 456 n. 105, 482,
Ghaṭotkaca, 417 n. 14, 454, 460–66, 486, 499, 504, 520, 536, 572 n. 12,
571, 574 615, 619, 639, 650 n. 75; Bālājī-
gleaners, gleaning. See uñchavṛtti Hanumān in Rajasthan, 439–40
gold coins, 206–7 Hiḍimbā, 460, 497, 532–34
Goldman, Robert P., 12 n. 4, 342 n. 10, Hīnayāna. See Nikāya schools
414 n. 6, 415 nn. 9 and 10, 416, Hopkins, Edward Washburn, 19 n. 28,
418–19, 421, 486, 503–6, 520, 543, 20, 192 n. 30, 196–97, 200 n. 56,
574, 600 n. 97 202–6, 592, 657 n. 97
Goldman, Sally Sutherland, 12 n. 4, hospitality, 32, 61, 78, 120–22, 185–89,
14–15, 343 n. 11, 414 n. 6, 483 n. 5, 220, 496, 553, 570–71, 579, 584–85,
484 n. 11, 501 n. 42, 503–6, 520, 574, 598, 603–22, 641, 651
600 n. 97 householder, 46, 108, 114 n. 34, 121–22,
González-Reimann, Luis, 244 n. 45, 252, 152 n. 39, 157. 163–66, 166, 172–75,
256–58, 268–72, 276 n. 11, 278, 186–89, 214, 217–19, 236, 249, 253,
288–90, 293, 296, 414 n. 6, 298, 308, 336, 393–94, 459, 528,
443 n. 79, 458 n. 55, 571 n. 9, 534 n. 29, 607, 609, 611–14, 630
591 n. 67, 9 n. 79 n. 16, 639, 643–44, 662–63, 668,
gratitude and ingratitude, 415 n. 10, 675–77; Buddhist dialogues with
463–64, 500–1 householder Brahmins, 119–21, 152,
Grünendahl, Reinhold, 12 n. 6, 163; householder ideal, 163–66;
260 n. 42, 543 n. 46, 668 n. 118 householder kings, 393, 38, 67–76;
guṇas, vitues, 49, 219, 318, 352, 355, houses built to conceal sex, 172–73, 179
377, 398, 415, 418, 452, 503, 576, humor, 122, 162, 167, 173, 177–79,
519, 521–22, 551; three guṇas. 196, 491, 638; satire, 167 n. 191, 248 n. 9
228 n. 45, 323–25, 556, 567 n. 82
Gupta dynasty, Guptas, 6–7, 12–14, 36, Ila/Ilā, 601–3
245 n. 4, 258, 287, 292–93, 311–12, Indra, 56–57, 60–65, 68–69, 72–74, 76,
318, 329–30, 336 n. 31, 666 80 n. 50, 82, 85–87, 90, 92, 200, 214,
Greeks, 30, 36, 42, 45, 50, 216, 264, 270, 239, 369, 377 n. 90, 379, 400 n. 142,
274–75, 282–88, 292, 295–98, 3022, 426, 459, 480, 488, 492, 555, 598,
309, 313–17, 324 n. 122, 328, 334, 601, 605, 637, 641; and Ahalyā, 421;
336, 534 n. 29, 575; Indo-Greeks, and Arjuna, 85, 260, 405, 572 nn. 12
274, 282 and 15; and Varuṇa, 56, 58, 69; and
̣ , 76–77, 85–87, 240, 605, 645;
Viṣnu
Halbfass, Wilhelm, 52 n. 1, 53, 61 nn. 20 enters Janamejaya’s sacrificial horse,
and 22, 74 n. 37, 97, 127, 147 281, 375 n. 86, 581–82; five former
Harivaṃśa, 11, 274 n. 1; as Mahābhārata Indras, 12 n. 8, 492–93, 574
appendix, 571, 577–84; dating of, itihāsa, history, 4–5, 8 n. 25, 57, 83, 191
584; known by Aśvaghoṣa, n. 25, 199, 204–5, 255, 274 n. 1, 277,
646; known in Yuga Purāṇa, 296–97, 335, 337–38, 352, 363, 425
280–82, 284–86, 288–89, 291, n. 31, 508, 596, 666; itahāsaṃ
294, 335, 336 purātanam, 204, 508
736 index

Jainism, 40, 50, 134, 162, 241, 245, 275 karma, 93–95, 123, 145, 195, 207, 221
n. 7, 282 n. 25, 292, 294, 511, 513, n. 123, 226 n. 136, 359, 3397, 433
592 n. 79 n. 52, 454, 458, 464 n. 122, 470,
Jamison, Stephanie, 54 n. 8, 58 n. 16, 208 475, 492–93, 500–1, 506, 510–11,
n. 77, 338–39, 357 n. 43, 359–61, 519, 528–30, 533, 543 n. 48, 561,
374–83, 386, 493 n 29, 605 n. 111 576, 581, 585–89, 631, 636, 662,
Janamejaya, 297–80, 282, 368, 420, 672; and dharma, 52–53, 183, 519;
577–79; aśvamedha of, 280–81, 375 as staining, 511, 515; fruition of,
n. 86, 379, 581–84; hears 31, 509–14, 523, 660 n. 104,
Mahābhārata from Vaiśaṃpāyana, 662. See also karmayoga, svakarma
211, 254–55, 393–94, 425, 436, karmayoga, 207, 214, 351, 517, 519–20,
4422–43, 445–406, 490, 573, 533–42, 553, 570, 662
669–70, 672; last questions of, Karṇa, 22, 279, 384 nn. 103 and 104,
417, 464 n. 22, 510 n. 61, 576, 387–89, 391, 394, 404, 406, 439,
578–80, 583, 610 n. 130, 616; 444, 454–55, 460–66, 620, 646;
snake sacrifice of, 200, 261, 371, a good man, satpuruṣa, 457–60,
373, 616, 617 466; and Kṛṣṇa, 622–24; feet of,
Jāṇussoṇi, 111 n. 26, 112 n. 30, 166 454, 458; Pāṇḍavas’ older brother,
n. 185, 187 454; regrets harsh words to
Jātakas, 245 n. 4; Devadhamma Jātaka, Draupadī, 458
432, 435, 443 n. 77, 444 n. 80; Kauravas, 9–10, 279, 311; birth of,
Jayaddisa Jātaka, 136 n. 99; 386, 390–92; host Kṛṣṇa;
Kurudhamma Jātaka, 135, 137; 620–22; hundred, 311, 337,
Mahāsutasoma Jātaka, 136; Vaṭṭaka 344 n. 14, 394
Jātaka, 634 n. 18; Vessantara Jātaka, Kauśāmbī, 298–332; Kauśāmbī myth,
178, 640 298, 300, 302, 305–6, 312–23, 326,
329, 331–36; alpha-cluster versions
Kālī, 40, 367–70, 372, 382, 440 n. 70. of, 299, 302–5, 307, 310, 312–17, 321,
See also Satyavatī-Kālī 331, 334–35; M-cluster versions of,
Kali, 578. See also yuga 300, 318–33, 335–36; S-cluster
Kalinga, 40, 46–49, 135, 287, 284, versions of, 300, 305–17, 323,
560, 655 326–331, 334–36
Kalki, 32, 288–91, 538, 590–94 Kātyāyana, Prophesy of, 8–9, 302, 306,
kalpa, kappa, 7, 30, 32, 39 n. 11, 42–43, 311–12, 314–15, 336; in prose, 598;
229, 243–73, 276, 292, 296–97, 332, in verse, 303, 311, 315
36, 517, 563, 592 n. 80, 614; Khandhaka, Skandhaka, section
antarakalpas, 246–47, 252; of Vinaya, 106 n. 11,
asaṃkhyeya kalpas, 246–47, 252; 153–54
bhadrakalpa, 246; end of, 43, 249, Kosambī, 300; dispute at, 301, 308
251, 255–58, 670: in Yuga Purāṇa, kings, kingship, 27, 30, 51, 66–78,
276; mahākalpa, 246, 251–52, 256. 189, 227–41, 286, 357, 408, 453,
See also Brahmā, yuga 553 n. 66, 604, 622, 676,
Kalmāṣapāda, 136, 422 n. 25, 679–80; a king’s ideal day, 223–41;
439 n. 66 and tribal chiefs, 56–57, 69, 71,
 index 737

80, 86; founding a kingdom, Kurus, Kuru dynasty, Kuru line, 78,
229–33; legal procedure in royal 80, 82, 134, 193, 213, 265, 284–85,
courts, 46, 79, 92, 122, 158, 213, 294, 296, 345, 358, 367–72, 377,
224, 339 n. 5, 362, 397, 423, 427, 388–410, 525, 555–56, 571, 596–97,
439–40, 454, 498 n. 34, 534, 619, 646, 667
597–99, 619–20, 622; Manu’s Kurukṣetra, 56, 58, 193 n. 34, 313, 349,
king, 229–33; royal fort, 224, 454–59, 572 n. 15, 575 n. 24, 580,
229–33; should be Kṣatriyas, 594, 597–99
240–41; statutory purity of Kuru-Pañcāla, Kurus and Pañcālas, 80,
kings, 239; what a king 82–84, 86, 109, 488
should think, 236–37. Kuru state, 57–58, 75, 598
See also yuga Kuśa and Lava, 377, 419, 483, 613,
Kloetzli, Randolph, 248, 249 n. 12, 615, 636
271 n. 65, 276 n. 10, 336 n. 144, Kuṣāṇas, 6–8, 11 n. 3, 275, 286, 288 n. 35,
353 n. 32 298, 314–16, 330, 625 n. 2
Kristeva, Julia, 340 n. 6, 343 n. 13, 355,
387 n. 113 Lacan, Jacques, 341–43, 349, 358
Kṛṣṇa, 10–13, 21–29; as authority on Lakṣmī, 263, 482, 644, 647–48,
dharma, 21, 463, 491; as avatar, 650–51
413–14, 579 n. 11, 577, 590–91, Law of the Girl, 340
497–99; as guest or host, 620–24; Law of the Mother, 340–44,
embassy to Kauravas, 388, 534, 549, 357–58, 410
597, 620–22; in yatas tatas Laws of Manu. See Manu,
formulae, 452–53, 606; not a Mānavadharmaśāstra
king, 553; restorer and protector legal precedent, 204–5, 420–29
of dharma, 562–63; See also Arjuna, Lévi, Sylvain, 17, 39 n. 11, 545–46
Bhagavad Gītā, Draupadī, Lubin, Timothy, 182 n. 3, 184–86
Yudhiṣṭhira
kṣatriyadharma, kṣatradharma, madhyadeśa, the middle region, middle
kṣātradharma, 361, 476, 526–28, country, 183 n. 34, 283, 312–15, 324,
534–35, 546, 594 n. 84; as role 335, 357, 600–3, 653; majjhimadesa,
models for other classes’ svadharma, 313 n. 94
561. See also svadharma Mādrī, 207, 343 n. 71, 349 n. 15, 374,
kuladharma, 48, 532, 558, 630, 651 382–87, 397–98, 405–9, 449,
Kuntī, 291, 341, 344–45, 374, 382–410, 453–54, 466, 481, 646; called
444, 449, 454, 457–59, 466, Bāhlakī, 207
468, 474, 481, 489–92, 497, Magadha, 86 n. 69, 109, 112 n. 29,
525, 532, 534 158–59, 240, 282–85, 292–96,
Kuru, 83, 555, 598; ordained that 313–14, 334–35, 392, 619, 641,
slain warriors go to heaven, 646–56, 683; “greater Magadha,”
555, 598 94, 137, 193 n. 34, 217 n. 108,
Kurudhamma, Kuruvatta-dhamma, 244–45, 273–74, 282, 294–95. See
135–37; Kurudharma, 558, 599 also Jarāsaṃdha, Paṭaliputra,
Kurujāṅgala, 232, 240, 292 Rājagṛha, Udāyin
738 index

Mahābhārata, 4, 8–9, 29–31, 58, 61, 333 n. 142; Mahāyāna

82–83, 85–87, 90, 198–200, 203–4, Mahāparinirvāṇa Sūtra, 318–19,
210, 245, 337–414, 416–429, 431–55, 332–33; Ugra Paripṛcchāsūtra, 167
457–516, 528, 542–53, 569–99, n. 75, 459, 630 n. 16, 679 n. 143;
602–24; and Bhagavad Gītā, 181, Vimalakīrti Sūtra, 6, 326, 621 n. 144.
512, 519, 553, 557–58; Buddhacarita, See also Candragarbha Sūtra
645–83; and Buddhism, 310–14, 324, Mahendrasena, king of Kauśāmbī, 309,
334–36, 567; and Harivaṃśa, 11, 294, 314, 323 n. 121
571, 577–84; and Mānavadharmaśāstra, mahiṣī, chief queen, 368, 400, 427 n. 38;
11, 100, 181, 184–91, 196–201, in Aśvamedha, 375–77, 605–6; two
204–8, 215–16, 228–29, 232–33, mahiṣīs, 364, 380–81, 394, 400–1
238–40, 292, 524, 529, 535, 594 Malamoud, Charles, 87–89, 234 n. 160,
n. 84, 604; and Rāmāyaṇa, 11, 136, 518, 524, 529
181, 189–201, 210, 429–31, 455–57, Malinar, Angelika, 507 nn. 53–55, 508
599–602; and Yuga Purāṇa, 5, n. 60, 512 n. 68, 540 n. 41, 543 nn. 46
273–97; chronometry in, 245, and 47, 553 n. 66, 557 n. 74, 573 n. 18
259–69, 286–92, 334–35; paradigm Mālyavān, 600
shift on dharma in, 20–29; Pune Mānavadharmaśāstra, Manu Smṛti,
Critical Edition of, 11–20, 253 n. 28, Laws of Manu, Manu, 8, 30, 444;
346, 433, 442, 451 n. 95, 453 n. 99, and Rāmāyaṇa, 205; in verse,
460, 469, 472 n. 134, 528, 576, 578, 196–98, 201–6; critical edition of,
581, 591, 602, 668, 676 n. 132, 688; 20; frame story of, 196–97, 199,
“Vulgate,” 15–18, 433–34, 442, 451 207–12, 214, 528; innovations of,
n. 45, 472 nn. 134 and 136, 653. 196–208; mokṣa as renunciatory
See also bhakti swerve, divine plan, asceticism in, 217–18, 660,
itihāsa 662–63, 667, 670; narrative in,
Mahadevan, Thennilappuram (T. P.), 227, 229–30, 233, 238, 339–40,
19–20, 22 n. 35, 296, 613 518, 524; on karmayoga, 535–40,
Mahāparinibbāna Sutta, 106, 123–24, 128, 553; on mixed castes, 526–28; on
158, 175, 249 n. 16, 411. See also Ten-Point Law, 522–23; on slain
Mahāyāna Mahāparinirvāṇa Sūtra warriors going to heaven, 560;
Mahāsammata, 108–9, 161, 174, 179 regarding women, 224–26,
Mahāsatipaṭṭhāna Sutta, 134–35, 139 339–40, 524–25; resistance to
Mahāyāna, 6, 40 n. 12, 103, 107, 126, 141, bhakti swerve, 569, 585; revolution
151, 154–55, 162 nn. 173 and 175, 246, in, 224, 226; Śū d ras in, 224–26 ,
248–49, 252, 273, 298–309, 318–33, 527 , 565 . See also dharmasū t ras
460, 630 n. 16, 680 and dharmaśā s tra , kings ,
Mahāyāna sūtras, 103, 107, 155, 305 n. 73, Mahā b hā r ata , Manu , Olivelle
327, 330; Laṅkāvatāra Sūtra, 244 Man-lion, Narasiṃha, 590
n. 3; larger Sukhāvatī Vyūha Sūtra, mantra, 52, 65, 80–81, 99, 100, 113, 116,
250 n. 16, 298 n. 52; Lotus Sūtra, 6, 118, 123, 177, 320, 327, 358, 620
298 n. 52, 331 n. 137; Mantra period, 75, 78–80, 210
Mahāprajñāpāramitā Sūtra, 141 Manu, 10, 182; as primal lawgiver,
n. 112; Mahāyāna Sūtra, 299 n. 56, 200, 209, 268; Manu Prācetasa,
 index 739

521 n. 69; Manu Svāyaṃbhu, son Mokṣadharmaparvan, third anthology of

of Svayaṃbhū, 209, 267, 360 Śāntiparvan, 218 n. 110, 220, 234
manvantara, interval of Manu, 7, 245, n. 160, 254, 264, 428, 446, 497
260, 267–70, 289 n. 38, 571 n. 34, 529, 534, 536, 570 n. 6, 607,
Māra, 133 n. 13, 156, 629 n. 15, 641–42, 611 n. 119, 660–73
646, 659, 677–78, 680
Mārkaṇḍeya, 278, 281, 288, 434, 436, Nāciketas, dialogue with Death,
538, 590–92, 611, 615, 682–84 98–100, 213
Matilal, Bimal Krishna, 21–29, 443 n. 53, Naimiṣa Forest, 199, 210–11, 253, 292,
461 n. 118, 471 n. 133, 445 n. 5, 294, 373, 425, 556, 613, 623
556 n. 71, 559 n. 77 Nakula (Pāṇḍava), 115 n. 37, 189 n. 22,
Mauryans, Mauryan dynasty, 6–7, 14 279, 443, 449, 451, 466, 468
n. 14, 35, 38, 45–50, 57 n. 13, 79, 104, Nara and Nārāyaṇa, 259–61, 548,
106 n. 11, 151 n. 36, 161, 165, 168, 633, 629
175–78, 184, 190, 258, 277, 283, 286 Nārada, in Harivaṃśa, 577, 595 n. 88; in
n. 32, 294, 297–98, 570 n. 4, 584, Mahābhārata, 188 n. 19, 234 n. 157,
604, 623–24, 650 n. 73 252, 259, 354 n. 53, 428 n. 44, 436
meditation, Buddhist, 109, 125–26, 130, n. 62, 481, 544, 550–51, 553, 573
134, 140–42, 14, 148–49, 161, 175, 177; n. 17, 580, 590, 598–99, 602,
Aśoka’s emphasis on, 30, 38, 43, 607–8, 611, 619–21, 658, 668–69;
48–49; Manu on, 214 in Rāmāyaṇa, 415–18, 424, 519
Megasthenes, 45, 292, 294 Nārāyaṇa, 251–53. 255–57. 259–60, 262,
Mīmāṃsā, Mīmāṃsāsūtra, 7, 181, 183, 265, 267, 270–71, 277, n. 13, 379,
208 n. 17, 512 390, 510 n. 62, 590–91, 607–8, 615,
mirror, as Mahāyāna symbol, 325–26, 619, 669–71; Nārāyaṇa Gopā
333, 336 warriors, 620; Nārāyaṇa weapon,
Mitchiner, John E., 274–91, 345, 460, 478, 615
584 n. 46 Nārāyaṇīya, 25264, 270, 2272, 277 n. 13,
Mitra, 58, 64, 77–78, 214 n. 98; and 335 n. 36, 379, 420, 428 n. 44,
Varuṇa, 56, 66–72, 76–77, 81–82, 523, 539, 543, 548 n. 55, 570 n. 6,
371, 378–79 579, 5883, 590, 657 n. 97,
Mitra kings, possibly including 664–74
Puṣyamitra Śuṅga, 284–86, 295–97; Nāstikas, nāstikya, heterodoxies, heresy,
Mitra period, 275 25, 86 n. 69, 88, 194, 224, 240,
Munidharma. See Ṛṣidharma, 244–45, 289, 339, 506, 508–15, 518,
Munidharma 522, 569, 643
Mūlasarvāstivādins, 154–57, 162 n. 73, Nattier, Jan, 44 n. 27, 154 n. 149, 162
300, 305, 307 n. 78, 314. See also nn. 173 and 175, 244, 246–50,
Vinaya. Sarvāstivādins 298–333, 336, 457 n. 110, 630 n. 16,
mokṣadharma, 254–55, 428, 681–82; in 663 n. 26, 679 n. 143
Aśvaghoṣa’s Buddhacarita, 626, Nikāya schools, 104–7, 124, 153–55,
656–60, 673–79, 682–83; in last 162–63, 248, 273, 305, 330
four units of Mokṣadharmparvan, Nīlakaṇṭha, 17, 199. n. 53, 576.
660–73 See Mahābhārata “Vulgate”
740 index

nirvāṇa, nibbāna, 10, 121–22, 146 n. 122, 103–6, 194 n. 42, 344 n. 1, 249, 273,
149, 250–51, 254, 319–20, 327, 637, 298 n. 53, 300, 305, 313 nn. 93 and
640, 644, 646 n. 63, 659, 663, 94, 321–23, 519 n. 2, 648, 678
670, 682–83; Brahma-nirvāṇa, ̣ ạ vas, 9–10, 207, 232, 240, 279, 337–38,
Pānd
565, 567 355, 393, 413, 428, 436–37, 442–43,
nivṛtti, nivṛtti dharma, 253–57, 260, 498, 452, 461, 467–70, 482–83, 489–93,
523, 534, 539–40, 567 n. 82, 669–73. 506, 538, 551, 572, 591, 598, 615,
See also pravṛtti, pravṛtti dharma 648–55, 678, 682; fatherless, 342–43;
niyoga, 349, 358, 364–72, 376, 378, 380, introduced at Hāstinapura, 403–10;
397–403, 406, 426, 496–97 Mount Pānd ̣ ạ va, 648–53. See also
non-cruelty. See ānṛśaṃsya Draupadī, svayaṃvara and polyandry
non-independence of women, 309, of; Fitzgerald, “invention” of
338–39; free females, svatantrā, pārājikā, four offenses requiring
387–88, 402–3, 410, 498 expulsion, 155–56, 172, 300, 328
nonviolence. See ahiṃsā Parthians, Indo-Parthians, Pahlavas,
nuns, 40, 44, 107, 121, 151, 154–57, 274–75, 288 n. 35, 298, 302, 309,
169, 174 n. 214, 250, 321 n. 113, 314–17, 328, 335
339–40, 498 paṭa (scenes of Mahābhārata as seen on a
woven cloth), 344 n. 14
Oberlies, Thomas, 12 n. 6, 64 n. 27, Paṭaliputra, Magadha capital, 282–86,
582 n. 39, 592 n. 71, 613 n. 125, 309, 324 nn. 123 and 125
669 n. 118 pativedana, husband-finding rite/spell,
Olivelle, Patrick, 3, 9, 11 n. 3, 20, 3–38, 336, 381–82, 385 n. 106, 398
43–55, 66, 79, 81–88, 91–100, 108, pativratā, devoted wife, 262 n. 45, 344
115, 119 n. 51, 151–52, 154, 165, n. 15, 401, 427 n. 38, 470 n. 110, 481
181–86, 189–203, 206–30, 233, 235, n. 1, 507 n. 54
239, 241, 266, 316 n. 100, 338 n. 2, Placer and Ordainer. See Dhātṛ and Vidhātṛ
520 n. 6, 525, 535, 537, 539, 628, Pokkharasāti, Puṣkarasārin, 112, 114,
643–48, 652, 660–63, 667, 675; 116–18, 120–21, 125, 166–67, 329
early dharma linked with king and Pollock, Sheldon, 18 n. 25, 64 n. 26, 100
Varuṇa, 66, 92 n. 81, 108; innovative n. 23, 222 nn. 124 and 124, 412 n. 2,
character of early Buddhist usage 506 n. 51, 590 n. 72; on Rāvaṇa’s boon,
of dharma, 50–51, 54, 108, 189; 222 n. 125, 414 n. 6, 487, 501 n. 42, 571
on dating dharmasūtras, 189, 623 possession, 4, 40, 340, 367, 393 n. 128,
n. 148; on dating Manu, 206–7; on 401, 435, 438–42, 445–47, 452,
“excurses” in Manu, 12. n. 6, 20, 211 463, 465, 498, 501, 558, 581
n. 87, 214 n. 95, 266, 537, 539 prādurbhāva, 256, 277 n. 15, 590 n. 71, 671
orality and writing, 14, 18, 20, 111 n. 26, prāpti, possession, 143, 145–46; aprāpti,
176–78, 615. See also writing 146; Prāpti, wife of Tranquility
(Śama) and daughter-in-law of
Pāli, 6, 97, 106–8, 114, 125, 140, 142, 145, Dharma, 263–64; a daughter of
147, 152, 154–55, 177, 186–88, 246, Jarāsaṃdha, 264 n. 48
305, 313, 318 n. 103, 322, 407, 411, Prātimokṣa, Pātimokkha, 152–58, 163,
678 n. 140, 683; Pāli canon, 6, 40, 301–7, 326, 328
 index 741

pratirūpaka-dharma, semblance dharma, 603, 605–6, 616–19; departures of,

imitation dharma, 318–23, 333 570 n. 3, 603, 617 n. 136; friendship
n. 240; saddhamma paṭirūpaka, with Sugrīva, 430–31, 623 n. 147;
counterfeit dhamma, 321–22 gives refuge to Vibhīṣaṇa, 626;
pravṛtti, pravṛtti dharma, 253–58, 260, in divine plan, 265 n. 54; killing of
523, 539–40, 567 n. 82, 670–71, 673. Vālin, 431, 453–56, 477, 479–80, 623
See also nivṛtti, nivṛtti dharma n. 147; kills the Śūdra Śaṃbūka, 20,
prison, imprisonment, 42, 46, 48, 225, 519, 569 n. 2; knows himself, 500,
504, 649, 652–53 503; perfect man, 10, 21, 24, 29, 211,
Puruṣa, 65, 74–76, 78, 113, 337 n. 1, 544, 414–15, 419, 501, 516, 569; split
551, 563, 587, 605; Supreme Puruṣa, personality of, 414. See also Sītā,
214, 267 Yudhiṣṭhira
Puruṣamedha, 84, 187 n. 16 Rāmāyaṇa, 10–14, 18–21, 26, 164, 181,
Puruṣasūkta, 54, 58, 78, 84, 87, 113, 167, 191, 198, 200, 206 n. 74, 209–11,
216, 221, 532, 594 219, 232, 234, 286, 328, 429–31, 438,
Puṣyamitra Śuṅga, 283–86, 311 n. 88, 584 453, 485–86, 498, 507, 519, 526, 532,
putrikā, 262, 264, 344 n. 15, 602 n. 101 536, 576, 586–97, 600–3, 611, 613,
615, 619, 623–24; and Buddhacarita,
Quinquennial assembly, 310, 324 626, 636, 638–46, 650 n. 75, 658,
660, 675 n. 136, 682; and
rājadharma, dharma of kings, 197, 221, Mānavadharmaśāstra, 11, 26, 190,
226, 228–33, 238, 353–54, 466 200–1, 204, 569, 585, 662; Baroda
n. 124, 522, 535, 594, 607, 672, Critical Edition of, 12 n. 4, 20; divine
676–79, 681–82 plan of, 571–75. See also
Rājadharma Parvan, 26 n. 44, 56 n. 13, Mahābhārata, upākhyānas
205, 227, 238 n. 19, 354 n. 33, 428, Rāvaṇa, 10, 281 n. 20, 350, 388 n. 90,
521 n. 9, 656 n. 94, 672, 679 380 n. 97, 414, 419, 430, 455, 482,
Rājagṛha, Rājagaha, Magadha capital, 40, 486–87, 499, 501–2, 504, 532,
159, 647–49, 680 573–74, 595, 600–1, 637, 639; boon of
Rājasūya, royal consecration ceremony, being invulnerable to men, 222 n. 125,
81–83, 85, 86–87, 91–92, 109 n. 22, 414 n. 6, 487, 501 n. 42, 570–71, 573
175 n. 216, 188 n. 19, 231–32, 285, republics, republican polities, 122,
311, 378, 572, 580, 583, 598, 605, 158–61, 216
607 n. 64, 651, 656; dangers of, 232, right effort, 49, 128. 132–34,
337 n. 90, 580, 599–600, 619 139–40, 146
Rājpūts, 240, 280, 286, 295 n. 46, right mindfulness, mindfulness, 49, 128,
342 n. 10, 655 132–34, 137–40, 160
Rāma, 10, 12, 21, 24, 29, 31–32, 164, 277, Ṛgveda, 5, 30, 51–84, 86, 90, 100, 109,
412–24, 429–31, 453, 455–57, 477, 201, 293, 379, 605–6, 617, 633;
479–80, 482–84, 486–88, 493–506, family books of, 55–58, 60, 68–69,
516, 587, 595–96, 599–602, 615, 83, 165, 617
619–21, 623, 637–46, 659–60, Ṛ ṣidharma, Munidharma, 607, 609–12,
675 n. 30; as avatar, 10, 222 n. 125, 615–16, 682 n. 15
585, 590, 599–600; as guest or host, Ṛṣisaṅgha, 608–9, 613–15, 621
742 index

ṛta, 52–53, 55, 59, 61 n. 21, 66, 70 399–400, 426; reversal of law on
Rudra, 92, 254, 375–77, 381–82, 651 women’s sexual freedom, 426; sex
change, 490 n. 21, 601; sexual
saddharma, true dharma, 4, 44 n. 27, 249, continence of Pāṇḍu, 392
307 n. 28, 318–25, 629, 643; in a Shulman, David, 15, 18 n. 25, 210 n. 85,
Brahmanical usage, 556 420, 424, 432–37, 441–44, 448–52,
Śakas, 274–75, 284, 287–88, 295, 583 n. 8
297–98, 302, 314–16, 328, 554 n. 29 Śiṣyaka, 301–4, 306–9, 324–25, 333; as
Śakuntalā, 19 n. 30, 82–83, 213, 337 n. 1, Tripiṭaka master, 307, 324,
339 n. 5, 360, 380 n. 97, 400 n. 142, 333 n. 142; son of Agnidatta, 308,
410, 427, 459 n. 99, 589 n. 61, 602. 314, 324 n. 125
See also upākhyānas Sītā, 10, 482–84, 571, 573–74, 587 n. 53,
Sakyans, 108–10, 114–15, 121–22, 133, 158, 600, 603, 605 n. 11, 616–18;
160, 168, 179 abduction of, 412–13, 430, 71,
Śākyamuni, 179, 246, 249–50, 320 587 n. 52, 617; as Vedavatī in previous
Śaṃtanu, 346–58, 646 n. 51. See also life, 486–87; banishment of, 10, 377,
Bhīṣma Śāṃtanava 400 n. 142, 417–19, 483, 536 n. 32,
satstrī, good woman, 457 n. 110. See also 587, 589 n. 60, 602; birth of, 483–84,
Karṇa, as satpuruṣa 486–38, 494 n. 30; enters into the
Sarvāstivādins, 124, 12 n. 67, 127, earth, 211, 417, 420, 516; in captivity,
140–47, 149, 142 m. 193, 154–55, 438 n. 65, 486, 498–506; incarnation
157 n. 159, 250 n. 17, 263–64, 300, of Lakṣmī, 482; svadharma of, 495;
322; and S-cluster version of svayaṃvara of, 493–94
Kauśāmbī myth, 305–17; fivefold Śiva, 26 n. 44, 40, 188 n. 19, 193 n. 33,
taxonomy of, 143. See also 200 n. 55, 226 n. 138, 251, 263,
Mūlasarvāstivādins 276, 281–84, 289, 295, 335–36,
satya, true, truth, real, 62, 66–67, 92, 346, 348, 367–69, 381–83, 386,
453, 462, 465, 503, 522 n. 13, 610 389–90, 395 n. 132, 442 n. 74,
Satyavatī, Satyavatī-Kālī, 354, 367–74, 448 n. 90, 488–89, 492–93,
382–83, 646 n. 91 610 n. 62, 528–29, 533, 572,
Schopen, Gregory, 104 n. 5, 106 n. 11, 595, 601, 607–9, 611–15, 618.
151–53, 177, 308–9, 318 n. 102, See also Rudra, Umā-Maheśvara
329 n. 133, 435 n. 58, 460 n. 114 Saṃvāda
sekkhiya, śaikṣa, monastic “training” six jobs of Brahmins, 188, 475, 520–21,
rules, 121, 124, 152, 157, 159, 161, 525, 539, 594
169, 248, 304 n. 71, 321 n. 113, 325 snātaka, bath-graduate, 114, 144, 165, 187,
sex, 110, 135, 151 n. 137, 154–56, 163, 168, 208 n. 77, 211, 234–35, 241; and
171–74, 187, 225, 237, 249, 263, 281, king, 235–36, 239–40, 518; Pāli
339, 355, 375, 378, 421, 484, 491, n(a)hātaka, 187; Kṛṣṇa, Bhīma, and
536 n. 32, 605; as gāmadhamma, Arjuna disguised as, 240, 650–51
village dhamma, 155; in Aśvamedha, Soma, soma, 55–58, 60, 64, 66, 70–73,
375–82, 400, 605; as visaladhamma, 76, 78, 82, 91–92, 94, 185 n. 8,
vile dhamma, 155; maithunadharma, 262–64, 359, 400, 580, 601 n. 98,
law of sex, 264, 407; necrophilia, 606, 609, 634
 index 743

śramaṇas, samaṇas, 41, 45–47, 49–50, 195, 248, 300, 411, 671. See also
113 n. 33, 117, 160 n. 169, 163, 167, Aggañña Sutta, Ambaṭṭha Sutta,
175, 218, 247, 632, 640; from soles of Mahāparinibbāna Sutta,
Brahmā’s feet, 113, 117, 167–68, 329 Mahāsatipaṭṭ̣hāna Sutta
Śrī, 85, 482–83, 488, 491–93, 516, svabhāva, 533 n. 26, 536; in Bhagavad
523 n. 25, 548, 599, 607, 651 Gītā, 517, 536, 540–41, 554, 559,
strīdharma, women’s dharma, 207, 561, 565; of demons, 53 nn. 26 and
276 n. 136, 338–40, 481. See Law 27; of Kuntī, 389; of women; 533
of the Mother, śuśrūṣā n. 27, 538 n. 36; of Yudhiṣṭhira, 372
Stietencron, H. von, 230 n. 142, 239, n. 80, 452, 464, 510, 534;
270 n. 59, 227, 292–94 Sarvāstivādin usage, 141, 143–45,
Strong, John, 36, 43, 143–47, 160 n. 169, 517 n. 1
246, 305, 310–11, 684 Sutta Vibhaṅga, Sūtra Vibhaṅga, section
stūpas, dharmarājikās, 109, 152 n. 139, of Vinaya, 153–56
304 n. 71, 310, 592, 693–84. See also svadharma, 26, 39, 53, 216, 517–35; in
eḍūkas Bhagavad Gītā, 527, 535–36, 540,
Śūdras, 22 n. 35, 75, 87, 92 201–2, 542, 553–54, 449–60; in Buddha-
205 n. 71, 215–16, 220, 224–26, 241, carita, 677–78; in Kṛta yuga, 274
264, 274, 288 n. 37, 291 n. 40, 293, n. 2; of animals, 535; of Arjuna,
339, 343, 366, 372, 382, 395, 519–21, 478, 535–36, 553, 560; of āśramas,
527–30, 532, 536 n. 32, 548, 561, 565, 521 532; of Brahmins, 211–12, 216,
569, 578, 589 n. 59, 592, 594, 638, 474, 476–77, 524, 529; of Death,
656 n. 96; Śūdradharma, 527–28, 533; of demons, 529–30, 532–33;
561. See also śuśrūṣā of kings, 534–35; of Kṣatriyas, 352,
Śuka, 218, 234 n. 157, 392 n. 127, 521, 525, 535, 561–62, 565–67, 594,
446 n. 85, 550 n. 43, 664, 666–70, 678; of jātis, 532; of trimūrti, 533;
672, 675 n. 131 of women, 372 n. 80, 481, 484,
Sulabhā, 497–98, 534, 664–66, 675 495–98, 533–34; of Yudhiṣṭhira,
Śuṅgas, Śuṅga Dynasty, 8–9, 42 n. 20, 65, 449–50 (in plural), 461 n. 118;
191, 193 nn. 34, 283–86, 296 n. 48, overseen by kings, 221–22,
311 n. 88, 584, 594, 650 n. 73. See also 227, 521–22, 535, 570 n. 4;
Puṣyamitra Śūnġ a Śrautasūtra usage of, 88, 183–84;
Sūrata, the last arhat, 301 n. 63, 303, svadharman in Ṛgveda, 58, 73 n. 68,
306–10, 313–14, 324–25, 334 519 n. 2
śuśrūṣā, obedience, liking to hear, virtue svakarma, 188, 216, 274 n. 2, 517–18, 527,
of women and Śūdras, 226 n. 136, 548, 554, 588; of Brahmins, 474,
355 n. 35, 594 520–21, 524–25, 562, 594. See also
Sūtra Piṭaka of northern Nikāya Schools, six jobs of Brahmins
107, 162 n. 73, 273, 305, 314, 323. svādhyāya, personal Vedic recitation,
See also Mahāyāna Sūtras 87–88, 185, 189
Sutta Piṭaka, 106–7; Suttas, 105–8, 111,
113–15, 118–24, 128, 134 n. 92, 135, Thapar, Romila, 6, 18 n. 25, 35–44, 57,
140, 142, 152, 158, 161–68, 158 n. 162, 283 n. 26, 295 n. 46,
125 n. 216, 178, 186–88, 191 n. 25, 422 n. 25
744 index

Theravāda, 103, 107, 129, 142, 145–56, 428 n. 41, 429–31, 436, 482, 487, 493,
148, 152, 154–56, 162 n. 173, 495, 576, 590, 595, 600, 609, 617–18,
163 n. 178, 301, 304–5, 318 n. 103, 640–41; Śakuntalā-Upākhyāna,
322, 627, 640 n. 42, 680 19 n. 30, 82, 427; Yakṣa’s questions
thirty-two marks of a Great Man, 112–13, as upākhyāna clearing house, 435
118–19 Upaniṣads, 5, 79, 85, 91–101, 452, 512 n. 69,
tripiṭaka, three baskets, 30, 104–7, 522, 556, 633; Bṛhadāraṇyaka Upaniṣad,
142, 300, 304, 307, 324, 80 n. 50, 91, 93–95; Chāndogya
333 n. 142, 330 Upaniṣad, 91–94, 98, 100 n. 105;
Tokunaga, Muneo, 12 n. 7, 203 n. 64, Kaṭha Upaniṣad, 98–100, 433, 556;
205, 508 n. 58, 657–58, 663 kings in, 97–98, 101; Śvetāśvatara
trivarga, triple set, 223 n. 131, Upaniṣad, 516 n. 76, 613–14
228 n. 145, 234–35, 237, 383, 392, upāyadharma, dharma of strategy, 656
630, 645, 651–52. See also four urbanization, second, 81, 86, 91, 109,
Puruṣārthas 158, 160; attitudes in dharma texts
Tryambaka, 376–77, 381 regarding cities, 109, 175, 185–86,
Tsuchida, Ryutaro, 109 n. 22, 111–12, 188, 200, 297, 324 n. 81, 512, 549,
114, 117–20, 136 n. 95, 187–89, 577 n. 30, 578, 610. See also villages
193 n. 37, 584
Vaiśyas, 75, 92, 115, 216, 264, 291 n. 40,
Udāyin, friend of prince Siddhārtha, 634, 385, 493 n. 29, 521, 525–30, 548, 561,
637–38, 640, 680 594; vaiśyadharma, 528–29
Udāyin, Śiśunāga king, founder of Vajjīs, 108 n. 18, 109 n. 19; laws of,
Magadha capital at Paṭaliputra, 275, 158–60
282, 285–86 Vālin, 10, 400, n. 142, 413; ambushed by
Ujjain, 275, 282, 285–88, 290, 295 Rāma, 414, 416, 418, 423, 431, 453,
Umā-Maheśvara Saṃvāda, 528, 607, 611, 455–56, 477, 479–80, 623 n. 147
614–15 Vālmīki, 10, 200, 209–11, 400 n. 42, 421,
uñchavṛtti, mode of gleaning, gleaners, 482–83, 486–88, 493–94,
436, 609, 611–14, 618, 664, 501 n. 92, 587, 600, 605 n. 111,
666–69, 671 n. 124, 682 n. 151 616–17; in Aśvaghoṣa’s Buddhacarita,
upākhyānas, Mahābhārata subtales, 635–36, 639
423 n. 28, 425–28, 435–37, 446, 450, Vāmana, Dwarf, 590
459, 476, 483–84, 492, 596, 612, vaṃśānucarita, 277–78, 282,
658, 682; legal precedents and 292–94, 297
postulates in, 420–21, 426–28; mirror Vapuṣṭamā, wife of Janamejaya, 280–81,
stories, certain upākhyānas as, 395, 581–82, 584
426, 484; Nakula-Upākhyāna, 425 varṇa, varṇadharma, see caste, class, social
n. 35, 435; Pañcendra-Upākhyāna, 437 class
n. 64, 483, 574 n. 23; Pativratā- varṇāśramadharma, class and life-stage
Upākhyāna, 201 n. 60, 226 n. 136, dharma, 39 n. 11, 215–27, 254, 559,
450, 476 n. 132, 480 n. 145, 682; 612, 616, 672; king as protector of,
puzzle pieces, some upākhyānas as, 221–27
378, 428, 436, 438; Rāmopākhyāna, Varuṇa, 51–52, 55–58, 63–64, 66–73,
Rāma-Upākhyāna, 417 n. 12, 76–79, 81–83, 85–87, 92, 213 n. 93,
 index 745

240, 371, 378–79, 431–32, 570, n. 25, 493, 569 n. 1, 577–78, 585,
601 n. 98 587–88, 590–91, 593, 595, 599–600,
Vasiṣṭha Dharmasūtra, 189, 192–93, 605–6, 608, 611, 12, 618,
202–7, 230 n. 48, 239–40 623 n. 147, 645, 652, 654; and
Veda, 51–104; Vedic canon, 30, 51, 79, 83, Brahmā, 251, 267; and Śiva, 251, 493,
91, 108, 297, 333, 519; Vedic ritual, 7, 620 n. 142; as Horse’s head,
53, 99, 157 n. 157, 171 n. 209, 193 n. 35, 257 n. 36; in Bhagavad Gītā, 563; in
203, 377, 586, 605; Vedic “universe of Ṛgveda, 64, 72, 76–78; vyūhas of,
sounds,” orality, 19, 54, 56–57, 93 259–61. See also avatar
Vidura, 18 n. 25, 237, 344 n. 14, 372–74, Viśvakarman, 73–74; fashioned Sītā in
379, 383–84, 386, 392, Rāmopākhyāna, 487
399 n. 14, 408–9, 439, 448 n. 92, Vratyas, 240, 275
547, 573 n. 18, 604, 620–22, Vrātya Kṣatriyas, 216, 295 n. 6
638, 682; incarnation of Vyāsa, Kṛṣṇa Dvaipāyana Vyāsa, 9–13,
Dharma, 344, 426, 439, 547, 253, 255–56, 344 n. 14, 354–58,
574, 578, 597, 620 366–93, 397, 404, 408–9, 468,
villages, 111–12, 117, 121, 138, 175, 177, 187, 483, 535, 543, 547–55, 572 n. 15,
441, 534, 593, 597; village dharma, 574–75, 579–84, 590–91, 608,
155, 407; village pastoral ideal, 671–72, 678 n. 141; advises
185–86, 577 n. 30. See also Yudhiṣṭhira, 26, 420, 454, 463,
urbanization 465–66, 472, 476, 681–82; and
Vinaya, Vinaya Piṭaka, 124, 105–8, 121, Kṛṣṇa, 10, 13 n. 11, 24–26, 637; and
150–79, 190, 298, 300–1, 304, Śuka, 218, 234 n. 157, 446 n. 85,
307–8, 322, 324–25, 328–29, 435, 666–67, 669–70; as author-poet,
648 n. 68, 679 n. 143; and 9–10, 136 n. 96, 190, 199, 209–10,
dharmaśāstra, 151–52, 161, 164–65, 238, 255, 259 n. 39, 352, 427, 482,
167, 175–76, 179, 328; 536 n. 32, 555, 563; as Veda-Vyāsa,
Dharmaguptaka Vinaya, 366, 379, 672; at Janamejaya’s
152 n. 139, 154–55; jurisprudence in, Aśvamedha, 579–84; at
160; Mahāsaṅghika-Vinaya, Janamejaya’s snake sacrifice,
152 n. 139, 154–57, 159; Mahiśāsaka- 373, 420, 579, 615–16; birth of, 355,
Vinaya, 151, 154–55; 368; foretells bad times in
Mūlasarvāstivāda-Vinaya, 109, 152, Mahābhārata, 409; in Aśvaghoṣa’s
154–57, 162 n. 173, 300 n. 59, 308, Buddhacarita, 635–38, 646, 675
329 n. 133, 330, 435 n. 58, n. 131; prompts Draupadī’s
648 n. 68; Sarvāstivāda Vinaya, polyandry, 483,
152 n. 139, 154–55, 157 n. 159, 389–93, 510 n. 62; prophesies
300 n. 159, 322; Secret Vinaya Kali yuga in Harivaṃśa,
(Guḷhavinaya), 322; six extant 274 n. 1, 281–82, 284, 286,
Vinayas, 151, 153–55; Theravāda 288–89, 355, 580–83, 591;
Vinaya, 154, 301 sires Dhṛtarāṣṭra, Pāṇḍu,
Viṣṇu, 26 n. 44, 32, 214 n. 98, and Vidura, 136 n. 96,
222 n. 12, 251, 257, 262–65, 201 n. 60, 342–43, 370–73;
269–70, 272, 279, 281 n. 20, 377 unveils secret of the gods, 573
n. 89, 390, 414, 483, 486–87, 491 Vyuṣitāśva, 400–1, 426, 581 n. 37
746 index

women and dharma, 40, 44, 107, 440–41, 445; 447; dialogue with
121, 151, 154–57, 169, Naciketas, 98–100, 213
174 n. 214, 224–26, 250, 262, 264, Yogasūtra, 100 n. 106, 149, 538, 540
276 n. 136, 309, 321 n. 113, Yudhiṣṭhira (Pāṇḍava), 4, 9, 12, 21–29,
334–35, 338–40, 347, 356, 362, 131, 164, 188n. 19, 189 n. 22, 194
364, 368, 375–77, 380–81, 387–88, n. 39, 200 n. 5, 219 n. 113, 236–
394, 398, 400–3, 410, 426–27, 37, 279–80, 285, 291, 372 n. 80,
457 n. 10, 470 n. 38, 481–82, 379, 393–94, 420, 449–50, 452,
498, 507 n. 54, 524–25, 454, 461, 464–66, 476, 510, 534,
533 n. 27, 538 n. 36, 594, 602 580, 591, 594, 603 n. 105,
n. 11, 605–6. See also dharmapatnī, 604 n. 107, 637 n. 35, 658, 678
non-independence of women, n. 141, 681–82; and Arjuna, 22–23,
Mānavadharmaśāstra, nuns, 223, 266, 403, 405, 422 n. 23, 426
pativratā, putrikā, satstrī, n. 62, 455, 462–71, 474, 479–80,
strīdharma, śuśrūṣā, svadharma 492, 535, 544, 547, 550–51, 588,
of women 590, 659 n. 102, 665; and Aśoka
Witzel, Michael, 19, 55–58, 75, 78–80, Maurya, 412; and Draupadī,
83–86, 91, 108–9, 161, 165, 490–92, 496, 506–16, 586; and his
182 n. 3, 191, 2286, 297, father Dharma, 21, 24; 260, 343,
318 n. 100 512 n. 12, 523, 574, 578; and Karṇa,
Writing, 13–14, 18–19, 35, 37, 57–58, 454–55, 459, 464–65, and Kṛṣṇa,
177–78, 190–91, 198, 201–2, 21–29, 463–65, 468–69, 478, 597,
208–9, 286–87, 292, 603, 622, 624; and Rāma, 226, 238,
297 n. 49, 307 n. 78, 330, 334, 411–80, 484; and Vidura, 372 n. 80,
504 n. 46, 608, 615, 627, 635, 574, 578, 597, 620; as Dharmarāja,
637–38, 648–49, 674. See orality 4, 9, 27–28, 253, 417,431, 435, 440,
and writing 475, 552, 651, 655, 677; Aśvamedha
Wulff Alonso, Fernando, 296–97, of, 583; birth of, 391, 404–5, 409;
379 n. 96, 575–76 gambler, 228, 231–32; lies to
Droṇa, 453–55, 457–79; postwar
Yājñavalkya, 74, 80 n. 50, 84 n. 62, curriculum of, 227, 229, 252–53,
85 n. 66, 93–97, 99–100, 125; and 349, 596, 607, 610–12, 658–75;
Maitreyī, 95–96, 125 677–79, 681–82; praises the “Ṛṣis’
Yajurveda, 278–81, 83, 208 standard,” 512; Rājasūya of, 580,
Yakṣas, 303, 307, 530 598, 600, 603 n. 105, 605, 619,
Yakṣapraśna, the Yakṣa’s Questions, 25, 647, 651, 654–56
413, 425–26, 428–29, 431–53, yuga, yugas, ages, 5, 13, 72, 200 n. 57,
466, 484, 492 n. 28, 511 n. 64, 523, 260, 264, 267–97, 333–36; and
573, 598 dice, 203 n. 45, 312, 350, 548–49,
Yama, 9, 92, 239–40, 260 n. 42, 364, 562–63, 571, 575–77, 590; and
386 n. 109, 393 n. 128, 407, 417 Greeks, 296–97, 336, 575; and
n. 15, 426, 445–47, 417 n. 15, 451, history, 259, 274, 295–96; and
492–93, 556; as Pretarāja, a kalpas, 7, 39 n. 11, 229, 244–45,
possessing spirit, 435, 437–39, 250 n. 19, 252, 258, 265, 268,
 index 747

270–73, 276 n. 11, 296, 517; 238, 260, 271, 291, 535, 600; Kṛta
Buddhist usages and references to, yuga, 29 n. 11, 271 n. 62, 277,
244–45, 31, 335; caturyuga, 287–91, 295, 312 n. 91; pre-epic
mahāyuga, 244 nn. 2 and 3, 267, usages of, 293–94; Tretā yuga,
271, 274, 295, 312, 594; Dvāpara 269–70, 273 n. 2, 277, 291, 591,
yuga, 269–70, 273 n. 2, 277, 291, 594; twilights and transitions
296, 407 n. 153, 577–78, 591, 594; between, 31, 265, 282, 312, 338,
end of a yuga, 244 n. 1, 256, 265, 548, 571
271, 278–80, 286–90, 295, 443, Yuga Purāṇa, 5, 8, 11, 243, 271, 273–97,
583, 591–92, 671; in land of 316, 578, 580, 584, 591, 594, 648
Bhārata, 271; Kali yuga, 32, 245–47, yugadharma, 32, 189 n. 20, 265–66,
269–97, 335, 337, 434, 437 n. 63, 268–70, 294–95, 336, 437 n. 63,
520, 548, 571, 577–83, 591, 619, 535, 583
683; kalivarjyas, 270 n. 59;
king makes the yuga, 217, 230, Zen, 541