Yungang

The first-ever comprehensive examination of its kind in any western language,

this unique volume provides a social art history of Yungang: a 5th-century
rock-cut court cave complex, UNESCO World Heritage site, and one of the
greatest Buddhist monuments of all time. Yungang asks why, when, and
under what circumstances this impressive cave sanctuary was made, who
played significant roles at various stages, and what was the construction
dating sequence.
Recent economic changes in China including the expansion and
reconstruction of roadways have led to unprecedented numbers of objects
being unearthed on site and near the cave-chapels. Archaeological excavations
above the caves in 2009 and 2010 have shed significant new light on the
architectural configurations of monasteries in the capital and in Yungang
itself, and the functions of different sections of the cave complex, as well as
monastic life within it. For the first time, it is possible to reconstruct where
the monks lived and translated sacred literary texts, and to fully understand
that freestanding monasteries are an important component of the rock-cut
cave complex.
Illustrated with remarkable, rare full-colour photographs, this
re-examination of the cave-chapels, which brings together previous
scholarship, primary documentation, and more than a decade of first-hand
field research, will not only fill in the gaps in our knowledge about Yungang,
but also raise, and answer, new questions in art history and archaeology.

Professor Joy Lidu Yi is a historian of Chinese art and archaeologist. Her

current research is focused on Buddhist rock-cut cave art and archaeology
in medieval China. She is particularly interested in the issues of art, power,
politics, patronage, and liturgy of Buddhist sanctuaries, as well as the rela-
tionship between cave-temples and tombs.
Yungang
Art, History, Archaeology, Liturgy

Joy Lidu Yi
First published 2018
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
and by Routledge
711 Third Avenue, New York, NY 10017
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2018 Joy Lidu Yi
The right of Joy Lidu Yi to be identified as author of this work has been
asserted by her in accordance with sections 77 and 78 of the Copyright,
Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced
or utilised in any form or by any electronic, mechanical, or other
means, now known or hereafter invented, including photocopying and
recording, or in any information storage or retrieval system, without
permission in writing from the publishers.
Trademark notice: Product or corporate names may be trademarks
or registered trademarks, and are used only for identification and
explanation without intent to infringe.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
Names: Yi, Lidu, author.
Title: Yungang : art, history, archaeology, liturgy / Lidu Yi.
Description: New York : Routledge, 2017. | Includes bibliographical
references and index.
Identifiers: LCCN 2017020521 (print) | LCCN 2017028722 (ebook) |
ISBN 9780203732830 (Master) | ISBN 9781351402408 (ePub) |
ISBN 9781351402415 (Web PDF) | ISBN 9781351402392
(Mobi / Kindle) | ISBN 9781138049901 (hardback : alk. paper)
Subjects: LCSH: Yungang Caves (China)—History. | Art and society—
China—Yungang Caves—History. | Buddhism and culture—China—
Yungang Caves. | Buddhism—China—Yungang Caves—Ritual.
Classification: LCC DS797.75.Y865 (ebook) | LCC DS797.75.Y865 Y5
2017 (print) | DDC 294.3/435095117—dc23
LC record available at https://lccn.loc.gov/2017020521
ISBN: 978-1-138-04990-1 (hbk)
ISBN: 978-0-203-73283-0 (ebk)
Typeset in Sabon
by Apex CoVantage, LLC

VN 1 December 2017
To My Parents
Contents

Acknowledgments ix
Chronology of Chinese Dynasties xi

1 Introduction 1
Chinese antiquarian documentation 2
Pioneering Japanese expeditions 4
Chinese scholarship 6
Western research 8
Purpose and organization of the book 10

2 The sacred site of Yungang 20

Making the sacred cave temples of Yungang 20
Auspicious geographical environment of Yungang 25
Archaeological excavations and related issues 27
Art and architecture 38

3 Phase one – emperor as Tathāgata 53

The imperial five Tanyao caves 54
Dates of excavation of the five Tanyao caves 59
Emperor as the living Tathāgata in the cave temple 64
Tanyao and the significance of the translation of sacred
books in Yungang 67
Concluding remarks 72

4 Phase two – political struggles and chronicles reconsidered 78

Dates of caves 11–13 reconsidered 79
Reconstructing a chronological sequence of the
second-phase cave temples 90
viii Contents
Summary of the chronology of the second-phase
cave temples 103

5 Considering karmic narratives and liturgical functions 110

Reading the iconography 111
Modes of narratives and viewers’ response 114
Ritual and function in a rock-cut cave sanctuary 116
Folk faith, karmic practice, and the Tiwei Boli jing 121
Concluding remarks 125

6 Phase three – the remaining splendor 132

Architectural structure 133
Iconographic composition 140
Telling tales: narrative stories and visual representations 143
Iconographic style 154
Classification, dating, and chronological sequence 156
Concluding remarks 170
One final remark: Buddhist sinicization reconsidered 173

7 Postscript 189

Classification of the caves in the third phase 192

Groups and periodization of the third-phase caves 193
Character glossary 195
Selected bibliography 202
Index 223
Acknowledgments

An old Chinese saying goes that it takes ten years to sharpen a sword. My
sword is the current book. In the process of “sharpening the sword” to figure
out Yungang, many people offered generous and kind help. Without them,
this book would not have been possible.
My heartfelt thanks first go to the Yungang Research Academy for their
full support in this endeavor over the last decade. I would particularly like to
thank my friends Liu Jianjun, Wang Yanqing, Zhang Zhuo, and Zhao Kunyu
at the academy for the many conversations we had and kind assistance when-
ever I was in Yungang, whether it was for three days or three weeks. Their
indispensible help and generosity will be forever cherished.
My deepest gratitude goes equally to the late Professor Ma Shichang,
Dr. Li Chongfeng from Beijing University, Dr. Li Yuqun from the Institute
of Archaeology of the China Academy of Social Sciences, and Dr. Wei Wen-
bin from Lanzhou University for their field research with me and inspiring
academic conversations whenever I am in China. I am grateful for Dr. Li
Chongfeng and Dr. Li Yuqun for suggesting that I take on this rather ambi-
tious task to work on Yungang. I must also admit that my conversations with
Professor Su Bai and his challenging questions were very helpful.
I would like to express my sincere gratitude to Dr. Okamura Hidenori
for his kind assistance when I was conducting research at the Institute for
Research in Humanities at the University of Kyoto.
The suggestions from the anonymous readers, Dr. Nancy Steinhardt,
and Dr. Dorothy Wong for the earlier version of the book have been
very useful. I also would like to thank Angela Howard, Zheng Yan, Amy
McNair, James Benn, Katherine Tsiang, Onishi Makiko, Yagi Haruo, Jen-
nifer Purtle, Keith Knapp, Susan Erickson, Martin Powers, Steven Heine,
Rongdao Lai, Lin Fan, Victor Hori, Grace Fong, John Stowe, and Ihor
Pidhainy for their moral support over the years and many wonderful con-
versations, and for coming to my presentations and providing feedback. I
am very thankful to Eric Henry who read and edited the second draft of
the manuscript chapter by chapter.
I want to thank my editors Matthew Gibbons and Molly Marler and
project manager Kerry Boettcher for their professional job.
x Acknowledgments
My time at the University of Toronto and McGill University was a great,
inspiring journey. My colleagues at Florida International University have
been a great support since I came here five years ago.
My thanks also go to many people from various grotto research institutes
for their support whenever I was there for field research over the last decade.
There are too many to name, but you know who you are.
I am deeply grateful for the unswerving support from Richard Lynn, Klass
Ruitenbeek, William Straw, Richard Guisso, and Graham Sanders. Without
their support, what I have achieved today would not have been possible.
Finally, a special thanks to my family, and, in particular, my husband for
being always there for me over the years. My husband literally read each
version of the manuscript patiently with me. Since English is not my first
language, I sometimes wrote and rewrote drafts in both languages to make
sure that I was expressing my ideas and arguments exactly the way I wanted
to in my mother tongue. We read each version aloud. I am deeply grateful
for this shared journey.
Chronology of Chinese Dynasties

Xia Dynasty ca. 2070–1600 BCE

Shang Dynasty ca. 1600 – ca. 1046 BCE
Zhou Dynasty ca. 1046–221 BCE
Qin Dynasty 221–206 BCE
Han Dynasty 206 BCE–220 CE
Three Kingdoms 220–280
Jin Dynasties 265–420
Northern Dynasties 386–581
Northern Wei 386–534
Emperor Daowu (386–409)
Emperor Mingyuan (409–423)
Emperor Taiwu (424–452)
Emperor Wencheng (452–465)
Emperor Xianwen (466–471)
Emperor Xiaowen (471–499)
Emperor Xuanwu (500–515)
Emperor Xiaoming (516–528)
Emperor Xiaozhuang (528–530)
Emperor Jiemin (531–)
Emperor Xiaowu (532–534)
Eastern Wei 534–550
Western Wei 535–557
Northern Qi 550–577
Northern Zhou 557–581
Southern Dynasties 420–589
Liu Song 420–479
Southern Qi 479–502
Liang 502–557
Chen 557–589
Sui Dynasty 581–618
Tang Dynasty 618–907
Five Dynasties 907–979
Ten Kingdoms 902–979
xii Chronology of Chinese Dynasties
Liao Dynasty 907–1125
Song Dynasty 960–1279
Xi Xia Dynasty 1038–1227
Jin Dynasty 1115–1234
Yuan 1271–1368
Ming Dynasty 1368–1644
Qing Dynasty 1644–1911
1 Introduction

Yungang, a fifth-century rock-cut cave-temple complex, a UNESCO World

Heritage site, and one of the greatest Buddhist monuments of all time, is situ-
ated south of Mount Wuzhou 武周山 and north of the Wuzhou River (current
Shili River 十里河), and is 10 miles west of Datong city in Shanxi Province in
north China. The complex consists of 45 major caves that contain more than
1,000 niches and 51,000 images.1 Traditionally, the caves are divided into three
sections – east (caves 1 to 4), middle (caves 5 to 20), and west (caves 21 to 45)
(fig. 1.1). They were hewn from the mountain cliff surface and stretch out from
east to west for more than half a mile. The caves arose from the imperative to
carry on the Buddha Dharma infinitely and the wish to pray for blessings for the
Northern Wei (386–534) imperial family under whose patronage the excava-
tions were initiated.
The Northern Wei geographer and essayist Li Daoyuan 酈道元 (d. 527)
once described the spectacle of Yungang as such:

Stones were chiseled and the mountain was hewn according to the struc-
ture of the cliff surface. The images are realistic and grandiose. They are
rare by the standards of this time. The [Buddha] Halls on the moun-
tain and over water, and the smoke [of the incense]-filled temples, look
toward each other. The grove and pond are like a bright mirror. Looking
into the distance, a new vista dazzles your eyes.2

This is the earliest extant literary record of Yungang. At that time, Yungang
was called the Lingyan Cave-monastery 靈巖寺.3 Later, in the Weishu (History
of the Wei), Yungang was also called the Mount Wuzhou Buddhist Cave-
monastery.4 By the Tang dynasty (618–907), a large cave in Yungang could
be higher than 200 feet and hold some 3,000 people.5 The eminent monk
Daoxuan (596–667) observed that the carving of the images was fantastic,
the beauty of the ornamentation was unparalleled, and each cave was unique.6
Yungang was referred to as Beitai Cave-monastery 北台石窟 or Heng’an Cave-
monastery 恆安石窟 at that time.7 The name Yungang was not used until the
Ming dynasty (1368–1644), when it occurs for the first time in an inscription
recording the repairs to the Yungang Fortress 雲岡堡, which derived its name
2 Introduction

Fig 1.1 Plan of Yungang caves

from the peak of Mount Wuzhou, as did the Yungang cave complex. The
complex got its name from the peak of the sacred mountain in which all the
caves were excavated. “Yungang” means “cloud ridges.”
In 1902, Japanese architect Itō Chūta (1867–1954) “rediscovered” Yun-
gang accidently and published two articles thereafter, introducing Yungang
to the world.8 In 1907, French sinologist Émmanuel-Édouard Chavannes
(1865–1918) investigated Yungang and other caves, recording them with his
lens. His Mission archéologique dans la Chine septentrionale (Archaeologi-
cal Expeditions in Northern China, 1909–15, 13 vols.) contains 78 valuable
photographs of Yungang. After this, the study of Yungang entered a new era,
that of visual images.9 These early expeditions to China at the beginning of
the 20th century opened a new chapter in the scholarship of Yungang, which
can be classified into four categories: Chinese antiquarian documentation,
pioneering Japanese expeditions, Chinese scholarship, and Western research.

Chinese antiquarian documentation

The written record on Yungang is scanty, and the inscriptions in the caves
have become increasingly illegible over the centuries. For many years this
hindered the study of Yungang, particularly its history after the Northern
Wei capital Pingcheng 平城 (present-day Datong) was relocated southward
to Luoyang in 494. The earliest extant piece of literature that mentions
Yungang is the Shuijing zhu (Annotated Classic of the Waterways). To date,
the Shuijing zhu, Weishu, and Xu gaoseng zhuan (Further Biographies of
Eminent Monks) are the three primary relevant antiquarian documents, of
which the Weishu is of particular importance.10 The fragmentary inscriptions
in the caves are also valuable in connecting the dots for a coherent picture
of Yungang. The inscription on a Jin-dynasty (1114–1234) stele discovered
in 1947 remains, due to lack of more significant evidence to this day, one of
the most important literary records on Yungang.11
The Weishu, compiled between 551 and 554, is an important source,
especially the chapter “A Treatise on Buddhism and Daoism,” providing
important clues for our understanding of the historical and social context
of Yungang.12 It records, for instance, the revival of Buddhism after seven
years of persecution (446–52), the significant meeting of Emperor Wencheng
(r. 452–65) with the eminent monk Tanyao, who pleaded with the emperor
to construct the rock-cut cave temples along the cliff surface of Mount
 Introduction 3
Wuzhou, and the royal visits to Yungang and the sacred mountain dur-
ing the process of the construction in Yungang. The Weishu also mentions
factors that might have, through trade, tribute, war, and foreign relations,
played a significant role in the formation and evolution of the iconography,
style, and architecture of the caves. These literary records (discussed fur-
ther in the relevant chapters), without a doubt, have significant academic
value for contextualizing Yungang in our interpretation and for piecing its
puzzles together.
The Shuijing zhu by Li Daoyuan not only describes the magnificence of
Yungang and the geographical environment surrounding it, the observation
is also sharp and sufficiently accurate, especially in noting that “stones were
chiseled and the mountain was hewn according to the structure of the cliff
surface” (zaoshi kaishan, yinyan jiegou 鑿石開山, 因岩結構).13 The famous
architect Liang Sicheng believes that this applied not only to the caves but
also to the wooden temples connected to their façades.14 Furthermore, Li
Daoyuan also observed that there were, by the river, stone monasteries and
cave chambers where nuns lived.15 Later, Daoxuan, too, mentioned that there
were temples for nuns at the west end of the Yungang complex.16 The investi-
gation of these nearby monasteries and their geographical surroundings may
well shed new light on Yungang and on Buddhist activities in Pingcheng, the
Northern Wei capital (see Chapter Two).
The Xu gaoseng zhuan by Daoxuan is another important primary source
among the scanty records on Yungang. It not only contains additional useful
information on Yungang itself but, more importantly, gives an account of Tan-
yao 曇曜 (d.u.), the founder of Yungang. We know very little about the eminent
monk considering his significant role in bringing Buddhism to a greater height
of development after the catastrophic persecution (452) and his supervision of
the excavation of Yungang, where he was also in charge of the translation of the
sacred texts. The Weishu does not provide us enough information, not even
his date of birth or death. He is only briefly mentioned in association with
the biography of Xuangao 玄高 in the Gaoseng zhuan 高僧傳 (Biographies of
Eminent Monks).17 Moreover, it records and comments on a then-influential
Buddhist text, the Tiwei Boli jing 提謂波利經 (the Sūtra of Trapusa and Bhal-
lika), which was supposed to have been fabricated in Yungang by a certain Tan-
jing for the mass laity.18 The scholar Chen Yuan has suggested that Tanjing 曇靖
translated sūtras with Tanyao at the same time and same place.19 If this is true,
the reason why a fabricated sutra was produced in a court cave-temple complex
where sacred texts were translated deserves a deeper inquiry. This might shed
some new light on the role of Yungang in instructing the mass laity as a court
religious sanctuary (see Chapter Five).
In addition to the aforementioned important written records, the fragmen-
tary inscriptions in the caves are valuable sources helping to join the scat-
tered pieces of information together. Some, especially those with an explicit
date, can bring significant light to the development of iconographic styles
and are indispensible material. They provide an important basis for dating
4 Introduction
and periodization, employing the methodology of archaeological typology,
of the cave temples.20 Some inscriptions clearly indicate the purposes of the
patrons in commissioning caves, niches, or images, and even suggest their
religious beliefs. The scholar Hou Xudong notes that the inscriptions can
sometimes even reflect the family life of the patrons.21 During the Northern
Dynasties (386–581) when Buddhism was very popular, Buddhist devo-
tional groups called yiyi often commissioned Buddhist images, either as
individuals, families, or a Buddhist community.22 These can be found in
many of the caves in Yungang. Thus far, some 60 inscriptions from the
Northern Wei to the Qing dynasty (1644–1911) have been identified and
recorded.23 These inscriptions are scattered in 15 caves: cave 11 has the
most; caves 17–20 each have one; and the rest are spread among the
third-phase caves in the western section. Two points need to be made
here. First cave 11 has the most inscriptions, and second, there are no
inscriptions in any of the so-called paired caves (caves 1 and 2; 5 and 6; 7
and 8; and 9 and 10). These provoke further inquiry and will be treated
in Chapter Four.24

Pioneering Japanese expeditions

The study of Yungang in the 20th century can be divided into two stages:
before 1950, and after. The essential difference lies in whether archaeo-
logical methodology was employed to investigate and record Buddhist cave
remains. In the first half of the 20th century, Chinese caves were mainly
studied, both in the West and the East, from the perspective of art history.
The introduction of archaeological method by Su Bai in the 1950s was
a turning point in the study of Buddhist cave temples.25 The archaeological
typology methodology provides scientific basis in dating and periodization
of the caves. Chinese scholars, however, did not contribute much to the
study of Yungang, or Buddhist art for that matter, in the early 20th century.
Buddhist art and architecture were not yet appreciated as an object of study,
and poverty and political turmoil impeded investigation. In contrast, Japa-
nese scholars, architects, photographers, and artists conducted pioneering
explorations on Yungang and other Buddhist caves.
The earliest expedition to Yungang was made in 1902 by Itō Chūta,
as noted earlier. He pointed out that the Yungang style had been directly
influenced by the Gandhara style and that it had no direct connection
with the Western Regions.26 These observations, however, as Itō himself
acknowledged, are only from the perspective of architecture because he had
knowledge neither of Buddhism nor Buddhist art.
Of the early Japanese scholarship on Chinese Buddhist art, Shina bijut-
sushi chōsō hen (A History of Chinese Art: Sculpture Volume, 1915) by
Ōmura Seigai (1868–1927) is among the most significant. Using largely pri-
mary sources, including some 2,600 inscriptions, Ōmura examined the his-
tory of Chinese sculpture, both religious and secular, from antiquity to the
 Introduction 5
Five Dynasties period (907–60), arranging them chronologically. Some of
Ōmura’s observations are insightful, including his discussions of the origins
of the yiyi Buddhist associations frequently recorded on Buddhist steles.27
Alexander C. Soper, calling the book “an anthology of source material on
Chinese sculpture, and an inexhaustible mine of quotations from texts or
inscriptions,” translated, with annotation, a large portion of the records
into English.28 The translation, focusing mainly on the Buddhist portion, is
in turn a great contribution to Chinese studies, especially to Buddhist art.
Sekinō Tadashi (1868–1935) and Tokiwa Daijō (1870–1945) made seven
archaeological expeditions to China (1906 to 1925) and complied a six-
volume survey of Buddhist monuments in China (1925 to 1931).29 In this
survey, Buddhist monuments (caves, temples, and steles), with an emphasis
on Buddhist cave temples, are carefully examined and arranged geographi-
cally. The section on Shanxi Province is dedicated mainly to the Yungang
caves (1–20). It gives a general description of their iconographic styles and
argues that, stylistically, Yungang was inspired indirectly by Central India
and Dunhuang.30 Tokiwa also pointed out that the iconography of Yungang
is primarily based on the translations of Kumārajīva (d. 413), which proved
to be oversimplified.31 Apparently, early study of Yungang was very limited.
Early studies of Chinese Buddhist art are generally preliminary surveys.
At the time, there were not as many archaeological materials available for
reference as there are today. Many studies, lacking in-depth contextual and
comprehensive analysis, were limited to the framework of a visual under-
standing of the imagery and focused only on part of the Yungang caves for
their investigation.
Two scholars, Mizuno Seiichi (1905–71) and Nagahiro Toshio (1905–90),
however, deserve special recognition. They undertook painstaking inves-
tigations of Buddhist cave temples at Xiangtangshan (1937), Yungang
(1938–45), and Longmen (1941). During seven years of investigation,
they examined and documented caves 1–20 of Yungang thoroughly,
taking measurements, photographs, drawings, and rubbings. They also
conducted several archaeological excavations (1938–40) and published the
monumental Yungang: The Buddhist Cave-Temples of the Fifth Century
A.D. in North China (16 volumes, 32 books), which has been a valuable
source for the study of Yungang, even though it contains misinterpretations
and has its limitations.32 As Harrie Vanderstappen aptly noted, “any study
on Yungang is forever indebted to the monumental work.”33 Mizuno and
Nagahiro’s systematic analysis and close reading of the imagery of each cave
make the work irreplaceable, and the exhaustive description of each plate
is admirable. The photographs and records of the inscriptions have become
even more valuable due to the deterioration of the images and the erosion
of the inscriptions over time.
Each of the 16 volumes contains four parts: texts (introduction, description,
and conclusion), plates, plans, and rubbings. The introduction to each vol-
ume discusses a particular topic, which is followed by a detailed description
6 Introduction
of the cave, and a conclusion that deals only with the cave described.34 The
themed introductory essays focus mainly on two aspects: the historical and
social context, and iconographical styles and visual analysis. The former is
the necessary basis for Buddhist art scholarship, whereas the latter reflect
Mizuno and Nagahiro’s methodology and research focus, which emphasize
the significance of iconographic styles.
Further, Mizuno and Nagahiro divided the caves into three phases (15
years for each phase, from 460 to 505) and put the caves in sequence of
construction.35 The first phase (460–75) includes caves 16 to 20 (the five
Tanyao caves), and caves 8 and 7; the second phase (475–90) starts with
caves 9 and 10, followed by caves 1 and 2, caves 11–13, then caves 5 and 6
(cave 3 was unfinished); and the third phase (490–505) includes caves 21–45
(the western-end caves) and the small caves scattered all over.
Dating and periodization are central to the study of Buddhist art and
archaeology. They have also been an issue of debate among scholars in the
study of Yungang. Su Bai, disagreeing with Mizuno and Nagahiro, suggested
a considerably different scheme of dating. This, along with the heated debate
it aroused, pushed the study of Yungang forward remarkably. These are the
main topics of the Chinese Scholarship section below.
Before we dive into the discussion, however, it should be noted that addi-
tional Japanese scholars continued to work on the Yungang caves, focusing
on certain themes. Yoshimura Rei, Yagi Haruo, Hinako Ishimatsu, and oth-
ers have written about stylistic changes, sinicization, ornamentation, and
dating issues. These will be discussed in relevant chapters.36

Chinese scholarship
Chen Yuan (1880–1971) started Chinese scholarship on Yungang in 1919.
He sorted through the antiquarian literature and pointed out that both the
Weishu and the Shuijing zhu, which, Chen believed was written during the
Taihe era (477–99), are reliable sources. He argued that Tanyao was not only
the founder of Yungang, but also the first to translate sutras there.37
Liang Sicheng’s (1901–72) investigation of Yungang mainly focused on
architecture and provided a survey discussing a range of topics, from col-
umns and rafters to pagodas.38 He noted that the architecture in Yungang
is primarily Chinese indigenous and not much affected by foreign influence.
Not only that, but foreign architecture was often assimilated by Chinese
culture. The pagoda is such an example.39 Liang also observed the ten
wooden temples built at various times before the Qing and referenced them
to those recorded in the Shuopingfu zhi (Gazetteer of Shuoping Prefecture,
juan 3, 1733).40
Yan Wenru (1912–94) investigated Yungang early in the 1940s and was
the first scholar in the 20th century to suggest that Yungang was actually
started earlier than the first year of Heping (460), as is traditionally accepted.
He suggested that the five Tanyao caves were started not in 460, but in the
 Introduction 7
second year of Xing’an 興安 (453), right after the revival of Buddhism.41 In
the light of literary record, which indicates that Tanyao had already begun
to translate sutras in Yungang in 462, it is not impossible that the Yungang
complex might have started earlier than 460.42 This is something worth con-
sideration. With the new archaeological discoveries above the rock-cut caves,
it is now clear that there were freestanding Buddhist temples in Yungang.
This will be further discussed in Chapter Two.
Dating and periodization of the caves, due to lack of clear evidence,
had long puzzled scholars. In 1947, exactly 800 years after Cao Yan 曹衍
initially composed the Jin stele inscription in 1147, Su Bai came across it
in the Beijing Library.43 Some 2,100 words long, the inscription records
the repairs and reconstructions of Yungang through the Tang, Liao, and
Jin dynasties within the context of the initial excavation, providing fresh
evidence and a new opportunity to reevaluate the caves. The inscription
has two features of academic value. First, it adds substantial information
about Yungang after the Northern Wei, thus filling in the blank space of our
knowledge. Second, and more importantly, it quotes now-lost inscriptions
and records from the Northern Wei concerning the dates and patronage
of the cave temples. The inscription is of particular importance because
it touches upon the ten wooden temples attached to the rock-cut cave
façades.44 They are the key to solving the knotty problems of dating, since
one of them (the Chongjiao temple 崇教寺) bears an explicit date. Accord-
ing to the Jin stele inscription (quoting the now-lost Northern Wei records),
the Chongjiao temple was constructed in the eighth year of Taihe (484)
and completed in 489 under the commission of the eunuch-official Qian’er
Qingshi 鉗耳慶時 (Wang Yu 王遇).45 The linchpin, then, is to associate
this temple with a certain rock-cut cave to which it is connected. Once the
Chongjiao temple can be identified and associated with a rock-cut cave, the
date of the relevant cave is therefore clear.
Based on the Jin stele inscription and his observation of the cliff surface
with square rafter holes and gable traces outside the caves, Su Bai inferred
that the ten temples are essentially the same as the one attached to caves 5
and 6 today. He further speculated that the ten temples were initially associ-
ated with caves 1 and 2, cave 3, caves 5 and 6, caves 7 and 8, caves 9 and
10, caves 11–13, caves 15 and 16, caves 17 and 18, cave 19, and cave 20.46
Most importantly, Su associated the Chongjiao temple (484–89) with caves
9 and 10, and the Huguo temple 護國 with caves 7 and 8.47 This indicates
that caves 9 and 10 were constructed between 484 and 489. For the first
time, a Yungang cave can be dated with explicit dates of construction regard-
less of the precise identity of the cave or date (this will be treated fully in
Chapter Four). This has advanced the study of Yungang. On the basis of this
new evidence and employing archaeological typology method, Su further
reexamined the caves and shed significant fresh light on periodization and
chronological sequence of the caves.48 He provided a new scheme of dating
while dividing the caves into three periods. The first phase (460–65) includes
8 Introduction
caves 16–20.49 The second phase (465–94) contains 12 caves (caves 7 and 8
were the earliest, followed by caves 9 and 10; caves 11–13;50 caves 1 and 2;
and, finally, caves 5 and 6; cave 3 is unfinished). If this is true, it only left
five years to finish seven caves (11–13, 1 and 2, and 5 and 6; the issue will
be addressed in Chapter Four). The third-phase caves (494–524) refer mainly
to the western-end caves (caves 21–45).
Su’s scheme of dating was, in turn, challenged by Mizuno and Naga-
hiro who divided the caves into three 15-year phases spanning from 460 to
505, putting the last year of Yungang almost 20 years earlier. Refuting Su,
Nagahiro argued that the authentication of the Jin inscription was highly
questionable and that the periodization of Yungang should not be based
on untested written material.51 Su rebutted with a detailed history of the
inscription and the ten temples, arguing that Nagahiro’s dating of the caves
in line with the dates of the imperial visits to the caves is not reliable.52
Ding Mingyi, supporting Su’s theory, argued that, without even reading the
inscription, Nagahiro’s argument challenging the dating and authenticity
cannot be convincing.53 Nagahiro later changed his position and accepted
Su Bai’s theories of dating and periodization.54 The scheme of periodization
of the first- and third-phase caves proposed by Su Bai is commonly accepted
in large part, but the second-phase caves remain problematic.55 I attempt to
treat these issues fully in Chapter Four.
The dispute between Japanese and Chinese scholars essentially reflects the
fact that the methodology they employ differs considerably. Chinese schol-
ars prefer to use literary sources and archaeological evidence to interpret
and analyze visual images and architectural structure, whereas the Japanese
scholars emphasize visual impression and interpretation, focusing more on
stylistic evolution and using it as an anchor for reading the imagery. In a
broader sense, the kernel of the dispute touches upon two basics in the
archaeological method: (a) how to use antiquarian primary sources, and
(b) how to use the typology approach to analyze iconographic styles.56

Western research
The earliest leading Western scholars of Chinese Buddhist art were Édouard
Chavannes, Osvald Sirén (1879–1966), and Alexander Coburn Soper (1904–93).
Chavannes started his expeditions to Buddhist cave temples (Yungang, Longmen,
and the caves in Shandong) in 1907. His survey of Buddhist caves and temples,
with invaluable photographic records, introduced Yungang and other caves visu-
ally.57 Some of the Yungang images documented by him were the first visual
sources scholars depended on during the early study of Yungang.
Sirén conducted important surveys of Chinese Buddhist art in the early
1920s. His Chinese Sculptures from the Fifth to the Fourteenth Centuries
(1925, 4 vols.) analyzed the Chinese Buddhist monuments, arranging them
according to geographical location and chronological order.58 He briefly dis-
cussed Yungang in the Shanxi section and argued, as did Chavannes, that
 Introduction 9
Yungang was started in the first year of Shenrui (414) and ended in Zheng-
guang (524).59
Soper’s great contribution to the field was his translation of Ōmura’s book,
as mentioned earlier. He also wrote several important articles concerning
patronage, dates, and iconographic style changes.60 Soper suggested that
caves 7 and 8 in Yungang were dedicated to Wenchengdi and his consort
Lady Li, caves 9 and 10 to Emperor Xianwen (Xianzu, r. 465–70) and Lady
Li, and cave 12 to Emperor Xiaowen (Gaozu, r. 471–99). He further pointed
out that caves 5 and 6 represent the intense ambition and aggressiveness,
and probably also the resentful malice, of the Empress Dowager Wenming,
Lady Feng. He believed cave 6 was dedicated to Lady Feng to “demonstrate
her unique position by the same sort of overwhelming display of power
and wealth that began to take form.”61 Cave 5 was dedicated in memory of
her father, Lang.62 Soper argued that the sinicizing style had nothing to do
with the reform carried out by Emperor Xiaowen: “The new fashion used
at Yungang was borrowed ready-made from southern practice, and in the
south, prior to the transfer, had been familiar for a century.”63 He attributed
the stylistic change in Yungang to the Buddhist imagery reform promoted
by Dai Kui (?–396).64
The Written and Unwritten (1988) by James Caswell revolves primarily
around two aspects: dating and patronage. Based on the Jin inscription as
an important source, Caswell provides a “new history of Yungang” with a
new scheme of dating, periodization, and patronage.65 The book has been
extensively reviewed.66 These are delicate issues for any scholar to take on
because the caves, especially the five Tanyao caves, do not bear any inscrip-
tions referring to their own excavation, nor do we have any cave-making
records concerning them.
Nancy Steinhardt’s several books are useful for the study of Chinese archi-
tecture. Her Chinese Architecture in an Age of Turmoil, 200–600 (2014)
examines the architectural history of China in the age of disunity.67 The sec-
ond part of the book focuses on “distinctive monuments or moments” from
stūpas to monasteries, ritual architecture, and Buddhist cave-chapels. It is of
particular importance for the study of Buddhist art and architecture.68 Chinese
Imperial City Planning (1990) provides an explanation of the multifaceted
evolutionary process of Chinese imperial city planning. It explores the major
imperial capitals of China including Pingcheng, outside of which Yungang is
situated. Steinhardt argues that an imperial city is more than the ruler’s capital;
it is an institution and an articulated concept for which a design is drawn.69
Archaeological and Visual Sources of Meditation in the Ancient Monaster-
ies of Kuca (2015) by Angela F. Howard and Giuseppe Vignato investigates
the cave temples in the Kucha region focusing on the monastic practices.
Vignato concentrates on “meditation cells” of the cave temples, whereas
Howard examines the paintings, especially the “visual language of medita-
tion.” The practice of meditation is recognized as “a source, if not the main
source of the decoration of central pillar caves, monumental Buddha caves
10 Introduction
and square caves.” The book on the art, architecture, and meditation prac-
tice is extremely invaluable for the study of Buddhist caves.
Dorothy Wong’s informative Chinese Steles (2004) is a significant work
in the study of Buddhist art. Although the focus of the book is primarily
on Buddhist steles of different regions, Wong’s meticulous examination of
the objects within their social, historical, and religious contexts gives this
study a greater scope. It provides useful information on regional beliefs,
practices, and even Buddhist activities. Furthermore, the diversified motifs
and inscriptions on the steles shed light on other Buddhist objects of similar
dates and styles.70
The Donors of Longmen by Amy McNair investigates the Longmen cave
temples, with an emphasis on donors in social, political, and religious con-
text. This is an important book for the study of Yungang because the author
frequently compares the architecture, iconographic style, and original pro-
gram of the Longmen caves to those of Yungang. The investigation of the
cost of a grotto is useful since we have little to no record on the subject.
The Weishu only indicates the expenditure for the three Bingyang caves as
802,366, without giving any monetary unit.71 Based on the cost of the Bing-
yang caves and that of the other caves indicated in the inscriptions and other
sources, we might arrive at an estimate of the cost of a cave in Yungang.72
Due to limited space, many other scholars’ important work on Buddhist
cave art, architecture, and archaeology cannot be mentioned here, but will
be discussed in related chapters.

Purpose and organization of the book

My purpose in this study is to revisit Yungang in the light of new archaeological
findings unearthed on site and in Pingcheng in recent years, and to offer a
social art history of Yungang based on previous scholarship, primary docu-
mentation, and more than a decade of firsthand field research. Many of these
findings were not available to earlier scholars. With the recent earth-shaking
economic changes in China, unprecedented numbers of objects have been
unearthed and archaeological remains discovered in the course of roadways
being expanded and cultural sites being rebuilt. The archaeological discoveries
in 2010 and 2011 above the caves have shed significant new light on monastic
life and liturgical functions in each different area in Yungang, of which we had
no clear picture before. For the first time we began to have an understanding
of where Buddhist monks lived and translated sacred texts in Yungang, and
why there were no vihara (dwelling) caves in Yungang. In addition, discov-
eries of the architectural ruins and tombs in and around Pingcheng have
provided important clues about the reciprocal impact between the religious
and secular material culture of the time, and indicate that the local influ-
ence on the construction of Yungang was greater than we had previously
believed. Moreover, a new approach of looking into the caves from their
outside wall surface enabled us to reconsider the relationships between caves,
 Introduction 11
and between caves and niches in time and space. Therefore, an updated inves-
tigation of the Yungang complex is much needed.
The essential questions of this study are why, when, and under what cir-
cumstances this impressive Buddhist rock-cut cave monument was made,
and who played a significant role at various stages. How did the caves and
temples function in each different area? Yungang, a court Buddhist cave
complex, was constructed soon after the revival of Buddhism in order to
protect the religion, pray for blessings, and avert disaster for the royalty.
For the Northern Wei rulers, Buddhism was a vehicle to proclaim power and
beneficence. Yungang was therefore a political as well as a religious sanctu-
ary. Meditation practice was not the primary consideration for function of
the caves, as it was the case for many of Buddhist cave temples. Yungang
cave temples, especially the five colossal-image caves (16–20), were essen-
tially memorials for the royal family.
It is believed, based on the literary sources and new archaeological ruins
discovered recently, that the whole Yungang complex started with Buddhist
temples above where the rock-cut caves are now, but the caves themselves
only began to be hewn at around 460 CE (discussed in Chapter Three).
Buddhist culture had penetrated to every corner of the Northern Wei capital
before the first-court cave temple was excavated.
It is also suggested that cave 13 is the earliest of the second-phase caves,
and that both cave 13 and cave 11 were excavated under Emperor Xianwen
(r. 465–71 CE). Cave 12 was added later, spatially between cave 13 and cave
11. Thereby a new chronological sequence of the excavation of the second-
phase caves and a dating scheme will be constructed. In addition, the third-
phase caves will be examined in detail. No previous attempt has been made
to explore these relatively small caves situated to the west in the Yungang
complex (caves 21–45).
It is argued that the subject matter of Yungang is associated with literary
texts – specifically, with those translated by the eminent monk Tanyao. It is
also my belief that the subject of the Three Buddhas (of the Past, Present,
and Future), which is ubiquitous throughout Yungang, was primarily based
on the Fufazang yinyuan zhuan (A History of the Indian Patriarchs) and the
Dajiyi shenzhou jing (Sūtra of the Great Auspicious Significance of Magic
Spell), which were translated by Tanyao and others in Yungang.
Yungang: Art, History, Archaeology and Liturgy is the first comprehen-
sive western-language work on Yungang, a long overdue book. It is hoped
that this re-examination of the cave temples in Yungang will not only fill in
the gaps in our knowledge about Yungang or reconstruct the site, but also
raise and perhaps answer new questions in art history and archaeology.
Throughout my research, I often pondered the question of the liturgical
functions of the cave temples, which cannot be emphasized enough but
which have, to date, been largely neglected. In this contextual study, I con-
stantly ask how the rock-cut Buddhist cave sanctuary of Yungang was used
as a religious space.
12 Introduction
Chapter One introduces the state of the field, and the purpose and arrange-
ment of this volume.
Chapter Two provides an overview of Yungang in historical, political,
social, religious, and artistic context. It discusses the art, architecture,
archaeology, cross-cultural, and local influence and factors exerted in the
excavation and formation of the caves. It is a panoramic perspective of
Yungang. An important part of the focus of the chapter is on the discoveries
of archaeological ruins in Yungang, which have not been discussed despite
the fact that Japanese scholars made several attempts to do so in the begin-
ning of the 20th century. New discoveries in recent years have shed great
new light on the functions of Yungang in each area, of which we had no
knowledge before.
Chapter Three focuses on the five Tanyao caves (caves 16–20), exploring
the architecture, subject matter, and function. Important questions to be
asked are how these five imperial caves with colossal images were used, and
which one of the five memorials was constructed first and why.
Chapters Four and Five examine the second-phase caves (11–13 and the
four paired caves) from different perspectives. They not only reconstruct a
chronological order of the excavation of the caves, but also investigate their
liturgical functions. Caves 13 and 11 best exemplify the close associations
between art and politics, art and power, and also between art and patrons,
not just in Yungang but in Buddhist cave art generally. The dating, chrono-
logical sequence, and patronage of the four paired caves have been much
debated. These will be the treated fully in Chapter Four. Chapter Five is
devoted to the liturgical functions of the second-phase caves using cave 12
as a case study.
Chapter Six is devoted entirely to the third-phase caves, which have
skipped scholars’ attention. Even basic issues such as what the visual rep-
resentations tell us, who they represent, what the chronological order of
the caves is, and how much the “Yungang style” influenced Buddhist art
in central plain China have gone unexplored. It is these questions that this
chapter attempts to answer. I will attempt to date the third-phase caves
using the extant images bearing explicit dates as an anchor, thereby put-
ting the caves in chronological sequence as well. This part of the study
is crucial to our understanding of the continuity of the “Yungang style”
and to tracing the iconographic origin of the Shanxi minor caves (plate 1)
and their interrelations with Yungang. It will provide a clear picture of the
development of Buddhist cave temples in the central plain region of China.
This is a major contribution to the field since these caves have been largely
neglected to date.
Chapter Seven concludes the study. It is impossible to give, either in words
or photographs, any adequate idea of the enormous impressiveness of Yun-
gang. It is my hope that this new investigation will advance the study of
Yungang and inspire further research on this, one of the most spectacular
Buddhist cave complexes.
 Introduction 13
Notes
1 There are more than 200 caves but only 45 major caves (21 large and 24 medium)
are numbered (from east to west). From the beginning of the 20th century to date,
the caves have been numbered five times. The numbers used here are from the last
time (1987) by the Yungang Research Academy itself. These are current conven-
tion. For detailed information about the numbers of the caves given at different
times, see Li Xueqin, “Yungang shiku xinbian kuhao shuoming,” in Zhongguo
shiku Yungang shiku. Ed. Yungang shiku wenwu baoguansuo (Beijing: Wenwu
chubanshe, 1991), vol. 1, 209–11. Also see Li Xueqin, “Guanyu Yungang shiku
xinbian kuhao de buchong shuoming,” Wenwu, no. 5 (2001), 87–9.
2 Li Daoyuan, Shuijing zhu jiaozheng, annotated by Chen Qiaoyi (Beijing: Zhon-
ghua shuju, 2007), 316.
3 Ibid.
4 Wei Shou (506–72), comp., Weishu (Beijing: Zhonghua shuju, 1974), 6:130,
7:151.
5 Daoxuan, Xu gaoseng zhuan, T50:2060:425c26 and T50:2060:427c23 (T refers
to Taishō Shinshū Daizōkyō 大正新修大藏經. Eds. Takakusu Junjirō and Watanabe
Kaigyoku. 85 vols. Tokyo: Taishō issaikyō kankōkai, 1924–32. Hereafter T for all
Chinese Buddhist canons).
6 Ibid., T50:2060:427c27.
7 Ibid., T50:2060:425c26 and T50:2060:427c23.
8 Chūta Itō, “Shina sansei Unkō no sekkutsu-ji,” Kokka, 17, nos. 197/198 (1906).
Also see Dongfang zazhi, 16, nos. 2/3. This article was translated into Chinese and
collected together with Chen Yuan’s article “Ji Datong Wuzhoushan shikusi,” in
Chen Yuan. Chen Yuan xueshu lunwenji , vol. 1 (Beijing: Zhonghua shuju, 1980),
398–409. Chūta Itō was the first to give numbers to the caves (current caves 5 to 13).
9 Édouard Chavannes, Mission archéologique dans la Chine septentrionale (Paris:
Ernest Leroux, 1909).
10 Other sources such as Guang hongming ji (Expanded Collection of the Propaga-
tion and Clarification [of Buddhism]) and his Da Tang neidian lu (Inner Record of
the Great Tang) Daoxuan also mentioned Yungang; see Daoxuan, Guang hong-
ming ji, T52:2103:103c26, and Daoxuan, Da Tang neidian lu, T55:2149:267b28.
They will be discussed later in the relevant chapters.
11 The stele, the Dajin xijing Wuzhoushan chongxiu dashikusi bei (A Record of the
Repairs to the Grand Cave Temple of Mount Wuzhou in the Western Capital
of the Great Jin; hereafter “the Jin stele inscription”), is no longer extant. The
inscription survives in the Yongle dadian. In the 23rd year of Zhizheng (1363),
during the Yuan dynasty, Xiong Mengxiang 熊夢祥 (Xiong Zide), a Confucian
scholar and poet, visited Yungang personally and transcribed the entire inscrip-
tion. He later included it in his Xijing zhi 析津志 (Xijing Gazetteer), the earliest
Beijing gazetteer, which, in turn, was included in the Shuntianfu zhi 順天府志
(Shuntianfu Gazetteer), a Beijing gazetteer compiled in 1885 by Miu Quansun
繆荃蓀 (1844–1919), a Qing bibliophile, in the fragmentary Yongle dadian. This
is the only source that we have today about the inscription.
12 Weishu, “Shilaozhi,” 114:3025.
13 Li Daoyuan, Shuijing zhu jiaozheng, 316.
14 Liang Sicheng, “Yungangshiku zhong suo biaoxian de Beiwei jianzhu,” in Yun-
gang bainian lunwen xuanji. Ed. Yungang shiku wenwu yanjiusuo (Beijing:
Wenwu chubanshe, 2005), 11–40.
15 Li Daoyuan, Shuijing zhu jiaozheng, 316.
16 Daoxuan, Guang hongming ji, T52:2103:0103c27(00).
17 Hui Jiao, Gaoseng zhuan, T50:2059:398b10.
18 Daoxuan, Xu gaoseng zhuan, T50:2060:428a.
14 Introduction
19 Chen Yuan, “Ji Datong Wuzhoushan shikusi,” in Chen Yuan xueshu lunwenji,
vol. 1, 398–409. Also see Chen Yuan, “Yungang shikusi de yijing yu Liu Xiao-
biao,” in Chen Yuan xueshu lunwenji, vol. 1, 443–48.
20 For instance, the inscriptions in cave 11–14 and on the east side of the window in
cave 17 are both dated to the thirteenth year of Taihe (489). The images in the
two caves, however, demonstrate completely different iconographic styles. The
images in cave 17 exhibit earlier characteristics, whereas those in cave 11–14 are in
a more sinicized style. This shows that there had been a stylistic change by 489,
but that both styles existed simultaneously before the older one disappeared.
21 Hou Xudong, Beichao cunmin de shenghuo shijie: chaoting, zhouxian yu cunli
(Beijing: Shang wu yin shu guan, 2005), 267.
22 “Yishe 邑社” and “lishe 裏社” were the local religious and social organizations
associated with the worship of She (the earth god). These yish and yili became
models for the later Buddhist devotional societies, called yiyi, yi, or she. See
Dorothy Wong, Chinese Steles: Pre-Buddhist and Buddhist Use of a Symbolic
Form (Honolulu: University of Hawai‘i Press, 2004), 9 and 15.
23 Mizuno Seiichi and Nagahiro Toshio recorded about 66 inscriptions and
arranged them in chronological order. See Mizuno Seiichi and Nagahiro Toshio,
Unkō sekkutsu: seireki goseiki ni okeru Chūgoku hokubu Bukkyō kutsuin no
kōkogakuteki chōsa hōkoku: Tōhō Bunka Kenkyūjo chōsa (Yungang: The Bud-
dhist Cave-Temples of the Fifth Century A.D. in North China: detailed report
of the archaeological survey carried out by the Mission of the Tōhōbunka
kenkyūsho 1938–45, hereafter Unkō sekkutsu), vol. 2 (Kyoto: Kyōto Daigaku
Jinbun Kagaku Kenkyūjo, 1951). The 27 inscriptions of the Northern Wei are pri-
marily dedicative niche inscriptions; the single Liao (907–1125) commemorative
inscription is about the repairs of the images; the 11 Yuan-dynasty (1271–1368)
inscriptions are travel accounts; and the Ming and Qing inscriptions are mostly
commemorative inscriptions on steles recording repairs of roads or cave temples.
Also, the inscription dedicated by a certain Tanmei 曇媚 in the fourth year of
Jingming 景明 (503) was not included in the record since it was only discovered
in 1956 in front of cave 20, in a pile of earth.
24 The aforementioned Jin stele inscription is discussed in detail below, in the Chi-
nese scholarship section.
25 Xu Pingfang, “Zhongguo shikusi kaoguxue de chuangjian licheng – Reading Su
Bai’s Zhongguo shikusi yanjiu,” Wenwu, no. 2 (1998): 54–63.
26 See Itō Chūta, “Shina sansei Unkō no sekkutsu-ji (The Cave-Temples of Yungang
in Shanxi, China),” Kokka, nos. 197/198 (1906). Also see Tōyō kenchiku no
kenkyū (Tōkyō: Ryūginsha, 1943), and Kengaku kikō (Tōkyō: Ryūginsha, 1936).
27 Wong, Chinese Steles, 6.
28 Ōmura Seigai, Ed. Literary Evidence for Early Buddhist Art in China. Alexander
C. Soper, trans. (Ascona, Switzerland: Artibus Asiae, 1959), 7.
29 Sekino Tadashi and Daijō Tokiwa, Shina Bukkyō shiseki (Tokyo: Bukkyō Shiseki
Kenkyūkai, 1925). The English version Buddhist Monuments in China came out
in 1930.
30 Tokiwa Daijō and Sekinō Tadashi, Buddhist Monuments in China (Tokyo: Bukkyō
Shiseki Kenkyūkai, 1930), text volume, 14–21. In 1918, Sekinō paid a visit to
Yungang and later wrote two articles, “Saiyū zasshin” and “Unkō sekkutsu no
nendai to sono yōshiki no iigen nitsuite,” which were collected in his book Shina
no kenchiku to geijutsu (Tokyo: Iwanami Shoten, 1938). Tokiwa also made a
number of archaeological expeditions to China on his own, and from 1920 to
1928, he made five visits to Buddhist sites, Yungang included. These exploration
accounts appeared later, in his book Shina bukkyō shiseki tōsaki (Expeditions to
Chinese Buddhist Sites) in 1937 (Tokyo Kokusho Kankōkai, 1972). Arranging
these sites geographically, Tokiwa introduced each of them briefly. Many other
 Introduction 15
Japanese scholars and photographers also visited and explored Yungang in the
early 20th century. Matsumoto Bunzaburō (1869–1944) visited Yungang on his
way home from a journey to India in 1917. In his book Shina Bukkyō ibutsu
(Buddhist Remains in China), (Tokyo: Daitōkaku, 1919), Matsumoto argued that
the Yungang caves are more influenced by the Gupta style, not the Gandhara
style. In September 1920, Kinoshita Mokutarō 本下杢太郎 (1885–1945), who
was a medical doctor, went to Yungang with the artist Kimura Sōhachi 木村莊八
(1893–1958). Kinoshita wrote about this journey, and included Kimura’s sketches,
under the titles “Unkō nichiroku 云岡日錄 (Account of Yungang Journey)” and
“Daidō sekibutsu zatsuwa 大同石佛雑話 (Miscellaneous Accounts of the Datong
Stone Buddhas).” Later on, these two short works were collected in his book
Daitō sekibutsuji 大同石佛寺 (The Datong Cave Temples), (Tōkyō : Nihon Bijutsu
Gakuin, 1922). This book, though not as scholarly as that of Ōmura or those by
Sekinō and Tokiwa, was remarkably influential and widely known in Japan.
31 Tokiwa Daijō, Shina Bukkyō no kenkyū (Tokyo: Shunjūsha, 1938), 512–14.
32 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu. The original 16 volumes
were republished and expanded to 20 volumes, with previously unpublished
old photographs and additional scholarship, by the Institute for Research in
Humanities (Kyoto University) and the Institute of Archaeology (China Academy
of Social Sciences), (Beijing: Kexue chubanshe, 2014–). A detailed account of
this indefatigable expedition is given in Mizuno’s article “Unko sekkutsu chōsaki
(Notes on the Investigation of Yungang Caves),” which later was collected in
Mizuno’s Chūgoku no Bukkyō bijutsu (Chinese Buddhist Art), (Tokyo: Heibon-
sha, 1968), which even provides the schedule of the measuring, photographing,
making rubbings, and so on. Nagahiro also published a book on both Yungang
and Longmen caves, Unkō to Ryūmon (Yungang and Longmen), in which he
briefly compares these two sites and traces the origin of the style of Yungang. He
analyzed the styles of Gandhara and Gupta as well as other Central Asian sites
and their relations. Nagahiro notes that cave 5 displays a combination of different
styles from Gandhara, central India, and Central Asia. However, he cautions that
this does not mean that cave 5 was directly influenced by the West; rather, it was
the Liangzhou artists who played a more important stylistic role. See Nagahiro
Toshio, Unkō to Ryūmon (Tokyo: Chūō Kōron Bijutsu Shuppan, 1964).
33 Harrie A. Vanderstappen, “Book Review: Written and Unwritten: A New History
of the Buddhist Caves at Yungang,” Ars Orientalis, 19 (1989): 125–7.
34 The topics of each introduction – for a total of 14 essays in all – are: “general
observations on the Yungang cave-temple,” “Shi-fo-si of Yungang,” “the Yun-
gang caves and their historical background,” “significance of ornamentation in
the cave-temples of Yungang,” “the investigation of the cave-temples of Yungang,”
“genealogy of the Yungang caves,” “representations of scenes from Buddha’s life in
the Yungang caves,” “iconography of the Yungang caves,” “Yungang and Long-
men Styles,” “Buddhist images prior to the Yungang caves,” “Western styles in the
Yungang caves,” “Tanyao and the Yungang caves,” “the cave-temples in China,”
and “chronology of Yungang cave.”
35 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vol. 16, 30.
36 See Yagi Haruo, Chūgoku Bukkyō bijutsu to kan minzokuka: Hokugi jidai kōki
o chūshin to shite (Sinicization of Chinese Buddhist Art), (Kyōto-shi: Hōzōkan,
2004); Yagi Haruo, Unkō sekkutsu mon’yō ron (Study of Yungang Ornamenta-
tion), (Kyoto: Hōzōkan, 2000); Yoshimura Rei, Chūgoku Bukkyō zuzō no kenkyū
(Study of Chinese Buddhist Art), (Tokyo: Tōhō Shoten, 1983); Yoshimura Rei,
Tennin tanjōzu no kenkyū: Higashi Ajia Bukkyō bijutsushi ronshū (Study of the
Celestial-Born Iconography: Selected Essays on East Asian Buddhist Art), (Tokyo:
Tōhōshoten, 1999); Hinako Ishimatsu, Hokugi Bukkyo zozoshi no kenkyu (Bud-
dhist images in the Northern Wei), (Kunitachi-shi: Buryukke, 2005).
16 Introduction
37 Chen Yuan, “Ji Datong Wuzhoushan shikusi,” 398–409.
38 Liang Sicheng, “Yungangshiku zhong suo biaoxian de Beiwei jianzhu,” 11–28.
39 Ibid., 27–8.
40 See Liu Shiming and Yang Wang, Shuopingfu zhi, 12 juan (China: s.n., 1733).
Noteworthy is that the ten temples (Tongsheng, Lingguang, Zhenguo, Huguo,
Chongfu, Tongzi, Nengren, Huayan, Tiangong, and Doushuai) mentioned in
Shuopingfu zhi are different from those recorded in the Jin stele inscription.
41 Yan Wenru, “Yungang shiku zaoxiang zhong de yixie ticai de kaoshi,” Xiandai
foxue, no. 2 (1963), 11–21 and Yan Wenru, “Yungang shiku de kaichuang he ticai
de fenxi,” Shehui kexue jikan, no. 5 (1980), 112–18, and no. 6 (1980), 110–14.
42 Fei Changfang 費長房 (597), comp., Lidai sanbao ji 歷代三寶記 (Records of the
Three Jewels through All Dynasties), T49:2034:85b05.
43 We know nothing about Cao Yan except that he wrote the Jin stele inscription.
44 The names of the temples are Tongle, Lingyan, Jingchong, Zhenguo, Huguo,
Tiangong, Chongjiao, Tongzi, Huayan, and Doushuai. Some of them are men-
tioned in other sources. For instance, the Tongle temple is recorded in Xu gaoseng
zhuan, juan 1; Kaiyuan shijiao lu, juan 6; Gujin yijing tuji, juan 3; Zhenyuan
xinding shijiao mulu, juan 9; the Lingyan temple is mentioned in Weishu, juan
114; Xugaoseng zhuan, juan 1; Kaiyuan shijiao lu, juan 6; Zhenyuan xinding shi-
jiao mulu, juan 9; and the Tiangong temple in Weishu, juan 114. The rest of the
temples are not mentioned in any extant sources before the Song dynasty. Su Bai
thus inferred that the names of all ten temples started in the Liao dynasty and that
they were still in use when Cao Yan inscribed the stele in 1147. The temple names
were changed later. In the Chenghua Shanxi Tongzhi (1474), they are named:
Tongsheng, Lingguang, Zhenguo, Huguo, Chongfu, Tongzi, Nengren, Huayan,
Tiangong, and Doushuai. See Su Bai, “Dajin xijing Wuzhoushan chongxiu dashi-
kusi bei jiaozhu (hereafter “Jinbei Jiaozhu”),” in Su Bai, Zhongguo shikusi yanjiu
(Beijing: Wenwu chubanshe, 1996), 60. The article was first published in Beijing
daxue xuebao, no. 1 (1956). Judging from the square rafter holes on the façade
of the caves and the archaeological findings, the wooden temples are believed to
have been constructed during the Liao dynasty, as evidence proves that there were
large-scale reconstructions and repairs in front of the caves during the Liao. See
Su Bai, “Jinbei Jiaozhu,” 69–70; and Yungang shiku baoguansuo and wenwu
baoguan kexue jishu yanjiusuo, “Yungang shiku jianzhu yiji de xin faxian,”
Wenwu, no. 4 (1976), 89–93.
45 Su Bai, “Jinbei Jiaozhu,” 53.
46 Ibid., 56–7.
47 Ibid., 60–1. Zhang Zhuo argued that the Huguo Temple should be caves 1 and 2,
not caves 7 and 8. See Zhang Zhuo, Yungang shiku biannianshi (The Chronicle
of the Yungang Caves), (Beijing: Wenwu chubanshe, 2006), 437.
48 Su Bai wrote five articles on Yungang after the discovery of the Jin stele inscription
and they are all collected in Su Bai, Zhongguo shikusi yanjiu: (1) “Jinbei Jiaozhu”;
(2) “Dajin xijing wuzhoushan chongxiu dashikusi bei de faxian yu yanjiu” (Discovery
and Study of the Stele on Repairs of the Mount Wuzhou Stone Cave-temples in the
Great Jin Western Capital, hereafter “Jinbei yanjiu”); (3) “Yungang shiku fenqi
shilun” (Preliminary Investigation on the Periodization of the Yungang Caves;
hereafter “fenqi”), (4) “Pingcheng shili de jiju he ‘Yungang moshi’ de xingcheng yu
fazhan” (The Steady Empowerment of Pingcheng and the Formation of “Yungang
Style,” hereafter “Yungang moshi”); and (5) “Heng’anzhen yu Heng’an shiku
(Heng’an Town and Heng’an Caves).”
49 Su Bai’s article “fenqi” appeared first in Kaogu xuebao, no. 1 (1978), and later
was included in his Zhongguo shikusi yanjiu, 80.
50 The earliest (483) and latest (496) inscriptions in cave 11 suggest that cave 11 is close
in time to caves 9 and 10. See Su Bai, “fenqi,” in Zhongguo shikusi yanjiu, 80.
 Introduction 17
51 Nagahiro Toshio, “Shuku Haku Shi no Unkō Sekkutsu bunkirono bakusu”
(Rebuttal of the Periodization of the Yungang Caves by Su Bai), Tōhō gaku,
no. 60 (1980): 30–44.
52 Su Bai, “Jinbei yanjiu,” in Zhongguo shikusi yanjiu, 89. The imperial visits to
Mount Wuzhou started with Emperor Mingyuan (r. 409–23) in 411. Emperor
Xiaowen (Gaozu) visited the “spiritual mountain” twice, in 475 and 477. The
imperial visits to the actual caves only started from Emperor Xianwen (Xianzu) in
467, and then Gaozu visited in 480, 482, 483, and 484. See the Weishu, “Lizhi”
181:2733, and “Diji” 6:125 and 7:135. From the eighth year of Taihe (484) to
the eighteenth year of Taihe (494), Gaozu did not visit either the caves or the
“spiritual mountain.”
53 Ding Mingyi, “Guanyu Yungang shiku fenqi de jige wenti,” in Yungang Bainian
Lunwen xuanji. Ed. Yungang shiku wenwu yanjiusuo (Beijing: Wenwu, 2005),
197–211. Also see Ding Mingyi, “Yungang shiku yanjiu wushinian,” in Zhong-
guo shiku Yungang shiku, vol. 1, 176.
54 Nagahiro Toshio and Seiichi Mizuno, Unkō sekkutsu: Chūgoku bunka shiseki
(The Yungang Caves: Historical Landmarks of Chinese Culture), (Tokyo: Sekai
Bunkasha, 1976), 41–4; also see Nagahiro Toshio, “Yungang shiku de jiushi
shuangku de tezheng” (The Features of the Twin Caves of 9 and 10), in Zhong-
guo shiku Yungang shiku, vol. 2. Ed. Yungang shiku wenwu baoguansuo (Beijing:
Wenwu chubanshe, 1994), vol. 2, 193–207.
55 Some scholars questioned the date of commencement in Yungang as the 60s of
the 5th century. Harrie A. Vanderstappen implied that it was too short of time to
complete so many caves, and caves 7 and 8 are different in style from other caves
and they should be earlier. See Vanderstappen, “Book Review,” 125–7. Follow-
ing the same line, Katherine R. Tsiang argued that caves 7 and 8 originated in the
earliest formative phase of cave-making at Yungang when foreign models were
considered authoritative and that they predate the conventionally accepted peri-
odization of the site. See Tsiang, “Reconsidering Early Buddhist Cave-Making of
the Northern Wei in Terms of Artistic Interactions with Gansu and the Western
Regions,” Journal of Inner Asian Art and Archaeology, 4 (2009): 101–18.
56 Xu Pingfang, “Zhongguo shikusi kaoguxue de chuangjian lichen–Du Su Bai xian-
sheng de ‘Zhongguo shikusi’,” Wenwu, no. 2 (1998): 61. There is much more
Chinese scholarship on Yungang, but due to limited space, it is only cited in the
following chapters in reference to related topics. For example, Zhang Zhuo col-
lected and annotated historical records concerning Yungang and Pingcheng; see
Zhang Zhuo, Yungang shiku biannianshi (The Chronicle of the Yungang Caves),
(Beijing: Wenwu chubanshe, 2006). Also, to give just a single additional example,
Liu Huida wrote an influential article on meditation and Buddhist caves; see Liu
Huida, “Beiwei shiku yu chan” (Northern Wei Caves and Chan), collected in
Zhongguo shikusi yanjiu, 338.
57 Édouard Chavannes, Mission archéologique dans la Chine septentrionale (Paris:
Ernest Leroux, 1909).
58 Osvald Sirén, Chinese Sculptures from the Fifth to the Fourteenth Century (New
York: Hacker Art Books, 1970), reprint of the 1925 edition by Ernest Benn, Ltd,
London.
59 Ibid., vol. 1, 8.
60 Alexander C. Soper, South Chinese Influence on the Buddhist Art of the Six
Dynasties Period, vol. 32 (Stockholm: B. Museums Far East. Antiq, 1960),
47–112. Also see Alexandra Soper, “Imperial Cave-Chapels of the Northern
Dynasties: Donors, Beneficiaries, Dates,” Artibus Asiae, vol. 28, no. 4 (1966):
241–270; and Alexandra Soper, “Northern Liang and Northern Wei in Gansu,”
Artibus Asiae, vol. 21, no. 2 (1958): 131–64.
61 Soper, South Chinese Influence, 65–6.
18 Introduction
62 Ibid., 64–72. Also see Soper, “Imperial Cave-Chapels of the Northern Dynasties,”
243–44.
63 Soper, South Chinese Influence, 58.
64 Ibid. The issues of stylistic changes and sinicization have been a much-discussed
topic. Soper is not alone in believing the southern influence on the development of
style in the late Northern Wei; Stanley K. Abe, Laurence Sickman, Emma Bunker,
Su Bai, Wen Fong, Katherine R. Tsiang, Yoshimura Rei, Yagi Haruo, Hinako
Ishimatsu, and others have also addressed this. Su Bai argues that the stylistic
changes in the Northern Dynasties are closely associated with those of the paint-
ers (Dai Kui 戴逵, Gu Kaizhi 顧愷之 [345–406], and Lu Tanwei 陸探微 [?–485]
and Zhang Sengyou 张僧繇 [d.u.]) in the south; see Su Bai, “Beichao zaoxing
yishu zhong renwu xingxiang de bianhua,” in Su Bai, Zhongguo shikusi yanjiu,
349–54. Also see Fong Wen, “Chuan Gu Kaizhi ‘Nüshizhen tu’ yu zhongguo
gudai yishushi (Admonitions of the Instructress to the Court Ladies attributed to
Gu Kaizhi and Chinese art history),” Wenwu, no. 2 (2003): 82–96. Stanley Abe
dedicated an entire chapter to discussing sinicization, with a focus on Yungang
and early Longmen caves; see Stanley Abe, Ordinary Images (Chicago: University
of Chicago Press, 2002), 173–257. Also see Laurence Sickman and Alexander C.
Soper, The Art and Architecture of China (New York: Penguin Books, 1971), 90;
Emma C. Bunker, “Early Chinese Representations of Vimalakīrti,” Artibus Asia,
30, no. 1 (1968): 28–52; Katherine R. Tsiang, “Changing Patterns of Divinity
and Reform in the Late Northern Wei,” The Art Bulletin, 84, no. 2 (1 June 2002):
222–45; Yoshimura Rei, “Donyō Gokutsu zōei shidai,” Bukkyō geijutsu, 212
(1994): 11–36; Yagi Haruo, Chūgoku Bukkyō bijutsu to kan minzokuka: Hokugi
jidai kōki o Chūshin to shite (Sinicization of Chinese Buddhist Art), (Kyōto-shi:
Hōzōkan, 2004); Yoshimura Rei, Chūgoku Bukkyō zuzō no kenkyū (Study on
Chinese Buddhist Art), (Tokyo: Tōhō Shoten, 1983); Hinako Ishimatsu, Hokugi
Bukkyo zozoshi no kenkyu (Buddhist Images in the Northern Wei), (Kunitachi-
shi: Buryukke, 2005). In contrast, Albert Dien objected the theory of southern
influence and argued that the stylistic change was northern rather than southern;
see Albert E. Dien, Six Dynasties Civilization (New Haven, CT: Yale University
Press, 2007), 315.
65 James O. Caswell, The Written and Unwritten: A New History of the Buddhist
Caves at Yungang (Vancouver: University of British Columbia Press, 1988).
66 See Vanderstappen, “Book Review,” 125–7, Stanley Abe, “Book Review,” Jour-
nal of the American Academy of Religion, 58 (Spring 1990): 116–18; Robert F.
Campany, “Book Review,” Journal of Asian History, 24, no. 1 (1990): 96–7;
Victor Mair, “Book Review,” Harvard Journal of Asiatic Studies, vol. 52, no. 1
(June 1992): 345–61; Victor Mair, “Reply to James O. Caswell’s ‘Sinology or Art
History,’ Pt II: A Response to Victor H. Mair,” Early Medieval China, 1 (1995):
147–49; and T. H. Barrett, “Book Review,” Bulletin of the School of Oriental
and African Studies-University of London, 55 (1992): 359–60.
67 Nancy S. Steinhardt, Chinese Architecture in an Age of Turmoil, 200–600 (Hono-
lulu: University of Hawai‘i Press, 2014).
68 Nancy Steinhardt’s Liao Architecture (Honolulu: University of Hawai‘i Press,
1997) also provides important analysis of the Liao and Jin monasteries and Bud-
dha halls.
69 Nancy S. Steinhardt, Chinese Imperial City Planning (Honolulu: University of
Hawai‘i Press, 1990), 5.
70 Dorothy C. Wong, Chinese Steles: Pre-Buddhist and Buddhist Use of a Symbolic
Form (Honolulu: University of Hawai‘i Press, 2004).
71 Weishu, 113:3043. Soper believed that the figure 802,366 refers to the “work
days” on the caves, but other scholars suggest that it indicates the cost of the
caves. See Soper, Literary Evidence, 102; Jacques Gernet, Buddhism in Chinese
 Introduction 19
Society: An Economic History from the Fifth to the Tenth Centuries (New York:
Columbia University Press, 1995), 15; Caswell, The Written and Unwritten, 190;
and Amy McNair, Donors of Longmen: Faith, Politics, and Patronage in Medi-
eval Chinese Buddhist Sculpture (Honolulu: University of Hawai‘i Press, 2007),
59 and n. 46.
72 The Longmen caves are filled with dedicatory inscriptions. Thus far, more than
3,600 inscriptions have been identified. See Gong Dazhong, Longmen shiku yishu
(Shanghai: Shanghai renmin chubanshe, 1981), 34.
2 The sacred site of Yungang

Making the sacred cave temples of Yungang

The tradition of rock-cut Buddhist cave temples originated in ancient India,
where they were used for liturgical purposes and daily use. The use of rock-
cut caves, however, traces its origins to the cave tombs of the powerful nobles
of the Middle Kingdom (mid 11th to 13th Dynasties, ca. 2030–1640 BCE)
in ancient Egypt. In ancient Iran, Darius the Great (c. 550–486 BCE) and
his successors in the Persian Achaemenid Empire also carved monumental
tombs into cliffs. Inspired by the cave tombs, people of the New Kingdom
period (18th to 20th Dynasties, ca. 1570–1069 BCE) built the great rock-
cut cave temples of Ramses the Great (19th Dynasty, ca. 1304–1237 BCE),
the Abu Simbel Temple, in southern Egypt. The four colossal statues of the
deified Ramses seated at the entrance to the temple are the world’s earli-
est works of monumental sculpture. Because the interior of the rock-cut
cave temples followed the standard layout of ancient Egyptian temples with
gradually diminishing vestibules and hypostyle halls, it can be said that the
cave temples of ancient Egypt not only evolved out of cave tombs, but also
were constructed in imitation of freestanding temples. The same is true of
the Buddhist rock-cut cave temples in Asia. Although the Buddhist rock-cut
cave temples in India are not directly linked to the royal cave tombs in the
Valley of the Kings of the New Kingdom period in Egypt or the cliff tombs
of the Achaemenid kings at Naqsh-i Rustam in ancient Iran, there is little
doubt as to the influence of these cave tombs on the cave temples in India. In
both cases we are dealing with works of sculpture rather than architecture,
and in both cases there was an appeal in the very permanence that was
promised in the carving out of tombs or temples from the very bones of the
earth.1 The façades of the royal tombs are carved in much the same way as
those of the Indian chaitya (A Buddhist shrine with a stūpa) halls. Buddhist
rock-cut caves were influenced not only by ancient cave tombs but also by
freestanding temples. As Li Yuqun observes, Buddhist rock-cut cave temples
were actually modeled after freestanding monasteries, and a Buddhist cave
temple is in fact a monastic structure carved into the cliffs that run along
a river.2 Although the structural layout of a cave temple differs from that of a
 The sacred site of Yungang 21
freestanding monastery, large cave temples usually have the same liturgical
function.3 They are essentially the same, except that a cave temple is remotely
located along a river. In his Da Song sengshi lüe (Brief History of The Clergy
of The Great Song), monk Zanning of the Song dynasty (960–1279) dis-
tinguishes between six types of monastic appellations: ku 窟 (cave temple),
yuan 院 (cloister), lin 林 (forest), miao 廟 (temple), lanruo 蘭若 (hermitage),
and putong 普通 (ordinary).4 He further notes that ku are cave temples, such
as those carved into the mountain cliffs in the Northern Wei to shelter monks
and sacred images.5 It is thus evident that rock-cut cave temples are a type of
monastery with a different appellation. From this, it can be seen that Yun-
gang is actually a rock-cut monastery hewn from the mountain cliffs. That
is, it is essentially a grand stone monastery with sculptures.
The primary causes for the construction of the rock-cut cave temples of
Yungang were two: the decline of belief in the Dharma (mofa 末法), and a
desire to secure blessings for the imperial family of the Northern Wei.
First, the wide influence of the mofa belief, which had spread to Pingcheng,
and the first persecution of Buddhism by Emperor Taiwu (r. 424–52) gave
immediate impetus to the creation of the cave temples in Yungang. In 446,
Emperor Taiwu, who was enraged at the lawlessness of the śramanas, pro-
mulgated a decree calling for Buddhist persecution, ordering the killing of
the śramaṇas, and burning and breaking of Buddha images. “If from now
on there be any who dare serve the barbarian gods or make images, statues
or figures in clay or bronze, they shall be executed with their whole house-
hold.”6 “Buildings and stupas, and all places where the promulgation of Bud-
dhism was carried out were completely destroyed.”7 This further deepened
belief in the decline of the Dharma, which would entail catastrophes that
would destroy the world at the end of a great aeon and confirmed the impres-
sion that the Way of Salvation was fading. The mofa belief in China started
as early as the Sixteen Kingdoms (304–439) and the Northern Liang (397–
439) era. It was closely related to the translation of the Mahayana text, the
Nirvana Sutra, in Liangzhou (present-day Wuwei 武威 in Gansu), and also
to a historical background of social unrest.8 It was believed that 1,500 years
after the nirvana of the Buddha, demons would leap and cheer, destroying
stūpas and monasteries, killing bhiksus, ruining the Buddha Dharma, thus
driving Buddhism into a final period of darkness.9 Buddhist devotees were
therefore eager to protect the Buddha Dharma by preserving Buddhist scrip-
tures and reliquaries. However, stūpas, monasteries, and images can easily be
destroyed, and would not escape from being ruined if another catastrophe
ever happened. Buddhist cave temples, carved out of the very bones of earth,
on the other hand, are solid and durable, and thus the creation of Buddhist
cave temples became an important means of protecting the Buddha Dharma.
Secondly, the cave temples in Yungang were constructed in order for the
imperial family of the Northern Wei to accumulate merit and virtue, and
pray for blessings. The endeavor had an intensely political character. The
“Treatise on Buddhism and Daoism” in the Weishu records that from the
22 The sacred site of Yungang
time that the capital was transferred to Pingcheng until the persecution of
Buddhism, all of the emperors, including Emperor Taiwu, were Buddhist
believers, but the virtuous action was the first priority of their belief.10 This
was especially true after the persecution of Buddhism, which functioned as
a powerful stimulus to the rapid development of Buddhism and Buddhist
monasteries in Pingcheng. The rulers of the Northern Wei supported the for-
eign religion to monumentalize themselves, and Buddhist cave temples were
constructed essentially as memorials. The cave temples themselves provide us
with the political and social history of the epoch. In addition, after experienc-
ing catastrophic suppression, Buddhists were further convinced that it would
be difficult to establish Buddhism without the support of the rulers.11 They
firmly believed that the rulers of the Northern Wei would directly influence
the rise and fall of Buddhism. Therefore, image-making and construction of
the cave temples were intimately tied to the Northern Wei imperial family.
In 452, Emperor Wencheng succeeded to the throne. He wasted no time in
restoring Buddhism and rebuilding monasteries and statues in repentance for
what his grandfather had done. When the entire empire received the man-
date, desecrated Buddhist images and monasteries were promptly restored
and repaired. Soon, Buddha images, scriptures, and treatises all reappeared
openly.12 That same year, officials were commanded by imperial edict to
make a stone Buddha image resembling the emperor himself. When it was
finished, on the face and the soles of the feet were black pebbles, resembling
the moles on the upper and lower parts of the emperor’s body.13 This aston-
ishing resemblance appeared to confirm the doctrine that the monk Faguo
had promoted: that the emperor was the Tathāgata incarnate.14 It can be said
that image-making was not only related to the emperor, but also was literally
modeled upon him. The following year (453), Tanyao, having assumed the
official position of monastic comptroller (shamentong 沙門統), petitioned
the emperor to have colossal Buddha images and cave temples carved along
the cliff of Wuzhou Mountain outside the Northern Wei capital, Pingcheng.
The tallest image would be 64 feet, and the next 55 feet, in height. Carvings
and embellishments would be wondrous and magnificent, and the best of
the time.15 These are the so-called five Tanyao caves (presently numbered as
caves 16–20) of the first-phase in Yungang.
It can be seen that the cave temples were constructed during this fervent
revival of Buddhism as the Northern Wei court strove to expiate its misdeeds.
The first persecution of Buddhism and the mofa belief were the direct causes
of this. And since the main purpose of the Northern Wei rulers’ belief in
Buddhism was to accumulate personal merit, making images and buildings
came to be characteristic of Northern Wei Buddhism.16
Finally, the relationship between the cave temples and meditation practice
must be addressed. Scholars usually emphasize the close association between
the two, as described by Gao Yun 高允 of the Northern Wei: “One excavates
a transcendent cave in order to reside there for meditation practice 鑿仙窟
以居禪.”17 However, I would propose, based on the recent archaeological
 The sacred site of Yungang 23
findings and other factors, that the cave temples of Yungang are not directly
related to meditation practice, which, in fact, was conducted in the monks’
quarters above the caves. This will be discussed in the Archaeological
Discoveries section later in the chapter.
One may argue that the construction of the rock-cut cave temples in Yungang
was intimately related with Buddhist dhyāna practice (meditation concentration)
considering the fact that the founder of Yungang (Tanyao) was a well-known
dhyāna master, and came, like many other eminent monks (Huishi 惠始, Xuan-
gao 玄高, and Shixian 師賢), from Liangzhou where meditation practice was
popular,18 and that Buddhism in the Northern Wei especially emphasized medi-
tation. Ding Mingyi even suggested that the colossal images in the five Tanyao
caves were the primary objects for visualization, and the large caves were used
to gather monks for meditation purposes.19 Su Bai also proposed that the large
caves in Yungang were used to gather Buddhist monks for meditation.20 The
fact is that there is no space for meditation at all in these colossal image caves.
One may also argue that the increased construction of Buddhist cave temples is
related to the emphasis on meditation during the Northern Wei, which developed
simultaneously, and that because of the close relationship between cave temples
and meditation, when the latter flourished during the period of the Sixteen King-
doms and the Northern Dynasties, cave temples also flourished in the northern
central plains.21 On the other hand, we must not forget the important fact that
meditation practice usually requires quiet or silent surroundings, and was thus
associated with deep valleys between mountain cliffs. Much extant literature indi-
cates that meditation practice should be conducted in rock-cut caves, where the
atmosphere is tranquil and quiet. The Chan miyao fa jing (The Scripture on the
Esoteric Essential Methods of Meditation) observes that, “when one is emerging
from meditative concentration, one should be in a quiet place such as a grave-
yard, beneath a tree, or in an aranya (forest).”22 Similarly, the Zuochan sanmei
jing (The Scripture on Concentration in Sitting Meditation) notes that meditative
concentration should be practiced in a tranquil place such as a forest.23 The fu
fazang yinyuan zhuan (A History of the Indian Patriarchs), translated by Tanyao
in Yungang, also pointed out that monks should practice meditation in caves and
between mountain cliffs in deep valleys so that the wind, cold, and other hard-
ships could be tolerated.24 As such, I would argue that the rock-cut cave temples
of Yungang, in fact, are not directly associated with meditative concentration,
since the cave temples are not situated in the valleys between mountain cliffs.
Rather they are on the vital communication line between the old (Shengle 盛
樂, present-day Helinge’er 和林格尔 in Mongolia) and new (Pingcheng) capitals
by the road. The remains of ancient road ruts in front of the cave temples in the
east section can still be seen today (fig. 2.1). Moreover, most of the caves in Yun-
gang are cave temples devoted to images or stūpas for worship (libaiku 禮拜窟);
that is, they are intended for liturgical practices such as worshipping, visualiza-
tion, and making offerings, etc. Therefore they are not ideal for concentration.
The monks’ quarters (僧房) on top of the cave temples, away from the commu-
nication road, on the other hand, offered us new clues. Recent archaeological
24 The sacred site of Yungang

Fig 2.1 Ancient road ruts

findings reveal that there were freestanding temples above the rock-cut cave
temples with monks’ cells.25 There is some distance between the road and the
cells and thus they would be better for meditation. “The Guanfo sanmei fa
(Methods of Visualizing the Buddha Image and Dhyāna Samādhi)” in The
 The sacred site of Yungang 25
Siwei Lüeyao Fa (Outlining the Methods of Contemplation) mentions that after
one visualizes the Buddha image, one should return to a quiet place and close
one’s eyes to contemplate (還於靜處, 閉目思維), keeping the Buddha image in
mind without distracting thoughts.26 This shows that visualization and meditation
took place in two different places. After visualizing the Buddha image (in a cave
temple), one would return to a quiet place to meditate. In Yungang, one would
return to monks’ cells above the caves in the mountain to concentrate. Buddhist
monks and devotees would conduct ritual practices down below in the cave
temples and then “return to a quiet place, and close their eyes to contemplate”
in the cells above. The cave temples for worship in Yungang essentially function
in tandem with monks’ quarters rather than with cave temples for meditation, as
in many other cave temples. In Kizil, for example, as Angela Howard observed,
the ubiquity of monastic cells and their unique typology clearly show that medita-
tion was given special consideration in the construction of these monasteries.27 In
addition, the Luyeyuan 鹿野苑 cave temples in Pingcheng and the Tuyuk 吐峪溝
cave temples in Turfan 吐魯番 (Xinjiang) both also have meditation caves near
the worship cave temples.28 The structure conforms to the need for visualization,
worship, and meditation. This, however, does not fit that of the Yungang cave
complex. Meditation practice was carried out in monks’ quarters above the cave
temples.

Auspicious geographical environment of Yungang

Wuzhou Mountain was an important northern frontier as far back as
the Han dynasty. The Weishu noted that about 1,100 Xiongnu horsemen
attacked Wuzhou Fortress in the first year of Yuanguang 元光 (134 BCE).29
The flat top of the mountain extends north and south for more than a half
mile. The north end of the mountain connects with high mountains, but the
south and west sides face the river. The east side of the mountain is separated
from the low mountains by a wide north-south dale. The western section of
the mountain is high, and is now divided into western and middle sections
by a dale. All the cave temples were perpendicularly carved into the south
cliff of Wuzhou Mountain on the upper terrace landform (plate 2), which
had been naturally formed through erosion by the river, while the lower
terrace was transformed into a road that ran along the bank connecting
the new and old capitals of the Northern Wei. The top of the mountain
was flat and vast, which allowed it to be an active site even before the Yun-
gang caves were excavated. Literary documents reveal that Buddhism in
Pingcheng flourished considerably even before Emperor Taiwu’s persecution
of Buddhism. In the first year of Tianxing 天興 (398), an edict commanded
that the five-story (tall) Buddha, the Mount Grdhrakuta (Vulture Peak), and
Mount Sumeru Halls be built, painted, and decorated.30 In addition, a lecture
hall, a Dhyāna hall, and residential cells for śramanas were constructed.31
These were the earliest Buddhist temple halls. When Emperor Mingyuan
ascended the throne, he held the doctrine of the Buddha in high regard.
26 The sacred site of Yungang
Thereupon, images and statues were set up in all corners of the capital,
and the śramanas were commanded to guide the laity (fudao minsu 輔導
民俗).32 Mention was also made that as early as the third year of Yongxing
永興 (411), Emperor Mingyuan prayed to the gods of heaven and earth on
Wuzhou Mountain.33 After ascending the throne, the emperor visited the
mountain annually to pray and offer sacrifices.34 The Jin stele inscription
states that Emperor Mingyuan built the Tongle Temple (of the ten temples in
Yungang) first, that Emperor Wencheng continued with Lingyan Temple, and
that the Huguo and Tiangong Temples were built under Emperor Xiaowen.
It can be seen that Wuzhou Mountain was an active sacred mountain. Its
unique topography with a keystone of smooth terrain on top was not only
ideal for praying and making offerings but also suitable for opening the
mountain, chiseling stone, and making rock-cut cave temples. This spiritual
and numinous mountain was believed to have brought blessings and protec-
tion, so it was chosen to be the site of the Yungang cave complex.
Written records and recent archaeological excavations both indicate that
Yungang was not the only cave temple on Mount Wuzhou; mention was
also made of other cave temples along the Wuzhou River. Li Daoyuan noted
that beside the Wuzhou River, there were stone monasteries where Bhikk-
hunis (nuns) lived. From there, the water turned southeast, passing south of
Lingyan (i.e., Yungang).35 On this basis, it can be deduced that the Yungang
complex was situated to the southeast of the nunnery. Daoxuan of the Tang
also mentioned that the monastery in the east was the Lingyan cave temple,
and the one at the west was a nunnery (東為僧寺, 名曰靈巖, 西頭尼寺).36
For a long time, this had puzzled scholars, and there was no clear evidence
suggesting the nunnery’s location. In 1956, a dedicative stone slab dated to
the fourth year of Jingming 景明 (503) by a certain nun Tanmei was dis-
covered in the pile of collapsed rocks in front of cave 20 (plate 3). Based on
this finding, researchers in the Yungang guji baoyangsuo inferred that “the
nunnery was perhaps located near cave 20.”37 Su Bai later adopted the same
view.38 However, all the cave temples in Yungang were for worshipping and
liturgical practices. The primary focus of a rock-cut cave was either images
or stūpas. No residence cells associated with the caves have been found. In
addition, all the archaeological excavations in front of the caves from the
early 20th century (1938) until now prove that there was no space in front
of the caves for residence cells (archaeological excavations will be discussed
later in the chapter). Furthermore, the geographic environment described
by Li Daoyuan in the Shuijing zhu has not actually changed much since his
time except that the river was pushed southward, and some images eroded
and some cave roofs collapsed over time. The rest has remained largely the
same.39 The two archaeological excavations (1940 and 1992) suggest that
the original river was only 60–90 feet south of the caves.40 The river was
very close to the caves initially. According to the Jin stele inscription, the
river was diverted about a half mile south of the caves only in 1131 during
the Jin dynasty. In addition, there are other dedicative inscriptions by nuns
 The sacred site of Yungang 27
(cave 11, dated to the 19th year of Taihe, and cave 17, dated to the 13th
year of Taihe), but we cannot thereby establish that these two caves were
the nunneries. Similarly, we cannot conclude that the nunnery is near cave
20 simply because a stone slab dedicated by a nun was found in front of
it. Most importantly, both Li Daoyuan and Daoxuan mentioned that the
nunnery was situated to the west of the Yungang complex. We must look
elsewhere for the nunnery, not in front of the caves. The nunnery should be
outside the current Yungang complex.
In recent archaeological excavations, three surviving cave temple remains
(the Lubanyao cave temple complex 魯班窯, Jiaoshansi cave temple com-
plex 焦山寺, and Wuguantun cave temple complex 吳官屯) were found
along the river (plate 4). Judging from its location and distance from the
river and Yungang, we now believe that the Lubanyao cave complex, situ-
ated to the west of Yungang, is likely to be the site of the vihara caves
where the nuns lived.41 From here the water runs southeast, passing Yun-
gang, exactly as described by Li Daoyuan.42 More recent archaeological
excavations of the Lubanyao cave complex indicate that the residence cells
were likely above the caves, as in Yungang.43 Some scholars inferred that
the Lubanyao cave complex should have been part of the Yungang cave
complex during the Northern Wei, and thus should not be considered an
individual monastery.44 This is not entirely impossible since these two cave
complexes demonstrate great resemblance in iconographic style, whether in
iconography or architectural layout; only the Lubanyao cave complex was
constructed as a nunnery beyond the west end.
The Xu gaoseng zhuan mentions also that the cave temples of Yungang
were aligned closely with one another like the teeth of a comb, and spanned
10 miles. At the east end of Yungang, says this text, is the monastery for
monks, which can hold at least 1,000 people.45 But the Yungang cave tem-
ples in the current complex span only 1 mile, not 10. Scholars questioned
the credibility of this record until the 1950s, when the Jiaoshansi cave
temple remains were discovered 10 miles to the west of Yungang.46 Judging
from both the location and distance from the Yungang complex, scholars
have been led to believe that the Jiaoshansi monastery remains are the cave
temples which spanned 10 miles, as described by Daoxuan, eastward to
Yungang. Thus, the mystery that had long puzzled scholars was solved.

Archaeological excavations and related issues

From the viewpoint of archaeology, the investigation of rock-cut cave tem-
ples requires close attention not just to all sorts of phenomena regarding the
remains of the cave temples themselves such as the cave structure, composi-
tion of the images, subject matter, and iconographic styles, but also to the
grounds associated with the cave temples themselves and the remains sur-
rounding them. These remains are usually associated with the initial design,
excavation, repairs, and reconstruction of the cave temples, and are of great
28 The sacred site of Yungang
academic significance. Therefore, archaeological excavations all around the
Buddhist cave temples, and inventory of the caves themselves, are significant
components of the systematic investigation of any cave temple. The archaeo-
logical excavations in the Yungang cave complex are good examples.47 Since
the 1940s, many archaeological excavations have been conducted in the
Yungang complex. Among them, the excavations conducted in the 1990s
in front of caves 9 and 10, and the two in 2010 and 2011 above cave 39,
and caves 5 and 6 are especially important. These were not only the largest
excavations in scale but also shed fresh light on the functions of different
sections of the Yungang complex. The excavations reveal why there are no
residential cells either within or in front of the cave temples. All of the exca-
vations were primarily in front of and above the caves: four areas above the
caves (east of caves 1, 3, 5 and 6, as well as in the area above cave 39), and
four in front of the caves (caves 3, 8, 9–13, the five Tanyao caves), among
which the excavation in front of caves 9 and 10 is of particular significance.

1. Early excavations of the remains by the Japanese

Japanese scholars first started the excavations between 1938 and 1940.48 In
the process of investigating Yungang, they noted the surviving remains in front
of and above the cave temples and paid attention to the relationship between
these remains and the cave temples. Between 1938 and 1940, they purpose-
fully conducted archaeological excavations in five sites: (a) the remains in front
of cave 8 and Wuhuadong 五華洞 (caves 9–13); (b) the remains in front of the
five Tanyao cave temples (caves 16–20); (c) the remains of the monasteries in
the western section of the complex above the cave temples; (d) the remains of
the monasteries in the eastern section of the complex above the cave temples;
and (e) the remains of the monastery beside the Longwang Shrine 龍王廟.
The excavation in front of cave 8 was conducted in June 1938. The trench
was 3 feet wide and 52 feet long, running north-south. Nothing significant
was found, only a row of piled stones, and their time period has yet to be
determined. The archaeologists suggested that they might be associated with
the stones found in front of caves 9 and 10. The excavation proved that the
bedrock in front of cave 8 extended 50 feet southward.
The excavation in front of caves 9 and 10 took place in May 1938. First, they
dug a north-south trench in front of cave 9 and discovered the base of a column.
Then they excavated along the base of the column to cave 10, in front of which
they discovered traces of lotus and tortoise patterns (plate 5). In the process, a
carving of the base of the western pagoda of cave 10 was found, proving the
earlier existence of a pagoda on the east side of cave 9 and the west side of cave
10 (fig. 2.2). In September 1940, a north-south trench (157 feet long and 6 feet
wide) was dug between caves 9 and 10, and a stratum of piled tiles and bricks
from the Liao dynasty was found at the north end of the trench. At the south
end of the trench, 6 feet below, a stratum of crushed rocks was found. Nothing
significant was found in front of caves 9 and 10, however, until the 1990s.
 The sacred site of Yungang 29

Fig 2.2 East column base, cave 10

Provided by the Yungang Research Academy

The excavation in front of cave 12 was conducted in 1940, when a

north-south trench (154 feet long and 6 feet wide) was dug. Square bricks
from the Liao dynasty were discovered, paving the ground up to 34 feet
south of the cave entrance. The large-scale excavation in front of the five
Tanyao caves took place after the local residences were torn down in Sep-
tember 1940. The excavation was mainly in front of caves 19 and 20. An
inventory was made of cave 17. Floor bricks from the Liao dynasty were
discovered in front of cave 19. Tiles of the Northern Wei were discovered
in front of cave 20. The excavation proved only that the architectural
remains date from the Northern Wei, but could not confirm the entire
cultural stratum. The trench was not dug to the bottom of the cave either.
The ground level of cave 20 is the same as in the side caves of cave 19, but
the height of the stratum containing the Northern Wei tiles only reaches to
the ground level of cave 19. So, in the past, stairs must have been built in
order to climb to cave 20.
Japanese scholars’ excavations in front of the caves revealed architectural
remains from the Liao dynasty in front of the Wuhua caves and cave 19. In
these remains, no Northern Wei structures were found. This indicates that
the structures in front of the caves were started by the Liao. A large quantity
of Northern Wei tiles in front of cave 20 was found, which led archaeologists
to believe that they were from Northern Wei structures.
30 The sacred site of Yungang

Fig 2.3 Tile engraved with characters

Beside the aforementioned excavations in front of the caves, the Japanese

also conducted two excavations above the caves: one in the west above cave 37,
and the other above cave 3 on the eastern terrace. These two surviving groups
of remains proved to belong to freestanding monasteries that had existed in
the Northern Wei.
The monastery remains above the cave temples on the western terrace
are situated to the west of Yungang Fortress. Excavations in October 1940
unearthed scattered tiles and other objects on the ground. The strata are in
an H shape and are primarily divided into four areas: A, B, C, and D. In areas
A and B, at 1 and 1.5 feet below the earth’s surface were piles of fragmented
tiles, unearthed tiles from eaves with a lotus design, and tiles engraved with
the characters “Support the imperial rulers, and extend the prosperity infi-
nitely (chuanzuo wuqiong 傳祚無窮),” as well as the rarely seen green-glazed
tiles (fig. 2.3). To the east of areas A and B were board-shaped rocks paving
the southeast portion of the extended area. Area C had piles of tiles as well.
In areas D and E were remains of brick walls. In addition, to the east of the
H-shaped stratum, excavations were made in areas F and G. More piled tiles
with lotus design and chuanzuo wuqiong characters were found in Area F.
No remains were found in area G.
In short, the excavations were unable to provide a clear understanding
of the buildings’ structure, but showed that monasteries were built there
 The sacred site of Yungang 31
during the Northern Wei, and that the eaves with lotus designs as well as
tiles with the characters chuanzuo wuqiong were used. Also, the flat tiles
near the eaves have undulating forms which indicate they were shaped by
hand. It should be particularly noted that green-glazed tiles were found,
although where they were used remains unknown. These flat tiles are
painted on each side with dark green glaze, and are thus far the oldest
known green-glazed tiles.
The remains of the Northern Wei monasteries on the eastern terrace are
situated on the flat ground above cave 3, and were excavated in August of
1940. An east-west stone-covered platform base was found with the head of
a lion at one end. To the west of the lion, stone-layered ground was found
that seems to have been the base of a structure. To the northwest is a north-
south lime wall painted vermillion. Outside the stone-layered ground is a
large quantity of flat and cylindrical tiles, some with the characters chuanzuo
wuqiong, others with handmade undulating patterns. Also, there are many
Y-shaped tiles. In short, the scale of the structure is moderate, and the exca-
vation found only a red wall structure above cave 3, which can be traced as
far back as the Northern Wei period.
The aforementioned excavations above the cave temples are moderate in
scale, and do not provide a complete picture of the remains of the freestand-
ing monasteries but they do prove that there were monasteries above the
caves during the Northern Wei.

2. Excavations by the Chinese after the 1970s and 1990s

The excavations conducted in the 1970s and 1990s were important and
fruitful, especially those in front of caves 9 and 10. In order to fully under-
stand the scale of these structures, the Institute of Cultural Relics Protection
in Yungang conducted large-scale excavations and made inventories in front
of the caves in 1972. They discovered architectural remains of two phases.49
Eight square holes for columns were found in the first-phase remains, which
line up with the square holes for rafters on the outside wall of the caves.
One of the column bases overlaps with the lotus floor of the Northern Wei
in front of cave 10, crushing it and ruining the integrity of the carved traces
of the floor. This suggests that this structure was seven bays wide (a bay is
the span between two columns) and was built after the Northern Wei and
before the Liao. Archaeologists surmise that this belonged to a “structure
rebuilt by an official in the 15th year of Zhenguan 貞觀 (641) of the Tang”
as recorded in the Jin stele inscription. Six square column bases remain
from the second phase, structures of the Liao dynasty. At the same time,
the remains of a roof and supports in the shape of an inverted V were dis-
covered on the outside wall on top of the badly eroded columns of caves 9
and 10 (plate 6).
The excavations of the above-mentioned architectural remains and the dis-
coveries of the remains of the eaves testify that in the Northern Wei period,
32 The sacred site of Yungang
the exterior of caves 9 and 10 featured a spectacular building in imitation of
a wooden structure five bays wide. This solved a long-standing puzzle about
the eaves of the caves. But it still lacks sufficient evidence to prove whether
or not the structure was rebuilt in the Tang. In 1973, on the upper part of
cave 12, again, a hip roof (庑殿顶) was found. The structure of roof eaves in
imitation of wooden buildings is not only important material evidence in the
study of the Northern Wei architecture in the Pingcheng era, but also had a
direct influence on the excavations of the later cave temples.
From 1992 to 1993, Shanxi Archaeological Institute and the Research
Academy of Yungang conducted thorough excavations and inventories of
the remains in front of the caves and unearthed gilded fragments of Buddha
images as well as four remains of structures. The excavation of the remains
in front of cave 3 was the most important and had great academic value.
Though cave 3 is unfinished, it is the largest court cave in Yungang. In
front, the excavation discovered remains of a grand-scale Liao dynasty struc-
ture nine bays wide. Inside, remains were found of a floor surface chiseled
from the bedrock of the Northern Wei but unfinished. For instance, inside the
east part of the cave, on the uneven floor were traces of chiseled rocks using
notches to divide the rocks into square or round pieces (fig. 2.4). The rock
may have been chiseled into different shapes for the use of column bases or
other structural components used in the construction of the palaces and other
buildings in Pingcheng. These remains fully reveal the process and method
of cave temple construction, all of which were previously unknown. This
is of great significance in the study of the carving process and the method
employed in the construction of large cave temples such as Yungang.50

Fig 2.4 Ruins of excavated stone bricks in cave 3

 The sacred site of Yungang 33
Through the excavations of the ’90s, the remains in front of the five Tan-
yao caves and the Wuhuadong caves were fully revealed and column holes
were discovered in the ground corresponding to the holes for the rafters on
the outside wall. Also, stairs and construction tiles of the Northern Wei were
found in front of cave 20. The remains of large lotus floor patterns carved in
the Northern Wei in front of caves 9 and 10 were discovered, but the detailed
archaeological report has yet to come out.51
In addition, two important excavations above the caves brought to light
the functions of the Yungang complex in each area. For the first time, we can
explain why there were no residential cells in the Yungang caves, and learn
where the monks lived and translated the sacred texts.
For several years, an archaeological team composed of The Shanxi
Archaeological Institute, The Yungang Research Academy, and Datong
Municipal Archaeological Institute conducted investigations and excava-
tions. From 2009 to 2010, they excavated and fully unearthed the monas-
tery remains of the Northern Wei above cave 39 in the western cave section;
in 2011, they excavated the monastery remains above caves 5 and 6 in the
central section caves.52
The surviving remains above cave 39 are those of the monastery that
the Japanese had excavated previously on the western terrace. They face
south, and the south part has been destroyed, but the north part is well
preserved. The remains are 200–205 feet wide from west to east, and 144
feet from south to north. The basic layout of the monastery consists of a
stūpa and residential cells (fig. 2.5). More than 20 residential cells were
found, among which the cells in the north are better preserved – a row of

Fig 2.5 Ruins of a monastery above cave 39

Provided by Zhang Qingjie
34 The sacred site of Yungang
13 cells. Connected to these north cells are three to the east and two to the
west. Only the northern part of these cells has been preserved. The cells were
built above ground and are divided into single cells and suite cells. They are
square or rectangular in shape, all with an entrance. All cells have corridors
too. The column-base stones of the corridors still remain. There are surviving
brick-beds, kitchen stoves, and flues in the cells. Moreover, at the western
end of the northern cells towards the south, there are rows of east-west cells
juxtaposed. The western residential cells connect with the western section of
the northern cells, dividing the monastery into big and small courtyards. The
front porch of the northern cells connects the two courtyards. The east court-
yard is bigger, in the center of which, slightly towards the south, is the base
of a stūpa. This should be a stūpa courtyard; however, there are no remains
in the western courtyard. The base of the stūpa faces south. It is square in
shape (48 feet wide, and 1–2 feet in height) and was tamped. Because the
south part of the monastery was damaged and formed a slope, the situation
of the entrance to the monastery and of the southern cells connected with
the entrance is unclear.
The primary surviving objects unearthed from the excavations include
the construction materials, tiles with lotus patterns, the engraved characters
chuanzuo wuqiong, and flat green-glazed tiles. Remnants of Buddha images
and worshippers, as well as a piece of pottery with the characters 西窟 (west
cave) were found.
In addition, the excavation of monastery remains above caves 5 and
6 is extremely important. It shows that the surviving remains contain a
large octagonal stūpa base of the Liao and Jin dynasties as well as iron
casting remains (fig. 2.6); however, enveloped by the stūpa base is the
Northern Wei square stūpa base 46 feet wide. The dimensions are roughly
the same as the stūpa in the western section. It can thus be deduced that
this area also belonged to the Northern Wei monastery, and when it was
rebuilt during the Liao and Jin dynasties, a new stūpa was constructed
on the base of the Northern Wei stūpa. Aside from the stūpa, the ground
surface of the monastery and the attached structures (residential cells and
corridor) were all destroyed in the reconstruction. Further investigation
is expected, since the scale of the monastery and the extent of its walls
are still unclear.
Many objects from the Northern Wei were unearthed at the two sites,
among which are tiles with the characters chuanzuo wuqiong, fugui
wansui 富貴萬歲 (wealth and longevity), and green-glazed tiles, all of
which are characteristics of Northern Wei imperial buildings. Also found
were remnants of Buddha images and worshippers, as well as a piece of
pottery with the characters “west cave.” The Yungang cave complex was
built in the beginning of the Heping era (460) under the imperial patron-
age of the Northern Wei and was the center for worship and making
offerings. In the 18th year of Heping (494), Emperor Xiaowen moved the
capital southward to Luoyang. As a result, the large-scale construction of
 The sacred site of Yungang 35

Fig 2.6 Ruins of a monastery above caves 5 and 6

Provided by Zhang Qingjie

Yungang, such as cave 3 (the largest cave in Yungang), had to be stopped.

Thus it appears that the two sets of remains were most probably con-
structed before the transfer of the capital.

3. Significance of the excavations in Yungang

The archaeological excavations in Yungang are of great academic signifi-
cance. They provide scientific evidence for the study of the area around
Yungang before and after the excavation and construction of the caves.
First, the excavations reveal the locations of the ancient road and the course
of the river. Yungang is situated to the west of Pingcheng on the arterial road
of communications to the old capital Shengle. It was where the Northern Wei
imperial court prayed for blessings before the Yungang cave temples were
built. The archaeological excavations show that the road passed by the front
of the Yungang cave complex; the remains of ruts can be seen in front of the
eastern section of the caves. Furthermore, the Wuzhou River was not as far
away as it is today; it ran right in front of the caves. Using the excavation in
front of Wuhuadong as an example, the original form of the sloping mountain
in front of caves 9 and 10 before they were created can be seen in the sectional
drawing.53 It was connected to the slope in the south of the stratum before
the mountain was chiseled. Below the area 6 feet south of the stratum is the
36 The sacred site of Yungang
layer of crushed rocks, which were deposited there in the process of chiseling
the mountain. South of the crushed rocks is the sand stratum. We can deduce,
then, that the river was initially very close to the caves.
Secondly, the excavations shed light on the process of construction. The
project of constructing the cave temples required the design to be well-
organized and preplanned. The unfinished court cave, cave 3, provides a
good example of the process of excavating such large-scale caves. The cave
is composed of an anteroom and main chamber. It should be noted that
the excavation process started by chiseling the internal space first, not, as
some scholars have imagined, with caves in the central section being chiseled
in different layers, i.e., chiseling first the window from which the upper-level
space was created and the images carved, and then moving downward. At
the same time, traces of the rock debris left inside and outside the caves indi-
cate that in the process of excavation, efforts were made to make comprehen-
sive use of the rocks removed by chiseling as material in palace and temple
construction. This is similar to the large-scale constructions in Pingcheng
during the Xiaowen era in the Northern Wei.
Thirdly, the excavations help us assess whether there were structures
in front of the caves. Scholars have long wondered whether monasteries
stood in front of the central section of cave temples in the Northern Wei
period in Yungang. The excavations found no remains of Northern Wei
structures in front of the Wuhuadong and caves 16–19; cave 20, however,
is an exception. Many Northern Wei tiles were discovered in front of cave
20, which suggests that a Northern Wei monastery existed there. It is certain
that the collapse of cave 20 took place soon after the cave was completed
(this will be further discussed in Chapter Four). Therefore, in order to avoid
exposing the big Buddha to the elements, a structure had to be built. Thus,
the outer appearance of the five Tanyao caves is asymmetrical.54

4. Functions allocated to areas as reflected in monastery remains

The Yungang cave temples themselves are all shrine chapels with images
and stūpas. There are no caves allocated to other functions (e.g., provision
of residential space) as in the Kizil caves in Xinjiang. The caves were used
primarily for worshipping, repentance, making offerings, chanting, and pos-
sibly jiangjing 講經 (sutra lecture), changdao 唱導 (vernacular sutra singing
and preaching), or merely for merit and virtue accumulation in some caves,
as I shall argue. For example, the five Tanyao caves, as mentioned previously,
functioned as a family shrine, which differentiates them from other caves.
Also, caves 7 and 8, and cave 12, which have spacious anterooms and narra-
tives on either side, may have functioned as lecture halls, and caves 9 and 10,
with their vaulted passageway for circumambulation, could have been used
for repentance. The functions of these spaces are worth further consideration.
So where are the buildings in Yungang in the Northern Wei that could have
accommodated more than 3,000 monks with living quarters, as recorded in
 The sacred site of Yungang 37
the documents?55 Also, where were the temples in the Yungang complex that
Tanyao translated the sutras during the third year of Heping (462)? As men-
tioned before, the Wuzhou River flowed right in front of the Yungang cave
temples, so there was no space for freestanding monasteries. The excavations
above the cave temples provide answers to these questions. Of the three mon-
astery remains, the western section included a stūpa courtyard and a side
courtyard. Beside the stūpa in the courtyard, which had a liturgical function,
some of the residential cells with anterooms and main chambers in the stūpa
and side courtyards were equipped with brick-beds, flue-stoves, and kitchen
stoves. This indicates that the residential cells were in fact living quarters. This
differentiates them from the cave temples themselves, which were merely for
worship, offerings, and visualization, etc.
Cao Yan in fact mentioned, in the 1147 Jin stele inscription, the ten tem-
ples and the “stone chambers 石室” in which monks translated the sutras in
the Yungang complex, but for a long time, we had no hard evidence to verify
the authenticity of the inscription before these new excavations revealed the
full picture of the site. According to the Jin stele, there were ten temples in
the great rock-cut cave-temple complex in the western capital, and they were
constructed during the Northern Wei. Mention was also made of several
stone chambers above (the cave temples) where the Indian monks translated
sutras.56 Also mention was that Emperor Mingyuan had the Tongle Temple
built first, after which Emperor Wencheng followed and had the Lingyan
built; whereas the Huguo and Tiangong temples were built by Emperor
Xiaowen, and the Chongfu temple completed by Qian’er (Wang Yu).57 The
Xu gaoseng zhuan also mentions that Tanyao lived at the Tongle Temple
in 460 within the Yungang cave complex as the chief administrator, and in
the third year of Heping (463) had already gathered many eminent monks
in the Beitai 北台 (North Terrace) cave-temple where he translated sutras.58
The Gujin yijing tuji (Seal of Ancient and Modern Sūtra Translation) also
mentions that Tanyao was the monk of the Heng’an cave-temple (Yungang),
and that when Buddhism was revived in the third year of Xing’an (453), he
gathered the Indian monks to translate the Jingtu sanmei jing (Samādhi Sūtra
on the Pure Land), and the Fufazang yinyuan zhuan.59 Now, with the new
archaeological discoveries in the area above the caves, it has become clear
that the northern terrace (Beitai) should be the vast terrain above the caves
and that the “stone chambers” were the monasteries.
The differentiation between these functions suggests that freestanding
monasteries were allocated to Yungang, as the grand Northern Wei royal
rock-cut cave monastery. These monasteries were where Tanyao translated
sutras and where the monks lived. Therefore, the remains of the monastery
above the cave temples are important and inseparable components of the
Yungang complex. At the same time, they also confirm what the Shuijing
zhu described as the spectacular scene of Yungang: “the Buddha Halls on
the mountain and over water, and the smoke of the incense-filled temples
overlook each other 山堂水殿, 烟寺相望.”
38 The sacred site of Yungang
5. Configuration of the monasteries in Pingcheng
as reflected through remains
The monastery above the caves in Yungang features a double-courtyard
configuration. The stūpa courtyard is centered around a stūpa, with residen-
tial cells radiating outward. Because residential cells divided the space, no
courtyard walls were discovered in the excavations. During the Pingcheng
era, Buddhism flourished, and in the beginning of the Heping era under the
reign of Xiaowen, there were more than 100 monasteries in the capital, old
and new, and more than 2,000 monks and nuns.60 However, the records are
usually unclear and vague; thus, the configuration of the monastery is dif-
ficult to verify. In the past, only the Siyuan monastery 思遠靈圖 (plate 7),
constructed under the patronage of Empress Dowager Feng in the third year
of Taihe (479) in the south of Fangshan 方山 in Datong, was investigated
and excavated.61 The layout of the monastery features the stūpa at the front
and halls at the rear, with walls surrounding them. This layout is different
from the monastery above the caves in Yungang. Thus we can see that in
the Pingcheng period, aside from the layout with a stūpa in front and halls
at the rear (e.g., Siyuan monastery), the early monastery configuration with
a single stūpa was also popular.

Art and architecture

Yungang involved the infusion of various cultural, material, and religious
exchanges between the west and east, north and south, as well as religious
and secular influences, all incorporated within a Buddhist rock-cut cave tem-
ple complex, therefore, the art, architecture, and liturgical rituals in Yun-
gang are the result of rather complicated cross-cultural phenomena. Such
intricacy does not appear in any other Buddhist cave temples in China, or
even the whole of Asia. The architecture, imagery, and liturgy in Yungang
all bear marks of multicultural origins. One can observe Greek and Roman
influences, and Indian and Gandhara influences, as well as the implantation
of elements from Xinjiang and Liangzhou, an influence that is especially
significant. Early in the Tai’an 太安 (455) period, five barbarian śramanas,
Yasagupta, Buddhanandi, and others, from the Land of the Lion (Sri Lanka)
arrived at the capital (Pingcheng) bearing three Buddha images.62 They all
claimed that they had traveled the countries of the Western Regions and
seen the Buddhist images.63 A barbarian monk from Shale 沙勒 (Kashiga 喀
什) also came to the capital and presented the Buddha’s begging bowl and
paintings of the Buddha image.64 The Weishu noted that when Emperor
Taiwu conquered Liangzhou in the fifth year of Taiyan 太延 (439), some
30,000 people were forced to move to Pingcheng.65 The śramanas and Bud-
dhist practices moved eastward, and thus the images and doctrines both
flourished more and more, and Buddhist image-making rapidly increased.66
The people of Liangzhou had believed in Buddhism for generations, starting
 The sacred site of Yungang 39
from the time of Zhang Gui 張軌 (255–314).67 It had been the center of
Buddhism in the northwest. The infusion into Pingcheng of Liangzhou Bud-
dhism, which had undergone great development since the Sixteen Kingdoms
period, marked a turning point in the development of Northern Wei Bud-
dhism in the capital. There can be little doubt that migrants from Liangzhou
brought not only images with them but Buddhist liturgical practices as well.
When Buddhism was introduced to China, it not only brought the faith to
the country, it brought religious practices too. Thus, Buddhism introduced
not only images and temples to China, but ways of worshiping as well.
The influence from Xinjiang and Liangzhou co-existed, but each had a
different impact on Yungang. Xinjiang had more impact on art and archi-
tecture, whereas Liangzhou exerted more influence on religious beliefs and
practices. The similarity between the architectural structure of the Kizil
caves in Xinjiang and those in Yungang strongly suggests the close asso-
ciation between them. The passage behind the seated Buddha in the main
chamber of caves 5, 9, and 10 shares similarities with that in Xinjiang;
also, the chaitya caves are undeniably influenced by the central pillar caves
in Xinjiang. Furthermore, most of the caves in the second phase have an
anteroom and main chamber, which indicates a strong connection with the
cave structures in Xinjiang. All demonstrate that Xinjiang had a strong
impact on Yungang. One important question is where the tradition of colos-
sal images comes from.
It should be pointed out that the population in Liangzhou was diverse and
had multiple ties with other regions. Many people had come from other king-
doms before being forced to migrate there. Between Liangzhou and many
places in central Asia, there were close political and economic ties. Chinese
Buddhist rock-cut cave art started from ancient Kucha in Xinjiang. When
it was disseminated to Liangzhou, it became Liangzhou Buddhist art with
strong characteristics of the Western Regions. As mentioned earlier, Liang-
zhou Buddhism exerted more influence on religious beliefs and liturgical
practices, and more on the second-phase caves than the Tanyao caves (first-
phase caves). This is clearly reflected in the architecture, subject matter and
liturgy. The devotional focus in the second phase was either on Maitreya
Buddha or on the stūpa, which basically were the emphases of Liangzhou
Buddhism. In addition, the cave structure during this period differed. It broke
from the precedent of focusing on the colossal Buddha image in the cave and
focused more on devotional rituals and practices, both of which were empha-
sized by Liangzhou Buddhism. It can be said that all cave temples in the
second phase are associated with them. Caves 5, 9, and 10, for example, are
all fodianku 佛殿窟 (image shrine caves) for worship, and behind the main
image in each cave is a vaulted passage; also, Buddhist devotees depicted
along the passage all face rightwards, indicating that one of the key functions
of the caves was circumambulation. In addition, five chaitya caves (1, 2, 3,
6, and 11) were built during this period for worship and circumambula-
tion, etc. Evidently, different architectural layouts served different functional
40 The sacred site of Yungang
purposes. As Walter Spink observes, different architectural layouts reflect
differences among the rituals performed in the caves.68 By the second phase
of Yungang, religion, ritual, practice, architecture, economy, and art had
all become one integrated and interrelated entity. Each supported, and was
interrelated with, the others. Moreover, the subject matter in caves 7, 8, 9,
and 10 was influenced by the Maitreya Sutra, as indicated by the main image
(Maitreya) of the caves, symbolizing the well-known ascent to and descent
from the Tushita Heaven of Maitreya.
After the Northern Wei unified north China, envoys began to be sent to
the capital from the Western Regions and interaction between the countries
began.69 Mention was made in the Beishi that many merchants from Sogdia
traveled to the Northern Liang to trade, and after Guzang 姑臧 (modern
Wuwei 武威, Gansu) was attacked, they were all captured.70 It can be seen
that Sogdians were among the migrants to Pingcheng. In the archaeological
findings from the tomb of Sima Jinlong 司馬金龍 (484 AD) and the tombs
excavated at Yanbei shiyuan, some huren 胡人 (ethnic groups) images, clay
camels, and small clay tents in imitation of huren houses with windows on
top, were unearthed. Not only did Sogdian merchants go to Pingcheng, as
indicated in the iconography of Yungang, but musicians were among the
migrants as well. By the time the Northern Wei captured the Northern Yan
北燕 (407–36) in 436, in the Taiyan era (435–40) of Emperor Taiwu, Sog-
dian musicians had already arrived at Pingcheng. In Yungang, many caves
have the depictions of musicians playing konghou 箜篌, pipa 琵琶, paixiao
排簫, yaogu 腰鼓, hengdi 橫笛 and other instruments, which are associated
with India, central Asia, and many other regions. Sogdian dance was intro-
duced to Pingcheng as well. The ink stone discovered in Datong in the 1970s
shows a hutengwu 胡騰舞 dancer accompanied by a man playing a pipa.71
The ethnic dancers are also seen in the Shaling 沙嶺 tomb mural paintings
(plate 8).
It must be emphasized, however, that the most direct influences on Yun-
gang were of a local nature. This can be seen in the objects excavated in
recent years outside Pingcheng. As Su Bai aptly points out, the most direct
influence on Yungang came from local Buddhist monasteries, which in turn
were influenced more and more by the elements deriving from the central
plain (Chang’an and Hebei) and from southern China.72 From the Xing-
guang 興光 era (454–55) to the first year of Taihe (477), 25 years after the
revival of Buddhism, more than 100 monasteries were built in Pingcheng.73
In addition, influences between secular and religious material culture should
not be overlooked. The images and mural paintings from the Sima Jinlong,
Shaling, Song Shaozu 宋紹祖 and many other tombs unearthed in and
around Pingcheng show that the relationship between secular and religious
culture, whether in images or architecture, was strong and reciprocal.74 The
clay figures unearthed from the Yanbei shiyuan tombs show striking simi-
larities with those in the five Tanyao caves (fig. 2.7). The Manes jewels 摩
尼寶珠 (plate 9) discovered in the Shaling tombs are also observed in the
 The sacred site of Yungang 41
second-phase caves (plate 10). The southern influence primarily came from
well-known painters in the south such as Dai Kui, Gu Kaizhi, Lu Tanwei, and
Zhang Sengyou.75 In the 5th and 6th centuries, Buddhist imagery underwent
stylistic transformations, which took place simultaneously with changes of
style in painting.76 The styles of the painters in the south had an impact on
Buddhist imagery in China during this time.
Finally, the extant individual images (some of which bear an explicit date)
made before Yungang display striking iconographic similarities with the early
images in Yungang. The image dedicated by a certain Wan Shen 菀申 and
dated to the fourth year of Taipingzhenjun 太平真君 (443)77 closely resem-
bles the standing Buddha on the south wall of cave 19. The Buddha’s robe
clings tightly, revealing the bodily contours of the figure. The end of the robe

Fig 2.7 Female figure, Yanbei tombs

42 The sacred site of Yungang
is layered with inner and outer robes. These two images are so similar that
they appear to come from the same prototype. Also, the Zhang Yong 張
永 (455) and Song Dexing 宋德興 (457) images78 show great similarities to
the seated Buddha in cave 20 at Yungang. Integrating different cultures and
artistic motifs, the Yungang artists made their own innovations when they
created the five Tanyao images of the first phase.
One more point to be made here is that Yungang artisans never wholeheart-
edly embraced or copied Western workmanship. They not only preserved the
Han tradition, but also rejected some “inappropriate” elements, using instead
images that were better suited to Chinese taste and culture, so when they came
to Yungang, their original style gradually disappeared. For instance, Yungang
artists put a light covering – the hem of a robe – over the right shoulder of the
Buddha image, while in Mathura, Gandhara, and the Western Regions the right
shoulder of the Buddha was often bare. Strictly speaking, there is no tanyou (bare
right shoulder) garment in Yungang. The term is used but the hem of garment
in fact provides a light covering for the right shoulder. The right shoulder of the
Buddha image in China is never bare, as is the case in Gandhara or Indian art.
Many literary sources suggest that the Indian orthodox Three Garments
三衣, upon arriving in China, were considered inappropriate for Chinese
culture even though the Chinese monastic sangha (Buddhist community)
adopted the tradition.79 Indian orthodox clothing was altered so as to make
it suitable for assimilation into Chinese culture.
The Indian Three Garments are often in the “open mode” with the right
shoulder of the Buddha uncovered. This does not fit Chinese ethical norms.
Therefore, a pianshan 偏衫 (patch) was added to the monastic dress so that
the physical body would not be revealed, because “that would be inappro-
priate (不以為善).” One of the Yuan sources, the Zengxiu jiaoyuan qinggui
(expanded rule for teachings monasteries) has the following notation:

People from the West in general have their arms uncovered. [Monks] were
afraid that criticism of this practice would arise, and so the arm needed to
be covered. . . . In the Northern Wei period, people from the Palace saw the
bared arm of the monks. They thought this was inappropriate. Then a right
sleeve was added, both sides of which were sewn. It was called pianshan.
It was open from the collar in the front, so the original appearance was
maintained. Therefore, it is known that the left part of the pianshan was
actually just the inner robe, while the right part is to cover the shoulder.80

Clearly, the Indian Three Garments did not look appealing to Chinese monks
and were considered “inappropriate.” Thus, the pianshan was added. The
Sifenlü shanbu suiji jiemo shuzhengyuanji (commentary on the revised
monastic procedures of the four-part Vinaya) also mentioned that there was
initially no such thing as pianshan in the West. People in the Northern Wei
likely felt that monks with a bare shoulder looked unseemly. They accord-
ingly added this kind of clothing (with pianshan) over the inner robe. It is
also called pianxiuyi.81 The Fozhi biqiu liuwutu 佛制比丘六物圖 (Diagram
 The sacred site of Yungang 43
of Six Buddhist Essentials) notes that the pianshan was added only to cover
the “ugly physical body.”82
The earliest extant evidence of pianshan is in the Northern Liang pagoda
images.83 By the 5th century, most images had pianshan, as may be seen in
several extant individual images such as the Zhang Yong, the Song Dexing
images, and the seated Buddha in caves 19 and 20 in Yungang.
It should be noted that the earliest images with baoyi bodai (loose robe
and wide girdle) in Yungang, the seated Buddhas in cave 11-14 (dated to
489), had no pianshan (plate 11); nor did the images clad in Chinese-style
garments, seated or standing, in cave 6 in the second phase. This further
proves that pianshan initially was added only to the open mode robes with
the right shoulder revealed to cover the “ugly physical body,” as suggested
in the sources above; however, we find that in Longmen the baoyi bodai
images have a pianshan on the right shoulder as may be observed in most
of the Northern Wei caves (Binyang central 宾阳中洞, Weizi 魏字洞, Putai
普泰洞 and Huangfugong 皇甫公 caves) and on the images discovered in
Wanfo Temple 万佛寺 in Chengdu. It seems that the Chinese did not know
what pianshan was, and were confused by it so more layers were added. As
is mentioned in the Sifenlü xingshi chao zichi ji (a commentary on the Dhar-
maguputaka Vinaya), “people at that time did not know that the inner robe
and fujian (pianshan) are two separate things. Therefore, they added another
one on top of it.”84 The pianshan, in essence, is a phenomenon brought about
by the process of sinicization in the 5th century Buddhist art, which funda-
mentally was influenced more by Chinese tradition than by foreign elements
in architecture and iconography.
As a political, religious, and cultural center, Pingcheng amassed quantities of
material resources and manpower in the form of hundreds of skilled migrant
craftsmen. This laid a foundation for the creation of Yungang, a process of
large-scale innovation in the tradition of Buddhist rock-cut cave temple art and
architecture in the Northern Wei capital. In imagery and architectural struc-
ture, early Yungang exemplifies a relatively archaic state of creative sensibility;
western influences still make themselves felt. During the second phase, Chinese
elements gradually begin to seep through. Thorough sinicization eventually
takes place in the last stage of the second phase, and reaches its peak during
the third phase.85 The art and architecture of Yungang reflect an intermixture
and fusion of multiple cultures and peoples, both local and foreign.
Buddhist architectural structures are only fossils of the culture, and their
meaning resides in how people experienced them. In a Buddhist sanctuary,
architectural space is a very important entity. The layout of a cave generally
determines the arrangement and composition of the images and the subject
matter; and this layout is often suggestive of function. Buddhist architec-
tural space is the basis for the study of typology in Buddhist archaeology.
In architecture, the first-phase cave temples are all elliptical in plan with a
half-domed roof and window. The cave’s focus is on the main colossal image,
which occupies almost the entire space. The subject matter at this time was
primarily centered on the three Buddhas and the thousand Buddhas, both of
44 The sacred site of Yungang
which are the main objects for visualizations. The main Buddha images in
this period appear majestic and solemn with a round face and broad chest
and shoulders. The robes of the seated Buddhas in large niches are usually
carved in raised folds with incised lines. Garments with step-like folds and
incised lines are similar to those represented in the early images of Dun-
huang, Maijishan, and Binglingsi caves. Among standing Buddhas, the two
Buddhas on the south wall of cave 19 are masterpieces, clad in robes that
closely cling to the supple bodies with low step-like folds (plate 12). The Bod-
hisattvas seated in the large niches of the five Tanyao caves are, for the most
part, later additions to the cave. Many were executed in the second phase.
They are elegantly carved with excellent proportions. The facial expression
is mild and dignified with a gentle smile. For instance, the seated Bodhisattva
with ankles crossed, along with the seated Buddhas, on the east reveal of the
window in cave 17 (dated to the 13th year of Taihe, 489), are fine examples
of the graceful images added to the first-phase caves (plate 13). The upper
body of the Bodhisattva is naked with one part of the celestial robe hanging
from the left shoulder and across to the right and another part fluttering to
the back and hanging from the arms. The skirt draped over the lower body
is carved in sharp incised lines. The attendant figures, usually carved in low
relief, appear elegant and dignified with plump face, and with long eyes and
brows. The image on the south wall of cave 17 represents one of the finest
attendant figures. The celestial robe hangs from the left shoulder to the right
waist covering the upper body, and another celestial robe flutters behind the
shoulders and swoops in graceful curves over the crooks of the arms. The
skirt clings to the supple body with an undulating line at the bottom. Overall,
the five Tanyao caves demonstrate a simple and vigorous style, and reflect a
strong western influence.
The most distinguishing feature of the second-phase caves is the rapid prog-
ress of sinicization. The cave structure in this phase differs considerably from
that of the first phase, with the exception of cave 13, which is occupied by a
large seated Bodhisattva with ankles crossed, a round face, and broad shoul-
ders. The caves are either rectangular or square in plan with flat, coffered, or
corbelled ceiling. The half-domed ceiling of the first-phase caves gradually
disappeared. Paired caves (1 and 2; 5 and 6; 7 and 8; and 9 and 10) emerged,
most of which are composed of an anteroom and main chamber (caves 5
and 6; 7 and 8; and 9 and 10; as well as cave 12). At the same time, roofed
niches in imitation of Chinese wooden structures began to appear in the caves,
as exemplified in the upper niches in the anteroom of caves 9 and 10. The
roofed niche usually consists of chiwei 鴟尾 (roof ridge ornaments), a bird in
the middle of the roof, rafters, brackets, triangular ornaments, and side pil-
lars dividing the niche into three parts: a central niche, which houses a seated
Bodhisattva with ankles crossed, and two side compartments. Beneath the
eaves with round rafters are brackets made of arms with three bearing blocks
and supports in the shape of an inverted V (plate 14). The east wall of the
antechamber in cave 9 (plate 17) corresponds with the west wall of cave 10
 The sacred site of Yungang 45

Fig 2.8 West and north walls, anteroom, cave 9

(plate 16), forming a paired cave with a passageway connecting the two caves
near the north wall of the anterooms. Influenced by the format of visual
representations of the Han stone reliefs, caves (e.g., 7 and 8) began to have
storied niches on each wall: the north wall contains two stories, whereas the
sidewalls contain four. Also, the narratives (Jataka Syama) are illustrated in
continuous panels like a Chinese traditional scroll painting, as may be seen
in the lower level of the anteroom in cave 9 (fig. 2.8). The second-phase caves
focused more on liturgical practices; thus, the cave structure was modified
accordingly. For instance, caitya caves (1, 2, 6, and 11) began to appear for
the purpose of pradaksina (circumambulation) practice. The emphasis on the
stūpa in temples and the construction of chaitya caves shares the same origin,
with the layout of the monasteries centered on the stūpa in Pingcheng. The
stūpa itself had come to be regarded as an outward and visual manifestation
of the Buddha. For a similar liturgical purpose, caves 5, 9, and 10 contain an
arched passage behind the main image for circumambulation.
The subject matter in the second phase is more diversified, and not lim-
ited to the three Buddhas and the thousand Buddhas as is the case in the
first phase. The twin Buddhas (Śākyamuni and Prabhutaratni), the Dharma
protectors, the narrative of the Buddha’s life and his previous incarnations,
as well as the discourse of Vimalakirti and Manjusri all begin to appear
in the second phase. The Maitreya Buddha had by this time become an
increasingly popular subject, and was the primary image in many caves.
The narratives are rich in content and form during this period. The Great
Departure, Banishing of the Demons, Conversion of the Three Kasyapas,
and the Offering of Food by the Merchants are favored subjects. The sto-
ries are depicted in various forms: in mono-scene, continuous scenes, or in
successive scenes as in a scroll painting. In mono-scene, the figures appear
only once and the composition depicts certain characters’ actions frozen in
46 The sacred site of Yungang
a single moment of the narrative. A continuous composition depicts a nar-
rative cycle in consecutive scenes, which are woven into one organic entity
with a clear sense of continuity over space and time.86 Temporal progression
is indicated by the recurrence of certain figures, sometimes in different set-
tings, sometimes against the same background. Through this compositional
device, the continuous pictorial flow of a narrative can be represented more
articulately. The successive scenes illustrate a narrative in progression, and
the story unfolds in temporal sequence, as when one reads a scroll painting.
The Syama story depicted on the lower level of the anteroom walls in cave 9
is the only story depicted in successive scenes to emphasize the significance
of Confucian ideology. In cave 6 alone, 36 scenes from the life story of the
Buddha (plate 15) are depicted. It is worth noting that the life story ends
with the first sermon of the Buddha; the Nirvana scene is not depicted. This
may be associated with the sutras that served as a basis for the series. The
Puyao jing (Lalitavistara),87 the Fo suoxing zan (Acts of the Buddha),88 and
the Guoqu xianzai yinguo jing (Sutras on Causes and Effects of the Past and
Present)89 are closest to the time of the construction of Yungang and were
most influential at the time. It therefore seems likely that these sutras may
have influenced the narrative; however, this does not necessarily exclude
the possibility of influences from other sutras. Judging from the depictions
and selection of subjects, it is almost certain that most of the depictions of
Buddha’s life story were based on the Guoqu xianzai yinguo jing. The visual
narratives, at any rate, can all be found in the sutra. But the artists who were
responsible for the depiction of the narratives in cave 6 most likely used
Lalitavistara, which ended the life story with the first sermon. The fact that
the Buddha’s life story in cave 6 ended immediately after his first sermon
testifies to the prominent influence of this text. It is impossible to pinpoint
which sutras directly inspired the Yungang artists when they used images to
narrate the tales. The same story sometimes appears in different texts, and
the same story can be depicted differently in literature. As Joanna Williams
aptly notes, the point of our textual comparisons is never to pinpoint the
literary basis of a particular theme but rather to build up from a statistical
vantage point an impression of general connections with various traditions
of Buddhist literature.90 The Buddha’s life is both history and legend, truth
and fiction. Some biographies are more historically based, whereas others
are full of miracles. It should also be noted that the jataka and nidana stories
(the celestial maidens dedicating the canopies, the two brothers who became
monks, the Nirgrantha-putras redeemed by Buddha in the fire Samadhi, the
offering of a meal by Sudatta’s wife, the defeat of Brahmana in a debate
by Nagarjuna, and Hariti losing her son) are represented only in caves 9
and 10, not in any others. The concentration of these narratives in caves 9 and
10 is believed to be associated with the Za baozang jing (Sutra of the Mis-
cellaneous Treasures), translated by Tanyao in Yungang.91 The discourse
between Vimalakirti and Manjusri also became increasingly popular. The
debate wonderfully set the stage for the Chinese elite to engage in qingtan
 The sacred site of Yungang 47
(pure conversation). The Śākyamuni Buddha and Many-Jeweled Buddha
gradually but steadily became one of the most important motifs. During the
third phase, the twin Buddhas are represented in almost every cave and often
are represented as the primary images of the cave on the north wall. They
became the visual symbols of the most popular sutra of the time, the Lotus
Sutra. Also, it is worth noting that the Dharma protectors became popular
from the second-phase caves onward. They are represented in every cave,
either on the ceiling or flanking the entrance. The ceilings of caves 13 and
11 are covered with intertwined dragons and many-armed protector deities.
Evidently, protecting the Dharma was a primary concern at that time.
During the second phase, stern and dull-looking images were replaced by
more human images with gently smiling faces. By the end of the second phase,
the images in cave 6 underwent fundamental stylistic changes. They were no
longer in the open mode with one shoulder covered as seen in the first phase.
Instead, all images were clad in traditional Chinese robes with both shoul-
ders covered. The standing Buddha on the west wall is clad in a long, thick
Chinese robe with the entire body covered (plate 18). The hem flares out like
a fishtail, accentuating the weightiness of the body. The V-shaped neckline
is low. The cordlike folds were replaced by flattened step-like folds. This is
a more naturalistic rendition. The tensions shown in the first-phase images
are not apparent in cave 6. Instead, the images appear relaxed and more
human, with a gentle smile and smoothly flowing robes. Beginning with
cave 6, the iconographic style underwent fundamental changes, and subse-
quently the images in the small caves between caves 11 and 13 continued
to develop stylistically. Also, the images became more slender and human
and the rendition became more naturalistic, as is apparent in the seated
Buddha in cave 11-10. This is the transitional period between the second and
third phase. In the third phase, the images became even more slender and
graceful, with long necks and sloping shoulders. The innovations in icono-
graphic style were continued, and new, sinicized images were created. The
process of sinicization was brought to full fruition during the third phase.
The third-phase caves were relatively moderate in size and either square or
rectangular with a flat ceiling. They are mostly fodian caves with only one
chaitya cave (39) and one colossal image cave (21). The iconographic com-
position and arrangement of the niches differ from the previous phases. A
new layout, sanbi sankan (three niches on three walls), emerged and came
to be predominant during this phase. The iconographic composition became
much more complicated. In the later caves of the third phase, five images may
appear on one wall, and disciples are represented near the Buddha, with two
Bodhisattvas further away. The subject matter became rich and diversified.
Some subjects appear only in third-phase caves. The Śākyamuni and Many-
Jeweled Buddhas were represented more and more, and became the main
subjects of the caves. All images are clad in baoyi bodai and appear slim and
elegant. They are typical xiugu qingxiang (graceful and refined) images.
The seated Buddha image on the left wall of cave 30 (plate 19) is a fine
48 The sacred site of Yungang
example of a third-phase image. The elegant seated Bodhisattva with ankles
crossed in cave 5-11 (plate 20) is one of the finest in the third phase. Here, the
sinicization process was fully realized.
The art and architecture of Yungang reflect the influence of multiple
entities with innovations from local artisans and migrants. The creation of
Yungang was a highly complicated process. The first-phase cave temples
functioned in practical terms as memorials to the Northern Wei court. Wor-
shiping a Buddha in the five Tanyao caves was much the same as demon-
strating respect and loyalty to the emperor. The second-phase cave temples
were designed principally for liturgical practices. The third-phase caves were
relatively small in scale and were mainly commissioned by local officials and
Buddhist devotees to pray for the imperial family and for their ancestors.
The archaeological findings in Yungang and Pingcheng, especially the
recent discoveries, not only contribute to our knowledge, but also allow us
to have a better understanding of Yungang, one of the most splendid stone
sculptural monasteries of all time. Previously, we had no concrete evidence to
prove where the monks lived and translated scriptures in Yungang and where
each particular architectural and artistic element is derived from. We had no
clear picture how secular tomb art and architecture, and Buddhist rock-cut
cave sanctuaries are related to each other. These new findings should not
only enrich our knowledge but also provoke new questions. It is hoped that
more monastery ruins in Pingcheng will be discovered to shed new light on
the components of the Yungang complex.

Notes
1 Benjamin Rowland, The Art and Architecture of India: Buddhist, Hindu, Jain
(London: Penguin Books, 1953), 69.
2 Li Yuqun, “Classification, Layout, and Iconography (Translated by John Kie-
schnick),” in Early Chinese Religion. Ed. Lagerwey John and Lü Pengzhi. Part
One (Leiden: Brill, 2010), 575.
3 Ibid., 575–8.
4 Da Song sengshi lüe, compiled by Zanning, T 2126:54: 237.
5 Ibid.
6 Wei Shou (506–572), comp., Weishu (Beijing: Zhonghua shuju, 1974), 114:3035.
7 Ibid.
8 Li Yuqun, “Classification, Layout, and Iconography” 582–3.
9 Mohe moye jing, Tanjing, trans. (Southern Qi dynasty, 479–502). T 12:383:
1013–14.
10 Weishu, 114:3030–2.
11 Hui Jiao, Gaoseng zhuan, T50:2059:352a.
12 Weishu, 114:3036.
13 Ibid.
14 Ibid., 114:3031.
15 Ibid., 114:3037.
16 Tang Yongtong. Hanwei liangjin nanbei chao fojiao shi (Beijing: Beijing daxue
chubanshe, 1997), 362.
17 Gao Yun, “Luyuan fu,” in Guang hongming ji. Comp. Daoxuan, T 52, no. 339.
18 Weishu, 114:3037.
 The sacred site of Yungang 49
19 Ding Mingyi, “Yungang shiku de kaizao lichen (The Excavation of Yungang),”
in Yungang shiku bainian lunwen xuanji, vol. 2. Ed. Yungang shiku wenwu yan-
jiusuo (Beijing: Wenwu chubanshe, 2005), 128.
20 Su Bai, “fenqi,” 78.
21 Liu Huida, “Beiwei shiku yu chan,” Kaoku xuebao, no. 3 (1978): 337–50.
22 Chan miyao fa jing. Kumarajiva, trans. T613:15:252. English translation by John
Kieschnick with minor revision in Early Chinese religion. Ed. Lagerwey John and
Lü Pengzhi, 580.
23 Zuochan sanmei jing, Kumarajiva, trans. T15:614:270.
24 Fu Fazang yingyuan zhuan, T2058:50:304.
25 Archaeological findings will be discussed later in the archaeological discovery
section of the chapter.
26 The sutra was translated by Kumārajīva. T15:617:299.
27 Angela Falco Howard and Giuseppe Vignato, Archaeological and Visual Sources
of Meditation in the Ancient Monasteries of Kuca, (Leiden: E. J. Brill, 2014), 97.
28 The Weishu records that the Luyeyuan cave temples were situated on the west-
ern mountain within the Northern Park (北苑), and there were rock cells and
meditation chambers (嚴房禪堂), in which meditation monks lived, see 114:3038.
Emperor Xianwen visited the cave complex on the fourth year of Huangxing
(470); see 114:128. The caves were discovered in 1980. Archeologists examined
the caves in 1987. There are 11 caves. Cave 6 contains images; the rest of them
are meditation cells, five on each side of cave 6. See Li Zhiguo and Liu Jian-
jun, “Beiwei Pingcheng Luyeyuan shiku diaochaji,” in Yungang Bainian lunwen
xuanji. Ed. Yungang Shiku Wenwu Yanjiusuo (Beijing: Wenwu chubanshe, 2005),
150–6. For the Tuyuk cave temples, see Li Yuqun, “Tulufan tuyugou shiku kaogu
xinfaxian: Shilun wushiji Gaochang fojiao tuxiang,” in Yishushi zhong de Han
Jin yu Tang, Song zhi bian. Ed. Shi Shouqian and Yan Juanying (Taibei: Shitou
Chuban gufen youxian gongsi, 2014), 95–124.
29 Si Maqian (c.145–186 BCE), Shiji, 180:2861.
30 Weishu, 114:3030.
31 Ibid.
32 Ibid.
33 Ibid., 181:2736.
34 Ibid.
35 Li Daoyuan, Shuijing zhu, T50:2060:427c27.
36 Daoxuan, Guang hongming ji, T52:2103:0103c27.
37 Yungang guji baoyangsuo, “Yungang xinfaxian de yikuai beiwei shike,” Wenwu
caikao ziliao, no. 9 (1957), 48.
38 Su Bai, “Jinbei Jiaozhu,” 63.
39 Su Bai, “Yungang moshi,” 120.
40 The first archaeological excavation was conducted by Japanese scholars, see
Mizuno and Nagahiro, Unkō sekkutsu, vols. 7 and 15 for archaeological reports.
The second was conducted by the Yungang Research Academy.
41 Lu Yifeng and Yuan Hairui, “Yungang shiku nisi kao,” in Yungang bainian lun-
wen xuanji. Ed. Yungang shiku wenwu yanjiusuo (Beijing: Wenwu chubanshe,
2005), 257.
42 Li Daoyuan, Shuijing zhu, T50:2060:427c27.
43 The Lubanyao cave complex contains three caves, all facing east, to the south of
the Wuzhou river, west of the Yungang caves across the river. Outside the caves
on the wall, there are, as in Yungang, holes for structure-supporting rafters. This
suggests that structures once stood in front of the caves. Above the caves lies flat
and vast terrain at different levels. Recent archaeological excavations show that
above the caves (first tier), there used to be residence quarters (second tier) and a
stūpa further above it to the east (third tier). This, to some extent, reconstitutes
50 The sacred site of Yungang
the original layout of the caves, the residence quarters above and the stūpa
further up as a cave complex entity. Recently, a huge Bodhisattva image was
unearthed in cave 2 in the Lubanyao complex. The image demonstrates a typi-
cal early Yungang style closely resembling the images of the five Tanyao caves.
This indicates the stylistic connection between the two cave complexes. The
thousand Buddhas in the Lubanyao caves were also created in an early Northern
Wei style. The donors and flying angels at the arched front of the entrance are
in the early Yungang style as well. The stone monastery (石祗洹舍) observed by
Li Daoyuan should be on the second tier above the caves since there are traces
of columns believed to be for the residential cells. Also, a large quantity of tiles
from the Liao dynasty was discovered. This suggests that in the Liao dynasty,
wooden eaves must have been added in front of the caves. It should be noted,
however, that the official archaeological report has yet to come out. Hence, this
is merely a simple analysis of the environmental situation near Yungang. The
author visited the site in 2013 and 2015, and was able to conduct field research
on the caves near Yungang.
44 See Lu Yifeng and Yuan Hairui, “Yungang shiku nisi kao,” 257.
45 Daoxuan, Xu gaoseng zhuan, T50:2060:425c26 and T50:2060:427c23.
46 In 1950, Yanbei wenwu kanchatuan conducted archaeological excavations and
discovered the remains of the Jiaoshansi cave complex, see Wang Xun, “Yungang
yidai kancha ji,” in Yanbei wenwu kanchatuan baogao. Ed. Yanbei wenwu kan-
chatuan. vol. 2 (Beijing: Zhongyang renmin zhengfu wenhuabu wenwuju, 1951),
28. The caves are divided into four areas.
47 Since the official report on the archeological excavations has not come out yet, the
following is a brief analysis of the excavations since the 1940s based on previous
investigations and my field research.
48 See Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vol. 7, 57–68, 123–9;
and vol. 15, 91–9, 185–90.
49 Jiang Huaiying, Yuan Hairui and Xie Tingfan, “Yungang shiku xinfaxian de
jichu jianzhu yizhi,” in Zhongguo shiku Yungang shiku. Ed. Yungang shiku
wenwu baoguansuo (Beijing: Wenwu chubanshe, 1991), 198–201. In 1938, Jap-
anese scholars Hibino Takeo and Ono Katsutoshi dug a deep trench but were
not able to tell from this what the structure and configuration of the buildings
in front of caves 9 and 10 was. Nor did they pay attention to the remains of a
large wudianding (hip roof). Cf. Mizuno Seiichi and Nagahiro Toshio, Unkō
sekkutsu, vol. 7.
50 See “Yungang kuqian yizhi fajue huo zhongda chengguo (The Great Result of the
Excavation of the Ruins in Front of the Caves),” Zhongguo wenwubao (1994):
1, 16. Also see Yungang shiku wenwu yanjiusuo, Shanxisheng kaogu yanjiusuo
and Datongshi bowuguan, “Yungang shiku disanku yizhi fajue jianbao,” Wenwu,
no. 6 (2004): 65–88.
51 Cao Chenming, “Yungang shikude kaogu diaocha, kaogu fajue jiqi yiyi,” in 2005
nian Yungang guoji xueshu yantaohui lunwenji yanjiujuan. Ed. Yungang shiku
yanjiuyuan (Beijing: Wenwu chubanshe, 2006), 153–6.
52 Zhang Qingjie, Li Baijun and Jiang Weiwei, “Shanxi Yungang shiku kuding beiwei
fosi yizhi,” in 2010 nian Zhongguo zhongyao kaogu faxian. Ed. Guojia wenwuju
(Beijing: Wenwu chubanshe, 2011), 127–30. Also see Zhang Qingjie, “Yungang
shiku kuding xiqu beiwei fojiao siyuan yizhi,” in Kaogu xuebao, no. 4 (2016),
533–62.
53 See fig. 29 General Map of the Excavation in front Wuhuadong caves in Mizuno
and Nagahiro, Unkō sekkutsu, vol. 7, 58.
54 The initial plan of the five Tanyao caves will be discussed in the next chapter.
55 Daoxuan, Xu Gaoseng zhuan records that Tanyao lived in the Tongle Temple in the
Heng’an caves. The temple was constructed in the Northern Wei. “Thirty kilometers
 The sacred site of Yungang 51
northwest of Heng’an, in the Wuzhou valley on the north cliff, mountains were
chiseled and monasteries constructed. This place is called Lingyan. The big cave is
more than 20 zhang deep and can hold three thousand people. The monastery in
the east can hold a thousand people. Tanyao gathered eminent monks in the Beitai
caves (Yungang) translated Fufazang yinyuan and the Jingtu jing. His work will be
carried down to later virtuous people, and the meaning will be kept without end.”
56 See Su Bai, “Jinbei Jiaozhu” for the record of the ten temples and the stone cham-
bers above the caves in the Jin stele inscription, 52–75.
57 The ten temples include Tongle 通樂, Lingyan 靈岩, Jingchong 鯨崇, Zhenguo
鎮國, Huguo 護國, Tiangong 天宮, Chongfu 崇福 (Chongjiao 崇教), Tongzi
童子, Huayan 華嚴, and Doushuai 兜率.
58 Xu gaoseng zhuan, T50:2060:427c24.
59 See Chen Yuan, “Yungang shikusi zhi yijing yu Liu Xiaobiao,” Chen Yuan, Chen
Yuan xueshu lunwen ji, vol. 1 (Beijing: Zhonghua shuju, 1980), 443–448.
60 Weishu, 114:3039.
61 See Datongshi bowuguan, “Datong beiwei fangshan siyuansi yizhi fajue baogao,”
Wenwu, 4 (2007): 4–26.
62 Weishu, 114:3036.
63 Ibid.
64 Ibid., 114:3037.
65 Ibid., 114:3032.
66 Ibid.
67 Ibid.
68 Walter M. Spink, Ajanta: History and Development, vol. 4 (Leiden: Brill, 2009), 34.
69 Weishu, “Xiyu zhuan,” 102:2259–60.
70 Li Yanshou, comp., Beishi, “Xiyu zhuan,” 53.
71 The ink stone is now housed at the Shanxi Provincial Museum. It was discovered
in the late 1960s in the Northern Wei remains in Zhouchengchang in Datong.
See Wenhuadageming chutu wenwu (The Cultural Relics Unearthed during the
Cultural Revolution). Ed. Chutuwenwu zhanlan gongzuozu (Beijing: Wenwu chu-
banshe, 1972), 81. In the archaeological findings, there were other dancers from other
central Asian countries. Among the migrants were merchants, musicians, and monks.
72 Su Bai, “fenqi,” 79.
73 The number of monks and nuns reached more than 2,000. There were 6,478
Buddhist monasteries in China. The number of monks and nuns reached 77,258.
Among them, Jianming Temple 建明寺 was constructed in the first year of Cheng-
ming 承明 (476), and the Siyuan Temple 思遠寺 on Fangshan 方山 in the third
year of Taihe (479), Weishu, 114:3039.
74 In the Simajinlong tombs, lacquer screen paintings and images were found. Some
gilded bronze images were also unearthed. See Shanxisheng Datongshi bowu-
guan (Datong Municipal Museum Shanxi), “Shanxi Datong Shijiazhai beiwei
Simajinlong mu (The Sima Jinlong tomb of the Northern Wei in Shijiazhai vil-
lage, Datong, Shanxi),” Wenwu, no. 3 (1972): 20–33; and Ma Yuji, “Datong-
shi Xiaozhancun Huagetatai beiweimu qingli jianbao (Inventory Report on the
Northern Wei Tomb Unearthed in Huagetatai Xiaozhan Village, Datong),”
Wenwu, no. 11 (1983): 1–4.
75 Su Bai, “Beichao zaoxing yishu zhong renwu xingxiang de bianhua,” in Su Bai,
Zhongguo shikusi yanjiu (Beijing: Wenwu chubanshe, 1996), 349–54.
76 The transformation in style, the issues of sinicization, and the relationships
between painters and sculpture from the 4th to 6th centuries will be discussed in
detail in Chapter Six in the investigation of the caves in the third phase.
77 See fig. 10 in Jin Shen, Zhongguo lidai jinian foxiang tudian (Beijing:
Wenwu chubanshe, 1994).
52 The sacred site of Yungang
78 See figs. 13 and 14 in Zhongguo lidai jinian foxiang tudian.
79 Three Garments, Pali, ticivara Sanskrit, tricivara and Chinese sanyi. They consisted
of Antarvasaka 安陀會, Uttarasanga 鬰多羅僧 and Samghati 僧伽梨, namely an
undercloth, inner robe, and outer robe. The outer robe is actually a thicker or heavier
version of the inner one. They are essentially the same. These three pieces of raglike
material are not supposed to be tailored or sewn but are only wrapped around the
physical body. Usually people used Kasaya (jiasha) 袈裟 to refer to the totality of
the habit. Chinese monastic garments adopted and followed Indian orthodox “Three
Garments” practice and were mainly inspired by both Gandhara and Mathura pro-
totypes. They can generally be divided into two types: fafu 法服 and changfu 常服.
Fafu is the formal monastic dress. It is also called sengyi 僧衣 (monastic dress), which
are composed of three pieces of oblong clothing: an undercloth, an inner robe and
an outer robe, similar to the orthodox Indian version of the “Three Clothes.” They
are called sanyi (Three Garments) in Chinese, which are essentially the same as the
Indian. They are merely worn in a different manner to suit Chinese taste and climate,
and the material is quite different from either Gandhara or Mathura.
80 Zengxiu jiaoyuan qinggui, Comp. Ziqing in the seventh year of Zhizheng 至正
(1347), X57, 968:32:17. The codes are divided into ten disciplines. It is an impor-
tant source for the study of the Tiantai monastic life of the Song and Yuan dynasties.
81 Sifenlü suiji jiemoshu zhengyuanji, annotated by Yunkan of the Song dynasty.
X40:0726:847:c24.
82 Fozhi biqiu liuwutu , Comp. Yuanzhao (?-1116) of the Song dynasty.
T45:1900:901:b17.
83 Yang Hong suggested that the earliest images with pianshan were the Zhang Yong
image dated to the first year of Taian (455) and Feng Aiai image dated to the first
year of Tian’an (466). See Yang Hong, “Shilun Nanbeichao qianqi foxiang fushi
de zhuyao bianhua (A Preliminary Study on the Evolution of Draperies of Buddhist
Images from the Southern and Northern Dynasties), Kaogu, no. 6 (1963): 330–7.
84 Sifenlü xingshi chao zichi ji, T40:1085:157a01.
85 This will be discussed fully in Chapter Six.
86 Junjie Huang and E. Zürcher, Time and Space in Chinese Culture (Leiden: E. J.
Brill, 1995), 244.
87 It is also called Fangdeng Benqi jing, eight juan. Translated by Dharmaraksa in
308 during the Western Jin period (265–317).
88 Also called Fobenxing jing, five juan. It was written by Asvaghosa, a Buddhist
scholar and gifted poet who lived in the 1st century CE and was translated by
Dharmaraksa (385–433) during the Northern Liang dynasty.
89 Translated by Gunabhadra (394–468) of the Liu Song period (420–479). Other
important Chinese translations of sutras, such as Zhongbenqi jing (Madhyama-
ityukta Sutra) translated by Tanguo and Kang Mengxiang of the Later Han dynasty
(25–220), Xiuxing benqi jing (Sutra on the Origin of the Practice of the Bodhisattva)
translated by Zhu Dali and Kang Mengxiang of the Later Han, the Taizi ruiying
benqi jing (Sutra on the Origin of the Lucky Fulfillment of the Crown Prince) trans-
lated by Zhi Qian of the Wu (222–280), Yichu pusa benqi jing (Abhinikramaa Sutra)
translated by Nie Daozhen of the Western Jin (265–317), Fobenxing jing (Bud-
dhapurvakarya Sutra) translated by Bao Yun of the early Liu Song dynasty, might
possibly have influenced visual narratives of Yungang as well. Sutras on Causes and
Effects of the Past and Present, Sutra on the Origin of the Practice of the Bodhisat-
tva and Sutra on the Origin of the Lucky Fulfillment of the Crown Prince are in fact
different versions of Chinese translations based on the same original text.
90 Williams Joanna, “Sarnath Gupta Steles of the Buddha’s Life,” Ars Orientalis,
no. 10 (1975): 171–92.
91 This will be discussed in detail in Chapter Four.
3 Phase one – emperor as
Tathāgata

The first-phase cave temples are symbols of the imperial rulers of the North-
ern Wei. They are the earliest colossal statue cave temples in Yungang. Inside
each of them is a giant Buddha image occupying most of the space. The
Weishu records that shamentong (the chief of śramana) Tanyao suggested
to the emperor that five cave temples be excavated to the west of the capital
soon after the restoration of Buddhism in 452.1 The edict of Buddhist revival
explicitly stated that Buddhism supports the prohibitions and regulations of
kingly government, and enriches the good nature and wisdom of human-
ity. It banishes all evil and brings about perfect Enlightenment.2 Evidently,
Buddhism was embraced to support the Northern Wei imperial court, and
the restoration of Buddhism was to maintain Northern Wei power. The pur-
pose was very clear. Soon after the revival of Buddhism, a stone Buddha
image was made to physically resemble the emperor. When it was finished,
it was seen to have black pebbles on both the face and soles of the feet,
which mysteriously resembled the moles on Emperor Wencheng’s body.3 The
motive underlying the image-making was also made clear in a new edict
issued in the first year of Xingguang (454), which ordered the officials to
cast five standing Śākyamuni images within the Grand Five-Storeyed Tem-
ple, representing the five emperors from the Emperor Taizu onwards. Each
image would be 6.8 feet in height and all together 25,000 pounds of copper
would be used.4 It can be seen that the emperor was portrayed as the present
incarnation of Tathāgata.5 In light of the traditional practice of making cave
temples for the late emperors in the Northern Wei, it has been suggested that
the five cave temples in Yungang were excavated for the same reasons and
for the same five founding emperors.6 “The five great caves carved out on
the rock of cliff at Yungang must have also been undertaken from the same
filial motive.”7 This theory was later widely accepted in academia, and the
five imperial great caves are now believed to be caves 16 to 20, tradition-
ally called the five Tanyao cave temples. It is clear that the excavation of the
five cave temples is intimately associated with the imperial family and their
welfare. The primary focus of the imperial five was upon kingship, the praise
of kingship, and worship of kingship. It was to “commemorate those who
were able to bestow compassion on the people and benefit the beings”8 of
54 Phase one – emperor as Tathāgata

Fig 3.1 East wall, cave 16

the founding emperors of the Northern Wei and promote the Buddhist ideal
of the emperor being the living Tathāgata of the time that caves 16 to 20
were excavated and the five colossal Buddha statues were carved out of liv-
ing rock. The court also desired to expiate previous wrongdoings. “Emperor
Wencheng wished to turn the inauspicious event of the Buddhist persecution
carried out by his predecessor into spiritual welfare.”9 It can be said that
the five Tanyao caves are essentially the imperial family shrine temples for
worshipping and praying.
What, then, is the relationship between the five emperors and the five cave
temples? Which of the five emperors is enshrined in the caves? When exactly
were they started, and when finished? What is the chronological sequence of
the caves? The positioning of the five caves now seems to be in disorder. What
was the initial grand plan and design of the caves? A yet more interesting ques-
tion is the meaning of a row of lower legs of worshippers depicted in cave 16
(fig. 3.1). The upper body of the worshippers seems to be cut and was left like
that purposefully. These are basic questions, but there is no consensus on them.
They are still worth careful investigation, especially in light of the ever-increasing
archaeological findings today within and around the Yungang complex. Before
we fully treat these issues, let us first examine the five cave temples and their
related problems briefly in order to have a fresh point of view.

The imperial five Tanyao caves

All the cave temples are elliptical in plan like a horseshoe with a colossal
Buddha image in the center of the main wall (fig. 3.2). The giant Buddha
image forms the focus of each cave. The east, west, and north walls lean for-
ward and become narrower as they ascend upward towards the ceiling. The
horseshoe shape of the caves with a window is usually believed to have its
origin in Indian thatch-covered freestanding wooden buildings.10 The recent
archaeological findings near Pingcheng, however, point to the half-dome
shaped tents used by the nomads including the Xianbei people. The clay
 Phase one – emperor as Tathāgata 55

Fig 3.2 Diagram of the five Tanyao caves

Fig 3.3 Clay tent with windows, Yanbei shiyuan tombs

models of the tents unearthed in the Yanbei shiyuan 雁北師院 tombs (fig. 3.3)
show striking similarities with the structure of the five cave temples.11 The
use of tents by the nomads has a long history and is mentioned in many his-
torical sources in the Northern dynasties; even the Hou Hanshun (Standard
History of the Later Han) noted that Wuhuan and Xianbei people did not
live in permanent dwellings, but instead used tents as their residence; all face
east.12 In addition, not only do we find clay models of tents in the tombs,
but we see tents depicted in the scenes of everyday life in the mural paint-
ings (plate 21). Evidently, the clay models of the tents, together with other
unearthed objects of everyday use, are the models of the residences that local
people were still using at the time. The artisans who created and designed
56 Phase one – emperor as Tathāgata
the five cave temples absorbed ideas from the tombs into a religious space.
In architecture, the five Tanyao cave temples were more influenced by local
mortuary and Xianbei culture than that of India and Central Asia. The famed
architect Liang Sicheng surmised, long before the tent models were unearthed,
that the imperial five might be the unique creation of Tanyao, the monks,
and local artisans of Pingcheng.13 Su Bai also pointed out that not a single
precedent for this has been found anywhere, including the South, Central
Asia, Xinjiang, and the Gansu area.14 It can be seen that the excavation of the
imperial five in Yungang was a process of making innovations and carrying
on traditions, and drawing inspirations from the regional culture of the time.
The five cave temples can be, based on subject and structure, further divided
into two groups: caves 20, 19, and 18 to the left belong to one group; and
caves 17 and 16 on the right are another group.15 Outside cave 16, the cliff
surface turns inward at the east end (fig. 3.4), and then starts a new cliff surface
receding deeper into the mountain. It is clear that the outside wall of caves
16–20 ends at the east side of cave 16, and when the outside wall was chiseled,
caves 16 to 20 were made as one grant group in phase one. The caves then
were further divided into two groups for excavation. These two groups were
not even on the same ground level. The ground level of caves 16 and 17 is
lower than that of caves 18−20, and the main Buddha image in caves 16 and
17 is different from that in the first group caves. In cave 17, the main statue is a
Maitreya Bodhisattva with ankles crossed, not a Buddha, and cave 16 contains
only one standing Buddha, with no attendant Buddhas. These are clearly two
different groups, but all are part of the pre-conceived plan of the first phase.

Fig 3.4 External wall of caves 16–20

 Phase one – emperor as Tathāgata 57
Cave 19, the largest of the five, is unique. This is the only cave that has
two side chambers (plate 22), facing slightly inward toward each other; inside
each an attendant Buddha seated with legs pendant is represented. The main
chamber contains a colossal seated Buddha (56 feet high, the tallest of the
five Buddhas) occupying most of the space in the cave (plate 23). The rest
of the walls of the cave are mostly covered with small thousand Buddhas,
which are rendered in different modes, some with both shoulders covered,
others with right shoulder slightly bare; only a few images are not clad in any
robes. The thousand Buddhas are carved with crude workmanship. On the
upper level of the south wall, beside the window, two standing Buddhas are
represented. The west Buddha is intended to represent the Buddha touching
his son Rahula’s head who is kneeling down on the side of the Buddha.16 The
two standing Buddhas in the south wall are the “most splendid sculptures in
the whole of Yungang caves.”17 The thin robe clings to the body of the Buddha
closely, creating an exquisite sensation (see plate 12). It should be noted that
the window of the cave is larger than the entrance gateway below and is
placed at the same height as the main Buddha’s face. This allows sufficient
light to illuminate the image and the cave. The architects were well aware of
the significance of making use of lighting. The reveals of window are covered
with small niches made in later times. The ceiling of the cave is not carved or
decorated. The concentration was on the main Buddha, representing one of
the emperors. The main chamber in the center recedes further back into the
cliff than other caves. This should be the original cliff structure since several
other caves (caves 30 and 31, etc.) are also set back deeper into the mountain.
The cave should have been excavated based on the original cliff structure. As
Li Daoyuan once mentioned, the cave temples of Yungang were excavated
according to the natural structure of the cliff (yinyan jiegou 因岩結構).18
Through investigation and observation of the outside wall of the caves, it
appears that some caves were excavated based on the natural structure of
the cliff surface, but others were man-made and carved to the current surface
according to need. The overall space of the cave provides a sense of majestic
solemnity and dignity. The key question concerning this cave is the identity
of the founding emperor enshrined in this largest cave as the living Tathāgata.
This is pivotal in figuring out the relationship between the five founding
emperors and the five cave temples since cave 19 is a unique cave.
Cave 20 has become the symbol of the Yungang complex. The seated
Buddha (46 feet high) in dhyāna position is full of power and grandeur with
round face and broad chest (plate 24). Behind it, the rest of the main wall is
covered with a large flaming nimbus over the head of the Buddha. The flame
is beautifully designed like a traditional Chinese cloud pattern. The robe of
the Buddha is in the so-called tanyou mode with the hem of the robe like a
pianshan (patch) covering the right shoulder. The inner robe top border is
decorated with a band of floral-scroll pattern, which has a half-palmette of
three leaves joined symmetrically with a small band. In front of the cave,
many Northern Wei tiles were unearthed in the archaeological excavations
58 Phase one – emperor as Tathāgata
in the 20th century. This suggests that there was Northern Wei architectural
remains in front of cave 20. Nothing hitherto has been found in front of
caves 16 to 19. That means that there were no buildings in front of caves 16
to 19 in the Northern Wei. Cave 20 is an exception. It is also the only cave in
which the south wall and part of the ceiling collapsed. As mentioned previ-
ously, the Northern Wei building in front of the cave was in all probability
related to this. In order to protect the main Buddha by shielding it from
exposure to the outside, a building had to be constructed in front of cave 20.
Cave 18 is the most majestic and dignified of all the caves. The three stand-
ing Buddhas and Ten Great disciples are depicted with refined craftsmanship.
The massive main Buddha (52 feet high) in the center occupies nearly all
of the interior space (plate 25). The great Buddha has a round face, broad
chest, and wide shoulders. The robe, which clings to the body closely, is in
the so-called tanyou mode like that of cave 20 with the right shoulder slightly
covered by a pianshan. The upper body is covered with small dhyāna Bud-
dhas and aupapadukas (lotus-born Buddhas) in relief. The right hand hangs
downward and the left hand is held up high to the breast. The window of the
cave is extremely big, even much bigger than the entrance gateway. Through
the window most of the main Buddha can be seen. Above the attendant
Bodhisattvas, ten great disciples, five each side, are represented. These ten
figures have vivid facial expressions with realistic and gentle smiles. These
are among the finest examples of Yungang workmanship. As with other
caves, there used to be small niches for thousand Buddhas outside, most of
which are eroded now, so that only traces can be seen.
Cave 17 uniquely contains a seated Bodhisattva with ankles crossed and
two attendant Buddhas. The space seems rather restricted with the colossal
Bodhisattva, whose extended knees and elbows occupy almost the entire
space of the cave (plate 26). The rest of the wall space is covered with niches
in various sizes that were added later at different times. The north wall
has a curved area indicating the cave was excavated in different horizontal
levels. This is the only cave that contains a seated Bodhisattva with ankles
crossed, rather than a Buddha. The important question then is which of the
five emperors is represented here as the future Buddha, rather than the liv-
ing Tathāgata. The reveals of the window are covered by the later niches,
one of which on the east reveal bears an inscription dated to the 13th year
of Taihe (489).
Cave 16 is different from the other four. The space in the cave is not strictly
controlled. There is more room, with only a Buddha standing in the center,
without attendant Buddhas next to it. The rest of the space is primarily cov-
ered with thousand Buddhas and a few large niches amidst them and on the
reveals of the window. The images inside the niches are carved in the incised
lines employed in the Han stone slabs and in full volume, resembling those
in caves 7 and 8 closely in style. They are fine examples of Yungang sculpture of
seated Buddha images. The gentle smile, the calm facial expression, and elon-
gated face with wavy hair make the standing Buddha look more realistic and
 Phase one – emperor as Tathāgata 59
personal like a young nobleman of the Northern Wei (plate 27). The sculpture
is endowed with a distinct sense of realism and naturalism. The Buddha image
appears more like the living Tathāgata than any other main Buddhas in the
imperial five. The robe of the standing main Buddha covering both shoulders
is rendered in highly formalized step-like pleats, and is represented in the
baoyi bodai Chinese garment style of cave 6.

Dates of excavation of the five Tanyao caves

Due to lack of explicit literary records and hard evidence, the dates of the
five caves have been a topic of debate among Yungang scholars. There have
been three suggestions about the inception of Yungang: (1) the first year of
Heping 和平 (460), which is the most widely accepted, (2) the second year of
Xing’an 興安 (453), and (3) the first year of Shenrui 神瑞 (414). Each theory,
based on different sources, looks at the beginning of the five cave temples
from a different perspective. Let us first look at the last point of view (414)
regarding the inception of Yungang, which was primarily based on the Da
Tang neidian lu. It reads:

Emperor Daowu (r. 386–409), Taizu of the Northern Wei changed the
name of his reign era to the first year of Shenrui (414–16). This was the
first year of Taiyuan of the Jin (376) in the reign of Emperor Xiaowu
(r. 372–396). . . . In the western outskirts of Heng’an (Pingcheng in the
Tang), in the deep valley on the stone cliff, caves were hewn throughout.
They are more than ten zhang (more than 100 Chinese feet) high and
spread over an area of 30 li (9 miles). They are numerous, and aligned
closely, like the teeth of a comb. To the east end of the valley, a stele can
still be seen. It states that the accomplishment [of constructing the cave
temples] was too great for calculation.19

Based on this record, many literary sources in later periods including the Jin
stele inscription suggest that the construction of Yungang started in 414. The
theory was later widely accepted in Ming and Qing times, and was quoted
in scholarship in the early 20th century.20 In recent years, some scholars
revisited the issue and also suggested that the beginning of Yungang should
predate the conventionally accepted date (460) and should be in 414.21
However, if we carefully examine the source, a problem appears immedi-
ately. First, during the reign of Emperor Daowu (r. 386–409), there was no
such reign title as “shenrui”; this was actually the reign title of Emperor
Mingyuan. Emperor Daowu only had two reign names (Tianxing 天興 and
Tianci 天賜) after he transferred the capital to Pingcheng in 398. Shenrui
was not used as his reign title. Secondly, Northern Wei Buddhism directly
followed that of the Northern Liang, and did not flourish until the large
migration of the Northern Liang people to Pingcheng in 439. The Weishu
records that in the fifth year of Taiyan 太延 (439), Liangzhou was pacified
60 Phase one – emperor as Tathāgata
by the Northern Wei, and some 30,000 people were forced to migrate to
the capital.22 The śramanas and Buddhist practices both went along with
them to the east, and from then on Buddhist images and doctrines grew
rapidly.23 Buddhism in Pingcheng entered a new era from that time on. If
Buddhism did not flourish before 439, it is hard to imagine the flourishing of
Buddhist rock-cut cave art on such a grand scale. Thirdly, during the reign
of Emperor Daowu, “the empire had just settled down. War chariots were
still moving constantly. All kinds of state affairs had just started; therefore
buildings of Buddhist reliquaries were yet to be built and Buddhist clergy
were not yet invited.”24 It was too soon to consider huge constructions like
the Yungang complex. Fourthly, most crucially, we have thus far found no
trace of the impact of Buddhist persecution on Yungang. As is widely known,
when the suppression of Buddhism occurred in 446, “buildings and stupas,
and all places where the edict reached were completely destroyed.”25 The
Weishu recorded many accounts of the destruction and construction of Bud-
dhist temples and images in Pingcheng during the persecution and revival of
Buddhism, but not one single source mentioned any damage, destruction,
or even rebuilding of the Yungang caves. This suggests that the Yungang
caves had not been constructed when the suppression of Buddhism hap-
pened. Therefore, it is hardly possible that Yungang was started as early as
the Shenrui era. At the beginning of the 20th century, the scholar Chen Yuan
had already questioned the possibility that the beginning of Yungang could
be as early as 414 and argued that it was impossible that the construction
could have begun then.26 After all, it was too short a time from the transfer
of the capital to Pingcheng (398) to the first year of Shenrui (414), 16 years
only, to build a new capital, political seat, and a monumental cave temple
complex. Moreover, Buddhism in Pingcheng was the Buddhism of migration,
and Buddhist art was the art of migrants. Through migration, Buddhism was
made to permeate the region, and Buddhist art was introduced.
Let us now turn our attention to the Weishu passage based upon which
both Su Bai and Yan Wenru drew their conclusions about the inception of
Yungang (460 and 453 respectively). The Weishu has the following account:

1). In the early Heping Era (460–65), Shixian passed away. Tanyao
replaced him and his title was changed to the chief administrator of
śramanas. 2). Earlier, in the second year of the restoration of the Dharma
of the Buddha (453), Tanyao was ordered to leave Zhongshan for the
capital (Pingcheng). The emperor happened to be out [when Taoyao
arrived] and met with him on the road. The Imperial Horse moved
forward to nibble Tan’s robe. At the time, it was believed that even the
horse recognized a benevolent man. 3). The emperor later honored him
by observing the rites that a student would extend to a teacher. Tanyao
petitioned the emperor to hew the mountain cliff of the Wuzhou Fortress
in the west of the capital and excavate five cave temples, inside each of
which would be carved one Buddha image. The tallest one would be
 Phase one – emperor as Tathāgata 61
seventy chi and next one sixty chi. Carvings and embellishments would
be wondrous and magnificent. It would be the best of the time . . . 4).
Tanyao and the Indian śramana Changnayeshe 常那邪捨 and others
translated and published fourteen new scriptures.27

The passage tells us four things: (1) Tanyao replaced Shixian in 460; (2)
before this, in the second year of the restoration of Buddhism in 453, Tan-
yao was ordered to go to the capital to meet with the emperor; (3) the
emperor later honored him with respect as a teacher, and Tanyao requested
the excavation of five cave temples along the mountain cliff surface; and
(4) he translated sacred sutras with others.
Based on the passage above, Su Bai suggested that the five Tanyao cave
temples were begun in 460 after Tanyao replaced Shixian as the chief admin-
istrator of śramanas.28 Mizuno Seiichi accepted the theory and also suggested
that point two in the passage above was a flashback inserted into the original
narrative.29 Yan Wenru, however, disagreed with Su, and argued that Su Bai
was mistaken in supposing that 460, the year of Tanyao’s replacement of
Shixian, was also the year that Tanyao’s five caves began to be constructed,
and that these were two different events that happened at different times.
Yan suggested that the five caves were begun in the second year of the resto-
ration of Buddhism (453) soon after Tanyao earnestly requested the emperor
to excavate the caves with colossal images. Thus, the five caves were started
in 453; not 460 as proposed by Su Bai.30 Yoshimura Rei, in turn, disagreed
with Yan and argued that the beginning of the Tanyao’s five cave temples
should be 460. The crux of the argument lies in point three, Tanyao’s peti-
tion to the emperor to excavate the five caves. Did this occur right after
Tanyao was summoned to the capital in 453? Or after Tanyao became the
shamentong in 460?
Turning back to the Weishu passage, it seems, from the structure of the
passage, that point two was indeed a flashback, and had nothing to do with
points one, three, or four. It was a recollection of what had happened in 453
when Tanyao was summoned to the capital, but the account of the passage
is mainly focused on what happened after Tanyao became the shamentong
in 460. Point two had nothing to do with the petition or with the excava-
tion of the five cave temples. Tanyao petitioned the emperor to excavate the
rock-cut caves outside the capital only after he became the shamentong in
460. As Rei pointed out “it should have been one of the key events of elevat-
ing Buddha Dharma after Tanyao became the shamentong in 460. He then
requested the excavation of the five caves and was granted permission.”31 It
is more likely that the petition was made when Tanyao was in charge after
Shixian, who was the Daorentong 道人統 (religious superintendent) and had
been in charge before 460, passed away.32 It would be hard to explain why
Tanyao made his petition in 453, when Shixian was still in charge as the
Daorentong. Moreover, if it had happened in 453, a notation concerning the
construction of the five caves would have been included in the earlier record
62 Phase one – emperor as Tathāgata
of the Weishu, which lists the important events after the revival of Buddhism,
such as the restoration itself (452), the imperial edict commanding officials
to have a stone image of the emperor made (also 452), and the edict order-
ing the casting of five images from Taizu onwards (454). The records are
arranged chronologically. The excavation of the five Tanyao cave temples,
instead, is recorded in the section after Tanyao became the shamentong in
460. Hence, it appears that the five cave temples were begun in 460.
Having this date in mind, let us take a look at literary records on Tanyao
and new archaeological evidence that may shed fresh light on the beginning
of Yungang complex as a whole, not just on the inception of the imperial five.
The Xu gaoseng zhuan notes that in the first year of Heping (460), Tanyao
was appointed as the chief administrator and that he lived at the Tongle
Temple within the Heng’an caves (the name of Yungang in the Tang).33 “In
the third year of Heping (463), Tanyao had already assembled many virtu-
ous monks in the Beitai cave temple (Yungang) and translated sutras.” This
is also mentioned in the Weishu passage, which says that Tanyao translated
14 sutras with others.34
We must admit that if he lived in the temple in Yungang in 460, the
Tongle Temple must have had been completed before that. Also, if we hold
that the five Tanyao cave temples were started in 460, then we accept at
the same time that the Tongle Temple within the Yungang complex must
have been built before 460, so Tanyao was able to live there at that time.
The translation of the sutras could not have taken place inside the caves. It
is not possible that the excavation of the caves and the translation of the
sutras took place at the same time in the caves. None of the caves were as yet
completed in 463. This forces us to look elsewhere for the place where the
translations were carried out and for Tanyao’s residence. All the findings in
the 20th century prove that there were no archaeological remains in front of
the wuhuadong (caves 9 to 13) and caves 16 to 19 (cave 20 is an exception)
during the Northern Wei period. Besides, the initial Wuzhou river waterway
was too close to the cave at that time for additional buildings. The only suit-
able location within the Yungang complex would have been above the rock-
cut caves. This assumption is supported by recent discoveries in 2010 and
2011. Archaeological remains of Buddhist monasteries of the Northern Wei
in Yungang consisting of residential halls and pagodas were found on top of
the caves on Mount Wuzhou, as we discussed previously. In 2010, another
Buddhist monastery of the Northern Wei was discovered in the area above
caves 39–45. These discoveries prove that the Yungang complex consists of
rock-cut cave temples and freestanding monasteries, and further explain why
there are no vahara caves. The residential quarters were above the rock-cut
cave temples. Su Bai once speculated that “several stone rooms above the
caves 上方一位石室數間,” mentioned in the Jin stele, might have been the
monasteries discovered by Mizuno and Nagahiro in the ’40s in the east and
west sections above the caves.35 Now the archaeological evidence corrobo-
rates literary sources. Evidently, all point to our inference that the place for
 Phase one – emperor as Tathāgata 63
translation and Tanyao’s residence were located above the rock-cut cave
temples. This also explains why Tanyao was able to live in a temple in the
Yungang complex already in 460, and in 463 was able to assemble monks
to translate sutras. It also proves that it is not impossible that the inception
of the five Tanyao caves was in 460, as Tanyao lived in the monastery that
we now know was above the caves. This does not contradict our inference,
but rather supports it.
The completion dates of the imperial five are another topic of debate
among scholars. Su Bai suggested that they might have been completed with
the death of Emperor Wencheng in 465, but the main Buddha images in
caves 16 and 19-2 were postponed to the second phase, and could have
been as late as 494, right before the capital was transferred southward to
Luoyang.36 This means that most of the five caves were constructed under
Emperor Wencheng. Nagahiro and Mizuno initially dated the completion of
the imperial five to 475.37 Then they changed their position and proposed
that the five caves were finished in the first year of Huangxing 皇興 (467),
and Emperor Xianwen’s visit to Yungang was to celebrate the completion of
the caves.38 Rei, however, suggested that it took about ten years to complete
the imperial five.39 He further argued that the main Buddha image in cave 16
was also completed within this time frame, which is substantially different
from what Su Bai and Nagahiro had suggested.40 If the five caves were com-
pleted in 470, as suggested by Yoshimura Rei, this means that they were only
half completed when Emperor Wencheng passed away in 465. Also, if the
main Buddha image in cave 16 was completed together with the five caves,
it implies that the images in cave 6 were completed much earlier than they
are conventionally believed to have been since, stylistically, they resemble
one another closely. But this is difficult to establish. The consensus is that
cave 6 was completed near the time of the transfer of the capital in 494.41
In addition, the baoyi bodai Chinese garment style with flattened step-like
pattern began to emerge only in the second phase in the 480s. The earliest
evidence we have for the step-like rendition of Chinese robes in Yungang is
dated to 489 in cave 11-14 on the outside wall of cave 11 (see plate 11). This
new style is rather different from the earlier cordlike folds represented in
caves 18–20. The early ones resemble more closely the Tai’an 太安 images
(dated to 455 and 45742). The step-like pattern is not rendered in any of the
early images from either Taipingzhenjun era 太平真君 (440–51) or Tai’an
era (455–59). We have no surviving objects to support the hypothesis that
the Chinese garment style seen in cave 16 could be as early as the period of
the five Tanyao caves. We therefore cannot accept the inference that the main
images in cave 16 and 19-2 were finished with the primary construction of the
five caves. “The main image of cave 16 is carved in a less deep relief which is
far away from the early style of Yungang. To judge by the fact alone, it would
be dangerous to identify cave 16 as one of the five Tanyao caves.”43 I believe
that the architectural structure and the colossal images in caves 17–20 were
well underway and near completion when Emperor Wencheng died in 465.
64 Phase one – emperor as Tathāgata
But the rest of the walls and windows of the caves were continued through
the eras of Emperor Xianwen and even to Emperor Xiaowen. Most of the
five cave temples, strictly speaking, were never really finished; the ceiling,
for example, is crude and undecorated, the floor is uneven, and small niches
were continuously added during different phases. The completion of the
caves here refers only to the completion of the main images in each cave and
the main architectural structure. In cave 17, one niche on the east reveal of the
window is dated to 489 (see plate 13). That means cave 17 was still under
the construction as late as at least 489.

Emperor as the living Tathāgata in the cave temple

It is still debatable who each of the five colossal images specifically repre-
sents. What is the relationship between the five emperors and the five cave
temples? An attempt to explore these questions will help us to understand
the initial grand plan of the five cave temples. The current asymmetrical
arrangement of the imperial five appears to lack the grand master plan of a
royal project.
Sekinō Tadashi and Tokiwa Daijō suggested that the five founding emper-
ors represented in the five Tanyao caves should be Emperors Daowu (Tuoba
Gui, r. 386–409), Mingyuan (Tuoba Si), Taiwu (Tuoba Tao, r. 424–52),
Jingmu (the Crown Prince, Tuoba Huang), and Emperor Pingwen (平文).44
They did not, according to them, include the reigning Emperor Wencheng.
Tsukamoto Zenryu, however, argued that the five emperors should be the
four late emperors (Daowu, Mingyuan, Taiwu, and Jingmu) plus the current
emperor, Wencheng (Tuoba Jun).45 He further noted that only in this way
(with the ruling emperor included), could the characteristics of Buddhism
in the Northern Wei be reflected.46 Ōmura Seigai made the same inference:
that the five emperors represented in the imperial five caves included the four
previous emperors and Emperor Wencheng.47 This theory was later widely
accepted. The imperial court strictly controlled the development of Bud-
dhism in the Northern Wei, and image-making and cave temple excavations
were closely associated with politics and power. As a passage in Gaoseng
zhuan observes, “without depending on the ruler of the state, it is difficult
to establish the Buddha Dharma.”48
Alexander Soper suggested cave 20 was built for Emperor Jingmu, cave
19 for Taiwu, cave 18 for Mingyuan, cave 17 for Daowu, and cave 16 for
Pingwen.49 Rei, however, associates the five caves with five different rulers.
He linked cave 16 to Emperor Wencheng, cave 17 to Jingmu, cave 18 to
Taiwu, cave 19 to Daowu, and cave 20 to Mingyuan; Su Bai did not assign
these images to any specific ruler. Instead, he simply suggested that these five
images represent the five emperors from Emperor Daowu onwards (Daowu,
Mingyuan, Taiwu, Jingmu, and Wencheng), and that each main image in
caves 16, 18, 19, and 20 is the historical Buddha Śākyamuni, while the
seated figure with ankles crossed in cave 17 is the Maitreya Bodhisattva,
 Phase one – emperor as Tathāgata 65
the future Buddha.50 It is appropriate to associate cave 17 with Emperor
Jingmu, the Crown Prince, who never became the emperor, implying he
would be the future incarnation of the Buddha and emperor. It is also appro-
priate to connect cave 16 with the ruling emperor of the five, emperor
Wencheng. The standing Buddha in this cave does not have the solemnity
and heaviness of the other Buddhas in other caves. Rather, it looks more like
a “noble prince of the Northern Wei.”51 This is the only Buddha sculpted
with a great sense of realism and naturalism. The gentle smile and gaze of
the eyes personalize the Buddha. If we believe that it is reasonable to con-
nect caves 16 and 17 to emperors Wencheng and Jingmu, which cave then
is more appropriate for emperor Daowu, the first emperor of the five? We
have caves 18, 19, and 20 left.
Cave 19, a unique cave, is the largest of the five, and it is the only cave
that has two side caves. Earlier, it was mentioned that the five Tanyao caves
can be divided into two groups: caves 18−20 belong to one group and caves
16 and 17 to the other, and caves 18−20 were constructed first. It is safe to
say that caves 16 and 17 were no earlier than caves 18−20, and one factor
that causes cave 17 to have a later appearance is that the main image is a
Bodhisattva and not a Buddha.52 The fact that the interior floor is deeper
than the exterior may indicate a later beginning, and a delay in the work.53 If
caves 18−20 were constructed as a group, it would have been logical to start
with cave 19 in the center, and to begin with Emperor Daowu, the founding
emperor of the Northern Wei. Emperor Daowu, who ascended the throne
as the King of Dai in 386, held the Law of Buddha in great reverence and
had the Five-storied Reliquary, Halls of Vulture Peak, and Sumeru built.54 In
addition, he also had the lecture hall, meditation hall, and cells for śrama-
nas built.55 It can be seen that Buddhism had well developed and Emperor
Daowu supported the religion before the excavation of the cave temples in
Yungang. Following the actions of Emperor Daowu, Emperor Mingyuan
also supported Buddhism. Under his reign Buddhist images and statues were
erected in all corners of the capital.56 He commanded the śramanas to guide
the laity.57 It is widely accepted that the five Tanyao cave temples were exca-
vated for the five emperors from Emperor Daowu onwards, and the pur-
pose was to eulogize the forefathers of Emperor Wencheng. They fulfilled
the same role as the family shrine temples of a state founder. Therefore, it
would be more reasonable that the five cave temples, in each of which an
emperor revered as the living Tathāgata is enshrined, would be positioned in
an ingenious arrangement following the Zhao Mu system 昭穆制, an ancient
procedure for ordering ancestral tablets named after two kings of the West-
ern Zhou (1046–771 BCE). The founding Emperor Taizu (Daowu) would
be in the center and the succeeding emperors would be situated alternately
to the left and right sides of the founder, who would retain its prime place in
the center. Michael Loewe noted that the system served to affirm genealogi-
cal origins and continuity.58 It implicitly asserted the legitimacy of royal and
imperial rule, and it provided for the perpetual continuation of sacrificial
66 Phase one – emperor as Tathāgata
rites offered to the founder of the house.59 As Hang Kan aptly pointed out,
the Northern Wei court attached great importance to the Zhao Mu system.60
This is mentioned in many literary sources. For Emperor Wencheng, and per-
haps to Tanyao for that matter, implementing the Zhao Mu system not only
provided for sacrifices to his forefathers, but also legitimized the worship of
the emperor as the Buddha. Cave 19 is a unique cave temple, and the seated
Buddha is the “largest and most powerful of all Yungang sculptures.”61 It
would be most suitable for Emperor Daowu, the founding emperor.62 It was
argued earlier that cave 17 was intended for Emperor Jingmu as the future
Buddha and cave 16 was for Emperor Wencheng taking the last place of
the family. As such, this would leave cave 18 to the left and cave 20 to the
right of the central cave (19) not connected with the two remaining emper-
ors (Mingyuan and Taiwu). According to the Zhao Mu system, the left is
superior to the right, and the prime position is the center for the founding
emperor. In this way, cave 18 would have been made for emperor Mingyuan,
and cave 20 for Emperor Taiwu. It is more plausible that the positioning of
the imperial five followed the Zhao Mu system rather than the reverse con-
figuration proposed by Rei.63
Up to now, we have not had any concrete statistics on how many gong
would have been used in building a rock-cut cave in Yungang.64 But in the
Longmen complex it is recorded that 823,066 gong were used from the
first year of Jingming (500) to the fourth year of Zhengguang 正光 (523)
to construct the three Binyang caves 賓陽三窟 dedicated to Emperor Gaozu
(Xiaowen), Empress Dowager Feng, and Emperor Shizong (Xuanwu).65 In
the Dazhusheng cave temples 大住聖窟 in Anyang, which were patronized
by the eminent monk Lingyu 靈裕 of the Sui Dynasty (581–618), the Record
of Excavating a Cave Temple 造窟記 carved in the cave indicates that 1,640
gong were used to excavate a cave, and it took up to 900 gong to make a
Buddha image.66 These records are extremely important because they not
only demonstrate the tremendous workload required to excavate a cave,
but also suggest the total amount of work and time involved. As such, the
workload of a gong can be calculated. According to the meticulous analysis
and calculation of Rei, cave 19 occupies the largest excavation space, more
than 8,870 m2, and took 190,000 people 2,332 days to excavate. Caves 20
and 18 took 110,000 and 100,000 people to excavate, respectively, so they
are medium sized (M) compared to the large size of cave 19 (L).67 It took
80,000 and 70,000 people to excavate caves 17 and 16, respectively; thus,
they are small-sized (S).68 With these statistics, the initial plan (shown in
table 3.1, below) proves to be more symmetrical and balanced with the larg-
est cave in the center, two medium-sized caves flanking it, and the two small
caves at each end. Therefore, cave 19 took the prime position with bilateral
symmetry. This is a more reasonable plan than the current arrangement. The
difference between this configuration and the current positioning of the caves
is that cave 16 has been placed to the left of cave 17. Something dramatic
might have happened to change the initial grand plan. The only clue we have
 Phase one – emperor as Tathāgata 67
Table 3.1 Initial plan for caves 16–20

Emperor Wencheng Taiwu Daowu Mingyuan Jingmu

Cave Cave 16 Cave 20 Cave 19 Cave 18 Cave 17

Sequence of 5 3 1 2 4
Construction
Height of 13.5m 13.7m 16.8m 15.5m 15.5m
the colossal
Buddha
Size S M L M S

had up to now is the collapse of the west and south walls of cave 20. As
was discussed earlier, the walls collapsed soon after the cave was completed.
Mizuno and Nagahiro also mentioned:

The damage may have been caused by the bad and brittle stratum of
rock. The outside wall, after the collapse, seems to have been repaired
with brick-work but of this, at present, only the bottom part on the
east side remains. The front court was excavated in 1940. Many rock
fragments carved with figures that had undoubtedly formed parts of the
walls were discovered during the excavation.69

The front (south) wall collapsed soon after the cave was completed, thus the
initial grand plan of having a small cave to the west of cave 20, i.e., on the
site of what is now cave 21, was out of necessity revised.70 Rei argued that
the change of the grand plan could only have been caused by the collapse
of the south wall of cave 20, and because of this, cave 17 had to be moved
to the current location.71 But it would be more reasonable to move cave 16,
not cave 17. This would also be a convincing explanation for the fact that
cave 16, supposedly for the ruling emperor, was delayed to a time as late as
the 480s, when the boyi bodai Chinese garment style had emerged. This is
precisely why cave 16 was postponed to a much later period.72 A more likely
initial plan for the ritual positioning of caves 16−20 as the family shrine
temple, would have been to follow the Zhao Mu system, but cave 16 had to
be changed when the walls of cave 20 collapsed.

Tanyao and the significance of the translation

of sacred books in Yungang
We do not know much about Tanyao, considering his significance in bringing
Buddhism to a greater height of development after the persecution, persuad-
ing the emperor to excavate the Yungang cave temples, and carrying out
the translations of Buddhist sutras. We do not even know his birth or death
68 Phase one – emperor as Tathāgata
dates, and have only a very small number of fragmentary literary accounts
concerning him. “Śramana Tanyao was virtually unknown. He left house-
hold life when he was young, and was virtuous, pure, and strictly followed
the vinaya vow.”73 He was only mentioned in passing in the biography of
Xuangao in the Gaoseng zhuan. “Tanyao was known for dhyāna practice
and was revered as a teacher (shili 師禮) by the Honorary Grand Mentor
Zhang Tan 張潭 when Juqu Mujian 沮渠牧犍 (r. 433–39) ruled western
Gansu.”74 According to the meticulous research of Tsukamoto Zenryu,
Tanyao was already well known during the Northern Liang period.75 “He
had virtuous qualities and was respected by Emperor Gongzong 恭宗 with
proper rites.”76 After the Buddhist persecution, most of the monks made
their living by their trades, and preferred to be seen as secularized persons.
But Tanyao made a vow to be faithful to the religion unto death. He still
wore Buddhist robes secretly and carried Buddhist objects from which he
was never separated. Emperor Gongzong personally made an effort to dis-
suade him many times, but eventually gave up.77 After Shixian passed away
in the early Heping era (460–65), Tanyao replaced him and became the chief
administrator of śramanas.78 He remained in this position during the reigns
of emperors Wencheng, Xianwen, and Xiaowen, under whom the Yungang
cave temples were excavated.
The Guang hongming ji records that Sengxian 僧顯, head of the Siyuan
Monastery 思遠寺, was appointed to be the chief administrator of monks
to succeed Tanyao, and Sengyi 僧義 from the Huangjiu Monastery 皇舅寺
was to become the vice administrator to assist him.79 From this, it can be
assumed that up until that time, Tanyao was still active and was solely in
charge without the assistance of a vice administrator. It was believed Tanyao
was then about 70 years of age.80 As it is known that the construction of
the Siyuan Monastery on Mount Fangshan (方山) was started in 477, and
Sengxian was the head of the temple when he succeeded Tanyao, this means
that his appointment as the administrator of monks in succession to Tanyao
was no earlier than this date. Tanyao would still have been alive then, or at
least have just passed away. No mention was made of Tanyao after this. The
death of Tanyao might not have been earlier than the year of the completion
of the construction of the Siyuan Monastery in 479, and not later than Liu
Xiaobiao’s 劉孝標 (462–521) flight in 486.81 It was probably very close to
this time. Without further evidence, it is hard to know the exact date of his
birth or death with full confidence, but it can be ascertained that his death
could not have been earlier than 477 when the construction of the Siyuan
Monastery was started.
Tanyao’s greatest contribution and what he is best known for was per-
suading the emperor to hew the cliff of Mount Wuzhou and build the Yun-
gang cave-temple complex soon after the Buddhist persecution. Not only
did he want to build one of the greatest Buddhist monuments with colossal
Buddha images so they could not be destroyed in another Buddhist persecu-
tion, but he also made great contributions to the translation of sacred texts of
 Phase one – emperor as Tathāgata 69
Buddhism so that Buddha Dharma would be carried down forever to subse-
quent generations. These translated scriptures were instrumental in influenc-
ing the visual representations of the second-phase cave temples in Yungang.
Tanyao, together with the śramana Changnayeshe and others, translated
14 new sutras.82 He was the chief administrator in Yungang, and while liv-
ing in the Tongle Temple within the Yungang complex, Tanyao assembled
a great number of monks and skillfully earned their trust.83 The purpose of
his assembling monks in Yungang to translate Buddhist sacred texts is clear:

By the third year of Heping (462), the chief administrator śramana Tan-
yao was still bitter about the persecution and was overjoyed that Bud-
dhism was restored. Therefore he assembled many Buddhist monks to
translate sutras in the Beitai cave temples, so that the sutras could be
carried down to future generations and that the Buddha Dharma would
be carried down forever.84

Apparently Tanyao still lamented the destruction that had occurred and had
a great sense of urgency to spread Buddhism among the sages of later times,
and to allow the Dharma treasury to endure and not be cut off. He collabo-
rated with Kekaya 吉迦夜 in 472 and translated the Zabaozang jing and the
Fufazang yinyuan zhuan.85 He translated scriptures to preserve the Buddha
Dharma and prevent it from falling into decline. It is worth mentioning that
Tanyao, together with Kekaya, translated and re-translated the Fufazang
yinyuan zhuan a second time because he was not satisfied with the previous
translation. The significance of the scripture is obvious.86
Among the 14 new scriptures he translated, three (the Fufazang yinyuan
zhuan, the Zabaozang jing, and the Dajiyi shenzhou jing are still extant.
These translations were transformed into pictorial narratives in the process of
image-making in Yungang, and reflect association between words and images,
and between sacred texts and pictorial interpretation. In the first phase, pro-
tecting the Dharma-pitaka and transmitting it everlastingly to the following
generations, was an urgent mission to conduct after the Buddhist persecution,
thus the three Buddhas of the three kalpas became the primary focus, and the
Buddhist ideal of the emperor as the living Tathāgata was promoted with the
patronage of the Northern Wei imperial court. In the second phase, narra-
tive stories derived from the translated texts became important subjects. For
instance, jataka and nidana narrative stories from the texts began to appear
in large quantity in the second-phase caves. The stories depicted in caves 9
and 10 are mostly based on the Zabaozang jing. Though they are extremely
popular motifs, they are not found in any other cave temples in Yungang.
The well-known stories of Hariti losing her son, the Brahmana (the seller
of human skulls), and the story of the woman who hated desire and became
a nun, etc., are not observed anywhere in Yungang other than in caves 9
and 10. This unique phenomenon not only indicates the close connection
between the Zabaozang jing and image-making in Yungang, but also, more
70 Phase one – emperor as Tathāgata
importantly, displays the significance of the sacred texts translated by Tanyao
in Yungang.
One may argue that it was the Lotus Sutra, a then-influential text, that influ-
enced the subject matter in caves 9 and 10, and it was because of the popularity
of the Lotus Sutra that some motifs (especially the three Buddhas of the three
kalpas, and the Śākyamuni Buddha and Prabhutaratna) were represented
throughout the three phases.87 It is true that the Lotus Sutra describes how
the three Buddhas succeeded one after the other, without intervention, and, in
due time, expounded the great Dharma.88 Before the Buddha entered nirvana
he assured the countless multitudes concerning the succeeding Buddha:

When I enter the extinction, you must not be concerned or fearful. This
Bodhisattva Virtue Storehouse has already fully understood in his mind
the true entity that is without outflows. He will be next to become a
Buddha, bearing the name Pure Body, and he too will save immeasur-
able multitudes.89

However, the fact is that the key concept of having a succession of Bud-
dhas promoted in the Lotus Sutra was, in essence, Tanyao’s primary con-
cern. He had just gone through a bitter experience of Buddhist persecution
by Emperor Taiwu, who believed that Buddhism and its teachings were
untrue.90 As stated earlier, the purpose of translating and compiling sutras
was to assure that the great law would be carried down forever without
interruption. In addition, both the Fufazang yinyuan zhuan and the Dajiyi
shenzhou jing vigorously advocate the significance of the succession of
Buddhas and the transmission of the Dharma. It was with the same pur-
pose that he urged Emperor Wencheng to carve the colossal images out
of living rock since it better withstands destruction. Protecting and trans-
mitting the Buddha Dharma became a key concept after the first Buddhist
persecution.
The Fufazang yinyuan zhuan is a compilation, excerpted from a number
of other sources, of biographical details of the 24 Indian patriarchs who
succeeded the Buddha after he entered nirvana. These biographies were
carefully chosen by Tanyao in order to establish a legitimate and uncor-
rupted line of the succession of the Buddha and of the transmission of the
Dharma-pitaka. When the Buddha entered nirvana, he was succeeded
by Mahakasyapa who, in turn, was succeeded by Ananda. The succession
continued until the 24th patriarch.91 It is as such the Buddha’s disciples
succeeded him and transmitted the Dharma one after the other without
interruption. To the Northern Wei imperial rulers and Tanyao, it was criti-
cal that there be only one patriarch at any given time, just as there could
only one legitimate Chinese emperor at any given time. Tanyao chose the
biographies of the 24 eminent monks to testify the authenticity of the
Dharma-pitaka and the legitimate succession of the Buddha. The first four
chapters of the Fufazang yinyuan zhuan were derived from the Ayuwang
 Phase one – emperor as Tathāgata 71
zhuan (The Biography of King Asoka) in which the transmission of the
Dharma-pitaka by the five disciples of the Buddha was recounted.92 Tanyao
was still haunted by the fear of persecution or any unpredictable events
that might interrupt the succession and transmission of the great law. He
excerpted biographical materials from many other sutras in order to bring
the Dharma closer to Chinese culture and make it accepted as authentic,
arguing against the edict issued by Emperor Taiwu in the seventh year of
Taipingzhenjun (446) when Buddhism was persecuted. It was against the
claim of “using the concept of the Emptiness of Lao and Zhuang and graft-
ing it (to Buddhism) to benefit from it” that Tanyao compiled the Fufazang
yinyuan zhuan to ascertain the authenticity of the Dharma and to protect
it. Clearly Tanyao was striving vigorously to promote the teachings of Bud-
dhism after the first persecution, and for that very reason he translated the
Fufazang yinyuan zhuan twice.
The Fufazang yinyuan zhuan was not the only sutra translated by Tanyao
that promotes a succession of Buddhas and the transmission of the Dharma-
pitaka. The Dajiyi shenzhou jing advocates the three Buddhas in three kal-
pas as well. In the beginning of the sutra, it emphasizes the significance of
the seven Buddhas of the Past, the Śākyamuni Buddha, and the Maitreya
Buddha. It also mentions the Buddhas in the Past, Present, and Future.93 In
addition, the scripture promotes that the Dharma protectors will protect the
rulers and avert disasters.94 This was a highly esteemed ideological concept
at the time after the persecution of Buddhism. Thereupon, it is not surprising
that in the second phase, as is observed, the Dvarapalas began to appear and
became an increasingly important subject in Yungang.
It can be seen that the Lotus Sutra is by no means the only Buddhist text
that played a part in visual images. The scriptures translated by Tanyao and
others were important sources of inspiration for pictorial narratives as well.
Tanyao’s contributions to Buddhism consisted not only of initiating the
excavation of Yungang and translating the sutras, but also his lesser known
establishment of the Sangha-household (sengzhihu 僧祇戶) and Buddha-
household (fotuhu 佛圖戶) economic systems. The former ensures a chari-
table fund to be used in times of disaster, and the latter provides human
resources for monastic farming, thereby ensuring a stable source of income
for monastic communities. Tanyao petitioned that the households of Pingqi
平齊 and those of the people who could yearly convey 60 hu 斛 of grain and
present them to the clerical officials constitute Sangha-households, and their
grain be designated Sangha-grain, to be used in lean years to relieve famine-
stricken people.95 He also requested that those of the people who committed
great crimes, as well as the public slaves, be formed into Buddha-households,
to serve the temples as sweepers and sprinklers, and also manage the fields
and transport the grain. Emperor Gaozu granted all these requests.96 There-
fore Sangha-households and Sangha-grain and temple-households were to
be found everywhere in the prefectures and garrisons.97 The system was an
ambitious enterprising way of managing a Buddhist community.98
72 Phase one – emperor as Tathāgata
Concluding remarks
The five Tanyao caves promote Buddhist ideal that the emperor is the liv-
ing Tathāgata and that the merits of the Sakya Tathāgata saved the Grand
Chiliocosm and His compassion flows out to the grimy regions.99 They best
exemplify the close association between imperial power, image-making, poli-
tics, and Buddhist rock-cut cave art under the patronage of the imperial fam-
ily. The most striking feature of these cave temples is the gigantic size of the
image inside each cave temple. The largest and most powerful is the seated
Buddha in cave 19. “Its grandeur and dignity combine to create an impres-
sion of sanctity and heroic might.”100 It would be impossible to provide,
either in words or visual images, any adequate idea of the enormous impres-
siveness of these Buddhist cave temples. This impressiveness comes not from
the builders’ providing a sense of space, for here space is completely con-
trolled and restricted, but from the beauty and austerity of the architectural
members and the mystery provided by the twilight which in these interiors
seems to make everything melt and almost disappear, so that the visitor feels
he is in a magic world of unreality.101 The rays of light through windows
of the cave temples shine on the face of the Buddha, illuminating the entire
space and increasing the sense of surreal mystery. In cave 18, the window is
larger than the gateway, revealing most of the Buddha image.
The subject matter of the caves is primarily focused on the three Buddhas
of the three kalpas and the thousand Buddhas, indicating that the Buddha
Dharma would be carried down without interruption in the light of that Bud-
dhism which was just revived after seven years of catastrophic suppression.
The belief in Dharma protection was stronger than ever after seven years of
Buddhist persecution.
The colossal images of the imperial five best represent Buddhist sculpture
and Buddhist art, as it gradually made its way eastward in China. This set
of five caves can be regarded as the masterpiece of Buddhist image-making.
The scale of the caves and the images is unprecedented and has never subse-
quently been matched. The majestic and stately impressiveness are brought
out not only by the grandeur of the images but also by refined craftsmanship.
The Buddha images are carved either in “closed mode” with both shoulders
covered or in “open mode” with right should slightly covered with a piah-
shan. The faces are round, and heads are big with broad chests, full of vigor.
Some are dressed in thick robes characteristic of Gandhara images; others are
in “wet clothes” of Indian style. Some Buddha images appear personal with
gentle smiles, and others look more dignified. The Ten Great Disciples of
Śākyamuni Buddha in cave 18 are so vividly carved that one standing in front
of them is immediately attracted to them and lost in gazing at them. After all,
the imperial five were constructed under imperial patronage in Pingcheng,
the capital of the Northern Wei, where different cultures and peoples from
different regions and civilizations met and contributed to building one of the
greatest Buddhist monuments in all history.
 Phase one – emperor as Tathāgata 73
The new archaeological findings not only fill in the blanks in our knowl-
edge about Yungang itself. They broaden our vision in examining the
rock-cut caves, and the religious and secular art. Essentially, the surviving
archaeological findings lead to new directions connecting the rock-cut Bud-
dhist sanctuaries with the secular material culture in the capital, and linking
religious art with mortuary art. This is significant in terms of study of Bud-
dhist art and architecture. It not only broadened the approaches of Buddhist
art and architecture, it also indicated that the influence between secular and
religious is reciprocal. Religious culture had an impact on the secular culture.
The most significant reference from the tent structure for the cave temples is
the use of huge windows, which allow sufficient rays of light to illuminate a
sacred space cut out of living rocks. It gives the spiritual gaze of humankind
at a Buddhist image a great value.

Notes
1 Wei Shou (506–572), comp., Weishu (Beijing: Zhonghua shuju, 1974),
114:3037.
2 Ibid., 114:3035.
3 Ibid., 114:3036.
4 Ibid. The Grand Five-Storeyed Temple is no longer in existence and the original
location is unknown.
5 Ibid., 114:3031. The claim that the emperor is the present incarnation of Tathāgata
was initially made by Faguo, a śramana of Prefecture of Zhao. He stated that
Taizu is enlightened and loves the Dao (Way). He is the very person of the Thus-
Having-Come-One. Śramana should pay him homage to the utmost. He should
often do obeisance and tell others that he who promotes the teachings of Buddha
is the Lord of men. He was in effect saying, “I am not doing obeisance to the
emperor, I am merely worshipping the Buddha.”
6 Tokiwa Daijo and Tadashi Sekino, Chugoku bunka shiseki. Kaisetsu 001(01)
(Kyoto: Hozokan, 1975), 2–4.
7 Sekino Tadashi and Daijo Tokiwa, Shina Bukkyo shiseki (Tokyo: Bukkyo
Shiseki Kenkyukai, 1926). English version, 16.
8 Weishu, 114:3035.
9 Tokiwa Daijo and Tadashi Sekino, Buddhist Monuments in China: Text Part
(Tokyo: Bukkyo-shiseki Kenkyu-kwai, 1926). English version of Shina Bukkyō
shiseki hyōkai, 16.
10 Su Bai, “fenqi,” 76; and Ding Mingyi, “Yungang shiku de kaizao licheng,” in
Yungang shiku bainian lunwen xuanji, vol. 2. Ed. Yungang shiku wenwu yan-
jiusuo (Beijing: Wenwu chubanshe, 2005), 128.
11 In the Shaling tomb mural paintings, the banquet scene on the south wall shows
the wooden structure buildings and the tents used by the people. The tents are
both square and round in shape.
12 Fan Ye (398–445), comp., Hou Hanshu, “Wuhuan, Xianbei liezhuan (Treatise on
the Wuhuan and Xianbei),” (Beijing: Zhonghua shuju, 1976), 90:2979. The tent
was called zhanzhang 毡帳 (felt dwelling) or baizizhang 百子帳 (hundred son tent)
during the Northern Wei. It was mainly used by the nomads, but Chinese people
used them in weddings, hence the name baizizhang, tent of one hundred sons.
13 Liang Sicheng, “Yungangshiku zhong suo biaoxian de Beiwei jianzhu,” in Yun-
gang bainian lunwen xuanji. Ed. Yungang shiku wenwu yanjiusuo (Beijing:
Wenwu chubanshe, 2005), 17.
74 Phase one – emperor as Tathāgata
14 Su Bai, “Yungang moshi,”122.
15 Su Bai, “fenqi,” 77.
16 The story is derived from the Zabaozang jing (the Sutra of the Miscellaneous
Treasures) translated by Tanyao and others in Yungang.
17 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vols. 13–14, 135. The rendi-
tion of Buddha’s robe is a perfect example of the so-called caoyi chushui 曹衣出水
style created by the Northern Qi (550–577) painter Cao Zhongda 曹仲達 who was
known for making the garment of his figures look drenched, as if the figures had
just emerged from the water. The garments cling closely to the body of the figures.
See Guo Ruoxu 郭若虛, Tuhua jianwen zhi in Huashi Congshu. Ed. Yu Anlan
(Shanghai: Shanghai renmin meishu chubanshe, 1963), 1:10.
18 Li Daoyuan, Shuijing Zhu jiaozheng, annotated by Chen Qiaoyi (Beijing: Zhon-
ghua shuju, 2007), 316.
19 Dao Xuan, Da Tang neidian lu, T55:2149:267b28.
20 Osvald Sirén, Chinese Sculptures from the Fifth to the Fourteenth Century,
vol. 1 (New York: Reprinted by Hacker Art Books, 1970), 8; and Édouard
Chavannes, Mission archéologique dans la Chine septentrionale (Paris: Ernest
Leroux, 1909).
21 Katherine R. Tsiang, “Reconsidering Early Buddhist Cave-Making of the North-
ern Wei in Terms of Artistic Interactions with Gansu and the Western Regions,”
Journal of Inner Asian Art and Archaeology, no. 4 (2009): 101–18.
22 Weishu, 4:90.
23 Ibid., 114:3032.
24 Ibid., 114:3030.
25 Ibid., 114:3030.
26 Chen Yuan, “Ji Datong Wuzhoushan shikusi,” in Chen Yuan, Chen Yuan xue-
shu lunwen ji, vol. 1 (Beijing: Zhonghua shuju, 1980), 398–409.
27 Weishu, 114:3037.
28 Su Bai, “fenqi,” 76.
29 Mizuno Seiichi, Unkō sekkutsu, vols. 13–14, 92.
30 Yan Wenru, “Yungang shiku de kaichuang he ticai de fenxi,” Shehui kexue
jikan, no. 5 (1980): 112–18, and no. 6 (1980): 110–14; Also see Yan Wenru,
Yungang shiku yanjiu (Guilin: Guangxi shifan daxue chubanshe, 2003), 7.
31 Yoshimura Rei and Liqiang Bian. Tianren danshengtu yanjiu: Dongya fojiao
meishushi lunwenji (Shanghai: Shanghai guji chubanshe, 2009), 444.
32 The Weishu mentioned that Shixian returned immediately to the life of a śramana
on the very day of the restoration of Buddhism. The emperor himself shaved
off his hair and Shixian became the daorentong. Wei Shou, comp., Weishu,
114:3037.
33 Xu gaoseng zhuan, T50:2060:427c24.
34 Ibid., T50:2060:428a08; Kaiyuan shijiao lu, T55:2154:539c28; Lidai sanbao
ji, T49:2034:85b5; Fayuan zhulin, T53:2122:875b18. Also see the Wei Shou,
comp., Weishu, 114:3037.
35 Su Bai, “Jinbei Jiaozhu,” 65.
36 Su Bai, “fenqi,” 77.
37 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vol. 16, 30.
38 Ibid.
39 Yoshimura Rei and Liqiang Bian, 388.
40 Ibid., 455.
41 Su Bai, “Jinbei,” 104.
42 Jin Shen, Zhongguo lidai jinian foxiang tudian (Beijing: Wenwu chubanshe,
1994), figs. 13 and 14.
43 However, they contradicted themselves later and concluded that the excavation
of cave 16 is likely to have been started in the period of the Tanyao five caves,
 Phase one – emperor as Tathāgata 75
i.e., the Heping era (460–465). The carving of the main Buddha of the north
wall would naturally have proceeded to a certain degree during these periods.
Although there is no trace of carving of early Yungang type in the image, the
broad contours of the shoulders, even though not well-rounded, are one of
the characteristics of the early Yungang type and may afford some support for
the theory. Furthermore, the fact that cave 16 has a larger space in front of the
main image than the other caves of Tanyao leads one to imagine an original
earlier main Buddha occupying a much larger space. See Mizuno Seiichi and
Nagahiro Toshio, Unkō sekkutsu, vol. 11, 109. Mizuno and Nagahiro often
have different views, and they contradict each other.
44 Tokiwa Daijo and Tadashi Sekino, Chugoku bunka shiseki. Kaisetsu (01)
(Kyoto: Hozokan, 1975), 4–6. They suggested that all five emperors should be
previous late emperors and believed that they were emperors Daowu (Tuoba
Gui), Mingyuan (Tuoba Si), Taiwu (Tuoba Tao), Jingmu (the Crown Prince,
Tuoba Huang), plus emperor Pingwen.
45 He did not include Emperor Pingwen. See Tsukamoto Zenryu, Shina Bukkyō shi
kenkyu Hokugi hen (Research on the History of Chinese Buddhism. Northern
Wei Section), (Tokyo: Kobundo, 1942), 225.
46 Tsukamoto Zenryu, Shina Bukkyō shi kenkyu Hokugi hen, 222.
47 Ōmura Seigai first proposed the theory in 1915 and later the theory was widely
accepted. See Ōmura Seigai, Shina bijutsushi choso hen (Tokyo: Bussho kanko-
kai zuzobu, 1915), 179.
48 Hui Jiao, Gaoseng zhuan, T50:2059:352a.
49 Alexandra Soper, “Imperial Cave-Chapels of the Northern Dynasties: Donors,
Beneficiaries, Dates,” Artibus Asiae, vol. 28, no.4 (1966), 242; Yoshimura Rei,
Tennin tanjozu no kenkyu: Higashi Ajia Bukkyo bijutsushi ronshu (Tokyo:
Tohoshoten, 1999), 256; and Su Bai, “fenqi,” 76.
50 Su Bai, “Yungang moshi,” 121.
51 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu vol. 11, English text, 108.
52 Ibid., vols. 13–14, English text, 125.
53 Ibid.
54 Weishu, 114:3030.
55 Ibid.
56 Ibid.
57 Ibid.
58 Michael Loewe, “The Heritage Left to the Empires,” in The Cambridge History
of Ancient China: From the Origins of Civilizations to 221 B.C. Ed. Loewe
Michael and Edward L. Shaughnessy. (Cambridge: Cambridge University Press,
1999), 1030.
59 Ibid.
60 Hang Kan 杭侃, “Yungang di ershi ku tanta de shijian yu Tanyao wuku zuichu
de Buju sheji,” in Yungang Bainian lunwen xuanji. Ed. Yungang shiku wenwu
yanjiusuo (Beijing: Wenwu chubanshe, 2005), 345.
61 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vols. 13–14, 108.
62 Yoshimura Rei and Liqiang Bian, 458.
63 Ibid.
64 Gong is a measurement of a working day of an artisan involved in the construc-
tion of the caves. In ancient China this was the way to calculate the volume of
work.
65 Weishu, 114:3043.
66 According to the Da Tang Liudian 大唐六典 (The six statutes of the Tang
dynasty) compiled by Li Linfu 李林甫 (d. 752) et al., in the Tang dynasty, based
on the length of the working hours, a gong is divided into three types. From the
fourth to seventh months, the amount of work is considered a long gong 長功;
76 Phase one – emperor as Tathāgata
in the second and third months, and eighth and ninth months, the workload is
considered a medium gong 中功; and in the tenth to the 12th months, the amount
of work is only a short gong 短功. See Li Linfu, et al. The Da Tang Liudian.
Annotated by Chen Zhongfu 陳仲夫 (Beijing: Zhonghua shuju, 1992), 595.
67 This does not include the excavation of the side caves since accurate measure-
ments of the caves are not available. See Yoshimura Rei, Tennin tanjozu no
kenkyu, 449.
68 Yoshimura Rei and Liqiang Bian, 449.
69 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vols. 13–14, 115.
70 Yoshimura Rei and Liqiang Bian,, 451.
71 Ibid., 450–1.
72 Hang Kan, “Yungang di ershi ku tanta de shijian yu Tanyao wuku zuichu de
Buju sheji,” 345.
73 Xu gaoseng zhuan, T50:2060:0427c22.
74 Gaoseng zhuan, T50:2059:398b10. Here “revered as a teacher” means to be
respected as a reader of the Buddhist sutras or as a leader in the observance of
religious commandments and in religious devotions, see Mizuno Seiichi and
Nagahiro Toshio, Unkō sekkutsu, vols. 13–14, 94.
75 Tsukamoto Zenryu, Shina bukkyōshi kenkyū (Tōkyō: Shimizu Kōbundō, 1969),
79.
76 Ibid, 114:3037.
77 Ibid.
78 Ibid, 114:3037.
79 Dao Xuan, Guang hongming ji, T52:2103:27b18.
80 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vols. 13–14, 101.
81 Ibid.
82 Weishu, 114:3035 and 114:3037.
83 Xu gaoseng zhuan, T50:2060:0427c22.
84 Fei Changfang, comp. in 597, Lidai sanbao ji, T49:2034:85b05.
85 Sengyou (445–518), Chu sanzang ji ji (A Catalogue of the Tripitaka),
T55:2145:13b06.
86 Evidently this is a rather important sutra to Tanyao and he was not satisfied
with his first translation. Thus he invited Kekaya to collaborate with him in 472.
The promotion of Buddhism and the protection of Buddha Dharma were so
important to Tanyao who just survived the seven years of Buddhist persecution
that he felt great need to translate the sutra twice to assure the transmission of
the Dharma and continuation of the relay.
87 Li Jingjie, “Guanyu Yungang dijiu, dishi ku tuxiang goucheng,” in Yishushi
yuanjiu, Ed. Zhongshan daxue yishushi yanjiu zhongxin, vol. 10 (Guangzhou:
Zhongshan daxue chubanshe, 2008), 327–59.
88 “When the Buddha Sun Moon Bright had finished preaching this sutra over a
period of sixty small kalpas, he spoke these words to the Brahmas, devils, shrama-
nas and Brahmans, as well as to the heavenly and human beings and asuras in the
assembly, saying ‘tonight at midnight the Thus Come One will enter the nirvana
with no remainder.’” “At this time there was a Bodhisattva named Virtue Store-
house. The Buddha Sun Moon Bright bestowed a prophecy on him, announcing
to the monks, ‘this Bodhisattva Virtue Storehouse will be the next to become a
Buddha. He will be called the Pure Body, tathagata, arhat, samyak-sambuddha.’”
“After the Buddha had finished bestowing this prophecy, at midnight, he entered
the nirvana of no remainder.” T 9:262:5a17. Eng. translation by Watson Burton,
see Watson Burton and Kumārajiva, The Lotus Sūtra (New York: Columbia Uni-
versity Press, 1993), 16.
89 Burton and Kumārajiva, The Lotus Sūtra, Watson translation, 20.
90 Weishu, 114:3034.
 Phase one – emperor as Tathāgata 77
91 As we are told in the sutra that the World-Honored One transmitted the Dharma
to Mahakasyapa and entered nirvana; Mahakasyapa transmitted it to Ananda
and entered nirvana; Ananda transmitted it to Madhantika and entered nirvana;
Madhyantika transmitted it to Sanakavasin and entered nirvana; Sanakava-
sin transmitted it to Upagupta and entered nirvana; and Upagupta transmit-
ted it to Dhitika . . . until the twenty-fourth disciple. T50:2058:297b06 and
T50:2058:301a23, etc.
92 The Ayuwang zhuan was translated by An Faqin 安發欽, a certain monk from
An Xi 安息 (Parthia), who came to Luoyang in 281 during the reign of Emperor
Wu of the Western Jin Dynasty (265–316). There is another Chinese text nar-
rating the legendary account of the life of King Asoka called the Ayuwang jing
阿育王經, which was translated into Chinese by Samghapala (459–524) 僧伽
婆羅, who was a monk from the kingdom of Funan (eastern part of present
Thailand). He came to China during the Qi dynasty (479–501) and stayed
in the Zhengguan Monastery 正觀寺 in the capital and studied Mahayana
texts with the Indian monk Gunabhadra and “comprehended the languages
of several countries,” see Gaoseng zhuan, T50:2059:345b09 and Xu gaoseng
zhuan, T50:2060:426a09. Judging from the time of the translation of the two
sutras, evidently Tanyao excerpted the accounts of the transmission of the
Dharma-pitaka by the disciples of the Buddha from the Ayuwang zhuan not
the Ayuwang jing which is a much later translation.
93 Tanyao, trans., Dajiyi shenzhou jing, T21:1335:568a06.
94 Ibid., T21:1335:571c16.
95 Weishu, 114:3037.
96 Mistakenly, the original text mentioned that it was Gaozong.
97 Weishu, 114:3037.
98 Serfs and former aristocrats from Shandong were indentured to Buddhist mon-
asteries, and became a labor force for the economic advancement of the Sangha.
Grains thus produced were ostensibly for distribution in years of famine. Bud-
dha households were composed of criminals made into serfs of monasteries.
Monastic households were required to submit 60 shi (approx. 30kg) of millet to
monasteries and thus were no longer state registered, exempted from state taxes
and corvée labor. Millet (which was also used as cash during this period) submit-
ted to and stored by the monasteries could be used as disaster relief in years of
low yield. The number of those households was not specified, but as recorded
in the Weishu, this new category of households soon pervaded every prefecture
and town. See Weishu, 114:3037. Regarding the monasteries’ taking on of the
task of disaster relief, Gernet asked the question “for what reason was a public
service (disaster relief), which should normally have come under the authority
of secular officials, under these circumstances entrusted to the Church”. See
Jacques Gernet in Buddhism in Chinese Society: An Economic History from
the Fifth to the Tenth Century (New York: Columbia University Press, 1995),
102. In 316, the recorded number of Buddhist monasteries was 182. It rose to
more than 8,000 at the beginning of the 6th century, and 30,000 to 40,000 in
the middle of 6th century.
99 Weishu, 114:3035. The Grand Chiliocosm is the trisahasramahasasralokadhatu
(sanqian daqian shijie 三千大千世界 in Chinese). There are also Middle Chilio-
cosm and a Small Chiliocosm. It is an Indian cosmological expression, found in
Dazhidulu Agamas and other works.
100 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vols. 13–14, 106.
101 Benjamin, The Art and Architecture of India, 71. The same impressiveness
applies to Chinese Buddhist rock-cut sanctuaries.
4 Phase two – political struggles
and chronicles reconsidered

Dating and periodization are an important part of the study of Buddhist

archaeological remains. This is particularly true of the Yungang cave temples.
If the issue can be settled, it would advance the study of Yungang to a higher
level. Since the beginning of the 20th century, many scholars have devoted
themselves to the investigation of the cave temples in Yungang. However,
there is still a great divergence of views on the construction dates and chrono-
logical sequence of the second-phase cave temples (caves 1 and 2, 5 and 6,
7 and 8, 9 and 10, and caves 11−13, as well as cave 3) (fig. 4.1). Scholars,
especially Chinese and Japanese, have debated these issues for decades, as I
discussed briefly in Chapter One, and there is still no consensus. The Japanese
scholars Mizuno Seiichi and Nagahiro Toshio suggested that the construc-
tion of the caves in the second phase started with caves 7 and 8, followed by
caves 9 and 10, caves 11–13, caves 1 and 2, and caves 5 and 6.1 Cave 3 was
left largely unfinished. The Chinese scholar Su Bai, however, argued a differ-
ent chronological sequence, suggesting that caves 1 and 2 were constructed
immediately before caves 11−13.2 The suggested dates of construction of
the caves differ as well. One of the keys to the point in controversy lies in
the dates of caves 11−13. Solving this problem would settle the fundamental
issue of the chronological history of the second-phase cave temples. Su Bai,
being cautious, avoided the issue skillfully in his discussion of the chronicle

Fig 4.1 Layout of the second-phase caves

 Phase two – political struggles 79
of the caves. I have pointed out elsewhere that caves 11−13 belong to the
earlier caves in the second phase, and of these, caves 13 and 11 are the earli-
est, but the conclusion was preliminary.3 There is still much more concern-
ing the second-phase caves that warrants careful, in-depth reconsideration.
Therefore, this chapter attempts once again to determine the dates of caves
11−13 from the perspective of the probable and rational use of the outside
cliff surface in the middle section of Yungang during excavation. On the
basis of this, I will reconsider the dates and reconstruct a chronological his-
tory of the caves in the second phase. I have noted in my field research that
the construction of a cave temple usually involved both chiseling man-made
excavations into the mountain and incorporating the natural cliff surface into
the design. Li Daoyuan mentioned that the caves were excavated according to
the nature of the cliff surface (yinyan jiegou).4 Daoxuan in Xu Gaoseng zhuan
also observed that the caves were chiseled based on the nature of the moun-
tain cliff (jiu’er juezhi 就而镌之).5 The use made of the cliff surface outside
the caves and the excavations made inside them can usually reveal the spatial
relationships between adjacent caves, and between the caves and niches. They,
in turn, can shed some light on the time sequence of the excavation.

Dates of caves 11–13 reconsidered

Before we dive into the discussion of the dates of caves 11–13, let us review
briefly the cave structures. Caves 11–13 are situated in the west section of
the central group in Yungang (fig. 4.2). To the east are caves 9 and 10, and
to the west is the shallow mountain valley, beyond which are caves 14 and
15, and the Tanyao five caves starting the western group of cave temples.

Fig 4.2 Caves 11−13

80 Phase two – political struggles
Cave 13 is a colossal statue cave (大像窟) (plate 28). The plan of the
cave, like the five Tanyao caves, is elliptical with a flat ceiling rounded at the
southern corners. Two intertwined dragons, surrounded by clouds and flying
celestials, form a symmetrical pattern on the ceiling and symbolize kingship,
thus indicating royal patronage. There is a window (15 feet high and 13 feet
wide) and an entrance gateway (12 feet high and 10 feet wide) in the cave.
Above the window outside, there is a row of three rectangular holes, above
which a line is cut into the rock horizontally, suggesting that there used to be
a wooden structure that continued over to cave 12.6 The majority of the cave,
including the colossal Bodhisattva image, the dragons and flying celestials on
the ceiling, and the two Dvarapalas in the lower level of the reveals of the
entrance gateway, was finished according to a pre-conceived plan. The focus
of the cave is upon the colossal Maitreya Bodhisattva with crossed ankles.
The south, east, and west walls are all filled with storied niches, and the
images in and around them show that they are later works. It is interesting
to note that each side of the bottom story of the south wall contains a row of
seven worshippers with palms together facing towards the entrance gateway.
Cave 12 is a Buddha shrine cave temple that imitates a wooden structure. It
is one of the most colorfully decorated caves in the entire Yungang complex.
The cave, rich in subject matter with a variety of illustrations of stories ranging
from Buddha’s life to his previous existences, comprises an anteroom and a
main chamber.7 The façade is divided into three spans by four octagonal pillars
tapering toward the top. The cave is well planned as an entity, with symmetri-
cal walls divided into upper and lower stories. On the reveals of the window,
two large meditation bhiksus in dhyāna position are depicted. The anteroom
is mainly covered by illustrations of episodes from the life of the Buddha and
his previous existences (the jataka stories). At the top of the east and west
walls in the coved area above are colorful musicians, each with a traditional
musical instrument. These musical figures enliven the cave and endow it with
the character of a concert chamber. The main chamber is rectangular in shape,
and the walls are divided into two stories like the anteroom.
Cave 11 is a caitya (stūpa) cave 塔廟窟. The stūpa in the center is divided
into two stories, the lower with a canopy on each face (plate 29). The top
of the pillar is covered with a five-layered pedestal adorned with floral orna-
ments, in the center of which, on each side, the upper body of a three-faced
and four-armed Asura appears, a figure who lived on the slopes of Mount
Sumeru, holding a sun and moon in its hands.8 The cave is not completed;
only the rough shape of the central stūpa-pillar, the window, the entrance,
and the intertwined dragons are finished. The rear part of the stūpa-pillar was
never carved down to the floor. Most of the north wall was left undecorated
and is even in the original rough chiseled state. The east, west, and south
walls are covered with small niches that lack unity in style or arrangement.
The entire cave, including the height, width, and sizes of the window and
entrance gateway, is out of balance. The cave is small but the window and
entrance are huge. At the bottom of the east reveal of the window, there is
 Phase two – political struggles 81
an inscription dated to 19th year of Taihe (495 AD). Above the inscription a
seated Buddha is represented inside a pointed-arch niche (plate 30). Mizuno
noted that in view of the known date of this niche it would appear safe to
date the rest of the niches, which resemble it so strongly, to approximately
the same time.9 The basic fact to be recognized is that these additional niches
are all closely related to cave 6.10 On the east wall there is another inscrip-
tion dated to the seventh year of Taihe (483 AD) surrounded by thousand
Buddha niches (plate 31). Above the inscription, a row of three Bodhisattvas
is represented; above them are two Buddhas seated side by side. The niches
may have been carved between 483 and 495. In the middle area of the west
wall seven large standing Buddhas are represented underneath a roof niche.
They are more elegantly carved than those in cave 13.
It should be pointed out that cave 11 contains the greatest number of
inscriptions in Yungang. The inscription on the east wall (which bears the
longest description) is dated to 483, the earliest found so far. This indicates
that the cave started to be reused around 483. The inscription on the out-
side wall dated to the 13th year of Taihe (489) is a significant datum that
divides the second-phase caves into early and late periods. The inscription
on the east reveal (dated to 495) is important evidence that enables us to
divide the caves into middle and late periods. Therefore, research on the
history of the construction of cave 11 would be of value in the dating and
periodization of the Yungang cave temples.

1. The problem of caves 11−13 as a group

Scholars usually regard caves 11−13 as a group.11 If this could be estab-
lished, the three caves would then be centered on cave 12, a rectangular
fodianku (Buddha-hall cave temple), with a caitya cave (11) to its left, and
a daxiangku (colossal statue cave, 13) to its right. The composition of the
three caves would then be a caitya cave, a Buddha-hall cave, and a statue
cave. Even though this conforms to the basic configuration of a stūpa and a
Buddha hall as a monastery, the size of cave 12 is too small, compared with
caves 11 and 13, to match the design of a royal cave temple. Therefore, it
is probably not appropriate to consider these three caves as a group. As
for the outside wall and the scale of the caves, the entrance gateways and
the windows of caves 11 and 13 are roughly the same in height and size
(fig. 4.3). The distance between these two caves is 66 feet, almost the same
distance between each of the five Tanyao caves. Caves 11 and 13 display
characteristics more typical of paired caves than caves that form parts of a
group of three. It is more appropriate to consider these two as paired caves.
Cave 12 is clearly low: the top of the cave is only at the same level as the
bottom of the windows of caves 11 and 13. It seems that the cliff surface
was chiseled deeper into the mountain in order to carve the top of the cave.
As a result, the upper part of the outside wall is not in the same plane, and
the entire external surface of caves 11 and 13 cannot be brought into line.
82 Phase two – political struggles

Fig 4.3 External view of caves 9–13A (right to left)

In addition, the anteroom of cave 12 is wider than the main chamber. The
central axis line deviates eastward slightly, and the middle part of the west
wall swells out in the main chamber. This is different from caves 7 and 8,
and caves 9 and 10, which have the same width in both anteroom and main
chamber. If the anteroom and main chamber in cave 12 did have the same
width, the adjoining sidewalls of caves 11 and 13 would have been pushed
inward accordingly. Judging from the layout of cave 13 and the fact that its
east wall is shared by the west wall of cave 12 (they are opposite sides of
the same wall), it seems that the common wall was too thin to carve a large
niche like those on the east wall of cave 12. Those who worked on the west
wall had to avoid piercing through to the other side. Given the fact that the
sidewalls of cave 12 deliberately avoided caves 11 and 13, we may infer that
cave 12 was inserted in between caves 11 and 13 after they had been con-
structed, and was not part of the original plan of a group of caves composed
together. It is hard to demonstrate any compositional relationship of cave 12
with caves 11 and 13.

2. Reconsidering the dates of caves 11−13

The dates of the three caves have long puzzled scholars; their excavation his-
tory is not mentioned in any literary documents. Cave 11 bears at least three
inscriptions with clear dates (483, 489, and 495), from which we can see that
the reconstruction of the cave began at the latest in 483, and continued on
until the very end, even after the capital was transferred to Luoyang. Mizuno
 Phase two – political struggles 83
and Nagahiro suggest that they were dug out many years after the begin-
ning of the Taihe era (477).12 Su Bai, on the other hand, does not provide a
clear answer for the dates of caves, only suggesting that caves 11−13 were
excavated after caves 9 and 10, which were constructed between 484 and
489 according to the Jin stele transcription.13
The cliff surface outside the caves is important evidence for our inquiry
into the dates of the three caves. To excavate a cave temple, the mountain
cliff surface must first be chiseled into a perpendicular plane according to
the scale of the conceived plan of the cave, then the cave is carved into the
mountain plane. Chiseling the mountain is a tremendous amount of work,
therefore the area outside the cave on both sides that does not need to have
caves dug into it will remain in the original sloping form of the mountain.
This may be seen in the western side of cave 13A, and the sloping form of
the mountain on both sides of cave 3.
The cliff surface outside caves 11−13 measures 89 feet in width and 56
feet in height. The cliff protruding southward at the eastern end of cave 11
was originally the stūpa pillar at the western end of cave 10. It is a sym-
metrical configuration with the stūpa pillar on the eastern end of cave 9, but
the body of the western stūpa pillar collapsed as early as in the Northern
Wei period. Now only the contour traces of both sides of the stūpa taper-
ing upward and small niches and figures in the style of the second phase
can be seen. According to the examination of the front part of caves 9 and
10 made by Mizuno and Nagahiro in 1938 and 1940, beneath the ground
surface of the western stūpa, there were remnants of the stūpa pillar base
and animal-shaped carvings above the base.14 The eastern stūpa is a square
nine-storied pagoda gradually tapering upward. Each story contains a small
niche; the south of the stūpa is in the same perpendicular plane as cave 8
next to it. This indicates that caves 9 and 10, and caves 7 and 8 are in the
same perpendicular surface plane. The two large stūpa pillars of caves 9 and
10 protrude out farther from the surface plane of caves 9 and 10 than from
that of caves 11−13. This indicates that caves 11−13 and caves 9 and 10
are not in the same perpendicular surface plane. There is also a protrubance
at the lower part of the cliff to the west of cave 13, at the western end of
which is cave 13A. The cave is not finished; only the exterior appearance of
the four pillars with three openings in imitation of wooden structure, the
shape of the cave, and the ceiling were completed. The pillars and the roof
that were originally planned to be carved clearly project further south than
the cliff surface of cave 13, but the section above the roof of cave 13A is
on the same surface plane as cave 13 and connected to it. This shows that
in order to create cave 13A, the surface plane was extended to the west of
cave 13 above cave 13A. Therefore, the initial cliff surface was limited to
the space between the eastern end of cave 11 and the western end of cave
13, and was not formed at the same time as the surface of caves 9 and 10.
In other words, there is a time difference between the creation of caves
11−13 and that of caves 9 and 10. Noticing this, Nagahiro points out that
84 Phase two – political struggles
the cliff surface of cave 11 was formed earlier than that of cave 10.15 He
further notes,

The appearance of the entire outer wall leads to the conclusion that the
outside wall containing caves 11 to 13 was cut earlier, and that the out-
side wall of caves 9 and 10 was made later, since they are set back from
the general surface of the cut made into the hill-side.16

Nagahiro reaffirmed his position later:

If we consider its relationship with the adjacent cave 10, it can be seen
that the outside wall of cave 11 was finished earlier than that of cave
10. But if we compare the iconographic style of the two caves, it is then
clear that cave 10 is earlier; cave 11 is merely following in step. This,
then, naturally suggests that the images in cave 10 are earlier than the
7th year of Taihe (483).17

There cannot be any doubt that Nagahiro is right to point out that the exterior
wall of cave 11 was cut earlier than that of cave 10, and that the images in
cave 10 are earlier than those in cave 11. It is curious that he later contradicted
himself and proposed that caves 11−13 were constructed after caves 9 and
10.18 As explained above, caves 11 and 13 share the same cliff surface of exca-
vation, which indicates that these two caves were excavated first, and cave 10
was excavated after them. Thus, Mizuno and Nagahiro’s assignment of caves
11−13 to a time subsequent to caves 9 and 10 contradicts the appearance of
the archaeological remains.
Su Bai also proposed that caves 11−13 were dug out later than caves 9
and 10, which he believed to be the Chongfu cave temple built between 484
and 489 according to the Jin stele inscription. If we accept his theory that
caves 11−13 were constructed after caves 9 and 10, that means caves 11−13
were constructed after 484. However, there is a fundamental problem here:
cave 11 bears an inscription with an explicit date (the seventh year of Taihe,
i.e., 483), which indicates that the cave was built, at the latest, before 483,
not after. It is evident that caves 9 and 10 were excavated after caves 11−13,
not before. It therefore does not seem possible to assign caves 11−13 to a
time subsequent to caves 9 and 10.
As for the layout of cave 13, this is the only cave that adheres to the
elliptical plan of the five Tanyao caves. The east, west, and north walls taper
gradually as they ascend toward the ceiling forming a half dome, but the ceil-
ing itself is almost flat, and has a smaller elliptical shape. In iconography, the
colossal Bodhisattva seated with ankles crossed appears basically the same
as the Bodhisattva in cave 17 even though it was repaired and colored at a
later time. It belongs to the early iconographic style. In other words, cave 13
is still a colossal statue cave like the five Tanyao caves. What differentiates
it is that the ceiling contains patterns of intertwined dragons, and a row of
 Phase two – political struggles 85
worshippers that emerges along the bottom story of the walls – details that
do not appear in the five Tanyao caves. This indicates that the impetus and
nature of the cave construction had changed.19
The reason why the construction of cave 11 was stopped is important
evidence in the investigation of the date of the cave. In the Yungang cave
complex, it is rare for a large cave commissioned by the royal family to be
aborted. Except for cave 3, which was abandoned perhaps due to the trans-
fer of the capital to Luoyang, cave 11 is the only example in Yungang. The
reason is worth in-depth investigation.
Nagahiro, out of caution, does not probe into the reason why cave 11
was abandoned halfway.20 In recent years, some scholars have suggested
that the abandonment of cave 11 was associated with Tanyao, the founder
of Yungang, losing his power and influence, but the date and patronage of
the cave have not been discussed.21 However, whether or not Tanyao indeed
lost power and influence, or whether the loss was associated with the murder
of Emperor Xianwen, are merely conjecture. There is not enough evidence
to support either claim. Some scholars have also suggested that the abrupt
abandonment is related to the Gongwei coup d’état (宫闱之变) and that the
excavation of cave 11 was around the end of the Yanxing (471–76) era and
the beginning of the Taihe (477–99) era, i.e., before or after 476, under the
patronage of Emperor Xianwen’s allies during his reign.22 Although the argu-
ments touch upon the crucial point, the evidence is not sufficient to establish
anything about the patronage of the cave. It is hard to associate cave 11, a
royal project, with the allies of the emperor only.
I shall argue here that the abandonment of cave 11 was associated with
significant political changes at the court. During the Xianwen era, the power
struggle between the Emperor and Empress Dowager Wenming (Lady Feng)
reached a critical point. In the Pingcheng period of the Northern Wei, Wen-
ming was the epitome of an Empress Dowager intervening in politics.23 The
Beishi records that Lady Feng, the imperial consort Wenming of Emperor
Wencheng, was from Xindu in the commandery of Changle (長樂, present-
day Jixian county in Hebei). When she was 14, Wencheng was enthroned
and chosen as guiren 貴人 (secondary wife), and later she was designated
as the imperial consort. When Emperor Wencheng passed away, his impe-
rial garments and implements were all burned after three days of imperial
mourning, as was customary. Hundreds of officials and palace people all
cried aloud, and the empress, deeply grieved, screamed and attempted to
throw herself into the fire. People around rescued her, and it took her a long
time to regain consciousness. When Xianwen came to throne, she assumed
the honored status of Empress Dowager Wenming. When the counselor-
in-chief Yihun 乙浑 attempted to usurp the throne, Emperor Xianwen was
only 12 and was still mourning the loss of his father. The Empress Dowager
Wenming then secretly made the important decision to have Yihun executed,
and by this means was able to govern from the imperial throne. At the time
when Xiaowen was born, the Empress Dowager Wenming raised him herself
86 Phase two – political struggles
and stopped handling political affairs. Her behavior became improper; she
accepted one Li Yi 李弈 as her lover, and because of this, Xianwen had him
executed. The Empress Dowager Wenming, displeased at this, had him mur-
dered. In the first year of Chengming 承明 (476) she was honored as Grand
Empress Dowager Wenming and resumed power.24
From this we can see that the first time Empress Dowager Wenming ruled
from the imperial throne was after Yihun was executed in February of the
first year of Tian’an (466), and that after Xiaowen was born (467), she
herself raised him and stopped intervening in political affairs. According
to the chronicle of Xianzu in the Weishu, in the eighth month of autumn in
the first year of Huangxing 皇興 (467), on a dingyou 丁酉 day, the emperor
(Xianwen) visited the Wuzhoushan shikusi (Yungang). On a wushen 戊申
day, Hong (Xiaowen) was born, general pardon was issued, and the reign
title was changed (from Tian’an to Huangxing).25 It is thus evident that
Empress Dowager Wenming “reigned from behind a curtain” for over a
year and a half. That Empress Dowager Wenming stopped intervening in
political affairs suggests that Emperor Xianwen himself controlled the state
with his supporters and allies. Emperor Xianwen changed the reign title
to Huangxing in the eighth month of the second year of Tian’an (467).
On the surface, this was because of the birth of Xiaowen, but in fact, it
reflects a political struggle that was then going on in the palace. The reign
title Huangxing appears to refer to the restoration and rejuvenation of
the emperor’s power. After Emperor Xianwen was enthroned, a series of
Buddhist activities flourished, and the building of Buddhist temples began.
But Empress Dowager Wenming’s personal elevation of Xiaowen, and her
absence from the political scene, were only a temporary reprieve from
the severe power struggle in the palace. She held the trump card in the
person of Xiaowen and was preparing for her return to power. Emperor
Xianwen’s abdication of the throne in the first year of Yanxing (471) is an
early indication of this.
The Weishu notes that Emperor Xianwen was not really interested in state
affairs and often had thoughts of parting from worldly cares. He intended to
abdicate in favor of his eldest surviving uncle, Prince Kang of Jingzhao 京兆
王 (Zitui 子推), but the state officials begged him not to, so he stopped.26 On
the surface it appeared that Emperor Xianwen was weary of worldly affairs
and wished to pursue his interest in Buddhism and Daoism, and therefore
preferred to abdicate and yield the throne to others. But the real situation
was that he was being forced off the stage by Empress Dowager Wenming
and had no choice, and thus thought of abdicating to Prince Kang, not
to Xiaowen, with the intention of casting off her control. 27 When Gaozu
(Xiaowen) ascended the throne, Xianzu (Emperor Xianwen) moved to the
Chongguang Palace 崇光宮 in the Northern Park, and poured over Dao-
ist and Buddhist texts. He had the Luyeyuan 鹿野苑 built on the western
hill within the park. There were stone houses and meditation halls where
dhyāna monks lived.28 But he rarely heard anything about state affairs. In
 Phase two – political struggles 87
the Chronicle of Gaozu in the Weishu, it recounts that in the fifth year of
Huangxing (471), in the eighth month on a bingwu 丙午 day, Xiaowen was
enthroned in the front Hall of Taihua 太華前殿, and granted a general par-
don.29 The reign title was changed in 471 to Yanxing 延興. This signified the
intention to prolong the preceding Huangxing reign period, and shows that
Xianwendi was not willing to resign himself to abdication yet. Of course,
Empress Dowager Wenming would not tolerate any of these actions. Xian-
wen’s execution of Li Yi, the empress’s favorite, was the trigger that caused
the power struggle to explode. Eventually in the sixth month of the sixth year
of Yanxing (476), the empress had him murdered.30 She thereupon resumed
her reign, and reached the pinnacle of her power. It is not surprising, there-
fore, that the building of cave 11, which had been supported by Xianwendi,
was aborted.
Cave 11 is not the only example showing that excavation of cave temples
and the creation of images were closely related to the social environment and
to political power struggles. Cave 16 is another case, as discussed previously.
In their investigation, Nagahiro noticed traces of a row of worshippers’
lower legs that were exposed but their upper bodies were replaced by a row
of thousand Buddhas:

In the summer of 1944, when I was conducting an investigation there, in

the slightly dark cave temple, I noticed that beneath the three rows of the
thousand Buddhas, there was a row of worshippers’ feet. In the center a
bit of an incense burner could be seen. I felt a kind of indescribable trem-
bling. To the east of the incense burner were six male worshippers. Thus
it could be inferred that to the west were six female worshippers. Why
was it necessary to cut out the row of worshippers, plainly memorable
or significant to the patrons, and replace them with the thousand Bud-
dhas? Were the patrons who added the thousand Buddhas later ordered
to eliminate the initial worshippers? Were they always hostile or were
they expressing resentment? I cannot find a clue to the mystery.31

It is still an enigma today. It seems that the “amputated” legs were intention-
ally displayed and left there, otherwise they could have been totally oblit-
erated by the later addition of the thousand Buddhas, which could have
been represented just a little bit lower to remove the lower part of the legs
completely. It has been argued that eliminating the worshippers’ images to be
replaced by thousand Buddhas is an indication of the fierce power struggle
between Emperor Xianwen and Empress Dowager Wenming.32

3. Emperor Xianwen and caves 11 and 13

Based on the analysis above, we may infer that cave 11 was excavated
under the patronage of Emperor Xianwen. From the perspective of the use
made of the cliff surface, caves 13 and 11 are situated closest to the five
88 Phase two – political struggles
Tanyao caves. Cave 13 essentially remains the same cave structure, with a
colossal image and elliptical cave plan, as those of the Tanyao caves. We
may thus infer that cave 13 was excavated first in the second phase, and is
earlier than cave 11. This also corresponds with the excavation sequence
from west to east, as in the Northern Wei, the west cardinal direction being
superior to the east. Also, cave 13 is in the same perpendicular plane as
cave 11. Hence, we may deduce that caves 13 and 11 were both patronized
by Xianwendi. This also corroborates the account quoted in the Jin stele
from the Yunzhongtu (雲中圖), which mentions that in the first year of
Heping (460) in the Wencheng era and the first year of Tian’an (466) in the
Xianwen era, innovations were made, and cave temples were excavated.33
We know two things from the record: (1) during the Xianwen era, cave
temples continued to be dug, and (2) innovations were made. It is impor-
tant to emphasize that during the Xianwen era, cave temples continued to
be created.34
Emperor Xianwen was the eldest son of Emperor Wencheng and assumed
the throne in the sixth year of Heping (465) when he was 12. Like Emperor
Wencheng, Emperor Xianwen was a deeply devout Buddhist. The Weishu
notes that when Xianzu was enthroned, his faith was earnest and profound.
He examined the various scriptures and treatises and delighted in Lao and
Zhuang (i.e., Daoism). He always played host to śramanas and gentlemen
who were able to discourse on the mysterious, and discussed the essence
of the Great Principle with them.35 In the first year of Huangxing (467),
Gaozu (Emperor Xiaowen) was born. It was at this time that the Yongning
Temple was built and the seven-story reliquary erected.36 It was more than
300 feet high, its base and frame were vast, and it was considered the first
temple under Heaven. In addition, a standing Śākyamuni figure (43 feet in
height) was erected in the Tiangong Temple (Temple of the Heavenly Palace,
Devapura in Sanskrit, City of Gods). One hundred thousand pounds of cop-
per and 600 pounds of gold were used.37 In the Huangxing era (467–71),
a three-story stone reliquary was built. The beams, chevrons, lintels, and
pillars, joined together from top to bottom, large and small, were all made
of stone. Firm, solid, skillful, and meticulous, it was a spectacular scene in
the capital.38 One can well perceive that with Emperor Xianwen’s enthusi-
astic promotion, Buddhism in Pingcheng was at the height of its vigor and
splendor.
The Weishu also mentions that Emperor Xianwen visited the cave temples
twice, going to Wuzhoushan shikusi (Yungang) in the eighth month of the
first year of Huangxing (467), and going to Luyeyuan park and the caves
in the 12th month of the fourth year of Huangxing (470).39 His visits might
have been related to the excavation of caves 13 and 11. It may accordingly
be inferred that the excavation of caves 13 and 11 might have started from
his visit in 467, and lasted until he was murdered in 476.
In addition, we may note that, among the sacred texts translated by Tanyao
in Yungang, the Dajiyi shenzhou jing (Sutra of Great Auspicious Significance
 Phase two – political struggles 89
of Magic Spell), a Tantric Dharaṇi text, translated in the third year of Hep-
ing (462), is especially worthy of emphasis.40 According to Otsuka Nobuo,
Tantric texts in the fifth and mid-sixth centuries are characterized by protec-
tion spells (Dharaṇi) and mudra-mantra-mandala based systems. The texts
are centered upon rituals for Buddha images and “mandalic” representations
of the primary object of devotion. These texts are fundamentally rooted in
Mahayana thought and the path of the Bodhisattva, though we also see a
thorough integration of “Hindu” rituals.41 The purpose of the Dajiyi shen-
zhou jing is to emphasize the worship of the seven Buddhas of the Past, the
three Buddhas, Maitreya, the four celestial guardians, Asura, and the Naga
Kings, and to emphasize that this Dharaṇi sutra is worshipped by the four
celestial guardians. It, more importantly, also emphasizes that one should
“support the imperial rulers, and extend their prosperity infinitely,” “be able
to protect the ruler absolutely,” “pray and make sacrifice to avert calam-
ity, and be filled with virtuous and charitable benevolence,” “endorse the
ruler’s benefits,” and “protect the ruler’s supremacy and attainment of great
luck.”42 It is worth pondering the reason for the particular emphasis placed
on the ability of this Dharaṇi sutra to bring about all sorts of protections and
benefits, including the infinite prolongation of the ruler’s prosperity. Tanyao
was the key figure in the creation of the Yungang cave temples. He was at
the same time in charge of the translations of the sacred texts there. It is
natural that the sacred texts translated by him should have some association
with the subject matter of the iconography in Yungang. The main image of
the Maitreya Bodhisattva seated with ankles crossed in cave 13, the double
dragons on the ceiling, the celestial guardians on the reveals of the entrance
gateway, and the seven Buddhas added later, as well as the Dharma protec-
tors with multiple arms and heads in the ceiling of cave 11, were all, except
for Maitreya, new subjects that emerged in the second phase. They conform
with what the Dajiyi shenzhou jing text emphasizes, reflecting a strong con-
sciousness of Dharma protection and a desire to support the power of the
imperial ruler. During Emperor Xianwen’s era, the power struggle between
the emperor and Empress Dowager Wenming was fierce. Thus, the emphasis
on benefitting and protecting the emperor seems to reflect a political purpose:
to strengthen public opinion in favor of Emperor Xianwen. This explains
why motifs related to Dharma protection, including deities with multiple
arms and heads, appeared in caves 13 and 11, which were supported by
Emperor Xianwen.

4. Dating cave 12
Cave 12 in the Yungang complex is splendidly decorated with dazzling col-
orful carvings of Buddhas, Bodhisattvas, deities, and musicians. Not a single
unused space can be found on the walls in the entire cave. The magnificent
carving is refined and unmatched in any other Buddhist chapels. Mizuno
and Nagahiro pointed out, in their description of the cave, that cave 12
90 Phase two – political struggles
imitates caves 9 and 10 but on a smaller scale.43 For instance, the pillars in
imitation of woodwork, the layout of layered niches in the anteroom, the
structure of the coffered ceiling, and the configuration of the images are
all identical to those in caves 9 and 10. In the rectangular main chamber,
the main images are arranged in the layered niches on the main wall. The
entire upper level is a curtained niche, and the lower niche is placed in the
center. This in the main chamber imitates features of caves 7 and 8. On
either side of the reveals of the window, there is a meditation monk, which
also replicates a feature of caves 7 and 8. Thus, cave 12 has features both of
caves 7 and 8 and caves 9 and 10. The builders were aware of the features
of both pairs of caves, so it is reasonable to place the creation date of cave
12 after them. This corresponds with what was suggested earlier: that cave
12 was “inserted” between caves 13 and 11, and was not part of the system
of caves excavated according to the initial plan. It was constructed after
caves 11 and 13.44 The excavation date of cave 12 should not be far from
the time when the construction of cave 11 was resumed, namely, around
the seventh year of Taihe (483), latest, under emperor Xiaowen’s reign. In
sum, the chronological sequence of the three caves should be cave 13, cave
11, and cave 12.45

Reconstructing a chronological sequence

of the second-phase cave temples
I argued earlier that caves 13 and 11 were constructed under Emperor Xian-
wen, and were the earliest in the second phase, and that cave 12 was a little
bit later than caves 7 and 8 and caves 9 and 10. I also suggested that cave
12 was constructed around the seventh year of Taihe (483) in the Xiaowen
era. Hence the dates of the creation of caves 7 and 8 and caves 9 and 10 will
be the key to the question of the chronological sequence of the second-phase
cave temples. Before engaging in this investigation, let us first review these
caves briefly to fully comprehend them.
There is ample evidence that caves 7 and 8 are paired caves (fig. 4.4).
Both consist of an anteroom and a main chamber, and both have the same
plan and scale. Between the two caves in the anteroom, close to the north
wall, there is an archway leading from one cave to the other. At each end of
the east and west walls of the anterooms are two symmetrically positioned
stūpa-shaped pillars. In the anteroom, both the west wall of cave 8 and the
east wall of cave 7 have pictorial narratives and symmetrical arrangements of
the niches.46 Both the west wall of cave 7, and the reverse side of it, the east
wall of cave 8, are covered by sets of thousand Buddhas in the anterooms.
There is uniformity in the main chamber of each cave: the north wall has
two stories, and the east, west, and south walls are divided into four stories.
More importantly, the distribution of the main images in the north niches
 Phase two – political struggles 91

Fig 4.4 Caves 7 and 8

echo each other. In cave 7, on the north wall (fig. 4.5), there are two large
niches, one on top of the other. The upper niche is trabeated and bears five
images, and the lower pointed-arch niche contains two Buddhas seated side
by side. In the middle of the upper niche is a seated Bodhisattva with ankles
crossed, attended by two seated Buddhas with legs pendant, which are each,
in turn, attended by a pensive image with one foot resting on the other knee.
Corresponding to this arrangement, in the upper niches of the north wall in
cave 8 (fig. 4.6), in the middle is a seated Buddha with legs pendant attended
by two seated Bodhisattvas with crossed ankles, each in turn attended, at the
far end of the niche, by a pensive image with one foot on the other leg. In the
lower niche is a seated Buddha. Also, in between the caves in the anteroom,
at the base of the wall, there is a carving of a gigantic tortoise, which reminds
one of guifu 龜趺. This leads us to believe that there was a stele between the
Fig 4.5 North wall, cave 7

Fig 4.6 North wall, cave 8

 Phase two – political struggles 93
two and that this initially was its base.47 It can be seen that the north walls
of the two caves echo one another. This is true of the east, west, and south
walls as well. The two caves were excavated as individual cave temples, but
it is evident that they were designed and constructed as a pair. The walls of
each cave echo each other, as do the caves themselves.
We have sufficient evidence to believe that caves 9 and 10 were made as
a pair as well (fig. 4.7). The façades of the two have similar structures with
columns, and share a central pillar in the middle between them. There is an
archway in the anterooms close to the north wall connecting the two, as in
caves 7 and 8. The structure of the anteroom and the main chamber in cave
9 corresponds symmetrically to that in cave 10. In the anteroom, the east
wall of cave 9 echoes the west wall of cave 10 more closely than the west
wall of cave 9 itself. In other words, the two caves were designed as a pair
corresponding to each other in structure. In the center of the main chamber,
each cave has a large statue behind which is a vaulted passage. It goes round
through the back of the main image for “practicing the rite of pradaksina
(circumambulation).”48 This is the outstanding characteristic of the two

Fig 4.7 Caves 9 and 10

94 Phase two – political struggles
caves, and they are the only paired caves that both have a passage behind
the main image.49 The main image in cave 9 is of a large seated Buddha with
legs pendant (plate 32); its counterpart in cave 10 is a seated Bodhisattva
with ankles crossed.50 The relationship between the main images of caves
9 and 10 is the same as between those of caves 7 and 8, in which on the
upper level of the north wall a seated Bodhisattva with ankles crossed and
a seated Buddha with legs pendant, respectively, appear. The archaeological
excavation in the 1940s revealed that, “the court in front of the two caves is
common to both and in the same design.”51 Looking from the outside cliff
surface, caves 7 and 8 and caves 9 and 10 are on roughly the same perpen-
dicular plane. These two are also the only caves that extensively illustrate
the narrative stories derived from the Zabaozang jing, translated by Tanyao
in Yungang. This is a rather important phenomenon that deserves scholarly
attention.

1. Dates of caves 7 and 8 and caves 9 and 10 reconsidered

It is generally agreed that caves 7 and 8 were constructed earlier than caves 9
and 10.52 But inferences about the dates of the caves are different. Mizuno
and Nagahiro argued that caves 7 and 8 were constructed during the Hep-
ing and Huangxing (460–71) period, and caves 9 and 10 were constructed
during the Huangxing and Yanxing periods (467–75).53 Su Bai deduced that
caves 7 and 8 were the original Huguo cave temple, and were constructed in
the early Xiaowen period, and that caves 9 and 10 were constructed between
the eighth and 13th year of the Taihe (484–89) and were initially the Chongfu
cave temple commissioned by Qian’er Qingshi (Wang Yu).54
Mizuno and Nagahiro based the dating of the cave temples primarily on
the analysis of iconographic style.55 As I emphasized earlier, iconographic
style can be used as a means of dating and periodization, but cannot be the
only method. As noted earlier, caves 7 and 8 and caves 9 and 10 share the
same perpendicular cliff surface. Although the cave structure of these two
sets of paired caves is very different, the subjects of the main figures on the
north (main) wall of the caves are interlinked. For instance, the main wall
in cave 8 is divided horizontally into upper and lower stories. The upper
level contains a seated Buddha with legs pendant and apart in the middle
flanked by two crowned Bodhisattvas with crossed ankles; the lower-level
niche bears a seated Buddha. The upper-level trabeated niche in cave 7 con-
tains a Bodhisattva seated with ankles crossed wearing a crown in the middle
flanked by two seated Buddhas with legs pendant; the lower-level niche bears
two seated Buddhas (Śākyamuni and Prabutaratna). Within each cave, the
composition of the subjects of the main images on the main wall is that of
Śākyamuni Buddha (and Prabutaratna Buddha) and Maitreya reflecting the
continuation of Buddha Dharma; but the composition of main subjects of
the two paired caves is that of the ascending and descending of Maitreya,
expressing the desire to save all beings. This is the same as the composition
 Phase two – political struggles 95
of the main images in caves 9 and 10. Moreover, in each pair, there is a pas-
sageway connecting the anterooms. This is not seen in any other caves in
the second phase. Therefore, based on the above analysis of the outside wall
elevation of the caves and the subject matter, these two sets of paired caves
can be considered to have been constructed at roughly the same time. There
may not, in fact, be a time difference.56
Dating caves 9 and 10 to 467 (Mizuno and Nagahiro) apparently is some-
what too early because many narrative illustrations derived from the sutra of
Zabaozang jing translated by Tanyao appear in these two caves. For instance,
the narratives of the celestial maiden dedicating lamps (east wall, anteroom),
the conversion of Nanda (east wall, anteroom), the story of the Svastika bird
(plate 33), the woman who hated desire and became a nun (plate 34), and
Mahaprabhasa and the elephant trainer are represented in cave 10, and the
narrative illustrations of Gautama Buddha meeting Rahula (west wall, ante-
room), the celestial maiden dedicating the canopies (south wall, main room),
two brothers becoming monks (south wall, main room), the Nirgrantha-putras
redeemed by Buddha in the Fire Samadhi (west part, south wall, main room),
the eight celestials taught by Buddha (west part, south wall, main room), Hariti
losing her son Pingala (west part, south wall, main room), and the offering of
a meal by Sudatta’s wife (west wall, main room) in cave 9, are all inspired by
the stories in the Zabaozang jing.
According to the Chu sanzang ji ji (Collected Works of Buddhism from All
Dynasties) by Sengyou of the Liang, the Zabaozang jing contains 13 juan, the
Fufazang yinyuan zhuan has six juan, and the fangbian xinlun (Essay on the
core of expedient means) contains two juan. The three texts have 21 juan in
total, and they were translated in the second year of Yanxing (472) by Tan-
yao and Kekaya.57 If the text was translated in 472, the dates of the pictorial
representations derived from it in the caves should not be earlier than that
year. This means that dates of caves 9 and 10 should not be earlier than 472.
Moreover, the emergence of the paired-cave structure is associated closely
with the historical context of the joint name of ersheng (two sages) referring
to Empress Dowager Wenming and Emperor Xiaowen.58 In other words, the
emergence could only be possible after Empress Dowager Wenming resumed
power again in 476. Thus, these two sets of paired caves should not be dated
to the Empress Dowager Wenming or Emperor Xianwen periods.59 It is more
plausible to date caves 7 and 8 to the early Xiaowen era, as suggested by
Su Bai.60
It is also not too far afield to deduce the dates of caves 9 and 10 to be
between the eighth and 13th year of Taihe (484–89). The time frame is about
right. However, Su Bai’s suggestion that caves 9 and 10 should be associated
with the Chongfu cave temple, and were commissioned by Qian’er Qingshi,
needs to be further substantiated.
The Huguo and Chongfu cave temples are mentioned amongst the ten
temples in the Jin stele inscription, as I mentioned previously. Due to lack of
historical documents on Yungang, the Jin stele inscription is still one of the
96 Phase two – political struggles
most valuable literary sources to date, until we find new evidence. According
to the Jin stele:

The Great cave temple of the western capital (Pingcheng) was con-
structed by the Later Wei. Altogether there are ten names: Tongle;
Lingyan; Jingchong; Zhenguo; Huguo; Tiangong; Chongjiao; Tongzi;
Huayan and Doushuai. . . . Emperor Mingyuan started with the Tongle
temple; Emperor Wencheng continued to erect the Lingyan temple; the
Huguo and Tiangong temples were created by Emperor Xiaowen; and
the Chongfu then was completed by Qian’er Qingshi.61

It thus can be known that the Huguo and Tiangong temples were patronized
by Emperor Xiaowen. The Jin stele further notes that:

Within the temples at present there are two remaining traces of inscrip-
tions. One is in the paired Huguo cave(s); it is large but not complete
and is without any date that can be ascertained. The other is in the
Chongjiao (fu) cave(s), and it is small but complete. It roughly says
that the General Pacifying the West, Cavalier Attendant-in-Ordinary,
and Secretary General of the Ministry of Civil Office, the Duke of
Dangchang, Qian’er Qingshi, had them carved. His carved inscription
says: ‘In gratitude for much good fortune, in supplication for future
blessings, to honor the imperial household, and to herald the eternal
age, Qingshi thus built this place to seek prosperity for the nation.’ It
concluded by saying: ‘Established in the 8th year of the Taihe (484)
of the Great Dai (i.e., Northern Wei) and completed in the 13th year
(489). The two Huguo caves opened up by themselves without exter-
nal effort, therefore (when the floor is) struck, it makes the sound of
miracle clocks. When people hear the sound their hearts become tran-
quil. While the stones were being carved, sweet spring water poured
out. It cured one after drinking it. . . . On the east wall of one of the
Huguo cave temples, there is an equestrian representation of a Tuoba
prince.’62

Based on the record, Su Bai inferred that caves 7 and 8 were the Huguo cave
temple, and caves 9 and 10 were the Chongfu cave temple.63 He further
pointed out, “as for the Tiangong temple among the ten temples in Yungang,
we are unable to deduce anything about it now, as the stele does not have
any other records.”64
Some scholars suggested that the Huguo er kan (two Huguo cave temples
护国二龛) in the Jin stele referred to the Shigudong and Hanquandong
(石鼓洞和寒泉洞) from the Ming and Qing dynasties at the eastern end of
the Yungang complex, i.e., the current caves 1 and 2. Even today, inside
the caves, there are pounding sounds when one touches the ground, and
a babbling of running water. Hence, it would be appropriate for caves 1
 Phase two – political struggles 97
and 2 to be the two Huguo cave temples.65 But this inference still needs to
be questioned. First of all, the Huguo cave temple recorded in the Jin stele
has a big Huguo stele. The height of the cliff surface outside caves 1 and
2 is low, insufficient for a big stele, and there are not any corresponding
remains found so far. Secondly, the Huguo er kan were built under Emperor
Xiaowen, in a grand royal construction project, and so should be a paired
cave of great scale. Judging from their scale, caves 1 and 2 appear to be
too small to reflect the scale of an imperial project. In Yungang, the only
caves that show a huge stone stele between the paired caves are caves 7
and 8 and caves 5 and 6. Both have a huge turtle base supporting a stele.
Therefore, considering that we do not have enough evidence, we cannot
repudiate the previous deduction that caves 7 and 8 are the Huguo cave
temple. I believe it is more plausible that caves 7 and 8 are the Huguo
cave temple. But it is still questionable whether or not caves 9 and 10 are
the Chongfu cave temple constructed between the eighth and 13th year
of Taihe under Qian’er Qingshi.66 Although Qian’er Qingshi was favored
privately by Wenming, excelled at exquisite design (of building structures),
and embezzled countless sums of money, the caves he commissioned to
“pray for the state” were, after all, not an imperial project, so the scale
ought to be smaller.67 If the inference that caves 9 and 10 are the Chongfu
cave temple can be established, then they would be juxtaposed in a similar
scale to caves 7 and 8 that were constructed under Emperor Xiaowen.
This apparently goes against the rigidly stratified relationships between
monarch and subject in a feudal society; it goes against the social logic of
that period. Besides, in regard to the chronological sequence, if we date
cave 11 after caves 9 and 10 (484–89), as was suggested, then we cannot
explain the earliest inscription dated to the seventh year of Taihe (483) on
the east wall in cave 11, namely, the date when the construction of cave 11
was resumed. Furthermore, if we accept that caves 11−13 were constructed
after caves 9 and 10, this would mean that there would have been too
many cave temples (caves 11−13, caves 1 and 2, and caves 5 and 6) left to
be completed in a rather short period before the transfer of the capital to
the south in 494. We therefore cannot yet draw the conclusion that caves
9 and 10 are the Chongfu cave temple.
The mention in the Jin stele inscription that the Tiangong cave temple
was created under Xiaowen, on the other hand, provides us with some
clues.68 The Tiangong (Heavenly Palace), here, should refer to the Tushita
Heaven where Maitreya resides in the present until his later rebirth as the
next Buddha. We should consider the Tiangong Temple to be the caves
associated with the main image of Maitreya. Caves 9 and 10 are both colos-
sal image caves. The main image in cave 10 is damaged. A seated Buddha
was sculpted in front of the main image in the Qing dynasty, flanked by
two Bodhisattvas (plate 35). The left Bodhisattva was made into a seated
Bodhisattva with legs pendant (plate 36). Based on the traces of the celestial
scarf of the Bodhisattva in the main wall, and the trabeated niche above the
98 Phase two – political struggles
head of the large image, it is believed that the main image was originally a
large seated Bodhisattva with ankles crossed. The main image in cave 9 is
a seated Buddha with legs pendant, flanked with two standing Bodhisat-
tvas. The main images in these two caves constitute a representation of
Maitreya’s ascent toward, and descent from, the Tushita Heaven.69 At the
same time, the main image in the roof niche on the east wall of the anteroom
in cave 9 is a seated Bodhisattva with ankles crossed (see plate 17), and a
seated Buddha with ankles crossed symmetrically faces it on the west wall
(plate 37). Accordingly, in cave 10, in the same position on the east wall,
is a seated Buddha with ankles crossed (plate 38) and on the west wall is a
seated Bodhisattva with ankles crossed, a reversed replica of the configura-
tion of cave 9 (see plate 16).70 Thus, the primary subjects in the anterooms
of these two caves are also a representation of Maitreya’s ascent toward,
and descent from, the Tushita Heaven, echoing the main images of Mai-
treya in both main chambers. Thus, we have good reason to believe that
caves 9 and 10 are the Tiangong temple. The Maitreya Buddha became
increasingly popular at this time. As Dorothy Wong aptly pointed out, “the
Maitreya faith in China peaked from the late 5th to the early 6th century,
and Buddhist cave-chapels such as Yungang and Longmen fully document
the popularity of Maitreya as a devotional icon.”71
Based on the analysis above, the Huguo and Tiangong temples created
during the Xiaowen reign that are referred to in the Jin stele inscription can
be identified with caves 7 and 8 and caves 9 and 10, respectively. These two
sets of paired cave temples are adjacent to each other. The excavation date
should be after the first year of Chengming (476) when Empress Dowager
Wenming resumed power. As for the date of completion, the collapse of the
stūpa at the west end of cave 10 provides us a significant clue.
The Yungang complex is situated within the sandstone of a Jurassic stra-
tum with soft horizontal layers. The excavation of rock-cut cave temples
usually changes the structure of the stratum. The collapse of cave 20 was
caused both by external and internal forces, which should also have caused
the collapse of the stūpa at the western end of cave 10.72 The approximate
date of the collapse of the stūpa can be deduced based on the image and
niche added onto the body of the stūpa. The niche contains a Buddha image
outside cave 11 (fig. 4.8) that has a round face, and a broad chest, and
resembles the images in caves 9, 10, and 12. It appears to exemplify the
old style before the stylistic change in the tenth year of Taihe (486). It can
be determined that the collapse of the stūpa occurred in the early phase of
the Taihe era. The Weishu records that on a xinyou 辛酉 day in the fourth
month and on a dingmao 丁卯 day in the seventh month in the first year of
Taihe (477) during Xiaowen period, there were earthquakes in the capital.73
There were also two earthquakes in the capital in the tenth year of Taihe
(486). The first earthquake happened in 477 when cave 10 had just started
and construction was still in progress, so it can be disregarded. The cave
should have been finished by the tenth year of Taihe (486). The collapse
 Phase two – political struggles 99

Fig 4.8 Seated Buddha outside cave 11

of the stūpa at the western end of cave 10 might have been related to the
second earthquake in 486. Thus, the completion of cave 10 most probably
occurred before 486.
Based on the analysis above, I argue here that the creation of caves 9 and
10 most probably occurred between the first year of Chengming and the
tenth year of Taihe (476–86). Considering the shared outside cliff elevation,
caves 7 and 8 should have been excavated at roughly the same time. This
is the period when Emperor Xiaowen visited the Wuzhoushan cave temple
frequently, and was also the time when Lady Feng (Empress Dowager Wen-
ming) controlled the state.74 Thus the sponsors of these caves were most
probably Empress Dowager Wenming and Emperor Xiaowen.
Let us now turn to caves 1 and 2 and caves 5 and 6. Caves 1 and 2 are
another set of paired caves belonging to the second phase (fig. 4.9). Both caves
100 Phase two – political struggles

Fig 4.9 Caves 1 and 2

are of medium size, and both contain a central stūpa pillar (plate 39). The two
caves are similar in size and architectural structure, but they do not resemble
one another as closely as do caves 7 and 8 or caves 9 and 10. The eastern
roofed niche of the south wall (plate 40) in cave 1 is occupied by the images
of Vimalakirti and Manjusri, whereas the western niche contains a seated Bud-
dha with a Brahmana. The figures in both roofed niches show Chinese style
robes with step-like pleats and both shoulders covered. Caves 1 and 2 share
many similarities with caves 5 and 6 in style and structure. The central stūpa
pillar resembles that of cave 6, the only difference being that this is a smaller
version. Most notably, two different styles appear in these two caves.
Caves 5 and 6 share a common outside perpendicular cliff surface
(fig. 4.10). Three large pagodas with seven or more stories are carved at
each end, and in the middle of the outside wall of the two caves. Cave 5,
like the five Tanyao imperial caves (16–20), is a colossal image cave, and
contains a large seated Buddha occupying most of the space in the cave
(plate 41). But behind the seated Buddha, as in caves 9 and 10, there is a
passage for the circumambulation ritual. This may suggest the functional
distinctions between cave 5 and the imperial five caves, and something of
the motivation that lay behind the excavation of the caves. The passage in
the rock-cut caves is also found in the caves of Bamiyan and Kizil. Cave 6
is a caitya cave with a large stūpa pillar in the center. It is the richest and
 Phase two – political struggles 101

Fig 4.10 Caves 5 and 6

most complete of all the caves, and also the most dignified and magnifi-
cently carved. In cave 6 alone, there are more than 30 pictorial narrative
stories from Buddha’s birth, to the great departure (plate 42), and to his
first sermon, which, interestingly, was the end of the life of the Buddha in
the illustrations. No other cave in the entire Yungang complex contains so
many scenes from the life of the Buddha and depicts them so exquisitely.
The whole cave is based on a uniform design with the four walls, as well
as the stūpa pillar, divided at the level of the window into upper and lower
stories. The stylistic transformation to Chinese style garments is clearly
reflected in the Buddhist images of cave 6.

2. Dates of caves 1 and 2 and caves 5 and 6 reconsidered

Soper suggested that caves 5 and 6 represent the intense ambition and
aggressiveness, and probably also the resentful malice, of Lady Feng. He
believed cave 6 was dedicated to Lady Feng to “demonstrate her unique
position by the same sort of overwhelming display of power and wealth”
that was beginning to take form in the political realm.”75 Cave 5 was dedi-
cated in memory of her father, Lang.76 Soper also proposed that caves 1 and
102 Phase two – political struggles
2 were “opened by Liu Chang 劉昶 (436–97), a refugee prince of the blood
from the Song court at Nanking, in memory of his imperial parents,” but
did not explain his reasoning.77 Mizuno and Nagahiro suggested that caves
1 and 2 started from the first year of Taihe (477), and caves 5 and 6 were
constructed simultaneously with them.78 But it was also supposed that caves
1 and 2 were actually the Chongfu cave temple constructed between 484
and 489 under Qian’er Qingshi.79 This is more plausible and worth further
exploration.
First of all, these two caves are rectangular in plan with a flat ceiling. The
main room in cave 1 is 23 feet wide, 31 feet deep, and 19 feet high; cave 2
is 25 feet wide, 34 feet deep, and 20 feet high. The scale of each cave is not
huge and appears to fit the status of court officials.
Secondly, caves 1 and 2 are the only paired caitya cave temples in all of
Yungang. Both stūpa-pillars are square and built in imitation of wooden
structures. The pillars have slots suggesting that at some period they
may have had railings, possibly made of wood. The cave 2 stūpa pillar
(plate 43) has three stories with tiled roofs and is surmounted with a
canopy; the cave 1 stūpa has two stories (see plate 39). The two stūpas are
reminiscent of the Huifusi Temple (晖福寺) built under the patronage of
Qian’er Qingshi. According to the records on the Dadai Dangchanggong
Huifusi Temple stele, Wang Qingshi (Qian’er Qingshi) had the two stūpa
pillars built for the “two sages (emperor Xiaowen and Empress Dowager
Wenming).” It took three years to complete them, and they were built in
the 12th year of Taihe (488).80 From this, we can surmise that the Huifusi
Temple was constructed in 488. Caves 1 and 2 were most probably con-
structed at a time close to that of the Huifusi Temple, and with the same
motives.
Thirdly, it is noticeable that caves 1 and 2 demonstrate both old and
new styles existing simultaneously. Cave 2 is badly eroded with very few
well-preserved images, but the tongjian (both shoulders covered) robe of
the main Buddha image, in the second niche on the east wall, is still discern-
able. The seated Buddha on the third level of the stūpa is clad in the tanyou
(right shoulder bare) robe. The Bodhisattva in the center is seated with ankles
crossed and sweeping celestial robe at the back, falling down to the front on
the arms and spreading out. The above images, and the celestial musicians
in the cave, all exemplify the old style in iconography, but the seated Buddha
and the standing Bodhisattva flanking it in the first niche on the south side
of the east wall are both in the new style.
Cave 1 is not well preserved either. The main image is of a seated Bod-
hisattva with ankles crossed and the celestial robe crossed at the front; this
and the attendant Bodhisattvas on the side, the robes of the celestial figures
flowing through the air, and the seated Buddha on the south side of the west
wall are all in the new style. The folds of the dress are similar to those in
caves 5 and 6. The seated Bodhisattva with ankles crossed on the upper level
of the east face of the stūpa is the only one in the old style.
 Phase two – political struggles 103
In brief, cave 2 has more old-style representations whereas cave 1 has more
in the new style. This, then, indicates either that cave 2 was excavated a little
bit earlier than cave 1, or caves 1 and 2 were excavated during the transition
of iconographic styles from old to new. Some scholars have suggested that
the iconographic transformation in style was related to the reform of cloth-
ing promoted by Emperor Xiaowen in the tenth year of Taihe (486).81 This is
a rather complicated issue and will be discussed along with the investigation
of the third-phase caves during the construction of which the sinicized style
was perfected.
On the basis of the three points made above, it is more reasonable to con-
nect caves 1 and 2 with the Chongfusi Cave temple commissioned by Wang
Yu. The only problem is that the outside cliff between the caves collapsed
badly so the “small and complete 小而完” stele remains cannot be observed
anymore.
Caves 5 and 6 are colossal-image and caitya caves respectively. This is the
largest set of paired caves in Yungang. Cave 6 was completed according to
the plan, but cave 5 was not; only the three Buddhas and two attendant Bod-
hisattvas were finished. The rest of the images and niches on the walls were
added later one after another until after the capital was moved southward.
The most outstanding transformation in this set of paired caves is the cloth-
ing style, namely, the sinicization of Buddhist images with baoyi bodai robes.
Nagahiro suggested that caves 5 and 6 were constructed in the first year of
Taihe (477), and finished in the seventh year of Taihe (483) in the Xiaowen
era when the emperor visited the Wuzhoushan cave temples.82 Su Bai pointed
out that Nagahiro’s proposal is not based on documentary sources, and that
“all the large Buddha images in the caves are clad in the baoyi bodai robes,
replacing the old garments. The baoyi bodai robe style is an echo of the
reform in clothing promoted by emperor Xiaowen between the tenth and
19th years of Taihe (486–95). Thus, the completion of cave 6 should not be
too far away in time from the transfer of the capital to Luoyang in 494.”83
Accordingly, caves 5 and 6 are dated to the tenth and 18th years of Taihe
(486–94). It is reasonable to date the caves to the last stage of Yungang right
before the capital was moved.

Summary of the chronology of the second-phase

cave temples
The second-phase cave temples were excavated between 467 and 494, begin-
ning with cave 13. The chronological order of the construction is: cave 13,
cave 11 (Emperor Xianwen, 467–76), caves 7 and 8, caves 9 and 10 (Empress
Dowager Wenming and Emperor Xiaowen, 476–86), cave 12 (patron ?, 483),
caves 1 and 2 (Qian’er Qingshi, 484–89), caves 5 and 6 (Emperor Xiaowen,
486–94), and cave 3, unfinished (Emperor Xiaowen, before 494).
Caves 7 and 8, caves 9 and 10, caves 1 and 2, and caves 5 and 6, four
sets of paired caves, were constructed under the reign of Emperor Xiaowen
104 Phase two – political struggles
(r. 471–99) and Empress Dowager Wenming. A fundamental innovation
took place in cave 6, where the images began to be clad with traditional
Chinese robes with both shoulders covered. Before that, all the second-phase
caves continued the style of the first-phase caves with minor changes.
The conclusion of the chronology of the second phase drawn above is
consistent with the use of the cliff surface in the excavation of the entire
complex, the observation of which is significant in the study of Buddhist cave
art. Essentially, the purpose of the discussion of the chronology lies in recon-
structing the process of the construction of the caves in each phase. From the
discussion, it can be seen that the second-phase cave temples were constructed
in accordance with the Chinese tradition that the west cardinal direction is
superior. Therefore, the cave temples were constructed from west to east; even
with the paired caves, the cave in the west was constructed earlier than the
one in the east. For instance, the main part of cave 13 was finished, but only
the layout of cave 11 was finished; cave 6 was finished according to the plan
conceived; but only the main body of cave 5 was finished; many old styles
appeared in cave 2, while new styles appeared in cave 1. This phenomenon
deserves scholars’ further attention.
The second phase of Yungang amounts to an illustrated history of politics,
faith, and political struggle. The development of architecture, subject mat-
ter, and iconography at this time marked a new era in the short 100-year
history of Pingcheng as the capital. The scale of the caves, the innovation
in architecture, the richness of the subject matter and the complexity of the
iconography, are all unprecedented. The history of the art of Yungang during
this period is marked both by the discovery of new forms and the assimila-
tion of various other elements and styles, both indigenous and foreign.
In addition to dating and periodization of the second-phase caves, another
important issue that has not been settled and is worth exploring is the litur-
gical function of the caves in Yungang. As mentioned previously, with more
recent archaeological excavations above the caves, it is clear now that the
Yungang complex had different functions in each area. The residential cells
and translation centers are all above the caves, but some liturgical ritual such
as image worshipping and circumambulation took place in the rock-cut caves
down below. In the following chapter, I attempt to discuss the liturgical func-
tion of the cave temples in the second phase using cave 12 as a case study.

Notes
1 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu: seireki goseiki ni okeru
Chūgoku hokubu Bukkyō kutsuin no kōkogakuteki chōsa hōkoku, vol. 16, 30.
2 Su Bai, “fenqi,” in Su Bai, Zhongguo shikusi yanjiu (Beijing: Wenwu chubanshe,
1996), 80. James Caswell put forward a new proposal on the second-phase cave
temples suggesting that caves 7 and 8 are the Chongfu cave temple, and caves 9
and 10 are the Huguo cave temple, which is the reverse of what Mizuno and Su
Bai had suggested. Caswell also suggested that there was no construction for 15
years after 465 at Yungang, and that the royal family did not pay any attention
 Phase two – political struggles 105
to the construction of the caves. He further argued that the construction was
resumed after 483; therefore, all the second-phase cave temples were dug into the
cliff between 465 and 495. See James O. Caswell, The Written and Unwritten: A
New History of the Buddhist Caves at Yungang (Vancouver: University of British
Columbia Press, 1988), 78–80.
3 Lidu Yi, The Third-Phase of the Yungang Cave Complex: Its Architectural Struc-
ture, Subject Matter, Composition and Style, Ph.D. Dissertation (University of
Toronto, 2010), 25. I also presented a paper on the issue during the AAS confer-
ence, “Re-examining Caves 11 to 13 at Yungang,” Association for Asian Studies
conference paper, 2013, San Diego.
4 Li Daoyuan, Shuijing Zhu jiaozheng, annotated by Chen Qiaoyi (Beijing: Zhon-
ghua shuju, 2007), 316.
5 Daoxuan, Xu gaoseng zhuan, T50:2060:425c26 and T50:2060:427c23.
6 These holes indicate that there used to be buildings. The presence of a Jin
stele suggests that the building was from the Jin dynasty. See Su Bai, “Jinbei
Jiaozhu,” 69.
7 Above these pillars there seems once to have existed a deeply carved roof prob-
ably with eaves and brackets represented on it but this is no longer visible. See
Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vols. 8 and 9, 95.
8 Mount Sumeru is not seen here, but it is seen in caves 1, 2 and 39.
9 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vols. 8 and 9, 90.
10 Ibid.
11 Mizuno and Nagahiro pointed out that caves 11−13 share a common outside
wall. The front of cave 12 with its pillars is found in the center and those of
caves 11 and 13 on either side, each having a window and entrance arch; they are
roughly speaking symmetrically placed. See Mizuno Seiichi and Nagahiro Toshio,
Unkō sekkutsu, vols. 8 and 9, 89; Su Bai noted that cave 12, with an anteroom
and a main chamber, is in the center; and above the entrance of the main chamber
is the window, while caves 11 and 13 on either side each have a window on the
entrance gateway. The symmetry of this arrangement, with cave 12 as the center,
is evidently intentional. See Su Bai, “fenqi,” 80.
12 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vol. 16, 15. Rei put the
dates of caves 11−13 between 480 to 489. See Yoshimura Rei and Liqiang Bian.
Tianren danshengtu yanjiu: Dongya fojiao meishushi lunwenji (Shanghai: Shang-
hai guji chubanshe, 2009), 404.
13 Su Bai, “Jinbei Jiaozhu,” 52–75.
14 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vol. 7, 67 (Japanese ver-
sion); see remains in front of caves 9 and 10, plates 37, 39, and fig. 85. These
details caused Mizuno and Nagahiro to believe that it used to be a more than
five-story multi-storied stūpa pillar; the only thing is that regrettably we cannot
see any original traces of the west end of the stūpa base; this may be due to the
digging out of cave 11.
15 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vols. 8 and 9, cave 11, 107.
16 Ibid, vol. 7, 83–4.
17 Nagahiro Toshio, “Yungang shiku chuzhongqi de teli dakan,” in Zhonguo shiku
Yungang shiku. Ed. Yungang shiku wenwu baoguansuo (Beijing: Wenwu chuban-
she, 1994), 240–3.
18 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vol. 16, 15.
19 The most prominent change in the second-phase caves is that the representations
of worshippers began to emerge, such as in the lower part of east, west and south
walls in cave 7, and in the west wall of the anteroom, and the south wall of the
main chamber in cave 9.
20 Nagahiro Toshio, “Unko sekkutsu no nazo,” Bukkyo geijutsu, no. 134 (1981):
11–42.
106 Phase two – political struggles
21 Shisong Rinaizi, “Yungang di shiyi ku Taihe qinian yiyi zaoxiang he wuzhoushan
shikusi de bianhua,” in 2005 nian Yungang guoji xueshu yantaohui lunwenji
yanjiujuan. Ed. Yungang shiku Yuanjiuyuan (Beijing: Wenwu chubanshe, 2012),
301–12; and “Yungang zhongqi dongku xinlun–Shamentong Tanyao de diwei
sangshi he hufu gongyangren xiang de chuxian,” Kaogu yu wenwu, 5 (2004):
81–92.
22 Zhao Kunyu, “Yungang di shiyi ku yingzao de jige wenti,” in 2005 nian yungang
guoji xueshu yantaohui lunwenji yanjiujuan. Ed. Yungang shiku yanjiuyuan (Bei-
jing: Wenwu chubanshe, 2006), 313–23.
23 For the discussion of intervention of political affairs by the Empress Dowager
Wenming, see Chapters 3 and 4 in Li Ping, Beiwei Pingcheng Shidai (Beijing:
Shehui kexue wenxian chubanshe, 2000), 124.
24 Li Yanshou, Beishi, “Houfei Liezhuan,” (Beijing: Zhonghua shuju, 1974),
13:495.
25 Wei Shou (506–572), comp., Weishu (Beijing: Zhonghua shuju, 1974), 6:125.
26 Ibid., 6:131.
27 It is recorded in the Treatise on Astronomy in the Weishu that in the fifth year
of Huangxing (471), Emperor Xianwen was forced by Empress Dowager Wen-
ming to abdicate to the prince, Xiaowendi, see the Wei Shou, comp., Weishu,
105:2412.
28 Weishu, 114:3036.
29 Ibid., 7:135.
30 Li Yanshou, Beishi, “Houfei liezhuan,” 13:495. Also, in the Weishu, in the Trea-
tise of Astronomy, it clearly notes that Xianzu suddenly died. It was said that the
empress was responsible for this.
31 Nagahiro Toshio, “Unko sekkutsu no nazo 雲岡石窟の謎,” Bukkyō geijutsu, no.
1 (1981), 11–14.
32 Yoshimura Rei and Liqiang Bian, 462.
33 See Su Bai, “Jinbei Jiaozhu,” 53. Here the Jin stele inscription bears a date: the
eighth year of the Heping reign-period. There were, however, only five years in the
Heping era. It appears, then, that this date resulted from an error in transcription:
the word 元 (first) was mistaken for 八 (eighth). The first year of Heping when
the five Tanyao cave temples were excavated, is the likely date.
34 Some scholars have suggested that the innovation is reflected in the architectural
structure of the caves since each of caves 11−13 display a different architec-
tural layout (stūpa pillar, Buddha Hall and colossal statue). See Zhao Kunyu,
“Yungang shiku yingzao de jige wenti,” in 2005 nian Yungang guoji xueshu
yantaohui lunwenji yanjiujuan. Ed. Yungang shiku Yanjiuyuan (Beijing: Wenwu
chubanshe, 2012), 313–23. Whether or not innovation refers to architectural
change, one thing that is certain is that cave temples were made during the
Xianwen era.
35 Weishu, 114:3037. The translation is that of Leon Hurvitz with minor revision.
36 Ibid.
37 Ibid.
38 Ibid, 114:3037.
39 Ibid., 6:130.
40 Kaiyuan shijiao lu 开元释教录, compiled by Zhi Sheng 智升 in the 18th year of
Kaiyuan (730), records that the Dajiyi shenzhou jing contains 2 or 4 juan (see
fashanglu 法上录), the Jingdu sanmei jing 净度三昧经 has 1 juan, and the Fufa-
zang zhuan has 4 juan, and that Tanyao gathered eminent monks and translated
the Dajiyi shenzhou jing and other sutras at the Northern Terrace (Yungang).
These works can thus come down to later generations, and the significance of
this is infinite, T 55: 2154:539.
 Phase two – political struggles 107
41 Otsuka Nobuo 大塚伸夫, “Shoki mikkyo no zentaizo: Shoki mikkyo no hoga
kara tenkai, kakuritsu he,” in Shoki mikkyo: shiso, shinko, bunka. Eds. Taka-
hashi Hisao 高橋尚夫, Hideaki Kimura, Keiya Noguchi, and Nobuo Ōtsuka
(Tokyo: Shunjusha, 2013), 11–13.
42 Da jiyi shenzhou jing, T21:1335:568–79.
43 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vols. 8 and 9, 102.
44 In my earlier study of caves 11–13, I proposed that they are the earliest of the
second phase caves. Li Yuqun questioned the date of cave 12 and suggested that
it was not constructed as early as cave 13.
45 Alexander Soper suggested that cave 12 was dedicated to Emperor Xiaowen. See
Alexander Soper, “Imperial Cave-Chapels of the Northern Dynasties: Donors,
Beneficiaries, Dates,” Artibus Asiae, 28 (1966): 241–70.
46 One story is identified as the Mugapakkha (soul of graveyard 墓魂) jataka story
on the east wall of cave 7 in the anteroom; see Mizuno Seiichi and Nagahiro
Toshio, Unkō sekkutsu, vol. 4, 100.
47 The stele was roughly 41 feet high, and 9 feet wide; see Mizuno Seiichi and
Nagahiro Toshio, Unkō sekkutsu, vol. 5, 72.
48 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vol. 6, 101.
49 It should be noted that on the north wall of the passage, donors (male and
female), monks, and soaring celestials are carved. The donors are divided into
two groups in the middle, each group looking outward. They are symmetri-
cally arranged on both sides of the tunnel. The worshippers all hold their palms
together at the breast holding stems of lotus flowers.
50 Both main images in caves 9 and 10 were repaired and repainted in the Qing
dynasty. From certain remaining traces, it is believed that the original images
were presumably a seated Buddha with legs pendant and a seated Bodhisattva
with ankles crossed in caves 9 and 10 respectively.
51 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vol. 16, 30.
52 But in recent years, some scholars have suggested that early cave-making at Yun-
gang involved a greater variety of initiatives than has previously been acknowl-
edged and offer perspectives for a broader reconsideration of the beginnings
of cave-making at Yungang, and that caves 7 and 8 contain motifs and design
concepts from Indian ivory carving that are not evident to any such extent in the
other Yungang caves. It has been further suggested that caves 7 and 8 originated
in the earliest formative phase of cave-making at Yungang when foreign models
were considered authoritative, and that they predate the conventionally accepted
periodization of the site, see Katherine R. Tsiang, “Reconsidering Early Buddhist
Cave-Making of the Northern Wei in Terms of Artistic Interactions with Gansu
and the Western Regions,” Journal of Inner Asian Art and Archaeology, no. 4
(2009): 107.
53 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vol. 16, 30.
54 Su Bai, “fenqi,” 79.
55 Soper suggested that caves 7 and 8 were dedicated to Wenchengdi and his con-
sort Lady Li, caves 9 and 10 to Xianwendi (Xianzu) and Lady Li, and cave 12
to Xiaowendi (Gaozu), see Alexander C. Soper, South Chinese Influence on the
Buddhist Art of the Six Dynasties Period, vol. 32 (Stockholm: B. Museums Far
East. Antiq, 1960), 65–6.
56 If we hold that caves 7 and 8 are earlier than caves 9 and 10, and the cliff surface
of caves 9 and 10 is later than cave 11, this would mean that the cliff surface
space for caves 9 and 10 was set aside when caves 7 and 8 were excavated. The
resulting chronological sequence is not persuasive.
57 Chu sanzang ji ji, T55:2145:13b06
58 Su Bai, “Jinbei Jiaozhu,” note 34, 106.
108 Phase two – political struggles
59 Some scholars suggested that caves 7 and 8 might be the earliest caves in the entire
Yungang complex, and that they could have been constructed as early as Shenrui
era. See Tsiang, “Reconsidering Early Buddhist Cave-Making of the Northern
Wei in Terms of Artistic Interactions with Gansu and the Western Regions,”
101–18.
60 Su Bai, “fenqi,” 79.
61 See transcription of the Jin stele, cf. Su Bai, “Jinbei Jiaozhu,” 53–4.
62 Su Bai, “Jinbei Jiaozhu,” 53–4. English translation by James Caswell with revi-
sion. See Caswell, The Written and Unwritten, 30.
63 Su Bai, “Jinbei Jiaozhu,” 60. James Caswell dated caves 7 and 8 “at the critical
juncture of the last decade of Yungang activity before the transfer of the capital to
Luoyang,” and associated them with the Chongfu cave temple, i.e., caves 7 and
8 were constructed between 484 and 489. Harrie A. Vanderstappen, however,
argues “the placement of caves 7 and 8 in the eighties of the 5th century as argued
by Caswell does seem very difficult.” See Caswell, The Written and Unwritten, 81.
Also see Harrie A. Vanderstappen, “Book Review: Written and Unwritten: A New
History of the Buddhist Caves at Yungang,” Ars Orientalis, 19 (1989): 125–7.
64 Su Bai, “Jinbei Jiaozhu,” note 16, 64.
65 Zhang Zhuo, “Dajin xijing wuzhoushan chongxiu dashikusi bei xiaoyi,” in Yun-
gang bainian lunwen ji, vol. 2. Ed. Yungang shiku wenwu yanjiusuo (Beijing:
Wenwu chubanshe, 2005), 350–4. Su Bai mentioned that Hanquandong is cave
2, see Su Bai, “Jinbei Jiaozhu,” note 26, 67.
66 Su Bai, “Jinbei Jiaozhu,” notes 7 and 8, 60–1, Ishimatsu Hinako assigned Chong-
fusi to cave 6, but this proposal lacks evidence, see Ishimatsu Hinako, Beiwei
fojiao zaoxiang shi yanjiu (Beijing: Wenwu chubanshe, 2012), 113–14.
67 According to the biography of Wang Yu in the Weishu, Wang Yu, styled Qing
Shi, had the original name of Ta E 他恶 and was the Pingyi of the Lirun Defense
Command and was a Qiang national, 94:2023–4.
68 Jin Weinuo once considered caves 5 and 6 to be the Tiangong temple among the
ten temples mentioned in the Jin stele inscription, see Su Bai, “Jinbei Jiaozhu,”
104. However, the main image in cave 5 is Śākyamuni Buddha, and the main
image in cave 6 is the seated Buddha with legs pendant, the two constituting the
composition of Śākyamuni Buddha and Maitreya Buddha. It is hard to associate
them with the Tiangong temple.
69 Nagahito and Minuzo believe that there is no evidence to enable a Buddha with
legs pendant to be identified as Maitreya. It would be safer to identify it as
Śākyamuni Buddha since the main Buddha with legs pendant in cave 9 is a com-
plement of the main Bodhisattva with ankles crossed in cave 10, see Nagahiro Sei-
ichi and Minuzo Toshio, Unkō sekkutsu, vols. 8 and 9, English text, 77; Nagahiro
held the same view in his article “The Characteristics of the Paired Caves 9 and
10,” in Zhongguo shiku-Yungang shiku. Ed. Yungang shiku wenwu baoguansuo
(Beijing: Wenwu chubanshe, 1994), 193–207. Li Yumin and Li Jingjie, on the
other hand, believe that it is the Maitreya Buddha, See Li Yumin, “Jintasi shiku
kao,” Gugong xueshu jikan. Ed. Guoli gugong bowuyuan, no. 22 (2004): 2,
33–66; Li Jingjie, “Yungang 9, 10 ku de tuxiang goucheng,” in Yishushi yanjiu.
Ed. Zhongshan daxue yishushi yanjiu zhongxin (Guangzhou: Zhongshan daxue
chubanshe, 10, 2008), 327–59.
70 See Nagahiro Seiichi and Mizuon Toshio, Unkō sekkutsu, vol. 6, Plan 6 and 7,
Elevation of the east and west walls of cave 9; vol. 7, Plan 5 and 6, Elevation of
the east and west walls of cave 10.
71 Dorothy C. Wong, Chinese Steles: Pre-Buddhist and Buddhist Use of a Symbolic
Form (Honolulu: University of Hawaii Press, 2004), 89.
 Phase two – political struggles 109
72 Hang Kan argued that the west wall of cave 20 collapsed soon after the attendant
Buddha was finished, see Hang Kan, “Yungang di 20 ku xibi tanta de shijian yu
Tanyao wuku zuichu de buju sheji,” Wenwu, no. 10 (1994): 55–63.
73 Weishu, “Lingzheng zhi 靈征志,” 114: 2894.
74 The Gaoxu chronicle records that in the fourth year of Taihe (480), eighth month,
the emperor, on a wushen day, visited the Wuzhoushan cave temple, and returned
to the palace on gengxu, see Weishu, 7:149. The emperor visited the Wuzhoushan
cave temple on a xinsi day in the third month in the sixth year of Taihe (482), and
granted clothes to the poor and old; see Weishu, 7:151. The emperor also visited
the Wuzhoushan cave temple in the seventh year of Taihe (483); see Weishu,
7:152. It should be noted that this is the last visit to the Yungang cave temple
that he ever made. In the eighth year of Taihe (484), he visited the Fangshan cave
temple 方山石窟寺; see Weishu, 7:154.
75 Soper, South Chinese Influence, 65–6.
76 Ibid., 64–72. Also see Soper, “Imperial Cave-Chapels of the Northern Dynasties,”
243–4.
77 Soper, South Chinese Influence, 69–72; and “Imperial Cave-Chapels of the
Northern Dynasties,” 242–5.
78 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vol. 16, 3.
79 Yoshimurai Rei believes that the caves should be the paired stūpa caves 1 and
2, see Yoshimurai Rei, “Unko sekkutsu hennenron shukuhaku cho kogakusetsu
hihan,” Kokka, 1140 (1993), 7–29; Li Jingjie noted that the old and new styles
co-existed in caves 1 and 2 in the celestial robes of the Bodhisattvas seated with
ankles crossed and thus argued “this phenomenon echoes well with the reform
of clothing from the tenth year of Taihe (486) to the 18th year of Taihe (494),
therefrom the Chongjiao Temple built by Qianer qingshi between 484 and 489
mentioned in the Jin stele inscription, would possibly, if the argument that it is a
paired cave is very sound, be caves 1 and 2,” see Li Jingjie, “Yungang 9, 10 ku
de tuxiang goucheng,” note 4, 327–59.
80 Tokiwa Daijo and Tadashi Sekino, Shina Bukkyo shiseki hyokai: Buddhist monu-
ments in China (Tokyo: Bukkyo Shiseki Kenkyukai, 1925), Pl. 68 (2), the rubbing
of the stele of Huifusi. The stele is now housed at the Xi’an Beilin Museum.
81 The stylistic change in iconography is a very complicated topic. I will treat this
issue fully in Chapter Six.
82 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vol. 16, 13.
83 Su Bai, “fenqi,” 109.
5 Considering karmic
narratives and liturgical
functions

An important question to be asked here is how Buddhist rock-cut cave tem-

ples function? What kind of liturgical performances might have happened
in the caves? And what can they tell us? When Buddhism was introduced to
China, it not only brought the faith to the country, but religious practices as
well. It brought not only images and temples, but ways of worshipping and
performing, into China. The religious ritual conducted in Yungang, how-
ever, is seldom investigated. The focus is usually on the identification of the
imagery, the iconographic style, or the relationship between cave temples and
meditation. Consequently, the function of cave temples beyond meditation is
often neglected – we merely assume that the purpose of the caves was devo-
tional in a broad and general sense. In recent years, as discussed previously,
with more and more archaeological excavations in and around Yungang, we
have begun to have a better understanding of the components and functions
of the Yungang complex in each area, in particular, the excavations above
the caves. It was made clear by the excavations that there were freestanding
monasteries and residence cells above the rock-cut caves. I have suggested
earlier that monks conducted meditative concentration up in monks’ cells,
not in the cave temples. I also argued that one of the primary functions of
the five Tanyao caves was to serve as family shrine. They were erected as
memorials of the Northern Wei imperial rulers, and the five colossal images
were represented as the present living Tathāgata. Then how did each of the
second phase caves function and how were they used? I have also proposed
that both Xinjiang and Liangzhou exerted an impact on Yungang: Xinjiang
had more impact on art and architecture, whereas Liangzhou, where stūpa
worshipping and meditation were especially emphasized, exerted more influ-
ence on religious beliefs and practices, and on the second-phase caves, which
were either occupied by a large image (caves 5, 7, 8, 9, 10, 12, and 13) or a
stūpa (caves 1, 2, 3, 6, and 11). The function of the latter is more revealing
with a stūpa in the center of the cave. It is traditionally believed that one of
their key functions was for the ritual practice of pradaksina (circumambu-
lation). Caves 5, 9, and 10 have a vaulted passage behind the main image
in the cave, and all the devotees represented along the passage are facing
east, making one believe that the caves also had such a function: devotees
 Considering karmic narratives 111
were to walk around the image. Of course, circumambulation was not the
only activity in each space. Other karmic activities such as worshiping and
making offerings, etc. were common practices as well. But how did other
cave temples without a vaulted passage or a central pillar function? The
architectural layout of the caves does not in itself clearly reveal this. How
did the imagery function in a religious space? How did the imagery and oral
ritual activities complement each other to serve the functional purpose of a
cave sanctuary? How did they perform? How did Buddhist cave architecture
function in Yungang? These issues, which from time to time skipped schol-
ars’ attention, are precisely what this chapter aims to examine. It is possible
to establish a hypothesis on the function of a cave sanctuary if the visual
images depict edifying narratives, the architectural space is suggestive, and
the extant literary sources are revealing. It is significant to explore the func-
tion of a cave-chapel since it not only reveals the local faith, but also brings
to light the popular ritual practices of the time. Besides, this is an important
social art history approach in Buddhist studies. We can never emphasize
enough the significance of the function. With hard evidence and literary
sources, I wish to address below several issues through the examination of
the functions of the second-phase caves using cave 12 as case study. I would
argue that most of the second-phase cave temples had liturgical functions,
which in some cave temples, are easy to be recognized, but in others are
difficult to confirm. I believe that in addition to the ritual practice of cir-
cumambulation, image and/or stūpa worshipping, and making offerings (as
in caves 1, 2, 3, 6, 11, and 5, 9, and 10), caves 7, 8, and 12, none of which
have either a central pillar or a vaulted passage behind the primary Buddha
image, also might have had a liturgical function for the then popular ritual
practices such as changdao and sujiang 俗講 (vernacular oral presentation),
based on the technique by which the narratives unfold and the ubiquity of
musical deities in the antechamber of cave 12. The narrative pictorials in the
cave had a liturgical function, and fostered an emphasis on karmic rebirth
and Buddha’s determination to achieve Enlightenment. I would also suggest
that it is quite possible that a then-popular text, the Tiwei Boli Jing (the
Sūtra of Trapusa and Bhallika, hereafter the TWBLJ), believed to have been
produced in Yungang, might have had an impact on the image-making there.
The text was instrumental to the mass laity in understanding Buddhist teach-
ing, and very often, a fabricated text 疑偽經 was even more important and
widely accepted than the orthodox canon since it was easier to comprehend
for the laity as an expedient means. The significance of the sutra has been,
however, greatly overlooked.

Reading the iconography

Cave 12, as I have suggested earlier, was constructed at the time when cave 11
was reused around 483. It is one of the most ornate Buddha Hall (fodian) cave
temples in Yungang. The symmetric space of the antechamber is completely
112 Considering karmic narratives
covered with karmic stories, and edifying narratives for achieving Enlighten-
ment. The upper level of the east and west walls with large niches face each
other symmetrically, but the bottom level on the east wall was repaired and
now contains one niche with two seated Buddhas side by side (plate 44). The
west wall contains two niches in the lower level, which illustrate one of the
earliest miracles in Buddha’s life: the Conversion of the Three Kasyapas of
Uruvilva (plate 45).1 The south niche shows the meditating Buddha seated in
the fire temple of Uruvilva Kasyapa surrounded by the Kasyapa brothers and
their many brahmana disciples trying to put out the flame of the fire serpent.
The story continues to the north niche where the standing Buddha is holding
his alms bowl in which the fire serpents were subdued. Both north and south
narratives are illustrated in conflated mode in which the protagonist appears
only once. When the story is continued from the south to the north niche,
they together form a continuous narrative mode.2 For a story to be presented
in two separate niches is a rare occurrence in Buddhist narratives in Southern
and Northern Dynasties Buddhist art.
At the top of the east and west walls in the coved area above are the col-
orful musical deities, each with a traditional musical instrument (plate 46).
One rarely sees so many musical deities concentrated in one cave, and given
such prominence to their role. As I shall demonstrate below, these musical
figures, beyond any doubt, were not displayed there randomly. They not
only enliven the cave, giving it the appearance of a concert chamber, but also
have a liturgical performance function. At the top of the east wall, one dwarf
musician is playing a pipa, another a pipe, and between them are narrative
scenes of the Jataka stories: the Dipamkara story to the north and Mara’s
Assault to the south (plate 47). Corresponding to this arrangement, the top
of the west wall presents narrative scenes as well. The north niche presents
the Buddha’s former life as King Asoka3 and the south niche contains a
seated Buddha accompanied by two Brahmanas; one is holding a skull, the
other a bird (?) (plate 48).4
The north wall of the anteroom also presents a symmetrical pattern on
both sides of the entrance and window, and is divided into two main stories
by a well-preserved lotus petal band as is seen on the east and west walls.
The lower story contains two levels (plate 49): the upper level is a trabeated
niche, the lower a roofed niche, inside of which are four seated Buddhas on
each side of the entrance. Together, the eight seated Buddhas represent the
seven Buddhas of the past plus the Maitreya Buddha. The Maitreya Buddha
became increasingly popular at this time.
The top-level illustrations on either side of the entrance window pres-
ent narratives of significant moments in Buddha’s life. The east side niche
portrays the First Sermon of the Buddha in Deer Park, as indicated by the
seated Buddha with the triple wheels (dharma-cakra) and two reclining deer
(plate 50). The west niche contains a seated Buddha holding an alms bowl
in his hands and two celestials holding bowls on either side (plate 51).5 The
narrative evidently depicts the story of the four celestial guardians offering
 Considering karmic narratives 113
four alms bowls to the Buddha who, after refusing the gold and silver bowls,
took the stone ones.6 In the end, the Buddha transformed four bowls into
one, as is shown in the narrative scene.7
The main chamber is rectangular in shape and the walls are divided into
two stories, as is the anteroom. The main wall, which is badly damaged and
was painted during the Qing dynasty (1644–1911), contains two niches
(plate 52). The top niche bears a seated Maitreya Buddha (Mizuno identi-
fies it as Bodhisattva) with legs pendant.8 The arch-pointed lower niche now
contains a seated Buddha. Traces inside the niche indicate that it once con-
tained two Buddhas seated side by side. Apparently, the main wall initially
represented the three Buddhas, the most prominent subject in the Northern
and Southern dynasties.9 On the reveals of the window, two large meditation
bhiksus in dhyāna position (plate 53), seated on either side under a tree with
an alms bag on its branch, are depicted. It should be noted that each of these
images (plate 54) is severed with a vertical cut that removes one kneecap of
the bhiksu and the remainder of the tree in order to accommodate a new
niche beside them. There is a dapuo (打破 breaking space) spatial relation-
ship between the bhiksus and the new niches.
The east wall of the main chamber (plate 55) is covered by two stories
and four niches with three seated Buddhas and a Bodhisattva with ankles
crossed. The lower part of the west wall (plate 56) bears, for some reason,
only a row of 10 seated Buddhas and small niches. Judging from the layout
of cave 13 and given the fact that its east wall is protruding into the west
wall of cave 12 (they are opposite sides of the same wall), it seems that the
common wall was too thin to carve a large niche like those on the east wall
of cave 12. The west wall had to avoid piercing through to the other side.
The east and west sides of the south wall are symmetrical. Both contain
four niches in two vertical stories. The upper niches bear a seated Buddha
with Vitarka mudra (teaching mudra), while the lower niches illustrate nar-
rative scenes of Buddha’s life. Mizuno and Nagahiro suggested that both
niches illustrate the same narrative, in which two rich merchants offer food
to the Buddha.10 It is widely accepted that honor and merit are attributed
to the two merchant brothers, Trapusa and Bhallika. Here, the narrative of
offering food to the Buddha for his first meal right after Enlightenment is
highlighted by the merchants’ caravan, represented by camels and horses
beside the Buddha (plate 57). He is sitting beneath the Tarayana tree with
shining flames around him like the rising sun, while a kneeling deity leads
the merchants and the caravan to the Buddha.11 Therefore Mizuno and
Nagahiro are correct in identifying this narrative as Trapusa and Bhallika
offering food in the east niche.12
The west niche, however, tells a different story, but a related one (plate 58).
The narrative, I would suggest, is intended to represent the 500 merchants
worshipping the Buddha after they were rescued. The 500 merchants are
symbolically implied by five lay people who are led by a bhiksu, the first
figure on the right, who convinced the merchants to worship the Buddha
114 Considering karmic narratives
in order to be saved. Above them, devotees, after being saved, worship the
Buddha with flowers and a boshanlu 博山爐. It would be more reasonable to
display these two coherent narratives side by side rather than repeating the
same legend of Trapusa and Bhallika, as suggested by Mizuno and Nagahiro.
The two narratives on the south wall of the main chamber and the two on the
north wall of the anteroom all center around the Buddha’s life immediately
after he became Enlightened. These illustrations are coherent. I believe these
narratives are related to the then influential sutra, the Sūtra of Trapusa and
Bhallika. One may argue that these narratives are likely based on the canoni-
cal Lalitavistara, which introduces the forest goddess and gives details of
the wheels sinking, harness straps coming loose, and parts coming undone,
as well as the vain efforts of the merchants to get the carts moving, but we
do not see the forest goddess or the sinking wheels here. Nor do we find the
harness straps coming loose, we only find camels and horses on either side of
the Buddha in the east niche symbolically representing the legend of offering
the food to the Buddha. The legend of the 500 followers is only highlighted
by worshipping figures, flowers, and the incense burner in the west niche of
the south wall. In brief, the narratives in cave 12, as well as in caves 7 and 8
are more concerned with karmic actions and achieving Enlightenment, which
conform with the tenet of the Sūtra of Trapusa and Bhallika.

Modes of narratives and viewers’ response

The subjects in the anteroom of cave 12 are richly diversified. The concentra-
tion is primarily on the narratives of the significant moments in Buddha’s life
and his previous existences, in particular, on his karmic acts and persever-
ance for Enlightenment. These edifying narratives were an effective means
of proselytizing among the mass laity. However, most of these narratives
are depicted within one frame, mono-scene, and to the mass laity, they are
illegible. This point leads to the question of the modes of narratives and the
viewers’ response.
Narrative is a universal phenomenon, which, as Roland Barthes remarks,
is international, transhistorical, transcultural; it is simply there, like life
itself.13 It is present in every age, in every place, in every society; it begins
with the very history of mankind. There is not, nor has there ever been, a
people without narrative.14 Narrative is the point at which communication
occurs: there is a giver of the narrative and a receiver of the narrative. In
linguistic communication, je and tu (I and you) are absolutely presupposed
by one another; similarly, there can be no narrative without one who relates
and one who listens (or reads).15 This is the very case in Buddhist narratives:
the stories cannot be comprehended without a narrator. This is precisely
what I attempt to establish in this study. I would argue that the givers of
the narrative in cave 12 could be changdao sengren 唱導僧人 (the clergy
who perform vernacular sutra singing and preaching), jiangjing fashi 講經
法师 (the Dharma masters who give sutra lectures), or the sujiangshi 俗講師
 Considering karmic narratives 115
(vernacular oral presenters), and the receivers would be the devotees, either
clergy or laity. Buddhist narratives in a cave-chapel usually had a didactic
function; they were a form of instruction used to proselytize audiences. The
stories usually begin with the biographies of the historical Buddha, Gautama
Siddhartha, a mixture of history and legend, of truth and fiction. In Buddhist
narratives, we are not only confronted with an extraordinary personality, but
indeed with one who has two faces, depending on whether we look at the
man of daily life or at the figure that has grown within the imagination of
the faithful.16 His previous births are presented to us as a “dramatic legend
in a thousand scenes.” His past existences were innumerable, and he climbed
one by one all the rungs of the ladder of all beings, animals, as well as human
and superhuman, from ant to god.17 However, there is no “beginning” for
Buddha’s existence. He proclaimed:

The transmigration of beings [samsara], O my disciples, has its origin in

the remote past. It is impossible to discover a beginning for the beings
caught in ignorance, enmeshed by the desire for life, wandering from
rebirth to rebirth, moaning and weeping and shedding more tears than
there are drops of water in the great ocean.18

There is no beginning of the biography of the Buddha in cave 12 either. We

do not know where to trace the beginning of his life stories even though the
stories are primarily concerned with his life. There is no beginning here, nor
an end. It is the karma of each of his lives that continues the transmigration,
so there is no need for beginning or end. There is no illustration of his birth
or attainment of Nirvana in the cave. It is his karmic acts and perseverance to
become a Buddha that are emphasized. The precept is that one good deed calls
for another, and everyone has a Buddhahood. Karmic deed and proselytiza-
tion are the focus and aspiration of this religious sanctuary. Thus, there is no
need for beginning or end. Life continues as long as karmic action continues.
In the cave, the narratives are presented in various modes: mono-scene, con-
flated mode, and continuous mode. These are the most common modes to tell
a narrative tale.19 The miracle of the Burning of the Fire Dragon’s Shrine and
the Imprisonment of the Fire Dragon in the Bowl (west wall, lower niches,
anteroom) are uniquely narrated in two different niches in continuous mode.
The continuous mode usually can provide a more detailed story for the view-
ers. Yet the artists in Yungang particularly favored the mono-scene and the
conflated mode.20 The Asoka Jataka (west wall, top niche, north, anteroom)
is a typical mono-scene tale, whereas the Dipamkara Jataka (east wall, top
niche, north, anteroom) and Mara’s Temptation (east wall, top niche, south,
anteroom), as usual, are typical conflated narratives.21 In the conflated mode,
the protagonist appears only once, but in synoptic mode the protagonist
appears repeatedly.22 In Yungang narratives, the protagonist mostly appears
once only. The Syama Jataka in cave 9 is the only story narrated in a succes-
sive series of panels to emphasize the Confucian concept of filial piety.
116 Considering karmic narratives
Let us first focus on the various modes of the narratives in the cave under
examination, so as to see how viewers would respond to them. This is a
rather important issue since it will help us understand the possible ritual
practices and activities in the cave, and ultimately reconstruct the initial func-
tion of this religious space. It is important to explore not only the material
details, but the function of a Buddhist cave, since this will cast a broader
light on the social role of a Buddhist community.
It should be pointed out first that many narrative modes, whether mono-
scenic, conflated, continuous, or synoptic, require a viewer’s knowledge of
the texts in order to grasp the meaning of the narrative. In other words, the
artist relied on a knowing viewer.23 Yet even the knowing viewer must closely
scrutinize the scenes to read the story accurately.24 Only those who previ-
ously knew the story and could read the episodes in their correct sequence,
supplying the missing narrative elements from their memory, could read the
“illegible” narratives.25 They were able to add the missing pieces of the story
to complete the narrative. This inference leads to three conclusions. One: the
cave was probably used exclusively by the Buddhist elite who were familiar
with the canonical sources and were able to decipher the narratives. Two:
some sort of guidance was made available in the cave to allow lay people to
understand the narratives and their implications. Three: the caves thus had
a practical function: to help viewers comprehend the edifying visual images
together with their implicit messages.

Ritual and function in a rock-cut cave sanctuary

First, it is unlikely that cave 12 served only the court and the elite when it
was being used. Yungang, as a court cave, was constructed to make sure that
the Dharma teaching would be carried down eternally without interruption.
One of the key functions of Yungang and other Buddhist temples is to fudao
minsu (instruct the mass laity) and to encourage their religious faith.26 In
addition, the inscriptions found in Yungang indicate that the yiyi community
was actively involved in Yungang even during the second phase.27 Therefore,
it is highly unlikely that the cave was constructed exclusively for the elite or
the court. This leads us to believe that there might have been some kinds of
practices to guide the mass laity in the cave.
The most common explanatory formats of the time were sengren chang-
dao, fashi jiangjing, and sujiang.28 And the most effective and direct expedi-
ent means for guidance and proselytization were the pictorial jataka, nidana,
and avadana stories and parables. These illustrated, edifying narratives of
karma and rebirth were often used in canonical texts. The Buddha used
innumerable expedients and various types of nidanas and avadanas while
he preached the Lotus Sutra. These Indian Buddhist literary genres usually
functioned either in a scriptural or in a performative setting as explanatory
devices, ideal materials for changdao or sujiang. As narratives, they pro-
vide the auditor with a variety of situations modeling karmic acts and their
 Considering karmic narratives 117
consequences.29 In addition, they function as ethical guides; above all, they
are a stimulus to action.30 Stanley Abe suggested that the narrative paintings
in cave 254 in Dunhuang were possibly used to illustrate an oral presen-
tation.31 Jiangjing lectures and visual images (either painting or sculpture)
would complement one another. Both carried the responsibility of spreading
Buddhism among all classes of people.
During the Six Dynasties period and early Tang, the most popular pros-
elytizers were “singing and leading preachers (changdaoshi)” who recited
narrative tales in conjunction with sutra exegesis while preaching to diverse
audiences. In his Gaoseng zhuan, Hui Jiao devoted an entire chapter to a
description of changdao.32 This is a demonstration of the prevalence and
significance of changdao during the Northern and Southern Dynasties.
He started the chapter by explaining that changdao was used to promote
Dharma teaching by singing in order to instruct the mass laity.33 According
to Hui Jiao:

This type of preaching came into practice in former times when Bud-
dha’s law was first being transmitted. At maigre feasts there were assem-
blies that only recited Buddha’s name and scriptures. When midnight
came, [the congregation] would be exhausted. In order to revive [every-
one], elder monks would be invited to ascend the dais and preach the
law.34 These monks would ‘either narrate miscellaneous nidanas or cite
avadanas.’35

The magically convincing ability of the changdaoshi was awe-inspiring. Hui

Jiao describes their performances as follows:

The cadence of their voices rises and falls, their arguments are strong,
their words without end. When they talk of death [impermanence], it
makes everyone’s heart tremble with fear. When they talk of hell, it
makes everyone weep. When they inquire after past causes, it is as if
everyone sees their past deeds. When they investigate fruits of present
[deeds], each person’s future rewards are demonstrated. When they talk
about joy, everyone is happy. When they narrate sorrowful things, then
everyone weeps and is sad.36

含吐抑揚辯出不窮言應無盡. 談無常則令心形戰慄, 語地獄則使怖淚

交零, 徵昔因則如見往業, 覈當果則已示來報. 談怡樂則情抱暢悅, 敘哀
慼則灑淚含酸.各各彈指人人唱佛, 爰及中宵後夜鍾漏將罷.

Hui Jiao further explains that these monks ought to possess four valuable
skills to be good orators for changdao: a sonorous voice 聲 (sheng), the
ability to argue 辯 (bian), talent 才 (cai), and broad knowledge 博 (bo).37
He also noted that these preachers were not necessarily experts in canoni-
cal scriptures.38 Rather, they were storytellers who used narrative tales to
118 Considering karmic narratives
educate the laity, and they reached as wide an audience as possible, adjust-
ing their preaching to diverse groups, popular and elite, lay and religious.39
The stories narrated by changdao specialists were part of a larger liturgical
program. Both types of performances of karmic narratives complemented the
more formal use of sutras through their function as didactic entertainment
designed to elicit interest in the general populace. They also share the same
temporal placement as the evening or night service, held either before or after
the lecture or recitation of scripture. Monks specializing in the enactment of
these karmic stories participated in a long tradition of formalized storytelling
with its roots on the Indian subcontinent.40
Cave 12 provides an ideal venue for changdao activities, preaching, and
singing with a variety of lively musicians from all regions and cultures play-
ing all sorts of pipes and lutes.41 The musical divinities are by no means
represented randomly without any symbolic purpose in a functional religious
space. They had never been represented in such a high concentration in one
cave as they were here. Besides, these musical deities are represented in a
very peculiar way: they are right beside the narratives as if each story is told
in accompaniment with musical composition. Cave 12 would be perfect for
such performances judging from the architectural structure, the subjects,
and visual pictorials.
In addition to changdao, I would propose, that other common practices
such as jiangjing and sujiang were performed in the cave, in particular,
jiangjing, which prevailed in Pingcheng, the capital. Jiangjing here is con-
ducted by way of changdao. It was a common practice of the time. Tandu
曇度, the famous jiangjing fashi, began to lecture widely and actively as
soon as he arrived in Pingcheng, and had thousands of auditors.42 Many
Xuzhou eminent monks are known to have given lectures in Pingcheng and
they played a significant role in the prominence of jiangjing activities there
during the Taihe era of Emperor Xiaowen.43 The Guang Hongming ji also
tells us that Xuzhou monks conducted changdao in lecturing on the Noble
Truths of Buddhist teaching.44 Many other Buddhist caves, such as those in
Kucha in the Northern dynasties, have lecture halls (jiangjingtang 講經堂) as
well.45 Ding Mingyi and others suggested that the square caves in Kizil are
all jiangjingtang.46 Giuseppe Vignato pointed out that jiangjing lecture-hall
caves are typical in Kucha and many ruins of the caves indicate the same
feature.47 Clearly jiangjing in caves was a common practice in the Northern
dynasties. According to an important Vinaya text, the Genben shuo yiqie
youbu pinaiye zashi (Mulasarvastiva davinaya), the purpose of narrative
illustrations of the Buddha’s life and Jataka stories in lecture halls was to
spread Dharma teaching and proselytize, to make sure the Dharma teaching
was carried down eternally.48 This further proves that the narratives had a
didactic function and were used in a religious space as an expedient means
to proselytize the mass laity. The narratives were useful visual aids for the
oral practices of changdao and jiangjing. Rock-cut caves are essentially the
same as Buddhist monasteries, and can be used as lecture halls.
 Considering karmic narratives 119
A sutra lecture (jiangjing) was a religious service given by a fashi (Dharma
master) to monks and laymen.49 It was then explained in detail by disciples.
Dao’an 道安, for example, often repeated the lectures given by the mas-
ter, Fotucheng 佛圖澄, and then gave lectures himself afterwards. He was
called a qidaoren 漆道人 (black monk), and his lectures sometimes shocked
audiences in the Buddhist community.50 The rituals of sutra lecturing were
usually supported by the imperial house and welcomed by the common
people.51 Many literary sources mentioned such religious activities. For
example, Yao He (early 9th century) wrote that when a sutra lecture was
held in a town, the wine shops and markets emptied and all the fishing
boats disappeared from nearby lakes.52 The abundant records of sutra lec-
tures attest to their great popularity even in the Tang. Sutra lecturing was
as popular as sutra painting in the Tang.53 The goal was to proselytize and
make Buddhist teaching comprehensible to different audiences, using either
visual narratives or performative explanatory activities, as the eminent
monk Hui Jiao suggested:

In propagating Buddhist doctrines, it is not necessary to speak incisively

of impermanence and to discourse trenchantly on repentance. For rul-
ers and elders it is necessary to cite popular allusions and interweave
set phrases. For the numerous masses of commoners, it is necessary to
point to events and construct shapes, to speak directly of what is seen
and heard. For mountain folk and desert dwellers, it is necessary to use
neighborly words and to reproach them with terms that arouse feelings
of guilt.54

Evidently even during the time of Huijiao and Yungang, making Buddhist
teaching comprehensible to people from all walks of life was just as impor-
tant as it was to be in the Tang. To have sutra lectures and oral presenta-
tions to ensure that the precepts are fully understood by everyone was the
primary goal of any religious space including Yungang, especially the cave
in question. Therefore, it is not unreasonable to believe that Yungang, as
an active Buddhist center, and in the capital, had fashi to give sutra lectures
and oral presentations. Fashi jiangjing and sengren changdao were prevail-
ing practices of the time. Based on the iconography, subject matter, and the
architectural space, I believe that it is highly likely that both changdao and
jiangjing oral presentations were used as an expedient means to support and
guide the laity (fudao minsu) and to proselytize diverse audiences.55
Another prevailing expedient technique used to fudao minsu was sujiang,
popular lectures for lay people. Sujiang was an articulated program of ritual
texts and actions performed by monks for the well-being of both lay people
and clergy. Victor Mair argues that sujiang actually started as early as the
Six Dynasties period.56 Zanning in the Seng Shi Lüe 僧史略 also proved an
earlier date than Tang for sujiang and recorded that sutra lectures given by
nuns began with a certain nun named Daoxin (道馨) in the third year of
120 Considering karmic narratives
Taihe (368) during the Eastern Jin (317– 420) in the Eastern Temple 東寺 in
Luoyang.57 Scholar Xiang Da proposed that sujiang is similar to changdao
and is actually derived from the latter.58 They both are intended to prosely-
tize the laity and make the profound Buddha Dharma more comprehensible.
Changdao focused on explaining the Dharma theories, whereas sujiang was
centered on canonical scriptures 唱道以說理為主, 而俗講則根本經文.59 Suji-
ang was often accompanied by music. Two Dunhuang manuscripts (P. 3849
and S. 4417) provide detailed sequences of the steps necessary to perform a
popular lecture:

The Dharma master would explain the sutra, the cantor would sing
the title and text passages of the sutra (it starts with the introductory
phase ‘please sing (chang jiang lai)’ and would direct questions (nan) to
be answered (tong) by the Dharma master. The typical format of sutra
lecture texts is an initial quote from the sutra in question followed by a
passage of explanatory prose, reiterated as explanatory verse, with the
pattern then repeated throughout the work. Contents range from exposi-
tion of such complexities as the relationship between the transcendental
Buddha (fashen) and the Buddha of transformations (huasheng) to such
basic tenets for laymen as accepting the Three Refuges (sangui) and tak-
ing the Five Vows (wujie).60

These liturgies typically centered on the exegesis of a particular sutra, such

as the Lotus Sutra, Vimalakirti Sutra, and Amitabha Sutra. These lectures
would begin in the morning and take place on festival days, or at maigre
feasts (zhai), or by imperial command.61 The Lotus Sutra and Vimalakirti
Sutra were influential in image-making in Yungang, as we see the visual rep-
resentations of Śākyamuni Buddha and Prabhutaratna seated side by side
and the legendary debate between Vimalakirti and Manjusri (the visual sym-
bols of the Lotus Sutra and the Vimalakirti Sutra respectively) throughout
Yungang. They were, however, not the only influential sacred texts there.
The sutras translated by Tanyao, the chief in command of śramanas in Yun-
gang, who gathered hundreds of monks to translate the sutras there, played
a significant role in image-making as well.62 Scholars have not grasped,
and consequently have failed to sufficiently emphasize, the impact of the
sutras translated or produced in Yungang on the visual representations. The
narratives represented in caves 9 and 10 reflect strongly the close associa-
tion between the images and the sutra translated by Tanyao and others,
the Zabaozang jing, a collection of the karmic narratives, which was, as
mentioned previously, compiled and translated in Yungang in 472.63 The
narrative stories on the south wall of the main chamber in cave 9 visually
highlight the tenet of the sutra. As Mizuno aptly noted, “they serve to illus-
trate the close connection between the construction of the caves and the
translation of the sutras by Tanyao.”64 These stories are represented only
in caves 9 and 10, which were constructed, as we discussed in the previous
 Considering karmic narratives 121
chapter, between 484 and 489. The fact that the narrative stories from the
Zabaozang jing are only represented in caves 9 and 10 can further prove
that the sutras translated in Yungang had an impact on image-making there.
These stories cannot be found in any other caves earlier than 472, the year
the sutra was translated.
Another source that deserves scholarly attention is the aforementioned
fabricated sutra, the Tiwei Boli jing. This text, drawn from the Lalitavis-
tara text, I would like to suggest, possibly had an impact on the image-
making in Yungang, especially in cave 12 since what cave 12 presents us
with, both in visual images and subject matter, conforms to the tenet of the
text. Both emphasize karmic acts and rebirth, and both advocate persever-
ance to Enlightenment. Several sources suggest that the sutra was fabricated
in Yungang during the series of reforms carried out by Tanyao who was still
recovering from the Buddhist persecution.65 Daoxuan notes that the sutra
was instrumental for common lay people in guiding their faith and karmic
practice.66 It was more easily accepted and comprehended by the mass laity.
It would be more reasonable to use a text that was influential and produced
in Yungang while making images as a visual aid. However, scholars have
not explored the possible connections between image-making in a rock-cut
sanctuary space and the fabricated sutras, which sometimes are more influ-
ential than the orthodox canons in educating the mass laity. We should not
neglect the significance of the fabricated sutras. This leads to the next topic:
the TWBLJ.

Folk faith, karmic practice, and the Tiwei Boli jing

The sutra TWBLJ is primarily concerned with karmic acts and rebirth, and it
was produced for lay people.67 The emphasis on karmic action is prominent
in texts that were written expressly for common lay Buddhists such as the
TWBLJ.68 According to Kenneth Chen, the TWBLJ is the earliest expression
of the Plebeian Buddhist tradition and represents the first known attempt to
align the pancasila (the five precepts in Buddhism) with the five Confucian
permanents (virtues) and the five elements theory in Daoism. The sutra there-
fore signaled a Buddha-Daoist synthesis in early Chinese Buddhist history.69
The Buddha said: “those who want good fortune in the next life, read this
sutra; those who want a long life in the next rebirth . . . who want to follow
the way of Buddha . . . acquire wisdom . . . and attain nirvana, read this
sutra.”70 Plebeian Buddhism is also known as ren tian jiao (human-heaven
teaching) and its social organization is known as yiyi or yihui. The Buddha
was said to have taught the ren tian jiao to the two merchant brothers Tiwei
and Boli during the first 21 days after his Enlightenment:

The first period is when the Buddha, soon after his enlightenment,
preached to the two merchants, Tiwei (Trapusa) and Boli (Bhallika),
along with their five hundred followers, the five precepts (pancasila) and
122 Considering karmic narratives
the ten good deeds, as well as the law of karmic retribution in samsara.
Buddha taught the wu jie jing, i.e., the TWBLJ. This was because the
recipients of his teaching did not have as yet the necessary roots condu-
cive to the attainment of the trans-samsara path. 71

This passage provides two visual messages for us: one, the Buddha
preached the TWBLJ to the two brother merchants, the protagonists of
karmic acts in the text; and two, their 500 followers were in the audi-
ence. The tales of the sutra conform to the visual narratives of cave 12
nicely. The visual and literary stories coincide. One is thus led to believe
that there is possibly a link between the text and the visual images. The
karmic narratives are all highlighted in prominent positions and portrayed
in large niches. The two brother merchants and the 500 followers are
presented on the south wall of the main chamber in large niches. It is
hard to believe that the representative karmic tales shown in the cave are
by random choice or by accident. The overall arrangement is too coher-
ent to be random. The tales must have been selected to serve the purpose
of the cave sanctuary as a functional religious space. Besides, the tales
of Trapusa and Bhallika and 500 merchants are not the most favored
subjects that appear frequently in Buddhist cave sanctuaries. But here in
cave 12 the stories are stressed in large niches and prominent positions.
There must be a reason that these narratives are emphasized. I would
suggest the possible association between the influential fabricated sutra
TWBLJ and image-making. Another factor that must be considered is
that the text was produced in Yungang and it was very popular at the
time.72 It was particularly instrumental in teaching to the lay people.73 It
is therefore not unreasonable to infer that the sutra produced in Yungang
had an impact on the image-making there. After all, the tenet of the text
on karmic acts and good deeds, and on proselytization is precisely what
the cave advocates. Both emphasized the determination and perseverance
of the Buddha before his Enlightenment, and both advocate good deeds
and their consequences. The philosophy of Buddha’s preaching to the two
brothers and the 500 followers in the sutra is essentially the same: the per-
formance of good deeds for the sake of better rebirth as humans (ren) or
in heaven (tian) as gods. The TWBLJ was an attempt, initiated by Tanyao,
to educate a significant number of common people attached to Buddhist
institutions as a result of the Sangha and Buddha households.74 Cave 12
would provide an ideal venue to serve this purpose. The diversified and yet
unified subject matter and visual representations in the cave all function
together to educate, inspire, and proselytize the common people.
Tanyao, the leader of the śramanas, who lamented over the destruction
caused by Buddhist persecution (446–52) and rejoiced over the revival and
prosperity, gathered a group of monks in the Beitai Grotto Cave (Yungang)
to translate various sutras so that the Dharma would be transmitted to later
generations and continue unbroken.
 Considering karmic narratives 123
At that time, a certain monk named Tanjing, seeing that the older trans-
lations of various sutras were lost in the fire of the Buddhist persecution
and that there was nothing to instruct the common people with during
the days of the Buddhist revival, produced the TWBLJ in two scrolls.
The purpose was to lead the people to higher understanding, but there
was much that was false in its teachings. . . . In the old catalogue of
sutras, there was a TWBLJ in one scroll, the teaching of which is in keep-
ing with other sutras. But Tanjing mixed stone with gold and therefore
the work is said to be a fabrication. Its whereabouts are unknown. In the
early years of the Kaihuang period (581–600), the people around Guan-
nei (the Wei River) still frequently practiced the Ti-wei cult. Members
of this yiyi would put on robes and hold bowls every fortnight, keeping
the precepts and watching over each other. The cult is widespread with
many followers.75

Daoxuan tells us several things here. One, the sutra was produced in Yun-
gang. Two, Tanjing knew Tanyao. Three, the text was written for lay people.
Four, the ti-wei cult was popular and had many followers. Thus, it is clear that
the sutra was known to the Yungang community, and that the sutra was sig-
nificant in instructing the yiyi common people. Many yiyi community people
were actively involved in the activities in Yungang, as we observe in inscrip-
tions. An inscription in cave 11 explicitly indicates that a group of about 54
yiyi people were dedicated to a certain niche.76 All these reasons made us
believe that it is highly likely that the sutra inspired the artists of Yungang
who were involved in the construction of cave 12. It would be odd not to use
this instrumental text, which was produced in Yungang, and was popular, as
a source to depict visual narratives, or not to use pictorial stories and parables
as an aid to educate the mass community. It has been suggested that Buddhist
literary texts indeed exerted an influence on the image-making of Yungang. Li
Jingjie, for example, proposed that the Lotus Sutra influenced all the subject
matter in Yungang.77 It was, however, argued that the image-making at Yun-
gang was heavily influenced by sutras such as the Dajiyi shenzhou jing, the
Zabaozang jing and others translated or fabricated in Yungang.78
It is hard to believe that a popular and influential sutra, produced in Yun-
gang, would not be used, but other sutras would, as a reference for artistic
representation. Furthermore, the sutra was produced to instruct the common
lay people for a “higher understanding.” Pictorial stories would be a great
way as an “innumerable expedient means” to achieve the goal. Therefore, I
would propose that it is entirely possible that the sutra was instrumental for
image-making in Yungang, in particular, in the cave in question. It enriched the
subject matter of the cave, and at the same time, visual presentations make the
text more comprehensible. The narratives on the north wall of the anteroom
and south wall of the main chamber vividly depict what the sutra narrates and
philosophically promotes. The two merchants’ stories are illustrated in detail
and are shown in prominent positions. The emphasis on the story is obvious.
124 Considering karmic narratives
The connection between the text and image-making is convincing. Likely, both
the TWBLJ and the Zabanzang Jing were used (cave 12 and caves 9 and
10, respectively) in preaching and singing activities, as Mizuno aptly argued,
“They serve to illustrate the close connection between the construction of the
caves and the translation of the sutras by Tanyao.”79 Mizuno further sug-
gested that Tanyao translated the sutras in Yungang itself and the caves must
therefore have been excavated under his direct and active leadership, and the
forged scripture, the TWBLJ, was intended to induce men to practice his faith.
Tanyao’s simple approach to the faith as reflected in the TWBLJ and Jingtu
Jing easily gained a hold over the population.80 This is so true regarding the
forged sutra TWBLJ that I would argue the sutra was a major visual source
for the subject matter of cave 12, and likely caves 7 and 8.
As I have suggested, much of the subject matter in cave 12, the Dipamkara
Jataka, the Asoka Jataka, and the four divinities offering bowls to the Bud-
dha, just to name a few, are all related to karmic rebirth, the attainment of
merit, and good deeds. All the narrative illustrations in the cave are centered
on karmic concern, the core teaching of Buddhist philosophy and the TWBLJ.
Karmic concern is one of the most prominent characteristics of Liang-
zhou Buddhism, which had a profound impact on Northern Wei Buddhism
and practices. The work of Dharmaksema and Liangzhou Buddhism as a
whole shaped many important characteristics of Buddhist practice in north-
ern China during the 5th century.81 The Weishu notes that when Liangzhou
was pacified in 359 by the Northern Wei, people from the capital of the
Northern Liang migrated to Pingcheng. Sramanas and Buddhist practices
both went east, and both images and Buddhist Law prospered more and
more.82 Therefore, as has been suggested, Buddhist practices found in the
area of Pingcheng after the relocation must have been transplanted from
Liangzhou.83 The focus of cave 12 is karmic actions, as can be said of other
second-phase caves. The function was on image worshipping, circumambu-
lation, and other karmic actions and practices.
The Mingseng Zhuan 名僧傳 provides an account of a Maitreya-related
meditative practice of a Dunhuang monk, Daofa:

Whenever Daofa begged for food or was served at a maigre feast, he

would use part of his own portion to feed the insects and birds. He used
to sit in meditation and chant, day and night without a halt. At night
he would strip off his clothes before the image of Maitreya, to give the
mosquitoes something to feed on. After several years had gone by in this
way, he saw the Maitreya emit lights of various colors.84

This account provides us with a vivid picture of the significant features of

Liangzhou Buddhism and practices in the 5th century: the utilization of
images to assist practices, the central role of Maitreya, and most importantly,
the emphasis on the results of karmic actions.85 It should be noted that the
key characteristics of Liangzhou Buddhism are consistent with those of cave
 Considering karmic narratives 125
12 and of the TWBLJ. This association indicates that both Liangzhou Bud-
dhism and the TWBLJ played a role while the cave in question was being
conceived.
Another characteristic of Liangzhou Buddhism is the circumambulation
practice. The practice is also emphasized in the TWBLJ. However, the sutra
does not mention the stūpa, around which the circumambulation ritual is
usually practiced; rather, the Buddha gives instructions on the ritual of cir-
cumambulating the statue and the benefit of the ritual practice:

Circumambulation will produce five rewards: the person will gain a

good complexion in his next life; his voice will be fine; he may be reborn
in heaven; he may be reborn into families of lords and nobles; he may
gain nirvana. What causes the good complexion? It is his rejoicing in
seeing the Buddha. What causes the fine voice? It is his reciting the sutra
while circumambulating. What causes the rebirth in heaven? It is his
will and intentional faithfulness to the precepts while circumambulating.
What causes rebirth into aristocratic families? It is his act of honoring
the feet of Buddha with bowed head and face. What causes the attain-
ment of nirvana? It is the accumulation of good karma.86

The Buddha further explained to the elder brother:

In circumambulating the statue, three acts should be present. In raising

your foot, reflect upon the act of raising your foot. In putting down
your foot, be mindful of the act of putting down your foot. Thirdly, you
should not look left and right or spit within the temple compound.87

Evidently, this emphasizes the effort of accumulating good karma and the
ritual of circumambulating the statue. Caves 5, 9, and 10 in the second phase
are image caves and all of them have a vaulted passage behind the primary
Buddha image in the cave for circumambulation. Circumambulating images
to accumulate good karma was a common practice of the time.

Concluding remarks
Cave 12 is richly diversified both in art décor and subject matter and every
single space on the walls is illustrated. One can easily feel the splendor of this
religious space in the 5th century: the ingenuity of the multicolored images,
the refined workmanship and the charming narratives of karma and rebirth.
The carving is so fine and meticulous, and the narratives are so appealing
that a devotee would be drawn into the devotional space immediately and
focus on religious aspiration.
As a Buddhist cave sanctuary, cave 12 is an example of the perfect
integration of folk faith, practice, and function in a religious space. As
I have argued earlier, sutra lectures by Dharma masters (fashi jiangjing),
126 Considering karmic narratives
vernacular sutra singing and preaching by clergy (sengren changdao), and
vernacular oral presentation (sujiang) were highly likely to have been prac-
ticed in the cave based on the arguments presented above. It is through
the visual narratives and these ritual practices, as guiding aids, that the
key concept of samsara philosophy was comprehended by the mass laity.
Changdao and sujiang played significant roles in proselytization and edu-
cating lay people. The basic teaching of the five precepts (not to kill, steal,
drink, lie, or be adulterous) and the ten virtues as a code of conduct in
China has functioned as the main guidance to daily conduct for lay Bud-
dhists, and the tenet of the teaching was promoted through these ritual
practices mentioned above. It is important to emphasize the roles of oral
rituals in a religious space.
The performance of good deeds will guarantee a better rebirth to the devo-
tee. The Pure Land faith or Chan wisdom are not seen leading to Enlighten-
ment or Buddhahood. Only worship and acts of donation counted.88 The
strong belief in karmic acts and good deeds in the yiyi community led to a
zeal for image-making and donation during the 5th century. It is the karmic
optimism that inspired the common people to become devout and convert.
The attainment of merit through hard work and dedication is promised to
promote the rebirth as ren in tian. The rentian cult at that time was power-
ful and popular, and it was still popular during the Sui dynasty.89 Daoxuan
described the yiyi members as such: “members would put on robes and hold
bowls every fortnight, keeping the precepts and watching each other.”90 It
has been suggested that the Chinese peasants in the 6th century, at the urg-
ing of the Sila-conscious Liangzhou monks, were literally putting on coarse
clothing and holding bowls in imitation of the monks’ uposatha and confess-
ing their sins against the pancasila.91 Both Shixian and Tanyao were from
Liangzhou and both were leaders of śramanas who “presumably adopted
the ten full silas.”
In summation, in this study I have attempted to bring scholars’ close
attention to a unique Buddhist cave-chapel in the Yungang cave complex
and emphasized the functions served by cave 12. I have argued that, as a
religious space, cave 12 had a liturgical function for jiangjing, changdao,
and sujiang, prevalent then in order to achieve a higher understanding of
Buddhist teaching by the mass laity, and of the implications of the nar-
ratives presented in various modes. The narrative stories were the most
convincing and direct means for proselytization and conversion. The aim
of this study is to draw scholars’ attention to these issues and open another
avenue to reconstruct and re-examine Buddhist caves in general. I wish to
deepen our understanding of the function of the caves in the second phase,
and draw the attention of scholars to the then instrumental sutra produced
in Yungang, the TWBLJ. The visual representations, oral presentations,
and architecture in a Buddhist sanctuary such as caves 5, 9, 10, and 12
of Yungang function as an integral entity to achieve the goal of spreading
Buddha Dharma.
 Considering karmic narratives 127
Notes
1 According to the doctorial texts, the disciples of the three Kasyapas of Uruvilva
poured water on the fire. The figures are shown in the usual forms of Brahmana
with an emaciated body dressed in only a loin-cloth, hair dressed in a high knot
and with a pointed beard.
2 For the modes of Buddhist visual narratives, see Dehejia Vidya. “On Modes of
Visual Narratives,” in Dehejia Vidya, Discourse in Early Buddhist Art: Visual
Narratives of India (New Delhi: Munshiram Manoharlal Publishers, 1997), 6.
3 See the Sutra of the Wise and the Foolish, trans. Hui Jue et al. in 445. T4:
202.368c.
4 The same illustration is seen in cave 9 as well. The identification of these two
figures has been problematic and is still the subject of scholarly debate. Wang
Huimin argues that the figure with a skull represents Mrgasirsa, while the other
with a bird represents Nirgrantha-putra rather than Vasistha, as is usually sug-
gested, see Wang Huimin, “Identifying the Heretic Figure Who Has Holding a
Bird,” Dunhuang yanjiu, 1/119 (2010): 1. About the subject, also see Li Jingjie,
“Guanyu Yungang dijiu, dishi ku de tuxiang goucheng,” in Yishushi yanjiu. Ed.
Zhongshan daxue yishushi yanjiu zhongxin, vol. 10 (Guangzhou: Zhongshan
daxue chubanshe, 2008), 330.
5 The tradition has it that the Buddha remained four weeks at the Bodhi-tree, the
two traveling from Ukkula and being warned by a deity, approached Buddha
and offered him the first meal: rice and honey cakes. The Buddha questioned,
“Tathagatas do not accept food in their hands. With what shall I accept the rice
and honey cakes?” Therefore, the four great deities brought four stone bowls,
which he accepted, and from which he ate his first meal. According to the Lali-
tavistara, bowls of gold, silver, and various kinds of precious stones were offered
and refused. The jataka says that the gods first offered him four sapphire bowls,
which Buddha refused. Then he took the four stone ones, which fitted together
and became one. See Edward J. Thomas, The Life of Buddha as Legend and
History (London: Routledge & Kegan Paul, 1949), 86. This story is not only
narrated in the primary textual source the Lalitavistara, several Pali texts such
as Mahavagga, the Nidana Katha, and the Mahavastu also recorded the story.
Lalitavistara, though, is the only one that mentions the magical stoppage of the
carts. It introduces the forest goddess and gives details of the wheels sinking, har-
ness straps coming loose, and parts coming undone, as well as the merchants’ vain
efforts to get the carts moving, also see Dehejia, “Narrative Cycles at Gandhara,”
in Dehejia Vidya, Discourse in Early Buddhist Art: Visual Narratives of India,”
203.
6 Some say six or even seven other precious materials: silver, jasper, crystal,
amethyst, sapphire, and emeralds, see Alfred Foucher, The Life of the Buddha:
According to the Ancient Texts and Monuments of India (Middletown, CT: Wes-
leyan University Press, 1963), 132.
7 This subject is often seen in Gandhara sculpture, from which the original edges of
the four bowls can be easily observed. The bowl was revered in India as a wor-
shipping object. See Lyons Islay and Harald Ingholt, Gandharan Art in Pakistan
(New York: Pantheon Books, 1957). Also cf. Foucher, The Life of the Buddha,
132; the bowl was revered in India as a worshipping object. Fa Xian (c. 337–422)
noted that the Buddha’s bowl was being worshipped in Peshawar. See Fa-hsien,
The Travels of Fa-Hsien (399–414 AD); or, Record of the Buddhistic Kingdoms
(London: Routledge & Kegan Paul, 1959), 104. Twice a day the precious relic
was exhibited before the faithful, probably placed on a throne and protected by a
dais such as is shown on a number of the bas-reliefs from Gandhara. See Foucher,
The Life of the Buddha, 132. Fa-hsien, The Travels of Fa-Hsien, 127.
128 Considering karmic narratives
8 Mizuno Seiichi and Nagahiro Toshio, Unkō sekkutsu, vols. 8 and 9, 99.
9 The Maitreya cult was also very popular, and meditation practice is usually asso-
ciated with it. See Abe Stanley, “Art and Practice in a Fifth-Century Chinese
Buddhist Cave Temple,” Ars Orientalis, 20 (1990): 8.
10 Mizuno and Nagahiro, Unkō sekkutsu, 8 and 9: 136–7. About the story of the
two merchants offering food to the Buddha, also see Phyllis Granoff, “The Gift
of the Two Merchants: Defining the Buddhist Community through Story,” East
and West, 55, no. 1/4 (December 2005): 129–38.
11 Apparently, the artists did not use bulls or carts to represent the caravan, see
Foucher, The Life of the Buddha, 131. The meal was something sweet made with
honey and some peeled sugar cane.
12 When the Buddha attained enlightenment under the tree, he entered into a state
of mystic meditation and remained plunged in a profound trance so long that
on the seventh day the Tree-God thought of giving food to him in his trance.
Halting them by causing all to stumble, he suggested to Trapusa and Bhallika,
the leaders of the caravan, that they offer food to the Buddha and obtain great
blessings in return. They mixed wheat flour with honey and came to the tree and
offered the food to the Buddha. But the Buddha reflected that all Buddhas should
receive offerings only in a bowl and would not eat from the hand like the com-
mon pious. Just then, the Four Deva Kings knew what was in the mind of the
Buddha and immediately appeared on Mount Jamuna and out of a rock obtained
four bowls and each of the Four Deva Kings offered a bowl to the Buddha. Fear-
ing that if he were to accept only one of them, he might offend the other three
kings, the Buddha accepted all four bowls. Placing them all on his left palm, he
touched them with his right hand, made the four bowls into a nest of bowls, and
then received the wheat cakes, and after converting the merchants to Buddhism
he taught them to be devoted to Sangha by offering food and other necessities
of life and thereby obtaining boundless blessings. See Taizi ruiying benqi jing,
T3:185:479:23, translated by Zhi Qian, juan 2. Eng. translation, see Mizuno and
Nagahiro, Unkō sekkutsu, vol. 5, 79.
13 Roland Barthes and Stephen Heath, Image, Music, Text (New York: Hill, 1977), 79.
14 Ibid.
15 Ibid., 109.
16 Foucher, A, 13.
17 Ibid., 16.
18 Ibid., 14.
19 This is the case for most of Chinese Buddhist caves. In Yungang, only Syama
Jataka is depicted in continuous mode like a series of pictorials. The rest of the
narratives are mostly highlighted in one scene. I believe the reason why only
Syama Jataka is depicted in serial pictorial is that the visual representation is
greatly influenced by Confucian filial piety.
20 Dehejia suggested that there are seven modes of visual narration that ancient
artists used to tell tales of Buddhist legends. Monoscenic, continuous, synoptic,
conflated, sequential and networked modes were typically used. Some of these
modes are closely allied while others are vastly different, see Dehejia, “On Modes
of Visual Narration,” 6.
21 Foucher notes that the Mara’s Temptation and the offering of the four bowls
were two popular representations in Gandhara, and the two scenes were used
interchangeably according to the donor’s desires, see Foucher, The Life of the
Buddha, 132.
22 See Dehejia about conflated and synoptic modes, chapter one.
23 Dehejia, “On Modes of Visual Narration,” 23.
 Considering karmic narratives 129
24 Ibid., 21.
25 Ibid., 23.
26 Wei Shou (506–572), comp., Weishu (Beijing: Zhonghua shuju, 1974), 114:3030.
27 Local religious and social organizations responsible for worship activities are
called yiyi or yishe. This social organization is for the plebeian Buddhist faith, i.e.,
ren tian jiao 人天教, Man and Heaven Teachings. See Dorothy C. Wong, Chinese
Steles: Pre-Buddhist and Buddhist Use of a Symbolic Form (Honolulu: University
of Hawai‘i Press, 2004), 10.
28 Sujiang, jiangjing and changdao were all popular activities in Buddhist communi-
ties at that time, see Xiang Da. “Tangdai Sujiang Kao,” in Xiang Da, Tangdai
Changan yu Xiyu wenming (Beijing: Shenghuo, dushu, xinzhi sanlian shudian,
1957), 295–336. For a detailed description of the process of Changdao, see Hui
Jiao (497–554), Gaoseng zhuan, T2059, 50:521a10.
29 Schmid David Neil, Yuanqi: Medieval Buddhist Narratives from Dunhuang,
Ph.D. Thesis (Philadelphia: University of Pennsylvania, 2002), 13.
30 Schmid further suggested that striking similarities between the practice of chang-
dao and the performance of Dunhuang Yuanqi indicate that the latter is a con-
tinuation of the former. He pointed out that both practices center on preaching
through established genres of Buddhist narratives, nidana and avadana. As with
karmic narratives from Dunhuang, stories were told by a single, unaccompanied
monk in a specific religious framework, 22 and 85.
31 Stanleyk Abe. “Art and Practice in a Fifth-Century Chinese Buddhist Cave Tem-
ple.” Ars Orientalis. vol. 20 (1990), 11.
32 Hui Jiao, Gaoseng zhuan, T2059:50:415–418a.
33 T2059:50:417c.
34 T2059:50:417c. English translation see Schmid, Yuanqi, 156.
35 T2059:50:417c.
36 T2059:50:418a. English translation by Schmid with revision, 164.
37 T2059:50:417c
38 T2059:50:417a.
39 T2059:50:417c. Also see Schmid, Yuanqi, 165.
40 See Schmid chapter 5, note 49, 157. Also see Strong John, “The Buddhist Ava-
danists and the Elder Upagupta,” in Tantric and Taoist Studies in Honour of
R.A. Stein, Mélanges Chinois et bouddhiques XXII. Ed. Michael Strickmann
(Bruxelles: Institut Belge des Hautes Etudes Chinoises, 1985), 862–81. He
pointed out the existence of such a tradition in India centering round “avadan-
ists” (avadanika) and experts in “avadanas” (avadanarthakovida); Lamotte Eti-
enne also discusses monks who generated interest before the actual teachings of
the law by telling stories of generosity, morality, and heaven, known as amapur-
vikatha or purvakalakaraniya. They often employed illustrations, particularly in
the form of the wheel of life (samsaramandala), in their preaching, see Schmid,
Yuanqi, 157.
41 Based on my observation, there are 35 different kinds of musical instruments in
cave 12 alone.
42 T2059:50:375b.
43 For eminent monks jiangjing in Pingcheng, see Tang Yongtong, Hanwei Liangjin
Nanbeichao fojiaoshi (Changsha: Shangwu yinshuguan, 1938), 600. Also, see
Zhang Zhuo, Yungang Shiku Biannianshi, (Beijing: Wenwu chubanshe, 2006),
424–5.
44 T.2103.52:273a.
45 See Angela F. Howard and Giuseppe Vignato, Archaeological and Visual Sources
of Meditation in the Ancient Monasteries of Kuča (Laiden: E. J. Brill, 2014).
130 Considering karmic narratives
46 Ding Mingyi and Ma Shichang, “Kezi’er shiku de fozhuan gushi,” in Zhongguo
shiku Kezi’er shiku. Ed. Xinjiang weiwuer zizhiqu wenwu guanli weiyuanhui,
vol. 1 (Beijing: Wenwu chubanshe, 1989), 213.
47 Giuseppe Vignato, Districts and Groups: An Archaeological Investigation of the
Rock Monasteries of Kucha (Shanghai: Shanghai guji chubanshe, 2013), 81.
48 Mulasarvastiva davinaya (Genbenshuoyiqieyou bu pinaiye zashi). Translated by
Yi Jing in the beginning of the 8th century. T24n1451: 0283b05(08).
49 Wu Hung, “Born in Paradise: A Case Study of Dunhuang Sutra Painting and Its
Religious, Ritual and Artistic Context,” Orientations 23, no. 5 (May 1992): 55.
50 T2059:50:351c
51 Wu, “Born in Paradise,” 55.
52 Ibid.
53 Ibid.
54 T2059:50:417. English translation by Wu Hung.
55 On Changdao, see Gaoseng zhuan, T50:2059:418b01. Also see He Jianping and
Zhou Xin, “Nanbeichao fojiao changdao de diben,” Xinan minzu daxue xuebao,
vol. 9 (2013): 64–70; Lu Lizhi, “‘Changdao’ bianzhang,” Zongjiaoxue yanjiu,
vol. 1 (2012): 160–6. Chen Shuo, “Dunhuang yishu zhong de ‘changdao’ yishi
yu changdaowen zhi guanxi tanwei,” Gansu shehui kexue, vol. 4 (2012): 244–7.
56 Victor H. Mair, Tun-huang Popular Narratives (Cambridge: Cambridge Univer-
sity Press, 1983), 6–7.
57 See Xiang Da, “Tangdai Sujiang Kao,” 299. Also see Zanning, Da Song sengshi
lüe, T54:2126:239b.
58 Ibid., 305.
59 Ibid., 301.
60 English translation, see Schmid, Yuanqi, 153.
61 Ennin gives thorough descriptions of the presentation of popular lectures, see
Ennin Reischauer and Edwin O. Reischauer, Diary: The Record of a Pilgrimage
to China in Search of the Law (New York: Ronald Press Co., 1955), 151, 152,
154, 298–299, 310, and 316.
62 T2060:50:428a.
63 For instance, the narratives of celestial maiden dedicating lamps (east wall, ante-
room), the conversion of Nanda (east wall, anteroom), the woman who hated
desire and became a nun (east wall, main room), the story of Svastika bird and
Mahaprabhasa and elephant trainer (south wall, main room) represented in cave
10 and the narrative illustrations of Gautama Buddha meeting Rahula (anteroom,
west wall), celestial maiden dedicating the canopies (main room, south wall), two
brothers became monks (main room, south wall), Nirgrantha-putras redeemed by
Buddha in the Fire Samadhi (main room, south wall, west part), eight celestials
taught by Buddha (main room, south wall, west part), Hariti lost her son, Pingala
(main room, south wall, west part), and offering of meal by Sudatta’s wife (main
room west wall) in cave 9 are all derived from the stories in the Zabaozang jing.
According to the record of the Chu sanzang ji ji by Sengyou of the Liang, the
Za baozang jing contains 13 juan, the Fufazang yinyuan zhuan has 6 juan, and
the fangbian xinlun contains 2 juan. The three texts have 21 juan in total, and
they were translated in the second year of Yanxing (472) by Tanyao and Kekaya.
Sengyou, Chu sanzang ji ji. Annotated by Su Jinren and Xiao Lianzi (Beijing:
Zhonghua shuju, 1995), 62–3.
64 Mizuno and Nagahiro, Unkō sekkutsu, vol. 13, 96.
65 Cao Ling, Zhongguo Fojiao Yiweijing Zonglu (Shanghai: Shanghai guji chuban-
she, 2011), 84.
66 Daoxuan, T50:428a.7–12.
67 The first significant scholarship on the subject is by Japanese scholar Tsukamoto
Zenryū, see Shina bukkyō no kenkyū: Hokugi hen (Tokyo: Kobuntō, 1942), 293–354.
 Considering karmic narratives 131
68 Abe, 3.
69 Kenneth K. S. Chʻen, The Chinese Transformation of Buddhism (Princeton, NJ:
Princeton University Press, 1973), 56–8.
70 Makita Tairyō, “Tonkohon Daii-kyō no kenkyū,” Bukkyō Daigaku daigakuin
kenkyū kiyō, I (1968): 149–150.
71 Whalen W. Lai, “The Earliest Folk Buddhist Religion in China: Ti-wei Po-li Ching
and Its Historical Significance,” in Buddhist and Taoist Practice in Medieval
Chinese Society. Ed. Chappell David W (Honolulu: University of Hawaii Press,
University of Hawaii Press, 1987), 15. Makita Tairyo identified a Dunhuang text
(Stein no. 2051 in the British Museum) as the second scroll of the TWBLJ. See
Makita Tairyō, “Tonkohon Daii-kyō no kenkyū,” in Bukkyō Daigaku daigakuin
kenkyū kiyō, I (1968), I:137–85; and II (1971), 165–97.
72 Daoxuan, T2060:50:428a.
73 Lai, “The Earliest Folk Buddhist Religion in China,” 13.
74 Ibid., 13–14.
75 Daoxuan, T2060:50:428a. English translation with minor revision, see Lai, “The
Earliest Folk Buddhist Religion in China,” 15.
76 The inscription is on the east wall of cave 11 in the upper level.
77 Li Jingjie, “Guanyu Yungang dijiu, dishi ku tuxiang goucheng,” 327–59.
78 See Lidu Yi, The Third-Phase Yungang Cave Complex: Its Architecture, Subject
Matter, Composition and Style, Ph.D. Dissertation (Toronto: The University of
Toronto, 2010), 118–19.
79 Mizuno and Nagahiro, Unkō sekkutsu, vol. 13, 96.
80 Ibid., 96.
81 Abe, 2.
82 Weishu, 114:3037.
83 Sato Chisui, “The Character of Yün-Kang Buddhism: A Look at the Emergence
of a State-Supported Religion in China Under the Northern Wei,” The Memoirs
of the Research Department of the Toyo Bunko, vol. 36 (Tokyo: Toyo Bunko,
1978), 67–73. Also see Lai. “The Earliest Folk Buddhist Religion in China,”
13–14; and Abe, 2.
84 The Mingseng zhuan was complied by Baochang 寶唱 in the early 6th century.
It is a collection of Buddhist monks’ biographies. Unfortunately, only a fraction
copied in 1235 by the Japanese monk Shusho was preserved. Shusho, comp.,
Meisodensho, reprint from Dai Nihon zoku zokyu (Taibei, 1975), 13. Transla-
tion see Coburn Soper A. “Literary Evidence for Early Buddhist Art in China,”
Artibus Asiae, Supplementum, 19 (1959): 54.
85 He died in 472. Also see Abe, 4.
86 Lai, “The Earliest Folk Buddhist Religion in China,” 26.
87 Ibid.
88 Ibid., 16.
89 Tang Yongtong, 586.
90 T2060:50:428a.
91 Lai, “The Earliest Folk Buddhist Religion in China,” 20.
Plate 1 Shanxi caves
Plate 2 Yungang complex on Mount Wuzhou
Provided by the Yungang Research Academy
Plate 3 Tanmei inscription discovered in front of Cave 20, dated 503
Plate 4 Yungang Area satellite map

Plate 5 Ruin trances of lotus and tortoise patterns

Plate 6 Ruins of roof support, Caves 9 and 10
Provided by the Yungang Research Academy

Plate 7 Remains of the Siyuan monastery

Plate 8 Ethnic dancers, Shaling tombs, Datong, Shan’xi province
Provided by the Yungang Research Academy
Plate 9 Manes jewels, Shaling tombs, Datong, Shan’xi province
Provided by the Yungang Research Academy

Plate 10 Manes jewels, Cave 9

Plate 11 Seated Śākyamuni and Prabhutaratni, Cave 11-14, dated 489
Plate 12 Cave 19, south wall, west side, upper level, right attendant Buddha
Plate 13 Seated Śākyamuni and Prabhutaratni with a Bodhisattva (above), east
reveal of window, Cave 17, dated 489, commissioned by a nun named Huiding
Plate 14 North wall, anteroom, Cave 9

Plate 15 Narrative scenes, east wall, Cave 6

Plate 16 West wall, anteroom, Cave 10

Plate 17 East wall, anteroom, Cave 9

Plate 18 Standing Buddha, west wall, Cave 6
Plate 19 Seated Buddha with legs crossed, west wall, Cave 30
Plate 20 Seated Bodhisattva with ankles-crossed, east wall, Cave 5-11
Plate 21 Mural paintings, Shaling tombs, Datong, Shan’xi province
Plate 22 External view of Cave 19

Plate 23 Seated Buddha, Cave 19

Plate 24 Seated Buddha, Cave 20

Plate 25 Standing Buddha, Cave 18

Provided by the Yungang Research Academy
Plate 26 Seated Bodhisattva with ankles-crossed, Cave 17

Plate 27 Standing Buddha, Cave 16

Plate 28 Seated Bodhisattva with ankles-crossed, Cave 13
Provided by the Yungang Research Academy

Plate 29 Central pillar, Cave 11

Plate 30 Seated Buddha and Bodhisattva (above) with Many Treasures stūpas, east reveal, window,
Cave 11, dated 495
Plate 31 Inscription on the upper east wall, dedicated by an yiyi of fifty-four men and
women, Cave 11, dated 483
Plate 32 Seated Buddha with legs pendant, north wall, Cave 9
Provided by the Yungang Research Academy
Plate 33 Svastika bird story, south wall, main room, Cave 10

Plate 34 Woman who hated desire and became a nun, east wall, main room, Cave 10
Plate 35 North and east walls, Cave 10

Plate 36 North and west walls, Cave 10

Plate 37 West wall, anteroom, Cave 9

Plate 38 East wall, anteroom, Cave 10

Plate 39 Central pillar, Cave 1

Plate 40 South wall, Cave 1

Plate 41 Seated Buddha with legs crossed, Cave 5

Plate 42 Great departure narrative story, south wall, Cave 6

Plate 43 Central pillar, Cave 2

Plate 44 Lower level, east wall, anteroom, Cave 12

Plate 45 Lower level, west wall, anteroom, Cave 12

Plate 46 Coved area, anteroom ceiling, Cave 12. Provided by the Yungang Research Academy
Plate 47 Upper level, east wall, anteroom, Cave 12

Plate 48 Upper level, west wall, anteroom, Cave 12

Plate 49 North wall, anteroom, Cave 12

Plate 50 East side, north wall, anteroom, Cave 12

Plate 51 West side, north wall, anteroom, Cave 12

Plate 52 North wall, main chamber, Cave 12

Plate 53 East reveal, window, Cave 12

Plate 54 West reveal, window, Cave 12

Plate 55 East wall, main chamber, Cave 12

Plate 56 West wall, main chamber, Cave 12

Plate 57 South wall, east side, main chamber, Cave 12

Plate 58 South wall, west side, main chamber, Cave 12

Plate 59 East wall, Cave 27

Plate 60 West wall, Cave 27

Plate 61 East wall, Cave 29

Plate 62 West wall, Cave 29

Plate 63 Entrance, Cave 35, dated 515
Plate 64 Seated Buddha with legs crossed, north wall, cave 5-11

Plate 65 Seated Buddha with legs crossed, west wall, cave 5-11
Plate 66 Xu Xianxiu tomb mural painting, Taiyuan, Shan’xi province
6 Phase three – the remaining
splendor

The third-phase caves include primarily the cave temples that are situ-
ated at the western end of Yungang (caves 21–45, to the west of cave 20),
caves 4, 14, and 15, niches scattered outside caves 11 and 13 on the cliff
plain, and caves 5-10 and 5-11 above cave 5 on the hill. All these caves and
niches were excavated after the capital was transferred southward to Luoy-
ang in 494, and the excavation was continued through the Zhengguang
正光 reign period (520–25). During this period, Yungang still remained a
Buddhist Mecca with fervent support of local official patrons and devotees.
But, at the end of the Northern Wei, Pingcheng fell into a low ebb with
constant riots near the capital involving, or even being led by, Buddhist
monks; therefore, the construction of Yungang gradually came to an end
(the latest inscription found thus far in Yungang is dated to 524 in cave 4).1
As the Jin stele records, “examining the surviving inscriptions, [the excava-
tions of the cave temples] came to an end during the Zhengguang era 驗
其遺刻, 終乎正光.”
The third-phase cave temples are relatively small in scale compared to
the caves in the first two phases. Some can only accommodate one person,
nothing as grand or monumental as the earlier caves that can “hold as
many as three thousand,” as described in the Xu gaoseng zhuan.2 There are
neither elliptical colossal image cave temples of the first-phase, nor paired
caves with front chamber in imitation of wooden structure, characteristic
of the second-phase cave temples. In spite of this, however, much innova-
tion was brought forth based on the tradition of the cave temples of the
first two phases. For example, rectangular or square cave plans with flat
ceilings and three niches on three walls (sanbi sankan 三壁三龕) became
the dominant architectural structure. A new trabeated niche style with a
folding screen front emerged, as seen in cave 5-11 (see plate 20). In addition,
the subject matter was rich, and iconographic composition was diversified.
For instance, symmetrically opposed scenes appeared illustrating such nar-
ratives as “Great Departure” and “Outing on the Elephant.” In style, the
images appear sinicized: graceful and refined. All of these demonstrate the
distinctive features and infinite charm of the third-phase cave temples. To
 Phase three – the remaining splendor 133
a certain degree, the cave temples of the third phase still played a pivotal
role in the excavation of Buddhist rock-cut cave temples. They exerted great
influence on the excavation of cave temples in north China. Their role in the
process of the development of cave temples in the Northern Wei cannot be
neglected. Without fully understanding the third-phase cave temples, it is
hard to comprehend the process of development of Buddhist rock-cut cave
temples in north China.
Though deserving of study, however, they somehow eluded the attention
of scholars. After nearly 100 years of academic research and investigation
of Yungang, the primary focus was on the cave temples of the first two
phases. Very little attention has been paid to these caves. One may well say
that there are very few research results thus far. The architectural struc-
ture, subject matter, iconographic composition, and artistic style form an
important part of the study of Buddhist rock-cut cave art. Therefore, I will
attempt to examine the features of the third-phase cave temples from these
perspectives and construct a dating scheme of the excavation, in hopes of
bringing them sufficiently to the attention of the academic community. It
should be explained here that the primary focus of this study is placed on
the cave temples located to the west of cave 20, as well as caves 5-10 and 5-11
(35 caves in total).3 Other caves and niches that are severely damaged, dif-
ficult to climb up to on the cliff, or otherwise difficult of access, will not be
touched upon. It should also be mentioned, since this is the first thorough
examination of the third-phase cave temples, that the cave numbers used
here follow the numbers compiled and edited by the Academy of Yungang
Research.4

Architectural structure
Buddhist cave temples can be divided into two basic types: liturgical service
cave temples (libaiku 禮拜窟) and vihāra cave temples (sengfangku 僧房窟).
There are, however, only liturgical service cave temples in Yungang, which
can be subdivided into Buddha Hall image caves (fodianku) and caitya caves.
All third-phase cave temples are image caves except cave 39, which is a caitya
cave. It should be noted that the architectural structure here is divided into
cave temples, niches, and ceilings. In fact, niches and ceilings belong to the
part of cave structure, but since they are complicated and diversified, I will
therefore treat them separately.

1. Cave structure
Based on niche arrangement, the cave temples can be divided into six
types: (1) three niches on three walls, (2) one niche on the north wall and
storied niches on other walls, (3) storied niches on three walls, (4) no niches,
(5) two niches on three walls, and (6) thousand Buddhas caves.
134 Phase three – the remaining splendor

Fig 6.1 Plan, cave 25

Three niches on three walls 三壁三龕

Three niches on three walls is the prevalent architectural structure in the
third phase. It constitutes almost half of the third-phase cave temples. The
cave temples include caves 22-1, 23-1, 24, 25, 27-2, 28, 32-11, 32-12, 33, 33-6,
34, 35, 37, 38-3, 40, and 41 (16 caves in total). To take cave 25 (fig. 6.1) as
an example, a niche with seated Buddhas appears in the center of each wall
in the cave.5

One niche on the north and storied niches on the

other walls 北壁一龛, 侧壁重龛
Type two includes cave temples 21, 22, 24-1, 26, 33-3, 33-4, and cave 38
(seven caves in total). The north wall contains one large niche in which the
Śākyamuni and Prabhutaratna are represented. The west and east walls are
covered by storied niches. The top one is usually a trabeated niche contain-
ing a Bodhisattva with ankles crossed, and the lower niche, a pointed arch,
 Phase three – the remaining splendor 135
houses the Śākyamuni and the Prabhutaratna. Cave 26 is an example of this
type. Cave 38 is unique (as will be discussed in detail later): the east wall is
covered with storied niches, but the north and west wall contain only one
niche on each wall.

Storied niches on three walls 三壁重龛

This configuration is seen in caves 27, 29, 30, 31, 32-9, and cave 36
(six caves in total). All three walls bear storied niches. According to the
arrangements of the niches on the north wall, they can be subdivided into
two types: either two niches or six niches on the north wall. The former
includes caves 27 (fig. 6.2), 29 (fig. 6.3), and 30 (fig. 6.4), the north wall
of which contains six niches arranged vertically in two tiers, three on each
tier. The standing Buddha to the north side of the west wall in cave 30 has
been looted (fig. 6.5).6 The latter group of this type includes caves 31, 32-9,
and 36.

Fig 6.2 Plan, cave 27

Fig 6.3 Plan, cave 29

Fig 6.4 Plan, cave 30

 Phase three – the remaining splendor 137

Fig 6.5 West wall, cave 30

Caves without niches 无龛窟

It includes only three caves (23, 25-2, and 28-2). In front of each wall in the
cave, there is an altar on which a Buddha image appears. Caves 25-2 and
28-2 were mostly left unfinished, except for the main Buddha. The niches
on west and east walls of cave 28-2 were added later. It is worth noting that
this cave contains two inscriptions with explicit dates, one the fourth year
of Zhengshi 正始 (507) and the other the third year of Yanchang 延昌 (514)
on the west and east walls respectively. This implies that the excavation of
cave 28-2 should not be later than 507.

Two niches on three walls 三壁二龛

Cave 5-10 (fig. 6.6) and cave 5-11 (fig. 6.7) can be regarded as the best exam-
ples of the third-phase cave temples in visual aesthetics and representation.
These are among the very few examples of colored sculpture images in the
entire Yungang complex. The main wall contains no niches. There is a low
altar on which all the images are standing. The east wall contains a trabeated
niche and the west wall has a pointed arch niche.
138 Phase three – the remaining splendor

Fig 6.6 Plan, cave 5-10

Fig 6.7 Plan, cave 5-11

Thousand Buddha cave 千佛龛

Cave 36-2 is the only thousand Buddha cave. All three walls are filled with
small thousand Buddha niches. At the bottom of each wall is the depiction
of donors in procession. The north wall bears seven Buddhas between the
thousand Buddha niches and the donors.
 Phase three – the remaining splendor 139
2. Niche types
There are four kinds of niches: (1) pointed-arch, (2) trabeated, (3) roofed,
and (4) canopied niches, among which the pointed-arch and trabeated niches
are the most popular and are represented in every cave, while the other two
rarely appear. The niches are simpler and less ornate than those in the second
phase. Some of the trabeated fronts are plain, without any ornament.

The pointed-arch niche

The pointed-arch niche usually comprises arched fronts, arched ends, and arched
posts. The arched front is often decorated with narrative stories of the life of
the Buddha or a row of seated Buddhas, normally seven, above which is a row
of ornate garlands held by a row of aupapadukas (youths born by spontaneous
transformation). The best representation of the arched front is exemplified by
those in the west walls of cave 34, cave 28, and cave 25. The arched ends are
normally terminated with animals (dragons, phoenixes, and unidentified crea-
tures) that turn their heads inward as if looking towards the Buddha. The best
examples of arch ends appear in cave 27 and cave 29 (discussed later).

The trabeated niche

This type consists of trabeation and posts that separate the niche into three
parts, with the central niche usually containing the cross-ankled Bodhisattva,
and left and right compartments bearing the seated Bodhisattva with one leg
pendant. The trabeation is divided into a central part, rhomboidal parts, and
wings. A looped curtain sometimes appears underneath the trabeation. The
trabeated front is usually decorated with flying celestials and musicians as
well as seated Buddhas. The most magnificent trabeated front is seen in the
eastern wall of cave 21, where the flying celestials and musicians are soaring
gracefully fluttering with movement. The trabeated front on the eastern wall
of cave 5-11 is like a folding screen, in the center is a seated Buddha beneath
which aupapadukas holding a row of garlands appear.

The roofed niche

This consists of chiwei (roof ridge ornaments), rafters, brackets, triangular
ornaments, and side pillars, which divide the niche into three parts: a central
niche, which usually houses a seated Buddha with both legs crossed, and two
side compartments. In a typical case, underneath the roofed niche, a trabe-
ated niche appears, as seen in the western side of the south wall in cave 38.

The canopied niche

This consists of inverted triangular and round ornaments under which
looped curtains usually appear. Canopied niches can be found in caves 30,
32-12, 33-3, and 36-6.
140 Phase three – the remaining splendor
3. Ceiling structure
According to the structural shape, the ceilings can be broadly divided into
two types: rectangular or square.7 According to the decorative patterns, they
can be subdivided into: coffered (pingqi 平棊) and neither coffered nor cor-
belled (zaojing). Compared with the most ornate second-phase ceilings such
as the pingqi zaojing 平棊藻井 ceiling in the anteroom of cave 12, ceilings
became much simpler in the third phase. Some ceilings are not even coffered,
but only decorated with lotuses and/or flying celestials. The pingqi pattern
consists usually of either square, rectangular, or trapezoid forms with 4, 5,
6, 8, 9, or 12 chessboard-like panels, inside of which are flying celestials,
lotuses, and sometimes intertwined dragons. For instance, the ceiling of cave
29 has nine trapezoid panels with a lotus in the center. During this time, the
pingqi design usually has nine panels with lotus pattern in the center and
eight flying celestials surrounding it, for example, the ceiling of cave 24.
Sometimes there are five lotus patterns and four flying celestials in the nine
panels as exemplified in cave 33-1. The ceiling of cave 39 is the only one in
the third phase that bears the themes of the Dharma protector deities.

Iconographic composition
The iconographic composition in the caves is primarily focused on the main
images. They are the primary objects of worship. Surrounding them, in some
caves, the biography of the Buddha and jataka stories, etc., appear as well.

1. Composition of the primary images in cave temples

Three niches on three walls

The three main images are all Buddhas. The designs are of seven types: (1) three
seated Buddhas with legs crossed (for example, caves 23-1 and 43), (2) a cross-
legged seated Buddha together with a cross-ankled Bodhisattva and a seated
Buddha with legs pendant (caves 32-12, 35, and 40), (3) the Śākyamuni and
the Prabhutaratna on the north wall with a seated Buddha and a cross-ankled
Bodhisattva on side walls (caves 33-6, 33, 38-3, and 42-4), (4) the Śākyamuni
and the Prabhutaratna on the north wall with a cross-ankled Buddha and a
cross-ankled Bodhisattva on side walls (cave 33), (5) the Śākyamuni and the
Prabhutaratna represented with two seated Buddhas on side walls (caves 25,
28, and 34), (6) a cross-ankled Bodhisattva on the north wall with two seated
Buddhas on the side (cave 37), and (7) a cross-legged Buddha on the main wall
and Śākyamuni and Prabhutaratna on both east and west walls (cave 24).

One niche on the north and storied niches on the side walls
All north wall niches bear images of the Śākyamuni and Prabhutaratna in
this type, but the composition of the storied niches on the east and west
 Phase three – the remaining splendor 141
walls is complicated. For instance, in cave 26, on each of the east and west
walls, the upper niche contains a Bodhisattva with ankles crossed, and the
lower niche bears the Śākyamuni and Prabhutaratna. Each wall then in
turn bears the three-Buddha design, with no compositional association with
the Śākyamuni and Prabhutaratna on the north wall. Some niches on the
sidewalls contain a Bodhisattva with ankles crossed and a seated Buddha.
For example, caves 24-1, 33-3, and 33-4. The iconographic composition in
cave 24-1 is unique: the very center of each sidewall has two-storied niches,
the top level contains a cross-ankled Bodhisattva flanked by two pensive
Bodhisattvas, the bottom level bears a seated Buddha flanked by two stand-
ing Bodhisattvas. These six images in the center, in turn, are flanked by six
images arranged vertically in three levels. The top is a seated Buddha with
legs pendant, the bottom one is a standing Buddha and the central figure is
a seated Buddha. The vertical arrangement of the three images is associated
with the concept of the three Buddhas.

Storied niches on three walls

The arrangement of the primary images on the north wall is complicated in
this type. For instance, in cave 27, the top level contains three cross-ankled
Bodhisattvas and the bottom level contains a seated Buddha flanked by two
seated Buddhas with legs pendant. Each of the top levels of caves 29 and
30 bears a seated Buddha with legs pendant in the center flanked by two
Bodhisattvas with ankles crossed, the bottom level is the same as those in
cave 27. Hence, in these three caves, the upper and lower niches in the center
on the north wall constitute the composition of Śākyamuni and Maitreya.
In caves 31 and 32-9, on the north wall, the top niche is a cross-ankled Bod-
hisattva, and the bottom niche bears the Śākyamuni and Prabhutaratna, a
configuration of the three Buddhas. The two stories on the east and west
walls of cave 31 are also intended to show the three Buddhas. The top niches
contain a seated Buddha with both legs pendant and a Bodhisattva with
ankles crossed; and the bottom niche bears a seated Buddha cross-legged.
They are meant to show aspects of Maitreya’s ascent and descent. In the
case of cave 36, the bottom niche bears a seated Buddha and the top niche
has a Bodhisattva with ankles crossed on the north wall. The top niches
bear a seated Buddha, but the lower level niche has a seated Buddha with
legs pendant and a Bodhisattva with ankles crossed on east and west walls
respectively. The iconographic design of each wall consists of images of the
Śākyamuni Buddha and Maitreya. The main images on the three walls in
each of the caves discussed above do not have iconography associated.

Caves without niches

The main images in cave 23 are unique: the north wall bears the Śākyamuni
and Prabhutaratna, and the side walls have two standing Buddhas flanked by
142 Phase three – the remaining splendor
a standing Bodhisattva. Caves 25-2 and 28-2 were not completed. The north
wall of cave 25-2 bears a seated Buddha, and on each sidewall is a Bodhisat-
tva and a disciple, making a total of five images.

Two niches on three walls

The primary images in caves 5-10 and 5-11 are the same: a seated Buddha
on an altar on the north wall, a seated Bodhisattva with ankles crossed on
the east wall, and a seated Buddha in the west niche. Together, they form a
three-Buddha design.

2. Composition of the images on each wall

The main (north) wall is usually covered with three or four images, i.e.,
Śākyamuni and Prabhutaratna together with two standing Bodhisattvas, as
seen in caves 23, 28, 33-6, 38-3, 24-1, and 33-3; or a seated Buddha with two
standing Bodhisattvas, as seen in caves 23-1. Later cave temples began to have
five or six images, as seen in caves 33, 34, 35, and 33-4. The Śākyamuni and
Prabhutaratna are seen together with Bodhisattvas and disciples.
The composition on the west walls displays a rich variety. The most popu-
lar one is, as on the main wall, a seated Buddha flanked by two standing Bod-
hisattvas. The composition of one seated Buddha with legs pendant flanked
by two standing Bodhisattvas often appears as well. The composition on
the storied west walls is more complicated. For example, in cave 24-1, the
top level in the center contains a cross-ankled Bodhisattva flanked by two
pensive Bodhisattvas, the bottom level bears a seated Buddha flanked by
two standing Bodhisattvas. In cave 26, the upper niche in the center is a
Bodhisattva with ankles crossed with two pensive Bodhisattvas, the lower
niche contains the Śākyamuni and Prabhutaratna with two standing Bod-
hisattvas. The west wall of cave 33 also presents a unique iconography. Here
the cross-ankled figure, usually a Bodhisattva, is a Buddha, and it is flanked
by two standing Buddhas of the jataka stories. The composition in cave 30
is unusual as well. The top niche contains a seated Buddha flanked by two
standing Buddhas, not the usual Bodhisattvas.
The iconographic composition on the east wall is usually simpler. The
composition that appears most often is a cross-ankled Bodhisattva in the
center attended by two standing Bodhisattvas, e.g., cave 5-11. In the case of a
storied east wall, the cross-ankled Bodhisattva in the center flanked by two
pensive Bodhisattvas is often represented together with a cross-legged Bud-
dha flanked by two standing Bodhisattvas one above the other (caves 24-1
and 38). Sometimes the bottom niche is replaced by the Śākyamuni and
Prabhutaratna with two attendant Bodhisattvas (caves 21 and 26).
In sum, the iconographic composition of the main image in each third-
phase cave and on each wall displays a rich diversity and is much more
complicated than that of the first two phases. Amidst this diversity, the
 Phase three – the remaining splendor 143
composition of the Śākyamuni and Prabhutaratna on the main wall became
increasingly popular and constitutes nearly half of the iconography. This
reflects the popularity of the Lotus Sutra in Yungang and its faith. The
Śākyamuni and Prabhutaratna eventually became the visual symbol of the
popular sutra. They are often represented together with a cross-ankled Bod-
hisattva to manifest the three Buddhas symbolizing the perpetuation of
the Buddha Dharma infinitely. The composition of the cross-ankled seated
Bodhisattva and the seated Buddha with legs pendant on sidewalls also
emerged. This is associated with Maitreya’s ascent toward, and descent
from, the Tushita Heaven. The composition on each wall is mainly com-
posed of one seated Buddha with two Bodhisattvas, or a cross-ankled seated
Bodhisattva attended by two pensive Bodhisattvas.
It should be noted that the images of disciples appeared rather late in Yun-
gang, and are not represented in every cave. They made regular appearances
only in the later Longmen caves, and were usually represented in bas-relief.
There are normally two disciples on each wall representing the two great
disciples of Mahakasyapa and Ananda. In cave 13-10 in Yungang and Lu
cave 路洞 in Longmen, however, there are four disciples, an arrangement
not observed in the other Yungang caves.8

Telling tales: narrative stories and visual representations

In addition to the main Buddha images in the cave temples, the supplemen-
tary subjects of the life stories of the Buddha, the jataka and nidana stories,
as well as the legendary debate between Vimalakirti and Manjusri, and oth-
ers, are rich and diversified in form and content. Among them, the significant
moments in the life of the Buddha, from his miraculous descent from Tusita
heaven to his serene realization of Nirvana, are especially favored subjects.
To date, as far as I can ascertain, more than 90 scenes with 33 themes in
Yungang can be identified as illustrations of narratives concerned with the
Buddha’s life.9 In the third phase alone, more than 20 of them are depicted
including: his descent from the Tusita Heaven in the form of an elephant; his
miraculous birth from his mother’s hip in the form of a human being; his first
bath by nagas; his riding on an elephant; archery competition and tossing the
elephant; his great departure from the palace and his bidding farewell to his
white horse Kanthaka; his first meditation under the tree; his defeat of the
demons and conversion of the three Kasyapas; his first sermon; his ascent to
Tavatimsa to preach for his mother Mayadevi; and his final serene realization of
Parinirvana. These narratives constitute an important part of the subject matter
of the third-phase cave temples. Some of them were never depicted in the earlier
cave temples. For example, the prince tossing the wild elephant depicted in cave
37,10 and Buddha’s ascent to Tavatimsa to preach for his mother seen in cave
38.11 These subjects are rarely seen in Chinese Buddhist caves, and certainly
only once in Yungang. Most of these narratives are represented in a mono-
scene, and only the Syama jataka, influenced by Chinese traditional Confucian
144 Phase three – the remaining splendor
ideology, is represented in sequential order like the unrolling of a handscroll
painting. Some scholars believe that Buddhist narratives in successive scenes is
influenced by Chinese tradition, in particular by the compositional structure of
handscroll paintings.12 Jin Weinuo stated that multiple, linked scenes organized
into horizontal registers obviously show the influence of Chinese handscroll
painting and are a sign of the sinicization of Buddhist art.13 Pao-chen Chen
also noted that the presentation of successive scenes in a linear sequence is an
innately Chinese preference.14 However, Julia K. Murray argues,

The belief that Chinese handscroll shaped the representation of Bud-

dhist narratives is based on yet another supposition for which there is
little supporting evidence, namely, that the handscroll was a standard
format for Chinese painting in the period before Buddhism interacted
significantly with Chinese art and culture.15

She concluded,

Rather than assume that the existence of the handscroll in China induced
Buddhist narrative subjects to be illustrated in a sequential presentation,
I am proposing that the development of the handscroll as a format for
painting in China was stimulated by the advent of Buddhism.16

Several other scholars also hold that the format of successive scenes in a
narrative was introduced when Buddhism came to China, and that the for-
mat of handscroll cannot explain the reversal of directions of the narrative
scenes in Dunhuang. As Murray observes “if the handscroll format influ-
enced the creation of this illustration (Mahasattva jataka scenes in cave 428
in Dunhuang, and the Sudhana jataka on the other side of the cave door),
there should be no reversal of direction.”17 Michael Sullivan also states that
the technique of continuous, linear narration appears to have come from
India with the introduction of Buddhism, for there is no evidence of it in
Han art.18 This has been a much-debated and unsettled issue. However, the
discovery of a painted lacquer box (34 inches long and 2 inches high) from
Baoshan 包山 tomb 2 at Jingmen 荊門, Hubei province, shed new light on
the issue.19 It was previously believed that there were no scenes placed in
narrative sequence before the Qin dynasty in the history of Chinese art,
however, the new archaeological findings of the narratives on the lacquer
box not only filled in a gap in our knowledge, but also push us to reconsider
the long accepted theory. As Wu Hung points out,

This work shows remarkable advances in both spatial conception and

temporal representation. Much like a later handscroll painting, it must
be viewed section by section in a sequential order. The horizontal format
and sequential reading invite the viewer to read figures and scenes as
components of a continuous narrative.20
 Phase three – the remaining splendor 145
The story on the round box is highlighted by the procession and the reception
leading the viewers to imagine the beginning and/or the end of the story. Simi-
larly, in Buddhist narratives, sometimes there is neither beginning nor end. The
beginning can be the end and vice versa, like a circle of continuous narratives.
The narrative stories in the third phase are primarily depicted in mono-scene
with only very few continuous scenes, as will be seen below in the discussion of
the form, content, and related issues of the narratives depicted in the third phase.

The Great Departure and elephant riding

These are the two most popular motifs,21 and are usually represented together
as a pair on the front wall over either side of the entrance of the cave.22 The
latter displays the prince riding on an elephant; the attendants follow him
holding a long handle parasol. The former, the Great Departure through
the city gates at night, always shows the prince, the groom, Kanthaka (the
horse), and the four guardian kings, illustrated either in profile or in full
face. The groom usually carries the parasol by its long handle and the four
guardian kings hold Kanthaka’s hooves in their hands to soften their touch
on the ground. The story denotes the Buddha’s renunciation of this worldly
life when the prince was 19 years old. He convinced himself that it was time
to leave home although he knew his father would be against it. He wanted
to relieve the pain of rejection and separation of all living things from the
ones they love. On the night of the departure during the second watch,
the guardian kings and Devas fully filled the sky and said to the prince,
“inside and outside the palace everyone is sound asleep. Now it is time to
leave.” The prince himself then went to wake up Chandaka, who asked
“why do you want the horse saddled suddenly in the middle of the night?”
The prince explained that he wanted to overcome life’s troubles for human-
ity, and wanted to leave the affairs of this world to follow the path of truth.
In his lion’s voice, he roared: “in the past the various Buddhas left home, and
now so do I.”23 The Devas then carried Kanthaka, the white horse, uphold-
ing his four hooves, and Indra, holding the canopy, followed. The Devas
then commanded the north gate to open without making any sound. So the
prince departed through the gate.24
The scene of the prince riding the elephant is often depicted with the Great
Departure in the third phase. In ancient India, all noblemen at the time of the
Buddha learned four skills: elephant riding, horsemanship, chariot-driving,
and handling of arms. Elephant riding and horsemanship were mandatory
skills for a nobleman, they must have been important activities in the life of
the prince. Although the Great Departure is an important turning point of
the Buddha’s life, it is not directly associated with elephant riding; however,
interestingly, the two are often paired and repeatedly represented in the third-
phase caves. Both show important episodes in his youth. It can be assumed
that these two important events are represented together for aesthetic purpose
to balance the visual space: one side showing a horse and the other side the
146 Phase three – the remaining splendor
elephant (both animals are central to a young nobleman’s life). The athletic
training of elephant riding and departing from the royal city are the two
significant moments in the lives of youths in those times. In visual art, espe-
cially in Buddhist cave sanctuaries, illustrations of these two motifs appear
in prominent positions so as to emphasize them and facilitate their use as
inspiring tools for proselytization. Elephant riding, however, is not associated
with the birth of the prince, which is discussed next.

Maya’s dream
This episode is normally illustrated with Maya lying down asleep, dreaming
of an elephant entering her body with the prince not even shown. The story
was a very popular motif in Gandhara art, but in Yungang the Buddha’s
early life is not depicted as much as his later life in the pictorial narratives.
The scene on the east wall in cave 37 is the only example, and is depicted
in a peculiar format. Here the Bodhisattva is seen descending from Tusita
Heaven mounted on an elephant while entering his mother through her right
flank (fig. 6.8). This is a rather unique illustration since the prince otherwise
never appears in this episode. It seems that Chinese artisans tried to solve
the puzzle of how and when the Bodhisattva transformed himself from a
six-tusked elephant, in which form he entered his mother’s body, to a human
being, and in which form he emerged from her womb.25

Fig 6.8 East wall, cave 37

 Phase three – the remaining splendor 147
An issue related to the episode of the conception is that it is often mis-
interpreted as the elephant-riding episode.26 As I have suggested above, the
elephant-riding episode demonstrates one of his skills in youth, it is thus not
intended to illustrate his conception, which is depicted with images of Maya
lying down with the elephant beside her, without the prince. The portrait of
the episode in cave 37 with the Bodhisattva mounting an elephant to enter
his mother’s body is the only exception.27 Another piece of evidence to sup-
port our inference that the elephant-riding illustration is the prince’s athletic
training, not the conception, can be found in cave 6, in which there are 37
depictions of the Buddha’s historical life. The stories on the central pillar are
represented clockwise, with the depiction of the elephant-riding on the east
side of the north face in the lower level and the ascetic Asita divining the
prince’s future, which happened right after the birth, depicted on the west
side before the riding of the elephant. The narratives in the central pillar are
read in sequential order clockwise. It is obvious then that the elephant-riding
episode could not be Maya’s dream of conception and her subsequent preg-
nancy. Rather, an event happened after Asita’s prophecy and it represents the
prince’s essential athletic training in youth.

Birth of the Buddha

The miraculous birth of the Buddha under the trees in the Lumbini garden
is one of the four most important events (birth, Enlightenment, first sermon,
and Parinirvana) in his life. It did not, however, seem to appeal to the Yun-
gang artists as much as it should have. The episode is usually depicted in
a small panel in bas-relief (caves 33-2, 33-3, and 41) and is represented in a
slightly different manner in Yungang than in Gandhara art. Although Maya
is still seen standing up grasping the branch of a tree, in Gandhara art, the
infant Buddha comes forth from the right side of her body, whereas in Yun-
gang, the baby sometimes comes out from her left side, as seen in cave 33-3.28

Seven steps
Immediately after the Buddha’s mystical birth, he slipped out of the hands
that received him and took seven steps and proclaimed in a lion’s voice:
“Among all beings, I am the first and I am the best. I have come here through
infinite births and deaths for the benefit of mankind.” As soon as the new-
born left the hands that received him, the lotus miraculously emerged from
the earth under his feet to protect the newborn from contact with the ground.
The episode is only portrayed on the north wall in cave 41.

Nagas bathing the Buddha

Another immediate activity after the Buddha’s miraculous birth was his first
bath. We are told in one of the literary sources that two Naga (water spirit)
148 Phase three – the remaining splendor
brothers, named Kala and Akala, welcomed the infant Buddha and gave him
the first bath. One spirit, on the left, rained warm water down on him, while
the other, on the right, poured cold spring water.29 The first bath is vividly
depicted on the east wall in cave 37.

Asita’s prophecy
The great sage Asita’s prophecy of the newborn is often depicted with Asita
holding the infant in his hands as seen in cave 41 where the episode is illus-
trated together with the birth of the Buddha and the seven steps.

Farewell to his white horse Kanthaka

The subject of the Buddha’s farewell to his white horse Kanthaka is usually
represented in small panels in bas-relief.30 The Yungang artists did not seem
to favor the subject as much the Gandhara artists, but in both representa-
tions, Kanthaka is seen bending his legs, licking the prince’s foot, and crying
with tears falling down like rain.31 There are no illustrations of the return of
Chandaka and Kanthaka in Yungang.

Banishing of the demons and the conversion

of the three Kasyapas
These are among the most favored subjects, usually depicted as a pair, and
portrayed in a large pointed-arch front on the south wall on either side of the
entrance or the window of the cave, as seen in cave 35.32 The conversion of
the three Kasyapas are usually characterized by images of Brahmins holding
bottles in their hands attempting to put out the fire and the Buddha holding
a bowl with the three Kasyapas inside. The conflict between the Bodhisattva
and Mara almost became a metaphor, like the temptation, of the ordeals one
has to endure before ultimately becoming Enlightened. The Assault of Mara
(Banishing of the Demons) is about temptation, resistance against the lure
of lust and desire, and most importantly, the difficulties encountered during
meditation. The ultimate goal of the Bodhisattva is to detach himself from
this world, but Mara, the deity of death and desire, attempted to prevent
the Buddha from Enlightenment. He sent his three daughters (Loving Sex,
Constant Pleasure, and Great Gratification) to seduce the Bodhisattva and
sent his army to threaten him as well. Each one of his soldiers had a differ-
ent kind of body and head: some held scabbards and swords or wore large
trees on their heads; others carried the golden batons; some had the head
of a pig, fish, donkey, horse, lion, dragon, bear, tiger, or some other animal.
All kinds of ferocious-looking soldiers surrounded the Bodhisattva shouting
noisely. The ground shook, but Bodhisattva’s heart was peaceful. He was like
a lion sitting among a flock of deer. After the defeat of Mara, the Bodhisattva
continued in meditation until he was without any doubts.
 Phase three – the remaining splendor 149
First sermon at Sarnath near Benares
The story is one of the four most important episodes in Buddha’s life, and
is usually illustrated in a large niche (as seen in caves 29, 36, and 38).33 The
emblem of the event is the triple wheels flanked by reclining deer as seen on
the south wall of cave 38. In cave 36, only triple wheels are shown. Neither
the five bhiksus (Kaundinya and others) nor the 500 merchants who offered
food to the Buddha right after his Enlightenment are illustrated in the third
phase.

First meditation
The events of the prince’s palace life, such as skill learning, his engagement
and his wedding, etc., are not popular subjects in Yungang; only the first
meditation was favored. The prince’s first meditation is the first sign foretell-
ing the Bodhisattva’s religious vocation, and is often depicted in small caves
(32-11, 32-12, 32-13, 33-3, 33-4, and 35-1). Since the episode took place in the
“plowmen’s village,” the prince usually sits under a jambu (rose-apple) tree
which crooks its branches to provide shade for him. The prince is usually
depicted with a small kneeling figure, his father Suddhodhana, in front of
him as seen in cave 33-3 and the king, startled by the miracle of the branches,
is worshipping the prince whom he had not recognized from a distance.34
It should be mentioned here that many scholars mistakenly identify the
pensive Bodhisattva as the prince’s first meditation.35 Although the pensive
Bodhisattva and the contemplative prince share some iconographic simi-
larities in posture (both figures have one leg pendant and the right hand is
raised to the chin), the first meditation, however, usually has a small kneeling
figure in front of the prince. The pensive Bodhisattva, on the other hand,
is often represented in the side compartments of a trabeated niche flanking
the Bodhisattva with ankles crossed in the center as seen in cave 35. The
first meditation episode is not depicted in the side compartments as flanking
figures and the prince’s head is straight up, not lowered in a downward gaze.
In Gandhara, the first meditation episode is often depicted with plowing
men and oxen. In Longmen, the first meditation motif became very popular,
appearing regularly on either side of the niche front. The prince is often seen
with a kneeling figure, the king, with his entourage holding the canopy for
him. A first meditation tableau must always include an image of his father,
who was looking for his son; otherwise it is just an illustration of a pensive
figure.

Physical training in youth

Elephant riding is not the only depiction of prince’s physical training in
youth. Other athletic contests such as the archery competition and tossing
the elephant are illustrated as well. Both activities, recounted together in
150 Phase three – the remaining splendor
the Guoqu xianzai yinguo jing, are represented together visually, as seen
on the west wall over the pointed arch front in cave 37.36 The illustration
of tossing the elephant here shows the prince and his two cousins, Nanda
and Devadatta. The figure on the south side with the elephant on his out-
stretched right arm is Devadatta, the standing figure pulling the elephant’s
trunk is Nanda and the person in the middle, eroded but still discernible,
holding the elephant to his chest, is the prince. The archery competition on
the south side of the arch front now shows only one bow, one drum and one
figure instead of three (the prince and his two cousins). The athletic contests
are a favorite subject of the Gandhara artists but they do not appear in Yun-
gang. The wrestling contest is not depicted here, and tossing the elephant is
depicted only once. It is especially precious.

Ascent to Tavatimsa to preach for his mother Maya

The event of the Buddha’s ascent to Tavatimsa to preach for his mother
Maya is illustrated only once even though it is an interesting subject. The
narrative is shown on the south wall in cave 38 (fig. 6.9). The subject is rarely
depicted in Buddhist art, so this illustration is of unusual interest.

Fig 6.9 South wall, cave 38

 Phase three – the remaining splendor 151
Parinirvana scene
The narratives are depicted only in caves 35 and 38 since the death of the
Buddha is not a favorite subject in Yungang, as mentioned previously. This
is perhaps because the initial impetus of the excavation of Yungang led by
Tanyao was to disseminate the Buddha Dharma and carry forward the Bud-
dhist doctrines. It was considered too pessimistic to show the death of the
spiritual leader. It is also said that due to indigenous distaste for scenes of
human suffering, the death scene appears infrequently in China.37 In Yun-
gang the Parinirvana scene is thus far less frequent than other important
events in his life.38

Jataka tales
The core tenets of karma and reincarnation in Buddhism constitute an
important part of all Buddhist iconography. The Wheel of Life shows vari-
ous realms, in which living beings can be reborn through karma, which is
the result of willful intention, and good or bad actions are what determine
karmic outcome. Therefore, rebirth to various realms of existence from
animal to human to divine depends on the good or bad actions in one’s
previous life. With these key concepts in mind, Chinese artists favored the
subject of the jataka stories of the Buddha’s previous lives, and among
the 17 jataka motifs in Yungang, four are illustrated in the third phase
including: the Buddha’s revisit to his hometown Kapilavastu and his son
Rahula, the Gridhrakuta Hill episode, the Asoka story, and the Dipamkara
legend.39

The Dipamkara legend

The emphasis on reincarnation made the Dipamkara legend a much-
depicted subject in Buddhist art. Tradition has it that Dipamkara was the
first Buddha of an earlier cycle of the universe in a distant epoch, many
thousands of millions of years ago. The legend of the Gautama Buddha
usually begins with the meeting of the Dipamkara Buddha with the future
Buddha, a Brahmin youth then named Sumadhi who was just a Bodhisattva
living a secluded life in the mountains meditating. When the Dipamkara
Buddha came to the city, the boy spent 500 pieces of silver, which was all
he had, to buy five flowers from a girl to present to the Buddha. When the
Bodhisattva approached the Buddha to toss his five flowers, they stopped
in the air and stayed suspended there above the Buddha. The Buddha who
knew the reason for this, said to the Bodhisattva: “for innumerable kalpas,
you have practiced the way of a pure recluse. You disciplined your heart,
rejected life, you cast off desire, kept your mind empty with no desire,
with mercy even for a mad dog. Because you accumulated virtue and held
fast to your vows, now you have achieved the goal. Thus, know that you
152 Phase three – the remaining splendor
yourself will be a Buddha by the name of Śākyamuni.”40 The Bodhisattva
heard these words and his questions were answered, his desire stopped. His
body became light and he ascended to kowtow at the Buddha’s feet. When
he saw the ground wet with puddles, he laid out his deerskin clothing to
cover them. Seeing they were insufficient to cover the puddles, he spread
out his hair on the ground and told the Buddha to tread on it and walk
by.41 The Buddha said: “Your spirit had progressed undauntingly, you will
be a Buddha.”42 The Buddha Dipamkara promised that all sentient beings
by practicing the six paramitas will be able to attain Enlightenment and
enter nirvana. The spreading of the hair story is a promise of achieving
one’s ultimate goal. After the last previous lives on earth, the Bodhisattva
was reborn among Tusita gods where he enjoyed divine bliss, resting and
relaxing. The legend is easily identified by the standing Buddha and the
learned Brahmin youth spreading his deerskin cloak and his own hair over
the marshy ground for the passage of Dipamkara, as seen on the south
wall of cave 35.43

Asoka jataka
The Asoka jataka is often depicted in a large niche on the south wall.44 The
illustration usually shows the Buddha holding a bowl and slightly bending
over towards three boys, squatting or standing, who are trying to offer earth
named “millet” to the Buddha. The boy standing on the shoulder of another
boy and closest to the Buddha is King Asoka. The illustration is without a
doubt based on the Xianyu jing (the Sutra of the Wise and the Foolish).45
Among the Buddha’s various previous reincarnations from animals to human
being, and tales of his extreme compassion and self-sacrifice, the Dipamkara
jataka and the Asoka jataka are most favored subjects in Yungang. The
Dipamkara prophecy promises us a future Buddha when the present Bud-
dha enters final nirvana, whereas Asoka promises that good karmic practices
will result in favorable reincarnation. The narrative pictorials emphasize
the similarity between King Asoka and Chinese emperors who therefore
could validate their claim to resemble King Asoka who earned his favorable
rebirth by practicing karmic deeds. The story implies the close relationship
between politics and religion, and politics and art. Rulers, particularly in
the Northern Dynasties, attempted to gain good karma and to imitate King
Asoka as the Dharma king by extending their support to Buddhist activities
and Buddhist sanctuaries.
In the Asoka story, it says that when Buddha was living in the city of
Sravasti at the Jetavana monastery in Anathapindika’s park, upon a certain
occasion, the Buddha and Ananda were traveling on the alm-round, and
there were some children playing along the roadside, building toy treasure
houses. One of the children, seeing the Buddha approaching, rejoiced, and
 Phase three – the remaining splendor 153
thinking that he would acquire merit, took a handful of earth and tried to
offer it to the Buddha, but because he was so tiny could not reach the alms-
bowl and asked another boy to let him climb upon his shoulders. The boy
did this and when the Buddha lowered his bowl, he dropped the handful
of earth into it. The Buddha accepted it, and handed the bowl to Ananda,
and said,

Take this earth to the monastery and scatter it. Ananda, a hundred years
after I have attained final nirvana, this little boy, by virtue of having
made an offering in great joy and by virtue of its having been scattered
at the monastery, will be reborn as the Emperor Asoka. The child who
let him climb on his shoulders will be reborn as a minister. The emperor
will rule all India and will make known to all the blessings of the Three
Treasures. He will venerate my relics by erecting eighty-four thousand
stupas.46

Rahula jataka
The depiction of the Rahula jataka is usually recognized by the Buddha pat-
ting a little boy’s head. In cave 19, the standing Buddha is patting the kneel-
ing boy’s head with his left hand and boy is seen to his left side. In cave 9,
the kneeling boy is in front of the Buddha. In cave 38, the story is illustrated
twice and in both cases the Buddha is seated and the boy is kneeling down
to the left of the Buddha. Evidently the Yungang artists have created their
own iconography rather than strictly following the doctrine,47 in which,
according to the record of the Zabaozang jing, the boy is usually standing
to the left side of the Buddha.

Gridhrakuta Hill episode

The vivid depiction of the Gridhrakuta Hill episode on the south wall in
cave 38 shows the Buddha and Ananda seated side by side in two niches
on Gridhrakuta Hill, where the Buddha used to dwell for some 50 years
and deliver the Dharma, with a huge vulture behind them. The Buddha
is, through the wall, patting the head of Ananda who is frightened by the
vulture (fig. 6.10).48 The story is only depicted once, and is a rare scene in
Buddhist art.
In summary, from the above discussion we can conclude that the life of
the Buddha and his previous lives constitute a major part of the subject
matter in Yungang. Together with the eternal theme of the three Buddhas,
the Śākyamuni and Prabhutaratna, the discourse between Vimalakirti and
Manjusri, and the Dharma protectors, etc., enrich the iconography of Yun-
gang and present a kaleidoscopic imagery.
154 Phase three – the remaining splendor

Fig 6.10 South wall, cave 38, Gridhrakuta Hill

Iconographic style
Stylistically, the images in the third phase do not appear as grandiose and
masculine as those in the Tanyao cave temples, nor do they appear as power-
ful and plump as the images of cave 6. Rather, they look slender, elegant, sup-
ple, and lifelike with amiable smiles on their faces. The imagery demonstrates
a new visual representation, a sinicized visual form. However, because they
were made in different times during the third phase, the images of Buddha,
Bodhisattva, and flying celestials display a clear transformation in style. Early
images appear slender but with a sense of roundness, e.g., the face is slightly
rounded, neck is shorter and the body is fuller, whereas later images show
elongated face and neck with slightly sloping shoulders. The bodies became
more slender; some were even elongated to create an exaggerated visual effect.
 Phase three – the remaining splendor 155
All Buddha images in the third phase are clad in a loose robe and a wide
girdle (baoyi bodai) with both shoulders covered. The early popular style
with only one shoulder covered, seen in the imperial five caves, disappeared
completely. The early Buddha images (e.g., cave 21), as observed in those
in cave 6, are with a “V” shaped neckline, but the late images mainly show
a “U” shape. The cordlike raised folds and incised lines observed in the
imperial five and early second-phase cave temples are not found during this
time. Rather, we only see step-like folds, some flattened, and others bulgy.
In general, the early third-phase images have flattened and well-balanced
step-like folds, as seen in caves 21, 22, 23 and 28, 33, 25, 27, and 29, etc.
The later images have bulgy and uneven folds, which is best exemplified in
the seated Buddha image, dated to the fourth year of Yanchang (515) in cave
19-2 (fig. 6.11) and those in caves 27-2, 5-10, 5-11, 33-3, and 40-1. The imagery
before 515 contrasts considerably from that made afterward.

Fig 6.11 Cave 19-2, dated 515

156 Phase three – the remaining splendor
The drapery of the Buddha image is generally rendered in two ways: one
large series of arcs in the center covering both feet, and two series of arcs
side by side covering both legs but leaving the right foot exposed. Gener-
ally speaking, the earlier images have only one large arc and the later ones
have two arcs; and the longer the drapery falls down, the later the rendition
is. The drapery from the last stage of the third phase falls down low like a
waterfall. But the one large arc style did not disappear entirely when the style
of two arcs emerged. For example, the same drapery style with one arc can
be observed in both the image in the east reveal of the window of cave 11
dated to the 19th year of Taihe (495) (see plate 30), and in the seated Buddha
images dated to the 13th year of Taihe (489) in cave 11-14 (see plate 11). The
two-arced style is best exemplified by the seated Buddha in cave 34.
The Bodhisattvas in the third phase appear slender and graceful with long
neck and narrow waist. The heaviness of the Bodhisattvas with short neck
and thick skirt shown in cave 6 is not seen. Rather, we see supple and elegant
figures with fluttering celestial robes, which form an “X” in the center and
taper from the shoulder area down to the end. The left hand usually holds
one end of it and the other hangs over the right arm. The end of the skirt
undulates, giving a strong sense of movement. The earlier Bodhisattva does
not have a large ring in the center for a celestial robe to cross through.
The flying celestials in the third phase look supple and elegant with long
drapery fluttering backwards. The worshippers are slender and elegant as
well, with long necks and high chignons.

Classification, dating, and chronological sequence

There are very few inscriptions in the third-phase cave temples, and they are
mainly in small niches that were added to the caves later, thus the inscrip-
tions often lack direct associations with the excavation of the cave temples.
Therefore in the following, the classification of the caves, the assignment
of dates and the construction of a chronological sequence will be primar-
ily based on the above discussion and analysis of cave structure, as well
as iconographic composition and style. Furthermore, the investigation will
also be based on my field observation of the spatial relationships between
the caves themselves, and between caves and niches: some caves make space
for others (birang 避讓) and some encroach on others (dapuo 打破). This
is an important approach for identifying the time sequence of the excava-
tions. The spatial relations between cave temples are useful in constructing
a chronological sequence.
It must be pointed out that iconographic style should not be used as the
sole criteria to construct a dating scheme. However, its significance becomes
more prominent when there are not many written records or inscriptions.
In such cases, iconographic style becomes more valuable and even indispen-
sible evidence. Although certain stylistic features sometimes continue, in gen-
eral, each period exhibits its own style. Based on field research and critical
 Phase three – the remaining splendor 157
analysis, the third-phase cave temples will be classified into four groups and
examined in chronological sequence. This is the very first comprehensive
study of these cave temples thus far.

1. Group one
The first group includes 13 caves (caves 21, 31, 39, 22, 23, 33, 25, 28, 24,
27, 29, 30, and 28-2). These caves resemble one another in style, yet have
distinguishing details. They are chronologically close to each other.

Cave 21
The cave, rectangular in plan, is situated to the west of cave 20 and is the
largest of the third-phase caves. The refined workmanship suggests that
the cave was patronized by a high official. Stylistically, the Buddha images,
though damaged, appear fuller, with broad chests. The seated Buddha on the
east wall resembles closely the image dated to the 19th year of Taihe (495) in
the east reveal of the window of cave 11. The drapery has only one arc in the
front instead of two. The seated Bodhisattva with ankles crossed resembles
closely that of cave 11-7 on the east wall. It should be noted that the cave was
not finished. Soon after the north and east walls were completed, the con-
struction was stopped for unknown reasons. It was perhaps related to the
transfer of the capital in 494, which may have caused the patron to move to
the south. Nevertheless, cave 21 is the earliest cave in the third phase and the
construction could have been started as early as before the transfer.

Cave 22
This is a small cave. Only the north wall is preserved. The seated Buddhas
on the north wall closely resemble those in cave 11-14. The round face with
gentle smile, the smooth usnisa, and the rendition of the robes all appear
similar. The Buddha images also resemble those in cave 21. The iconographic
similarity suggests that cave 22 is close to cave 21 in time and is also one
of the earliest caves. The east and west walls, as well as the south wall, are
damaged.

Cave 23
Cave 23 has no niches. It should not be far in time from caves 21 and 22
judging from the style of the seated Buddhas on the north and east walls.
The drapery of the seated Buddhas on the east wall still has the archaic single
arc in the front, covering both legs and feet. The standing Buddhas on the
west wall look similar to the seven Buddhas on the west wall in cave 11, so
the excavation dates should be close to each other in time. In overall shape
and attitude, the images in these two caves resemble one another. The flying
158 Phase three – the remaining splendor
celestials are clumsy compared to those made later, which are supple and
elegant with a great sense of movement.

Caves 27, 29, and 30

Caves 27 (see fig. 6.2), 29 (see fig. 6.3), and 30 (see fig. 6.4) are among the
larger of the third-phase caves. They resemble one another closely in subject
matter, style, and structure with storied niches on three walls. This makes it
appear likely that they were carved from the same design and by the hands
of the same artisans. Caves 27 and 29, like caves 7 and 8, are likely to have
been excavated as paired caves, judging from the architectural structure,
arrangement of the niches, and iconography. The images on the sidewalls of
cave 27 (plates 59 & 60) and cave 29 (plates 61 & 62), which are divided
into two stories and contain four niches, are represented delicately, and drap-
ery is exquisitely rendered with two series of arcs side by side. These images
are considered the masterpieces of the western-end caves and well represent
the third-phase work of Yungang. The masterpiece among masterpieces is
the image housed in the upper story canopied niche in the west wall of cave
30 (see plate 19). It is a perfect visual representation with well-proportioned
body and refined rendition. The image, gazing into the distance, appears
calm, graceful, and stately. It should be noted that the Asoka narrative story
is depicted on the south wall of cave 29.

Cave 31
This large, badly eroded cave is the only one in the third phase that has both
an anteroom and a main room. There is a small cave dug into the west wall
of the anteroom. The images appear slender and archaic, and the seated
Buddhas are reminiscent of those dated to the 13th year of Taihe (489) of
cave 11-14. Another piece of evidence, besides the architectural and stylistic
indications, to prove that this cave is earlier than many other third-phase
caves is that the Dvarapalas on each side of the entrance closely resemble
those on cave 9.

Cave 33
Cave 33, a sanbi sankan (three niches on three walls) cave (fig. 6.12), is the
only cave that contains seated Buddhas with ankles crossed, rather than
seated Bodhisattvas with ankles crossed as is usual, and the disciples are
represented beside the Bodhisattvas, not the Buddha. The representation
of the Buddhas is refined with flattened and well-balanced folds, smooth
texture, and soft material. The solidness of the images in the cave is remi-
niscent of those in cave 6 but the rendition is more natural. The facial fea-
tures of the disciple Mahākāśyapa are striking with high cheekbones, bulgy
forehead, deep eyes, and exaggerated wrinkles on his neck. The gentle
 Phase three – the remaining splendor 159

Fig 6.12 Plan, cave 33

smile makes this image lifelike. In a similar manner, the Bodhisattva looks
kind and humane with a mild smile. Stylistic features and iconographic
similarities to earlier images suggest that cave 33 is one of the early caves
in the third phase.

Cave 25
Cave 25 is a typical sanbi sankan cave with a square floor plan. The images
appear slender and graceful; in particular, the standing Buddhas in the south
wall and the attendant Bodhisattvas on the sidewalls. The seated Buddha
with long neck and slender face on the east wall is reminiscent of those in
cave 29. It is worth noting that the Bodhisattva on the south side of the east
wall has a ring in the abdomen to have the celestial robes to cross through
it. The ring is not, with the one exception of cave 33, represented in any of
the aforementioned caves. We may thus infer that cave 25 was excavated
later than those caves.
160 Phase three – the remaining splendor
Cave 28
Cave 28 is a typical sanbi sankan cave. It shares close similarities with cave
25 in cave structure and iconography. The two caves are possibly close in
time. The flattened and wide folds of the robes of the two Buddhas on the
north wall accentuate the elegance of the images. The inner robe covers
only one shoulder and the sash is represented with a bowknot with stream-
ers hanging down outside of the right hem of the robe. The rendition of the
robes in cave 28 proves the mastery of sculpting by Chinese artisans at that
time. The images can be considered representative works of Yungang third-
phase caves.

Cave 24
Cave 24 contains three niches on three walls (fig. 6.13). The south wall is
damaged. The west and east walls contain twin Buddhas, and the north wall

Fig 6.13 Cave plan, cave 24

 Phase three – the remaining splendor 161
bears a canopied niche, both iconographically rare. The images appear slim
and elegant with long necks and sloping shoulders. The overall rendition is
smooth and natural, and the seated Buddhas on the west wall, in particular,
appear graceful and supple. The drapery has two series of arcs side by side,
however, the main image on the north wall has one series of arcs in the center
of the drapery which falls down over the pedestal and the hem flares out
forming an angular 八.

Cave 28-2
Cave 28-2, situated above cave 28 on the right, is not finished. The seated
Buddha shares considerable iconographic similarities with the masterpiece of
cave 30 and the images in caves 27, 29, and 24, except that the workmanship
is not as refined as in those. The overall impression given by this image is
stately and powerful; it has a well-built body and a full face. The Bodhisat-
tvas on both sides appear slender and animated with one hand holding the
celestial robe and the other raised to the chest.
It should be noted that cave 28-2 has two inscriptions with explicit dates.
The inscription on the east wall niche (fig. 6.14) is dated to the fourth year of
Zhengshi 正始 (507) and the one on the west wall niche (fig. 6.15) is dated to
the third year of Yanchang 延昌 (514).49 The images in the niches are crude,

Fig 6.14 East wall niche, cave 28-2, dated 507

162 Phase three – the remaining splendor

Fig 6.15 West wall niche, cave 28-2, dated 514

and do not seem to have been completed, thus cannot be used as typological
evidence. But the inscription can at least prove that the main image in the
cave was started before 507, or the latest in 507.
More significantly, cave 28-2 encroaches onto the Bodhisattva in the south
niche of the west wall in cave 27. This not only indicates that cave 27 is
earlier than cave 28-2, but also accordingly provides an important clue for
constructing a chronological sequence of the caves in this group. The cave
structure, subject matter, and style in caves 27, 29, and 30 are all the same
and belong to the same time period of construction. Cave 28 is situated
between caves 27 and 29. From the observation of the outside cliff, it is clear
that cave 28 used the space left by the other two. In turn, cave 28-2 used the
space on top of cave 28. Therefore, we can infer that cave 28 is earlier than
cave 28-2 but later than caves 27 and 29.
The images in caves 28 and 25 share similarities in the rendition of the
robes and the arrangement of the draperies. They should have been exca-
vated within the same time period. The images in caves 22 and 23 resemble
 Phase three – the remaining splendor 163
the images in cave 21, although they do not appear as full, but they should
have been constructed close in time.
The anteroom in cave 31 was not finished as a conceived plan. The
arrangement of the niches and the subject matter of the north wall in the
main chamber are similar to those in cave 12. Although the images are
severely eroded, they display iconographic similarities with caves 27 and 29.
Based on the analysis above, I wish to argue that the caves in this group
were excavated before the fourth year of Zhengshi 正始 (507), approxi-
mately from before the transfer of the capital in 494 (cave 21) to the first year
of Zhengshi (504). The earliest in this group is cave 21, followed by caves
22, 23, 33, 27, 29, 30, 25, 28, 24, and 28-2 in succession.

2. Group two
The second group includes only four caves: 34, 35, 37, and 39. The imagery
in these caves appears different from that in earlier caves, transformed from
slender and elegant to slightly solid and fleshy.

Cave 34
Cave 34 is a sanbi sankan cave (the east wall is damaged). The seated Bud-
dhas on the north wall appear solid and robust compared to the images in
caves 27, 29, and 30, but the seated Buddha on the west wall appears more
elegant particularly with a more refined rendition of the robe. The step-like
folds are flattened and rhythmical. The drapery is in the two parallel series
of arcs as seen in the images in caves 27 and 29, etc.

Cave 35
Cave 35 (fig. 6.16) is a sanbi sankan cave with square plan and window.
Two Dvarapalas outside the entrance and Dharma protectors on the reveals
of the window are depicted. The overall impression of the images is solid
and chunky with large series of arcs. The cross-ankled Bodhisattva in the
east wall trabeated niche, adorned with necklaces, earrings, and crown, is
the most elaborate image in Yungang (fig. 6.17). A series of sharp arcs is
rendered on both legs and gives a strong sense of rhythm. The two standing
Buddhas on the south wall look fleshy and tall, disproportionate with small
heads and broad chests. On the other hand, the seated Buddhas appear
slender with sloping shoulders and long neck compared with the images on
the sidewalls.
It should be noted that the cave entrance bears an inscription dated to the
fourth year of Yanchang (515) (plate 63). It is thus clear that cave 35 was
executed before that date.
Fig 6.16 Plan, cave 35

Fig 6.17 Cross-ankled Bodhisattva, east wall, cave 35

 Phase three – the remaining splendor 165

Fig 6.18 Plan, cave 37

Cave 37
Cave 37 (fig. 6.18) is a unique cave in iconography and subject matter, but is
the same in structure as cave 35 with three niches on three walls and a win-
dow, the reveals of which bear the Dharma protectors. The main image is a
rare one: a cross-ankled Bodhisattva. The subject matter is richly diversified,
and some of the themes (e.g., the prince tossing the wild elephant depicted
on the west wall) are depicted only in this cave. The images appear solid and
stern with round face, full cheeks, and voluminous drapery hanging down
low over the pedestal.

Cave 39
Cave 39 is the only caitya cave in the third phase and bears Dharma pro-
tectors on the ceiling. Judging from the arrangements of the niches on the
south wall, it is assumed that the cave was not completed as it was initially
166 Phase three – the remaining splendor
planned, and only the stūpa and the ceiling were finished. The thousand
Buddha niches and the south wall niches were added later. Judging from the
styles of the images, the cave was completed within a short period and was
executed early in the third phase.
As discussed above, the Buddha images in cave 34 resemble those in
caves 27 and 29. The iconographic style in cave 39 is the same, so the time of
construction is close. Cave 35 was constructed before the fourth year of Yan-
chang (515). The sizes of caves 35 and 37 are the same and the cave structures
are the same, with three niches on three walls and the south wall with storied
niches. Their iconographic composition is similar as well: each has Dharma
protectors on the reveals of the window, two standing Buddhas, and Vimala-
kirti and Manjusri on the south wall. At the same time, there are differences
reflecting the complementary nature of the two caves. For example, in cave
35, on the north wall a seated Śākyamuni Buddha appears, whereas in cave
37 a seated Bodhisattva with ankles crossed appears instead, so that together
the caves represent Śākyamuni and Maitreya. The sidewalls of cave 35 show
a Bodhisattva with crossed ankles and a seated Buddha with legs pendant,
and the sidewalls in cave 37 contain seated Buddhas, forming the same icono-
graphic complementarity (i.e., the Śākyamuni Buddha and Maitreya). It can
therefore be assumed that the two caves were based on the same design, with
images from the same prototype, so the possibility that they are paired caves
cannot be ruled out. Hence, the chronological sequence for this group is cave
34, caves 35 and 37, and cave 39. The dates are approximately between the
first year of Zhengshi (504) and before the fourth year of Yanchang (515).

3. Group three
Caves 5-10 (see fig. 6.6) and 5-11 (see fig. 6.7), hidden deep in the mountain
above cave 5, are unique and far from the other caves, and thus are classified
separately as a group. They are the most delicate and refined examples of
workmanship in the third-phase caves; in particular, the seated Bodhisattva
with ankles crossed in cave 5-11.

Caves 5-10 and 5-11

The caves are situated on both sides below the window of cave 5 on the
outside cliff. The cave structures and arrangements of the subject matter
are the same (both have two niches on three walls and multi-armed deities
outside, and both have three Buddhas). Possibly, they are paired caves. The
structure of the caves with two niches on three walls might be related to the
thickness of the cliff; they had to be excavated without a north niche. Even
so, the north wall of cave 5-11 encroached on the south wall of cave 5 (there
is a huge hole in between). Each main Buddha (plate 64) appears graceful
and dignified. The generous drapery hangs down low over the pedestal and
the edges flare out at each side. The west wall Buddha (plate 65) appears
 Phase three – the remaining splendor 167
even more slender and elegant with sloping shoulders and long neck, and the
east wall contains a cross-ankled Bodhisattva in a trabeated niche, the front
of which is the so-called folding screen. The south wall bears two standing
Buddhas and the narrative stories in the life of the Buddha.
Cave 5-11 is perhaps one of the most refined caves in all of Yungang. The
main Buddha is draped in a thin robe with both shoulders covered. The
folds are dense and flattened, and the neckline is double hemmed. It is rich
and rhythmical like a waterfall flowing over the pedestal. The west wall
also contains the depiction of the debate between Vimalakirti and Manjusri.
The seated Bodhisattva on the east wall is a masterpiece of workmanship.
The image appears extremely elegant and stately with a well-proportioned
slender body and gentle expression on the face. The south wall echoes the
layout of the south wall in cave 5-10.
The images in these two caves closely resemble the image in cave 19-2,
dated to 515 (see fig. 6.11). The only difference is that the cave 5-11 images
are more refined works. This suggests that the two caves were constructed
approximately at that date (before or after). The celestial robes of the Bod-
hisattvas with an “X” in the front and the trabeation adorned with folding
screen resemble those of the Bodhisattva in cave 35, which was excavated
before 515. This further suggests that these two caves are later in time than
cave 35, and were excavated around the fourth year of Yanchang (515).

4. Group four
This group includes 17 caves: 22-1, 23-1, 24-1, 25-2, 26, 27-2, 32-11, 32-12, 32-9,
33-3, 33-4, 33-6, 36, 36-2, 38, 40, and 41. The caves here are smaller and
more numerous. The cave structure is mostly the same: three niches with
three walls (22-1, 23-1, 27-2, 32-11, 32-12, 33-6, 40, and 41), or one niche on
north wall and sidewalls with storied niches (24-1, 26, 33-3, 33-4, and 38).
The iconographic composition and style are the same or similar. Therefore,
in what follows, I shall not discuss the caves one by one, but shall instead
analyze several representative caves only.

Cave 23-1
Cave 23-1 is a small and elegant cave with one pointed arch in the north
wall and a trabeated niche in each sidewall. The rest of the space is filled
with thousand Buddha niches neatly arranged. The seated Buddha in the
north wall appears serene and stern, but the body is disproportioned with
small head and heavy drapery. The Bodhisattvas look slender and elegant.

Cave 27-2
Cave 27-2 is a small cave with three niches on three walls. The main Buddha
on the north wall appears slender. The trabeated niche with folding screen on
168 Phase three – the remaining splendor
the east wall contains a cross-ankled Bodhisattva, disproportioned with large
torso and short legs. The robe of each Buddha, carved in bas-relief on the fold-
ing screen, is unique. Instead of covering the left shoulder, as is usual, it covers
the right shoulder with incised vertical lines rendered on the robe. This unique
rendition is also represented in all the so-called folding screen trabeations and
some of the arch fronts (such as in caves 24-1, 26, 35, 38, 5-10, and 5-11).

Cave 33-3
Cave 33-3 is a small cave with storied niches on the sidewalls and one cur-
tained niche on the north wall. The images resemble those in cave 27-2 and
are crudely carved. The two Buddhas on the main wall appear slender and
elegant. The rich drapery hangs over the pedestal voluminously and resem-
bles those of caves 5-10 and 5-11. The seated Bodhisattva with one leg pendant
under a tree and a kneeling figure in front of him in the lower level of the
west wall depicts the First Meditation of the prince.

Cave 24-1
Cave 24-1 contains slender and elegant images. The two Buddhas on the
main wall look slender and dignified with sloping shoulders and long necks.
The attending Bodhisattvas appear elegant and humane with a gentle smile.
The west wall contains two stories. The vertical composition of the images
in the west wall is noteworthy as previously mentioned. It is a rarely seen
composition: that of a seated Buddha with both legs pendant, a seated
Buddha with both legs crossed, and a standing Buddha. The niches on the
sidewalls are arranged to correspond to each other.

Cave 26
Cave 26 resembles cave 24 closely in style, structure, composition, and sub-
ject matter. The main wall contains a trabeated niche in which two seated
Buddhas side by side are attended by two Buddhas with both legs pendant,
a rare representation, observed elsewhere only in cave 7, the anteroom of
cave 12, and caves 27, 29, and 30.

Cave 36
Cave 36 is a small rectangular cave with storied niches on three walls. The
seated Buddha appears stately and dignified, and the workmanship is refined.
The outer robe neckline is in a “V” shape with a thick bottom across the
front to hang over the left arm forming the series of arcs in the abdominal
area. The First Sermon episode in the upper niche on the west wall is por-
trayed. The seated Buddhas with legs crossed in the upper level niches on the
east wall are similar to those in the sidewalls of cave 27, slender and serene
with flat and well-balanced folds.
 Phase three – the remaining splendor 169
Cave 33-4
Cave 33-4 is a small cave. The north wall bears only a pointed-arch niche
inside of which twin Buddhas are represented. The sidewalls contain storied
niches with compact composition. The motifs are diversified, and the Asoka
episode is represented twice. The seated Buddha on the left wall resembles
the images in the east wall of the entrance gateway of cave 35 (dated to 515)
and the image in the west wall of cave 28-2 (dated to 514).

Cave 38
Cave 38 (fig. 6.19) is so small and inaccessible that it is difficult to photo-
graph a full view of each wall, but it contains subject matter that is among
the richest and the most diversified of all the caves of Yungang. Some motifs
represented in this cave (Buddha’s ascent to Tavatimsa to preach for his
mother Queen Maya and the Gridhrakuta Hill episode) are found nowhere
else. The north wall contains a pointed-arch niche and many narrative pic-
torials of Buddha’s life (the Great Departure, the Prince Riding an Elephant,
the Buddha’s Meeting with His Son Rahula, and the Parinirvana). The east
wall contains three stories (the Asoka narrative and the Dipamkara legend
on the second level), and the west wall contains only one niche, inside of
which is a seated Buddha with legs pendant. The south wall is filled with
episodes in the life of the Buddha (the Conversion of the Three Kāśyapas,

Fig 6.19 Plan, cave 38

170 Phase three – the remaining splendor
Buddha’s Ascent to Tavatimsa to Preach for His Mother, Gridhrakuta Hill,
Mara’s Assault, and Buddha’s First Sermon).
In sum, the caves in this group are numerous, but were excavated mainly
in the space left between the caves of the first two groups. For instance,
cave 38 was excavated using cliff space allocated for cave 39, and therefore
should be dated later than cave 39; also, cave 22-1 was excavated using the
space above and between caves 22 and 21. Thus, without a doubt, this group
is later than the first two groups. Cave 39 is at the west end of the Yungang
mountain body, beyond which the stratum of rock is bad and not suitable for
excavation, so caves 40 and 41, etc., were hewn here only because there was
no available space to the east of cave 39. Hence, these caves should belong
to the latest group of the third phase.
Also, the iconographic style in all of cave 27-2 is similar to the south wall
of cave 35. This indicates that the construction date is close, so it should also
have been excavated around the fourth year of Yanchang (515), or a little bit
later. Furthermore, the trabeated niche front with folding screen is the same
as in caves 24-1, 26, 38, 5-10, and 5-11. This style is also seen in cave 35 in the
trabeated niche on the east wall and on the east entrance niche which is dated
to 515. The seven Buddhas on the north wall in cave 36-2 are exceedingly
similar to the Buddha in cave 4,50 which bears an inscription dated to the □
year of Zhengguang (520–24).51 It may thus be inferred that the excavation
date of cave 36-2 is not too far from this.
It is worth noting that some of the caves in this group were not completed.
For example, in cave 36 the carving is crude in some images, and hastily
done; and in cave 33-4, the east niche images are refined but those in the west
niche have hardly any carved lines on the robes. These phenomena suggest
that these caves are near the end of the construction of Yungang. In brief,
the iconographic style in this group is extremely similar, the excavation dates
were close, and should belong to the last phase of Yungang (515–24). Among
them, 23-1, 27-2, 32-12, 32-9, 32-11, 33-3, and 35-1 are probably early, and 24-1,
26, 36, 33-4, 38, 40, and 41 belong to the final phase of the caves.

Concluding remarks
As discussed above, the third-phase caves have been divided into four groups
and three time periods of construction based on the perspective of archaeo-
logical periodization. Dating and chronological sequence are two important
aspects of Buddhist art study. It is fundamental in the examination of Bud-
dhist rock-cut cave temples. The features of each period in the third phase
can be summarized as follows:
The first period, i.e., the first group, is located in the eastern section of
the west-end caves (21 to 31). They are mainly medium-sized, and are the
representatives of the exquisite craftsmanship in the third phase, especially
caves 27, 29, and 30. They can be called masterpieces of the third phase.
The patrons were probably higher-level officials. This group of caves was
 Phase three – the remaining splendor 171
excavated approximately at the beginning of the transfer of the capital, but
cave 21 was probably constructed before then. The cave structure is pri-
marily storied niches on three walls, continuing the cave structure of caves
7 and 8; there is also cave structure of one niche on the north wall with
storied niches on sidewalls; the cave structure of three niches on three walls
(sanbi sankan) emerged. The structure of the niches is a continuation of
the second phase with pointed and trabeated niches. The curtained niche
emerged (on the west wall of cave 30). The Śākyamuni and Prabhutaratna
become iconographically popular and are the main images used for the sub-
ject of the three Buddhas. There even appeared the unique composition of the
Śākyamuni Buddha on the north wall and Śākyamuni and Prabhutaratna on
the sidewalls (cave 24). This implies that the Lotus Sutra was widely popu-
lar and deeply rooted in the heart of the people. Next is the composition of
Śākyamuni (upper niche) and Maitreya (low niche) expressing the passing
on of the Buddha Dharma. The iconographic style continued that of cave 6
and caves 1 and 2, but appear more slender and elegant. The neckline of the
robes is mainly in a “V” shape, and the drapery is short. The “U” shaped
neckline began to appear and the drapery has two arcs. The celestial robes
of the Bodhisattva has an “X” crossing in the front.
The second period, i.e., the second group caves (37, 35, 34, and 39) are
mainly situated at the western section of the western end. They are larger in
scale, thus the images are larger than those in other caves. The excavation
dates of this group of caves were before the first year of Zhengshi (504) up
to the fourth year of Yanchang (515). The patrons, as with the first-group,
were higher level officials. Sanbi sankan is the popular cave structure. Caves
35 and 37 are probably paired caves. The caitya cave (39) began to emerge.
The central pillar is in imitation of wooden pagoda with nine stories. Com-
pared with caves 1 and 2, it is more like the structure of a real pagoda, mod-
eled after the monasteries in Pingcheng. The niche structure of this group
is the same as the first group, but the trabeated niche front with the folding
screen (cave 35) began to emerge. At this time, there appeared Maitreya
Bodhisattva (cave 37) as the main image in the iconographic composition
of the three Buddhas, as well as Śākyamuni and Prabhutaratna (cave 34) as
the main image. There also appeared the iconographic composition of three
Buddhas in the form of Śākyamuni as the main image flanked by seated
Buddha with legs crossed and seated Bodhisattva with ankles crossed. The
latter is the representation of Maitreya’s ascent toward, and descent from,
the Tushita Heaven, a manifestation of passing on the Buddha Dharma in
three phases. The appearance of the disciples as the attendants of the Bud-
dha is significant. Stylistically, this group followed the older tradition, but at
the same time, innovations were made. The images appear robust and solid
with broad chests and full faces. The workmanship is not as refined as the
previous group.
The third period includes the third and fourth groups. These two groups
are close in time. All caves were excavated between the fourth year of
172 Phase three – the remaining splendor
Yanchang (515) and the fifth year of Zhengguang (524). These two groups
are mainly small cave temples, and were excavated using the leftover cliff
space. The excavation was extended to the west of cave 39. The social status
of the patrons was apparently lower than those of the first two groups. They
were middle or low officials, or ordinary devotees. But caves 5-10 and 5-11 are
exceptions. The images in these caves are refined sculpture and can be con-
sidered the masterpieces of the group. The patrons possibly had high social
status and were able to invest abundant financial resources and use excellent
craftsmen. The sanbi sankan and a design with one niche on the north wall
with storied niches on the sidewalls was popular in this group, but caves 5-10
and 5-11 are different due to the thickness of the north walls; also, the design
of the cave includes two niches with an altar on the north wall. The trabe-
ated niche with folding screen is a prominent and popular characteristic of
this group. Iconographically, Śākyamuni and Prabhutaratna are primary,
and Śākyamuni and Maitreya is often seen as well. The iconographic style is
elegant and slender with some images elongated. Overall, the workmanship
is not as exquisite as the first two groups; some are even unduly crude, with
disproportioned bodies and rough lines. All of these imply that the excava-
tion of the caves in Yungang had begun to decline. This perhaps was related
to the fact that after the southward transfer of the capital to Luoyang, a large
number of excellent craftsmen moved south.
In the end, mention must be made of the relationships between scrip-
tures and visual images, and image-making and politics, and in particular,
the sacred texts translated by Tanyao and others in Yungang. The eminent
monk Tanyao from Liangzhou had just experienced the disastrous persecu-
tion of Buddhism by Emperor Taiwu and was appointed as Shamentong
after Buddhism was revived. After the painful persecution, Tanyao deeply
felt that without the support of the court it would be impossible to carry
the Buddha Dharma forward in an unbroken fashion. It was due to such
feelings that he persuaded Emperor Wencheng to carve five colossal images
of the five emperors of the Northern Wei on Mount Wuzhou. While he
was supervising the excavation of the caves, Tanyao gathered many eminent
monks in the Beitai caves (Yungang) to translate sutras, among which, the
Fufazang yinyuan zhuan and Dajiyi shenzhou jing especially emphasize the
significance of passing on the Buddha Dharma.52 The most popular subject
matter of the three Buddhas in the caves is evidently related to the influ-
ence of these two scriptures. The Lotus Sutra and the Vimalakirti Sutra also
inspired image-making in the rock-cut cave temples. The Śākyamuni and
Prabhutaratna are the primary Buddha images in the third-phase caves, and
the debate between Vimalakirti and Manjusri is illustrated in almost every
cave. They became the visual symbols of the sutras. One can well perceive
the profound influence of the two classics.
The subject matter of the imagery in Yungang reflects the social and politi-
cal climate, as well as ideology. The images in the first-phase caves were
modeled after the five founding emperors to validate the claim by Faguo that
 Phase three – the remaining splendor 173
the emperor was the living Tathāgata.53 The emperor represented the mani-
festation of the Buddha. On one hand, Buddhism was supported actively by
the court; on the other hand, in return, it served the imperial family. In the
second phase, paired caves became the prominent arrangement, reflecting
the reign of the “two sages” of the time: Emperor Xiaowen and Lady Feng.
The sinicized style emerged first in cave 6, was further developed, and was
eventually perfected in the third phase. That art, politics, and society were
closely connected, and the subjects served as a political tool is clearly reflected
in the art of Yungang. Many prominent narrative scenes (the Great Departure,
the Defeat of Mara, the Dipamkara Buddha, and the Asoka jataka, etc.) from
the life of the Buddha and his previous existences were represented as didactic
tools. These favored subjects were carefully chosen by the Yungang craftsmen
and/or donors to be in line with political exigencies. The third-phase cave tem-
ples are an integral and important part of the “Yungang style.” Even though
the excavation of the caves at this time fell into decline, they are still endowed
with some of the glory of the preceding phases. The innovative styles created
in the third phase, the iconography and cave structure, etc., had a profound
impact in the Luoyang area, even the entire central plain area, in particular,
the caves in southeast Shanxi, with the southward transfer of the capital.54 The
cave structure, and the subject matter and the style of the caves in these areas
are strikingly similar to those in the third-phase caves in Yungang. Su Bai notes
that the excavation of the cave temples in the Luoyang area primarily reflected
the “Yungang Style.”55 The Guyang cave temple in Longmen excavated dur-
ing Emperor Xiaowen and Emperor Xuanwu imitated the cave temples in the
second phase of Yungang. Mention was made in the Weishu that Emperor
Shizong commanded Dachangqiuqing 大長秋卿 (the Grand Chief Autumnal
Prince) Bai Zheng 白整 to excavate the Binyang caves in Longmen modeling
on the caves of Yungang.56 Thus it is clear that the third-phase caves had a
significant role in the development of the Buddhist rock-cut cave temples in
north China. This will be my final remark below.

One final remark: Buddhist sinicization reconsidered

One final point must be made to emphasize the significance of the third-
phase caves in the process of Buddhist sinicization, which was fully real-
ized in the third-phase caves before Longmen. Chinese Buddhist imagery
underwent a stylistic transformation during the late 5th century. It went
from the early images with one shoulder covered only to the icons with
baoyi bodai and both shoulders covered. This can be best reflected in cave
6 in which Buddhist images began to wear baoyi bodai Chinese style gar-
ments. Numerous scholars have explored the issue and most of these same
scholars suggest that the process was not improved until Longmen.57 Since
little to no attention has been paid to the third-phase caves before this
research, the full process of sinicization from the baoyi bodai images in cave
6 to the elongated and elegant images of the third phase has been wholly
174 Phase three – the remaining splendor
overlooked. The earliest datable example of Buddhist imagery in this new
trend can first be observed in cave 11-14 outside cave 11 at Yungang. The
inscription below indicates that the figures were executed in the 13th year of
the Taihe era (489). In view of the analysis of the third-phase caves above, it
is my belief that the process of sinicization was perfected by the end of the
third phase of Yungang. This can be seen clearly in the images of the third-
phase caves. The sinicized images are best exemplified in caves 28 and 30 in
the west-end portion of Yungang, as well as in caves 5-10 and 5-11. It is also
my belief that the sinicization did not result merely from the 486 edicts by
Emperor Xiaowen or from the close influence from the south.58 The reason
Buddhist images in the late 5th century began to wear baoyi bodai and to
appear attenuated and angular is far more complicated than scholars have
previously supposed. Also, the stylistic transformation in Buddhist imagery
is an important visual phenomenon. In the following, I will discuss the
complicated process based on the previous studies and new research in the
hope to emphasize the significance of the “Yungang Style 雲崗模式” and its
influence in the image-making in central plain China especially in Shan’xi
and Henan. I would also like to emphasize the mutual influence between
north and south in the process of sinicization. Finally, I will attempt to
demonstrate how painting was synchronized with sculpture during the 4th,
5th, and 6th centuries. These stylistic innovations perfectly correspond to
what Zhang Huaiguan 張懷瓘 (active 713–41) summarized in the evolution
of Chinese painting: Zhang gets the flesh right, Lu gets the bones right, and
Gu gets the spirit right 張得其肉, 陸得其骨, 顧得其神.59
First of all, the baoyi bodai robes appeared before the 486 edicts adopting
Chinese culture. The Maitreya-Amitayus stele discovered in 1920 in Maox-
ian 茂縣, northwest of Sichuan, was dated to the first year of Yongming 永明
(483). Both Maitreya Buddha and Amitayus Buddha wear baoyi bodai robes
with both shoulders covered. Apparently the sinicized baoyi bodai style existed
before the edict of 486. Besides, the reform of Xianbei people adopting Chinese
culture actually started as early as Daowudi in the beginning of the Northern
Wei, and it was continued through Xiaowendi and Xiaomingdi (r. 516–28).
The edict in 486 just reinforced the development. In the first year of Tianxing
天興 (398), Daowudi commanded Dong Mi 董謐 to compose different kinds
of ritual ceremonials.60 In the sixth year of Tianxing (404), he commanded
that caps and gowns be made according to rank.61 In the Taihe era of Emperor
Xiaowen’s reign, he commanded that a new style of caps and gowns be made
according to established rules. The Suishu (The Book of Sui) records that even
from the beginning of the Northern Wei during the reign of Emperor Daowu,
the adoption of Han clothing by the Northern Wei Xianbei people had started
and after him, through the emperor Xiaowen and emperor Xuanwu, the sini-
cization was continued.62 Through this long process of adoption of Chinese
culture, Emperor Xiaowen’s reform was perhaps most effective. It includes
using Chinese names, speaking the Chinese language, wearing Chinese clothes,
and transferring the capital to the south, from Pingcheng to Luoyang.
 Phase three – the remaining splendor 175
The development of Buddhism, a foreign religion, and its art, from its
arrival in Chinese soil in the Han dynasty, had been in a process of adjust-
ing to and adopting Chinese culture, and Buddhist art and architecture are
reflections of this cultural assimilation. The changes of the garments of the
Buddhist images were inevitably influenced by Han culture from the begin-
ning. The earliest extant Han Buddha images discovered in the 1940s in
Pengshan and Leshan tombs in Sichuan wear tongjian robes. These, though
reminiscent of Gandhara art, clearly demonstrate Chinese Han cultural influ-
ence, with loose sleeves and a strong sense of linear design. Chinese art is
characterized by linearity. The figures on the frescos in cave 169 in the Bin-
gling cave-temple complex are another good example of Chinese linear art
reflected in Buddhist iconography. Besides, the life of the Buddha depicted
in consecutive panels also shows the influence of Han brick reliefs. It is
clear that from the very beginning, Buddhist art is a hybrid of Chinese and
western culture.
The second interpretation of the stylistic shift, which assigns much impor-
tance to southern influence, cannot be wholly denied. There is, however,
little evidence that southern culture became dominant. The images in the
Qixiashan 棲霞山 cave-temple complex in Nanjing do not wear baoyi bodai
robes. “This baoyi bodai style was apparently northern rather than southern,
as is generally assumed.”63 A more valid position is that the well-known
Chinese tradition of syncretism, of blending seemingly divergent traditions,
was operative in the cave-art we are discussing. In this regard, we find an
interesting discussion between Chen Qingzhi 陳慶之 (d.u.), a southern offi-
cial, and Zhu Yi 朱异 (483–549), a scholar well versed in classics, history,
and literature. The conversation, from the Luoyang qielan ji goes like this:

Xiaoyan appointed (Chen Qingzhi) Governor of the Metropolitan prov-

ince of Sizhou. Chen showed great respect for northerners and relied
heavily on them. This attitude was highly unusual. Zhu Yi found it
strange and asked him about it. Chen replied: ‘Ever since the Jin and
Song periods, people have been calling Luoyang a desolate land and say-
ing that anyone north of the Yangtze is a barbarian. On my recent visit
to Luoyang, I at once realized that the capped and gowned scholars all
lived in the central plains where ritual and etiquette flourish and people
are prosperous. I can’t find words to express how what I saw there
impressed me. This is why we say that the imperial capital is magnificent
and the model of the four quarters. This is like those who have climbed
Mt. Tai think little of mere hills, and those who have been on the Yang-
tze or open sea despise the Xiang and Yuan. How can we not assign
heavy responsibilities to the northerners?’ Qingzhi therefore adopted
the Northern Wei style of feathered canopies for insignias and costumes.
Gentry and commoners south of the Yangtze quickly competed in imi-
tating him. The loose robe with wide girdle (baoyi bodai) was worn in
Moling (Nanjing).64
176 Phase three – the remaining splendor
What Yang Xuanzhi tells us in his Luoyang qielan ji, confirms that the north-
ern culture actually inspired the south in the style of dress and that the baoyi
bodai style is from the central plain, not from the south, as scholars have
suggested. But this is not to say that the south did not have an impact on
the north. It certainly did, but in Buddhist iconography, North and South
interacted, and the influence is mutual.
To summarize the points I have made above, the process of sinicization
started as early as the Han Dynasty. The 486 edicts were only one of the
factors which enhanced the process, and it is not the major reason for sinici-
zation. In this process, northern culture and southern interacted, and influ-
enced each other. The southern influence is undeniable particularly in the
late stages of Yungang and throughout the construction of Longmen. The
influence of the South, due mostly to migration to the district of the North-
ern Wei capital, made itself felt in Buddhist images of the caves. Supple and
slender bodies replaced the stiff formalism of the earlier images. I would
suggest that northern culture also had a profound influence on south China.
The baoyi bodai dress style started from the north rather than south. The
robes of Buddhist figures were modified in stages to fit more closely into the
Chinese tradition and aesthetic value. This led to a greater degree of natural-
ism in the depiction of the body in Chinese Buddhist sculptures.
As I mentioned earlier, many scholars have taken the position that the
sinicization of Buddhist iconography was not completed until the Long-
men period. I have argued that sinicization was perfected during the third
phase of Yungang, not during the Longmen time as other scholars have
suggested. The images in the third phase of Yungang appear much more
elegant and graceful. They are the genuine xiugu qingxiang icons demon-
strating a strong sense of linear design, one of the outstanding features
of Chinese art. The elongated seated Buddha on the west wall of cave 30
can be considered the masterpiece of the third phase. It is delicate and
supple. The gentle smile makes the image more personal. Besides, it is
widely accepted that the Longmen artisans imitated Yungang in iconogra-
phy and architecture while creating their own regional art. The images in
the earliest cave, Guyang cave 古陽洞, in Longmen closely resemble those
in caves 7 and 8 in Yungang. Also, it is traditionally believed that some of
the Yungang artisans went to Luoyang when the capital was transferred
southward.
My final argument against simplistic explanations for the iconographic
shift in cave sculptures is based on the largely unexplored relationship
between painting and sculpture in China during the period under discus-
sion. It is this relationship, and the fact that many renowned artists like Dai
Kui 戴逵 (?−396), Gu Kaizhi (345–406), Lu Tanwei 陸探微 (?−485), Zhang
Sengyou 張僧繇 (?−?), a contemporary of Emperor Wudi (r. 502–49) of the
Liang, and Cao Zhongda 曹仲達 of the Northern Qi (550–77) were both
painters and sculptors, that lead me to believe that the art of the brush had
a profound influence on the shaping of the figures in the Buddhist caves of
 Phase three – the remaining splendor 177
Yungang and Longmen. Below, I will examine more closely some of the
representative exemplars of this phenomenon.
The sinicization is not only reflected in the changes of Buddhist garments,
but is also reflected in the iconographic transformations from the non-
Chinese masculine colossal images of five Tanyao icons to the attenuated
and elegant Buddhas seen in the masterful Buddha image on the left wall
of cave 31 in Yungang; the slender seated Bodhisattvas with ankles crossed
in caves 5-10 and 5-11; and the fleshy and heavy images of Xiangtangshan
caves and Qingzhou. These iconographic transformations took place
simultaneously with the transformation of style in painting. As sculptures
were mostly painted, the sculptors were often required, as a foundation for
sculpture, to be adept at painting as well. There is sufficient evidence that
distinguished sculptors, in general, excelled both in painting and sculpture.
Chinese painters thus had a profound impact on the transformation of
Buddhist sculpture.
Dai Kui, a contemporary of Gu Kaizhi, is well-known for being good at
both painting and sculpture. Dai was also skilled at casting Buddhist images
and carving them. Whenever he sculpted a Buddha image he analyzed his
task and pondered the means to carry it out until it achieved brilliance. Dai
also sculpted five “image of Buddha taken in procession” 行像 images. Dai
Kui’s son, Dai Yong 戴顒, inherited his father’s talent. “The simplicity of his
rendering of images was charming and profound.”65 “The images made by
both Dais were unequalled in all generations.”66
Jiang Shaoyou 蔣少游 (?−501) of the late Wei period was said to be skilled
in calligraphy and painting. According to Lidai minghua ji, he excelled in
both painting the human figure and carving.67
That good sculptors were also adept at painting is well demonstrated in
all of the examples mentioned above, and a close inspection of cave 328 at
Dunhuang gives one a better idea of how painting and sculpture are closely
associated as an integral whole. The consistency of the configuration and a
uniformity of modeling and style are clear. We are even led to believe that in
many cases it is one single artist creating both painting and sculpture since
they echo one another so remarkably.
Both from literary evidence and observation of actual work, it can be seen
that painting and sculpture are interconnected and synchronized. When the
style of painting changes, the style of sculpture changes as well. To dem-
onstrate this point we might look at some of the renowned painters, each
representative of his own time: Gu Kaizhi of 4th century, Lu Tanwei of the
5th century, and Zhang Sengyou of the 6th century and their painting style
while at the same time looking at the style of contemporary sculpture.
Gu Kaizhi remains a legend in China even today, revered as the father of
Chinese painting. Gu, it is said, often left the eyes of his figures undotted
for years to decide how to transmit the spirit. “Transmitting the spirit in
sketching a portrait lies precisely in the eyes,” he wonderfully put it when he
was asked.68 He made the eyes of his figures speak. The eye contact between
178 Phase three – the remaining splendor
the figures in the Nüshizhen tu (Admonitions to the Court Ladies) and the
direction of their gazes bring them to life and make the viewer feel at one
with them. In his Lidai minghua ji, Zhang Yanyuan commented on Gu’s
brushstrokes as follows:

(They are) firm and powerful connecting with one another without inter-
ruption. The strokes follow each other in succession and go beyond the
mundane. (In this way he made) his style lofty and unique. His strokes
are as swift as the wind and like a flash of lightening. His intent existed
in his mind before it emerged from his brush. When the painting was
completed, his intent remained in it. This is entirely because of the vital-
ity of spirit. . . . Master Gu understood the subtle principles so well in
painting the worthies of antiquity that even after looking at them all day
one did not feel weary.69

Gu Kaizhi created a revolutionary innovation when he depicted the Indian

layman Vimalakirti because it satisfied Chinese eyes and taste by making
this Indian layman a Chinese scholar. Master Gu was the first to create the
portrait of Vimalakirti. His face gives the impression of emaciation and deli-
cacy suggesting his illness. He is shown resting against a small table having
forgotten how to speak.70
He created this delicate and emaciated style to fit the Chinese taste, and
more importantly, he transformed the Indian layman into an idealized Chi-
nese scholar to satisfy Chinese eyes and taste, so much so that the Chinese
elite hoped to find reflections in the image of Vimalakirti. This became the
model in creating iconic Buddhist images.
It should be mentioned that Gu was not alone in creating this new style.
His contemporary, Dai Kui, played an important role as well in this stylis-
tic innovation. Dai was perhaps more influential after all since he sculpted
many Buddhist icons. “Dai’s style was delicate and his integrity was lofty.”71
His famous The Seven Sages of the Bamboo Grove and Rong Qiqi, now a
lost piece, was, according to Gu Kaizhi, a masterpiece beyond comparison
with the work of any other painters.72 Gu also painted “Seven Sages” but it
was not as good as Dai’s. Although unfortunately this work did not come
down to us, we can still nevertheless trace the very root from the extant
stamped-brick tomb murals of the “Seven Sages,” which was excavated in
1960 at Xishanqiao 西善橋 near Nanjing and dated as a work of the 5th
century. The authorship of this work may never be known, but it perhaps
is not too unreasonable to infer that the designer was influenced by Dai Kui
since it is so close to Dai Kui both in time and in place. And since Dai was
the “leader of all craftsmen,” who wouldn’t want to copy his work? Xie He
謝赫 of the sixth century expressed the view that “Dai excelled in depicting
Wise Men and Sages. They have become the models for all craftsmen. He
became a true leader.”73 The biography of Dai Kui in Songshu (The Book of
Song) records that Buddha images began to emerge from the Han dynasty,
 Phase three – the remaining splendor 179
but that the rendering of the forms was not yet skillful. Dai Kui uniquely
excelled in this matter.74 “Both Dais, father and son, excelled in painting, and
believed in Buddhism. In sculpting and painting, they created new models
in every move.”75 These new models were copied by painters and sculptors
of all times. After him, Cao Zhongda of the Northern Qi, Zhang Sengyou
of the Liang dynasty, and Wu Daozi and Zhou Fang of the Tang dynasty
all showed their advantages and disadvantages.76 But the root was Dai Kui.
The later styles, Cao style, Zhang style, Wu style, and Zhou style initially
copied Dai Kui’s model though they each later invented their new mode. Lu
Tanwei was the third painter in the tradition of Dai Kui and Gu Kaizhi and
also known for “transmitting the spirit” and portraying figures of elegance.
He is usually regarded as the key figure in the stylistic evolution of Chinese
plastic arts: “Lu embraced what went before him and gave birth to what
succeeded him. He stands alone from ancient times up till now.”77 Both Gu
Kaizhi and Lu Tanwei were esteemed as the foremost apractitioners of their
art by Zhang Yanyuan:

Master Lu brought subtlety into the souls of his figures. Each brush
stroke joined together with spirit. His strokes are powerful and sharp as
if chiseled with a knife. The appearances of his figures look elegant and
refined. They seem to be actually living. In capturing the whole person,
Lu gets the bones right, but Gu gets the spirit right.78

It is this “elegant and refined appearance” that was favored by Chinese con-
noisseurs and soon copied by Chinese artists in both painting and sculpture.
As mentioned earlier, the earliest datable example of Baoyi bodai appeared
in cave 11-14 outside cave 11 at Yungang (dated to 489 AD). Buddhist iconic
figures thereafter all showed emaciated bodies and elongated faces: they were
transformed into Chinese scholars. Most of the images from the third phase
of Yungang demonstrate these features and the masculine, non-Chinese look-
ing Buddha disappeared from Chinese soil from the late 5th century. The
essential qualities of the elongated style are also exemplified by the images
in the Guyang cave at Longmen.
This trend of slender and elegant figures is not only illustrated in Buddhist
sculpture, but is also observed in other, non-religious, art, and became a
dominant feature of 5th century Chinese art. The earliest known examples
of non-Buddhist images of this type are illustrated on the lacquer screen in
the tomb of Sima Jinlong and his wife, discovered in the village of Shijiazhai,
outside Datong, dated from the fourth year of Yanxing 延興 (474) to the
eighth year of Taihe (484).79
A different trend gradually began to emerge in the late 5th century and
would become dominant throughout the 6th century. Typical examples of
this trend are found in the figures from the stone steles unearthed from the
Wanfo Temple dated to 523 AD and 529 AD. The figures are fleshy with
plump shoulders and full-moon faces.80
180 Phase three – the remaining splendor
From these images we do not see Gu Kaizhi’s “transmitting the spirit”
through eyes shown in Nüshizhen tu, or the emaciated Vimalakirti; nor
do we see Lu Tanwei’s clear-cut elegant figures reflected in the images in
the third-phase cave temples. Rather we seem to observe Zhang Sengyou’s
“fleshy” figures. In rendering Buddhist images, Zhang Yanyuan tells us,
Zhang was skilled in:

A thousand changes and myriad transformations, (creating) strange

shapes and unique forms. He let them pass before his eyes and then
transferred to his hand; receiving them with his mind, he echoed them
with his hand [he masters them with his mind and hand]. In reality it is
not only that he reached perfection in the Six Laws (of Xie He); in truth
he mastered all the wonders of myriad categories.81

The trend toward simpler draperies and fleshier bodies was increasingly
noticeable at the end of Northern Wei. It continued into the mid-6th century
under the Eastern Wei (534–50), Western Wei (535–57), and Northern Qi
dynasties and gradually underwent a process of softening. The bodies and
faces of the figures became less narrow and the pleats of their robes lost their
sharp, angular character. The decorative linearity which characterized the
Northern Wei treatment of drapery lost its tension and rhythmic dynamism,
becoming somewhat mechanical and repetitive, while the images’ counte-
nances became milder.
The simplicity in the rendering of the images further demonstrates the
influence of Zhang Sengyou who created “spare style 疏体.” Zhang Sengyou
“acquired his talent from Heaven, so that after only one or two strokes the
image already existed in them.”82 Mi Fu (1050–1107) of the Northern Song
described his images as “short of face and heavily painted 面短而艷.”83
Zhang’s style is also clearly reflected in the mural paintings in Xu Xianxiu
徐顯秀 tomb discovered in Shanxi province. The figures on the mural appear
rather fleshy and are simply sketched (plate 66).
Another example is the cavalrymen in the murals at the tomb of Prince
Lou Rui 婁叡 dated to the first year of Wuping (570 AD) of the Northern
Qi dynasty. They were depicted with fuller, rounder faces and bodies. The
murals are drawn with simple but forceful lines, showing great vitality.
The Jiankang shilu 建康實錄 records that Zhang adopted the three-
dimensional method (凹凸法) from Indian and Western Regions painters
who were invited to Jiankang by Emperor Liang Wudi. Zhang was also often
invited to paint Buddhist icons in temples for Emperor Liang Wudi:

Three-dimensional flowers [with bright and dark colors] were painted all
over the temple gate. They are said to be Zhang Sengyou’s hand traces.
These flowers were [painted] in Indian traditions with red, green and
blue colors. Viewed from a far distance they look three-dimensional
 Phase three – the remaining splendor 181
making one dizzy; viewed from a near distance they look flat. People of
the time all think they are very unique, therefore named it the Temple of
Flower of Three-dimensions.84

Zhang’s “three-dimension method” is not only reflected in the technique

of the “light-heavy” turns used in the calligraphy, it is also displayed in the
skills of the “bright-dark” technique influenced by the West.
It is significant to note that Zhang Sengyou is not alone in contributing to
the second transformation in the evolution of Chinese Buddhist sculpture.
Cao Zhongda of the Northern Qi dynasty played a fairly important role in
this innovation.
According to Lidai minghua ji, Cao Zhongda of the Northern Qi was by
origin a native of the realm of Cao (in today’s Uzbekistan near Samarkand)
and acknowledged his supremacy in painting icons of the Indian type. Guo
Ruoxu of the Northern Song made a vivid description of Cao Chongda’s and
Wu Daozi’s styles in his Tuhua jianwen zhi:

Wu’s strokes in form were round and curving and his robes fluttered
upward. Cao’s strokes were dense and layered. His robes were tight and
clinging. Therefore later generations said Wu’s girdles flutter in the wind,
while Cao’s garments had just come out of the water.85

Since none of Cao’s works survive today, we must rely on Guo Ruoxu’s
record and some recently discovered (1996) Buddhist sculptures of Qingzhou
in Shandong Province. The Buddhist icons found there are primarily from
Northern Wei and Northern Qi dynasties. Luckily most of them are still
painted. The robes of the Northern Qi images all cling tightly to the bodies.
They appear to be wet and just out of the water, exactly as Guo Ruoxu had
observed in his comments on Cao Zhongda’s style. The Qingzhou Buddhist
icons, as they continue to be studied, will provide invaluable and abundant
evidence for scholars to investigate Zhang Sengyou’s and Cao Zhongda’s
styles further and to study the cultural exchange between the south and the
north.
By examining these renowned painters and their styles, I have attempted
to demonstrate how painting was synchronized with sculpture during the
4th, 5th, and 6th centuries. These stylistic innovations perfectly correspond
to what Zhang Huaiguan suggested in the evolution of Chinese painting:
Zhang gets the flesh right, Lu gets the bones right, and Gu gets the spirit
right.86 In other words, in the 4th century, Gu (Kaizhi) gets the spirit right;
in the 5th century, Lu (Tanwei) gets the bones right; and in the 6th century,
Zhang (Sengyou) gets the flesh right. These three great painters had a proud
impact on the evolution of Buddhist sculpture.
These three preoccupations of the three painters conform perfectly to the
Buddhist images produced in their times. In the first phase the images by Gu
182 Phase three – the remaining splendor
emphasize the spirit; the second phase, by Lu, emphasizes the bones; and
the third phase, by Zhang, emphasizes fleshiness. In these periods, Chinese
sculpture experienced two successive stylistic changes. It first emerged from
the simple, powerful, and virile forms of the first phase of Yungang to the
elongated, refined, and slender figures of the Yungang third phase and Long-
men. We find, in this process, that alien Buddhist icons were transformed
into Chinese scholar clad in baoyi bodai. Then in turn, this changed to the
full-moon faced, fleshy figures of Qingzhou and Xiangtangshan.
In summary, the sinicizaiton of Buddhist images was a long process.
Buddhism and its art underwent assimilation from the time of its arrival
on Chinese soil in order to survive. It was a process of absorption and
innovation. Chinese art is the art of linear design, and Buddhist art became
linear under Chinese influence. The garments of Buddhas are in a way a
reflection of monastic life. They are not just the results of political reform
enforcing sinicization or the southern influence. The cultural influence is
mutual through interaction as our literary sources tell us. The sinicization
is not just reflected in the change of the garments of the icons. It is also
reflected in the elegant and attenuated iconography. This transformation
is closely related to the famous painters. As we have pointed out Zhang
Huaiguan perceptively suggested: Zhang gets the flesh right, Lu gets the
bones right, and Gu gets the spirit right.87 The iconographic transforma-
tions of painting and sculpture took place simultaneously in China. The
great painters had a profound impact on the evolution of the sculpture
from the 4th to the 6th centuries. From Yungang to Longmen, from north
to south, there is a clear map of evolution and yet they each have their
own aesthetic principles and each created a distinctive art form. As I have
suggested, these art forms did not solely result from the two factors usu-
ally mentioned by art historians. I suggest that the sinicization edict of
Emperor Xiaowen and the growth of southern influences after the court
moved to Luoyang played a part. I have attempted to amplify these expla-
nations by suggesting an additional factor: the synchronization of painting
and sculpture. Together these four factors combined to create some of the
great monuments of world art history.

Notes
1 In the history of peasants’ rebellions, only at the end of the Northern Wei did
many Buddhist monks participate and even lead some of them. They destroyed
Buddhist temples and caves, and burned sutras. For details, see Su Bai, “fenqi,”
in Zhongguo shikusi yanjiu (Beijing:wenwu chubanshe,1996), note 23, 88.
Also see. Weishu, “Shizong ji” (Beijing: Zhonghua shuju, 1974), 8:191 and
“Suzong ji,” 9:221.
2 Daoxuan, Xu gaoseng zhuan, T50:2060:425c26 and T50:2060:427c23.
3 These include caves 21, 22, 22-1, 23, 23-1, 24, 24-1, 25, 25-2, 26, 27, 27-2, 28, 28-2,
29, 30, 31, 32-9, 32-11, 32-12, 33, 33-3, 33-4, 33-6, 34, 35, 36, 36-2, 37, 38, 38-3, 40,
and 41, as well as 5-10 and 5-11.
 Phase three – the remaining splendor 183
4 Li Xueqin, “Yungang shiku xinbian kuhao shuoming,” in Zhongguo shiku Yun-
gang shiku. Ed. Yungang shiku wenwu baoguansuo (Beijing: Wenwu chubanshe,
1991), 209–11.
5 In some cases, the Śākyamuni and Prabhutaratna seated side by side are the pre-
dominant images on the main wall. They constitute almost half of the primary
images on the main wall. This demonstrates the popularity of the Lotus Sutra
during the third phase.
6 The author discovered this lost standing Buddha in the Musée Guimet in 2005
when doing research in Paris.
7 Mizuno and Nagahiro divided the ceilings into: (1) elliptical and slightly domed,
(2) flat and rectangular, always coffered (zao jing), (3) coffered and corbelled,
and (4) coffered, corbelled, and coved, four different types, see Mizuno Sei-
ichi and Nagahiro Toshio, “Unkō sekkutsu: seireki goseiki ni okeru Chūgoku
hokubu Bukkyō kutsuin no kōkogakuteki chōsa hōkoku: Tōhō Bunka Kenkyūjo
chōsa (Yungang: The Buddhist Cave-Temples of the Fifth Century A.D. in North
China), (Kyoto: Kyōto Daigaku Jinbun Kagaku Kenkyūjo, 1951), vol. 16, 65–6.
However, types 3 and 4 should belong to type 2 since they are just two differ-
ent kinds of decorative designs and the structural pattern of types 3 and 4 is the
same. Mizuno and Nagahiro also mistook pingqi design as pingqi zaojing. Pingqi
is an architectural term of the Song dynasty. In Ming and Qing dynasties it was
called tianhua 天花, and was used for plafond decorations. The design looks like
a chessboard pattern with 4, 5, 6, 8, 9, or 12 panels. Zaojing is a term used for
plafond decorative art and structure. The zaojing panel can be dousi 斗四 (quad-
rilateral), or douba 斗八 (octagon). Zaojing plafond decoration can be seen as
early as in the Han dynasty. In Yungang, pingqi and zaojing plafond decorative
designs are combined together with zaojing inside pingqi.
8 In some of the images on steles discovered in Chengdu, four disciples carved in
bas-relief appear as well. See Li Yuqun, “Sichuan nanchao zaoxiang de ticai yu
beifang shiku de guanxi,” in Sichuan chutu nanchao fojiao zaoxiang. Ed. Sichuan
bowuyuan, chengdushi wenwu kaogu yanjiusuo, and Sichuan daxue bowuguan
(Beijing: Zhonghua shuju, 2013), 228–42.
9 According to Zhao Kunyu’s statistics, there are more than 190 narratives of Bud-
dhist sculpture. See Zhao Kunyu, Yungang – Fojiao gushi diaoke yishu (Nanjing:
Jiangsu meishu chubanshe, 2010), 3.
10 The story is told in the Guoqu xianzai yinguo jing (Sutra on the Causes and the
Effects in the Past and Present), translated by the Indian Buddhist monk Gunab-
hadra (Qiunabatuoluo 求那跋陀羅) in 435–443 CE, see T03:189:628:c21.
11 See Foshuo guanfo sanmei haijing, translated by Buddhabhadra (Fotuobatuoluo
佛陀跋陀罗) of the Eastern Jin, “guan siweiyi pin,” T15:643:667:c14; also see
Mokemoye jing (Mahāmāyā sutra), also known as Buddha’s Ascent to Tavatimsa
to preach for his mother Mayadevi, translated by unknown Tanjing 释昙景 of
the late 5th century, T11:383:1008.
12 Jin Weinuo, “fobensheng tu xingshi de yanbian,” in Jin Weinuo, Zhongguo mei-
shushi lunji (Taibei: Mingwen shuju, 1984), 389–96.
13 Ibid.
14 Pao-chen Chen, The Goddess of the Lo River: A Study of Early Chinese Narrative
Handscrolls, Ph.D. dissertation (Princeton, 1987), 139–40.
15 Julia K. Murray, “Buddhism and Early Narrative Illustration in China,” Archives
of Asian Art, 48 (1995): 24.
16 Ibid.
17 Ibid.
18 Michael Sullivan, The Arts of China (Berkeley, CA: University of California Press,
2009), 107. Also see Karetzky Patricia Eichenbaum, Early Buddhist Narrative
184 Phase three – the remaining splendor
Art: Illustrations of the Life of the Buddha from Central Asia to China, Korea,
and Japan (Lanham, MD: University Press of America, 2000), 72.
19 Baoshan Chu mu, 2 vols. (Beijing: Wenwu, 1991), vol, 1, 144–6, 501–3. Also see
Cui Renyi, “Jingmen Baoshan erhaomu chutu de yingbin chuxing tu chulun,”
Jianghan kaogu, 2 (1988): 72–9; Chen Zhenyu, “Chu guo chema chuxingtu
chulun,” Jianghan kaogu, 4 (1989): 54–63.
20 Wu Hung read the story from right to left: an official in a white robe is taking
a tour on a horse-drawn chariot; the horses increase speed and attendants run
ahead of them; the chariot then slows down and is greeted by a kneeling figure.
Meanwhile, a gentleman wearing a dark robe is on his way to meet the official.
In the final scene, the official has descended from the chariot and meets the host;
but somehow he is now dressed in a dark robe while the host wears a white
robe. See Wu Hung, “The Art and Architecture of the Warring States Period,”
in The Cambridge History of Ancient China: From the Origins of Civilization
to 221 B.C. Ed. Michael Loewe, and Edward L. Shaughnessy (Cambridge, UK:
Cambridge University Press, 1999), 705.
21 The great departure episode is depicted in caves 5-11, 5-33, 5-38, 28, 31, 35, 38, and
41, eight depictions in total during the third phase. Since these episodes – the
great departure and the prince riding on an elephant – often appear together as
a pair, I believe cave 5-10 originally had a depiction of the great departure as well
but it was looted leaving the big hole that now appears on the west side of the
front wall, but the prince riding on an elephant is still there on the east side. It
is the same situation as in caves 33-3 and 33-4. In cave 33-4, the two episodes are
illustrated outside the cave above the entrance. The prince riding on an elephant is
illustrated in caves 5-10, 5-11, 5-33, 5-38, 31, 33-3, 33-4, and 38, eight discernable now
in total. I also believe that the north side of the west wall in cave 35 originally
had the depiction of the prince riding on an elephant as well, but it is eroded now.
The south side still bears an illustration of the great departure.
22 In cave 28, the great departure episode is depicted on the west wall over the niche,
but it is not illustrated together with the story of the prince riding on an elephant
as a pair. Instead, it is depicted together with the prince’s farewell to his white
horse Kanthaka. In cave 38, these two episodes are illustrated together as a pair,
as usual, but they are not illustrated on the front wall over the entrance. Rather
they are depicted on the north wall. In cave 31, the two episodes are depicted on
the either side of the reveals of the window of the rear chamber, not on the front
wall.
23 Guoqu xianzai yinguo jing, T3:189:632c27.
24 The prince then made several solemn vows: “if I do not eliminate the grief of
old age, sickness and death, I will not return to the palace. If I do not achieve
complete Enlightenment, if I cannot turn the wheel of the Dharma, I will not
return to my father. Until I exhaust all feelings of love and attachment, I will not
go home to see Mahaprajapati nor Yashodhara.” See Guoqu xianzai yinguo jing,
T3:189:633a18.
25 Alfred Charles Auguste Foucher, The Life of the Buddha: According to the
Ancient Texts and Monuments of India (Middletown, CT: Wesleyan University
Press, 1963), 24.
26 Yan Wenru, Yungang shiku yanjiu (Guilin: Guangxi shifan daxue chubanshe,
2003), 332; Wang Heng, Yungang fojing gushi (Sutra Stories in Yungang) (Tai-
yuan: Shanxi renmin chubanshe, 2002), 11.
27 In both India and Gandhara narratives, Maya is often seen lying down, while in
the distance, a descending elephant prepares to enter her side. Besides, it is more
suitable to depict the great departure and the elephant-riding as paired images,
for they are closer in time in historic life and in sutra texts.
 Phase three – the remaining splendor 185
28 The episode is depicted in caves 32-3, 33-3 and 41.
29 Xiuxing benqi jing, T3:184:463a17.
30 The episode is depicted in caves 28, 30 and 41.
31 Guoqu xianzai yinguo jing tells that after they departed the palace and by dawn,
the prince patted Kanthaka’s back and said: “you have accomplished some dif-
ficult task.” He said to Chandaka: “now that I have reached this place of refuge,
you need to lead Kanthaka and return to the palace.” Chandaka cried after he
was told to leave. He fell lame on the ground and could not control himself.
Kanthaka, when hearing that they would be left behind, knelt down and licked
the prince’s foot, his tears fell like rain, see T3:189:633b1.
32 The episode of the banishing of the demons is depicted in caves 31, 35, and 38.
The episode of the conversion of the three Kasyapas is illustrated in caves 35 and
38. Judging from the illustration of the banishing the demons in cave 31, it seems
that the story of the conversion of the three Kasyapas was originally portrayed
as well.
33 The other three are the birth, the Enlightenment, and the attainment of Nirvana.
34 According to the Xiuxing benqi jing (Sutra on the Origin of Practice of the Bod-
hisattva), translated by Zhu Dali and Kang Mengxiang in 197 C. E., the earliest
extant literary source on the biographies of the life of the Buddha, tells us that the
king asked his advisers how to prevent the prince from leaving home to pursue his
spiritual life. One minister suggested that the prince should be ordered to oversee
the ploughing so he can be distracted. So he was sent out with a thousand ser-
vants to oversee the farming and told how to inspect their work. The prince then
sits under a jambu tree watching them till the land. When they broke new soil,
worms came out from the soil, and then the birds swallowed them. Then frogs
chased and ate the birds and snakes ate the frogs. Peacocks then ate the snakes
and in turn the hawks ate the peacocks. The vultures then ate the hawks. Watch-
ing these creatures devouring each other, the adolescent’s heart was wounded.
He sat under the tree and attained the first meditation. When the king heard that
his effort to prevent the prince from practicing his spiritual life was in vain, he
went out to the field to meet with his son. Looking from the distance, the king
witnessed the miracle of the branches of the trees moving to make shade for the
prince. He recognized the existence of the spirit after watching the miracle, but
he didn’t realize that it was his son. So he dismounted the horse and worshipped
his son, see T3:184:467b18.
35 Mizuno and Nagahiro, Unkō sekkutsu, vol. 7, 79. Also Wang Heng, Yungang
fojing gushi, 8.
36 Guoqu xianzai yinguo jing, T3:0189:0628c21. The story is slightly different in
the Xiuxing benqi jing which begins with the elephant event: “when they tried to
go out from the city gates, they realized that an elephant was stuck between the
gates so the prince and his cousins decided to have a test of strength to see who
was the strongest,” see T3:184:465c7.
37 Karetzky Patricia Eichenbaum, Early Buddhist Narrative Art: Illustrations of the
Life of the Buddha from Central Asia to China, Korea, and Japan (Lanham, MD:
University Press of America, 2000), 180.
38 The moment is illustrated only in caves 35 and 38.
39 The Rahula jataka is depicted in caves 9, 19, and 38; the Asoka story is illustrated
in caves 25, 28, 29, 33, 33-4, 34, and 5-11; the Dipamkara story is depicted in caves
35 and 38 and the Gridhrakuta Hill legend is depicted on the south wall in cave
38. The story is accounted in Xuan Zang’s Da Tang xiyu ji 大唐西域記 (Buddhist
Records of the Western World), see T51:2087:921b6.
40 Guoqu xianzai yinguo jing, T3:189:622b14.
41 Ibid., T3:189:622b24.
186 Phase three – the remaining splendor
42 Ibid., T3:189:622b27.
43 Many sutras, for example, the Sutra on the Causes and Effects of the Past and
Present, Sutra on the Origin of Practice of the Bodhisattva and Lalitavistara, start
with the former existence of the Buddha in the time of the Buddha Dipamakara.
44 The story can be found in caves 5-11, 5-38, 25, 28, 29 33, 33-4, and 34. It is depicted
twice in cave 33-4.
45 Xianyu jing, translated into Chinese by Hui Jue and others in 445. Accord-
ing to Chu sanzang ji ji, Tanxue, Weide and eight others went westward from
northwestern China to Khotan to seek Buddhist scriptures. They unexpectedly
encountered a Buddhist Council at the Great Monastery which happens every five
years. During the Council, Buddhist monks expounded the Buddhist scriptures
and rules of monastic discipline in the language of Khotan. Tanxue and other
monks translated and recorded the lectures in Chinese. Later they returned from
Khotan and, upon reaching the region of Turfan, compiled the lectures given by
the monks in Khotan as a single sutra, which is known as the Sutra of the Wise
and the Foolish, T4:202:368:c14.
46 Xianyu jing, T4:202:368c7.
47 In Zabaozang jing (Sutra of the Miscellaneous Treasures), from which the story
is derived, we are told that the boy is standing to the left side of the Buddha but
the Yungang artists portrayed a kneeling boy and he is not always on the left side
of the Buddha, see T4:023:0497:b15.
48 According to Xuan Zang, “. . . to the south of the vihara, by the side of the
mountain cliff, is a great stone house in which Tathagata, when dwelling in the
world long ago, entered Samadhi. To the north-west of the stone house and in
front of it is a great and extraordinary stone. This is the place where Ananda
was frightened by Mara. When the venerable Ananda had entered the Sama-
dhi in this place, Mara, assuming the form of a vulture, in the middle of the
night, during the dark portion of the month, took his place on this rock, and
flapping his wings and uttering loud screams, tried to scare the venerable one.
Ananda, filled with fear, was at a loss to know what to do, then Tathagata, by
his spiritual power, seeing his state, stretched out his hand to compose him. He
pierced the stone wall and patted the head of Ananda, and with these words of
great love he spoke to him: ‘you need not fear the assumed form which Mara
has taken.’ Ananda in consequence recovered his composure, and remained
with his heart and body at rest and peace. Although years and months have
elapsed since then, yet the bird traces on the stone and the hole in the rock still
remain visible,” see T51:2087:921b6. Eng. trans. by Samuel Beal with minor
revision, see Buddhist Records of the Western World (Delhi: Motilal Banar-
dass, 1994), 9:154. Also see Faxian zhuan 法顯傳 (A Record of Buddhistic
Kingdoms) by Faxian of the Eastern Jin, which has similar record of the story,
T51:2085:859:c16.
49 See Mizuno and Nagahiro, Unkō sekkutsu, vol. 2, Appendix, “epigraphy” for
inscriptions.
50 See Mizuno and Nagahiro, Unkō sekkutsu, vol. 1, plates 109 A, B, and 110 A, B.
51 See Mizuno and Nagahiro, Unkō sekkutsu, vol. 2 for the inscription detail.
52 Fourteen in total (the Fufazang yinyuan zhuan, dajiyi shenzhou jing, and zabao-
zang jing are the three in extant), see Weishu, 114:3037.
53 Weishu, 114:3031.
54 These caves are mainly situated within the triangle-shaped Buddhist centers and
capitals of the Northern dynasties (Pingcheng, Luoyang and Ye) where Buddhism
flourished quickly, and many Buddhist caves were commissioned for merit accu-
mulation, good fortune and for meditation. Some areas in southeastern Shanxi
have more caves than others. Gaoping 高平 County, for example, has many
 Phase three – the remaining splendor 187
Buddhist caves and monasteries (Yangtoushan cave complex 羊頭山, Shitanghui
cave complex 石堂會, Gaomiaoshan cave 高廟山 and Fushan cave 釜山, and
Dinglin monastery 定林 and Kaihua monastery 開化 etc.). Besides that, Changzhi
長治 County, Lingchuan 陵川 County, and many other counties are also filled
with Buddhist caves and monasteries. I have examined and documented more
than 30 unexplored caves in Shanxi including Lianghoudian caves 良侯店 in
Wuxiang County 武鄉; Zihongzhen caves 子洪鎮 in Qixian 祁縣; Yuanzishan
caves 圓子山 and Xiangtangsi caves 響堂寺 in Yushe 榆社; Shifosi caves 石佛寺
in Zuoquan 左權; Beishanqianfo caves 北山千佛, Wangqing caves 王慶, and Jia-
odingshan caves交頂山 in Changzhi 長治; Yunlongshan caves 云龍山 in Heshun
和順; Fushan 釜山, Shitanghui 石堂會, Yangtoushan 羊頭山, Dinglinsi 定林寺,
Gaomiaoshan 高廟山 in Gaoping 高平; Baoyingsi caves 寳應寺 in Lingchuan 陵
川; Shimasi Cave 石馬寺 in Xiyang 昔陽; and Kaihesi Cave 開河寺 in Pingding
平定, etc.
55 Su Bai, “Yungang moshi,” 142.
56 Weishu, 114:3043.
57 See endnote 64 in Chapter One for the scholarship of sinicization.
58 Alexander Soper, “Southern Chinese Influence on the Buddhist art of the Six
Dynasties Period,” Bulletin of the Museum of Far Eastern Antiquities, 32 (1960):
78–9; also Yang Hong, “Shilun Nanbeichao qianqi foxiang fushi de zhuyao
bianhua,”Kaogu, 6 (1963): 335.
59 Zhang Yanyuan 張彥遠, Lidai minghua ji 歷代名畫記 (Beijing: Renmin meishu
chubanshe, 1963), 6:77. English trans. see Bush Susan and Hsio-yen Shih, Early
Chinese Texts on Painting (Cambridge, MA: Published for the Harvard-Yenching
Institute by Harvard University Press, 1985), 79–80.
60 Weishu, 108: 2817.
61 Ibid.
62 Wei Zheng (580–643), Suishu (History of the Sui dynasty) (Beijing: Zhonghua
shuju, 1973), 11:238, and Weishu, 91:1971.
63 Albert E. Dien, Six Dynasties Civilization (New Haven: Yale University Press,
2007), 315.
64 Yang Xuanzhi, Luoyang qielan ji (Record of the Monasteries in Luoyang),
(Shanghai: Shanghai shudian chubanshe, 2000), 107–8. English translation with
revision, see William John Francis Jenner, Memories of Lo-yang: Yang Hsuan-
chih and the Lost Capital (493–534) (New York: Clarendon Press, Oxford Uni-
versity Press, 1981), 203.
65 Daoshi 道世, Fayuan zhulin 法苑珠林 (Beijing: Zhonghua shuju, 2003), 16:543.
66 Ibid.
67 Zhang, Lidai minghua ji, 8:95.
68 Ibid., 5:68.
69 Ibid., 2:24.
70 Ibid., 2:25.
71 Daoshi, Fayuan zhulin, 16:543.
72 Gu Kaizhi, “Lun Hua,” in Lidai minghua ji, 5:73.
73 Xie He (Southern Qi), “Guhua pinlu 古畫品錄 (Classification record of ancient
paintings),” in Congshu jicheng chubian. Ed. Wang Yunwu (Shanghai: Shangwu
yinshuguan, 1936), 1:7.
74 Shen Yue 瀋約 (441–513), Songshu (History of the Southern Song), (Beijjing:
Zhonghua shuju, 1974), 93:53:2274.
75 Zhang, Lidai minghua ji, 5:75.
76 Ibid.
77 Xie, “Guhua pinlu,” 1:2.
78 Zhang, Lidai minghua ji, 6:77.
188 Phase three – the remaining splendor
79 Another example is the slender figures on the brick-relief walls, excavated in
1958, in the tomb at the village of Xuezhuang 學庄, Dengxian Prefecture 邓县,
Henan province and the terracotta figures there exhibit the same feature. All of
them are typical of the linear emphasis in Chinese painting.
80 Liu Zhiyuan, Chengdu Wanfosi shike yishu (Beijing: Zhongguo gudian yishu
chubanshe, 1958), 12.
81 Zhang, Lidai minghua ji, 7:91.
82 Ibid.
83 Tao Zongyi 陶宗儀 (1329–1410), Shuofu 說郛, 92:1226.
84 Xu Song 許嵩, Jiankang shilu (Shanghai: Shanghai guji chubanshe, 1987),
17:481.
85 Guo Ruoxu, “Tuhua jianwen zhi (Experiences in painting),” in Huashi congshu.
Ed. Yu Anlan (Shanghai: Shanghai renmin meishu chubanshe, 1963), 1:10.
86 Zhang, Lidai minghua ji, 6:77.
87 Ibid.
7 Postscript

It was over 80 years ago that the well-known Japanese scholars Mizuno
Seiichi and Nagahiro Toshio conducted the first comprehensive studies of
the Yungang cave complex in Shanxi. Inspired by their meticulous research,
and my curiosity about Yungang, I spent over a decade investigating this
splendid Buddhist monument. My fascination began with these questions:
Why, when, and under what circumstances was Yungang produced? Who
played significant roles at various stages and paid for them? Which cave was
built first? What were the essential components of this Buddhist monument?
Where did the monks live and translate the scriptures? What kinds of liturgi-
cal practices were conducted there? Where did the tradition of making colos-
sal images come from? How long did it take to make a rock-cut cave space?
Yungang was built as a royal memorial sponsored by Northern Wei rulers,
high-ranking officials, eminent clergy, and devotees in a rather short period.
At the time, Yungang amassed human resources and material from different
places, both foreign and domestic. During the second phase of Yungang, the
“Yungang moshi (Yungang style)” had been well established. After the trans-
fer of the capital southward to Luoyang in 494, the splendour of Yungang
remained. During the Liao and Jin periods, repairs and reconstructions were
made. The new discovery above caves 5 and 6, for instance, reveal that a
freestanding Buddhist monastery was reconstructed during the Liao dynasty
on the foundation of a Northern Wei monastery.
One of the goals in this book has been to date and propose a construction
sequence of the second- and third-phase cave temples. It is my belief that
cave 13 is the earliest cave temple in the second-phase caves. It is followed by
cave 11. They were likely to have been constructed as a pair under Emperor
Xianwen. Cave 12 was added to the space between the two. The four paired
caves were commissioned by Emperor Xiaowen (caves 7 and 8, 9 and 10,
and 5 and 6) and Wang Yu (caves 1 and 2). I have also divided the third-
phase cave temples into four groups and three construction periods. In doing
so, observation of the external walls of the caves was an important approach
in this research in reconstructing a timeline for the caves.
The recent archaeological discoveries in and around Yungang have shed
great fresh light on a comprehensive view of the Yungang complex as a
190 Postscript
significant rock-cut Buddhist monastery, and corroborate the mention in
the Jinbei stele inscription of the “stone chambers on the northern terrace”
(i.e., the monks’ cells above the caves) in Yungang. These discoveries, there-
fore, provide a better picture of the important components of the Yungang
complex, as well as the monastic life in Yungang as to where the monks
lived and worked. The archaeological evidence also indirectly testifies to the
authenticity of the records in the Jin stele inscription. Without the discov-
ery of further written documentation, the Jin stele inscription will remain
a valuable source in our study of Yungang. In addition, the discovery of
the tombs in Pingcheng has provided important clues about the recipro-
cal impact between the religious and secular material culture at the time
Yungang was constructed. It is hoped that more freestanding monastery
remains will be excavated in the near future to shed significant light on art
and architecture in Yungang. Freestanding monasteries were an important
resource for Yungang.
One of the key questions inquires what the liturgical function of the caves
are. If the five Tanyao cave temples were constructed as family memorial,
how did the other caves function? I have argued that the cave temples in
Yungang were not ideal for meditation since they were situated on the road
between the two capitals, and were primarily statue shrines. This is one of
the most distinguishing features of the Yungang cave temples. Many of the
caves were constructed not merely for the sake of having them built, for
merits and virtues; rather, some, such as caves 1 and 2, 5 and 6, 9 and 10,
and caves 11 and 12, etc., had liturgical functions such as Changdao. The
function of the cave temples will remain my inquiry.
The sinicization of Buddhist iconography in northern China was a long
process, and far more complicated than previous explanations, such as
reforms by the court or influences from the south, would have us believe.
I have tried to show that over the three phases of Yungang, western influ-
ence gradually weakened and finally disappeared altogether during the third
phase.
The difficulties in studying Yungang are that we do not have much liter-
ary or visual evidence. Yungang cave temples carry very few inscriptions,
much less compared with those in Longmen. As such, it is very difficult to
construct a dating scheme with complete confidence. Therefore, dating and
periodization of the Yungang cave temples will remain a topic of scholarly
debate. Part of the problem has been that sculpture and architecture are
considered merely labor work in Chinese civilization. Great sculptors and
architects have rarely been as recorded and appreciated as they are in the
west, such as the “Renaissance Man” Leonardo Da Vinci (1452–1519) and
his contemporary Michelangelo (1475–1564). The technological and scien-
tific ingenuity of the workers involved in making the Yungang caves is still
largely unknown and remains a big question for scholars.
I believe that there remain other fruitful lines of inquiry. Among them is a
cluster of questions around the donors who sponsored the niches and images,
 Postscript 191
and the nature and training of the artists, artisans, and laborers who built
them. And undoubtedly, a more comprehensive study of the Shanxi caves
and their intimate relationship with Yungang is overdue. The distributions
of these caves (see plate 1), all of which are situated within the triangle of
the three capitals (Pingcheng, Luoyang, and city of Ye) of the Northern
Dynasties, well reflect the social and political situation. Further questions to
explore are: how far did the “Yungang style” spread both temporally and
geographically? Did the tradition of making colossal images come from Kizil,
Gaochang, and the Hexi corridor?
I very much hope that this study will inspire further research on Yungang,
one of the most magnificent Buddhist monuments of all time.
Classification of the caves in the
third phase

Cave Three One Two No Three Thousand Chaitya

Structure niches on niche on storied niche walls Buddha cave
three walls the north niches 无龛 with niches 支提窟
三壁三龛 wall and on three two 千佛龛
storied walls niches
niches on 三壁重龛 三壁二
sidewalls 龛
北壁一龛,
侧壁重龛
Cave 16 caves: 7 caves: 6 caves: 3 caves: Caves Cave 36-2 Cave 39
Numbers 22-1, 23-1, 21, 22, 27, 29, 23, 25-2 5-10 and
24, 25, 27-2, 24-1, 26, 30, 31, and 28-2 5-11
28, 32-11, 33-3, 33-4, 32-9, and
32-12, 33, and 38 36
33-6, 34, 35,
37, 38-3, 40,
and 41
Groups and periodization of the
third-phase caves

Groups Periodization Caves Dates Features

Group Phase One Caves 21, Approximately Many medium-sized caves.

One 22, 23, 33, before the In architecture, caves with
27, 29, 30, transfer of the storied niches vertically
31, 25, 28, capital until the arranged on three walls
24, and 28-2 first year of predominate. Niches are
Zhengshi 正始 mainly either pointed-arch
(494–504) or trabeated. Canopied
niches with looped
curtains appeared. The
iconographic composition
of the Śākyamuni and
the Prabhutatna became
increasingly popular.
Other compositions
(the Śākyamuni and the
Prabhutaratna with a
seated Buddha and a
cross-ankled Bodhisattva;
the Śākyamuni and the
Prabhutaratna with a
cross-ankled Buddha and a
cross-ankled Bodhisattva;
the Śākyamuni and the
Prabhutaratna with two
seated Buddhas; and a
cross-ankled Bodhisattva
with two seated Buddhas)
also appeared. The images
appear slender and
continue the style in caves
6, 1, and 2.
(Continued )
(Continued)

Groups Periodization Caves Dates Features

Group Phase Two Caves 34, Approximately All medium-sized caves.
Two 35, 37, and between the Three niches on three
39 first year of walls predominate in
Zhengshi and architecture. Chaitya
before the cave appeared. The
fourth year of folding-screen trabeated
Yanchang 延昌 niches appeared. The
(504–15) iconographic composition
is primarily composed of
the three Buddhas. Also,
a cross-legged seated
Buddha together with a
cross-ankled Bodhisattva
and a seated Buddha with
legs pendant. Disciples
began to appear for the
first time. The images
continued the style of
the first group with
innovation. The drapery of
the Buddha images began
to have two series of arcs
side by side. Bodhisattva
began to have a large ring
in the center for a celestial
robe to cross through.
Group Phase Three Caves 5-10 Around the Mainly small caves. Three
Three and 5-11 fourth year of niches on three walls
Yanchang (515) predominate. One niche
Group Caves 22-1, Between the on the main wall with
Four 23-1, 24-1, fourth year of storied niches on sidewalls
25-2, 26, Yanchang and were also popular. The
27-2, 32-11, the fifth year folding screen trabeated
32-12, 32-9, of Zhengguang niches were popular.
33-3, 33-4, (515–24) The composition of
36, 36-2, the Śākyamuni and the
36-6, 38, 40, Prabhutatna or Śākyamuni
and 41 became the primary
images of the caves. The
composition of Śākyamuni
and Maitreya is often seen
as well. The images appear
elegant and slender with
elongated body.
Character glossary

An Faqin 安發欽
An Xi 安息
Antarvasaka 安陀會
Ayuwang jing 阿育王經
Ayuwang zhuan 阿育王經傳
Bai Zheng 白整
Baizizhang 百子帳
Baochang 寶唱
Baoshan 包山
Beitai 北台
Bian 辯
Bingwu 丙午
Bingyangdong 賓陽
Birang 避讓
Bo 博
Boshanlu 博山爐
Cai 才
Cao Zhongda 曹仲達
Caoyi chushui 曹衣出水
Chan miyao fa jing 禪秘要法經
Changdao 唱導
Changdao sengren 唱導僧人
Changfu 常服
Changle 長樂
Changnayeshe 常那邪捨
Changzhi 長治
Chen Qingzhi 陳慶之
Chengming 承明
Chiwei 鴟尾
Chongfu 崇福
Chongguang Palace 崇光宮
Chongjiao temple 崇教寺
Chuanzuo wuqiong 傳祚無窮
196 Character glossary
Da Song sengshi lüe 大宋僧史略
Dachangqiuqing 大長秋卿
Dai Kui 戴逵
Dai Yong 戴顒
Dajiyi shenzhou jing 大吉義神咒經
Dao’an 道安
Dao’xin 道馨
Daorentong 道人統
Daoshi 道世
Dapuo 打破
Dazhusheng 大住聖
Dinglin 定林
Dingmao 丁卯
Dingyou 丁酉
Dong Mi 董謐
Douba 斗八
Doushuai 兜率
Dousi 斗四
Fafu 法服
Fangshan 方山
Fashanglu 法上錄
Faxian 法顯傳
Fayuan zhulin 法苑珠林
Fei Changfang 費長房
Fo suoxing zan 佛所行讚經
Fodianku 佛殿窟
Foshuo guanfo sanmei haijing 佛說觀佛三昧海經
Fotucheng 佛圖澄
Fotuhu 佛圖戶
Fotuobatuoluo 佛陀跋陀罗
Fozhi biqiu liuwutu 佛製比丘六物圖
Fudao minsu 輔導民俗
Fufazang yinyuan zhuan 付法藏因緣傳
Fugui wansui 富貴萬歲
Fushan 釜山
Gao Yun 高允
Gaomiaoshan 高廟山
Gaoping 高平
Gaoseng zhuan 高僧傳
Gongwei zhibian 宮闈之變
Gongzong 恭宗
Gu Kaizhi 顧愷之
Guan siweiyi pin 觀四威儀品
Guanfo sanmei fa 觀佛三昧法
Guannei 關內
 Character glossary 197
Guhua pinlu 古畫品錄
Guifu 龜趺
Guiren 貴人
Gujin yijing tuji 古今譯經圖記
Guo Ruoxu 郭若虛
Guoqu xianzai yinguo jing 過去現在因果經
Guzang 姑臧
Hanquandong 寒泉洞
Helinge’er 和林格尔
Heng’an 恆安
Hengdi 橫笛
Heping 和平
Hou Hanshun 後漢書
Hu 斛
Huangfugong 皇甫公
Huangjiu si 皇舅寺
Huangxing 皇興
Huayan 華嚴
Huguo 護國
Huifusi 晖福寺
Huishi 惠始
Huren 胡人
Hutengwu 胡騰舞
Jiang Shaoyou 蔣少游
Jiangjing 講經
Jiangjing fashi 講經法师
Jiangjingtang 講經堂
Jiankang shilu 建康實錄
Jiaoshansi 焦山寺
Jiasha 袈裟
Jinbei 金碑
Jingchong 鯨崇
Jingdu sanmei jing 净度三昧经
Jingmen 荊門
Jingming 景明
Jingtu sanmei jing 淨土三昧經
Jingzhao 京兆王
Jiu’er juezhi 就而镌之
Juqu Mujian 沮渠牧犍
Kaihua 開化
Kaihuang 開皇
Kaiyuan shijiao lu 開元释教錄
Kashiga 喀什
Konghou 箜篌
Lanruo 蘭若
198 Character glossary
Li Daoyuan 酈道元
Li Yi 李弈
Liangzhou 涼州
Libaiku 禮拜窟
Lidai sanbao ji 歷代三寶記
Lingchuan 陵川
Lingyan 靈岩
Lingyu 靈裕
Lishe 裏社
Liu Chang 劉昶
Liu Xiaobiao 劉孝標
Longwangmiao 龍王廟
Lou Rui 婁叡
Lu Tanwei 陸探微
Lubanyao 魯班窯
Luyeyuan 鹿野苑
Maoxian 茂縣
Mingseng Zhuan 名僧傳
Miu Quansun 繆荃蓀
Mofa 末法
Mokemoye jing 摩诃摩耶經
Moni 摩尼
Nanqishu 南齊書
Nanshi 南史
Niepan jing 涅槃經
Otsuka Nobuo 大塚伸夫
Paixiao 排簫
Pianshan 偏衫
Pingcheng 平城
Pingqi 平棊
Pingqi zaojing 平棊藻井
Pingwen 平文
Pipa 琵琶
Putai 普泰
Putong 普通
Puyao jing 普曜經
Qian’er Qingshi 鉗耳慶時
Qidaoren 漆道人
Qiunabatuoluo 求那跋陀羅
Qixiashan 棲霞山
Ren tian jiao 人天教
Samghapala 僧伽婆羅
Samghati 僧伽梨
Sanbi sankan 三壁三龕
Sanqian daqian shijie 三千大千世界
 Character glossary 199
Sengfangku 僧房窟
Sengxian 僧顯
Sengyi 僧衣
Sengzhihu 僧祇戶
Shale 沙勒
Shaling 沙嶺
Shamentong 沙門統
Sheng 聲
Shengle 盛樂
Shenrui 神瑞
Shigudong 石鼓洞
Shili 師禮
Shilihe 十里河
Shitanghui 石堂會
Shixian 師賢
Shuijing zhu 水經注
Shuntianfu zhi 順天府志
Shuofu 說郛
Shuopingfu zhi 朔平府志
Sifenlu shanbu suiji jiemo shuzhengyuanji 四分律隨機羯磨疏正源記
Sifenlu xingshichaozichiji 四分律行事鈔資持記
Sima Jinlong 司馬金龍
Siwei lüeyao fa 思維略要法
Siyuan lingtu 思遠靈圖
Song Dexing 宋德興
Song Shaozu 宋紹祖
Songshu 宋書
Sujiang 俗講
Sujiangshi 俗講師
Ta E 他恶
Tai’an 太安
Taihua 太華
Taipingzhenjun 太平真君
Taiyan 太延
Takahashi Hisao 高橋尚夫
Tandu 曇度
Tanjing 昙景
Tanmei 曇媚
Tanyao 曇曜
Tao Zongyi 陶宗儀
Tianci 天賜
Tiangong 天宮
Tianhua 天花
Tianxing 天興
Tiwei Boli jing 提謂波利經
200 Character glossary
Tongle 通樂
Tongzi 童子
Tulufan 吐魯番
Tuyuk 吐峪溝
Uttarasanga 鬰多羅僧
Wan Shen 菀申
Wanfosi 万佛寺
Wang Yu 王遇
Weishu 魏書
Weizi caves 魏字洞
Wudianding 庑殿頂
Wuguantun cave temple complex 吳官屯
Wuhuadong 五華洞
Wushen 戊申
Wuwei 武威
Wuzhoushan 武周山
Xie He 謝赫
Xijing zhi 析津志
Xing’an 興安
Xingguang 興光
Xinyou 辛酉
Xiong Mengxiang 熊夢祥
Xishanqiao 西善橋
Xu Song 許嵩
Xu Xianxiu 徐顯秀
Xuangao 玄高
Xuezhuang 學庄
Yanbei shiyuan 雁北師院
Yanchang 延昌
Yangtoushan 羊頭山
Yanxing 延興
Yaogu 腰鼓
Yihun 乙浑
Yijing 義淨
Yinyan jiegou 因岩結構
Yishe 邑社
Yiweijing 疑偽經
Yongming 永明
Yuan 院
Yuanguang 元光
Yuanzhao 元照
Yungang bu 雲岡堡
Yungang moshi 雲岡模式
Yunzhongtu 雲中圖
Za baozang jing 雜寶藏經
 Character glossary 201
Zengxiu jiaoyuan qinggui 增修教苑清規
Zhang Gui 張軌
Zhang Huaiguan 張懷瓘
Zhang Sengyou 張僧繇
Zhang Tan 張潭
Zhang Yong 張永
Zhanzhang 毡帳
Zhao Mu zhi 昭穆制
Zhengguang 正光
Zhengguansi 正觀寺
Zhengshi 正始
Zhenguan 貞觀
Zhenguo 鎮國
Zhi Qian 支谦
Zhi Sheng 智升
Zhizheng 至正
Zhu Yi 朱异
Zitui 子推
Zuochan sanmei jing 坐禪三昧經
Selected bibliography

Primary sources
Ayuwang zhuan 阿育王傳 (Aśokāvadana). Trans. An Faqin 安法欽 (Western Jin
dynasty, 265–316 C.E.). T 50, no. 2042.
Chan miyao fa jing 禪秘要法經 (The Scripture on the Esoteric Essential Methods of
Meditation). Trans. Kumārajīva (Jiumoluoshi 鳩摩羅什, 344–413). T 613, no. 252.
Chu sanzang ji ji 出三藏記集 (Collected Works of Buddhism from All Dynasties).
Comp. Sengyou 僧佑 (445–518). T 55, no. 2145.
Da Song sengshi lüe 大宋僧史略 (Song Dynasty Compendium of Monastic History).
Comp. Zanning 贊寧 (919–1001). T 54, no. 2126.
Da Tang liudian 大唐六典 (The Six Statutes of the Tang Dynasty). Comp. Li Linfu
李林甫 (d. 752), et al., in the Tang dynasty. Anno. Chen Zhongfu 陳仲夫. Beijing:
Zhonghua shuju, 1992.
Da Tang neidian lu 大唐內典錄 (Catalogue of Buddhist Works in the Great Tang).
Comp. Daoxuan 道宣 (596–667). T 55, no. 2149.
Da Tang xiyuji 大唐西域記 (Records from the Western Regions of the Great Tang),
Comp. Xuanzang 玄奘 (602–664). T 51, no. 2087.
Daban niepan jing 大般涅槃經 (Mahāparinirvāna Sūtra). Trans. Dharmaksema (Tan-
wuchen 曇無讖, 385–433). T 12, no. 374.
Dajiyi shenzhou jing 大吉義神咒經 (Sutra of Great Auspicious Significance of Magic
Spell). Trans. Tanyao 曇曜. T 21, no. 1335.
Dasheng zaoxiang gongde jing 大乘造像功德經 (Tathāgatapratibimbapratistha
Nuśamsā). Trans. Devendraprajna (Tiyunbanruo 提雲般若, fl. 689–691). T 16, n. 694.
Fang Xuanling 房玄齡 (578–648), et al., Jin Shu 晉書 (History of Jin). Beijing: Zhon-
ghua shuju, 1974.
Fayuan zhulin 法苑珠林 (The Forest of Gems in the Garden of the Dharma). Comp.
Daoshi 道世 (d. 683). T 53, no. 2122.
Fobenxing jing 佛本行經 (Buddhapurvakarya-Sutra). Trans. Baoyun 寶雲 (Liu Song
dynasty). T 4, no. 193.
Foguo ji 佛國記 (Records of the Buddhist Countries). Comp. Faxian 法顯. T 51,
no. 2085.
Foshuo zaota gongdejing 佛說造塔功德經 (Sūtra of merit and virtue accumulation
by making a stūpa). Trans. Divākara (地婆訶羅 7th c.). T 16, no. 699.
Fo suoxing zan 佛所行讚 (Acts of the Buddha). By Asvaghosa 馬鳴 (Lived in the First
Century C.E.). Trans. Dharmaraksa 曇無讖 (385–433). T 04, no. 192.
Fozhi biqiu liuwutu 佛制比丘六物图 (Diagram of Six Buddhist Essentials). Trans.
Yuanzhao 元照. T 45, no. 1900.
 Selected bibliography 203
Fufazang yinyuan zhuan 付法藏因緣傳 (A History of the Indian Patriarchs). Trans.
Tanyao. T 50, no. 2058.
Gaoseng zhuan 高僧傳 (Biographies of Eminent Monks). Comp. Hui Jiao 慧皎.
T 50, no. 2059.
Genben shuo yiqie youbu pinaiye zashi 根本說一切有部毘奈耶雜事 (Mūlasarvāstivāda).
Trans. Yijing 義淨 (635–713). T 23, no. 1443.
Guan Mile pusa shangsheng doushuai tianjing 觀彌勒菩薩上生兜率天經 (Visualizing
the Maitreya Bodhisattva’s Ascent to the Tusita Heaven). Trans. Juqu Jingsheng
沮渠京聲 (Liu Song dynasty, 420–79 C.E.). T 14, no. 452.
Guan Wuliangshoufo jing 觀無量壽佛經 (Amitāyur-dhyāna-sūtra; Amitāyur-
buddha-dhyāna-sūtra). Trans. Kālayśas 畺良耶舍 (Liu Song dynasty, 420–79
C.E.). T 12, no. 365.
Guanfo sanmei hai jing 觀佛三昧海經 (Buddha-dhyāna-samādhi-sāgara-sūtra). Trans.
Buddhabhadra 佛陀跋陀羅 (Eastern Jin dynasty, 317–420 C.E.). T 15, no. 645.
Guang hongming ji 广弘明集 (Expanded Anthology of Extended Brilliance). Comp.
Daoxuan. T 52, no. 2103.
Gujin yijing tuij 古今譯經圖紀 (Seal of Ancient and Modern Sūtra Translation).
Comp. Jingmai 靖邁 (fl. 645–665). T 55, no. 2151.
Guoqu xianzai yinguo jing 過去現在因果經 (Sutras on Causes and Effects of the Past
and Present). Trans. Gunabhadra 求那跋陀羅 (394–468). T 3, no. 189.
Hongming ji 弘明集 (Collected Essays on Buddhism from the Eastern Han to the
Southern Liang). Comp. Sengyou 僧祐 (445–518). T 52, no. 2102.
Jin guangming jing 金光明經 (Dharmaksema). Trans. Tanwuchen 曇無讖. T 16,
no. 663.
Kaiyuanshijiaolu 開元釋教錄 (A Catalogue of Buddhist Scriptures Edited in the Kai-
yuan Era). Comp. Zhisheng 智昇 (669–740). T 55, no. 2154.
Kang Pian 康騈. Jutan lu 劇談錄 (Records of Talks on Drama). Shanghai: Shangwu
yinshuguan, 1922.
Li Delin 李德林 (530–90) and Li Baiyao 李百藥 (565–648). Beiqishu 北齊書 (History
of the Northern Qi). Beijing: Zhonghua shuju, 1972.
Li Yanshou 李延寿 (fl. 618–76). Beishi 北史 (History of the Northern Dynasties).
Beijing: Zhonghua shuju, 1974.
———. Nanshi 南史 (History of the Southern Dynasties). Beijing: Zhonghua shuju,
1975.
Lidai sanbao ji 歷代三寶記 (Records of the Three Jewels through All Dynasties),
Comp. Fei Changfang 費長房. T 49, no. 2034.
Linghu Defen 令狐徳棻 (583–666). Comp. Zhoushu 周書 (History of the Zhou).
Beijing: Zhonghuashuju, 1971.
Luoyang qielan ji 洛陽伽藍記 (Record of the Monasteries in Luoyang). Comp. Yang
Xuanzhi 楊衒之. T 51, no. 2092.
Miaofa lianhua jing 妙法蓮華經 (Saddharmapundāríka Sūtra). Trans. Kumārajīva.
T 9, no. 262.
Mile xiasheng chengfo jing 彌勒下生成佛經 (Maitreyavyākaraṇa). Trans. Kumārajīva
(Yao Qin dynasty, 384–417 C.E.). T 14, no. 423.
Mohe moye jing 摩诃摩耶经 (Sūtra of Great Māyā). Trans. Tanjing 曇景 (Southern
Qi dynasty, 479–502). T 12: 383, no. 1013.
Puyao jing 普曜經 (Lalitavistara). Trans. Dharmaraksa 竺法護 (Western Jin dynasty,
265–317). T 3, no. 186.
Shen Yue 沈約 (441–513). Songshu 宋書 (History of the Southern Song). Beijjing:
Zhonghua shuju, 1974.
204 Selected bibliography
Shuijing zhu 水經注 (Annotated Commentaries on the Classics of Waterways) by
Li Daoyuan 酈道元. Anno. Chen Qiaoyi 陳橋驛. Beijing: Zhonghua shuju,
2007.
Shuoping fuzhi (Shanxi) 朔平府志 (Gazetteer of Shuoping Prefecture, Shanxi Prov-
ince). Eds. Liu Shiming 劉士銘, et al. China: s.n., 1731.
Sifenlü juiji jiemoshu zhengyuanji 四分律隨機羯磨疏正源記 (Commentary on the
Revised Monastic Procedures of the Four-Part Vinaya). Comp. Yunkan 允堪 (of
the Song dynasty). X 40, no. 726.
Sifenlü xingshi chao zichi ji 四分律四分律行事鈔資持記 (A Commentary on the Dhar-
maguputaka Vinaya). Comp. Yuanzhao 元照 (of the Song dynasty). T 40, no. 1805.
Sita ji 寺塔記 (Records of Monasteries and stūpas) by Duan Chengshi 段成式. T 51,
no. 2093.
Siwei Lüeyao Fa 思維略要法 (Outlining the Methods of Contemplation). Trans.
Attributed to Kumārajīva. T 15, no. 617.
Song Gaoseng zhuan 宋高僧傳 (Eminent Monks of the Song Dynasty). Comp. Zan-
ning 贊寧. T 50, no. 2061.
Taizi ruiying benqi jing 太子瑞應本起經 (Kumāra-kuśalaphala-nidāna-sūtra). Trans.
Zhiqian 支謙 (Wu dynasty, 222–280). T 3, no. 185.
Tiwei Boli Jing 提謂波利經 (The Sūtra of Trapusa and Bhallika). Pelliot 3732 and
Stein 2051, et al.
Wei Shou 魏收 (505–572). Weishu 魏書 (History of the Northern Wei). Beijing:
Zhonghua shuju, 1974.
Wei Zheng 魏徵 (580–643). Suishu 隋書 (History of the Sui dynasty). Beijing: Zhon-
ghua shuju, 1973.
Xianyujing 賢愚經 (Damamūka-nidana-sūtra). Trans. Huijue (Northern Wei dynasty,
386–534 C.E.), et al. T 3, no. 202.
Xiao Zixian 蕭子顯 (487–537). Nanqishu 南齊書 (History of the Southern Qi). Bei-
jing: Zhonghua shuju, 1974.
Xie He 謝赫 (Southern Qi). Guhua pinlu 古畫品錄 (Classification Record of Ancient
Paintings), in Congshu jicheng chubian. Ed. Wang Yunwu 王雲五. Shanghai:
Shangwu yinshuguan, 1936, vol. 1, 7.
Xiuxing benqi jing 修行本起經 (Caryā-nidāna-sūtra). Trans. Mahābal 竺大力 and
Kang Mengxiang 康孟祥 (Eastern Han dynasty, 25–220 C.E.). T 3, no. 184.
Xu gaoseng zhuan 續高僧傳 (The Continuation of the Biographies of Eminent
Monks). Comp. Daoxuan. T 50, no. 2060.
Xu Song 許嵩. Jiankang shilu 建康實錄 (Veritable Records of Jiangkang [reign]).
Shanghai: Shanghai guji chubanshe, 1987.
Yichu pusa benqi jing 異出菩薩本起經 (Abhiniṣkramaṇa). Trans. Nie Daozhen 聶道
真 (Western Jin dynasty, 265–317). T 3, no. 188.
Za baozang jing 雜寶藏經 (Sūtra of the Miscellaneous Treasures). Trans. Tanyao.
T 4, no. 203.
Zengxiu jiaoyuan qinggui 增修教苑清規 (Expanded Rule for Teachings Monaster-
ies). Comp. Ziqing 自慶. X 57, no. 968.
Zhang Yanyuan 張彥遠. Lidai minghua ji 歷代名畫記 (Record of Famous Painters
of Successive Dynasties) in Huashi congshu (10 juan). Ed. Yu Anlan. Shanghai:
Shanghai renmin meishu chubanshe, 1963.
Zhao Yi 赵翼. Ershi’er shi zhaji 廿二史箚記 (36 juan). Shanghai: Shanghai guji chu-
banshe, 1995.
 Selected bibliography 205
Zhongbenqi jing 中本起經 (Madhyama-ityukta Sutra). Trans. Tanguo 曇果 and
Kang Mengxiang 康孟祥 (Later Han dynasty 25–220). T 4, no. 196.
Zuochan sanmei jing 坐禪三昧經 (The Scripture on Concentration in Sitting Medita-
tion). Trans. Kumārajīva. T 15, no. 614.

Secondary sources

Works in Western languages

Abe, Stanley. “Art and Practice in a Fifth Century Chinese Buddhist Cave Temple.”
Ars Orientalis 20 (1990): 1–31.
———. “Book Review.” Review of Written and Unwritten: A New History if the
Buddhist Caves at Yungang, by James O. Caswell. Journal of the American Acad-
emy of Religion 58 (Spring 1990): 116–18.
———. Ordinary Images. Chicago and London: University of Chicago Press, 2002.
Barrett, Tim H. “Book Review.” Review of Written and Unwritten: A New History
of the Buddhist Caves at Yungang, by James O. Caswell. Bulletin of the School of
Oriental and African Studies-University of London 55 (1992): 359–60.
Barthes, Roland and Stephen Heath. Image, Music, Text. New York: Hill, 1977.
Bunker, Emma. “Early Chinese Representations of Vimalakīrti.” Artibus Asiae 30, 1
(1968): 28–52.
———. “The Spirit Kings in Sixth Century Chinese Buddhist Scupture.” Archives of
the Chinese Art Society of America 18 (1964): 26–37.
Bush, Susan and Hsio-yen Shih. Early Chinese Texts on Painting. Cambridge, MA:
Published for the Harvard-Yenching Institute by Harvard University Press, 1985.
Campany, Robert F. “Book Review.” Review of Written and Unwritten: A New His-
tory of the Buddhist Caves at Yungang, by James O. Caswell. Journal of Asian
History 24, 1 (1990): 96–7.
Caswell, James O. “The ‘Thousand-Buddha’ Pattern in Caves XIX and XVI at Yun-
kang.” Ars Orientalis 10 (1975): 35–54.
———. Written and Unwritten: A New History of the Buddhist Caves at Yungang.
Vancouver: University of British Columbia Press, 1988.
Chavannes, Édouard. Mission archéologique dans la Chine septentrionale. Paris:
Ernest Leroux, 1909.
———. Six monuments de la sculpture chinoise. Bruxelles and Paris: G. van Oest &
Cie, 1914.
Chʻen, Kenneth K. S. The Chinese Transformation of Buddhism. Princeton, NJ:
Princeton University Press, 1973.
Copp, Paul F. The Body Incantatory: Spells and the Ritual Imagination in Medieval
Chinese Buddhism. New York: Columbia University Press, 2014.
Dehejia, Vidya. Discourse in Early Buddhist Art: Visual Narratives of India. New
Delhi: Munshiram Manoharlal Publishers, 1997.
Dien, Albert. Six Dynasties Civilization. New Haven: Yale University Press, 2007.
———. Ed. State and Society in Early Medieval China. Stanford: Stanford University
Press, 1990.
Foucher, Alfred. The Life of the Buddha: According to the Ancient Texts and Monu-
ments of India. Middletown, CT: Wesleyan University Press, 1963.
206 Selected bibliography
Fowler, Alastair. Renaissance Realism: Narrative Images in Literature and Art.
Oxford: Oxford University Press, 2003.
Fu Xinian. “The Three Kingdoms, Western and Eastern Jin, and Northern and
Southern Dynasties,” in Chinese Architecture. Ed. Nancy S. Steinhardt. New
Haven, CT: Yale University Press and Beijing: New World Press, 2002, 61–89.
Gernet, Jacques. Buddhism in Chinese Society: An Economic History from the Fifth
to the Tenth Centuries. New York: Columbia University Press, 1995.
Granoff, Phyllis. “The Gift of the Two Merchants: Defining the Buddhist Commu-
nity through Story.” East and West 55, 1/4 (December 2005): 129–38.
Greene, Eric. Meditation, Repentance, and Visionary Experience in Early Medieval
Chinese Buddhism. Ph.D. Dissertation, University of California, Berkeley, 2012.
Gregory, Peter N., ed. Traditions of Meditation in Chinese Buddhism. Honolulu:
University of Hawaii Press, 1986.
Griswold, Alexander B. “Prolegomena to the Study of the Buddha’s Dress in Chinese
Sculpture.” Artibus Asiae 26, 2 (1963): 89–90.
Hogan, Patrick C. Narrative Discourse: Authors and Narrators in Literature, Film,
and Art. Columbus: Ohio State University Press, 2013.
Holliday, Peter J. Narrative and Event in Ancient Art. Cambridge: Cambridge Uni-
versity Press, 1993.
Holmgren, Jennifer. Annals of Tai-Early To’pa History According to the First Chap-
ter of the Weishu. Canberra: Australian National University Press, 1982.
———. “Empress Ling of the Northern Wei and the Tuoba Sinicization Question.”
Papers on Far Eastern History 18 (1978): 160–70.
Howard, Angela Falco. “Liang Patronage of Buddhist Art in the Gansu Corridor
during the Fourth Century and the Transformation of a Central Asian Style,” in
Between Han and Tang: Religious Art and Archaeology in a Transformative
Period. Ed. Wu Hung. Beijing: Wenwu chubanshe, 2000, 235–75.
———. “Miracles and Visions among the Monsstic Communities of Kucha, Xinji-
ang.” Journal of Inner Asian Art and Archaeology 2 (2007): 77–87.
Howard, Angela Falco and Giuseppe Vignato. Archaeological and Visual Sources of
Meditation in the Ancient Monasteries of Kuča. Leiden and Boston: Brill, 2015.
Huntington, John C. “Iconography of the Tanyao Caves at Yungang.” Oriental Art
33, 2 (1986): 142–57.
Hurvitz, Leon. Weishou: A Treatise on Buddhism and Daoism. Kyoto: Jimbun
Kagaku Kenkyusho, Kyoto University, 1956.
Jenner, William John Francis. Memories of Lo-yang: Yang Hsuan-chih and the Lost
Capital (493–534). New York: Clarendon Press, Oxford University Press, 1981.
Karetzky, Patricia Eichenbaum. Early Buddhist Narrative Art: Illustrations of the
Life of the Buddha from Central Asia to China, Korea, and Japan. Lanham, MD:
University Press of America, 2000.
Kieschnick, John. The Eminent Monk: Buddhist Ideals in Medieval Chinese Hagiog-
raphy. Honolulu: University of Hawaii Press, 1997.
———. The Impact of Buddhism on Chinese Material Culture. Princeton: Princeton
University Press, 2003.
Lai, Whalen W. “The Earliest Folk Buddhist Religion in China: Ti-wei Po-li Ching and
Its Historical Significance,” in Buddhist and Taoist Practice in Medieval Chinese
Society. Ed. David W. Chappell. Honolulu: University of Hawaii Press, 1987, 13–14.
Ledderose, Lothar. Ten Thousand Things: Module and Mass Production in Chinese
Art. Princeton, NJ: Princeton University Press, 2000.
 Selected bibliography 207
Lewis, Mark E. China between Empires: The Northern and Southern Dynasties.
Cambridge, MA: Belknap Press of Harvard University Press, 2009.
———. “From Gandhāra to Yungang: Design of a Free-Standing Buddhist Monas-
tery.” Ancient Pakistan 23, 12 (2012): 13–54.
———. “The Image of Maitreya in the Chētiyagharas of Kizil, Kucha.” Gandhara
Studies 7 (2013): 11–21.
Li Yuqun. “Classification, Layout, and Iconography (Trans. John Kieschnick),” in
Early Chinese Religion: Part One. Ed. Lagerwey John, and Lü Pengzhi. Leiden:
Brill, 2010, 575.
Lin Wei-Cheng. Building a Sacred Mountain: The Buddhist Architecture of China’s
Mount Wutai. Seattle and London: University of Washington Press, 2014.
Loewe, Michael. “The Heritage Left to the Empires,” in The Cambridge History of
Ancient China: From the Origins of Civilizations to 221 B.C. Eds. Michael Loewe
and Edward L. Shaughnessy. Cambridge: Cambridge University Press, 1999,
976–1032.
Lyons, Islay and Harald Ingholt. Gandharan Art in Pakistan. New York: Pantheon
Books, 1957.
Mair, Victor. “Book Review.” Review of Written and Unwritten: A New History of
the Buddhist Caves at Yungang, by James O. Caswell. Harvard Journal of Asiatic
Studies 52 (June 1992): 345–61.
———. “Reply to James O. Caswell’s ‘Sinology or Art History,’ Pt II – A Response
to Victor H. Mair.” Early Medieval China 1 (1995): 147–49.
———. Tun-huang Popular Narratives. Cambridge: Cambridge University Press, 1983.
Martin, Emily and Arthur P. Wolf, eds. Religion and Ritual in Chinese Society. Stan-
ford, CA: Stanford University Press, 1974.
McNair, Amy. Donors of Longmen: Faith, Politics, and Patronage in Medieval Chi-
nese Buddhist Sculpture. Honolulu: University of Hawaii Press, 2007.
Mizunō, Seiichi. “Archaeological Survey of the Yun-kang Grottoes.” Archives of the
Chinese Art Society of America 4 (1950): 39–60.
Murray, Julia K. “Buddhism and Early Narrative Illustration in China.” Archives of
Asian Art 48 (1995): 17–31.
Nagahiro, Toshiō. The Representational Art of the Six Dynasties Period. Tokyo:
Seikosha Printing Co. and General Printing Co., 1969.
Pao-chen Chen. The Goddess of the Lo River: A Study of Early Chinese Narrative
Handscrolls. Ph.D. Dissertation, Princeton, 1987.
Reischauer, Ennin and Edwin O. Reischauer. Diary: The Record of a Pilgrimage to
China in Search of the Law. New York: Ronald Press Co., 1955.
Rosenfield, John M. The Dynastic Arts of the Kushans. Berkeley: Univeristy of Cali-
fornia Press, 1967.
Rowland, Benjamin. The Art and Architecture of India: Buddhist, Hindu, Jain. Lon-
don: Penguin, 1953.
Sato Chisui. “The Character of Yün-Kang Buddhism: A Look at the Emergence of a
State-Supported Religion in China under the Northern Wei.” The Memoirs of
Toyo Bunko 36 (1978): 38–83.
Schmid, David Neil. Yuanqi: Medieval Buddhist Narratives from Dunhuang. Ph.D.
Dissertation, Philadelphia: University of Pennsylvania, 2002.
Schopen, Gregory. Bones, Stones, and Buddhist Monks: Collected Papers on the
Archaeology, Epigraphy, and Texts of Monastic Buddhism in India. Honolulu:
University of Hawaii Press, 1997.
208 Selected bibliography
———. Buddhist Monks and Business Matters: Still More Papers on Monastic Bud-
dhism in India. Honolulu: University of Hawaii Press, 2004.
Sharf, Robert. “Art in the Dark: The Ritual Context of Buddhist Caves in Western
China,” in Art of Merit: Studies in Buddhist Art and Its Conservation. Ed. David
Park, Kuenga Wangmo, and Sharon Cather. London: Archetype Publications,
Courtauld Institute of Art, 2013, 38–65.
———. “Ritual,” in Critical Terms for the Study of Buddhism. Ed. Donald S. Lopez,
Jr. Chicago: University of Chicago Press, 2005, 245–69.
Sickman, Laurence and Alexander Soper. The Art and Architecture of China. New
York: Penguin, 1971.
Siren, Osvald. Chinese Sculptures from the Fifth to the Fourteenth Century. Reprint
of the 1925 edition by Ernest Benn, Ltd., London. New York: Hacker Art Books,
1970.
Soper, Alexander. “Imperial Cave-Chapels of the Northern Dynasties: Donors, Ben-
eficiaries, Dates.” Artibus Asiae 28 (1966): 241–70.
———. Literary Evidence for Early Buddhist Art in China. Ascona, Switzerland:
Artibus Asiae, 1959.
———. “Northern Liang and Northern Wei in Gansu.” Artibus Asiae 21 (1958):
145–6.
———. “South Chinese Influence on the Buddhist Art on the Six Dynasties Period.”
Bulletin of the Museum of Far Eastern Antiquities, Stockholm 32 (1960): 47–112.
Spink, Walter. Ajanta: History and Development, Volume Four. Leiden: Brill, 2009.
Stansbury-O’Donnell, Mark. Pictorial Narrative in Ancient Greek Art. Cambridge:
Cambridge University Press, 1999.
Steinhardt, Nancy Shatzman. Chinese Architecture in an Age of Turmoil, 200–600.
Honolulu: University of Hawaii Press, 2014.
———. Chinese Architecture: The Culture and Civilization of China. New Haven:
Yale University Press, 2002.
———. Chinese Imperial City Planning. Honolulu: University of Hawaii Press, 1990.
———. Chinese Traditional Architecture. New York City: China Institute in Amer-
ica, China House Gallery, 1984.
———. Liao architecture. Honolulu: University of Hawaii Press, 1997.
———. “Liao Archaeology: Tombs and Ideology along the Northern Frontier of
China.” Asian Perspectives 37, 2 (Fall 1998): 224–44.
Strong, John. “The Buddhist Avadanists and the Elder Upagupta,” in Tantric and
Taoist Studies in Honour of R.A. Stein, Mélanges Chinois et bouddhiques XXII.
Ed. Michael Strickmann. Bruxelles: Institut Belge des Hautes Etudes Chinoises,
1985, 862–81.
———. Relics of the Buddha. Princeton, NJ: Princeton University Press, 2004
Sullivan, Michael. The Arts of China. Berkeley: University of California Press, 2009.
Thomas, Edward. The Life of Buddha as Legend and History. London: Routledge
& Kegan Paul, 1949.
———. “Changing Patterns of Divinity and Reform in the Late Northern Wei.” The
Art Bulletin 84, 2 (1 June 2002): 222–45.
———. “Reconsidering Early Buddhist Cave-Making of the Northern Wei in Terms
of Artistic Interactions with Gansu and the Western Regions.” Journal of Inner
Asian Art and Archaeology 4 (2009): 101–18.
———.“The Sramana Superintendent Tanyao and his Time.” Trans. Galen Eugene
Sargent. Monumenta Serica, no. 16 (1957): 376–8.
 Selected bibliography 209
Vanderstappen, Harrie A. “Book Review.” Review of Written and Unwritten: A New
History of the Buddhist Caves at Yungang by James O. Caswell. Ars Orientalis 19
(1989): 125–7.
Wang, Eugene. “Grotto-Shrine as Chronotope and the Working of Analogous Iconog-
raphy: The Sixth Century Sculptureal Program in Cave 38 at Yungang in Perspec-
tive,” in Between Han and Tang: Religious Art and Archaeology in a Transformative
Period. Ed. Wu Hung. Beijing: Wenwu chubanshe, 2000, 279–312.
———. Shaping the Lotus Sutra: Buddhist Visual Culture in Medieval China. Seattle:
University of Washington Press, 2005.
Wang, Yi-t’ung. Trans. A Record of Buddhist Monasteries in Lo-yang. Princeton:
Princeton University Press, 1984.
Weiner, Sheila L. Ajaṇṭā: Its Place in Buddhist Art. Berkeley: University of California
Press, 1977.
Williams, Joanna. “Sarnath Gupta Steles of the Buddha’s Life.” Ars Orientalis,
no. 10 (1975): 171–92.
Wong, Dorothy C. Chinese Steles: Pre-Buddhist and Buddhist Use of a Symbolic
Form. Honolulu: University of Hawaii Press, 2004.
———. “Four Sichuan Buddhist Steles and the Beginnings of Pure Land Imagery in
China.” Archives of Asian Art 51 (1998/99): 56–79.
———. “Maitreya Buddha Statues at the University of Pennsylvania Museum.” Ori-
entations 32, 2 (2001): 24–31.
———. “Women as Buddhist Art Patrons During the Northern and Southern Dynas-
ties 386–581,” in Between Han and Tang: Religious Art and Archaeology in a
Transformative Period. Ed. Wu Hung. Beijing: Wenwu chubanshe, 2000, 535–66.
Wu Hung. “Art in Ritual Context: Rethinking Mawangdui.” Early China 17 (1992):
111–44.
———. “Born in Paradise: A Case Study of Dunhuang Sutra Painting and its Reli-
gious, Ritual and Artistic Context.” Orientations 23, 5 (May 1992): 54–60.
———. Monumentality in Chinese Art and Architecture. Stanford: Standford Uni-
versity press, 1995.
———. Reinventing the Past: Archaism and Antiquarianism in Chinese Art and
Visual Culture. Chicago: Center for the Art of East Asia, University of Chicago,
2010.
———. “The Transparent Stone: Inverted Vision and Binary Imagery in Medieval
Chinese Art.” Representations, no. 46 (Spring 1994): 58–85.
Yi, Lidu. “Art, Ritual and Patrons: An Unknown Rock-Cut Buddhist Cave Shrine in
Shanxi.” Eastern Buddhist 12 (2013): 1–25.
Zhao Chao. “Stone Inscriptions of the Wei, Jin Nanbeichao Periods (Trans. Victor
Xiong).” Early Medieval China 1 (1994): 84–96.
Zürcher Erik. The Buddhist Conquest of China: The Spread and Adaptation of Bud-
dhism in Early Medieval China. Leiden: E. J. Brill, Sinica Leidensia, 1972.
Zürcher Erik and Junjie Huang. Time and Space in Chinese Culture. Leiden: E. J.
Brill, 1995.

Works in Chinese and Japanese

Cao Chenming 曹臣明. “Pingcheng fujin xianbei ji beiwei muzang fenbu guilu kao
平城附近鮮卑及北魏墓葬分佈規律考 (Study on the distribution pattern of the
Xianbei and Northern Wei tombs near Pingcheng).” Wenwu, no. 5 (2016): 1–9.
210 Selected bibliography
———. “Yungang shiku de kaogu diaocha, kaogu fajue jiqi yiyi 雲崗石窟的考古調查,
考古發掘及其意義 (Archaeological excavations and investigations in Yungang and
their significance),” in 2005 nian Yungang guoji xueshu yantaohui lunwenji yanjiu-
juan. Ed. Yungang shiku yanjiuyuan. Beijing: Wenwu chubanshe, 2006, 153–6.
Chang Qing 常青. “Beichao shiku shenwang diaoke de jieshao 北朝石窟神王雕刻的
介紹 (Introduction to the spirit kings in the Northern dynasties caves).” Kaogu,
no. 12 (1994): 1127–41.
Chang Renxia 常任俠. “Yungang shike yishu 雲岡石刻藝術 (Yungang stone sculp-
ture),” in Fojiao yishu lunji 佛教藝術論集. Ed. Zhang Mantao 張曼濤. Taibei:
Dacheng wenhua chubanshe, 1980, 155–60.
Changzhoushi bowuguan 常州市博物館. “Changzhou nanjiao qijiacun huaxiang
zhuan mu 常州南郊戚家村畫像磚墓 (Bricks unearthed from Qijiacun tomb in
Southern suburb of Changzhou).” Wenwu, no. 3 (1979): 32–41.
Chen Bingying 陳炳應. “Beiwei Cao Tianhu Zao fangshita 北魏曹天護造方石塔
(Square stone stupa dedicated by Cao Tianhu in the Northern Wei).” Wenwu,
no. 3 (1988): 83–5.
Chengdushi wenwu kaogudui and Chengdushi wenwu kaogu yanjiusuo 成都市文物
考古隊與成都市文物考古研究所. “Chengdushi xi’anlu nanchao shike zaoxiang
qingli jianbao 成都市西安路南朝石刻造像清理簡報 (Report on the inventory of
the stone images of the Southern dynasties from the Xi’an road, Chengdu).”
Wenwu, no. 11 (1998): 4–20.
Chen Shuo 陳爍. “Dunhuang yishu zhong de ‘changdao’ yishi yu changdaowen zhi
guanxi tanwei 敦煌遺書中的唱導儀式與唱導文之關係探微 (The relationship
between the rituals and texts of Changdao in the manuscripts of Dunhuang).”
Gansu shehui kexue, no. 4 (2012): 244–7.
Chen Yuan 陳垣. Chen Yuan xueshu lunwen ji 陈垣学术论文集 (Collected essays of
Chen Yuan), 2 vols. Beijing: Zhonghua shuju, 1980.
———. “Ji Datong Wuzhoushan shikusi 記大同武州山石窟寺 (A record of the cave
temples on Mount Wuzhou in Datong),” in Chen Yuan. Chen Yuan xueshu lun-
wenji, vol. 1. Beijing: Zhonghua shuju, 1980, 398–409.
———. “Yungang shikusi zhi yijing yu Liu Xiaobiao 云岡石窟寺之譯經與劉孝標
(Sūtra translations in Yungang and Liu Xiaobiao),” in Chen Yuan xueshu lun-
wenji, vol. 1. Beijing: Zhonghua shuju, 1980, 443–448.
Chutuwenwu zhanlan gongzuozu 出土文物展覽工作組, ed. Wenhuadageming chutu
wenwu 文化大革命出土文物 (The cultural relics unearthed during the cultural
revolution). Beijing: Wenwu chubanshe, 1972, 81.
Datongshi bowuguan 大同市博物館. “Datong beiwei fangshan siyuansi yizhi fajue
baogao 大同北魏方山思遠寺遺址發掘報告.” Wenwu, no. 4 (2007): 4–28.
Ding Mingyi 丁明夷. “Guanyu Yungang shiku fenqi de jige wenti 關於雲崗石窟分期
的幾個問題 (Several problems on periodization of the Yungang caves),” in Yun-
gang bainian lunwen xuanji, vol. 1. Ed. Yungang shiku wenwu yanjiusuo. Beijing:
Wenwu chubanshe, 2005, 197–211.
———. “Yungang shiku de kaizao licheng 雲岡石窟的開鑿歷程 (The excavation of
theYungang caves),” in Yungang shiku bainian lunwen xuanji, vol. 2. Ed. Yungang
shiku wenwu yanjiusuo. Beijing: Wenwu chubanshe, 2005, 125–40.
———. “Yungang shiku yanjiu wushinian 雲崗石窟研究五十年 (Fifty years of
research on Yungang),” in Zhongguo shiku Yungang shiku. Ed. Yungang shiku
wenwu baoguansuo. Beijing and Tokyo: Wenwu chubanshe and Heibonsha, 1994,
174–86.
 Selected bibliography 211
Ding Mingyi and Ma Shichang. “Kezi’er shiku de fozhuan bihua 克孜爾石窟的佛傳
壁畫 (Mural paintings of the biography of the Buddha in Kizil caves),” in Zhong-
guo shiku Kezi’er shiku. Ed. Xinjiang weiwuer zizhiqu wenwu guanli weiyuanhui,
vol. 1 Beijing: Wenwu chubanshe, 1989, 185–222.
Du Doucheng 杜斗城. Beiliang fojiao yanjiu 北涼佛教研究 (Study on the Beiliang
Buddhism). Taibei: Xinwenfeng chuban gufenyouxian gongsi, 1998.
Feng Hanji 馮漢驥. “Chengdu wanfosi shike zaoxiang 成都万佛寺石刻造像 (Stone
sculpture from Wanfosi, Chengdu).” Wenwu cankao ziliao, no. 9 (1954):
110–20.
Fong Wen 方聞. “Chuan Gu Kaizhi ‘Nüshizhen tu’ yu zhongguo gudai yishushi 傳顧愷
之<女史箴圖> 與中國古代藝術史 (Admonitions of the instructress to the court ladies
attributed to Gu Kaizhi and Chinese art history).” Wenwu, no. 2 (2003): 82–96.
Fu Xinian 傅熹年. “Maijishan shiku zhong suo fanyin de beichao jianzhu 麥積山石
窟中所反映的北朝建築 (Architecture of the Northern dynasties reflected in the
Maijishan Cave Complex).” Wenwu ziliao congkan, no. 4 (1981): 156–83.
Gao Yun 高允. “Luyuan fu 鹿苑賦,” in Guang hongming ji. Comp. Daoxuan. T 52,
no. 339.
Guo Ruoxu 郭若虛. “Tuhua jianwen zhi 圖畫見聞志 (Experiences in painting),” in
Huashi congshu. Ed. Yu Anlan. Shanghai: Shanghai renmin meishu chubanshe,
1963, 1–10.
Hang Kan 杭侃. “Yungang di 20 ku xibi tanta de shijian yu Tan Yao wuku zuichu
de buju sheji 雲岡第20窟西壁坍塌的時間與曇曜五窟最初的佈局設計 (The col-
lapsing time of the west wall of cave 20 at Yungang and the initial designs and
compositions of the five Tanyao caves).” Wenwu, no.10 (1994): 55–63.
Han Youfu 韓有富. “Beiwei Cao Tiandu qianfota tacha 北魏曹天度千佛塔塔刹
(Thousand Buddha stupa dedicated by Cao Tiandu in the Northern Wei).” Wenwu,
no. 7 (1980): 65.
He Jianping 何劍平 and Zhou Xin 周欣. “Nanbenchao fojiao changdao de diben 南
北朝佛教唱導的底本 (Original texts for Buddhist Changdao in the Southern and
Northern Dynasties).” Xinan minzu daxue xuebao, 9 (2013): 64–70.
He Shizhe 賀世哲. “Guanyu Beichao shiku qianfo tuxiang zhu wenti 關於北朝石窟
千佛圖像諸問題 (Several problems of the iconography of the thousand Buddhas
during the Northern Wei dynasty).” Dunhuang yanjiu, no. 4 (1992): 1–18 and
no. 1 (1993): 1–10.
Henansheng wenhuaju Wenwu gongzuodui 河南省文化局文物工作隊. Dengxian
huaxiangzhuan mu 鄧縣彩色畫像磚墓 (Tomb with painted bricks at Dengxian).
Beijing: Wenwu chubanshe, 1958.
Hou Xudon 侯旭東. Beichao cunmin de shenghuo shijie: chaoting, zhouxian yu cunli
北朝村民的生活世界－朝廷州縣與村里 (Life of the Northern Dynasties villager-
Court, state, county and village). Beijing: Shang wu yin shu guan, 2005.
Hu Mi 胡謐 and Li Kan 李侃. Shanxi tongzhi 山西通誌 (General Gazetteer of Shanxi
Province), 17 vols. Jinan: Qilu chubanshe, 1997.
Ishimatsu Hinako 石松日奈子. Beiwei fojiao zaoxiang shi yanjiu 北魏佛教造像史研究
(Studies on Buddhist images in the Northern Wei). Beijing: Wenwu chubanshe,
2012.
———. Hokugi Bukkyo zozoshi no kenkyu 北魏仏敎造像史の硏究 (Buddhist
images in the Northern Wei). Kunitachi-shi: Buryukke, 2005.
Itō Chūta 伊東忠太. Itō Chūta kenchiku bunken 伊東忠太建築文献 (Itō Chūta’s
documents on architecture). Tōkyō: Ryūginsha, 1936–37.
212 Selected bibliography
———. Kengaku kiko 見学紀行 (Diary of things studied and observed). Itō Chūta
kenchiku bunken 伊東忠太建築文献, 5. Tōkyō: Ryūginsha, 1936.
———. “Shina sansei Unkō no sekkutsu-ji (The Cave-temples of Yungang in Shanxi,
China).” Kokka, nos. 197 and 198 (1906): 104.
———. Tōyō kenchiku no kenkyū 東洋建築の研究 (Research on East Asian archi-
tecture), Reprint of 1937 Tōyō kenchiku no kenkyū. Tōkyō: Ryūginsha, 1936–37.
Tōkyō: Hara Shobō, 1982.
Jiang Huaiying 姜懷英, Yuan Hairui 員海瑞 and Xie Tingfan 解廷凡. “Yungang shiku
xinfaxian de jichu jianzhu yizhi 雲岡石窟新發現的幾處建築遺址 (Several architec-
tural remains newly discovered in Yungang),” in Zhongguo shiku Yungang shiku. Ed.
Yungang shiku wenwu baoguansuo. Beijing: Wenwu chubanshe, 1991, 198–201.
Jiang Renhe 蔣人和. “Zaoqi foxiang huoyan shiwen shenguang zhi yanbian ji Guy-
ang dong qiyuan de yixie tanxiang 早期佛像火焰飾紋身光之演變及古陽洞起源的
一些探想 (Some thoughts on the origin of the Guyang cave and the evolution of
flame patterns in early Buddhist mandorlas),” in Longmen shiku yiqian wubai
zhounian guoji xueshu taolunhui lunwenji (Collected works from the 1500th year
anniversary international scholarly conference on the Longmen Caves). Ed. Long-
men shiku yanjiusuo. Beijing: Wenwu chubanshe, 1996, 220–51.
Jin Hua 晉華 and Qu Shengrong 瞿盛榮. “Shanxi xiyang Shimasi Shiku ji moya
zaoxiang 山西昔陽石馬寺石窟及摩崖造像 (Shimasi caves and cliff images in
Xiyang, Shanxi Province).” Wenwu, no. 4 (1999): 69–76.
Jin Shen 金申. Fojiao meishu congkao 佛教美術叢考 (Essays on Buddhist art). Bei-
jing: Kexue chubanshe, 2004.
———. Zhongguo lidai jinian foxiang tudian 中國歷年佛像圖典 (Buddhist images
from all times). Beijing: Wenwu chubanshe, 1994.
Jin Weinuo 金維諾. “Caojiayang yu Yang Zihua fengge 曹家樣與楊子華風格 (The
Cao school and the Yang Zihuan style).” Meishu yanjiu, no. 1 (1984), 39–40.
———. “Fobensheng tu xingshi de yanbian 佛本生圖形式的演變 (Evolution of the
forms of Buddhist biography narratives),” in Jin Weinuo, Zhongguo meishushi
lunji. Taibei: Mingwen shuju, 1984, 389–96.
Li Chongfeng 李崇峰. Fojiao kaogu: cong yindu dao zhongguo 佛教考古: 從印度到
中國 (Buddhist archaeology from India to China), 2 vols. Shanghai: Guji chuban-
she, 2014.
———. Zhongyin fojiao shikusi bijiao yanjiu 中印佛教石窟寺比較研究 (A compara-
tive study of Chinese and Indian Buddhist caves). Beijing: Beijing daxue chuban-
she, 2003.
Li Jingjie 李靜傑. “Guanyu Yungang dijiu, dishi ku de tuxiang goucheng 關於雲崗
第九第十窟的圖像構成 (Composition of the iconography of caves 9 and 10 in
Yungang),” in Yishushi yanjiu 藝術史研究. Ed. Zhongshan daxue yishushi yanjiu,
vol. 10. Guangzhou: Zhongshan daxue chubanshe, 2008, 327–59.
Li Ping 李凭. Beiwei Pingcheng shidai 北魏平城時代 (Pingcheng during the Northern
Wei). Beijing: Shehui kexue wenxian chubanshe, 2000.
Li Xueqin 李学勤. “Guanyu Yungang shiku xinbian kuhao de buchong shuoming
關於雲崗石窟新編窟號的補充說明 (Further explanations on the new cave num-
bers of Yungang caves).” Wenwu, no. 5 (2001): 87–9.
Li Yumin 李玉珉. “Jintasi shiku kao 金塔寺石窟考 (On the Jintasi cave temple).”
Gugong xueshu jikan, no. 22 (2004): 33–66.
Li Yuqun 李裕群. Beichao wanqi shikusi yanjiu 北朝晚期石窟寺研究 (A study of
Buddhist caves of the late Northern dynasties). Beijing: Wenwu chubanshe, 2003.
 Selected bibliography 213
———. “Shanxi beichao shiqi xiaoxing shiku de kaocha yu yanjiu 山西北朝小型石
窟的考察與研究 (Brief examination on the small caves of the Northern dynasties
in Shanxi),” in Between Han and Tang: Religious Art and Archaeology in a
Transformative Period 漢唐之間的宗教藝術與考古. Ed. Wu Hung. Beijing:
Wenwu chubanshe, 2000, 27–60.
———. “Shanxi Zuoquan shifosi shiku yu ‘Gao Huan Yundong’ shiku 山西左權石
佛寺石窟與高歡云洞 (Shifosi caves and Gao Huan caves in Zuoquan Shanxi Prov-
ince).” Wenwu, no. 9 (1995): 58–71.
———. “Sichuan nanchao zaoxiang de ticai yu beifang shiku de guanxi 四川南朝造
像的題材與北方石窟的關係 (The relationship between the subjects of the South-
ern dynasty images and the Northern cave temples),” in Sichuan chutu nanchao
fojiao zaoxiang. Ed. Sichuan bowuyuan, chengdushi Wenwu kaogu yanjiusuo, and
Sichuan daxue bowuguan. Beijing: Zhonghua shuju, 2013, 228–42.
———. “Tulufan tuyugou shiku kaogu xinfaxian: Shilun wushiji Gaochang fojiao tux-
iang 吐魯番吐峪溝石窟考古新發現試論五世紀高昌佛教圖像 (On Buddhist iconog-
raphy of the 5th century newly discovered in Gaochang in the archaeology of Tuyuk
cave temples),” in Yishushi zhong de Han Jin yu Tang, Song zhi bian. Ed. Shi Shouq-
ian and Yan Juanying. Taibei: Shitou Chuban gufen youxian gongsi, 2014, 95–124.
———.“Wutaishan nanchansi jiucang beiwei jingang baozuo shita 五臺山南禪寺舊
藏北魏金剛寶座石塔 (Northern Wei stupa with vajra pedestal formerly housed in
Nanchan temple on Wutai Mountain).” Wenwu, no. 4 (2008): 82–9.
Li Yuqun and Lidu Yi. “Shanxi Gaoping Shitanghui shiku 山西高平石堂會石窟 (The
Shitanghui Cave Complex in Gaoping Shanxi Province).” Wenwu, no. 5 (2009): 67–85.
———. “A study of the iconography of the great Buddha cliff in Gaoping Shanxi 山
西高平大佛山摩崖造像考.” Wenwu, no. 3 (2015): 82–94.
Li Yuqun, “Tulufan tuyugou shiku kaogu xinfaxian: Shilun wushiji Gaochang fojiao
tuxiang,” in Yishushi zhong de Han Jin yu Tang, Song zhi bian 藝術史中的漢晉與唐
宋之變 (Changes in Han and Jin, and Tang and Song in art history). Ed. Shi Shouqian
石守謙 and Yan Juanying 顏娟英. Taibei: Shitou chuban gufen youxian gongsi, 2014.
Li Zhiguo 李治国 and Liu Jianjun 刘建军. “Beiwei Pingcheng Luyeyuan shiku dia-
ocha ji 北魏平城鹿野苑石窟调查记,” in Yungang bainian lunwen xuanji. Ed. Yun-
gang shiku wenwu yanjiusuo. Beijing: Wenwu chubanshe, 2005, 150–6.
Liang Sicheng 梁思成. “Yungangshiku zhong suo biaoxian de Beiwei jianzhu 雲岡石
窟中所表現的北魏建築 (The Northern Wei architecture demonstrated in the Yun-
gang caves),” in Yungang bainian lunwen xuanji. Ed. Yungang shiku wenwu yan-
jiusuo. Beijing: Wenwu chubanshe, 2005, 11–28.
Liu Huida 劉慧達. “Bei Wei shiku yu Chan 北魏石窟與禪 (Northern Wei caves and
Chan).” Kaogu xuebao, no. 3 (1978): 337–50.
———. “Bei Wei shiku zhong de ‘sanfo’ 北魏石窟中的三佛 (Three Buddhas in the
Northern Wei caves).” Kaogu xuebao, no. 4 (1958): 91–101.
Liu Jianhua 刘建华. “Hebei zhangjiakou xiahuayuan shiku 河北張家口下花園石窟
(Caves from Xiahuayuan Zhangjiakou Hebei).” Wenwu, no. 7 (1998): 60–6.
———. Yixian wanfotang 义县万佛堂石窟 (Ten thousand Buddha hall caves in Yi
County). Beijing: Kexue chubanshe, 2001.
Liu Jinglong 劉景龍. “Longmen shiku kaizao niandai 龍門石窟開鑿年代 (Date of the
execution of the Longmen caves),” in Guyang dong. Ed. Liu Jinglong. Beijing:
Kexue chubanshe, 2001, 94–5.
Liu Shufen 劉淑芬. “Cong minzushi de jiaodu kan Taiwu miefo 從民族史的角度看
太武滅佛 (On Buddhist persecution by Taiwu from the perspective of the history
214 Selected bibliography
of ethnic minorities).” Zhongyang yanjiuyuan lishi yuyan yanjiusuo jikan, no. 72.1
(2001): 1–47.
———. Liuchao de chengshi yu shehui 六朝的城市與社會 (Cities and society during
the six dynasties). Taibei: Taiwan xuesheng shuju, 1992.
———. “Wu zhi liu shiji huabei xiangcun de fojiao xinyang 五至六世紀華北鄉村的
佛教信仰 (Buddhist belief in rural north China during the fifth and sixth centu-
ries).” Zhongyang yanjiuyuan lishi yuyan yanjiusuo jikan, no. 3 (1993):
497–544.
Liu Zhiyuan 劉志遠 and Liu Tingbi 劉廷壁. Chengdu wanfosi shike yishu 成都萬佛
寺石刻藝術 (Art of stone carving in Wanfosi Chengdu). Beijing: Gudian yishu
chubanshe, 1958.
Lu Yaodong 逯耀東. Cong Pingcheng dao Luoyang – Tuoba Wei wenhua zhuanbian
de licheng 從平城到洛陽 (From Pingcheng to Luoyang). Taibei: Lianjing chuban
shiye gongsi, 1979.
Lu Yifeng 陸屹峰 and Yuan Hairui 員海瑞. “Yungang shiku nisi kao 雲崗石窟尼寺考
(A study of the nunnery in the Yungang),” in Yungang bainian lunwen xuanji
雲崗百年論文選集. Ed. Yungang shiku wenwu yanjiusuo 雲岡石窟文物研究所.
Beijing: Wenwu chubanshe, 2005, 255–8.
Luo Shiping 羅世平. “Beiqi xinhuafeng 北齊新畫風 (New painting trend of the
Northern Qi).” Wenwu, no. 10 (2003): 63–5.
———. “Taiyuan beiqi Xu Xianxiu mu bihua zhong de huhua yinsu 太原北齊徐顯
秀墓壁畫中的胡化因素 (Barbarian elements in the murals of Xu Xianxiu tomb of
the Northern Qi in Taiyuan).” Yishushi yanjiu, no. 5 (2003): 223–41.
Luo Zongzhen 罗宗真. Liuchao kaogu 六朝考古 (Archaeology of six dynasties).
Nanjing: Nanjing daxue chubanshe, 1994.
———. “Nanjing xishanqian youfanfcun nanchao damu de fajue 南京西善橋油坊村
南朝大墓的發掘 (The excavation of the Xishanqiao tomb in Youfang village, Nan-
jing).” Kaogu, no. 6 (1963): 340–2.
Luoyang bowuguan 洛陽博物館. “Luoyang beiwei yuanshao mu 洛陽北魏元邵墓
(The Northern Wei Yuanshao tomb in Luoyang).” Kaogu, no. 4 (1973): 218–24.
Ma Changshou 馬長壽. Wuhuan yu Xianbei 烏桓與鮮卑 (Wuhuan and Xianbei).
Beijing: Zhonghua shuju, 1962, 30–3.
Ma Yuji 馬玉基. “Datongshi Xiaozhancun Huagetatai beiweimu qingli jianbao 大同
史小站村花麧塔台北魏墓清理簡報 (Inventory report on the Northern Wei tomb
unearthed in Huagetatai Xiaozhan village, Datong).” Wenwu, no. 11 (1983): 1–4.
Maeda, Masana 前田正名. Heijōno rekishi chirigakuteki kenkyū 平城の歴史地理学
的研 (Research on the history and geography of Pingcheng). Tokyo: Kazama
Shobō, 1979.
Maeda, Masana, Li Ping, Sun Yao and Sun Lei, eds. Pingcheng li shi di li xue yan jiu
平城历史地理学研究 (Research on Pingcheng history and geography). Beijing:
Shumu wenxian chubanshe, 1994.
Matsubara, Saburō 松原三郎. Chūgoku bukkyō chōkoku shiron 中国仏教彫刻史論
(On the history of Chinese Buddhist sculpture), 4 vols. Tokyo: Yoshikawa
Kōbunkan, 1995.
———. Chūgoku bukkyō chōkokushi kenkyū: tokuni kondōbutsu oyobi sekkutsu
zōzō igai no sekibutsu ni tsuite no ronkō 中國佛敎彫刻史研究 : 特に金銅仏及び
石窟造像以外の石仏に就いての論考 (Chinese Buddhist sculpture: A study based
on bronze and stone statues other than works from cave temples). Tokyo:
Yoshikawa Kōbunkan, 1966.
 Selected bibliography 215
Matsumoto Bunzaburō 松本文三郎. Gokuraku jōdoron 極樂淨土論 (On Pure Land
Paradise). Tokyo: Kinkōdō Shoseki Kabushiki Kaisha, 1904.
———. Miroku jōdoron 彌勒淨土論 (On Maitreya Pure Land). Tokyo: Heigo Shup-
pansha, 1911.
———. Shina Bukkyō ibutsu 支那佛教遺物 (Buddhist remains in China). Tokyo:
Daitōkaku, 1919.
Mi Wenping. “Xianbei shishi de faxian yu chubu yanjiu 鮮卑石室的發現與初步研究
(Discovery and preliminary research on the stone chamber of the Xianbei).”
Wenwu, no. 2 (1981): 1–7.
Mizunō Seiichi 水野清一. Chūgoku no Bukkyō bijutsu 中國の佛敎美術 (Buddhist art
of China). Tokyo: Heibonsha, 1968.
———. Chūgoku no chōkoko: sekibutsu, kondōbutsu 中國 の彫刻 (Chinese sculp-
ture). Tokyo: Nihon Keizai shinbunsha, 1960.
Mizunō Seiichi and Nagahirō Toshiō 長廣敏雄. Unkō sekkutsu 雲岡石窟: seireki
goseiki ni okeru Chūgoku hokubu Bukkyō kutsuin no kōkogakuteki chōsa hōkoku:
Tōhō Bunka Kenkyūjo chōsa (Yungang: The Buddhist Cave-Temples of the Fifth
Century A.D. in North China: detailed report of the archaeological survey carried
out by the Mission of the Tōhōbunka kenkyūsho 1938-45), 16 vols. Kyoto: Kyōto
Daigaku Jimbun Kagaku Kenkyūjo, 1951-56.
Nagahirō, Toshiō 長広敏雄. Chūgoku no sekkutsuji 中国の石窟寺 (Cave temples in
China). Sekai no bunka shiseki, 7. Tokyo: Kōdansha, 1969.
———. Daidō sekibutsu geijutsuron 大同石佛藝術論 (On art of stone Buddhas at
Datong). Kyōto: Kōtō Shoin, 1946.
———. Hiten no geijutsu 飛天の藝術 (Art of flying celestials). Kyōto Daigaku Jin-
bun Kagaku Kenkyūjo hōkoku. Osaka: Asahi Shinbunsha, 1949.
———. Unkō nikki: Taisenchū no Bukkyō sekkutsu chōsa 雲岡日記: 大戦中の仏教
石窟調查 (Diary of Yungang: Investigations of Buddhist stone caves during the
great war). Tokyo: Nihon Hoso Shuppan Kyōkai, 1988.
———. “Unko sekkutsu no nazo.” Bukkyo geijutsu, no. 134 (1981): 11–42.
———. Unkō to Ryūmon 雲岡と龍門 yungangshiku (Longmen and Yungang).
Tokyo: Chūō Kōron Bijutsu Shuppan, 1964.
———. “Yungang shiku chuzhongqi de teli dakan 雲崗石窟初,中期的特例大窟 (Big
caves in the first two phases in Yungang),” in Zhonguo shiku Yungang shiku. Ed.
Yungang shiku wenwu baoguansuo. Beijing: Wenwu chubanshe, 1994, 240–3.
———. “Yungang shiku di jiu shi shuangku de tezheng 雲崗石窟第九十雙窟的特徵 (The
features of the twin caves of 9 and 10),” in Yungang shiku wenwu baoguansuo, vol. 2.
Ed. Zhongguo shiku Yungang shiku. Beijing: Wenwu chubanshe, 1994, 193–207.
Nagahirō, Toshiō and Seiichi Mizuno. Unkō Sekkutsu: Chūgoku Bunka Shiseki 雲岡
石窟: 中国文化史蹟 (The Yungang cave temples: Historical landmarks of Chinese
culture). Tōkyō: Sekai Bunkasha, 1976.
Nanjing bowuyuan 南京博物院. “Jiangsu Danyang huqiao jianshan liangzuo nan-
chao muzang 江蘇丹陽胡橋建山兩座南朝墓葬 (Two tombs from Jiangsu Danyang
Huqiaojianshan of the Southern dynasties).” Wenwu, no. 2 (1980): 1–17.
———. “Jiangsu danyang huqiao nanchao damu ji zhuanke bihua 江蘇丹陽胡橋南
朝大墓及磚刻壁畫 (Tomb and brick murals from Danyang Huqiao in Jiangsu).”
Wenwu, no. 2 (1974): 44–56.
———. “Nanjing Raohuamen nanchao Liangmu fajue jianbao 南京蕘化門南朝梁墓
發掘簡報 (The excavation report on a Liang dynasty tomb in Raohuamen, Nan-
jing).” Wenwu, no. 12 (1981): 14–23.
216 Selected bibliography
———. “Nanjing xishanqiaomu jiqi zhuanke bihua 南京西善橋南朝墓及其磚刻壁畫
(Nanjing Xishanqiao tomb and brick murals from the Southern dynasties).”
Wenwu, no. 8 (1960): 37–42.
Ogawa Seiyō 小川晴暘. Daidō no sekibutsu 大同の石佛 (Stone Buddhas of Datong).
Arusu bunka sōsho, 14. Tokyo: Arusu, 1941.
———. Daidō Unkō no sekkutsu 大同雲岡の石窟 (The caves of Yungang, Datong).
Tokyo: Nikko Shoin, 1944.
Ogawa Seiyō 小川晴暘 and Torao Miyagawa 宮川寅雄. Unkō no sekkutsu 雲岡の石窟
(Yungang cave temples). Tokyo: Shinchōsha, 1978.
Ōmura Seigai 大村西崖. Shina bijutsushi chōso hen 支那美術史: 彫塑篇 (History of
Chinese art: Sculpture), 2 vols. Tokyo: Bussho Kankōkai Zuzōbu, 1915.
Onō Genmyō 小野玄妙. Daijō bukkyō geijutsushi no kenkyū 大乗仏敎芸術史の研究
(Research on the history of Mahayanna Buddhist art). Tokyo: Kanaobun’endō,
1944.
———. Kyokutō no sandai geijutsu 極東の三大藝術 (Three great arts of the Far
East). Tokyo: Hinoeuma Shuppansha, 1924.
———. Onō Genmyō Bukkyo geijutsu chosakushu 小野玄妙佛敎藝術著作集 (Onō
Genmyō’s collected W works on Buddhist art), 9 vols. Tokyo: Kaimei Shoin, 1977.
Ōtsuka Nobuo 大塚伸夫. “Shoki mikkyō no zentaizō: Shoki mikkyō no hōga kara
tenkai, kakuritsu he,” in Shoki mikkyō: shisō, shinkō, bunka. Eds. Takahashi
Hisao, Hideaki Kimura, Keiya Noguchi, and Nobuo Ōtsuka. Tokyo: Shunjusha,
2013, 11–13.
Qu Chuanfu 渠傳福. “Xu Xianxiu yu Beiqi Jinyang 徐顯秀墓與北齊晉陽 (Xu
Xianxiu tomb and Northern Qi Jinyang).” Wenwu, no. 10 (2003): 50–2.
Rong Xinjiang 榮新江. “Lüetan Xu Xianxiu mu bihua de pusa lianzhuwen 略談徐
顯秀 墓壁畫的菩薩連珠紋 (A brief discussion of the rings of pearls on the Bod-
hisattvas in the murals of Xu Xianxiu tomb).” Wenwu, no. 10 (2003): 66–8.
Sekino Tadashi 関野貞. Shina no kenchiku to geijutsu 支那の建築と藝術 (Art and
architecture of China), 3rd ed. Tokyo: Iwanami Shoten, 1941.
Shanxisheng Datongshi bowuguan 山西省大同市博物館. “Shanxi Datong Shijiazhai bei-
wei Simajinglong mu 山西大同石家寨北魏司馬金龍墓 (The Sima Jinlong tomb of the
Northern Wei in Shijiazhai village, Datong, Shanxi).” Wenwu, no. 3 (1972): 20–33.
Shanxisheng jianzhu baohu yanjiusuo 山西省建築保護研究所. “Shanxi pingding
kaihesi shiku 山西平定開河寺石窟 (Kaihe cave temples in Pingding Shanxi Prov-
ince).” Wenwu, no. 1 (1997): 73–85.
Shanxisheng kaogu yanjiusuo 山西省考古研究所. “Datongshi beiwei Song Shaozu mu
fajue jianbao 大同市北魏宋紹祖墓發掘簡報 (A brief report on the excavation of the
Song Shaozu tomb of the Northern Wei in Datong).” Wenwu, no. 7 (2001): 19–39.
———. “Taiyuan Beiqi Xu Xianxiu mu fajue jianbao 太原北齊徐顯秀墓發掘簡報 (A
brief report on the discovery of Xu Xianxiu tomb of the Northern Qi).” Wenwu,
no. 10 (2003): 4–40.
Shi Shuqing 史樹青. “Beiwei Cao Tiandu zao qianfo shita 北魏曹天度造千佛石塔
(Thousand Buddha stone stupa dedicated by Cao Tiandu in the Northern Wei).”
Wenwu, no. 1 (1980): 68–71.
Shisong Rinaizi. “Yungang di shiyi ku Taihe qinian yiyi zaoxiang he wuzhoushan
shikusi de bianhua 雲岡第11窟太和七年邑義造像和武周山石窟寺的變化 (The
image of the seventh year of Taihe in cave 11 and the changes of the Wuzhoushan
cave temples),” in 2005 nian Yungang guoji xueshu yantaohui lunwenji yanjiujuan.
Ed. Yungang shiku Yuanjiuyuan. Beijing: Wenwu chubanshe, 2012, 301–12.
 Selected bibliography 217
Su Bai 宿白. “Luoyang diqu beichao shiku de chubu kaocha 洛陽地區北朝石窟的初
步考察 (Preliminary research on the Northern dynasties caves in Luoyang area),”
in Zhongguo shiku Longmen shiku 中國石窟龍門石窟. Eds. Longmen wenwu bao-
guansuo and Beijing daxue kaoguxi. Beijing: Wenwu chubanshe, 1991, 225–38.
———. Zhang Yanyuan he Lidai minghua ji 張彥遠和《歷代名畫記》(Zhang
Yanyuan and famous paintings from all dynasties). Beijing: Wenwu chubanshe,
2008.
———. Zhongguo fojiao shikusi yiji 中國佛教石窟寺遗址 (Ruins of Buddhist cave
temples in China). Beijing: Wenwu chubanshe, 2010.
———. Zhongguo shikusi yanjiu 中國石窟寺研究 (Study of Chinese Buddhist cave
temples). Beijing: Wenwu chubanshe, 1996.
Sun Ji 孫機. Handai wuzhi wenhua ziliao tushuo 漢代物質文化資料圖説 (Illustra-
tions of material culture of the Han dynasty). Beijing: Wenwu chuban she,
1990.
———. Sun Ji tan wenwu 孫機談文物 (Su Ji on cultural relics). Taibeishi: Dongda,
2005.
———. Yang guan ji 仰觀集 (Collected essays on appreciation and authentication of
ancient Chinese cultural relics). Beijing: Wenwu chubanshe, 2015.
Sun Tongxun 孫同勛. Tuobashi de hanhua 拓跋氏漢化 (Sinicization of the Tuoba).
Taibei: Guoli Taiwan daxue wenxueyuan, 1962.
Tang Changru 唐長儒. Weijin nanbei chao shi luncong 魏晉南北朝史論叢 (Discus-
sions on History of Wei, Jin and Northern and Southern dynasties). Beijing: San-
lian shudian, 1962.
———. Weijin nanbei chao shi shiyi 魏晉南北朝史論拾遺 (Supplement history of
Northern Wei, Jin and Northern and Southern dynasties). Beijing: Zhonghua
shuju, 1983.
Tang Yongtong 湯用彤. Hanwei Liangjin nanbei chao fojiao shi 漢魏兩晉南北朝佛敎
史 (History of Buddhism from the Two Jins and the Southern and Northern dynas-
ties). Beijing: Beijing daxue chubanshe, 1997.
Tian Yuqing 田餘慶. Qin Han Wei Jin shi tanwei 秦漢魏晉史探微 (Study of the his-
tory of Qin, Han, Wei, and Jin dynasties). Beijing: Zhonghua shuju, 1993.
Tokiwa Daijo 常盤大定. Shina Bukkyō no kenkyū (Studies on Buddhism in China).
Tokyo: Shunjūsha, 1938.
———. Shina bukkyō shiseki tōsaki (Expeditions to Chinese Buddhist sites). Tokyo:
Kokusho Kankōkai, 1937.
———. Shina Bukkyō shiseki tōsaki 支那佛敎史蹟踏查記 (Investigation of the mon-
uments of Chinese Buddhism). Tokyo: Ryūginsha, 1938.
Tokiwa Daijo 常盤大定 and Tadashi Sekino 関野貞. Buddhist monuments in
China:Text part (English version of Shina Bukkyō shiseki hyōkai). Tokyo: Bukkyo-
shiseki Kenkyu-kwai, 1926.
———. Chūgoku bunka shiseki: kaisetsu 支那文化史蹟: 解說 (Monuments of Chi-
nese culture: commentary). Kyōto: Hōzōkan, 1975.
———. Shina Bukkyo shiseki 支那佛教史蹟 (Monuments of Chinese Buddhism), 5
vols. Tokyo: Bukkyo Shiseki Kenkyukai, 1925–27.
———. Shina bunka shiseki: zuhan 支那文化史蹟: 圖版 (Monuments of Chinese
culture: plates). Tokyo: Hōzōkan, 1939–41.
Tsukamoto Yasushi 塚本靖, Itō Chuta 伊東忠太 and Tadashi Sekino 關野貞. Sekai
kenchiku shūsei, Shina kenchiku 世界建築集成, 支那建築 (Compendium of world
architecture: Chinese architecture). Tokyo: Kenchiku Gakkai, 1928.
218 Selected bibliography
Tsukamoto Zenryu 塚本善隆. Chūgoku Bukkyō tsūshi 中国仏教通史 (Comprehen-
sive history of Chinese Buddhism). Tokyo: Suzuki Gakujutsu Zaidan, 1968.
———. Daisekibutsu 大石仏 (Great stone statues of Buddhism). Tōkyō: Kōbundō,
1953.
———. Shina bukkyōshi kenkyū 支那佛敎史硏究: 北魏篇 (Research on the history
of Chinese Buddhism, Northern Wei chapter). Tōkyō: Shimizu Kōbundō, 1969.
———. Hokuchō bukkyōshi kenkyu 北朝仏敎史研究 (Research on the history of
Northern dynasties Buddhism). Tsukamoto zenryū chosakushū, 2. Tokyo:
Daitōshuppansha, 1974.
———.“The Sramana Superintendent Tanyao and his Time.” Trans. Galen Eugene
Sargent. Monumenta Serica, no. 16 (1957): 376–8.
———. Tsukamoto Zenryū chosakushu 塚本善隆著作集 (The works of Zenryu Tsu-
kamoto). Tokyo: Daitō Shuppan, 1974.
Wan Shengnan 万繩楠. Wei-jin nanbeichao shi lungao 魏晉南北朝史論稿 (Essays on
the history of Wei-Jin and Southern and Northern dynasties). Heifei: Anhui jiaoyu
chubanshe, 1983.
———. Wei-jin nanbeichao wenhua shi 魏晉南北朝文化史 (Cultural history of Wei-
Jin and Southern and Northern dynasties). Huangshan: Huangshan shushe, 1989.
Wang Huimin 王惠民. “Zhique waidao fei posouxian bian 執雀外道非婆薮仙辯
(Identifying the heretic figure who has holding a bird).” Dunhuang yanjiu, no. 1,
119 (2010): 1–7.
Wang Xun 王遜. “Yungang yidai kancha ji 雲崗一帶勘察記 (Records on investiga-
tion in the Yungang area),” in Yanbei wenwu kanchatuan baogao. Ed. Yanbei
wenwu kanchatuan. vol. 2. Beijing: Zhongyang renmin zhengfu wenhuabu wen-
wuju, no. 2, 1951, 28.
Wang Yintian 王银田 and Liu Junxi 刘俊喜. “Datong zhijiabu beiwei mu shidun
bihua 大同智家堡北魏墓石椁壁画 (Stone bed murals from the Northern Wei tomb
at Zhijiabu in Datong).” Wenwu, no. 7 (2001): 40–51.
Wang Zhongluo 王仲犖. Wei jin nanbeichao shi 魏晉南北朝史 (History of Wei-jin
and Southern and Northern dynasties). Shanghai: Renmin chubanshe, 1976.
Wei Zhengzhong 魏正中. Quduan yu zuhe – qiuzi shiku siyuan yizhi de kaoguxue
tansuo 區段與組合－龜茲石窟寺院遺址的考古學探索區段與組合 (Districts and
groups: An archaeological investigation of the rock monasteries of Kucha). Shang-
hai: guji chubanshe, 2013.
Wu Hung, ed. Han Tang zhijian de zongjiao yishu yu kaogu 漢唐之間的宗教藝術與
考古 (Between Han and Tang: Religious art and archaeology in a transformative
period), 3 vols. Beijing: Wenwu chubanshe, 2000.
Xiang Da 向達. “Tangdai sujiang kao 唐代俗講考 (Study of vernacular oral presenta-
tion in the Tang dynasty),” in Xiang Da, Tangdai Changan yu xiyu wenming 唐代
長安與西域文明. Beijing: Sanlian shudian, 1957, 295–336.
Xiangfanshi Wenwu guanlichu 襄樊市文物管理處. “Xiangyang jiajiachong huax-
iangzhuanm 襄陽賈家沖畫像磚墓 (Tomb with brick carvings from Jiajiachong,
Xiangfan).” Jianghan kaogu, no. 1 (1986): 16–33.
Xinjiang weiwuer zizhiqu wenwu guanli weiyuanhui, baichengxian kezier qian-
fodong wenwu baoguansuo and Beijing daxue kaoguxi, eds. Zhongguo shiku
kezi’er shiku 中國石窟, 克孜爾石窟 (Chinese caves: Kizil), 3 vols. Beijing and
Tokyo: Wenwu chubanshe and Heibonsha, 1997.
Xu Pingfang 徐蘋芳. “Zhongguo shikusi kaoguxue de chuangjian licheng – Du Su
Bai’s Zhongguo shikusi yanjiu 中國石窟寺考古學的創建歷程讀宿白的中國石窟寺
 Selected bibliography 219
研究 (The process of establishing the archaeological methodology in the study of
caves in China—reading Su Bai’s Studies on Chinese cave temples),” Wenwu, no. 2
(1998): 54–63.
Yagi Haruo 八木春生. Chūgoku Bukkyō bijutsu to Kan minzokuka: Hokugi jidai
kōki o chūshin to shite 中国仏敎美術と漢民族化: 北魏時代後期を中心として
(Sinicization in Chinese Buddhist art). Kyoto: Hōzōkan, 2004.
———. Unko sekkutsu mon’yoron 雲岡石窟文様論 (On the ornament of Yungang
caves). Kyoto: Hōzōkan, 2000.
Yan Wenru 阎文儒. “Yungang shiku de kaichuang he ticai de fenxi 雲崗石窟的開創
和題材的分析 (Analysis of the excavation and subjects of the Yungang cave tem-
ples).” Shehui kexue jikan, no. 5 (1980), 112–18, and no. 6 (1980), 110–14.
———. “Yungang shiku zaoxiang zhong de yixie ticai de kaoshi 雲崗石窟造像中的
一些題材的考釋 (Study of some subjects in the image of the Yungang cave tem-
ples).” Xiandai foxue, no. 2 (1963), 11–21.
———. Yungang shiku yanjiu 云岡石窟研究 (The study of Yungang caves). Guilin:
Guangxi shifan daxue chubanshe, 2003.
———. Zhongguo shiku yishu zonglun 中國石窟藝術總論 (General introduction to
Chinese grottoes). Tianjin: Guji chubanshi, 1987.
Yang Hong 楊泓. “Guanyu nanbeichao qingzhou kaogu de sikao 關於南北朝青州考
古的思考 (Reflections on the archaeology in Qingzhou of the Southern and North-
ern dynasties).” Wenwu, no. 2 (1998): 2.
———. “Guonei xiancun zuigu de jizun fojiao zaoxiang shiwu 國内現存最古的幾尊佛
教造像實物 (Earliest extant Buddhist images).” Xiandai foxue 現代佛學, no. 4
(1962): 31–4.
———. Han Tang meishu kaogu he fojiao yishu 漢唐美術考古和佛教藝術 (Art
archaeology and Buddhist art between the Han and Tang). Beijing: Kexue chuban-
she, 2000.
———. Meishu kaogu banshiji 美術考古半世紀中國美術考古發現史 (Half century
of archaeological discoveries in Chinese arts). Beijing: renmin meishu chubanshe,
2015.
———. “Shilun Nanbeichao qianqi foxiang fushi de zhuyao bianhua 試論南北朝期間
佛像 服飾的主要變化 (A preliminary study on the evolution of draperies of Buddhist
images from the Southern and Northern dynasties).” Kaogu, no. 6 (1963): 330–7.
Yi, Lidu. “Handan Cheng’anxian chutu de beiwei Taihe liunian shijia sanzunxiang
邯鄲成安縣 出土的北魏太和六年釋迦三尊像 (Diciphering a newly discovered
image in Cheng’an).” Dunhuang yanjiu, no. 3 (2012): 7–14.
Yin Guanming 殷光明. “Beiliang Shita Fenqi Shilun 北梁石塔分期 試論 (On the
periodization of the Northern Liang stone pagodas).” Dunhuang yanjiu, no. 3
(1997): 84–92.
———. “Shilun mofa sixiang yu Beiliang fojiao jiqi yingxiang 試論末法思想與北涼
佛教及其影響 (On mofa belief and Northern Liang Buddhism as well as its influ-
ence).” Dunhuang yanjiu, no. 2 (1998): 89–102.
Yoshimura Rei 吉村怜. Chūgoku Bukkyō zuzō no kenkyū (Study of Chinese Buddhist
art). Tokyo: Tōhō Shoten, 1983.
———. Chūgoku Bukkyō zuzō no kenkyū 中國仏敎図像の研究 (A study on the
iconography of Buddhist art in China). Tokyo: Tōhō Shoten, 1983.
———. “Donyō Gokutsu zōei shidai.” Bukkyō geijutsu, no. 212 (1994): 11–36.
———. Tennin tanjōzu no kenkyū: Higashi Ajia Bukkyō bijutsushi ronshū 天人誕生
図の研究東アジア仏教美術史論集. Tokyo: Tōhōshoten, 1999.
220 Selected bibliography
———. Tennin tanjōzu no kenkyū: Higashi Ajia Bukkyō bijutsushi ronshū (Study of
the celestial-born iconography: Selected essays on East Asian Buddhist art). Tokyo:
Tōhōshoten, 1999.
Yungang guji baoyangsuo 雲岡古蹟保養所. “Yungang xinfaxian de yikuai beiwei
shike 雲岡新發現的一塊北魏石刻 (A Northern Wei stone inscription just discov-
ered in Yungang).” Wenwu caikao ziliao, no. 9 (1957): 48.
“Yungang kuqian yizhi fajue huo zhongda chengguo 雲崗窟前遺址發掘獲重大成果
(The great result of the excavation of the ruins in front of the caves).” Zhongguo
wenwubao (1994): 1, 16.
Yungang shiku baoguansuo 雲岡石窟保管所 and Wenwu baoguan kexue jishu yan-
jiusuo 文物保管科學技術研究所. “Yungang shiku jianzhu yiji de xinfaxian 雲岡石
窟建築遺址的新發現 (New findings of architectural remains in Yungang),”
Wenwu, no. 4 (1976): 89–93.
Yungang shiku wenwu baoguansuo, ed. Zhongguo shiku yungang shiku 中國石窟云
岡石窟 (Chinese caves: Yungang), 2 vols. Beijing: Wenwu chubanshe, 1994.
Yungang shiku wenwu yanjiusuo 雲岡石窟文物研究所, Shanxisheng kaogu yanji-
usuo 山西省考古研究所 and Datongshi bowuguan大同市博物館. “Yungang shiku
disanku yizhi fajue jianbao 雲崗石窟第三窟遺址發掘簡報 (Report on the excava-
tions in front of cave 3 in Yungang),” Wenwu, no. 6 (2004): 65–88.
Yungang shiku wenwu yanjiusuo, ed. Yungang bainian lunwen xuanji 雲崗百年論文
選集 (Selected essays on Yungang in the last hundred years). Beijing: Wenwu
chubanshe, 2005.
Zhang Qingjie 張慶捷, Li Baijun and Jiang Weiwei. “Shanxi Yungang shiku kuding
beiwei fosi yizhi 山西雲崗石窟窟頂北魏佛寺遺址 (Northern Wei monastery ruins
discovered on top of the cave-temples in Yungang of Shan’xi),” in 2010 Zhongguo
zhongyao kaogu faxian. Ed. guojia wenwuju. Beijing: Wenwu chubanshe, 2011,
127–30.
Zhang Qingjie, “Yungang shiku kuding xiqu beiwei fojiao siyuan yizhi 雲岡石窟窟
頂西區北魏佛教寺院遺址 (The Ruins of Buddhist Temples in the Western Section
above the Yungang Caves),” in Kaogu xuebao, no. 4 (2016), 533–62.
_______. “Shanxi Datong caochangcheng beiwei erhao yizhi fajue jianbao 山西大同
操場城北魏二號遺址發掘簡報 (Important excavation on no. 2 ruins of the North-
ern Wei in Caochangcheng Datong Shanxi).” Wenwu, no. 4 (2016): 4–25.
Zhang Zhuo. “‘Dajin xijing wuzhoushan chongxiu dashikusibei’xiaoyi 大金西京武周
山重修大石窟寺碑小譯 (On the Jin stele inscription on the repairs of the great stone
caves on Mount Wuzhou of the Jin dynasty),” in Yungang bainian lunwen ji. Ed.
Yungang shiku wenwu yanjiusuo. Beijing: Wenwu chubanshe, 2005, 350–4.
———. Yungang shiku biannianshi 雲崗石窟編年史 (The chronicle of the Yungang
cave temples). Beijing: Wenwu chubanshe, 2006.
Zhao Chao 趙超. Hanwei nanbei chao muzhi huibian 漢魏南北朝墓誌彙編 (Collec-
tions of tomb inscriptions from the Han, Wei, and the Southern and Northern
dynasties). Tianjin: guji chubanshe, 1996.
Zhao Kunyu 趙昆雨. “Yungang shiku yingzao de jige wenti 雲岡石窟營造的幾個問題
(Issues on the excavations of Yungang cave temples),” in 2005 nian Yungang guoji
xueshu yantaohui lunwenji yanjiujuan. Ed. Yungang shiku Yuanjiuyuan. Beijing:
Wenwu chubanshe, 2012, 313–23.
Zhao Shuguang 趙曙光. “Longwanggou xice gudai yizhi qingli jianbao 龍王廟溝西
側古代遺址清理簡報 (Report on the inventory of the ancient ruins to the west of
the Dragon King temple).” in Zhonguo shiku Yungang shiku, 2 vols. Ed. Yungang
 Selected bibliography 221
shiku wenwu baoguansuo. Beijing and Tokyo: Wenwu chubanshe and Heibonsha,
1994, 219–30.
Zhao Wanli 趙萬里. Hanwei nanbeichao muzhi jishi 漢魏南北朝墓誌集釋 (Anno-
tated collections of tomb inscriptions from the Han, Wei and the Southern and
Northern dynasties). Beijing: Kuxue chubanshe, 1956.
Zhou Yiliang 周一良. Weijin nanbeichao shilunji 魏晋南北朝史论集 (Collected
essays on the history of Wei-Jin and Northern and Southern dynasties). Beijing:
Beijing daxue chubanshe, 1987.
———. Weijin nanbeichao shilunji xubian 魏晉南北朝史論集續編 (Continued essays
on the history of Wei-Jin and Southern and Northern dynasties). Beijing: Beijing
daxue chubanshe, 1991.
Zhou Zheng 周錚. “Xiwei Ju Shiguang zaoxiangbei 西魏巨始光造像碑 (A Western
Wei stele by Ju Shiguang).” Zhongguo lishi bowuguan guankan, no. 11 (1985):
90–4.
Index

Page numbers for entries dealing with specific cave numbers are prefixed by p or pp
to avoid confusion between cave numbers and page numbers.

Abe, Stanley K. 18n64, 117, 128n9, 33, 57, 64; in front of caves 9 and
129n31, 131n68, 131n81, 131n83, 10 pp28–9, 31, 50n49, 83, 105n14;
131n85 Jiaoshansi caves 50n46; Lubanyao
Abu Simbel Temple 20 caves 27; by Mizuno and Nagahiro
accumulation of merit and virtue 5, 27–31, 49n40, 94; monastery
126; Asoka 153; cave function remains above caves 5 and 6 pp28,
36, 186n54; imperial family 21–2; 33–4, 189; Sima Jinlong 40; tiles
Trapusa and Bhallika 113; see also 28–30, 33, 34–6, 50n43, 57; Yanbei
merit shiyuan 40
Achaemenid kings 20 archaeological expeditions 2, 4–6, 8,
Admonitions to the Court Ladies see 14n30, 15n32
Nüshizhen tu archaeological materials 5
Amitabha Sūtra 120 archaeological method 4, 8
Ananda 70, 77n91, 143, 152–3, archaeological periodization 170
186n48 archaeological remains: in front of
Anathapindika’s park 152 Wuhuadong caves 62; Northern Wei
An Faqin 77n92, 195 monasteries 62
Antarvasaka 52n79, 195 archaeological typology methodology 4
anterooms: cave 3 p36; cave 12 pp36, archery competition 143, 149–50
80, 81, 105n11, 112–16, 130n63, architectural space of Buddhist
140, 168; cave 31 pp158, 163; caves sanctuaries 43, 73, 111, 119
7 and 8 pp36, 90, 105n19, 107n46; architectural structure of Buddhist
caves 9 and 10 pp90, 93, 94, 96, sanctuaries 3, 4, 6, 8, 10, 12, 27, 29,
123; passageway connecting 94; 31, 38–40, 43, 48, 56, 58, 63, 73,
residential cells with 37; second- 100, 106n34, 111, 118, 126, 132–40,
phase caves 39, 44–5 158, 190; functions of 39, 43
An Xi 77n92, 195 architecture 4, 6, 9, 20, 38–40, 43, 48,
aranya 23 56, 104, 110, 175, 176, 183n7
archaeological excavations 12, 26–37; archways: caves 7 and 8 p90; caves 9
in 1970s and 1990s 31–5; above the and 10 p93
Yungang caves 7, 10–11, 23, 26, 27, Asita’s prophecy of 147, 148
28, 29–31, 33, 36–7, 49n43, 65, 73, Asoka, King 71, 77n92, 152–3; jataka
104, 110, 189–90; in and around narrative 115, 124, 151–3, 158, 169,
Yungang 10, 48, 110, 189; in front 173, 185n39; merit accumulation 153
of cave 3 p32; in front of cave 8 p28; Asura 80, 89
in front of cave 12 pp29, 105n11; in aupapadukas (lotus-born Buddhas) 58,
front of cave 20 pp14n23, 26, 29, 139
224 Index
avadana narratives 117, 129n30, broad knowledge (bo) 117
129n40 Buddha: alms bowl of 38, 112, 127n7,
Ayuwang jing 77n92, 195 152–3; Amitayus 174; archery
Ayuwang zhuan (The Biography of competition 143, 149–50; ascent
King Asoka) 70–1, 77n92, 195 to and descent from Tusita Heaven
40, 94, 97, 141, 143, 146, 171;
Bai Zheng see Grand Chief Autumnal ascent to Tavatimsa to preach for
Prince Bai Zheng his mother Mayadevi 143, 150, 169,
baizizhang (tent of a hundred sons) 183n11; athletic training 146–7,
73n12, 195 149–50; birth of 101, 115, 143,
Bamiyan 100 146–8, 185n33; chariot-driving
Banishing of the Demons see Mara, 145; conception 147; contemplative
Assault of prince 149; Dharma 1, 21, 61, 64,
Baochang 131n84, 195 69–70, 72, 76n86, 94, 120, 126,
Baoshan tomb 2 144, 195 143, 151, 171–2; Dipamkara 112,
baoyi bodai robes 43, 47, 59, 63, 103, 115, 124, 151, 169, 173, 185n39;
155, 173–4, 179, 182; cave 11-14 elephant-riding 95, 130n63, 132,
pp43, 179 143, 145–7, 149, 169, 184n21–2,
Barthes, Roland 114 184n27; farewell to His White
bas-relief 127n7, 143, 147–8, 168, Horse Kanthaka 143, 145, 148,
183n8 184n22, 185n31; first meditation
Beishi 40, 85 143, 149, 168, 185n34; historical
Beitai caves 1, 37, 51n55, 62, 69, 122, biographies 46, 115, 185n34;
172, 195; see also Yungang horsemanship 145; law of 65, 117,
Benares 149 124; life stories 115, 143; Maitreya
Bhallika 3, 111, 113–14, 121, 128n12; 39, 45, 71, 89, 94, 97, 108n68–9,
see also Trapusa and Bhallika 112, 124, 128n9, 141, 143, 166,
Bhikkhunis (nuns) 26 171–2, 174, 194; manifestation of
bhiksus 21, 80, 113, 149 45, 173; Many-Jeweled 47; meeting
bian (talent) 117, 195 his son Rahula 95, 130n63, 151,
bilateral symmetry 66 153, 169, 185n39; Prabhutaratna
Binglingsi caves 44, 175; cave 169 p175 45, 70, 94, 120, 134–5, 140–5,
bingwu day 87, 195 153, 171–2, 183n5, 193; previous
Binyang caves (Bingyangdong) 10, 43, incarnations 45; return to his
66, 173, 195 hometown Kapilavastu 151;
Biographies of Eminent Monks see Śākyamuni 45, 47, 53, 64, 70, 71,
Gaoseng zhuan 88, 94, 108n68–9, 120, 134–5,
Biography of King Asoka see Ayuwang 140–3, 152–3, 166, 171–2, 183n5,
zhuan 193–4; worshiping 48
birang (making room for another space) Buddha-Daoist synthesis 121
156, 195 Buddha Dharma 1, 21, 61, 64, 69,
birth of Buddha 101, 115, 143, 146–8, 72, 76n86, 94, 120, 126, 143, 151,
185n33 171–2
black pebbles on a stone Buddha 22, 53 Buddha Hall image caves (fodianku)
bo (knowledge) 117, 195 81, 133
Boli see Trapusa and Bhallika Buddha Halls (fodian) cave temples pp1,
Book of Song see Songshu 18n68, 37, 81, 106n34, 111
boshanlu 114, 195 Buddha-household economic system 71
brackets 44, 105n7, 139 Buddhanandi 38
Brahmana 46, 69, 100, 112, 127n1 Buddha of transformations (huasheng)
Brahmins 148, 151 120
Brahmin youth 152 Buddhism: five precepts of 121–2,
Brief History of The Clergy of The 126; Liangzhou 38–9, 124; Noble
Great Song see Da Song sengshi lüe Truths 118; persecution of 2, 21–2,
 Index 225
25, 54, 60, 67, 70, 172; Plebeian cave 7 pp90–1, 94, 105n19, 107n46,
121–2, 129n27; reliquaries 21, 60; 168; see also caves 7 and 8
restoration of 2, 7, 11, 22, 37, 40, cave 8 pp28, 83, 90–1, 94; excavation
53, 59–61, 69, 72, 74n23, 76n86, in front of 28; see also caves 7 and 8
172; sinicization 18n64, 173–82; cave 9 pp28, 44, 45–6, 83, 93–5, 98,
texts 21, 69, 122 105n19, 108n69–70, 120, 127n4,
Buddhist archaeology 4, 43 130n63, 158; excavation in front
Buddhist architecture 4, 6, 9–10, of 28; initial function 116; Rahula
73, 175, 190; caves 10, 31, 176; jataka 153; Syama jataka 115; see
Yungang 3, 6, 10, 12, 38–9, 41, 43, also caves 9 and 10
48, 56, 104, 110, 126, 190 cave 10 pp28–9, 44, 83, 93–5, 98,
Buddhist art 4–6, 8–10, 12, 60, 73, 108n69–70, 130n63; excavation in
112, 170; sinicization 39, 43, 144, front of 28–9, 31; see also caves 9
173–4, 182 and 10
Buddhist centers 119, 186n54 cave 11 pp63, 81–2, 83, 97, 103,
Buddhist community 4, 116, 119 105n14, 107n56, 111, 156–7;
Buddhist doctrines 119, 151 abandonment of 85; dating 11, 88,
Buddhist iconography 151, 175–6, 190 90, 97, 99, 104, 189; inscriptions
Buddhist image-making 22, 38, 53, 72, 16n50, 27, 81, 83, 84, 123, 131n76;
120, 122, 174 see also caves 11 to 13
Buddhist narratives see narratives cave 11-14 pp14n20, 43, 63, 156–7,
Buddhist sūtras 6–7, 36–7, 46–7, 158, 174; baoyi bodai 43, 179;
52n89, 61–2, 69–70, 76n74, 77n92, inscription 14n20, 174
106n40, 118–19, 172–3, 186n43 cave 12 pp9, 11–12, 32, 80, 81–2,
Burning of the Fire Dragon’s Shrine 115 90, 104, 105n11, 107n55, 111–18,
121–6, 163; anteroom 37, 80, 82,
cai (talent) 117, 195 105n11, 112–16, 130n63, 140,
caitya (stūpa) caves: cave 6 pp45, 102, 168; archaeological excavation in
103; cave 11 pp45, 80, 81; cave 39 front of 29, 105n11; configuration
pp133, 165–6, 172; caves 1 and 2 of images 90; dating 11, 89–90, 90,
pp45, 101 103, 107n44; function 37, 118, 122,
canonical scriptures 117, 120; see also 124–6, 189–90; insertion between
Buddhism, texts caves 13 and 11 p89; musical deities
Cao style 179 80, 89, 111, 112, 118, 129n41;
Cao Yan 7, 16n43–4, 37; see also Jin patronage 9, 102, 107n45, 189;
stele inscription square bricks 29; see also caves 11
caoyi chushui style 74n17, 195 to 13
Cao Zhongda 74n17, 176, 179, 181 cave 13 pp11, 44, 80, 81, 82–4;
Caswell, James 9, 105n2, 108n63 colossal image 80, 81, 84, 88;
cave 1 pp101–3; see also caves 1 and 2 elliptical cave plan 80, 84, 88;
cave 2 pp40n43, 102–3, 108n65; intertwined dragons 47, 80, 81, 84;
see also caves 1 and 2 patronage 11, 80, 87; see also caves
cave 3 pp6–8, 32, 35–6, 78, 83, 85, 11 to 13
103; excavations above 29–31, 35; cave 16 pp54, 56, 58, 63–6, 74n43, 87;
excavations in front of 32 worshippers legs depicted 54, 87; see
cave 4 pp132, 170 also caves 16 to 20
cave 5 pp9, 15n32, 100–4, 108n68, cave 17 pp14n20, 44, 56, 58, 63–6,
166; see also caves 5 and 6 84; iconographic style 14n20, 84;
cave 5-10 pp137–8, 166, 167, 184n21 inscription 14n20, 27, 58; inventory
cave 5-11 pp48, 132, 137–9, 142, 166–7 29; see also caves 16 to 20
cave 6 pp9, 43, 46–7, 49n28, 59, 63, cave 19 pp29, 41, 44, 57, 64, 72, 153;
81, 100–3, 108n66, 108n68, 147, excavation in front of 29; floor bricks
154–6, 158, 171, 173; see also caves 29; see also caves 16 to 20
5 and 6 cave 19-2 pp155, 167
226 Index
cave 20 pp26–7, 29, 36, 42, 58, 62, 108n66, 108n68, 147, 154–6, 158,
64–7, 98, 109n72; excavation in 166, 172, 174, 189; cliff surface 100;
front of 14n23, 26, 29, 64; tiles dating 6, 102; monastery remains
found in front of 29–30, 33, 57; see above 28, 33–4, 189; pagodas 100;
also caves 16 to 20 transformation of clothing style 103
cave 21 pp67, 139, 155, 157, 163, 171 caves 7 and 8 pp7–9, 16n47, 17n55,
cave 22 p157 28–9, 58, 83, 90, 94–5, 96–7, 98,
cave 22-1 p170 103, 104n2, 105n19, 107n46,
cave 23 pp141, 157 107n52, 107n55–6, 108n59, 108n63,
cave 23-1 p167 114, 124, 158, 158, 171, 176, 189;
cave 24 pp140, 160–1, 168, 171 anterooms 36–7, 82, 90, 105n19,
cave 24-1 pp141–2, 168 107n46; archways 90; dating 78, 90,
cave 25 pp134, 139, 159 94–5, 95, 99, 103, 107n56, 108n59,
cave 25-2 p142 108n63; stele between 91, 97; stūpa-
cave 26 pp135, 141–2, 168 shaped pillars 90
cave 27 pp135, 139, 141, 158, 162, 168 caves 9 and 10 pp6–9, 16n50, 28–9,
cave 27-2 pp167–8, 170 31–2, 35, 44, 44, 46, 70, 84, 90,
cave 28 pp139, 160–2, 162, 184n22 93–101, 104n2, 105n19, 107n50,
cave 28-2 pp137, 161–3, 169; 107n55–6, 108n69–70, 115, 120–1,
inscriptions 137 124, 127n4, 130n63, 153, 158;
cave 29 pp136, 139–40, 158–9, 161 anterooms 44, 82; dating 7–8, 78–9,
cave 30 pp47, 135–7, 142, 158, 161, 83, 93–101, 103; excavations in front
176 of 28–9, 31, 50n49, 83, 105n14; façade
cave 31 pp141, 158, 163, 177, 184n22, 93; traces of lotus and tortoise patterns
185n32 28, 33; wooden structure 31–2
cave 33 pp140, 142, 158–9, 163 caves 11 to 13 pp9, 11–12, 16n50, 27,
cave 33-1 p140 28–9, 31–2, 36–7, 44, 47, 63, 79–90,
cave 33-3 pp147, 149, 168 97, 103, 105n11, 105n14, 107n44–5,
cave 33-4 pp168, 169–70, 184n21, 107n55–6, 111, 112–16, 117–18,
186n44 120–6, 129n41, 130n63, 131n76,
cave 34 pp139, 156, 163, 166, 171 140, 156–7, 189–90; archaeological
cave 35 pp148–9, 152, 163–6, 167, excavation in front of 28–9, 105n11;
169–71, 184n21 cliff surface 83, 87; dating 78–90, 97,
cave 36 pp135, 141, 149, 168, 170 98, 103, 189, 107n44
cave 36-2 pp138, 170 caves 16 to 20 pp6, 8, 12, 22, 28,
cave 37 pp140, 143, 146–8, 150, 53–67; see also Tanyao caves
165–6, 171 cave sanctuaries 48, 122, 146
cave 38 pp134–5, 139, 143, 149–50, cave structure 27, 39, 44, 45, 79, 88,
153–4, 169–70, 184n22, 185n39 94, 95, 133, 156, 160, 162, 166,
cave 39 pp133, 140, 201–2, 208–9; 167, 171, 173
excavations above 62 cave tombs 20
cave 169 in the Bingling cave-temple celestial guardians 89, 112
complex 175 celestial maidens dedicating canopies
cave 254 in Dunhuang 117 46, 95, 130n63
cave 328 in Dunhuang 177 celestial musicians 80, 89, 102, 112,
caves 1 and 2 pp4, 6–8, 16n47, 49n43, 129n41
96, 99–100, 101–3, 108n65, 109n79, celestial robes 44, 102, 109n79, 156,
171, 189; architectural structure 100; 159, 167, 171, 194
central stūpa pillar 100, 102–3; cliff central Asia 15n32, 39–40, 51n71, 56;
surface 97, 100; dating 6, 78, 101–3; culture 56
Qian’er Qingshi (Wang Yu) 102–3, 189 central India 5, 15n32
caves 5 and 6 pp6–9, 15n32, 44, 44, 45, central pillar 9, 39, 80, 93, 93, 100–2,
49n28, 58–9, 63–4, 80, 97, 99–103, 105n14, 106n34, 111, 147, 171;
 Index 227
cave 11 pp80, 82; cave 12 pp80, City of Gods see Tiangong Temple
89–90, 105n7, 105n11, 106n34; clay camels 40
cave 13A p83; cave 39 p172; caves clay tents 40; see also tents
1 and 2 pp99–100, 101–3; caves 5 and cliff surface 1–3, 7, 56–7, 79, 83, 94,
6 pp100–1, 147; caves 7 and 8 p90; 97, 100, 104, 107n56, 132; natural
caves 9 and 10 pp83, 93, 105n14; structure (yinyan jiegou) 57, 79
function 111; see also stūpa caves closed mode garment 72
central plain China 12, 23, 40, 173–6 cloud pattern 57
chaitya caves pp20, 39, 45, 47, 194 coffered ceilings (pingqi) 44, 90, 140,
Chandaka 145, 148, 185n31 183n7
Chang’an 40 colossal images 22, 39, 43, 47, 53–4,
changdao (vernacular sūtra singing and 68, 189, 191; Abu Simbel Temple 20;
preaching) 36, 111, 116–20, 126, cave 5 pp100, 102; cave 13 pp80,
129n28, 129n30, 190, 195 81, 84, 88; caves 9 and 10 p97;
changfu (monastic dress) 52n79, 195 daxiangku (colossal statue cave) 81;
Changle (present-day Jixian county in Maitreya Bodhisattva with crossed
Hebei) 85, 195 ankles 80; Tanyao caves 11–12, 23,
Changnayeshe 61, 69, 195 42, 54, 64, 70, 72, 84, 100, 110,
Changzhi 187n54, 195 210, 215
Chan miyao fa jing (Scripture on column holes 33
the Esoteric Essential Methods of configuration of caves 66, 81, 128n20,
Meditation) 23, 195 128n22
chanting 36 conflated narrative mode 112, 115–16
Chan wisdom 126 Confucian ideology 46, 143–4; filial
chariot-driving 145 piety 115, 128n19; five permanents
Chavannes, Émmanuel-Édouard 2, 8 121
Chen, Kenneth 121 consecutive scenes narrative 46, 175
Chengming era 51n73, 86, 98–9, 195 Constant Pleasure 148
Chen Qingzhi 175, 195 construction materials from
Chen Yuan 3, 6, 13n8, 60 excavations 34
chi 61 continuous narrative mode 45–6, 112,
Chinese architecture 6, 9, 31, 43 115–16, 128n19–20, 144–5
Chinese culture 6, 42, 71, 174 Conversion of Nanda 95, 130n63
Chinese ethical norms 42 Conversion of the Three Kasyapas of
Chinese linear art 175 Uruvilva 45, 112, 127n1, 143, 148,
Chinese style garments 43, 100, 100, 169, 185n32
173 corbelled ceilings 44, 140, 183n7
chiwei (roof ridge ornaments) 44, 139, cord-like folds 47, 63, 155
195 cross-ankled Bodhisattva 58, 80,
Chongfu cave temple 37, 84, 95, 95–7, 139–43, 163–4, 165, 167, 168, 193
97–8, 104n2, 108n63, 108n66, 195 cross-legged Buddha 140, 142
Chongfusi see Chongfu cave temple Crown Prince see Jingmu
Chongguang Palace 86, 195
Chongjiao cave temple 7, 96, 109n79, Dachangqiuqing 196
195; inscription 7, 96 Dai Kui 9, 18n64, 41, 176–9, 196
Chronicle of Gaozu 87 Dai Yong 177, 196
chronicle of Xianzu 86 Dajiyi shenzhou jing (Sūtra of Great
chuanzuo wuqiong characters Auspicious Significance of Magic
(extending prosperity infinitely) 30–1, Spell) 11, 70–7, 89, 106n40, 123,
34, 195 172, 186n52, 196
Chu sanzang ji ji 95 Dangchang, Duke of 96
circumambulation (pradaksina) 36, 39, Dao’an 119, 196
45, 93, 100, 104, 110–11, 124–5 Daofa 125
228 Index
Daoism 2, 21, 86, 88, 121 Devas 145
Daorentong 61, 74n32, 196 devotional rituals and practices 4,
Daoshi 187n65, 196 14n22, 39, 89, 110, 125
Daowu, Emperor (Tuoba Gui) ix, 59, dharaṇi sūtra 89
64–5, 75n44, 174; see also Taizu Dharma 1, 60–1, 64, 68–9, 77n91, 94,
Daowudi see Daowu 116–17, 119, 122, 126, 143, 151,
Daoxin 119, 196 153, 171–2, 184n24; Asoka Dharma
Daoxuan 1, 3, 13n10, 26, 27, 50n50, king 152; masters (fashi jiangjing)
79, 121, 123, 126 116–19, 125, 197; mofa 21–2, 198;
dapuo (breaking space) spatial protectors 21, 45, 69–72, 77n92, 89,
relationship 113, 156, 196 140, 153, 163, 165–6
Darius the Great (c. 550–486 BCE) 20 Dharma-cakra 112
Da Song sengshi lüe (Brief History of Dharma-pitaka 70, 77n92
The Clergy of The Great Song) 21, Dharmaksema 124
196 dhyāna: Buddhas 57; halls 25; masters
Da Tang neidian lu 13n10, 59 23; monks 86; position 57, 80, 113;
dating and periodization 4, 6–8, 78, practice 23, 68; Samādhi 24; see also
81, 94, 104, 170, 190; cave 11 pp11, meditation
87, 90, 97, 99, 103, 189; cave 12 Dinglin monastery 187n54, 196
pp11, 82, 89, 103, 107n44; caves 1 dingmao day 98, 196
and 2 pp6, 78, 101–3; caves 5 and Ding Mingyi 8, 23, 118
6 pp6, 103; caves 7 and 8 pp78, 90, dingyou day 86, 196
94–5, 96, 99–100, 103, 107n56, Dipamkara 151–2; jataka 112, 115,
108n59, 108n63; caves 9 and 10 124, 151–2, 169, 173, 185n39;
pp7–8, 78–9, 83, 93–101, 103; caves prophecy 152
11 to 13 pp78–90, 97, 99, 104, 189, Dong Mi 174, 196
107n44; second-phase chronology 8, douba (octagon) 183n7, 196
11–12, 78–9, 90–103; Tanyao caves double-courtyard configuration 37
6, 9, 59–67, 75n43, 106n33; third- Doushuai cave temple 16n40, 16n44,
phase chronology 12, 156–70, 189, 96, 196
193; Yungang 3–4, 6–9, 11, 59–67, dousi (quadrilateral) 183n7, 196
75n43, 78–90, 93–101, 103, 106n33, dragons 139, 148; Fire Dragon 115;
107n44, 107n56, 108n59, 108n63, intertwined 47, 80, 84, 89, 140
189 drapery 156–7, 158, 161, 163, 165–8,
Datong 1–2, 38, 40, 51n61, 51n71, 171, 180, 194
179; see also Pingcheng Dunhuang caves 5, 44, 124, 144; cave
Datong Municipal Archaeological 254 p117; cave 328 p177; cave 428
Institute 33 p144; manuscripts 120
daxiangku (colossal statue cave) 81; see Dvarapalas 71, 80, 158, 163
also colossal images dwarf musician 112
Dazhusheng cave temples 66, 196
death (impermanence) 117 earthquakes in the capital 98
debate between Vimalakirti and Eastern Jin 120
Manjusri 120, 143, 167, 172 Eastern Temple in Luoyang 120
decline of belief in the Dharma see Eastern Wei dynasty ix, 180
mofa belief edicts: of 398 (building of Buddhist
dedicative stone slab 26 temple halls) 26; of 446 (destruction
Deer Park 112 of Buddhist temples) 53, 68; of 452
deerskin cloak 152 (Buddhist revival) 53, 62; of 452
Defeat of Mara see Mara, Assault of (making of a stone Buddha image)
Devadatta 150 22, 62; of 454 (making of five
Devapura (City of Gods) see Tiangong images) 53, 62; of 486 (adoption of
Temple Chinese culture) 174, 176, 182
 Index 229
edifying narratives 111, 114, 116 flaming nimbus 57
Egyptian temples 20 floral ornaments 80
eight celestials taught by Buddha 94, floral-scroll pattern 57
130n63 flying celestials 50n50, 80, 139–40,
elephants: Buddha in elephant form 154, 156–7
143, 146–7, 184n27; Buddha riding fodian caves (fodianku) 39, 47, 81,
143, 145–7, 169, 184n21–2, 184n27; 111, 133, 196
Buddha tossing 143, 150, 165, folding screen on niche 132, 139, 167,
185n36; Mahaprabhasa and elephant 168, 170–2, 194
trainer 95, 130n63; Outing on the folk faith 121, 125
Elephant 132; six-tusked 146 foreign relations 3
enlightenment 2, 111–12, 114, 121–2, Foshuo guanfo sanmei haijing 183n11,
126, 128n12, 147–9, 152, 184n24, 196
185n33 Fo suoxing zan (Acts of the Buddha)
ersheng see two sages 46, 196
ethnic dancers 40 Fotucheng 119, 196
explanatory prose 120 Fotuhu 71, 196
extending prosperity infinitely see Fotuobatuoluo 183n11, 196
chuanzuo wuqiong founding emperors of the Northern Wei
53, 57, 64–6, 172
fabricated sūtras 3, 111, 121–2 four celestial guardians 89, 112
fafu (formal monastic dress) 52n79, 196 four divinities offering bowls to Buddha
Faguo 22, 73n5, 172 112, 124, 127n5, 128n12, 128n21
fangbian xinlun 95, 130n63 four guardian kings 145
Fangshan, Mount 38, 51n73, 68, Fozhi biqiu liuwutu 42, 196
109n74, 196 freestanding monasteries 20–1, 30, 31,
fashanglu 106n40, 196 37, 62, 110, 190
fashi 116–19, 125, 197 frescos 175
Faxian 186n48, 196 fudao minsu (instructing mass laity) 26,
Fayuan zhulin 74n34, 187n65, 187n71, 116, 119, 196
196 Fufazang yinyuan zhuan (A History
Fei Changfang 16n42, 76n84, 196 of the Indian Patriarchs) 11, 23, 37,
Feng, Empress Dowager (Lady) see 69–70, 95, 130n63, 172, 186n52, 196
Wenming fugui wansui (wealth and longevity)
filial motive 53 34, 196
filial piety 115, 128n19 fujian see pianshan
first meditation of Buddha 143, 149, function of cave temples 10–12, 25, 39,
168, 185n34 43, 48, 100, 104, 110–12, 116, 122,
first phase 6–7, 22, 42, 43, 44–5, 45, 124, 126, 190; circumambulation
47, 48, 53, 56, 69, 104, 132, 181 39, 100, 110, 124; liturgical 10–12,
first-phase caves 39, 43–4, 48, 53, 104, 21, 37, 104, 110–11, 112, 116,
172 126, 190; meditation 11, 36; merit
First Sermon of Buddha 46, 101, 112, accumulation 36, 186n54; monastic
143, 147, 149, 168–9 remains 12, 27, 33, 36–7
five Confucian permanents (virtues) 121 Further Biographies of Eminent Monks
Five Dynasties era ix, 5 see Xu gaoseng zhuan
five elements theory 121 Fushan cave 187n54, 196
five hundred merchants worshipping future incarnation of the Buddha 11,
the Buddha 113–14, 121, 149 58, 64–5, 71, 151–2
five precepts (pancasila) 121–2, 126;
see also Buddhism, five precepts of Gandhara 38, 42, 52n79, 128n21, 149,
Five-storeyed Reliquary 53, 65, 73n4 184n27; art 42, 72, 127n7, 146–7,
Five Vows (wujie) 120 175; artists 148, 150; style 4, 15n30
230 Index
Gansu 21, 40, 56, 68 Guyang cave 173, 176, 179
Gaomiaoshan cave 187n54, 196 Guzang 40, 197
Gaoping County 186–7n54, 196
Gaoseng zhuan (Biographies of half-domed ceiling 43, 44
Eminent Monks) 3, 64, 68, 117, 196 half-palmette of three leaves 57
Gao Yun 22, 196 Hall of Taihua 87, 199
Gaozu, Emperor 9, 17n52, 66, 71, 86, Han art 144
88, 107n55; see also Xiaowen Han brick reliefs 175
garments see robes Han clothing 174
Gautama Buddha 151; meeting Rahula handling of arms 145
95, 130n63; Siddhartha 115 handscroll paintings 45–6, 144
Gazetteer of Shuoping Prefecture see Han dynasty ix, 25, 175–6, 178
Shuopingfu zhi Han stone reliefs 45
Genben shuo yiqie youbu pinaiye zashi Han stone slabs 58
(Mulasarvastiva davinaya) 118 Hang Kan 66, 109n72
giver of the narrative 114 Hanquandong 96, 108n65, 197
gong 66, 75n64, 75n66 Hariti losing her son, Pingala 46, 69,
Gongwei coup d’état (Gongwei zhibian) 95, 130n63
85, 196 heaven (tian) 122
Gongzong, Emperor 68, 196 Hebei 40, 85
Grand Chief Autumnal Prince Bai Helinge’er 23, 197; see also Shengle
Zheng 173 Henan 174
Grand Chiliocosm 72, 77n99 Heng’an (Pingcheng in the Tang) 59, 197
Great Dai 96 Heng’an cave-temple 1, 37, 50n55, 62;
Great Departure 45, 101, 132, 143, see also Yungang
145, 169, 173, 184n21–2, 184n27 hengdi 40, 197
Great Gratification 148 Heping era 6, 34, 37–8, 59–60, 62, 68,
Great Principle 88 88, 94, 106n33, 197
Greek and Roman influences 38 Hindu rituals 89
Gridhrakuta Hill 151, 153–4, 169, hip roof 32, 50n49
185n39 historical Buddha Śākyamuni 64
Group Four caves 167–70, 194 History of Chinese Art: Sculpture Volume
Group One caves 157–63, 193 (see Shina bijutsushi chōsō hen)
Group Three caves 166–7, 194 History of the Indian Patriarchs see
Group Two caves 163–6, 194 Fufazang yinyuan zhuan
Guanfo sanmei fa (Methods of Honorary Grand Mentor Zhang Tan
Visualizing the Buddha Image) 24, 196 68, 201
Guang hongming ji 68, 118 horizontal format narrative 144
Guannei (the Wei River) 123, 197 horseshoe-like plan 54
Guan siweiyi pin 183n11, 196 Hou Hanshun 55, 197
Guhua pinlu (Classification record of Hou Xudong 4
ancient paintings) 187n73, 196 Howard, Angela F. vii, 9, 25
guide the laity see fudao minsu hu 71, 197
guifu 91, 197 Huangfugong caves p43, 197
guiren (secondary wife) 85, 197 Huangjiu Monastery 68, 197
guji baoyangsuo 26 Huangxing era 63, 86–7, 94, 197
Gujin yijing tuji (Seal of Ancient and huasheng 120
Modern Sūtra Translation) 37, 197 Huayan cave temple 16n40, 16n44,
Gu Kaizhi 18n64, 40, 176–7, 196 51n57, 197
Guoqu xianzai yinguo jing (Sūtras on Hubei province 144
Causes and Effects of the Past and Huguo cave 7, 16n40, 16n44, 16n47,
Present) 46, 150, 183n10, 197 26, 37, 51n57, 95, 95–6, 98, 104n2,
Guo Ruoxu 181, 197 197; inscription 95–6
 Index 231
Huifusi Temple stele 102 108n68, 109n79, 132, 190; by nuns
Hui Jiao 117–19 26–7
Huishi 23, 197 intermixture and fusion of multiple
humans (ren) 122 cultures and peoples 43
huren (ethnic groups) 40, 197 inverted V shaped roof brackets 31, 44
hutengwu dancers 40, 197 Iran 20
Itō Chūta 2, 4
iconographic composition of images 27,
43, 45–6, 47, 94, 108n68, 132–3, jambu (rose-apple) tree 149, 185n34
140–4, 156, 166, 167, 171, 193–4 jataka narratives 45, 46, 69, 80, 112,
iconographic shift 176 116, 118, 140, 142–3, 151–4; Asoka
iconographic style 3, 5–6, 8, 9–10, 115, 124, 151, 152–3, 158, 169,
14n20, 27, 47, 84, 94, 102, 110, 173, 185n39; Dipamkara 112, 115,
154–6, 166, 170–2 124, 151–2, 169, 173, 185n39;
iconographic transformations 177, 182 Mahasattva 144; Mugapakkha
identification of imagery 110, 127n4 107n46; Sudhana 144; Syama 45–6,
image making 22, 38, 53, 64, 69, 72, 115, 128n19, 143
120–4, 126, 172, 174 Jetavana monastery 152
imperial consort 85 jiangjing (sūtra lectures) 36, 114, 116,
imperial family of Northern Wei 21–2, 126, 129n28, 197
53, 88, 173; accumulation of merit jiangjing fashi (Dharma masters who
21–2; ancestors 48; patronage 1, 72; give sūtra lectures) 114, 118, 126, 197
shrine temples 54 Jiang Shaoyou 177
imperial five see Tanyao caves Jiankang 180
impermanence 117, 119 Jiankang shilu 180, 197
Imprisonment of the Fire Dragon in the Jiaoshansi caves 27, 50n46, 197
Bowl 115 jiasha (Kasaya) 52n79, 197
incense burner 87, 114 Jinbei 190, 197
incised lines 44, 58, 155; vertical 168 Jin dynasty x, 26, 77n92
India 5, 15n30, 20, 42, 52n79, 54, 118, Jingchong cave temple 16n44, 51n57,
144–5, 153, 178; culture 27–8, 56; 96, 197
monks 37; painters 180; patriarchs Jingdu sanmei jing 106n40, 197
70; style 72 Jingmen 144, 197
Indian and Gandhara influences 15n32, Jingming era 14n23, 26, 66, 197
38, 42, 184n27 Jingmu, Emperor (Tuoba Huang) 64–6,
Indra 145 75n44
ink stone 40, 51n71 Jingtu sanmei jing (Samādhi Sūtra on
inner robe 42–3, 52n79, 57, 160 the Pure Land) 37, 51n55, 124, 197
innovation in cave temples 42–3, 48, Jin stele inscription 7, 13n11, 16n40,
56, 88, 104, 106n34, 132 16n43, 16n48, 26, 31, 37, 51n56, 59, 2,
innovation of Chinese Buddhist 83, 84, 88, 95–7, 98, 106n33, 108n68,
sculpture 171, 173, 194 109n79, 132, 190; Cao Yan 7, 16n43–4,
inscriptions 2–5, 7, 9–10, 12, 14n20, 37; Miu Quansun 13n11, 198
14n23, 19n72, 116, 132, 156, 190; Jin Weinuo 108n68, 144
cave 4 pp132, 170; cave 11 pp16n50, jiu’er juezhi (based on the nature of the
27, 80–1, 83, 84, 97, 123, 131n76, mountain cliff) 79, 197
211; cave 11-14 pp14n20, 174; cave Juqu Mujian 68, 197
17 pp14n20, 27, 58; cave 28-2 pp137, Jurassic stratum 98
161–2; cave 35 p163; Chongjiao cave
7, 96; Huguo cave 95–6; Jin stele Kaihua monastery 187n54, 197
1–2, 7–9, 13n11, 16n40, 16n43, Kaihuang period 123
16n48, 26, 31, 37, 51n56, 59, Kaiyuan shijiao lu 16n44, 74n34,
62, 83, 84, 88, 95–7, 98, 106n33, 106n40, 197
232 Index
kalpas 69, 72, 76n88, 151 linked narrative scenes 144
Kanthaka, Buddha’s horse 143, 145, lishe 14n22, 198
148, 184n22, 185n31 Liu Chang 102, 198
Kapilavastu 151 Liu Xiaobiao 68, 198
karma 115, 116, 125, 151–2; core Li Yi 86–7, 198
tenets of 151; retribution in samsara Li Yuqun 20, 107n44
115, 122 Loewe, Michael 65
Kashiga (Shale) 38, 197 Longmen 5, 176; Binyang 173; caves
Kasyapas 45, 112, 127n1, 143, 148, 8, 10, 15n32, 18n64, 19n72, 43, 66,
169, 185n32 98, 143, 149, 173, 176, 182, 190;
Kaundinya 149 Guyang cave 173, 176, 179; Lu cave
Kekaya 69, 76n86, 95, 130n63 143; period 176
King of Dai 65 Longwang Shrine (Longwangmiao)
Kizil caves p25, 36, 39, 100, 118, 191 28, 198
kneeling figure 57, 113, 149, 153, 168, lotus-born Buddhas see aupapadukas
184n20, 186n47 lotus design: cave 10 column 28, 31–2;
knowing viewer 116 ceilings 140; petal bands 112; tiles
konghou 40, 197 30, 34
Kucha 9, 39, 118 Lotus Sūtra 47, 70–1, 116, 120, 123,
143, 171–2, 183n5
lacquer box 144 Lou Rui, Prince 180, 198
Lady Feng see Wenming Loving Sex 148
Lalitavistara 46, 114, 121, 127n5, 186n43 lower terrace 25
Land of the Lion (Sri Lanka) 38 Lubanyao caves 27, 49n43, 198;
Lang 9, 101 archaeological excavations 27
lanruo (hermitage) 21, 197 Lumbini garden 147
Lao and Zhuang 88; emptiness of 71 Luoyang 77n92, 120, 172–6, 186n54,
Later Wei 96 191; transfer of capital to 2, 35,
lecture halls (jiangjingtang) 36, 118–19 63, 82, 85, 103, 108n63, 132, 172,
Leshan tombs 175 173–4, 182, 189
Liang dynasty 9, 95, 176, 179 Luoyang qielan ji 175
Liang Sicheng 3, 6, 56 Lu Tanwei 41, 176–7, 179, 198
Liang Wudi, Emperor 180 Luyeyuan park caves 25, 49n28, 86,
Liangzhou 15n32, 21, 23, 38–9, 59, 88, 198
110, 124, 126, 172, 198; Buddhism
38–9, 124 Mahākāśyapa 70, 77n91, 143, 158
Liao dynasty x, 7, 16n44, 31, 32, 189; Mahaprabhasa and the elephant trainer
grand-scale structure remains 29, 32, 95, 130n63
34; tiles and brick remains 28, Mahasattva jataka 144
50n43 Mahayana 21, 77n92, 89
Libaiku 23, 133, 198 maigre feasts (zhai) 117, 120, 124
Lidai minghua ji 177–8, 181 Maijishan caves 44
Lidai sanbao ji 16n42, 74n34, 76n84, Mair, Victor 119
198 Maitreya 39, 89, 94, 98, 124, 128n9,
Li Daoyuan 1, 3, 26–7, 50n43, 57, 79, 141, 143, 166, 171–2, 194; ascent
198 and descent 40, 94, 98, 141, 143,
Li Jingjie 108n69, 109n79, 123 146, 171–2; Bodhisattva 56, 65,
linear design 175–6, 180, 182, 188n79 80, 89, 171; Buddha 45, 71, 98,
linear narrative sequence 144 108n68–9, 112, 174; Śākyamuni
Lingchuan County 187n54, 198 and Maitreya 108n68–9, 141, 166,
Lingyan cave temple 1, 16n44, 26, 37, 171–2, 194; Sūtra 40
51n55, 51n57, 96, 198 Maitreya-Amitayus stele 174
Lingyu 66, 198 Manes jewels 40
 Index 233
Many-Jeweled Buddha 47; see also mofa belief (the decline of belief in the
Prabhutaratna Dharma) 21–2, 198
Maoxian 174, 198 Mokemoye jing 183n11, 198
Mara, Assault of (Mara’s Temptation, Moling (Nanjing) 175
Banishing of the Demons) 112, 115, monasteries 3, 9, 21–2, 25–6, 28, 30–8,
128n21, 143, 148, 169–70, 173, 40, 48, 50n43, 51n55, 51n73, 62,
185n32, 186n48 68, 77n98, 118, 153, 171, 186n45,
Mara’s Temptation see Mara, 187n54; freestanding 20–1, 30, 31,
Assault of 37, 62, 110, 190, 189–90; layout of
Mathura 42, 52n79 33, 38, 45, 82
Maya 146–7, 150, 169, 184n27; monastery remains in Yungang 28, 30,
Buddha’s ascent to Tavatimsa 143, 33–4, 36–8, 189–90; anterooms 36;
150, 169, 183n11; Maya’s Dream brick beds 34, 37; brick walls 30;
146–7 courtyards 34, 37; flues 34, 37; iron
Mayadevi see Maya casting remains 34; kitchen stoves 34,
McNair, Amy 10 37; residential cells 33–4, 37, 104;
meditation 9, 11, 22–5, 65, 110, 113, stūpa base 34, 37–8; suite cells 34,
124, 128n9, 128n12, 148, 186n54, 37; tiles 29–31, 34
190; cells 9, 25, 49n28; first 143, monastic comptroller see shamentong
149, 168, 185n34; halls 65, 86; Mongolia 23
see also dhyāna mono-scene narratives 45, 114–15,
merchants 40; caravan 113; five 128n20, 143, 145
hundred 113–14, 122–3, 149; mortuary art and architecture 56, 73
offering food to Buddha 45, 113, Mount Grdhrakuta (Vulture Peak) 25
128n10; Trapusa and Bhallika (Tiwei Mount Sumeru 25, 65, 80, 105n8
and Boli) 113–14, 121–2 Mount Wuzhou 1–3, 17n52, 22, 25–6,
merit 21–2, 36, 113, 124, 126, 153, 62, 68, 172
186n54; accumulation 21–2, 36, mudra-mantra-mandala 89
113, 126, 153, 186n54; Trapusa and multi-armed deities 166
Bhallika 113 mural paintings 40, 55, 73n11, 178,
Methods of Visualizing the Buddha 180
Image see Guanfo sanmei fa musical deities 80, 89, 111, 112, 118,
Mi Fu 180 129n41
migration: Buddhism 60; migrants musical instruments 40, 80, 112,
39–40, 43, 48, 51n71, 60; to 129n41; hengdi 40, 197; konghou
Pingcheng 39–40, 59–60, 124, 176; 40, 197; lutes 118; paixiao 40, 198;
Sogdian 40 pipa 40, 112, 198; yaogu 40, 200
Ming dynasty x, 1, 59, 96, 183n7
Mingseng Zhuan 124, 131n84, 198 nadanas 116
Mingyuan, Emperor (Tuoba Si) ix, Naga (water spirit) brothers, Kala and
17n52, 26, 37, 59, 64, 65–6, 75n44, Akala 148
96; prayer to gods of heaven and Nagahiro Toshio 8, 15n32, 63, 84, 87,
earth 26 103; see also Mizuno Seiichi and
Miu Quansun 13n11, 198; see also Jin Nagahiro Toshio
stele inscription Naga Kings 89
Mizuno Seiichi 15n32, 61, 104n2, 113, Nagarjuna defeats Brahmana in a
120, 124; see also Mizuno Seiichi and debate 46
Nagahiro Toshio Nagas Bathing the Buddha 143, 147–8
Mizuno Seiichi and Nagahiro Toshio nan (questions) 120
5–6, 8, 14n23, 15n32, 61, 62–3, 67, Nanda 95, 130n63, 150
78, 81, 82–3, 89, 94, 102, 105n11, Nanjing (Nanking) 175, 178; see also
105n14, 113–14, 183n7, 189; Moling
excavations of 28–31 Nanking 102; see also Nanjing
234 Index
Nanqishu 198, 204 proselytizing mass laity 115–26, 126,
Nanshi 198, 203 146; Rahula 95, 130n63, 151, 153,
Naqsh-i Rustam 20 169, 185n39; receiver of 114; riding
narratives 36, 45–6, 52n89, 69, 71, 90, an elephant 95, 130n63, 132, 143,
94–5, 101, 110–26, 129n30, 132, 145–7, 149, 169, 184n21–2, 184n27;
139, 143–54, 158, 167, 169, 173, Seven Steps 147; successive scenes
183n9; archery competition 143, 45–6, 144; Svastika bird 95, 130n63;
149–50; ascent to and descent from synoptic mode 115, 128n22; tossing
Tusita Heaven 40, 95, 98, 141, 143, the elephant 143, 150, 165, 185n36;
146, 171; ascent to Tavatimsa 143, viewers’ response to 114–16; woman
150, 169, 183n11; avadanas 116, who hated desire and become a
129n30, 129n40; Banishing of the nun 69, 95, 130n63; see also jataka
Demons 112, 115, 128n21, 143, 148, narratives
169, 173, 185n32, 186n48; birth naturalism 59, 65, 176
of Buddha 101, 115, 143, 146–8, nature of the mountain cliff see jiu’er
185n33; celestial maidens dedicating juezhi
canopies 46, 95, 130n63; conflated niches 1, 4, 11, 14n23, 44, 45, 47,
mode 112, 115–16; consecutive 57, 58, 64, 79, 80–1, 82–3, 90, 91,
scenes 46, 175; continuous mode 94, 98, 100, 102, 103, 112–13,
45–6, 112, 115–16, 128n19–20, 114–15, 122–3, 132–42, 149, 152–3,
144–5; Conversion of the Three 156–73, 190, 192–3; canopied 139,
Kasyapas of Uruvilva 45, 112, 158, 161; caves without niches 137,
127n1, 143, 148, 169, 185n32; 141; compartments 44, 139, 149;
didactic function 115; Dunhuang curtained 90, 168, 171, 193; folding
paintings 117; eight celestials taught screen front 132, 139, 167–8, 170–2,
by Buddha 95, 130n63; as ethical 194; pointed-arch 81, 91, 113, 193;
guides 117; Farewell to Kanthaka roofed 44, 100, 112, 139; storied 45,
143, 148, 184n22, 185n31; Fire 80, 133–5, 140–2, 158, 166, 167–8,
Dragon’s Imprisonment 112, 115; 171, 172, 192–3; trabeated 91, 94,
first meditation 143, 149, 168, 97, 112, 132, 134, 137, 139, 149,
185n34; First Sermon 46, 101, 112, 163, 167–8, 170–2, 193–4
143, 147, 149, 168–9; as form of nidana narratives 46, 69, 117, 129n30,
instruction 115, 119; four divinities 143
offering bowls to Buddha 112, 124, Niepan jing 198, 202
127n5, 128n12, 128n21; giver of Nirgrantha-putras redeemed by Buddha
114; Great Departure 45, 101, 132, in the fire Samadhi 46, 95, 130n63
143, 145, 169, 173, 184n21–2, nirvana 21, 70, 76n88, 77n91, 115,
184n27; Gridhrakuta Hill 151, 121, 125, 143, 152–3
153–4, 169, 185n39; Hariti losing nirvana scene 46
her son, Pingala 46, 69, 95, 130n63; Nirvana Sūtra 21
horizontal format 144; linear 144; Noble Truths of Buddhism 118
linked scenes 144; Mahaprabhasa nomads 54–5, 73n12
and elephant trainer 95, 130n63; north China caves p133
Mara’s Assault 112, 115, 128n21, Northern Dynasties ix, 4, 18n64, 23,
143, 148, 170, 173, 185n32, 55, 112, 118, 152, 191
186n48; modes 114–16, 126, Northern Liang 21, 43, 43, 59, 68, 124
128n20; mono-scene 45, 114–16, Northern Park 49n28, 86
128n20, 143, 145; nagas bathing Northern Qi dynasty ix, 176, 179–81
the Buddha 143, 147–8; nidanas Northern Song 180–1
46, 69, 117, 129n30, 143; offering Northern Wei ix, 1, 4, 7, 11, 21–3, 25,
of a meal by Sudatta’s wife 46, 95, 27, 29–30, 31–2, 34–9, 40, 43, 53,
130n63; Parinirvana 143, 147, 151, 57–8, 59–60, 65, 72, 88, 96, 124,
169; pictorials 114–24, 123, 146; 132, 174–5, 176, 180–1; era 31, 36,
 Index 235
42, 62, 83, 85; five founding emperors Pengshan 175
53, 57, 64–6, 172; monasteries 29–31, pensive Bodhisattvas 141–3, 149
33–6, 62, 189; rulers 21–2, 35, 48, pensive figure 91
53, 64, 69, 110, 189 performative explanatory activities 110,
Northern Yan 40 112, 119
North Terrace see Beitai caves perpendicular plane 83, 88, 94, 100
nunnery 26, 27 Persian Achaemenid Empire 20
nuns 3, 26; dedicative inscriptions by pianshan (patch) 42–3, 52n83, 57–8,
26; see also Bhikkhunis 198
Nüshizhen tu (Admonitions to the pianxiuyi 42
Court Ladies) 178, 180 pillars 39, 44, 90, 111, 139; cave
11 pp80–1, 83; cave 12 pp80, 90,
octagonal stūpa base 34 105n7, 105n11, 106n34; caves 5 and
offering of a meal by Sudatta’s wife 46, 6 pp100–1, 147; caves 7 and 8 p90;
95, 130n63 caves 9 and 10 pp83, 93, 105n14;
Offering of Food by the Merchants 45, see also stūpa pillars
113 Pingcheng 2–3, 9–10, 21–3, 25, 38,
Ōmura Seigai 4–5, 9, 64 40, 56, 59–60, 72, 88, 95, 104,
one good deed calls for another precept 118, 124, 132, 171, 174, 191, 198;
115 archaeological excavations near
One Niche on the North and Storied 40–1, 48, 54, 190; configuration of
Niches on the Side Walls 133, 134, the monasteries 38, 45; era 32, 38,
140, 167, 171, 172, 192, 194 85; monastery remains 38; palaces
open mode garment 42–3, 72 and other buildings 32, 35
oral presentations in Buddhist pingqi 71, 140, 183n7, 198; see also
sanctuaries 111, 115, 117, 119, 126 coffered ceilings
orthodox canon 111, 121 pingqi zaojing 140, 183n7, 198
Otsuka Nobuo 89, 198 Pingwen, Emperor 64, 75n44, 198
Outing on the Elephant 132; see also pipa 40, 112, 198
elephants, Buddha riding please sing (chang jiang lai) 120
Outlining the Methods of Plebeian Buddhism 121–2, 129n27
Contemplation see Siwei lüeyao fa plowmen’s village 149
power struggle between Xianwen and
pagodas 6, 43, 62, 171; caves 5 and 6 Wenming 85–7, 89
p100; caves 9 and 10 pp12, 83 Prabhutaratna 45, 70, 94, 120, 134–5,
paired caves 4, 12, 44–5, 95, 96–7, 98, 140–3, 153, 171–2, 183n5, 193; see
103, 132, 173, 189 also Many-Jeweled Buddha
paixiao 40, 198 pradaksina see circumambulation
pancasila see Buddhism, five precepts of Prince Kang of Jingzhao 86, 197
Pao-chen Chen 144 proselytizing mass laity: TWBLJ 3, 111;
parasol 145 visual narratives 114–24, 126, 146
Parinirvana narrative 143, 147, 151, 169 protection spells (Dharaṇi) 89
passageways (passages) 36, 39, 45, Pure Land faith 37, 126
93–4, 100, 107n49, 110–11, 125; Putai caves 43, 198
connecting anterooms 45, 95 putong (ordinary) 21, 198
patronage 4, 7, 9, 12, 69, 85, 87, Puyao jing 46, 198; see also
96, 132, 157, 170–2, 190; cave 13 Lalitavistara
pp11, 80, 87; Northern Wei imperial
family 1, 34, 69, 72, 80, 189; Qian’er Qian’er Qingshi (Wang Yu) 7, 37, 95–6,
Qingshi 102; social status of patrons 102–3, 198; caves 1 and 2 p101–2,
172; Wenming (Lady Feng) 38; 189; patronage 102
Xianwen 87, 95; Xianwen’s qidaoren (black monk) 119, 198
allies 85 Qin dynasty ix, 144
236 Index
Qing dynasty x, 4, 6, 59, 96–7, Śākyamuni and Prabhutaratna 45, 70,
107n50, 113 94, 120, 134–5, 140–3, 153, 171–2,
qingtan (pure conversation) 46 183n5, 193–4
Qingzhou 177, 181 Sakya Tathāgata 72
Qixiashan caves 175, 198 Samādhi Sūtra on the Pure Land 37; see
Qiunabatuoluo (Gunabhadra) 183n10, also Jingtu sanmei jing
198 Samarkand 181
Samghapala 77n92, 198
rafters 6, 44, 139; holes for 7, 16n44, Samghati 52n79, 198
31, 33, 49n43 samsara 115, 122, 126
Rahula 57; Buddha meeting 95, sanbi sankan 47, 132, 134, 158–60,
130n63, 151, 153, 169; jataka 151, 163, 171–2, 198; see also three
153, 185n39 niches on three walls
Ramses the Great 20 sangha 42
receiver of the narrative 114 Sangha-household economic system 71,
Record of Excavating a Cave Temple 66 77n98, 122, 128n12
Rei see Yoshimura Rei Sanqian daqian shijie
reincarnation 151–2 (trisahsramahasasralokadhatu)
religious and secular, reciprocal 77n99, 198
influence 10, 38, 40, 73, 190 Sarnath 149
ren (humans) 122, 126 Scripture on Concentration in Sitting
rentian cult 126 Meditation see Zuochan sanmei jing
ren tian jiao (human-heaven teaching) Scripture on the Esoteric Essential
121, 129n27, 198; see also Plebeian Methods of Meditation see Chan
Buddhism miyao fa jing
roads: ancient ruts 23–4, 35; between scroll paintings 45–6, 144
Pingcheng and Shengle 23–6, 35, 190 Seal of Ancient and Modern Sūtra
robes: Antarvasaka 52n79, 195; Translation see Gujin yijing tuji
baoyi bodai 43, 47, 59, 63, 103, second phase 6, 8, 31, 39, 40, 43, 45,
155, 173–6, 179, 182; celestial 44, 47, 63, 79, 88–9, 116, 171, 173,
102, 109n79, 156, 159, 167, 171, 182, 189
194; changfu 52n79, 195; Chinese second-phase caves 6, 8, 44, 45, 47,
style garments 43, 100, 100, 173; 69, 78–9, 90, 105n2, 105n19, 124,
cord-like folds 47, 63, 155; fafu 132, 139–40, 155, 171–2, 189;
52n79, 196; flattened pattern 47, 63, chronological sequence 8, 11–12,
155, 158, 160, 163, 167; imperial 78–9, 90–103; inscriptions 81, 116;
garments 85; inner 42–3, 52n79, 57, liturgical function 12, 39, 45, 48,
160; Samghati 52n79, 198; step-like 104, 110–11, 116, 126; narratives
folds 44, 47, 59, 63, 100, 155, 163; 69; patronage 11, 69; visual
Three Garments 42, 52n79; tongjian representations 69
(both shoulders covered) 102, 175; second transformation in the evolution
traditional Chinese 47, 104; wet of Chinese Buddhist sculpture 181
clothes style 72 secular and religious, reciprocal
rock-cut cave temples 20–1, 27, 37, 98, influence 10, 38, 40, 73, 190
133, 170, 172–3, 189; function 11, secular material culture 10, 73, 190
110–11, 116; layout 20; origins 20 secular tomb art and architecture 48
roof eaves 30–2, 44, 50n49, 105n7 Sekinō Tadashi 5, 64
Sengfangku 199
Śākyamuni 45, 47, 53, 64, 70, 71, 88, sengren changdao (the clergy who
94, 108n68–9, 120, 134–5, 140–3, perform vernacular sūtra singing)
152–3, 166, 171–2, 183n5, 193–4 116, 119–21, 126, 195
Śākyamuni and Maitreya 108n68–9, Seng Shi Lue 119, 199
141, 166, 171–2, 194 Sengxian 68, 199
 Index 237
Sengyi 52n79, 68, 199 Siwei lüeyao fa (Outlining the Methods
Sengyou of the Liang 95 of Contemplation) 25, 199
Sengzhihu 199 Six Dynasties era 117, 119
seven bay wide span 31 Six Laws 180
seven Buddhas of the Past 71, 89, 112 six paramitas 152
Seven Sages of the Bamboo Grove and Sixteen Kingdoms era 21, 23, 39
Rong Qiqi 178 Siyuan monastery 38, 68, 199
seven steps 147 Sizhou 175
seven-story reliquary 88 social status of patrons 172
Shale (Kashiga) 38, 199 Sogdia 40; dance 40; merchants 40;
Shaling tomb mural paintings 40, 73n11 musicians 40
shamentong (monastic comptroller) 22, Song Dexing 42–3, 199
53, 61, 172, 199 Song dynasty x, 16n44, 21, 102, 175,
Shandong Province 8, 77n98, 181 183n7
Shan’xi 173–4 Song Shaozu 40, 199
Shanxi 1, 5, 8, 51n71, 180, 189, 191; Songshu (The Book of Song) 178, 199
minor caves p12, 173, 187n54, 191 sonorous voice (sheng) 117, 199
Shanxi Archaeological Institute 32, 33 Soper, Alexander C. 5, 8–9, 18n64,
Shengle 23, 35, 199 18n71, 64, 101, 107n45, 107n55
Shenrui era 9, 59–60, 108n59, 199 Southern Dynasties ix, 113, 117
Shigudong 96, 199 southern influence 18n64, 41,
Shijiazhai 179 175–6, 182
shili (teacher) 68, 199 spare style 180
Shilihe (Shili River) 1, 199; see also spatial conception, narrative 144
Wuzhou River spatial relationships of caves p79, 156
Shina bijutsushi chōsō hen (A History Spink, Walter 40
of Chinese Art: Sculpture Volume) 4 śramanas 21, 26, 38, 53, 60–1, 65, 68,
Shitanghui cave 187n54, 199 73n5, 74n32, 88, 120, 124, 126
Shixian 23, 60–1, 68, 74n32, 126, 199 Sravasti 152
Shizong, Emperor 66, 173; see also state of the field 12
Xuanwu Steinhardt, Nancy 9
Shuijing zhu 2–4, 6, 26, 37, 199 steles 2, 5, 10, 13n11, 14n23, 59, 91,
Shuntianfu zhi 13n11, 199 96, 102, 107n47, 174, 179, 183n8;
Shuofu 188n83, 199 see also Jin stele inscription
Shuopingfu zhi (Gazetteer of Shuoping step-like folds 44, 47, 59, 63, 100, 155,
Prefecture) 6, 199 163; flattened pattern 47, 63, 155,
Sichuan 174–5 158, 160, 163, 167
Sifenlü shanbu suiji jiemo stone chambers 37, 190
shuzhengyuanji (Commentary on the stone monasteries 3, 21, 26, 50n43
revised monastic procedures of the stone reliquary 89
four-part Vinaya) 42, 52n81, 199 stone slab 26, 27, 58
Sifenlü xingshi chao zichi ji (A Storied Niches on Three Walls 133,
commentary on the Dharmaguputaka 135, 141, 158, 168, 171, 192
Vinaya) 43, 199 story of the Svastika bird 95, 130n63
Sila-conscious Liangzhou monks 126 story of the woman who hated desire
silas 126 and become a nun 69, 95, 130n63
Sima Jinlong 40, 51n74, 179, 199 stratum: excavations 29–30, 35; rock
singing and leading preachers 36, 67, 98, 170
(changdaoshi) 117 stūpa caves 23, 26, 39, 45; cave 6 pp46,
single stūpa monastery configuration 38 100, 102; cave 11 pp45, 80–1, 83,
sinicization 6, 18n64, 43, 44, 47, 51n76, 106n34; cave 39 pp133, 165–6, 172;
103, 144, 173–4, 176–7, 182, 190 caves 7 and 8 pp90; caves 9 and 10
Sirén, Osvald 8 pp120–1, 98–9
238 Index
stūpa pillars 9, 39, 80, 90, 98–100, of the Wise and the Foolish) 152,
101–2, 105n14, 106n34, 111, 147, 224n45
171; cave 6 pp100–1, 147; cave 11 Syama jataka 45–6, 115, 128n19, 143
pp80, 82; cave 39 p170; caves 1 and symbols of imperial rulers 53, 80
2 pp99–100, 101–2; function 111; synoptic narrative mode 115, 128n22
see also stūpa caves
stūpas 9, 21, 38, 45, 60, 110, 125, Ta E 108n67, 199
153; courtyards 34, 36–8; Siyuan Tai’an era 38, 63, 199; images 63
monastery 38; Yungang monastery Taihe era 6–7, 14n20, 17n52, 27, 38,
33–4, 37, 49n43 40, 44, 51n73, 58, 81, 83, 83–5,
style 3, 9–10, 12, 15n30, 15n32, 90, 95, 94–6, 98, 102–3, 109n74,
17n55, 18n64, 41, 42, 44, 50n43, 109n79, 118, 120, 156–7, 158,
63, 67, 72, 83, 98–101, 132–3, 157– 174, 179
8, 162, 166–8, 173–81, 189, 191, Taipingzhenjun era 41, 63, 71, 199
193, 194; iconographic 3–6, 8–10, Taiwu, Emperor (Tuoba Tao) ix, 21–2,
14n20, 27, 47, 84, 94, 154–7, 166, 25, 38, 40, 64, 66, 71, 75n44, 172
170–2; transformation 41, 51n76, Taiyan era 38, 40, 59, 199
101, 103–4, 173–4, 177 Taizu, Emperor 53, 59, 62, 65, 73n5;
Su Bai 4, 6–8, 16n44, 16n48, 18n64, see also Daowu
23, 26, 40, 56, 61–4, 78, 84, 94, 96, Takahashi Hisao 107n41, 199
103, 105n2, 105n11, 108n65, 173 talent (cai) 117
successive narrative scenes 45–6, 144 Tandu 118, 199
Sudatta 46, 95, 130n63 Tang dynasty ix, 1, 7, 26, 31, 59, 62,
Suddhodhana 149 75n66, 117, 119, 179
Sudhana jataka 144 Tanjing 3, 123, 183n11, 199
Sui dynasty ix, 66, 126 Tanmei 14n23, 26, 199
Suishu (The Book of Sui) 174 Tantric Dharaṇi text 89
sujiang (vernacular oral presentation) Tanyao 2–3, 6–7, 11, 22–3, 36–7, 46,
111, 118–19, 120, 126, 199 50n43, 53, 56, 60–1, 67–71, 74n16,
sujiangshi (vernacular oral presenters) 76n86, 77n92, 84, 89, 94–5, 94,
114 120–4, 126, 130n63, 151, 172, 199
Sullivan, Michael 144 Tanyao caves p6, 9, 12, 22–3, 28–9, 39,
Sumadhi, a Brahmin youth 151 40, 44, 48, 50n43, 53–67, 72, 79,
Sumeru, Mount 25, 65, 80, 105n8 84, 88, 106n40, 154; architecture 12,
sūtra lectures see jiangjing 56; asymmetry 36; colossal images
sūtra lecture texts 120 11–12, 23, 43, 54, 64, 72, 84, 100,
Sūtra of the Great Auspicious 110, 177; configuration 36, 50n43,
Significance of Magic Spell see Dajiyi 65–7; dating 6, 9, 59–67, 75n49,
shenzhou jing 106n33; excavations in front of
Sūtra of Trapusa and Bhallika 3, 28–9, 33; as family shrine 36, 54, 65,
111, 114; see also Tiwei Boli Jing 72, 110, 190; functions of 12, 36,
(TWBLJ) 110, 190; see also caves 16 to 20
sūtras 3, 6–7, 37, 46–7, 49n26, tanyou robes 42, 57–8, 102
52n89, 61–2, 67, 70, 76n86, Tao Zongyi 188n83, 199
106n40, 117, 120–1, 125–6, 172, Tathāgata incarnate 22, 53–4, 57,
182n1; Amitabha 120; dharaṇi 89; 58–9, 64–5, 69, 72, 73n5, 76n88,
fabricated 3, 111, 121; Lotus Sūtra 110, 173
47, 70, 116, 120, 123, 143, 171–2, Tavatimsa, Buddha’s ascent to 143,
183n5; Maitreya 39; Nirvana 21; 150, 183n11; cave 38 pp169–70
Tiwei Boli jing (TWBLJ) 3, 111, Temple of the Heavenly Palace see
121–6, 131n71, 200; Trapusa and Tiangong Temple
Bhallika 3, 111, 113–14; Vimalakirti temptation 148
120, 172; Xianyu jing (the Sūtra ten good deeds 122
 Index 239
Ten Great Disciples of Śākyamuni 58, 72 tomb murals 40, 73n11, 178
tents: baizizhang 73n12, 195; clay tombs 10, 20, 40, 48, 51n74, 55–6,
model 40, 54–6; half-dome shaped 73n11, 144, 175, 178–80, 188n79, 190
54; murals 55, 73n11; nomads 55, tong (answers) 120
73n12; structure of caves 73 tongjian robes (both shoulders covered)
ten virtues 126 102, 175
third phase 4, 43, 47–8, 103, 132–82, Tongle Temple 16n44, 26, 37, 50n55,
183n5, 184n21, 190; iconographic 51n57, 62, 69, 96, 200
composition 132–3, 140–4, 156, Tongzi Temple 16n40, 16n44, 51n57,
166–8, 171 96, 200
third-phase caves 6, 8, 11–12, 132–82, tossing the elephant 143, 150, 165,
192–3; architectural structure 47, 185n36
132–4, 158; chronological sequence traditional Chinese robes 47, 104
12, 156–70, 189, 193; subject matter transcendental Buddha (fashen) 120
47, 132–3, 153, 143, 158, 162, 163, transfer of the capital to Luoyang 2, 35,
165–6, 168–9, 172–3 63, 82, 85, 103, 108n63, 132, 172,
thousand Buddhas 43, 45, 50n43, 57, 174–6, 182, 189
58, 72, 81, 87, 90, 133, 138, 166, translation centers 104
167, 192 transmigration 115; see also samsara
three Buddhas 11, 38, 43, 45, 69, 72, Trapusa and Bhallika 3, 111, 113–14,
89, 103, 113, 141–3, 153, 166, 171, 122, 128n12
172, 194 triangular ornaments 44, 139
three Buddhas in three kalpas 69, 72–3 Tsukamoto Zenryu 64, 68
three daughters of Mara (Loving Tuhua jianwen zhi 181
Sex, Constant Pleasure, and Great Tulufan 49n28, 200
Gratification) 631 Tuoba Gui see Daowu
Three Garments 42, 52n79 Tuoba Huang see Jingmu
three kalpas 69, 70–1 Tuoba Jun see Wencheng
three niches on three walls 47, 132–4, Tuoba prince 96
140, 158, 160, 166–7, 171, 192, 194; Tuoba Si see Mingyuan
see also sanbi sankan Tuoba Tao see Taiwu
Three Refuges (sangui) 120 Turfan 25, 224n45
Three Treasures 153 Tushita Heaven 40, 97; see also Tusita
tian (gods) 122, 126 Heaven
Tian’an era 52n83, 86, 88 Tusita Heaven: ascent to and descent
Tianci 59, 199 from 113, 141, 143, 146, 171; gods
Tiangong Temple (Heavenly Palace) 152; see also Tushita Heaven
16n40, 16n44, 26, 37, 51n57, 88, Tuyuk caves 25, 200
96–7, 98, 108n68, 199 TWBLJ 111, 121–6, 131n71; see also
tianhua 183n7, 199 Tiwei Boli Jing
Tianxing era 25, 59, 174, 199 twin Buddhas (Śākyamuni and
tiles 28–31, 33, 34–6, 50n43, 57; with Prabhutaratni) 45, 47, 160, 169
characters 30–1, 34; green glazed 30, two brothers who became monks 46,
34; with lotus patterns 30, 34 113, 130n63
Tiwei (Trapusa) 121 two niches on three walls 133, 137,
Tiwei and Boli, two merchant brothers 142, 166
121; see also Trapusa and Bhallika two pensive Bodhisattvas 141–3
Tiwei Boli jing (TWBLJ) 3, 111, 121–6, two sages (ersheng) 95, 102, 173
131n71, 200; proselytizing mass laity typology, archaeological 7–8, 25, 43;
3, 111; see also Sūtra of Trapusa and methodology of 4
Bhallika
Ti-wei cult 3, 123 uposatha 126
Tokiwa Daijō 5, 14n30, 64, 75n44 upper terrace landform 25
240 Index
usnisa 157 102–3, 173; patronage 38, 98;
Uttarasanga 52n79, 200 political affairs of 86, 106n23,
Uzbekistan 181 106n27; power struggle with
Xianwen 85–7, 89–90; two sages 95,
vahara (residential) caves 62 102, 173
Vanderstappen, Harrie 5, 17n55, western-end caves 4, 6, 8, 25, 158,
108n63 170–1, 174
vaulted passages 36, 39, 93, 110–11, western hill, Northern Park 86
125 Western Regions 4, 38–9, 42, 180
vernacular oral presentation (sujiang) Western Wei dynasty ix, 180
111, 115, 119, 126, 126; see also Western Zhou era 65
changdao wet clothes style 72
vernacular sūtra singing and preaching Wheel of Life 129n40, 151
by clergy (sengren changdao) Williams, Joanna 46
116–18, 126, 195 woman who hated desire and became a
viewers’ response to narratives 114–16 nun 69, 113–14, 130n63
Vignato, Giuseppe 9, 118 Wong, Dorothy 10, 98
vihāra caves 10, 27, 133, 224n48 wooden structures 32, 44, 73n11, 80,
Vimalakirti 45, 46, 100, 120, 143, 154, 83, 102, 132; caves 9 and 10 pp31, 33
166–7, 172, 178, 180 worshippers, images of 34, 80, 85, 87,
Vimalakirti and Manjusri 45, 46, 100, 105n19, 107n49, 156; lower legs
120, 143, 154, 166–7, 172 only 54, 87
Vimalakirti Sūtra 120, 172 wrestling contest 150
vinaya 42–3, 68, 118 Wu Daozi 179, 181
Virtue Storehouse 70, 76n88 Wudi, Emperor 176
visual narratives see narratives, wudianding (hip roof) 50n49, 200
pictorials Wuguantun caves 27, 200
Vitarka mudra 113 Wuhua caves see Wuhuadong
Vulture Peak 25, 65; see also Mount Wuhuadong (caves 9 to 13) 28, 33, 36,
Grdhrakuta 62, 82
vultures 153, 185n34, 224n48 Wuhuan people 55
Wu Hung 144, 184n20
Wanfosi 200 wu jie jing see Tiwei Boli Jing
Wanfo Temple 179 Wuping era 180
Wang Qingshi see Qian’er Qingshi wushen day 86, 109n74, 200
Wang Yu see Qian’er Qingshi Wu style 179
Wan Shen 41, 200 Wuwei see Liangzhou
wealth and longevity see fugui wansui Wuzhou Fortress 25, 60
Wei era 177 Wuzhou Mountain see Mount Wuzhou
Weishu 1–3, 6, 10, 25, 38, 49n28, 53, Wuzhou River 1, 26, 37, 49n43, 62
59–62, 74n34, 77n95, 86–8, 98, Wuzhoushan cave temple 99, 103,
106n27, 106n30, 124, 173, 200; “A 109n74, 200
Treatise on Buddhism and Daoism” Wuzhoushan shikusi (Yungang) 86, 88
chapter 2, 21
Weizi caves 43, 200 Xianbei people 54–6, 174
Wencheng, Emperor (Tuoba Jun) ix, 2, Xiang Da 120
9, 22, 26, 37, 53–4, 63, 64–6, 68, Xiangtangshan caves 5, 177, 182
70, 85, 88, 96, 107n55, 172; black Xianwen, Emperor ix, 9, 11, 17n52,
pebbles on a stone Buddha 22, 53; 49n28, 63, 68, 85–9, 103, 106n27;
era 88 era 64, 85, 88, 95, 106n34, 107n55,
Wenchengdi see Wencheng 189; imperial garments 85; patronage
Wenming, Empress Dowager (Lady of 87; patronage of allies 85; power
Feng) 9, 38, 66, 85, 95, 97, 98–9, struggle with Wenming 85–7, 89
 Index 241
Xianwendi see Xianwen yishe 14n22, 129n27, 200
Xianzu see Xianwen Yiweijing 130n65, 200
Xianyu jing (the Sūtra of the Wise and yiyi 4–5, 14n22, 116, 121–3, 126,
the Foolish) 152, 186n45 129n27
Xiaomingdi 174 Yongming era 174, 200
Xiaowen, Emperor (r. 471–499) ix, Yongning Temple 88
9, 17n52, 26, 34, 37–8, 64, 66, 68, Yongxing era 26
85–7, 90, 95–7, 98–9, 102, 107n55, Yoshimura Rei 6, 63, 67, 109n79
118, 173, 174, 182, 189; two sages youths born by spontaneous
95, 102, 173; see also Gaozu transformation see aupapadukas
Xiaowendi see Xiaowen yuan (cloister) 21, 200
Xiaowen era 36, 90, 94, 95–6, 99, 103 Yuanguang era 25, 200
Xiaowu, Emperor ix, 59 Yuanzhao 202
Xiaoyan 175 Yungang: archaeological excavations
Xie He 178, 180, 200 above caves 7, 10–11, 23, 25, 27,
Xijing zhi 13n11, 200 28, 29–31, 33, 36–7, 49n43, 62, 73,
Xindu 85 104, 110, 189–90; archaeological
Xing’an era 7, 37, 59, 200 excavations in front of caves 14n23,
Xingguang era 40, 53, 200 26, 28–30, 31–2, 50n49, 57, 64–5,
Xinjiang 25, 36, 38–9, 56, 110; see also 83, 105n11, 105n14; architectural
Turfan structure 3–4, 6, 8–10, 12, 38–9,
xinyou day 98, 200 40, 43, 56, 58, 63, 72, 100, 104,
Xiong Mengxiang 13n11, 200 106n34, 111, 118–19, 126, 133–40,
Xiongnu 25 158, 176, 190, 193–4; as Buddhist
Xishanqiao 178, 200 Mecca 132; colossal images 11–12,
xiugu qingxiang (graceful and refined) 22–3, 39, 43, 43, 47, 53–4, 64, 68,
images 47, 176 70, 72, 80, 81, 84, 88, 97, 100,
Xuangao 3, 23, 68, 200 103, 110, 172, 177, 189, 191;
Xuanwu, Emperor ix, 62, 173–4; components of 37, 48, 110, 189;
see also Shizong dating 3–4, 6–9, 11, 59–67, 75n49,
Xuezhuang 188n79, 200 78–90, 93–101, 103, 106n33,
Xu gaoseng zhuan (Further Biographies 107n44, 107n56, 108n59, 108n63,
of Eminent Monks) 2–3, 27, 37, 62, 189; function of caves 10–12, 21,
79, 132 25, 28, 33, 36–7, 39, 43, 48, 100,
Xu Song 188n84, 200 104, 110–12, 116, 122–4, 126, 190,
Xu Xianxiu 180, 200 186n54; geographical environment
Xuzhou eminent monks 118 1, 3, 25–7; inscriptions 2–5, 7,
9–10, 12, 14n20, 14n23, 16n48,
Yanbei shiyuan tombs 40–1, 55, 200 18n65, 25–7, 58, 81–2, 84, 97, 116,
Yanchang era 137, 155, 161, 163, 123, 131n76, 132, 137, 156, 161–2,
166–7, 170–1, 194, 200 163, 170, 174, 190; patronage 1,
Yang Xuanzhi 176 4, 7, 9–12, 34, 38, 66, 69, 72, 80,
Yangtoushan caves 187n54, 200 85, 87, 96, 102, 132, 157, 170–2,
Yangtze 175 189–90
Yanxing era 85, 87, 94, 95, 130n63, Yungang bu 1, 200; see also Yungang
179, 200 Fortress
Yan Wenru 6, 60–1 Yungang Fortress 1, 30, 200; see also
yaogu 40, 200 Yungang bu
Yao He 119 Yungang moshi 16n48, 189, 200;
Yasagupta 38 see also Yungang style
Yijing 130n48, 200 Yungang style (Yungang moshi) 4, 12,
yihui 121 50n43, 173–4, 189, 191, 200
Yihun 85–6, 200 Yunzhongtu 88, 200
242 Index
Zabaozang jing (Sūtra of the Zhang Yong image 42–3, 52n83, 194, 201
Miscellaneous Treasures) 69, 94–5, Zhao Mu system 65–7
121, 123, 130n63, 153, 224n47, 201 Zhengguang era 9, 66, 132, 170–1, 201
Zanning 21, 119 Zhengshi era 137, 161, 163, 166, 171,
zaoshi kaishan yinyan jiego 3 193–4, 201
Zengxiu jiaoyuan qinggui (Expanded Zhenguan era 31, 201
rule for teachings monasteries) 42, Zhenguo cave temple 16n40, 16n44,
201 51n57, 96, 201
zhang 50n52, 59 Zhongshan 60
Zhang Gui 39, 201 Zhou Fang 179
Zhang Huaiguan 174, 182, 201 Zhou style 179
Zhang Sengyou 18n64, 41, 140n63, Zhu Yi 175, 201
176–7, 179–81, 201 Zuochan sanmei jing (The Scripture on
Zhang style 179 Concentration in Sitting Meditation)
Zhang Yanyuan 178–80 23, 201