Exploring Multilingualism and Multiscriptism in Written Artefacts

Studies in Manuscript Cultures

Edited by
Imre Galambos, Konrad Hirschler, Caroline Macé,
Cécile Michel, Jörg B. Quenzer and Eva Wilden

Volume 38
Exploring
Multilingualism and
Multiscriptism in
Written Artefacts
Edited by
Szilvia Sövegjártó and Márton Vér
The publication of this volume was funded by the Deutsche Forschungsgemeinschaft (DFG, German
Research Foundation) under Germany’s Excellence Strategy – EXC 2176 ‘Understanding Written
Artefacts: Material, Interaction and Transmission in Manuscript Cultures’ – project no. 390893796.

ISBN 978-3-11-138048-3
e-ISBN (PDF) 978-3-11-138054-4
e-ISBN (EPUB) 978-3-11-138105-3
ISSN 2365-9696
DOI https://doi.org/10.1515/9783111380544

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0

International License. For details go to https://creativecommons.org/licenses/by-nc-nd/4.0/.
Creative Commons license terms for re-use do not apply to any content (such as graphs, figures,
photos, excerpts, etc.) not original to the Open Access publication and further permission may be
required from the rights holder. The obligation to research and clear permission lies solely with
the party re-using the material.

Library of Congress Control Number: 2024933102

Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie;
detailed bibliographic data are available on the internet at http://dnb.dnb.de.

© 2024 with the author(s), editing © 2024 Szilvia Sövegjártó and Márton Vér, published by
Walter de Gruyter GmbH, Berlin/Boston.
This book is published with open access at www.degruyter.com.

Printing and binding: CPI books GmbH, Leck

www.degruyter.com
Contents
Szilvia Sövegjártó, Márton Vér
Introduction | 1

Practical Multilingualism

Jost Gippert
Written Artefacts from the Maldives: 1,500 Years of Mixing Languages and
Scripts | 13

Leah Mascia
A Journey Through the Multilingual Landscape of Greco-Roman and Late Antique
Oxyrhynchus | 41

Jochen Hermann Vennebusch

Commemoration, Explication, and Obligation: The Baptismal Font of St Reinoldi in
Dortmund | 75

Apiradee Techasiriwan, Volker Grabowsky

Multi-scriptural and Multilingual Inscriptions in Lan Na | 111

Scholarly Multilingualism

Szilvia Sövegjártó
On the Emergence and Development of Middle Babylonian Bilinguals | 151

Antonio Manieri
Vernacular Terms in Sinitic Texts: Multilingualism in Eighth-century Japanese
Documents | 169

Emmanuel Francis
Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 205

Peera Panarut, Manasicha Akepiyapornchai

Sanskrit Prayers in a Theravada Kingdom: A Multilingual Siamese Grantha
Manuscript from Munich | 235
VI | Contents

Dmitry Bondarev, Darya Ogorodnikova

How to Spell Loanwords? Integration of Arabic Etymons in Bilingual Islamic
Manuscripts of West Africa | 255

Propagandistic Multilingualism

Gábor Zólyomi
Mesopotamian Bilingual Royal Inscriptions from the Third Millennium BCE:
Texts with a Primary and Secondary Context | 309

Zsolt Simon
Anatolian Theonyms in the Aramaic Version of the Letoon Trilingual | 331

Viola Allegranzi
Multilingualism in the Epigraphic Culture of the Persianate World (Eleventh to
Thirteenth Century) | 347

Eva Orthmann
Bilingual Inscriptions from India: Combining Arabic and Persian with Indic
Languages | 371

Contributors | 439

General Index | 443

Szilvia Sövegjártó, Márton Vér
Introduction
Multilingualism is an intricate concept that reflects the interconnected nature of
peoples and languages. Languages are not merely layered but internally com-
plex and geographically diverse. They exist within a web of interactions and are
shaped by diverse influences over time. The enduring traces of spoken lan-
guages, manifested in written artefacts, bear witness to the vibrant tapestry of
human communication and cultural exchange that has profoundly influenced
societies throughout history. These written artefacts serve as valuable testa-
ments to the multifaceted essence of language, encapsulating the thoughts,
ideas, stories and experiences of countless individuals and communities across
time and space. They provide a glimpse into the evolution, adaptation and in-
terplay of languages, offering valuable insights into the past while also paving
the way for future linguistic developments.
Multiscriptual written artefacts also highlight the rich diversity and com-
plexity of human languages. Languages often manifest themselves through
multiple scripts, each with its own unique set of characters and symbols.
Multiscriptual written artefacts encapsulate the historical trajectory of nu-
merous influences and cultural encounters, providing tangible evidence of
the dynamic exchange of ideas and knowledge across societies. These arte-
facts bear witness to the multiplicity of scribal traditions and the vast array of
human experiences they represent. They serve as tangible reminders of the
enduring power of scripts and writing systems, transcending time and space
to communicate the thoughts, stories and heritage of diverse cultures.
Through their existence, multiscriptual written artefacts offer valuable in-
sights into the development, adjustment and mutual influences of scripts,
illuminating the path towards further linguistic exploration and cultural
appreciation.
Multilingual and multiscriptual written artefacts serve as reflective mirrors
of a vibrant and diverse society that embraces multiculturalism and linguistic
plurality. These artefacts not only provide evidence of the coexistence of multi-
ple languages and scripts but also bear witness to the innovative ways in which
individuals navigate elements that may be foreign to a particular language or
writing system but familiar to its users. The interplay of various languages and
scripts within written artefacts necessitates a contextual understanding of the
social and cultural practices of those who created them. Exploring the practices
of scribes proves especially enlightening when studying written artefacts pro-

Open Access. © 2024 the authors, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-001
2 | Szilvia Sövegjártó, Márton Vér

duced by multilingual and multicultural communities or in regions where dis-

tinct writing cultures intersect. We can uncover valuable insights into the dy-
namics of linguistic and cultural exchange by delving into the mapping of
scribal practices, shedding light on the complex relationships and interactions
that shape these diverse written expressions. Such investigations provide a
fruitful avenue for unravelling the rich tapestry of multilingual and multiscrip-
tual societies and deepening our appreciation of the nuanced complexities of
language and writing systems.
Despite the widespread presence of corresponding written artefacts in al-
most all societies, multilingualism and multiscriptism are often considered
exceptional phenomena when compared to the more prevalent monolingualism
or monoscriptism. The present volume addresses this and other fundamental
matters, aiming to broaden horizons by establishing connections and under-
standing the underlying coherences that extend beyond the scope of case stud-
ies focusing solely on individual written artefacts.
The studies included in this volume span a wide chronological range, from
the third millennium BCE to the present day, and encompass diverse regions
across Asia, Africa and Europe. Rather than focusing solely on individual writ-
ten artefacts, the articles concentrate on the cultures of writing and their contri-
butions to the realm of multilingual manuscripts and inscriptions. By adopting
this broader perspective, the volume seeks to provide a comprehensive under-
standing of the complexities inherent in multilingualism, even within the realm
of writing.
However, it is important to note that the collection of articles represents
mere snapshots, offering glimpses into the multifaceted nature of multilingual-
ism as it manifests in writing. The intention is to showcase the complexity and
richness of multilingualism and multiscriptism, highlighting its varied forms
and manifestations throughout history. This volume strives to shed light on the
interplay between languages and scripts by examining the cultures of writing
and their impact, offering valuable insights into the dynamic nature of linguis-
tic expression in diverse societies.
Multilingual written artefacts can be classified by examining the various
multilingual scenarios they represent. Broadly speaking, these artefacts can be
categorised based on the underlying purposes of their creation: practical,
scholarly or propagandistic multilingualism.1 The papers in this volume are

||
1 Proposed by Peter Toth: https://www.bl.uk/greek-manuscripts/articles/multilingualism-in-
greek-manuscripts (accessed on 22 August 2023).
 Introduction | 3

organised into these three groups, aiming to contribute to the study of multilin-
gualism and multiscriptualism within these distinct domains.
In the case of practical multilingualism, written artefacts incidentally
contain texts in multiple languages, serving practical aims in a multicultural
environment. The study of practical multilingualism is initiated by Jost Gip-
pert’s comprehensive overview of multilingual and multiscriptual written
artefacts from the Maldives. The paper sheds light on the interplay between
the local Dhivehi language and other languages, such as Sanskrit, Prakrit,
Persian, Arabic and even English, within a dynamic multicultural environ-
ment. Various inscriptions, paper documents, inscribed metal plates, and
charters on wood and stone are presented to illustrate the diverse manifesta-
tions of multilingualism and multiscriptism, spanning from the sixth century
up to the present day. Another aspect that provides insight into the practical
aspects of written language intertwining is the eighteenth-century technical
innovation of replacing the local Dives akuru script with the right-to-left
directed Thaana script, which resolved the challenge of incorporating Arabic
elements into Dhivehi texts.
Leah Mascia takes us on a captivating journey through the intricately in-
scribed multilingual landscape of Oxyrhynchus. This town’s social and reli-
gious panorama was moulded by the presence of Egyptians, Romans, Jews
and Nabataeans during the Greco-Roman and late antique periods. Although
this paper maintains a focused approach in terms of space and time, it pro-
vides a comprehensive and illuminating review of multilingualism and mul-
tiscriptism within the realms of administration and religion, while also delv-
ing into various aspects of the funerary landscape. The author adeptly
portrays how the coexistence and constant interaction between native and
foreign inhabitants contributed to the shaping of the settlement’s cultural
horizon.
Jochen Hermann Vennebusch’s paper presents a compelling case study
centred on the bronze baptismal font in the church of St Reinoldi in Dortmund.
This remarkable font bears the preservation of two Latin inscriptions and a Low
German vernacular inscription, each tailored to address distinct audiences and
vividly showcasing the diversity within the medieval church. Additionally,
these inscriptions potentially represent a case of limited scriptural presence,
emphasising the importance of their mere existence on the object rather than
their readability or comprehension.
In the concluding contribution to the exploration of practical multilingual-
ism, Apiradee Techasiriwan and Volker Grabowsky delve deeply into a thought-
provoking exploration of the manifestation of various scripts and languages
4 | Szilvia Sövegjártó, Márton Vér

within the epigraphical culture of the Lan Na region in Thailand. The authors
shed light on the rich cultural history of the region by presenting compelling
examples of script amalgamation within the Lan Na writing culture, as well as
the intriguing interaction between neighbouring writing cultures in the Thai
world and more distant writing cultures, such as Chinese and English. Addi-
tionally, they make a valuable contribution by providing an edition of previous-
ly unpublished multilingual and multiscriptual inscriptions, which have not yet
been made accessible to English-speaking audiences.
The focus of contributions in the second section of the volume shifts to
scholarly multilingualism which gave rise to written artefacts intended for
scholarly or educational purposes. Textbooks designed for language teaching,
as well as bilingual or multilingual glossaries and dictionaries, exemplify this
category. However, the realm of scholarly multilingual manuscripts is much
broader and more diverse. It encompasses manuscripts that employ established
termini technici from a different language, as well as those that include glosses,
commentaries or additions in other languages or vernaculars.
The focus in Szilvia Sövegjártó’s article is on Middle Babylonian Sumeri-
an literary manuscripts, often accompanied by Akkadian glosses or transla-
tions. These manuscripts blur the line between the Old and Middle Babyloni-
an periods due to their hybrid palaeography. While conventionally ascribed
to the distinct practices of Middle Babylonian scribes, Sövegjártó proposes an
alternative hypothesis: the influence of Old Babylonian manuscripts on the
stylistic features of Middle Babylonian texts. This proposition re-evaluates
previous assumptions regarding the transition from glossed manuscripts to
bilingual ones, emphasising a diachronic viewpoint. Furthermore, Sövegjár-
tó’s research highlights a noteworthy shift in the practices of Middle Babylo-
nian scribes – the preservation of glosses and translations. This shift reflects
the changing significance of the Sumerian language and its literary heritage
during this period.
Antonio Manieri presents case studies examining the technical and artisan-
al terminologies of ancient Japan, with the aim of providing insights into these
terms and the underlying concepts within the realms of craftsmanship and
technology. The author argues that, despite the established conventions favour-
ing Sinitic languages in document compilation, there are instances where spe-
cialized vernacular terms are utilised, disregarding the presence of equivalent
Sinitic terms. Interestingly, the use of these vernacular terms does not stem
from a distinct Japanese cultural specificity that necessitates the use of phono-
grams, such as culturemes or proper names. The paper concludes that the
blending of bureaucratic Sinitic and artisanal vernacular in the texts arises from
 Introduction | 5

the interaction between two distinct knowledge systems: the Chinese scholar-
ship and literary tradition, on the one hand, and the local practices transmitted,
learned and disseminated orally and experientially through work experience,
on the other.
Emmanuel Francis’s paper draws attention to an intriguing observation
regarding Tamil inscriptions, where Sanskrit loanwords are frequently writ-
ten using the specific Grantha alphabet. This phenomenon highlights the
historical interplay between Sanskrit and Tamil in South India. The paper
also examines Tamil manuscripts with novel script-mixing practices, notably
the creation of conjunct graphemes that blend elements from both the Gran-
tha and Tamil alphabets. Such developments emphasise the dynamic nature
of script evolution and the cultural exchange within the region. Furthermore,
the paper stresses the importance of incorporating these findings into edito-
rial practices within the realm of digital humanities. It suggests that not only
should the phenomena of language and script mixing be recorded in the
metadata, but it is also crucial to encode this information in the edited texts
themselves. This integration of information aligns with the objectives of the
‘The Domestication of “Hindu” Asceticism and the Religious Making of South
and Southeast Asia’ (DHARMA) project, ensuring a comprehensive under-
standing of the linguistic and scriptural complexities present in Tamil in-
scriptions and manuscripts.2
Peera Panarut and Manasicha Akepiyapornchai present a case study on a
previously unpublished Siamese Grantha manuscript, which sheds light on the
use of both the Sanskrit and Thai languages in the predominantly Buddhist
culture of Thailand. The manuscript, containing Sanskrit mantras and Thai
ritual instructions, challenges the notion that the Theravada culture of Siam
was solely dominated by Thai and Pali. It reveals that the Sanskrit and the
Brahmanical tradition were also part of the cultural profile, showcasing the
coexistence of multiple linguistic and religious traditions in Siamese society.
This discovery highlights the cultural diversity and influences present in the
region, providing valuable insights into the historical dynamics of Siamese
culture.
Finally, Dmitry Bondarev and Darya Ogorodnikova’s article investigates the
integration of Arabic loanwords into West African languages through an analy-
sis of spelling patterns found in two distinct manuscript traditions: Old Kanem-
bu and Soninke. The study presents a preliminary typology of these spellings,
revealing that orthographic practices within educational circles play a crucial

||
2 https://dharma.hypotheses.org/project (accessed on 18 September 2023).
6 | Szilvia Sövegjártó, Márton Vér

role, overshadowing scribes’ individual preferences. The study emphasises that

relying solely on spelling as a diagnostic tool falls short in assessing loanword
integration. However, when considered in conjunction with other linguistic and
cultural factors, Ajami loanwords offer valuable insights into the chronology
and pathways of borrowing. Ajami data should be approached cautiously, as
they often reflect learned environments with distinct lexicons, influenced by
established conventions within specific circles of scholars. This complexity
sheds light on the diverse history of Arabic loanwords in West African lan-
guages, previously unrecognised as a significant factor in their multiple forms.
The case studies in the third section refer to propagandistic multilingualism
which is more restricted to the realm of inscriptions. It involves the production
of written artefacts in a multilingual format with a polemic or propagandistic
agenda, which can be either concealed or overt.
In the initial paper, Gábor Zólyomi explores bilingual inscriptions that have
been transmitted in Sumerian and Akkadian versions, albeit solely preserved in
later copies and not in their original form. The author skilfully reconstructs the
distinctive characteristics of these inscriptions, in both their primary and sec-
ondary contexts, and asserts that these bilingual inscriptions were initially
composed in Akkadian. The author demonstrates significantly that the use of
bilingual inscriptions was driven by ideological motives, as the ruler aimed to
portray himself as the legitimate authority over both parts of Babylonia, adroitly
manipulating the medium to suit the intended audience.
Zsolt Simon’s case study investigates the topic of the formal adaptation of
local theonyms in the Aramaic version of the decree called the Letoon Trilin-
gual. The composer of the Aramaic version faced several challenges during this
process, leading to the implementation of diverse solutions depending on the
specific divinity being referenced. When Old Persian equivalents were available,
they opted to substitute them. In other instances, the names were omitted, and
blanket references were used instead. However, this approach was not viable for
significant deities, resulting in the decision to transcribe the original names.
These strategic choices were employed to maintain the cohesiveness of the orig-
inal version of the decree.
Viola Allegranzi’s paper explores language usage in the epigraphic cul-
ture of the Persianate world during the eleventh to thirteenth centuries. Fo-
cusing on a range of monumental inscriptions and inscribed objects, she
presents case studies that combine Arabic and Persian, examining the evolu-
tion of linguistic and stylistic choices in the absence of a reference model for
Persian inscriptions. The paper investigates whether language-related deci-
sions were influenced by the patron, scribe or recipient of the inscription,
 Introduction | 7

aiming to provide a contextual understanding of multilingual inscriptions

within the cultural production of medieval Persianate societies. The presence
of Persian poetic inscriptions on both royal and privately commissioned
monuments and objects is highlighted. Additionally, Persian’s usage in con-
struction texts reveals a greater influence from Arabic models. Allegranzi
ultimately concludes that the distinctive features of Islamic epigraphy in the
Persianate world lie in its multilingualism, as well as the development of new
writing styles and decorative elements.
Eva Orthmann’s article delves into the world of bilingual inscriptions in
Arabic, Persian and various Indic languages found in the Indian subconti-
nent. These inscriptions, discovered in a diverse range of locations such as
mosques, epitaphs, walls and copper plates, offer a valuable glimpse into the
era of Muslim rule in India. The author presents a comprehensive overview of
this unique corpus in this article, shedding light on how these languages
were skilfully combined to create visually striking effects and express the
complex hierarchy between them. Unlike mere translations, these bilingual
inscriptions often present two distinct versions of a text, each adhering to its
own stylistic norms. The study suggests that the production of these inscrip-
tions may have involved expert scribes fluent in both languages, who were
provided with clear content guidelines and entrusted with the task of crafting
culturally appropriate texts, resulting in a rich tapestry of linguistic and cul-
tural interplay.
By considering these classifications, we gain a deeper understanding of the
multilingual nature of written artefacts, shedding light on the practical, scholar-
ly and propagandistic contexts in which they were created. Such investigations
contribute to the broader study of multilingualism and multiscriptualism, en-
riching our knowledge of language dynamics and cultural expressions within
these distinct domains.
The present volume delves into the exploration of written artefacts that ex-
hibit the distinctive use of multiple languages and scripts. The papers within
this collection focus specifically on the interaction between languages and writ-
ing systems found in both manuscripts and inscriptions across various do-
mains, including everyday life, scholarship and political propaganda. By adopt-
ing a cross-disciplinary approach, this volume aims to shed light on a wide
range of phenomena encompassed by the comprehensive terms of multilingual-
ism and multiscriptism.
The papers included in this volume embrace an interdisciplinary perspec-
tive, incorporating disciplines such as manuscript studies, epigraphy, textual
criticism, linguistics, palaeography and cultural history. This diverse approach
8 | Szilvia Sövegjártó, Márton Vér

enables a comprehensive examination of the subject matter, drawing on the

collective expertise of various fields. The volume seeks to transcend traditional
conceptual boundaries and encourage a deeper understanding of the dynamics
at play in multilingual and multiscriptual written artefacts by undertaking a
comparative study of inscriptions and manuscripts spanning ancient, premod-
ern and modern times.
This volume aims to contribute to the broader discourse surrounding multi-
lingualism and multiscriptism through the exploration of these written artefacts
and the multifaceted methodologies employed. It seeks to foster new insights,
bridge gaps in understanding and offer a fresh perspective on the complex in-
terplay between languages and writing systems throughout history by bringing
together scholars from diverse disciplines.

Acknowledgements

The research for this note was funded by the Deutsche Forschungsgemein-
schaft (DFG, German Research Foundation) under Germany’s Excellence
Strategy, EXC 2176 ‘Understanding Written Artefacts: Material, Interaction
and Transmission in Manuscript Cultures’, project no. 390893796, at Univer-
sität Hamburg (2019–2025). The research was conducted within the scope of
the Centre for the Study of Manuscript Cultures (CSMC), at Universität Ham-
burg. The topic was first discussed in Research Field B (Inscribing Spaces)
and D (Formatting Contents), where Kaja Harter-Uibopuu and Eva Wilden
developed the idea of a workshop. The event, ‘Mixing Languages and Scripts:
Material from Manuscripts and Inscriptions’, took place in a hybrid form at
the CSMC on 19–21 May 2022.3 The peer reviewers of the articles that appear
in the volume are warmly thanked for their generous and constructive con-
tributions: Ignasi-Xavier Adiego, Paola Buzi, Giovanni Ciotti, Steffen Döll,
Michael Feener, Jost Gippert, Volker Grabowsky, Peera Panarut, Jürgen Paul,
Lameen Souag, Iris Wendelholm, Eva Wilden, Kamran Vincent Zand, and

||
3 On that occasion, eighteen speakers presented seventeen papers, most of them are authors
of papers presented in this volume. Further speakers whose contributions should be acknowl-
edged here are Sara Chiarini (UHH), ‘Polyglossia in the Ancient Western Mediterranean’; Tibor
Porció (University of Szeged), ‘Uyghur Buddhist Pilgrim Inscriptions Revisited’; Ishayahu
Landa (University of Bonn), ‘Multilingualism and Multiscriptism in Mongol Eurasia’s Numis-
matics’; Grzegorz Ochała (University of Warsaw / Leiden University), ‘One People, One Script,
Three Languages: Multilingual Experience of Medieval Nubia’; Erin McCann (UHH), ‘Multilin-
gual Manuscripts in a Multilingual Tradition: The Case of Śrīvaiṣṇava Manipravalam’; and
Polina Yaroslavtseva (UHH), ‘Multilingualism in BnF, Ms Fr. 375’.
 Introduction | 9

Elyze Zomer. Warm thanks also go to the two English copy-editors, Philip
Saunders and Kristen de Joseph, for their patient and competent effort of
putting the thoughts of the authors into correct English. The same thanks go
to Francesca Panini for conscientiously taking care of the typesetting and
laying out this quite complex and demanding volume as well as for preparing
the index. Lastly, we wish to acknowledge Caroline Macé for her meticulous
editorial guidance, which has greatly influenced the development of this
volume. Additionally, we extend our gratitude to all the authors for their
unwavering commitment in contributing to this project.
|
Practical Multilingualism
Jost Gippert
Written Artefacts from the Maldives:
1,500 Years of Mixing Languages and Scripts
Abstract: The archipelago of the Maldives, which extends over c. 1,000 km in a
north-south direction in the Indian Ocean, looks back on an uninterrupted pro-
duction of written artefacts since about the sixth century CE. In their entirety,
the monuments that have been preserved give a clear picture of the historical
development of both the local language, Dhivehi (lit. ‘Islanders’ [language]’),
and the scripts that were used to write it down, as well as the religious and cul-
tural conditions in which the artefacts emerged (Buddhist vs Islamic). At the
same time, the religious and cultural setting, which changed considerably with
the conversion of the islanders from Buddhism to Islam in the middle of the
twelfth century CE, caused a specific interaction of languages and scripts that
manifested itself in a remarkable amount of mixing, with more or less impact on
the graphical appearance of the artefacts. In the present article, I illustrate the
interplay of Dhivehi with the other languages concerned (Sanskrit, Prakrit,
Persian, Arabic and, lastly, English) and its reflection in written form based on a
few examples.

1 The Buddhist pre-stage: Prakrit vs Sanskrit

The culture of the Maldives was predominantly determined by Buddhism before
their inhabitants were converted to Islam in the middle of the twelfth century
CE, a fact that is witnessed by a great bulk of archaeological findings throughout
the archipelago.1 Only a very few written artefacts have been preserved from
those times which, however, display a principle that would remain typical for
Maldivian literacy throughout the centuries, namely, the mixing of languages
and, later, scripts.

||
1 For a survey of the remnants of Buddhism in the Maldives, see Gippert forthcoming a.

Open Access. © 2024 the author, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-002
14 | Jost Gippert

1.1 The Landhoo inscription

The oldest autochthonous written artefact that has been unearthed on the Mal-
dives2 is a brick-shaped block of coral stone detected on the island of Landhoo,
which is inscribed on its four longitudinal faces in a Brāhmī script similar to the
varieties used in southern India in about the sixth to eighth centuries. Its con-
tents have been determined to be a dhāraṇī spell, i.e. a series of mantras pro-
tecting against demons; the stone probably pertained to a Buddhist monastery,
not as part of its building construction but as a relic that was enshrined in its
stupa.3 In contrast to comparable dhāraṇī spells, such as the Sitātapatrādhāraṇī,
the language it is written in is not just (Buddhist) Sanskrit; instead, it is a mix-
ture of Sanskrit with a peculiar form of ‘Insular’ Prakrit, namely, the Middle
Indic vernacular that is likely to be the historical ancestor of Dhivehi. In the
excerpt of face 1 (Fig. 1) displayed in Table 1, the Sanskrit elements are indicated
in red and the Prakrit elements in blue; the words in black in the last line are so-
called bīja syllables, i.e. interjections that are believed to have magic power in
themselves but are not necessarily attributable to one of the two languages. As
the Table shows, it is the element grāha meaning ‘seizing’ or ‘possession’ that is
usually spelt in its Sanskrit form (the Insular Prakrit equivalent would have
been *gahǝ or *gā), whereas the formula for ‘I smash into pieces’, toṭa(ṃ)
bhidāmi, is rather Prakrit (vs Sanskrit *troṭayaṃ bhindāmi), as are most of the
names of demons, even when compounded with grāha. The name preta denot-
ing the evil spirit of a dead person (the Prakrit form would be *peyǝ), nāga for
‘snake’ (vs Prakrit *nā) and bhūta (spelt bhuta), a designation of another type of
demon (vs Prakrit *bu(v)ǝ) are exceptions to this. Interestingly enough, the
latter term contrasts with bhui-, which reflects, in its Prakrit form, the female
equivalent of the same demon, in Sanskrit named bhūtī-; and of all demons men-
tioned here, only this one has survived into modern Dhivehi as a suffix in the
name of a fairy, Santi Mariyam-bu, which obviously reflects the Christian Saint
Mary.4 Sanskrit forms are also svāha, the formulaic greeting ‘Hail!’ (also occur-
ring in the spelling svaha in the inscription), which would be *sahǝ (or simply

||
2 We leave aside here Roman and other coins of Antiquity that must have been imported by
travellers (see, e.g. Forbes 1984).
3 See Gippert 2004 for the first edition of the inscription and further details. In the present
treatise, personal and geographical names and other terms that are in everyday use are ren-
dered in the English-based Roman transcription for Dhivehi as it is used on the Maldives; other
linguistic elements are transcribed in the system used in Indology.
4 The name was probably introduced into the Maldives by the Portuguese conquerors in the
middle of the sixteenth century; see Gippert 2004, 93–94.
 Written Artefacts from the Maldives | 15

*sā) in Prakrit, and sarva (here spelt sarvva) ‘all’, whose Prakrit equivalent sava
(for *savǝ) appears elsewhere in the text (partly with the spelling savva).
Whether or not this variation is to be regarded as a case of true bilingualism or if
it was just an indication of uncertainty in rendering Sanskrit graphically by
speakers of the local Prakrit, must remain open; we should keep in mind, how-
ever, that Buddhist Sanskrit was notorious for being ‘hybrid’, in the sense of
being mixed with elements from spoken vernaculars and that in the given case,
Sanskrit grammar seems to be ignored at large (e.g. there are no accusative
endings).5

Table 1: The Landhoo inscription, face 1 (excerpt).

pisacchagrāha toṭa bhidāmi I smash the possession caused by piśācas into pieces.
vasmāra<grāha toṭa bhidāmi> I smash the possession caused by apasmāras into pieces.
<***grāha> toṭa bhidāmi I smash the possession caused by ??? into pieces.
bhutagrāha toṭaṃ bhidāmi I smash the possession caused by bhūtas into pieces.
bhuigrāh<a toṭa> bhidāmi I smash the possession caused by bhūtīs into pieces.
pretagrāha toṭa bhidāmi I smash the possession caused by pretas into pieces.
kālamaṭṭagrāha toṭaṃ bhidāmi I smash the possession caused by akālamṛtyu(s) into pieces.
<**>rakkusagrāha toṭaṃ bhidāmi I smash the possession caused by ??rakṣasas into pieces.
kummaṇḍagrāha toṭa bhidāmi I smash the possession caused by kumbhāṇḍas into pieces.
suvaṇṇagrāha toṭa bhidāmi I smash the possession caused by suparṇa(s) into pieces.
duṭṭanāgagrāha toṭaṃ bhidāmi I smash the possession caused by wicked nāgas into pieces.
sarvva toṭaṃ bhidāmi I smash all (of them) into pieces.
svāha Hail!
ili mili khili khili khili khili <****> gili g*li hili hili hi<li hi>li hili hili hili ||

||
5 See <https://sketchfab.com/3d-models/inscription-1-from-landhoo-maldives-ea6c56ddf41c4
320a5c76961bb2958ff> (accessed on 22 July 2023) for a 3D-model of the Landhoo stone, which is
registered as MAHS-MDV-COL-001-O-0033 in the database of the ‘Maritime Asia Heritage Sur-
vey’ project (hereafter MAHS; <https://maritimeasiaheritage.cseas.kyoto-u.ac.jp/manuscript-
viewer/>; Feener s.a.). A similarly inscribed artefact has recently been detected on the same
island (registered as MAHS-MDV-COL-029-O-0001 in MAHS); see <https://sketchfab.com/3d-
models/inscription-2-from-landhoo-maldives-5c7e84b41e274a7a9f676c67c9228af2> (accessed on 22
July 2023) for a 3D-model.
16 | Jost Gippert

Figs 1a–b: The Landhoo inscription, face 1; photograph and sketch Jost Gippert, 2003.

1.2 Inscribed statues

In contrast to the inscription from Landhoo, which, as we have seen, can be
regarded as being mostly in Prakrit, two inscribed artefacts from a few centuries
later (c. ninth to tenth centuries CE) must be considered as being basically com-
posed in Sanskrit. Both these artefacts are shaped as multifaced statues with
features that are typical for the so-called Vajrayāna Buddhism; their inscrip-
tions, which are applied across the surface, are written in a much more ‘cursive’
script than that of Landhoo, resembling that of medieval Sinhalese and, thus,
obviously representing an antecedent of the so-called Dives akuru script that
was used for Dhivehi until the eighteenth century.6 Even though the shape of
the two statues and the arrangement of the inscriptions is totally different (Figs

||
6 See Mohamed 1999 and 2004 and Gippert 2013 for details as to the script and its development.
 Written Artefacts from the Maldives | 17

2a–b), they have been proven to contain the same text, namely, a mantra ad-
dressing Yamāntaka, the ‘destroyer of death’ in Vajrayāna belief, which is also
known from other sources.7

Fig. 2a: Statue with Sanskrit inscriptions; photograph and markings Jost Gippert, 1993.

||
7 These are the Guhyasamājatantra (Chap. 14) and the Mañjuśrīmūlakalpa (Chaps 1 and 52); the
mantra also exists in Chinese and Tibetan translations. See Gippert 2013–2014 for further details.
18 | Jost Gippert

Fig. 2b: Statue with Sanskrit inscriptions; photograph and markings Jost Gippert, 1993.
 Written Artefacts from the Maldives | 19

In contrast to the Landhoo inscription, the Sanskrit wording is here aimed at

being at least grammatically correct (with adequately formed genitives, accusa-
tives and vocatives); however, the spoken vernacular shimmers through every-
where in the form of spelling inconsistencies that are caused by the reduced
phonological inventory of spoken Dhivehi. Sanskrit vyāghracarma- ‘tiger’s
skin’, for example, appears in both inscriptions as vyagrasarmma, the impera-
tive chinda ‘break!’ as sinda or ṣinda, and vajra, the ‘thunderbolt’, whose name
underlies the given Buddhist tradition, as vadra.8

2 Islamisation and the arrival of Persian and Arabic

Even after the conversion of the Maldives to Islam in the middle of the twelfth cen-
tury, the influence of Sanskrit on the islands remained notable. This can be well
illustrated with the first written artefacts that have been preserved from this time
(end of the twelfth century), namely, sets of copperplates documenting the kings’
endowments for newly constructed mosques. These documents (usually styled
lōmāfānu, i.e. ‘great excellency of copper’ in the Maldives) are generally written in
an early form of Dhivehi, in a script (sometimes called Evēla akuru)9 which clearly
continues that of the Buddhist statues. Regarding the self-introduction of the king,
however, they usually contain a seal that is written in a Nāgarī-type script, with the
language being Sanskrit (albeit with graphical distortions); this is true, for exam-
ple, for the so-called Isdū lōmāfānu, which was issued by a king named Ga-
ganāditya for the island of Isdhoo in Haddunmathi (or Laamu) Atoll in c. 1194 CE
(see Figs 3 and 4). The seal reads sva-sti śrī ga-gʰa-nā-dī-tʰtʰa de-va-s'-ya dā-nāṃ,
rendering svasti śrī gaganāditya devasya dānam ‘Hail! Grant of the brilliant God
(-like) Gaganāditya’. In a similar way, the king’s signature on the last plate of the
set was written in Nāgarī script, as visible in Fig. 5, where we read āṃ-dī-nta śrī ma-
hā-rā-c' as a rendering of āditya śrī mahārāja ‘Brilliant great-king Āditya’.

||
8 See Gippert 2013, 90–91, for the relation between Dhivehi phonology and Dives akuru script.
9 The term was coined ‘provisionally’ by Henry Bell (1940, 166); it means ‘script of that time’ or
‘of yore’.
20 | Jost Gippert

Fig. 3: First plate of the Isdū lōmāfānu with king’s seal; photograph Jost Gippert, 2002.

Fig. 4: King’s seal on the first plate of the Isdū lōmāfānu (enlarged); photograph Jost Gippert, 2002.

Fig. 5: King’s signature on the last plate of the Isdū lōmāfānu (line 1); photograph Jost Gippert, 2002.

However, the use of Sanskrit terms in these documents is not restricted to seals
and signatures. As a matter of fact, the grants abound in Sanskritisms, even
where they are written in the local script, thus, being indistinguishable from
Dhivehi words proper. This is not only true for most of the names mentioned in
 Written Artefacts from the Maldives | 21

them, at least those of kings, which, like gaganāditya ‘sun of the heaven’,10
reflect purely Sanskrit formations. The co-occurrence of Dhivehi and Sanskrit
elements is illustrated in Table 2, which shows the content of the first plate of
the Isdū lōmāfānu, with Sanskrit elements in red and those of Dhivehi in blue;
note that, in some cases, the Sanskrit genitive ending -sya is attached to a
Dhivehi word form (rasunu-sia and rasun-sya ‘of the king’, quasi Sanskrit
*rājanasya, with Dhivehi rasun- representing the historical outcome of the San-
skrit n-stem rājan-). Besides names, it is typically invocations and titles, but also
numbers, such as ekavińśatī (for ekaviṃśati) ‘twenty-one’ (the modern Dhivehi
equivalent would be ekāvīs, ‘twenty’ alone would be vihi)11 or panṣatrińśati (for
pañcatriṃśati) ‘thirty-five’ (vs Dhivehi tirīs fas) that are in Sanskrit.12

Table 2: Beginning of the Isdū lōmāfānu with Sanskrit elements in red.

svasti śrī ṣomavańśa ādipati Hail! When the lord of the brilliant Moon Dynasty,
śrī manābaruṇa mārasun the brilliant mahārāja Mānābharaṇa,
puṇa dese reda vuṇa dai after of his reigning over all the country
ekavińśatī avurodun ikit vī kalu had passed twenty one years,
reda opun side had left his reign,
mi rasunusia beni (and) this king’s younger brother,
svasti śrī tribuvana ādītya mārasun hail! the brilliant mahārāja Tribhuvanāditya,
reda vuṇa dʰai after of his reigning
panṣatriṅśati avurodun ikit vī kal had passed thirty-five years,
mi rasunsya malu … (and) this king’s nephew …

At the same time, the conversion of the islands introduced new terms that were
related to the Islamic faith and not readily available in either Dhivehi or San-
skrit. These are typically Arabic or Persian words concerning mosques and peo-
ple associated with them; on plate 13r of the Isdū lōmāfānu,13 for example, we

||
10 See Gippert 2003, 34 with n. 13.
11 See Fritz 2002, vol. 1, 114.
12 There is no established transliteration system for Old and Middle Dhivehi texts written in
Evēla and Dives akuru. The transcriptions provided here and in the following tables are simpli-
fied to warrant readability.
13 An excellent image of the plate can be found on the MAHS website (Feener s.a.) under the
designation MAHS-MDV-COL-001-O-0024 - Isdhoo Loamafaanu (MLE-NMM-MS1) as page 34
(image 35). Unfortunately, the images of the lōmāfānu plates are arranged in disorder there.
22 | Jost Gippert

find masudid(h)u ‘mosque’ representing the Arabic masǧid, mālimu ‘Qurʼan

teacher’ ~ the Arabic muʽallim, and mūdimu ‘muezzin’ ~ the Arabic muʼaḏḏin. As
these examples show, the Arabic terms were adapted to Dhivehi phonetically
and grammatically (e.g. by adding the dative ending aṭa as in masudidhaṭa ‘to
the mosque’) and could, thus, be written in the local script; their integration is
illustrated in Table 3 where the Arabic elements are marked in green.

Table 3: Text of plate 13 (recto) of the Isdū lōmāfānu with Arabic elements in green.

śrī isuduvu masudidu For the mosque of brilliant Isdū

ṣāna pangeāi reed for thatching the roof,
veṭṭelāi sadakā saḍulāi oil for the lamps, alms rice,
mālim mūdimun kana bogaāi the food share for the teacher (and) the muezzin,
mi emme kamakemmaṭa māruvai and measures for providing all these things
masudidʰaṭa dinu were allotted to the mosque.
mālimu iduna ge bimaṭa For the ground of the house of the teacher
vaṭai sia bamai koṭṭakāi a parcel of a hundred fathoms and a parcel of
vaṭai panasu bamai koṭṭakāi fifty fathoms and for the housing of the
mūdimu idinaṭa vaṭai sia bamai koṭṭāi muezzin, a parcel of a hundred fathoms
mi de dʰarunaṭa … dinu were allotted to these two … people.

On the other hand, even for religious concepts, Sanskrit terms that are likely to
have been associated with Buddhist thought previously continued to be used. A
striking example is provided by a second copperplate grant issued by King Ga-
ganāditya in 1194, the so-called Gamu lōmāfānu, which concerns a mosque on
Gan, another island of Haddunmati (or Laamu) Atoll. Unfortunately, this grant,
which comprises a lengthy account of the Creation and the life of Muḥammad,14
is not as well preserved as its counterpart from Isdhoo (see Figs 6a–b showing
the remnants of plates 2 and 3); however, the remarkable interplay of Sanskrit
words such as devatā ‘deity’ (also in ekadevatā- ‘one deity’), utthara svargga
ḷoka- (for uttara-svarga-loka) ‘the place of utmost heaven’, narakā- (for naraka-)
‘hell’, ādītya- (for āditya) ‘sun’, pṛtthivi loka (for pṛthivī loka-) ‘the place of earth’
or sind (for sindhu) ‘Indus, Panjab’ with Arabic terms such as rahimat- (for
raḥmat) ‘mercy’ and kautar (for kauṯar) ‘abundance’ (here denoting a river in

||
14 See Gippert 2003, 37–40 for details. Images of the Gamu lōmāfānu can be found on the
MAHS website under the designation MAHS-MDV-COL-001-O-0026 - Gan Loamafaanu (MLE-
NMM-MS3).
 Written Artefacts from the Maldives | 23

paradise),15 and names such as ṣaur ‘Syria’ (Arabic sūriya), erāk ‘Iraq’ (ʽIrāq),
bābhil ‘Babylon’ (bābil), pārīṣ ‘Persia’ (fāris) and dadhīrā ‘Mesopotamia’ (ǧazīra)
can be taken for certain. A peculiar trait of the Gamu lōmāfānu, not attested
elsewhere in historical documents of Dhivehi, are the listings of Islamic para-
dises (beginning with [dā]ralu salāmu for Arabic dār al-salām ‘House of Peace’)
and hells (beginning with sairu for Arabic saʽīr ‘flaming fire’), as illustrated in
Table 4. Note that only one of the geographical names involved, jabuduv for
Jambudvīpa ‘(southern) India’, appears in a local form (the modern Dhivehi
equivalent would be dаmbidū).16
The picture becomes even more colourful if we consider that some Islamic
terms are not of Arabic but of Persian origin. This is true, for example, of the
word for the ‘prophet’, petāmbaru, which obviously represents a dialectal vari-
ant of Persian pay(ġ)āmbar17 ‘messenger’ and which is used in the Gamu
lōmāfānu for both the founder of Islam, Muḥammad, and his (biblical) prede-
cessors; other Persian terms that occur regularly are roda ‘fasting’ (for Persian
rōza, today rūze) and namādu ‘prayer’ (Persian namāz), as well as a second
denomination of mosques, miskitu, which must reflect Early New Persian mazgit
‘id.’ and which substitutes masudid(h)u after the first copperplate grants.18 In
addition, the text contains a few elements that have retained a Prakrit shape.
These are higher numbers, such as eklakka ‘100,000’ (cf. Sanskrit eka-lakṣa),
ṣauvīsu ‘twenty-four’ (cf. Sanskrit catur-viṃśa(ti)), and ṣāḷīsu ‘forty’ (cf. Sanskrit
catvāriṃśa(ti)), contrasting with cāśra ‘1,000’ which probably represents San-
skrit sahasra- directly (vs Prakrit *sǝhassǝ > Dhivehi *sāsu > hās). Persian ele-
ments are indicated in violet and Prakrit ones in orange in Table 5.

Fig. 6a: Gamu lōmāfānu, plate 2; photograph Jost Gippert, 2002.

||
15 For this use see the Qurʼan, Surah 108.1.
16 See Gippert 2003 and 2015, for further details as to the Gamu lōmāfānu.
17 See Gippert 2003, 42, for details.
18 See Gippert 2003, 41–42 and forthcoming b for further details.
24 | Jost Gippert

Fig. 6b: Gamu lōmāfānu, plate 3; photograph Jost Gippert, 2002.

Table 4: Gamu lōmāfānu, plate 2 (excerpt).

… ekadevatāinge rahimatun upe[dduvi …] … [crea]ted by the mercy of the one deity …

… devatāinaṭa ga]t birun …] … [by fear gai]ned towards the deity …
… pen ve kekuṇi dumu seduṇu … … made of steam emerging from boiled water …
… [dā]ralu salāmu dāralu karālu dāralu … dār al-salām ‘House of Peace’, dār al-qarār
damālu dāralu kuludu dannatalu adin ‘House of Steadfastness’, dār al-ǧamāl ‘House of
dannatalu na[īmu …] Beauty’, dār al-ḫuld ‘House of Eternity’, ǧannat
al-ʿadan ‘Paradise of Eden’, ǧannat al-na[ʽīm
‘Paradise of Happiness’ …]
… usu ran risi maṇikkatun ṣeduṇu raṭu … from utmost heaven, a land made from gold,
uttʰara svargga ḷokaen … silver and jewels …
… sairu sakaru ṣāvvā mi sat narakāen … al-saʽīr ‘flaming fire’, saqar ‘hell’, al-hāwiyat
ādītya mulu koṭu ṣadān pṛttʰivi loka ‘depth, nethermost hell’, from these seven hells,
upaduv[ai … making the sun first (lit. as the root), creat[ing]
the earth …
… svarggai kautar eviana gagu svarggai … in heaven and the river named Kauṯar in heaven,
ṣaur eviana raṭu erāk eviana raṭu bābʰil the land named Syria, the land named Iraq, [the
evi[ana raṭu … land nam]ed Babylon, …
… pā]rīṣ eviana raṭu dadʰīrā eviana raṭu … the land named [Per]sia, the land named Mes-
sind eviana raṭu jabuduv eviana raṭu opotamia, the land named Panjab, the land
mitak ra[ṭu … named India (Jambudvīpa), all these lands …
 Written Artefacts from the Maldives | 25

Table 5: Gamu lōmāfānu, plates 2–3 (excerpt).

… upaduva]i prāṇa dī mīn aṛddʰa aṅgain … creati]ng (Adam), giving him breath, creating
śauā upaduvai mi de mapirīn svargga Eve from his half body, lodging this couple of man-
vaṣai … and-woman (lit. mother-and-man) in heaven …
… ādamu petāmbarun suktra bīdain ṣauā … the children born from Adam the prophet’s
baḍun upan genīn dari ve dunie mīsun ve vī sperm drop by the womb of Eve having become
mankind on earth,
mi ādamu petāmbarun ādi koṭu paṭai after 124,000 prophets had been created with
eklakka ṣauvīsu cāśra petāmbarun upede this Adam the prophet as the first, (and after
dʰunien nivana gat pase they had) departed (again) from the world,
upan mahammadu petāmubarun ṣāḷīsu in the fortieth year of Muḥammad the prophet,
āvurodun born after them,
mi petāmbarun kraṭa dabarīlu aisu buṇe Gabriel coming to this prophet, talking to him
gosu and going away (again),
mi dabarailāi mīkailai mi de malāikatun this Gabriel and Michael, the two angels, ap-
peṇe pearing (again),
burāk eviana asu puṭe mahammadu causing Muḥammad the prophet to mount the
petāmbarun aruvai back of the horse named Burāq,
baitalu magadeṣaṭa gene gosu … taking him to Jerusalem (Bait-al Maqdis) …

3 Mixing Dives akuru with Arabic script

More than a hundred years after these copperplate grants were issued, the first
artefacts that are inscribed in Arabic script appear on the Maldives. Leaving aside
building and funeral inscriptions that are written in Arabic all over,19 the most
relevant monuments for our investigation are those where Arabic writing appears
along with Dhivehi. The first such artefact that has been preserved is a copperplate
grant of the year 758 AH (1356–1357 CE), which is about the endowment of a mosque
named Boḍugalu (‘big rock’) in the islands’ capital, Male.20 There is still a clear
distinction between the Arabic and Dhivehi elements in this document in that its
first plate, comprising a Shahāda formula, the seal of King Ǧalāl ad-dīn bin Ṣalāḥ
ad-dīn (r. 1306–1341), and a few lines summarising the construction of the mosque,
is all in Arabic, while the remaining plates are all in Dhivehi.21 In later decrees,

||
19 See Kalus and Guillot 2005; Gippert forthcoming c, § 2.2.
20 See Mālēge Miskittaʼ 1980, 91–93 as to the mosque.
21 See Gippert forthcoming c, § 2.1 for further details.
26 | Jost Gippert

however, we begin to meet with Arabic elements that are inserted, in right-to-left
directed Arabic script, into a Dhivehi context, in its turn, written in left-to-right
Dives akuru. The first specimen of this type of mixed-script document is a copper-
plate grant of similar appearance to the Boḍugalu lōmāfānu; unfortunately only
one plate has been preserved of this grant, containing neither a date nor the name
of the issuing king (or queen).22 We find here, integrated into the text passage that
explains the ruler’s purposes, two quotations of God and the Prophet that are taken
from the Qurʼan and the Ḥadīth, respectively; the context reads (cf. Fig. 7):

Causing for (every) Muslim king, royal prince, queen, high-born noble-man (or) noble-
lady, minister, mandatee, or other person whosoever who intends to maintain this sacred
parcel of land as a sacred parcel (by saying) ‘God the Almighty said, “May the curse of God
be on the unjust”’, the nine-fold blessing that is there for all persons who are well-
intentioned towards Muslims, (and by saying) ‘The Prophet – may God’s prayers and peace
be upon him – said, “None of you becomes a true believer until he likes for his brother what
he likes for himself”’, in this way, Muḥammad the commander-in-chief …23

Fig. 7: Lōmāfānu (single plate) with insertions in Arabic (highlighted) in lines 1–2 and 4–5; Bell
1930, pl. I A.

While the Arabic elements of this grant are still full sentences in their own right,
later documents exhibit the tendency towards shorter phrases or even individual
words in Arabic script being inserted into the Dhivehi context. This is true, first of
all, for two further fragments of copperplate grants that are undated. In one of
them, also consisting of only one plate (see Fig. 8), we meet, within a series of taxes
that were levied for a newly endowed mosque, the expression qalam govibai denot-

||
22 The fragment is only known via Bell 1930, 541–552 with pl. I; its present whereabouts are
unknown.
23 For a first translation see Bell 1930, 543, where the Arabic elements are omitted, however.
The Arabic quotations are ‫ﻰ اَﻟﻈَﺎﻟِﻤﯿﻦ‬ ْ َ‫ﺖ ُﷲ َﻋﻠ‬ ُ ‫( َﻗﺎَل ُﷲ َﺗَﻌﺎﻟَْﻲ اًَﻻ ﻟَْﻌَﻨ‬see Surah 7.44 and 11.19 of
the Qurʼan) and ‫ﺴﮫ‬ ِ ‫ﺐ ﻟِﻨْﻔ‬
ُ ‫ﺐ ِﻷِﺧْﯿِﮫ َﻣﺎ ُﯾِﺤ‬ ‫ﺻﻠ ﱠْﻲ ُﷲ َﻋﻠَْﯿِﮫ َو َﺳﻠ ﱠﻢ َﻻ ُﯾْﻮ ِﻣُﻦ آََﺣُﺪُﮐْﻢ َﺣﺘﱠْﻲ ُﯾِﺤ ﱡ‬
َ ‫َﻗﺎَل اَﻟَﻨﺒًﻲ‬
(Ḥadīth 183).
 Written Artefacts from the Maldives | 27

ing a tax for registrars and registration, with qalam ‘scriptor(ium)’ written in Arabic
(‫)ﻗﻠﻢ‬.24 The other one, of which only half a plate has survived (see Fig. 9), contains
two instances of the Arabic title al-ġāzī ‘conqueror, warlord’, once associated with
King Ibrahim (III, r. 1585–1609 CE), son of Muhammad Boḍu Takurufānu
(r. 1573–1585 CE) who liberated the Maldives from the short Portuguese rule
(1558–1573), and once, with his uncle Ali Katīb Takuru, the latter’s brother.25

Fig. 8: Arabic qalam (highlighted) in a Dhivehi copperplate grant (fragment); Bell 1930, pl. II B.

Fig. 9: Arabic al-ġāzī (highlighted) in a Dhivehi copperplate grant (fragment); Bell 1930, pl. III A.

||
24 See Bell 1930, 557–560 with pl. II for a first account. The other taxes in question are ras boli
‘royal cowries’, rahu rā ‘toddy tax’, goi mas ‘fish tax’ and pālabba ‘tree tax’; in contrast to Bell,
who takes qalam and govibai as two distinct taxes (‘Tax Collector’s Fee’ and ‘Cultivation Tax’),
the two terms are taken together here as they are demarked as constituting a phrase by a dou-
ble vertical stroke (daṇḍa) preceding and following them.
25 See Bell 1930, 560–567 with pl. III for a first account. According to Bell’s reconstructions,
about fourteen characters (akṣaras) per line must have been lost with the right half of the plate.
28 | Jost Gippert

3.1 Paper documents

Even though the change of writing directions must have caused severe prob-
lems for writers here and there, the practice of inserting Arabic words in Arabic
script into Dhivehi contexts written in Dives akuru continued for at least two
centuries, in all kinds of written artefacts that were produced, beginning with
the first charters on paper which replaced the copperplate grants from the end
of the sixteenth century onwards.26 A typical example of such charters, which
are usually called fatkoḷu (lit. ‘leaflet’),27 is a document issued in 1108 AH (1696 CE)
by King Muḥammad son of Ḥāǧǧī ʽAlī Tukkala (r. 1692–1701) concerning the
renewal of the mosque on Gan island that had been the object of the Gamu
lōmāfānu some centuries before. The Arabic elements in this voluminous fatkoḷu
(cf. Fig. 10) comprise, firstly, the king’s seal, attached at the top of the leaf, then
the basmala (introductory formula ‘In the name of God, the Most Gracious, the
Most Merciful’, ‫ﺣﯿِﻢ‬ِ ‫ٱﻟﱠﺮْﺣَٰﻤِﻦ ٱﻟﱠﺮ‬ ‫) ِﺑْﺴِﻢ‬, and five and a half lines of continuous
Arabic text (prayers quoted from the Qur’an and the Ḥadīth) with the name of
God and other important words highlighted in red ink. Arabic elements usually
also appear in red ink in the Dhivehi text that follows from line 6 onwards; the
words and phrases concerned range from the name and designations of the
Prophet, muḥammad (combined with Dhivehi petāmbaru ‘prophet’ in line 40),
rasūl-allāh (lit. ‘messenger of God’) and sulṭān-al- anbiyāʼ (lit. ‘king of prophets’,
in line 10), via names of other persons such as al-sulṭān muḥammad (referring to
the issuing king in line 28), al-sulṭān iskandar ibrāhīm (referring to King Ibrahim
Iskandar I, r. 1648–1687, in lines 27 and 32), a vizier named hāǧǧī muḥammad,
and the scribe, muḥammad ḫaṭīb (in line 49), references to the Qur’an (ḥaqq
qurʼānā, lit. ‘truth of the Qur’an’, in line 10 and simply qurʼān combined with
Dhivehi pot ‘book’ in line 44), the Kaaba (kaʽbat-ullāh in line 45), the Hijra
(hiǧra- in line 11), and other terms that pertain to Muslim thought (e.g. raḥmat
‘mercy’ in line 40, ṯiqalain ‘the two species of genies and men’ in line 9, and
ṣalawāt ‘prayer rituals’ in line 35), up to mere abstract nouns (e.g. manfaʽa ‘ben-
efit’ in line 34 and qabūl ‘acceptance’, combined with Dhivehi kuravvai ‘making’
to form a complex verb meaning ‘accept’ in line 19). In several cases, the Arabic
terms are equipped with Dhivehi endings, in turn, written in Dives akuru (and

||
26 The oldest paper document that has been preserved can be dated to c. 1580 CE; it concerns
the endowment of a mosque on the island of Kolhufushi in Mulaku atoll and was issued by
Muḥammad Boḍu Takurufānu (see Bell 1940, 187–190 with pl. L).
27 The Maritime Asia Heritage Survey website (Feener s.a.) displays about fifty Dhivehi char-
ters in Dives akuru script; an immense increase of material in comparison with Bell’s list of
eleven specimens (Bell 1940, 187).
 Written Artefacts from the Maldives | 29

usually not in red); this is the case, for example, of rasūl-allāh and kaʽbat-ullāh,
which are followed by the genitive ending -ge (lines 10–11 and 45), manfaʽa (line
34) with the ablative ending -in (‘from the benefit’) or ǧāhil ‘ignorant (person)’
with the indefinite suffix -aku (line 44). However, the Dhivehi endings could
also be integrated into the Arabic spelling, as in the case of hiǧrain (line 11) with
the ablative ending -in (‘from the Hijra’) or in muʼmin musulmanaku ‘a Muslim
believer’ with the indefinite suffix -aku (line 41). In contrast to this, Arabic terms
are sometimes rendered in Dives akuru; this is true, for example, of the name of
Medina spelt madhīnayaṭ with a dative ending in line 11, rasūl ‘messenger,
prophet’ appearing with the comitative suffix -ayi in line 46, and insī ‘human
(being)’ occurring in the form insīnayi with a plural ending and the comitative
suffix in line 9, contrasting with ǧinī-nayi ‘genies’ where the word stem is writ-
ten in Arabic. Beyond this, Sanskrit terms also continue to be used (in a spelling
strongly influenced by Dhivehi pronunciation); this is especially true of the long
list of epithets with which the kings used to be ornated.28 The interplay of all
these elements is illustrated with some excerpts of the document in Table 6.

Table 6: The Gamu fatkoḷu of 1696 CE (excerpts).

… sulṭān-al-anbiyāʼ eve nan dʰevvai : … from the Hijra that (our) lord, (his) majesty the
ḥaqq-qurʼānā dʰī poṇuvvi kau rasūl- Messenger of God undertook to Medina, sent out
allāhu-ge pānun' madʰīnayaṭ voḍigat (by God) who had given him the name King of
hiǧrain Prophets and bestowed on him the truth of Qur’an,
ekusāstura ekusattʰa aṭvana ahar etere : in the 1108th year,
svastī srīmata mahāsrī purasituru audʰāna (I), Hail! the Maharaja of the world, glorious,
srīkula ramvahāaudʰakīriti kattʰiri bvana with great splendour, excessively bright, a
mahāradʰum vidʰāḷīn : kshatriya of a brilliant race, with a great golden
fame of power, have ordered:
… mage redʰali kurana hataruvana averudʰu … in the fourth year of my reign …
…
… srīkula ramvahāaudʰakīriti kattʰiri bvana … After the Maharaja of the world, a kshatriya of
mahāradʰunge kattʰiri mudʰigesvuman mi brilliant race, with a great golden fame of power
pat lievvumuge huddʰayi vidʰāḷu bahayi had left with the command to write down this
nikume al-wazīr hāǧǧī muḥammadu document with the kshatriya’s seal, and (his)
dʰoṭimenayi aḷā kī hidʰu muḥammad ḫaṭīb subject, the vizier Muḥammad Doṭimena had
aḷā lī :: read it, (his) subject Muḥammad the scribe
wrote it down.

||
28 Geiger (1919, 176) and Bell (1940, 67) apply the Sinhalese term biruda < Skt. viruca. Given
the distortion of the Sanskrit terms over the centuries, these epithets are not always restorable.
30 | Jost Gippert

Fig. 10: The Gamu fatkoḷu of 1696 CE; photograph Jost Gippert, 1992.
 Written Artefacts from the Maldives | 31

3.2 Charters on wood and stone

The extensive mixture of Dhivehi text with Arabic elements in official docu-
ments is not restricted to paper charters. It is especially two inscriptions on
wood that deserve our interest for the period in question. As a matter of fact,
the two inscriptions, both dated to the year 1062 AH (1652 CE), are probably
copies of charters that also existed on paper; they concern the endowment of
mosques by King Ibrahim Iskandar I, in quite the same wording as what we
find in fatkoḷus. One of them is engraved in five longitudinal lines on a beam
of about 3 m length (Fig. 11 may give an idea of its size), which was used in
the construction of the so-called ‘Old Mosque’29 of the Galolhu quarter of
Male until its demolition in about 1980; it is today preserved in the National
Museum in Male.

Fig. 11: Endowment inscription of the Galolhu Old Mosque (total); photograph Jost Gippert, 1993.

In the excerpt of line 4 shown in Figs 12a–b, we see the combination of Arabic
ʽimārat ‘construction’ with Dhivehi kuḷaimaṭ, the verbal noun ‘making’ in the
dative used to express a purpose; the sentence reads: mi dʰanāru huṭi galoḷu
emve aḷutakun mi danāru ʽimārat kuḷaimaṭ adʰai dʰannavai gat hidʰu ‘when (I)
received the information that all subjects of Galolhu where this building stands
desired to construct this building’. The construction is dated to Friday, 11 Ram-
adan of the fifth regnal year of the king, with both the month name, ramaḍān,
and the Dhivehi word for ‘month’, mahu, written in Arabic (mahāradʰunge
pasuvana averudʰu ramaḍān mahu egāra vī hukuru dʰā); the date equals 16 Au-
gust 1652.

||
29 Bā Miskit, also known under the names Kuṇḍikoi-Miskit and Masjid al-Hadith.
32 | Jost Gippert

Figs 12a–b: Same, excerpt with line 4 highlighted; photographs Jost Gippert, 1993.

The second ‘fatkoḷu on wood’ is a large rectangular board measuring 1.7 ×

0.54 m; its eighteen-line inscription attests the endowment of a mosque on the
island of Gan in the southernmost atoll, Aḍḍū (or Seenu), on the 23 Muharram
of the same year, a Friday as well, corresponding to 5 January 1652. This inscrip-
tion is peculiar in that it begins with a lengthy account of the Creation and the
life of Muḥammad, as in the Gamu lōmāfānu; in addition, it provides important
information on the circumstances of the conversion of the Maldives to Islam.30
The board, today also kept in the National Museum in Male, was probably ex-
posed in the mosque concerned as long as it existed.31
During the construction of mosques, fatkoḷu-like endowment acts were
sometimes also engraved in the building material itself. This is true, for exam-
ple, of the Dives akuru inscription that was applied to the four inner walls of the
so-called ‘Middle Mosque’ (Medu Miskit) in Male, which provides details con-
cerning both the construction of the mosque and its properties. Here, we read,
among others, the Arabic words ʽādat ‘custom’ (with the ablative ending -in
added in Dives akuru), miḥrāb ‘niche for praying’ (once with the comitative

||
30 See Gippert 2003 and forthcoming b for further details.
31 See Bell 1940, 190 according to whom the board was still ‘in the possession of the inhabit-
ants of Gan Island’ in his times.
 Written Artefacts from the Maldives | 33

ending -āi, written in Dives akuru together with the stem-final b), Ḥadīth and
ziyārat (‘shrine’).32

3.3 Funerary inscriptions

A considerable number of artefacts written in Dives akuru with Arabic insertions
consist of epitaphs which, however, are less numerous than the bulk of funerary
inscriptions that are all in Arabic. It is usually datings (by Hijri year and month),
personal names, and mentions of God and the Prophet that are written in Arabic
in the mixed inscriptions. Of the sixteen such inscriptions that have been rec-
orded for the big cemetery of the Friday Mosque in Male,33 three are undecipher-
able today due to erosion caused by the maritime climate; the other ones are
datable between 1662 and 1782 CE. The outer appearance of the Dhivehi script is
partly similar to that of the contemporary charters, as in the case of the grave-
stone of a certain Ḥussain Boḍu Doṭimena Kiṇagepānu, son of Wazīr
Muḥammad Pāmuladeri Maṇikupānu, who died in the night of Monday, 6 Dhu
al-Hijja 1178 AH, corresponding to 27 May 1765 CE (see Fig. 13); its textual content
is illustrated in Table 7.34 In a few cases, however, the script on the gravestones
is extremely ornamental, thus, reminiscent of Arabic calligraphy. This is true,
for example, of the epitaph of a certain Ḥussain Āmin (?) Kilagepānu, grandson (?)
of ʽUmar Rannabaderi Kilage, who died on Friday, 18 Jumada al-awwal 1089 AH,
corresponding to 8 July 1678 CE (see Fig. 14).35 As far as it can be ascertained, its
text is displayed in Table 8.36

||
32 A reproduction of the inscription is published in Mālē Hukuru Miskit 1984, 173–180; an
edition of the text in a Thaana transcription in the same book, 97 and 165–168.
33 Epitaphs with mixed inscriptions are also found in graveyards on other islands of the ar-
chipelago; they are especially numerous in the Koagannu graveyard on Hulhumeedhoo island
(Seenu atoll).
34 See Mālē Hukuru Miskit 1984, 372 (group XIII, no. 142).
35 See Mālē Hukuru Miskit 1984, 346 (group IV, no. 9).
36 See Gippert forthcoming c, § 3.1, for further information on Maldivian funerary inscriptions.
34 | Jost Gippert

Fig. 13: Gravestone of 1765; photograph Jost Gippert, 1993.

Table 7: Epitaph of 1765 CE.

hiǧrain mitakvana ave|rudʰu sanat 1178 gai ḏū- This is to memorise that in the year 1178 after
al-ḥiǧǧa mahu ha | rean vvī angāra vīle rei | Hijra, in the month of Ḏū’l-ḥiǧga, in the fourth
hataruvana sāʽatu al-sulṭān | al-ġāzī ḥassan hour of the sixth night which dawned into
ʽizz-al-dīn | sirīkʰula raṅmība kattri bvana | Tuesday, passed away Ḥussain Boḍu Doṭimena
mahāradʰunaṭu ek kihun upan | mī radʰunme Kiṇagepānu, born from the same womb as the
diapurasūtā vi al-wazīr muḥammadu pāmu- Maharaja of the world, Sultan al-Ġāzī Ḥassan
ladʰeri mani|kupānaṭ | me upan ḥussayn boḍu ʽIzz al-dīn, a kshatriya of brilliant race with
dʰoṭimenāyi ki|ṇagepānu | purauttʰa vi | kamu golden elephants, and born to the vizier
hadʰān : Muḥammad Pāmuladeri Manikupānu by whom
was begotten this same king.
 Written Artefacts from the Maldives | 35

Fig. 14: Gravestone of 1678; photograph Jost Gippert, 1993.

Table 8: Epitaph of 1678 CE.

hiǧrain | ekusas|tura | uṇavai|vana ahar eture | This had to be written down to memorise that
svastī srīma|ta mahāsirīkula ra|mvība kattri in the 1089th year after Hijra, in the 31st year
ba|vana maṣāradʰunge | ektirīsvana | aharu of Hail! the glorious Maharaja of the world, a
ǧumād al-awwal mahu | hukuru aṭāra vi|le rei kshatriya of very brilliant race, with golden
mi radʰuna ek-|kihun upan ʽumar rannaba|ḍīri elephants, in the month of Ǧumād al-awwal, in
kilagon upan | koṭari kalegepānuge | dʰi the night dawning into Friday, the 18th, on this
kamanāpānu | baḍun upan ḥussayn | āminnāi day passed away Ḥussain Āmin Kilagepānu,
kilage|pānu mi dʰva|hu me purauta | veddʰye born from the womb of Kamanāpānu, the
kamu hadʰā|naṭ līpanti daughter of Koṭari Kalegepānu who was born
to ʽUmar Rannabaḍeri Kilage, born from the
same womb as that king.
36 | Jost Gippert

4 The Invention of Thaana

With the replacement of Dives akuru by a newly invented right-to-left directed
script in the eighteenth century, the problem of inserting Arabic elements into a
Dhivehi context was solved for the Maldivian scribes. Curiously enough, the new
Thaana script was not based on Arabic letters but on digits: as shown in Table 9,
nine characters derive from the Arabic or, rather, Persian digits for 1 to 9, nine
further characters from the inherited digits of Dives akuru.37 Considering that in
Arabic, numbers are written against the regular script direction, the use of the
digits for the purpose of representing consonants in a right-to-left directed script is
perplexing. This is all the more true since the ratio of the assignment of sound
values to the digits is still unknown. Nevertheless, the influence of Arabic writing
habits is clear from another characteristic: Thaana uses the Arabic vocalisation
marks as far as possible for the assignment of vowels (including sukūn, the circle-
shaped mark denoting the absence of a vowel). The interplay of Thaana and Arabic
script is illustrated in Fig. 15 and Table 10 with a page from the Rādavaḷi, a chroni-
cle covering the time from Creation up to the eighteenth century;38 note that the
inherited digits are used for denoting numbers, not the Arabic ones, and that both
the Arabic names and the numbers are usually in red ink.

Table 9: The development of Thaana characters.

Thaana character ‫ހ‬ ‫ށ‬ ‫ނ‬ ‫ރ‬ ‫ބ‬ ‫ޅ‬ ‫ކ‬ ‫އ‬ ‫ވ‬
Sound value h ś n r b ḷ k ˀ w
Persian digit ١ ٢ ٣ ۴ ۵ ۶ ٧ ٨ ٩
Numeric value 1 2 3 4 5 6 7 8 9

Thaana character ‫މ‬ ‫ފ‬ ‫ދ‬ ‫ތ‬ ‫ލ‬ ‫ޏ ގ‬/‫ޱ‬ ‫ސ‬ ‫ޑ‬
Sound value m F d t l g ñ/ṇ s ḍ
Dives akuru digit 𑥑 𑥒 𑥓 𑥔 𑥕 𑥖 𑥗 𑥘 𑥙
Numeric value 1 2 3 4 5 6 7 8 9

||
37 See Gippert 2013, 96–98, for more details.
38 The present copy of the chronicle (styled ‘C’ in Bell 1940, 200) was published in facsimile in
Rādavaḷi 1979, 2; fragments of two further copies (‘A’ and ‘B’, the latter in Dives akuru script),
which were partially edited by Bell (1940, 198–200 with plates N and O), do not contain the
passage in question.
 Written Artefacts from the Maldives | 37

Fig. 15: Rādavaḷi ‘C’, page 2; Mohamed 2004, 5, Fig. 5.

38 | Jost Gippert

Table 10: Rādavaḷi ‘C’, (excerpt).

…’ādamu fētānbarunaʼ : kulavati kuraʼ|vai ʼe … after having made her (Eve) a family mem-
suvāmīnge reʼsavaʼigenfi hataʼ | dāku su- ber of the prophet Adam, having entrusted (?)
varuge laʼi tiʼbavaʼifaʼi : mi de ʼaḷun | bo- her to him as (her) husband, He (God) lodged
vanaʼaʼ : vaʼdavaʼifī māʼi : duniʼa matī : | them for seven days in heaven and (then)
fētānbarukamu huśi ʼaharu . 300 . caused these two servants to enter earth; the
years of his (Adam’s) prophethood in the
world (were) 300.
miʼaʼ fahu | ʼaneʼka : 1200 : ʼaharu duvas veʼje When after this another 1200 years had seen
fahu . | lamakuge fut nūḥu fētānbaru kalōge the light, He created the noble prophet Nūḥ
haʼdavaʼi | mi kalōge : fuśāʼi samundurē naʼu (Noah), son of Lamech; the years this noble-
maʼcā՚i | huśi ʼaharu : 950 : man passed in a boat on the ocean after
having embarked (were) 950.
miʼaʼ fahu : ʼāzaruge fut |ʼibrahīm fētānbaru After this, the years of the lifetime of the
kalōge : vanavaruge ʼa|haru : 1142 : noble prophet Ibrahīm (Abraham), son of
Azar, (were) 1142.
mī fahu ʼimrānge fut mūsā | fētānbaru After this, the years of the lifetime of the
kalōgośʼ : vanavaruge ʼaharu . 505 . noble prophet Mūsā (Moses), son of Imrān
(Amram), (were) 505.
mī fahu den dāwūdu fētānbaru : … After this then, of the prophet Dāwūd (David) …

5 Outlook
Dhivehi has witnessed a strong influence of another foreign language, namely,
English, since the nineteenth century, and especially from 1887 to 1965 when
the Maldives were a British protectorate. Indeed, there was an attempt to intro-
duce Latin script for Dhivehi, based on English orthographical premises; how-
ever, this ‘Male Latin’ was only officially used for two years (1976–1978), and
only a few written artefacts from that period have survived (see Fig. 16 showing
a metal plate with a Roman inscription from the grave of a certain Ali Didi on
Hithadhoo island). In contrast to Arabic, it has remained rather uncommon to
insert English words in Latin script in a Dhivehi context, which would, of
course, bring about the problem of mixing script directions again; instead, Eng-
lish elements are usually entered in a Thaana transcription today, as in the
online news headlines illustrated in Table 11;39 the only Dhivehi elements in this
text are, except for nominal endings, the auxiliary verbs vumun in feʼil-vumun

||
39 Mihaaru, 9 May 2023 (<https://mihaaru.com/news/121098>, accessed 9 January 2024).
 Written Artefacts from the Maldives | 39

‘by failure’, lit. ‘by be(com)ing fail’, and kuri in saspeṇḍ-kuri ‘suspended’, lit.
‘suspend-made’. As a Dhivehi word, we might also count ra’īs ‘president’; this,
however, is a loan from Arabic.

Table 11: News headline.

.‫ނ‬ ަ ‫ރޯމ‬
ް ‫ޝ‬ ޮ ‫ރްށ ްޕ‬
ަ ‫ސ‬
ަ ‫ރ ޮއިފ‬
ި ‫ންޑ ުކ‬
ް ‫ސެޕ‬
ް ‫ސ‬
ަ ‫ނ‬
ް ‫ލުވުމ‬
ް ‫ޓ ެފިއ‬ ި ‫ސިކުއ‬
ީ ‫ރ‬ ެ ‫ސެގ‬
ް ‫ރީއ‬
ަ
raʼīsge sekiʼuriṭī feʼil-vumun saspeṇḍ kuri Promotion for officer suspended for Presiden-
ofisaraʼ promōšan tial security failure.

Fig. 16: Grave inscription in ‘Male Latin’; photograph Jost Gippert, 1999.

References
Bell, Henry Charles Purvis (1930), ‘Excerpta Maldiviana. No. 9.–Lómáfánu’, Journal of the Cey-
lon Branch of the Royal Asiatic Society, 31 (83): 539–578 with pl. I–IV; repr. in Henry
Charles Purvis Bell, Excerpta Máldiviana, New Delhi: Asian Educational Services, 1998.
Bell, Henry Charles Purvis (1940), The Maldive Islands: Monograph on the History, Archaeology
and Epigraphy, Colombo: Ceylon Government Press.
Feener, R. Michael (ed.) (s.a.), Maritime Asia Heritage Survey, Kyoto: Center for Southeast Asian
Studies, Kyoto University, <https://maritimeasiaheritage.cseas.kyoto-u.ac.jp> (accessed on
14 May 2023).
40 | Jost Gippert

Forbes, Andrew D. W. (1984), ‘A Roman Republican Denarius of c. 90 B.C., from the Maldive
Islands, Indian Ocean’, Archipel, 28: 53–60.
Fritz, Sonja (2002), The Dhivehi Language: A Descriptive and Historical Grammar of Maldivian
and Its Dialects, 2 vols, Würzburg: Ergon.
Geiger, Wilhelm (1919), ‘Máldivian Linguistic Studies’, Journal of the Ceylon Branch of the Royal
Asiatic Society, 27 – Extra number.
Gippert, Jost (2003), ‘Early New Persian as a Medium of Spreading Islam’, in Ludwig Paul (ed.),
Persian Origins – Early Judaeo-Persian and the Emergence of New Persian: Collected Pa-
pers of the Symposium, Göttingen 1999, Wiesbaden: Harrassowitz, 31–47.
Gippert, Jost (2004), ‘A Glimpse into the Buddhist Past of the Maldives. I. An Early Prakrit
Inscription’, Wiener Zeitschrift für die Kunde Südasiens, 48: 81–109.
Gippert, Jost (2013), ‘An Outline of the History of Maldivian Writing’, in Chen Shu-Fen and
Benjamin Slade (eds), Grammatica et Verba: Glamor and Verve, Studies in South Asian,
Historical, and Indo-European Linguistics in Honor of Hans-Henrich Hock on the Occasion
of His Seventy-fifth Birthday, Ann Arbor, MI: Beech Stave Press, 81–98.
Gippert, Jost (2013–2014), ‘A Glimpse into the Buddhist Past of the Maldives. II. Two Sanskrit
Inscriptions’, Wiener Zeitschrift für die Kunde Südasiens, 55: 111–144.
Gippert, Jost (2015), ‘An Unusual Account of the Miʿrāǧ’, in Elisabetta Ragagnin and Jens Wil-
kens (eds), Kutadgu Nom Bitig: Festschrift für Jens Peter Laut zum 60. Geburtstag, Wies-
baden: Harrassowitz, 163–182.
Gippert, Jost (forthcoming a), ‘Buddhism in the Maldives’, Brill’s Encylopedia of Buddhism, vol.
4: History.
Gippert, Jost (forthcoming b), ‘The role of Iran in the Islamicisation of the Maldives’, to appear
in the proceedings of the conference ‘Iran and Persianate Culture in the Indian Ocean
World’, St Andrews, 7–8 September 2022.
Gippert, Jost (forthcoming c), ‘Epigraphy of the Maldives (c. 6th–19th century CE)’, in Kaja Harter-
Uibopuu, Jochen Hermann Vennebusch and Ondřej Škrabal (eds), Handbook of Epigraphic
Cultures, Berlin: De Gruyter.
Kalus, Ludvik and Claude Guillot (2005), ‘Inscriptions islamiques en arabe de l’archipel des
Maldives’, Archipel, 70: 15–52.
Mālē Hukuru Miskit (1984), Male: Divehi bahāi tārīḫaś ḫidmat kurā qaumī marukazu [‘National
Center for Linguistic and Historical Research’].
Mālēge Miskittaʼ (1980), Male: Dipārṭmenṭ of infōrmēšan enḍ brōḍkāsṭing [‘Department of
information and broadcasting’].
Mohamed, Naseema (1999), Dhivehi Writing Systems, Male: National Centre for Linguistic and
Historical Research.
[Mohamed, Naseema] (2004), Scripts of Maldives, Male: National Centre for Linguistic and
Historical Research, <http://www.qaumiyyath.gov.mv/docs/whitepapers/history/Scripts
%20of%20Maldives.pdf> (accessed on 14 May 2023).
Rādavaḷi (1979), Male: Divehi bahāi tārīḫaś ḫidmat kurā qaumī marukazu [‘National Center for
Linguistic and Historical Research’].
Leah Mascia
A Journey Through the Multilingual
Landscape of Greco-Roman and Late
Antique Oxyrhynchus
Abstract: This paper aims to provide an insight into the multicultural and
-lingual panorama of Oxyrhynchus (modern el-Bahnasa), integrating textual
and archaeological evidence. The present study combines the examination of
papyrological and epigraphic sources with the analysis of the results of over
thirty years of archaeological investigations carried out by the archaeological
mission of the University of Barcelona. Despite the wealth of Greek papyri pub-
lished so far, whose significance remains indisputable for understanding the
society inhabiting this ancient site, new discoveries highlight the importance of
the Egyptian language and culture up until the late antique phase and beyond,
as well as the existence of foreign communities (Romans, Jews, Nabataeans)
who have introduced their own culture into Oxyrhynchus and shaped the local
social and religious panorama. Furthermore, through the lens of the documen-
tation discussed, the linguistic transformations which occurred in the Christian
period and particularly the essential importance of the Coptic language in the
local society will be highlighted.

1 Introduction
Oxyrhynchus represents an interesting case study for having a glimpse into the
variety of linguistic interactions which might have existed in many Egyptian
settlements during the Greco-Roman and late antique periods. This site, laying
on the western bank of the Bahr Yusuf River, c. 190 km from the city of Cairo, is
known predominantly in light of the extraordinary corpus of Greek papyri dis-
covered here by Bernard Grenfell and Arthur Hunt more than one hundred years
ago. However, the essential role played by these Greek papyri in the under-
standing of Oxyrhynchite society and beyond has inevitably overshadowed a
wide range of papyrological and epigraphic evidence attesting to a variety of
other languages.

Open Access. © 2024 the author, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-003
42 | Leah Mascia

Combining the research conducted in museums and other research institu-

tions with the results of over thirty years of archaeological investigations at the
site carried out by the mission of the University of Barcelona,1 this paper aims to
provide an insight into the different degrees of linguistic interaction traceable in
the multicultural panorama of Oxyrhynchus between the Greco-Roman and the
late antique phases. This complex intertwining between languages and scripts
in this ancient city will be observed from two distinct perspectives. On the one
hand, analysing the coexistence of writings in different languages and scripts in
the same spatial contexts and/or cultural settings. On the other, observing how
different languages and scripts were mixed in the production of a variety of
written sources, from inscribed portable objects made of papyrus, wood, clay
and stone to texts inscribed or painted on the walls of various buildings discov-
ered at the site. It remains essential, when possible, to consider these sources in
their original archaeological context for understanding how this linguistic com-
presence shaped the cultural landscape of this city, marking phenomena of
continuity and transformation throughout the centuries. For this reason, as
much as the sources available allow, this paper will discuss the collected docu-
mentation in its original archaeological settings or try to recontextualise them
according to their specific function. However, major limits of contextualisation
cannot be overlooked dealing with other inscribed materials discovered in the
past which were predominantly unearthed in the rubbish dumps surrounding
the ruins of this site. Their nature as discarded material poses numerous ques-
tions regarding their provenance and function. In this regard, their discussion
according to the cultural and/or social settings to which they could have origi-
nally belonged is meant to provide an overview of the surviving evidence rather
than offering a comprehensive solution to these issues.

2 A brief introduction to bilingual practices in the

administration of Greco-Roman Oxyrhynchus
Before examining the contexts and social settings in which we can trace multi-
lingual practices in Oxyrhynchus, a side note must be devoted to the admin-

||
1 My deepest gratitude goes to the directors of the mission of the University of Barcelona,
Esther Pons Mellado and Maite Mascort Roca, for having welcomed me into the team since 2020
and for their constant support in studying the unpublished Oxyrhynchite epigraphic and papy-
rological material.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 43

istration of this settlement in the Greco-Roman phase and what we can recon-
struct according to the evidence offered by the papyrological documentation.
Papyri dating before the Ptolemaic period are virtually absent due to the ris-
ing underground water level in the entire archaeological area, a phenomenon
already begun in the early twentieth century with the expansion of the agricul-
tural landscape.2 Nevertheless, documents especially in Demotic are occasional-
ly found,3 and various pieces of evidence seems to remark that aside from
Greek, which had become the administrative language of the country after the
conquest of Egypt by Alexander the Great, Egyptian played a key role in the
Oxyrhynchite administrative settings during the Ptolemaic period.4
The native language seems to have been largely abandoned in administra-
tive settings after the Roman conquest, substantially replaced, as in most Egyp-
tian settlements, by the Greek language. Despite this apparent predominant
switch from Egyptian to Greek, especially in the administrative sphere, a small
corpus of Demotic documents dating between the late Ptolemaic and early Ro-
man periods has been recently identified. The Oxyrhynchite Egyptian adminis-
trative collection includes, among others, annuity contracts and documents
about real estate.5 Within the scope of this paper, is interesting to note that
many of them included a Greek translation,6 which perhaps was also meant to
make these documents comprehensible to the imperial officers, unable to read
the Egyptian version.

||
2 The corpus of papyrological evidence dating even to the Ptolemaic phase is very limited.
3 Several ostraca inscribed in Demotic, for instance, were found during the investigations of a
series of rubbish dumps surrounding the area of the so-called High Necropolis of Oxyrhynchus
in the excavation campaign held between November and December 2022. For a brief overview
of this unpublished material, see Mascia 2023b.
4 A corpus of c. 1800 Demotic ostraca, for instance, dated mainly between 170 and 116 BCE, is
housed in the collection of the universities of Pisa and Cologne. Most of these documents dealt
with the provision of a desert outpost between Oxyrhynchus and the Bahariya Oasis in the
Western Desert; see Thissen 2013.
5 Among the latest might be one document preserving the name of Emperor Trajan; see Quack
2016b, 107–108. However, Edward Love also reports the existence of a document dating to the
nineteenth year of Hadrian (135/136 CE); see Love 2021, 177. In addition, the presence of several
Demotic accounts, probably dating from the first century CE, should be mentioned; see Quack
2016b, 108.
6 Quack 2016b, 108.
44 | Leah Mascia

3 Multilingualism in the Egyptian temples of

Oxyrhynchus and beyond
The papyrological documentation offers a glimpse into the religious panorama
of this city between the Roman and late antique phases. Despite the existence of
sanctuaries devoted to the cult of a variety of Greco-Roman and Near Eastern
deities, Egyptian temples still played a central role in the lives of the local in-
habitants. Since the Ptolemaic phase, the Great Thoereion was beyond doubt
the major religious centre of this settlement. This sanctuary hosted the cult of
Thoeris,7 the goddess patroness of the city, here worshipped together with Isis,
Serapis and all the associated gods (i.e. synnaoi theoi). The central importance
of this religious organisation is underlined by the administrative documentation
dating predominantly to the early Roman period, which seems to demonstrate
that this temple and its priesthood had a chief function in the administration of
the other religious institutions of this city and the temples of the villages of the
Oxyrhynchite nomos.8 It is often in association with this sanctuary that we find
the name of the local highest priestly ranks9 and the personnel traditionally in
charge of the production of ritual texts essential for the endurance of the activi-
ties of the temples and their personnel.10 Considering this evidence, it might be
hypothesised that the Great Thoereion hosted a ‘house of life’ (per-ankh), ‘The
institution of an ancient Egyptian temple in which the priests formed, trans-
formed and transmitted the religious traditions of their country’, as defined by
Andreas Stadler.11

||
7 On the cult of this deity at Oxyrhynchus from the Late Dynastic to the Greco-Roman phase,
see Mascort Roca and Pons Mellado 2019.
8 For an introduction to the role of the Great Thoereion and its priesthood, as well as other insti-
tutions associated with this deity, in the religious panorama of Oxyrhynchus, see Mascia 2023a.
9 Like the prophetes Harthonis (SB X 10256, 55–67 CE). Another priest named Aurelius
Osorapis, holding the titles of archiprophetes and protostolistes (PSI IX 1039, third century CE),
and who was probably also associated with this sanctuary is the most important religious
authority of Roman Oxyrhynchus known so far. On these religious offices in Greco-Roman
Egypt, see Clarysse 2010, 288.
10 The papyrological documentation records only one priest holding this service in the Roman
phase; Thoonis in service at the temple of Athena-Thoeris and the divine Augustus Caesar in
the second century CE (P.Mich. XVIII 788). He bears the title of pteraphoros, the lector priest
with a key role in conveying and translating temple knowledge. On this religious office in
Greco-Roman Egypt, see Clarysse 2010, 288.
11 Stadler 2015, 190.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 45

Some of the Egyptian documentary, literary and paraliterary texts recently

attributed to this settlement, being located in the Great Thoereion or another
sanctuary, should be linked to the production of the ‘house of life’ of Oxyrhyn-
chus, presumably active at least until the late second century CE. Indeed, while
research on the ‘Egyptophone’ history of Oxyrhynchus is still in its infancy, the
in-depth studies carried out over the last few years by Joachim Quack in the
Egyptian Exploration Society papyrus collection in Oxford12 have provided a first
idea of the essential importance of the Egyptian textual sources for understand-
ing the religious panorama of this city.13
The corpus of Egyptian texts identified by Quack counts around forty hierat-
ic, seven hieroglyphic and ninety Demotic papyri;14 other evidence recently
discussed by Edward Love15 and some specimens spread in museum collections
should be added to this number.16 The majority of these texts seem to date from
the second century CE, with a minor number of exemplars dating from the first
century CE and, less likely, to the third century CE.17 A similar chronology can be
attributed to a large part of the administrative texts recording the activities of
the Egyptian temples and the priesthood of Oxyrhynchus.18 To some extent,

||
12 For an introduction to this corpus, see Quack 2016b.
13 While many of these texts might be attributed to the Egyptian temples of Oxyrhynchus, it
should be remarked that all of them were found scattered in the rubbish dumps surrounding
the ancient settlement. In particular, Kahle MS 2/3 belongs to the discoveries of the excavation
campaign 1897–1898; Love 2021, 187–188. P.EES Oxyrhynchus 3b 4/9 was found in the 1903–
1904 excavation campaign; Love 2021, 185. P.EES Oxyrhynchus 29 4B.42/H (13) and P.BM EA
10808 were discovered in the 1904–1905 excavation campaign; Love 2021, 184, 191. And Clack-
son MSS 8.2.5 in the 1905–1906 investigations; Love 2021, 198.
14 Quack 2016b, 106.
15 Kahle MS 2/3 and Clackson MSS 8.2.5; see Love 2021, 187–189, 198–199.
16 Several unpublished Demotic and hieratic ostraca are now kept in the collection of Colum-
bia University; they formerly belonged to the findings donated to the Metropolitan Museum of
Art in the early twentieth century. I owe my gratitude to Niv Allon for providing this infor-
mation on the de-accessed materials transferred in the early 1950s. In addition, numerous
unpublished inscribed objects in the Egyptian language have been discovered in the course of
the last years of investigations carried out by the mission of the University of Barcelona.
17 On the problems behind the dating of the aforementioned fragments, see Quack 2016b, 107;
Love 2021, 176, 183. PSI III 177Ro, probably dating between the second and third century CE, is a
hieratic text containing an offering liturgy, i.e. a water libation, destined to a private benefi-
ciary. For a preliminary description of this text, see Pintaudi 2011–2012 [2013], 161; Quack has
provided a more precise interpretation of the text; Quack 2016b, 109.
18 Although it must be noted that the number of administrative texts associates with the
Egyptian priesthood dating from the third century CE is also substantial. For a discussion of this
documentation, see Mascia 2023a.
46 | Leah Mascia

some of these testimonies, in particular the hieratic ritual texts, finds parallels
among textual sources predominantly from Tebtynis, but no extensive compari-
sons from contemporary Egyptian settlements renowned for the discovery of
textual corpora associated with temple activities.19 Comprehensively, the Ox-
yrhynchite corpus encompasses features which remark on the unique peculiari-
ties of this textual production epitomised by the attestation of distinctive phe-
nomena of script-shifting in the composition of texts belonging to the ritual
domain.20
An introduction to the bilingual textual production in large part presumably
associated with these sanctuaries is deemed necessary before examining written
artefacts in which we can trace a direct combination between different lan-
guages and scripts.21 While no ‘mixing’ phenomenon occur in these composi-
tions, they, nonetheless, hold essential importance in understanding how the
long interaction between indigenous and foreign communities had irremediably
influenced many aspects of the organisation and, therefore, the textual produc-
tion of the Egyptian temples. The multicultural setting in which these texts were
produced is remarked by the variety of solutions adopted to answer different
liturgical and administrative needs.
Egyptian temples in the Roman period were still organised according to a
system of rules which had been established since the Pharaonic period. Texts of
central importance in the traditional cult administration continued to be pro-
duced in Greek and Egyptian until the Roman period. The Book of the Temple,22 a
manual that provides insight into the laws regulating the native religious organ-
isations and describes the appearance, as well as the functioning of the ideal
Egyptian temples is illustrative in this sense.23 While this composition was only
witnessed for decades in Oxyrhynchus by two Greek fragments dating from the

||
19 Like also Soknopaiou Nesos and Narmouthis, see Love 2021, 183.
20 Love 2021, 200–204.
21 Among others, the following discussion comprehends both texts found in Egyptian and
Greek versions at Oxyrhynchus (i.e. The Book of the Temple), as well as compositions attested
in the two languages, but with only one version identified among the Oxyrhynchite papyri (i.e.
P.Oxy. XLVI 3285).
22 For an introduction to this handbook, see Quack 2000 and 2003.
23 The study conducted by Quack on numerous fragments of this handbook found at Tebtynis
has made it possible to reconstruct the content of this ancient manual. The manuscript con-
tained a treatise on the ideal temple and a detailed list of rights and duties that all the members
of the temple personnel must follow. See Quack 2000.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 47

second century CE,24 recent studies have led to the discovery of a discrete num-
ber of copies in the hieratic script from this site.25 Although probably following
an older archetype, the manual is known from manuscripts dating exclusively
from the Roman period.26 It is important to note that some of the hieratic frag-
ments, such as those transmitted in Greek, might date to the second century
CE;27 thus, suggesting a possible coexistence of the Greek and Egyptian versions.
While the reasons behind the choice of Greek may reside in the bilingual
nature of the local community28 and decrease of temple personnel having a
strong command of the Egyptian writing systems, it is likely that this change
also reflects the control exercised on the Egyptian temples and the ritual activi-
ties of these institutions by the imperial authorities.29 The verso side of the same
papyrus roll preserving The Book of the Temple contains a collection of laws
regulating the conduct of the priestly class and listing the punishments des-
tined for transgressors of temple law.30 This collection of rules could also have
been translated into Greek to provide the necessary information to the Roman
authorities, who partially controlled the Egyptian temples’ administration from
the time of Augustus.31
Another interesting testimony although not necessarily directly linked to
the temples’ domain is P.Oxy. XLVI 3285, dating to the second half of the second
century CE.32 The papyrus preserves a Greek translation of a legal code known as
The Legal Manual of Hermopolis,33 which was identified for the first time in a

||
24 P.Washington University inv. 138 + P.Oslo 2; both fragments have been the object of several
studies, although Quack was the first to provide a reliable interpretation of the texts preserved
in these fragments recently. For an introduction to the various interpretations proposed over
the last century, see Quack 2016a, 268–271.
25 Quack 2016b, 109.
26 Quack 2016a, 268.
27 Quack 2016b, 107.
28 Quack 2016b, 116.
29 Mascia 2023a.
30 For an overview on the recent identification of this text, see Quack 2016a, 277–278.
31 Some of the regulations present in this manuscript find parallels in the Gnomon of the Idios
Logos (regulations of the emperor’s private account), which was a handbook containing a list
of legal rulings pertinent to the duties of the idios logos. In particular, it seems that the sections
of this manual dedicated to the rights and duties of the Egyptian priesthood were composed
following the traditional temple law. For an introduction to this handbook, see Speidel 2013.
32 See Rea 1978, 30–38.
33 For an overview of this manual, see Mattha 1975.
48 | Leah Mascia

papyrus roll written in Demotic dating from around the third century BCE.34 The
existence of a Greek translation of this manual concerning mainly Egyptian civil
law, as remarked by Edda Bresciani, was probably meant to allow its consulta-
tion by the Ptolemaic and later Roman authorities.35 The Greek version found at
Oxyrhynchus seems to have been a compendium rather than a literal transla-
tion of this Demotic manual.
A coexistence between texts written in both Egyptian and Greek languages
is also attested in divination practices, which combined elements peculiar to
both foreign and native oracular traditions. Aside from the numerous Greek
oracle tickets published over the last century,36 and a Greek fragment of a dream
book,37 several divinatory compositions written in Demotic have been identified
recently.38 At least some of these textual sources seems to be contemporary,39
underlining how the temple personnel of the Sarapeion,40 the Thoereion41 and
perhaps other sanctuaries42 performed oracular services in both languages
throughout the early Roman phase. The coexistence between Egyptian and

||
34 This handbook found during the investigations of the ancient site of Hermopolis in the
winter between 1938 and 1939 is a collection of rules concerning civil law, including instruc-
tions for the redaction of contracts and other documents. On the story of the transmission of
this text, see Pestman 1983.
35 ‘Non sorprende che il νόμος τῆς χώρας esistesse in traduzione greca, a disposizione e per la
comprensione dei funzionari e degli amministratori di lingua greca, in età tolemaica e poi in
quella romana’; see Bresciani 1981, 202.
36 Like P.Oxy. IX 1213, P.Oxy. VIII 1149, P.Oxy. XLII 3078, P.Oxy. XXXI 2613, P.Oxy. LXXIV 5018
and P.Köln. IV 202 date from the second century CE. P.Oxy. LV 3799, P.Oxy. L 3590, P.Mich. inv.
1258, P.Oxy. LXV 4470 and P.Oxy. VI 923 date between the second and third century CE. A
Demotic oracular question addressed to Osiris and Serapis, possibly dating to the Roman peri-
od, has probably been identified among the papyri kept in Oxford, Quack 2016b, 108.
37 P.Oxy. XXXI 2607 (third century CE). For a re-examination of this text and a possible inter-
pretation of this composition as a treatise of Egyptian origin, see Prada 2016.
38 Quack 2016b, 108.
39 Quack 2016b, 107.
40 Most of the oracular tickets from this city are addressed to the deity patron of this temple.
41 One oracle ticket associated with Thoeris PSI Congr. XVII 14, dating from the Ptolemaic
phase (second–first century BCE), is known from Oxyrhynchus. Despite the absence of direct
attestations ascribable to the Roman period, a document P.Oxy. XLI 2976 (second century CE)
seems to attest the continuity of divinatory practices associated with this deity up to the second
century CE.
42 One oracular question is addressed to the god Thonis (P.Köln. IV 202, second century CE),
which might indicate the existence of a sanctuary associated with this deity. An alternative
solution might be the presence of a chapel or an altar in another temple of this city. The wide
attestation of priests bearing theophoric names associated with Thonis in the Great Thoereion
might suggest the worshipping of this god in the major religious institution of this city.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 49

Greek in the performance of these ritual practices might find various explana-
tions. On the one hand, it could depend on the typology of divination practices
performed, being oracle tickets predominantly attested in Greek, not only from
this site, while divination handbooks continued to be produced to some extent
still in Egyptian. On the other hand, the linguistic choice might have also been
dictated by the necessary adaptation to the needs of an ever-growing multicul-
tural clientele requesting divinatory services.
The production of such a variety of texts also implies the existence of an
educational setting that, as has already been noted, must have been strictly
linked to a local centre of textual production, perhaps the ‘house of life’. Several
are, indeed, compositions among the unpublished Oxyrhynchite material which
might have pursued their function in an educational context.43 Clackson MSS
8.2.5, an Old Coptic wordlist (second century CE), for instance, although barely
intelligible, provides at least an idea of the typology of ‘compendia of lexical
information’ used by the local priests in the early Roman period.44
Aside from the coexistence between languages in the production of docu-
mentary, literary and paraliterary texts associated with the domain of the local
Egyptian temples,45 the site attests to the recurrent interaction between various
scripts in the production of specific textual genres: compositions associated
with the inner ritual domain of temples. The practice at Oxyrhynchus took
shape especially in the combination of a linear text in hieratic script with gloss-
es in Old Coptic.46 These glosses were used to supply punctuation, translitera-
tion and full vocal pronunciation, as testified by various findings offered by
other contemporary Egyptian settlements. In other words, many were probably
pronunciation notes for priests who needed to fully comprehend the text, as
well as recite and perform the religious ritual in the most accurate way. Correct
pronunciation was indeed mandatory for the procedure efficacy; therefore, we
can imagine how these glosses played a key role, especially in a historical phase

||
43 Sections of a Demotic schoolbook, for instance, have been identified. As noted by Quack,
the text includes, among others, a list of birds used for memorising the Late Egyptian alphabet-
ic sequence; see Quack 2016b, 109.
44 Love 2021, 198–199. Old Coptic is based principally on the Greek alphabet with a collection
of signs derived from the Demotic script used to write sounds absent in the Greek language to
transliterate any stage of the Egyptian language. For an introduction to Old Coptic scripts and
texts, see Quack 2017.
45 Although, as already discussed, we cannot be sure that some of the texts examined were
produced outside temples’ settings, like in the case of P.Oxy. XLVI 3285.
46 It is important to consider that 33% of the Oxyrhynchite hieratic texts known so far appear
to have been glossed in Old Coptic; see Love 2021, 176.
50 | Leah Mascia

in which high proficiency in the Egyptian language and scripts, especially hier-
atic, was probably restricted to a minority of temple personnel. In this sense,
according to the evidence offered by Oxyrhynchus, a priest illiterate in hieratic
but having certain knowledge of Old Coptic together with the pronunciation and
intonation of Middle/Classical Egyptian would have been able to recite the text
correctly.47
While glosses had, in principle, a similar function, especially in ritual texts,
the system of glossing was adopted according to different solutions reflecting
the phase of experimentation characterising the Egyptian temple scriptoria
between the late Ptolemaic and early Roman phases. It is especially at Oxyrhyn-
chus that the Egyptian glossed documentation, written mainly in hieratic but
transmitting texts in Middle/Classical Egyptian, remarks the variety of employ-
ments of supralinear glossing in Old Coptic. As noted by Love:

At Oxyrhynchus, script shift in the ritual domain was from hieratic to Old Coptic, but at
one other it was from hieratic to demotic. What can only be stated is that (an) Egyptian
priesthood(s) at Oxyrhynchus considered Old Coptic to have the highest utility in aiding
decipherment, perhaps thereby coupled with ensuring correct pronunciation. This ‘shift
in priority’, however, had the consequence that ‘while the phonetic realisation becomes
even easier’, i.e., literacy in Old Coptic is considerably easier to acquire than literacy in hi-
eratic due to orthographic depth, comprehension of the text’s semantics became even
harder.48

Among the hieratic texts attesting the presence of Old Coptic glosses, several are
included in the corpus of sources classified by Love as texts showing ‘subse-
quent selective supralinear glossing’, namely, concerning only designated sec-
tions of the ritual composition.49 Others attest to what is defined by Love as a
‘subsequent comprehensive supralinear glossing’, which seems part of a pro-
cess generally posterior to the composition of the linear text.50 Another group, in
Love classification, consists of ritual texts characterised by the presence of a

||
47 As recently remarked by Love 2021, 190.
48 Love 2021, 197.
49 This group includes P.EES Oxyrhynchus 29 4B.42/H (13), comprising five fragments of a
ritual for protection featuring no punctuation (second century CE). P.EES Oxyrhynchus 3b 4/9
counts one fragment of a ritual text of unclear content preserving partial glossing with linear
punctuation (second century CE); see Love 2021, 184–187.
50 Kahle MS 2/3, a hieratic ritual text with Old Coptic glosses (early second century CE?) of
uncertain content belongs to this category; see Love 2021, 187–189.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 51

layout planned to leave space for the glossing of the text; therefore, allowing
the inclusion of the glosses contemporary to the linear hieratic composition.51
Aside from hieratic texts, linear compositions in Old Coptic are also attest-
ed. The texts in this script could have found their origins in copies of Old Coptic
glosses, which provided a transliteration of the original linear hieratic or per-
haps even Demotic compositions.52 P.BM EA 1080853 is certainly the most well-
known example of Old Coptic linear text from this settlement. The papyrus pre-
serves a collection of magical texts in Egyptian probably dating to the second
century CE, which production setting remains still debated.54 The composition
consists of a poorly preserved first column written in Demotic, Greek script and
Old Coptic, and a second column in Old Coptic,55 ‘complemented by certain
group-writings in Demotic’, as described by Love.56 The text comprises different
spells interpreted by Quack as incantations to acquire favour and love.57 Among
them, one meant to obtain a πάρεδρος, an ‘assistant in ritual practices’; another
possibly interpreted as a dream-sending invocation, probably calls for the assis-
tance of a deceased, a ‘ghost’ from the necropolis, followed by three spells in
which a deity is invoked. It is important to notice that the title and indication for
the actions are written mainly, as noted by Quack, in Late Demotic. The use of
Demotic was doubtless aimed at clarifying the content of the ritual instructions
to the user of the spell. On the other hand, the spell itself, written in Old Coptic,
probably remained mostly unintelligible to the reader. The interaction between
scripts in this context shows the dual component peculiar to the Greco-Egyptian
ritual production. As observed by Quack:

||
51 P.EES Oxyrhynchus 25 3B.58 M(a), which consists of three fragments of a composition
comprising a ritual for protection involving body parts; see Love 2021, 189–191.
52 On the possible origins of Old Coptic linear texts, see Love 2021, 191.
53 A first interpretation of this composition has been provided, among others, by Dieleman
2004. For further studies in light of the identification of another fragment among the Oxyrhyn-
chus papyri kept in the Sackler Library, see Quack 2010, 83; Quack 2017, 64. The presence of a
reference to Nephthys seems to justify its association with a ritual or temple setting, which, as
we will discuss in the following pages, might be proposed for other texts recently discovered in
the necropolis of Oxyrhynchus.
54 For instance, Love (2021, 198), while stating that the text ‘is safely ascribable to either to the
ritual or temple domain’ rejects Renate Dekker’s interpretation (Dekker 2013, 64) of the text as
belonging to an Egyptian temple library considering the lack of evidence at our disposal; see
Love 2021, 191, n. 55.
55 The copying from an Old Coptic Vorlage seems suggested, according to Love, by ‘the fluent
copying of the hand and precise layout of the manuscript’; see Love 2021, 197.
56 Love 2021, 191.
57 Further bibliography in Love 2021, 194.
52 | Leah Mascia

Ritual texts have two different components. The one is the indication of the aims achieved
by it as well as performance guidelines. For both, semantic clarity is imperative, and in
this manuscript, this is guaranteed by the use of the contemporary vernacular. The second
one is the spell, and that functions by the power of its sound, not by its meaning.58

Alongside the ritual domain, some texts also help us to gain more information
about the use of Egyptian and the intertwining between distinguished writing
systems in a private setting, although most likely restricted to the Egyptian
temples personnel.59 Several letters have been identified as largely written in
hieratic, but with the addition of a few Demotic signs, even though the language
appears as being quite certainly Demotic.60

4 Scripts interactions in context: The Osireion of

Oxyrhynchus
As we have seen, the evidence offered by materials kept in museums’ collec-
tions and other research institutions already provide essential information on
the multicultural context characterising the religious institutions of Roman
Oxyrhynchus and the activities of various ritual specialists, which might have
been linked, at least to some extent, to these organisations.61 However, the ab-
sence of any information regarding their provenance inevitably limits our
knowledge of the settings in which these ritual texts were produced and used. It
is in this sense that the investigations of the temple-catacomb of Osiris, carried
out by the mission of the University of Barcelona in the early 2000s, play an
essential role in concretely observing how different scripts physically coexisted
in an Egyptian temple between the Ptolemaic and Roman phases. The compres-
ence of hieroglyphic, hieratic and Demotic inscriptions in an Egyptian sanctu-
ary certainly does not come as a surprise since each script traditionally an-
swered specific functions in the fulfilment of both administrative and ritual

||
58 Quack 2016b, 118–119.
59 Quack (2016b, 108) reflected on the peculiar use of the writing system and proposed to
identify priests ‘equally at home in both writing systems’ as the authors of these documents. In
particular, a reference to ‘The Scribe of the Book of Per-Medja’, namely, Oxyrhynchus, is pre-
sent on the recto side of one of these documents; see Love 2021, 177, n. 4.
60 See Quack 2016b, 108; Love 2021, 177.
61 See, in particular, the problems in establishing the authorship and production setting of
P.BM EA 10808 discussed in the previous section.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 53

activities. Nonetheless, the Osireion of Oxyrhynchus offers us the rare occasion

to imagine how these different scripts interacted spatially and examine a variety
of inscribed objects in their original context.
The underground building complex, constructed in the early Ptolemaic
phase62 and enlarged in the Roman period, comprised two halls and two galler-
ies meant to host the daily offerings to Osiris and the rituals performed during
the Khoiak festival.63 The investigations conducted over the last few decades
have allowed the identification of a rich repertoire of inscribed materials, which
comprise texts in hieroglyphic, hieratic and Demotic scripts.
The hieroglyphic script was used by the priesthood of Osiris to scribble
short inscriptions on the walls of the underground structure. We might imagine
one of these priests carving a short prayer in honour of Osiris64 in the main hall,
hosting the colossal statue of this deity, perhaps during a break from the daily
religious services. Ritual formulae in hieroglyphic script were painted on the
back of a group of Osirian figurines made of clay associated with the liturgical
activities performed inside the building.65
While the latter evidence helps us in visualising the temple-catacomb of
Osiris as an inscribed space, it is only by entering the galleries of this sanctuary
that we would have seen a clear physical interaction between scripts in the
production of single ritual texts.
Gallery 1A, built at the beginning of the Ptolemaic phase, hosts thirty niches
progressively erected, or at least inscribed, under the reign of several Lagid
rulers.66 The gallery had a precise function in daily cultic activities and specific
annual festivities. The last stages of the Khoiak festival, which was celebrated
on the thirtieth day of the month of Khoiak, the fourth month of the season of
flood, were performed in this area. On this occasion, the priests carried out the

||
62 Although some evidence seems to suggest that the Ptolemaic structure had been built upon
the ruins of another religious complex, perhaps associated with a temple known as the Pr Khf
dating back to the Late Dynastic phase; see Mascort Roca 2018, vol. 1, 25–26.
63 For an introduction to the evidence for this religious celebration in the New Kingdom, see
Eaton 2006.
64 The various graffiti identified preserve invocations or brief prayers addressed to the main
deity of the sanctuary. An inscription left on the door-jamb on the south-east wall at the 2C and D
halls entrance is the best preserved of these. The text recites a religious formula alternatively
interpreted as ‘The life belongs to the god Osiris’ or ‘The divine life of Osiris’, see Mascort Roca
2018, vol. 1, 48.
65 See Mascort Roca 2018, vol. 1, 92–93.
66 The last recorded are Berenice IV (c. 79–55 BCE) and Cleopatra V–VI Tryphaena (c. 95–57 BCE);
see Coulon 2018, 187.
54 | Leah Mascia

burial of mud figurines of Osiris, which had been crafted in the previous year,
probably using the silt of the sacred lake located in the precinct of this sanctu-
ary.67 This ritual is described in the Mysteries of Osiris, which offers detailed
instructions on how the ritual was performed.68 As theorised by Laurent Coulon,
the thirtieth day of the month of Khoiak marked the end of a monthly cycle,
which echoed the cycle symbolised by the thirty niches of the galleries. The
temple-catacomb of Oxyrhynchus functioned as the Osirian necropolis of Kar-
nak, the niches being inscribed in the staging of the cyclical rebirth of the god of
the dead.69 Texts are painted in black on the lintels of the thirty niches placed at
the beginning of the gallery and indicate the date of burial of the Osirian figu-
rines.70 The use of a standard chronological system, based on the year of the
reign of the ruling Lagid king, also marked the assimilation between Osiris and
the ruling pharaoh. The ritual texts are written in hieratic script, but the date of
the deposition of the Osirian figurines is occasionally rendered in Demotic.71
Dates are also recorded in hieratic script in the latest niches, where only the
formula ḥꜣ t-sp is expressed in Demotic.72
Plaques sealing the niches were destined to close the burials of the Osirian
figurines;73 the texts, which also record the statuettes deposition, are traced in
black ink and combine hieratic, hieroglyphic and Demotic scripts.74 The date is
generally expressed in cursive hieroglyphs; however, it is sometimes associated
with a version rendered in Demotic. The switching to Demotic might depend on
the specific function that this script assumed since its creation. As indicated by
the same Greek term Demotic (‘popular’), this script was originally employed for
writing administrative texts, while hieratic and hieroglyphic scripts remained
tied to the production of literary and paraliterary compositions. Although al-
ready in the Ptolemaic phase, Demotic was often used in the production of ritual

||
67 For a description of the sacred lake of the Osireion of Oxyrhynchus and associated evi-
dence, see Mascort Roca 2018, vol. 1, 34–36.
68 For an introduction to this religious celebration, see Chassinat 1966 and 1968.
69 Coulon 2018, 187 and Coulon 2015.
70 For an in-depth study of these inscriptions, see Coulon 2018.
71 Coulon 2018, 171–172.
72 As noted by Coulon 2018, 172, the joint use of the hieratic and Demotic scripts in the pro-
duction of the ritual texts traced in this gallery is quite remarkable, although attested by evi-
dence from other settlements. For a detailed bibliography see Coulon 2018, 172, n. 8.
73 For a discussion of these testimonies, see Coulon 2018, 182–187.
74 As remarked by Coulon 2018, 183, ‘le hiératique est progressivement concurrencé par
l’emploi des hiéroglyphes cursifs. La date, donnée une première fois en hiéroglyphes cursifs,
peut être doublée par une ligne de démotique’.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 55

texts,75 this script switching might reflect the adherence to a more archaic scrib-
al habit.
The small funerary furniture buried inside these niches provides additional
evidence.76 The objects are predominantly small limestone boxes, which pre-
served a magical bulla inside77 and limestone cones featuring representations of
the goddess Neith.
As noted by Jean-Claude Goyon, the interpretation of the script remains
problematic:

The cursive inscriptions are not hieratic of the classical type but have a ductus sometimes
close to that of ‘abnormal’ hieratic, either ‘old’ demotic, or the range of demotic of the
second and first centuries BCE.78

The poor state of conservation of most of these exemplars does not allow a clear
interpretation of this ensemble of evidence. However, the ductus suggests the
production of these texts between the beginning of the Ptolemaic and the early
Roman period. According to the inscriptions preserved on the cones and the
limestone boxes containing bullae and sealed by pyramidal lids, it appears that
most of these objects were associated with the arrangement of the orientation of
each loculus containing the Osirian figurines. They refer to the deities protectors
of the four cardinal points, and others to Osiris as the guardian of the West.79
Slabs preserving inscriptions mixing hieroglyphic, hieratic and Demotic scripts
could have also sealed the three niches of the nearby Roman gallery (2E).80 This
area, built at the beginning of the imperial phase, was meant to host the contin-
uation of the Khoiak rituals in the early Roman period. Inscribed limestone

||
75 On the use of Demotic for the production of ritual texts in the Ptolemaic and Roman peri-
ods, see Quack 2012.
76 For a preliminary study of these objects, see Goyon 2018.
77 According to the Mysteries of Osiris, the bullae preserved inside these containers were meant
to protect Osiris and served to keep away Seth and other malevolent entities, thus, promoting the
completion of the ritual of the renaissance of the deity; see Chassinat 1966, 51–52.
78 Goyon 2018, 197: ‘Les inscriptions cursives ne sont pas hiératiques de type classique mais
ont un ductus parfois proche de celui du hiératique “anormal”, soit démotique “ancien”, soit
du ressort du démotique des IIe/Ier siècles avant notre ère’ (my translation).
79 Various deities associated with different cardinal points are mentioned in the inscriptions
preserved on these portable written artefacts: to the south, Ammon and Montu; to the north,
Shu and Tefnut; to the west, Neith and Wadjet; to the east, Sekhmet and Bastet; see Goyon
2018, 197.
80 However, given the poor state of preservation of this area, this hypothesis cannot be prov-
en. A description of this temple area is in Mascort Roca 2018, vol. 1, 74–78.
56 | Leah Mascia

boxes containing bullae of the same typology as those traced in the Ptolemaic
gallery were also found inside one of these niches (no. 2).81

5 The multilingual funerary landscape of Roman

Oxyrhynchus and beyond
The so-called High Necropolis of Oxyrhynchus offers a unique perspective on
the complex interaction between languages and scripts in the same spatial con-
text. We should begin looking at the appearance of the funerary landscape of
this city after the Roman conquest to fully comprehend how foreign and native
cultures were intricately intertwined in the Roman period.82 Imagine walking
through this cemetery in ancient times: we would have been struck by the het-
erogeneity of funerary monuments dotting the landscape. Barrel-vaulted tombs,
platform tombs with pyramidal superstructures, hypogea and coffins inspired
by Roman models reflected the multicultural panorama of this Egyptian settle-
ment. The intermingling between different ethnic groups might have also been
perceived by looking at the compresence of Greek,83 Latin84 and Egyptian85 fu-

||
81 Mascort Roca 2018, vol. 1, 74–75.
82 The choice of concentrating the discourse to this period is dictated by the fact that the
Roman phase of this cemetery is certainly archaeologically more well-preserved than the Ptol-
emaic phase. Nevertheless, it should be remarked that many tombs were probably in use al-
ready in the Ptolemaic period, and various Ptolemaic and Roman funerary complexes probably
coexisted for centuries.
83 For an overview of the first funerary inscriptions discovered at the site by Bernard Grenfell
and Arthur Hunt, although probably quite late in date (fourth to fifth century CE?), see Parlasca
2007, 95, 100–101. An inscribed female statue on a pedestal found in Sector 2D of the necropo-
lis is certainly earlier in date (second/third century CE), see Padró et al. 2012, 4. A date to the
late second or early third century CE might be attributed to an unpublished funerary statue
preserving a short epitaph painted in red discovered in the excavations carried out in 2020; a
brief mention of this discovery is in Mascort Roca et al. 2020, 8. Another inscribed stela found
at the entrance of the funerary chamber of Tomb 12 in memory of a child name Theon, probably
dating between the third and fourth century CE, was discovered in 2009. On this discovery, see
Padró et al. 2009, 12. A study of this inscribed stela has been published in Campillo and Pie-
drafita 2010. Further reflections on the inscribed funerary stelae discovered at Oxyrhynchus in
Codina Reina 2015, 172–173.
84 Although only a few in number, several exemplars of funerary epitaphs written in Latin
have been discovered during the investigations of this settlement. Fragments of Latin inscrip-
tions, for instance, possibly funerary stelae although too fragmentary to identify their exact
function, have been discovered in the High Necropolis (Sector 2D); see Mayer-Olivé 2015. A
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 57

nerary inscriptions standing in front of the tombs86 or located inside the funer-
ary chambers.87 However, it was entering inside these burial spaces, in the inner
funerary halls of these structures, where we would have perceived more closely
how the long coexistence of Egyptians, Greeks and Romans and other ethnic
communities have concurred in shaping the local funerary customs. This syn-
cretic repertoire of beliefs is reflected in all the elements characterising the mor-
tuary landscape of Roman Oxyrhynchus. Decorative programmes still betrayed
the adherence to models finding their roots in the Pharaonic tradition, ex-
pressed in the broad attestation of offering scenes depicting the deceased stand-
ing in front of an enthroned Osiris surrounded by a court of Egyptian deities.88
We would have often seen Greek graffiti witnessing the names of the deceased,89
sometimes associated with brief invocations, on the walls of several of these
tombs.90 At least on one occasion, the owner (or owners) of the Roman Tomb 40
seems to have commissioned the production of a Latin inscription painted in
red, which originally surrounded the walls of the main funerary chamber of this

||
stela representing the deceased dressed as a Roman citizen and preserving a Latin inscription
painted in red was found in Sector 26; see Padró et al. 2014, 14–15. Another fragmentary Latin
funerary inscription mentioning a libertus was discovered in the High Necropolis during the
investigations carried out in March 2020. Other unpublished Latin inscriptions are currently
kept in the archives of the Egyptian Antiquity Service in el-Bahnasa.
85 In truth, only a few inscriptions are known so far, which are predominantly dated to the
Ptolemaic phase. The most well-preserved is a stela which was originally part of the Metropoli-
tan Museum of Art collection and acquired in the early 1950s by Macquarie University (Sidney).
I am grateful to Dr Marsha Hill and Dr Niv Allon for providing information on this artefact.
Apart from these inscriptions, at least one fragmentary stela possibly originally preserving an
Egyptian text has been discovered in a Roman tomb in Sector 26; see Pons Mellado 2016, 167.
86 A funerary statue representing the deceased reclined on a triclinium was found in front of a
tomb structure in the course of Evaristo Breccia’s investigations in the early 1930s. Photograph-
ic documentation of this discovery is now kept in the Archivio Breccia of the University of Pisa.
I am grateful to Dr Flora Silvano for allowing me to examine the documentation collected
during Breccia’s investigations at Oxyrhynchus.
87 As in the case of the stela of the young Theon; see Campillo and Piedrafita 2010.
88 Similar to those in Tomb 3, Tomb 18 and Tomb 21; see Pons Mellado 2016, 163–168, 170–171.
89 Tomb 2, also known as ‘Uraeus Tomb’, preserved a charcoal graffito recording the name
Demetrios on the northern wall of chamber no. 1. Two carved graffiti in chamber no. 2 pre-
served the name Didymas. A preliminary discussion of these graffiti is in Piedrafita 2011, 64–66.
Two charcoal inscriptions preserving the name Pausanias were discovered at the eastern and
western sides of the entrance to Tomb 11 in 2009; see Piedrafita 2011, 71–72. A description of
this tomb preserving the mummified bodies of three individuals is in Padró et al. 2009, 7–11.
90 A graffito recording the formula ‘Marcus, be brave’ was individuated on the southern wall
of the funerary chamber in Tomb 11. Further discussions on this inscription are in Piedrafita
2008, 136, 143–144.
58 | Leah Mascia

structure.91 However, as has already been mentioned, Egyptian traditional mo-

tifs were still privileged in the decoration of most of these tombs. A short in-
scription written in Demotic, painted on the margins of a scene depicting the
deceased surrounded by Egyptian gods in Tomb 29,92 represents the only evi-
dence found in its original archaeological context of the use of this script in the
funerary spaces of Roman Oxyrhynchus. Getting closer to the deceased and
their funerary paraphernalia, we would have seen in the iconography of coffins
and mummy cases the undeniable influence of both indigenous and Greco-
Roman motifs. Aside from a few exceptions,93 these funerary artefacts are main-
ly inscribed in Greek and preserve epitaphs in honour of the deceased having
the same formulary of the funerary inscriptions attested on stelae.94 Interesting-
ly, these texts might have also been found scribbled on the leg95 or foot96 of
some Oxyrhynchite mummy cases. Even votive objects deposited inside the
tomb, from the traditional wedjat eyes97 to terracotta figurines,98 underline the
overall inspiration to a broad multicultural repertoire of religious beliefs.

||
91 Several inscribed plaster fragments preserving what seems to be a Latin inscription painted
in red were discovered among the debris of Tomb 40 in the course of the excavations season
2020. A brief discussion on this funerary structure is in Mascort Roca et al. 2020, 9–11.
92 For a report on the discovery of this barrel-vaulted tomb in the High Necropolis of Oxyrhyn-
chus, see Padró et al. 2010, 7–8.
93 Several painted wooden coffins preserving a rich decoration featuring short hieroglyphic
inscriptions have been identified in the hypogeum (Tomb 17) of Sector 2A. However, their
chronology remains still uncertain and might date no later than the late Ptolemaic or early
Roman phase; for an introduction, see Castellano i Solé 2011–2012, 44–45.
94 The slab sealing the anthropoid coffin buried in Tomb 18 (Sector 26) identifies the deceased
as Polydeukes. For a description of the tomb and discovery of the inscription, see Pons Mellado
2016, 163–167. For further discussions on this text, see Piedrafita 2011, 67–69. An embossed
inscription on a plaster mummy case has been discovered in the course of the 2020 excavation
campaign. For a preliminary description of this discovery, see Mascort Roca et al. 2020, 14–15.
The text is currently under study by the writer.
95 As in the case of the plaster mummy case discovered in 2020, preserving a Greek inscription
scribbled on the right leg of the deceased.
96 Similar to the mummy of Harpochratiaina; a short inscription recorded the name, patro-
nymic, and age of death of the young woman is in the area of the cartonnage mummy case
corresponding to the feet, see Piedrafita 2008, 136, 138.
97 A faience amulet of this type was discovered in a mudbrick structure sealing the entrance of
Tomb 17 in Sector 2A. The building was reused as a burial space in the early Christian period
(fourth to fifth century CE?), see Castellano i Solé 2011–2012, 43.
98 Several terracotta figurines have been discovered over the last years of the archaeological
investigations carried out in the High Necropolis, most of which remain unpublished. Other
terracotta figurines discovered during Grenfell and Hunt’s excavations were distributed among
various museums and research institutions.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 59

Despite the poor evidence known to date, the Egyptian language seems to
have been still employed in the early Roman period in the local funerary admin-
istration.99 Aside from the numerous Greek mummy labels spread among vari-
ous museum collections,100 one wooden fragment preserving a bilingual text
testifies to a practice well known in the Roman phase,101 which foresaw the writ-
ing of prayers or specific sections of the text only in the Egyptian language.102 It
is for this reason that this ‘mummy label’103 records information essential for the
identification of the deceased in Greek, but seems to abruptly switch to the use
of Demotic in writing a closing formula.104 These objects were traditionally at-
tached to the body of the departed, and they seem to have been found partly in
the area of the necropolis at Oxyrhynchus,105 but mostly discarded in the rub-
bish dumps surrounding the ancient settlement.106
Defining the cultural and ethnic identity of the individuals buried inside
these tombs remains extremely difficult. The wide attestation of Greek names
recorded in graffiti, on mummy cases, labels and coffins, probably reflects the
existence, aside from Greek settlers and their offspring, of a wealthy and influ-
ential group of native inhabitants who have partly embraced the customs and
ways of living of the Greek and Roman conquerors. It is less difficult to identify
Roman citizens, at least when we can trace names of Latin origin in funerary
artefacts. The Romans largely adopted the Egyptian funerary habits; however,

||
99 On the use of mummy labels as identification tools see, for instance, Martín Hernández 2011.
100 Several are now kept in the Penn Museum, Philadelphia: Inv. no. E11734; Inv. no. E11735;
Inv. no. E 11736; Inv. no. E 11737. An important corpus, soon to be published by the writer, is
also preserved in the Metropolitan Museum of Art, New York: Inv. no. 97.4.36; Inv. no. 97.4.37;
Inv. no. 05.4.186; Inv. no. 05.4.187; Inv. no. 05.4.180; Inv. no. 05.4. 188. My gratitude goes to Dr
Niv Allon for allowing me to study and publish the Oxyrhynchite mummy labels kept in the
Metropolitan Museum collection. Another uninscribed mummy label (Inv. no. EA 50146) is now
kept in the British Museum, in London.
101 Inv. no. 05.4.179; the artefact is currently under study by the writer.
102 However, it is important to stress that both the Greek and Demotic versions generally
provided similar information on the deceased identity.
103 It still remains unclear whether the fragment belonged to a mummy label or a coffin.
104 Still, the interpretation of this last section remains tentative given the poor state of con-
servation of the text; further studies might lead to future revisions.
105 Inv. no. 97.4.36 and Inv. no. 97.4.37 were discovered at the beginning of the excavation
campaign 1896–1897 when the investigations were concentrated in the area of a necropolis.
106 The Penn Museum corpus, for instance, was discovered in the excavation campaign of
1904–1905, which was concentrated in the investigation of several of these dumps. Inv. no.
05.4.179, Inv. no. 05.4.188, Inv. no. 05.4.180 and Inv. no. 05.4.186 were also discovered in the
1904–1905 excavation campaign and are among the artefacts kept in the Metropolitan Museum
of Art, New York.
60 | Leah Mascia

they did not miss the chance to occasionally reaffirm their identity, as testified
by a ‘mummy label’ preserving a Latin epitaph107 that still does not find parallels
among the objects of the same typology published so far. If the interpretation of
this object as a mummy label is proven to be correct, this artefact would repre-
sent an outstanding and unique example of the assimilation between the Ro-
man and Egyptian traditions, which found its expression in a variety of declina-
tions.
Mixing languages and scripts might have also taken the form of a strange
combination of documentary, literary and paraliterary texts in Greek and Egyp-
tian folded together and deposited on the chest, abdomen or pelvic area of sev-
eral deceased identified in the Roman Tomb 42.108 While this odd textual mé-
lange might suggest, at first glance, their identification as cartonnage papyri,
their discovery in small packages sealed by clay devices identified as embalm-
ers seals, one of which preserves a hieroglyphic inscription,109 suggests the per-
formance of a practice mimicking the traditional deposition of funerary compo-
sitions.110 Another inscribed clay embalmer seal was found in association with a
papyrus, this time still found accurately sealed, preserving a Greek magical
text.111 Indeed, when speaking of the necropolises of Oxyrhynchus, particular
attention should be devoted to the magical artefacts discovered in this area. The
presence of these objects in the cemeteries of this city112 is certainly not a sur-
prise since magical practices are attested in funerary contexts across various

||
107 Inv. no. 05.4.188. The identification of the object as a mummy label, while suggested by
both the shape of the object and the content of the inscription, is still uncertain. The artefact
will be published soon by the writer.
108 However, several deceased were found in association with only one text, like Potamon
(UE 36192) who was buried with a Greek sealed documentary text placed on his abdomen. For a
similar discovery see Grenfell 1897. Among the deceased buried with papyri are 36181, 36182,
36186, 36192 and 36204.
109 36204.
110 A comprehensive study of this corpus of texts and their context of discovery will be pub-
lished soon by the writer.
111 The papyrus (36345) was discovered in Tomb 52 (Sector 36) on 28 November 2021. The
folded packet was found on the abdomen of the mummy of a child buried in the barrel-vaulted
tomb sometime between the first and second century CE. The identification of the composition
as a Greco-Egyptian magical text has been confirmed by studies carried out by the writer be-
tween November and December 2022 in the laboratory of the mission. Over 20 charaktêres
visible through the folder fractures have been identified; the text is currently under study by
the writer. For a short introduction on this discovery, see Mascort Roca et al. 2021, 11–12, 23.
112 For an overview of the inscribed objects associated with magical practices performed in
the necropolis of this city, see Mascia 2023c.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 61

provinces of the Roman Empire. The reason behind the choice of the mortuary
landscape for the performance of magical procedures resides in the idea that
necropolises represented a liminal space between the world of the living and
the dead. Furthermore, being outside the borders of the urban space and the
administrative centre of the city, they were probably seen as more conceivable
places for the activities of ritualists.113 Among the artefacts found in the archaeo-
logical investigations of this settlement114 is the discovery in 2020 of a wooden
amulet in Tomb 40 of the High Necropolis dating to the late antique phase.115
The object can be interpreted as a so-called ‘Bous’ amulet, a type of protective
device widely attested in Egypt from the late antique period onwards.116 The
Greek voces magicae are surrounded by a series of charaktêres and what seems,
at first glance, to be the name of God written in the Coptic language.117 Apart
from these recent discoveries, at least one other written artefact found during
Grenfell and Hunt’s 1904–1905 excavation campaign could be associated with a
mortuary context in light of the content of its text. The object, measuring 7.1 × 2 cm,
is a wooden tablet,118 which seems to preserve a small section of what was origi-
nally a much more complex and larger text of applied magic, possibly a binding
spell used in an execration rite.119 The coercive nature of the text and its use in a
burial context is suggested by the adoption of a terminology recurrent in the
production of such incantations. The most well-preserved section of the ritual
text is written on the recto side in Coptic. Nonetheless, the composition presents
several misspellings, which suggest that the writer, having a certain knowledge
of Greek, mixed up the two languages.

||
113 On this aspect, see Wilburn 2012, 249.
114 In particular, a clay tablet preserving a drawing of the god Seth associated with Greek
magical voices was discovered in 1993, see Piedrafita 2010. For the reinterpretation of the text
as a sequence of magical voices, see Mascia 2023c, 1140. An ostracon preserving a Greek protec-
tive spell was discovered in Tomb 40 in the course of the excavation campaign held in 2020;
the object is currently under study by the writer.
115 The artefact found on 26 February 2020 is currently under study by the writer; for a brief
mention of this discovery, see Mascort Roca et al. 2020, 29.
116 For an introduction to this typology of objects, see Sijpesteijn 1981; Menci 2007.
117 Some doubts remain on the interpretation of the text given its poor state of conservation.
118 Inv. no. 05.4.182. As with the other aforementioned objects preserved in the Metropolitan
Museum of Art, my gratitude goes to Dr Niv Allon for allowing me to study this object. An
edition of this text will be published soon by the writer; for a short introduction to this artefact,
see Mascia 2023c, 1139.
119 To the best of my knowledge, only one wooden object preserving an aggressive spell
meant to be deposited in a burial context has been published so far, see Boyaval 1974.
62 | Leah Mascia

6 A brief note on the Jewish and Nabataean

communities of Oxyrhynchus
Papyri, particularly documentary texts, offer significant evidence of the interac-
tion between Greek and other foreign languages in this city.120 This is the case of
a receipt likely belonging to a Nabataean merchant living at Oxyrhynchus
around the second century CE.121 The presence of traces of Greek letters on the
upper margin of the papyrus indicates that the fragment belongs to a bilingual
contract, where the Nabataean text probably consisted of a subscription or a
statement closing the document. A contract BL Or 9180D dating from the fourth
century CE written in Greek preserving a note in Hebrew also comes from the
same city.122

7 The multilingual landscape of Christian

Oxyrhynchus
The integration of archaeological and textual evidence provides a new perspec-
tive on the multilingual panorama of this settlement at time of the rise of Chris-
tianity.
Despite the predominance of Greek sources published so far, which, none-
theless, offer essential information on the social and religious panorama of this
settlement in the late antique phase, recent findings underline the importance
that the Coptic language played in the local textual production. Sarah Clackson
stressed the need to revise our understanding of the role played by the Egyptian
language in Christian Oxyrhynchus in an essential contribution in the early

||
120 Despite the importance of Greco-Latin documentary texts for understanding the impact of
the arrival of the Roman settlers in this city, these bilingual sources are excluded from the
present discussion.
121 Healey 2004.
122 For a short description of this text see https://www.bl.uk/manuscripts/FullDisplay.aspx?
ref=Or_9180_D (last accessed 12 March 2023). For an overview of the literary and paraliterary
texts in the Hebrew language associated with the Jewish community of Oxyrhynchus, see
Piotrkowski 2018.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 63

2000s devoted to the Oxyrhynchite Coptic papyrological material.123 This

perspective can be now substantiated by the wealth of material discovered in
over more than thirty years of archaeological investigations carried out at the
site by the mission of the University of Barcelona.124 These discoveries help us to
understand how Coptic and Greek coexisted and interacted in the same envi-
ronment.
As in other contexts and cultural settings discussed so far, we will speak
predominantly in terms of spatial interaction; even though several testimonies
will underline the recurrent linguistic intertwining in the production of single
written testimonies. This phenomenon, attested throughout the Christian reli-
gious landscape of Oxyrhynchus, remarks how pilgrims visiting this cultic cen-
tre from various areas of the Egyptian lands, local inhabitants and members of
the Oxyrhynchite religious institutions were frequently at home in using the two
languages.

7.1 The funerary houses

A large building was identified in the High Necropolis of Oxyrhynchus in the
early 2000s. The archaeological interventions conducted in this complex led to
its interpretation as a structure associated with the cult of the dead dating be-
tween the early fifth and seventh century CE.125 A pavilion might had the func-
tion of a triclia or apparatorium, i.e. funerary banquet hall. These structures
were traditionally devoted to events held in honour of the deceased in the Ro-
man culture, also attested in the Christian tradition.126 While this type of build-
ing finds parallels across the Mediterranean area,127 the Oxyrhynchite complex
can be compared to only a few structures discovered in other Egyptian settle-
ments.128

||
123 Clackson 2007. Over four hundred Coptic papyri have been identified by Clackson in the
Sackler Library of Oxford and other institutions. The earliest evidence counting private letters,
documents, and literary texts date to the fourth century CE.
124 A consistent number of unpublished Coptic papyrological and epigraphic evidence kept in
the archives of the Egyptian Antiquity Service at el-Bahnasa should be added to the materials
discussed in the following pages, currently under study by the writer.
125 For an introduction to this religious complex, see Subías Pascual 2008.
126 Subías Pascual 2008, 55. On this subject, also see Krautheimer 1965, 34, 36, 38.
127 For an introduction on the ritual practice of funerary banquets in early Christianity, see
Jastrzębowska 2019.
128 In the necropolis of the nearby Antinoopolis (modern el-Sheikh Ibada), for instance, a
building probably erected to answer similar religious needs, the so-called ‘Peristylbaus’, has
64 | Leah Mascia

The discovery of the so-called Funerary Houses of Oxyrhynchus is not only

archaeologically relevant since it offers a deeper understanding of the funerary
practices peculiar to this historical phase, but it is also vital for the identifica-
tion of an important corpus of graffiti (i.e. dipinti)129 painted on the walls of this
cultic space.130 Most of them consist of votive inscriptions often enclosed in
tabulae ansatae left by various devotees visiting the building and short epitaphs
in memory of numerous deceased, which are presumably associated with the
main function of this complex.131 The corpus seems to include mainly Greek
inscriptions; however, several graffiti appear to bear clear influences of the
Coptic language.132

7.2 The basilica of St Philoxenos

The investigations of the mission of the University of Barcelona led to the dis-
covery of another Christian building identified in the Sector 24 of the High Ne-
cropolis in 2009.133 Further analyses of archaeological, papyrological and epi-
graphic evidence led to the identification of this complex as the basilica of St
Philoxenos,134 built around the fifth century CE,135 presumably on the ruins of a
pre-existing religious structure.136 This religious institution, previously known

||
been identified over the last few years. For an introduction, see Grossmann 2008. Another
building meant to answer similar functions, known as the ‘Edificio Comunitario’, has also been
identified in the necropolis of el-Bagawat; see Grossmann 1982, 78–79; Cipriano 2008, 74–83.
129 This type of evidence is traditionally classified in classical epigraphic studies, as dipinti.
However, the term graffiti, especially in Egyptological studies, is often employed for both
scratched and painted inscriptions. While bearing in mind this traditional terminological
distinction, I will here adopt the term graffiti to discuss painted, charcoal and scratched in-
scriptions.
130 See Piedrafita 2003.
131 Piedrafita 2003, Piedrafita 2008, 142–143.
132 Piedrafita 2003, 37–38.
133 Padró et al. 2009, 15–18.
134 Padró, Martínez and Piedrafita 2018. For further reflection on the identification of this
religious building see Martínez García and Mascia 2023.
135 On the dating of this religious building see Martínez García and Mascia 2023.
136 A possible identification of this pre-existing building with the Sarapeion of Oxyrhynchus
has been advanced in the past few years; see Padró, Martínez and Piedrafita 2018, 716–717. This
hypothesis seems suggested by the presence of an oracular cult inside the Christian building
that could have inherited the role played by the Sarapeion in the Greco-Roman period. Howev-
er, the surviving archaeological evidence do not provide sufficient information to verify this
interpretation.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 65

only in light of several pieces of papyrological evidence,137 was an important

oracular centre attracting pilgrims from all over the Egyptian territory.138 The
existence of an oracular cult associated with this saint at Oxyrhynchus was
suggested by the identification of several Greek oracular tickets.139 The perfor-
mance of divination practices in the recently discovered building, entitled to
this saint, is suggested by the examination of the Coptic and Greek graffiti rec-
orded in the crypt and annexed chambers. The presence of Coptic graffiti pre-
sumably witnessing the performance of divination practices in the basilica
opens up a new perspective on the ritual procedures performed in this sanctu-
ary, suggesting that divination services were not only conducted in Greek, but
also in Egyptian.140 This linguistic compresence and interaction is reflected in
various practices, from graffiti recording the visit of devotees to this cultic
space, to the production of funerary stelae.
Over two hundred figural and textual graffiti have been recorded in the area
surrounding the basilica superstructure141 and the underground area.142 The
current study carried out on this broad corpus of sources seems to demonstrate
how the walls of the basilica superstructure were originally covered by char-
coal, painted and scratched graffiti made by pilgrims in the course of their visit
to this religious complex. Apart from a few Greek testimonies, the use of Coptic
seems to prevail in this area, although it was occasionally used to transliterate
anthroponyms in other languages.143 On the other hand, Greek and Coptic graffi-

||
137 P.Oxy. LXVII 4617 (400–499 CE); P.Oxy. LXVII 4620 (475–550 CE); P.Oxy. XVI 1950 (487 CE);
PSI VII 791 (500–599 CE); P.Oxy. XI 1357 (535–536 CE); P.Oxy. XVI 2041 = P.Cairo 10122 (500–699 CE);
P.Lond. V 1762 (500–699 CE).
138 For an introduction on the oracular cult of St Philoxenos, see Papaconstantinou 2001,
336–338. The presence of pilgrims arriving from various areas of the Egypt is indicated, among
others, by the attestation of various Coptic dialects.
139 P.Oxy. VIII 1150 (sixth century CE); P.Oxy. XVI 1926 (sixth century CE); P.Rendel Harris 54
(sixth century CE).
140 A Coptic divinatory composition from Oxyrhynchus has been recently identified. The
editor has proposed a possible association of this ritual text with the local oracular cult of
St Philoxenos; see Kocar 2019.
141 The stone blocks pertaining to the main building were predominantly reused to fill the
underground area after its destruction. For an introduction to the graffiti (dipinti) identified on
these stone slabs see Martínez García and Mascia 2023.
142 Graffiti have been recorded in the crypt (Espacio I), the hallway (Espacio II) and two of the
four annexed chambers (Espacio III and IV). The other two annexes were completely destroyed
and only a few graffiti, predominantly pertaining to the pre-existing Greco-Roman building,
have been recorded.
143 Coptic was used to transliterate an Arabic name, for instance, on the stone slab inv. no. 470.
66 | Leah Mascia

ti in the inner chambers of the basilica appear to testify to the coexistence of the
two languages in the production of graffiti in the earliest stages of the life of this
religious institution. In this sense, graffiti in both languages could have been
made at the same moment by individuals simply choosing to write in different
languages.
The area of the crypt where the balance is, nevertheless, still in favour of
Coptic sources seems to have been a privileged space for graffitiing practices. It
is in this area that we find the most consistent number of pieces of figural and
textual evidence. The variety of writing skills of the individuals engaging in
graffitiing practices seems to indicate the authorship of members of this reli-
gious institution and lay people. This relationship is overturned in the annexed
chambers, where Greek seems to prevail over Coptic sources. The partial chron-
ological overlap between Greek and Coptic graffiti is also indicated by the attes-
tation of writings in which elements of Coptic and Greek are combined. The
intertwining between the two languages is attested, among others, in the inclu-
sion of closing formulae in which a switch into another script occurs. A Greek
graffito scratched on the wall of one of the annexed chambers of the under-
ground gallery of the basilica is a votive inscription in which Biktor, Doxia and
Phaustine ask for St Philoxenos’s protection.144 While the text is written in
Greek, the graffitist uses the Coptic letter ϩ to write the word ‘amen’, which is
used here as a closing formula.145 Tracing the influence of the Greek language in
text written in Coptic is more common, reflected in misspellings and the inclu-
sion of elements alien to Coptic. The frequent occurrence of this phenomenon
seems to demonstrate that many of the authors of these graffiti were at home
using both Greek and Coptic, and even the more experienced writers146 occa-
sionally mixed up the two languages when writing these ephemeral texts.
A compresence between Greek and Coptic would have also been probably
perceived visiting the necropolis built in proximity to the basilica of St Philo-
xenos. Indeed, apart from the Greek evidence published so far,147 the corpus
counts a considerable number of Coptic epitaphs. Seven Coptic funerary stelae
have been identified in this area, several of which seems to date between the

||
144 For a first mention of this graffito, see Padró, Martinez and Piedrafita 2018, 710. An edition
of the text has been provided in Delattre, Dijkstra and van der Vliet 2020, 401.
145 As noted in the re-edition, commenting on the script-switching choice of our scribe, ‘He
shows himself to be only moderately familiar with Greek orthography’; see Delattre, Dijkstra
and van der Vliet 2020, 401.
146 This phenomenon is also recurrent in graffiti allegedly written by trained writers.
147 Mascia and Martínez García 2021.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 67

sixth and the seventh century CE.148 According to the palaeographic analyses
carried out to date, at least some Greek and Coptic inscriptions are probably
contemporary, a consideration that might be extended, as we have seen, to the
written evidence provided by the main religious building.

7.3 The so-called Byzantine Fortress

At the end of this overview of the multilingual panorama of Oxyrhynchus be-
tween the Greco-Roman and late antique phases, I will briefly examine another
archaeological context which provides additional evidence of the spatial coex-
istence between languages in the Oxyrhynchite religious institutions. This reli-
gious complex149 was identified in the early 2000s in the suburban area of this
site. The so-called Byzantine Fortress (i.e. Fortaleza Bizantina) was then the
object of extensive investigations between 2005 and 2010.150 Despite its original
interpretation, the area preserves the remains of a cemeterial and reliquary
church located at the centre of the fortress, two chapels alongside its eastern
façade, and the residence of a monastic community.151 In addition to the discov-
ery of several funerary inscriptions in the Greek language152 and Arabic graffiti,
the latter certainly later in date,153 the area preserves an important corpus of
Coptic graffiti (dipinti).154 Seventeen textual graffiti were discovered in the area
of the apse during the 2010 excavation campaign. However, the religious com-
plex certainly counts a broader collection of figural and textual graffiti, which
only future investigations will lead to a comprehensive evaluation.155

||
148 This epigraphic material is currently under study by the writer, in collaboration with José
Javier Martínez García, who is overseeing the examination of the archaeological context.
149 On the date of this religious structure, see Subías Pascual 2020.
150 For a preliminary report, see Subías Pascual 2012.
151 Subías Pascual 2020, 77.
152 One of which carries a date in the 14th indiction and the year 402 of the era of Diocletian
(i.e. 686 CE). For a first publication of this stone slab, see Piedrafita 2015. For a further discus-
sion on the inscribed materials from this archaeological context, see Subías Pascual 2020, 77–78.
153 On one of the Cufic texts discovered in the area of the chapel, see Subías Pascual 2020, 88.
154 Probably dated to the mid-fifth century CE; for a comprehensive edition of this corpus, see
Bosson 2015.
155 Recent surveys of the area, particularly the monastic cells, remark the presence of numer-
ous figural and textual graffiti, which still await publication.
68 | Leah Mascia

8 Conclusion
This paper aimed to provide a new and comprehensive perspective of the multi-
cultural and -lingual society of Oxyrhynchus between the Greco-Roman and the
late antique phase. The comparative examination of textual and archaeological
evidence allows us to integrate and widen the perspective offered so far by the
Greek papyrological documentation. The heterogeneous corpus of written tes-
timonies found or attributed to this city helps us to perceive how the interaction
between languages and scripts could take different forms. From no more than
their simple coexistence in the same physical space narrating the transfor-
mation of this city’s cultural landscape throughout the centuries to their inter-
twining in the production of specific written artefacts. This latter phenomenon
especially marks the existence of a considerable part of society at home in using
different languages, although how each individual could master them varied
greatly. The Egyptian language maintained a central importance in a social
panorama subject to continuous transformations, especially in the production
of ritual texts up until the early Roman period, latterly regaining its role in reli-
gious settings soon after the affirmation of the Christian religion. Through the
bilingual documentation combining Hebrew or Nabataean to the Greek lan-
guage, we gain a closer look at other communities inhabiting this city, which
have unfortunately left almost no traces in archaeological records. Observing
the compresence of different languages (Greek, Egyptian, Latin) in this city’s
funerary landscape, we can grasp a glimpse of how the coexistence and con-
stant interaction between the native and foreign inhabitants shaped the cultural
panorama of this settlement. Beyond doubt, in a city such as Oxyrhynchus,
standing out for the richness and variety of the surviving material evidence, it is
possible to trace the distinguishing features of the multilingual landscape of an
Egyptian town in the long transition to Christianity.

Acknowledgements

The research for this article, which is partly based on the results of my PhD
thesis entitled The Transition from Traditional Cults to the Affirmation of Chris-
tian Beliefs in the City of Oxyrhynchus, was funded by the Deutsche For-
schungsgemeinschaft (DFG, German Research Foundation) under Germany’s
Excellence Strategy – EXC 2176 ‘Understanding Written Artefacts: Material,
Interaction and Transmission in Manuscript Cultures’, project no. 390893796.
The research was conducted within the scope of the Centre for the Study of
Manuscript Cultures (CSMC) at Universität Hamburg.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 69

References
Bosson, Natalie (2015), ‘Inscriptions d’Oxyrhynque provenant du secteur 19’, in Anne
Boud’hors and Catherine Louid (eds), Études coptes XIII: Quinzième journée d’études (Ca-
hiers de la bibliothèque copte, 20), Paris: De Boccard, 69–89.
Boyaval, Вernard (1974), ‘Une malédiction pour viol de sépulture’, Zeitschrift für Papyrologie
und Epigraphik, 14: 71–73.
Bowman, Alan K., Revel A. Coles, Nikolaos Gonis, Dirk Obbink and Peter J. Parsons (eds)
(2007), Oxyrhynchus: A City and Its Texts (Graeco-Roman Memoirs, 93), London: Egypt Ex-
ploration Society.
Bresciani, Edda (1981), ‘Frammenti da un “prontuario legale” demotico da Tebtuni nell’Istituto
Papirologico G. Vitelli di Firenze’, Egitto e Vicino Oriente, 4: 201–215.
Campillo, Jordi and Concepció Piedrafita (2010), ‘L’estela del nen oxirrinquita Teó’, Nilus, 20:
18–22.
Castellano i Solé, Núria (2011–2012), ‘Un hypogée gréco-romain à Oxyrhynchos’, Journal of
Society for the Study of Egyptian Antiquities, 38: 35–53.
Castellano i Solé, Nuria, Maite Mascort, Concepció Piedrafita and Jaume Vivó (eds) (2015), Ex
Aegypto Lux et Sapientia: Homenatge al professor Josep Padró Parcerisa (Nova Studia Ae-
gyptiaca, 9), Barcelona: Publicacions i Edicions de la Universitat de Barcelona.
Chassinat, Émile (1966), Le Mystère d’Osiris au mois de Khoiak, vol. 1, Cairo: Imprimerie de
l’Institut français d’archéologie orientale.
Chassinat, Émile (1968), Le Mystère d’Osiris au mois de Khoiak, vol. 2, Cairo: Imprimerie de
l’Institut français d’archéologie orientale.
Cipriano, Giuseppina (2008), El-Bagawat: Un cimitero paleocristiano nell’alto Egitto, Todi: Tau
Editrice.
Clackson, Sarah (2007), ‘Coptic Oxyrhynchus’, in Bowman et al. 2007, 332–341.
Clarysse, Willy (2010), ‘Egyptian Temples and Priests: Graeco-Roman’, in Alan B. Lloyd (ed.), A
Companion to Ancient Egypt, Chichester: Wiley-Blackwell, 274–290.
Codina Reina, Dolors (2015), ‘La mort dels més joves en el món grecoromà: la representació de
l’infant a les esteles oxirrinquites’, in Castellano i Solé et al. 2015, 165–180.
Coulon, Laurent (2015), ‘Du périssable au cyclique: les effigies annuelles d’Osiris’, in Sylvia
Estienne, Valérie Huet, François Lissarrague, Francis Prost (eds), Figures de dieux: cons-
truire le divin en images, Rennes: Presses universitaires de Rennes, 295–318.
Coulon, Laurent (2018), ‘Les inscriptions associées aux niches de l’Osireion’, in Maite Mascort
(ed.), Oxyrhynchos IV: L’Osireion d’Oxirrinc, vol. 1 (Nova Studia Aegyptiaca, 10), 171–196.
Dekker, Renate (2013), ‘Recension of: Van Hinckley Sederholm, Papyrus British Museum 10808
and Its Cultural and Religious Setting (Probleme der Ägyptologie, 24), Leiden: Brill, 2006’,
Bibliotheca Orientalis, 70/1–2: 61–64.
Delattre, Alain, Jitse Dijkstra and Jacques van der Vliet (2020), ‘Christian Inscriptions from
Egypt and Nubia 7 (2019)’, Bulletin of the American Society of Papyrologists, 57: 397–420.
Dieleman, Jacco (2004), ‘Ein spätägyptisches magisches Handbuch: Eine neue PDM oder
PGM?’, in Friedhelm Hoffmann and Heinz-Josef Thissen (eds), Res severa verum gaudium:
Festschrift für Karl-Theodor Zauzich zum 65. Geburtstag am 8. Juni 2004 (Studia Demotica,
6), Leuven: Peeters, 121–128.
70 | Leah Mascia

Eaton, Katherine J. (2006), ‘The Festivals of Osiris and Sokar in the Month of Khoiak: The Evi-
dence from Nineteenth Dynasty Royal Monuments at Abydos’, Studien zur Altägyptischen
Kultur, 35: 75–101.
Goyon, Jean-Claude (2018), ‘Les inscriptions du petit mobilier funéraire’, in Mascort Roca 2018,
vol. 1, 197–210.
Grenfell, Bernard (1897), ‘Oxyrhynchus and Its Papyri’, in Francis Llewellyn Griffith (ed.), Egypt
Exploration Fund: Archaeological Report 1896–1897, London: Egypt Exploration Fund, 1–12.
Grossmann, Peter (1982), Mittelalterliche Langhauskuppelkirchen und verwandte Typen in
Oberägypten: eine Studie zum mittelalterlichen Kirchenbau in Ägypten, Glückstadt: Verlag
J. J. Augustin GMBH.
Grossmann, Peter (2008), ‘Antinoopolis: der Komplex des Peristylbaus’, in Rosario Pintaudi
(ed.), Antinoupolis I: Scavi e Materiali, Firenze: Firenze University Press, 41–46.
Healey, John (2004), ‘A Nabataean Papyrus Fragment (Bodleian MS Heb. d. 89)’, Zeitschrift für
Papyrologie und Epigraphik, 146: 183–188.
Jastrzębowska, Elżbieta (2019), ‘La comunità a tavola in memoria dei morti e martiri nel IV
sec.’, in Jordi López Vilar (ed.), Actes 4t Congrés Internacional d’Arqueologia i Món Antic:
VII Reunió d’Arqueologia Cristiana Hispànica, El cristianisme en l’Antiguitat Tardana, No-
ves Perspectives, Tarragona: Universitat Rovira i Virgili, 83–89.
Kocar, Alexander (2019), ‘Oxyrhynchus and Oracles in Late Antiquity’, in AnneMarie Luijendijk and
William E. Klingshirn (eds), My Lots Are in Thy Hands: Sortilege and Its Practitioners in Late
Antiquity (Religions in the Graeco-Roman World, 188), Leiden: Brill, 196–210.
Krautheimer, Richard (1965), Early Christian and Byzantine Architecture, Harmondsworth:
Penguin Books.
Love, Edward O. D. (2021), Script Switching in Roman Egypt: Case Studies in Script Conven-
tions, Domains, Shift, and Obsolescence from Hieroglyphic, Hieratic, Demotic, and Old
Coptic Manuscripts (Archiv für Papyrusforschung und verwandte Gebiete, 46), Berlin: De
Gruyter.
Martín Hernández, Raquel (2011), ‘El transporte de momias a través del Nilo: los textos de las
etiquetas de momia’, in Juan Antonio Belmonte Marín and Juan Carlos Oliva Mompeán
(eds), Esta Toledo, aquella Babilonia: convivencia e interacción en las sociedades del
Oriente y del Mediterráneo Antiguo (Colección Estudios, 131), Cuenca: Universidad de
Castilla–La Mancha, 543–554.
Martínez García, José Javier and Leah Mascia (2023), ‘Figural Graffiti from the Basilica of St.
Philoxenos at Oxyrhynchus (el-Bahnasa, Egypt)’, Imafronte, 30: 34–47.
Mascia, Leah (2023a), ‘The Egyptian Temples of Oxyrhynchus and Their Personnel Under the
Roman Empire’, in Alexandre Loktionov (ed.), Compulsion and Control in Ancient Egypt:
Proceedings of the 3rd Lady Wallis Budge Symposium, Oxford: Archaeopress Publishers,
27–41.
Mascia, Leah (2023b), ‘Preliminary Report on the Examination of the Papyrological and Epi-
graphic Sources Found at the Site of Oxyrhynchus’, Nilus, 31: 19–21.
Mascia, Leah (2023c), ‘Practices of Ritual Power in the Necropolis of Oxyrhynchus: A New
Assessment’, in Ola el-Aguizy and Burt Kasparian (eds), ICE XII: Proceedings of the Twelfth
International Congress of Egyptologists, vol. 2, Cairo: Institut français d’archéologie ori-
entale, 1137–1143.
Mascia, Leah and José Javier Martínez García (2021), ‘A New Christian Epitaph from Oxyrhyn-
chus’, Zeitschrift für Papyrologie und Epigraphik, 218: 139–141.
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 71

Mascort Roca, Maite (2018) (ed.), L’Osireion d’Oxirrinc, vol. 1–2 (Nova Studia Aegyptiaca, 10),
Barcelona: Publicacions i Edicions de la Universitat de Barcelona.
Mascort Roca, Maite and Esther Pons Mellado (2019), ‘Tueris–Oxirrinco: la diosa protectora de
Per-Medyed’, Trabajos de Egiptología, 10: 241–256.
Mascort Roca, Maite, Esther Pons Mellado, Bibiana Augustí Farjas, Eloy Algorri García, Hassan
Ibrahim Amer, Ibrahim Hassan Amer, Núria Castellano Solé, Marguerite Erroux-Morfin, An-
tonio López Cano, José Javier Martínez García, Leah Mascia, Margalida Munar Grimalt and
Adriana Recasens Escardó (2020), ‘Memòria dels treballs desenvolupats per la Missió
Arqueològica de la Universitat de Barcelona-IPOA, en el jaciment d’Oxirrinc (Mínia),
Egipte. Campanya de 2020’, Nilus, 29: 3–35.
Mascort Roca, Maite, Esther Pons Mellado, Bibiana Augustí Farjas, Eloy Algorri García, Hassan
Ibrahim Amer, Ibrahim Hassan Amer, Bernat Burgaya Martínez, Núria Castellano Solé, De-
lia Eguiluz Maestro, Marguerite Erroux-Morfin, Antonio López Cano, José Javier Martínez
García, Leah Mascia, Margalida Munar Grimalt, M. Francesca Pullia, Adriana Recasens Es-
cardó, Irene Riudavets González and Susana Soler Polo (2021), ‘Memòria dels treballs
desenvolupats per la Missió Arqueològica de la Universitat de Barcelona–IPOA, en el
jaciment d’Oxirrinc (Mínia), Egipte. Campanya 2021’, Nilus, 30: 3–36.
Mattha, Girgis (ed.) (1975), The Demotic Legal Code of Hermopolis West, vol. 1 (Bibliothèque
d’étude, 45), Cairo: Publications de l’Institut français d’archéologie orientale du Caire.
Mayer-Olivé, Marc (2015), ‘Una inscripció romana trobada a Oxirrinc (Oxyrhynchus)’, in Castel-
lano i Solé et al. 2015, 393–398.
Menci, Giovanna (2007), ‘Un amuleto “Bous” da Antinoe’, Zeitschrift für Papyrologie und
Epigraphik, 159: 249–252.
Padró, Josep, Eloy Algorri, Hassan Ibrahim Amer, Jordi Campillo, Núria Castellano, Dolors
Codina, Marguerite Erroux-Morfin, María Luz Mangado, José Javier Martínez, Maite Mas-
cort, Esther Pons and Eva Subías (2009), ‘Memòria provisional dels treballs realitzats en
el jaciment d’Oxirrinc (El-Bahnasa, provincia de Minia, Egipte) durant la campanya de
2009’, Nilus, 18: 3–20.
Padró, Josep, Hassan Amer, Jordi Campillo, Núria Castellano, Marguerite Erroux-Morfin, María
Luz Mangado, José Javier Martínez, Maite Mascort, Esther Pons, Irene Riudavets, Eva
Subías and Anaïs Tillier (2010), ‘Informe preliminar dels treballs d’excavació i restauració
realitzats al jaciment d’Oxirrinc (El-Bahnasa, Mínia) durant la campanya de 2010’, Nilus,
19: 3–12.
Padró, Josep, Hassan Amer, Jordi Campillo, Dolors Codina, Marguerite Erroux-Morfin, José
Javier Martínez, Maite Mascort, Esther Pons, Irene Riudavets (2012), ‘Informe preliminar
dels treballs arqueològics realitzats a Oxirrinc (El-Bahnasa, Mínia, Egipte), durant la cam-
panya de 2011–2012’, Nilus, 21: 3–29.
Padró, Josep, Bibiana Agustí, Hassan Amer, Dolors Codina, Marguerite Erroux-Morfin, Jérôme
Gonzalez, José Javier Martínez, Maite Mascort, Esther Pons, Irene Riudavets, Wim van
Neer (2014), ‘Memòria preliminar de les excavacions de la campanya de 2014 a El-
Bahnasa, Oxirrinc (Mínia, Egipte)’, Nilus, 23: 3–18.
Padró, Josep, José Martínez, Concepció Piedrafita (2018), ‘Historia de un edificio religioso en
Oxirrinco, desde el s. IV a.C. hasta el s. VII d.C.’, in Alejandra Guzmán Almagro and Javier
Velaza (eds.), Miscellanea philologica et epigraphica Marco Mayer oblata (Anuari de
Filologia: Antiqua et Mediaevalia, 8), Barcelona: Universitat de Barcelona, 702–718.
72 | Leah Mascia

Papaconstantinou, Arietta (2001), Le culte des saints en Égypte des Byzantins aux Abbassides:
L’apport des inscriptions et des papyrus grecs et coptes, Paris: Centre national de la re-
cherche scientifique.
Parlasca, Klaus (2007), ‘Grave-Reliefs and Architectural Sculpture’, in Bowman et al. 2007, 91–103.
Pestman, Peter W. (1983), ‘L’origine et l’extension d’un manuel de droit égyptien: quelques
réflexions à propos du soi-disant Code de Hermoupolis’, Journal of the Economic and So-
cial History of the Orient, 26/1: 14–21.
Piedrafita, Concepció (2003), ‘Lectura i traducció dels textos murals grecs’, in Eva Subías Pascual
(ed.), La corona immarcescible: Pintures de l’antiguitat tardana de la necròpolis alta d’Oxirinc
(Mínia, Egipte) (Sèrie Documenta, 1), Tarragona: Institut Català d’Arqueologia Clàssica,
37–42.
Piedrafita, Concepció (2008), ‘L’épigraphie grecque d’Oxyrhynchos’, in Marguerite Erroux-
Morfin and Josep Padró Parcerisa (eds), Oxyrhynchos: Un site de fouilles en devenir, Col-
loque de Cabestany Avril 2007 (Nova Studia Aegyptiaca, 6), Barcelona: Publicacions i Edi-
cions de la Universitat de Barcelona, 135–145.
Piedrafita, Concepció (2010), ‘Una tauleta màgica del déu Set procedent d’Oxirrinc (Proposta
d’identificació i lectura)’, Nilus, 19: 29–31.
Piedrafita, Concepció (2011), ‘Quatre breus inscripcions onomàstiques gregues d’Oxirrinc
(campanyes 2007–2010)’, Sylloge epigraphica Barcinonensis, 9: 63–73.
Piedrafita, Concepción (2015), ‘El prior Menas i una inscripció amb l’era de Dioclecià a Oxirrinc’, in
Castellano i Solé et al. 2015, 455–466.
Pintaudi, Rosario (2011–2012 [2013]), ‘Materiali per una riflessione su indirizzi, prescritti e
protocolli: Note di lettura e nuove edizioni’, Analecta Papyrologica, 23/24: 143–170.
Piotrkowski, Meron M. (2018), ‘“Literary Jews:” The Jewish Community of Oxyrhynchus in Light
of Non-documentary Texts on Papyrus’, in Meron M. Piotrkowski, Geoffrey Herman and
Saskia Dönitz (eds), Sources and Interpretation in Ancient Judaism: Studies for Tal Ilan at
Sixty (Ancient Judaism and Early Christianity, 104), Leiden: Brill, 143–173.
Pons Mellado, Esther (2016), ‘Tombs of the Roman Period in Sector 26 of the High Necropolis:
Archaeological Site of Oxyrhynchus, El-Bahnasa’, Trabajos de Egiptología, 7: 161–178.
Prada, Luigi (2016), ‘P.Oxy. XXXI 2607 Re-Edited: A Greek Oneirocriticon from Roman Egypt’, in
Tomasz Derda, Adam Łajtar and Jakub Urbanik (eds), Proceedings of the 27th International
Congress of Papyrology: Warsaw 29 July–3 August 2013 (Journal of Juristic Papyrology:
Supplement, 28), Warsaw: University of Warsaw, 623–646.
Quack, Joachim Friedrich (2000), ‘Das Buch vom Tempel und verwandte Texte: Ein Vorbericht’,
Archiv für Religionsgeschichte, 2: 1–20.
Quack, Joachim Friedrich (2003), ‘Le manuel du temple: une nouvelle source sur la vie des
prêtres égyptiens’, Égypte, Afrique & Orient, 29: 11–18.
Quack, Joachim Friedrich (2010), ‘Was ist das “Ptolemäische”?’, Die Welt des Orients, 40/1:
70–92.
Quack, Joachim (2012), ‘Old Wine in New Wineskins? How to Write Classical Egyptian Rituals in
More Modern Writing Systems’, in Alex de Voogt and Joachim Friedrich Quack (eds), The
Idea of Writing: Writing Across Borders, Leiden: Brill, 219–243.
Quack, Joachim (2016a), ‘Translating the Realities of Cult: The Case of the Book of the Temple’,
in Ian Rutherford (ed.), Greco-Egyptian Interactions: Literature, Translation, and Culture,
500 BCE–300 CE, Oxford: Oxford University Press, 267–286.
Quack, Joachim (2016b), ‘The Last Stand? What Remains Egyptian in Oxyrhynchus’, in Kim
Ryholt and Gojko Barjamovic (eds), Problems of Canonicity and Identity Formation in An-
 A Journey Through the Multilingual Landscape of Oxyrhynchus | 73

cient Egypt and Mesopotamia (Carsten Niebuhr Institute Publications, 43), Copenhagen:
Museum Tusculanum Press, 105–126.
Quack, Joachim (2017), ‘How the Coptic Script Came About’, in Eitan Grossmann, Peter Dils,
Tonio Sebastian Richter, Wolfgang Schenkel (eds), Greek Influence on Egyptian-Coptic:
Contact-Induced Change in an Ancient African Language (Lingua Aegyptia Studia Mono-
graphica, 17), Hamburg: Widmaier Verlag, 27–96.
Rea, John (ed.) (1978), The Oxyrhynchus Papyri, vol. 46 (Graeco-Roman Memoirs, 65), London:
Egypt Exploration Society.
Sijpesteijn, Pieter Johannes (1981), ‘Objects with Script in the Collection Moen’, Zeitschrift für
Papyrologie und Epigraphik, 42: 111–112.
Speidel, Michael Alexander (2013), ‘Gnomon of the idios logos’, in Roger S. Bagnall, Kai
Brodersen, Craige B. Champion, Andrew Erskine and Sabine R. Huebner (eds), The Ency-
clopedia of Ancient History, Hoboken: Wiley–Blackwell Publishing, 3390–3391.
Stadler, Martin Andreas (2015), ‘Archaeology of Discourse: The Scribal Tradition in the Roman
Fayyum and the House of Life at Dime’, in Mario Capasso and Paola Davoli (eds), Soknopaios,
the Temple and Worship: Proceedings of the First Round Table of the Centro di Studi Papiro-
logici of Università del Salento (Edaphos, 1), Lecce: Pensa Multimedia, 187–232.
Subías Pascual, Eva (2008), La maison funéraire de la nécropole haute à Oxyrhynchos (el
Minyâ, Égypte): du tombeau à la diaconie (Nova Studia Aegyptiaca, 5), Barcelona: Univer-
sitat de Barcelona.
Subías Pascual, Eva (2012), ‘La fortaleza bizantina del suburbio noroccidental de Oxirrinco (El-
Minia, Egipto)’, in Luís Manuel Araújo and José das Candeias Sales (eds), Novos trabalhos
de Egiptología ibérica, vol. 2, Lisboa: Instituto Oriental e Centro de História da Faculdade
de Letras da Universidade de Lisboa, 1163–1178.
Subías Pascual, Eva (2020), ‘Further Reflections on the Byzantine Fortress at Oxyrhynchos:
Martyrial and Funerary Church, Monastery and Arab Fort’, in Paola Buzi (ed.), Coptic
Literature in Context (4th–13th cent.): Cultural Landscape, Literary Production, and
Manuscript Archaeology (Percorsi, Strumenti e Temi di Archeologia, 5), Rome: Edizioni
Quasar, 77–92.
Thissen, Heinz-Josef (2013), ‘Donkeys and Water: Demotic Ostraca in Cologne as Evidence for
Desert Travel Between Oxyrhynchos and the Bahariya Oasis in the 2nd Century BC’, in
Frank Förster and Heiko Riemer (eds), Desert Road Archaeology in Ancient Egypt and
Beyond (Africa Praehistorica, 27), Cologne: Heinrich Barth Institut, 391–398.
Wilburn, Andrew T. (2012), Materia Magica: The Archaeology of Magic in Roman Egypt, Cyprus,
and Spain, Ann Arbor: The University of Michigan Press.
Jochen Hermann Vennebusch
Commemoration, Explication, and
Obligation: The Baptismal Font of St Reinoldi
in Dortmund
Abstract: The bronze baptismal font cast by Johann Winnenbrock in 1469 in the
church of St Reinoldi in Dortmund is an almost unique testimony to multilin-
gualism in late medieval churches. Instead of extensive figural decoration, there
is a Low German inscription around the foot and a Latin inscription around the
cuppa. These texts bear witness to the origin of the artefact and interpret the
sacrament of baptism theologically. The third inscription on the upper rim of
the cuppa is almost singular. The godparents are addressed and reminded of
their duties towards the baptismal candidate in vernacular language. These
texts are taken or derived from various sources and imply specific addressees or
intentions through the choice of language and their placement.

1 Introduction
Three inscriptions in relief are particularly decorative. These are written in minuscule script
and are not difficult to decipher despite the abbreviations. The minuscule script (monk’s
script), which consists of angular small letters (Fraktur script), is the prevailing artist’s script
from the second half of the fourteenth century until into the sixteenth century.1

With these words, Otto Stein described the inscriptions on the bronze baptismal
font in the church of St Reinoldi in Dortmund (Fig. 1). Although he noted in the
further course of his remarks that the inscriptions are written in different lan-
guages and transcribed the texts, he did not consider why there could be one
Latin and two Low German inscriptions on the sacred vessel, which turn the
baptismal font into a multilingual written artefact. This phenomenon, which
has only been marginally discussed in further research on the bronze baptismal

||
1 Stein 1906, 142: ‘Zu ganz besonderem Schmucke gereichen ihm drei Reliefinschriften. Diese
sind in Minuskelschrift geformt und trotz der Abkürzungen nicht schwer zu entziffern. Die
Minuskelschrift (Mönchsschrift), welche aus eckigen kleinen Buchstaben (Fraktur) besteht, ist
die seit der zweiten Hälfte des 14. Jahrhunderts bis ins 16. Jahrhundert hinein herrschende
Künstlerschrift’.

Open Access. © 2024 the author, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-004
76 | Jochen Hermann Vennebusch

font, is the central topic of this article. Firstly, a brief introduction to the inscrip-
tions on medieval bronze baptismal fonts is given, before the object preserved
in Dortmund is described in detail. Finally, the inscriptions are analysed from a
formal point of view, before the origin of the texts that can be read on the bap-
tismal font is examined and the corresponding use of the respective language in
connection with the particular place of attachment is discussed.

Fig. 1: General view of the baptismal font, Johann Winnenbrock, 1469, bronze, St Reinoldi,
Dortmund; © Andreas Lechtape.
 Commemoration, Explication, and Obligation | 77

2 The phenomenon of multilingual inscriptions

on late medieval bronze baptismal fonts
Numerous, but not all, medieval baptismal fonts made of bronze in northern
Germany bear inscriptions, which in most cases were applied before casting.
These inscriptions sometimes interpret the sacrament of baptism and its effects
on the baptised person theologically, and sometimes they also provide infor-
mation about the casting of the baptismal font or its caster or donor.

Fig. 2: Detail of the inscription, baptismal font, unknown caster, first half of the fourteenth
century, bronze, St Mary’s Church, Rendsburg; © Jochen Hermann Vennebusch.
78 | Jochen Hermann Vennebusch

A quite early example of an inscription distinctly related to the sacrament of bap-

tism can be found on the bronze baptismal font of St Mary’s Church in Rendsburg
in Schleswig-Holstein (Fig. 2). An inscription band runs below the upper rim
around the cuppa of the baptismal font, which was created at the beginning of the
fourteenth century. The beginning of the Apostles’ Creed is rendered in Latin here,
however, due to a production error, upside down and in a left-hand reading direc-
tion: ‘+ CREDO + IN + DEVM + PATREM + OMNIPOTENTEM + CREATOREM +
CELI’.2 The reference to the sacrament of baptism is obvious because of the profes-
sion of faith expressed in the text. During the baptismal liturgy in early Christiani-
ty, the originally adult candidates for baptism were asked by the baptismal minis-
ter about their faith, which was immediately followed by baptism.3 Adult baptism
receded into the background in the Middle Ages and infant baptism became the
rule. Nevertheless, even in the medieval baptismal liturgy, questions were asked
about the faith of the person to be baptised, although it was not the infant who
answered these questions but the godparents who took on this obligation on behalf
of the child.4 Thus, the affixing of the beginning of the Latin Credo as an inscription
on the baptismal font is extremely meaningful. By contrast, the baptismal font cast
by Hans Apengeter in 1337 in St Mary’s Church in Lübeck gives background infor-
mation on its manufacture in Low German (Fig. 3):

ANNO ⋅ D(OMI)NI ⋅ Mo ⋅ CCC ⋅ XXXo VIIo ⋅ JN ⋅ UI//GILIA ⋅ PE(N)THECOSTES ⋅

PERFECTVM ⋅ EST ⋅ PRESENS ⋅ OPVS ⋅
MARIA ⋅ WES ⋅ T//O ALLEN ⋅ GMALEN ⋅
GNEDICH HERN ⋅ EVERDE UAN ⋅ ALEN ⋅
CRIST(US) ⋅ DI DI ⋅ MART(ER) ⋅ HEFT ⋅ GELEDEN ⋅
GNADE ⋅ HERN ⋅ IOH(ANN)E ⋅ UAN · SCHEPENSTEDEN ⋅ /
UNDE ⋅ UERSEGTEG ⋅ NICHT ⋅ HEMELRIKE ⋅
IWME ⋅ TRWEN ⋅ DIENER DARTWIKE ⋅
(CHRIST)E ⋅ UERGIF ⋅ ALLE ⋅ MISSEDAT ⋅
DEME ⋅ DI ⋅ DIT ⋅ VAT ⋅ GEMAKET ⋅ HAT ⋅
HANS ⋅ APENGITER ⋅ WAS ⋅ HE ⋅ GENANT ⋅
VND ⋅ WAS ⋅ GEBORN ⋅ UAN ⋅ SASSENLANT ⋅5

||
2 ‘I believe in God the Father, the Almighty, the Creator of Heaven’. The translations of the
inscriptions and texts have been provided by the author, unless otherwise stated. See on this
baptismal font, Haupt 1888, 207; Rauterberg 2006, 24; see on the production errors of the
inscription on this baptismal font, Vennebusch 2022a, 156.
3 Fürst 2008, 132.
4 Angenendt 1987, 317–318; Müller 2012, 105.
5 ‘In the year of Our Lord 1337 on the evening before Pentecost [7 June] this work was complet-
ed. May Mary be gracious to Mister Everd van Alen in all cases. Christ, who suffered the torture,
 Commemoration, Explication, and Obligation | 79

Fig. 3: General view of the baptismal font, Hans Apengeter, 1337, bronze, St Mary’s Church,
Lübeck; © Kathrin Ulrich.

||
be merciful to Mister John of Schepenstede and do not deny the Kingdom of Heaven to his
faithful servant Hartwich. Christ, forgive all evil deeds to him who made this baptism: Hans
Apengeter was his name and he came from Saxony’. See on this baptismal font, especially
Lampe 2022, 344–346; Vennebusch 2022a, 157–158; Vennebusch 2022b.
80 | Jochen Hermann Vennebusch

We can already observe here that there are some Latin set pieces in the Low
German text that refer to the date of the casting and seem like a solemn dec-
laration of the completion of the baptismal font. This observation is hardly
surprising since this date is based on the liturgical calendar generally used
and is an object in a sacred context. Similar phenomena of multilingual texts
can be found on some medieval bronze baptismal fonts, on which either Lat-
in dates appear in Low German vernacular inscriptions in the context of
longer donor or caster’s inscriptions, or Latin phrases commonly used appear
as short passages. An example is the bronze baptismal font cast at the end of
the fourteenth century in St Bartholomew’s church in Neuenkirchen near
Soltau, whose inscription calls for prayers for the salvation of the presumed
donors (Fig. 4): ‘orate ♦ p(ro) ♦ Ludeke ♦ mils ♦ vnde ♦ sin ♦ vrowe ♦’.6 We
can assume here that the Latin passage ‘orate pro’ was anchored in the gen-
eral cultural consciousness, so that it was understood as a liturgical phrase,
although the general legibility of the inscription is made more difficult by the
left-hand reading direction – also due to a production error.7 We find another
case on the baptismal font cast by Heinrich Kock in 1505 in St Andrew’s
church in Geversdorf (Fig. 5): ‘fons kristi bin ik ghe nant hinrik [foundry
mark] kock mi ghe ghoten hat got gheve deir sele rat / in dem iar unses [lily]
heren do me schref [lily] omo ccccc vnde v’.8 In this Low German inscription,
in addition to the Latin date, the expression ‘fons kristi’, ‘baptismal fountain
of Christ’, is used as a self-designation of the baptismal font that speaks for
itself, with the name of Christ rendered as a phonetic spelling of the Latin
phrase.

||
6 ‘Pray for Ludeke Miles and his wife’. See on this baptismal font, Mithoff 1878, 70; Deckert et
al. 1939, 39; Kämmer 1980, 48; von Poser und Groß-Naedlitz 1983, 51–55; Pantel 2001, 241.
7 See on the method of production of the inscription on this baptismal font and the errors that
occurred in the process, Vennebusch 2022a, 160–161.
8 ‘I am called Fons Christi. Heinrich Kock cast me. God grant counsel to the soul. / In the year
of Our Lord, when it was written 1505’. See on this baptismal font, Mithoff 1878, 42; Kiecker,
Lenz and Rüther 1956, 132, pl. 115; Böker 1997, 194.
 Commemoration, Explication, and Obligation | 81

Fig. 4: General view of the baptismal font, unknown caster, before 1400, bronze, St Bartholo-
mew’s church, Neuenkirchen; © Jochen Hermann Vennebusch.
82 | Jochen Hermann Vennebusch

Fig. 5: General view of the baptismal font, Heinrich Kock, 1505, bronze, St Andrew’s church,
Geversdorf; © Jochen Hermann Vennebusch.
 Commemoration, Explication, and Obligation | 83

3 The baptismal font of the church of St Reinoldi

in Dortmund
After these introductory remarks, we come to a bronze baptismal font that has a
remarkable solution in the context of multilingualism on this type of written
artefacts: unfortunately, the baptismal font of the St Reinoldi church in Dort-
mund is only a fragment now, as it was severely damaged during the air raids
during the Second World War.9 The baptismal font,10 reminiscent of a chalice,
stands on a circular base tapering towards the top, from the centre of which
emerges a similarly round and tapering shaft divided by whorls (Fig. 6). On top
of this rests the broadly sprawling cuppa, which is divided into various bands
by rings running around it (Fig. 7). The broad, widely protruding ring, which
forms the upper end of the font, rests on lions’ heads,11 whose throats held six
separately cast pinnacles surrounding the baptistery and resting on small
crouching lion figures before the damage during the Second World War
(Fig. 8).12 Two busts of youthful-looking figures sprout from the ring, which
formerly served to fasten a font cover. Shallow reliefs of eagles and griffins –
according to Horst Appuhn ‘nach der Art einer Glocke’,13 in the manner of a bell,
are recognisable on the cuppa, which go back to wax appliqués that were re-
produced with the help of models and applied to the ‘clay shirt’ and melted out
before casting. In addition, a repeat of round arches from which drooping leaves
sprout runs below the rim. In contrast to the cuppa, the ornamentation of the
foot is composed of floral motifs, with the exception of a small dragon and a
bird in a tendril of leaves and flowers, and is mainly limited to signs separating
words. Beyond that, the shaft, apart from the whorls, remained free of any deco-
ration.

||
9 See on this baptismal font, Fritz 1933, 48–51; Fritz 1956, 88–89; Lindemann s.a., 16; Lübke
1853, 418–419; Ludorff 1894, 30; Stein 1906, 140–143; Rinke 1985, 26–27; Rinke 1987, 52; Ohm,
Schilp and Welzel 2006, 232–233, no. 145 (Judith Zepp).
10 Measures: h: 113 cm / Ø: 105.5 cm. See on the measurements, Rinke 1985, 26.
11 Wolfgang Rinke wrongly interprets these heads as the heads of dragons or demons. See
Rinke 1985, 27; Rinke 1987, 52.
12 Lübke 1853, 418; Appuhn 1970, 33.
13 Appuhn 1970, 33.
84 | Jochen Hermann Vennebusch

Fig. 6: Foot of the Dortmund baptismal font; © Andreas Lechtape.

Fig. 7: Cuppa of the Dortmund baptismal font; © Andreas Lechtape.

 Commemoration, Explication, and Obligation | 85

Fig. 8: Pre-war photograph of the Dortmund baptismal font; © Bildarchiv Foto Marburg.

Let us now turn to the inscriptions on the baptismal font, which are of particular
interest for this volume, and begin with the inscription on the foot. Here, the
founder and the year of casting are mentioned in Low German (Figs 9a–c):

♦ Jn : dem : iar : vns ⋅ herr(n) ♦ [pinnacle] ♦ Mo CCCCo ⋅ LXIXo ⋅ doe ♦ goet : [pinnacle] ♦
iohan ♦ winnenbrock : [pinnacle] : klockengeiter ♦ borger ♦ [pinnacle] : tho ♦ dorpmun-
de : ♦ ♦ [pinnacle] ♦ disse : dope ♦ [pinnacle]14

||
14 ‘In the year of Our Lord 1469, Johann Winnenbrock, bell founder and citizen of Dortmund,
cast this baptismal font’. See on the founder, Mithoff 1885, 344.
86 | Jochen Hermann Vennebusch

Figs 9a–c: Inscription on the foot of the Dortmund baptismal font; © Andreas Lechtape.
 Commemoration, Explication, and Obligation | 87

This inscription, enclosed by two bars and written in Gothic minuscule, was
placed on the foot that tapers upwards, thus, it is turned towards the viewer and
can be read from above. The individual words are separated from each other by
colons, elongated blossoms or smaller blossoms partly arranged one above the
other without any recognisable system. The beginning of the inscription is
marked by a slightly wider strip occupied by a small dragon and a bird in dense
tendrils turning to the right, thus, implying a reading direction. It can be as-
sumed that the leaf and flower tendril after the section ‘: tho ♦ dorpmunde :’ is
to be understood as a section filler, because the following word ‘♦ disse’ could
not have been inscribed there without disturbing the symmetrical arrangement
of the pinnacles. The wider and now empty spaces between some sections of the
inscription can also be traced back to the pinnacles destroyed during the Sec-
ond World War, once being placed on small lions bearing the whole baptismal
font. Thus, the inscription band on the base was originally interrupted by these
architectures, but here too without any recognisable system or division of the
text into sections of meaning.
The second inscription, also in Gothic minuscule, but now in Latin, is
placed around the cuppa of the baptismal font (Figs 10a–f). A theological inter-
pretation of the sacrament of baptism is given here:

Baptismi ♦ fructus ♦ est [dragon and bird in tendrils] plena ♦ remissio : culpe [dragon
and bird in tendrils] gracia : confertur ♦ et [tendrils] pena ♦ remictitur ♦ omnis [dragon in
tendrils] qui ♦ crediderit ♦ et ♦ baptisat(us) ♦ fuerit ♦ salvus ♦ erit ⋅ i(n) ⋅ et(er)nu(m) ⋅
[tendrils]15

Again, the individual words are separated from each other without any recog-
nisable system either by elongated flowers, smaller flowers arranged one above
the other, by the figuration already known from the inscription on the foot,
consisting of a dragon in a network of tendrils either with or without a bird, or
leaf and flower tendrils. The beginning of the inscription is particularly distin-
guished by the leafy tendrils that merge into the initial letter. The limited space
available was almost completely used, and numerous abbreviations were made
especially at the end of the text. Additionally, no consideration was given here
to the pinnacles formerly surrounding the baptismal font; the inscription con-
tinued behind them without the areas occupied by the tendril work or the word
separators corresponding to the rod-like architectures.

||
15 ‘The fruit of baptism is the complete forgiveness of guilt. Grace is granted and sin is remit-
ted. Whoever believes and is baptised will be saved for eternity.’
88 | Jochen Hermann Vennebusch

a b

c d

e f
Figs 10a–f: Inscription on the cuppa of the Dortmund baptismal font; © Andreas Lechtape.

Furthermore, we can observe an almost singular location of an inscription for

medieval baptismal fonts on the Dortmund object. We can read a surrounding
inscription in Gothic minuscule on the upper side of the flat rim, which is again
formulated in Low German (Fig. 11):

⋅ O ⋅ geestilike ⋅ vader ⋅ vnde ⋅ mod(er) ⋅ nempt ⋅ dat ⋅ to sin(n)e ⋅ juwe(n) ⋅ pade(n) ⋅ to ⋅

lere(n) ⋅ de(n) ⋅ gelow(en) ⋅ i(n) ⋅ r(e)cht(er) ⋅ min(n)e ⋅ [bust] ⋅ Unde ⋅ dat ⋅ pat(er) ⋅ nost(er) ⋅
dei ⋅ x ⋅ gebode ⋅ gods ⋅ to ⋅ gader ⋅ Jh(e)s(us) ⋅ sal ⋅ sin ⋅ iu ⋅ loen ⋅ vnd ⋅ sin ⋅ hemelsche ⋅
vader ⋅ ⋅ [bust]16

||
16 ‘O spiritual father and mother, take this to mean to teach your godchildren the faith in right
love and the Lord’s Prayer, the Ten Commandments of God together. Jesus shall be your reward
 Commemoration, Explication, and Obligation | 89

Fig. 11: Inscription on the upper rim of the Dortmund baptismal font; © Andreas Lechtape.

In contrast to the inscriptions on the foot and the cuppa, the word separators
are kept very uniform and consist of six-petalled blossoms, which, however,
show a different casting quality. The inscription itself begins after one of the
two busts and is highlighted by a majuscule O; the end is marked by two suc-
cessive blossoms which are not found elsewhere. The interruption of the text by
the second bust occurs between the words ‘min(n)e’ and ‘Unde’, without a cal-
culated co-ordination of the layout and the text becoming clear. The bronze
baptismal font of St Lamberti church in Coesfeld also has very similar inscrip-
tions and in the same places (Fig. 12).17 According to the inscription on the foot,
it was cast in 1504 by the Dortmund citizens Reinhold Wiedenbrock and Klaus

||
and his heavenly Father’. The further shaft in the word juwe(n) after the w is probably due to an
error.
17 Lübke 1853, 419; Wenning 1996, 13; see on the founder, Mithoff 1885, 340–341.
90 | Jochen Hermann Vennebusch

Potgeiter, from which it can be concluded that Reinhold Wiedenbrock is proba-

bly a descendant, possibly the son, of Johann Winnenbrock, who took over his
workshop and also the casting models, although the ornamentation on the
Coesfeld baptismal font is more restrained.18

Fig. 12: General view of the baptismal font, Reinhold Wiedenbrock and Klaus Potgeiter, 1504,
bronze, St Lamberti church, Coesfeld; © Andreas Lechtape.

||
18 Lübke 1853, 419.
 Commemoration, Explication, and Obligation | 91

Let us briefly summarise the formal observations on the inscriptions on this

baptismal font: the beginning of the text is particularly emphasised in each
inscription, on the cuppa, by a stripe decorated with vegetal ornaments, and on
the foot, by a dragon and a bird in tendrils, whereas on the upper rim the end is
marked by two flowers. In all three cases, majuscules introduce the inscriptions
and, thus, stand out from the minuscules of the texts. The word separators on
the foot and on the cuppa – in comparison to the inscription on the rim – do not
follow any systematic. It is noteworthy that the caster’s inscription on the foot
and the inscription addressed to the godparents on the rim are placed in such a
way that they can be read without any problems, whereas the sacramental-
theological text on the cuppa has been partially covered by the pinnacles for-
merly placed in front of it. Since this inscription is directed slightly downwards
it was quite difficult to read as well.

4 Praising the work: The vernacular caster’s

inscription on the base
After this detailed description, the following considerations will now focus on
the question of the extent to which the place of attachment, the content and the
language of the respective inscriptions are co-ordinated and what implications
are connected with this. As has already been mentioned at the beginning, the
inscription on the upwardly tapering foot is slightly inclined so that it can be
seen and read by the viewer of the baptismal font from a certain distance in top
view. Even if it cannot be supposed that all viewers were able to read, it can be
assumed based on the vernacular text that Johann Winnenbrock intended the
lowest possible hurdles to understanding the inscription.19 The fact that the
inscription took into account the former pinnacles also contributes to this, so
that no part of the inscription was partially obscured by the architecture in front
of it. It is also likely that Winnenbrock, as the caster, wrote the inscription in a
language familiar to him, possibly even appropriate to or befitting his status,
and affixed it to the baptismal font. In comparison with other medieval bronze
baptismal fonts, it is conspicuous that there is no call for intercession for the
caster. As we have already seen, these texts, which serve the memoria of the
caster, sometimes also of the donor, are to be found quite frequently. In addi-
tion to the baptismal font by Hans Apengeter in Lübeck, this occurs, for exam-

||
19 See on vernacular foundry inscriptions, Neumüllers-Klauser 1984, 73–81.
92 | Jochen Hermann Vennebusch

ple, at the baptismal font in St Alexandri Church in Einbeck, which was cast by
Henning Regner in 1427 (Fig. 13).20

Fig. 13: General view of the baptismal font, Henning Regner, 1427, bronze, St Alexandri Church,
Einbeck; © Andreas Lechtape.

There, among other texts, are the subsequently engraved inscriptions ‘got ⋅
gheue ⋅ de(n) ⋅ sele(n) ⋅ // rat ⋅ de ⋅ dit ⋅ ghe ⋅ m(a)k(e)t ⋅ h(a)t ⋅ regner(us) ⋅/
hen(n)y(n)g(us) ⋅’21 and, in connection with a depiction of the donor,

||
20 Heege 2000, 19; Kellmann 2017, 262–263; see on the inscriptions Hülse 1996, 14–15, no. 10.
21 ‘God grant counsel to the soul of, who made this font, Henning Regner’.
 Commemoration, Explication, and Obligation | 93

‘d(omi)n(u)s ⋅ dege(n)hard(us) ⋅ ree ⋅ orate ⋅ pro ⋅ dato(r)e ⋅’.22 A remarkable cor-

respondence with the Dortmund baptismal font can be seen on the object in
Einbeck, for in both cases – as on the baptismal font in Lübeck – the caster’s
inscription is actually in Low German, while in Einbeck, the donor’s inscription
is in Latin. Admittedly, it is not very likely that Johann Winnenbrock knew this
baptismal font, but a fundamental strategy is possibly discernible here, which
consisted of a kind of Low German ‘advertising’ for the objects he produced as
well as publicity for his workshop and which was sometimes combined with an
appeal for intercessory prayer.

5 Interpreting the sacrament: The Latin

theological inscription on the cuppa
The Latin inscription on the cuppa refers to the effect of the sacrament of bap-
tism, which according to theological doctrine provides the baptised with com-
plete forgiveness of sins. Remarkably, it has not yet been recognised that the
first part of this text (‘Baptismi fructus est plena remissio culpe gracia confertur
et pena remictitur omnis’) is a quotation from the Liber Floretus, an anthology of
1160 rhymed Latin hexameters attributed to the Cistercian abbot Bernard of
Clairvaux, but probably compiled in the fourteenth century by an unknown
author.23 The second part (‘qui crediderit et baptisat(us) fuerit salvus erit i(n)
(a)et(er)nu(m)’) quotes the Gospel according to Mark (Mark 16:16) and, thus,
continues this theological interpretation of baptism. This text – only differing in
details – is also found on the bronze baptismal font in St Bartholomew’s in Wit-
tenburg in Mecklenburg, which was created by a caster named Wilkinus in 1342,
revealing a certain tradition of the use of this text on medieval bronze baptismal
fonts.24
Without being able to go into the development of baptismal theology in
detail at this point, only a brief classification of this text shall follow: the
characterisation of the effect of baptism made here goes back to the view
formulated by the Apostle Paul in the Letter to the Romans. In baptism, the
person who was to be baptised was originally immersed in the water and
then lifted out of the water, symbolically expressing dying, being buried and

||
22 ‘The Venerable Degenhard Ree. Pray for the founder’.
23 Liber Floretus, V, 366–367, ed. Orbán 1979, 17.
24 See on this baptismal font, Bartels and Waack s.a., s.p. [4]; Schlie 1899, 55; Mundt 1908, 34–36.
94 | Jochen Hermann Vennebusch

rising with Christ (immersion baptism), to thereby die to sin and be resur-
rected to a new life without sin.25 For this purpose, the baptismal fonts de-
signed as piscinas were sometimes but not always so deep that an adult per-
son could be immersed in them.26 Before this, however, the adult candidates
for baptism in the early days of Christianity had to answer the questions ad-
dressed to them about the faith of the church, which, in the course of the
Middle Ages, was passed on to the godparents, after infant baptism had be-
come established.27 It is conceivable that even in the high and late Middle
Ages the children were completely immersed in the baptismal water, as the
baptismal font also had a correspondingly large diameter and a sufficient
depth of the cuppa, but depictions created in the middle of the sixteenth
century, such as a miniature in a ritual book from St. Gallen, show that only
the head of the baptised person was doused with water (infusion baptism)
(Fig. 14).28
Let us return to the inscription on the cuppa and the question of why this
inscription in Latin was placed there. First of all, it must be remembered that
the text is a quotation from a Latin anthology and from the Gospel according to
Mark, which already explains the Latin wording. The theological content and
liturgical relevance of the inscription suggests the use of the Latin language, the
prescribed language of the Roman church, although numerous bronze baptis-
mal fonts created at about the same time also have a sacramental-theological
inscription written in the vernacular, although these are not (translated) quota-
tions from theological treatises or aphorism collections. An example of this is
the baptismal font in St Lawrence’s church in Kirchgellersen near Lüneburg,
made around 1450 probably by the Lüneburg caster Cord Vribusch (Fig. 15).

||
25 Koch 1910, 3–5, 10–11; Schwarzmann 1950, 16–27; Schneider 1952, 44–48; see on the bap-
tismal theology in early Christianity, Jensen 2012, 30–51.
26 See on piscinas, Kleinheyer 1989, 59–61.
27 Angenendt 1987, 317–318; Fürst 2008, 132.
28 See on this manuscript, Schmid 1954, 147.
 Commemoration, Explication, and Obligation | 95

Fig. 14: Depiction of an infant baptism ritual, parchment, around 1555, Abbey library of St.
Gallen, Cod. Sang. 442, p. 43; © Stiftsbibliothek St. Gallen.
96 | Jochen Hermann Vennebusch

Fig. 15: General view of the baptismal font, Cord Vribusch, around 1450, bronze, St Lawrence’s
church, Kirchgellersen; © Jochen Hermann Vennebusch.
 Commemoration, Explication, and Obligation | 97

The following sacramental-theological text, which reminds us of the text in

Dortmund in terms of content, is on the surrounding inscription band below the
rim: ‘In de ere svnte lavrencii bin ik gheghoten [2 coins] de dope to alle stv(n)en
maket reyne de(n) mi(n)chen van allen svnden [3 coins]’.29 Because of its loca-
tion on the cuppa, the Latin inscription on the baptismal font of the St Reinoldi
church in Dortmund is not as easy to read as the Low German inscriptions on
the foot or the upper rim. This is also due to the fact that the sacramental-
theological inscription slopes slightly downwards and is always obscured by the
widely protruding rim when the viewers are standing quite close to the baptis-
mal font, so that it can only be read from a certain distance. In addition, the
pinnacles formerly surrounding the baptismal font impaired the general legibil-
ity of the inscription. Whether it was intended that the inscription could be read
and understood at all, however, is questionable; the use of Latin alone severely
restricts the circle of those who received information about the effect of the
sacrament of baptism through the text. The affirmative statement about the
forgiveness of sins conveyed by baptism could, therefore, also be a phenome-
non of restricted scriptural presence.30 The inscription generally identified the
artefact as a vessel used for the liturgy of baptism and irrevocably inscribed on
it the effects associated with the sacrament administered on it, admittedly with-
out being able to influence the efficacy of the ritual act itself or even be constitu-
tive of it. While some inscriptions on bronze baptismal fonts recite prayers or
liturgical texts which are addressed to God and perpetuated again and again
due to their placement, this is not the case with the Dortmund baptismal font,
since a statement is made here about the effect of the sacrament without being
addressed to an addressee who would have any influence on this themselves.31
Nevertheless, by casting this inscription in bronze, a permanent expression was
given to the baptismal grace bestowed on this object.

||
29 ‘I was cast in honour of St Lawrence. Baptism makes a person clean from all sins at all
hours’. See on this baptismal font, St. Laurentius Kirche Kirchgellersen, Kirchgellersen, s.n., s.a.,
s.p.; Maria-Magdalenen-Kirche Lauenburg/Elbe, Lauenburg/Elbe, s.n., s.a., 8–9; Mithoff 1877,
110; Wrede 1908, 25–28; Erhardt 1939, pl. 36; Weiß 1981, 87; Hessing 1987, 30; Vennebusch
2022c, 24–27.
30 See Frese, Keil and Krüger 2013, 241–242; Keil et al. 2018, 2–6.
31 See on the mediality of medieval baptismal fonts and their inscriptions, Vennebusch 2023.
98 | Jochen Hermann Vennebusch

6 Teaching the duties: The vernacular admonition

of the godparents on the edge of the cuppa
The inscription on the upper edge, however, seems to be different. In contrast to
the Latin text and also to the caster’s inscription, an addressee is clearly recog-
nisable here, which means that the inscription begins immediately: ‘O ⋅ gees-
tilike ⋅ vader ⋅ unde ⋅ mod(er) ·’, ‘O spiritual father and mother’. The inscription
uses the Germanised Latin phrase pater spiritualis or mater spiritualis here, by
which the godparents were designated as spiritual teachers.32 The office of god-
parents itself can be traced back to early Christianity, with godparents accom-
panying and vouching for the originally adult candidate in preparation for bap-
tism during the catechumenate.33 In the Middle Ages, the godparents brought
those to be baptised, now babies, to the church, asked for baptism and an-
swered the baptismal questions on behalf of the candidate and opposed evil.34
The further duties of the godparents are formulated in the text on the Dortmund
baptismal font: they are obliged to teach their godchildren the faith and teach
them the Lord’s Prayer and the Ten Commandments.35 In return for this service,
they are figuratively promised the Kingdom of Heaven.
A remarkable correspondence with the Dortmund baptismal font can be
found in a liturgical book from the diocese of Schleswig, printed in 1512 and
preserved in various places, which will be discussed in more detail below. Alt-
hough the city of Dortmund belonged to the archbishopric of Cologne in the
Middle Ages, the relative temporal proximity of the baptismal font and the ritual
book allows a reference to be made. The liturgical book from northern Germany
formulates this as follows: ‘Tunc tam sacerdos atque patrini dicant orationem
dominicam. Aue maria. et Credo. quod est simbolum apostolorum.’36 Both the
rubrics and the prayers are written here in Latin, but then there is a remarkable
instruction in the further course:

||
32 Dannecker 2005, 113; Angenendt 2009, 475.
33 Angenendt 1987, 314–321; Daschner 2006. See on the office of godparents in the early Mid-
dle Ages, Dick 1939; Bailey 1952, 1–26; Lynch 1986; Müller 2012, 103–105.
34 Bailey 1952, 37–50.
35 As contemporary written sources show, the expression ‘de(n) gelow(en)’ could also stand
for the text of the Apostles’ Creed. See on these texts within the Church’s catechesis, Wei-
denhiller 1965, 17–20.
36 ‘Then the priest and godparents thus say the Lord’s Prayer, the Ave Maria and the Credo,
which is the Apostles’ Creed’ (emphasis in the original); Liber Agendarum ecclesie, ed. Freisen
1898, 45.
 Commemoration, Explication, and Obligation | 99

Deinde dicat sacerdos vulgariter ad leuantes puerum sic laica lingua Ick ghebede Iu by iu-
wer zelen salicheyt / Wen dat kindt komet so synen iaren der vornuft. dat gy dat suluighe
kindt. vnderwisen in den rechten cristen louen. Dat Pater noster. Aue maria. vnd
tuchnisse ghewen dat id hebbe de hillighen cristlijen dope vntphangen.37

The recitation of the Lord’s Prayer, the Ave Maria and the Apostles’ Creed by the
priest and the godparents equally represents the highly abbreviated and, above
all, completely recontextualized reminiscence of Early Christian preparatory
rites for baptism. At the end of the catechumenate, which usually lasted three
years,38 was the time of competence, which the candidates for baptism under-
went within the forty-day pre-Easter penitential period (Quadragesima) before
the date of baptism in the Easter Vigil. Within this period of preparation, known
today as Lent, scrutiny took place as part of the celebration of Mass, during
which candidates for the sacrament of baptism were examined, exorcised sev-
eral times and blessed at the end of each celebration.39 The church father Am-
brose, who served as a bishop in Milan in the fourth century, reports on the
local customary practice of Traditio, the handing over of the profession of faith
within the competent period before baptism.40 This only took place after the
completion of the scrutiny, during which the baptismal minister could convince
himself of the sanctification of the body and soul of the baptismal candidates by
means of exorcisms.41 The Traditio took place on the Sunday before Easter after
the Liturgy of the Word of the Mass and consisted of an address in the course of
which the Creed was recited several times by the bishop and commented on in
sections.42 Ambrose also inculcated in this context that the text of the Creed
should be learnt by heart and not written down, for as an expression of arcane
discipline it should not be known to heretics or to catechumens not yet deemed
worthy of this step in the context of baptismal preparation on the basis of the

||
37 ‘Then the priest says in everyday language to those who hold up the child, as follows in
secular language (lay language): I command you by the blessedness of your souls, when the
child comes to his years of reason, that you instruct the same child in the right Christian faith,
the Lord’s Prayer, Ave Maria, and bear witness that he has received Christian baptism’ (empha-
sis in the original); Liber Agendarum ecclesie, ed. Freisen 1898, 45; see on this ritual, Spital
1968, 90–95; see on vernacular addresses (exhortationes) in late medieval ritual printings,
Kopp 2016, 19–37.
38 Kleinheyer 1989, 40.
39 Angenendt 2009, 463.
40 See on this ritual Lange 2008, 9; cf. on other local practices of the Traditio, Fürst 2008, 118–119;
see on the Traditio in general, Corblet 1881, 351–353.
41 See on the scrutiny, Ambrosius, De sacramentis, ed. Schmitz 1990, 23–24.
42 Ambrosius, De sacramentis, ed. Schmitz 1990, 25; Kleinheyer 1989, 69–70.
100 | Jochen Hermann Vennebusch

preceding scrutiny.43 It is not known when the Redditio, the ceremonial recita-
tion of the Creed by the baptismal candidates as the concluding rite of the peri-
od of competence before baptism, took place in Milan, because Ambrose only
mentions that there was such a liturgical act. By contrast, it was customary ‘in
Rome for the redditio symboli to be performed in a prominent place in full view
of all the faithful’.44 It can be assumed that the Redditio served to check the
knowledge of the faith of the adult candidates for baptism and, thus, equally
the success of the preceding period of preparation for the reception of the sac-
rament of baptism.45 Furthermore, the usual Redditio in Rome in the face of the
assembled congregation could indicate that the greatest possible number of
witnesses to the baptismal candidate’s confession of the Christian faith was to
be ensured.
In the seventh century, as liturgical sources attest, the long catechumenate
was condensed to the earlier period of competence during Lent.46 According to
the Sacramentarium Gelasianum and the Ordo Romanus XI, the Traditio of the
central texts of the Christian faith took place in the fourth week of this peniten-
tial season.47 Among these texts were the four Gospels, the beginnings of which
were sung by deacons, the Apostles’ Creed and the Lord’s Prayer.48 In addition,
all texts were explained by the presider of the liturgy through an already pre-
formulated explanation. Finally, on the morning of Holy Saturday, the Redditio
or Recitatio symboli took place, which in early Christianity, originally ‘showed
that one had successfully completed the catechetical instruction’49 by testing
the knowledge of the text of the Creed.50 Furthermore ‘[i]t served as a statement
of the personal faith of the candidate and admitted the candidate to baptism’.51
In the early Middle Ages, however, when the Sacramentarium Gelasianum and
the Ordo Romanus XI were compiled, this catechetical instrument had already
been transferred to a priest, since infant baptism had meanwhile become estab-
lished, so that the liturgical sources no longer presupposed that the candidates

||
43 Fürst 2008, 120.
44 Fisher 1965, 6 (emphasis in the original).
45 Ambrosius, De sacramentis, ed. Schmitz 1990, 27–28.
46 Angenendt 1987, 288–289.
47 See on this ritual, Fisher 1965, 9–10; Angenendt 1987, 277, 289; Kleinheyer 1989, 110–112;
Kunzler 2003, 400.
48 Gelasian Sacramentary, ed. Wilson 1894, XXXV–XXXVI; see also Wahle 2008, 32.
49 Old 1992, 7.
50 Angenendt 1987, 278.
51 Old 1992, 7.
 Commemoration, Explication, and Obligation | 101

for baptism could recite the text of the Creed by heart.52 It is noteworthy, howev-
er, that although the rite originally conceived for adult baptismal candidates
was greatly compressed in time, it remained almost unchanged in itself in terms
of the individual elements, even though it was now infants who were to be bap-
tised.53 The ‘distribution of roles’ within the liturgy especially changed and,
thus, the actual character of the time of preparation for baptism as an institu-
tionalized and clearly regulated period of instruction in the faith and consolida-
tion of knowledge about the faith gave way to a sequence of barely comprehen-
sible rituals. It was no longer the candidate for baptism who was introduced to
the mysteries of the faith and examined on them, but rather various ritual acts
were performed on him, so that one can speak of ‘a displacement of the cate-
chetical by the liturgical and thus of the intelligible by the ceremonial’.54 While
the liturgical books from the early Middle Ages still assume, in principle, that
the baptised recite the texts given to them, this task is transferred to the godpar-
ents in ninth century sacramentaries.55 Arnold Angenendt, therefore, describes
the obligation to be responsible for the further Christian education of the bap-
tised child as the most important task of the godparents in the course of the
Middle Ages.56 Possibly the reference of the ritual text from Schleswig ‘Wen dat
kindt komet so synen iaren der vornuft’ represents a recourse to the criticism of
the missionary Winfried-Bonifatius, who complained about the lack of catech-
esis during the preparation for and administration of baptism, since false priests
(pseudo-sacerdotes) do not even teach the solemn vows which every disciple of
the faith, if they are old enough to have the understanding, should grasp and
understand in their heart, nor interrogate them to those whom they are to bap-
tise.57 Therefore, finally, in Carolingian times, special emphasis was placed on
the fact that the Lord’s Prayer and the Apostles’ Creed should be part of the
knowledge of faith of every believer.58 Even in the late Middle Ages, the central

||
52 Angenendt 1987, 289–290; Old 1992, 7–8; Wahle 2008, 32.
53 Old 1992, 6–9.
54 Bakhuizen van der Brink 1970, 68: ‘Verdrängung des Katechetischen durch das Liturgische
und damit des Verständigen durch das Zeremonielle’; see on this development, Schlegel 2012,
75–81.
55 Angenendt 1987, 317.
56 Angenendt 1987, 318; see also Weidenhiller 1965, 14.
57 Bonifatii epistolae, ed. Rau 1968, Ep. 80, 262: ‘nec ipsa sollempnia verba, quae unusquisque
caticuminus, si talis aetatis est, ut iam intellectum habeat, sensu cordis sui percipere et intelle-
gere, nec docent nec quaerent ab eis, quos baptizare debent’. See Angenendt 1987, 290; Wahle
2008, 40; Angenendt 2009, 470.
58 Jungmann 1941, 170–171.
102 | Jochen Hermann Vennebusch

texts of the Christian faith were repeatedly taught to the faithful, as they were
also among the texts that were a prerequisite for absolution in confession: in
1451, for example, the papal legate and cardinal Nicholas of Cusa (1401–1464)
had a ‘catechism board’ hung in St Lamberti church in Hildesheim, on which
the vernacular texts of the Lord’s Prayer, the Ave Maria, the Apostles’ Creed and
the Ten Commandments were listed.59 The introduction before the texts, which
is also in the vernacular, states that the cardinal had discovered during a visita-
tion that the faithful were unable to say the Lord’s Prayer, so he wanted to en-
sure that these texts, which are essential for salvation, were taught. In sum-
mary, then, it can be said that by the time the Dortmund baptismal font was
created, the rite of the baptismal liturgy had been contracted to the point that,
according to Hughes Oliphant Old, it was a ‘telescoped rite’. He characterises it
as follows:

It was a rite made up of many rites which in earlier centuries had extended over a period
of several weeks and then in the Middle Ages had been compressed into a single rite
which could be performed in about a half an hour.60

Unfortunately, the surviving late medieval and early modern rituals from the
Archdiocese of Cologne do not give any indication that the godparents were
instructed about their obligations in the vernacular in connection with the bap-
tismal liturgy. The oldest surviving printed ritual containing the baptismal lit-
urgy, the Agende in chatolicis ecclesiis observande61 manufactured around 1482
by Ludwig von Renchen, lists the text of the Creed, the Lord’s Prayer and the
Ave Maria after the rubric ‘Deinde ponat manu(m) super caput baptisa(n)di et
cathec(izat) dice(n)s simbolu(m)’62 and the Latin rubric ‘Et com(m)ittat patrinis
q(uod) informent baptisandu(m) de fide et or(ati)one d(omi)nica’.63 Whether this
is a modification of the Traditio or the Redditio of the texts of faith cannot be
determined with certainty. According to Alois Stenzel, this rite is more likely to
be interpreted as a Traditio, since in the early medieval liturgical sources this
act is accompanied by the laying on of hands, which is still mentioned in the

||
59 See on this ‘catechism board’, Wulf 2003, 394–397, no. 167. The author thanks Dr Christine
Wulf and Dr Jörg H. Lampe for their reference to this object.
60 Old 1992, 1.
61 See on this liturgical book, Vollmer 2000, 46; Vollmer 1994, 45–62.
62 ‘Then the priest puts his hand on the head of the child to be baptised and instructs [him] in
the faith by saying the Creed’, Agende in chatolicis ecclesiis observande s.a., s.p.; see also
Vollmer 2000, 50; Vollmer 1994, 221–223.
63 ‘And he should entrust to the godparents that they teach the child to be baptised the faith
and the Lord’s Prayer’, Agende in chatolicis ecclesiis observande s.a., s.p.
 Commemoration, Explication, and Obligation | 103

late medieval ritual.64 A vernacular text, however, addressed directly to the

godparents and recited by the priest, as in the ritual from the diocese of Schles-
wig, cannot be found in the print of the liturgical book used in Cologne, nor are
there any vernacular exhortations and addresses in manuscripts of rituals in
general.65 A vernacular address to the godparents, on the other hand, can be
found in the 1591 printed ritual Agenda ecclesiastica sive legitima ac solennis
sacramentorum Ecclesie administratio from the diocese of Münster in Westphalia:

Ick vermane juw Paden / dat ghy dit Kind juw laten befollen syn / neuen synen Natuerli-
chen Olderen / Also / wenn eth tho synen verstendigen Jahren kommen werdt / dat ghy
idt (da desseluigen Olderen daran seumich weren oder verstoruen) mit ganzem Flyte un-
derrichten vnd lehren / stede und fast in der hylligen Christlichen Catholischen vnd Apos-
tolischen Kercken (darin idt idzunder de Doepe entfenget) to blyuen. Derseluigen wahren
vnd vngetwyuelten Gelouen alletydt standthafftiglich to holden / vnd darbeneuen oick
dat Vader vnse / de Engelsche Grote / de twelff Artickel deß hilligen Geloues / vnd de tein
Gebodde Gods.66

Although this was printed after the final reformation of the city of Dortmund in
157067 and the Council of Trent (1545–1563), the geographical proximity of Dort-
mund to the diocese of Münster, on the one hand, and the expected continua-
tion of local rites and customs in the liturgy, on the other hand, could reflect a
tradition of admonishing the godparents in the vernacular within the baptismal
liturgy that was possibly already recognisable in the Middle Ages.
Even if there is no absolute literal correspondence between the inscription
on the Dortmund baptismal font and vernacular liturgical formulas from late
medieval ritual printings, the clear correspondence in content between the texts
cannot be denied. Finally, the question remains to whom the inscription on the
upper rim of the bronze baptismal font was addressed: although it is conceiva-
ble that the very brief exhortation to the godparents served as a reminder to the
priest as baptismal minister during the ceremony, it is more likely that this was
intended to directly address the godparents who knew the scriptures. Neverthe-

||
64 Stenzel 1958, 277–278.
65 Kopp 2016, 19.
66 ‘I admonish you, godparents, to make this child your charge, next to its natural parents, so
that when it comes to its years of understanding (should its parents fail to do so or die) you will
teach and instruct it with all diligence, to abide steadfastly and firmly in the holy Christian,
Catholic and Apostolic Church (in which he now receives baptism), to hold fast to her true and
undoubted faith at all times, and besides this, to learn the Lord’s Prayer, the Angelic Salutation
[the Hail Mary], the Twelve Articles of the Holy Faith [the Apostles’ Creed] and the Ten Com-
mandments of God’, Agenda ecclesiastica 1592, 17.
67 See on the reformation in Dortmund, Freitag 2017, 212–215.
104 | Jochen Hermann Vennebusch

less, in this case, the group of people who were able to understand this text at
all remains small, despite the vernacular wording. It is possible that the priest
pointed out the inscription and its content to the godparents during the course
of the baptismal ceremony, and that they were perhaps even able to literally
grasp this text haptically, which was intended to make their duties abundantly
clear to them. Since it can be assumed that the baptismal font was formerly
placed in the west of the church and that one usually passed this artefact when
entering the church, the godparents were implicitly reminded of their duties
again and again. This assumption is not invalidated by the fact that, as the two
busts show, a font cover was used to lock this vessel when it was not being used
for baptism.

7 Conclusion
In summary, we will now discuss the potential reasons that might have led to
the affixing of the various inscriptions on the baptismal font of St Reinoldi in
Dortmund, making the object a multilingual written artefact: first of all, the
question of the recipients is certainly central to the choice of the respective lan-
guage. The addressees are only explicitly mentioned in the inscription on the
upper rim, which addresses the godparents directly and refers to their obliga-
tions. But the question remains to whom the inscriptions on the foot and on the
cuppa are directed. The vernacular caster’s inscription on the foot can probably
be seen as the self-confident signature of Johann Winnenbrock, who presents
himself as a bell founder and Dortmund citizen who made the baptismal font for
the local council church and possibly also tried to generate further commissions
due to the special quality of his casting. It can also be assumed that Winnen-
brock implicitly wanted to encourage the readers to commemorate him through
the signature and the vernacular inscription, but this aspect is not nearly as
clear as on other baptismal fonts due to the lack of reference to God and the
omitted call to intercessory prayer. However, it must at least be conceded that
his name is permanently present on an object used liturgically, which is highly
significant for the question of memoria, the provision for the heavenly afterlife.68
By contrast, the determination of an addressee for the inscription on the cuppa
is considerably more complicated. The statement made there about the effects

||
68 Johannes Tripps addresses the question of the presence of donor names on liturgical ob-
jects and their legibility using the example of medieval chalices. See Tripps 2018, 335–344.
 Commemoration, Explication, and Obligation | 105

of the sacrament of baptism is formulated in the indicative, it is an affirmative

statement of fact in the theological sense. In this context, the quotation from the
Gospel according to Mark, the second part of the inscription, could be ad-
dressed to the person being baptised or even to those who have already been
baptised; after all, the baptismal grace is here associated with faith. In this
sense, baptism can, thus, be seen as a fundamental condition for salvation,
which, however, only finds its fulfillment through the faith of the person being
baptised. For this reason, it seems to be meaningful to point out the firmness in
faith, which – according to the addressees in the printed rituals – the godpar-
ents should also point out to their godchildren repeatedly. But even if these two
Latin quotations from the Liber Floretus and the Gospel according to Mark, re-
spectively, could be read, they certainly could not be comprehended by all be-
lievers or baptised people, which is not only due to the theological background
but, above all, the Latin language of the inscription. Again, it could be assumed
that the celebrant, when administering baptism, referred to this saying and the
promise of the forgiveness of sins and the promise of salvation expressed
through it. With the baptismal font, these quotations were once again before the
eyes of those entering the church and reminded them of the reception of the
sacrament. As has already been mentioned, the inscription could be a phenom-
enon of restricted scriptural presence, so that it is not necessarily important that
an inscription is read and understood, but primarily that it is present on the
object. While this is obvious for the founder’s inscriptions, since the presence of
the corresponding name on the liturgical device can be interpreted as a contri-
bution to memoria, it is not nearly as clear for the complex theological inscrip-
tion.69 In this respect, the Latin text could mark the baptismal font as a sacred
object and, because it was written in the official language of the church and
liturgy, it could receive a higher authority through the Latin quotations which
can be regarded as an additional authentication of the statement in the inscrip-
tion about the salvation of the soul obtained through baptism and the for-
giveness of sins that had taken place. Finally, the baptismal font stands in a
field of tension between liturgical change and theological continuity: while the
form of the baptismal font clearly reflects the practice of infant baptism that has
prevailed since the early Middle Ages, since the immersion baptisms of adults
practiced originally are no longer possible in the cuppa, the Latin inscription
reflects the eternally valid theology of baptism that is already grounded in nuce
in the New Testament Pauline epistolary literature. By contrast, the Low German
inscription on the upper rim testifies to the change in the office of godparent. It

||
69 Tripps 2018, 347.
106 | Jochen Hermann Vennebusch

is no longer the adult candidates for baptism who receive an introduction to the
Christian faith and learn the most important texts during the catechumenate,
but the task of imparting faith and knowledge has passed to the godparents, to
whom this duty was clearly inculcated by the priest during the baptismal litur-
gy, as can be seen in some late medieval prints of rituals.

Acknowledgements

The research for this article was funded by the Deutsche Forschungsgemein-
schaft (DFG, German Research Foundation) under Germany’s Excellence Strate-
gy – EXC 2176 ‘Understanding Written Artefacts: Material, Interaction and
Transmission in Manuscript Cultures’, project no. 390893796. The research was
conducted within the scope of the Centre for the Study of Manuscript Cultures
(CSMC) at the Universität Hamburg. The author is grateful to Kaja Harter-
Uibopuu (Hamburg), Klaus Gereon Beuckers (Kiel), Jörg H. Lampe, Christine
Wulf and Katharina Kagerer (Göttingen), Claus Peter (Hamm) and Benjamin
Heu (Cologne) for their advice and support.

References
Primary sources
Agenda ecclesiastica sive legitima ac solennis sacramentorum Ecclesie administratio, Münster:
Raesfeldt, 1592.
Agende in chatolicis ecclesiis observande, Cologne: Ludwig von Renchen, s.a. [around 1482].
Ambrosius, De sacramentis – De mysteriis. Über die Sakramente – Über die Mysterien, ed.
Josef Schmitz (Fontes Christiani, 3), Freiburg: Herder, 1990.
Bonifatii epistolae: Willibaldi vita Bonifatii quibus accedunt aliquot litterarum monumenta
coaeva. Briefe des Bonifatius: Willibalds Leben des Bonifatius nebst einigen zeitgenössi-
schen Dokumenten, ed. Reinhold Rau (Ausgewählte Quellen der deutschen Geschichte
des Mittelalters, Freiherr vom Stein-Gedächtnisausgabe, 4b), Darmstadt: Wissenschaftli-
che Buchgesellschaft, 1968.
Gelasian Sacramentary. Liber sacramentorum Romanae Ecclesiae, ed. Henry Austin Wilson,
Oxford: Clarendon, 1894.
Liber Agendarum ecclesie et diocesis Sleszwicensis. Katholisches Ritualbuch der Diözese
Schleswig im Mittelalter, ed. Joseph Freisen, Paderborn: Junfermann, 1898.
Liber Floretus, Herausgegeben nach der Hs. Utrecht, U. B. 283, ed. Árpád Orbán (Beihefte zum
Mittellateinischen Jahrbuch, 16), Kastellaun: Henn, 1979.
 Commemoration, Explication, and Obligation | 107

Secondary literature
Angenendt, Arnold (1987), ‘Der Taufritus im frühen Mittelalter’, in Ovidio Capitani (ed.), Segni e
riti nella chiesa altomedievale occidentale (Settimane di studio del Centro italiano di studi
sull’alto medioevo, 33), Spoleto: Centro italiano di studi sull’alto medioevo, 275–321.
Angenendt, Arnold (2009), Geschichte der Religiosität im Mittelalter, 4th edn, Darmstadt:
Wissenschaftliche Buchgesellschaft.
Appuhn, Horst (1970), Dortmund (Westfälische Kunst), Munich: Deutscher Kunstverlag.
Bailey, Derrick Sherwin (1952), Sponsors at Baptism and Confirmation: An Historical Introduc-
tion to Anglican Practice, London: S.P.C.K.
Bakhuizen van der Brink, Jan Nicolaas (1970), ‘Sakrament und Ethik in der frühen Kirche’, in
Patrick Granfield and Josef Andreas Jungmann (eds), Kyriakon: Festschrift Johannes Quas-
ten, Münster: Aschendorff, 59–68.
Bartels, Gisela and Martin Waack (s.a.), Evangelische Stadtkirche St. Bartholomäus zu Witten-
burg, Wittenburg: s.n.
Böker, Doris (1997), Landkreis Cuxhaven (Baudenkmale in Niedersachsen, 19), Hameln: Vieweg.
Corblet, Jules (1881), Histoire dogmatique, liturgique et archéologique du sacrement de baptême:
Tome premier, Paris: Victor Palmé / Brussels: Joseph Albanel / Geneva: Henri Trembley.
Dannecker, Klaus Peter (2005), Taufe, Firmung und Erstkommunion in der ehemaligen Diözese
Konstanz: Eine liturgiegeschichtliche Untersuchung der Initiationssakramente (Liturgie-
wissenschaftliche Quellen und Forschungen, 92), Münster: Aschendorff.
Daschner, Dominik (2006) ‘Art. Pate’, in Walter Kasper (ed.), Lexikon für Theologie und Kirche,
vol. 7, Freiburg/Breisgau: Herder, col. 1450.
Deckert, Hermann, Oskar Kiecker, Hans Lütgens and Lothar Pretzell (1939), Die Kunstdenkmale des
Kreises Soltau (Die Kunstdenkmale der Provinz Hannover, 3.6), Hannover: Theodor Schulze.
Dick, Ernst (1939), ‘Das Pateninstitut im altchristlichen Katechumenat’, Zeitschrift für katholi-
sche Theologie, 63: 1–49.
Erhardt, Alfred (1939), Mittelalterliche Taufen aus Erz und Stein, Hamburg: Ellermann.
Fisher, John D. C. (1965), Christian Initiation: Baptism in the Medieval West, A Study in the
Disintegration of the Primitive Rite of Initiation (Alcuin Club Collections, 47), London:
S.P.C.K.
Freitag, Werner (2017), Die Reformation in Westfalen: Regionale Vielfalt, Bekenntniskonflikt und
Koexistenz, Münster: Aschendorff.
Frese, Tobias, Wilfried E. Keil and Kristina Krüger (2013), ‘Zur Problematik restringierter
Schriftpräsenz: Zusammenfassung dieses Bandes’, in Tobias Frese, Wilfried E. Keil and
Kristina Krüger (eds), Verborgen, unsichtbar, unlesbar: Zur Problematik restringierter
Schriftpräsenz (Materiale Textkulturen, 2), Berlin: De Gruyter, 233–242.
Fritz, Rolf (1933), Dortmunder Kirchen und ihre Kunstschätze, Dortmund: Krüger.
Fritz, Rolf (1956), ‘Kunstwerke der Reinoldi-Kirche’, in Hans Lindemann (ed.), St. Reinoldi in
Dortmund, Dortmund: Crüwell, 80–95.
Fürst, Alfons (2008), Die Liturgie der Alten Kirche: Geschichte und Theologie, Münster:
Aschendorff.
Haupt, Richard (1888), Die Bau- und Kunstdenkmäler der Provinz Schleswig-Holstein mit Aus-
nahme des Kreises Herzogtum Lauenburg, vol. 2, Kiel: Ernst Homann.
Heege, Elke (2000), Münsterkirche St. Alexandri Einbeck (Kleiner Kunstführer, 2434), Regens-
burg: Schnell & Steiner.
108 | Jochen Hermann Vennebusch

Hessing, Erich (1987), Die Kirchen im Landkreis Lüneburg (Großer Kunstführer, 138), Munich:
Schnell & Steiner.
Hülse, Horst (1996), Die Inschriften der Stadt Einbeck (Die Deutschen Inschriften, 42), Wiesba-
den: Reichert.
Jensen, Robin M. (2012), Baptismal Imagery in Early Christianity: Ritual, Visual, and Theologi-
cal Dimensions, Grand Rapids, MI: Baker Academic.
Jungmann, Josef Andreas (1941), Gewordene Liturgie: Studien und Durchblicke, Innsbruck:
Felizian Rauch.
Kämmer, Hans-Joachim (1980), ‘Wertvolle Gegenstände und Abendmahlsgeräte in unserer
Kirche’, in Ev.-luth. Kirchengemeinde Neuenkirchen (ed.), 100 Jahre St.-Bartholomäus-
Kirche Neuenkirchen 1880–1980, Soltau: Mundschenk, 47–57.
Keil, Wilfried E., Sarah Kiyanrad, Christoffer Theis and Laura Willer (2018), ‘Präsenz, Sichtbar-
keit und Unsichtbarkeit von Geschriebenem und Artefakten’, in Wilfried E. Keil, Sarah
Kiyanrad, Christoffer Theis and Laura Willer (eds), Zeichentragende Artefakte im sakralen
Raum: Zwischen Präsenz und Unsichtbarkeit (Materiale Textkulturen, 20), Berlin: De Gruy-
ter, 1–15.
Kellmann, Thomas (2017), Stadt Einbeck (Baudenkmale in Niedersachsen, 7/3), Petersberg:
Imhof.
Kiecker, Oskar, Wilhelm Lenz and Heinrich Rüther (1956), Die Kunstdenkmale des Kreises Land
Hadeln und der Stadt Cuxhaven (Die Kunstdenkmale des Landes Niedersachsen), Munich:
Deutscher Kunstverlag.
Kleinheyer, Bruno (1989), Sakramentliche Feiern I: Die Feiern der Eingliederung in die Kirche
(Gottesdienst der Kirche: Handbuch der Liturgiewissenschaft, 7/1), Regensburg: Pustet.
Koch, Wilhelm (1910), Die Taufe im Neuen Testament (Biblische Zeitfragen, 10), Münster:
Aschendorff.
Kopp, Stefan (2016), Volkssprachliche Verkündigung: Die Modellanreden in den Diözesanritua-
lien des deutschen Sprachgebietes (Studien zur Pastoralliturgie, 42), Regensburg: Pustet.
Kunzler, Michael (2003), Die Liturgie der Kirche (AMATECA: Lehrbücher zur katholischen Theo-
logie, 10), 2nd edn, Paderborn: Bonifatius.
Lampe, Jörg H. (2022), ‘Die Inschriften auf den Werken Hans Apengeters und Johannes von
Halberstadts: Eine epigraphische Analyse’, in Klaus Gereon Beuckers and Jochen Her-
mann Vennebusch (eds), Hans Apengeter: Norddeutscher Bronzeguss des 14. Jahrhun-
derts im Kontext (OPERA BOREALIA: Beiträge zur norddeutschen Kunst des Mittelalters
und der Frühen Neuzeit, 1), Regensburg: Schnell & Steiner, 325–349.
Lange, Christian (2008), ‘Gestalt und Deutung der christlichen Initiation in der Alten Kirche’, in
Christian Lange, Clemens Leonhard and Ralph Olbrich (eds), Die Taufe: Einführung in Ge-
schichte und Praxis, Darmstadt: Wissenschaftliche Buchgesellschaft, 1–28.
Lindemann, Margot (s.a.), St. Reinoldi Dortmund (DKV-Kunstführer, 272/9), Munich: Deutscher
Kunstverlag.
Lübke, Wilhelm (1853), Die mittelalterliche Kunst in Westfalen, Leipzig: Weigel.
Ludorff, Albert (1894), Die Bau- und Kunstdenkmäler des Kreises Dortmund-Stadt (Die Bau- und
Kunstdenkmäler von Westfalen, 2), Münster: Schöningh.
Lynch, Joseph H. (1986), Godparents and Kinship in Early Medieval Europe, Princeton: Prince-
ton University Press.
Mithoff, Hector Wilhelm Heinrich (1877), Fürstenthum Lüneburg (Kunstdenkmale und Alterthü-
mer im Hannoverschen, 4), Hannover: Hellwing.
 Commemoration, Explication, and Obligation | 109

Mithoff, Hector Wilhelm Heinrich (1878), Herzogthümer Bremen und Verden mit dem Lande
Hadeln, Grafschaften Hoya und Diepholz (Kunstdenkmale und Alterthümer im Hannover-
schen, 5), Hannover: Hellwing.
Mithoff, Hector Wilhelm Heinrich (1885), Mittelalterliche Künstler und Werkmeister Niedersach-
sens und Westfalens lexikalisch dargestellt, 2nd edn, Hannover: Hellwing.
Müller, Andreas (2012), ‘Tauftheologie und Taufpraxis vom 2. bis zum 19. Jahrhundert’, in
Markus Öhler (ed.), Taufe (Themen der Theologie, 5), Tübingen: Mohr Siebeck, 83–135.
Mundt, Albert (1908), Die Erztaufen Norddeutschlands von der Mitte des XIII. bis zur Mitte des
XIV. Jahrhunderts: Ein Beitrag zur Geschichte des deutschen Erzgusses, Leipzig: Klink-
hardt & Biermann.
Neumüllers-Klauser, Renate (1984), ‘Schrift und Sprache in Bau- und Künstlerinschriften’, in
Karl Stackmann (ed.), Deutsche Inschriften: Fachtagung für mittelalterliche und neuzeitli-
che Epigraphik Lüneburg 1984, Göttingen: Vandenhoeck & Ruprecht, 62–81.
Ohm, Matthias, Thomas Schilp and Barbara Welzel (eds) (2006), Ferne Welten – Freie Stadt:
Dortmund im Mittelalter (Dortmunder Mittelalter-Forschungen, 7), Bielefeld: Verlag für
Regionalgeschichte.
Old, Hughes Oliphant (1992), The Shaping of the Reformed Baptismal Rite in the Sixteenth
Century, Grand Rapids, MI: William B. Eerdmans.
Pantel, Etta (2001), Landkreis Soltau-Fallingbostel (Baudenkmale in Niedersachsen, 25), Ha-
meln: Niemeyer.
Rauterberg, Claus (2006), Die St. Marienkirche in Rendsburg (DKV-Kunstführer, 263), 4th edn,
Munich: Deutscher Kunstverlag.
Rinke, Wolfgang (1985), Kunst der Spätgotik in St. Reinoldi zu Dortmund, Dortmund: Ruhfus.
Rinke, Wolfgang (1987), Dortmunder Kirchen des Mittelalters: St. Reinoldi – St. Marien – St.
Johannes Bapt. (Propstei) – St. Petri, Dortmund: Ruhfus.
Schlegel, Silvia (2012), Mittelalterliche Taufgefäße: Funktion und Ausstattung (Sensus: Studien
zur mittelalterlichen Kunst, 3), Cologne: Böhlau.
Schlie, Friedrich (1899), Die Amtsgerichtsbezirke Hagenow, Wittenburg, Boizenburg, Lübtheen,
Dömitz, Grabow, Ludwigslust, Neustadt, Crivitz, Brüel, Warin, Neubukow, Kröpelin und
Doberan (Die Kunst- und Geschichts-Denkmäler des Grossherzogthums Mecklenburg-
Schwerin, 3), Schwerin: Bärensprungsche Hofdruckerei.
Schmid, Alfred A. (1954), Die Buchmalerei des XVI. Jahrhunderts in der Schweiz, Olten: Urs Graf.
Schneider, Johannes (1952), Die Taufe im Neuen Testament, Stuttgart: Kohlhammer.
Schwarzmann, Heinrich (1950), Zur Tauftheologie des hl. Paulus in Röm 6, Heidelberg: F. H.
Kerle.
Spital, Hermann Josef (1968), Der Taufritus in den deutschen Ritualien von den ersten Drucken
bis zur Einführung des Rituale Romanum (Liturgiewissenschaftliche Quellen und For-
schungen, 47), Münster: Aschendorff.
Stein, Otto (1906), Die Reinoldikirche in Wort und Bild, Dortmund: s.n.
Stenzel, Alois (1958), Die Taufe: Eine genetische Erklärung der Taufliturgie, Innsbruck: Felizian
Rauch.
Tripps, Johannes (2018), ‘Teilnahme, Erlösung und ewiges Gedenken: Strategien der Sichtbar-
keit von Fürbittinschriften auf liturgischen Geräten’, in Wilfried E. Keil, Sarah Kiyanrad,
Christoffer Theis and Laura Willer (eds), Zeichentragende Artefakte im sakralen Raum:
Zwischen Präsenz und UnSichtbarkeit (Materiale Textkulturen, 20), Berlin: De Gruyter,
329–353.
110 | Jochen Hermann Vennebusch

Vennebusch, Jochen Hermann (2022a), ‘In Lehm geritzt, in Wachs eingekerbt, in Bronze gra-
viert: Produktionsweisen von Inschriften auf norddeutschen Bronzetaufen des Mittelalters
und ihre ästhetischen Implikationen’, Das Münster: Zeitschrift für christliche Kunst und
Kunstwissenschaft, 75: 154–165.
Vennebusch, Jochen Hermann (2022b), ‘Innovation – Variation – Rezeption: Das Taufbecken in
der Lübecker Marienkirche im Kontext des nord- und mitteldeutschen Bronzegusses’, in
Klaus Gereon Beuckers and Jochen Hermann Vennebusch (eds), Hans Apengeter: Nord-
deutscher Bronzeguss des 14. Jahrhunderts im Kontext (OPERA BOREALIA: Beiträge zur
norddeutschen Kunst des Mittelalters und der Frühen Neuzeit, 1), Regensburg: Schnell &
Steiner, 97–133.
Vennebusch, Jochen Hermann (2022c), ‘Die Taufbecken des Bronzegießers Cord Vribusch:
Rezeption und Fortschreibung Lüneburger Traditionen’, Lüneburger Blätter, 38: 13–46.
Vennebusch, Jochen Hermann (2023), ‘“Es sei eine lebendige Quelle, belebendes Wasser, eine
reinigende Welle!” – Mittelalterliche Bronzetaufbecken im Spannungsfeld von Interme-
dialität und Intermaterialität’, in Jan Stellmann and Daniela Wagner (eds), Materialität und
Medialität: Aspekte einer anderen Ästhetik (Andere Ästhetik – Koordinaten, 5), Berlin De
Gruyter, 431–461.
Vollmer, Thomas (1994), Agenda Coloniensis: Geschichte und sakramentliche Feiern der ge-
druckten Kölner Ritualien (Studien zur Pastoralliturgie, 10), Regensburg: Pustet.
Vollmer, Thomas (2000), ‘Die Kölner Ritualien: Eine bis ins 19. Jahrhundert beständige Diözesanli-
turgie’, in Albert Gerhards and Andreas Odenthal (eds), Kölnische Liturgie und ihre Ge-
schichte: Studien zur interdisziplinären Erforschung des Gottesdienstes im Erzbistum Köln
(Liturgiewissenschaftliche Quellen und Forschungen, 87), Münster: Aschendorff, 46–54.
Von Poser und Groß-Naedlitz, Hasso (1983), ‘Zur alten Kirche in Neuenkirchen’, in Gemeinde
Neuenkirchen (ed.), Neuenkirchen 1283–1983: Beiträge zur älteren Geschichte eines
Kirchspiels im ehemaligen Stift und Herzogtum Verden, Neuenkirchen: s.n., 50–67.
Wahle, Stephan (2008), ‘Gestaltung und Deutung der christlichen Initiation im mittelalterli-
chen lateinischen Westen’, in Christian Lange, Clemens Leonhard and Ralph Olbrich (eds),
Die Taufe: Einführung in Geschichte und Praxis, Darmstadt: Wissenschaftliche Buchge-
sellschaft, 29–48.
Weidenhiller, Egino (1965), Untersuchungen zur deutschsprachigen katechetischen Literatur
des späten Mittelalters: Nach den Handschriften der Bayerischen Staatsbibliothek
(Münchner Texte und Untersuchungen zur deutschen Literatur des Mittelalters, 10),
Munich: Beck.
Weiß, Gerd (1981), Landkreis Lüneburg (Baudenkmale in Niedersachsen, 22/2), Braunschweig:
Vieweg.
Wenning, Wilhelm (1996), St. Lamberti in Coesfeld (Kleiner Kunstführer, 1477), 2nd edn, Re-
gensburg: Schnell & Steiner.
Wrede, Hermann (1908), ‘Die Glocken des Landkreises Lüneburg’, Lüneburger Museumsblät-
ter, 5: 3–53.
Wulf, Christine (2003), Die Inschriften der Stadt Hildesheim, vol. 2: Die Inschriften, Jahreszah-
len und Initialen (Die Deutschen Inschriften, 58), Wiesbaden: Dr. Ludwig Reichert.
Apiradee Techasiriwan, Volker Grabowsky
Multi-scriptural and Multilingual Inscriptions
in Lan Na
Abstract: ‘Lan Na’ is the name of an ancient kingdom situated in the upper
north of present-day Thailand. Founded in the late thirteenth century, it became
fully integrated into the modern Thai state only in the late nineteenth and early
twentieth centuries. The Tai Yuan possess their own distinct language and have
developed three sets of scripts: Tham (Dhamma), Fak Kham and Thai Nithet
script. These scripts (mainly Fak Kham) were used for inscriptions and manu-
scripts (mainly Tham and Thai Nithet). In our paper, we analyse how different
scripts and languages (such as Burmese, Thai/Siamese, Chinese, Pali and Ro-
man scripts/languages, in addition to the Tai Yuan vernacular language/script)
are documented on various kinds of artefacts and what this tells us about the
cultural history of the region.

1 Objectives
The epigraphical culture of Lan Na is in particular diverse and richly document-
ed compared to other regions of the Tai world. Moreover, the various Lan Na
scripts have also been applied by other (Tai) ethnic groups who have migrated
to Lan Na in substantial numbers either voluntarily or as the result of forced
resettlements caused by war. The resulting processes of acculturation and inte-
gration of these groups into a coherent Lan Na society is also reflected by the
use of multiple scripts and languages on the same inscribed object. We pay
special attention to the social and cultural contexts which furthered the use of
multiple languages and scripts on the same inscribed artefact. We differentiate
between the presence of languages and scripts widely used within the larger
Lan Na cultural domain and those which were derived either from neighbouring
Tai writing cultures (e.g. Thai and Lao) and more distant areas (e.g. Burma,
China and Europe).

Open Access. © 2024 the authors, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-005
112 | Apiradee Techasiriwan, Volker Grabowsky

2 Historical background
The kingdom of Lan Na emerged after the conquest of the Mon polity of
Hariphunchai by King Mangrai (1292)1 and the founding of Chiang Mai (literally,
‘New City’) as the kingdom’s capital (1296). Lan Na, literally meaning ‘[the land
of] a million rice fields’, comprises the eight provinces of Thailand’s upper
north. More than 80% of its population belong to a Tai ethnic group called the
(Tai) Yuan. During the reign of King Tilokarat, the ninth king of the Mangrai
dynasty (1441–1487), the borders of Lan Na expanded into southern Yunnan
(Sipsòng Panna), the eastern part of the Shan region in present-day Burma
(Chiang Tung, Müang Nai and Müang Yòng) and north-western Laos.2 With the
extension of political power to areas further to the north, Theravada Buddhism,
which was firmly established in the main Buddhist centres of Chiang Mai,
spread throughout the upper and middle Mekong basin, and Lan Na finally
became the major centre of Buddhist scholarship in South East Asia for more
than half a century.3 Lan Na remained an independent polity until the conquest
of Chiang Mai by a Burmese army (1558). By the end of the sixteenth century,
Lan Na had ceased to exist as a unitary Burmese vassal state and split into con-
tending polities.4
Lan Na was occupied by Burmese forces for more than two and a half centu-
ries and remained under Burmese suzerainty until the late eighteenth century.
From 1775 (liberation of Chiang Mai) until 1804 (conquest of Chiang Saen), Lan
Na was liberated from Burmese rule by a joint effort of the Tai Yuan elite of Lan
Na and the resurgent Siamese kingdom. From that time, Lan Na consisted of five
autonomous vassal states: Chiang Mai, Lamphun and Lampang in the western
part, all ruled by members of the Kawila dynasty, and Phrae and Nan in the
east, each of them staying under Siamese suzerainty. Between 1874 and 1933,
the five vassal kingdoms became gradually integrated into the emerging mod-
ern nation-state of Thailand.

||
1 Unless otherwise stated all dates are CE.
2 The political dynamics of the Lan Na kingdom in its early phase is discussed in detail in
Sarassawadee Ongsakul 2005, 89–105; Grabowsky 2004, 116–125; Grabowsky 2005, 3–19.
3 The eighth Buddhist Council was held in Wat Chet Yòt or Wat Photharam Maha Wihan in the
city of Chiang Mai in 1477.
4 For the political developments in Lan Na during the period of Burmese domination, see
Sarassawadee Ongsakul 2005, 109–128; Grabowsky 2004, 149–176.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 113

Fig. 1: Map of Lan Na kingdom, c. 1450 CE; © Volker Grabowsky.

The period of Burmese domination shows a significant decline in inscriptional

activities and the gradual appearance of the Tham script as the dominant one in
Lan Na epigraphy.5 The post-Burmese period of Lan Na epigraphy is character-

||
5 The dearth of Lan Na (stone) inscriptions during the period of Burmese domination is dis-
cussed in Buchmann 2018, 124–127, who also observes: ‘Only four inscriptions form the pre-
Burmese period show a combination of Fak Kham and Dhamma-Lanna scripts. This implies
114 | Apiradee Techasiriwan, Volker Grabowsky

ised by the exclusive use of Tham script and the increased preference of writing
material other than stone, such as metal and wood.6 The forced resettlement of
numerous war captives from Tai Khün and Tai Lü inhabited areas in the eastern
Shan areas of Myanmar, north-western Laos and south-western Yunnan to the
core areas of the five Lan Na principalities (Chiang Mai, Lamphun, Lampang,
Phrae and Nan) in the late eighteenth and early nineteenth centuries. This re-
sulted in significant ‘foreign’ Tai communities – making up more than one-third
of the total populations – which maintained particular features of their original
written cultures.7

3 Lan Na scripts
The Tai of Lan Na have their own language which is distinct, and developed
three different scripts. These scripts were used separately for two different pur-
poses: one script was used predominantly for the writing of religious texts, an-
other script was used for the writing of secular, or non-religious, texts. A third
script, which appeared at a much later stage, mixed elements of the first two
scripts. The three types of Lan Na scripts are as follows:
(1) The Tham script (or Dhamma script of Lan Na, Thai: aksòn tham lan na
อักษรธรรมล้านนา): This script was developed from the old Mon script of the pre-
Tai kingdom of Hariphunchai (present-day Lamphun province). In Buddhism,
dhamma (Skt.: dharma) means the cosmic law and order as expressed in the
Teachings of the Buddha. Thus, as its name already implies, the Tham script
was first used for the writing of Buddhist scriptures; these texts were mostly
written in Pali but, in some cases (for instance the commentaries or the popular
Jataka tales), also in the Northern Thai vernacular. The earliest evidence of the
Tham script is from 1376. It is a bi-scriptural and bilingual inscription combin-
ing Sukhothai and Tham scripts, discovered in the early 1980s on a golden leaf
in a cetiya in Sukhothai, which is situated outside the Lan Na area (see Fig. 11)

||
that Dhamma-Lanna script was not used in pre-Burmese stone inscriptions’, see Buchmann
2018, 128.
6 Lorrillard 2022, 25.
7 Up to two-fifths of the Tai-speaking population of Lan Na by the end of the nineteenth centu-
ry were descendants of war captives who had been deported from the Tai Lü, Tai Khün and
Shan inhabited areas of the north and resettled in the southern parts of Lan Na during the late
eighteenth and early nineteenth century, the era of kep phak sai sa, kep kha sai müang (‘gather-
ing vegetables in baskets, gathering people in the polities’). For details, see Grabowsky 1999.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 115

However, an inscription on the pedestal of a Buddha image from Wat Chiang

Man in Chiang Mai city, dated 1465, is the earliest evidence of that script used
for the writing of texts in the Tai vernacular and Pali found in Lan Na territory.8
At a later period, notably since the late eighteenth century, this script was also
used for the writing of secular texts, such as folktales, chronicles, and astrologi-
cal and medical treatises. Eventually, it became the only script of the country
and, thus, it is colloquially called tua müang (ตัวเมือง), which means ‘script of
the country’.

Fig. 2: Tham (Dhamma) script; inscription 1.2.1.1: Pratu Tha Phae, 2010 BE / 1467 CE; photo-
graph © ALI, CMU.

(2) Fak Kham script (อักษรฝักขาม): The script’s name was chosen because the
letters’ curves resemble the shape of a tamarind pod.9 The Fak Kham script was
developed from the script that had been adopted from Sukhothai, the immedi-
ate southern neighbour of Lan Na, in the late thirteenth century. The Sukhothai
script is also the prototype of the modern Thai script, which is nowadays used
as the only and official script in Thailand. The Fak Kham script was normally
used for secular texts written in the vernacular Tai Yuan language, but we also
find Pali words or short Pali phrases written in this script. The vast majority of
Lan Na stone inscriptions prior to the seventeenth century are written in Fak

||
8 The inscription of Wat Chiang Man has been edited, translated and analysed by Prasert na
Nagara and Griswold 1992, 699–732.
9 The tamarind is a tropical fruit indigenous to tropical Africa and also naturalised in South
East Asia. The tamarind tree produces brown, pod-like fruits which contain a sweet, tangy pulp
used in cuisines around the world. The pod-like fruits resemble the shape of the largely square
characters of the script named thereafter.
116 | Apiradee Techasiriwan, Volker Grabowsky

Kham script.10 While the famous inscription of Wat Phra Yün, dated 1370, in
Lamphun was still written in Sukhothai script,11 the inscription of Wat Suwanna
Maha Wihan of 1411 from Phayao province was already written in Fak Kham
script. This inscription that records the donation of rice fields and 246 house-
holds as temple serfs to support this monastery under royal patronage is, there-
fore, considered to be the oldest extant Lan Na inscription written in the North-
ern Thai vernacular and a genuine Lan Na script.12

Fig. 3: Fak Kham script; inscription 1.5.1.1: Suwannamahawihan, 1974 BE / 1411 CE; photograph
© ALI, CMU.

(3) Thai Nithet script (อักษรไทยนิเทศ): This script was created by mixing the
Dhamma and Fak Kham scripts. Most characters resemble those of the Fak
Kham script but, unlike the latter, they have a round shape similar to the Tham
script characters. Thai Nithet was used mainly for secular poetry written on
palm-leaf manuscripts. However, we also find this script in royal edicts and
decrees inscribed on silver plates. The silver plate of Ban Pae in Chòm Thòng
district, Chiang Mai province, recording a royal decree of Queen Wisuttha Thewi
(r. 1564–1578), the last female ruler of the Mangrai dynasty who ruled over the
whole of Lan Na as a Burmese vassal, dated 5 June 1567, is the oldest evidence

||
10 For a glossary of these inscriptions, see Buchmann 2011; with a descriptive catalogue
(Buchmann 2012) and a grammar (Buchmann 2015) being published separately.
11 The Wat Phra Yün inscription has been edited, translated and analysed by Prasert na Naga-
ra and Griswold 1992, 605–623.
12 The endowment of land and temple serfs to monasteries is discussed in detail by Penth
2003 and Grabowsky 2004 and 2005. For an edition of the Wat Suwanna Maha Wihan inscrip-
tion (Lamphun 9), see Sujit Wongthes 1995, 49–57.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 117

of the Thai Nithet script.13 The latter also appears on a few other inscriptions.
This script was not widely spread. It appeared mainly in manuscripts of the first
half of the nineteenth century and not much later disappeared into oblivion.
There are only fewer than ten extant palm-leaf manuscripts and inscriptions
written in this hybrid script.

Fig. 4: Thai Nithet script; palm-leaf manuscript entitled Ongkan chaeng thian (‘Royal command
cursing the candle’); Microfilm no. 90.166.03/023-023, Social Research Institute, Chiang Mai
University; photograph © ALI, CMU.

4 Multi-scriptural and multilingual inscriptions in

Lan Na
Our ongoing survey of Lan Na epigraphic culture brought to light a small but by
no means insignificant number of inscriptions which reflect the mixing of
scripts and languages in the same inscribed object. For the sake of clarity, we
have divided these multi-scriptural and multilingual inscriptions into three
categories. The first category consists of inscriptions written in the Northern
Thai (Tai Yuan) vernacular – in some cases with occasional Pali phrases inter-
spersed – but with at least two of the three ‘indigenous’ Lan Na scripts used.
The second category includes inscriptions in which scripts and languages from
neighbouring writing cultures – such as that of Burma – were integrated into
Lan Na inscriptions. The third category pertains to inscriptions after the eight-
eenth century which reflect the influence of more distant writing cultures, such
as Thai (Siamese), Chinese and European.

||
13 The silver plate of Ban Pae has been translated and analysed by Kraisri Nimmanhaeminda
1967; for a German translation of both the inscription (dated 1567) and a palm-leaf manuscript
confirming the contents of the inscription by confirming the villagers’ privileged status as
temple serfs regarding events that happened in 1632, see Grabowsky 2004, 448–453.
118 | Apiradee Techasiriwan, Volker Grabowsky

4.1 The mixing of scripts within the Lan Na writing culture

The first example shows the mixing of scripts within the Lan Na written cul-
ture. It is an inscription on the pedestal of a Buddha image, dated c. 1490,
kept at the National Museum in Nan province close to the border to Lao
(1.7.3.2 Museum Nan c. 2030 BE / c. 1490 CE , ALI inventory number).14 The
inscription is written partly in Tham script and partly in Fak Kham script.
This example makes the original role of the Tham script to be used for the
writing of Pali language very clear and the use of Fak Kham script for writing
texts in the local Tai vernacular. The main content of the inscription is a mag-
ic spell called Gathā Paṭhamaṃ, which was popularly inscribed on Buddha
statues made of bronze in Lan Na during the fourteenth to fifteenth centu-
ries.15
The text starts with the Gathā inscribed in Tham script, but then within
the Gāthā, the scribe inserted the name of the sponsor, พระหัวเมืองซ้ายฟ่อน
(Phra Hua Müang Sai Fòn), i.e. the person who initiated the casting of this
Buddha statue, in Fak Kham script. The name of the sponsor inscribed in Fak
Kham script is framed in red in Fig. 6. The inscription then continues with
the Gāthā, inscribed in the Tham script again. This text is framed in blue in
the figure.

||
14 The Bronze Buddha statue has been published in Penth, Phanphen Khrüathai and Silao
Ketphrom 2001, 203–214.
15 Interview with Silao Ketphrom, a specialist of Lan Na inscriptions, on 11 May 2023 by
Apiradee Techasiriwan. Yijing or I-tsing was a Chinese Buddhist monk who spread Bud-
dhism from India to China during the Tang dynasty in the seventh century. He recorded
that the casting of Buddha statues entailed the enshrining of two types of elements inside
the Buddha statues, namely, relics of the Gotama Buddha and/or principles of the Buddha.
It is presumed that this idea to insert internal organs inside the Buddha statues and in-
scribe Buddhist verses on the pedestals of Buddha statues also spread to Lan Na until the
belief became popular. In Lan Na, the most popular Buddhist verse inscribed on the pedes-
tal of Buddha statues is Gathā Phra Buddha Sihing’s heart (คาถาหัวใจพระสิหิง ค์) or Gathā
Paṭhamaṃ, in which the meaning corresponds to the Four Noble Truths. See Surasawas
Suksawadi 2012, 21–22.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 119

Fig. 5: Mixing of Fak Kham script and Tham script on the pedestal of a Buddha image; inscrip-
tion 1.7.3.2: Museum Nan c. 2030 BE / c. 1490 CE; photograph © ALI, CMU.

Fig. 6: Detail of the pedestal of a Buddha image (Fig. 5); inscription 1.7.3.2: Museum Nan
c. 2030 BE / c. 1490 CE; photograph © ALI, CMU.
120 | Apiradee Techasiriwan, Volker Grabowsky

The second example is from a much later period; it is an inscription on a wood-

en board, dated 18 April 1848, which has been kept at Chang Khoeng monastery
(วัดช่างเคิง่ ) in Mae Caem district west of Chiang Mai city (1.2.2.1 Wat Chang
Khoeng 2391 BE / 1848 CE, ALI inventory number). This inscription runs over
seventeen lines. The main content is inscribed in Fak Kham script and Tai Yuan
language running over sixteen lines, while the last line is a colophon recording
the name of the scribe, Saen Siri (แสนสิร)ิ ; this last line is inscribed in Tham
script, in a smaller size than the rest of the inscribed text. The inscription high-
lights the simultaneous use of Fak Kham and Tham scripts in the same in-
scribed object. It seems evident that the scribe used two different scripts as a
tool to separate the main text visually from the colophon. This is frequently the
practice in Tai manuscript cultures where the main text could be written in one
of the major religious scripts (f.e. Tham or Khòm), while the colophon is in a
secular script (f.e. Thai or Lao).16 Moreover, it might be assumed that using Fak
Kham script for the main content of this inscription reflects the fact that this
script was commonly used for inscribing Lan Na stone inscriptions from the
early fifteenth to the nineteenth centuries, though after the seventeenth centu-
ry, Tham script had superseded Fak Kham script for the writing of manu-
scripts.17
Finally, the last example of the first category is an inscription on a pedestal
of a hermit statue made of bronze currently kept at Chiang Saen National Muse-
um, Chiang Rai province (1.4.3.2 Doi Tung 2148 BE / 1605 CE, ALI inventory num-
ber).18 The inscription is dated 1605, thus, it derives from the early period of
Burmese rule in that area which had become the main military and administra-
tive centre of Burma in the upper Mekong valley by the end of the century. The
inscription is inscribed mainly in Tham script and Tai Yuan language. The text
starts with the date and names of sponsors, with the local governor of Chiang
Saen and high-ranking monks presiding over the casting of the statue, then the
text continues with a chant in Pali, a phrase recited for worshiping Buddha
relics and referring to a chronicle of Phrathat Dòi Tung.

||
16 As Peter Skilling (2009, 3) aptly remarks, it was a ‘feature of Thai, Lanna Thai, and Isan
bilinguals is the use of two (or very rarely three) scripts in the same inscription or text, for
example Thai script for Thai language combined with Khom script for Pāli, or Fak Kham for
Lanna language and Tham for Pāli’.
17 Kannika Wimonkasem 1981, 18.
18 The bronze hermit has been published in Penth, Phanphen Khrüathai and Silao Ketphrom 1997,
189–210.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 121

Fig. 7: Wooden board inscription; inscription 1.2.2.1: Wat Chang Khoeng, Chiang Mai, 2391 BE /
1848 CE; photograph © ALI, CMU.

Fig. 8: Detail of inscription 1.2.2.1 (Fig. 7) using Tham script in the last line of the wooden board
inscription: Wat Chang Khoeng, Chiang Mai, 2391 BE / 1848 CE; photograph © ALI, CMU.
122 | Apiradee Techasiriwan, Volker Grabowsky

Fig. 9: A bronze hermit; inscription 1.4.3.2: Dòi Tung, Chiang Rai, 2148 BE / 1605 CE; photograph
© ALI, CMU.

Fig. 10: Tham and Thai Nithet scripts on the bronze hermit, detail of inscription 1.4.3.2 (Fig. 9):
Dòi Tung, Chiang Rai, 2148 BE / 1605 CE; photograph © ALI, CMU.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 123

It is worth mentioning that the scribe inserted a short text in Thai Nithet script
running from the second half of the second line until the first half of the third
line (highlighted in yellow in Fig. 10) for inscribing the Pali phrase, followed by
a short text in the Tai Yuan vernacular marking the beginning of the chronicle,
then continuing the text in Tham script until the end.
This inscription seems to break the original convention of using Thai Nithet
script for the Pali language instead of Tham script. We are tempted to speculate
that the scribe might have preferred to use Tham script for inscribing the main
text, but, at the same time, also wanted to keep the distinction between the
vernacular and Pali languages. To solve this problem, the scribe decided to use
the Thai Nithet script for the short Pali phrase. If we consider the year of produc-
tion of the inscription, 1605, it occurred thirty-eight years after the oldest evi-
dence of Thai Nithet script, which appears on a silver-plate inscription made in
1567 in Chiang Mai, as an order of the last queen of Lan Na’s Mangrai dynasty
(Queen Wisutthathewi, r. 1564–1578).19 Therefore, it is likely that Thai Nithet
script was still popular to use in the royal court that the ruler of Chiang Saen
presided over the casting of this statue.

4.2 Mixing of scripts and languages of neighbouring writing

cultures
In the past, Lan Na had maintained close ties with neighbouring kingdoms,
such as Ayutthaya and Lan Sang, in several ways: culturally, politically and
economically, ties which did not exclude military confrontations from time to
time. Therefore, there was a close cultural exchange between the various
polities of the Tai world and Tai scripts were arguably borrowed from one
another in the writing cultures of neighbouring Tai polities.
The first example of this category is very important because it is consid-
ered the oldest evidence of the appearance of the Tham Lan Na script. Inter-
estingly, it was found outside the Lan Na area, in Sukhothai, the older king-
dom that had had a close relation with Lan Na since the time of the
foundation of the Lan Na golden plate mentioned previously which was pro-
duced by a high-ranking monk from Sukhothai in 1376 (สท. 52, Inventory
number of the National Library of Thailand). The golden plate is inscribed
mainly in Sukhothai script and the Thai language. The inscription is about

||
19 For a translation of this inscription and of a palm-leaf manuscript related to it, see
Grabowsky 2004, 448–453.
124 | Apiradee Techasiriwan, Volker Grabowsky

the construction of a temple hall (vihāra), Buddha statues and relics being
placed in the monastery by the Venerable Cudhāmuṇi. The scribe then sepa-
rated the content by using a small circle and continued the inscribed text in
Tham Lan Na script and the Pali language until the end. However, until now,
we cannot find any other evidence to substantiate the hypothesis that the
Tham Lan Na script was used for writing Pali texts in the Sukhothai kingdom.
However, we can also not exclude the possibility that the high-ranking Su-
khothai monk was a native of Lan Na.

Fig. 11: Golden leaf of Maha Thera Cuthamuni (Cudāmuṇi) from 1376; photograph © Apiradee
Techasiriwan.

Another example is an inscription on a bronze Buddha image. This inscrip-

tion is very important because the image was produced by a Burmese ruler
during the first years in which Burma occupied Chiang Mai. The Buddha
image was cast in Lan Na art, in 1566, and has been sitting in Chai Phra Kiat
(ชัยพระเกียรติ) monastery in Chiang Mai town until today (1.2.3.2 Wat Chai
Phra Kiat 2108 BE / 1605 CE , ALI inventory number). The text on the pedestal
is divided into two parts. The first part is inscribed in Tham Lan Na script,
while the second part is written in Burmese script. The two parts contain the
contents of the same narrative. However, the text inscribed in Tham Lan Na
script and Tai Yuan language provides more details. The inscription text
mentions the background of casting the image, stating that the king of Burma
appointed a Burmese nobleman as the new ruler of Chiang Mai. The Burmese
ruler then asked the last queen of Chiang Mai and the local Tai people to gain
merit together by collecting damaged and scattered bronze Buddha images
with the intention of melting the bronze for casting a new Buddha statue,
and, finally, named the Buddha image Müang Rai Cao (เมืองรายเจ้า ), which
refers to the name of the founding king of Lan Na, Mangrai (มังราย).20 The use
of Lan Na art in casting the Buddha image and combining the two scripts

||
20 Penth 1976, 98–102.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 125

together in this inscription reflects the relative tolerance the Burmese

showed towards the cultural traditions of their Northern Thai vassal state, at
least during the early period of Burmese suzerainty which lasted until the
early seventeenth century.21

Fig. 12: A bronze Buddha image; inscription 1.2.3.2: Wat Chai Phra Kiat, Chiang Mai, 2108 BE /
1605 CE; photograph © ALI, CMU.

||
21 According to Sarassawadee Ongsakul (2005, 111), the period of tolerance lasted from 1558
until c. 1664.
126 | Apiradee Techasiriwan, Volker Grabowsky

Fig. 13: Detail showing Burmese script on the Buddha image; inscription 1.2.3.2: Wat Chai Phra
Kiat, Chiang Mai, 2108 BE / 1605 CE; photograph © ALI, CMU.

Fig. 14: Detail showing Tham script on the Buddha image; inscription 1.2.3.2: Wat Chai Phra
Kiat, Chiang Mai, 2108 BE / 1605 CE; photograph © ALI, CMU.

4.3 Mixing of scripts and languages of distant writing cultures

After the Tai Yuan elite, especially in the south-western parts of Lan Na, accom-
plished the expulsion of the Burmese from Lan Na through the strategic alliance
with and help of King Taksin of Thonburi and King Rama I of Bangkok from 1774
until 1804,22 Lan Na became associated with the Siamese kingdom in the south
and was eventually known as ‘Northern Siam’. The economy and society in Lan
Na prospered, particularly in Chiang Mai, the largest of the five Northern Thai

||
22 Sarassawadee Ongsakul 2018, 238–239.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 127

kingdoms. In the course of the nineteenth century, various new technologies

entered the area and a number of foreigners – such as Burmese, Chinese, West-
erners – came to settle in Chiang Mai, a city known for its multicultural envi-
ronment.
This characteristic is also reflected in some Lan Na inscriptions, for ex-
ample, an inscription on a bronze bell at Wat Dòi Suthep in the city of Chiang
Mai originally cast in May 1868 CE (1.2.3.1 Wat Phra That Dòi Suthep 2473 BE /
1930 CE , ALI inventory number). More than sixty years later, in 1930, this
large bronze bell was recast by a high-ranking monk and Cao Kaeo Nawarat
(1910–1939), the then ruler (เจ้าหลวง cao luang) of Chiang Mai, joined by a
wealthy family of the city, because the old bell was severely damaged. There-
fore, the sponsors cooperated to cast the bell and added more bronze, then
donated the bell to Suthep Cetiya on Dòi Suthep mountain in Chiang Mai. It
is worth mention that the inscription notes Princess Dara Ratsami (1873–
1933), daughter of King Inthawichayanon of Chiang Mai and consort of late
Siamese King Chulalongkorn (r. 1868–1910), as a royal co-sponsor of the
casting of the new bell.
The new bronze bell is inscribed in three scripts and languages. The first
part (Part I),23 written exclusively in the Tham script, presumably recalls the
text of an earlier inscription which existed before the recasting of the dam-
aged bell. It is followed by an additional text referring to the reasons for the
recasting. This text is written in two almost identical versions, one in Tham
script and the vernacular Tai Yuan language (Part II) and the other in modern
Thai script and Thai (Siamese) language (Part III). Both parts record the date
of casting the new bell, the names of the sponsors, their purpose in recasting
the bell, and their wishes for the benefit derived from the deed gaining merit.
As for the Chinese characters (Part IV), we assume that they were used to be
inscribed on the bell because a wealthy family of Chinese descent who had
been living in Chiang Mai were among the group of lay sponsors. The mem-
bers of that family might have wanted to add a short text recorded the date of
casting the bell in Chinese style, accompanied by words of blessing to all
Chiang Mai people.

||
23 This numbering refers to the edition presented in Appendix 3 of this paper.
128 | Apiradee Techasiriwan, Volker Grabowsky

Fig. 15: A large bronze bell of Dòi Suthep; inscription 1.2.3.1: Wat Phra That Dòi Suthep, 2473 BE
/ 1930 CE; photograph © ALI, CMU.

Fig. 16: Detail showing Tham script on the bronze bell of Dòi Suthep; inscription 1.2.3.1: Wat
Phra That Dòi Suthep, 2473 BE / 1930 CE; photograph © ALI, CMU.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 129

Fig. 17: Detail showing Thai script on the bronze bell of Dòi Suthep; inscription 1.2.3.1: Wat
Phra That Dòi Suthep, 2473 BE / 1930 CE; photograph © ALI, CMU.

Fig. 18: Detail showing Chinese characters on the bronze bell of Dòi Suthep; inscription 1.2.3.1:
Wat Phra That Dòi Suthep, 2473 BE / 1930 CE; photograph © ALI, CMU.
130 | Apiradee Techasiriwan, Volker Grabowsky

The last example is an inscription engraved on a cement board, 1964,

วัดหมื่นเงินกอง (Mün Ngoen Kòng monastery), Chiang Mai (1.2.4.1 Wat Mün
Ngoen Kòng 2507 BE / 1964 CE, ALI inventory number). The content of this
inscription is divided into four parts: the first part is inscribed in Tham Lan
Na script. It provides the name and the purpose of the principal sponsor, Nai
Tha Thipphotha (นายทา ทิพย์โพธา), who donated money to build an
Ubosatha hall and dedicated the benefits derived from the meritorious deed
to his deceased wife, Nang Lan (นางหลาน); the second part is inscribed in
modern Thai script, the contents of which is the same as that of the first part;
the third part is inscribed in Roman script, the meaning also corresponds
largely with the first and second parts; and in the last part, the scribe turned
to using Tham script again to add more details, including the names of the
sponsors who initiated the building of the wall surrounding the monastery.
Moreover, at the end of this inscription, the name of the abbot of this monas-
tery is inscribed in Thai script. We surmise that it was the monastery’s abbot
himself who acted as the scribe or at least ordered the engraving of this ce-
ment inscription.
However, a connection between the principal sponsor (Nai Tha Thipphotha)
and the use of the Roman script is still unclear. However, a modern Thai educa-
tion had gradually been introduced in the north since the late nineteenth centu-
ry, but the role of American missionaries of the Presbyterian Mission should be
mentioned as well. The teaching of English became popular in the region, open-
ing the opportunity to local people, laity and monks similarly to learn the Eng-
lish language. The abbot and scribe of our inscription might also have acquired
some knowledge of English. Moreover, scribe and sponsor might have wished
people from several Western nationalities to be able to understand and know
the merit they had accumulated, or it exhibits the proficiency of the scribe in a
Western foreign language and script.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 131

Fig. 19: Cement inscription; inscription 1.2.4.1: Wat Mün Ngoen Kòng 2507 BE / 1964 CE; photo-
graph © Apiradee Techasiriwan.
132 | Apiradee Techasiriwan, Volker Grabowsky

Fig. 20: Detail showing Thai, Roman and Tham scripts of the cement inscription; inscription
1.2.4.1: Wat Mün Ngoen Kòng 2507 BE / 1964 CE; photograph © Apiradee Techasiriwan.

5 Conclusion
These characteristics mentioned above are not only found in the Lan Na epi-
graphic culture, but also in other epigraphic and manuscript cultures in South
East Asia. Exemplary are the Tai groups in Yunnan, such as the Tai Nüa in the
autonomous Tai counties of Moeng Laem and Jinggu who normally use only
their locally developed secular Lik To Ngòk (Bamboo Shoot Script). However,
some Tai Nüa settlements closer to Chiang Rung in China and Chiang Tung in
eastern Burma came under the cultural influence of Lan Na Buddhism and
adopted the Tham script which they use alongside the Lik script. Moreover, in
the case of the manuscript culture of the Tai Lü in northern Laos, we also ob-
serve that Tai Lü scribes often copy the main contents in Tai Lü script following
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 133

older extant manuscripts, then use modern Lao script to write colophons.24 An-
other case are bilingual – Pali and Thai – manuscripts in central Thailand,
where scribes normally use Khòm script, which is a variant of Khmer script, for
writing Pali, and Thai script to write vernacular texts. Similarly, Burmese script
is often used for the writing of Pali language texts in the Shan manuscript cul-
ture because Shan consonants are insufficient for the writing of Pali, and Shan
orthography does not support the writing of Pali, which has a different way of
writing with the vernacular language.25
In conclusion, the use of several scripts and languages in the same in-
scribed object is testified in Lan Na epigraphy and can be conceptualised by
differentiating three categories: firstly, the mixing of scripts widely used within
the Lan Na writing culture; secondly, the mixing of scripts and languages of Lan
Na with those of neighbouring writing cultures; and thirdly, the mixing of the
scripts and languages of Lan Na with more ‘distant’ writing cultures. Multi-
scriptism and multilingualism in Lan Na are influenced by, at least, the follow-
ing four factors: the individual knowledge of scribes; the role of languages and
scripts in public spheres; the ethnic background of sponsors and commission-
ers; and the popularity and social value of scripts at the time when the inscrip-
tions were produced. The multilingual and multiscriptual inscriptions present-
ed in this preliminary study reflect the vicissitudes of Lan Na history and the
transformation of the region from an independent Buddhist kingdom to a vassal
state of Burma, and later Siam, and its final integration into the modern Thai
nation-state.

Abbreviations
ALI = Archive of Lan Na Inscriptions (CMU)
BE = Buddhist Era
CE = Common Era
CMU = Chiang Mai University
CS = Culasakarāja, lunar calendar used in Thailand

||
24 For colophons in Tai Lü manuscripts in northern Laos, see the catalogue of manuscripts found
at Vat Pak Chaek, a Tai Lü village monastery in Luang Prabang province, situated roughly 50 km
to the north of Luang Prabang town. See Khamvone Boulyaphonh and Grabowsky 2022.
25 For the round-shaped secular Shan script, also called Lik Tou Moan (literally, ‘circular
letters’), see Sai Kam Mong 2004, 119–146.
134 | Apiradee Techasiriwan, Volker Grabowsky

References
Buchmann, Marek (2011), Northern Thai Stone Inscriptions (14th–17th Centuries): Glossary,
Wiesbaden: Harrassowitz Verlag.
Buchmann, Marek (2012), Northern Thai Stone Inscriptions (14th–17th Centuries): Catalogue,
Wiesbaden: Harrassowitz Verlag.
Buchmann, Marek (2015), Northern Thai Stone Inscriptions (14th–17th Centuries): Grammar,
Wiesbaden: Harrassowitz Verlag.
Buchmann, Marek (2018), ‘An Introduction to the Northern Thai Stone Inscription Corpus’, in
Daniel Perret (ed.), Writing for Eternity: A Survey of Epigraphy in Southeast Asia, Paris:
École française d’Extrême-Orient, 123–134.
Grabowsky, Volker (1999) ‘Forced Resettlement Campaigns in Northern Thailand during the
Early Bangkok Period’, Journal of the Siam Society, 87/1–2: 45–86.
Grabowsky, Volker (2004), Bevölkerung und Staat in Lan Na: Ein Beitrag zur Bevölkerungsge-
schichte Südostasiens, Wiesbaden: Harrassowitz.
Grabowsky, Volker (2005), ‘Population and State in Lan Na prior to the Mid-sixteenth Century’,
Journal of the Siam Society, 93: 1–68.
Grabowsky, Volker (ed.) (2022), Manuscript Cultures and Epigraphy in the Tai World, Chiang
Mai: Silkworm Books.
Khamvone Boulyaphonh and Volker Grabowsky (2022), Tai Lü Manuscripts in the Lao Diaspora
(Hamburger Südostasienstudien, 19), New York / Luang Prabang: Anantha Publishing.
Kannika Wimonkasem (กรรณิการ์ วิมลเกษม) (1981), Aksòn fak kham thi phop nai phak nüa
อักษรฝั กขามที่พบในภาคเหนื อ (‘Fak Kham Scripts Found in Northern Thai Inscriptions’), MA
thesis, Silpakorn University, Bangkok.
Kongkaeo Wirapracak (ก่องแก้ว วีระประจักษ์) (2014), Carük lan thòng somdet phra maha thera
cuthamuni จารึกลานทองสมเด็จพระมหาเถรจุทามุณิ (‘Inscription on the golden leaf of Maha
Thera Cuthamuni (Cudāmuṇi)’), in Nana saratthakhadi cak carük lae ekkasan boran
นานาสารัตถคดีจ ากจารึกและเอกสารโบราณ (‘Substance of inscriptions and ancient docu-
ments’), Bangkok: National Library, Department of Fine Arts, Ministry of Culture, 166–170.
Kraisri Nimmanhaeminda (1967), ‘The Lawa Guardian Spirits of Chiangmai’, Journal of the Siam
Society, 60/2: 185–225.
Lorrillard, Michel (2022), ‘The Inscriptions of the Lan Na and Lan Xang Kingdoms: A New Ap-
proach to Cross-Border History’, in Grabowsky 2022, 21–43.
Penth, Hans (ฮันส์ เพนธ์) (1976), Kham carük thi than phraphuttharup nai nakhòn chiang mai
คําจารึกที่ฐานพระพุทธรูปในนครเชียงใหม่ (‘Inscriptions on Buddha Figures from Chiang
Mai’), Bangkok: Office of Prime Minister.
Penth, Hans (2003), ‘On Rice and Rice Fields in Old Lān Nā – Texts. Translations. Interpreta-
tions’, Journal of the Siam Society, 91: 90–188.
Penth, Hans, Phanphen Khrüathai and Silao Ketphrom (1997), ฮันส์, เพนธ์; พรรณเพ็ญ เครือไทย;
ศรีเลา เกษพรหม Prachum carük lan na (‘Corpus of Lān Nā Inscriptions’), vol. 1: Carük nai
phiphitthaphan chiang saen เล่ม ๑: จารึกในพิพิธภั ณฑ์ฯ เชียงแสน (‘Inscriptions in the
Chiang Sän Museum’), Chiang Mai: Social Research Institute, Chiang Mai University.
Penth, Hans, Phanphen Khrüathai and Silao Ketphrom (2001), ฮั นส์, เพนธ์; พรรณเพ็ญ เครือไทย;
ศรีเลา เกษพรหม; ศราวุธ ศรีทา Prachum carük lan na (‘Corpus of Lān Nā Inscriptions’), vol. 5:
Carük nai phiphitthaphan nan lae carük müang nan thi na son cai เล่ม ๕: จารึกในพิพิธภัณฑ์ฯ
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 135

น่ าน และจารึกเมื องน่านที่น่ าสนใจ (‘Inscriptions in the Nān Museum and other Nān Inscrip-
tions’), Chiang Mai: Social Research Institute, Chiang Mai University.
Penth, Hans, Silao Ketphrom and Sarawut Sitha (2004), ฮันส์, เพนธ์; ศรีเลา เกษพรหม; ศราวุ ธ
ศรีทา Prachum carük lan na (‘Corpus of Lān Nā Inscriptions’), vol. 8: Carük nai cangwat
chiang mai เล่ม ๘: จารึกในจังหวัดเชียงใหม่ (‘Inscriptions in the Chiang Mai Province’),
Chiang Mai: Social Research Institute, Chiang Mai University.
Penth, Hans, Silao Ketphrom and Apiradee Techasiriwan (2007), ฮันส์, เพนธ์; ศรีเลา เกษพรหม;
อภิรดี เตชะศิ รวิ รรณ Prachum carük lan na (‘Corpus of Lān Nā Inscriptions’), vol. 12: Carük
nai cangwat chiang mai phak 4 เล่ม ๑๒: จารึกในจังหวัดเชียงใหม่ ภาค ๔ (‘Inscriptions in the
Chiang Mai Province, Part IV’), Chiang Mai: Social Research Institute, Chiang Mai University.
Prasert na Nagara and Alexander Brown Griswold (1992), Epigraphic and Historical Studies,
Bangkok: The Historical Society under the Royal Patronage of H.R.H. Princess Maha Cha-
kri Sirindhorn.
Sai Kam Mong (2004), The History and Development of the Shan Scripts, Chiang Mai: Silkworm
Books.
Skilling, Peter (2009), ‘Manuscripts and inscription, languages and letters’, in Claudio Cicuzza
(ed.), Buddhism and Buddhist Literature of South-East Asia: Selected Papers, Bangkok /
Lumbini: Fragile Palm Leaves Foundation, Lumbini International Research Institute.
Sarassawadee Ongsakul (2005), History of Lan Na, tr. Chitraporn Tanratanakul, Chiang Mai:
Silkworm Books.
Sarassawadee Ongsakul (สรัสวดี อ๋องสกุล) (2018) Prawatsat lan na chabap sòmbun
ประวัติศาสตร์ล้านนา ฉบับสมบูรณ์ (‘Lan Na History, complete version’), 2nd edn [1st edn:
1996], Bangkok: Amarin Printing.
Sujit Wongthes (สุจิตต์ วงษ์เทศ) (ed.) (1995), Prachum carük müang phayao
ประชุมจารึกเมืองพะเยา (‘Inscriptional History of Phayao’), Bangkok: Matichon.
Surasawas Suksawadi, M.L. (2012), ‘Buddhist Sects in Lān Nā between the Reigns of Phayā
Tilōk to Phayā Kaeo (1441–1525): Studies from Dated Bronze Buddha Images in Chiang
Mai’, Silpakorn University Journal, 32/2: 8–25.
136 | Apiradee Techasiriwan, Volker Grabowsky

Appendix 1: Wooden board inscription of Wat

Chang Khoeng, Chiang Mai
1.2.2.1 Wat Chang Khoeng 2391 BE / 1848 CE (ALI inventory
number)
Source: Penth, Silao Ketphrom and Sarawut Sitha 2004, 137–148.

Transcription into modern Thai:

ศุภมัสตุ ศักราชได้ 1209 ตัว ปีมะเมียฉนํ า กัมโพชคาม ตามขอมพิสย ั เข้ามาในคิมห

จิตรมาส ไทยว่าปีเมืองเม็ด เดือน 7 ออก 15 คํ่า วันอังคาร ไทย (ว่าวัน) กัดเป้า ฤกษ์ 12
ตัว ชื่ออุตรผลคุณ บุณมี
พระเป็นเจ้าตนชื่อแก้วเมืองมา (พร้อมทัง้ ) อรรคชายาชื่อศรีบญ ุ ยวง มีจต ิ ประสาทเป็น
ปฐมมหามูลศรัทธาอันยิง่ จิ่งมีพระบัณฑูรใส่หัวนายนายก จุ๊คน หนภายในมี
- ครูบาสีลมณี เป็นเค้า และ
- ตุ๊หลวงนั นทา
- ตุ๊หลวงโพธา
- ตุ๊หลวงเทพ
หนภายนอก มีแสนสิรเิ ป็นเค้า (พร้อมทัง้ ) ฝาเมือง ตงเมือง จุ๊คน เยนศมเยน
มีจิตสมัครสัมโมทมาน (ทีจ ่ ะ) สร้างแป๋งยังสุวรรณรัชตฉัตร ขึ้นต่างไว้ยัง
จอมเกศพระชินมาร (พร้อม) กับทัง้ (สร้าง) พระหาร 2 หลัง เป็นที่สถิตสําราญ (ของพระ)
พุทธรูปเจ้า และ (ของรูป) พระมหากัจจายนเถรเจ้า หื้อก้านกุ่ง รุง่ เรือง (แก่) ศาสนา ดีหลี
เต๊อะ
แสนสิรต ิ ้องแต้ม แปลจารึกนี้ ไว้ แล

Translation:

Be you blessed. In 1209 CS, the year // of the horse, according to the Cambo-
dian tradition, entering the hot season, // a müang met year as the Tai say, //
on the fifteenth waxing day of the month of Caitra, the seventh lunar month,
on a Tuesday, a kat pao day as the Tai say, at the auspicious moment of 12, //
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 137

which was a full-moon day,26 // His Royal Highness Kaeo Müang Ma, along
with his queen named Si Bun Yuang, // had the intention to be the royal
leading supporter // and, thus, gave the royal order // to all chiefs; among
the heads of the Sangha these were Khruba Sinlamani, who takes the lead, //
Tu Luang Nantha, Tu Luang Photha, and Tu Luang Thep; // among the heads
of the laypersons, these are Saen Siri who takes the lead, as well as Fa Müang
// and Tong Müang. They all are delighted unanimously // to sponsor the
making of this golden and silver parasol // which is placed above the head of
the Phra Chinaman (Braḥ Jinamāra) Buddha statue, // along with the [con-
struction of] two temple halls (vihāra) which are the joyful place of Buddha
images and images of Phra Maha // Kaccāyana Thera Cao so that the Teach-
ings of the Buddha will prosper surely. // Saen Siri Tòng Taem (i.e. the
royal scribe) translated this inscription.27

Appendix 2: Golden leaf of Maha Thera Cuthamuni

(Cudāmuṇi)
สท. 52 (Inventory number of National Library of Thailand)
Source: Kongkaeo Wirapracak 2014.

Transcription into modern Thai:

1) ๐ ศักราชได้ ๗๓๘ โรงนั กษัตร เดินหกแรมคํ่าหนึ่ งสุกราพาร ๐ สมเด็จพระมหาเถรจุ

ทามุณิปลูกพิหาร
2) นี้ ได้เดินสิบห้าวันบริบูรณ์ แล้วเมือหน้ า เดินสิบ แรมสี่ค่าํ พุธพาร กรรดิการิกสจึง
ปราดิสถา
3) พระธาตุ และพระพุทธประติมาอันท่านกระทําด้วยเงิน ทอง เหียก งา ทัง้ หลายจึง
กระทํา
4) พระอถารสเหนื ออัน ้ ๐ อิมินาปุญฺญกมฺเมน พุทฺโธ โหมิ อนาคเต สงฺสาราโมจนตฺถาย
สพฺเพ สตฺเต อเสสโตฯะะะ

||
26 This date corresponds to Tuesday, 18 April 1848 (1210 Caitra 15 CS). This day was indeed a
kat pao day. However, on that day, the solar year had already moved to 1210 CS, a poek san
year.
27 This last sentence is inscribed in the Tham script and, thus, highlighted in bold letters.
138 | Apiradee Techasiriwan, Volker Grabowsky

Translation:

In Culasakkarat 738, the year of the dragon, on the first waning day of the
sixth [lunar] month, a Friday,28 Somdet Phra Maha Thera Cuthamuni com-
pletely built this temple hall (vihāra) for a month and fifteen days. // Passing
to the tenth [lunar] month, on the fourth waning day, on a Wednesday, the
auspicious moment of Karttika29 (…), therefore it establishes // the stupa and
the Buddha statue which is made of silver, gold, zinc, ivory. Thus it cele-
brates // Phra Atharot (Braḥ Athārasa). Iminā puññakammena buddho homi
anāgate saṃsārāmocanatthāya sabbe satte asesato (‘By virtue of the merit
(obtained by) this deed I ask for becoming a Buddha in the future in order to
ensure that all sentient beings escape from the cycle of rebirth without any
exception’).30

Appendix 3: A large bronze bell of Dòi Suthep

1.2.3.1 Wat Phra That Dòi Suthep 2473 BE / 1930 CE (ALI
inventory number)
Source: Penth, Silao Ketphrom and Apiradee Techasiriwan 2007, 287–306.

Part I: Tham script

Transcription into modern Thai script:

สุทินฺนํ วต เม มหาฆณฺ โฑ ทานํ มงฺคลวุฑฺฒิ สิรส

ิ ุภมสฺสตุ
จุลศักราชได้ 1230 ตัว มะโรงฉนํ า กัมโพชพิสย ั ในคิมกันตฤดู มาสวิสาขา สุกลปักษ์
ปุณมี พุธวารไถง ไทยภาษาว่าปีเปิกสี เดือน 8 เพ็ญ เม็ง (ว่า)วันพุธ ไทย(ว่าวัน) เต่าสี
ดิถี 15 ตัว นาทีดิถี 1 ตัว พระจันทร์จรณยุตเิ สด็จเข้าเทียวเทียมนั กขัตฤกษ์ตวั ถ้วน 16
ชื่อวิสาขเทวดา นาทีฤกษ์ 4 ตัว ปรากฏในตุลวาโยราศี อดีตวรพุทธศาสนาคลาล่วงแล้วได้
2412 อนาคตวรพุทธศาสนา(ที่)จักมาภายหน้าบ่น้อยยัง 2588 พระวรรษาเต็มบ่มเี ศษ

||
28 This date corresponds to Tuesday, 24 February 1377 (738 Phalguna 16 CS).
29 This date corresponds to Sunday, 6 July 1376 (738 Ashadha 19 CS).
30 The Pali phrase was translated by Achan Sa-aem Kaeokhlai.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 139

เหตุนั้นหมายมี พระเจ้าชีวต ิ ตนเป็นเจ้าแก่รต ั นติงสา อภินวบุรศ

ี รีพิงไชย เจ้ากาวิโลรส
สุรยิ วงษ์ เจ้าพระนครเชียงใหม่ เป็นเค้า ซึ่งเป็นกับด้วยอรรคมเหสี และราชบุตรา ราชบุตรี
หมายมี เจ้าทิพ(เกสร) และเจ้าอุบลวัณณา และราชบุตรา ราชบุตรี จุ๊ตน จุ๊คน
ก็ได้สร้างยังมหาฆัณโฑเดงหลวงลูกนี้ ถวายไว้ให้เป็นทาน (เพื่อ)สว่ายตีบูชา กับพระมหา
ชินธาตุเจ้าสุเทพทีนี้ ตราบ 5000 พระวรรษา นํ้ าหนั กทองมี (1) ล้าน 5 แสนปลาย 2 หมื่น
ขอจงเป็นปัจจัย(แก่)โลกิยะ โลกุตระ มรรคะ มีผลนิ พพานเป็ นยอด แด่
นิ พฺพานํ ปรมํ สุขํ

Translation:

Sudinnaṃ vatame mahāghaṇḍo dānaṃ maṅgalavuḍḍhi sirisubhamastu (‘The

donation of the great bell is given by me. Let there be glorious and auspi-
cious!’)
In Culasakarāja 1230, in the year of the dragon in the hot season, on the full
moon-day of the month of Visākhā, a Wednesday, [or] as the Tai say, in a
poek si year, on the eighth [lunar] month, a Wednesday as the Mon say, a tao
si day as the Tai say, at the tithī (auspicious time) of 15 and nādī tithī of 1,31
when the moon entered the sixteenth nakkhataṛkṣa, the zodiacal sign of Li-
bra, with 2412 years of the Buddhist Era already elapsed and 2588 years of the
Buddhist Era still remaining: the Lord of Life, Protector of the Three Gems,
Supreme Ruler over the Great City of the Ping River, King Kawilorot Suri-
yawong of Chiang Mai, as principal initiator, along with his queen, royal
sons and royal daughters, namely Princess Thip[kesòn] and Princess Ubon-
wanna, and all their sons and daughters, sponsored the making of the Mahā
Ghaṇḍo Deng Luang bell and donated it to worship the Phra Mahā Jinadhātu
Cao Suthep (statue) so that it will last [until the end of] 5000 years. The
weight of its bronze is over one million and five hundred and twenty thou-
sand [units]. May this be a support for the mundane (lokiya) and super-
mundane (lokutta), a path to reach Nibbāna.
Nibbānaṃ paramaṃ sukhaṃ (‘Nibbāna is the highest [stage of] happiness’).

||
31 This date corresponds to Wednesday 6 May 1868.
140 | Apiradee Techasiriwan, Volker Grabowsky

Part II: Tham script; Pali and Tai Yuan languages

Transliteration into modern Thai script:

มงฺคลวุฑฒ ฺ ิ สิรสิ ุภมสฺสตุ

จุลศักราชได้ 1292 ตัว มะเมียฉนํา กัมโพชพิสย ั เข้ามาในคิมหันตฤดู อุตุนฤกษ์สตุ สุกละ
ปักษ์สต ั ตมี ฤดู พุธวารไถง ไทยภาษาว่าปีกดสะง้า เดือน 10 ขึ้น 7 คํ่า เม็ง (ว่า) วัน 4
ไทย (ว่าวัน) ก่าเป้า ดิถี 7 ตัว นาทีดิถี 12 ตัว นาทีฤกษ์ 18 ตัว พระจันทร์จรณยุตโยส
โสดเสด็จเข้าเทียวเทียมเสมียงเมียงม่อ อว่ายหน้าราชรถล่อ พระนั กขัตฤกษ์ตวั ถ้วน 12
ชื่ออุตตราผลคุณเทวดา มาปรากฏในการกัฎอาโปราศี อดีตวรพุทธศาสนา อันคลาล่วง
แล้วได้ 2473 พระวรรษา ธิกาปลาย 1 เดือนกับ 18 วัน อนาคตวรชินศาสนาอันจักมา
ภายหน้ า ยังมีมากบ่น้อยยังอยู่ 2526 พระวรรษา ธิกาปลาย 10 เดือนกับ 8 วันบ่เศษ
เหตุเอามาบวกสมกันก็หากเต็ม 5,000 พระวรรษาบ่เศษ
เหตุนั้นหมายมีมหาสมณศรัทธา และมูลศรัทธาทัง้ 2 คณะคณา ทัง้ ภายในและภาย
นอกภายใน หมายมีเจ้าคุณอภัยสารทะ เจ้าคณะจังหวัดเชียงใหม่ และเจ้าคณะแขวงทัง้
หลายทุกๆ องค์ ก็พร้อมด้วยศิษย์ สานุศษ ิ ย์ทง ๊ งค์ (และ) ศรัทธาคุรุอุปัฏฐาก
ั้ หลาย จุ๊ตน จุอ
จุ๊ผู้ จุ๊คน ๆ หนภายนอกหมายมีพ่อแก้วนวรัฐ เจ้าผู้ครองนครเชียงใหม ก็พร้อมด้วยอรรค
มเหสี และบุตรา บุตรีทง ั้ หลาย จุ๊ผู้ จุ๊คนๆ ถัดเรียงนั ้นมาหมายมี พระราชชายาเจ้า
ดารารัศมี ก็พร้อมด้วยเด็กสาว ชาวแม่นาง เถ้าแก่เรือนหลวงทัง้ หลาย จุ๊ผู้ จุ๊คน ๆ
ถัดเรียงนั ้นมาหมายมีหลวงอนุสารสุนทร ก็พร้อมด้วยภริยา (ชื่อ) นางคําเที่ยง และบุตรา
บุตรี ทัง้ หลาย จุ๊ผู้ จุ๊คน ๆ
ได้เล็งหันเด็งหลวงอันเก่านั ้น ก็มาชํารุดทรุดโทรมแตกแหงไปเสีย เสียงต่อยก็บด ่ ี ตีก็บ่ดัง
ไปดัง่ นี้ นํ้ าหนั กเด็งหลวงเก่านั ้นมีล้าน 5 แสน 2 หมื่นเท่านี้ (จึง) พร้อมด้วยสมณศรัทธา
และมูลศรัทธา ทัง้ 2 ฝ่าย คณะคณา ก็พร้อมเพรียงกันมาปฏิสงั ขรณ์ ซ่อมแซมขึ้นใหม่
เตื่อมทองใส่แถม 5 แสน รวมนํ้ าหนั กทอง (ของ) เด็งเก่า (และ) ใหม่ มีน้ํ าหนั ก 2 ล้าน 2
หมื่นทอง เท่านี้ แล เด็งหลวงแก่นนี้ ถวายเป็นทานไว้กับ พระธาตุเจ้าสุเทพ ตราบอายุเด็ง
นี้ เทอญ
ขอให้ศรัทธาผูข ้ า้ ทัง้ หลาย ทัง้ 2 ฝ่ายคณะคณา ทัง้ ภายในและภายนอก ขอหื้อได้ข้าม
พ้นจากโอฆวัฏ สงสาร หื้อได้ถงึ ฝั่ งกลํ้าหน้ าทีเ่ มืองฟ้านิ พพานเจ้านั ้น แท้ดีหลี
เชยฺยตุภวํ เชยฺยตุภวํ เชยฺยมงฺคลํ

Translation:

Maṅgala vuḍḍhi siri subhamastu. In Culasakarāja 1292, the year of the horse,
(…) the Tai say a kot sanga year, on the seventh waxing day of the tenth [lu-
nar] month, the fourth day of the week as the Mon say, a ka pao [day] as the
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 141

Tai say,32 at the tithī (auspicious time) of 7, and nādī tithī of 1, the nādīṛkṣa of
18, the moon entering the twelfth nakkhataṛkṣa, the zodiacal sign named
Uttarā phala guṇa devatā, with 2473 years, one month and twenty-one days
of the Buddhist Era already elapsed and 2526 years, ten months and eight
days of the Buddhist Era still remaining, adding up to the full 5,000 years of
the Buddhist Era: the two leading monastic supporters and all principal sup-
porters (mūlasaddhā) of the two groups – the ‘inner group’ led by Cao Khun
Aphai Santha, the principal abbot of Chiang Mai province, along with the
principal abbots of the ecclesiastical districts and all their disciples and at-
tendants, the ‘outer group’ led by Phò Cao Kaeo Nawarat, the ruler of Chiang
Mai, together with his queen, and all his children, joined by the royal consort
Princess Dara Ratsami33 and her daughter, (…) all people in the palace, fol-
lowed by Luang Anusan Sunthòn and his consort named Nang Kham Thiang,
and all their children, saw that the old bell was dilapidated, its sound was
bad and not loud.
The original weight was one million five hundred and twenty thousand
[units]. Therefore, the leading monastic supporters and principal supporters
of the two groups unanimously renovated the bell by adding five hundred
thousand [units of] copper. Thus the total weight of the bell – old and new –
was two million and twenty thousand [units of] copper. This great bell was
donated to the stupa (phra that) of Dòi Suthep until [the end of] this bell’s
life.
May we all, religious faithful of the two groups, both the inner and the outer
group, escape from the cycle of rebirth (oghavaṭasaṃsāra) and reach the bor-
ders of the celestial city of Nibbāna with certainty.
Jeyyatubhavaṃ jeyyatubhavaṃ jeyyamaṅgalaṃ (‘Let there be victorious and
auspicious!’).

||
32 It is the tenth lunar month of the northern Thai calendar corresponding to the month Asha-
da. This date corresponds to Wednesday, 2 July 1930 which was indeed a kap pao day.
33 The first part of the inscription states that King Kawilorot of Chiang Mai (r. 1854–1870),
along with his queen, presided over the original casting of the bell in 1869.
142 | Apiradee Techasiriwan, Volker Grabowsky

Part III: Thai script; Thai language

Transliteration into modern Thai script:

มังคลวุฑฒิสิรส ิ ุภมัสตุ จุลศักราชได้ 1292 ตัว มะเมียฉนํ ากัมโพชพิสัย เข้ามาในคิม

หันตฤดู สุกลปักษ์ สัตตมีฤดู พุทธวารไถง ไทยภาสาว่าปีกดสะง้า เดือน 8 ขึ้น 7 คํ่า เมง
(ว่า)วัน 4 ไทย (ว่าวัน) ก่าเป้า ดีถี 7 ตัว นาทีดิถี 12 ตัว นาทีฤกษ์ 18 ตัว พระจันทร์
จรณะยุตโิ ยดโสดเสด็จเข้าเทียวเทียมเสมียงเมียงม่อ อว่ายหน้าราชรถล่อพระนั กขัตต
ฤกษ์ตัวถ้วน 12 ชื่ออุตตราผละคุณเทวดา ปรากฏในการกัฎอาโปราศี อตีตวรพุทธศาสนา
อันกลาล่วงแล้วได้ 2473 พระวรรษา ธิกาปลาย 1 เดือนกับ 21 วันอนาคตวรชินศาสนา
อันจักมาภายหน้ายังมีมากคามบ่น้อยยังอยู่ 2526 พระวรรษา ธิกาปาย 10 เดือนกับ 8
วันบ่เศษ เหตุเอามาบวกสมกันก็หากเต็มถ้วน 5,000 พระวรรษาบ่เศษ
เหตุนั้นหมายมีมหาสมณศรัทธา และมูลศรัทธาทัง้ 2 คณะคณะทัง้ ภายในและภายนอก
ภายในมีเจ้าคุณอภัยสารทะ เจ้าคณะจังหวัดเชียงใหม่ และเจ้าคณะแขวงทุก ๆ องค์
พร้อมด้วยศิษย์สานุศษ ิ ย์ทุกองค์ๆ ศรัทธาอุปัฏฐานทุกคนๆ
หนภายนอกมีพ่อเจ้าแก้วนวรัฐ เจ้าผู้ครองนครเชียงใหม่ พร้อมด้วยอรรคมเหสี บุตรา
บุตรี ทุกคน ๆ ถัดเรียงนั ้น มีพระราชชายาเจ้าดารารัศมี ก็พร้อมด้วยเด็กสาว ชาวแม่
นาง เถ้าแก่เรือนหลวงทุกคน ๆ ถัดนั ้นหลวงอนุสารสุนทร พร้อมภรรยา (ชื่อนาง) คําเที่ยง
บุตรา บุตรี ทุกๆ คน ได้เล็งหันยังเด็งหลวงอันเก่านั ้น ชํารุดทรุดโทรมแตกเสีย เสียงต่อยก็
่ ัง นํ้ าหนั กเด็งเก่า มีลา้ น 5 แสน 2 หมื่น เพราะฉะนั ้นสมณศรัทธา และ
บ่ดี ตีก็บด
มูลศรัทธาทัง้ 2 คณะคณา พร้อมเพรียงกันปฏิสังขรณ์ ซ่อมแซมขึ้นใหม่ เตื่อมทองใส่แถม
5 แสน รวม (นํ้าหนั กของเด็ง) ทัง้ เก่า (และ) ใหม่ มีน้ําหนั ก 2 ล้าน 2 หมื่นทอง เท่านี้ แล
เด็งหลวงแก่นนี้ ถวายเป็นทานไว้กับพระธาตุดอยสุเทพ ตราบอายุเด็งนี้เทอญ
ขอหื้อศรัทธาผูข ้ า้ ทัง้ หลาย ได้ถึง นิ พพ
ฺ านปจฺจโย โหตุ โนนิ จฺจํ

Translation:

Maṅgala vuḍḍhi siri subhamastu (‘Let there be glorious and auspicious’). In

Culasakarāja 1292, the year of the horse, (…) the Tai say a kot sanga year, on
the seventh waxing day of the eighth [lunar] month, the fourth day of the
week as the Mon say, a ka pao [day] as the Tai say,34 at the tithī (auspicious
time) of 7, and nādī tithī of 1, the nādīṛkṣa of 18, the moon entering the
twelfth nakkhataṛkṣa, the zodiacal sign named Uttarā phala guṇa devatā,
with 2473 years, one month and twenty-one days of the Buddhist Era already
elapsed and 2526 years, ten months and eight days of the Buddhist Era still

||
34 It is the eighth lunar month of the central Thai calendar corresponding to the month Asha-
da. This date corresponds to Wednesday, 2 July 1930 which was indeed a kap pao day.
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 143

remaining, adding up to the full 5,000 years of the Buddhist Era: the two
leading monastic supporters and all principal supporters (mūlasaddhā) of the
two groups – the ‘inner group’ led by Cao Khun Aphai Saratha, the principal
abbot of Chiang Mai province, along with the principal abbots of the ecclesi-
astical districts and all their disciples and attendants, the ‘outer group’ led
by Phò Cao Kaeo Nawarat, the ruler of Chiang Mai, together with his queen,
and all his children, joined by the royal consort Princess Dara Ratsami and
her daughter, (…) all people in the palace, followed by Luang Anusan Sun-
thòn and his consort named Nang Kham Thiang, and all their children, saw
that the old bell was dilapidated, and its sound bad and not loud.
The original weight was one million five hundred and twenty thousand
[units]. Therefore, the leading monastic supporters and principal supporters
of the two groups unanimously renovated the bell by adding five hundred
thousand [units of] copper. Thus the total weight of the bell – old and new –
was two million and twenty thousand [units of] copper. This great bell was
donated to the stupa (phra that) of Dòi Suthep until [the end of] this bell’s
life.
May this be for all of us, religious faithful, a condition to reach Nibbāna with
certainty.

Part IV: Chinese script; Chinese language

Transliteration into modern Thai script:

1) จง หัว หมิน กัว๋ อี้ สือ

2) จิ่ว เหนี ยน ซุ่ย ชื่อ เกิง อู่
3) ลิ่ว เยว่ ชู ชิ ยึ เจี้ยน
4) ชิง ม่าย เหอ เซิ่ง ผิง อาน

Translation:

On the sixth month, the seventh day, in the nineteenth year (after the change of
government in China, i.e. 1930 CE), this bell has been cast at noon time.
May the people of Chiang Mai live all in happiness.
144 | Apiradee Techasiriwan, Volker Grabowsky

Appendix 4: Inscription of Wat Mün Ngoen Kòng

1.2.4.1 Wat Mün Ngoen Kòng 2507 BE / 1964 CE (ALI inventory
number)
Source: Unpublished inscription kept at the Archive of Lan Na Inscriptions,
Social Research Institute, Chiang Mai University.

Part I: In Tham script

Transcription into modern Thai script:

1) ปฺลสสอศกจุรสกราก 1326 พุทฺธสกราช 2507

2) [วง]นี้ หมายมีพ่อธาทิบโพธาบ้านอยู่หลังวัดบวกหงตําบรภละสิงค์ อําเภอเมือง
3) จังหวัดเชียงใหม่ เพื่อปียี ภ.ศ. 2507 นี้ อายุ 63 ปี มีสัทธาเชื่อในสาสนาพระพุทธ
4) เจ้า ได้สละเงินเปนจํานวน 2750034 บาท สองหมื่นห้าพันเจ็ดร้อยบาท ได้ส้างยัง
อุโบสถ
5) หลังนี้ ภ้อมกับด้วยลูกอันมีนางบัว[ซุ]ม นางสีมอย นางบัวแก้ว เพื่อไว้ค้าํ ชูพระ
สาสนาไว้กับวัด
6) หมื่นเงินกอง ภ้อมกันนี้ ค็ได้ยกยอทานไพหาแม่หลานได้ตายไพเมื่อวันที่ [3] ธันวา
คม ภ.ศ.
7) 2505 . กับ 3 [เหนื อ]ขึ้น 7 ฅํ่า เวลา 20[.20] นาที วันจัน นางหลานเกิดปีเป้าเปิ้ งงัว
เมื่อตายอายุ 62 ปี
8) กุสลส่วนบุญทีไ่ ด้กทํามานี้ ใพรอดใพเถิ่งอุ้มปกยกออกที่รา้ ยฅ้ายใพสู่ทด ี่ ี
9) ตามภูมทีต ่ นปากฏ จิ่มมีผู้บริจากเงินร่วมปิดธองพระพุทธรูปในโบถนี้ มีธ่านเจ้า
10) ฅุรสุเทพ . . . . . เจ้าคณะจังหวัดเชียงใหม่วด ั พระธาตุสุเทพ แลวัดทุงยูเปน
11) เงิน 2[9]0 บาท ภ้อมกันนี้ พ่อน้ อยแสนบริจาก 200 บาท พ่อน้ อยสม 27 บาท
พ่อหนานต . 27
12) . . . [เงิน] . . [417] บาท . . . สร้าง . . . . 26467 บาทถ้วน ทานเดือน 6 ออก 2[4]
ภ,ศ 2408

Translation:

In this year 1326 CS, 2507 BE, // Phò Thathipphotha whose house is situated be-
hind (the monastery of) Wat Buak Hong in Phra Sing sub-district, Müang dis-
trict, // Chiang Mai province, who is 63 years old in the year 2507 BE, had the
faith in the Teachings of the Buddha // spending 25,700 Baht of money to build
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 145

this ordination hall (ubosot) // along with his daughters Nang Bua Sum, Nang Si
Mòi, and Nang Bua Kaeo, with the intention to support the Teachings of the
Buddha at Wat // Mün Kòng. At the same time, he makes the donation for his
wife Mae Lan who had passed away on 3 December // 2505 BE, [i.e.] on the sev-
enth waxing day of the third [lunar] month of the northern [calendar] at 20
o’clock, on a Monday.35 Nang Lan was born in the year of the cow and died at
the age of 62. // This share of merit derived from this donation [shall help] move
her from an unpleasant place to a pleasant one // depending on her current
situation. The persons donating the money, together with putting golden leaves
on the Buddha statue placed in the ordination hall, are Cao // Khunsuthep, the
head of the Sangha of Chiang Mai province [residing at] Wat Phrathat Suthep
and Wat Thung Yu // amounting at 290 Baht. Besides, Phò Nòi Saen donated
200 Baht, Phò Nòi Som 27 Baht, Phò Nan (…) 27 Baht // (…) 417 Baht. (…) making
up a total of 26,467 Baht. This donation was made on the 24th of the sixth
month, 2408 BE.

Part II: In Thai (Siamese) script

13) นายทา ทิพโพธา ได้บริจาคทรัพย์สร้างอุโบสถวัดหมื่นเงินกอง ตํา

14) บลพระสิงห์ ๑ อ. เมือง จ. เชียงใหม่ สร้าง ๑๕ มีนาคม พ.ศ. ๒๕๐๗
15) เสร็จ ๓๑ ธันวา ๒๕๐๗ การสร้างสิ้นทุนทรัพย์ ๒๕๗๐๐ บาท
16) ขออุทิศกุศลให้นางหลานผู้เป็นภรรยาที่ตายไป ทําบุญฉลอง
17) วันที่ ๑๕-๑๖-๑๗ มิถุนายน ๒๕๐๘ ตรงเดือน ๖ ขึ้น ๑๓-๑๔-๑๕ คํ่า

Translation:

Mr Ta Thipphotha donated property for the construction of the ordination

hall (ubosot) at Wat Mün Ngoen Kòng // in Phra Sing sub-district, Müang
district, Chiang Mai province. The construction [started] on 15 March // and
finished on 31 December 1964. A capital of 25,700 Baht was spent. // May I
dedicate the merit to Nang Lan, my deceased wife. The merit-making is cele-
brated // on 15–16–17 June 1965 corresponding to the 13th–14th–15th waxing
days of the sixth month.

||
35 This date corresponds to Monday, 8 February 1965 (1326 Magha 7).
146 | Apiradee Techasiriwan, Volker Grabowsky

Part III: In Roman script

18) This church is dedicated to the memory of Mrs Lan

19) . . . . . . Money for the construction of this
20) building was donated by her husband
21) Mr Ta Tippota. The construction began
22) on 15 March and was finished on
23) 31 December 1964. The cost of this monument
24) [w]as 25,700 Baht. The dedication [cermon]
25) was held on 15, 16 and 17 March 1965

Part IV: In Tham script

Transcription into modern Thai script:

26) ภ้อมกันนี้ ค็ได้มีพ่อน้ อยสมริด ปัญจะ บ้านตัง้ อยู่หน้ าวัดหมื่นเงินกอง เกิดปี ใค้ ปี 25 . .
27) อายุได้ 66 มีสัทธาบริจาคเงินเปนจํานวน 6,700 บาท ส้างกําแพงด้นหน้ าวัด
28) ภ้อมด้วยลูกนายจัน น้ อยฅํา นางนวลน้ อย นาง[โ]ถน้ อยพี่นายบุญมี ไว้กับพระ
สาสนาภ้อม
29) กันนี้ ขอยกยอทานไพหาแม่บัวจีนผู้เปนเมียตายไพวันที่ 2 ธันวาคม ภ.ศ. 2504
30) เดือน 3 ออก 6 ฅํ่า วันอาทิด เวลา 13.15 นาที เมื่อก่ายอายุ 63 ภ้อมกันนี้ ทานไพหา
31) นายมงคล ขอกุสลบุญอุ้มปกยกออกที่รา้ ยข้ายใพสู่ทพ ื หื้อได้เสวยฅวามสุก .
ี่ ้น
32) ภ้อมกันนี้ สัทธาทังหลายบริจากส้างกําแพงด้านหลังแลเหนื อ เงิน 7939 บาท
33) รวมกานส้างภ้อมสาลา 2 หลัง 74,724 บาท (ต่อไปเป็นอักษรไทย) พระปลัดทอง
อินทร์ สีลสํวโร
เจ้าอาวาส . . . . . .

Translation:

At the same time, Phò Nòi Somrit Panca whose house is situated in front of Wat
Mün Ngoen Kòng and who was born in the year 25[…] // and 66 years old, has
the religious faith to donate an amount of 6,700 Baht to construct the wall on
the front side of the monastery. // [This donation is made] together with his
children Nai Can, Nòi Kham, Nang Nòi, Nang Tho Nòi who is elder than Nai
Bunmi [in support of] the Teachings of the Buddha. // Besides, I ask that this
donation is to the benefit of my wife, Mae Bua Cin, who passed away on 2 De-
cember 1961, // on the sixth waxing day of the third [lunar] month, a Sunday, at
 Multi-scriptural and Multilingual Inscriptions in Lan Na | 147

13.15 o’clock, at the age of 63. In addition, the donation is also dedicated to //
Nai Mongkhon. May this merit help them to move from an unpleasant place to a
place where they enjoy happiness. // All religious faithful have donated 7,939
Baht to build the wall on the backside of the monastery. // Together with the
construction of the two pavilions (sala) this total up to 74,724 Baht.
(followed by a colophon in Thai)
Phra Palat Thòng In Sinlasngwaro (Indra Sīlasaṃvaro), the abbot.
|
Scholarly Multilingualism
Szilvia Sövegjártó
On the Emergence and Development of
Middle Babylonian Bilinguals
Abstract: Sumerian literary manuscripts from the Middle Babylonian period
were often accompanied by Akkadian glosses or partially translated into Akka-
dian. However, determining whether a specific manuscript belongs to the Old or
Middle Babylonian period is not a straightforward task: Middle Babylonian
manuscripts frequently display a hybrid palaeography, incorporating elements
of both earlier and later grammar and sign forms. Scholars typically attribute
this phenomenon to the innovative and conservative tendencies of Middle Baby-
lonian scribes. Nonetheless, it is also plausible that Old Babylonian manu-
scripts were available and copied during the Middle Babylonian period, directly
influencing the stylistic characteristics of the final product. This article aims to
examine the origins and evolution of the corpus of Middle Babylonian bilingual
manuscripts from this perspective.

1 Introduction
The corpus of Middle Babylonian (c. 1595–1155 BCE) literary manuscripts1 is sig-
nificantly smaller than the Old Babylonian (c. 2004–1595 BCE). Despite several
recent publications that have utilised this corpus, it has not yet been fully pub-
lished.2 Another important distinction is that while most compositions are still
written in Sumerian, the manuscripts containing complete compositions, rather
than excerpts,3 are mostly bilingual. These manuscripts include either Akkadian
glosses or a more or less complete Akkadian translation.4

||
1 The present study refers with the term Middle Babylonian manuscripts specifically to manu-
scripts of the Kassite period.
2 See Viano 2016, 33, on the details.
3 These excerpts are mostly school exercises, published by Bartelmus 2016. On the format of
school tablets, see Bartelmus 2018. An overview to this corpus was also provided by Veldhuis
2000.
4 Viano 2016, 33–36, lists seventy-nine Middle Babylonian manuscripts, including thirty-three
bilinguals. Bilingualism also seems to be prominent in the scholarly corpus of the Middle
Babylonian period. Zomer 2018, 124 reports forty-six bilingual incantations from the Middle

Open Access. © 2024 the author, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-006
152 | Szilvia Sövegjártó

When distinguishing between Old Babylonian and Middle Babylonian literary

manuscripts, their ductus and visual organisation provide the most informative
clues, since explicit dating or colophons are generally not preserved on the
manuscripts. The typical format of Middle Babylonian bilingual manuscripts
was described by Jeremiah Peterson:

The Akkadian is entered below the Sumerian, and the translation is typically incomplete,
with only select lines rendered in Akkadian or sporadic glossing in Akkadian in smaller
script within the same line underneath the Sumerian it pertains, which is analogous to the
predominant method of glossing in Old Babylonian literary texts.5

This format will be referred to as the incomplete interlinear format.

Another format frequently observed in this period is the two-column format,
with the Sumerian version on the left and the Akkadian version on the right
column. The Akkadian version is generally more complete in this two-column
bilingual format6 compared to the other two types of visual organisation.
A third format provides an Akkadian version without any demarcation or
line break, as a direct continuation of the Sumerian text. This format is an inno-
vation of the Middle Babylonian period7 as it is virtually unattested in the Old
Babylonian period, except for a manuscript from Tell Harmal.8 It is also rare
within the Middle Babylonian bilingual corpus. This manuscript type will be
referred to as the continuous bilingual format.9

||
Babylonian period, compared to the five to seven bilingual manuscripts known from the Old
Babylonian period.
5 Peterson 2017, 260–261.
6 Peterson 2017, 261.
7 Zomer 2018, 133, refers to a precursor of this format in the Old Babylonian period.
8 The only example for this irregular format dating to the Old Babylonian period known to me
is the manuscript IM 53977 (RIME 3/2.1.2.38), a bilingual composition containing a royal in-
scription of Šulgi in a syllabic Sumerian and an Akkadian version. In this case, the unique
format is not the only irregularity.
9 The typology presented here is simplified compared to the very elaborated categorisation
provided by Zomer 2018, 127–138. Her typology considers not only Middle Babylonian but also
Middle Assyrian bilingual manuscripts. On the latter corpus, also see Wagensonner 2018. Here,
I only relate to the presentation of the Sumerian and Akkadian text regarding their position,
while Zomer considers several other features, including script size, rulings and indentation
techniques. I would argue that the choice of whether and how to apply ruling and indentation
was determined more by the scribes’ individual preferences, while the positioning of the Sume-
rian and Akkadian texts or text versions has praxeological implications. Therefore, the rough
typology presented here is justified as I do not ascribe any specific function to the aforemen-
tioned elements of visual organisation. Nevertheless, it should be pointed out that the variety
 On the Emergence and Development of Middle Babylonian Bilinguals | 153

Although these formats are also found in the Old Babylonian period, the
visual appearance of the tablets is distinct, as suggested by Peterson’s compari-
son of Old Babylonian and Middle Babylonian manuscripts from Nippur.10 The
ductus can be assessed based on diagnostic sign forms, including the KUR, NE
KA, RU, KUG, SIKIL, HUR, UB, MUNUS and HE2 signs.11 The use of Middle Baby-
lonian forms is sufficient to date a given manuscript to the Middle Babylonian
period. However, as has been observed by several scholars, manuscripts often
exhibit inconsistency in the use of Old Babylonian and Middle Babylonian sign
forms. This phenomenon is usually explained by the use of contemporary sign
forms while still drawing on the archaic and conservative Old Babylonian
forms.12
Mixing older and newer sign forms is not unique to this corpus. Neverthe-
less, in a culture where copying13 has played a significant role in the transmis-
sion of scholarly and literary content, there must be comprehensible reasons for
the combination of various sign forms.14 Contrary to the argument for conserva-
tism, it could be proposed that scribes literate in cuneiform also had their indi-
vidual writing styles and, as a consequence, they applied sign forms with some
consistency in their manuscripts. Therefore, the mixture of sign forms may re-

||
of formats attested in the Middle Babylonian period shows that the standardisation of bilingual
formats is a later phenomenon, as suggested by Zomer 2018, 138.
10 Peterson 2017, 261: ‘In contrast to prevailing O[ld] B[abylonian] Nippur formats, formal
column rulings are frequently used to establish the middle of the line and to establish a small
empty margin between the end of the first columns and the beginning of the second column.
Rulings on M[iddle] B[abylonian] literary texts are frequently distinguishable by their quality
and their depth, which is often quite shallow and rounded in comparison to O[ld] B[abylonian]
rulings.’
11 See Peterson 2017, 261. Considering diagnostic sign forms, those of the late Old Babylonian
First Sealand dynasty texts from Nippur should also be mentioned, including RI, UZ, ḪU, AR,
KA and GI, where the Winkelhaken of the first four signs is written behind the vertical wedge.
On this topic, see Nougayrol 1971, 68–69 with n. 4; George 2013, 129–130; Gabbay 2014, 148. A
more comprehensive overview and recent discussion is provided by Gabbay and Boivin 2018,
24.
12 See e.g. Peterson 2017, 261. He also mentions that such an admixture of various sign forms
‘is frequently encountered in Babylonian scripts’, however, I argue that whenever such incon-
sistencies occur, the reason behind them should be investigated.
13 On the cultural practice of copying, see Brita et al. 2020.
14 Krebernik 2001, 241, was more careful when assessing these features in a manuscript: ‘Ob
die sumerischen Zeichenformen tatsächlich eine ältere Vorlage widerspiegeln oder aber künst-
liche Archaismen sind (was mich wahrscheinlicher dünkt), wage ich nicht zu entscheiden. Eine
umfassende Studie zu dem bisher kaum beachteten Phänomen der Koexistenz älterer und
jüngerer Zeichenformen im selben Text steht noch aus.’
154 | Szilvia Sövegjártó

sult from the use of other scribes’ manuscripts as a source in the production
process, which involved copying, updating, redacting and complementing older
manuscripts, including the addition of (further) glosses or even an Akkadian
translation.

Fig. 1: CBS 3558 obverse (Nippur, Middle Babylonian Period, Penn Museum, Philadelphia), an
incomplete interlinear bilingual manuscript; courtesy of the Penn Museum.

It was initially observed by Andrew R. George in relation to the bilingual frag-

ment CBS 3558+, a manuscript of the royal hymn Lipit-Ištar A, that distinct fea-
tures of both Old Babylonian and Middle Babylonian dialects coexist in the
 On the Emergence and Development of Middle Babylonian Bilinguals | 155

Akkadian version of the composition.15 Based on this observation, he proposed a

multi-phase translation process that may have made use of previous Old Baby-
lonian glosses.16 The Old Babylonian status of this particular composition, the
royal hymn Lipit-Ištar A, indeed, suggests the existence of earlier versions with
only sporadic glosses. This hymn was among the few curricular texts studied in
elementary education during the Old Babylonian period, which could have
prompted the inclusion of a few or even additional glosses, depending on the
skills of the apprentice scribe. However, a complete Akkadian translation is
unlikely due to the predominant use of the Sumerian language in schools dur-
ing that era.
In the following, I will argue that while this manuscript may appear unique,
it is not truly one-of-a-kind. Additionally, the various types of bilingual manu-
scripts provide valuable insights into the production process of Sumerian liter-
ary manuscripts during the Middle Babylonian period.

2 Changes in the institutional background of

literary manuscript production
Scholarly education in the Old Babylonian period was the subject of several
studies, catered by the substantial amount of extant manuscripts related to this
institution, as well as relevant archaeological discoveries, including school
houses.17 Scribal education in the Middle Babylonian period is not documented
that extensively and, therefore, it also received moderate attention until recent-
ly. The most important studies were carried out by Leonhard Sassmannshausen
(1997, 2002), Niek Veldhuis (2000), Veldhuis and Hermann Hilprecht (2003–
2004) and in the recent monograph of Alexa Bartelmus (2016). A further im-
portant corpus, that of the incantations, was published by Elyze Zomer (2018).18

||
15 George 2012, 370.
16 George 2012, 370: ‘Probably individual words of the Sumerian text were glossed in Akkadi-
an in the Old Babylonian period, and these glosses only later joined up to yield a full transla-
tion.’
17 On a school house discovered in Nippur, see Robson 2001. Another school house, ‘No. 7,
Quiet Street’ is known from Ur, see Charpin 1986, 52, 69 and 434. A school in Babylon was
described by Pedersén 1998, 332; Pedersén 2005, 19.
18 Zomer 2018, 142, argues that especially bilingual incantations suggest that the genre was
incorporated into the scribal curriculum during the late Old Babylonian or early Middle Baby-
lonian period and, therefore, they are also relevant for the present paper.
156 | Szilvia Sövegjártó

The fundamental difference between Old Babylonian and Middle Babyloni-

an scribal education is not the corpus of Sumerian texts they used for this pur-
pose. School texts still mostly consisted of lexical and literary manuscripts, in
many cases, excerpts. These two genres provided the necessary skills to under-
stand scholarly texts, namely, access to the Sumerian lexicon and the complexi-
ties of Sumerian grammar. The lexical tradition could be also important to use
the language actively, by composing new texts analogous to the older ones.19
However, the status of the Sumerian language appears to be very different
compared to the Old Babylonian period. Though literary bilingualism is attested
from the Old Babylonian period onwards, the low number of bilinguals prove
that the bilingual format was more experimental than established in this era.20
While scribal and scholarly education was carried out in Sumerian, on an indi-
vidual basis and quite infrequently, Akkadian aids and hermeneutic tools mak-
ing use of the scribes’ native tongue were also applied. In the Middle Babyloni-
an period, by contrast, Sumerian was no longer the mere device of scholarly
activities and the language of education. Akkadian had already been imple-
mented in schools as a teaching language as it was appropriate and encouraged
to comment and translate Sumerian literary compositions more extensively than
before. Moreover, Sumerian enjoyed a status of high prestige and, as a venera-
ble language, was also the carrier of a cultural heritage.21

3 Case studies of Middle Babylonian bilingual

manuscripts
In the following, I would like to present some case studies of Middle Babylonian
bilingual manuscripts. I will focus on incomplete interlinear bilinguals with
some typical features related to mixing Old Babylonian and Middle Babylonian
characteristics. These hybrid features were also pointed out by the respective
editors of these manuscripts. As I will argue, such hybrid manuscripts invite
further investigation because they probably had a similar setting to manuscript
CBS 3558+. The latter is not the only one where the Middle Babylonian scribe
possibly used an Old Babylonian master copy when preparing his own exem-
plar.

||
19 Bartelmus 2016, 249.
20 Cooper 1969, 13; Cooper 1993, 79.
21 Bartelmus 2016, 5.
 On the Emergence and Development of Middle Babylonian Bilinguals | 157

My aim is to shed new light on the direct influence of Old Babylonian liter-
ary manuscripts on their Middle Babylonian counterparts. The close relation-
ship of the two corpora has been proven and extensively discussed by scholars,
however, mostly restricted to the level of contents, especially to literary patterns
and models. Here, I would suggest a more direct relationship apparent only in a
few cases, namely, where exemplars of Old Babylonian literary manuscripts
could have served as master copies when scribes recopied and redacted them in
the Middle Babylonian period.

3.1 BM 78164
The manuscript BM 78164 (CT 58, 70) contains a bilingual liturgical composition
dedicated to the god Enki in Sumerian22 and Akkadian, written in interlinear
format.23 The Akkadian version was written with a somewhat smaller script and
provides a partial translation to the Sumerian composition. The Sumerian and
the Akkadian version of each line is separated from the previous and subse-
quent text line by a ruling. The manuscript probably comes from Sippar, and,
thus, forms part of the northern Babylonian tradition.
The dating of the manuscript to the Middle Babylonian period is secured by
the Middle Babylonian sign forms of RU (obv. 7’ and 9’) and KUR (rev. 20’).24
Maurizio Viano provides an in-depth discussion of this manuscript25 and points
out that the Akkadian version of the text exhibits a mixture of Old Babylonian
and Middle Babylonian features. He also claims that the Sumerian version was
redacted in the Middle Babylonian period because of some features, however,
two of his three remarks are rather invalid26 and the use of the terminative with
a human referent (obv. 4’ passim and obv. 24’) is already common in the Old
Babylonian corpus.

||
22 In the Emesal dialect of Sumerian.
23 The text was edited by Geller 1992.
24 The line numbers provided here deviate from those given by Viano 2016 and refer to the edi-
tion of the Electronic Babylonian Library (<https://www.ebl.lmu.de/fragmentarium/BM.78164>
accessed on 9 August 2023).
25 Viano 2016, 69–70.
26 Viano reads a sign as [ka] in rev. 20’ but it is probably a [de3!] and, thus, correct. In rev. 12’, I
would read mur7 še10-ba, phonetic mur7-ku-ba or prefer solutions considering a dittography, a
scribal lapse or a frozen gloss here, but certainly not an imperative prefixed erroneously with
the /ba/-.
158 | Szilvia Sövegjártó

a b

Figs 2a–b: BM 78164 obverse (Sippar, Middle Babylonian Period, The British Museum, Lon-
don), an incomplete interlinear bilingual manuscript; © The Trustees of the British Museum,
shared under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International
(CC BY-NC-SA 4.0) licence.

The Old Babylonian features of the Akkadian version are the mimation pre-
served in two nominal forms (obv. 15’: i-na e-re-em pa-nim; rev. 15’: su-mu-˹uk-
ta˺-am), though missing in all other contexts,27 as well as the use of the Old
Babylonian possessive suffix -šu in the text.28 Further Old Babylonian peculiari-
ties are observable on the level of the script, namely, the use of the signs ŠA and
QA2 (rev. 7’), as expected in Old Babylonian manuscripts, instead of their Middle
Babylonian counterparts ŠA2 and QA. Nevertheless, several CVC signs (obv. 25’:
PIŠ; rev. 11’: ṬIR, ŠIṬ; rev. 19’: LAK; rev. 21': LID2, LUL) are typical of the Middle
Babylonian orthography.29
While most Middle Babylonian traits are restricted to the Akkadian version,
the Middle Babylonian sign form for KUR is also attested in the Sumerian ver-
sion of the text, suggesting that the scribe was copying an Old Babylonian man-
uscript, probably one containing a few glosses, without being completely rigid

||
27 On the loss of the mimation by the Middle Babylonian period, see Aro 1955, 32.
28 On the orthography of the pronominal suffixes in the Middle Babylonian period with the
loss of the final vowel see Aro 1955, 54–58.
29 On this matter, see Aro 1955, 22.
 On the Emergence and Development of Middle Babylonian Bilinguals | 159

in keeping to the archaic orthography. All in all, this manuscript is a further

candidate for an Old Babylonian manuscript, originally with sporadic glosses,
copied and supplemented with a more complete Akkadian translation in the
Middle Babylonian period.

3.2 HS 1512
The manuscript HS 1512 contains a bilingual Ki-utu incantation.30 The tablet was
probably unearthed in Nippur and dates to the Middle Babylonian period, as-
sured by the typical sign form of KUR (obv. 4b and 9a) both in the Sumerian and
the Akkadian version.31 The manuscript exhibits the interlinear bilingual for-
mat: the Sumerian and the Akkadian version of each line are separated from the
previous and subsequent text line by a ruling. The Akkadian version is written
with smaller signs than the Sumerian text. While the obverse of the tablet, or at
least the preserved fragment of it, gives a complete Akkadian translation to the
Sumerian text, the reverse only contains a few glosses arranged under the cor-
responding Sumerian expressions. A remarkable exception is the first line of the
obverse where the Akkadian version is written above the Sumerian line, on the
upper edge of the tablet.32
The palaeography of the tablet is characterised by Old Babylonian sign
forms in the Sumerian version, while the Akkadian version contains several
later sign forms. Manfred Krebernik compared the palaeography of the sign
forms which appear in both versions, namely KUG, RU and two signs from the
GIRI3 group, and concluded that the newer forms appear consequently in the
Akkadian version.33
The editor’s further observations include the presence of the mimation in
most cases, typical of the Akkadian grammar of the Old Babylonian period. The
use of the Old Babylonian signs QA2 and ŠA (obv. 10b) is also attested in the

||
30 The tablet was published by Krebernik 2001.
31 See Viano 2016, 75. Krebernik 2001, 240, proposed either a late Old Babylonian or Middle
Babylonian date. Zomer 2018, 143, proposed a late Old Babylonian or early Middle Babylonian
date, listing some of the archaic and Middle Babylonian features of the manuscript on pp. 144–
145. The lines of the tablet are quoted according to Krebernik 2001, however, (a) refers to the
Sumerian, (b) to the Akkadian version of each text line.
32 Krebernik 2001, 240. Zomer 2018, 82, suggested that this deviation might correspond to
archival practices, though the archival context of the manuscript remains unknown.
33 Krebernik 2001, 241. He also raised attention to further archaic sign forms in the Sumerian
version without counterparts in the Akkadian text.
160 | Szilvia Sövegjártó

Akkadian version, suggesting an earlier dating of parts of the Akkadian text. By

contrast, the use of several CVC signs (obv. 1b: NAP, ŠA2; 9b: KUN, MIT, U; 10b:
BIŠ) is a Middle Babylonian trait of the Akkadian text.34
Assessing the status of this manuscript, conclusions are more problematic
than in the previous case. First of all, the provenance of the tablet is only se-
cured by the fact that most manuscripts in the Hilprecht Collection, but espe-
cially those of literary content, come from Nippur. Krebernik probably men-
tioned on purpose that no information is provided on the provenance of this
tablet,35 since Ki-utu incantations typically come from and are related to the
temple of the sun god in Sippar.36 Viano explicitly argues for reconsidering the
provenance of the manuscript based on these arguments.37 Even if the manu-
script’s provenance was established correctly, whether the tablet was produced
in or was copied based on a master copy from the Sippar area should be consid-
ered.
The present manuscript seems to be more problematic at first sight when
assessing the status of the Sumerian and Akkadian versions because the Akka-
dian version shows a high intertwining of Old and Middle Babylonian peculiari-
ties. While the Sumerian version is characterised by archaic orthography and a
fairly archaic grammar, except for a few, late Old Babylonian features, the Ak-
kadian version is a mixture of Old Babylonian grammatical traits and Middle
Babylonian orthographic peculiarities. Nevertheless, these peculiarities do not
co-occur in single words but in some lines,38 and, thus, it is possible that, in-
deed, Old Babylonian glosses or even a rather extensively glossed manuscript
was completed by the scribe to provide a full translation to the Sumerian text.
Interestingly, the sporadic glosses of the reverse include no Middle Babylonian
peculiarities at all, which further reinforces the hypothesis of glosses taken over
by the scribe from an Old Babylonian manuscript. A further, even if minor, hint
of a master copy used in the course of the production process of the present
manuscript can be found in obv. 3b. The Sumerian verb luh-ha-ab is provided
with two equivalents in Akkadian, namely mi-si ‘wash!’ and ub-bi-ib ‘clean!’,
separated by a double wedge. As, in this case, two translations are allocated to a

||
34 Krebernik 2001, 240, and, generally, Aro 1955, 22.
35 Krebernik 2001, 238.
36 Compare the list of manuscripts in Krebernik 2001, 238, and the close parallel CBS 1529,
also probably from Sippar, mentioned in Krebernik 2001, 242, n. 15.
37 Viano 2016, 75. On the corpus of the Ki-utu incantations in general, see Baragli 2022, 27–31.
38 See e.g. obv. 1b: [dUTU š]ar-rum da-a-a-nu ša2 AN u KI. In this line, the Old Babylonian trait is
the mimation on šarrum, while dayyanu lacks mimation and is followed by the sign ša2, both
peculiar for the Middle Babylonian period.
 On the Emergence and Development of Middle Babylonian Bilinguals | 161

lexeme, both probably come from an original manuscript and, thus, the scribe
wished to keep them in the copy instead of making a decision of which transla-
tion to take over.

4 The emergence and development of literary

bilinguals in the Middle Babylonian period
The archaic sign forms discovered in the Middle Babylonian literary corpus can
be interpreted in two ways. Firstly, they may have been inherited from older
manuscripts. Secondly, they could be deliberate attempts by scribes to imitate
older manuscripts, aiming to harmonise the form and the content, namely, the
Sumerian language, considered as an archaic relict, with the sign forms used for
writing the manuscript. I will argue that both of these strategies can be traced in
the corpus of Middle Babylonian literary bilinguals. Although archaisms may
appear similar on the surface, distinguishing their function requires careful
consideration of their scope and presence in either or both versions of the bilin-
gual text.
To conclude regarding the emergence and development of Middle Babylo-
nian bilinguals, it is important to carefully assess all three main types of bilin-
gual formats. What I will present here is a hypothesis based on the variety of
visual organisations found in the Middle Babylonian literary and scholarly cor-
pus, which provide insights into how Middle Babylonian manuscripts became
bilingual.
During the Middle Babylonian period, scribes continued to copy Sumerian
manuscripts from the Old Babylonian period, as they were part of a literary
heritage written in a highly esteemed language. In some cases, it can be proven
that not only the content was transmitted from the Old Babylonian to the Middle
Babylonian period, but Old Babylonian manuscripts also served as master cop-
ies for producing new tablets. However, Middle Babylonian scribes took an
innovative approach when creating their own manuscripts, providing them with
more or less complete or incomplete Akkadian translations.
Middle Babylonian literary bilinguals exhibit several similarities in their
microstructure, including formatting, ruling and indentation techniques, use of
innovative sign forms and grammar. However, when examining the sources of
text production, specifically the various types of Old Babylonian manuscripts
available to Middle Babylonian scribes, a key distinguishing factor is the visual
162 | Szilvia Sövegjártó

organisation, namely, the arrangement of the Sumerian and, where present, the
Akkadian text on the tablet in relation to each other.
The three dominant formats of bilinguals known from this period might
have been influenced by their respective master copies. This does not necessari-
ly mean that Middle Babylonian scribes adhered strictly to the format of the
master copy, but rather that their formatting choices were influenced by the
resources available to them when preparing their manuscripts. When they could
only draw on the Sumerian version and had to create their own translation, they
preferred either continuous bilinguals or parallel columns. The first option in-
volved translating the Sumerian text line by line, placing the Akkadian version
directly after the corresponding Sumerian phrase. Some interlinear bilinguals
may also have been inspired by monolingual Sumerian manuscripts, in case the
scribe chose to establish a clear separation between the Sumerian and Akkadian
versions, either through ruling or indentation techniques, or by keeping the two
versions on separate text lines – a practice commonly observed in cuneiform
text production. These formats could also invite archaisms, both in the script
and grammar, although they manifest differently compared to the case studies
presented above. Importantly, the archaisms in these manuscripts primarily
characterise the Sumerian version, while the Akkadian version adheres to Mid-
dle Babylonian standards.39 Therefore, the hybridity of the Akkadian version
becomes a distinct characteristic of those manuscripts where the Akkadian
version draws on pre-existing contents preserved in the master copy.
The second format, arranging the two versions in parallel columns, allowed
the scribes to create a complete copy first and then proceed with the translation
in the second column. This formatting technique provided a clear boundary
between the two versions, making them accessible separately for performative
or reference purposes. Moreover, the parallel arrangement facilitated direct
comparison, as corresponding lines were positioned next to each other.
When the format of parallel columns was employed, it can generally be as-
sumed that the Akkadian version would be free of archaisms, as the translation
was the work of the Middle Babylonian scribe. However, it is worth noting that
exceptions exist. The manuscript CBS 11341 (PBS 1/1 11), for instance, a two-
column tablet found in Nippur, contains a bilingual text. While the reverse of
the tablet is well-preserved, the obverse has nearly been lost. The Sumerian text
is presented in the left column, and the Akkadian version is found in the right
column. However, the lines do not evenly match, and only the Sumerian lines

||
39 Representative examples of this type of visual organisation are the Nippur manuscripts CBS
13905 and N 3395.
 On the Emergence and Development of Middle Babylonian Bilinguals | 163

have rulings. This particular composition is probably a ritual description which

once belonged to a hymn dedicated to King Šulgi, with no known duplicates
from earlier or later periods. The dating of the tablet has been a subject of de-
bate, Veldhuis even suggested that the bilingual composition is entirely a prod-
uct of the Middle Babylonian period.40 In this case, the archaisms present in the
composition41 would be deliberate attempts by the scribe to align the form and
content with the older tradition. By incorporating archaic elements into the
newly composed hymn, the scribe demonstrates their erudition and knowledge
of the Sumerian literary heritage, both in terms of its form and content. Notably,
the archaic elements in this specific case are present in both versions, but pri-
marily on the level of script and orthography, while grammatical archaisms are
completely absent in both the Sumerian and Akkadian versions. Additionally,
the Sumerian version displays some unique and innovative lexical elements
that are not found in the extensive Old Babylonian literary corpus. This suggests
that the author drew more from the Middle Babylonian lexical tradition when
composing the hymn, rather than relying solely on the established literary style
and phraseology of the Old Babylonian period.
Lastly, the (incomplete) interlinear format, which was discussed in detail in
two case studies above, facilitated the simultaneous copying and translation of
a text line by line. Furthermore, this format proved useful for preserving the
glosses from the master copy and supplementing them with additional annota-
tions when necessary, resulting in manuscripts with more complete, albeit
sometimes still partially incomplete, Akkadian versions. One characteristic of
these manuscripts, especially when using a glossed manuscript as the master
copy, is the mixture of Old Babylonian and Middle Babylonian grammar and
orthography within the Akkadian version.
In summary, the high formal variety within the relatively small corpus of
Old Babylonian bilinguals can be attributed to the different affordances provid-
ed by various visual organisations, which were probably advantageous in an

||
40 Veldhuis 2008, 31, n. 11, dated the tablet to the Middle Babylonian period, while West-
enholz 2005, 345, argued for an Old Babylonian date. For the discussion of their arguments pro
and contra as well as for further reasoning for a Middle Babylonian date, see Viano 2016, 47–
48. On the dating of the composition, but not this specific manuscript, see Sövegjártó 2022.
41 Especially the mixing of Old Babylonian and Middle Babylonian sign forms throughout the
composition could be such a feature, but archaisms are also identifiable on the level of the
grammar. Furthermore, the lexicon can be also used as an argument for a later dating, as
several unorthographic writings and terms only known from lexical compositions were inte-
grated into the text. On the exact lexemes, see Viano 2016, 47, n. 175. All in all, Viano also
argues for a Middle Babylonian dating.
164 | Szilvia Sövegjártó

educational context. However, in the case of the Middle Babylonian corpus, the
variety of formats stems more from scribal practices. The formatting choices of
Middle Babylonian scribes were probably also influenced by the range of Old
Babylonian master copies available to them, including monolingual Sumerian
manuscripts, manuscripts with sporadic glosses, and even complete or incom-
plete bilinguals. Additionally, translation practices and the Sumerian proficien-
cy of the scribe, allowing for direct line-by-line translation, could also have
influenced the format chosen. Nevertheless, it cannot be entirely ruled out that
other factors, such as specific assignments in a teaching context, might have
constrained scribal choices.42

5 Conclusion
Glossed manuscripts and bilinguals emerge contemporaneously in the Old Bab-
ylonian period, therefore, the concept proposed by Joachim Krecher43 that bilin-
guals once developed from glossed manuscripts should be refuted as I have
argued elsewhere.44 In the end, apparently, some glossed manuscripts were, in
fact, precursors of bilingual manuscripts, though not from a synchronic but
from a diachronic perspective.
As was shown through a few examples, at least some bilingual manuscripts
dated to the Middle Babylonian period probably used Old Babylonian glossed
manuscripts as their models. While scribes in the Old Babylonian period did not
usually preserve the glosses of a manuscript while copying as they were mostly
intended for the personal use of the scribes, this practice changed in the course
of the Middle Babylonian period. Most probably, as soon as Sumerian did not
played such a central role in education and apprenticeship, as it was the case in
the Old Babylonian period, any hints surviving in the form of glosses and trans-
lations proved to be useful and worth preserving for the copyist.
The fact that Middle Babylonian scribes appreciated and safeguarded the
glosses by taking them over to the newly produced manuscript prove the differ-

||
42 See also Cooper 1993, 80–83, who argues for the practicality of the interlinear format re-
garding the format of the clay tablet and the use of the multicolumn format in an educational
context. Jerrold S. Cooper only provides a rough overview without referring specifically to any
period. However, the affordances of bilingual literary manuscripts are quite different in the
early and late second millennium BCE.
43 Krecher 1980, 127.
44 Sövegjártó 2020, 91.
 On the Emergence and Development of Middle Babylonian Bilinguals | 165

ent status of the Sumerian language and literary heritage compared to the early
second millennium BCE. Sumerian was still an estimated language of erudition,
but also an ancient and prestigious relict subject to preservation together with
its written heritage. Therefore, unlike Old Babylonian scribes, the literates of the
Middle Babylonian period opted for preserving the heritage of the past in its
entirety, and complemented the extant manuscripts according to their best
knowledge of what they certainly considered inferior.

Acknowledgements

The research for this paper was funded by the Deutsche Forschungsgemein-
schaft under Germany’s Excellence Strategy – EXC 2176 ‘Understanding Written
Artefacts: Material, Interaction and Transmission in Manuscript Cultures’, pro-
ject no. 390893796. The research was conducted within the scope of the Centre
for the Study of Manuscript Cultures at the Universität Hamburg.

Abbreviations
BM = British Museum, London
CBS = Catalogue of the Babylonian Section, University Museum, Philadelphia, PA
CT = Cuneiform Texts, Babylonian Tablets, British Museum
HS = Hilprecht Collection, Jena
N = Nippur, University Museum, Philadelphia, PA
PBS = Publications of the Babylonian Section, University of Pennsylvania, Philadelphia, PA
RIME = Royal Inscriptions of Mesopotamia, Early Periods

References
Aro, Jussi (1955), Studien zur mittelbabylonischen Grammatik (Studia Orientalia, 20), Helsinki:
Societas Orientalis Fennica.
Baragli, Beatrice (2022), Sonnengrüße: Die sumerischen Kiutu-Gebetsbeschwörungen (Ancient
Magic and Divination, 19), Leiden: Brill.
Bartelmus, Alexa (2016), Fragmente einer großen Sprache: Sumerisch im Kontext der Schrei-
berausbildung des kassitenzeitlichen Babylonien, 2 vols (Untersuchungen zur Assyriolo-
gie und Vorderasiatischen Archäologie, 12), Berlin: De Gruyter.
Bartelmus, Alexa (2018), ‘Formale Besonderheiten mittelbabylonischer Schülertafeln’, in Eva
Cancik-Kirschbaum and Babette Schnitzlein (eds), Keilschriftartefakte: Untersuchungen
zur Materialität von Keilschriftdokumenten (Berliner Beiträge zum Vorderen Orient, 26),
Gladbeck: PeWe, 63–94.
166 | Szilvia Sövegjártó

Brita, Antonella, Giovanni Ciotti, Florinda De Simini and Amneris Roselli (eds) (2020), Copying
Manuscripts: Textual and Material Craftsmanship (Series Minor, 93), Naples: UniorPress.
Charpin, Dominique (1986), Le clergé d’Ur au siècle d’Hammurabi (XIXe XVIIIe siècles av. J.-C.)
(Hautes études orientales, 22), Genève: Droz.
Cooper, Jerrold S. (1969), Sumero-Akkadian Literary Bilingualism, PhD thesis, University of
Chicago.
Cooper, Jerrold S. (1993), ‘Bilingual Babel: Cuneiform Texts in Two or More Languages from
Ancient Mesopotamia and Beyond’, Visible Language, 27: 68–96.
Gabbay, Uri (2014), ‘A Balaĝ to Enlil from the First Sealand Dynasty’, Zeitschrift für Assyriologie
und Vorderasiatische Archäologie, 104: 146–170.
Gabbay, Uri and Odette Boivin (2018), ‘A Hymn of Ayadaragalama, King of the First Sealand
Dynasty, to the Gods of Nippur: The Fate of Nippur and Its Cult during the First Sealand
Dynasty’, Zeitschrift für Assyriologie und Vorderasiatische Archäologie, 108: 22–42.
Geller, Markham J. (1992), ‘CT 58, No. 70. A Middle Babylonian Eršahunga’, Bulletin of the
School of Oriental and African Studies, 55: 528–532.
George, Andrew R. (2012), ‘Recension of: Jeremiah Peterson, Sumerian Literary Fragments in
the University Museum, Philadelphia (Biblioteca del Próximo Oriente Antiguo, 9), Madrid:
Consejo Superior de Investigaciones Científicas, 2011’, Bulletin of the School of Oriental
and African Studies, 75: 368–370.
George, Andrew R. (2013), Babylonian Divinatory Texts Chiefly in the Schøyen Collection (Cor-
nell University Studies in Assyriology and Sumerology, 18), Bethesda, MD: CDL Press.
Krebernik, Manfred (2001), ‘Ein ki-dutu-Gebet aus der Hilprecht-Sammlung’, Zeitschrift für
Assyriologie und vorderasiatische Archäologie, 91: 238–252.
Krecher, Joachim (1980), ‘Interlinearbilinguen’, in Michael P. Streck (ed.), Reallexikon der
Assyriologie und Vorderasiatischen Archäologie, vol. 5, Berlin: De Gruyter, 124–128.
Nougayrol, Jean (1971), ‘Nouveaux textes sur le ziḫḫu (II)’, Zeitschrift für Assyriologie und Vor-
derasiatische Archäologie, 65: 67–84.
Pedersén, Olof (1998), Archives and Libraries in the Ancient Near East 1500–500 BC, Bethesda,
MD: CDL Press.
Pedersén, Olof (2005), Archive und Bibliotheken in Babylon: Die Tontafeln der Grabung Robert
Koldeweys 1899–1917 (Abhandlungen der Deutschen Orient-Gesellschaft, 25), Berlin:
Deutsche Orient-Gesellschaft.
Peterson, Jeremiah (2017), ‘A Middle Babylonian Sumerian Fragment of the Adapa Myth from
Nippur and an Overview of the Middle Babylonian Sumerian Literary Corpus at Nippur’, in
Lluis Féliu, Fumi Karahashi and Gonzalo Rubio (eds), The First Ninety Years: A Sumerian
Celebration in Honor of Miguel Civil (Studies in Ancient Near Eastern Records, 12), Berlin:
De Gruyter, 262–283.
Robson, Eleanor (2001), ‘The Tablet House: A Scribal School in Old Babylonian Nippur’, Revue
d’Assyriologie et d’Archéologie Orientale, 95: 39–67.
Sassmannshausen, Leonhard (1997), ‘Mittelbabylonische runde Tafeln aus Nippur’, Baghdader
Mitteilungen, 28: 185–208.
Sassmannshausen, Leonhard (2002), ‘Zur babylonischen Schreiberausbildung’, Baghdader
Mitteilungen, 33: 211–228.
Sövegjártó, Szilvia (2020), Sumerische Glossenhandschriften als Quellen des altbabylonischen
hermeneutischen Denkens (dubsar, 18), Münster: Zaphon.
Sövegjártó, Szilvia (2022), ‘Source texts as authority constructions: A conceptual approach to
the Old Babylonian literary discourse’, in Chiara Meccariello and Jennifer Singletary (eds),
 On the Emergence and Development of Middle Babylonian Bilinguals | 167

Uses and Misuses of Ancient Mediterranean Sources: Erudition, Authority, Manipulation

(Studies in Education and Religion in Ancient and Pre-Modern History in the Mediterrane-
an and Its Environs, 12), Tübingen: Mohr Siebeck, 163–179.
Veldhuis, Niek (2000), ‘Kassite Exercises: Literary and Lexical Extracts’, Journal of Cuneiform
Studies, 52: 67–94.
Veldhuis, Niek and Hermann Hilprecht (2003–2004), ‘Model Texts and Exercises from the
Temple School of Nippur: BE 19’, Archiv für Orientforschung, 50: 28–49.
Viano, Maurizio (2016), The Reception of Sumerian Literature in the Western Periphery (Antichi-
stica, 9; Studi Orientali, 4), Venice: Edizioni Ca’Foscari.
Wagensonner, Klaus (2018), ‘Sumerian in the Middle Assyrian Period’, in Jens Braarvig and
Markham J. Geller (eds), Multilingualism, Lingua Franca and Lingua Sacra, Berlin: Max-
Planck-Gesellschaft zur Förderung der Wissenschaften, 225–297.
Westenholz, Joan Goodnick (2005), ‘Sing a Song for Šulgi’, in Yitzhak Sefati, Pinhas Artzi,
Chaim Cohen, Barry L. Eichler and Victor Avigdor Hurowitz (eds), ‘An Experienced Scribe
Who Neglects Nothingʼ: Ancient Near Eastern Studies in Honor of Jacob Klein, Bethesda,
MD: CDL Press, 343–373.
Zomer, Elyze (2018), Corpus of Middle Babylonian and Middle Assyrian Incantations (Leipziger
Altorientalistische Studien, 9), Wiesbaden: Harrassowitz.
Antonio Manieri
Vernacular Terms in Sinitic Texts:
Multilingualism in Eighth-century Japanese
Documents
Abstract: The paper proposes the analysis of manuscript Japanese documents in
Sinitic (on wooden tablets or on paper) in which specialised vernacular Japa-
nese terms are employed. These terms are used regardless of the availability of
equivalent Sinitic terms, their referents are not culturally specific and, there-
fore, their usage may not be justified on the basis of cultural particularity. The
paper explores the factors that motivate such departures from established
norms and their implications for the communicative efficacy of the documents
in question, and aims to elucidate the rationale behind the compiling of such
multilingual texts. Additionally, the study touches tangentially upon the pro-
cess of sharing and learning such terminologies among different actors of the
eighth-century state.

1 Introduction
In this paper, I present some results of ongoing research into the technical and
artisanal terminologies of ancient Japan, specifically from the late seventh to
the eighth centuries CE, with the aim of shedding light on such terms and their
underlying concepts in the domains of craftsmanship and technology. As no
Japanese treatises on technical knowledge from this period have been found,
and probably none were produced,1 reconstructing these matters involves rec-
ognising terms, actors and practices in other types of texts, such as legal
sources, dictionaries and encyclopaedias, and administrative and practical
documents. I use the term ‘practical documents’ to refer to texts containing
highly technical content, composed for practical purposes.

||
1 It goes without saying that Chinese technical books circulated widely in Japan, since, as is
well known, the reception of the Chinese civilisation had a bookish character. For a synthesis
of the problems regarding the history of technical knowledge in pre-1600 Japan, refer to Buhr-
man 2017. For more information on the reception of Chinese civilisation and its bookish charac-
ter, see von Verschuer 1985, 251–256.

Open Access. © 2024 the author, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-007
170 | Antonio Manieri

The present study entails the examination of five distinct case studies that
showcase a diversity of examples regarding the typology of the documents, the
context of their production and consumption, and the related terminology.
Through the discussion of these sources, the paper contends that, despite the
established conventions governing the compilation of documents in Sinitic,
there are instances where specialised vernacular terms are employed,2 regard-
less of the availability of an equivalent Sinitic term. It is notable that referents of
these terms are not culturally specific, and, therefore, their usage may not be
justified based on cultural particularity.
The paper, thus, explores the factors that motivate such departures from es-
tablished norms and their implications for the communicative efficacy of the
documents in question, and aims to elucidate the rationale behind the compila-
tion of such multilingual texts. Additionally, the study touches tangentially
upon the process of sharing and learning such terminologies among different
actors of the eighth-century state.

2 Multilingual texts in ancient Japan

The linguistic repertoire3 of ancient Japan was characterised by a macroscopic
variation related, first of all, to the presence of sources in Sinitic versus those in
Japanese vernacular. Sinitic, as the cosmopolitan language of prestige and
learning, and widespread throughout East Asia, was a conservative written
medium over a large and shifting area that went far beyond the lands where
Chinese languages were spoken, and Japanese was certainly also a spoken lan-
guage of the archipelago.4 Actually, scattered evidence suggests that spoken
Chinese was also used on the archipelago, at least in the early eighth century,
but only at the level of cultural elites. One passage, for example, reports that

||
2 I use the word ‘term’ in the sense of modern terminological science and particularly as ‘a
conventional symbol that represents a concept defined within a particular field of knowledge’
(Cabré 1999, 81).
3 I use the term ‘linguistic repertoire’ to refer to the collection of linguistic resources available
to a language community, which individuals within that community use to communicate and
convey meaning.
4 It was particularly the obligatory language of diplomacy and statecraft throughout East Asia
until the late nineteenth century, but it was not imposed by China, since each East Asian socie-
ty used the language of its neighbour voluntarily. On the choice of the term ‘Sinitic’ rather than
‘Classical’ or ‘Literary Chinese’, see Mair 1994 and 2001.
 Vernacular Terms in Sinitic Texts | 171

five scholars were each ordered to teach ‘Han speech’ (kango 漢語),5 in other
words, the spoken language of the Han country (i.e. Chinese), to two pupils
apiece. In any case, the phenomenon was transitory and not solidified as in the
case of written Sinitic.6
The limited use of Sinitic as an oral language speaks against the existence
of ‘diglossia’ (much less that of ‘bilingualism’) in Charles Ferguson’s terms,
leading some scholars to state the ‘fallacy of bilingualism’ in favour of a continu-
um in terms of written language/script between Sinitic and Japanese vernacular.7
This script continuum exhibits several variations. In the realm of Sinitic, the
language of translated Buddhist scriptures, laws and official documents, and
literary poetry displays significant differences. Moreover, certain texts exhibit a
distinctive style that lies somewhere between the two ends of the scriptural
continuum. The Kojiki 古事記 (‘Record of Ancient Matters’, 712) and the Harima
no kuni fudoki 播磨国風土記 (‘Record of the Province of Harima and its Cus-
toms’, c. 714), for instance, feature a sort of ‘multi-grammatical prose’ in which
logograms are arranged in both Chinese and Japanese syntax.8 But Japanese
had prestigious uses too, such as being the language of royal edicts (shō 詔) that
‘conveyed the august word’ (mikotonori) of the sovereign9 – the only official
documents written in Chinese characters used as phonograms – and the lan-
guage of norito 祝詞 prayers and formulas, with the magic implications they
possess. The script used for edicts and norito prayers is called senmyōgaki 宣命
書 (‘writing in the style of edicts’).10
Moreover, local variation is also attested in sources, since, besides the so-
called Western Old Japanese, we also find some languages of fragmentary attes-

||
5 In this paper, Japanese names and bibliographic references are transcribed in the revised
Hepburn system, and Chinese names and bibliographic references in pinyin. Linguistic data
from Old Japanese are presented following Vovin 2020, from Medieval Chinese following Bax-
ter and Sagart 2014. Characters are furnished only for titles of books and when useful to the
argument.
6 The passage is in the Shoku Nihongi 続日本紀 (‘Chronicle of Japan. Continuation’, 797), Book 10,
Tenpyō 2.3.27 (Aoki Kazuo et al. 1990, 132–133). See also Duthie 2014, 208–209, Manieri 2022a, 182.
7 Lurie 2011, 323–334.
8 Lurie 2011, 227–228; Sema Masayuki 2011, 35–48; Palmer 2016, 13, 55–56.
9 In Japanese the character shō 詔 (‘edict’) was spelled as mikotonori, lit. ‘relating the august
word’, as it was intended to be read publicly. On the form of shō, cf. Migliore 2011, 18–21.
10 On senmyōgaki, see Bender 2009; Lurie 2011, 250–253; on norito, see Philippi 1990. On the
prestige usage of Japanese in general, see Hayakawa Shōhachi 1997, 3–21, which highlights the
passage from the ‘world of orality’ (口頭の世界) of pre-ritsuryō Japan to the ‘world of docu-
ments’ (文書の世界) established by Nara’s strong administrative state.
172 | Antonio Manieri

tation, such as Eastern Old Japanese and the Kyūshū dialects.11 When discussing
the languages of ancient Japan, scholars usually refer to a passage from the
somewhat cryptic text titled Tōdaiji fuju monkō 東大寺諷誦文稿 (‘Tōdai Monas-
tery Recitation Draft Text’, early ninth century), which refers to languages spo-
ken in Japan, Great Tang, Silla, Persia, India and South East Asia, as well as
several local dialects spoken in the archipelago, namely those of ‘this land’, the
Emishi, Hida (to the north of Yamato) and the Azuma (eastern regions).12 Alt-
hough the passage does not aim to present the linguistic scenario and focuses
instead on the Buddha’s universal ability to communicate, it indirectly provides
information about the languages spoken during the period under examination,
suggesting a sort of awareness of linguistic diversity.
Various terms have been proposed to describe this linguistic repertoire of
ancient Japan where, rather than a speech community, it is the emergence of a
script community centred on the use of written Sinitic that is relevant.13 One
convincing framework is the ‘Sinographic cosmopolis’, originally elaborated by
Ross King in his study of Sinitic and vernacular Korean,14 which may also be
fruitfully applied to Japan, Vietnam and other peripheries in the Sinic world.
The concept highlights the ‘supraregional dimension’ (‘cosmo-’) of this sphere,
with a focus on the political dimension and the common aesthetics of political
culture (‘-polis’). The term ‘Sinographic’ is used to emphasise the graphological

||
11 Western Old Japanese is the language of songs in the Kojiki and the Nihon shoki 日本書紀
(‘Chronicle of Japan’, 720), as well as of almost any book of the Man’yōshū 万葉集 (‘Collection
of Myriad Leaves’, post 759). Eastern Old Japanese is attested in Books 14 and 20 of the
Man’yōshū and in some poems of the Hitachi no kuni fudoki 常陸国風土記 (‘Record of the
Province of Hitachi and its Customs’, c. 713–718), while sparse attestations of the Kyūshū dia-
lect are found in the fudoki (geographical records) of the Saikaidō region. For Western Old
Japanese, see Vovin 2020; for Eastern Old Japanese, see Vovin and Ishisaki-Vovin 2021; for the
old dialect of Kyūshū, see Manieri 2022b.
12 For a detailed and updated introduction to the manuscript and its complex history, see
Whitman 2022. The text is reproduced in Nakada Norio 1969. For a translation of the passage
and an analysis in the wider context of the construction of imperial power, refer to Duthie 2014,
209–211. Terada Akira argues that the passage shows no awareness of the ‘difference among
the languages’, since both the language group outside the archipelago and that within are
referred to as hōgen 方言 (‘dialects’) (Terada Akira 2009, 167–168). However, in my view, the
fact that the text distinguishes two different language groups and statutes – ‘the language of
this land’ in relation to ‘foreign’ languages – indicates the opposite.
13 Cf. ‘hieroglossia’, discussed by Robert 2006, and ‘Sinoglossia’, by Saitô Mareshi 2021. Cf.
also the concept of ‘brush talk’ in Denecke 2014. For a detailed review of other terms, see Kor-
nicki 2018, 33–41.
14 King 2014. See also Handel 2019 and Kin Bunkyō 2021.
 Vernacular Terms in Sinitic Texts | 173

and scriptological dimension that distinguishes this sphere from the mega-
regions studied by Sheldon Pollock in his ‘Sanskrit cosmopolis’.15
This diverse array of texts often incorporates multiple languages. It is not
uncommon to encounter texts written in Sinitic that use phonograms to indicate
vernacular pronunciation for the proper names of people or places. Further-
more, some works in Sinitic, such as diaries, annals and fiction, include poems,
edicts and prayers in vernacular. In some cases, such as in reports from the
provinces, even local variants of certain words are documented.16 Finally, texts
in Sinitic with lexical and/or grammatical glosses in vernacular are not rare in
the eighth century.
The focus of this paper, however, is not the attestation of local varieties or
Japanese poetic diction in texts in Sinitic, but rather, the occurrence of vernacu-
lar terms in bureaucratic and practical texts in Sinitic, whose models are more
or less standardised or even defined by rule.
The Yōrōryō 養老令 (‘Administrative Code of the Yōrō Era’, 718), for exam-
ple, which is the official code of the eighth century, includes the Kūjikiryō 公式
令 (‘Law on Official Documents’), where twenty-two out of eighty-nine articles
are specifically devoted to establishing the formats of official documents, such
as the edict, petition, notification, appointment, register and pass, providing
information on their opening and closing formulas, the character of the main
text and the procedure to follow when compiling them.17 All the text forms are in
Sinitic, with the sole exception of the edict (shō 詔). Other kinds of documents,
however, even ones not possessing a format determined by law, follow a tradi-
tion of procedures and precedents, as shown by the wide availability of practi-
cal texts in Japanese archives and later legislation.

||
15 King 2014, 6. On the Sanskrit cosmopolis, see Pollock 2006.
16 See, for example, the occurrences of the Kyūshū dialect of Western Old Japanese in records
from the provinces of the Saikaidō region (Manieri 2022b, 40–45).
17 The text of the law has been preserved in two commentaries from the ninth century, name-
ly, the Ryō no gige 令義解(‘Official Commentary on Administrative Laws’, 833) and the Ryō no
shūge 令集解 (‘Collection of Commentaries on Administrative Laws’, compiled between 859
and 877). The reconstructed text of the Kūjikiryō 公式令 can be found in Inoue Mitsusada et al. 1976,
365–406. For translations and commentaries into European languages, see Dettmer 2010, 329–415
and Migliore 2011. An introduction to the law is in Migliore 2018, 183.
174 | Antonio Manieri

3 Multilingualism in practical texts

This paper focuses on instances of multilingualism in practical texts on wooden
tablets or paper.
Wooden tablets are called mokkan 木簡 in modern Japanese and they are,
by a simple definition, documents written down by the use of ink on wooden
slips, made of Japanese cypress or cedar, of varying size, usually not very big.18
Most tablets date back from the late seventh to the eighth century and originate
from Nara, the ancient capital, but a significant number also come from the
surrounding regions of other former capitals such as Fujiwara, Nagaoka and
Heian, from the headquarters of certain provinces and from Buddhist monaster-
ies, that were, in any case, areas of power with a more advanced level of literacy.
Japanese scholars at the Nara National Research Institute for Cultural Prop-
erties have elaborated a taxonomy in fifteen types of surviving tablets according
to shape, showing a diversity that does not emerge in their Chinese counter-
parts.19 In terms of their contents, they can be classified into three major types:
labels or tags, which were attached to tax goods; documents in the strict sense;
and slips for calligraphy practice and learning.20
Regarding the paper documents, this study refers to some Shōsōin monjo 正
倉院文書 (lit. ‘Shōsōin documents’), which are the documents contained in
around 660 scrolls preserved in the Shōsōin 正倉院, the repository of Tōdai
Monastery 東大寺 in Nara. The Shōsōin consists of three sections: the North
Section contains artefacts connected with the sovereign Shōmu 聖武 (701–756,
r. 724–749), donated by his consort Kōmyō 光明 (701–760) in 756, forty-nine
days after his demise; the Middle Section, documents and objects related to the
Office for the Construction of the Tōdai Monastery (Zō Tōdaiji shi 造東大寺司);
and the South Section, Buddhist and ritual implements used at the Tōdai Mon-
astery.21 The documents of the Office for the Construction of the Tōdai Monas-

||
18 Tōno Haruyuki 1983, 4. The initial significant collection of wooden tablets was excavated at the
Heijō Palace site in Nara in 1961. Subsequently, more than 150,000 slips have been uncovered
across approximately five hundred sites spanning from northern Honshū to Kyūshū. For a general
introduction to mokkan, see Tōno Haruyuki 1977; Tōno Haruyuki 1983; Tōno Haruyuki 2005; and
Satō Makoto 1997. For some remarks in English, see Piggott 1990; Farris 1998; Manieri 2020.
19 Farris 1998, 201.
20 A corpus of about 50,000 tablets is freely available online at the Mokkanko – Wooden Tab-
let Database (<https://mokkanko.nabunken.go.jp/>, accessed 25 June 2023) by the Nara Na-
tional Research Institute for Cultural Properties.
21 After being maintained by Tōdai Monastery for over a thousand years, and under the re-
sponsibility of the Imperial Household Agency (former Imperial Household Ministry) since
 Vernacular Terms in Sinitic Texts | 175

tery were actually inscribed on the reverse side of discarded documents that
were originally created by other offices in the capital or the provinces. Conse-
quently, this practice of utilising them as ledger paper ensures the preservation
of various distinct documents.22
The practical texts, which are the focus of this article, are ‘service’ texts that
were entrusted to writing supports that – not being intended for preservation,
such as wooden tablets or loose sheets of paper – were more likely to be lost.
Those that do remain available to us have been preserved for often accidental
reasons, such as being reused on the back for transcribing other documents.
The use of these texts associated with temporary material support offers a
glimpse into a form of writing that maintains a sense of naturalness. Moreover,
it sheds light on the practice of multilingualism, even within contexts character-
ised by limited institutional formality.
Although these sources have been widely studied for their contents, histori-
cal reconstruction or document genres, scant attention has been paid to their
linguistic features, and particularly to the co-presence of Sinitic syntax and
specialised terms in Japanese vernacular. Thus, insights into technical and
artisanal terminology contribute to the ongoing and enthusiastic debate sur-
rounding the linguistic repertoire of seventh- and eighth-century Japan.

4 Attestations of vernacular terms in practical

texts in Sinitic
In this section, I will provide five different sources in Sinitic containing special-
ised terms in vernacular Japanese. I will introduce each source in terms of its
philological features and textual genre; I will then focus on the terms, eliciting
their scriptural features and meaning by consulting ancient Japanese dictionar-

||
1884, as of 31 March 2010, the catalogued collection held 8,932 items. The details of the items
belonging to Shōmu are recorded in the Kokka chinpōchō 国家珍宝帳 (‘Record of Rare Treas-
ures of the Country’, 756). For a general introduction to the Shōsōin, see Hayashi Ryoichi 1975,
in particular on the Shōsōin monjo, see Tōno Haruyuki 1977; Tōno Haruyuki 2005.
22 The digitised version of the documents is available at the database of the Shōsōin official web-
site (<https://shosoin.kunaicho.go.jp/>, accessed 25 June 2023). Most of Shōsōin monjo have been
published as Dai Nihon Komonjo 大日本古文書 (‘Old documents of Japan’) (1901–1940), also avail-
able at the open access databases of the Historiographical Institute of the University of Tokyo
(<https://wwwap.hi.u-tokyo.ac.jp/ships/>, accessed 25 June 2023).
176 | Antonio Manieri

ies or, when no lexicographic information is available, by linguistic reconstruc-

tion with reference to later sources or lexical data.
I will mainly consult the Wamyōruijushō 和名類聚抄 (‘Classified Notes on
Japanese Nouns’, c. 930), compiled by the scholar-official Minamoto no Shitagō
源順 (911–988). It is a bilingual Sinitic-Japanese dictionary, and though it dates
to c. 930 CE – thus, much later than the period of compilation of the texts under
investigation – it is an essential tool because it is organised into categories and
composed of quotations from previous sources, including eighth-century dic-
tionaries, as we will see more thoroughly below. The work survives in a twenty-
book version (nijikkanbon 廿巻本) and a ten-book version (jikkanbon 十巻本)
that differ in their respective number of books and internal arrangement; there
is not always complete correspondence between the lemmas of the two ver-
sions, and it is difficult to determine which of the two is the oldest or most faith-
ful to the author’s intentions.23
I will also occasionally refer to other dictionaries, such as the Shinsen jikyō
新撰字鏡 (‘Mirror of Characters, Newly Edited’, c. 898–901) by the monk Shōjū
昌住 (?–?), which is a character dictionary organised by character radical – but
also includes an appendix where words are listed by semantic category – and to
the later Ruijumyōgishō 類聚名義抄 (‘Notes on Classified Meanings’, late elev-
enth century), which is also arranged by radical.24

4.1 Source A: A bay horse in mokkan 1926

The first example comes from the mokkan numbered 1926, unearthed at excava-
tion SD1900 at the Heijōkyū Suzakumon site in Nara (Sakichō), on the Shimo-

||
23 There are several testimonies of the two versions in the form of both manuscripts and print
editions. All of the manuscripts are either incomplete or have lacunas, but the two versions
have vulgates in the form of printed editions: for the twenty-book version, the Genna sannen
kokatsujiban nijikkanbon 元和三年古活版廿巻本 (‘Movable-Type Edition in Twenty Books of
the Third Year of Genna Era’), an edition printed with movable type and published in 1617 by
the Confucian scholar Naba Kassho 那波活所 (1595–1648); and, for the ten-book version, the
Senchū Wamyōruijushō 箋注倭名類聚抄 (‘Annotated Commentary on the Wamyōruijushō’), the
edition annotated by Kariya Ekisai 狩谷棭斎 (1775–1835), completed in 1823 but published only
in 1883. In this paper, I will mainly consult the Genna sannen kokatsujiban nijikkanbon, for
which I follow Nakada Norio 1978. Other testimonies are in Kyōto daigaku bungakubu kokugo
kokubungaku kenkyūshitsu 1968. For a general introduction to the Wamyōruijushō, see Lin
Zhongpeng 2002.
24 For the Shinsen jikyō, I follow Kyōto daigaku bungakubu kogaku bungaku kenkyūshitsu
1967; for the Ruijumyōgishō, I follow Tenri daigaku fuzoku Tenri toshokan 2018.
 Vernacular Terms in Sinitic Texts | 177

tsumichi, one of the three main roads that ran north-south through the ancient
Yamato plain, and which extended northward from the Fujiwara capital leading
to Nara. It is a rectangular plate measuring 656 × 36 × 10 mm.25 It is written on
both the recto (one line) and the verso (two lines). Around seventy-three charac-
ters are readable, while one character may be delineated as missing.
The text of the mokkan reads as follows:26

関々司前解 近江国蒲生郡阿伎里人大初上阿□勝足石許田作人
大宅女右二人左京小治町大初上笠阿曾彌安戸人右二
同伊刀古麻呂
送行乎我都 鹿毛牡馬歳七 里長尾治都留伎

A tentative translation is as follows:

Petition to the official of frontiers

Aki nǝ Suŋguritaruiwa, upper great initial rank, a man from the village of A[ki], district of
Kamapu, province of Apumi, allowed the tillers of rice fields.
The tillers of rice fields [are] the two men Itokomarǝ and Opoyakɛme. The two aforemen-
tioned [are] men of the residence unit of Kasa nǝ Asǝmi Yasu, upper great initial rank, of
Woparimati, in the Left Capital.
Accompanied by Wokatu. Male bay horse, seven years old.
– Village Chief Wopari nǝ Turuŋgi

The mokkan is a permit issued by the village chief Ohari no Tsurugi (Wopari nǝ
Turuŋgi) to allow two men, Itokomaro (Itokomarǝ) and Ōyakeme (Opoyakɛme),
members of the residence unit (ko 戸) of Kasa no Yasu in Oharimachi (Wopari-
mati), to return to the capital from the village of Aki. They were working as rice
field tillers for Aki no Suguritaruiwa (Aki nǝ Suŋguritaruiwa) in the village of
Aki, district of Kamō (Kamapu), in the province of Ōmi (Apumi).
The mokkan does not contain a date, but several hints suggest that it may have
been issued in the early eighth century. Firstly, the institution of the village chief
was regulated by the Taihō ritsuryō 大宝律令 (‘Penal and Administrative Codes of
the Taihō Era’) of 701, but the character 里 was used for ‘village’ until 715 and then
again after 740. Between these two years, the more complex gōrisei (郷里制) sys-
tem came into effect, in which the character 郷 (Ch. xiang, Sino-Jp. gō) was used for
‘village’, while 里 (Ch. li, Sino-Jp. ri) was used to indicate its subunits. Secondly, in

||
25 For the excavation report, see Nara kokuritsu bunkazai kenkyūjo 1974b. A photographic repro-
duction is found in Nara kokuritsu bunkazai kenkyūjo 1974a, pl. 1. See also Kiyota Yoshiki 1980.
26 All translations are my own unless otherwise noted. In the transcriptions of the texts and
the translations, round brackets indicate portions of text printed in characters of smaller size in
the original; square brackets in translations indicate my insertions to facilitate understanding.
178 | Antonio Manieri

715, the use of wood and bamboo for transit permits was prohibited, and they be-
gan to be produced on paper instead.27 Thirdly, some textual features, such as 前,
were present in the seventh-century mokkan, but became less common in the
eighth century,28 showing that the mokkan was inscribed in a transitional period,
possibly the early eighth century. Finally, several places in Fujiwara, the capital
from 694 to 710, are called by proper nouns, such as Oharimachi, while there are
no such cases for Nara, the capital from 710.29 Therefore, we can assume that this
mokkan was created between 701, the year of the promulgation of the Taihō
ritsuryō, and 710, the year of the capital’s transfer from Fujiwara to Nara.
The fact that the mokkan was unearthed near Nara is due to the likelihood
that the travellers using the ancient pass crossed Yamashiro Province via the
Tōkaidō road to enter Yamato Province, where the pass was abandoned as it
was no longer necessary. The two men were accompanied by another man
named Okatsu (Wokatu), likely a serf, and a seven-year-old male bay horse.
The ‘Law on Official Documents’ of the Yōrōryō of 718 contains an article on
the form of the transit pass, probably also present in the Taihō ritsuryō of 701,
since the Ryō no shūge 令集解 (‘Collection of Commentaries on Administrative
Laws’, compiled between 859 and 877), in the section regarding the law, con-
tains some quotations from the Koki 古記 (‘Ancient Records’, early eighth cen-
tury), which was a commentary to the code of 701. The article reads as follows:

過所式
其事云云。度某関往其国。
某官位姓。(三位以上。称卿。)資人。位姓名。(年若干。若庶人称本属。)従人。某国某
郡某里人姓名年。(奴名年。婢名年。)其物若干。其毛牡牝馬牛若干疋頭。
年 月 日 主 典 位 姓 名
次 官 位 姓 名
右過所式。並令依式具録二通。申送所司。々々勘同。即依式署。一通留為案。一通判
給。30

Transit Pass: Form.

Purpose of travel. Crossing such barriers and reaching such province.
The official, rank, honorific title (if holder of a third rank or higher: family name and lord).
Assistants: Rank, honorific title, name (age. If a commoner, [the family register] he be-
longs to shall be indicated). Accompanying people: Province, district, village, honorific ti-

||
27 Kiyota Yoshiki 1980. For the prohibition on using bamboo and wooden tablets for transit
passes, see Shoku Nihongi, Book 6, Reiki 1.5.1 (Aoki Kazuo et al. 1989, 224–227); Ryō no shūge,
Book 34 (Kyūjikiryō), ‘Tenshi shinji jō’ (Kuroita Katsumi 1955, 852).
28 Nara bunkazai kenkyūjo 2017, 1.
29 Nara bunkazai kenkyūjo 2017, 2.
30 Inoue Mitsusada et al. 1976, 389.
 Vernacular Terms in Sinitic Texts | 179

tle, name, age (male slaves: name and age; female slaves: name and age). Objects and
their quantities. Number of heads of horses and cattle, their coat colour and sex.
Year, month, day. Official of the fourth class, rank, honorific title, name.
Vice director: rank, honorific title, name.
The aforementioned transit pass forms shall be prepared in two copies according to this
format and sent to the relevant office. The relevant office shall carry out the checks. Then,
the copies shall be signed according to the format. One copy shall be filed [in the relevant
office], and the other shall be delivered [to the interested party].31

The transit pass on mokkan 1926 deviates from the law article in some respects,
but it does not fail to record the names of the travellers, their origin, purpose,
the accompanying people and the heads of horses, along with their coat colour
and sex.
In the mokkan, the horse’s coat colour, which we have translated as ‘bay’, is
expressed by the two characters 鹿毛, respectively meaning ‘deer’ and ‘coat’.
This compound seems to refer to a colour similar to the ‘coat of a deer’, but it is
not attested in Chinese sources as a term for an equine coat. Additionally, the
two characters are also often used as phonograms of the kungana 訓仮名 type
for ka and kɛ.
The Wamyōruijushō, Section 16 ‘Bovines and Horses’ (牛馬部), Subsection
149 ‘Bovine and Equine Coats’ (牛馬毛) has the largest early organised nomen-
clature of equine and bovine coat colours. Among the various lemmas, we find
the compound 鹿毛 in the following entry:

Wamyōruijushō, Book 11, 16/149

騮馬[…] 毛詩注云騮(音留漢語抄云騮馬鹿毛也[…])赤身黒𩮊馬也[…]32

Bay horse. The Maoshizhu states [that] bay 騮 (the sound is [that of the character] ljuw 留;
the Kangoshō states [that] a bay 騮 horse is a kakɛ horse […]) is a horse with a brown body
and black hair.

The entry explains that the Kangoshō 漢語抄 (‘Notes on Chinese Words’), which
is a lost bilingual Sinitic-Japanese dictionary from the eighth century, furnishes
the compound 鹿毛 as an equivalent of the Sinitic 騮, which is defined by a
quotation from the Chinese source Maoshizhu 毛詩注 (‘Mao’s Commentary on
the Classic of Poetry’, pre-221 BCE) as a horse with a reddish-brown body and
black extremities (called ‘bay’ in English). In the ‘Bovine and Equine Coats’
section of the Wamyōruijushō, there are many examples of vernacular equiva-

||
31 A translation into Italian can be found in Migliore 2011, 81; into German, in Dettmer 2010,
370–371.
32 Nakada Norio 1978, 128.
180 | Antonio Manieri

lents with the morphological structure ‘name + kɛ 毛’, none of which is attested
in the varied Sinitic nomenclature. Finally, these vernacular terms with phono-
grams that also have a semantic value do not occur in contemporary poetry,
such as in the poetic anthology Man’yōshū 万葉集 (‘Collection of Myriads of
Leaves’, post-759), where the horse-coat colours are only basic terms, such as
aka ‘red’, kuro ‘black’ and awo ‘white’.33

4.2 Source B: Unusable horses in the Suō no kuni shōzeichō

The second text I would like to present is a passage from the Suō no kuni
shōzeichō 周防国正税帳 (‘Register of Correct Taxes for the Province of Suō’).
The shōzeichō 正税帳 was the register that each province compiled to rec-
ord the amount of annual regular taxes collected, expenditures borne in the
previous year and balances. The form of this register has varied, but without
straying significantly from the model later recorded in Book 27 of the Engishiki
延喜式 (‘Procedures of the Engi Era’, 927) related to the Bureau of Public Re-
sources (Shuzeiryō 主税寮).34 About twenty-five shōzeichō from the Tenpyō era
(729–749) have been preserved among the documents of the Shōsōin, some
being more or less complete, others fragmentary.
Shōzeichō were prepared in three copies: one copy was kept at the provin-
cial administration headquarters as a reference for subsequent governors and
the drafting of the following year’s document; two copies were submitted to the
Great Council of State (Daijōkan 太政官) by the end of the second month,35 for-
warded by provincial officials called shōzeichōshi 正税帳使 or shōzeishi 正税帳
使 (lit. ‘messengers of the registers of regular taxes’). The shōzeichō were in-
spected for errors or irregularities in each category at the Bureau of Public Re-
sources of the Ministry of Popular Affairs (Minbushō 民部省), and, if any were
found, they were returned, and adjustments would be ordered. The entries in
the register are, therefore, related to quantities of rice, millet or salt, expenses
for repairing weapons or transportation expenses, and it is also possible to
come across expenses related to livestock, particularly post and relay horses.36

||
33 For other attestations of ‘name + 毛’ types of coat colours in ancient Japanese sources, see
Manieri 2012.
34 Kuroita Katsumi 1938, 671–685. As for the Engishiki see Bock 1970; Bock 1972; Bock 1985.
35 The provinces of the Saikaidō (present-day Kyūshū) were required to send them to the
Dazaifu before the thirtieth day of the second month, and, once checked, to send them to the
Council of State before the thirtieth day of the fifth month.
36 On the shōzeichō, see Inoue Tatsuo 1967; Hérail 1966, particularly 106–107.
 Vernacular Terms in Sinitic Texts | 181

The extant registers held at the Shōsōin also include that of the province of
Suō, a territory corresponding to the eastern part of present-day Yamaguchi
Prefecture. The Suō no kuni shōzeichō dates to 738.37 The document survives on
15 sheets that are part of manuscript scroll Seishū 正集 35 (paper, 26.2 × 617.2
cm, consisting of 18 glued sheets in total) and Seishū 36 (paper, 26.6 × 556 cm,
consisting of 15 glued sheets in total).38 These sheets have survived because the
Office of Sutra Transcription (Shakyōshi 写経司) used their reverse sides as
second-hand paper for several types of registers and documents from the eight-
eenth (746) to the twentieth (748) year of the Tenpyō era. The Suō no kuni
shōzeichō from 738 occupies fols 12–18 of Seishū 35 and 1–8 of Seishū 36,39
which are marked with the province’s stamp. The document is incomplete, but
it provides various pieces of information regarding the province’s expenditures
and rice reserves. It also includes a brief section on horses that are no longer
usable (fuyō uma 不用馬), which reads as follows:

不用馬陸匹 (一匹天平三年買、歯七、経傳八歳、左前足宇弖。二匹天平五年買、歯六、
経傳五歳、左後足多利。一匹天平四年買、歯七、経傳八歳、右前足宇弖。一匹天平六年
買、歯四、経傳五歳、左前足宇弖。一匹天平六年買、歯五、経傳五歳、右前足宇弖。)

Unusable horses: six. (One horse purchased in the third year of the Tenpyō era, seven
years old, eight years old according to the manual, with a bruise on the left front leg. Two
horses purchased in the fifth year of the Tenpyō era, six years old, five years old according
to the manual, with lameness on the left hind leg. One horse purchased in the fourth year
of the Tenpyō era, seven years old, eight years old according to the manual, with a bruise
on the right hind leg. One horse purchased in the sixth year of the Tenpyō era, four years
old, five years old according to the manual, with a bruise on the left front leg. One horse
purchased in the sixth year of the Tenpyō era, five years old, also five years old according
to the manual, with a bruise on the right front leg.)

Other shōzeichō – such as that of the province of Echizen in 733, of Owari in 734,
and of the administration of the eastern part of the capital (Sakyōshiki) in 738 –
also give an indication of the ‘unusable horses’, though recording only the

||
37 On various questions about the Suō no kuni shōzeichō, see Inoue Tatsuo 1967, 247–294.
38 The manuscripts Seishū 35 and 36 are both held at the Shōsōin, Section Centre, 15. Their
digitised versions are at the following URLs, respectively: <https://tinyurl.com/4278fmh8>;
<https://tinyurl.com/kzywvtc3> (accessed on 25 June 2023). A diplomatic edition of the docu-
ment is in DNK, vol. 2, 130–146.
39 In Seishū 35, the height of the sheets in is 26.6 the length of each sheet, from 12 to 18, is
14.2 + 50.8 + 50.8 + 56.2 + 56.0 + 56.1 + 29.3. In Seishū 36, the height of the sheets is 26.6 the
length of each sheet, from 1 to 8, is 29.9 + 40.6 + 55.3 + 56.2 + 56.3 + 5.2 + 17.5 + 28.
182 | Antonio Manieri

number of heads.40 The Suō shōzeichō under examination provides additional

information, including the year of acquisition, age (expressed by the character
歯, lit. ‘tooth’, as usual in the technical field of hippology) and, notably, the
reasons for the animals’ lack of utility.
Specifically, all six heads of cattle are reported to have a leg issue, which in
two cases is attributed to a condition known as tari 多利, and in the remaining
four cases, to a condition called ute 宇弖. A peculiarity of the text is the fact that
although the document is written in Sinitic, the names of the two diseases are
conveyed by the phonograms 多利 and 宇弖. The term ute has no other attesta-
tions in sources. It could have the same root as the verb utsu 打つ [Old Jp. utu],
meaning ‘to strike, to hit’, and it could, thus, indicate a contusion (bruise) oc-
curring when a blow strikes part of the body. The term tari is, etymologically,
the converb form of the intransitive quadrigrade verb taru 垂る, which means
‘to droop’ or ‘to sag’, and could refer to the formation of fleshy warts on a part of
the body that droop and sag. The term is also attested in several ancient dic-
tionaries. In the Shinsen jikyō, it occurs in the following lemma in Section 27
‘Radical Sickness’ (疒部):

Shinsen jikyō, 27
疣 三形同有流反平腫也。伊比保又太利又比志比子。41

Lameness. It is the same as the word ‘three-formed’ 三形. [The sound is that of hjuw, like
the initial of the character] hjuwX 有 plus [the final of the character] ljuw 流. It is an ordi-
nary ‘swelling’ 腫. [Vernacular terms:] ihiho, and tari, and hisihisi.

A precise definition is found in the Wamyōruijushō, Section 16 ‘Bovines and

Horses’, Subsection 151 ‘Bovine and Equine Maladies’ (牛馬病):

Wamyōruijushō, Book 11, 16/151

驇 唐韻云驇(陟利反、興到同、俗云驇多利。)馬脚屈重也。42

Lameness. The Tangyun states that lameness 驇 (the sound is that of [the initial of the
character] trik 陟 plus [the final of the character] lijH 利, the same [as that of the character]
trijH 致; the vernacular of 驇 is tari) is [the malady in which] the horse’s leg bends and
trudges.

||
40 The shōzeichō of Echizen Province, dating to the fifth year of the Tenpyō era (733), and that
of Owari Province, dating to the sixth year of the Tenpyō era, are in DNK, vol. 1, 461–469 and
607–622, respectively, while that of Sakyōshiki, dating to the tenth year of the Tenpyō era, is in
DNK, vol. 2, 106.
41 Kyōto daigaku bungakubu kogaku bungaku kenkyūshitsu 1967, 44.
42 Nakada Norio 1978, 130.
 Vernacular Terms in Sinitic Texts | 183

The later Ruijumyōgishō, Section ‘Clergy/Part 2’ (僧中), Subsection 102, reads as

follows:

Ruijumyōgishō, Book Clergy/Part 2, 102

驇 今正音至馬重皃音致 タリ ナツム ナへク ツマツク43

Lameness. Present-day correct sound is the same as [that of the character] tsyijH 至. Trou-
blesome appearance of a horse. [Vernacular terms:] tari, natumu, naheku, tumatuku.

Tari is, therefore, the vernacular equivalent of 驇, a pathology of the horse’s leg
most likely identifiable with what is known as ‘lameness’, a disease in which a
deformed hoof curves backwards, generating a defect in the flatness of the
horse’s foot, due to which the animal proceeds by resting predominantly on the
toes of the hooves.
In conclusion, ancient dictionaries attest that tari is the vernacular equivalent
of the Sinitic veterinary term 驇 or 疣. For the purpose of this analysis, it is interest-
ing to note that the Wamyōruijushō specifically mentions zoku 俗 in its twenty-book
version (but zokujin 俗人, ‘common people’ in its ten-book version). Within the
Wamyōruijushō, Japanese equivalents are marked with wamyō 和名 and zoku (or
alternatively zokujin or zokugo 俗語) where there is no attestation in previous dic-
tionaries. As Tsukishima Hiroshi has pointed out, wamyō refers to a term for which
the compiler Shitagō recognises a previous attestation, while zoku is used when he
does not recognise it and takes it from spoken language.44

4.3 Source C: Struts and beams in the Zō Ishiyamain shō

yōdochō
The third text is a passage from the Zō Ishiyamain sho yōdochō 造石山院所用度
帳 (‘Register of Expenses of the Institute for the Construction of the Ishiyama
Monastery’), which is a budget document listing necessary supplies.
The monastery was originally built in 747 at the request of sovereign
Shōmu, when the monk Rōben 良弁 (689–773), founder of Tōdai Monastery,
enshrined Nyoirin Kannon 如意輪観音.45 Later, from 761, the expansion of the

||
43 Tenri daigaku fuzoku Tenri toshokan 2018, vol. 2, 257.
44 Tsukishima Hiroshi 1963, 57.
45 Established as a temple of the Kegon lineage, its foundation and history is told in the
Ishiyamadera engi emaki 石山寺縁起絵巻 (‘Illustrated Scroll of the Origins of Ishiyama Monas-
tery’, first version: c. 1325; enlarged version: 1805) (Aizawa Masahiko and Kuniga Yumiko 2016).
184 | Antonio Manieri

halls and the maintenance of the temple complex were promoted as a state
project and carried out under the government office known as the Institute for
the Construction of the Ishiyama Monastery (Zō Ishiyamadera sho 造石山寺所),
to which Buddhist monks and other staff members were dispatched from the
Office for the Construction of the Tōdai Monastery.
Several documents preserved at the Shōsōin date to 761 (i.e. the sixth year
of the Tenpyō hōji era) are related to works of expansion and maintenance,
including the Zō Ishiyamain sho yōdochō under examination. The document
survives on 14 sheets: 19v–7v of the manuscript scroll Zokuzokushū 続々集 38.9
(paper, 29.1 × 921 cm, consisting of 19 glued sheets) and 9v of Zokuzokushū 43.9
(paper, 29.0 × 287 cm, consisting of 9 glued sheets).46
The document provides various pieces of information regarding the supplies
necessary for the expansion of the complex, including a section on the coating to
apply to the double-level roof of the pulpit (kōza 高座), which reads as follows:

三斗六升高座蓋二覆塗料
一斗五升蓋二覆𡑮料
九升一蓋三重𡑮料(重別三升)
六升一蓋二重𡑮料(重別三升)
九升二蓋裏於一度土漆料(蓋別四升五合)
五升二蓋枚桁丸桁垂木并多々理形及波佐目等《塗二度土》塗二度土漆
料(蓋別二升五合)
七升二蓋枚桁丸桁垂木多々理波佐目等塗三度墨漆料(蓋別三升五合)47

Here is a translation of the passage:

3 to and 6 shō as the quantity of coating to apply to the two levels of the roof of the pulpit.
1 to and 5 shō as the quantity of dry lacquering to apply the two levels of the roof.
9 shō as the quantity for three layers of dry lacquering for one level of the roof (three shō
for each layer).
6 shō as the quantity for two layers of dry lacquering for one level of the roof (3 shō for
each layer).

||
46 The manuscripts Zokuzokushū 38.9 and Zokuzokushū 43.9 are both held at the Shōsōin,
Section Centre, 20. In Zokuzokushū 38.9, the length of each sheet from 19v to 7v is 56 + 57 +55 +
56 + 58 + 56 + 38 + 44 + 26 +56 + 57 +57 + 57; in Zokuzokushū 43.9, the length of sheet 9v is 45 cm.
Their digitised versions are, respectively, at the following URLs: <https://tinyurl.com/
yzvjsn5p>; <https://tinyurl.com/6esptbtv> (accessed on 25 June 2023). A diplomatic edition of
the document is in DNK, vol. 16, 263–274. Another document with the same name, but dated to
the twelfth (intercalary) month of the sixth year of the Tenpyō hōji era (761), is preserved in
Zokuzokushū 45.5, 1v–5v; 45.6, 5r; 45.7, 1r–3r, 1v–3v. The diplomatic edition of this second docu-
ment is in DNK, vol. 16, 219–252.
47 Zokuzokushū 38.9, 11v.
 Vernacular Terms in Sinitic Texts | 185

9 shō as the quantity of clay lacquering to apply once each to the two roof insides (4 shō
and 5 gō for each inside).
5 shō as the quantity of clay lacquering to apply twice to flat beams, round beams, rafters,
struts, pasamɛ, etc. of the two roofs (2 shō and 5 gō for each roof).
7 shō as the quantity of black lacquering to apply three times to flat beams, round beams,
rafters, struts, and pasamɛ of the two roofs.

In the passage, quantities are expressed by the system of measurement in to 斗,

shō 升 and gō 合,48 and refer to lacquering to apply to each part of the cover of
the pulpit. Therefore, the recorded specialised terms are related to lacquering,
such as ‘dry lacquering’ 𡑮 and ‘black lacquering’ 墨漆,49 and to carpentry and
building construction, such as ‘flat beams’ 枚桁, ‘round beams’ 丸桁 and ‘raft-
ers’ 垂木.50 Among these terms, which are all written in Sinitic, we also find two
words indicating other parts of the construction framework spelled in phono-
grams: tatarikata 多々理形, which, as we see below, is the ‘strut’, and pasamɛ
波佐目, a term that is not yet clear. The Wamyōruijushō, Book 10, Section 13
‘Residences’ (居処部), Subsection 137 ‘Parts of the Dwelling’ (居宅具), records
the vernacular term tatarikata:

Wamyōruijushō, Book 10, 13/137

栭 爾雅注云梁上謂之栭(音而文選師説多々利加太)欂櫨也。説文云欂櫨[薄盧二音]柱上
枅也51

Strut. Eryazhu states [that what is] on the beams is called a ‘strut’ 栭 (the sound is that of
nyi 而; the master’s explanation to the Wenxuan [states it is] tatarikata), and [it] is the
bracket [supporting the beams]. The Shuowen states [that] the bracket (the sound is that of
bak lu 薄盧) is the capital at the top of the columns.

Thus, the vernacular tatarikata is furnished as the equivalent of the term 栭

(‘strut’), which is explained as 欂櫨 (‘bracket’); this, in turn, is explained – with
reference to the Shuowen jiezi 説文解字 (‘Explaining the Graphs and Unravelling
the Characters’, 100 CE) – as 柱上枅, the ‘capital’, the square wooden block at

||
48 According to Article 1 of the Zōryō 雑令 (‘Miscellaneous Law’) of the Yōrōryō, 1 to is equiva-
lent to 10 shō, and 1 shō, in turn, is equivalent to 10 gō (Inoue Mitsusada et al. 1976, 475; Dett-
mer 2010, 536). In the present-day usage of these traditional units, 1 gō is equivalent to 0.18 litres.
49 Dry lacquering is a lacquer craft technique in which a lacquer-coated linen cloth is applied
to the bare surface, allowed to dry, and then coated over and over again.
50 A ‘flat beam’ 枚桁 (hirageta) is a light, horizontal, rectangular structure that is inserted into
the corner-post between the top and bottom rails of a balustrade; a ‘round beam’ 丸桁 (maro-
geta) is a circular beam that supports the rafters; and a ‘rafter’ 垂木 (taruki) is the simplest type
of structure, extending from the ridge to the end of or beyond the eave.
51 Nakada Norio 1978, 118.
186 | Antonio Manieri

the top of the column. In the Wamyōruijushō, the word is indicated by the onga-
na 音仮名 phonograms 多々利加太, while in the attestation of the word in the
Zō Ishiyamain shō yōdochō, the last two syllables, kata, are rendered by the
disyllabic kungana 形. The source that the compiler Shitagō uses is marked by
the expression 師説, meaning ‘master’s explanation/comment’, and refer to the
explanations the master provides while commenting on a certain text.
Unfortunately, ancient dictionaries do not record the vernacular term
pasamɛ 波佐目, and it is not clear what part of the construction it indicates. It
could be related to the verb pasamu (Modern Jp. hasamu 挟む or 挿む), meaning
‘to insert’, ‘to put between’, and which also has a bigrade conjugation. Thus,
pasamɛ could refer to an ‘insert’ to be put between the beams. In any case, it is
worth noting that this additional word, so specific as not to be recorded else-
where, is recorded in the vernacular, using only phonograms.
These two terms related to carpentry are not the only ones recorded in
Shōsōin documents in vernacular, as we can infer from the list of architecture
terminology collected by Fukuyama Toshio, where eight of the eighty-five terms
enlisted are written in phonograms,52 showing a trend of making use of vernacu-
lar terms in the domain of carpentry and architecture.

4.4 Source D: Things to arrange in the Shasho zōyōchō

The text in this subsection is a Shasho zōyōchō 写書雑用帳 (lit. ‘Register of Mis-
cellaneous Expenses for Copying Books’). It is a document on paper preserved
at the Shōsōin, inscribed on the verso of the manuscript scroll Zokushū 続集 16
(paper, 29.1 × 397.3 cm, consisting of 10 glued sheets), in particular on sheets 2v
and 3v, respectively 45 and 27.7 cm long. The document is cut off at the begin-
ning and the end.53
This budget document dates to the fourth day, seventh month, second year
of the Tenpyō era, which is 730 CE. Its compiler is Takaya no Akamaro 高屋赤麻
呂 (fl. 730), an official of the Office of Sutra Transcription – established around

||
52 Fukuyama Toshio 1986. The list includes tatarikata 多々理形, but not pasamɛ 波佐目. The
useful article of Fukuyama lacks a discussion of the typology of terms.
53 The manuscript Zokushū 続集 16 is held at the Shōsōin, Section Centre, 16. Its digitised
version is at the following URL: <https://tinyurl.com/398dder9> (accessed on 25 June 2023). A
diplomatic edition of the document is in DNK, vol. 1, 303–394.
 Vernacular Terms in Sinitic Texts | 187

729 under Queen Consort Kōmyō – who was already engaged at the scriptorium
when it was still part of Kōmyō’s household organisation.54
The text is a list of entries with quantities of supplies needed at the office,55
which can be grouped into four main sections: (1) wrappers with the number of
scrolls they contain; (2) ink, paper sheets and rollers for transcribing the texts;
(3) garments (e.g. the ‘pure robe’, a hemp mantle, trousers and caps) and tools
(e.g. fragrances, a cauldron and a small knife) useful to the transcription prac-
tice and rite; and (4) things to arrange and set up.
This fourth section, titled ‘Things to Arrange and Set Up’ (舗設物), includes
fifteen entries:

舗設物
長畳二枚 短畳五枚 立薦二枚 苫二帙
簀四枚 長席一枚 短机九枝(四寺送附安宿熊取見五足)
辛櫃七合(又須利一合) 瓼三口 由加六口 叩戸二口
缶一口 壺二口 長机二枝 缶一口56

A translation, useful to the current argument, is as follows:

Things to arrange and set up

Long mats: two; short mats: five; ‘vertical’ woven mats: two; straw mats: two.
Bamboo-plaited mats: four; long seats: one; short desks: nine (four sent by the temple;
Asaka no Kumatori now has five of them).
Chests: seven (and bamboo basket: one); shallow-bottomed jars: three; crocks: six; con-
tainers [for liquids]: two.
Vases: one; jars: two; long desks: two.

Among the several types of mats and pieces of furniture listed, there are two entries
on boxes, counted by the classifier 合, and five entries on kinds of containers, all
counted by the classifier 口 (lit. ‘mouth’), used for things with ‘mouths’ or ‘open-
ings’. While 辛櫃 (‘chest’), 瓼 (a ‘shallow-bottomed jar’ used for carrying water or
making liquors), 缶 (a ‘vase’ for holding liquids, large-bodied and small-mouthed),
and 壺 (‘jar’) are represented by the Sinitic words, the three words appearing in the
third line, 須利, 由加, and 叩戸, are rendered by phonograms.

||
54 For the reconstruction of the various institutions dedicated to the copying of scriptures and
related to Queen Kōmyō, as well as to aristocratic households and provincial organisations, see
Lowe 2017, 106–145.
55 There are similar extant budget documents from the same office, as in Zokuzokushū 39.4v
(DNK, vol. 21, 121). See also Lowe 2017, 106.
56 This entry is struck through, a correction necessary because the entry had already been
inserted at the beginning of this line.
188 | Antonio Manieri

The compound 須利 is recorded twice in the Wamyōruijushō, both in the

subsection on ‘Travel Utensils’ (行旅具) in Book 14 and in that on ‘Bamboo Im-
plements’ (竹器類) in Book 16.

Wamyōruijushō, Book 14, 22/189

簏 説文云(音鹿楊氏漢語抄云簏子須利)竹篋也57

Bamboo basket. The Shuowen states (the sound is that of luwk 鹿; the Yōshi kangoshō
states [that] ‘bamboo basket’ 簏子 is the suri) [that it] is a small bamboo box.

Wamyōruijushō, Book 16, 23/205

簏 考声切韻云簏(音祿和名須里)箱類也58

Bamboo basket. The Kaosheng qieyun states that the ‘bamboo basket’ 簏 (the sound is that
of luwk 祿; the Japanese name is suri) is a type of box.

In both cases, suri 須利 is attested as the equivalent of 簏子 or 簏, defined as a

‘small box made of bamboo’.
The term yuka 由加 is attested in the Wamyōruijushō’s subsection on ‘Earth-
en Implements’ (瓦器類) in Book 16.

Wamyōruijushō, Book 16, 23/204

游堈 唐韻云堈(音剛楊氏漢語抄云游堈由賀)甕也(今案俗人呼大桶為由加乎介是弁色立
成云於保美加)59

Urn. The Tangyun states [that] a ‘crock’ 堈 (the sound is that of kang 剛; the Yōshi kangoshō
states [that] ‘crock’ 游堈 [corresponds to the vernacular] yuka) is a [kind of] urn 甕. (It is possi-
ble that people call a big bucket 大桶 yuka woke. The Benshiki rissei states [it is] opomika.)

The lemma yuka is rendered by the phonogram 由 and the phonogram 賀, which has
the same value as 加, and is presented as the equivalent of an earthen crock. The
same phonograms also appear in mokkan 159 (198 × 26 × 2 cm), unearthed in Nara:

r. 移 務所 経師分由加六口
v. 附秦忌寸万呂60

Notification: Scripture master divides six crocks.

Submitted by Pata nǝ Imiki Marǝ.

||
57 Nakada Norio 1978, 164.
58 Nakada Norio 1978, 179.
59 Nakada Norio 1978, 179.
60 The mokkan is an item with a square end and a hole perforated. The digitised version is at
<https://tinyurl.com/3wtnny2y> (accessed on 25 June 2023). A diplomatic edition is found in
Terasaki Yasuhiro 1989, 9.
 Vernacular Terms in Sinitic Texts | 189

This mokkan is an i 移, i.e. a notification between equivalent offices regulated

by Article 12 of the Kūjikiryō.61 Though only a few characters are inscribed, we
can see that 経師 ‘scripture master’ occurs, and the name Hata no Imiki Maro
(Pata nǝ Imiki Marǝ) 秦忌寸万呂 is that of the same Hata no Maro 秦麻呂
attested in another mokkan, where he is declared a ‘proofreading attendant’ (
文校帳内), thus, again involved in scripture transcription.62
Finally, the third word, 叩戸, though not attested as a Sinitic word, has a
slightly different aspect than the other two terms. The character 戸 is a
kungana-type phonogram for pɛ, while the first one, 叩, is used as a trisyl-
labic kungana-type tataki. Also attested in the Engishiki, Book 1 ‘Festival to
the Four Deities of Hiraoka’ (Hiraoka no kami shiza no matsuri 平岡神四座祭),
where it is glossed as tataihe タタイヘ, it is a synonym of 叩瓫 (tataibe),
which is a type of pottery container for water or wine.63
To sum up, also in the passage from the Shasho zōyōchō, vernacular
terms rendered by phonograms, including kungana-type ones, are used in a
Sinitic text and even inserted among other terms in the same semantic family
of containers, which are written in Sinitic (semantograms). Finally, it is also
worth noting that in furnishing the Japanese equivalents of these terms, the
Wamyōruijushō quotes some previous dictionaries, such as the Yōshi kan-
goshō 楊氏漢語抄 (‘Notes on Chinese Words by Master Yako’, 720 c.) and the
Benshiki rissei 弁色立成 (‘Compendium of Classifications’, early eighth cen-
tury).

4.5 Source E: Weaving tools in the Hizen no kuni fudoki

The last example I shall provide is drawn from a received document, namely the
Hizen no kuni fudoki 肥前国風土記 (‘Record of the Province of Hizen and its
Customs’). This bureaucratic record was compiled in eighth-century Japan pur-

||
61 Inoue Mitsusada et al. 1976, 379; Dettmer 2010, 358–359; Migliore 2011, 62–63.
62 The term is also attested in the Engishiki, for example, in Book 1, Kasuga no kami shiza no
matsuri 春日神四座祭; Hiraoka no kami shiza no matsuri 平岡神四座祭; and Hirano no kami
shiza no matsuri 平野神四座祭 (Kuroita Katsumi 1938, 12, 16, 21). See also Bock 1970, 66, 71, 75.
63 Engishiki, Book 1, Hiraoka no kami shiza no matsuri (Kuroita Katsumi 1938, 16). See also
Bock 1970, 71; Arai Hideki 2019, 380.
190 | Antonio Manieri

suant to a decree promulgated by the sovereign Genmei 元明 (660–721, r. 707–715)

in 713, which stipulated that the governors of each province should prepare an
elaborate report on the geographic features of their provinces, with particular
emphasis on natural resources and legends of bygone times. The primary objec-
tive of this edict was to establish the legitimacy of the Yamato court’s authority.64
We do not have the precise date of compilation or completion of the text,
but several features suggest dating it between 732 and 739.65 The authors of the
Hizen no kuni fudoki remain unknown, as is the case for the majority of the fudo-
ki 風土記 (geographic records) corpus. However, it is possible that the compila-
tion of the text involved a layered approach. The initial phase of writing may
have been conducted by district officials who provided sources of a purely bu-
reaucratic nature on their lands. These officials probably played an important
role in consulting with the elders, as mandated by Genmei’s edict. The second
phase of writing probably consisted of revision by provincial officials. It is evi-
dent that the texts were the product of officials dispatched to the provinces on
behalf of the central government, who were strongly aligned with the central
government and supported its prerogatives and needs. It is worth noting that
the province of Hizen was also subject to the authority of a supra-provincial
institution known as the Dazaifu, located on the island of Kyūshū. Consequent-
ly, it is plausible that the Hizen no kuni fudoki was further reviewed by this of-
fice, as suggested by the numerous similarities shared with reports from other
provinces.66
The oldest extant manuscript of the Hizen no kuni fudoki is the Inokumabon
猪熊本 from 1297, preserved at the private archive of Inokuma Nobuo 猪熊信男
(1883–1963), who discovered it at a book market. Nowadays recognised as a
‘national treasure’ (kokuhō 国宝), it is acknowledged as a reliable version of the

||
64 The edict is in the Shoku Nihongi, Book 6, Wadō 6.5.2. For the text see Aoki Kazuo et al. 1989,
196–199.
65 Manieri 2022b, 37.
66 Manieri 2022b, 37–38. If we accept the dating of the text between 732 and 739, the governor-
general of the island would be Fujiwara no Umakai 藤原宇合 (694–737), who was the son of
Fuhito 不比等 (659–720), Genmei’s powerful minister of the right, an expert in Sinitic both in
terms of vocabulary and sentence structure, especially after his participation in the 717 mission
to the Tang. Umakai became governor-general after 734 (the exact date is not known) until 737,
so his contribution to the writing of the text appears very probable and, according to Akimoto
Kichirō 1958, 29, even inevitable.
 Vernacular Terms in Sinitic Texts | 191

text. The few other existing manuscripts of the text are much later and do not
show consistent variations from Inokumabon.67
The passage to be examined in this study, for which no variant has been
found among the several testimonies, pertains to a narrative section describing
the village of Himekoso, located in the District of Ki. This section is valuable for
elucidating the origin of the place name, which is a customary feature of the
fudoki. The compilers specifically record the legend of the shrine attendant
Kazeko, who was able to appease a cursing deity after receiving an oracle.
Kazeko is visited by a dream where he sees weaving tools that dance, push him
around and indicate that the cursing deity is Orihime, the ‘princess of weaving’.
As a result, the popular etymology recorded in the passage suggests that the
name Himekoso means ‘princess’s shrine’. The passage I will be focusing on
reads as follows:

其夜、夢見臥機謂(久豆比岐)絡垛謂(多々理)。儛遊出来、圧驚珂是古。於是、知織女神。68

That night, in a dream, [Kazeko] saw that a heddle cord of the loom (called kutupiki) and a
warping reel (called tatari) were dancing and pushing him around. From this, he under-
stood that [the deity] was Orihime.

The two tools Kazeko sees in the dream are a heddle cord of the loom 臥機 (MC
ngwaH kj+j), which is the tool used to wind the sewing threads, and a warping reel
絡垛 (MC lak thwaX). The compilers used the two Sinitic terms 臥機 and 絡垛, but
they needed to gloss the two terms for weaving tools with notations in man’yōgana
万葉仮名 phonograms to furnish the vernacular equivalents: kutupiki 久豆比岐
and tatari 多々理, respectively. The glosses are introduced by the character 謂 (MC
hjw+jH, Ch. wei, Jp. iware), meaning ‘referred to as, termed’. Both terms are record-
ed in the Wamyōruijushō, Section 22 ‘Utensils. 2’ (調度部), Subsections 185 ‘Parts of
the Loom’ (織機具) and 186 ‘Sericulture’ (蚕糸具):

Wamyōruijushō, Book 14, 22/185

臥機 楊氏漢語鈔云臥機(久豆比岐) […]69

Heddle cord of the loom. The Yōshi kangoshō states ([that this is called] kutupiki).

||
67 The oldest printed editions date back to 1800, when the kokugaku 国学 philologist Arakida
Hisaoyu 荒木田久老 (1746–1804) published the annotated edition with a preface by Hasegawa
Sugao 長谷川菅緒 (d. 1848) at the Yanagihara Kihee 柳原喜兵衛 publisher in Ōsaka.
68 The text is from the Inokumabon in the critical edition by Okimori Takuya, Satō Makoto and
Yajima Izumi 2008, 70.
69 Nakada Norio 1978, 161.
192 | Antonio Manieri

Wamyōruijushō, Book 14, 22/186

絡垛 楊氏漢語鈔云(多々理) […]70

Warping reel. The Yōshi kangoshō states ([that this is called] tatari).

As these two entries show, the Wamyōruijushō quotes the lost Yōshi kangoshō in
turn. In Saikaidō (Kyūshū) fudoki, glosses with vernacular equivalents are usu-
ally inserted just to provide the local variant of a word. But in this case, the
compilers feel the need to clarify these Sinitic terms by means of vernacular
equivalents that are used to facilitate the understanding of difficult technical
words related to the specialised domain of the textile sector.

5 Discussion
In the previous section, we have analysed five different texts in Sinitic from
which the following vernacular terms have been elicited (Table 1). The eight
terms belong to different lexical domains, from hippology and veterinary (texts
A and B) to carpentry and architecture (text C), from bamboo and pottery
craftsmanship (text D) to the textile sector (text E). The attestations provided are
not isolated or rare cases, since some terms are also found in other eighth-
century texts or in later documents, such as the term yuka, also occurring in a
mokkan, or the same yuka and tatakipe found in the Engishiki. Previous litera-
ture confirms trends in this direction, given the large repertoire of coat colours71
attested in documents on paper or wooden tablets, or the list of architectural
terminology in the Shōsōin monjo.72
It is useful to evaluate the nature of the terms from two different perspec-
tives. On the one hand, we need to evaluate the intrinsic character of the words
elicited in terms of their classification and exposition in ancient dictionaries. On
the other hand, we have to consider the register and usage of the words, focus-
ing on the semantic domains to which they belong, the textual genres in which
they occur, the aim of the texts and the actors involved in the communication.

||
70 Nakada Norio 1978, 163.
71 Manieri 2012.
72 Fukuyama Toshio 1986.
 Vernacular Terms in Sinitic Texts | 193

Table 1: Vernacular terms.

Text Term Attestation in the Wamyōruijushō

A kakɛ 鹿毛 (‘bay’) Kangoshō
B tari 多利 (‘lameness’) zoku (vernacular)
ute 宇弖 (‘bruise’?) not attested
C tatarikata 多々理形 (‘strut’) Yōshi kangoshō
pasamɛ 波佐目 (‘insert’?) not attested
D suri 須利 (‘bamboo basket’) Yōshi kangoshō
yuka 由賀 (‘crock’) Yōshi kangoshō
tatakipe 叩戸 (‘container’) not attested
E kutupiki 久豆比岐 (‘heddle cord’) Yōshi kangoshō
tatari 多々理 (‘warping reel’) Yōshi kangoshō

5.1 Dictionary description of the terms

Verifying the attestation of terms in phonograms within the Wamyōruijushō has
been necessary not only to search for the meanings of these words but also to
ascertain the treatment they received in Shitagō’s dictionary. The Japanese
equivalents are always provided, either through a citation from a previous dic-
tionary or glossary or through a label indicating a lexicographic marker.
The labels zoku 俗 and shisetsu 師説 occur in the entries analysed in this
paper.
The character zoku can indicate different meanings in different contexts, as
recently demonstrated by Baba Mariko.73 Generally, zoku can be interpreted as
referring to what is typical of everyday life: what is common, usual or not par-
ticularly noteworthy. Incidentally, zoku also appears in the compound sezoku 世
俗, meaning ‘the world of common people’, in the preface of the Wamyōruiju-
shō, where it is contrasted with fūgetsu 風月, a well-known and widely used
metaphor for belles lettres, intended as both poetry and prose in Sinitic. Moreo-
ver, zoku is often associated with spoken language and idiomatic expressions
such as ‘the talks of the streets and discussions of the alleys’ (街談巷説). This
expression, which is still used as a four-character idiom in both China and Ja-
pan, refers to what is not canonical, has practical utility and is performed orally.
In summary, the concept of zoku (and sezoku), as outlined in the preface, per-

||
73 Baba 2022.
194 | Antonio Manieri

tains to orality, everyday life and what is not canonical or officially transmitted,
such as popular conversations and discussions.
The label shisetsu refers to explanations offered by the master while com-
menting on a certain text in public or private lectures. In the Wamyōruijushō, it
is applied to several Chinese texts, such as the Wenxuan 文選 (‘Literary Selec-
tion’, c. 530), the Youxianku 遊仙窟 (‘The Dwelling of Playful Immortals’, early
eighth century) and the Yanshi jiaxun 顔氏家訓 (‘Family Instructions of Master
Yan’, second half of the sixth century), as well as some Japanese works, such as
the Nihon shoki 日本書紀 (‘Chronicle of Japan’, 720). The lectures were common-
ly conducted in private learning contexts outside the official state curriculum,
thus, merging philological activity with oral transmission.74 Therefore, as has
been noted previously, these terms have a strong oral character and are not
typically associated with written language. Additionally, none of these terms
are found in contemporaneous poetry, not even the ancient songs of the Kojiki
or Nihon shoki or the Man’yōshū. Incidentally, of the eight terms, only tari, suri,
yuka, kutupiki and tatari are recorded in the dictionary of Old Japanese lan-
guage Jidaibetsu kokugo daijiten. Jōdaihen 時代別国語大辞典―上代編, while
only tatari is attested in the ONCOJ – Oxford-NINJAL Corpus of Old Japanese.75

5.2 Context of use of the terms

The texts exist on wooden tablets or paper, and represent different genres, in-
cluding brief documents required by common people, such as transit passes;
long documents consisting mostly of lists of words and quantities, such as regis-
ters of expenses and necessities; or geographical reports, where compilers
choose to explain technical terms even when used in a narrative passage related
to a myth.
Consequently, some texts have a predominantly bureaucratic purpose, such
as recording goods, registering budgets and reporting the situation of provinc-
es, while others have practical aims, such as allowing the crossing of frontiers
and organising an office. However, regardless of their purpose, the texts are not
private communications or individual notes, but are ‘public’, if not even ‘offi-

||
74 The shisetsu label applied to the Wenxuan has been studied by Yin Xianhua 2009, to the
Youxianku by Kuranaka Susumu 1967, to the Yanshi jiaxun by Kuranaka Shinobu 2011, and to
the Nihon shoki by Kuranaka Shinobu 1988.
75 Manieri 2022a has shown how very few vernacular terms that the Wamyōruijushō quotes
from the Nara-period Yōshi kangoshō are effectively recorded in the Jidaibetsu kokugo daijiten.
Jōdaihen or the ONCOJ – Oxford-NINJAL Corpus of Old Japanese.
 Vernacular Terms in Sinitic Texts | 195

cial’, and presume one or more readers, not necessarily in a close relationship.
As such, the authors are aware that their texts would be read and must be un-
derstood. Thus, the authors did not use vernacular terms on their own whims or
as a personal stylistic choice.
In terms of the relevant offices, there are various levels and responsibilities
involved. Text A, for instance, was issued by the village chief and addressed to
officials stationed at the internal frontiers for inspection of the passage of travel-
lers and goods. The district-level officials, probably aided by the village chiefs,
were responsible for the initial drafting of the information in the fudoki (text E),
which was subsequently refined by the provincial officials, including the pro-
vincial governor. Regarding the Hizen no kuni fudoki, the text was further scru-
tinized and endorsed by the governor-general of Dazaifu before being submitted
to the central government. The provincial officials were also the authors of the
shōzeichō (text B), which had to undergo review by the Bureau of Public Re-
sources before being presented to the Great Council of State. Texts C and D were
compiled by officials from the Institute for the Construction of the Ishiyama
Monastery and the Office of Sutra Transcription, respectively, both in connec-
tion with the Office for the Construction of the Tōdai Monastery.
In all of these offices, as in any office, bureau or ministry of the state,
fourth-level officials played an important role in assisting higher officials in the
compiling of documents. This type of petty official, holding a rank of 7 to 9 on a
ranking scale of 1 to 10 (where 10 is the lowest), was indicated by different char-
acters according to the office, but all refer to the so-called reishi 令史 (clerks),
who were in charge of drafting texts, such as correspondence and records, and
reading, selecting and summarising official documents.
However, clerks and their superiors, as well as the village and the district
chiefs, were not expected to be experts in the fields of knowledge treated in the
documents that they compiled.
In other words, the official who had to compile the document did not neces-
sarily have a command of the knowledge, sources and related terminology to be
used in the document. At the same time, there were offices of shinabe 品部 (or
be 部), professional groups of skilled artisans, who were the actual experts and
possessed the skilled know-how required by the developed bureaucratic system
of the so-called ‘code-based state’ (ritsuryō kokka 律令国家).76
Therefore, apparently specialised terminologies were employed at two dis-
tinct levels of the state system: firstly, in the production and management of

||
76 Shinabe were a relic of pre-ritsuryō Japan. See Kanō Hisashi 1960. In general, on the ritsuryō
state, see Enomoto Jun’ichi 2010.
196 | Antonio Manieri

activities for the benefit of the state, where labourers undertook hands-on work;
and secondly, in the compilation of various document types, where users rang-
ing from clerks to high-ranking officials engaged in managerial or intellectual
work.77
The two categories of actors using terminologies had differing typologies of
training in and command of technical knowledge and its lexicon. Skilled la-
bourers held a procedural form of technical knowledge, and would learn termi-
nologies orally and by performative acts, consisting, for example, of naming
things while indicating them, or describing procedures while showing how to
do them. Conversely, document compilers were not necessarily immersed in
such a learning environment. They were expected to possess a descriptive, al-
most exclusively conceptual and lexical form of specialised knowledge, not
imparted by the official system of education established by the Gakuryō 学令
(‘Law on Education’) or represented by the Bureau of Higher Education
(Daigakuryō 大学寮, i.e. ‘State Academy’).78 They needed to know only the terms
to fill out the documents, and it seemed essential for them to acquire such
words, or, at least, know the tools to search for the words.
In my previous research, it has been demonstrated that the Wamyōruijushō
cites dictionaries such as the Yōshi kangoshō for words that frequently appear in
the types of texts under investigation. Unfortunately, the Yōshi kangoshō is now
lost and survives only in indirect transmission. It is one of the fragmentary dic-
tionaries collectively known as kangoshō (‘notes on Chinese words’), which also
includes the Kangoshō found for the lemma kakɛ, and the Benshiki rissei. The
textual reconstruction and analysis of their lexicographic macro- and micro-
structure show that they are bilingual, monodirectional dictionaries with Sinitic
as the source language and vernacular Japanese as the target language. They
collect terms covering technical domains of knowledge: generally, low-frequency
words or hapax, not attested in refined literature, but mostly in handwritten

||
77 The presence of clerks and skilled artisans in most offices and bureaus can be ascertained
by looking at the Shiki’inryō 職員令 (‘Law on Officialdom’). See Inoue Mitsusada et al. 1976,
157–196. An introduction to the system in English is in Sansom 1932.
78 An exception is the field of medicine, since the Ishichiryō 医病令 (‘Law on Medicine’) also
regulated the official education of medicine and acupuncture students, and had a special
institution, the Bureau of Medicine (Ten’yakuryō 典薬寮), dedicated to their training. For an
introduction to the official educational system in ancient Japan, see Momo Hiroyuki 1994; the
translation of the Gakuryō into English is in Crump 1952; for medical education, see Maruyama
Yumiko 1998.
 Vernacular Terms in Sinitic Texts | 197

practical documents on wooden tablets or paper. The contents of these diction-

aries are arranged thematically to provide easy access to their contents.79
Therefore, they were intended to serve as ‘passive’ dictionaries, as they
were consulted for passive assistance in comprehending unfamiliar Sinitic
terms encountered in reading. Additionally, due to their thematic organisation,
they were surely used as pedagogical instruments for learning technical termi-
nologies across various semantic domains.80 Specifically, they proved valuable
for the technical instruction of state clerks, whose cultural backgrounds may
not have encompassed the competencies and specialised vocabulary necessary
for their assigned office. Their compilers, thus, mediated the passage of tech-
nical-artisanal knowledge from the oral and performative layer of specialised
labourers to the conceptual and literary layer of the clerks.
The learning of terminologies and vocabulary in ancient Japan is an area of
research that has been largely overlooked and deserves greater attention. While
this paper touches on this issue only briefly, it is also essential to consider some
wooden tablets that were used in educational settings, called shūsho mokkan 習
書木簡 (‘wooden tablets for learning texts’). Previous research has focused on
tablets where Sinitic words are accompanied by their vernacular equivalents,
which are known as ongi mokkan 音義木簡 (literally, ‘wooden tablets with pro-
nunciation’);81 where characters are repeatedly written down along with similar
characters or those with the same radical component;82 or where texts from
classics are copied.83 However, a third group of wooden tablets for learning also
exists; this consists of items containing words that share a semantic relation-
ship, such as meronymy, co-hyponymy and antonymy, in a way that resembles
the arrangement found in the kangoshō dictionaries. These various types of
shūsho mokkan show how part of education focused on learning terminologies
that did not occur in the classics or in belles lettres, but were useful for practical
and bureaucratic work.

||
79 Kuranaka Susumu 2001; Kuranaka Susumu 2002; Kuranaka Susumu 2003; Lin Zhongpeng 2012;
Manieri 2022a.
80 Manieri 2022a.
81 Yamamoto Takashi 2020.
82 Watanabe Akihiro 2009; Inoue Miyuki 2017.
83 Satō Makoto 1997, 429–465; Watanabe Akihiro 2009.
198 | Antonio Manieri

6 Conclusion
This paper has presented an analysis of five multilingual texts from eighth-
century Japan employing Sinitic syntax and lexicon but also containing some
vernacular Japanese terms. These terms cover a range of semantic domains,
including hippology, carpentry, craftsmanship and the textile sector. As evi-
denced by the dictionaries consulted, these terms have Sinitic counterparts,
which, however, were not made use of in the texts themselves. Therefore, their
use in the vernacular is not based on a Japanese cultural specificity of the con-
cept or object that would require the use of phonograms, such as culturemes or
proper names. Instead, the referents of the terms are known in both Japan and
China, and, thus, the use of the Sinitic equivalent in the texts does not appear
meaningless. Moreover, other specific words in Sinitic are also used in the same
texts. The lexicographic exposition of the words by means of labels, such as
zoku or shisetsu, has shown a solid relation to orality, and the usual contexts of
usage of these words confirm this. Incidentally, viewed from a transcultural
perspective, this phenomenon does not appear exceptional: as in some Europe-
an traditions, the origins of the artisanal lexicon are rooted in dialects, which
are vernacular and popular means of expression of craftsmanship, in contrast to
the learned and educated varieties used by cultural elites.84
The artisanal lexicon and specialised terminology in ancient Japan were,
therefore, based on a dense social and economic network, whose actors were
the workers, who possessed the know-how and, in fact, used the vernacular
terminology; the officials, who mostly had a role of management and control
through the drafting of documents modelled on Sinitic formats; and the schol-
ars (‘lexicographers’), who connected theoretical and bookish erudition from
Chinese sources with practical and procedural instruction from local profes-
sional groups. The latter were particularly able to promote broader access to
technical knowledge through the compilation of dictionaries and glossaries.
The attestations of Sinitic – vernacular word pairs in the dictionaries – suggest
that officials effectively learned the terminologies in both versions.
Therefore, the choice of vernacular terms in documents with formalized Si-
nitic models does not invalidate the function of the texts or affect communica-
tion, since the terms are expressions of a residual orality justified by the envi-
ronment in which they were used more, and because related tools, such as the
kangoshō-type dictionaries, were produced in the same period.

||
84 See e.g. Zanola 2018.
 Vernacular Terms in Sinitic Texts | 199

In conclusion, the code-mixing of bureaucratic Sinitic and artisanal vernac-

ular found in the texts results from the interaction between two different sys-
tems of knowledge: that of Chinese scholarship and literature (in the broader
sense of ‘writing texts’), which were described by books and approached in
formalized settings, and that of local practices, which were transmitted, learned
and disseminated orally and performatively via work experience. These two
systems are not intended in a strictly binary opposition, as both were interrelat-
ed in the management of the relevant offices to which officials and labourers
belonged, and both contributed to the good governance of the state.

Acknowledgements

The author wishes to thank Maria Chiara Migliore, Inoue Miyuki and the anon-
ymous reviewer for their insightful comments and suggestions, and Kristen de
Joseph for her help in revising the English text.

Abbreviation
DNK = Tōkyō daigaku shiryō hensanjo 東京大学史料編纂所 (eds) (1901–1940), Dai Nihon
komonjo 大日本古文書, Tokyo: Tōkyō daigaku shuppankai.

Dictionaries, corpora, databases

Jōdaigo jiten henshū iinkai 上代語辞典編集委員会 (eds) (2001), Jidaibetsu kokugo daijiten.
Jōdaihen 時代別国語大辞典―上代編, Tokyo: Sanseidō [1st edn: 1967].
Mokkanko – Wooden Tablet Database (Nara National Research Institute for Cultural Properties)
<https://mokkanko.nabunken.go.jp/ja/> (accessed on 25 June 2023).
ONCOJ – Oxford-NINJAL Corpus of Old Japanese <https://oncoj.ninjal.ac.jp> (accessed on 25
June 2023).
Shōsōin Database (Imperial Household Agency) <https://shosoin.kunaicho.go.jp/> (accessed
on 25 June 2023).

References
Aizawa Masahiko 相澤正彦 and Kuniga Yumiko 國賀由美子 (eds) (2016), Ishiyamadera engi
emaki shūsei 石山寺縁起絵巻集成, Tokyo: Chūō kōron bijutsu shuppan.
Akimoto Kichirō 秋元吉郎 (ed.) (1958), Fudoki 風土記, Tokyo: Iwanami shoten.
200 | Antonio Manieri

Aoki Kazuo 青木和夫, Inaoka Kōji 稲岡耕二, Sasayama Haruo 笹山晴生 and Shirafuji Noriyuki
白藤禮幸 (eds) (1989), Shoku Nihongi 続日本紀, vol. 1, Tokyo: Iwanami shoten.
Aoki Kazuo 青木和夫, Inaoka Kōji 稲岡耕二, Sasayama Haruo 笹山晴生 and Shirafuji Noriyuki
白藤禮幸 (eds) (1990), Shoku Nihongi 続日本紀, vol. 2, Tokyo: Iwanami shoten.
Arai Hideki 荒井秀規 (2019), ‘Hotogi to motai ni kansuru oboegaki’ 瓫(盆)(ホトギ)と瓮・缶(
モタイ)に関する覚え書き, Kokuritsu rekishi minzoku hakubutsukan kenkyū hōkoku 国立
歴史民俗博物館研究報告, 218: 375–398.
Baba, Mariko (2022), ‘The Transformation of the Meaning and Concept of Zoku 俗. The “Sa-
cred” and “Profane” in Japanese Translation’, Journal of Religion in Japan, 11: 165–195.
Baxter, William H. and Laurent Sagart (2014), Old Chinese: A New Reconstruction, New York:
Oxford University Press.
Bender, Ross (2009), ‘Performative Loci of the Imperial Edicts in Nara Japan, 749–70’, Oral
Tradition, 24/1: 249–268.
Bock, Felicia G. (1970), Engi-shiki: Procedures of the Engi Era, Books I–V, Tokyo: Sophia University.
Bock, Felicia G. (1972), Engi-shiki: Procedures of Engi Era, Book VI–X, Tokyo: Sophia University.
Bock, Felicia G. (1985), Classical Learning and Taoist Practices in Early Japan, Tempe: Arizona
State University.
Buhrman, Kristina (2017), ‘Knowledge of Nature and Craft: Researching the History of Science,
Mathematics, and Technology in Japan before 1600’, in Karl F. Friday (ed.), Routledge
Handbook of Premodern Japanese History, London: Routledge, 365–370.
Cabré, Maria Teresa (1999), Terminology: Theory, Methods, and Applications, Amsterdam: John
Benjamins.
Crump, James I. Jr. (1952), ‘“Borrowed” T’ang Titles and Offices in the Yōrō Code’, Occasional
Papers of the Center for Japanese Studies, University of Michigan, 2: 35–58.
Denecke, Wiebke (2014), ‘Worlds without Translation: Premodern East Asia and the Power of
Character Scripts’, in Sandra Bermann and Catherine Porter (eds), A Companion to Trans-
lation Studies, New York: John Wiley & Sons, 204–216.
Dettmer, Hans A. (2010), Der Yōrō-Kodex: Die Gebote, Einleitung und Übersetzung des Ryō no
gige, Bücher 2–10, Wiesbaden: Harrassowitz.
Duthie, Torquil (2014), Man’yōshū and the Imperial Imagination in Early Japan, Leiden: Brill.
Enomoto Jun’ichi (2010), ‘Japan Ritsuryō System in the “East Asian World”’, Acta Asiatica, 99:
1–17.
Farris, William Wayne (1998), Sacred Texts and Buried Treasures: Issues in the Historical Ar-
chaeology of Ancient Japan, Honolulu: University of Hawai‘i Press.
Fukuyama Toshio 福山敏男 (1986), ‘Shōsōin monjo ni mieru kenchiku yōgo’ 正倉院文書に見え
る建築用語, Shōsōin nenpō 正倉院年報, 8: 1–14.
Handel, Zev (2019), Sinography: The Borrowing and Adaptation of the Chinese Script, Leiden:
Brill.
Hayakawa Shōhachi 早川庄八 (1997), Nihon kodai no monjo to tenseki 日本古代の文書と典籍,
Tokyo: Yoshikawa kōbunkan.
Hayashi Ryoichi (1975), The Silk Road and the Shoso-in, New York: Weatherhill/Tokyo:
Heibonsha.
Hérail, Francine (1966), Yodo no tsukai ou le système des Quatre Envoyés = Bulletin de la Mai-
son franco-japonaise, 8/2, Paris: Presses universitaires de France.
Inoue Mitsusada 井上光貞, Seki Akira 関晃, Tsuchida Naoshige 土田直鎮 and Aoki Kazuo 青木
和夫 (eds) (1976), Ritsuryō 律令, Tokyo: Iwanami shoten.
 Vernacular Terms in Sinitic Texts | 201

Inoue Miyuki 井上幸 (2017), ‘Kodai no shūsho mokkan ni okeru renshū haikei tenbyō. Tōji no
shiyō jitai kara no suisoku’ 古代の習書木簡における練習背景点描―当時の使用字体から
の推測, Asia: Shakai, Keizai, bunka Asia: 社会・経済・文化, 4: 192–180 [sic].
Inoue Tatsuo 井上辰雄 (1967), Shōzeichō no kenkyū. Ritsuryō jidai no chihō seiji 正税帳の研究
― 律令時代の地方政治, Tokyo: Hanawa shobō.
Kanō Hisashi 狩野久 (1960), ‘Shinabe zakko sei no saikentō’ 品部雑戸制の再検討, Shirin 史林,
43/6: 871–903.
Kin Bunkyō (2021), Literary Sinitic and East Asia: A Cultural Sphere of Vernacular Reading, ed.
Ross King, tr. Ross King et al., Leiden: Brill.
King, Ross (2014), ‘Introduction. Koh Jongsok’s Infected Language’, in Koh Jongsok, Infected
Korean Language, Purity Versus Hybridity: From the Sinographic Cosmopolis to Japanese
Colonialism to Global English, Amherst, NY: Cambria Press, 1–16.
Kiyota Yoshiki 清田善樹 (1980), ‘Nara-Heijōkyū seki (dai16-17ji)’ 奈良・平城宮跡(第 16 17 次),
Mokkan kenkyū 木簡研究, 2: 64–65.
Kornicki, Peter F. (2018), Languages, Scripts, and Chinese Texts in East Asia, Oxford: Oxford
University Press.
Kuranaka Shinobu 藏中しのぶ (1988), ‘Wamyōruijushō shoin Nihongi, Nihongishichū no
saikentō. Jikkanbonkei/nijikkanbonkei no idō o chūshin ni’ 『和名類聚抄』所引「日本紀」
「日本紀私記」の再検討－十巻本系・廿巻本系の異同を中心に, Minato. Kotoba to
rekishi 水門ー言葉と歴史, 16: 22–66.
Kuranaka Shinobu 藏中しのぶ (2011), ‘Ritsuryō, bukkyō, bungaku no kōsaku. Tōdai kōgo goi
“ganmen” o meguru kōsetsu no ba’ 律令・仏教・文学の交錯―唐代口語語彙「顔面」を
めぐる講説の場, Nihon bungaku 日本文学, 60/5: 31–39.
Kuranaka Susumu 藏中進 (1967), ‘Wamyōruijushō to Youxianku’ 『和名類聚抄』と『遊仙窟』,
Kōbe gaidai ronsō 神戸外大論叢, 18/2: 29–45.
Kuranaka Susumu 藏中進 (2001), ‘Wamyōruijushō shoin Benshiki rissei kō’ 『和名類聚抄』所
引『弁色立成』考, Tōyō kenkyū 東洋研究, 141: 1–40.
Kuranaka Susumu 藏中進 (2002), ‘Wamyōruijushō shoin Yōshi kangoshō kō’ 『和名類聚抄』
所引『楊氏漢語抄』考, Tōyō kenkyū 東洋研究, 145: 1–37.
Kuranaka Susumu 藏中進 (2003), ‘Wamyōruijushō shoin Kangoshō kō’ 『和名類聚抄』所引
『漢語抄』考, Tōyō kenkyū 東洋研究, 150: 1–37.
Kuroita Katsumi 黒板勝美 (1938), Kōtaishiki. Kōninshiki. Engishiki 交替式・弘仁式・延喜式,
Tokyo: Yoshikawa kōbunkan.
Kuroita Katsumi 黒板勝美 (1955), Ryō no shūge 令集解, Tokyo: Yoshikawa kōbunkan.
Kyōto daigaku bungakubu kogaku bungaku kenkyūshitsu 京都大学文学部語学文学研究室
(eds) (1967), Shinsen jikyō Tenjibon wahon, Gunsho ruijubon 新撰字鏡天治本附享和本・
群書類從本, Kyoto: Rinsen shoten.
Kyōto daigaku bungakubu kokugo kokubungaku kenkyūshitsu 京都大学文学部国語学国文学
研究室 (eds) (1968), Shohon shūsei Wamyōruijushō 諸本集成倭名類聚抄, Kyoto: Rinsen
shoten.
Lin Zhongpeng 林忠鵬 (2002), Wamyōruijushō no bunkengakuteki kenkyū 和名類聚抄の文献学
的研究, Tokyo: Bensei shuppan.
Lin Zhongpeng 林忠鵬 (2012), ‘Nara jidai no jisho Kangoshō oyobi sono shūroku goi ni tsuite’
奈良時代の辞書『漢語抄』及びその収録語彙について, in Kōno Kimiko 河野貴美子 and
Wang Yong 王勇 (eds), Higashi Ajia no kanseki isan. Nara o chūshin to shite 東アジアの漢
籍遺産―奈良を中心として, Tokyo: Bensei shuppan, 305–320.
202 | Antonio Manieri

Lowe, Bryan D. (2017), Ritualized Writing: Buddhist Practice and Scriptural Cultures in Ancient
Japan, Honolulu: University of Hawai‘i Press.
Lurie, B. David (2011), Realms of Literacy: Early Japan and the History of Writing, Cambridge,
MA: Harvard University Asia Center / Harvard University Press.
Mair, Victor H. (1994), ‘Buddhism and the Rise of the Written Vernacular in East Asia: The Mak-
ing of National Languages’, The Journal of Asian Studies, 53/3: 707–751.
Mair, Victor H. (2001), ‘Language and Script’, in Victor H. Mair (ed.), The Columbia History of
Chinese Literature, New York: Columbia University Press, 19–57.
Manieri, Antonio (2012), ‘Wamyōruijushō “gyūba no ke” mon to Narachō no kakyū kanjinsō.
Kangoshō, Yōshi kangoshō, Benshiki rissei o megutte’ 『和名類聚抄』牛馬毛門と奈良朝
の下級官人層―『漢語抄』『楊氏漢語抄』『弁色立成』をめぐって, Higashi Ajia hikaku
bunka kenkyū 東アジア比較文化研究, 11: 72–85.
Manieri, Antonio (2020), ‘Document Production by Low-rank Officials in Ancient Japan: Notes
on Some Public Advisory Texts on Wooden Tablets’, in Antonella Brita, Giovanni Ciotti,
Florinda De Simini and Amneris Roselli (eds), Copying Manuscripts: Textual and Material
Craftsmanship, Naples: UniorPress, 241–270.
Manieri, Antonio (2022a), ‘Technical Education in Nara Japan: Text and Context of the Yōshi
kangoshō (ca. 720)’, Annali – Serie Orientale, 82: 172–210.
Manieri, Antonio (2022b), Cronache del Saikaidō, Rome: Carocci.
Maruyama Yumiko 丸山裕美子 (1998), Nihon kodai no iryō seido 日本古代の医療制度, Tokyo:
Meicho kankōkai.
Migliore, Maria Chiara (2011), I documenti ufficiali del periodo Nara (710–784), Rome: Nuova
Cultura.
Migliore, Maria Chiara (2018), ‘Notes on Diplomatics and Palaeography in Ancient and Medie-
val Japan’, in Arianna D’Ottone Rambach (ed.), Palaeography between East and West =
Supplemento n. 1 alla Rivista degli Studi Orientali, Nuova Serie, 90: 179–191.
Momo Hiroyuki 桃裕行 (1994), Jōdai gakusei no kenkyū 上代学制の研究, Kyoto: Shibunkaku
shuppan.
Nakada Norio 中田祝夫 (1969), Tōdaiji fuju monkō no kokugogakuteki kenkyū 東大寺諷誦文稿
の国語学的研究, Tokyo: Kazama shobō.
Nakada Norio 中田祝夫 (ed.) (1978), Wamōruijushō Genna sannen kokatsujiban nijikkanbon 倭
名類聚抄元和三年古活字版二十巻本, Tokyo: Bensei shuppan.
Nara bunkazai kenkyūjo 奈良文化財研究所 (eds) (2017), Heijōkyū seki shiryōkan shūki tokube-
tsuten: Chika no Shōsōinten 2017. Kokuhō Heijōkyū seki shutsudo mokkan dainiki tenshi
mokkan 平城宮跡資料館秋期特別展：地下の正倉院展 2017：国宝平城宮跡出土木簡第Ⅱ
期展示木簡, Nara: Nara bunkazai kenkyūjo.
Nara kokuritsu bunkazai kenkyūjo 奈良国立文化財研究所 (1974a), Heijōkyū mokkan 2 平城宮
木簡二, Nara: Nara kokuritsu bunkazai kenkyūjo.
Nara kokuritsu bunkazai kenkyūjo 奈良国立文化財研究所 (1974b), Heijōkyū mokkan 2. Kaise-
tsu 平城宮木簡二 解説, Nara: Nara kokuritsu bunkazai kenkyūjo.
Okimori Takuya 沖森卓也, Satō Makoto 佐藤信 and Yajima Izumi 矢嶋泉 (eds) (2008), Bungo
no kuni fudoki. Hizen no kuni fudoki 豊後国風土記 肥前国風土記, Tokyo: Yamakawa
shuppansha.
Palmer, Edwina (2016), Harima fudoki: A Record of Ancient Japan Reinterpreted, Translated,
Annoted, and with a Commentary, Leiden: Brill.
Philippi, Donald L. (1990), Norito: A Translation of the Ancient Japanese Ritual Prayers, Prince-
ton, NJ: Princeton University Press.
 Vernacular Terms in Sinitic Texts | 203

Piggott, Joan (1990), ‘Mokkan. Wooden Documents from Nara Period’, Monumenta Nipponica,
45: 449–470.
Pollock, Sheldon (2006), The Language of the Gods in the World of Men: Sanskrit, Culture, and
Power in Premodern India, Berkeley: University of California Press.
Robert, Jean-Noël (2006), ‘Hieroglossia. A Proposal’, Bulletin of the Nanzan Institute for Reli-
gion & Culture, 30: 25–48.
Saitô Mareshi (2021), Qu’est-ce que le monde sinographique? Quatre conférences du profes-
seur Saitô Mareshi au Collège de France, tr. Arthur Defrance, Paris: Collège de France, Ins-
titut des hautes études japonaises.
Sansom, George B. (1932), ‘Early Japanese Law and Administration. Part I’, Transactions of the
Asiatic Society of Japan, Series 2, 9: 67–109.
Satō Makoto 佐藤信 (1997), Nihon kodai no kyūto to mokkan 日本古代の宮都と木簡, Tokyo:
Yoshikawa kōbunkan.
Sema Masayuki 瀬間正之 (2011), Fudoki no moji sekai 風土記の文字世界, Tokyo: Kasama
shoin.
Tenri daigaku fuzoku Tenri toshokan 天理大学附属天理図書館 (eds) (2018), Ruijumyōgishō.
Kanchiinbon 類聚名義抄 観智院本, 3 vols, Tokyo: Yagi shoten.
Terada Akira (2009), ‘L’évolution des idées sur la langue dans le Japon ancien’, Histoire,
Épistémologie, Langage, 31/2: 163–173.
Terasaki Yasuhiro 寺崎保広 (1989), ‘Nara-Heijōkyō seki’ 奈良 平城京跡, Mokkan kenkyū 木簡
研究, 11: 7–21.
Tōno Haruyuki 東野治之 (1977), Shōsōin monjo to mokkan no kenkyū 正倉院文書と木簡の研究,
Tokyo: Hanawa shobō.
Tōno Haruyuki 東野治之 (1983), Nihon kodai mokkan no kenkyū 日本古代木簡の研究, Tokyo:
Hanawa shobō.
Tōno Haruyuki 東野治之 (2005), Nihon kodai shiryōgaku 日本古代史料学, Tokyo: Iwanami
shoten.
Tsukishima Hiroshi 築島裕 (1963), ‘Wamyōruijushō no wakun ni tsuite’ 和名類聚抄の和訓につ
いて, Kuntengo to kunten shiryō 訓点語と訓点資料, 25: 28–60.
Von Verschuer, Charlotte (1985), Les relations officielles du Japon avec la Chine aux VIIIe et IXe
siècles, Geneva: Droz.
Vovin, Alexander (2020), A Descriptive and Comparative Grammar of Western Old Japanese,
Revised, Updated, and Enlarged Second Edition, Leiden: Brill.
Vovin, Alexander and Sambi Ishisaki-Vovin (2021), The Eastern Old Japanese Corpus and Dic-
tionary, Leiden: Brill.
Watanabe Akihiro 渡辺晃宏 (2009), ‘Nihon kodai no shūsho mokkan to kakyū kanjin no kanji
kyōiku’ 日本古代の習書木簡と下級官人の漢字教育, in Takata Tokio 高田時雄 (ed.), Kanji
bunka sansennen 漢字文化三千年, Kyoto: Rinsen shoten, 91–110.
Whitman, John (2022), ‘Contested Vernacular Readings, c. 800–830 CE: The Satō-bon Kegon
mongi yōketsu and the Tōdaji fujumonkō’, in Michael Clarke and Máire Ní Mhaonaigh
(eds), Medieval Multilingual Manuscripts Case Studies from Ireland to Japan (Studies in
Manuscript Cultures, 24), Berlin: De Gruyter, 13–32.
Yamamoto Takashi 山本崇 (2020), ‘Iwayuru ongi mokkan to sono ikyo genten. Ritsuryō kokka
seiritsuki no keiten shōrai o megutte’ いわゆる音義木簡とその依拠原典―律令国家成立
期の経典将来をめぐって, Kokugakuin zasshi 國學院雑誌, 121/11: 362–378.
204 | Antonio Manieri

Yin Xianhua 尹仙花 (2009), ‘Wamyōruijushō shoin Monzenchū ni okeru Minamoto no Shitagō
in’yō taido’ 『和名類聚抄』所引『文選』注における源順の引用態度, Gaikokugogaku
kenkyū 外国語学研究, 10: 1–15.
Zanola, Maria Teresa (2018), ‘La terminologie des arts et métiers entre production et commer-
cialisation: une approche diachronique’, Terminàlia, 17: 16–23.
Emmanuel Francis
Mixing Languages and Scripts in Tamil
Inscriptions and Manuscripts
Abstract: This chapter focuses on two periods in the history of writing texts in
the Tamil language in the present-day Tamil Nadu state of South India. The first
period starts around 600 CE, when two different alphabets – the Grantha and
the Tamil alphabets – were designed to write texts in Sanskrit and Tamil, re-
spectively. One can observe for several centuries onwards that Sanskrit loan-
words are often written in their specific Grantha alphabet in Tamil inscriptions.
The second period is attested in Tamil manuscripts, most of which are dated to
the eighteenth and nineteenth century CE. These manuscripts evince new prac-
tices of script-mixing, the most conspicuous being the creation of conjunct
graphemes mixing Grantha and Tamil alphabets.

1 Introduction
The Tamil region in South India, the present-day state of Tamil Nadu, has
known a peculiar situation as far as alphabets have been concerned for approx-
imately 1200 years, from the seventh to the nineteenth century CE. A single al-
phabet has been used to write texts in the regional language and in Sanskrit, the
Indian cosmopolitan language, in most parts of India. By contrast, a distinct
alphabet, called Grantha, designed around 600 CE, was used to write texts in
Sanskrit in the Tamil south, whereas other distinct alphabets (called the Tamil
and Vaṭṭeḻuttu alphabets) were used for writing texts in the Tamil language.
Given the multilingual (Sanskrit and Tamil) culture of the learned, it often
happened that multilingual texts were produced in the Tamil area, with the use
of specific alphabets for the different languages. This has resulted in interesting
cases of language and alphabet mixing attested early in Tamil inscriptions and
later in the surviving manuscripts.
The present chapter will focus mainly on two different cases of language and
alphabet mixing in the Tamil area, separated by approximately a millennium:
− The use of Grantha graphemes for Sanskrit loanwords in otherwise fully
Tamil inscriptions, with examples dated between the seventh and the thir-
teenth century CE.

Open Access. © 2024 the author, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-008
206 | Emmanuel Francis

− The use of the Grantha grapheme m as the final element of conjunct graph-
emes mixing Grantha and Tamil alphabets, in otherwise fully Tamil manu-
scripts dated mostly to the eighteenth and nineteenth century CE.

2 Scripts and languages in the Tamil region

The earliest attested alphabet for writing Tamil is the Tamil-Brāhmī alphabet
(also called the Tamiḻi alphabet). This is an adaptation of the northern Brāhmī
alphabet (fourth century BCE, at the latest), initially designed for writing Indo-
Aryan languages. This alphabet has been adapted in the Tamil area to write
inscriptions in Tamil, a Dravidian language, from the third century BCE. Slightly
later, another alphabet called Vaṭṭeḻuttu, also derived from the Brāhmī alpha-
bet, is attested to write down Tamil.1
A new alphabet for Tamil appeared around the beginning of the seventh
century CE, simply called the Tamil alphabet, from which the modern Tamil
alphabet derives. At the same time, a specific alphabet, called Grantha, was
designed for writing Sanskrit texts, as this language knows phonemes (e.g.
aspirate) unknown in Tamil and as, in writing, Sanskrit consonant clusters are
generally dealt with ligature letters rather than using a ‘vowel-killer’ (see below)
on the first consonant.2 The Vaṭṭeḻuttu alphabet continued to be used, up to the
eighth century CE, in countryside areas, for instance, for the specific type of hero
stone inscriptions, before later being confined to Kerala and the southern dis-
tricts of present-day Tamil Nadu.
The Tamil and Grantha alphabets, designed around 600 CE, however, have
some graphemes in common, as shown in Table 1, which is only an attempt to
describe the configurations of Tamil and Grantha alphabets at the time of their
invention in the seventh century CE.3

||
1 On these small corpora of Tamil-Brāhmī and early Vaṭṭeḻuttu inscriptions, and on the rela-
tions between these alphabets, see Mahadevan 2003 and 2014.
2 Grantha is also called Grantha Tamil (Grünendahl 2001), Tamilian Grantha (Franceschini 2022),
to distinguish it from other varieties of Grantha used in regions of South India, contiguous to
the Tamil area, e.g. Grantha Malayalam (Grünendahl 2001).
3 Some graphemes listed in lines 2 and 3 might, in fact, be changed from one line to the other.
As such, lines 2 and 3 are provisional.
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 207

Table 1: Graphemes of the Grantha and Tamil alphabets (seventh century CE).4

1 Graphemes Grantha alphabet Tamil alphabet

2 Common to both alphabets ṭa, ṇa, ta, na
ya, va
3 Specific in each alphabet initial vowels initial vowels
ña ña
ka, ra ka, ra
pa pa
ma ma
la la
4 Specific to the Grantha kha, ga, gha
alphabet ca, cha, ja, jha
ṭha, ḍa, ḍha
tha, da, dha
pha, ba, bha
ṣa, sa, śa
ha
ṃ, ḥ
5 Specific to the Tamil ca
alphabet ḻa, ḷa, ṟa, ṉa
ḵ

Line 2 shows the graphemes common to both alphabets. Line 3 shows graph-
emes for phonemes common to both languages but differentiated in their re-
spective alphabet. The letters ka and ra, for instance, have a single stroke in the
Tamil alphabet and a double stroke in the Grantha alphabet (see Fig. 1d); the
Grantha pa is larger than the Tamil pa. Line 4 shows graphemes found only in
the Grantha alphabet for specific Sanskrit phonemes not attested in the Tamil
language, for instance, aspirated consonants, and for phonemes not distin-
guished by script in Tamil language, such as voiced and unvoiced consonants.5
Conversely, as shown in line 5, the Tamil alphabet also has exclusive graph-
emes, specifically for Tamil Dravidian phonemes, such as alveolar consonants.

||
4 For the conventions of visualisation of different scripts and languages, see Conventions,
below.
5 The reason why Tamil does not distinguish voiced/unvoiced consonants in its dedicated
script is that this quality is indicated by the immediate environment of the consonant: for
instance, the Tamil t, between two vowels or after n, is a voiced dental, but is unvoiced when
duplicated or at the beginning of a word (unless, in the latter case, the final grapheme of the
preceding word is n or t).
208 | Emmanuel Francis

Furthermore, Grantha also regularly resorts to ligature writing, using con-

junct graphemes, i.e. one single grapheme made by merging the graphemes of
two or more consonants following each other without being vocalised. Such a
feature only occasionally appears in the Tamil alphabet, where it mostly affects
graphemes common to both the Tamil and Grantha alphabets, showing the
probable influence of the Grantha alphabet upon the Tamil alphabet.
The distribution above of graphemes between the two alphabets is, in fact,
tentative and shows the situation in the seventh century CE. In the course of
time, Tamil and Grantha scripts came to be more sharply distinguished. Initial
vowels might not be well differentiated, at least in the early period, and were
clearly differentiated only later. Further differentiation between the two scripts
may be observed quite early in some inscriptions:6
− In the inscription IP 205 (c. 877 CE), the Tamil ṭa is longer than the Grantha
ṭa, which is here found also used in a conjunct grapheme and, as such,
marked as Grantha. Compare Eṭṭāvatu (Fig. 1a) and bhāṭṭanukku (Fig. 1b).
− In the same inscription IP 205, in the word Ācāriryyanēn· (Fig. 1c), the final
n, a grapheme common to both scripts, is marked as Grantha by a super-
script wavy line, which is a virāma (i.e. a vowel-killer, indicating that the
consonant is not vocalised).7
− In the inscription IP 96 (late ninth century CE), in the word vallaṉ āy
uktaṉākiya (Fig. 1d), the form of the vowel sign -u in yu marks this graph-
eme as Grantha, as it has two strokes (instead of one stroke in the Tamil
script). Also note the difference mentioned above between Grantha k (two
strokes) and Tamil script k (a single stroke).
− The inscription IT 4 (c. 1048 CE) shows two types of tu (Fig. 1e): śrīcoḷentra-
siṃhaccatu<r>vvedimaṅkalattu. The first tu is in Grantha and the second in
Tamil script, as indicated by the alternative ways the vowel sign -u is at-
tached to the consonant.

The Tamil script does not normally use conjunct graphemes. But there are also
cases of Tamil conjunct graphemes observed in Tamil words. They concern
graphemes common to both scripts (e.g. n and t) but also Tamil graphemes
proper (e.g. k). Here are a few examples:

||
6 Roman, grey-highlighted roman, and italic, respectively mark graphemes in the Tamil al-
phabet, graphemes in the Grantha alphabet, and graphemes common to the Tamil and Grantha
alphabets.
7 The same wavy line is used in this inscription as a vowel-killer above Tamil graphemes,
instead of the usual puḷḷi (‘dot’). Also see the Centalai inscriptions below.
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 209

− n=ti (Fig. 2a) in the phrase svasti śrī nan=tippō⟨t⟩taraiyarkku (IP 144,
c. 849 CE). This syllable nti can be considered to be written here in Tamil
script as it represents the Tamil pronunciation and notation of Sanskrit ndi.
− k=ku (Fig. 2b) in the word putuk=ku (IP 120, supplementary inscription,
c. 852 CE). This is possibly the first example of a common Tamil conjunct
grapheme. We see here that the single vertical strokes of both k are very
close, that both graphemes seem to share a single horizontal stroke, while
the vowel sign -u surrounds both of them.
− p=pa (Fig. 2c) in the word tirup=pati (IT 30, c. 1241 CE). That this is not a Gran-
tha conjunct grapheme is indicated by the fact that the p is not large and the
two graphemes are merged horizontally (instead of vertically, as in Grantha).
− pa=ṭ=ṭa (Fig. 2d) in the word tēvapa=ṭ=ṭaṉ (IT 30, l. 15, c. 1241 CE).

The coexistence of specific alphabets for Tamil and Sanskrit languages implies
that both languages might be marked by their alphabets in a text written in the
Tamil area. This practice, attested by some of the inscriptional examples above,
has been apparently acknowledged by Tamil grammarians, as early as the elev-
enth century,8 and comes, in the Tamil area, in different configurations.9
On the one hand, there are inscriptions which consist of at least two distinct
portions of text, one in Sanskrit and Grantha, the other in the Tamil language
and script. These are diglossic inscriptions, as there is a division of labour10
between the languages: Sanskrit for the eulogy and Tamil for the business part,
i.e. what Timothy Lubin calls ceremonial diglossia11 and Sheldon Pollock, hy-
perglossia.12 There are also cases of bilinguals, where the Tamil language, on
par with Sanskrit, is also used for the eulogy, which I have suggested calling
amphiglossic (ceremonial) bilinguals.
On the other hand, there are numerous inscriptions that I have called Tamil
mixed-language inscriptions, where Tamil is the main language of the record
whereas Sanskrit loanwords, in a variable quantity, are integrated into the text

||
8 On viraviyal, ‘that which has the nature of mixing’, mentioned in the eleventh-century CE
Tamil grammar called Vīracōḻiyam, see Francis 2021a, 129.
9 The following three paragraphs summarise what is described at length in Francis 2021a.
10 Pollock 2006, 117.
11 Lubin 2013, 411.
12 Such inscriptions are false or non-strict bilinguals, i.e. the two texts in different languages
are not translations one of the other, but differ in content, as opposed to strict bilinguals,
where the two texts in different languages are (almost exact) translations one of the other. The
latter type is virtually inexistent in the Tamil region until the nineteenth century CE.
210 | Emmanuel Francis

and marked as such by the use of Grantha. This is another type of diglossia,
legal13 or technical.
Finally, there is also the language called Maṇippiravāḷam – applied to texts,
especially in Vaiṣṇava commentarial tradition, from the twelfth century CE on-
wards – transmitted in manuscripts, where Tamil and Sanskrit are mixed, with-
out hierarchy. The label Maṇippiravāḷam has also been applied to inscriptions
from the Tamil area. I would restrict this label only to inscriptions where San-
skrit and Tamil are both languages of eulogy, mixed in the same text, i.e. in
cases of amphiglossic (ceremonial) mixed-language inscriptions, but not in
cases of legal or technical diglossia in Tamil mixed-language inscriptions,
where Sanskrit is a term of higher register, i.e. hyperglossic.

3 Tamil inscriptions (600–1300 CE)

We are concerned with the period, starting around 600 CE, when the Grantha
alphabet, specifically designed for writing Sanskrit, appeared and coexisted
with alphabets designed for writing Tamil: an already existing Vaṭṭeḻuttu al-
phabet and a newly-designed Tamil alphabet. This period lasted for several
centuries and also witnessed the progressive assimilation of Sanskrit loanwords
to Tamil phonetics and script, with only a few residual Grantha graphemes still
used.14

3.1 Trichy royal epithets

The first examples of these two new alphabets are possibly to be found, around
600 CE, in the rock-cut cave at Trichy.15
Among the inscriptions of this cave are a series of royal epithets (birudas,
‘glorifying soubriquets’) of the Pallava king Mahendravarman I, distributed on

||
13 Following Lubin 2013.
14 These Grantha graphemes, i.e. ja, śa, ṣa, sa, ha, kṣa and śrī, eventually came to be consid-
ered as Grantha elements of the modern Tamil alphabet, but are found only in Sanskrit loan-
words.
15 Lockwood 2008.
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 211

different parts of the monument. This list is multilingual and multiscriptual, as

we find:16
− Sanskrit royal epithets in Grantha script.
− Tamil royal epithets in Tamil script.
− Tamil royal epithets, exceptionally, in Grantha script.
− Telugu royal epithets in Grantha script, with a specific grapheme for the
Dravidian alveolar ṟa.

3.2 Auspicious beginnings

Similar to any Indian text, an inscription usually starts with an auspicious be-
ginning. Inscriptions in the Tamil region, from the eighth century CE onwards,
whether in Sanskrit or in Tamil, start with the Sanskrit blessing svasti śrī (‘Pros-
perity! Fortune!’) in Grantha (Fig. 3).17 This almost compulsory initial blessing
svasti śrī appears to be typical of South Indian Tamil inscriptions and only occa-
sionally occurs in North Indian inscriptions.
Alternative blessings of the same purport in Tamil script are attested at an
early time. (Āyu)ḷ cīr (IP 262; ninth century CE), for instance, which, in fact, is
made of two Sanskrit loanwords, āyus and śrī, assimilated to Tamil phonetics
and script.
Another pan-Indian Sanskrit blessing formula, śubham astu, became ubiq-
uitous in the Tamil region and practically displaced svasti śrī subsequently,
from the fifteenth century CE onwards.

3.3 Sanskrit loanwords

Tamil inscriptions contain loanwords from Prakrit, and later from Sanskrit, from
the early beginnings. These words are usually assimilated into the Tamil pho-
netics and morphology in the early Tamil-Brāhmī inscriptions.18
However, from the time the Grantha alphabet was invented to write down
Sanskrit texts, we regularly find Sanskrit loanwords written, fully or partially, in
Grantha script in a Tamil inscription, denoting, similar to our modern use of the

||
16 For details, illustrations, and translations, see Francis 2013a, 363–368; Francis 2021a, 112–115.
For another list of royal epithets, bilingual only, and multiscriptual, see the Centalai inscrip-
tions below.
17 Also see IP 144 above (Fig. 2a).
18 See Mahadevan 2003, 109; Mahadevan 2014, 146.
212 | Emmanuel Francis

italics, the awareness of the writer that the word is a loanword. Here are a few
examples of selected Sanskrit loanwords found in such Tamil mixed-language
inscriptions:
− brahmāṇiyār (IP 78, l. 2; c. 755 CE).
− brahmatēyan (IP 171, l. 3; c. 890 CE).
− branmatēyattukku (IP 119, l. 11; c. 852 CE).

In the last two cases, note that the dental in tē is already assimilated into Tamil
script, but there are instances where this is not the case, such as:
− brahmadeyam (SII 19, no. 357, ll. 15–16; late tenth century CE).

3.3.1 Allographs

Various allographs are found in Tamil inscriptions, across places at the same
time or across time in the same place, because, as just seen above, the same
Sanskrit loanword can be instantiated using a variable amount of Grantha
graphemes. Allographs are also attested in the same record by the same hand.
Compare brahmadeyam (l. 50; Fig. 4a) and brahmatēyam (l. 73; Fig. 4b) in the
Bāhūr plates (IP 155, c. 877 CE). Similarly, we find [mahā]devarkku (l. 1),
mātēvaṭikaḷārāna (l. 2), Iddevarkku (l. 2) and Idevaruṭaiya (l. 2) in SII 19, no. 292
(late tenth century CE), all instances of various ways of actualising the Sanskrit
word deva in a Tamil inscription.

3.3.2 Mixing morphology

The examples above additionally show that mixing alphabets and languages
also implies mixing morphology. The root of Sanskrit loanword is normally in
Grantha, whereas the Tamil morphological suffix is in Tamil letters. Thus, the
transition of one script to another is often at the morpheme boundary.
An interesting case is that of Sanskrit personal names of the a-stem declen-
sion. In the following examples, the stem of the Sanskrit loanword in Grantha is
suffixed with the purely Tamil morphological ending of the third person mascu-
line -aṉ, in Tamil letters:
− pākaśācanaṉ = Sanskrit pākaśāsana- (‘punisher of Pāka’, a name of the god
Indra) + Tamil aṉ (SII 3, no. 206, l. 89; tenth century CE). The letter na, at the
transition between Grantha and Tamil scripts, is considered here as a
grapheme common to both alphabets.
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 213

− madhyastaṉ (IP 152, pl. 7v, l. 4 = l. 125; c. 875 CE) = Sanskrit madhyastha
(with aspirated tha changed to non-aspirated ta) + Tamil aṉ.

The first example comes from the Tamil portion of a bilingual ceremonial am-
phiglossic copperplate charter, the second one from the Tamil portion, evincing
technical diglossia, of a bilingual ceremonial diglossic copperplate charter.
Interestingly, we also have, although rarely, examples where the Sanskrit
personal name is fully written in Grantha, including the purely Tamil morpho-
logical suffix of the third person masculine, ‘Granthaised’ as n, a grapheme, in
fact, common to Tamil and Grantha scripts, but marked here as Grantha in con-
trast with the Tamil grapheme ṉ:
− śrīvaran· śrīmanoharan· ciṉac·coḻaṉ· puṉap·pūḻiyaṉ· vītakan·maṣan· vinaya-
viśṛutan· vikramapārakan· vīrapurokan· (Vēḷvikkuṭi plates, EI 17, no. 16,
ll. 98–99; c. ninth century CE). This is a long string of royal epithets of King
Neṭuñcaṭaiyaṉ, where Tamil ones, written in Vaṭṭeḻuttu script19 and ending
with the Tamil ending of the third person masculine -aṉ, alternate with
Grantha ones, written in Grantha and ending with Tamil suffix of the third
person masculine, ‘Granthaised’ as n and followed by a virāma vowel-killer.
This text portion is an example of epigraphical Maṇippiravāḷam from the
Tamil portion of a bilingual amphiglossic copperplate charter.
− bhāṭ=ṭanukku (see IP 205 above and Fig. 1b). The letter nu, at the transition
between Grantha and Tamil scripts, is considered here as a grapheme com-
mon to both alphabets.

In the same Vēḷvikkuṭi plates (EI 17, no. 16, c. ninth century CE) we also find
personal names mixing Tamil and Sanskrit, fully in Grantha, in portions of texts
otherwise fully in Tamil script:
− māravarmman· (plate 4v, l. 2 = l. 48) = Tamil māṟaṉ + Sanskrit varman.
− śrīmāravarmman· (plate 5r, l. 6 = l. 62).

There is an interesting case in Centalai of two mixed-language amphiglossic

inscriptions found each inscribed twice on four pillars. Each inscription is a list
of four royal epithets of a Muttaraiyar chief (eighth century CE; EI 13, no. 10).20

||
19 This is, thus, a case of Vaṭṭeḻuttu and Grantha scripts being mixed, examples of which are
found in the southern parts of the Tamil region, where Vaṭṭeḻuttu subsides. Also see, for in-
stance, a cave inscription at Āṉaimalai (EI 8, no. 33.II).
20 For further details, translations and illustrations, see Francis 2013a, 376–382; Francis 2021a,
114–115.
214 | Emmanuel Francis

The first series reads [śrīmāṟaṉ·] śrīśatrukesarī śrīkaḷvarakaḷvan· śrīAtisāhasan·,

whereas the second series reads śrītamarālayaṉ· śrīAbhimānadhīran· śrīkaḷ·vara-
kaḷ·vaṉ· śrīśatrukesarī. The following traits are noticeable here:
− The virāma on the final Grantha n of the Grantha royal epithets and the puḷḷi
on the final Tamil letter ṉ of the Tamil royal epithets, which are both a simi-
lar wavy line.
− The two royal epithets in the Sanskrit language and Grantha script śrīAtisā-
hasan· and śrīAbhimānadhīran·, which both end with the Tamil personal
suffix of the third personal singular (-aṉ), ‘Granthaised’ as n.
− The Tamil royal epithet kaḷvarakaḷvaṉ found in Tamil script except for the
honorific prefix śrī in Grantha (śrīkaḷ·varakaḷ·vaṉ·), but also written entirely
in Grantha (śrīkaḷvarakaḷvan·), with the alveolar final ṉ of the third person
singular ‘Granthaised’ as n.

3.3.3 Sociolinguistic aspects

Several factors, not mutually exclusive, might explain the choice of the Grantha
alphabet for the Sanskrit loanwords:
− An almost ‘mechanical’ writing of words in their appropriate script – like
we, today, use quotation marks or italics for foreign words – by a copyist
knowledgeable in both scripts and languages.
− A pedantic writing reflected in the script.
− An identity statement by a Brahmin writer, as Sanskrit is the Brahmins’
preserve.
− A reference to a pan-Indian concept otherwise theorised in Sanskrit texts.
This is a diglossic situation (legal, technical) at the level of a Tamil mixed-
language inscription.
− An identity statement of a Tamil writer, in the case of amphiglossic Tamil
mixed-language inscriptions or Maṇippiravāḷam, such as the multilingual
lists of royal epithets, where both Sanskrit and Tamil are used as both fit to
proclaim the grandeur of kings.

Sanskrit loanwords are progressively assimilated into the phonetics and script of
the Tamil receiving language. The Sanskrit word brahmadeya, that we have seen
written as brahmadeyam or brahmatēyam above, could also be and was indeed
written piramatēyam. This spelling spread and became the norm in Tamil inscrip-
tions, but for reasons sometimes unclear, the older spellings brahmadeyam or
brahmatēyam could, across time and place, subside longer or resurface.
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 215

4 Sanskrit inscriptions from the Tamil region

Regarding Sanskrit inscriptions in Grantha script produced in the Tamil region,
they sometimes evince other instances of mixing, but not that of alphabets and
languages.
The spelling of Sanskrit words written in Grantha in a Sanskrit inscription
might evince the influence of the Tamil language and phonetics as they were
composed by a Tamil speaker. This would be a case of mixing phonetics.
The spelling, for example, in a list of royal epithets in Mahābalipuram (IP 39;
seventh century CE), betrays that the mother tongue of the composer is Tamil, as
he writes bāchana for the Sanskrit bhājana and pridhivi for the Sanskrit pṛthivi.21
A current Tamil alteration of Sanskrit pronunciation is reflected in writing
tṣa instead of kṣa. An instance of this is the Bāhūr plates (IP 155; c. 877 CE),
where the writer is, however, not consistent as he writes the Sanskrit phoneme
kṣa, sometimes kṣa and sometimes tṣa.
The substitution of an unvoiced consonant for a voiced consonant (e.g. t for
d before a consonant) is also common:
− kulotbhava instead of kulodbhava (IP 91, l. 5; c. 793 CE and IP 181, l. 11;
c. 893 CE).
− sakara instead of sagara (IP 91, l. 15; c. 793 CE).

There are also cases of biscript monolingual Sanskrit inscriptions, where the same
text is found twice in two different alphabets, the local Grantha and a North Indi-
an alphabet, in a statement of cosmopolitism and claim to universality.22

5 Tamil manuscripts
Let us now jump to a time approximately five hundred years later than the end
of the inscriptional period concerned above23 and deal with practices observed
in Tamil manuscripts, which usually date to the eighteenth century at the earli-

||
21 See EI 10, no. 7, n. 1. For other examples, see the Bāhūr plates (IP 155; c. 877 CE, see EI 18,
no. 2, p. 6).
22 See Francis 2013b; Francis 2021b, 157–160.
23 The period between the fourteenth and the seventeenth century CE is not the area of exper-
tise of the present contributor and certainly deserves a closer study.
216 | Emmanuel Francis

est and, most of the time, to the nineteenth century CE.24 This period was also
marked by polyglossy: Sanskrit, the pan-Indian cosmopolitan language, and
regional languages coexisted and were used and read by the same people. One
often finds, for instance, complete Sanskrit sentences, blessings or colophons,
i.e. scribal paratexts attesting the copyist’s knowledge of these two languages
and their respective scripts, in Tamil manuscripts.
However, cases of legal or technical diglossia comparable to those dis-
cussed above from Tamil mixed-language inscriptions are only rarely found:
Sanskrit loanwords appear to have been fully assimilated into Tamil phonetics
and script; their Sanskrit origin is now unrecognisable from their script. Grantha
and Tamil scripts are also more differentiated than at their origins.
One novelty that can be observed is the creeping of the Grantha grapheme
m, so far not observed in earlier inscriptions, but possibly in inscriptions con-
temporary to the manuscripts, into the writing of Tamil texts, in the form of
conjunct graphemes.

5.1 Final m
The main phenomenon observed concerns the grapheme for the final m in
words, which are mostly Sanskrit loanwords, but also, by extension, in Tamil
words. Basically, there are three options for writing the ending m in a Sanskrit
loanword for a copyist:
− Tamil letter m, as an independent grapheme.
− Grantha letter m, as an independent grapheme.
− Grantha letter m, as the final element of a conjunct grapheme, the first ele-
ment being a Tamil grapheme.

The final m in Tamil manuscripts can, thus, be variously instantiated, besides

the expected standard Tamil letter m.

5.1.1 Independent final Grantha m

Here are a few instances of the use of a final Grantha m, not in a conjunct but as
an independent final grapheme, with or without an ascending stroke, that could

||
24 The manuscripts considered here are mostly on palm leaves, but there are also some paper
manuscripts.
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 217

be considered a virāma (vowel-killer), meaning that the consonant is plain,

without vocalisation:
− śrīrā=maceyam· (BnF Indien 1039, U2b, fol. 167r, left margin; Fig. 5a).
− virut=tam (BnF Indien 303, U2c, fol. 77v, l. 2; Fig. 5b).
− tirucciṟṟampalam· (BnF Indien 303, U2c, fol. 77v, l. 2; Fig. 5c).

Note that this final Grantha m is used mostly, if not exclusively, in paratexts,
such as scribal blessings (in the margins or at junctures of the root text, such as
ends of chapters or colophons) or total verses (at junctures of the root text), but
rarely in the root text itself.

5.1.2 Tamil-Grantha conjuncts with a final Grantha m

Another option is to use a Tamil-Grantha conjunct grapheme, where the final m

is appended below the base of the first (Tamil) grapheme of the conjunct. This
appended final m, noted here as =m, appears, from its physical aspect, to be, in
fact, a Grantha ma. It should, as pointed out to me by Charles Li, be read as m,
i.e. even in the absence of an explicit vowel-killer.
There are also cases where, similar to the independent final Grantha m (Sec-
tion 5.1.1 above), the final m bears what could be interpreted as a vowel-killer in
the form of an ascending stroke. In the absence of an explicit vowel-killer, it is,
however, sometimes to be read ma and not m, as we will see below (Section 5.2).
Here are some commonly found examples of Tamil-Grantha conjunct
graphemes ending with a below-base Grantha letter m, as compared to the
standard Tamil script graphemes:
− ṇam vs ṇa=m (Fig. 6a and Fig. 6b).
− ṉam vs ṉa=m (Fig. 6c and Fig. 6d).
− yum vs yu=m (Fig. 6e and Fig. 6f).
− yam vs ya=m (Fig. 6g and Fig. 6h).

Note that yu=m and ya=m are basically the same grapheme that can be read yum
(in pure Tamil words, e.g. aṟiyu=m) or yam (in Sanskrit loanwords, e.g. ceya=m).
Other Tamil-Grantha conjunct graphemes attested are, for instance, ka=m,
k=ka=m, ta=m, t=ta=m, tu=m, mu=m, ra=m, la=m, vu=m and ḷa=m. There
might be more, not yet observed.25

||
25 See growing list of examples on <https://tst-project.github.io/palaeography/below-base-
ligatures> (accessed on 20 July 2023).
218 | Emmanuel Francis

5.1.3 Allographs

As such variant spellings coexist with standard Tamil spellings, there are sever-
al allographs for certain words. The syllable ṇam, for instance, could be written:
− ṇam (two graphemes, fully in Tamil script).
− ṇa=m (Tamil-Grantha conjunct grapheme).
− ṇam (one Tamil grapheme + Grantha m without virāma).
− ṇam· (one Tamil grapheme + Grantha m with virāma).

The Tamil word caraṇam, from Sanskrit śaraṇa, ‘shelter, refuge, protection’, for
instance, encountered frequently in the blessing formula X caraṇam, is found
variously spelled, with further spelling options, such as ṟa for ra, as follows:
− caraṇam (BnF Indien 246, U2, fol. 77v, l. 7 and BnF Indien 362, U3, fol. 17r,
col. 3, l. 3; Fig. 7a and Fig. 7b).
− caṟaṇam (BnF Indien 362, U3, fol. 17r, col. 2, l. 2; Fig. 7c). Note that this is by
the same hand as that of the second of the two examples above, in other
words, an inconsistent hand.
− caṟaṇa=m (BnF Indien 89, U2a, fol. 1r, l. 3; Fig. 7d).
− caraṇam=௳ (BnF Indien 247, U2, fol. 1v, l. 5; Fig. 7e), with piḷḷaiyār cuḻi
merged with the final Grantha m.

The same could be shown for other Sanskrit loanwords, in blessings adapted
from Sanskrit. Sanskrit sakāya (‘support’) in the X-cakāyam blessing formula, or
Sanskrit kaṭākṣa (‘glance, side look’, i.e. ‘grace’) in the X-kaṭākṣam blessing
formula (see examples below) are instances.

5.2 rā=m / rā=ma conjunct grapheme

The grapheme rā=ma is a case of a Tamil-Grantha conjunct grapheme where the
appended m should be read vocalised, that is ma and not m. We read this
grapheme, which looks like two merged Tamil kāls26 + Grantha ma attached
below its base, as rāma. This could alternatively be read rām, but in all its oc-
currences, it makes more sense to read rāma rather than rām, except when it is
followed by a medial ā, so as to note rāmā (see below).

||
26 The kāl is an ambiguous Tamil grapheme, which represents either the medial vowel ā or the
consonant r. Only the context indicates the relevant reading.
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 219

In practice, this conjunct grapheme rā=ma has been found so far almost ex-
clusively for the word rāma, as in the blessing śrīrāmaceyam (from Sanskrit śrī +
rāma + jaya), for which there are numerous variant spellings.27 Several factors
explain why this blessing is found spelt with so many possible variants, as the
copyist could use:
− The conjunct rā=ma.
− A final independent Grantha m, with or without virāma, to write yam.
− The final below-base Grantha m in a conjunct grapheme, with or without
virāma, to write the same yam.
− Several variant Tamil spellings for Sanskrit jaya.
− The allograph ṟ for r.
− The double kāl merged for rā.

What follows is just a sample of the various possible spellings:28

− śrīrāmaceyam (BnF Indien 390, U2c, fol. 2v, l. 2; Fig. 8a), an example of full
assimilation to the Tamil script except for the Grantha śrī.
− śrīrā=maceyam (BnF Indien 265, U2b, fol. 601r, l. 2; Fig. 8b), where the con-
junct grapheme rā=ma is used.
− śrīrā=maceya=m=௳ (BnF Indien 3, U2c, fol. 14v, l. 5; Fig. 8c), where the
conjunct graphemes rā=ma and ya=m are used, the latter being merged
with the punctuation in the form of a piḷḷaiyār cuḻi.
− śrīrā=maceyam (BnF Indien 1037, fol. 1v, left margin; Fig. 8d), where the
conjunct grapheme rā=ma and the final Grantha m are used.
− śrīrā=maceyam· (BnF Indien 1039, U3, fol. 167r, left margin; Fig. 8e), where
the conjunct grapheme rā=ma and the final independent Grantha m with a
vertical stroke, possibly functioning as a vowel-killer, are used.
− śrīṟāmaceyam (BnF Indien 143, fol. 1r, l. 1; Fig. 8f), where the Tamil graph-
eme ṟā is used instead of the Tamil grapheme rā.
− śrīṟāmajayam (BnF Indien 431, U2a, fol. 1r, col. 1, l. 1; Fig. 8g), where the
Tamil grapheme ṟā is used instead of the Tamil grapheme rā, whereas the
Grantha ja is kept.
− śrīṟāmajeyam (BnF Indien 143, fol. 271r, l. 1; Fig. 8h), where the Tamil
grapheme ṟā is used instead of the Tamil grapheme rā, the Grantha j is kept,
but vocalised -e instead of -a.

||
27 BnF Indien 143, Ariel papers, made of papers by various hands bound in one volume, used
here, provides a good sample of various possible spellings.
28 Furthermore, the final grapheme can be merged with the following punctuation.
220 | Emmanuel Francis

− śrīrāmaceyam (BnF Indien 143, fol. 464r, l. 1; Fig. 8i), where the allograph rā
(double kāl fused instead of two discrete successive kāls) is used.

As in the case of epigraphical allographs, such allographs can be found in the

same manuscript by the same hand. This might be due to the inconsistency of
the copyist, but one can wonder if sometimes allographs were not used on pur-
pose. We find this blessing śrīrāmaceyam three times in succession, each with a
different spelling, for instance, in BnF Indien 947, U4, fol. 112r, l. 3: śrīrāma-
ceya=m śrīrā=maceyam śrīrā=maceya=m. Perhaps the copyist used three differ-
ent spellings here on purpose, so that the blessing would be more operative by
being repeated three times in three variant spellings.
There are other instances where the very same grapheme is not to be read
rā=ma but rā=m instead, that is, when it is followed by a medial -ā. We have
found this so far only for the name Rāmānuja and for the name Rāma in its voc-
ative form rā=mā, as in the following example: śrīrā=m=ā (BnF Indien 335, U5,
fol. 476r, left margin; Fig. 9), where the medial -ā looks like a part of the whole
conjunct (thus, preceded here by ‘=’).

5.3 Grantha-Tamil conjunct graphemes

The examples above can be considered as regular Tamil-Grantha conjunct
graphemes. As it happens, there are further, rare or unusual, examples of con-
junct graphemes where these two distinct alphabets are also mixed, but in the
reverse order, i.e. the first element of the conjunct is a Grantha grapheme and
the below-base element is a Tamil grapheme. This practice concerns, unexpect-
edly, Sanskrit loanwords. Note that the Grantha graphemes involved in the
examples observed so far are among those to become Grantha elements of the
modern Tamil alphabet. This is as if the copyist, unfamiliar with the full-fledged
Grantha script, was aware of the Sanskrit origin and reflected this awareness in
the script, using a Grantha grapheme still commonly used in Tamil writing,
even considered as a Tamil grapheme (a kind of loan grapheme), to create a
Grantha-Tamil conjunct grapheme. Here are two examples:
− ṣ=ca for Grantha kṣa, usually noted ṭca in Tamil script. Compare kaṭāṣ=cam
(BnF Indien 294, U2, fol. 1r, col. 2, l. 3; Fig. 10a) and kaṭāṣcam, with two suc-
cessive graphemes instead of a Grantha-Tamil conjunct grapheme (BnF In-
dien 237, U3, fol. 211r, l. 5; Fig. 10b).
− s=cu for Grantha su: s=cuvāmi (BnF Indien 28, U3, fol. 181r, l. 2; Fig. 10c).
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 221

The existence of these further Grantha-Tamil conjuncts imply further cases of al-
lography, involving, besides the final m, the spelling of kṣa, as in the X-kaṭākṣam
blessing formula, where the Sanskrit kaṭākṣa is found variously spelled, for exam-
ple, kaṭāṭcam, kaṭāṭcam, kaṭāṣcam, kiṭāṣcam, kiṭāṣṣam, kaṭāṣṣam or kaṭāṣ=cam.

5.4 Vowel-killer virāma

Another interesting phenomenon is the use of a t-like virāma (Fig. 11) with the
Grantha grapheme s (one of those that had become integrated in the modern
Tamil alphabet). This specific type of virāma (vowel-killer) looks like a below-
base t, but some examples indicate it could indeed be a virāma.
In a first example, this grapheme looks exactly like a t, and the reading is
uncertain:
− pos·takattukku or posttakattukku (BnF Indien 339, U1, fol. 1r, l. 3; Fig. 11a).

In other examples, it does not look so much like a t, and we are inclined to read
it as a virāma, as reading a t would amount to read three ts successively, which
is improbable:
− namas· ttu NOT namas tttu (BnF Indien 449, U1, fol. 1r, l. 3; Fig. 11b).
− Akas·t=tiyar NOT Akastt=tiyar (BnF Indien 112, U2, fol. 1r, col. 1; Fig. 11c).

The reading as t, given the environment, does not make much sense in two fur-
ther examples, although it might not be precluded:
− Aḻakēs·paraṉ rather than Aḻakēstparaṉ (BnF Indien 417, U2, fol. 1r, l. 1; Fig. 11d).
− tecamas·kkantam rather than tecamastkkantam (BnF Indien 256, U2, fol. 276v,
l. 4; Fig. 11e).

5.5 Sanskrit phrases in the Tamil alphabet

As has been mentioned above, given the polyglossic milieu of copyists, Sanskrit
phrases occasionally appear in otherwise fully Tamil manuscripts.
The blessing hariḥ Om is a telling example as it has many spelling variants
which show different mixings of Grantha and Tamil graphemes. From the origi-
nal standard Sanskrit formula hariḥ Om (with typical Grantha r and final m) to
the standard Tamil adaptation hari Ōm (where only the Grantha grapheme ha
subsists, now integrated into the Tamil script), there are, in between intermedi-
ary allographs, with or without visarga (the final Sanskrit phoneme ḥ), or with
hariḥ in Grantha and Ōm in Tamil, as shown in the following examples:
222 | Emmanuel Francis

− hariḥ Om (BnF Indien 291, U2, fol. 1r, left margin; Fig. 12a).
− hariḥ Ōm (BnF Indien 433, U2, fol. 1r, left margin; Fig. 12b), where r is Gran-
tha (as it has a middle dot) and Ōm alone is in the Tamil alphabet.
− hari Om (BnF Indien 294, U4a, fol. 1r, left margin; Fig. 12c), where r appears
to be in Tamil script and ḥ is dropped.

Sanskrit namaskāra (‘homage’) formulae are also found entirely or almost en-
tirely (with a few Grantha graphemes only) written in Tamil script, as in the
following examples:
− viṉāyakāya namam, with Grantha ma (BnF Indien 983, U2, fol. 1r, left mar-
gin; Fig. 13a), i.e. Sanskrit vināyakāya namaḥ.
− śrīmatē ramāṉucāya nama namaṣ ttu, with a Grantha śrī and ṣ (whereas a
Grantha s would have been more faithful to the original Sanskrit word) (BnF
Indien 381, U3c, fol. 26v, col. 1, l. 1; Fig. 13b), i.e. Sanskrit śrīmate rāmānujāya
namas tu.
− Ōm catācivāya nama (BnF Indien 516, U2, fol. 2v, col. 2; Fig. 13c), i.e. San-
skrit Oṃ sadāśivāya namaḥ.
− kuruppiyō nama (BnF Indien 307, U2a, fol. 1v, l. 4; Fig. 13d), i.e. Sanskrit
gurubhyo namaḥ.

Was this due to copyists not having the command of the Grantha script and
using the Tamil script to write down a Sanskrit phrase as they heard it pro-
nounced? Or was it an early statement of Tamil regionalism of a copyist assimi-
lating Sanskrit phrases to Tamil script?

6 Sanskrit manuscripts from the Tamil region

Finally, mention must be made of a phenomenon comparable to the conjunct
grapheme rā=ma discussed above, i.e. the use of the Grantha conjunct graph-
eme n=ma for writing the Sanskrit word namas. This is regularly found in San-
skrit manuscripts from the Tamil region (and possibly from other parts of South
India) and, occasionally, in Tamil manuscripts which include Sanskrit bless-
ings. Compare the following two allographs examples:
− śrīvedavyāsāya namaḥ ௳ (BnF Sanscrit 290, U4, fol. 342r, l. 5; Fig. 14a)
without the conjunct grapheme.
− śrīmahāsarasvatyai n=maḥ ௳ (BnF Sanscrit 290, U3, fol. 1r, l. 25; Fig. 14b)
with the conjunct.
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 223

7 Conclusion
Various phenomena of mixing have been observed from the provisional survey
of two periods in the history of writing Tamil texts containing Sanskrit loan-
words above. The mixing of languages and alphabets, as well as the mixing of
morphology, occur in bilingual diglossic/amphiglossic inscriptions, in diglossic
mixed-language inscriptions and in amphiglossic mixed-language inscriptions.
The mixing of alphabets, but not of languages, occurs in biscript monolingual
inscriptions, for Tamil conjuncts in Tamil inscriptions (influenced by Grantha
practices) and for Tamil-Grantha conjuncts in manuscripts. The existence of
many allographs, variant spellings of the same word, are attested in inscrip-
tions, for several centuries, before the progressive assimilation of Sanskrit
loanwords to Tamil phonetics was achieved (with the conservation of a set of
specific Grantha graphemes for Sanskrit loanwords, up to, let us say provision-
ally, circa 1300 CE, but inscriptional practices after this date still have to be ex-
plored in depth). Manuscripts of the eighteenth and nineteenth century CE also
show a great variety of allographs. In both cases, inconsistencies by the same
hand have been observed: the same inscriptional writer may use various spell-
ings for the same Sanskrit loanword in the same record, whereas a manuscript
copyist may use different spellings of the same word or syllable in the very same
manuscript.
Regarding our present editorial practices for inscriptions and manuscripts
in the era of digital humanities, it seems important not only to record such phe-
nomena in the metadata but also to encode these in our edited texts, so as to be
able to make quantitative and qualitative studies. Such an effort is currently
being made in the framework of two research projects.
The project ‘The Domestication of “Hindu” Asceticism and the Religious
Making of South and Southeast Asia’ (DHARMA) has developed transliteration
and encoding conventions that reflect the script peculiarities in the editions of
inscriptions. Marking up Grantha graphemes in Tamil mixed-language inscrip-
tions will hopefully one day make it possible to study the progressive assimila-
tion of Sanskrit loanwords to Tamil phonetics and script across time and place.
The project ‘Texts Surrounding Texts: Satellite Stanzas, Prefaces and Colo-
phons in South-Indian Manuscripts’ (TST) similarly not only developed translit-
eration and encoding conventions for Tamil script phenomena in Tamil manu-
scripts, but also started to provide, thanks to Charles Li, specifically designed
224 | Emmanuel Francis

Tamil font graphemes for script phenomena, such as Tamil-Grantha conjuncts

with a final Grantha m.29
The road is still long, but, at least, the path and direction are taken.

Acknowledgements

This paper, written at the invitation of the editors of this volume, is a result of
the projects DHARMA, ‘The Domestication of “Hindu” Asceticism and the Reli-
gious Making of South and Southeast Asia’ and TST, ‘Texts Surrounding Texts:
Satellite Stanzas, Prefaces and Colophons in South-Indian Manuscripts’ (collec-
tions of the Paris BnF and Hamburg Stabi). The DHARMA project (2019–2025)
has received funding from the European Research Council under the European
Union’s Horizon 2020 research and innovation programme (grant agreement no.
809994). The TST project (2019–2023) has received funding from the ANR and
the DFG in the framework the FRAL programme (Franco-German research pro-
jects in humanities and social sciences).
Thanks to Charles Li for sharing with me many insights about the peculiari-
ties of Tamil alphabet and for commenting upon a draft of this paper.

Conventions
The original texts are provided in transliteration – so as to clearly demarcate
them and use the distinction between roman, grey-highlighted roman, and italic
– according to the following transliteration conventions:

Roman for graphemes in the Tamil alphabet

Roman (grey-scale for graphemes in the Grantha alphabet
highlighted)
Italic for graphemes common to the Tamil and Grantha alphabets
Upper-case (vowel) initial vowel
C=C Tamil conjunct consonant (‘C’ for consonant, ‘=’ to demarcate the
consonants fused in the conjunct grapheme; e.g. k=ku, t=ta)
=m / =ma below-base Grantha consonant m or ma in a conjunct grapheme
(abc) graphemes not clearly legible
[abc] lost graphemes supplied by conjecture

||
29 See <https://tst-project.github.io/editor/entities.html> (accessed on 20 July 2023), under
construction. These graphemes are used in the Tamil display of the TST catalogue <https://tst-
project.github.io> (accessed on 20 July 2023).
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 225

⟨abc⟩ graphemes omitted by the copyist and supplied

௳ piḷḷaiyār cuḻi (short and long forms)
· explicit puḷḷi or virāma

The BnF manuscripts are referred to above by their current BnF shelfmarks, that
is, ‘Indien’ or ‘Sanscrit’, followed by accession number. In the case where the
manuscript is a composite manuscript, a letter follows immediately the acces-
sion number.

Abbreviations
BnF = Paris, Bibliothèque nationale de France.
EI = Epigraphia Indica, 42 vols (1892‒1992), New Delhi: Archaeological Survey of India.
IP = Inscriptions of the Pallavas; see Mahalingam 1988.
IT = Inscriptions of Tirunaḷḷāṟu; see Viyavenugopal 2017.
SII = South Indian Inscriptions, 27 vols (1890‒2001), New Delhi: Archaeological Survey of India.
U = unit of a manuscript, followed by a number or a number + letter to indicate the codicologi-
cal/textual unit (e.g. ‘U1’, ‘U2a’ and ‘U2b’).

References
Franceschini, Marco (2022), ‘A Modular Framework for the Analysis of the Dates Found in Man-
uscripts Written in the Tamil and Tamilian Grantha Scripts’, in Nalini Balbir and Giovanni
Ciotti (eds), The Syntax of Colophons: A Comparative Study across Pothi Manuscripts
(Studies in Manuscript Cultures, 27), Berlin: De Gruyter, 171–207.
Francis, Emmanuel (2013a), ‘Praising the King in Tamil during the Pallava Period’, in Whitney
Cox and Vincenzo Vergiani (eds), Bilingual Discourse and Cross-cultural Fertilisation:
Sanskrit and Tamil in Medieval India, Puducherry: Institut français de Pondichéry / École
française d’Extrême-Orient, 359–409.
Francis, Emmanuel (2013b), ‘North Indian Scripts & South Indian Kings’, available at
<https://rcsi.hypotheses.org/288> (accessed on 30 March 2023) [first published 1 Febru-
ary 2013].
Francis, Emmanuel (2021a), ‘Multilingualism in Indian Inscriptions: With Special Reference to
Inscriptions of the Tamil Area’, in Giovanni Ciotti and Erin McCann (eds), Linguistics and
Textual Aspects of Multilingualism in South India and Sri Lanka, Puducherry: Institut fran-
çais de Pondichéry / École française d’Extrême-Orient, 42–163.
Francis, Emmanuel (2021b), ‘Imperial Languages and Public Writings in Tamil South India: A
Bird’s Eye View in the very Longue Durée’, in Peter C. Bisschop and Elizabeth A. Cecil (eds),
Primary Sources and Asian Pasts (Beyond Boundaries, 8), Berlin: De Gruyter, 152–183.
Grünendahl, Reinhold (2001), South Indian Scripts in Sanskrit Manuscripts and Prints: Grantha
Tamil, Malayalam, Telugu, Kannada, Nandinagari, Wiesbaden: Harrassowitz.
226 | Emmanuel Francis

Lockwood, Michael (2008), ‘The Creation of Pallava Grantha Tamil Script’, in R. Kalaikkovan et
al. (eds), Airāvati: Felicitation Volume in Honour of Iravatham Mahadevan, Chennai: Vara-
laaru.com, 77‒110.
Lubin, Timothy (2013), ‘Legal Diglossia: Modeling Discursive Practices in Premodern Indic Law’,
in Whitney Cox and Vincenzo Vergiani (eds), Bilingual Discourse and Cross-cultural Fertili-
sation: Sanskrit and Tamil in Medieval India, Puducherry: Institut français de Pondichéry /
École française d’Extrême-Orient, 411‒455.
Mahadevan, Iravatham (2003), Early Tamil Epigraphy: From the Earliest Times to the Sixth
Century A.D., 1st edn (Harvard Oriental Series, 62), Chennai: Cre-A / Harvard: Harvard Uni-
versity Press.
Mahadevan, Iravatham (2014), Early Tamil Epigraphy, vol. 1: Tamil-Brāhmī Inscriptions, 2nd
edn, Chennai: Central Institute of Classical Tamil.
Mahalingam, Therazhundur Venkataraman (1988), Inscriptions of the Pallavas, New Delhi:
ICHR and Agam Prakashan.
Pollock, Sheldon I. (2006), The Language of the Gods in the World of Men: Sanskrit, Culture,
and Power in Premodern India, Berkeley: University of California Press.
Viyavenugopal, G. , . (2017), ் ் ்
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 227

Appendix: Images
All photos of BnF manuscripts, either retrieved from Gallica (<https://gallica.
bnf.fr>) or taken by Emmanuel Francis, by courtesy of BnF.

a b

c d

Figs 1a–e: Graphemes differing between Grantha and Tamil alphabets from inscriptions: (a) the
word Eṭṭāvatu at Uttiramērūr c. 877 CE; IP 205, l. 1 (© E. Francis); (b) the word bhāṭṭanukku at
Uttiramērūr c. 877 CE; IP 205, l. 4 (© E. Francis); (c) the word Ācāriryyanēn· at Uttiramērūr c. 877 CE;
IP 205, l. 7 (© E. Francis); (d) the word vallaṉ āy uktaṉākiya at Tirunaḷḷāṟu, late ninth century CE; IP 96,
l. 2 (© E. Francis); (e) the word śrīcoḷentrasiṃhaccatu<r>vvedimaṅkalattu at Tirunaḷḷāṟu, c. 1048 CE;
IT 4, ll. 6–7 (© Babu N. Ramaswamy).
228 | Emmanuel Francis

a b

c d

Figs 2a–d: Tamil conjunct graphemes from inscriptions: (a) the phrase svasti śrī nan=tippō⟨t⟩tarai-
yarkku at Guḍimallam, c. 849 CE; IP 144, ll. 1–3 (© Valérie Gillet); (b) the word putuk=ku at Vanta-
vāci, c. 852 CE; IP 120, supplementary inscription, l. 2 (© Valérie Gillet); (c) the word tirup=pati
at Tirunaḷḷāṟu, c. 1241 CE; IT 30, line 12 (© Babu N. Ramaswamy); (d) the word tēvapa=ṭ=ṭaṉ IT 30,
line 12, c. 1241 CE; IT 30, l. 15 (© Babu N. Ramaswamy).

Fig. 3: Initial blessing svasti śrī at Kōṉērirājapuram, late tenth century CE; SII 3, no. 146, l. 1
(© E. Francis).
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 229

a b

Figs 4a–b: Allographs in the Bāhūr plates: (a) the word brahmadeyam, c. 877 CE; IP 155, l. 50;
(b) the word brahmatēyam, c. 877 CE; IP 155, l. 73 (courtesy of BnF).

a b c

Figs 5a–c: Independent final Grantha m in manuscripts: (a) the phrase śrīrā=maceyam·; BnF
Indien 1039, U2b, fol. 167r, left margin; (b) the word virut=tam; BnF Indien 303, U2c, fol. 77v, l. 2;
(c) the word tirucciṟṟampalam·; BnF Indien 303, U2c, fol. 77v, l. 2 (courtesy of BnF).

a b

c d

e f

g h

Figs 6a–h: Tamil-Grantha conjunct graphemes with appended final Grantha m in manuscripts:
(a) ṇam; (b) ṇa=m; (c) ṉam; (d) ṉa=m; (e) yum; (f) yu=m; (g) yam; (h) ya=m (courtesy of BnF).
230 | Emmanuel Francis

a b c

d e
Figs 7a–e: Allographs of the word caraṇam (standard form) in manuscripts: (a) the word caraṇam in
BnF Indien 246, U2, fol. 77v, l. 7 (courtesy of BnF); (b) the word caraṇam in BnF Indien 362, U3, fol. 17r,
col. 3, l. 3 (© E. Francis); (c) the word caṟaṇam in BnF Indien 362, U3, fol. 17r, col. 2, l. 2 (© E. Fran-
cis); (d) the word caṟaṇa=m in BnF Indien 89, U2a, fol. 1r, l. 3 (courtesy of BnF); (e) the word caraṇam
in BnF Indien 247, U2, fol. 1v, l. 5 (courtesy of BnF).

a b

c d

e f g

h i

Figs 8a–i: Allographs of the phrase śrīrāma in manuscripts: (a) the phrase śrīrāmaceyam in
BnF Indien 390, U2c, fol. 2v, l. 2 (courtesy of BnF); (b) the phrase śrīrā=maceyam in BnF Indien 265,
U2b, fol. 601r, l. 2 (courtesy of BnF); (c) the phrase śrīrā=maceya=m௳ in BnF Indien 3, U2c, fol. 14v,
l. 5 (courtesy of BnF); (d) the phrase śrīrā=maceyam in BnF Indien 1037, fol. 1v, left margin (© E.
Francis); (e) the phrase śrīrā=maceyam· in BnF Indien 1039, U3, fol. 167r, left margin (courtesy
of BnF); (f) the phrase śrīṟāmaceyam in BnF Indien 143, fol. 1r, l. 1 (© E. Francis); (g) the phrase
śrīṟāmajayam·in BnF Indien 431, U2a, fol. 1r, col. 1, l. 1 (© E. Francis); (h) the phrase śrīṟāmajeyam·
in BnF Indien 143, fol. 271r, l. 1 (© E. Francis); (i) the phrase śrīrāmaceyam in BnF Indien 143, fol. 464r,
l. 1 (© E. Francis).
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 231

Fig. 9: The phrase śrīrā=mā in BnF indien 335, U5, fol. 476r, left margin (courtesy of BnF).
232 | Emmanuel Francis

a b

Figs 10a–c: Tamil-Grantha conjunct graphemes in manuscripts: (a) the word kaṭāṣ=cam in BnF
Indien 294, U2, fol. 1r, col. 2, l. 3 (© E. Francis); (b) the word kaṭāṣcam in BnF Indien 237, U3,
fol. 211r, l. 5 (courtesy of BnF); (c) the phrase śrīkumārascuvāmi in BnF Indien 28, U3, fol. 181r, l. 2
(courtesy of BnF).

Figs 11a–e: Vowel-killer (virāma) in manuscripts: (a) the word pos·takattukku or posttakattukku
in BnF Indien 339, U1, fol. 1r, l. 3 (courtesy of BnF); (b) the phrase namas· ttu NOT namas tttu in
BnF Indien 449, U1, fol. 1r, l. 3 (© E. Francis); (c) the word Akas·t=tiyar NOT Akastt=tiyar in BnF
Indien 112, U2, fol. 1r, col. 1 (© E. Francis); (d) the word Aḻakēs·paraṉ rather than Aḻakēstparaṉ
in BnF Indien 417, U2, fol. 1r, l. 1 (courtesy of BnF); (e) the phrase tecamas·kkantam rather than
tecamastkkantam in BnF Indien 256, U2, fol. 276v, l. 4 (courtesy of BnF).
 Mixing Languages and Scripts in Tamil Inscriptions and Manuscripts | 233

a b c

Figs 12a–c: Allographs of hariḥ Om in manuscripts: (a) the phrase hariḥ Ōm in BnF Indien 291,
U2, fol. 1r, left margin (courtesy of BnF); (b) the phrase hariḥ Ōm in BnF Indien 433, U2, fol. 1r,
left margin (© E. Francis); (c) the phrase hari Om in BnF Indien 294, U4a, fol. 1r, left margin
(courtesy of BnF).

a b

Figs 13a–d: Sanskrit homage formulae (namaskāra) in Tamil alphabet in manuscripts: (a) the
phrase viṉāyakāya namam in BnF Indien 983, U2, fol. 1r, left margin (courtesy of BnF); (b) the
phrase śrīmatē ramāṉucāya nama namaṣ ttu in BnF Indien 381, U3c, fol. 26v, col. 1, l. 1 (courte-
sy of BnF); (c) the phrase Ōm catācivāya nama in BnF Indien 516, U2, fol. 2v, col. 2 (© E. Fran-
cis); (d) the phrase kuruppiyō nama in BnF Indien 307, U2a, fol. 1v, l. 4 (courtesy of BnF).
234 | Emmanuel Francis

Figs 14a–b: Allographs nama and n=ma in manuscripts: (a) the phrase śrīvedavyāsāya namaḥ ௳ in
BnF Sanscrit 290, U4, fol. 342r, l. 5 (© E. Francis); (b) the phrase śrīmahāsarasvatyai n=maḥ ௳ in
BnF Sanscrit 290, U3, fol. 1r, l. 25 (© E. Francis).
Peera Panarut, Manasicha Akepiyapornchai
Sanskrit Prayers in a Theravada Kingdom: A
Multilingual Siamese Grantha Manuscript
from Munich
Abstract: This article explores the multilingual character of a Siamese Grantha
manuscript preserved at the Bayerische Staatsbibliothek in Munich (Cod. siam 99).
The manuscript, presumably dated between the late nineteenth and the early
twentieth centuries, is the only known Siamese Grantha manuscript that is kept
in Germany. Manuscript Cod. siam 99 collects a variety of Sanskrit mantras,
written in the South-Indian-derived Siamese Grantha, as its title Wet Tang Tang
(‘Various mantras’) suggests. In addition, ritual instructions in Thai consistently
accompany their Sanskrit counterparts throughout the manuscript; the multi-
lingual dynamic between Sanskrit and Thai is, thus, a salient feature of this
manuscript. The manuscript, therefore, deserves more scholarly attention, and
its texts should be published in the future.

1 Introduction
The traditional writing culture of Thailand – the country that, until 1939, was
known as Siam – has long been dominated by Theravada Buddhism, much like
its neighbours in mainland South East Asia: Laos, Burma and Cambodia. Pali,
an Indo-European language and the canonical idiom of Theravada Buddhism,
was commonly used in various Buddhist manuscript cultures throughout this
region, becoming the religious lingua franca of the Theravada world, or what
Steven Collins calls the ‘Pali imaginaire’:1 the trans-local and -temporal premod-
ern world of Buddhism, which prospered amid the coexistence of diverse ver-
nacular languages from different linguistic families.2 The Khòm script, which
developed from the Old Khmer script in Thailand, has mainly been employed in
the Siamese manuscript culture of central and southern Thailand for preserving

||
1 Collins 1998.
2 For example, the Thai (the official language of modern Thailand), Tai Lan Na (from northern
Thailand), Shan (from north-eastern Burma) and Lao languages from the Tai-Kadai (Kra-Dai)
ethnolinguistic group; Burmese from Tibeto-Burman; and Khmer and Mon from the Aus-
troasiatic group. See further in Goddard 2005.

Open Access. © 2024 the authors, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-009
236 | Peera Panarut, Manasicha Akepiyapornchai

religious texts, especially those written in Pali, and has, thus, been perceived as
the sacred script of the Siamese Buddhist tradition.3 On the other hand, the Thai
script has been used for writing vernacular Thai from the early phase of Thai
literacy in the thirteenth century to the present day.
Despite the dominance of Theravada Buddhism, the Brahmanical tradition
continued to be practiced in Thailand for many centuries. Brahmanical priests
conducted a number of royal ceremonies and rituals alongside Buddhist monks
in the Siamese royal court. These Brahmanical rituals aimed at establishing the
sacred identity of the king and the hierarchy between the king and his subjects.4
Sanskrit texts were, therefore, also uttered by Brahmanical priests in Siam. After
the fall of the Siamese kingdom in Ayutthaya in 1767, the Brahmanical lineage
descending from the Angkorian Khmer was believed to have ended. Therefore,
after King Rama I founded Bangkok as the new capital of the later Siamese
kingdom in 1782, Brahmanical priests from Nakhòn Si Thammarat (southern
Thailand) were recruited to Bangkok to perform Brahmanical rituals at the royal
court.5 Due to their strong connection with the Brahmanical tradition in Tamil
Nadu, priests from the south brought with them their specific script, originating
from South Indian Grantha, most possibly Grantha Tamil, which has been wide-
ly used to write Sanskrit texts in Tamil Nadu. This Grantha script used among
the Siamese Brahmanical priests is known in Thai as chiang khrün or chiang
phram script, which is referred to here in English as Siamese Grantha. The latter
script features shapes that obviously differ from those of the Thai and Khòm
scripts, and is, thus, to be considered a specific script used uniquely by the
Brahmanical priests of Siam. The script was used to write ritual texts in Sanskrit,
Tamil and Thai, forming a unique multilingual manuscript corpus in Siam.
This article presents one particular manuscript to demonstrate the multilin-
gual character of Siamese Grantha. It is the only Siamese Grantha manuscript to
have found its way to Germany, and is now preserved at the Bayerische Staatsbib-
liothek (Bavarian State Library) in Munich under the inventory number Cod. siam
99 (Fig. 1). As the manuscript was digitised and made available online only in 2019,
modern scholars have been less aware of it. It has scarcely been mentioned in any
scholarly works from the twentieth century. However, manuscript Cod. siam 99
offers us a case study on the use of the Sanskrit and Thai languages within a single
manuscript, a phenomenon that has not often been found in the Buddhist-
dominated culture of Thailand, part of the ‘Pali imaginaire’.

||
3 Grabowsky 2011, 146; Santi Pakdeekham and Nawarat Pakdeekham 2018, 45.
4 See more in Wales 1931.
5 Kanjana Suwanwong 1996, 53.
 Sanskrit Prayers in a Theravada Kingdom | 237

Fig. 1: A Siamese Grantha manuscript from the Bayerische Staatsbibliothek in Munich, original-
ly from Thailand, between the late nineteenth and the early twentieth centuries; © Munich,
Bayerische Staatsbibliothek, Cod. siam 99, fols 5r–6r.

2 Background of the Siamese Grantha manuscript

corpus
Originating from South Indian Grantha, the Siamese Grantha script does not
resemble the usual scripts of the Siamese, such as Thai and Khòm, and is, thus,
unintelligible to normal Siamese readers. On the other hand, Siamese Grantha
has embraced the use of certain elements of the Thai and Khòm scripts, for ex-
ample, tone marks, numerals and the use of some Khòm consonants. As a re-
sult, the script created a restricted literacy accessible to the limited group of
Siamese Brahmanical priests who were trained to read and write it. It is note-
worthy that the Brahmanical priests of twenty-first-century Thailand still learn
Siamese Grantha and perceive it as the sacred script of the Brahmanical tradi-
tion in the Thai context.
238 | Peera Panarut, Manasicha Akepiyapornchai

In contrast to the large number of Siamese Buddhist manuscripts from cen-

tral and southern Thailand, Brahmanical manuscripts written in Siamese Gran-
tha are very limited in number, with only seventeen manuscripts found so far.6
All are khòi paper leporello manuscripts, which were employed to preserve
secular and vernacular texts of the Siamese tradition. Most of them are now
preserved either at the National Library of Thailand in Bangkok or the National
Library of Thailand in Nakhòn Si Thammarat. Only one of these manuscripts,
Cod. siam 99, has been preserved in Munich, Germany. However, there is also
an unknown number of manuscripts in the possession of the Brahmanical tem-
ple of Bangkok, which has served as the main temple of the Brahmanical priests
at the Siamese royal court since the early days of Bangkok. The manuscripts
from the temple are not accessible to the public, except one manuscript on the
allegiance oath (Thai: Ongkan Chaeng Nam), published in 1988.7
Owing to the small handful of manuscripts, the dating of the emergence of
Siamese Grantha script is uncertain. The earliest extant manuscript can be dat-
ed from the mid-nineteenth century, while the latest dated Siamese Grantha
manuscript is from 1919 CE, copied from an exemplar taken from Nakhòn Si
Thammarat.8 However, the date of 1693 CE is briefly mentioned in one damaged
Siamese Grantha manuscript from the National Library in Nakhòn Si Thamma-
rat.9 Though the date is arguably more likely to refer to the composition of the
text than the date of the manuscript’s production, it indicates that Brahmanical
rituals must have existed in southern Thailand at least since the late seven-
teenth century, supporting the scholars’ suggestion that Siamese Grantha origi-
nated in southern Thailand before thriving at Bangkok’s royal court.
Though the corpus available is rather small, it consists of different collec-
tions of ritual texts used and uttered by the Brahmanical priests on different
occasions, covering priestly rituals, such as the daily purification rite, and those
used in royal ceremonies, for example, the coronation and oath of allegiance
ceremonies. Therefore, as an anthology of ritual texts, a Siamese Grantha man-
uscript often consists of multiple texts in Sanskrit, Thai and sometimes even
Tamil.10 The compilation of texts differs in each manuscript, suggesting differ-
ent editorial processes and text selections by the compilers or editors, probably

||
6 See Akepiyapornchai and Panarut 2022, 131–133.
7 See Wudhichai Kosolkanjana 1988, 305–316.
8 Akepiyapornchai and Panarut 2022, 128–130.
9 Nakhòn Si Thammarat, National Library of Thailand, no. 461, fol. 4r.
10 For a discussion of Tamil texts from the Siamese Grantha manuscript corpus, see Akepiya-
pornchai and Panarut 2022, 137–140.
 Sanskrit Prayers in a Theravada Kingdom | 239

the Brahmanical priests. Sanskrit texts constitute the bulk of the Siamese Gran-
tha corpus among the extant manuscripts, while Tamil texts appear in only
three manuscripts of the corpus,11 collected alongside Sanskrit and Thai texts.
Due to the various languages of their ritual texts, Siamese Grantha manuscripts
can be considered a unique multilingual corpus, in which three languages from
three linguistic families appear together.
Siamese Grantha has never been a script widely learned and studied by
modern scholars and students. Still, scholarly works on Siamese Grantha can
be found from the early twentieth century onwards. P. Subrahmanya Shastri
provides a survey (in Thai) on manuscripts and their texts,12 while Quaritch
Wales has discussed (in English) the use of the script in the context of royal
ceremonies in Siam.13 Both works are based on seven Siamese Grantha manu-
scripts. John Ralston Marr describes the forms and orthography of Siamese
Grantha in writing Sanskrit and Tamil in his first article,14 and investigates
some Sanskrit and Tamil texts in the second one.15 On the other hand, the
forms and orthography in writing Thai are specifically elaborated in the
works of Wudhichai Kosolkanjana and Niyada Lausoonthorn.16 Neelakanta
Sarma has produced modern transliterations of Siamese Grantha manu-
scripts, provided with an introduction in French by Jean Filliozat,17 through
utilising facsimiles of the manuscripts kept at the National Library of Thai-
land in Bangkok. The twentieth-century scholars of Thai and Tamil studies
mentioned above may not have known of the whereabouts of Cod. siam 99
from the Bayerische Staatsbibliothek in Munich, thus, it has never been
properly studied in the previous scholarship on Siamese Grantha.
After having been acquired by the Bayerische Staatsbibliothek in Munich
in 1978, manuscript Cod. siam 99 became one of the Asian manuscripts
shown at an exhibition by the library between 16 November 1982 and 5 Feb-
ruary 1983. Therefore, the exhibition book, Buch im Orient,18 published in
1983, also includes a description of this manuscript by Alfons Dufey, who

||
11 Namely, Bangkok, National Library of Thailand, Lilit Subsection, no. 360; and Bangkok,
National Library of Thailand, Royal Ceremonial Treatises Subsection, nos 667 and 668.
12 P. Subrahmanya Shastri 1931.
13 Wales 1931.
14 Marr 1969.
15 Marr 1972.
16 Wudhichai Kosolkanjana 1988; Niyada Lausoonthorn 1992.
17 Filliozat 1972.
18 Dachs 1983.
240 | Peera Panarut, Manasicha Akepiyapornchai

proposes the date of this manuscript to the end of the nineteenth century.19
This might be the only publication of the twentieth century that mentions
this manuscript at all. While labelling the language of the manuscript as
corrupted Sanskrit (‘Korruptes Sanskrit’), Dufey, nonetheless, focuses his
description on Brahmanical priests in Thailand and their culture rather than
the content of this particular manuscript:

Vetatam 3. Leporello-Handschrift aus pappeartigem Papier. 80 Bl., 11,5 × 36,5 cm. Korrup-
tes Sanskrit, weiße Grantha-Schrift auf schwarzem Grund. – Thailand, Ende 19. Jh. Cod.
siam 99. Obwohl die offizielle Religion Thailands der Buddhismus ist, war das siamesi-
sche Hofzeremoniell schon seit jeher hinduistisch. Seit 1786, als König Cakri den Namen
Rama I. annahm, knüpft das Herrscherhaus sogar an das hinduistische Epos Ramayanam
an. Für die Staatszeremonien, soweit sie nicht rein buddhistischen Charakter tragen, le-
ben seit Jahrhunderten eigene Brahmanen in der jeweiligen Regierungshauptstadt. Da
diese überwiegend aus Südindien stammenden Priester keine Frauen mitbrachten, hat
sich der Anteil indischen Blutes in ihren Adern stark verringert. Eine ähnliche Entwick-
lung haben auch die Sanskrit- und Tamilkenntnisse dieser Vishnu- bzw. Shiva-Anhänger
durchgemacht. Immerhin tragen sie noch die lange Haartracht ihrer Kaste, eine Art Chig-
non, legen bei feierlichen Anlässen die Brahmanenschnur um und rezitieren religiöse in-
dische Texte. Dazu lesen sie aus Handschriften wie der hier aufgelegten, deren sehr kor-
ruptes Vedisch in einem vom südindischen Tamil-Grantha abgeleiteten Alphabet
geschrieben ist. Erstaunlicherweise haben sich aber die Bezeichnung für den musikali-
schen Akzent des Vedischen bei diesen Zaubersprüchen, Mantras, erhalten.

Vetatam 3. Leporello manuscript on khòi paper. 80 pages, 11.5 × 36.5 cm. Corrupted San-
skrit; white Grantha script on black background. – Thailand, late nineteenth century. Cod.
siam 99. Although the official religion of Thailand is Buddhism, for a long time the Sia-
mese court ceremonies were Hindu. Since 1786, when King Cakri took the name ‘Rama I’,
the royal house has even connected itself with the Hindu epic Ramayana. Brahmanical
priests have lived in the capital city for centuries for [conducting] state ceremonies, which
do not bear a purely Buddhist character. Since these priests, most of whom came from
South India, did not bring any wives with them, the proportion of Indian blood in their
[descendants] veins has been greatly reduced. The Sanskrit and Tamil knowledge of these
Vishnu and Shiva devotees has also developed in a similar way. After all, they still wear
the long hairstyle of their caste, a kind of chignon; put on the Brahmanical threads on cer-
emonial occasions; and recite Indian religious texts. To do this, they read from manu-
scripts such as the one published here, whose very corrupt Vedic is written in an alphabet
derived from the Grantha Tamil script of South India. Surprisingly, the indication of the
chanting accent of the Vedic in these magic spells, mantras, has been preserved [through
this script].20

||
19 Dufey 1983.
20 Dufey 1983, 283 (our translation).
 Sanskrit Prayers in a Theravada Kingdom | 241

Barend J. Terwiel briefly mentioned this manuscript from Munich more recently
in 2017, as one of the few Thai manuscripts preserved in Germany.21 Therefore,
Manasicha Akepiyapornchai and Peera Panarut also include the manuscript
Cod. siam 99 as one of the seventeen manuscripts in their updated overview of
the extant Siamese Grantha corpus.22 However, the manuscript and its text, a
collection of Brahmanical formulas or mantras in the domain of the Pali imagi-
naire, deserve more elaboration.

3 Cod. siam 99: A Siamese Grantha manuscript

from Munich
The manuscript in question, Cod. siam 99, is a blackened khòi paper leporello
manuscript, the form that is commonly found in traditional Siamese manuscript
culture. The writing is written in white ink on a black paper background. Ac-
cording to the handwritten catalogue of Oriental manuscripts at the Bayerische
Staatsbibliothek,23 the library acquired the manuscript in 197824 by purchasing it
from Dr W. Burger of Hong Kong.25 The catalogue provides a brief description as
follows: ‘Leporello manuscript. 11.5 × 36.5 cm. 4 lines [per page]. White writing
on black paper. So-called Grantha script. Text of the court Brahmanical priests.
Vetatam3’26 (see Fig. 2). The manuscript is in fine condition, except for the front
cover, on which the title (in Grantha) is rather faded. However, it is identifiable
as Thai written in Siamese Grantha, and reads Wet Tang Tang (‘Various man-
tras’).27 Though the Thai word wet originates from Sanskrit veda, the title here
also suggests that this manuscript collects the verses or mantras in general ra-
ther than any direct or precise citations of Vedic Sanskrit texts, compiling mate-
rial from different sources into an anthology.

||
21 Terwiel 2017.
22 Akepiyapornchai and Panarut 2022.
23 Bayerische Staatsbibliothek 2008.
24 The catalogue indicates that the entries from Cod. siam 91 to Cod. siam 99 were ac-
quired in 1978.
25 Noted in the catalogue as ‘Gekauft von Dr. W. Burger, Hongkong’.
26 Originally in German: ‘Faltbuch. 11,5 × 36, 5 cm. 4 Zeilen. Weiße Schrift auf schwarzem
Papier. Sog. Grantha-Schrift. Text der Hofbrahmanen. Vetatam3’.
27 Thai เวตตางๆ. As the writing is faded, the title could easily have been read and misread as
‘Vetatam3’, which does not make proper sense in the Thai language, as recorded in the catalogue.
242 | Peera Panarut, Manasicha Akepiyapornchai

Fig. 2: Information on Cod. siam 99 as registered in the catalogue of Oriental manuscripts at

the Bayerische Staatsbibliothek in Munich; Bayerische Staatsbibliothek 2008.

Even though the writings inside the manuscript appear undamaged, the cop-
ying process seems incomplete. The entire verso side has been left blank,
while what is written on the recto side does not seem to have reached its
proper end. The text runs from the beginning of the manuscript to fol. 73r,
and then the running text, seemingly unfinished, is interrupted by a few
blank pages (fols 76r–81r). The writing continues on fol. 82r and ends on fol. 86r.
The latter part (fols 82r–86r) preserves the related ritual texts and was un-
doubtedly written by the same hand. Nonetheless, it does not seamlessly
continue from the previous part. The manuscript is undated, but is probably
from the period between the late nineteenth century and the early twentieth
century, in which Siamese Grantha manuscripts were still produced by the
Siamese Brahmanical priests.28 No scribal colophon has been found. There-
fore, we unfortunately do not have any further information on its scribe,
compiler or editor, apart from the presupposition that Siamese Grantha has
been used only by Siamese Brahmanical priests.
The entirety of manuscript Cod. siam 99 is written in Siamese Grantha
script. Nevertheless, this single script is employed for writing two languages
here: Sanskrit and Thai. The manuscript is comprised of various sets of San-
skrit mantras. It is unclear whether all the mantras collected in this manu-
script were used together in the same series of rituals, but the different man-
tras must have been collected in Cod. siam 99 for the practical purpose of
performing rituals, as each of them is often interrupted by short ritual in-
structions in vernacular Thai, telling the priests how to conduct the rituals.
The use of two languages in this manuscript and their relationship will be
discussed further below.

||
28 See more in P. Subrahmanya Shastri 1931.
 Sanskrit Prayers in a Theravada Kingdom | 243

4 Sanskrit texts in Cod. siam 99

Approximately half of this manuscript is parallel to Sarma’s transliteration of
the section he labels Ācāryapūjā Devarūpa Śrī Umā Śrī Nārāyaṇa Śrī Ma-
hāvighneśvara.29 Although the title includes various Hindu deities, such as Umā
and Nārāyaṇa, the texts based on Sarma’s transliteration seem to be devoted
mainly to Śiva. This is also the case of manuscript Cod. siam 99. It should be
noted that a great deal of the Sanskrit in this manuscript is unintelligible. The
structure of the entire manuscript, especially the parts that are intelligible, is as
follows:30

Table 1: The structure and texts in the manuscript Cod. siam 99.

Text Folio no. (recto)

1 Prayer to Śiva and Umā 4–9
2 Prayer to the different directions 10
3 The eightfold praise poem to Śiva (stotrāṣṭaka) and 11–20
4 Praise of the five-syllable mantra (śivapañcākṣarastotra)
5 Self-purification (ātmaśuddhi) 20–23
6 Prayer to Īśvara and Śiva 23–27
7 Prayer to Vighnanāyaka (Śiva’s son) 27–29
8 Offerings 29–39
9 Worship of the teachers (ācāryapūjā) 39–53
10 Prayer to destroy faults 53–67
11 Prayer to Umā (incomplete) 68
12 Prayer to the Śiva liṅga (?) 68–71
13 Unintelligible parts 71–84
14 A poem from Bāṇa’s Harṣacarita 84–85
15 Prayer to Kailāsa 85–86

As we can see, the Sanskrit texts could be categorized differently, namely by

genre, if we were to look at them through the lens of the Sanskrit literary para-
digm. Most of the Sanskrit texts have no label attached to them. Nevertheless,

||
29 Sarma 1972, 137–163.
30 The structure and identification of each text is based on Sarma’s transliteration.
244 | Peera Panarut, Manasicha Akepiyapornchai

they usually refer to a particular Hindu deity or more than one deity at a time,
and can, thus, be understood as prayers or devotional and expressive state-
ments dedicated to a deity, especially Śiva in this context.31 It is also probable
that these prayers had been circulated in mainland India and even South East
Asia by the time they were incorporated into this collection. Interestingly, one
text evidently identifies itself as a praise poem (stotra): the eightfold praise
poem (stotrāṣṭaka). Furthermore, there is one poem from the Harṣacarita, the
biography of King Harṣa, by the important Sanskrit poet Bāṇa (seventh century).32
Despite the differences in genre, it can be argued that these texts all have
the same function as mantras. It is difficult to determine precisely what a man-
tra is, as it is so culturally specific; still, it may be loosely defined as a formula
that is used in a certain ritual practice or action and has innate efficacy regard-
less of its linguistic meaning.33 Some of the texts in the manuscript have a form
that is obviously mantric, especially those that include oṃ at the beginning or
svāhā at the end. Examples are Oṃ pañcākṣarāya siddhiḥ jaya34 (‘Oṃ to the five-
syllable [mantra of Śiva]. May there be success. Triumph!’)35 and Hari oṃ namaḥ
śivāya (‘Hari oṃ, honour to Śiva’)36 and Oṃ namo nirṛtidiśāya […] svāhā (‘Oṃ,
honour to the south-west direction […] svāhā’).37 On the other hand, it is less
common in India for a poem like that of Bāṇa to be used as a mantra. In any
case, the ritual instructions in Thai that surround these Sanskrit texts also serve
to contextualize them, indicating that they function as mantras that accompany
various ritual practices.38 Importantly, the title on the front cover of this manu-
script, Wet Tang Tang, also suggests that it is a collection of various wet or
Brahmanical mantras. The words wet (or veda in Sanskrit) and mon (mantra in
Sanskrit) are usually paired in the compound wet-mon (เวทมนตร) to refer to
sacred formulas in Brahmanical culture in contemporary Thai.39 In this manu-
script, however, the word wet alone is used to refer to a mantra. Although the
Sanskrit word veda, which became wet in Thai, usually indicates the most au-
thoritative scriptural corpus in Brahmanical orthodoxy, it could be argued that

||
31 For more on prayers, see Stainton 2019, 159–196.
32 For more information on Bāṇa and his works, see Bronner, Shulman and Tubb 2014.
33 See Alper 1989 for other definitions of mantra.
34 Cod. siam 99, fol. 20r.
35 All of the translations of the Sanskrit texts are Manasicha Akepiyapornchai’s unless indi-
cated otherwise.
36 Cod. siam 99, fol. 38r.
37 Cod. siam 99, fol. 10r.
38 See the next section on Thai instructions.
39 See McGovern 2022, 259.
 Sanskrit Prayers in a Theravada Kingdom | 245

once it entered the Thai context, its lexical scope was broadened to Brahmanical
mantras in general, both Vedic and non-Vedic.
Mantras dedicated to Śiva form the majority of the Sanskrit texts. They are of-
ten marked by opening passages that mention the names of Śiva: exemplarily, the
first statement is for Śiva and his consort, Umā, namas tubhyam umāsvāmibhyām
(‘Honour to Umā and the Husband [Śiva]’).40 In some cases, they are also ac-
companied by closing statements dedicated to Śiva, such as hara rudrarūpa
mahādeva śambho (‘O, the destroyer, the one with the form of Rudra, the great
god, the benevolent one’), which occurs twelve times at the end of each set of
mostly unintelligible mantras.41 We also find a declaration of the result (phalaśruti)
of reciting this set of mantras, namely, that it may save one from going to hell,
which is Yama’s abode: idaṃ puṇyaṃ yaḥ paṭhet śivasaṃnidhau yamalokam
ayaṃ naiti (‘the one who recites this in the presence of Śiva does not reach the
abode of Yama’).42 Further mantras to Śiva can be identified through his various
paradigmatic attributes and epithets in the text, most evidently, ‘the three-eyed
one’, ‘the destroyer of the three cities’, ‘the one with the blue neck’, and so on.
This can be seen, for example, in the passage, Oṃ namaḥ satataṃ bhagavad-
viśvaviśveśvarāya tryambakāya tripurāntakāya trikamalanayanāya […] nīlakaṇ-
ṭhāya sarveśvarāya sadāśivāya śivamahādevāya namaḥ (‘Oṃ, honour to the
great god, Śiva, the lord of the universe, the three-eyed one, destroyer of the
three cities […] the one with the blue neck, the lord of all, the eternal Śiva’).43
In addition to Śiva, some of the Sanskrit mantras are meant for other, minor
deities. This latter includes Śiva’s son, Vighnanāyaka or Vighneśvara, as in the
line, mahādevāya ḍamaruśivaputrāya namo namaḥ (‘Honour to the great god,
the son of Śiva who has the sacred drum’).44 Interestingly, all the mantras in the
self-purification ritual (ātmaśuddhi) relate solely to Viṣṇu except the first one,
which is the beginning of the Gāyatrī mantra: Oṃ bhūr bhuvaḥ svaḥ (‘Oṃ, o
earth, space and heaven’).45 The mantras in this ritual context are paired with
the different body parts that should be purified. The first mantra, the Gāyatrī, is
to be recited when one plunges into the water three times. While washing the
hands, one should recite, Oṃ padmanābhāya namaḥ (‘Oṃ, honour to the lotus-

||
40 Cod. siam 99, fol. 4r.
41 Cod. siam 99, fols 12r–18r. Sarma 1972, 149 identifies this text as the Śivaparādhakṣamāpa-
ṇastotra, but the two texts are far from identical.
42 Cod. siam 99, fol. 18r.
43 Cod. siam 99, fols 23r–24r.
44 Cod. siam 99, fol. 29r.
45 Cod. siam 99, fols 20r–23r.
246 | Peera Panarut, Manasicha Akepiyapornchai

navelled one’);46 for the arms, Oṃ mādhavāya narāya namaḥ (‘Oṃ, honour to
the primeval one, Mādhva’);47 for the body, Oṃ narāya namaḥ (‘Oṃ, honour to
the primeval one’);48 and for the head, Oṃ keśavanarāya namaḥ (‘Oṃ, honour to
the primeval one, Keśava’).49
Furthermore, we find a more extensive list of the deities, along with non-
divine elements, such as mountains, continents, oceans, aeons and scriptures,
in the section that corresponds to the one Sarma identifies as Ācāryapūjā:50

Oṃ, may there be success; oṃ, honour to the good one; oṃ, honour to the god, the god of
gods and demons; honour to the great lord, whose wife is Umā, who is superior to all
kinds of goodness, the fruit of the four classes, the powerful god, Saṅkarṣaṇa, Pradyum-
na, Aniruddha, the young one who is invincible on the battlefield […] earth, water, fire,
wind, space; the sun, the moon, Indra, the four guardians of the directions, the carrier of
the vase of wealth, the mountains that are the middle one, Kailāsa, Meru and Himalaya,
along with the continents, the four oceans, the four aeons, the four Vedas, the four kinds
of grease.51

Note that these elements and their numbers reflect common tropes usually found
in Sanskrit literature in mainland India. There are other elements, for example, the
seven (sapta) rivers (nadī), continents (dvīpa), oceans (samudra), kings (rāja) and
sages (ṛṣi), followed by the ten guardians of the directions (daśalokapāla) and
thirty-three deities (trayastriṃśaddevatā). We then have the names of the mytho-
logically well-known snakes or nāgas, such as Takṣaka, Ānanta and Vāsuki. Un-
fortunately, the rest is unintelligible. Importantly, this is not the only place we
find reference to forces other than the deities. There is another set of mantras at
the beginning of the manuscript devoted to the eight directions, such as the east
(pūrva), south-east (āgneya), south (dakṣiṇa), south-west (nirṛti) and north-west
(vāyavya). However, the other directions, namely, the north, north-east and
west, cannot be definitively identified.52
The two most distinctive texts in this manuscript are the one that is about
the destruction of faults and the poem of Bāṇa. The first set can be found imme-
diately after the Ācāryapūjā section, and runs until the interruption in the man-

||
46 Cod. siam 99, fol. 21r.
47 Cod. siam 99, fol. 21r.
48 Cod. siam 99, fol. 21r.
49 Cod. siam 99, fol. 21r.
50 Sarma 1972, 138–142; Cod. siam 99, fols 39r–50r.
51 This translation is based on the text in Sarma 1972, 138.
52 Cod. siam 99, fols 9r–10r.
 Sanskrit Prayers in a Theravada Kingdom | 247

uscript.53 This set usually begins with ‘all’ (sarva) and ends with ‘for the destruc-
tion of faults’ (doṣavināśāya), as in ‘for the destruction of all karmic faults’ (sar-
vakarmadoṣavināśāya). It claims that the recitation or ritual in each context has
the aim of destroying various kinds of faults: for example, those related to dif-
ferent astrological aspects, namely, the conjunction of qualities based on the
position of the moon when one is born (gaṇa), the conjunction of the planets
(saṃgati) and the zodiac sign (rāśi).54 The other set merely constitutes a popular
poem from the Harṣacarita of Bāṇa. As noted above, this poem stands out in
this manuscript of ritual instructions and mantras, yet, it seems to have the
same mantric function as other Sanskrit texts. It should be noted that this poem
particularly was collected in Bhagadatta Jalhana’s thirteenth-century Sūkti-
muktāvali, edited by Krishnamacharya.55 Based on this, one could argue that
this poem circulated outside of the Harṣacarita and attained its status as part of
a collection of well-known passages before it was incorporated into this Siamese
ritual corpus. In any case, it is devoted to Śiva as the cosmic pillar and describes
him in the universal city:

Homage to him,
beautiful with the chowrie
of the crescent moon
kissing his high head,
to the one Pillar
put in place at the founding of the city
that is the universe –
to Śambhu.56

As has been mentioned above, this manuscript shares some of the Sanskrit texts
found in Sarma’s transliteration. The texts in this manuscript also present a
similarly high degree of corruption and incompleteness when compared to
those recorded by Sarma, to the extent that it is arguably the most corrupted of
the Siamese Grantha manuscripts we have explored so far.57 The textual corrup-
tion makes it quite difficult to identify some of the texts precisely and under-
stand what they convey, even for those with a good knowledge of Sanskrit, like
Sarma. Thus, we still do not know the sources of the various Sanskrit texts aside
from those already identified. Given their unintelligibility, these Sanskrit texts

||
53 Cod. siam 99, fols 51r–67r.
54 Cod. siam 99, fol. 64r.
55 Krishnamacharya 1938.
56 Bronner, Shulman and Tubb 2014, 27. See the same source for more analysis of the verses.
57 Sarma 1972, 137.
248 | Peera Panarut, Manasicha Akepiyapornchai

are not just variations or representatives of what we find in South East Asia,
which, in some cases, differ only slightly from Sanskrit texts in mainland India.
Moreover, it seems unlikely that the Brahmanical priests understood the
meaning of these texts. They are, as has been stated previously, meant to be
used as mantras in a ritual or rituals, and the recitation is, thus, more important
than comprehension. However, contrary to the general practice of Sanskrit reci-
tation in mainland India, which prioritizes precise and invariant pronunciation,
the pronunciation of the texts in this manuscript deviates greatly from the San-
skrit pronunciation, and seems to align more with Thai and Pali. We have sug-
gested in the overview of Siamese Grantha manuscripts that the alignment of
the Sanskrit pronunciation with the Thai and Pali ones could contribute to tex-
tual corruption,58 which might be the case in this particular manuscript. In addi-
tion to corruption, some of the texts and the manuscript itself are also incom-
plete, suggesting that they might not have been recorded or handed down
properly.

5 Ceremonial instructions in Thai: Text or

paratext?
The Siamese Grantha script was commonly employed to write texts in the Thai
language. Ritual poetry texts in Thai, such as Ongkan Chaeng Nam and The
Collection of Old Elephant Treatises, have been preserved as the main texts of
several Siamese Grantha manuscripts from the corpus. However, the paratexts
of the manuscripts often appear in vernacular Thai, though written in Siamese
Grantha, as in the case of the front cover title of the manuscript Cod. siam 99,
Wet Tang Tang (‘Various mantras’). The vernacular Thai here seems to be the
language the scribes used to communicate with their readers, while the main
texts of a Siamese Grantha manuscript can appear in foreign languages, such as
Sanskrit and sometimes Tamil. Apart from scribal paratexts, such as cover titles,
ritual instructions in Thai also accompany the Sanskrit texts in many cases,
including Cod. siam 99.
Ritual instructions in Thai consistently accompany each mantra from the
beginning to the unfinished end of Cod. siam 99. Even when some longer San-
skrit texts run to more than ten folded pages, Thai instructions still appear at
the end of each part. Although the written texts are interrupted by blank pages

||
58 Akepiyapornchai and Panarut 2022, 136.
 Sanskrit Prayers in a Theravada Kingdom | 249

(fols 76r–81r), the writing that resumes on fols 82r–86r also bears Sanskrit man-
tras, along with the corresponding Thai instructions. These brief instructions,
forming a significant feature of this manuscript, inform readers not only of the
purpose and context of each mantra, but also what to do while uttering these
mantras in the actual rituals. Furthermore, the instructions also mark the end of
each mantra before the following ones begin, thereby assuming a structural
function and helping readers to navigate through the Sanskrit text. Interesting-
ly, the instructional texts of this manuscript, despite being written in Thai, ap-
pear only in Siamese Grantha script throughout the manuscript. Thus, unless
the average reader reads Siamese Grantha, they would not be able to recognise
the Thai texts or distinguish them from the Sanskrit texts at all. This may be
considered a way to prevent the average reader from making sense of passages
written in the common vernacular language.
Sometimes the name of the specific ritual is also mentioned in the in-
struction; for example, a Thai instruction text from page 13r reads อนอาราต
ธนาครปรจมเนอษษะ แลวจงธมอาตมาสต (‘These verses are for inviting teachers
over [our] heads. [After that,] then conduct the Attama Sut [ritual]’).59 This short
passage explains that the previous Sanskrit mantras were used to invite and pay
homage to teachers; it then names the ritual to be further performed as Attama
Sut (Sanskrit ātmaśuddhi ‘self-purification’), a purification ritual for Brahmani-
cal priests. Correspondingly, other Thai instructions found in this manuscript
explicitly state the purpose of the Sanskrit mantras, namely, to which god they
were recited. The instructions พรคาถาสามบตนสงนามพระอสวนแล (‘These
three verses are for consecrating [the statue of] God Śiva with water’)60 and
อนนสงนามพรวคเนกวฆนายแล ฯ (‘These verses are for consecrating [the statues
of] God Ganesha’)61 were found, for example. These Thai instructions are essential
for traditional readers to recognise the purpose of each individual Sanskrit text.
Furthermore, a number of Thai instructions also inform readers on how to
perform the ritual when reciting the related Sanskrit mantras. The manuscript
(Cod. siam 99, fol. 4r), for example, begins with the reverential verse namatasa-
va ulaśvāmyeṇaṃ (Sanskrit: namas tubhyam umāsvāmibhyām), followed by an
instruction in Thai: ๓ ท (‘three times’), marking how many times the prayer

||
59 The passage can be rendered in modern Thai orthography as อนนอาราธนาครประจาเหนอศรษะ
แลวจงทาอาตมาสตร; Munich, Bayerische Staatsbibliothek, Cod. siam 99, fol. 20r.
60 In modern Thai orthography: พระคาถาสามบทนสรงนาพระอศวรแล ฯ; Munich, Bayerische
Staatsbibliothek, Cod. siam 99, fol. 27r.
61 In modern Thai orthography: อนนสรงนาพระวฆเนศรวฆนายแล ฯ; Munich, Bayerische Staats-
bbliothek, Cod. siam 99, fol. 29r.
250 | Peera Panarut, Manasicha Akepiyapornchai

should be uttered. Then the instruction continues: กราบเบนบนจายคปรทตตามเพ

ตชแลววานมษกาน (‘Prostrate in the manner of five parts, according to the tradi-
tion, and then utter the reverencing verses’),62 telling readers which gestures
should be performed in the ritual. The most obvious case of Thai instructions
describing how to conduct the ritual is found in the text of the self-purification
ritual, in which Brahmanical priests purify each part of the body while uttering
the mantras. The Sanskrit mantras in this part are short, followed by the Thai
instructions marking which part of the body the priests have to wash for purifi-
cation (see Fig. 3).63

Figs 3a–b: The text of the Attama Sut (ātmaśuddhi) purification from Cod. siam 99 (Munich,
Bayerische Staatsbibliothek, Cod. siam 99, fols 20r–21r).

||
62 In modern Thai orthography: กราบเบญจางคประดษฐตามเวชชแลววานมสการ; Munich, Bayeri-
sche Staatsbibliothek, Cod. siam 99, fol. 4r.
63 For the Sanskrit mantras and translation, see the previous section.
 Sanskrit Prayers in a Theravada Kingdom | 251

The text of these two folded pages (fols 20r–21r) reads:

kam | oṃpancanamaṣakāriyaṣidhijaiyya (× 2) ‘Oṃ to the five-syllabled [mantra of Śiva]. May

there be success. Triumph!’ (× 2) yamahesra oṃbhagha vattivipanca vetṣanukancevayo-
gitepraḥṣime [this part is unintelligible] | อนอาราตธนาครปรจมเนอษษะ แลวจงธมอาตมาสต
‘These verses are for inviting teachers over [our] heads. [After that,] then conduct the Attama
Sut [ritual].’ |
ombhuva (× 2) vaśavaṣṣavāha ‘Oṃ, o earth, space, and heaven.’ (× 2) อนชบนามตวตวเฐต ๓ท
‘This [verse] is for rinsing the body with water three times.’ oṃgukkhāṣināyanam [unintelligi-
ble part] อนลางตน ‘This is for washing the feet.’ | ompaṣṣamāṇarāyyanam ‘Oṃ honour to the
lotus-navelled One.’ อนลางมอ ‘This is for washing the hands.’ |
ommādhivāyanarāyanam ‘Oṃ honour to Mādhva.’ อนลางแขน ‘This is for washing the arms.’ |
omṇarāyyanam ‘Oṃ honour to Nara.’ อนลางตว ‘This is for washing bodies.’ | omkottapāy-
yanarāyyanam ‘Oṃ honour to Keśava.’ อนลางหว ‘This is for washing heads.’ |

The question of whether these Thai ritual instructions should be considered part
of the main text of the manuscript or its paratext might be answered differently
depending on the perspective. If we follow what the manuscript title, Wet Tang
Tang (‘Various mantras’) suggests, then the main text should cover the Sanskrit
mantras, which are the only parts recited in the actual rituals. The instructional
texts, on the other hand, are not supposed to be uttered at all, but give the read-
ers further details concerning the rituals (i.e. the purpose of mantras, the ges-
tures for performing the rituals). Nevertheless, these instructional texts were
probably not composed by the scribe of this particular manuscript, given the
fact that instructions for the same ritual texts sometimes appear in multiple
manuscripts, as in the case of Attama Sut (ātmaśuddhi). The Sanskrit mantras
and their Thai instructions as found in Cod. siam 99 correspond closely to the
ātmaśuddhi preserved in a Siamese Grantha manuscript from Nakhòn Si Tham-
marat, suggesting that the Thai instructions might have been transmitted along
with the Sanskrit ritual texts. The use of Siamese Grantha in Cod. siam 99 also
allows these Thai instructions to assimilate visually to the Sanskrit texts by
means of the same writing. In addition, these Thai instructions may have been
necessary for Brahmanical priests in Thailand, who did not use Sanskrit in their
everyday lives outside ritual contexts. From this perspective, these Thai instruc-
tions may also be considered part of the main text and should not be missing
from the manuscript, although they might not originally have been conceived
for their Sanskrit counterparts, but added later in order to assist readers, or in
this case ritual practitioners, with practical details of the rituals.
Regardless of whether the Thai instructional texts of manuscript Cod. siam 99
fall under the category of main text or that of paratexts, these Thai instructions
are undeniably a significant part of the manuscript, interacting with their corre-
252 | Peera Panarut, Manasicha Akepiyapornchai

sponding Sanskrit texts, and, thus, should not be overlooked by any traditional
reader when reading this manuscript.

6 Concluding remarks
Although the Theravada culture of Siam has been dominated by Thai and Pali,
Sanskrit and the Brahmanical tradition are still part of its cultural profile. This is
evident in manuscript Cod. siam 99, in which Thai and Sanskrit coexist, per-
forming different but closely related functions. In this manuscript, the Thai
texts play a role as ritual instructions, while the Sanskrit texts constitute the
majority and serve as mantras. The two languages create multilingual dynamics
in a single-script artefact, and represent the multicultural identity of this Sia-
mese Grantha corpus, which is situated between the Pali imaginaire, in the
Theravada world of the Thai-speaking kingdom, and the Sanskrit-language
Brahmanical culture of India. The manuscript Cod. siam 99 should, therefore,
be considered a significant manuscript due to its multilingual character, despite
the fact that it has long been excluded from scholarship on the Siamese Grantha
manuscript corpus. This article merely serves as a preliminary study of this
particular manuscript and its multilingual dynamic, aiming to be a first step
towards its further investigation and the future publication of its contents.
Although the Sanskrit texts of manuscript Cod. siam 99 could potentially
come from mainland India, they differ and are detached from the standard or
widely known literary corpus of Sanskrit. The corruption evident in these
texts also suggests that they had lost contact with the Sanskrit corpus in
mainland India and were handed down and circulated solely in the Pali im-
aginaire of Thailand long enough to attain this level of unintelligibility, espe-
cially in terms of pronunciation. However, the corruption indicates that these
texts had assimilated to the Siamese Theravada domain, in which Thai and
Pali were linguistically predominant. Being accompanied by the Thai instruc-
tions, the Sanskrit texts, as collected and preserved in Cod. siam 99, also
attained a new life, distinct and independent from the Sanskrit corpus of
India. They are not merely Sanskrit texts but part of Siamese Brahmanical
rituals in Thailand. Importantly, their function as mantras still resonates
with Sanskrit verses in the Indian or even South Asian context. Thus, the
Sanskrit mantras in a Siamese Grantha manuscript such as Cod. siam 99 rep-
resent the persistence of Indian influence in the Theravada kingdom of Thai-
land as much as Siamese Brahmanical culture.
 Sanskrit Prayers in a Theravada Kingdom | 253

Acknowledgements

The research for this paper was funded by the Deutsche Forschungsgemein-
schaft (DFG, German Research Foundation) under Germany’s Excellence Strate-
gy – EXC 2176 ‘Understanding Written Artefacts: Material, Interaction and
Transmission in Manuscript Cultures’, project no. 390893796. The research was
conducted within the scope of the Centre for the Study of Manuscript Cultures
(CSMC) at Universität Hamburg.

References
Akepiyapornchai, Manasicha and Peera Panarut (2022), ‘Sacred Literacy: An Overview of Multi-
lingual Siamese Grantha Manuscripts’, in Volker Grabowsky (ed.), Manuscript Cultures
and Epigraphy in the Tai World, Chiang Mai: Silkworm Books, 115–165.
Alper, Harvey (1989), Understanding Mantras, Albany: SUNY Press.
Bayerische Staatsbibliothek (2008), Repertorium der Codices orientales der Bayerischen
Staatsbibliothek, Kapsel 2: K-Z – BSB Cbm Cat. 40 d(2, München, 19.-21. Jh. [BSB-Hss Cbm
Cat. 40 d(2], <http://daten.digitale-sammlungen.de/bsb00026276/image_207> (ac-
cessed on 19 May 2023).
Bronner, Yigal, David Shulman and Gary Tubb (2014), ‘Introduction’, in Yigal Bronner, David
Shulman and Gary Tubb (eds), Innovations and Turning Points: Toward a History of Kāvya
Literature, New Delhi: Oxford University Press, 1–32.
Collins, Steven (1998), Nirvana and Other Buddhist Felicities: Utopias of the Pali Imaginaire,
Cambridge: Cambridge University Press.
Dachs, Karl (ed.) (1983), Das Buch im Orient: Handschriften und kostbare Drucke aus zwei
Jahrtausenden, Wiesbaden: Ludwig Reichert.
Dufey, Alfons (1983), ‘206 Ritualtext siamesischer Hofbrahmanen’, in Dachs 1983, 283.
Filliozat, Jean (1972), ‘Introduction’, in Neelakanta Sarma, Textes Sanskrits et Tamouls de
Thaïlande, Puducherry: Institut français d’Indologie, vii–xi.
Goddard, Cliff (2005), The Languages of East and Southeast Asia: An Introduction, Oxford:
Oxford University Press.
Grabowsky, Volker (2011), ‘Manuscript Culture of the Tai’, manuscript cultures, 4: 145–156.
Kanjana Suwanwong กาญจนา สุวรรณวงศ์ (1996), Withi chiwit phithikam lae kan thamrong
ekkalak thang watthanatham khòng klum phram ratchasamnak nai sangkhom thai: süksa
kòrani thewasathan bot phram phra nakhòn วิถีชีวิต พิธีกรรมและการธํารงเอกลักษณ์ ทางวัฒ
นธรรม ของกลุ่มพราหมณ์ ราชสํานักในสังคมไทย: ศึกษากรณี เทวสถานโบสถ์พราหมณ์พ ระนคร
(‘Ways of Life, Rituals and Cultural Identity of Court Brahmins in Thai Society: A Case Study
of Bangkok Devasthan Botsbrahmana’), MA thesis, Chulalongkorn University.
Krishnamacharya (ed.) (1938), The Suktimuktavali of Bhagadatta Jalhana, Baroda: Oriental
Institute.
Marr, John Ralston (1969), ‘Some Manuscripts in Grantha Script in Bangkok’, Bulletin of the
School of Oriental and African Studies, 32/2: 281–322.
254 | Peera Panarut, Manasicha Akepiyapornchai

Marr, John Ralston (1972), ‘Some Manuscripts in Grantha Script in Bangkok II’, Journal of the
Siam Society, 60/2: 61–87.
McGovern, Nathan (2022), ‘On the Utility of the Category of “Hinduism” in Thailand’, The Jour-
nal of Hindu Studies, 15/3: 244–267.
Niyada Lausoonthorn นยะดา เหลาสนทร (1992), ‘Ongkan chaeng nam chabap aksòn chiang
phram phasa thai’ โองการแชงนา ฉบบอกษรเฉยงพราหมณ ภาษาไทย (‘A Study of the Thai
Version of Chiang Bhramh Script: Ongkan Chaengnam’), in Phinit wannakam: ruam bot
khwam wichakan dan wannakhadi lae phasa (süksa cak chabap tua khian) พนจวรรณกร
รม: รวมบทความวชาการดานวรรณคดและภาษา (ศกษาจากตนฉบบตวเขยน) (‘Phinit Wanna-
kam: A Collection of Academic Articles on Literature and Language (Studied from Manu-
scripts)’), Bangkok: Mae Kham Phang, 12–19.
P. Subrahmanya Shastri ป. สพรหมณย ศาสตร (1931), Rai ngan samruat tamra phra ratcha
phithi phram sayam รายงานสารวจตาราพระราชพธพราหมณสยาม (‘Survey on the Siamese
Brahmanical Ceremonial Treatises’), Bangkok: Phra Can Printing House.
Santi Pakdeekham ศานต ภกดคา and Nawarat Pakdeekham นวรตน ภกดคา (2018), Prawattisat
ayutthaya cak carük: carük samai ayutthaya ประวตศาสตรอยธยาจากจารก: จารกสมยอยธยา
(‘History of Ayutthaya from Inscriptions: The Inscriptions of the Ayutthaya Period’), Bang-
kok: The Historical Society.
Sarma, Neelakanta (1972), Textes Sanskrits et Tamouls de Thailande, Puducherry: Institut
français d’Indologie.
Stainton, Hamsa (2019), Poetry as Prayer in the Sanskrit Hymns of Kashmir, New York: Oxford
University Press.
Terwiel, Barend Jan (2017), ‘Cultural Goods and Flotsam: Early Thai Manuscripts in Germany
and Those Who Collected Them’, Manuscript Studies: A Journal of the Schoenberg Institute
for Manuscript Studies, 2/1: 82–105.
Wales, H. G. Quaritch (1931), Siamese State Ceremonies, London: Quaritch.
Wudhichai Kosolkanjana วฒชย โกศลกาญจน (1988), Kan chamra wannakam rüang ongkan
chaeng nam การชาระวรรณกรรมเรองโองการแชงนา (‘A Textual Criticism of Ongkan Chaeng
Nam’), MA thesis, Silpakorn University.
Dmitry Bondarev, Darya Ogorodnikova
How to Spell Loanwords? Integration of
Arabic Etymons in Bilingual Islamic
Manuscripts of West Africa
Abstract: Different stages of the integration of Arabic loanwords into West Afri-
can languages are studied on examples of spelling patterns identified in two
distinct manuscript traditions of Old Kanembu and Soninke. Both languages are
written in Arabic script (Ajami) providing interlinear translations of Arabic
texts. These Ajami writings exhibit a high degree of Arabic loanwords, some
spelled etymologically and some deviating from the Arabic graphemic source to
the point of complete dissimilarity. The paper outlines a preliminary typology of
the spelling of Arabic loanwords in interlinear Ajami writings and suggests that
retentions and divergences from etymological spelling are probably motivated
by established orthographic practices specific to teaching–learning circles,
whereas the individual linguistic sensitivity of scribes does not seem to play a
prominent role in the selection of spelling features.

1 Introduction
Arabic came to sub-Saharan Africa with the spread of Islam, starting from the
ninth century CE.1 Lexicons of regional languages have since been under the
significant influence of Arabic. Many sub-Saharan societies with a long history
of Islam or contact with Muslims had parts of their vocabulary changed and
expanded by Arabic borrowings covering a wide semantic range, including
religion, administration, warfare, science, trade, time and counting systems and
mathematics. Some Arabic borrowings have been entirely integrated into the
lexicon of the target languages, so that the Arabic etymons have changed con-
siderably, for example, Soninke sèyìdì from Ar. shahīd ‘martyr’, Hausa lādân
from Ar. al-ʾaðān ‘muezzin’ or Kanuri ashâm from Ar. aṣ-ṣiyām ‘fast(ing)’. How-
ever, a large part of Islamic vocabulary remains transparent as to its source in
Arabic.

||
1 On the chronology of propagation of Islam in sub-Saharan Africa, see Hiskett 1984; Levtzion
and Pouwels 2000; Insoll 2003; Salvaing 2020.

Open Access. © 2024 the authors, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-010
256 | Dmitry Bondarev, Darya Ogorodnikova

The relative chronology, different mechanisms and types of borrowing from

Arabic into sub-Saharan languages have been predominantly studied based on
spoken data. How the process and result of borrowing were reflected in writing
is largely unknown. This is partially due to the paucity of early written data.
Epigraphic evidence, for example, of the interaction between Arabic and
Songhay on funeral inscriptions dates to the thirteenth to fourteenth centuries.
But the type of texts incised onto cliffs and tombstones is too short and too Ara-
bic-oriented to be of significance for investigating the borrowing processes.2
Writing on paper opens up a much more generous resource, which, however, is
of much later time. The most ‘ancient’ manuscripts where Arabic is accompa-
nied by local languages in Arabic script (Ajami) go back to the mid-seventeenth
century, and even this date is exceptional – only Qur’an manuscripts from Bor-
no have such a noble age. The rest of the sub-Saharan region of West Africa
feature Arabic-Ajami manuscripts dating, at best, to the late eighteenth century.3
This kind of bilingual and multilingual Islamic manuscripts from West Afri-
ca have only recently started emerging from obscurity, and the novelty of the
material explains its absence from the previous research on Arabic loans in sub-
Saharan languages. This paper is just a first step to complement studies of lexi-
cal borrowing in languages of West Africa by manuscript data by taking a pre-
liminary look into some select Ajami manuscripts from two unrelated regions of
the Senegambia and Lake Chad. Our slightly more ambitious aim is to sketch
some tendencies of loanword spelling observable in manuscripts and map dis-
tances from the original Arabic graphemic words to their spelling in the recipi-
ent language (what we call a ‘gradient of etymological spelling’). We then relate
the types of spelling thus obtained to the question of graphemic visual interac-
tion between the donor and recipient language (spelling affected by the visual
proximity to the original, and/or by the level of the scribe’s literacy) and will ask
whether the typology may have the potential for the study of the relative chro-
nology of borrowing and degrees of incorporation of the loanwords into the
recipient language.
The paper is structured as follows. We sketch out a brief history of adapta-
tion of Arabic script to the languages of West Africa in the next section. Section 3
deals with previous studies of channels and patterns of Arabic loanwords in
these languages. Section 4 starts with the specification of the kind of linguistic
data represented by the interlinear Ajami material under study (Subsection 4.1).
We then introduce our methods and principles for the study of loanwords in

||
2 De Moraes Farias 2003. On epigraphy in Essouk-Tadmakka, see Nixon 2017.
3 See a summary on types of Ajami manuscripts in Bondarev 2021, 708.
 How to Spell Loanwords? | 257

interlinear Ajami (Subsection 4.2). Section 5 presents an analysis of selected

etymons in Old Kanembu (Subsection 5.1) and Soninke (Subsection 5.2) manu-
scripts. Section 6 provides a comparative summary of our findings, and Section 7
draws some conclusions.

2 Adaptation of Arabic script: Ajami

Inscriptions on cliffs and tombstones provide the earliest evidence of the inter-
action between Arabic and local languages, for example, Songhay. But the
words in Songhay are very few, the texts in Arabic do not betray any linguistic
influence of Songhay, and nothing can be said about the process of borrowing
and adaptation of the Arabic lexicon at the time when the inscriptions were
made. We should, thus, start with much later written practices witnessed by
manuscripts on paper. The first such evidence comes with the Qur’an manu-
scripts of the Borno sultanate, what is now roughly north-east Nigeria and
south-east Niger. We find an enormous amount of annotations in Old Kanembu,
a language that was exclusively used for the translation of the Qur’an and other
Arabic texts used in traditional Islamic education and scholarship, in the inter-
linear and marginal space of the manuscripts written from the mid-seventeenth
century to the early nineteenth century.4
Soninke, the language spoken in north-eastern Senegal, south-eastern Mau-
ritania and western Mali, is another sub-Saharan language that has an early
written attestation comparable to Old Kanembu. At least one manuscript with
interlinear annotations in Soninke to a theological treatise by Muḥammad b.
Yusūf al-Sanūsī (d. 1486) is datable to the late eighteenth century.5 Writing in
Soninke must have been practiced much earlier however, because it was proba-
bly one of the earliest languages exposed to Islam in the western Sahel.6

||
4 Old Kanembu was a Western Saharan language closely related to Kanuri and Kanembu of
the same sub-branch in the Saharan languages. Circumstantial evidence points to much earlier
practices of writing in Old Kanembu, very possibly before the fifteenth century. On this topic
and also on relationship between Old Kanembu, its modern-day descendant exegetical transla-
tional language Tarjumo and Kanuri/Kanembu dialects, see, inter alia, Bondarev 2013a; 2013b;
2014a; and 2014b.
5 See Ogorodnikova 2023.
6 Soninke was the language of the ruling elite of ancient Ghana, one of the first West African
polities to adopt Islam; see Cuoq 1984.
258 | Dmitry Bondarev, Darya Ogorodnikova

There is no certain periodisation of Ajami writing in the other languages of

West Africa, and various publications give different chronological pictures.
Thus, Tamasheq, a Tuareg Berber language mostly spoken in the north of Mali,
has been reported to exist in a sixteenth-century manuscript, although the proof
of this claim is yet to be seen.7 Songhay, after its prominent, if fragmented, ap-
pearance in medieval stone epigraphy, seems to reappear in manuscripts only
in the late eighteenth century, but even this dating is speculative because the
scribes are typically tacit about the time of their writing. The chronology does
not improve in the case of Fulfulde (Fula), the language of the Fulbe (Fulani)
who live across West Africa from Senegal to Chad and were the major propaga-
tors of Islam across the whole region in the nineteenth century. Possible earliest
hard evidence has been found in manuscripts going back to the eighteenth
century.8 The Wolof language of Senegal, known for its significant body of con-
temporary literature in Ajami (Wolofal), was used in 1817 in a bilingual treaty
between the Bar kingdom and France.9 Incidentally, 1817 is also the date for the
first extant example of Ajami in Hausa, the language spoken in the past in sev-
eral urban kingdoms to later evolve into the most spoken language of West
Africa, centred in northern Nigeria and southern Niger.10
As is the case of Old Kanembu and Soninke, the other languages must have
existed in written form much earlier than the earliest manuscript evidence sug-
gests. A French travel account of the mid-seventeenth century, for example,
reports that Mandinka spoken in the lower Gambia region was written in Arabic
characters.11
In spite of the chronologically limited range of the extant manuscript data,
there is one clear tendency in the development of various Ajami orthographies,
especially visible in Old Kanembu, Soninke and Hausa writings: the older the
manuscripts, the more the conservative approach is manifest in the adaptation
of Arabic script. Thus, the Arabic characters in the earliest Ajami writings were
rarely modified to better represent the sounds of the target languages and no
innovation is visible before the late nineteenth century. In Old Kanembu, for
example, only nineteen of twenty-eight Arabic consonants were actively used

||
7 Gutelius 2000. On Ajami writing in the other Berber languages (beyond the scope of this
survey), see van den Boogert 1997.
8 Salvaing and Hunwick 2003.
9 Ngom 2017.
10 ‘The first dated example of Hausa Ajami (in a multilingual text) was written in the Caribbe-
an diaspora in 1817’, Bondarev and Dobronravin 2019, 254.
11 Cultru 1913.
 How to Spell Loanwords? | 259

for the reconstructed twenty-three consonants.12 The main selection principle

was homorganic proximity, i.e. the phonetic similarity in the place and manner
of articulation. The Arabic letter <d>, for example, was chosen for the sound [d].
One or more distinctive features are omitted in less straightforward cases. Thus,
for a bilabial voiceless plosive [p] absent in Arabic, the choice was either the
letter <b>, which encodes a phoneme with two features shared with the [p], i.e.
‘bilabial’ and ‘plosive’ (but lacking ‘voiceless’), or the Arabic letter <f>, sharing
(in its phonetic representation) the features ‘bilabial’ and ‘voiceless’ (but lack-
ing ‘plosive’).13
Some Arabic letters that represent sounds absent in Soninke, are used for
specific Soninke sounds in Soninke manuscripts. Thus, the letter ʕayn is con-
ventionally used for the Soninke /ŋ/. In rare instances, the letter is used as a
graphemic support for a vowel diacritic. The letter ðāl, which represents᷼ the
interdental fricative [ð] in Arabic, is ‘reinterpreted’ for the sound [t᷼ ʃ ] (hence-
forth [c] in simplified transcription), or sometimes for the sound [d ʒ] (hence-
forth [j]), the latter sound being also represented by the letter jīm. The same
letter ðāl is used in some manuscripts to encode the voiced palatal approximant
[j] (henceforth [y] in the conventional notation of the Soninke phonology).
The Arabic consonants that were only sparsely used in non-Arabic Ajami
content range from nine in Old Kanembu Ajami (ṣād, ḍād, ṭāʾ, ẓāʾ, ʕayn, qāf,
xāʾ, ðāʾ and šīn) to eight in Soninke (ṣād, ḍād, ṭāʾ, ẓāʾ, zāy, xāʾ, θāʾ and šīn).14
We will call these letters Arabic-specific. As will become obvious in the follow-
ing sections, the scribes of respective manuscript traditions had the letters in
reserve for spelling Arabic loanwords.
The special status of the Arabic-specific letters implies a conscious attitude
to spelling in multilingual writing which goes hand in hand with scribes’ self-

||
12 Bondarev and Dobronravin 2019, 244–245.
13 The encoding of vowels in early Ajami texts exhibits the same conservative principle. The
picture depicted here is intentionally simplified to emphasise the tendencies. However, there is
more to the selection of Arabic graphemes than can be sketched out in this short overview. For
more details on orthographies in various Ajami traditions, see Souag 2010 and Mumin and
Versteegh 2014; and for specific languages, see Bondarev 2014b on Old Kanembu; Dramé 2021
and Ogorodnikova 2023 on Soninke; Newman 2000 and Bondarev and Dobronravin 2019 on
Hausa; McLaughlin forthcoming on Atlantic languages, such as Fula and Wolof; Vydrine 1998
and 2014, and Tamari 2017 on Manding.
14 In Soninke manuscripts analysed by Dramé 2022, xāʾ is regularly used for the Soninke /x/.
In the Soninke manuscripts discussed in this study, the sound phoneme /x/ is usually ex-
pressed by the letter qāf, but can also be denoted by kāf. A rarer set of graphemes for this
sound is ɣayn and xāʾ.
260 | Dmitry Bondarev, Darya Ogorodnikova

perception, evident in how they labelled non-Arabic content with the term
ʕajami and other metalinguistic means.
The scribes were not just conscious of differences between their language
and Arabic – this is well expected in translational context. They were marking
their Ajami writing by metalinguistic means, such as a short tag signalling that
the writing is in Ajami, or describing their phrases as written ‘in our language’.15
The spelling choices made by the scribes in writing loanwords may have been
prompted by the use of such conscious metalanguaging, as will be discussed
later.16

3 Arabic influence on lexicons of West African

languages and types of borrowing
The Arabic lexical load on vocabularies of West African languages is substantial
and the languages spoken by communities with a long history of Islam have a
particularly high number of Arabic loans. Lexical borrowing comes about by a
variety of routes. The earliest attempt to map different channels and patterns of
Arabic borrowings across the breadth of languages of West Africa is by Paul
Wexler.17 With the exception of Sergio Baldi, later lexicographical studies dedi-
cated to Arabic loans have focused mostly on individual contagious or co-
regional languages.18
Our study is, in many ways, consonant with Wexler’s approach to disentan-
gle different patterns of borrowing from Arabic, and we briefly outline some of
his concepts here. In an attempt to ‘distinguish between borrowed vocabulary
defined by the speakers themselves as “Arabic” (regardless of the origin) and

||
15 Ogorodnikova 2017; Ogorodnikova 2023; Bondarev 2021.
16 On the concept of metalanguaging, see Maschler 1994.
17 Wexler (1980) examined data available at the time for sixteen languages: Tamasheq (Berber
family); Kanuri, Teda (Saharan); Dangaléat, Hausa, Mandara (Chadic); Maba; Fula, Wolof
(North-Central Atlantic); Temne (Mel, of Atlantic-Congo macro phylum); Songhay (Nilo-
Saharan?); Dagbani, Moore (Gur, of Atlantic-Congo); Bambara, Busa, Manding (Mande) and
Mbay (Central Sudanic).
18 Baldi 1999 deals with the periodisation of Arabic loans in Chadic languages and Baldi 2008
provides large-scale surveys of Arabic loans across African languages. Kossmann 2005 is an
exemplary study of Arabic loans in Hausa and Kanuri contrasted to loans from Berber. An
excellent overview of literature and topics dealing with the Arabic impact on sub-Saharan
languages is given in Souag 2016.
 How to Spell Loanwords? | 261

genuine Arabic loans no longer identified as such’,19 Wexler draws a distinction

between ‘Arabisms’ and ‘Arabic elements’. Thus, ‘Arabisms’ are (1) the words
purposefully borrowed from Arabic, precisely because of their Arabic origin,
and (2) the words recognised by the speakers to be of Arabic origin (even if
wrongly). ‘Arabic elements’ are Arabic loans ‘no longer recognised as such by
speakers’20 as a result of a ‘chance diffusion of Arabic elements’.21
Islamisation was a gradual process, initially involving elite social groups
and only later reaching out to wider populations, therefore, a useful distinction
has to be made between the loans coming at an earlier stage of demographically
narrower interaction with Islam and those introduced later into predominantly
Muslim societies. However, the identification of such different stages of borrow-
ing is complicated by ‘the possibility that the original norms of Arabic in the
target language may have undergone change’ and, therefore, ‘the age of the
borrowing cannot easily be determined simply from the type of formal integra-
tion in the target languages’.22 Another complicating factor is the possibility of
different or multiple contacts of the target languages with the secondary inter-
mediate channels of Arabic etymons. The eastern regions of the Hausa lan-
guage, for example, were exposed to indirect contact with Arabic from the east
through its contact with Kanuri, whose speakers were influenced by Islam from
the eleventh century, and then later, around the fifteenth century (and simulta-
neously with the ongoing Kanuri contacts), the western regions of Hausa came
into contact with Songhay. At the same time, Islamic education was constantly
developing, encompassing a greater number of people, leading to new or re-
newed direct borrowing from written Arabic.

The result of widespread reciprocal impact is that West and Central African languages fre-
quently have doublet forms of a single Arabic etymon – one form borrowed directly from
Arabic and one form introduced through a neighbouring lingua franca.23

A classification similar to Wexler’s has been advanced by Nico van den Boogert
in his study of the Berber literary tradition of the Sous.24 Van den Boogert distin-
guishes between four types of borrowing in written Berber: quotations, classi-

||
19 Wexler 1980, 524–525.
20 Wexler 1980, 525.
21 Wexler 1980, 526.
22 Wexler 1980, 529.
23 Wexler 1980, 531.
24 Van den Boogert 1997.
262 | Dmitry Bondarev, Darya Ogorodnikova

cisms, colloquialisms and Berberised loans.25 Quotations are loans from Classi-
cal Arabic that retain the standard Arabic orthography. Classicisms are borrow-
ings of nouns with their morphological properties, such as plural and singular
forms, the definite article and case affixes. Colloquialisms are borrowings from
Moroccan colloquial Arabic. Finally, the Berberised loans consist of the oldest
Arabic layer of borrowed verbs, and are distinguished from the other loans by
their complete integration into the Sous Berber morphology.
Types of borrowing discussed by Wexler and van den Boogert are commen-
surate with the general framework of lexical borrowing. Gerrit J. Dimmendaal,
drawing on the examples of African languages, outlines different types and
mechanisms of borrowing.26 In order to avoid terminological ambiguity, we first
list some of the relevant concepts discussed by Dimmendaal, and then align the
terminology we use in our study with that of Wexler, van den Boogert and Dim-
mendaal.
Our data (in Section 5) is introduced from the less to maximally integrated
loans, therefore, our selection of terms mentioned by Dimmendaal is also given
in a similar order. The first term is unadapted borrowing, which refers to the type
of borrowing when items are not phonologically integrated into the target lan-
guage. Unadapted borrowing occurs ‘in particular when there is widespread
bilingualism’.27 The second type is morphosyntactically integrated borrowings
without phonological integration of the borrowed item.28 The third type is loan
blends or hybrids, which refers to ‘partial morpheme substitution in loan-
words’.29 The fourth type is phonological adaptation, that makes the loanword
completely integrate into the target language by various mechanisms, from the
insertion and deletion of sounds to phone substitution and phonological re-
structuring.30
In our analysis of Arabic loans, we try to ascertain to what extent the scribes
were making intentional decisions in their spelling choices (from less integrated
to more integrated). This sociolinguistic dimension is also discussed by Dim-
mendaal under the notion of an ‘act of perception’,31 whereby the speakers are
conscious of their multilingualism, prompting unadapted borrowings.

||
25 Van den Boogert 1997, 223–224.
26 Dimmendaal 2011, 179–188.
27 Dimmendaal 2011, 182.
28 Dimmendaal 2011, 182.
29 Dimmendaal 2011, 185.
30 Dimmendaal 2011, 182.
31 Dimmendaal 2011, 183–184.
 How to Spell Loanwords? | 263

We are now in a position to introduce the terminology we use in our study

in comparison to the concepts by the three authors mentioned above. Note that
not all concepts are fully compatible with each other.

Table 1: Terminology in comparison.

This study Wexler van den Boogert Dimmendaal

insertion Arabisms (1) quotations unadapted borrowing
the words borrowed
from Arabic purposefully
copy-spelling; Arabisms (2) classicisms morphosyntactically
etymological spelling the words recognised by integrated
the speakers as of
Arabic origin
partial etymological loan blends or hybrids
spelling
complete integration; Arabic elements Berberised phonological adaptation
adapted spelling

4 Types and mechanisms of borrowing

discernible in manuscripts
All data on Arabic loans in African languages, except for van den Boogert’s
research into the Berber literary tradition,32 have come so far from spoken
African languages. The general assumption that written Classical Arabic was
a significant (or even the primary) point of contact between West African
languages and Arabic has never been corroborated by the study of Ajami
writing.33
The type of written data presented in this study has the potential for testing
various assumptions about borrowing processes related to written Arabic. Un-
like independent Ajami texts that exist on their own without direct relationship
with Arabic source texts, the interlinear annotations are in constant interplay

||
32 Van den Boogret 1997.
33 Wexler’s important conclusion that ‘written Arabic can be considered the most important
contributor of Arabic elements to the West and Central African languages spoken by Muslims’
(Wexler 1980, 556) results from his study of grammars and dictionaries that do not deal with
literary variants of the languages.
264 | Dmitry Bondarev, Darya Ogorodnikova

with the Arabic in the translational source-target frame. At the same time, this
kind of written texts has its difficulties and limitations. We, firstly, discuss the
peculiarities of such written data and then introduce our method of dealing with
the spelling of Arabic loans.

4.1 Peculiarities of Ajami texts in interlinear annotations

Linguistic data represented by the Ajami manuscripts under study are highly
specialised. The Ajami texts consist of glosses that translate the Arabic texts,
mostly of a religious genre which deal with various subjects of the classical
Islamic curriculum. The manuscripts were produced in the domain of inter-
mediate and advanced classical Islamic education.34 Ajami glosses represent
a written counterpart of oral exegetical practices based on the translation of
small units of the source text, most typically parsed into short noun phrases
and verb phrases. In spite of such a source-dependent relationship between
the units of the parsed Arabic texts and the units of translation, the transla-
tional output is rarely influenced by the grammatical structures of Arabic,
and the phrases of the target language are usually grammatically well-
formed.35 At the same time, the Ajami glosses represent specialised exegetical
varieties of the respective languages which differ from the ordinary spoken
languages and may, therefore, be incomprehensible to the speakers outside
the learned circles.36 One specific characteristic of these exegetical and trans-
lational registers is their technical vocabulary, drawn largely from Arabic.37
In this context, loanwords are often preferred over the language-internal
vocabulary to unambiguously convey the concepts of the main text. Such an
increased load of the Arabic vocabulary in exegetical spoken and written
practices makes it difficult to draw a line between the borrowings already
incorporated into the target language and the specialist ‘technical’ terms
used exclusively in the educational and learned domain.
A similar problem of differentiating borrowings from specialist vocabulary
was identified by Maarten Kossmann in relation to the linguistic content of Ber-
ber manuscripts:

||
34 Bondarev 2017; Ogorodnikova 2021.
35 Tamari and Bondarev 2013, 9; Bondarev 2022.
36 Tamari and Bondarev 2013, 15–22; Bondarev 2022.
37 Tamari and Bondarev 2013, 15–22.
 How to Spell Loanwords? | 265

In Islamic treatises and admonitions, the text genre entails the usage of much Arabic vo-
cabulary. Some of this is no doubt genuine borrowing, while others are necessary inser-
tions in order to explain concepts not nameable otherwise. There are also many terms that
seem to be inserted from Arabic, even though there are Berber forms available.38

Van den Boogert also stressed the increased ‘use of Arabic loans instead of an
Existing [sic] Berber word’ as a characteristic of Berber manuscript verse texts.39
Another complicating factor in the analysis of borrowings in the glosses is
that literate scribes were conversant with the phonology and morphology of the
source languages. Therefore, they probably adhered to accurate pronunciation40
and orthographic rules, both of which may trigger the etymological spelling of
Arabic borrowings. As a result, manuscript glosses are likely to capture and
reflect scholarly attitudes towards borrowings rather than general tendencies or
patterns of integration into ordinary language. Moreover, the scribes’ awareness
of the words’ etymologies may result in ‘original/authentic’ spellings even when
the words had been well-integrated and undergone phonological adaptations.

4.2 Method: Identification of distance from etymological

spelling
Aware of peculiarities of the linguistic material represented by Ajami annota-
tions, we set out the principles employed in our analysis of Arabic borrowings
discernible in manuscripts. The basic procedure is to compare the spelling of a
given Arabic loanword in Ajami with the original Arabic orthography of the
word. The comparison is carried out following what we call a gradient of etymo-
logical spelling schematised in Table 2. The types of spelling are organised in
descending order from the spelling identical to Arabic to the spelling unrelated
to the Arabic original orthography.
Working with the etymological scale of graphemic representation, we pay
close attention to the visual proximity of a given Ajami loanword to the origi-
nal Arabic word. This is important to ascertain the extent to which the scribes
are influenced in their spelling choices by the immediate graphemic prompts
of the source orthography, rather than by their mental map based on the
knowledge of Arabic orthography acquired during their previous cycles of
learning.

||
38 Kossmann 2013, 47.
39 Van den Boogert 1997, 52.
40 Tamari and Bondarev 2013, 16.
266 | Dmitry Bondarev, Darya Ogorodnikova

Table 2: Types of etymological spelling of Arabic loans in Ajami: a scalar.

Type of spelling
1 Identical: all graphemic segments retained
2 Partial
2a All consonants retained
2b Some consonants retained
2c Vowels retained
2d Some vowels changed
3 Dissimilar spelling: none of the original spelling is retained
4 Sensitivity to Arabic morphology: Arabic definite article
4a retained
4b absent

We also try to distinguish various characteristics of borrowing, partly drawing

on Wexler’s work discussed in Section 3, in our analysis of selected examples
presented in Section 5.41 The characteristics include relative chronology, chan-
nels of borrowing, the impact of written Arabic, integration into the target lan-
guage, scribes’ recognition of Arabic elements in their language, and semantic
fields of loanwords. These characteristics are presented in Table 3.

Table 3: Characteristics of borrowing.

Earlier vs later layers of borrowed words (including multiple cycles)

1 Direct vs intermediary source (including multiple sources)
2 Impact of scholarly and written Arabic vs non-written, colloquial Arabic
3 Non-integrated vs fully integrated etymons
4 Recognised vs unrecognised Arabic elements
5 Semantic fields related to Islamic concepts vs unrelated/non-bound to Islamic concepts

A systematic investigation of all these characteristics is beyond the scope of the

present study. However, we will take them into consideration where possible in
order to relate a preliminary typology of borrowing in Ajami texts to more gen-
eral questions of processes of borrowing.

||
41 Wexler 1980.
 How to Spell Loanwords? | 267

5 Case studies
We now turn to the analysis of loanwords in Old Kanembu (henceforth, OKb)
and Soninke Ajami manuscripts. Differences in the type of manuscripts and
linguistic practices of translation represented by each Ajami tradition dictate a
slightly different organisation of data. Old Kanembu glosses of the Qur’anic
manuscripts retain much more archaic features not present in modern Kanuri.
The glosses in Soninke are translations of non-Qur’anic texts and the similarity
between Soninke glosses and modern spoken Soninke is much greater than that
between OKb and Kanuri.

5.1 Old Kanembu

As already mentioned, OKb is a written language exclusively used for the trans-
lation of Arabic texts. In this paper, we deal with the earliest written attestations
of OKb preserved in Qur’anic manuscripts produced in the Borno sultanate from
around the mid-seventeenth century to the early nineteenth century. Old
Kanembu or, more precisely, its later written variants outlived the epoch of
written Qur’anic translation practiced in Borno until the beginning of the nine-
teenth century.42 The non-Qur’anic texts have continued to be translated into
written OKb. The written translational language has functioned as a downsized
counterpart of the oral exegetical language known locally as Tarjumo, also ap-
plied in exclusively translational contexts.43 Although it is possible to trace a
continuum of incremental linguistic changes from the earliest witnesses of OKb
to the modern-day Tarjumo, the term ‘Old Kanembu’ refers to the language of
the Qur’anic manuscripts before the early nineteenth century. Both written OKb
and spoken Tarjumo are closely related to the Kanuri language spoken in north-
east Nigeria and south-east Niger (around and off the western shores of Lake
Chad) and to the Kanembu language spoken in the areas north, north-east and
east of the lake.44 It will, therefore, be necessary to compare some of the exam-
ples of Arabic loans in OKb with their equivalents in one of the major Kanuri
dialects, Yerwa Kanuri. A wide-scale comparison across available data of the
other Kanuri and Kanembu dialects is outside the scope of this paper.

||
42 Bondarev 2024 and forthcoming.
43 Bondarev and Tijani 2014.
44 Kanuri and Kanembu are usually considered a continuum of dialects.
268 | Dmitry Bondarev, Darya Ogorodnikova

Kanuri is one of the West African languages with a large number of Arabic
loans, and various aspects of a borrowed lexicon in Yerwa Kanuri have been
studied in previous literature.45 John P. Hutchison described the Kanuri lexicon
as chronologically layered (a notion applicable to almost any human language,
no doubt) ‘in the sense that it is possible to recognize words in the language that
are older as distinct from words that have not been in the language as long’.46 As
our primary focus is written attestations of Arabic borrowings, we will assess
the extent to which their spelling may help to unravel such chronological layers.
The discussion of Arabic loans in OKb is organised along the scale from et-
ymologically identical (copy-spelling) to completely divergent spelling. The
scale also takes into consideration the visual relationship between the target
and source etymons. The spelling scale is, thus, subdivided into the following
components.

Table 4: Spelling scale of loanwords.

Type of spelling and visual proximity to Subtype

the source word
1 Copy-spelling (of entire lexeme) above the Copy-spelling, no source word present
corresponding source word
2 Partial etymological spelling above the
corresponding source word
3 Partial etymological spelling, not written
above the source word
4 No traces of etymology, but written above
the corresponding source word
5 No traces of etymology, no source word
present

These types of spelling are exemplified below by select etymons. Each etymon is
introduced as a vocabulary entry followed by explanations. Each entry is pro-
vided with a reference to a manuscript and the Qur’anic text (e.g. 1YM/91:9
reads ‘from the manuscripts 1YM, chapter 91, verse 9’).

||
45 Greenberg 1960; Baldi 2002; Baldi 2003; Baldi 2007; and Baldi 2020.
46 Hutchison 1981, 10.
 How to Spell Loanwords? | 269

5.1.1 Etymological spelling written above corresponding Arabic words

Most of the etymons in this category consist of the terms related to theology,
religious practice and conduct, nouns designating specific entities of geograph-
ic and botanical nature unknown in the host culture, and proper names (the
latter is, however, not dealt with here).
{1} <aflaḥa> ‘to prosper’.47 The verb is a direct borrowing from Arabic ʔaflaḥa
‘to prosper’. In the manuscript, the word is spelled without the initial hamza
according to the Warsh reading of the Qur’an and the word is, thus, an exact
orthographic copy of the original. At the same time, it functions as a typical OKb
verb with its own inflectional morphology, which is characterised in the exam-
ple given by the suffix <-jī> denoting the 3rd person subject morpheme and the
perfective category: <aflaḥa-jī> ‘he has gained reward’.
{2} <ṭūr> ‘Mount Sinai’,48 written above Ar. ṭūr ‘mount (Sinai)’. Since Mount
Sinai is the place of revelation of the Torah, the passage about Sinai is com-
mented upon by another OKb term borrowed directly from the orthography of
Arabic: <tawrē> ‘Torah’, the pronunciation specific to the Warsh reading. The
same OKb spelling is found written next to the original Arabic term, as in
1YM/3:50.
{3} <aṣiyām> ‘fasting’,49 above Ar. al-ṣiyām. This is a special case of phonetic
spelling of the assimilated definite article before the coronal consonants (‘sun
letters’). The Arabic word ṣiyām ‘fasting’ when used with the definite article al-
is written as <al-ṣiyām> and pronounced as [aṣ-ṣiyām]. The OKb loan <aṣiyām>
is written above the Arabic <al-ṣiyām>, but, nevertheless, reflects a phonetic
form. At the same time, the OKb written word does not indicate the phonetic
gemination [ṣṣ] in its spelling, although the Arabic spelling clearly has the gem-
ination marked by the diacritic shadda.50
{4} <kun-ẓūlm> ‘wrongdoing, injustice’,51 above Ar. ẓulm. This loanword en-
tered the language at the time when there was a productive nominal derivation
through the prefix k- (and its variant kVn-, where V = vowel). Such prefixed
forms are among the oldest integrated loans, and, nevertheless, the Arabic
source of some of these is obvious to the scribes who copy the orthography of

||
47 1YM/91:9. Note that in transliteration of OKb the letter ɣayn is represented as <g>.
48 1YM/2:93.
49 1YM/2:183, 187.
50 That OKb <aṣiyām> is borrowed from the definite article form (rather than being a phonetic
adjustment by insertion of an initial prosthetic vowel [a-]) is confirmed by the OKb loan that
starts with the consonant <ṣ> <ṣūm> ‘to fast’ borrowed from the Arabic verb ṣūm ‘to fast’.
51 1YM/2:61, 114; 4:160.
270 | Dmitry Bondarev, Darya Ogorodnikova

the corresponding Arabic etymon. A similar case of scribal awareness is attested

in the loanword <kaṣūm> ‘fasting’, which is, however, not found in visual prox-
imity to the source word (and, therefore, discussed with category 3 below), un-
like its more frequent and supposedly later loan <aṣiyām> ‘fasting’ discussed
above.

Fig. 1: The OKb loan <aṣiyām> is written above the Arabic <al-ṣiyām>.

Many words in this category start with the letter ʕayn <ʕ>, which is rarely used
in non-Arabic etymons.
{5} <ʕaṣ> ‘to disobey’,52 above Ar. ʕaṣā. The OKb verb form <ʕaṣ-inmi> (diso-
bey-AUX.2s.PRF) ‘you disobeyed’ is written above Arabic ʕaṣay-ta (disobey-
2s.m.PRF) ‘you disobeyed’. The verb has been fully integrated into the Kanuri
lexicon, having the phonological shape ashi.
{6} <ʕaðāb> ‘punishment’,53 above Ar. al-ʕaðābu. Although the OKb form is a
copy of the Arabic orthography, it does not retain the case ending -u necessarily
present in the written Qur’an, but not always pronounced in recitation.

(1) 1YM/2:86 tandīkan ʕaðāb-yi itskāybō

3p.on punishment-SJ lighten.3s.IMPF.NEG

‘the punishment will not be lightened for them’

||
52 1YM/10:91.
53 1YM/2:86.
 How to Spell Loanwords? | 271

However, the etymon is spelled as <ʕaðābu>, i.e. with the final vowel -u, in other
places of the same manuscript. This probably shows a process of phonetic ad-
justment to a more typical OKb syllable structure that avoids plosive consonants
at the end of the syllable. This three-syllable form has been retained in Modern
Kanuri: azáwu.
{7} <ʕinab> ‘grape’,54 above Ar. ʕinab-an (-an is the suffix of the accusative
case). The OKb word is borrowed from a singular Arabic form without case end-
ings and is invariable in OKb. Thus, the same singular form is used to translate
the plural ʔaʕnāb-an ‘grapes-ACC’ in 1YM/78:32. The OKb item has a final vowel
after the consonant: <ʕinab-e>, probably due to the same phonetic adjustment
described above.
{8} <ʕadas> ‘lentil’,55 above Ar. ʕadas. In the other manuscripts, the same
Arabic agricultural term is translated by vernacular generic terms for beans:
3ImI <gālū>, 1YM <gālgāla>.

5.1.2 Partial retention above corresponding Arabic words

These OKb words are probably well integrated, and, at the same time, the
scribes are conscious about their origin, which shows in the selection of specifi-
cally Arabic graphemes for the spelling of the OKb words.
{9} <ʕājab> ‘wonder’.56 This etymon is a good example of a fully integrated
word that does not always copy the source vowels, but invariably retains the
etymological spelling of the consonants. The visual proximity to the source
etymon does not, therefore, have a defining effect on the orthography of the
vowels, as explained below.
The OKb loan <ʕājab> is based on the Arabic noun ʕajab ‘wonder’ and is used
for not only the translation of this noun but also the Arabic adjective ʕajīb ‘wonder-
ful, amazing’. In 1YM/72:1, the Arabic phrase innā samiʕnā qurʔānan ʕajaban ‘In-
deed, we have heard an amazing Qur’an’ is translated into OKb as <alqurʔān
ʕajabū=ka krniyē> (the Qur’an amazing=DO we have heard). Although the conso-
nants of both qurʔān ‘the Qur’an’ and ʕajab ‘wonder’ are retained in the OKb
spelling, the OKb phrase is morphologically independent from the source text.
Thus, the absence of the article al- in the source word ‘Qur’an’ is ignored and the
word is given in its borrowed form with the article: al-qurʔān. The Arabic noun (in

||
54 1YM/80:28.
55 2ShK/2:61.
56 1YM/11:72, 50:21, 72:1.
272 | Dmitry Bondarev, Darya Ogorodnikova

attributive function) ʕajab-an is translated as an adjective in OKb, and has the final
vowel <u> to form a permissible phonotactic structure in OKb (i.e. avoiding sylla-
bles with final plosive consonants).
The Arabic adjective ʕajīb(un) ‘wonderful’ (derived from the noun ʕajab
‘wonder’) in 1YM/11:72 is again translated by the OKb word <ʕajab(u)>, which (1)
ignores the second vowel <ī> in the source word and (2) functions as a noun (in
the OKb associative construction ‘with’) rather than an adjective.

(2) 1YM/11:72 attī agō ʕājab-wa

this thing wonder-with

‘this is an amazing thing’ (lit. ‘this is a thing with wonder’)

Fig. 2: Arabic adjective ʕajīb(un) ‘wonderful’ is translated by the OKb word <ʕājab>.

The scribe’s awareness of the Arabic source of this otherwise fully integrated
lexeme becomes more obvious when compared to the genuine OKb word <agō>
written next to it. If <agō> starts with an ʔalif and the vowel sign [a] fatḥa above
it, the initial grapheme of the loan word is a copy of the source letter ʕayn that
corresponds to the guttural voiced pharyngeal fricative sound [ʕ] in Arabic but
is absent in Saharan languages including OKb. However, the rest of the OKb
word is spelled deviating from the original. The most significant difference is in
the vowel of the first syllable. It is a short vowel [a] in Arabic, whereas it is writ-
ten as <a> + <ʔalif> in OKb, the combination representing a long [ā] in Arabic,
but probably used here to mark a high tone á (cf. Kanuri cognate ájap ‘wonder’,
ájabba ‘amazing’). Similarly, the vowel of the second syllable in Arabic is
spelled as <i> + <ya’>, standing for a long [ī], but it is <a> in OKb.
Other OKb words written with the letter ʕayn, such as <ʕabat> ‘to worship’
and <ʕalam> ‘to teach’, are in the same category of spelling.
{10} <ʕabat> ‘to worship, serve’, from Ar. ʕabada. The spelling of the final con-
sonant in the OKb lexeme denotes a voiceless stop [t] rather than the voiced [d] of
the Arabic source. The non-voiced feature shows that the verb had long been inte-
 How to Spell Loanwords? | 273

grated into the language. The OKb phonotactics observable in the manuscripts
does not require a syllable-final voiced /d/ to become voiceless in any environ-
ment. Consider the inflectional forms of the verb <ʕabat> ‘worship’ (3) and <wud>
‘pour’ (4) for comparison, both having the auxiliary verb n suffixed to the final root
consonant. In the case of <wud>, there is no change from [d] to [t].

(3) 1YM/3:51 nadīyi tīka ʕabat-n-ūgū

2p.SJ 3s.DO worship-AUX-2p.IMV

‘worship Him!’

(4) 1YM/80:25 andiye wud-n-ē

1p.SJ pour-AUX-1p.PRF

‘we poured (water)’

At the same time, there are other Arabic verbs with the last consonant /d/ that
changed in OKb to the root-final /t/, as in the OKb verb <wat> ‘promise’ from
Arabic wāʕada ‘promise’. The change of the final /d/ of the source word to the
final /t/ in the target should then be due to some earlier sound change rules no
longer productive at the time of the OKb of the Qur’anic manuscripts.57
Conversely, the Arabic verbs with the final voiced plosive consonant /b/ re-
tained the voiced feature of the consonant when borrowed into OKb: <jarab>
from Ar. jarraba ‘to test’, <jab> from Ar. jāwaba ‘to answer’.
{11} <ʕalam> ‘to teach’ from Ar. ʕallama. The OKb form is invariable in its
syllabic and vocalic structure (i.e. CV-CVC), irrespective of the Arabic form that
has the geminated /ll/ and changes vowels according to inflection. In
1YM/2:102, for example, this OKb verb <ʕalam> is written above Arabic yuʕal-
limāni ‘two (angels) were teaching’:

(5) 1YM/2:102 malayka dīro trgsā yamka ʕalam-kisa-ḥalan

angels two were sent to people teach-ing-ADV

‘two angels were sent to teach people’

||
57 The final voiced ‘emphatic’ /ḍ/ in some Arabic nouns also changed to [t] in the process of
borrowing, but not systematically: thus, the OKb loan from Arabic al-ʔarḍ ‘earth’ is sometimes
written as <lart>, sometimes as <larḍ> (see below item {30}).
274 | Dmitry Bondarev, Darya Ogorodnikova

{12} The last example in this category of loans with a recognisably Arabic-
specific letter is a lexeme that starts with the latter šīn, rarely used outside Ara-
bic borrowings. This is the verb <šawar> ‘to consult’ in 1YM/3:159, that trans-
lates the Arabic shāwir-hum (consult.2s.m.IMV-them) ‘consult them!’, as in <ni-
ye tandīka shawar-nē> (you.SJ they.DO consult-2s.IMV) ‘consult them!’. The
same lexeme is used as a noun in 1YM/2:233, where it translates the Arabic
phrase wa tashāwurin ‘(by mutual consent) and consultation’ as follows:
<šawari-bi gābukan> (consultation-GEN after.from) ‘after consultation’.

5.1.3 Borrowing from Arabic, not written above the source word

This group of loanwords retains some etymological graphemes but is not writ-
ten in proximity to the corresponding source word. The Arabic corresponding
terms exist in the Qur’an elsewhere, and the same OKb lexeme is sometimes
also used to translate its etymological source.
{13} <kaṣūm> ‘fasting’, from Arabic ṣūm ‘to fast’. This is another k-prefixed
borrowing comparable to <kun-ẓūlm> ‘wrongdoing’ in {4} that, in spite its an-
cient integration into the language, was recognised by the scribes as an Arabic
loan. But, unlike <kun-ẓūlm>, the etymological spelling is not prompted by the
corresponding Arabic word (e.g. ṣūm ‘to fast’): no such Arabic word is present in
the verse (Q2:51) where OKb <ka-ṣūm> ‘fasting’ occurs. However, the event de-
scribed in the verse refers to Moses’s fasting on Mount Sinai and, thus, the OKb
phrase explains this as follows:

(6) 1YM/2:51 būnēkami fīdegbi kaṣūm-nn watkīnīyē

night.from forty.GEN fasting-ADV we.appointed

‘we appointed forty nights of fasting (for Moses)’

{14} <ḥaṣm>/<ḥaṣum> ‘to argue’ from Ar. xaṣm ‘opponent’ or xāṣama ‘to argue’.
In 1YM/4:109, the verb <ḥaṣm> translates a different Arabic lexeme jadala ‘to
argue’, rather than the verb xāṣama ‘to argue’.

(7) 1YM/4:109 ʔālagin amū dū ḥaṣm-nū

God.with persons who argue-3p.FUT

‘who will argue (on their behalf) with God?’

 How to Spell Loanwords? | 275

In 1YM/22:19, a participle form of <ḥaṣum> is used to translate its etymological

source, also in nominalised form. Thus, the Arabic xaṣmāni ‘two adversaries’
corresponds to the OKb <ḥaṣum-kita-ndi> (argue-ing-two) ‘two arguing ones’.
The OKb word is fully integrated in both cases, but retains the etymological
connection through the use of the letter ṣād. Notably, the first consonant is
represented by the letter ḥāʾ (it denotes a fricative voiceless pharyngeal sound
in Arabic) rather than by the original xāʾ (used for the same set of articulatory
features, except for the place of articulation, which is velar/uvular rather than
pharyngeal). This is interesting in light of the existence of the OKb loans that
copy the etymological x, e.g. <xalq> ‘to create’ from Ar. xalq. In modern Kanuri,
the lexeme corresponding to the OKb <ḥaṣm> does not have the initial conso-
nant: ásǝm ‘argument’.
{15} <safr> ‘to travel’ from Ar. safar ‘voyage’.58 The OKb loan verb is used to
translate a different Arabic verb ḍaraba with the synonymous meaning ‘to travel’:
<nadīye safr-nīyūya> (above Ar. wa ʔiðā ḍarab-tum ‘and when you travel’).

Fig. 3: The OKb loan verb <safr> is used to translate a different Arabic verb ḍaraba.

The same loan may function as a noun ‘journey’ in OKb and is used to interpret
the corresponding Arabic noun. Thus, the Ar. asfarinā ‘our journeys’ in
1YM/34:19 is translated into OKb as <safr-dē> ‘journey-our’. Note that the Arabic
plural is not expressed in OKb.
The Arabic term safar ‘travel’ is sometimes translated by a genuine OKb word
<blāgrū> (Kanuri bǝláwúró), rather than by the loan <safr> (as in 1YM/2:184).
It is difficult to assess whether the scribes meant to copy the original or-
thography with items such as <safr>. This is because all consonantal letters of
the original word are equally used in the genuine OKb lexemes. Therefore, such
words overlap with those in the following two categories, but, at the same time,
represent items in their own right because the Arabic words to which they corre-

||
58 1YM/4:101.
276 | Dmitry Bondarev, Darya Ogorodnikova

spond do not have specific consonants that can be used as a diagnostic of the
scribal orthographic awareness.

5.1.4 Written above, but no etymological spelling

The orthography of the lexemes in this category do not have any signs of
original etymology, even though they are used as translations of the original
Arabic etymons and are actually written above or in close vicinity to the
source words.
{16} <ka-sūgu> ‘market’, from Arabic -suq ‘market’, written above Arabic
ʔaswāq ‘markets’.59 This k-prefixed word has been discussed extensively in the
literature on lexical contact in West Africa as an unambiguous example of bor-
rowing pathways from Arabic into a language of wider communication (lingua
franca), such as Kanuri, and the k-prefixed word’s subsequent spread into many
other languages.60 The fact that the spelling of OKb <ka-sūgu> is not sensitive to
the source word suq, let alone its plural form ʔaswāq, seems to betray the
scribes’ oblivion of its Arabic connection.
{17} <alū> ‘writing board’ from Ar. al-lawḥ. The OKb item has almost no
traces of the original word, except for the element of the definite article manifest
in the initial vowel <a>. The consonant /l/ could probably have been pro-
nounced as a geminated [ll], which would make it closer to the source word,
and this is the case in Kanuri: alló. The example below (8) and Fig. 4 demon-
strate an interplay between this well-integrated word and an orthographic
copy/insertion, although the latter has the long vowel misplaced (to the third
syllable instead of the second).

(8) 1YM/85:22

Q. lawḥ-in maḥfūẓ-un

OKb alū maḥfuẓū=gin kaθikō

tablet preserved=in be.3s.AFP

‘in the preserved tablet’

||
59 1YM/25:7, 20.
60 Greenberg 1960; Kossmann 2005; Baldi 2020.
 How to Spell Loanwords? | 277

Fig. 4: OKb <alū> ‘writing board’ and maḥfūẓ ‘preserved’ written above the corresponding
Arabic phrase.

The simultaneous use of the non-etymological and etymological spelling of two

etymons in the same short phrase suggests that the word <alū> was not per-
ceived by the scribe as a loan from Arabic.
{18} <jagaru> ‘blast’,61 from Arabic zajra ‘shout, cry, blast’ (variants: <jgaru>
(2ShK), <jugaru> (Arabe 402)). The lexeme is found in the phrase <jagaru tilō>
(blast one) ‘one blast’ written above Ar. zajratun wāḥidatun (blast first) ‘the first
blast’ (of the Day of Judgement). The first consonant of the OKb <jagaru> repre-
sented by <j> probably stands for the voiced non-palatal affricate [dz],62 which
would be a close approximation to the Arabic [z] zajra. The second consonant
<g> points to a source other than Classical Arabic, to a dialect where the Classi-
cal Arabic /j/ corresponds to /g/. But this alone does not explain why a collo-
quial Arabic source would be considered for a very specific word used in the
context of the Day of Judgement.63
{19} <ʔala> from Ar. allāh ‘God’. The etymon has the most transparent con-
nection to the original Arabic word, and it is remarkable that the OKb employs
this very specific spelling convention for the most significant religious concept
God. However, this non-etymological spelling only occurs in the first part of
1YM, about 65% of the entire 470-folio manuscript (see item {30} for more de-
tails). The remaining 35%, and indeed many other manuscripts, exhibit the
etymological spelling <allāh>.64

||
61 1YM/79:13.
62 Bondarev 2014b.
63 Lameen Souag confirms the doubts about a colloquial source (Souag’s comments on an
earlier draft of the paper).
64 The manuscripts that have the etymological spelling <allāh> or, much more often, non-
vocalised spelling, are T.Kano, 3ImI, 4MM, Kad.Ar.33, 5.Konduga.
278 | Dmitry Bondarev, Darya Ogorodnikova

5.1.5 Borrowing from Arabic, but not associated with the source language

The Arabic words in this group of borrowings are either not used in the Qur’an
or used only once or twice.
{20} <algāma> ‘wheat’, from Ar. al-qamḥ (Kanuri algáma, lǝgáma,
laáma). The Arabic word qamh ‘wheat’ does not occur in the Qur’an, and the
OKb term is written above the Qur’anic Arabic fūm (Q.2:61), which may mean
either ‘garlic’ or ‘wheat’.65 The OKb scribes interpreted the Arabic word fūm
as ‘wheat’ following certain exegetical sources.66 This is also corroborated by
the very choice of the OKb word, or rather, by the absence of the OKb term for
‘garlic’, which would resemble the modern Kanuri kǝngálmú ‘garlic’ (unat-
tested in OKb).
Due to its significance for the reconstruction of the scribes’ perception of
the integrated loanwords, the whole sentence translating various agricultural
terms in this Qur’anic verse is illustrated below in (9). The two Arabic loans
<algāma> ‘wheat’ and <albāsar> ‘onion’ are integrated into the list of the ver-
nacular botanical terms, and it is obvious that both terms are not perceived as
loans. Thus, <algāma> ‘wheat’ does not have its original equivalent in the verse
and elsewhere in the Qur’an, whereas <albāsar> ‘onion’ is written above the
corresponding Arabic term in the spelling, entirely disconnected from the origi-
nal word (baṣal).

(9) 1YM/2:61

Q. min baqli-hā wa qabaθāʔi-hā wa fūmi-hā wa ʕadasi-hā wa baṣali-hā

OKb kālō klī-jī fālī-jī kāmi algāma-jī gālgāla-jī kāmi albāsar-

kāmi kāmi ji=kami

leaf green-its watermelon-its wheat-its beans-its from onion-its from

from from from

‘from its (earth’s) herbs, its cucumbers, its wheat, its beans, its onions’

||
65 Badawi and Haleem 2008, 721.
66 Ibn Kathīr (d. 1373) referred to this ambiguity in his tafsīr: ‘Others said that fūm is wheat,
the kind used for bread. Al-Bukhari said, “Some of them said that fūm includes all grains or
seeds that are eaten”’ (https://quran.com/2:61/tafsirs/en-tafisr-ibn-kathir; accessed on 12
January 2024). Similar explanations in Arabic are found in various OKb manuscripts.
 How to Spell Loanwords? | 279

{21} <kn-jikl> ‘hardship’, from Ar. shikāl (pl. shukul) ‘fetter(s) for shackling the
feet of riding animal’ or shakl ‘state of confusion’.67 The k-prefixed word does
not exist in modern Kanuri, and the cognate Kanuri term shíwol ‘distress, trou-
ble’ must have entered the language at a later stage and possibly from a source
different to <knjikl>.68
{22} <fahama> ‘hear’, from Ar. fahima ‘understand’. Some obvious loans
that retain Arabic orthography seem so well integrated into the language that
their spelling peculiarities should be considered part of the core graphemic
inventory of OKb. Such is the word <fahama>, with the letter hāʾ normally used
in borrowings and as a word-spacing device.69 The lexeme is used as a verb ‘to
hear’ in the OKb manuscripts. It is unlikely that the OKb term was borrowed
from the Qur’an since there is only one occurrence of the Arabic fahima in the
Qur’an, where it occurs in the causative derivation fahhama ‘to make under-
stand, explain’ (Q.21:79). What the Okb term <fahama> consistently translates is
the Arabic samiʕa ‘to hear’ and, thus, the loan word has its own path of seman-
tic derivation from ‘understand’ to ‘hear’, making it completely independent
from the source language (however present the latter may be in such a transla-
tional relationship). The only occurrence of the Arabic fahima (a causative
fahhama) is translated into OKb by another OKb verb fam, <fam/fan> also with
the meaning ‘to hear’ (<itfamgiye> ‘we made him understand’).70

5.1.6 Definite article retained

Many Arabic words came into OKb and Kanuri-Kanembu with the Arabic defi-
nite article al-. The article in such loans may be written as either <al-> or only

||
67 1YM/90:4. As pointed out by Lameen Souag in his comments on an earlier draft of the
paper, the Arabic shakl as a borrowing source raises the problem of the difference in vowel, /a/
in Arabic vs /i/ in OKb. It is also possible that the OKb lexeme is internal Saharan rather than a
loan from Arabic. Thus, in Tubu, the verb kakal ‘belästigen’ has been attested by Lukas (1957,
185), with the basic root kal. This would not, however, explain the affricate <j> in the OKb <kn-j-
kl>, and the matter, therefore, is left for future research.
68 Apart from the prefix k-, the major difference here is in the vowel quality in the second
syllable of shíwol, rather than in the quality of the second consonant, because the historical
sound change k > w in the intervocalic position is common in Kanuri.
69 Bondarev and Dobronravin 2019, 249.
70 It is unclear whether the OKb <fam/fan> (Kanuri fan) is a different borrowing of the same
Arabic fahima or an internal Saharan root with such cognate forms as bas in Teda-Daza and
Bakoore Beria kɛbɛnô given regular correspondences between Kanuri-Kanembu f- and Teda-
Daza and Beria b- (Chonai 1999, 21, 183, 217–218).
280 | Dmitry Bondarev, Darya Ogorodnikova

retaining the second segment – the consonant /l/, or as reflecting a form assimi-
lated to the following coronal consonants/graphemic ‘sun letters’ (see etymon
{3} <aṣiyām> ‘fasting’). We only consider the two first types of the retained arti-
cle, <al-> and <l->, in this subsection.
The borrowings with the definite Arabic article are listed here in a way simi-
lar to the previous examples, from the closest etymological spelling to the items
spelled independently of Arabic source words. Given that the Arabic definite
article is an easily identifiable grammatical item with a very high occurrence
frequency, it is expected that the scribes who constantly deal with the Arabic
and OKb content would naturally analyse the OKb words containing the initial
element al- as Arabic loans. Some of the article-based OKb loans have other
distinctive Arabic graphemic elements and, thus, plausibly represent the con-
scious Arabic-related repertoire of scribal choices. However, the orthography of
some OKb words with the retained definite article is unrelated to the original
Arabic words, making such loans more difficult to classify from the point of
view of scribes’ etymological awareness. This discrepancy between the clear
presence of the Arabic definite article and non-transparent spelling of the rest of
the word also complicates categorisation of the loanwords that retain the Arabic
article in a clipped form, with the initial vowel lost, such as <laqabur> ‘grave’
from Arabic al-qabr. If the word ‘grave’ is spelled with the Arabic letter <q>,
irrespective of the changed (phonetic and graphemic) shape of the article al-,
why does a more obvious article-based loan <al-bāsar> ‘onion’ from Arabic al-
baṣal not reflect the original ṣad <ṣ> and is spelled instead with the letter sīn
<s>? A plausible answer to this question could only be obtained through the
analysis of a much larger set of data than that selected for this study.

5.1.6.1 Words with the Arabic article al- retained in full

{23} <alaxira> ‘afterlife; the hereafter’,71 from Ar. al-ʔaxira ‘the afterlife’ and
written above the Arabic word used in genitive case (al-ʔaxira(ti) ‘of afterlife’).
This genitive grammatical relationship is translated into OKb with the respec-
tive genitive marker -be (<alaxira-be> ‘of afterlife’). The only deviation from the
original spelling is the disregard of the tāʾ marbūṭah ending -ti in the Arabic al-
ʔaxira-ti, but this is typical in almost all borrowings in OKb or Kanuri-Kanembu.
{24} <alqurʔān> ‘the Qur’an’, from Arabic al-qurʔān. This OKb word denot-
ing the sacred scripture always occurs with the definite article integrated into
the lexeme, irrespective of the grammatical status of the original word. Thus,

||
71 1YM/79:25.
 How to Spell Loanwords? | 281

the OKb <al-qurʔān> in 1YM/72:1 is written above the Arabic word without the
article (qurʔān). The form with the whole definite article is, however, less fre-
quent in this manuscript than the one with the clipped initial vowel (see
<luqurān> and its variants in the next list). However, the form with the whole
article is preferred in the manuscript T.Kano: <alqurān> (Q.46:2). At the same
time, the T.Kano variant lacks the original hamza <ʔ> in its spelling.
{25} <albāsar> ‘onion’. Ultimately from Arabic al-baṣal, via an intermediary
language. The same etymon in Kanuri has the article integrated into the clipped
form: luwásar.
{26} <algāma> (repeated {19}) ‘wheat’. Ultimately from Ar. al-qamh, but bor-
rowed via an intermediary source. The etymon in Kanuri occurs in three (idio-
lect/dialect-specific) forms: algáma, lǝgáma, laáma.72

5.1.6.2 Words with the definite article without the initial vowel a, i.e. <l->
{27} <luqurān>/<luqrān> ‘the Qur’an’.73 Unlike the form with the unchanged
definite article al- (etymon {24}), this is fully integrated into the language and
the only sign of the original etymology is the letter <q> rather than the Okb
standard <k> used for the voiceless velar stop /k/. It is this clipped form that
exists in modern Kanuri, with the velar being elided between vowels: luwurân.
{28} <laqabur>/<laqabr> ‘grave’,74 from Ar. al-qabr (pl. al-qubūr). The form
with the clipped article is not known in Kanuri, where the cognate (borrowed)

||
72 See Wexler 1980, 536, on the problem of borrowing paths of both etymons (onion and
wheat). Baldi (2008, 244–245) provides an extensive list of the etymon ‘onion’ across forty
different languages, the majority of which have either the whole or clipped article in the first
syllable. The etymon ‘wheat’ with the integrated Arabic article is attested in twenty languages,
see Baldi 2008, 415–416. The form with the voiced [g] after the article al- is only reported in
Kanuri and Munjuk (a Bio-Mandara [Central Chadic] language spoken in Cameroon and Chad).
Lameen Souag’s suggestion that the voiced [g] (g < q) ‘would readily reflect Libyan or Sudanese
Arabic’ may well explain the voiced feature of the velar in this OKb/Kanuri borrowing (Souag’s
comments on an earlier draft of the paper).
73 1YM/17:105 and 1YM/2:176 respectively.
74 <laqabur> in 1YM/9:84, 82:4; <laqabr> in 1YM/35:22, 100:9. In the case of <laqabur>, the
insertion of a vowel into the consonant cluster (here, -br- in the source word al-qabr) to form a
permissible syllable structure is a typical mechanism in Kanuri. Exemplarily, K. laadar (lagadar)
‘capacity’ < Ar. al-qadr ‘extent’; K. lámar ‘event, situation’ < Ar. al-ʔamr ‘matter, affair’. The quality
of the inserted vowel /u/ in <laqabur> is conditioned by the preceding bilabial /b/. Lameen
Souag suggested an alternative source from a Maghribi colloquial form *l-ǝqbǝṛ (Souag’s com-
ments on an earlier draft of the paper). However, since the common pattern of vowel insertion
plausibly justifies Classical Arabic as a direct source of borrowing, a distant link to Maghribi
sources seems unnecessary.
282 | Dmitry Bondarev, Darya Ogorodnikova

word is káwar. However, the OKb variant has been preserved in Tarjumo:
lawura.
{29} <lqalam>, variant <lqālm> ‘pen’,75 from Ar. al-qalam. All three known
instances of the word in 1YM exhibit a peculiar combination of the copy-spelling
and phonetic encoding. The letter <q> is a reference to the source word, but the
integration of the article in the clipped form rather than the whole <al-> sug-
gests an adapted pronunciation [lǝkalam] or, possibly, [lǝkálam] if the long
vowel in the variant <lqālm> encodes a high tone. This is plausible, since the
Kanuri cognate also has a high tone on the second syllable: alkálam. However,
unlike OKb, the Kanuri form (and its numerous variants alkǝrǝm, alláram,
ar’álam and arkǝrǝm) has the whole article integrated into the word.
{30} <larḍ>/<lart> ‘earth’, 76 from Ar. al-ʔarḍ ‘the earth’. The word al-ʔarḍ oc-
curs 461 times in the Qur’an, and only with the article, which is always preced-
ed by a vowel. This results in the pronunciation [l-arḍ]. The prominence of this
phonetic form in the Qur’an suggests that the OKb (and Kanuri/Kanembu) bor-
rowed the term directly from the Qur’an. Note, however, two distinct types of
spelling: one with the copy of the original letter <ḍ> for the emphatic voiced
alveolar stop [ḍ] and the other with the letter <t> used for the voiceless alveolar
[t]. The former spelling convention <larḍ> is consistent in the first part of the
manuscript 1YM, whereas the latter <lart> corresponds to the second part. The
watershed between the two spelling types runs in Q.34. In Q.34:1, the etymon is
written as <larḍ> and in the next verse, Q.34:2 (and in different hand!), the same
word is written as <lart>. It is hard to ascertain the significance of this ortho-
graphic difference. On the one hand, being an obvious deviation from the Ara-
bic original, the variant <lart> seems to suggest an adapted pronunciation inde-
pendent from the source word, something like [lartǝ]. This is plausible given
that the modern Kanuri word is lárdǝ ‘earth’ and that historically, the voiceless
alveolar stop /t/ became voiced between sonorant/vowel and vowel (e.g. atǝ > adǝ
‘this’). On the other hand, the same second part of the manuscript consistently
copies the orthography of the etymon ‘God’ <allah> ( ). This co-occurrence is,
however, reversed in the first part of the manuscript. There, the etymon ‘earth’
is copy-spelled as <larḍ>, but ‘God’ is written as <ʔāla> (‫)ءﺍﻝ‬, probably reflecting
pronunciation (cf. Kanuri Ála, with the first vowel having a high tone). The two

||
75 <lqalam> in 1YM/68:1, 96:4; <lqālm> in 1YM/31:24.
76 <larḍ> is found in the first part of 1YM (e.g. Q.2:30, 36, 116, 117, 164, 3:50, 4:97, 18:14, 27:25),
whereas <lart> occurs in the second part of 1YM (e.g. 50:7, 44, 51:20, 57:4, 5, 10, 79:30, 81:2
84:3, 86:12, 88:20, 89:21, 91:6). Other spelling variants are rare, e.g. <lārd> in 1YM/9:2 and
<lard> in 1YM/78:6. In 2ShK/91:14, the word is also written as <lart>.
 How to Spell Loanwords? | 283

parts of the manuscript 1YM are different in many other ways as well,77 and it is
plausible that they were produced by different students taught within different
schools of Qur’anic interpretation, each of which had specific preferences for
the choice of phonetic- and etymology-based spelling.
{31} <k-l-īslām> ‘Islam’. Unlike the other k-prefixed loan nouns, the source
for this word was the Arabic lexeme with the definite article al- (al-ʔislām). The
orthography of the OKb <klīslām> is consistent and invariable, irrespective of
the form and proximity of the source word. Thus, in 1YM/3:19, it is written above
al-ʔislām, in 1YM/2:208, above al-silm ‘submission to God, peace’ and in 1YM/2:257,
it is used to specify the Arabic word nūr ‘light’ in the sense of the ‘light of Islam’:
<klīslām-bi nūr>. The OKb k-prefixed form based on the source word with the
article (a)l- did not find its way into Kanuri, where the concept of Islam is ex-
pressed by a prefixed form kǝr-mǝsǝlǝm ‘Islam’, which derives from the Arabic
word without the article muslim ‘Muslim’.

5.2 Soninke
Soninke was the language of the Ghana empire, which was among the first sub-
Saharan polities to adopt Islam.78 Despite the long history of contact with Islam
and, hence, the Arabic language, there was relatively scarce research on Arabic
loans in Soninke. Although examples of Arabic borrowings into Soninke have
been documented in lexicographical works79 and quoted in comparative sur-
veys,80 the only work (we are currently aware of) addressing this topic in detail
is that of Seydina-Ousmane Diagana.81
The present section examines how Arabic borrowings are represented in
Soninke Ajami writings.82 The materials discussed here appear in manuscripts
mainly from the early nineteenth century. These manuscripts were produced by
speakers of Mandinka,83 who resorted to a related language, Soninke, as a
means of exegesis.84 Theological treatises and Islamic law manuals are among

||
77 Bondarev forthcoming.
78 Cuoq 1984.
79 Dantioko 2003; Baldi 2008; Ousmane Moussa Diagana 2013; Bathily 2017.
80 Souag 2016.
81 Seydina-Ousmane Diagana 1992.
82 On spelling of Arabic borrowings in Soninke Ajami, see also Dramé 2021.
83 The scribes used their native Mandinka language to write colophons and some glosses, which
they marked with the phrase fī kalāminā ‘in our language’. See Ogorodnikova 2017 and 2023.
84 See Ousmane Moussa Diagana 1995, 19; Tamari and Bondarev 2013, 18; Tamari 2016, 45.
284 | Dmitry Bondarev, Darya Ogorodnikova

the texts annotated by the scribes. Because of the texts’ specific nature, the
scribes extensively incorporated words of Arabic origin.
The examples of borrowings in this section are divided into three groups,
depending on their spellings: (1) etymological, (2) partially etymological and (3)
non-etymological. The spelling of borrowed words does not seem to depend on
their proximity to the source word.
The spellings of borrowings are compared to their assumed source, on the
one hand, and to the modern-day Soninke form, on the other. It is frequent
across the manuscripts under examination that the same word is spelled differ-
ently by either the same or different scribes. As a result, the spellings range from
less to more distant from the assumed etymon. Although spelling variants and
their frequency are considered when categorising examples into a particular
group, the statistical analysis of these variations and occurrences of each item
will be carried out in the future.

5.2.1 Etymological spelling

This group includes borrowings that retain the orthographies of the source in
the Ajami glosses, even though many of these words in modern-day language
have undergone phonological changes. It comprises concepts and words central
to Islam, such as God and the holy scripture, words from the religious exegesis
and practices.
{32} <alqurʔān> al-qurʔān ‘the Qur’an’. The term for the holy scripture of Is-
lam is one example of complete etymological spelling throughout the manu-
script corpus, with consonantal base and vocalisation rendered exactly the
same as in the source. The lexeme is attested in two forms in modern-day spo-
ken language – with and without the definite article – àlixùràané and
qùráanà/qùràané.
{33} <allāh> Állà ‘God’. The word ‘God’ in Soninke Ajami glosses is often
written in proper Arabic spelling. However, spellings reflecting the phonetically
adjusted form Állà are also frequent: <alā> or <ala> (the medial consonant gem-
ination being underrepresented). The latter graphical representation with ʔalif-
lām is usually attached to the following segment, similar to that in <alataʕalā>85
Állà tāʕalā ‘God the Exalted’ or <alamakiri>86 Állà mà gìrí ‘God is eternal’ (lit.
‘God did not come to an end’). Some spellings differ from the source entirely,

||
85 EAP 1042/9/2 p. 22.
86 BULAC MS.ARA.219bis fol. 2r.
 How to Spell Loanwords? | 285

replacing the initial ʔalif with ʕayn, followed by lām, i.e. <ʕala>87, or lām-ʔalif
maqsurah, i.e. <ʕalā>88 (corresponding to the spelling of the Arabic preposition
ʕalā ‘on, upon’).

5.2.2 Partial etymological spelling

Several etymologically spelled borrowings exemplified below reproduce the

consonantal base (including long vowels) but may differ in vocalic notation.
{34} <ṣāḥibi> saahibe ‘companion’ from Ar. ṣāḥib (pl. aṣḥāb, ṣaḥb) ‘associ-
ate, companion’. The Soninke plural suffix -nu is often attached to the core
form, spelled etymologically, i.e. <fāri ṣāḥibinu> fàaré saahibe-nu ‘prophet’s
companions’. Word formation with indigenous suffixes suggests borrowing
rather than insertion or copy-spelling. This word is not listed in dictionaries,
which suggests that its usage is specific to exegetical contexts.
{35} <tālibi> táalíbè ‘student’ from Ar. ṭālib (pl. ṭullāb) ‘student’. The main
phonological difference between the source and the borrowing is the emphatic
/ṭ/ transformed into an alveolar voiceless plosive /t/. Most Ajami spellings re-
tain the initial consonant of the Arabic original, but sporadic spellings with tāʾ
for the adapted form also occur. Similar to the previous example, the word pro-
duces derivatives with Soninke suffixes (e.g. <ṭālibinu> táalíbè-nú ‘students’),
indicating its integration into the Soninke lexicon. The term refers to learners at
intermediate and advanced stages of classical Islamic education and has several
synonyms, such as qàrànlénmè ‘pupil, student’ (a compound formed with an
Arabic borrowing; see below).
The next group of examples contain the sounds /ħ/ (fricative voiceless
pharyngeal) transliterated as <ḥ> and the glottal fricative /h/. While the former
is outside of the Soninke phonemic inventory, the latter exists in eastern dia-
lects, regularly corresponding to f in western dialects.89 The /ħ/ in the Arabic
borrowings is usually adjusted to /h/ (without being changed to the dialectal f)
in the spoken domain.90 In the written domain, however, the tendency is to
retain the graphemes of the original: <ḥaramu> hàráamè ‘forbidden’ from Ar.
harām ‘forbidden, unlawful’, <ḥaqi> háqqè ‘right’ from Ar. ḥaqq ‘right’, also

||
87 BL Or. 6473 fol. 206v; UbL Or. 14.052(5) fol. 1r.
88 BL Or. 6473 fol. 221r; PGL ORI 11/2 fol. 12r–v.
89 Creissels 2016, 13. Ousmane Moussa Diagana (1995, 19) notes that due to intellectual migra-
tion, the language of religious exegesis is characterised by systematic usage of f (and not h).
90 See Seydina-Ousmane Diagana 1992, 207; Ousmane Moussa Diagana 2013, 77–78.
286 | Dmitry Bondarev, Darya Ogorodnikova

with attached Soninke suffixes:91 <ḥijāna> híijáanà ‘pilgrim’ from Ar. ḥajj/ḥijjah
‘pilgrimage’ + -áanà (NMAG). At the same time, the glottal fricative /h/ is repre-
sented by the letter hā: <jāhilāku> jàahìláaxù ‘ignorance’ from Ar. jāhil ‘igno-
rant’ + -áaxù (ABSTR), <ẓāhirinkīti> jàahìrìnkíitè ‘transparent judgement’92 from
Ar. ẓāhir ‘visible, clear’ + Soninke kíitè ‘judgement’, and <halaki> hàlákì/fàlákì
‘ruin, destroy’ from Ar. halaka ‘to perish, to be destructed’. However, these
graphemes are sometimes used interchangeably. Yet, there is one word that
appears consistently across the corpus in the source spelling with hāʾ even
though in the phonetically adjusted form it is dropped:
{36} <fahamu> fàámù ‘understand’ from Ar. fahima/fahm ‘under-
stand/understanding’.93 While the modern Soninke form does not have the me-
dial /h/,94 the corresponding Arabic grapheme hāʾ is systematically retained in
writing. At the same time, the verb behaves as fully integrated, taking the
Soninke grammatical items, such as aspectual suffixes, for example, the imper-
fective gerundive suffix -nV in the example below:95

(10) Gloss to [kull mā] yafhamu al-jāhil [minhi] ‘[everything that] the ignorant understands
[from it (the Qur’an)]’: <jāhilinun kwʾ fahamunu>

jàahìli-nú-n gà wá fàámù-nú

ignorant-PL-D SBD INACP understand-GER

‘the ignorant (ones) understand’

The following examples demonstrate approaches to spelling Arabic words with

voiced alveolar /z/ (encoded with the grapheme zāy) and voiced interdental ð
(encoded with ðāl), both of which correspond to /j/ in Soninke borrowings.
{37} <lazimi> làajímí ‘become obligatory’ from Ar. lazima ‘to be necessary’. This
is another word of the exegetical domain in which the original consonant (zāy) of
the source word is retained in the Ajami spellings. Although in a few cases, it is

||
91 Morphological categories are marked based on Creissels 2016.
92 See this example in Dramé 2021, 195, and also MAAO AF 14722(87) fol. 154v.
93 Seydina-Ousmane Diagana (1992, 128) quotes fahama as the basis of the Soninke borrow-
ing. For the cognate Bamana verb fàamu, Zappa (2011, 237) suggests either fahm (maṣdar)
‘understanding’ or ifham (2sg m imperative) ‘understand!’.
94 Creissels (2016, 25) argues that the vowel sequence à-á of the word hàámù/fàámù resulted
from the dropping of the h in the intervocalic position.
95 MAAO AF 14722(87) fol. 190v.
 How to Spell Loanwords? | 287

also spelled with jīm, thus, betraying its oral counterpart pronounced as /j/. De-
spite the etymological spelling, <lazima> behaves as a genuine Soninke verb:

(11) Gloss to bi-mā yalzamu ‘with what is necessary, required’: <ʾa ka wu sirī lazimini>

a gà wó sèré làajímí-ní

3S SBD INACP person become.obligatory-GER

‘that [what] is obligatory to a person’

It also produces various derivatives, such as làajímí-yé ‘obligation’ (‘become

obligatory’ + NMLZ) and làajímí-nd-áanà (‘become obligatory’ + ANTP + NMAG)
‘necessary, requiring’.96 The core-element <lazimi> tends to be spelled etymolog-
ically in all these forms.
The retention of etymological spelling in the verb làajímí contrasts with the
spellings of jíidì ‘to increase’ from Ar. zāda or zid.97 The latter is much more
frequently spelled with jīm or ðāl, rather than the original zāy.
{38} <ðunubin> jùnúbù ‘sin’ from Ar. (s. ðanb) ðunūb ‘sins’. The target spell-
ings retain the initial consonant ðāl of the source, even though it is pronounced
as an affricate /j/ (for which the grapheme jīm is a better match). The vocaliza-
tion pattern of the Soninke item suggests that it derives from the Arabic plural
form ðunūb. Contrary to the attested modern form (for singular and plural)
jùnúb-ù,98 the spellings in the glosses mark the last vowel with kasra <ðunu-
bin>,99 giving /junub-i/ or /junub-e/. Thus, the scribes possibly reinterpreted the
borrowed junub-u as the Soninke plural and, using the -e ending, derived its
singular form.100

||
96 The gloss làajímíndáanà corresponds to the main text’s (active participle) lāzim ‘necessary,
requisite’.
97 Zappa (2011, 237) suggests the imperative zid of the verb zāda as the etymon of the Bamana
jíidi ‘increase’.
98 The Soninke plural is formed by the suffixes -o, -u and -nu. The singular word forms ending
with -e regularly form the plural by changing the final vowel to -u, while singular forms ending
with -u form the plural by attaching the suffix -nu. The Soninke words with the singular marker
-u, including Arabic borrowings, have the lowest rate of occurrence. See Ousmane Moussa
Diagana 1995, 54–58. Thus, jùnúb-ù seems to have an exceptional and irregular number mor-
phology.
99 AAN1 p. 100; BL Or. 4897 fol. 168rv; JRL MS 780[825] fol. 37r.
100 However, the glosses are sometimes ambiguous as to whether the form <ðunubin> is to be
interpreted as singular or plural. Additionally, I could not find out whether there are forms
such as /junubu/ or /jubuni-nu/ for the plural.
288 | Dmitry Bondarev, Darya Ogorodnikova

{39} <ḥarafu> harafe ‘letter’ from Ar. ḥarf ‘letter, consonant’. An opposite
case is the word harafe ‘letter’, in the modern language attested with the final -e
but spelled in manuscripts with the final ḍamma for /u/, i.e. <ḥaraf-u>. This
corresponds to the Arabic source form in the nominative case ḥarf-un. From the
examples below in the glosses, it follows that <ḥarafu> is used in singular (12),
and forms the plural by attaching the suffix -nu (13):101

(12) Gloss to [mā xaṭṭa bi yadihi] ḥarfun ‘[he (the Prophet) did not write with his hand] a
letter’: <ḥarafu bāni ya>

haraf-u báané yá ní

letter-SG one FOC COPEQ

‘a single letter’

(13) Gloss to bi-al-ḥurūf ‘with the letters’: <tiḥurufūnu>

tí harafu-nu

with letter-PL

‘with letters’

It is plausible that the item haraf-u was at some point reinterpreted as a plural,
producing the modern-day singular haraf-e.
{40} <xibārindiyi> qìbáarìndíyè ‘information’ from Ar. xabar pl. ʔaxbār
‘news’. The Soninke qìbáarì-ndí-yè (qìbáarí-ndì ‘to inform’ + -yé NMLZ) ‘infor-
mation’ seems to be another borrowing which derived from the broken Arabic
plural noun. The Soninke qìbáarì/qìbáarè, with the long vowel /ā/ in the second
syllable, corresponds to the Arabic plural ʔaxbār with the initial vowel omitted
(i.e. xbār). Interestingly, in one instance, the scribe retained the entire Arabic
plural form (i.e. retaining the initial vowel and adjusting vocalic notation) in the
Soninke borrowing <axbārin di yin dū> /axbaarindiyen du/ qìbáarìndíyèn dí ‘in
the news, information’ when translating the respective Arabic phrase fī al-
ʔaxbār.102 Given the close proximity of the source word, copy-spelling of the
form /axbaarindiyen/ cannot be ruled out. However, it is then unclear as to why
the scribe omitted the definite article al- and added the Soninke suffixes.

||
101 Examples 12–13 are from MAAO AF 14.722(87) fol. 208r and fol. 199r.
102 MAAO AF 14722(87) fol. 105r.
 How to Spell Loanwords? | 289

In this and other instances, the scribes retain the etymological grapheme
xāʾ in the spelling of qìbáarè/qìbáarìndíyè (without the initial vowel). Although
the Arabic grapheme xāʾ is a good match for the Soninke uvular /q/ or its allo-
phone [χ], it is not a common convention to use it in writing indigenous lexemes
in the Ajami material under examination, and the grapheme qāf is typically
used. Another example of retaining the source orthography <xayran>103 with the
grapheme xāʾ is qéerì ‘joy, fortune’ from Ar. xayr ‘prosperity’.
The graphemic representations of borrowings in the following group are
somewhat ‘hybrid’, since they reflect phonetically adjusted forms while retain-
ing traces of the source orthographies, especially conspicuous with Arabic-
specific graphemes.
{41} <qara>/<karā> qàrá ‘study, read’ from Arabic qaraʔa ‘read, recite’.
Some of the scribal spellings replicate the orthography of the etymon, retaining
the final ʔalif (although without hamza). The final ʔalif may appear even when
the first consonant is not spelled etymologically, also in composites: <ka-
rānmūdi>104 qàrànmóodì ‘teacher’. The usage of the long vowel in the second
syllable of <qarā>/<karā> could be explained by the scribal intention to mark
the high tone. However, the example of the compound word <karānmūdi> shows
that a long vowel is marked even when it bears a low tone, and it is probably a
reference to the source orthography.
{42} <ṣali> sállì/-è ‘pray/prayer’ from Ar. al-salāh/ṣalla ‘prayer/pray’. This
word could possibly count among the oldest borrowings due to the antiquity of
Islam among the Soninke. The scribal spellings often conform to the original
orthography by using the emphatic ṣād, even in compounds, such as <ṣali yi>105
sállí-jí ‘ablutions (lit. prayer-water)’ or <ṣalifana> sállìfàná ‘afternoon prayer’.106
However, as indicated earlier, the etymological and adapted spellings coexist,
and one finds spellings with ṣād or sīn even on the pages of the same manu-
script.107

||
103 BULAC MS.ARA.219bis fol. 9v; TCD MS 3499 fol. 47v.
104 BL Or. 4897 fol. 154v.
105 UbL Or. 14.052(8) fol. 21r–v. The grapheme yāʾ in this spelling apparently represents a
palatal glide /y/ rather than rather than the affricate /j/ expected in the Soninke word jî ‘water’.
This may evidence a specific Soninke dialect or an influence of a scribal native language.
106 EAP 1042/9/2 p. 3. The Soninke gloss sállìfàná corresponds to the Arabic ẓuhr ‘midday
prayer’ in the main text. As Lameen Souag explains, the Soninke term sállìfàná ‘lit. prayer-first’
is in fact a calque of Arabic or Berber term ṣalat al-ʕūlā ‘the first (among the obligatory) pray-
er(s)’, with it another appellation for ẓuhr. See Souag 2015, 361–362.
107 BnF Arabe 5675 fol. 71v, fol. 74r, fol. 75r; TCD MS 2179 fol. 20v, fol. 21v.
290 | Dmitry Bondarev, Darya Ogorodnikova

{43} <laṣili>/<lasili> lásìlì ‘origin’ from Ar. al-ʔaṣl ‘origin’. The word retains
two etymological features: the definite article (even though with the clipped
initial vowel) and the emphatic <ṣ>, that fairly often appears in the spellings of
the otherwise adapted CV-CV-CV form. Spellings with sīn, indicating a phonetic
adjustment to /s/, are also found in the glosses.
{44} <laʕada> láadà ‘custom, tradition’ from Ar. al-ʕādah ‘custom’. Similar
to the previous example, the word was borrowed with the (clipped) definite
article. The manuscripts display remarkably consistent spelling of this item with
lām-ʕayn-dāl (‫ )ﻟﻌﺪ‬vocalized with fatḥas.
{45} <sariʕa>/<šariʕā> sàríyà ‘law’ from Arabic šarʕ ‘Revelation, divine law’
or šarʕīah ‘canonical Islamic law’. Interestingly, the phonetical form of the
Soninke borrowing sàríyà corresponds more to the Arabic šarʕīah, whereas its
graphical representation reproduces the Arabic šarʕ almost identically. The
variants with sīn at the beginning of the word likely signal the phonetically
adapted form, while those with šīn evoke the original etymon. As for the reten-
tion of the final ʕayn, it does not seem to have any phonological motivation
since it usually represents the velar nasal /ŋ/ or serves as a support for a vowel,
which would then give /sariŋa/ or /saria/. Thus, the retention of the ʕayn seems
to be a purely visual reminder of the source, and, in this case, is to be interpret-
ed as /y/ for /sariya/.
{46} <niʕima> néemà ‘peace, grace’ from Ar. niʕmah ‘benefit, blessing’. The
etymological spellings with the medial <ʕ> appears alongside an adjusted or-
thography of the word with yāʾ and kasra for /ī/: <nīma>. Therefore, the ʕayn is
conceivably a way of encoding the long vowel, while, at the same time, referring
to the word’s source orthography.
Similar to examples 42–45, the feminine suffix -ah encoded by tāʾ marbūṭah
is omitted in the spelling, and the last syllable is vocalised with fatḥa (for /a/).108
By contrast, the word umatunu ‘peoples’109 from Ar. ummah (pronounced [um-
matun]) ‘community’, is a borrowing derived from the literary Arabic form (with
/t/ at the end) followed by the Soninke plural suffix -nu. The phonetically inte-
grated tāʾ marbūṭah ending resulted in spelling the last syllable with tāʾ <t>:
<ʾumatunu> or <ʾumantūnu>.110

||
108 The word sóora ‘sura’ from Ar. sūrah may occasionally be spelled with tāʾ marbūṭah <ẗ>,
i.e. <sūraẗ> (e.g. UbL Or. 14.052(8) fol. 72v). However, spelling <sūra> without tāʾ marbūṭah also
occur (e.g. BL Or. 4897 fol. 155r).
109 The word unmatoonuu (pl.) ‘prophet disciples’ is listed in Mody Bathily’s dictionary as a
borrowing from Arabic ummah, see Bathily 2017, 362. A cognate word mànton ‘people, nation,
group’ is attested in Bamana.
110 BL Or. 6473 fol. 214r; TCD MS 2179 fol. 13r.
 How to Spell Loanwords? | 291

5.2.3 No etymological spelling

Several words of Arabic origin display spellings quite dissimilar to the source
orthography. Some of these words possibly came via an intermediary language,
and the Arabic etymology of some is even disputable.
{47} <sʔumu>/<sumu> súumè/súumù ‘fast’ from Ar. ṣaum ‘abstention, fast-
ing’.111 The adapted spelling of the word appears twice in the proximity of the
source word.
In the first case (14), it explains taṣūmu ‘she is to fast’.

(14) <ā nā sʔumu kafunu ki furu yā>

à nà súumì kàfínì ké fòró yá

3S SUBJ fast with DEM blood FOC

‘she should fast with this bleeding’

In the second instance (15), the gloss is linked to the Arabic phrase wa lā
yaqṭaʕu-hā ‘and he does not interrupt it’, in which the pronoun -hā relates to the
ṣiyām ramaḍān ‘the fast of Ramdan’ written two lines above.

(15) <ʾa nti ramaḍāna sumuyi kutunu>

à ntá Ramaḍān súumìyé kútú-nú

3S INACP.NEG Ramadan fasting interrupt

‘he does not interrupt the Ramadan fasting’

Interestingly, the name of the month in the phrase <ramaḍāna sumuyi>

Ramaḍān súumìyé ‘Ramadan fasting’ is spelled etymologically, while the word
súumìyé appears as well integrated, both in spelling (no etymological conso-
nant) and derivation (with the Soninke nominalising suffix -yé).
{48} <bataqin>/<batāqi>/<batāki> bàtáaxè ‘letter, message’ from Arabic
biṭāqah ‘slip of paper, card’.112 The Soninke lexeme does not appear in the prox-
imity of the source word in any instances and is used to explain different Arabic

||
111 Seyidina-Ousmane Diagana (1992, 201) suggests the 2 pl. imperative form ṣūmū ‘you all
fast!’ as a basis for the Soninke borrowing.
112 Seydina-Ousmane Diagana 1992, 197.
292 | Dmitry Bondarev, Darya Ogorodnikova

words, such as ṣaḥāʾif ‘leaves, pages’ or jarīdah ‘list, register’. The spellings of
the Ajami forms – with slight variations: <bataqin>, <batāqi> or <batāki>113 –
reflect the phonetically adjusted form. Thus, the initial bāʾ is vocalised with
fatḥa (for /a/) and the second consonant is written as ṭāʾ as opposed to the em-
phatic ṭāʾ of the original, followed by the long or short /a/. The consonant of the
last syllable is ambiguously represented by the letters qāf or kāf, each of which
can encode both /x/ and /k/. The tāʾ marbūṭah is omitted (cf. examples above).
{49} <hakili> háqílè ‘mind, intelligence’ from Ar. ʕaql ‘sense reason’. The
spellings of this word range from etymologically transparent to obscurer forms,
i.e. from <ʕaqili> or <ʕakili> to <ḥaqili>/<ḥankili> or <hankili>, with the latter spell-
ings predominant in the corpus.114 The variation between kāf and qāf may be pure-
ly graphical since both graphemes are used for uvular /q/, which occurs both in
the source and Soninke borrowing. On the other hand, spellings with kāf <ḥakili>
may, in fact, reflect Mandinka pronunciation with the velar /k/. Regarding the
initial letters ḥāʾ or hāʾ, they stand for the Soninke /h/, which, in its turn, is a
result of the phonological adaptation of the initial /ʕ/ of the source word.115
{50} <ḥariziki>/<ḥarziqi> ‘favour, fortune’ from Arabic al-rizq ‘livelihood,
blessing of God’. Although <ḥariziki>116 corresponds to Arabic yurzaqūna ‘they
are bestowed/blessed’ in one instance, its spelling only remotely resembles the
original. On the other hand, the second spelling <ḥarziqi>,117 without the source
word visible nearby, reproduces the etymon al-rizq (pronounced [ar-rizq]) al-
most identically. While the zāy in the third syllable for /j/ hints at the word’s
etymology, the first ḥā represents /h/, apparently inserted to avoid an initial
vowel-only syllable.118 Unlike the word háqílè examined above, the forms

||
113 TCD MS 3499 fol. 46r.
114 MAAO AF 14722(87) fol. 5v; PGL ORI 11/2 fol. 13r.
115 The change of initial /ʕ/ to /h/ is not very common. Seydina-Ousmane Diagana (1992, 210),
for instance, cites only one further example in addition to the one discussed here. Interestingly,
forms for the etymon ʕaqil with the initial /h/ are attested in a few other sub-Saharan lan-
guages, see Baldi 2008, 347–348.
116 TCD MS 2179 fol. 16v.
117 MAAO AF 14722(87) fol. 137r.
118 It is noteworthy that Creissels (2016, 24) cites /w/ and /ŋ/ as being more typically added to
‘regularise the proper syllable structure CV’. He also notes that many Arabic borrowing are
retained in Soninke with the initial /a/ unpreceded by a consonant. However, /h/ appears at
the beginning of the compound hádámárénmé ‘human being’, which is a calque from the
Arabic banī Adama ‘sons of Adam’. Interestingly, this word is mainly written etymologically
with the initial ʔalif, and only exceptionally with ḥāʾ, in the glosses.
 How to Spell Loanwords? | 293

<ḥariziki>/<ḥarziqi> seem to have survived only in manuscripts’ margins since

this item is attested as àríjaxè/wàrjàxé in modern-day Soninke.119
{51} <ɣalabi>120 xàlíbè ‘writing tool’. Next to the word <ɣalabi>, which is only
found in one instance, the scribe added the Arabic al-qalam ‘pen’. However, the
vowels of the modern form suggest a derivation from Tamasheq a-ɣanib ‘(reed)
pen’ rather than from the Arabic qalam ‘(reed) pen’. The manuscript form
<ɣalabi> may, thus, represent a hybrid between a-ɣanib and qalam, motivated
by the scribe’s knowledge of Arabic.121
{52} <ʕākin>/<ankin> from Ar. ʔax ‘brother’. The word áaxí/wáaxì ‘brother’ in
Soninke is a borrowing from Arabic, but not always recognised as such in scrib-
al spellings. In two occurrences,122 for instance, both linked to the Arabic text yā
ʔaxī ‘oh, my brother’, the glosses are as follows: <ʕā kin ū> áaxìn wó and <yā an
kinwū> ya áaxìn wó. Neither spelling resembles the etymon. Interestingly, in two
other instances,123 the scribes chose the native Soninke kinship term máarénmè
‘brother’ to translate the same Arabic phrase. Finally, in one instance, the scribe
wrote the word <axi> áaxí in the same spelling as Arabic.124
Another group of examples written in deviating orthographies are probably
borrowings from Arabic that underwent successive phonetic changes in trans-
mission from one linguistic community to another. Their spellings vary from
etymological to ‘native’ in the glosses:
{53} <šuɣula> ~ <suqula>/<sūla>125 súxùlà ‘preoccupy, trouble’ from Ar.
šaɣala/šuɣl ‘occupy, concern’.

<naqasi> ~ <nakasi>/<nāsi>126 nàqásì ‘decrease’ from Ar. naqaṣa ‘decrease’.

<waqati> ~ <wakati>/<wati>127 wáxátì ‘time’ from Ar. waqt ‘time, moment’.
<lawḥi> ~ <walaḥa>/<wulā>128 wáláhà/wáláxà ‘wooden board’ from Ar. (al-)lawḥ ‘board, tablet’.

||
119 However, Maninka has the form hàrijɛɛ. See Vydrine 1999, 249.
120 TCD MS 2179 fol. 11r.
121 We are grateful to Lameen Souag for his interpretation of the manuscript form as a hybrid and
for elucidating the borrowing path of the Soninke word. He further added that ‘Despite appearanc-
es, this [xàlíbè] is at best very distantly cognate with Arabic [qalam]: a-ɣanib is a Tuareg-specific
variant of pan-Berber a-ɣanim “reed”, a loanword from Phoenican *qān-īm’ (Lameen Souag’s
comments on an earlier draft of this paper).
122 EAP 1042/9/2 p. 24 and BULAC MS.ARA.219bis fol. 5v.
123 BULAC MS.ARA.219bis fol. 10r; UPenn Lewis O35 fol. 34r.
124 UPenn Lewis O35 fol. 86v.
125 AAN1 p. 179; PGL ORI 11/3 fol. 102v; BnF Arabe 5675 fol. 77v and fol. 80r; TCD MS 2179 fol. 9r.
126 TCD MS 2179 fol. 16r; UbL Or. 14.052(8) fol. 58r and fol. 59r; BL Or. 6473 fol. 210v and fol. 212r;
TCD MS 3500 fol. 21v.
127 UbL Or. 14.052(8) fol. 63r; BmT 2234 p. 104; ZOC1 di 5302.
294 | Dmitry Bondarev, Darya Ogorodnikova

This spelling variation reflects phonetic changes in (1) consonants: ɣ, ḥ → x or

k/g; (2) syllabic structures: metathesis l-w-ḥ → w-l-x; C → CV; and (3) dropping of
the intervocalic uvular/velar. While the adaptation (1) and (2) probably hap-
pened primarily due to borrowing into Soninke, (3) is a clear secondary adapta-
tion of Soninke into Mandinka, where the velar between two vowels is rare.
The next group of examples is concerned with different sources, channels
or chronological layers of borrowing.
{54} làgàré ‘last, youngest’. This word is regularly linked in the glosses to
the Arabic al-ʔāxīr ‘last, extreme’, as well as other derivatives from the same
root, such as muʾaxxar ‘rear part, end’. Although it is possible that the Arabic
word is the source for Soninke,129 the etymological connection is not evident
from spellings: <lākarin> and <laʾri>,130 <lāri>131, or <laqari>.132 The graphemes kāf
and qāf may stand either for /g/, /k/ or /x/, as in /lagare/, /lakare/ or /laxare/,
whereas the spellings with ʕayn and ʔalif most likely indicate the dropping of
the intervocalic velar/uvular, as in /laare/. Thus, both the graphic and phonetic
realisations of this borrowing are relatively distant from its source. Interesting-
ly, the word láaxàrá ‘afterlife, hereafter’, which is a phonetic adaptation of the
Arabic al-ʔāxirah, in writing is typically spelled etymologically (i.e. retaining the
definite article and the consonant xāʾ). Therefore, the different channel and/or
time of borrowing might be assumed for these two words deriving from the
same root.
{55} <ḥāli>/<ḥāri>133 hárì ‘even, until, so that’. This Soninke function word
corresponds systematically to its potential source, the Arabic preposi-
tion/particle ḥattā ‘until, even, so that’. The initial consonant ḥāʾ suggests a
borrowing.134 The consonant of the second syllable is represented either by lām
or rāʾ, and never by tāʾ of the assumed etymon. The transformation from /t/ in
Arabic to /l/ or /r/ in Soninke is unusual and not attested in other examples.
Hence, the grapheme ḥāʾ seems inspired by its Arabic counterpart rather than

||
128 TCD MS 3499 fol. 167v and MAAO AF 14722(87) fol. 196v and fol.197v, respectively.
129 Vydrine (1999, 271) suggests the Arabic al-ʔāxīr ‘the last’ as the source for the cognate
Maninka word làgare ‘youngest child’.
130 TCD MS 2179 fol. 31r and fol. 32r.
131 BL Or. 6473 fol. 222v.
132 BnF Arabe 5675 fol. 69v; UbL Or. 14.522(8) fol. 18r and fol. 26r.
133 PGL ORI 11/1 fol. 15r; TCD MS 3499 fol. 164v.
134 The initial consonant in the word hárì is not explained by the dialectal variation f~h, but
rather that this word is a borrowing; see Ousmane Moussa Diagana 2013, 77–78.
 How to Spell Loanwords? | 295

dictated by an etymological connection. Instead, Tamasheq can be suggested as

a source for the Soninke particle hárì.135
{56} mìsídè ‘mosque’. The Arabic word al-masjid ‘mosque’ in the glosses is
explained in two ways: <mišijidi>136 and <misdi>137. While the first spelling clear-
ly mimics the Arabic, the second, with the omitted <j>, reflects the Soninke
modern-day mìsídè. However, the omission of /j/ is unlikely to be a result of
phonetic change.138 The form <misdi> and modern-day Soninke mìsídè ‘mosque’
could instead be derived from Ḥassānīya Arabic msīd ‘mosque’.139
{57} <šafāri> sáfàaré ‘medication’ used in translating the Arabic dawāʾ or
ṭibb ‘medicine’. In a few instances, the scribes encoded the first consonant with
šīn, a grapheme not frequently used in indigenous words, which led to previous
assumptions that sáfàaré derived from Arabic šafā ‘cure’.140 However, this word
was most certainly borrowed from another source. The word safari ‘cure, treat’
exists in Songhay, where it was borrowed from Tamasheq.141 In this case, the
Soninke Ajami spellings of sáfàaré with šīn is either another (untypical) way of
encoding /s/, an awareness of the scribes that the word is ultimately a borrow-
ing, or (as suggested by Lameen Souag) a mistaken assumption by the scribes
that it derives from šifāʾ.

5.2.4 Definite article retained

Some borrowed Arabic nouns retain the definite article, integrated into the lex-
eme without any grammatical function. This section presents examples of such
words, looking at the relationship between graphemic representation and pho-
netic realisation in Arabic and Soninke items.

||
135 Vydrine (1999, 252) suggests Tamasheq as a source for the Maninka hálì ‘even’. Whether
the Tamasheq har has derived from the Arabic ḥattā is inconclusive. Heath (2006, 215), for
instance, indicates a potential relation of the Tamasheq hàr ‘until’ to the Arabic ḥattā.
Kossmann (2005, 137–138), on the other hand, considers such a derivation unlikely.
136 Ubl Or. 14.522(8) fol. 60v.
137 BnF Arabe 5675 fol. 78r.
138 See Seydina-Ousmane Diagana 1992, 206.
139 See Wexler 1980, 540–541. Lameen Souag also points out that ‘msīd is a pan-Maghrebi
form attested at least from Tunisia and Morocco, sometimes meaning “Quranic school”’
(Souag’s comments on an earlier draft of the paper).
140 Ogorodnikova 2023, 179. Note that Ousmane Moussa Diagana (2013, 180) considers the
word sáfàaré to be an Arabic borrowing.
141 See Heath 1998, 212; Baldi 2005; Souag 2016.
296 | Dmitry Bondarev, Darya Ogorodnikova

The article is kept intact in some words, and it is represented in writing the
same way as in Arabic by ʔalif-lām with sukūn, even though some modern forms
feature a vowel added after l: aljannà <aljana>142 from Ar. al-jannah ‘Paradise’,
alhaalà <alḥāla>143 from Ar. al-ḥāl ‘condition, state’, and álíxíyáamà <alqi-
yāma>144 from Ar. al-qiyāmah ‘resurrection, final judgment’.
The definite article is typically retained in words borrowed from the words
with the initial ʔ and ʕ. The borrowed forms in modern Soninke begin with the
consonant l-: liimànàaxù from Ar. al-ʔīmān ‘faith’; lásílì from Ar. al-ʔaṣl ‘origin’;
láaxàrá from Ar. al-ʔāxirah ‘the hereafter’; and láadà from Ar. al-ʕādah ‘cus-
tom’. The initial ʔalif of the article may still be written in the glosses, but such
spelling is infrequent, and the words are mostly written with the initial lām,
representing phonologically adapted forms.145 Peculiarly, on one occasion, the
/l/ of the definite article in the word l-Àráabù ‘Arabs’ <ḍaʕarābu>146 was encoded
with the Arabic ḍād <ḍ> (which is sometimes used for /l/).
The phonetic assimilation is ignored in a few words beginning with the ‘sun
letters’, adhering to the conventional writing of the article with lām: <al-
dālīlin>147 from Ar. al-dalīl ‘sign, proof’, <alðāti>148 from Ar. al-ðāt ‘essence’,
<[a]lṭālibinu>149 from Ar. al-ṭālib ‘student’, <alšayx>150 from Ar. al-šayx ‘shaykh,
elder’ or <al-tawḥīdi>151 from Ar. al-tawḥīd ‘asserting oneness (of God)’. Never-
theless, the forms of these same words without the article occur much more
frequently in the glosses: for example, dàlíilè ‘proof’, jàatí ‘divine essence’ or
táalíbènù ‘students’.
There are two exceptions to the previous habit – both beginning with nūn
(which is the ‘sun letter’) – where the article is represented in its ‘assimilated’
rather than ‘graphical’ form. Thus, the most common spelling for the borrowing
ànnábìnɲínmè ‘prophet’ (from Ar. al-nabīy pronounced as [an-nabī]) in the Aja-
mi glosses is with the initial ʔalif-nūn <anabi yūmi> (the reduplication of the
first consonant is underrepresented),152 or sometimes with ʔalif-nūn-nūn <anna-

||
142 BULAC MS.ARA.219bis fol. 10v; PGL ORI 11/3 fol. 102r.
143 BnF Arabe 5657 fol. 53r; PGL ORI 11/1 and ORI 11/3 fol. 19r.
144 BnF Arabe 5675 fol. 67r; EAP 1042/9/6 p. 52.
145 TCD MS 3499 fol. 62v.
146 BULAC MS.ARA.219bis di 1233.
147 BL Or. 6473 fol. 205v; EAP 1042/9/3 p. 13; MAAO AF 14722(87) fol. 52v.
148 EAP 1042/9/3 p. 20.
149 EAP 1042/9/8 p. 18; PGL ORI 11/3 fol. 10r.
150 PGL ORI 11/3 fol. 4r.
151 BL Or. 6473 fol. 222v; PGL ORI 11/3 fol. 20r and fol. 25r.
152 DNN2 di 0001; MAAO AF 14722(87) fol. 2r, fol. 190v, fol. 198r; TCD MS 3499 fol. 45r.
 How to Spell Loanwords? | 297

biyunmi>.153 However, even the first ʔalif of the article can sometimes be ‘re-
placed’ with an ʕayn as a support for the initial vowel, with spellings such as
<ʕanabiy>154 ànnábì ‘prophet’ or <ʕaniya>155 ànníyà ‘intention, desire’ (<Ar al-
niyyah pronounced as [an-niyyatu]).

6 Towards a typology of spelling loanwords in

annotated manuscripts
The OKb and Soninke data presented in Section 5 have both distinct and similar
characteristics of loanwords in terms of their spelling and relationship with the
original Arabic orthography. The main difference is in the parameter of the vis-
ual proximity between the Ajami word and the corresponding Arabic word. The
proximity of the Arabic original in OKb seems to play a more significant role
than in Soninke. It is possible that a conservative environment imposed on OKb
by the Qur’an text created tighter connections between Ajami and Arabic. Such
connections are best seen in OKb presented in Subsections 5.1.2 and 5.1.3. At the
same time, visual proximity is a good diagnostic for the integration of a loan-
word when it does not exhibit etymological orthography while written close to
the source word, as discussed in Subsections 5.1.4 (OKb) and 5.2.3 (Soninke).
Both OKb and Soninke data are similar in how they show some predictable
tendencies along the scale from etymological to non-etymological spelling.
Thus, the closest (identical) orthography is displayed by the etymons in the
semantic fields of Islamic terminology – a tendency noticed in previous litera-
ture.156 At the same time, both OKb and Soninke Ajami show that some promi-
nent religious concepts and terms related to literacy (OKb items 17–19; Soninke
items 47, 48, 50, 51) are graphemically very distant from the source words.
The orthography divergent from the original is often applied to well-
integrated words, usually with a longer history of phonological adaptation. But,
at the same time, the undoubtedly oldest loans that had undergone significant
morphological adjustments (e.g. OKb items 4, 13) are written with clear refer-
ences to the original spelling.

||
153 DNN2 di 0028 and 0029.
154 MAAO AF 14722(87) fol. 190v, fol. 208v.
155 BnF Arabe 5675 fol. 74r; BULAC MS.ARA.219bis di 1247.
156 Greenberg 1960; Wexler 1980, 539–540; van den Boogert 1997, 223–229.
298 | Dmitry Bondarev, Darya Ogorodnikova

Such spelling retentions and divergences are probably motivated by estab-

lished orthographic practices specific to a teaching-learning circle, whereas the
individual sensitivity of scribes does not seem to play a prominent role in the
selection of spelling features. The orthography of some loanwords (e.g. OKb
items 9, 11, 15; Soninke items 36, 43, 44, 49) display a high degree of consisten-
cy, even when written by different scribes. Such standardisation tendencies
suggest that the writing of these items predated the manuscripts in which they
are found and that the words in question had long ago been integrated into the
target language.
Remarkably, the frequency of loanwords used in manuscripts does not
seem to be a spelling factor, which is in stark contrast to frequency-conditioned
standardisation of language-internal words.157
The spelling of loanwords borrowed with the Arabic definite article al-
sometimes indicates a kind of graphemic hypercorrection. Thus, an otherwise
integrated word with the initial vowel clipped, and, therefore, retaining only the
consonant /l/, may be written with the whole article restored to the original
(OKb Subsection 5.1.6.1; Soninke Subsection 5.2.4). At the same time, a reverse
process has been observed, whereby a written word is spelled with the clipped
article but the spoken counterpart attested in modern language retains the
whole article (e.g. OKb item 29).
We have tried to use the terminology introduced in Table 1 in our discussion
(in Section 5) of spelling types based on Table 2. The analysis presented allows
us to clarify some of the terms. The major difference is between identical and
divergent spelling, which is trivial. However, the term copy-spelling should
probably be reserved to denote conscious copying of what the scribe sees in the
original Arabic word. However, given our assumption that individual groups of
teachers-learners constitute a stronger factor of orthographic conventions, we
should acknowledge that copy-spelling, in the strict sense, is rather difficult to
identify.
In light of the analysis presented in Section 5, we should also clarify the
term insertion suggested by Kossmann.158 It is useful to distinguish between the
narrower and wider meaning of the term. In a narrow sense, the insertion is a
word strictly used to denote an Arabic concept not available in the target lan-
guage. In a wider sense, however, the insertion refers to the use of a word in the
exegetical register associated with practices of translation. The exegetical inser-
tion has then a higher potential to be passed onto the wider circles of speakers

||
157 Bondarev and Dobronravin 2019.
158 Kossmann 2013, 47; cited in Section 4.1.
 How to Spell Loanwords? | 299

through register-induced contacts. Due to this potential path of transmission

from insertion to integrated borrowing, it is often hard to establish whether we
are dealing with the (exegetical) insertion or the etymological spelling of an
already integrated word. This is because of a lack of evidence of contemporane-
ous use (synchronic with our data) of the given word in the spoken language of
the time when manuscripts were produced.
In summary, types of spelling presented in Table 2 and exemplified in Sec-
tion 5 are difficult to subdivide into neat categories. However, our preliminary
typology opens up a novel perspective on Arabic loans and we hope that it may
also be a useful heuristic tool to look into the patterns of borrowing in addition
to previous research based on common historical linguistic principles. These
types can be used further in future research, involving more data, to investigate
the characteristics of borrowing outlined in Table 3.

7 Conclusion
Our paper has examined different stages of the integration of Arabic loan-
words into West African languages based on examples of spelling patterns
identified in two distinct manuscript traditions of OKb and Soninke. Both
languages, written in Arabic script (Ajami), were used for interlinear transla-
tions of Arabic texts. These interlinear Ajami writings exhibit a high degree of
Arabic loanwords, some spelled etymologically, some deviating from the
Arabic graphemic source to the point of complete dissimilarity. The question
‘To what extent such spelling variation reflects the integration of Arabic
loanwords into the target language?’ was a primary motivation for our study.
We started with an analysis of the spelling of Arabic loans in OKb and Sonin-
ke annotations by using the principle of gradient etymological spelling, in-
troduced in Subsection 4.2. Different spelling types were, thus, examined
following a scalar from identical to partial etymological (further subdivided
into four subtypes) to dissimilar spelling. Based on this comparison of
spelling patterns, we have sketched out some tendences for a preliminary
typology of spelling of Arabic loanwords in interlinear Ajami writings (Sec-
tion 6). However, our findings are far from being conclusive. Both OKb and
Soninke data demonstrate that spelling alone is a weak diagnostic for the
degree of integration of loanwords, but together with other linguistic and
cultural factors, loanwords in Ajami open up an additional perspective on
the chronology and paths of borrowing.
300 | Dmitry Bondarev, Darya Ogorodnikova

The Ajami data should, nevertheless, be taken with caution. This is be-
cause the learned environment in which Ajami writings were produced has
its own lexical base, often independent from the lexical history of the spoken
languages. Although the Ajami texts in our corpus are translations bound to
the source texts, with a high potential of Ajami–Arabic interplay, ortho-
graphic choices seem to be largely dictated by the conventions established
within the individual learned circles, which transmit such conventions
across time and manuscripts. Such circles are usually clustered around cer-
tain scholars,159 who could be responsible for an orthographic consistency
different from the other learned circles of the same speech community. The
multiple and parallel channels of this kind of individualised transmission
must have amplified the divergent history of Arabic loanwords – the process
that has so far been unrecognised as a factor behind multiple forms of loan-
words in West African languages.

Acknowledgements

This paper is the result of joint work by the two authors. However, for academic
purposes, Dmitry Bondarev is responsible for Sections 1–3, 4.2 and 5.1; Darya
Ogorodnikova for Section 5.2; Sections 4.1, 6 and 7 were co-written.
We are grateful to Lameen Souag for reading an earlier draft of the paper
and for his valuable comments, most of which have been incorporated into the
paper. Our thanks are also due to Valentin Vydrine for confirming Lameen
Souag’s doubts about the etymology of the Soninke lexeme bátù ‘to worship’
which was subsequently removed from the paper. All errors are our own re-
sponsibility.
The research for this article was funded by the Deutsche Forschungsge-
meinschaft (DFG, German Research Foundation) under Germany’s Excellence
Strategy – EXC 2176 ‘Understanding Written Artefacts: Material, Interaction and
Transmission in Manuscript Cultures’, project no.390893796. The research was
conducted within the scope of the Centre for the Study of Manuscript Cultures
(CSMC) at Universität Hamburg.

||
159 See Ogorodnikova 2023 on individual distinct learning groups.
 How to Spell Loanwords? | 301

Abbreviations
1, 2, 3 = 1st, 2nd, 3rd person INACP = incomplete aspect
1s, 2s, 3s = 1st person singular, etc. IPFV = imperfective
1p, 2p, 3p = 1st person plural, etc. IMV = imperative
AAN = Aliou Ndiaye, Adéane, manuscript JRL = John Rylands Library
ABSTR = abstract MAAO = Musée national des Arts d’Afrique
ADV = adverbial (operator) et d’Océanie
AFP = argument focus perfective NEG = negative
ANTP = antipassive NMAG = agent noun
Arabe 402 = the manuscript in the Biblio- NMLZ = nominalization
thèque nationale de France (N334 OKb = Old Kanembu
Arabe 402) PL = plural
AUX = auxiliary PFV = perfective
BL = British Library PGL = Palace Green Library
BmT = Bibliothèque municipale de Tours 2ShK = the ‘Shetima Kawo’ manuscript
BnF = Bibliothèque nationale de France SBD = subordination marker
BULAC = Bibliothèque Universitaire des SG = singular
Langues et Civilisations SJ = subject marker
COPEQ = equative copula SUBJ = subjunctive
DEM = demonstrative TCD = Trinity College, Dublin
DNN = N.N., Dembancane, manuscript T.Kano = the ‘Tahir Kano’ manuscript
DO = direct object 1YM = the ‘Yerima Mustafa’ manuscript
EAP = Endangered Archives Programme, Q.2:72 etc. = chapter and verse of the
the British Library Qur’an
FOC = focus marker UbL = Universitätsbibliothek Leiden
FUT = future tense UPenn = University of Pennsylvania
GEN = genitive ZOC = O.C., Ziguinchor, manuscript
GER = gerundive

References
Badawi, Elsaid M. and Muhammad Abdel Haleem (2008), Arabic–English Dictionary of Qur’anic
Usage, Leiden: Brill.
Baldi, Sergio (1999), Ancient and New Arabic Loans in Chadic (University of Leipzig Papers on
Africa, Languages and Literature, 7), Leipzig: Institut für Afrikanistik, Universität Leipzig.
Baldi, Sergio (2002), ‘On Arabic Loans in Kanuri’, Studi Magrebini, 25: 45–57.
Baldi, Sergio (2003), ‘Some Statistical Analysis on Arabic Loans in Kanuri’, in Robert Nicolai
and Peter Zima (eds), Lexical and Structural Diffusion: Interplay of Internal and External
Factors of Language Development in the West African Sahel, Nice: Sophia Antipolis, 9–15.
Baldi, Sergio (2005), ‘Les emprunts berbères, surtout d’origine arabe, dans les langues de
l’Afrique Occidentale’, Studi Magrebini, n.s., 3: 103–118 <https://unora.unior.it/retrieve/
handle/11574/35563/1636/10Baldic.pdf> (accessed on 8 November 2023).
302 | Dmitry Bondarev, Darya Ogorodnikova

Baldi, Sergio (2007), ‘Kanuri’, in Kees Versteegh et al. (eds), Encyclopedia of Arabic Language
and Linguistics, vol. 2, Leiden: Brill, 552–555.
Baldi, Sergio (2008), Dictionnaire des emprunts arabes dans les langues de l’Afrique de
l’Ouest et en swahili, Paris: Karthala.
Baldi, Sergio (2020), ‘Arabic Influence in West Africa: An Overview’, Folia Orientalia, 57: 11–23.
Bathily, Mody (2017), Le grand dictionnaire: soninké–français, français–soninké, s.l.: s.n.
Bondarev, Dmitry (2013a), ‘Language Change Induced by Written Codes: A Case of Old Kanembu
and Kanuri Dialects’, in Esther-Miriam Wagner, Ben Outhwaite and Bettina Beinhoff (eds),
Scribes as Agents of Language Change (Studies in Language Change, 10), Boston: De Gruyter
Mouton, 291–323.
Bondarev, Dmitry (2013b), ‘Qur’anic Exegesis in Old Kanembu: Linguistic Precision for Better
Interpretation’, in Tal Tamari and Dmitry Bondarev (eds), Qur’anic Exegesis in African Lan-
guages = Journal of Qur’anic Studies, 15/3: 56–83.
Bondarev, Dmitry (2014a), ‘Multiglossia in West African Manuscripts: A Case of Borno, Nigera’,
in Jörg B. Quenzer, Dmitry Bondarev and Jan-Ulrich Sobisch (eds), Manuscript Cultures:
Mapping the Field (Studies in Manuscript Cultures, 1), Berlin: De Gruyter, 113–155.
Bondarev, Dmitry (2014b), ‘Old Kanembu and Kanuri in Arabic Script: Phonology through the
Graphic System’, in Mumin and Versteegh 2014, 107–142.
Bondarev, Dmitry (2017), ‘Islamic Education and Ample Space Layout in West African Manu-
scripts’, in Brigaglia and Nobili 2017, 105–142.
Bondarev, Dmitry (2021), ‘A Typology of West African Ajami Manuscripts: Languages, Layout
and Research Perspectives’, in Jörg B. Quenzer (ed.), Exploring Written Artefacts: Objects,
Methods, and Concepts (Studies in Manuscript Cultures, 25), Berlin: De Gruyter, 707–728.
Bondarev, Dmitry (2022), ‘Grammar and Art of Translation as Expression of Muslim Faith:
Translational Practices in West Africa’, in Hephzibah Israel (ed.), Routledge Handbook on
Translation and Religion, London: Routledge, 415–431.
Bondarev, Dmitry (2024), ‘Shifting Regimes, Reshaping Manuscripts: Disappearance of Anno-
tated Qurʾans in Borno’, in Christiane Gruber and Bihter Esener (eds), Regime Change:
New Horizons in Islamic Art and Visual Culture (Gingko Library Art Series), London: Gingko,
32–46.
Bondarev, Dmitry (forthcoming), Old Kanembu: A Saharan language of the Qur’anic glosses.
Bondarev, Dmitry and Nikolay Dobronravin (2019), ‘Standardisation Tendencies in Kanuri and
Hausa Ajami Writings’, in Dmitry Bondarev, Alessandro Gori and Lameen Souag (eds),
Creating Standards: Interactions with Arabic Script in 12 Manuscript Cultures (Studies in
Manuscript Cultures, 16), Berlin: De Gruyter, 237–269.
Bondarev, Dmitry and Abba Tijani (2014), ‘Performance of Multilayered Literacy: Tarjumo of the
Kanuri Muslim Scholars’, in Kasper Juffermans, Yonas Mesfun Asfaha and Ashraf K. Ab-
delhay (eds), African Literacies: Ideologies, Scripts, Education, Newcastle-upon-Tyne:
Cambridge Scholars Publishing, 118–146.
Brigaglia, Andrea and Mauro Nobili (eds) (2017), The Arts and Crafts of Literacy: Manuscript
Cultures in Muslim Sub-Saharan Africa (Studies in Manuscript Cultures, 12), Berlin: De
Gruyter.
Chonai, Hasan (1999), Gruppa teda-daza (tsentral’nosaharskaya semya yazikov) i yeyo genet-
ichesckiye vzaimodeystviya (etimologichesckiy i fonologichesckiy aspekt) Группа теда-
даза (центральносахарская семья языков) и ее генетические взаимодействия
(этимологический и фонологический аспект) (‘Genealogical Relationships in Teda-Daza
 How to Spell Loanwords? | 303

Language Group (Central Saharan Linguistic Family): Phonology and Etymology’), un-
published PhD thesis, Russian State Humanitarian University.
Creissels, Denis (2016), ‘Phonologie segmentale et tonale du soninké (parler du Kingi)’, Man-
denkan, 55: 3–174.
Cultru, Prosper (1913), Premier voyage du Sieur de La Courbe fait à la Coste d’Afrique en 1685,
Publié avec une carte de Delisle et une introduction, Paris: Edouard Champion / Emile Larose.
Cuoq, Joseph (1984), Histoire de l’Islamisation de l’Afrique de l’Ouest des origines à la fin du
XVIe siècle, Paris: Geuthner.
Dantioko, Oudiary Makan (2003), Dictionnaire soninké–français, Bamako: Jamana.
De Moraes Farias, Paulo F. (2003), Arabic Medieval Inscriptions from the Republic of Mali:
Epigraphy, Chronicles, and Songhay-Tuareg History (Fontes Historiae Africanae, New Se-
ries, 4), Oxford: Oxford University Press / New York: The British Academy.
Diagana, Ousmane Moussa (2013), Dictionnaire soninké–français (Mauritanie) Paris: Karthala.
Diagana, Ousmane Moussa (1995), La langue soninké: morphosyntaxe et sens à travers le
parler de Kaédi (Mauritanie), Paris: Harmattan.
Diagana, Seydina-Ousmane (1992), Contact de langues: Approche sociolinguistique des em-
prunts du soninké au français, à l’arabe et au pulaar, PhD thesis, Université de Paris V.
Dimmendaal, Gerrit J. (2011), Historical Linguistics and the Comparative Study of African Lan-
guages, Amsterdam: John Benjamins.
Dramé, Djibril (2021), ‘Le soninké dans les manuscrits islamiques: Corrélation entre les corres-
pondances graphématiques et phonémiques’, Mandenkan, 66: 187– 203.
Dramé, Djibril (2022), Mapping Dialectal Variation in Soninke Islamic Manuscripts, PhD thesis,
University of Hamburg.
Greenberg, Joseph H. (1960), ‘Linguistic Evidence for the Influence of the Kanuri on the Hausa’,
The Journal of African History, 1/2: 205–212.
Gutelius, David (2000), ‘Newly Discovered 10th/16th Century Ajami Manuscript in Niger and
Kel Tamasheq History’, Saharan Studies Association Newsletter, 8/1–2: 6.
Heath, Jeffrey (1998), Dictionnaire songhay–anglais–français, vol. 1: Koyra Chiini, Harmattan:
Paris.
Heath, Jeffrey (2006), Dictionnaire touareg du Mali: tamachek–anglais–français, Paris: Khartala.
Hiskett, Mervyn (1984), The Development of Islam in West Africa, London: Longman.
Hutchison, John P. (1981), The Kanuri Language: A Reference Grammar, Madison: University of
Wisconsin.
Insoll, Timothy (2003), The Archaeology of Islam in Sub-Saharan Africa, Cambridge: Cambridge
University Press.
Kossmann, Maarten (2005), Berber Loanwords in Hausa (Berber Studies, 12), Cologne: Rüdiger
Köppe.
Kosmann, Maarten (2013), The Arabic Influence on Northern Berber (Studies in Semitic Lan-
guages and Linguistics, 67), Leiden: Brill.
Levtzion, Nehemia and Randall L. Pouwels (eds) (2000), The History of Islam in Africa, Athens,
OH: Ohio University Press / Oxford: James Currey / Cape Town: David Philip.
Lukas, Johannes (1957), Die Sprache der Tubu in der Zentralen Sahara (Deutsche Akademie der
Wissenschaften zu Berlin, Institut für Orientforschung, 14), Berlin: Akademie-Verlag.
McLaughlin, Fiona (forthcoming), ‘Ajami Writing Practices in Atlantic-Speaking Africa’, in Fried-
erike Lupke (ed.), The Oxford Guide to the Atlantic Languages (Oxford Guides to the
World’s Languages), Oxford: Oxford University Press; for an online prepublication version,
304 | Dmitry Bondarev, Darya Ogorodnikova

see <https://people.clas.ufl.edu/fmcl/files/AjamiCIRCRED.pdf> (accessed on 20 Septem-

ber 2023).
Maschler, Yael (1994), ‘Metalanguaging and Discourse Markers in Bilingual Conversation’,
Language in Society, 23/3: 325–366 <doi:10.1017/S0047404500018017>.
Mumin, Meikal and Kees Versteegh (eds) (2014), The Arabic Script in Africa: Studies in the Use
of a Writing System (Studies in Semitic Languages and Linguistics, 71), Leiden: Brill.
Newman, Paul (2000), The Hausa Language: An Encyclopedic Reference Grammar (Yale Lan-
guage Series), New Haven: Yale University Press.
Ngom, Fallou (2017), ‘West African Manuscripts in Arabic and African Languages and Digital
Preservation’, in Thomas Spear (ed.), Oxford Research Encyclopedia of African History,
<https://doi.org/10.1093/acrefore/9780190277734.013.123>.
Nixon, Sam (ed.) (2017), Essouk-Tadmekka: An Early Islamic Trans-Saharan Market Town (Jour-
nal of African Archaeology Monograph Series, 12), Leiden/Boston: Brill.
Ogorodnikova, Darya (2017), ‘ʿAjamī Annotations in Multilingual Manuscripts from Mande
Speaking Areas: Visual and Linguistic Features’, Islamic Africa, 8: 111–143.
Ogorodnikova, Darya (2021), ‘“I Heard It from My Teacher”: Reflections on Transmission of
Knowledge in Islamic Manuscripts from Senegambia and Mali’, in Stefanie Brinkmann,
Giovanni Ciotti, Martin Delhey and Stefano Valente (eds), Education Materialized: Recon-
structing Teaching and Learning Contexts through Manuscripts (Studies in Manuscript
Cultures, 23), Berlin: De Gruyter, 127–150.
Ogorodnikova, Darya (2023), Soninke Ajami Manuscripts and Islamic Education: A Reconstruc-
tion of Scribal Practices, PhD thesis, Universität Hamburg.
Salvaing, Bernard (2020), ‘Islam in Sub-Saharan Africa’, in Fallou Ngom, Mustapha H. Kurfi and
Toyin Falola (eds), The Palgrave Handbook of Islam in Africa, Cham: Springer International
Publishing.
Salvaing, Bernard and John Hunwick (2003), ‘Writers of Guinea’, in John Hunwick (ed.), Arabic
Literature of Africa, vol. 4: The Writings of Western Sudanic Africa (Handbuch der Oriental-
istik, 4), Leiden: Brill, 491–529.
Souag, Lameen (2010), ‘Ajami in West Africa’, Afrikanistik Aegyptologie Online,
<http://www.afrikanistik-aegyptologie-online.de/archiv/2010/2957> (accessed on 22 Au-
gust 2023).
Souag, Lameen (2015), ‘Archaic and Innovative Islamic Prayer Names around the Sahara’,
Bulletin of the School of Oriental and African Studies, 78/2: 357–374.
Souag, Lameen (2016), ‘Language Contact in the Sahara’, in Thomas Spear (ed.), Oxford Research
Encyclopedia of African History, <https://doi.org/10.1093/acrefore/9780199384655.013.141>.
Tamari, Tal (2016), ‘Styles in Islamic Education: Perspectives from Mali, Guinea, and The Gam-
bia’, in Robert Launay (ed.), Islamic Education in Africa: Writing Boards and Blackboards,
Bloomington / Indianapolis: Indiana Univ. Press, 29–60.
Tamari, Tal (2017), ‘Bamana Texts in Arabic Characters: Some Leaves from Mali’, in Brigaglia
and Nobili 2017, 207–278.
Tamari, Tal and Dmitry Bondarev (2013), ‘Introduction and Annotated Bibliography’, in Tal
Tamari and Dmitry Bondarev (eds), Qur’anic Exegesis in African Languages = Journal of
Qur’anic Studies, 15/3: 1–55.
Van den Boogert, Nico (1997), The Berber Literary Tradition of the Sous, with an Edition and
Translation of ‘The Ocean of Tears’ by Muhammad Awzal (d. 1749), Leiden: Nederlands In-
stituut voor het Nabije Oosten.
 How to Spell Loanwords? | 305

Vydrine, Valentin (1998), ‘Sur l’écriture Mandingue et Mandé en caractères arabes (mandinka,
bambara, soussou, mogofin)’, Mandenkan, 33: 1–87.
Vydrine, Valentin (1999), Manding–English Dictionary: Maninka, Bamana, vol. 1, Saint Peters-
burg: Dimitry Bulanin.
Vydrin, Valentin (2014), ‘Ajami Scripts for Mande Languages’, in Mumin and Versteegh 2014,
199–224.
Wexler, Paul (1980), ‘Problems in Monitoring the Diffusion of Arabic into West and Central African
Languages’, Zeitschrift der Deutschen Morgenlandischen Gesellschaft, 130: 522–556.
Zappa, Francesco (2011), ‘When Arabic Resonates in the Words of an African Language: Some
Morphological and Semantic Features of Arabic Loanwords and Calques in Bambara’, in
Guiliano Lancioni and Lidia Bettini (eds), The Word in Arabic (Studies in Semitic Lan-
guages and Linguistics, 62), Leiden: Brill, 229–249.
|
Propagandistic Multilingualism
Gábor Zólyomi
Mesopotamian Bilingual Royal Inscriptions
from the Third Millennium BCE: Texts with a
Primary and Secondary Context
Abstract: This paper discusses inscriptions transmitted in both Sumerian and
Akkadian versions, originally prepared for the Akkadian-speaking rulers of the
dynasty of Sargon in the twenty-third century BCE. They are known from manu-
scripts made in the first part of the second millennium BCE. The first part of the
paper discusses features of the manuscripts that originate in the cuneiform
manuscript culture of the Old Babylonian period, the time when their manu-
scripts were prepared. The second part examines features deriving from the time
when they were at first composed. Its main conclusion is that these bilingual
inscriptions were first composed in Akkadian. The Sumerian versions may have
been prepared depending on the Akkadian ones by bilingual people who were
familiar with the writing conventions of southern Babylonia.

1 Introduction
This paper discusses Old Akkadian royal inscriptions transmitted both in Sume-
rian and Akkadian versions. These texts are known from manuscripts written
with cuneiform script on clay tablets in the Old Babylonian period, i.e. in the
first part of the second millennium BCE. The manuscripts are scholarly editions
recording not only the inscriptions but also other features of the original objects
that carried them. The inscriptions, therefore, have a double character: they
have both a primary and secondary context. The primary context is the period of
the dynasty of Agade, around the twenty-third century BCE, when the original
inscriptions were composed and carved on monumental and votive objects,
while the secondary is the Old Babylonian period, when they were copied from
the original objects and collected with a number of exclusively Akkadian in-
scriptions into some kind of ancient anthologies.
Both Sumerian and Akkadian were spoken as vernaculars by their native
speakers in the third millennium BCE, and we may safely assume that a number
of individuals existed who were capable of communicating in both languages in
both written and spoken form. By the Old Babylonian period, on the other hand,
Sumerian had ceased to be a vernacular, its users were native speakers of Akka-

Open Access. © 2024 the author, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-011
310 | Gábor Zólyomi

dian, who did not learn Sumerian from their mothers but in some kind of insti-
tutional context, which we may call a scribal school.1
Accordingly, this paper is in two parts. In the first part, I will discuss fea-
tures of the inscriptions that originate in their secondary context, the cuneiform
manuscript culture of the Old Babylonian period.2 In the second part, I will ex-
amine features that derive from their primary context, the bilingual environ-
ment that felt it important to have texts describing the deeds of the Akkadian-
speaking rulers of Agade in both Akkadian and Sumerian.

2 The secondary context

Modern editions of these inscriptions recognise three of them. Two record the
deeds of Sargon, the founder of the dynasty of Agade in the twenty-third centu-
ry BCE, and one belongs to Rimuš, the elder son of Sargon.3 They will be referred
as Sargon 1,4 Sargon 115 and Rimuš 186 in this paper, following the edition of
Douglas Frayne (1993).
However, Walter Sommerfeld argues in a number of publications that Sar-
gon 1 and Sargon 11, in fact, constituted one continuous text written on the back
of the same monument, a statue of Sargon.7 According to him, the separation of
this inscription into Sargon 1 and 11 is due the editorial work of the scribe(s)
who prepared the collections preserved on the Old Babylonian MTMs (also see
below).
Let me start with a short characterisation of the four manuscripts that pre-
served the royal inscriptions.8 Table 1 lists below the four manuscripts:9

||
1 For a review on the relationship between Sumerian and Akkadian over the millennia, see
Crisostomo 2020; Hasselbach-Andee 2020b; Van Dijk-Coombers 2021; Cancik-Kirschbaum and
Schrakamp 2022.
2 For more Old Babylonian collections of royal inscriptions, see Sövegjártó 2023.
3 For the dynasty of Agade, see Westenholz 1999; Foster 2016.
4 RIME 2.1.1.1 (CDLI Q000834) recorded on manuscripts A and B; labelled as ‘Sargon C1’ in
Gelb and Kienast 1990.
5 RIME 2.1.1.11 (CDLI Q001403) recorded on manuscripts A and B; labelled as ‘Sargon C2’ in
Gelb and Kienast 1990.
6 RIME 2.1.2.18 (CDLI Q000842) recorded on manuscripts A, C and D; labelled as ‘Rīmuš C9’ in
Gelb and Kienast 1990.
7 Cf. Sommerfeld 2008, 234; Sommerfeld, 2021, 560, n. 142.
8 For a more detailed description of the features of manuscripts A and B based on his own
collations, cf. Sommerfeld 2012, 197–199.
 Mesopotamian Bilingual Royal Inscriptions from the Third Millennium BCE | 311

Table 1: Manuscripts containing Sargon 1, 11 and Rimuš 18.

Museum no. CDLI no. Type Texts

MS A CBS 13972 + CBS 14545 P227509 MTM, 28 cols Sargon 1, 11,
Rimuš 18
MS B Ni 3200 P227510 MTM, 20 cols Sargon 1, 11
MS C CBS 02344 (+) CBS 14547 P227513 MTM, ? cols Rimuš 18
+ N 3539
MS D AO 5477 P220619 single-text MS, Rimuš 18
2 cols

The first tablet, manuscript A, is now in the Penn Museum in Philadelphia. It is

an unbaked tablet, 25 × 22 cm, and about 2 cm thick. It has been reconstructed
from two main fragments. It is not complete, as you can see on Fig. 1 below,
small parts are missing here and there.
It originally contained twenty-eight narrow columns, fourteen on each side,
which preserved a total of twenty-four royal inscriptions. Fig. 2 shows the approx-
imate arrangement of the texts on the tablet. The bilingual Sargon texts are at the
beginning of the obverse, and the Rimuš text is at the very end of the reverse.10
The inscriptions are separated with colophons referring to parts of the ob-
ject on which the text was originally carved. The captions that identify people
on reliefs of the original monument are also recorded after some texts.
Manuscript A is an MTM, containing all bilingual inscriptions which are the
subject of this paper. The arrangement of the texts is chronological: Sargon,
Rimuš and Maništusu ruled in this order. The texts kept the archaic orthography
and even reproduced the sign forms of the originals.11
Manuscript B is a slightly smaller MTM that originally contained twenty
columns, preserving twenty-three inscriptions. It is now in the Arkeoloji
Müzeleri of Istanbul. Among our texts, it contained only the Sargon 1 and 11;
Rimuš 18 was not part of this collection.
Manuscript C is a very fragmentarily preserved MTM that originally con-
tained probably only Rimuš 18 among our inscriptions, as it seems to be a col-

||
9 For copies and/or photos of the manuscripts, visit CDLI’s search page (<https://cdli.mpiwg-
berlin.mpg.de/search>) and search for the manuscripts with the P-numbers provided in Table 1.
10 Note that the columns run from left to right on the obverse and right to left on the reverse
on cuneiform tablets.
11 Cf. Sommerfeld 2021, 556, n. 121 and 561–564.
312 | Gábor Zólyomi

lection of inscriptions of Rimuš and Naram-Sin. It is now in the Penn Museum in

Philadelphia.
Manuscript D is a single-text manuscript now in the Louvre in Paris. It con-
tains only Rimuš 18 written in two parallel columns: the Sumerian on the left,
the Akkadian version on the right.

Fig. 1: Obverse of manuscript A (CBS 13972 +); courtesy of the Penn Museum, image 296512 and
object B13972.
 Mesopotamian Bilingual Royal Inscriptions from the Third Millennium BCE | 313

Fig. 2: Approximate arrangement of the inscriptions on manuscript A (numbering after Fray-

ne 1993).

There is a colophon on the left edge of manuscript A relating to the whole col-
lection written on it:

Manuscript A, left edge i 1–iii 2

[… …], ⸢šar⸣-ru-GI, ri₂-mu-uš, ⸢ma-an-iš⸣-[tu]-šu, šag₄ e₂-kur-⸢ra⸣, a-na me-a-be₂12
‘The … … of Sargon, Rimuš and Maništusu, within the temple Ekur, as many as there are.’

It states that the collection on the tablet is a copy of all the inscriptions of the
first three rulers of the dynasty of Agade which were on display in the Ekur tem-
ple of Nippur. Ekur was the main temple of the god Enlil, the head of the Sume-
rian pantheon. Unfortunately, the text is broken where we would find the des-
ignation of the objects located in the Ekur.
No colophon relating to the collection is preserved on manuscript B. There
is another colophon relating to the whole of the collection recorded on manu-
script C, at the end of the last two columns of the reverse:

||
12 In the graphemic transliteration, subscript numerals distinguish homophonic graphemes,
and superscript graphemes are semantic classifiers; graphemes that constitute a word are
linked by hyphens.
314 | Gábor Zólyomi

Manuscript C, rev. vi’ and vii’ 46–47

murub₄ kisal-ka al-ŋar, šag₄ kisal e₂-kur-ra
‘Set up in the middle of the courtyard, within the courtyard of the Ekur.’

It confirms the information of the colophon on manuscript A: the texts copied to

these tablets are from objects originally on display in the Ekur temple, particu-
larly in the courtyard of the temple complex. They probably still stood there
during the Old Babylonian period when the copies were made, some 400 to 600
years after they had originally been set up.
Other colophons relate to individual texts of the collections. Table 2 below
shows the preserved colophons of Sargon 1 and 11:13

Table 2: The preserved colophons of Sargon 1 and 11.

Colophons
Sumerian version Akkadian version
Sargon 1 MS A, obv. iii 43–44 MS A, obv. iv 44–45
mu-sar-ra, ki-gal-ba mu-sar-⸢ra⸣, ⸢murgu₂-na⸣
‘inscription on its pedestal’ ‘inscription of his back/shoulder’
Sargon 11 MS B, obv. vii 11–12 MS A, obv. vi 48–49
mu-sar-ra, ki-gal-ba ⸢mu-sar-ra⸣ murgu₂-na, ki-gal-be₂ nu-sar
‘inscription on its pedestal’ ‘inscription of his back/shoulder, its
pedestal is uninscribed’

On the basis of these colophons, Sommerfeld assumes that on the original mon-
ument, a statue of Sargon, there was one continuous Akkadian text situated on
the back of the statue, which was then protected with one curse formula, pre-
served at the end of Sargon 1 (manuscript A obv. iii 35–41); leaving the pedestal
of the rear uninscribed. According to him, the corresponding Sumerian texts
were written in two parts (as Sargon 1 and 11) on separate sections of the pedes-
tal of the statue, explaining that both end with a protecting curse formula.14

||
13 Note that the colophons of Sargon 1 on manuscript B are not preserved; the text is broken
where they would be expected to occur. The colophon after the Sumerian version of Sargon 11
is preserved only on manuscript B, and the colophon after the Akkadian version of Sargon 11 is
preserved only on manuscript A. See also n. 32 and n. 33 below.
14 ‘Bei der akkadischen Fassung FAOS 7 Sargon C 1 und C 2 (RIME 2.1.1.1 und 11) handelt es
sich um einen einzigen zusammenhängenden Text auf der “Rückseite”, zu deren Schutz dann
auch eine Fluchformel ausreichend war. Die selbständige [sich] sumerische Version wurde
geteilt, an zwei verschiedenen Stellen auf dem “Sockel” angebracht und folglich auch mit zwei
 Mesopotamian Bilingual Royal Inscriptions from the Third Millennium BCE | 315

This reconstruction entails that the division of the original texts into Sargon
1 and 11 were the result of the editorial work of the Old Babylonian scribe(s) who
started from the separate Sumerian texts and adjusted the Akkadian version
accordingly.
The main argument for Sommerfeld’s reconstruction is the lack of a curse
formula in the Akkadian version of Sargon 11, as one would not expect separate
colophons for the Akkadian versions if there was one continuous Akkadian text
situated on the back of the statue. This argument, however, is undermined by
the fact that a monolingual, Akkadian version of Sargon 11 is also copied to both
manuscripts A and B,15 and the monolingual version also lacks a curse formula,
indicating that the lack of the curse formula in the Akkadian version of Sargon
11 cannot be explained by the assumption that Sargon 1 and 11 were inscribed
on the same monument.
Additionally, Sargon 1 is followed by seven fragmentary captions on both
manuscripts. These captions refer to the leaders involved in the conflicts narrat-
ed in the text, who were most probably carved into a relief on some part of the
original monument. The captions are distributed into two columns and their
presence suggests that the copy of the monument, i.e. the monument that car-
ried both versions of the text of Sargon 1, ends here and what follows belongs to
another one.16
So, in my view, pace Sommerfeld, there are good arguments for assuming
that Sargon 1 and 11 were also originally inscriptions belonging to separate

||
separaten Fluchformeln versehen’ (Sommerfeld 2008, 234, also see Sommerfeld 2021, 560). Cf.
also Buccellati’s (1993) paper in which he reconstructs a statue of Rimuš with numerous sur-
faces and pedestals, which, according to him, carried reliefs and inscriptions of Rimuš divided
by modern editions into five separate texts.
15 The fragmentarily preserved inscription begins in obv. xiii 43 on manuscript A and in rev. ii
17 on manuscript B. Frayne considers this inscription (Sargon 12 in his edition) ‘distinct from,
but similar to’ the Akkadian version of Sargon 11 on the ground that one ‘would not find an
original inscription copied twice on one Sammeltafel’ (Frayne 1993, 30). One could, however,
argue that the Old Babylonian scribes copied monuments but not texts, and indeed Sargon 12
ends with captions that were not part of Sargon 11. According to Sommerfeld, ‘(e)s handelt sich
nicht um direkte Duplikate, sondern um gleichlautende Inschriften, die auf zwei verschiede-
nen Monumenten wiederholt sind’ (Sommerfeld 2012, 198, n. 3). The edition of Gelb and
Kienast 1990 considered the monolingual inscription (Frayne’s Sargon 12) as another manu-
script of the bilingual version (Frayne’s Sargon 11, their Sargon C 2). Their edition is, therefore,
somehow misleading, the captions following Sargon 12 are not part of Sargon 11, so, the text
reconstructed by them and labelled as Sargon C 2 never existed in this form in reality.
16 For another, textual argument against Sommerfeld’s reconstruction, cf. the discussion of
Examples 7 and 8 below.
316 | Gábor Zólyomi

monuments. The monument that carried Sargon 1 may have been a statue with
the Akkadian inscription on its back or shoulder, and with the Sumerian text on
its pedestal. There must have been reliefs on it as well. The position of the in-
scriptions on the original monument of Sargon 11 is less clear on the basis of the
colophons, but Sommerfeld’s idea that the clause ‘its pedestal is uninscribed’
refers only to the rear is not unrealistic.
We have two colophons for Rimuš 18. On manuscript A, the colophon says:

Manuscript A rev. xviii 30 + rev. xvii 31

mu-⸢sar-ra urud⸣šen ⸢za-hum⸣
‘inscription on a zahum cauldron’

This colophon runs over two columns, so, it must refer to both versions. The
colophon on manuscript C, written under each of the versions, involves a word
whose reading is still unsolved.

Manuscript C rev. vii’ 43–45 rev. vi’ 43–45

(end of Sum. text) (end of Akk. text)
mu-sar-ra, mu-sar-ra,
TI URUD? X-ga du₃?/ni?, TI URUD? X-ga du₃?/ni?,
ri₂-mu-uš-kam ri₂-mu-uš-kam
‘inscription on … of Rimuš’ ‘inscription on … of Rimuš’

Now some words on the arrangement of the inscriptions on the manuscripts.

The two versions are always in two parallel columns, and not after each other.
This indicates that the scribe who made these collections wanted to present the
versions as those of the same text. This arrangement, Sumerian to the left, Ak-
kadian to the right, is well-known from other kind of scholarly texts of the Old
Babylonian period, such as lexical or grammatical lists. It is meant to demon-
strate that a given part of the Sumerian text corresponds to a given part of the
Akkadian texts.
The principle of parallel columns is so strong that it can even overwrite
the chronological arrangement on manuscript A. The two versions of Rimuš 18
are placed at the beginning of the last two columns on the reverse, with the
result that the text Maništusu 1 is, in fact, placed before them, see Fig. 2
above.
The desire to place corresponding semantic units next to each other is
demonstrated clearly on Fig. 3 below; see also Example 2 below.
 Mesopotamian Bilingual Royal Inscriptions from the Third Millennium BCE | 317

Fig. 3: The arrangement of the lines obv. i 33–37 and ii 35–37 on manuscript A; copy of the
corresponding lines from Poebel 1914, pl. XX–XXI (no. 34).

The Akkadian idiom ina tahāzim šaʾārum ‘to be victorious over/to defeat
someone/something in battle’ is translated with two finite verbs in Sumerian,
requiring five lines to write. The corresponding Akkadian clause is written in
three lines but there is deliberately so much space left between the lines that
they occupy approximately the same space as the corresponding five lines in
Sumerian.17
The arrangement of these texts on the tablets, therefore, clearly indicates
that the scribes considered these texts as each other’s translations which corre-
spond to each other exactly, unit by unit.18 The purpose of these arrangements

||
17 Cf. also Sommerfeld 2008, 234 for a description of similar adjustments in another part of the
tablet.
18 Also note that the arrangement of the texts in parallel columns seems to refute Lucrezia
Menicatti’s conclusion ‘that the Sumerian texts were original compositions rather than transla-
tions from the Akkadian ones’ (Menicatti 2019, 23). The Old Babylonian scribes obviously
considered these texts corresponding to each other unit by unit. She concludes that ‘the two
versions differ to a much larger extent, though. As for the writing system, the logographic
writings of verbal forms in the two versions show substantial discrepancies. The logograms
used in the Akkadian versions do not correspond to those of the Sumerian ones, and thus the
verbal stems are not semantic equivalent. Moreover, these verbs often require different gram-
matical constructions, and as a result, the whole structure and meaning of the Sumerian and of
the Akkadian texts may not be equivalent’ (Menicatti 2019, 22). Regarding the differences in
grammatical constructions, one wonders why it should be a requirement to use the construc-
tions of the source language but not its own for a text to qualify as a translation. Concerning
318 | Gábor Zólyomi

was probably educational. The scribes, native speakers of Akkadian, learnt the
Sumerian language by studying the equivalent constructions of the two lan-
guages.

3 The primary context

This part of the paper will discuss features relating to the primary context of the
bilingual inscriptions. Firstly, one can be certain that these texts, both the Ak-
kadian and the Sumerian versions, were composed in around the twenty-third
century BCE. Their orthography and grammar comply with those of the contem-
porary texts.19
In the case of the Sumerian versions, one can also notice, somehow unex-
pectedly, that these show linguistic features characteristic of texts from south-
ern Babylonia. One of the few linguistic features indicating dialectal differences
in third millennium BCE Sumerian is the phenomenon of vowel harmony. It op-
erates on certain verbal prefixes. Depending on the following syllable in some of
the verbal prefixes, the vowel may be either /i/ or /e/. This assimilation, the so-
called ‘Old Sumerian vowel harmony’, was an isogloss dividing cities in south-
ern Babylonia (Lagaš, Umma, Ur and Uruk, which exhibit the assimilation) from
those further north in Babylonia (Nippur, Adab, Šuruppag and Isin) in the twen-
ty-fifth to twenty-third centuries BCE. The assimilation disappeared in subse-
quent centuries.20
Neither a scribe from Nippur nor one from Agade would be expected to
write texts showing this phenomenon, but the vowel harmony characteristic of
texts from southern Babylonia is in operation in all three Sumerian versions.
The use of the southern dialect may have been more prestigious than the local
dialect.
Another, so far neglected, indication of the southern background of the
scribe(s) responsible for the Sumerian versions is the writing of the Sumerian
word murub ‘middle’ in Rimuš 18, see Example 1 below.

||
the writing system, the second part of this paper will argue that it was a deliberate choice to
create a Sumerian text that follows the writing conventions of southern Babylonia.
19 See Sommerfeld 2021, 561–564, who shows that the Akkadian texts use the orthography
and sign forms of the local Nippur tradition, while the Sumerian versions use the orthography
and sign forms of southern Babylonia.
20 See Zólyomi 2017, 29–33.
 Mesopotamian Bilingual Royal Inscriptions from the Third Millennium BCE | 319

Example 1: Rimuš 18 16–19

mu₅-ru₅, diŋir-re-ne-ka, me-te-ne₂, ni-šid
‘(Rimuš, …, fashioned a tin statue of himself and set it up before Enlil.) He counted himself
among the gods.’

This writing, mu₅-ru₅(NI.UL), is attested so far only in administrative documents

from Lagaš written in the twenty-fourth century BCE.21 It is attested twenty-three
times in the administrative texts.22 Its use in Rimuš 18 suggests a southern back-
ground for the composer of the text.
A remarkable feature of Sargon 1 is the use of the Sumerian /nga/- prefix, cf.
Example 2, l. 41, and Example 3, l. 101 below, written with the sign GA in the
text.23

Example 2: Sargon 1 (= MS A obv. i 33–41 and ii 35–43)

33. lu₂ urim₂ki-ma-da 35. in tāḫazim(REC169)
34. ŋeštukul
35. e-da-sag₃ 36. urim₂ki
36. aga₃-kar₂
37. e-ne₂-seg₁₀ 37. iš₁₁-ar
38. ⸢iri⸣-ne₂ 38–39. u₃, ālam(URUki)
39. e-ḫul 40. inīr(SAG.GIŠ.RA)
40. bad₃-be₂ 41–42. u₃, dūr(BAD₃)-šu
41. e-ga-[seg₁₀] 43. iqqur(I₃.GUL.GUL)
‘(Sargon, king of Agade) fought with the ‘He was victorious over Ur in battle, and
man (= leader) of Ur, defeated him, con- (then, as a consequence of this victory) he
quered the city, and also destroyed its conquered the city, and (then, as a conse-
walls.’ quence of this conquest) he destroyed its
walls.’ (translation of Kogan 2014, 51)

||
21 For reading the signs NI UL as mu₅-ru₅, see Krecher 1985, 171, n. 76. The reading is con-
firmed by the following clause, where the same word is written as mu-ru: mu-be₂ mu-ru diŋir-
re-ne-ka, gu₃-de₂-a ensi₂ lagaški-ke₄, pa e₃ ba-ni-a ‘Gudea, the ruler of Lagaš, made their names
appear among those of the gods’ (Gudea Cylinder A xxvi 17–19).
22 Cf. DP 403 rev. i 5 (CDLI P221053), DP 406 rev. i 3 (CDLI P221056), DP 408 obv. iii 3 (CDLI
P221058), DP 437 obv. iv 5 (CDLI P221087), DP 461 obv. i 2, rev. ii 6 (CDLI P221111), DP 577 rev. ii
6 (CDLI P221227), DP 596 obv. ii 2–3 (CDLI P221246), DP 641 obv. iii 5, rev. i 9, ii 4 (CDLI
P221291), DP 654 rev. i 1 (CDLI P221304), Nik 1, 46 rev. i 3 (CDLI P221815), Nik 1, 35 obv. i 3 (CDLI
P221742), TSA 40 obv. i 1 (CDLI P221401), VS 14, 52 obv. iii 2 (CDLI P020067), VS 14, 177 obv. ii 2
(CDLI P020188), VS 25, 39 obv. iii 2 (CDLI P020246), VS 25, 102 obv. i 1–2 (CDLI P020309), VS 27,
8 obv. i 1, rev. iii 3 (CDLI P020325).
23 On the use of /nga/- prefix in Old Akkadian bilingual texts, cf. Jagersma 2010, 514–515.
320 | Gábor Zólyomi

Example 3: Sargon 1 (= MS A obv. iii 37–41 and iv 39–43)

97. dutu 105. d
utu
98. suḫuš-ne₂ 106. išid(SUHUŠ)-šu
99. ḫe₂-burx(KAxŠU)-re₆ 107. li-su₃-uḫ
100. numun-na-ne₂ 108. u₃ zēra(ŠE.NUMUN)-šu
101. he₂-ga-degx(RI)- degx(RI)-ge 109. li-il-qu₃-ut
‘(Whoever obliterates this inscription), may Utu uproot him and also destroy his lineage!’

This prefix occurs five times in this text and its use always correlates with that
of the conjunction u₃ ‘and’ in Akkadian. There is, however, a difference in use
between the two morphemes, as you can also see in Example 2: in Akkadian, all
three clauses are co-ordinated with u₃, but the /nga/- prefix occurs only in the
last finite verb of a series of co-ordinated clauses.
The Sumerian prefix appears to have a conjunctive function here, co-
ordinating clauses. In other texts, however, the /nga/- prefix has an adverbial,
additive function meaning something like ‘also, too’, as in Example 4 below:

Example 4: Gudea Cylinder A vii 10 (CDLI Q000377) (Lagaš, twenty-second century)

gal mu-zu gal i₃-ga-tum₂-mu
‘(The true shepherd Gudea) is wise, and able too to realise things.’
‘(The true shepherd Gudea) is wise, and therefore able to realise things.’

Clause co-ordination in Sumerian is unmarked most of the time, so apparently the

translator here applies the /nga/- prefix only to have a morpheme in the Sumerian
text that corresponds to the Akkadian conjunction u₃.24 Rimuš 18 ends with the
same curse formula as Sargon 1 (cf. Example 3 above), but no /nga/- prefix is used
there, indicating that this solution was the invention of an individual.
Clauses such as the one in Example 4, however, may also show that this in-
vented use of /nga/- might have a basis in Sumerian. Example 4 may also be
interpreted so that the clause with the /nga/- prefix expresses a logical conclu-
sion of the predicate expressed in the first clause: ‘(The true shepherd Gudea) is
wise, and therefore able to realise things’.
And this is exactly the function that is attributed to the conjunction u₃ in the
Sargonic corpus by Leonid Kogan in an article about this morpheme,25 as his
translation under the Akkadian version in Example 2 demonstrates.

||
24 A similar principle is in use in the so-called Old Babylonian Grammatical Texts (Black 1991).
These are long lists containing Sumerian-Akkadian verbal paradigms which try to establish one-to-
one correspondences between Sumerian and Akkadian morphological elements, resulting, there-
fore, sometimes in obviously incorrect Sumerian verbal forms, see Zólyomi 2005, 353–357.
25 Kogan 2014.
 Mesopotamian Bilingual Royal Inscriptions from the Third Millennium BCE | 321

The use of the /nga/- prefix in Sargon 1 is, therefore, a kind of Akkadism in
the Sumerian version and its presence may indicate that the relation between
the two versions of Sargon 1 is asymmetric: the Sumerian version has been pre-
pared with a knowledge of the Akkadian version. In other words, the Sumerian
version is dependent on the Akkadian version and was prepared by a competent
bilingual scribe using the Akkadian version as a starting point.
Consider again Example 2 above. The two versions are clearly not literal
translations. In ll. 40 and 43 of the Akkadian version, the text uses logograms to
write the Akkadian verbs, written in modern transliteration in brackets and with
capitals. These logograms are essentially Sumerian words. One would expect
the Sumerian version to use the same verbs. But this is not the case. It uses dif-
ferent verbs, which correspond only semantically to the Akkadian ones. The two
versions deliberately applied two different writing traditions; and the Sumerian
version was prepared by someone familiar with the writing conventions of
southern Babylonia.26
The difference between the two versions is even bigger in the first part of the
excerpt. In the Sumerian version, a sequence of two clauses corresponds to the
Akkadian clause in ll. 35–37. The Akkadian idiom ina tahāzim šaʾārum ‘to be vic-
torious over/to defeat someone/something in battle’ has no literal correspondence
in Sumerian; it is translated with the combined use of two idioms: ŋeštukul — sag₃
‘to fight with’ and aga₃-kar₂ — seg₁₀ ‘to defeat someone/something’.27
The idiom ‘to fight with someone’ is also attested in an inscription of Irikagi-
na, see Example 5 below, and Sumerian also has an expression dam-ha-ra — ak
‘to do a battle with someone’, see Example 6 below.

Example 5: Irikagina 4 iii 4’–5’ (CDLI Q001132) (Lagaš, twenty-fourth century)

iri-ka-gen₆-na-ke₄, tukul e-da-sag₂
‘Iri-kagina fought with him.’

Example 6: En-metena 1 i 25–27 (CDLI Q001103) (Lagaš, twenty-fourth century)

ummaki-da, dam-ḫa-ra, e-da-ak
‘(Ningirsu, Enlil’s warrior) did battle with Umma.’

The sequence of clauses used in ll. 33–37 of the Sumerian version is not attested
in any other Sumerian text; its first part (ŋeštukul e-da-sag₃) functions only as the

||
26 On features of these two traditions, called Nippur and Uruk tradition by him, see Sommer-
feld 2021, 561–565.
27 For a detailed discussion of the idiom aga₃-kar₂ — seg₁₀ ‘to defeat, to win over’, see Zólyomi 2018,
146–148.
322 | Gábor Zólyomi

counterpart of the Akkadian adverbial phrase ina tahāzim ‘in battle’; the bi-
clausal construction was invented to cover all semantic nuances of the corre-
sponding Akkadian clause by the scribe who translated the original Akkadian
into Sumerian.
One may also notice a dependence of the Sumerian version on the Akkadian
in Example 7 below:

Example 7: Sargon 1 (= MS A obv. iii 1–6 and iv 1–6)

Sum. 62–67 ⸢šar⸣-[um-GI], ⸢lugal⸣, kalam-ma-ra, d⸢en⸣-lil₂-le, lu₂-⸢erim₂⸣, nu-na-
[šum₂]
Akk. 67–72 [šar-ru-GI], šar(⸢LUGAL⸣), mātim(⸢KALAM⸣.MA[ki]), den-⸢lil₂⸣, ma-[ḫi-
ra], [la i-di₃-šum]
Sum. and Akk. ‘Enlil did not give Sargon, king of the Land, an opponent.’

Here, the Akkadian expression ‘not to give someone an opponent’ meaning

‘making someone to be a person without an opponent’ is almost certainly, de-
spite the break in the text, translated with the verb šum₂ ‘to give’ in Sumerian. It
is, therefore, a literal translation of the Akkadian idiom, which also uses the
verb ‘to give’ nadānum.
However, the same Akkadian expression in Sargon 11, see Example 8 below,
is translated with a causative form of the Sumerian verb tuku ‘to have’.

Example 8: Sargon 11 (= MS A obv. v 32–36 and vi 39–43)

Sum. 29–33 šar-um-GI, lugal, ⸢d⸣en-lil₂-le, lu₂ gaba-ru, nu-mu-ni-tuku
‘Sargon, the king whom Enlil made a man without opponent.’
(lit. ‘did not let him have an opponent’)
Akk. 36–40 šar-ru-GI, šarrum(LUGAL), šu den-lil₂, ma-ḫi-ra, la i-di₃-šum
‘Sargon, the king to whom Enlil did not give an opponent.’

This idiom is also attested in an inscription of Lugalzagesi, see Example 9 below,

and inscriptions of Gudea, ruler of Lagaš, from the twenty-second century.28

Example 9: Lugal-zagesi 1 ii 14–16 (CDLI Q001379) (Nippur, twenty-third century)

⸢den⸣-lil₂-le, [gaba]-⸢šu⸣-ŋar, nu-mu-ni-tuku
‘Enlil made him (i.e. a man without opponent’ (lit. ‘did not let him have an opponent’)

The scribe who prepared the Sumerian version of Sargon 1 might have also chosen
this idiom, but apparently, he chose another one nearer to the Akkadian one. The
different Sumerian translations of the same Akkadian idiom in Sargon 1 and 11

||
28 The expression gaba-šu-ŋar nu-tuku ‘who has no opponent’ also occurs in Gudea Cylinder
A ii 10 and xxiii 15.
 Mesopotamian Bilingual Royal Inscriptions from the Third Millennium BCE | 323

provide another argument against Sommerfeld’s assumption that Sargon 1 and 11

originally constituted one single inscription; see the discussion above.
Another example in which the Sumerian version might have been influ-
enced by the Akkadian one is Example 10 below, in which the Sumerian version
uses the verb keše₂ ‘to bind’ in the meaning ‘to moor’ instead of the more com-
mon us₂ ‘to move next to something’, see Example 11 below. In fact, this is the
only known example of using the verb keše₂ in this meaning. In Akkadian, the
verb rakāsum ‘to bind’ is attested in at least two more texts with this meaning.29

Example 10: Sargon 11 (= MS A obv. v 12–6 and vi 20–25)

9. ma₂ me-luḫ-ḫaki 11. elep(MA₂) me-luh-ḫa
10. ma₂ ma₂-ganki 12. elep(MA₂) ma₂-ganki
11. ma₂ dilmunki 11. elep(MA₂) dilmunki
12. kar ag-ge-de₃ki-ka 14–15. in ka₃-ri₂-im ši a-ka₃-de₂ki
13. bi₂-keše₂ 13. ir-ku-us
‘He moored the ships of Meluhha, Magan and Dilmun at the quay of Agade.’

Example 11: Gudea Cylinder A iv 4 (CDLI Q000377) (Lagaš, twenty-second century)

kar niŋin₆ki-na-ke₄ ma₂ bi₂-us₂
‘He moored the boat at the quay of Nigin.’ (lit. ‘he moved the boat to be next to the quay of
Nigin’)

In Example 12 below, the Sumerian verb dim₂ ‘to fashion’ is used with the abla-
tive prefix -ta-. This construction is unattested in any Sumerian text.

Example 12: Rimuš 18

6. […] 6. ⸢ma⸣-na-ma
7. alan an-na 7. ṣalam(DUL₃) amūtim(KUG.AN)
8. nu-ta-dim₂ 8. la ib-ni
‘(No one) had ever fashioned a tin statue.’

In Sumerian texts, the expression ‘to fashion a thing from a material’ requires
the THING in the terminative case (-še₃) and the MATERIAL functions as the object
of the verb ‘he fashioned the MATERIAL into a THING’, as showed in Example 13
below:

Example 13: Gudea Statue B v 45–47 (CDLI P232275) (Lagaš, twenty-second century)
ŋeš
eren-be₂, ig gal-še₃, mu-dim₂
‘He manufactured the cedar-beams into big doors.’

||
29 Cf. CAD R, p. 94 (rakāsu 1d).
324 | Gábor Zólyomi

Apparently, the Sumerian version in Example 12 reflects the syntax of the Akka-
dian version, in which the THING is the object of the verb, and the MATERIAL is
expressed as the possessor of the THING.
As the context is broken, it is difficult to explain the presence of the ablative
prefix, but one tentative explanation may be that the choice of the -ta- prefix
was influenced again by the Akkadian usage which used the preposition ina
with materials, thus, ‘from gold’ may be expressed as ina ḫurāṣim. A corre-
spondence between the Sumerian ablative and the Akkadian ina preposition is
well attested in later periods.

4 Conclusion
The arrangement of the texts on the Old Babylonian manuscripts discussed in
the first part of the paper shows that the scribes considered the Sumerian and
Akkadian versions of the inscriptions as translations which correspond to each
other exactly unit by unit. Their arrangement in parallel columns is well known
from other kinds of scholarly texts, such as lexical or grammatical lists. The
purpose of this arrangement was probably educational. The scribes, native
speakers of Akkadian, learnt the Sumerian language by studying the equivalent
constructions of the two languages.
The grammatical and lexicographic peculiarities listed in the second part of
the paper suggest that these bilingual inscriptions were first composed in Akka-
dian. The Sumerian versions may have been prepared depending on the Akka-
dian ones by bilingual people who were familiar with the writing conventions of
southern Babylonia.30
The need for the bilingual versions was ideological. Sargon established his
empire by defeating and capturing Lugalzagesi, king of Uruk, who had already
united the mainly Sumerian-speaking southern part of Babylonia under his rule.
By displaying a monument with inscriptions written in both main languages of
his territorial state, he and his successor Rimuš represented themselves as legit-
imate rulers of both parts of Babylonia.
This message may also have been strengthened by deliberately preparing a
Sumerian inscription with clearly southern features, mirroring the language
used by Lugalzagesi, Sargon’s enemy, on his inscriptions. In other words, the

||
30 For a partly similar conclusion, cf. Sommerfeld 2021, 560–561: ‘The Sumerian version is a
literal, in part awkward, translation of the Akkadian.’
 Mesopotamian Bilingual Royal Inscriptions from the Third Millennium BCE | 325

Sumerian version with its southern linguistic and orthographic features reflects
the targeted audience and demonstrates a sophisticated sense of manipulating
the medium according to the intended audience.31 It may also indicate that the
southern version of Sumerian, both in its written and spoken form, was consid-
ered more prestigious than the local version.

Acknowledgements

The research for this article was funded by the Nemzeti Kutatási, Fejlesztési és
Innovációs Hivatal (National Research, Development and Innovation Office): ‘A
Dictionary of the Sumerian Royal Inscriptions of the Third Millennium’, project
no. NKFI-135325.
I am most grateful to Flóra Rétlaki for her kindness in preparing Fig. 2 using
photos of the tablet.

Abbreviations
AO = Antiquités orientales, Musée du Louvre, Paris
CAD = The Assyrian Dictionary of the Oriental Institute of the University of Chicago, vols 1–21
CBS = Catalogue of the Babylonian Section, Penn Museum, Philadelphia, PA
CDLI = Cuneiform Digital Library Initiative (<https://cdli.mpiwg-berlin.mpg.de/>, accessed on 9
August 2023)
DP = Allotte de la Fuÿe 1908–1920
ETCSRI = Electronic Text Corpus of Sumerian Royal Inscriptions
(<https://oracc.museum.upenn.edu/etcsri>, accessed on 9 August 2023)
MTM = multiple-text manuscript
N = Nippur, Penn Museum, Philadelphia, PA
Ni = Nippur, Arkeoloji Müzeleri, Istanbul
Nik 1 = Nikolskij 1908
obv. = obverse
rev. = reverse
RIME 2 = Frayne 1993
TSA = Genouillac 1909
VS 14 = Förtsch 1916
VS 25 = Marzahn 1991
VS 27 = Marzahn 1996

||
31 Accordingly, in my view, the concerns of Hannes D. Galter that ‘in all three cases the two
versions are separate and independent texts that even show different orthographies in logo-
graphic writing of the same words’ (Galter 1995, 31) may not support the assumption that the
original objects did not carry a bilingual inscription.
326 | Gábor Zólyomi

References
Allotte de la Fuÿe, François-Maurice (1908–1920), Documents présargoniques, Paris: Leroux.
Black, Jeremy (1991), Sumerian Grammar in Babylonian Theory, 2nd rev. edn (Studia Pohl,
Series maior, 12), Rome: Editrice Pontificio Istituto Biblico.
Buccellati, Giorgio (1993), ‘Through a Tablet Darkly. A Reconstruction of Old Akkadian Monu-
ments Described in Old Babylonian Copies’, in Mark E. Cohen, Daniel C. Snell and David B.
Weisberg (eds), The Tablet and the Scroll: Near Eastern Studies in Honor of William W. Hal-
lo, Bethesda, MD: CDL Press, 58–71.
Cancik-Kirschbaum, Eva and Ingo Schrakamp (2022), ‘Transfer, Adaption und Neukonfiguration
von Schrift- und Sprachwissen in den Keilschriftkulturen Vorderasiens: Eine Annäherung’,
in Eva Cancik-Kirschbaum and Ingo Schrakamp (eds), Transfer, Adaption und Neukonfigu-
ration von Schrift- und Sprachwissen im Alten Orient (Episteme in Bewegung, 25), Wies-
baden: Harrassowitz, 1–79.
Crisostomo, C. Jay (2020), ‘Sumerian and Akkadian Language Contact’, in Hasselbach-Andee
2020a, 403–420.
Förtsch, Wilhelm (1916), Altbabylonische Wirtschaftstexte aus der Zeit Lugalandas und Uruka-
ginas (Vorderasiatische Schriftdenkmäler der Königlichen Museen zu Berlin, 14/1),
Leipzig: J. C. Hinrichs.
Foster, Benjamin (2016), The Age of Agade: Inventing Empire in Ancient Mesopotamia, London:
Routledge.
Frayne, Douglas (1993), Sargonic and Gutian Periods (The Royal Inscriptions of Mesopotamia,
Early Periods, 2), Toronto: University of Toronto Press.
Galter, Hannes D. (1995), ‘Cuneiform Bilingual Royal Inscriptions’, Israel Oriental Studies, 15:
25–50.
Gelb, Ignaz J. and Burkhart Kienast (1990), Die altakkadischen Königsinschriften des dritten
Jahrtausends (Freiburger altorientalischen Studien, 7), Stuttgart: Franz Steiner.
Genouillac, Henri de (1909), Tablettes sumériennes archaïques: Matériaux pour servir à
l’histoire de la société sumérienne, Paris: Geuthner.
Hasselbach-Andee, Rebecca (ed.) (2020a), A Companion to Ancient Near Eastern Languages,
Hoboken, NJ: John Wiley & Sons.
Hasselbach-Andee, Rebecca (2020b), ‘Multilingualism and Diglossia in the Ancient Near East’,
in Hasselbach-Andee 2020a, 457–470.
Jagersma, Bram (2010), A Descriptive Grammar of Sumerian, PhD thesis, Universiteit Leiden,
<https://openaccess.leidenuniv.nl/handle/1887/16107> (accessed on 25 March 2023).
Kogan, Leonid (2014), ‘Waw Sargonicum. On parataxis in Sargonic Royal Inscriptions’,
Zeitschrift für Assyriologie, 104: 42–55.
Krecher, Joachim (1985), ‘Die /m/-Präfixe des sumerischen Verbums’, Orientalia, 54: 133–181.
Marzahn, Joachim (1991), Altsumerische Verwaltungstexte aus Girsu/Lagaš (Vorderasiatische
Schriftdenkmäler der Staatlichen Museen zu Berlin, 25), Berlin: Akademie Verlag.
Marzahn, Joachim (1996), Altsumerische Verwaltungstexte und ein Brief aus Girsu/Lagaš (Vor-
derasiatische Schriftdenkmäler der Staatlichen Museen zu Berlin, 27), Mainz: Philipp von
Zabern.
Menicatti, Lucrezia (2019), Bilingualism and Kingship: An Analysis of Mesopotamian Bilingual
Inscriptions from the Old Akkadian until the Old Babylonian Period, Master of Arts thesis,
Universiteit Leiden, <https://hdl.handle.net/1887/77875> (accessed on 25 March 2023).
 Mesopotamian Bilingual Royal Inscriptions from the Third Millennium BCE | 327

Nikolskij, Mikhail Vasil’evich (1908), Dokumenty chozjajstevnnoj otcetnosti drevnejšej epochi

chaldei iz sobranija N. P. Lichaceva (Drevnosti Vostocnyja, 3/2), St Petersburg: s.n.
Poebel, Arno (1914), Historical and Grammatical Texts (Publications of the Babylonian Section, 5),
Philadelphia, PA: The University Museum.
Sommerfeld, Walter (2008), ‘Große Zahlen in den altakkadischen Königsinschriften’, Altorien-
talische Forschungen, 35: 220–237.
Sommerfeld, Walter (2012), ‘Recension of Rebecca Hasselbach, Sargonic Akkadian: A Historical
and Comparative Study of the Syllabic Texts, Wiesbaden: Harrassowitz Verlag, 2005’,
Wiener Zeitschrift für die Kunde des Morgenlandes, 102: 193–284.
Sommerfeld, Walter (2021), ‘Old Akkadian’, in Juan-Pablo Vita (ed.), History of the Akkadian
Language, vol. 1: Linguistic Background and Early Periods (Handbook of Oriental Studies,
152/1), Leiden: Brill, 513–663.
Sövegjártó, Szilvia (2023), ‘A Collector’s Edition of the Past: Personal Collections of Mesopo-
tamian Royal Inscriptions from the Old Babylonian Period’, in Jürgen Paul and David Du-
rand-Guédy (eds), Personal Manuscripts: Copying, Drafting, Taking Notes (Studies in
Manuscripts Cultures, 30), Berlin: De Gruyter, 447–476.
Van Dijk-Coombers, Renate Marian (2021), ‘The Use of Sumerian and Akkadian during the
Akkadian Period: The Case of the “Elites”’, in Louis C. Jonker, Angelika Berlejung and Izak
Cornelius (eds), Multilingualism in Ancient Contexts: Perspectives from Ancient Near East-
ern and Early Christian Contexts, Stellenbosch: SUN PReSS, 128–144.
Westenholz, Aage (1999), ‘The Old Akkadian Period: History and Culture’, in Walter Sallaberger
and Aage Westenholz (eds), Mesopotamien, Annäherungen, vol. 3: Akkade Zeit und Ur III-
Zeit (Orbis Biblicus et Orietnalis, 160/3), Freiburg, CH: Universitätsverlag / Göttingen:
Vandenhoeck & Ruprecht, 15–117.
Zólyomi, Gábor (2005), ‘Structural Interference from Akkadian in Old Babylonian Sumerian’,
Acta Sumerologica, 22: 335–360.
Zólyomi, Gábor (2017), An Introduction to the Grammar of Sumerian, Budapest: Eötvös Kiadó,
<https://www.eltereader.hu/kiadvanyok/english-gabor-zolyomi-an-introduction-to-the-
grammar-of-sumerian/> (accessed on 25 March 2023).
Zólyomi, Gábor (2018), ‘E-ana-tum and the Ruler of Arawa’, Journal of Ancient Civilisations,
33/2: 141–163.
328 | Gábor Zólyomi

Appendix: The Sumerian texts in translation

Sargon 1 (cf. <http://oracc.museum.upenn.edu/etcsri/Q000834>, accessed on 9
August 2023)32

1–20
Šarrukin, king of Agade, commissioner of Inana, king of the world, išib priest of An,
king of the Land, chief governor of Enlil, conquered the city of Uruk and demolished its
city walls. He fought with the leader of Uruk and defeated him.
21–29
He fought with Lugal-zagesi, king of Uruk, captured him, and took him to the gate of
Enlil in a neck stock.
30–41
Šarrukin, king of Agade, fought with the leader of Ur and defeated him. He conquered
his city and demolished its city walls.
42–52
He conquered the temple of Ninmarki, and demolished its walls. From Lagaš until the
sea he conquered all lands, all of them, and washed his weapons in the sea.
53–61
He fought with the leader of Umma and defeated him. He conquered his city and de-
molished its city walls.
62–80
Enlil made Šarrukin, king of the Land, a man without opponent, and gave him (all
the people) from the Upper sea to the Lower sea. From the Lower sea the citizen of Agade
exercise rulership.
81–93
Mari and Elam stood (in obedience) before Šarrukin, king of the Land. Šarrukin, king
of the Land, restored Kiš, and the city provided him with ....
94–101
Whoever obliterates this inscription, may Utu uproot him and destroy his lineage!

Sargon 11 (cf. <http://oracc.museum.upenn.edu/etcsri/Q001403>, accessed on

9 August 2023)33

1–13
Šarrukin, king of the world, was victorious in thirty-four battles. He demolished all city
walls as far as the shore of the sea. He moored the ships of Meluhha, Magan, and Dilmun
at the quay of Agade.
14–28
In Tuttul, Šarrukin, the king, prostrated himself before Dagan and prayed to him.
(Dagan then) gave him the Upper land, (including) Mari, Yarmuti, and Ebla, as far as the
cedar forests and the mountains of precious metal.

||
32 About the Akkadian version, MS A, obv. iv 44–45 (see Table 2): l. obv. iv 45 is read both by
Gelb and Kienast (1990, 160) and Frayne (1993, 12) as ⸢ki⸣-[gal]-⸢ba⸣ ‘on its pedestals’. Sommer-
feld readings are based on collations carried out on manuscripts A and B (cf. Sommerfeld 2012,
197). Note, however, that Sommerfeld’s new reading creates a grammatical problem not dis-
cussed and solved by him. Namely, the statue is referred to both by a non-human possessive
enclitic =/be/ (kigal=be=ak ‘of its pedestal’) and by a human one =/ane/ (murgu=ane=ak ‘of
his back’).
33 About the Akkadian version, MS A, obv. vi 48–49 (see Table 2): the second part of obv. vi 48
is read both by Gelb and Kienast (1990, 165) and Frayne (1993, 29) as ⸢alan⸣-na ‘of a statue’.
Sommerfeld readings are based on collations carried out on manuscripts A and B (cf. Sommer-
feld 2012, 197).
 Mesopotamian Bilingual Royal Inscriptions from the Third Millennium BCE | 329

29–37
In the presence of Šarrukin, the king whom Enlil made a man without opponent, thir-
teen (units of) troops eat daily.
29–37
Whoever obliterates this inscription, may An obliterate his name, may Enlil put an
end to his lineage, may Inana cut his … short!

Rimuš 18 (cf. <http://oracc.museum.upenn.edu/etcsri/Q000842>, accessed on

9 August 2023)

1–19
[Rimuš, king of the world: since the dawn of time no one had ever] fashioned a tin
statue [for Enlil], (but now) Rimuš, king of the world fashioned a statue of himself, which
is of tin, and set it up before Enlil. He counted himself among the gods.
20–27
Whoever obliterates this inscription, may Enlil and Utu uproot him and destroy his
lineage!
Zsolt Simon
Anatolian Theonyms in the Aramaic Version
of the Letoon Trilingual
Abstract: This paper discusses the formal adaptation of the local theonyms in
the Aramaic version of the Letoon trilingual. It provides a new, regular explana-
tion of the theonym KNDWṢ from Carian and an improved explanation of the
Lydian origin of ˀRTMWŠ.

1 Introduction
The so-called Letoon trilingual (N 320) is a decree introducing a new cult in
Xanthos (Lycia), that of the deities ‘the King of Kaunos’ and ‘ArKKazuma, the
king’. The inscription, discovered by the French archaeological mission to Xan-
thos in 1973, was written on three different sides of a 1.35 m high and 0.575 m
wide stele of local limestone,1 erected in the temple complex of Leto and her
children called Letoon, about 4 km south-west to the Lycian city of Xanthos
(Lycian Arñna), in 337 BCE.2 The stele now stands in Fethiye Museum.
The inscription was written in three different languages and alphabets: Ly-
cian A (the local vernacular, henceforth, ‘Lycian’) and Greek (a widespread
lingua franca in this region) on the front and backsides of the stele, and Arama-
ic (the official language in the Achaemenid empire), on one of the side faces of
the stele. The present paper does not deal with the general sociolinguistic issue
of the usage of these languages in Lycia; instead, it focuses only on a single
aspect, which, however, does contribute to the larger picture: the adaptation
strategies of the Aramaic version (KAI 319)3 regarding the local theonyms. The
description of the Aramaic adaptation strategies is hindered by the fact that
some of the theonyms have not been explained satisfactorily. Therefore, the
goal of this paper is to provide an improved explanation of these theonyms and,
thus, clarify the adaptation strategies of the composer of the Aramaic text.

||
1 Metzger 1979b, 31.
2 See most recently Christiansen 2021, 48 with references.
3 Editions: Dupont-Sommer 1979; Lemaire 1995.

Open Access. © 2024 the author, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-012
332 | Zsolt Simon

Fig. 1: The Letoon trilingual stele, fourth century BCE (Fethiye Museum, Fethiye, Türkiye). Source:
Wikimedia Commons (<https://www.flickr.com/photos/sarah_c_murray/4914812770/>).
 Anatolian Theonyms in the Aramaic Version of the Letoon Trilingual | 333

In addition to the newly introduced deities, the King of Kaunos and ArKKazu-
ma, the king, the Lycian version, more precisely its final section (lines 38–39),
also mentions other local deities: ‘the Mother of this precinct’ and her children
(i.e. Leto, Artemis and Apollo), and the Elijãnas.4 In the Aramaic version, the
King of Kaunos is one of the theonyms to be explained (see Section 2). ‘ArK-
Kazuma, the king’ is never mentioned by his name but consistently mentioned
only as ‘his companions’ (W-KWNTH, lines 8 and 23, possibly an Elohim-type
pluralis majestatis in divine names),5 i.e. of the King of Kaunos. Whatever this
means, the underlying deity and its name are completely obscure, including its
reading as visible in the conventional transliteration with -KK-,6 and, therefore,
it must be left for further investigations. Setting aside the Elijãnas, who are
simply mentioned as the ‘others’ (W-ˀḤWRN, line 25), Leto and her children
appear in an interesting mixture: Leto as LˀTW, Artemis as ˀRTMWŠ, and Apollo
as ḤŠTRPTY (lines 24–25). While LˀTW points to an underlying Dorian Greek
form,7 the spelling of Artemis is completely unexpected, and, thus, it will be
discussed in Section 3. The term ḤŠTRPTY is linguistically clear, being the Ara-
maic transcription of Old Persian *xšaθra-pati- ‘Lord of Power’,8 and, thus, it
will not be discussed further here.9 The paper closes with a summary of the
observations (Section 4).

||
4 Laroche 1979. Unsurprisingly, the Greek version (Metzger 1979b) shows nothing unexpected:
Greek theonyms were used, if Greek equivalents were available (Leto in line 34 [Artemis and
Apollo are referred to as ‘(her) children’] and the Nymphs for the Elijãnas, line 34–35); in the
other cases, a Greek translation, if applicable (Basileus of Kaunos, lines 7, 15–16, 22–23) and
otherwise, a simple transcription, i.e. a phonetically adapted form (Arkesima, lines 7–8, 16, 23).
5 Dupont-Sommer 1979, 145–146.
6 For overviews, see Melchert 2004, 16–17; Raimond 2004, 401–407; Neumann 2007, 26; Moli-
na Valero 2016, 36–37; Vernet forthcoming; all with detailed references.
7 Dupont-Sommer 1979, 155. Leto as such is not attested yet in Lycian, only its genitival adjec-
tive, if this is the correct analysis of leθθi in TL 44b, 61 (for different proposals see the overviews
in Melchert 2004, 35; Neumann 2007, 185–186, with references), which, however, does not tell
us anything about the Lycian stem of Leto. Nevertheless, Lycian did not have /o/ and, there-
fore, sound substitution is expected. While /u/ cannot a priori be excluded, Lycian did not have
u-stems. Exact parallels do not exist, but Greek names in -ων were reinterpreted as a-stems: e.g.
Ijera* from Ἰέρων and Pa[r]mna from Παρμένων (Melchert 2004, 97, 101), which argues for a
substitution with /a/. Therefore, it is more probable that LˀTW with its -W reflects the Greek
form.
8 Mayrhofer 1979, 184–185.
9 Its employment for Apollo is a complicated problem of religious history, which is beyond the
scope of this paper.
334 | Zsolt Simon

2 The King of Kaunos

The god called ‘the King of Kaunos’ in the Lycian and the Greek versions ap-
pears as KNDWṢ ˀLHˀ ‘KNDWṢ, the god’ in the Aramaic version (lines 7–8, 12,
16, 21, 22–23 [in line 22, spelt as KNDWS]), which is qualified in line 8 with
KBYDŠY. Setting aside ˀLHˀ ‘god’, the phrase is not in Aramaic, the question is
therefore, from which language it originates and what it actually means.
The publishers believed that KNDWṢ was the transcription of the Lycian
term (‘manifestement’).10 As for KBYDŠY, André Dupont-Sommer explained the
suffix -ŠY as Anatolian, more precisely, Lycian or Carian.11 Considering the gen-
eral setting of the inscription, a Lycian derivation was certainly logical (the
Carian idea could not have been checked, since at that time it had not yet been
deciphered).12 Nevertheless, Ignasi-Xavier Adiego rightly objected that both
words are incompatible with the Lycian words:13 S/Ṣ cannot be reconciled with
the Lycian word for ‘king’, χñtawat(i)-,14 nor -ŠY with the Lycian suffix -ẽñni.
Mariona Vernet also argued that KNDWṢ/S cannot be a Lycian word, since none
of the cases of χñtawat(i)- show a sibilant ending.15 The phonetic criticism is
fully justified, assuming that the Lycian word was i-mutating: although the

||
10 Dupont-Sommer 1979, 145 (cf. already Dupont-Sommer 1974, 142–143 with n. 2); Lemaire
1995, 425 with n. 12 (who, however, even claimed ‘or rather Caunian/Carian’); followed by
Garbini 1977, 270; Fried 2004, 147, the latter ignoring the critical literature to be discussed
below.
11 Dupont-Sommer 1979, 145 (and not Carian only, contra Carruba 1977, 293; Adiego 1995, 19,
it was only the preliminary edition [Dupont-Sommer 1974, 143] that considered only Carian
[with a question mark]).
12 Onofrio Carruba criticised Dupont-Sommer’s description, claiming that there are many
Anatolian possibilities, such as the -assi-/-assa- toponym suffix. This solution, however, can
certainly be excluded, since this would have given Aramaic †-SY or †-S.
13 Adiego 1995, 18 (followed by Molina Valero 2004, 1015; Molina Valero 2016, 29–30, without
reference).
14 Dupont-Sommer 1979, 145 claimed that the Lycian <t> corresponds to the Aramaic sibilant,
but, firstly, as Adiego 1995, 18, n. 1, rightly pointed out, there is absolutely no evidence for the
assibilation of Lycian <t> /t/. Secondly, although Aramaic <Ṣ> could have represented an inter-
dental fricative, this is not the phonetic value of the Lycian <t> and, anyway, this possibility
was present only in Old Aramaic and not in Imperial Aramaic (see e.g. Segert 1983, 92–93).
Lemaire 1995, 425, n. 12 wondered whether Aramaic -Ṣ transcribes the nominative ending ‘-tis >
t(i)s > ṣ’, but there is no such change in the nominative ending.
15 Vernet 2021a, 87, n. 11.
 Anatolian Theonyms in the Aramaic Version of the Letoon Trilingual | 335

word is not attested yet in nom./acc. sg./pl., which would clearly decide the
question, the closely related (Hieroglyphic) Luwian form is i-mutated.16
After the decipherment of Carian, Adiego provided a new analysis, in which
he argued that KNDWṢ/S KBYDŠY is nothing else but the Aramaic spelling of
the Carian phrase ‘el rey caunio’.17 He argued that KNDWṢ/S is identical to the
Carian word for ‘king’, attested as kδowš in esaγδowš of AS 7 (from *esan-kδoúš
with a secondary voicing of [k] after the nasal), ‘cuya coincidencia con licio
χñtawat(i)-, luvita handawatti(i)- resulta difícil de negar’.18 He emphasised that
the Carian word fully matches the Aramaic spelling, including the final sibilant,
which is, in his view, either the result of a stem-final -t with the nominative
ending -s or a suffix, the Carian equivalent of Lycian -za- building professions.19
As for the suffix -ŠY, he identified it with the Carian ethnic suffix -si-, and ex-
plained the discrepancy between the sibilants with the argument that the differ-
ent Carian sibilants were not very clear to the adaptors, as the vacillation in this
very name shows.20 Later, he alternatively suggested that it could reflect the
Carian genitival adjective with the i-mutation.21
A similar argumentation was presented by Diether Schürr regarding
KNDWṢ/S, but on a much broader Carian basis.22 He identified the underlying
Carian word as *kδusi, which would even be attested in kδxśi from Abydos,
which he read as kδuśi and assigned the meaning ‘commander’. As supporting
evidence, he, furthermore, quoted the words kδuśolš (nom. pl.) as ‘königliche’;
kδuśo- from Hyllarima; and rmkδuśioś from Abydos, i.e. a patronym kδuśio in
gen. sg., in which he saw an earlier form of kδuśo- (from *k(i)ndusyó-).
He had, however, very different views on KBYDŠY: according to him, this
would show the same suffix -ŠY as the Aramaic ˀPŠŠY ‘Ephesian’ from Sardes,

||
16 Bauer 2020. Presumably the same applies to the similarly closely related Pisidian form as
well, if the word underlying the personal name Gdebetis is identical to this word (see most
recently Sasseville 2020b with references). Note that the interpretation of χñtawati in TL 35, 1 is
debated (nom. sg. according to Melchert 2004, 84, but Neumann 2007, 128 proposed dat.-loc.
sg., followed by Sasseville 2020a). To be sure, there is yet another local language, Lycian B, the
relation of which to Lycian A is unclear. Nevertheless, Lycian B can surely be excluded as a
candidate, since, although the word for ‘king’ itself is not attested in it yet, Lycian B has -b-
instead of -w- in the word family of χñtawat(i)- (cf. e.g. Melchert 2004, 136).
17 Adiego 1995, 19–21 (followed by Carruba 1999, 51–52; Kottsieper 2001, 198, n. 7b; and Rai-
mond 2004, 397–399).
18 Adiego 1995, 19.
19 Adiego 1995, 19–20.
20 Adiego 1995, 20–21.
21 Adiego apud Molina Valero 2004, 1016; Molina Valero 2016, 31.
22 Schürr 1998, 145–147.
336 | Zsolt Simon

translating Lydian ipsimšiš ‘dto’23 and reflecting the Lydian suffix -ši-, directly
on the stem of the toponyms.24 He did not explain, however, why a Lydian suffix
would appear on a Carian term such as in KBYDŠY, especially in Lycia, far away
from Lydia without Lydian speakers. In my view, Vernet offered a more logical
solution for ˀPŠŠY:25 she argued that we are dealing with a regular Aramaic
nisba in -Y and ˀPŠŠ- simply reflects ipsimši-, with a trivial assimilation of the
first sibilant to the second. In other words, I add, we would be dealing with the
Aramaic spelling of the Lydian term, re-characterised by the Aramaic nisba,
presumably because the composer of the Aramaic version had limited
knowledge of Lydian, the local vernacular. We will see below that this is not an
isolated phenomenon.
Another attempt to explain KBYDŠY was presented by André Lemaire, who
suggested that KBYDŠY is a regular ethnic name in -Y from KBYDŠ, which could
‘perhaps’ be a transcription of ‘Lydian chbide [sic] with a (Luwian nominative?)
ending -is’.26 However, all aspects of this idea are completely ad hoc. What kind
of ending is this ‘-is’? Why does it appear as Š instead of S? Why would any
Luwian ending be attached to a Lydian toponym, especially in an inscription from
Lycia? And why would the Lydian name have been used for a Carian toponym?
Finally, the Carian interpretation was picked up and further developed by
Vernet.27 She argued that KNDWS appears only once (line 22), while KNDWṢ
appears elsewhere (four times) and, thus, KNDWS is a scribal error and it is
KNDWṢ that needs explanation.28 She accepted both the idea that KNDWṢ is
based on the Carian word for ‘king’ and the Carian words provided by Adiego
and Schürr with the meaning ‘king’ or similar to it, except rmkδuśioś, which she
did not mention.29 Furthermore, she identified the word kdow- ‘king’ extracted
from esaγδowš of E.AS 7 with kδouś (gen. sg.) in E.Bu 1 (‘probably the same’).30
In both cases, she defended the meaning ‘king’, claiming that they ‘appear very

||
23 Dupont-Sommer 1979, 145.
24 Schürr 1996, 152, n. 4; Schürr 2018, 14.
25 Vernet 2021a, 89, n. 23.
26 Lemaire 1995, 426.
27 Vernet 2021a.
28 Vernet 2021a, 85, n. 9, referring to further palaeographic errors in the Aramaic text, for
which see esp. the list of Lemaire 1995, 431. For Garbini 1977, 271 it was only a phonetic variant,
which hardly explains anything.
29 Vernet 2021a, 87–88.
30 Carian inscriptions are quoted according to the standard reference system of Adiego 2007.
In this system, the sigla refer to the inscriptions’ findspot (E = Egypt, C = Caria, Ab = Abydos,
AS = Abu Simbel, Bu = Buhen, H = Hyllarima, xx = findspot unknown).
 Anatolian Theonyms in the Aramaic Version of the Letoon Trilingual | 337

near’ to the pharaoh’s name in both inscriptions.31 She claimed that KNDWṢ can
represent a gen. sg. kdouś ‘king’, ‘since Carian -ś, in its use in pismaśk could
represent a palatal fricative [ç] or even an affricate [ts]’,32 or a dat. sg. in -s (cf.
the spelling variant KNDWS), which would fit the Aramaic syntax showing L-
‘to’ before this phrase. Nevertheless, considering the discussion around the
Carian dat. sg. ending -s, she is cautious regarding the latter solution.
As for KBYDŠY, Vernet rightly pointed out33 that, firstly, we have learnt in
the meanwhile that ‘Kaunian’ is kbdyn- in Carian and not *kbdsi-; secondly, the
stem of the city’s name is KBYD only; and, thirdly, -ŠY cannot be an Aramaic
ethnic suffix as there is no such Aramaic ethnic suffix.34 Therefore, she sug-
gests35 that KBYDŠ° represents the Aramaic spelling of the gen. sg. of the city
name, fitting both semantically (completely equivalent to the Greek and Lycian
formulations) and formally (the Carian gen. sg. is -ś, a sort of palatal sibilant).36
She explained the remaining °Y as a re-characterisation by the Aramaic nisba
suffix -Y in order to make it clear that the last word of the sequence KNDWṢ ˀLHˀ
KBYDŠY should be understood as an adjective indicating the god’s origin for an
Aramaic reader unfamiliar with Carian.
Vernet’s solution for KBYDŠY convincingly solves all the problems.37 I
doubt, however, that her explanation for the re-characterisation is satisfactory,
since for a reader unfamiliar with Carian this is but little help, since the first
word, KNDWṢ, is also opaque, not to mention that the presence of the toponym,
and thus, the ‘affiliation’ of the god, could also have been heard from this
phrase even by a non-Carian speaker. Therefore, I prefer the explanation I have
already provided in Vernet’s paper:38 the Aramaean scribe did not know Carian
well enough and did not recognise the gen. sg. ending and, consequently, add-
ed the nisba-suffix.

||
31 Vernet 2021a, 87–88.
32 Vernet 2021a, 89.
33 Vernet 2021a, 89–90.
34 See Lemaire 1995, 426.
35 Vernet 2021a, 90–91.
36 See esp. Schürr 2001, 116–118; Melchert 2002, 310–312; Adiego 2007, 250–251. There could
be an objection that according to some proposals, Kaunos was a plurale tantum in Carian:
nevertheless, this proposal is far from proven and generally accepted, since the phrase in
which it appears is opaque (see most recently the detailed discussion in Simon 2022b with
references).
37 Add also that the similar Carian suffixes, -si- and the genitival adjective -s-, both proposed
previously as per above, do not have a palatal sibilant and, thus, they do not fit anyway.
38 Simon apud Vernet 2021a, 91, n. 24.
338 | Zsolt Simon

Whatever the explanation of the re-characterisation is, KBYDŠY seems to

have found a fitting explanation. This is, however, not the case with KNDWṢ.
We saw that it is neither Lycian nor Lydian, but all three Carian solutions, i.e.
those of Adiego, Schürr and Vernet, have two common problems: they do not
explain the spelling with -Ṣ and they assume a Carian word for ‘king’ based on
words that have no demonstrated meaning ‘king’.
As for the spelling, even if we assume that the Carian word for ‘king’ was a
dental stem and not an i-mutating stem (contrary to the closest relatives, as per
above), Adiego’s explanation of -Ṣ from the dental stem and the -s nom. sg.
ending still cannot be maintained, since the Carian nom. sg. ending is a zero
morpheme. His alternative, the Carian equivalent of -za- building professions,
suffers from the lack of any similar Carian suffix and, due to the only partly
known Carian historical phonology, from the uncertainty about precisely what
it should look like in Carian. Similarly, even if we assume with Schürr that
*kδusi means ‘king’, for which we have no evidence, it should also have been
consistently spelt as KNDWS.
The same applies to Vernet’s alternative as a dat. sg., which, contra her
suggestion, is not supported by the syntax: why would a Carian name get a
Carian ending in an Aramaic inscription triggered by an Aramaic prefix? Her
main solution, KNDWṢ as a gen. sg., is unfortunately incompatible with her
own explanation of KBYDŠY, since both cannot reflect the same sibilant: while
KBYDŠY is spelt consistently and, thus, implies an underlying Carian palatal
sibilant (which, if it is indeed a sibilant and does not represent a sound substitu-
tion, can only be ś or š), KNDWṢ/S obviously cannot have the same underlying
sibilant. Furthermore, it does not make sense morphologically either: why
would KNDWṢ be in the genitive singular and not in the nominative?
Turning to the alleged words meaning or supporting a meaning of ‘king’,
unfortunately none of them can be accepted:39
1) rmkδuśioś: Schürr referred to Ševoroškin40 as the source for this word, but
this inscription is known only from the transcription and claim of Še-
voroškin; it has never been published or documented, and, thus, it cannot
be considered.
2) esaγ?δowš (E.AS 7): Adiego’s assumption of a secondary voicing from
*kδowš is completely ad hoc. But even if this were correct, the beginning of
the word, esa-, and the final -š provide problems for which there is currently

||
39 See also Nunn and Simon 2023, 170–171.
40 Ševoroškin 1994, 143.
 Anatolian Theonyms in the Aramaic Version of the Letoon Trilingual | 339

no solution, as admitted by Adiego himself.41 Vernet’s claim that it stands

close to the name of the pharaoh does not prove anything since we do not
understand the inscription and the pharaoh’s name is, in fact, separated by
another word (mýqudem) of unknown meaning.42 Moreover, esaγ?δowš
seems to stand in agreement with the preceding word, býš,43 and if this is
correct, then they are in either nom. or acc. pl., which cannot be reconciled
with the meaning ‘king’.
3) kδouś (E.Bu 1): Unfortunately, the meaning and even the grammatical defi-
nition of this word are unclear; gen. sg. is only a possibility.44 Vernet’s claim
that it stands close to the name of the pharaoh does not prove anything
since we do not understand the inscription and the pharaoh’s name is in
fact separated by three words. Moreover, it is probably not the pharaoh’s
name at all: it seems to appear in an onomastic formula: (…) | ar[ĺ]iš |
psmškś | urmś | (…) ‘Arĺiš, (son) of Psmšk, (son) of Urm’,45 in which Psmšk is
obviously not the pharaoh Psammetichus.
4) kδuśolš (C.xx 4 & 5): Since these inscriptions, each one word on a bracelet,
are discussed in detail elsewhere,46 it suffices to quote here a summary of
that investigation: none of the several analyses provide a formally and se-
mantically fitting solution, with the exception of Günter Neumann’s theory
of a personal name (Kδ-uśol, with the well-known Carian name element
-uśol- and the widespread Luwic name element kδ-, here from *ḫanti(ya)-
‘first’), which can now be reinterpreted as an appurtenance adjective ‘of
Kδuśol’, referring to the owner of the bracelet.
5) kδụsiś (E.Ab 35): This word provides a papponym or a title or the origin of
the deceased. Adiego’s cautious choice of a papponym or an adjective of
origin from a toponym *Kδu-47 fits the context better than Schürr’s ‘com-
mander’, since he is forced to assume that the owner of the tomb, interred

||
41 Adiego 2007, 294, 364. Note that the reading esak?δowš was introduced by Schürr 2001, 108
without any reasoning (followed by Adiego 2007, 118); it would solve the phonetic problem but
is not supported by the photograph in Masson 1979, pl. IV/3.
42 See most recently Simon 2020b with references.
43 See the critical overview in Simon 2021a.
44 See the detailed discussion in Simon 2020a with references.
45 Cf. the phrase of the closely related inscription E.Bu 2: | arĺiš pdtomś uromś | ‘Arĺiš, (son) of
Pdtom, (son) of Urom’.
46 Nunn and Simon 2023, 168–172.
47 Adiego 2007, 269, 372; on the identification of the toponym see most recently Simon 2021c,
197–198 with references.
340 | Zsolt Simon

in Abydos (Egypt) was a ‘king’ in Caria, which he was clearly not, or a sort
of ‘commander’ in Egypt, clearly a petitio principii.48
6) kδuśo (C.Hy 1, 1): This word undoubtedly means ‘reign’,49 but its morpholog-
ical relation to the Luwic words for ‘king’ is completely unclear and thus, it
is not helpful.

All in all, we still do not have the Carian word for ‘king’ and none of the Carian
interpretations of KNDWṢ/S can be upheld. Therefore, we are looking for a solu-
tion that explains the spelling with -Ṣ in the nominative (as per Vernet above,
KNDWṢ seems to be the correct form) and, preferably, comes up with a fitting
Carian word for ‘king’, since both the Greek and Lycian traditions call this divin-
ity ‘King’ and Kaunos belonged to the Carian-speaking region (and, as we saw
above, it cannot be a Lycian A or B word). Stephan of Byzantium claimed that
gela means ‘king’ in Carian, but it does not help further and we cannot generally
contextualize this claim historically.50 We also saw that the Carian word for
‘king’ is not yet attested in the Carian inscriptions and the term ‘reign’ does not
help, since it is morphologically opaque. It is worth recalling, however, that the
closely related Lycian A and Luwian languages share the same word for ‘king’,
as per above. Therefore, it is logical to assume that the same word would have
been used in Carian as well. Nevertheless, we have to apply the known Carian
sound laws, and they lead to a nom. sg. form *k(V)δ(V)w(V)τ. The history of the
Carian vowels remains elusive, but they do not have a role in this case.51 <δ> is
phonologically /nd/ or /n.d/,52 naturally spelt as <ND> in the Aramaic version.
The nom. sg. -s had regularly disappeared.53 The key change is *ti > τ, in which τ
is a palatal or dental affricate,54 the closest equivalent of which is precisely Ṣ in
Imperial Aramaic. Therefore, KNDWṢ faithfully spells the expected Carian form
of the Luwic word for ‘king’.

||
48 One could, of course, assume that the personal name underlying the papponym or the
profession of the deceased goes back to a word that once meant ‘king’. For this, however, the
existence of the homonymous word for ‘king’ should independently be demonstrated, which is
not yet the case.
49 Adiego 2019a, 18–19, 23–24.
50 See the critical discussion in Simon 2022a.
51 Several scholars assumed the loss of unaccented vowels, but the details were defined very
differently, see e.g. Hajnal 1995, 17–21 and Simon 2023.
52 See the discussion between Kloekhorst 2008, 138–139; Adiego 2019b, 105.
53 See e.g. Adiego 2007, 312.
54 See most recently Simon 2021b, 56–60 with discussion and references.
 Anatolian Theonyms in the Aramaic Version of the Letoon Trilingual | 341

In other words, the composer of the Aramaic text chose to use the original
Carian name of the deity. According to Vernet,55 one possible reason for this fact
– instead of translating it from Lycian – is that the term ‘king’ is not used for
gods in Aramaic and Phoenician.56 However, considering the probable lack of
knowledge of Carian of the scribe writing in Aramaic (he did not seem to recog-
nise a simple genitive either), it is more probable that he simply did not under-
stand the term, and thus, used the original name of the deity.

3 ˀRTMWŠ
The case of ˀRTMWŠ ‘Artemis’ mentioned in line 24 of the Aramaic version is
considerably simpler, although it is remarkable from two points of view: firstly,
as the spelling with -WŠ shows, this is not the Greek form. Secondly, surprising-
ly enough, it is not the Lycian form either: the Lycian form of Artemis is well
attested and she is called Erteme/i- and Ertẽme/i-,57 neither of which is compati-
ble with the Aramaic spelling,58 and the same applies to her Phrygian name,
attested as Artimitos (gen. sg.), whence *Artimis, too59 (on Carian, cf. below).
However, the editor, Dupont-Sommer, has already compared60 the Lydian form
of Artemis, the single by-form of her name with -u- (Artimus,61 transliteration
modernised), and it was also identified as the Lydian form by Schürr62 and Ingo
Kottsieper.63 The solution is, formally speaking, undoubtedly correct.64 Howev-
er, Carlos Molina Valero objected that if this is correct, we have the problem of

||
55 Vernet 2021a, 86–87.
56 See already Teixidor 1978, 183, not quoted by Vernet.
57 Melchert 2004, 17; Neumann 2007, 72.
58 Pointed out also by Vernet 2021b, 533, n. 12.
59 See most recently Obrador-Cursach 2022, 136–138.
60 Dupont-Sommer 1979, 155.
61 See Gusmani 1964, 63–64 for the attestations.
62 Schürr 1996, 152, n. 4 (here without referring to Dupont-Sommer); Schürr 2018, 14.
63 Kottsieper 2001, 199, n. 24c, without references.
64 Vernet 2021b, 533, n. 12 claims that this name is attested in Carian, too, as ‘artm’ and ‘rtim’.
These are, firstly, personal names, but even if they are theophoric names from Artemis, they
obviously have no -u-. Moreover, based on Vernet’s own research (2016), rtim seems to be of
Iranian origin (*Rtima-) and it is not assured that ‘artm’ (recte artmi, C.Tr 2) is a theophoric
name (see Adiego 2007, 356–357 for the problems involved). This requires further investigation,
but if it turns out to be correct, it would show that the Carian form does not fit either. The claim
of Molina Valero 2016, 33 that the reading of this word is (a)rtimu and thus, comparable to the
Letoon form, is false.
342 | Zsolt Simon

whether we are dealing with a Lydian term or a Lydian loan in Carian.65 The
latter question is not a problem, since there is no reason to assume any Carian
transmission,66 and thus, the problems with the Lydian origin can more precise-
ly be formulated as follows: the spelling with -Š and the question of why an
Aramaic scribe would have used the Lydian form in Lycia.67 It is the spelling
with -Š that answers these questions.
The Lydian form has [s]68 and since Aramaic did have [s], ˀRTMWŠ cannot
be a direct spelling of the Lydian form. However, for historical phonological
reasons, the nominative of u-stems (such as Artimus) ends in -uš – in Old Per-
sian.69 Therefore, Lydian Artimus regularly became Old Persian *Artimuš, which
was precisely spelt as ˀRTMWŠ in the Aramaic version. There is nothing surpris-
ing in the fact that a scribe in the Achaemenid empire writing in Imperial Ara-
maic, the official language of the empire, uses the Old Persian forms of divini-
ties, if available, which was, in contrast, not the case with the other gods of the
inscription. This is precisely what happened to Apollo as well, who appears as
ḤŠTRPTY. But why was Lydian the source of the Old Persian form? If we look at
the expansion of the Achaemenid empire in Anatolia, we can find a logical sce-
nario: the first Anatolian region conquered by the Achaemenids was none other
than Lydia, and one of the main gods of Lydia was Artemis.70 Therefore, the
officers of the Achaemenid empire encountered Artemis for the first time in
Lydia and, logically, they adopted her Lydian name, which they used later
elsewhere.

||
65 Molina Valero 2004, 1017; Molina Valero 2016, 33.
66 Molina Valero 2004 and 2016, 22–34 argues that the major part of the non-Semitic onomas-
ticon in the Aramaic text is of Carian (and not Lycian) origin, but this is not correct, as Carian
origin could have logically been ascertained only in the case of Carian people and gods, and
Artemis is not one of them.
67 Schürr 2018, 14 only stated that ‘der aramäische Part der Trilingue spiegelt zumindest nicht
nur einen regionalen Sprachgebrauch wieder’, and that it was adopted in Aramaic from Lydia,
later used in the Aramaic of the Achaemenid chancellery (Schürr apud Molina Valero 2016, 33),
which does not explain anything.
68 Note that the earlier transcription with <ś> was phonetically incorrect, see e.g. Melchert
1994, 335.
69 See e.g. de Vaan and Lubotsky 2012, 200.
70 Compare with Payne 2019, esp. 240.
 Anatolian Theonyms in the Aramaic Version of the Letoon Trilingual | 343

4 Conclusion: Adaptation strategies in the

Aramaic version of the Letoon trilingual
We can summarise our findings regarding the uneasy situation of the composer
of the Aramaic version on how to deal with the local divine names as follows:
firstly, he substituted those divine names that were known to him with the Old
Persian equivalents (ˀRTMWŠ and ḤŠTRPTY).71 As for the remaining ones, he
simplified his task by dropping the names and referring to the divinities in a
blanket way, wherever he could (‘his colleagues’ and ‘the others’). This was not
possible in the case of the King of Kaunos and Leto, in the first case, due to ob-
vious reasons (the very goal of the decree) and in the latter case, perhaps due to
the deity’s importance. In these cases, he chose to transcribe the original name.72

Acknowledgements

This paper was written within the framework of the research project ‘The gods
of Anatolia and their names (continuity, importation, interaction): a philologi-
cal and linguistic approach’ (PID2021-124635NB-C31) financed by the Spanish
Ministry of Economy and Competitiveness.

Abbreviation
eDiAna = Olav Hackstein et al. (eds), Digital Philological-Etymological Dictionary of the Minor
Ancient Anatolian Corpus Languages, <https://www.ediana.gwi.uni-muenchen.de/> (ac-
cessed on 8 July 2023).

References
Adiego, Ignacio-J. (1995), ‘Contribuciones al desciframiento del cario’, Kadmos, 34: 18–34.
Adiego, Ignacio J. (2007), The Carian Language (Handbook of Oriental Studies, Section 1: The
Near and Middle East, 86), Leiden: Brill.

||
71 These are Old Persian loans in Aramaic. On the frequent Old Persian terminology in
Achaemenid Imperial Aramaic, see the overview of Tavernier 2013, 652.
72 Molina Valero 2016, 32–33 rightly wonders why the Greek form was used in the case of Leto
instead of a local form. This requires further investigation.
344 | Zsolt Simon

Adiego, Ignasi-Xavier (2019a), ‘A Kingdom for a Carian Letter’, in Ignasi-Xavier Adiego et al.
(eds), Luwic dialects and Anatolian: Inheritance and diffusion (Barcino Monographica Ori-
entalia, 12, Series Anatolica et Indogermanica, 1), Barcelona: Edicions de la Universitat de
Barcelona, 11–50.
Adiego, Ignasi-Xavier (2019b), ‘Consonant Clusters, Defective Notation of Vowels and Syllable
Structure in Caromemphite’, in Ronald I. Kim, Jana Mynářová and Peter Pavúk (eds), Hroz-
ný and Hittite: The First Hundred Years, Proceedings of the International Conference Held
at Charles University, Prague, 11–14 November 2015 (Culture and History of the Ancient
Near East, 107), Leiden: Brill, 95–119.
Bauer, Anna Henriette (2020), ‘Hieroglyphic Luwian /hantawatt(i)-/, REX-ti- ‘king’,’ in eDiAna,
<https://www.ediana.gwi.uni-muenchen.de/dictionary.php?lemma=1083> (accessed on
8 July 2023).
Carruba, Onofrio (1977), ‘Commentario alla trilingue licio-greco-aramaica di Xanthos’, Studi
Micenei ed Egeo-Anatolici, 18: 273–318.
Carruba, Onofrio (1999), ‘Ar/w/wazuma’, Kadmos, 38: 50–58.
Christiansen, Birgit (2021), ‘Palaeographical Dating of Lycian Inscriptions. A Critical Review of
Former Studies and a New Approach’, Hungarian Assyriological Review, 2: 27–63.
De Vaan, Michiel and Alexander Lubotsky (2012), ‘Old Persian’, in Holger Gzella (ed.), Lan-
guages from the World of the Bible, Boston: De Gruyter, 194–208.
Dupont-Sommer, André (1974), ‘La stèle trilingue récemment découverte au Létôon de Xan-
thos: le texte araméen’, Comptes rendus des séances de l’Académie des Inscriptions et
Belles-Lettres, 118: 132–149.
Dupont-Sommer, André (1979), ‘L’inscription araméenne’, in Metzger 1979a, 129–178.
Fried, Lisbeth S. (2004), The Priest and the Great King: Temple-Palace Relations in the Persian
Empire (Biblical and Judaic Studies, 10), Winona Lake, IN: Eisenbrauns.
Garbini, Giovanni (1977), ‘Osservazioni sul testo aramaico della trilingue di Xanthos’, Studi
Micenei ed Egeo-Anatolici, 18: 269–272.
Gusmani, Roberto (1964), Lydisches Wörterbuch, Heidelberg: Winter.
Hajnal, Ivo (1995), ‘Das Vokalsystem des Karischen. Eine provisorische Bestandsaufnahme’,
Die Sprache, 37: 12–30.
Kloekhorst, Alwin (2008), ‘Studies in Lycian and Carian Phonology and Morphology’, Kadmos,
47: 117–146.
Kottsieper, Ingo (2001), ‘Aramäische und phönizische Texte’, in Otto Kaiser (ed.), Texte aus der
Umwelt des Alten Testaments: Ergänzungslieferung, Gütersloh: Gütersloher Verlagshaus,
176–202.
Laroche, Emmanuel (1979), ‘L’inscription lycienne’, in Metzger 1979a, 49–127.
Lemaire, André (1995), ‘The Xanthos Trilingual Revisited’, in Ziony Zevit, Seymour Gitin and
Michael Sokoloff (eds), Solving Riddles and Untying Knots: Biblical, Epigraphic, and Semi-
tic Studies in Honor of Jonas C. Greenfield, Winona Lake, IN: Eisenbrauns, 423–432.
Masson, Olivier (1979), ‘Remarques sur les graffites cariens d’Abou Simbel’, in Jean Vercoutter
(ed.), Hommages à Serge Sauneron, vol. 2: Égypte post-pharaonique (Bibliothèque
d’étude, 82), Cairo: Institut français d’archéologie orientale du Caire, 35–49.
Mayrhofer, Manfred (1979), ‘Die iranischen Elemente im aramäischen Text’, in Metzger 1979a,
179–185.
Metzger, Henri (ed.) (1979a), Fouilles de Xanthos VI: La stèle trilingue du Létôon, Paris: Klincksieck.
Metzger, Henri (1979b), ‘L’inscription grecque’, in Metzger 1979a, 29–48.
 Anatolian Theonyms in the Aramaic Version of the Letoon Trilingual | 345

Melchert, H. Craig (1994), Anatolian Historical Phonology (Leiden Studies in Indo-European, 3),
Amsterdam: Rodopi.
Melchert, H. Craig (2002), ‘Sibilants in Carian’, in Matthias Fritz and Susanne Zeilfelder (eds),
Novalis indogermanica: Festschrift für Günter Neumann zum 80. Geburtstag, Graz: Ley-
kam, 305–313.
Melchert, H. Craig (2004), A Dictionary of the Lycian Language, Ann Arbor, MI: Beech Stave.
Molina Valero, Carlos (2004), ‘Usos onomásticos no semíticos en la versión aramea de la
inscripción trilingüe de Janto’, Interlingüística, 15: 1011–1019.
Molina Valero, Carlos (2016), ‘What Do We Know about the Letoon? A Study of a Sanctuary
during the Period of Achaemenid Rule over Lycia’, in María-Paz de Hoz, Juan Pablo
Sánchez Hernández and Carlos Molina Valero (eds), Between Tarhuntas and Zeus Polieus:
Cultural Crossroads in the Temples and Cults of Graeco-Roman Anatolia (Colloquia Anti-
qua, 17), Leuven: Peeters, 15–45.
Neumann, Günter (2007), Glossar des Lykischen: Überarbeitet und zum Druck gebracht von
Johann Tischler (Dresdner Beiträge zur Hethitologie, 21), Wiesbaden: Harrassowitz.
Nunn, Astrid and Zsolt Simon (2023), ‘Zwei Armreife mit karischer Inschrift aus der Archäologi-
schen Staatsammlung in München’, in Mark B. Garrison and Wouter F. M. Henkelman
(eds), The Persian World and Beyond: Achaemenid and Arsacid Studies in Honour of Bruno
Jacobs (Melammu Workshops and Monographs, 6), Münster: Zaphon, 163–176.
Obrador-Cursach, Bartomeu (2022), ‘The Gods of the Phrygian Inscriptions’, in Bartomeu Ob-
rador-Cursach and Ignasi-Xavier Adiego (eds), Phrygian Linguistics and Epigraphy: New
Insights (Barcino Monographica Orientalia, 20; Series Anatolica et Indogermanica, 3),
Barcelona: Edicions de la Universitat de Barcelona, 103–154.
Payne, Annick (2019), ‘Native Religious Traditions from a Lydian Perspective’, in Sandra
Blakely and Billie Jean Collins (eds), Religious Convergence in the Ancient Mediterranean
(Studies in Ancient Mediterranean Religions, 2), Atlanta, GA: Lockwood Press, 231–248.
Raimond, Éric (2004), ‘Dieux-rois du sud-ouest anatolien, de Kos et de Karpathos: le roi kau-
nien et arĝĝazuméen’, Res Antiquae, 1: 389–408.
Sasseville, David (2020a), ‘Lycian A xñtawate/i- ‘king, ruler’, in eDiAna, <https://www.ediana.
gwi.uni-muenchen.de/dictionary.php?lemma=1083> (accessed on 8 July 2023).
Sasseville, David (2020b), ‘Pisidian *γδεβετι- ‘king’, in eDiAna, <https://www.ediana.gwi.uni-
muenchen.de/dictionary.php?lemma=1083> (accessed on 8 July 2023).
Schürr, Diether (1996), ‘Zur karischen Felsgrabinschrift Si. 62 F’, Kadmos, 35: 149–156.
Schürr, Diether (1998), ‘Kaunos in lykischen Inschriften’, Kadmos, 37: 143–162.
Schürr, Diether (2001), ‘Karische und lykische Sibilanten’, Indogermanische Forschungen, 106:
94–121.
Schürr, Diether (2018), ‘Anmerkungen zum Namen der Karer’, Gephyra, 16: 13–17.
Segert, Stanislav (1983), Altaramäische Grammatik mit Bibliographie, Chrestomathie und
Glossar, 2nd edn, Leipzig: VEB Verlag Enzyklopädie.
Simon, Zsolt (2020a), ‘Carian kδouś “(unknown)”’, in eDiAna, <https://www.ediana.gwi.uni-
muenchen.de/dictionary.php?lemma=1071> (accessed on 8 July 2023).
Simon, Zsolt (2020b), ‘Carian †m=, †mý=, †=m’, in eDiAna, <https://www.ediana.gwi.uni-
muenchen.de/dictionary.php?lemma=967> (accessed on 8 July 2023).
Simon, Zsolt (2021a), ‘Carian bý “(an adjective)”’, in eDiAna, <https://www.ediana.gwi.uni-
muenchen.de/dictionary.php?lemma=2136> (accessed on 8 July 2023).
Simon, Zsolt (2021b), ‘Die karische Endung -τ’, Indogermanische Forschungen, 126: 53–64.
346 | Zsolt Simon

Simon, Zsolt (2021c), ‘Ḫinduwa: Kindye oder Kandyba?’, Hungarian Assyriological Review, 2:
197–203.
Simon, Zsolt (2022a), ‘Carian gela “king”’, in eDiAna, <https://www.ediana.gwi.uni-
muenchen.de/dictionary.php?lemma=2438> (accessed on 8 July 2023).
Simon, Zsolt (2022b), ‘Carian uiomλn “(unknown)”’, in eDiAna, <https://www.ediana.gwi.uni-
muenchen.de/dictionary.php?lemma=2440> (accessed on 8 July 2023).
Simon, Zsolt (2023), ‘Egyptian in Carian transmission: Towards a better understanding of
Carian vocalism’, in José Virgilio García Trabazo et al. (eds), New approaches on Anatolian
linguistics (Barcino Monographica Orientalia, 22; Series Anatolica et Indogermanica, 4),
Barcelona: Edicions de la Universitat de Barcelona, 241–252.
Ševoroškin, Vitaly (1994), ‘Carian ‒ Three Decades Later’, in M. E. Giannotta et al. (eds), La
decifrazione del cario: Atti del 1° Simposio internazionale, Roma, 3–4 maggio 1993, Rome:
Consiglio nazionale delle ricerche, 131–166.
Tavernier, Jan (2013), ‘Old Persian’, in Daniel T. Potts (ed.), The Oxford Handbook of Ancient
Iran, Oxford: Oxford University Press, 638–657.
Teixidor, Javier (1978), ‘The Aramaic Text in the Trilingual Stele from Xanthus’, Journal of Near
Eastern Studies, 37: 181–185.
Vernet Pons, Mariona (2016), ‘The Lycian PN Artimas and Arteimas: A New Proposal for an
Iranian and Epichoric Etymology’, Glotta, 92: 280–294.
Vernet, Mariona (2021a), ‘The Letoon Trilingual Revisited: Some Notes on the “King of Kaunos”’,
in Federico Giusfredi and Zsolt Simon (eds), Studies in the Languages and Language Contact
in Pre-Hellenistic Anatolia (Barcino Monographica Orientalia, 17; Series Anatolica et In-
dogermanica, 2), Barcelona: Edicions de la Universitat de Barcelona, 81–100.
Vernet, Mariona (2021b), ‘The Theonyms “Leto, Artemis and Ḥšatrapati” in the Letoon Trilin-
gual: A Parallelism with a Passage from the Xanthos Stele’, in Lluís Feliu, Adelina Millet
and Jordi Vidal (eds), ‘Sentido de un empeño’: Homenatge a Gregorio del Olmo Lete (Bar-
cino Monographica Orientalia, 16) Barcelona: Edicions de la Universitat de Barcelona,
527–540.
Vernet, Mariona (forthcoming), ‘Who Was ArKKazuma? Some Considerations that Point to a
Carian Origin of this Divinity’, News from the Lands of the Hittites.
Viola Allegranzi
Multilingualism in the Epigraphic Culture of
the Persianate World (Eleventh to Thirteenth
Century)
Abstract: Arabic and Persian coexisted in the eastern Islamic regions in medie-
val times, and their use can be related to specific contents and contexts. Inscrip-
tions seem to have remained more anchored in the Arabic models than manu-
scripts, since Persian rarely appears in epigraphic sources dating from the pre-
Mongol period. Moreover, epigraphic Persian is never independent but either
juxtaposed or mixed with Arabic. This paper examines a selection of monumen-
tal inscriptions and inscribed objects combining the two languages, and ex-
plores the ways in which these texts deviate from the standard. In most cases,
the linguistic interplay is intentional and reflects a specialisation of languages,
but linguistic interferences and scribal errors are also attested. An attempt is
made to define for each case study whether language-related choices depended
on the patron, the scribe or the recipient of the inscription, in order to contextu-
alize multilingual inscriptions better within the cultural production of medieval
Persianate societies.

1 Introduction
Successive conquests and political upheavals had a strong impact on the cul-
tural and linguistic background of the Iranian region in medieval times.1 The
main geo-political transformations can be outlined in three main phases: the
Arab conquests and annexation to the Caliphate (seventh–eighth centuries), the
rise of semi-independent dynasties of Iranian or Turkic origin (ninth–twelfth
centuries), and the Mongol invasions and incorporation into the Mongol empire
(thirteenth century). The linguistic situation in the region was equally complex
and changing, and its development cannot be completely reconstructed due to

||
1 We refer to a region wider than contemporary Iran, often designated as the ‘Persianate zone’,
which includes the lands between eastern Iraq, the Central Asian Steppe and northern India,
where Persian has long been the main language of literature and trade. Fragner 1999; Amanat
and Ashraf 2019.

Open Access. © 2024 the author, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-013
348 | Viola Allegranzi

the scarcity of primary sources, especially for the period preceding the tenth
century. It is clear, however, that the medieval Persianate society was multilin-
gual (as it still is today), since Arabic, New Persian and Turkic languages had
different uses across places and social classes, and coexisted with numerous
local languages and dialects of Iranian, Turkic or Indic origin.2 Although we
know from the sources that political and military elites of Central Asian origins
used to speak turk and often had Turkic names and honorifics, Arabic and Per-
sian established themselves as the main written languages of the eastern Islam-
ic regions in the pre-Mongol period.3
This paper looks at the use of languages in inscriptions on monuments and
objects from the medieval Persianate world.4 It focuses on the introduction of
Persian, its interplay with Arabic, and the progressive transformation of the
epigraphic models, which seem to have remained more grounded in the Arabo-
Islamic tradition compared with manuscript culture.5 Indeed, New Persian texts
in both prose and verse dating from the ninth century onwards are known,6
while the earliest preserved inscriptions in this language date back to the elev-
enth century. Even after this date, Arabic remained dominant in the epigraphic
culture of the eastern Caliphate, while Persian was used sporadically and usual-
ly juxtaposed or mixed with Arabic.7 The following sections explore the ways in

||
2 New Persian is a descendant of Old Persian (sixth–fourth centuries BCE) and Middle Persian
(third century BCE–ninth century CE). It developed after the Arab conquest and is written in
Arabic characters. On the rise of New Persian and the context in which it took place, see Lazard
1995; Ludwig Paul 2000; Perry 2009.
3 On the interaction between Arabic and Persian, see Bosworth 1978–1979. Only two eleventh-
century literary works in Qarakhanid Turkic written in Arabic letters have come down to us, see
Vásáry 2015. I follow the system of the International Journal of Middle East Studies (IJMES) for
the transliteration of Arabic and Persian.
4 Only after submitting the first draft of the paper did I become aware of Andrew C. S. Pea-
cock’s article on the interplay of Persian, Turkish and Arabic in eastern Islamic epigraphy
(Peacock 2023). However, Peacock examines documents from a wider chronological and geo-
graphical range, focusing on the epigraphic uses of Turkish, which makes our works rather
complementary.
5 On the introduction of Persian in Islamic epigraphy, see O’Kane 2009; Allegranzi 2019, vol. 1,
207–224.
6 Lazard 1963 and 1964. These texts are copied in later manuscripts. About the earliest known
handwritten documents in New Persian (tenth–eleventh centuries), see the preliminary results
of the project ‘Invisible East’ (PI, Arezou Azad, University of Oxford) at <https://invisibleeast.
web.ox.ac.uk/timeline-landing-page> (accessed on 5 May 2023).
7 Sheila Blair’s catalogue of Islamic monumental inscriptions from Iran and Transoxiana up to
the early twelfth century lists sixty-nine out of seventy-nine inscriptions in Arabic alone; five in
Arabic with the insertion of Persian words (Blair 1992, 34, 115, 118, 130, 174, nos 7, 42, 43, 48,
 Multilingualism in the Epigraphic Culture of the Persianate World | 349

which these two languages combined in epigraphic sources, and their associa-
tion with specific contents and styles.8 The purposes of this investigation are to
define the rationale of language-related choices and measure the success of
written and visual expressions experimented in eastern Islamic epigraphy be-
tween the eleventh and the thirteenth century.

2 Multiscriptual inscriptions
2.1 Arabic and Middle Persian
Epigraphic sources confirm that the Middle Persian (or Pahlavi) language and
script were still in use in some Persian-speaking milieux for a few centuries after
the Arab conquests.9 A small group of bilingual inscriptions in Arabic and Mid-
dle Persian dating from the eleventh century is particularly noteworthy as it
bears witness to the impact of the Arabic models and the hybridization of epi-
graphic language. The inscriptions appear on three tomb towers erected in the
historical region of Tabaristan (on the Caspian coast of northern Iran) by some
members of the Bavandid dynasty, who ruled locally between the eight and
fourteenth century.10 The bilingual inscriptions on the tomb towers at Radkan
(Iran, Golestan province, dated 407–411 AH / 1016–1021 CE) and Lajim (Iran,
Mazandaran province, dated 413 AH / 1022–23 CE) are located in epigraphic
bands encircling the cylindrical structure below the roof. In both cases, the
Arabic and Pahlavi texts record similar – but not identical – foundation texts
and give the date of construction according to the Islamic (hijrī) and Zoroastrian
(yazdigirdī) calendars. The foundation text on the tomb tower at Risgit (or Res-

||
66); three bilingual in Arabic and Middle Persian (Blair 1992, 85, 88, 208, nos 31, 32, 79); and
two in New Persian (Blair 1992, 128, 153, nos 47, 58). The use of Persian increased slightly dur-
ing the twelfth century, but the general trend remained unchanged.
8 Only selected excerpts from the inscriptions are quoted in the article, exemplifying the
linguistic choices of their authors; for the full texts of the inscriptions, please refer to the publi-
cations cited in the footnotes.
9 For an overview of Middle Persian inscriptions, their genres, chronology and geographical
distribution, see Engeskaug 2020.
10 On the monuments and their Arabic inscriptions, see Blair 1992, 85–90, 208–210, nos 31, 32,
79; on the Middle Persian inscriptions, see Bāghbīdī 1383 SH / 2004 CE and Cereti 2015, 161–169
(these authors reference previous studies by Max van Berchem, David H. Bivar, Ernst Herzfeld
and André Godard). See also an overview of the bilingual inscriptions in Allegranzi 2017, vol. 1,
367–374, Allegranzi 2019, vol. 1, 208–209. On the history of the Bavandids, see Madelung 1984.
350 | Viola Allegranzi

ket, Iran, Mazandaran province) is inscribed above the entrance door and con-
sists of three and a half lines in Arabic, followed by half a line in Middle Persian.
The end of the inscription is badly damaged, but it seems plausible that it origi-
nally recorded the date of construction, according to the Islamic and Zoroastri-
an calendars.11 The patrons and dedicatees in the inscriptions at Radkan and
Lajim bear Arabic names and honorifics (e.g. al-amīr ‘emir’, mawlā amīr al-
muʾminīn ‘client of the Commander of the Faithful’) along with others of Persian
origin (e.g. al-ispahbad ‘army chief’, shahriyār ‘king’). Moreover, the epigraphic
band below the roof of the Risgit tomb tower contains Quranic quotations
(Quran 12:36 and Quran 112). The hybrid titulatures and epigraphic formulary,
together with the juxtaposition of Arabic and Pahlavi script, are well-suited to
the cultural policy of the Bavandid rulers, who claimed descent from the Sasa-
nian kings but had converted to Islam as early as the ninth century.
Some choices related to the arrangement and palaeography of the bilingual
inscriptions are also noteworthy. The Arabic text in Radkan is written on one
line and occupies the longer section of the epigraphic band, while the Middle
Persian version is distributed on two lines within a shorter section. The Arabic
letters are traced in an ornate knotted and floriated Kufic style, and some orna-
mental motifs inspired by this style are appended to the smaller Pahlavi letters.12
The inscriptions in Lajim each run on one line on two superimposed bands: the
lower band contains the Arabic text in simple Kufic script, the upper band the
Middle Persian text, traced in smaller, undecorated letters.13 Both Middle Per-
sian inscriptions display a script derived from cursive Book Pahlavi, and their
graphic style seems to be influenced by the Arabic inscriptions next to them.14
The distribution and relative dimensions of the texts seems to confirm a kind of
hierarchy between the two languages and the predominance of the Arabo-
Islamic epigraphic model.

||
11 The chronological attribution of the monument is debated, but it seems likely that this
tomb tower was founded, like the other two, in the first half of the eleventh century. See Alle-
granzi 2017, vol. 11, 371.
12 See Flury 1921 and Cereti 2015, 165–166 for a palaeographic analysis of the Arabic and the
Pahlavi inscription, respectively.
13 One may admit that the different medium and technique (carved stucco in Radkan versus
cut brick in Lajim) influenced the different degree of plasticity and ornamentation of the in-
scriptions. The inscriptions are illustrated in Allegranzi 2017, vol. 2, pl. LXIII.
14 Cereti 2015, 162. Most funerary inscriptions in Middle Persian known to us employ cursive
Book Pahlavi instead of Inscriptional Middle Persian characters, and are, therefore, attributed
to the late Sasanian or early Islamic period (Engeskaug 2020, 177–178).
 Multilingualism in the Epigraphic Culture of the Persianate World | 351

Another noteworthy evidence of the survival but progressive decline of

Middle-Persian epigraphy is a short funerary inscription, undated, incised on a
Byzantine sarcophagus in Constantinople and dedicated to an Iranian immi-
grant. The use of a variety of cursive script similar to Book Pahlavi, together
with the fact that several words follow the phonetic spelling of New Persian,
point to a late dating (probably ninth or tenth century).15
The inscriptions mentioned so far are among the last examples of the use of
Pahlavi script in epigraphy and show a process of hybridization of the language
and writing style. They constitute an exception to the epigraphic practices de-
veloped in Islamic Iran, where the vast majority of inscriptions make use of the
Arabic (or Perso-Arabic) alphabet to transcribe texts in both Arabic and Persian,
abolishing the visual differences between the two languages.16

2.2 Arabo-Persian and Sanskrit

The earliest inscription in New Persian language and Arabic script known to
date is engraved on a rectangular limestone slab (55 × 36 cm) retrieved from the
village of Zalamkot in the lower Swat valley (close to Batkhela, Pakistan).17 The
inscription on the slab is multilingual and multiscriptual: it consists of seven
lines in Arabo-Persian,18 traced in simple Kufic script, followed by three (or four)
lines in Sanskrit, inscribed in Śāradā script. The Persian text commemorates the
construction of a building, probably a mosque, patronised by two political and
military authorities: the amīr (‘emir’) Arslān al-Jādhib, governor of Tus, and the
sarhang (‘general’) Biktāsh, castellan of Ajaypālnaghar (?).19 The date of comple-

||
15 See the reading and chronological attribution proposed by de Blois 1990.
16 New Persian has four more consonants than Arabic, which are transcribed adding three
diacritical points or a stroke to the basic shape of some letters (‫‘ پ‬p’, ‫‘ چ‬ch’, ‫‘ ژ‬zh’, ‫‘ گ‬g’).
However, diacritical marks are often omitted in medieval Islamic inscriptions, especially – but
not exclusively – in those executed in Kufic script.
17 The slab entered a private collection in Peshawar and was published by Abdur Rahman in
1998 (the only available illustration of the object is appended to this article); its current loca-
tion is unknown.
18 The main text is in Persian, but is interspersed with Arabic phrases (particularly the basma-
la, the shahāda or Profession of Faith, two benedictory formulae and the name of the month).
19 Rahman 1998, 469–470 first read the Persian inscription; see also Allegranzi 2017, vol. 1,
374–377; Allegranzi 2019, vol. 1, 209–211. Shavarebi 2022 proposes a new version, where he
convincingly emended the reading of the building type (mazgid ‘mosque’ instead of markad,
supposed to be an erroneous form for marqad ‘tomb’) and the name and titles of the second
patron. Arslān al-Jādhib was a Ghaznavid military commander and local governor in Khura-
352 | Viola Allegranzi

tion is given in a mixture of Arabic and Persian as dhū‘l-qaʿda 401 AH (6 June–5

July 1011 CE). Moreover, the construction verbs banā kard (‘he built’) and tamām
kard (‘he completed’) are both placed at the beginning of a sentence, as is
standard in Arabic foundation inscriptions. One can argue that the verb-initial
word order of Arabic influenced the syntax of the text, which deviates from the
subject-object-verb order common in Persian.20 The Sanskrit text is barely legi-
ble and has only been partially deciphered.21 It seems to correspond to an adap-
tation of the foundation text and conform to the model of Śāradā inscriptions; in
fact, the text opens with the date and the Sanskrit honorific śri (‘great honoura-
ble’) precedes the names of the patrons.
Although no information is available on the archaeological context of the
inscription, its content clearly indicates that it originally belonged to an Islamic
building founded in the lower Swat valley in the early eleventh century, possi-
bly following the conquest of this area by the Ghaznavid armies.22 The scribe
was probably acquainted with the Arabic language and epigraphic models, as
evidenced by the inclusion of Arabic formulae and the general structure of the
text. Why he composed the text in Persian is uncertain. His and/or his patrons’
intention was perhaps to facilitate the readers’ comprehension, as is the case in
the Sanskrit part, probably addressing the local community. Indeed, Persian,
besides being the favoured language of Ghaznavid court poetry and historiog-
raphy, may have been more widespread than Arabic among the soldiers and
immigrants settled in this peripheral area of the Ghaznavid state. At any rate,
the choice of composing a construction text in New Persian is rather unusual for
the time (the second known foundation inscription in Arabo-Persian is dated
547 AH / 1152 CE, see below, Subsection 4.1) and the scribe was probably forced
to experiment with new expressions due to the lack of an established repertoire
of epigraphic texts in this language. Finally, observing the stylistic features of
the Zalamkot inscription, one is struck by the lack of ornamentation: the writing

||
san, known from historiographical sources; on the contrary, the identity of Biktāsh remains
uncertain, although his name suggests that he was a military commander of Turkic origin,
possibly appointed in the Swat region. The identification of the toponym Ajaypālnaghar (per-
haps a corrupted form of Jayapālnagar) is similarly uncertain.
20 Note, however, that verb-initial phrases occur in later Persian epigraphic texts and manu-
script colophons, where an Arabic influence need not be assumed.
21 See Ingo Strauch’s reading in Shavarebi 2022.
22 The Ghaznavid dynasty (366–582 AH / 977–1186 CE), a Turko-Iranian line of Central Asian
origin, settled in Ghazni (in present-day Afghanistan) and founded an influential state stretch-
ing from Khurasan to northern India. On the history of the Ghaznavids and their conquests in
the first half of the eleventh century, see Bosworth 1963.
 Multilingualism in the Epigraphic Culture of the Persianate World | 353

style of the text engraved on the slab is closer to that of rock graffiti than to the
carefully designed and ornate monumental inscriptions retrieved from the
Ghaznavid capital city, Ghazni.23
The production of multiscriptual texts including Arabic versions is not en-
tirely new in the Pakistani area: indeed, two inscribed stones retrieved from the
Tochi valley in Waziristan bear an Arabic inscription (one dated 243 AH / 857 CE),
accompanied by a Bactrian and a Sanskrit text, respectively.24 Furthermore, a
series of bilingual coins (dammas) in Arabic and Sanskrit/Nagari were issued in
Multan in the ninth century. They set precedents for the bilingual coins issued
by the Ghaznavid ruler Maḥmūd b. Sebüktegin (r. 388–421 AH / 998–1030 CE) in
Lahore, which display an Arabic legend on the obverse and a Sanskrit transla-
tion on the reverse.25 The minting of such coins seems to be limited to two con-
secutive years (418–419 AH / 1027–1028 CE), which suggests that this model was
soon abandoned in favour of a return to more traditional coinage.
All in all, the appearance of multiscriptual inscriptions in peripheral areas
of the eastern Islamic regions appears to be a restricted phenomenon. The spe-
cific cultural policy and linguistic situation of their context of production and
reception in the case of both the Bavandid tomb towers in northern Iran and the
Ghaznavid bilingual inscriptions from Pakistan, seem to be the main reasons for
the deviation from the standard models of Arabo-Islamic epigraphy.

3 Persian and Arabic juxtaposed

New Persian seems to have spread as an epigraphic language during the elev-
enth century, when it appears on the royal monuments sponsored by the Qa-
rakhanid and Ghaznavid rulers in the eastern Iranian lands.26
An inscription entirely composed in the Persian language and in verse
frames the entrance arch of the brick portal of the complex known as Ribāṭ-i

||
23 For an overview of the epigraphic repertoire from Ghazni, see Flury 1925; Giunta 2003; and
below, Section 3.
24 Sims-Williams and de Blois 2018, 85–94.
25 Flood 2009, 39–42.
26 The Qarakhanids were a Turko-Muslim dynasty who ruled in Central Asia from the late
tenth to the early thirteenth century, see Jürgen Paul 2021. The monuments discussed here are
attributed to the western branches of the dynasty, based in Samarqand (Uzbekistan) and Uz-
gend (Kyrgyzstan).
354 | Viola Allegranzi

Malik (Navoi, Uzbekistan) (Fig. 1).27 The surviving sections praise a ‘building’
(banāī) comparable to paradise, erected by the ‘sultan of the world’ (sulṭān-i
jahān) with God’s blessing. The text appears to be a panegyric poem composed
on the occasion of the construction of Ribāṭ-i Malik itself, but does not provide
any details on its function, dating or patron. This architectural complex located
on the road connecting Bukhara and Samarqand was long considered to be a
caravanserai built in the times of the Qarakhanid ruler Shams al-Mulk Naṣr b.
Ibrāhīm (r. 460–472 AH / 1067–1080 CE); however, more recent studies revealed
that the structure underwent several transformations and propose that the
monument initially served as an extra-urban royal residence, eventually turning
into a caravanserai in post-Mongol times.28 Inscriptions in Arabic certainly
adorned the complex as well; in fact, Quranic verses (Quran 3:16–19) were rec-
orded on the lost minaret and other undeciphered fragments (probably later) on
the interior structures. Yet, the position of the Persian text above the main en-
trance gives it an undisputed prominence.
Three Persian inscriptions are sculpted on stucco at different heights on the
interior of the so-called Shāh Fażl mausoleum at Safid Buland (Kyrgyzstan,
447–451 AH / 1055–1060 CE).29 The upper text was the best preserved at the time
of the first surveys and contains an elegy for a deceased ruler, identified as the
Qarakhanid governor of Ferghana Sayf al-Dawla Muḥammad b. Naṣr. His son,
Muʿizz al-Dīn ʿAbbās, was probably the patron of the mausoleum (khvābgāh,
literally ‘sleeping place’), mentioned in the middle inscription. The latter has
only been partially deciphered; it was also composed in verse and possibly
functioned as a construction text, as suggested by the use of the verb farmūd
(‘he ordered’). The lower inscription is very fragmentary, but the surviving sec-
tions seem to address the reader a moralising message. Arabic formulae (the
shahāda and the doxology al-mulk li-llāh ‘sovereignty [belongs] to God’) occur at
the end of the two upper Persian inscriptions and around one decorative roun-
del; small-sized Quranic inscriptions frame other roundels adorning the inner
walls. It appears, however, that Arabic played a secondary role in the epigraph-
ic decoration of this building.

||
27 See on the monument, McClary 2020, 202–213, and on the inscription, Giunta and Allegran-
zi 2020, 121–122 (both studies reference previous works).
28 See especially Karev 2013, 125–126.
29 See Nastič 2019, 33–77; McClary 2020, 23–52 (both studies reference previous works).
 Multilingualism in the Epigraphic Culture of the Persianate World | 355

Fig. 1: The Persian inscription on the portal of Ribāṭ-i Malik (Navoi, Uzbekistan, probably sec-
ond half of the eleventh century). Photograph by Viola Allegranzi, 2015.

Numerous fragments of Persian inscriptions were discovered during the excava-

tions and surveys in Ghazni (Afghanistan), capital city of the once-mighty
Ghaznavid state.30 A series of more than two hundred marble panels (c. 75 × 40 cm)
with geometric and vegetal decoration, topped by short poetic inscriptions in
Persian is particularly noteworthy (Fig. 2).31 Most panels were excavated in a
Ghaznavid royal palace and once adorned the dado of the antechambers that
opened onto the building’s central courtyard (probably late eleventh–early
twelfth century). Their inscriptions composed a continuous text, made up of two
or more poems praising the Ghaznavid rulers, their military exploits and reli-

||
30 The Italian Archaeological Mission in Afghanistan investigated the Islamic and Buddhist
sites of Ghazni between 1957 and 1978. The archaeological and photographic materials collect-
ed have undergone systematic analysis in more recent times, and research on premodern
Ghazni is still ongoing. A digital archive and up-to-date information on the mission’s activities
are available at <https://ghazni.bdus.cloud/> (accessed on 5 May 2023).
31 Alessio Bombaci devoted a first study to the inscriptions carved on 115 dado panels (Bom-
baci 1966); 113 more inscribed panels of the same type were recorded during subsequent ar-
chaeological and research activities. The entire corpus is examined in Allegranzi 2017 and 2019.
356 | Viola Allegranzi

gious policy. The poetic language is consistent with that used in the surviving
poetic anthologies of the Ghaznavid court poets; it mixes common tropes of
Persian panegyric poetry with Arabic royal titles and notions borrowed from
Islamic religious sciences.32

Fig. 2: Marble dado panel with a fragment of a Persian inscription excavated in the Ghaznavid
palace (Ghazni, Afghanistan, probably late eleventh–early twelfth century, H 75, W 40, D 7 cm).
Text: … ʿulūm-i sharīʿat ‘… the sciences of the canonic Law’ (reading by Bombaci 1966, 14, no. 93).
Archives of the Italian Archaeological Mission in Afghanistan, inv. no. C2791.

||
32 Arabic loanwords are Persianized and adapted to the prosody. The Arabic article, for in-
stance, is generally omitted, the genitive case replaced by the Persian iżāfa, and the tāʾ mar-
būta (‫ )ﺓ‬by a final tāʾ maftūḥa (‫)ﺕ‬.
 Multilingualism in the Epigraphic Culture of the Persianate World | 357

One should note that several other inscriptions on different materials (marble,
baked brick, stucco) were part of the architectural decoration of the Ghaznavid
palace: they are all in Arabic and record construction texts, benedictory
phrases, doxologies and Quranic verses. The large epigraphic repertoire collect-
ed throughout the city confirms that Arabic remained the main epigraphic lan-
guage in premodern Ghazni. Nevertheless, a small group of marble finds of
different types bear traces of Persian inscriptions: they are all attributable to the
eleventh and twelfth centuries and probably belonged to lost buildings of the
Ghaznavid city, which demonstrates the significant role this dynasty played in
the development of Persian epigraphy.33
The Qarakhanid and Ghaznavid Persian inscriptions discussed so far share
several common features: they are part of the architectural decoration of royal
buildings, occupy prominent positions, and the best preserved among them
prove to be composed in verse. In addition, in all the sites mentioned above,
Persian poetic inscriptions coexist with Arabic texts recorded in separate epi-
graphic registers and having different contents, including Quranic quotations
and rather standardised formulae. It appears that a linguistic specialisation in
monumental epigraphy – titulatures, foundation and religious inscriptions in
Arabic versus poetic inscriptions in Persian – was established by the late elev-
enth century, which reflects the bilingual cultural environment of the eastern
Islamic courts. The visibility given to poetic inscriptions in royal buildings also
testifies to the important role of Persian poetry in the legitimisation strategies of
the Turko-Iranian dynasties and raises the question of a possible link with per-
formative practices. Indeed, the narrative sources describe court ceremonies in
which poets recited panegyric poems, competing one against the other to deliv-
er the most pleasing praise of the ruler and obtain the greatest reward.34
From the palaeographic point of view, most Persian inscriptions from the
eleventh and early twelfth century are written in Kufic letters,35 although the
writing style and ornaments vary from text to text. The lack of diacritical points
makes the reading of some Persian passages ambiguous, and although the in-
scriptions were well visible to the observer, their actual legibility is an open

||
33 Allegranzi 2019, vol. 1, 193–206.
34 Bosworth 1963, 129–139.
35 Some fragments of Persian inscriptions in curvilinear letters come from Ghazni (Allegranzi 2019,
vol. 1, 195–200), where curvilinear script seems to have spread in epigraphy in the first half of
the eleventh century, i.e. earlier than at any other Islamic site known to us, see Giunta 2001;
Giunta 2003, 431.
358 | Viola Allegranzi

issue.36 As a rule, inscriptions in various writing styles (in Kufic or curvilinear

script, with different degrees and types of ornamentation) are placed side by
side in the epigraphic programme of medieval Iranian monuments. While a
fixed relationship between the use of language and the content of the text is
generally observed, there is no established correspondence between the lan-
guage and the writing style, the choice of which seems determined more by
artistic trends and the visual harmony of the architectural decoration.
The tradition of inscribing Persian verses on monuments, juxtaposing them
with Arabic inscriptions of different content, was destined to continue over
time. A Persian poem inscribed in curvilinear letters is partially preserved on
the arch framing the portal of the southern mausoleum in Uzgend (Kyrgyzstan,
582 AH / 1186–1187 CE), the capital of the Qarakhanid line of Ferghana.37 Fur-
thermore, fragments of curvilinear inscriptions in Persian, probably versified,
have emerged among the wall paintings of a Qarakhanid royal pavilion exca-
vated in Samarqand and attributed to the late twelfth or early thirteenth centu-
ry.38 In both contexts, several Arabic inscriptions varying in length, content and
style were also recorded.
Persian poetic inscriptions are increasingly found on elements of architec-
tural decoration and inscribed objects from the early thirteenth century on-
wards, and this epigraphic tradition bears witness to the continuity of artistic
practices between the pre-Mongol and Mongol periods. The large corpora of
lustre-painted ceramic vessels and tiles produced in north-western Iran between
the late twelfth and the fourteenth century are a case in point. Persian love
quatrains, excerpts from Firdawsī’s Shāhnāma (‘Book of Kings’) and well-
wishing verses are frequently inscribed on them, alternating with Quranic pas-
sages, hadiths and benedictory inscriptions in Arabic.39 The inscriptions are
painted on lustre-painted ceramics in curvilinear letters inspired by handwrit-
ing and they probably addressed a larger audience than the royal court. Indeed,
while some of the lustre-painted tiles originally decorated state-sponsored

||
36 The progressive replacement of Kufic with curvilinear script in twelfth-century inscriptions
did not solve the issue of ambiguity, since Persian inscriptions continued to be written without
diacritics and using some confusable letters, see Allegranzi 2018, 106; Allegranzi 2019, vol. 1,
241–242.
37 Nastič 2019, 99–100; see also Allegranzi 2019, vol. 1, 215–216; McClary 2020, 88–96.
38 Karev 2005, 69; see also Allegranzi 2019, vol. 1, 216–217.
39 Qūchānī (1371 AH / 1992 CE); O’Kane 2006. Several scholars have contributed to the study of
inscriptions on lustre-painted ceramics, including Mehdi Bahrami, Leon T. Gjuzal’jan, ʿAbdullah
Gouchani, Assadullah S. Melikian-Chirvani and Manijeh Bayani; nevertheless, numerous
inscriptions on these materials are still unpublished.
 Multilingualism in the Epigraphic Culture of the Persianate World | 359

buildings (such as the Ilkhanid palace at Takht-i Sulaymān), other tiles and
most vessels were probably destined for members of the upper classes. This
gives us reason to assume that their epigraphic repertoire mirrors the literary
taste of educated people in the premodern Persian-speaking societies.40

4 Mixing Persian and Arabic

4.1 The inclusion of Persian in construction texts
As mentioned above, a standardised Arabic formulary prevails in the construc-
tion texts of medieval Iranian monuments. However, in addition to the Zalam-
kot inscription (see above, Subsection 2.2), there are other exceptions of pre-
thirteenth-century construction inscriptions partially composed in Persian. A
foundation text that mixes Persian and Arabic is inscribed on the entrance arch
of the northern mausoleum in the Qarakhanid site of Uzgend; it gives the start-
ing date of construction as the fourth of rabīʿ II 547 AH / 9 July 1152 CE.41 The main
text is in Persian and some Arabic words inserted in the date are Persianized
(rabīʿ al-ākhir > rabīʿ-i ākhir; hijra > hijrat). However, the verb is placed at the
beginning of the text, as in standard Arabic texts, and the last third of the in-
scription contains a benediction of the prophet Muḥammad and his family and
a doxology in Arabic. The text is sculpted on terracotta tiles and features curvi-
linear letters on a background of vegetal scrolls. An inscription that differs in
style and content runs inside the entrance vault of the same building: it is exe-
cuted in bordered Kufic and records the titulature of the alleged patron, the
Qarakhanid Jalāl al-Dīn Ḥusayn (r. c. 531–551 AH / 1137–1156 CE), mixing Arabo-
Islamic and Turkic titles and names.42
The Ghurids, successors to the Ghaznavids in eastern Khurasan and north-
ern India, do not seem to have continued the tradition of adding Persian poetic
inscriptions to monuments, at least as far as one can judge from the architectur-

||
40 By contrast, the largest group of inscribed vessels produced in Khurasan and Transoxiana
in earlier times, namely Samanid epigraphic pottery (probably ninth to eleventh century),
exclusively contained Arabic inscriptions (benedictory formulae, proverbs, hadiths). The lan-
guage and the ornated Kufic styles employed in this production probably made the inscriptions
challenging for Persian-speaking consumers to understand. Ghouchani 1364 AH / 1986 CE;
Pancaroğlu 2002; Hillenbrand 2015.
41 Nastič 2019, 82–83; McClary 2020, 74–76.
42 Nastič 2019, 83–87; McClary 2020, 77–78. See also Peacock 2023, 291–295.
360 | Viola Allegranzi

al complexes patronised by this dynasty that have come down to us.43 Mention
should be made, however, of a passage of the inscriptions in sculpted stucco
adorning the interior of a Ghurid domed structure in Chisht-i Sharif (Hari Rud
valley, Afghanistan), recording in Persian the date of the tenth of jumāda I
562 AH / 4 March 1167 CE. Similar to the inscription in the northern mausoleum at
Uzgend, the Arab name of the month is altered (jumāda al-awwal > jumīd al-
awwal) and a benediction of the prophet in Arabic follows the date.44 Interest-
ingly, the Persian section occurs at the end of a long inscription in Kufic letters
containing several Quranic quotations (Quran 3:18–19; 2:255–257; 112). The
renovation text and the titulature of the Ghurid ruler Shams al-Dīn (later known
as Ghiyāth al-Dīn) Muḥammad b. Sām (r. 558–599 AH / 1163–1203 CE), by con-
trast, appear in a separate epigraphic register, more easily visible to the visitor,
composed in Arabic and an ornamented curvilinear script.
A final example of the oscillation between the adherence to the Arabic
model and the experimentation of new formulae in Persian is that of the inscrip-
tions referring to the foundation of the Quwwat al-Islām mosque (592–594 AH /
1195–1197 CE) in the Quṭb Minār complex in Delhi. Two foundation inscriptions
in Arabic, located on the northern and eastern portal, respectively, mention the
Ghurid ruler Muʿizz al-Dīn Muḥammad b. Sām (r. 569–602 AH / 1173–1206 CE)
and his military commander Quṭb al-Dīn Aybak (d. 607 AH / 1210 CE), responsible
for the Ghurid expansion in India and actual patron of the building.45 Two more
inscriptions adorning the eastern entrance mention Quṭb al-Dīn Aybak, his
victory over the infidels and role as the founder of the mosque: they are com-
posed in Persian, in two different varieties of curvilinear script.46 The intricate
writing style suggests that the longer text, located on the inner lintel, is a later
addition to the original epigraphic programme, but the shorter text, in the lu-

||
43 The Ghurids (or Shansabanis) were a Muslim confederation originating from central Af-
ghanistan, who conquered and ruled over a vast area stretching from eastern Iran to northern
India between the mid-twelfth and the early thirteenth century, see O’Neal 2015. On Ghurid
architecture in Afghanistan and Pakistan, see Patel 2021 (with an Appendix on monumental
inscriptions: pp. 342–356).
44 Blair 1985, 81–82, pl. 12; Giunta 2010, 177. Earlier inscriptions attest to the insertion of some
Persian words in a date, see, for instance, a Buyid text incised in Persepolis (Iran, 438 AH /
1046 CE), Blair 1992, 118–120, no. 43.
45 Horovitz 1914, 14–15, nos 3, 5; Blayac 2013, 218–219, fig. 12.2.
46 Horovitz 1914, 13–14, nos 1, 2; Blayac 2013, 221, figs 12.4 and 12.5. Other royal inscriptions in
Arabic mentioning the Ghurid sultans and Quṭb al-Dīn Aybak appear on the lower storey of the
adjoining minaret, together with a highly fragmentary inscription in Kufic letters, probably in
Persian (Horovitz 1914, 19, no. 12).
 Multilingualism in the Epigraphic Culture of the Persianate World | 361

nette above the gateway, is probably contemporary with the Arabic foundation
texts (late twelfth century).
This review has shown that the use of Persian in pre-Mongol monumental
epigraphy is restricted to scattered sites. Besides the evidence mentioned above,
a few other examples are related to late twelfth-century Islamic buildings in the
Caucasus.47 Although some inscriptions composed in a mixture of Persian and
Arabic are recorded on monuments dating back to the second half of the twelfth
century, it seems that construction texts did not cease to refer to the Arabo-
Islamic model, and there was no consistent development of an epigraphic for-
mulary in Persian at that time.

4.2 The uses of Persian on inscribed objects

Following its introduction in monumental inscriptions, Persian epigraphy be-
gan to be sporadically used on luxury objects, as evidenced by a few pieces of
metalwork produced between the mid-twelfth and the early thirteenth century.48
These are four dated and signed objects made of copper alloy or brass, decorat-
ed and inscribed with silver and copper inlays. The earliest artefact is a rectan-
gular pen-box dated 542 AH / 1148 CE, kept at The State Hermitage Museum (inv.
no. SA-12688).49 Several inscriptions decorate the object, among which is a text
in curvilinear script included in cartouches on both long sides, containing a
quatrain in Persian followed by the name ʿUmar b. al-Faḍl b. […].50 The full
name occurs in a passage of the Arabic dedicatory inscription on the upper face:
ʿamal ʿumar b. al-faḍl b. yūsuf al-bayyāʿ (‘work of ʿUmar son of al-Faḍl son of
Yūsuf, the seller’).51 The epithet al-bayyāʿ (‘the seller’) suggests that the signa-
ture does not refer to the maker or decorator of the pen-box, as one would ex-

||
47 O’Kane 2009, 33, 59; Allegranzi 2019, vol. 1, 218–220. Furthermore, a fragmentary founda-
tion text in Persian, probably dating back to the mid-thirteenth century, has recently been
identified in the Great Mosque of Faryumad in north-eastern Iran (Aghajani forthcoming, no. 46).
48 A silver bowl inscribed with Persian verses has been attributed to the Samanid period (late
ninth–early tenth century); however, O’Kane (2009, 11–13) questioned this early dating.
49 See an in-depth analysis of the object and its inscriptions in Giuzalian 1968. The inscrip-
tions are also recorded in TEI, no. 16862.
50 Giuzalian 1968, 115–116, figs 1, 9. This author translated the poem as follows: ‘I was satis-
fied with Izad’s pre-determination, I cut off the hope of all creatures, Izad regarded me in
favour while I lived, He will favour me in the future until I die’. Note the use of the Persian
word īzad for ‘God’ instead of the Arabic allāh, more widespread at that time.
51 Giuzalian 1968, 98–99 and pl. 1, fig. 2. The inscription is unconventionally designed in a
mixture of Kufic and curvilinear script.
362 | Viola Allegranzi

pect in an inscription that begins with ʿamal ‘work of’, but to the person respon-
sible for its sale.52 It is unusual for a signature to be repeated twice on the same
object and the implications of the mention following the Persian poem are un-
certain: it is unlikely that the signatory was the author of the pietistic verses,
while it is more than plausible that their function was to draw divine favour on
him.53
A longer Persian poem appears on the ribbed body of a ewer produced in
Herat in 577 AH / 1181–1182 CE and held by the Simon Janashia Museum of Geor-
gia in Tbilisi (inv. no. MS 135).54 The verses praise the extraordinary beauty of
the ewer (āftābah) and mention its place of production (Herat) and its maker,
upon whom they invoke heavenly mercy, generous reward and good fortune.
One line attached to the versified inscription contains the signature – al-ʿamal
al-naqsh maḥmūd b. muḥammad al-harawī (‘work [and] design (?) of Maḥmūd
son of Muḥammad al-Harawī’) – and the date. This section is composed in Ara-
bic but has some inaccuracies: the article (al-) preceding the word ʿamal is not
justified by the syntax, since ʿamal is usually employed as the first term of an
iḍāfa construction (construct case) and, thus, without the definitive article. This
also applies to the following al-naqsh (‘the painting’ or ‘the design’), which
should additionally be preceded by the conjunction wa ‘and’. The use of al-
naqsh is in itself unusual, as it replaces the agent noun derived from the same
root, al-naqqāsh (‘the painter’ or ‘the designer’), commonly used in inscriptions
on various media to designate the artisan responsible for the decoration. The
signature probably refers to the inlayer and, as in the case of the pen-box dis-
cussed above, the question can be asked whether the signatory was also the
author of the poem, which explicitly demands for recognition and reward for
the artisanal work. Doubts have been raised about the authenticity of the Per-
sian inscription on the ewer, due to the inconsistency in the spelling of the rela-
tive pronoun, alternatively transcribed kī or kih (the latter being a later form,
rarely attested at that time). If we accept a dating from the late twelfth century,
we must admit that the inscription represents a diversion from the established

||
52 This was already the hypothesis of Giuzalian (1968, 99–102); note, however, that profes-
sional epithets and nisbas did not necessarily reflect the actual social role of a named person
and should be treated with caution. On the varied roles of signatories of Iranian metalwork and
on the multiple meanings of the introductory formula ʿamal, see Allegranzi forthcoming.
53 Two more Persian poems incised on the lower face of the pen-box were probably added
later (Giuzalian 1968, 116–117). These verses personify the pen and inkwell (qalam and dawāt),
thus, allegorically referring to the function of the artefact.
54 Loukonine and Ivanov 1996, 136–137, no. 117; Canby et al. 2016, 155–156, no. 85.
 Multilingualism in the Epigraphic Culture of the Persianate World | 363

models, since the scribe seems to be unfamiliar with the standard Arabic formu-
lary and more acquainted with Persian language.
The last two inscribed objects discussed, both held by The State Hermitage
Museum (inv. nos IR-2268 and AZ-225), provide examples of fabrication texts
composed in Persian. The first is a bucket of globular shape produced in Herat
in 559 AH / 1163 CE, the body of which is finely decorated with figurative scenes
and Arabic benedictory inscriptions in ornate styles.55 The fabrication text, in-
scribed in curvilinear letters on the upper side of the rim, is in Persian. Howev-
er, as in some examples discussed above (see Subsections 2.2 and 4.1), the syn-
tax follows the verb-object-subject word order and the phrase farmūdan-i īn
khidmat rā ‘ordered this service (or: gift)’ (literally ‘the ordering of this service’)
introduces the names of three individuals involved in the production process,
followed by the mention of the owner.56 All the people mentioned bear Arabic
names and titles, and the owner’s name is followed by a benedictory phrase in
Arabic, which shows the strong influence of the Arabic model.
We finally focus attention on a zoomorphic aquamanile dated 603 AH /
1206 CE, featuring a zebu suckling her calf and a lion-shaped handle, which
originally had movable pieces (the zebu’s bell, harness and tail) activated by an
internal mechanism (Fig. 3).57 The artefact is unique in its design, and the fabri-
cation inscription, traced on the zebu’s neck, is equally unconventional.58 The
first part of the text is in Persian: it unusually describes the object and its fabri-
cation process (īn gāv va gūsāla va shīr har si yak bāra rīkhta shud-ast ‘This cow,
calf and lion were all cast at the same time’), and invokes God through a mag-
niloquent Persian phrase (yāzdān-i dādgar parvardgār ‘God, the all-just judge
and the nourisher’). Conversely, the second part, giving the name of the owner
and the decorator, adheres to the Arabic syntax and formulary. However, both
the patron (Rūzba b. Afrīdūn [b.] Barzīn) and the owner (Shāh Barzīn b. Afrīdūn
b. [Bar]zīn) bear Persian names and seem to be members of the same family,
probably belonging to the Iranian elites. The context of reception could, there-
fore, justify the choice of composing ad hoc a ‘Persianized’ inscription for the
precious automaton.

||
55 Ettinghausen 1943. See photos at <https://www.kornbluthphoto.com/BobrinskyBucket.html>
(accessed on 5 May 2023).
56 Ettinghausen 1943, 196, 198, fig. 3; Giuzalian 1968, 105. See also TEI, no. 7979 and a full
transliteration and translation in Allegranzi forthcoming. I thank Prof. Jürgen Paul who point-
ed out that the term khidmat could be used in this context with the meaning of ‘present, gift’.
57 Giuzalian 1968, 103–109; Loukonine and Ivanov 1996, 144, no. 127.
58 Giuzalian 1968, 103–104. See also TEI, no. 2890 and a full transliteration and translation in
Allegranzi forthcoming.
364 | Viola Allegranzi

Fig. 3: Zoomorphic aquamanile (probably from Khurasan, 603 AH / 1206 CE, H 35 cm). The State
Hermitage Museum, St Petersburg, inv. no. AZ-225. Photograph © The State Hermitage Muse-
um / photo by Vladimir Terebenin.

The Persian inscriptions on this small group of objects reflect trends already
observed in monumental epigraphy. On the one hand, Persian verses are juxta-
posed to Arabic inscriptions of different natures (cf. the pen-box and the ewer);
on the other hand, Arabic syntax and vocabulary exert a strong influence on
fabrication texts, composed in a mixture of Persian and Arabic (cf. the bucket
and the aquamanile). The four inlaid pieces of metalwork are all extremely val-
uable and the explicit mentions of their patron and/or owner indicate that they
 Multilingualism in the Epigraphic Culture of the Persianate World | 365

were the result of special commissions. The choice to deviate from the standard
model and use Persian in the epigraphic decoration could respond to a specific
request of the patron or represent an adaptation to the context of reception and
consumption. It is good to remember that these pieces constitute an exception
and that the great majority of medieval Iranian metalwork, presumably pro-
duced to be sold on the market, display standard benedictory inscriptions in
Arabic, bestowing protection to an anonymous owner.

Conclusion
Based on surviving evidence, the Persian language has been progressively in-
troduced into the epigraphic culture of the eastern Islamic regions from the
eleventh century onwards. In this period and up to the thirteenth century, in-
scriptions that deviate from the Arabo-Islamic model are few and geographical-
ly dispersed. Although it is likely that some vestiges of this probably wider phe-
nomenon have been lost, it can be assumed that there was still no reference
model for Persian inscriptions, as linguistic and stylistic solutions vary from
case to case. Furthermore, at this epoch, Persian is never completely independ-
ent in the epigraphic decoration of monuments and objects, but always alter-
nated or mixed with Arabic.59 This is made possible by the fact that New Persian
shares the same alphabet with Arabic. In fact, multiscriptism is observed in
some eleventh-century bilingual inscriptions, but they constitute isolated at-
tempts, produced in peripheral areas of the Iranian region.
The contents of the earliest inscriptions in New Persian allow us to define
two main trends. On the one hand, Persian verses are inscribed on royal monu-
ments (especially palaces and mausoleums), juxtaposed to Arabic inscriptions
of different contents. The prominent position and panegyric tone of these in-
scriptions reflect the importance of Persian poetry in the political discourse and
court culture of the Turko-Iranian dynasties. The use of Persian poetic inscrip-
tions on monuments and objects is destined to continue and spread in the Ira-
nian world, in association with both royal and private patronage, through the
premodern and subsequent periods. On the other hand, Persian is sometimes
used in construction texts, which, however, reveal a greater influence of the

||
59 The present study did not consider funerary inscriptions, which reveal a similar tendency, that
is, a general adherence to the Arabo-Islamic formulary, rarely supplemented with Persian expres-
sions (especially verses or invocations). See Giunta 2010; Allegranzi 2018; Babadjanov et al. 2019,
258–259, 262–263, 286–287, 304–305, 308–309, 446, nos Q-095, Q-097, Q-108, Q-117, Q-119, Q-180.
366 | Viola Allegranzi

Arabic models. In particular, the verb usually occupies the first position, in
accordance with Arabic verb-initial syntax (thus, contradicting the Persian sub-
ject-object-verb word order); moreover, Arabic words or phrases (especially
benedictory formulae and doxologies) often complement the inscriptions. As a
general rule, Arabic remains the standard language for construction texts in
eastern Islamic epigraphy. The reasons behind the unconventional choice to
compose a construction text in Persian (or rather, in a mixture of Persian and
Arabic) are sometimes difficult to establish with certainty, but they seem to be
mostly related to the particularity of the context of production and/or reception.
From a stylistic point of view, the Persian inscriptions attributed to the
eleventh century were mainly composed in Kufic script without diacritical
marks, in more or less ornate styles depending on the visual effect sought and,
perhaps, on the medium. From the twelfth century onwards, the use of curvilin-
ear script (generally not diacritised) took root. However, this shift is not limited
to Persian inscriptions, but reflects a general trend in the artistic production of
the eastern caliphate, where epigraphic cursive became widespread much earli-
er than in the western regions. Thus, there does not appear to be a preferential
association between the language of composition and the writing style. Howev-
er, multilingualism and the development of new writing styles and ornaments
are, in effect, the most original features of Islamic epigraphy in the Persianate
world.

Abbreviations
EI³ = Kate Fleet, Gudrun Krämer, Denis Matringe, John Nawas and Devin J. Stewart (eds), Ency-
clopaedia of Islam, THREE, Leiden: Brill, 2021 (online).
TEI = Ludvik Kalus, Frédéric Bauden and Frédérique Soudan, Thesaurus d’Épigraphie Islamique
<www.epigraphie-islamique.uliege.be> (accessed on 5 May 2023).

References
Aghajani, Iman (forthcoming), ‘The Friday Mosque of Faryūmad and Its Inscriptions’, Iran:
Journal of the British Institute of Persian Studies.
Allegranzi, Viola (2017), Les inscriptions persanes de Ghazni, Afghanistan: Nouvelles sources
pour l’étude de l’histoire culturelle et de la tradition épigraphique ghaznavides (Ve-VIe /
XIe-XIIe siècles), 2 vols, PhD thesis, Sorbonne Nouvelle-Paris 3 and Università degli studi
di Napoli ‘L’Orientale’, available at <https://shs.hal.science/tel-01720484> (accessed on
5 May 2023).
 Multilingualism in the Epigraphic Culture of the Persianate World | 367

Allegranzi, Viola (2018), ‘Les inscriptions persanes du tombeau d’Abū Jaʿfar Moḥammad,
Ghazni (Afghanistan), début du VIe / XIIe siècle. La poésie comme lieu de mémoire et
d’histoire’, Studia Iranica, 47/1: 89–118.
Allegranzi, Viola (2019), Aux sources de la poésie ghaznavide: Les inscriptions persanes de
Ghazni (Afghanistan, XIe–XIIe s.), 2 vols, Paris: Presses Sorbonne Nouvelle.
Allegranzi, Viola (forthcoming), ‘Noteworthy Signatures on Medieval Iranian Metalwork: Reflec-
tions on Artisans and the Organisation of the Art’, in Rebecca Müller, Nikolaus Dietrich
and Mandy Telle (eds), Proceedings of the conference Signaturentragende Artefakte:
Schriften, Materialien, Praktiken im transkulturellen Vergleich, Heidelberg, 11–12 March
2022 (KEMTE: Kulturelles Erbe: Materialität – Text – Edition), Heidelberg: Heidelberg Uni-
versity Publishing.
Amanat, Abbas and Assef Ashraf (eds) (2019), The Persianate World: Rethinking a Shared
Sphere (Iran Studies, 18), Leiden: Brill.
Babadjanov, Bakhtiyar, Lola Dodkhudoeva, Ashirbek Muminov and Ulrich Rudolph (eds) (2019),
Epitafii musul’maniskix učenyx Samarkanda (X–XIV veka) / Epitaphs of Muslim Scholars in
Samarkand (10th–14th centuries), Istanbul: IRCICA.
Bāghbīdī, Ḥasan R. (1383 SH / 2004 CE), ‘Katība-yi pahlavī-kūfī-yi burj-i Lājīm’, Nāma-yi Irān-i
bāstān, 4/1: 9–21.
Blair, Sheila S. (1985), ‘The Madrasa at Zuzan: Islamic Architecture in Eastern Iran on the Eve of
the Mongol Invasions’, Muqarnas, 3: 75–91.
Blair, Sheila S. (1992), The Monumental Inscriptions from Early Islamic Iran and Transoxiana
(Muqarnas, Supplements, 5), Leiden: Brill.
Blayac, Johanna (2013), ‘Sovereign Epigraphy in Location. Politics, Devotion and Legitimization
around the Quṭb Minār, Delhi’, in Mohammad Gharipour and İrvin C. Shick (eds), Calligra-
phy and Architecture in the Muslim World, Edinburgh: Edinburgh University Press, 217–229.
Bombaci, Alessio (1966), The Kūfic Inscription in Persian Verses in the Court of the Royal Palace
of Masʿūd III at Ghazni (Reports and memoirs, Istituto Italiano per il Medio ed Estremo Ori-
ente, Centro studi e scavi archeologici in Asia, 5), Rome: Is.M.E.O.
Bosworth, Clifford E. (1963), The Ghaznavids: Their Empire in Afghanistan and Eastern Iran,
994–1040, Edinburgh: Edinburgh University Press.
Bosworth, Clifford E. (1978–1979), ‘The Interaction of Arabic and Persian Literature and Culture
in the 10th and Early 11th Centuries’, al-Abhath, 27: 59–75.
Canby, Sheila R., Deniz Beyazit, Martina Rugiadi and Andrew C. S. Peacock (2016), Court and
Cosmos: The Great Age of the Seljuqs, New York: The Metropolitan Museum of Art.
Cereti, Carlo G. (2015), ‘Lingue e linguaggi nell’Iran altomedievale’, in Maria C. Benvenuto and
Paolo Martino (eds), Linguaggi per un nuovo umanesimo, Vatican City: Libreria Editrice
Vaticana, 153–296.
de Blois, François (1990), ‘The Middle-Persian Inscriptions from Constantinople: Sasanian or
Post-Sasanian?’, Studia Iranica, 19: 209–218.
Engeskaug, Aleksander (2020), ‘Quantifying Middle Persian inscriptions: A New Approach to
the Epigraphic Culture of Sasanian Iran’, in Ingvar B. Mæhler, Per B. Ravnå and Eivind H.
Seland (eds), Methods and Models in Ancient History: Essays in Honor of Jørgen Christian
Meyer, Athens: Norwegian Institute, 173–202.
Ettinghausen, Richard (1943), ‘The Bobrinsky Kettle, Patron and Style of an Islamic Bronze’,
Gazette des Beaux Arts, 24: 193–208.
Flood, Finbarr B. (2009), Objects of Translation: Material Culture and Medieval ‘Hindu-Muslim’
Encounter, Princeton: Princeton University Press.
368 | Viola Allegranzi

Flury, Samuel (1921), ‘Bandeaux ornementés à inscriptions arabes: amida-dyarbekr, XIe siècle
(Troisième article)’, Syria, 2/1: 54–62.
Flury, Samuel (1925), ‘Le décor épigraphique des monuments de Ghazna’, Syria, 6/1: 61–90.
Fragner, Bert G. (1999), Die ‘Persophonie’: Regionalität, Identität und Sprachkontakt in der
Geschichte Asiens (ANOR, 5), Berlin: Das Arab.
Giunta, Roberta (2001), ‘The Tomb of Muḥammad al-Harawī (447/1055) at Ġaznī (Afghanistan)
and Some New Observations on the Tomb of Maḥmūd the Ġaznavid’, East and West,
51/1–2: 109–126.
Giunta, Roberta (2003), Les inscriptions funéraires de Ġaznī (IVe-IXe / Xe-XVe siècles) (Series
Maior, 8), Naples: Università degli studi di Napoli ‘L’Orientale’ / IsIAO / Fondation Max
van Berchem.
Giunta, Roberta (2010), ‘Les inscriptions persanes dans l’épigraphie monumentale de la ville
de Ghazni (Afghanistan) aux 6e–7e / 12e–13e siècles’, in Carlo G. Cereti (ed.), Iranian
Identity in the Course of History: Proceedings of the Conference Held in Rome, 21–24 Sep-
tember 2005 (Orientalia Romana, 9), Rome: IsIAO, 163–180.
Giunta, Roberta and Viola Allegranzi (2020), ‘Ghaznavid, Qarakhanid and Seljuq Monumental
Inscriptions and the Development of Royal Propaganda’, in Sheila Canby, Deniz Beyazit
and Martina Rugiadi (eds), The Seljuqs and Their Successors: Art, Culture and History, Ed-
inburgh: Edinburgh University Press, 113–128.
Giuzalian, Leon T. (1968), ‘The Bronze Qalamdan (Pen-Case) 542 / 1148 from the Hermitage
Collection (1936-1965)’, Ars Orientalis, 7: 95–119.
Ghouchani, Abdullah (1364 AH / 1986 CE), Inscriptions on Nishabur Pottery / Katibahā-yi sufāl-i
Nayshābūr, Tehran: Reza Abbasi Museum.
Hillenbrand, Robert (2015), ‘Content versus Context in Samanid Epigraphic Pottery’, in Andrew
C. S. Peacock and Deborah G. Tor (eds), Medieval Central Asia and the Persianate World:
Iranian Tradition and Islamic Civilisation, London: I. B. Tauris, 56–107.
Horovitz, Josef (1914), ‘The Inscriptions of Muḥammad Ibn Sām, Quṭbuddīn Aibeg and Iltut-
mish’, Epigraphia Indo-Moslemica, 1911–1912: 12–34.
Karev, Yuri (2005), ‘Qarakhanid Wall Paintings in the Citadel of Samarqand: First Report and
Preliminary Observations’, Muqarnas, 22: 46–84.
Karev, Yuri (2013), ‘From Tents to City. The Royal Court of the Western Qarakhanids between
Bukhara and Samarqand’, in David Durand-Guédy (ed.), Turko-Mongol Rulers, Cities and
City Life, Leiden: Brill, 99–147.
Lazard, Gilbert (1963), La langue des plus anciens monuments de la prose persane, Paris: C.
Klincksieck.
Lazard, Gilbert (1964), Les premiers poètes persans (IXe-Xe siècles): Fragments rassemblés,
édités et traduits, 2 vols, Tehran: Institut franco-iranien / Paris: Librairie d’Amérique et
d’Orient.
Lazard, Gilbert (1995), La formation de la langue persane, Paris: Institut d’études iraniennes.
Loukonine, Vladimir and Anatoli Ivanov (eds) (1996), Lost Treasures of Persia: Persian Art in the
Hermitage Museum, Washington, DC: Mage.
Madelung, Wilferd (1984), ‘ĀL-E BAVAND’, in Encyclopaedia Iranica, online edition,
<https://www.iranicaonline.org/articles/al-e-bavand> (accessed on 5 May 2023).
McClary, Richard P. (2020), Medieval Monuments of Central Asia: Qarakhanid Architecture of
the 11th and 12th Centuries (Edinburgh Studies in Islamic Art, 1), Edinburgh: Edinburgh
University Press.
 Multilingualism in the Epigraphic Culture of the Persianate World | 369

Nastič, Vladimir N. (2019), Musul’manskaja epigrafika Fergany i Semireč’ja: Pamiatniki ar-

abskogo pis’ma XI–XVII vv. na territorii Kyrgyzstana / Islamic Inscriptions of Ferghana and
Semirechye: Arabic-written Monuments of 11–17th Centuries from Kyrgyzstan, St Peters-
burg: The Herzen State Pedagogical University.
O’Kane, Bernard (2006), ‘Persian Poetry on Ilkhanid Art and Architecture’, in Linda Komaroff (ed.),
Beyond the Legacy of Genghis Khan, Leiden: Brill, 346–354.
O’Kane, Bernard (2009), The Appearance of Persian on Islamic Art, New York: Persian Heritage
Foundation.
O’Neal, Michael (2015), ‘Ghūrids’, in EI³, <http://dx.doi.org/10.1163/1573-3912_ei3_COM_27477>,
first published online 2015 (accessed on 12 July 2023).
Pancaroğlu, Oya (2002), ‘Serving Wisdom: The Contents of Samanid Epigraphic Pottery’, in
Rochelle L. Kessler, Mary Anderson McWilliams, Oya Pancaroglu and David J. Roxburgh (eds),
Studies in Islamic and Later Indian Art, Cambridge, MA: Harvard University Art Museum,
59–75.
Patel, Alka (2021), Iran to India: The Shansabanis of Afghanistan, c. 1145–1190 CE, Edinburgh:
Edinburgh University Press.
Paul, Jürgen (2021), ‘Karakhanids’, in EI³, <http://dx.doi.org/10.1163/1573-3912_ei3_COM_30250>,
first published online 2021 (accessed on 12 July 2023).
Paul, Ludwig (2000), ‘Persian Language, I: Early New Persian’, in Encyclopaedia Iranica, online
edition, <https://iranicaonline.org/articles/persian-language-1-early-new-persian> (ac-
cessed on 5 May 2023).
Peacock, Andrew C. S. (2023), ‘The Rise of New Epigraphic Languages in the Medieval Islamic
East: The Interplay of Persian, Turkish and Arabic on Inscriptions’, in Bernard O’Kane, An-
drew C. S. Peacock and Mark Muehlhaeusler (eds), Inscriptions of the Medieval Islamic
World, Edinburgh: Edinburgh University Press, 283–323.
Perry, John R. (2009), ‘The Origin and Development of Literary Persian’, in Johannes T. P. de
Bruijn (ed.), A History of Persian Literature, vol. 1: General Introduction to Persian Litera-
ture, London: I. B. Tauris, 43–70.
Qūchānī, ʿAbdullāh (1371 AH / 1992 CE), Ashʿār-i fārsī-i kāšīhā-yi Takht-i Sulaymān, Tehran:
Markaz-i nashr-i dānishgāhī.
Rahman, Abdur (1998), ‘The Zalamkot Bilingual Inscription’, East and West, 48/3–4: 469–473.
Shavarebi, Ehsan (2022), with a contribution by Ingo Strauch, ‘The Mosque of the Forgotten City:
The Bilingual Inscription of Zalamkot Revisited’, East and West, n.s. 3 [62] / 2: 189–204.
Sims-Williams, Nicholas and François de Blois (2018), with contributions by Harry Falk and
Dieter Weber, Studies in the Chronology of the Bactrian Documents from Northern Afghan-
istan, Vienna: Verlag der Österreichischen Akademie der Wissenschaften.
Vásáry, István (2015), ‘Two Patterns of Acculturation to Islam: The Qarakhanids versus the
Ghaznavids and Seljuqs’, in Edmund Herzig and Sarah Steward (eds), The Age of the Sel-
juks (The Idea of Iran, 6), London: I. B. Tauris, 9–28.
Eva Orthmann
Bilingual Inscriptions from India: Combining
Arabic and Persian with Indic Languages
Abstract: Multilingualism and multiscriptuality are omnipresent in today’s
India: different languages in various scripts are found on street signs and bank-
notes, in metro stations, public buildings and many other places.1 However,
despite the omnipresence of contemporary multilingual texts, examples from
premodern India are much less frequent. The following article will shed light on
one specific type of multilingual text from the premodern and early modern
period: epigraphic documents. There are multiple combinations of languages
and scripts in Indian epigraphy, but this article will focus only on combinations
of Indic languages and scripts with texts in Arabic or Persian.

1 Introduction
Arabic and Persian inscriptions are an important source of information on Mus-
lim rule in India. They are found on epitaphs, wells, walls, tablets, pillars and
copper plates, in mosques and other places. The vast majority of the several
thousand inscriptions so far discovered and published are monolingual or in a
mixture of Persian and Arabic. The latter combination will not be considered as
bilingual or multilingual, since the insertion of Arabic expressions into Persian
is very frequent.2 Rather, the bilingual texts studied in this article are combina-
tions of Arabic or Persian with Indic languages, mainly Sanskrit. In a prelimi-
nary survey of the inscriptions published in the Epigraphia Indo-Moslemica and
its continuation, Epigraphia Indica: Arabic and Persian Supplement, from 1907
to 2011, sixty-three bilingual and three trilingual inscriptions have been record-
ed. Another bilingual and one more trilingual inscription are included in the

||
1 For good examples, see e.g. the images in Brandt 2014.
2 I am not referring here to the many Arabic loanwords in Persian, but to Arabic phrases – in
many cases from the Quran or Ḥadīth (records of the words and deeds of Muḥammad) – that
have been included in Persian texts; Arabic eulogies or numbers are also common. In some
cases, we even find Arabic poetry. See also Sheikh 2014, 189 for the same approach.

Open Access. © 2024 the author, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-014
372 | Eva Orthmann

Siddham database,3 and eight further inscriptions are contained in other publi-
cations. The total number of inscriptions considered here is thus seventy-six.
Most of the examples I will present contain texts in two different scripts and
two different languages. In spite of the use of two different scripts, these texts
are not biscriptal, but bilingual, as only monolingual texts in two different
scripts are biscriptal by definition;4 the only texts in this category are direction
stones with Persian in Arabic and Nāgarī characters.5 In the inscriptions consid-
ered below, one script is always Arabic, the other an Indian script.
Unfortunately, not all of these inscriptions have been fully edited, and in
quite a few cases, we either lack photos of part of the inscription, or the quality
of the reproduction is very poor. In many cases, the texts of the inscriptions
have been published separately by language, with the Persian or Arabic text in
one journal, the Sanskrit text in another. Because only rarely, all the necessary
linguistic competences are found with one scholar, the texts have also been
studied by different scholars,6 and in most cases, the relationship between the
two texts of the same bilingual inscription has not been examined. Two recent
studies however address this question: while Soroni’s study examines the cor-
pus of inscriptions in Persian and Marathi,7 Sheikh looks at the corpus of bilin-
gual inscriptions from Gujarat.8
In my article, I attempt to give an overview of the entire corpus of bilingual
inscriptions that join Arabic/Persian with an Indic language from the subconti-
nent. The goal is to formulate a broader idea of how these languages have been
combined, what visual effects were produced, and how the hierarchy between
languages was expressed. The comparison will show that in spite of the vast
geographical and historical provenance of the inscriptions, their visual design
was often similar. Two points are striking: given the dominance of Persian as an
official language in many regions of India over a vast period, we would expect
Persian and Arabic to be visually dominant in the inscriptions as well. While
this is often the case, we also find many inscriptions in which the Persian and
Arabic text is at the top of the inscription, but of equal size or even shorter than
the Indic text. This broad overview cannot address the political and social con-
text in which the single inscriptions were produced; it is rather meant as a pre-

||
3 https://siddham.network/ (accessed on 4 September 2023).
4 Bunčić 2016, 51–54.
5 See below, Subsection 4.6.
6 Soroni 2022, 2.
7 Soroni 2022.
8 Sheikh 2014.
 Bilingual Inscriptions from India | 373

liminary attempt to present the corpus and a starting point for further studies on
bilingual inscriptions from India. It has not been possible to consider every
archaeological publication on inscriptions in India, so the corpus can certainly
be enlarged, especially by including unpublished inscriptions. The article will
describe some general features of the corpus and look at some of the most inter-
esting inscriptions in greater detail, considering especially the relations be-
tween the texts involved. The historical and social conditions of the inscrip-
tions’ production are not examined.

2 Languages
Twenty-eight of the inscriptions studied here – i.e. the majority – are in Persian
(with or without some Arabic) and Sanskrit.9 They are spread over a period of some
one thousand years, and come from many different regions of the subcontinent:
the Swat valley, Delhi, Gujarat, Rajasthan, Mumbai, the Deccan and other places.
Persian-Sanskrit bilingual inscriptions have been produced by various dynasties
and rulers, but the corpus contains no Persian-Sanskrit inscriptions from the peri-
od of Bābur (r. 1526–1530 CE), Humāyūn (r. 1530–1540 CE, 1555–1556 CE) or Akbar
(r. 1556–1605 CE), and only one from Jahāngīr (r. 1605–1627 CE).
There are furthermore five inscriptions in Arabic and Sanskrit.10 Though few
in number, these are among the most interesting bilingual inscriptions, both
with regard to their content and their placement (see below). The earliest Ara-
bic-Sanskrit bilingual inscription, from the Tochi valley in today’s Pakistan,
dates to 243 AH / 857 CE; the latest, from Burhānpūr in contemporary Madhya
Pradesh, to 997 AH / 1588–1589 CE.
The second largest group are inscriptions in Persian and Marathi: a total of
nineteen inscriptions belong to this group.11 We must however consider that
seven of these are directional inscriptions, containing no more than place
names, and that among these directional inscriptions, five consist mainly of
transliterated Persian, written in Nāgarī script. One inscription in Arabic and a
mixture of Gujarati and Marathi has also been preserved.12 The majority of Per-

||
9 Persian and Sanskrit: 3, 5, 6, 8, 9, 10, 13, 14, 15, 17, 19, 20, 21, 22, 27, 29, 30, 32, 37, 58, 70, 74,
76; Persian, Arabic and Sanskrit: 18, 24, 38, 63, 70.
10 Arabic and Sanskrit: 2, 4, 26, 44, 46.
11 Persian and Marathi: 7, 35, 42, 43, 47, 48, 49, 56, 57, 65, 72, 73, 75; transliterated Persian:
50, 51, 52, 53, 55; Persian, Marathi and Kannada: 36, 60.
12 Arabic with a mixture of Marathi and Gujarati: 11.
374 | Eva Orthmann

sian and Marathi inscriptions belongs to the Niẓām Shāhī period (1490–1636 CE)
and comes from this dynasty’s realm.13
We furthermore find inscriptions in Persian and Telugu,14 Persian and Raja-
sthani,15 Persian and Kannada,16 Persian and Kanarese,17 Persian and Moṛi,18 and
Persian and Hindi.19 For some inscriptions, the language has not been specified.
In most of these cases, only the script is indicated (as Nāgarī), with no specifica-
tion of the language.20 In some cases, the second language is a local dialect.21
Four multilingual inscriptions also deserve to be mentioned: first and fore-
most, the so-called Tarisāppaḷḷi copper plates, which are mainly inscribed in
Tamil, but contain signatures in Syriac, Arabic and Pahlavi (no. 1). Another
trilingual inscription comes from Sri Lanka; this is in Tamil, Chinese and Per-
sian. Further trilingual inscriptions are in Persian, Old Gujarati and Sanskrit
and Persian, Marathi and Kannada.22

3 Dating
The inscriptions under study span a period of some one thousand years, with
the earliest dating to 850 CE, the latest to 1830 CE. Of course, bilingual inscrip-
tions have also been produced after 1830 CE – they are omnipresent in today’s
India – but the time frame considered here ends with the beginning of direct
British rule in 1857 CE. There are only two inscriptions from the ninth century,
and one from the eleventh century. The number of inscriptions significantly
increases in the fifteenth century, with a peak in the sixteenth century. There is
a significant number of inscriptions related to the Niẓām Shāhī dynasty,23 and

||
13 For a study of this corpus, see Soroni 2022.
14 Persian and Telugu: 33, 66, 68.
15 Persian and Rajasthani: 61, 69.
16 Persian and Kannada: 28; Persian, Marathi and Kannada: 36, 60.
17 Persian and Kanarese: 25.
18 Persian and Moṛi: 40.
19 Persian and Hindi: 23.
20 Arabic, Persian and Nāgarī: 38; Persian and Nāgarī: 34, 39, 41, 67; language not specified:
31, 45.
21 Local dialects: 54, 59, 62, 64, 71.
22 Chinese, Tamil and Persian: 16; Persian, Old Gujarati and Sanskrit: 12; Persian, Marathi and
Kannada: 36, 60. No. 36 is however rather triscriptal than trilingual, since the Persian is written
in Arabic and Nāgarī characters.
23 Nos 37, 38, 39, 41, 42, 43, 47, 48, 49, 50, 51, 52, 53, 55, 56.
 Bilingual Inscriptions from India | 375

quite a few to different rulers of Gujarat;24 otherwise, the inscriptions are equally
distributed over the different dynasties. The number of bilingual inscriptions
from the Mughals is rather low, with the majority of them belonging to the peri-
od of Shāh Jahān (r. 1628–1658 CE)25 and Aurangzēb (r. 1658–1707 CE).26
The inscriptions use a number of dating systems. The Arabic and Persian
part of the inscriptions is usually dated in Hijrī years. In most cases, the day and
the month are indicated. The numbers are mostly written out, but, in some cas-
es, they are in Arabic numerals. In a few cases, we also find a chronogram. In-
scriptions from the Deccan, namely from the Niẓām Shāhī and ʿĀdil Shāhī
(1490–1686 CE) dynasties, indicate the Shuhūr-San, a solar year introduced in
the Deccan in 741 AH / 1340 CE.27 A document from the period of Aurangzēb men-
tions the Faṣlī year, a lunisolar year introduced by the third Mughal ruler, Ak-
bar, in 1556 CE, mainly for accounting (no. 72).28
In the period of Shāh Jahān and Aurangzēb, inscriptions are dated in regnal
years. In the Indic versions of the inscriptions, the most widespread dating sys-
tem is the Vikrama saṁvat calendar, a Hindu lunisolar calendar named after
King Vikramāditya, whose era began in 57 CE. The Śaka saṁvat, also known as
the Śālivāhana Śaka, is also used. Named after King Śālivāhana, it is a lunisolar
calendar beginning in 78 CE. The earliest inscription mentions, in Sanskrit, the
year [39]32, which belongs to the Laukika or Śāstra era used in the Punjab and
Kashmir, beginning in 3076 BCE. The inscription from Prabhas Pātan (no. 4) also
contains the Valabhī era, which began 375 years after the Vikrama saṁvat era
and for which this inscription is an important source.29 The Siṁha era, which
begins in 1113/1114 CE, is also mentioned in the same inscription.30 The dates are
usually very precisely indicated in the Indic version. There are sometimes dis-
crepancies, however, of a matter of days or even weeks between the Vikrama or
Śaka date and the Hijrī date. This could indicate that the respective inscriptions
were installed on different dates.
Another significant difference between the indication of dates in the Arabic
or Persian and Indic versions is the placement of the date. In Indic inscriptions,
it usually comes at the very beginning, while in the Arabic or Persian version,
different places are possible, but it often comes at the end.

||
24 Nos 4, 6, 12, 13, 15, 19, 20, 24, 29, 31.
25 Nos 61, 62, 63, 64, 69.
26 Nos 67, 69, 70, 71, 72, 73.
27 On the Shuhūr-San, see Martin 1971.
28 On the Faṣlī year, see Blake 2013, 117–118.
29 Virji 1952, 106.
30 On this era, see Indraji 1896, 176.
376 | Eva Orthmann

4 Contexts and types of inscriptions

Most of the inscriptions are related to building activities. We thus find inscrip-
tions referring to the construction of a tomb;31 a mosque or minaret;32 a step
well, well or spring;33 doors and gates;34 military structures like walls, fortifica-
tions and bastions;35 cities;36 gardens;37 and other buildings.38 One specific group
of inscriptions served as sign posts.39 Other inscriptions specify the details of
endowments and grants,40 refer to the cultivation of land41 or indicate the con-
tent of farmāns (royal orders), tax regulations and other instructions.42 We also
find an epitaph (no. 11) and a description of offerings by the Chinese emperor
(no. 16). The types of inscriptions do not differ very much from monolingual
inscriptions. However, religious inscriptions are not very prominent, and nei-
ther epitaphs nor inscriptions in mosques are present in large number.
The original placement of the inscriptions was related to their content.
Those that are still in situ are usually located in close proximity to the object
referred to in the inscription, like on the wall of a well, the battlement of a forti-
fication, or somewhere in the mosque.

5 Stylistic features and arrangement

The placement of the two or sometimes three languages in the inscriptions is
not always the same, but one can identify some recurrent patterns. These differ
in terms of period, the languages involved and the respective dynasty. The in-
scriptions also vary in terms of how they are made. Traditionally, Indic texts are
incised, and Persian/Arabic texts are in relief. In the bilingual inscriptions, we

||
31 Nos 3, 32, 62.
32 Nos 4, 7, 9, 22, 29, 40, 44, 46, 66.
33 Nos 5, 10, 12, 14, 18, 21, 26, 27, 29, 30, 32, 33, 39, 57, 60, 61, 63, 64, 70.
34 Nos 13, 67, 75.
35 Nos 15, 20, 31, 37, 38, 41, 43, 45, 60, 68, 69.
36 Nos 24, 58, 64, 73.
37 Nos 32, 35.
38 Nos 2, 42.
39 Nos 47, 48, 49, 50, 51, 52, 53, 55.
40 Nos 6, 8, 36, 56, 74, 76.
41 Nos 25, 33.
42 Nos 1, 17, 19, 23, 28, 34, 54, 59, 65, 71.
 Bilingual Inscriptions from India | 377

find three different arrangements: the texts may both be incised or in relief, or
the Indic text may be incised and the Persian/Arabic text in relief.
In the following, the inscriptions will be placed into four different catego-
ries. These categories are defined according to the size and prominence allotted
to the languages on the slab. The categories thus do not reflect the length of
each text, but its outward appearance. Since the characters used for one lan-
guage might be smaller than those used for the other, texts of different sizes
might have the same content, whereas in texts of the same size, one language
might contain more information. Some of the cases are difficult to decide, either
because one of the texts has only been partially preserved, or since no photo of
the Indic text has been published.

5.1 Indic text marginalized

In several inscriptions, the Persian/Arabic text is entirely dominant, while the
Indic text is very small and in some cases almost invisible in the margins of the
slab. The Indic text is in most cases in Nāgarī, and the language is Sanskrit or
Hindi, but sometimes unspecified. Often, the Indic text is almost invisible in the
published image and no translation is provided; this is the case for eight in-
scriptions.43 In two cases (nos 31, 45), the Indic text is not mentioned in the edi-
tion at all. For no. 31, it is possible that the Persian text was inscribed on a spolium
from an earlier building, and that the Indic text was not related to the Persian.
In this category of bilingual inscriptions, the Persian text is always promi-
nently in the centre, and in most cases in relief. Inscriptions nos 37, 38 and 39
are all related to a certain Aflāṭūn Khān, who was the local governor for Mur-
taḍā Niẓām Shāh I (r. 1565–1588 CE) of Ahmadnagar. The Indic version is very
short, and in all three cases very damaged. It contains the construction date. In
one case, Arabic numerals are spelled out in Nāgarī script (no. 38). Inscriptions
nos 67, 69 and 70 were produced in the period of Aurangzēb, but differ in their
arrangement of the two texts: in the first case, the Indic text is below the Per-
sian; in the second, above it; and in the third, in the right and left margins. All
three inscriptions deal with construction works in the realm of a rāja or
mahārāja; the Indic versions contain the names of the persons involved.
We may wonder who the addressees of the Indic part of the inscriptions
were. It was certainly only possible to read them from up close, while from a
distance, one might not even have seen them. In the case of constructions from

||
43 Nos 31, 37, 38, 39, 45, 67, 69, 70.
378 | Eva Orthmann

the period of Aurangzēb, it is reasonable to assume that the Indic texts were
added because the construction works were carried out in the territory of a rāja
or mahārāja. It remains unclear, however, why even in such an environment,
the Indic versions were entirely marginalized.

5.2 Indic text in smaller size

In many inscriptions, the Indic text, in most cases in Nāgarī script, is smaller in
size than the Persian/Arabic text. The letters are smaller, and – with the excep-
tion of one inscription (no. 76) – the Indic text is below the Persian/Arabic.
There is thus a clear hierarchy between the two texts, with the Persian/Arabic in
the dominant position.44 All three types – with both the Persian and Indic text
incised,45 the Persian in relief and the Indic text incised,46 and both texts in re-
lief47 – feature this arrangement.
The comparison of the two versions is not easy, since in many cases, the In-
dic version is too weathered to be properly deciphered, and the Persian/Arabic
version is also often damaged. The inscriptions are from many different areas
and periods. But two general tendencies can be observed: first, in many cases in
which the Indic version is shorter, it focuses on the essential information, like
the name of the builder, the governor and the ruler and the purpose of the in-
scription. The basmala,48 eulogies and blessing formulas, as well as threats,
have often been deleted; this implies that specifically Islamic elements have
been omitted, and only in a few cases replaced by Hindu phrases. Second, in
spite of the reduction of space, many Indic inscriptions contain more chronolog-
ical information.

5.3 Indic and Persian/Arabic texts of the same size

The second largest group of bilingual inscriptions has Indic and Persian texts of
the same size.49 There are four subcategories in this group: the first and predom-
inant group consists of inscriptions in which the Persian/Arabic and Sanskrit

||
44 Nos 3, 6, 8, 9, 10, 15, 17, 19, 21, 22, 28, 34, 35, 42, 43, 47, 58, (65), 66, 71, 76.
45 Nos 3, 8, 15, 66, 71, 76.
46 Nos 6, 9, 10, 17, 19, 21, 22, 28, 34, 54, 58.
47 Nos 42, 43.
48 The phrase ‘In the name of Allāh, the Beneficent, the Merciful’ is called the basmala.
49 Nos 2, 4, 5, 7, 11, 13, 23, 27, 29, 30, 32, 41, 44, 56, 59, 61, (64), 68, 72, 73.
 Bilingual Inscriptions from India | 379

text are on the same slab and placed one below the other.50 In these cases, the
Persian/Arabic inscription is always on top and thus optically prioritized. This is
not the case for the other three types, in which no priority can be established. In
the second type, the texts are placed next to each other, on the right and left
sides of the same slab.51 Here, the Indic text is on the left side. This arrangement
is probably related to the different writing directions, because Persian/Arabic is
written from right to left, and the Indic scripts from left to right. The third type
consists of texts that have been placed on both sides of the same slab. Only one
inscription of this type is found in the corpus (no. 5). In the fourth category, the
two texts are on different slabs.52 In these cases, it is often not easy to under-
stand the intended relative position of the two inscriptions to each other, since
many of these inscriptions have not been found in their original context. There
are too many different relations between the two versions of the text to summa-
rize them here; for some examples, details of the relationship between the two
versions of the text are given below.

5.4 Indic text larger than Persian/Arabic text

There is also a certain number of inscriptions in which the Indic text is longer
and more prominent.53 The order of the texts varies; in most cases, the Per-
sian/Arabic text is on the top;54 in others, it is on a different slab.55 In one in-
scription, the two languages are on different sides of a copper plate (no. 74).
There is one single inscription in which the Persian is below the Indic text (no.
62), and because of the layout of the two texts, it deserves special attention. It
will be described in detail below (Subsection 6.9).
Here again, the relationship between the texts in different languages varies,
and no general rule can be established. Some examples will therefore also be
discussed.

||
50 Nos 2, 11, 13, 23, 27, 29, 32, 41, 44, 59.
51 Nos 30, 61, 64.
52 Nos 4, 26, 56, 68, 72, 73.
53 Nos 18, 24, 25, 33, 46, 57, 62, 63, 74.
54 Nos 18, 25, 33, 46, 63.
55 Nos 24, 57.
380 | Eva Orthmann

5.5 Tri- and multilingual inscriptions

Altogether five examples of inscriptions with more than two languages are in-
cluded in the corpus. The first is the Tarisāppaḷḷi copper plates, the oldest in-
scription included in the corpus (no. 1). On these plates, the main language is
Tamil, while the other languages – Hebrew, Arabic and Pahlavi – have only
been used for signatures. The majority of the space is therefore occupied by
Tamil, which was the language of the Chera Parumals, the ruling dynasty of
Kerala at that time (r. 844–1124 CE).
The second inscription is from Sri Lanka and is in Tamil, Persian and
Chinese (no. 16). The Tamil version is on the top left, the Persian on the bot-
tom left, and the Chinese fills the entire right side. The size of all three texts
is almost the same, with none dominating the slab. This slab was not pro-
duced locally, but in Nanjing, and was transported to Sri Lanka to be placed
in a temple. Due to its location in Sri Lanka, this inscription is exceptional in
many regards.
The trilingual inscription in Persian, Marathi and Kannada (no. 60) from
Karnataka belongs to the ʿĀdil Shāhī dynasty. The size of the three versions is
more or less the same, with Persian first, then Marathi, and Kannada last. The
Marathi and Kannada versions additionally report the construction of a well for
the god Śambhu, while the Persian version has more praise formulas. In the
trilingual inscription in Persian, Old Gujarati and Sanskrit, the Persian version
is dominant.56

5.6 Direction stones

In a category of their own are the direction stones, or signs indicating the way to
specific places.57 Almost all of these were erected around the year 1000 AH /
1591–1592 CE by Burhān Niẓām Shāh II (r. 1591–1595 CE), and one by Burhān
Niẓām Shāh III (r. 1609–1631 CE). On these direction stones, the majority of the
words written in Nāgarī are Persian or Arabic, like jamān for zamān (Persian),
‘time’; syā for shāh (Persian), ‘king’; and aḷapha for alf (Arabic), ‘thousand’.
They are thus rather biscriptal than bilingual, with Persian and Arabic written
in two different alphabets. However, in between the Persian words, we also find

||
56 No. 12; see Sheikh 2014, 202.
57 Nos 47, 48, 49, 50, 51, 52, 53, 55.
 Bilingual Inscriptions from India | 381

some terms in Marathi.58 We may wonder who was supposed to read these
direction stones, and why they were produced in this way. In the case of Latin
inscriptions written in Greek, Martti Leiwo has suggested that the Greek char-
acters were used because the stonecutter did not know the Latin characters,
and because the person who set up the inscription did not know the difference
between Latin and Greek.59 This explanation cannot be transferred to the Ma-
rathi–Persian case, since the direction stones have Persian characters in addi-
tion to the Nāgarī ones. The reason for writing Persian in Nāgarī must there-
fore have been related to the addressees. According to Soroni, there also exist
letters and farmāns with Persian written in Moṛi script, meant to be read
aloud.60 We must therefore assume that there were people who were able to
read Nāgarī or Moṛi, but not Persian, and who knew Persian only orally. Be-
sides the direction stones, two further inscriptions are known in which Nāgarī
script is used to write Persian or Arabic: a grant for the village of Shirol (now
Jaisinghpur, Maharashtra), which, besides an inscription with Persian in Ara-
bic and Nāgarī characters, also has an inscription in Kannada, and an inscrip-
tion from Galna fort (Nasik district, Maharashtra), in which the Arabic year is
written in Nāgarī.61

6 Examples of bilingual inscriptions

The number of bilingual inscriptions in the corpus is too high to discuss them
all. In the following, only some examples of bilingual inscriptions are therefore
examined. The selection was guided by the wish to choose inscriptions that are
distinguished by particular differences between the two versions, or that stand
out due to certain historical or stylistic features.62

||
58 See Sohoni 2022, 1–2, who also gives a good transcription and transliteration of one of these
direction stones.
59 Leiwo 2002, 178.
60 Soroni 2022, 6.
61 Nos 36 and 38.
62 The translations of the Indic texts are taken from the editions, with some corrections of the
English. The translations of Arabic and Persian usually also follow the editions, but are occa-
sionally modified according to the present author’s understanding or – in cases of passages
from the Quran – according to contemporary standard translations.
382 | Eva Orthmann

6.1 A Persian and Marathi inscription from the Niẓām Shāhī

period (no. 56)
Persian version:

‫ ﺍﺯ ﺍﻳﻦ ﺑﺎﻍ ﺯﻣﻴﻦ‬،‫ ﺑﺴﺎﻁ ﺧﺎﻥ ﮐﻪ ﻫﺴﺖ ﺍﺯ ﺟﺎﻥ ﭼﺎﮐﺮی‬،‫ﺩﺭ ﻋﻬﺪ ﺷﺎﻩ ﻋﺎﺩﻝ ﮐﺎﻣﻞ ﻧﻈﺎﻡ ﺷﺎﻩ‬
‫ ﺑﺮ ﺯﻥ ﺍﻭ ﺧﺮ‬، ‫ﺍﻧﻌﺎﻡ ﺑﺪﻝ ﭼﺮﺍﻍ ﻣﺴﺠﺪ ]ﻭ[ ﮔﻨﺒﺪ ﺩﺍﺩ ﺍﺳﺖ – ﻫﺮ ﮐﻪ ﻣﻨﻊ ﮐﻨﺪ‬

Translation of the Persian:

In the reign of the just and perfect king Niẓām Shāh, Bisāṭ Khān, who is (His Majesty’s)
most loyal servant, has endowed land from this garden as inʿām for the lights of the
mosque and the dome. Whoever contravenes, may a donkey be on his wife!

Transliteration of the Marathi:63

nījāma syā kā kadīma … vīsāta — b(v?)āgaṁ

jamīna īnāma vadala dīvā vattī. masīdu va
gubhata dīdhalā śrāhe jo koyahī manā karola
————māṃ para gacchava

Translation of the Marathi:

Niẓām Shāh’s old (?) … Bisāṭ … has given the garden land in inʿām for lighting the mosque
and the dome. Whosoever forbids … donkey.

The two versions of the inscription are remarkably close to each other. Accord-
ing to Pushkar Sohoni’s categorization scheme,64 the same text is rendered in
the two languages. Even with some of the Marathi part missing, the purpose of
the grant is evidently the same: in both cases, the endowed object is the garden;
the name of the ruler and the donor are similar. The only difference seems to be
at the beginning, where the Persian text refers to the reign of Niẓām Shāh. The
Persian term inʿām (‘grant, endowment’) has been kept in the Marathi version,
probably because the composer wanted to keep this culturally specific term
with all its juridical implications. From the Persian side, the typical western
Indian curse ‘may a donkey be on his wife’ has been adopted.65 Based on the
textual evidence, it is impossible to decide which inscription was composed
first, and which is a translation. Placed on two separate slabs, none of them is

||
63 Transliteration by Lingli Li.
64 Sohoni 2022, 8. This scheme is a modification of the one proposed in Leiwo 2002, 173–174.
65 This inscription is also discussed in Sohoni 2022, 9.
 Bilingual Inscriptions from India | 383

optically dominant. Since the content is about a typical Islamic grant, I assume
however that the Marathi is a translation of the Persian version.

6.2 Inscription from Junagarh, Khaljī period (1290–1329 CE)

(no. 22)
Persian version:

‫( ﺑﺴﻢ ﷲ ﺍﻟﺮﺣﻤﻦ ﺍﻟﺮﺣﻴﻢ ﺑﻨﺎ ﮐﺮﺩ ﺍﻳﻦ ﻣﻴﻨﺎﺭ ﻣﻠﮏ ﺳﻴﺪ ﻣﺤﻤﺪ ﻣﺒﺎﺭک ﻋﺰ ﺳﻠﻄﺎﻧﭙﻮﺭی‬1)
‫( ﺩﺭ ﻋﻬﺪ ﺳﻠﻄﺎﻥ ﻗﻄﺐ ﺍﻟﺪﻳﻦ ﺑﻦ ﻣﺤﻤﺪ ﺷﺎﻩ ﺍﻟﺴﻠﻄﺎﻥ ﺑﺘﺎﺭﻳﺦ ﻳﺎﺯﺩﻫﻢ ﻣﺎﻩ ﺭﺑﻴﻊ ﺍﻻﻭﻝ‬2)
[‫ ﺑﺎﻧﯽ ﺭﺍ ﺩﻋﺎ ﺍﻳﻤﺎﻥ ﻳﺎﺩ ﮐﺮ]ﺩ‬...‫ﻫﺮ ﮐﻪ‬... (3)

Translation of the Persian:

(1) In the name of Allāh, the Beneficent, the Merciful. This mīnār (tower, minaret) was
built by Malik Sayyid Muḥammad Mubārak ʿAzz of Sulṭānpūr,
(2) during the reign of Sulṭān Quṭb ud-Dīn son of Muḥammad Shāh, the sulṭān, on the 11th
of Rabīʿ I.66
(3) …Whoever … offers a prayer for (the soundness of) the faith of the builder.

Transliteration of the Sanskrit:

(1) saṁvat 1514 barṣe

(2) śrāvaṇa-badi [2 rtha ravau] sulatāna-śrī-kutabadina-viji-rāje śrī devapatana-
(3) taḥ sulatāna-[pahī] malika-śrīḥ mubāra[ka-su]ta-malika-śrī-mahaṁmada-suta-malika-
śrī-[makada ...].
(4) ji … ipati vajepa … ā … na ha punya-tī a ...ī ... ṣa ... na…i ... bīrāsila
(5) ... ī … lā sūtra[dha]ra ... sūya...i ... tta ... su ... la ... tī saṁvat
(6) [15]14 ba[r]ṣa ...

Translation of the Sanskrit:

In the (Vikrama) year 1514, on Sunday, the second day of the dark fortnight of (the month
of) Śrāvaṇa,67 during the victorious reign of the illustrious Sulatāna, Kutabadīna (Sulṭān
Quṭb ud-Dīn), from the glorious (city of) Devapatana, the illustrious Malika (Malik)
[Makada …], son of the illustrious Malika Mahaṁmada, son of the illustrious Malika Mu-
bāraka, … Sulatāna …Pious [holy place] … the mason … the year [15]14…

||
66 The third month of the Islamic lunar year.
67 The fifth lunar month of the Hindu calendar.
384 | Eva Orthmann

Even with the incomplete reading of the Sanskrit, the differences between the
Persian and Sanskrit versions are obvious. The Sanskrit begins and ends with
the date, which appears in the Persian version only once, towards the end, be-
fore the call to pray for the builder. This is missing in the Sanskrit version. The
names and the genealogical information differ slightly, and Sulṭānpūr is called
by its old name Devapatana. Due to the lacunae, we do not know which word
was used for the minaret. The Sanskrit version contains more adjectives, like
‘victorious’ and ‘illustrious’. Although the two texts convey the same message,
they are therefore quite different. The inscription belongs to Sohoni’s Category 3
(two different texts of nearly the same content). The size and placement of the
Persian text give it a dominant position, but the Sanskrit text is not a translation
of the Persian.

6.3 Quṭb Shāhī inscription from the Nalgonda district,

Telangana (no. 33)
Persian version:

‫ﻗﻄﺒﺸﺎﻩ‬
‫ﺍﺑﻮ ﺍﻟﻤﻈﻔﺮ ﺳﻠﻄﺎﻥ‬
‫ﺩﺭ ﺗﺎﺭﻳﺦ ﭼﻬﺎﺭﺩﻫﻢ ﺷﻬﺮ ﺭﻣﻀﺎﻥ‬
‫ ﺣﻀﺮﺕ ﭘﻨﺎﻫﯽ‬٩٥٨ ‫ﺍﻟﻤﺒﺎﺭک ﺳﻨﻪ‬
‫ﻧﻘﺎﺑﺖ ﺩﺳﺘﮕﺎﻫﯽ ﺳﻴﺪ ﺷﺎﻩ ﻣﻴﺮ ﺑﻦ‬
‫ﺍﻟﻤﺮﺣﻮﻡ ﺳﻴﺪ ﺍﺣﻤﺪ ﻁﺒﺎﻁﺒﺎﺋﯽ ﺍﺻﻔﻬﺎﻧﯽ‬
‫ﮐﺎﻟﻮﻩ ﺣﻮﺽ ﭘﺎﻧﮕﻞ ﮐﻪ ﺑﻌﺪ ﺍﺯ ﺳﺎﻟﻬﺎ ﺧﺮﺍﺏ ﺷﺪﻩ‬
‫ﺑﻮﺩ ﺍﺯ ﺟﻬﺔ ﺗﻮﺍﺏ ﺯﺭ ﺧﺮﺝ ﮐﺮﺩ ﻭ ﻣﻌﻤﻮﺭ ﺳﺎﺧﺖ‬
‫ﻭ ﮐﺴﯽ ﮐﻪ ﮐﺎﻟﻮۀ ﻣﺬﮐﻮﺭﺭﺍ ﻣﺸﻘﺖ ﮐﺮﺩ ﻭ ﺍﻫﺘﻤﺎﻡ ﻧﻤﻮﺩ‬
‫ﺭﺣﻤﺖ ﷲ ﺑﻦ ﻋﺒﺪ ﺍﻟﮑﺮﻳﻢ ﺧﻮﺍﻧﺸﺎﻫﯽ )؟( ﻭ ﺍﺯ ﺍﺑﺘﺪﺍی‬
...‫ ﺗﺎ ﺣﺪ ﻧﺪی ﮐﺮﺷﻨﺎ ﺍﺯ ﮐﺎﻟﻮﻩ ﻭ ﺣﻮﺽ‬...‫ ﺍﻧﺪ ﻭ‬...‫ﮐﺘﻮﻩ‬
‫ﺩﺭ ﻗﺼﺒﮥ ﭘﺎﻧﮕﻞ ﭼﮑﻴﺪﻩ ﮔﺸﺖ‬...
...‫ﻭ ﺩﺭ ﺗﺤﺖ ﺣﻮﺽ ﺣﺼﻪ )؟( ﻣﺴﻠﻤﺎﻥ ﻭ‬
٩٥٨ ‫ﺑﺎﺩ ﺳﻨﻪ‬...

Translation of the Persian:

[In the reign of] Quṭb Shāh Abū l-Muẓaffar Sulṭān [Ibrāhīm]. On the 14th of the blessed
month of Ramaḍān 958 [15 September 1551 CE], His Honour [lit. ‘His Refuge’], the wielder
of authority, Sayyid Shāh Mīr, son of the late Sayyid Aḥmad Ṭabāṭabāʾī Iṣfahānī’, rebuilt
the embankment of the Pangal tank, which had fallen out of repair through age, and spent
money [on this work] to obtain [God’s] forgiveness. The person who worked hard and su-
pervised [the repairs] was Raḥmat Ullāh, son of ʿAbd ul-Karīm Khwān Shāhī [?]. The em-
bankment was breached from … to the boundary of the river Krishna, and waters flowed
 Bilingual Inscriptions from India | 385

from the tank through the breach (of the dyke) to the town of Pangal … The share of the
Muslims from [the lands of] the tank … remain …! 985 AH.

Transliteration of the Telugu:

(1) śrī [II*] svasti śrī jayâbhyudaya śālivāha

(2) na-śaka-varuṣaṁbulu 1432 yagunāṭi pramō
(3) da-saṁvvatsara-māgha-śu.15 Bhaumavāraṁ | Śrīma
(4) n-mahāmaṁḍḍalēśvara Yiburāhim Kutubu
(5) Śa rājyamu śeyaṁgānu vāri maṁnnana-sabhi
(6) kuṁḍu Sayidu Sādātu Sayidu Śahā-Mī-
(7) ruku puṇyamugānu āyana nija-hitu
(8) ḍu vivēka-bhūṣaṇuḍu maṁnnana-tējō
(9) nidhi ayinaṭhavaṁṭi Rāmattullāgāru Pānu
(10) gaṁṭṭi Vudaya-samudraṁ-kāluva Yiṁddu
(11) pukela-sīmalō Namile-vadda Musi-yēṭi kattuva khi
(12) lamai vuṁḍḍaṁgānu punar-ōddharakamugānu
(13) kattuva gaṭṭi kāluva śēyiṁcci ā nīḷlu U
(14) dayasamudramu niṁcci āügi nīḷlu Kr
(15) ṣṇa-gāmini śēyiṁci Namile-kattuva moda
(16) lu koni Krṣṇa yīmadhya cheruvulu kuṁṭalu
(17) kāluvalu niṁcci yiṁdula pāla-paḍḍa dhānyānaku da
(18) śabaṁddamu Pānugaṁṭṭi kindda naḍachēnu ā-chamdr-ā
(19) rka-sthāyigā | Yī Vudayasamudraṁ ve
(20) nuka pāḷla vivaramu, rājuku pālu
(21) 1 prajaku palumnn-ara 11/2-ṁ Turukala-
(22) ku Brāṁhmalaku pāḷlu reṁḍḍu 2 yī.
(23) choppana ā-chamdr-ārkka-sthāyigā i-
(24) stimi || Yī dharmaṁ pratipāliṁcinavāri-
(25) ki yaṁttō puṇyaṁ | sva-dattādvi-guṇaṁ pu-
(26) ṇya[ṁ] para-datt-anupālanaṁ | para-datt-āpa-
(27) hārēṇa | sva-dattaṁ niṣphalaṁ bhavēttu || Śrī [||*]
(28) naṣṭaṁ kulaṁ bhinna-taṭāka-kūpaṁ | babhra-para-
(29) rājyaṁ śaraṇâgataṁ cca | gāṁ Brāhmaṇāṁ dē
(30) va-gṛihâlayaṁ cca yō[d*]dharē[t*] pūrva-catur-guṇa[h*].
(31) syat || Maṁgala mahā-śrī śrī śrī jēyu-
(32) nū || Śrī[||*]

Translation of the Telugu version:

Hail! On Tuesday, the 15th day of the bright half of Māgha in the Jovian year Pramoda, the
year of the Śālivāhana Śaka era being 1432, while Mahāmaṇḍalēśvara Yiburahiṁ Kutubu
Śa (Ibrāhīm Quṭb Shāh) was ruling. Ramattulla (Raḥmat Ullāh), who was kind to his own,
had prudence as his ornament and who was like a revered treasure of splendour, saw that
the channel from the Udayasamudram (tank) in Pānugallu and the dam of the river Musi
near Namile in the Yindupukela-area had been dilapidated. For the sake of their recon-
386 | Eva Orthmann

struction he fixed the dam and had a channel made and filled the Udayasamudram (tank)
with its waters. He then made the water flow into the Krṣṇā river and filled the tanks,
ponds and canals between the Namile dam and the Krṣṇā river. (It was also ordered) that
there be a daśabandham on the shared crops there (around the tanks, ponds canals, etc.)
below Pānugallu as long as the sun and the moon last. (These acts of charity) were done
for the merit of Sayidu Sādātu Sayidu Śāhā-Mīru (Sayyid us-Sādāt, Sayyid Shāh Mīr), who
was an honoured councillor (of the king). The shares (of income) from [the lands irrigated]
by the Udayasamudram (tank) are explained:
We (i.e. Raḥmat Ullāh) have given, at the following rates, one share to the king, one share
and a half to the subjects, and two shares to the Turukas (i.e. Muslims) as well as to the
Brahmans for so long as the sun and the moon endure. Those who protect this charitable
act (dharmamu) will acquire immeasurable merit.
Maintaining the gift of another is twice as meritorious as one’s own gift. By stealing
another’s gift, one’s own gift becomes fruitless.68
He who should support a family ruined after their ponds and wells have burst, and a
kingdom that has sought protection, and a cow and a Brahman on the grounds of a tem-
ple, will have four times [as much merit] as before [??].

Here, the differences between the two versions are even more remarkable. First of
all, the Telugu version is longer. Although the Persian is at the top of the slab, the
Telugu occupies the majority of the space and is more prominent. The sun and
moon engraved on the top of the stone further enhance the dominance of the non-
Persian elements. This dominance is also reflected in the content of the text. Both
versions deal with the same event, namely the rebuilding of the Pangal tank and
the building of a channel. The exact shares and the taxation of the crops
(daśabandham tenure) are recorded in greater detail in the Telugu. The Telugu
version mentions that the charity was done for the merit of Sayyid Shāh Mīr, prom-
ises merits for those who protect the endowment, and praises the virtue of giving.
In the Persian version, the purpose of the endowment consists in obtaining God’s
forgiveness and thus in securing recompense for the next life – a typical Islamic
concept. What is perhaps most remarkable is the difference in prominence be-
tween the person who financed the building activities and the one who supervised
the work. In Telugu, the supervisor is not only mentioned earlier, but also praised
more greatly. Was he involved in the wording of the Telugu inscription, or did he
himself perhaps understand Telugu? Was he a locally prominent person whose
merit was more meaningful to the Telugu-speaking community?
In conclusion, we also have to consider the relation between the two texts: is
the Persian version a summarized translation of the Telugu, corresponding to So-
honi’s Category 4? Or are these rather two texts that have been prepared on the

||
68 Language of this verse and the next is Sanskrit, not Telugu.
 Bilingual Inscriptions from India | 387

same topic in ‘dynamic equivalence’?69 Given the difference of emphasis between

the builder and the donor in this inscription, what David G. K. Taylor has described
for Aramaic and Greek inscriptions from Palmyra seems to be quite fitting here:

There are numerous examples of bilingual inscriptions where the two texts are not dra-
matically different, and yet it is clear that each is an independent product conforming to
accepted conventions.70

6.4 Inscription from Bidar about the construction of a step

well (no. 18)
The following inscription from Bidar is also longer in its Indic version. Again,
the Persian text is at the top, but the Sanskrit text is much longer, taking almost
two-thirds of the slab.

Persian version:

‫ﺧﻴﺮ ﺣﺎﻓﻈﺎ ﻭ ﻫﻮ ﺍﺭﺣﻢ ﺍﻟﺮﺍﺣﻤﻴﻦ ﻳﺎ ﻏﻔﻮﺭ‬ ‫( ﺑﺴﻢ ﷲ ﺍﻟﺮﺣﻤﻦ ﺍﻟﺮﺣﻴﻢ‬1)

‫( ﺫﮐﺮ ﺑﻨﺎء ؟ ﺗﺎﺭﻳﺦ ﭼﻬﺎﺭﺩﻫﻢ ﻣﺎﻩ ﺟﻤﺎﺩﺍﻻﺧﺮ ﺳﻨﻪ ؟ ﺛﻤﺎﻥ ﻭ ﺍﺭﺑﻌﻴﻦ ﻭ ﺛﻤﺎﻧﻤﺎﻳﺔ ﻣﺨﺪﻭﻣﻪ ﺟﻬﺎﻥ‬2)
‫( ﺑﻴﺒﯽ ﺷﺎﻫﻨﺎﺯ ﻭﺍﻟﺪۀ ﺧﺎﻥ ﻣﻌﻈﻢ ﺧﻠﻒ ؟ ﺍﻋﻈﻢ ﻣﺤﻤﻮﺩ ﺧﺎﻥ‬3)
‫( ﺷﺪ ﺑﻨﺎ ﺍﻳﻦ ﺑﺎﺋﻴﻦ ﺍﺯ ﻋﻮﻥ ﺧﺪﺍی ﮐﻦ ﻓﮑﺎﻥ ﺑﻮﺩ ﺳﺎﻝ ﺍﺯ ﻫﺠﺮﺕ ﭘﻴﻐﺎﻣﺒﺮ ﺁﺧﺮ ﺯﻣﺎﻥ‬4)
‫( ﭼﻬﻞ ﻭ ﻫﺸﺖ ﻭ ﻫﻴﺼﺪ )= ﻫﺸﺘﺼﺪ( ﻭ ﻋﻬﺪ ﺷﺪﻩ ﺷﻪ ﻋﺎﺩﻝ ﮐﯽ ﺑﻮﺩ ﺍﻝ ﺑﻬﻤﻦ ﺷﺎﻩ ﻭ ﺩﺍﺭﺍ‬5)
‫ﺷﺎﻩ ﺍﺣﻤﺪ ﺫﻭ ﺍﻻﻣﺎﻥ‬
... ‫ ﺷﺎﻩ‬...‫( ﻣﺎﺩﺭ ﻣﺤﻤﻮﺩ ﺧﺎﻥ ﻣﻬﻮﺭی؟‬6)
‫( ﻋﻮﺭﺕ ﺻﺎﺣﺐ ﺳﻌﺎﺩﺕ ﭘﺎک ﺩﺍﻣﻦ ﺑﻮﺩ ﺗﺎ ﺩﺳﺖ ﺩﺍﺩ ﺍﻳﻦ ﺧﻴﺮ ﺟﺎﺭی ﻫﺮ ﮐﻪ ﺑﺨﻮﺭﺩ ﺁﺏ ﺁﻥ‬7)
‫( ﺩﺭ ﺩﻋﺎء ﻋﺎﻗﺒﺖ ﺧﻴﺮ ﺍﻳﻦ ﺳﻪ ﺗﻦ ﺩﺭ ﻫﺮ ﺩﻣﯽ ﺑﺮ ﮐﺸﺎﻳﺪ )= ﮔﺸﺎﻳﺪ( ﺍﺯ ﻣﻴﺎﻥ ﺟﺎﻥ ﺑﺼﺪﻕ‬8)
‫ﺩﻝ ﺯﺑﺎﻥ‬

Translation of the Persian:

In the name of Allāh, the Beneficent, the Merciful. ‘Allāh is the best preserver, and He is
the Most Merciful of those who show mercy.’71 O Forgiver!
Account of the construction, on the 14th of the month of Jumādā II (of the) year eight and
forty and eight hundred, (by) Makhdūma-yi Jahān Bībī Shahnāz, mother of the magnifi-
cent khān and great royal scion Maḥmūd Khān.
This step well was constructed with help from the lord of the universe
the year [counting] from the Hijra of the last of the prophets
was forty-and-eight and eight hundred, and (it was under) the reign of the just king,

||
69 Taylor 2002, 321.
70 Taylor 2002, 321.
71 Quran 12:64, tr. Pickthall 1930.
388 | Eva Orthmann

who is directly descended from Bahman Shāh and Dāra Shāh, namely Aḥmad (Shāh), the
protector.
(The builder is) the mother of Maḥmūd Khān of Mahūr (?) … king
…
She was a fortunate woman, and chaste; hence,
such a flowing bounty came into existence (through her). May whosoever drinks water
therefrom
in prayer for the happy ending of these three personages with every breath
loosen his tongue from the grid of soul with sincerity of heart.

Transliteration of the Sanskrit:72

(1) 1 śriḥ | ajāya jagad-utpatti-sthiti-saṁhāra-kāriṇe | triguṇā[ya] guṇâtīta-mūrtaye brah-

mane na
(2) maḥ || [1*] tad-ājñayâjāyata bhūtale smin bhūri-pratāpâ[na]la-tāpi[tâ]riḥ | ahaṁmado
drpta-nareṁdra-pakṣi-śye-
(3) no mahān bahmani-pātiśāhaḥ || [2*] maheṁdrasya kuberasya nagarī na garīyasī |
apekṣā bedaraṁ yasya na-
(4) garaṁ nagaraṁ[jitam] || [3*] muktāmayaṁ pravālâ[ḍhyaṁ] puraṁ śra-bhavanaṁ hi
yat | ucchair aṁbhodhivad bhā[ti] citraṁ nānaka-
(5) bhūdharaṁ || [4*] vāditrair badhirāyate tribhuvanaṁ [dhūlī]bhir aṁdhāyate yasmin
rāṁjy abhiṣeṇayaty atibharād bhū
(6) miś ca na[mrā]yate yad-durgâdhikrtā malīkaᴗcanāḥ khānāḥ puraḥ koṭayaḥ kas ta-
syâhmada-bahma
(7) nī-narapateḥ saṁkhyātum iṣte camūṁ || [5*] [jīva]d-rāja-maheṁdra-durga-nrpa[tiṁ]
yasy[âgra]-senāpatiḥ kārāmaṁ-
(8) dira-gaṁ [karoti] ᴗᴗ— śrī-devarājaṁ nrpaṁ | vikrāṁtaṁ karadīkaroti sacivo yaś caika-
kaḥ sarvadā
(9) kas tasyâhmada-bahmanī-narapateḥ saṁkhyātum īśo guṇān || [6*] tasyâsty abhimatâbhū
(10) pa-vanitā paryupāsitā | bībī śrī-sahanājh-ākhyā hrepayantī ratiṁ śriyā || [7*] kiṁ
(11) śīta-dyuti-maṁḍalād abhipatatkuṁdôjvalā kaumudī smerâṁbhoruhataḥ prabhūta-
madhuno
(12) niṣyanda-dhārā kim u | kiṁ vā caṁdramaṇeḥ sravan-nava-sudhā-ve[lā] manohāriṇī
kiṁ vā ratna-sa
(13) mūhato bahir asau sphārībhavaddīdhitiḥ || [8*] saubhāgya-suṁdara-satītva-kalā-
vilāsa-cātu
(14) rya-drāna-kuśalatva-suśīlatā[dyā] | yasyāṁ sadā saha-bhuvaḥ prathitā guṇâughā ratnāni
(15) rohaṇa-prthu-dyutimanti santi || [9*] mahamūda-khāna-nā[mā tanayo] māṁ.. [pu]raṁ
yasyāḥ | pāla
(16) yati smâvani-pati-kirīṭa-koṭi-pramrṣṭâṁghriḥ || [10*] sā câhmada-sulutrāṇa-tanaye
naya-śā-

||
72 Metres: verses 1, 3–4, 7, 11–13 Anuṣṭubh; verse 2 Upajāti; verses 5–6, 8 Śārdūlavikrīḍita;
verse 9 Vasantatilaka; verses 10, 15 Upagīti; verse 14 Āryā.
 Bilingual Inscriptions from India | 389

(17) lini | allābadī-pātiśāhe śāsaty ūrvīṁ mahodaye || [11*] tasmin puravare ’gaṇya-tuṁga-
maṁdi-
(18) ra-maṁḍite | vāpīm akārayad ramyāṁ bhūri-sopāna-saṁpadāṁ || [12*] niṁdaty
upavanaṁ ramyaṁ yasyā ... va-
(19) naṁ vanaṁ(nam) | manute câpi pānīyaṁ taṁ payaḥsāgaraṁ garaṁ(ram) | | [13*] idaṁ
ca | tarka-rasânala-caṁdraiḥ 1366 śā
(20) ke raktākṣi-vatsarâśvayuji | māse śukla-daśamyāṁ some śravaṇe dhruve yoge ||
[14*]… n=āhmada-ba-
(21) hmanī-bībī-[sa]hanājhayā vihitā | ā-caṁdrârkaṁ ... s ānaṁdinī jagataḥ || [15*] śrīḥ ||

Translation of the Sanskrit:

[Verse 1] Bow to Brahmā, who is the cause of the creation, sustenance and destruction of
the Universe, who is full of all the three qualities and transcends all the three qualities.
[Verse 2] As ordained by him [Brahmā], the great Bahmani king Ahaṃmada [Aḥmad] was
born on earth, a proud king like a hawk, his enemies scorched by the fire of his majesty.
[Verse 3] Not greater is the capital city of lord Kubera, who looks to the city of Bedara [Bi-
dar] delighted by mountains [?].
[Verse 4] His city is laden with pearls and rich in coral. The lofty splendid palace appears
like the sea; [the city] in which a wondrous drum like a mountain
[Verse 5] (accompanied by) instruments deafens the three worlds; when the king is on the
march they [the three worlds] are blinded by the dust; and because of the extreme burden
the earth bowed down. Crores of kings, Khans were put in charge of his forts/citadels.
Who is able to calculate [the size of] the army of Ahmada Bahmanī.
[Verse 6] His chief general makes the living ruler of the fort at Rājamahendra a prisoner;
and his minister, all by himself, makes the powerful King Devarāja pay tribute in perpetui-
ty. Who can count the virtues of King Ahamada Bahmanī?
[Verse 7] He has a beloved queen, much revered, named Sahanājh who puts to shame Rati
with her splendour.
[Verse 8] Is she moonlight, shining like a white water lilly descending from the orb of the
moon or a torrent of juice from a fully-opened lotus abounding in nectar, or a charming
stream of fresh nectar flowing from a moon stone? That [queen]’s external brilliance grows
brighter from the abundance of gems [that she wears].
[Verse 9] She was beautiful by good fortune, in possession of chastity, skill in the arts and
flirtatious grace, possessed a penchant for generosity, and refined habits. In her abide in-
nate multitudes of virtues, far-famed, always abide, jewels with the abundant lustre of
Mt Rohaṇa.
[Verse 10] Her son named Mahmūd Khān protected the city of Māṃ[.], his feet rubbed by
the tips of the crowns of kings.
[Verse 11] And she, while the prosperous Allābadī Pātishah, son of Āhmad Sultān, gov-
erned the earth with good policy,
[Verse 12] had made a lovely tank/reservoir with plenty of excellent stairs in that best of
cities which innumerable lofty mansions adorned.
[Verse 13] Its lovely grove condemns the forest [of the gods?] as a normal forest [?] and its
drinking water considers the milk-ocean to be poison.
390 | Eva Orthmann

[Verse 14] And this [reservoir] was constructed on the 10th day of the bright fortnight in
the month Āśvina73 of the Jovian Year Raktākṣi74 in 1366 of the Śālivahana Śaka, a Monday
in the nakṣatra Śravaṇa, in the yoga (half-nakṣatra) of Dhruva (the polar star). It will last
as long as the sun and the moon are there,
[Verse 15] by Ahmad Bahmanī Bībī Sahanājh … as long as the sun and moon … delighting
for the world …

The Persian version mainly mentions Bībī Shahnāz, the mother of Maḥmūd
Khān, as the builder of the step well, records its construction date, and links the
genealogy of the Bahmanids (r. 1347–1527) with the legendary hero Bahmān and
his son Dārā. It begins with the basmala and a quotation from the Quran, and
ends with the wish that those who drink from the well will pray for the builder
Bībī Shahnāz, Maḥmūd and Aḥmad.
The Sanskrit inscription is much more detailed. After referring to Brahma,
the creator god of Hinduism, the current ruler, ʿAlā ud-Dīn Aḥmad Shāh
(r. 1436–1458 CE), is praised at length. It then provides some historical infor-
mation and mentions Aḥmad’s victory over Deva Rāi of Vijayanagara. Related to
these events, the inscription refers to a conflict within the royal family. It is too
enigmatic, however, to fully understand the conflict without prior knowledge of
the participants and events. What makes this inscription interesting is the fact
that it conveys a very different message in the two languages. While in the Per-
sian version, the building of the step well is no more than a charitable act, it has
a political dimension in Sanskrit. The victory over Deva Rāi, the averted danger
and the representation of the king as ordained by god and as so powerful that
nothing against his interest can happen – all this is mentioned only in the San-
skrit and thus conveyed only to those able to read this language. Was it meant
as a warning to those who had been involved in the intrigue and the supporters
of Deva Rāi – that is, to insiders who knew the context?
Besides the information on the political intrigue, it is worth considering the
adaptations of the religious phrases, like the invocation of Brahma and the
references to the gods Indra and Kubera, which appear only in the Sanskrit text.
Such changes are quite frequent, and in particular, the invocation of Hindu
goddesses is found in several Indic versions of inscriptions.75 Similarly, refer-
ences to the king as a conqueror of infidels and shelter of Islam76 are often omit-

||
73 The seventh lunar month of the Hindu calendar.
74 The fifty-eighth year of the Jupiter cycle of sixty years.
75 Nos 10, 24, 32, 60, 61.
76 Nos 4, 75.
 Bilingual Inscriptions from India | 391

ted from the Indic version, just as references to libations are from the Per-
sian/Arabic text.77 We can again see here an adaptation to accepted conventions.

6.5 Inscription from Gujarat about a religious endowment

Another interesting adaptation is related to curses and blessings, as already
seen in some of the inscriptions discussed so far. These usually pertain to ad-
herence to regulations and rules. Curses and blessings are used in all the lan-
guages considered here, often not translated verbatim, but rather adjusted to
the cultural context. One good example comes from Gujarat; the inscription
provides information on a religious endowment (waqf).78

Persian version:

‫( ﺩﺭ ﺗﺎﺭﻳﺦ ﺭﻭﺯ ﺳﻪ ﺷﻨﺒﻪ ﻏﺮﻩ ﻣﺤﺮﻡ ﺳﻨﻪ ﺍﺭﺑﻊ ﻭ ﺳﺒﻌﻤﺎﺋﻪ‬1)

‫( ﺍﺯ ﺻﺪﻗﺎﺕ ﺑﺎﺩﺷﺎﻩ )=ﭘﺎﺩﺷﺎﻩ( ﻋﺎﺩﻝ ﺭﺍی ﮐﺮﻧﺪﻳﻮ ﺯﻳﺪ ﻣﻠﮑﻪ‬2)
‫( ﻭ ﻣﻠﮏ ﺍﻻﮐﺎﺑﺮ ﺑﻠﭽﻖ )ﻭ( ﻣﻠﮏ ﺍﻻﻣﺮﺍ ﺷﺎﺩی ﻋﺰ ﻧﺼﺮﻫﻢ‬3)
‫( ﻭﻗﻒ ﮐﺮﺩ ﺑﺮﺍی ﻣﺴﺠﺪ ﺟﺎﻣﻊ ﮐﻨﺒﺎﻳﺔ ﺗﺎﺝ ﺍﻟﺪﻳﻦ‬4)
[‫( ﺣﺴﻦ ﺑﻦ ﻭﺯﻳﺮ ﺍﻟﻮﺯﺭﺍ ﻧﺠﻢ ﺍﻟﺪﻧﻴﺎ ﻭﺍﻟﺪﻳﻦ ﻣﺤﻤﻮﺩ ﺧﻂﻴ]ﺏ‬5)
...‫( ﺩﻩ ﺳﺎﻧﺒﺎ )=ﺳﺎﻧﭙﺎ( ﺩﺭ ﺯﻣﻴﻦ ﮐﻨﻢ ﺑﺎ ﺟﻤﻠﻪ ﺣﺪﻭﺩ‬6)
...‫( ﻭ ﺩﺭ ﺟﻨﺎﻥ )ﺟﻬﺎﺕ؟( ﺍﻧﭽﻪ ﺩﺭ ﺗﻌﻠﻖ ﺍﻳﻦ ﺩﻫﺴﺖ ﺍﺯ ﺣﺎﺻﻞ ﻭ‬7)
‫ ﻣﻪ؟ ﺑﺎﻳﺪ ﻣﺴﻠﻤﺎﻧﺎﻥ‬... (8)
‫ ﮐﻨﻨﺪ ﺗﺎ ﻫﻤﻪ ﺩﺭ‬... (9)
‫ ﻳﺎ؟ ﺗﻌﺮﺽ ﻭ ﺣﮑﻢ ﮐﻨﺪ ﻭ ﻧﻘﺾ‬... (10)
‫ﺩﻫﺪ ﮔﺮ ﺑﮕﺮﺩﺍﻧﺪ ﺩﺭ ﻟﻌﻨﺖ‬.....‫( ﺍﻳﻦ ﻭﻗﻒ ﻭﺻﻴﺖ ؟‬11)
‫( ﺧﺪﺍ ﻭ ﺭﺳﻮﻝ ﺑﺎﺷﺪ ﻭ ﻟﻌﻨﺔ ﺍﻟﻼﻋﻨﻴﻦ ﻭ ﺍﻟﻤﻼﺋﮑﺔ ﺍﺟﻤﻌﻴﻦ‬12)

Translation of the Persian version:

(1) On the day of Tuesday, the first of Muḥarram,79 of the year four and seven hundred,
(2) from the estates dedicated to the pious use (ṣadaqāt) of the just king Rāi Karṇ Dev, may
his kingdom increase,
(3) and Malik ul-Akābir Balchaq (and) Malik ul-ʿUmarā Shadī, may their victory be glorified,
(4–5) Tāj ud-Dīn Ḥasan, son of Wazīr ul-Wuzarā Najm ud-Dunyā wa d-Dīn, made in laud-
able words an endowment for the Friday mosque of Kambāya,
(6) in the village of Sāmbā, in the land of Kānam, with all its boundaries
(7) and whatever in the vicinities is related to this village and the income and …
(8) … the Muslims should …

||
77 Nos 4, 30.
78 No. 6. See also Sheikh 2014, 191–192.
79 First month of the Islamic year.
392 | Eva Orthmann

(9) … do so that they all, in

(10) … should object to or pass any judgement against it or seek to breach
(11) this endowment (and) attestation, or change it, will incur the curse
(12) of God and (His) Prophet and (also) the curse of the cursers and angels, all of them.

Transliteration of the Sanskrit:

(1) [śrī saṁ] vata 1360 varṣe bhādravā vadi

(2) [dvitīya-bhau] me Mahārājādhirāja śrī Ka-
(3) [rṇadeva]… Maliṣa śrī Baicaka Maliṣa śrī
(4) [Śādī]…… va śrī Hasaṇa prabhrtibhiḥ
(5) [Kāname Sāṃ] pâbhidhāna-grāmo ’yaṃ Staṃ
(6) [bhatīrthe jā] me misi kiyā dharmme
(7) ……[pra] dāttaḥ || samastarāṇakaiḥ
(8) [ranu maṃtavya śva] pālanīyaḥ || ठ ||

Translation of the Sanskrit:

(1) In the auspicious year saṁvat 1360, in the dark half of Bhādrapada80,
(2) [the second day, Tuesday] Mahārājādhirāja Ka
(3) [ṛna Deva Ma]lik Śrī Baḍchak, Malik Śrī
(4) [Shādī …] and Śrī Ḥasaṇa and the like,
(5–7) gave the village named Sāmpā in Kānam for religious purposes to the Stam(bhatīrtha
Jā)miʿ mosque. All the Rānās should honour and
(8) execute it. That is all.

The Sanskrit version is not only smaller in dimension, but also shorter in length,
and the extent of the waqf in particular is described much more briefly. Similar
to our first example, the Sanskrit records the older name of the city, Stambhatīr-
tha. Further, the Persian version contains several eulogies that are not included
in the Sanskrit. The most interesting difference, however, relates to the protec-
tion of the waqf. The Persian version threatens all those who do not respect the
endowment with the curses of God, the Prophet and the angels. The Sanskrit is
much more down to earth and addresses the local chiefs (rānās), asking them to
honour and execute the endowment. It is likely that the Sanskrit version was
meant especially for these local chiefs and was supposed to explain the provi-
sions of the waqf to them in a language they could understand. We can consider
the Sanskrit version a summarized translation corresponding to Sohoni’s Cate-
gory 4, but at the same time, we here see an effort to address those who were
responsible on site.

||
80 The sixth month of the Hindu calendar.
 Bilingual Inscriptions from India | 393

6.6 A bilingual inscription from Baroda

A similar local context can be seen in the bilingual inscription from the tomb of
Bābā Arjun Shāh in Petlad, Baroda, who died in 633 AH / 1236 CE.81 It is in Per-
sian and Sanskrit, and is one of a total of three inscriptions from the tomb. Both
the Persian and the Sanskrit versions are damaged. The Persian version is at the
top and occupies much more space than the Sanskrit version, which is below it
in smaller letters.

Persian version:

[‫ﺑﺴﻢ ﺍﻟﺮﺣﻤﻦ ]ﺍﻟﺮﺣﻴﻢ‬

[‫( ﺑﺎﺗﻤﺎﻡ ﺭﺳﻴﺪ ﻋﻤﺎﺭﺕ ﭼﺎﻩ ﺍﺯ ﺻﺪﻗﺎﺕ ﺧﻠ]ﻳﻔﻪ‬2).....‫ﺗﺎﺭﻳﺦ ﻣﺎﻩ ﻣﺒﺎﺭک ﺫی ﺍﻟﺤﺠﻪ ﺳﻨﻪ‬....(1)
‫( ﻭ ﺍﻟﺤﺎﺗﻢ ﻏﻴﺎﺙ ﺍﻟﺪﻧﻴﺎ ﻭ ﺍﻟﺪﻳﻦ ﻏﻮﺙ‬4)......‫( ﺍﻟﺴﻼﻁﻴﻦ ﺧﺪﺍﻳﮕﺎﻥ ﻋﺎﻟﻢ ﺑﺎﺩﺷﺎﻩ‬3)......
‫( ﻣﺪ ﷲ ﻋﻤﺮﻩ ﻭ ﺧﻠﺪ‬6)....‫( ﮐﻬﻒ ﺍﻟﺜﻘﻠﻴﻦ ﻅﻞ ﷲ ﻓﯽ ﺍﻟﺨﺎﻓﻘﻴﻦ ﺍﺑﻮ ﺍﻟﻤﻈﻔﺮ‬5).....[‫ﺍﻻﺳﻼ]ﻡ‬
‫( ﻧﻮﺭ ﷲ ﻗﺒﺮﻩ ﺻﺎﺣﺐ ﻋﻤﺎﺭﺕ‬7).....‫ﺩﻭﻟﺘﻪ ﺩﺭ ﻗﺼﺒﻪ ﭘﻴﺘﻼﻭﺩﺭ ﺟﻮﺍﺭ ﺭﻭﺿﻪ ﺷﻴﺦ ﺍﻟﻤﺸﺎﻳﺦ‬
‫( ﮐﺮﺩ ﻟﻮﺟﻪ ﷲ‬8) (‫ﺍﺿﻌﻒ ﻋﺒﺎﺩ ﷲ ﺣﺎﺟﯽ ﺍﺳﻤﻌﻴﻞ ﻋﺜﻤﺎﻥ ﺷﻴﺮﺍﺯی ﺍﻳﻦ ﻋﻤﺎﺭﺕ ﻭﻗﻒ )؟‬
‫( ﺍﺑﻮ‬9) ‫ﺗﻌﺎﻟﯽ ﻭ ﺑﻴﺴﺖ ﮐﺒﻪ ﺯﻣﻴﻦ ﺑﺠﻬﺖ ﭼﺎﻩ ﺍﺯ ﻣﻘﻄﻊ ﻗﺼﺒﻪ ﭘﻴﺘﻼﻭﺩﺭﺳﻴﺪ ﺍﻻﻣﺮﺍ ﺑﺪﺭ ﺍﻟﺪﻳﻦ‬
‫ﺩﺍﺭﺩ ﻣﺴﻠﻢ ﺍﺳﺖ ﺗﺎ ﻏﺮﻳﺐ ﻭ ﺷﻬﺮی ﺭﺍ ﺍﺯ ﺍﻳﻦ ﻭﺿﻊ ﺭﺍﺣﺘﯽ ﺭﺳﺪ ﻭﺍﺟﺐ‬....‫ﺑﮑﺮ ﺍﻣﻴﺮ ﮐﻮﻩ‬
‫( ﺑﺮ ﺍﻣﺮﺍ ﻭ ﻣﻠﻮک ﻭ ﺣﺎﮐﻤﺎﻥ ﮐﻪ ﺑﻌﺪ ﺍﺯ ﺍﻳﻦ ﺁﻳﻨﺪ ﺍﻳﻦ ﺧﻴﺮﺭﺍ ﻗﺎﻳﻢ ﺩﺍﺭﻧﺪ ﻭ ﻫﻴﭻ‬10) ‫ﺍﺳﺖ‬
‫( ﻭ ﺑﻴﺴﺖ ﮐﻨﺒﻪ ﺯﻣﻴﻦ ﮐﻪ ﺫﮐﺮ ﺭﻓﺖ ﻣﺴﻠﻢ ﺩﺍﺭﻧﺪ ﺗﺎ ﺛﻮﺍﺏ ﺩﻭ ﺟﻬﺎﻧﯽ ﺣﺎﺻﻞ‬11) ....(‫ﻟﻐﺰﺷﯽ )؟‬
‫ﮔﺮﺩﺩ ﺑﺎﻟﻨﺒﯽ ﻭ ﺁﻟﻪ ﺍﺟﻤﻌﻴﻦ‬......‫( ﻣﺤﻤﺪ ﻋﻠﻴﻪ ﺍﻟﺴﻼﻡ‬12) ‫]ﺷﻔﺎ[ﻋﺖ‬.... ‫ﮐﺮﺩﻩ ﺑﺎﺷﻨﺪ ﻭ‬

Translation of the Persian:

In the name of Allāh, the Beneficent, the Merciful.

On the … [20th] of the auspicious month of Dhū l-Ḥijja82 [723 AH], the construction of the
well was completed. [It is one] of the charitable works (ordered) by the Caliph – … of
kings, the lord of the world, the monarch … and Ḥātim, Ghiyath ud-Dunyā wa d-Dīn, the
defender of Islam … the refuge of men and demons, the shadow of God in the east and
west, Abū l-Muẓaffar … may God prolong his life and perpetuate his glory! – in the town of
Petlawad, in the vicinity of the tomb of Shaikh ul-Mashāʾikh [Arjūn Shāh], may God illu-
mine his grave. The builder, the humble servant of God Hājī Ismāʿīl ʿUthmān of Shiraz,
dedicated this building to the cause of God, be He exalted. And 20 kubhas of land in the
town of Petlawad, from the fief of Sayyid ul-Umarāʾ Badr ud-Dīn Abu Bakr, the chief of the
mountain … are granted for the maintenance of the well: so that the stranger and the peo-
ple of the town may gain relief through this charitable institution. It is incumbent upon
the chiefs, kings and governors who succeed in the future to protect this gift and not fail …
and (also) to preserve the grant of 20 kubhas of land mentioned above, so that they may
secure rewards in both worlds and … intercession of Muḥammad, may peace be upon him!
… through the Prophet and all his descendants.

||
81 No. 8. See also Sheikh 2014, 201–202.
82 The twelfth month of the Islamic calender.
394 | Eva Orthmann

Transliteration of the Sanskrit:

(1) oṁ saṁvat 1380 varṣe pauṣa-vadi 7 duliheja chaṁ 20

(2) [bh]omav adyeha yoginīpurâdhiṣṭhita-mahārājâdhirāja-
(3) śrīmat-suratrāṇa-śrī-gayāsadīna-vijiya-rājye tan-niyukt …
(4) ṇahillapattana-śri-dīvan-ādeśena peṭilāpadra-maṇḍala-karaṇi
(5) śrī-badara-dīna avubaka ahamad-amīra-koha-pratipattau petila…
(6) * ṣa-śrī-a(rjuna) ghori sannidhau skambha-tīrtha-vāstavyen=[el
(7) ... [i]la-ūshamāna-sīrājena jirṇa-uddhāra-kupa vahā(*)e
(8) ... k(e)na-āghāṭe kṣepita bhūmi kubha 20 viṁ [śati ka] ...
(9) ... sthale śrī-thakkurai[ḥ] pālanīyāni
pārasī likhitaṁ

Translation of the Sanskrit:

Oṁ! In (Vikrama) saṁvat (year) 1380, on the 7th day of the dark (fortnight) of Pauṣa,83 on
the 20th (day) of the lunar month Duliheja (Dhū l-Ḥijja), on a Tuesday – on this day in the
victorious reign of the glorious sulṭān, the illustrious Gayāsadīna (Ghiyath ud-Dīn), the
paramount king of great kings, by the order of the Diwan at Aṇahilapaṭṭana … appointed
by him (i.e. the sulṭān), under the dispensation of the glorious Badaradīna Avubaka
Ahamada Amīru-koha (Badr ud-Dīn Abu Bakr Amīr-i Kūh), agent in the circle of Petilāpa-
dra, in the proximity of the revered Arjuna-ghori, (at) Petila … 20 kubhas of land marked
off with boundary were given by (Ismāila?) Ushmana Siraj (Ismāʿīl ʿUthmān Shirazī), an
inhabitant of Skambhatīrtha … repaired well…. The illustrious thakkuras should protect
(these gifts). Written in Persian.

The two inscriptions cannot be compared in full due to some illegible parts in
the Sanskrit. They bear many similarities, but also differ: all the religious
phrases of the Persian text have been eliminated from the Sanskrit version,
which instead begins with ‘Oṁ’. At the end, the Persian version calls upon no-
table figures, kings and rulers to preserve the land grant. In the Sanskrit ver-
sion, this request is addressed to the thakkuras, a Sanskrit title for members of a
landowning caste. One may wonder if the Sanskrit version of the inscription
was addressing them: were the thakkuras supposed to read the text and obey it?
It is the only inscription that ends in Sanskrit with a reference to the Persian text
immediately above it. Perhaps the composer of the Sanskrit version did not
know that the two versions would be placed on the same slab – or he wanted to
indicate that the original text was in Persian, and that the Sanskrit is a transla-
tion of it.

||
83 The tenth month of the Hindu calendar.
 Bilingual Inscriptions from India | 395

6.7 The Veraval inscription of Chaulukya-Vaghela Arjuna

Many of the differences so far observed are also found in the famous bilin-
gual inscription from Somanātha Pāṭan in Gujarat; it is in Sanskrit and Ara-
bic and is about a Muslim shipowner and merchant, Nūr ud-Dīn Fīrūz from
Hormuz, who acquired land next to the Somnath temple to build a mosque
(no. 4). The lengthy text specifies the names of the people involved in the
property transaction and the establishment of the waqf related to the build-
ing of the mosque and its maintenance. Since this text has been edited and
translated several times and also discussed in the literature,84 I will focus
here on the bilingual aspects of the document. Unfortunately, the original
location of the inscription and the relative position of the Arabic and the
Sanskrit text are unknown. Both inscriptions have been displaced: the San-
skrit one is currently on a wall of the temple of Harshada Mātā in Verāval,
next to Somnath, and the Arabic inscription is on the façade of the Qāḍī
mosque in Somnath. The two texts were written on two separate stones, and
we can only assume that they had originally been placed side by side or –
somewhat less likely, due to access and readability – one atop the other.85
Both scripts were engraved in black granite, with no ornamentation. There is
no obvious hierarchical difference between the two inscriptions, but the date
of the Sanskrit inscription is two months earlier (25 May 1264) than that of the
Arabic inscription (23 July 1264), thus making it the primary one. Given that
they are related to the building of a mosque, it is highly probable that the
inscriptions were originally placed on a wall of said mosque. They obviously
did not have a decorative function, but an informative character. We may
assume that the addressees were those who had supported the construction
of the mosque. Since the majority of the persons mentioned in the text were
Hindus, one may wonder if the inscriptions were placed outside, perhaps on
the façade – similar to the current placement of the Arabic inscription in
another mosque – or close to one of the doors.
The Arabic version is not a translation of the Sanskrit. The content of the
two inscriptions overlaps, but there are also significant differences: in the
Sanskrit version, the details of the waqf are enumerated, both its sources and

||
84 Hultzsch 1882; Desai 1961, 10–15 and pl. II b; Sircar 1961–1962, 141–150; Chattopadhyaya
1998, 70–78, Thapar 2008, 88–99. The Arabic text is edited in Desai 1961, the Sanskrit text (in
Nāgarī script) in Hultzsch 1882 and (in transliteration) Sircar 1961–1962. It is too long to repro-
duce here.
85 The size of the Arabic inscription, which is not fully preserved, is 17″ × 25″ (43.18 × 63.5 cm);
the size of the Sanskrit inscription is nowhere properly indicated.
396 | Eva Orthmann

intended use. The Sanskrit version furthermore contains four different dating
systems, the first being the Hijrī year, referred to as year 662 of ‘rasūla Ma-
hammada’ (rasūl Muḥammad, i.e. the Prophet Muḥammad). This inscription
also mentions that Fīrūz performed a libation with water, a ceremonial act
rooted in Indian customs. The inscription generally shows a remarkable
transfer of Islamic notions to the Hindu sphere, searching for equivalents
within the other’s religious concepts. The mosque is thus called a dhar-
masthāna, and ‘Śrī Viśvanātha’ most probably denotes Allāh. Usually, this
name is applied to Śiva, the god of the Somnath temple. The activities to
which the waqf is dedicated are described with a combination of terms used
in the context of temple rituals – like pūjā (reverence, worship) – and Arabic
religious terms.86 The Arabic inscription is much less detailed with regard to
the waqf. Instead, we find long eulogies related to Fīrūz’s father and Fīrūz
himself. Furthermore, the inscription expresses the hope that Somnath might
become a city of Islam, and that infidelity and idols might be banned from it.
Since all other details of this bilingual inscription point to a friendly interac-
tion between Hindus and Muslims, we may assume that this part of the Ara-
bic inscription is rather formulaic and follows standard patterns.87 Interest-
ingly, the ruler addressed in the two inscriptions is also not the same: while
the Sanskrit version refers to Arjunadeva, the Chaulukya-Vāghela king who
ruled in Gujarat (r. 1261–1274 CE), the Arabic version instead mentions
Maḥmūd b. Aḥmad, most probably the Delhi sultan Nāṣir ud-Dīn Maḥmūd
Shāh (r. 1246–1265 CE).88 These two versions of the inscription can again best
be explained by the need to address the different expectations of the pre-
sumed readers, and to fulfil different needs. The eulogies for Fīrūz and his
father, for example, might have been meant to raise his status in the eyes of
his co-religionists, and fulfilled the conventions of this genre. On the other
hand, the need to elaborate the details of the waqf and its use might be relat-
ed to the fact that it was most likely temple property that was purchased for
the endowment. The Hindu authorities thus may have wanted the details of
this endowment to be disclosed and promulgated.

||
86 Chattopadhyaya 1998, 74–76.
87 Thapar 2008, 95.
88 The identification is not entirely clear. Nāṣir ud-Dīn Maḥmūd Shāh was a son or grandson
of the Mamluk ruler Shams ud-Dīn Iltutmish (r. 1211–1236). It is not evident why he is called
Maḥmūd b. Aḥmad here.
 Bilingual Inscriptions from India | 397

6.8 Two Arabic-Sanskrit inscriptions from Asīrgarh and

Burhānpūr
The two inscriptions from the Friday mosques (Jāmaʿ Masjid) in Asīrgarh and
Burhānpūr89 in today’s Madhya Pradesh resemble the inscription from Somnath
insofar as they also translate central notions of Islam into a Sanskrit terminolo-
gy that Hindus could understand. However, the design and location of these
inscriptions are very different, causing us to wonder who the addressees were.
Both mosques were erected by the Fārūqī dynasty (r. 1382–1601 CE) and date
from 992 AH / 1584 CE and 997 AH / 1589 CE, respectively. The inscriptions are in
Arabic and Sanskrit and are placed inside the prayer hall on the qibla wall. The
qibla wall is the wall towards which Muslims turn for the ritual prayer; it con-
tains one or more prayer niches (miḥrāb) that indicate the direction of Mecca. It
is thus the most important wall in a mosque. The inscription in Asīrgarh is
placed in the northernmost miḥrāb, and the one in Burhānpūr in the southern-
most miḥrāb. Although they are not placed over the central miḥrāb, their posi-
tion is very prominent and quite unusual. All the inscriptions are in relief.
The Asīrgarh inscription consists of two lines in Arabic and three lines in
Sanskrit. Although the text is longer in Sanskrit, the size of the two inscriptions
is equal; this is achieved by using slightly smaller letters in Nāgarī. The Arabic
dominates insofar as it is on the top, but the Nāgarī might have been easier to
read since it was a little bit lower and in a very clear and easily readable script.

Arabic text:

‫( ﺑﺴﻢ ﷲ ﺍﻟﺮﺣﻤﻦ ﺍﻟﺮﺣﻴﻢ ﺍﻥ ﺍﻟﻤﺴﺎﺟﺪ ﻓﻼ ﺗﺪﻋﻮﺍ ﻣﻊ ﷲ ﺍﺣﺪﺍ ﻗﺎﻝ ﻋﻠﻴﻪ ﺍﻟﺴﻼﻡ ﻣﻦ ﺑﻨﯽ‬1)
‫ﻣﺴﺠﺪﺍ ﻟﻮ ﮐﻤﻔﺤﺺ ﻗﻄﺎﺓ ﺑﻨﯽ ﷲ ﻟﻪ ﺑﻴﺘﺎ ﻓﯽ ﺍﻟﺠﻨﺔ ﺍﻣﺮ ﺑﺒﻨﺎء ﻫﺬﺍ ﺍﻟﻤﺴﺠﺪ ﺍﻟﻤﺒﺎﺭک ﺍﻟﺬی ﻫﻮ‬
‫ﻣﻦ ﺣﺴﻨﺎﺕ ﺍﻟﺰﻣﺎﻥ ﻭ ﮐﺎﻟﺸﺎﻣﺔ ﻋﻠﯽ ﻭﺟﻪ ﺍﻟﺤﺴﺎﻥ ﺳﻴﺪﻧﺎ ﻭ ﻣﻮﻻﻧﺎ ﺍﻟﺴﻠﻄﺎﻥ ﺍﺑﻦ‬
‫( ﺍﻟﺴﻠﻄﺎﻥ ﺍﺑﻦ ﺍﻟﺴﻠﻄﺎﻥ ﻋﺎﺩﻟﺸﺎﻩ ﺑﻦ ﻣﺒﺎﺭﮐﺸﺎﻩ ﺑﻦ ﻋﺎﺩﻟﺸﺎﻩ ﺍﻟﻔﺎﺭﻭﻗﯽ ﺍﻟﻌﻤﺮی ﺍﻟﻌﺪﻭی ﺧﻠﺪ ﷲ‬2)
‫ﺗﻌﺎﻟﯽ ﻣﻠﮑﻪ ﻭ ﺳﻠﻄﺎﻧﻪ ﻭ ﺍﻓﺎﺽ ﻋﻠﯽ ﺍﻟﻌﺎﻟﻤﻴﻦ ﺑﺮﻩ ﻭ ﺍﺣﺴﺎﻧﻪ ﺧﺎﻟﺼﺎ ﻣﺨﻠﺼﺎ ﻟﻮﺟﻪ ﷲ ﺍﻟﮑﺮﻳﻢ ﻭ‬
‫ﻁﻠﺒﺎ ﻟﻤﺮﺿﺎﺗﻪ ﺍﻟﺠﺴﻴﻢ ﺗﻘﺒﻞ ﷲ ﻣﻨﻪ ﺻﺎﻟﺢ ﺍﻻﻋﻤﺎﻝ ﺑﻤﺤﻤﺪ ﻭ ﺻﺤﺒﻪ ﻭ ﺍﻵﻝ ﻭ ﮐﺎﻥ ﺍﺑﺘﺪﺍء ﺑﻨﺎﻳﻪ‬
‫ ﻭ‬٩٩٢ ‫( ﻓﯽ ﺍﻳﺎﻡ ﺍﻟﺴﻠﻄﻨﺔ ﺍﻟﻔﺎﺭﻭﻗﻴﺔ ﺍﻟﻌﺎﺩﻟﻴﺔ ﻓﯽ ﻳﻮﻡ ﺍﻟﺴﺒﺖ ﺭﺍﺑﻊ ﺷﻬﺮ ﺷﻌﺒﺎﻥ ﻓﯽ ﺳﻨﺔ‬sic)
‫ﺍﺗﻤﺎﻣﻪ ﻓﯽ ﻳﻮﻡ‬

Translation of the Arabic text:

In the name of Allāh, the Beneficent, the Merciful. ‘And the places of worship are only for
Allāh, so pray not unto anyone along with Allāh’.90 And, says the Prophet, blessings of

||
89 Nos 44 and 46.
90 Quran 72:18, tr. Pickthall 1930.
398 | Eva Orthmann

God be upon him, ‘whoever builds a mosque[;] be it small as the nest of the sandgrouse,91
God will build him a house in Paradise’. The construction of this mosque, which is one of
the meritorious acts of the age and like a mole on a beautiful face, was ordered by our lord
and master the sulṭān, son of the Sulṭān ʿĀdil Shāh, son of Mubārak Shāh, son of ʿĀdil
Shāh, al-Fārūqī l-ʿUmarī l-ʿAdawī – may God perpetuate his kingdom and sovereignty,
and spread his goodness and munificence all over the world – purely and sincerely for the
sake of Allāh the Merciful and with the object of obtaining His great pleasure. May Allāh
accept his pious actions through the holy Prophet and his companions and descendants!
Its construction was commenced in the days of the Fārūqyya ʿĀdiliyya Kingdom, on Sat-
urday, the fourth of the month of Shāʿbān,92 in 992 AH, and was completed on the day …

Transliteration of the Sanskrit text:

(1) śrī kartṛpurūṣāya namaḥ | guṇātmane nirguṇāya vyaktāvyaktasvarūpiṇe cidānandāt-

mane nityaṃ viśvādhārayate namaḥ | 1 | candrārkatārāgaṃgādi* tiṣṭhanti gagane bhuvi
tāvat phārukivaṃśo ’sau*
(2) vimalo bhuvi tiṣṭhatu | 2 | *kṣaṃdrārkka-tārāgaṃgādi śrīmatphāruṣikulo hita-prīta-
pratāpa-dinakara-mitra-jana-cakora-ānandakara-pūrṇacandro (?) nirguṇe sacintana*pa-
rāyaṇapātaśāha śrī śrī Ādilaśāha bina Mubārakhaśāha bina
(3) Ādilaśāha vijayarājya śrīṣānadeśa *adhipateḥ || śrīvikramādityasamayātīta saṃvat
1641 varṣe śālavāhana (śālivāhana) kṛta śāke 1506 prakṛmine(?)* śrāvaṇama* śukala (śukra)
paṣye (pakṣye) tithi 6 śani-dine citrā-naṣya (nakṣatre)

Translation of the Sanskrit text:

Let obeisance be paid to the creator of the world, the possessor of all qualities and yet
destitute of them, manifest yet hidden, inherent in citti (mind) and ānanda (happi-
ness), eternal, and upholder of the universe. May this pure Phāruki (Fārūqī) family
endure on the earth as long as the sun, the moon, the stars, the Ganges and the like
remain on heaven and earth. (Glorious) be the royal fortune of the pādshāh, ʿĀdil Shāh
of the family of Phārushi,93 son of Mubārak Shāh, son of ʿĀdil Shāh, the king of
Khāndeś, (who is beneficial) to his people like the heat (of the sun)[or: the king’s
splendour] is beneficial and beloved to the lotuses, and just as the full moon makes
delight for the cakora birds, (and who is always) absorbed in meditation upon the Su-
preme Being. (Written) on Saturday, the 6th of the bright half of the early (?) Śrāvaṇa94
in the Citrā nakṣatra95 of the Vikrama saṁvat, 1641, (corresponding to) 1506 of the
Śālivāhana Śaka era.

||
91 Verbatim translation; often, it is translated as ‘the nest of a sparrow’.
92 The eighth month of the Islamic calendar.
93 ष instead of क, probably a mistake. See Ḳuraishi 1925–1926b, 2, n. 1.
94 The fifth lunar month of the Hindu calendar.
95 That is, the fourteenth nakṣatra or lunar mansion according to Hindu astronomy.
 Bilingual Inscriptions from India | 399

The religious phrases from the Quran and Ḥadīth have not been translated into
Sanskrit, but replaced by a call to obey God, who is then qualified by a number
of characteristics. The two terms citti and ānanda could be taken from Advaita-
Vedantic concepts. The wishes for the dynasty to last long have also been ex-
pressed with images common in a Hindu environment. The king’s kindness
towards his people is compared to the sun’s kindness towards the lotus, and he
is also considered a delight to his people, similar to the moon. He is said to be in
constant meditation upon the Supreme Being.
In the Burhānpūr mosque, the design of both inscriptions is more sophis-
ticated. The Arabic inscription consists of three lines. They are at the top, but
due to the curvature of the arch, there is much less space for them; the first
line in particular is very short. The letters are very elongated, so the single
lines are significantly higher than the lines in Nāgarī below. The Sanskrit
inscription is considerably longer: it not only fills six lines, but these lines
also have more space, since they are less affected by the curvature of the
arch. The śirorekhā or headline that is drawn above the Nāgarī letters is so
thick that the vowel signs are not above it, but incised into it. Though this
visual effect, the vowel signs are more eye-catching. Most remarkably, in the
middle of the third line from the top, a sign in the shape of the tashdīd ( ◌ ّ )
has been added. The tashdīd is used in Arabic to indicate the gemination of a
consonant. No such sign is used or needed in Nāgarī. I therefore postulate
that through the graphic design of the Nāgarī text, an assimilation to the
Arabic script was attempted, with the tashdīd as the most obvious element,
but also the eye-catching vowel signs as the counterpart of the Arabic vowel
signs. The script would thus be the graphic realization of what is happening
in the text as well – a translation of Islamic concepts into terms and notions
that were intelligible for Hindus.
The text of the Arabic inscription is very close to the one in Asīrgarh. The
only major difference is that the genealogy of ʿĀdil Shāh is mentioned. The ge-
nealogy is likewise mentioned in the Sanskrit version. The translation of Islamic
concepts is similar to that of Asīrgarh, but the terminology is not exactly the
same. The main factor in the length of the Sanskrit inscription is the very elabo-
rate chronological data, indicating not only the date, but also the exact time the
mosque began to be constructed.
400 | Eva Orthmann

Fig. 1: Sanskrit-Arabic inscription from the qibla wall of the Jāmaʿ Masjid in Burhanpur, ʿĀdil
Shāhī dynasty, 1588–1589 (Arabic inscription) / 1590 (Sanskrit inscription); public domain.

Transliteration of the Sanskrit, last section:

L. 5 virjayaṃte bhūpālacūḍāmaṇiḥ || 5 || svasti śrī saṃvat 1646 varṣe śākre 1511 virodhi-
saṃvatsare pauṣamāse śukl(v?)apakṣe 10 ghaṭi 23 sahaikā-daśyāṃ tithau some [kṛ?]ttikā-
ghaṭī 33 rāha rohi -
L. 6 ṇayāṃ śubha ghaṭī 42 yoge vaṇijakaraṇesmin dine rātrigataghaṭī 11 samayo kanyālagna
śrīmubārakhaśāhasutaśrī 7 edalaśāharājño masītiriyaṃ nirmitā svadharmapālanārthaṃ

Translation of the Sanskrit, last section:96

Hail! Prosperity! This mosque was built by the king Śrī Ādil Shāh, son of the illustrious
Mubārakh Shāh, for fostering his own religion97, in saṁvat (year) 1646, Śaka year 1511, in
the year Virodhin,98 in the month of Pauṣa, in the bright fortnight, on the 10th tithi (lunar
day), 23 ghaṭīs before the 11th tithi, on Monday, in the Kṛttikā, 33 ghaṭīs before Rohīṇī,99 in
the Śubha yoga at 42 ghaṭīs, in the Vaṇija karaṇa (lunar half-day), at the time when 11
ghaṭīs of the night on this day had passed and in the Kanyā lagna (entrance of the sun in
the sign of Virgo).

||
96 I would like to express my sincere thanks to Prof. Jost Gippert, who has corrected the read-
ing of the Sanskrit calendrical and astronomical information.
97 Not transliterated.
98 The twenty-fifth year in the Jupiter cycle of sixty years.
99 Kṛttikā and Rohiṇī are the third and the fourth nakṣatras.
 Bilingual Inscriptions from India | 401

This exact information on the construction date was related to astrological cal-
culations for finding the best such moment, the so-called catarchai. Since the
exact moment is only indicated in Sanskrit, the calculations were obviously
done by an Indian astronomer or astrologer.
By virtue of their length, their position on the miḥrāb wall, and their at-
tempt to adapt the religious content to a Hindu context, these two inscriptions
are very exceptional. Who were the addressees: Muslim converts who knew
Sanskrit? Muslim converts who did not know Sanskrit, but should be accommo-
dated with a familiar script? Local notables and Brahmans who came to visit the
mosque, without being Muslims themselves? And who translated the text? At
least in the Asīrgarh version, there are some mistakes – were they due to the
translator, or rather the mason? Was the inscription produced by the same ma-
son who also wrote the Arabic one, or were different masons employed? And
why was the astrological information only given in Sanskrit – was it considered
inappropriate to display it in Arabic on a religious building?

6.9 A bilingual inscription from Bilara in the Jodhpur district

The next inscription to be discussed is interesting because of its design. The
inscription comes from Bilara in the Jodhpur district and was prepared in the
time of Shāh Jahān; it is in Persian and a local dialect, written in Nāgarī (no. 62).
The text in Nāgarī is much larger and therefore more prominent than the Per-
sian version. It refers to the construction of a tomb for the rulers of Jodhpur and
of a well and indicates the cost of building them. This inscription is remarkable
insofar as the Nāgarī text on top is divided into two halves by a T-shaped central
field. This field is filled with Persian text, with the basmala in the upper bar of
the T, and the name of the ruler, Shāh Jahān, and the year in the vertical bar. In
the same hand, the main Persian text is placed below the Nāgarī inscription.

Persian text:

‫( ﻻ ﺍﻟﻪ )ﺍﻻ( ﷲ ﻣﺤﻤﺪ ﺭﺳﻮﻝ ﷲ‬1)

‫( ﺷﺎﻩ ﺷﻬﺎ‬2)
‫( ﺏ ﺍﻟﺪﻳﻦ‬3)
‫( ﻗﺮﺍﻧﯽ ﺷﺎﻩ‬4)
‫( ﺟﻬﺎﻥ ﺑﺎﺩ‬5)
‫( ﺷﺎﻩ ﻏﺎﺯی‬6)
١٠٤٩ ‫( ﺳﻨﻪ‬7)
‫( ﻣﺎﻩ ﻣﺤﺮﻡ‬8)
402 | Eva Orthmann

[‫( ﺗﺎﺭﻳﺦ ﻏﺮﻩ ﻣﺎﻩ ﻣﺤﺮﻡ ﻳﻮﻡ ﭘﻨﺠﺸﻨﺒﻪ ﺑﺎﻣﺮ ﻣﻬﺎﺭﺍﺝ ﺭﺍﺟﻪ ﺟﺴﻮﻧﺖ ﺳﻨﮕﻪ ﺑﺤﮑﻢ ﭼﻮ]ﺩﻫﺮی؟‬9)
‫ ﻭ ﭼﺸﻤﻪ ﮔﻨﮕﺎ ﺭﻭﭘﻴﻪ‬٢٥٠١ ‫ﺭﻭﭘﻴﻪ‬.....‫ ﻳﺎ ﺩﺍﺷﺖ‬.....‫؟[ ﻧﻤﻮﺩ ﭼﻮﺩﻫﺮی‬...] ‫ﻟﮑﻬﻤﻴﺪﺍﺱ ﺭﺍﺱ‬
‫ ﺧﺮچ‬١٥٠١
‫( ﺷﺪ ﺳﻨﮓ ﺗﺮﺍﺵ ﻻﺩ ﻣﺤﻤﺪ ﻭ ﺩﻭﺩﺍ ﮐﺎﺭ ﻓﺮﻣﺎ ﻭﻳﺘﻬﺎ ﺑﻦ ﺩﺍﺭﮐﺎ ]ﺩﻭﺍﺭﮐﺎ؟[ ﺷﻌﺮ ﺑﺎﺯ‬10)
‫ ﺑﺎﺷﺪ ﺩﺭ ﮐﻨﺎﺭ ﮐﺎﺗﺐ ﺍﻟﻌﺒﺪ‬..... ‫ﺻﺮﻑ ﺭﺍ‬.....‫ﺑﻌﺪ[ ﻣﺮﺩﻥ ﻧﺎﻡ ﻧﮑﻮ ﺩﺭ ﺟﻬﺎﻥ ﺁﻳﺪ ﺑﮑﺎﺭ‬....]
‫ﻋﺎﺭﻑ ﻣﺤﻤﺪ‬

Translation of the Persian text:

There is no God but Allāh. Muḥammad is the Prophet of Allāh. Shāh Shihāb ud-Dīn
[Ṣāḥib] Qirān [-i Thā]nī, Shāh Jahān Pādshāh Ghāzī. Year (AH) 1049, month of Muḥarram.
On the first of the month of Muḥarram, the day of Thursday, under the orders of Mahārāj
Rāja Jaswant Singh (and) at the instance of Cho(udhary?) Lakhmī Dās, Choudharī … con-
structed (?) the tomb… 2501 rupees were spent (on …), and on (?) Gangā spring, 1501 ru-
pees. The stone carvers are Lād Muḥammad and Dūdā, and the supervisor Vīthā, son of
Dārkā (Dwārkā?) (Verse:) After death, a good name will remain in the world. Only … will
be in one’s embrace (i.e. possession). The writer is the servant ʿĀrif Muḥammad.

Although the Nāgarī has not been fully read, it very much looks as if the Persian
script has been written over the Nāgarī, both in the middle of the first line and
the central part, and possibly also at the bottom of the slab. In the central part,
single words are spread over two lines due to the shortness of the line. Since
Persian words are usually not hyphenated, this clearly indicates that no Persian
text was planned there. Given the fact that the Persian or Arabic version is typi-
cally placed on top, and that the text has been arranged in such an unusual
way, I assume that the original version of this inscription was in Nāgarī only,
and that the Persian version was added later. The superimposition of some Per-
sian text, and especially the placement of the basmala in the centre of the upper
line, might have been a means to symbolically restore the normal order. The
dates would contradict this interpretation, since the saṁvat date of the Nāgarī
text – Āshāḍhā100 1696 – comes after Muḥarram 1049, the date indicated in the
Persian version. But this could also be a deliberate misdating to give priority to
the Persian text.

6.10 A bilingual inscription from the lower Swat valley

The last inscription to be considered also refers to the erection of a tomb and is
one of the oldest inscriptions in the corpus (no. 3). It allegedly comes from
Zalamkot, a place in the lower Swat valley. The languages used are Persian in

||
100 The fourth lunar month of the Hindu calendar.
 Bilingual Inscriptions from India | 403

Arabic script and Sanskrit in Śāradā script. The two texts are not of equal status:
by its length, the line spacing and its placement at the top, the Persian text is
dominant, while the Sanskrit text is secondary. On the left, it is aligned with the
Persian inscription, while on the right, it does not adhere to the space limit. This
is a consequence of the different writing directions of Persian and Sanskrit, but
also of the limited space available at the bottom of the limestone slab.

Persian version:

‫( ﺑﺴﻢ ﷲ ﺍﻟﺤﻤﻦ ﺍﻟﺮﺣﻴﻢ‬1)

‫( ﻻ ﺍﻟﻪ ﺍﻻ ﷲ ﻣﺤﻤﺪ ﺭﺳﻮﻝ ﷲ‬2)
‫( ﺑﻨﺎ ﮐﺮﺩ ]ﺍﻳﻦ[ ﻣﺮﮐﺪﺭﺍ ﺍﻣﻴﺮ ﺟﻠﻴﻞ ﺍﻣﻴﺮ ﻁﻮﺱ‬3)
‫( ﺍﺭﺳﻼﻥ ﺍﻟﺠﺎﺫﺏ ﺍﻁﺎﻝ ﷲ ﺑﻘﺎﻩ ﺗﻤﺎﻡ ﮐﺮﺩ‬4)
‫( ﺳﭙﻬﺒﮏ ﺧﻠﻴﻞ ﺑﮏ ﺍﺑﻦ ﮐﻮﺗﻮﺍﻝ ﺍﺣﻤﺪ ]ﺍ[ﻟﺒﻐﺮ‬5)
‫( ﺍﺩﺍﻡ ﷲ ﺩﻭﻟﺘﻪ ﻓﯽ ﻣﺎﻩ ﺫﻭ ﺍﻟﻘﻌﺪﻩ‬6)
‫( ﺳﺎﻝ ﻣﺮ ﭼﻬﺎﺭ ﺻﺪ ﻳﮏ ﺑﻮﺩ‬7)

Translation of the Persian:

(1) In the name of Allāh, the Beneficent, the Merciful.

(2) There is no God but Allāh; Muḥammad is the Prophet of Allāh.
(3) This tomb was founded by the grand amīr, the governor of Ṭūs,
(4) Arsalān ul-Jādhib – may Allāh prolong his life. It was completed by
(5) the sipahbak (i.e. commander) Khalīlbak, son of the Kotwāl Aḥmad ul-Baghr –
(6) may Allāh perpetuate his fortune – in the month of Dhū l-Qaʾada;101
(7) the current (marr) year was 401.

No translation or transliteration of the Sanskrit text have been provided.

By the dates indicated in both languages, the inscription can be dated to the
year 401 AH / 1011 CE. It is thus one of the oldest clearly datable attestations of
Persian in Arabic script. Geographically, it can be related to the other earliest
testimony of Persian in Arabic script, which has been found in today’s Afghani-
stan.102 It thus provides further evidence of the use of the Arabic script in the
eastern parts of Iran up to the border region with India. The early date of the
inscription is probably the reason for some orthographic peculiarities, especial-
ly the spelling of markad instead of marqad (‘tomb’), although this variant is not

||
101 The eleventh month of the Islamic calendar.
102 There in one older document, in the so-called Afghan Genizah; see Haim 2019. Some even
older Persian notes are found in Quranic booklets; these were done by a person from Tus. See
Kariminia 1396 AH / 2018 CE.
404 | Eva Orthmann

attested to in the contemporary legal documents from Bamiyan.103 The inscrip-

tion consists of seven lines in Persian in Kufi script, and three lines of Sanskrit
in Śāradā script. Unfortunately, only the Persian inscription has been tran-
scribed and translated in full in the publication. It contains the Islamic creed
(shahāda), the name of the governor of Tus who ordered the tomb to be built,
and the name of the commander who had it completed. The name of the buried
is not mentioned. The Persian text ends with the date; in the Sanskrit text, the
date comes at the beginning and is indicated more precisely than in the Persian.
The inscription is unusual insofar as it is one of very few bilingual tomb in-
scriptions, but also since nothing about the dead person has been revealed.
Neither his or her name, nor the date of his or her death have been mentioned,
at least not in Persian. We can only speculate that the deceased had a local
origin – was this the reason for adding some lines in Śāradā?

Conclusion
In none of the inscriptions studied in this corpus is either text a verbatim trans-
lation of the other. It is therefore more appropriate to speak about two versions
of a text conveying, in most cases, the same message. Even in cases in which the
texts do not vary significantly, each version follows its own stylistic norms. The
majority of the inscriptions can be compared to what Taylor has described for
Aramaic and Greek inscriptions from Palmyra, which he considers to be ‘inde-
pendent products conforming to accepted conventions’.104 These conventions
can be understood from how the details of the dates are indicated; the position
of the dates in the text; the gods who have been invoked; specific phrases that
are included or omitted in the text, like e.g. the basmala or quotations from the
Ḥadīth and Quran; blessings and imprecations; the indication of other or addi-
tional people responsible for adherence to a waqf regulation; etc. Taylor speaks
here of a ‘dynamic equivalence’,105 which means that the text in one language is
not repeated verbatim, but transposes the meaning to its own cultural context.
In the case of the inscriptions from Palmyra, Taylor concludes that the bi-
lingual inscriptions were not translated, but produced by the same bilingual
speaker.106 For the corpus of bilingual inscriptions considered here, such a con-

||
103 Haim 2019, 418–419.
104 Taylor 2002, 321.
105 Taylor 2002, 323.
106 Taylor 2002, 323.
 Bilingual Inscriptions from India | 405

clusion cannot generally be drawn, since it is much too heterogenous. At least

for the Sanskrit-Persian bilingual inscriptions, it is highly improbable that they
were composed by the same bilingual speaker, since we know from studies on
translations from Sanskrit into Persian that they were produced by a team of
Sanskrit and Persian experts, and that there was often even an intermediate
language involved in the process of translation.107 I also doubt that there were
speakers of both Arabic and Sanskrit. The production of the inscriptions certain-
ly followed various models in the different geographical and historical contexts,
but one might imagine that an order to produce an inscription was issued to one
expert in each language, with clear instructions about the content, upon which
the expert was left to formulate an appropriate text fitting the conventions of the
language and the culture associated with it. Although it is possible to categorize
the inscriptions according to the scheme provided by Sohoni, this gives only a
rough orientation, and I doubt that this scheme is very helpful in understanding
the complexity of goals, interests, needs and conventions that influenced the
production of these inscriptions. This can be achieved only by further examin-
ing the historical and social circumstances of their creation in detailed studies
for the different dynasties, epochs and regions. Joined editions of the different
versions of the texts, together with high-quality images in a database like Sid-
dham, would also be helpful.

Acknowledgements

I sincerely thank Giovanni Ciotti, Victor D’Avella, Kristen de Joseph, Jost Gip-
pert, Szilvia Jáka-Sövegjártó, Eva Wilden and Lingli Li for helping with the
transliteration of Sanskrit and Telugu inscriptions as well as their translation,
for improving the English and for all critical remarks and help.

References
Adams, James Noel, Mark Janse and Simon Swain (eds) (2002), Bilingualism in Ancient Society:
Language Contact and the Written Text, Oxford: Oxford University Press.
Ahmad, Khwaja Muhammad (1925–1926), ‘An Inscription of Mahmūd Shāh I of Gujarat found at
Dohad’, Epigraphia Indica: Arabic & Persian Supplement, 1925–1926: 20–21.
Ahmad, Qeyamuddin (1973), Corpus of Arabic & Persian inscriptions of Bihar (A.H. 640-1200),
Patna: K.P.Jayaswal Research Institute.

||
107 Truschke 2016, 104–105; Orthmann and Speziale 2020.
406 | Eva Orthmann

Bendrey, Vasudeo Sitaram (1944), A study of Muslim inscriptions: with special reference to the
inscriptions published in the Epigraphia Indo-Moslemica, 1907–1938, together with sum-
maries of inscriptions chronologically arranged, Bombay: Karnatak Pub. House.
Blake, Stephen (2013), Time in Early Modern Islam: Calendar, ceremony, and chronology in the
Safavid, Mughal, and Ottoman empires, Cambridge: Cambridge University Press.
Brandt, Carmen (2014), ‘The identity politics of language and script in South Asia’, Depart, 17:
24–31.
Bukhari, Y. (1953–1954), ‘Some inscriptions from Kalpi and Jatara’, Epigraphia Indica: Arabic &
Persian Supplement, 1953–1954: 34–41.
Bukhari, Y. (1955–1956), ‘Inscriptions from the Archaeological Museum, Amber’, Epigraphia
Indica: Arabic & Persian Supplement, 1955–1956: 57–60.
Bukhari, Y. (1959–1960), ‘Inscriptions from the Archaeological Museum, Red Fort, Delhi’,
Epigraphia Indica: Arabic & Persian Supplement, 1959–1960: 1–22.
Bunčić, Daniel (2016), ‘A heuristic model for typology’, in Daniel Bunčić, Sandra L. Lippert and
Achim Rabus (eds), Biscriptality: A sociolinguistic typology, Heidelberg: Universitätsver-
lag Winter, 51–71.
Chaghtai, M. Abdulla (1949–1950), ‘Some inscriptions from Jodhpur State, Rajputana’,
Epigraphia Indica: Arabic & Persian Supplement, 1949–1950: 18–53.
Chattopadhyaya, Brajadulal (1998), Representing the Other? Sanskrit Sources and the Muslims
(Eighth to Fourteenth Century), New Delhi: Manohar.
Datta, B. and C. L. Suri (1962), ‘Sanskrit portion of the Bidar Inscription in the Hyderabad Mu-
seum’, Epigraphia Indica: Arabic & Persian Supplement, 1962: 81–84.
Desai, Ziauddin Abdul Hayy (1953–1954a), ‘Inscriptions of the Sultans of Gujarat from Saurash-
tra’, Epigraphia Indica: Arabic & Persian Supplement, 1953–1954: 49–77.
Desai, Ziauddin Abdul Hayy (1953–1954b), ‘A Persian inscription of Adil Shahi period from
Siruguppa’, Epigraphia Indica: Arabic & Persian Supplement, 1953–1954: 41–44.
Desai, Ziauddin Abdul Hayy (1955–1956a), ‘Inscriptions from the Museum of Antiquities, Juna-
garh’, Epigraphia Indica: Arabic & Persian Supplement, 1955–1956: 89–102.
Desai, Ziauddin Abdul Hayy (1955–1956b), ‘Two Inscriptions from the Government Museum,
Madras’, Epigraphia Indica: Arabic & Persian Supplement, 1955–1956: 103–106.
Desai, Ziauddin Abdul Hayy (1957–1958), ‘Inscriptions from the Prince of Wales Museum,
Bombay’, Epigraphia Indica: Arabic & Persian Supplement, 1957–1958: 1–28.
Desai, Ziauddin Abdul Hayy (1959–1960), ‘Inscriptions from the State Museum, Hyderabad’,
Epigraphia Indica: Arabic & Persian Supplement, 1959–1960: 27–37.
Desai, Ziauddin Abdul Hayy (1961), ‘Arabic Inscriptions of the Rajput period from Gujarat’,
Epigraphia Indica: Arabic & Persian Supplement, 1961: 1–24.
Desai, Ziauddin Abdul Hayy (1962), ‘Khaljī and Tughluq inscriptions from Gujarat’, Epigraphia
Indica: Arabic & Persian Supplement, 1962: 1–40.
Desai, Ziauddin Abdul Hayy (1963), ‘Inscriptions of the Gujarat Sultans’, Epigraphia Indica:
Arabic & Persian Supplement, 1963: 5–50.
Desai, Ziauddin Abdul Hayy (1970), ‘Some Mughal inscriptions from Gujarat’, Epigraphia Indi-
ca: Arabic & Persian Supplement, 1973: 63–92.
Desai, Ziauddin Abdul Hayy (1975a), ‘Persian-Sanskrit inscription of Karṇa Deva Vāghelā of
Gujarat’, Epigraphia Indica: Arabic & Persian Supplement, 1975: 13–20.
Desai, Ziauddin Abdul Hayy (1975b), ‘Dohad Inscription of Maḥmūd Begḍā’, Epigraphia Indica:
Arabic & Persian Supplement, 1975: 26–30.
 Bilingual Inscriptions from India | 407

Desai, Ziauddin Abdul Hayy (2011), ‘Inscriptions of Shah Jahan from Rajasthan’, Epigraphia
Indica: Arabic & Persian Supplement, 2011: 1–61.
Diskalkar, Dattatraya Balkrishna (1939–1940), ‘Inscriptions of Kathiawad’, New Indian Anti-
quary, 2: 591–606.
Gyani, R. G. (1944), ‘Karakhaḍi stone inscription of the reign of Sulṭān Muḥammad b. Tughluq
dated 740 (A.D. 1340)’, in Ghulam Yazdani and R. G. Gyani (eds), Muslim Inscriptions from
the Baroda State, Baroda: Baroda State Press, 1–2 and pl. 1.
Haig, Thomas Wolseley (1907–1908), ‘Some inscriptions in Berar’, Epigraphia Indo-Moslemica,
1907–1908: 10–20.
Haim, Ofir (2019), ‘Acknowledgement Deeds (iqrārs) in Early New Persian from the Area of
Bāmiyān (395–430 AH / 1005–1039 CE)’, Journal of the Royal Asiatic Society, 3rd series,
39/3: 415–446.
Harmatta, János (1966), ‘New evidences for the history of early medieval Northwestern India’,
Acta Antiqua, 14: 423–462.
Humbach, Helmut (1966), with contributions by Adolf Grohmann, Baktrische Sprachdenkmäler,
vol. 1, Wiesbaden: Harrassowitz.
Hussain, S. S. (1973), ‘Two New Inscriptions of the Mughal Period from Ranthambor in Raja-
sthan’, Epigraphia Indica: Arabic & Persian Supplement, 1973: 44–51.
Hultzsch, Eugen (1882), ‘A Grant of Arjunadèva of Gujarat, dated 1264 A.D.’, The Indian Anti-
quary, 11: 241–245.
Indraji, Pandit Bhagavánlál (1896), with the collaboration of A. M. T. Jackson, ‘Early History of
Gujarát (B.C. 319–A.D. 1304)’, Gazetteer of the Bombay Presidency, 1/1: 1–206.
Ishaque, M. (1955–1956), ‘Inscriptions from the Sardar Museum, Jodhpur’, Epigraphia Indica:
Arabic & Persian Supplement, 1955–1956: 63–66.
Joshi, R. M. (1959–1960), ‘Sanskrit version of the bilingual inscription from the State Museum,
Hyderabad’, Epigraphia Indica: Arabic & Persian Supplement, 1959–1960, 38–40.
Kadiri, A. A. (1962), ‘Bahmani inscriptions from Raichur district’, Epigraphia Indica: Arabic &
Persian Supplement, 1962: 53–66.
Kadiri, A. A. (1964), ‘Inscriptions of the Bahmanīs of Deccan’, Epigraphia Indica: Arabic &
Persian Supplement, 1964: 21–44.
Kadiri, A. A. (1965), ‘Three Adil Shahi inscriptions from Goa’, Epigraphia Indica: Arabic & Per-
sian Supplement, 1965: 39–48.
Kadiri, A. A. (1967), ‘Nizam Shahi Inscriptions from Galna’, Epigraphia Indica: Arabic & Persian
Supplement, 1967: 43–60.
Kadiri, A. A. (1970), ‘Some mote direction stones of Niẕām Shāhī dynasty’, Epigraphia Indica:
Arabic & Persian Supplement, 1970: 45–62.
Kariminia, Morteza (1396 AH / 2018 CE), ‘Kuhantarīn maktūb-i tārīkhdār-i farsī: dastnawishtī
farsī az Aḥmad b. Abi ’l-Qāsim Khayqānī dar pāyān-i Qurʾānī az qarn-i siwwum’, Āyina-yi
mīrāth, 61: 9–26.
Khan, M. F. (1964), ‘Two Persian Inscriptions of Jahāngīr from Madhya Pradesh’, Epigraphia
Indica: Arabic & Persian Supplement, 1964: 79–82.
Khan, M. F. (1968), ‘Three Grants of the time of Aurangzeb from Kota district’, Epigraphia Indi-
ca: Arabic & Persian Supplement, 1968: 69–78.
Ḳuraishi, Muhammad Hamid (1925–1926a), ‘A Kufic Sarada Inscription from the Peshawar
Museum’, Epigraphia Indica: Arabic & Persian Supplement, 1925–1926: 27–29.
408 | Eva Orthmann

Ḳuraishi, Muhammad Hamid (1925–1926b), ‘Some Persian, Arabic and Sanskrit Inscriptions
from Asīrgarh in Nimār District, Central Provinces’, Epigraphia Indica: Arabic & Persian
Supplement, 1925–1926: 1–7.
Lal, Hira (1908), ‘Burhanpur Sanskrit Inscription of Adil Shah’, Epigraphia Indica, 9: 306–310.
Leiwo, Martti (2002), ‘From Contact to Mixture: Bilingual Inscriptions from Italy’, in Adams,
Janse and Swain 2002, 168–194.
Martin, Marie H. (1971), ‘The Shuhūr San: date equivalencies, origins and special problems’,
Epigraphia Indica: Arabic & Persian Supplement, 1971: 81–106.
Nazim, M. (1933–1934), ‘Inscriptions from the Bombay Presidency’, Epigraphia Indica: Arabic
& Persian Supplement, 1933–1934: 1–61.
Nazim, M. (1936), Bijapur Inscriptions (Memoirs of the Archaeological Survey of India, 49),
Delhi: Manager of Publications.
Orthmann, Eva and Fabrizio Speziale (2020), ‘Tarjuma-yi Sālōtar’, in Fabrizio Speziale and Carl
W. Ernst (eds), Perso-Indica: An Analytical Survey of Persian Works on Indian Learned Tra-
dition, <http://www.perso-indica.net/work/tarjuma-yi_salotar> (accessed on 4 Septem-
ber 2023).
Pickthall, Marmaduke (1930), The Meaning of the Glorious Koran: an Explanatory Translation,
London: Knopf.
Prakash, Satya (1968), ‘A Bilingual Inscription from Jaipur District’, Epigraphia Indica: Arabic &
Persian Supplement, 1968: 67–68.
Prasad, Pushpa (1990), Sanskrit Inscriptions of Delhi Sultanate 1191–1526, Delhi: Oxford Uni-
versity Press.
Quraishi, Muhammad Hamid (1923–1924), ‘Inscriptions of Sher Shāh and Islām Shāh’,
Epigraphia Indica: Arabic & Persian Supplement, 1923–1924: 26–30.
Rahim, S. A. (1961), ‘Inscriptions of the Fārūqī kings from Burhanpur’, Epigraphia Indica: Arabic
& Persian Supplement, 1961: 49–58.
Rahman, Abdur (1998), ‘The Zalamkot bilingual inscription’, East and West, 48/3–4: 469–473.
Saksena, Ram Singh (1937–1938), ‘Muslim inscriptions from Bhonrasa, Gwalior State’,
Epigraphia Indica: Arabic & Persian Supplement, 1937–1938: 22–34.
Samadi, S. R. (1955–1956), ‘Inscriptions from the Archaeological Museum, Bijapur’, Epigraphia
Indica: Arabic & Persian Supplement, 1955–1956: 71–86.
Sankalla, H. D. (1937–1938), ‘Dohad Stone inscription of Mahamuda (Begarha): V.S. 1545, Saka
1410’, Epigraphia Indica, 24: 212–225.
Shaikh, C. H. (1939–1940), ‘Some unpublished inscriptions from Ahmadnagar’, Epigraphia
Indica: Arabic & Persian Supplement, 1939–1940: 30–32.
Sheikh, Samira (2014), ‘Languages of Public Piety’, in Francesca Orsini and Samira Sheikh
(eds), After Timur Left: Culture and Circulation in Fifteenth-Century North India, Oxford: Ox-
ford University Press, 186–209.
Siddiqui, Mazher Alam (1995), ‘A bilingual Ādil Shāhī inscription from Sholapur, Maharashtra’,
Journal of the Epigraphical Society of India, 21: 98–99.
Sircar, Dinesh Chandra (1961–1962), ‘Veraval inscription of Chaulukya-Vaghela Arjuna, 1264 A.D.’,
Epigraphia Indica, 34: 141–150.
Sohoni, Pushkar (2022), ‘Cultural Translation and Linguistic Equivalence: Persian-Marathis
Bilingual Inscriptions’, Journal of Persianate Studies, 15/1: 1–15.
Taylor, David G. K. (2002), ‘Bilingualism and Diglossia in Late Antique Syria and Mesopotamia’,
in Adams, Janse and Swain 2002, 298–331.
 Bilingual Inscriptions from India | 409

Thapar, Romila (2008), Somanatha: The many voices of a history, New Delhi: Penguin Books
[1st edn: 2004].
Truschke, Audrey (2016), Culture of Encounters: Sanskrit at the Mughal Court, New York: Co-
lumbia University Press.
Tintu, K. J. (2019), ‘The Syrian Christian copper plate of Tarisāpaḷḷy, and the Jewish and Muslim
merchants of early Malabar’, Proceedings of the Indian History Congress, 80: 184–191.
Verma, B. D. (1955–1956), ‘Inscriptions from the Central Museum, Nagpur’, Epigraphia Indica:
Arabic & Persian Supplement, 1955–1956: 109–118.
Virji, Krishnakumari J. (1952), Ancient History of Saurashtra (Being a Study of the Maitrakas of
Valabhī), Bombay: Konkan Institute of Arts and Sciences.
Watson, John W. (1879), ‘Notes on the seacoast of Saurâshṭra, with a few remarks on the extent
of the Chuḍâsamâ rule’, Indian Antiquary, 8: 181–187.
Yazdani, Ghulam (1913–1914), ‘Inscriptions in Golconda Fort’, Epigraphia Indica: Arabic &
Persian Supplement, 1913–1914: 47–59.
Yazdani, Ghulam (1915–1916), ‘Inscriptions in the Tomb of Bābā Arjun Shāh, Petlād (Baroda
State)’, Epigraphia Indica: Arabic & Persian Supplement, 1915–1916: 15–18.
Yazdani, Ghulam (1921–1922), ‘Inscriptions from the Bīḍ (Bhīr) district’, Epigraphia Indica:
Arabic & Persian Supplement, 1921–1922: 13–30.
Yazdani, Ghulam (1925–1926), ‘Inscription of Ibrāhīm Quṭb Shāh from the Pangal Tank,
Nalgonda District’, Epigraphia Indica: Arabic & Persian Supplement, 1925–1926: 23–24.
Yazdani, Ghulam (1930), ‘Annual Report of the Archaeological Department, Hydearbad State,
for 1337 F. (1927–28 A.C.)’, Annual Report of the Archeological Department of His Exalted
Highness the Nizam’s Dominions, 1930: 1–16.
Yazdani, Ghulam (1935–1936a), ‘The Bilingual Inscription of Quṭbu’d-Dīn Khaljī from the Rasul
Khanji Museum, Junagarh’, Epigraphia Indica: Arabic & Persian Supplement, 1935–1936:
48–49.
Yazdani, Ghulam (1935–1936b), ‘Inscriptions from Mudgal’, Epigraphia Indica: Arabic & Per-
sian Supplement, 1935–1936: 14–19.
Yazdani, Ghulam (1937–1938a), ‘Inscription of Sulṭān Balban from Bayana, Bharatpur State’,
Epigraphia Indica: Arabic & Persian Supplement, 1937–1938: 5–6.
Yazdani, Ghulam (1937–1938b), ‘Five new inscriptions from the Bidar district’, Epigraphia
Indica: Arabic & Persian Supplement, 1937–1938: 1–4.
Yazdani, Ghulam (1937–1938c), ‘Some Muslim inscriptions from the Madras Presidency and
Orissa’, Epigraphia Indica: Arabic & Persian Supplement, 1937–1938: 52–59.
Yazdani, Ghulam (1939–1940), ‘The Bililngual Inscription of Quṭbu’d-Dīn Khaljī from the Rasul
Khanji Museum, Junagarh’, Epigraphia Indica: Arabic & Persian Supplement, 1939–1940: 47.
410 | Eva Orthmann

Appendix: List of inscriptions

No. 1
Year: 5th regnal year of King Sthanu Ravi [849–850 CE]
Names: Ayyanadikal; Tiruvadikal; Sthanu Ravi; Mar Sapir Iso
Location: Kollaṁ, Kerala
Languages: Tamil, with Hebrew, Arabic and Middle Persian
Material aspects: Vaṭṭeḻuttu, Pahlavi, Kufic and Hebrew scripts
Context: Copper plate
Content: Conferring of rights and privileges to Christian merchants
Salient features: Main text in Tamil, other languages and alphabets used for
signatures
References: Tintu 2019

No. 2
Year: 13 Jumādā I 243 AH; (39)32 Śāstra or Laukika era [857 CE]
Location: Tochi valley, North Waziristan
Languages: Arabic and Sanskrit
Material aspects: Kufic, Śāradā and Nāgarī; all engraved; Arabic on top, same size
Context: Out of context, Peshawar Museum
Content: Building inscription
Salient features: Sanskrit text with dating only, incomplete
References: Ḳuraishi 1925–1926a, 27–28 and pl. XI b; Harmatta 1966, 427–448;
Humbach 1966, 11–18

No. 3
Year: Dhū l-Qaʿda 401 AH; Saṃvat 189; 1 Āśaḍha-vadi [1011 CE]
Names: Arsalān al-Jādhib (governor of Ṭūs); Khalīlbak; Ghaznavid
Location: Zalamkot, lower Swat valley
Languages: Persian and Sanskrit
Material aspects: Kufic and Śāradā, both engraved
Context: Tomb; currently in Peshawar Museum
Content: Construction of a tomb by two governmental functionaries
Salient features: Oldest Persian inscription in Arabic script; Sanskrit text not
edited and translated; name of deceased not mentioned
References: Rahman 1998
 Bilingual Inscriptions from India | 411

No. 4
Year: 27 Ramaḍān 662 AH; Saṃvat 1320; 13 Āśaḍha-vadi; 945 Valabhī; 151 Siṁha
[1264 CE]
Names: Nākhudā Fīrūz; Arjunadeva, king of Gujarat; Maḥmūd b. Aḥmad; Amīr
Rukn ud-Dīn, ruler of Hormuz
Location: Prabhaṣ Patan/Somnath, Gujarat
Languages: Arabic and Sanskrit
Material aspects: Engraved on two separate black granite stones
Context: Mosque; texts since relocated to the wall of another mosque and a temple
Content: Building of a mosque; waqf
Salient features: Details of waqf in Sanskrit; translation of Allah into Hindu
concepts
References: Hultzsch 1882; Desai 1961, 10–15 and pl. II b; Sircar 1961–1962, 141–150;
Chattopadhyaya 1998, 70–78; Thapar 2008, 88–99; Sheikh 2014, 190–191

No. 5
Year: 1 Ramaḍān 669 AH [1271 CE]
Names: Nuṣrat Khān; Ghiyāth ud-Dīn Balban
Location: Bayana, Bharatpur, Rajasthan
Languages: Persian and Sanskrit
Material aspects: Naskh Sanskrit inscription on the back of the tablet
Context: Well belonging to Goculchandramājī temple in Kāman
Content: Clearance and renewed digging of a well that had been filled with stones
Salient features: Sanskrit not edited or translated
References: Yazdani 1937–1938a, 5–6 and pl. III a

No. 6
Year: 1 Muḥarram 704 AH; Saṃvat 1360; 2 Bhādrapada-vadi [1304 CE]
Names: Rāi Karna Deva Vāghelā; Balchaq; Shādī; Tāj ud-Dīn Ḥasan (Khaljī period)
Location: Sāmpā, 45 km from Vaḍodara, Gujarat
Languages: Persian and Sanskrit
Material aspects: Naskh, Nāgarī; Persian on top; both incomplete
Context: Close to a pond; no edifice left
Content: Endowment for the village of Sāmpā for the mosque of Kambāyat; in-
vocation to abide to the terms
412 | Eva Orthmann

Salient features: Sanskrit similar to Persian; Sanskrit version invokes Rānās to

honour and execute the gift
References: Desai 1975a, 13–20 and pl. II a; Sheikh 2014, 191–192

No. 7
Year: (720 AH); Śaka 1242 [1320 CE]
Names: Revaiya/Rībya; Karīm ud-Dīn; Quṭb ud-Dīn Khaljī
Location: Bijapur, Karnataka
Languages: Persian and Marathi
Material aspects: Nāgarī, Persian; Persian in hesitant script
Context: Malik Karīm ud-Dīn mosque; Marathi on a pillar in the north wall; Per-
sian on the second pillar in second row from the north
Content: Building of the mosque; mentions the carpenter who built it and his
remuneration
Salient features: Marathi version much more detailed than Persian version; date
indicated only in Marathi
References: Nazim 1936, 25; Sohoni 2022, 8–9

No. 8
Year: 20 Dhū l-Ḥijja (723 AH); Saṃvat 1380; 7 Pausa-vadi [1323 CE]
Names: Bābā Arjun Shāh; Hājī Ismāʿīl; ʿUthmān of Shiraz; Ghiyāth ud-Dīn Tughluq
Location: Petlad, Anand district, Gujarat
Languages: Persian and Sanskrit
Material aspects: Naskh, Nāgarī
Context: In the tomb of Bābā Arjun; originally in the vicinity of the tomb
Content: Dedication of a well; grant
Salient features: In the Sanskrit version, the thakkuras are asked to protect the gifts
References: Yazdani 1915–1916, 16–18 and pl. XIV b; Sheikh 2014, 201–202

No. 9
Year: 25 Dhū l-Qaʿda 740 AH [1340 CE]
Names: Mokha Mehta; Malik Muẓaffar Sulṭān; Muḥammad b. Tughluq
Location: Karkhaḍi, Vaḍodara district, Gujarat
Languages: Persian and Sanskrit
 Bilingual Inscriptions from India | 413

Material aspects: Naskh, Nāgarī; Sanskrit in more lines

Content: Construction of a mosque and a well
Salient features: Sanskrit has more details
References: Gyani 1944, 1–2 and pl. 1; Sheikh 2014, 193–194

No. 10
Year: 1 Rabīʿ I 765 AH; Saṃvat 1420; 1286 Śaka; 14 Mārgaśīrṣa-vadi [1363 CE]
Names: Bāmdev, son of Nathū; Fīrūz Shāh Tughluq
Location: Originally from Sambhar, Jaipur district, Rajasthan; today in Archaeo-
logical Museum, Amber, Jaipur, Rajasthan
Languages: Persian and Sanskrit
Material aspects: Naskh, Nāgarī; Persian in relief, Nāgarī incised and in very
poor condition; Sanskrit much smaller and below the Persian
Context: Fixed on a well outside the town of Sambhar
Content: Construction of a step well
Salient features: In Sanskrit; invocation of Varuna at the beginning
References: Bukhari 1955–1956, 57–59 and pl. XV a

No. 11
Year: 15 Safar 779 AH [1377 CE]
Names: Nākhudā Ghafūr, son of Nākhudā Aḥmad
Location: Originally probably from the west coast, somewhere to the north of
Mumbai; today in Chhatrapati Shivaji Maharaj Vastu Sangrahalaya (CSMVS)
museum, Mumbai
Languages: Arabic; mixture of Marathi and Gujarati; Portuguese insignia on the
back
Material aspects: Arabic in verses in Naskh script, carved in relief, illegible;
Nāgarī throughout the lower fourth, incised, text width broader than for the
Arabic part
Context: Removed from original position by the Portuguese
Content: Epitaph, indicating the name of the deceased
Salient features: Only Sanskrit version read and translated; Perso-Arabic portion
badly damaged
References: Desai 1957–1958, 12–13 and pl. IV a; Sohoni 2022, 9
414 | Eva Orthmann

No. 12
Year: (784 AH) [1382–1383 CE]
Names: Dungar; Tikam; Natha; Punji, the mother of Dungar
Location: Sathod, Vaḍodara district, Gujarat
Languages: Persian, Old Gujarati, Sanskrit
Material aspects: Naskh, Nāgarī; versified Sanskrit
Content: Repair of a step well; grant of land for the maintenance of the well
Salient features: Genealogy of brothers in Sanskrit; different curses in the three
languages
References: Sheikh 2014, 202–204

No. 13
Year: 797 AH; Saṃvat 1452 (15 Vaiśākha-vadi) [1394–1395 CE]
Names: Malik Yaʿqūb at-Tamīmī; Nuṣrat Shāh; Qāḍī Badr Rāṇig, son of Virdha-
vala; Tughluq period
Location: Mangrol, Junagadh district, Gujarat
Languages: Persian in verse; Sanskrit in prose
Material aspects: Persian arranged in four columns; Sanskrit below; scripts in
about the same size; Persian in relief; Sanskrit incised
Context: Small room at the Gādī gate, today at the record office of Mangrol
Content: Construction of two iron doors for the two gates
Salient features: No translation of the Sanskrit text; Persian begins with the
creation of earth; Sanskrit text indicates who wrote the Torki (= Persian) text
and who engraved the Sanskrit text
References: Desai 1962, 34–37 and pl. XI a; Diskalkar 1939–1940, 592–593;
Sheikh 2014, 192–193

No. 14
Year: (807 AH) [1404–1405 CE]
Names: Amir Nathu, son of Takhir; Ẓafar Khān, son of Wajīh ul-Mulk; Malik
Adam b. Sulaimān
Location: Vaḍodara, Gujarat
Languages: Persian and Sanskrit
Material aspects: Not indicated
Context: At the entrance of a well; today at the entrance of the Jamiʿ mosque
 Bilingual Inscriptions from India | 415

Content: Construction of a step well

Salient features: Not indicated
References: Sheikh 2014, 204

No. 15
Year: 14 Shawwāl 810 AH; 27 Rabīʿ I 811 AH; Saṃvat 1464; 2 Caitra-vadi; 13 Śrāvana-
vadi [1408 CE]
Names: Faḍlallāh Aḥmad; Abū Rajā; Muẓaffar Shāh, sultan of Gujarat
Location: Veraval, Gir Somnath district, Gujarat
Languages: Persian (prose) and Sanskrit
Material aspects: Sanskrit below Persian, much smaller
Context: Inner face of the west wall of the mortuary of a small tomb of Maghrībī
Shāh; not in situ
Content: Construction of a city wall and guard-room
Salient features: Sanskrit contains mostly dates and names
References: Desai 1953–1954a, 50–52 and pl. XV a; Diskalkar 1939–1940, 598–599;
Sheikh 2014, 194

No. 16
Year: (811–812 AH) [1409 CE]
Names: Zheng He
Location: Galle, Sri Lanka; today in Colombo National Museum
Languages: Tamil, Persian and Chinese
Material aspects: Tamil at the top left; Persian at the bottom left; Chinese in the
right portion of the slab; two dragons facing each other carved at the top
Context: Made in Nanjing and carried to Sri Lanka; placed in a temple
Content: Offerings of the Chinese emperor to a god/saint in Sri Lanka; list of
offerings
Salient features: Beneficiary of the offerings is different in each language
References: Siddham database, https://siddham.network/object/ob03125/ (ac-
cessed on 4 September 2023)
416 | Eva Orthmann

No. 17
Year: 15 Muḥarram 847 AH [1443 CE]
Names: Muḥammad Shāh
Location: Vaḍodara district, Gujarat
Languages: Persian and Sanskrit
Material aspects: With figures of animals engraved at the top; Persian (in Naskh)
much longer
Context: Pillar, originally from Sathod
Content: Order of remission of certain taxes
Salient features: Sanskrit not edited
References: Desai 1963, 22–23 and pl. VII a

No. 18
Year: 14 Jumādā II 848 AH; Śālivāhan Śaka 1366, 10 Aśvin-śudi [1444 CE]
Names: Makhdūma-yi Jahān Bībī Shahnāz; Maḥmūd Khān; ʿAlā ud-Dīn Aḥmad
Bahmānī
Location: Bidar, Karnataka; today in the Telangana State Archaeology Museum,
Hyderabad
Languages: Persian, Arabic and Sanskrit
Material aspects: Naskh in relief; Nāgarī engraved; prose and verse; Sanskrit
much longer
Context: Step well
Content: Construction of a step well for public use
Salient features: Sanskrit version has much more information, inter alia about a
conflict within the royal family
References: Desai 1959–1960, 33–34 and pl. VIII a; Joshi 1959–1960, 38–40;
Datta and Suri 1962, 81–84 and pl. XXVI

No. 19
Year: 11 Dhū l-Qaʿda 859 AH [1067 CE]
Names: Quṭb ud-Dīn Aḥmad II, sultan of Gujarat
Location: Prabhaṣ Patan, Gir Somnath district, Gujarat
Languages: Persian and Sanskrit
Material aspects: Naskh, Nāgarī; Naskh much more prominent; Sanskrit small
and at the bottom; Naskh in relief; Sanskrit incised
 Bilingual Inscriptions from India | 417

Context: Inner face of the city wall, to the left of the Moṭā Darwāza
Content: Order by the sultan to put a stop to the oppression of the public by
some officials
Salient features: Sanskrit considered illegible; in Persian, admonition of Mus-
lims and non-Muslims not to commit illegal acts
References: Desai 1955–1956a, 94–96 and pl. XXVI b; Sheikh 2014, 196–197

No. 20
Year: 15 Rabīʿ I 862 AH; Saṃvat 1514; 2 Śrāvaṇa-vadi [1458 CE]
Names: Malik Asad; Quṭb ud-Dīn Aḥmad II, sultan of Gujarat
Location: Museum of Pasnavada, close to Prabhaṣ Patan; Gir Somnath district,
Gujarat
Languages: Persian and Sanskrit
Material aspects: Nāgarī
Content: Construction of a fort
Salient features: Sanskrit version more informative than Persian; indicates names
of artisans
References: Desai 1955–1956a, 96–99; Sanskrit version in Watson 1879, 183;
Sheikh 2014, 194–195 (no plate)

No. 21
Year: 862 AH [1457–1458 CE]
Names: Humāyūn Shāh Bahmānī
Location: Gulbarga, Gulbarga district, Karnataka
Languages: Persian and Sanskrit
Material aspects: Naskh and Nāgarī; Persian in relief, Sanskrit incised; Sanskrit
much smaller than Persian
Context: Eastern wall of the Bāolī Qalandar Shāh, Gulbarga
Content: Construction of a step well
Salient features: Sanskrit version mentions Ijalidevi as builder; Sanskrit not
edited or translated
References: Kadiri 1964, 40–41 and pl. XIII b
418 | Eva Orthmann

No. 22
Year: 862 AH; Saṃvat 1514; 2 Śrāvaṇa-vadi [1458 CE]
Names: Quṭb ud-Dīn Aḥmad
Location: Darbar Hall Museum (?), earlier Rasul Khanji Museum, Junagadh
Languages: Persian and incorrect Sanskrit
Material aspects: Thulth in relief; Nāgarī; Nāgarī in much smaller characters
Content: Construction of a mīnār in Sulṭānpūr
Salient features: Sanskrit inscription mentions the earlier name Deopattan
References: Yazdani 1935–1936a, 48–49 and pl. XXXV b; Yazdani 1939–1940, 47

No. 23
Year: (887 AH); Saṃvat 1540; 5 Phālguna-vadi [1482 CE]
Names: Mahmūd Shāh, king of Malwa
Location: Bhonrasa, Bhopal district, Madhya Pradesh
Languages: Persian and Hindi with a few words in defective Sanskrit
Material aspects: Naskh; Nāgarī of a late period; writing style crude in both
languages; both texts of equal size; Persian at the top
Context: Stone post near the main gate
Content: Text only partially preserved; remission of taxes
Salient features: In Hindi version, revival of worship and sin of killing cows is
mentioned
References: Saksena 1937–1938, 23–26 and pl. VII a

No. 24
Year: 893 AH; Saṃvat 1545; Śaka 1410; 13 Vaiśākha-śudi [1488 CE]
Names: Maḥmūd Begarhā Āṣaf, Malik ush-Sharq Maḥmūd Shāh b. Shāh b.
Maḥmūd Shāh
Location: Mahamuda, Dohad district, Gujarat
Languages: Arabic (with some Persian phrases) and Sanskrit
Material aspects: Naskh, in relief; date given in figure; Nāgarī, in verse, en-
graved; both versions on two different slabs
Context: Basket Bond Dohad
Content: Genealogy of Gujarat sultans; Maḥmūd’s deeds and conquests; com-
pletion of a city
 Bilingual Inscriptions from India | 419

Salient features: In Sanskrit invocation of a goddess residing in Kashmir; two

versions of the same record
References: Khwaja Muhammad Ahmad 1925–1926, 20–21 and pl. IX b; Desai 1975b,
26–30 and pl. III a; Sankalla 1937–1938, 212–225

No. 25
Year: 12 Muḥarram 901 Shuhūr-San; 1422 Śaka; 3 Bhādrapada-vadi [1500 CE]
Names: Khān Ghalib Khān
Location: Bijapur, Archaeological Museum (Gol Gumbaz Museum)
Languages: Persian and Kanarese
Material aspects: Naskh, incised; two scripts separated by different figures (sun
and moon, scales etc.); Kanarese part much longer
Context: No location indicated
Content: Qaulnāma (agreement, promise) issued by Khān Ghalib Khān
Salient features: The two inscriptions complement each other: the Persian part
asks people to obey the content of the Qaul-i Hindawī, regulations for Muslims
and Hindus in case they do not adhere to it; Kanarese version (not edited or
translated) refers to cultivation of fallow lanReferences: Samadi 1955–1956, 77–78
and pl. XIX c

No. 26
Year: (904 AH) [1499 CE]
Names: Bai Harir Sulṭānī; Maḥmūd Shāh
Location: Asarwa, Ahmadabad, Gujarat
Languages: Arabic and Sanskrit
Material aspects: Arabic briefer than Sanskrit
Context: On opposite walls of the wall, facing each other
Content: Construction of a step well by a slave attendant
Salient features: Sanskrit inscription begins with invocation of Varuṇa, the lord
of water; Sanskrit indicates position of Bai Harir and costs of the well as well as
the names of the builders
References: Sheikh 2014, 204–205
420 | Eva Orthmann

No. 27
Year: Rajab 906 AH; Saṃvat 1557; 3 Phālguna-śudi [1501 CE]
Names: Shyām Kunwar Kalāwantī; Tātār Khān; Lodī dynasty
Location: Daulatpura village, Jatara, Tikamgarh district, Madhya Pradesh
Languages: Persian and Sanskrit
Material aspects: Naskh, upper half, in relief; Nāgarī incised; Persian and Nāgarī
are the same size
Context: Shaikhonwālī Bāolī, embedded in the wall
Content: Construction of a step well
Salient features: Sanskrit text cannot be fully read
References: Bukhari 1953–1954, 40–41 and pl. XII c

No. 28
Year: 14 Dhū l-Qaʿda 918 AH [1513 CE]
Names: Malik ush-Sharq; Malik ʿAmbar; Bahmanid dynasty
Location: Maliabad village Raichur district, Karnataka
Languages: Persian and Kannada
Material aspects: Naskh, Kannada; Naskh in relief, Kannada incised; with ani-
mal drawing
Context: Lying loose in the local mosque
Content: Qaulnāma issued by Malik ʿAmbar, reductions in the levies on the vil-
lage communities and prohibition of forced labour
Salient features: Kannada only contains imprecatory portion
References: Kadiri 1962, 63–65 and pl. XIX

No. 29
Year: 13 Rajab 921 AH; Saṃvat 1572; 1437 Śaka [1515 CE]
Names: Fatḥ Khān; Abū n-Naṣr Muẓaffar Shāh (Gujarat)
Location: Sankheda, Chhota Udaipur district, Gujarat
Languages: Persian and Sanskrit
Material aspects: Persian letters in uneven size; Sanskrit in an even hand; Per-
sian at top
Context: Right bastion of the south gate of the fort
Content: Construction of a well and a mosque
Salient features: Sanskrit is very damaged
References: Desai 1963, 43–44 and pl. XIV a
 Bilingual Inscriptions from India | 421

No. 30
Year: 20 Ṣafar 923 AH; Saṃvat 1573; 7 Caitra-vadi [1517 CE]
Names: Shaikh Sikandar; Bībī ʿĀʾisha; Sikandar Shāh Lodī
Location: Mubarakpur Kotla, Delhi
Languages: Persian and Sanskrit
Material aspects: Naskh in relief, Nāgarī incised; Persian at right, Sanskrit at
left; same size
Content: Construction of a well
Salient features: Contents very similar, but Sanskrit omits reference to the king
as conqueror of infidels, shelter of Islam and warrior on the path of God
References: Bukhari 1959–1960, 8–10 and pl. II b; Prasad 1990, 37–40

No. 31
Year: 1 Rabīʿ I 940 AH [1533 CE]
Names: Khāqān Ulugh Khān; Bahādur Shāh, son of Muẓaffar Shāh
Location: Bharuch (Broach), Bharuch district, Gujarat
Languages: Persian and ?
Material aspects: Persian in relief in the centre; another script on the margins,
incised, perhaps an earlier text that has been overwritten in the central panel
Context: On a stone fixed to the Furza Mosque
Content: Construction of a new ditch for a fortress; names and dates
Salient features: The second script is not mentioned in the edition – it could be a
spolium from an earlier temple or other building
References: Nazim 1933–1934, 30–31 and pl. XVIIb

No. 32
Year: Ṣafar 94(8) AH; Saṃvat 1599; 7 Āsāḍha-śudi [1541–1542 CE]
Names: Yūsuf; Hemārdrasēn; Shēr Shāh
Location: Bhabua, Kaimur district, Bihar; today in the Indian Museum Kolkata
Languages: Persian and Sanskrit
Material aspects: Naskh, Nāgarī; letters raised in both scripts
Content: Construction of a sepulchre, a tank and a garden
Salient features: Sanskrit with invocation to Ganesha; mentions Hemārdrasēn as
somebody who made an endowment; Hijrī date indicated wrongly in Sanskrit
References: Quraishi 1923–1924, 26; Qeyamuddin Ahmad 1973, 136–140 and pl. 27
422 | Eva Orthmann

No. 33
Year: 14 Ramaḍān 958 AH; 1432 Śaka; 15 Māgha-śudi [1551 CE]
Names: Abū l-Muẓaffar Sulṭān Quṭb Shāh; Sayyid Shāh Mīr Iṣfahānī, son of
Sayyid Aḥmad Ṭabāṭabāʾī
Location: Nalgonda, Telangana
Languages: Persian and Telugu
Material aspects: Naskh, Telugu, both incised; sun and moon carved at the top;
Persian on top; Telugu version longer
Context: Fixed on the dyke of the Pangal tank
Content: Rebuilding of the dam and the tank; attribution of benefits from the
irrigated land
Salient features: Telugu close to Persian, with more details; also better pre-
served; Muslims and Brahmans receive benefits from irrigated lands
References: Yazdani 1925–1926, 23–24 and pl. XI a; Yazdani 1930, 4–7

No. 34
Year: 12 Rabīʿ I 959 AH [1552 CE]
Names: Shaikh Sulaimān; Yūsuf Daulat Khān Ḥusain Sūr; Sūr dynasty
Location: Originally from Nagaur, Rajasthan; today in Sardar Government Mu-
seum, Jodhpur
Languages: Persian and Nāgarī
Material aspects: Naskh, Nāgarī; Nāgarī script is indistinct, text is much shorter;
Persian in relief; Nāgarī incised; Persian much more prominent
Context: From a pūsāl at Nagaur
Content: Declaration of redemption and restoration of a pūsāl (school) by the
intervention of a Muslim shaikh
Salient features: Sanskrit text very damaged; purport not clear
References: Ishaque 1955–1956, 63–64 and pl. XVI a

No. 35
Year: (962 AH) [1555 CE]
Names: Nūr Khān; Ibrāhīm I ʿĀdil Shāh
Location: Shirol, Kolhapur district, Maharashtra; today in Irwin Agricultural
Museum, Kolhapur
Languages: Persian and Marathi
 Bilingual Inscriptions from India | 423

Material aspects: Nāgarī, Persian

Content: Grant of a garden near the village of Shirol (Jaisinghpur) to Nūr Khān
by Ibrāhīm I ʿĀdil Shāh
Salient features: Marathi much shorter than Persian; both inscriptions have
donkey curse
References: Sohoni 2022, 9–10

No. 36
Year: (962 AH) [1555 CE]
Names: Nūr Khān; Ibrāhīm I ʿĀdil Shāh
Location: Shirol, Kolhapur district, Maharashtra; today in Irwin Agricultural
Museum, Kolhapur
Languages: Persian, Marathi and Kannada
Material aspects: Persian in Arabic script and Nāgarī
Context: Originally affixed to Nūr Khān’s Gombād in Shirol
Content: Grant of the village of Shirol to Nūr Khān
Salient features: Not indicated
References: Sohoni 2022, 10

No. 37
Year: 977 AH; 1491 Śaka [1569–1570 CE]
Names: Aflāṭūn Khān; Murtaḍā Niẓām Shāh I of Ahmadnagar
Location: Galna fort, Nasik district, Maharashtra
Languages: Persian and Sanskrit
Material aspects: Persian verse; only one line of Nāgarī
Context: Found between two central arches of a bastion situated in the wall of
the fort
Content: Construction of a bastion for the king
Salient features: No image or translation of the Sanskrit part; Sanskrit part par-
tially covered by a wooden frame
References: Desai 1957–1958, 13–14 and pl. IV b
424 | Eva Orthmann

No. 38
Year: (977 AH); 970 Shuhūr-San; 1491 Śaka [1569–1570 CE]
Names: Aflāṭūn Khān; Niẓām Shāhī
Location: Galna fort, Nasik district, Maharashtra
Languages: Arabic, Persian and Nāgarī (no indication of the language)
Material aspects: Naskh, Nāgarī; Nāgarī on the left and lower borders; Persian
prose and verse
Context: Inner side of the fort wall to the right of the iron gate
Content: Date of the construction of a fort
Salient features: No translation of the Nāgarī Arabic year written in Nāgarī letters
References: Kadiri 1967, 46–47 and pl. XI a

No. 39
Year: 978 AH; 1492 Śaka [1570–1571 CE]
Names: Aflāṭūn Khān; Niẓām Shāhī
Location: Galna fort, Nasik district, Maharashtra
Languages: Persian and Nāgarī (no indication of the language)
Material aspects: Naskh, Nāgarī; Nāgarī on the margins; Persian verse
Context: Western wall of the enclosure of a spring called Chashma-yi Khiḍr
Content: Excavation of a spring, associated with the Water of Life
Salient features: Nāgarī version much shorter
References: Kadiri 1967, 47–48 and pl. XII a

No. 40
Year: 978 AH [1570–1571 CE]
Names: Bābājī Ḍābit Khān; ʿAlī ʿĀdil Shāh
Location: Jumma Peth, Solapur, Maharashtra
Languages: Persian and Moṛi
Material aspects: Naskh, Nāgarī
Context: Fixed into a pillar near the steps leading towards the ablution tank in
the Jāmiʿ mosque
Content: Construction of the Friday mosque
Salient features: Moṛi version not edited and described
References: Siddiqui 1995, 98–99 (no plate)
 Bilingual Inscriptions from India | 425

No. 41
Year: 1 Rabīʿ II 985 AH; 1499 Śaka; 10 Jyēṣṭha-śudi [1577 CE]
Names: Ẓahīr Muḥammad Dāʿī; Khān-i Aʿẓam Haibat Khān; Niẓām Shāh period
Location: Galna fort, Nasik district, Maharashtra
Languages: Nāgarī (no indication of the language) and Persian
Material aspects: Persian in Naskh; Nāgarī; both scripts in relief; Nāgarī better
preserved; some lines of Persian lost
Context: A niche between the battlements fronting the north of the bastion
Content: Building inscription for a bastion
Salient features: Sanskrit not edited or translated
References: Desai 1957–1958, 14–15 and pl. V a

No. 42
Year: No Hijra date; 1503 Śaka (1 Vaiśākha-vadi) [1581 CE]
Names: Pūlād Khān Niẓām Shāh
Location: Galna, Nasik district, Maharashtra
Languages: Persian and Marathi
Material aspects: Nastaliq and Nāgarī, both in relief
Context: Fixed on a wall to the right of the iron gate
Content: Some construction
Salient features: Marathi version contains some more details about names and
the date
References: Kadiri 1967, 52–54 and pl. XIII c; Sohoni 2022, 10

No. 43
Year: Rabīʿ I 991 AH (Shuhūr-San); 1505 Śaka; 1 Chaitra-śudi [1583 CE]
Names: Ẓahīr Muḥammad; Pūlād Khān Ghāzī Niẓām Shāh
Location: Galna, Nasik district, Maharashtra
Languages: Persian and Marathi
Material aspects: Naskh and Nāgarī, both in relief; only one line in Marathi
Context: Galna fort
Content: Foundation of a bastion in Galna fort
Salient features: Marathi version contains only the date
References: Kadiri 1967, 55–56 and pl. XIV b; Sohoni 2022, 10
426 | Eva Orthmann

No. 44
Year: 4 Shaʿbān 992 AH; Saṃvat 1641; 1506 Śaka [1584 CE]
Names: ʿĀdil Shāh b. Mubārak Shāh al-Fārūqī
Location: Asīrgarh, Burhanpur district, Madhya Pradesh
Languages: Arabic and Sanskrit
Material aspects: Naskh, Nāgarī; very beautiful; letters raised
Context: Jāmiʿ Masjid, northernmost miḥrāb
Content: Construction of a mosque; dating
Salient features: Translation of central concept of Islam into Sanskrit; rulers
shall endure as long as sun, moon, stars and the Ganges remain
References: Ḳuraishi 1925–26b, 1–2 (no plate)

No. 45
Year: 12 Rajab 996 AH [1588 CE]
Names: ʿAlī ʿĀdil Shāh; Bijapur rulers
Location: Mudgal, Raichur district, Karnataka
Languages: Persian and ?
Material aspects: Ṭughrā style; inscription in other letters below the Persian
text, incised
Context: Bastion
Content: Building of the Fatḥ-i Jang bastion
Salient features: Nothing about another inscription is indicated in the edition
References: Yazdani 1935–1936b, 15 and pl. IX a

No. 46
Year: 997 AH; Saṃvat 1646; 11 Pausha-śudi; 1511 Śaka [1588–1589 CE / 1590 CE]
Names: ʿĀdil Shāh, son of Mubārak Shāh Fārūqī
Location: Burhanpur, Madhya Pradesh
Languages: Arabic and Sanskrit
Material aspects: Naskh, with Ṭughrā flourishes; Nāgarī; very beautiful; letters
raised in both scripts
Context: Jāmiʿ Masjid, southernmost miḥrāb
Content: Construction of a mosque; dating; genealogical information
Salient features: Exact moment of building indicated in Sanskrit; references to
astral bodies; translation of Islamic concepts
References: Rahim 1961, 56–58 and pl. XIX b; Lal 1908, 306–310 (with photo)
 Bilingual Inscriptions from India | 427

No. 47
Year: 1000 AH, but 1000 Shuhūr-San indicated [1591–1592 CE]
Names: Burhān Niẓām Shāh Sulṭān II
Location: Kalamb, Yavatmal district, Maharashtra; today in Central Museum
Nagpur
Languages: Persian and Marathi
Material aspects: Persian at top; very bad condition; Marathi in Nāgarī, but with
some letters in Moṛi style
Context: Stone pillar, which was a direction stone
Content: Indication of directions
Salient features: Beginning like a transliteration of Persian in Nāgarī script;
Marathi version seems to repeat Persian text; directions preserved only in Marathi
References: Verma 1955–1956, 114–115 and pl. XXIX c

No. 48
Year: 1000 AH; but 1000 Shuhūr-San indicated [1591–1592 CE]
Names: Burhān Niẓām Shāh II
Location: Bir, Beed district, Maharashtra
Languages: Persian and Marathi
Material aspects: Naskh; one line in Persian and three lines in Marathi; both
carved in relief
Context: Pillar, locally called rankhamb, fixed in the ground by the side of the road
Content: Indication of direction from Bir to Ahmadnagar
Salient features: Beginning like a transliteration of Persian in Nāgarī script; texts
identical in Persian and Marathi, only one word differs
References: Kadiri 1970, 49–50 and pl. X a

No. 49
Year: 1000 AH [1591–1592 CE]
Names: Burhān Niẓām Shāh II
Location: Nimbait, Nashik district, Maharashtra
Languages: Persian and Marathi
Material aspects: Inscribed on all four sides with bilingual texts
Context: Stone pillar in local graveyard
Content: Indications of direction
428 | Eva Orthmann

Salient features: Marathi version much shorter, without name of ruler and reli-
gious formula
References: Kadiri 1970, 50–53 and pl. XI a, XI b, XII a, XIII b

No. 50
Year: 1000 AH [1591–1592 CE]
Names: Burhān Niẓām Shāh II
Location: Sommaripeṭh, close to Khedla, Betul district, Madhya Pradesh
Languages: Persian and Marathi
Material aspects: Nastaliq and Nāgarī
Context: Somewhere in the village of Sommaripeṭh
Content: Indications of direction
Salient features: Beginning of Marathi like a transliteration of Persian, esp. yek
(one)
References: Kadiri 1970, 53–54 and pl. XII b

No. 51
Year: 1000 AH [1591–1592 CE]
Names: Burhān Niẓām Shāh II
Location: Rawanbari, close to Khedla, Betul district, Madhya Pradesh
Languages: Persian and Marathi
Material aspects: Nastaliq
Context: Lying loose by the side of a cart track
Content: Indications of direction
Salient features: Marathi is transliteration of Persian; no indication of directions
in Marathi
References: Kadiri 1970, 55–56 and pl. XIII a

No. 52
Year: 1000 AH [1591–1592 CE]
Names: Burhān Niẓām Shāh II
Location: Mehkar, Buldana district, Maharashtra
Languages: Marathi (and originally also Persian)
Material aspects: Persian has broken off, only Marathi has been preserved; Nāgarī
 Bilingual Inscriptions from India | 429

Content: Indications of direction

Salient features: Marathi is transliteration of Persian, without directions
References: Kadiri 1970, 57 and pl. XIV b

No. 53
Year: 1000 AH [1591–1592 CE]
Names: Burhān Niẓām Shāh II
Location: Pavnar, Wardha district, Maharashtra
Languages: Persian written in Nāgarī
Material aspects: Persian section is lost, but Nāgarī part also contains Persian
Context: Fixed in a gateway
Content: Indication of direction
Salient features: No Marathi version, but Persian written in Nāgarī
References: Kadiri 1970, 57–58 and pl. XIV a

No. 54
Year: 1 Ramaḍān 1006 AH; Saṃvat 1655; 1 Chaitra-śudi [1598 CE]
Names: Nawwāb Raja Jagannathji; Khwāja Bhoginder Singh; Rao Chondaji;
Miyan Burhān; Mughal Period
Location: Ranthambore, Sawai Madhopur district, Rajasthan
Languages: Persian, local dialect
Material aspects: Nastaliq, Nāgarī; Persian occupies more space than Nāgarī,
which is below
Context: Eastern face of a dwarf pillar of marble in the right side near the foot-
steps of a small ruined mosque
Content: Official order discontinuing the levy in the form of commodities and
goods received for Muslim and Hindu charitable funds
Salient features: Nāgarī version has not been translated
References: Hussain 1973, 45–48 and pl. V a

No. 55
Year: (1018–1019 AH) 1010 Shuhūr-San [1609 CE]
Names: Burhān Niẓām Shāh III
Location: Shrivardhan Taluka, Raigad (earlier: Kolaba) district, Maharashtra
430 | Eva Orthmann

Languages: Persian and Marathi

Material aspects: Naskh and Nāgarī, both in relief; Nāgarī much longer than
Persian
Context: Fixed on the side of the road from Dive Āgar to Borlai Panchāyatan
Content: Sign post with indication of directions
Salient features: Marathi version begins with a transliteration of Persian. Indica-
tion that there is a ferry on the route to Ḍanḍā in the Marathi version
References: Kadiri 1970, 59–62 and pl. XV

No. 56
Year: No date
Names: Bisāṭ Khān; Niẓām Shāh
Location: Ahmadnagar, Maharashtra
Languages: Persian and Marathi
Material aspects: On two separate stones, about the same size
Context: Aṛhhāʿī Guṃbad Hājī Ḥamīd mosque and tomb complex, on the outer
face of the walls; Persian in relief, Marathi incised
Content: Endowment of land, especially a garden, for the lights of the mosque
and the dome
Salient features: Two versions very similar, including the formulation ‘whoever
contravenes may a donkey be on his wife’
References: Shaikh 1939–1940, 30 and pl. XIII b and c; Sohoni 2022, 9

No. 57
Year: 1018 AH; 1010 Shuhūr-San; 1531 Śaka [1609 CE]
Names: Jagapat Rāo Daulatī; Mirza Walī Amīr Barīd
Location: Bidar, Karnataka
Languages: Persian and Marathi
Material aspects: Thulth, on two different slabs; Marathi version much longer
Context: From a well at Āshtūr
Content: Building of a well with steps
Salient features: Both versions are very close to each other
References: Yazdani 1937–1938b, 2–3 and pl. II a and b
 Bilingual Inscriptions from India | 431

No. 58
Year: 20 Jumādā I 1033 AH; Saṃvat 1681; 1546 Śaka; 12 Vaiśākha-śudi [1624 CE]
Names: Laʿl Beg; Prince Dāwar Bakhsh; Jahāngīr
Location: Mangrol Junagadh district, Gujarat
Languages: Persian and Sanskrit
Material aspects: Nastaliq; Nāgarī below Persian. Much less space for Nāgarī
Context: Tablet lying loose in a house, now in the mosque
Content: Population of a suburb called Laʿlpūr, seeking cooperation of all; offi-
cials should look after it
Salient features: Sanskrit version refers to Jahāngīr by his name Salīm Shāh; not
fully deciphered
References: Desai 1970, 91–92 and pl. XXIII

No. 59
Year: 7 Dhū l-Qaʿda 1035 AH [1626 CE]
Names: Rāja Rām Dās; Jahāngīr
Location: Shivpuri, Shivpuri district, Madhya Pradesh
Languages: Persian and local dialect
Material aspects: Naskh and Nāgarī
Context: On a pillar lying in front of a temple in Baṛā Bazār Maḥalla in Purani
Shivpuri
Content: Farmān of the emperor about the remission of taxes
Salient features: No information about the Nāgarī version
References: Khan 1964, 79–81 and pl. XXV a; no photo of Nāgarī part

No. 60
Year: (1036 AH); 1018 Shuhūr-San (Persian); 1028 Shuhūr-San (Marathi); 1549
Śaka (Kannada) [1627 CE]
Names: ʿAbd ul-Muḥammad, son of Malik Raiḥān; ʿĀdil Shāh period
Location: Siruguppa, Ballari district, Karnataka
Languages: Persian, Marathi and Kannada
Material aspects: Naskh, Nāgarī, Kannada
Context: In the wall of the bastion near the Śambhulingaswāmin temple
Content: Erection of a bastion in the midst of the river (Persian); Marathi and
Kannada also mention erection of a well
432 | Eva Orthmann

Salient features: Marathi almost identical with Kannada; erection of well for god
Śambhu only in Marathi and Kannada
References: Desai 1953–1954b, 41–44 and pl. XIII a; Sohoni 2022, 11

No. 61
Year: Saṃvat 1694; 13 Chaitra-vadi [1627 CE]
Names: Dhā’u Chatrā and his wife Dhā’ī Lakmī Mahārāja Jai Singh; Shāh Jahān
period
Location: Bhaupura near Renwal, Jaipur district, Rajasthan
Languages: Persian and Rajasthani
Material aspects: Nastaliq, Nāgarī; Persian on the right, Nāgarī on the left; same
number of lines, similar arrangement; some additional lines in Nāgarī added later
Context: Wall of a step well at Bhaupura
Content: Building of a step well by two nurses (male and female = dhā‘u and dhā‘ī)
Salient features: Invocation of Rāma; dedication of the step well in Rajasthani to
Śrījī (Lakshmi or Vishnu), whose devotees the donors perhaps were
References: Prakash 1968, 67–68 and pl. XIV a

No. 62
Year: 1 Muḥarram 1049 AH; Saṃvat 1694; 3 Vaiśākha-śudi; Saṃvat 1695; 1 Phāl-
guna-vadi [1639 CE]
Names: Cho. Lakhmī Dās; Mahārāja Rāo Jaswant Singh; Shāh Jahān
Location: Bilara, Jodhpur district, Rajasthan
Languages: Persian and local dialect
Material aspects: Nāgarī was carved above the Persian, contrary to general prac-
tice; Nāgarī divided into two parts by Persian
Context: Inner face of the western arch of main cenotaph of Jodhpur rulers, on
the bank of the Tank at Banganga
Content: Construction of a tomb; amount of rupies spent
Salient features: Building of two tombs in Nāgarī; information on another 1501
rupies spent; Persian uses the abbreviation Cho., common in Sanskrit, but not
in Persian
References: Desai 2011, 16–19 and pl. III b
 Bilingual Inscriptions from India | 433

No. 63
Year: 14 Dhū l-Qaʿda 1054 AH; Saṃvat 1701; 1 Māgha-vadi; 1566 Śaka [1645 CE]
Names: Gokuladāsa and Dāmodaradāsa, sons of Qānungo Haridāsa of Māthura
kāyastha family; Shāh Jahān period
Location: Udaipur, Rajasthan
Languages: Arabic and Persian, Sanskrit
Material aspects: Nastaliq, Nāgarī; Nāgarī much longer and more prominent
Context: Right side wall of a step well known as Qanungo Baoli, eastern out-
skirts of Udaipur
Content: Construction of the step well by Gokuladasa and Damedaradasa
Salient features: In Sanskrit, reference to Ganesha and more exact information
on date, otherwise, very similar to Persian
References: Siddham database, https://siddham.network/inscription/inap00021/
(accessed on 4 September 2023)

No. 64
Year: 1061 AH; 25th regnal year of Shāh Jahān; Saṃvat 1708; 11 Bhādrapada
[1651 CE]
Names: Pahad Khān
Location: Gunaoti, near Makrana, Nagaur district, Rajasthan
Languages: Persian, local dialect
Material aspects: Nastaliq, Nāgarī; Nāgarī to the left of the Persian
Context: Western wall of a well known as Pahād Kunwā
Content: Excavation and construction of a well and foundation and settlement
of a village through the kindness of God and efforts of Pahad Khān in a locality
containing mines
Salient features: Nāgarī not deciphered no information on it
References: Desai 2011, 39–42 and pl. VIII a

No. 65
Year: 1062 AH [1651–1652 CE]
Names: Malik Shaikh ʿAlī; ʿĀdil Shāh period
Location: Mustafabad, Dabhol, Ratnagiri district, Maharshtra; today in the storage
of Chhatrapati Shivaji Maharaj Vastu Sangrahalaya (CSMVS) museum, Mumbai
Languages: Persian and Marathi
434 | Eva Orthmann

Material aspects: Bi-faced stele; inscriptions on the two sides

Context: Tablet opposite Jāmiʿ Masjid
Content: End of confiscation of property for men without son
Salient features: Reference to the Mokāshī, the Thānādār, the Ināmdār who
should not confiscate the property
References: Bendrey 1944, 151 (no plate); Sohoni 2022, 11

No. 66
Year: 20 Shawwāl 1063 AH; 1578 Śaka; 13 Bhādrapada-śudi (1575 Śaka would be
correct) [1653 CE]
Names: Mīr Jumla; Rustam, son of Dhū l-Fiqār; ʿAbdullāh Qutb Shāh
Location: Sriperumbudur taluqa, Poonamallee, Tiruvallur district, Tamil Nadu
Languages: Persian and Telugu
Material aspects: Nastaliq; Persian much more prominent than Telugu, which is
written in rather small letters at the bottom
Context: Carved on a mosque at Poonamallee; perhaps on a base that originally
belonged to a temple
Content: Building and completion of a mosque
Salient features: In Telugu, invocation to protect the mosque; promise of the
sixteen great gifts for those who protect it; those who cause obstruction will
incur the sin of killing a cow at Varanasi
References: Yazdani 1937–1938c, 52–54 and pl. XVIII b

No. 67
Year: 29 Muḥarram 1076 AH [1665 CE]
Names: Rāja Rāi Singh; Dūngarsī Kotwāl Rājpūt Gahalot, son of Rāo Amar
Singh; Aurangzēb
Location: Nagaur, Rajasthan
Languages: Persian and Nāgarī
Material aspects: Nastaliq in relief, Nāgarī incised; Nāgarī below, small, in the
margin
Context: Over an arch leading to the Zanāna Tāl at Nagaur
Content: Construction of a gateway called Darwāza-yi Islām
Salient features: Nāgarī defective in many places; seems to be similar to Persian
References: Chaghtai 1949–1950, 47–48 and pl. XVI c
 Bilingual Inscriptions from India | 435

No. 68
Year: 1077 AH; 1578 Śaka; 7 Māgha-śudi; 1589 Śaka [1666 CE]
Names: ʿAbdullāh Khān; Bukhārī Mīr-i Mīrān; Mūsā Khān; ʿAbdullāh Quṭb Shāh
Location: Golconda fort, Hyderabad, Telangana
Languages: Persian and Telugu
Material aspects: Naskh, Telugu; Persian in relief, Telugu engraved; Persian on
a greenish stone, Telugu on black basalt
Context: Mūsā Burj, back of the wall forming the western wing of the modern
steps, probably relocated; Telugu part on a basalt tablet broken into three piec-
es, in the southern wall of the bastion steps
Content: Shooting of Mīr-i Mīrān; conclusion of peace; building of a bastion
Salient features: Telugu version with some more details about Quṭb Shāhī officer
in charge of the bastion during the war
References: Yazdani 1913–1914, 51–55 and pl. XVII a and b

No. 69
Year: 5 Ramaḍān of 11th regnal year (of Aurangzēb) = 1079 AH [1669 CE]
Names: Rāja Gopāl Dās Goṛ; Rāja Manohar Dās; Shāh Jahān; Aurangzēb
Location: Thane, Maharashtra; today Chhatrapati Shivaji Maharaj Vastu San-
grahalaya (CSMVS) museum, Mumbai
Languages: Persian and a variation of Nāgarī/Rajasthani
Material aspects: Nastaliq, Nāgarī; Persian by far longer; Nāgarī on top of it and
on the right side of the slab
Context: Probably Maholi fort, found lying in the district court of Thane
Content: Services of Rāja Gopāl Dās Goṛ and Rāja Manohar Dās; reparation of
walls, granaries, etc.
Salient features: Nāgarī version much shorter; only about repairs at the three
forts of Maholi, Bhaṇdar Durg and Palāsgaḍh and who carried them out
References: Desai 1957–1958, 16–20 and pl. VI

No. 70
Year: 9 Rabīʿ II 1083 AH; Saṃvat 1729 [1672 CE]
Names: Muḥammad Dānish; Mahārāja Rām Singh; Aurangzēb
Location: Archaeological Museum, Amber, Jaipur, Rajasthan
Languages: Persian verse with Arabic prose; Sanskrit
436 | Eva Orthmann

Material aspects: Nastaliq, Nāgarī; Persian in relief, Sanskrit engraved on the

margins
Context: No exact location indicated
Content: Construction of a well by the eunuch Muḥammad Dānish
Salient features: No translation of the Sanskrit is given, only dates and names;
mentions a Thākūra and Rām Singh
References: Bukhari 1955–1956, 59–60 and pl. XV b

No. 71
Year: 14 Ramaḍān of the 23rd regnal year (of Aurangzēb) = 1090 AH [1679 CE]
Names: Randaula Khān; Aurangzēb
Location: Shahabad, Kota district, Rajasthan
Languages: Persian and local dialect
Material aspects: Nastaliq and Nāgarī; Persian above; Nāgarī part not complete-
ly preserved
Context: A loose pillar, previously fixed to a platform in front of the Kotwālī police
Content: Farmān of Aurangzēb regarding taxes and dispensation of specific
taxes after complaints by people of the town; text addresses mahājans, Brah-
mins and traders who had complained
Salient features: No information about edition of local dialect version
References: Khan 1968, 69–74 and pl. XV a

No. 72
Year: 1102 Faṣlī [1690–1691 CE]
Names: Dhondajī Kishan; Aurangzēb
Location: Beed district, Maharashtra
Languages: Persian and Marathi
Material aspects: Nāgarī
Context: Dhonda gate
Content: Name of the humble servant, year
Salient features: Texts both very short and identical to each other
References: Yazdani 1921–1922, 21–22 and pl. VIII b and pl. IX a
 Bilingual Inscriptions from India | 437

No. 73
Year: (1113 AH); 1625 Śaka; 1 Āsīn-śudi [1701–1702 CE]
Names: Ghāzī ud-Dīn Khān Bahādur; Aurangzēb
Location: Beed district, Maharashtra
Languages: Persian and Marathi; Marathi on first panel in semi-Sanskrit
Material aspects: Marathi carved on two panels, one beside the other; relative
position of the Persian and Marathi inscriptions not clear
Context: Purāna Pura
Content: Laying out of a new street and quarter, Ghāzī ud-Dīn Nagar
Salient features: First panel with list of titles of a person, described as having the
various titles of Śaiva divinity and representing the different phases of the god
Śiva; second panel not translated
References: Yazdani 1921–1922, 19–20 and pl. III b and IX b

No. 74
Year: 1 Jumādā I of the 2nd regnal year of Farrūkh Siyār = 1126 AH; 1637 Śaka;
14 Vaiśākha-vadi [1714 CE]
Names: Todar Mal Srī Niwās Dās son of Ṭekchand son of Hans Gopāl; Nawwāb
Saʿādatullāh Khān
Location: Chennai, Tamil Nadu
Languages: Persian and Sanskrit
Material aspects: Nastaliq; Sanskrit in late medieval Telugu characters; in-
scribed on both sides: one and a half occupied by Sanskrit, the rest by Persian;
three figures carved at the top, above the Sanskrit
Context: Copper plate from South Arcot district in Madras
Content: Five villages offered as a gift for the maintenance of a temple; Nawwāb
also visits the temple
Salient features: According to the Sanskrit version, Srī Niwās Dās is a protégé of
Todar Mal, while according to the Persian, this name is an alias of Todar Mal;
Sanskrit version mentions libation of gold and water; dates disagree
References: Desai 1955–1956b, 104–106 and pl. XXVII b; no photo of Sanskrit
version
438 | Eva Orthmann

No. 75
Year: 17 Rabīʿ II 1201 AH; 1708 Śaka [1787 CE]
Names: Govind Āppā
Location: Akola, Maharashtra
Languages: Persian and Marathi
Material aspects: Marathi inscription is much longer; Persian on the northern
side and Marathi on the southern side of the gate
Context: Delhi gate of Mokāsa Vēs
Content: Building of a gate
Salient features: Marathi mentions name of the city, pious men and Brahmans;
gate is described as large and beautiful; building of a ghāt is also mentioned
References: Haig 1907–1908, 17 (no plate)

No. 76
Year: 11 Shaʿbān 1245 AH; Saṃvat 1886; 13 Mah-śudi [1830 CE]
Names: Rāja Sohan Laʿl Bahādur Singh
Location: Delhi, today Archaeological Museum, Red Fort, Delhi
Languages: Persian and Sanskrit
Material aspects: Nastaliq, Nāgarī; both incised; carved cow feeding a calf en-
graved at the top; Sanskrit much shorter
Context: Sandstone pillar, originally belonging to a temple close to Purānā Qal‘a
Content: Endowment of land together with two pakka wells for a temple of Lord
Krishna; details of endowment
Salient features: Saṃvat dating also in Persian; Sanskrit text not translated;
temple for Śiva in Sanskrit and for Krishna in Persian
References: Bukhari 1959–1961, 19–21 and pl. VI a
Contributors
Manasicha Akepiyapornchai is currently an assistant professor in the Department of Asian
Studies at The University of Texas at Austin. Her primary research focuses on the medieval and
early modern intellectual history of South and South East Asia, particularly Thailand, with a
particular interest in religion, philosophy, and multilingualism. Her noteworthy past research
projects include the ‘Manuscript Culture of the Siamese Brahmins’ project, which was funded
by the OVPR-COLA Partnership, College of Liberal Arts, The University of Texas at Austin.

Viola Allegranzi is a postdoctoral researcher at the Institute of Iranian Studies of the Austrian
Academy of Sciences, Vienna. Her dissertation was devoted to the Persian inscriptions and
cultural history of the Ghaznavid period (Afghanistan, eleventh–twelfth century). Her research
focuses on epigraphy and architectural decoration from the premodern eastern Islamic regions
and explores the intersections between archaeological and textual sources.

Dmitry Bondarev is head of West Africa projects at the Centre for the Study of Manuscript
Cultures (CSMC) and senior lecturer at the Department of African Studies, Universität Ham-
burg. His research interests cover African linguistics (Saharan, Chadic and Mande languages),
literacy studies, especially in the context of Islamic exegetical traditions in African languages,
the history of writing in Arabic-based scripts (Ajami), and the palaeography and codicology of
West African manuscripts.

Emmanuel Francis is a research fellow at the Centre national de la recherche scientifique and
is affiliated to the Centre d’études sud-asiatiques et himalayennes (CESAH, UMR 8077, EHESS
& CNRS) in Paris. He is a historian of premodern Tamil Nadu, using inscriptions and manu-
scripts for his research on the social and cultural history of the Tamil language.

Jost Gippert was professor of comparative linguistics at the Goethe Universität Frankfurt from
1994 to 2022; since 2022, he has been senior professor at the CSMC, Universität Hamburg, and
PI of the ERC project ‘DeLiCaTe’ (‘The Development of Literacy in the Caucasian Territories’).

Volker Grabowsky has been professor of Thai and Lao studies at the Asia-Africa Institute,
Universität Hamburg, since 2009. He has published widely on the history, literary traditions
and manuscript cultures of Tai ethnic groups in Mainland South East Asia and South West
China. Since 2011, he has directed various projects pertaining to the manuscript cultures of the
Tai peoples in the upper Mekong valley. He is currently involved in a project documenting the
epigraphic landscape of Luang Prabang, Laos.

Antonio Manieri received his PhD in Japanese philology from Daito Bunka University, Tokyo, in
2012, sponsored by a MEXT scholarship. He is currently a senior researcher of Japanese stud-
ies (civilization of ancient and medieval Japan) at the University of Naples ‘L’Orientale’. His
research interests include lexicography and encyclopaedism, legal texts, and practi-
cal/technical knowledge in ancient Japan.
440 | Contributors

Leah Mascia is a postdoctoral researcher in Egyptology and Coptology at the CSMC, Universität
Hamburg. She is an archaeologist and epigraphist member of the archaeological mission of
the University of Barcelona working at the archaeological site of Oxyrhynchus (Egypt) and of
the archaeological mission of the University of Urbino, Italy, investigating ancient Cyrene
(Libya). She specialises in the study of written artefacts associated with religion, mortuary
practices, and magic in Greco-Roman and late antique Egypt.

Darya Ogorodnikova completed her PhD at the Universität Hamburg in 2021 and is currently a
research associate at the CSMC. Her research focuses on manuscripts with Soninke and Man-
dinka Ajami writings, their codicological characteristics, linguistic configurations, and con-
texts, and the networks of scholars behind their production and use.

Eva Orthmann is professor of Iranian studies at the University of Göttingen. She has been the
director of the Institute of Iranian Studies at the University in Göttingen since 2018. Her re-
search interests are related to the Mughal Empire, occult sciences, especially astrology, and
the Indo-Persian transfer of knowledge and culture.

Peera Panarut wrote a dissertation concerning the paratexts and transmission history of clas-
sical Thai literature. He is now working as a postdoctoral researcher at the CSMC, Universität
Hamburg, in a project on royal scribes and archival practices of the Thai royal court. His re-
search interests and publications cover Thai manuscript studies, Thai palaeography and epig-
raphy, textual criticism of classical Thai literature, and Thai book history.

Zsolt Simon is a Hittitologist, research fellow at the HUN-REN Hungarian Research Centre for
Linguistics, co-author of the Digital Philological-Etymological Dictionary of the Minor Ancient
Anatolian Corpus Languages, co-editor of the Hungarian Assyriological Review, and former re-
search fellow at the Koç University (Istanbul) and the Ludwig-Maximilians-Universität, München.
He is currently working for the Spanish research project ‘The gods of Anatolia and their names
(continuity, importation, interaction): a philological and linguistic approach’ in Barcelona.

Szilvia Sövegjártó is an Assyriologist and researcher at the CSMC, Universität Hamburg. Her
academic pursuits encompass the study of literary and lexical manuscripts from the third and
second millennia BCE, focusing on written language intertwining and multilingual contexts in
Old Babylonian Mesopotamia. Her current research focuses on the exploration of the earliest
colophons within the Sumerian literary and lexical tradition.

Apiradee Techasiriwan has been a scholar at the Archive of Lan Na Inscriptions, Chiang Mai
University since 2005. She has published on Lan Na (Northern Thai) epigraphy. She is currently
involved in the Fang river basin epigraphic and archaeological research project.

Jochen Hermann Vennebusch is a postdoctoral researcher at the CSMC, Universität Hamburg,

and is currently working on a habilitation project on north German bronze baptismal fonts from
the Middle Ages. After studying art history, history, Catholic theology and university didactics,
he received his doctorate in art history in 2020 with a thesis on the Ottonian and early Salian
Gospel books from the scriptorium of Reichenau monastery. His research focuses on medieval
 Contributors | 441

goldsmiths’ art and metal casting as well as book illumination, and he is also interested in
liturgical studies and the history of piety.

Márton Vér is currently a postdoctoral researcher at the CSMC, Universität Hamburg. His re-
search centres on the philology of written artefacts in Old Uyghur and Middle Mongolian, as
well as the social, cultural and administrative history of the Silk Roads, with a focus on the
ninth to fifteenth centuries CE. His present research project focuses on the cultural history of
Old Uyghur literacy in the light of multilingual and multiscriptual manuscripts.

Gábor Zólyomi is professor of Assyriology at the Eötvös Loránd University, Budapest, Hungary.
He specialises in Sumerian grammar, literature and cultural history. His latest project is ‘A
Dictionary of the Sumerian Royal Inscriptions of the third Millennium BCE’.
General Index
The primary aim of this general index is to assist the reader in finding concepts and terms of
interest, rather than creating a concordance. This is also why terms that appear too frequently
and would therefore be impractical as index headings are omitted as well as most personal and
geographical names. Page numbers with an asterisk refer to illustrations.

Francesca Panini

Abydos 335–336, 340 archival/archive 57, 63, 133, 144, 159, 173,
Achaemenid empire 331, 342–343 190, 301, 355–356*
(to) adapt/adaptation 1, 6, 22, 49, 206, Arñna 331
218, 221, 256–258, 262–263, 265, artisanal lexicon/terminology/vernacular
282, 285, 289–292, 294, 296–297, 4, 169, 175, 198–199
331, 333, 343, 352, 356, 365, 390–391, Asīrgarh 397, 399, 401, 426
401 audience 3–4, 6, 325, 358
administration 3, 42, 44, 46–47, 59, 101,
180–181, 255 Bactrian 353
Ajami 6, 255–260, 263–267, 283–286, baked brick 357
289, 292, 295–297, 299–300 Bambara 260
Akkadian 4, 6, 151–152, 154–163, 309–310, baptismal font 3, 75–76*, 77*–79*,
312, 314–224, 328 80–82*, 83–84*, 85–86*, 87–90*,
allograph 212, 218–223, 229*–230*, 91–92*, 93–94, 96*–98, 101–106
233*–234* Baroda 393
alphabet/alphabetic 5, 49, 205–215, basmala 28, 351, 378, 390, 401–402, 404
220–224, 227*, 331*, 233*, 240, 351, bell 26, 28, 83, 85, 104, 127–128*, 129*,
365, 380, 410 138–139, 141, 143, 363
amphiglossic 209, 210, 213–214, 223 Berber 258, 260–265, 289
amulet 58, 61 Berberised 262–263
Anatolia/Anatolian 331, 334, 342–346 Bidar 387, 389, 416, 430
anthology 93–94, 180, 238, 241, 309, 356 Bilara 401, 432
Arabic 3, 5–7, 13, 19, 21–23, 25–26*, biscript monolingual inscription 215, 223
27*–29, 31–33, 36, 38–39, 65, 67, blessing 211, 216–222, 228*, 290, 292,
255–270*, 271–272*, 273–275*, 354, 378, 391, 397, 404
276–277*, 278–300, 347–454, Book Pahlavi 350–351 see also Pahlavi
356–366, 371–381, 391, 395–397, Borno 256, 257, 267
399–400*, 401–403, 405–411, 413, borrowing 6, 255–257, 260–266,
416, 418–419, 423, 424, 426, 433, 435 268–269, 273–274, 276, 278–281,
Arabisms 261, 263 283–296, 299
Aramaic 6, 331, 333–338, 340–343, 387, Brahmanical 5, 236–242, 244–245,
404 248–252
archaisms 153, 161–163 Brāhmī alphabet/script 14, 206, 211

Open Access. © 2024 the author, published by De Gruyter. This work is licensed
under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
https://doi.org/10.1515/9783111380544-015
444 | General Index

Buddha figure/image/statue 115, 119*, colloquialisms 262

124–126*, 137–138 see also statue colophon 120, 133, 147, 152, 216–217,
Buddhism 13, 16, 112, 114, 132, 235–236, 223–225, 242, 283, 311, 313–316, 352
240 (to) commemorate 104, 351
Buddhist 5, 8, 13–15, 19, 22, 112, 114, 133, commissioner 133, 328
139, 141–143, 171, 174, 184, 235–236, compiler 183, 186, 191–192, 194, 196–197,
238, 240, 355 238, 242
bulla 55–56 conjunct grapheme 5, 205–206, 208–209,
bureaucratic 4, 173, 189–190, 194–195, 216–220, 222, 224, 228*–229*, 232*
197, 199 see also grapheme
Burhānpūr 373, 397, 399 conservatism 153
Burmese 111–113, 116, 120, 124–126*, 127, convention 123, 277, 282, 289
133, 235 copper 7, 19, 141, 143, 361, 371, 374, 379,
Busa 260 380, 410, 437
copperplate charter see charter
calligraphy 33, 174 copperplate grant 22–23, 25–27*, 28 see
caption 311, 315 also grant
caravanserai 354 Coptic 41, 49–51, 61–67
carved 57, 309, 311, 315, 350, 355, 413, 415, copy/(to) copy 36, 132, 156, 160–163,
422, 427, 432, 434, 437–438 179–180, 269–272, 275–276, 282,
carvers 402 313, 315, 317*
carving 53 copy-spelled/copy-spelling 263, 268, 282,
(to) cast 75, 78, 80, 83, 85, 89, 92, 97, 124, 285, 288, 298
143, 363 cultural heritage 156
cement board 130 culturemes 4, 198
Centalai 208, 211, 213 cuneiform 153, 162, 309–311, 325
ceramic 358 curricular/curriculum 155, 194, 264
character dictionary 176 curse 26, 314–315, 320, 382, 391–392,
charaktêres 60, 61 414, 423
charter 213 curvilinear letter/script 357–361, 363, 366
Chiang Mai 112, 114–117, 120–121*,
122*–126*, 127, 130, 132–136, 139, Dagbani 260
141, 143–145, 236 damma 353
Chinese 4, 5, 17, 111, 117, 127, 129*, 143, Dangaléat 260
169–171, 174, 179, 189, 194, 196, decree 6, 25, 116, 190, 331, 343
198–199, 374, 376, 380, 415 Demotic 43, 45, 48–55, 58–59
Christian period 41, 58 Dhivehi 3, 13–14, 16, 19–23, 25–27*,
Christianity 62–63, 68, 78, 94, 98, 100 28–29, 31, 33, 36, 38–39
classical 50 diacritic 259, 269, 358
classicisms 262–263 diacritical point 351, 357
clay 42, 53, 60–61, 83, 164, 185, 309 dialect/dialectal 23, 65, 154, 157, 172–173,
cliffs 256–257 198, 257, 267, 277, 281, 285, 289, 294,
code-mixing 199 318, 348, 374, 401, 429, 431–433, 436
coffin 56, 58–59
coin 14, 97, 353
 General Index | 445

dictionary 4, 169, 175–176, 179, 182–183, Fak Kham script 111, 113, 115–116*,
186, 189, 192–194, 196–198, 263, 285, 118–119*, 120
290, 325 fatkoḷu 28–30*, 31–32
digit 36 format/ formatting 6, 8, 151–153, 156–157,
digital humanities 5, 223 159, 161, 162–164, 173, 179, 198
diglossia/diglossic 171, 209–210, 213–214, formulaic 14, 396
216, 223 foundry inscription 91
direction stone 372, 380–381, 427 fudoki 171–172, 189–192, 195
Dives akuru script 3, 16, 19, 21, 25–26, Fula 258–260
28–29, 32–33, 36 Fulfulde 258
donor 77, 80, 91–93, 104, 256, 382, 387, funeral 25, 256
432 funerary chamber 56–57
ductus 55, 152–153 funerary inscription 33, 56–58, 67,
350–351, 365
Eastern Old Japanese see Japanese
edict 116, 171, 173, 190 Galna 381, 423–425
editor 65, 242 Gan island 22, 28, 32, 398
education 101, 130, 155–156, 164, 196–197, Germanised 98
257, 261, 264, 285 Ghaznavid 351–353, 355–356*, 357, 410
Egypt/Egyptian 41, 43–52, 56–60, 61–63, Ghurid 359–360
65, 68, 336, 340 gloss/(to) gloss 4, 49–51, 151–152,
(to) emboss 58 154–155, 157–160, 163–164, 173,
Emesal dialect 157 189, 191–192, 264–265, 267,
English 3–4, 9, 13–14, 38, 130, 174, 179, 283–284, 286–288, 290–296
196, 236, 239, 381 glossary 4, 116, 193, 198
epigraphic band 350 golden plate 123
epigraphic register 357, 360 Gothic 87–88
epigraphy 7, 113, 133, 256, 258, 348–349, graffito 53, 57, 59, 64–67, 353
351, 353, 357, 361, 364, 366, 371 grant 19–20, 22–23, 25–28, 80, 92, 224,
epitaph 7, 33–35, 56, 58, 64, 66, 371, 376, 376, 381–383, 393–394, 412, 414, 423
413 Grantha 5, 205–227*, 229*, 232*,
erudition 163, 165, 198 235–237*, 238–242, 247–249,
etymological/etymology 191, 255–256, 251–254
263, 265–266, 268–269, 271, Granthaised 213–214
274–277, 280–281, 283–285, 287, grapheme 5, 205–213, 216–225,
289–291, 292, 293–295, 297, 227*–229*, 232*, 259, 271–272, 274,
299–300 285–287, 289, 292, 294–295, 313
etymon 255, 257, 261, 266, 268–271, graphic style 350
276–277, 280–282, 284, 287, gravestone 33–34*, 35*
289–290, 292–294, 297 Greco-Roman period/phase 3, 41–44, 64,
Evēla akuru script 19, 21 67–68 see also Roman period
excerpt 14–15, 24–25, 29, 31–32*, 38, 151, Greek 2, 41, 43, 46–49, 51, 54, 56–68, 331,
156, 321, 349, 358 333–334, 337, 340–341, 343, 381,
387, 404
446 | General Index

Gujarat 372–373, 375, 391, 395–396, 283–285, 289–290, 297, 347–353,

411–412, 414–421, 431 355–357, 361, 365–369, 378, 383,
Gujarati 373–374, 380, 413–414 386, 391, 393, 396–399, 403–404,
421, 426
Ḥadīth 26, 28, 33, 358–359, 371, 399, 404 isogloss 318
hapax 196
Hausa 255, 258–261 Japan 4, 169–172, 175, 189, 193–198
Hebrew 62, 68, 380, 410 Japanese 4, 169–176, 179–180, 183,
Herat 362–363 188–189, 193–194, 196, 198
hermeneutic 156 Jewish 62
hermit statue 120, 122* see also statue
hierarchy 7, 210, 236, 350, 372, 378 Kanarese 374, 419
hieratic 45–47, 49–55 Kanembu 5, 255, 257–259, 267–270*,
hieroglyphic/hieroglyphs 45, 52–55, 58, 271–272*, 273–275*, 276–277*,
60, 335 278–283, 297–299, 301
Hindi 374, 377, 418 Kannada 373–374, 380–381, 420, 423,
Hithadhoo island 38 431–432
Hulhumeedhoo island 33 Kanuri 255, 257, 260–261, 267–268,
hybrid/hybridity/hybridization 4, 8, 15, 270–272, 275–276, 278–283
117, 151, 156, 162, 262–263, 289, 293, Karnak 54
349–351 Kassite period 151
Hyllarima 335–336 Khoiak festival 53, 55
hyperglossia/hyperglossic 209–210 Khòm script 120, 133, 235–237
hyphenated 402 Koagannu graveyard 33
Kufic script 350–351, 357–361, 366, 410
idiom 193, 235, 317, 321–322 Kyūshū dialect 172–174, 180, 190, 192
incantation 51, 61, 151, 155, 159–160
(to) incise 256, 351, 360, 362, 376–378, Lagid rulers 53–54
399, 413–414, 416–417, 419–422, 426, Lajim 349–350
430, 434, 438 Lan Na kingdom/region 4, 111–113*,
indentation 152, 161–162 114–118, 120, 123, 126, 130, 132–135,
Indic 7, 14, 348, 371–372, 375–379, 381, 144–146, 235
387, 390–391 Landhoo inscription 14–16*, 19
ink 28, 36, 54, 174, 187, 241 Lao script 111, 120, 133, 235
insertion 26*, 33, 177, 262–263, 265, 269, late antique period 3, 41–42, 44, 61–62,
276, 281, 285, 298–299, 348, 360, 371 67–68
instruction 98, 101, 197–198, 249 Latin 3, 38–39*, 56–60, 68, 75, 78, 80, 87,
integration 5–6, 22, 62, 111, 133, 255, 93–94, 97–98, 102, 105, 381
261–263, 265–266, 274, 282, 285, layout 51, 89, 379
297, 299 legibility 80, 97, 104, 357
interlinear 152, 154*, 156–158*, 159, legitimisation 357
162–164, 255–257, 263–264, 299 leporello manuscript 238, 240–241
Isdhoo island 19, 21–22 Letoon trilingual inscription 6, 331–332*,
Islam/Islamic 7, 13, 19, 21, 23, 32, 341, 343
255–258, 260–261, 264–266,
 General Index | 447

lexical 49, 156, 163, 173, 176, 192, 196, Mesopotamia 23

245, 256, 260, 262, 276, 300, 316, 324 metal 3, 38, 114, 328
lexicographic marker 176, 193, 196, 198, metalanguaging 260
324 metal plate see metal
lexicographical/lexicography 260, 283 metalwork 361–362, 364–365
lexicon 6, 156, 163, 196, 198, 255, 257, Middle Ages 78, 94, 98, 100–103, 105
268, 270, 285 Middle Babylonian period 4, 151–154*,
limestone 55, 331, 351, 403 155–158*, 159–165
lingua franca 235, 261, 276, 331 Middle/Classical Egyptian 50
linguistic community 293 Middle Persian see Persian
literacy 13, 50, 174, 236–237, 256, 297 miḥrāb 32, 397, 401, 426
literary tradition 5, 261, 263 miniature 94
liturgical book 98, 101–103 minuscule script 75, 87–88
loanword 5–6, 205, 209–212, 216–218, missionary 101
220, 223, 255–256, 259–260, 262, misspelling 61, 66
264–270, 274, 278, 280, 293, model 6–7, 56–57, 83, 90, 157, 164, 173,
297–300, 356, 371 180, 198, 347–350, 352–353,
lōmāfānu plate 19–20*, 21–23*, 24*–26*, 360–361, 363, 365–366, 405
28, 32 see also copperplate grant mokkan 174, 176–79, 188, 189, 192, 197
Low German 3, 75, 78, 80, 85, 88, 93, 97, monastery 14, 116, 120, 133, 144, 146–147,
105 172, 174, 183–184, 195
lustre-painted 358 monolingual 162, 164, 215, 223, 315,
Luwian/Luwic 335–336, 339–340 371–372, 376
Lycian 331, 333–335, 337–338, 340–342 monument 7, 13, 25, 56, 146, 211, 310–311,
Lydian 331, 336, 338, 341–342 314–316, 324, 348–350, 353–354,
358–359, 361, 365
Maba 260 Moore 260
magical artefact/text 51, 55, 60 Moṛi 374, 381, 424, 427
majuscule 89, 91 morpheme 212, 262, 269, 320, 338
Maldives 3, 13–14, 19, 25, 27, 32, 38 morphology 211–212, 223, 262, 265–266,
Male 25, 31–33, 39* 269, 287
Mandara 260 mosque 7, 19, 21–23, 25–26, 28, 31*–33,
Manding 259–260 295, 351, 360–361, 371, 376, 382,
Maṇippiravāḷam 210, 213–214 391–392, 395–401, 411–414, 420–
mantra 5, 14, 17, 235, 240–242, 243–252 421, 424, 426, 429–431, 434
Marathi 372–374, 380–383, 412–413, Muḥammad, king 28
422–423, 425, 427–433, 436–438 multiple-text manuscript (MTM) 310–311,
marble panel 355–356* 325
margin 58, 62, 153, 217, 219–220, 222, mummy label 59–60
229*–231*, 233*, 293, 377, 421, 424, Muslim era 7
434, 436
master copy 156–157, 160–64 Nabataean 62, 68
Mbay 260 Nāgarī 19, 353, 372–374, 377–378, 380–
medieval church 3, 75 381, 395, 397, 399, 401–402, 410–414,
memoria 91, 104–105 416–418, 420–436, 438
448 | General Index

necropolis 43, 51, 54, 56–61, 63–64, 66 Persian 3, 6–7, 13, 19, 21, 23, 36, 333,
New Persian see Persian 342–355*, 356*–366, 371–384,
Nippur 153–154*, 155, 159–160, 162, 313, 386–387, 390–394, 401–438
318, 321–322, 325 Persianate world 6–7, 348, 366
norm 7, 169–170, 214, 261, 404 Persianized 356, 359, 363
Phoenician 341
Old Akkadian see Akkadian phoneme 206–207, 215, 221, 259
Old Babylonian period 4, 151–161, 163– phonogram 4, 171, 173, 179–180, 182,
165, 309–310, 314–317, 324 185–189, 191, 193, 198
Old Coptic see Coptic phonology 19, 259, 265, 338
Old Kanembu see Kanembu pillar 213, 247, 371, 412, 416, 424, 427,
Old Khmer script 133, 235–236 429, 431, 436, 438
Old Persian see Persian Pisidian 335
oral/orality 171, 194, 197–198, 264, 267, polyglossic/polyglossy 216, 221
287 Prakrit 3, 13–16, 23, 211
ornamentation 83, 90, 350, 352, 358, 395 printed 98, 102–103, 105, 176–177, 191
orthography 66, 133, 158–60, 163, 239, (to) produce/production 6–7, 13, 28, 42,
249–250, 258–259, 262, 265, 44–46, 49, 51–55, 57, 61–63, 65–66,
269–271, 275–276, 279, 280, 68, 78, 80, 93, 123, 133, 154–155,
282–284, 289–291, 293, 297–298, 160–162, 164, 169–170, 178, 195, 198,
311, 318, 325 205, 215, 238–239, 242, 264, 267,
Osireion 52–54 283, 285, 287, 299–300, 347, 353,
ostracon 43, 45, 61 358–359, 361–363, 365–366,
Oxyrhynchite/Oxyrhynchus 3, 41–46, 372–374, 377, 380–381, 401,
48–54, 56–60, 62–65, 67–68 404–405
pronunciation 29, 49–50, 173, 197, 209,
Pahlavi 349–351, 374, 380, 410 215, 248, 252, 269, 282, 292
Pakistan 351, 353, 360, 373 prose 171, 193, 348, 414–416, 424, 435
palaeography 4, 7, 151, 159, 217, 350 provenance 42, 52, 160, 372
Pali 5, 111, 114, 115, 117, 120, 123, 133, 138, Ptolemaic period/phase 43–44, 48, 50,
140, 235–236, 241, 248, 252 52–58
Pali imaginaire 235–236, 241, 252 see also public sphere 133
Pali
palm-leaf manuscript 116–117*, 123 Qarakhanid 348, 353–354, 357–359
paper 3, 28, 31, 169, 174–175, 178, 181, qibla wall 397, 400*
184, 186–187, 192, 194, 216, 219, 238, quotation 26, 93, 94, 105, 176, 178–179,
240–241, 256–257, 291 214, 262–263, 350, 357, 360, 390, 404
papyrus 41–43, 45–48, 51, 60, 62–63 Quran/Qurʼan 22–23, 26, 28–29, 256–257,
paratext 216–217, 248, 251 269–271, 274, 278–282, 284, 286,
patron/patronage 6, 48, 116, 347, 351, 354, 297, 301, 350, 354, 360, 371, 381, 387,
359, 360, 363–365 390, 397, 399, 404
patronym 335 Quranic/Qur’anic 267–268, 273, 278,
(to) perform/performance 49, 52, 60–61, 283,295, 301, 350, 354, 357–358,
65, 236, 249 360, 403
Quṭb Minār complex 360
 General Index | 449

Radkan 349–350 scholar/scholarly 2, 4, 6–8, 149, 151, 153,

Rajasthani 374, 432, 435 155–157, 161, 171–172, 174, 176, 198,
readability 3, 21, 395 235–236, 238, 239, 265–266, 300,
reader 51, 104, 195, 237, 248–252, 337, 309, 316, 324, 340, 358, 372
354, 396 school 151, 155–156, 283, 295, 310, 422
recipient 6, 104, 256, 347 scribal choice 164, 280
(to) recite 49–50, 53, 97, 99, 101, 103, 120, scribal tradition 1
240, 245, 249, 251, 289, 357 (to) scribble 53, 58
(to) redact/redaction 48, 154, 157 script amalgamation 4
relative position 379, 395, 437 script community 172
relict 161, 165 scriptorium 50, 187
relief 75, 83, 311, 315–316, 376–378, 393, script size 152
397, 413–414, 416–418, 420–422, script switching 55, 66
425, 427, 430, 434–336 scroll 174, 181, 183–184, 186–187, 359
religion/religious 3, 5, 13, 22, 41, 44–46, seal 19–20*, 25, 28–29, 60
48–49, 52–54, 58, 62–68, 114, 120, secular 99, 114–116, 120, 132–133, 238
141, 143, 146–147, 223–224, 235–236, semantogram 189
240, 255, 264, 269, 277, 284–285, Semitic 342
297, 333, 356–357, 376, 390–392, Shāh Fażl mausoleum 354
394, 396, 399, 401, 428 Shan script 133, 235
Ribāṭ-i Malik 354–355* Shirol 381, 422, 423
Risgit 349, 350 Siam/Siamese 5, 111–112, 117, 126, 133,
ritual 5, 44, 46–47, 49–55, 61, 63, 65, 68, 145, 235–237*, 238–252
94–95, 97–103, 163, 174, 235–236, silver bowl 361
238–239, 242, 244–245, 247–252, silver plate 116–117, 123
397 Sinitic 4, 169–173, 175–176, 179–180,
ritual book 94, 98 see also ritual 182–183, 185, 187, 189–193, 196–199
Roman period 43–44, 46–49, 53, 55–56, Sinographic 172
59, 68 see also Greco-Roman period Sippar 157–158*, 160
Roman script 111, 130, 132*, 146 slab 55, 58, 65, 67, 351, 353, 377, 379–380,
royal monument 353, 365 382, 386–387, 394, 402–403, 415,
ruling 47, 152–153, 157, 159, 161–163 418, 430, 435
sociolinguistic 214, 262, 331
Safid Buland 354 Somnath 395–397, 411, 415–417
Samanid period 361 Songhay 256–258, 260–261, 295
Sanskrit 3, 5, 13–17*, 18*–23, 29, 173, Soninke 5, 255, 257–259, 267, 283–300
205–207, 209–223, 233*, 235–236, spelling 5–6, 14–15, 19, 29, 80, 214–215,
238–252, 351–353, 371–375, 377–378, 218–220, 221, 223, 255–256,
380, 383–384, 386–390, 392–400*, 259–260, 262–266, 268–269,
401, 403–423, 425–426, 431–433, 271–272, 274, 276–287, 289–299,
435–438 333, 335–338, 340–342, 351, 362,
Śāradā 351–352, 403–404, 410 403
Sarapeion 48, 64 sponsor 120, 127, 130, 133, 137
Sargon 309–311, 313–316, 319–324, 328 St Reinoldi, church of 3, 75–76*, 83, 97,
Sasanian period 350 104
450 | General Index

standardisation 153, 298 136–138, 140–147, 235–239,

statue 16–17*, 18*–19, 53, 56–57, 120, 123, 241–242, 244–245, 248–254
137–139, 145, 249, 310, 314–316, 319, Thailand 4–5, 111–112, 115, 123, 133, 137,
323, 328–329 235–237*, 238–240, 251–252
stela/stele 56–58, 65–66, 331–332, 434 Thai Nithet script 111, 116–117*, 122*–123
stone 3, 14–15, 31, 42, 65, 67, 113–115, 120, Tham script 111, 113–115*, 116, 118–119*,
206, 258, 353, 372, 380–381, 386, 120–121*, 122*–124, 126*–128*, 132*,
389, 395, 402, 411, 418, 421, 427, 430, 137–138, 140, 144, 146
435 theonym 6, 331, 333
stucco 350, 354, 357, 360 Theravada Buddhism 112, 235–236
Sukhothai 114–116, 123 Thoereion 44–45, 48
Sukhothai script 115–116, 123 tile 358–359
Sumerian 4, 6, 151–153, 155–167, 309–310, tomb 56–61, 339, 349–351, 353, 376, 393,
312–328 401–404, 410, 412, 415, 430, 432
sun letter 269, 280, 296 tombstone 256–257
tomb tower 349–350, 353
tablet 61, 151, 153, 159–164, 169, 174–175, transcription 14, 21, 33, 38, 136–138, 144,
178, 192, 194, 197, 276, 293, 309, 146, 177, 181, 186–187, 189, 195, 259,
311, 313–314, 317, 325, 371, 411, 431, 333–334, 336, 338, 342, 381
434–435 translation 4, 7, 17, 26, 43, 47–48, 55, 78,
Tai Yuan 111–112, 115, 117, 120, 123–124, 117, 123, 136, 138–140, 142–146, 151,
126, 140 152, 154–155, 157, 159–164, 172–173,
Tamasheq 258, 260, 293, 295 177, 179, 184, 187, 196, 209, 211, 213,
Tamil 5, 205–227*, 228*–229*, 240, 244, 246, 250, 255, 257, 264,
232*–233*, 236, 238–240, 248, 374, 267, 271, 276, 298–300, 317, 319–322,
380, 410, 415, 434, 437 324, 328, 333, 353, 363, 377, 381–387,
target language 196, 255, 258, 261–262, 389, 391–395, 397–400, 402–405,
264, 266, 298–299 411, 414, 423–424, 426, 436
technical diglossia 213, 216 see also di- transmission/(to) transmit 8, 48, 68, 106*,
glossia/diglossic 153, 194, 196, 293, 299–300, 342
technical terminology 197 Trichy 210
Teda 260 typology 5, 49, 56, 60–61, 152, 170, 186,
Tell Harmal 152 196, 255–256, 266, 297, 299
Telugu 211, 374, 385–386, 422, 434–435,
437 Uzgend 353, 358–360
Temne 260
temple 44–53, 55, 116–117, 137–138, 160, Vaṭṭeḻuttu 205–206, 210, 213, 410
183–184, 187, 238, 313–314, 328, 331, vernacular 3–4, 14–15, 19, 52, 75, 80, 91,
380, 386, 395–396, 411, 415, 421, 431, 94, 98–99, 102–104, 111, 114–117, 123,
434, 437–438 133, 169–173, 175, 179–180, 182–183,
textbook 4 185–186, 188–189, 191–199, 235–236,
Thaana script 3, 33, 36, 38 238, 242, 248–249, 271, 278, 309,
Thai 4–5, 111, 114–117, 120, 122*–123, 331, 336
125–127, 129*–130, 132*–134, verse 217, 241, 247, 249–252, 358,
361–362, 364–365, 274, 278, 282,
 General Index | 451

301, 348, 353–354, 357, 386,

388–390, 402, 413–414, 416, 418,
423–424, 435
version 6–7, 43, 46–48, 54, 59, 152, 155,
157–163, 175–176, 181, 183–184, 186,
189–190, 198, 309, 312, 314, 315–318,
320–325, 328, 331, 333–334, 336,
340–343, 350–351, 353, 375, 377–387,
390–396, 399, 401–405, 412–413,
416–419, 422, 424–425, 427–431,
435–437
virāma 208, 213–214, 217–219, 221, 225,
232*
visual organisation 152, 161–163
vocabulary 190, 197, 255, 260, 264–265,
268, 364
vocalisation mark 36
votive inscription/object 58, 64, 66, 309
vowel harmony 318
vowel-killer 206, 208, 213, 217, 219, 221,
232*

wall 7, 32, 42, 53, 57, 64–66, 146–147, 319,

328, 354, 358, 371, 376, 395, 397,
400–401, 411–412, 415, 417, 419–420,
423–425, 430–433, 435
wall painting 358
waqf 391–392, 395–396, 404, 411
well 371, 376, 380, 386–387, 390, 393,
401, 411–417, 419–421, 430–433, 436,
438
West Africa 256, 258, 260, 276
Western Old Japanese see Japanese
Wolof 258–260
wood/wooden 3, 31–32, 42, 58–59, 61,
114, 120–121*, 136, 169, 174–175, 178,
185, 192, 194, 197, 293, 423
wooden tablet see tablet
writing direction 28, 379, 403

Xanthos 331

Zalamkot 351, 352, 359, 402, 410