Abhidharma Across Buddhist Scholastic Traditions

i
Text, History, and Philosophy
© Koninklijke Brill NV, Leiden, 2016 | doi 10.1163/9789004318823_001

ii
Brill’s Indological Library
Edited by
Johannes Bronkhorst
In co-operation with
Richard Gombrich, Oskar von Hinüber,

Katsumi Mimaki, Arvind Sharma
VOLUME 50
The titles published in this series are listed at brill.com/bil

iii
Text, History, and Philosophy

Abhidharma across Buddhist Scholastic Traditions
Edited by
Bart Dessein
Weijen Teng
LEIDEN | BOSTON
iv
Cover illustration: Mogao Grottoes, cave 341, northern wall. Illustration of Maireya’s Pure Land, Early Tang
dynasty (courtesy of Dharma Drum).
Library of Congress Cataloging-in-Publication Data
Names: Conference "From Abhidhamma To Abhidharma. Early Buddhist

Scholasticism in India, Central Asia, and China" (2013 : Ghent
University), author. | Dessein, Bart, editor. | Teng, Weijen, editor.
Title: Text, history, and philosophy : Abhidharma across Buddhist scholastic
traditions / edited by Bart Dessein, Weijen Teng.
Description: Boston : Brill, 2016. | Series: Brill's Indological library,
ISSN 0925-2916 ; VOLUME 50 | "Papers ... of the conference 'From
Abhidhamma To Abhidharma. Early Buddhist Scholasticism in India, Central
Asia, and China' held on 8 and 9 July 2013 at Ghent University,
Belgium"--Preface. | Includes bibliographical references and index.
Identifiers: LCCN 2016013224 (print) | LCCN 2016022593 (ebook) | ISBN 9789004316669 (hardback :
alk. paper) | ISBN 978-90-04-31882-3 (e-book) | ISBN 9789004318823 (E-book)
Subjects: LCSH: Abhidharma--Study and teaching--India--History--Congresses. |
Abhidharma--Study and teaching--Asia, Central--History--Congresses. |
Abhidharma--Study and teaching--China--History--Congresses.
Classification: LCC BQ4195 .C66 2016 (print) | LCC BQ4195 (ebook) | DDC 294.3/824--dc23
LC record available at https://lccn.loc.gov/2016013224
Want or need Open Access? Brill Open offers you the choice to make your research freely accessible
online in exchange for a publication charge. Review your various options on brill.com/brill-open.
Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill-typeface.
issn 0925-2916
isbn 978-90-04-31666-9 (hardback)
isbn 978-90-04-31882-3 (e-book)
Copyright 2016 by Koninklijke Brill nv, Leiden, The Netherlands.

Koninklijke Brill NV incorporates the imprints Brill, Brill Hes & De Graaf, Brill Nijhoff, Brill Rodopi and
Hotei Publishing.
All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system,
or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise,
without prior written permission from the publisher.
Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill nv provided
that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive,
Suite 910, Danvers, ma 01923, usa. Fees are subject to change.
This book is printed on acid-free paper and produced in a sustainable manner.

v
In memory of Lance S. Cousins (7 April 1942–14 March 2015)
⸪
vi
Contents
Contents vii
Contents
Preface ix
List of Figure and Tables xi
Notes on Contributors xii
Introduction 1
Part 1
Mātṛkā and Abhidharma Terminologies
1 Abhidharma and Indian thinking 29

Johannes Bronkhorst
2 Abhidharmic Elements in Gandhāran Mahāyāna Buddhism: Groups

of Four and the abhedyaprasādas in the Bajaur Mahāyāna Sūtra 47
Andrea Schlosser and Ingo Strauch
3 Interpretations of the Terms ajjhattaṃ and bahiddhā: From the Pāli

Nikāyas to the Abhidhamma 108
Tamara Ditrich
4 Some Remarks on the Proofs of the “Store Mind” (Ālayavijñāna) and

the Development of the Concept of Manas 146
Jowita Kramer
Part 2
Intellectual History
5 Sanskrit Abhidharma Literature of the Mahāvihāravāsins 169

Lance S. Cousins
6 The Contribution of Saṃghabhadra to Our Understanding of

Abhidharma Doctrines 223
KL Dhammajoti
viii Contents
7 Pratītyasamutpāda in the Translations of An Shigao and the Writings

of His Chinese Followers 248
Eric M. Greene
8 Abhidharma in China: Reflections on ‘Matching Meanings’ and

Xuanxue 279
Bart Dessein
9 Kuiji’s Abhidharmic Recontextualization of Chinese Buddhism 296

Weijen Teng
10 Traces of Abhidharma in the bSam-gtan mig-sgron (Tibet, Tenth

Century) 314
Dylan Esler
Part 3
Philosophical Studies
11 Madhyamaka in Abhidharma Śāstras: The Case of Harivarman’s

*Tattvasiddhi 353
Goran Kardaš
12 Svalakṣaṇa (Particular) and Sāmānyalakṣaṇa (Universal) in

Abhidharma and Chinese Yogācāra Buddhism 375
Chen-kuo Lin
13 Perspectives on the Person and the Self in Vasubandhu’s

Abhidharmakośabhāṣya 396
Yao-ming Tsai
Index 413
PrefacePreface ix
Preface
The papers included in this volume are the outcome of the conference From
Abhidhamma To Abhidharma. Early Buddhist Scholasticism in India, Central
Asia, and China held on 8 and 9 July 2013 at Ghent University, Belgium. The
conference was initiated to commemorate the 120th anniversary – two 60-year
cycles in Chinese chronology – of the appointment of Louis de La Vallée Pous-
sin (1869–1938) as professor at Ghent University. During the almost 30 years he
was affiliated with Ghent University (he resigned in 1929), Louis de La Vallée
Poussin, son of a French father and a Belgian mother who chose to become a
Belgian national, laid the foundations of Belgian Buddhology. Louis de La Val-
lée Poussin is generally recognized as one of the greatest Abhidharma scholars
in the history of Buddhology. The contribution “Rétrospective: L’oeuvre de
Louis de La Vallée Poussin” in the Bibliographie bouddhique of 1955 enumerates
no less than 323 works of his hand, published between 1891 and his death in
1938.
The conference From Abhidhamma To Abhidharma. Early Buddhist Scholas-
ticism in India, Central Asia, and China was a joint initiative of Bart Dessein
and Weijen Teng, and was organized by the Ghent Centre for Buddhist Studies,
Belgium, and the Department of Buddhist Studies of Dharma Drum Institute
of Liberal Arts (Formally Dharma Drum Buddhist College), Taipei, Taiwan. Re-
searchers on Abhidhamma / Abhidharma work at different institutes all over
the globe, and their precise research activities concern different epochs of
Buddhist history, spanning from the life time of the historical Buddha to the
contemporary period; deal with Abhidharmic developments in different geo-
graphical regions, extending from India, over Southeast Asia, Central Asia, Chi-
na, Mongolia, Japan, and Tibet; and concern different types of materials, with
some researchers working on (recent) manuscript founds, and others working
on edited editions of texts in Pāli, Sanskrit, Chinese, Japanese, Mongolian or
Tibetan. While some scholars investigate the Abhidharma of one particular
school, others work in a comparative framework, or work cross-culturally.
Bringing together scientific papers of scholars dealing with such a wide va-
riety of subjects and materials in one coherent conference did not announce
itself to be an easy task. To our surprise, however, the conference papers did
show a surprising coherence, thus proving that despite the long chronological
development and the large geographical dissemination of Abhidhamma /
Abhidharma traditions, old Abhidharmic materials that are at the foundation
of the tradition have continued to determine and impregnate later develop-
x Preface
ments. The ‘World of Abhidharma’ proved to be a more interwoven complex

than one might expect.
This prompted us to publish an edited volume of a selection of conference
papers. The papers collected here concern the very beginnings of Abhidham-
ma / Abhidharma in India, and discuss the development of the genre and
Abhidharmic notions and concepts in India, Central Asia, China and Tibet
from the life time of the historical Buddha to the 10th century CE. As such this
volume forms a geographical (continental South, Central and East Asia + Sri
Lanka) as well as a philosophical unity, determined by the development of the
Sarvāstivāda Abhidharmic tradition and its ramifications.
The editors of this volume are grateful to the Fund for Scientific Research,
Flanders, the Faculty of Arts and Philosophy of Ghent University, and Dharma
Drum Institute of Liberal Arts for providing funds to organize the conference.
Thanks also go to Ann Heirman and Christophe Vielle for supporting our ap-
plication for funding with the Funds for Scientific Research, Flanders.
The editors of this volume especially want to thank Bhikkhunī Dhammadinnā
for her help and valuable suggestions during the editorial process.
Bart Dessein
Sint Amandsberg, Belgium
Weijen Teng
Jinshan, Taiwan
List of Figure and Tables xi
List of Figure and Tables
Figure
1 Extract from BajC2, part 1 (frame 2) 52
Tables
1 Development of lists based on the bodhipakṣya-dharmas after Bronkhorst,

1985 54
2 Sequence of dhyāna, apramāṇa and ārūpya in the different versions of the
Saṃgītisūtra/-paryāya 55
3 Comparison of the list in BajC2 with the groups of four occurring in the
Saṃgītisūtra/-paryāya 57
4 Listings in Prajñāpāramitā texts in comparison to the list in BajC2 61
5 Modular composition of the list in BajC2 67
xii notes on Contributors Notes On Contributors
Notes on Contributors
Johannes Bronkhorst
(PhD Pune 1979, doctorate Leiden 1980) is professor emeritus at the University
of Lausanne. He has published numerous research papers and books, includ-
ing Greater Magadha (Brill, 2007) and Buddhism in the Shadow of Brahmanism
(Brill, 2011). His book How the Brahmins Won (Brill, 2016) has recently come out.
Lance S. Cousins
was active at Oxford University until his death on March 14, 2015. He taught
Pali and Middle Indian at the Faculty of Oriental Studies, and Buddhism in the
Faculty of Theology. He is the author of numerous publications on Pali texts,
early Buddhist schools, Buddhist meditation and Buddhist ethics.
Dhammajoti KL Bhikkhu
is currently the Glorious Sun Professor at the Centre of Buddhist Studies, Uni-
versity of Hong Kong. He has taught Buddhism in several institutes around the
world including the University of Kelaniy, Sri Lanka; the University of Calgary,
Numata Chair in Buddhist Studies; and the International Buddhist College,
Thailand. Prof. Dhammajoti specialized in Theravāda and Sarvāstivāda Abhi
dharma, as well as Mahāyāna doctrines of the Madhyamaka and Yogācāra
schools. He has published many journal articles and is the author of The Chi-
nese Dharmapada, The Sarvāstivāda Abhidharma, and is the Editor of Journal
of Buddhist Studies, Sri Lanka.
Bart Dessein
is Professor at the Department of Languages and Cultures of Ghent University,
Belgium. His research focus is on philosophical developments in early Bud-
dhism and early school formation. He has published mainly on the
Sarvāstivādins and the Mahāsāṃghikas.
Tamara Ditrich
has been researching and lecturing Sanskrit, Pali and a variety of academic
subjects related to Asian religions and languages at several universities in Eu-
rope and Australia. Her research areas include Buddhist studies (mainly in
Pāli), Sanskrit linguistics and Vedic philology.
Notes on Contributors xiii
Dylan Esler
is a scholar and translator of Tibetan Buddhist texts. He holds an MA in Bud-
dhist Studies from the School of Oriental and African Studies (SOAS), London,
and is currently a research scholar at the Institut Orientaliste of the Université
Catholique de Louvain (Belgium), where he is preparing a PhD. He is working
on a translation and study of an important tenth-century Tibetan text on the
subject of meditation. His research interest focuses on early rNying-ma exposi-
tions of rDzogs-chen and Tantra. He is also an associate member of the Ghent
Centre for Buddhist Studies, Ghent University.
Goran Kardaš
is Assistant Professor at the Department of Philosophy and the Department of
Indology and Far Eastern Studies, Faculty of Philosophy and Social Sciences,
University of Zagreb, Croatia. His research interests are Buddhist philosophy
(mainly Abhidharma and Madhyamaka), Nyāya epistemology and Indian the-
ories of meaning.
Eric M. Greene
is Assistant Professor of Religious Studies at Yale. His research focuses on the
early history of Buddhism in China, with a particular emphasis on the trans-
mission of Buddhism to China, early translation practices, and the history of
Buddhist meditation. His articles have appeared in T’oung Pao, Journal of Chi-
nese Religions, Artibus Asiae, and History of Religions.
Jowita Kramer
completed a doctorate (Hamburg, 2004) and habilitation (Munich, 2010) in In-
dology. The main focus of her research lies on Indian and Indo-Tibetan Bud-
dhism. She has held positions at the Universities of Oxford, Heidelberg and
Göttingen in the past and is currently a research fellow at the University of
Munich.
Chen-kuo Lin
is a Distinguished Professor in both the Department of Philosophy and the
Graduate Institute of Religious Studies at National Chengchi University. His
research interest includes Buddhist philosophy (Buddhist logic and epistemol-
ogy, Yogācāra, Mādhyamika), Chinese philosophy (Neo-Confucianism, Dao-
ism), and comparative philosophy. In addition to three books and many
articles, recently he has published (co-edited with Michael Radich) A Distant
Mirror: Articulating Indic Ideas in Sixth and Seventh Century Chinese Buddhism
(Hamburg University Press, 2014).
xiv Notes On Contributors
Andrea Schlosser
studied Indian philology, Indian art history, and culture sciences at the Free
University in Berlin. She received her Master (Magister) degree in 2008 with an
edition of a South Indian copperplate donation record of the Western Gaṅgas.
As a research assistant she focused on Buddhist manuscripts, working with the
Sanskrit fragments of the Berlin Turfan Collection in cooperation with the In-
ternational Dunhuang Project, as well as participating in a project about the
edition of Kharoṣṭhī manuscripts from the Bajaur Collection. For her disserta-
tion, submitted in 2013, she edited two Gāndhārī manuscripts that illustrate
the practices of a bodhisattva. Currently, she is a research associate in the
project “Buddhist Manuscripts from Gandhāra”, where her focus lies on early
Mahāyāna texts.
Ingo Strauch
is Professor for Sanskrit and Buddhist Studies at the University of Lausannne.
He received his PhD (2000) and his habilitation (2010) in Indian Philology at
Freie Universität Berlin. His current research focusses on early Buddhist manu-
scripts from Gandhāra. He is currently preparing an edition of Vinaya and
Sūtra texts from the Bajaur Collection of Kharoṣṭhī manuscripts. Together with
Andrea Schlosser (Munich) he is editing an early Mahāyāna sūtra from
Gandhāra. His research also covers early Buddhist epigraphy in Brāhmī and
Kharoṣṭhī scripts.
Weijen Teng
received his PhD in the Study of Religion from Harvard University. He is cur-
rently teaching at Dharma Drum Institute of Liberal Arts, Taiwan. His interests
of research include Chinese Buddhist intellectual history, the study of Sanskrit
grammar in Chinese Buddhism, and Pāli Abhidhamma theory of meditation.
Yao-ming Tsai
Professor of Philosophy at National Taiwan University and editor-in-chief of
Taiwan Journal of Buddhist Studies, received his Ph.D. in Buddhist Studies in
1997 from the University of California, Berkeley. His teaching and research fo-
cus on Buddhist philosophy as well as philosophy of life. He is the author of
four books: The Teachings of Prajñāpāramitā and the Purification of the Bud-
dha-field (2001), An Open Path for Constructing Buddhology (2006), Research
Methods and Academic Resources for Buddhist Studies (2006), Philosophy of Life
and Worldview from the Perspective of Buddhist Teachings (2012), as well as of
dozens of scholarly articles and book chapters on Buddhist studies.
Introduction
Introduction 1
Introduction
Abhidharma as Rational Inquiry (Bart Dessein)
Although discussion exists on the precise meaning of the word Abhidhamma /

Abhidharma, the compounds abhi and dhamma / dharma suggest that Abhi
dhamma / Abhidharma is a historical development of the Buddha’s doctrine
(dhamma / dharma).1 The word Abhidhamma / Abhidharma is used both to
refer to the method in which the Buddhist doctrine is set forth, and to the genre
of literature in which this is done. In what follows, focus will not primarily be
on the creation of the so-called Abhidhammapiṭaka / Abhidharmapiṭaka, the
third of the three collections (Tipiṭaka / Tripiṭaka) in which Buddhist literature
is traditionally categorized (Sutta/Sūtra, Vinaya, Abhidhamma/Abhidharma),
nor on the development of the textual format of the works of the Abhidharma
genre, but on the Abhidharma as exegetical method.2
Discussing Abhidhamma / Abhidharma as a specific exegetical method,
this volume contains three parts. The first part, “Mātṛkā and Abhidharma
Terminologies,” contains contributions on the development of the Buddhist
argumentative technique and on the creation of lists of technical elements
(mātikā / mātṛkā) that are the outcome of the Buddhist exegetical method and
that are the argumentative material used in philosophical discussions. The sec-
ond part, “Intellectual History,” investigates the importance of the Buddhist
rational tradition for the development of Buddhist philosophy in the home-
land of Buddhism, as well as for the peculiar developments that were the result
of the contact of Buddhism with the philosophical traditions of Central and
East Asia. The third part, “Philosophical Studies,” focuses on some peculiar
doctrinal issues that resulted from rational Abhidharmic reflections.
For a tradition of rational inquiry to develop, thinkers have to accept the
legitimacy of questions and critique, even if these are directed against convic-
tions that are sanctioned by intuition, generally accepted truth, or revealed
truth. Rather than accepting ‘revealed truth,’ rational inquiry is aimed at
1 For an overview of scholarship on the meaning of the word ‘Abhidhamma / Abhidharma’: see
Malalasekera, 1961: 38–40; Willemen, Dessein & Cox, 1998: 13–15 and the contribution by Dylan
Eslar in this volume. For a recent study of the meaning of the term: see Anālayo, 2014: 70–71,
78–79, 171.
2 For discussions on the development of the Abhidharmic genre and the Abhidharmapiṭaka:
see Gethin, 1992 and Dessein, 2013. For a discussion of the development of the textual format
of the Abhidharma texts: see Cox, 1995: 29–37.

2 Introduction
‘revealing truth’.3 It is through rational thinking and argumentation that this

aim can be achieved. It is obvious that such a ‘rational attitude’ is stimulated
when a thinker “seriously contemplates the pervasiveness of the possibility
that he may be wrong, that he needs reasons and arguments to support the
validity of his views,”4 and that this situation especially occurs when con-
fronted with an intellectual opponent. The question as to the reason why a
tradition of rational inquiry – and of ‘philosophy’ for that matter – develops is
therefore a contextual one,5 and can be rephrased as follows: Where and when
were circumstances such that ‘philosophers’ were confronted with the possi-
ble fallibility of (their) traditional concepts, and / or with the need to convince
others of the correctness of their views? It is, further, precisely because rational
inquiry does not take convictions that are sanctioned by intuition, generally
accepted truth, or revealed truth as granted, that systems of rational thinking
have the possibility to cross the borders of the cultural context in which they
originated and first developed. The latter is important with respect to the fol-
lowing: scholarly opinion differs on the number of scientific traditions, that is,
traditions of rational inquiry as defined here, that have developed in the his-
tory of mankind.6 While the West Eurasian tradition (which includes the Greek
and Islamic traditions) and the Indian cultural tradition are generally accepted
to be characterized by an attitude of rational inquiry,7 the Chinese tradition
would arguably be devoid of this attitude.8 Buddhism is then accepted to have
considerably influenced the Chinese ‘philosophical’ tradition.9 These charac-
teristics of rational inquiry – its aim to reveal truth through rational thinking
and argumentation, and its possibility to cross cultural boundaries – are partic-
ularly important with respect to the present volume: Abhidharma as exegetical
method and its importance for the development of Buddhist ‘philosophy’.
3 See Bronkhorst, 2001: 34.

4 Harbsmeier, 1998: 261.
5 For a study of ‘context’: see Scharfstein, 1989. For reflections on the possible justification of
Abhidharma as philosophy: see Malalasekera, 1961: 48–49.
6 See Dessein, 2001: 97.
7 Staal, 1993: 16. For arguments for Egyptian, Babylonian, Hittite and Phenician influence on the
Greek tradition: see Needham, 1974: 55. For Greek philosophy’s indebtedness to Indian phi-
losophy: see Garbe, 1987: 36–46 and Conger, 1952: 103, 105, 107, 109–11. Przyluski, 1932: 286 has
laid emphasis on the Iranian (Persian) borrowings in both the Greek and the Indian culture.
8 Staal, 1989: 308.
9 See Zürcher, 1972; Ch’en, 1973; Frankenhauser, 1996; and Harbsmeier, 1998.
Introduction 3
Judging from early Buddhist texts in which especially the Jains are referred
to,10 it appears that in the region of Magadha where the historical Buddha lived,
the Jains (and sometimes also the Ājīvikas) were important religious competi-
tors of the Buddhists.11 This competitive context gives Buddhism a peculiar
position in the Indian religious/philosophical landscape, and is important for
our understanding of the development of Abhidharma as exegetical method
– above, we have claimed that the argumentative skills that religions/philos-
ophies develop in a competitive context are important for them to become
conversant and cross-cultural.12 Indeed, because texts of the Vedic oral/aural
tradition were primarily aimed at delivering a message to the realm of the Vedic
gods, who, by definition, did not need to be convinced of the truth revealed in
the Vedic texts, these texts do not show a development of argumentative tech-
niques. In contradistinction to the Vedic texts, the Brāhmaṇa prose texts – as
do the Upaniṣads – explain the offers that are performed. This suggests that,
with the development of Brahmanism, the religious audience were no longer
the mere passive spectators they had been in the Vedic period, but became
– at least passive – participators in a religious dialogue. It therefore may be
the development of Brahmanism that, in the Indian context, laid the basis for
what eventually would become a tradition of rational inquiry.13 This transfor-
mation should not be overstressed though, as the debates that are included in
the Brāhmaṇas and the Upaniṣads are not ‘rational’ in the sense that they want
to convince someone of the Brahmanic or Upaniṣadic truth. Their explana-
tions are for internal use.
While on the one hand pointing to the Indian oral/aural tradition in which
Buddhism originated,14 the introductory formula to many Buddhist sūtras
“evaṃ me sutaṃ” (“Thus have I heard”) also alludes to it that, for the early
Buddhists too, the word spoken by the Buddha primarily had the value of
‘revealed truth’ and was the guideline to be followed.15 The following passage
included in the Mahāparinirvāṇasūtra testifies of this: When Ānanda lamented
10 See, e.g. T.1.12: 227b7–10; T.26.1536: 367b8–24.

11 For some reflections on the Buddhist-Jain encounter: see Bronkhorst, 2011: 130–142. For
the different religious groups who were active contemporaneous with the Buddha: see
Hirakawa, 1990: 16–18.
12 See Assmann, 2003.
13 See von Simson, 1965: 139–141.
14 See Cousins, 1982: 1; von Hinüber, 1989: 22; and Dessein, 2012.
15 For the suggestion that also early Buddhist debates were primarily for internal use: see
Bronkhorst, 2001: 32–33. Bronkhorst, 2000: 23 remarks that, indeed, no matter how inter-
connected the Buddhist doctrine and Buddhist philosophy may be, what the Buddha
taught was a ‘doctrine,’ not a ‘philosophy’.
4 Introduction
that the doctrine would be without teacher after the Buddha had died – thus
urging the Buddha to appoint a successor – the Buddha is reported to have
replied that the doctrine and discipline he had taught them should also serve
as teacher after his demise.16 Also the following two viewpoints, attributed to
the Mahāsāṃghikas in the Samayabhedoparacanacakra (Yibuzong lun lun 異
部宗輪論) testify of this acceptance of the word of the Buddha as ‘revealed
truth’: “Everything that has been preached by the World-honored One is in
conformity with the truth (yathārtha),”17 and “The sūtras proclaimed by the
Buddha are all perfect in themselves (nītārtha)”.18 In his Yibuzong lun lun shuji
異部宗輪論述記, his commentary to the Chinese version of the Samayabhedo
paracanacakra, Kuiji 窺基 (632–682) explains these claims as that nothing the
Buddha has said is without benefit for living beings, and that, therefore, every-
thing he has said should be considered as the turning of the wheel of the
doctrine.19
Mentioning of the Samayabhedoparacanacakra (Treatise on the Development
of the Different Sects) brings us to one of the major peculiarities connected with
Abhidharma as exegetical method: the formation of Buddhist (Abhidharma)
sects. The territorial expansion of the Buddhist doctrine and the concomitant
different living conditions for Buddhist monks and nuns, must have made it
impossible for all monastics to live according to exactly the same set of rules
(Vinaya), and the Buddha’s choice not to appoint a successor must have invoked
different interpretations of the precise meaning of his sermons. Debates with
non-Buddhists must further have encroached on this tendency and must be
responsible for it that non-Buddhist – particularly Vedic and Jain – concepts
and elements were introduced in the Buddhist doctrine. All this eventually
resulted in the formation of different schools and sects.20 It is especially in the
Abhidhamma / Abhidharma literature that developed from the Sutta / Sūtra
and the Vinaya literature that these developments can be discerned.21
There is evidence that not only the formation of a corpus of sūtra texts, but
that also the Abhidharma practice originated as part of an oral/aural tradi-
16 Mahāparinibbāna suttanta, chap. 6, v.1.

17 T.49.2031: 15b28–29. See also Bareau, 1955: 58. See also T.27.1545: 912b.8–9
18 T.49.2031: 15c24. See also Bareau, 1955: 67.
19 See Kuiji: I 16b-17a and II 42a. For the discussion on which event should be considered as
the first turning of the wheel of the doctrine: see Dessein, 2007.
20 For the distinction between monastic ‘schools’ and philosophical ‘sects’: see Bechert, 1961.
21 For the development of Abhidharma from the Sūtra and Vinaya literature: see T.24.1451:
408b2–8.
Introduction 5
tion.22 It has hereby been suggested that it are the mnemotechnical skills of
the Brahmans who joined the Buddhist order that made the memorization of
the Buddhist texts possible.23 Although we may indeed assume that the oral/
aural tradition of the early Buddhist community with synods (saṃgīti) meant
to recite the words of the Buddha, occurred in a context very similar to that of
the Brahmanic tradition – early saṃgha members claimed that they had main-
tained in their memories the words as they had been spoken by the Buddha
himself and had, generation after generation, been publicly recited – it is very
likely that, contrary to general acceptance, Buddhism did not start as a reac-
tionary movement against Vedic Brahmanism. There is evidence that
Brahmanism only started to reach the northeastern domains of the Indian sub-
continent during the Aśokan period. It are the territorial expansion and
enhanced possibilities for adherents of all beliefs and faiths to freely travel the
country that characterizes this period that must have brought Buddhists in
contact with Brahmins.24 This claim is supported by the following: tradition
connects king Aśoka with the so-called Buddhist ‘synod’ (saṃgīti) that resulted
in the first schism in the Buddhist community – the one between the
Mahāsāṃghikas and the Sthaviravādins.25
Contrary to what was the case for Brahmanism, however, Buddhists, as
adherents of a new faith, were quickly confronted with the fact that, for their
religion to survive, non-Buddhists had to be convinced of the Buddhist truth
– one may be born a Brahmin, however, one is not born a Buddhist.26 The
necessity to convert non-Buddhists helps to explain why early Buddhists culled
the most important elements of the Buddhist doctrine from the Buddha’s ser-
mons, and organized them in lists of technical elements, the so-called mātṛkās
that are the fundament of all later Abhidharmic developments.27 These
22 See T.22.1428: 968b15 ff. This oral recitation is connected to Ānanda, one of the first five
disciples of the Buddha, in Sumaṅgala (Rhys Davids and Carpenter, 1968, vol.1: 17),
Aṭṭhasālinī (Bapat and Vadekar, 1942: 3), Samantapāsādikā (Takakusu and Nagai, 1975: 18),
T.1.1: 1a9–10; T.49.2030: 14b8; T.22.1428: 968b25–26; T.24.1463: 818a28–29.
23 See von Hinüber, 1989: 68.
24 See Bronkhorst, 2001: 2–4 and 8–11.
25 This event is likely to have occurred 116 years after the demise of the Buddha. See Nattier
and Prebish, 1976–77: 239, 271–272.
26 It can also be recalled here that many of the first converts used to be adherents of Jainism.
27 See Bronkhorst, 2000: 77. The Sarvāstivāda Vinaya, Shi song lü 詩誦律, T.23.1435: 449a19 ff.
identifies the Abhidharma with teachings that are presented in list form. For arguments
that the first mātṛkās predate the ‘age of Abhidhama’ and that the beginnings of the
Abhidharma seem to go back to an early period before the formation of schools: see
Anālayo, 2014: 21–25; 167–168.
6 Introduction
mātṛkās were doctrinally interpreted, and became the argumentative material

used in philosophical discussions. Mnemotechnical skills were undoubtedly a
major asset in remembering and expounding these lists. Also the questions-
and-answers format – illustrative for a tradition of rational debate – that is
frequently used in Abhidharma texts to explain the elements that make up a
mātṛkā facilitates the memorization of such lists.28 This questions-and-
answers format echoes the dialogue form of most of the Pāli Nikāyas and
Sanskrit (Chinese) Āgamas in which not only the Buddha, but also his chief
disciples had discussions with other monks and with non-Buddhists.29
The famous Milindapañha (Questions of King Milinda) records a debate that
allegedly took place between the Greek king Milinda / Menandros, member of
the Hellenistic tradition, and the Buddhist monk Nāgasena. Although Édith
Nolot (1995: 9) has argued that there has been a historical King Menandros
who reigned in the northwest of India around the second century BCE, but that
this historical figure is not the actual participant in the alleged discussion with
Nāgasena who, moreover, might be a fictitious person as well, this does not
diminish the importance of the fact that a text written in a language of north-
western India, records this debate and, thus, witnesses of the confrontation of
Buddhists with members of the Oriental Hellenistic tradition that flourished
in Central Asia starting from around 185 BCE. The date suggested by Édith
Nolot (1995: 9) of the second century BCE for the earliest version of such a text
might be too early – committing Buddhist texts to writing is likely to only have
started in the second century BCE,30 it remains a fact that the Milindapañha
testifies of an early philosophical debate.31 In its Chinese rendering, this work
that is likely to be dated in the beginning of the common era, belongs to the
Sarvāstivāda school of Buddhism that had Gandhāra as one of its strongholds.32
Gandhāra and Sri Lanka have become of particular importance for the two
main Abhidhamma / Abhidharma traditions: the southern Theravāda tradi-
tion, and the northwestern Sarvāstivāda tradition.
The encounter of the Sarvāstivādins of Gandhāra with members of the
Greek rational tradition of which the Milindapañha testifies, must have
inspired the Sarvāstivādins to systematize the word of the Buddha into a sound
28 See Norman, 1995: 309; von Simson, 1965: 142; Stache-Rosen, 1968: 8.
29 See Watanabe, 1983: 76. For the early appearance of the questions-and-answers format in
Buddhist literature: see Anālayo, 2014: 28; 167–168.
30 See Norman, 1992: 248 and Norman, 1993: 280.
31 The final version of this record was rendered into Pāli in the fifth century CE. See Finot,
1992: 7–9.
32 See Willemen, Dessein & Cox, 1998: 104–105.
Introduction 7
philosophical system, as creating their peculiar system was to all likelihood

seen as the only way to defend their faith against the well-developed Greek
tradition of rational debate. These particular circumstances may explain why
the development of a Buddhist rational tradition was especially prominent
among the Sarvāstivādins, why this tradition may have started in Gandhāra,
and why it are the Sarvāstivādins who have the most exhaustive collection of
Abhidharma texts.33 It is to the development of Gandhāran Sarvāstivāda think-
ing in its confrontation with the Hellenistic world, as well as to the influence
this development has had on other currents of thought in India – the gram-
marians, Jains, and Brahmins – that Johannes Bronkhorst devotes his attention
in his contribution “Abhidharma and Indian Thinking”.
Also Ingo Strauch and Andrea Schlosser study Abhidharmic developments
in the region of Gandhāra. In “Abhidharmic Elements in Gandhāran Mahāyāna
Buddhism. Groups of Four and the abhedyaprasādas in the Bajaur Mahāyāna
Sūtra,” they discuss the data contained in the so-called Bajaur Mahāyāna Sūtra,
a text in Kharoṣṭhī that was allegedly discovered in the ruins of a Buddhist
monastery near the village of Mian Kili at the Dir-Bajaur border. The text
belongs to one of three genres that can be considered as authentic products of
a Gandhāran literary tradition: Avadāna texts, scholastic and commentarial
texts, and early Mahāyāna sūtras. Investigation of this text that contains two
levels – a dialogue between the Tathāgata and his disciple Śāriputra that is
reminiscent of early Buddhist sūtras, and a ‘Mahāyānistic’ description of the
bodhisattva path – sheds light on the peculiar development of Mahāyāna
notions (especially the character of dharmas and of saṃjñā) in the region of
Gandhāra. The particular Abhidharmic developments in Kashmir, Gandhāra,
and Bactria (together also referred to as ‘Greater Gandhāra’34) can therefore be
interpreted as the coming to full maturation of a tendency that was already
present in early Buddhism.
In “Interpretations of the terms ajjhattaṃ and bahiddhā: From the Pāli
Nikāyas to the Abhidhamma,” Tamara Ditrich discusses the terms ajjhattaṃ
(internally), bahiddhā (externally) and ajjhattabahiddhā (internally and exter-
nally) as they are used in the Satipaṭṭhānasutta (Sutta on the Applications of
Mindfulness) of the Pāli canon, that is, as possible modes to practice application
of mindfulness on the body (kāyānupassanā), on feelings (vedanānupassanā),
on the mind (cittānupassanā), and on mental objects (dhammānupassanā).35
33 Beckwith, 2012: 24 remarks that “it is highly improbable that the Greeks and the early Bud-
dhists had no influence at all on each other”. See also Beckwith, 2012: 55–56.
34 See Salomon, 1999: 2.
35 MN I 55–63, DN II 290–315.
8 Introduction
A study of the Theravāda Abhidhamma literature shows that Buddhaghosa’s

(fifth century CE) interpretation of this early mātṛkā has, with some modifica-
tion, been followed by most modern scholars and practitioners who appear to
have developed a common understanding of the meaning of ‘internally’ and
‘externally’. This, however, does not seem to be the case for the compound term
ajjhattabahiddhā. Tamara Ditrich therefore suggests that diverging interpreta-
tions of the compound term may be due to the fact that the “Indian Buddhist
oral transmission may have adopted the already established stylistic and lin-
guistic patterns and markers (that is: patterns of the Vedic language), such as
the usage of dvandva compounds, for indicating specific contexts such as the
modes of contemplation, especially in relation to the four satipaṭṭhānas”. The
connection of a mātṛkā of three types of practice with a mātṛkā of four applica-
tions of mindfulness illustrates how, already in an early period of Abhidharmic
development, existing mātṛkās were brought into an ever more intricate con-
nection, with elements of one mātṛkā used to explain the elements of one or
another mātṛkā. The list of four avetyaprasādas discussed in Ingo Strauch and
Andrea Schlosser’s contribution mentioned above, is further a fine illustra-
tion of it that these mātṛkās as oldest fundament of the mature Abhidharma
texts,36 are a common heritage of both the Theravāda and the Sarvāstivāda
Abhidharma traditions.37 While the authors prove that the term abhedya was
first used in Gandhāra, and that there is good reason to believe that Sarvāstivāda
circles were the first to refer to the ‘four avetyaprasādas,’ the Bajaur Mahāyāna
Sūtra connects this concept with the intentional reflection on the positive
qualities of Buddha, Dharma, Saṃgha and morality (śīla). This reflection is
qualified as to its either having originated internally (adhyātmasamutthita), or
having originated externally (bahidhāsamutthita), or having originated both
internally and externally (adhyātmabahidhādamutthita). In this text, the well-
known Śrāvakayāna concept of the āryaśrāvaka and his avetyaprasādas is thus
linked with a Mahāyāna type of notions.
That Sarvāstivādin yogācāra masters inherited the Sarvāstivādin analysis of
dharmas and contributed, together with a certain section of the Sautrāntikas,
36 This explains the occurrence of the term ‘Mātṛkāpiṭaka’. See T.50.2042: 113c2–9; T.50.2043:
152a15. That a separate authoritative collection of mātṛkās – a Mātṛkāpiṭaka – must have
existed prior to the moment the originally orally transmitted texts were submitted to writ-
ing, is also suggested in the Mahāsāṃghikavinaya. See T.22.1425: 334c20–22.
37 See Frauwallner, 1964. In his contribution to this volume, Bhikkhu KL Dhammojoti sug-
gests that it could even be said that, to some extent, the Theravāda Abhidhamma develop-
ment cannot be satisfactorily understood without at least a basic understanding of the
possible influence from the broad Sarvāstivāda lineage. Also the contribution by Lance S.
Cousins hints in this direction.
Introduction 9
to the establishment of the Mahāyāna Yogācāra school is evident from the

contribution by Jowita Kramer. In “Some Remarks on the Proofs of the ‘Store
Mind’ (Ālayavijñāna) and the Development of the Concept of Manas,” she
discusses the change in interpretation of the mātṛkā of five aggregates (skan-
dha) under influence of the innovative Mahāyāna concept of the “store
mind” (ālayavijñāna). This is, more precisely, done through an examination
of Vasubandhu’s Pañcaskandhaka and its commentaries. Illustrative for the
process of debate and inquiry that characterizes later Abhidharma texts, it is
shown how the authors of these works sometimes try to compromise on the
varying teachings of Vasubandhu’s Abhidharmakośabhāṣya and Abhidharmic
works of the Yogācāra tradition, and sometimes appear to develop their own
interpretations.
It may be clear from the above that the Sarvāstivādins who lived in a differ-
ent context than the one that existed in the region of Vidiśā that was the
stronghold of the Theravādins before they migrated to Sri Lanka,38 enjoyed a
long period of philosophical development, and that their doctrines had an
immense impact on the development of Ṥrāvakayāna and Mahāyāna Buddhist
philosophy and methodology. Moreover, the influence of the Sarvāstivādins
was not restricted to India and Central Asia, but their method of rational
inquiry also had an impact on the Chinese tradition. It is to the issue of philo-
sophical influence that the second part of this volume, “Intellectual History,” is
devoted.
The mutual influence of different Buddhist traditions in India is addressed
in Lance S. Cousins’s contribution. In “Sanskrit Abhidharma Literature of the
Mahāvihāravāsins,” he examines the Sanskrit citations in the thirteenth cen-
tury sanne to the Pāli Visuddhimagga. He argues that at least some of these
citations prove that the seventh century Jotipāla, a major figure of the
Mahāvihāravāsin tradition who most likely came from the major center of
non-Mahāyāna tradition in the Tamil country, wrote in Sanskrit as well as in
Pāli. Confronted with the difficulty that they had to establish their legitimacy
as a reform movement – Mahāyāna doctrine differing from the early doctrine
in some essential points of theology – the early Mahāyānists of around the first
century BCE were to all probability the first to turn to the use of script to
expound and spread their views.39 It is then suggested that, in these circum-
stances, also the Theravādins started to write down their texts.40 The language
38 See Frauwallner, 1956: 18.

39 See McMahan, 1998: 251. Gombrich, 1990: 29 dates the earliest surviving Mahāyāna texts
back to the second or first centuries BCE.
40 See Norman, 1993: 280.
10 Introduction
in which this was done was Pāli. The use of Sanskrit in southern India is inter-
esting because it to all probability is Brahmanism that introduced the use of
Sanskrit in virtually the whole of South Asia and larger parts of Southeast Asia,
a development that is likely to be related to the fact that Brahmanism had
become a socio-political ideology, with Brahmins offering their services as
advisor to the political elites.41 With respect to the question of philosophical
interaction, the evidence provided in the sanne to the Visuddhimagga thus
opens a new perspective on the activities of the Brahmins, as well as of the
Theriya in South Asia in the later part of the first millennium CE.42
It is an important trait of Vasubandhu’s Abhidharmakośa that it criticizes
Vaibhāṣika viewpoints as they are expounded in the so-called vibhāṣā compen-
dia, the commentarial literature that was produced by the so-called Vaibhāṣika
Sarvāstivādins of Gandhāra. According to the Chinese monk and translator
Xuanzang 玄奘 (602–664), Saṃghabhadra (ca. fifth century CE) determined
to refute Vasubandhu’s critique on the Vaibhāṣikas.43 Bhikkhu KL Dhammajoti
discusses the importance of Saṃghabhadra in “Saṃghabhadra’s Contribution
to our Understanding of Abhidharma”. Kuiji, a disciple of Xuanzang, even
records and identifies Vaibhāṣika theories as “neo-Sarvāstivāda”.44 This makes
Saṃghabhadra, who is by Xuanzang’s pupils taken to represent the correct
Vaibhāṣika viewpoint, an important exponent of Abhidharma thinking. In this
way, Saṃghabhadra’s doctrinal expositions allow us insight in at least that sec-
tion of the Dārṣṭāntika Sarvāstivādins who developed into the Sautrāntikas.
The work of Saṃghabhadra is also important for the study of the doctrinal
evolution of the Mahāyāna, as his importance was acknowledged by the
Yogācāra masters (as also discussed in Jowita Kramer’s contribution) who took
him as representative of the true Vaibhāṣika viewpoint in their Sarvāstivāda
studies.
It is through the intellectual context of Greater Gandhāra that both
Śrāvakayāna and Mahāyāna Buddhism reached China around the beginning of
the common era. While the Chinese Buddhist monastic code developed to be
essentially the Dharmaguptaka code, the doctrinal school that had the largest
impact in China was without any doubt the highly rationalistic Sarvāstivāda
41 See for the latter: Bronkhorst, 2011: 42–43, 51–61.

42 Lance S. Cousins suggests that it must have been after the fourteenth century that the
Sanskrit literature from which Jotipāla drew was lost and could, probably, not be recov-
ered from Southeast Asia where Pāli rather than Sanskrit were more used for Theriya Bud-
dhist purposes, either having never been taken there or not having been preserved.
43 T.51.2087: 891c16ff.; T.50.2053: 232c22ff. See also T.41.1821: 11a20ff. and T.41.1822: 457c26ff.
44 T.43.1830: 271a10.
Introduction 11
school that, in its Vaibhāṣika interpretation, flourished in the Kuṣāṇa empire.

After Emperor Ming 明 (r. 58–75 CE) of the Han 漢 dynasty (206 BCE–220 CE)
had sent Zhang Qian 張騫 and Qin Jing 秦景 as envoys to Central Asia to estab-
lish an alliance against the Xiongnu 匈奴, thus opening the road to Central
Asia,45 Buddhist missionaries from the regions of Gandhāra, Parthia, Sogdiana,
Khotan, and Kučā reached China.46 After an initial period in which Buddhism
mainly flourished among the immigrants and merchants who had settled in
China, the second half of the second century CE saw the first translations of
Abhidharma texts into Chinese. Especially the Parthian monk An Shigao 安世高
(fl. 148–168) has to be mentioned here. It is to the translations of An Shigao and
his followers that Eric M. Greene devotes his attention in “Pratītyasamutpāda
in the Translations of An Shigao and the Writings of his Chinese Followers”.
This contribution shows how pre-Buddhist Chinese understandings of sense
perception – and particularly the relationship between sense perception and
desire – decisively influenced the early translations of Abhidharma texts into
Chinese. It is, more precisely, shown how the notion of sense contact (sparśa)
as a neutral process occurring prior to the arising of desire conflicted with the
traditional Chinese understanding that desire is located in the sense organs
themselves and seek out their desired objects independently from the will.
This contribution thus allows us to see how key Buddhist ideas were (or were
not) initially understood by the Chinese audience.
When Buddhism entered China, Confucianism had been elevated to official
orthodoxy. It has already been said repeatedly by different scholars that the
unification of the empire in the third century BCE and the subsequent victory
of Confucianism as orthodoxy in the Han dynasty have had a devastating
impact on the development of ‘rational inquiry’ as defined here.47 Scholarship
had become highly institutionalized and bureaucratized, and the philosophi-
cal profession had become organized in ‘schools’ (jia 家) of thought. In China,
books came to be seen as texts that were accepted by and commented upon by
a peculiar ‘school’ in accordance with that particular school’s interpretation of
the words of old sages.48 Membership of a ‘school’ was the prerequisite to gain
45 See Shiji 史記, 63.123: 345a2–11 and Han Shu 漢書, 61: 250b35-c1.
46 See Gernet, 1990: 188.
47 See Lloyd and Sivin, 2002: 27.
48 See Lloyd and Sivin, 2002: 73. Cheng, 1997: 318, note 4 remarks: “Concernant la notion
d’écoles dans la Chine ancienne, Nathan Sivin (cf. Philosophy East and West, 42, 1, 1992, 27)
remarque que, contrairement à la conception grecque de l’école formée d’orateurs et de
polémistes sur la place publique, elle correspond bien plus à des classifications bibli-
ographiques qu’à des groupements de personnes. En Chine, les écoles se distinguaient
entre elles en ce qu’elles préservaient et transmettaient des corpus différents de textes
12 Introduction
knowledge and understanding of this interpretation. Traditional Chinese phi-

losophy was therefore not characterized as a “search for truth,”49 as the truth of
the words of antiquity was considered to stand beyond doubt. Rather, tradi-
tional Chinese philosophy was aimed at engaging with the world, conceived as
a holistic whole, the good order of which is dictated by the past, and consists
in bringing a specific and transmitted interpretation of the words of the past
into practice.
The fall of the Han dynasty in 220 CE was a major challenge for the Confucian
scholar’s identity. This, on the one hand, as argued by Eric M. Greene resulted
in it that Buddhist writings of the period posterior to the Han reflect an effort
to find ways of talking about Buddhist theories that would be resonant with a
more general Chinese audience. On the other hand, this period also witnesses
a ‘conservative’ Confucian movement, with Confucians blaming a degener-
ate Confucianism for the fall of the Han dynasty. In “Abhidharma in China:
Reflections on ‘Matching Meanings’ and Xuanxue,” Bart Dessein describes how,
when the originally Abhidharmic exegetical technique of ‘geyi’ – explaining
Buddhist mātṛkās through matching them with traditional Chinese numeri-
cal lists – had proven unsuccessful to make clear the meaning of the Buddhist
doctrine, the technique might have continued to be used by ‘conservative’
Confucians who tried to redefine Chinese culture in a context of a growing
influence of Daoism and Buddhism. That the Abhidharma technique was
thus reinvented as a technique to redefine Confucian orthodoxy shows the
influence of the Abhidharmic rational debate on Chinese traditional philoso-
phy. That is to say, a technique that was originally used in a Buddhist context
was integrated in Confucian orthodoxy in so far as the social cohesion of the
Chinese world view was maintained.50
For the integration of Buddhist elements in Confucian doctrine also a
development within Buddhism itself was important: The Mahāyāna Buddhist
position that salvation can be reached in the sentient world did no longer
confine Buddhist adepts to monasteries. This opened the road for Buddhist
adherents to also become politically active – in this taking the path also
the Brahmins had done earlier.51 While some Buddhist monks thus became
increasingly involved in ‘Confucian’ politics, other monks continued to devote
écrits, dans une lignée de transmission qui ressemblait fort à une filiation (d’où le mot jia
qui désigne le clan)”.
49 Bauer, 2006: 17.
50 See Seiwert, 1994: 532.
51 See Bronkhorst, 2011: 236–237. For some concrete examples: see Forte, 2000: 9–10, 51; Des-
sein, 2003: 329–332.
Introduction 13
their time to Abhidharma studies. In this way, an Abhidharma School started

to flourish in fourth century South China. This Abhidharma School became
superseded by the Kośa School after Xuanzang and his translation team had
translated the famous Abhidharmakośa in 653 CE.52 The importance of these
Chinese Abhidharma and Kośa Schools as ‘conservative’ schools that tried to
set Chinese understanding of Buddhism aright by reorienting it toward Indian
Buddhist systems of thought – in this undoubtedly inspired by the direct con-
tacts with the Indian world that existed in Tang 唐 dynasty China (618–907),
is discussed by Weijen Teng. His “Kuiji’s Abhidharmic Recontextualization”
shows that Kuiji employs Abhidharma as an exegetical strategy to recontex-
tualize Chinese Buddhism, that is, to prove that this methodology can offer
a common and normative foundation for discussion and argument about
Buddhist doctrines and concepts.
Dylan Esler’s “Traces of Abhidharma in the bSam-gtan mig-sgron (Tibet,
Tenth Century),” outlines how the tenth century bSam-gtan mig-sgron, a trea-
tise on the subject of meditation and the first indigenous Tibetan doxography
is concerned with knowledge of the ultimate nature of reality and with the
methods of spiritual realization. Although not an Abhidharma treatise per se,
through its implicit and explicit references to the Abhidharmic framework, the
text reveals that it is Abhidharma that provides the metaphysical, cosmologi-
cal and phenomenological backdrop against which any form of Buddhist
meditative praxis takes place. Abhidharma therefore is not only the philosoph-
ical construct to explain the Buddhist doctrine and the framework within
which philosophers interpreted internal developments and external influ-
ences, but also the context that informs and defines Buddhist meditative
praxis.
In a period in which sectarian disagreements and discussions between dif-
ferent Buddhist groups had become prominent, the practice to sanction one
specific interpretation of the meaning of the elements enumerated in a given
mātṛkā against another interpretation is likely to have gained importance.
Sectarian disagreement and a transition from an oral/aural transmission of
the word of the Buddha to a written transmission with new means to build
up argumentation, also opened new perspectives for the development of the
textual format of the texts in which ‘philosophical’ positions were outlined.
Developed from simple mātṛkās that had been culled from sūtras and Vinaya
texts, Abhidharma texts increasingly became polemical, and became char-
acterized by an increasing “attempt to cogently summarize salient doctrinal
52 On the history of the Chinese Abhidharma and Kośa Schools: see Dessein, 2010: 57–62,
66–67.
14 Introduction
positions and yet also refute, point by point, positions thought to represent
rival groups”.53 In this process, the original mātṛkās were not only adapted
to suit doctrinal interpretations and developments, but they also became
the force of doctrinal development themselves.54 New interpretations of the
transmitted material were also included in commentarial texts on earlier
Abhidharma texts, and sectarian self-identification also culminated in the cre-
ation of well-structured pedagogical manuals of Buddhist doctrine in which,
through sophisticated methods of argument, the own position is established
and the view of others is refuted.55 This development of peculiar philosophical
positions is the subject of the third part of this volume: “Philosophical Studies”.
Harivarman lived in the third-fourth century and is likely to have belonged to
the Bahuśrutīyas, a sub-group of the Mahāsāṃghikas.56 He is credibly attested
as being a pupil of Kumāralāta who, himself, was connected to the Sautrāntika
development of the Sarvāstivādins.57 By the time of Harivarman, the early
Buddhist dharma theory – the Sarvāstivādin atomic world view that all that
exists is composed of phenomena (dharma) that are themselves subject to
constant change, and that therefore all that exists is devoid of intrinsic nature
(svabhāva) – had experienced a major reinterpretation. Rational inquiry could
not but lead to the advanced opinion that when all that exists is composed
of phenomena, it should necessarily be so that these phenomena (themselves
dharmas too) should equally consist of further dharmas. The only logical out-
come was that all that exists was taken to be empty (śūnya).58 It especially
were the Madhyamaka philosophers, the most important of whom are the sec-
ond-third century Nāgārjuna and his disciple Āryadeva, who took this thinking
to its logical and methodological extreme. As masters of logic argumenta-
tion, they even dared to use logical argument to discuss the proper dogmas of
Buddhism. “Madhyamaka in Abhidharma Śāstras: The Case of Harivarman’s
*Tattvasiddhi” is Goran Kardaš’s investigation of the Madhyamaka view-
points in Harivarman’s *Tattvasiddhiśāstra. In this Abhidharma treatise that
is one of the very few to refer to the Mādhyamikas, Harivarman is seen to put
Madhyamaka argumentation – that most likely stems from Āryadeva – in favor
of the non-existence of things in the mouth of his opponent who, through this,
appears as a Madhyamaka adherent. This contribution thus witnesses of the
53 Cox, 1995: 30–31.

54 See Gethin, 1992: 161.
55 See Cox, 1995: 35.
56 See Potter, 1999: 255–256; Dessein, 2009: 39–40.
57 T.55.2145: 78c9f.; T.45.1852: 3c11–14. See also Willemen, Dessein & Cox, 1998: 107–110.
58 For more on this development: see Bronkhorst, 2000: 139–148.
Introduction 15
third-fourth century intellectual debate between adherents of the Ṥrāvakayāna

and early Mahāyānists.
One of the consequences of the growing importance and popularity of
Mahāyāna Buddhism in China was that Abhidharma studies rapidly started to
decline after the eighth century.59 Although Abhidharma studies disappeared
as a separate branch of Buddhist activity, the Abhidharmic methodology con-
tinued to live on in the Chinese Mahāyāna schools. In “Svalakṣaṇa (Particular)
and Sāmānyalakṣaṇa (Universal) in Abhidharma and Chinese Yogācāra
Buddhism,” Chen-kuo Lin outlines how the seventh-eighth century Chinese
Yogācārins disagreed with the theory the Ābhidhārmikas Kuiji and Huizhao 慧
沼 (632–682) proposed on whether liberation through meditation can be
attained when taking the universal characteristics of things as object or when
taking the particular characteristics as object, and how they appropriated
Abhidharma teachings in their Yogācāra theory of truth and liberation. In its
focus on Yogācāra Buddhism’s relation to the Sarvāstivāda, this chapter cor-
roborates the importance of the latter for the former also Bhikkhu KL
Dhammajoti discerned, and its subject – the mātṛkā of meditation leading to
liberation – is also the subject matter discussed in the contributions by Tamara
Ditrich and by Ingo Strauch and Andrea Schlosser.
Another pedagogical manual is the already mentioned Abhidharmakośa
by Vasubandhu. In “Perspectives on the Person and the Self in Vasubandhu’s
Abhidharmakośabhāṣya,” Yao-ming Tsai examines Vasubandhu’s (ca. fourth-
fifth century CE) approach to the nature and identity of the person and the self.
It is, more precisely, shown how Vasubandhu’s thesis of the selflessness of a per-
son (pudgala-nairātmya) is a critique of the Pudgalavādins (those who accept
the existence of a person) or the Tīrthikas (heretics). Discerning a develop-
ment from an ‘Abhidharmic’ approach to a ‘not-dharmic’ approach, Yao-ming
Tsai argues that the ninth chapter of the Abhidharmakośabhāṣya reveals the
significance of the role played by soteriology in Buddhist intellectual thought.
He makes clear that the issue of not-self is not about an existential differentia-
tion, but about metaphysical identification: while it is conventional knowledge
to speak of the five aggregates that constitute a person and that explain why
there is no such thing as a ‘self,’ it is wise to state that these five aggregates
are in reality not the five aggregates and are only designations of provisional
naming without any substantial or ontological distinction per se. It may also
be recalled here that Vasubandhu was a native of Puruṣapura in Gandhāra,
the region we characterized as having been important for the development
59 See Dessein, 2010: 67.

16 Introduction
of ‘Buddhist philosophy’ and for the maturation of Mahāyāna Buddhism.60

This development towards the metaphysical might have been stimulated by
the development of the dharma theory outlined above, and by the fact that,
as mentioned, the Brahmins had taken up a political role, a possibility that
was unavailable to the adherents of Śrāvakayāna Buddhism that is overall
characterized by a denunciation of mundane life. This lack might have made
it possible that Buddhism became more oriented towards the metaphysical.
It is thus clear that Abhidharma studies, their long history and diverse
contexts of development notwithstanding, continued to deliberate over the
same fundamental Buddhological questions. Abhidharma was at once the
philosophical construct to explain the Buddhist doctrine as the framework
within which philosophers interpreted internal developments and external
influences.
Methodological Remarks (Weijen Teng)
Thus far, we have signified the very nature of the Abhidharma tradition as an
enterprise of rational thinking, sketched its historical transmission and trans-
formation across South and East Asia, and outlined the arrangement and the
main points of the chapters. The following part of the introduction highlights
the methodological significance of this edited volume as a whole, and teases
out some of the methodological features reflected in the individual chapters.
The compilation of this volume, which covers cross-regional and longue
durée studies of Abhidhamma / Abhidharma traditions, appears to encapsu-
late a methodological intuition – a demonstration of how the Abhidharma
study, which is often envisioned as a lofty, hypertechnical, and relatively mar-
ginal subject within Buddhist Studies, could be approached in a way that could
contribute to a broader inquiry into the Humanities. Given the intention of
offering some new lights to the study of the Abhidharma tradition method-
ologically, this volume does not claim to offer a methodological ‘paradigm
shift’. As the three thematic divisions of this volume suggests, the authors have
adopted the methods that are fundamental to the study of Abhidharma study,
namely a meticulous textual criticism and historical philology, contextualized
intellectual history, and doctrinal philosophy. Nonetheless, in their studies,
these authors take up issues that go beyond the Abhidharma tradition itself
and therefore reveal verily some methodological significances and potentials.
We will point out these methodological features shortly, but before that, we
60 See Willemen, Dessein & Cox, 1998: 270.

Introduction 17
will try to signify some methodological advantage this volume as a whole

encapsulates.
One methodological advantage of this volume apparently lies in its cover-
age of the birth, growth, and changes over time and space of this Buddhist
scholastic enterprise we call ‘the Abhidharma tradition’.61 This volume can be
imagined as a satellite in space: in its surveying down the river system of the
Abhidharma, an overall picture is produced that enables us to connect its
remote sources with its ends, and to trace its turns and stops across regions. In
this way, through Johannes Bronkhorst’s delineating of the intellectual net-
work of Gandhāran Abhidharma, we can connect the Hellenistic philosophical
world not only to the South Asian world of the Brahmanical philosophies and
to the Tamil country and Ceylon addressed in Lance S. Cousins’s study, but also
across East Asia, to the seventh century Abhidharmic recontextualization by
Kuiji in Weijen Teng’s study, and to the Chinese seventh-eighth century Chinese
Yogācāra Buddhism that Chen-kuo Lin studied. We can even connect the
Hellenistic philosophical methodology to the otherwise untraceable end of
“geyi Confucianism” in which the Abhidharmic technique was adopted by
Confucians to redefine their orthodoxy, a case studied in Bart Dessein’s chap-
ter. In Ditrich’s trace of the genealogy of the Pāli syntagms ajjhattaṃ, bahiddhā
and ajjhattabahiddhā, a stylistic current that originated in the Vedic textual
tradition flows into the Pāli Nikāyas through the oral practice of mātikā, and
was later formulized and formalized by Buddhagosa in the later Theravāda
Abhidhamma literature. Through the vintage point this volume stands upon,
the complex nature of these intellectual connections, that has an enormous
coverage in time and space, reveals, in a much fuller manner, the cross-cultural
influence and confluence of the Abhidharma tradition in the intellectual his-
tory of the globe.
Another methodological feature this volume presents, is that many chapters
offer potential contributions to the areas of studies that go beyond the subject
of Abhidharma scholasticism itself. In general, the subject of Abhidharma is
studied in its own right, whereby the origins of the Abhidharmic enterprise
are sought, its philosophies and doctrines are interpreted, and its intellectual
61 The enterprise of Abhidharma is universally treated as the Buddhist “Scholasticism”. To

name a few book titles or chapters that contain this usage: Kalupahana, 1967, Chapter 9
“Scholasticism, Theravāda, Sarvāstivāda, and Sautrāntika”; Aviv, 2014, Chapter 9 “Ouyang
Jinwu: From Yogācāra Scholasticism to Soteriology”; Willemen, 2006; Willemen, Dessein
& Cox, 1998. As a scholastic enterprise, Abhidharma studies could naturally fit in the
broader field of Comparative Philosophy of Religion. See José Cabezón (ed.), 1998, Scho-
lasticism: Cross-Cultural and Comparative Perspectives.
18 Introduction
dynamics are revealed. This is surely a legitimate and necessary approach.

However, the study of Abhidharma also has the potential to go beyond its
own philosophical and religious significance. Through Johannes Bronkhorst’s
as well as Ingo Stauch and Andrea Schlosser’s studies of the rational form of
intellectual debate and its reception and development in “Greater Gandhāra,”
the cultural richness and complex intellectual network of this ancient region
can be inferred. To some extent, Lance S. Cousins’s study of the “Sanskrit
Abhidharma Literature of the Mahāvihārāvāsins” likewise enhances our
knowledge of the intellectual history of the Theriya school of Southern India
and Ceylon. The materials these authors document and analyze constitute a
cogent case for Randall Collins’ “sociology of philosophies,” and thus engage in
the dialogue with the “global theory of intellectual change” that he proposed.62
A varied methodological move of going beyond Abhidharma itself can be
discerned in the history of East Asian Abhidharma. As demonstrated in this
volume, Abhidharma is taken not only as a scholastic system of thought per se.
It is also viewed as a peculiar genre of Buddhist literature as is the case in Eric
M. Greene’s chapter, as an exegetical technique by Bart Dessein, and as an exe-
getical strategy by Weijen Teng. Methodologically speaking, Eric M. Greene
rightly takes the ‘problematic’ Chinese translations of Buddhist terms seri-
ously, a methodological move that contrasts with the studies that simply
dismiss such problematic Chinese translations as doctrinal deficiency, and
thrive for the ‘authentic’ meanings of the Indic terms. In his seeking out seman-
tic and/or cultural explanations for the “mistranslated” terms, Greene makes
an interesting case for the study of “cultural translation,”63 vis-à-vis literal ren-
dering, and he hence offers an alternative approach to the study of Chinese
translations of Buddhist texts.
Bart Dessein’s and Weijen Teng’s approaches to Abhidharma are similar in
that they do not focus on any particular Abhidharma literature or theory.
Instead, they approach Abhidharma as an exegetical technique or strategy
employed by fourth-fifth century and seventh century Chinese intellectuals
respectively, and they try to bring about the historical significance of this tech-
nique or strategy for the intellectual world of medieval China. Dessein’s study
is particularly interesting in that the Chinese Buddhist pedagogy in relation to
Abhidharma, that is, geyi, unintendedly helped some ‘conservative’ Confucian
literati to redefine the Confucian orthodoxy of the time. In the same vein,
Teng’s approach to Abhidharma looks into the methodological nature of
62 For example, he proposes that “conflict is the energy source of intellectual life, and con-
flict is limited by itself”. Collins, 1998: 1.
63 For a general discussion of the term, see Pym, 2010, Chapter 8 “Cultural Translation”.
Introduction 19
Abhidharmic scholasticism, and shows how Kuiji attempted to construct a

normative foundation for the Chinese Buddhist intellectual practice.
Rigorous philosophical and philological approaches to Abhidharma stud-
ies are not lacking from this volume. Due to their close investigations into
important Abhidharma texts that have not received due attention from
Abhidharma scholars, the discoveries and arguments advanced in these stud-
ies improve our knowledge of the Abhidhamma / Abhidharma tradition. In
the case of Jowita Kramer, the primary text of the Pañcaskandhakavibhāṣā
was recently edited and published by the same author.64 Her study of this text
with a comparison primarily of the Pañcaskandhaka, the Yogācārabhūmi, and
the Mahāyānasaṃgraha, enables us to trace the various lines of the develop-
ment of the concepts of kliṣṭamanas and ālayavijñāna from early Buddhism to
Mahāyāna via Abhidharma developments.
It is worth mentioning that the text Goran Kardaš studies, the *Tattvasiddhi,
is a Sanskrit text that has been reconstructed from its Chinese translation,
the Chengshi lun 成實論. The *Tattvasiddhi probably is the only extant Mahā
sāṃghika Abhidharma text, whose importance in terms of the intellectual
history of the Abhidharma tradition is evident. Methodolgocially speaking,
Kardaš’s choice and study of the Sanskrit reconstruction, instead of the
‘original’ Chinese translation, offers some interesting perspectives for textual
criticism when the Sanskrit text is compared with its Chinese translation.
We have already seen that several papers in this volume work on Chinese mate-
rials. How Chinese materials can be important for more historically informed
and philosophically engaged Abhidharma studies is made explicit in Bhikkhu
KL Dhammajoti’s and Chen-kuo Lin’s studies. They bring to our attention,
though not for the first time in this field of studies, the Abhidharma literature
persevered only in Chinese translation as well as the Abhidharmic discussions
found in Chinese Buddhist commentaries. Through his close examination of
the Chinese Abhidharma translations, in his chapter, the *Nyāyānusāra (Shun
zhengli lun 順正理論) and the *Abhidharmasamayapradīpikā (Apidamo xian-
zong lun 阿毗達磨顯宗論) of Saṃghabhadra, Dhammajoti demonstrates how
Abhidharma scholars can be misled by relying solely on the Indic Abhidharma
literature, for instance, the Abhidharmakośa, and by overlooking the vast cor-
pus of Chinese translations of Abhidharma texts, in particular Xuanzang’s
玄奘 translation of the gigantic Abhidharmamahāvibhāṣā, a fundamental
Abhidharma text of the Sarvāstivāda school.
In his philosophical study of the Chinese Yogācārins of the seventh-eighth
century, Chen-kuo Lin introduces the scholastic and philosophical discussions
64 Kramer, 2014.
20 Introduction
found in the indigenous commentaries to the Abhidharma and Yogācāra texts.

Lin’s study cogently demonstrates the value of these Chinese commentaries to
the extent that they do not only continue some core philosophical debates of
the Abhidharma and Yogācāra traditions, but also offer new and plausible phil-
osophical insights which might be informed by the Chinese world-view. In
terms of the diversity of research material as a methodological gesture, we
should also mention Dylan Esler’s study of a tenth century Tibetan treatise, the
bSam-gtan mig-sgron composed by gNubs-chen Sangs-rgyas ye-shes. That also
literature that does not necessarily possess an Abhidharmic outlook might
contain useful references to the Abhidharma philosophy, is shown in his
chapter.
The methodological remarks presented so far, fall into the categories of
‘descriptive’ or ‘interpretative’ (or explanatory). From a methodological per-
spective, Yao-ming Tsai’s approach stands out uniquely. The main thesis in his
study on the Person and the Self in the Abhidharmakośabhāṣya relates to an
interesting methodological method less common in the field of religious stud-
ies. Tsai’s study expresses an urge to re-orientate the perspective in the study of
Buddhist philosophy from ‘Abhidharmic’ to ‘non-Abhidharmic’. With his
emphasis on ‘first hand’ observation and his caution for an “Abhidharma ori-
entation” which tends to “organize, classify, analyze and compare conceptual
factors,” Yao-ming Tsai takes side with the domain of the phenomenology of
religion. His urge to side with a phenomenological approach for an apprecia-
tion of the Buddha-dharma seems to be less a philosophical interpretation or
description than a ‘normative’ proposition. It will be quiet interesting to see
how Yao-ming Tsai’s approach could re-open a discussion on the problematics
and the appropriateness of the ‘normative discourse’ in Buddhist studies as an
academic discipline.65
Abbreviations
DN Dīghanikāya (ed. Rhys Davids and Carpenter, [1890–1911] 1995–2007)

MN Majjhimanikāya (ed. Trenckner and Chalmers, 1888–1902)
T. Taishō shinshū daizōkyō 大正新脩大藏經 (ed. Takakusu 高楠, Watanabe
渡邊 and Ono 小野, 1924–1934)
65 For a previous discussion on this issue of “normative discourse”, see Cabezón, 1995: 231–
268.
Introduction 21
Bibliography
Anālayo (2014). The Dawn of Abhidharma. Hamburg Buddhist Studies 2. Hamburg:

Hamburg University Press.
Assmann, Jan (2003). Die Mosaische Unterscheidung oder der Preis des Monotheismus.
München/Wien: Carl Hanser.
Aviv, Eyal (2014). “Ouyang Jinwu: From Yogācāra Scholasticism to Soteriology”, In
Transforming Consciousness: Yogācāra Thought in Modern China. Edited by John
Makeham. Oxford: Oxford University Press.
Bapat, Puruṣottama V. and R.D. Vadekar (eds.) (1942). Aṭṭhasālinī. Bhandarkar Oriental
Series, no.3. Poona.
Bareau, André (1955). Les sectes bouddhiques du petit véhicule. Paris: École Française
d’Extrême-Orient.
Bauer, Wolfgang (2006). Geschichte der chinesischen Philosophie. München: Verlag C.H.
Beck.
Bechert, Heinz (1961). “Aśoka’s Schismenedikt und der Begriff ‘Saṃghabheda’”, Wiener
Zeitschrift für die Kunde Süd- und Ost-Asiens 5, pp. 18–53.
Beckwith, Christopher I. (2012). Warriors of the Cloisters. The Central Asian Origins of
Science in the Medieval World. Oxford: Oxford University Press.
Bronkhorst, Johannes (2000). “Die buddhistische Lehre”, in Der Buddhismus I. Der indi-
sche Buddhismus und seine Verzweigungen. Edited by Heinz Bechert et al. Stuttgart:
W. Kohlhammer, pp. 23–213.
Bronkhorst, Johannes (2001). “Pourquoi la philosophie existe-t-elle en Inde?” La ratio-
nalité en Asie / Rationality in Asia. Études de Lettres 2001/3, pp. 7–48.
Bronkhorst, Johannes (2011). Buddhism in the Shadow of Brahmanism. (Handbook of
Oriental Studies 2/24). Leiden – Boston: Brill.
Cabezón, José (1995). “Buddhist Studies as a Discipline and the Role of Theory”, Journal
of the International Association of Buddhist Studies. Volume 18. On method,
pp. 231–268.
Cabezón, José (ed.) (1998). Scholasticism: Cross-Cultural and Comparative Perspectives.
(SUNY Series, Towards a Comparative Philosophy of Religions). Albany NY: State
University of New York Press.
Ch’en, Kenneth ([1964] 1973). Buddhism in China. A Historical Survey. Princeton NJ:
Princeton University Press.
Cheng, Anne (1997). Histoire de la pensée chinoise. Paris: Seuil.
Collins, Randall. (1998). The Sociology of Philopsophies: A Global Theory of Intellectual
Change. Cambridge MA: Harvard University Press.
Conger, G.P. (1952). “Did India Influence early Greek Philosophies?”, Philosophy East and
West II,2, pp. 102–128.
22 Introduction
Cousins, Lance S. (1982). “Pali Oral Literature”, In Buddhist Studies – Ancient and Modern
(Collected Papers on South Asia No. 4). Edited by Philip Denwood and Alexander
Piatigorsky. London: Curzon Press, pp. 1–11.
Cox, Collett (1995). Disputed Dharmas. Early Buddhist Theories on Existence. An Annotated
Translation of the Section on Factors Dissociated from Thought from Saṅghabhadra’s
Nyāyānusāra. (Studia Philologica Buddhica. Monograph Series XI). Tokyo: The
International Institute for Buddhist Studies.
Dessein, Bart (2001). “Climbing a tree to catch fish: Some reflections on Plato, Aristotle,
and China”. La rationalité en Asie / Rationality in Asia. Études de Lettres 2001/3,
pp. 97–125.
Dessein, Bart (2003). “The Glow of the Vow of the Teacher Samantabhadra ‘Puxian pusa
xing yuanzan’ (T.297) *Samantabhadrācāryapraṇidhānarāja”, Acta Orientalia 56/2–4,
pp. 317–338.
Dessein, Bart (2007). “The First Turning of the Wheel of the Doctrine – Sarvāstivāda and
Mahāsāṃghika Controversy”, In The Spread of Buddhism. (Handbuch der Orientalistik.
8. 16). Edited by Ann Heirman and Stephan Peter Bumbacher, pp. 15–48.
Dessein, Bart (2009). “The Mahāsāṃghikas and the Origin of Mahayana Buddhism:
Evidence Provided in the *Abhidharmamahāvibhāṣāśāstra”, The Eastern Buddhist
New Series 40/1–2, pp. 25–61.
Dessein, Bart (2010). “The Abhidharma School in China and the Chinese Version of
Upaśānta’s Abhidharmahṛdayasūtra”, The Eastern Buddhist, New Series 41/2,
pp. 49- 69.
Dessein, Bart (2012). “‘Thus Have I Heard’ and Other Claims to Authenticity: Development
of Rhetorical Devices in the Sarvāstivāda Ṣaṭpādābhidharma Texts”, In Zen Buddhist
Rhetoric in China, Korea, and Japan. Edited by Christoph Anderl. Leiden: Brill,
pp. 121–162.
Dessein, Bart (2013) “Lists as Dynamic Devices in Early Buddhist Doctrine and Textual
Tradition”, Antiquorum Philosophia. An International Journal 7, pp. 29–48. (Special
Issue: Forme del catalogo).
Finot, Louis (1992). Les Questions de Milinda. Paris: Gaillimard.
Forte, Antonino (2000). A Jewel in Indra’s Net. (Italian School of East Asian Studies.
Occasional Papers 8). Kyoto: Istituto Italiano di Cultura. Scuola di Studi sull’Asia
Orientale.
Frankenhauser, Uwe (1996). Die Einführung der buddhistischen Logic in China. Opera
Sinologica 1. Wiesbaden: Harrassowitz.
Frauwallner, Erich (1956). The Earliest Vinaya and the Beginnings of Buddhist Literature.
(Serie Orientale Roma 8). Roma: Istituto Italiano per il Medio ed Estremo Oriente.
Frauwallner, Erich (1964). “Abhidharma Studien”, Wiener Zeitschrift für die Kunde Süd-
und Ost-Asiens 8, pp. 59–99.
Garbe, Richard (1987). The Philosophy of Ancient India. Chicago IL: The Open Court.
Introduction 23
Gernet, Jacques ([1972] 1990). Le monde chinois. Paris: Armand Colin.

Gethin, Rupert M.L. (1992). “The Mātikās: Memorization, Mindfulness, and the List”, in
In the Mirror of Memory. Reflections on Mindfulness and Remembrance in Indian and
Tibetan Buddhism. Edited by Janet Gyatso. Albany NY: State University of New York
Press, pp. 149–172.
Gombrich, Richard (1990). “How the Mahāyāna Began”, In The Buddhist Forum. Vol. I.
Seminar Papers 1987–1988. Edited by Tadeusz Skorupski. London: School of Oriental
and African Studies, pp. 21–30.
Han shu 漢書, Ershiwu shi 二十五史 (1988). Shanghai: Shanghai guji chubanshe.
Harbsmeier, Christoph (1998). Science and Civilisation in China. Vol. 7: Pt. 1: Language
and Logic. Cambridge: Cambridge University Press.
Hinüber, Oskar von (1989). Der Beginn der Schrift und frühe Schriftlichkeit in Indien.
(Abhandlungen der Geistes- und Sozialwissenschaftlichen Klasse; Jahrgang 1989,
Nr.11). Mainz/Stuttgart: Franz Steiner Verlag.
Hirakawa, Akira (1990). A History of Indian Buddhism. From Ṥākyamuni to Early
Mahāyāna. (Asian studies at Hawaii 36). Honolulu HI: University of Hawaii Press.
(Translated and Edited by Paul Groner).
Kalupahana, David (1967). Buddhist Philosophy: A Historical Analysis. Honolulu, HI:
University of Hawai’i Press.
Kramer, Jowita (2014). Sthiramati’s Pañcaskandhakavibhāṣā, Part 1: Critical Edition. Part
2: Diplomatic Edition. 2 volumes. Wien/Beijing: Verlag der Österreichischen
Akademie der Wissenschaften/China Tibetology Publishing House.
Kuiji 窺基, Yibuzong lun lun shuji 異部宗輪論述記, Dainihon zokuzōkyō 大日本続蔵経,
1905–1912. 150 vols. Kyoto: Zōkyō shoin [Reprint: Shanghai: Shangwu yinshuguan,
1923]. ZZ1–8.
Lloyd, Geoffry E.R. and Nathan Sivin (2002). The Way and the Word. Science and Medicine
in Early China and Greece. New Haven CN and London: Yale University Press.
Makeham, John (1994). Name and Actuality in Early Chinese Thought. (SUNY Series in
Chinese Philosophy and Culture). Albany NY: State University of New York Press.
Malalasekera, G.P. (1961). Encyclopedia of Buddhism (Fascicule A-Aca). Published by the
Government of Ceylon.
McMahan, David L. (1998). “Orality, Writing, And Authority in South Asian Buddhism:
Visionary Literature and The Struggle for Legitimacy in the Mahāyāna”, History of
Religions 37/3, pp. 249–274.
Nattier, Jan and Charles S. Prebish (1976–77). “Mahāsāṃghika Origins: The Beginnings
of Buddhist Sectrianism”, History of Religions 16, pp. 237–272.
Needham, Joseph (1974). La tradition scientifique chinoise. (Collection Savoir). Paris:
Hermann.
Nolot, Édith (1995). Entretiens de Milinda et Nāgasena. Traduit du pâli, présenté et annoté.
Paris: Gallimard.
24 Introduction
Norman, Kenneth R. (1992) “The Development of Writing in India and its Effect upon
the Pāli Canon”, Wiener Zeitschrift für die Kunde Südasiens und Archiv für indische
Philosophie 36, pp. 239–249.
Norman, Kenneth R. (1993). “Der Beginn der Schrift und frühe Schriftlichkeit in Indien.
By Oskar v. Hinüber. (Abhandlungen der Geistes- und Sozialwissenschaftlichen
Klasse; Jahrgang 1989, Nr.11) pp. 75. Mainz, Akademie der Wissenschaften und der
Literatur; Stuttgart, Franz Steiner Verlag GmbH, 1990)”, Journal of the Royal Asiatic
Society (3rd series), vol.3, part 2, pp. 277–281.
Norman, Kenneth R. (1995). “Oskar von Hinüber: Untersuchungen zur Mündlichkeit
früher mittelindischer Texte der Buddhisten. Untersuchungen zur Sprachgeschichte
und Handschriften-Kunde des Pali III. (Abhandlungen der Geistes- und Sozial
wissenschaftlichen Klasse, Jahrgang 1994, No. 5). Mainz/Stuttgart (Akademie Der
Wissenschaften und der Literatur, Mainz. Franz Steiner Verlag), 1994, 45 pp.”, Acta
Orientalia LVI, pp. 302–311.
Potter, Karl H. (1999). Encyclopedia of Indian Philosophies. Volume VIII: Buddhist
Philosophy from 100 to 350 ad. Delhi: Motilal Banarsidass.
Przyluski, Jean (1932). “L’influence iranienne en Grèce et dans l’Inde”, Revue de l’Univer-
sité de Bruxelles XXXVII, pp. 283–294.
Pym, Anthony (2010). Exploring Translation Theories. London and New York NY:
Routledge.
Rhys Davids, Thomas W. and J. Estlin Carpenter (eds.) (1968). The Sumaṅgala-Vilāsinī,
Buddhaghosa’s Commentary on the Dīgha Nikāya. 3 volumes. Pali Text Society.
London: Luzac & Co.
Rhys Davids, Thomas W. and Carpenter, Joseph E. (ed.) ([1890–1911] 1995–2007).
Dīghanikāya. 3 vols. London: The Pāli Text Society.
Salomon, Richard (1999). Ancient Buddhist Scrolls from Gandhāra. The British Library
Kharoṣṭhī Fragments. London: The British Library.
Scharfstein, Ben-Ami (1989). The Dilemma of Context. New York NY and London: New
York University Press.
Seiwert, Hubert (1994). “Orthodoxie, Orthographie und Zivilreligion im vorneuzeitli-
chen China”, In Gnosisforschung und Religionsgeschichte. Festschrift für Kurt Rudolph
zum 65. Geburtstag. Edited by Holger Preißler, Hubert Seiwert and Heinz Mürmel.
Marburg: Diagonal-Verlag, pp. 529–541.
Shiji 史記, Ershiwu shi 二十五史 (1988). Shanghai: Shanghai guji chubanshe.
Simson, Georg von (1965). Zur Diktion einiger Lehrtexte des buddhistischen Sanskritkanons.
München: J. Kitzinger.
Staal, Frits (1989). “The Independence of Rationality from Literacy”, European Journal
of Sociology 30, pp. 301–310.
Staal, Frits (1993). Concepts of Science in Europe and Asia. Leiden: International Institute
for Asian Studies.
Introduction 25
Stache-Rosen, Valentina (1968). Dogmatische Begriffsreihen im älteren Buddhismus. Das

Saṅgītisūtra und sein Kommentar Saṅgītiparyāya. 2 vols. Berlin: Akademie-Verlag.
Takakusu, Junjirō 高楠順次郎, Watanabe, Kaigyoku 渡邊海旭, and Ono Gemmyō 小野
玄妙 (ed.) (1924–1934). Taishō shinshū daizōkyō 大正新脩大蔵經, Tokyo: Taishō
issaikyō kankōkai.
- T.1.1: Dīrghāgama, Chang ahan jing 長阿含經, Buddhayaśas and Zhu Fonian 竺
佛念.
- T.1.12: Saṃgītisūtra, Da ji famen jing 大集法門經, Shihu 施護.
- T.22.1425: Mahāsāṃghikavinaya, Mohesengqi lü 摩訶僧祇律, Buddhabhadra,
Faxian 法顯.
- T.22.1428: Dharmagupta[ka]vinaya 四分律, Si fen lü, Buddhayaśas, Zhu Fonian 竺
佛念, et. al.
- T.23.1435: Sarvāstivādavinaya, Shi song lü 十誦律, Puṇyatara, Dharmaruci,
Kumārajīva.
- T.24.1451: [Mūlasarvāstivāda]Vinayakṣudrakavastu, Genben shuo yiqieyou bu pi
naiye zashi 根本說一切有部毘奈耶雜事, Yijing 義淨.
- T.24.1463: Vinayamātṛkā, Pinimu jing 毘尼母經.
- T.26.1536: Śāriputra, [Abhidharma]saṃgītiparyāya[pādaśāstra], Apidamo ji yi
men zu lun 阿毘達磨集異門足論, Xuanzang 玄奘.
- T.27.1545: 500 arhats, [Abhidharma]mahāvibhāṣā[śāstra] Apidamo da piposha
lun 阿毘達磨大毘婆沙論, Xuanzang 玄奘.
- T.41.1821: Puguang 普光, Jushe lun ji 俱舍論記.
- T.41.1822: Fabao 法寶, Jushe lun shu 俱舍論疏.
- T.43.1830: Kuiji 窺基, Cheng weishi lun shuji 成唯識論述記.
- T.45.1852: Jizang 吉藏, San lun xuan yi 三論玄義.
- T.49.2030: [Ārya]Nandimitrāvadāna, Da aluohan nantimiduolo suo shuo fa zhu ji
大阿羅漢難提蜜多羅所說法住記, Xuanzang 玄奘.
- T.49.2031: Vasumitra, Samayabhedoparacanacakra, Yibuzong lun lun 異部宗輪論,
Xuanzang 玄奘.
- T.50.2042: Aśokarājāvadāna, Ayu wang zhuan 阿育王傳, An Faqin 安法欽.
- T.50.2043: Aśokarājasūtra, Ayu wang zhuan 阿育王傳, Saṃghabhara.
- T.50.2053: Da Tang Daci’en si sanzang fashi zhuan 大唐大慈恩寺三藏法師傳, Huili
慧立, Yancong 彥悰.
- T.51.2087: Da Tang xiyu ji 大唐西域記, Bianji 辯機, Xuanzang 玄奘.
- T.55.2145: Sengyou 僧祐, Chu sanzang ji ji 出三藏記集.
Takakusu, Junjirō and Nagai Makoto (eds.) (1975). Samantapāsādikā. Buddhaghosa’s
Commentary on the Vinaya Piṭaka. 3 volumes. Pali Text Society. London: Routledge
& Kegan Paul.
Trenckner, Vilhelm and Chalmers, Robert (ed.) (1888–1902). Majjhimanikāya. London:
The Pāli Text Society.
26 Introduction
Watanabe, Fumimaro (1983). Philosophy and its Development in the Nikāyas and
Abhidharmma. Delhi: Motilal Banarsidass.
Willemen, Charles (2006). The Essence of Scholasticism: Abhidharmahṛdaya. T1550.
Delhi: Motilal Banarsidass.
Willemen, Charles; Dessein, Bart & Cox, Collett (1998). Sarvāstivāda Buddhist
Scholasticism. (Handbook of Oriental Studies 2/11). Leiden etc.: Brill.
Zürcher, Erik ([1959] 1972). The Buddhist Conquest of China. The Spread and Adaptation
of Buddhism in Early Medieval China. 2 vols. Leiden: E.J. Brill.
27
Part 1
Mātṛkā and Abhidharma Terminologies
∵
28
Abhidharma and Indian thinking 29
Chapter 1
Abhidharma and Indian thinking

Johannes Bronkhorst
1 Introduction
Much could be said, and has been said, about the influence of Buddhism on
other currents of thought in India. This chapter will deal with this same topic,
but limit its attention to one specific form of Buddhism: Abhidharma. And
even here, it will only take into consideration the form of Abhidharma that was
created in the northwestern corner of the Indian subcontinent, in an area that
nowadays belongs to three different political entities: Pakistan, Afghanistan
and Kashmir. This form of Abhidharma is primarily, but not exclusively, associ-
ated with the Sarvāstivādins, and the region in which it arose is often referred
to as “Greater Gandhāra,” a somewhat vague term, to be sure, but appropriately
so, for it is hard to determine with certainty which regions can legitimately be
included in it, and which cannot.1 Kashmir and Gandhāra may be taken to
belong to it, and if we add Bactria, we can call it, with David Gordon White
(2012), KGB.
The Abhidharma of Greater Gandhāra (I will call it Gandhāran Abhidharma)
exerted a major influence on Buddhist thought in India, to be sure. Most, if not
all, of the philosophical schools of Indian Buddhism – both Main Stream and
Mahāyāna – are based on the foundations laid here. I have argued elsewhere
that already the very earliest Mahāyāna texts we possess are based on
Abhidharma thought; these texts may date from the first century BCE.2
* This paper brings together a number of observations made and conclusions drawn in other
publications, duly referred to in the footnotes, along with new observations. In its present
form it has profited from various critical remarks, most notably by Collett Cox and Shoryū
Katsura. Bronkhorst, 1996 overlaps to some extent with parts of this paper, but concentrates
on Buddhist notions of language and their influence on Brahmanical philosophy.
1 A Buddhist presence in this region from at least the 2nd century BCE seems certain.
Cp. Behrendt, 2004: 256: “[Phase I] began with the founding of the earliest Buddhist centers
in Greater Gandhāra: Butkara I and the Dharmarājikā stūpa in Taxila […]. An early 2nd century
bce date seems a conservative benchmark for the beginning of this period.” On the extent of
Greater Gandhāra, see also note 6, below.
2 Bronkhorst, 2013.

30 Bronkhorst
In this chapter, I will concentrate on Abhidharma influence outside

Buddhism. I will begin with a short sketch of some important features of
Abhidharma thought, then turn to a number of Indian schools of thought that
took over one or more of those features.
Gandhāran Abhidharma, it should be emphasized right from the beginning,
was different from other forms of Abhidharma, most noticeably so from the
Abhidharma preserved in Pāli by the Theravāda Buddhists of Sri Lanka and
some countries of Southeast Asia. It is not my intention to give an exhaustive
characterization of the two Abhidharmas and their differences, even if I had
the competence to do so. I will rather concentrate on Gandhāran Abhidharma,
emphasizing those of its features that will play a role in the remainder of this
chapter.
Gandhāran Abhidharma, like Theravāda Abhidharma, is based on old lists
of items, and therefore on old material. Gandhāran Abhidharma dealt with
this old material in its own manner, and in doing so went far beyond the old
heritage. Already in our oldest documents, it presents itself as the result of an
attempt to rethink this traditional material, to systematize and modify it so as
to make it part of a coherent philosophy. The emphasis in this philosophy is on
ontology: the question “what exists?” is central to this form of Abhidharma,
and inspired a number of highly original, and surprising, answers.3
What does exist? Put very briefly, the answer is: the dharmas. Dharmas are
the items that had been enumerated in lists, apparently with the purpose of
preserving the essential elements of the Buddha’s teaching. Unsurprisingly,
given the nature of the Buddha’s teaching, many dharmas were of a psycho-
logical nature; they referred to mental states. But not all of them did so. It is not
always easy to figure out why certain dharmas were included in the lists, and
presumably the early Gandhāran scholiasts faced the same problem. They
offered an easy yet radical solution: the dharmas are the ultimate elements of
existence. Everything that exists is made up of dharmas. This applies to human
beings, of course. After all, we have numerous psychological constituents. But
the same vision also applies to material objects; fortunately there were a num-
ber of dharmas in the traditional lists that are of a material nature and can
therefore account for the existence of material objects, too.
3 It is true that this question is rarely, if ever, explicitly formulated in the technical texts of early
Abhidharma. However, this aspect is emphasized, presumably at least from the time of King
Menander onward, in texts that draw the consequences of this form of Abhidharma, includ-
ing the Milindapañha and Prajñāpāramitā texts right from the beginning. See Bronkhorst
2013.
This, then, was the first step taken by the Gandhāran Ābhidharmikas. They
introduced an atomic vision of reality, in the sense that they maintained that
there are ultimate constituent elements of all that exists, and that these ulti-
mate elements are the dharmas which they had collected in lists. But they did
not stop there. The next step was the claim that dharmas are the only things
that exist. The objects that they constitute – this covers all there is, including
the macroscopic objects of our everyday experience – do not exist. And the list
of dharmas is an exhaustive enumeration of all that exists.
This second step can easily be understood in the light of the persistent
Buddhist tradition that the self, or the person, does not exist. If the person is
conceived of as the aggregate of all its constituent dharmas (and this became
the generally accepted way in Buddhism to think about it), this can easily be
interpreted to mean that the person does not exist because it is an accumula-
tion of dharmas. Other accumulations of dharmas therefore do not exist either.
The Gandhāran Buddhist scholiasts found themselves in this way in the pos-
session of an exhaustive enumeration of all there is, viz., the inherited and
slightly adjusted list of dharmas. These dharmas were now thought of as
momentary, i.e. as each lasting no more than one single moment. Once again,
a traditional Buddhist doctrine could be invoked in defense of this new view:
the Buddha had taught that all conditioned factors are non-eternal. This could
easily be interpreted to mean that all dharmas are momentary, and this is what
happened.4
Gandhāran Abhidharma reduced in this manner the whole world, both ani-
mate and inanimate, to an uninterrupted sequence of some seventy-five
momentary dharmas; more precisely, some seventy-two momentary dharmas,
plus three eternal ones, the so-called asaṃskṛta ‘non-conditioned’ dharmas.
The world of our experience thus turned out to be ultimately unreal, with the
real world of momentary dharmas hidden below it.
How could we possibly be misled into believing that we live in a world of
persons and other macroscopic objects? The Buddhist texts frequently respond
by pointing out that this or that macroscopic object is a mere name and does
not really exist. This is what, in the Milindapañha, Nāgasena pointed out with
regard to King Milinda’s chariot, and numerous other Buddhist texts state the
same. The objects we are familiar with in our everyday world owe their relative
existence – or rather our mistaken conviction that they exist – to the words of
language.
There is one more point I wish to add to this brief characterization of
Gandhāran Abhidharma. Words, as we have seen, are responsible for our
4 On the beginning of Buddhist momentariness, see Rospatt, 1995, along with Bronkhorst, 1995.
32 Bronkhorst
mistaken belief in the reality of a world of persons, chariots and much else. But
words, one might object, are themselves ultimately non-existing entities. Does
this not undermine the system in some vital manner?
It is possible that the Gandhāran scholiasts were aware of this objection. All
we know for certain is that they introduced some dharmas in their list that we
might call linguistic dharmas. There are three of them, originally perhaps only
two. They correspond to individual speech sounds, words and sentences. If we
concentrate on words, this means that, beside the sequence of speech sounds
that make up a word, there is a momentary dharma (or perhaps better: a series
of identical momentary dharmas) that are the word. The word exists in this
way beside, and independently of, the speech sounds. This is a highly remark-
able conception, and the only justification for its existence I can think of is to
save the reality, the real existence, of words and with it the fundamental coher-
ence of the Abhidharma system that was being developed.
2 Grammar
It is time to leave the details of the Abhidharma system and turn to the schools
of non-Buddhist thought on which it exerted an influence. The first to be con-
sidered is the tradition of Pāṇinian Sanskrit grammar, and more in particular
one specific text that belongs to it, Patañjali’s Mahābhāṣya “Great Commentary”.
It is known that Patañjali wrote this work during the decades following the
middle of the second century BCE, and that he was at some time in the service
of King Puṣyamitra, one of the successor kings of the Mauryan empire. He was
already alive during the inroads of Indo-Greeks into northern India, around
150 BCE. There is some evidence to believe that he settled in Kashmir, and this
may explain his interest for us at present.5 For Kashmir was on the one hand a
part of Greater Gandhāra,6 with a substantial Buddhist presence. On the other
hand, Kashmir was conquered by Puṣyamitra, a Brahmanical ruler who, as we
saw, had scholars like Patañjali in his service. Kashmir may therefore have been
the place par excellence for Buddhist philosophy, i.e. Gandhāran Abhidharma,
and Brahmanical scholarship to meet. Judging by certain features of Patañjali’s
Mahābhāṣya, this is what actually happened. Let me explain.
Gandhāran Abhidharma was preoccupied with ontology; this we have seen.
There are reasons to think that the Buddhists who created it were the first to
5 Aklujkar, 2008; Bronkhorst, 2016: 43–46 and § III.3.2.

6 The map given by Salomon, 1999: 2 suggests that he includes Kashmir in “Greater Gandhāra”;
Behrendt, 2004: 16, 22 does so explicitly.
introduce this preoccupation. It is absent from all contemporary Indian

sources, which includes grammatical treatises. Pāṇini does not show the slight-
est interest in the question what exists. This changes with Patañjali. His
Mahābhāṣya discusses the ontological status of words and speech sounds, and
proclaims that both have independent existence and are eternal. This is highly
remarkable. To begin with, there is no discernible reason why Patañjali should
be interested in the ontological status of sounds and words. As a matter of fact,
his claim that words exist eternally obliges him to reinterpret Pāṇinian deriva-
tions of words. No longer do words come about by adding suffixes to stems, as
had been the case in Pāṇini. No, since words are eternally existing independent
entities, there can for Patañjali be no question of parts of words, whether they
be stems or suffixes. He therefore has to justify grammatical derivations differ-
ently. The details do not concern us at present. The main thing is that Patañjali
had come to look upon words as ontologically existing entities, and that this
forced him to interpret grammatical derivations differently.7
But why had Patañjali come to look upon words as ontologically existing
entities? And why do his ontological interests stop at words? The obvious
answer is that Gandhāran Abhidharma exerted an influence on him. This does
not only explain his ontological preoccupations, but also the fact that he took
the audacious step of postulating that words are not just a sequence of sounds,
but independently existing things. This is exactly what Gandhāran Abhidharma
had done. There the preoccupation with ontology was not surprising, for its
philosophy was deeply interested in ontology. In the case of Patañjali, on the
other hand, ontology played no useful role in his grammatical reflections; it
even made them more complicated.
A closer look at grammatical derivations as conceived of by Patañjali reveals
that Abhidharma influence had gone further, and caused him greater and
more subtle difficulties than the ones just described. Patañjali, it turns out,
imposed upon grammatical derivations a scheme that is completely parallel to
the vision the Buddhist scholiasts had imposed on reality in general.
What is the Abhidharma vision of reality in general? Basically it is a linear
vision. All the things we know are successions of momentary dharmas. Each
dharma is replaced, after a moment, by another dharma. This next dharma is
usually similar to the preceding one. After all, a cow does not become a horse
from one moment to the next; this continuity is explained by the regularity
with which dharmas are succeeded by identical or similar dharmas. Moreover,
the nature of each next dharma is largely determined by the immediately pre-
7 For details, see Bronkhorst, 1987: 46 ff.

34 Bronkhorst
ceding dharma, not by dharmas that occurred earlier and that have gone long
since, nor indeed by dharmas that are still to appear.8
A grammatical derivation, too, is a succession of stages. What determines
each next stage? For Pāṇini this could be any of the preceding stages, or occa-
sionally even a stage that was still to come. Patañjali changed all this. For him,
each next stage was completely determined by the immediately preceding
one. He imposes in this way a linear scheme that is in all essential respects
parallel to the linearity of Gandharan Buddhism. This imposition confronted
him with major difficulties, for Pāṇini had never intended anything of the kind.
Patañjali was therefore obliged to introduce new procedures, add metarules
and use various other tricks to make derivations conform to his vision.
But this vision was extraneous to Pāṇini’s grammar. Patañjali imposed it,
without telling us why. However, it seems safe to assume that the explanation
of this strange change imposed by Patañjali lies in his acquaintance with the
Abhidharma vision of the world.9
3 Jainism
One of the old texts of the Śvetāmbara Jaina canon, the Sūyagaḍa (in Sanskrit:
Sūtrakṛtāṅga), shows awareness of a number of characteristics of Buddhism. It
mentions the five skandhas, but also – and this is more interesting in the pres-
ent context – the notion of momentariness. Momentariness, as we have seen,
is one of the innovations introduced by Gandhāran Abhidharma. We may
assume that the author of this text (or of this portion of the text) was acquainted
with the developments that had taken place in north-western Buddhism.
Other texts of the Śvetāmbara canon – all of them no doubt younger than
the Sūyagaḍa – do not only know the notion of momentariness: they have
adopted it themselves. The moment (samaya) as the smallest unit of time
appears to occur for the first time in the Uttarajjhayaṇa, and this same text
further knows the notion of santati, the sequence of moments that is also com-
mon in Buddhism. Beside moments, other significant notions are found in the
Uttarajjhayaṇa, among them pradeśa (the smallest unit of space) and
paramāṇu (atom). In other words, we find here an atomic vision of time, space
and matter. It is true that Buddhism does not appear to have accepted the
8 The justification for this way of viewing the succession of dharmas lies in the Buddhist doc-
trine of dependent origination (pratītyasamutpāda).
9 For details, see Bronkhorst, 2004. See further Bronkhorst, 1994; 2002b.
atomic nature of space, it did accept and elaborate the atomic nature of mat-
ter, and of course of time.
The Sūyagaḍa is also acquainted with the Buddhist notion of the person as
a collection of skandhas. It tells us that the Buddhist person is neither different
nor not different from the skandhas; this is a position that was held by the
Buddhist Pudgalavādins. The Sūyagaḍa, we learn from this passage, was
already aware of the issue regarding wholes and their parts that occupied the
Buddhists. Most Buddhists, with the exception of the Pudgalavādins, rejected
the existence of wholes; subsequent texts from the Śvetāmbara canon accept
it.
Perhaps the most surprising feature that Jainism took over from Abhidharma
Buddhism, if my reflections are correct, is the use of the word pudgala. In
Buddhism the pudgala is the person, conceived of as the combination of the
skandhas, i.e. of the dharmas, that make up a person. Many Buddhists thought
that this conception does not correspond to any reality; the Pudgalavādins
were of a different opinion, and specified that the pudgala is neither different
nor not different from the skandhas. All this we know.
Interestingly, Jainism too adopted this word pudgala. For them, it does not
refer to the person, but rather to material objects. A closer study of the relevant
passages brings to light that one of the earliest occurrences of this word in the
Jaina canon does refer to a person. In subsequent Jaina developments the
emphasis is more and more on the aggregate. We know that in Buddhism pud-
gala refers to an aggregate, but only to one special kind of aggregate: the
person. In Jainism it comes to include other aggregates as well: early texts use
it in the sense of “portion, quantity”. It appears indeed that within Jainism the
meaning of this word developed from “person” (the meaning also used in
Buddhism) to “material object”. This development only makes sense if we start
from the Buddhist notion of the person as an aggregate.
It is not possible to enter more deeply into this discussion. Further reflec-
tions and textual references can be found in another publication, to which I
must refer for details.10 Here a few words must be said about the time and
place of the interaction between Gandhāran Abhidharma and Jainism. Jainism
was not present in Gandhāra.11 Gandhāra is beyond the lands where a Jaina
monk is allowed to travel, which extend westward until Thaneshwar.12 However,
10 Bronkhorst, 2000.
11 See however Pal, 2007, which shows that Jaina merchants may have ventured into
Gandhāra and Afghanistan.
12 See Jain, 1984: 23–24 (with note 2), 337 ff.; further PPN II s.v. Saṃpai (Samprati), with refer-
ences. Kalpasūtra 1.50 reads: “monks and nuns may wander eastward as far as
36 Bronkhorst
Jainism was strongly present in Mathurā under the Kuṣāṇas, i.e. at a time when
Buddhism, too, had a strong presence in that region. There are also reasons to
think that Buddhism and Jainism interacted in other respects in this region
during this period. It seems, for example, that Jainism at that time abandoned
the worship of relics of the Jina and of stūpas that contained them. This would
not be an example of Jainism borrowing from Buddhism but rather of Jainism
consciously differentiating itself from Buddhism. Once again, there is no space
to pursue this issue further at present.13
4 Brahmanical Philosophy
Let us now turn to Brahmanical philosophy. The influence of Gandhāran

Abhidharma is here most obvious in the case of the Vaiśeṣika school of thought.
Vaiśeṣika philosophy imposes an ontological scheme on the world, and it can
easily be seen that this scheme has been inspired by Gandhāran Abhidharma.
Remember the important features of Gandhāran Abhidharma I enumer-
ated earlier, and consider the following three:
1. Gandharan Abhidharma interpreted its lists of dharmas as exhaustive
enumerations of all there is.
2. These dharmas, these elements of existence, can form aggregates, but
these aggregates do not themselves exist; no wholes exist, only their ultimate
parts.
3. These aggregates yet play roles in our daily lives and experience, but ulti-
mately they owe their relative existence to words. There are no chariots beyond
their ultimate parts, yet we believe there are on account of the word “chariot”.
These features reappear in Vaiśeṣika, though the details are different.14
1. Like Abhidharma, Vaiśeṣika has an exhaustive enumeration of all there is.
These are its well-known categories (of which there are six, or seven, or ten,
depending on the sub-school concerned), and a large number of well specified
subcategories.
Anga-Magadha, southward as far as Kosambī, westward as far as Thūṇā and northward as

far as Kuṇālā. They may wander thus far, (for) thus far there are Āryan countries, but not
beyond unless the Dhamma flourishes there” (tr. Bollée, 1998: xxiv). According to PPN I s.v.
Thūṇā (Sthūṇā), this place is Thaneshwar, north-west of Mathurā.
13 See Bronkhorst, 2016: Appendix VIII.
14 For a different opinion, see Lysenko, 2011.
2. Vaiśeṣika, too, recognizes the importance of aggregates; but contrary to

Gandhāran Abhidharma, it considers these aggregates as really existing.
Chariots therefore exist, as do their parts.
3. In Vaiśeṣika, as in Gandhāran Abhidharma, words correspond to aggre-
gates. But where these words ultimately did not correspond to anything real in
Abhidharma, words correspond to reality in Vaiśeṣika. This is true to the extent
that the existence of certain entities is concluded from the fact that there is a
word for them.
Once again, it is not possible to enter into details.15 It seems however clear
that the Vaiśeṣika philosophy is indebted to Gandhāran Abhidharma, not just
in some details, but in its very structure. It is further appropriate to recall that
Vaiśeṣika ontology was taken over by the Nyāya school of thought, virtually
wholesale. It was taken over in part by numerous other thinkers, and we can
safely state that Vaiśeṣika ontology is the most important Brahmanical ontol-
ogy of classical India. The fact that it was more or less a mirror image of
Gandhāran Abhidharma ontology shows the historical importance of the
latter.
5 Influence Elsewhere
It cannot be the purpose of this chapter to give a complete survey of the ways
in which Gandhāran Abhidharma has influenced non-Buddhist traditions in
India. Some of its notions, such as that of momentariness and that of the per-
son as an aggregate, appear unexpectedly in texts such as the medical Caraka
Saṃhitā.16 Instances could no doubt be multiplied.
Here I will mention one more feature of Gandhāran Abhidharma that was
enthusiastically taken over by certain Brahmanical thinkers. Recall that in this
philosophy the world of our experience is ultimately unreal. This became a
central element of Buddhist systematic philosophy, one that distinguished it
for a long time from the Brahmanical systematic philosophies with which it
coexisted.17 This changed around the middle of the first millennium CE, when
philosophies like Advaita Vedānta joined the inter-philosophical debate. One
of its most important early thinkers was Śaṅkara, and it is not surprising that
some of his contemporaries accused him of being a crypto-Buddhist:18
15 See Bronkhorst, 1992.

16 See Bronkhorst, 2002a.
18 Bronkhorst, 2009: 187, with note 440.
38 Bronkhorst
Śaṅkara’s philosophy had adopted the essentially Buddhist notion that the
world of our experience is not ultimately real. There is no need to recall that
Advaita Vedānta became in due time India’s most popular philosophy, and
Śaṅkara its most famous thinker.
There is one more case of Abhidharma influence that deserves to be men-
tioned. However, for reasons that will become clear, it makes most sense to
deal with this case after a brief discussion of the origin of Gandhāran
Abhidharma. That is therefore the topic to which I will turn now.
6 Whence Gandhāran Abhidharma?
It should be clear from what has been mentioned so far that Gandhāran
Abhidharma has been extremely influential in the subsequent development of
Indian thought. Not only did it to a large extent determine the shape of subse-
quent Buddhist philosophy in all of its forms. Non-Buddhist philosophies and
other forms of Indian thought, too, were profoundly influenced by it. I think
therefore that we can state without hesitation that Indian philosophy – or at
any rate systematic, rational philosophy in South Asia – began in Greater
Gandhāra.
How did this happen? Hard-headed philologists rarely ask this question.
They often feel that it is difficult enough to extract from the texts what their
authors thought, and that those same texts rarely, if ever, tell us why these
authors thought the way they did. There are some exceptions. T.R.V. Murti –
whose book The Central Philosophy of Buddhism (1960) was once widely
admired, but whose influence seems to be beyond its peak – emphasized the
importance of intellectual strife and the resolution of contradictions in the
development of new ideas. According to him, “the Ābhidharmika schools […]
grew as the rejection of the ātmavāda of the Brāhmanical systems”.19 This can-
not be right. To the best of our knowledge there were no, or few, Brahmins in
Greater Gandhāra at that time.20 Gandhāran Abhidharma cannot therefore
have been based on a rejection of the Brahmanical ātmavāda.
I have elsewhere dealt with Murti’s ideas in general, which I do not accept.21
However, it does seem right to assume that revolutionary new ideas tend to
arise in appropriate surroundings, especially in challenging intellectual sur-
roundings. It does therefore make sense to ask, if there were no Brahmins in
19 Murti, 1960: 8.
20 Bronkhorst, 2011: 202–205; 2016: § I.1.3.
Greater Gandhāra at that time, who else may have been there to challenge the
Buddhists into creating a vision of the world that was to inspire Indian think-
ers, directly or indirectly, for the next two thousand years.
The answer to this question seems simple and straightforward. Gandhāra
and most of what we call Greater Gandhāra (with the exception of Kashmir,
see above) fell again in the hands of the Indo-Greeks around 185 BCE, at the
time of the collapse of the Mauryan empire. At that time, and perhaps already
before that time, the Buddhists of that part of the subcontinent had to deal
with Hellenistic rulers. And we know that Hellenistic rulers had a tendency to
surround themselves with philosophically cultivated sages, and that discus-
sions of a philosophical nature had become part of Hellenistic tradition. It is
hard to imagine that representatives of the clearly numerous Buddhists who
inhabited those regions were not sometimes challenged to take part in such
discussions. However, to take part in a sophisticated discussion, you better
present, and represent, a coherent position yourself. Presumably the Buddhist
scholiasts of Gandhāra realized this, and the result is that they created, for the
first time in Buddhist history, a coherent philosophical ontology: Gandhāran
Abhidharma was born.
I realize that these few words about the possible origin of Gandhāran
Abhidharma do not but scratch the surface of an important and complex his-
torical question. I have tried to do it more justice in some other publications,22
and I remain aware that the question has not been fully explored even there.
Others may take it up and, who knows, they may come to different conclusions.
However, I do wish to emphasize that, if it is true that Indian systematic philos-
ophy began in the north-western region we call Greater Gandhāra, reflections
as to how this happened, and why, cannot be ignored.
After these summary reflections about the origin of Gandhāran Abhidharma,
we are ready to address a theory recently launched by Christopher Beckwith in
his book Warriors of the Cloisters: The Central Asian Origins of Science in the
Medieval World (2012). Beckwith’s theory concerns what he calls the recursive
argument method that, he claims, came to be used in Indian philosophical
texts, but not only there. The recursive argument method, according to
Beckwith, ended up being used in medieval Europe which borrowed it from
the Islamic world. The Islamic world itself, still according to Beckwith, took it
from Buddhism in Central Asia. This recursive argument method supposedly
played a crucial role in the development of European science. This is why the
22 Bronkhorst, 1999; 2001.

40 Bronkhorst
subtitle of Beckwith’s book is The Central Asian Origins of Science in the

Medieval World.23
Where did this recursive argument method ultimately come from? Beckwith
writes the following about it (p. 56):
The earliest text so far identified that uses a primitive version of the
method, and indeed uses it throughout the text, is the Central Asian
Aṣṭagrantha [or Aṣṭaskandha; Taishō 1543]. In this work, each topic argu-
ment is followed by a list of arguments about it – usually a rather long list
– and then they are repeated and disputed, one by one, in order.24 By
contrast, the later Jñānaprasthāna [Taishō 1544] does not use the recur-
sive argument method at all. It strictly follows the two-part Question:
Answer format.
The Vibhāṣā, a scholastic work of the Bactrian-Gandhāran branch of
the Sarvāstivāda school dated possibly to the first century AD, during the
Kushan Empire, contains the earliest known example of what eventually
became the fully developed recursive argument method. The method
apparently thus developed specifically within the Bactrian-Gandhāran
branch of the Sarvāstivāda school, and was only later partially adopted by
the Kashmiri Vaibhāṣika sect of Sarvāstivāda.
Examples of the recursive argument do not occur in earlier Buddhist
texts, earlier non-Buddhist Indian texts, or earlier texts connected to
other branches of the Sarvāstivāda school.
23 Beckwith’s theory does not only concern the recursive argument method, but also the
development of medieval European colleges under the influence of Islamic madrasas,
which themselves presumably arose under the influence of Buddhist vihāras. Beckwith
claims, with a reference to Dutt, 1962: 62ff. and 211 ff., that “the plan of the vihāra is strik-
ingly different from that of the saṅghārāma, the typical earlier, strictly Indian, Buddhist
monastic design” and he adds that “[t]he vihāra design is […] a specifically Central Asian
innovation developed under the Kushans and spread by them” (p. 41). These claims, and
especially the second one, are not substantiated but cannot here be further examined; see
however Schopen, 2004: 73–80; 2006.
24 Beckwith refers here to Willemen, Dessein & Cox, 1998: 223, where we read: “the two Chi-
nese translations of the *Aṣṭaskandhaśāstra (30 fascicles) and Jñānaprasthāna (20 fasci-
cles) do differ in length, at least in part, as a result of a difference in format: the
*Aṣṭaskandhaśāstra lists the questions that will be addressed at the beginning of each
section and then repeats the questions with each answer; the Jñānaprasthāna gives the
questions only once prior to the answer.”
This passage raises a number of questions, among them the following: Does
Beckwith not overemphasize the role of Central Asia? Is the Aṣṭagrantha (or
Aṣṭaskandha) really a Central Asian text?25 And is the Vibhāṣā really a scholas-
tic work of the Bactrian-Gandhāran branch of the Sarvāstivāda school?26 At
present we can leave these questions aside, and reformulate Beckwith’s theory
slightly, so that it now states that the recursive argument method originated in
Abhidharma Buddhism. Put this way, Beckwith’s claim is that the origins of
science in the Medieval European world have to be looked for in Gandhāran
Abhidharma.
What is this recursive argument method? Beckwith dedicates a chapter
(Chapter Two) of his book to explaining and illustrating what he means by this.
Here I will merely repeat his statement of its essence, which occurs on page 89
and reads:27
I. Argument (the Main Argument, Question, or Topic)

II. Subarguments1 about the Argument
III. Subarguments2 about the Subarguments1 about the Argument
Suppose now, for argument’s sake, that Beckwith is right in maintaining that
examples of the recursive argument do not occur in earlier texts, whether
25 Beckwith refers to Willemen, but all Willemen says is, 2006: 6: “Kātyāyana’s Aṣṭagrantha
[…], the main text of the Gandhārans, [was] probably written in Gāndhārī and Kharoṣṭhī,
in the late first century BCE. In the second century CE this text was rewritten in Sanskrit,
and called Jñānaprasthāna.” (my emphasis, JB). See also Willemen, 2012: 163–164: “The
Sarvāstivāda ‘orthodoxy’ rewrote the old Gandharan Aṣta-grantha in Sanskrit, now called
Jñāna-prasthāna. Because the old text had many Vibhāṣās, commentaries, the new text
needed a new commentary. This is the Mahāvibhāṣā.”
26 On p. 59, Beckwith specifies that “[t]he earliest vibhāṣā preserved (in Chinese translation)
[…] is the Central Asian work known as the Vibhāṣā [Taishō 1547]”. And p. 62: “the Bac-
trian Aṣṭagrantha and Vibhāṣā are the models for the later Jñānaprasthāna and
Mahāvibhāṣā.” Contrast this with Cox’s remark: “It is with the composition of the vibhāṣā
compendia that the Sarvāstivāda school within Kaśmīra comes to be defined both doc-
trinally and textually.” (Willemen, Dessein & Cox, 1998: 229). See also Willemen, 2006:
6–7: “The western [i.e. non-Kashmirian, JB] Sarvāstivādins, a very heterogeneous group,
seem to have had more than one Vibhāṣā on the Aṣṭagrantha.” Beckwith’s concern with
Central Asia finds expression in his earlier book Empires of the Silk Road: A History of
Central Eurasia from the Bronze Age to the Present (2009).
27 See also pp. 25–26: “The recursive argument is, minimally, an argument that is disputed
by an argument that is disputed by an argument, or more simply (but in reverse order), an
argument about an argument about an argument.”
42 Bronkhorst
Buddhist or non-Buddhist, whether Asian or European. It will yet be hard to

believe that this way of arguing was unknown before the Aṣṭagrantha and the
Vibhāṣā. Is this stylistic expression not simply a literary reflection of oral dispu-
tations in which the different participants are given the occasion to present
their arguments in full and to refute, point by point, those of their opponents?
After all, the recursive argument “is at heart a way to examine a problem sys-
tematically, logically, and in great detail” (Beckwith, 2012: 25). And do such oral
disputations – systematic, logical, and detailed – not constitute the back-
ground against which Gandhāran Abhidharma arose and could arise? If so, the
literary feature to which Beckwith draws attention is no more and no less than
the reflection of the real life situation that allowed Gandhāran Abhidharma –
and not just the Aṣṭagrantha and the Vibhāṣā – to arise.28
Beckwith may not be right in thinking that recursive arguments do not
appear in texts older than the Aṣṭagrantha and the Vibhāṣā. The Mahābhāṣya
of Patañjali is such a text, and it appears to use the recursive argument method.
Here is what specialists say about it. According to George Cardona (1976: 253),
the Mahābhāṣya is composed “in the form of dialogues in which take part a
student (śiṣya) who questions the purpose […] of rules and their formulations,
an unaccomplished teacher (ācāryadeśīya) who suggests solutions which are
not fully acceptable, and a teacher (ācārya) who states what is the finally
acceptable view.” Hartmut Scharfe (1977: 156) analyzes a passage from the
Mahābhāṣya and concludes: “With great stylistic art Patañjali has created the
impression of a freely progressing debate with new disputants butting in now
and then in which all possibilities of an interpretation are scrutinized”. This,
one might think, is precisely what Beckwith finds in the Aṣṭagrantha and so
many other texts.
28 Beckwith emphasizes the difference between structure and content, but believes that the
former can influence the latter (2012: 35–36): “The overt, explicit, formal structure of the
recursive argument is its most crucial factor. It is not quite true that ‘the medium is the
message’ in recursive method books, but because they typically consist exclusively of lists
of recursive arguments, each of which contains many contrasting views on the same
problem, they clearly did encourage scepticism and speculation by the authors. In that
respect, therefore, it is true that the form of the recursive argument did have a significant
indirect impact on the content of works written according to it. Nevertheless, it must be
stressed that the specific overt structure, per se, of a recursive argument is not directly or
even implicitly connected, structurally or semantically, to its specific overt content or to
the implicit logical structure of the internal content. In other words, in a recursive argu-
ment method, the way it is said has essentially nothing to do with what is said. It does,
however, have a great deal to do with the general way the content is approached and
understood […].”
Recall now that the Mahābhāṣya underwent the influence of Abhidharma

Buddhism. Since it is older than the Aṣṭagrantha and the Vibhāṣā, it cannot
have undergone the direct influence of these texts. However, it may have
undergone the influence of earlier Buddhist Abhidharma texts that displayed
this style, or perhaps not the influence of any particular text but rather the
influence of the specific way of discussing that was the background of
Gandhāran Abhidharma. In this way the Mahābhāṣya may add an interesting
dimension to Beckwith’s theory.
We have come to the end of this chapter. We have briefly touched upon
many details related to the history of Gandhāran Abhidharma. Each of these
may merit further discussion and critical assessment. But there is one conclu-
sion that in my opinion cannot be seriously doubted: the historical importance
of Gandhāran Abhidharma is beyond dispute. Anyone interested in the intel-
lectual history of South Asia – and perhaps in the intellectual history of Eurasia
in general – will have to pay serious attention to this system of thought, both in
its origin and in its development, because of the tremendous influence it has
exerted over the centuries.
Abbreviation
PPN Āgamic Index, vol. I: Prakrit Proper Names (ed. Malvania, 1970–1972)
Bibliography
Aklujkar, Ashok (2008). “Patañjali: a Kashmirian”, In Linguistic Traditions of Kashmir.

Essays in memory of Paṇḍit Dinanath Yaksha. Edited by Mrinal Kaul & Aklujkar
Ashok. New Delhi: D.K. Printworld, pp. 173–205.
Beckwith, Christopher I. (2009). Empires of the Silk Road: A History of Central Eurasia
from the Bronze Age to the Present. Princeton NJ & Oxford: Princeton University Press.
Beckwith, Christopher I. (2012). Warriors of the Cloisters. The Central Asian Origins of
Science in the Medieval World. Princeton NJ & Oxford: Princeton University Press.
Behrendt, Kurt A. (2004). The Buddhist Architecture of Gandhāra (Handbook of Oriental
Studies, 2/17). Leiden etc.: Brill.
Bollée, Willem B. (1998). Bhadrabāhu Bṛhat-kalpa-niryukti and Sanghadāsa Bṛhat-kalpa-
bhāṣya. Romanized and metrically revised version, notes from related texts and a
selective glossary. Part One: Pīṭhikā and Uddeśa 1. Stuttgart: Franz Steiner.
44 Bronkhorst
Bronkhorst, Johannes (1987). Three Problems Pertaining to the Mahābhāṣya (Post-

Graduate and Research Department Series, No. 30. “Pandit Shripad Shastri Deodhar
Memorial Lectures” [Third Series]). Poona: Bhandarkar Oriental Research Institute.
Bronkhorst, Johannes (1992). “Quelques axiomes du Vaiśeṣika”, Les Cahiers de Philosophie
14 (“L’orient de la pensée: philosophies en Inde”), pp. 95–110.
Bronkhorst, Johannes (1994). “A note on Patañjali and the Buddhists”, Annals of the
Bhandarkar Oriental Research Institute 75, pp. 247–254.
Bronkhorst, Johannes (1995). Review of Rospatt, 1995. Études Asiatiques / Asiatische
Studien 49(2), pp. 513–519.
Bronkhorst, Johannes (1996). “Sanskrit and reality: the Buddhist contribution”, In
Ideology and Status of Sanskrit: Contributions to the history of the Sanskrit language.
Edited by Jan E.M. Houben. Leiden etc.: E.J. Brill, pp. 109–135.
Bronkhorst, Johannes (1999). Why is there philosophy in India? Amsterdam: Royal
Netherlands Academy of Arts and Sciences (Sixth Gonda lecture, held on 13 November
1998 on the premises of the Royal Netherlands Academy of Arts and Sciences).
Bronkhorst, Johannes (2000). “Abhidharma and Jainism”, In Abhidharma and Indian
Thought. Essays in honor of Professor Doctor Junsho Kato on his sixtieth birthday.
Edited by the Committee for the Felicitation of Professor Doctor Junsho Kato’s
Sixtieth Birthday, Nagoya. Tokyo: Shuju-sha, pp. 598–581 ([13]–[30]).
Bronkhorst, Johannes (2001). “Pourquoi la philosophie existe-t-elle en Inde?” La ratio-
nalité en Asie / Rationality in Asia. Études de Lettres 2001/3, pp. 7–48.
Bronkhorst, Johannes (2002a). “A note on the Caraka Saṃhitā and Buddhism”, In Early
Buddhism and Abhidharma Thought: In Honor of Doctor Hajime Sakurabe on His
Seventy-seventh Birthday. Kyoto: Heirakuji Shoten, pp. 115–121.
Bronkhorst, Johannes (2002b). “Patañjali and the Buddhists”, In Buddhist and Indian
Studies in Honour of Professor Sodo Mori. Hamamatsu: Kokusai Bukkyoto Kyokai
(International Buddhist Association), pp. 485–491.
Bronkhorst, Johannes (2004). From Pāṇini to Patañjali: the search for linearity (Post-
graduate and Research Department Series, 46). Pune: Bhandarkar Oriental Research
Institute.
Bronkhorst, Johannes (2006). “T.R.V. Murti’s Reason”, Asiatische Studien / Études
Asiatiques 60(4), pp. 789–798.
Bronkhorst, Johannes (2009). Buddhist Teaching in India. Boston MA: Wisdom
Publications.
Bronkhorst, Johannes (2011). Buddhism in the Shadow of Brahmanism (Handbook of
Oriental Studies 2/24). Leiden – Boston: Brill.
Bronkhorst, Johannes (2012). “Buddhist thought versus Brahmanical thought”, In World
View and Theory in Indian Philosophy (Warsaw Indological Studies Series, 5). Edited
by Piotr Balcerowicz. New Delhi: Manohar, pp. 21–28.
Bronkhorst, Johannes (2013). “Reflections on the origins of Mahāyāna”, Septimo Cen

tenario de los Estudios Orientales en Salamanca. Salamanca: Ediciones Universidad
(Estudios Filológicos, 337), pp. 489–502.
Bronkhorst, Johannes (2016). How the Brahmins Won (Handbook of Oriental Studies
2/30). Leiden-Boston: Brill.
Cardona, George (1976). Pāṇini. A survey of research. Reprint: Motilal Banarsidass, Delhi,
1980.
Dutt, Sukumar (1962). Buddhist Monks and Monasteries of India: Their history and their
contribution to Indian culture. London: George Allen & Unwin.
Jain, Jagdishchandra (1984). Life in Ancient India as Depicted in the Jain Canon and
Commentaries: 6th century BC to 17th century AD. Second rev. and enl. edition. New
Delhi: Munshiram Manoharlal.
Lysenko, Victoria (2011). “Buddhist motives in some doctrines of Praśastapāda”, In
Vacaspativaibhavam. A Volume in Felicitation of Professor Vachaspati Upadhyaya.
Edited by Radha Vallabh Tripathi et al. Delhi: D.K. Printworld, pp. 1223–1233.
Malvania, Dalsukh (ed.) (1970–1972). Āgamic Index, vol. I: Prakrit Proper Names. Compiled
by Mohanlal Mehta & K. Rishabh Chandra. 2 parts (Lal Dalpatbhai Series, 28 & 37).
Ahmedabad: L.D. Institute of Indology.
Murti, Tirupattur Ramaseshayyer Venkatachala (1960). The Central Philosophy of
Buddhism. A study of the Mādhyamika system. Second edition. Reprint: Unwin
Paperbacks, London, 1980.
Pal, Pratapaditya (2007). “Evidence of Jainism in Afghanistan and Kashmir in ancient
times”, Bulletin of the Asia Institute 21, pp. 25–33.
Rospatt, Alexander von (1995). The Buddhist Doctrine of Momentariness. A survey of the
origins and early phase of this doctrine up to Vasubandhu (Alt- und Neu-Indische
Studien, 47). Stuttgart: Franz Steiner.
Salomon, Richard (1999). Ancient Buddhist Scrolls from Gandhāra. The British Library
Kharoṣṭhī Fragments. London: The British Library.
Scharfe, Hartmut (1977). Grammatical Literature (A History of Indian Literature, 5/2).
Wiesbaden: Otto Harrassowitz.
Schopen, Gregory (2004). Buddhist Monks and Business Matters. Still more papers on
monastic Buddhism in India. Honolulu HI: University of Hawai’i Press.
Schopen, Gregory (2006). “The Buddhist ‘monastery’ and the Indian garden: aesthetics,
assimilations, and the siting of monastic establishments”, Journal of the American
Oriental Society 126(4), pp. 487–505.
White, David G. (2012). “Netra Tantra at the crossroads of the demonological cosmopo-
lis”, The Journal of Hindu Studies 5, pp. 145–171.
Willemen, Charles (2006). The Essence of Scholasticism: Abhidharmahṛdaya. T 1550.
Revised edition with a completely new introduction. Delhi: Motilal Banarsidass.
46 Bronkhorst
Willemen, Charles (2012). “Remarks about the history of Sarvāstivāda in Northwestern

India”, In Buddhism in Kashmir (Śata-Piṭaka Series, 639). Edited by Nirmala Sharma.
New Delhi: Indian Council for Cultural Relations & Aditya Prakashan, pp. 162–164.
Willemen, Charles; Dessein, Bart & Cox, Collett (1998). Sarvāstivāda Buddhist Scholas
ticism (Handbook of Oriental Studies 2/11). Leiden etc.: Brill.
Abhidharmic Elements in Gandhāran Mahāyāna Buddhism 47
Chapter 2 Schlosser and Strauch
Abhidharmic Elements in Gandhāran Mahāyāna

Buddhism: Groups of Four and the
abhedyaprasādas in the Bajaur Mahāyāna Sūtra
Andrea Schlosser and Ingo Strauch
1 Introduction
From the various collections of Gāndhārī manuscripts, an increasing number

can be ascribed to the Mahāyāna branch of Buddhism. As of now seven early
Mahāyāna sūtras have been identified, supplemented by some scholastic texts,
which also appear to bear a Mahāyāna character.
The Gāndhārī Mahāyāna sūtras can be divided into two chronologically –
and probably also regionally – different groups.1 The more recent one is
represented by manuscripts which most likely originate from Bamiyan.
According to the paleography of their script and the advanced stage of their
orthography and language, they can be ascribed to the later period of
Gāndhāran literature, i.e. the third, early fourth century CE. This date is also
confirmed by radiocarbon dating. This more recent group comprises frag-
ments of Gāndhārī versions of already known Mahāyāna texts:
Skt. Bhadrakalpika-sūtra (ca. 60 fragments, Schøyen Collection, see Allon and

Salomon, 2010: 6f.; Baums, Glass and Matsuda, forthcoming)
Skt. Bodhisattvapiṭaka-sūtra (MS17, see Allon and Salomon, 2010: 8)
Skt. Sarvapuṇyasamuccayasamādhi-sūtra (MS89, see Allon and Salomon, 2010:
7f.)
The older manuscripts preceding this group were written most likely in the late
first, early second century CE. Again this date could be confirmed by radio
carbon dating (for the Prajñāpāramitā cf. Falk, 2011: 20). Although two of these
early manuscripts contain texts which can also directly be linked to extant ver-
sions of Mahāyāna works, two of them seem to represent texts which are
hitherto unknown and have not been transmitted in any of the known Buddhist
literary traditions. These texts promise new insights into the formative phase
1 For more details cf. Strauch, forthcoming. See also Allon and Salomon, 2010.

48 Schlosser And Strauch
of early Mahāyāna, when texts had yet to be harmonized into authoritative

versions. The texts of this second group comprise:
“Bajaur Mahāyāna sūtra” (BajC2, see Strauch, 2010; Strauch, forthcoming)

Skt. Prajñāpāramitā-sūtra (G. prañaparamida, SplitC5, see Falk, 2011; Falk and
Karashima, 2012, 2013)
Skt. *Sucitti-sūtra (unpublished private collection, see Allon and Salomon,
2010: 11)
Skt. Pratyutpannabuddhasaṃmukhāvasthitasamādhi-sūtra (unpublished pri-
vate collection, see Harrison and Hartmann, 2014: xvi, note 19)
At least two of these four early texts – the “Bajaur Mahāyāna sūtra” and the
Prajñāpāramitā – hail, according to reliable records, from the region along the
Pakistan-Afghanistan border, in the modern-day districts of Bajaur and Dir, i.e.
east of the Hindukush range.2
Among these early texts, the Bajaur Mahāyāna sūtra is of special interest.
First, it is by far the largest of these early Mahāyāna texts, and the longest text
in Gāndhārī known so far, comprising around six hundred lines on a large com-
posite birch bark scroll of about 2 meters length. Second, it belongs to those
texts, for which no parallel in another language is known.
The Bajaur Mahāyāna sūtra is part of the Bajaur Collection, which was alleg-
edly discovered in the ruins of a Buddhist monastery near the village Mian Kili
at the Dir-Bajaur border. The collection comprises texts of various Buddhist
literary genres, such as āgama, vinaya, rakṣā and stotra texts. A considerable
number of texts belong to the genre of scholastic literature, some of which
have a distinctive Mahāyāna tendency.3 Even non-Buddhist texts such as a
rājanīti verse anthology and a loan contract could be identified among the
birch barks of the Bajaur Collection.4
The study of the large Bajaur Mahāyāna sūtra is still ongoing. Although the
edition still needs some further research regarding certain passages, a more
general discussion and summary of its contents will be published soon.5
2 For the origin of the Bajaur Collection cf. Strauch, 2008; for the Split Collection see Falk, 2011.
3 The best preserved texts of this group were edited by Andrea Schlosser in her dissertation “On
the Bodhisattva Path in Gandhāra – Edition of Fragment 4 and 11 from the Bajaur Collection
of Kharoṣṭhī Manuscripts” (2013, revised version 2016).
4 For a general survey of the collection see Strauch, 2008. Separate texts are dealt with in
Strauch, 2011; Strauch, 2014a and 2014b.
5 The editing of the text is carried out by Ingo Strauch and Andrea Schlosser within a coopera-
tion between the Chair of Buddhist Studies at Lausanne University and the project “Early
Buddhist Manuscripts from Gandhāra” of the Bavarian Academy of Sciences, Munich.
The entire sūtra can be divided into two different narratives. The first, frame
narrative represents a dialogue between the Buddha and Śāriputra. This dia-
logue occurs at the Vulture Peak in Rājagṛha. The Bajaur sūtra shares this
location with other early Mahāyāna sūtras, as e.g. the Aṣṭasāhasrikā
Prajñāpāramitā and the Gāndhārī Prajñāpāramitā (cf. Falk and Karashima,
2012: 28).
The second narrative, which is mainly devoted to the description of the
bodhisattva path, reports the dialogue between the Buddha and 84,000
gods (devaputra). This part contains many elements which are known from
other early Mahāyāna texts, as e.g. the prediction of a future buddha land
(vyākaraṇa). It is a distinctive feature of the text that this prediction refers to
the buddha land Abhirati of the Buddha Akṣobhya. Moreover, the text cele-
brates dharmakṣānti, “endurance towards the [non-arising of the] factors of
existence,” as the major characteristic of a bodhisattva. As a characteristic pas-
sage of the sūtra’s approach we cite the following passage:6
(*e)[va vuto] bhag̱ava aï[śpa] (*śa)[r](*ip)u(*tro edad oya sarvadharma)

[ṇa] śariputra · ṇa as̱i prañayati · ṇa maje prañayati · ṇa p(*r)ayos̱aṇo pra-
ñayati ◦ yado ya · śariputra sarvadharma[ṇa] (*ṇa as̱i praña)yati ṇa maje
prañayati [‧] ṇa prayos̱aṇo prañayati · ṇa tas̱a śariputra dha(*r)
mas̱a [haṇi] praña[yati] ṇa ṭ́hi[di] (*pra)[ñayati ṇa veul](*o)[do pra]
(*ñayati ◦) yado ya śariputra ◊ sarvadharmaṇa ◊ · ṇa haṇi prañaïdi · ◊ ṇa
ṭ́hidi prañaya[d]i ◊ ṇa veulodo prañayadi ◊ ida ta śariputra · prag̱idie
(*acalo aṇalao dha)[rm](*o · ya) [śa]riputra ◊ acalo aṇalao ◊ dharma ◊
ida ta śariputra · imasvi dharmaviṇae ◊ saro (BajC2: 3H.44+1F.33–1F.36)
Thus addressed, the Blessed One said to the Venerable Śāriputra: (*Of all
dharmas), Śāriputra, a beginning (ādi) is not conceived, a middle (mad-
hya) is not conceived, an end (paryavasāna) is not conceived. And
because, Śāriputra, of all dharmas a beginning is not conceived, a middle
is not conceived, an end is not conceived, of this [single] dharma,
Śāriputra, a decrease (hāni) is not conceived, a stability (sthiti) is not con-
ceived, an increase (vaipulyatā) is not conceived. And because, Śāriputra,
6 The quotations and translations in this article are based on the ongoing edition by Ingo
Strauch and Andrea Schlosser. The conventions are those of the series Gandhāran Buddhist
Texts, i.e. [ ] uncertain reading, (*) editorial restoration of lost text, ⟨* ⟩ editorial addition of
omitted text, { } editorial deletion of redundant text, ? illegible akṣara, + lost akṣara, /// textual
loss at left or right edge of support (cf. <http://gandhari.org/a_dpreface.php>). ◊ signifies an
intentional space.
of all dharmas a decrease is not conceived, a stability is not conceived, an

increase is not conceived, this [single] dharma, Śāriputra, is by nature
immovable (acala) and baseless (anālaya). Which dharma, Śāriputra, is
immovable and baseless, this, Śāriputra, is the essence of this Dharma
and discipline (dharmavinaya).
Passages like this are not rare in early Mahāyāna sūtras, as is, for example,
shown by an almost literal parallel from the Aṣṭasāhasrikā (see Aṣṭa §2, ed.
Vaidya, 1960: 32).
The focus of early Mahāyāna sūtras towards the character of dharmas was
interpreted by Johannes Bronkhorst as clear evidence for the influence of
Gandhāran scholasticism on early Mahāyāna.7 As Bronkhorst (forthcoming)
argues:
It was in Greater Gandhāra, during this period, that Buddhist scholasti-

cism developed an ontology centered around its lists of dharmas. Lists of
dharmas had been drawn up before the scholastic revolution in Greater
Gandhāra, and went on being drawn up elsewhere with the goal of pre-
serving the teaching of the Buddha. But the Buddhists of Greater
Gandhāra were the first to use these lists of dharmas to construe an ontol-
ogy, unheard of until then. They looked upon the dharmas as the only
really existing things, rejecting the existence of entities that were made
up of them. Indeed, these scholiasts may have been the first to call them-
selves śūnyavādins. No effort was spared to systematize the ontological
scheme developed in this manner, and the influence exerted by it on
more recent forms of Buddhism in the subcontinent and beyond was to
be immense. But initially this was a geographically limited phenomenon.
It may even be possible to approximately date the beginning of this intel-
lectual revolution. I have argued in a number of publications that various
features of the grammarian Patañjali’s (Vyākaraṇa-)Mahābhāṣya must be
explained in the light of his acquaintance with the fundamentals of the
newly developed Abhidharma. This would imply that the intellectual
revolution in northwestern Buddhism had begun before the middle of
the second century BCE. If it is furthermore correct to think, as I have
argued elsewhere, that this intellectual revolution was inspired by the
interaction between Buddhists and Indo-Greeks, it may be justified to
situate the beginning of the new Abhidharma at a time following the
renewed conquest of Gandhāra by the Indo-Greeks; this was in or around
7 On Sarvāstivāda scholastisicm in the northwest (Gandhāra and Bactria) see Willemen, Dessein
and Cox, 1998: 255–285.
185 BCE. The foundations for the new Abhidharma may therefore have
been laid toward the middle of the second century BCE.
Based on this statement, it seems worthwhile to have a closer look at the

interrelationship of both literary genres. The present paper will confine the
discussion to the evidence of the Bajaur Mahāyāna sūtra as one of the ear-
liest attested manuscripts of a Gandhāran Mahāyāna text. It will focus on a
passage within the introductory portion that is devoted to the description
of an ideal disciple of the Buddha (āryaśrāvaka) and culminates in a list of
Buddhist scholastic terms, which are grouped as four. The list of fours is fol-
lowed by the discussion of another category of fours, called abhejapras̱ada,
Skt. abhedyaprasāda. According to the amount of text, which is devoted to this
category (ca. 53 out of 121 lines of the introduction), they play a dominant role
among the characteristics of an āryaśrāvaka as conceived in this text.
The first part of this paper – written by Andrea Schlosser – investigates the
relationship of the list of groups of four to Abhidharma and Prajñāpāramitā lit-
erature.8 The second part – written by Ingo Strauch – is particularly devoted to
the abhedyaprasādas, which seem to represent the raison d’être which caused
the inclusion of this list into the text of the Bajaur Mahāyāna sūtra. Both parts
try to establish the position of the Bajaur Mahāyāna sūtra within the debates of
early Abhidharma discourses and to find the mechanisms that accompanied
the transition of Abhidharma thinking into a Mahāyāna context.
2 Groups of Four
2.1 Position in the Text

Immediately preceding the list consisting of groups of four, Śāriputra enumer-
ates several things, such as rūpasthiti or vedanotpāda, that are not perceived by
a tathāgata. The bhagavant approves and says that also his disciple does not
perceive anything of it. He goes on to ask:
ta ki mañas̱i śariputra ◊ vida⟨*vi⟩di9 te dharma ya ma ? ? + + + + +

What do you think, Śāriputra, … these dharmas, which …
8 Thanks go to Paolo Visigalli and Lin Qian for reading and commenting upon an earlier draft
of this part.
9 The meaning of this word is yet uncertain. The vi is inserted based on the spelling in the
preceding passage, where the word occurs twice as vidavidi.
Figure 1 Extract from BajC2, part 1 ( frame 2).
Since the end of the line is not preserved any more (1B.14, see fig. 1), the exact
wording cannot be established and also the verbal connection to the following
passage is unclear, but it should still be spoken by the bhagavant and the miss-
ing portion should only comprise the end of the rhetorical question addressed
to Śāriputra.
2.2 The List

The following passage in response to the preceding question contains a long
list of items. All of them consist of four members. The reconstruction is as fol-
lows (BajC2, 1B.15–17+E.27–28):
(*catvari ś̱paḏovaṭ́haṇa ·) (*Four establishments of mindfulness),

(*catvari) samapra[s̱aṇa] · (*four) right endeavours,
[catvari] irdh[ipada] · four bases of [supernormal] power,
catvari jaṇa10 · four [stages of] meditation,
catvari saca · four truths,
catvari apramaña · four unlimited,
catvare ? + + + + +11 four …,
[catvare] (*pa)ḍ̱isabhida · four analytical knowledges,
catvare va[ś]ida · four masteries,
ca[tvare veharaja] · four self-confidences,
10 Possibly written with a stroke above the j (G. j̄aṇa) as etymologically expected, but the
manuscript is folded here, concealing the upper part.
11 Perhaps a term corresponding to the four ārūpyasamāpattis or the four samādhibhāvanās.
catvare ñaṇamulea dharma · four things rooted in knowledge,

catvare so[ḏavati](*aga ·) four factors of stream entry,
++++++??+ …,
catvare taṣ̄amulea dharma (*·) four things rooted in craving,
catvare paḍ̱i[va](*da ·) four kinds of progress,
++++++++++ …,
+++++++??? …,
[catvare] ? ? ? ? + +12 [four] (*stations of consciousness),
+ + + + + + ? ? ? …,
[ca]tvare sakṣig̱araṇia dharma four things to be realized,
catvari as̱aharia dharma four unconquerable things.
2.3 Context and Meaning

2.3.1 Similar Lists in Buddhist Literature
Parts of the list in the Bajaur manuscript are familiar from other texts, but
the whole set is, to my knowledge, not found in any other Indic text source.
The first three items are identical with the first three items of the thirty-seven
bodhipakṣya-dharmas, the factors conducive to awakening. These are: four
smṛtyupasthānas, four samyakpradhānas, four ṛddhipādas, five indriyas, five
balas, seven bodhyaṅgas, and the eightfold mārga. This listing was studied by
Johannes Bronkhorst in his article “Dharma and Abhidharma” (1985), where
he compares its occurrences throughout Buddhist literature and divides
them into four phases of development.13 The sequence in BajC2 corresponds
to Bronkhorst’s categories III or *IV with addition of the four dhyānas (II14)
and the four apramāṇas (III) and possibly also the four ārūpya(samāpatti)s
(*IV15), see table 2.1. Since at the crucial point in BajC2 the birch bark is broken
off, it cannot be determined if the ārūpyas had been included or not, but the
remaining traces of the first akṣara do not suggest an a, thus speaking against
ārūpya.
12 Most probably this is G. viñaṇaṭ́hiḏio, based on the remaining traces of ink and the
sequence in two versions of the Saṃgītisūtra/-paryāya, namely G Cm and T.1.1 (cf.
table 3).
13 Cf. also Gethin, 2001, especially pp. 264–283 about the ‘seven sets expanded’.
14 Cf. Bronkhorst, 1985: 306, note 8: “It is remarkable that the Dīrghāgama preserved in Chi-
nese seems to have only list II, not I”. Also a list in another Gāndhārī manuscript (frag-
ment 5 of the Senior Collection) contains this extended list (II) corresponding to the
Chinese Dīrghāgama (see Glass, 2007: 35).
15 This combination is not extant in any text but a “hypothetical construction” (Bronkhorst,
1985: 308).
Table 1 Development of lists based on the bodhipakṣya-dharmas after Bronkhorst, 1985.
I II III *IV
smṛtyupasthāna smṛtyupasthāna smṛtyupasthāna smṛtyupasthāna

samyakpradhāna samyakpradhāna samyakpradhāna samyakpradhāna
ṛddhipāda ṛddhipāda ṛddhipāda ṛddhipāda
dhyāna dhyāna dhyāna
apramāṇa apramāṇa
ārūpya
indriya indriya indriya
bala bala bala
bodhyaṅga bodhyaṅga bodhyaṅga
ārya aṣṭāṅga mārga ārya aṣṭāṅga mārga ārya aṣṭāṅga mārga
bodhipakṣya-dharmas > some canonical > Dhātukathā, groups of four

> sūtras and Vinaya sūtras Vibhaṅga > Saṃgītisūtra/-
> Ch Dīrghāgama paryāya
(only II, not I)
The same set of items (III, excluding the ārūpyas) also occurs in the Dhātukathā16
or the Vibhaṅga. In the latter, the next (and last) group comprising four items
are the four paṭisambhidās, quite similar to the BajC2 list: (4) ariyasaccāni,
(7) satipaṭṭhānā, (8) sammappadhānā, (9) iddhipādā, (12) jhānāni, (13)
appamaññāyo, (15) paṭisambhidā (the numbering reflects the chapters, cf.
also Frauwallner, 1995: 17f.).17 Thus, all seven fourfold categories discussed in
16 Bronkhorst, 1985: 306 (page 1 of the PTS edition, cf. also Narada, 1962: xlviii).
17 The ārūpyas have been included under the heading jhāna, but this is thought to be a
later addition. According to Bronkhorst, 1985: 308, a part of the Vibhaṅga (pp. 193–305)
is based on the following list: (1) 4 satipaṭṭhāna, (2) 4 sammappadhāna, (3) 4 iddhipāda,
(4) 7 bojjhaṅga, (5) 8-aṅgika magga, (6) 4 jhāna, (7) 4 appamaññā, (8) 5 sikkhāpada,
(9) 4 paṭisambhidā, thus excluding the 4 sacca. The Vibhaṅga is believed to “have
developed out of an earlier work [before 200 BCE] which also underlay the Dharma
skandha of the Sarvāstivādins” (Bronkhorst, 1985: 308). There, the ārūpyas as well as
the satyas are contained (T.26.1537: 453b24–514a10, 阿毘達磨法蘊足論, Apidamo
fayun zu lun, tr. by Xuanzang 玄奘). However, the sequence of the chapters is differ-
ent (Frauwallner, 1995: 15f. (= 1964: 73–74)): (2) srotāpattyaṅgāni, (3) avetyaprasādāḥ,
(4) śrāmaṇyaphalāni, (5) pradipadaḥ, (6) āryavaṃśāḥ, (7) samyakpradhānāni, (8)
ṛddhipādāḥ, (9) smṛtyupasthānāni, (10) āryasatyāni, (11) dhyānāni, (12) apramāṇāni, (13)
ārūpyāṇi, (14) samādhibhāvanāḥ (only chapter (1) [the 5 śikṣāpadāni] and chapter (15)
[the 7 bodhyaṅgāni] of the first part (–494b29) are not groups of four).
the Vibhaṅga are contained within the BajC2 list under the first eight items,
leaving no indication for the uncertain item no. 7 though. Nevertheless, one
tentative conclusion might be that the listing in BajC2 is based on a list also
occurring in the Vibhaṅga and Dhātukathā, but with the inclusion of the
ārūpyas or another additional group of four. The satyas, however, are posi-
tioned not at the beginning but in the middle of two categories pertaining to
meditation. Interestingly, they are placed likewise in the Sanskrit version of
the Saṃgītisūtra (‘Skt’) and its Chinese commentary (T.26.1536), additionally
followed by saṃjñā / xiang 想, see table 2. The Pāli version (‘P’) lists samādhi-
bhāvanā instead, the satyas (P. sacca) or saṃjñās (P. saññā) are not contained
at all; in the Gāndhārī commentary and the other Chinese versions they are
inserted later (satya: G Cm 23., T.1.1: 23., T.1.12: 9.; saṃjñā: G Cm 34., T.1.1: 34.,
T.1.12: –).18
Table 2 Sequence of dhyāna, apramāṇa and ārūpya in the different versions of the Saṃgītisūtra/-
paryāya.
G Cm T.1.1 T.1.12 P BajC2 Skt T.26.1536
13.a jaṇa 14. 禪 04. 禪定 04. jhānāni

04. dhyānāni 04. jaṇa 04. 靜慮
05. samādhi-
05. ārya- 05. saca 05. 聖諦
bhāvanā satyāni
06. saṃjñāḥ 06. 想
15. apravaṃñā 15. 梵堂 b 05. 無量 06. 06. 07. 07. 無量
appamaññāyo apramaña apramāṇāni
16. arupa 16. 無色定 06. 無色 07. arūpā 07. ? 08. ārūpyāṇi 08. 無色
[sa]⸨ma⸩vatie 定 (v.l. āruppā)
a G Cm interchanges idhivaḏa (14.) and j̱aṇa (13.).

b ~ brahmavihāra.
Thus, G Cm, T.1.2, and T.1.12 seem to represent an older version of this particu-
lar sequence, where the first three fourfold items of the bodhipakṣya-dharmas
plus the dhyānas are immediately followed by the ‘unlimited’. The list in BajC2
should be more recent than the list preserved in these versions, but older than
the one in Skt/T.26.1536, however being part of the same strand of develop-
ment that included the satyas at this position (so far only attested in
18 For references and more information about the different versions of the Saṃgītisūtra/-
paryāya see table 3.
Sarvāstivāda versions of the Saṃgītisūtra/-paryāya). After the ‘unlimited,’ all

versions of the Saṃgīti◦ agree again in listing the ‘formless,’ but it has to be left
open what is to be reconstructed in BajC2.
All other legible groups of four occurring in the Bajaur manuscript are
mostly contained in the extant versions of the Saṃgītisūtra and its commen-
taries, even though not in the same sequence, see table 2.3.19
The (Gāndhārī) terms only occurring in some versions are:
- saca Skt, T.26.1536, T.1.12, T.1.1, G Cm.

- (*pa)ḍ̱isabhida T.1.12, T.1.1, G Cm.
- taṣ̄amulea dharma Skt, T. 26. 1536, P, T.1.12.
- sakṣig̱araṇia dharma Skt, T. 26. 1536, P, T.1.1, G Cm.
No unambiguous pattern can be observed that would show a distinctive affili-

ation of BajC2 to one or other version of the Saṃgītisūtra/-paryāya. The list in
BajC2 resembles the Skt/T.26.1536/P/T.1.12 versions in that it likewise begins
with the smṛtyupasthānas. Among these, it seems somehow connected with
T.1.12 in that it includes the (*pa)ḍ̱isabhidas (wu’aijie 無礙解), although on the
other hand T.1.12 strangely lacks the sakṣig̱araṇia-dharmas and also adds the
sacas (shengdi 聖諦) at a later position. Regarding the sequence, BajC2 seems
most similar to P (though with gaps) but includes, as already said, the sacas
and the (*pa)ḍ̱isabhidas.
Terms only occurring in BajC2 are:
- va[ś]ida
- [veharaja]
- as̱aharia dharma
19 In the Sanskrit Dīrghāgama manuscript from Gilgit, the relevant passages of the
Saṃgītisūtra are too fragmentary to be taken into consideration here. Apparently, only
IV.12 and IV.20 are preserved partly (thanks to Jens-Uwe Hartmann for sharing unpub-
lished information). In the table, ~ indicates that the equivalence is uncertain.
Table 3 Comparison of the list in BajC2 with the groups of four occurring in the Saṃgītisūtra/-paryāya.
BajC2 Skt T.26.1536 P T.1.12 T.1.1 G Cm
01. (*catvari ś̱paḏovaṭ́haṇa ·) 01. smṛtyupasthāna 01. 念住 01. satipaṭṭhānā 01. 念處觀 11. 念處 11. [ś̱paḏ]ova[ṭ́ha]ṇa
02. (*catvari) samapra[s̱aṇa] · 02. samyakprahāṇa 02. 正斷 02. sammappadhānā 02. 正斷 12. 意斷 12. saṃmepras̱aṇa
03. [catvari] irdh[ipada] · 03. ṛddhipāda 03. 神足 03. iddhipādā 03. 神足 13. 神足 14. idhivaḏa
04. catvari jaṇa · 04. dhyāna 04. 靜慮 04. jhānāni 04. 禪定 14. 禪 13. j̱aṇa
05. catvari saca · 05. āryasatya 05. 聖諦 – 09. 聖諦 23. 聖諦 23. arias̱aca
06. catvari apramaña · 07. apramāṇa 07. 無量 06. appamaññāyo 05. 無量 15. 梵堂 15. apravaṃña
07. catvare ? + + + + +
08. [catvare] (*pa)ḍ̱isabhida · – – – 26. 無礙解 27. 辯才 27. paḍisaṃbiḏa
09. catvare va[ś]ida · – – – – – –
10. ca[tvare veharaja] · – – – – – –
11. catvare ñaṇamulea dharma · ~ 13. jñāna ~ 13. 智 ~ 11. ñāṇāni ~ 07. 智 ~ 26. 智 ~ 26. ñaṇa
12. catvare so[ḏavati](*aga ·) 12. aṅgaiḥ samanvāgataḥ 12. 證淨 14. sotāpannassa aṅgāni 18. 預流身 20. 須陀洹支 20a. soḏavatiaga
srotāpannaḥ
13. + + + + + + ? ? +
14. catvare taṣ̄amulea dharma (*·) ~ 24. tṛṣṇotpāda ~ 34. 愛 ~ 20. taṇhuppādā ~ 30. 愛生 – –
Abhidharmic Elements in Gandhāran Mahāyāna Buddhism
15. catvare paḍ̱i[va](*da ·) 31. pratipad 21. 行 21. paṭipadā 17. 神通道 22. 道 22. paḍivaḏa
16. + + + + + + + + + +
17. + + + + + + + ? ? ?
18. [catvare] ? ? ? ? + + 23. vijñānasthiti 33. 識住 18. viññāṇaṭṭhitiyo 13. 識住 28. 識住處 28. viñaṇaṭ́hiḏio
(= viñaṇaṭ́hiḏio?)
57
58
Table 3 Comparison of the list in BajC2 with the groups of four occurring in the Saṃgītisūtra/-paryāya (cont.).
BajC2 Skt T.26.1536 P T.1.12 T.1.1 G Cm
19. + + + + + + ? ? ?
20. [ca]tvare sakṣig̱araṇia dharma 20. sākṣīkaraṇīya dharma 20. 應證法 30. sacchikaraṇīyā – 21. 受證 21. sakṣiḵatava
dhammā
21. catvari as̱aharia dharma – – – – –
G Cm: BL15, ca. 0–100 CE, not published yet (groups of four to be edited by Stefan Baums).
P: Saṃgītisuttanta, DN III 221–233.
Skt: Saṃgītisūtra, Central Asian manuscripts, ca. 7th c. CE, ed. Stache-Rosen, 1968.
T.1.1: ~ Saṃgītisuttanta (眾集經, Zhongji jing, DĀ, sūtra no. 5), T.1.1.50b23–51b4,
tr. Buddhayaśas, ca. 5th c. CE, tr. in Behrsing, 1930.
T.1.12: ~ Saṃgītiparyāyasūtra (大集法門經, Daji famen jing = *Mahā-saṃgīti-sūtra), T.12.1.228b16–230a5, tr. by Dānapālaa, ca. 1000 CE.
T.26.1536: ~ Saṃgītiparyāya (阿毘達磨集異門足論, Apidamo jiyimen zulun = *Abhidharma-saṃgīti-paryāya-pāda-śāstra), T.26.1536.26.391b11–411c11), tr. by
Xuanzang, 660–663 CEb, tr. in Stache-Rosen, 1968.
Skt is closely connected to T.26.1536 (Sarvāstivāda), which is a commentary on it. P (Theravāda) and T.1.12 (affiliation unknown) seem to stand for themselves,
although being connected to Skt/T.26.1536 due to the same beginning. G Cm is closely connected to T.1 (Dharmaguptaka).
a Also BajC1 (*Gautamīsūtra) has a closely related, though not identical Chinese version that has been translated by Dānapāla (T.1.84, ca. 980–1000 CE, cf.
Strauch, 2007/2008: 19–21, and Strauch, 2014a: 33). He was from Uḍḍiyāna (Swat), his school affiliation is unknown (cf. Strauch, 2014 a: 26 and 35).
b Willemen, Dessein and Cox, 1998: 177.
Schlosser And Strauch
G va[ś]ida. The vaśitās (“masteries / powers”) are normally classified as being

ten20, but they are different from the ten balas (of a tathāgata or bodhisattva)21.
So far, it seems that the only Sanskrit texts mentioning only four vaśitās are the
Mahāyānasūtrālaṃkāra (Msa) and the Madhyāntavibhāga (MAV) transmitted
by Asaṅga, and Vasubandhu explains them as “the masteries in the absence of
conceptual discrimination, in the purification of a field, in awareness, and in
action” (caturdhā vaśitā nirvikalpavaśitā kṣetrapariśuddhivaśitā jñānavaśitā
karmavaśitā ca, tr. D’Amato, 2012: 140, commentary on MAV 2.15, ed. Nagao,
1964: 35; cf. Msa 11.45–46). In the Chinese version of the Madhyāntavibhāga,
this is rendered as si zizai 四自在 (four kinds of unhinderedness / mastery),
explained as being the unhinderedness of non-discrimination, wufenbie zizai
無分別自在, pure land, jingtu zizai 淨土自在, knowledge, zhi zizai 智自在, and
karma, ye zizai 業自在.22 The four vaśitās (si zizai 四自在) seem far more fre-
quent in Chinese than in Sanskrit, but a more detailed study of this group of
four has to be postponed.23
20 Dhsgr 74: āyur-, citta-, pariṣkāra-, dharma-, ṛddhi-, janma-, adhimukti-, praṇidhāna-,
karma-, jñāna-◦. A partly different explanation is given in the Abhisamayālaṃkāravṛttiḥ
sphuṭārthā (AAV, ed. Tripathi, 1977: 3–44) on Abhisamayālaṃkāra (AA) 8.4: āyuś-citta-
pariṣkāra-karmopapatty-adhimukti-praṇidhāna-rddhi-jñāna-dharma-vaśitā iti daśa vaśi
tāḥ. See Brunnhölzl, 2011: 114 and also Brunnhölzl, 2010: 659 (chart 12) for a translation.
The same list is given e.g. in the PvsP (fol. 532b; ed. Kimura, 2006 [VI–VIII]: 59) and the
Sāratamā (ed. Jaini, 1979: 176), and – slightly varied – the Catuḥstavasamāsārtha (ed.
Tucci, 1956: 239).
21 Cf. e.g. Dhsgr 75: bodhisattvānāṃ daśa balāni / tadyathā // adhimuktibalaṃ pratisaṃ
khyānabalaṃ bhāvabalaṃ kṣāntibalaṃ jñānabalaṃ prahāṇabalaṃ samādhibalaṃ
pratibhānabalaṃ puṇyabalaṃ pratipattibalaṃ ceti // 76. tathāgatasya daśa balāni /
tadyathā // sthānāsthānajñānabalaṃ karmavipākajñānabalaṃ nānādhātujñānabalaṃ
nānādhimuktijñānabalaṃ sattveṃdriyaparāparajñānabalaṃ sarvatragāminīpratipatti
jñānabalaṃ dhyānavimokṣasamādhisamāpattisaṃkleśavyavadānavyutthānajñānabalaṃ
pūrvanivāsānusmṛtijñānabalaṃ cyutyutpattijñānabalam āsravakṣayajñānabalaṃ ceti.
22 T.31.1599: 455a7–8 (MAVBh, Paramārtha), T.31.1600: 468b5–6 (MAVBh, Xuanzang); root
text: T.31.1601: 478b25, MAV, Xuanzang). Another explanation of the ‘four sovereign
powers’ is: jie 戒 the moral law; shentong 神通 supernormal powers; zhi 智 knowledge;
and hui 慧 wisdom (Soothill, according to the Digital Dictionary of Buddhism, www.
buddhism-dict.net).
23 A worthwile start would be the passages in T.1.13 (Chang ahan shi bao fa jing 長阿含十報
法經), T.2.125 (Zengyi ahan jing 増一阿含經 ~ Ekottarāgama), T.9.272 (Da sazhe niganzi
suoshuo jing 大薩遮尼乾子所説經 ~ Mahāsatyanirgrantha-sūtra), T.10.279 (Dafang-
guang fo huayan jing 大方廣佛華嚴經).
G [veharaja]. In the Pāli canon, the four vesārajjas are explained as the
self-confidences or fearlessnesses of a buddha, because of which he cannot
be reproved by an “ascetic or brahmin or deva or Māra or Brahmā or anyone
in the world” of (1) not having reached full enlightenment, of (2) not having
destroyed all taints, of (3) not having understood the obstructions, of (4) not
having taught the correct way to the destruction of suffering.24 Also, in the
Mahāvastu, they are enumerated as one of the characteristics of a buddha.25
They are more often mentioned in Sanskrit texts, most of which are Mahāyāna-
related, where they likewise determine characteristics of an awakened being.
The four vaiśāradyas are also included in similar lists in Prajñāpāramitā texts,
but here they are always preceded by the (ten tathāgata-) balas and not by
vaśitās (cf. table 4).26 The apparently only text listing the (ten) vaśitās is the
Abhisamayālaṃkāra, where they occur adjacent to and precede the (ten) balas
(cf. table 4, A).27 This list is similar to the list in BajC2 in regard to the sequence
of the fours, even though some of the groups are not mentioned (jaṇa = Skt.
dhyāna, saca = Skt. satya, and the uncertain one). It is however different from
other lists in Prajñāpāramitā texts (cf. table 4, B).
24 AN II 9, book of fours, sutta 8 (tr. Bodhi, 2012: 394f.); MN I 7, Mahasīhanāda-sutta (tr.

Bodhi/Ñāṇamoli, 1995: 167f.). Cf. also AN IV 83, book of sevens, sutta 55 (tr. Bodhi, 2012:
1056f.); here they are listed as the “three things about which he is irreproachable” (the
three things are the four vesārajja related to the dhamma, sutta, and saṃgha).
25 Cf. Binz, 1980: 81 and 88. The characteristics are: 32 marks (lakṣaṇa), 80 secondary marks
(anuvyañjana), 18 special characteristics (āveṇikadharma), 10 powers (bala), 4 self-confi-
dences (vaiśāradya), setting the wheel of Dharma in motion (dharmacakrapravartana),
and the harmonic leading of the saṃgha.
26 Also in a Kharoṣṭhī manuscript of the first century the four vaiśāradyas are preceded by
the ten balas, cf. BL9 r3: vriṣavida‧daśabalada ca‧caduveharajada ca‧“mastery and
the state of possessing the ten powers and the state of possessing the four confidences”
(Baums, 2009: 329).
27 AA 8.4 sarvākārāścatasro ’tha śuddhayo vaśitā daśa / balāni daśa catvāri vaiśāradyāny
arakṣaṇam, which is part of a list of the 21 features of a dharmakāya (AA 8.2–6, cf. Conze,
1954: 96f.). The Abhisamayālaṃkārāntaḥ, AAV and Sāratamā refer to this passage and thus
contain the vaśitā as well.
Table 4 Listings in Prajñāpāramitā texts in comparison to the list in BajC2.

a.
BajC2 AA 8.2–6 / AAV / Abhisamayālaṃkārāntaḥ (ed. Tripathi, 1977: 1–67),

the last one is being cited.
01. (*catvari ś̱paḏovaṭ́haṇa ·) (1.) smṛtyupasthānādyārabhya āryāṣṭāṅgamārgaparyantā

02. (*catvari) samapra[s̱aṇa] · saptatriṃśad bodhipakṣāḥ,
03. [catvari] irdh[ipada] ·
04. catvari jaṇa · (2.) catvāryapramāṇāni maitryādicaturbrahmavihārāḥ,

05. catvari saca · (3.) aṣṭau vimokṣāḥ,
06. catvari apramaña · (4.) navasamāpattayaḥ,
07. catvare ? + + + + + (5.) kṛtsnāyatanāni daśa
(6.) aṣṭau abhibhvāyatanāni,
(7.) araṇāsamādhiḥ,
(8.) praṇidhijñānam,
(9.) ṣaḍabhijñāḥ,
08. [catvare] (*pa)ḍ̱isabhida · (10.) catasraḥ pratisaṃvidaḥ,

09. catvare va[ś]ida · (11.) āśrayālambanacittajñānapariśuddhaya iti cataśraḥ śuddhayaḥ,
10. ca[tvare veharaja] · (12.) daśa vaśitāḥ,
(13.) daśa balāni,
(14.) catvāri vaiśāradyāni,
11. catvare ñaṇamulea dharma · (15.) trīṇi arakṣaṇāni,

12. catvare so[ḏavati](*aga ·) (16.) trīṇi smṛtyupasthānāni,
13. + + + + + + ? ? + (17.) asammoṣadharmatā,
14. catvare taṣ̄amulea dharma (*·) (18.) kleśajñeyāvaraṇānuśayarūpabījaprahāṇāt vāsanāsamudghātaḥ,
15. catvare paḍ̱i[va](*da ·)
16. + + + + + + + + + + (19.) sakalajanahitāśayatā mahākaruṇā,
17. + + + + + + + ? ? ? (20.) aṣṭādaśāveṇikā buddhadharmāḥ,
18. [catvare] ? ? ? ? + + (21.) sarvākārajñatāditrisarvajñatā
19. + + + + + + ? ? ?
20. [ca]tvare sakṣig̱araṇia dharma
21. catvari as̱aharia dharma

Table 4 Listings in Prajñāpāramitā texts in comparison to the list in BajC2 (cont.).

b.
Aṣṭasāhasrikā, Larger PP (T.8.222) Larger PP (LPG), Larger PP (PvsP), Śatasāhasrikā,

ed. Vaidya 1960: 97 [Lokakṣema, ed. Conze 1962, ed. Kimura, ed. Kimura,
268 CE] 1974 1986–2009 2009–2010
(四禪 ~ 4 dhyāna)* catvāri dhyānāni catvāri catvāri

smṛtyupasthānāni smṛtyupasthānāni
(四等心 catvāry catvāri samyak- catvāri
~ 4 apramāṇa)* apramāṇāni prahāṇāni samyakprahāṇāni
(四無色三昧 catasraḥ ārūpya- catvārariddhipādā catvāra ṛddhipādāḥ
~ 4 arūpya)* samāpattayo
pañcendriyāṇi pañcendriyāṇi
saptatriṃśad- pañcabalāni pañcabalāni
bodhipakṣyā
saptabodhyaṅgāny saptabodhyaṅgāni
dharmā
*= catvāri 四意止 ~ 4 catvāri āryāṣṭāṅgo mārgo āryāṣṭāṅgo mārgaḥ
smṛtyupasthānāni smṛtyupasthāna smṛtyupasthānāni
*= catvāri 四意斷 ~ 4 catvāri
samyakprahāṇāni samyak-prahāṇa samyakprahāṇāni
(catvāry āryasatyāni) catvāry āryasatyāni
*= 四神足 ~ 4 catvārariddhipādā
catvārariddhipādā ṛddhipāda pañcendriyāṇi
*= pañcendriyāṇi 五根 ~ 5 indriya pañcabalāni
*= pañcabalāni 五力 ~ 5 bala saptabodhy-
catvāry apramāṇāni catvāri dhyānāni
*= saptabodhy- 七覺意 ~ 7 aṅgāny
aṅgāny bodhyaṅgāny āryāṣṭāṅgo mārgo catvāri dhyānāni catvāry apramāṇāni
*= āryāṣṭāṅgo 八由行 ~ 8-mārgo catasraḥ ārūpya- catasra ārūpya-
mārgo samāpapramāṇāni
catvāry attayo samāpdhyānāni
catvāri attayaḥ
catvāri dhyānāni catvāry apramāṇāni
(triṇi vimokṣa-
mukhāni)
(aṣṭau vimokṣā) (aṣṭau vimokṣā) aṣṭau vimokṣāḥ
(navānupūrva- (navānupūrva- navānupūrvavihāra
samāpattī) samāpattī) samāpattayaḥ
(…) (śūnyatānimittāpraṇi śūnyatānimittāpraṇi-
hitavimokṣamukhāni) hitavimokṣamukhāni
(abhijñāḥ) pañcābhijñāḥ
(sarvaśūnyatāḥ)
(sarvasamādhayaḥ) sarvasamādhayaḥ
(sarvadhāraṇī- sarvadhāraṇī-
mukhāni) mukhāni
Aṣṭasāhasrikā, Larger PP (T.8.222) Larger PP (LPG), Larger PP (PvsP), Śatasāhasrikā,

ed. Vaidya 1960: 97 [Lokakṣema, ed. Conze 1962, ed. Kimura, ed. Kimura,
268 CE] 1974 1986–2009 2009–2010
balāni 怛薩阿竭十種力 daśatathāgata daśatathāgatabalāni daśatathāgatabalāni

~ 10 tathāgatabala balāni
vaiśāradyāni 四無所畏 ~ 4 catvāri catvāri vaiśāradyāni catvāri vaiśāradyāni
vaiśāradya vaiśāradyāni
pratisaṃvido 四分別辯 ~ 4 catasraḥ catasraḥ catasraḥ
pratisaṃvid pratisaṃvido pratisaṃvido pratisaṃvidaḥ
aṣṭādaśāveṇikā 十八不共諸佛之 mahāmaitrī * ((mahāmaitrī)) mahāmaitrī

buddhadharmāḥ 法 ~ 18 mahākaruṇā * ((mahākaruṇā)) mahākaruṇā
aveṇika-
buddhadharma
(大慈 ~ aṣṭādaśāveṇikā aṣṭādaśāveṇikā aṣṭādaśāveṇika-

mahāmaitrī)* buddhadharmā buddhadharmā buddhadharmāḥ
(大悲 ~
mahākaruṇā)*
* *
154b19–22: without placed here or
dhyāna etc. at the end of
153a16–19: with the list
dhyāna etc.
149b08–09: with
maitrī/karuṇā
These lists can be analysed as consisting of several modules (consist-

ing themselves of several terms), the positions of which can change and in
between of which additional terms can be added. The most basic list is
found in the presumably oldest Prajñāpāramitā text, the Aṣṭasāhasrikā (8th
chapter, viśuddhiparivarta, ed. Vaidya: 97). It consists of the 37 bodhipakṣya-
dharmas, followed by the bala / vaiśāradya / pratisaṃvid, and concluded
by the 18 āveṇikabuddhadharmas.28 In the earliest Chinese translation of
the Larger Prajñāpāramitā by Lokakṣema (268 CE), the same list is found
(T.8.222.1.154b19–22), but so is an enlarged version which adds the four
28 saptatriṃśad bodhipakṣā dharmā balāni vaiśāradyāni pratisaṃvido aṣṭādaśāveṇikā

buddhadharmāḥ. In other passages (ed. Vaidya: 37, 103, 246) these categories are already
combined with other terms (like the abhijñās or the three vimokṣamukhas).
dhyānas / apramāṇas / ārūpyas at the beginning (T.8.222.1.153a16–19). This

list is further expanded by the addition of the mahāmaitrī and mahākaruṇā
(T.8.222.1.149b8–9). In the Gilgit manuscript of the Larger Prajñāpāramitā
(LPG), for example, this list was still further expanded by the insertion of sev-
eral terms between the bodhipakṣya-dharmas and the daśatathāgatabalas.
Moreover, the two ‘mahās’ (maitrī and karuṇā) could change their position
with the 18 āveṇikadharmas.29 Furthermore, as for the instance preserved
in the Nepalese manuscript of the Pañcaviṃśatisāhasrikā (PvsP) and in the
Śatasāhasrikā, the first block (dhyāna / apramāṇa / ārūpya) was moved, so
that the list would begin again with the traditional bodhipakṣya-dharmas.30
Although more details are to be taken into account in studying the develop-
ment of these lists, it becomes clear that they consisted of several blocks or
modules. The list in BajC2 seems to be based on the same module-based sys-
tem, beginning with the fourfold groups of the bodhipakṣya-dharmas and
adding the fourfold categories related to meditation (up to the ‘unlimited’ and
probably also the ‘formless absorptions’) plus adding the four ‘truths.’
Subsequently, three Mahāyāna- or Prajñapāramitā-typical categories that
characterize a tathāgata (pratisaṃvid / vaśitā / vaiśāradya) were added,
although the (four) vaśitās are replaced by (ten) balas in other texts.
In a passage in the PvsP (ed. Kimura, 2009 [I-2]: 27, also 32) the bodhipakṣya-
dharmas and the tathāgatabalas etc. are characterized as anāsrava- and
asādhāraṇa-dharma (together with the three vimokṣamukhas), while the
dhyāna-block is analysed as sāsrava- or sādhāraṇa-dharma (together with the
five abhijñās). Furthermore, the tathāgatabalas etc. are called lokottara-
kuśaladharma. Also, the bodhipakṣya-dharmas and everything up to the
tathāgatabalas are dharmas of a śrāvaka and meant to be practiced, while the
tathāgatabalas up to the āveṇikabuddhadharmas are dharmas which are to be
29 References for the LPG parts edited by Conze: ed. Conze, 1962: 57 (fol. 229b), 142f. (fol.
251a–b), 162 (fol. 255b), 180 (fol. 260a), 185f. (fol. 261b); ed. Conze, 1974: 11 (fol. 268b), 24
(fol. 273a), 29 (fol. 274b), 46 (fol. 279a–b), 80 (fol. 290a), 126 (fol. 305a). Four times the list
does not begin with dhyāna etc.; thrice these terms are missing, once they are inserted
after the 37 bodhipakṣya-dharmas (ed. Conze, 1974: 29, fol. 274b). The mahāmaitrī etc. can
be placed after the āveṇikadharmas or before it, but they are – with one exception –
always included.
30 The text references for the PvsP are too numerous to list here, one example is ed. Kimura,
1990 [IV]: 13. In some instances, near the beginning and the end of the whole text, also the
“old” sequence is given, beginning with dhyāna etc. (e.g. ed. Kimura, 2007 [I-1]: 149, and
ed. Kimura, 1992 [V]: 151). The same with the Śatasāhasrikā (ed. Kimura, 2009–2010, see
e.g. Kimura, 2010 [II-3]: 39). In contrast to the LPG and the PvsP, the list in the Śatasāhasrikā
is stable. Cf. also Advayaśatikā, ed. Shakya, 1988: 82–84.
possessed by a tathāgata or by which a tathāgata is distinguished (cf. PvsP, ed.

Kimura, 1990 [IV]: 58: […] ebhiś ca subhūte dharmais tathāgata iti prabhāvyate).31
It is noteworthy that all groups of four occurring in these Prajñapāramitā-
related-texts are enumerated within the first ten items of the BajC2 list. With
the exception of the four ārūpyasamāpattis, there is no group of four left that
is not represented in the Gāndhārī manuscript. This could be a further argu-
ment in reconstructing this group as item no. 7 in BajC2. However, the
Arthaviniścaya contains a very similar list (named the dharmaparyāya)32,
including additionally the four samādhibhāvanās, which could be another
option for the reconstruction, even though the traces of the first akṣara in
BajC2 do not suggest a reading of sa either. Another supporting fact is that the
P Saṃgītisuttanta inserts this term between the jhānas and the appamaññas,
thus approximately at the same position (the Sanskrit version places it very
late at position 33., T.26.1536 at position 23 (xiuding 修定); T.1.12 at position 21
(sanmodi xiang 三摩地想); T.1.1 and G Cm do not include it).
To conclude, as is often the case in studies of Gāndhārī manuscripts, the
intertextual relation to other Buddhist texts in Pāli, Sanskrit or Chinese is
not an easy one. What is common to all of them are the first three items
(smṛtyupasthāna, samyakpradhāna, ṛddhipāda). This is the beginning of the
bodhipakṣya-dharmas, a list that precedes the first schism (Willemen, Dessein
and Cox, 1998: 11). This basic list was expanded by terms related to meditation
(dhyāna, apramāṇa), which served as a basis for Abhidharma texts such as the
Vibhaṅga and the Dharmaskandha, both of which supposedly go back to a
common source that predates the splitting of the two schools (Theravāda and
Sarvāstivāda). This source is therefore dated earlier than the mission under
Aśoka (Willemen, Dessein and Cox, 1998: 69). The beginning of the list of four-
fold groups got further expanded by the ārūpya(samāpatti)s in the
Saṃgītisūtra/-paryāya versions (extant in G Cm, T.1.1 (Dharmaguptaka), and
T.1.12). In some of the versions (Skt, T.26.1536 (Sarvāstivāda), and P (Theravāda))
further categories had been inserted before the ‘unlimited’. In this respect,
31 Cf. Migme Chodron, 2001: 1314f.

32 The 27 items (ed. Samtani, 1971: 2) are: 5 skandhāḥ, 5 upadānaskandhāḥ, 18 dhātavaḥ, 12
āyatanāni; 12 pratītyasamutpādaḥ; 4 āryasatyāni, 22 indriyāṇi, 4 dhyānāni, 4 ārūpya-
samāpattayaḥ, 4 brahmavihārāḥ, 4 pratipadaḥ, 4 samādhibhāvanāḥ; 4 smṛtyupasthā
nāni, 4 samyakprahāṇāni, 4 ṛddhipādāḥ, 5 indriyāṇi, 5 balāni, 7 …, 8 …; 16 …, 4 srota
āpattyaṅgāni, 10 tathāgatabalāni, 4 vaiśāradyāni, 4 pratisaṃvidaḥ, 18 …, 32 …, 80 …
(parallel terms are set in bold, possible candidates for reconstruction are set roman).
Similarly, the list can be subdivided into semantic modules (cf. Samtani, 2002: xx): First,
four traditional lists of terms, then the pratītyasamutpāda, then satya and meditation, the
bodhipakṣya-dharmas, and finally characterizations of a tathāgata.
BajC2 seems to agree more closely with the Sarvāstivāda and Theravāda ver-
sions than the Dharmaguptaka. Between these two, the enumeration in BajC2
shares some features with the Pāli (Theravāda) on account of the sequence. At
the same time, however, there are exceptions, where it fits better to the Skt ver-
sion and T.26.1536 (Sarvāstivāda), and it also contains a category (G. saca) not
extant in the Pāli version. A special connection is given to T.1.1, T.1.12 and G Cm
due to the term (*pa)ḍ̱isabhida (T.1.1: biancai 辯才, T.1.12: wu aijie 無礙解, G
Cm: paḍisaṃbiḏa), which only occurs in those versions, although at later posi-
tions. Among all versions, the list in BajC2 is perhaps most similar to T.1.12 in
representing an intermediate state between the Theravāda / Sarvāstivāda and
the Dharmaguptaka versions. A link to Prajñāpāramitā texts is indicated by the
four veharajas (Skt. vaiśāradya), a term that is not known from the Saṃgītisūtra.
Also the va[ś]idas (Skt. vaśitā) point to an early Mahāyāna affiliated context.
The groups that have been mentioned so far are represented as the first ten
items of the list in BajC2, as far as they are characterized as being fourfold. The
subsequent ten items are a seemingly random selection of fourfold groups, also
known from the Saṃgītisūtra. The last of the altogether twenty-one items are
the asaṃhārya-dharmas, a term peculiar to the Aṣṭasāhasrikā, which will be
discussed later. The modular composition of the list is summarized in table 5.
2.3.2 Explanation of the groups and its items
The several items of each group are:33
1. (*catvari ś̱paḏovaṭhaṇa): smṛtyupasthānāni; satipaṭṭhānā; establish-

ments of mindfulness.
1. body (kāye kāyānupaśyana◦),
2. feeling (vedanāyāṃ …),
3. mind (citte …),
4. mind-objects (dharmeṣu …).
2. (*catvari) samapra[s̱aṇa]: samyakprahāṇāni (= samyakpradhānāni);

sammappadhānā; right endeavours.
1. for the abandoning of unwholesome mental states that have arisen
33 The sequence of terms under each point is: »G: Skt; P; E«. Unless otherwise stated, the
citations are taken from the reconstructed Sanskrit version given in the edition of Stache-
Rosen, 1968. If there are significant differences to the other versions in Pāli or Chinese this
is noted.
Table 5 Modular composition of the list in BajC2.
01. (*catvari ś̱paḏovaṭ́haṇa ·) fourfold groups of the śrāvakadharmas,

02. (*catvari) samapra[s̱aṇa] · bodhipakṣya-dharmas (anāsravā / to be practised
03. [catvari] irdh[ipada] · saṃskṛtā / asādhāraṇā dharmāḥ)
04. catvari jaṇa · fourfold groups related to meditation

05. catvari saca · (sāsravā / asaṃskṛtā / sādhāraṇā
06. catvari apramaña · dharmāḥ)
+ satyas
07. catvare ? + + + + +
08. [catvare] (*pa)ḍ̱isabhida · fourfold groups characterizing a buddhadharmas,

09. catvare va[ś]ida · tathāgata/buddha, related to be possessed
10. ca[tvare veharaja] · to Prajñāpāramitā or early
Mahāyāna texts (anāsravā /
asādhāraṇā / lokottarāḥ
kuśaladharmāḥ)
11. catvare ñaṇamulea dharma · fourfold groups, also occurring in the dharmas to be
12. catvare so[ḏavati](*aga ·) Saṃgītisūtra, apparently random known
13. + + + + + + ? ? + selection
14. catvare taṣ̄amulea dharma
(*·)
15. catvare paḍ̱i[va](*da ·)
16. + + + + + + + + + +
17. + + + + + + + ? ? ?
18. [catvare] ? ? ? ? + +
19. + + + + + + ? ? ?
20. [ca]tvare sakṣig̱araṇia
dharma
21. catvari as̱aharia dharma fourfold category, peculiar to the buddhadharmas

Aṣṭasāhasrikā, synonym to
awakening and a characteristic
of a tathāgata/buddha
(utpannānāṃ pāpakānām akuśalānāṃ dharmāṇāṃ prahāṇāya …),34

2. for the non-arising of unwholesome mental states that have not yet
arisen (anutpannānāṃ pāpakānām akuśalānāṃ dharmāṇāṃ
anutpādāya …),
3. for the arising of wholesome mental states that have not yet arisen
(anutpannānāṃ kuśalānāṃ dharmāṇāṃ utpādāya …),
4. for the stabilizing, increase, etc. of wholesome mental states that have
arisen (utpannānāṃ kuśalānāṃ dharmāṇāṃ sthitaye …).
3. [catvari] irdh[ipada]: ṛddhipādāḥ; iddhipādā; bases of [supernormal]

power.
1. … through will (chanda-◦),
2. … through energy (vīrya-◦),
3. … through mind (citta-◦),
4. … through investigation (mimāṃsā-◦ / P vimāṃsā-◦).
4. catvari jaṇa: dhyānāni; jhānāni; [stages of] meditation.

1. with initial thought and sustained contemplation, born from detach-
ment, experiencing joy and happiness (savitarkaṃ savicāraṃ vivekajaṃ
prītisukhaṃ),
2. without initial thought and sustained contemplation, born from
concentration, experiencing joy and happiness (avitarkam avicāraṃ
samādhijaṃ prītisukhaṃ),
3. characterized by equanimity and mindfulness, not experiencing joy but
happiness (upekṣakaḥ smṛtimān sukhaṃ viharatīti niṣprītikaṃ),
4. characterized by being purified due to equanimity and mindfulness,
experiencing neither pain nor happiness (aduḥkhāsukham
upekṣāsmṛtipariśuddhaṃ).
5. catvari saca: āryasatyāni; ariyasaccāni; [noble] truths.

1. suffering (duḥkha),
2. the origin of suffering (duḥkha-samudaya),
3. the cessation of suffering (duḥkha-nirodha),
4. the path that leads to the cessation of suffering (duḥkha-nirodha-gāminī
pratipad).
34 P interchanges (1.) and (2.).

6. catvari apramaña: apramāṇāni; appamaññāyo; unlimited.35

1. loving-kindness (maitrī),
2. compassion (karuṇā),
3. sympathetic joy (muditā),
4. equanimity (upekṣā).
7. catvare ? + + + + +
Uncertain. Most probably either the ārūpyasamāpattis36 or the samādhibhā
vanās.37
8. [catvare] (*pa)ḍ̱isabhida: pratisaṃvidaḥ; paṭisambhidā; analytical

knowledges.38
1. … of the meaning of things (artha°),
2. … of the condition of things (dharma°),
3. … of their linguistic explanation (nirukti°),
4. … of eloquence/perspicuity (pratibhāna°).
This category is only extant in G Cm, T.1.1, and T.1.12:

- G Cm: catvari paḍisaṃbiḏa: atha, dhaṃma, niruti, parivhaṇa;
- T.1.1: 謂四辯才。法辯義辯詞辯應辯。 ~ dhamma, attha, nirutti, paṭibhāna;
- T.1.12: 復次四無礙解。是佛所說。謂義無礙解。法無礙解。樂說無礙解。
辯才無礙解。 ~ artha, dharma, nirukti, pratibhāna.
In G Cm, paḍisaṃbiḏa goes back to √vid,39 and also in T.1.1, biancai 辯才 (“tal-
ent for debating”) is connected with √vid rather than with √bhid. T.1.12
35 In T.1.1 these are called fantang 梵堂 (~ brahmavihāra).

36 Skt IV.8, T.26.1536 IV.8, P IV.7, T.1.12 IV.6, T.1.1 IV.16, G Cm IV.16.
37 Skt IV.33, T.26.1536 IV.23, P IV.5, T.1.12 IV.21. The four ‘concentrative meditations’ are char-
acterized by (1) leading to happiness in the present life (dṛṣṭadharmasukhavihārāya), (2)
obtaining knowledge-and-vision (jñānadarśanapratilābhāya), (3) analysis through
understanding (prajñāprabhedāya) / P mindfulness and clear awareness (satisampa
jaññāya), and (4) the destruction of [all] defilements (āsravakṣayāya).
38 Cf. e.g. the explanations in Aung and Davids, 1915: 377–381 (related to the Kathāvatthu and
the Vibhaṅga); Ñāṇamoli, 2011: 436 (related to the Visuddhimagga); de La Vallée Poussin /
Pruden, 1988–1990: 1151ff. (related to the Abhidharmakośabhāṣya); or, for a Mahāyāna con-
text, see e.g. Migme Chodron, 2001: 1322ff. (related to the MPPŚ, with a summary of rele-
vant references); Apple, 2009: 164–165; or Brunnhölzl, 2010: 659 (chart 12): ‘of dharmas
(knowing the individual characteristics of all phenomena), meanings (knowing the clas-
sifications of all phenomena), semantics (knowing the languages, terms, etc.), self-confi-
dence (hearing and explaining the dharma without doubts)’.
39 It is explained by ki pa[ḍisaṃ](*biḏa) [39] ? as̱a va paḍivijaṇati (preliminary unpublished
transliteration), thus giving the synonym √jan for √vid.
combines both possibilities by writing wu aijie 無礙解 (“unobstructed under-

standing/knowledge”). By using (*pa)ḍ̱isabhida, BajC2 seems to be closer to
the P tradition.40
9. catvare va[ś]ida: vaśitā; –; masteries.

It is uncertain what exactly the four vaśitās refer to (see earlier discussion).
Looking at the Saṃgītisūtra, the vaśitās could be a synonym of the four balas
(missing only in G Cm and T.1.1).41 Since other groups of ten vaśitās and ten
balas exist and do not overlap, and since the Abhisamayālaṃkāra lists both
terms side by side (see above), it seems unlikely that they refer to the same
group here. Therefore, the four vaśitās mentioned in the Bajaur sūtra can per-
haps be related to the ones in the MAV(Bh) and the Msa discussed above:
nirvikalpa-/avikalpa-, kṣetra(pariśuddhi)-, jñāna-, karma-vaśitā (“mastery in
the absence of conceptual discrimination, in the purification of a field, in
awareness, and in action”).
10. ca[tvare veharaja]: vaiśāradyāḥ; vesārajjā; self-confidences.42

1. regarding supreme awakening (abhisaṃbodhi),
2. … destruction of [all] defilements (āsravakṣaya),
3. … [understanding of all] obstructing factors (antarāyikadharma),
4. … [knowing and teaching the correct] way to salvation
(nairyāṇikapratipada).
11. catvare ñaṇamulea dharma: ~ jñānāni; ~ ñāṇāni; things rooted in

knowledge.43
1. [true] doctrine (dharma),
2. the following (anvaya),44
3. other’s mind (paracitta),
4. common knowledge (saṃvṛti).
40 Cf. PTSD s.v. paṭisambhidā: “BSk. pratisaṃvid is a new formation resting on confusion
between bhid & vid”.
41 Skt IV.15 śraddhā, vīrya, samādhi, prajñā; P IV.26 sati, viriya, … There are also four other
balas “leading to a Bodhisattva’s cittotpāda, Bbh 13.22, listed 17.8–9 as adhyātma-, para-,
hetu-, prayoga-bala” (BHSD s.v. bala).
42 Source of explanation: Abhidharmasamuccaya (Abhidh-s 98).
43 Due to the different terminology it is not entirely certain that the Gāndhārī refers to the
four jñānas as given in the Saṃgītisūtra. Another secondary explanation of the four
knowledges relates to the four truths (Skt IV.14: duḥkha, samudaya, nirodha, mārga; also P
IV.12). G Cm seems to mix the two alternative explanations.
44 Cf. the explanation in T.26.1536 (Stache-Rosen, 1968: 100).
12. catvare so[ḏavati](*aga): śrotāpattyaṅgāni; srotāpattiyaṅgāni; factors of

[the state of] stream entry.45
1. being endowed with perfect / unbreakable46 faith in the Buddha
(buddhe ’vetya- prasādena saṃpannaḥ),
2. … in the Dharma (dharme …),
3. … in the Saṃgha (saṃghe …),
4. … in the ethics estimated by the noble ones (āryakāntaiḥ śīlaiḥ
saṃpannaḥ).
13. + + + + + + ? ? +
Uncertain.
14. catvare taṣ̄amulea dharma: ~ tṛṣṇotpādāḥ; ~ taṇhuppādā; things rooted

in craving.47
1. raiment (cīvara),
2. alms-food (piṇḍapāta),
3. lodging (śayanāsa),
4. existence or non-existence (bhavavibhava)48.
45 There are two different explanations. The above-mentioned one seems more likely here,
since the four avetyaprasādas / abhedyaprasādas (G. abhejapras̱ada) are referred to later
in the text, although this is no conclusive evidence. Most versions list both explanations,
the first is named Skt. srotāpattyaṅgaḥ / T.26.1536 yuliuzhi 預流支 / P sotāpattiyaṅgāni /
T.1.12 – / T.1.1 –; the second one is named Skt. caturbhir aṅgaiḥ samanvāgataḥ srotāpannaḥ
/ T.26.1536 zhengjing 證淨 / P sotāpannassa aṅgāni / T.1.12 yuliushen 預流身 / T.1.1 xutuo-
huanzhi 須陀洹支. G Cm calls both soḏavatiaga, and says that the second explanation
is favored “here” (G. iśa). The alternative (and probably older) explanation would be:
1. associating with good people (satpuruṣasaṃseva); 2. listening to the good doctrine
(saddharmaśravaṇa); 3. investigating it thoroughly (yoniśo manasikāra); correct behav-
iour according to the doctrine (dharmānudharmapratipatti).
46 Skt avetya◦ / T.26.1536 zhengjing 證淨 / P avecca◦ / G Cm aveca◦ / T.1.12 buhuai 不壞 /
T.1.1 buhuaixin 無壞信. In the subsequent text of BajC2, the term is spelled abhejopras̱ada
(= abhedyaprasāda). In the Saṃgītisūtra preserved in the Gilgit Dīrghāgama manuscript,
the form avetya◦ is used (thanks to Jens-Uwe Hartmann for the information on this
unpublished manuscript portion). For a more comprehensive discussion of the term
abhedyaprasāda cf. § 3 of this article.
47 Similar to the four jñānamūlaka-dharmas (11.), it is uncertain if the tṛṣṇāmūlaka-dharmas
equate to the tṛṣṇotpādas in the Saṃgītisūtra at all, since they denote things that have
tṛṣṇā as a cause and not as a result. There are however no four tṛṣṇāmūlaka-dharmas, but
only nine (taṇhāmūlaka-dhamma) in the Aṅguttaranikāya (AN IV 400–401, cf. DN II
58–61), the Paṭisambhidāmagga (Ps 130), and the Vibhaṅga (Vibh 390).
48 Mss. (Hoernle, Hs. 47/48): bhavatibhava◦ (cf. Stache-Rosen, 1968: 79, note 133); bhavā
bhava, for which Thomas W. Rhys Davids gives “dainty foods” like “oil, honey, ghee, etc.”
15. catvare paḍ̱i[va](*da): pratipadaḥ; paṭipadā; kinds of progress.49

1. painful progress with slow comprehension (duḥkhā pratipad
dhandhābhijñā),
2. painful progress with quick comprehension (duḥkhā pratipad
kṣiprābhijñā),
3. pleasant progress with slow comprehension (sukhā pratipad
dhandhābhijñā),
4. pleasant progress with quick comprehension (sukhā pratipad
kṣiprābhijñā).
16. + + + + + + + + + +
Uncertain.
17. + + + + + + + ? ? ?
Uncertain.
18. [catvare] ? ? ? ? + + (= viñaṇaṭ́hiḏio ?); vijñānasthitayaḥ; viññāṇaṭṭhitiyo;

stations of consciousness.
1. being directed to form (rūpopagā50),
2. … to feeling (vedanopagā),
3. … to perception (saṃjñopagā),
4. … to volition (saṃskāropagā).
19. + + + + + + ? ? ?
Uncertain.
according to the commentary of Buddhaghosa. T.26.1536 explains bhava as the five skan-
dhas, thus ‘existence’.
49 The translation is taken from Walshe, 1995: 492 (P). The Sanskrit and Pāli versions also
contain another explanation, that is however not contained in the Chinese versions,
which is why the given explanation has been preferred. The alternative would be (Skt
IV.32, P IV.22): 1. inability to endure (akṣamā), 2. taming / self-control (damā), 3. ability to
endure (kṣamā), 4. appeasement (chamā).
50 P has ◦upāya instead of ◦upaga; G Cm has ruovao / veḏaṇ[o]vao / saṃñ[o]ao /
saṃkharovao (preliminary unpublished transliteration), which can be both.
20. [ca]tvare sakṣig̱araṇia dharma: sākṣīkaraṇīyā dharmāḥ; sacchikaraṇīyā

dhammā; things to be realized.51
1. by the body (kāyena), i.e. the eight deliverances (P vimo(k)kha)52,
2. by mindfulness (smṛtyā), i.e. former lives (P pubbenivāsa),
3. by the [heavenly] eye (cakṣuṣā), i.e. decease and rebirth (P
cutūpapāta)53,
4. by understanding (prajñayā), i.e. destruction of intoxicants (P āsavānaṃ
khaya).
21. catvari as̱aharia dharma: asaṃhāryā dharmāḥ; asaṃhāriyā dharmā;

unconquerable/insuperable things.
This term has no parallel in the Saṃgītisūtra. The asaṃhārya-dharmas are
mentioned in the Aṣṭasāhasrikā, where they are associated with the attain-
ment of the highest form of awakening:
May that thought of enlightenment which they have wished for, thought
over and taken hold of, bring to fulfillment in them the dharmas of a
Buddha, and dharmas associated with all-knowledge, the dharmas of the
Self-Existent, the insuperable dharmas [asaṃhārya-dharmāṇāṃ]! (tr.
Conze, 1973; xxvi 434, ed. Vaidya: 215).
They signify a state of mind or knowledge, by which a bodhisattva becomes

irreversible (avinivartanīya) and can no longer be overcome by disciples or
pratyekabuddhas (xviii 341, ed. Vaidya: 170; cf. xxii 401, ed. Vaidya: 199) or others
(xx 380, ed. Vaidya: 188), especially not Māra (xvii 329, ed. Vaidya: 164; xvii 332,
ed. Vaidya: 165; xvii 337, ed. Vaidya: 168):
An Arhat, a monk whose outflows are dried up, does not go by someone
else whom he puts his trust in, but he has placed the nature of dharma
directly before his own eyes, and Mara has no access to him [asaṃhāryo
bhavati māreṇa]. Just so an irreversible [avinivartanīyo] Bodhisattva
cannot be crushed by persons who belong to the vehicle of the Disciples
and Pratyekabuddhas, he cannot, by his very nature, backslide into the
level of Disciples or Pratyekabuddhas, he is fixed on allknowledge, and
51 The sequence varies: Skt 1–2–3–4; P 2–3–1–4; T.1.1 3–1–2–4 (cp. Behrsing, 1930: 75–76 note
169); G Cm 3–2–1–4.
52 T.1.1 and G Cm have here “cessation [of perception and feeling]” (G. ṇiros̱o; T.1.1 shenshou
mie zheng 身受滅證 ~ P vedayita-nirodha…).
53 T.1.1 and G Cm have here “forms” (~ rūpa).
ends up in perfect enlightenment. It is quite certain that a Bodhisattva

who stands firmly in the element of irreversibility cannot possibly be led
astray by others (tr. Conze, 1973; xvii 329, ed. Vaidya: 164).
This state of irreversibility is closely connected with the realisation of empti-

ness (ix 205, ed. Vaidya: 102) and the perception of all elements as a dream
(svapnopamāḥ sarvadharmā iti, xx 380, ed. Vaidya: 188). As a further example,
another passage reads:
He can no longer be led astray by others, and on the stage which is his by
right he cannot be crushed. For, as he has stood firm on it, his mind
becomes insuperable, his cognition becomes insuperable (aparapraṇeyo
bhavati, anavamardanīyaś ca bhavati svasyāṃ bhūmau / tatkasya hetoḥ?
tathā hi sa sthito ’saṃhāryeṇa cittena asaṃhāryeṇa jñānena samanvāgato
bhavati, tr. Conze, 1973; xvii 337, ed. Vaidya: 168).
Also, a passage in the Larger Prajñāpāramitā from Gilgit (fol. 253b) circum-
scribes the practice of the perfection of wisdom as a state, in which
[one] cannot be overpowered by Mara or the deities of his host, or by the

persons who belong to the vehicle of the Disciples and Pratyekabuddhas,
nor can this perfection of wisdom of the Bodhisattva, the great being, be
taken away by any heretics or bad spiritual friends. And why? Because all
these cannot be apprehended in this perfection of wisdom, on account of
the emptiness of own-marks (tr. Conze, 1975: 521, asaṃhāryā mārair vā
mārakāyikābhir devatābhiḥ śrāvakapratyekabuddhayānikair vā pudgalair
yāvan na kaiścid anyatīrthikaiḥ pāpamitrair iyaṃ prajñāpāramitā śakyam
ācchetuṃ bodhisattvasya mahāsattvasya. tat kasya hetos? tathā hi te sarve
’tra prajñāpāramitāyāṃ nopalabhyante svalakṣaṇaśūnyatām upādāya,
ed. Conze, 1962: 152).54
In other words, the list given in BajC2 culminates in the asaṃhārya-dharmas,

which are synonym to the perfect awakening of a tathāgata, and represent his
all-encompassing knowledge. When one is endowed with the asaṃhārya-
dharmas, one becomes unconquerable. Thus, the succeeding passage (BajC2,
1E.28–32) states:
54 Similarly LPG fol. 247a (tr. Conze, 1975: 479f.).

[ye]hi caduhi ⸨as̱a⟨*ha⟩r[i]ehi⸩ dhamehi samuṇag̱ada b[o]s[o] ṇa

sahariadi ◊
añatithiecarag̱aparivayag̱ehi ṇiaṭ́hapariva(*yag)e[h]‍(*i ṇa
sa)[hariadi]
caduraghimaras̱e[ṇa](*e) ◊ ṇa sahariadi ◊
+ + [dehi] ◊ ṇa sahariaṃti
duhavedaṇehi ◊ ṇa sahariati
adukham asuehi ? ? ? ?55 [ṇa sa](*hari)[a]di
triṭhi ṭ́haṇehi ◊ ṇa sahariati
+ ? ? ṇa [sahariadi]
aṇuśea ṭ́haṇ[ehi] ◊ ṇa pa[ḍ̱i]śe-ṭ́ha[ṇe]hi ◊ [ṇa] sahariati ◊
sa[s̱a]ve[hi] ◊ puṇa bhaviehi ◊ kudha56dhaduaïdaṇehi + + + + + + + +57
(*ṇa saha)‍[riati] ◊
yava sarva bosa-pa[kṣia] dha(*rma)
sarva sa[kil](*eśa) [pa]kṣia dha(*r)[ma] ◊
sarva vodaṇa-pakṣia dharma [va
sarva] + + + + + + + + + + + ? [s̱]i ·
[kas̱a] deśati ? ? ? ? dha[rma]58 v[i]d[i]da · ◊ pruṭho me [sa]martho ·
An awakened one59 who is endowed with these four unconquerable

things (asaṃhārya-dharma) is not conquered.
55 Reconstruct vedaṇehi ?
56 Clearly written ku, but perhaps kaṃ was intended, like in other Gāndhārī manuscripts
(next to kadha or ḱadha).
57 Maybe pratītyasamutpāda is to be inserted here, as it follows after skandhadhātvāyatana
(and precedes the bodhipakṣya-dharma) in lists in the Larger Prajñāpāramitā (LPG). Also,
those lists are concluded and analysed by terms like saṃkleśa and vyavadāna, just as in
BajC2 a few words later.
58 Maybe tasagadadharma (Skt. tathāgatadharma) is to be reconstructed.
59 G. b[o]s[o]: Skt. bodho (?). Skt. buddho is excluded, since it should be written budho or
bodho. Nevertheless, since a translation as “awakening” seems rather unlikely due to the
associated verbal forms (G. samuṇag̱ada = samanvāgata and sahariadi = saṃharīyate), it
appears to be an unusual bahuvrīhi (“possessing awakening”). Alternatively, the transla-
tion would be “A state of awakening, which is endowed with these four unconquerable
things is not conquered”.
[He] is not conquered by adherents of other sects like the caraka mendi-
cants60 or the nirgrantha mendicants61,
[He] is not conquered by the fourfold army of Māra62,
[He] is not conquered by …,
[He] is not conquered by feelings of suffering,
[He] is not conquered by (feelings ?) [such as] non-suffering [or]
non-happiness,
[He] is not conquered by states of [wrong] views,
[He] is not conquered by …,
[He] is not conquered by states of propensity [or] by states of aversion
(?)63,
[He] is not conquered by defiled (sāsrava) aggregates, elements, [or]
sensory bases (skandhadhātvāyatana) leading to rebirth (punarbhavika)
[…]
[etc.] up to all characteristics (dharma) associated with awakening
(bodhipākṣika),
all characteristics associated with defilement (saṃkleśapākṣika),
or all characteristics associated with purification (vyavadānapākṣika),
all …
60 G. añatithiecarag̱aparivayag̱ehi: Skt. anyatīrthikacarakaparivrājakaiḥ. It remains un-

clear whether carakaparivrājaka refers to a specific religious group or to non-settled
mendicants in general (cf. e.g. BHSD s.v. caraka, SWTF s.v. nānā-tīrthya-śramaṇa-
brāhmaṇa-caraka-parivrājaka). A contextually similar passage can be found in the
Suvikrāntavikrāmiparipṛcchā (ed. Vaidya: 56) or the PvsP (ed. Kimura, 1990: 149), where
the bodhisattva also cannot be overcome by Māra and his assembly nor by non-Bud-
dhist mendicants (Suvikrānta◦: anyatīrthika, carakaparivrājaka, PvsP anyatīrthika,
parivrājaka), because he courses in the perfection of wisdom, i.e. he does not perceive
any dharma (na kaṃcid dharmaṃ samanupaśyati).
61 G. ṇiaṭ́hapariva(*yag)e[h](*i): Skt. nirgranthaparivrājakaiḥ, usually referring to Jainas.
62 Cf. e.g. Dhsgr 80: catvāro mārāḥ / tadyathā // skaṃdhamāraḥ kleśamāro devaputramāro
mṛtyumāraś ceti. Interestingly, in BajC2 there is no specific mention of the bodhisattva
being insuperable in regard to disciples or pratyekabuddhas (which is the case in the
preserved Sanskrit versions of the Aṣṭasāhasrikā and LPG), but only in regard to Māra and
non-Buddhists. BajC2 thus represents a stage of development, where the opposition to
śrāvakas has not been established yet. An observation that holds true throughout the text.
63 G. aṇuśeaṭ́haṇ[ehi] ◊ ṇa pa[ḍ̱i]śeṭ́ha[ṇe]hi ◊ [ṇa] sahariati. These terms were not found in
other texts. They could refer to anuśaya◦, “propensity,” and – based on that – perhaps to
*pratiśaya◦ in the meaning of pratigha◦, “aversion” (possibly G. pa[ḍ̱i]śe◦ can also
directly be derived from pratigha◦ with ś < h < gh in analogy to the development h < ś in
veharaja < vaiśāradya). The position of the first ṇa is grammatically odd and it probably
has to be elided.
Why has it been shown? [Now] the characteristics (dharma) (of a tathā
gata ?) are known. Having been asked, I have answered adequately
(samartha).
2.3.3 Meaning of the list(s)

The list in BajC2 summarizes the characteristics or constituents (dharma) of
awakening explained by the bhagavant after having been asked about it by
Śāriputra. In the Saṃgītisūtra, similar groups are listed as items that should be
known as the Dharma and Vinaya of the Tathāgata, and the whole text is said
to have been recited in order to memorize the teaching that leads to awaken-
ing. Generally, such lists function as “succinct compendia of the Dhamma”
(Gethin, 1992: 157).
More important to BajC2, however, are the Prajñāpāramitā texts, as they are
likewise dealing with the concept of emptiness in general. In the Aṣṭasāhasrikā
(ed. Vaidya: 97), the list is used within an enumeration of things that a bodhi-
sattva should not be attached to in order to course in the perfection of
understanding (prajñāpāramitā).64 In the Larger Prajñāpāramitā, the listings
are summarized as the “gift of the Dharma” (dharmadāna, cf. e.g. LPG fol. 279a-
b, ed. Conze, 1974: 46, VIII 5,2) or more often as “wholesome dharmas”
(kuśala-dharma, e.g. LPG fol. 20a, ed. Zacchetti, 2005: 214; kuśalā bodhipakṣā
dharmāḥ, PvsP, ed. Kimura, 2009 [I-2]: 136) that are conducive to awakening
and that constitute the path of a bodhisattva to reach omniscience (sarvajñatā)
(e.g. PvsP, ed. Kimura, 2007 [I-1]: 171; 2009 [I-2]: 115 or 171; 1986 [II–III]: 71 or 168;
2006 [VI–VIII]: 119).
The list in the Bajaur Mahāyāna sūtra pursues the same purpose: It is used
to describe the state of awakening, either by the qualities that are part of this
state or that lead thereto. The main difference between the Bajaur list and the
lists in Prajñāpāramitā texts is that the latter also include groups of five or
seven or more. The restriction to fourfold groups in BajC2 may be explained by
the passage that follows them, where the abhedyaprasādas are discussed at
length.65 Thus, it appears that the list of groups of four is merely a rhetorical
64 […] evaṃ saptatriṃśad bodhipakṣā dharmā balāni vaiśāradyāni pratisaṃvido aṣṭādaśā

veṇikā buddhadharmāḥ sasaṅgāsaṅgā iti na carati, carati prajñāpāramitāyām […] (viii
194, cf. tr. Conze, 1973: 146).
65 As the next part of this article will show, it is a peculiar feature of Sarvāstivāda traditions
that they consist of four (and not three) items. In this context, it is also worthy to note that
the four noble truths apparently had been a prevalent organizational feature in Abhi
dharma texts of the northwest, as has recently been indicated by Collett Cox (2014: 38f.).
Thus, the number four might have been important or at least popular in this region, or
even specifically among the Sarvāstivādin.
device to introduce the four abhedyaprasādas, while at the same time illustrat-
ing the author’s knowledge of certain lists and categories, of which he
enumerated all those fourfold ones that came to his mind in order to represent
the Dharma.
3 The Four abhedyaprasādas
The discourse about the groups of four culminates in a long exposition about
the abhedyaprasāda “unbreakable confidence/trust”. Despite its obviously dif-
ferent etymology this term has to be related to its Pāli equivalent aveccapasāda
or its Sanskrit representative avetyaprasāda which are usually translated as
“perfect confidence/trust/faith” or “confidence/trust/faith based on under-
standing,” respectively. As in our text, in canonical literature, these terms
describe one of the characteristic features of an āryaśrāvaka or srotāpanna
“stream-enterer”. But not only the altered etymology distinguishes our text
from these parallels, the inclusion of the abhedyaprasādas/avetyaprasādas
among the group of four is similarly remarkable.
Based on the amount of text devoted by the Bajaur Mahāyāna sūtra to this
issue, the abhedyaprasādas played a key role in the concept of an āryaśrāvaka.
In order to determine the specific role of the abhedyaprasādas in the Bajaur
Mahāyāna sūtra, our exposition will focus on three major points:
3.1 From aveccapasāda to abhedyaprasāda: shifting etymologies
3.2 The abhedyaprasādas as a group of four
3.3 The reinterpretation of this category in the context of the Bajaur
Mahāyāna sūtra
3.1 From aveccapasāda to abhedyaprasāda: Shifting Etymologies

The etymology of the term in Gāndhārī seems to be quite clear: abheja has to
be derived from Old Indian abhedya “unbreakable”. This is also supported by
the text’s own explanation:
yado ṇa samaṇupaśati tado ṇa bhijati ta vucati abhejeṇa pras̱adeṇa

samuṇag̱ada
[And] because he does not perceive [anything], he is not broken.
[Therefore] it is said: ‘[he is] endowed with unbreakable confidence’.
In Pāli texts, the same term regularly occurs as aveccapasāda. The etymology of
the first member of this compound is doubtful. Modern Pāli dictionaries,
including the Critical Pāli Dictionary, derive avecca from the root ava-(ā)-√i “to
understand, to know”. This explanation appears to be based on two aspects.66
First, some of the Pāli commentaries paraphrase the initial avecca with
words meaning “to understand, to know”. Thus, Buddhaghosa’s commentaries
explain avecca by paññāya ajjhogahetvā, paṭivijjhitvā, ñatvā or jānitvā (cf. CPD
s.v. avecca). Secondly, an etymology based on Skt. ava-(ā)-√i is also indicated
by the Sanskrit variant of this term, which is usually given as avetyaprasāda (cf.
BHSD s.v.). Consequently, modern studies on the Buddhist concepts of faith
and belief characterize this scholastic category as “confidence/trust/faith
based on understanding”. Thus, Rupert Gethin writes:
There is some reason for thinking that pasāda is often thought of as

denoting a more refined and developed stage of saddhā; it is used espe-
cially in contexts where this seems appropriate. In this case pasāda is
especially aveccapasāda, that is full-trust, trust that results from a certain
degree of understanding (Gethin, 2001: 113, my emphasis).
In his monograph on the “Early Buddhist Theory of Knowledge,” Kulatissa

Nanada Jayatilleke also underlines the fact that the term prasāda is specifically
associated with the act of intellectual understanding (1963: 386). Consequently,
he translates aveccappasāda as “faith born of understanding”. As Jayatilleke
points out, the Pāli commentators give sometimes a different explanation
for the initial word avecca, using attributes like acala “immovable” or acyuta
“firm, solid”. Thus, Buddhaghosa paraphrases aveccappasāda repeatedly by
acalappasāda (see CPD s.v.), not regarding this as contradictory to his alterna-
tive explanation. This is, for instance, indicated by his commentary on DN II
93,27 (Sv (II) 544,22): buddhaguṇānaṁ yathābhūtato ñātattā acalena accutena
pasādena. Here, he combines both possible meanings, “understanding” and
“immovable,” by explicitly deriving the immovable, solid character of pasāda
from the true knowledge of the qualities of the Buddha (yathābhūtato ñātattā).
Based on this alternative explanation given by Buddhaghosa, many modern
translators prefer the connotation “unwavering” for avecca.
It seems that at a certain point, the origin and background of this term
became obscure.67 This uncertainty probably paved the way for different expla-
66 Cf. also the detailed note by Samtani in his edition of the Arthaviniścaya (1971: 241).
67 During the discussion at the 1st Lausanne Gāndhārī Workshop in June 2013, Harry Falk
suggested an alternative etymology of the Pāli word avecca based on the root √vic “to sift,
separate”. In this case one would certainly have to distinguish the absolutive avecca used
in isolated position in a phrase and the gerundive avecca used as first member of a
nations and interpretations. One of these alternative interpretations, which is

semantically very close to Buddhaghosa’s acala or acyuta is represented by the
connotation abhedya “unbreakable” attested in our Gāndhārī text. But is this
occurrence the only instance for this variant of reinterpretation?
As far as I could ascertain, there are also some other, although very few, Skt.
texts which use this very variant. One of them is the Daśabhūmikasūtra. In its
description of the third bhūmi, the Brilliant One (arciṣmatī), the text repeatedly
refers to the abhedyaprasādas, e.g. triratnābhedya-prasāda-niṣṭhā-gama-na-
tayā (ed. Vaidya, 1967: 24; ed. Rahder, 1926: 38) “by certainty with regard to the
unbreakable confidences in the Three Jewels”.
Describing the ten ways by which the career of a bodhisattva (bodhisattva
caryā) is to be considered with regard to his invincibility (asaṃhāryatā), the
text lists one feature for each of the ten bhūmis. With regard to the arciṣmatī
bhūmi it says according to Vaidya’s edition (1967: 66):
arciṣmatyāṃ bodhisattvabhūmau buddhabhedyaprasādaikarasataḥ
There can be little doubt that the text has to be corrected into buddhābhe
dyaprasād°68 and can be translated as:
On the arciṣmatī bodhisattva level [he is invincible] because of the single

affection towards the unbreakable confidence in the Buddha.
The same form abhedya also occurs in the summarizing verses (upasaṃhāra
gāthā) devoted to the fourth bhūmi (ed. Vaidya, 1967: 79–80):
sahaprāptu arciṣmati bhūmi mahānubhāvaḥ saṃvṛttu śāstu kuli bhūyu

vivartiyatve/
abhedya buddharatane tatha dharmasaṃghe udayavyayasthiti nirīhaka
prekṣamāṇaḥ // 8 //
Immediately at reaching the arciṣmatī level, the powerful [bodhisattva]

becomes member of the Buddhas’ family – and does [not] return any-
more [from that status].
compound word. According to this explanation, the meaning of the Gāndhārī variant
abheja would nearly correspond to the original meaning of the term. In any case, the com-
mentaries of Buddhaghosa as well as the Sanskritized term avetya leave no doubt that the
etymological origin of this term was no longer understood.
68 This is in fact the reading given in the older edition by Rahder, 1926: 97.
Unbreakable with regard to the Buddha jewel and the Dharma and the
Saṃgha, he sees that things are inactive with regard to their production,
cessation, and stability.
A closely-related term in the Daśabhūmikasūtra is bodhisattvasyābhedyāśayatā,

i.e. “a bodhisattva’s unbreakable resolve” (bhūmi 6: ed. Vaidya, 1967: 34; ed.
Rahder, 1926: 53). According to Rahder’s glossary (1928: 18), Skt. abhedya is ren-
dered in all these instances in the Tibetan version as mi phyed pa, in Śīladharma’s
Chinese translation as buhuai 不壞.69
The same coherence between the Sanskrit version and the later translations
can be observed in the Vimalakīrtinirdeśa.70 The only preserved Sanskrit man-
uscript of this text conveys the attributes dṛḍhavajrādhyāśayābhedyabuddha-
dharmaprasādapratilabdhaiḥ, i.e. “who have obtained unbreakable confidence
in the Buddha and the Dharma by their diamond-firm resolve” (1.3, ed. Study
Group, 2006: 2), and buddhe ’bhedyaprasādaratiḥ (3.64, ed. Study Group, 2006:
40) as qualities of bodhisattvas. A synopsis of these two Sanskrit terms and
their translations in the Tibetan and Chinese versions71 yields the following
picture:
Skt dṛḍhavajrādhyāśayābhedyabuddha dhar- buddhe ’bhedyaprasādaratiḥ

maprasādapratilabdhaiḥ
Tib hag pa’i bsam pa rdo rje ltar sra bas sangs rgyas la mi phyed par dad
sangs rgyas dang | chos dang | dge ’dun cing dga’ ba dang
la mi phyed pa’i dad pa rnyed pa
T.14.474 有金剛志得佛聖性樂於喜不離佛
T.14.475 深信堅固猶若金剛樂常信佛
T.14.476 於諸佛法得不壞信流法苑樂者謂於諸佛不壞淨樂
It is obvious that both the Tibetan and the Chinese translation by Xuanzang 玄
奘 (T.14.476) confirm the reading of the Skt. version and use the already men-
tioned translations for abhedya (Tib. mi phyed pa, Chin. buhuai 不壞). The two
earlier Chinese translations are more difficult to evaluate. Whereas T.14.474 at
69 The glossary’s reference “(10M)” for one of the occurrences seems to refer to the Parīndanā
section, which corresponds in Rahder’s edition to ch. C.
70 The reference to the Vimalakīrtinirdeśa I owe to Dan Stuart.
71 The Tibetan translation from the Derge edition of the Kangyur and the Chinese transla-
tions T.14.474 by Zhiqian 支謙 (223–228 CE), T.14.475 by Kumārajīva 鳩摩羅什 (406 CE)
and T.14.476 by Xuanzang 玄奘 (650 CE) can be easily accessed in Jens Braarvig’s excel-
lent Thesaurus Literaturae Buddhicae (URL: <http://www2.hf.uio.no/polyglotta/index.
php?page=volume&vid=37>, accessed 2/11/2105).
least in the second case seems to render abhedya by buli 不離, the other trans-
lations leave the element abhedya untranslated or altered its meaning. Thus,
abhedyaprasāda is represented in T.14.475 either as changxin 常信 “eternal
faith” or as shenxin 深信 “profound faith.”72
Another text which uses this variant is the Sarvatathāgatādhiṣṭhāna
vyūhasūtra, the Sanskrit version of which is only partially preserved in two
manuscripts from Gilgit.73 The text describes the effects of a meditation practice
called sarvatathāgatādhiṣṭhāna-sattvāvalokana-buddhakṣetra-sandarśana-
vyūho nāma samādhiḥ. One of these effects is described as follows:
[…] smṛtimantaḥ prajñāvantaḥ buddhe dharme sa(ṃ)ghe (’)bhedya

prasādena samanvāgat(ā)ḥ […] (transliterated from Raghu Vira and
Lokesh Chandra, 1995, plates 1751–1752)74
[…] being mindful [and] knowledgeable, they [will] be endowed with
unbreakable confidence in the Buddha, the Dharma, the Saṃgha.
Again the reading of the Sanskrit manuscript is confirmed by the Tibetan

translation mi phyed pa (Dutt, 1984: 53, note 1).75 It is, however, interesting to
note that despite the use of abhedya, the text maintains the association of the
term with mindfulness and knowledge.
The three texts cited above consistently rendered this term as abhedya in
the manuscripts and in the corresponding Tibetan translations. It can there-
fore be assumed that their original versions or at least one or several of their
rather early recensions did show this variant.
There is some evidence that the variant abhedya could also replace an origi-
nal avetya/avecca. Such a case is probably represented by the Gilgit manuscript
of the so-called Larger Prajñāpāramitā76 which replaces the conventional
72 For this last variant cf. the Digital Dictionary of Buddhism, s.v. (<http://www.buddhism-
dict.net/>).
73 The facsimiles of both manuscripts are reprinted in Raghu Vira and Lokesh Chandra,
1995: plates 1746–1815 and plates 1816–1837. The passage, which mentions abhedyaprasāda
is found only in the former.
74 The text given by Dutt is not entirely correct: smṛtimantaḥ prajñāvantaḥ buddhe dhama
saṃghe abhedyaprasādena samanvāgatā (ed. Dutt, 1984: 53).
75 The late Chinese translations T.19.1022 by Amoghavajra (eighth century CE) and T.19.1023
by Dānapāla (tenth/eleventh century CE) cannot confirm this reading. They use dingxin
定信 “firm faith” (T.19.1022A: 710 a23) and shanxin 善信 “good faith” (T.19.1023: 715 a19)
instead.
76 For a detailed description of the Gilgit version of the Larger Prajñāpāramitā see Zacchetti,
2005: 19–28.
avetyaprasāda found in other manuscripts by the distinct reading abhedya77

(fol. 143 recto, lines 6–8, transliterated from Raghu Vira and Lokesh Chandra,
1995: plate 453):
bhagavā āha. tat kiṃ manyase kauśika kiyantah jāṃbūdvīpakā manuṣyā

ye buddhe abhedyaprasādena samanvāgatā//s te saṃghe abhedyapra
sādena samanvāgatā […]?
śakra āha. alpakās te bhagavaṃ jāṃbūdvīpakā manuṣyā ye buddhe
abhedyaprasādena samanvāgatāḥ//s te saṃghe abhedyaprasādena
samanvāgatā
The Blessed One said: What do you think, Kauśika, how many people of
Jambūdvīpa are endowed with unbreakable confidence in the Buddha,
[are endowed with unbreakable confidence in the Dharma,] are endowed
with unbreakable confidence in the Saṃgha. Śakra said: Few people of
Jambūdvīpa, Blessed One, are endowed with unbreakable confidence in
the Buddha, [are endowed with unbreakable confidence in the Dharma,]
are endowed with unbreakable confidence in the Saṃgha.
According to the majority of later manuscripts78 and the commentaries,

abhedya was most likely not the original reading of the Larger Prajñāpāramitā
text.
Within their commentaries ad Abhisamayālaṃkāra 2.18–19 on adhimukti
(cf. ed. Stcherbatsky and Obermiller, 1929: 13; tr. Conze, 1954: 37), both Ārya
Vimuktasena and Haribhadra quote this same passage from the Pañcaviṃśati
sāhasrikā. According to Sparham’s translation of Vimuktisena’s commentary,
which is based on an unpublished single Nepali manuscript,79 Vimuktisena
seems to quote the text by replacing the avetya of the mūla text by abhedya:
77 This variant was already indicated by Conze, 1973: 87.

78 See Kimura, 1986: 59 for reference. The text edited by Kimura states: evam ukte bhagavān
śakraṃ devānām indram etad avocat: tat kiṃ manyase kauśika kiyantas te jāmbūdvīpakā
manuṣyā ye buddhe ’vetyaprasādena samanvāgatā, ye dharme ’vetyaprasādena saman
vāgatā, ye saṃghe ’vetyaprasādena samanvāgatā […]? atha khalu śakro devānām indro
bhagavantam etad avocat: alpakās te bhagavan jāmbūdvīpakā manuṣyā ye buddhe ’vetya
prasādena samanvāgatā ye dharme ’vetyaprasādena samanvāgatā ye saṃghe ’vetyaprasā
dena samanvāgatā […].
79 Sparham’s translation is based “on a photocopy of a single manuscript kept in the National
Archives in Kathmandu (Ms. No. 5–55, Reel No. A37 / 9)” (Sparham, 2006: vii). For the
catalogue entry see <http://catalogue.ngmcp. uni-hamburg.de/wiki/A_37–9_Abhisamay
%C4%81la%E1%B9%85k%C4%81ravy%C4%81khy%C4%81> (accessed 2/11/2015).
There they have ‘unbroken faith’ [(abhedyaprasāda), i.e. ‘knowledgeable

faith’ (avetyaprasāda)] when having destroyed doubt, they have faith
that a knowable (avagamya) good quality is possible […] (Sparham, 2008:
22).
However, the microfilm copy of this manuscript at the Berlin State Library con-
firms the variant avetyaprasāda. This reading is clearly the preferable one since
it corresponds to the following explanation of avetya as avagamya.
Vimuktisena’s commentary was taken up by the later commentator
Haribhadra,80 who comments on the same passage in his ālokā as follows (ed.
Wogihara, 1932: 213):
avagamyaguṇasambhāvanāpūrvakaḥ prasādo ’vetyaprasādo vicikitsāpra

hāṇād ity eke. dṛṣṭatattvasya śraddhā triṣu ratneṣv āryakāntaṃ ca śīlaṃ
caturtham avetyaprasāda ity anye
Some [say], ‘perfect confidence’ is confidence accompanied by the reali-

sation of knowable good qualities resulting from giving up doubts. Others
say ‘perfect confidence’ (avetya-prasāda) is a confidence in the Three
Jewels and fourth, morality pleasing to noble beings, of one who has seen
the true reality (modified from Sparham, 2008: 160).
According to the available editions, Haribhadra uses the conventional variant

avetyaprasāda throughout.81 Another aspect of interest in Haribhadra’s com-
mentary is that the text cites an opinion, which lists four a°prasādas82 including
morality (śīla). This development of a fourfold list of a°prasādas will be inves-
tigated in the next paragraph.
In the selected examples above, we have seen that the term abhedyaprasāda
was rendered in Tibetan as mi phyed pa’i dad pa / mi phyed par dad cing and in
some of the Chinese translations as buhuaixin 不壞信 or buhuaijing 不壞淨.
Both translations leave little doubt about the association of the compound’s
80 Haribhadra wrote his commentary on the Abhisamayālaṃkāra “about the year 800 dur-
ing the reign of Dharmapāla (rg. C. 770–810), the greatest of the Pāla kings” (Sparham,
2006: xv).
81 Once more, Sparham’s translation suggests that the commentary used the term
abhedyaprasāda: One, [i.e. Ārya-Vimuktisena] says, “they have ‘unbroken faith’ [(abhedya-
prasāda), i.e. ‘knowledgeable faith’ (avetya-prasāda)]” (Sparham, 2008: 160). However,
neither the text edited by Tucci, 1932: 182, nor Wogihara’s edition of the AAA (cf. above)
refer to this reading.
82 In the following the term a°prasāda is used to designate both variants of this term:
avetyaprasāda and abhedyaprasāda.
first member with Skt. abhedya. There seems to be some evidence that this
variant was far more widely used than our small survey suggests. Thus, the
Mahāvyutpatti (ed. Sakaki, 1916: 440) lists not only the more common term
avetyaprasāda, but also adds the variant abhedyaprasāda:
6823. (562). śes nas dad pa: avetyaprasādaḥ

6824. (563). dad pa mi phyed pa: abhedyaprasādaḥ
The Tibetan mi phyed pa (< ’byed ba “to split, to break”) clearly points to Skt.
abhedya. It seems that the majority of the early Chinese translators up to
Paramārtha (563 CE) even preferred this variant. The first element of the com-
pound a°prasāda is here usually represented by Chinese buhuai 不壞
“indestructible, unbreakable, incorruptible.” Thus, we find the form buhuaixin
不壞信 in the Chinese Dīrghāgama (T.1.1), in the separately translated
Saṃyuktāgama (T.2.100), in both translations of the Larger Buddhāvataṁsaka
(T.9.278 and T.10.279), and in several Prajñāpāramitā texts translated by
Kumārajīva (e.g. T.8.227, T.25.1509). Only from the late sixth century onwards
(and in particular in the translations by Xuanzang) is the first element of the
compound usually represented by zheng 證 “to realize,” which may be related
to Skt. avetya.83
It is hardly probable that in all these aforementioned cases the underlying
Indic text really contained the reading abhedyaprasāda. But at least in cases
where the older variant buhuai 不壞 is used by post-6th c. translators (such as
some of Xuanzang’s translations), there is good reason to argue that the origi-
nal text contained abhedya rather than avetya. A systematic investigation of
this question is beyond the scale of the present study.
There is some evidence that certain Abhidharma texts are aware of this
alternative interpretation of the term. Discussing this issue with Lin Qian, he
drew my attention to an important passage from the Mahāvibhāṣā (T.27.1545,
534c14–29) and provided the following translation:
Question: Why [they] are referred to as avetya-prasāda? What is the

meaning of avetya-prasāda?
Answer: (1) [They are referred to as] ‘purities’ (prasāda, jing 淨) because
[they refer to] faith (śraddhā, xin 信) and virtue (sīla, jie 戒) removed
from defilements. Having contemplated, pondered, and apprehended
the four noble truths one after another, [one] attains these purities,
therefore [they] are referred to as avetyaprasāda.
83 I am most grateful to Lin Qian and Jan Nattier, who kindly provided this evidence for me.
(2) The Venerable Pārśva (xizunzhe 脇尊者) says that it should be

‘unbreakable purity’ (*abhedya-prasāda, buhuaijing 不壞淨). It is referred
to as ‘unbreakable’ (*abhedya) because it is not to be broken by faithless-
ness (*aśraddha, buxin 不信) and those false virtues (*duḥśīla, ejie 惡戒).
‘Purity’ (prasāda) means pure faith (śraddhā, xin 信), because it is the
pure characteristic of the mind, and virtue (śīla) is the pure characteristic
of the great elements (mahābhūta, dazhong 大種).
(3) The Venerable Vasumitra (Shiyou 世友) says thus, they should be
referred to as ‘uninterrupted purities’ (*nitya or *abhedya, buduanjing 不
斷淨), namely, once attained, they are not to be led astray by the power of
any śramaṇa, brāhmaṇa, etc., interrupted or destroyed. As it is said in the
sūtra, ‘This is referred to as faith having [right] view as its root, and asso-
ciated with the knowledge of comprehension, śramaṇas and brāhmaṇas
etc. of this world are not able to lead astray and cause it to be interrupted
and destroyed.’
(4) The Bhadanta [Dharmatrāta] says, if [one] cannot contemplate,
ponder, and apprehend the Buddha dharmas, the faith and virtue attained
can be easily moved like a boat on water. If [one] can carefully contem-
plate, ponder, and apprehend the Buddha dharmas, the faith and virtue
attained are immovable like an *indradhvaja (dichuang 帝幢). Correctly
it should be ‘immovable-purity’ (budongjing 不動淨).
(5) The Venerable Ghoṣaka says that these four should be referred to as
the ‘purities of view’ (*dṛṣṭi-prasāda, jianjing 見淨), because these puri-
ties are attained after seeing the four noble truths. Or [they] should be
referred to as the ‘purities of understanding’ (*prajñā-prāsāda, huijing 慧
淨), because they function together with the noble understanding (*ārya-
prajñā, shenghui 聖慧).
It seems that at least two of the five explanations given here, numbers 2
and 3, point to abhedya as the underlying form rather than to avetya.84 A
slightly different explanation based on the same etymology is given by the
Saṃyuktābhidharmahṛdaya:
Question: What is known?

Answer: The four noble truths. It is further said that they are called ‘per-
fect faith’: just as the increase of sūra (= strength, power, Skt. śūra).
Furthermore, some say that what is not abandoned because of agitation
is called ‘perfect faith’: just as the increase of confidence (pratiśaraṇa).
84 Dharmatrāta’s interpretation (no. 4) recalls Buddhaghosa’s acala.

These two kinds of increase both acquire the first path. Māra cannot
destroy or break [it]. Each is named by depending on the specific expla-
nation (Dessein, 1999,1: 681, my emphasis).
As Bart Dessein points out, Saṃghavarman’s Chinese translation of Dhar

matrāta’s Saṃyuktābhidharmahṛdaya (T.28.1552) uses the Chinese term
buhuaijing 不壞淨 (Skt. abhedyaprasāda) throughout (see Dessein, 1999,3:
31, s.v. avetyaprasāda). In his note on stanza 169, Dessein argues that this
Chinese rendering buhuaijing 不壞淨 is “a wrong translation of the Sanskrit”
(Dessein, 1999,2: 201–202). As stated above, the preference of pre-Xuanzang
translators for this variant cannot prove that the original text contained this
variant. But in light of the explanation given in the text, one might assume
that the variant abhedya was not completely unknown to the author of the
Saṃyuktābhidharmahṛdaya.85
Without a doubt, the earliest extant attestation of the term abhedya comes
from our manuscript from Gandhāra. Like the examples from the Mahāvibhāṣā
and, probably, the Saṃyuktābhidharmahṛdaya, the Gāndhārī text not only
uses this term, but it even tries to explain the specific meaning of the attribute
abhedya as “unbreakable, indivisible,” based on its etymology. In addition, the
text uses the simile of space (ākāśa), which is described as “indivisible as a hole
pierced by a hundredfold split tip of a hair:”
[ṇa sukaro ag̱aśo chidido vi bhidido vi ◊] (*chidro vi) [sakato] atamaśo ◊

śadadha chiṇa vi valagrakoḍ̱ie ◊ (BajC2, 2A.4–5)
It is not easy to split or break the space, (*just like a hole) that was pierced
by an even hundredfold split tip of a hair.
Is it therefore possible to argue that the new term was introduced in a north-
western environment before it was introduced into other contexts including
the translational practice of early Chinese translators?
There is no easy way to explain the sound change from Middle Indic avecca
to Gāndhārī abheja. Such a transformation can only be justified by an inten-
tional reinterpretation of an inherited, but obscure term. This transformation
might be at least partially due to a hypersanskritisation based on the language
of the Indian northwest. Only in the orthography of the northwest is Old Indian
bh- regularly represented by the sign for the labio-dental v or its aspirated vari-
ant vh. Besides that, the “historical” spelling bh is frequently attested (e.g. Skt.
85 The variant buhuaijing 不壞淨 is also used in the other Hṛdaya works by Dharmaśreṣṭhin
and Upatrāta. See footnote 96 below.
prabhā > prava, pravha, prabha). Inherited intervocalic bh was obviously pro-
nounced as a fricative with or without aspiration, in clear departure from other
Middle Indic languages where we observe the change bh > h (von Hinüber,
2001: §§ 190f.). In a Gāndhārī environment a term avecca could easily be mis-
understood as a word containing an aspirated labial. The change from cca to
j( j)a is more complicated. It could be explained as an intentional shift from a
no longer comprehensible form aveca / abheca to a hypersanskritized form
abheja (Skt. abhedya). However, it cannot be completely excluded that this
shift also had a phonetical background. That the pronounciation of c and j was
sometimes confused, is demonstrated by some Prakrit grammarians (von
Hinüber, 2001: 155, § 177). Moreover, Kenneth R. Norman (1970: 134–135) lists a
number of words where this change obviously occurred. The interchangeabil-
ity of c and j is also occasionally attested in a Gāndhārī environment, as shown
by one of the Senior fragments where OI añjali is written as G acali (GD, Index
s.v. acali). Thus, both changes (v > bh, c > j) are at least hypothetically within
the range of possible phonetical developments of Gāndhārī. Especially the
characteristic shift from v to bh makes a Gāndhārī influence on the emergence
of this variant highly probable.
3.2 The abhedyaprasādas as a Group of Four

The Abhidharma sources and commentaries cited above refer to a tradition
which knows four varieties of a°prasādas. According to Haribhadra, the “‘per-
fect faith’ (avetyaprasāda)” comprises the “perfect faith in the Three Jewels”
and, as fourth, the “morality pleasing to noble beings in those who have seen
true reality”. This fourfold list is in accordance with the text of the Bajaur
Mahāyāna sūtra, which clearly refers to four abhedyaprasādas (cadu
abhejapras̱ada), namely:
1. towards the Buddha

2. towards the Dharma
3. towards the Saṃgha
4. towards the noble virtues
This fourfold list is not attested in the earliest layers of Buddhist literature but
appears to belong to a specific scholastic tradition. Usually, the early texts refer
only to three such items, namely the three jewels. These three a°prasādas are
arranged together with (ārya)śīla to another fourfold list: that of the
srotāpattyaṅgas, the constituents of stream entry. As locus classicus for the
definition of the a°prasādas as a part of the srotāpattyaṅgas in canonical lit-
erature, I quote a passage from the Pāli Saṃgītisuttanta (DN 33 III 227):
Cattāri sotāpannassa aṅgāni: idhāvuso ariyasāvako buddhe avecca-

ppasādena samannāgato hoti: iti pi so bhagavā arahaṃ sammāsambuddho
vijjācaraṇasampanno sugato lokavidū anuttaro purisadammasārathī
satthā devamanussānaṃ buddho bhagavā’ti. Dhamme aveccappasādena
samannāgato hoti: svākkhāto bhagavatā dhammo sandiṭṭhiko akāliko
ehipassiko opanayiko paccattaṃ veditabbo viññūhī’ti. Saṅghe avec
cappasādena samannāgato hoti: supaṭipanno bhagavato sāvakasaṅgho,
ujupaṭipanno bhagavato sāvakasaṅgho, ñāyapaṭipanno bhagavato
sāvakasaṅgho, sāmīcipaṭipanno bhagavato sāvakasaṅgho, yadidaṃ cattāri
purisayugāni, aṭṭha purisapuggalā, esa bhagavato sāvakasaṅgho
āhuṇeyyo pāhuṇeyayā dakkhiṇeyyo añjalikaraṇīyo anuttaraṃ puñña
kkhettaṃ lokassā’ti. Ariyakantehi sīlehi samannāgato hoti akhaṇḍehi
acchiddehi asabalehi akammāsehi bhujissehi viññūppasatthehi aparā
maṭṭhehi samādhisaṃvattanikehi.
Four characteristics of a Stream-Winner: Here, the Ariyan disciple (ariya

sāvaka) is possessed of unwavering confidence (aveccapasāda) in the
Buddha, thus: ‘This Blessed Lord is an Arahant, a fully-enlightened
Buddha, endowed with wisdom and conduct, the Well-Farer, Knower of
the worlds, incomparable Trainer of men to be tamed, Teachers of gods
and humans, enlightened and blessed.’ He is possessed of unwavering
confidence in the Dhamma, thus: ‘Well-proclaimed by the Lord is the
Dhamma, visible here and now, timeless, inviting inspection, leading
onward, to be comprehended by the wise each one for himself.’ He is pos-
sessed of unwavering confidence in the Sangha, thus: ‘Well-directed is
the Sangha of the Lord’s disciples, of upright conduct, on the right path,
on the perfect path; that is to say the four pairs of persons, the eight kinds
of men. The Sangha of the Lord’s disciples is worthy of offerings, worthy
of hospitality, worthy of gifts, worthy of veneration, an unsurpassed field
of merit for the world.’ And he is possessed of morality dear to the Noble
Ones, unbroken, without defect, unspotted, without inconsistency, liber-
ating, praised by the wise, uncorrupted, and conducive to concentration
(tr. Walshe, 1995: 490–491).
This fourfold list of srotāpattyaṅgas is also part of the Sanskrit, Gāndhārī and
Chinese versions of the Saṃgītisūtra and its commentaries (see § 2.3.2). As
seen above, the srotāpattyaṅgas (G so[ḏavati](*aga)) are also mentioned
among the groups of four listed in the respective section of the Bajaur
Mahāyāna sūtra.
Apparently, this well established and widely known arrangement of srotā

pattyaṅgas influenced the list of the a°prasādas and resulted in the inclusion
of the additional element ‘morality’ (śīla). It is difficult to ascertain when and
in which environment this altered, fourfold, list of a°prasādas originated, but
there appears to be good reason to believe that Sarvāstivāda Abhidharma cir-
cles were the first to refer to the ‘four avetyaprasādas’.
Already the Dharmaskandha, according to Frauwallner “the Sarvāstivādin’s
earliest Abhidharma work after the Saṃgītiparyāya” and composed in the time
“before Aśoka’s missions” (Frauwallner, 1995: 20),86 lists the four a°prasādas
(T.26.1537: 460 a21–28) in the first part. The Dharmaskandha begins its discus-
sion by quoting a sūtra passage that corresponds to the 836th sūtra of the
Chinese Saṃyuktāgama (T.2.99: 214 b7–19), which is a parallel to Pāli SN 55.17
(V 365–366).87 The Chinese version of this sūtra clearly speaks of four
a°prasādas (T.2.99: 214 b10+12: si buhuaijing 四不壞淨). Consequently, the
Dharmaskandha takes up this Āgama passage and states (T.26.1537: 460
a21–24):
What are the four *avetyaprasādas (si zhengjing 四證淨)? They are: bud-
dha-avetyaprasāda, dharma-avetyaprasāda, saṅgha-avetyaprasāda, and
the virtue favored by the nobles. Why? The four great elements, namely,
the elements of earth, water, fire, and air, are capable of change; those
noble disciples who have achieved the four avetyaprasādas definitely will
not change (translation: Lin Qian).88
It seems therefore that the transformation of the a°prasādas into a group of

four was also introduced into the canonical text of the (Mūla-)Sarvāstivāda
Saṃyuktāgama.89 The original understanding of this sūtra was probably a dif-
ferent one, as indicated by the Pāli version that refers instead to the four
sotāpattiyaṅgas (SN V 365–366). As in the Saṃgītisuttanta, the three avecca
pasādas are mentioned as the first three. It is therefore possible that the text
86 But cf. Willemen, Dessein and Cox, 1999: 69. Based on the quotations in the Saṃgītiparyāya,
they consider the Dharmaskandha as “the oldest of the seven Abhidharma works”.
87 I am once more indebted to Lin Qian, who guided me through the Chinese texts of the
Dharmaskandha and the SĀ. For more parallels to this sūtra see Chung, 2008: 185.
88 Although the Chinese translation of the SĀ (Guṇabhadra, 443 CE) uses the term buhuai-
jing 不壞淨 (Skt. abhedyaprasāda), the Dharmaskandha (Xuanzang, 659 CE) refers to
zhengjing 證淨 (Skt. avetyaprasāda). This again shows the difficulties in making any con-
clusions that are solely based on the terminology of early Chinese translations.
89 For the school affiliation of the Chinese SĀ see the discussion by Chung (2008: 11–20).
underlying the Chinese SĀ had already replaced the reference to the four
srotāpattyaṅgas by that to the four a°prasādas.
The character of a°prasādas as a group of four had become firmly estab-
lished in the later Sarvāstivāda Abhidharma tradition, although the distinction
between the first three members of this group and śīla as its fourth element
had been a matter of discussion. For example, the Mahāvibhāṣā discusses this
problem in the passage directly preceding the one cited above (T.27.1545:
534c5–10, quoted after Dessein, 1999: 513, note 450):
Question: How are the four forms of perfect faith established? Is it

because of uniqueness or is it because of that which is taken as support-
ing object? When because of uniqueness, there are only two: faith and
restraint. When because of that which is taken as supporting object,
there are only three: perfect faith in Buddha, in the doctrine and in the
order: because moral precept does not have that which it takes as sup-
porting object. Answer: This statement should be made: it is both by the
uniqueness and by that which is taken as supporting object. Of the forms
of perfect faith, the one established by uniqueness is perfect faith in
moral precept: because moral precept has nothing it takes as supporting
object. Produced by that which is taken as supporting object, are the
other three forms of perfect faith: because faith takes the three treasures
as supporting object.
The same sort of discussion is also found in Vasubandhu’s Abhidharmakośa. In

the kārikās we find the following passage (6.73–74):
trisatyadarśane śīladharmāvetyaprasādayoḥ
lābho mārgābhisamaye buddhatatsaṃghayor api (ed. Pradhan, 1975: 386)
The relevant expression in the kārikā text, i.e. śīladharmāvetyaprasādayoḥ, is

not quite clear and did pose certain problems to its later commentators and
translators depending on whether the dual dvandva compound is dissolved as
“morality and the Perfect Confidence in the dharma” or “the two Perfect
Confidences in morality and dharma”. Thus, Louis de La Vallée Poussin trans-
lates (1925: 292):
Quand on voit trois vérités, on obtient la moralité et l’avetyaprasāda rela-

tivement au Dharma: quand on comprend le chemin, aussi l’avetyaprasāda
relativement au Bouddha et à son Saṃgha.
In his note to this passage he remarks (293):
Hiuan-tsang dit: ‘Le Sūtra dit qu’il y a quatre avetyaprasādas: à l’endroit

du Bouddha, du Dharma, du Saṃgha, de l’āryaśīla.’ On peut dire qu’il y a
āvetyaprasāda en ce qui concerne la moralité, śīla, car prasāda = pureté
[…]. Mais Paramārtha et le tibétain montrent qu’il ne faut pas entendre
notre kārikā: ‘obtention de l’avetyaprasāda relativement à la moralité et
au Dharma.’
However, the auto-commentary makes quite clear that Vasubandhu indeed

refers to four prasādas without ignoring the fundamental differences between
the first three (in Buddha, Dharma, Saṃgha) and the fourth one (in morality).
Thus, the bhāṣya commenting upon kārikā 74 and 75 states:
ta ete śraddhādhiṣṭhānabhedān nāmataś catvāro ’vetyaprasādā ucyante

dravyatas tu dve śraddhā śīlaṃ ca,
buddhadharmasaṃghāvetyaprasādāḥ śraddhāsvabhāvāḥ,
āryakāntāni ca śīlāni śīlam iti dve dravye bhavataḥ
[…]
avetyaprasādā iti ko’rthaḥ? yathābhūtasatyāny avabudhya sampratyayo
’vetyaprasādaḥ |
yathā ca vyutthitaḥ saṃmukhīkaroti tathaiṣām ānupūrvīm |
kathaṃ vyutthitaḥ saṃmukhīkaroti? samyaksaṃbuddho vata bhagavān,
svākhyāto ’sya dharmavinayaḥ, supratipanno ’sya śrāvakasaṃgha iti;
vaidyabhaiṣajyopasthāpaka-bhūtatvāt |
cittaprasādakṛtaś ca śīlaprasāda ity ucyate caturtha uktaḥ |
evaṃ prasannasyaiṣā pratipattir iti;
ārogyabhūtatvād vā, deśikamārgasārthikayānavad vā | (ed. Pradhan, 1975:
387)90
On a donc, vu la variété de l’objet du prasāda, quatre prasādas distingus

au point de vue des noms.
Au point de vue des choses, ces quatre sont deux choses, foi et moralité.
Les avetyaprasādas relativement au Bouddha, au Dharma, au Saṃgha,
sont, de leur nature, foi (śraddhā). Les moralités chères aux Āryas, sont de
leur nature, moralité (śīla). Donc deux choses. […]
90 Cited after the improved text of the Bibliotheca Polyglotta (<http://www2.hf.uio.no/poly-

glotta/index.php?page=fulltext&vid=511&view=fulltext>, access 29.11.2015).
Quel est le sense de cette expression avetyaprasāda? Foi consécutive à la

compréhension exacte des vérités. Les avetyaprasādas sont rangés dans
l’ordre où, en sortant de la contemplation des vérités (vyutthita), on se les
rend présents (saṃmukhīkaroti). – Comment se les rend-on présents en
sortant de la contemplation? – “Oh! Bhagavat est parfait Bouddha! Bien
prêché son Dharma-Vinaya! Bien en route son Śrāvakasaṃgha !” c’est
ainsi qu’on se les rend présents, car le Bouddha, le Dharma et le Saṃgha
sont, dans l’ordre, le médicin, le remède, l’infirmier.
Comme le prasāda de la moralité résulte du prasāda de la pensée, il est
nommé, quatrième, à la fin: c’est quand la pensée est ainsi croyante (pra-
sanna) qu’on possède la moralité chère aux Āryas (de La Vallée Poussin,
1925: 294–295).
Although Vasubandhu clearly admits the difference between these two types
of avetyaprasāda, his commentary leaves no doubt that āryakāntaśīla-
avetyaprasāda has to be regarded as part of a fourfould list of avetyaprasādas.
For his Abhidharmakośa Vasubandhu used “the Abhidharma system as it
had been systematized by Dharmaśreṣṭhin and revised and enlarged by
Upaśānta and Dharmatrāta” (Willemen, Dessein and Cox, 1998: 270). The
*Abhidharmahṛdayaśāstra, written by Dharmaśreṣṭhin/Dharmaśrī from
Bactria probably between 220 BCE and 220 CE91 and translated into Chinese in
391 CE (cf. Willemen, Dessein and Cox, 1998: 255–256) represents the earliest
preserved systematic compilation of Sarvāstivāda dogmatics. Dharmaśreṣṭhin’s
work was the basis of the two Gandhāran Abhidharmahṛdaya works by
Upaśānta and Dharmatrāta who lived in the third and early fourth centuries
(cf. Willemen, Dessein and Cox, 1998: 259 and 261). All these works consistently
refer to a fourfould list of a°prasādas.92
91 For a detailed discussion of the different opinions regardings Dharmaśreṣṭhin’s disputed

life-time cf. Willemen, 1975: v-viii.
92 Moreover, all Chinese translations of these Hṛdaya treatises use the variant abhedya
prasāda (buhuaijing 不壞淨 / buhuaixin 不壞信). Since these translations belong the
pre-Xuanzang phase (cf. for the exact dates Willemen, Dessein and Cox, 1998 : 253–263),
the value of this terminological usage is restricted. In Dharmaśreṣṭhin’s Abhidhar
mahṛdaya (T.28.1550: 827c) we find the following passage: “Question: The World-Hon-
oured One has expounded four perfect faiths (si buhuaijing 四不壞淨): perfect faith in
the Buddha, perfect faith in the dharma, in the saṃgha, and in noble morality. What
about these? Answer: (188) Pure and stainless faith in the qualities of the self-awakened
and of the disciple, in deliverance and in the remaining causality, and noble morality have
attained certainty. Pure and stainless faith in the qualities of the self-awakened and of the
disciple, in deliverance and in the remaining causality: a self-awakened one is a Buddha.
In Dharmatrāta’s Saṃyuktābhidharmahṛdaya, we find a discussion of the

list that closely resembles that of the Mahāvibhāṣā, without leaving any doubt
about the integrity of the group as a whole:
Question: How many actual entities (vastu) do these forms of perfect

faith have?
Answer: ‘There are two forms of these actual entities’: Faith and moral
precept. Faith is awarenesses that are clean; moral precept is the four ele-
ments that are clean.
‘It is said that there is the name of four forms’: Because of being estab-
lished by the actual entities as supporting object, there are four [forms];
because of the difference of being with faith as supporting object, there
are three forms (Dessein, 1999,1: 681).
All these references demonstrate that the tradition, which refers to the
a°prasādas as a group of four was well established in Sarvāstivāda Abhidharma,
although the different character of its three original members (Buddha,
Dharma, Saṃgha) and the later incorporated moral (śīla) continued to be
discussed.
Vasubandhu, as well as his direct predecessors Upaśānta and Dharmatrāta,
lived in Gandhāra. It might, therefore, be hardly surprising that the author of
the Bajaur Mahāyāna sūtra referred in his work to this fourfold list of a°prasādas
that had become commonplace at least in the northwestern Abhidharma
traditions.
3.3 The Reinterpretation of this Category in the Context of the Bajaur

Mahāyāna sūtra
In the “Mainstream Buddhism” traditions the concept of avetyaprasādas is
usually based on the intentional reflection on the positive qualities of Buddha,
That Buddhahood is comprised within the fruit of being without attachment. The quali-
ties of one who has no more training to do are the qualities of a Buddha. When one has
pure faith in these qualities, it is called perfect faith in the Buddha. Having taken up the
realization of that which is right, one is a disciple. The qualities of one in training and of
one who has no more training to do are said to be the qualities of a disciple. When has
pure faith in these qualities, it is called perfect faith in the saṃgha. Pure faith in nirvāṇa
and faith in the remaining formed dharmas, such as the truth of suffering and the truth of
origination, faith in the pure qualities of the bodhisattva, and faith in the qualities of the
pratyekabuddha who is in training or who has no more training to do, this is called perfect
faith in the dharma. Noble morality is pure morality. This is called perfect faith in moral-
ity” (Willemen, 1975: 135–136).
Dharma, Saṃgha and morality. This is obvious from the above quoted passage
from the Saṃgītisuttanta and it is also evident from the passage extracted from
the Abhidharmakośabhāṣya.
The author of the Gāndhārī sūtra chooses a different approach, which is,
however, in accordance with the general message of the text, which is based on
the notion of emptiness (śūnyatā). Usually the notion of emptiness is expressed
by the phrase na samanupaśyati, “does not perceive”. In accordance with this
rhetoric, the abhedyaprasādas are defined. For sake of briefness I quote only
few characteristic extracts for each of the four items:93
Buddha
utamaṭ́haṇaṭ́hido vi tasag̱ado ◊ ṇa samaṇupaśati ◊ paramaṭ́haṇaṭ́hida vi ◊
tasag̱ada ṇa sa(*maṇupaśati ·) (BajC2, 1A3–4) […]
yado ya śariputra ◊ mamo ṣ̱avag̱a · edehi ca ◊ añehi ca karaṇehi ◊ ṇa
samaṇupaśati · tado budho abhejapras̱a(*deṇa samuṇaga)[d]a bho[di]
(BajC2, 1.A7–8 + 1CD.18)
He also does not perceive the Tathāgata as being in the highest place
(uttamasthāna-sthita). He also does not perceive the Tathāgata as being
in the supreme place (paramasthāna-sthita). […]
And because, Śāriputra, my disciple does not perceive [the Tathāgata (?)]
out of these and other reasons, he is endowed with unbreakable confi-
dence in the Buddha.
Dharma
yoda94 [vi] dharmaviharam eva ṇa samaṇupaśati ◊ tado vi dharmo ◊
abh[e]jopras̱{e}deṇa samuṇag̱ado bhodi (BajC2, 1.A8 + 1CD.18)
And also because he does not perceive a dwelling in the Dharma

(dharmavihāra), he is endowed with unbreakable confidence in the
Dharma.
Saṃgha
ya[s̱a] yeva tu[a] (*śariputra) dharma ṇa samaṇupaśas̱i ◊ yeṇa dharmeṇa
samuṇag̱ado raha di voharias̱i ◊ evam eva śariputra ◊ yeṇa dharmeṇa ◊
mama ṣ̱avag̱a-sagho ṣ̱avag̱a(*sa)[gha] saṃkho gachati ◊ ta dharmo aria ·
ṣ̱avag̱o ◊ yoṇiśo vavarikṣata ◊ · ṇa as̱ig̱achadi ◊ yado ya ṇa as̱ig̱achadi tado
ya (*sagho a)[bhejo]pras̱adeṇa samu{s}ag̱ado bhoti ◊ (BajC2, 1CD.18–21)
93 Some of these passages are also discussed in Strauch, forthcoming.

94 Read: yado.
Just as you, Śāriputra, do not perceive a dharma by [lit. “endowed with”]

which you are called an arhat, just so, Śāriputra, an āryaśrāvaka does not
realise (adhigacchati) a dharma by which my assembly of disciples is
called [“reckoned as”] an assembly of disciples, even when thoroughly
investigating it. And because he does not realise it, he is endowed with
unbreakable confidence (*in the saṃgha).
Śīla
yado ya śariputra ◊ mamo ṣ̱avag̱o ◊ ṇa aj̄atvo samaṇupaśati ◊ kudo ⟨*bha⟩
hidho · tado ya (*ṇa aj̄atvasamu)[ṭ́hi]da śilo samaṇupaśati ṇa bhahidha
samuṭ́hida śilo samaṇupaśati ◊ ṇa ajatvabhahidhasamuṭ́hido śilo
(*samaṇupaśati ·) […] (*yavado a)[ria]ṣ̱avag̱o ◊ aribhutehi śilehi samuṇa
g̱ado bhoti · etavado śariputra ◊ caduhi abhejapras̱adehi ◊ samuṇagado
bhoti (BajC2, 1CD.21–23; 2B.13)
And because, Śāriputra, my disciple does not perceive [anything] inter-

nal let alone [anything] external, he does not perceive morality having
originated (*internally) (adhyātmasamutthita), he does not perceive
morality having originated externally (bahidhāsamutthita), he does not
perceive morality having originated both internally and externally
(adhyātmabahidhāsamutthita). […] (*The extent to which) the
āryaśrāvaka is endowed with noble virtues (āryabhūta śīla), to this
extent, Śāriputra, he is endowed with the four unbreakable confidences.
The whole treatment of the four abhedyaprasādas is concluded by the

statement:
evam eva (*śariputra) [mamo ṣ̱avag̱o] ◊ edehi ca añahi ca karaṇahi ◊

{samuṇag̱ado} budho ṇa samaṇupaśati · dharma sagho ◊ ṇa samaṇupaśati
◊ śilo samasi praña vimuti (*vimutiñaṇadarśaṇa sa)[vado sava] ◊ ṇa
samaṇupaśati · yado ṇa samaṇupaśati tado ṇa bhijati ◊ ta vucati ◊ abhejeṇa
pras̱adeṇa samuṇag̱ada · (BajC2, 2.D36–39)
Just so, (*Śāriputra), out of these and other reasons my disciple does not
perceive a Buddha, does not perceive a Dharma [or] a Saṃgha. He does
not perceive morality (śīla), concentration (samādhi), understanding
(prajñā), release (vimukti), he does not perceive anything at all. [And]
because he does not perceive [anything], he is not broken. [Therefore] it
is said: ‘[he is] endowed with unbreakable confidence’.
It becomes evident that the Bajaur text explicitly links the well-known
śrāvakayāna concept of the āryaśrāvaka and his a°prasādas with a Mahāyāna
type of notions. The confidence in the Buddha, the Dharma, the Saṃgha and
the śīla does not arise out of reflexion upon their positive qualities, but out of
their complete non-perception. The entire concept of an āryaśrāvaka and his
characteristic confidences are, thus, clearly reinterpretated in terms of the
theory of emptiness. As in other passages, the text uses well-established cate-
gories of Buddhist thinking and re-defines them according to its own
ideological needs.
The same attitude towards the a°prasādas can be observed in Nāgārjuna’s
Mahāprajñāpāramitāśāstra. In its 20th chapter, the sūtra quotation (and its
commentary) regarding the seventh bhūmi lists twenty things a bodhisattva
should avoid (viṃśatidharmā na kartavyāḥ). Among them are the following
four elements (ed. and tr. Lamotte, 1980: 2421–2422):95
17. buddhaniśrayadṛṣṭyabhiniveśo na kartavyaḥ. tathā hi na buddhadṛṣṭi

niśrayād buddhadarśanam utpadyate.
18. dharmaniśrayadṛṣṭyabhiniveśo na kartavyaḥ. dharmasyādṛṣṭatvāt.
19. saṃghaniśrayadṛṣṭyabhiniveśo na kartavyaḥ. saṃghanimitta
syāsaṃskṛtatvāt aniśrayatvāc ca.
20. śīlaniśrayadṛṣṭyabhiniveśo na kartavyaḥ. āpattyanāpattitām
anabhiniveśāt.
17. Ne pas s’attacher à la vue du recours en Buddha. En effet ce n’est pas de

cette vue que provient la [vraie] vision des Buddha.
18. Ne pas s’attacher à la vue du recours en Dharma. Car le Dharma est
invisible.
19. Ne pas s’attacher à la vue du recours en Saṃgha. Car le Saṃgha est en
soi inconditionné et ne constitue pas un support.
20. Ne pas s’attacher à la vue du recours dans les [hautes] moralités. Car
le Bodhisattva ne s’attache pas [à distinguer arbritrairement] la culpabi-
lité de l’innocence.
95 The text of the quotation roughly corresponds to the following passage from the
Pañcaviṃsatiprajñāpāramitā: punar aparaṃ subhūte bodhisattvasya mahāsattvasya
saptamyāṃ bhūmau vartamānasya viṃśatidharmā na bhavanti. katame viṃśatiḥ? yad
uta ātmagrāho ‘sya na bhavati sattvagrāho jīvagrāhaḥ pudgalagrāha ucchedagrāhaḥ
śāśvatagrāho nimittasaṃjñā hetudṛṣṭiḥ skandhābhiniveśo dhātvabhiniveśaḥ, āyata
nam ṛddhis traidhātuke pratiṣṭhānaṃ traidhātukādhyavasānaṃ traidhātuke ālayo
buddhaniśrayadṛṣṭyabhiniveśo dharmaniśrayadṛṣṭyabhiniveśaḥ saṃghaniśrayadṛṣṭyabhi
niveśaḥ śīlaniśrayadṛṣṭyabhiniveśaḥ śūnyā dharmā iti vivādaḥ śūnyatāvirodhaś cāsya na
bhavati, ime subhūte viṃśatidharmā bodhisattvasya mahāsattvasya saptamyāṃ bhūmau
vartamānasya na bhavanti (ed. Kimura, 2009: 90).
As Étienne Lamotte rightly remarks, “les articles 17 à 20 sont une critique dis-
crète contre le noble disciple animé d’une foi éclairée à l’endroit du Buddha, du
Dharma et du Saṃgha, et doué des moralités chères aux saints” (Lamotte, 1980:
2422, note 1). The sūtra text and Nāgārjuna explicitly justify these twenty avoid-
able things by referring to the notion of emptiness. The same kind of critique
against the traditional view of an āryaśrāvaka based on the doctrine of empti-
ness can certainly be stated for the treatment of the abhedyaprasādas in the
Bajaur Mahāyāna sūtra.
4 Conclusion
Within the introductory passage of the “Bajaur Mahāyāna sūtra” (BajC2, Frag
ment 2 of the Bajaur Collection) which is concerned with the emptiness of all
dharmas, the text stresses two categories: the four asaṃhārya-dharmas and
the four abhedyaprasādas. Both are explained in longer passages.
The four asaṃhārya-dharmas, “unconquerable things,” are the last item in a
list of altogether twenty-one groups of four. They circumscribe the highest per-
fect awakening and omniscience of a tathāgata, which make him invincible in
regard to Māra or other adversaries. While the asaṃhārya-dharmas are spe-
cific to Prajñāpāramitā texts, especially the Aṣṭadaśasāhasrikā, all other
categories of the list occur in both Abhidharma and/or Prajñāpāramitā texts,
where they represent the teaching or the characteristics of a buddha. The first
three items of this list are part of the bodhipakṣya-dharmas, and the next four
are related to meditation and include the four noble truths. Up to here every-
thing belongs to the śrāvakadharmas and is also known from canonical or
Abhidharma texts. The next three items characterize a tathāgata, an awakened
being. Elsewhere these items are also called buddhadharmas or lokottara-
kuśaladharmas and they are only known from Mahāyāna texts. This seems to
indicate that the original Abhidharma list had been expanded in order to fit
into an explict Prajñāpāramitā or rather early Mahāyāna context.
The following ten groups of four (as far as they are preserved on the manu-
script) appear to be random selections of terms, which also occur in the
Saṃgītisūtra/-paryāya. None of the different versions of the Saṃgītisūtra/-
paryāya shows a particular parallel in regard to the sequence, selection or
spelling of the terms. However, the Gāndhārī text (BajC2) is in principle closer
to the Pāli version of the Theravādin or the Sanskrit version of the Sarvāstivādin
rather than to the Chinese and Gāndhārī versions of the Dharmaguptaka.
Despite certain parallels, it was not possible to determine a close connec-
tion to any of the Prajñāpāramitā texts. Furthermore, its restriction to groups
of four distinguishes the list of the Bajaur sūtra from all extant parallels.
Nevertheless, a common background of the lists occurring in the Prajñāpāramitā
literature and the Bajaur Mahāyāna sūtra is clearly discernible. This is con-
firmed by the general diction of the text that uses terms and concepts typical
to Prajñāpāramitā. However, the term itself, prajñāpāramitā, is not mentioned
even once in the preserved portions of the text.
The list of the groups of four leads to a discussion of the four abhedyaprasādas,
“unbreakable confidences”. The extensive treatment of the abhedyaprasādas
appears to be an original trait of the sūtra that cannot be found in other early
Mahāyāna texts. The peculiar variant abhedya replacing the more common
avecca / avetya of other traditions as well as the arrangement of the traditional
three a°prasādas together with morality (śīla) in a group of four can be traced
back to early Abhidharma traditions attested for the Sarvāstivādins. While the
arrangement as a fourfold group seems to be a pan-Sarvāstivādin feature, the
specific interpretation of the a°prasādas as abhedyaprasādas, “unbreakable
confidences,” could have its origins in the circles of early Gandhāran
Abhidharma specialists. From there, however, it seems to have spread out to
various traditions including the translational Chinese literature.
The Bajaur Mahāyāna sūtra uses this well-established concept of
“Mainstream Buddhism” and reinterprets, or to use an expression by Paul
Harrison, “Mahāyāna-ises” it “in terms of the doctrines of Śūnyatā” (1978: 55). It
thus follows a strategy that can also be observed in other early Mahāyāna texts.
As Johannes Bronkhorst (forthcoming) correctly noticed, the dogmatic dis-
courses of early Mahāyāna literature presupposed the existence of a
well-developed Abhidharma tradition, and there seems to be good evidence
that the rich Sarvāstivāda Abhidharma, which was particularly influential in
the Indian northwest, was one of the main sources for the upcoming Mahāyāna
and its terminology and scholarly debates in early Gandhāra.
Abbreviations
AA Abhisamayālaṃkāra (ed. Wogihara, 1932)

AAA Abhisamayālaṃkārālokā of Haribhadra (ed. Wogihara, 1932)
AAV Abhisamayālaṃkāravṛtti Sphuṭārthā of Haribhadra (ed. Tripathi, 1977)
AN Aṅguttaranikāya
BajC Bajaur Collection, fragment no.
BHSD Buddhist Hybrid Sanskrit Dictionary (Edgerton, 1953)
BL British Library Collection, fragment no.
Ch. Chinese
Cm Commentary
CPD Critical Pāli Dictionary (Trenckner et al., 1924–)
DĀ Dīrghāgama
DN Dīghanikāya
Dhsgr Dharmasaṃgraha of Nāgārjuna (ed. Müller and Wenzel, 1885)
G. Gāndhārī
GD A Dictionary of Gāndhārī, Stefan Baums and Andrew Glass, eds., <http://
gandhari.org/dictionary.php>
LPG Larger Prajñāpāramitā from Gilgit (partly ed. Conze, 1962, 1974, Zacchetti,
2005)
MAV(Bh) Madhyāntavibhāga-bhāṣya (ed. Nagao, 1964)
MPPŚ Mahāprajñāpāramitāśāstra (T.25.1509)
MS Martin Schøyen Collection, fragment no.
Msa Mahāyānasūtrālaṃkāra (ed. Lévi, 1907)
MN Majjhimanikāya
P. Pāli
PP Prajñāpāramitā
Ps Paṭisambhidāmagga (ed. Taylor, 1905)
PTSD Dictionary of the Pali Text Society (Rhys Davids and Stede, 1921–1925)
PvsP Pañcaviṃśatisāhasrikā Prajñāpāramitā (Ms. Cambridge, ed. Kimura,
1986–2009)
SĀ Saṃyuktāgama
Skt. Sanskrit
SN Saṃyuttanikāya
SplitC Split Collection, fragment no.
SWTF Sanskrit-Wörterbuch der buddhistischen Texte aus den Turfan-Funden
und der kanonischen Literatur der Sarvāstivāda-Schule. Begonnen von
Ernst Waldschmidt. Im Auftrage der Akademie der Wissenschaften zu
Göttingen hrsg. von Heinz Bechert u.a. Göttingen: Vandenhoeck &
Ruprecht.
渡邊, and Ono 小野, 1924–1934)
Tib. Tibetan
Vibh Vibhaṅga (ed. Rhys Davids, 1904)
Bibliography
Allon, Mark, and Salomon, Richard (2010). “New Evidence for Mahayana in Early
Gandhāra”, The Eastern Buddhist 41, pp. 1–22.
Apple, James (2009). “‘Wordplay:’ Emergent Ideology through Semantic Elucidation. A

Rhetorical Technique in Mahāyāna Buddhist Formations”, Bulletin for the Institute of
Oriental Philosophy 25, pp. 161–173.
Aung, Shwe Zang and Rhys Davids, Caroline A.F. (1915). The Points of Controversy. Oxford:
Pāli Text Society.
Baums, Stefan (2009). “A Gāndhārī Commentary on Early Buddhist Verses: British
Library Kharoṣṭhī Fragments 7, 9, 13 and 18”. PhD.
Baums, Stefan; Glass, Andrew and Kazunobu, Matsuda (forthcoming). “Fragments of a
Gāndhārī Version of the Bhadrakalpikasūtra”, In Buddhist Manuscripts in the Schøyen
Collection, Volume IV. Edited by Jens Braarvig. Oslo: Hermes Academic Publishing.
Behrsing, Siegfried (1930). “Das Chung‐tsi‐king (衆集經) des chinesischen Dīrghāgama”.
PhD.
Binz, Wolfgang (1980). “Praṇidhāna und Vyākaraṇa. Untersuchungen zur Geschichte
der Entwicklung des Bodhisatva-Ideals”. PhD.
Bodhi Bhikkhu (2012). The Numerical Discourses of the Buddha. A Translation of the
Aṅguttara Nikāya. Boston MA: Wisdom Publications. Bodhi Bhikkhu and Ñāṇamoli
Bhikkhu (1995). The Middle Length Discourses of the Buddha. A New Translation of
the Majjhima Nikāya. Boston MA: Wisdom Publications in association with the Barre
Center for Buddhist Studies.
Bronkhorst, Johannes (1985). “Dhamma and Abhidhamma”, Bulletin of the School of
Oriental and African Studies 48, pp. 305–320.
Bronkhorst, Johannes (forthcoming). “Abhidharma in early Mahāyāna”, In Early
Mahāyāna. Edited by Paul Harrison. Sheffield: Equinox Publishing.
Brunnhölzl, Karl (2010). Gone Beyond. Vol. 1. The Prajñāpāramitā Sūtras, the Ornament
of Clear Realization, and Its Commentaries in the Tibetan Kagyü Tradition. Ithaca NY:
Snow Lion.
Brunnhölzl, Karl (2011). Gone Beyond. Vol. 2. The Prajñāpāramitā Sūtras, the Ornament
of Clear Realization, and Its Commentaries in the Tibetan Kagyü Tradition. Ithaca NY:
Snow Lion.
Chung, Jin-il (2008). A survey of the Sanskrit fragments corresponding to the Chinese
Saṃyuktāgama. Tokyo: Sankibo Busshorin.
Conze, Edward (1954). Abhisamayālaṅkāra (Serie Orientale Roma 6). Roma: Istituto
Italiano per il Medio ed Estremo Oriente.
Conze, Edward (1962). The Gilgit Manuscript of the Aṣṭādaśasāhasrikāprajñāpāramitā
(I). Chapters 55 to 70, Corresponding to the 5th Abhisamaya (Serie Orientale Roma 26).
Roma: Istituto Italiano per il Medio ed Estremo Oriente.
Conze, Edward (1973). The Perfection of Wisdom in Eight Thousand Lines & Its Verse
Summary. Bolinas CA: Four Seasons Foundation; distributed by Book People, Berkeley.
Conze, Edward (1974). The Gilgit Manuscript of the Aṣṭādaśasāhasrikāprajñāpāramitā

(II). Chapters 70 to 82, Corresponding to the 6th, 7th and 8th Abhisamayas (Serie
Orientale Roma 46). Roma: Istituto Italiano per il Medio ed Estremo Oriente.
Conze, Edward (1975). The Large Sutra on Perfect Wisdom, with the Divisions of the
Abhisamayālaṅkāra. Berkeley CA: University of California Press.
Cox, Collett (2014). “Gāndhārī Kharoṣṭhī Manuscripts: Exegetical Texts”, In From Birch
Bark to Digital Data: Recent Advances in Buddhist Manuscript Research. Papers
Presented at the Conference, Indic Buddhist Manuscripts: The State of the Field,
Stanford, June 15–19 2009. Edited by Paul Harrison and Jens-Uwe Hartmann. Wien:
Verlag der Österreichischen Akademie der Wissenschaften, pp. 35–49.
D’Amato, Mario (2012). Maitreya’s Distinguishing the Middle from the Extremes
(Madhyāntavibhāga). Along with Vasubandhu’s Commentary (Madhyāntavibhāga-
bhāṣya). A Study and Annotated Translation. New York NY: American Institute of
Buddhist Studies.
Dessein, Bart (1999). Saṃyuktābhidharmahṛdaya. Heart of Scholasticism. 3 vols.
(Buddhist Tradition Series 33–35). Delhi: Motilal Banarsidass.
Dutt, Nalinaksa (1984). Gilgit Manuscripts. Vol. I. Delhi: Sri Satguru Publications (2nd
ed.).
Edgerton, Franklin (1953). Buddhist Hybrid Sanskrit Grammar and Dictionary. New
Haven CN: Yale University Press.
Falk, Harry (2011). “The ‘Split’ Collection of Kharoṣṭhī Texts”, Annual Report of The
International Research Institute for Advanced Buddhology at Soka University 14,
pp. 13–23.
Falk, Harry and Karashima, Seishi (2012). “A First‐Century Prajñāpāramitā Manuscript
from Gandhāra – parivarta 1 (Texts from the Split Collection 1)”, Annual Report of The
pp. 19–61.
Falk, Harry and Karashima, Seishi (2013). “A First-Century Prajñāpāramitā Manuscript
from Gandhāra – parivarta 5 (Texts from the Split Collection 2)”, Annual Report of
The International Research Institute for Advanced Buddhology at Soka University 16,
pp. 97–169.
Frauwallner, Erich (1995). Studies in Abhidharma Literature and the Origins of Buddhist
Philosophical Systems. Translated from the German by Sophie Francis Kidd as Translator
and Under the Supervision of Ernst Steinkellner as Editor. Albany NY: State University
of New York.
Gethin, Rupert (1992). “The Mātikās: Memorization, Mindfulness, and the List”, In In the
Mirror of Memory: Reflections on Mindfulness and Remembrance in Indian and Tibetan
Buddhism. Edited by Janet Gyatso. Albany NY: State University of New York Press,
pp. 149–172.
Gethin, Rupert (2001). The Buddhist Path to Awakening. Second edition. Oxford:
Oneworld Publications.
Glass, Andrew (2007). Four Gāndhārī Saṃyuktāgama Sūtras. Senior Kharoṣṭhī Fragment
5. Seattle WA: University of Washington Press.
Harrison, Paul (1978). “Buddhānusmṛti in the Pratyutpanna-Buddha-Saṃmukhāvasthita-
Samādhi-Sūtra”, Journal of Indian Philosophy 6, pp. 35–57.
Harrison, Paul and Hartmann, Jens-Uwe (2014). “Introduction”, In From Birch Bark to
Digital Data: Recent Advances in Buddhist Manuscript Research. Papers Presented at
the Conference, Indic Buddhist Manuscripts: The State of the Field, Stanford, June 15–19
2009. Edited by Paul Harrison and Jens-Uwe Hartmann. Wien: Verlag der
Österreichischen Akademie der Wissenschaften, pp. vii-xxii.
Hinüber, Oskar von (2001). Das ältere Mittelindisch im Überblick (Österreichische
Akademie der Wissenschaften, philosophisch‐historische Klasse, Sitzungsberichte,
467. Band / Veröffentlichungen der Kommission für Sprachen und Kulturen
Südasiens, Heft 20). Wien: Verlag der Österreichischen Akademie der Wissenschaften.
Jaini, Padmanabh S. (1979). Sāratamā. A Pañjikā on the Aṣṭasāhasrikā Prajñāpāramitā
Sūtra by Ratnākaraśānti (Tibetan Sanskrit Works Series). Patna: Kashi Prasad Jayaswal
Research Institute.
Jayatilleke, Kulatissa Nanada (1963). Early Buddhist Theory of Knowledge. London:
Routledge.
Kimura, Takayasu (ed.) (1986). Pañcaviṃśatisāhasrikā Prajñāpāramitā II–III. Tokyo:
Sankibo Busshorin.
Kimura, Takayasu (ed.) (1990). Pañcaviṃśatisāhasrikā Prajñāpāramitā IV. Tokyo: Sankibo
Busshorin.
Kimura, Takayasu (ed.) (1992). Pañcaviṃśatisāhasrikā Prajñāpāramitā V. Tokyo: Sankibo
Busshorin.
Kimura, Takayasu (ed.) (2006). Pañcaviṃśatisāhasrikā Prajñāpāramitā VI–VIII. Tokyo:
Sankibo Busshorin.
Kimura, Takayasu (ed.) (2007). Pañcaviṃśatisāhasrikā Prajñāpāramitā I-1. Tokyo:
Sankibo Busshorin.
Kimura, Takayasu (ed.) (2009a). Pañcaviṃśatisāhasrikā Prajñāpāramitā I-2. Tokyo:
Sankibo Busshorin.
Kimura, Takayasu (2009b). Ś atasāhasrikā Prajñāpāramitā II-1. Tokyo: Sankibo Busshorin.
Kimura, Takayasu (2010a). Ś atasāhasrikā Prajñāpāramitā II-2. Tokyo: Sankibo Busshorin.
Kimura, Takayasu (2010b). Ś atasāhasrikā Prajñāpāramitā II-3. Tokyo: Sankibo Busshorin.
La Vallée Poussin, Louis de (1925). L’Abhidharmakośa de Vasubandhu. Cinquième et si-
xième chapitres. Société Belge d’Etudes Orientiales. Paris: Paul Geuthner.
La Vallée Poussin, Louis de / Pruden, Leo M. (1988–1990). Abhidharmakośabhāṣyam of
Vasubandhu. Translated into English from the French Translation of Louis de La Vallée
Poussin. English Translation by Leo M. Pruden. 4 vols.: I.1988, II.1988, III.1989, IV.1990.
Berkeley CA: Asian Humanities Press.
Lamotte, Étienne (1980). Le Traité de la Grande Vertu de Sagesse de Nāgārjuna
(Mahāprajñāpāramitāśāstra). Tome V. Chapitres XLIX–LII, et Chapitre XX (2e série)
(Publications de l’Institut Orientaliste de Louvain 24). Louvain-la-Neuve: Université

de Louvain, Institut Orientaliste.
Lefmann, Salomon (1874). “Lalita Vistara: Erzälung von dem Leben und der Lere des
Çâkya Simha”. PhD.
Lévi, Sylvain (ed.) (1907). Mahāyāna-Sūtrālaṃkāra. Exposé de la doctrine du Grand
Véhicule selon le système Yogācāra. Vol. I. Paris: Librarie Ancienne Honoré Champion.
Migme Chodron, Gelongma (2001). The Treatise on the Great Virtue of Wisdom of
Nāgārjuna (Mahāprajñāpāramitāśāstra), Ētienne Lamotte, Translated from the French
by Gelongma Migme Chodron. 5 vols. Gampo Alley.
Müller, Max F. and Wenzel, Heinrich (ed.) (1885). The Dharma-Saṃgraha. Amsterdam:
Oriental Press (reprint 1972).
Nagao, Gajin (ed.) (1964). Madhyāntavibhāga-bhāṣya. A Buddhist philosophical treatise.
Tokyo: Suzuki Research Foundation.
Ñāṇamoli Bhikkhu (2011). The Path of Purification (Visuddhimagga). Revised, third, on-
line edition. Buddhist Publication Society.
Narada, U. (1962). Discourse on Elements. Bristol: Pāli Text Society.
Norman, Kenneth R. (1970). “Some Aspects of the Phonology of the Prakrit Underlying
the Aśokan Inscriptions”, Bulletin of the School of Oriental and African Studies 33,
pp. 132–143.
Pradhan, Prahallad (ed.) (1975). Abhidharmakośabhāṣyam of Vasubandhu (Tibetan
Sanskrit Works Series 8). Patna: K.P. Jayaswal Research Institute (rev. 2nd ed.).
Raghu, Vira and Lokesh, Chandra (1995). Gilgit Buddhist Manuscripts. Revised and en-
larged compact facsimile edition. Vols. 1–3. Delhi: Satguru Publications.
Rahder, Johannes (1926). Daśabhūmikasūtra. Leuven: J.-B. Istas. PhD.
Rahder, Johannes (1928). Glossary of the Sanskrit, Tibetan, Mongolian and Chinese
Versions of the Daśabhūmika-Sūtra. Paris: Geuthner.
Rhys Davids, Caroline A.F. (ed.) (1904). The Vibhaŋga. Being the Second Book of the
Abhidhamma Piṭaka. London: Henry Frowde.
Rhys Davids, Thomas W. and Stede, William (1921–1925). The Pali Text Society’s Pali-
English Dictionary. Chipstead: Probsthain.
Sakaki, Ryōzaburō (1916). Mahāvyutpatti. Vol. 1. Kyōto: Shingonshū Kyōto Daigaku.
Samtani, Nārāyana Hemanadāsa (1971). The Arthaviniścaya-Sūtra & Its Commentary
(Nibandhana) (Written by Bhikṣu Vīryaśrīdatta of Śrī-Nālandāvihāra). Critically edited
and annotated for the first time with Introduction and several Indices (Tibetan Sanskrit
Works Series 13). Patna: K.P. Jayaswal Research Institute.
Schlosser, Andrea (2016). “On the Bodhisattva Path in Gandhāra. Edition of Fragment
4 and 11 from the Bajaur Collection of Kharoṣṭhī Manuscripts”. PhD. <http://www.
diss.fu-berlin.de/diss/receive/FUDISS_thesis_000000101376>.
Shakya, M.B. (1988). Advayaśatikā Prajñāpāramitā-sūtra. Nepal: Nagarjuna Institute of
Exact Methods.
Shukla, Karunesha (1973). Śrāvakabhūmi of Ācārya Asaṅga (Tibetan Sanskrit Works

Series 14). Patna: K.P. Jayaswal Research Institute.
Sparham, Gareth (2006). Abhisamayālaṃkāra with vṛtti and ālokā. Vṛtti by Ārya
Vimuktisena. Ālokā by Haribhadra. Volume 1: First Abhisamaya. Fremont CA: Jain
Publishing Company.
Sparham, Gareth (2008). Abhisamayālaṃkāra with vṛtti and ālokā. Vṛtti by Ārya
Vimuktisena. Ālokā by Haribhadra. Volume 2: Second and Third Abhisamaya. Fremont
CA: Jain Publishing Company.
Stache-Rosen, Valentina (1968). Das Saṅgītisūtra und sein Kommentar Saṅgītiparyāya.
Berlin: Akademie Verlag.
Stcherbatsky, Theodor and Obermiller, Eugene (1929). Abhisamayālankāra-prajñāpāra
mitā-upadeśa-śāstra: the work of Bodhisattva Maitreya (Bibliotheca Buddhica 23).
Leningrad.
Strauch, Ingo (2007/2008). The Bajaur Collection: A New Collection of Kharoṣṭhī
Manuscripts – A Preliminary Catalogue and Survey – Online Version 1.1 (May 2008).
URL: <http://gretil.sub.uni-goettingen.de/gretil_elib/Str007__Strauch_Bajaur_
Collection-Preliminary_Catalogue_and_Survey_v_1-1_2008.pdf> (accessed 2/11/2015).
Strauch, Ingo (2008). “The Bajaur Collection of Kharoṣṭhī Manuscripts – A Preliminary
Survey”, Studien zur Indologie und Iranistik 25, pp. 103–136.
Strauch, Ingo (2010). “More Missing Pieces of Early Pure Land Buddhism: New Evidence
for Akṣobhya and Abhirati in an Early Mahayana Sutra from Gandhāra”, The Eastern
Buddhist 41, pp. 23–66.
Strauch, Ingo (2011). “The character of the Indian Kharoṣṭhī script and the ‘Sanskrit
revolution’ writing system between identity and assimilation”, In The Idea of Writing.
Writing Across Borders. Edited by Alex de Voogt and Joachim Quack. Leiden: Brill,
pp. 131–168.
Strauch, Ingo (2014a). “Mahāprajāpatī Gautamī and the Order of Nuns in a Gandhāran
version of the Dakṣiṇāvibhaṅgasūtra”, In Women in Early Indian Buddhism:
Comparative Textual Studies. Edited by Alice Collett. New York NY: Oxford University
Press, pp. 17–45.
Strauch, Ingo (2014b). “The Evolution of the Buddhist rakṣā Genre in the Light of New
Evidence from Gandhāra: The *Manasvi – nāgarāja – sūtra from the Bajaur Collection
of Kharoṣṭhī Manuscripts”, Bulletin of the School of Oriental and African Studies 77,
pp. 63–84.
Strauch, Ingo (forthcoming). “Early Mahāyāna in Gandhāra. New evidence from the
Bajaur Mahāyāna Sūtra”, In Early Mahāyāna. Edited by Paul Harrison. Sheffield:
Equinox Publishing.
Study Group on Buddhist Sanskrit Literature (2006). Vimalakīrtinirdeśa. A Sanskrit
Edition Based upon the Manuscript Newly Found at the Potala Palace. Tokyo: The
Institute for Comprehensive Studies of Buddhism, Taisho University.

玄妙 (ed.) (1924–1934). Taishō shinshū daizōkyō 大正新脩大藏經, Tokyo: Taishō
- T.1.1: Dīrghāgama, Chang ahan jing 長阿含經, Buddhayaśas, Zhu Fonian 竺佛念.
- T.1.12: *Saṃgītisūtra, Daji fa men jing 大集法門經, Shihu 施護.
- T.1.13: Chang ahan shi bao fa jing 長阿含十報法經, An Shigao 安世高.
- T.1.84: Fenbie bu shi jing 分別布施經, Shihu 施護.
- T.2.99: Saṃyuktāgama, Za ahan jing 雜阿含經, Guṇabhadra.
- T.2.100: Saṃyuktāgama, Bieyi za ahan jing 別譯雜阿含經.
- T.2.125: Ekottarāgama, Zengyi ahan jing 增一阿含經, Saṃghadeva.
- T.8.222: *Pan͂ caviṃśatisāhasrikāprajn͂ āpāramitā, Guang zan jing 光讚經, Zhu Fahu
竺法護.
- T.8.227: *Aṣṭasāhasrikāprajn͂ āpāramitā, Xiaopin boruopoluomi jing 小品般若波羅
蜜經, Kumārajīva.
- T.9.272: Bodhisattvagocaropāyaviṣayavikurvāṇanirdeśa, Da sazheniganzi suo shuo
jing 大薩遮尼乾子所說經, Bodhiruci.
- T.9.278: [Buddha]avataṃsakasūtra, Dafangguang fo huayan jing 大方廣佛華嚴經,
Buddhabhadra.
- T.10.279: [Buddha]avataṃsakasūtra, Dafangguang fo huayan jing 大方廣佛華嚴經,
Śikṣānanda.
- T.14.474: Vimalakīrtinirdeśa, Weimojie jing 維摩詰經, Zhi Qian 支謙.
- T.14.475: Vimalakīrtinirdeśa, Weimojie suo shuo jing 維摩詰所說經, Kumārajīva.
- T.14.476: Vimalakīrtinirdeśa, Shuo wugou cheng jing 說無垢稱經, Xuanzang 玄奘.
- T.19.1022: Yiqie rulai xin bimi quanshen sheli baoqie yin tuoluoni jing 一切如來心祕
密全身舍利寶篋印陀羅尼經, Amoghavajra.
- T.19.1022A: Yiqie rulai xin bimi quanshen sheli baoqie yin tuoluoni jing 一切如來心
祕密全身舍利寶篋印陀羅尼經, Amoghavajra.
- T.19.1023: Yiqie rulai zhengfa bimi qie yin xin tuoluoni jing 一切如來正法祕密篋印
心陀羅尼經, Shihu 施護.
- T.25.1509: Nāgārjuna, Mahāprajn͂ āpāramitopadeśa, Da zhidu lun 大智度論,
Kumārajīva.
- T.26.1536: Śāriputra, [Abhidharma]saṃgītiparyāya[pādaśāstra], Apidamo jiyimen
zulun 阿毘達磨集異門足論, Xuanzang 玄奘.
- T.26.1537: Mahāmaudgalyāyana, [Abhidharma]dharmaskandha[pādaśāstra],
Apidamo fayun zulun 阿毘達磨法蘊足論, Xuanzang 玄奘.
- T.27.1545: 500 arhats, [Abhidharma]mahāvibhāṣā[śāstra], Apidamo da piposha lun
阿毘達磨大毘婆沙論, Xuanzang 玄奘.
- T.28.1550: Dharmaśreṣṭhin, Abhidharmahṛdaya[śāstra], Apitan xin lun 阿毘曇心
論, Saṃghadeva, Huiyuan 慧遠.
- T.28.1552: Dharmatrāta, Saṃyuktābhidharmahṛdaya[śāstra], Za apitan xin lun 雜
阿毘曇心論, Saṃghavarman.
- T.31.1599: Vasubandhu, Madhyāntavibhaṅgaṭīkā, Zhongbian fenbie lun 中邊分別

論, Paramārtha.
- T.31.1600: Vasubandhu, Madhyāntavibhaṅga, Bian zhongbian lun 辯中邊論,
Xuanzang 玄奘.
- T.31.1601: Maitreya, Madhyāntavibhaṅga, Bian zhongbian lun song 辯中邊論頌,
Xuanzang 玄奘.
Taylor, Arnold (ed.) (1905). Paṭisambhidāmagga. Vol. I. London: Henry Frowde.
Trenckner, Vilhelm et al. (1924-). A Critical Pāli Dictionary. Copenhagen: The Royal
Danish Academy of Sciences and Letters / The Department of Asian Studies, Univer
sity of Copenhagen.
Tripathi, Ram Shankar (ed.) (1977). Prajñāpāramitopadeśaśāstre Abhisamayālaṅkāravṛttiḥ
Sphuṭārthā Ācāryaharibhadraviracitā. Sarnath: Central Institute of Higher Tibetan
Studies.
Tucci, Giuseppe (1932). The Commentaries on the Prajñāpāramitās. Volume 1: The
Abhisamayālaṃkārāloka (Gaekwad’s Oriental Series 62). Baroda: Oriental Institute.
Tucci, Giuseppe (1956). Minor Buddhist Texts. Roma: Istituto italiano per il Medio ed
Estremo Oriente.
Vaidya, Paraśurāma Lakṣmaṇa (1960). Aṣṭasāhasrikā Prajñāpāramitā with Haribhadra’s
Commentary called Āloka (Buddhist Sanskrit Texts 4). Darbhanga: The Mithila
Institute.
Vaidya, Paraśurāma Lakṣmaṇa (1961). Mahāyānasūtrasaṃgraha. Part 1 (Buddhist
Sanskrit Texts 17). Darbhanga: The Mithila Institute.
Vaidya, Paraśurāma Lakṣmaṇa (1967). Daśabhūmikasūtram (Buddhist Sanskrit Texts 7).
Darbhanga: The Mithila Institute.
Walshe, Maurice (1995). The Long Discourses of the Buddha. A Translation of the Dīgha
Nikāya. Boston MA: Wisdom Publications.
Willemen, Charles (1975). The Essence of Methaphysics. Abhidharmahṛdaya. Translated
and annotated (Série Etudes et Textes 4). Bruxelles: Institut Belge des Hautes Études
Bouddhiques.
Scholasticism (Handbuch der Orientalistik 2/11). Leiden etc.: Brill.
Wogihara, Unrai (ed.) (1932). Abhisamayālaṃkārālokā Prajñāpāramitāvyākhyā. The
Work of Haribhadra. Vol. 1. Tokyo: The Toyo Bunko (repr. 1973 Sankibo Buddhist Book
Store).
Zacchetti, Stefano (2005). In Praise of the Light. A Critical Synoptic Edition with an
Annotated Translation of Chapters 1–3 of Dharmarakṣa’s Guang Zan Jing, Being the
Earliest Chinese Translation of the Larger Prajñāpāramitā. Tokyo: The International
Research Institute for Advanced Buddhology, Soka University.
108 Ditrich
Chapter 3
Interpretations of the Terms ajjhattaṃ and

bahiddhā: From the Pāli Nikāyas to the
Abhidhamma
Tamara Ditrich
1 Introduction
The terms ajjhattaṃ “internally,” bahiddhā “externally,” and ajjhattabahiddhā

“internally and externally” are attested repeatedly in the recurring refrain of
the Satipaṭṭhānasutta as well as in many other Pāli canonical and post-canoni-
cal texts, indicating their significance in the practice of the four satipaṭṭhānas.
Although they appear frequently, the texts provide very scant information
on their meaning, especially for the compound ajjhattabahiddhā. This chap-
ter systematically surveys occurrences of the three terms in the Pāli Canon,
aiming to outline specific functions and meanings of the terms in various
contexts. Firstly, it identifies the main syntagmatic formulae comprising these
terms in specific collocations in the Nikāyas. Secondly, it investigates how
these formulaic stylized expressions, occurring in a variety of syntagms, are
presented and explicated in the Abhidhamma, and how those interpretations
are linked to the explications established in the later Pāli commentarial litera-
ture which serve today as the standard interpretations of the terms. Through
a delineation of the semantic spectrum of ajjhattaṃ and bahiddhā, additional
information on the interpretation of the four satipaṭṭhānas is indicated which
is of particular significance in the light of the current prominence given to the
Satipaṭṭhānasutta, often regarded as the Ur-text for the practice of mindful-
ness in modern Buddhism.
2 Ajjhattaṃ, bahiddhā, and ajjhattaṃbahiddhā in the Suttapiṭaka
2.1 Syntagms Comprising ajjhattaṃ and bahiddhā

There are relatively numerous attestations of ajjhattaṃ, bahiddhā and
ajjhattabahiddhā in the Tipiṭaka, appearing on their own or as a component
of compounded words. Having examined all the occurrences of ajjhattaṃ,
bahiddhā and ajjhattabahiddhā in the Nikāyas, attested as they are in several

Interpretations of the Terms ajjhattaṃ and bahiddhā 109
types of collocations, a number of patterns emerge which may indicate specific

meanings and usages of the terms in various contexts. Based on an analysis of
lexical collocations of the three terms, the main typological and semantic pat-
terns identified in the Nikāyas are overviewed below as a way to mark out their
functions and meanings.
2.1.1 ajjhattaṃ, bahiddhā and ajjhattabahiddhā

The terms ajjhattaṃ, bahiddhā and ajjhattabahiddhā are attested together in
collocation in several suttas in the Nikāyas,1 always in a specific context of
instructions given about meditation, especially on the practice of the four
satipaṭṭhānas, prescribing or describing the practice in three modes of con-
templation.2 In all instances, the compound ajjhattabahiddhā is attested only
together with ajjhattaṃ and bahiddhā, forming a group of three, often con-
nected by the correlative conjunction vā. Typically, this triad appears in a
formulaic structure, as exemplified in the Satipaṭṭhānasutta3 where the same
stylized formulaic refrain follows each of its fourteen sections; e.g. the first
satipaṭṭhana on contemplation of the body (kāyānupassanā)4 always begins:
iti ajjhattaṃ vā kāye kāyānupassī viharati bahiddhā vā kāye kāyānupassī

viharati, ajjhatta-bahiddhā vā kāye kāyānupassī viharati.5
1 DN II 290–315; MN I 55–63; MN III 111–112; SN V 143, 294–297; AN III 450; Nidd I 1, 28, 63, 72,
78, 99; 340, 354, 370, 387; Nidd II 78, 124, 128.
2 The only exception appears in the Niddesa section of the Khuddakanikāya where the trip-
let is not attested within the framework of contemplation but instead in the context of
explication of the doctrine; e.g. in the Kāmasuttaniddesa, the three terms are listed in the
discussion on kāmā: “…ajjhattā kāmā bahiddhā kāmā ajjhattabahiddhā kāmā …” (Nidd
I 1). These specific occurrences of the triad in the Niddesa seem to reflect the commen-
tarial nature of the text.
3 Here references are given for the shorter version in the Majjhimanikāya (MN I 55–63),
with parallel passages from the Dīghanikāya (DN II 290–315) noted when relevant; the
longer Mahāsatipaṭṭhānasutta in the Dīghanikāya (DN II 290–315) differs only in giving a
longer explanation of the four noble truths and hence, this difference is not consequen-
tial for the study of the three modes of contemplation.
4 The same refrain appears after each of the following sections: breathing, once (MN I 56,
DN II 292); postures of the body, once (MN I 57, DN II 292); bodily activities, once (MN I 57,
DN II 293); parts of the body, once (MN I 57, DN II 294); elements, once (MN I 58, D II 294);
cemetery contemplations, nine times (MN I 58, DN II 295–298).
5 MN I 56–58; DN II 292–298); trans. Ñāṇanamoli and Bodhi, 1995: 146: “In this way he abides
contemplating the body as a body internally, or he abides contemplating the body as a
body externally, or he abides contemplating the body as a body both internally and exter-
nally”. For all Pāli quotations in this chapter, the existing English translations are given in
110 Ditrich
The formulaic refrain occurs also after the three other sections – the contem-
plation of feelings (vedanānupassanā),6 the mind (cittānupassanā),7 and
mental objects (dhammānupassanā);8 in total it is attested in 21 places in the
sutta, suggesting very clearly that these three modes were an important or per-
haps even essential aspect of the practice of the four satipaṭṭhānas.
The three terms appear also connected asyndetically (without vā) in close
or distant collocation; e.g. in the Dhammānupassīsutta, when describing how
six dhammas are to be abandoned in order to practice contemplation of the
body, feelings, the mind and mental objects.
Cha bhikkhave dhamme appahāya abhabbo ajjhattaṃ kāye kāyanupassī

viharituṃ … pe … bahiddhā kāye … ajjhattabahiddhā kāye … pe …
vedanāsu … ajjhattaṃ vedanāsu … bahiddhā vedanāsu …
ajjhattabahiddhā vedanāsu … citte … ajjhattaṃ citte … bahiddhā citte …
pe … ajjhattabahiddhā citte … dhammesu … ajjhattaṃ dhammesu …
bahiddhā dhammesu … ajjhattabahiddhā dhammesu …9
The triad appears in the Nikāyas consistently only in the context of meditation
instructions, as three different modes or ways of contemplation in the practice
of the satipaṭṭhanas. Since the suttas discussing the four satipaṭṭhānas seem to
be prescriptive, it means that the monks were instructed to contemplate in one
or another of the three modes (vā). In the Satipaṭṭhānasutta the terms
ajjhattaṃ, bahiddhā and ajjhattabahiddhā are linked by the conjunction “or”
(vā); there is no clear indication that they refer to three successive stages of
contemplation respectively, to be practiced in progression, starting with the
internal, followed by the external and then both the internal and external
mode, which has been suggested by some modern interpreters.10 Rather the
the footnotes (usually from Pāli Text Society versions). However, when those translations
are problematic in relation to the terms and topics discussed, alternative options are
suggested.
6 MN I 59; DN II 298.
7 MN I 59; DN II 299.
8 MN I 60–62; DN II 301–304, 313–314; this refrain follows each of the sections comprising
dhammānupassanā, i.e. five hindrances, five aggregates, six sense-spheres, seven factors
of enlightenment and four noble truths.
9 AN III 450; trans. by Bodhi, 2012: 988: “Bhikkhus, without having abandoned six things,
one is incapable of contemplating the body in the body internally … externally … both
internally and externally … contemplating feelings in feelings … mind in mind … phe-
nomena in phenomena … internally … externally … both internally and externally”.
10 Anālayo, 2006a: 94; cf. Gethin, 2001: 54.
implication seems to be that each mode is a distinctive and valid way of con-
templation in its own right.
Neither the Satipaṭṭhānasutta nor other suttas in the Nikāyas explain the
meaning of the three terms nor do they inform on their actual practice, espe-
cially of the third mode. To my knowledge, there is only one exception in the
Nikāyas, in the Janavasabhasutta:11
Ajjhattaṃ kāye kāyānupassī viharanto tattha sammā samādhiyati, sammā

vippasīdati. So tattha sammā samāhito sammā vippasanno bahiddhā
para-kāye ñāṇa-dassanaṃ abhinibbatteti …12
Here the term ajjhattaṃ is not explained, whereas bahiddhā is explained as

referring to the body of another (para-kāye).13 It is noteworthy that this is the
only sutta in the Suttapiṭaka that gives instructions on contemplation, yet the
terms ajjhattaṃ and bahiddhā occur, atypically, in a pair and not in a group
of three (together with the compound ajjhattabahiddhā). This may indicate
that the sutta reflects a different layer of textual transmission; in compari-
son, the relevant section from the Chinese (Dharmaguptaka) parallel to the
Janavasabhasutta also interprets the external mode as referring to another and
includes only two modes of contemplation.14
2.1.2 ajjhattaṃ and bahiddhā

Throughout the Tipiṭaka, the terms ajjhattaṃ and bahiddhā appear most fre-
quently in close or distant collocation, often as a pair linked asyndetically
or with the correlative conjunction vā. They function as complementary
antonyms, i.e. externally versus internally, usually rendered into English as
11 This sutta is referred to by most scholars discussing the interpretations of the internal and
external; e.g. Gethin, 2001: 54; Schmithausen, 2012: 292.
12 DN II 216; trans. by Walshe, 1995: 298: “As he thus dwells contemplating his own body as
body, he becomes perfectly concentrated and perfectly serene. Being thus calm and
serene, he gains knowledge and vision externally of the bodies of others”. The translation
adds “his own” body which is not in the Pāli original (ajjhattaṃ kāye kāyānupassī); the
translation suggested here is to use the word “internally”.
13 Explained in the same way for the other three satipaṭṭhānas (DN II 216).
14 Dīrghāgama fascicle 5, T.1.1: 34c25–35a4; trans.: “There being internal body contempla-
tion, knowledge of other bodies arises. There being internal feeling contemplation,
knowledge of other feelings arises. There being internal mind contemplation, knowledge
of other minds arises. There being internal mind-object contemplation, knowledge of
other mind-objects arises”. I am grateful to Rod Bucknell for providing this reference and
translation from the Chinese. Cf. Schmithausen, 2012: 292, note 9.
112 Ditrich
ajjhattaṃ “externally” and bahiddhā “internally,” or “oneself” and “others;”

no clarification for the particular choices in translating the terms is given by
translators.15 In the Nikāyas, the pair repeatedly occurs in the context of expli-
cating or describing various aspects of the teachings such as the aggregates
and the hindrances, discussing their external and (or) internal aspects; e.g. in
the Mahāpuṇṇamasutta:
Yaṃ kiñci, bhikkhu, rūpaṃ atītānāgatapaccuppannaṃ ajjhattaṃ vā

bahiddhā vā, oḷārikaṃ vā sukhumaṃ vā hīnaṃ vā paṇītaṃ vā yaṃ dūre
santike vā, ayaṃ rūpakkhandho.16
This sutta explicates, among other attributes, two aspects for each of the five
aggregates, internal or external (ajjhattaṃ vā bahiddhā vā) but does not talk
about their contemplation. Only in the context of contemplation is the third
mode (ajjhattabahiddhā) given as well; for example, in the Satipaṭṭhānasutta,
in the section on dhammas,17 the contemplation of the arising and passing
away of each aggregate is to be conducted in one of the three modes, as reiter-
ated in the refrain.
Similarly, in the Bojjhaṅgasaṃyutta the five hindrances are described as
internal and external:
katamo ca, bhikkhave, pariyāyo, yaṃ pariyāyaṃ āgamma pañca nīvaraṇā

dasa honti? yadapi, bhikkhave, ajjhattaṃ kāmacchando tadapi
nīvaraṇaṃ, yadapi bahiddhā kāmacchando tadapi nīvaraṇaṃ.
‘kāmacchandanīvaraṇan’ti iti hidaṃ uddesaṃ gacchati. tadamināpetaṃ
pariyāyena dvayaṃ hoti.18
15 The translations “oneself” and “others” are more common rendering for the pair whereas
“externally” and “internally” are used for the triad; e.g. in the English translation of the
Satipaṭṭhānasutta (DN II 290–315) by Walshe, 1995: 335–350, the triad is translated
ajjhattaṃ “internally,” bahiddhā “externally,” and ajjhattabahiddhā “internally and exter-
nally;” whereas in the same Nikāya, in the Sangītisutta (DN III 249) the pair ajjhattaṃ and
bahiddhā is translated as “yourself” and “others” respectively (ibid., p. 500).
16 MN III 19; trans. by Ñāṇanamoli and Bodhi, 1995: 888: “Bhikkhu, any kind of material form
whatever, whether past, future, or present, internal or external, gross or subtle, inferior or
superior, far or near – this is the material form aggregate”.
17 MN I 60–63.
18 SN V 110; trans. by Bodhi, 2000: 1603: “And what, bhikkhus, is the method of exposition by
means of which the five hindrances become ten?” Whatever sensual desire there is for the
internal is a hindrance; whatever sensual desire there is for the external is also a hin-
But when instructions on the contemplation of the hindrances are given these
are to be contemplated in one or another of the three modes.19 The pair
ajjhattaṃ and bahiddhā in the Nikāyas most commonly refers to a range of
pairs in opposition, such as internal-external, inward-outward, inside-outside.20
The English rendering of the terms as pairs of the opposites personal–imper-
sonal or oneself-others seems to draw mainly on the interpretation of the
aṭṭhakathā texts, to be discussed below (2.2). It has to be noted here that since
the pair and the triad occur in specifically different contexts, the semantic
ranges inferred from the usage of the pair (ajjhattaṃ, bahiddhā) should not be
simply, and without caution, inferred and applied to the interpretations of the
meaning of the triad (ajjhattaṃ, bahiddhā, ajjhattabahiddhā), especially of the
compound ajjhattabahiddhā which seems to refer to some other, a specific
mode or method of contemplation.
2.1.3 bahiddhā
The term bahiddhā is most frequently attested in a pair with ajjhattaṃ. When
bahiddhā occurs on its own, it usually refers to things being external, outside,
or aside.21 Another common occurrence of bahiddhā is in conjunction with ito,
referring to those people that are outside the Buddhist teachings or commu-
nity (ito bahiddhā); e.g. in the Brahmajālasutta:
Ye hi keci, bhikkhave, samaṇā vā brāhmaṇā vā sassata-vādā sassataṃ

attānañ ca lokañ ca paññāpenti, sabbe te imeh’ eva catuhi vatthūhi etesaṃ
vā aññatarena, n’ atthi ito bahiddhā.22
2.1.4 ajjhattaṃ
The term ajjhattaṃ occurs more frequently on its own than bahiddhā does. It
usually signifies “inwardly, internally,” often referring to states of mind; e.g.
drance. Thus what is spoken of concisely as the hindrance of sensual desire becomes, by
this method of exposition, twofold”.
19 For example, MN I 60–63.
20 Cf. PED s.v. ajjhatta.
21 For example, in the Janavasabhasutta: … na cassa bahiddhā ghoso niccharati… “… its
sound does not carry outside …” (DN II 211).
22 DN I 16; trans. by Walshe, 1995: 75: “And whatever ascetics or Brahmins are Eternalists and
proclaim the eternity of the self and the world, they do so on one or other of these four
grounds. There is no other way”. Cf. DN II 151, 283; AN III 206, 372; AN IV 25, 27, 136; MN I
323; MN II 121, 122; Nidd I 249, 397; SNII 133; SN V 229, 230; Th I 141.
114 Ditrich
ajjhattaṃ kathaṃkathī “internally uncertain,”23 or ajjhattaṃ vūpasantacitto

“inwardly peaceful,”24 or ajjhattaṃ abyāseka-sukhaṃ paṭisaṃvedeti “… he
experiences inwardly an unimpaired happiness”.25 When attested on its own,
ajjhattaṃ appears often in collocation with atthi, specifically referring to the
presence or absence of various mental factors. The phrase atthi me ajjhattaṃ is
commonly translated as “I have in myself,” pointing to an internal presence;
e.g. in the Satipaṭṭhānasutta ajjhataṃ occurs in this sense in two sections con-
cerning the contemplation of dhammas (dhammānupassanā) – the five
hindrances (nīvaraṇapabbam) and the seven factors of enlightenment
(bojjhaṅgapabbam). For each of the five hindrances,26 the sutta instructs that
he (i.e. a monk) contemplates whether the hindrance in question is present in
him (atthi me ajjhattaṃ) or not (natthi me ajjhattaṃ);27 e.g. concerning the
hindrance of sensual desire:
idha, bhikkhave, bhikkhu santaṃ vā ajjhattaṃ kāmacchandaṃ ‘atthi me

ajjhattaṃ kāmacchando’ti pajānāti, … asantaṃ vā ajjhattaṃ
kāmacchandaṃ ‘natthi me ajjhattaṃ kāmacchando’ti pajānāti …28
The entire section on hindrances is followed by the refrain, reiterating that

they are to be contemplated either internally or externally or both internally
and externally.29 This implies that a monk knows that a particular hindrance is
present in him (internally) and it is to be contemplated in one of the three
modes; however, the sutta does not inform – and it seems that it was not its
purpose in the first place – about the actual practice, especially in the case of
the third mode (i.e. he knows a hindrance which is in him internally both inter-
nally and externally).
23 MN I 8.
24 DN III 49.
25 MN I 181.
26 Namely, sensual desire (kāmacchanda), aversion (byāpāda), sloth and torpor (thīna
middha), restlessness and worry (uddhaccakukkucca) and doubt (vicikicchā).
27 Then the sutta continues instructions on how an unarisen hindrance can arise, an arisen
hindrance can be removed and the future arising of the removed hindrance can be pre-
vented.
28 MN I 60; trans. by Ñāṇanamoli and Bodhi, 1995: 151: “Here, there being sensual desire in
him, a bhikkhu understands: ‘There is sensual desire in me’; or there being no sensual
desire in him, he understands: ‘There is no sensual desire in me”.
29 iti ajjhattaṃ vā dhammesu dhammānupassī viharati, bahiddhā vā dhammesu dhammānu
passī viharati, ajjhattabahiddhā vā dhammesu dhammānupassī viharati (MN I 60, DN II
301).
Similarly, ajjhataṃ occurs on its own under the contemplation of the seven
factors of enlightenment (bojjhaṅgapabbam)30 where for each factor a monk
is instructed to know whether the factor in question is present in him (atthi me
ajjhattaṃ) or not (natthi me ajjhattaṃ).31 Once again, each of the seven factors,
perceived as internal phenomena, are to be contemplated in one or another of
the three modes, as instructed in the refrain following this section.
2.1.5 ajjhattikabāhira
The two stems also appear frequently in the adjectival forms ajjhattika
“internal” and bāhira “external”. They are attested either on their own or
in a dvandva compound, in each case with specific meanings and usages of
the terms, depending on the context. When compounded in a dvandva they
always refer to the six internal and external sense-spheres (āyatana);32 e.g. in
the Satipaṭṭhānasutta the compound ajjhattikabāhira occurs in the section
on sense-spheres (āyatanapabbaṃ), referring to the six internal and external
sense-spheres (āyatana) respectively. The sutta lists all the six senses and their
objects (cakku and rūpa, etc.) and the fetters arising dependent on them; e.g.
for the eye sense:
Kathañ-ca bhikkhave bhikkhu dhammesu dhammānupassī viharati chasu

ajjhattikabāhiresu āyatanesu: Idha bhikkhave bhikkhu cakkhuñca pajānāti
rūpe ca pajānāti, yañca tadubhayaṃ paṭicca uppajjati saṃyojanaṃ tañca
pajānāti …33
The Satipaṭṭhānasutta does not directly explain the meaning of the two adjec-
tives; however, ajjhattika “internal” is commonly interpreted as referring to the
30 Namely, mindfulness (sati), investigation of dhamma (dhammavicaya), energy (viriya),

joy (pīti), tranquillity (passaddhi), concentration (samādhi), equanimity (upekkhā).
31 For example: idha, bhikkhave, bhikkhu santaṃ vā ajjhattaṃ satisambojjhaṅgaṃ ‘atthi me
ajjhattaṃ satisambojjhaṅgo’ti pajānāti, asantaṃ vā ajjhattaṃ satisambojjhaṅgaṃ ‘natthi
me ajjhattaṃ satisambojjhaṅgo’ti pajānāti (MN I 62, DN II 304); trans. by Ñāṇanamoli and
Bodhi, 1995: 153: “Here, there being the mindfulness enlightenment factor in him, a bhik-
khu understands: ‘There is the mindfulness enlightenment factor in me’; or there being no
mindfulness enlightenment factor in him, he understands: ‘There is no mindfulness
enlightenment factor in me’”.
32 DN II 302–303; DN III 102; AN V 109; MN I 61, 190, 191; MN III 32, 63; Nidd I 430, 431, 441.
33 MN I 61, D II 302; trans. by Ñāṇanamoli and Bodhi, 1995: 152–153: “And how does a bhikkhu
abide contemplating mind-objects as mind-objects in terms of the six internal and exter-
nal spheres? Here a bhikkhu understands the eye, he understands forms and he under-
stands the fetter that arises dependent on both …”.
116 Ditrich
six senses and bāhira “external” to their respective objects.34 In the case of the
sixth sense-sphere, the internal sense-sphere is the mind (mano) and the
external sense-spheres are dhammas which are the objects arising through the
mind-door (mano).35 The sutta instructs that the entire set of sense-spheres –
internal (the senses) and external (objects of the senses) – are to be
contemplated either internally, or externally, or both internally and externally;36
e.g. in the case of the sixth sense-sphere, both the internal sense, the mind
(mano), and its corresponding external sense-sphere, phenomena (dhamma),
are to be contemplated either internally, or externally, or internally and
externally.
When the two adjectives are not compounded but occur in close or distant
collocation they are usually used as attributes describing or explaining various
aspects of the phenomena; e.g. in the Mahāhatthipadopamasutta, the adjec-
tives refer to two aspects of the element of earth as internal or external:
pathavīdhātu siyā ajjhattikā, siyā bāhirā “The earth element may be internal,
may be external”.37
2.1.6 ajjhattika
The adjective ajjhattika always occurs in close or distant collocation with
bāhira; e.g.:
Aṭṭhārasa kho pan’ imāni bhikkhave taṇhāvicaritāni ajjhattikassa upādāya,

aṭṭharasa taṇhāvicaritāni bāhirassa upādāya.38
34 Namely, internal: eye (cakkhu), ear (sota), nose (ghāna), tongue (jivhā), body (kāya), mind
(mano); external: visible form (rūpa), sound (sadda), odor (gandha), flavor (rasa), tangi-
ble object (phoṭṭhabba), phenomena (dhamma); cf. DN III 243; Ps I 287.
35 The later subcommentary, the Līnatthapakāsanā Ṭīkā by Dhammapala, further explicates
that the sixth internal sense-base is the bhavaṅga, the so-called life-continuum: chaṭṭassa
pana bhavaṅgamanasaṅkhāto manāyatanekadeso uppatti dvāraṃ (Soma, 1981: 133). Mod-
ern interpreters such as Ñāṇamoli, 1980: 159, suggest that ajjhatikāyatana represents the
“organisation of experience” and bahiddhāyatana the “experience as organised”. Cf.
Anālayo, 2006a: 215.
36 iti ajjhattaṃ vā dhammesu dhammānupassī viharati, bahiddhā vā dhammesu dhammānu
passī viharati, ajjhattabahiddhā vā dhammesu dhammānupassī viharati (MN I 61).
37 MN I 185.
38 AN II 212; trans. by Bodhi, 2012: 586: “There are, bhikkhus, these eighteen kinds of craving
related to the internal and eighteen kinds of craving related to the external”.
2.1.7 bāhira
The adjective bāhira, when attested on its own, does not refer to the senses but
mostly refers to “external, outside, outward;” e.g. bāhiraṃ cammakāyaṃ “outer
hide;”39 bāhirāni vā kammantāni paṭivekkhituṃ “or attending to outside
works;”40 bāhirehi paccatthikehi paccāmittehi “by external adversaries and
enemies”.41 The adjectival stem bāhira appears very frequently in compounds
(e.g. bāhira-assāda, bāhira-samaya, santara-bāhira) and in secondary nominal
derivations (e.g. bāhiraka, bāhirima). As an adverb bāhire, it usually refers to
those who are outside the Buddhist order or teachings.
Some semantic inferences may be drawn for the context dependent use of
the terms ajjhattaṃ and bahiddhā in the Nikāyas, studied at the syntagmatic
level. The usage of the triad is very specific, with a very restricted semantic
range: it is attested only in the instances where instructions are given on the
three modes of contemplations and this is also the only context for the occur-
rence of the compound ajjhattabahiddhā. Similarly, the use of the compound
ajjhattikabāhira is restricted only to the sense-spheres. The use of the pair
(ajjhattaṃ and bahiddhā) is less restricted, appearing in a range of contexts,
explicating various aspects of the teachings. The stylistic patterns of formulaic
repetitions of the terms – in dvandva, asyndeta or constructions with the con-
junction vā – seem to reflect the style of the ancient Indian tradition of oral
transmission of sacred texts, particularly the transmission of the Vedas where
the specific stylistic and linguistic patterns in the Vedic language act as semi-
otic markers, related to the religious beliefs and ritualistic acts.42 The Indian
Buddhist oral transmission may have adopted the already established stylistic
and linguistic patterns and markers, such as the usage of dvandva compounds,
for indicating specific contexts such as the modes of contemplation, especially
in relation to the four satipaṭṭhānas.43
Although the study of lexical collocational patterns informs on the use of
the two terms in distinct contexts, most evidently in the case of the triad, the
Nikāyas do not explicate the meanings of the terms in question, nor do they
39 MN III 275.
40 AN I 69.
41 AN IV 106.
42 Ditrich, 2010; and forthcoming.
43 Although Buddhist oral literature may not have fully followed the rather precise conven-
tions of the Vedic oral transmission and exhibits, through variations and divergences,
many characteristics similar to ancient epic transmissions (as pointed out by Cousins,
1983: 1–11), it is likely that many mnemonic formulae, stylistic and linguistic devices were
adopted and adapted by early Buddhists from the already well established Vedic tradition.
118 Ditrich
provide clarification of or comment upon their significance in the practice of

the satipaṭṭhānas.
2.2 Ajjhattaṃ, bahiddhā and ajjhattabahiddhā in the Pāli

Commentaries
The interpretations transmitted by Buddhaghosa – presumably the fifth
century44 author of the commentaries on the four primary Nikāyas and
the Visuddhimagga – became the standard reading of the terms ajjhattaṃ,
bahiddhā and ajjhattabahiddhā in Theravāda doctrine. These are also reflected
in contemporary interpretations of the practice of the satipaṭṭhānas.45 In the
Papañcasūdanī, the commentary on the Majjhimanikāya, Buddhaghosa com-
ments upon the section on breathing from the Satipaṭṭhānasutta:46
iti ajjhattaṃ vāti evaṃ attano vā assāsapassāsakāye kāyānupassī viharati.

bahiddhā vāti parassa vā assāsapassāsakāye. ajjhattabahiddhā vāti kālena
attano, kālena parassa assāsapassāsakāye … ekasmiṃ kāle panidaṃ
ubhayaṃ na labbhati.47
The Papañcasūdanī as well as the other commentaries on the Nikāyas consis-

tently explicate the three terms in the same manner: ajjhattaṃ refers to oneself
(attano), bahiddhā to another (parassa), and ajjhattabahiddhā alternatively to
oneself and another (kālena attano, kālena parassa),48 the latter actually being
only a reiteration of the first two modes to be applied alternatively. It is stated
that both modes cannot occur at the same time (ekasmiṃ kāle panidaṃ
ubhayaṃ na labbhati).
The terms are commented upon in a similar manner in the Visuddhimagga;
e.g. in the description of the aggregates (khandhaniddesa):
44 von Hinüber, 1996: 103, proposes 370–450 CE.

45 For example, in the writings of Mahāsi Sayādaw (stemming partly from the teachings of U
Narāda) who was instrumental in popularizing vipassanā in the twentieth century.
46 The passage being commented upon is: Iti ajjhattaṃ vā kāye kāyānupassī viharati
bahiddhā vā kāye kāyānupassī viharati, ajjhattabahiddhā vā kāye kāyānupassī viharati
(MN I 56).
47 Ps I 249; trans. by Soma, 1981: 51–52: “He dwells in contemplation of the body in his own
respiration body, or he dwells in contemplation of the body in another’s respiration body,
or dwells at one time in his own and at another [time] in another’s respiration body …
both cannot occur at once”. The Sumaṅgalavilāsinī, the commentary on the Dīghanikāya,
gives identical commentary (Sv III 765).
48 Ps I 252.
Ajjhattabahiddhā-bhedo vuttanayo eva. Api ca idha niyakajjhattam pi

ajjhattaṃ, parapuggalikam pi ca bahiddhā ti veditabbaṃ.49
Ajjhattaṃ is explained as niyakajjhattam “one’s own,”50 bahiddhā as parapug-

galika “another person”. The compound ajjhattabahiddhā receives a comment
in the Visuddhimagga only in one instance:
Catutthattike attano khandhe gahetvā āraddhā vipassanā-paññā

ajjhattābhinivesā, parassa khandhe bāhiraṃ vā anindriyabaddharūpaṃ
gahetvā āraddhā bahiddhābhinivesā; ubhayaṃ gahetvā āraddhā
ajjhattabahiddhābhinivesā ti evaṃ ajjhattābhinivesādivasena tividhā.51
In distinction to the commentaries on the Nikāyas which interpret the third

mode as an alternation between the first two (kālena attano, kālena parassa),
here it is referred to as ubhayaṃ “both,” without specifying whether the con-
templation in both modes is to take place alternatively (which is more likely)
or not.
Buddhaghosa’s interpretation of the three terms has been generally accepted
in the Theravāda tradition and has remained the standard interpretation in a
large section of secondary modern works on the practice of the satipaṭṭhānas.
There have been several attempts by modern scholars and practitioners to
interpret ajjhataṃ and bahiddhā, especially in terms of meditation practice.
Most of them follow Buddhaghosa, occasionally with modifications or drawing
from the interpretations in the Abhidhamma.
3 Ajjhattaṃ, bahiddhā, and ajjhattabahiddhā in the Abhidhamma
In contrast to the Nikāyas, the Abhidhamma provides interpretations for

ajjhataṃ as “oneself” and bahiddhā as “another,” however, there is no clear
explanation of the third mode, i.e. ajjatthabahiddhā. The analysis of collo-
cation of the three terms shows that they follow the same main typological
49 Vism 473; trans. by Ñāṇanamoli, 1994: 537: “It is internal in the sense of one’s own that
should be understood here as internal and that of another person as external”.
50 Cf. As 46.
51 Vism 440; trans. by Ñāṇanamoli, 1994: 484: “In the fourth triad, insight-understanding ini-
tiated by apprehending one’s own aggregates is interpreting the internal. That initiated by
apprehending another’s aggregates or external materiality not bound up with the facul-
ties, is interpreting the external. That initiated by apprehending both is interpreting the
internal and external. So it is of three kinds as interpreting the internal, and so on”.
120 Ditrich
and semantic features as identified in the Nikāyas and overviewed above

(2.1.1–2.1.7). Nearly all occurrences of the three terms in collocation are in
the Dhammasaṅgaṇi and Vibhaṅga which are examined here in conjunc-
tion with their commentaries, i.e. the Atthasālinī and the Sammohavinodanī
respectively.52
3.1 The Dhammasaṅgaṇi and its Commentaries

3.1.1 Dhammasaṅgaṇi
In the Dhammasaṅgaṇi the words referring to internal and external (ajjhatta,
bahiddhā, ajjhattika, bāhira) generally follow the same typology identified for
the Nikāyas (2.1.1–2.1.7). The mātikā section, which provides the classification
of dhammas that are discussed and analyzed in the Abhidhamma texts,
includes the three terms in collocation under the section of 22 threefold desig-
nations (tikamātikā), listed in two sets of triplets (tikas):
Ajjhattā dhammā, bahiddhā dhammā, ajjhattabahiddhā dhammā.

Ajjhattārammaṇā dhammā, bahiddhārammaṇā dhammā,
ajjhattabahiddhārammaṇā dhammā.53
The dhammas are classified as “internal” (ajjhattā), “external” and “internal

and external” (ajjhattabahiddhā); and “dhammas having internal object(s),”
“dhammas having external object(s),” and “dhammas having internal and
external object(s)”.54 The three terms are listed (as adjectives) at the very
beginning of the Dhammasaṅgaṇi, among the triplets (tika) in the mātikā, in
reference to all dhammas.55 In contrast, in the Nikāyas the three terms (as
adverbs) occur in collocation exclusively in passages giving instructions on the
practice of the four satipaṭṭhānas, only referring to selected categories listed as
52 The largest number of attestations of the three terms in collocation is in the Vibhaṅga,
followed by the Dhammasaṅgaṇi, and a very few instances in the Dhātukathā and the
Paṭṭhāna.
53 Dhs 2.
54 Trans. by Rhys Davids, 2012: M2: “States that belong to one’s self; are external to one’s self;
are belonging or external to one’s self. States that have for an object one’s self; an object
external to one’s self; an object that is both”. This translation is based on paragraphs 1044–
1049 (Dhs 250).
55 Frauwallner, 1964: 67, comments on the list of the attributes (“Eigenschaften Mātṛkāḥ”) as
being artificially contrived, and claims that ajjhattabahiddhā was added later as the third
aspect to the original pair. This explanation is problematic in relation to the evidence in
the suttas (see 2.1) as well as in the Vibhaṅga where the group of the three terms (as
adverbs) is consistently incorporated into the texts on the four satipaṭṭhānas.
objects of contemplation (e.g. the sixteen states of the mind, five hindrances,
five aggregates, seven factors of enlightenment etc.),56 indicating that these
were viewed as either essential, most relevant or sufficient for the develop-
ment of meditation practice. It is noteworthy that out of the twenty-two
triplets listed in the mātikā there is only one triplet (ajjhattā, bahiddhā,
ajjhattabahiddhā) that occurs adverbially, as the three modes of contempla-
tion, in the refrain of the Satipaṭṭhānasutta, indicating that for some reason it
was viewed as an essential aspect of the satipaṭṭhāna practice; this may be fur-
ther underlined by the positioning of the three modes at the forefront in the
beginning of the refrain in the Satipaṭṭhānavibhaṅga (Vibh 193–207) (dis-
cussed in 3.2.1 below).
The two triplets listed in the mātikā are further explained in the Nikkhepa
kaṇḍa section of the Dhammasaṅgaṇi:
katame dhammā ajjhattā? ye dhammā tesaṃ tesaṃ sattānaṃ ajjhattaṃ

paccattaṃ niyatā57 pāṭipuggalikā upādiṇṇā, rūpaṃ, vedanā, saññā,
saṅkhārā, viññāṇaṃ – ime dhammā ajjhattā.
katame dhammā bahiddhā? ye dhammā tesaṃ tesaṃ parasattānaṃ
parapuggalānaṃ ajjhattaṃ paccattaṃ niyatā pāṭipuggalikā upādiṇṇā,
rūpaṃ, vedanā, saññā, saṅkhārā, viññāṇaṃ – ime dhammā bahiddhā.
katame dhammā ajjhattabahiddhā? tadubhayaṃ – ime dhammā
ajjhattabahiddhā.58
The internal dhammas are explained as those of one’s own (sattānaṃ) which
are internally, individually bound – the five khandas. By way of contrast, the
external dhammas are presented as those for other beings, other people
(parasattānaṃ parapuggalānaṃ), and dhammā ajjhattabahiddhā are simply
presented as those dhammas that are both (tadubhayaṃ). Although the
Dhammasaṅgaṇi, as distinct from the Nikāyas, comments on the meaning of
56 Cf. MN I 55–63; DN II 290–315.

57 Rhys Davids, 2012: 250, note 5, suggests reading niyakā instead, presumably based on niya-
kajjhatte as explained in As 46.
58 Dhs 187–188; trans. by Rhys Davids, 2012: 250: “Which are the states that are personal?
Those states which, for this or that being, are of the self, self-referable, one’s own, indi-
vidual, the issue of grasping; [in other words,] the five skandhas”. “Which are the states
that are external? Those states which, for this or that other being, for other individuals,
are of the self, self-referable, their own, individual, grasped at; [in other words,] the five
skandhas. Which are the states that are personal-external? States which are both [per-
sonal and external]”.
122 Ditrich
the internal and external dhammas, there is no clarification of dhammā

ajjhattabahiddhā apart from the very brief reference to both (tadubhayaṃ).59
The dhammas with the three types of objects are explained in the
Nikkhepakaṇḍa:
katame dhammā ajjhattārammaṇā? ajjhatte dhamme ārabbha ye uppajj-

anti cittacetasikā dhammā – ime dhammā ajjhattārammaṇā.
katame dhammā bahiddhārammaṇā? bahiddhā dhamme ārabbha ye
uppajjanti cittacetasikā dhammā – ime dhammā bahiddhārammaṇā.
katame dhammā ajjhattabahiddhārammaṇā? ajjhattabahiddhā dhamme
ārabbha ye uppajjanti cittacetasikā dhammā – ime dhammā
ajjhattabahiddhārammaṇā.60
The dhammas having an internal object (ajjhattārammaṇā) are those having

the mind and mental factors (cittacetasikā) arising with reference to the inter-
nal dhammas; the dhammas having an external object (bahiddhārammaṇā)
are those arising with reference to the external dhammas; and the dhammas
with ajjhattabahiddhārammaṇā objects are those arising with reference to the
internal and external dhammas.
Apart from the triplet, the pair ajjhattaṃ and bahiddhā appears frequently
in the Dhammasaṅgaṇi in close or distant collocation, functioning as com-
plementary antonyms, i.e. externally versus internally, similar to the Nikāyas
(2.1.2). They are attested in two contexts: in the Cittuppādakaṇḍa section
(rūpāvacarakusalaṃ), where kusala dhammas are discussed in the context
of the jhānas, they repeatedly occur in the phrase … ajjhattaṃ arūpasaññī
bahiddhā rūpāni passati parittāni …;61 and in the Rūpakaṇḍa section where
59 This explanation is adopted in the commentaries, e.g. Vism 440; As 46, 325.
60 Dhs 188; trans. by Rhys Davids, 2012: 251: “Which are the states that have an object of
thought concerning the self? Conscious states and their mental properties, which arise in
connection with states of the self. Which are the states that have an object of thought
concerning that which is external [to the self]? Conscious states and their mental proper-
ties, which arise in connection with that are external [to the self]. Which are the states
that have an object of thought concerning that which is ‘personal-external’? Conscious
states and their mental properties, which arise in connection with states that are ‘per-
sonal-external’”.
61 Dhs 42–53; trans. by Rhys Davids, 2012: 54: “When, that he may attain to the heavens of
Form, he cultivates the way [thereto], and unconscious of any part of corporeal self, but
seeing external objects to be limited …”. This translation which follows Buddhaghosa’s
comments (As 188, 189, 191) is problematic and a clearer rendering of ajjhattaṃ may be
simply “internally”.
they are listed among attributes of the elements, e.g. in the description of
the earth element: … kakkhaḷattaṃ kakkhaḷabhāvo ajjhattaṃ vā bahiddhā vā
upādiṇṇaṃ vā anupādiṇṇaṃ vā – idaṃ taṃ rūpaṃ pathavīdhātu.62 The term
bahiddhā has several occurrences on its own in the Nikkhepakaṇḍa,63 referring
to people outside the Buddhist teachings, whereas ajjhattaṃ does not appear
outside its collocation with bahiddhā.
The adjectives ajjhatika and bāhira are listed in the mātikā section among
twofold designations (dukamātikā): ajjhattikā dhammā, bāhirā dhammā.64
The two terms are attested only in the Rūpakaṇḍa section, in the same seman-
tic range as in the Nikāyas (2.1.5–2.1.7): they refer to internal and external
sense-spheres respectively,65 or appear among the pairs of attributes related to
rūpa.66 In distinction to the Nikāyas, ajjhatika and bāhira do not appear com-
pounded (ajjhattikabāhira) but only in asyndetic collocation.
3.1.2 Commentaries
In the last section of the Dhammasaṅgaṇi, in a supplement of commentarial
nature, called the Atthuddhāra67 or Aṭṭhakathākaṇḍa,68 the terms ajjhattā,
bahiddhā and ajjhattabahiddhā are further discussed:
anindriyabaddharūpañca nibbānañca ṭhapetvā, sabbe dhammā siyā

ajjhattā, siyā bahiddhā, siyā ajjhattabahiddhā. anindriyabaddharūpañca
nibbānañca bahiddhā.69
It is stated that form (rūpa) which is not bound up with the faculties, and
nibbāna are external whereas all other dhammas may be internal (siyā ajjhattā),
external (siyā bahiddhā), or internal and external (siyā ajjhattabahiddhā).
Then dhammas in relation to objects are explained:
62 Dhs 177; trans. by Rhys Davids, 2012: 222: “That which is hard, rough, hardness, rigidity,
whether it be of the self, or external, or grasped at, or not grasped at”.
63 For example, Dhs 183.
64 Dhs 5.
65 Dhs 129.
66 Dhs 125.
67 As 6, 38.
68 As 409, 422.
69 Dhs 241; trans. by Rhys Davids, 2012: 340: “With the exception of form which is not bound
up with faculties, and Nirvana, all states may be personal or external or personal-external.
[Material] form which is not bound up with faculties, and Nirvana, are both external”.
124 Ditrich
katame dhammā ajjhattārammaṇā? viññāṇañcāyatanaṃ,

nevasaññānāsaññāyatanaṃ – ime dhammā ajjhattārammaṇā.70
Dhammas having an internal object (ajjhattārammaṇā) are the sphere of infi-

nite consciousness (viññāṇañcāyatana) and the sphere of neither perception
nor non-perception (nevasaññānāsaññāyatana), whereas those with an exter-
nal object include the following:
katame dhammā bahiddhārammaṇā? rūpāvacaratikacatukkajjhānā

kusalato ca vipākato ca kiriyato ca, catutthassa jhānassa vipāko,
ākāsānañcāyatanaṃ, cattāro maggā apariyāpannā cattāri ca
sāmaññaphalāni – ime dhammā bahiddhārammaṇā.71
Dhammas with an external object (bahiddhārammaṇā) are the threefold and

fourfold jhānas related to the form realm (rūpāvacaratikacatukkajjhānā) (aris-
ing from wholesome, resultant, inoperative), the result of the fourth jhāna
(catutthassa jhānassa vipāko), the sphere of infinite space (ākāsānañcāyatana),
the four paths (cattāro maggā apariyāpannā),72 and the fruitions (sāmañña
phalāni). The text does not comment separately on the third component, i.e.
ajjhattabahiddhārammaṇā, but rather continues with a discussion of those
dhammas that have any one of the three objects:
rūpaṃ ṭhapetvā sabbeva kāmāvacarā kusalākusalāvyākatā dhammā,

rūpāvacaracatutthajhānaṃ kusalato ca kiriyato ca – ime dhammā siyā
ajjhattārammaṇā, siyā bahiddhārammaṇā, siyā
ajjhattabahiddhārammaṇā.73
70 Dhs 241; trans. by Rhys Davids, 2012: 340: “Which states have a personal [internal] object
of thought? The sphere of infinite consciousness and the sphere where there is neither
perception nor non-perception – these are states that have a personal object of thought”.
71 Dhs 241–242; trans. by Rhys Davids, 2012: 340: “Which states have an external object of
thought? The threefold and fourfold Jhāna relating to the heavens of Form, whether it
arise as good (karma), as result (of good karma), or as completed thought, also results of
Fourth Jhana, the sphere of infinite space, the four Paths that are the Unincluded and the
four Fruits of the life of the recluse: these states have an external object of thought”.
72 Apariyāpannā means here: not including the worlds of the senses, form and the formless.
73 Dhs 242; trans. by Rhys Davids, 2012: 340–341: “Excepting form, states, good, bad and inde-
terminate relating to the sensuous universe, and the fourth Jhāna relating to the worlds of
Form, whether it arise as good (karma), or as completed thought: all these may be either
a personal [internal] object of thought, or an external object of thought, or a personal-
external object of thought”.
Apart from form, all dhammas (wholesome, unwholesome and indetermi-

nate) related to the desire realm, and the fourth jhāna, related to the form
realm (wholesome and inoperative), may have an internal object, an external
object, or an internal and external object. The text continues:
ākiñcaññāyatanaṃ na vattabbaṃ ajjhattārammaṇanti pi,

bahiddhārammaṇanti pi, ajjhattabahiddhārammaṇanti pi-rūpañca
nibbānañca anārammaṇā.74
The section concludes that it should not be said that the sphere of nothingness
has an internal, external or both internal and external object.75 Form and
nibbāna are without an object.
To summarize, in the commentary on the first triplet, form (rūpa) and
nibbāna are external, all other dhammas may be internal, external, or inter-
nal and external. Apart from form, all dhammas related to the desire realm
and the fourth jhāna (related to the form realm) may have an internal, exter-
nal, or internal and external object. Dhammas having an internal object are
the sphere of infinite consciousness and the sphere of neither perception nor
non-perception. Dhammas having an external object are the threefold and
fourfold jhānas, the result of the fourth jhāna, the sphere of infinite space,
and the four paths and fruitions. Although the text does provide more details
on the meaning of internal and external, it does not give any new informa-
tion on ajjhattabahiddhā and ajjhattabahiddhārammaṇa dhammas; apart
from rūpa and nibbāna, all cittas and cetasikas may be ajjhattabahiddhā, and
all dhammas, apart from rūpa and the fourth jhāna, may have as an object
ajjhattabahiddhārammaṇa.
In the Atthasālinī, the commentary on the Dhammasaṅgaṇi by Buddha
ghosa,76 the comments on the first triplet (dhammā ajjhattā, dhammā
bahiddhā, dhammā ajjhattabahiddhā) focus almost entirely on ajjhatta:
Ajjhattattike evaṃ pavattamānā ‘mayaṃ attā ti gahaṇaṃ gamissāmā ti’

iminā viya adhippāyena attānaṃ adhikāraṃ katvā pavattā ti ajjhattā.
74 Dhs 242; trans. by Rhys Davids, 2012: 341: “But it is not proper to say that the sphere of
nothing whatsoever is all three. Form and Nirvana are without objects of thought”.
75 Cf. Ajjhattadhammāpagamamattatova ākiñcaññāyatanārammaṇassa ajjhattabhāvampi
bahiddhābhāvampi ajjhattabahiddhābhāvampi ananujānitvā ākiñcaññāyatanaṃ na
vattabbaṃ ajjhattārammaṇantipītiādi vuttaṃ (As 423).
76 Three commentaries on the Abhidhamma texts are attributed to Buddhaghosa:
Atthasālinī, Sammohavinodanī and Pañcappakaraṇāṭṭhakathā; cf. von Hinüber 1996: 149–
153; Norman, 1983: 120–130.
126 Ditrich
Ajjhattasaddo panāyaṃ gocarajjhatte niyakajjhatte ajjhattajjhatte visay-

ajjhatte ti catūsu atthesu dissati.77
Ajjhatta is here interpreted as referring to oneself and its four meanings

are listed: internal in field/sphere (gocarajjhatte), internal in oneself (niya-
kajjhatte), internal in internal (ajjhattajjhatte), internal in range/object
(visayajjhatte). These four terms are not attested in the Tipiṭaka but only in the
commentaries ascribed to Buddhaghosa. They are discussed in the Atthasālinī,
starting with gocarajjhatte78 exemplified by the mind focused on the sign of
concentration (samādhinimitte),79 inwardly rapt and composed (ajjhattarato
samāhito).80 The term niyakajjhatte is explained as internal to oneself, illus-
trated by inner tranquility (ajjhattaṃ sampasādanaṃ)81 when contemplating
dhammas internally; this term is usually translated as subjective, personal or
referring to oneself.82 The third term, ajjhattajjhatte is explained as referring to
the six sense organs (cha ajjhattikāni āyatanānī).83 The last term, visayajjhatte
is referred, in a sense of dominion (visayajjhatte issariyaṭṭhāne ti attho), to inner
emptiness (ajjhattaṃ suññataṃ),84 as achieved by the Tathāgata by disregard-
77 As 46; trans. by Maung Tin, 2013: 60: “In the triplet of ‘personal,’ this word refers to states
which occur after making a locus of selves as though with the understanding ‘we shall
consider or take things thus existing to be we ourselves’. The word ‘personal’ (ajjhatta) has
a fourfold content, namely personal in field, in self-reference, (just) personal, personal in
range”.
78 Ten’ Ānanda, bhikkhunā tasmiṃ yeva purimasmiṃ samādhinimitte ajjhattam eva cittaṃ
saṇṭhapetabbaṃ. Ajjhattarato samāhito ti ādīsu hi ayaṃ gocarajjhatte dissati (As 46).
Trans. by Maung Tin, 2013: 60: “In such sentences as, ‘Ānanda, mind should be well
focussed by that bhikkhu as ajjhattam, namely, only in that symbol of concentration
which has been practised before;’ ‘inwardly rapt (ajjhattarato) and concentrated’ ajjhatta
means ‘personal in field’”.
79 Cf. samādhinimitta (MN I 249) is explained in the commentary (Ps II 292) as suññata
phalasamāpatti, fruition of arahatship; see also MN III 112.
80 Cf. DN II 107–108; SN V 263; AN IV 312; Dhp 101; Ud 64; Th 89.
81 ajjhattaṃ sampasādanaṃ occurs in numerous places in the Tipiṭaka in the context of
jhānas (e.g. DN I 37; DN I 74; MN I 21; SN II 211; AN I 53).
82 Ajjhattaṃ sampasādanaṃ ajjhattaṃ vā dhammesu dhammānupassī viharatī ti ādīsu niya-
kajjhatte (As 46). Trans. by Maung Tin, 2013: 60: “In such passages as, ‘He lives contemplat-
ing states, even among states which are pleasing as ajjhatta, ajjhatta means ‘subjective’”.
83 Cha ajjhattikāni āyatanānī ti ādīsu ajjhattajjhatte (As 46); trans. by Maung Tin, 2013: 60:
“[In] the six ajjhatika sense-organs, ajjhatta means ‘personal’”.
84 Cf. the Mahāsuññata Sutta (MN III 111) where ajjhattaṃ suññataṃ leads to the fruition of
arahatship. 
ing all signs and symbols (sabbanimittānaṃ amanasikārā).85 The commentary

concludes that internal dhammas are to be understood as those in the continu-
ity of oneself (attano santāne), pertaining to the individual (pāṭipuggalikā).86
While the Atthasālinī gives significant attention to the term ajjhata,
the other two terms of the triplet are only briefly touched upon. It depicts
bahiddhā dhammā as those that are external (bāhirabhūtā), bound or
unbound by the faculties (indriyabaddhā vā anindriyabaddhā vā).87 For the
ajjhattabahiddhā dhammā it is only stated that these are states by virtue of
both (tadubhayavasena).88 The compound is attested again only once, in the
commentary on the Cittuppādakaṇḍa, in the context of jhāna attainments,
where the text mentions ajjhattabahiddhā in reference to the kasiṇas having
internal and external bases.89
The next triplet listed in the mātikā (i.e. ajjhattārammaṇā dhammā,
bahiddhārammaṇā dhammā, ajjhattabahiddhārammaṇā dhammā)90 is only
briefly commented upon in the Atthasālinī, saying that these three refer to
the dhammas which attend to the three kinds of objects.91 In the commentary
on the Atthuddhāra, dhammas having an internal object (ajjhattārammaṇā)
are said to be those arising with reference to objects of oneself (attano), the
dhammas having an external object (ajjhattārammaṇā) arise with refer-
85 Ayaṃ kho pan’ Ānanda vihāro Tathāgatena abhisambuddho yad idaṃ sabbanimittānaṃ
amanasikārā ajjhattaṃ suññataṃ upasampajja viharatī ti visayajjhatte issariyaṭṭhāne ti
attho. Phalasamāpattīhi buddhānaṃ issariyaṭṭhānaṃ nāma (As 46). Trans. by Maung Tin,
2013: 60: “In such passages as, ‘This, Ānanda, is the life fully attained by the Tathāgata, to
wit, that he, by disregarding all provocative signs and symbols, has reached the ajjhatta
Void and therein abides, ajjhatta means ‘range’ in the sense of ‘dominion’. The attainment
of Fruition is named the dominion of the Buddhas’”.
86 Tasmā attano santāne pavattā pāṭipuggalikā dhammā ajjhattā ti veditabbā (As 46). Trans.
by Maung Tin, 2013: 60–61: “Hence states occurring in one’s own continuity and pertain-
ing to each individual are to be understood as ‘personal’”.
87 Tato bāhibhūtā pana indriyabaddhā vā anindriyabaddhā vā bahiddhā nāma (As 46).
Trans. by Maung Tin, 2013: 61: “But states outside that personality, whether bound up with
the controlling faculties or not, are termed ‘external’”.
88 Tatiyapadaṃ tadubhayavasena vuttaṃ (As 46). Trans. by Maung Tin, 2013: 61: “The third
term is spoken by virtue of both”.
89 For example: Rūpāni passatī ti bahiddhā pi nīlakasiṇādirūpāni jhānacakkhunā passati.
Iminā ajjhattabahiddhāvatthukesu kasiṇesu jhānapaṭilābho dassito (As 191).
90 Dhs 2.
91 Anantarattiko te yeva tippakāre pi dhamme ārammaṇaṃ katvā pavattanavasena vutto (As
46). Trans. by Maung Tin, 2013: 61: “The immediately following triplet refers to states
(dhamme) occurring in the act of attending to just these three kinds of states (i.e. per-
sonal, external, externo-personal) as objects”.
128 Ditrich
ence to those of another (parassa), and those having internal and external
object (ajjhattabahiddhārammaṇā) with reference to both (tadubhaya).92
Ajjhattabahiddhārammaṇā is consistently interpreted in the commentary as
“both,” except in one instance when it is said that it refers to objects sometimes
internally, sometimes externally (kālena ajjhattaṃ kālena bahiddhā, pavattiyaṃ
ajjhattabahiddhārammaṇaṃ).93 In other commentaries by Buddhaghosa the
same interpretations are given.94
3.2 The Vibhaṅga and Sammohavinodanī

3.2.1 Vibhaṅga
Although the Theravāda tradition positions the Vibhaṅga after the Dhamma
saṅgaṇi,95 the chronology of the two texts is problematic and some scholars
suggest an alternative one, viewing the Vibhaṅga as the oldest among the
seven canonical texts of the Abhidhammapiṭaka.96 However, as in the case of
most ancient Indian texts, establishing a textual history of Abhidhamma texts
requires considerable circumspection in drawing conclusions since the texts
may have gone through several revisions and may reflect various historical lay-
ers.97 In relation to the three terms investigated here, it will be shown that the
92 For example: … ime terasa cittuppādā attano rūpādīni ārabbha pavattiyaṃ ajjhattāram
maṇā, parassa rūpādīsu pavattā bahiddhārammaṇā, tadubhayavasena ajjhattabahiddhā
rammaṇā (As 425).
93 This passage discusses powers (iddhi) manifesting internal and external objects: Iddhi
vidhacatutthaṃ kāyavasena cittaṃ cittavasena vā kāyaṃ pariṇāmanakāle bahiddhā
rammaṇaṃ attano kumārakavaṇṇādinimmānakāle ca sakāyacittānaṃ ārammaṇakaraṇato
ajjhattārammaṇaṃ, bahiddhā hatthi assādidassanakāle bahiddhārammaṇaṃ, kālena
ajjhattaṃ kālena bahiddhā, pavattiyaṃ ajjhattabahiddhārammaṇaṃ (As 426); cf. Vibh-a
375.
94 For example, in Ps I 249: ajjhattabahiddhā refers alternatively to oneself and to another
(kālena attano, kālena parassa); in Vism 440: ajjhattabahiddhā refers to both (ubhayaṃ).
95 Budddhaghosa lists the texts in the following order: Dhammasaṅgaṇi, Vibhaṅga, Dhātu
kathā, Puggalapaññatti, Kathāvatthu, Yamaka, and Paṭṭhāna (As 3, 21–23).
96 Frauwallner, 1971: 106–121, suggests that the Abhidhamma texts were written in the period
between 200 BCE and 200 CE; he proposes a relative chronology with the Vibhaṅga as the
earliest and the Dhammasaṅgaṇi as the latest text. The supposition by Frauwallner, 1964:
73–80, that the Vibhaṅga may have developed out of an earlier text which was also the
foundation of the Dharmaskandha of the Sarvāstivādins, is developed further by Bronk-
horst, 1985: 305–320, who argues, with rather convincing evidence, that “original” layers of
the Vibhaṅga may well stem back to before the Sūtrapiṭaka that we know at present.
97 See Kragh, 2002: 123–168; cf. Frauwallner, 1971: 69–12; Willemen, Dessein, and Cox, 1998:
10–16, 139–145; Wynne 2005: 35–70.
Vibhaṅga may be considered to reflect the oldest layers of the Abhidhamma

tradition.
Among the seven Abhidhamma texts, the terms ajjhattaṃ, bahiddhā and
ajjhattabahiddhā occur by far the most frequently in the Vibhaṅga. In its sev-
enth division, the Satipaṭṭhānavibhaṅga, the three modes of contemplation
are positioned at the forefront, at the very beginning of the suttanta part of the
text, delineating the four satipaṭṭhānas:
Cattāro satipaṭṭhānā: idha bhikkhu ajjhattaṃ kāye kāyānupassī viharati,

bahiddhā kāye kāyānupassī viharati, ajjhattabahiddhā kāye kāyānupassī
viharati …; ajjhattaṃ vedanāsu vedanānupassī viharati, bahiddhā …
ajjhattabahiddhā …; ajjhattaṃ citte cittānupassī viharati, bahiddhā …
ajjhattabahiddhā …; ajjhattaṃ dhammesu dhammānupassī viharati,
bahiddhā … ajjhattabahiddhā dhammesu dhammānupassī viharati …98
The text presents the four satipaṭṭhānas in five sections (as opposed to 23 in
the Satipaṭṭhānasutta99): contemplation of the body comprising one section
only, i.e. the repulsiveness of the body (paṭikkūlamanasikāra); contemplation
of feelings; contemplation of the mind; contemplation of phenomena com-
prising two sections, i.e. the hindrances (nīvaraṇa) and the factors of
enlightenment (bojjhaṅga). Under each section, the three terms are explained:
idha bhikkhu ajjhattaṃ kāyaṃ … paccavekkhati: atthi imasmiṃ kāye …

idha bhikkhu bahiddhā kāyaṃ … paccavekkhati: atthissa kāye …100
A monk contemplates the body internally (ajjhattaṃ) as “it is in this body”

(atthi imasmiṃ kāye), and externally (bahiddhā) “it is in the body of this one”
(atthissa kāye), and then the text lists the body parts. The distinction between
the two modes lies in the use of two different cases of the pronominal stem
idam: in the locative (imasmin) it can be interpreted in a deictic adjectival
meaning “in this [body],” usually translated, following the commentary,101 as
referring to one’s own body. The pronominal stem in the genitive (assa) is com-
monly interpreted as the third person pronoun, translated “of him” or “that
one’s”.102 The phrase atthissa is attested only in the Satipaṭṭhānavibhaṅga, in
98 Vibh 193.
99 MN I 55–63; DN II 290–315.
100 Vibh 193–194.
101 Vibh-a 219.
102 For example, the translation by Thiṭṭila, 2010: 251–252.
130 Ditrich
the kāyānupassanā and dhammānupassanā sections in the presentation of

contemplation of the body, the five hindrances and the seven factors of enlight-
enment in the external mode. It is later, in Buddhaghosa’s commentary, that
the reading of assa as referring to “other” (parassa)103 became established and
adopted; in the Satipaṭṭhānavibhaṅga there is no direct indication that assa
would refer to the “other” and hence, interpretation of this phrase requires fur-
ther exploration. One suggestion here is that the external mode (bahiddhā)
indicates that the meditator’s body (and not another’s) is contemplated exter-
nally, i.e. in a dissociated impersonal state or mode.104
The third mode is explained in the Vibhaṅga:
idha bhikkhu ajjhattabahiddhā kāyaṃ … paccavekkhati: atthi kāye …105
A monk contemplates the body internally and externally (ajjhattabahiddhā)

thus: “it is in the body” (atthi kāye), and then the text lists the body parts. This
is the only instance in the Tipiṭaka that represents the third mode as “there is”.
No text provides any further clarification of the third mode and hence, the
question how this mode was practiced remains an open one.106 Based on this
phrase from the Satipaṭṭhānavibhaṅga, the object in the third mode is contem-
plated without reference to anything internal or external, to one’s own or other
(i.e. without the usage of any pronoun); it may be viewed as experiencing the
body as such, which may suggest that contemplation is perhaps occurring
while experiencing anattā.
The Satipaṭṭhānavibhaṅga continues in the section on feelings (vedanā):
kathañca bhikkhu ajjhattaṃ vedanāsu vedanānupassī viharati? idha

bhikkhu sukhaṃ vedanaṃ vedayamāno: sukhaṃ vedanaṃ vedayāmīti
pajānāti … . bahiddhā vedanāsu vedanānupassī viharati? idha bhikkhu
sukhaṃ vedanaṃ vedayamānaṃ: sukhaṃ vedanaṃ vedayatīti pajānāti …
ajjhattabahiddhā vedanāsu vedanānupassī viharati? idha bhikkhu sukhaṃ
vedanaṃ: sukhā vedanā ti pajānāti …107
103 Vibh-a 219.

104 The author was told that this was a method of satipaṭṭhāna practiced in the mid-twenti-
eth century in Burma and handed down as such by the meditation teacher Premasiri of
Sri Lanka (personal interview, December 2014). It is said to go back to the predecessors of
Mahasi Sayadaw; however, no further information on the origins of this method has yet
been traced.
105 Vibh 194; cf. Burmese edition: atthi imasmiṃ kāye (Be 201).
106 Cf. Dharmaskanda (T.26.1537: 475c24–479b23), noted by Schmithausen, 2012: 293.
107 Vibh 195–196.
In the contemplation of feeling internally (ajjhattaṃ), a monk experienc-

ing feeling (vedayamāno) knows: “I feel pleasant feeling” (sukhaṃ vedanaṃ
vedayāmīti); externally (bahiddhā), he knows the feeling experienced (vedanaṃ
vedayamānaṃ): “he feels a pleasant feeling” (sukhaṃ vedanaṃ vedayatīti).
Here the modes ajjhattaṃ and bahiddhā are distinguished through the use of
the verb in the first person (vedayāmi) and the third person (vedayāti), and may
refer to oneself and to another respectively. Alternatively, the external mode
(bahiddhā) could be interpreted here (again) as an external, dissociated mode
of contemplation of feeling. In the case of the third mode (ajjhattabahiddhā),
the text simply says: “[there is] a pleasant feeling” (sukhā vedanā ti), indicat-
ing, as in the section on the kāyānupassanā, experience of the feelings as such,
without any reference which may, again, suggest that contemplation is occur-
ring while experiencing anattā.108
In a similar way the three terms are presented in the section on the mind
(citta):
kathañca bhikkhu ajjhattaṃ citte cittānupassī viharati? idha bhikkhu

sarāgaṃ vā cittaṃ: sarāgaṃ me cittan ti pajānāti … bahiddhā … sarāgaṃ
vāssa cittaṃ: sarāgamassa cittan ti pajānāti … ajjhattabahiddhā …
sarāgaṃ vā cittaṃ: sarāgaṃ cittan ti pajānāti.109
A monk contemplates the mind internally by knowing the lustful mind “my
mind is lustful” (sarāgaṃ me cittam); externally by contemplating the mind of
this one (or his) (assa cittam) “the mind of this (one) is lustful” (sarāgamassa
cittam); and internally and externally by contemplating “the mind is lustful”
(sarāgaṃ cittam), without relating it to anyone.
In the section on dhammas, only hindrances and factors of enlightenment
are discussed, starting with the hindrances:
Kathañ ca bhikkhu ajjhattaṃ dhammesu dhammānupassī viharati? Idha

bhikkhu santaṃ vā ajjhattaṃ kāmacchandaṃ: atthi me ajjhattaṃ
kāmacchando ti pajānāti …
Kathañ ca bhikkhu bahiddhā …? Idha bhikkhu santaṃ vāssa
kāmacchandaṃ: atthissa kāmacchando ti pajānāti …
108 Thiṭṭila (2010: 255) omits ajjhattabahiddhā in his translation of the sentence kathañca
bhikkhu ajjhattabahiddhā vedanāsu vedanānupassī viharati: “And how does a Bhikkhu
dwell contemplating feeling in feelings?”
109 Vibh 197–198.
132 Ditrich
Kathañ ca bhikkhu ajjhattabahiddhā …? Idha bhikkhu santaṃ vā

kāmacchandaṃ: atthi kāmacchando ti pajānāti …110
In internal (ajjhattaṃ) contemplation of the first hindrance, a monk, having

sensual desire internally (ajjhattaṃ kāmacchandaṃ), knows “there is in me
sensual desire” (atthi me ajjhattaṃ kāmacchando); in external (bahiddhā) con-
templation, a monk knowing sensual desire of this one (of him) (assa
kāmacchandaṃ) knows “there is sensual desire of this (one)” (atthissa
kāmacchando). In contemplation internally and externally (ajjhattabahiddhā),
a monk knows “there is sensual desire” (atthi kāmacchando), indicating an
impersonal mode. In this way the text continues through all the five hindrances
and the seven factors of enlightenment.
In the Vibhaṅga nearly all attestations of the three terms occur in the
Satipaṭṭhānavibhaṅga, in the context of the practice of the four satipaṭṭhānas,
thus consistently following the stylistic pattern for the three modes identified
in the Nikāyas (2.1.1). Consequently, concerning the presentation of the three
terms, the Vibhaṅga may reflect an older layer of the Abhidhamma literature
than the Dhammasaṅgaṇi where the three terms are listed as three different
approaches or perspectives for examination of all dhammas (i.e. mental states
and material qualities). Furthermore, in the Dhammasaṅgaṇi, the difference
between the internal and external is explained quite unambiguously: the inter-
nal dhammas are one’s own (sattānaṃ) and the external dhammas are those of
others (parasattānaṃ parapuggalānaṃ),111 an interpretation that is followed
up in the commentaries. In comparison, although the Vibhaṅga indicates that
ajjhataṃ may refer to oneself and bahiddhā to another, there is more space for
ambiguity and alternative interpretations; e.g. atthi imasmiṃ kāye … versus …
atthissa kāye … (Vibh 193–194) cannot be translated with any certainty as
meaning only “one’s own” and “another’s body;” it may be interpreted as two
ways of viewing one’s own body.
In the Vibhaṅga the third mode is presented differently than in other Pāli
canonical texts. The interpretation of ajjhattabahiddhā in the Satipaṭṭhā
navibhaṅga may indicate contemplation accompanied by anattā; e.g. atthi
kāmacchando indicates only a presence of this particular hindrance, with-
out reference to an individual. It is noteworthy that this interpretation in
the Vibhaṅga is not followed up by the commentaries. In comparison, the
Dhammasaṅgaṇi explains ajjhattabahiddhā as “both” (tadubhayaṃ), internal
110 Vibh 199–201.

111 Dhs 187–188.
and external;112 this is reflected in the commentarial literature. Concerning

these three terms, the Vibhaṅga seems to preserve in their presentation a dif-
ferent, perhaps older layer of textual transmission than the Dhammasaṅgaṇi.113
The three terms are also attested in other sections of the Vibhaṅga. In the
Khandhavibhaṅga section, they always occur in relation to the dhammas
being internal, external, and internal and external, and having internal, exter-
nal, and internal and external object(s);114 e.g. in the presentation of feelings
(vedanākkhandha):
tividhena vedanākkhandho: … atthi ajjhatto, atthi bahiddho, atthi ajjhat-

tabahiddho, atthi ajjhattārammaṇo, atthi bahiddhārammaṇo, atthi
ajjhattabahiddhārammaṇo …115
The stems referring to internal and external (ajjhatta, bahiddhā, ajjhat-

tika, bāhira) generally follow the same typological and semantic features
as identified for the Nikāyas (2.1.1–2.1.7). The pair ajjhattaṃ and bahiddhā
often occurs among other attributes describing the five khandhas,116 and the
two terms are explained in the same way as in the Dhammasaṅgaṇi; e.g. in
the Khandhavibhaṅga section the two aspects of material quality (rūpa) are
discussed:
tattha katamaṃ rūpaṃ ajjhattaṃ? yaṃ rūpaṃ tesaṃ tesaṃ sattānaṃ

ajjhattaṃ paccattaṃ niyakaṃ pāṭipuggalikaṃ upādinnaṃ …
tattha katamaṃ rūpaṃ bahiddhā? yaṃ rūpaṃ tesaṃ tesaṃ
parasattānaṃ parapuggalānaṃ ajjhattaṃ paccattaṃ niyakaṃ
pāṭipuggalikaṃ upādinnaṃ … .117
The internal rūpa is explained as one’s own (sattānaṃ), the external as for any
other beings, other people (parasattānaṃ parapuggalānaṃ).118 The semantic
context for the occurrences of the terms in pair is different to the context for
the triad which is consistently within the framework of contemplation.
112 Dhs 188.

113 However, it has to be added that both texts are, in their present form, nevertheless mutu-
ally related and dependent. Cf. Gethin, 1992: 163–164.
114 There are numerous attestation; e.g. Vibh 19–21, 26–29, 38–41, 55–62, 75, 92, 115, 311–315,
327–328.
115 Vibh 16–17; cf. Dhs 187–188.
116 See Vibh 1, 3, 5, 7, 10.
117 Vibh 2; Cf. Dhs 187–188.
118 Vibh 3, 5, 8, 10.
134 Ditrich
The adjectival stems ajjhatika and bāhira occur in the Khandhavibhaṅga

section among the pairs of attributes related to the khandhas.119 In distinction
to the Nikāyas, ajjhatika and bāhira are not attested in a compound
(ajjhattikabāhira) but only in collocation.
3.2.2 Sammohavinodanī
The commentary on the Vibhaṅga, the Sammohavinodanī, ascribed to
Buddhaghosa, gives similar interpretations of the three terms as those in
the Atthasālinī. The three terms are discussed in the commentary on the
Satipaṭṭhānavibhaṅga, starting with the three modes of contemplation of the
body:120
Ajjhattaṃ ti niyakajjhattaṃ adhipettaṃ. Tasmā ajjhattaṃ kāye ti attano

kāye ti attho.121
Bahiddhā kāye ti parassa kāye. Ajjhattabahiddhā kāye ti kālena attano
kāye, kālena parassa kāye.122
The internal mode is understood as niyakajjhatta which is discussed further in

the Atthasālinī.123 The phrase ajjhattaṃ kāye is explained as referring to one’s
own body (attano kāye), bahiddhā kāye to another’s body (parassa kāye),124 and
ajjhattabahiddhā alternatively to one’s own and another’s body (kālena attano
kāye, kālena parassa kāye).125
The phrase atthi kāye which is an explanation of the third mode of the body
contemplation (ajjhattabahiddhā kāyaṃ) in the Vibhaṅga,126 is commented
upon in the Sammohavinodanī:
atthi kāye ti idaṃ yasmā na ekantena attano kāyo, nāpi parass’eva kāyo
adhippeto, tasmā vuttaṃ.127
119 See Vibh 13, 67, 79, 82–85.

120 See Vibh 193.
121 Vibh-a 217.
122 Vibh-a 219.
123 As 46. Similarly, in the commentary on the section rūpāvacarakusala (Dhs 31), in the con-
text of the second jhāna, ajjhattaṃ is presented as niyakajjhattam; here Buddhaghosa
adds: vibhaṅge pana ajjhattaṃ paccattan ti ettakam eva vuttaṃ (As 169).
124 Cf. Vibh-a 261.
125 Cf. Ps I 249, 252, 270–273, 279–280; 286–287.
126 Vibh 194.
127 Vibh-a 261; trans. by Ñāṇamoli, 1996: 321: “There is in the body: this is said because it is not
exclusively his own body nor another’s body that is intended”.
The commentary states that neither one’s own body (attano kāyo) nor anoth-
er’s (parass’eva kāyo) is meant exclusively. Similarly, the Sammohavinodanī
comments on the section on feelings:
Ajjhattabahiddhā ti kālena attano kālena parassa vedanāsu cittaṃ

upasaṃharati. Imasmiṃ vāre yasmā neva attā na paro niyamito, tasmā
vedanāpariggahamattam eva dassetuṃ.128
The third mode is presented here as neither oneself (neva attā) nor another
(na paro). There is no comment in the Sammohavinodanī on the three modes
in the sections on the mind (citta) and phenomena (dhamma).
The interpretation saying for the third mode to be neither one’s own nor
another’s exclusively is, to my knowledge, found only in the Sammohavinodanī
and is not further discussed or followed up by any other Pāli text of that period.
In other commentaries ascribed to Buddhaghosa the usual interpretation of
the third mode is “sometimes one’s own, sometimes another’s,” or “both”.
Other comments on the three terms in the Sammohavinodanī are presented
in a similar way as in other commentaries by Buddhaghosa. In the commen-
tary on the Khandhavibhaṅga the aggregates are discussed, having internal
and external objects (ajjhattabahiddhārammaṇā) as being sometimes inter-
nal, sometimes external: kālena ajjhattaṃ kālena bahiddhā dhammesu evaṃ
pavattentassa ajjhattabahiddhārammaṇā.129 The adjectival stems ajjhatika
and bāhira are attested in the Sammohavinodanī in reference to internal and
external senses respectively,130 or occur among the pairs of attributes.131
To summarize then, several interpretations for the three terms are
given in the Abhidhamma. In the last section of the Dhammasaṅgaṇi, the
Aṭṭhakathākaṇḍa, the terms are related to different stages of high absorptions:
e.g. dhammas having an internal object are the sphere of infinite consciousness
and the sphere of neither perception nor non-perception; dhammas having an
external object are the threefold and fourfold jhānas, the result of the fourth
jhāna, the sphere of infinite space, and the four paths and fruitions (3.1.2). The
third mode is given a different interpretation in the Vibhaṅga: ajjhattabahiddhā
128 Vibh-a 268; trans. by Ñāṇamoli, 1996: 330: “Internally and externally: he applies the mind
at one time to his own and at another time to another’s feelings. In the [last] section,
because neither self nor other is specified, therefore in order to point out the mere laying
hold of feeling …”.
129 Vibh-a 44; Cf. As 426, Ps I 249, Vibh-a 375.
130 See Vibh-a 51.
131 See Vibh-a 55, 276.
136 Ditrich
is presented as contemplation of phenomena without reference to oneself or

another; this is further implied in the commentary, the Sammohavinodanī,
explaining the third mode as neither oneself (neva attā) nor another (na paro)
(3.2.2). On the other hand, in the Dhammasaṅgaṇi, the internal dhammas are
presented as those of one’s own (sattānaṃ), the external as those of others
(parasattānaṃ parapuggalānaṃ), and dhammā ajjhattabahiddhā as those
of both (tadubhayaṃ) (3.1.1). Only the latter interpretation was followed by
Buddhaghosa in his Visuddhimagga.
4 Interpretations of ajjhattaṃ, bahiddhā, and ajjhattabahiddhā:

From Buddhist Traditions to Modern Scholarship
Buddhaghosa’s interpretation of the three terms (2.2) has become generally

accepted in the Theravāda tradition and has remained the standard interpreta-
tion in a large section of secondary modern works dealing with the practice of
the satipaṭṭhānas. As noted by several scholars,132 the rendering of bahiddhā as
“other” is problematic: while mindfulness of the body practiced by observing
the bodies of others may be viable, the contemplation of the mind and phe-
nomena of others is certainly questionable. Although in several instances in
the Nikāyas there is mention that knowledge of others’ mental states can be
obtained through seeing, hearing, inference or mind reading,133 there is no ref-
erence to these four means of knowledge in the texts giving instruction on
contemplation in the three modes. Anālayo suggests that another person’s
mental state can be contemplated indirectly through observation of their pos-
ture, tone of voice, or facial expression.134 He comments that mindfulness
practice focused only internally can cause self-centeredness and hence, exter-
nal mindfulness builds up a balance between introversion and extroversion.135
These comments are interesting yet still problematic. Even in the case of con-
templation of another person’s body, it has to be noted that the suttas generally
instruct the monks to go into solitude; e.g. in the Satipaṭṭhānasutta it is said:
“having gone to the forest or to the foot of a tree, or to an empty place”136 which
132 Gethin, 2001: 92–99; Anālayo, 2006a: 92–99; Kuan, 2008: 117–119; Schmithausen, 2012: 291–
303.
133 D III 103, noted by Anālayo, 2006a: 93; cf. Schmithausen, 2012: 294–295.
134 Anālayo, 2006a: 93–94; cf. Schmithausen, 2012: 295.
135 Anālayo, 2006b: 247.
136 Idha, bhikkhave, bhikkhu araññagato vā rukkhamūlagato vā suññāgāragato vā nisīdati (M
I.56).
indicates that there would be no other people to observe, especially since the
monks were instructed to sit alone in meditation and develop concentration
and mindfulness.137 The three modes of contemplation are an integral compo-
nent of most texts on contemplation, retained to a large extent in the early
schools of Buddhism, and seem to have originated from a very early stage of
Buddhist tradition and hence, it can be presumed that they were initially
addressing wandering mendicants rather than large groups in monasteries.
Thus the instructions on the three modes would probably target primarily
ascetics who would practice in silence and solitude, not observing other
people.
The interpretation of the term ajjhattabahiddhā is even more puzzling: nei-
ther the Nikāyas not the commentaries by Buddhaghosa shed any light on
what this mode of practice involves. Interpretation of the three modes, espe-
cially of the third one, seems to have been problematic already at early stages
of canonical transmission, as reflected not only in the Pāli Canon and its com-
mentaries but also in the Chinese translations of the Sarvāstivāda and
Dharmagupta texts.138 The overview by Schmithausen of several texts attesting
the three modes shows that their interpretations vary; e.g. in the Śrāvakabhūmi
the three modes are presented is several ways: internally as referring to one’s
live body, externally to observation of a dead body, and both internally and
externally referring to insight that the dead body was once alive and that one’s
own live body will also become a corpse; or internally refers to one’s live body,
externally to non-living matter, and internally and externally to the body of
another living being; or internal contemplation of feelings refers to one’s own
feelings emerging in one’s own body, external to one’s own feelings emerging
in relation to non-living matter, and internal and external to those emerging in
relation to another.139
A different interpretation of the two modes may be drawn from the Chinese
translation of the *Mahāprajñāpāramitāśāstra by Kumārajīva.140 In the section
on mindfulness of the body, one’s own body (svakāya) is inner, another’s body
(parakāya) is outer; the five organs (indriya) are the inner body whereas the five
137 There are many instances in the suttas where the delight of seclusion is recommended, e.
g. AN V 134–135.
138 For example, the Sarvāstivāda version of the Satipaṭṭhānasutta in Chinese, translated
from the Madhyamāgama by Kuan, 2008: 146–151, prescribes, unlike the Pāli Canon, only
two modes – internal and external; e.g. “Thus a monk contemplates body as a body, con-
templates the external body as a body” (2008: 147).
139 Schmithausen, 2012: 296–299.
140 Lamotte, 1970: 1173–1175 (cited by Gethin, 2001: 54), Anālayo, 2006a: 97, Schmithausen,
2012: 296.
138 Ditrich
objects of the senses (viṣaya) are the outer body. The text does not elucidate the
third mode of contemplation, however, it states that the body can be examined
internally and externally simultaneously (yugapat), at the same time (ekakāle),
or as two distinct operations (bhinna).141 In comparison, Buddhaghosa also
interprets the third mode as “at times internally, at times externally” (kālena
attano, kālena parassa)142 which may refer to two distinct operation, or as
“both” (ubhayam),143 however, it is unlikely that this would refer to simultane-
ous contemplation since he states that both cannot occur at once (ekasmiṃ
kāle panidaṃ ubhayaṃ na labbhati).144 The *Mahāprajñāpāramitāśāstra dis-
tinguishes bodily feeling as external and mental feeling as internal, and the
feelings associated with the [first] five consciousnesses as external, and those
linked to the mind consciousness as internal. The mind is external when it has
for its object an external dharma, and internal when its object is an internal
dharma. Mental consciousness (manovijñāna) is the internal mind, and the
[first] five consciousnesses (pañcavijñāna) are the external mind. Notably, the
text does not elucidate the third mode (external and internal).145
Most modern scholars and practitioners, attempting to interpret ajjha
taṃ and bahiddhā, especially in terms of meditation practice,146 follow
Buddhaghosa, often with some modifications: e.g. only the internal mode
can be practiced as direct contemplation;147 the external mode is to be con-
templated by inference;148 the external mode refers to bodily feelings at skin
level, and internal to the feelings deeper in the body.149 The meaning of the
141 Lamotte, 1970: 1173.

142 Ps I 249.
143 Vism 440.
144 Ps I 249.
145 Here, an alternative interpretation may be suggested in relation to the practice of mind-
fulness of the six sense-spheres, depending on which aspect becomes prominent during
the practice: (1) ajjhattaṃ: the internal sense-sphere is the object of mindfulness (e.g.
when seeing, one is primarily aware of the eye); (2) bahiddhā: the external sense-sphere
is the object of mindfulness (e.g. when seeing, the object of seeing is predominant); (3)
ajjhatta-bahiddhā: the specific consciousness (one of the six) is the object of mindfulness,
i.e. the knowing of the process (e.g. when seeing, one is primarily aware of consciousness
arising with the visual object and the eyes).
146 A good survey of various modern interpreters is given by Anālayo, 2006a: 95–99.
147 Ñāṇapoṇika, 1962: 59, says that “only internal objects are taken up” since only these are
“accessible to direct experience” (paccakkhañāṇa “direct experience”); cf. Anālayo, 2006a:
94.
148 Mahāsi Sayadaw, 1994: 41, follows Buddhaghosa, adding that bahiddhā refers to “contem-
plation of the life processes of others, by way of inference (anumāna)”.
149 Goenka, 1999: 22; cf. Śrāvakabhūmi noted by Schmithausen, 2012: 297, note 29.
antonyms internal-external can be expanded, drawing from the Sarvāstivāda

sources, to the correlation past-future and further to the Upaniṣadic correla-
tion the personal-the cosmic.150 The interpretations from the Vibhaṅga are
viewed to be among the more plausible ones.151
5 Conclusion
The prominence and frequency of occurrences of ajjhattaṃ “internally,”

bahiddhā “externally,” and ajjhattabahiddhā “internally and externally” in the
texts concerning meditation in the Pāli Canon indicate their significance in
the practice of the four satipaṭṭhānas. In this chapter, all occurrences of the
three terms in the Pāli Canon are surveyed, their semantic ranges outlined and
linked to a variety of contextual usages.
Firstly, and in terms of new findings, through identification of the main syn-
tagmatic formulae comprised of the three terms in various types of collocation,
their specific functions and typological and semantic patterns are outlined
(2.1.1–2.1.7). These passages have been problematic for interpreters and trans-
lators alike, hence new approaches seem needed. As revealed here, in the
Nikāyas the compound ajjhattabahiddhā always appears in collocation with
ajjhattaṃ and bahiddhā in a triad, consistently so only in the passages (prob-
ably of a prescriptive nature) where instructions on contemplation of the four
satipaṭṭhānas are given. Consequently, we have to assume that the three modes
were probably a significant or even essential aspect of the meditation practice,
as evidenced by their integration and many occurrences in the Pāli canonical
and other early Buddhist sources. The attestations of ajjhattabahiddhā in the
context of contemplation only signal that this dvandva may have served as a
semiotic marker, referring to a particular type of meditation practice, presum-
ably known to the audience of the time. The formulaic repetitions of the three
terms (in dvandva, asyndeta or constructions with the conjunction vā) seem
to follow the style of the Vedic tradition which cannot but have influenced
the early Buddhist oral transmission. This is further supported by occurrences
of another triad which appears, yet again, only in the context of satipaṭṭhāna
practice, namely samudayadhammā, vayadhammā, samudayavayadhammā.152
150 Sujato, 2005: 174–177.

151 Kuan, 2008: 119; Anālayo, 2006: 95.
152 For example, samudayadhammānupassī vā kāyasmiṃ viharati, vayadhammānupassī vā
kāyasmiṃ viharati, samudayavayadhammānupassī vā kāyasmiṃ viharati (D II 292).
140 Ditrich
In contrast, as identified in this chapter, ajjhattaṃ and bahiddhā occur

linked as a pair in close or distant syntactic collocation in a broader range of
contexts, explicating or describing various aspects of the teachings such as the
aggregates or the hindrances, but not when contemplation practice is dis-
cussed. The adverb bahiddhā is only rarely attested outside the syntactic
correlation to ajjhattaṃ; on its own it refers to things being external, outside,
or to people outside the Buddhist teachings. The term ajjhattaṃ is frequently
used without bahiddhā, mostly referring to the states of mind such as the pres-
ence or absence of various phenomena. The adjectival forms ajjhattika
“internal” and bāhira “external,” when compounded in a dvandva, always refer
to the six internal and external sense-spheres respectively, however, when not
compounded they are used among attributes describing various phenomena
or doctrinal aspects. The adjective ajjhattika always occurs in close or distant
collocation with bāhira whereas the adjective bāhira, when attested on its
own, refers to things that are external or outside. To summarize, each type of
collocation of the three terms has a rather well defined semantic range which
is signaled, as suggested here, by the particular syntagmatic formula; e.g. the
triad ajjhattaṃ, bahiddhā, ajjhattabahiddhā refers to a specific mode of con-
templation; the dvandva ajjhattikabāhira refers to the six internal and external
sense-spheres respectively. However, no sutta (except for the Janavasabhasutta,
2.1.1) in the Nikāyas provides information on the meanings of the three terms
in question, nor do any provide clarification of their significance in the
satipaṭṭhāna practice itself. It is in the Abhidhamma that the meanings for the
three terms are given or implied.
Secondly, this chapter investigates how formulaic stylized expressions,
occurring in a variety of syntagms comprised of the three terms, are presented
and interpreted in the Abhidhamma, and linked to the explications established
in the later Pāli commentarial literature which serve today as the standard
interpretations of the terms. As discussed, the instructions on the four
satipaṭṭhānas in the Nikāyas consist largely of very brief comprehensive lists; it
is very likely that oral commentaries and instructions on the actual practice
were handed down along with the texts and supposedly it was taken for granted
that the audience was well versed in the topic. It was only much later, through
the commentaries from the fifth century onwards ascribed to Buddhaghosa
that the interpretations of the terms ajjhattaṃ, bahiddhā and ajjhattabahiddhā
were established and have since then remained the standard reading in the
Theravāda tradition – ajjhattaṃ referring to oneself, bahiddhā to another and
ajjhattabahiddhā to both or alternatively to oneself and another (2.2). The
commentaries seem partly to draw from the Abhidhammapiṭaka where the
first two modes are explained in several ways. In the Dhammasaṅgaṇi, ajjhattā
dhammā is said to refer to oneself and bahiddhā dhammā to others while

ajjhattabahiddhā refers to both (3.1.1). The terms are further commented upon
in the Atthasālinī, largely reflecting the interpretations in other aṭṭhakathā
texts (3.1.1). Another interpretation is given in the last section of the
Dhammasaṅgaṇi, the Aṭṭhakathākaṇḍa, where the terms are related to differ-
ent stages of high absorptions (3.1.2).
The Vibhaṅga though contributes a different angle to the interpretation of
the three terms. The Satipaṭṭhānavibhaṅga may explicate ajjhataṃ in reference
to oneself and bahiddhā to another, as commonly read by most interpreters;
however, there is space for alternative readings, as suggested here (3.2.1). The
third mode ajjhattabahiddhā is presented in the Vibhaṅga as contemplation
of phenomena or processes without reference to oneself or another; this study
proposes that the term may refer to the experience of anattā, implied also in
the commentary Sammohavinodanī where the third mode is presented as nei-
ther oneself (neva attā) nor another (na paro) (3.2.2). It is noteworthy that other
canonical and post-canonical texts do not comment upon or follow up this pre-
sentation of ajjhatta bahiddhā in the Vibhaṅga. This may be an indication of
the Vibhaṅga reflecting a different, perhaps earlier layer of Abhidhamma texts,
which is further evidenced by the attestations of the triad in the Vibhaṅga,
occurring only in the context of the four satipaṭṭhānas (following the pattern
identified in the Nikāyas, 2.1.1), whereas in the Dhammasaṅgaṇi the three terms
are listed as three different perspectives for examination of all the dhammas.
From all the interpretations of the three terms given in the Abhidhamma, only
the explanation given in the Dhammasaṅgaṇi, referring ajjhattā dhammā to
oneself, bahiddhā dhammā to others, and ajjhattabahiddhā to both (3.1.1), has
been followed by Buddhaghosa and by the later Theravāda tradition.
In this chapter some semantic inferences are drawn from the context
dependent use of the terms in question, occurring in a range of formulaic
expressions. The complete lack of any explanation of, or comment upon
ajjhattabahiddhā in the Nikāyas, the very scant information given in the
Abhidhamma texts, along with the variety of interpretations found in the
Chinese translations of the texts originating from other early Buddhist schools,
indicate that interpretation of the three modes, especially the third one, may
have been problematic already at an early stage of the canonical transmission.
Since the triad (ajjhattaṃ, bahiddhā, ajjhattabahiddhā), the pair (ajjhattaṃ,
bahiddhā) and the single terms occur in specifically different contexts, the
semantic ranges inferred from the usage of the terms in pair, functioning as
complementary antonyms, should not be simply applied to the interpretations
of the terms in triad (ajjhattaṃ, bahiddhā, ajjhattabahiddhā), especially in
the case of the compound ajjhattabahiddhā which seems to refer to a specific
142 Ditrich
meditation practice (e.g. may refer to contemplation without reference, as sug-

gested before).
As evidenced by the consistent integration of ajjhattaṃ, bahiddhā and
ajjhattabahiddhā in the Pāli canon and other early Buddhist sources, the three
modes form a very important aspect of the satipaṭṭhāna practice and hence,
their significance requires investigation from several perspectives, and within
the broader context of ancient Indian traditions of the time. This chapter
exemplifies how the analysis of the context-dependent usage and the typologi-
cal and semantic patterns formed by these terms can be used as a way to mark
out their functions and meanings. Undoubtedly, the problem of the interpreta-
tion of the three terms in question remains open to further research.
Abbreviations
Abbreviations and the quotation system of Pāli sources follow the Critical Pāli
Dictionary (Epilegomena to vol. 1, 1948, pp. 5*–36*, and vol. 3, 1992, pp. II–VI). The
numbers in the quotations of Pāli sources refer to the volume and page of the Pāli Text
Society edition (e.g. MN I 21 refers to the Majjhimanikāya, vol 1, p. 21).
AN Aṅguttaranikāya (ed. Morris and Hardy, [1885–1900] 1999–2013)

As Atthasālinī (ed. Müller; rev. Cousins, [1897] 2011)
DN Dīghanikāya (ed. Rhys Davids and Carpenter, [1890–1911] 1995–2007)
Dhp Dhammapada (ed. von Hinüber and Norman, 1994)
Dhs Dhammasaṅgaṇi (ed. Müller, [1885] 2001)
MN Majjhimanikāya (ed. Trenckner and Chalmers, [1888–1902] 2013)
Nidd I Mahāniddesa (ed. de La Vallée Poussin and Thomas, [1916–1917] 2001)
Nidd II Cullaniddesa (ed. Stede, [1918] 1988)
PED Pāli-English Dictionary. Rhys Davids, Thomas W. and Stede, William. London:
The Pāli Text Society, 1921–1925.
Ps Papañcasūdanī, Majjhimanikāyāṭṭhakathā of Buddhaghosa (ed. Woods,
Kośambi, Horner, [1922–1938] 1976–1979)
SN Saṃyuttanikāya (ed. Feer, [1884–1898] 1975–2006)
Sv Sumaṅgalavilāsinī (ed. Rhys Davids, Carpenter, Stede, [1929–1932] 1968–1971)
渡邊 and Ono 小野, 1924–1934)
Th Theragāthā and Therīgāthā (ed. Oldenberg and Pischel, [1883] 1966)
Ud Udāna (ed. Steinthal, [1885] 1982)
Vibh Vibhaṅga (ed. Rhys Davids, [1904] 2003)
Vibh-a Sammohavinodanī (ed. Buddhadatta, [1923] 1980)
Vism Visuddhimagga (ed. Rhys Davids, [1920–1921] 1975)
Bibliography
Anālayo Bhikkhu (2006a). Satipaṭṭhāna: The Direct Path to Realization. Selangor:

Buddhist Wisdom Centre.
Anālayo Bhikkhu (2006b). “Mindfulness in the Pāli Nikāyas”, In Buddhist Thought and
Applied Psychological Research: Transcending the Boundaries. Edited by D.K. Nauriyal,
Michael S. Drummond, and Y.B. Lal. London and New York: Routledge, pp. 229–249.
Bodhi Bhikkhu (2000). The Connected Discourses of the Buddha: A New Translation of
the Saṃyutta Nikāya. Boston MA: Wisdom Publications.
Bodhi Bhikkhu (2012). The Numerical Discourses of the Buddha: A Translation of the
Aṅguttara Nikāya. Boston MA: Wisdom Publications.
Bronkhorst, Johannes (1985). “Dharma and Abhidharma”, Bulletin of the School of
Oriental and African Studies 48, pp. 305–320.
Buddhadatta, Ambalangoda Polwatte (ed.) ([1923] 1980). Sammohavinodanī. London:
Cousins, Lance. S. (1983). “Pāli Oral Literature”, In Buddhist Studies: Ancient and Modern.
Edited by Philip Denwood and Alexander Piatigorsky. London: Curson Press, pp. 1–11.
Ditrich, Tamara (2010). “The Variety of Expression for Heaven and Earth in the Ṛgveda”,
Crossroads 5(1), pp. 35–44.
Ditrich, Tamara (forthcoming). “Stylistic Analysis of Coordinative Nominal Constructions
for Dual Deities in the Ṛgveda”, Bulletin d’Études Indiennes.
Feer, Leon (ed.) ([1884–1898] 1975–2006). Saṃyuttanikāya. 5 vols. London: The Pāli Text
Society.
Frauwallner, Erich (1964). “Abhidharma-Studien, II”, Wiener Zeitschrift für die Kunde
Südasiens und Archiv für indische Philosophie 8, pp. 59–99.
Frauwallner, Erich (1971). “Abhidharma-Studien: III. Abhisamayavādaḥ“, Wiener
Zeitschrift für die Kunde Südasiens und Archiv für indische Philosophie 15,
pp. 69–121.
Frauwallner, Erich (1972). “Abhidharma-Studien: IV. Der Abhidharma der anderen
Schulen”, Wiener Zeitschrift für die Kunde Südasiens und Archiv für indische Philosophie
16, pp. 95–152.
Gethin, Rupert (1992). “The Mātikas: Memorisation, Mindfulness, and the List”, In In the
Mirror of Memory: Reflections on Mindfulness and Remembrance in Indian and Tibetan
Buddhism. Edited by Janet Gyatso. Albany NY: State University of New York Press,
pp. 149–172.
Gethin, Rupert (2001). The Buddhist Path to Awakening. 2nd Edition. Oxford: Oneworld.
Goenka, S.N. (1999). Discourses on the Satipaṭṭhāna Sutta. Igatpuri: Vipassanā Research
Institute.
Hinüber, Oskar von, and Norman, Kenneth R. (eds.) (1994). Dhammapada. London: The
Pāli Text Society.
144 Ditrich
Hinüber, Oskar von (1996). A Handbook of Pali Literature (Indian Philology and South
Asian Studies 2). Berlin: De Gruyter.
Kragh. Ulrich L. (2002). “The Extant Abhidharma-Literature”, The Indian International
Journal of Buddhist Studies 3, pp. 123–168.
Kuan, Tse-fu (2008). Mindfulness in Early Buddhism: New Approaches through Psychology
and Textual Analysis of Pali, Chinese and Sanskrit Sources. London and New York:
Routledge.
La Vallée Poussin, Louis de, and Thomas, Edward J. (eds.) ([1916–1917] 2001). Mahāniddesa.
London: The Pāli Text Society.
Lamotte, Étienne (1970). Le Traité de la Grande Vertu de Sagesse de Nāgārjuna
(Mahāprajñāpāramitāśāstra). Vol. 3. Louvain-la-Neuve: Université de Louvain,
Institut Orientaliste.
Mahāsi, Sayadaw (1994). The Progress of Insight: A Modern Pali treatise on Buddhist
Satipaṭṭhāna Meditation. Trans. by Nyānaponika. Kandy: Buddhist Publication
Society.
Maung Tin, Pe ([1920–1921] 2012–2013). The Expositor (Atthasālinī): Buddhaghosa’s
Commentary on the Dhammasaṅgaṇi, the first book of the Abhidhamma Piṭaka. Rev.
by Mrs Rhys Davids. 2 vols. Bristol: The Pāli Text Society.
Morris, Richard and Hardy, Edmund (eds.) ([1885–1900] 1999–2013). Aṅguttaranikāya.
5 vols. London: The Pāli Text Society.
Müller, Edward (ed.) ([1897] 2011). Atthasālinī. Rev. by L.S. Cousins. London: The Pāli
Text Society.
Müller, Edward (ed.) ([1885] 2001). Dhammasaṅgaṇi. London: The Pāli Text Society.
Ñāṇamoli Bhikkhu (1980). A Thinker’s Notebook. Kandy: Buddhist Publication Society.
Ñāṇamoli Bhikkhu ([1956] 1994). The Path of Purification (Visuddhimagga) by Bhantā
cariya Buddhaghosa. Kandy: Buddhist Publication Society.
Ñāṇamoli Bhikkhu (1996). The Dispeller of Delusion (Sammohavinodanī). Rev. by
L.S. Cousins; Nyanaponika Mahāthera and C.M.M. Shaw. 2 vols. Oxford: Pāli Text
Society.
Ñāṇanamoli Bhikkhu and Bodhi Bhikkhu (1995). The Middle Length Discourses of the
Buddha: A New Translation of the Majjhima Nikāya. Boston MA: Wisdom Publication.
Ñāṇapoṇika Thera (1962). The Heart of Buddhist Meditation. London: Rider.
Norman, Kenneth R. (1983). Pāli Literature (A History of Indian Literature VII, 2).
Wiesbaden: Otto Harrassowitz.
Oldenberg, Henrik and Pischel, Richard (eds.) ([1883] 1966). Theragāthā and Therīgāthā.
London: Pāli Text Society.
Rhys Davids, Caroline A.F. ([1900] 2012). A Buddhist Manual of Psychological Ethics.
Bristol: Pāli Text Society.
Rhys Davids, Caroline A.F. (ed.) ([1904] 2003). Vibhaṅga. London: The Pāli Text Society.
Rhys Davids, Caroline A.F. (ed.) ([1920–1921] 1975). Visuddhimagga. London: The Pāli
Text Society.
Rhys Davids, Thomas W., and Carpenter, Joseph E. (eds.) ([1890–1911] 1995–2007).
Dīghanikāya. 3 vols. London: The Pāli Text Society.
Rhys Davids, Thomas W., Carpenter, Joseph E. and Stede, William (eds.) ([1929–1932]
1968- 1971). Sumaṅgalavilāsinī, London: Pāli Text Society.
Schmithausen, Lambert (2012). “Achtsamkeit ‘innen’, ‘außen’ und ‘innen wie außen’”, In
Achtsamkeit: Ein buddhistisches Konzept erobert die Wissenschaft – mit einem Beitrag
S.H. des Dalai Lama. Edited by Michael Zimmermann Christof Spitz and Stefan
Schmidt. Bern: Hans Huber, pp. 291–303.
Soma Thera ([1941] 1981). The Way of Mindfulness. Kandy: Buddhist Publication Society.
Stede, William (ed.) ([1918] 1988). Cullaniddesa. London: The Pāli Text Society.
Steinthal, Paul (ed.) ([1885] 1982). Udāna. London: The Pāli Text Society.
Sujato Bhikkhu (2005). A History Of Mindfulness: How Insight Worsted Tranquility in the
Satipatthana Sutta. Taipei: The Corporate Body of the Buddha Educational
Foundation.
玄妙 (eds.) (1924–1934). Taishō shinshū daizōkyō 大正新脩大藏經, Tokyo: Taishō
- T.1.1: Dīrghāgama, Chang ahan jing 長阿含經, Buddhayaśas and Zhu Fonian 竺佛
念
Thiṭṭila, Sayadaw U. (2010). The Book of Analysis (Vibhaṅga). Pāli Text Society (PTS
Translation Series No. 39). Bristol: Pāli Text Society.
Trenckner, Vilhelm and Chalmers, Robert (eds.) ([1888–1902] 2013). Majjhimanikāya.
3 vols. London: The Pāli Text Society.
Walshe, Maurice (1995). The Long Discourses of the Buddha: A Translation of the Dīgha
Nikāya. Boston MA: Wisdom Publications.
Willemen, Charles, Dessein, Bart and Cox, Collett (1998). Sarvāstivāda Buddhist Scholas
ticism (Handbook of Oriental Studies 2/11). Leiden: Brill.
Woods, James H., Kośambi, Dharmananda, Horner, Isaline B. (eds.) ([1922–1938] 1976–
1979). Papañcasūdanī, Majjhimanikāyāṭṭhakathā of Buddhaghosa. 5 vols. London:
The Pāli Text Society,
Wynne, Alexander (2005). “The Historical Authenticity of Early Buddhist Literatue:
A Critical Evaluation”, Wiener Zeitschrift für die Kunde Südasiens 49, pp. 35–70.
146 Kramer
Chapter 4
Some Remarks on the Proofs of the “Store Mind”

(Ālayavijñāna) and the Development of the
Concept of Manas*
Jowita Kramer
This chapter is concerned with two main topics: The first part provides a com-
parison of the four proofs for the existence of the ālayavijñāna found in
Vasubandhu’s Pañcaskandhaka and Sthiramati’s Pañcaskandhakavibhāṣā with
the eight proofs included in the “Proof Portion”1 of the Yogācārabhūmi and the
six proofs of the Mahāyānasaṃgraha. The second part deals with the concept
of manas, trying to trace its development from early Buddhism to the Yogācāra
theory of kliṣṭamanas.
1 Proofs of the ālayavijñāna
The four arguments to establish the existence of the ālayavijñāna mentioned

in the Pañcaskandhaka (PSk 16,11–17,4) and explained in the Pañcaskandha
kavibhāṣā (PSkV 51b1–57b5) may be summarized as follows:2
1. Actual perception (pravṛttivijñāna) reappears after a person has risen

from unconscious states as for instance the equipoise of cessation
(nirodhasamāpatti).
2. Actual perceptions appear in different modes (prakāra) depending on
different kinds of object conditions (ālambanapratyaya).
3. Actual perception reappears after it has been interrupted by sleep
(middha) or a swoon (mūrchā).
* I would like to thank Ralf Kramer, Constanze Pabst von Ohain, Alexander von Rospatt,
Lambert Schmithausen, and Robert Sharf for offering very helpful comments and corrections
to previous drafts of this chapter. I am also grateful for the support received from the German
Research Foundation (DFG), which enabled me to complete this article.
1 The name “Proof Portion” for this part of the Yogācārabhūmi was introduced in Schmithausen,
1987: 299, note 226.
2 A detailed study of the four proofs is included in Kramer, 2014b: 316–319.

Some Remarks on the Proofs of the “Store Mind” 147
4. Without the ālayavijñāna an individual could not (a) arise in and (b) be
liberated from saṃsāra, because, on the one hand, the process of rebirth
would not be possible and, on the other, the contaminations (kleśa)
could not be removed.
These four arguments presented in the Pañcaskandhaka(vibhāṣā) appear con-

siderably different from the eight proofs provided in the “Proof Portion” of the
Yogācārabhūmi and in the Abhidharmasamuccayabhāṣya (which includes an
identical listing):3
1. appropriating (upātta; ASBh 12,1–13)

2. beginning (ādi; ASBh 12,14–19)
3. clarity (spaṣṭatva; ASBh 12,20–24)
4. seed (bīja; ASBh 12,25–13,3)
5. function (karman; ASBh 13,4–7)
6. corporeal sensation (kāyiko ’nubhavaḥ; ASBh 13,8–11)
7. two unconscious absorptions (acitte samāpattī; ASBh 13,12–15)
8. death (cyuti; ASBh 13,16–20)
Apart from these two obviously different traditions there is a third alternative
enumeration found in the Mahāyānasaṃgraha,4 which includes the following
six5 proofs:
1. defilement [consisting in] the contaminations (kleśasaṃkleśa; MSg

I.30–32)
2. defilement through deeds (karmasaṃkleśa; MSg I.33)
3. defilement through birth (janmasaṃkleśa; MSg I.34–42)
4. mundane purification (laukiko vyavadānaḥ; MSg I.43)
5. supramundane purification (lokottaro vyavadānaḥ; MSg I.44–49)
6. ālayavijñāna is necessary in nirodhasamāpatti (MSg I.50–55)
3 The edition of the Tibetan text of the “Proof Portion” is found in Hakamaya, 1978: 7–15. The
Sanskrit text is preserved in ASBh 11,15–13,20. See also Yid kun 40a-48b. For investigations of
these eight proofs, see Griffiths, 1986: 96–104 and 130–138; Schmithausen, 1987: 194–196; and
Waldron, 1995: 16–18. Moreover, a very insightful study of the eight proofs and their possible
meditative implications has recently been published by Nobuyoshi Yamabe (see Yamabe,
2015).
4 MSg I.29–55.
5 As already pointed out in Schmithausen, 1987: 402, note 710, the sixth proof is not included in
the systematic enumeration of five proofs (the first three concerning “defilement” [saṃkleśa]
and the latter two relating to “purification” [vyavadāna]) stated in advance in MSg I.29.
148 Kramer
Although these three lists of proofs are formally different, the arguments
they include are partly similar and may be classified into three main
categories.
1.1 First Category of Proofs

A first group of proofs comprises arguments concerned with issues related to
the presence of the mind in the body. These proofs deal for a great part with
what one could call the “vitalizing aspects” of the mind, i.e. with bringing life
to the body and keeping the latter alive, and with the ālayavijñāna’s function of
ensuring the continuity of the mind. The vitalizing function of the store mind
refers mainly to what Lambert Schmithausen calls the “somatic” aspects of the
ālayavijñāna,6 that is its appropriation of the body at the beginning of a new
existence and throughout life and its gradual abandonment of the body at the
point of death. The ālayavijñāna’s function of appropriating the body at the
moment of conception and throughout life is discussed in the first argument of
the “Proof Portion”. In the Pañcaskandhakavibhāṣā, only the first aspect of the
appropriation is mentioned, namely in the first part of the fourth proof.
Sthiramati explains that the ālayavijñāna is essential for the process of rebirth
since the vijñāna mentioned in the series of dependent origination cannot
have the nature of an actual perception (pravṛttivijñāna).7 Parallel arguments
are also expressed in the second and third proof of the Mahāyānasaṃgraha,
that is “defilement through deeds” (karmasaṃkleśa) and “defilement through
birth” (janmasaṃkleśa).8
Another proof that could be assigned to this first category is proof number
seven of the “Proof Portion,” the “two unconscious absorptions” (acitte
samāpattī). This argument expresses the view that the ālayavijñāna is neces-
sary to keep the body alive during meditative states, in which no actual
perceptions occur. This statement is related to Vasubandhu’s first proof, in
which the reappearance of actual perception after a person has risen from
unconscious states (or the state of being without conception [āsaṃjñika]) is
mentioned. Parallel arguments are also found in paragraphs 50–52 of the
6 Schmithausen, 1987:195.
7 PSkV 55b6ff. See also Kramer, 2014b: 318.
8 For a detailed discussion of these arguments, see Waldron, 1995: 26f. Notably, the reasons given
in the Mahāyānasaṃgraha to support the idea that the ālayavijñāna is necessary for the
process of rebirth in all the three spheres of the world and for sustaining the body during the
whole life are more multifaceted than those in the Pañcaskandhaka(vibhāṣā). MSg I.37, for
instance, mentions the concept of vijñāna “nourishing” the body, i.e. securing the subsistence
of corporeal matter (on this function, see also Schmithausen, 1987: 70), as one of the four
“nourishments” (āhāra).
Mahāyānasaṃgraha’s first chapter (= proof 6 mentioned above). Notably,

Vasubandhu’s third proof, actual perception “reappears after it has been inter-
rupted,” which points in a similar direction, has no parallels in the other two
lists. This argument refers to states in which actual perception is interrupted
due to incidents not connected to meditation but to phases of sleep or
fainting.9
Another proof that seems to be related to the somatic function of the store
mind is number six in the “Proof Portion”. This proof ascribes the fact that cor-
poreal sensation (kāyiko ’nubhavaḥ) can occur regardless of whether the person
is thinking properly (yoniśas) or improperly (ayoniśas) or whether the person’s
mind is in a state of contemplation or not to the existence of ālayavijñāna. This
argument is not entirely clear and has been interpreted differently by modern
scholars. Lambert Schmithausen understands it as pointing to the problem
that physical sensation can be present even in meditative states where there is
no functioning of the tactile sense perception.10 In contrast to this view, Paul
Griffiths seems to take this argument to indicate the fact that physical experi-
ence can vary in the same individual and that this variation cannot originate
from the sensation itself (and can only be explained if the existence of the
ālayavijñāna is accepted).11 Finally, the eighth proof of the “Proof Portion,”
stating that the existence of ālayavijñāna is necessary in order to make the
vijñāna’s gradual withdrawal from a dying body possible, may also be allocated
in the first category of proofs. The Mahāyānasaṃgraha does not discuss the
problem of bodily experience, but refers to the withdrawing vijñāna at death
in connection with its third proof, “defilement through birth (janmasaṃkleśa)”
(MSg I.42). Neither proof six nor eight are mentioned in any form in the
Pañcaskandhaka(vibhāṣā). It therefore seems most likely that these somatic
aspects of the store mind lost their relevance in the course of time.
It should also be noted that some of the arguments discussed above
appear similar but have to be distinguished with regard to their emphasis.
While the “Proof Portion” seems to be mainly concerned with the vitalis-
ing quality of the store mind in the strict sense of its being the “principle of
life,”12 the Pañcaskandhaka(vibhāṣā) places more weight on the aspect of
mental continuity after states of interruption. Both, the “Proof Portion” and
the Pañcaskandhaka(vibhāṣa) explain for instance that the ālayavijñāna is
necessary because perception is interrupted in meditative states. However,
9 PSkV 55a6ff.
10 Schmithausen, 1987: 44.
11 Griffiths, 1986: 102f. See also Yamabe, 2015: 153–155, and Schmithausen, 2014: 17–21.
12 On this function, see Schmithausen, 1987: 44.
150 Kramer
the “Proof Portion” emphasizes the fact that the ālayavijñāna keeps the body
alive during these states, whereas the Pañcaskandhaka(vibhāṣa) focuses on the
assumption that actual perception can reappear after a person has risen from
unconscious states only on the basis of the store mind. It is therefore possible
that the eighth proof of the “Proof Portion” (i.e. the indispensability of the
store mind in the moment of the mind gradually leaving the body at the point
of death) is not mentioned in the Pañcaskandhakavibhāṣā because this issue is
only related to the vitalizing mechanism (and not to continuity).
Summarizing the results of our investigation so far, the following proofs
have been subsumed under the first category concerned with the ālayavijñāna’s
relevance for the presence of the mind in the body:
a. Yogācārabhūmi proofs 1 and 6–8

b. Mahāyānasaṃgraha proofs 2, 3, and 6
c. Pañcaskandhaka(vibhāṣā) proofs 1, 3, and 4a
1.2 Second Category of Proofs

The common characteristic of the arguments which I have grouped in the sec-
ond category of proofs is their concern with the actual perceptions
(pravṛttivijñāna). While some of these arguments refer mainly to the function
of the ālayavijñāna as the store of impressions and seeds of the pravṛttivijñānas,
others deal with the possibility of the simultaneous arising of several vijñānas.
As for the first group, the “Proof Portion” describes in connection with its
fourth proof the condition that the varying pravṛttivijñānas cannot be the
seeds of each other. This argument is very closely related to Vasubandhu’s sec-
ond proof, that is the statement that ālayavijñāna is needed because actual
perceptions appear in different modes (prakāra) depending on different kinds
of object conditions (ālambanapratyaya). In his comments on this proof
Sthiramati explains that actual perceptions appear in various modes because
of various kinds of object conditions. Thus, it is usually impossible for a pre-
ceding perception to be the seed of the following one, for example a beneficial
perception is not appropriate as the seed of an unbeneficial perception.13
Therefore a “store,” namely the ālayavijñāna, is necessary to hold the seeds
from which the different perceptions may arise. These seeds, in turn, are nour-
ished by previous actual perceptions that leave imprints (vāsanā) in the
ālayavijñāna.14 As for the Mahāyānasaṃgraha, its first proof, the “defilement
[consisting in] the contaminations” (kleśasaṃkleśa), seems to deal with a rela
13 PSkV 55a1f.
14 PSkV 55a4f. See also Kramer, 2014b: 318.
ted issue, namely the condition that only an entity having the nature of the
ālayavijñāna can be the recipient of the imprints (vāsanā) made by the actual
sense perceptions (MSg I.30).
As mentioned above, some of the proofs which I have assigned to this sec-
ond group – because they are also concerned with the pravṛttivijñānas – refer
to a slightly different issue, namely the simultaneous arising of several vijñānas.
These proofs are only dealt with in the Yogācārabhūmi, namely in proofs 2, 3,
and 5, and have no parallels in the other texts. As already indicated by Lambert
Schmithausen, the arguments presented in these “proofs” are not directly prov-
ing the existence of the ālayavijñāna but are aiming to show that several
vijñānas, i.e. the pravṛttivijñāna(s) and the ālayavijñāna, may occur at the same
time.15 According to the first of these proofs, more than one actual perception
may occur at the same time because there are cases when someone wants to
see, hear, taste etc. simultaneously. Neither the meaning of this argument nor
the reasons for its being indicated as “beginning” (ādi) in the summarizing
verse at the beginning of the “Proof Portion” are entirely clear. Paul Griffiths
seems to take the latter to mean that ālayavijñāna is necessary as the basis for
the first moment (ādi?) of a particular actual perception. His example is “the
occurrence of visual consciousness in the mind of a man whose eyes have been
closed for hours”.16 Moreover, Griffiths understands the argument in the sense
that ālayavijñāna is needed as the immediate condition for simultaneously
arising sense perceptions.17 In contrast, Lambert Schmithausen considers it
not to be intended as a direct proof of the ālayavijñāna but of the appropriate-
ness of the assumption that the pravṛttivijñāna(s) and the store mind can exist
simultaneously since even several pravṛttivijñānas can occur at the same time.18
The second of this category’s proofs (i.e. proof 3) is adduced to show that
mental perception (manovijñāna) could not function clearly if it could not
arise at the same time as the other sense perceptions, the sense data of which
the manovijñāna analyses. In order to illustrate this issue, the thinking about
present objects of the sense perception (i.e. analyzing them by manovijñāna)
is compared with the thinking about remembered objects. While mental per-
ception is explained to function clearly with regard to the first, it is considered
to be unclear with regard to the latter objects.19 The last of the proofs of this
15 Schmithausen, 1987: 195f.

16 Griffiths, 1986: 99.
17 Griffiths, 1986: 99, 133. For a critical discussion of Griffiths’ understanding, see Yamabe,
2015: 145–149.
18 Schmithausen,1987: 46.
19 Griffiths, 1986: 134.
152 Kramer
category (proof 5) is concerned with the simultaneity of the actual percep-

tions, the notion of “I” (by the kliṣṭamanas), and the ālayavijñāna’s continuous
perceiving of the surrounding world (bhājana) and one’s own corporeal basis
(āśraya).20
Neither the Mahāyānasaṃgraha nor the Pañcaskandhaka(vibhāṣā) men-
tion any of the arguments of this category. Maybe the question of whether
simultaneous existence of subliminal kinds of vijñāna (i.e. the store mind and
the kliṣṭamanas) and actual perception is possible lost its relevance in the
course of time.21 However, the issue of the simultaneous arising of several
pravṛttivijñānas appears to have remained a controversial topic up to the times
of Sthiramati. In the Pañcaskandhakavibhāṣā, Sthiramati explicitly states that
manovijñāna follows the sense perception,22 whereas in his commentary on
the Triṃśikā he explains that manovijñāna occurs simultaneously with the
sense perceptions and that all six kinds of perception may arise at the same
time.23
Mental perception is traditionally described as the perception that is
based on dharmas, which are mental objects that cannot be sensed by the
other five indriyas, and on manas, the mental faculty.24 The essential differ-
ence between the five sense perceptions and manovijñāna results from the
assumption that the former cognize only their own respective objects being
free of conceptualization, whereas the latter is concerned with conceptual-
izing the sense data of all sense perceptions.25 In some Yogācāra sources, the
description of this process was further systematized. The first occurrence
of a more detailed explanation is to be found in the “Basic Section” of the
Yogācārabhūmi. There it is explained that after the actual sense perception,
which is called aupanipātika (“occurring directly”), follows the investigating
(paryeṣaka) and the determining (niścita) manovijñāna. By means of these two
20 See Griffiths, 1986: 101f, 135f., and Schmithausen, 1987: 90, 196, 386f., note 631a. Griffiths
considers this proof and the fourfold vijñapti mentioned in it to be parallel with a passage
in the Madhyāntavibhāga and seems to understand vijñapti here in the context of the
vijñaptimātratā (see especially p. 102). In contrast, Schmithausen points out that this pas-
sage in the “Proof Portion” does not presuppose the “representation only” concept and
refers merely to the cognitive functions of the eight vijñānas.
21 The fact that it was important in the early developmental stages of Yogācāra thought is
discussed in Schmithausen,1987: 45f.
22 PSkV 49b4.
23 TrBh 102, 9ff.
24 See, for instance, MN I 112.
25 See Schmithausen, 1967: 122f.
stages of manovijñāna, the object is conceptualized (vikalpyate).26 The model

described by Sthiramati in his commentary on the Pañcaskandhaka is proba-
bly to be considered a further elaboration of the system explained in the “Basic
Section”. According to Sthiramati, after the actual perception of an object, the
investigative (paryeṣaka) mental perception (manovijñāna) arises, being fol-
lowed by the classifying (vyavasthāpaka) and finally by the conceptualizing
(vikalpaka) manovijñāna.27 Although the terminology used is not identical
and the vikalpaka stage is not a separate level of the manovijñāna in the “Basic
Section,” there are clear parallels between these two models visible: As already
indicated, the pravṛttivijñānas correspond to the aupanipātika function, the
investigating manovijñāna is mentioned in both texts as paryeṣaka, and the
manovijñāna qualified as niścita in the “Basic Section” is parallel to the “clas-
sifying” (vyavasthāpaka) manovijñāna of the Pañcaskandhakavibhāṣā.
This understanding of cognition seems to be related to (though not identi-
cal with) the stages of the process of perceiving explained in Buddhaghosa’s
Visuddhimagga, which is described as follows. As soon as there is an object, the
mind becomes attentive or alert which is referred to as āvajjana and which is
the moment just before the actual perception. After the actual sense percep-
tion has arisen, the moment of mental reception or apperception of the
sense-data called sampaṭicchana follows. Both the moment before and the
moment after the actual sense perception are referred to in the Visuddhimagga
as manodhātu.28 Notably, the moment of becoming attentive has no corre-
spondence in the Yogācāra model.29 Moreover, the understanding of the
concept of manas seems to differ in the two traditions, and we do not find an
equivalent of the sampaṭicchana function in the Yogācāra description.
However, the following two moments of the Visuddhimagga, the investigating
(santīraṇa) and the determining (voṭṭhapana) manoviññāṇa are parallel to the
paryeṣaka and the vyavasthāpaka moments of the Pañcaskandhakavibhāṣā.
What follows is the stage of impulses (javana), which does not include the
“conceptualizing” aspect of the Yogācāra model, but is similar to the vikalpaka
aspect insofar as it represents the moment of intention (cetanā) and produc-
26 Y 10,2f. and 58,18f.

27 See Kramer, 2014b: 314f.
28 The whole perception process is described in Vism XIV, 458–460. See also Karunadasa,
2010: 138–143. On the concept of manodhātu in the Śāriputrābhidharma, see Schmithau-
sen, 2014: 222f. and 223, note 973. On the Theravāda theory of the heart (hadayavatthu) as
the material basis for the mind, see Karunadasa, 2010: 78–82.
29 Remarkably, the term āvarjana (which corresponds to Pāli āvajjana) is mentioned in
ASBh 5,1f. (see also PSkV 20b1f.) in the context of the definition of manaskāra: ālaṃbane
cittadhāraṇaṃ tatraiva punaḥ punar āvarjanaṃ veditavyam.
154 Kramer
tion of karmic activity. The next moment mentioned in the Visuddhimagga is

tadārammaṇa, which indicates “[the retaining of] the object [of the javana]”30.
There is again no direct equivalent in the Yogācāra system, which instead men-
tions the process of leaving an imprint (vāsanā) in the ālayavijñāna as the final
step of perception.31
To sum up, the relations between the three sources may be depicted in the
following way:
“Basic Section” Visuddhimagga Pañcaskandhakavibhāṣā

āvajjana
aupanipātika 5-fold sense perception sense perception
sampaṭicchana
paryeṣaka mano- santīraṇa mano- paryeṣaka

vijñāna vijñāna
(vikalpyate)
niścita voṭṭhapana vyavasthāpaka
javana >> cetanā << vikalpaka
tadārammaṇa vāsanā in ālayavijñāna
The sources compared above all assume the manovijñāna to arise subsequent
to the sense perception, the data of which it investigates, determines, and
conceptualizes.32 However, in a number of other Yogācāra sources, such as
the Viniścayasaṃgrahaṇī, the Saṃdhinirmocanasūtra, and, as already indi-
cated above, the Triṃśikā and its commentary by Sthiramati, an alternative
system is being taught. According to these sources, all six kinds of perception
may arise at the same time and manovijñāna is always active (except for some
30 My understanding of tadārammaṇa follows that of Lance S. Cousins (personal communi-

cation).
31 PSkV 49b6.
32 Another work that seems to advocate this idea is the Viṃśikā (see Schmithausen, 1967:
124f.; on the title Viṃśikā instead of Viṃśatikā, see Kano, 2008: 345). The concept of
manovijñāna following the sense perception is also mentioned in Sthiramati’s commen-
tary on the Mahāyānasūtrālaṃkāra (SAVBh mi 161b6f.). There, manovijñāna is only quali-
fied as rnam par rtog pa dang bcas pa (savikalpa).
unconscious and meditative states) with or without any of the other five sense
perceptions.33
Remarkably, the two views of a “multilayered” manovijñāna on the one hand
and the simultaneously arising pravṛttivijñānas on the other seem to exclude
each other. Texts which list the various stages of manovijñāna do not mention
the simultaneity of the perceptions, and works mentioning the simultaneity
do not analyze manovijñāna into various stages.34 What is even more strik-
ing is the fact that Sthiramati uses both systems in his commentaries without
discussing this obvious contradiction explicitly. As already mentioned, in the
Pañcaskandhakavibhāṣā he sticks to the first model whereas in his Triṃśikā
commentary he states the simultaneity of perceptions. In the latter he even
explains the reason why others do not accept the possibility of more than
one pravṛttivijñāna at the same time, without indicating that he himself
also propagates the concept of succeeding single moments of mind in the
Pañcaskandhakavibhāṣā.35 The argument that is adduced against the simul-
taneity of several pravṛttivijñānas says, according to Sthiramati, that in some
cases there would not be enough immediate conditions (samanantarapraty-
aya) for more than one vijñāna.36 This refers to the possibility that there is, for
instance, only one visual perception in the present moment, but two or more
perceptions in the following moment. In such a case, the samanantarapraty-
aya for only one perception would be available and the second would lack an
immediate condition. Sthiramati objects to this opposing view by stating that
the whole cluster of the preceding vijñānas functions as the samanantarapra-
tyaya for the following cluster of vijñānas, no matter how many perceptions it
includes.37
33 TrBh 104,13–15.
34 However, it should be noted that, for instance, in the Saṃdhinirmocanasūtra the mano
vijñāna that accompanies the sense perceptions is characterized as vikalpaka (Saṃdh V.4:
rnam par rtog pa’i yid kyi rnam par shes pa).
35 It should be noted, however, that Sthiramati does not explicitly state in his Pañcaskand-
haka commentary that the simultaneous existence of several vijñānas is impossible. For
passages in which such a simultaneous existence is explicitly denied, see Schmithausen,
1987: 16f., note 303.
36 TrBh 102,9–104,10.
37 Interestingly, both models of perception consider manas to be the samanantarapratyaya
for the following mental moment. Thus, the question arises whether the concept of
manas is used with two different connotations in the present context. For a more detailed
discussion of the concept of manas, see below.
156 Kramer
Coming back to our investigation of the second category of proofs for the
existence of the ālayavijñāna, it can be stated that the following proofs are
concerned with the ālayavijñāna’s relation to the pravṛttivijñānas:
a. Yogācārabhūmi proofs 2–5

b. Mahāyānasaṃgraha proof 1
c. Pañcaskandhaka(vibhāṣā) proof 2
1.3 Third Category of Proofs

The proofs which have not been assigned to any of the first two categories are
all related to the issue of “liberation from saṃsāra” and can therefore be
grouped in a third category. Arguments dealing with this topic are to be found
in the Mahāyānasaṃgraha and the Pañcaskandhaka(vibhāṣā), but not in the
Yogācārabhūmi:
a. Mahāyānasaṃgraha proofs 1 (only paragraph 32), 4, and 5

b. Pañcaskandhaka(vibhāṣā) proof 4b
The fact that the “Proof Portion” does not include a discussion of the topic of
purification and liberation in connection with the ālayavijñāna probably
shows that the debates surrounding the concept of the store mind have
changed in the course of time and that the aspect of purification became rel-
evant only at a later developmental stage.
The arguments of this category of proofs can be subdivided into two sub-
groups. The first of them is concerned with the problem that the storage of
seeds of contaminations would be impossible without the ālayavijñāna. It is
discussed in the first proof of the Mahāyānasaṃgraha and in the second part
of the fourth proof of the Pañcaskandhakavibhāṣā. The problem originates
from the assumption that the contaminations and their seeds cannot exist in a
moment of mind that counteracts the contaminations. Thus, as soon as the
counteracting mind has arisen in a person on the path to liberation, there is,
without the existence of the ālayavijñāna, no possibility to store the seeds of
those contaminations which have not been removed yet. Thus, a multi-layered
mind stream is needed for the removal of the contaminations, allowing a par-
allel existence of seeds and their antidotes. This seems to be the main concern
of the argument in the Mahāyānasaṃgraha. The argumentation in the Pañca
skandhakavibhāṣā appears to have a slightly different focus or at least to
include another closely related but somewhat different point, namely the con-
cept that in order to remove the contaminations, the seeds of the latter and
their antidotes have to be existent in the person simultaneously (otherwise
they could not be removed by the antidotes).38 Sthiramati explains that the
contaminations have to be abandoned either in their active forms or as seeds.
The first alternative is not acceptable since, as mentioned above, a moment of
the counteracting mind cannot contain a contamination. Thus, it must be the
seed of the contamination that is present, but the seed can neither exist in the
counteracting mind itself. Therefore a parallel “store” is needed that allows the
simultaneous existence of the seed of the contamination and its antidote.39
The second subgroup of arguments assigned to this third category com-
prises proofs 4 and 5 of the Mahāyānasaṃgraha. These proofs are also
associated with the process of liberation but go beyond the explanations pre-
sented in the context of the first subgroup. Proof 4 deals with the assumption
that a moment of mind belonging to the desire realm (kāmadhātu) cannot
be the seed of the mind belonging to the form realm (rūpadhātu). Therefore
there must be a seed for the latter stored in the ālayavijñāna. The fifth proof
of the Mahāyānasaṃgraha is concerned with the question where the supra-
mundane forms of the mind originate from, and it has already been discussed
in some detail by Lambert Schmithausen.40 In this context, the store mind is
regarded to be necessary in order to store the (mundane) “imprint of listen-
ing” (śrutavāsanā) which originates from the true reality (dharmadhātu) and
later becomes the seed of supramundane insight (or factors leading to it). This
imprint is left in the mental series of a person who listens to the Buddhist doc-
trine and “coexists” with the ālayavijñāna without becoming part of it like milk
coexists with water.41
1.4 Summary
Summing up the results of the comparison of the three lists of proofs the fol-
lowing distribution across the three categories emerges:
1. ālayavijñāna’s relevance for the presence of the mind in the body

– Yogācārabhūmi proofs 1 and 6–8; Mahāyānasaṃgraha proofs 2, 3, and
6; Pañcaskandhaka(vibhāṣā) proofs 1, 3, and 4a
2. ālayavijñāna’s relevance for the actual perceptions – Yogācārabhūmi
proofs 2–5; Mahāyānasaṃgraha proof 1; Pañcaskandhaka(vibhāṣā)
proof 2
38 PSkV 57a3ff.
39 See also Kramer, 2014b: 318f.
40 Schmithausen, 1987: 79f.
41 Schmithausen, 1987: 370, note 580.
158 Kramer
3. ālayavijñāna’s relevance for liberation from saṃsāra –

Mahāyānasaṃgraha proofs 1, 4, and 5; Pañcaskandhaka(vibhāṣā) proof
4b.
2 Evolution of the Concept of manas
As indicated above, the second part of this chapter is concerned with the
concept of manas. In the Pañcaskandhaka and in Sthiramati’s commentary,
manas is first explained as the basis for the following moment of mind.42
However, Vasubandhu also mentions that the main function of manas in the
context of the vijñānaskandha is to be the vijñāna that has the ālayavijñāna
as its object and that is always associated with the contaminations of “wrong
attitude towards the self” (ātmamoha), “false view of the self” (ātmadṛṣṭi),
“conceited conception of the self” (ātmamāna), and “self-love” (ātmasneha).43
In his comments on this passage, Sthiramati further specifies this vijñāna as
the “contaminated mind” (kliṣṭaṃ manaḥ) and explains that the latter con-
tinuously takes as its object the ālayavijñāna in the form of the self (ātman).44
According to Sthiramati, the manas which is the basis for mental perception
(manovijñāna) is different from the manas which has the nature of conceit
(manyanā).45
In canonical sources manas seems sometimes to be regarded as a kind of
a “mental sense faculty” that experiences its objects. But the main function
of this faculty was probably to reflect on the data provided by the other sense
faculties.46 In Abhidharmic context this view of manas as a mental sense organ
corresponding to the five material sense faculties has been given up, possibly
because of its close affinity to the concept of a permanent self or ātman. The
fact that such a mental sense organ was not accepted is obvious for instance
from statements in the Abhidharmakośabhāṣya. There it is explained that
there is no manas (or manodhātu) besides the six vijñānadhātus and that
manas is only the designation for one of the six vijñānas that has just passed
(anantarātīta).47 Moreover, the text states that manas is included in the
42 PSk 16,8: manaḥsanniśrayatām; PSkV 49a3f.: ṣaṇṇāṃ cakṣurādivijñānānāṃ yad yan nirud-
hyate / tat tad anantarasya vijñānasyotpadyamānasyāśrayabhāvena vyavatiṣṭhata ity ataḥ
samanantaracittasanniśrayatām upādāya mana ity ucyate.
43 PSk 17,7f.
44 PSkV 59a5f.
45 PSkV 59b1. See also Kramer, 2014b: 313.
46 See Vetter, 2000: 64, note 135, 101, note 7 and Boisvert, 1995: 120f.
47 AKBh 11,18–21.
structure of eighteen dhātus – although it is merely a designation for the past

moment of vijñāna – in order to establish a basis for the manovijñāna paral-
lel to the points of support of the other five perceptions, which are the five
sense faculties.48 The reason for introducing the idea of manas as an āśraya
for manovijñāna is explained as follows in the Abhidharmakośabhāṣya: While
the five sense perceptions are based on the respective sense faculties, the men-
tal perception was lacking such a basis.49 In the Mahāyānasaṃgraha, manas
is referred to as being “the immediate condition” (samanantarapratyaya) and
the basis for the arising (utpattyāśraya) of vijñāna.50
Although the function of manas as the immediately preceding basis for the
arising of the next moment of vijñāna in general (as it seems to be understood,
e.g. in MSg I.6.1)51 and its role as the support of manovijñāna in particular (as
explained in AKBh 11,25) appear similar, it remains unclear how exactly these
two aspects are related to each other. It is also uncertain how the concept of
manas as the basis for the arising manovijñāna is to be understood in view of
the above-mentioned idea of an investigating, determining, and conceptual-
izing manovijñāna that follows the sense perception. Is it right to assume that
the mental sense data provided by the actual perception was regarded to be
the manas for the following analyzing manovijñāna? Another doubtful issue
concerns the relation between the last stage of the conceptualizing process of
the manovijñāna and the next moment of actual sense perception. Strictly
speaking, the immediately preceding mental incident before the next moment
of actual perception (i.e. the manas) is the conceptualizing manovijñāna. But
it seems that the concept of manas as the samanantarapratyaya for the next
moment of actual perception does not refer to this manovijñāna in particular
but to the whole preceding vijñāna cluster in general, beginning with the
actual perception and ending with the conceptualizing manovijñāna. However,
since no explicit statements are made with regard to these issues in the texts
under discussion, it cannot be determined with certainty if this is the way in
which the authors understood the functions of manas. Moreover, it is unclear
whether authors who claim the simultaneity of manovijñāna and the sense
perceptions presumed that the vijñāna clusters follow each other without an
interstage of a separate manovijñāna. If we assume that this was the case, did
48 AKBh 11,25.
49 AKBh 11,26f.
50 MSg I.6.1.
51 An example for an explicit definition of manas as the immediately preceding basis
(samanantarāśraya) for the next moment of perception in general (i.e. of all six vijñānas)
is found in Y 4,6f., 6,5, 7,6 and 22, 8,14, and 11,9f. See also Schmithausen, 1987: 110f., 124.
160 Kramer
they think that the manovijñāna analyses the sense-data perceived in the last
moment while the other perceptions are busy perceiving the objects of the
present moment? Or did they assume that the manovijñāna is able to “access”
the sense-data immediately, that is in exactly the same moment as it is per-
ceived by the sense perceptions? Unfortunately, these questions have to remain
unanswered since there is no explanation given in this regard in the texts
under discussion here.52 It must also be taken into account that the differentia-
tion between the two concepts of a succeeding and a simultaneous manovijñāna
was not always considered in detail when the often stereo-typed definitions of
manas were created. In general, the statements regarding the nature of manas
found in Abhidharmic and Yogācāra works give the impression that the con-
cept was not thought through thoroughly and that several slightly inconsistent
models of manas existed concurrently.
The problem of a missing basis for the manovijñāna was only partly solved
by the assumption of manas as having this function in the form of the imme-
diately preceding mental moment. In the course of further organizational
development another desideratum became manifest in connection with the
manovijñāna, namely the need of a “simultaneously present basis” (sahabhūr
āśrayaḥ) or “dominant condition” (adhipatipratyaya) for the latter. While the
five indriyas provide a continuously present support for the five sense facul-
ties, the manovijñāna’s basis in the form of manas is present only in the instant
immediately before the occurrence of the manovijñāna itself. This inconsis-
tence must have been regarded a weak point of the system – especially as
manas was a mere designation and not a “real” vijñāna or mind on its own.
Some texts, as for instance the Pañcaskandhakavibhāṣā, seem to have ignored
this problem. Sthiramati classifies the four conditions necessary for a per-
ception in the following way: the imprints stored in the ālayavijñāna are the
“causal condition” (hetupratyaya), the vijñāna that has just passed, i.e. manas,
is the “immediate condition” (samanantarapratyaya), the sense objects, visible
matter etc., are “object conditions” (ālambanapratyaya), and the sense facul-
ties (indriya) are the “ruling condition” (adhipatipratyaya). Since the indriyas
most probably do not include manas but only the five sense faculties which
are mentioned in the rūpa chapter of the Pañcaskandhaka,53 Sthiramati does
52 As mentioned above, the third proof of the Yogācārabhūmi seems to presuppose the idea
of manovijñāna accessing the data it analyses directly. However, since the argument is not
explained in detail, it appears difficult to say with certainty if this statement is really aim-
ing at the concept of perception and analysis of a particular object happening at exactly
the same time.
53 PSk 2,3f.
not seem to designate any adhipatipratyaya for manovijñāna.54 Other scholars,

however, were obviously bothered by this deficiency and developed an alter-
native model for a basis being simultaneously present with the manovijñāna.
This alternative concept becomes visible (probably for the first time) in the
Mahāyānasaṃgraha, where the new form of manas, the kliṣṭamanas, is
declared to be the sahabhūr āśrayaḥ of manovijñāna.55 This functional combi-
nation of the old and the new manas is particularly interesting, as it manifests
in a rather obvious way how these two originally unrelated lines of thought
became intertwined with each other over time. As already mentioned, at the
beginning of this development we find manas as a “real” mental faculty – a
view which was evidently given up in Abhidharmic scriptures. What remained
was the concept of manas as an āśraya, a basis, which, however, was consid-
ered a mere designation and not a “real” entity. Thus, on the one hand manas
had always the connotation of being a basis, on the other hand – because it
was only a designation – it provoked the wish for a real āśraya, and especially
for an adhipatipratyaya for manovijñāna.
Besides this process, we have the conceptual development of an idea which
finally resulted in the creation of the term kliṣṭaṃ manaḥ. As has been shown
by Lambert Schmithausen, the origins of this concept do not go back to the
traditional manas but mainly to the idea of “the notion ‘I am [this]’” (asmimāna)
and of sahajā satkāyadṛṣṭiḥ, the inborn false view of the five skandhas as being
the self, which were considered to have the “basis of personal existence”
(ātmabhāva) as their object. The reason for these “ordinary” forms of the feel-
ing of identity having been remodeled into a subtle, subliminal kind of vijñāna
(i.e. the kliṣṭamanas) could be that the store mind, which had evolved into the
new fundamental layer of personality and thus the objective basis of the
notion of “I,” was considered to be a subliminal form of the mind itself.56 The
question of why the expression manas was chosen for the newly introduced
kind of mind is difficult to answer definitely, but one of the reasons was cer-
tainly the fact that the concept of manas had become such an empty and
unoccupied term in the meantime. Another favorable factor was doubtlessly
the possibility of etymologizing the term manas in the sense of manyanā
(“conceiving“), which expresses the nature of the new manas in a very appro-
54 Another example is Y 4,6ff., which specifies a sahabhūr āśrayaḥ for the five sense percep-
tions but omits this category silently with regard to the manovijñāna (Y 11,9). Note that the
*Pañcaskandhavivaraṇa (PSkViv 25a6) and the *Pañcaskandhabhāṣya (PSkBh 95a6), two
further commentaries on the Pañcaskandhaka, mention six indriyas as adhipatipratyaya.
55 MSg I.7.2. See also Schmithausen, 1987: 124.
56 See Schmithausen, 1987: 147–152.
162 Kramer
priate way.57 From the point of view of organization, the term manas might
have appeared suitable insofar as the triad of vijñāna, citta, and manas was
already an established structure in the Abhidharmic context.
The analogy of names of the “traditional” and the “new” manas resulted in an
increasing overlapping of their functions. Already in the context of what might
be the first explicit occurrence of the new manas in the Viniścayasaṃgrahaṇī,
the new manas with its specific functions of conceiving and establishing the
feeling of identity is also claimed to be the basis of manovijñāna.58 The text
explains that this is stated in the sense of the manovijñāna’s being bound
to cognizing its object as long as manas is not interrupted. A similar idea is
expressed in another part of the Yogācārabhūmi, in which manas is taught to be
the basis for the conceptualizing aspect (vikalpa) of manovijñāna,59 and in the
Mahāyānasaṃgraha, which describes manas as the basis for the defilement
(saṃkleśa) of vijñāna.60 All these statements are clearly based on the assump-
tion of a manas which has the nature of the “new” manas, being responsible
for the concept of “I” and “mine” in a person. This manas was obviously con-
sidered an important constituent of the individual’s conceptualizing process
of everyday phenomena and thus as accountable for the contamination of the
mind. At the same time these characteristics of the new manas were combined
with the traditional function of manas of being a basis, in particular a basis
of manovijñāna. Finally, all these developments and probably also the wish
for giving the kliṣṭamanas an appropriate position within the system of the
four conditions (pratyaya) resulted in the equation of the kliṣṭamanas with the
sahabhūr āśrayaḥ, setting the kliṣṭamanas on one level with the sense faculties
(indriya). Thus, it seems that the final concept of (kliṣṭa)manas resembles the
original idea of manas as a “real” mental faculty.
Finally, it should be mentioned that the Mahāyānasaṃgraha seems to be
the oldest (available) source containing proofs for the existence of the
kliṣṭamanas. The text provides six arguments which aim at showing that the
57 On this etymology, see Schmithausen, 1987: 149.

58 The edition of the Tibetan text of this passage is to be found in Hakamaya, 1979: 32f. It
should be noted that according to Schmithausen, 1987: 501, note 1351, this paragraph is a
later addition (see also pp. 488ff., note 1298).
59 On this passage, see Schmithausen, 1987: 487, note 1297.
60 MSg I.6.2. Note that in SAVBh mi 126b5 kliṣṭamanas is explicitly equated with the manas
which is the basis for manovijñāna in the same way as the visual faculty is the basis for
visual perception.
rejection of the kliṣṭamanas would result in unacceptable consequences.61

According to the Mahāyānasaṃgraha’s list, if there was no kliṣṭamanas
1. the “specific” ignorance (āveṇikī avidyā)62 could not exist;

2. manovijñāna would have no basis parallel to the sahabhūr āśrayaḥ of
the other five sense perceptions, which are the five indriyas;
3. the etymological interpretation (nirukti) of manas would not be
possible without the understanding of it as manyanā, “conceiving”
(which is the nature of kliṣṭamanas);
4. there would be no difference between asaṃjñi- and nirodhasamāpatti
(as the difference between the two states is that kliṣṭamanas is present
in the former but absent in the latter);
5. the existence of the unconscious beings (āsaṃjñika) could not be
continuously contaminated, because there would be no ahaṃkāra and
asmimāna there, of which kliṣṭamanas is the basis;
6. ahaṃkāra could not occur in neutral and good (but only in bad) mental
states.
It is notable that these arguments either are not mentioned at all in later
works or are adopted from the Mahāyānasaṃgraha without any substantial
modifications,63 whereas the proofs of the ālayavijñāna are presented in at
least three different systems. In the Pañcaskandhaka(vibhāṣā), kliṣṭamanas is
treated far less exhaustively than the ālayavijñāna, and neither Vasubandhu
nor Sthiramati mention the proofs for its existence.
3 Conclusion
In conclusion, it can be stated that there are (at least) three different tradi-
tions of ālayavijñāna proofs to be found in Yogācāra texts, which, taking into
consideration the differences in terminology and in the way the arguments
are presented, do not seem to be directly related. Nonetheless, the argu-
ments are dealing with similar issues in general and most of the proofs of
the Pañcaskandhaka(vibhāṣā) have parallels in the other lists. In some cases
61 See MSg I.7. The six proofs have been investigated by Lambert Schmithausen in an unpub-
lished study entitled “Satkāyadṛṣṭi, asmimāna und kliṣṭamanas”.
62 Sthiramati explains the ignorance characterized as āveṇika to be the “inborn” (sahaja)
ignorance, which is not associated with the other contaminations (PSkV 31a5f.).
63 See, for instance, Yid kun 48b-52a.
164 Kramer
the arguments appear to be similar at the first glance, but on closer look it
becomes obvious that their focus is different, as for instance in proofs 1 and
6–8 of the “Proof Portion,” which seem to be mainly concerned with vitalizing
aspects, whereas the corresponding passages in the Pañcaskandhakavibhāṣā
emphasize the aspect of continuity. All these divergences might reflect various
stages of development of the concept of ālayavijñāna. This probably results
from the fact that slightly different functions of the store mind were under
discussion when the relevant Yogācārabhūmi passages were composed and
when the Pañcaskandhaka and its commentary were produced. This idea is
also supported by the fact that some of the proofs, as for instance the proofs
of the “Proof Portion” concerning the simultaneous arising of various vijñānas
are not mentioned at all in the lists of the Pañcaskandhaka(vibhāṣā) and the
Mahāyānasaṃgraha.
The twofold understanding of the concept of manas which manifests itself
in the Pañcaskandhaka(vibhāṣā) is the result of the combination of various
lines of thought that developed within Buddhist philosophy with regard to the
basis of thinking, the contamination of the mind, and the notion of “I”. While
manas was probably considered to be a really existing “mental sense faculty” at
its earliest developmental stage, it was reduced to a mere designation of the
immediately preceding moment of mind in Abhidharmic sources. However, it
never lost its relevance as the basis for the arising of manovijñāna and finally
became, in combination with the Yogācāra concept of kliṣṭamanas (which,
according to Lambert Schmithausen, most probably developed from the desire
to remodel the concepts of asmimāna and the sahajā satkāyadṛṣṭiḥ into a sub-
liminal form of mind), the “simultaneously present basis” (sahabhūr āśrayaḥ)
of the manovijñāna.
Abbreviations
AKBh Abhidharmakośabhāṣya (ed. Pradhan, 1967)

ASBh Abhidharmasamuccayabhāṣya (ed. Tatia, 1976)
MN Majjhimanikāya (ed. Trenckner, 1888)
MSg Mahāyānasaṃgraha (ed. Lamotte, 1938)
PSk Pañcaskandhaka (ed. Li and Steinkellner, 2008)
PSkBh Pañcaskandhabhāṣya (author unknown, probably not *Pṛthivībandhu)
(Derge 4068)
PSkV Pañcaskandhakavibhāṣā (ed. Kramer, 2014)
PSkViv *Pañcaskandhavivaraṇa (Guṇaprabha) (Derge 4067)
Saṃdh Saṃdhinirmocanasūtra (ed. Lamotte, 1935)
SAVBh *Sūtrālaṃkāravṛttibhāṣya (Sthiramati) (Derge 4034)

TrBh Triṃśikāvijñaptibhāṣya (ed. Buescher, 2007)
Vism Visuddhimagga (ed. Rhys Davids, 1975)
Y Yogācārabhūmi (ed. Bhaṭṭācārya, 1957)
Yid kun Yid dang kun gzhi’i dka’ ba’i gnas rgya cher ’grel pa (reproduced in Kelsang and
Odani, 1986)
Bibliography
Bhaṭṭācārya, Vidhuśekhara (ed.) (1957). Yogācārabhūmi. Calcutta: University of Calcutta.

Boisvert, Mathieu (1995). The Five Aggregates: Understanding Theravāda Psychology and
Soteriology. Ontario: Wilfrid Laurier University Press.
Buescher, Hartmut (ed.) (2007). Triṃśikāvijñaptibhāṣya. Vienna: Austrian Academy of
Sciences Press.
Griffiths, Paul (1986). On Being Mindless: Buddhist Meditation and the Mind-Body
Problem. La Salle IL: Open Court.
Hakamaya, Noriaki (1978). “Āraya shiki sonzai no hachi ronshō ni kansuru shobunken
アーラヤ識存在の八論証に関する諸文献 (*Materials on the Eight Proofs of the
Existence of Ᾱlayavijñāna)”, Komazawa Daigaku Bukkyōgakubu Kenkyū Kiyō 駒澤大
學佛教學部研究紀要 36, pp. 1–26.
Hakamaya, Noriaki (1979). “Viniścayasaṃgrahaṇī ni okeru āraya shiki no kitei におけ
るアーラヤ識の規定 (*The Definition of Ᾱlayavijñāna in the Viniścayasaṃgrahaṇī)”,
Tōyō Bunka Kenkyū-jo Kiyō 東洋文化研究所紀要 79, pp. 1–79.
Kano, Kazuo (2008). “Two Short Glosses on Yogācāra Texts by Vairocanarakṣita:
Viṃśikāṭīkāvivṛti and *Dharmadharmatāvibhāgavivṛti”, In Sanskrit Texts from
Giuseppe Tucci’s Collection, part 1. Edited by Francesco Sferra. Roma: Istituto Italiano
per l’Africa e l’Oriente, pp. 343–380.
Karunadasa, Yakupitiyage (2010). The Theravāda Abhidhamma: Its Inquiry into the Nature
of Conditioned Reality. Hong Kong: Centre of Buddhist Studies, The University of
Hong Kong.
Kramer, Jowita (ed.) (2014a) Sthiramati’s Pañcaskandhakavibhāṣā. Beijing/Vienna:
China Tibetology Publishing House/Austrian Academy of Sciences Press.
Kramer, Jowita (2014b). “Indian Abhidharma Literature in Tibet: The Section on Vijñāna
in Sthiramati’s Pañcaskandhakavibhāṣā”, In Buddhism Across Asia: Networks of
Material, Intellectual and Cultural Exchange. Edited by Tansen Sen. Singapore:
Institute of Southeast Asian Studies, pp. 309–325.
Lamotte, Étienne (1935). Saṃdhinirmocana Sūtra: L’Explication des Mystères. Louvain:
Bureaux du Recueil / Paris: Adrien Maisonneuve.
166 Kramer
Lamotte, Étienne (1938). La Somme du Grand Véhicule d’Asaṅga. vol. 1, Louvain: Bureaux
du Muséon.
Li, Xuezhu and Steinkellner, Ernst (ed.) (2008). Pañcaskandhaka. Beijing/Vienna: China
Tibetology Publishing House/Austrian Academy of Sciences Press.
Pradhan, Prahallad (ed.) (1975). Abhidharmakośabhāṣyam of Vasubandhu (Tibetan
Rhys Davids, Caroline A.F. (ed.) ([1920–1921] 1975). Visuddhimagga. Pāli Text Society,
London: Routledge & Kegan Paul.
Schmithausen, Lambert (1967). “Sautrāntika-Voraussetzungen in Viṃśatikā und
Triṃśikā”, Wiener Zeitschrift für die Kunde Süd- und Ostasiens 11, pp. 109–136.
Schmithausen, Lambert (1967; unpublished paper). “Satkāyadṛṣṭi, Asmimāna und
Kliṣṭamanas”.
Schmithausen, Lambert (1987). Ālayavijñāna: On the Origin and the Early Development
of a Central Concept of Yogācāra Philosophy. Tokyo: The International Institute for
Buddhist Studies.
Schmithausen, Lambert (2014). The Genesis of Yogācāra-Vijñānavāda: Responses and
Reflections. Tokyo: The International Institute for Buddhist Studies.
Tatia, Nathmal (ed.) (1976). Abhidharmasamuccayabhāṣya. Patna: K.P. Jayaswal Research
Institute.
Trenckner, Vilhelm (ed.) (1888). Majjhimanikāya. London: The Pāli Text Society.
Vetter, Tilmann (2000). The ‘Khandha Passages’ in the Vinayapiṭaka and the Four Main
Nikāyas. Wien: Verlag der Österreichischen Akademie der Wissenschaften.
Waldron, William (1995). “How Innovative Is the Ālayavijñāna: The Ālayavijñāna in the
Context of Canonical and Abhidharma Vijñāna Theory (Part II)”, Journal of Indian
Philosophy 23/1, pp. 9–51.
Yamabe, Nobuyoshi (2015). “A Reexamination of on Being Mindless: Possible Meditative
Implications of the Eightfold Proof of Ᾱlayavijñāna”, In Buddhist Meditative Praxis:
Traditional Teachings and Modern Applications. Edited by KL Dhammajoti. Hong
Kong: Centre of Buddhist Studies, the University of Hong Kong, pp. 137–176.
Yid dang kun gzhi’i dka’ ba’i gnas rgya cher ’grel pa (Tsong kha pa), reproduced in Tsultrim
Kelsang and Nobuchiyo Odani (1986). アーラヤ識とマナ識の研究：クンシ・カン
テル (*A Study of Tsong kha pa’s Treatise on Ālayavijñāna and Manas, the Kun gzhi
dka’ ’grel), Kyoto: Bun’eido, pp. 147–184.
Part 2
Intellectual History
∵
168 Kramer
Sanskrit Abhidharma Literature of the Mahāvihāravāsins 169
Chapter 5
Sanskrit Abhidharma Literature of the

Mahāvihāravāsins
Lance S. Cousins
1 Introduction
The argument of this chapter can be summarized as follows:
1. We learn from the Cūlavaṃsa that a monk named Jotipāla played an

important role in the Buddhism of the island of Ceylon at the end of the
sixth century and during the early decades of the seventh century. In a
previous article, I have discussed his important role in the development
of the Pāli ṭīkā literature.1 In the first section of this chapter I briefly
review the evidence we have for his work in Sanskrit.
2. The Visudhimagga with the Commentary written by King Parākramabāhu
II (Vism-sn) cites passages from a work called the Āryasatyāvatāra (ĀS)
and from another which is named either Jñeya-sampatti-ṭīkā or Jñeya-
saptati-ṭīkā (JS-ṭ). Given the manner in which they are cited, it seems
likely that this refers to a single work and its commentary – the Ārya-
satyāvatāra would be the name of a work in kārikās, which would have
an accompanying commentary, both being associated with the name of
Jotipāla. The Jñeya-saptati-ṭīkā would be the name of this commentary.
3. One of the views given in the JS-ṭ passages in Vism-sn is attributed in
Sumaṅgala’s twelfth century commentary on the Abhidhammāvatāra
(Abhidh‑av‑ṭ) to Jotipāla. A passage in Abhidh-av-ṭ which gives a saying
of Jotipāla is described in the thirteenth century ṭīkā to the Sacca-
saṅkhepa as from the Aññeyya-sattati-ṭīkā.
4. Vism-sn contains around sixty-six Sanskrit citations. A few of these are
from general Indian sources, from the grammatical literature or from
the literature of non-Theravādin forms of Buddhism. The great majority
are specific to the Theriya Abhidha(r)mma tradition.
5. Conclusion: Jotipāla emerges as a major figure who defended the
position of the Mahāvihāravāsins in Sanskrit, very much the literary
1 Cousins, 2011.

170 Cousins
language of the day. It seems likely that the Sanskrit language played a
greater role in the history of the Theriya school of Southern India and
Ceylon than has been hitherto realized. This no doubt accounts for
some of the increasing North Indian awareness of Theravāda during this
period.
2 Who was Jotipāla?
It is clear that Jotipāla was a figure of major importance in the Abhidhamma

tradition of the Mahāvihāravāsins. He is frequently cited by Sumaṅgala in the
early thirteenth century and by Parākramabāhu II in the mid thirteenth.
Indeed at Abhidh-av-ṭ II 177 and 185 he is cited with Dhammapāla in second
place! His Anuṭīkā is a commentary on the Mūla-ṭīkā of Ānanda – either the
extant one or an earlier one on which the present Anuṭīkā is based. The ideas
of Ānanda are extremely important to the development of the Abhidhamma
component of the ṭīkās, but completely unknown even to Nidd-a and Paṭis-a.
If he is identified, as seems likely, with the Jotipāla mentioned in the Cūlavaṃsa,
then his floruit would be ca. 600 CE.
Heinz Bechert points out that Mahāyāna influences reached their zenith in
Ceylon in the eighth and early ninth centuries.2 This receives striking support
for the present period from the accounts of Xuanzang 玄奘, writing only a little
earlier. According to the Chinese pilgrim, there were 10,000 ‘Sthavira’ monks in
100 monasteries in the Tamil country and 20,000 ‘Mahāyāna Sthavira’ monks
in 200 monasteries in Ceylon.3 These are clearly intended as round figures
but even so there are obvious problems. The ‘Mahāyāna Sthaviras’ in Ceylon
and at Bodhgayā can only refer to the Dhammarucikas (and Sāgalikas?), but
the Mahāvihāra tradition cannot have been totally absent from the island!
Xuanzang must be getting his information on Ceylon from Mahāyānist sources,
probably directly or indirectly from the monks of Bodhgayā – possibly it was
simply assumed that all monks in the island were of the same school as those
present in the north of India. I take it that in ‘Mahāyāna Sthavira’ ‘Mahāyāna’
is an epithet, not part of a school name and simply means Sthaviras who also
practice Mahāyāna.4
The information concerning the Tamil country which Xuanzang may have
visited in person might be more reliable. Even so it cannot be literally true. We
2 Bechert, 1998: 2.
3 Lamotte, 1958: 597; Li, 1996a: 320 and 331; Li, 1995: 119.
4 A different view: Deeg, 2012.
should take it that this Chinese evidence indicates that the Mahāvihāra tradi-
tion was at this time predominant among ‘Sthaviras’ in South India (excluding
Mahāyānist Mahāsaṃghikas) and weak in its homeland or possibly strong only
in its hinterland in the south of the island. It is perhaps worth noting that
Xuanzang is using the terminology of the four fundamental Nikāyas; so when
he refers e.g. to Sāmitīyas, this probably means any of the four or so Pudgalavādin
schools. However, by this date the name ‘Sthavira’ or similar seems to be used
exclusively of the three schools which originate from the three major monas-
teries of Anurādhapura.
Xuanzang’s account is supported by the data provided by another Chinese
pilgrim, some decades later. In particular, Yijing 義淨 notes the complete rejec-
tion of the Mahāsaṃgha in Ceylon and does not record any presence in the
south of the four nikāyas which are grouped as the Mūlasarvāstivādin funda-
mental nikāya. The three ‘Sthavira’ schools are mentioned as predominating in
the south, especially Ceylon.5
According to the Nikāyasaṃgraha, Jotipāla came from Jambudīpa. That
is perhaps also implied in the Cūlavaṃsa, if we understand that the king is
specifically mentioned as arranging for him to dwell in the Mahāvihāra. It is
likely then that he came from the major center of non-Mahāyāna Theravādin
tradition in the Tamil country. Given the obvious similarity of the names
Dhammapāla and Jotipāla we should look to some kind of pupil-teacher or
monastic relation between the two.6
3 The Verses Ascribed to Jotipāla in Vism-sn
In AS I, eight Sanskrit stanzas are given as attributed to Jotipāla. Three of these

are explicitly named in Vism-sn and five are sufficiently closely connected
that I include them also. Although I gave them and translated them in AS I
together with some similar Pāli passages from the ṭīkā literature, I give them
again here so as to collect all of Jotipāla’s Sanskrit verses together in one place.
They concern three main topics. The first is a presentation of part of the argu-
ments in support of the Theravādin notion of the heart-base (hadayavatthu).
That of course is known in Sanskrit Buddhist circles as a view specific to the
Sthaviravādin tradition. One stanza is attributed specifically to Jotipāla and a
second occurs on the next page of Vism-sn, continuing the same discussion.
5 Deeg, 2012: 602; Li, 1996b: 12.

6 See also: Malalasekera, 1928: 210f.
172 Cousins
The stanza in defence of the heart-base that cites Jotipāla-māhimiyo by name

is the following one:
Vism-sn III 1060 (to Vism 447):

Vastv-āśrayor dvayor dhātvo rūpâvabaddha-vṛttitaḥ
hṛd dhi tau dvāv upādāya rūpâśrayo bhava-dvaye. (1)
Because the two elements that have the [heart-] base as their support
operate bound up with rūpa,
the heart is certainly the support for rūpa in two [kinds of] existence in
dependence upon those two [elements]. (cp. Abhidh-av 674; Sacc 12)
The same discussion is continued in:
Tad vastu-bhāvāt karmotthaṃ dṛśvat pratīyana-kriyaṃ

hṛd-khedāc cârtha-cintāyāṃ tatra-stham iti gamyate. (2)
That is understood as originated by karma because it is a base (vastu),

fixed in its function like the eye and positioned there [in the chest]
because the heart becomes tired when one reflects on matters.
The second topic concerns higher knowledge (abhijñā). Again at Vism-sn III
1098 (to Vism 456) the following gāthā are attributed to Jotipāla-māhimiyo:
Nâtrâpy abhijñā dhyānasya dānâder artha-lābhavat

ihânṛśaṃsa-bhūtatvāt phala-dâsambhavād api. (3)
Nânya-bhū-phala-daṃ karma rūpa-pākasya go-caraḥ
karmālamba[ḥ] parittâdi na cety ayam asambhavaḥ. (4)
Higher knowledge does not give results because it is the reward in this
existence of dhyāna just as the obtaining of wealth [is the reward of] giv-
ing and the rest; also because it is impossible.
The reason that it is impossible is that karma does not give fruit on a dif-
ferent level [to its own] and the object of resultant [citta] of the rūpa
[level] is [sign of] the karma, not such objects as small dhammas.
In AS I a number of closely related discussions in the ṭīkās are cited and some
are translated.
The third topic dealt with in AS I concerns five Sanskrit stanzas on the
pratisaṃvid. Again, in Sumaṅgala’s discussion prior to his final mention of
Jotipāla, we find:
nirutti māgadhā-bhāsā, atthato nāma-sammutī ti.
This is part of a treatment of the four paṭisambhidā/pratisaṃvid.7 Five verses

of this are separately cited in Vism-sn III 1034–1039 (to Vism 440–442):
Arthe dharme niruktau ca taj-jñānesu ca tatra tu

kramād bhedaṃ gatās saṃkhyā catasraḥ pratisaṃvidaḥ. (5)
Pratyayotpanna-vāg-artha-mukti-pāka-kriyā-vaśāt
pañcâṅgo ’rthaḥ punaḥ kāryajñāpyaprāpya iti tridhā. (6)
Dharma-pratyaya-vāṅ mārgaḥ śuklaṃ kṛṣṇañ ca pañcadhā;
sa punaḥ kārako hetur jñāpakaḥ prāpakas tridhā. (7)
Niruktir Māgadhī-bhāṣā sā cârthān nāma-saṃvṛtiḥ;
keci dhvāna iti prāhur vijñapty-ākāra-saṃyutaḥ. (8)
Lābhas tāsām asammohāc chaikṣâśaikṣa-pathaḥ kṣaṇe;
prayoge ’rthādim ālambya varttanatas tu sāśravaḥ. (9)
For ease of reference I add a continuous numbering of all the Sanskrit verses
which I attribute to Jotipāla. The first two pādas of stanza nine correspond to
the line cited in Abhidh-av-ṭ. I translate:
The discriminations number four [and] are divided in sequence into the
knowledges of artha, dharma, nirukti and the knowledge of those [knowl-
edges]. (5)
Artha is fivefold by way of (a) what has arisen by a condition; (b) the
meaning of speech; (c) liberation; (d) resultant [citta]; (e) kriyā [citta].
Furthermore, it has three kinds because it may be the result (artha)
which follows from doing, the meaning (artha) which has to be made
known or the goal (artha) which must be attained. (6)
Dharma is fivefold: (a) condition; (b) speech; (c) the path; (d) the white
and (e) the black. Furthermore it is the three kinds of cause: that which
does, that which finds out and that which attains.(7)
7 On the paṭisambhidā/pratisaṃvid, see: Pagel, 1995: 272ff.; 359ff. Also, Samtani, 1971: 53f.; 115ff.;
275 (refs); Griffiths, 1994: 116 note.
174 Cousins
Nirukti is the Māgadhī language; this means that it is what causes one to
understand names.8 Some say that it is sound joined with a form of com-
munication. (8)9
The obtaining [of the discriminations] is [nothing but] the śaikṣa or the
aśaikṣa path because it is due to the absence of confusion at that moment.
But [a discrimination] is subject to āśravas in operation (prayoge)
because it occurs taking artha and the rest as its object. (9)
Presumably, the original included a stanza on the fourth discrimination too.

We should note that the author of the Abhidhamma Commentary (but not
Buddhaghosa?) strongly defends the position of the Vibhaṅga that the third
paṭisambhidā has sounds as its object, not concepts (paññatti). The Mūla-ṭīkā
is more sophisticated here and somewhat avoids the two alternatives; it speaks
instead of a paññatti which is not speech, the point being that sound is indeed
a nāmapaññatti. Both here and in the Anuṭīkā there are rather elaborate dis-
cussions related to the whole issue. I cannot address this rather large topic
here, but it is important to note that it is intimately related to the on-going
debates with Mahāyānists wearing their Dārṣṭāntika-Sautrāntika hat. This may
well have much to do with Jotipāla’s role as defender of the tradition.
5 Sanskrit Quotations in Vism-sn
The Sinhalese sanne to the Visuddhimagga (Vism-sn) cites some sixty-four

Sanskrit stanzas and also four prose passages.10 This is a work which was com-
posed by King Parākramabāhu II who reigned from 1236 to 1270 CE. Given that
he cannot have been very old at his accession, it could not have been written
much before his accession and, more probably, was produced during the
course of his reign. Since Sumaṅgala was a pupil of Sāriputta, his Abhidh-av-ṭ
8 I translate nāmasaṃvṛti as the intended equivalent of nāmapaññatti. cf. Vism-mhṭ II 82:

Sā panâyaṃ sabhāva-nirutti Māgadha-bhāsā. Atthato nāma-paññattī ti ācariyā. Apare
pana yadi sabhāva-nirutti paññatti-sabhāvā, evaṃ sati paññatti abhilapitabbā, na vacanan
ti āpajjati.
9 Nett-a 121: Tattha idaṃ dukkhan ti ayaṃ paññattī ti kakkhaḷa-phusanâdi-sa-bhāve
rūpârūpa-dhamme atītâdi-vasena aneka-bheda-bhinne abhinditvā pīḷana-saṅkhata-
santāpa-vipariṇāmaṭṭhatā-sāmaññena yā kucchita-bhāvâdi-mukhena ekajjhaṃ gaha
ṇassa kāraṇa-bhūtā paññatti. Kā pana sā ti? Nāma-paññatti-nibandhanā taj-jā paññatti.
Viññatti-vikāra-sahito saddo evā ti apare.
10 These and other Sanskrit quotations in Sinhalese literature are discussed in Godakum-
bura, 1943.
must have been written either late in the twelfth century or early in the thir-
teenth, i.e. before Vism-sn.
Vism-sn was not the first sanne on Vism. Sumaṅgala refers explicitly to a
Sinhalese commentary on Vism and evidently understands Dhammapāla to
refer to it.11 So some or all of the Sanskrit passages in Vism-sn are likely to be
taken from there and may therefore be much earlier in date. Certainly, there is
nothing which has so far been dated with certainty to a period after the sev-
enth century. The same seems to be the case with the Pāli quotations in
Vism-sn, although there are a small number whose source cannot be identi-
fied.12 This kind of work is of course not intended as an original composition in
the modern sense. Rather the author seeks to present the tradition with subtle
modifications of detail on specific points.
Most of these Sanskrit passages are in verse. Examining these verses we find
three which are unequivocally from brahmanical literature in general and two
which are from grammatical writing.13 These do not name their source, pre-
sumably because it was obvious. From other Buddhist literature, there is one
stanza which apparently derives from the Saundarananda of Aśvaghoṣa, one
from the Vyākhyāvyukti of Vasubandhu, one attributed to Diṅnāgapāda (cor-
rected from Dīghanāgapāda), and two explicitly cited from the Abhidharmakośa.
I omit one other which could be based upon the Kośa.14
There remain fifty-four ślokas (including those already attributed to Jotipāla)
which mostly concern specifically Buddhist topics. There has been a tendency
to assume that these will be from otherwise unknown Buddhist Sanskrit, i.e.
non-Theravādin, authors. This must be mistaken. Many of them describe posi-
11 Abhidh-av-ṭ II 125: idha Sīhaḷa-saṃvaṇṇanāyaṃ, Visuddhi-maggassa Sīhaḷa-saṃvaṇṇa

nāyañ ca vutto; then later: Visuddhi-magga-sīhaḷa-saṃvaṇṇanāyaṃ pana; and on the next
page: Ācariya-Dhammapāla-ttherena pana tad eva Sīhaḷa-saṃvaṇṇanā-vacanaṃ anic-
chantena.
12 Vism-sn II 798 resembles Ap-a 530; II 798 = Vjb 415; III 1107 = Abhidh-av 444. Of course
there are numerous quotations and mentions of canonical and other aṭṭhakathā works,
not to mention frequent reference to Vism-mhṭ. But ? I 246; 497; II 798; III 1044; IV 1561.
13 Brahmanical: Vism-sn II 757 = Manu-smṛti IX 138; Vism-sn I 2 = Ślokavartika I 12; Vism-sn
III 1053 = Subhv 407. Grammatical: Vism-sn I 28 = Kāśikā ad Pāṇ VI 3 (in Pāli: Nidd-a I 264;
Ap-a 102; cf. Vism 210); Vism-sn II 522 = Vākya-padīya v.1. To these could be added Vism-sn
III 1089 (in Pāli: Vism-mhṭ II 118; Abhidh-av-ṭ I 251) which seems likely to be from a brah-
manical source and a grammatical stanza attributed to Rāhulapāda.
14 Vism-sn I 34 = Saundarananda XIV 36; Vism-sn II 495 = Vyākhyāvyukti; Vism-sn III 1053
(Diṅnāga); Vism-sn III 1043 = Abhidh-k I 21; Vism-sn III 1187 = Abhidh-k II 5. To these
could be added another verse at Vism-sn III 1043 which resembles material at Abhidh-k-
bh I 23; Vism-sn II 495 has a verse about the shape of the four continents; see Appendix
One.
176 Cousins
tions which seem to derive from the Theravādin Abhidhamma tradition.

Indeed they must do, unless some of the South Indian Buddhist schools pos-
sessed an Abhidharma tradition much closer to the Theravādin than available
evidence has so far suggested. Those attributed to Jotipāla must be specifically
Mahāvihāravāsin, but others could in theory be from the Abhayagiri or Jetavana
traditions, since these probably shared much with the Mahāvihāra. The above
count includes twelve stanzas in the āryā meter.15
Most of the remaining forty-five stanzas which I take to be mainly from a
Theravādin tradition are unattributed. I assume that many, if not most, of
these would in fact be the work of Jotipāla. But there are also three attributed
sources. One of these is the Āryasatyāvatāra to which I shall return, but the
others are a stanza attributed to Rāhulapāda16 and a pair of stanzas attributed
to Nāgabodhithera.17 They are quoted in a Pāli version by Vism-mhṭ and their
content expanded in detail.18 So Nāgabodhi must precede Vism-mhṭ in date.
The use of the title sthavira (in Sinhalese) makes it likely that he is Theravādin
– it is not usually, if ever, applied to non-Theravādin monks. Presumably the
name refers to the nāga tree as the bodhi tree of various of the twenty-eight
Buddhas (cf. Bv-a 161 etc.)
6 The Jñeya-saptati-ṭīkā
In his last paper on Buddhist Sanskrit literature in Ceylon, Heinz Bechert refers
to the quotations from Sanskrit works found in Vism-sn as follows:
In this context one might mention the quotation from an otherwise

unknown text, the Jñeya-saptati-ṭīkā (or Jñeya-sampati-ṭīkā), which evi-
dently belonged to the Abhidharma tradition of the Sarvāstivādins or the
15 Vism-sn II 479; III 1119 (×3); 1136 (×2); 1190 (×3); 1250; 1344; IV 1688. The first of these is the
grammatical stanza attributed to Rāhulapāda. Assuming that pāda is honorific, his name
would be Rāhula. See Cousins, 2012: 91ff.
16 See previous note.
17 Vism-sn III 1075. Nāgabodhi’s verses reject the notion that the water element is fluid-
ity (dravatā). cf. Mahāvyutpatti (cited LVP) p. 101; Śikṣāsamuccaya p. ; Buddha-c 9.56
& 13.57: apāṃ dravatvaṃ. Abhidh-k I 13: sneho ’bdhātuḥ; Abhidh-dīp p. 12: abdhātunā
dravasnehalakṣaṇena; Abhidh-sam 3: abdhātuḥ katamaḥ niṣyandatā.
18 Vuttañ h’ etaṃ Purātanehi: Vism-mhṭ II 109 = Abhidh-s-mhṭ 151, etc.
Mūlasarvāstivādins,19 as well as the quotation from the Āryasatyāvatāra

and certain others from Abhidharma texts.20
Bechert refers to “the quotation from […] the Jñeya-saptati-ṭīkā,” although in

fact he gives references to three passages. I shall look at each in turn.
The first reference is the following one: Vism-sn III 1069 (to Vism 450) refers
successively to JS-ṭ, then to the Bhāṣya and the Bhāṣya-ṭīkā, which I under-
stand as follows:21
Kārikā:
Ojas tu yāpana-hetu snehaḥ praśyâdi-vastunām22 (10)
nutritive essence on the other hand is a liquid which is the reason why
such things as edibles and so on nourish’, (10)
JS-bh:
iti tu-śabdaḥ kṛtyenābdhātos tasya viśeṣa-darśanârthaṃ prayuktaḥ; na hi
tasya yāpanaḥ kṛtyam ābandhana-kṛtyatvāt.
the expression ‘on the other hand’ is employed in order to show that it [i.e.
nutriment] has a different function to the water element; for nourishing
is not the function of the water element, since that has binding together
as its function.23
JS-ṭ:
Praśyādi-vastunām: praśyam annam. Ādi-śabdena pānaṃ parigṛhyate
ubhayenāpi yat kabalaṃ kṛtvâb[h]yavaharyyate,24 bhūta-bhautika-
saṅkhyātaṃ vastu tat pradarśayati. Yasmāt tat sarvaṃ rūpotpādanena
19 I have a slight problem with this terminology, as I would regard Mūlasarvāstivāda as a

vinaya term. For me the Mūlasarvāstivādavinaya is a relatively late recension of the
Vinaya produced in Mahāyāna circles (It is still an open question whether or not it was
based upon some earlier distinct Vinaya tradition.) The corresponding term for Abhi
dharma would be Sautrāntika or Dārṣṭāntika, i.e. the critique of Abhidharma produced in
Mahāyāna circles. Note that Vasubandhu’s Abhidharmakośa appears to cite the old canon
in its Mūlasarvāstivāda recension.
20 Bechert, 1998: 5–6. Earlier papers are: Bechert, 1976; Bechert, 1977.
21 Here they are slightly rearranged. See Appendix One for this material set out in order as
in Vism-sn.
22 ‘ŭ’ is m.c.
23 Pj II 54: yaṃ dvādasasu ābandhanakiccakaraṃ snehagataṃ, sā āpodhātu.
24 cf. Abhidh-k-bh I 30; III 39.
178 Cousins
sendriyaṃ kāyam anugṛhṇāti, tasmāt tasya sārañ ca dṛśyamāna-sneho

yāpana-hetutvād oja ity uktaḥ.
Such things as edibles and so on: an edible is food. Drinks are included
under the heading of ‘and so on’. By means of both he shows that that
which is eaten after making it into lumps is accounted among those
things (vastu) classed as elements and derivative elements. Because all of
that supports the body with its sensory faculties by arousing rūpa, there-
fore the liquid which appears as its sāra is called ‘nutritive essence’
because it is the reason why they nourish.
Charles E. Godakumbura (1955: 44) writes:
The work called the Jneya-sampatti-ṭīkā (sic) which has been cited sev-
eral times, as revealed from the quotations, appears to be a text-book of
this school [i.e. the Vijñānavādins]. A bhāṣya and a bhāṣya-ṭīkā of this
work are also cited.
However, JS(-ṭ) is plainly a work of the Theravādin Abhidhamma tradition and

not Vijñānavādin in any direct sense. Probably Godakumbura has been misled
by the literary format of kārikā with bhāṣya and a bhāṣya-ṭīkā. This is of course
characteristic of the Yogācārin school, but it is equally typical inter alia of
Sanskrit Abhidharma traditions which are not Mahāyānist. Compare, for
example, the Abhidharmadīpa which uses most of the kārikā of Vasubandhu’s
Abhidharmakośa but comments on them from a Vaibhāṣika standpoint. It is
also possible that the occurrence of the terms bhāṣya and bhāṣya-ṭīkā is a
scribal or editing error in Vism-sn. The text cited reads more like a single com-
mentary. It may be that there was no bhāṣya – the Āryasatyāvatāra could have
been regarded as the equivalent of an aṭṭhakathā and its commentary styled a
ṭīkā. This would be comparable to the case of the Mahāvaṃsa-ṭīkā and to the
ṭīkās of such works as the Abhidhammāvatāra and Sacca-saṅkhepa.
For these passages, we can compare the description of this type of material-
ity in the Pāli sources:
Dhs-a 330 ≠ Vism-mhṭ II 104 = Abhidh-av-ṭ II 140f. = Abhidh-s-mhṭ 152:

Kabaḷiṃ-kārâhāra-niddese kabaḷaṃ karīyatī ti kabaḷiṃ-kāro, āharīyatī ti
āhāro; kabaḷaṃ25 katvā, ajjhoharīyatī ti attho.
25 Dhs-a (Ee): kabaḷiṃ.

Dhs-mṭ 154 = Vism-mhṭ II 103:

Ojā-lakkhaṇo ti ettha aṅga-m-aṅgânusārino rasassa sāro upatthambha-
bala-kāro bhūta-nissito eko viseso ojā ti.
Has the characteristic of nutritive essence: the sāra of the rasa which flows
throughout the body is a specific nutritive essence dependent on the ele-
ments which gives support and strength.
Vism-mhṭ II 103 adds:

Idaṃ pana savatthukaṃ ojaṃ dassetuṃ vuttaṃ. bāhiraṃ āhāraṃ
paccayaṃ labhitvā eva ajjhattikāhāro rūpaṃ uppādeti. So pana rūpaṃ
āharatī ti āhāro. tenāha: rūpāharaṇa-raso ti. tato eva oj’-aṭṭhamaka-rūp’-
uppādanena imassa kāyassa upatthambhana-paccupaṭṭhāno. ojāya
rūpāharaṇa-kiccaṃ bāhirādhīnan ti āha: āharitabba-vatthu-pada-
ṭṭhāno ti.
But this was said to show nutritive essence together with its material
basis. Only after obtaining external food as its condition does internal
food arouse materiality. But it is called food (āhāra) because it nourishes
(āharati) materiality. That is why he said: it has the function of nourishing
materiality. After that [he said]: it has the manifestation of supporting this
body because it arouses [kalāpas of] materiality with nutriment as its
eighth component. Because nutritive essence’s function of nourishing
materiality is dependent on external [nutriment] he said: its footing is a
material basis to be nourished.
Abhidh-s-mhṭ 152:
Idañ ca savatthukaṃ katvā āhāraṃ dassetuṃ vuttaṃ. S’endriya-
kāy’opatthambhana-hetu-bhūtā pana aṅgamaṅgānusārī
rasa-hara-saṅkhātā ajjhoharitabbâhāra-sineha-bhūtā ojā idha āhāra-
rūpaṃ nāma. Tathā h’etaṃ
s’endriya-kāy’opatthambhana-hetu-bhāva-lakkhaṇaṃ, oj’-aṭṭhamaka-
rūpâharaṇa-lakkhaṇaṃ vā.
Ranjan Wijeratne and Rupert Gethin (2002: 222) translate:

And this, having been mentioned in connection with its basis, is stated to
indicate food. Nutriment which is the viscous liquid of edible food under-
stood as the essence of the fluid that flows through the various parts [of
the body] acting as the cause which sustains the body and its faculties, is
here called materiality of food. Thus it has the characteristic of being the
180 Cousins
cause of sustaining the body with its faculties, or of nourishing [the clus-
ters of] materiality which have nutriment as their eighth [constituent].26
6.1 The Aggregates and nirvāṇa

At Vism-sn 1042f. (to Vism 443) the following is attributed to JS-ṭ:
Vibhāgârthatāyān tu nirvāṇasya skandhântaratva-prasaṃgaḥ syāt? Yasya

câtītâdi-vibhāgo ’sti tad-apekṣatve nâsti prasaṃgaḥ.
This comes just after a śloka and a partial śloka on the subject of the relation
between nirvāṇa and the aggregates. The first of these is attributed to a work
named as the Āryasatyāvatāra ‘Entrance to the Noble Truths’.
Vism-sn III 1042 (attr. to the Āryasatyāvatāra):

Asti bhedo hi rūpâder atītâdi-vasena saḥ
rāśy-arthenaiva saṃkṣipya skandhatvenâbhidhīyate. (11)
There is division of materiality and the other [four] into past, [present
and future]. When [past, present and future] have been grouped together,
each is designated as being an aggregate in the specific sense of being a
grouping (rāśi). (11)
This is part of discussions on the nature of the aggregates found both in the
Pāli commentarial sources and in the Abhidharmakośa (see nn. 31 & 32). The
latter cites an authoritative source which states that the meaning of aggregate
is ‘heap’ (rāśi). This must be a citation from a recension of the Peṭaka, since
exactly this statement is found there.27 It also justifies this with reference to a
passage in a sūtra which states that the rūpa aggregate is so-named after group-
ing together (ekadhyaṃ abhisaṃkṣipya) all rūpa – past, present and future,
26 Abhidh-av-ṭ II 140f. (to Abhidh-av 675): Idañ ca savatthukaṃ katvā āhāraṃ dassetuṃ
vuttaṃ. Kabaḷī-kārâhārassa pana ojā idha āhāro nāma. Tenāha: yāya ojāyā ti-ādi. So ca
rūpāharaṇa-sabhāvo upatthambhana-bala-karo aṅga-m-aṅgânusārī rasâharaṇa-bhūto
bhūta-nissito eko viseso. Bāhirāhāra-paccayaṃ paṭilabhitvā eva ajjhattikāhāro rūpaṃ
uppādetī ti ayaṃ ajjhattikāhārassa upanissaya-bhāvenâpi rūpaṃ āharatī ti āha: “oj’-
aṭṭhamakaṃ rūpaṃ āharatī” ti. Ojā aṭṭhamī yassa, taṃ oj’-aṭṭhamakaṃ. Yāya ojāyā ti
attano udayânantaraṃ rūpa-jananato ojā-saṅkhātāya odana-kummāsâdi-vatthu-gatāya
yāya pharaṇa-ojāya. Yattha yatthā ti yamhi yamhi jana-pade, nagarādīsu ca kabaḷīkāro
āhāro ti pavuccati; tab-batthukattā ti adhippāyo.
27 Peṭ 112: rāsattho khandhattho. The Peṭaka was known in the north. See Zacchetti, 2002a,
Zacchetti, 2002b.
within or without, gross or subtle and so on. Most of this passage is found as
part of the stock description of the aggregate of materiality in many Pāli
sources, but the specific reference to ‘grouping together’ is found only in the
Vibhaṅga.28 Nothing in this stanza indicates whether we are dealing with a
Sanskrit śloka based on the Kośa or something which is rendering into Sanskrit
ideas from the Pāli commentaries. However, a few lines later we have another
Sanskrit stanza, probably also from the Āryasatyāvatāra:
Artha-yogāt tu nirvāṇaṃ dravya-sattve ’pi noditam

skandheṣu tan niruktārtha-yogād … (12)
iti rāśy-artha-yogāt; adravyatvāc ca artha-yogāc ceti samuccayati
Appended to this is a piece of commentary. This suggests that the name of the
work in kārikās may in fact be Āryasatyāvatāra (ĀS). Jñeya-saptati (JS) could be
an alternative name with the Jñeya-saptati-bhāṣya (JS-bh) being its ‘auto-com-
mentary’. Or, the name Jñeya-saptati-ṭīkā might strictly apply only to the
sub-commentary. There could well have been confusion as to this, as manu-
scripts of JS-ṭ would no doubt contain the kārikā together with the commentary.
As to the exact name of the work, it is sometimes given as Jñeya-sampatti in
editions of Vism-sn, but the more probable ‘saptati’ seems to be confirmed by
the form in Sacc-ṭ: Aññeyya-sattati- (see below). The form aññeyya- ‘to be
understood’ may be correct, although unexpected as a rendering of jñeya-.
Perhaps the technical uses of the Pāli word ñeyya- were felt to constrict the
intended meaning in some way.
Rearranged and translated we have:

Artha-yogāt tu nirvāṇaṃ dravya-sattve ’pi noditam
skandheṣu tan niruktārtha-yogād … (12)
iti rāśy-artha-yogāt; adravyatvāc ca artha-yogāc ceti samuccaye ’ti29 …

Vibhāgârthatāyān tu nirvāṇasya skandhântaratva-prasaṃgaḥ syāt? Yasya
câtītâdi-vibhāgo ’sti tad-apekṣatve nâsti prasaṃgaḥ.
kārikā = Āryasatyāvatāra
Yet even although it is a real entity, nirvāṇa is not spoken of [as one of the
aggregates] because the meaning is [not] applicable [and] because the
28 Vibh 1: tadekajjhaṃ abhisaññūhitvā abhisaṅkhipitvā.

29 Text: samuccayati.
182 Cousins
meaning which has been mentioned is applicable in the case of the

aggregates. (12)
Commentary = Jñeya-saptati-ṭīkā
i.e. because the meaning of grouping is applicable; both because it is not
a real entity and because the meaning is [not] applicable. ‘And’ is in the
conjunctive sense.
But if aggregate has the meaning of separate constituent, would not the
(unwanted) possibility then arise of nirvāṇa being another aggregate?
There is no such possibility because the aggregates are connected with
that which is temporally divided.
We can compare two stanzas on this topic in the Sacca-saṅkhepa:
Sacc 5–6:
Param’-attho sa-nibbāna-pañca-kkhandh’ ettha rāsito
khandh’-attho ca samāsetvā, vutto ’tītâdi-bhedanaṃ.
Vedanâdīsu p’ ekasmiṃ, khandha-saddo tu ruḷhiyā
samuddâd’-eka-dese tu, samuddâdi-ravo yathā.30
The supreme meaning is nibbāna and the five aggregates. But here such
divisions as past, [present and future] have been combined and the
meaning of ‘aggregate’ has been given as ‘heap’. (5)
Even when there is only a single item as in the case of [the aggregates of]
feeling and so on, the word ‘aggregate’ [has the meaning of ‘heap’] by
convention, just as in such cases as the sound of the sea referring to one
part of the sea. (6)
Vism-mhṭ II 85 has here:

Khandhā ti paricchinna-dhamma-nidassanaṃ. Yasmā c’ettha khandha-
saddo rās’-aṭṭho ‘mahā-udaka-kkhandho’ ti-ādīsu (SN V 400 etc.) viya,
tasmā atītâdi-vibhāga-bhinnaṃ sabbaṃ rūpaṃ rāsi-vasena buddhiyā
ekajjhaṃ gahetvā, rūpam eva khandho rūpa-kkhandho ti
samānâdhikaraṇa-samāso daṭṭhabbo. ‘Tīhi khandhehi iṇaṃ dassāmā’
ti-ādīsu viya koṭṭhās’-aṭṭhe pana khandha-sadde nibbānassâpi
30 cp. Vibh-a 2: svāyam idha rāsito adhippeto. ayañ hi khandhaṭṭho nāma piṇḍaṭṭho pūgaṭṭho
ghaṭaṭṭho rāsaṭṭho. tasmā ‘rāsilakkhaṇā khandhā’ ti veditabbā. koṭṭhāsaṭṭho ti pi vattuṃ
vaṭṭati; lokasmiñ hi iṇaṃ gahetvā codiyamānā ‘dvīhi khandhehi dassāma, tīhi khandhehi
dassāmā’ ti vadanti. iti ‘koṭṭhāsalakkhaṇā khandhā’ ti pi vattuṃ vaṭṭati. evam ettha
rūpakkhandho ti rūparāsi rūpakoṭṭhāso.
khandh’-antara-bhāvo āpajjatī ti? Nâpajjati, atītâdi-vibhāgâbhāvato; na

hi ekassa niccassa sato nibbānassa atītâdi-vibhāgo atthī ti. Paṭhamen’
atthena rūpa-rāsī ti attho, dutiyena rūpa-koṭṭhāso ti.
Aggregates indicates the dhamma to be defined. But because in this case

the word ‘aggregate’ has the meaning of ‘heap’ as in such cases as ‘great
mass (khandha) of water’, therefore, deliberately taking all materiality
which is divided up as past, [present and future] together as a heap,
aggregate of materiality means aggregate which is materiality, i.e. it is a
compound with the first member in apposition.
[Objection:] But if the word ‘aggregate’ has the meaning of ‘part’ as in
such cases as ‘I will repay the loan in three parts (khandha),’ does it not
entail the inclusion even of nibbāna within the aggregates?
[Reply:] It does not follow because there is no division into past, [present
and future]; for nibbāna which is single and constant has no division into
past, [present and future].
In the first sense the meaning is ‘heap of materiality;’ in the second [the
meaning is] ‘constituent of matter’.
Both these senses are allowed in the Pāli commentaries. The Abhidharmakośa
is familiar with this second sense but rejects it:
Alternatively the meaning is ‘piece;’ for there are those who say ‘we will
give what is to be given in three parts’. That is outside of the Sūtra; for the
Sūtra mentions only the meaning of ‘heap’ with the words: whatever
materiality past, present and future and so on.31
The underlying problem here is that if an aggregate is understood to be a

grouping,32 then it is difficult to speak of a single dhamma as an aggregate; so
a single dhamma can only be an aggregate in the sense that it is a part (prac-
cheda/koṭṭhāsa) of a larger grouping, i.e. in effect a separate entity. But nibbāna
31 Abhidh-k-bh p. 13: pracchedārtho vā; tathā hi vaktāro bhavanti tribhiḥ skandhakair deyaṃ
dāsyāma iti. tad etad utsūtram. sūtraṃ hi rāśyartham eva bravīti “yat kiṃcid rūpam
atītānāgatapratyutpannam” iti vistaraḥ.
32 The notion of khandha as equivalent to rāsi is found in Pāli before Buddhaghosa. See Peṭ
113: Tattha katamo khandh’-attho: samūh’-attho khandh’-attho, puñj’-attho khandh’-attho,
rās’-attho khandh’-attho, taṃ yathā dabba-kkhandho vana-kkhandho dāru-kkhandho aggi-
kkhandho udaka-kkhandho vāyu-kkhandho. Iti evaṃ khandhesu sabba-saṅgaho va evaṃ
khandh’-attho. In fact, it is probably implied by the treatment of the khandhas in the
Yamaka.
184 Cousins
too is a separate entity; so it could be classed as a sixth aggregate (very much

against the tenor of the tradition). This possibility is ruled out by pointing out
that even a single such dhamma is temporally divided and hence, viewed over
a period of time, a kind of grouping. Since nibbāna is the one dhamma which
is not temporally divided, this excludes it from the category of aggregate.
This line of argument could easily be originally of Vaibhāṣika origin, but as
the passage from Vism-mhṭ shows, it is readily acceptable within Theravāda
thought. At all events, it could have been accepted by Theravādins without dif-
ficulty at an early date as well as later on.
6.2 Visual Perception

There remains one further passage in Vism-sn which is attributed explicitly to
the JS-ṭ. This is the one mentioned by Heinz Bechert above and cited in a pub-
lication of Junkichi Imanishi (from information provided by Bechert).33 For
completeness, I give it again here.

Cakṣum pratītya rūpāṇi cotpadyate34 cakṣur-vijñānam iti vacanād
ubhayâśritatve kim arthaṃ cakṣur-vijñānam iti vyapadeśaḥ kriyate na
rūpa-vijñānam ity? Āha: yavâṃkuravad asādhāraṇa-kāraṇena vyapadeśa
iti; cakṣur hi cakṣur-vijñānasyâsādhāraṇa-kāraṇam. Rūpâyatanaṃ mano-
vijñānasyâpi kāraṇam iti sādhāraṇam.
Given that eye discrimination depends on both [the eye and visual
forms], (because of the canonical statement: ‘eye discrimination arises,
conditioned by eye and visual forms’), why is it designated ‘eye discrimi-
nation’ and not ‘form discrimination’?
He replied: what is called designation is made on the basis of a cause
which is not shared, as in the case of a shoot of barley; for the eye is the
unshared cause of eye discrimination. The form base is a shared [cause]
because it is the cause of mind discrimination also.
Closely related to this is another passage from earlier in Vism-sn, also cited by
Junkichi Imanishi:35
33 Imanishi, 1975: 25, note 48.

34 Imanishi has rūpāṇi cotpadyante for Vism-sn rūpāni cotpadyate. The Pāli source is: cak-
khuñ ca paṭicca rūpe ca uppajjati cakkhu-viññāṇaṃ, tiṇṇaṃ saṅgati phasso at M I 111; III
281–286; S II 72–74; IV 32; 67; 86–90 and elsewhere.
35 Imanishi, 1975: 28, note 55.
Vism-sn I 59 (to Vism 20):

Yasya cakṣuḥ paśyatîti darśanaṃ, tasya na sarvaṃ cakṣuḥ paśyati. Kiṃ
tarhi? Vijñānâdhiṣṭhitaṃ. Yasya tu vijñānaṃ paśya[tī]ti darśanaṃ, tasyâpi
na sarvaṃ vijñānaṃ paśyati. Kiṃ tarhi? Cakṣur-āśritam. Tac ca vyavahite
notpadyate yatrâlokasya kuḍyādau pratibandhaḥ. Yatra tv apratibandhaḥ
sphaṭikâbhra-paṭalâdau tatra vyavahite ’py utpadyata eva. Tasmāt tad-
anutpannād vyavahite na cakṣuḥ paśyati vyavahitântaḥ. Tad uktam
cakṣur-vijñānaṃ paśyatîti. Darśana-kṛtyaṃ hi cakṣur-vijñānam iti.
For one whose opinion is that the eye sees, it is not every eye that sees.
How so? It is just [eye that is] directed by consciousness. But for one
whose opinion is that consciousness sees, it is not every consciousness
that sees. How so? It is just [consciousness that is] supported by the eye.
And that consciousness does not arise when there is something inter-
posed where there is obstruction of light in such cases as a wall. But
where there is no obstruction [of light] it does arise even when there is
something interposed as in such cases as crystal or a thin layer of cloud.
Therefore it is because consciousness has not arisen that when there is
something interposed the eye does not see, being restricted by what is
interposed.36 Therefore it is said that eye consciousness sees. For eye con-
sciousness has the function of seeing.
Both these passages concern the controversy over vision.37 That seems to have
originally been a debate with the Mahāsāṃghikas who apparently held that it
is the sensitive matter in the physical eye that actually sees.38 Buddhaghosa
and other commentators cite a stanza from the old commentaries:
Vism 20: (the topic is indriya-saṃvara-sīla) ≠ Dhs-a 400f.: (indriyesu

agutta-dvāratā-niddese) = Nidd-a II 346; 390; III 141 = Paṭis-a III 676:
Cakkhunā rūpaṃ disvā ti kāraṇa-vasena cakkhū ti laddha-vohārena
rūpa-dassana-samatthena cakkhu-viññāṇena rūpaṃ disvā. Porāṇā
panâhu:
[cakkhu] rūpaṃ na passati acittakattā,
cittaṃ na passati acakkhukattā;
36 Ce1 reads: vyavahito vy utpādyata eva tasmā tad-anutpannād avyavahitantena cakṣur na

paśyati vicikṣitāntaḥ yad uktam cakṣūr-vvijñānaṃ.
37 See Dhammajoti, 1997; Pieris, 1997.
38 Kv-a 177.
186 Cousins
dvārârammaṇa-saṅghaṭṭ[an]e pana
[cakkhu]-pasāda-vatthukena cittena passati.39
Īdisī pan’ esā ‘dhanunā vijjhatī’ ti-ādīsu viya sasambhāra-kathā nāma hoti.
Tasmā cakkhu-viññāṇena rūpaṃ disvā ti ayam ev’ ettha attho” ti.
Seeing a visible object with the eye: seeing a visible object with eye
consciousness that is capable of seeing visible objects and which is
referred to as ‘eye’ because [the eye] is the cause [of vision]. But the
Porāṇas declared:
Matter does not see40 because it is not part of the mind.
Mind does not see because it is not part of the eye.
But when the object impacts on the [eye] door,
one sees by means of mind whose base is the sensory matter [of the
eye].
But that is what is known as speaking in reference to everything con-
nected41 as in such cases as when one says ‘he wounds with a bow’.
Therefore the only meaning in this case is: ‘seeing a visible object with
eye consciousness’.
The comments in the Jñeya-saptati-ṭīkā are closely paralleled in Dhammapāla’s

Mahāṭīkā to the Visuddhimagga:
Because [the eye] is the cause [of vision]: because it is a particular cause.
For the result is indicated by a particular cause as with a shoot of corn
[or] the sound of a drum. Or, this refers to what is supported by a word for
a support, as [in the expression] ‘the seats make a noise’. Seeing a visible
object with the eye: as to this if the eye could see a visible object, those
experiencing other consciousnesses could see [it]. But that is not the
case. Why? Because the eye is not conscious. Therefore [the Porāṇas]
said: The eye does not see a visible object because it does not have mind. If
consciousness could see a visible object, it could see it even in such con-
cealed places as the other side of a wall because there is no sensory
39 Pādas a & b are tuṭṭhubha; pādas c & d are jagatī. Cakkhu in pādas a & d is intrusive (prob-
ably from a commentary). It is omitted in the second case in some editions. In pāda c read
°saṅghaṭṭane pana with Nidd-a III 141 and Paṭis-a III 676. There is resolution of the first
syllable. In pāda d emend to °vatthukacittena.
40 Or, one does not see a visible object.
41 cf. Kv-a 178: sasambhārakathānayena vuttaṃ. yathā hi usunā vijjhanto pi ‘dhanunā vijjhatī’
ti vuccati, evaṃ cakkhuviññāṇena passanto pi ‘cakkhunā passatī’ ti vutto.
resistance. This is not the case because all consciousness lacks [the abil-
ity to] see. Therefore [the Porāṇas] said: The mind does not see because it
is not an eye. As to this, just as it is consciousness supported by the eye
which sees, not any consciousness at all (and it does not arise when any
such thing as a wall is in the way where there is an obstruction to light,
but where there is no obstruction in such cases [of transparency] as the
interior of a membrane42 or a crystal, it does arise even when there is
something in the way), in the same way it is eye supported by conscious-
ness which sees, not any eye. Hence the words: Seeing a visible object with
the eye were said with reference to the eye supported by consciousness.43
This establishes that the issue was alive in the writings of the early ṭīkā writers
in a form very similar to that we find in JS-ṭ. But it is even closer in Sumaṅgala’s
twelfth century ṭīkā to the Abhidhammāvatāra and there it is associated with
the name of Jotipāla.
7 Jotipāla and the Aññeyya-sattati-ṭīkā
Abhidh-av-ṭ II 120f. (to Abhidh-av 656):

Kiñcâpi cakkhu rūpaṃ na passati, kiñ carahi tan-nissitaṃ viññāṇam eva.
tathā hi “mañcā ukkuṭṭhiṃ karontī” ti-ādīsu viya nissita-kiriyaṃ nissaye
42 Or, read °abbha-paṭala- ‘thin layer of cloud’.

43 Vism-mhṭ I 47: Kāraṇa-vasenā ti asādhāraṇa-kāraṇassa vasena. asādhāraṇa-kāraṇa-
vasena hi phalaṃ apadisīyati, yathā yav’-aṅkuro bheri-saddo ti. nissaya-vohārena vā etaṃ
nissita-vacanaṃ, yathā mañcā ukkuṭṭhiṃ karontī ti. Rūpan ti rūpāyatanaṃ. Cakkhunā
rūpaṃ disvā ti ettha yadi cakkhu rūpaṃ passeyya, añña-viññāṇa-samaṅgino pi passeyyuṃ.
Na c’etaṃ atthi. Kasmā? Acetanattā cakkhussa. Tenâha: cakkhu rūpaṃ na passati
acittakattā ti. Atha viññāṇaṃ rūpaṃ passeyya, tiro-kuṭṭâdi-gatam pi naṃ passeyya
appaṭigha-bhāvato. Idam pi n’atthi sabbassa viññāṇassa dassanâbhāvato. Tenâha: cittaṃ
na passati acakkhukattā ti. Tattha yathā cakkhu-sannissitaṃ viññāṇaṃ passati, na yaṃ
kiñci (tañ ca kenaci kuṭṭâdinā antarite na uppajjati, yattha ālokassa vibandho, yattha pana
na vibandho phalika-gabbha-paṭalâdike, tattha antarite pi uppajjate va), evaṃ
viññāṇâdhiṭṭhitaṃ cakkhu passati, na yaṃ kiñcī ti viññāṇâdhiṭṭhitaṃ cakkhuṃ sandhāy’
etaṃ vuttaṃ: cakkhunā rūpaṃ disvā ti. cf. Vism-mhṭ II 176 ≠ Sp-ṭ III 220: Cakkhussa
viññāṇan ti cakkhussa kāraṇa-bhūtassa viññāṇaṃ. kāmaṃ rūpâloka-manasikārâdayo pi
tassa viññāṇassa kāraṇā, te pana sādhāraṇa-kāraṇaṃ, cakkhu asādhāraṇan ti
asādhāraṇa-kāraṇenâyaṃ niddeso yathā bheri-saddo, yav’-aṅkuro ti. tathā hi cakkhu
puggal’-antarâsādhāraṇaṃ, nīlâdi-sabba-rūpa-sādhāraṇañ cā ti sāmi-bhāvena niddiṭṭhaṃ
(From kāmaṃ on is cited in Vism-sn from Sp-ṭ immediately before the quotation from
JS-ṭ).
188 Cousins
viya katvā vohāra-sambhavato “cakkhu-pasādena passatī” ti vuttaṃ.

Ācariya-Jotipāla-ttherenâpi hi iminā va adhippāyena idaṃ vuttaṃ. Idam
ettha sanniṭṭhānaṃ: kiṃ cakkhu rūpaṃ passati, udāhu viññāṇan ti.
Kiñc’ettha – yadi cakkhu passeyya, añña-viññāṇa-samaṅgino pi cakkhu
passeyya. atha viññāṇaṃ, kuṭṭâdi-antaritam pi passeyya, tassa
appaṭighātattā? Nâyaṃ doso. Yassa ‘cakkhu passatī’ ti mataṃ, tassâpi na
sabbaṃ cakkhu passati, atha kho viññāṇâdhiṭṭhitam eva. Yassa pana
viññāṇaṃ passatī ti mataṃ, tassâpi na sabbaṃ viññāṇaṃ passati, atha
kho cakkhu-nissitam eva, tañ ca antarite n’uppajjati yattha ālokassa
kuṭṭâdīhi vinibandho. Yattha pana so n’atthi phalika-abbha-paṭalâdimhi,
tattha antarite pi uppajjati eva. Tasmā taṃ anuppannattā antaritaṃ na
passati. Idam eva ca sanniṭṭhānaṃ, yaṃ kho cakkhu-viññāṇaṃ passatī ti;
tañ hi dassana-kiccaṃ. Yadi evaṃ, kathaṃ cakkhunā rūpaṃ disvā ti. tena
dvārena karaṇa-bhūtenā ti ayam ettha abhisandhi. atha vā nissita-kiriyā
nissayassa pavuccati yathā “mañcā ukkuṭṭhiṃ karontī” ti.
Although the eye does not see a visible object, why in that case is it just
consciousness supported by the [eye]. Accordingly it is said that ‘he sees
by means of the sensitive matter of the eye’ because the expression arises
taking the activity of what is supported as like the supports, as in such
[expressions] as ‘the seats make a noise’. For this was said also by the
Elder Ācariya Jotipāla with just this specific meaning. The following is the
agreed answer to the question: does the eye see a visible object or is it
consciousness [which sees a visible object]?
What is more, as to this [it may be objected]: if the eye were to see, the
eye of a person experiencing another kind of consciousness would see.
And consciousness would see even what is concealed by a wall or other
obstruction because there is no sensory resistance. This is not a valid
objection. For someone who holds that the eye sees, it is not every eye
that sees, but only an eye that is directed by consciousness. But for some-
one who holds that consciousness sees, it is not every consciousness that
sees, but only a consciousness that is supported by the eye and that does
not arise in relation to something concealed in a place where there is
obstruction of the light by such things as walls. But in a place where there
is no such obstruction as in such cases as crystal or a thin layer of cloud,
it does arise even in relation to something concealed. Therefore [con-
sciousness] does not see what is concealed because it has not arisen. The
following is the agreed answer to the question: that is to say, eye con-
sciousness sees; for that has the function of seeing.
[Objection] If so, how come that it is stated [in the texts]: ‘seeing a visual
object with the eye’?
[Reply] The meaning in these cases is ‘by means of the [eye] door which
is the cause’. Or rather, the activity of what is supported is applied to the
supports, as in such [expressions] as ‘the seats make a noise’.44
Sumaṅgala is here surely referring explicitly to a work of Jotipāla which is

either identical to or closely resembles the Sanskrit citations from the JS-ṭ.
As Heinz Bechert indicates, the Jñeya-saptati-ṭīkā appears at first sight to be
otherwise unknown in published works. However, I find the following quota-
tion in a transcript of a ṭīkā to Sacca-saṅkhepa (verse 149) presumably made for
Helmer Smith:45
Sahetū46 pi hotī ti sahetukaṃ bhav’-aṅgaṃ ahetukassa bhav’-aṅgassa

anantara-paccayena paccayo ti (Dukap 45) vacanato sahetuka-vipākaṃ pi
tad-ārammaṇaṃ hoti. Ettha ca sahetuka-gahaṇena du-hetukam eva
gahetabban ti keci. Aññeyya-sattati-ṭīkāyaṃ pana: sahetukan ti avisesâ-
padesato dvi-hetukaṃ ti-hetukañ ca gahetabban ti vuttaṃ; tathā hi
Aṭṭha-kathāyaṃ: ahetukassâpi ti-hetukaṃ tad-ālambanam abhimataṃ.
Yam pana Attha-samāse ahetukānaṃ ti-hetuka-vipākāni na hontī ti
vuttaṃ, tam kathaṃ? So yeva pucchitabbo yo tassa kattā ti vatvā, taṃ
paṭikkhittaṃ.
[By a kamma other than the one which gives paṭisandhi] there are caused
[resultants] too, i.e. the tadārammaṇa is caused resultant because of the
words (in the Paṭṭhāna): ‘caused bhav’-aṅga is a condition for causeless
bhav’-aṅga by succession condition’. But some say that in this context
‘caused’ should be taken as referring to two-caused [resultant] only.
However, in the Aññeyya-sattati-ṭīkā it is said that, since ‘caused’ does not
specify [either two-caused or three-caused], it should be taken as both
44 Vibh-anuṭ 163: kilesuppattinimittatāya uppattirahaṃ kilesaṃ ārammaṇaṃ antogad-

hakilesan ti vuttaṃ. tañ ca kho gāhake labbhamānaṃ gahetabbe upacaritvā, yathā nissite
labbhamānaṃ nissaye upacaritvā ‘mañcā ukkuṭṭhiṃ karontī’ ti. idāni upacāraṃ muñcitvā
nippariyāyen’ eva atthaṃ dassento “nimittaggāha … pe … sadisā” ti āha.
45 Transcripts of two ṭīkās to Sacc are among the papers, etc. of Helmer Smith preserved in
Uppsala. My thanks are due to the library of the University of Uppsala which kindly pro-
vided photocopies for me some years ago.
46 Ms: sahetu.
190 Cousins
two-caused and three-caused; for in the Aṭṭhakathā47 a three-caused

tadālambana is allowed in this way even for a causeless [being]. [An
objection was offered:] ‘But how does this fit with the statement in the
Atthasamāsa that causeless [beings] do not have three-caused resul-
tants?’ The objection was rejected with the reply: ‘ask the author of the
Atthasamāsa himself’.
This can be compared with a view attributed to Jotipāla, which is cited in

Sumaṅgala’s Abhidhammāvatāra-ṭīkā (Abhidh-av-ṭ I 258):
Ācariya-Jotipāla-tthero pana: sahetukan ti avisesena vuttattā ahetukānam

pi ti-hetuka-tad-ārammaṇaṃ icchati. Vuttañ hi tena: sa-hetukan ti
avises’opadesena du-hetukaṃ, ti-hetukañ ca gahetabbaṃ; tathā hi
Aṭṭhakathāyaṃ ahetukassâpi ti-hetuka-tad-ārammaṇam abhihitaṃ. Yañ
carahi Attha-samāse ahetukānaṃ ti-hetuka-phalāni detī ti vuttaṃ, taṃ
kathaṃ? So eva pucchitabbo, yo tassa kattā ti. Apare pana: mūla-sandhiyā
jaḷattā tassa ti-hetuka-tad-ālambaṇaṃ na labbhati yevā ti vadanti.
(cp. Pm-vn 271: ñāṇa-pākā na vattanti, jaḷattā mūla-sandhiyā; cf. Dhs-a
416; Vibh-a 15)
But the Elder Ācariya Jotipāla teaches a three-caused tadārammaṇa even

for causeless [beings] because ‘caused’ is given [in Paṭṭhāna] without
specifying. For he said the following: since ‘caused’ does not specify
[either two-caused or three-caused], it should be taken as two-caused
and three-caused; for in the Aṭṭhakathā a three-caused tadārammaṇa is
mentioned in this way even for a causeless [being]. [An objection was
offered:] ‘How now does this fit with the statement in the Atthasamāsa
that it does not give three-caused fruits to causeless [beings]?’48 [The
47 The notion that a caused tad-ārammaṇa is possible for a being with a causeless recon-
necting by a kamma other [than the reconnecting] is found at Dhs-a 288, but it does not
seem that the fifth or sixth century commentaries ever specify whether caused here
means ‘two-caused’ or both ‘two-caused and three-caused’. Probably Jotipāla intends to
argue that if the author of the Abhidhamma Commentary had intended to make it
restricted, he would have specified ‘two-caused’. Since he did not so specify, his intent was
to allow ‘three-caused’ also. Abhidh-av 206f. and 215f. allow only 37 cittas for those reborn
in the four apāya – a figure which excludes both kinds of caused resultant. This would still
leave open the case of a causeless rebirth as a human being and in any case Abhidh-av-ṭ
takes the figure as excluding ‘other kamma’.
48 The Burmese edition must be in error here, by omitting a na or something similar. Com-
pare Abhidh-s-mhṭ 121 where Sumaṅgala gives a briefer version of the same debate: idha
objection was rejected:] ‘ask the author of the Atthasamāsa himself’. But
others say that the [causeless being] does not get a three-caused
tadālambaṇa because his fundamental connecting (i.e. paṭisandhicitta)
lacks intelligence.
This is probably a specifically Theravādin Abhidhamma debate, as it is linked

to the theory of the cittavīthi. In effect, the canonical Paṭṭhāna does not specify
beyond indicating that the sequence from caused to causeless bhav’-aṅga is
legitimate. Since this refers to what the later terminology calls the succession
from tadārammaṇa to the (mūla-) bhav’-aṅga, it establishes for the fifth cen-
tury Abhidhamma Commentary that a being whose bhav’-aṅga is causeless
does sometimes have a caused tadārammaṇa, i.e. one of the eight mahāvipākas.
This is explained as being the result of a kamma other than the one which led
to that particular rebirth. Some writers wished to limit this to the four
mahāvipākas without knowledge, but Jotipāla rejected that view.
Sumaṅgala is commenting here on Abhidh-av 443 which simply denies the
possibility of even a two-caused tadārammaṇa for a causeless being, i.e. one
who is reborn in an apāya or a human being who is incomplete in some major
way from conception. After citing Paṭṭhāna, he refers to teachers who say:
“causeless [beings] have caused [resultants] by means of other kamma (hoti
aññena kammena, sahetukaṃ ahetunan)” and permit caused for the causeless,
similarly three-caused [resultants] for two-caused [beings]”. The stanza he
cites is Sacc 149, which is the verse being commented on in the passage from
Sacc-ṭ cited above. The author of Sacc-ṭ simply accepts the authority of Jotipāla
to decide the question.
Not so Sumaṅgala. He refers to the view of others who reject the possibility
of a three-caused tadārammaṇa in this case because the fundamental (re)con-
necting mind is stupid. The argument would seem to be that, since the rest
state (bhav’-aṅga) to which the mind continually reverts throughout life is dull
and stupid, it will not support even temporary rest states (tadārammaṇa) with
wisdom. The ‘others’ in question must include Anuruddha, the author of
Param’-atthavinicchaya (Pm-vn), since the wording at Pm-vn 271 is very close.
That it is Anuruddha to whom Sumaṅgala refers is clear from his treatment of
this issue in his mahāṭīkā to the Abhidhamm’-atthasaṅgaha (Abhidh-s-mhṭ
121).
For Sumaṅgala, Anuruddha is also the author of the Abhidhamm’-
atthasaṅgaha (Abhidh-s). So the discussion here starts from the position of
ñāṇa-sampayutta-vipākâbhāva-vacanassa parihāsa-vasena, so eva pucchitabbo, yo tassa

kattā ti vuttaṃ.
192 Cousins
Abhidh-s which explicitly denies resultants with knowledge for two-caused

and causeless beings in a fortunate destiny. Indeed, it goes further and denies
that any caused resultants occur to beings in an apāya (Abhidh-s IV 41f.).
Sumaṅgala commences his comments by acknowledging that the Paṭṭhāna
source text allows the possibility of a two-caused tadārammaṇa due to miscel-
laneous kamma. He then points out that Anuruddha rejects this explicitly and
gives as the reason for this that the fundamental (re)connecting mind is stu-
pid. Thus far his argument expands the sanne of Sumaṅgala’s teacher Sāriputta
(Abhidh-s-sn 124). He now adds further material, drawn from Abhidh-av-ṭ
(cited above):49
But the Elder Ācariya Jotipāla said that there is a three-caused tadā
rammaṇa even for causeless [beings] because ‘caused bhav’-aṅga’ is
given [in Paṭṭhāna] without specifying. Then he said the following: ‘ask
the author of the [Atthasamāsa] himself,’ as a humorous rejection of the
claim that there are no resultants joined with knowledge in the case [of
causeless beings].
Sumaṅgala goes on to point out that, although this was said humorously, in
fact the right thing to do is to consult an (or the) ācariya i.e. Anuruddha, the
author of both Abhidh-s and Pm-vn and quotes from Pm-vn by name.50 He
then adds that others comment: “just as there is a caused tadārammaṇa for
causeless beings, similarly there is a three-caused tadārammaṇa for two-
caused beings. And in compliance with their understanding, people say that
the rejection of resultants joined with knowledge at this point [in Abhidh-s]
applies only to [the case of] the causeless”.
He then comments that, since there is no authoritative text for this, their
words should be accepted after investigation, as the ācariya has classified the
cittas in exactly the same way (samakam eva) after giving a reason common to
both [causeless and two-caused beings] for the absence of resultants joined
49 As Oskar von Hinüber points out (Hinüber 1996, §346), Abhidh-s-mhṭ was “finished
within the astonishingly short time of 24 days”. This is no doubt best accounted for by
supposing that Sumaṅgala is translating his teacher’s sanne into Pāli and adding material
from an already written Abhidh-av-ṭ.
50 Taṃ pana parihāsa-vasena vuttam pi ācariyaṃ pucchitvā va vijānan’-atthaṃ vutta-
vacanaṃ viya ṭhitaṃ. Tathā hi ācariyen’ev’ ettha kāraṇaṃ Param’-attha-vinicchaye
vuttaṃ: ñāṇa-pākā na vattanti, jaḷattā mūla-sandhiyā ti (Pm-vn 271).
with knowledge.51 Sumaṅgala refers again to Jotipāla’s view later in

Abhidh-av-ṭ.52
The two passages in Abhidh-av-ṭ and Sacc-ṭ, the one cited from Jotipāla, the
other from the Aññeyya-sattati-ṭīkā, are very close in meaning but not identical
in wording. This is just the result we might expect if they are independently
rendered into Pāli from a Sanskrit original.
The verses from Vism-sn make it clear that Jotipāla did indeed write in
Sanskrit. There seems then no reason to doubt that Jotipāla is indeed the
author of the Jñeya-saptati-ṭīkā, given that essentially the same passage (frag-
ment one) was attributed by Sumaṅgala to Jotipāla and by the author of Sacc-ṭ
to the Jñeya-saptati-ṭīkā. Noticeably, when Vism-sn refers to Jotipāla by name it
is invariably citing the kārikā; so this raises the question as to whether in fact
Jotipāla was the author of both commentaries. There would be nothing sur-
prising about an author after the sixth century writing both, but there seems
no way of being sure of this. However, it seems very unlikely that a
Mahāvihāravāsin verse text would have been written in Sanskrit without at
least one auto-commentary – the distinctive terminology of the Pāli tradition
would have needed explanation.
7.1 The Nature of the Five Aggregates

A few lines later in Vism-sn (III 1043) we have (unattributed):
Bhājanaṃ bhojanañ caiva vyañjanaṃ kartṛ-bhoktṛṇī

pañca dīpayituṃ skandhāḥ kramenânena deśitāḥ. (13)
The five aggregates were taught in this order to reveal: the vessel, the
food, the spice, the maker [of the meal] and its consumer. (13)
(cp. Abhidh-k-bh I 22: bhājanabhojanavyañjanakarttṛbhoktṛbhūtā hi
rūpādayaḥ skandhāḥ; Mahāvibhāṣā).
This comes closely after the passage attributed to the Āryasatyāvatāra and
seems likely to be part of an account of the first noble truth in that work. The
51 Tattha pana pamāṇa-pāṭhâbhāvato ācariyena ubhinnam pi sādhāraṇa-vasena ñāṇa-

sampayutta-vipākâbhāve kāraṇaṃ vatvā samakam eva citta-paricchedassa dassitattā
tesaṃ vacanaṃ vīmaṃsitvā sampaṭicchitabbaṃ.
52 Abhidh-av-ṭ I 355: Ahetukānaṃ paṭisandhi-sadisa-tad-ārammaṇa-vasena ‘satta-tiṃs’evā’ ti
vuttaṃ, añña-kammena pana dvi-hetuka-tad-ārammaṇassâpi sambhavato eka-cattālīsa
honti. Ācariya-Jotipāla-ttherassa adhippāyena ti-hetuka-vipākehi pi saddhiṃ pañca-
cattālīs’evā ti daṭṭhabbaṃ.
194 Cousins
actual simile involved is found in both Sarvāstivādin and Theriya sources. But
if the stanza is indeed part of a work of Jotipāla, its immediate source might be
expected to be the commentaries of the school of Buddhaghosa. But there we
have:
Vibh-a 32 = Vism 479 = Moh 123:

Api ca cāraka-kāraṇa-aparādha-kāraṇa-kāraka-aparādhikâpamā ete
bhājana-bhojana-byañjana-parivesaka-bhuñjakûpamā cā ti, evaṃ
upamāto vinicchaya-nayo viññātabbo.
Bhikkhu Ñāṇamoli (1956) translates:

Also they are (respectively) like the prison, the punishment, the offence,
the punisher, and the offender. And they are like the dish, the food, the
curry sauce [poured over the food], the server, and the eater.
This is actually closer to the version in the Abhidharmakośa.
Nāmar-p 655 renders this into Pāli:

Bhājanaṃ bhojanaṃ tassa, byañjanaṃ bhojako tathā.
bhuñjitā cā ti pañc’ete, upamenti yathā-kkamaṃ.
Vism-mhṭ II 85f. has:

Kasmā pan’ete khandhā pañc’eva vuttā iminā eva ca kamenā ti? Bhājana-
bhojana-byañjana-bhatta-kāraka-bhuñjaka-vikappa-dassanato
yath’oḷārika-yathā-saṃkilesûpadesato cā ti veditabbaṃ. Vivāda-mūla-
hetu-bhāvaṃ saṃsāra-hetutaṃ, kamma-hetutañ ca cintetvā,
vedanā-saññā saṅkhāra-kkhandhato nīharitvā, visuṃ-khandha-bhāvena
desitā. (cf. AK I 21 cited Vism-sn III 1043)
But why are the aggregates spoken of just as five and just in this order? It
should be understood that that it is to show them variously as the vessel,
the food, the spice, the cook and the one who eats, and to expound the
way in which they are gross and subject to defilement. Feeling and label-
ing are taught as separate aggregates, separating them from the aggregate
of constructions, considering them as being the cause of the bases of dis-
pute, the cause of saṃsāra and the cause of kamma [respectively].
Abhidh-s-mhṭ 175 explains and expands:

Te pan’ete khandhā bhājana-bhojana-byañjana-bhatta-kāraka-bhuñjaka-
vikappa-vasena pañc’eva vuttā ti āha: rūpa-kkhandho ty-ādi rūpañ hi
vedanā-nissayattā bhājana-ṭṭhāniyaṃ, vedanā bhuñjitabbattā bhojana-

ṭṭhāniyā, saññā vedanâssāda-lābha-hetuttā byañjana-ṭṭhāniyā, saṅkhārā
abhisaṅkharaṇato bhatta-kāraka-ṭṭhāniyā, viññāṇaṃ upabhuñjakattā
bhuñjaka-ṭṭhāniyaṃ. ettāvatā ca adhippet’-attha-siddhī ti pañc’eva vuttā.
Desanā-kkame pi idam eva kāraṇaṃ yattha bhuñjati, yañ ca bhuñjati,
yena ca bhuñjati, yo ca bhojako, yo ca bhuñjitā, tesaṃ anukkamena
dassetu-kāmattā.
Ranjan Wijeratne and Rupert Gethin (2002) translate:

These aggregates, considered as the vessel, the food, the seasoning, the
cook, and the one who eats, are said to be just five; hence he states the
words beginning ‘the aggregate of materiality’.
The body, being the support of feeling, represents the vessel; feeling,
being the thing enjoyed, represents the food; recognition, being the
means by which one savors the feeling, represents the seasoning; forma-
tions, as putting things together, represents the cook; consciousness,
being the one who enjoys [it all], represents the one who eats. With this,
the intended meaning is established; hence, precisely five [aggregates]
are stated. This [explanation] is also the reason for their being taught in
this order since [it reveals] the wish to indicate in due order which of
them it is that constitutes the place where one consumes, what one con-
sumes, what one consumes [it] with, who provides what is consumed,
and who does the consuming.
If we tabulate the different versions, they seem fairly close and it does not
seem possible to identify a source. The concept seems likely to predate both
Buddhaghosa and the Kośa; so the Sanskrit stanza may be quite old. Or, it could
be a version produced by Jotipāla or elsewhere.
Vism-sn Abhidh-k Vism etc. Vism-mhṭ Nāmar-p
Materiality bhājana bhājana bhājana bhājana bhājana

Feeling bhojana bhojana bhojana bhojana bhojana
Labelling vyañjana vyañjana byañjana byañjana byañjana
Constructing kartṛ kartṛ parivesaka bhattakāraka bhojaka
Consciousness bhoktṛ bhoktṛ bhuñjaka bhuñjaka bhuñjita(r)
196 Cousins
7.2 Another Passage from the Āryasatyāvatāra?

Abhidh-s-sn cites a stanza in sragdharā meter attributed by Hammalawa
Saddhātissa (Abhidh-s-mhṭ 110 and Upās 106; 109) to the Āryasatyāvatāra:53
Abhidh-s-sn 106:
Suptaś cūtasya mūle phala-patana-ravāt kaścid unmīlya cakṣur
dṛṣṭvā câdāya caitat phalam atha paritā pīḍayitvā kareṇa
jñatvā pakva-sva-bhāvaṃ kila paribubhuje tad rasâkṛṣṭa-cetāḥ
paścāt tālv-ādi-lagnaṃ rasam api ca saha-śleṣmaṇâsvādya śiśye.
Rupert Gethin (Abhidh-s-mhṭ p. 128n.) translates:

A man had gone to sleep at the foot of a mango tree; at the sound of a
falling fruit he opened his eyes; he saw the fruit, picked it up, pressed it all
round with his hand, and realized it was ripe; attracted by the thought of
the taste he ate it, and having enjoyed the final flavor that lingered in his
mouth with the saliva, he went to sleep. 54
This is a version of the mango simile, used to explain the consciousness pro-
cess in Pāli sources. It is quite close to the version given in the Abhidhamma
Commentary (Dhs-a 271) and even closer to the versions given in the Pāli ṭīkās.
It differs even from the version in the Chinese translation of the Vimuttimagga.
Nothing like it is known from any non-Theravādin source and if the indication
of the Vimuttimagga is anything to go by, it probably differs from any version
current in the Abhayagiri school.
So whether or not Saddhātissa had any source for the attribution to the
Āryasatyāvatāra, it seems certain that these two stanzas are of Mahāvihāravāsin
origin and they might well be the work of Jotipāla.
7.3 Other Citations of Possible Theravādin Origin

There remain thirty-eight stanzas. For this chapter, I will ignore some of those
which could originate from any Buddhist source and focus on those which
seem likely to have come from the Theriya tradition or something very
similar.
53 I do not know whether he had a source for this or merely inferred it.
54 Gethin reads paritaḥ.
7.3.1 Stanzas Concerning the Four Truths

There are a number of stanzas concerning the four truths, which could easily
be from the Āryasatyāvatāra but could equally come from other Buddhist lit-
erature. To begin with, we have three stanzas in the āryā meter providing
nirukti of the syllables of the words duḥkha, samudaya and nirodha.55 These
closely follow passages in the Visuddhimagga and are very likely to be Jotipāla’s
Sanskrit version of that, but it is not possible to rule out other Sanskrit litera-
ture or even a source earlier than Buddhaghosa.

duś-śabdaṃ kutsāyām kha-śabdam etañ ca tuccha-paryāyam.
sarvaṃ yojya caturṣûvāca duḥkhaṃ prathamaṃ satyam. (14)
The syllable dus has the sense of despicable, while the syllable kha is a
synonym for the empty. Combining both together, he spoke of the truth
of duḥkha first among the four.56 (14)
Vism-sn III 1190:

saṃyogotpatty arthau sam-udāv aya-iti ca kāraṇasyākhyā
trayam abhisamīkṣya Bhagavān samudaya iti jālinīm ākhyāt. (15)
The Lord spoke of the net [of craving] as samudaya, referring to three
things: sam and ud have the two senses of conjunction (saṃyoga) and
arising (utpatti), while aya is a term for cause.57 (15)
Vism-sn 1190:
Nirabhāvo nistṛṇavad bhavati ca rodha iti cārakasyākhyā
tāv abhisamīkṣya tādṛn nirodha ity āha nirvāṇam. (16)
Nir means absence as in ‘without grass’ (nistṛṇa) and confinement

(rodha) is a term for imprisonment (in saṃsāra). Referring to these two,
55 Kahrs, 1998.
56 Vism 494: Idha du iti ayaṃ saddo kucchite dissati, kucchitaṃ hi puttaṃ dupputto ti vadanti,
khaṃ-saddo pana tucche, tucchaṃ hi āsaṃ khan ti vuccati. Idañ ca paṭhama-saccaṃ
kucchitaṃ aneka-upaddavâdi-ṭṭhānato tucchaṃ bāla-jana-parikappita-dhuva subha-
sukh’-atta-bhāva virahitato, tasmā kucchitattā tucchattā ca dukkhan ti vuccati.
57 Vism 495: Saṃ iti ca ayaṃ saddo samāgamo sametan ti ādisu saṃyogaṃ dīpeti, u iti ayaṃ
saddo uppannaṃ uditan ti ādisu uppattiṃ, aya-saddo kāraṇaṃ dīpeti, idāñ câpi dutiya-
saccaṃ. Avasesa-paccaya-samāyoge sati dukkhass’ uppatti-kāraṇaṃ, iti dukkhassa
saṃyoge uppatti-kāraṇattā dukkha-samudayan ti vuccati.
198 Cousins
the holy one spoke of nirvāṇa as absence of confinement (nirodha).58

(16)
Two stanzas give the reason for the selection of craving as the cause of suffer-
ing. Buddhaghosa simply states that it is the chief cause (mūlakāraṇa), but the
ṭīkās elaborate exactly as in these Sanskrit stanzas:59
Vism-sn III 1213f. (to Vism 506):

Ajñānâdiṣu duḥkhasya santesv anyeṣu hetuṣu,
tṛṣṇā samudayas Sūtre proktâyaṃ mukhya-kāraṇāt. (17)
Karma-vaicitrya-hetutvāt sāhāyya-gamanāc ca yat
kurute duḥkha-vaicitryaṃ tṛṣṇā, tan mukhya-kāraṇam.60 (18)
Given that there are other causes of suffering such as ignorance

this craving has been proclaimed as the arising [of suffering] in Sūtra
because it is the chief cause. (17)
Craving which brings about many kinds of suffering is the chief cause
because it is the cause of many kinds of karma and because it follows as
the companion [of karma]. (18)
We can compare the closely parallel statement in the ṭīkā to the Abhidhamm’-
atthasaṅgaha (Abhidh-s-mhṭ 178):
58 Vism 495: Tatiya-saccaṃ pana yasmā ni-saddo abhāvaṃ, rodha-saddo ca cārakaṃ dipeti,
tasmā abhāve ettha saṃsāra-cāraka-saṅkhātassa dukkha-rodhassa sabba-gati-suññattā.
Samadhigate vā tasmiṃ saṃsāra-cāraka-saṅkhātassa dukkha-rodhassa abhāvo hoti tap-
paṭipakkhattā ti pi dukkha-nirodhan ti vuccati.
59 Vism-mhṭ I 229 = Sp-ṭ I 231: Tattha mūla-kāraṇa-bhāvenā ti santesu pi avijjâdīsu aññesu
kāraṇesu tesam pi mūla-bhūta-kāraṇa-bhāvena. taṇhā hi kammassa vicitta-bhāva-hetuto,
sahāya-bhāvûpagamanato ca dukkha-vicittatāya padhāna-kāraṇaṃ. Vism-mhṭ I 204 = Sp-ṭ
III 177: Kasmā pan’ettha taṇhā va samudaya-saccaṃ vuttā ti? Visesa-hetu-bhāvato. Avijjā hi
bhavesu ādīnavaṃ paṭicchādentī, diṭṭhi-ādi-upādānañ ca tattha tattha abhinivisamānaṃ
taṇhaṃ abhivaḍḍhentī dosâdayo pi kammassa kāraṇaṃ honti, taṇhā pana taṃ-taṃ-
bhava-yoni-gati-viññāṇa-ṭṭhiti-sattâvāsa-satta-nikāya-kula-bhog’-issariyâdi-vicittataṃ
abhipatthentī, kamma-vicittatāya upanissayataṃ kammassa ca sahāya-bhāvaṃ upa
gacchantī bhavâdi-vicittataṃ niyameti, tasmā dukkhassa visesahetu-bhāvato aññesu pi
avijjā-upādāna-kammâdīsu Sutte Abhidhamme ca avasesa-kilesâkusala-mūlâdīsu vuttesu
dukkha-hetūsu vijjamānesu taṇhā va samudaya-saccan ti vuttā ti veditabbaṃ.
60 Two lines of this are cited also at Vism-sn II 480 in a different order.
Kasmā pana aññesu pi dukkha-hetūsu santesu taṇhā yeva samudayo ti

vuttā ti? Padhāna-kāraṇattā. Kamma-vicittatā-hetu-bhāvena, hi kamma-
sahāya-bhāvûpagamanena ca dukkha-vicittatā-kāraṇattā taṇhā
dukkhassa visesa-kāraṇan ti.
Ranjan Wijeratne and Rupert Gethin (2002) translate this:
But why, since there are other causes of suffering, is craving alone stated
as the origin of suffering? Because it is the principal cause. For since it is
the reason for a variety of suffering by being the motivation for a variety
of kamma and by following as the companion of kamma, craving espe-
cially is the cause of suffering.
Coming to the third truth, we find three stanzas related to the debates as to
whether nirvāṇa is a really existing entity in some sense or just the mere
absence of greed, hate and delusion. These too are closely related to the discus-
sions of this topic in the Visuddhimagga and the Abhidhammāvatāra, as well
as in the ṭīkās to those works.61 The first of these is:

sākṣāt-kārī-vacas siddhe sattve ’nuttara-gocare
nirvāṇe ’nya-pratikṣepas sūryye jāty-andha-nihnavam. (19)
Since the existence of the supreme object has been established on the
authority of one who has realized it, the rejection of nirvāṇa by others is
[like] the denial by one born blind of the [existence of] the sun. (19)
At this point Buddhaghosa simply states that it cannot be said that what a fool-
ish puthujjana does not perceive, therefore does not exist (Vism 507).
Sumaṅgala, however, does give the same simile.62 Since he is clearly influenced
by Jotipāla, his source is most probably the Satyāvatāra.
61 See the summary of part of those discussions in Collins, 1998: 177ff.

62 Abhidh-av-ṭ II 192f.: Yathā hi ceto-ñāṇa-lābhino eva ariyā paresaṃ lokuttara-cittaṃ jānanti,
tatthâpi ca arahā eva sabbesaṃ cittaṃ jānāti, evaṃ nibbānam pi sīlâdi-sammā-paṭipatti-
bhūtena upāyen’eva upalabbhatī ti “n’atthī” ti na vattabbaṃ ariyehi upalabbhanīyato. yathā
hi jacc-andha-temirikâdīnaṃ adassanena “sūriyâdayo n’atthī” ti na vattabbā cakkhuman
tānaṃ gocara-bhāvena atth’evā ti siddhattā, evam etam pi bāla-puthujjanassa adassana-
mattena “n’atthī” ti na vattabbaṃ. ariyānaṃ pana paññā-cakkhuno padesen’eva
upaṭṭhānato ekantena sampaṭicchitabbaṃ “atthi nibbānan” ti.
200 Cousins
The second stanza is:
Vism-sn III 1219 (to Vism 508)

rāgâdi-kṣaya-hetutvān nirvāṇaṃ hi tadākhyayam
kṣayânta-jas tathârhattvan no cet syād etad apy asat. (20)
For nirvāṇa is referred to as the [destruction of greed, etc. because it is

the cause of the destruction of greed, etc. Similarly arahatship is born of
the ending [of greed, etc.]. If that were not so, arahatship too would be
non-existent. (20)
This second of our Sanskrit stanzas makes two points. The first is that
nirvāṇa is strictly speaking not the destruction of greed, hate and delusion,
but rather the cause of that.63 The second is that the same statement is made
about arahatship; so if that statement implies non-existence, arahatship
too would be non-existent. This second point is already made in the discus-
sion in the Visuddhimagga and the Abhidhamma Commentary.64 In fact, the
Visuddhimagga account is for the most part a rewritten version of the material
in the latter. Indeed, the account in the fourth or fifth century Abhidhamma
Commentary is almost certainly copied from earlier material, giving refuta-
tions of the position of the Vitaṇḍavādin(s). The first point, however, does
not seem to be explicitly made in this context until the account in Kassapa’s
Mohavicchedanī (ca. twelfth century).65
The third stanza is:

rāgâdi-kṣaya-sūtreṇa siddhāntaḥ kṣaya-mātrakam
nirvāṇam iti caitan na tad dhi neyârtham iṣyate. (21)
63 cf. Spk-pṭ II 162: āsavānaṃ khayā ti āsavānaṃ khayahetu ariyamaggena sabbaso

āsavānaṃ khepitattā.
64 Vism 508: “Yo kho, āvuso, rāga-kkhayo” ti ādi-vacanato (SN IV 251) ‘khayo nibbānan’ ti ce. na,
arahattassâpi khaya-mattâpajjanato. tam pi hi “yo kho, āvuso, rāga-kkhayo” ti ādinā (SN IV
252) nayena niddiṭṭhaṃ. More detail in Vibh-a 51f.
65 Moh 133: na pāṭekkaṃ nibbānaṃ nāma atthi; kilesakkhayamattam eva nibbānaṃ. Ten’eva
ca tāsu tāsu Suttābhidhammadesanāsu ‘rāgakkhayo dosakkhayo’ ti ādinā vuttan ti ce? Na,
arahattanibbānānaṃ ekatāpajjanato. Arahattam pi hi “yo kho, āvuso, rāgakkhayo dosak-
khayo mohakkhayo, idaṃ vuccati arahattan” ti (SN IV 252) rāgādīnaṃ parikkhayante
uppannattā upacārena arahattaṃ rāgakkhayādibhāvena vuttaṃ. Tathā rāgādīnaṃ
khayahetuttā nibbānam pi rāgakkhayādibhāvena vuttan ti gahetabbaṃ.
If it is said that the authoritative position in accordance with the sūtra on

the destruction of passion, etc. (SN IV 252f.) is that nirvāṇa is just the
destruction [of greed, hate and delusion], that is not correct; for that
[sūtra] is taught as one requiring elucidation. (21)
This third stanza is responding to the argument that the statement in the
Jambukhādakasutta is an authoritative one, i.e. the sutta is nītārtha – one
whose meaning is explicit. This point does not seem to be spelt out in the com-
mentarial literature, although it is certainly implied by their arguments. It
seems to be first spelt out in the Mūla-ṭīkā.66 It is expanded in the Mahāṭīkā.67
This again seems to fit with an attribution to Jotipāla.
Another stanza that could be part of a work on the truths is given at the
conclusion of the discussion of the aggregate of materiality:

sūtre kṛtaṃ savijñātvaṃ jāteḥ paryāya-bhāṣitam
saṃskṛtānān tu lakṣmatvāt saṃskṛtaṃ triṣu coditam. (22)68
Birth’s having birth given in Sūtra69 was stated metaphorically,

but because constructed things have characteristics, having birth is con-
structed and stated among the three. (22)
This continues a discussion found already in the Abhidhamma Commentary.70

It concerns the question as to how far notions such as ‘birth’ or ‘conditioned
origination’ can apply to ‘birth’ itself or to old-age and death for that matter.71
Since this question must have been long discussed among Ābhidharmikas,
66 suttato muñcitvā ti suttapadāni muñcitvā. añño suttassa attho “mātaraṃ pitaraṃ hantvā”
ti ādīsu (Dhp 294f.) viya āharitabbo, natthi suttapadeh’ eva nīto attho ti attho.
67 Vism-mhṭ II 208: tam pi hi … pe … niddiṭṭhaṃ: tattha yathā rāgādīnaṃ khīṇante
uppannattā arahattaṃ rāgakkhayo ti ādinā niddiṭṭhan ti suttassa neyyatthattā. evaṃ
nibbānaṃ āgamma rāgâdayo khīyantī ti nibbānaṃ rāgakkhayo ti ādinā niddiṭṭhan ti
imassâpi suttassa neyyatthatā veditabbā. cp. Abhidh-av-ṭ II 191.
68 Variant reading: śāstre kṛtaṃ. pāda a: read ajātatvaṃ; pāda c: read lakṣ[y]atvāt.
69 SN II 26: jāti, bhikkhave, aniccā saṅkhatā paṭiccasamuppannā; Paṭis I 51: jāti
paṭiccasamuppannā.
70 Dhs-a 341f.
71 Vism-mhṭ II 112: yadi evaṃ, kathaṃ “jarā-maraṇaṃ paṭicca-samuppannan” ti (SN II 26)
vuttaṃ? yasmā paṭicca-samuppannānaṃ dhammānaṃ paripāka-bhaṅgatāya tesu san-
tesu honti, na asantesu. na hi ajātaṃ paripaccati, bhijjati vā, tasmā taṃ jāti-paccayataṃ
sandhāya “jarā-maraṇaṃ paṭicca-samuppannan” ti pariyāyena suttesu vuttaṃ.
202 Cousins
there is no way to be sure that this stanza comes from Jotipāla’s work. But it is
perhaps more likely than not.
7.3.2 Stanzas Concerning the Consciousness Process

In the next section, I turn to seven stanzas which with the mango simile already
mentioned are indisputable proof of the existence of Sanskrit works on the
consciousness process (cittavīthi). Nothing like this is found in the Sarvāstivādin
Abhidharma literature. In principle, it could have been shared with other
South Indian Buddhist schools but in the absence of specific evidence we have
to assume that it is unique to the Theriya tradition of Ceylon. Indeed, when
concepts related to this are mentioned in later Mahāyāna śāstra literature,
they are specifically attributed to the Ceylon school.72 However, there is no way
of knowing whether they come from Jotipāla’s work, from other Mahāvihāravāsin
texts or from other schools in Ceylon.
I turn first to a stanza which concerns specifically initiation of the process
concerned with direct sensory perception:

yady artha-sannidhiḥ pañca-dvāre prasāda-ghaṭṭanam
katham anyâśritasyeñjā tat baddhāśrayatād bhavet. (23)
If there is striking of the sensitive matter in the five door process,

how is there movement of [the passive mind] which has a different resort
[i.e. the heart-base]? That would be because it is connected. (23)
The corresponding passage in the Mahāṭīkā is closely related:
Vism-mhṭ II 132:
Visaya-visayī-bhāva-siddhāya dhammatāya ārammaṇassa abhimukhī-
bhāvena pasādassa tāva ghaṭṭanā hotu, añña-sannissitassa pana
bhav’-aṅgassa calanaṃ kathaṃ hotī ti? Taṃ-sambandha-bhāvato.
bheri-tale ṭhapitāsu sakkharāsu ekissā sakkharāya ghaṭitāya tad-añña-
sakkharāyaṃ ṭhita-makkhikā calanaṃ c’ettha udāharaṇan ti.
Granted, firstly, that there is impact on the sensitivity owing to confronta-

tion with an object, since the necessity for that is established by the
72 Cousins, 1981.
existence of the objective field and the possessor of the objective field,
but how does there come to be disturbance (movement) of the life-con-
tinuum that has a different support? Because it is connected with it. And
here the example is this: when grains of sugar are put on the surface of a
drum and one of the grains of sugar is tapped, a fly sitting on another
grain of sugar moves.
A set of three stanzas addresses the conclusion of the consciousness process

when the karmically active moments (javana) can be followed by one or two
moments of retention (tadārammaṇa) before relapsing back into the default
state of the passive mind (bhavaṅga). All of this terminology seems to be spe-
cific to the Theriya tradition:
Vism-sn III 1106 (cf. Pāli version Abhidh-s-mhṭ 119 = Abhidh-s-sn 119f.):
javanaḥ tat-samāno vā tad-ālambana-cetasā
javo ’nubaddhyate nânyo bāla-dāraka-līlayā (24)
bījâbhāvāt na cordhveṣu tad-ālambanam asya tu
bījaṃ kāmâpta-pākâkhyaṃ pratisandhi-mano matam (25)
pravarttamāna-javanam ucite sthāna eva tat
anubadhnāti dṛṣṭântam āhur atrâpi dārakam (26)
The active mind that is similar to it is followed by a retention mind

[just as] the progenitor, not anyone else, is followed in the play of a young
boy. (24)
But because of the absence of its seed, retention does not occur in the
upper [realms].
The seed known as appropriate sense-sphere resultant is reckoned as the
reconnection mind. (25)
It follows an active mind that is occurring only in a familiar place.
They say that the simile for this is a boy. (26)
The illustrations given here are specific to the ṭīkā literature. In fact the
Abhidhamm’-atthasaṅgaha Commentary has what looks like a rewritten ver-
sion of these verses in Pāli:
Abhidh-s-mhṭ 119:
janakaṃ taṃsamānaṃ vā javanaṃ anubandhati,
na tu aññaṃ tadālambaṃ bāladārakalīlayā. ||
bījassābhāvato natthi brahmānam pi imassa hi;
paṭisandhimano bījaṃ kāmāvacarasaññitaṃ. ||
204 Cousins
ṭhāne paricite yeva taṃ idaṃ bālako viya

anuyātī ti nāññattha hoti taṇhāvasena vā ti. ||
But these illustrations are already found much earlier in the Mahāṭīkā.73 They
seem unlikely to be significantly earlier than the seventh century in date. Two
stanzas on the same general topic follow shortly after.
Vism-sn III 1106:

paritta-śukla-tṛṇ-moha-mūla-saumya-kriyā-javāt
pañcasv ekaṃ tad-ālambaṃ saumyavat svânulomataḥ. (27)
parittād aśubhāc chuklāt kriyā-cittāc ca nissukhāt
sopekṣeṣu tad-ālambaṃ ṣaṭsv74 ekam anurūpataḥ. (28)
After small active minds which are pure or rooted in greed and delusion
or kriyā with pleasant feeling,
[there arises] one corresponding retention with pleasant feeling from
among the five [possible]. (27)
After small impure, pure or kriyā minds which are without happiness,
[there arises] one appropriate retention from among the six with neutral
feeling. (28)
The account here is not different in essence to the commentarial account, but
the phrasing is quite close to the Mahāṭīkā.75
Two further verses concern the consciousness process at the time of rebirth:

sugatau jāyamānasya prāk cyuter jjavane śubhe
durgatau kṛṣṇam eveti niścinanti sumedhasaḥ. (29)
73 Vism-mhṭ II 133: Kāmâvacara-javanâvasāne ti kāmâvacara-javanass’eva avasāne. na hi

taṃ kāma-taṇhāhetuka-kamma-nibbattaṃ mahaggatânuttara-javanaṃ anubandhati
ajanakattā, janakâsadisattā ca. yathā gehato bahi gantu-kāmo taruṇa-dārako janakaṃ,
janaka-sadisaṃ vā anubandhati, na aññaṃ, evam idam pi.
74 text: ṣaṭṣv.
75 Vism-mhṭ II 133: tatthâpi na sabbasmā javanā sabbaṃ javanena tad-ārammaṇassa niyam-
etabbato, ārammaṇena ca vedanāya parivattetabbato. tatthâyaṃ niyamo: paritta-kusala-
lobha-moha-mūla-somanassa-sahagata-kiriya-javanānaṃ aññatarânantaraṃ atimahati
visaye pañcannaṃ somanassa-saha-gatānaṃ aññataraṃ tad-ārammaṇaṃ uppajjati,
tathā paritta-kusalâkusala-upekkhā-saha-gata-kiriya-javanānaṃ aññatarânantaraṃ
upekkhā-saha-gatānaṃ channaṃ tad-ārammaṇānaṃ aññataraṃ pavattati.
When there is a pure active mind before decease for one being reborn in
a fortunate destiny,
the very wise decide that there is only a dark active mind for one being
born in a bad destiny. (29)

pratikāle satṛṣṇasya svapnavat gati-cihnakam
karmaṃ karma-nimittaṃ vā mānasaṃ khyāti karmanā. (30)
at the time of [death]76 the mind of one with craving is said to have like a
dream,
the mark of the destiny, the karma or the sign of karma due to his karma.
(30)
In this case the ideas found do not go beyond those found in the commentaries
of the school of Buddhaghosa, but they remain specific to the Theriya tradition.
7.3.3 Stanzas Concerning Particular caitasikas

Six stanzas concern particular thought concomitants (cetasika/caitasika/
caitta). Since the thought concomitants which accompany the mind in various
of its states are lists largely derived from the shared canonical literature which
was inherited by all later Buddhist schools, it is not surprising that many of the
same terms are found in all the Abhidha(r)mma lists of thought concomitants
known to us.
First of all, I will mention a stanza concerning shamelessness (ahrī) and
disregard of consequence (anapatrapa):
Abhidh-s-sn 48 = Vism-sn 1130 (to Vism 470):

nâhrīr jugupsate pāpād amedhyād iva śūkaraḥ.
śalabho ’gner ivaitasmān na trasaty anapatrapaḥ. (31)
Abhidh-s-sn: nâhrī; trasas ty.

Shamelessness does not feel disgust at the bad, like a pig at excrement.
Disregard of consequence has no fear of that like a moth with fire. (31)
A Pāli version of this stanza is cited by Sumaṅgala:
76 Possibly emend to cutikāle, but if so it is used loosely. The reference is to the last javana
process before death.
206 Cousins
Abhidh-av-ṭ 315:
yathāhu Porāṇā:
jigucchati nāhiriko, pāpā gūthā va sūkaro.
na bhāyati anottappo, salabho viya pāvakā” ti ||
The comparisons with the pig and the moth are already found in the Mahāṭīkā.77
I have not so far found them in any Sanskrit source, but they clearly could
come from a non-Theriya source. But, given the passage in the Mahāṭīkā, a
Theriya origin again seems the most likely.
Three stanzas in the āryā meter are concerned with faith (śraddhā/saddhā).
They give comparisons of faith with a hand, wealth and seeds. These similes
are quite old with canonical roots and various mentions in the commentarial
literature.78 They are further expanded in the works of Dhammapāla, the ṭīkā
writer. I have not come across them so far in Northern Buddhist literature as
similes applied to śraddhā, but they may well be found there somewhere.79

bhatta-dattam api dravyaṃ yathā ca dṛṣṭavā naiva gṛhṇāti
śraddhāhastād dattaṃ tathaivāryya-dhanaṃ na gṛhṇāti. (32)
Just as one who has vision does not take hold even of what is given as
food as wealth,
so likewise he does not take hold of the ārya wealth given by the hand of
faith. (32)
Vism-sn III 1119:

dhanavaikalyāt kṛpaṇo yathâsamarthaḥ kriyāvidhau bhavati
śraddhā-dhanena hīnas tathâsamartho bhavati dharmme. (33)
77 Vism-mhṭ II 149: Gāma-sūkarassa viya asucito kilesâsucito ajigucchanaṃ ahirikena hoti,

salabhassa viya aggito pāpato anuttāso anottappena hotī ti evaṃ vutta-ppaṭipakkha-
vasena vitthāro veditabbo.
78 Dhs-a 120 = Vism 464: hatthavittabījāni viya daṭṭhabbā; cf. Mp IV 57; Pj II 144.
79 Vism-mhṭ II 143: Kusala-dhammānaṃ ādāne hatthaṃ viya, sabba-sampatti-sampadāne
vittaṃ viya, amata-kasi-phala-phalane bījaṃ viya daṭṭhabbā; Dhs-anuṭ 96: “saddhāhattho,
Mahānāma, ariyasāvako” ti (cf. Dhs-anuṭ 96; older sources cite: AN III 346 = Th 694), “sad-
dhîdha vittaṃ purisassa seṭṭhan” ti (SN I 42; 214; Sn 182), “saddhā bījaṃ tapo vuṭṭhī” ti (SN I
172; Sn 77) evamādivacanato kusaladhammānaṃ ādānâdīsu hatthâdayo viya saddhā
daṭṭhabbā; cf. Abhidh-av-ṭ I 299.
Just as someone wretched through insufficient wealth is unable to carry

out affairs (kriyāvidhi), so someone who lacks the wealth of faith is not
capable of dharma. (33)
Vism-sn III 1119:

bījān asato na bhavati yathâṅkuraṃ kṣiti-tale ’pi sunikṛṣṭe
śraddhā-bījān asato na tu bhavati tathā kuśala-mūlam. (34)
Just as one who eats the seeds will have no shoot even in a well-ploughed
area of land,
so one who consumes the seeds of faith will have no skilful root. (34)
Two stanzas in the āryā meter concern jealousy (īrṣyā/issā) and miserliness
(mātsarya/macchariya). The first is:

īrṣyā-vahni-śikhâgrād yeṣām para-sampad-indhanaṃ jvalayati
prāmodyasya rasaṃ te †na vindanti† brahma-bhogyasya. (35)?
Those for whom jealousy’s flame ignites the fuel of other people’s gains,
do not find the elixir of the gladness which is to be enjoyed by?brahmas.
(35)
Sumaṅgala gives a Pāli version of this:
Abhidh-av-ṭ I 319:
ten’ev’ etaṃ vuccati:
issânala-sikhā yesaṃ hadaye jalatîdha te
n’eva vindanti pāmojjaṃ sambuddhâdīhi sevitan ti.
Those for whom the flame of the fire of jealousy burns here in the heart,
do not find the gladness practiced by the Fully Awakened Ones, etc.
Vism-sn III 1136:

tyāgodbhavayā prītyā svapne ’pi mano hi sādhyate naiṣām
yeṣāṃ cetasi labhate mātsaryya-kaliḥ †pada-nyāsam†. (36)
The mind accomplishes nothing even in dream with the joy that springs
from relinquishing for those in whose mind the unlucky throw of miserli-
ness is obtained. (36)
208 Cousins
7.3.4 Stanzas concerning particular cittas

Four stanzas are related specifically to the classification and definition of types
of citta. Although not entirely absent, this is not a characteristic feature of the
Northern Abhidharma literature. So it is not surprising that these stanzas con-
cern issues known to us mainly from Pāli literature. Again, although it is not
possible to rule out completely sources from other Southern Buddhist schools
whose literature does not survive, these stanzas seem almost certainly to be of
Theriya origin.
I give first a stanza relating to the eight mahākusala cittas. This concerns a
specific issue. Skilful citta is explained as occurring with objects of sense or
mind that are intrinsically attractive, referred to as wished for (iṣṭa/iṭṭha)
objects. These can obviously be objects of greedy states of unskilful conscious-
ness. So the question is what brings about the shift to skilfulness.

yoniso viṣayâbhogāt parimityāc chubhasya ca
naiyyāmāt pariṇāmāc ca śubham iṣṭeṣu vastuṣu. (37)80
The beautiful [occurs] with wished for things [as its object] due to appro-
priate directing towards the object and due to [familiarization] and due
to restricting the mind to the skilful and changing it [to that]. (37)
This stanza summarizes a passage already found in the Abhidhamma Com

mentary. I translate:
But surely this wished for object is the basis for greed? How has this citta
which is called ‘skilful’ arisen? [The answer is:] due to restricting (niyam-
ita), due to changing (pariṇāmita), due to familiarization (samudācāra),
and due to directing (ābhujita). Someone’s mind is restricted to the per-
formance of the skilful with the idea ‘I should do only what is skilful,’ after
turning from the occurrence of the unskilful it is changed into doing just
the skilful, it makes just the skilful familiar by doing it again and again
and appropriate directing [of the mind] occurs due the prior conditions
(upanissaya) of dwelling in a suitable place, attendance upon good peo-
ple, hearing the saddhamma, having merit performed previously and so
80 pāda b.?parimityāc; pāda c. em. to nairyāmāt.

on. Because of this that which is called ‘skilful’ has arisen to him by means
of restricting, changing, performing and directing.81
The passage is slightly rewritten in the Mahāṭīkā with ‘directing’ in first place
as in our Sanskrit verse.82 It is again slightly rewritten by Sumaṅgala, returning
to the order of the Abhidhamma Commentary. But he then adds a Pāli version
of our Sanskrit verse:
Abhidh-av-ṭ I 182f.:
Hoti c’ettha:
niyāmapariṇāmehi, samudāciṇṇatāya ca
ñāṇapubbaṅgamābhogā iṭṭhe pi kusalaṃ siyā” ti ||
By means of restricting [the mind to the skillful] and changing it [from

the unskillful] and because [the skillful] is made familiar, there can be
skilful even on a wished for [object] as a result of directing [the mind]
which is preceded by knowledge.
It is striking that the version we have here in Sanskrit is closest to that in the
Mahāṭīkā.
Two stanzas address the question of precisely how many resultant cittas
arise as the results in another life of the various skilful cittas. Part of the issue
turns precisely on the difference between skilful cittas with three causes (hetu)
and those with two. In the Northern Abhidharma systems wisdom is a univer-
sal accompaniment of all cittas and the idea of skilful states without wisdom is
81 Dhs-a 75: Nanu c’etaṃ iṭṭhārammaṇaṃ lobhassa vatthu? Kathaṃ etaṃ cittaṃ kusalaṃ
nāma jātan ti? Niyamitavasena pariṇāmitavasena samudācāravasena ābhujitavasenā ti.
Yassa hi ‘kusalam eva mayā kattabban’ ti kusalakaraṇe cittaṃ niyamitaṃ hoti, akusalap-
pavattito nivattetvā kusalakaraṇe yeva pariṇāmitaṃ, abhiṇhakaraṇena kusalasamudācāren’
eva samudācaritaṃ, patirūpadesavāsasappurisūpanissayasaddhammasavanapubbekata
puññatādīhi ca upanissayehi yoniso ca ābhogo pavattati, tassa iminā niyamitavasena
pariṇāmitavasena samudācāravasena ābhujitavasena ca kusalaṃ nāma jātaṃ hoti.
82 Vism-mhṭ II 115: nanu ca iṭṭhārammaṇaṃ lobhassa vatthu, kathaṃ tattha kusalaṃ
hotī ti? na-y-idam ekantikaṃ iṭṭhe pi ābhogādivasena kusalassa uppajjanato. yassa hi
catusampatticakkasamāyogâdivasena yoniso va ābhogo hoti, kusalam eva ca mayā kat-
tabban ti kusalakaraṇe cittaṃ niyamitaṃ, akusalappavattito ca nivattetvā kusalakaraṇe
eva pariṇāmitaṃ, abhiṇhakaraṇavasena ca samudācaritaṃ, tassa iṭṭhe pi ārammaṇe
alobhādisampayuttam eva cittaṃ hoti, na lobhādisampayuttaṃ.
210 Cousins
not admitted. So this discussion too can only derive from the Theriya schools.
In fact, we can go further than this.
When commenting on the eight mahākusala cittas at their first occurrence
in the Dhammasaṅgaṇi, the Abhidhamma Commentary offers what it refers to
as the mātikā of the discourse setting out resultant [citta]. This is part of a
series of named topics in a sequence already established in an earlier com-
mentary or commentaries. In the course of this sequence the positions
attributed to three named elders are set out in brief.83 The positions of the first
two are set out in these two Sanskrit verses. The three Elders in question appear
to have lived in Ceylon during the first century BCE or just after.84 Their views
will have been known to the author of the Abhidhamma Commentary from
sources dating from the early centuries CE. These two stanzas also can only
have come from the Theriya tradition:

yat tri-hetukam utkṛṣṭaṃ sandhinā tv aṣṭakāv ubhau
phaled uktau nikṛṣṭan tu dvi-hetūn dhī-yutair vinā. (38)
tadvad dhī-hīnam utkṛṣṭaṃ hīnan tv ahetum aṣṭakam
dvādaśânye daśâṣṭau ca teṣām pākān pracakṣate. (39)
The superior kind which has three causes would result in both the sets of
eight which have been mentioned together with the connexion [citta].
But the inferior kind of citta would result in [cittas] with two causes lack-
ing those with intelligence. (38)
The superior kind of citta lacking knowledge is like that. But the inferior
kind would result in the set of eight without causes. Others declare that
their results are twelve, ten and eight. (39)
This is closely parallel to the account in the Sacca-saṅkhepa:
Sacc 124–126:
Ñāṇa-yutta-varaṃ tattha, datvā sandhiṃ ti-hetukaṃ
pacchā paccati pākānaṃ, pavatte aṭṭhake duve ||
tesu yeva nihīnaṃ tu, datvā sandhiṃ du-hetukaṃ.
deti dvādasa pāke ca, pavatte dhī-yutaṃ vinā ||
evaṃ dhī-hīnam ukkaṭṭhaṃ, sandhiyañ ca pavattiyaṃ
hīnaṃ pan’ ubhayatthâpi, hetu-hīne va paccati ||
83 Dhs-a 267–288; Abhidh-av 378ff.; Paṭis-a III 574 cites Paṭis against both the first two.
84 Mori, 1989: 126f., 129f., 134f.
Among these the excellent citta joined with knowledge, after giving
three-caused connexion, subsequently in process gives two sets of eight
results.85
But a lower kind of citta among those, after giving a two-caused connex-
ion, subsequently in process gives twelve results (without the [four] that
have intelligence).
This distinguishes inferior and superior kinds of these cittas. The supe-
rior kind can give results of its own kind, whereas the inferior kind cannot
give the best results.
The superior [kind of citta] lacking intelligence [gives results] in the
same way both in connexion and in process, but the inferior kind [of
ñāṇavippayuttacitta] gives in both [connection and process] only the
[eight results] without causes.
The use of dhīhīna and dhīyuta as verse equivalents of ñāṇasampayutta and

ñāṇavippayutta in both these sources is very unusual. It never seems to be used
by Anuruddha or Buddhadatta. The date and authorship of Sacca-saṅkhepa is
uncertain, but it must belong to the period between the sixth and tenth cen-
tury. Tradition variously attributes it to Ānanda the author of the Mūla-ṭīkā, to
Dhammapāla and, more recently to a hypothesized Culla-Dhammapāla. In any
case we find that the Mahāṭīkā also gives just the same two Theravādas as our
Sanskrit text; so this also supports a date in this period.86
Finally we have one Sanskrit stanza which concerns the eleventh type of
unskilful citta – that accompanied by agitation (auddhatya/uddhacca). This
85 i.e. eight great resultants and eight skilful resultant causeless cittas.
86 Vism-mhṭ II 118: Ta-y-idaṃ aṭṭha-vidham pi kāmâvacaraṃ kusala-cittaṃ rūpârammaṇaṃ
yāva dhammârammaṇan ti chasu ārammaṇesu yaṃ vā taṃ vā ālambitvā upekkhā-
saha-gatâhetuka-kiriyā-mano-viññāṇa-dhātânantaraṃ kāya-dvārâdīhi tīhi dvārehi
kāya-kammâdi-vasena uppajjatī ti veditabbaṃ. Tattha ñāṇa-sampayuttāni cattāri
yadā ti-hetuka-paṭisandhiṃ uppādenti, tadā soḷasa vipāka-cittāni phalanti. Yadā pana
du-hetukaṃ, tadā dvādasa ti-hetuka-vajjāni. Ahetukaṃ pana paṭisandhiṃ ti-hetukāni
na uppādent’eva; du-hetukāni pana du-hetuka-paṭisandhi-dāna-kāle dvādasa, ahetuka-
paṭisandhiṃ dāna-kāle aṭṭha phalanti. Ti-hetukā pana paṭisandhi duhetukehi na hoti
yeva. ‘Aṭṭha phalantī’ ti ce taṃ paṭisandhiṃ janaka-kamma-vasena vuttaṃ. Aññena pana
kammunā “sa-hetukaṃ bhav’-aṅgaṃ ahetukassa bhav’-aṅgassa anantara-paccayena
paccayo” ti (Paṭṭh III 42) vacanato sa-hetukam pi vipāka-cittaṃ ahetuka-paṭisandhikassa
hoti yeva. Imasmiṃ ca pakkhe balavatā paccayena uppannaṃ asaṅkhāraṃ, dubbalena
sasaṅkhāran ti veditabbaṃ. Ye pana āgamanato ca vipākassa asaṅkhāra-sasaṅkhāra-
bhāvaṃ icchanti, tesaṃ matena dvādasa, aṭṭha ca phalantī ti yojetabbaṃ. Evaṃ tidhā
phalaṃ dadantañ c’etaṃ kāmâvacara-sugatiyaṃ upapattiṃ, sugati-duggatīsu bhoga-sam-
padañ ca karoti.
212 Cousins
too does not seem to be something which could have come from a Northern
Abhidharma tradition:
Vism-sn III 1091 to Vism 454:

varṇṇayanti phalâbhāvam auddhatya-sahitasya tu
yataś cauddhatyam uddhṛtya tatpāke ’rtham itîritaḥ. (40)87
But they explain that there is [no] fruit of the citta which is accompanied
by agitation,
since it is said with reference to agitation that [discrimination of] the
meaning is [knowledge of] its result. (40) (cf. Vibh 297)
In its present form the stanza seems corrupt. The reference is the commen-
tarial discussions concerning the results of this eleventh kind of unskilful citta.
The discussion presupposes the statement in the Dhammasaṅgaṇi (Dhs §1391)
that classifies this citta arising (cittuppāda) as exclusively to be abandoned by
bringing into being (bhāvanā). In other words it can only be removed by the
development of the path after the stage of stream-entry. Most other citta aris-
ings are partially eliminated at this stage and partially eliminated in the
preceding stage of vision (dassana). So this is generally understood as meaning
that agitation consciousness is not able to give result as the connexion citta
which would initiate a future life (in one of the four apāyas).88
It is explicitly understood that it can give results during the course of such a
life. The basis for this is a statement in the Vibhaṅga. At Vibh 297, when clas-
sifying the four discriminations (paṭisambhidā), it is specifically indicated that
the discrimination of attha is knowledge of the result in respect of all twelve
unskilful cittas. It is this that is being referred to in the second line of this
Sanskrit stanza. We must therefore emend the first line to remove the
negation.
87 pāda a: read phalabhāvam?

88 Dhs-a 261: Imehi pana dvādasahi pi akusalacittehi kamme āyūhite, ṭhapetvā uddhaccasa
hagataṃ, sesāni ekādas’ eva paṭisandhiṃ ākaḍḍhanti. Vicikicchāsahagate aladdhādhi
mokkhe dubbale pi paṭisandhiṃ ākaḍḍhamāne uddhaccasahagataṃ laddhādhimokkhaṃ
balavaṃ kasmā nâkaḍḍhatī ti? Dassanena pahātabbâbhāvato (Dhs §1390) Yadi hi
ākaḍḍheyya dassanenapahātabba-padavibhaṅge āgaccheyya, tasmā, ṭhapetvā taṃ, sesāni
ekādasa ākaḍḍhanti. Tesu hi yena kenaci kamme āyūhite tāya cetanāya catūsu apāyesu
paṭisandhi hoti. Akusalavipākesu ahetukamanoviññāṇadhātu upekkhāsahagatāya paṭi
sandhiṃ gaṇhāti. itarassâpi etth’eva paṭisandhidānaṃ bhaveyya. Yasmā pan’ etaṃ n’ atthi,
tasmā dassanenapahātabba-padavibhaṅge nâgatan ti.
It is not specifically indicated in the stanza, but we may suspect that what
is being referred to is a debate in the ṭīkā literature, which refers to a modi-
fied view which allows for the existence in puthujjanas of a type of agitation
consciousness which can give reconnection in a future life.89 They seem to
have understood that it was already abandoned prior to the stage of stream-
entry. According to Sumaṅgala this was the position of Buddhamitta and
others. However, the name Buddhamitta does not seem to have been unusual
and no Pāli text authored by a Buddhamitta appears to be extant. Indeed it
is quite possible that what is being referred to is a position espoused by non-
Mahāvihāravāsin writers.
8 Conclusion
From the Cūlavaṃsa we learn that Jotipāla played a major role in the
Buddhism of Ceylon around the seventh century. From the citations in the
work of Sumaṅgala we know that he was considered a major authority in the
Polonnaruva period and we can deduce that he was the author of an Anuṭīkā
on the Abhidhammapiṭaka. From the citations of verses specifically attributed
to him it is clear that he also wrote in Sanskrit.
89 Vism-mhṭ II 122f.: Apare panâhu: puthuj-janassa uppajjamānaṃ uddhacca-saha-gataṃ

dassana-ppahātabba-sahāya-sabbhāvato ubhaya-vipākam pi deti, na sekkhassa tad-
abhāvato ti. Idam ettha vicāretabbaṃ, yassa vipāka-dānaṃ vuttaṃ, kiṃ taṃ bhāvanāya
pahātabbaṃ, udāhu no ti? Kiñ c’ettha – yadi tāva bhāvanāya pahātabbaṃ, Paṭṭhāne
bhāvanāya pahātabbassa nānā-kkhaṇika-kamma-paccaya-bhāvo vattabbo siyā. Atha na
bhāvanāya pahātabbaṃ, dassanena-pahātabba-ttike “n’eva-dassanena-na-bhāvanāya-
pahātabbam” icc assa vibhaṅge vattabbaṃ siyā. Yadi tab-biruddha-sabhāvatāya tattha na
vucceyya, evam pi tasmiṃ tike tassa na-vattabbatā āpajjatī ti? Nâpajjati. Kiṃ kāraṇaṃ?
Citt’-uppāda-kaṇḍe āgatānaṃ dvādasannaṃ akusala-citt’-uppādānaṃ dvīhi padehi
saṅgahitattā vibhajitvā dassetabbassa niyogato kassaci citt’-uppādassa abhāvā, yathā
uppanna-ttike atītâdīnaṃ navattabbatā na vuttā, evam etassâpi. Atha vā bhāvanāya
pahātuṃ asakkuṇeyyassâpi tassa puthujjane vattamānassa bhāvanāya pahātabba-
sabhāva-sāmaññato, sâvajjato ca bhāvanāya pahātabba-pariyāyo vijjatī ti n’atthi
na-vattabbatā-pasaṅga-doso. Nippariyāyena ca na bhāvanāya pahātabban ti tassa vasena
nānā-kkhaṇika-kamma-paccaya-bhāvo pi na vutto. Dassana-pahātabba-paccayassâpi
uddhacca-saha-gatassa sahāya-vekalla-mattam eva dassanena kataṃ, na koci pi bhāvo
anuppāda-dhammataṃ tassa āpādito ti ekantena bhāvanāya pahātabbatā vuttā. Atha
vā apāya-gamanīya-bhāvâpekkhaṃ dassana-ppahātabba-vacanan ti tad-abhāvato taṃ
vibhajanaṃ vuttan ti. See Dhs-mṭ 122f.; Vibh-mṭ 95; Dhs-anuṭ 129f.; in detail Abhidh-av-ṭ
270ff.
214 Cousins
In this chapter, citations from works known as the Āryasatyāvatāra and as

the Jñeya-saptati-ṭīkā are also examined. These prove to contain material
which is closely paralleled in Pāli. Most probably, they represent a single verse
text and its commentary. Additional citations of material in Sanskrit, mostly
from Vism-sn, are also examined. It is clear that most of them do not come
from a Northern Abhidha(r)mma tradition and many of them are likely to be
from the Āryasatyāvatāra and be the work of Jotipāla. It remains possible that
some may be from one of the other Theriya schools originating in Ceylon.
We should probably assume that a significant use of Sanskrit as a medium
for presenting its understanding of Buddhist teachings was a feature of all the
Theriya schools in Ceylon and South India from at least the seventh century
down to the thirteenth or later. The subsequent vicissitudes of Buddhism on
the island and the disappearance of Buddhism in South India after the four-
teenth century led to the loss of this Sanskrit literature. Probably it could not
be recovered from Southeast Asia where Pāli rather than Sanskrit was more
used for Theriya Buddhist purposes, either having never been taken there or
not having been preserved.
Appendix 1: References to Āryyasatyāvatāra and Jñeya-saptati-ṭīkā
Vism-sn III 1042 (to Vism 443) attr. Āryyasatyāvatāra:

Asti bhedo hi rūpâder atītâdi-vasena saḥ
rāśy-arthenaiva saṃkṣipya skandhatvenâbhidhīyate.90

Artha-yogāt tu nirvāṇaṃ dravya-sattve pi ’noditam.
skandheṣu tan niruktârtha-yogād …
iti rāśy-artha-yogāt adravyatvāc ca artha-yogāc ceti samuccayati91
Vism-sn III 1069 (to Vism 450) attr. Jñeya-saptati-ṭīkā:92

Ojas tu yāpana-hetu snehaḥ praśyâdi-vastunām93 (6)
iti tu-śabdaḥ kṛtyenābdhātos tasya viśeṣa-darśanârthaṃ prayuktaḥ94; na
hi tasya yāpanaḥ kṛtyam ābandhana-kṛtyatvāt. (Bhāṣya)
90 Ce1: skandhenaiva.
91 em. to samuccaya iti and to arthâyog- in the first and third occurrences.
92 Ce1: Jñeya-sampatti-ṭīkā.
93 Ce1: vastuna.
94 Ce1: prayuktā.
Praśyādi-vastunām: praśyam annam. Ādi-śabdena pānaṃ95 parigṛhyate ubhayenāpi

yat kabalaṃ kṛtvâb<h>yavaharyyate,96 bhūta-bhautika-saṅkhyātaṃ vastu tat
pradarśayati. Yasmāt tat sarvaṃ rūpotpādanena sendriyaṃ kāyam anugṛhṇāti, tasmāt
tasya sārañ ca dṛśyamāna-sneho yāpana-hetutvād oja ity uktaḥ. (Bhāṣya-ṭīkā)
Appendix 2: Unidentified Buddhist Material
Vism-sn II 495 (to Vism 207)97:

ardhendv-ādarśa-saṃsthānau98 Deha-Goyāna-nāmakau
Kurus tu pīṭha-saṃsthāno, Jambu-dvīpo rathâkṛtiḥ. (41)
Vism-mhṭ 243:
Jambu-dīpo sakaṭa-saṇṭhāno. Apara-goyānaṃ ādāsa-saṇṭhāno. Pubba-videho
addha-canda-saṇṭhāno. Uttara-kuru pīṭha-saṇṭhāno. Taṃ-taṃ-nivāsīnaṃ,
taṃ-taṃ-parivāra-dīpa-vāsīnañ ca manussānaṃ mukham pi taṃ-taṃ
saṇṭhānan ti vadanti.
Vism-sn III 1062 (to Vism 447f.):

viṣayatvam anāpannaiḥ śabdair nârthaḥ prakāśyate
na sattva-mātrato ’rthānām agrahītāḥ prakāśakāḥ. (42)
The meaning is not made known through sounds which have not become
sense-objects.
Nor do those who have not apprehended meanings make [anything] known
from the mere [apprehension of] a being. (42)
Vism-mhṭ II 99 = Abhidh-av-mhṭ II 145 = Sadd 37 (Porāṇehi):

tathā hi vadanti:
visayattam anāpannā saddā n’ev’ attha-bodhakā;
na sattā-mattato99 atthe te aññātā pakāsakā ti.

Sounds that have entered no objective field
95 Ce1: pānaḥ.
96 Ce1: ubhayenāpy akabalaṃ kṛtvāvahāryyate; cf. Abhidh-k-bh I 30; III 39.
97 cf. Vism-mhṭ I 243 = Ss 332 ≠ Sp-ṭ 273; also Lok-d 184.
98 (Ce1): a(rḍḍ?)endv-.
99 Sadd: pada-mattato.
216 Cousins
Do not awaken any kind of meaning;

And also beings merely recognized
As such communicate no meanings either.

kāme ’vacaraṇād etat kāmâvacaram ucyate
śeśe ’vacarad apy evaṃ saṃgrāmâvacarâdivat. (43)
That is called kāma-frequenting because it frequents the kāma realm

even when frequenting elsewhere,
just as [an elephant is called] battle-frequenting [when elsewhere]. (43)
Abhidh-av-ṭ I 192:100
hoti c’ettha:
kāme ’vacaratîty etaṃ, kāmāvacara-saññitaṃ.
sese avacarantam pi, saṅgāmâvacaro yathā ti

kṛtyânugrahayor101 bhedāt mānasetarayor dvidhā
bhitvā duḥkha-sukhe yuktir nety upekṣe pañcoditā. (44)
The reason they are given as five is the result of dividing painful and pleasant
[feeling] in two as mental and bodily,
because they differ in the performance of their function, but that is not the case
with neutral feeling. (44)
Vism-mhṭ II 136:
Kiṃ pana kāraṇaṃ mānas’etara-sātâsāta-vasena sukhaṃ, dukkhañ ca vibhajitvā,
vuttaṃ: sukhaṃ somanassaṃ dukkhaṃ domanassan ti, upekkhā pana mānasī,
itarā ca ekadhā va vuttā ti? Bhedâbhāvato; yathā hi anuggahûpaghātakatāya
sukha-dukkhāni aññathā kāyassa anuggaham upaghātañ ca karonti, aññathā
manaso, na evam upekkhā. tasmā bhedâbhāvato ekadhā va upekkhā vuttā ti.
100 Dhs-a 62 ≠ Paṭis-a I 234; cf. I 296: Kiñcâpi hi etaṃ rūpârūpa-bhavesu pi avacarati, yathā
pana saṅgāme avacaraṇato saṅgāmâvacaro ti laddha-nāmako nāgo nagare caranto pi
‘saṅgāmâvacaro’ tv eva vuccati, thala-jala-carā ca pāṇā athale ajale ca ṭhitā pi ‘thala-carā
jala-carā’ tv eva vuccanti, evam idaṃ aññattha avacarantam pi kāmâvacaram evā ti
veditabbaṃ.
101 Or em. to ghātyānugrahayor ‘because they differ in harming and facilitating’.

āddhyātmikāni darśyāni śūnya-grāma-tulyāni102 ṣaṭ
sāśravāni tu bāhyāni ṣaṅ grāma-ghāti-cauravat. (45)
The six internal [bases] should be seen as like an empty house,

but the six external ones which are subject to āśravas are like village-destroying
brigands. (45)
cf. Nāmar-p 672:

suñña-gāmo va daṭṭhabbam ajjhattikam asārato
gāma-ghātaka-corā va, taṃ hanantaṃ va bāhiraṃ.103

pratyaya-pratyayodbhūtair bheri-daṇḍa-svanâdibhiḥ
tulyatvan tad budho nītvā dhātūn paśyen nirātmanaḥ. (46)
After ascertaining the similarity of the elements to such things as a drum, a

drumstick and the sound [produced]
which are conditions and products of conditions, the intelligent person would
see the elements as without self.104 (46)

bāla-janair adhirūḍho janma-rathaḥ kleśa-karma-cakro ’yam.
kleśaika-cakranāsān105 na vahati karmaika-cakras san. (47)
The chariot of birth mounted by foolish folk is this wheel of afflictions and
actions.
After destruction of just the one wheel of afflictions, it does not operate, there
being [only] the single wheel of actions. (47)
102 Em. to: °tulāni.

103 cf. Abhidh-mhṭ II 176 ≠ Vibh-mṭ 36f.:“Suñño gāmo ti kho, bhikkhave, chann’etaṃ ajjhatti
kānaṃ āyatanānaṃ adhivacanan” ti (SN IV 174f.) vacanato suñña-gāmo viya daṭṭhabbāni.
Anna-pāna-sahitanti gahite suñña-gāme yaṃ yad eva bhājanaṃ parāmasīyati, taṃ taṃ
rittakaṃ yeva parāmasīyati, evaṃ dhuvâdi-bhāvena gahitāni yoniso upaparikkhiyamānāni
rittakān’eva etāni dissantī ti. tenâha: dhuva-subha-sukh’-atta-bhāva-virahitattā ti.
Cakkhâdi-dvāresu abhijjhā-domanass’-uppādaka-bhāvena rūpâdīni cakkhâdīnaṃ abhi
ghātakānī ti vuttāni.
104 Vibh-a 80 = Vism 489 = Moh 129: visesato pan’ettha bheri-talaṃ viya cakkhu-dhātu
daṭṭhabbā, daṇḍo viya rūpa-dhātu, saddo viya cakkhu-viññāṇa-dhātu.
105 Em. to °cakranāśān. The meter is āryā.
218 Cousins
Abhidh-av-ṭ II 104:
Tenâhu Porāṇā:
bāla-jan’-ajjhāruḷho
jāti-ratho kamma-kilesa-cakko ’yaṃ
kilesa-cakka-vidhamano
na yāti kamm’-eka-cakke pī ti.

dṛṣṭaṃ santāna-bandhe ’rthe tattvâtattva-parīkṣakaiḥ
sabhāga-janmanas tattvam atattvañ ca na bhedataḥ.106 (48)
?The reality of a shared birth is seen by one investigating the reality or unreality
of something in a stream of continuity and its unreality [is seen] due to the
difference of the moment. (48)
Vism 554 = Vibh-a 163; cf. Abhidh-av 602ff.:

siyuṃ nidassanānettha, paṭighosādikā atha.
santānabandhato natthi, ekatā nāpi nānatā. ||
Abhidh-av-ṭ II 101 (Poraṇā):

Diṭṭha-santāna-bandhasmiṃ ek’-aññatta-vicāraṇe
sabhāg’-uppattiy’ ekattam aññattaṃ khaṇa-bhedato ti |
Appendix 3: Unclear

eṣaḥ pantho ‘vitatho’devayāno
yanti putrino viśokāḥ107
taṃ paśyanti paśavo vayāṃsi
tena te mātary api maithunaṃ caranti. (49)
106 ? read ksaṇabhedataḥ.

107 Add yena and read viśokataṃ? Meter?
Vism-mhṭ II 118 = Abhidh-av-ṭ I 251108:

Ādi-saddena
esa pantho109 vitatho110 deva-yāne
yena yanti puttavanto visokā111
taṃ passanti112 pasavo pakkhino ca
tena te113 mātari pi mithunaṃ carantī
ti-ādinā nayena putta-mukha-dassanaṃ sagga-mokkha-maggo ti evam-ādikaṃ
micchā-diṭṭhiṃ saṅgaṇhāti.

lakṣaṇa-siddham asiddhaṃ yad api ca lakṣyâdibhis tad api siddham
skhalite tu lakṣaṇa-vidhau jñāpaka-siddham bhavati siddham. (50)
meter: āryā
Vism-sn IV 1688 (to Vism 684):

siddhatvaṃ vidyād atha siddhyati pūrvoktam uttaroktena
pūrvoktenottaram api caivaṃ kvacid iha bhavati siddham. (51)
meter: āryā
Abbreviations
Abbreviations used in this chapter are those of the Critical Pāli Dictionary. Page refer-
ences to Vism-sn are to the edition of Saddhātissa, 1949–1955. Citations are to the page
numbers of the editions published by the Pāli Text Society. Page numbers for other
texts are as given for the Burmese edition on the Vipassanā Research Institute (VRI) CD,
unless otherwise indicated.
Abhidh-s(-mhṭ) The Abhidhammatthasaṅgaha of Bhadantācariya Anuruddha and the

Abhidhammatthavibhāvinī-ṭīkā of Bhadantācariya Sumaṅgalasāmi
(ed. Saddhātissa, 1989)
108 Sv-pṭ III (Be) 49 = III (Ee) 62 (vll): Micchā-diṭṭhi-kammassā ti “esa saddhādhigato deva-
yāno, yena yanti puttino visokā” ti-ādinā pavattitassa micchā-diṭṭhi-saha-gata-kammassa.
Mss to Ee variously have: panethāvitato; pan’ esā vitato; saddhâdhigato which Ee recon-
structs to: esa pan’ evâvitatho. Many read vipattino for puttino.
109 Vism-mhṭ (Ce): saddo.
110 Abhidh-av-ṭ (Ce): vitato; (Be): vihito for vitatho; Vism-mhṭ inserts: pag eva before vitatho.
111 Abhidh-av-ṭ (Ce): visokataṃ.
112 Abhidh-av-ṭ (Ce): passanti taṃ.
113 Abhidh-av-ṭ (Ce): te is omitted in the old Be.
220 Cousins
Abhidh-s-sn Abhidhamm’-atthasaṅgaha with the Purāṇasanne of Sāriputta

Saṅgharāja (5th ed. Paññāmoli, 1960)
Vism-sn (Ce1) Visuddhimarga with Commentary of K.K.S.S. Paṇḍita Parākramabāhu
and new explanation by M. Dharmaratne (ed. Pahana, 1890–1946),
Vism-sn (Ce2) The Visuddhimagga with the Commentary written by King Parākrama
bāhu II (ed. Saddhātissa, 1949–1955)
VRI Text cited from version three of the Dhammagiri CD issued by the
Vipassana Research Institute
Bibliography
Bechert, Heinz (1976). “Sanskrit Literature in Sri Lanka as a Paradigm of Regional

Sanskrit Literatures”, In Malalasekera Commemoration Volume. Edited by Oliver
Hector de Alwis Wijesekera. Colombo: Malalasekera Commemoration Volume
Editorial Committee, pp. 23–35.
Bechert, Heinz (1977). “Mahāyāna Literature in Sri Lanka: The Early Phase”, In
Prajñāpāramitā and related systems: Studies in honour of Edward Conze. Edited by
Lewis Lancaster. Berkeley CA: The Group in Buddhist Studies etc., pp. 361–368.
Bechert, Heinz (1998). “Remarks on Buddhist Sanskrit Literature in Sri Lanka from the
9th Century until the End of the Dam̆ badeṇiya Period”, In Sūryacandrāya. Essays in
Honour of Akira Yuyama On the Occasion of His 65th Birthday. Edited by Paul Harrison
and Gregory Schopen (Indica et Tibetica. Monographien zu den Sprachen und
Literaturen des indo-tibetischen Kulturraumes 35). Swisttal-Odendorf: Indica et
Tibetica Verlag, pp. 1–8.
Collins, Steven (1998). Nirvana and other Buddhist felicities: Utopias of the Pali imaginaire
(Cambridge studies in religious traditions 12). Cambridge: Cambridge University
Press.
Cousins, Lance S. (1981). “The Paṭṭhāna and the Development of the Theravādin
Abhidhamma”, Journal of the Pali Text Society, pp. 22–46.
Cousins, Lance S. (2011). “Abhidhamma Studies I: Jotipāla and the Abhidhamma Anuṭīkā”,
Thai International Journal for Buddhist Studies II, pp. 1–36.
Cousins, Lance S. (2012). “The Teachings of the Abhayagiri School”, In How Theravāda
is Theravāda? Exploring Buddhist Identities. Edited by Peter Skilling et al. Chiang Mai:
Silkworm Books, pp. 67–127.
Deeg, Max (2012). “Sthavira, Thera and ‘*Sthaviravāda’ in Chinese Buddhist Sources”, In
How Theravāda is Theravāda? Exploring Buddhist Identities. Edited by Peter Skilling
et al. Chiang Mai: Silkworm Books, pp. 129–163.
Dhammajoti KL Bhikkhu (1997). “The Abhidharma Controversy on Visual Perception”,
In Recent Researches in Buddhist Studies. Essays in Honour of Professor Y. Karunadasa.
Edited by KL Dhammajoti; Asanga Tilakaratne; and Kapila Abhayawansa. Colombo:

Y. Karunadasa Felicitation Committee, Chi Ying Foundation (Hong Kong),
pp. 70–117.
Godakumbura, Charles E. (1943). “References to Buddhist Sanskrit Writers in Sinhalese
Literature”, University of Ceylon Review 1/1, pp. 86–93.
Godakumbura, Charles E. (1955). Sinhalese Literature. Colombo: Colombo Apothecaries.
Griffiths, Paul J. (1994). On being Buddha. The Classical Doctrine of Buddhahood (SUNY
Series, Towards a Comparative Philosophy of Religions). Albany NY: State University
of New York Press.
Hinüber, Oskar von (1996). A Handbook of Pali Literature (Indian Philology and South
Asian Studies 2). Berlin: Walter de Gruyter.
Imanishi, Junkichi (1975). Fragmente des Abhidharmaprakaraṇabhāṣyam in Text und
Übersetzung (Nachrichten der Akademie der Wissenschaften in Göttingen. Philo
logisch-historische Klasse) Göttingen: Vandenhoeck & Ruprecht.
Kahrs, Eivind (1998). Indian semantic analysis: the nirvacana tradition (University of
Cambridge Oriental Publications 55). Cambridge: Cambridge University Press.
Lamotte, Étienne (1958). Histoire du bouddhisme indien (Bibliothèque du Muséon 43).
Louvain: Publications Universitaires Institut Orientaliste.
Li, Rongxi (1995). A Biography of the Tripiṭaka Master of the Great Ci’en Monastery of the
Great Tang Dynasty (BDK English Tripiṭaka 77). Berkeley CA: Numata Centre for
Buddhist Translation and Research.
Li, Rongxi (1996a). The Great Tang Dynasty Record of the Western Regions (BDK English
Tripiṭaka 79). Berkeley CA: Numata Centre for Buddhist Translation and Research.
Li, Rongxi (1996b). Buddhist Monastic Traditions of Southern Asia. A record of the Inner
Law sent home from the South Seas by Śramaṇa Yijing (BDK English Tripiṭaka 93-I).
Berkeley CA: Numata Centre for Buddhist Translation and Research.
Malalasekera, Gunapala Piyasena (1928). The Pali literature of Ceylon (Prize Publication
Fund (Series) 10). London: The Royal Asiatic Society of Great Britain and Ireland.
Mori, Sodō (1989). Studies of the Pāli Commentaries: A Provisional Collection of Articles.
Saitama: Sodo Mori.
Ñāṇamoli Bhikkhu (1956). The Path of Purification (Visuddhimagga). Colombo: R.
Semage.
Pagel, Ulrich (1995). The Bodhisattvapiṭaka. Its Doctrines, Practices and their Position in
Mahāyāna Literature (Buddhica Britannica. Series Continua V). (Edited by Tadeusz
Skorupski). Tring: The Institute of Buddhist Studies.
Pahana, Lakmiṇi et al. (ed.) (1890), Visuddhimarga with Commentary of K.K.S.S. Paṇḍita
Parākramabāhu and new explanation by M. Dharmaratne. Colombo: Laksminipahana
Press (Vol. I 1890; II 1892–1895; III 1898; IV 1909; V 1917; VI 1926; VII 1946).
Paññāmoli, Toṭagamuva Tissa (ed.) (1960). Abhidhammatthasaṅgaha with the Purāṇa
sanne of Sāriputta Saṅgharāja, 5th ed., Colombo.
222 Cousins
Pieris, Aloysius, s.j. (1997). “Cakkhu-viññāṇa which is Dassanamatta: Visual Perception

or Non-Perceptual Vision?”, In Recent Researches in Buddhist Studies. Essays in Honour
of Professor Y. Karunadasa. Edited by KL Dhammajoti; Asanga Tilakaratne; and Kapila
Abhayawansa. Colombo: Y. Karunadasa Felicitation Committee, Chi Ying Foundation
(Hong Kong), pp. 540–566.
Saddhātissa, Bentota (ed.) (1949–1955). The Visuddhimagga with the Commentary written
by King Parākramabāhu II, 4 vols., Kalutara.
Saddhātissa, Hammalawa (ed.) 1989). The Abhidhammatthasaṅgaha of Bhadantācariya
Anuruddha and the Abhidhammatthavibhāvinī-ṭīkā of Bhadantācariya Sumaṅgala
sāmi. Oxford: Pāli Text Society.
Samtani, Narayan Hemandas (ed.), (1971). The Arthaviniścaya-sūtra & its Commentary
(Nibandhana). (Tibetan Sanskrit Works Series Vol. XIII). Patna: K.P. Jayaswal Research
Institute.
Wijeratne, Ranjan and Gethin, Rupert (2002). Summary of the Topics of Abhidhamma
(Abhidhammatthasaṅgaha). Oxford: Pāli Text Society.
Zacchetti, Stefano (2002a). “An early Chinese translation corresponding to Chapter 6 of
the Peṭakopadesa. An Shigao’s Yin chi ru jing T 603 and its Indian original: a prelimi-
nary survey”, Bulletin of the School of Oriental and African Studies 65/1, pp. 74–98.
Zacchetti, Stefano (2002b). “Some Remarks on the Peṭaka Passages in the Da zhidu lun
and their Relation to the Pāli Peṭakopadesa”, Annual Report of The International
Research Institute for Advanced Buddhology at Soka University 5 (March), pp. 67–85.
The Contribution of Saṃghabhadra 223
Chapter 6
The Contribution of Saṃghabhadra to Our

Understanding of Abhidharma Doctrines
KL Dhammajoti
1 Introduction
The whole range of Abhidhamma and Abhidharma extends in space from

India to Central Asia and China. One lineage that has the most significant
impact on virtually the whole of Indian and Chinese Abhidharma studies is
undoubtedly the Sarvāstivāda. From this broad lineage, the great contribution
of Vasubandhu is well known. But his junior contemporary, Saṃghabhadra,
equally, if not more brilliant and influential, is relatively little discussed.
The Sarvāstivāda enjoyed a long period of uninterrupted development.
Its doctrines had great impact on the doctrinal evolution of not only the
Śrāvakayāna schools, but also of the Mahāyāna. It were the Sarvāstivādins
– with their compilation of the gigantic Abhidharmamahāvibhāṣā (MVŚ)
completed around the middle of the second century CE – who comparatively
developed the most systematic doctrinal system, to which most, if not all, the
subsequent Śrāvakayāna doctrines were in part directly or indirectly indebted.
They were also most probably the earliest to systematically seek to account for
the experience of continuity against the backdrop of the early Buddhist teach-
ings of selflessness (nairātmya) and impermanence (anityatā) – and later on
momentariness (kṣaṇikatva).
The Dārṣṭāntikas within this broad tradition developed their epistemologi-
cal theories in dynamic interaction and contention with the Sarvāstivādin
Ābhidharmikas, evolving eventually into the Sautrāntikas with their distinc-
tive bīja (seed) theory.
The Sarvāstivādin Yogācāra masters inherited the Ābhidharmika analysis of
dharmas and their broad system of spiritual praxis and contributed, together
with a certain section of the Sautrāntikas, to the establishment of the Mahāyāna
Yogācāra school.
The early Prajñāpāramitā scriptures (e.g. the Pañcaviṃśatisāhasrikāpra
jñāpāramitā) too display, from the beginning, an unmistakable familiarity
with Sarvāstivāda Abhidharma doctrines. These include (at least, in respect
of enumeration, terminology and broad outline): the contrast between the

224 Dhammajoti
perfected wisdom (particularly the notion of sarvākārajñatā) and compassion

(mahākaruṇā) of the Buddha and those of the two yānas (pratyekabud-
dha and śrāvaka); the process of abandoning contaminations, divided into
darśanamārga (including the distinctive Sarvāstivādin scheme of the 16 kṣaṇas
and the 16 ākāras of the four noble truths) and bhāvanāmārga, the path of cul-
tivation; meditative attainment such as the nine anupūrvavihārasamāpattis;
contamination (kleśa) vs habitual residue (vāsanā), and that the Buddha alone
was able to completely abandon all kleśas together with the vāsanā; etc.
In brief, the doctrinal importance of the Abhidharma tradition can be
summed up as its all-pervasive impact – since the establishment of the
Sarvāstivāda system – on the development of practically every Buddhist school
of thought (and, to some extent, also on certain non-Buddhist tenets). Indeed,
we cannot stress more the significant contribution of Abhidharma doctrines
for the whole of the development of Indian Buddhism.
In ancient China too, particularly since Xuanzang 玄奘, their impor-
tance was well acknowledged. It was in fact thanks to the devoted effort of
Xuanzang et al., that the Indian Abhidharma texts came to be preserved in
Chinese translation, including the gigantic 200-fascicle translation of the
Abhidharmamahāvibhāṣā (MVŚ). Xuanzang and his pupils, all committed
Mahāyāna Yogācārins themselves, devoted a great amount of time and effort in
studying and expounding these texts. Now, in spite of their great importance,
it is only in relatively recent years that we can observe a keener interest in the
study of the Abhidharma tradition.
This is heartening; but much more is awaiting our further serious investi-
gation, and many inadequate or improper assumptions need to be rectified.
There are of course many reasons – many already well discussed by scholars
– as to why progress in this regard has not been so satisfactory, such as the
immensity of the Abhidharma material transmitted in Chinese translation.
One important reason that deserves to be highlighted is the over-reliance on
the Abhidharmakośabhāṣya (AKB) ever since Louis de La Vallée Poussin’s trans-
lation and, later, the publication of the Sanskrit texts of the AKB and Yaśomitra’s
Sphuṭārthā Abhidharmakośavyākhyā (Vy). The great reference value of these
two texts in their original Sanskrit is beyond doubt. But, important as they are,
it hardly needs belaboring the fairly well-known point that their expositions
of the Abhidharma doctrines are tainted with an anti-Abhidharmic stance
– notwithstanding the fact that Vasubandhu indeed, true to his own words,
presented these doctrines largely in accordance with what he believed to be
the “Vaibhāṣika” views.
Moreover, there had been further development and fine-tuning of the
Abhidharma doctrines since the time of the MVŚ, and the AKB and the Vy
alone cannot suffice as a source of information. The impact of the Vaibhāṣika

doctrines, particularly those via Saṃghabhadra, seems to have been under-
examined. On this occasion, therefore, I would like to specifically underscore
the importance of studying Saṃghabhadra’s expositions in this connection.
Owing to space constraint, I shall in this chapter only outline a few major
perspectives illustrating the doctrinal contributions of Saṃghabhadra and
their impact on both the Abhidharma tradition and other related Buddhist
traditions. In a couple of these illustrations, I provide relatively more details,
considering that the issues concerned have so far received little attention
elsewhere. On the whole, however, I am concerned with underscoring certain
important points, rather than offering a comprehensive survey.
I will first discuss and demonstrate how Saṃghabhadra’s views are exten-
sively cited, and reckoned with, as authoritative exposition of Abhidharma
doctrines in the Abhidharma and Yogācāra traditions (2). Then I will discuss
Saṃghabhadra as a major source for our understanding of the Dārṣṭāntika-
Sautrāntika doctrines, as well as the relationship between the Dārṣṭāntikas and
Sautrāntikas (3). This is followed by some illustrations (4) of Saṃghabhadra’s
contribution to the articulation of the Sarvāstivāda-Vaibhāṣika epistemologi-
cal and ontological doctrines, and also (5) of the so-called “neo-Sarvāstivāda”
doctrines of Saṃghabhadra. I will then end with a brief discussion of the
*Nyāyānusāra (Ny) as a source for Mahāyāna studies (6), and some remarks on
the comparative study of the Theravāda Abhidhamma and the Abhidharma
doctrines (7).
2 Saṃghabhadra’s Views Extensively Cited in Abhidharma and

Yogācāra Commentaries
2.1 Abhidharma Commentaries

Saṃghabhadra (ca fifth century CE) is undoubtedly one of the most bril-
liant Abhidharma masters in India. He left behind two valuable works: the
*Nyāyānusāra (Shun zhengli lun 順正理論) and the *Abhidharmasamaya
pradīpikā (Apidamo xian zong lun 阿毘達磨顯宗論), translated into Chinese
by Xuanzang in 80 and 40 fascicles, respectively. It is no exaggeration to state
that they are among the most precious sources for our understanding of not
only Sarvāstivāda Abhidharma, but the whole of the Abhidharma tradition,
including the doctrines of the Dārṣṭāntika-Sautrāntikas.
Louis de La Vallée Poussin has long noted their importance and, in
his translation of the AKB, he quite often refers to them, especially the
226 Dhammajoti
Ny, in his annotation. On Louis de La Vallée Poussin’s translation of the

Abhidharmakośabhāṣya, Étienne Lamotte comments:
I was not surprised to notice that the French translation […] based mainly
on the Chinese version of Hiuan-tsang, adhered literally to the original
Indian text. The honor of such an exceptional success must be given both
to the Chinese master of the seventh century and the Belgian master of
the twentieth century […].1
When we collate Xuanzang’s translation of the AKB and the Ny, we can find
evidence, at least in some places, that Xuanzang’s understanding of the AKB
was guided by the Ny. Moreover, several of his pupils – Puguang 普光, Fabao
法寶, Shentai 神泰 and Huaisu 懷素 – had all received instructions from him
on the AKB and wrote important commentaries on the text (Huaisu’s is no lon-
ger extant), citing the Ny extensively and indispensably. Puguang, for instance,
quotes it by name more than 2,600 times, and generally takes Saṃghabhadra
as representing the correct Vaibhāṣika exponent. In the following, I will give
two of many similar examples that show the influence of the Ny on Xuanzang’s
pupils, and the central position which Saṃghabhadra often holds in the con-
text of Abhidharmic controversial debates.
(1) On the nature of the non-defiled nescience (akliṣṭājñāna), Puguang first
enumerates eleven Indian opinions.2 Information on opinions (1) to (4) must
have been derived from his teacher, Xuanzang. For opinions (5) to (11), he cites
Saṃghabhadra’s Ny. He then proceeds to refute them one by one, the first four
with his own reasoning basing himself on Abhidharma tenets.3 For the refuta-
tion of (5) to (11), he simply quotes Saṃghabhadra’s arguments, offering some
elaboration in each case.4 Finally, he cites, as the “correct meaning” (zheng yi
正義), Saṃghabhadra’s own interpretation that the akliṣṭājñāna has, as its
intrinsic nature, all the inferior prajñās (understanding) that are impure
(sāsrava) and non-defiled (akliṣṭa), which are present in one who has not yet
attained buddha-hood.5
Puguang’s enumeration of the first four opinions (the third is ascribed to a
“western master” (pāścātīya), named Deguang 德光, *Guṇaprabha) shows that
1 Lodrö Sangpo, 2012, vol. 1: 73.

2 T.41.1821: 3c29–4b19.
3 T.41.1821: 4b19-c9.
4 T.41.1821: 4c9–5b24.
5 T.41.1821: 5a24–26. He cites Saṃghabhadra verbatim in full (T.41.1821: 5a24-b20). See also
T.29.1562: 502a21-b13.
the views of other Abhidharma masters were known to the Xuanzang’s tradi-
tion – which is only to be expected since Xuanzang studied in India for many
years. Besides, Sthiramati’s Abhidharmkośaṭīkā Tattvārthā (= Tattvārthā) was
translated into Chinese by Xuanzang.6 In fact, both Puguang and Fabao occa-
sionally also quote its rejoinder to Saṃghabhadra’s objection to Vasubandhu’s
explanation, as can be seen from the next example.
(2) Vasubandhu defines the avijñapti thus:
vikṣiptācittakasyāpi yo ’nubandhaḥ śubhāśubhaḥ |

mahābhūtāny upādāya sa hy avijñaptir ucyate | (AKB, 8)
Both Puguang and Fabao quote Saṃghabhadra’s several objections to this

definition. On Vasubandhu’s commentarial explanation of “api” in the above
verse, Puguang (followed by Fabao) quotes both Saṃghabhadra’s criticism7
and Sthiramati’s defence.8 He further quotes Saṃghabhadra’s objection to
Vasubandhu’s description of the avijñapti as a serial flow (prabandha) as well
as the defense by “the Kośa masters” (jushe shi 俱舍師).9
Thus, for Xuanzang and his pupils, Saṃghabhadra stood out as the most
authoritative master among the Ābhidharmikas to whose views they were well
exposed. In all the mentioned contexts, Saṃghabhadra’s authority is well
respected – even where objections by other masters to his criticism are quoted
alongside.
Likewise, for Sthiramati and other AKB commentators subsequent to the
compilation of the Ny, Saṃghabhadra’s views are certainly to be reckoned
with. As a matter of fact, Sthiramati’s Tattvārthā too cites Saṃghabhadra’s
views extensively (sometimes without mentioning his name), both as the
authoritative representative of the Vaibhāṣika tenets and as the source of the
views by other Abhidharma masters. Where he feels Vasubandhu’s explana-
tion is seriously challenged, he no doubt rises up in defence. The same is true
for Yaśomitra, who also cites Saṃghabhadra often and defends Vasubandhu
against Saṃghabhadra’s criticism. A case in point, in the context of Sthiramati,
6 Only a small portion of this translation is preserved. For the past many years, Professor Odani
Nobuchiyo, et al., have been working on an edition of the Sanskrit manuscript recently ob-
tained from China.
7 T.29.1562: 335b27–29: “ 又謂等言通無心者, 此言無用; 前已攝故: 亂心等言, 已攝一切
餘有心位; 第二等言, 復何所攝? …” The criticism covers more than what I have discussed
here.
8 T.41.1821: 22a15–28.
9 T.41.1821: 22b2–22.
228 Dhammajoti
is Saṃghabhadra’s criticism of Vasubandhu’s definition on avijñapti men-

tioned above.
But then again, where there is no such threat, Saṃghabhadra’s explanations
are actually often quoted by Sthiramati to refute other variant Abhidharma
views. For instance, in a discussion on whether the sensory consciousnesses
can be morally defined (i.e. good (kuśala) or bad (akuśala)), Saṃghabhadra
quotes a certain opinion which states that since they, unlike mental conscious-
ness, are completely without the capacity to conceptualize and since they
focus on an object-field for no more than a single moment, they cannot be
morally defined. He then proceeds to refute this view, and also describes the
Mahīśāsaka view, which he likewise refutes.10 Sthiramati quotes the above dis-
cussion, including the Mahīśāsaka view, virtually verbatim from Saṃghabhadra,
agreeing with him, albeit without acknowledgment.11
On the discussion as to whether a sensory consciousness cognizes a concep-
tualized whole or an assemblage of atoms, Sthiramati cites both Śrīlāta’s view
and his simile of the blind, followed by Saṃghabhadra’s refutation – again, all
verbatim from the Ny – in each case adding some elaborations, which are of a
non-committal nature.12 Finally, as to why, if the sensory consciousnesses
indeed cognize ultimate reals, they do not effectuate the abandonment of con-
taminations, Sthiramati once again simply records verbatim the brief dialogue
in the Ny, without further comments. The only addition from him is the posi-
tive identification of the question as being raised by Śrīlāta. (He begins the
dialogue with “Śrīlāta further says: …”).13
2 Yogācāra Commentaries
In their Yogācāra commentaries too, the Xuanzang tradition as well as
Sthiramati quote Saṃghabhadra as the authoritative representative of the
Vaibhāṣikas. (See also below). Thus, Kuiji 窺基, considered as Xuanzang’s
chief pupil, in his commentary on Xuanzang’s Cheng weishi lun 成唯識論
(CWSL Shuji 述記), describes Sthiramati as the master who “compiled the
*Abhidharmasamuccayavyākhyā, rescued/defended the AKB and refuted the
*Nyāyānusāra master (jiu jushe po zhengli shi 救俱舍破正理師)”.14 Even in this
description, the indication is evident that the Yogācāra master, Sthiramati, is
10 T.29.1562: 348c12–349b7.
11 Tattvārthā(C), 222 f. See also Dhammajoti, 2012: 244f.
12 Dhammajoti, 2012: 216f.
13 Tattvārthā(C), 239.
14 T.43.1830: 321c19–23: “ 梵云悉恥羅末底。唐言安慧。即糅雜集。救俱舍論破正理
師 …”.
familiar and quite concerned with Saṃghabhadra’s sophisticated views and

criticism. In both Sthiramati’s Triṃśikāvijñaptibhāṣya (TVB) and Xuanzang’s
CWSL, it is Saṃghabhadra’s expositions that are quoted as the Vaibhāṣika
(not directly so named) theories. Kuiji records and identifies them as
Saṃghabhadra’s “neo-Sarvāstivāda” tenets.15 (See also below, 5).
Tunnyun (遁倫, also known as 道倫), a Korean master in the Tang 唐 dynasty
period, in his commentary on the Yogācārabhūmiśāstra (Yujia lun ji 瑜伽論記)
– which is based on Kuiji’s – also cites the Ny numerously in his discussion on
Abhidharma terms and definitions,16 and when contrasting Abhidharma
explanations with those of the Yogācāra. For instance, on the sūtra statement
that consciousness (vijñāna) does not depart from the body of one in attain-
ment of cessation, Kuiji comments that – in contrast with the Yogācāra
explanation in terms of the continuous existence of the ālayavijñāna – the
Sarvāstivādins explain this to mean that the acquisition (prāpti) of conscious-
ness has not departed from the body, which, however, is not the proper
Sarvāstivāda defence. Their “proper/true defense” (zheng jiu 正救) is that while
there is no consciousness when one has entered into nirodhasamāpatti, one’s
consciousness necessarily arises upon one’s coming out of the meditation”.17
Tunnyun comments in exactly the same way.18 Elsewhere in the same CWSL
Shuji, Kuiji again refers to this latter (what he calls the “proper”) explanation,
adding immediately that this is like the case that a dormant disease, though
not presently manifesting, is said to not have departed from the body.19 This
example is, in fact, given by Saṃghabhadra in this very same context.20 It is
thus clear that both Kuiji and Tunnyun take Saṃghabhadra’s explanation as
the “proper” Sarvāstivāda explanation.
In a discussion on what contaminations “have basis” (savastuka) and what
contaminations “have no basis” (avastuka), Dunlun notes that there are diverse
interpretations in the Yogācāra texts. He then cites the Tripiṭaka Master who
15 Cf. T.43.1830: 271a10: “眾賢論師新薩婆多義”. Also in his commentary on Vasubandhu’s

Viṃśatikā: T.43.1834: 992a27–28: “新薩婆多順正理師”.
16 On the visible (sanidarśana): T.42.1828: 316a; on the definition of pratītyasamutpāda:
T.42.1828: 372b28-c4.
17 T.43.1829: 172b23–29.
18 T.42.1828: 595c15–29.
19 T.43.1830: 369c4–7.
20 T.29.1562: 403c13–17: “引契經說識不離身, 於定無心, 亦無違害. 以即於此所依身中,
識必還生, 故言不離. 謂一相續眾同分中, 識相續流, 非畢竟斷. 譬如鬼病暫不發
時; 由未永除, 仍名不離”. Also: T.29.1562: 771c23–26: “ 故滅定中無有心理. 非迷正理,
纔覽經文, 便能會通聖教深趣. 識非永滅, 言不離身. 如病未永除, 暫息亦
名有”.
230 Dhammajoti
asserts that this distinction should be coherently comprehended in accordance

with the Ny, “which states that, as spoken in the sūtra, the contaminations
abandonable through seeing (darśanaheya) take cognitive objects that have no
support; the contaminations abandonable through cultivation (bhāvanāheya)
take cognitive objects that have support […]”.21
Needless to say, all this does not mean that the Yogācāra masters’ under-
standing of their own tenets were determined by the Vaibhāṣika, especially
Saṃghabhadra’s, expositions. But it does show that they greatly valued and
made a great effort to study them, both by way of understanding their inheri-
tance from the Sarvāstivāda tradition from a fuller perspective, and by way of
contrasting and enriching the doctrinal interpretations within their own
tradition.
In the later period, we can still feel Saṃghabhadra’s impact on the Yogācāra
(or syncretic Yogācāra-Mādhyamika) masters’ exposition of the Vaibhāṣika
views. One example is Śāntarakṣita’s (eighth century CE) Tattvasaṃgraha and
Kamalaśīla’s commentary on it, i.e. the Pañjikā, where Saṃghabhadra’s articu-
lation of activity (kāritra) as projection of the fruit (phalākṣepa) is presented as
the final answer to the Sautrāntika criticism on the Sarvāstivāda-Vaibhāṣika
theory of temporality.22 The Pañjikā (508, 509) likewise explicitly attributes the
Vaibhāṣika claim of the neither-identical-nor-different relationship between a
dharma and its kāritra to Sahantabhadra (= Saṃghabhadra). In brief, the
account of the whole debate therein, on the issue of tri-temporal existence,
clearly indicates the author’s familiarity with the Ny account.
3 Dārṣṭāntika-Sautrāntika Doctrines and the Relationship between

the Dārṣṭāntikas and Sautrāntikas
Saṃghabhadra’s Ny is not only an indispensable source for the study of the

orthodox Vaibhāṣikas – often unfairly presented by Vasubandhu and Yaśomitra
– and those of other Abhidharma masters. It is also a most valuable source for
our understanding of the tenets of other schools of thought, particularly the
Dārṣṭāntika-Sautrāntikas, for which very little information is available in the
extant Sanskrit texts. Louis de La Vallée Poussin has long noted its importance
in this connection:
21 T.42.1828: 684a27–29: “… 此等諸文, 云何會釋? 三藏云: 准正理論文會釋. 謂彼論云:

如經曰見所斷惑無事所緣, 修所斷惑有事所緣 …”.
22 The Pañjikā, 506, explicitly attributes this definition to Sahantabhadra (= Saṃghabhadra).
Vasubandhu speaks the truth when he says that his book is an explana-
tion of the doctrines of the Vaibhāṣikas of Kaśmīr […]. But the Kośa, as
we know, is also the refutation of several essential doctrines of the
Vibhāṣā. Vasubandhu, who the orthodox Vaibhāṣika Saṃghabhadra calls
“the Sautrāntika”, sets the view of the Sautrāntikas or his own views
against the views of the Sarvāstivādin-Vaibhāṣikas: here we have very lit-
tle information on the sources that he uses; he makes use of data foreign
to and undoubtedly later than the Vibhāṣā, the Dārṣṭāntikas–Bhadanta–
Dharmatrāta cited by the Vibhāṣā. We will see more clearly when we have
read Saṃghabhadra who gathers information on Śrīlāta, who must be
one of the Sautrāntika sources of Vasubandhu.23
Indeed, among the extant Abhidharma texts, it is in the Ny that we find detailed
doctrinal descriptions and arguments of the Dārṣṭāntika-Sautrāntikas, along
with Saṃghabhadra’s lengthy refutations. Of particular importance are
Sthavira Śrīlāta’s doctrines, such as: the anudhātu (or *pūrvānudhātu) theory,
the cognitive process, the citta-caitta doctrine, the doctrine that the sense-ele-
ments (dhātu) alone are real existents (and not the aggregates (skandha) or the
sense-spheres (āyatana); the process of direct perception (pratyakṣa); etc.24 It
is also in this work that we are introduced to Śrīlāta’s explanation of the eight-
moment process of direct realization (abhisamaya).25
The Ny also clarifies the relationship between the Dārṣṭāntikas and the
Sautrāntikas. Basing ourselves on the MVŚ, the AKB and the Ny, we can see that
the early Dārṣṭāntikas figuring in the MVŚ were Sarvāstivādins: They shared the
school’s central thesis of the tri-temporal existence of dharmas. At least around
the time of the AKB, we see one sector of them evolving into Sautrāntikas, com-
ing to hold the present-only-exists standpoint. Śrīlāta, usually described – as in
Ny – as a Dārṣṭāntika master, was also sometimes called a Sautrāntika.
Xuanzang informs us that he composed the *Sautrāntikavibhāṣā (Jing bu pipo-
sha 經部毘婆沙), which Xuanzang himself studied in India. From this, we can
23 Lodrö Sangpo, 2012: 78. Had Louis de La Vallée Poussin lived several years longer, we
would have had the good fortune of his study of the Ny. Although he quotes Saṃghabhadra
frequently in his annotations, it appears that he probably had not made any detailed
study of the text. We find him – at an earlier stage, for instance – at a couple of places
attributing the anudhātu doctrine of the “Sthavira”, in the Ny, to the “the school of the
Sthaviras”! (See his French translation of the AKB: de La Vallée Poussin, 1971: 173, note 5,
and 246, note 2). These were later amended by him in his “Additions and Corrections”.
24 We have been able to make some study of these doctrines to some extent. See Dhamma-
joti, 2007a; 2007b; 2011b; 2013.
25 Cf. T.29.1562: 684a-b.
232 Dhammajoti
understand that he was a Sautrāntika leader at the time, for which reason
Saṃghabhadra refutes his doctrines extensively when refuting Vasubandhu. In
fact, as is well known, Yaśomitra states that “sautrāntika” means the same as
“dārṣṭāntika”.26
Thus, in this later period, the two labels, “Dārṣṭāntika” and “Sautrāntika” had
come to be interchangeable, and no pejorative implication (contrary to Jean
Przyluski’s theory) is detectable in the first label.
We can also observe the same in the “what sees” debate. Vasubandhu attri-
butes the criticism that the whole debate is in vain to the Sautrāntika (AKB, 31:
sautrāntikā āhuḥ …); but both the Abhidharmadīpa-Vibhāṣāprabhāvṛtti of the
Abhdharmadīpa and the Ny attribute it to the Dārṣṭāntikas.27
The Ny, which predates the Vy, provides further instances showing the
interchangeability between the two labels. For example, Saṃghabhadra
states that Śrīlāta and his pupils are those who assert that all sūtra is explicit
meaning (nītārtha).28 This is none other than the fundamental stance of the
“Sautrāntikas”.
On the issue of nītārtha and neyārtha sūtras (sūtras of implicit meaning),
Saṃghabhadra states more explicitly that Śrīlāta – and along with him
Vasubandhu as well – is not worthy of the Sautrāntika claim that they take the
sūtras as the final authority; thus implying that Śrīlāta was identified as a
Sautrāntika during Saṃghabhadra’s time:
[…] the Sthavira himself, from various perspectives, interprets the pro-
found implicit meanings of that sūtra. […] Which wise person is so
attached [as he], as to call a sūtra having an implicit meaning a nītārtha
sūtra? […] It is extremely illogical [for him] to claim: “we take the sūtras
as the decisive authority” (我用經為定量; *vayaṃ sūtraprāmāṇikāḥ29) –
while not discerning the difference between the nītārtha and neyārtha
sūtras. […] In this connection, the Sūtrakāra (i.e., Vasubandhu) asserts
thus: “…” This [assertion] should be fully refuted as in the case of the
Sthavira’s tenet.30
26 “na sautrāntiko na dārṣṭāntika ity arthaḥ |”.

27 Jaini, 1959: 31; T.29.1562: 367b24–27.
28 T.29.1562: 482a28-b3: “ 如是師徒 … 彼論說, 經皆了義 …”.
29 Cf. Wogihara, 1971: 11: “kaḥ sautrāntikārthaḥ | ye sūtra-prāmāṇikā na śāstra-prāmāṇikāḥ |”.
30 T.29.1562: 495c14–28: “… 彼上座, 自以多門, 解釋彼經深隱理趣, … 誰有智者, 執著
如斯 – 有別意經, 名為了義. 又經處處, 以種種門, 廣說緣起; 多非了義. 皆隨所
應, 當求意旨. 如是, 不達了不了義經差別相, 而稱 “我用經為定量”. 甚為非
理! … 經主於此, 假作是言, 經部諸師。作如是白. … 此如上座宗應廣遮遣”.
In the debate on whether envelopment (paryavasthāna) and proclivity

(anuśaya) are different or identical – for the Vaibhāṣikas, they are identical;
for the Dārṣṭāntika-Sautrāntikas, different31 – we see another instance where
Saṃghabhadra clearly refers to Śrīlāta’s explanation as a Sautrāntika tenet.
Saṃghabhadra here cites a sūtra in which it is stated that the purpose of
practicing the spiritual life (brahmacarya) is for the abandonment of attach-
ment (rāga), hatred (dveṣa) and delusion (moha), without any mention of the
abandonment of anuśayas. Śrīlāta replies that in this context, the rāga, etc,
themselves are the anuśayas.32 Saṃghabhadra, objecting to this, states:
It is not the case that, in the Sautrāntika tenets, the paryavasthānas can
be abandoned (fei jing bu zong (chan) you duan yi 非經部宗 (纏) 有斷義),
because, [according to them,] they are conjoined with thought
(cittasaṃprayukta) and the past and future are non-existent […].33
As yet another similar instance: Criticizing Vasubandhu’s exposition of the

process of karmic retribution in terms of distinctive transformation in the
serial continuity (santatipariṇāmaviśeṣa), Saṃghabhadra points out its logical
invalidity, again connecting it up with the Dārṣṭāntika tenets:
This is because, for the Dārṣṭāntikas, there can be no serial continuity in

terms of a before–after difference, nor can there be the tri-temporal con-
ditionings (saṃskāra) qua cause and effect, nor can there be any efficacy
that yields the fruit immediately […].34
31 Śrīlāta cites a sūtra (Cf. Mahāmāluṅkyasutta, Majjhima I, 433) to substantiate his claim
that “the Buddha, Bhagavat, himself explains the difference between the paryavasthānas
and the anuśayas: When a kleśa comes into manifestation, it is called a paryavasthāna,
because it presently binds the person’s serial continuity; the anudhātu of the kleśa is
called [its] anuśaya, for [its] causal efficacy always follows along and lays dormant”.
(T.29.1562: 597b27-c3: “ 上座於此, 謂佛世尊, 自說諸纏與隨眠異. 謂: 諸煩惱現起,
名纏; 以能現前縛相續故. 煩惱隨界, 說名隨眠; 因性恒隨而眠伏故. 以契經說:
幼稚童子嬰孩眠病, 雖無染欲, 而有欲貪隨眠隨增. 此唯說有諸隨眠性 …”).
32 T.29.1562: 598a1–5.
33 T.29.1562: 598a7–8.
34 T.29.1562: 398b12–17.
234 Dhammajoti
4 Saṃghabhadra’s Contribution to the Articulation of the

Epistemological and Ontological Doctrines of the
Sarvāstivāda-Vaibhāṣikas
The Ny is also a great treasure for our understanding of the epistemological

theories of the Abhidharma tradition. To begin with, Saṃghabhadra brilliantly
articulates the Sarvāstivāda-Vaibhāṣika doctrines of cognition. He offers a clear
definition of an existent (sat) as that which is capable of generating a cogni-
tion – which can be either an absolute real or a conceptual real based on the
former. This enables the Sarvāstivādins to argue more cogently for the exis-
tence of past and future dharmas. And in this, he demonstrates that the
Ābhidharmikas’ ontological theories proceed from, and are intrinsically bound
up with, their epistemological considerations.
Saṃghabhadra demonstrates the doctrinal importance of “co-occurring
causality” (sahabhūhetu) in the understanding of human experiences involv-
ing the various types of forces (i.e. the dharmas) in the universe, such as the
citta-caitta processes, bondage and liberation from contaminations, direct per-
ception (pratyakṣa), etc. It was Saṃghabhadra who most systematically
defended this major innovation of the Sarvāstivāda against the Dārṣṭāntika-
Sautrāntika objections. In this process, he articulates that there are in fact only
three situations where such a co-occurring causality obtains among co-nascent
dharmas:
[1] among those that share the same effect; or

[2] that are reciprocally effects; or
[3] where by the force of this, that dharma can arise (virile effect). Such
co-nascent [dharmas] have a cause–effect relationship, [i.e. are co-
existent causes]. (Ny, 419c26–29)
Significantly, this causal category came to be held by the Yogācāra school as the
only true type of causality – quite in spite of the dominant doctrinal influence
on this school by the Sautrāntikas who were (as evident in all the extant
Abhidharma texts) vehemently opposed to it – and, indispensably, was uti-
lized in their establishment of the vijñaptimātra thesis and of the doctrines of
bīja (seed), vāsanā (habitual residue), and the simultaneously occurring pro-
cess of bījas in the ālayavijñāna manifesting into phenomena and the
perfuming of these phenomena on the bījas.35
35 Cf. Dhammajoti, 2009: 157–164.

Saṃghabhadra also enunciates the Sarvāstivāda-Vaibhāṣika doctrine of

direct perception (pratyakṣa) in a distinct way as involving three necessarily
related aspects spanning two moments: (1) faculty-based, (2) qua experience,
(3) qua discernment. The first two aspects occur in the first moment of a sen-
sory perception, when the sensory faculty perceives the external object (1),
and at the very same time – thanks to co-occurring causality – the sensory
consciousness arises together with sensation (vedanā) and other thought-
concomitants (2). He argues that this fact of directly sensing/experiencing the
object outside in the very same present moment enables the person to have
the vivid impression of having directly and personally perceived the object. It
is only on the basis of this vivid, direct experience, that it becomes possible, in
the immediately following moment, to have a clear discernment of the object
as having been directly perceived (3). This third and last aspect is the stage of
the corresponding mental consciousness which is capable of discerning.
Saṃghabhadra clarifies the Sarvāstivāda definition of a dharma’s “activity”
as its efficacy, uniquely arising only in the present moment, of projecting/
grasping its own-fruit (phala-ākṣepa; phala-pratigrahaṇa). This definition –
often labeled as a neo-Sarvāstivāda doctrine (but see below) – coupled with his
employing of the below-mentioned theory of ‘difference in mode of being’
(bhāvānyathātva) of Dharmatrāta, greatly contributes to an articulate
Sarvāstivādin distinction among dharmas existing in the tri-temporal periods.
Dharmatrāta’s theory states that the svabhāva (intrinsic nature) of the tri-
temporal dharmas remains unchanged; but their bhāva (mode of being)
changes. In the MVŚ, this theory was criticized as being unsatisfactory. The rea-
son is: “Apart from a dharma’s svabhāva, what is it that is called its bhāva?
When a conditioned dharma reaches the present from the future, its preceding
mode of being ought to cease. When it reaches the past from the present, its
succeeding mode of being ought to arise: How can it be logical that what is
past can arise and what is future can cease?”36 Vasubandhu’s criticism is much
more severe, branding it a Sāṃkhya doctrine of transformation (pariṇāma).
Saṃghabhadra, however, argues that it is in part similar to Vasumitra’s
theory,37 which is the one endorsed by the MVŚ compilers. In fact, he consis-
tently employs the bhāvānyathātva theory along with Vasumitra’s ‘difference
in state’ (avasthānyathātva) theory in his exposition and defence of the thesis
of sarvāstitva. In this way, Saṃghabhadra is able to argue more cogently that
36 T.27.1545: 396b18–22.
37 T.29.1562: 631, b9–10.
236 Dhammajoti
the tri-temporal existence of a dharma implies neither its permanence

(nityatā)38 nor that its existence is identical in the three temporal periods:39
The essential nature of a dharma remains always; its mode of being

(bhāva) changes: When a conditioned dharma traverses time, it gives rise
to its activity (kāritra) in accordance with the causal conditions, without
abandoning its svabhāva; immediately after this, the activity produced
ceases. Hence it is said that the svabhāva exists always and yet it is not
permanent, since its bhāva changes.40
This is part of Saṃghabhadra’s response to Vasubandhu’s critique of the

Sarvāstivāda position – which is ridiculed as an act of the Almighty – that a
dharma’s svabhāva exists always and yet its bhāva is not acknowledged to be
permanent, nor is its bhāva acknowledged to be distinct from its svabhāva.41
This ridicule implies that between the completion of the MVŚ and the AKB,
such a position must have already become an integral part of the Sarvāstivāda-
Vaibhāṣika exposition of sarvāstitva. From this perspective, Saṃghabhadra’s
advocation of Dharmatrāta’s theory is not an innovation. Nevertheless, it was
he who most articulately (at least in the extant Abhidharma texts) demon-
strates its doctrinal significance for the thesis of sarvāstitva.
I quote here below another exposition of Saṃghabhadra in which he bril-
liantly argues that a dharma with the same intrinsic nature can have different
modes of being, and thus a distinct mode of being implies a distinct existence
of that dharma. Such a position, when successfully established, importantly
supports the defence that the claim of the tri-temporal existence of a dharma
does not amount to a doctrine of its permanence.
[Opponents:] – When a conditioned dharma is traversing the three peri-

ods of time, there being no variation in respect to its essential nature
38 See also Dhammajoti, 2009: 136.

39 See Dhammajoti, 2009: 71–72.
40 T.29.1562: 633c24–26. Further examples: T.29.1562: 628b26–27: “非去來有如現在, 以於
一切同實有中, 許有種種有性別故”. Also, establishing tri-temporality in terms of both
Vasumitra’s as well as Dharmatrāta’s theories: T.29.1562: 633b29-c11: “ 又略說者: 如諸有
為, 實體雖同而功能別; 如是三世實體雖同, 於中非無作用差別, 以有性類有無
量種 … 是故現在過去未來三種有性, 條然差別. 寧如現在, 去來亦然? 依有, 可
言有未生滅; 約所無故, 未生滅成. 謂: 於有中先闕作用, 彼未有故, 名未已生. 有
法後時復闕作用, 彼已無故, 名為已滅. 故唯有中, 有未生滅. 由斯建立三世理成”.
41 Pradhan: 1975: 298: “svabhāvaḥ sarvadā cāsti bhāvo nityaśca neṣyate | na ca svabhāvād
bhāvo ’nyo vyaktam īśvaraceṣṭitam |”.
(tixiang 體相; *svarūpa, *svabhāva), how can there be difference in

respect to existence (you 有; astitva) and mode of being (xing[lei] 性
[類]; bhāva)?
[Saṃghabhadra:] – Isn’t it observed that there are dharmas co-existing
simultaneously, whose essential natures do not vary, but whose exis-
tences and modes of being differ? For instance, the elements earth (dijie
地界), etc., differ as being internal and external; sensations (vedanā; shou
受), etc., differ as being [sensations] of oneself and [sensations] of others,
and as being pleasurable, etc.
[For a given entity], this mode of being and existence cannot be different.
When its mode of being changes, it necessarily exists differently. It is on
this account that earth, etc., while identical in respect to essential nature,
can be said to differ in respect to their modes of being – as internal and
external. [Likewise], sensations, etc., while identical in respect to their
essential nature of being experience (anubhava), can be said to be differ-
ent modes of being, pleasure, etc. Again, it is like the eye, etc., within the
same serial continuity (i.e. same sentient being), which, while identical
in respect to their essential nature of being derived matter of tranquility
(rūpaprasāda), differ among them as distinct species of existence42 –
since their efficacies (gongneng; 功能) of seeing, hearing, etc., are distinct.
It is not the case that, therein, the efficacy differs from the existence, so
that there can be the different efficacies, such as seeing, etc.43 Rather, the
efficacy of seeing, etc., is none other than the existence of the eye, etc. A
distinct efficacy necessarily implies a distinct existence.
Thus, we know that there are dharmas which, while existing simultane-
ously and not differing in respect to essential nature, [nonetheless] differ
in respect to mode of being.
Since it is seen that there are dharmas existing simultaneously, which
while not varying in respect to their essential natures, differ in respect to
their modes of being, we know that when a dharma is traversing the
three periods of time, while not varying in respect to essential nature, it
has different modes of being. In this way, the Abhidharma tenet comes to
be well established.44
42 you xinglei bie 有性類別. In Xuanzang’s translation of the Abhidharmakośabhāṣya, youx-

ing 有性 corresponds to both astitva and bhāva.
43 Considering the immediately following sentence, jian deng gongneng 見等功能, I believe
that xing 性 in 性等功能差別 is a script error for 見.
44 T.29.1562: 625, a19-b2: “諸有為法歷三世時, 體相無差, 有性寧別? 豈不現見, 有法同
時, 體相無差, 而有性別? 如地界等, 內外性殊; 受等, 自他, 樂等, 性別. 此性與有,
238 Dhammajoti
5 The So-called “Neo-Sarvāstivāda” Doctrines of Saṃghabhadra
This leads to our discussion of what are commonly called the “neo-Sarvāstivāda”
doctrines of Saṃghabhadra. Such a conception is very much influenced by
Vasubandhu’s exposition of the Sarvāstivāda-Vaibhāṣika doctrines, and often
carries the implication that these supposedly new or modified Sarvāstivāda
doctrines represent innovative views not held by the tradition prior to the AKB
– and generally attributed to Saṃghabhadra.
One example is the claim that under the criticism in the AKB, the Vaibhāṣikas
came to assert the simultaneous operation of examination (vitarka) and scru-
tiny (vicāra) (also so explained in the Ny). But, in actual fact, such an
explanation has long been presented in the MVŚ.45 Other similarly claimed
views include the above-mentioned definition of activity (kāritra);46 the defi-
nition of the cessation not due to deliberation (apratisaṃkhyānirodha);47 the
simultaneous operation of the three primary characteristics (saṃskṛtalakṣaṇa)
– enduring (sthiti), deterioration (jarā), passing away (vyaya); the contrast
between hehe 和合 and heji 和集; etc. Most of these doctrines have been dis-
cussed elsewhere. As to the simultaneous operation of the three lakṣaṇas:
Kuiji and others describe it as a neo-Sarvāstivāda doctrine of Saṃghabhadra.48
But it is in fact already found in the MVŚ.49
To save space, I shall here only discuss the last example, i.e. the contrast
between hehe and heji, which will also serve to further illustrate how
理定無差. 性既有殊, 有必有別. 由是: 地等, 體相雖同, 而可說為內外性別; 受等,

領等體相雖同, 而可說為樂等性別. 又如眼等, 在一相續, 清淨所造色體相同, 而
於其中, 有性類別; 以見聞等功能別故. 非於此中, 功能異有, 可有(見)等功能差
別. 然見等功能, 即眼等有; 由功能別故, 有性定別. 故知諸法, 有同一時, 體相無
差, 有性類別. 既現見有法體同時, 體相無差, 有性類別; 故知諸法歷三世時, 體相
無差, 有性類別. 如是善立對法義宗”.
45 See Dhammajoti, 2008: 9–11.
46 See further discussion in Dhammajoti, 2009: 126–134.
47 See Dhammajoti, 2009: 485–491.
48 E.g. T.43.1830: 284c27–285a2: “此上, 古薩婆多師, 四相用違, 前後別起故, 為此難. 正
理論師, 為救此義, 復別解云: 三相用俱一時, 所望別故 … 子段第三, 新宗背古
難, 論主非之”; T.43.1832: 708b29-c2: “ 論: 又生等相若體俱有等者 – … 初破古師,
相用前後. 又住異滅下, 破新薩婆多, 三相用俱”.
49 T.27.1545: 200a3–12: “或有執: 三有為相非一剎那; 如譬喻者. 彼作是說: 若一剎那有
三相者, 則應一法, 一時亦生亦老亦滅. 然無此理, 互相違故. 應說: 諸法初起名
生, 後盡名滅, 中熟名老. 為遮彼執, 顯一剎那具有三相. 問: 若如是者, 則應一法,
一時亦生亦老亦滅. 答: 作用時異, 故不相違. 謂: 法生時, 生有作用. 滅時, 老滅方
有作用. 體雖同時, 用有先後. 一法生滅作用究竟, 名一剎那, 故無有失”.
Saṃghabhadra’s explanations had been taken seriously by both the ancient

Abhidharma as well as Yogācāra masters.
On the nature of the sensory objects, Saṃghabhadra argues against Śrīlāta,
maintaining that a sensory perception necessarily takes a real (and not a con-
ceptual) cognitive object (ālambana). This he describes as a physical collection/
collocation of atoms, rendered by Xuanzang as heji.
In contrast, denying the possibility of co-occuring causality, Śrīlāta claims
that the cognitive object of a sensory perception is necessarily past and con-
sists of a conceptualized whole of the atoms, rendered by Xuanzang as hehe.
The same theory of Saṃghabhadra is cited verbatim in Sthiramati’s
Tattvārthā, though not refuted. It is likewise cited, but then also refuted, in his
Triṃśikāvijñaptibhāṣya and in Xuanzang’s Cheng weishi lun (CWSL). Kuiji (and
others), calls this a “neo-Sarvāstivāda” theory, attributed to Saṃghabhadra.
All these texts use the same term heji, as in the Ny, to describe the “physical
collection” theory. Although the corresponding TVB passage does not spell out
the theory as being Saṃghabhadra’s, the equivalence in content and near iden-
tical wording in Xuanzang’s rendering of the corresponding sentences in the
Ny and the CWSL passages clearly indicate that all the three texts are here
describing Saṃghabhadra’s theory. And noteworthily, we see that there is no
Sanskrit equivalent to the term heji in the TVB.
It appears probable, as suggested by both Louis de La Vallée Poussin50 and
Yin Shun,51 that Xuanzang translated the same Sanskrit term (probably
saṃcita) as both hehe (和合) and heji (和集).
Kuiji’s way of glossing heji in his commentary on the CWSL seems more
Chinese than Indian, being based entirely on the compounded Chinese term,52
and may well reinforce the probability that the interpretation of Saṃghabhadra’s
theory in terms of heji by the Xuanzang tradition is indeed influenced by
Xuanzang’s rendering the same Sanskrit term *saṃcita as heji in the case of
Saṃghabhadra’s theory – in addition to rendering it as hehe in the case of
Dārṣṭāntika-Sautrāntika theory – and then consistently stressing it as the key
term in explaining Saṃghabhadra’s position in the Ny and the CWSL. In this
50 de La Vallée Poussin, 1928: 44 (under c. Doctrines des Sarvāstivādins).

51 Study, 704f.
52 T.31.1585: 271a18–21: “he 和 means being collocated (‘being in one place’), juxtaposed; ji 集
means not being [unified into] a single entity. Because these very juxtaposed [atoms] are
distinct entities, and because they truly exist, they are capable of generating conscious-
ness. Because they are in a gross form, and because the consciousness bears this form,
they fulfill the requirements of being a condition qua cognitive object (ālambanapratyaya)”.
(一處相近名和; 不為一體名集. 即是相近體各別故, 是實法故, 有力生識. 以相麁
故, 識有此相故, 所緣緣理具足有).
240 Dhammajoti
way, the impression is created that Saṃghabhadra innovated a new doctrine

with the new key term, heji – when in fact the MVŚ already has spoken of a
physical collocation of atoms as constituting a sensory object.
Accordingly, we may summarize and restate the heji vs heji controversy in
terms of the same Sanskrit term, saṃcita, thus:
Śrīlāta claims that an individual atom in itself does not have the efficacy of
generating a perception. Only when saṃcita (conceptually collected together),
are they capable of being cognized as a whole unity by consciousness. Since
this saṃcita state, forming a unified whole, is a mental construction (an ākāra),
and since it arises only in the second moment in the perceptual process – at
which time the atoms have become past and are no more existent – the cogni-
tive object is necessarily a non-existent.
In contrast, Saṃghabhadra holds that each individual atom in itself is actu-
ally efficacious. But it is only when they are saṃcita (physically assembled
together) that their combined efficacy becomes strong enough to generate a
cognition. However, this saṃcita state is not a conceptualized whole but a
physical collocation or aggregation of the atoms. Moreover, since, from the
Sarvāstivāda-Vaibhāṣika perspective, this takes place in the very first moment,
thanks to co-occurring causality, the cognitive object is a real existent.53
Thus, it is more appropriate to understand Saṃghabhadra’s heji explanation
as essentially a fine-tuning of the Sarvāstivāda theory, rather than labeling it as
being “neo-Sarvāstivāda”. The same remark applies to other – similarly labeled
– explanations of Saṃghabhadra. Or rather: the “new” element in such expla-
nations consists essentially of his fine-tuning and further articulation.
Moreover, in such examples as the exposition of sarvāstitva in terms of
bhāvānyathātva, we may do better to be more aware of the continuous doctri-
nal developments of the Vaibhāṣika masters since the time of the MVŚ.
6 *Nyāyānusāra as a Source for Mahāyāna Studies
The gigantic MVŚ – whose compilation must have spanned over decades
and finally completed as early as circa the middle of the second century CE
– records Abhidharma doctrines which are extremely valuable for the study
of the emergence of the Mahāyāna. These include: the doctrine of the three
vehicles (yānatraya: śrāvaka-, pratyeka- and buddhayāna); of the Buddha
alone possessing perfect wisdom (sarvākārajñāna) and great compassion
(mahākaruṇā) whose intrinsic nature is understanding (prajñā); of the
53 Cf. KL Dhammajoti, 2012: 216ff.

Buddha’s fulfillment of the six pāramitās, of him alone being totally free from
contamination (kleśa) along with all habitual residues (vāsānā); of (contam-
ination-hindrance) kleśāvaraṇa and knowable-hindrance (jñeyāvaraṇa); of
the Buddha “journeying in the five planes of existence together with Ānanda,
incessantly day and night benefitting sentient beings”;54 etc.55
This text also provides considerable information on the early Dārṣṭāntikas
and the Sarvāstivādin Yogācāra masters, who are important for our under-
standing of the emerging Mahāyāna movement.
The Ny contains further elaboration, additions and articulation of many of
these doctrines, and in this way adds to the value of the Abhidharma textual
tradition as an indispensable source for the study of the doctrinal evolution of
the Mahāyāna.
An important instance is Saṃghabhadra’s elaborate discussion on the non-
defiled nescience (akliṣṭājñāna) and habitual residue (vāsanā) (see 2.1 above).
Such buddhological notions have influenced the whole of the Mahāyāna
development, particularly the Yogācāra.56 They also partly contributed to some
later Mahāyāna conceptions, such as that of the avidyāvāsabhūmi and the two
types of death – parichinnā cyuti and pariṇāmikī cyuti. This two-death doctrine
originated in the tathāgagarbha school of thought; but also came to be taken
in by the Yogācāra57 and the *Mahāprajñāpāramitopadeśa (Da zhidu lun 大智
度論)58 in their own ways.
We have already noted above the doctrinal importance of the doctrine of
co-occurring causality for the Yogācāras. Saṃghabhadra’s systematic and very
detailed exposition and brilliant defence of this doctrine in the Ny can be stud-
ied alongside with the Yogācāra discussion predating and post-dating the Ny
– for example, Asaṅga’s *Mahāyānasaṃgraha (and commentaries thereon by
Vasubandhu and Asvabhāva) and the CWSL, both of which claim this causal
category as the only valid basis for the establishment of causality.59 Such a
comparative study would yield useful information on the doctrinal interaction
between the Abhidharma tradition and the Yogācāra in this milieu. The same
remark applies to discussions and arguments in the Ny on the doctrines of the
Sautrāntika, the Vibhajyavāda and the Mahāsāmghika (such as the intrinsic
54 T.27.1545: 159c28–29: “… 或將阿難遊歷五趣, 晝夜無間, 饒益有情.”. Also, T.27.1545:

428c9–12: “佛世尊為眾生故, 捨尊貴位. 或作陶師, 或作力士, 或作樂人, 或作獵師,
或作婬女, 或作乞人, 或引難陀遍遊五趣”.
55 Cf. Dhammajoti, 2011a.
56 Cf. Dhammajoti, 1998.
57 See T.31. 1585: 45a11–28.
58 See T.25.1509: 714a9–15.
59 Cf. T.31.1594: 134c15–20.
242 Dhammajoti
purity of thought), three schools known to have significantly contributed to

the Mahāyāna development.
Śrīlāta’s anudhātu doctrine detailed in the Ny (mentioned above) also mer-
its special attention for its possible relation to the Yogācāra bīja and ālayavijñāna
theories – even though it is based on a model of successive arising of citta-
caittas. We learn from the Ny that Śrīlāta’s exposition of santatipariṇāmaviśeṣa
based on the present-only-exist stance leads to the distinction between “retri-
bution-born” (vipākaja) and retribution-fruit (vipākaphala), comparable in
some way to the Yogācāra distinction of the two. For Śrīlāta, all is retribution-
born; no dharma is classifiable as being emanational (naiḥṣyandika) or
cumulative (aupacayika).60
Other expositions in the Ny that are noteworthy in this context include: (1)
the seed of the impure (anāsrava) dharmas61 – including Rāma’s doctrine of
the “white-dharma perfuming (bai fa xiqi 白法習氣; *śukladharmavāsanā)”,62
which Yin Shun63 believes is related to the doctrine of the anāsravadharmas
arising from the perfuming (which are sāsrava) of the listening to the
true dharma, as taught in the *Mahāyānasaṃgraha and suggested in the
Viniścayasaṃgrahaṇī of the Yogācārabhūmi; (2) Saṃghabhadra’s questioning
of the validity of the Dārṣṭāntika-Sautrāntika notions of perfuming; (3) the
Dārṣṭāntika doctrine of the unreality of the external object; etc. For the last
doctrine mentioned, in addition to our discussion above, we may also note the
interesting Dārṣṭāntika argument for this unreality from the Buddhist teaching
that beings born into different planes of existence (gati) cognize purity and
impurity differently (e.g. with regard to water)64 – a commonly known argu-
ment in the Yogācāra texts for vijñaptimātratā.
60 T.29.1562: 359a8–14. See also Dhammajoti, 2011a.

61 See, for instance, T.29.1562: 421a16-b18, 712c18–713b25 (this contains a long controversy).
62 T.29.1562: 502b13–19: “大德邏摩作如是說: 有不染法, 名為習氣; 如不善因所招異
熟. 世尊昔在菩薩位中, 三無數劫修諸加行, 雖有煩惱, 而能漸除煩惱所引不染
習氣; 白法習氣漸令增長. 後於永斷諸漏得時, 前諸習氣有滅不滅. 以於長時修
加行故, 證得無上諸漏永盡; 然佛猶有白法習氣, 言習氣有滅不滅故”.
63 Yin Shun, 1974: 198.
64 T.29.1562: 639b4–10: “譬喻部師作如是說: 由分別力, 苦樂生故, 知諸境界體不成實.
以佛於彼摩建地迦契經中說: 諸癩病者觸苦火時, 以為樂故. … 又如淨穢不成實
故, 謂: 別生趣同分有情, 於一事中, 取淨穢異. 既淨穢相非定可得, 故無成實淨穢
二境”.
7 Comparative Studies between Abhidhamma and Abhidharma
Finally, I would like to briefly remark on the importance of comparative stud-

ies between Theravāda Abhidhamma and the Abhidharma tradition (including
the Dārṣṭāntika-Sautrāntikas).
It has often been assumed that Theravāda Abhidhamma developed – since
the introduction of Buddhism into Sri Lanka – in isolation from the Indian
continental tradition. But, as is well known, the northern Abhidharma and
Yogācāra texts are not entirely ignorant of the Theravāda development.
Thus, they know the latter’s theory of the hṛdayavastu as the mana-indriya
(/manas) or manodhātu.65 Their bhavāṅgavijñāna doctrine is cited by
*Karmasiddhiprakaraṇa and Asaṅga’s *Mahāyānasaṃgraha, etc., as compa-
rable to the ālayavijñāna doctrine.
In the Theravāda tradition, it is well known that Ācariya Dhammapāla often
offers commentarial explanations that reveal familiarity and affinity with the
northern tradition. For instance, his understanding of the distinction between
kilesa and vāsanā and the notion of ñeyyāvaraṇa are very similar to those in
the northern Abhidharma tradition.
Such familiarity is even more conspicuous in the ṭīkā and anuṭīkā litera-
ture. Thus, the Majjhimaṭīkā ascribes certain views to the “sabbatthivādino”.66
The Pañcapakaraṇa-anuṭīkā67 speaks of the four types of sabbatthivādā:
the bhāvaññattikā, the lakkhaṇañnattikā, the avatthaññattikā and the
aññathaññattikā.
It could even be said that, to some extent, the Theravāda Abhidhamma
development cannot be satisfactorily understood without at least a basic
understanding of the possible influence from the broad Sarvāstivāda lineage
– including the Sautrāntika tenets. The following are a few observations in this
connection:
(1) Theory of moments:68 In the canonical Abhidhamma texts, there is

reference only to two moments, moment of origination (uppā
dakkhana) and moment of cessation (nirodhakkhana). There is no
evidence to suggest that these canonical texts recognized a static
phase, a moment of duration. However, perhaps due to the influence
65 Wogihara, 1971: 39.

66 E.g. Majjhima-ṭīkā (Burmese edn) II, 360 (= AAṬ (Burmese edn) II, 112).
67 Cf., Pañcapakaraṇa-anuṭīkā (Burmese edn), 94–97.
68 Cf. Karunadasa, 2010: 252–267.
244 Dhammajoti
of the Sarvāstivādins, the Abhidhamma commentarial exegesis intro-

duces a moment of duration (ṭhitikkhana).
(2) Denial of motion: It is mostly in the Abhidhamma sub-commentar-
ies and compendiums that the Theravādins maintain that since
rūpadhammas are momentary they cannot move from one locus in
space to another (desantarasaṅkamana). What really takes place is
the successive arising of momentary rūpas of a given series in adja-
cent locations (desantaruppatti). Yakupitiyage Karunadasa thinks it
likely that this “is a theory adopted by the Theravādins from Sanskrit
Buddhism”.69
(3) It is a well-known fact that the Sautrāntikas were more aware of the
subjectivity of phenomena than the Sarvāstivādins. Therefore the
tendency among them for reification is very much less evident than
among the Sarvāstivādins. We find this trend surfacing in the
Theravāda Abhidhamma too, most probably due to Sautrāntika
influence:
– In the Dhammasaṅgaṇi, rūpāyatana is defined as consisting of
both color (vaṇṇa) and shape (santhāna). However, in the
Atthasālinī, it is observed that only color constitutes the
rūpāyatana, because shape is a mental construct superimposed
on the difference of coloration.70
– In the commentaries, the 28 rūpadhammas are divided into two
categories, 18 as nipphanna and the remaining 10 as anipphanna.
The purpose of this division is to show that what constitutes the
latter category are not true rūpadhammas, although they are so
called. What this amounts to is a reduction of the number of
rūpadhammas by 10 items.71
(4) In the case of the canonical texts, it is true that they developed more
or less separately in the Theravāda and the Sarvāstivāda traditions.
But given the fact that they both evolved out of the Sthaviravāda lin-
eage, it should be possible to trace at least some of their common
doctrinal inheritance and detect certain ancient elements of mutual
influences, which had taken place in India, through a comparative
study – especially since these canonical texts have been well pre-
served in the two traditions. Yin Shun is one of the very few modern
scholars who have made some effort to do this, basing himself par-
ticularly on the *Śāriputrābhidharma (Shelifo apitan lun 舍利佛阿毘
69 Cf. Karunadasa, 2010: 260f.

70 Cf. Karunadasa, 2010: 176–182.
71 Cf. Karunadasa, 2010: 152–160.
曇論), which is commonly taken as the fundamental Abhidharma

text by the Vātsīputrīya (and its branches) and the continental
Vibhajyavāda sects. Such an investigation is all the more meaningful
considering that the Theravāda is a branch of the Vibhajyavāda.
Analyzing the Śāriputrābhidharma and comparing it with the Pāli
Vibhaṅga and the Sarvāstivādin Dharmaskandha, Yin Shun discovers
12 topics common to them, and deduces that these topics – compris-
ing the 37 bodhipakṣya-dharmas (centered on the āryamārga),
skandhas, dhātus, āyatanas, the four truths and pratītyasamutpāda,
etc. – constitute the primitive form of the Sthaviravāda Abhidharma.72
8 Conclusion
Even though space is limited here, I hope I have been able, from various per-
spectives, to further clarify the importance and need for studying the Ny
beyond the AKB and the Vy. But daring to be more concrete and more specific:
Within our modern world where English functions as the lingua franca, it
would be a great step forward in the advancement of Abhidharma studies, if a
team of scholars could jointly devote some effort to the systematic translation
of the Ny into the English language.
Abbreviations
AKB Abhidharmakośabhāṣya (ed. Pradhan, 1975)

CWSL Cheng weishi lun
MVŚ Abhidharmamahāvibhāṣā
Ny *Nyāyānusāra
T. Taishō shinshū daizōkyō 大正新脩大藏經 (ed. Takakusu 高楠, Watanabe 渡
邊 and Ono 小野, 1924–1934)
TVB Triṃśikāvijñaptibhāṣya
Vy Sphuṭārthā Abhidharmakośavyākhyā
72 Yin Shun, 1968: 64–72, 125.

246 Dhammajoti
Bibliography
Dhammajoti, KL (1998). “The Defects in the Arhat’s Enlightenment: His akliṣṭājñāna

and vāsanā”, Bukkyō Kenkyū XXVII, pp. 65–98.
Dhammajoti, KL (2007a). “Ākāra and Direct Perception (pratyakṣa)”, Pacific World
Journal 9, pp. 245–272.
Dhammajoti, KL (2007b). Abhidharma Doctrines and Controversies on Perception. Hong
Kong: The University of Hong Kong.
Dhammajoti, KL (2008). Entrance into the Supreme Doctrine. Hong Kong: The University
of Hong Kong.
Dhammajoti, KL (2009). Sarvāstivāda Abhidharma. Hong Kong: The University of Hong
Kong.
Dhammajoti, KL (2011a). “From Abhidharma to Mahāyāna: Remarks on the early
Abhidharma doctrine of the three yāna‑s”, Journal of Buddhist Studies IX, pp. 153–169.
Dhammajoti, KL (2011b) “Śrīlāta’s anudhātu doctrine”, Bukkyō Kenkyū XXXIX,
pp. 49–57.
Dhammajoti, KL (2012). “Abhidharma Debate on the Nature of Sensory Perception”,
Journal of Buddhist Studies X.
Dhammajoti (2013). “Abhidharma Debate on the Nature of the Objects of Sensory
Perception”, JCBSSL X, pp. 203–234.
Jaini, Padmanabh S. (ed.) (1959). Abhidharmadīpa with Vibhāṣāprabhāvṛtti. (Tibetan
Sanskrit Works Series 4). Patna: Kashi Prasad Jayaswal Research Institute.
Karunadasa, Yakupityage (2010). The Theravāda Abhidhamma. Hong Kong: Hong Kong
University Press.
La Vallée Poussin, Louis de (1928). Vijñaptimātratāsiddhi – La Siddhi de Hiuan-tsang.
2 vols. Paris: Librairie Orientaliste Paul Geuthner.
La Vallée Poussin, Louis de (1971). L’Abhidharmakośa de Vasubandhu. Traduction et an-
notations. 6 vols. Bruxelles: Institut Belge des Hautes Études Chinoises.
Lodrö Sangpo, Gelong (2012). Abhidharmakośa-Bhāṣya of Vasubandhu. Translated into
French by Louis de La Vallée Poussin. Annotated English Translation by Gelong Lodrö
Sangpo. With a New Introduction by Bhikkhu KL Dhammajoti. 4 vols. Delhi: Motilal
Banarsidass.
Majjhima-ṭīkā (Burmese edn) II, 360 (= AAṬ (Burmese edn) II, 112). Mahāmāluṅkyasutta,
Majjhima I, 433.
Pradhan, Prahallad (ed.) (1975). Abhidharmakośabhāṣya of Vasubandhu. (Tibetan
Sanskrit Works Series 8). Patna: Kashi Prasad Jayaswal Research Institute.
- T.25.1509: Nāgārjuna. Mahāprajñāpāramitopadeśa. 大智度論, Kumārajīva.

- T.29.1562: Saṃghabhadra. [Abhidharma]Nyāyānusāraśāstra, Apidamo shun zhengli
lun 阿毘達磨順正理論, Xuanzang 玄奘.
- T.31.1585: Dharmapāla. Vijñaptimātrāsiddhi[śāstra], Cheng weishi lun 成唯識論,
Xuanzang 玄奘.
- T.41.1821: Puguang 普光, Jushe lun ji 俱舍論記.
- T.42.1828: Dunlun 遁倫, Yujia lun ji 瑜伽論記.
- T.43.1829: Kuiji 窺基. Yujia lun lüezuan 瑜伽論略纂.
- T.43.1830: Kuiji 窺基, Cheng weishi lun shuji 成唯識論述記.
- T.43.1832: Huizhao 慧沼, Cheng weishi lun liaoyi deng 成唯識論了義燈.
- T.43.1834: Kuiji 窺基, Weishi ershi lun shuji 唯識二十論述記.
Wogihara, Unrai (ed.) (1971). Sphuṭārtā Abhidharmakośavākhyā by Yaśomitra. 2 vols.
Tokyo: Anastatic reproduction of the 1933–1936 edition.
Yin Shun (1974). Weishixue danyuan 唯識學探源. Taipei: Yinshun wenjiao jijinhui
(reprint).
Yin Shun (1968). Shuo yiqieyou bu wei zhu lunshu yu lunshi zhi yanjiu 說一切有部為主
論書與論師之硏究. Taipei: Yinshun wenjiao jijinhui.
248 Greene
Chapter 7
Pratītyasamutpāda in the Translations of An Shigao

and the Writings of His Chinese Followers
Eric M. Greene
1 Introduction
It goes without saying that pratītyasamutpāda – “dependent origination” – is

one of the most important technical formulations of doctrinal Buddhism.1
According to some early sources, pratītyasamutpāda indeed encapsulates the
whole of the Buddha’s teaching, such that “one who sees dependent origination
sees the Dharma, and one who sees the Dharma sees dependent origination”.2
It is thus not surprising that pratītyasamutpāda figures prominently in
the works of the earliest known translator of Buddhist texts into Chinese,
An Shigao 安世高 (fl. 148–168 CE),3 whose oeuvre focused precisely on the
kinds of Āgama, early commentarial, and proto-Abhidharma literature where
pratītyasamutpāda plays such an important role. An Shigao’s translations can
be dated with reasonable precision, and they are thus important source mate-
rials for Indian Buddhism during the early years of the common era, a moment
in Indian Buddhist intellectual history for which we have little other direct
evidence. Since the Abhidharma itself was a project centered on the precise
and systematic definition and explication of the important technical termi-
nology found in the sūtras, An Shigao’s varied and often surprising Chinese
translations of this terminology (such as the pratītyasamutpāda formula) may
in some cases give us a window onto the currents of Indian scholastic thought
1 On pratītyasamutpāda, the classic study of Louis de La Vallée Poussin, 1913 remains one of the
best. See also Johansson, 1979; Nakamura, 1980; Stalker, 1987: 47–96; Cox, 1993; Bucknell, 1999;
Jurewicz, 2000; Shulman, 2008.
2 Yo paṭiccasamuppādaṃ passati so dhammaṃ passati, yo dhammaṃ passati so
paṭiccasamuppādaṃ passatīti (MN, 1:190–191).
3 Little is known about An Shigao himself (see Forte, 1995 for some interesting theories) or his
translation methods. He may have simply been the reciter of the Indian texts, as was the
normal role of the non-Chinese “translators” in later times (Zürcher, 1972: 31). For the sake of
convenience, I shall refer to “An Shigao” as an individual person, but we should remember
that the works associated with this name may have had a collective authorship.

Pratītyasamutpāda in the Translations of An Shigao 249
that were, in this same period of time, leading to the formation of the various
Abhidharma collections.
An Shigao’s translations are equally important – if so far underutilized by
scholars – from the perspective of Chinese Buddhism, as they allow us, in
some cases at least, to glimpse how key Buddhist ideas were initially under-
stood in China. Such questions can be asked, of course, about a broad range
of topics. However focusing our attention on a fixed list of terms such as the
pratītyasamutpāda formula allows us to be certain of the underlying Indic
vocabulary that An Shigao was translating, something difficult to achieve in
other cases.4 With this as a point of reference, we can then begin to observe
notable features of the Chinese terms An Shigao selected in his transla-
tions, as well as of the interpretations of these terms by subsequent Chinese
commentators.
As I will show, An Shigao had particular difficultly finding a suitable Chinese
translation for the term sparśa, “sense contact,” a concept that did not have a
ready equivalent within pre-Buddhist Chinese theories of sense perception.
The subsequent efforts of commentators to grasp this topic, which occupies a
central place within pratītyasamutpāda, shows that sense perception was an
area where pre-existing Chinese theories and technical vocabulary strongly
shaped the Chinese interpretation of Buddhist ideas.
2 The Terms
It will be convenient to first present a table with all known examples of the
members of the pratītyasamutpāda formula within the surviving translations
of An Shigao.5 In compiling this information, I have limited myself to the con-
4 Indic parallels have been identified for some of An Shigao’s translations (Harrison, 1997;
Harrison, 2002; Zacchetti, 2002). When possible I have used these to help establish An Shigao’s
usage (for a complete survey of known parallels to An Shigao’s translations, see Nattier, 2008:
35–72). But even without such parallels, it is usually possible to identify the pratītyasamutpāda
formula, safely allowing us to align the translated terms with the items in this list.
5 For the purposes of this chapter, “An Shigao’s translations” are those given by Nattier, 2008
and Zacchetti, 2010a. I cannot here address the important question of the reliability of the
received texts of An Shigao’s translations. Though there is no reason to doubt them on the
whole, there is some evidence that, within early Chinese Buddhist texts in general, later copy-
ists or editors occasionally replaced unusual technical terms with more familiar ones. Thus
in the Xiuxing benqi jing 修行本起經 (possibly a third-century text; Nattier, 2008: 104–109),
the second Kǒryo edition (the main Taishō text) gives shou 受 as the translation of upādāna,
which is standard in early translations (T.3.184: 470c2). However the Sixi and Puning editions
250 Greene
1. avidyā (“ignorance”) chi 癡 “confusion”

2. saṃskāra (“volitional formations”) xing 行 “activity”
3. vijñāna (“consciousness”) shi 識 “consciousness”
4. nāmarūpa (“name-and-form”) a. ming zi 名字a “name and style”
b. ming zi se 名字色 b “name and appearance”
5. ṣaḍāyatana (“six sense gates”) a. liu ru 六入 “the six entrances”c
b. liu shuai 六衰 “the six destroyers”d
c. liu rushou 六入受 “the six entrance-graspers”e
6. sparśa (“contact”) a. geng 更 f “to suffer”g
b. gengshou 更受h “suffering and grasping”
c. zhi 致 i “reaching”
d. si 思 j “longing”
e. siwang 思望 k “distantly longing”
f. zai 栽 l “planting” (or “a seedling”)
7. vedanā (“sensation”) a. tong 痛 m “pain”
b. tong yang 痛痒 n “pain and pleasure”
8. tṛṣṇā (“craving”) ai 愛 “craving”
9. upādāna (“grasping”) a. shou 受 “grasping”o
b. qiu 求 “seeking”p
10. bhava (“becoming”) you 有 “existence”
11. jāti (“birth”) sheng 生 “life”
12. jarāmaraṇa (“old-age and death”) lao si 老死 “old-age and death”
a Ren ben yu sheng jing 人本欲生經, T.1.14: 243b6 and passim. There is also a single case within this
text of the eventually standard translation ming se 名色 (T.1.14: 243b19); this may be a later
hyper-correction.
b Chang ahan shi bao fa jing 長阿含十報法經, T.1.13: 240a24–25 (reading 名字色 for 名字苦).
c T.1.13: 240a25; YCRJ, T.15.603: 174b25 (and throughout).
d T.2.150A: 876b20.
e YCRJ, T.15.603: 174c13–14; here followed by the more usual liu ru 六入. Note the similar ru shou 入
受 in the Dao di jing 道地經 (T.15.607: 231c5). On the prevalence of reversible compounds in Han-
dynasty Buddhist translations, see Zürcher, 1979: 180.
f T.1.14: 243a27-b3 (and throughout).
g On this meaning of geng 更, see below note 26.
h T.1.13: 240a25 (de Jong, 1979: 272).
i YCRJ, T.15.603: 174b25, b30.
j T.1.57: 853a3.
k T.1.57: 852a25; YCRJ, T.15.603: 174c14–17.
l T.2.150A: 875c1 and passim (see Vetter and Harrison, 1998: 205).
m YCRJ, T.15.603: 174c16 (and passim); T.1.13: 240a25; T.1.14: 243b4 and passim; T.1.57: 852a20–22 and
passim.
n YCRJ, T.15.603: 174b25.
o This is An Shigao’s standard translation (e.g. YCRJ, T.15.603: 175a13–14).
p T.1.14: 242a14–15 and passim.
text of pratītyasamutpāda itself or where pratītyasamutpāda is implicitly

invoked,6 as it is often only through their appearance in a standardized list
such as this that we can be certain, or nearly certain, of the words An Shigao
was actually translating. Exploring the different ways these same terms were
handled in other contexts would no doubt shed additional light on the issues
raised here.
For links 1–3, 8, and 10–12, An Shigao always used the same translation, and
I have not provided explicit references to all such instances. In the left hand
column of the table, I give the list of terms in their Sanskrit forms and a typical
English translation of them.7 In the right hand column, I give the Chinese
term(s) and an attempted English translation. In many cases the intended
meanings of the Chinese words is not clear, and these English translations are
not intended to be definitive.
Along with those included in the table, we should also consider certain
terms that appear only in the Ahan koujie shi’er yinyuan jing 阿含口解十二因
緣經 (T.25.1508; AHKJ hereafter). As Stefano Zacchetti has argued, this text is
most likely an oral teaching given by An Shigao, perhaps originally as a com-
mentary to his now-lost translation the Shi’er yinyuan jing 十二因緣經, which
judging from its title must have been devoted primarily to pratītyasamutpāda.6
Within the AHKJ we find many of the same translations for the links of
pratītyasamutpāda already given in the table.8 For others, we have but slight
modifications of more usual translations.9 Thus for saṃskāra we find not
simply “activity (xing 行), but “activity that is performed” (suo zuo xing 所作
(Song 宋 and Yuan 元 in the Taishō footnotes) give qu 取, which later replaced shou 受 as
the normal translation of upādāna. At some point an editor thus “updated” some copies of
this text. Although in this case the (presumably) correct reading survives in the second Kǒryo
edition, this raises the possibility that in some texts archaic terminology may have been re-
placed by later copyists and editors. This possibility raises important questions, especially
since scholars have often relied upon the translations of technical terms as markers of the
authenticity and likely date of early Chinese Buddhist translations.
6 In particular I include the Lou fen bu jing 漏分布經 (T.1.57) and the Qi chu san guan jing 七
處三觀經 (T.2.150A), both of which discuss “sensation” (vedanā) as originating in “contact”
(sparśa), thereby clearly invoking their relationship in pratītyasamutpāda.
7 The texts An Shigao translated were certainly not written in Sanskrit. For the sake of conve-
nience, however, I will use Sanskrit forms throughout, unless directly citing a Pāli or other
middle Indic text.
8 The translations used are not, however, always the most common ones. Thus the AHKJ uses
“seeking” (qiu 求) for upādāna (T.25.1508: 53a10), found in only a single other An Shigao trans-
lation (9b in the table).
9 On these terms from the AHKJ, see Zacchetti, 2004: 198, note 4.
252 Greene
行);10 for vijñāna not simply “consciousness” (shi 識) but “what one becomes
conscious of” (suo shi 所識);11 and similarly for sparśa “what is suffered” (suo
geng 所更) rather than simply “to suffer” (geng 更).12 We also find the seem-
ingly new “name and appearance” (ming se 名色) for nāmarūpa,13 though this
differs only slightly from the mingzi se 名字色 that An Shigao uses elsewhere
(4b in the table above).
A few of the new translations in the AHKJ are worthy of note, such as “appro-
priating” (de 得) for bhava.14 Several novel translations of saṃskāra – “seminal
factors of rebirth” (sheng si jing 生死精),15 and “seminal activities of rebirth”
(sheng si jing xing 生死精行)16 – are perhaps combinations of An Shigao’s
standard translation of saṃskāra in the case of pratītyasamutpāda, namely
xing 行 (“activities”), with the different translation he usually uses to render
saṃskāra when it is one of the five skandhas, namely shengsi 生死 (“[factors]
leading to rebirth”).17
The above table already allows us to draw some interesting conclusions
about the issues that An Shigao faced when translating the pratītyasamutpāda
formula. We may first notice the relative uniformity of the translations. Seven
of the twelve terms are always translated in exactly the same manner (one of
these, bhava, finds a variant only in the AHKJ). Three other terms occur with
only slight variations: nāmarūpa translated as both “name and style” (ming zi
名字) and “name and appearance” (ming zi se 名字色),18 and vedanā, usually
simply “pain” (tong 痛) but on one occasion “pain and pleasure” (tong yang 痛
痒),19 which is An Shigao’s normal translation for vedanā in the context of
10 T.25.1508: 53a8. Cf. xing zuo 行作 in Taizi ruiying benqi jing 太子瑞應本起經, T.3.185:
478b7.
11 T.25.1508: 53a9.
12 T.25.1508: 53a10. The use of suo 所 here is peculiar (as noted by Zacchetti, 2007: 412–413).
13 T.25.1508: 53a10.
14 T.25.1508: 53a12. This translation of bhava does not seem to appear outside this text.
15 T.25.1508: 53b20.
16 T.25.1508: 53b7.
17 An Shigao uses sheng si 生死 to translate saṃskāra only in the context of the skandhas,
never in pratītyasamutpāda, where we always find xing 行. However conversely, xing 行
does occasionally appear in the context of the skandhas (T.1.13: 234c24; T.15.607: 231b29-
c1).
18 The translation of nāmarūpa as ming zi 名字 is mysterious. As Vetter suggests, we are
perhaps meant to read ming 名 and zi 字 as opposed categories, “true name” versus “pub-
lic appearance” (Vetter, 2012: 55–56). Alternatively, perhaps An Shigao’s lists had only
nāma rather than nāmarūpa (as we see in, e.g., Vibh: 138–139).
19 Several second- or third-century texts associated with An Shigao explain tong yang 痛痒
(var. 痛癢) as “pain” and “pleasure” respectively (Da anban shouyi jing 大安般守意經,
either the five skandhas or the second smṛtyupasthāna.20 For the term upādāna,
we do find two different translations (“grasping” shou 受 and “seeking” qiu 求).
But even here, the meaning of the two words is rather similar.
Things become more interesting for link 5, the “six sense gates” (ṣaḍāyatana)
and link 6, “contact” (sparśa), two terms that are crucial for the Buddhist expla-
nation of sense perception. For āyatana, An Shigao used two quite distinct
Chinese expressions: the “six entrances” (liu ru 六入) and the “six destroyers”
(liu shuai 六衰). The first of these translations fits a common Indian Buddhist
explanation of the āyatanas as the gateways through which sense data arrive,21
an explanation that is expressed pictorially when the āyatanas are depicted as
the six doors or windows of a house in paintings of the so-called wheel of life
(bhavacakra).22
But the second of the two Chinese translations – the āyatana as the “six
destroyers” – is more mysterious. Certainly it is not a literal translation. Here
too, however, An Shigao may have been drawing on Indian commentarial tra-
ditions. Indeed, according to one well-known explanation, the six external
āyatanas (the six kinds of sense data) are like “village-raiding bandits,” “attack-
ers” of the sense organs that “destroy” them.23 The term liu shuai (“the six
T.15.602: 168a7–11; YCRJZ, T.33.1694: 11c2; SMJ, 297–298). This does seem like a good expla-
nation, as this would yield a typical Chinese abstract noun meaning “sensation”. However
some classical Chinese Buddhist lexicographers consistently explained yang 痒 to be syn-
onymous with tong 痛 (Yiqie jing yinyi 一切經音義, T.54.2128: 529a5). Modern dictionar-
ies too explain yang 痒 as only possibly meaning “sickness” or “unpleasant itching”
(GHYZD: 754; HYDZD: 2673; Morohashi: 8080). Furthermore this meaning is not necessar-
ily late, and we find examples from Han or Three Kingdoms period Buddhist texts where
tong yang 痛痒 must simply mean pain (Zhong benqi jing 中本起經, T.4.196: 152a8–10;
Bannihuan jing 般泥洹經, T.1.6: 186c6). It is not clear how we are to resolve this discrep-
ancy.
20 As he seems to have done with saṃskāra (see above note 17), An Shigao may have inten-
tionally distinguished vedanā in the context of pratītyasamutpāda from vedanā as a skan-
dha or a smṛtyupasthāna. With only one exception he uses tong 痛 in the former context
and tong yang 痛痒 in the later. Interestingly in that exceptional case (the YCRJ), a com-
parable distinction is still maintained, albeit reversed, with tong 痛 used for vedanā as a
skandha (T.15.603: 173b7) or a smṛtyupasthāna (T.15.603: 174a4), and tong yang 痛痒 in
pratītyasamutpāda (T.15.603: 174b25). It is unclear if there was here any intended differ-
ence in meaning. Given that the YCRJ preserves the distinction while reversing the terms,
the purpose may have been simply to mark the different contexts.
21 AKBh, 13.18
22 Hartmann, 1940: 359.
23 “The external [āyatanas] are to be likened to village-raiding bandits, because they are of
the nature to attack the internal [āyatanas].” Bāhirāni gāmaghātakacorā viya
254 Greene
destroyers”) might thus be called an “exegetical translation” – a translation in

which what is produced in Chinese is not the literal meaning of the word, but
rather a commonly accepted commentarial gloss.
Of course even if this was what An Shigao intended, we may still ask why he
used this kind of “exegetical” translation only in this particular case. Though
this is difficult to determine with any certainty, it is probably relevant that pre-
senting the sense objects as “destroyers” fits well with certain pre-Buddhist
Chinese understandings. As it is put in the Daodejing 道德經, the sense objects
harm the sense organs: “The five colors blind a person’s eyes […] the five flavors
injure a person’s mouth. The five tones deafen a person’s ears”.24 Though this is
perhaps the most famous passage in which such ideas are expressed, they are
found throughout pre-Buddhist Chinese literature and almost certainly would
have constituted a point of reference for any educated Chinese audience dur-
ing An Shigao’s time.
But whatever the origin of this imagery, more important is the mere fact
that in this case, unlike for most other terms in the pratītyasamutpāda for-
mula, An Shigao felt the need to go beyond a literal translation, to explain the
importance of the word rather than merely translate its meaning.
The second term in the above table where An Shigao’s translations arouse
our attention is sparśa, link 6. Here we notice first of all their sheer number. Six
different renderings are attested in An Shigao’s translations, plus one more in
the AHKJ. Some of these terms do overlap in meaning, such as “longing” (si 思)
and “distantly longing” (siwang 思望). But even accounting for this, there are
four completely distinct translations – “to suffer” (geng 更), “reaching” (zhi 致),
“longing” (si 思), and “planting” (zai 栽). And of these, only “reaching,” the least
common of them, is even plausibly a literal translation of the basic meaning of
sparśa, “touching”.
As we have seen already, An Shigao was often happy using more than one
translation for a given technical term, and he also may have occasionally used
what I have called “exegetical” translations that were informed by well-estab-
lished commentarial traditions. But even still, the number and diversity of An
Shigao’s translations of sparśa contrast dramatically with the relatively consis-
tent way he translated other key items in the pratītyasamutpāda formula (such
as, for example, tṛṣṇā), or even with his approach to the word āyatana, in
which a relatively literal translation is paired with a single more figurative one.
The diversity and range of meanings seen in An Shigao’s translations of
sparśa suggest that there was considerable debate, overtly or implicitly, about
ajjhattikānaṃ abhighātakattā. (Vism: 484). These images draw from a passage in

Saṃyuttanikāya (SN IV:172–175).
24 五色使人目盲 […] 五味使人之口爽，五音使人之耳聾 (Bo shu Laozi jiao zhu: 273).
how to best render this term into Chinese. On the one hand, this might indi-
cate that An Shigao attributed particular importance to this term – that for
him sparśa was the key concept within pratītyasamutpāda. But it is also pos-
sible that the notion of sparśa proved especially difficult for the Chinese to
grasp, and that the diversity of translations reflects a continuing struggle
to convey the meaning of this term. Indeed none of An Shigao’s choices for
sparśa ultimately proved satisfactory, and most were never used again by later
translators.25 And this, again, stands in sharp contrast with many of his other
translations from the pratītyasamutpāda formula – such as ai 愛 for tṛṣṇā and
shi 識 for vijñāna – that remained standard for the rest of Chinese Buddhist
history.
Looking at the specific Chinese terms in question, we can often make at least
some sense of them. The initially perplexing translation geng 更, for example,
likely derives from this word’s semi-colloquial meaning (during the period
when An Shigao worked) of “encounter something bad” (hence my translation
“to suffer”) – such as being smitten by disease, curses, or other misfortune.26
Meanwhile the puzzling translation “to plant” or “a sprout” (zai 栽), was per-
haps chosen based on an explanation of pratītyasamutpāda that An Shigao
himself presents in the AHKJ. As Stefano Zacchetti has observed, the AHKJ
draws an analogy between the twelve links of pratītyasamutpāda and an
“external” sequence of twelve causal factors in the growth of a seed.27 Thus
ignorance (avidyā), the first link, is correlated with the “external” condition
earth (soil), conditioned factors (saṃskāra) with water, and so forth.28 This
continues for all twelve links, finishing with “old age and death” as the fruit (shi
實). Notably within this sequence the sixth “external” factor is the “seed,” zhong
種, and this perhaps explains An Shigao’s use of “sprout” or “to plant” (zai 栽)
to translate the corresponding “internal” factor, sparśa.29
25 “To suffer” (更) was reasonably influential, but only until the late fourth century. Zai 栽
was occasionally used into the very late third century (e.g., Fangguang bore jing 放光般
若經, T.8.221: 12a22–23), but not beyond that.
26 Hu, 2002: 143–144. The basic meaning of geng 更 is “change” (the Shuo wen 說文 glosses
it as gai 改; see GHYZD: 447). The meaning is thus similar to bian 變, which also means
both change and, by extension, misfortune. That geng 更 was here being used in this
sense, and, not in a more neutral sense of “experience” or “undergo” (cf. Vetter, 2012: 146),
is suggested by the slightly later, and only occasionally used translation of sparśa as geng-
le 更樂 (e.g. Weimojie jing 維摩詰經, T.14.474: 536a16), seemingly an attempt to create an
abstract noun (“contact”) by joining two opposites (“suffering and pleasure”).
27 Zacchetti, 2004: 207–212.
28 T.25.1508: 53a18–24.
29 Zacchetti, 2004: 208, note 59. Cf. Yiqie jing yinyi 一切經音義, which correctly interprets
zai 栽 as meaning sparśa (chu 觸), but which explains it by analogizing sparśa, which
256 Greene
Here too, however, knowing the ultimate source of the explanation of sparśa
as a seed or sprout does not explain why An Shigao chose to base his Chinese
translation of this word (or rather, one of his translations) on this understand-
ing, as he certainly did not follow a similar strategy with his translations of the
other pratītyasamutpāda terms. As with An Shigao’s translation of āyatana as
“destroyers,” we see here evidence that the concept of sparśa was somehow felt
to require a different kind of explanation.
3 The Ambiguities of Contact
An Shigao’s approach to the term sparśa in particular raises a number of ques-

tions about his own understanding of pratītyasamutpāda and about the
manner in which his Chinese readers and students would have been led to
interpret it. Indeed, even if they were all originally chosen with some definite
intention, we might easily wonder whether a Chinese reader of An Shigao’s
texts would have been able to make sense of the many distinct terms for sparśa
and their divergent meanings.
We might notice, for example, a certain promiscuity in the use of the
Chinese word shou 受, “taking hold” or “grasping”. This word is, firstly, An
Shigao’s regular translation of upādāna, the “grasping” that accompanies crav-
ing and desire. But he also used shou 受 as part of one translation of āyatana
(rushou 入受), as well as in one translation of sparśa (gengshou 更受). In these
cases, the word thus seems intended to point to raw sense perception.
The Chinese commentaries to An Shigao’s texts written within the first cen-
tury after his period of activity reveal that in other contexts it was indeed
common to use shou 受 to refer to the basic operation of the sense organs. The
canonical Anban shouyi jing 安般守意經 (a third-century Chinese commen-
tary on an An Shigao translation)30 thus typically explains that sense perception
occurs when a sense organ “grasps” (shou 受) its object: “the eye grasps visible
forms” (yan shou se 眼受色) and so forth.31 The Yin chi ru jing zhu 陰持入經註
(YCRJZ hereafter), a third-century commentary to An Shigao’s Yin chi ru jing 陰
“grows into the mind and mental factors,” with a sprout, which grows into a tree (T.54.2128:
356c19).
30 I here use the title “canonical” to distinguish the transmitted text bearing this name from
that discovered at Kongō-ji 金剛寺, which is likely An Shigao’s original translation
(Deleanu, 2003; Zacchetti, 2008).
31 T.15.602: 169c29. The other senses are described similarly. See T.15.602: 167c11–12 (ear);
169c27 (nose).
持入經 (YCRJ hereafter), similarly comments on a list of the twelve āyatanas

(the six sense organs plus their corresponding objects) by noting that “the six
internal āyatanas [the sense organs] grasp (shou 受) the six external āyatanas
[the sense objects]”.32
Given that An Shigao regularly translated the term “grasping” (upādāna)
using a term that was also incorporated into some of his translations for “con-
tact” (sparśa), and given further that Chinese Buddhists themselves seem to
have often used this same word to describe the basic mode of interaction
between sense organs and sense objects, we might wonder whether those who
read and commented on An Shigao’s translations understood there to be much
of a distinction between these two ideas.
This is a question worth considering if only because in what we might call
the normative Indian Buddhist interpretation of pratītyasamutpāda, links 5, 6
and 7 (the six sense gates, contact, and sensation), and links 8 and 9 (craving
and grasping) invariably belong to fundamentally different categories; the for-
mer, the fruits of past karma, and the latter, the present karma that leads to
rebirth and suffering in the future.33 Although different Indian commentarial
traditions divide up pratītyasamutpāda in different ways, there is broad agree-
ment about this basic point.34 Indeed, early Buddhist scriptures consistently
posit an important analytical and practical distinction between raw sensory
activity and mental defilements such as craving or aversion. Buddhist mental
32 T.33.1694: 11a2–3. As noted by Zacchetti, 2010b: 144–147, the base text and inter-linear
commentary were transmitted together as a single unit, and the distinct title Yin chi ru
jing zhu 陰持入經註 is a creation of the editors of the (modern) Taishō canon. For ease
of reference, however, I will use this title to refer to the commentary.
33 That the division between “old” and “new” karma is the division between sense experi-
ence and one’s response to sense experience (or if not one’s response, then some other
mental factor that is distinguishable from sense experience itself) is already seen clearly
in the early scriptures (see for example SN IV:172).
34 The typical approach of the Pāli commentarial tradition marks a distinction between the
fruits (phala) of past karma (consciousness, name-and-form, the six sense gates, contact,
and sensation) and present karmic causes (hetu), namely craving, grasping, and becom-
ing (Vism: 579–580). The Vaibhāṣika system (AKBh, 134.6–135.6) similarly distinguishes
between consciousness, name-and-form, the six sense gates, contact, and sensation as
“bases” (vastu; explained as the results of past karma), and craving and grasping as “defile-
ments” (kleśa; see also Van Den Broeck, 1977: 131–132). The Abhidharmasamuccaya gives
yet another classification, distinguishing the “projected members” (ākṣiptāṅga) that are
the results of past actions (name-and-form, the sense gates, contact, and sensation) from
the “actualizing members” (abhinirvartakāṅga) that are causes for the future, namely
craving, grasping, and becoming (ADhS, 26.6–9; Kritzer, 1999: 27). For references to other
similar divisions, see de La Vallée Poussin, 1913: 35–36.
258 Greene
training, we are told in such texts, lies not in eradicating one’s present sensory
experience – which is the unavoidable fruit (vipāka) of past karma – but in
preventing the karmically productive mental defilements that arise in response
to whatever sensations (vedanā) one experiences.35 This formulation amounts
to a technical explanation of the basic Buddhist notion that it is possible to be
fully liberated while still alive (which was not something that all Indian ascetic
groups believed); that it is possible, in the language of pratītyasamutpāda, to
experience “contact” (sparśa) and “sensation” (vedanā) without then giving
rise to craving or attachment.
That we can separate sense perception into two such aspects is not neces-
sarily a phenomenological given. One of the primary functions of the central
section of the pratītyasamutpāda formula is thus to provide the vocabulary
that makes such a separation analytically possible. And from this point of view,
we may say that An Shigao’s use of overlapping Chinese terms for his transla-
tions would not have contributed to cementing a clear distinction between
sparśa as the fruit of past karma in contrast to upādāna as newly created
karma.
It is, further, interesting to observe that there are a number of other exam-
ples where An Shigao’s translations blur such distinctions in a potentially
similar way. Consider, for example, the seemingly strange translation of sparśa
as zai 栽, “to plant” or “a sprout”. As discussed above, this translation might have
been motivated by the analogy between “internal” pratītyasamutpāda and an
“external” sequence of the growth of a plant discussed in the AHKJ, in which
sparśa lines up with the “seed” (zhong 種)36 of the external sequence. However
within the AHKJ we find another passage where “planting” and “sprouts” are
associated not with sparśa, but with saṃskāra (xing 行), the second link of
pratītyasamutpāda.
This passage begins with reference to the previously given analogy of the
growth of a seed, and then comments on the first several links of the chain:
“The earth [in the previous analogy] corresponds to confusion (avidyā); activi-
ties (saṃskāra) refers to the flourishing of the dark elements (*upādānaskandha)”
(di ming chi xing wei sheng yin 地名癡行為盛陰). What is going on here is then
35 We thus find, in the early scriptures, innumerable statements to the effect that “the eye is
not the fetter of visible forms, nor are visible forms the fetter of the eye, but rather the
desire and lust that arise there in dependence on both: that is the fetter there”. (SN IV:163;
translation by Bodhi, 2000: 1230 with slight modifications). Conversely the Buddha is
reported to have criticized those who taught practices that merely involved ceasing sen-
sory perception itself (MN III:298–302). See also Bronkhorst, 1986.
36 Or perhaps, the “planting”; both 種 and 栽 functioning as either nouns or verbs.
further explained: “[This means] one carries out ‘activities of birth and death’
(saṃskāra); to plant the sprouts [of rebirth] is called ‘activity.’” (geng zuo sheng
si xing. zi zhong zai ming wei xing 便作生死行。自種栽名為行).37
The activity of the saṃskāras is thus said to be like “planting sprouts” (zhong
zai 種栽), and this indeed fits with typical explanations of saṃskāra (in this
context) as the past karmic activities that have given rise to present existence.
However in the immediately preceding passage, the AHKJ also links “planting
sprouts” to a different section of pratītyasamutpāda:
十二因緣有三38 事。一者癡。二者生死精行。是前世因緣。三者識。
從識受身生。四者名色。色身復成五陰。是今世因緣。五者六衰。復
作生死精行種栽。是後世因緣。39
The twelve causes and conditions [pratītyasamutpāda] have three divi-

sions. Link 1, confusion, and link 2, the seminal activities of birth and
death, are [the first division], causes and conditions from a previous life.
Link 3 is consciousness, from which one receives a new rebirth, and link
4 is name-and-form, the material body which further forms the five
aggregates – these are [the second division], causes and conditions [rel-
evant for] one’s present lifetime. Link 5 is the six destroyers [āyatana],
which once again give rise to the seminal activities of birth and death and
the planting of the sprouts [of rebirth] – these are [the third division],
the causes and conditions [that give rise to] one’s future life.
Dividing the members of the pratītyasamutpāda formula into past, present,

and future, as we see here, is an approach found throughout the Buddhist com-
mentarial tradition, and some of how the AHKJ does this in the above passage
is entirely expected. Thus, that links 1 and 2, “confusion” (avidyā) and “seminal
activities of birth and death” (saṃskāra), would be “causes and conditions
from a past life” accords with similar explanations seen in sources such as the
Visuddhimagga or Abhidharmakośabhāṣya.40 Quite unexpected, however, is
that the AHKJ separates links 3 and 4 from link 5 (āyatana; the “six destroyers”),
which it declares to be not the fruit of past karma (which is what the text seems
to mean by “causes and conditions [relevant for] one’s present lifetime”) but
37 T.25.1508: 53b16.
38 Following the reading 三 given in the Sixi (宋), Puning (元), and Kunaichō (宫) editions.
39 T.25.1508: 53b6–10.
40 See above note 34.
260 Greene
a cause of future rebirth, as something that “gives rise to the seminal activities
of birth and death and the planting of the sprouts [of rebirth]”.41
Reading this passage in keeping with a more typical understanding of
pratītyasamutpāda, we might take “the seminal activities of birth and death
and the planting of the sprouts [of rebirth]” as an explanation of how the
karmic activities that condition one’s future life (craving, grasping, becom-
ing) arise in the wake of the entire process of sense activity (āyatana, sparśa,
vedanā). To refer to presently performed karma using the same language that
earlier characterized the saṃskāras (the “sprouts”) would make perfect sense,
since the saṃskāras are usually explained as, precisely, the karmic activities of
a past lifetime.
However from the Chinese text of the AHKJ alone, such an understanding
could probably only be reached by someone with a certain amount of back-
ground knowledge. Indeed not only is the word “sprouts” (zai 栽) explicitly
given earlier in the text as a translation of sparśa, but in the above passage it is
explicitly said that “planting the sprouts” is produced from the six sense gates,
precisely where we would expect sparśa to occur. In other words, as a Chinese
text the AHKJ seems here to assert that saṃskāra and sparśa are similar if not
identical things.42
41 A very similar discussion immediately follows this passage: “Once there is name-and-
form, there then comes to exist the six sense gates. There then again arises the flourishing
of the dark elements. After activities plant the sprouts [of rebirth], one will once again
undergo rebirth based on the unfurling of causes and conditions”. 已有名色，便有六
入。復作盛陰。行種栽後，當復受轉相因緣生死。(T.25.1508: 53b17–19).
42 It is interesting to note that a very similar ambiguity occurs in the Liao ben sheng si jing 了
本生死經, an early Chinese translation of the Śālistambasūtra, a text that, as Zacchetti,
2004, has noted, may have been important in the formation of the AHKJ (the translator of
the Liao ben sheng si jing is unknown, but the text is cited in the YCRJZ, so it must pre-date
the mid-third century). In a passage listing the entire pratītyasamutpāda formula,
saṃskāra is initially translated as xing 行. Links 5 through 9 are then described as follows:
“In dependence on name-and-form, there are the six sense gates. Because of a union of
the three [organ, object, and consciousness] there is contact. Because of the activity of
contact, there is sensation. Delighting in this sensation, there is craving. When craving
comes to fullness, there is grasping. When there is grasping there is again activity, and
therefore there is becoming”. 猗名色根故為六入。三合故為更樂。更樂行故為
痛。痛而樂故為愛。愛彌廣故為受。受當復有行，故為有。 (T.16.708: 816a11–
17). Thus xing 行, which initially translates saṃskāra (as is usual), appears again in the
discussions of sparśa (geng le xing 更樂行) and upādāna (shou dang fu you xing 受當復
有行). In the case of upādāna, the point seems clear – grasping is precisely the occur-
rence of new karmic formations (saṃskāra). For sparśa it is much less easy to explain
things. Whatever the motivations of the translator, a Chinese reader here could easily see
There is yet a third example where one of An Shigao’s translations ends up,
intentionally or not, associating sparśa not with karmic fruits from the past but
with karmic causes in the present. The word geng 更, “to encounter something
bad” (translated in the table above as “to suffer”) was one of An Shigao’s most
frequently used translations of sparśa. In the YCRJ, however, it translates not
sparśa but cetanā, “volition,” a concept usually associated with karmically pro-
ductive mental activity,43 that is to say, with the factors that are the cause of
future rebirth as opposed to those that are the fruits of past karma.
The context in the YCRJ is an analysis of the fourth skandha, “conditioned
factors” (saṃskāra; xing 行). The relevant passage of the YCRJ, and the Pāli par-
allel from the sixth book of the Peṭakopadesa,44 are as follows:
Tattha cha cetanākāyā saṅkhārakkhandho, rūpasañcetanā yāva

dhammasañcetanā. Ime cha cetanākāyā; ayaṃ saṅkhārakkhandho. (Peṭ,
112/437)
行種名為身六更。一色所更，二聲所更，三香所更，四味所更，五
通45 所更，六法所更。是為身六更。是名為行種。(T.15.603: 173b16–
19)
that the mysterious “activities” (行) were active during (in their Indic terminology)
saṃskāra, sparśa, and upādāna. This is precisely the same overlap that we see in the
AHKJ’s discussion of the “sprouts” (zai 栽).
43 “It is intention, O monks, that I declare to be karma. Having intended, one makes karma
through body, speech, and mind”. cetanāhaṃ, bhikkhave, kammaṃ vadāmi. cetayitvā
kammaṃ karoti kāyena vācāya manasā (AN III 415). On the interpretations of this oft-
cited passage, see McDermott, 1980: 181–189.
44 On the Peṭakopadesa as a parallel to the YCRJ, see Zacchetti, 2002.
45 The Taishō faithfully reproduces the second Kǒryo edition reading chu 觸. The first Kǒryo
edition reads the same (<http://kb.sutra.re.kr/ritk>, accessed July 2012; photographs of a
copy of the first Kǒryo edition of this text were available here from 2011–2013, but have
since been removed). However the Taishō records the variant tong 通 in the Sixi (宋) and
Puning (元) editions, and this is shared by the Jin 金 edition (ZHDZJ, 36:130). This must be
correct because the commentary, in all editions, comments on the character tong 通, not
chu 觸. The word here must correspond to spraṣṭavya, “tangibles,” translated as 更 in the
previous paragraph. Chu 觸 becomes the translation of sparśa and spraṣṭavya beginning
only with the translations of Kumārajīva in the late fourth century. The versions of the
YCRJ represented in the Kǒryo editions have thus probably been “corrected” based on this
later usage. This anachronistic chu 觸 occurs in two other places in the text (T.15.603:
174c7, c18). Though the Jin edition shares these readings, the first Kǒryo edition reads, in
both cases, shen 身, agreeing with the Sixi, Puning, and Kunaichō (宫) editions. I suspect
that shen 身 is again correct. This might be confirmed by the commentator’s initial gloss
262 Greene
The correspondence between the texts is, as elsewhere, nearly exact.46 The
meaning of the Pāli version is straightforward:
Here the aggregate (skandha) of formations refers to the six collections of

volition (cetanā), namely volition in regards to visible objects (rūpa),
[volition in regards to sounds, smells, tastes, tangibles], and volition in
regards to mental objects (dharmas). These six collections of volition
comprise the aggregate of formations.
Examining the corresponding Chinese, it is clear that An Shigao here used

geng 更 to translate cetanā and/or sañcetanā (thus shen liu geng 身六更
for cha cetanākāyā, “the six collections of volition,” se suo geng 色所更 for
rūpasañcetanā, “volition in regards to visible objects,” and so forth).47 However
as we have seen, the word geng is, in other An Shigao texts, a translation of
sparśa.48 And of course, the context here is an explanation of the word
xing 行, An Shigao’s normal translation of saṃskāra, the second member of
pratītyasamutpāda, and a word that, as we have seen above, was itself some-
times conflated with sparśa through the translation “sprouts” (zai 栽).
Thus while the original Indic passage here clearly explains saṃskāra to
mean volitional formations in regards to sense objects, it is again not hard to
see how this would have been quite obscure to a reader of An Shigao’s Chinese
version. Perhaps especially to someone familiar with An Shigao’s other transla-
tions, where geng 更 is frequently used in the context of sense contact (sparśa),
the “six collections of volition” that are here equated with the “activities”
(saṃskāra) could easily be understood as the “six [kinds of] bodily contact”
(shen liu geng 身六更), an interpretation made even easier by An Shigao’s
of tong 通 as shen 身 (T.33.1694: 10a19), which seems to assume that the reader is familiar
with shen 身 as a translation of spraṣṭavya.
46 Tattha cha cetanākāyā saṅkhārakkhandho (xing zhong ming wei shen liu geng 行種名為
身六更) rūpasañcetanā (yi se suo geng 一色所更) yāva (er sheng suo geng, san xiang suo
geng, si wei suo geng, wu tong suo geng 二聲所更，三香所更，四味所更，五通所更)
dhammasañcetanā (liu fa suo geng 六法所更) ime cha cetanākāyā (shi wei shen liu geng
是為身六更) ayaṃ saṅkhārakkhandho (shi ming wei xing zhong 是名為行種).
47 In the expression se suo geng 色所更, the word suo 所 (“place”) might be an attempt to
explain rūpasañcetanā as a locative or genitive dependent compound (“sañcetanā with
respect to rūpa”).
48 That geng 更 was seen by An Shigao as a plausible translation for a term like cetanā is
confirmed elsewhere by his translation of samyaksaṃkalpa, the second member of the
eightfold path, as zhi geng 直更 (T.1.57: 852a13). Saṃkalpa is usually understood as similar
in meaning to cetanā.
etymologically correct, but in terms of normal Chinese usage entirely baffling

selection of the character shen 身 to translate “collection”.49
This reading would, moreover, have found support in a slightly later section
of the YCRJ, where saṃskāra is defined in the context of pratītyasamutpāda.
彼癡因緣行為何等？為六望受。何等為六？色、聲、香、味、身50 、
法，是為身六望受。是名為行。51
What are these “activities” that are conditioned by confusion? They are
the six “yearnings”. What are the six? Visible forms, sounds, smells, tastes,
bodily [sensations], and dharmas. These are the six yearnings of the body.
These are called the “activities”.
In this case, we cannot rely on the parallel from the Peṭakopadesa because
the corresponding section is missing.52 However the definitions given
immediately after the above passage for vijñāna and vedanā (as members of
pratītyasamutpāda) are identical to those found in the earlier passage where
these same terms were defined as members of the five aggregates (skandha).53
This strongly suggests the definition of saṃskāra here was also identical to its
earlier definition as one of the skandhas (the passage discussed above), as the
six “collections of volition” (cetanākāyā). Indeed once we take the otherwise
obscure term wangshou 望受, translated above as “yearning,” simply as a differ-
ent translation of cetanā, the two passages are effectively identical.
It is quite remarkable that the YCRJ here repeats an evidently identical
passage but translates its key terminology differently. And while we cannot
know for certain what motivated An Shigao, as the list of pratītyasamutpāda
49 An Shigao uses shen 身 in this manner throughout the YCRJ, and elsewhere (see for exam-
ple Vetter and Harrison, 1998, passim). It would seem that An Shigao’s Chinese readership
did not understand his idiosyncratic (to put it mildly) usage (see the YCRJZ’s explanations
of the “six collections of consciousness” liu shen shi 六身識; T.33.1694: 14a18–19; this has
been noted by Zacchetti, 2010b: 157).
50 Here following the first Kǒryo, Sixi, Puning, and Kunaichō editions. The second Kǒryo
edition reads chu 觸 (which as I suggest above must be wrong; see note 45).
51 T.15.603: 174c6–7.
52 The absence does not appear to be intentional (Ñāṇamoli, 1964: 160). In light of the YCRJ,
Ñāṇamoli’s suggestion that the missing passage might have referred to the three-fold divi-
sion of saṅkhāra as body, speech, and mind is probably not correct.
53 In the Peṭakopadesa, the definitions of vijñāna and vedanā as a members of pratītyasa
mutpāda are abbreviated (as the “six collections” of each term), thus referring back to the
full definition given under the analysis of the skandhas (Peṭ, 116/459 and 463).
264 Greene
continues and we come to sparśa, we find that it too now receives a new
translation, namely siwang 思望, “distantly longing”.54 It is again striking that
cetanā (wangshou 望受) and sparśa (siwang 思望) have been translated into
reasonably similar Chinese words. By itself the similarity between these two
terms would perhaps not be enough to attract our attention. But considered
together with the earlier translation of cetanā as geng 更 that also overlaps
An Shigao’s most common translation of sparśa, and the similar overlapping
use of “sprouts” (zai 栽) elsewhere, we might be justified in seeing the pair-
ing wangshou 望受 and siwang 思望 as another example of the way that An
Shigao’s translations tend to blur the distinction between sense contact itself
(sparśa) as the fruit of past karma, and karmically productive factors such as
craving, grasping, or volition (cetanā).
4 The Chinese Interpretations: Sense Perception and Desire
An Shigao’s translations thus often seem to create the potential for a Chinese
reader to see sparśa, sensory “contact,” as something very similar or even iden-
tical to things such as upādāna, saṃskāra, or cetanā, terms that in Indian
interpretations of pratītyasamutpāda are usually presented as forms of karmic
activity, and hence as strictly different from sense contact itself (understood as
the fruit of past karma).
This blurring of the line between sparśa and karmically productive desire
was, I believe, part of a more general pattern in the earliest Chinese interpreta-
tions of Buddhist theories of sense perception. This pattern is revealed most
clearly through the vocabulary that appears in third-century Chinese composi-
tions as glosses on, or even as alternative translations of, the word āyatana. Our
most important source in this regard is the YCRJZ, one of our best windows
onto mid-third-century Chinese Buddhism, and a text that attests to the con-
tinuing influence and legacy of An Shigao’s translations during this time.55
As mentioned above, within An Shigao’s translations the word āyatana is
rendered in two distinct ways: as the “entrances” (ru 入) and the “destroyers”
(shuai 衰). Within translated texts, these remain the preferred Chinese terms
54 The word siwang 思望 is attested from the mid-Han (HYDCD, 2:4254). Its meaning in the
late second and early third centuries comes out clearly in a passage from the Taizi ruiying
benqi jing 太子瑞應本起經: “[Beings] are pulled about by desire, confused by material
acquisitions and sex, and long for comfort and pleasure. From this arises the root of all
evils” 牽於愛欲，惑於財色，思望安樂，從是生諸惡本。 (T.3.185: 478b9–10).
55 On the YCRJZ, see Zacchetti, 2010b.
throughout the Three Kingdoms period (until ca. 280 CE). However the YCRJZ,
a Chinese composition rather than a translation, uses quite different terminol-
ogy. Both when glossing the more standard translations found in the root text
(the YCRJ), and in other, independent discussions, the “external” āyatanas (the
sense objects) are mostly commonly called the “six objects of desire” (liu yu
六欲)56 or the “six perversities” (liu xie 六邪),57 while the “internal” āyatanas
(the sense organs) are, almost invariably, termed the “six emotions” (liu qing
六情).58
The persistent use of the term “six emotions” to refer to the sense organs is
especially noteworthy because of the important role played by the concept of
qing 情 in pre-Buddhist Chinese discussions of human nature.59 At the risk of
glossing over some of the nuances, we may generally translate qing as “emo-
tion,” and early Chinese philosophical texts contain various lists of six or seven
such emotions such as joy, anger, sadness and so forth. There was much discus-
sion among early Chinese thinkers about the nature and status of the qing.
Most important for our purposes is that by the early Western Han dynasty (206
BCE–9 CE) the qing had come to be understood not simply as those human
emotions that were natural and unlearned (the primary meaning before this
time), but rather those desires that are problematic. Typical examples of the
qing thus came to include things such as “greed” (tan 貪) and other factors
associated with our baser, animal natures.60 By the Eastern Han (25–220 CE),
the qing had further come to be linked to the yin 陰 or “dark” aspects of human
qi 氣 (“vital energy”),61 which were understood as the factors within the body
that cause death and decay (this provided a physiological explanation for the
dangers of desire).62 The understanding thus came to be not simply that the
qing were human emotions in need of regulation or control, but that human
56 See T.33.1694: 11a3, 11a17, 19c15, among other examples.

57 See for example T.33.1694: 13c11, 14a15–16, 14b26, 14c11, 17a29. See also Kang Senghui’s pref-
ace to the Fa jing jing 法鏡經 (Chu sanzang ji ji 出三藏記集, T.55.2145: 46b22–23), and
his Liu du ji jing 六度集經 (T.3.152: 28a15–17). Within the YCRJZ, two other expressions for
the external āyatanas each appear only once: “the six objects of consumption” (liu fu 六
服; 21c20), and “the six confusions” (liu dao 六倒; 17c27).
58 The one exception appears to be the single instance of the term liu ti 六體, the “six bod-
ies” (T.33.1694: 9c15).
59 On the qing, see Graham, 1990: 59–66; Hansen, 1995; Ng, 1998; Cheng, 1999; Harbsmeier,
2004; Puett, 2004.
60 Xu, 2001: 2:315; Tang, [1957] 2001: 99.
61 Lun heng jiao shi: 139; the Shuo wen 說文 even straightforwardly defines qing 情 as the yin
qi 陰氣 of human beings (Morohashi, 4:1074).
62 Lewis, 2006: 47–51.
266 Greene
beings would be better off without them all together. This idea is embodied in
the famous dictum that “sages” (sheng 聖) such as Confucius were sages pre-
cisely because they were “without qing” (wu qing 無情), that is to say free of
problematic human desires and emotional responses.63
Qing 情 was thus a central concept in pre-Buddhist discussions of desire,
emotion, and human perfection, and its appearance as a technical term within
the YCRJZ reflects an effort to find ways of talking about Buddhist theories of
sense perception and desire that would resonate with a more general Chinese
audience. Many scholars have indeed noted the tendency of key third-century
Buddhist texts and authors – especially those associated with the Wu 吳
Kingdom period (220–280 CE), as is the YCRJZ – to freely draw from a range of
Chinese philosophical concepts in their attempts to explicate Buddhist ideas.
It is in Wu-Kingdom Chinese Buddhist texts that we thus find the first exam-
ples of interpretations of Buddhist idea of rebirth that make use of terms
derived from pre-Buddhist Chinese understandings of the fate of the spirit
after death such as shen 神, shishen 識神, hunling 魂靈, and shenling 神靈.64
Since qing referred, in the Chinese context, to the problematic desires that
perfected human beings have eliminated, in many ways it would have been a
perfect translation for terms such as tṛṣṇā or kleśa that refer to the craving and
other mental defilements that the practice of Buddhism is supposed to eventu-
ally eliminate. Yet qing was never used to translate or gloss such terms. When
used in a technical context, this word appears in early Chinese Buddhist litera-
ture exclusively as a collective designation of the six internal āyatanas, the
sense organs themselves. It is not clear how or when this word came to be used
in this manner, though it does not seem to have happened prior to the Wu
Kingdom period.65
Regardless of the details of its first appearance, what is most relevant about
the interpretation of the internal āyatanas as the “six emotions” is that it seems
63 The classic treatment of this subject remains Tang, [1957] 2001 (see especially p. 101–102).
See also Wang, 1987: 372–375; Wu, 2009: 9–30.
64 Zacchetti, 2010b: 171–176. Zacchetti aptly refers to all of these and other similar terms as
“soul-language.”
65 Using Nattier’s list of likely authentic Han and Three-Kingdoms Buddhist translations
(Nattier, 2008), there are only two possible examples earlier than the Wu Kingdom. First
is the Zhong benqi jing 中本起經 (T.4.196: 152a9), but as Nattier points out this text was
most likely revised in light of the other biographies of the Buddha that appeared during
the Three Kingdoms period (Nattier, 2008: 109). The other example is the AHKJ (T.25.1508:
53c13). However the AHKJ uses the expression liu qing 六情 not for the āyatanas, but for
the six sensory consciousnesses (vijñāna). It is possible that this usage eventually
morphed into what we find in later texts such as the YCRJZ.
to accord with, and indeed further reinforces, the ambiguity between sparśa as
raw sense “contact” and karmic forces such as craving, an ambiguity that, as we
have seen, was a persistent feature of An Shigao’s translation of the terms of
the pratītyasamutpāda formula.
The YCRJZ, following the root-text’s listing of the six internal āyatanas and
their corresponding objects, thus explains their interaction by noting that “the
six internal emotions grasp the six external objects of desire” (nei liu qing shou
wai liu yu ye 內六情受外六欲也).66 As noted above, merely describing the
basic interaction between sense organs and sense objects as “grasping” (shou
受) already suggests a certain conflation of sense contact and karmic desires
because “grasping” was also the standard translation of upādāna. The glosses
of “six emotions” and “six objects of desire” for the sense organs and their
objects reinforces this perspective – under this description sense perception is
not an event that makes available a possible object of desire, but rather simply
is the meeting of human desires with the objects of those desires.
Indeed, elsewhere the YCRJZ describes the same interaction using not shou
受 but ai 愛, “craving,” the universal Chinese translation of tṛṣṇā: “the internal
emotions [āyatanas] crave the objects of desire” (nei qing ai yu 內情愛欲).67 Or,
in another passage, “the emotions [āyatanas] of a human being hunger and
thirst for the six objects of consumption” (ren qing ji ke yu liu fu 人情飢渴于六
服), with the sense objects themselves here likened to the food and drink.68
And in keeping with this understanding, the YCRJZ often describes the hoped
for fruit of Buddhist practice as the opposite of this, as a state where the six
emotions “no longer grasp the six desirable objects” (bu fu shou wai liu yu 不復
受外六欲).69 Or, as another passage has it, where “the six emotions no longer
seek out anything in the three spheres” (liu qing bu fu yu san jie zhong you suo
qiu suo ye 六情不復於三界中有所求索也).70
The difficulty that arises when trying to understand these explanations in
terms of normative Indian Buddhist understanding is that by glossing the word
āyatana as “emotion” (qing 情), the YCRZJ seems to attribute desire and craving
to the sense organs themselves. Indian Buddhist presentations of sense per-
ception do, of course, often take on a similar tone, and there are certainly
66 T.33.1694: 11a2–3; some versions here read “crave” ai 愛 in place of “grasp” shou 受.
67 T.33.1694: 15a3.
68 Liu fu 六服 for the six sense objects is not attested elsewhere. The image here is perhaps
drawn from Liu du ji jing 六度集經, T.3.152: 15c23–24.
69 T.33.1694: 19c15; here the context is a description of the fourth dhyāna.
70 T.33.1694: 18b17. Note that “to seek for” (qiu 求), here describing how the sense organs find
and make contact with their objects, is also one of An Shigao’s translations of upādāna
(see 9b in the table above).
268 Greene
precedents for understanding the sense objects themselves as “objects of

desire” (kāmaguṇa). What we should note, however, is that the relevant Indian
Buddhist vocabulary, particularly the terms found within the central sections
of the pratītyasamutpāda formula, makes it possible, and indeed necessary, to
separate sense experience into two aspects, one that is potentially contami-
nated by desire, and the other that is karmically neutral (and hence neutral
with respect to craving or aversion). The distinctions afforded by the terms
āyatana, sparśa, vedanā, and tṛṣṇā, make it possible to speak intelligibly about
two people who have the same sensory organs, who encounter the same range
of objects, and who feel the same pleasant or unpleasant sensations, but who
then respond to these sensations differently, one person under the influence of
craving and the other not.71
But the language used by the YCRJZ does not allow for this possibility. The
sense organs simply are the “emotions” and the sense objects the “desires”.
Craving and grasping are here not defilements that might or might not accom-
pany sensory contact, but rather the very forces that produce sensory contact.
Sensory objects are not karmically neutral things towards which different
kinds of people might have different responses, but are already “perversions”
(xie 邪).72 The YCRJZ has glossed the Indian technical terminology (or rather,
An Shigao’s Chinese translations of the Indian technical terminology) in such
a way that it is impossible to describe sense contact happening in the absence
of craving and desire. It is thus not surprising that when the YCRJZ wishes to
describe the absence of desire or defilement, it does not do this by describing a
state in which one no longer feels desire or aversion towards objects that one
has sensed, but rather a state in which one “prevents the six sense objects from
making contact with the six sense organs” (que liu xie wu ling ru liu qing 却六邪
無令入六情).73
71 We find this, for example, in the extended treatment of pratītyasamutpāda within the Pāli
Vibhaṅga, where both wholesome (kusala) and unwholesome (akusala) states of mind
(citta) are analyzed; the formula remains the same up until taṇhā, which in wholesome
states of mind is replaced by other terms (Vibh: 135–192).
72 In the YCRZJ, this same principle seems to apply to mental objects, and the root text’s fa
法 (dharmas, the object of the mental organ) is glossed with terms like “defiled thoughts”
(xie nian 邪念; T.33.1694: 9c9–10, c15–16; in the first passage it seems likely that xie rong
邪榮 is a mistake for xie nian 邪念). Here the YCRJZ seems to imply that objects of the
mind are not simply mental objects in the abstract, but specifically those defiled objects
that one must eliminate (AHKJ similarly uses the term nianyu 念欲, “thoughts concern-
ing desire,” rather than fa 法; T.25.1508: 54b3).
73 T.33.1694: 17a29.
5 Conclusions
The material examined above has allowed us to move through the complete
cycle of translation and interpretation, from the original Indic terms of the
pratītyasamutpāda formula, to their Chinese translations in the works of An
Shigao, and finally to the interpretation of these translations by early Chinese
commentators. Tracing this process has allowed us to discern one key area
– notably the concept of sparśa, sensory “contact” – where An Shigao may
have struggled to find ready Chinese equivalents, and where, moreover, the
bewildering diversity of his translations may have encouraged interpreta-
tions potentially at odds with what I have styled the “normative” Indian
understanding.
However while something interesting was thus clearly going on here, in
both An Shigao’s translations and in Chinese commentaries such as the YCRJZ,
it is more difficult to determine precisely what accounts for these unusual
translations and interpretations. Do they represent some kind of Chinese mis-
understanding that was reflected even in An Shigao’s translations themselves
as his assistants struggled to find Chinese concepts to express the potentially
mysterious Indian Buddhist notion of sense “contact” (sparśa)? Or do they
reflect An Shigao’s own attempts to inflect his presentation of these ideas with
the understanding of certain Indian commentarial traditions? Indeed, we saw
some examples – such as the translation of āyatana as “destroyer” (shuai 衰) –
that might be understood as “exegetical” translations, translations that add a
layer of explanation or glossing within the translated words themselves.
Although further investigation of this must await another occasion, it might
be possible to explain An Shigao’s treatment of the term sparśa in a similar
fashion. We know that the precise nature of sparśa was the subject of consider-
able debate within the Abhidharma and early commentarial tradition,74 and
sense perception more generally was a topic of recurring controversy, between
Buddhist schools as well as between Buddhists and adherents of other Indian
philosophical systems.75 It is possible that An Shigao was well versed in some
version of these debates, and that his less-than-straightforward translations of
sparśa were attempts to somehow include the latest Indian scholastic under-
standings within the Chinese words he used. For example, in many Indian
Buddhist commentarial traditions the category of saṃskāra eventually comes
to include almost all otherwise unclassified mental or pseudo-mental fac-
tors, including sparśa. This could have something to do with the occasional
74 Sarachchandra, 1958: 27–31; Dhammajoti, 2007: 120–131.

75 Cox, 1988; Taber, 2005; Dhammajoti, 2007.
270 Greene
overlap between An Shigao’s translations of these words in the context of

pratītyasamutpāda.
But whatever their proximate cause, it is also true that An Shigao’s seem-
ingly peculiar translations resulted in a way of speaking about sense perception
that dovetailed easily with long-established Chinese ways of thinking about
this topic. The understanding seen in the YCRJZ that the sense organs them-
selves are the six “emotions” thus accompanies and encourages a view of the
sense organs as active agents that seek out what they desire. Rather than
merely the loci where sense contact passively happens, the “six emotions”
themselves reach out to their objects. As the YCRJZ puts it, “in accord with what
the mind intends, the six emotions make contact with the six objects of desire”
(liu qing cong yi suo nian zhi liu yu ye 六情從意所念致六欲也).76 It is precisely
this attribution of active agency to the sense organs, and to sense experience
more generally, that appears at odds with the Indian Buddhist perspective,
where at least part of the sensory process is usually presented as the more or
less passively received fruits of past karma.77
Yet this more “active” interpretation of sense perception was the norm in
China prior to the arrival of Buddhism. Early Chinese philosophical texts thus
typically describe the relationship between sense organs and their objects
using language that Indian Buddhists, and we ourselves, would more normally
associate with the active agency of a human mind or person. Sense organs
themselves are “fond of” (hao 好) their objects, they “desire” (yu 欲) them,
76 T.33.1694: 11a16–17. The Indian commentarial tradition does eventually feel the need to go
beyond the basic list of the pratītyasamutpāda formula and explain how sensory con-
sciousness and the material elements of perception (the organs and their objects) can
possibly come together at all. The need for such an explanation seems to arise from, and
reflect, the deep Buddhist intuition that consciousness cannot arise merely on the basis
of material causes (Griffiths, 1986). Thus in many commentarial traditions sparśa is said
to depend not only on the sense organs and their objects (as one might expect if sense
activity were an entirely “passive” affair), but also on what we might call “attention,” often
manasikāra or, in a term found as far as I know only in the Pāli commentaries, “adverting”
(āvajjana; for a discussion of these issues, see Sarachchandra, 1958: 26–27). However
despite this recognition that the causes and conditions of sense experience must also
include ontologically mental factors, those factors ultimately posited in such a role, such
as “adverting,” are never explained as karmically productive, and they are strictly distin-
guished from defilements such as craving.
77 This contrast between the “active” and “passive” members of the chain is even encoded
grammatically, as past participle versus agentive noun, within the Abhidharmasamuc-
caya’s categories of “projected members” (ākṣiptāṅga) versus “actualizing members”
(abhinirvartakāṅga; see above note 34)
“know” (zhi 知) them, or “long for” (si 思) them.78 Human desire for sense
objects is, in this view, a matter of the sense organs themselves desiring their
objects – the “eyes are fond of beauty” (mu hao se 目好色)79 – not a matter of
the mind desiring something seen, heard, or otherwise sensed with the organ
in question (grammatically, Indian Buddhist texts invariably phrase sense per-
ception by placing the organ of perception in the instrumental case).
In pre-Buddhist Chinese understanding the sense organs themselves were
thus thought of as active, willful agents. As Mark Edward Lewis puts it:
Unlike Western philosophy, which treats the senses as suspect because

unreliable in their perceptions, early Chinese texts usually treat them as
dangerous in their reckless desire for sensual stimulation […] [early
Chinese texts] portray these [sense organs] as agents with their own
motives that they will recklessly pursue unless forced by the mind to
serve the collective good.80
The independence of the desires of the sense organs is an integral part of the
standard metaphor for the body in early China, that of a state in which the
sense organs are the “officials” (guan 官) and the mind is the ruler.81 And as
Lewis again observes, “the political image of the body suggests that certain
parts of the body would, like certain officials, do as they pleased and create
chaos if not rigorously controlled”.82
In other words, in early China the senses were not seen merely as windows
onto the world, but as localized embodiments of particular forms of desire, as
independent centers of willful activity. In keeping with this understanding, the
normal way of discussing the problem of desire was by arguing that one willful
agent, the ruler (the mind), must be strong enough to control the various other
willful agents that are dispersed throughout the body. This model is, I would
suggest, fundamentally different from the typical Indian Buddhist “solution” to
this problem, in which there is only ever a single willful agent (usually the mind)
78 Geaney, 2002: 51. For “to long for” (si 思), see Xiang Xiu’s 向秀 (211–300) Nan Ji Shuye yang
sheng lun 難嵇叔夜養生論: “the mouth longs for the five flavors, the eye for the five
colors” (kou si wu wei, mu si wu se 口思五味，目思五色; QSGW, 2:1876). That si 思 also
figures among An Shigao’s translations of sparśa strongly suggests that he was, at least
occasionally, borrowing the vocabulary used in typical Chinese discussions of the sense
organs.
79 Xunzi ji jie: 338.
80 Lewis, 2006: 39.
81 Sato, 2003: 154–161.
82 Lewis, 2006: 40.
272 Greene
who must restrain the desires that arise in response to sensory perception.
This is the model that is encoded in the pratītyasamutpāda formula, and it is a
model that is reflected in – that is in fact enabled by – the conceptual distinc-
tions afforded by vocabulary such as āyatana, sparśa, vedanā, and tṛṣṇā.83
The extent to which the pre-Buddhist Chinese model of the senses and their
relationship to desire influenced An Shigao’s choice of translation terminology
must remain a topic for further research. What we may say at this point is that,
intentionally or not, certain ambiguities An Shigao introduced into his transla-
tions of key members of the pratītyasamutpāda formula potentially brought
the Buddhist model closer to the Chinese one by associating sparśa with kar-
mically productive desires and defilements. And this seems to have encouraged
or enabled early Chinese readers of An Shigao’s translations to make sense of
these arguably fundamental Buddhist classification systems by drawing from
established Chinese models of human desire and its relationship to sensory
perception.
It is important to again reiterate the still primitive state of our knowledge of
An Shigao’s doctrinal background. Our broader grasp of post-canonical Indian
Buddhist commentarial traditions is, of course, considerably better. Here too,
however, we must acknowledge that the relevant surviving material, from
which we can derive what I have called the “normative” Indian understanding
of a topic such as pratītyasamutpāda, is at most only a fraction of what once
circulated. It is thus far too early to judge which elements of the peculiar inter-
pretations found in a text such as the YCRJZ are best seen as distinctly Chinese
and which elements might simply accurately reflect, albeit using Chinese
vocabulary, An Shigao’s own understanding of the relevant Buddhist concepts
(an understanding perhaps quite different from what we might have excepted
for an Indian Buddhist). Nevertheless, as I have hinted throughout this chapter,
the commentaries of An Shigao’s Chinese followers, and even in some cases An
Shigao’s translations themselves, can be profitably examined as stemming
from a distinctly Chinese environment. This material can be fully explained
only as the product of the complex and multifaceted, yet still very little under-
stood process of negotiation, insight, confusion, surprise, and misunderstanding
– as well as genuine cross-cultural communication – that accompanied the
very earliest (and ultimately, perhaps, all) translations of Indian Buddhist texts
into Chinese.
83 In a greatly modified form, this same intuition seems even to underlie the much later
theories of Dignāga and Dharmakīrti concerning the non-conceptual (kalpanāpoḍha)
nature of sense perception.
Abbreviations
AKBh Abhidharmakośabhāṣyam of Vasubandhu (ed. Pradhan, 1975)

ADhS Abhidharmasamuccaya (ed. Gokhale, 1947)
AN Aṅguttaranikāya (ed. Morris and Hardy, [1885–1900] 1999–2013)
DhS-A Atthasālinī (rev. ed. Müller, 1979)
MN Majjhimanikāya (ed. Trenckner and Chalmers, 1888–1902)
Peṭ Peṭakopadesa (ed. Barua, 1982. References to the page numbers of this
edition, followed by the paragraph numbers of Nāṇamoli, 1964)
QSGW Quan shang gu Sandai Qin Han Sanguo Liuchao wen 全上古三代秦漢
三國六朝文 (ed. Yan, 1958)
SN Saṃyuttanikāya (ed. Feer, 1884–1898)
SMJ Shi’er men jing 十二門徑 (as reproduced in Ochiai 2004). Cited by line
number(s).
T Taishō shinshū daizōkyō 大正新脩大藏經 (CBETA edition). Based on
(with corrections) Taishō shinshū daizōkyō 大正新脩大藏經 (ed.
Takakusu 高楠, Watanabe 渡邊, and Ono 小野, 1924–1934)
ZHDZJ Zhonghua da zang jing 中華大藏經. Beijing: Zhonghua shuju,
1984–1996.
GHYZD Wang Li gu Hanyu zidian 王力古漢語字典 (ed. Wang, et al., 2002)
HYDCD Hanyu da cidian: suoyinben 汉语大词典: 缩印本 (ed. Luo, 2000)
Morohashi Morohashi Tetsuji 諸橋轍次. Dai Kan-Wa jiten 大漢和辞典. 13 vols.
Tokyo: Taishukan shoten, 1955–1960.
Vism Visuddhimagga (ed. Rhys Davids, 1975)
Bibliography
Barua, Arabinda (ed.) (1964). Peṭakopadesa. London: The Pāli Text Society.
Bo shu Laozi jiao zhu 帛書老子校注 (1996). Gao Ming 高明 (ed.). Beijing: Zhonghua
shuju.
Bodhi Bhikku (2000). The Connected Discourses of the Buddha. Somerville MA: Wisdom
Publications.
Bronkhorst, Johannes (1986). The Two Traditions of Meditation in Ancient India. Stuttgart:
Franz Steiner Verlag.
Bucknell, Roderick S. (1999). “Conditioned Arising Evolves: Variation and Change in
Textual Accounts of the Paticca-Samuppāda Doctrine”, Journal of the International
Association of Buddhist Studies 22/2, pp. 311–342.
274 Greene
Cheng, Anne (1999). “Émotions et Sagesse dans La Chine Ancienne: L’Élaboration de la

Notion de Qing dans les Textes Philosophiques des Royaumes Combattants jusqu’aux
Han”, Études Chinoises 18 1/2, pp. 31–58.
Cox, Collett (1988). “On the Possibility of a Non-Existent Object of Consciousness”,
Journal of the International Association of Buddhist Studies 11/1, pp. 31–87.
Cox, Collett (1993). “Dependent Origination: Its Elaboration in Early Sarvāstivādin
Abhidharma Texts”, In Researches in Indian and Buddhist Philosophy: Essays in Honour
of Professor Alex Wayman. Edited by Ram Karan Sharma. Delhi: Motilal Banarsidass,
pp. 119–142.
Deleanu, Florin (2003). “The Newly Found Text of the An Ban Shou Yi Jing Translated by
An Shigao”, Journal of the International College for Postgraduate Buddhist Studies 6,
pp. 170–189.
Dhammajoti, Bhikkhu KL (2007). Abhidharma Doctrines and Controversy on Perception.
Hong Kong: Centre of Buddhist Studies, The University of Hong Kong.
Feer, Leon (ed.) (1884–1898). Saṃyuttanikāya. London: The Pāli Text Society.
Forte, Antonino (1995). The Hostage An Shigao and His Offspring. Kyoto: Istituto Italiano
di Cultura. Scuola di Studi sull’Asia Orientale.
Geaney, Jane (2002). On the Epistemology of the Senses in Early Chinese Thought.
Honolulu: University of Hawaii Press.
Gokhale, V.V. (ed.) (1947). Abhidharmasamuccaya: “Fragments from the Abhidharmasa
muccaya of Asaṃga,” Journal of the Royal Asiatic Society: Bombay Branch 23 [1979
reprint], pp. 13–38.
Graham, Angus C. (1990). Studies in Chinese Philosophy and Philosophical Literature.
Albany NY: State University of New York Press.
Griffiths, Paul J. (1986). On Being Mindless: Buddhist Meditation and the Mind-Body
Problem. La Salle IL: Open Court.
Hansen, Chad (1995). “Qing 情 in Pre-Buddhist Chinese Thought”, In Emotions in Asian
Thought. Edited by Joel Marks and Roger T. Ames. Albany NY: State University of New
York Press, pp. 181–212.
Harbsmeier, Christoph (2004). “The Semantics of Qing in Pre-Buddhist Chinese”, In Love
and Emotions in Traditional Chinese Literature. Edited by Halvor Eifring. Leiden: Brill,
pp. 69–148.
Harrison, Paul (2002). “Another Addition to the An Shigao Corpus? Preliminary Notes
on the Early Chinese Saṃyuktāgama Translation”, In Studies in Honour of Dr. Sakurabe
Hajime on the Occasion of His Seventy-Seventh Birthday. Edited by Sakurabe Ronshu
Committee. Kyoto: Heirakuji Shoten, pp. 1–32.
Harrison, Paul (1997). “The Ekottarikāgama Translations of An Shigao”, In Bauddha
vidyāsudhākaraḥ: Studies in Honour of Heinz Bechert on the Occasion of His 65th
Birthday. Edited by Petra Kieffer-Pülz and Jens-Uwe Harmann. Swisttal-Odendorf:
Indica et Tibetica Verlag, pp. 261–284.
Hartmann, Gerda (1940). “Symbols of the Nidānas in Tibetan Drawings of the Wheel of
Life”, Journal of the American Oriental Society 60, pp. 356–360.
Hu, Chirui 胡敕瑞 (2002). Lun Heng yu Dong Han Fodian ciyu bijiao yanjiu 《論衡》與
東漢佛典詞語比較研究. Chengdu: Bashu shushe.
Johansson, Rune E.A. (1979). The Dynamic Psychology of Early Buddhism. London:
Curzon Press.
Jong, Jan Willem de (1979). “The Daśottarasūtra”, In Buddhist Studies by J.W. De Jong.
Edited by Gregory Schopen. Berkeley CA: Asian Humanities Press, pp. 251–273.
Jurewicz, Joanna (2000). “Playing with Fire: The Pratītyasamutpāda from the Perspective
of Vedic Thought”, Journal of the Pāli Text Society 26, pp. 77–104.
Kritzer, Robert (1999). Rebirth and Causation in the Yogācāra Abhidharma. Wien:
Arbeitskreis für Tibetische und Buddhistische Studien, Universität Wien.
La Vallée Poussin, Louis de (1913). Bouddhisme, Études et Matériaux, Théorie des Douze
Causes. Londres: Luzac & Co.
Lewis, Mark Edward (2006). The Construction of Space in Early China. Albany NY: State
University of New York Press.
Lunheng jiao shi 論衡校釋. Huang, Hui 黄暉 (ed.) (1990). Beijing: Zhonghua shuju.
Luo Zhufeng 罗竹风 (ed.) (2000). Hanyu da cidian: suoyinben 汉语大词典: 缩印本.
Beijing: Zhonghua shuju.
McDermott, James P. (1980). “Karma and Rebirth in Early Buddhism”, In Karma and
Rebirth in Classical Indian Traditions. Edited by Wendy Doniger O’Flaherty. Berkeley
CA: University of California Press, pp. 165–192.
Morohashi, Tetsuji 諸橋轍次 (1955–1960). Dai Kan-Wa jiten 大漢和辞典. 13 vols. Tokyo:
Taishukan shoten.
Morris, Richard and Hardy, Edmund (ed.) ([1885–1900) 1999–2013). Aṅguttaranikāya. 5
vols. London: The Pāli Text Society.
Müller, Edward (ed.) (1979). Atthasālinī. Revised edition. London: The Pāli Text Society.
Nakamura, Hajime (1980). “The Theory of Dependent Origination in Its Incipient Stage”,
In Buddhist Studies in Honor of Walpola Rahula. Edited by André Bareau. London:
Gordon Fraser, pp. 165–172.
Ñāṇamoli Bhikkhu (1964). The Piṭaka-Disclosure (Peṭakopadesa). London: The Pāli Text
Society.
Nattier, Jan (2008). A Guide to the Earliest Chinese Translations: Texts from the Eastern
Han and Three Kingdoms Periods. Tokyo: The International Research Institute for
Advanced Buddhology.
Ng, Ong-cho (1998). “Is Emotion (Qing) the Source of a Confucian Antinomy?”, Journal
of Chinese Philosophy 25/2, pp. 169–190.
Ochiai Toshinori 落合俊典 (2004). Kongōji Issaikyō no Kisoteki Kenkyū to Shinshutsu
Butten no Kenkyū: Heisei 16-nen 3-gatsu 金剛寺一切経の基礎的研究と新出仏典の
276 Greene
研究: 平成 16年 3月. Tokyo: International College of Postgraduate Buddhist

Studies.
Pradhan, Prahallad (ed.) (1975). Abhidharmakośabhāṣyam of Vasubandhu. (Tibetan
Puett, Michael (2004). “The Ethics of Responding Properly: The Notion of Qing 情 in
Early Chinese Thought”, In Love and Emotions in Traditional Chinese Litterature.
Edited by Halvor Eifring. Leiden: Brill, pp. 37–68.
Rhys Davids, Caroline A.F. (ed.) (1975). Visuddhimagga. London: The Pāli Text Society.
Sarachchandra, E.R. (1958). Buddhist Psychology of Perception. Dehiwala (Sri Lanka):
Buddhist Cultural Centre.
Sato, Masayuki (2003). The Confucian Quest for Order: The Origin and Formation of the
Political Thought of Xun Zi. Leiden: Brill.
Shulman, Eviatar (2008). “Early Meanings of Dependent-Origination”, Journal of Indian
Philosophy 36, pp. 291–317.
Stalker, Susan C. (1987). “A Study of Dependent Origination: Vasubandhu, Buddhagosa,
and the Interpretation of Pratītyasamutpāda”. Ph.D. dissertation, University of
Pennsylvania.
Taber, John A. (2005). A Hindu Critique of Buddhist Epistemology: Kumārila on Perception:
The “Determination of Perception” Chapter of Kumārila Bhaṭṭta’s Ślokavārttika.
London: RoutledgeCurzon.
Taishō shinshū daizōkyō 大正新脩大藏經 (CBETA edition). Based on (with corrections)
Taishō shinshū daizōkyō 大正新脩大藏經. Edited by Takakusu Junjirō 高楠順次郎,
Watanabe Kaigyoku 渡邊海旭, and Ono Gemmyō 小野玄妙. Tokyo: Taishō issaikyō
kankōkai, 1924–1934.
- T.1.6: Mahāparinirvāṇasūtra, Bannihuan jing 般泥洹經.
- T.1.13: Chang Ahan shi bao fa jing 長阿含十報法經 (Daśottarasūtra), An Shigao 安
世高.
- T.1.14: Mahānidānasūtra, Ren ben yu sheng jing 人本欲生經, An Shigao 安世高.
- T.1.57: Lou fen bu jing 漏分布經, An Shigao 安世高.
- T.2.150A: Qi chu san guan jing 七處三觀經, An Shigao 安世高.
- T.3.152: Liu du ji jing 六度集經, Kang Senghui 康僧會.
- T.3.184: Xiuxing benqi jing 修行本起經, Zhu Dali 竺大力 and Kang Mengxiang 康
孟詳.
- T.3.185: Taizi ruiying benqi jing 太子瑞應本起經, Zhi Qian 支謙.
- T.4.196: Zhong benqi jing 中本起經, Tanguo 曇果 and Kang Mengxiang 康孟詳.
- T.8.221: Pañcaviṃśatisāhasrikāprajñāpāramitā, Fangguang bore jing 放光般若經,
Mokṣala.
- T.14.474: Vimalakīrtinirdeśa, Weimojie jing 維摩詰經, Zhi Qian 支謙.
- T.15.602: Da anban shouyi jing 大安般守意經.
- T.15.603: Yin chi ru jing 陰持入經, An Shigao 安世高 (YCRJ).
- T.15.607: Saṃgharakṣa, Yoga[ācāra]bhūmi. Dao di jing 道地經, An Shigao 安世

高.
- T.16.708: Ṥalistamb[h]a[ka]sūtra, Liao ben sheng si jing 了本生死經.
- T.25.1508: Ahan koujie shi’er yinyuan jing 阿含口解十二因緣經 (AHKJ).
- T.33.1694: Yin chi ru jing zhu 陰持入經註 (YCRJZ), Chenhui 陳慧
- T.54.2128: Yiqie jing yinyi 一切經音義, Huilin 慧琳.
Tang Yongtong 湯用彤 ([1957] 2001). Wei Jin xuanxue lun gao 魏晉玄學論稿. Reprint.
Taipei: Foguang shuju.
Trenckner, Vilhelm and Chalmers, Robert (ed.) (1888–1902). Majjhimanikāya. London:
Van Den Broeck, José (1977). La Saveur de l’Immortel. Louvain-la-Neuve: Institut
Orientaliste Louvain-la-Neuve.
Vetter, Tilmann (2012). A Lexicographical Study of An Shigao’s and His Circle’s Chinese
Translations of Buddhist Texts. Tokyo: The International Institute for Buddhist
Studies.
Vetter, Tilmann and Paul Harrison (1998). “An Shigao’s Chinese Translation of the
Saptasthānasūtra”, In Sūryacandrāya: Essays in Honour of Akira Yuyama on the
Occasion of His 65th Birthday. Edited by Paul Harrison and Gregory Schopen. Swisttal-
Odendorf: Indica et Tibetica Verlag, pp. 197–216.
Wang, Baoxuan 王葆玹 (1987). Zhengshi xuanxue 正始玄學. Jinan: Qilu shushe.
Wang, Li 王力 et al. (ed.) (2002). Wang Li gu Hanyu zidian 王力古漢語字典. Beijing:
Zhonghua shuju.
Wu, Guanhong 吳冠宏 (2009). Wei Jin xuan lun yu shi feng xin tan 魏晉玄論與士風新
探. Taipei: Huamulan wenhua chubanshe.
Xu, Fuguan 徐福觀 (2001). Liang Han sixiang shi 兩漢思想史. Shanghai: Huadong shifan
daxue chubanshe.
Xunzi ji jie 荀子集解. Wangx Xianqian 王先謙 (ed.) (1988). Beijing: Zhonghua shuju.
Yan Kejun 嚴可均 (1958). Quan shang gu Sandai Qin Han Sanguo Liuchao wen 全上古
三代秦漢三國六朝文. Beijing: Zhonghua shuju.
Zacchetti, Stefano (2002). “An Early Chinese Translation Corresponding to Chapter 6 of
the Peṭakopadesa”, Bulletin of the School of Oriental and African Studies 65/1,
pp. 74–98.
Zacchetti, Stefano (2004). “Teaching Buddhism in Han China: A Study of the Ahan Koujie
Shi’er Yinyuan Jing T1508 Attributed to An Shigao”, Journal of the International
Research Institute for Advanced Buddhology at Soka University 7, pp. 197–225.
Zacchetti, Stefano (2007). “Inventing a New Idiom: Some Aspects of the Language of
the Yin Chi Ru Jing 陰持入經 T603 Translated by An Shigao”, Annual Report of the
pp. 395–417.
278 Greene
Zacchetti, Stefano (2008). “The Nature of the Da Anban Shouyi Jing 大安般守意經 T602
Reconsidered”, Journal of the International Association of Buddhist Studies 31/1–2,
pp. 421–484.
Zacchetti, Stefano (2010a). “Defining An Shigao’s 安世高 Translation Corpus: The State
of the Art in Relevant Research”, In Xiyu lishi yuyan yanjiu jikan disan ji 西域歷史語
言研究集刊第三輯. Edited by Shen, Weirong 沈衛榮. Beijing: Kexue chubanshe,
pp. 249–270.
Zacchetti, Stefano (2010b). “Some Remarks on the Authorship and Chronology of the
Yin Chi Ru Jing Zhu 陰持入經註 T 1694 – the Second Phase in the Development of
Early Chinese Buddhist Exegetical Literature”, In Buddhist Asia 2. Edited by Silvio
Vita. Kyōto: Italian School of East Asian Studies, pp. 141–198.
Zürcher, Erik ([1959] 1972). The Buddhist Conquest of China. The Spread and Adaptation
of Buddhism in Early Medieval China. Reprint. Leiden: Brill.
Zürcher, Erik (1979). “Late Han Vernacular Elements in the Earliest Buddhist Translations”,
Journal of the Chinese Language Teachers Association 12/3, pp. 177–203.
Abhidharma in China 279
Chapter 8
Abhidharma in China: Reflections on ‘Matching

Meanings’ and Xuanxue
Bart Dessein
1 Buddhism Entering a Confucian World
The earliest direct reference to the introduction of Buddhism in China is the

well-known account that relates how Emperor Ming 明 (r. 58–75 CE) of the
Han 漢 dynasty (206 BCE–220 CE) in a dream saw a golden deity flying in front
of his palace. When Fu Yi 傅毅, one of his ministers, had explained to him that
this deity was the Buddha, he is reported to have dispatched two envoys –
Zhang Qian 張騫 and Qin Jing 秦景 – abroad to learn more about this sage and
his teachings.1 Although this account is now recognized as apocryphal, during
the reign of this Emperor Ming, a Buddhist community of (foreign) monks and
Chinese layman must indeed have been present in the region of Pengcheng 彭
城, the most important city of the country of Chu 楚, south of the Yellow River.
This can be deduced from a decree Emperor Ming issued in 65 CE in which the
Buddha is mentioned with respect to the king of this country of Chu, i.e. Liu
Ying 劉英, one of the sons of Emperor Guangwu 光武 (r. 25–58 CE), Emperor
Ming’s predecessor.2 As Buddhism had been brought to China in the first cen-
tury BCE by merchants and pilgrims who traveled over the land routes that
crossed Central Asia into China, this decree has to postdate the actual intro-
duction of Buddhism in China with about one and a half centuries.3 This
means that Buddhism entered and traveled through China in a period in which
Confucianism was the official doctrine that was adhered to by the literati class.
1 See Zürcher, 2007: 22; Ch’en, [1972] 1973: 29–30; Nattier, 2008: 35. The accounts of the dream
of Emperor Ming derive from the “Preface to the Sutra in Forty-two Sections” (Sishi’er zhang
jing xu 四十二章經序) which was, in the sixth century, incorporated in Sengyou’s 僧祐 Chu
sanzang ji ji 出三藏記集 (T.55.2145: 42c15–19). As noted by Storch, 2014: 25, “The authorship
and the date of this sutra have been debated since the beginning of the twentieth century”.
For a first major study on this legend: See Maspéro, 1910. For more legends concerning the
introduction of Buddhism in China: see Ch’en, [1972] 1973: 27–29.
2 See Zürcher, 2007: 26–27.
3 See Loewe, 1986: 670–671; Zürcher, 2007: 26.

280 Dessein
The variant of Confucianism that had been elevated to the status of state
orthodoxy in 136 BCE was an interpretation of the early doctrine that was
highly influenced by numerological and ‘Daoist’ concepts.4 This highly syncre-
tistic from of Confucianism was called Jin wen jia 今文家 (New Text School
Confucianism).5 The importance of this peculiar type of Confucianism as
orthodoxy for the development of Buddhism in China is twofold. A first ele-
ment of importance is that this doctrine developed to be a part of Chinese
self-identification. By the time of the Zhou 周 dynasty (1046–256/221 BCE), a
hierarchical feudal order headed by the ‘Son of Heaven’ (tianzi 天子) had
developed in the central plains (zhongyuan 中原) along the Yellow River. This
‘Son of Heaven’ was thought to rule by divine mandate (tianming 天命).6 When
Confucianism was established as state doctrine, this Zhou concept of rulership
was given a Confucian interpretation: the now Confucian rule was seen as a
divine undertaking, and ‘governing’ was interpreted as ‘maintaining the har-
monious order between Heaven, Man and Earth’.7 This is clear from the
explanation of the character for ‘wang’ 王, ‘ruler’ (generally translated as ‘king’)
given in Xu Shen’s 許慎 (ca. 58–ca. 147) Shuo wen jie zi 說文解字, the oldest
extant etymological dictionary of the Chinese language, dated 100 CE. Here we
read that, according to Dong Zhongshu who had been a major exponent of the
development of New Text School Confucianism, it is so that “when depictions
(wen 文) were created in olden [times], three strokes that were connected
through the middle were called wang. The three are Heaven, Earth and Man,
and the one who connects them is the ‘ruler’ (wang)”.8 This concept also
4 Dull, 1994: 3 remarks that the reign of Emperor Wu 武 (140–87 BCE) of the Han dynasty is the
period in which Confucianism for the first time was recognized as the ‘ism,’ “to the exclusion
of all others, that was to be acceptable to the state and was to become the object of study for
those who hoped for official careers”. Schwartz, 1985: 377 calls Emperor Wu of the Han the
‘Constantine of Confucianism,’ who followed the momentous advice of Dong Zhongshu 董
仲舒 (179?–104? BCE) and others to establish the “five classics” of Confucianism as the founda-
tion of all official education and to proscribe all “unorthodox” doctrines. The term ‘daojia’ 道
家, ‘Daoist school’ only started to be used in the second century BCE. See Cheng, 1997: 103.
5 On the element ‘jia,’ ‘school’: see further and p. 11 note 48.
6 According to Michèle Pirazzoli-t’Serstevens, [2008] 2009: 302, although the doctrine of the
Mandate of Heaven (tianming) had not been taken into account before Emperor Chengdi 成
帝 (r. 33–7 BCE) or as full-fledged theory before the time of the usurper Wang Mang 王莽 (r.
5–23 CE), the concept had been operative even before the first century BCE.
7 Yu, 2005: 34 remarks that the politico-religious narrative of Confucianism, in fact, builds on
the system that was developed already in the Shang 商 dynasty (trad. 1766–1122 BCE), when
ancestors were transformed from kin to symbols of divine power.
8 Xu, [1981] 1988: 7b. This also explains why Han Confucianism has been labeled ‘cosmological
Confucianism’ by Joseph Needham (1958: 281–282).
implies that it is from Heaven that the ruler, the ‘Son of Heaven,’ obtains his
mandate to rule, and that it therefore also is Heaven that can, ultimately, with-
draw this mandate.9 Phrased differently, it is through divine approval of his
moral virtue – philosophically articulated in the Confucian concept ren 仁,
‘humaneness’ – that the ‘Son of Heaven’ maintains his legitimacy to rule.10 The
connection between the indigenous Confucian political ideology and the age-
old acceptance of a divine origin of the ‘Chinese’ people made it possible that
the inhabitants of the central plains imagined themselves and the territory
they inhabited (that is: Zhongguo 中國) as fundamentally different from the
people, creeds and customs of the regions surrounding the central plains.11
A second element of importance is that Confucian state orthodoxy in its
New Text School variant was, as mentioned, a syncretistic doctrine. When Han
imperial power was usurped by Wang Mang 王莽 in the first decade CE, a
period of gradual decline of central authority set in. This development must
have come as a disenchantment for the Confucian literati for whom the uni-
fied Confucian empire had gradually become perceived as normative. On a
political level, such new independent religious groups as the Wudoumi dao 五
斗米道 (Way of the Five Pecks of Grain) or the Taiping dao 太平道 (Way of
Great Peace) presented themselves as an alternative for the Confucian order.12
On an ideological level, these groups even managed to develop Daoism into a
unified state-sponsored religion, capable to compete with Buddhism. As the
Chinese intellectuals were, through the New Text School interpretation of
Confucianism, acquainted with Daoist concepts, it was possible for them to
reinterpret these concepts within a Confucian framework. This intellectual
9 See on this Schwartz, 1985: 23.

10 For a discussion of ren: see Schwartz, 1985: 75–85.
11 The divine origin of the Chinese people is described in the ode ‘Shengmin’ 生民 (The
birth of [our] people) of the Mao shi 毛詩 version of the Shijing 詩經 (Odes). That the
‘Prefaces’ to this Mao shi 毛詩 which were, when not authored in the transitional period
from Western to Eastern Han (i.e. ca. 50 BCE–50 CE), then at least expanded and revised
in that period, present the Shijing as one single history of the Zhou from its beginning in
the eleventh century BCE up to 599 BCE, a history of Zhongguo, inhabited by the people of
Zhou, and surrounded by barbarian peoples, illustrates this. In its approach, the Mao shi
version of the Odes therefore is a major rupture with earlier historical works such as the
Chunqiu 春秋 (Spring-and-Autumn Annals) or the Shujing 書經 (Documents). See Mittag,
[2008] 2009: 151–153.
12 An important event that signaled the successful creation of a state-sponsored Daoist
orthodoxy is Lu Xiujing’s 陸修靜 (406–477) collection and catalogue of the then scat-
tered Daoist scriptures, undertaken in response to an imperial order in AD 471. See Jansen,
[2008] 2009: 399–400.
282 Dessein
development was referred to as xuanxue 玄學 (‘mysterious learning’ or ‘study

of the mystery’) in the Chinese tradition, and has become known as Neo-
Daoism in Western literature. Furthermore, this era has also become known as
the era of qingtan 清談 (‘pure conversation’) culture, a phenomenon in which
former Confucian literati engaged themselves in witty conversations on non-
political subjects, indulging themselves in showing their literary prowess and
knowledge of Confucian, Daoist, and Buddhist works.13 The latter may explain
why a watered down New Text School Confucianism opened the way for the
Chinese intellectuals to shift their interest away from traditional Chinese phi-
losophies, toward Buddhism. It can in this respect be noted that the above
mentioned Liu Ying was, according to the Hou Han shu 後漢書, deeply inter-
ested in Daoism (Huanglao 皇老) and at the same time “observed fasting and
performed sacrifices to the Buddha”.14 Also Emperor Huan 桓 (r. 147–167) of
the Han dynasty is reported to have performed joint sacrifices to Laozi and the
Buddha – to all likelihood a basically Daoist ritual to which some Buddhist ele-
ments were added.15 Also illustrative for such a Confucian-Daoist-Buddhist
syncretism, is the memorial the scholar Xiang Kai 襄楷 submitted to the throne
in 166 CE, in response to the just mentioned Daoist-Buddhist ritual. In this
memorial, we also read that “Some people say that Laozi 老子 has gone into
the region of the barbarians and (there) has become the Buddha” – the famous
huahu 化胡 theory.16
2 Early Abhidharma Translations and the Chinese Abhidharma

School
Among the first translations of Buddhist scriptures into Chinese as the out-
come of a systematized translation project, we not only find popular sūtras on
the life of the Buddha or jātaka (birth) stories that were rendered in a simple
language,17 but also translations of the highly technical and intellectual
Abhidharma treatises. This hints to it that not only the common people had
become interested in Buddhism, but that also the former (New Text School)
13 See Jansen, [2008] 2009: 403.

14 Hou Han Shu, 72.4b.
15 Hou Han Shu, 7.15a. See also Zürcher, 2007: 36.
16 Hou Han Shu, 60B.18b. Translation Zürcher, 2007: 37. On the huahu theory and the ‘con-
version of the barbarians’: see Zürcher, 2007: 288–320.
17 See Mizuno [1982] 1995: 47–48.
Confucians had taken up an interest is this new doctrine and were in need of
Chinese versions of texts of this erstwhile ‘foreign’ creed in China.18
It is An Shigao 安世高 (fl. 148–168), a Parthian prince who renounced the
throne he was supposed to inherit from his father and who came to Luoyang 洛
陽 in the mid-second century CE, who is credited with the first Abhidharma
translations.19 According to Sengyou’s 僧祐 Chu sanzang ji ji 出三藏記集, it
more precisely is An Shigao’s translation of the Apitan wufa xing jing 阿琵曇五
法行經, done during the reign of Emperor Huan, that introduced the knowl-
edge of the Abhidharma in China.20
It would take a few centuries, however, before Abhidharma studies started
to gain popularity. Especially the period of Emperor Fu Jian 符堅 (357–384) of
the Former Qin 前秦 dynasty (351–394), one of the sixteen kingdoms that ruled
in Northern China after the Han dynasty had finally collapsed in 220 CE, was
important for the take-off of Abhidharma translation activities. One of the
translators who were active in the capital city Chang’an 長安, is the famous
Saṃghadeva. During the years of Emperor Fu Jian, he and Zhu Fonian 竺佛念
together had started the translation of the *Aṣṭagrantha (Apitan ba jiandu lun
阿毘曇八犍度論, T.26.1543), one of the major early Sarvāstivāda Abhidharma
works.21
After political turmoil in fourth century North China had urged the Chinese
intelligentsia to move to the South, Lushan 廬山 in Jiangxi 江西, South China,
gradually developed to be an important center of Abhidharma studies.22 It was
especially the arrival of Saṃghadeva on Mt. Lu that stimulated the study of
the Abhidharma.23 It is reported that Saṃghadeva’s lectures on Abhidharma
attracted huge audiences to Mt. Lu. Among these was his sponsor Wang Xun
王珣 (350–401), one of the most important dānapatis of the period.24 Wang
Xun belonged to the inner circles of Huan Wen 桓溫 of the Eastern Jin 晉
18 For reflections on the different styles of translation of Buddhist texts and their possible
relation to readership: See Nattier, 2008: 18.
19 See Zürcher, 1991: 282, 289. On the translations of An Shigao: see Mizuno, [1982] 1995: 45;
Tsukamoto, 1985: 78–112; Zacchetti, 2010. See the contribution in this volume by Eric M.
Greene.
20 T.55.2145: 6b4–5. Zürcher, 1991 does not list this text among what he considers to be genu-
ine translations by An Shigao.
21 See T.55.2154: 620 a23–24. On the *Aṣṭagrantha / Jn͂ ānaprasthāna: See Willemen et al.,
1998: 221–229.
22 See Tsukamoto, 1985: vol.1, pp. 401, 414–415, 444; vol.2, p. 760 ff.
23 See Dessein, 2010: 49–50.
24 See T.49.2034: 329a10–19, 361b23–27; T.50.2059: 329a15, 361b24; T.55.2145: 64a17. T.50.2059:
329a10–19, 361b23–27 also mentions Wang Min 王珉 (361–388), Wang Xun’s brother,
284 Dessein
dynasty (317–419), and of the Jin emperor Xiaowu 孝武 (r. 379–397).25 The
translation of Dharmaśreṣṭhin’s *Abhidharmahṛdaya (Apitan xin lun 阿毘曇
心論, T.28.1550) Saṃghadeva did together with Huiyuan 慧遠 (334–416) in 391
and his lecturing on this text, set off the flourishing of the Mt. Lu Abhidharma
School.26 Numerous monks studied the text and lectured on it,27 and the so-
called ‘group of Mt. Lu’ were all specialists in the *Abhidharmahṛdaya.28 The
translation of the *Abhidharmahṛdaya was followed by the final translation
of Dharmatrāta’s *Saṃyuktābhidharmahṛdaya (Za apitan xin lun 雜阿毘
曇心論, T.28.1552) by Saṃghavarman in 434.29 On this Chinese translation,
numerous commentaries were written.30 The Apitan xin lun and the Za api-
tan xin lun became the two most important texts of the Chinese Abhidharma
School into the Tang 唐 dynasty.31 The Abhidharma School was superseded
by the so-called Kośa School when Xuanzang 玄奘 translated Vasubandhu’s
Abhidharmakośa (Apidamo jushe lun 阿毘達摩俱舍論) in 653.32 From these
accounts of the Buddhist activities on Mt. Lu, we can thus deduce that also
the Chinese Abhidharma School that developed on Mt. Lu was shaped in the
pattern of traditional school formation in the Chinese cultural context: recit-
ing texts and accepting the teacher’s interpretation by the disciple.33 The third
among the audience on Mt. Lu. However, as Saṃghadeva arrived on Mt. Lu only in 397,
this Wang Min cannot be the brother of Wang Xun. See Mather, 1976, 124–125.
25 See Tsukamato, 1985, vol.1: 386. Huan Wen was a descendant from a low-level military
family that had risen to power after the Western Jin dynasty had moved to the South in 316
CE and had reestablished itself as Eastern Jin. In the South, they subdued the Daoist the-
ocracy that had been established there in the aftermath of the Wudoumi movement. See
Lewis, 2007: 203–204, and Lewis, 2009: 199–200. On Huan Wen: see Lewis, 2009: 64.
26 Saṃghadeva had already begun translating the*Abhidharmahṛdaya in 388, while still in
the North. See T.50.2059: 357c23–361b13.
27 See T.50.2059: 359b23–24, T.50.2060: 529c10, 599b23, 668a24–25, 675a23–24. See also: Des-
sein, 2010: 52–54.
28 See T.49.2035: 343a16–28.
29 See T.55.2145: 74c3–7. Also see Dessein, 1999,1: lxxvii-lxxxii.
30 See T.50.2059: 244a8–9, 368b28–29, 338a19, 369a23, 373a13, 373a22–23,373b18–20, 373c1,
375a4, 376a22, 376b28–29, 378b19–22, 381a25–26,382b23-c2, 401c3, 442b7, 460a26;
T.50.2060: 442a7–8, 461a24–25, 482c25–28, 483a11–12, 495c13–16, 501b8, 502a2, 504c25,
507c21, 508a12, 508c7–14, 509a6, 510b3–4, 524b27, 527b7, 529a6–7, 530a21, 532b27, 544b21–
22, 549c24, 572a19, 594c21. See also: Dessein, 2010: 5762.
31 See T.50.2059: 359b23–24.
32 See T.2149.311c12; T.2154: 557a22–23, 620c22–23, 695c10, 720b6. Also see Watanabe, Mizuno,
and Ōishi, 1932: 123.
33 It is also with Mt. Lu that Pure Land (Jingtu 淨土) Buddhism is connected. The fourth
century thus also is the period in which original Chinese ‘schools’ (jia) of Buddhism
started to develop. 
century BCE Lüshi chunqiu 呂氏春秋 (Spring-and-Autumn of the Lü Clan) testi-

fies of this traditional organizational pattern in the following words:
In studying, one has to progress, so that there is no blindness in the mind.

Recite [the texts] carefully. […] Observe whether [the teacher] is happy,
and [if so] ask him the meaning of the texts. Make your ears and eyes
obedient, and do not contradict his intention. Retreat from him, and
think about what he has said.34
When the highly technical Abhidharma texts had to be translated into Chinese,
equivalents for the Buddhist terms and concepts had to be found. As stated by
Whalen Lai (1979: 238–239):
[I]n any initial cultural encounter, epistemic translations of things for-

eign into things familiar are the norm […] all human understanding is
[…] an endless appropriation of new ideas by relying on the flexibility of
the old.
The early translators did not all follow the same translation technique. While
the already mentioned An Shigao preferred to transliterate proper names, he
sought the best matching equivalent in Chinese philosophical vocabulary for
Buddhist terms. The Yuezhi 月氏 Lokakṣema and his disciple Zhi Qian 支謙
(born in China in the third century CE) preferred to transliterate Indian techni-
cal terms instead of translating them.35 The Parthian An Xuan 安玄 (fl. 181)
34 Ershi’er zi 二十二子, [1986] 1988: 640a.

35 See Zürcher, 1991: 279–283; Harrison, 1993: 140, note 5; Nattier, 2008: 75, 120, 151; Mair, 2012:
55. Lokakṣema adopted many terms previously introduced by An Shigao. See Nattier,
2008: 75. On the use of a vernacular vs. literary style of translations and the possible read-
ership of these translations: see Harrison, 1993: 138; Nattier, 2008: 17–18. On the relation
between this vernacular and the actual speech of the population of Luoyang at the time,
see Zürcher, 1977 and Zürcher, 1991: 282, who comments on the language of the earliest
Chinese translations of Buddhist texts, as follows: “We can conclude that the language of
the earliest Buddhist translations is based on a substrate that in a somewhat formalized
way reflects the second century spoken language of the metropolitan area. This substrate
forms a clearly recognizable system that deviates from the literary language in a very con-
sistent way,” and, p. 289, that “in the context of the urban society of Later Han times, one
may think of a sub-elite of clerks and copyists, the lowest fringe of the bureaucracy, and
traders and artisans”.
286 Dessein
attempted to not only translate the doctrinal terms, but also the proper names.36
While, on the one hand, opting for transliterations may point to it that the
early translators were very cautious to preserve the original meaning of the
Buddhist texts and tried to avoid that rendering a Buddhist term with a Chinese
term that did not fully express the Buddhist term’s original meaning would
lead to misunderstandings of the Buddhist doctrine, the tremendous variety of
Chinese translations we find for one and the same technical term may equally
reveal the struggle of the individual translator to give the most perfect Chinese
term. The latter can be illustrated by the eight different translations and tran-
scriptions Zhi Qian gives for the term ‘arhat’: aluohan 阿羅漢, aluohe 阿羅訶,
luohan 羅漢, wusuozhuo 無所著, wuzhuo 無著, yingyi 應儀, yingzhen 應真,
zhenren 真人 and zhizhen 至真.37
Given the intellectual climate of second century China, it does not come as
a surprise that it was especially in the Daoist doctrine that equivalents for
Buddhist concepts and terms were sought. A well-known example is the use of
the word dao 道 to render the Buddhist term dharma (doctrine). This same
word dao was, however, also used to translate bodhi (enlightenment) or yoga.
The just mentioned term zhenren 真人 which was one of the words used to
translate arhat was, actually, a word used in Neo-Daoist philosophy to refer to
a Daoist immortal.38 Also the Buddhist final extinction (nirvāṇa), that is, the
state where all karmic activity has come to a halt, was easily paired with the
Daoist wuwei 無為 (non-action).39 The same wuwei was, however, also a com-
mon translation for the Buddhist notion of being ‘unconditioned’ (asaṃskṛta),
i.e. existing free from karmic influence. Not all equivalents were sought in the
Daoist sphere, though. In the translations done by Kang Senghui 康僧會 (third
century CE), a Sogdian who came to Luoyang toward the end of the reign of
Emperor Huan,40 we also find terms borrowed from popular religion and
Confucianism. Kang Senghui in this way refers to Taishan 太山 as recluse after
death, he discusses the fate of the hunling 魂靈 spirit(s), and he uses the
Confucian term ren 仁 (humaneness).41
Difficulties to find corresponding terms might have been even bigger in the
case of the ‘numerical lists’ (mātṛkā) of which the Abhidharma texts abound.
36 See Nattier, 2008: 91. On the translations by An Shigao: see the contribution by Eric M.
Greene in this volume.
37 See Nattier, 2008: 119, note 23.
38 See Wright, 1959: 36.
39 See Liu and Shao, 1992: 39–40.
40 T.55.2145: 95c24. For further biographical information on Kang Senghui: see T.55.2145:
96a29–97a17; T.50.2059: 325a13–18 and 325b4–326b13.
41 See Nattier, 2008: 151 and 151, note 113.
These lists form logical units, and, although also in the Chinese cultural con-
text numerical lists were common, these did not necessarily match the
Buddhist list with respect to content. In this way, the Buddhist series of ‘great
elements’ (mahābhūta: earth (pṛthivī), water (ap), fire (tejaḥ), wind (vāyu), and
matter derived (upādāyarūpa) from the four great elements) were matched
with the Chinese ‘five phases’ (wu xing 五行: water (shui 水), fire (huo 火),
wood (mu 木), metal (jin 金), and earth (tu 土)) known from ancient Chinese
numerology, and the Confucian five constituent virtues (wu chang 五常:
humaneness (ren 仁), righteousness (yi 義), ritual (li 禮), knowledge (zhi 智),
and faith (xin 信)) were paired with the five lay precepts (no harming living
beings, no stealing, no sexual misconduct, no lying, and no intoxication).
Liu Yiqing’s 劉義慶 (403–444) Shi shuo xin yu 世說新語, a collection of short
stories and conversations compiled in ca. 430 and the best known literary
reflection of the qingtan movement, alludes to the difficulty to understand
these often enigmatic lists in the following story:
When Yin Hao 殷浩 was dismissed and transferred to Tung-yang 東陽

(western Zhejiang) he read a large number of Buddhist sutras, gaining a
detailed understanding of them all. It was only when he came to places
where items were enumerated (shih-shu 事數), that he did not under-
stand. Whenever he chanced to see a monk he would ask about the items
he had noted down, and then they would become clear.42
Liu Jun’s 劉峻 (462–521) commentary to this passage cites several examples of

such lists: the five aggregates (skandha), the twelve sense-spheres (āyatana),
the four truths (āryasatya), the twelvefold cycle of dependent origination
(pratītyasamutpāda), etc.
3 Abhidharma and the ‘geyi’ Technique
The above described technique to render Buddhist terms with a term of the
Chinese – especially Daoist – philosophical tradition is generally referred to as
‘matching meanings’ (geyi 格義).43 We mentioned that difficulties in transla-
tion must have been particularly prominent when ‘numerical lists’ of
42 <http://ctext.org/shi-shuo-xin-yu/wen-xue/zh>. Translation: Mather 1976: 123.

43 Nattier, 2004: 10 in this respect refers to what she calls ‘Chinese cultural calques’: “transla-
tions that make no attempt to reflect the etymology of the Indian term, but instead
employ what was viewed as a suitable counterpart in Chinese”.
288 Dessein
Abhidharma texts had to be translated, and that the earliest Abhidharma

translations into Chinese were done in the second century CE. This brings us to
two main problems with the traditional explanation of ‘geyi’ as a technique
used in early Chinese translations. The first is that there is no single semantic
possibility to render the meaning of ‘ge’ 格 as ‘matching;’44 and the second is
that the qingtan movement and the ‘geyi’ technique that is associated with it
actually postdate the first translations of Abhidharma texts. That is, the ‘geyi’
technique only originated after the earliest translators already had equated
Buddhist with traditional Chinese concepts.45
The – scarce – historical evidence we have indeed points towards another
origin of the ‘geyi’ technique. In the biography of the Buddhist monk Zhu Faya
竺法雅, included in the Gaoseng zhuan 高僧傳, compiled around 530 CE by
Huijiao 慧皎 (497–554), e.g. we read that Zhu Faya (first half of the fourth cen-
tury) “took the numerical categories (shishu) of the sūtras and matched these
with [terms from] non-Buddhist literature (wai shu 外書), as a method to
explain them,”46 and that this method was called “matching meanings (geyi)”.47
This biography also informs us that Zhu Faya had excelled in non-Buddhist
studies (wai xue 外學) in his youth.48 As this Zhu Faya lived in the first half of
the fourth century, that is, at a time when such equations as the Buddhist
‘bodhi’ and the Daoist ‘dao’ 道 had already been established in the process of
the first Chinese translation activities, and as his biography mentions ‘geyi’ as
a technique to explain Abhidharma texts (pifutan 毘浮曇),49 it indeed is very
likely that ‘geyi’ originally referred to a particular method of exegesis, restricted
to an Abhidharmic context, and not to a more general method of expressing
Buddhist ideas in terms of Chinese philosophical terms, i.e. the meaning it was
given later. The connection of ‘geyi’ with the numerical lists of categories of
elements (shishu 事數) also justifies the use of the element ‘ge’ in the term
‘geyi’: ‘categorization’.50 There even is evidence that the technique was not
highly valued. Dao’an 道安 (312–385) would have opposed to ‘geyi’ because it
44 See Mair, 2012: 30–32.

45 Zürcher, 2007: 12.
46 T.50.2059: 347a20–21.
47 T.50.2059: 347a21–22. Note that “non-Buddhist literature” does not restrict the ‘matching’
with terms that were culled from Daoist literature. See also Mair, 2012: 30, 40; Zürcher,
2007: 12.
48 T.50.2059: 347a18.
49 T.50.2059: 347a22.
50 See Mair, 2012: 30–32. Mair, 2012: 37 remarks that ‘shishu’ 事數 may be equated with
‘fashu’ 法數 (which is linked to the Sanskrit dharmaparyāya: ‘discourse on dharma;’ or
with mingshu 名數: ‘numerical groups of related items’. He thus suggests (2012: 40) that
would “deviate from the principles [of Buddhism] (yu li duo wei 於理多違)”.51
The Chu sanzang ji ji informs us that Huirui 慧叡 (fourth century), a disciple of
Kumārajīva, reportedly said that the method of geyi “deluted and perverted the
doctrine”.52 Judging from Kumārajīva’s (344–413) biography, included in the
Chu sanzang ji ji, Sengyou (435–518) appears not to have understood the mean-
ing of the term any more.53
Controversy thus must have arisen within Buddhist circles on the suitability
of the ‘geyi’ method to be applied in other than purely Abhidharmic contexts.
An example of this application of the ‘geyi’ method that goes beyond the expla-
nation of Abhidharmic categories is the preface to Kumārajīva’s translation of
the *Dvādaśadvāraka, Shi’er men lun 十二門論 (T.30.1568) done by Kumārajīva’s
disciple Sengrui 僧叡 (born in 352 CE) and Jizang’s 吉藏 (549–623) commen-
tary on this text, titled Shi’er men lun shu 十二門論疏 (T.42.1852). It may not be
without importance that while Sengrui’s introduction only quotes from the
Zhuangzi 莊子,54 Jizang’s commentary expands the references with quotations
from the Laozi 老子, from Guo Xiang’s 郭象 (died 312) Neo-Daoist commen-
tary on the Zhuangzi, and even from the Confucian Lunyu 論語 and the already
mentioned Mao shi version of the Shijing.55
It is not unlikely that it was precisely at the moment that the use of ‘geyi’ as
an exegetical technique was expanded to non Abhidharma texts, that it became
criticized by learned monks such as Dao’an.56 The questionability of the tech-
nique may also explain why as good as none of the extant Chinese translations
of Abhidharma texts uses this ‘matching meanings’ technique.57
the term ‘shishu’ designates “enumerative categories (or categorized enumeration) of

things/items, i.e., (technical) terms”.
51 T.50.2059: 355a25. This account is included in the biography of Shi Sengguang 釋僧光.
Translation: Link, 1958: 43. Cf. Tang, 1987: 234–238.
52 T.55. 2145: 41b12.
53 T.55.2145: 101b15. See also Tang, 1987: 237–8.
54 T.30.1568: 159b17, 159b20–21.
55 References to the Zhuangzi: T.42.1825: 173c8–9, 173c23–24, 174b7; references to the Laozi:
T.42.1825: 173b8–9, 173b4, 173c13–14, 174a3; references to Guo Xiang’s commentary on the
Zhuangzi: T.42.1825: 172c8–9; references to the Lunyu: T.42.1825: 171a15, 173b8, 173b16; refer-
ences to the Mao shi: T.42.1825: 174b9. For a complete analysis of this preface: See Dessein,
1997.
56 Mair, 2012: 46 argues that it is especially through the critique Dao’an formulated on the
‘geyi’ that twentieth century scholarship knows about this unfruitful technique.
57 See Zürcher, 2007: 12.
290 Dessein
4 ‘Geyi’ Confucianism
The period between the fall of the Han dynasty in 220 CE and the reunification
of the country under the Sui 隋 dynasty in 589 CE has, in Chinese historiogra-
phy and in Western scholarship alike, generally been described as “an age of
religious escapism, metaphysical speculation, and uninhibited eccentricity”.58
Confucian scholars for whom the possibilities to pursue their official careers
had been halted are then thought to have indulged themselves in philosophical
discussions, a trend that is generally designated as ‘qingtan,’ ‘pure conversa-
tion’. Confronted with the advent of Daoism and Buddhism, also an altogether
different reaction by the Confucian intellectuals is therefore imaginable. It is
not unlikely that along with the ‘qingtan’ movement, a ‘conservative’ move-
ment that wanted to reformulate Confucianism developed. Some literati may,
indeed, have blamed a degenerate Confucianism for the fall of the Han, and
may have vowed to become Confucian models to emulate in their own right,
and therefore follow the principles of Confucianism more strictly than those
in power.59 It is therefore also possible that we should see the ‘geyi’ exegetical
method as a coping strategy deployed by those few ‘conservative’ Confucian
literati who wanted to redefine Chinese culture in a context of growing influ-
ence of Daoism and Buddhism.60
Having proven its unsuitability to explain/translate the numerical lists of
Buddhist texts, the ‘geyi’ technique must have disappeared as a Buddhist
method not much longer than one generation after it was first used.61 Hereafter,
in fourth and fifth century China, ‘geyi’ developed to be an instrument to rede-
fine Confucian orthodoxy. This may explain why it especially were “young
members of gentry families” who “were all well-versed in the secular canons
(shi dian 世典), but had not yet become conversant with the principles of
58 Jansen, [2008] 2009: 403.

59 As argued by Jansen, [2008] 2009: 402, also in the post-Han period “Confucian influence
remained strong and many thinkers would still have identified themselves first and fore-
most as Confucians”.
60 In the same way that, as shown by Shils, 1971: 258–260, in his study on the ideological and
political behavior of intellectuals in times of crisis, it typically are intellectuals who are
susceptible for feelings of humiliation of national and cultural proud and who feel the
need for action. A similar reformulation of Confucian orthodoxy happened with the
development of so-called Neo-Confucianism (daoxue jia 道學家) in the Tang and Song
periods, and with the development of so-called New Confucianism (xin rujia 新儒家) in
the contemporary period.
61 See Mair, 2012: 38.
Buddhism”62 who adhered to Zhu Faya for his classical knowledge and his
Buddhist instruction, and why Dao’an opposed to ‘geyi’ because it would “devi-
ate from the principles [of Buddhism]”.63 Seen from this perspective, so-called
fourth and fifth century ‘geyi Buddhism’ should rather be recoined as ‘geyi
Confucianism’.
5 Conclusion
From the above, we can deduce that the technique of ‘geyi’ most likely started
as a type of exegesis applied to the numerical lists that were prevalent in the
Abhidharma texts. This technique, further, was practiced in circles of monks
who had enjoyed a traditional schooling and were well-versed in the Chinese
classics. We also know that this peculiar technique was used up to the fourth
century, i.e. the period in which also such intellectual trends of qingtan
flourished.
Given the denouncement of the ‘geyi’ technique in Buddhist circles itself, it
is therefore possible to formulate a hypothesis on the nature of ‘geyi’ Buddhism
and a hypothesis on why this technique attained more attention than its fac-
tual importance allows, that differs from the generally accepted one. Rather
than that, as commonly accepted, Buddhist – especially prajñā – studies were
carried under the aegis of Chinese xuanxue thinking, it might have been, on
the contrary, Buddhism that stimulated creativity in Chinese philosophy, thus
forming xuanxue thinking and many devotees of qingtan also engaging in
xuanxue.64 This supposition is corroborated both in terms of chronology and
content. The similarities in psychological and metaphysical vocabulary
between qingtan and xuanxue facilitated this movement, and, gradually, also
Buddhist monks and Confucians who were knowledgeable in Buddhism must
have participated in these debates and conversations.65When Buddhist
Ābhidharmikas retreated from the ‘geyi’ method, it was adopted by Confucian
literati to reformulate their doctrine in the complex philosophical and reli-
gious atmosphere that preceded the reunification of the empire in the Sui
dynasty. This must have helped to pave the way for the development of so-
called Neo-Confucianism.
62 T.50.2059: 347a19–20.
63 T.50.2059: 355a25.
64 See also Lewis, 2009: 205.
65 Some eminent monks like Zhi Dun 支遁 (fourth century) even became leading figures in
the qingtan movement of the Jin dynasty. See Lewis, 2009: 205.
292 Dessein
That Confucian ‘geyi’ is an adaptation of an exegetical method – not a trans-

lation technique – that, before, had been applied to a peculiar type of Buddhist
texts might than also explain why the term is not mentioned in the Shi shuo xin
yu 世說新語: the term – not the technique – was tainted with a Buddhist con-
notation – precisely what the ‘conservative’ literati wanted to avoid.
Abbreviation

邊 and Ono 小野, 1924–1934)
Bibliography
Ch’en, Kenneth ([1972] 1973). Buddhism in China. A Historical Survey. Princeton NJ:
Cheng, Anne (1997). Histoire de la pensée chinoise. Paris: Seuil.
Dessein, Bart (1997). “The Chinese Reception of Indian Buddhism: Shih-erh Men Lun
Hsü (T.Vol. 30, No. 1568): Seng-jui’s Preface to Nāgārjuna’s ‘Twelve Gates Treatise’ and
Shih-erh Men Lun Hsü Shu (T.Vol. 42, No. 1825): Chi-tsang’s commentary on this pref-
ace – An analysis)”, The Indian Journal of Buddhist Studies 9/1–2, pp. 38–68.
Dessein, Bart (1999). Saṃyuktābhidharmahṛdaya. Heart of Scholasticism with Miscella
neous Additions. 3 vols. (Buddhist Tradition Series 33–35). Delhi: Motilal Banarsidass.
Dessein, Bart (2010). “The Abhidharma School in China and the Chinese Version of
Upaśānta’s *Abhidharmahṛdayasūtra”, The Eastern Buddhist. New Series 41/2,
pp. 49–69.
Dessein, Bart (2014). “Faith and Politics: (New) Confucianism as Civil Religion”, Asian
Studies II (XVIII)/1, pp. 39–64.
Dull, Jack L. (1994). “Determining Orthodoxy: Imperial Roles”, In Imperial Rulership and
Cultural Change in Traditional China. Edited by Frederick P. Brandauer and Chün-
chieh Huang. Seattle: University of Washington Press, pp. 3–27.
Fan, Ye (1976). Hou Han shu 後漢書. Beijing: Zhonghua shuju, 1978.
Harrison, Paul (1993). “The Earliest Chinese Translations of Mahāyāna Sūtras: Some
Notes on the Works of Lokakṣema”, Buddhist Studies Review 10/2, pp. 135–177.
Jansen, Thomas ([2008] 2009). “New Tendencies, Religious and Philosophical, in the
Chinese World of the Third through Sixth Centuries”, In Conveiving the Empire. China
and Rome Compared. Edited by Fritz-Heiner Mutschler and Achim Mittag. Oxford:
Oxford University Press, pp. 397–419.
Lai, Whalen (1979). “Limits and Failure of Ko-I (Concept-Matching) Buddhism”, History
of Religions 18/3, pp. 238–257.
Lewis, Mark E. (2007). The Early Chinese Empires. Qin and Han. Cambridge MA and
London: Harvard University Press.
Lewis, Mark E. (2009). China Between Empires. The Northern and Southern Dynasties.
Cambridge MA and London: Harvard University Press.
Link, Arthur E. (1958). “Biography of Shih Tao-an”, T’oung pao, Second Series, Vol. 46, 1/2,
pp. 1–48.
Liu, Jiahe and Shao, Dongfong (1992). “Early Buddhism and Taoism in China (ad 65–
420)”, Buddhist-Christian Studies 12, pp. 35–41.
Liu, Yiqing 劉義慶, Shi shuo xin yu 世說新語. <http://ctext.org/shi-shuo-xin-yu/zh>.
Loewe, Michael (1986). “The religious and intellectual background”, In The Cambridge
History of China. Volume 1: The Ch’in and Han Empires. 221 bc–ad 220, Edited by Denis
Twitchett and Michael Loewe. Cambridge: Cambridge University Press,
pp. 649–725.
Lüshi chunqiu 呂氏春秋 ([1986] 1988). In Ershi’er zi 二十二子. Shanghai: Changhai guji
chubanshe.
Mair, Victor (2012). “What is Geyi, After All?”. China Report, pp. 29–59.
Maspéro, Henri (1910). “Le songe et l’ambassade de l’empereur Ming: Étude critique des
sources”, Bulletin de l’École Française d’Extrême-Orient 10/1, pp. 95–130.
Mather, Richard B. (1976). Shih-shuo Hsin-yü. A New Account of Tales of the World. By Liu
I-ch’ing, with commentary by Liu Chün. Minneapolis MN: University of Minnesota
Press.
Mittag, Achim ([2008] 2009). “Forging Legacy: The Pact between Empire and
Historiography in Ancient China”, In Conceiving the Empire. China and Rome
Compared. Edited by Fritz-Heiner Mutschler and Achim Mittag. Oxford: Oxford
University Press, pp. 143–165.
Mizuno, Kōgen ([1982] 1995). Buddhist Sutras. Origin, Development, Transmission. Tokyo:
Kōsei Publishing Co.
Nattier, Jan (2004). “Beyond Translation and Transliteration: A New Look at Chinese
Buddhist Terms”. Unpublished paper delivered at the annual meeting of the Western
Branch of the American Oriental Society. Portland, 16 October 2004.
Nattier, Jan (2008). A Guide to the Earliest Chinese Buddhist Translations. Texts from the
Eastern Han 東漢 and the Three Kingdoms 三國 Periods. (Bibliotheca Philologica et
Philosophica Buddhica X). Tokyo: The International Research Institute for Advanced
Buddhology. Soka University.
Needham, Joseph (1958). Science and Civilisation in China. Vol. 2: History of Scientific
Thought. Cambridge: Cambridge University Press.
294 Dessein
Pirazzoli-t’Serstevens, Michèle ([2008] 2009). “Imperial Aura and the Image of the Other
in Han Art”, In Conceiving the Empire. China and Rome Compared. Edited by Fritz-
Heiner Mutschler and Achim Mittag. Oxford: Oxford University Press, pp. 299–317.
Schwartz, Benjamin I. (1985). The World of Thought in Ancient China. Cambridge MA and
London: Harvard University Press.
Shils, Edward (1971). “The Intellectuals in the Political Development of New States”, In
Political Development and Social Change. Edited by Jason L. Finkle and Richard W.
Gable. New York NY: John Wiley & Sons, pp. 249–276.
Storch, Tanya (2014). The History of Chinese Buddhist Bibliography. Censorship and
Transformation of the Tripitaka. Amherst: Cambria Press.
- T.26.1543: Kātyāyanīputra, *Aṣṭagrantha, Apitan ba jiandu lun 阿毘曇八犍度論,
Saṃghadeva, Zhu Fonian 竺佛念,.
- T.28.1550: Dharmaśreṣṭhin, *Abhidharmahṛdaya, Apitan xin lun 阿毘曇心論,
Saṃghadeva and Huiyuan 慧遠.
- T.28.1552: Dharmatrāta, *Saṃyuktābhidharmahṛdaya, Za apitan xin lun 雜阿毘曇
心論, Saṃghavarman.
- T.30.1568: *Nāgārjuna, *Dvādaśadvāraka, Shi’er men lun 十二門論, Kumārajīva.
- T.42.1825: Jizang 吉藏, Shi’er men lun shu 十二門論疏.
- T.49.2034: Fei Changfang 費長房, Lidai sanbao ji 歷代三寶記.
- T.49.2035: Zhipan 志磐, Fo zu tong ji 佛祖統記.
- T.50.2059: Huijiao 慧皎, Gaoseng zhuan 高僧傳.
- T.50.2149: Daoxuan 道宣, Da Tang neidian lu 大唐內典錄.
- T.55.2154: Zhisheng 智昇, Kaiyuan shi jiao lu 開元釋教錄.
Tang Yongtong 湯用彤 (1987). Han Wei Liang Jin Nanbeichao Fojiao shi 漢魏兩晉南北朝
佛教史. 2 vols. Shanghai, 1938. Reprint: Taipei: Luotuo Chubanshe.
Tsukamoto, Zenryū (1985). A History of Early Chinese Buddhism. From its Introduction
to the Death of Hui-yüan. 2 vols. Tokyo, New York, San Francisco: Kodansha
International. (English translation of the original Japanese work by Leon Hurvitz.
Original publication: Chūgoku Bukkyō tsūshi. Tokyo: Shunjūsha, 1979).
Watanabe, Baiyū; Mizuno, Kōgen; and Ōishi, H. (1932). Abidonshinron. (Kokuyaku
Issaikyō. Bidonbu vol.21). Tokyo: Daitō Shuppansha.
Scholasticism. (Handbook of Oriental Studies 2/11). Leiden etc.: Brill.
Wright, Arthur F. (1959). Buddhism in Chinese History. Stanford CA: Stanford University
Press.
Xu, Shen 許慎 ([1981] 1988). Shuo wen jie zi zhu 說文解字注. Shanghai: Shanghai guji
chubanshe.
Yu, Anthony C. (2005). State and Religion in China. Historical and Textual Perspectives.
Chicago IL and La Salle IL: Open Court.
Zacchetti, Stefano (2010). “Defining An Shigao’s 安世高 Translation Corpus: The State
of the Art in Relevant Research”, In Historical and Philological Studies of China’s
Western Regions 3. Edited by Weirong Shen. Institute for Historical and Philological
Studies of China’s Western Regions, Renmin University of China, pp. 249–270.
Zürcher, Erik (1977). “Late Han Vernacular Elements in the Earliest Buddhist Translations”,
Journal of the Chinese Language Teachers Association 12/3, pp. 177–203.
Zürcher, Erik (1991). “A New Look at the Earliest Chinese Buddhist Texts”, In From Benares
to Beijing. Essays on Buddhism and Chinese Religion. Edited by Koichi Shinohara and
Gregory Schopen. Oakville CA: Mosaic Press, pp. 277–304.
Zürcher, Erik (2007). The Buddhist Conquest of China. The Spread and Adaptation of
Buddhism in Ealry Medieval China. Third Edition with a Foreword by Stephen F. Teiser.
Leiden: Brill.
296 Teng
Chapter 9
Kuiji’s Abhidharmic Recontextualization of

Chinese Buddhism
Weijen Teng
1 Introduction: Contextualization and Recontextualization
Elsewhere I have tried to show that Kuiji 窺基 (632–682) who was recognized
historically as the successor of Xuanzang 玄奘 especially in the latter’s legacy
of the Chinese Yogācāra school, the Faxiang school (Faxiang zong 法相宗),
intended to set Chinese understanding of Buddhism aright by reorienting it
toward Indian Buddhism. I tried to demonstrate in that work that what consti-
tuted such a reorientation was not a mere promotion of a Yogācāra form of
Buddhism as has been generally understood by scholars of Chinese Buddhism.
I argued that Kuiji’s more significant potential contribution to such a reorien-
tation lay in his construction of a methodological foundation for Chinese
Buddhist intellectual practice, which can be interpreted as a response to the
much debated issue of the ‘sinification of Buddhism’. The three methodologi-
cal dimensions of Kuiji’s endeavors I emphasized are: recontextualization,
exegesis, and logic. The present study explores the ‘recontextualization’ dimen-
sion of Kuiji’s methodology.1
In the prolonged course of the spreading of Buddhism to China, Buddhist
texts and ideas were studied and interpreted divorced from their home cul-
tural, historical, and intellectual contexts. On the other hand, they were
received, digested and made use of the new contexts informed by the local
cultural and socio-political concerns. Echoing Yūki Reimon, Robert Gimello
has tried to show that in response to the socio-political context of the late sixth
and early seventh centuries, most of the leading Buddhists in that period were
concerned to relieve Buddhism of the onus of its foreignness and approached
Buddhist texts and teachings in such a context.2
1 See the “Introduction” of Teng’s unpublished doctoral dissertation (2011), “Recontextualization,

Exegesis, and Logic: Kuiji’s (632–682) Methodological Restructuring of Chinese Buddhism”.
This chapter is a modification of Teng, 2011’s Chapter Two, “Doctrinal and Historical
Recontexualization of Chinese Buddhism”.
2 Gimello, 1976: 95–97.

Kuiji’s Abhidharmic Recontextualization of Chinese Buddhism 297
According to Robert Gimello and others, the innovation of panjiao 判教 sys-

tems was a demonstration of the leading Chinese Buddhist thinkers’
determination to show that “these new Chinese modes of Buddhism were in
no way inferior to their Indian counterparts but were, in fact, superior rendi-
tions of the Buddha’s truth”.3 Gimello further cited several predominant
Chinese expressions of Buddhism that emerged as a response to the local con-
texts. He illustrates, for instance, that in a time of socio-political turmoil and a
‘Buddha-less’ age, Buddhists aspired to be reborn in the Pure Land of the West
by pure faith in Amitābha Buddha. More prominently, Zhiyi 智顗 (538–597),
the founder of the Tiantai 天臺 school, put forward that “ignorance is identical
with dharma nature and the defilements are identical with awakening (wum-
ing ji faxing, fannao ji puti 無明即法性煩惱即菩提)” and Zhiyan 智儼, the
Second Patriarch of the Huayan 華嚴 school, proposed that “saṃsāra and
nirvāṇa are not two and are without distinction of one another (shengsi ji nie-
pan, wu er wu ci bi 生死即涅槃，無二無此彼)”.4 Expressions such as these
were put forward by the Buddhist elites, who were “most sensitive to historical
indications of the inception of ‘the last phase of the Dharma (mofa 末法),’ least
bound to the alien institutional forms of Buddhism, and most concerned with
providing a less intimidating path to a more readily accessible religious goal”.5
Chinese contextualization of Buddhism as mentioned above was inevitable
and in a sense necessary.6 However, it also resulted in a ‘context-shifted’ reading
of Buddhism that was viewed as doctrinally unsystematic and indiscriminative
in the critical eyes of Xuanzang and his student, Kuiji. Like his master, Kuiji
was concerned that when such contextualized expressions of Buddhist con-
cepts, doctrines, and the texts themselves were taken by the Buddhists of later
3 Gimello, 1976: 95. Peter Gregory has a similar suggestion. (Gregory, 1991: 110–112). A compre-
hensive historical survey in English of the Chinese Buddhist ‘doctrinal classification’ system
(panjiao 判教 can be found in Mun’s The History of Doctrinal Classification in Chinese
Buddhism. However, Mun did not address the hermeneutical issue or exegetical nature of the
panjiao system.
4 Gimello, 1976: 111.
5 Gimello, 1976: 113. More Chinese expressions of the ‘new Buddhism’ can be found in the same
reference.
6 The necessity of a “contextual theology” has been argued for initially by Shoki Coe (Coe, 1993),
and more recently by Stephen Bevan in his Models of Contextual Theology: The Struggle for
Cultural Relevance, in which he puts forward five models of “contextual theory,” namely (1)
translation; (2) the anthropological; (3) praxis; (4) the synthetic; and (5) the transcendental
issues related to the localization of Chinese theology (Bevans, 2002). These five models pro-
vide a helpful analytical framework for thinking about various forms of encounter between
‘Buddhisms’ and ‘Chinese cultures’.
298 Teng
generations to be what was intended in their original contexts, there would

be no measurement against which to determine an ‘authentic’ for ‘erroneous’
reading of Buddha’s teaching. Such concerns are explicit in Xuanzang’s words:
When I examined the teachings (of the teachers), each of them professed
exclusive orthodoxy [of his teachings]. When I tried to test them against
the [available] textual testimony, the meanings were either implicitly or
explicitly inconsistent. I am left with no one to follow,7
and in Kuiji’s:
Interpretations (of Buddhist teaching) as diverse as these (earlier ones)

caused confusions and disputes among the Buddhist masters. This is due
to the lacking of a(n) [exegetical] standard. The students of later genera-
tions were even more confused and had no way to decide which to accept
and which to reject.8
Kuiji’s recontextualization as will be treated in this chapter constitutes a cor-

rective to this indigenously contextualized reading of Buddhism with its texts
and doctrines on the part of Chinese Buddhist elites.9 More precisely, Kuiji’s
recontextualization as understood here is a methodological move, which
emphasizes the paramount importance of Abhidharmic scholasticism, that is,
using the methodology of Abhidharma for the discernment of the nature of
scripture, doctrines, and fundamental concepts. A survey of Kuiji’s commen-
taries and especially his ‘theological summa,’ the Dasheng fayuan yilin zhang
大乘法苑義林章, finds that Kuiji attached paramount importance to the dis-
cernment of the nature of ‘D(d)harmas,‘ in both senses of the ‘Buddha’s
teaching (Dharma)’ and ‘phenomena (dharmas)’. For example, for each topic
that Kuiji explicated in his Fayuan yilin zhang, he would distinguish the nature,
bianxing 辨性, distinguish the characteristics, bianxiang 辨相, correlate and
categorize, xiangshe 相攝, and so on. As is to be pointed out shortly, these
methods are the methodological functions of Abhidharmic scholasticism in
the Ābhidharmika tradition for the decisive discernment of the Buddhist
7 Xuanzang’s biography, T.50.2053: 222c3–4, reads: “詳考其義各擅宗塗，驗之聖典，亦隱

顯有異，莫知適從”.
8 T.45.1861: 254c20–22: “如是種種解釋不同，率己情未為典據，以諸師紛亂互起異端，
後學徒無可從受”.
9 Elsewhere, I have studied Kuiji’s other methodological strategies, such as Sanskrit grammar,
and Indian logic. See Teng, 2011.
teachings and concepts. In what follows, let me introduce briefly such func-
tions relevant to this study.
2 The Role and Methodology of Abhidharma in the Buddhist

Tradition
The paramount importance of the Abhidharma mode of exposition is recog-

nized by the ‘Hīnayāna’ traditions,10 except the Sautrāntikas, and with some
qualifications, by the Mahāyāna tradition as well. According to the ‘Hīnayāna’
traditions, Abhidharma is either the direct words of the Buddha,11 or what the
Buddha originally intended.12 By and large, the basic function of an Abhidharma
exposition is to ascertain, determine (vi –niś-√ci), and discern directly (abhi-
sam-√i or sākṣāt-√kṛ) the nature and characteristics of the dharmas; it enables
one to properly realize fundamental Buddhist doctrines such as the twelve-
linked dependent origination (pratītyasamutpāda), the four noble truths, the
noble eightfold path, nirvāṇa, and so on.13
The service of Abhidharma to the Sūtra is attested in the Mahāyānasūtrā
laṃkāra, in which the Abhidharma is defined as “that by which the meaning of
the sūtra is best understood” (abhigamyate sūtrārtha etenety abhidharmaḥ).14
The Ābhidharmikas profess that the “Abhidharma exposition teaches absolute
10 The use of the pejorative term ‘Hīnayāna’ here is to highlight the Chinese Buddhist sectar-
ian conception of Abhidharma Buddhism as opposed to the Mahāyāna school, a point to
be discussed in this chapter.
11 According to the Theravāda tradition, all of the seven Abhidhamma texts are delivered by
the Buddha and transmitted through Śāriputra. Sammāsambuddhassa Abhidhamma
desanā pariyosānañ ca tesaṃ bhikkhūnaṃ sattappakaraṇauggahanañ ca ekappahāren’
eva ahosi. Abhidhammo cācanāmaggo nāma Sāriputtattherappabhavo. The tradition rec-
ognized that Buddha is first an Ābhidharmika, Sammāsambuddho va paṭhamataraṃ
Ābhidhammiko (Dhammasaṅgaṇi-atthakathā, p. 17).
12 The Sarvāstivāda school acknowledges Kātyāyanīputra and others’ authorship of the
Abhidharma texts, but believes that the Buddha was the original author. “na hi
vinābhidharmopadeśena śiṣyaḥ śakto dharmān pravicetum iti | sa tu prakīrṇa ukto
bhagavatā | bhadanta Kātyāyanīputraprabhṛtibhiḥ piṇḍīkṛtya sthāpito bhadanta Dharma
trātodānavargīyakaraṇava” (Pradhan, 1967: 3). See also Dhammajoti’s Sarvāstivāda
Abhidharma on a variety of scripture genres or features that lead to the development of
the Abhidharma exposition: abhidharma kathā, vaidalya, vibhaṅga, mātṛkā, and upadeśa
(Dhammajoti, 2007: 3–9).
13 For more illustrations of the functions of the Abhidharma exposition: see Dhammajoti,
2007: 9–13.
14 Potter et al., 1996: 74. See also the introduction to this volume.
300 Teng
truth (paramārthasatya), with fully drawn out explicit meanings (nītārtha). In

contrast, the Sūtra literature does not represent the Buddha’s true intention
(abhiprāyika). It generally represents the expedient (aupacārika) teachings
whose meanings are yet to be fully drawn out (neyārtha).15 The Pāli Aṭṭhasālinī
therefore remarks that:
The Abhidharma surpasses the teachings in the Sūtra, because the vari-
ous classifications of the elements of existence are listed haphazardly in
the sūtras, while the Abhidharma gives them in their definitive catechetic
forms. Thus, the sūtras are preached from the standpoint of conventional
truth (saṃvṛtisatya) according to specific worldly circumstances, but the
Abhidharma deals with absolute truth, and is concerned with the analy-
sis of mind and matter (nāmarūpapariccheda).16
In order to perform the services mentioned above, the Ābhidharmikas devel-

oped a set of methodological devices, which became the normative means for
a thorough and analytical understanding of the dharmas. This methodology
was adopted in Kuiji’s exposition of the basic dharma-taxonomies. These
devices are:
I. Subsumption/inclusion (saṃgraha; Ch. she 攝)

This device is basically a means of categorizing all dharmas by means of
investigating their intrinsic nature (svabhāva) or particular characteristics
(svalakṣaṇa) and universal characteristics (sāmānyalakṣaṇa). The basic taxon-
omies are the five aggregates (skandha), six sense-spheres (āyatana), eighteen
elements (dhātu), and the five-group taxonomy (pañcavastuka; Ch. wuwei 五
位). The taxonomy of the dharmas of the Mahāyāna Abhidharma/Yogācāra
system is based on the last taxonomy.
II. Association/conjunction (saṃprayoga; Ch. xiangying 相應)

This is a device for determining the interconnection of thought with thought-
concomitants, and among the concomitants themselves. To be associated they
must be supported by the same basis, that is, the sense organ (āśrayasamatā),
the same object (ālambanasamatā), cease simultaneously (kālasamatā), and
be singular in substance (dravyasamatā).
15 Dhammajoti, 2007: 19.

16 Potter et al., 1996: 74.
III. Discernment of the dharmas (dharmapravicaya)

This is probably the most important and relevant category for Kuiji’s method-
ological construction. The Ābhidharmikas adhere to the idea that “it is only by
a thorough and systematic examination of the true nature of all dharmas that
the true import of the Buddha’s Dharma can be ascertained and liberation
assured”.17 The object of the dharma investigation of the Abhidharma includes
Buddha’s teaching. According to the *Mahāvibhāṣāśāstra, Da piposha lun 大毘
婆沙論, one should skillfully distinguish between the provisional liaoyi 了義
(nītārtha) and ultimate buliaoyi 不了義 (neyārtha) teachings.
The wise should skillfully distinguish the meanings in the sūtras and not
simply explain in the manner that the words are uttered [therein]. If one
simply explains in the manner that the words are uttered, one will create
contradictions among the noble teachings and also give rise to topsy-
turvy attachment in one’s own mind.18
Given the superiority and importance of Abhidharma in the early Buddhist

tradition, how was Abhidharma received in the Chinese Buddhist context?
What, if any, did Chinese exegetes make of it? When Abhidharma was neglected
along with other Hīnayāna canonical texts by the Mahāyāna minds of Chinese
Buddhists, how did Kuiji, who was a Mahāyāna adherent himself, argue for the
importance of Abhidharma and make much of it?
3 Transmission of Abhidharma in China
Before the introduction to China of the Abhidharma literature proper, near the
end of the fourth century, a number of Abhidharmic taxonomies and numeri-
cal categories [of dharmas] in meditation (chanshu 禪數) were made known to
the Chinese Buddhists through a few meditation texts translated by the
Parthian monk An Shigao 安世高 who arrived in Luoyang 洛陽 in 148 CE. The
Abhidharmic taxonomies were mentioned in An Shigao’s translation of the
Discourse on the Aggregates, Elements, and Sense-spheres (Yin chi ru jing 陰持入
經), in which were introduced basic dharma-taxonomies such as the five aggre-
gates (skandha), six sense-spheres (āyatana), eighteen elements (dhātu), and
17 See Dhammajoti, 2007.

18 T.27.1545: 145c15–23: “於諸契經應善分別了不了義，復作是言，故於契經應分別
義。如世說 … 是故智者應於契經善分別義，不應如說而便作解。若如說而解
者，則令聖教前後相違，亦令自心起顛倒執”.
302 Teng
the thirty-seven factors conducive to awakening (bodhipakṣikā dharmāḥ).

However, due to the abstruseness of these terminologies and obscure transla-
tion terms, the actual content of these dharma-taxonomies remained alien to
the Chinese Buddhists. Moreover, such dry technical texts were soon overshad-
owed by the subsequent translation of the Prajñāpāramitā literature made
available by Lokakṣema and others who arrived some twenty years after An
Shigao.19 Erik Zürcher has pointed out the interesting anomaly that passages
from a Prajñāpāramitā text were quoted to explicate the Yin chi ru jing in the
commentary to the latter.20
The study of Abhidharma in China was initiated, in a more meaningful
sense, and supported by Dao’an 道安 (312–385). It was during his time that
Abhidharma texts were translated for the first time in China. The Abhidharma
texts were brought to China by the Abhidharma specialists *Saṃghadeva and
*Saṃghabhadra,21 who arrived in Chang’an 長安 around 381, and Kumārabodhi
who arrived in the following year.22 The Abhidharma texts translated, all of
which belong to the Sarvāstivāda tradition, were:
I. Apitan ba jiandu lun 阿毘曇八犍度論 (*Aṣṭagrantha / *Jñānaprasthāna),

introduced by *Saṃghadeva (T.26.1543)23
II. Piposha lun 鞞婆沙論 (*Vibhāṣāśāstra), introduced by *Saṃghabhadra
(T.26.1547)24
III. Zun Poxumi pusa suo ji lun 尊婆須蜜菩薩所集論 (*Āryavasumitrasaṃgīti),
introduced by *Saṃghadeva and Dharmanandi (T.28.1549)
IV. Apitan xin lun 阿毘曇心論 (Abhidharmahṛdaya) introduced by *Saṃgha
deva (T.28.1550).25
19 See Zürcher, 2007: 35.

20 The commentary was composed by certain Chen Hui 陳慧 in the second half of the third
century. For more detailed information about this commentary: see Zürcher, 2007: 54.
21 Demiéville restored 僧伽跋澄 as *Saṅghabhadra, Cox (1995) gave *Saṅghabhūti (or
Saṅghabhadra), and Dhammajoti (2007: 127) gave *Saṅghabhūti. The Sanskrit restoration
of *Saṅghabhūti was probably from zhongxian 眾現, a name that gives the Chinese
meaning instead of the pronunciation of the Sanskrit original. However, Paul Demiéville
(1951: 364 note 8) has shown that the character xian 現 was an error for xian 賢, and the
translation should be zhongxian 眾賢, which is the standard Chinese translation of
*Saṅghabhara.
22 Zürcher, 2007: 202.
23 For a Sanskrit fragment of this text: see de La Vallée Poussin, 1930.
24 A Sanskrit fragment of this text is available: see Enomoto, 1996.
25 For an annotated translation of this text: see Willemen, 1975.
The actual translations of these texts were probably made by Zhu Fonian 竺佛
念, who was a Chinese polyglot with little knowledge of Buddhist doctrine,
under the sponsorship and management of Dao’an.26 Dao’an himself studied
Abhidharma thought from An Shigao’s translations and was eager to under-
stand more about it. He remarked that the earlier commentaries made by Kang
Senghui 康僧會 were obscure, and took the initiative of composing another
commentary.27
Though a Prajñāpāramitā scholar, Dao’an was totally open to the newly
arrived Hīnayāna Sarvāstivādin teaching. He was keen to introduce it to the
world of Chinese Buddhism, which had been carried away by Prajñāpāramitā
scholarship.28 Thanks to Dao’an’s sponsorship, Abhidharma studies thrived for
a short period of time. There was probably some Abhidharma learning fever
even among non-Buddhist intellectuals that urged the use of a pedagogical
strategy known as “geyi 格義” by Zhu Faya 竺法雅. In Chapter Four, “Literature,”
of the New Account of Tales of the World 世說新語,29 we find an anecdote in
which “a middle-troop general by the name of Yin Hao 殷浩 took great interest
in studying Buddhism after he was ousted from his office. In general, Yin Hao
had a very good grasp of the meanings of Buddhist texts, but when he came
across shishu 事數, that is the numerical categories of dharmas, he was not
able to understand them.30
Geyi was a method of adopting numerical categories from non-Buddhist
texts as an analogy to the Buddhist numerical categories, for the purpose of
lecturing non-Buddhist intellectuals on Buddhist teachings.31 Dao’an found
26 Zürcher, 2007: 202–203.

27 T.55.2145: 43c19–24: “有安世高者，博聞稽古，特專阿毘曇，學其所出經，禪數
最悉。此經其所譯也，茲乃趣道之要徑，何莫由斯道也。魏初康會為之注義
義或隱而未顯者，安竊不自量。敢因前人為解其下，庶欲蚊翮以助隨藍，霧
潤以增巨壑也”.
28 Tsukamoto, 1985: 736–737.
29 The Shishuo xinyu is a collection of anecdotes of wise men, scholars, monks of Buddhist
and Daoist origin, even of emperors and their concubines, written by Liu Yiqing 劉義慶
in the first half of the fifth century).
30 “殷中軍被廢，徙東陽，大讀佛經，皆精解，唯至“事數”處不解” (世說新
語．文學第四).
31 Victor Mair has recently argued that the term geyi has been over-generalized by scholars
to mean using any traditional Chinese concept to translate or to understand Buddhist
concepts. See Mair’s most recent article, which aims at rectifying the understanding of
geyi, “What is Geyi, After All” in Philosophy and Religion in Early Medieval China, 2010. I
thank Michael Radich for pointing out this reference to me.
304 Teng
disadvantages in employing this method, and suggested abandoning it.32 The

geyi method was probably practiced for only a short period of time, since the
Abhidharma study fever did not last more than a few decades.
The short life of Abhidhama studies was probably due to the influence of
Kumārajīva’s new translations of the Mahāyāna texts, which attracted a good
number of learned Buddhists, partly due to his elegant literary style.33 Antago
nism to the Hīnayāna texts, including Abhidharma, soon emoted probably
right after the first generation of Kumārajīva’s disciples. A discontent about an
enthusiasm of Abhidharma learning and a negligence of some Mahāyāna
sūtras is found in the Hongming Ji 弘明集 compiled by Sengyou 僧祐 (445–
518).34
By the early sixth century, Abhidharma was discarded as an inferior teach-
ing, along with other Hīnayāna texts. This can be attested in the Dashengyi
zhang 大乘義章 (On the Meaning of the Mahāyāna) composed by Huiyuan 慧
遠 (523–592), an influential Dilun 地論 exegete:
The Hīnayāna tradition is equipped with three complete baskets. Why is

this not mentioned in the Mahāyāna tradition? It is from the point of
view of the Buddha’s primitive teaching that this is so. In order to teach
Hīnayāna followers who are dull in their faculties and have difficulty in
penetrating the Sūtra and Vinaya, the Tathāgata teaches Abhidharma to
explicate [the Sūtra and Vinaya] so that they can understand. That is why
there are three baskets in the Hīnayāna tradition. The followers of the
Mahāyāna are sharp in their faculties and they can have instant profound
realization as soon as they hear the teachings of the Sūtra and Vinaya.
They do need not to bother the Tathāgata to teach the Abhidharma par-
ticularly to explicate the teaching for them, and therefore, the Mahāyāna
has no Abhidharma.35
32 T.48.2023: 1047a1–3: “ 安曰：先舊格義於理多違。光曰：且當分析逍遙，何容是

非先達？安曰：弘贊理教宜令允愜，法鼓競鳴何先何後 ?”..
33 Huiyuan 慧遠 (334–416), the other most influential Buddhist leader contemporary to
Dao’an, is said to have promoted the study of Abhidharma. However, he abandoned the
study of Abhidharma and Hīnayāna teachings as a whole probably under the influence of
Kumārajīva’s translated Mahāyāna texts (Zürcher, 2007: 231).
34 T.52.2102: 78b19–25: “外國風俗還自不同。提婆始來。義觀之徒莫不沐浴鑽仰。
此蓋小乘法耳。便謂理之所極。謂無生方等之經皆是魔書”.
35 T.44.1851: 469b15–22: “小乘備明三藏。大乘不論何故如是。此就如來本教故爾。
如來所化。小乘眾生。鈍根難悟。聞說經律。不能廣解。是故如來。重以毘
曇分別開示。方能悟入。故有三藏。大乘眾生。利根易悟。聞說經律。即能
深解。不假如來重以毘曇分別解釋。是故不具”.
Although during the time of Dao’an, Chinese Buddhists began to take an inter-
est in Abhidharma, there had been no serious study and in-depth understanding
of the Abhidharma system. Despite of the fact that a good number of
Abhidharmic texts were continuously introduced to China and translated into
Chinese, there was no sign of employing the Abhidharma method by the indig-
enous Buddhist exegetes.
Apparently Xuanzang recognized the weakness of Abhidharma studies in
Chinese Buddhism. Even just from the sheer size and variety of the Abhidharma
texts translated by Xuanzang we could not miss the importance that Xuanzang
attached to the Abhidharma. In what follows I will show how Kuiji defended
the importance of the Abhidharma and how he recontexualized Buddhist doc-
trines and concepts in the Abhidharma system.
4 Kuiji’s Defence and Advocacy of Abhidharma
Kuiji argued for the importance and necessity of Abhidharma as a methodol-

ogy for a decisive discernment of Buddha’s teaching and Buddhist concepts
based on the characteristics of Abhidharma themselves. In his explication of
the nature of the scripture, Kuiji explained the superiority of the Abhidharma:
The upadeśa36 refers to the mātṛkā that repeatedly discerns the sūtras.
All ultimate (nītārtha) teachings are called mātṛkā. This (upadeśa) is
where the World-honored One thoroughly distinguishes the characteris-
tics of the dharmas, and this is where the Buddha’s noble disciples, who
have penetrated into reality, distinguish accurately the characteristics of
the dharmas by their own realization of the truth. What is named the
mātṛka is also called Abhidharma. On account of this, after the parinirvāṇa
of the Buddha, only with respect to upadeśa of the twelve divisions of the
teaching is there no difference between its being taught by the Buddha
and by the noble disciples. This is so with the Abhidharma of the three
baskets.37
36 According to the Sarvāstivādin (e.g. the Saṃyukāgama and the *Mahāvibhāṣā, Da pipo-
sha lun 大毘婆沙論) and several Mahāyana sūtras and śāstras (e.g. Pañcaviṃśatisāhas
rikāprajñāpāramitāsūtra, Yogācārabhīmiśāstra), upadeśa is one of the twelve divisions
of Buddhist scripture (dvādaśāṅga): 1. sūtra, 2. geya, 3 vyākaraṇa, 4. gāthā, 5. udāna,
6. nidāna, 7. avadāna, 8. itivṛttaka, 9. jātaka, 10. vaipulya, 11. adbhūtadharma, 12. upadeśa.
See Hirakawa, 1963: 61–62.
37 T.45.1861: 274 c1–12: “論議經者，謂諸經典循環研覈磨呾理迦，一切了義經皆名磨
呾理迦。謂於是處，世尊自廣分別諸法體相，又於是處，諸聖弟子已見諦
306 Teng
From this passage we see that Kuiji defended the Abhidharma by highlighting
its uniqueness and superiority to other teachings. Since the Abhidharma
teaches the ultimate teaching, it could be relied upon even by the Mahāyāna
followers.
Furthermore, as there is in fact an Abhidharma text in the Mahāyāna teach-
ing, Kuiji could defend the importance of Abhidharma from the Mahāyāna
point of view. Kuiji quoted Asaṅga’s Abhidharmasamuccaya and its Vyākhyā by
Sthiramati and, in his subcommentary, entitled Zajilun shuji 雜集論述記 to
Sthiramati’s Abhidharmasamuccayavyākhyā, Kuiji pointed out the indispens-
ability of Abhidharma for the discernment of the true nature of the dharmas:38
The purpose of composing the Mahāyāna-abhidharma-samuccaya was

because people are from time immemorial ignorant of the characteristics
of the dharmas as they really are. Because of this reason, obstinate as
they were, they committed two kinds of [unwholesome] actions and
received three kinds of sufferings as a result.
The Tathāgata taught the characteristics of the dharmas as they really
are, as occasion arose, in order to show that dharmas are in such and such
a way not empty and in such and such a way not of real existence. He
enabled people to get rid of ignorance and attachment and [assisted]
them according to their spiritual inheritance (zhongxing 種姓) to elimi-
nate completely all defilements and to attain the felicity of cessation.
After the parinirvāṇa of the Buddha, the deeds of Māra surged up cha-
otically, causing the competing emergence of the schools of thought
professing a view of the real existence of dharmas. [In such a context],
bodhisattva Nāgārjuna, drawing from the Mahāyāna scriptures, taught
迹，依自所證無倒分別諸法體相，此亦名為摩呾理迦，亦名阿毘達磨。由此
文故傳受釋云許佛滅後十二分中論議一分，三藏之內對法一藏通弟子說，不
許餘之十一分教，并餘二藏通弟子說。所以者何？唯此對法及論議教，弟子
所造義理與佛意無有異”.
38 This passage is found in Kuiji’s sub-commentary to Sthiramati’s Abhidharmasamucca
yavyākhyā: X48.796: 1c9–2a2: “教起所因者，瑜伽釋言：諸有情等無始時來於法實
相無知僻執，造二種業受三苦。果如來隨機為說實相，令知諸法如是非空，
諸法復由如是非有，既了諸法非有非空遠離無知疑惑，僻執隨種姓別起處中
行永滅諸障、證寂滅樂。佛涅槃後魔事糺紛，部執競興，多生有見，龍猛菩
薩採集大乘，遣相空教造中論等，究暢真宗除彼有執。聖提婆等諸大論師造
百論等，弘闡大義。由是眾生復起空見，然有菩薩懼廣文海未起聞思，預生
怯退。大聖無著具廣慧悲，集阿毗達磨經所有宗要，括瑜伽師地論一切法
門，令懼文海者初依略教易可受持、終耐多聞、能達大義、證菩提果、利樂
有情，亦兼為彼二乘種姓各依自乘，證自利果”. (Punctuation mine).
the doctrine of emptiness in order to reject the [real existence of] phe-
nomena. He composed the Mūlamadhyamakakārikā and other treatises
advocating his thought and rejecting attachment to the view of real exis-
tence. Subsequently, Āryadeva and other treatise masters composed
treatises such as the Catuḥśataka (Bai lun 百論) and so on, and propa-
gated extensively the meaning [of their teachings]. As a result, people
came to hold fast the view of emptiness. Then bodhisattva Asaṅga who
attained the first [bodhisattva] stage and obtained the concentration of
dharma-luminosity (faguang ding 法光定) and great supernatural pow-
ers, invited the Noble Maitreya to teach the Treatise of the Yogācārabhūmi
to reveal the teaching of the “middle path” that eradicates attachments to
the views of emptiness and of real existence.
However, there are bodhisattvas who were intimidated by the vastness
of this text and withdrew from learning it. Endowed with great wisdom
and compassion, the Noble Asaṅga composed a compendium on the
“Abhidharmasūtra” that includes all teachings in the Treatise of the
Yogācārabhūmi.
In this passage Kuiji pointed out that the teaching of the Abhidharma was
originally intended by the Buddha for discerning the true nature of the dhar-
mas. And it is only by a clear discernment of the true nature of the dharmas
that the eventual elimination of suffering is possible. Kuiji also pointed out the
potential danger of attachment to the view of emptiness.
The more important contextualization that Kuiji effected, in my view, is
doctrinal contextualization in the Abhidharmic systematization. Abhidharma
is the foundational analytic tool, with a complex system of categorization of
sight into the external world we experience, as well as into ourselves as the
experiencing subject. It is granted that different Abhidharmic systems would
likely present the external world and internal self differently, but at least an
effective and grounded dharma discussion and argument can be possible.
5 Abhidharmic Recontextualization of Buddhist Doctrines and

Concepts
In the previous section, we saw some of the basic methodologies of the

Abhidharma system for the thorough discernment of the dharmas. The most
fundamental methodology is ‘categorization or inclusion’ (saṃgraha). In his
exposition of dharma-taxonomy, Kuiji not only introduced a variety of dharma-
taxonomies, but he also demonstrated how these different taxonomies are
308 Teng
related to each other and mutually subsumed. The following is Kuiji’s mapping
of the five groups of the one hundred dharmas (wu wei bai fa 五位百法) taxon-
omy against the taxonomies of the five aggregates (skandha), six sense-spheres
(āyatana), and eighteen elements (dhātu):39
The eleven types of matter [of the five-skandha-taxonomy] are subsumed

under rūpaskandha. Sensation of the universal [thought-concomitants]
(sarvatraga [caittas]) is subsumed under vedanāskandha. Perception
(saṃjñā) of the universal [thought-concomitants] is subsumed under
saṃjñāskandha. Seventy-three dharmas, namely the forty-nine thought-
concomitants40 and the twenty-four conditionings disjoined from thought
(cittaviprayukta saṃskāra), are subsumed under saṃskāraskandha. The
eight types of thought (citta) are subsumed under vijñānaskandha. [Of
the one hundred dharmas], excluding the six types of unconditioned,
ninety-four are subsumed under the five-skandha-taxonomy. Thus the
Abhidharmakośa says: “The skandha-taxonomy does not subsume the
unconditioned, because the two are not mutually associable”.
於色蘊中攝十一種色，受蘊攝遍行中受數，想蘊攝遍行中想數，行蘊
攝相應心所四十九法，及攝不相應行二十四法，合七十三法，識蘊唯
攝心法八種，總攝九十四法，唯除六無為。故俱舍云：蘊不攝無為義
不相應故。
Against the twelve-āyatana-taxonomy, the ten types of matter41 are sub-

sumed under the five internal and the five external āyatanas [of the
twelve-āyatana-taxonomy]. The eight types of thought are subsumed
under mana-āyatana. Subsumed under dharmāyatana are four types of
dharmas: namely the five types of matter associated with thought-objects
(dharmāyatanikāni rūpāni), the fifty-one thought-concomitants, the
twenty-four conditionings disjoined from thought, and the six un-condi-
tionings, eighty-six dharmas in total.
39 T.45.1861: 334b25-c7.
40 In the Theravāda Abhidhamma system, there are 51 thought-concomitants, but two
thought-concomitants, sensation and perception, are excluded in this context, because
they belong to vedanā- and saṃjñāskandha respectively, instead of to saṃskāraskandha.
41 In the Sarvāstivāda Abhidharma system, there are eleven types of matter listed in the
‘five-group-taxonomy,’ of which matter associated with thought objects (dharmāyatanikāni
rūpāni) is not subsumed under the rūpāyatana of the ‘twelve-āyatana-taxonomy’.
十二處攝者，內五處外五處攝十種色，意處攝八識，法處攝四類法。
所謂色法法處所攝色有五種。是相應法有五十一，不相應法二十四，
無為法有六。合有八十六法。
Against the eighteen-dhātu-taxonomy, the mapping of the five-skan-

dha- taxonomy is the same as against the twelve āyatanas, except that
mana-āyatana is substituted by the six thought-dhātus.42
十八界攝者，唯開意處立六識界，餘竝與處同
In his commentary, Kuiji employed dharma-taxonomies to analyze key con-

cepts of Buddhist doctrines. For example, in his commentary to the
Sukhāvativyūhasūtra 阿彌陀經, he employed these dharma-taxonomies to
analyze the term, “Land of Bliss”:43
If we situate the term, ‘Land of Bliss’ in the ‘five-group-taxonomy,’ ‘blissful’

(sukhāvatin) is subsumed under the category of thought-concomitants
(caitta-dharma), and ‘land’ under the category of matter (rūpa). If we
situate it in the five-āyatana-taxonomy, ‘blissful’ is subsumed under
the category of saṃskāraskandha, and ‘land’ under the category of
rūpaskandha. In the ‘twelve-āyatana-taxonomy,’ ‘blissful’ is subsumed
under the category of dharmāyatana, and ‘land’ under the category of
the rūpāyatana. In the ‘eighteen-dhātu-taxonomy,’ ‘blissful’ is subsumed
under the category of dharmadhātu and ‘land’ under the six sense objects.
五位分別者，極樂者心所收。世界者即色法攝。三科收攝者，極樂者
即行蘊收。世界者即色蘊攝。十二處中極樂法處收。世界即色處攝。
十八界中法界及六境中攝。
We will see now how such Abhidharmic recontextualization helps to resolve

some of the dilemmas in the Pure Land practice. In his Xifang yao jueshi yi
tonggui 西方要決釋疑通規 (General Principles on the Resolving of the Doubts
about the Essential Teachings of the West), a question was posed to Kuiji regard-
ing the reciting of Buddha’s name as a means for rebirth in the Pure Land.
I paraphrase:
42 Kuiji’s equation of mana-āyatana with the six thought-dhātus is a mistake, because the
mana-āyatana should include manodhātu in addition to the six thought-dhātus.
43 T.37.1758: 338a6–10.
310 Teng
Question: the Maitreya Inquiry Sūtra says: “Chanting the name of the
Buddha by an ordinary person whose mind is still fettered and defiled
will not lead to the rebirth in the Pure Land. If this is so, how do people
who are still at the stage of practicing and are yet to attain the noble
stage, ever strive to be reborn in the ‘Pure Land,’ since their mind is yet to
be purified?”44
Kuiji replied: “At the moment of chanting the Buddha’s name, the mind
that is operative is the sixth consciousness. It has this characteristic that
when it is taking the chanting of Buddha’s name as its mental object, it is
only associated (saṃprayukta) with the five universal thought-concomi-
tants bianxing wu suo 遍行五所 (sarvatraga-caitta) and the eleven
wholesome thought-concomitants shan shiyi suo 善十一所 (kuśala-
caitta). This thought of chanting Buddha’s name will not activate the
defiled thought-concomitants because there is no other mental object to
serve as the condition for their arising.”45
It should be clear now how such profound puzzles could be resolved system-
atically with the Abhidharma methodology. First of all, Kuiji applied the
method of categorizing (saṃgraha) to identify the locus of the thought of
chanting Buddha’s name. Then he applied the theory of association (sam-
prayoga) to determine what types of thought-concomitants are associated
with the thought of chanting Buddha’s name.
6 Conclusion
For Kuiji, as I have tried to show, the recontextualization strategy is to situate

Buddhist doctrines and concepts in the Abhidharma systematization. The
analytic exposition of Abhidharma that aims at investigating Dharma and
dharmas systematically and methodically is what had been lacking in the
Chinese reading of Buddhist doctrines and texts. Although the importation of
Abhidharma to China was quite early, for various historical and religious rea-
sons, such as the Mahāyāna-Hīnayāna complex, the Abhidharma study was
never established firmly in China.
44 T.47.1964: 105a20–23: “ 第四彌勒問經云: 念佛者非凡愚念，不雜結使念，得生彌

陀佛國。疑曰: 准此經說，夫念佛者, 非是凡愚，不雜結使，方成淨業。今修
行者，聖位未登，結使不除，如何得往 ?”..
45 T.47.1964: 105b18–21: “念佛之心即第六識。心王正起，欲作惡事，結使煩惱，容
可得生？正念佛時，與遍行五所，及善十一所，不動諸結。無因起故”.
I have shown the characteristics of Abhidharma and its role in Buddhism as

seen by the earlier Indian Buddhist tradition. According to the tradition,
Buddhist doctrines and concepts could be better understood and understood
decisively by analyzing them according to the Abhidharmic theories, which is
a precise and definite teaching (nītārtha). This feature of Abhidharma, as this
chapter tries to demonstrate, was recognized by Kuiji as what had been lacking
in Chinese Buddhism before his time. Kuiji therefore was keen to show how
Abhidharmic recontextualization can offer a common and normative founda-
tion for discussion, and argue about Buddhist doctrines and concepts, whereas
the eclectic panjiao or sectarian doctrinal orthodoxy would only lead to the
predicament of “ge shan zong tu 各擅宗塗 (each of them professing exclusive
orthodoxy of one’s teachings)” or “bing shuai ji qing 竝率己情 (depending all
on one’s own intention)”.
Abbreviation

邊 and Ono 小野, 1924–1934)
Bibliography
Bevans, Stephen B (2002). Models of Contextual Theology. Revised and Expanded Edition.
Maryknoll NY: Orbis Books.
Coe, Shoki (1993). Recollections and Reflections. New York: Formosan Christians for
Self-Determination.
Cox, Collett (1995). Disputed dharmas. Early Buddhist theories on existence: an annotated
translation of the section of factors dissociated from thought from Saṅghabhadra’s
Nyāyānusāra. (Studia philologica Buddhica, Monograph Series 11). Tokyo: The
International Institute for Buddhist Studies.
Demiéville, Paul (1951). “La Yogācārabhūmi de Saṅgharakṣa”, Bulletin de l’École française
d’Extrême-Orient 44, pp. 339–436.
Dhammajoti Bhikkhu KL (2007). Sarvāstivāda Abhidharma. Hong Kong: University of
Hong Kong.
Enomoto, Fumio (1996). “A Sanskrit Fragment from the Vibhāṣā Discovered in Eastern
Turkestan”, In Sanskrit-Texte aus dem buddhistischen Kanon: Neuentdeckungen und
Neueditionen III. Edited by Fumio Enomoto, Jens-Uwe Hartmann, and Hishashi
Matsumura. Göttingen: Vandenhoeck & Ruprecht, pp. 135–143.
312 Teng
Gimello, Robert (1976). “Chih-Yen (智儼 602–668) and the Foundation of Hua-Yen
Buddhism”. Unpublished PhD dissertation, Columbia University.
Gregory, Peter N (1991). Tsung-mi and the sinification of Buddhism. Princeton NJ:
Hayashima, Osamu, et al. (2003). Abhidharmasamuccaya-bhāṣya & vyākhyā. Sanskrit,
Tibetan, and Chinese texts Compared. <http://www.shiga-med.ac.jp/public/yugagyo/
AS/AS_ETEXT_V3_ALL.pdf> (accessed 08/27/2014).
Hirakawa, Akira (1963). The Rise of Mahāyāna Buddhism and its Realtionship to the
Worship of Stūpas. (Memoirs of the Research Department of the Toyo Bunko 22).
Tokyo: The Toyo Bunko.
La Vallée Poussin, Louis de (1930). “Documents d’Abhidharma”, Mélanges chinois et boud-
dhiques I, pp. 65–121.
Lopez, Donald S. (1988). Buddhist Hermeneutics. (Studies in East Asian Buddhism 6).
Honolulu HI: University of Hawaii Press.
Lusthaus, Dan (2002). Buddhist phenomenology: a philosophical investigation of Yogacara
Buddhism and the Ch’eng Wei-shih lun. (Curzon critical studies in Buddhism series).
London, New York NY: RoutledgeCurzon.
Müller, Edward (ed.) (1897). Dhammasaṅgaṇi-atthakathā (Atthasālilī). London: Pāli Text
Society.
Potter, Karl H.; Buswell, Robert E.; and Jaini, Padmanabh S. (1996). “The Development
of Abhidharma Philosophy”, In Encyclopedia of Indian Philosophie. Vol. 7: Abhidharma
Buddhism to 150 ad. Delhi: Motilal Banarsidass, pp. 73–119.
- T.37.1758: Kuiji 窺基, Amituo jing tongzan shu 阿彌陀經通贊疏.
- T.44.1851: Huiyuan 慧遠, Dasheng yizhang 大乘義章.
- T.45.1861: Kuiji 窺基, Dasheng fayuan yilin zhang 大乘法苑義林章.
- T.47.1964: Kuiji 窺基, Xifang yaojue shiyi tonggui 西方要決釋疑通規.
- T.48.2023: Rujin 如卺, Zimeng jingxun 緇門警訓.
- T.50.2053: Huili 慧立 and Yancong 彥琮, Da Tang Daci’en si sanzang fashi zhuan 大
唐大慈恩寺三藏法師傳.
- T.52.2102: Sengyou 僧祐, Hongming ji 弘明集.
Teng, Weijen (2011). Recontextualization, Exegesis, and Logic: Kuiji’s (632–682) Metho
dological Restructuring of Chinese Buddhism. Unpublished PhD dissertation, Harvard
University.
Tsukamoto, Zenryū (1985). A History of Early Chinese Buddhism. From its Introduction
to the Death of Hui-yüan. 2 vols. Tokyo, New York, San Francisco: Kodansha
International. (English translation of the original Japanese work by Leon Hurvitz.
Original publication: Chūgoku Bukkyō tsūshi. Tokyo: Shunjūsha, 1979).
Willemen, Charles (1975). The Essence of Methaphysics. Abhidharmahṛdaya. Translated
and annotated. (Série Etudes et Textes 4). Bruxelles: Institut Belge des Hautes Études
Bouddhiques.
Zürcher, Erik (2007). The Buddhist Conquest of China. The Spread and Adaptation of
Buddhism in Ealry Medieval China. Third Edition with a Foreword by Stephen F. Teiser.
Leiden: Brill.
314 Chapter 10
Chapter 10
Traces of Abhidharma in the bSam-gtan mig-sgron

(Tibet, Tenth Century)
Dylan Esler
1 Introduction: Metaphysical Premises
It is well known that the Abhidharma provides the metaphysical, cosmological

and phenomenological backdrop against which any form of Buddhist medita-
tive praxis takes place, the category of ‘contemplation’ (dhyāna) being
fundamental in the various Abhidharma lists.1
Written in the early tenth century by the master and scholar gNubs-chen
Sangs-rgyas ye-shes (844–early half of the tenth century),2 the bSam-gtan mig-
sgron is a Tibetan treatise on the subject of contemplation and one of the first
indigenous Tibetan doxographies.3 As such, it presupposes the context of the
Abhidharma, to which it makes many implicit references. These will be exam-
ined in the course of this chapter.
Before we begin elucidating these, a few comments concerning the term
abhidharma may be in order. Etymologically, the word dharma means that
which holds (dhāraṇa) a particular characteristic (svalakṣaṇa).4 The prefix
abhi expresses the idea of ‘towards,’ as well as that of ‘superiority’ or ‘pre-emi-
nence;’ it also can signify the notion of ‘in front of,’5 hence of being ‘face to face’
or ‘evident’ in the case of abhimukha.6 In the context that concerns us here, the
Abhidharma is the doctrine (dharma) that makes evident (abhi) the character-
* I would like to thank the late Lance S. Cousins and Rupert Gethin for their comments which
helped me to improve my paper in various ways.
1 Cousins, 1973: 115.
2 On gNubs-chen Sangs-rgyas ye-shes’ dates and the time of composition of the bSam-gtan
mig-sgron, see Esler, 2014: 5–27.
3 gNubs-chen Sangs-rgyas ye-shes, rNal-’byor mig-gi bsam-gtan or bSam-gtan mig-sgron: A trea-
tise on bhāvanā and dhyāna and the relationships between the various approaches to Buddhist
contemplative practice, Reproduced from a manuscript made presumably from an Eastern
Tibetan print by ’Khor-gdong gter-sprul ’Chi-med rig-’dzin, Smanrtsis shesrig spendzod, vol.74
(Leh: Tashigangpa 1974). Henceforth abbreviated as C.
4 de La Vallée Poussin, 1971, vol.1, ch.1: 4.
5 Monier-Williams, 2001: 61.
6 Conze, 1973: 63.

Traces of Abhidharma in the bSam-gtan mig-sgron 315
istics of phenomena (dharma); it is the teaching that systematizes the many

phenomena spoken of haphazardly in the Buddhist sūtras. Indeed, accord-
ing to Vasubandhu’s Abhidharmakośa, the Abhidharma is so-called because
it envisages (abhimukta) the supreme principle (dharma), i.e. transcendence
(nirvāṇa), or else because it envisages the particular and universal characteris-
tics of phenomena.7 Hence, the Abhidharma refers to the detailed analysis of
the characteristics and meaning of phenomena, including the effect of their
maturation (vipākaphala).8 It further refers to the understanding (prajñā) that
investigates phenomena, which can signify both their absolute and relative
truth.9 Far from being an innovation or a deviation, the Abhidharma must in
fact be considered as an elucidation of the full ramifications of the Buddha’s
doctrine carried out by scholars working far from the hustle and bustle of the
world, who sought to draw all the subtle consequences of their master’s teach-
ing.10 Its formal trait is its almost exclusive concern with doctrinal exposition
divested of any literary or narrative embellishment.11
If the Abhidharma systematizes phenomena and makes evident their char-
acteristics, what, then, are these ‘phenomena’? The above definition of dharma
should make it clear that it is futile to try to determine what dharmas are in
terms of content,12 since they only make sense in relation to their function,
which is to hold a particular characteristic. A particular characteristic may be
illustrated, for example, by the solidity of the earth, the humidity of water, the
heat of fire and the motility of wind.13 Neither can dharmas be conceived of in
terms of the opposition between psychic and physical phenomena, nor can
they be said to be attributes of a substance.14 A dharma is a theoretically
descriptive term for any object to be described while simultaneously signifying
the main object of description.15 It is an existent that cannot be reduced to any
component parts, a substantial existent (dravyasat). Such substantial existents
are to be distinguished from linguistic existents (prajñaptisat) – such as chairs,
tables, etc. – which exist merely by virtue of their being the referents of con-
ventional designations. These linguistic existents are complex compounds
made of dharmas, i.e. substantial existents. It is possible to see the Abhidharmic

8 Lamotte, vol.2, 1949: 1078.
9 Hirakawa, 1980: 167–168.
10 Lamotte, 1958: 686.
11 Lamotte, 1958: 197.
12 Piatigorsky, 1984: 180–181.
13 Lamotte, vol.5, 1980: 2186–2187.
14 Griffiths, 1999: 56.
15 Piatigorsky, 1984: 174.
316 Esler
enterprise as an attempt to translate coherently and consistently all discourse

about linguistic existents into a discourse about substantial existents.16
While we cannot discuss here the challenges of translating the word dharma,
suffice it to say that it seems important to render it in such a way that the pro-
cess-character (or ‘non-thingness’) of this term be preserved.17 For example,
whereas generally speaking Edward Conze avoids translating the word dharma,
he implicitly suggests that it can accurately be rendered as ‘event,’ which
indeed does justice to the functional and dynamic character of this term in the
Buddhist philosophical context.18 For similar reasons, I have, like many others,
chosen the word ‘phenomenon’ to render dharma, because etymologically this
term signifies that which appears, which shows itself, presences itself, or lights
up,19 and hence points to a process of appearing rather than a static thing, and
because semantically it covers a wide range of applicable meanings. At times,
I have opted for ‘principle’/‘principial’ or ‘doctrine,’ according to context.
While some scholars hold the Buddha’s doctrine to be fundamentally anti-
metaphysical and radically empirical,20 this characterization seems to be quite
problematic,21 not the least because it reduces Buddhism to a system of ethics
devoid of any metaphysical basis and because it reduces metaphysics to a par-
ticular branch of philosophy characterized by ratiocination, which, following
Martin Heidegger,22 is to be overcome.23 Addressing the first of these errors, to
present Buddhism as a mere moral teaching designed to uplift man from his
mediocre sufferings is nothing short of travesty and profanation. Indeed, from
its inception, Buddhism has been characterized by a radical orientation
towards the unconditioned, towards that which transcends both life and death.
This is what enabled this doctrine to point out with unmatched rigor that the
source of suffering (duḥkha) is the continual thirsting (tṛṣṇā) for existence, a
thirst which is quenched in extinction (nirvāṇa).24 Moreover, the prominent
place accorded to metaphysics by Buddhist philosophers is illustrated by the
16 Griffiths, 1999: 50–51.

17 In this regard it may be profitable to recall Herbert V. Guenther’s emphasis on the dynamic
and process-orientated character of much of Buddhist thought, which is reflected in this
doctrine’s insistent refusal to remain satisfied with an eternal and static ‘self’. Cf. Guen-
ther, 1989: 1–8, 245–248.
18 Conze, 1962: 99.
19 Heidegger, 2006: 51. See also Partridge, 1982: 199.
20 Cf. Kalupahana, 1994: 43–44, 50.
21 See Conze’s critique of Kalupahana’s position in Conze, 1980: 51.
22 Cf. Heidegger, 2000: 41, 217; and Heidegger, 2003: 14–15, 56, 81.
23 Cf. Levin, 1985: 56, 61, 121, 307.
24 Evola, 1982: 138.
fact that they held an accurate knowledge of the phenomena comprising real-
ity to be the necessary condition to effectively remove the contaminations
(kleśa).25
Turning to the second error, the tendency in the West to reduce metaphysics
to a mere branch of philosophy has led to the widespread opinion that it is
little more than mental activity concerned with abstractions. More often than
not, the rejection of metaphysics betrays a lack of understanding concerning
both its nature and scope.26 Traditionally speaking, metaphysics is the science
which deals with the nature of reality, and hence which enables man to discern
the real from the illusory. It is, moreover, integrally wed to methods enabling
the spiritual realization of this knowledge and, as scientia sacra, lies at the
heart of all the other traditional sciences.27 In the Buddhist case, the close con-
nection between metaphysics and meditative praxis is illustrated by the fact
that the taxonomic enterprise of the Abhidharma probably had its origin in
certain early Buddhist techniques of meditative introspection, where the prac-
titioner deconstructed through analysis his everyday experience, and labeled
and identified its individual components.28
While the ontological difference, i.e. the distinction between the ontic (the
concern with beings and with existence) on the one hand, and the ontological
(the concern with Being) on the other, articulated by Martin Heidegger29 is
crucial philosophically, it fails metaphysically to take into account that which
lies beyond Being and transcends manifestation.30 This failure means that it
does not go beyond the domain of philosophy. What distinguishes philosophy
from metaphysics is precisely its systematic character, which makes all philo-
sophical systems, whatever their orientation, of little value metaphysically,
because they are closed and seek to analyze a particular aspect of reality, all
the while excluding those aspects that do not conform to their limited point of
view. While it may be legitimate to investigate a particular possibility (which is
what amounts to a science), it is certainly illegitimate to affirm that this is the
whole of reality and to deny all that goes beyond the narrow confines of this
limited understanding.31
25 Griffiths, 1999: 49.

26 Borella, 2004: 211, note 12.
27 Nasr, 1989: 132.
28 Griffiths, 1999: 54–55.
29 Heidegger, 2006: 26, 31. See also Levin, 1985: 51.
30 Nasr, 1989: 136.
31 Guénon, 1984: 19.
318 Esler
Furthermore, whereas the organ of philosophical investigation is reason, in

knowledge of a metaphysical order rational thought can only have the second-
ary role of circumstance to awaken intellectual intuition, which always occurs
simultaneously and suddenly.32 Hence, intellection does not reach to absolute
certitude through reasoning, but rather through an a priori direct intuition of
the supra-formal truth. Therefore, the rationalistic arguments against meta-
physical knowledge presented by those in whom reason is severed from the
intellect and hence operates on false premises cannot in the least negate intel-
lection, which does not depend on reason in the first place. This is not, however,
to say that intellection is illogical or irrational, since the very categories of logic
are reflections on their own plane of intellectual certitudes. This is why reason
can serve intellection, not as a cause, but as a means of expressing principles of
a metaphysical order.33
If the perspective of the Abhidharma can be said to be metaphysical, it also
includes cosmology. This in itself is not surprising, since cosmology is itself
derived from metaphysics. It therefore always takes account both of the mac-
rocosm and the microcosm, since the two dimensions are interdependent and
analogous to each other. Furthermore, when cosmology looks at the physical
dimension, it never loses sight of its link to the subtle manifestation which
stands immediately above it.34 The implications of this are illustrated by the
four contemplations (dhyāna) and four formless absorptions (samāpatti),
which are both meditative states and conditions of existence brought about by
contemplation.35 At the same time, a metaphysical perspective like that of the
Abhidharma will tend to emphasize the relativity of all that pertains to the
cosmos,36 as is apparent for instance in its teaching of multilayered universes
and in its cyclical conception of time.
Before we turn to the actual references to the Abhidharma in the bSam-gtan
mig-sgron, a final word of caution is in order. The apparent absurdity of certain
cosmological descriptions stems from the fact that the human mind is inca-
pable of understanding multifarious aspects simultaneously and of expressing
that which surpasses its limits. It can only contain a broken and somewhat
absurd image of the realities beyond it, an image that is bound to be marked by
certain contradictions and over-simplifications. This situation is analogous to
trying to describe a polyphonic melody or to explaining the play of the ocean’s
32 Schuon, 1995: 108–109.

33 Nasr, 1989: 148.
34 Guénon, 2005: 261.
36 Schuon, 1993: 53.
waves to a blind person who has never seen the sea.37 Having laid these meta-
physical premises, we may now proceed to examine the individual traces of
Abhidharma in the bSam-gtan mig-sgron, which we have classified under the
rubrics of cosmological, phenomenological and meditational categories.
2 Cosmological Categories
2.1 The Third Order Chiliocosm

The first cosmological reference occurs in Chapter I (§3)38 and concerns the
third order chiliocosm. The passage in question appears in the context of dis-
cussing the qualities of the retreat place suitable to contemplative praxis. After
giving various specifications in this regard, gNubs-chen Sangs-rgyas ye-shes
assumes the perspective of the Mantrayāna and Atiyoga paths, declaring “that
the third order chiliocosm’s genuine abode is the wisdom of intrinsic aware-
ness, the great expanse of mother-like space”.39 What this indicates is that
solitude is not to be sought outside, since for the adept who has realized inner
detachment, primordial solitude is always present, even within the external
world signified by the third order chiliocosm.
Of course, the third order chiliocosm (trisāhasramahāsāhasralokadhātu;
Tib. stong-gsum-gyi stong chen-po’i ’jig-rten-gyi khams) is a concept derived
from the Abhidharma cosmology, where the universe is divided into three
chiliocosms, of which this is the third. The first order chiliocosm encompasses
one thousand worlds, each of them identically structured, with Mount Meru
at the center, surrounded by four major continents and eight subcontinents.
The second order chiliocosm has one thousand times as many worlds as the
first, and the third order chiliocosm has one thousand times as many worlds
as the second.40 As we have seen, while the passage in question evokes the
internalization of solitude and its paradoxical presence within cyclic existence
(saṃsāra), it evidently assumes the cosmological structure of the chiliocosm
as its backdrop. 
37 Schuon, 1995: 40.

38 Paragraph numbers correspond to editorial divisions introduced in the English transla-
tion and critical edition I have been preparing of the bSam-gtan mig-sgron.
39 C 8.2–3: […] rang rig pa’i ye shes nyid […] yum gyi mkha’ klong chen po dang/ […] stong
gsum nyid yang dag pa’i gnas su bka’ stsal to/.
40 de La Vallée Poussin, 1971, vol.2, ch.3: 170; Cornu, 2006: 158–160.
320 Esler
2.2 The Seven Mountains

In a catalogue of dreams in Chapter IV indicating spiritual progress on the
gradual path,41 there is a further allusion to this cosmological framework. In
a dream where one stands on a mountain peak, depending on which of the
seven mountains one stands on, this signifies that one will view one of the
first seven bodhisattva stages (§2.4.71).42 This is a reference to the seven moun-
tains surrounding Mount Sumeru, viz. Yugaṃdhara, Īṣādhara, Khadiraka,
Sudarśana, Aśvakarṇa, Vinataka and Nimiṃdhara.43 One of these seven moun-
tains, Yugaṃdhara, is also mentioned by name in Chapter III (§3.2.3), where
a brief criticism of the approach of Mahāyoga is offered: it is explained that
through overly yearning for non-duality and focusing on the tantric means of
generating emanations, this approach obscures the spontaneous complete-
ness which is beyond all action. Such a condition is likened to not seeing
the peak of Mount Meru from the flank of Yugaṃdhara.44 Now, according to
the Abhidharma cosmology, upon the golden earth are Mount Meru and the
seven mountains, surrounded by Cakravāḍa; the latter is so-called because it
encircles the universe, thereby giving it the form of a wheel. The seven moun-
tains are golden, whereas Cakravāḍa is made of iron. Mount Meru has four
flanks, which are made of gold in the north, of silver in the east, of beryl in the
south and of crystal in the west. The beryl color of the southern flank is what
makes the sky appear the way it does in Jambudvīpa, which corresponds to
the human world. These various jewels are said to have originated when the
waters, endowed with numerous potentialities, fell upon the earth and were
subsequently blown away by the winds.45
2.3 The Inhabitants of the Buddhist Cosmos

The continents surrounding Mount Meru are inhabited by various beings. Some
of these are briefly touched upon in the bSam-gtan mig-sgron in a passage in
Chapter VI concerning one of the views of the tantric approach, the view of
non-duality (§1.5.6). Listing the faults of duality, Sangs-rgyas ye-shes mentions
the fact that if there were a fundamental duality between Buddhas and sen-
tient beings, they would be unable to appear to each other. He compares this
41 This catalogue of dreams, itself derived from the Ratnakūṭasūtra, has been studied in:
Esler, 2012a: 317–328.
42 C 102.1–2; see rMi-lam bstan-pa, in dKon-mchog brtsegs-pa’i mdo, in KD, vol.39, 412.1,
449.2–4.
43 Csoma de Koros, 1980: 45.
44 C 63.
45 de La Vallée Poussin, 1971, vol.2, ch.3: 141–142.
hypothetical situation to that which exists for the humans of the southern con-
tinent of Jambudvīpa, “who are unable to show the path to the ghastly people of
the northern continent of Uttarakuru”.46 The presumable reason the people of
Uttarakuru are here termed ‘ghastly’ is found in the Abhidharmakośa, where it
is explained that these beings cannot detach themselves from the desire realm
(kāmadhātu) and so are unable to enter contemplation. Furthermore, they are
also incapable of taking up the vows of individual release (pratimokṣa).47
2.4 The Gods of the Abhidharma Cosmology

Buddhist Abhidharma cosmology integrated the gods of the Brahmanical pan-
theon and effectively neutralized them by depriving them of any transcendent
status. In other words, they are exalted inhabitants of exalted worlds, whose
delusion and pride binds them to the more refined levels of cyclic existence.
Nonetheless, they continue to play a significant role in Buddhist discourse,
whether in the biographical accounts of the Buddha48 or in the metaphysical
Prajñāpāramitā sūtras,49 their main purpose being to elevate the Buddha and
his doctrine through their own (relative) mediocrity.
In the above mentioned catalogue of dreams, several of the oneiric signs
concern encounters with these gods. For instance, depending on whether one
sees the Thus-gone One (Tathāgata) surrounded by the four great kings, by the
Śakras, by the Brahmās or by the gods of the ethereal classes, this is interpreted
to mean that one will view the sixth, seventh, eighth or ninth stages of the
bodhisattva path (§2.4.36).50 In another dream, it is depending on whether
one sees one of the four great kings, a god of the Trāyastriṃśa heavens, a god
from the Tuṣita heaven, a god from the Brahmakāyika heavens, or a god of the
ethereal abodes (śuddhāvāsa) that one will view one of the sixth to tenth bod-
hisattva stages (§2.4.81).51 A further allusion to these gods is found in Chapter
VI in a passage dealing with twenty defects pertaining to the gradual tantric
46 C 209.6–210.1: […] sangs rgyas sems can tha dad na ’brel ba med pas snang du mi btub ste/
lho’i ’dzam bu gling gi mi sgam pos/ byang gi sgra mi snyan gyi mi ma legs pa’i lam bstan du
mi btub la/ […].
47 de La Vallée Poussin, 1971, vol.2, ch.3: 183; vol.3, ch.4: 104.
48 For instance, Brahmā’s repeated pleas addressed to the Buddha in the weeks following his
enlightenment finally convince him to teach his doctrine. See Lamotte, 1947: 52, 54.
49 Note for example Śakra’s role as interlocutor of the Buddha in these sūtras. See Conze,
1994: ch.3: 106; ch.31: 283, 296.
50 C 93.5–94.3; see rMi-lam bstan-pa, in dKon-mchog brtsegs-pa’i mdo, in KD, vol.39, 410.5–6,
428.5–429.4.
51 C 104.4–6; see rMi-lam bstan-pa, in dKon-mchog brtsegs-pa’i mdo, in KD, vol.39, 412.3,
454.6–455.3.
322 Esler
approach (§2.2.4.1.4.2.1).52 These defects cause one to fall into the first to fourth
levels of the gods of the desire realm. We will now turn to the Abhidharmakośa
to see how it classifies these different categories of divine beings.
There are six gods of the desire realm: the Cāturmahārājakāyikas, the
Trāyastriṃśas, the Yāmas, the Tuṣitas, the Nirmānaratis and the Paranirmita
vaśavartins. Like their human counterparts, the gods of the first four levels
enjoy the desirable sense-qualities that are at their disposal. It is only the gods
of the fifth and sixth levels of the desire realm who can emanate the objects
of their desire (kāma).53 The head of the Paranirmitavaśavartins is Māra, who
rules the desire realm and constantly seeks to oppose the Buddha.54
The four great kings (cāturmahārājakāyikas) are the rulers of the first and
lowest of the six classes of gods of the desire realm; they are Ḍhṛtarāṣṭra, king
of the east and lord of the scent-eaters (gandharva); Virūḍhaka, king of the
south and lord of the kumbhaṇḍa demons, who are pictured with huge stom-
achs and genitals; Virupākṣa, king of the west and lord of the serpent-spirits
(nāga); and Vaiśaravaṇa, king of the north and lord of the goblins (yakṣa). As
protectors of the world (lokapāla), their function is to examine the religious
conduct of human beings and report back to the Trāyastriṃśa gods.55
The Trāyastriṃśa heavens represent the second among the six classes of gods
of the desire realm and are located in the upper half of Mount Meru.56 They
are in a sense intimately associated with the Abhidharma, since it is believed
that the Buddha first taught the Abhidharma to his deceased mother who had
taken birth in the Trāyastriṃśa heavens. It is only later that Śāriputra received
these teachings and passed them on to his disciples.57 Śakra is an epithet of
Indra,58 who, in Buddhist cosmology, is the lord of the gods of the Trāyastriṃśa
heavens and of the lower levels of the desire realm.59
Tuṣita is the fourth among the six classes of gods of the desire realm and
is located in the space above Mount Meru; it is here that a bodhisattva of the
tenth stage dwells while waiting to take his final birth as a Buddha.60
To gain an impression of the decreasing materiality of these different gods, we
may recall that the Abhidharmakośa informs us that the Cāturmahārājakāyikas,
52 C 233.2–5.
53 de La Vallée Poussin, 1971, vol.2, ch.3: 164, 166.
54 Lamotte, 1958: 761.
55 Lamotte, 1958: 759–760.
57 Lamotte, 1958: 200.
58 Cf. Amarasiṃha, 2008, ch.1, svargavarga, v.42: 23.
59 Lamotte, 1958: 760.
60 Lamotte, 1958: 761.
along with the next category, the gods of the Trāyastriṃśa heavens, abide
on the earth and engage in sexual union just like their human counterparts.
However, they appease their desire by emitting wind, since they do not
carry a seminal substance. The Yāmas appease the fire of desire by embrac-
ing, the Tuṣitas by holding hands, the Nirmāṇaratis by laughing, and the
Paranirmitavaśavartins by simply looking at each other. Furthermore, the gods
above the Trāyastriṃśa heavens (viz. the Yāmas, the Tuṣitas, the Nirmāṇaratis,
the Paranirmitavaśavartins, and the gods of the form realm) dwell in seraphic
mansions (vimāna; Tib. gzhal-yas-khang).61
The Brahmās inhabit the Brahmakāyika heavens, the first among the three
abodes of the first contemplation (bsam-gtan dang-po’i gnas-gsum) of the
form realm (rūpāvacara), which will be examined below. The form realm
(rūpadhātu), located above the desire realm, consists of four worlds, corre-
sponding to the first contemplation (two or three levels), the second and third
contemplations (three levels each) and the fourth contemplation (eight lev-
els). The different levels of the form realm are superposed as terraces.62 The
beings in the Brahmakāyika heavens have cleaned away their sins and dwell
on the four abodes of Brahmā (brahmavihāra); these are synonyms of the four
immeasurables when taught within the context of the Hīnayāna. The mea-
sure of the merit of the Brahmakāyika gods is that they abide blissfully in this
heaven for an aeon (kalpa).63
In these heavens is found a high abode inhabited by a single sovereign,
Mahābrahmā, which corresponds to the third abode of the first contempla-
tion. The gods inhabiting the Brahmakāyika heavens are characterized by the
diversity of their bodies on the one hand, and by the uniformity of their per-
ceptions on the other. For example, although there may be differences in their
height, etc., if Brahmā thinks that the other gods are created by him, his aco-
lytes too will think that they are created by Brahmā.64
The ethereal classes (gtsang-rigs), also called ethereal levels (śuddhāvāsa
kāyika; Tib. gtsang-ris) or ethereal abodes (śuddhāvāsa; Tib. gnas gtsang-ma),
which are five in number, are located just under Akaniṣṭha, which is at the
peak of the form realm (rūpadhātu; Tib. gzugs-khams) and corresponds to
the fourth contemplation. Only those having perfected the path of seeing
(darśanamārga; Tib. mthong-lam) can be born there. The gods of the ethereal
61 de La Vallée Poussin, 1971,vol.2, ch.3: 164.

62 de La Vallée Poussin, 1927: 71–72.
64 de La Vallée Poussin, 1971, vol.2, ch.3: 3–4, 17.
324 Esler
classes are unable of going to the formless realm (ārūpyadhātu; Tib. gzugs-
med-kyi khams) or of falling to the lower realms.65
3 Fundamental Phenomenological Categories
Aside from these cosmological references, the bSam-gtan mig-sgron also

contains several allusions to some of the Abhidharma’s fundamental phenom
enological categories.
3.1 The Four Conditions

Such is for instance a reference to the four conditions which occurs in Chapter
I (§7) in a gloss to the text. It concerns the gradual approach (called Sautrāntika
Madhyamaka), which holds that the nature of reified entities is to assemble
interdependently in relative truth. The gloss in question then specifies that the
reified entities appear due to the assemblage of the four conditions.66
The four conditions alluded to in passing in this gloss are as follows: (1) the
causal condition, (2) the immediate condition, (3) the condition qua cognitive
object and (4) the ruling condition. They are explained as follows in Vasu
bandhu’s Abhidharmakośa:
(1) The causal condition (hetupratyaya; Tib. rgyu’i rkyen) refers to five types
of causes (viz. the co-occurring cause, the homologous cause, the concomitant
cause, the ubiquitous cause and the cause of maturation). To understand what
these five causes are, we need to take a brief look at Vasubandhu’s description
of the six types of causes, viz. (i) the instrumental cause, (ii) the co-occurring
cause, (iii) the homologous cause, (iv) the concomitant cause, (v) the ubiqui-
tous cause and (vi) the cause of maturation.
(i) The instrumental cause (kāraṇahetu; Tib. byed-pa’i rgyu): all phenom-
ena may be the instrumental cause of all phenomena, excepting themselves.
(ii) The co-occurring cause (sahabhūhetu; Tib. lhan-cig ’byung-ba’i rgyu)
consists of those phenomena which produce an effect on each other, such as
the elements (bhūta), the mind and that which follows the mind (cittānuvartin),67
as well as the characteristics and that which is characterized.
(iii) The homologous cause (sabhāgahetu; Tib. skal-ba mnyam-pa’i rgyu)
means that the cause is homologous with the effect. For instance, phenomena
that are contaminated are the cause of contaminated effects.

66 C 22.3 (gloss): rkyen bzhi ’dus kyi dngos po/.
67 This includes the mental events (Skt. caitta).
(iv) The concomitant cause (saṃprayuktakahetu; Tib. mtshungs-par ldan-

pa’i rgyu) refers to the mind and mental events.
(v) The ubiquitous cause (sarvatragahetu; Tib. kun-tu ’gro-ba’i rgyu) refers
to the cause of contaminated phenomena only. Past or present ubiquities of a
given stage are the causes of future contaminated phenomena of the same
stage.
(vi) The cause of maturation (vipākahetu; Tib. rnam-par smin-pa’i rgyu)
refers to the cause of impure phenomena as well as to the cause of phenomena
that are virtuous yet contaminated.68
It will be clear that the five causes referred to by the term ‘causal condition’
are numbers ii-vi in the above classification and exclude the instrumental
cause (which is included, as will be seen below, under the ruling condition).
Hence, it can be said that the causal condition designates the root springs of
action in cyclic existence, these springs being the three poisons of passion,
hatred and stupidity. We may now proceed with our analysis of the three
remaining conditions.
(2) The immediate condition (samanantarapratyaya; Tib. de-ma-thag-pa’i
rkyen) refers to the mind and mental events that immediately precede (anan-
tara) and are equal to (sama), in the sense of being of the same type as, a
subsequent moment of consciousness.
(3) The condition qua cognitive object (ālambanapratyaya; Tib. dmigs-pa’i
rkyen) can refer to all phenomena, since they all can be an object of conscious-
ness. For example, visual consciousness has as its condition qua cognitive
object all visual objects, i.e. the sense-element of form. Even if an object is not
referentially imaged (ālambyate; Tib. dmigs-pa) by consciousness, it is still an
imaged object of cognition, since its nature remains the same, just as a com-
bustible is always a combustible, whether or not it is being burnt.
(4) The ruling condition (adhipatipratyaya; Tib. bdag-pa’i rkyen) refers to
the instrumental cause (kāraṇahetu); for example, in the production of the
various consciousnesses, the visual faculty is the ruling condition for the pro-
duction of the visual consciousness. Its activity (kāritra) consists in not
obstructing a past, present or future phenomenon. Hence, its purpose is to
account for a condition’s continued existence after it has produced an effect
due to other conditions, such as the causal condition mentioned above.69
These four conditions must be understood as the relations existing between
various phenomena. It is precisely because of the Ābhidhārmikas’ exhaustive
enumeration and classification of phenomena that they were compelled to
68 For the six causes, see de La Vallée Poussin, 1971, vol.1, ch.2: 245–271.
69 For the four conditions, see de La Vallée Poussin, 1971, vol.1, ch.2: 299–309; and Kalupah-
ana, 1994: 148–149.
326 Esler
focus on their relations, failing which they would have been left with lists of
disconnected phenomena.70 Incidentally, it might be noted that Nāgārjuna
himself, despite his deconstruction of Abhidharma phenomenological catego-
ries, had no objection to this formulation of conditional relations, as long as
the links in the causal chain were not taken to have an intrinsic nature
(svabhāva).71
3.2 Temporal Divisions

Several of the Abhidharma’s temporal divisions are evoked in the bSam-gtan
mig-sgron. The most basic of these is the moment (kṣaṇa): it is defined as the
time it takes for a phenomenon to be produced if the given conditions are
assembled. It may also be defined as the time it takes for a phenomenon in
motion to move from one atom (paramāṇu) to another.72 Chapter V of the
bSam-gtan mig-sgron explains that although in the simultaneous approach of
Chan the mode of purifying the obscurations is indefinite, the flaws are puri-
fied in the moment of completing an action (§4).73 A reference to the moment
of completing an action is also found in Chapter VI, where it is explained that
for those on the Hīnayāna this is one of the possible durations it takes to attain
the fruition (§4.4).74
The moment of completing an action (bya-rdzogs-kyi skad-cig-ma) is a tech-
nical term of Buddhist metaphysics. It refers to the period of time it takes to
accomplish any given action. Depending on the action in question, the time
needed to complete it will vary.75
This is in fact one of two temporal divisions of the moment, the other being
the infinitesimal moment (dus-mtha’i skad-cig-ma). According to the Vaibhā-
ṣikas, the infinitesimal moment is defined as follows: in the time it takes for a
young man in optimal health to snap his fingers, there are sixty-five infinitesi-
mal moments.76
The opposite of the moment is, it might be said, the aeon, four of which are
counted in the Abhidharma: (1) the intermediate aeon (antarakalpa; Tib. bar-
gyi bskal-pa), or small aeon, eighty of which make up a larger aeon; (2) the
aeon of destruction (saṃvartakalpa; Tib. ’jig-pa’i bskal-pa), which lasts from
70 Kalupahana, 1994: 149.

73 C 183.4: ’di ni ’dag lugs nges pa med kyang bya ba rdzogs pa’i skad cig ma la ’dag go/.
74 C 275.5–6.
the time when no more denizens of hell are born up to the destruction of the
universe; (3) the aeon of formation (vivartakalpa; Tib. chags-pa’i bskal-pa),
which lasts from the primordial wind (prāgvāyu) up to the time when are born
denizens of hell; and (4) the great aeon (mahākalpa; Tib. bskal-pa chen-po),
defined as the period of one universe, lasting from its inception to its
destruction.77
Several allusions to aeons are found in the bSam-gtan mig-sgron, again in
the context of defining the time it takes to accomplish the fruition. We find a
reference to the so-called incalculable aeon (asaṃkhyeyakalpa) in the same
passage in Chapter VI just mentioned, which gives the various possible dura-
tions required to attain the fruition on the Hīnayāna.78 In the same chapter,
a passing allusion to the aeon is also made in the section describing the
Mahāyoga’s relative superiority to Chan: the latter approach is here depreci-
ated for taking aeons in order to reach the eleventh bodhisattva stage, the stage
of universal light (§5.1.11).79 We have yet another reference in a quotation from
the Saṃdhinirmocanasūtra cited in Chapter V (§4):80
To cleanse the imprints of the atrophies, it may take three incalcula-

ble aeons, a juncture (ṛtu),81 a year, a month, twenty-four hours, a day,
a session, half a session, an instant, a fleeting second, a minute or an
immeasurable aeon.
Since it is said that it takes three incalculable aeons (asaṃkhyeyakalpa; Tib.

bskal-pa grangs med-pa) to reach awakening, the question arises as to which
type of aeon is being referred to. The answer is given unequivocally: it refers to
a great aeon.82 Somewhat paradoxically, according to the Abhidharmakośa, an
incalculable aeon is not actually incalculable, but merely represents an
extremely high number. If one takes 1 as a starting point, the second point
would be 10 x 1 (=10), the third point 10 x 10 (=100), the fourth point 10 x 100
(=1,000), etc., up to the sixtieth point (lit. ‘extrinsic place’, sthānāntara; Tib.
gzhan-gnas), which is the incalculable aeon. If this were not the case, it would

78 C 275.5.
79 C 287–288.
80 Quoted as San-de’i mdo in C 184.2–3. Lamotte, 1935, ch.9, §30: 146 (Tibetan text), 255
(French translation). This Saṃdhinirmo-canasūtra citation is also found in Pelliot tibé-
tain 116 (fol. v36.7-r37.1).
81 Note that dus has been emended to dus-tshigs (‘juncture’) according to the Saṃdhinirmo
canasūtra.
328 Esler
indeed be impossible to speak of the bodhisattva’s career as lasting three incal-

culable aeons.83
3.3 The Psycho-physical Makeup of the Concept ‘Person’

The Buddhist analysis of the person reveals that what is mistakenly assumed to
be an individual is in fact a complex compound of psycho-physical phenom-
ena. These are variously classified as the five aggregates (skandha), twelve
sense-spheres (āyatana) and eighteen sense-elements (dhātu), depending on
the dispositions and inclinations of the doctrine’s recipients. Vasubandhu
explains that the reason that the aggregates, sense-spheres and sense-elements
are spoken of by the Hallowed One is that there are (1) three types of stupidity,
viz. mistaking mental events for a self, mistaking form for a self, and mistaking
both the mental events and form for a self; (2) three types of faculties; and (3)
three types of preferences (ruci; adhimokṣa), viz. for a summarized exposition,
a middling exposition, and an extensive exposition.
In this way, the teaching on the aggregates is intended for those auditors
who mistake the mental events for a self, who are of sharp faculties and who
prefer a summarized exposition; the teaching of the sense-spheres is intended
for auditors who mistake form for a self, who are of middling faculties and who
prefer a medium-length exposition; and the teaching of the sense-elements is
intended for those auditors who mistake both form and the mental events for
a self, who are of average faculties and who prefer an extensive explanation.84
The bSam-gtan mig-sgron makes reference to the five aggregates in Chapter
IV in the section on genuine relative truth (§2.2.1.1): here it is explained that all
the various enumerations of relative truth, including the five aggregates – form
(rūpa), feelings (vedanā), perceptions (saṃjñā), conditionings (saṃskāra) and
consciousness (vijñāna) – are all in fact spurious and void.85 Furthermore,
Chapter V contains an allusion to the eighteen sense-elements in a citation
from the Chan master Han (called Lang in Chinese) (§1.1):86
Mind is the essence of the religious path. The body is the vessel for the
religious path. The virtuous friend originates due to conditions. If the
mind is comfortable, objects too are comfortable. If one is without accep-
tance and rejection towards the two limitations, one views the eighteen
sense-elements.

85 C 67–68.
86 C 128. For a translation from the Chinese, see Broughton, 1999: 52.
The idea expressed here is that the faculty of understanding (prajñā) permits
one to pierce through the illusory nature of relative truth (“acceptance and
rejection”) to reveal that what we habitually take to be concrete entities are in
fact but the assemblage of the psycho-physical sense-elements. Given the
Chan orientation of this quotation, we can take the eighteen sense-elements
mentioned in this passage as shorthand for ‘emptiness’ (śūnyatā), the ultimate
nature of phenomena.
In Chapter VI, we find several references to the aggregates, sense-elements
and sense-spheres taken together to signify that they are in fact the mind’s
luminous clarity arising as gods and goddesses. We find such statements in the
sections concerning the tantric views of the inseparability of means and
understanding (§1.3)87 and of non-duality (§1.5),88 as well as in the section on
the tantric approach’s fruition (§4.1).89 Later in Chapter VII, this notion is criti-
cized from the perspective of Atiyoga for still being contrived (§2.3.4.11).90
Let us here take a look at the Abhidharma’s classification of the eighteen
sense-elements and twelve sense-spheres. The eighteen sense-elements (dhātu;
Tib. khams) are divided into three groups of six. (I) The first group consists of
the faculties: (1) the sense-element of the visual faculty (cakṣurindriyadhātu),
(2) the sense-element of the auditory faculty (śrotrendriyadhātu), (3) the
sense-element of the olfactory faculty (ghrāṇendriyadhātu), (4) the sense-
element of the gustatory faculty (jihvendriyadhātu), (5) the sense-element of
the tactile faculty (kāyendriyadhātu) and (6) the sense-element of the psychic
faculty (manendriyadhātu). (II) The second group consists of the objects: (7)
the sense-element of form (rūpadhātu; Tib. gzugs-khams), (8) the sense-ele-
ment of sound (śabdadhātu), (9) the sense-element of odor (gandhadhātu),
(10) the sense-element of taste (rasadhātu), (11) the sense-element of tangi-
bles (spraṣṭavyadhātu) and (12) the sense-element of [psychic] phenomena
(dharmadhātu; Tib. chos-khams). (III) The third group consists of the conscious-
nesses: (13) the sense-element of visual consciousness (cakṣurvijñānadhātu),
(14) the sense-element of auditory consciousness (śrotravijñānadhātu), (15)
the sense-element of olfactory consciousness (ghrāṇavijñānadhātu), (16)
the sense-element of gustatory consciousness (jihvāvijñānadhātu), (17) the
sense-element of tactile consciousness (kāyavijñānadhātu) and (18) the
sense-element of psychic consciousness (manovijñānadhātu). According to
Vasubandhu’s Abhidharmakośa, groups I (i.e. sense-elements 1–6) and III (i.e.
87 C 199.
88 C 205.
89 C 268.
90 C 435.
330 Esler
sense-elements 13–18) are internal, whereas group II (i.e. sense-elements 7–12)

consists of the external sense-elements; ‘internal’ and ‘external’ are designated
in relation to consciousness.91 To exemplify the working of these sense-ele-
ments in connection with the condition qua cognitive object mentioned
above, we may take the example of the visual consciousness. Just as fire can
only burn where there is fuel, the visual consciousness can only arise when
there is the coming together of the visual faculty and of form. If its imaged
object of cognition (ālambana) is absent, visual consciousness does not arise
in that particular moment.92
Returning to our main discussion, groups I and II taken in isolation from
group III are referred to as the twelve sense-spheres (āyatana; Tib. skye-
mched), in which case the word ‘sense-element’ (dhātu; Tib. khams) is
replaced by the term ‘sense-sphere’ (āyatana; Tib. skye-mched). The Sanskrit
term āyatana literally means ‘entrance’ (āyam tanoti), since it is an entrance
for the mind and the mental events (cittacaitānām). Consciousness cannot
arise alone, but must be supported by a sense-faculty and a corresponding
sense-object. Hence, these are given the name sense-spheres. Thus we have:
(I) (1) the sense-sphere of the visual faculty (cakṣurindriyāyatana), (2) the
sense-sphere of the auditory faculty (śrotrendriyāyatana), (3) the sense-
sphere of the olfactory faculty (ghrāṇendriyāyatana), (4) the sense-sphere of
the gustatory faculty (jihvendriyāyatana), (5) the sense-sphere of the tactile
faculty (kāyendriyāyatana) and (6) the sense-sphere of the psychic faculty
(manendriyāyatana); (II) (7) the sense-sphere of form (rūpāyatana; Tib. gzugs
skye-mched), (8) the sense-sphere of sound (śabdāyatana), (9) the sense-
sphere of odor (gandhāyatana), (10) the sense-sphere of taste (rasāyatana),
(11) the sense-sphere of tangibles (spraṣṭavyāyatana) and (12) the sense-
sphere of [psychic] phenomena (dharmāyatana; Tib. chos skye-mched). The
sense-spheres of group I (i.e. sense-spheres 1–6) are referred to as the ‘internal
sense-spheres’ (adhyātmāyatana), whereas those of group II (i.e. sense-spheres
7–12) are called the ‘external sense-spheres’ (bāhyāyatana).93

92 Lamotte, vol.4, 1976: 2108.
93 Cf. Stcherbatsky, 1970: 7–10, 58.
4 Meditational Categories
4.1 Defects of Calm Abiding Meditation and Their Antidotes

Turning to themes more specifically linked to meditation, we now come to the
defects which occur during the practice of calm abiding (śamatha; Tib. zhi-
gnas) and their antidotes. This discussion occurs as part of the exposition of
the gradualist approach to contemplation (Chapter IV, §2.3.1.1).94 There are five
defects, viz. laziness (kausīdya; Tib. le-lo), forgetfulness of imaging (ālambana
saṃpramoṣa; Tib. dmigs-pa brjed-pa), sinking (laya; Tib. bying) and agitation
(auddhatya; Tib. rgod), non-effort (anābhoga; Tib. mi rtsol-ba), and effort
(ābhogatā, Tib. rtsol-ba). As pointed out by Miyazaki Izumi,95 by listing sink-
ing and agitation together, gNubs-chen enumerates only five defects, rather
than the six of the third Bhāvanākrama.96 In this respect, the five defects of
Sangs-rgyas ye-shes’ exposition are in accord with Vasubandhu’s Madhyānta
vibhāgabhāṣya (ch.4, v.4).97
The antidotes against laziness are faith (śraddhā; Tib. dad-pa), inclination
(chanda; Tib. ’dun-pa), effort (vyāyāma; Tib. rtsol-ba) and veritable refinement
(praśrabdhi; Tib. shin-tu sbyangs-pa); the antidote to forgetfulness of imaging
is recollection (smṛti; Tib. dran-pa), whereas attentiveness (saṃprajanya; Tib.
shes-bzhin) is the antidote to sinking and agitation. There is a clear relation
between recollection (smṛti; Tib. dran-pa), often translated as ‘mindfulness,’
and attentiveness (saṃprajanya; Tib. shes-bzhin), which is why they are some-
times counted as one. The difference between them is that attentiveness has a
more constant character than recollection, in that it informs the mind when
recollection is lost and the mind wanders astray in meditation.98 Furthermore,
it should be noted that non-effort – or, in the Madhyāntavibhāgabhāṣya,
absence of conditionings (anabhisaṃskāra)99 – is a defect when putting to rest
sinking and agitation. However, once they have been put to rest, it is effort that
is a defect. The same point is made in the bSam-gtan mig-sgron, which explains
intention as an antidote to non-effort and equanimity as an antidote to effort.
94 C 77.2–6.
95 Miyazaki, 2004: 899.
96 Kamalaśīla (Gyaltsen Namdrol, 1997: 156–158 (Tibetan text), 259 (Sanskrit text)).
97 Anacker, 2008: 247 (English translation), 447 (Sanskrit text).
98 Guenther, 1989: 76–77.
99 Anacker, 2008: 247.
332 Esler
4.2 The Four Contemplations

When perfecting the practice of calm abiding, the adept passes through the
four contemplations (dhyāna). The passage describing them in the bSam-gtan
mig-sgron begins as follows (Chapter IV, §2.3.1.2):100
When one attains the qualities of having completed calm abiding, [medi-
tation] is given the name ‘contemplation’. By attaining the other special
qualities of imaging, [meditation] is given the names ‘contemplation,’
‘[formless absorption of the] form[less] realm,’ and ‘release in the form-
less realm’.
It is apparent from this last sentence that, as mentioned above, each of the four
contemplations can be either an absorption (samāpatti) or a state of existence
(upapatti). According to Vasubandhu, the former is contemplation-as-cause
(kāraṇadhyāna), the latter contemplation-as-effect (kāryadhyāna).101 As is
explicitly stated in the passage below, it is only when one is free from the thirst
of the desire realm (kāmadhātu) that one can enter these four contemplations,
which pertain to the form realm (rūpadhātu). gNubs-chen Sangs-rgyas ye-shes’
description proceeds as follows:102
Thus, only when one is endowed with equanimity of feelings, as well

as examination and scrutiny, is there stability without inadequacy
(anāgamya), as well as the intention of application. Also, only when one
is free from the thirst of the desire [realm] and is endowed with joy and
bliss, does one speak of the first contemplation. With respect to the first
100 C 77.6–78.2: zhi gnas rdzogs pa’i yon tan thob pas bsam gtan gyi ming thob pas ni/ de la
dmigs pa’i yon tan khyad par can gzhan thob pas/ bsam gtan dang/ gzugs [med pa’i snyoms
par ’jug pa] dang/ gzugs med pa’i rnam par thar pa’i ming thob bo/. The words in square
brackets have been added according to the corresponding passage in Kamalaśīla (Gyalt-
sen Namdrol, 1997: 45 (Tibetan text), 214 (Sanskrit text)).
102 C 78.2–6: de ltar gang gi tshe tshor ba btang snyoms dang ldan zhing rtog dpyod bcas par
ldan tsam na/ mi lcogs pa med pa’i brtan pa dang sbyor ba’i sems so/ /yang ’dod pa’i sred pa
dang bral [bral em : dbral C]/ dga’ ba bde [bde em. : de C] dang ldan [ldan em. : tsan C] par
gyur tsam na/ bsam gtan dang po zhes bya’o/ /bsam gtan dang po de la rtog pa tsam med
pa ni/ bsam gtan khyad par can no/ /yang rtog dpyod dang bral bas bsam gtan dang po sred
[sred em. : brjod C] pa dang bral ba de/ dga’ ba bde nang [nang em. : na C] yongs su dang
ba dang ldan pa de bsam gtan gnyis pa’o/ /de’i sred pa dang bral ba btang snyoms shes bzhin
dang ldan na bsam gtan gsum pa’o/ /yang de’i sred pa dang bral te sdug bsngal yang ma
yin/ bde ba yang ma yin pa’i btang snyoms dang dran pa de bsam gtan bzhi pa’o/.
contemplation, absence of mere examination is the distinctive contem-

plation (dhyānāntara).
Furthermore, being free from examination and scrutiny, when one is
free from thirsting for the first contemplation and is endowed with joy,
bliss and thorough limpidity, that is the second contemplation.
When one is free from thirsting for [the second contemplation] and
is endowed with equanimity and attentiveness, this is the third con-
templation.
When one is free from thirsting for [the third contemplation] and is
possessed of equanimity which is neither suffering nor pleasure, and rec-
ollection, this is the fourth contemplation.
These two passages follow quite closely the wording found in Kamalaśīla’s first
Bhāvanākrama, so much so that they appear to be a summary or paraphrase
thereof. Here is a translation of the section in the Bhāvanākrama:103
103 Kamalaśīla (Gyaltsen Namdrol, 1997: 44–46 (Tibetan text), 214–215 (Sanskrit text)); Van
den Broeck, 1977: 27–28 (French translation). Here is the Tibetan with the parallel words
in Roman type: /sems rtse gcig pa de phyi phyir las su rung ba nyid dang ldan zhing dmigs
pa la sogs pa’i yon tan gyi khyad par thob pas bsam gtan dang/ gzugs med pa’i snyoms par
’jug pa dang/ rnam par thar pa la sogs pa’i ming thob bo/ /’di ltar gang gi tshe btang snyoms
kyi tshor ba dang ldan zhing rtog pa dang bcas pa dang/ dpyod pa dang bcas par gyur pa
de’i tshe mi lcogs pa med pa zhes bya ste/ bsam gtan dang po’i sbyor ba’i sems so/ /gang gi
tshe ’dod pa’i sred pa dang/ sdig pa’i chos rnams dang bral zhing rtog pa dang/ dpyod pa
dang/ dga’ ba dang/ bde ba dang ldan par gyur pa de’i tshe bsam gtan dang po zhes bya’o/
/bsam gtan dang po de nyid las rtog pa tsam med pa de ni bsam gtan khyad par can zhes
bya’o/ /gang gi tshe rtog pa dang dpyod pa dang bral te bsam gtan dang po’i sa’i sred pa
dang bral bar gyur nas dga’ ba dang bde ba dang/ nang yongs su dang ba dag dang ldan
par gyur pa de’i tshe bsam gtan gnyis pa zhes bya’o/ /gang gi tshe bsam gtan gnyis pa’i
sa’i sred pa dang bral bar gyur te bde ba dang/ btang snyoms dang/ dran pa dang/ shes
bzhin dag dang ldan par gyur pa de’i tshe bsam gtan gsum pa zhes bya’o/ /gang gi tshe
bsam gtan gsum pa’i sa’i sred pa dang bral te sdug bsngal yang ma yin bde ba yang ma yin
zhing/ btang snyoms dang/ dran pa dag dang ldan par gyur pa de’i tshe bsam gtan bzhi pa
zhes bya’o/. The Sanskrit reads: eṣā ca cittaikāgratā uttarottarakarmaṇyatāsampra-yogād
ālambanādiguṇaviśeṣayogācca dhyānārūpisamāpattiḥ vimokṣādivyapadeśaṃ labhate/
tathā hi yadopekṣāvedanāsamprayuktā savitarkasavicārā sā bhavati, tadā’nāgamyā
ucyate [prathamadhyāna-prayogacittatvāt]/ yadā ca kāmatṛṣṇayā [pāpadharmaiśca]
viviktā bhavati, [vitarkavicāra] prītisukhādhyātma-samprasādaiḥ samprayuktā
bhavati, tadā prathamaṃ dhyānam ucyate/ ata eva prathamadhyānaṃ vitarkamātra-
rahitaṃ dhyānāntaramucyate/ yadā vitarkavicārarahitā prathamadhyānabhūmitṛṣṇayā
viviktā ca bhavati, prītisukhādhyātmasamprasādaiḥ samprayuktā bhavati, tadā dvitīyaṃ
dhyānamucyate/ yadā tu dvitīyadhyānabhūmitṛṣṇayā viviktā bhavati, sukhopekṣā-
smṛtisamprajanyasamprayuktā bhavati, tadā tṛtīyaṃ dhyānam ucyate/ yadā
334 Esler
When one-pointedness of mind (cittaikāgratā) is endowed with further

competence (karmaṇyatā), and when it has attained the special qualities
of imaging, etc., it is given the names ‘contemplation’, ‘formless absorp-
tion of formlessness’, ‘release’, and so forth.
Thus, when it is endowed with the feeling of equanimity, as well as
examination and scrutiny, it is called ‘absence of inadequacy,’ and the
intention of application of the first contemplation [is present]. When it is
free from the thirst of the desire [realm] and from sinful phenomena, and
is endowed with examination, scrutiny, joy and bliss, it is called the ‘first
contemplation’. Within the first contemplation, absence of mere exami-
nation is called the ‘distinctive contemplation’.
Being free from examination and scrutiny, when it is free from thirst-
ing for the stage of the first contemplation and is endowed with joy, bliss
and thorough limpidity, it is called the ‘second contemplation’.
When it is free from thirsting for the stage of the second contempla-
tion and is endowed with bliss, equanimity, recollection and attentiveness,
it is called the ‘third contemplation’.
When it is free from thirsting for the stage of the third contemplation,
when there is neither suffering nor pleasure, and when it is endowed
with equanimity and recollection, it is called the ‘fourth contemplation’.
Let us now look at the Abhidharmakośa’s presentation of the four contempla-

tions. The first contemplation has five ancillaries: examination (vitarka; Tib.
rtog-pa), scrutiny (vicāra; Tib. dpyod-pa), joy (prīti; Tib. dga-ba) [i.e. rapture
(saumanasya)], bliss (sukha; Tib. bde-ba) [i.e. veritable refinement (praśrabdhi)]
and concentration (samādhi; Tib. ting-nge-’dzin). While in their fully devel-
oped state these five factors constitute the first contemplation, it should be
remembered that they are present, albeit to a weaker degree, in normal con-
sciousness.104 This in itself already indicates that contemplation is not an
altered state of trance-like unconsciousness.
The second contemplation has four ancillaries: thorough limpidity (sampra
sāda; Tib. yongs-su dang-ba) [sometimes listed as sureness (prasāda; Tib.
rab-tu dad-pa), i.e. the faculty of faith (śraddhendriya)], joy, etc. [i.e. bliss and
concentration].
The third contemplation has five ancillaries: equanimity (upekṣā; Tib. btang-
snyoms) [i.e. not the equanimity of feelings (vedanopekṣā), since here joy, or
tṛtīyadhyānabhūmitṛṣṇayā viviktā bhavati, aduḥkhāsukhopekṣāsmṛtyabhisamprayuktā

bhavati, tadā caturthaṃ dhyānamucyate/.
104 Cousins, 1973: 122.
rapture, is still present, but the equanimity of conditionings (saṃskāropekṣā)],

recollection (smṛti; Tib. dran-pa), attentiveness (saṃprajanya; Tib. shes-bzhin),
bliss and concentration.
The fourth contemplation has four ancillaries: recollection, equanimity
[this again refers to the equanimity of conditionings], neither pleasure nor suf-
fering (aduḥkhāsukha; Tib. bde-min sdug-min) [i.e. the equanimity of feelings]
and concentration.105
The definition of the four contemplations offered by Asaṅga’s Abhidharmasa
muccaya is identical, with concentration being replaced by the synonymous
one-pointedness of mind (cittaikāgratā).106
Some of the ancillaries alluded to here need to be further elucidated.
Examination (vitarka) and scrutiny (vicāra) are defined as the crude (audāri
katā) and subtle (sūkṣmatā) states of conceptual thought.107 Examination
serves to understand the historical background, whereas scrutiny concentrates
on the present situation;108 both factors are also considered to be the requi-
sites for the faculty of speech.109 To illustrate the difference between the two,
examination may be said to be investigation, whereas scrutiny refers to the
investigation’s resolution. For instance, to investigate a great number of pots in
order to determine which ones are well baked and which ones are unbaked is
examination. Scrutiny, on the other hand, refers to the conclusion that such a
number of pots belong to each category.110 In the Milindapañha the venerable
Nāgasena explains that examination is characterized by fixing the mind (P.
appanā) and scrutiny is defined as pondering (P. anumajjana). He compares
examination to striking a bronze gong, and scrutiny to the reverberating sound
that lingers on.111 These two factors of examination and scrutiny inherently
disturb the mind’s concentration,112 which is why they are abandoned in the
second contemplation.
The importance of joy (prīti) in the accounts of Buddhist contemplative
praxis have often been overlooked in Western writings on the subject, with the
result that Buddhism (especially in its early form) is frequently presented as a
dry rationalistic religion. Lance S. Cousins has drawn attention to the
105 de La Vallée Poussin, 1971, vol.5, ch.8: 147–149, 147, note 1; the information given in [square
brackets] is supplied by de La Vallée Poussin in his note to this passage.
106 Rahula, 1971: 111.
109 Cousins, 1973: 122.
110 de La Vallée Poussin, 1971, vol.1, ch.2: 175, note 2.
111 Horner, 1969: 86.
112 Lamotte, vol.3, 1970: 1488.
336 Esler
Visuddhimagga’s careful distinction between five types of joy: (1) minor (P.
khuddikā) joy makes the body’s hairs stand on end; (2) momentary (P. khaṇikā)
joy is comparable to sudden flashes of lightning that occur from time to time;
(3) descending (P. okkantikā) joy descends upon the body like waves hitting the
seashore; (4) buoyant joy (P. ubbegā) is sufficiently powerful to cause the body
to jump in the sky; and (5) pervading (P. pharaṇā) joy pervades the entire body
just as a torrent of water would inundate a grotto.113 With regard to transport-
ing joy, Lance S. Cousins notes that this form of levitation is particularly linked
to strong forms of religious devotion.114 The emotional feeling-tone of the term
is also evoked by the fact that in common non-technical Indian usage, prīti
signifies “the joyful and blissful aspects of family and marital love”.115
As far as bliss (sukha) is concerned, it is differentiated from the relative
excitement of joy in that it is a calm and steady state of well-being. Furthermore,
the bliss pertaining to the first two contemplations must be distinguished from
the bliss of the third contemplation, in that in the former case it is the bliss of
veritable refinement (praśrabdhisukha), whereas in the latter case it is a feel-
ing of bliss (sukhā vedanā). This feeling of bliss is, in the case of the third
contemplation, purely mental in that it does not rely on any of the other sense-
faculties. Such a purely mental bliss is beyond the desire realm.116
As for recollection (smṛti), in the context of the contemplations it is defined
as the ability not to forget the sign (nimitta, here meaning motive or reason) of
equanimity.117
As has already been pointed out, equanimity is of two kinds: on the one
hand, equanimity of conditionings (saṃskāropekṣā) is characteristic of the
third contemplation, where joy is still present. On the other hand, equanimity
of feelings (vedanopekṣā), which is neither pleasure nor suffering, is peculiar to
the fourth contemplation.118
To summarize in a less technical language, what we have then is a pro-
cess of gradual simplification as the adept progresses along the various
contemplations: first, examination and scrutiny are abandoned in the second
contemplation; then, joy is abandoned in the third contemplation; and finally,
113 Buddhaghosa (Maës, 2002, ch.4, §§94–99: 172–174).

114 Cousins, 1973: 120.
115 Cousins, 1973: 121.
116 de La Vallée Poussin, 1971, vol.5, ch.8: 150, incl. Note 2.
118 de La Vallée Poussin, 1971, vol.5, ch.8: 147–149; Cornu, 2006: 184.
bliss itself is abandoned in the fourth contemplation, where pure equanimity

remains.119
It is evident that on the whole the descriptions of the four contem-
plations found in the Abhidharmakośa and in the bSam-gtan mig-sgron
overlap. Furthermore, we may note the bSam-gtan mig-sgron’s allusion to the
precursory level (sāmantaka) of the first contemplation, when our author
mentions “stability without inadequacy” as being endowed with equanimity
of feelings, examination and scrutiny.120 The precursory level of the first con-
templation is indeed classically defined as anāgamya (‘not arrived’) in Sanskrit
and mi-lcogs med-pa (‘without inadequacy’) in Tibetan. Each of these defini-
tions brings out a different aspect of the meditative level in question. The term
anāgamya is used because one has not yet arrived (√an-ā-gam) at the main
contemplation (mauladhyāna); nonetheless, even though the precursory level
is devoid of the qualities of the first contemplation, it is without inadequacy
(mi-lcogs med-pa), since it is capable of cutting through the contaminations.121
The precursory level, like the first contemplation per se, is said to possess both
examination and scrutiny (savitarkaḥ savicāraḥ)122 and is also characterized
as being endowed with the faculty or feeling of equanimity (upekṣendriya).123
Reference is also made to the so-called distinctive contemplation: “With
respect to the first contemplation, absence of mere examination is the distinc-
tive contemplation”.124 This distinctive (or intermediate) contemplation
(dhyānāntara) is indeed defined as being “without examination but with scru-
tiny alone” (avitarko vicāramātraḥ).125 It is therefore superior to the first
contemplation but inferior to the second one; hence, it is called ‘distinctive’ (or
‘intermediate’) contemplation (dhyānāntara),126 which implies that it is ‘dis-
tinct from’ (and superior to) the first contemplation due to the absence of
examination (vitarka). In other words, the first contemplation acquires a dis-
tinction (i.e. becomes dhyānāntara) through the absence of examination.127
119 Cousins, 1973: 124.

120 C 78.2–3: […] tshor ba btang snyoms dang ldan zhing rtog dpyod bcas par ldan tsam na/ mi
lcogs pa med pa’i brtan pa […].
121 Cf. de La Vallée Poussin, 1971, vol.5, ch.8: 179, note 6.
122 Lamotte, vol.3, 1970: 1487.
123 Lamotte, vol.2, 1949: 1042.
124 C 78.3–4: /bsam gtan dang po de la rtog pa tsam med pa ni/ bsam gtan khyad par can no/.
125 Lamotte, vol.3, 1970: 1488.
127 de La Vallée Poussin, 1971, vol.5, ch.8: 181, note 1.
338 Esler
Hence, the distinctive contemplation is to be understood as a distinction

within the first contemplation.128
All the concentrations below the distinctive contemplation, i.e. all those
pertaining to the first contemplation, have both examination and scrutiny,
whereas above the distinctive contemplation, all the concentrations are with-
out examination and scrutiny.129
The fruition of the distinctive contemplation is rebirth as Mahābrahmā,130
so that, cosmologically speaking, it is the residence of the Mahābrahmās and
goes up to the Brahmapurohitas (the second of three abodes pertaining to the
first contemplation).131
Despite the bSam-gtan mig-sgron’s overall agreement with the Abhidharma
kośa, elsewhere in the text (Chapter VI, §2.2.4.1.4.2.1) the second contemplation
is said to be without examination but to have scrutiny and joy.132 This of
course contradicts the classical presentation, which invariably characterizes
the second contemplation by the removal of scrutiny – and, by implication,
examination, since the former is always accompanied by the latter.133 What is
most intriguing about this inconsistency is that, as we have seen in the passage
quoted above, the classical presentation is accepted by gNubs-chen Sangs-rgyas
ye-shes.134 The only solution would be to surmise that in the above-mentioned
incongruous passage from Chapter VI, Sangs-rgyas ye-shes in fact has the dis-
tinctive contemplation in mind, which, as already discussed, is endowed with
scrutiny and is without examination.135 This explanation, however, can hardly
be admitted, since in the passage in question the concentration is explicitly
referred to as the second contemplation and occurs, moreover, between the
first and the third contemplations. If the distinctive contemplation were
intended, it would surely be mentioned by name (as it is in the main passage
from Chapter IV quoted here) and would occur between the first and second
contemplations. We are therefore obliged to leave this incongruity as it stands.
There are specific feelings pertaining to each of the states of existence
associated with the four contemplations: the first contemplation is charac-
terized by the feeling of bliss associated with the visual, auditory and tactile
128 Lamotte, vol.3, 1970: 1185, note 1.

132 C 233.5–6: gnyis pa’i rtog pa med pa la dpyod pa dang dga’ ba yod pa/.
133 de La Vallée Poussin, 1971, vol.5, ch.8: 133. Cf. also the formulation according to the
Theravāda Abhidhamma tradition in Cousins, 1973: 124.
134 C 78.4–5.
consciousnesses; by the feeling of rapture associated with the psychic con-

sciousness; and by the feeling of equanimity associated with the visual,
auditory, tactile and psychic consciousnesses.
The second contemplation is characterized by rapture and equanimity,
both associated with the psyche.
In the third contemplation, there is bliss and equanimity, both pertaining
to the psyche. In the fourth contemplation, there is only one feeling, namely
equanimity.136
4.3 The Four Formless Absorptions

The bSam-gtan mig-sgron’s rather laconic statement: “Likewise, this should
also be applied to the formless realm,”137 hints at the fact that the four form-
less absorptions (samāpatti) are but an extension of the four contemplations
to the formless realm (ārūpyadhātu). Indeed, they differ from the four con-
templations in that here the aggregate of form (rūpaskandha) is absent.138 The
various formless absorptions of formlessness are at once cosmological realms
and altered states of consciousness. Being the product of meditative states,
these realms are held to be inhabited by beings without physical form, who
live there for one or several lives.139 While Sangs-rgyas ye-shes does not elab-
orate further, we will take our cues from the Abhidharmakośa’s elucidation
of the four formless absorptions. They are named after the four increasingly
subtle perceptions (samjñā; Tib. ’du-shes) which are their objective focus or,
as here, after the four spheres which apprehend these perceptions: the sphere
of infinite space (ākāśānantyāyatana; Tib. nam-mkha’ mtha’-yas skye-mched),
the sphere of infinite consciousness (vijñānānantyāyatana; Tib. rnam-shes
mtha’-yas skye-mched), the sphere of nothing whatsoever (ākiṃcanyāyatana;
Tib. ci-yang med-pa’i skye-mched) and the sphere of neither perception nor
non-perception (naivasaṃjñānāsaṃjñāyatana; Tib. ’du-shes med ’du-shes med-
min skye-mched). Each of the four formless absorptions proceeds through a
separation from the stage below. For instance, the sphere of infinite space
(ākāśānantyāyatana) proceeds through the separation (viveka) from the
fourth contemplation. The same applies to each of the four formless absorp-
tions, until the sphere of neither perception nor non-perception.140

137 C 78.6: /de bzhin du gzugs med pa yang sbyar ro/.
139 Griffiths, 1999: 17.
340 Esler
When practicing the four formless absorptions (samāpatti), the adept is

supposed to actively think of, and indeed to verbalize, the perception for
instance that space is infinite. He eventually comes to an actual experience of
the infinity of space, whereby it becomes impossible for him to see physical
form.141 This would indicate that even in the enstatic states of contemplation,
the discursive functions of examination and scrutiny still have a role to play, so
that the states of calm abiding (śamatha) become the retrospective objects of
insight (vipaśyanā).
In view of the fact that the four formless absorptions are without form, it
might be asked how a being who for long periods of time remains in these
absorptions can, when he is reborn in a lower stage, recover a form. The
problem arises because the state of cessation of perceptions and feelings
(samjñāvedayitanirodha), to which the four formless absorptions lead, is free
from all mental events, and the emergence from cessation must have an imme-
diately antecedent and similar condition. According to the Sautrāntikas, who
believe that an event cannot be separated in time from its immediate
condition,142 the answer is that his form is produced from the mind. This
means that in the past he has committed an action which must mature as
form, which action has left an imprint on his mind.143 On the other hand, the
answer proposed by the Vaibhāṣikas is twofold: (1) the immediate condition of
an event can be separated in time from that event; and (2) the immediate con-
dition for the emergence from cessation is the last moment of consciousness
before entering the state of cessation.144
4.4 The Concentration of Cessation

The concentration of cessation (nirodhasamādhi) is mentioned at the end of
Chapter IV, in the context of the gradual approach’s fruition. By perfecting the
two accumulations (of merit and wisdom) and purifying his obscurations, the
bodhisattva realizes the principial body (dharmakāya), which corresponds to
his own purpose (svārtha) and to the state of cessation. Emerging from that
concentration is exemplified as waking up from sleep. While the text does not
explicitly say so, the implication is that this is where the Hīnayāna path ends,
whereas the bodhisattva intent on the purpose of others (parārtha) emerges
from that state as the twofold body of form (rūpakāya) through the power of
141 Griffiths, 1999: 18.

142 Griffiths, 1999: 64.
144 Griffiths, 1999: 63.
his compassion and his aspirations.145 It must be understood that ‘cessation’

here refers to cessation due to deliberation (pratisaṃkhyānirodha), which is
the definitive eradication through deliberation (pratisaṃkhyā) of the latencies
(anuśaya) producing new existences. This state effectively corresponds to
transcendence (nirvāṇa) and must be distinguished from cessation not due to
deliberation (apratisaṃkhyānirodha), where an existence ceases due to the
mere absence of particular causes.146
The cessation of perceptions and feelings (samjñāvedayitanirodha), which
occurs after having transcended the sphere of neither perception nor non-per-
ception, is free from all mental events, including of course examination and
scrutiny denoted by the terms ‘understanding’ and ‘deliberation’. Hence, (in
the case of cessation due to deliberation,) understanding is produced after the
attainment of cessation. It might then be asked why the state of cessation is at
all necessary as a prelude to the production of understanding and the atten-
dant destruction of the contaminations. The answer appears to lie in the
bringing together of two distinct soteriological goals: on the one hand, the
state of cessation (which by definition precludes all forms of intellectual activ-
ity) as the antidote to the contaminations; and on the other hand, understanding
as the antidote to ignorance.147
4.5 The Cognitive Aspect of Realization

The cognitive aspect of this realization is accorded great importance in
Buddhist thought. A brief reference to this is found in a citation from the rDo-
rje bkod-pa in Chapter VI, in the context of the tantric view of equality (§1.6.1):148
The path of equality does not image a center or limit;

Having only assimilated equality,
One is without objects and is free from the searching mind.
Having relinquished the limitations of acceptance and rejection,
One abides in a non-abiding manner.
145 C 118.1–2.
147 Griffiths, 1999: 22–23.
148 C 211, quoting rDo-rje bkod-pa, in NGM, vol.15/ba, 358.3–5: dbus mtha’ mi dmigs mnyam pa’i
lam/ /gcig tu [tu C: pu NGM] mnyam par chud nas ni/ /yul med tshol ba’i sems bral te/ /
blang dor gnyis kyi mtha’ [mtha’ NGM: mnya’ C] spangs nas/ /gnas pa med pa’i tshul gyi [gyi
C: gyis NGM] gnas/ /chos nyid dbyings su ro gcig phyir/ /shes dang shes bya gnyis su med
[gnyis su med C: gnyis med do NGM]/ /ye nas shes shing [shing C: nyid NGM] chos shes na/
/chos nyid gzhan nas bcos mi dgos [bcos mi dgos C: btsal ci dgos NGM]/.
342 Esler
Since this is the single taste of the open dimension of phenomena’s

beingness,
Knower and knowable are non-dual;
When knowing primordially and cognizing a principle,
Phenomena’s beingness (Skt. dharmatā) need not be contrived from
elsewhere.
‘Cognizing a principle’ (chos-shes) carries important connotations. In the sūtra

system, each of the four noble truths (suffering, its origin, its cessation and the
path leading to its cessation) is realized in a sequence of four moments. These
are acknowledging the cognition of the principle [of the truth in question]
(dharmajñānakṣānti; Tib. chos shes-pa’i bzod-pa), cognizing the principle
(dharmajñāna; Tib. chos shes-pa), acknowledging the cognition of the princi-
ple in its subsequent realization (anvayadharmajñānakṣānti; Tib. rjes-su
rtogs-par shes-pa’i bzod-pa) and cognizing it in its subsequent realization
(anvayadharmajñāna; Tib. rjes-su rtogs-par shes-pa). Taking for example the
truth of suffering, the first moment, acknowledging the cognition of the prin-
ciple, is still tainted by doubt and contamination; the yogin considers the fact
that in the desire realm all phenomena are impermanent, suffering, empty and
selfless. Since he thereby removes a certain type of doubt and contamination,
this acknowledging is called path of relinquishing (prahāṇamārga; Tib. spong-
ba’i lam) and path without obstacle (ānantaryamārga; Tib. bar-chad med-pa’i
lam). Through this relinquishing he stops being an ordinary person and
becomes a noble one (ārya). In the second moment, cognizing the principle,
the yogin truly comes to possess this certain knowledge without doubt and
contamination; this cognition is hence called path of utter liberation
(vimuktimārga; Tib. rnam-par grol-ba’i lam). Whereas the first two moments
acknowledge and cognize the truth in the desire realm, the latter two acknowl-
edge and cognize it in the two higher realms, i.e. the form realm and the
formless realm. This entire process is realized through the operation of wis-
dom (jñāna; Tib. ye-shes). Since there are four moments for each of the four
noble truths, this yields sixteen moments, which are collectively termed the
sixteen moments of wisdom on the path of seeing (ṣoḍaśadarśanamārgakṣaṇa;
Tib. mthong-lam-gyi ye-shes skad-cig bcu-drug).149 However, Vasubandhu holds
that only the first fifteen moments – i.e. acknowledging the cognition of the
principle with regard to suffering (duḥke dharmajñānakṣānti) up to acknowl-
edging the cognition of the principle in its subsequent realization with regard
to the path (mārge ’nvayadharmajñānakṣānti) – pertain to the path of seeing
149 Lamotte, 1958: 681–682; Cornu, 2006: 709; and Negi, 2006: 11.
(darśanamārga) per se. The sixteenth moment, cognizing the principle in its
subsequent realization with regard to the path (mārge ’nvayadharmajñāna),
belongs as such to the path of cultivation (bhāvanāmārga), since here there is
no longer anything unseen to be seen, but merely meditation on the truth as it
has been seen.150
4.6 Transcendence
Although the bSam-gtan mig-sgron is written from a tantric, not to say rDzogs-
chen perspective, there are several references to the classical Abhidharma
formulations of the path’s fruition. In Chapter VI, which is devoted to the tant-
ric approach of Mahāyoga, in the passage cited above concerning the time
required to accomplish the fruition in the tantric vehicle, a comparison is
drawn with the followers of the Hīnayāna, who hold that its maturation can be
felt in this life (dṛṣṭadharmavedanīya), in the next (upapadyavedanīya), after
any number of rebirths (aparaparyāyavedanīya), or be of uncertain ripening
(aniyatavipāka) (§4.4).151
It will be noted that according to the Abhidharmakośa, a deed (karma) may
be either of uncertain ripening (aniyatavipāka) – i.e. whose ripening may
or may not be felt – or of certain ripening (niyatavipāka). In the latter case,
there are three possibilities: (1) it may be felt as a phenomenon seen [in this
life] (dṛṣṭadharmavedanīya); (2) felt in the next life (upapadyavedanīya); or
(3) after any number of rebirths (aparaparyāyavedanīya).152 Furthermore,
in the Abhidharmakośa, the transcendence obtained visibly in this life
(dṛṣtadharmanirvāṇaprāpta) is defined as the transcendence with a remain-
der of the aggregates (sopadhiśeṣanirvāṇa; Tib. phung-po dang-bcas-pa’i
mya-ngan-las ’das-pa); this is the transcendence which an arhat obtains in
this life, while still living out the karmas which make up his physical body.
It is contrasted with the transcendence without remainder of the aggregates
(nirupadhiśeṣanirvāṇa; Tib. phung-po lhag-ma med-pa’i mya-ngan-las ’das-pa),
which is obtained after death.153 While the transcendence with a remainder of
the aggregates is not explicitly mentioned in our text, gNubs-chen makes sev-
eral allusions to the transcendence without remainder of the aggregates and
150 Cf. de La Vallée Poussin, 1971, vol.4, ch.6: 192.

151 C 275.4–6.
344 Esler
clearly sees both forms of transcendence as a pair, as is demonstrated by this

sentence (§4.5):154
There are two manners of transcendence that are found in the oral
injunctions. In [the case of] the transcendence without remainder of the
aggregates, it is doubtless that the yogin [will attain it] as soon as he is
free from the knotted net of the body.
He thereby seems to implicitly suggest a correspondence between the tantric

stage of an awareness-holder dominating life,155 which he discusses at length
in the previous passage,156 and the transcendence with a remainder of the
aggregates.157 There are further references to the transcendence without
remainder of the aggregates in Chapter VII, which exposes the rDzogs-chen
approach of self-liberation. In the section containing the view of spontaneous
presence (§1.2.2.1),158 Sangs-rgyas ye-shes describes the awakening without
remainder “as the nature of the great state of the spontaneously present being-
ness of phenomena”.159 It is further explained that it is by virtue of the state of
spontaneous presence that those of middling and outstanding faculties can
gain the awakening without remainder.160
5 Concluding Remarks
While gNubs-chen Sangs-rgyas ye-shes’ bSam-gtan mig-sgron is not an

Abhidharma treatise per se, the perspective of this early Tibetan doxographical
154 C 278.4–5: mya ngan las ’das pa’i tshul gnyis ni bka’ nyid las bzhugs pas/ phung po lhag ma
med pa mya ngan las ’das na’ang/ […].
155 There are four types of awareness-holder counted in Mahāyoga, among which this is the
second. The four are the awareness-holder of maturation (rnam-smin rig-’dzin), the
awareness-holder who dominates life (tshe-dbang rig-’dzin), the awareness-holder of the
great seal (phyag-rgya chen-po’i rig-’dzin) and the awareness-holder of spontaneous pres-
ence (lhun-grub rig-’dzin). See Cornu, 2006: 694. It might be recalled that gNubs-chen
Sangs-rgyas ye-shes himself is credited with having attained the accomplishment of an
awareness-holder dominating life. See Guru bKra-shis, 1990: 246.
156 C 276–277.
157 See Meinert, 2002: 301.
158 On this and the other views of rDzogs-chen, see Esler, 2012b: 88–91.
159 C 320.3: […] lhun gyis grub pa’i chos nyid ngang chen po’i rang bzhin du lhag ma med par
sangs rgyas so/.
160 C 322.
text on contemplation is eminently metaphysical, in the sense described above

of being concerned with knowledge of the ultimate nature of reality and with
the methods of its spiritual realization. Furthermore, as we have attempted to
trace in this chapter, its discourse is embedded in the Abhidharmic framework,
to which it makes both implicit and explicit references. These remind us of the
fact that Buddhist meditative praxis – whether pertaining to the Sūtrayāna,
Mantrayāna or rDzogs-chen – does not take place in a vacuum, but within a
context informed and defined by the Abhidharma’s metaphysical, cosmologi-
cal and phenomenological premises.
It is not easy to determine exactly which Abhidharma text gNubs-chen had
in mind, since he does not cite specific Abhidharma treatises but makes use of
numerous ideas and categories listed in such sources. Whereas both Asaṅga’s
Abhidharmasamuccaya and Vasubandhu’s Abhidharmakośa were translated
into Tibetan during the eighth century, Dan Martin has pointed out that,
according to Nyang-ral nyi-ma ’od-zer’s (1124–1192) Chos-’byung me-tog snying-
po,161 the transmission of the Abhidharmakośa was interrupted until the late
tenth century, when it was renewed by Smṛti.162 This figure would seem to be
identical to Smṛtijñānakīrti, an Indian scholar who was active in Tibet during
the latter part of the tenth century (i.e. after Sangs-rgyas ye-shes’ demise).163
While gNubs-chen does not quote directly from the Abhidharmasamuccaya
or from the Abhidharmakośa, the passage concerning the development
of the four contemplations that we have traced above to Kamalaśīla’s
first Bhāvanākrama contains several elements that seem to point to the
Abhidharmakośa as their prototype. The passage in question mentions both
the precursory level (sāmantaka) of the first contemplation, which it alludes to
through the expression mi-lcogs med-pa (‘absence of inadequacy’), and the dis-
tinctive contemplation (dhyānāntara). Since neither of these are discussed in
the Abhidharmasamuccaya, but are on the contrary explained in some detail
in the Abhidharmakośa, we may infer that by the early tenth century, when
gNubs-chen Sangs-rgyas ye-shes composed the bSam-gtan mig-sgron, the
ideas of the Abhidharmakośa (notably its intricate meditative and cosmologi-
161 Nyang-ral nyi-ma ’od-zer, 1988: 472: chos mngon pa mdzod ni/ sngon ka cog zhang gsum nas
brgyud pa (bar) skabs su stongs pa la/ dus phyis smri tis ’gyur gsar mdzad de nas thog ma’i
bshad pa rnams mchims rgyud kyis dar bar mdzad/. This may be translated as follows: “As
for the Abhidharmakośa, it was previously transmitted through the triad Ka-ba dpal-
brtsegs, Cog-ro klu’i rgyal-mtshan and sNa-nam zhang Ye-shes-sde, yet in the meanwhile
[its transmission] became extinct. Later, it was renewed by Smṛti, after whom the former
explanations were propagated through the lineage of mChims”.
162 Martin, 2002: 337.
163 On some of Smṛtijñānakīrti’s literary accomplishments, see Tanzin, 2013: 25–26.
346 Esler
cal classifications and correspondences) had gained wide enough circulation

and acceptance in Tibet for them to be referred to and understood without
needing to quote the specific texts that expounded them. They had, in other
words, become part of the general Buddhist lore that any learned master
would have been familiar with and could have used at will. Even though the
Abhidharmakośa itself was probably unknown to gNubs-chen, since its line of
transmission had already been interrupted by his time,164 its complex of ideas
and categories infiltrated the bSam-gtan mig-sgron, partly through intermedi-
ary texts like the Bhāvanākrama. The bSam-gtan mig-sgron thus offers some
tantalizing glimpses of the ways these ideas survived the period of fragmenta-
tion (sil-bu’i dus) that followed the Tibetan dynasty’s disintegration in 842 CE,
even while one of the principal texts expounding them, the Abhidharmakośa,
had been apparently forgotten. These glimpses are all the more revealing in
view of the pivotal role gNubs-chen Sangs-rgyas ye-shes played in codifying
and preserving Buddhist teachings during this tumultuous and obscure period.
Abbreviations
C gNubs-chen Sangs-rgyas ye-shes. bSam-gtan mig-sgron, Chhimed Rigdzin

Rinpoche edition.
KD bKa’-’gyur, Facsimile of the sDe-dge 1733 edition prepared by Si-tu paṇ-chen
Chos-kyi ’byung-gnas, 103 Volumes. Delhi: Karmapae Chodhey Gyalwae
Sungrab Partrun Khang, 1976–1979.
NGM rNying-ma rgyud-’bum, mTshams-brag edition, 46 Volumes. Thimphu: Royal
National Library, 1982.
P. Pāli
Tib. Tibetan
164 The extinction of the Abhidharmakośa’s transmission mentioned by Nyang-ral (see

above, note 161) presumably took place either during or shortly after Dar-ma Khri-’u dum-
btsan’s reign (alias Glang-dar-ma, r. 836–842 CE), since this was a particularly turbulent
time. It must be recalled that, contrary to traditional accounts, gNubs-chen Sangs-rgyas
ye-shes was not a contemporary of Glang-dar-ma, but was probably born two years after
the latter’s assassination and witnessed the troubled period that ensued, notably the
revolt of 904 CE. See Esler, 2014: 22.
Bibliography
Amarasiṃha (2008). Amarakoṣa. Varanasi: Chaukhamba.

Anacker, Stefan (2008). Seven Works of Vasubandhu, The Buddhist Psychological Doctor.
Delhi: Motilal Banarsidass.
Borella, Jean (2004). Histoire et Théorie du Symbole. Lausanne: L’Âge d’Homme.
Broughton, Jeffrey L. (1999). The Bodhidharma Anthology: The Earliest Records of Zen.
Berkeley CA: University of California Press.
Buddhaghosa (2002). Visuddhimagga: Le Chemin de la Pureté, Traduit du Magadhi (Pāli)
par Christian Maës. Paris: Fayard.
Conze, Edward (1962). Buddhist Thought in India: Three Phases of Buddhist Philosophy.
London: George Allen & Unwin.
Conze, Edward (1980). “Contradictions in Buddhist Thought”, In Indianisme et
Bouddhisme: Mélanges offerts à Mgr Étienne Lamotte. Louvain-la-Neuve: Institut
Orientaliste, pp. 41–52.
Conze, Edward (1973). Materials for a Dictionary of the Prajñāpāramitā Literature. Tokyo:
Suzuki Research Foundation.
Summary. Delhi: Sri Satguru Publications.
Cornu, Philippe (2006). Dictionnaire Encyclopédique du Bouddhisme. Paris: Éditions du
Seuil.
Cousins, Lance S. (1973). “Buddhist Jhāna: Its Nature and Attainment according to the
Pali Sources”, Religion 3/2, pp. 115–131.
rDo-rje bkod-pa = Kun-’dus rig-pa’i mdo = De-bzhin gshegs-pa thams-cad-kyi thugs gsang-
ba’i ye-shes don-gyi snying-po/ khro-bo rdo-rje’i rigs/ kun-’dus rig-pa’i mdo rnal-’byor
bsgrub-pa’i rgyud. In NGM, vol.15/ba, 321–672.
Esler, Dylan (2012a). “Note d’oniromancie tibétaine: réflexions sur le Chapitre 4 du bSam-
gtan mig-sgron de gNubs-chen sangs-rgyas ye-shes”, Acta Orientalia Belgica 25,
pp. 317–328.
Esler, Dylan (2012b). “The Exposition of Atiyoga in gNubs-chen Sangs-rgyas ye-shes’
bSam-gtan mig-sgron”, Revue d’Études Tibétaines 24, pp. 81–136.
Esler, Dylan (2014). “On the Life of gNubs-chen Sangs-rgyas ye-shes”, Revue d’Études
Tibétaines 29, pp. 5–27.
Evola, Julius (1982). Le Chemin du Cinabre. Milan: Archè.
Griffiths, Paul J. (1999). On Being Mindless: Buddhist Meditation and the Mind-Body
Problem. Delhi: Sri Satguru Publications.
Guénon, René (1984). Les États Multiples de l’Être. Paris: Guy Trédaniel.
Guénon, René (2005). Aperçus sur l’Initiation. Saligny: Éditions Traditionnelles.
Guenther, Herbert V. (1989). From Reductionism to Creativity: rDzogs chen and the New
Sciences of Mind. Boston MA: Shambhala.
348 Esler
Guru bKra-shis (1990). Gu-bKra’i chos-’byung: bsTan-pa’i snying-po gsang-chen snga-’gyur

nges-don zab-mo’i chos-kyi byung-ba gsal-bar byed-pa’i legs-bshad mkhas-pa dga’-byed
ngo-mtshar gtam-gyi rol-mtsho. Beijing: Krung-go’i bod-kyi shes-rig
dpe-skrun-khang.
Heidegger, Martin (2000). Introduction to Metaphysics. New Translation by Gregory Fried
and Richard Polt. New Haven CN: Yale Nota Bene.
Heidegger, Martin (2003). The End of Philosophy. Translated by Joan Stambaugh. Chicago
IL: University of Chicago Press.
Heidegger, Martin (2006). Being and Time. Translated by John Macquarrie and Edward
Robinson. Malden: Blackwell Publishing.
Hirakawa, Akira (1980). “The Meaning of ‘Dharma’ and ‘Abhidharma’”, In Indianisme et
Bouddhisme: Mélanges offerts à Mgr Étienne Lamotte. Louvain-la-Neuve: Institut
Orientaliste, pp. 159–175.
Horner, Isaline Blew (1969). Milinda’s Questions. 2 vols. London: Luzac & Co.
Kalupahana, David J. (1994). A History of Buddhist Philosophy: Continuities and
Discontinuities. Delhi: Motilal Banarsidass.
Kamalaśīla (1997). Bhāvanākrama: Tibetan Version, Sanskrit Restoration and Hindi
Translation. Restored, translated and edited by Gyaltsen Namdrol. Sarnath: CIHTS.
Koros, Alexander Csoma de (1980–1982). Sanskrit-Tibetan-English Vocabulary: Being an
Edition and Translation of the Mahāvyutpatti. 2 vols. Delhi: Sri Satguru Publications.
Lamotte, Étienne (1935). Saṃdhinirmocanasūtra: L’Explication des Mystères. Louvain:
Bureaux du Recueil / Paris: Adrien Maisonneuve.
Lamotte, Étienne (1947). “La légende du Buddha”, Revue de l’Histoire des Religions 134,
pp. 37–71.
Lamotte, Étienne (1958). Histoire du Bouddhisme Indien: Des Origines à l’Ère Śaka.
Louvain: Institut Orientaliste.
Lamotte, Étienne (1944–1980). Le Traité de la Grande Vertu de Sagesse de Nāgārjuna
(Mahāprajñāpāramitāśāstra). 5 vols. Louvain & Louvain-la-Neuve: Institut
Orientaliste.
La Vallée Poussin, Louis de (1927). La Morale Bouddhique. Paris: Nouvelle Librairie
Nationale.
La Vallée Poussin, Louis de (1971). L’Abhidharmakośa de Vasubandhu, Traduction et
Annotations. 6 vols. Bruxelles: Institut Belge des Hautes Études Chinoises.
Levin, David Michael (1985). The Body’s Recollection of Being: Phenomenological
Psychology and the Deconstruction of Nihilism. London: Routledge.
Martin, Dan (2002). “Gray Traces: Tracing the Tibetan Teaching of the Mngon pa kun
btus (Abhidharmasamuccaya) through the Early Period of Disunity”, In The Many
Canons of Tibetan Buddhism. Edited by Helmut Eimer and David Germano. PIATS
2000. Leiden: Brill, pp. 335–357.
Meinert, Carmen (2002). “Chinese Chan and Tibetan rDzogs chen: Preliminary Remarks
on Two Tibetan Dunhuang Manuscripts”, In Religion and Secular Culture in Tibet:
Tibetan Studies II. Edited by Henk Blezer. PIATS 2000, vol.2. Leiden: Brill,
pp. 289–307.
rMi-lam bstan-pa (Svapnanirdeśa), In dKon-mchog brtsegs-pa’i mdo (Ratnakūṭa-sūtra).
In KD, vol.39, 406–473.
Miyazaki, Izumi (2004). “The Gradualist Chapter of the bSam gtan mig sgron and the
Teaching of Kamalaśīla”, Journal of Indian and Buddhist Studies 52/2, pp. 899–902.
Monier-Williams, M. (2001). A Sanskrit-English Dictionary. New Delhi: Asian Educational
Services.
Nasr, Seyyed Hossein (1989). Knowledge and the Sacred. Albany NY: State University of
New York Press.
Negi, J.S. (2006). Dharmasaṅgraha-Kośaḥ. Sarnath: CIHTS.
gNubs-chen Sangs-rgyas ye-shes (1974). rNal-’byor mig-gi bsam-gtan or bSam-gtan mig-
sgron: A treatise on bhāvanā and dhyāna and the relationships between the various
approaches to Buddhist contemplative practice. Reproduced from a manuscript made
presumably from an Eastern Tibetan print by ’Khor-gdong gter-sprul ’Chi-med rig-
’dzin, Smanrtsis shesrig spendzod, vol.74. Leh: Tashigangpa.
Nyang-ral nyi-ma ’od-zer (1988). Chos-’byung me-tog snying-po sbrang-rtsi’i bcud. Lhasa:
Bod-ljongs mi-dmangs dpe-skrun-khang.
Partridge, Eric (1982). Origins: A Short Etymological Dictionary of Modern English.
London: Routledge & Kegan Paul.
Pelliot tibétain 116.
Piatigorsky, Alexander (1984). The Buddhist Philosophy of Thought: Essays in Interpretation.
London: Curzon Press.
Rahula, Walpola (1971). Le Compendium de la Super-Doctrine (Philosophie)
(Abhidharmasamuccaya) d’Asaṅga. Paris: École Française d’Extrême-Orient.
Schuon, Frithjof (1993). Treasures of Buddhism. Bloomington IN: World Wisdom Books.
Schuon, Frithjof (1995). L’Œil du Coeur. Lausanne: L’Âge d’Homme.
Stcherbatsky, Theodore (1970). The Central Conception of Buddhism and the Meaning of
the Word “Dharma”. Delhi: Motilal Banarsidass.
Tanzin, Lopon P. Ogyan (2013). “Assessing the Greatness of Tibet’s Early Translations
according to Rong-zom Mahāpaṇḍita”, Translated by Dylan Esler. Temenos Academy
Review 16, pp. 21–42.
Van den Broeck, José (1977). La Progression dans la Méditation (Bhāvanākrama de
Kamalaśīla), Traduction du sanscrit et du tibétain. Bruxelles: Institut Belge des Hautes
Études Bouddhiques.
350 Teng
Part 3
Philosophical Studies
∵
352 Teng
Madhyamaka in Abhidharma Śāstras 353
Chapter 11
Madhyamaka in Abhidharma Śāstras: The Case of

Harivarman’s *Tattvasiddhi
Goran Kardaš
1 Introduction
There are only a few very brief references to the Madhyamaka philosophi-
cal standpoint in the preserved Abhidharma śāstras. In the ninth chapter of
the Abhidharmakośabhāṣya (AKBh), Vasubandhu accuses those Buddhists
who propound the existence of pudgala (Vātsīputrīyas) and those who con-
tend that nothing exists (sarvanāstitāgrāhaka) of introducing heresy into
the Buddhadharma. These doctrines are wrong and do not lead to libera-
tion (mokṣābhāvadoṣa) (Pradhan, 1975: 472, lines 13–14). In his commentary
(Vyākhyā) on the AKBh (Wogihara, 1932–1936: 710, 31f.), Yaśomitra explicitly
identifies those who contend that nothing exists with the Madhyamaka school.
Vaibhāṣika master Saṃghabhadra, speaking in his *Nyāyānusāra about those
Buddhists who are not worthy of being called members of Sarvāstivāda, men-
tions those who “under the empire of distorted mirror hold that all dharmas
are devoid of own nature and that even present existence of dharmas is illu-
sory and wrong” (de La Vallée Poussin, 1937: 89). Another Vaibhāṣika śāstra, the
Abhidharmadīpa (kārikā 299 and vṛtti), criticizes those propounders of empti-
ness (śūnyatāvādinaḥ) who contend that nothing exists (sarvaṃ nāsti) in any
of the three time periods, calling them “upholders of destruction” (vaināśika)
and “propounders of the emptiness of the connection (between cause and
effect)” (ayogaśūnyatāvādins).1
This situation is slightly different in the case of the *Tattvasiddhi2 (hence
forth: TS) of Harivarman (third to fourth century CE), who, according to Shōryū
1 In my translation of ayogaśūnyatāvādin, I have followed Jaini’s understanding (cf. Jaini, 1977:

123) of this compound, not to be found in any other Buddhist text. Jaini understands ayoga as
absence (“disconnection”) of causal interrelatedness. Therefore, following his understanding,
the translation could be either as I have done, or alternatively, as “propounders of the doctrine
that emptiness is (implies or means) absence (“disconnection”)” of any actual causation or
conditioning activities or so.
2 In the present chapter, I will utilize N. Aiyaswami Sastri’s Sanskrit (re)translation of the text
(“Satyasiddhi” in his reconstruction) from the classical Chinese, and I will also follow his
pagination.

354 Kardaš
Katsura, was a member of the Bahuśrutīya school.3 If so, this Śāstra is the only
surviving representative of Mahāsāṃghika Abhidharma literature. The Śāstra
also discloses a strong Sautrāntika affiliation on many doctrinal issues. The TS
was translated into Chinese in the early fifth century, and had a huge influence
on Chinese Mādhyamikas at some point, as it was wrongly understood as a
proper Mādhyamika work. Nevertheless, although composed as a typical
Abhidharma śāstra centered around the four noble truths, some of its key
philosophical views have an undeniable Madhyamaka tinge.
The relevant portion of the Śāstra I will focus on in this chapter, is situated
in the fourth chapter called “Nirodhasatya,” sections 142–151. I will focus on
three issues:
1. What could be the main source for Harivarman’s presentation of an

apparently Mādhyamika position?
2. What is the nature of his criticism of the Mādhyamika position, bearing
in mind his overall philosophical standpoint?
3. What is the general Mādhyamika answer to the kind of criticism
Harivarman exposes in his treatise?
2 The Main Source for Harivarman’s Presentation of an Apparently

Mādhyamika Position
Our part of the text begins with the enumeration of the four wrong doctrines
that presumably concern the existential status of objects of everyday experi-
ence, namely the doctrine of unity or identity (ekatva), of diversity or difference
(nānātva), of unspeakability (anirvacanīyatva) and of non-existence (abhāva)4
(142). According to the first doctrine, parts of, e.g. a pot (color, smell etc.) are
identical to the pot itself; according to the second, they are different from it;
according to the third, it cannot be said whether they are identical to or differ-
ent from the pot; according to the fourth, the pot does not exist (at all).
3 Cf. his argumentation on this point as well as on the TS’s doctrinal affiliations and disagree-
ments with other traditional Buddhist schools in Katsura, 1974: 29–49.
4 Katsura, 1974: 181, is of the opinion that Harivarman here presents four arguments concerning
“the relationship between concepts and dharmas”; Priestley, 1970: 31, thinks so also (concepts
and real phenomena). This would hold for the first three positions (the relationship between
parts and the whole), but not for the fourth position (“the pot does not exist”). I think that
Harivarman here more generally provides four possible explanations of the existential status
of everyday objects.
A brief note on the first three doctrines: the doctrine of identity is wrong
because the idea (buddhi) of parts (color etc.) and the idea of the pot are differ-
ent. Only the parts exist in reality, while the pot exists only by name since its
existence is “borrowed,” being merely (conceptually) based on its (real) parts
(143); the doctrine of diversity or difference is wrong because there is no pot
apart from its color etc., since no idea of it arises in their absence, and it is
therefore illogical for the pot to exist apart from its color etc. (144); the doctrine
of unspeakability is wrong because no substantial dharma (dravyadharma) is
unspeakable in terms of unity and diversity – each phenomenon (dharma) has
particular characteristics (svalakṣaṇa), which implies its identity with itself
and its difference with regard to other (dharmas), and so it is speakable: this is
rūpa, etc. (145)
For Harivarman, all of these doctrines, including the fourth (the pot does
not exist) to which we will now turn, are wrong because the pot etc. is nomi-
nally or conceptually existent (prajñaptisat).5
Harivarman dedicates the most space by far to the presentation and criti-
cism of the fourth doctrine, that of non-existence (abhāva). He firstly provides
a general criticism of this doctrine, or better said, outlines some unwanted
consequences of this doctrine that are very similar in nature to the objections
of Nāgārjuna’s anonymous opponent from the beginning of the Madhyama
kaśāstra (MMK) 24 (1–6), as well as elsewhere, e.g. in the Vigrahavyāvartanī
(VV). A free translation of some of these objections follows (TS, 146, 366–367):
If all dharmas are abhāva, then there is neither virtue nor its opposition
nor deliverance. If it is thought that what exists does not exist, nor will
that position (‘apprehension,’ grāha) exist, because in that case neither
speaker nor listener exist. Doctrines about existence and non-existence
are expressed through faith either in perception or in inference or in
scriptures. That there is nothing (kiñcana nāstīti) does not come under
any one of the above categories … You think that the idea of a dharma
arises as a result of a delusion. But if everything is non-existent
(abhāvatva), this illusion will also be non-existent. How will one act? If
you think that all dharmas are non-existent, in dependence on what
basis does this knowledge arise? No knowledge arises in dependence on
the non-existent. That which knows dharmas is knowledge. If dharmas
were in the nature of absolute non-existence (atyantabhāvā), then
5 Harivarman defines prajñapti as a “mental construct” (vikalpa) based on the five aggregates
(pañcaskandhān upādāya) (141). For other definitions of prajñapti: see Katsura, 1974:
179–181.
356 Kardaš
people would do anything they like. But wise people are interested in
charity, etc. … The pot, etc. are cognized through the perceptive faculty,
and yet you think that every perceived dharma is a non-entity. None will
have faith in the sūtras if they teach non-existent dharmas (abhāvadharma
katvācca na sūtre śraddhādhīta).
Then Harivarman provides the abhāvavādin (or nāstika, “nihilist”) the oppor-
tunity to philosophically elaborate on his position. Although not sharply
distinguished, it seems that the advocates of abhāva put forward three argu-
ments, epistemological and ontological in nature, for the non-existence of
dharmas. Here is a brief account of some of these arguments (cf. also Katsura,
1974: 183–184 and Priestley, 1970: 31–32):
a) Refutation of the parts and the whole

The first argument for the non-existence of dharmas is centered around the
impossibility of grasping a sense faculty’s object field (indriyaviṣaya) because
objects have no separate whole (avayavin) that can be grasped. Contrary to the
general standpoint of the Ābhidharmikas, parts (dharmas) have no particular
characteristics, but are (conventionally) considered to be “dependent” upon
the whole just as the whole is “dependent” upon (its) parts. Beyond this inter-
dependence (and inter-designation), there are no real parts (with particular
characteristics), and consequently, nor is there any whole. Moreover, all parts
are non-existent because they can be further analyzed (reduced) over and over
into smaller parts, and this reduction can finally end only in ultimate non-exis-
tence (atyantābhāva). Therefore, “the idea of emptiness is necessarily (and
finally) generated regarding all dharmas” (sarveṣāṃ dharmāṇāṃ niṣṭhā
śūnyatābuddhijananam avaśyam). (TS, 147, 368)
b) Refutation of perception
Most arguments against the possibility of perception utilize the Ābhidharmika
concept of momentariness, turning it against their upholders:
b1) (Question:) Visual consciousness first grasps color, and then non-sen-
sory consciousness remembers it. (Answer:) … Visual consciousness, after
seeing color, disappears, and then non-sensory consciousness arises.6 This
6 This Sautrāntika position was criticized by Saṃghabhadra in his defence of the Vaibhāṣika
position according to which visual objects, the visual faculty, and visual consciousness
necessarily arise simultaneously because otherwise the visual faculty and visual objects
could not serve as conditions for visual consciousness; cf. Dhammajoti, 2007: 137. The
Sautrāntika solution to this problem is that contact between visual objects and the cor-
responding sense-faculty in the first moment produces an “image” (ākārā) of the object,
consciousness does not see color. If it does not see it, how can it recollect it? In
that case, the blind would also recollect it [without seeing it].7
b2) [You may say that] non-sensory consciousness arises from visual con-
sciousness, and hence recollects it. This cannot be. Why? Because in that case,
all consciousness of the last [moment] that originated from visual conscious-
ness would recollect it and never forget it.8
b3) The eye cannot see rūpa regardless of whether it reaches (prāpya) it or
not (aprāpya). In the first case, as has already been shown, visual conscious-
ness disappears after (possibly) reaching rūpa, having the characteristic of the
past (atītalakṣaṇam), and so cannot “deliver” its content to non-sensory con-
sciousness. In the second case, the absurd consequence would be that visual
consciousness would grasp any rūpa.9
b4) If the eye consisted of four great elements and would see color, the ear
etc. would also see, because it similarly consists of four great elements.10
b5) Because of the momentariness (kṣaṇika) of consciousness (citta) and
sound (śabda), none can actually hear the unity (ekatva) of e.g. the word
“puruṣa,” but only the momentary arising and disappearance of syllables one
after another (‘pu,’ ‘ru,’ ‘ṣa’).11
b6) Non-sensory consciousness does not grasp dharmas; for it does not
grasp color, taste, odor and touch in the present moment…12
which is an exact mental representation of the object and serves as a condition for the
origination of corresponding visual consciousness in the second moment.
7 (pṛ) cakṣur vijñānena rūpe gṛhīte tato manovijñānam anusmarati … (u) cakṣur vijñānaṃ
rūpaṃ dṛṣṭvā niruddham eva | tata ūrdhvaṃ manovijñānam utpadyate | manovijñānam
idaṃ na rūpaṃ paśyati | adṛṣṭvā katham anusmaret … andho ’pi rūpam anusmaret, TS, ibid.
8 (pṛ) cakṣur vijñānān manovijñānam utpadyate | ato ’nusmarati | (u) maivam | kasmāt |
sarvāṇi caram acittāni cakṣur vijñānam upādāya samutpannāni anusmareyuḥ | na eva
vismareyuḥ, TS, ibid.
9 yadi vadasi cakṣuḥ paśyati iti | kiṃ rūpaṃ prāpya paśyati kiṃ vāprāpya paśyati | yadi
prāpya [paśyati iti] tadā na paśyati | cakṣur nātītalakṣaṇam iti idaṃ pūrvam eva
pratipāditam | yady aprāpya paśyati iti | tadā sarvasthaṃ rūpaṃ paśyet, TS, ibid., 369.
10 kiñca cakṣuś caturmahābhūta[mayam] | yadi cakṣuḥ paśyati | śrotrādīni api paśyeyuḥ |
caturmahābhūtasāmyāt, TS, ibid.
11 cittaṃ hi kṣaṇikam | śabdo ’pi kṣaṇikaḥ | yathā vadanti puruṣa iti | aya [mekatva]vādo na
śrāvyaḥ | kasmāt | «pu» śravaṇam anu vijñānaṃ na «ruṃ» śṛṇoti | «ruṃ» śrutvā na «ṣaṃ»
śṛṇoti, TS, 148, 370. This was essentially Vasubandhu drawing an absurd consequence,
obviously taken over from some earlier source (probably the Mahāvibhāṣāśāstra), of the
Vaibhāṣika’s understanding of syllables (along with names and phrases) as real existing
entities (dravyadharmas); cf. Bhāṣya to AK 47ab and also Cox,1995: 168.
12 Manovijñānam api dharmān na gṛhṇāti | kasmāt | manovijñānaṃ hi na pratyutpannān
rūparasagandhasparśān gṛhṇāti | [yat] atītamanāgataṃ tan na asti iti …, TS, 150, 371.
358 Kardaš
c) Refutation of cause and effect

c1) If the fruit exists, it should arise as a quality already existent in its cause
or non-existent in its cause. Both are defective. If the fruit is already existent in
its cause (satkārya), it will not arise (again). How can the existent arise (again)?
If it is non-existent (asatkārya), then it will not arise again. How can the non-
existent ever arise? (TS, 151, 372)
c2) The cause may be either anterior to (pūrvaṃ) the fruit, posterior to
(paścāt) it, or simultaneously (samakālam) with it, but all of these alternatives
do not hold (na yujyate). If it is anterior to the fruit, then how does the fruit
arise after the cessation (niruddhe) of the cause? If it is posterior to the fruit,
then how will it produce the fruit if it is non-produced (svayamanutpannaḥ)
itself? If both are simultaneous, then how can they be distinguished? E.g. two
horns appearing at the same time (yugapadudbhūte) are not regarded as mutu-
ally (anyonya) caused (TS, ibid.)
c3) If it is considered that the cause and the fruit are one (eka) or different
(nānā), both are wrong. If they are different, then garments would appear
without threads. If they are one, the garment and threads would not be distin-
guished (vibhāga). (TS, ibid.)
c4) If a fruit exists, it is produced by itself, by another, by both, or without
cause (svakṛtaṃ parakṛtam ubhayakṛtam ahetukṛtaṃ vā). The first alternative
is wrong because no dharma can make itself (na kaścid dharmaḥ svātmānaṃ
karoti); the second is wrong because all dharmas are agentless, being devoid of
the idea of agenthood (kartṛtvasaṃjñābhāvāt sarve dharmā akartṛkāḥ); the
third is wrong because it includes the defects of both alternatives; the fourth is
wrong because, in the absence of a cause, a fruit does not exist either (hetāvāsati
phalam api nāstīti). (TS, ibid., 372–373)
At least two of these three groups of arguments for the non-existence of
dharmas cannot be found in Nāgārjuna’s work, at least not explicitly: the
parts-whole argument13 and the argument on the refutation of the process
of perception as described above. Of course, this does not necessarily mean
that Nāgārjuna was not familiar with them.14 Be that as it may, in MMK and
VV, Nāgārjuna only deals with the impossibility of the establishment of
13 There is, however, a hint of this argument in Vaidalyaprakaraṇa (32–39) where Nāgārjuna
(if he is the author of this work at all) refutes the Naiyāyika’s five-membered (avayava)
syllogism since they are not subject to a whole (avayavin); and even if they were one with
a whole, they would be identical with it, etc. (cf. Lindtner, 1990: 89–90).
14 For example, he uses the argumentative structure “x cannot make contact with y either by
reaching (prāpya) it or not reaching (aprāpya) it” (cf. b3 above) in the context of light and
darkness (MMK 7.10–11; also VV 38–39).
pramāṇas (instruments of valid cognition) (VV 30–33 and 40–51) or dialecti-

cally investigates the impossibility of the establishment of any of the triad of
seer-seeing-seen (MMK, chapter 3).
As for the refutation of causality, although Nāgārjuna utilizes most of the
above mentioned arguments, they are elaborated upon much further here, at
least in some cases, than in any of his other works, and in fact look like further
commentarial clarifications (i.e. drawing unacceptable consequences) of con-
tradictions he found in various theses that are stated in the form of general and
abstract prasaṅga type inferences.
Thus, e.g. although Nāgārjuna employs the general argument that “a state
(or entity) x can be neither anterior, nor posterior, nor simultaneous with a
related state (or entity) y” (cf. c2 above), even in the case of cause and effect (cf.
MMK 11.6–8, 9.12, relationship between an entity and the function of seeing
etc.), he does not draw unacceptable consequences issuing from any of these
three temporal positions between two entities as TS does.15 Likewise, Nāgārjuna
employs the initial argument that cause and effect can be neither identical nor
different (cf. c3 above) (MMK 18.10 and 20.19–20), but, unlike TS, he does not
elaborate this further.16
As for arguments c1 and c4 (see above) against causation, Nāgārjuna does
not deal straightforwardly with the Sāṃkhya or Nyāya-Vaiśeṣika theory of cau-
sation (satkārya- and asatkāryavāda), nor does he mention these doctrines by
name. The c4 general argument, of course, can be found in MMK 1.1, but further
reasoning as to why the production of an entity is not defensible in any of the
four possible ways is, unlike in TS, absent. Note, however, that the unaccept-
able consequences drawn in TS do not correspond with Buddhapālita’s
reasoning as cited in the Prasannapadā.17
15 The commentaries (e.g. Candrakīrti), of course, do: tatra yadi pūrvaṃ kāraṇaṃ paścāt
kāryaṃ syāt, akāryakaṃ kāraṇaṃ nirhetukaṃ syāt | atha pūrvaṃ kāryaṃ paścāt kāraṇam,
evam api kāraṇāt pūrvaṃ kāryaṃ nirhetukam eva syāt | atha yugapat kāryakāraṇe syātām,
evam ubhayam apyahetukaṃ syāt. PP 225.6–8 (vṛtti to MMK 11.7–8); cf. also vṛtti to MMK
9.12 (the same type of investigation between seeing and perceiving a subject, ātman).
16 Cf. PP 376.10–15 (vṛtti to MMK 18.10): naiva yad eva bījaṃ sa eva aṅkuraḥ, janyajanakayor
ekatvaprasaṅgāt … ananyatvāc ca aṅkurāvasthāyām aṅkuravadbījagrahaṇam api syāt,
bījavac ca aṅkurasya api grahaṇaṃ syāt … na ca anyad api tattasmāt | nāpi bījādaṅkura
syānyatvam, bījamantareṇa api aṅkurodayaprasaṅgāt; cf. also vṛtti to MMK 20.19–20.
17 Ācāryabuddhapālitas tv āha – na svata utpadyante bhāvāḥ, tadutpādavaiyarthyāt,
atiprasaṅgadoṣāc ca | na hi svātmanā vidyamānānāṃ padārthānāṃ punarutpāde prayoja-
nam asti | atha sannapi jāyeta, na kadācin na jāyeta iti … na parata utpadyante bhāvāḥ,
sarvataḥ sarvasaṃbhavaprasaṅgāt … dvābhyām api nopajāyante bhāvāḥ, ubhayapakṣā
360 Kardaš
The parts-whole argument and the refutation of existence on the basis of

the impossibility of the process of perception in the Abhidharma framework,
together with the refutation of causality, became standard arguments for
śūnyatā and were heavily utilized in the middle period of the Madhyamaka
school. However, we can find all of these arguments elaborated upon or at least
touched upon to some extent in the works of Nāgārjuna’s famous pupil
Āryadeva. Not only this, but there is also a striking similarity between ŚŚ (chap-
ters 3–8; cf. Tucci, 1929: 37–72) and TS 143–151 (especially 143–144, 146–147 and
150–151) in the presentation and refutation of false doctrines, specifically the
doctrines of unity and diversity (ŚŚ chapters 3–4 and TS 143–144), based on the
investigation of the relationship between parts and the whole,18 the presenta-
tion and refutation of the process of perception (ŚŚ chapters 5 and TS 147), and
the presentation and refutation of satkārya- and asatkāryavāda (ŚŚ chapters
7–8 and TS 151).
I will now briefly call attention to similarities between Āryadeva’s argu-
ments found in ŚŚ and CŚŚ with some of the arguments described above (a-c),
which Harivarman ascribes to the abhāvavādins (or nāstikas) in their argu-
ment for the non-existence of dharmas.
c argument (refutation of parts and the whole): “If the foot [i.e. a part] and
the body [i.e. the whole] are not different, how is it that the foot is not the
head?” (ŚŚ, Tucci, 1929: 40) “The color is a part of the pot and therefore it is not
the pot. Since the whole [avayavin] does not exist, the part [avayava] does not
exist either”. (CŚŚ XIV 9; cf. Lang, 1983: 479; cf. also XIV 2, 7, 12 and 15).
b1 argument: “A dharma that has been seen does not reappear. Moreover, the
cognition of it does not occur. Therefore, to talk about memory is false”.19 (CŚŚ
XI 25; cf. Lang, 1983: 407)
b3 argument: “Were the eye to possess motion, then it should perceive that
[object] at a distance over a long period of time”. (CŚŚ XIII 13; cf. Lang, 1983:
bhihitadoṣaprasaṅgāt pratyekam utpādāsām arthyāc ca … ahetuto ‘pi notpadyante-

hetāvasati kāryaṃ ca kāraṇaṃ ca na vidyate; PP 14.10 (vṛtti to MMK 1.1).
18 Harivarman refutes the doctrines of identity and diversity (the relationship between
parts and the whole) in TS 143–144 within the framework of Abhidharma philosophy, i.e.
parts (dharmas) are real while the whole is not, and the latter has only provisional exis-
tence (and no particularly characterized existence) based upon parts (TS 143–144), while
the abhāvavādin (in TS 147) refutes the parts and the whole altogether.
19 The underlying idea here is obviously that when visual consciousness grasps a visual
dharma, it disappears in the first moment along with its content, so the latter cannot
“reappear” (na punar bhavet) as the object field, i.e. the cognition of non-sensory con-
sciousness in the second moment. Consequently, the memory of the supposedly cog-
nized object cannot occur.
455) “Were the eye to apprehend [the object] without going, then it should
perceive this entire world!” (CŚŚ XIII 15; cf. Lang, 1983: 456) “If the eye goes far
and far, it sees by and by”. (ŚŚ, cf. Tucci, 1929: 51)
b6 argument: “Deprived of the sense faculties, what will the mind (citta) do,
even after it has gone [to the object]?” (CŚŚ XIII 21; cf. Lang, 1983: 460)
c1 argument: “If the fruit exists, then, because the fruit is pre-existing in the
cause, the cause would be without effect on account of the non-existence of
the fruit”. (ŚŚ; cf. Tucci, 1929: 61; cf. also Tucci, 1929: 66, 70) “An existent dharma
does not arise from an existent dharma … a non-existent dharma does not
arise from an existent dharma”. (CŚŚ XV 14; cf. Lang, 1983: 511; cf. also CŚŚ XV 15,
Lang, ibid. and CŚŚ XI 12d and 15c)
In the light of these similarities in reasoning, and especially in the textual
topical ordering of false doctrines (ŚŚ chapters 3–8 and TS 143–151), my conten-
tion (which, of course, cannot be completely conclusive) is that the works of
Āryadeva, not those of Nāgārjuna, were the main source for Harivarman’s pre-
sentation (and knowledge) of Madhyamaka philosophical views and reasoning,
especially bearing in mind the more highly developed style of argumentation
presented in Āryadeva’s works compared to Nāgārjuna’s.20 This advancement,
as has been shown above, is also reflected in Harivarman’s treatise.21 Moreover,
there is one explicit (and affirmative) citation in TS (105) of CŚŚ (II 8), albeit in
a context that is not relevant to our presentation, but that nevertheless explic-
itly suggests that Harivarman was at least familiar with Āryadeva’s works.
3 The Nature of his Criticism of the Madhyamaka Position
The next issue I would like to discuss is the nature of Harivarman’s criticism of
an apparently Madhyamaka position which he, like the other Abhidharma
sources mentioned, terms abhāvavāda, calling its followers abhāvavādins (or
nāstikas). I believe that his criticism streams from his total ignorance of the
actual (Nāgārjuna’s) Madhyamaka position regarding the existential status of
objects of everyday experience and the conventional (vyavahāra, saṃvṛti) in
20 Frauwallner (1956: 218) has noted that although Āryadeva agrees with Nāgārjuna in all
important views, he exceeds him in his “style of presentation” (Art der Darstellung).
Unlike Nāgārjuna, Āryadeva treats the philosophical views of his opponents more thor-
oughly and in greater detail.
21 We cannot identify any major figure in the development of Madhyamaka philosophy
between the time of Nāgārjuna (ca. second century CE) and the classical commentarial
tradition (Buddhapālita, Bhāvaviveka and Candrakīrti, from the fifth century CE onward)
other than Āryadeva.
362 Kardaš
general, and that Āryadeva is to blame for this.22 However, before providing
some textual evidence for this thesis, it would not be out of place here to briefly
outline Harivarman’s ontological position, which is partly based on the stan-
dard Abhidharma ontology and is obviously crucially influenced in part by the
Mādhyamika position itself.
In certain parts of the text, Harivarman posits the standard Abhidharma
theory of the two truths (saṃvṛti- and paramārthasatya), whose “originator” is
the Buddha himself (TS 3, 12) but whose actual preaching was based on praj-
ñapti (“concepts”) (TS 4, 16), i.e. on saṃvṛtisatya. TS 141, 354–355 gives the
standard Abhidharma account of the content of the two truths:
Paramārthasatya are dharmas, rūpa, etc. and nirvāṇa, while saṃvṛtisatya

is mere prajñapti that is devoid of own nature. [As for the latter], just as
the [concept] ‘jar’ is established on the basis of rūpa, etc., so the [con-
cept] ‘person’ is established on the basis of the pañcaskandhas.23
One of the characteristics of the conceptually existent (prajñaptisat) is that it

is established in dependence on some other dharma, e.g. the jar is established
in dependence on the rūpa, etc.,24 while the absolute dharma (paramārtha
dharma), e.g. feeling (vedanā), is not (i.e. it is self-established). (TS 142, 358).
Furthermore, this means that the generation of knowledge cannot be based on
prajñapti-, but on paramārthadharma:
22 Although, unlike his commentators, Nāgārjuna never discusses the actual “content” of the
conventional, he is more than explicit in that his emptiness-based philosophical analysis
should be understood within the realm of the conventional, cf. e.g. his auto-commentary
to VV 28: “We, however, do not say ‘all dharmas are empty’ without relying on the conven-
tional truth, rejecting the conventional truth. For it is not possible to teach the Dharma
without having recourse to the conventional truth” (api ca na vayaṃ vyavahārasatyam
anabhyupagamya vyavahārasatyaṃ pratyākhyāya kathayāmaḥ śūnyāḥ sarvabhāvā iti / na
hi vyavahārasatyam anāgamya śakyā dharmadeśanā kartuṃ). On the other hand, this
emphasis on the conventional grounding of the teaching of emptiness is completely
absent in Āryadeva, and that could give his opponents an impression that for the Mādhya
mika “does not exist ultimately” essentially means “does not exist at all” (abhāvavāda).
23 “paramārthasatyaṃ yaduta rūpādayo dharmā nirvāṇañ ca | saṃvṛtisatyaṃ yat prajñapti
mātraṃ niḥsvabhāvam | yathā rūpādipratyayo ghaṭaḥ sidhyati | tathā pañcaskan
dhapratyayaḥ puruṣaḥ sidhyati”.
24 That does not mean, according to Harivarman, that what exists conceptually does not
exist at all. It simply means that it does not exist substantially (na ca prajñaptisannāsti iti
vastumātraṃ na bhavati, TS, 35, 93) and that it is devoid of own nature (niḥsvabhāva; TS
142, 360).
The arising of knowledge is not in prajñapti, but rather in rūpa, etc. that
is present before man’s eyes. After that, [a man conceptually] construes
with [his] non-sensory consciousness: “I see the jar” … In prajñapti, doubt
arises as to whether it is a pillar or a man. But in [the case of] rūpa, etc.,
no doubt arises as to whether it is rūpa or śabda.25
Paramārthasat differs from prajñaptisat (or saṃvṛtisat), among other things,

because it “can be neither burned nor destroyed”. (paramārthadharmā
syādāhyatvāt avināśyatvāt, TS 142, 361). Likewise, paramārtha is not based on
anything (else), e.g. rūpa not being dependent upon other dharmas is (self-)
established (paramārthastu nāpekṣikaḥ … rūpaṃ nānyapadārthamapekṣya
… sādhayatîti, TS, ibid.). All of these characterizations of the two truths are
in complete accord with, for example, the Sarvāstivāda Abhidharma view.
In other places, however, Harivarman presents a different view regarding
the “content” of the two truths. Thus discussing the purpose of writing a
(Buddhist) “scientific treatise” (śāstra), he states that it should be based on
two principles (or should have two “faces,” mukha), namely lokamukha (or
saṃvṛti-) and paramārthamukha. The former states that ātman exists (asty
ātmā ity ucyate, TS 14, 40), while the latter states that everything is empty and
non-existent. (paramārthamukhatastūcyate sarvaṃ śūnyamasat iti, TS, ibid.)
Paramārthasat is then identified elsewhere in the text with the third noble
truth (nirodhasatya, the truth of cessation), and no longer with the skandhas,
rūpa, etc. (nirodhaḥ paramārthasan, na skandhā iti, TS 153, 377) Cessation itself
is nothing else than non-existence, and so Harivarman finally concludes that
“on the paramārtha level all dharmas (saṃskāras) are non-existent in nature,
while they exist on the level of saṃvṛti”. (nirodhaścābhāva … paramārthataḥ
saṃskārāḥ sarve ’bhāvātmakāḥ saṃvṛtitas tu santi saṃskārā iti, ibid., 379)
These two different “versions” of the two truths are prima facie mutually
exclusive. There obviously cannot be “two two truths”. Harivarman was aware
of this apparent inconsistency when he answered the question posed by his
imagined interlocutor:
(Question): If the pañcaskandhas are of the level of saṃvṛtisatya, why is

it said that dharmas such as rūpa, etc. are paramārthasatya?
25 prajñaptau ca na jñānotpādo bhavati | purovartirūpādiṣu jñānotpādo bhavati | taduttaraṃ

manovijñānena vikalpayati ghaṭamahaṃ paśyāmi iti … prajñaptau ca bhavati saṃśayaḥ
yathā sthāṇurvā puruṣo vā iti | rūpādiṣu tu na saṃśayo bhavati rūpaṃ vā kiṃ vā śabda iti
(TS 142, 360).
364 Kardaš
(Answer): It is said [so] for the good of men [because] there are some
[men] who harbor the notion of paramārtha towards the pañcaskandhas.26
These men are obviously Ābhidharmikas, who “investigate dharmas” (dhar-

mapravicaya) as the only means by which to reach nirvāṇa (cf. AK 1.3). Just like
Candrakīrti in PP (39. 1–44. 15), albeit far less systematically, Harivarman goes
on to provide canonical citations and the manner in which they should be
understood in their real purport, e.g.:
It is said that conditioned by birth [there is] old age-and death. This is
called the middle way. This should be understood [here as] that there is
no old age-and death from the paramārtha level. That conditioned by
birth [there is] old age-and death is said from the saṃvṛti level.27
It thus follows that teachings based on the investigation of dharmas are noth-
ing but an “expedient means” (upāya) by which to reach nirvāṇa devised for
those who “harbor the notion of paramārtha towards the pañcaskandhas”:
“Meditation on skandhas, dhātus, āyatanas, etc. are [only] different doors and
expedient means for entering nirvāṇa”.28
This “shift” in the meaning of the two truths places the Ābhidharmika’s
paramārtha analysis on the level of saṃvṛti analysis29 together with the “origi-
nal” saṃvṛtisat (or prajñaptisat), leaving the emptiness or non-existence of all
dharmas as the sole “content” of paramārthasatya. Harivarman obviously con-
siders the Ābhidharmika’s paramārtha “context-bound,” i.e. relative to the
26 (pṛ) yadi pañca skandhāḥ saṃvṛtisatyataḥ santi | kasmād ucyante rūpādayo dharmāḥ
paramārthasatyā iti | (u) sattvānāṃ kṛta ucyante | santi kecit pañcaskandheṣu samutpanna
paramārthasaṃjñāḥ (TS 153, 376).
27 kiñcāha- jātipratyayaṃ jarāmaraṇaṃ madhyamā pratipadityucyate | nāsti jarāmaraṇaṃ
paramārthata ity uktamiti jñātavyam | saṃvṛtita ucyate jātipratyayaṃ jarāmaraṇam iti (TS
153, 377). In this example, Harivarman seems to use the same argument as the
Mādhyamikas do in their pointing to the emptiness of dharmas – the fact of the condi-
tionality of all dharmas. There are even more explicit examples that corroborate this ‘bor-
rowing’ to which we will turn soon.
28 skandhadhātvāyatanādīnāṃ bhāvanā nānādvārāṇyupāyā nirvāṇe ’vataraṇasya (TS 201,
537).
29 Cf. TS 153, 376: “Then the pañcaskandhas do not exist substantially? They do not. They
exist from the saṃvṛti level” ((pṛ) … kimidānīṃ na santi vastutaḥ pañcaskandhāḥ | (u) na
santi vastutaḥ | santi tu saṃvṛtitaḥ).
Ābhidharmika’s saṃvṛti level, while the last and final paramārtha is ‘context-
free,’ i.e. it is absolute.30
Following this analysis, Harivarman develops a theory of three basic types
of consciousness with three corresponding types of content: prajñapticitta
(makes e.g. a pot its content), dharmacitta (makes rūpa, etc. its content, thus
suppressing the idea of the pot), and śūnyatācitta (makes the non-existence or
emptiness of the rūpa, etc. its content, thus suppressing the idea of dharma
lakṣaṇa). The last citta is suppressed by entering into nirodhasamāpatti (cf. TS
141, 354). Harivarman seems to clearly distinguish here between śūnyatā as the
(still conceptual) content of śūnyatācitta and the experiential “state” of
śūnyatā, which is beyond any relational designation.
There is no doubt that Harivarman adopted the Madhyamaka position
regarding the final nature of all dharmas “as they are” (yathābhūta).31 Many of
his characterizations of the real nature of dharmas appear to have been copied
from a Madhyamaka śāstra. Here are some examples:
If [skandhas, dhātus and āyatanas] were real, then asaṃskṛtadharmas,

the Tathāgata, etc. would also be real. But in fact they are non-existent.
Therefore, it is known that dharmas like skandhas, dhātus and āyatanas
are not grasped [as if they possess self-] characterization.32
To he who pursues the functions and characteristics of skandhas, the
thought of ātman is not utterly destroyed because [the chain of] causes
and conditions is not destroyed.33
30 Priestley, 1970: 35, believes that Harivarman here “imperfectly formulated [the] theory of
three truths: conventional truth, which is truth based on concepts; conventional ultimate
truth, which is based on phenomena; and ultimate truth, which is the truth of Nirvāṇa”.
However, I believe he is incorrect, as his “conventional ultimate truth” (corresponding to
dharmacitta) is nothing but the “Buddhist conventional truth” especially “designed” for
those, as Harivarman puts it, who “harbor” the notion of paramārtha towards the pañca
skandhas” (see above). “Conventional truth” (corresponding to prajñapticitta) is, on the
other hand, conventional truth for all others, i.e. ordinary men (pṛthagjanas). What is
crucial in this scheme, is that the respective content of all levels of consciousness appears
as paramārtha, which can be downgraded to the level of saṃvṛti only by a subsequent
and, so to speak, more investigative type of consciousness. Consequently, and opposed to
Priestley, 1970: 36, I believe that Harivarman does not analyze pots, rūpas, etc. ontologi-
cally, but rather epistemologically as the content of consciousness.
31 Cf, TS 16, 50: śūnyataiva yathābhūtam and TS 192, 506: nirodhaḥ sa niyataṃ yathābhūta ity
ataḥ paramārthasan iti.
32 TS 20, 69: tathā cet tathāgatādayo ’saṃskṛtadharmā api santaḥ syuḥ … vastutastu asan-
taste | ato jñāyate skandhadhātvāyatanasaṅgṛhītā dharmā na sallakṣaṇā iti.
33 TS, 130, 317: skandhavṛttilakṣaṇavṛttim anusarato nātmam atiratyantaṃ prahīyate |
hetupratyayānām anirodhāt.
366 Kardaš
The vision of emptiness is simply the non-grasping of [any] sign regard-

ing [all] dharmas. Because of the cessation of [all] signs, there is no more
birth….34
It is said in the sūtra: “The five skandhas are empty and are like an illu-
sion.” (coktaṃ sūtre- pañcaskandhāḥ śūnyā māyāvat iti, TS 192, 506)
All saṃskṛtadharmas are said to be illusions because they undergo modi-
fications. Being illusions, they are not real. Being unreal, they are not
paramārthasat.35
The Buddha said: “Saṃskāras exist according to the saṃvṛti level, like
illusions or some magical being, but not according to the paramārtha
level”.36
The sons of the Conqueror who are disgusted with the world observe
the nature of dharmas as [originally] unproduced and as non-entities.37
There are even some passages that strongly resemble the classical Mādhyamika
argument for the śūnyatā of all dharmas – their dependent origination which
prevents the generation of mental and linguistic proliferations about them:
Because [all dharma]s are dependently arisen, their [existential] charac-

teristics are necessarily of the nature of a conceptual construction, and it
cannot be said that [this existential characteristic] can be reached
[objectively].38
Therefore, ignorance is destroyed through the knowledge of the real
state of affairs, and the real state of affairs is obtained by rightly under-
standing dependent origination.39
34 TS 140, 353: dharmāṇāṃ śūnyatādarśanam evānimittapratilābhaḥ nimittanirodhān na

punarbhavaṃ….
35 TS 153, 377: sarve saṃskṛtadharmā vipariṇāmitvāt māyā ity ucyante | māyātvād abhūtāḥ |
abhūtatvān na paramārthasantaḥ.
36 TS 153, 376: bhagavānāha-saṃskārāḥ saṃvṛtitaḥ santi māyāvat nirmitavat | na tu para
mārthataḥ.
37 TS 153, 378: ye jinaputrāḥ saṃsārāt paramanirviṇṇāḥ | [te]dharmāṇāṃ prakṛtito ’nutpā
dam ākiñcanyañ ca paśyanti.
38 TS 20, 70: pratītyasamutpannatvād dhāvalakṣaṇaṃ niyatavikalparūpaṃ nopalabhyavaca-
nam.
39 TS 127, 307: atastattvajñānādavidyā kṣīyate | pratītyasamutpādaṃ samyak prajānan
tattvajñānaṃ pratilabhate.
Those ignorant of dependent origination [conceptually] discriminate

against knowledge of a particular vedanā, etc. All heretics … are unable to
still their minds.40
Where there is a common practice [of talking] about ātman through
the [perspective] of dependent origination, this [talk] is free of men-
tal [and linguistic] proliferation. If one perceives being as empty and
dharmas as empty as well, one is then free from mental [and linguistic]
proliferation.41
In the light of the cited examples and their close affinity with the Mādhyamika
position, it is difficult to decide what exactly the point of contention between
Harivarman and the Madhyamaka is as the former sees it. My contention is, as
has already been stated, that he exhibits a total ignorance of the actual
Mādhyamika position regarding the existential status of objects of everyday
experience (including dharmas, rūpas, etc.).
Firstly, let us recall Harivarman’s initial objections (see above) regarding the
nāstika thesis that everything is non-existent. These objections are much the
same as those that appear in MMK and VV, and apparently belong to a realist
standpoint that does not necessarily solely include Nyāya realism, but also the
realism of e.g. Sarvāstivāda. However, in one passage of TS, Harivarman once
again seems to take the side of the Mādhyamika in answering a question that
smacks of his abhāvavādin or nāstika position:
(Question:) A dharma that does not exist ultimately does not exist at all.
Why is it then said that it exists according to saṃvṛti [‘empirically’]?
(Answer:) It is in agreement with all mundane [practices to say] “it exists;”
that is to say, action and retribution, or bondage and deliverance. [But]
all this is born out of illusion. Why? These pañcaskandhas are empty, illu-
sory, and because they continuously arise, they are like a flame. [But] for
the sake of the crossing over of the worldlings, [they] are spoken of as if
they would exist. If it were said that [they] do not exist [at all], then the
worldlings, being confused [frightened], would fall into nihilism…42
40 TS 170, 421: pratītyasamutpādasyājñātā vedanāviśeṣādijñānaṃ vikalpayati | sarve tīrthikā

… na cittaṃ nirodhayitum alam.
41 TS 184, 460: yat pratītyasamutpādenātmavyavahāraḥ sa niṣprapañcaḥ | yadi paśyati
sattvaṃ śūnyaṃ dharmāśca śūnyā iti sa niṣprapañcārāmaḥ.
42 TS 130, 316: (pṛ) yo dharmaḥ paramārthato nāsti sa sutarāṃ nāsti iti syāt | kena punarucy-
ate saṃvṛtito ’sti iti | (u) sarvair laukikair vyavahriyate asti iti yaduta karma karmavipāko
yadi vā bandho yadi vā mokṣa iti | ime sarve mohajāḥ | kasmāt | ime pañcaskandhāḥ śūnyā
māyopamāḥ jvālāvac ca santānenotpannatvāt | pṛthagjanānāṃ titīrṣayā astīti vacanam
368 Kardaš
What is slowly being revealed here is that the nāstikas, as Harivarman pictures
them, deny the existence of dharmas straightforwardly, even on the saṃvṛti
level. Dharmas for the nāstikas simply do not exist in any way or any sense (or
any level) whatsoever. They hence do not (and do not need to) make a distinc-
tion between saṃvṛtisatya and paramārthasatya. This is unacceptable for
Harivarman, as in this case all our activities would be paralyzed, and the grad-
ual revelation of the (original) emptiness or non-existence of dharmas would
be based on nothing. Emptiness can be reached only through the analytical
investigation of the actual content of the corresponding consciousness that
should be considered to be existent at that level.43 It cannot be reached by
simple denial of that content:44
Just as a man who is ignorant of painting, etc. says that these diverse
dharmas do not exist, in the same way, you, having not proven anything,
say “these dharmas do not exist”. They exist for he who knows [them],
they do not exist for he who does not know [them]. It is just as when one
who is blind from birth would say that black and white do not exist
because he does not see them. It is not proper [to say] that colors do not
exist just because they are not seen.45
We can now summarize in two points Harivarman’s main objections regarding

the nāstikas’ contention that all dharmas are non-existent:
anuvartate | yadi nāsti iti vadet | tadā pṛthagjanā vyāmuhya yadi vā ucchedavāde
pateyuḥ….
43 Cf. TS 152, 374: “Only tathāgatas are capable of analytical knowledge of dharmas …
tathāgatas are, on the other hand, [able to] dispose of the non-substantiality of all dhar-
mas in their nature, all of their aspects as well as their characteristics, particular and uni-
versal … the knowledge of emptiness is easy to obtain, but the analytical knowledge of
dharmas is difficult to develop” (tathāgatāḥ kevalaṃ dharmavivecanajñānasamarthāḥ …
tathāgatāḥ paraṃ sarvadharmāṇāṃ sarvākāraṃ prakṛtito naissvābhāvyaṃ viśeṣasāmāny
alakṣaṇāni sarvāṇi pratividhyanti … evaṃ śūnyatājñānaṃ sulabham dharmāṇāṃ
pravicayajñānaṃ durutpādam).
44 Cf. TS 130, 313: “One who does not analyze does not enter emptiness” (yo na vivecayati | ko
’vakrāmati śūnya[tāyā]m).
45 TS 152, 374: yathā kaścit citrāṅkanādidharmavikalpamajñātvā vadati tāni na santi iti | tathā
bhavānapi yat kim apy asādhayitvā vadati nāstīdaṃ- vastu iti | jñātus tu asti | ajñātuḥ
punar nāsti | yathā jātyandho vadati nāsti kṛṣṇamavadātaṃ vā, mayādṛṣṭatvāt | na
cādṛṣṭatvād rūpāṇi na santīti sambhavati).
1. Because the nāstikas are not in possession of the knowledge or vision of

emptiness, their negating dharmas is purely nihilistic and is based on
nothing. This vision can only be revealed and gained gradually through
the process of proper analytical (contemplative) investigation of the
actual content of the corresponding consciousness (prajñapti-, dharma-
and śūnyatācitta).46
2. The nāstikas consequently apply the qualification “non-existent” or
“empty” to all content of consciousness (prajñapti- and dharmacitta)
that, however, does not appear as non-existent or empty on these levels
of consciousness, but rather as qualified47 – and this is not admissible as
a path to pure nihilism.
How then is one to resolve the obvious puzzle of the fact that Harivarman first
borrows a fundamental Madhyamaka ontological position nearly verbatim –
including a good deal of its standard argumentation, while simultaneously
criticizing the nāstikas or abhāvavādins who promote the same Madhyamaka
prasaṅga type arguments for the fundamental non-existence or emptiness of
dharmas as we know from their works? There are, I believe, at least two solu-
tions to this puzzle. One is hypothetical, and I cannot present any reliable
evidence for it as of yet. According to this hypothetical solution, there was a
brief tendency within early Mahāyāna philosophical circles centered around
the use of Prajñāpāramitā sūtras to interpret their straightforward negation of
all dharmas literally without reliance on any interpretative device on which
such a negation might be based. That is to say, they did not employ such inter-
pretative devices as distinguishing the two truths or pointing to the emptiness
of all dharmas through a proper understanding of pratītyasamutpāda, as the
Mādhyamikas (Nāgārjuna) did. With the latter, they could share only the prac-
tice of drawing unacceptable consequences (prasaṅgavākya) from all theses
that tried to establish the existence or validity of anything, as is evident from
46 “If one says without reaching the intuition of emptiness that a being does not exist, it is a
wrong view” (yadapratilabdhaśūnyatāprajño vadati nāsti sattva iti | iyaṃ mithyādṛṣṭiḥ, TS
141, 356, and also TS 141, 357: “If one who has not reached the reality of emptiness unfortu-
nately harbors the view that nothing exists, he falls into a wrong view, namely nihilism”
(apratilabdhaśūnyatāt attvo nāsti kiñcanetidṛṣṭyā durgatau patati yadutocchedadṛṣṭi[rūpa]
mithyādṛṣṭiḥ).
47 This was hinted at by Harivarman when he stated that “there is no mundane emptiness
(laukikī śūnyatā)” (nāsti tu laukikī śūnyatā, TS 191, 505. I hence believe Priestley to be
incorrect when he states that Harivarman distinguishes between “conceptual emptiness
and phenomenal emptiness” (1970: 34) for there is not, so to speak, any “mark” of empti-
ness on that level of consciousness.
370 Kardaš
Harivarman’s account. But this practice, if not backed by the conceptual struc-
ture of the two truths (MMK 18.8) and by the equation of pratītyasamutpāda
with śūnyatā (MMK 24.18), inevitably leads to nihilism.48 It was Nāgārjuna then
who performed this task, and so Harivarman was actually criticizing this philo-
sophical tendency within Mahāyāna philosophy and not the Mādhyamikas.
The second solution is, on the other hand, more factually based. In the first
part of this chapter, I attempted to provide some evidence of the fact that
Āryadeva’s works, and particularly ŚŚ, were probably the main source for
Harivarman’s presentation of the nāstikas’ or abhāvavādins’ position. If we
carefully read Āryadeva’s work, we will find that it strangely lacks the most
fundamental Mādhyamika argument and conceptual (or hermeneutical)
framework, those being the equation of pratītyasamutpāda with śūnyatā and
the theory of the two truths. These two are, as already mentioned, the only
“positive tools” preventing the degradation of the Mādhyamika position into
pure nihilism. CŚŚ is wholly and solely dedicated to the utter deconstruction of
all concepts and dharmas in terms of their referent or existence, leaving no
room for an affirmation of their provisional or conventional status, much in
the same spirit of Harivarman’s nāstikas. On the other hand, Nāgārjuna and
Candrakīrti especially, were careful enough to qualify their statements about
the non-existence of dharmas with qualifications such as svabhāvatas (“in
terms of their own being” or “substantially”).49
I conclude, therefore, that Āryadeva’s highly negativistic approach in
defending the Mādhyamika position significantly or even crucially contrib-
uted to the subsequent misunderstanding and improper appreciation of the
Mādhyamika standpoint regarding the non-ultimate, i.e. saṃvṛti nature of
dharmas reflected both in Harivarman’s work and in later Abhidharma śāstras
such as AKBh, *Nyāyānusāra and Abhidharmadīpa, which unanimously qual-
ify the apparently Mādhyamika position as nihilistic.
48 Arguing against the equation of utter non-existence (abhāva) with emptiness in the con-
text of the Mādhyamika position, Candrakīrti rhetorically asks: “And who criticizes us in
the manner just described? It is he who does not know the infallible distinction between
the two truths as taught in the words of the Buddha” (kaścāsmākaṃ yathoktam
upālambhaṃ karoti? yo bhagavatpravacanopadiṣṭāviparītasatyadvayavibhāgaṃ na jānāti,
kevalaṃ granthamātrādhyayanapara eva iti (PP 492.28–29).
49 It, of course, goes without saying that Āryadeva was familiar with his Master’s theory of
the two truths, as well as with the “diamond argument” for emptiness, that being the
conditional existence of dharmas (pratītyasamutpāda). However, even in his major work,
CŚŚ, these notions rarely occur and are never mentioned straightforwardly (cf. CŚŚ VIII,
19–20; IX, 2–3, 25; XIV, 23).
4 The General Mādhyamika Answer to the Kind of Criticism

Harivarman Exposes in His Treatise
It was the second generation of Mādhyamika philosophers, beginning with

Buddhapālita (vṛtti to MMK 18.7), who responded to the accusation that they
were nāstikas by attempting to clearly differentiate their position from that of
the nāstikas.
Buddhapālita highlights the following major points of divergence between
the Mādhyamikas and nāstikas (Cf. Lindtner: 1981: 206–207): The latter fail to
discern the meaning of emptiness (śūnyatārtha). The nāstikas’ statement that
the world does not exist is purely verbal and “doxic,” and hence fettered with
ignorance, while the same statement in the case of the Mādhyamikas is backed
by cognition and insight into the emptiness of all entities. Finally, the
Mādhyamikas’ argument for the non-existence of dharmas does not transgress
the limits of language imposed by linguistic conventions, while the nāstikas’
baseless statements about existence and non-existence violate these conven-
tions. Mādhyamikas regard existence and non-existence as noema qua image
(pratibimba) since they arise dependently (pratītyasamutpanna).
Expanding on Buddhapālita’s commentary to MMK 18.7, Candrakīrti first
states that, unlike the Mādhyamikas who are “proponents of pratītyasam
utpāda” (pratītyasamutpādavādino hi mādhyamikāḥ) and describe the
entire world as essenceless (niḥsvabhāva) because it comes into being
dependent upon causes and conditions (hetupratyayān prāpya pratītya samut
pannatvāt, PP 368.27–28), the nāstikas are essentially (dogmatic) “realists”
(svarūpavādins) because they do not understand that dharmas are empty of
their own being (bhāvasvabhāvaśūnyatva) on account of their interdepen-
dence (pratītyasamutpannatvād, PP ibid.). Furthermore, unlike the nāstikas,
the Mādhyamikas accept the existence of dharmas conventionally (saṃvṛtyā
mādhyamikair astitvenābhyupagamān na tulyatā, PP 369.4). Candrakīrti con-
cludes this discussion by stating that the nature of dharmas is properly known
by the Mādhyamikas who are able to discuss and understand it (yathāvad
viditavastusvarūpāṇāṃ mādhyamikānāṃ bruvatāmavagacchatāṃ ca) and
that, although there is no difference between the two parties regarding the real
nature of dharmas (vastusvarūpābhede), there is no similarity between the
Mādhyamikas who know it and can discuss it, and the nāstikas who do not
properly know the nature of dharmas. (yathāvad aviditavastusvarūpair nāsti
kaiḥ saha jñānābhidhānayor na asti sāmyam, PP 369.9–10).
In other words, the Mādhyamikas, unlike the nāstikas, know how to acquire
insight into emptiness, how to philosophically ground and conduct the analy-
sis leading to that insight, and how to linguistically elaborate upon such an
372 Kardaš
analysis without transgressing conventional limits imposed by language. All of

these steps are pregnantly expressed in MMK 24.18 (pratītyasamutpāda=śūnya-
tā=prajñaptir upādāya), and they can be undertaken only if the notion of the
two truths is accepted. Otherwise, if it is nonreferentially said that dharmas do
not exist, then this very statement is prima facie devoid of any meaning simply
because there is nothing to be qualified by any predicate, “does not exist”
included.
5 Conclusion
We are still far from any promising clue as to who Harivarman’s nāstikas actu-
ally are. As presented by him, they are definitely Mādhyamika-like Buddhists
who abundantly employ prasaṅga argumentation for the non-existence of
dharmas, and this type of argumentation backed by nothing seems to be dis-
carded by Harivarman. If Harivarman’s attack on his nāstikas is in fact an
attack on the Mādhyamikas whose fundamental position he misunderstood
due to Āryadeva, then how is one to explain that, in other parts of the text, he
exhibits his knowledge of the conceptual structure of the two truths much in
the same line as the Mādhyamikas, including the understanding of pratītyasam
utpāda as a door to the śūnyatā-insight? Shōryū Katsura (1974: 184) believes
that “the fact that nāstikas’ attack is mainly directed at the Vaiśeṣikas and
Sāṃkhyas … seems to suggest that Harivarman is here intentionally copying
some arguments set by Nāgārjuna or his followers”. However, the nāstikas’
attack is in fact mainly directed at Buddhist (Abhidharma) doctrines, and only
in a few cases at these non-Buddhist schools (refutation of satkārya- and
asatkāryavāda, and, to some extent, refutation of parts and wholes). Our con-
fusion is further increased by the fact that Buddhapālita’s and Candrakīrti’s
nāstikas are probably not Buddhists, but are more likely Cārvākas/Lokāyatas or
Ajñānikas (agnostics), since these philosophical schools or trends (and only
they) are generally known in later Buddhist works as nāstikas.
Nevertheless, as stated in the first solution to our puzzle, could there have
been quasi-Mādhyamika Buddhist philosophers prior to or contemporaneous
with the early Mādhyamikas proper who propounded the non-existence of
dharmas and delighted only in reductio ad absurdum arguments to this end?
Abbrevations
AK Abhidharmakośabhāṣyam of Vasubandhu (ed. Pradhan, 1975)

AKBh Abhidharmakośabhāṣyam of Vasubandhu (ed. Pradhan, 1975)
CŚŚ Catuḥśatakaśāstra (ed. Vaidya, 1923)

MMK Madhyamakaśāstra of Nāgārjuna with the Commentary: Prasannapadā by
Candrakīrti (ed. Vaidya, 1960)
PP Madhyamakaśāstra of Nāgārjuna with the Commentary: Prasannapadā by
Candrakīrti (ed. Vaidya, 1960)
ŚŚ Pre-Diṅnaga Buddhist Texts on Logic from Chinese Sources (ed. Tucci, 1929)
VV The Dialectical Method of Nāgārjuna: Vigrahavyāvartanī (ed. Bhattacharya et
al., 2002)
Bibliography
Bhattacharya, Kamaleswar et al. (2002). The Dialectical Method of Nāgārjuna:

Vigrahavyāvartanī (Text and Translation). Delhi: Motilal Banarsidass.
Cox, Collett (1995). Disputed Dharmas. Early Buddhist Theories on Existence: An Annotated
Translation of the Section on Factors Dissociated from Thought from Saṅghabhadra’s
Nyāyānusāra. Tokyo: The International Institute for Buddhist Studies.
Dhammajoti Bhikkhu, KL (2007). Abhidharma Doctrines and Controversies on Perception.
Hong Kong: The University of Hong Kong, Centre of Buddhist Studies.
Frauwallner, Erich (1956). Die Philosophie des Buddhismus. (Philosophische Studien
Texte 2). Berlin: Akademie-Verlag.
Jaini, Padmanabh S. (1977). Introduction to Abhidharmadīpa with Vibhāṣāprabhāvṛtti.
(2nd ed.). Patna: Kashi Prasad Jayaswal Research Institute.
Katsura, Shōryū (1974). A Study of Harivarman’s Tattvasiddhi. Ph.D.
La Vallee Poussin, Louis de (1937). “Documents d’Abhidharma: La Controverse du
Temps”, Melanges chinois et bouddhiques 5, pp. 7–158.
Lang, Karen C. (1983). Āryadeva on the Bodhisattva’s Cultivation of Merit and Knowledge.
Ph.D.
Lindtner, Christian (1981). “Buddhapālita on Emptiness”, Journal of Indian Philosophy
23, pp. 187–217.
Lindtner, Christian (1982): Nagarjuniana. Studies in the Writings and Philosophy of
Nāgārjuna, Copenhagen: Akademisk Forlag
Priestley, Leonard C.D.C. (1970). “Emptiness in the Satyasiddhi”, Journal of Indian
Philosophy 1, pp. 30–39.
Sastri, Aiyaswami N. (ed.) (1975). Harivarman: Satyasiddhiśāstra. Vol. I. Baroda: Oriental
Institute.
374 Kardaš
Tucci, Giuseppe (ed.) (1929). Pre-Diṅnaga Buddhist Texts on Logic from Chinese Sources
(Part I: Texts. pp. 1–89. English translation of ŚŚ together with Vasu’s Commentary).
Baroda: Oriental Institute.
Vaidya, Paraśurāma Lakṣmaṇa (ed.) (1923). Catuḥśatakaśāstra: Études sur Āryadeva et
son Catuḥśataka. chapitres VII–XVI. Paris: Librairie Orientaliste Paul Geuthner.
Vaidya, Paraśurāma Lakṣmaṇa (ed.) (1960). Madhyamakaśāstra of Nāgārjuna with the
Commentary: Prasannapadā by Candrakīrti. Darbhanga: The Mithila Institute of post-
graduate studies and research in Sanskrit learning.
Wogihara, Unrai (ed.) (1932–1936). Sphutārthā Abhidharmakośavyākhyā by Yaśomitra.
2 vols. Tokyo: Publishing Association of Abhidharmakośavyākhyā.
Svalakṣaṇa (Particular) and Sāmānyalakṣaṇa (Universal) 375
Chapter 12
Svalakṣaṇa (Particular) and Sāmānyalakṣaṇa

(Universal) in Abhidharma and Chinese Yogācāra
Buddhism
Chen-kuo Lin
1 Introduction
I am sitting in the study room and experiencing the surrounding things, such
as table, pen, music, and bell sound from the elementary school not far from
the apartment where I live.1 There are also many thoughts bubbling up even
without notice. But I am still able to say that “I see a pen,” “I am reading Martin
Heidegger’s Being and Time,” “I have heard the bell ringing,” and so on. The
experience I just described contains both cognitive and psychological tones.
Sometimes I am in good mood when I am sensing and thinking. Sometimes
I am blue. If I go further to analyze my experience of, say, a pot of tulip on the
table, I see the color of cup shaped pedals, even though I am not able to pre-
cisely identify a correct term for that shade of color. But I do see something. In
Buddhist terms, that “something” is called “particular” (svalakṣaṇa). If I go on
to say, “The color of this tulip is purple,” then since the words “color” and “pur-
ple” in the statement are applicable to the qualities that are exemplified in
many individual entities, they are called “universal” (sāmānyalakṣaṇa).2
Imagine another scenario. If the Buddha, an enlightened being, sits in the
same room, does he have the same cognitive and psychological experience as
mine just described above? As claimed in many Yogācāra texts, the enlight-
ened cognition is pure perception only. As for the Buddha himself, he does not
need to engage in reasoning at the moment of enlightenment. All he experi-
ences is nothing but perception in the momentary flux. The difference of the
enlightened experience from the ordinary experience is that he does not con-
ceptualize the momentary flux of perception within the dichotomy of ‘”subject”
and “object,” “I” and “mine”.
1 The author of this article is grateful for Ernest Brewster’s proofreading.

2 “Sāmānyalakṣaṇa” can literally be rendered by “general characteristic,” and “svalakṣaṇa” by
“individual characteristic”.

376 Chen-kuo Lin
However, the puzzle has not been completely solved. Is it true that the
enlightened one does not need to employ concepts and reasoning for commu-
nication after enlightenment? Or, does he only experience reality in perception
without any aid of conception and reasoning? These puzzles were discussed in
Kuiji’s 窺基 (632–682) Cheng weishi lun shuji 成唯識論述記 (Commentary on
the Treatise on the Establishment of Consciousness-Only), a foundational trea-
tise in the Chinese Yogācāra school in the seventh century.3 Concerning this
question, I will explore in this paper the different theories of cognition with
regard to svalakṣaṇa (particular) and sāmānyalakṣaṇa (universal) in the
Abhidharma and Yogācāra contexts. The main question is, what is the correct
cognition, the cognition of universal or the cognition of particular, which can
be employed to eliminate the psycho-cognitive hindrances for attaining final
liberation? This is the basic question raised by both the Abhidharma and
Yogācāra Buddhists.
As the Abhidharma literature shows, the above question is raised in the con-
text of mental and cognitive cultivation. The practitioner is guided to realize
the truth of impermanence, suffering, and selflessness, which will in turn help
free one from ignorance and contamination, through following the various
stages of meditation. “Impermanence,” “suffering” and “no-self” are considered
to be universals for the reason that they are universal truth of phenomena,
while the individual objects in the categories of five aggregates, twelve sense-
spheres, and eighteen elements are identified as particulars. When the question
is raised again in the same context, asking which kind of cognition, particular
or universal, leads to the final elimination of contaminations for attaining lib-
eration, the Abhidharma answers that it is the cognition of universal which
leads to liberation. This conviction is firmly stated in the Abhidharmama
hāvibhāṣa: “Only in the path which takes universal as the object, one is able to
eliminate the contaminations”.4 However, there are disagreements regarding
this issue among Yogācāra commentators.
Another important issue is about the ontology of universals. In both Western
and Indian (including Buddhist) philosophy, generally speaking, there are
three positions with regard to the ontological status of universals: (1) meta-
physical realism, which holds the view that universal is real in the sense that it
exists independently of mind; (2) nominalism, claiming that universal is not
real, while particular is real; (3) conceptualism, which argues that while uni-
3 The same passages in Kuiji’s Cheng weishi lun shuji is also found in his Dasheng fayuan yilin
zhang 大乘法苑義林章 and Huizhao’s Dasheng fayuan yilin zhang buque 大乘法苑義林
章補闕, which will be examined in the following of this article.
4 T.27.1545: 820a1–2: “唯共相境道能斷煩惱”.
versal is not ultimately real, it does exist in the mind in the form of concept.
Regarding the ontological status of universal, most of the Buddhists stand for
nominalism and reject metaphysical realism. However, when we consider the
various ontological stances taken by Sarvāstivāda, Sautrāntika, Madhyamaka,
and Yogācāra, the label of “nominalism” cannot be satisfactorily applied to all
Buddhist schools without further clarification. Could conceptualism be an
option for Yogācāra? In Yogācāra commentaries, we find some discussion rele-
vant to this issue. An important note with regard to the ontology of universals
is that those theories are not resulted from the metaphysical speculation only.
On the contrary, the debates on the ontological status of universals are mainly
concerned with the soteriological consequences. For underlying the various
theories of universal there is a fundamental belief that only truth will set free
those who are entrapped in ignorance and contamination: “Only mental atten-
tion to the reality (tattvamanaskāra) is capable of cutting off the afflictions”.5
2 Three Theories of Particular and Universal
In response to the above questions, firstly, we come to Kuiji who cited three
theories from the Buddhabhūmyupadeśa:6 (1) The first theory holds that all
characteristics of objects, including particular and universal, are cognized by
direct perception (pratyakṣa) in the concentrated state of mind, whereas
according to Buddhist logic (hetuvidyā) (i.e. Nyāyamukha and Nyāyapraveśa)
– particular and universal are experienced in the non-concentrated state of
mind by direct perception and inference respectively.7 (2) The second theory
holds that only particular is experienced in the concentrated state of mind.
5 T.27.1545: 819c28–29: “唯真實作意能斷煩惱”.

6 T.26.1530: 318b16-b15.
7 T.43.1830: 584b8–12: “[Answer:] There are three theories [in regard to this question]. (1) Some
hold that two instruments of valid cognition (pramāṇa) [i.e. pratyakṣa and anumāṇa] appear
in the non-concentrated state of mind; they are established in accordance to two character-
istics [i.e. svalakṣaṇa and sāmānyalakṣaṇa]. The mind in the concentrated state is excluded
[in this theory]. If all characteristics are cognized in the concentrated state, they are cognized
by direct perception. According to this theory, the cognition of the combined support-object
(samastālambanajñāna) is also classified as direct perception, which is definitely capable of
eliminating the afflictions. The concentrated mind is said to be direct perception simply be-
cause [the mind] in the concentrated state illuminates the sāmānyalakṣaṇa itself (svarūpa)”.
(“ 彼有三說。有義，二量是散心位，依二相立。不說定心。若在定心，緣一切相
皆現量攝。由此總緣智亦現量攝，斷惑無失。即由定照共相自體故，說定心為現
量也”).
378 Chen-kuo Lin
However, in the progress of meditative cultivation, universal is used as an expe-

dient means (upāya) that helps the practitioner to experience the truths in the
form of particulars. Hence, one can provisionally say that the universal is also
cognized in the concentrated state of mind. According to this theory, the cog-
nition of universal is taken as a crucial step to eliminate the afflictions. In the
final analysis, however, one eliminates the afflictions through the cognition of
particular.8 (3) The last theory attempts to elaborate the positions favored by
both Buddhist logicians and the Buddhabhūmyupadeśa. That is, since the mind
in the concentrated state is free of conception, all of the objects in this state are
particulars, which are perceived by pratyakṣa only. Even if a practitioner is able
to cognize impermanence, suffering, and so forth, in the concentrated state,
since these universals are experienced as something within the individual
object, they are also taken as particular because they are not separate from the
individual object. Accordingly, truth/thusness (tathatā) should be also viewed
as particular only, instead of being taken as universal.9 This theory is opposed
8 T.43.1830: 584b12–18: “(2) Some hold that in the concentrated state of mind only the par-
ticular characteristic is cognized. However, it is introduced by means (upāya) of the uni-
versal characteristic. What is cognized is the truth/principle manifested in the various
characteristics of universal. It is provisionally called ‘cognition of universal characteristic’.
It is called ‘cognition of particular characteristic’ when it is not so. According to this rea-
son, some name ‘suchness’ (tathatā), which is also named ‘emptiness’ (śūnyatā) and ‘self-
lessness’ (nairātmya), as the universal characteristic of existents. Some say that ‘suchness,’
which is manifestation of two kinds of emptiness, is not a universal characteristic.
According to this theory, it is a provisional theory held by the Abhidharma to claim that
the cognition of universal characteristics is capable of eliminating the afflictions. As a
matter of fact, it is the discerning of particular characteristics that is capable of elimina-
tion”. (“第二說，有義，定心唯緣自相，然由共相方便所引，緣諸共相所顯理
者，就方便說名知共相，不如是者名知自相。由此道理，或說真如名空無我
諸法共相，或說真如二空所顯非是共相。由此義故，對法等說緣共相智能斷
惑者依方便說，實自相觀方能斷之”).
9 T.43.1830: 584b18–29: “(3) “As regards the true meaning [of dharmas] explained here, the
treatises on logic (hetuvidyā) offer a different account of svalakṣaṇa and sāmānyalakṣaṇa.
They claim that the real objects (tattvārtha) of all dharmas are named svalakṣaṇa for the
reason that the svalakṣaṇa and sāmānyalakṣaṇa of dharmas exist separately by them-
selves and hence are not shared with each other. When the conceptual mind establishes
the classes (jāti) of signifier and signified which is applied to dharmas in the similar way
that the thread is used to string flowers together, they [class, etc.,] are called “sāmānyala
kṣaṇa”. They are conceptual construct in the non-concentrated state. They are the object
of inference. All minds in the concentrated state, which are free from conceptualization,
are called “pratyakṣa”. Even if [in the concentrated state,] suffering, impermanence, and
so forth, of existents are cognized, they exist in each individual existent. Hence they are
called “svalakṣaṇa”. Although suchness (tathatā) is disclosed through sāmānyalakṣaṇa, it
to the Abhidharma teaching; the latter maintains that only the cognition of
universal is capable of eliminating the contaminations before realizing the
final enlightenment.
After having cited three theories of particular and universal, Kuiji asserts his
own view which disagrees with the Abhidharma theory. He says:
According to the above analysis, some Abhidharma [texts] hold the view
that the cognition which takes universal as the object is capable of
destroying the contaminations.10 The universal is thus named for the rea-
son that the conceptual mind uses class (jāti) to string the various
existents on [the level of] truth/thusness (tathatā). Some hold that uni-
versal is named for the meaning (artha) of existents which are grounded
in tathatā. The cognition of universal is thus named for that which takes
this [i.e. universal] as the object. As a matter of fact, the existence of
tathatā is real by nature, so it is not universal. In reality, the cognition of
particular [i.e. svalakṣaṇa] is capable of destroying the contaminations.
As explained above, [it is because] universal is conceptually designated.11
is not sāmānyalakṣaṇa, because it is the real nature of existents, and because it possesses
its own characteristic (lakṣaṇa). It cannot be named “sāmānyalakṣaṇa” simply because it
is neither identical nor different from all existents. It is because svalakṣaṇa is also neither
identical nor different from all sāmānyalakṣaṇas”. As a result, the treatises [on logic] hold
a view distinct from this (i.e. Buddhabhūmyupadeśa – T.26.1530.318b4–14)”. (“第三說，
如實義者，彼因明論立自共相，與此少異。彼說一切法上實義皆名自相。以
諸法上自相共相。各附已體不共他故。若分別心立一種類能詮所詮通在諸法
如縷貫華名為共相。此要散心分別假立是比量境。一切定心離此分別皆名
現量。雖緣諸法苦．無常等。亦一一法各別有故名為自相。真如雖是共相所
顯。以是諸法自實性故。自有相故。亦非共相。不可以其與一切法不一不異
即名共相。自相亦與一切共相不一異故。是故彼論與此不同”).
10 In Kuiji’s Cheng weishi lun biechao 成唯識論別抄, “Abhidharma” is identified as referring
to Asaṅga’s Abhidharmasamuccaya and Sthiramati’s Abhidharmasamuccayabhāṣyā. Cf.,
Sthiramati’s Abhidharmasamuccayabhāṣyā: T.31.1606: 727a10–12: “Question: By what kind
of intentional attention can one eliminate [the contaminations]? Answer: One can elimi-
nate the contaminations through the intentional attention to sāmānya by which all exis-
tents are realized to be the lack of self-nature. By ‘intentional attention to sāmānya,’ it
means that one applies the intentional attention to the universals of all existents”. (“問：
何等作意能斷耶？答：總緣作意觀一切法皆無我性，能斷煩惱。總緣作意
者，謂令緣一切法共相行作意”).
11 T.43.1830: 584b29-c5: “由此義故，對法等說緣共相智能斷惑者，依分別心於一種
類，真如之上，通在諸法，說名共相。或真如體諸法皆有義，名共相。緣此
之智，名共相智。論實，真如法實性故，非是共相。據實而言，即別相智能
斷惑也。共相假立，已如前辨”.
380 Chen-kuo Lin
Kuiji distinguishes two types of universal: (1) concepts, such as class and spe-
cies, which are superimposed upon the real existents by the conceptual mind;
(2) the meaning/property of tathatā, which also describes sixteen modes
(ākāras) of the four noble truths. Both types of universal are the cognitive
object of inference. In the case of tathatā, tathatā in itself is viewed as sub-
stance, while “suffering,” “impermanence,” and so on, are viewed as property/
meaning (artha). For Kuiji, tathatā itself cannot be conceptually cognized.
Hence it cannot be taken as universal. In short, as regards the issue whether the
cognition of universal or particular leads to the elimination of afflictions, Kuiji
clearly disagrees with the Abhidharma view. The main reason for his disagree-
ment is that the Abhidharma confuses the two aspects of tathatā: tathatā in
itself (ti 體, *svarūpa) and tathatā as property (yi 義, artha). Tathatā in itself is
the ultimate reality which can be experienced in direct perception. However,
one has to cognize “tathatā as property” before one directly perceives tathatā
in itself which leads to the final enlightenment.
In the following I will examine how the Chinese Yogācāra Buddhists during
the seventh and the eighth centuries are indebted to the Abhidharma theory of
svalakṣaṇa and sāmānyalakṣaṇa, even though their ontological positions are
different.12 I will also point out that the continuity between Abhidharma and
Yogācāra is grounded in the same matrix of cultivation and theorization. For
both Abhidharma and Yogācāra, the ontology of particular and universal can-
not obtain its full meaning unless it is understood within the context of
meditative practice. In the following section, I will first examine the crucial
passages in Kuiji’s Cheng weishi lun shuji to see how critically the Chinese
Yogācāra scholars adapted the Abhidharma doctrine to a more complex
framework.
3 Particular and Universal in the Abhidharma and Yogācāra Paths
It should be noted that the above three theories cited from Kuiji’s Cheng weishi
lun shuji appear in the commentary on the path of cultivation. Adopting the
Abhidharma model, which is believed to have been accepted and practiced in
the entire community of Buddhists, the path of cultivation in the Cheng weishi
12 Part of materials in this paper has been presented in another paper, “How to attain
enlightenment through cognition of particulars and universals – Huizhao on svalakṣaṇa
and sāmānyalakṣaṇa,” at the International Workshop on “The Ontology of Asian Philoso-
phy – Perspectives from Buddhist Studies and Analytic Philosophy”, 13–14 April, 2013, Ryu-
koku University and Kyoto University, Kyoto.
lun is divided into five stages: (1) the stage of equipment (sambhāra, ziliang wei
資糧位), (2) the stage of application (prayoga, jiaxing wei 加行位), (3) the stage
of penetration (nirvedha, tongda 通達), which is also called “the stage of seeing
(darśana, jiandao 見道)” in the Abhidharma literature, (4) the stage of cultiva-
tion (bhāvanā, xiuxi wei 修習位), and (5) the stage of completion (niṣṭhā,
jiujing wei 究竟位).13 In each stage, several sub-stages of cultivation are also
listed. Roughly speaking, the whole system of cultivation is designed to achieve
the final liberation by eliminating psychological and cognitive hindrances,
including the contamination-hindrances (kleśāvaraṇa, fannao zhang 煩惱障)
and the hindrances of cognitive objects (jñeyāvaraṇa, suo zhi zhang 所知障).14
According to the Yogācāra doctrine, the hindrance which should be gradually
eliminated includes those in forms of mental phenomena (samudācāra, xian
xing 現行 (lit., “presentation” – “present manner/feature”), seeds (bīja, zhongzi
種子), and habitual residues (vāsanā, xiqi 習氣). The mental phenomena are
the form that should be tamed first. However, a more crucial step should be
further taken to eliminate the seeds and habitual forces which are contained in
the store-consciousness. For the Yogācāra, the elimination of the contamina-
tion-hindrances leads to the realization of nirvāṇa, while the elimination of
cognitive hindrances leads to the realization of enlightenment (bodhi).
Although a long course of cultivation is required to eliminate the contamina-
tion-hindrances, the elimination of cognitive hindrance is rather considered
more foundational. It is precisely in the context of eliminating the cognitive
hindrances that the cognition of particular and universal is brought into the
agenda of practice, for both particular and universal are the aspects of objects
which should be correctly cognized.
In the Cheng weishi lun, the first four stages of practice are further divided
into the mundane path and the sagely path. The mundane path consists of
the first two stages, the stage of equipment and the stage of application, while
the sagely path consists of the latter two stages, the stage of penetration and
the stage of cultivation. Since the entire system of cultivation is rather com-
plex, it is not practical to lay it out in excruciating details. At this juncture we
merely need to point out that, as for the discernment of particular and univer-
sal, the practice at the stage of application is most crucial before one attains
the insight of emptiness at the stage of penetration. The stage of application is
further divided into four sub-stages: “warmth” (ūṣmagata, meaning lit., “gone
13 For a brief account of the Abhidharma path, see Dhammajoti, 2007: 564–612.
14 Cf., Muller, 2004: 207–235. A brief version of interpretation can be viewed at the entry
“two hindrances” by the same author at Digital Dictionary of Buddhism (<http://www.
buddhism-dict.net>).
382 Chen-kuo Lin
to [the state] of warmth [ūṣman]”), “summit” (mūrdhan), “receptivity” (kṣānti)

and “the worldly supreme dharmas” (laukikāgradharma). As we will see in
the following, the Abhidharma analysis of particular and universal is intro-
duced at the stage of “warmth”. In the Cheng weishi lun, the similar analysis
is also found in two contexts. One is found in the clarification of the notion
of “post-enlightened cognition” (pṛṣṭhalabdhajñāna, hou de zhi 後得智), which
claims that the post-enlightened cognition is able to differentiate particular
and universal for communication through discerning the difference of sentient
beings’ inborn nature.15 The other instance is found in the exposition of one
of the Buddha’s four transformative insights, “the insight of skillful investiga-
tion” (pratyavekṣajñāna, miao guancha zhi 妙觀察智), which means that the
Buddha’s mind of thinking is capable of skillfully discerning both the partic-
ular aspect and universal aspect of objects.16 These instances show that the
epistemological issue of particular and universal in Yogācāra philosophy is also
raised within the context of meditative practice. For Yogācāra, however, they
are more concerned with the Buddha’s experience of cognition than ordinary
cognition. The best textual evidence is also found in the Buddhabhūmyupadeśa.
On the side of the Abhidharma, a yogin is instructed to practice the discern-
ing of particular and universal at the stage of application (prayoga) in order to
acquire three types of wisdom: (1) wisdom acquired through hearing, (2) wis-
dom acquired through thinking, (3) and wisdom acquired through cultivation.
The cognition of particular and universal culminates in the sub-stage of culti-
vation, which is described as follows:
1. Learn the essentials of teaching either from a master or by self-learning.

The essentials of teaching can be classified by different categories, such
as (i) eighteen elements, (ii) twelve sense-spheres, (iii) five aggregates,
(iv) four bases of mindfulness, and (v) four noble truths, which serve as
the taxonomical lists for classifying the objects of discernment.
2. When a yogin picks up a taxonomical category for discerning, he should
learn three constituents: name, particular, and universal. First, name is
employed to identify the object of discernment, such as “This is a visual
object,” “This is an audible object,” etc. According to this method of
classification, all objects of discernment should be exhausted by the
category. Second, “svalakṣaṇa” is used to identify the nature of each
category, e.g. the nature of the eye-faculty defined as “purity of matter
which serves as the basis of visual consciousness (cakṣurvijñānāśrayo
15 T.31.1585: 50b21–23: “此智分別諸法自共相等，觀諸有情根性差別而為說故”.

16 T.31.1585: 56a21–22.
rūpaprasādaḥ).17 Third, the practitioner goes on to discern the universal

characteristics (sāmānyalakṣaṇa) of objects, which consist of sixteen
modes of cognitive activities (ākāras): (1) suffering (duḥkha), (2)
impermanence (anitya), (3) emptiness (śūnya), (4) no-self (anātmaka),
(5) cause [of suffering] (hetu), (6) gathering (samudaya), (7) arising
(prabhava), (8) conditions (pratyaya), (9) extinction (nirodha), (10)
serenity (śānta), (11) sublimity (praṇīta), (12) solicitude (niḥsaraṇa), (13)
path (mārga), (14) accordance (nyāya), (15) practice (pratipatti), and (16)
transcendence (nairyāṇika).18
3. The whole procedure of discernment consists of three stages: learning,
thinking, and cultivation. The same procedure of practice is also applied
to three realms: the desire realm, the form realm, and the formless
realm. The objects in three realms can be discerned either separately or
simultaneously. The manual of meditation concludes: “When he
discerns the four noble truths in this way, [his discerning] is like some-
one who sees the images of body through the covering of silk. The
wisdom acquired through hearing can be perfected only if one follows
the above instruction. It follows that the wisdom acquired through
thinking occurs. As soon as the wisdom acquired through thinking is
perfected, the wisdom acquired through cultivation occurs. [The first
stage] is called ‘warmth’. It follows ‘summit,’ ‘receptivity,’ ‘the worldly
supreme dharmas,’ ‘path of seeing,’ ‘path of cultivation,’ and ‘path of no
learning’ sequentially. Following the above paths the wholesome
capacity is thus fulfilled”.19
17 Abhidharmakośabhāṣya of Vasubandhu ed. by Pradhan, 1967, 5, 24–6, 6, Chapter. I v.9cd,

cited from Bauddha Kośa, eds., Akira Saito, accessed at <http://www.l.u-tokyo.ac.jp/~b_
kosha/html/index_75dharma.html#, 6/29/2013>.
18 According to KL Dhammajoti’s interpretation, the sixteen ākāras of the four noble truths
are not “images or ‘aspects’ of the objects, but are in the active sense of the mental func-
tion of understanding”. Those ākāras are also identified as sāmānyalakṣaṇa, which are
“the universal principles of all dharma-s intuited by spiritual insight pertaining to the
absolute truth, not universals abstractly constructed by the mind as in the case of mental
inference”. See Dhammajoti, 2007: 352 for conceptual clarification; ākāra can be referring
to subjective mental function, while the objective principle of an existent is called
sāmānyalakṣaṇa.
19 T.27.1545: 34c20–26: “如是觀察四聖諦時，猶如隔絹觀諸色像，齊此修習聞所成
慧，方得圓滿。依此發生思所成慧，修圓滿已，次復發生修所成慧，即名為
煖。煖次生頂，頂次生忍，忍次生於世第一法，世第一法次生見道，見道次
生修道，修道次生無學道，如是次第，善根滿足”.
384 Chen-kuo Lin
Commenting on the path of seeing, where two hindrances are said to be elimi-
nated through discerning the sixteen modes or ākāras of the four noble truths,
Kuiji followed the Abhidharma path in pursuing the issue of the cognition of
particular and universal. The questions are raised again: In the path of cultiva-
tion, when one meditates on four noble truths as the object of cognition
(ālambana), should he cognize four noble truths as a whole? Or, should he
cognize four noble truths individually in each object? Which way of discerning
is correct practice? If the four noble truths are cognized as a whole, then the
object of cognition should be none other than the universal. Here the question
arises again: Is the cognition of universal capable of eliminating the cognitive
hindrance? Since the cognition of universal relies on inference (anumāṇa),
which is conceptual activity (vikalpa), the same question can be posed as such:
Can one attain enlightenment through inference? The question can be also
stated conversely: Does one realize enlightenment through direct perception
(pratyakṣa)?20 These questions are considered by Chinese Yogācāra scholars to
be essential in their epistemological and soteriological project.
4 Particular and Universal in the Abhidharma Course of Cultivation
When we come to the ontological problem of particular and universal, as

explained above, we found that in the Abhidharma literature the issue does
not stand alone. Rather, it is treated in a phenomenological way. That is, there
is no fixed ontology of particular and universal without referring to the variable
mental states which are further corresponding to the various stages of cultiva-
tion. For the Abhidharma, particular and universal are the objects in cognition,
while the state of cognition varies in the course of meditation. A line of
20 T.43.1830: 584a29-b4: “Question: As to the term ‘elimination [of affliction],’ is the cogni-
tion (jñāna) which takes the whole [of five aggregates, four noble truths, etc.] as a sup-
port-object (samastālambana) capable of eliminating [affliction]? Or, is the cognition
which takes the individual [item of five aggregates, four noble truths, etc.] as a support-
object (vyavacchinnālambana) capable of eliminating [affliction]? What is wrong with
either [theory]? [Answer:] The cognition that takes the whole as a support-object is not
the cognition of particular characteristics (svalakṣaṇa). How could the cognition of infer-
ence, which takes sāmānyalakṣaṇa [as the support-object], be capable of eliminating
affliction? If one holds that the cognition of the individual support-object is capable of
elimination, this will contradict the statements in fascicle fifty-nine and the theory in
fascicle seven of the Abhidharma [treatise]”. (“問：此言斷者，為總緣智能斷？為別
緣智能斷？此有何失？總緣之智非自相智，如何共相比量之智能斷惑耶？若
別相智能斷，即違對法五十九等文，對法七等說”).
experiential/phenomenological distinction is often drawn between the con-

centrated state and the non-concentrated state.
In the course of meditation, a practitioner is instructed to investigate the
characteristics of objects. As far as the property in the discerning act belongs to
the single object, the characteristic is called “svalakṣaṇa”. When the character-
istic under discernment is shared by many objects, it is called “sāmānyalakṣaṇa”.21
This is to say that both particular and universal as the property of objects are
the correlate in the mental act of discerning. They are noematic by nature.
While the above distinction between particular and universal is drawn in terms
of object, the other distinction is explained from the side of subject, i.e. three
types of understanding. According to the Abhidharma, the understanding
acquired through hearing and thinking often cognizes the particular as object,
while the understanding acquired through cultivation often cognizes the uni-
versal as object. The latter refers to sixteen modes of ākāras. At first glance, it
seems that particular is distinct from universal in terms of the content of
understanding. However, the content of cognition is determined by the associ-
ated act of understanding. When the property of object as a whole is cognized,
it is called universal. When the property of object is experienced individually, it
is called particular. When the objects are viewed as a whole, their property is
given in sixteen modes of ākāras. According to Dhammajoti’s interpretation,
instead of referring to “images or ‘aspects’ of the object,” the sixteen modes of
ākāras should be understood “in the active sense of the mental function of
understanding,” while universal as the universal principles of all dharmas is
cognized by the understanding (prajñā) acquired through cultivation.22 It is
the same as saying that the understanding of universal occurs at the stage of
cultivation, while the understanding of particular occurs at the stages of hear-
ing and thinking. In the stages of hearing and thinking, sixteen modes of ākāras
are discerned individually, while in the stage of cultivation, sixteen modes of
ākāras as a whole are discerned collectively. In this sense, the nature of the
property of existent remains the same. The distinction is rather drawn upon
21 T.27.1545: 217a12–16: “[Question:] What is the understanding/wisdom (prajñā) capable of

discerning the sāmānyalakṣaṇa of existents? Answer: Discerning the property (lakṣaṇa)
of one entity means to discern svalakṣaṇa, while discerning the property of many entities
means to discern sāmānyalakṣaṇa. Furthermore, discerning an individual aggregate
(skandha) means to discern svalakṣaṇa, while discerning two or three or more aggregates
means discerning sāmānyalakṣaṇa”. (“何等慧能分別諸法共相耶？答：分別一物相
者，是分別自相。分別多物相者，是分別共相。復次，分別一一蘊等者，是
分別自相。分別二蘊、三蘊等者，是分別共相”).
22 Dhammajoti, 2007: 352.
386 Chen-kuo Lin
the different way of discerning. This is explained in the following passage in

the Abhidharmamahāvibhāṣā:
[Question:] What is the understanding (prajñā) capable of discerning the

particular of existents? What is the understanding capable of discerning
the universal of existents? Answer: Discerning the property (lakṣaṇa) of
one entity means to discern the particular, while discerning the property
of many entities means to discern the universal. Furthermore, discerning
an individual aggregate (skandha) means to discern the particular, while
discerning two or three or more aggregates means discerning the univer-
sal. Furthermore, the particular is often discerned by the understanding
acquired through hearing and thinking, while the universal is often dis-
cerned by the understanding acquired through cultivation. Furthermore,
the particular is often discerned by the understanding in which sixteen
ākāras are not included, while the universal is often discerned by the
understanding in which the sixteen ākāras are included. Furthermore,
the particular is often discerned by the understanding of practicing the
truth, while the universal is often discerned by the understanding in direct
realization (abhisamaya). Furthermore, the particular is often discerned
by the understanding of discerning the truths individually, while the uni-
versal is often discerned by the understanding of discerning the truths as
a whole.23
[Question:] How should one know [the difference between] these two
types of understanding? Answer: When the various entities are placed
close to the blue gem, their color looks like that [gem] without manifest-
ing their own particular property. Thus should be known is the
understanding of discerning the universal. When the various entities are
placed in distance from the blue gem, the colors of blue, yellow, and so
on, are manifest individually. Thus should be known is the understanding
of discerning the particular. Furthermore, one should know that the
understanding of discerning the universal is like the shining everywhere
of the rising sun: all shadows are dispelled immediately. One should also
23 T.27.1545: 217a8–21: “問：何等慧能分別諸法自相？何等慧能分別諸法共相耶？

答：分別一物相者是分別自相，分別多物相者是分別共相。復次，分別一一
蘊等者是分別自相，分別二蘊三蘊等者是分別共相。復次，聞思所成慧多分
別自相，修所成慧多分別共相。復次，十六行相所不攝慧多分別自相，十六
行相所攝慧唯分別共相。復次，行諦時慧多分別自相，現觀時慧唯分別共
相。復次，別觀諸諦慧名分別自相，總觀諸諦慧名分別共相”.
know that the understanding which discerns the particular is like the sun
that has arisen, gradually shining on the various things: the gaps between
walls and the valleys between mountains are all manifest. Furthermore,
one should know that the understanding of discerning the universal is
like someone who enters a dark room with a lamp, the darkness of all
corners is broken in an instant. After entering the room, [the light] gradu-
ally shines towards jars, clothes, utensils, baskets, etc. Thus should be
known is the understanding which cognizes the particular. Furthermore,
the understanding of discerning the universal is like a mirror that does
not make the particular property visible from afar. When the mirror is
placed close to the object, the individual property is revealed. Thus
should be known is the understanding of the particular. Furthermore,
one should know that the understanding of discerning the universal is
like someone who watches a mountain and forest from afar. One should
also know that the understanding of discerning the particular is like
someone who closely watches a mountain and forest.24
As shown in the above lengthy passages, especially in the analogies of a blue

gem, sunshine, a lamp in a dark room, and a mirror, the nature of particular
and universal are understood as the correlate to the subjective act of cognition.
They are noematic by nature. In this sense, the Abhidharma ontology of par-
ticular and universal can be better appreciated in light of the phenomenological
description of parts and wholes, identity and difference, and presence and
absence.25
5 Kuiji and Huizhao’s Ontology of Particular and Universal
When we come to Kuiji’s Commentary, we see that in addition to contextu-

alizing the issue of particular and universal within the Buddhist path, Kuiji
24 T.27.1545: 217a22-b6: “問：此二種慧如何應知？答：如種種物近帝青寶，自相不

現，皆同彼色，分別共相慧，應知亦爾。如種種物遠帝青寶，青黃等色，各
別顯現，分別自相慧應知亦爾。復次，如日出時，光明遍照，眾闇頓遣，分
別共相慧，應知亦爾。如日出已，漸照眾物，牆壁竅隙，山巖幽藪，皆悉顯
現，分別自相慧，應知亦爾。復次，如人持燈，初入闇室，頓破諸闇，分別
共相慧，應知亦爾。如燈入已，漸照瓶衣，器篋諸物，分別自相慧，應知亦
爾。復次，如鏡遠照，別相不顯，分別共相慧，應知亦爾。如鏡近照，別相
明了，分別自相慧，應知亦爾。復次，如人遠觀山林等物，分別共相慧，應
知亦爾。如人近觀山林等物，分別自相慧，應知亦爾”.
25 Cf., Sokolowski, 2000: Chapter 3.
388 Chen-kuo Lin
also adopted the Buddhist logician’s theory of cognition. In the Cheng weishi
lun shuji, Kuiji defined “particular” by referring it to “the existent itself which
is only known by direct cognition, but not designated by verbal expression”.
He also defined “universal” by referring it to “the nature of existent which can
be both designated by the verbal expression and cognized by the concep-
tual mind”.26 This definition is adopted from Dignāga’s epistemology, taking
particular and universal as the objects known through direct perception and
inference respectively.27 Following Dignāga’s ontology, Kuiji and Huizhao 慧
沼 also adopted his theory of apoha. In this regard, we have reason to assume
that their knowledge of apoha must come from Xuanzang’s 玄奘 oral teach-
ing, because Xuanzang did not translate Dignāga’s Pramāṇasamuccaya into
Chinese.
Now, let us examine some other textual sources. In the Cheng weishi lun
shuji, Kuiji continued to examine the ontological status of universal:
[Question:] If all existents are [said to be] beyond verbal expression, is it

not contradictory to claim that “universal” refers to that which can be
verbally designated? Answer: Universal refers to the meaning (*artha, yi
義) [which is shown] on the substance/entity (*svarūpa, ziti 自體). It
does not refer to any other substance/entity. For example, when the
name “fire” is employed to refer to fire, it is used to exclude (zhe 遮) non-
fire. By this way, the meaning [of fire through negation] can be applied to
all fires. Hence we say that universal has its meaning. It does not refer to
the universal principles, i.e. universal as truth (gongxiangli 共相理), such
as suffering, emptiness, etc.28
26 T.43.1830: 288a20–23: “Question: Why are ‘svalakṣaṇa’ and ‘sāmānyalakṣaṇa’ so named?

[Answer:] ‘Svalakṣaṇa’ refers to the existent itself which is only known in direct/immedi-
ate perception (zhengzhi 證智 = pratyakṣa) and is not designated by verbal expression. If
the nature of an existent can be both designated by verbal expression and cognized by
conceptual mind (jiazhi 假智), it is called ‘sāmānyalakṣaṇa’”. (“問曰：何故名自相、共
相？答曰：法自體唯證智知，言說不及，是自相。若法體性言說所及，假智
所緣，是為共相”).
27 T.32.1629: 3b10–11: “As to the correct cognition for oneself, there are only perception and
inference as the means of valid knowledge. Those verbal testimony and examples are
included [in inference]. Hence there are merely two means of valid knowledge for the
reason that one is capable of cognizing two characteristics [i.e. particular and universal]
through these [i.e. perception and inference]”. (“為自開悟，唯有現量及與比量。彼
聲喻等，攝在此中，故唯二量。由此能了自共相故”).
28 T.43.1830: 288a20–27: “問曰：如一切法皆言不及，而復乃云言說及者是為共相，
一何乖返？答曰：共相是法自體上義，更無別體。且如名詮火等法時，遮非
In Kuiji’s ontology of universal, a concept (sāmānyalakṣaṇa) is not real in itself.

It does not possess any form of substance. Following Dignāga’s theory of apoha,
Kuiji holds the view that universal is the meaning of X which can be explained
only through negation of non-X.29 In Kuiji’s words, the universal is lack of sub-
stance (wuti 無體). Hence, the universal is ontologically unreal. If there is any
ontological quality of universal, it is the meaning (artha) of the existent which
is constructed through exclusion of other, the function of apoha. Of course, the
whole issue will become much more complicated when Kuiji’s ontology of uni-
versal is understood within the paradigm of consciousness-only. That is, the
ontological status of universal as meaning remains indeterminate with regard
to the meaning of meaning.30
The same ontological problem of universal was taken up by Huizhao again.
According to Huizhao’s interpretation in the Treatise on Two Pramāṇas (Erliang
zhang 二量章), the universal as the object of inference is provisionally desig-
nated by the conceptual mind.31 That is, universal is not real, for it is signified
by verbal expression (prajñapti) to that which is excluded of others in the same
class. Like Kuiji, Huizhao takes a nominalist position with regard to the onto-
logical status of the universal.
However, we should be more cautious about this conclusion. The question
can be pushed further: if the universal is not real, what causes the conceptual
mind to arise? The conceptual mind cannot arise from something unreal, for
the arising of something real must be caused by something real too. Now, what
is this real something? According to the Yogācāra causal theory of cognition,
the universals cannot serve as the cause of conceptual mind, because they are
prajñapti in the sense that they are lack of substance. Following Xuanzang and
Kuiji’s Yogācāra-Vijñaptimātra system, Huizhao claims that the conceptual
mind takes the noematic part (xiangfen 相分) of mind as its object, an internal
object in the mind. The internal object (noema) is further divided into two
types: “noema qua image” (pratibimba, yingxiang xiangfen 影像相分) and
“noema qua basic stuff” (bimba, benzhi xiangfen 本質相分). The noema qua
火等。此義即通一切火上。故言共相得其義也。非苦空等之共相理”.
29 Cf. Katsura, 2014: 99–118.
30 For the Yogācāra, the meaning (artha) of verbal designations is the conceptual construc-
tion (parikalpita) superimposed upon the mental phenomena (paratantra); the latter
arises depending on the cause and conditions. Ontologically speaking, while parikalpita
is not real at all, paratantra is real in the sense of phenomena.
31 Huizhao, “Erliang zhang” 二量章 in Dasheng fayuan yilinzhang buque 大乘法苑義林章
補闕, X55, No. 0882. For the study of this text, cf. Lin, 2014 (forthcoming): “How to attain
enlightenment through cognition of particulars and universals? – Huizhao on svalakṣaṇa
and sāmānyalakṣaṇa”.
390 Chen-kuo Lin
image serves as the content of cognition, which is also termed “the direct
object of cognition” (qin suoyuanyuan 親所緣緣), while the noema qua basic
stuff is named “indirect object of cognition” (shu suoyuanyuan 疏所緣緣),
which cannot be directly known by the seeing part of mind (noesis). As
Huizhao explains in The Lamp for Illuminating the Definite Meaning of Cheng
Weishi lun (Cheng weishi lun liaoyi deng 成唯識論了義燈)32, universal refers to
the conceptual construct which is created and superimposed by the concep-
tual mind upon the “noema qua original stuff”. This type of conceptual
construct is not real in itself, for it is parikalpita (conceptually-constructed) in
view of the theory of threefold nature (trisvabhāva). Direct perception in the
structure of three or four parts (*bhāga), including the seeing part (noesis) and
the image part (noema), however, is real in the sense of paratantra (phenom-
ena as causally arising), while universal is not real in the sense of parikalpita.
Hence, universal as prajñapti or parikalpita is said to be lack of substance,
therefore not real.
In the Treatise on Two Pramāṇas, Huizhao’s explanation is somehow differ-
ent.33 He says that universal can still be qualified as something real, if it is
32 T.43.1832: 716a23–26: “Since sāmānyalakṣaṇa is lack of substance, under what condition

does the cognition arise? Answer: Although sāmānyalakṣaṇa superimposes the illusory
appearance (siyou 似有) upon the svalakṣaṇa qua original stuff (bimba) (benzhi zixiang
本質自相), when the conceptual mind cognizes, it transforms the original stuff (bimba)
into the image part of mind (xiangfen). Both bimba and xiangfen are lack of substance of
the universals. Cognition merely takes the image part of mind (xiangfen, noema) as its
object. It cannot take the sāmānyalakṣaṇa of that [xiangfen] as object”. (“共相體無，智
緣何起？答：共相雖依本質自相增益似有，假智緣時，還依本質以變相分，
質相俱無彼共相體，智但緣相，不能緣著彼之共相”).
33 X55.163a11-a21: “Question: In the list of hundred dharmas, under which category is sub-
sumed the image part carried by the universal character as depicted in the logical treatise
and permanence and impermanence cognized in the concentrated state of mind?
Answer: There are two [explanations]. According to the first explanation, owing to non-
substantiality [i.e. noema] it is not subsumed under any one of the hundred dharmas.
According to the second explanation, it is subsumed under [the category of] commonal-
ity (sabhāga) of existents. Now we classify [the noematic part] under the category of
commonality [of existents] for the reason that it can be taken in the category of five
aggregates and it follows the changing mind in the way that it cannot be separated from
the mind. Commonality is established on the basis of various resembling existents. It
does not have substance. The universal character of this sort can be said to have sub-
stance if it refers to the noematic part as the result of the transformation of conditions,
because the condition of the direct object of cognition must exist. It [i.e. universal char-
acter] is also different from the noematic part which is substantially real, being the result
of the same seeds with the noetic part, and serving as the object of the noetic part. It is
viewed in terms of the “noema qua image” as the direct object of cognition.
Nevertheless, it is void of substance, if it is viewed in terms of “original stuff”
(bimba). This explanation seems to be used for explaining the cognition of
impermanence, suffering, etc., in the concentrated state of meditation, for
impermanence and suffering are vividly and intuitively perceived as the direct
object of experience, which therefore cannot be said to be not real at all. If
impermanence, suffering, etc., can be taken as universal, as the Abhidharma
holds, then they cannot be something entirely unreal. Again, the context of
meditation is taken into account with regard to the ontological issue of
universal.
To sum up, if Huizhao exclusively takes a nominalist position, he will be
forced to answer another difficult question: How is it possible for the universal
to be cognized when it is not real? If he takes a conceptualist position, on the
contrary, he will also have difficulty to explain Dignāga’s theory of anyāpoha,
which he learned in the Xuanzang School.34 According to the theory of apoha,
the universals can be known only through the exclusion of the rest of the
members in the same class, e.g. “blue” means the exclusion of the other colors
that are not blue. Nevertheless, due to the ontological commitment to the
non-substance if it is viewed from the “original matter”. Cognizing permanence, imper-

manence and so on, which are cognized in the concentrated state of mind is merely
clearly intuiting the transformation of mind. If the universal character is taken [as the
object of cognition] by the mind of inference, it is provisionally established by the mind
of reference. In that case, the Buddha would not take [as the object of cognition] the
universal character which is the object of the mind of inference. As it is said in the
Buddhabhūmiśāstra, the cognition of the Buddha does not take the universal character of
inference as object. Answer: The cognition of the Buddha does not take the sameness (i.e.
universal) provisionally established by the mind of inference. [However,] the Buddha has
omniscience through which he apprehends the sentient beings’ mind of inference as well
as establishes all universal characteristics for the sake of practice and understanding. If
one is able to penetrate the imagined [character], in truth what does one lose?” (“ 問：此
論共相及定心緣常無常等所帶相分，百法何攝？答：一云，以無體故，百法
不取。一云，法同分攝。今謂可通五蘊所攝。隨能變心，不可離心，判屬同
分。同分依相似眾多法立，而全無體。此之共相，若據緣所變相分，可言有
體。親所緣緣，定應有故。亦不同於相分體實，與見同種。義分所緣，若據
本質，即是無體。定心所緣常無常等，但現觀心變。若比量心所緣共相，但
比量心假所安立。若爾，此比量心所緣共相，佛應不緣。《佛地論》說，佛
智不緣比量共相。答：作[若?]比量心，假立無異，佛智不緣。佛有遍智，緣
彼有情比量之心，行解安立所有共相，如達遍計所執，於理何失 ?”).
34 Traditionally, Kuiji is regarded as the founder of the Faxiang School of Chinese Yogācāra
Buddhism. However, owing to the fact that Kuiji’s system is almost entirely inherited from
Xuanzang’s teaching, it is more reasonable to rename this school as “Xuanzang School”.
392 Chen-kuo Lin
Vijñaptimātra doctrine, Huizhao seems inevitably to be forced to compromise

the position between nominalism and conceptualism.
6 Concluding Summary
In the following, some tentative remarks are concluded from the above
expositions:
1. The continuity of Abhidharma and Yogācāra is grounded in the same

paradigm of mental cultivation and theorization. For both Abhidharma
and Yogācāra the ontology of particular and universal cannot fully make
sense unless they are understood within the context of meditative prac-
tice. The textual evidences can be found in both Abhidharma and
Yogācāra canons, such as the Abhidharmamahāvibhāṣā and the Cheng
weishi lun. In the former, the analysis of particular and universal is intro-
duced at the stage of warmth. In the Cheng weishi lun, Chinese Yogācāra
Buddhists are deeply concerned with the content of the Buddha’s experi-
ence of enlightenment. The discussion in the Cheng weishi lun is largely
indebted to the influence of the Buddhabhūmyupadeśa.
2. Regarding the question about the experience that leads to the final elimi-
nation of contaminations for attaining liberation, the Ābhidharmikas’
answer is that the cognition of universal leads to liberation. This convic-
tion is firmly stated in the Abhidharmamahāvibhāṣā: “Only in the path
which cognizes universal as the object one is able to eliminate the
contaminations”.
3. One might consider phenomenology as a better approach to account for
the Ābhidharmika position with regard to the ontology of particular and
universal. The reason is that in the Abhidharma literature the nature of
particular and universal are varied in terms of the mental states which
are in correspondence with the various stages of cultivation. For the
Ābhidharmikas, particular and universal are the objects of cognition, cor-
relating to the subjective act, while cognition varies by depending on the
various mental states in the course of meditation, either the concen-
trated state or the non-concentrated state of mind. In short, particular
and universal are noematic by nature.
4. In contrast to the Ābhidharmika stance, Kuiji argues that the cognition of
particular is capable of destroying the contaminations. Kuiji’s own view
was already shown in the third theory he adopted from the Buddhabhū
myupadeśa, claiming that even the sixteen modes of ākāras as universals
are perceived directly by abhisamaya in the concentrated state of mind.

Nevertheless, it does not mean that Kuiji was not aware of the crucial
function of universal in the course of meditation. On the contrary, he
emphasized the instrumental role of the cognition of universal as a nec-
essary upāya in the progressive path that finally leads to the cognition of
particular. His pragmatic conception of universal can also explain why in
the Yogācāra school, including its development in seventh-eighth cen-
tury China, logico-epistemology played a crucial role in the Buddhist
soteriological project.
5. A unique feature in Kuiji’s contribution is that he employs the method of
apoha to account for the nominalist ontology of universal. For Kuiji, uni-
versal is not real in itself. It does not possess any form of substance, such
as tree, pot, etc. However, being not real does not imply being pragmati-
cally useless.
6. Following Kuiji’s steps, on the one hand, Huizhao tends to explain the
ontology of universal within the Yogācāra framework, such as the theory
of three natures. On the other hand, owing to the commitment to the
Vijñaptimātra ontology, Huizhao compelled to negotiate the position
between nominalism and conceptualism. For Huizhao, universal can be
properly taken as prajñapti or parikalpita. It is lack of substance, there-
fore not real. However, universal can still be qualified as mentally real, if it
is taken as the “noema qua image,” which is the direct object of cognition,
and therefore must be real to a certain degree. No matter how plausible
this interpretation could be, Huizhao’s theory can be helpful for rethink-
ing the possibility of conceptualism as an alternative interpretation in
the Yogācāra theory of universals.
Abbreviation

邊 and Ono 小野, 1924–1934)
Bibliography
Cox, Collett (2004). “From Category to Ontology: The Changing Role of Dharma in
Sarvāstivāda Abhidharma”, Journal of Indian Philosophy 32, pp. 543–597.
Cox, Collett (1992). “Attainment through Abandonment: The Sarvāstivādin Path of
Removing Defilements”, In Paths to Liberation: The Mārga and its Transformations in
394 Chen-kuo Lin
Buddhist Thought. Edited by Robert E. Buswell Jr. and Robert M. Gimello. Honolulu
HI: University of Hawaii Press, pp. 63–105.
Dhammajoti Bhikkhu, KL (2007). Sarvāstivāda Abhidharma. Hong Kong: Centre of
Buddhist Studies, The University of Hong Kong.
Huizhao 慧沼, Dasheng fayuan yilin zhang buque 大乘法苑義林章補闕. X55.882.
Katsura, Shōryū (2011). “From Abhidharma to Dharmakīrti – With a special reference
to the concept of svabhāva”, In Religion and Logic in Buddhist Philosophical Analysis.
Proceedings of the Fourth International Dharmakīrti Conference (Vienna, August
23–27, 2005) (Denkschriften der phil.-hist. Klasse 424. Beiträge zur Kultur- und
Geistesgeschichte Asiens 69). Edited by Helmut Krasser, Horst Lasic, Eli Franco and
Birgit Kellner. Wien: Verlag der Östereichischen Akademie der Wissenschaften,
pp. 271–279.
Katsura, Shōryū (2014). “The theory of apoha in Kuiji’s Cheng weishi lun shuji”, In
A Distant Mirror: Articulating Indic Ideas in Sixth and Seventh Century China. Edited
by Chen-kuo Lin and Michael. Radich. Hamburg: University of Hamburg Press,
pp. 99–118.
Kuiji 窺基, Cheng weishi lun biechao 成唯識論別抄. X48.808.
Lin, Chen-kuo (forthcoming). “How to attain enlightenment through cognition of par-
ticulars and universals? – Huizhao on svalakṣaṇa and sāmānyalakṣaṇa”, In Dharma
and Dao: Chinese Buddhist Philosophy. Edited by Sandra Wawrytko and Wang Youru.
Springer.
Muller, A. Charles (2004). “The Yogācāra two Hindrances and their Reinterpretations in
East Asia”, Journal of the International Association of Buddhist Studies 27,
pp. 207–235.
Saito, Akira (ed.). Bauddha Kośa, <http://www.l.u-tokyo.ac.jp/~b_kosha/html/index_75
dharma.html#>.
Sokolowski, Robert (2000). Introduction to Phenomenology. Cambridge: Cambridge
University Press.
Sponberg, Alan (1986). “Meditation in Fa-hsiang Buddhism”, In Traditions of Meditation
in Chinese Buddhism. Edited by Peter Gregory. Honolulu HI: University of Hawaii
Press, pp. 15–43.
- T.26.1530, Bandhuprabha et al., Buddhabhūmyupadeśa, Fodijing lun 佛地經論,
Xuanzang 玄奘.
- T.31.1585: Dharmapāla, et al., Vijn͂ aptimātratāsiddhi[śāstra], Cheng weishi lun 成唯
識論, Xuanzang 玄奘.
- T.31.1606, Sthiramati, Abhidharmasamuccayavyākhyā, Dasheng apidamo zajilun 大

乘阿毘達磨雜集論, Xuanzang 玄奘.
- T.32.1628, Dignāga, Nyāyamukha, Yinming zhenli men lun 因明正理門論, Xuanzang
玄奘.
- T.43.1830, Kuiji 窺基, Cheng weishi lun shuji 成唯識論述記.
- T.43.1832, Huizhao 慧沼, Cheng weishi lun liaoyi deng 成唯識論了義燈.
396 Yao-ming Tsai
Chapter 13
Perspectives on the Person and the Self

in Vasubandhu’s Abhidharmakośabhāṣya
Yao-ming Tsai
1 Introduction
This chapter examines Vasubandhu’s approach to the ideas of the person

and the self in his Abhidharmakośabhāṣya and explores the usefulness of
his approach to the study of sentient beings or human beings in the world.
With the thesis of the “selflessness of the person” (pudgala-nairātmya), Vasu
bandhu was operating within a Buddhist philosophical tradition, and his
critique of the Pudgalavādins and the Tīrthikas provides a clearing that reveals
important aspects of other traditions and cultures. Analysis of the problem
of wrong views finds one of its clearest expressions in Buddhism in the ninth
chapter of the Abhidharmakośabhāṣya. Through a critical examination of this
important work, this chapter will illustrate the significance of the role played
by soteriology in Buddhist intellectual thought. This, in turn, may open up
new ways of understanding the life-and-world through a transition from an
“Abhidharmic” to a “not-dharmic” orientation.
Among all the Buddha’s teachings, those on the nature and identity of the
self are considered basic to Buddhist practices. The Buddha’s teachings are
often summarized in dharma-seals1, the characteristic principles of the factors
of the sentient world. One of the three or four dharma-seals is formulated as
follows: All factors are not-self (anātmanaḥ sarva-dharmāḥ / sabbe dhammā
anattā).2 This principle sets the factors of the sentient world as the subject
matter of exploration and states that all the explored factors are not the self.
Although the principle of the not-self of factors (dharma-nairātmya) consti-
tutes one of the most distinctive features of the Buddha’s teachings, its meaning
is not fully conveyed by its summarized proposition, and is therefore open to
1 Hakamaya discussed the Indic origin of this term and suggested “dharmoddāna” rather than
the often assumed one, “dharmamudrā” (Hakamaya, 1979: 60–81).
2 For the rendering of the term anātman / anattā as “not-self,” see Chapter 2 “Impermanence,
Not-self and Suffering” in David Burton, 2004: 11–30; Chapter 6 “An Interpretation of the Not-
Self Doctrine” in Gowans, 2003: 63–75; Chapter 1 “Setting the Scene” in Hamilton, 2000: 18–32;
Tsai, 2006: 145–147; Tsai, 2013: 121–122.

Perspectives on the Person and the Self 397
various interpretations and thoughts. For example, the Indian Buddhist phi-
losopher Vasubandhu (ca. fourth/fifth century CE), offers just such an
interpretation in the ninth chapter, entitled “Refutation of the Doctrine of the
Self” (Ātmavāda-pratiṣedha) or “An Examination of the Person (or Individual)”
(Pudgala-viniścāya), which he added to his masterwork of Abhidharma stud-
ies: the Commentary on the Treasury of the Abhidharma (Abhidharmakośabhāṣ
ya). In the chapter “Refutation of the Doctrine of the Self” (henceforth abbrevi-
ated as “Refutation”), Vasubandhu briefly presents his perspectives on the
person and the self, raises objections to the viewpoint of the Pudgalavādins (or
Vātsīputrīyas, i.e. followers of Vātsīputra), replies to the Pudgalavādins’ objec-
tions to his own view, and then replies to the objections put forward by the
Tīrthikas (Forders), i.e. non-Buddhist thinkers of India, specifically the Nyāya-
Vaiśeṣika tradition.
Vasubandhu’s “Refutation,” together with his other works, has been a subject
of veneration and disputation across the Buddhist traditions, and has attracted
much attention in contemporary philosophical circles. Numerous scholarly
works have been written on the “Refutation,” especially James Duerlinger’s
Indian Buddhist Theories of Persons, a translation mainly based on the Sanskrit
text of the Abhidharmakośabhāṣya, which provides a comprehensive intro-
duction and analytical commentary.3 Yet, even with such an abundance of
commentaries, there are still several critical points that bear scrutiny to clarify
both what is intended by the concept that all factors are not-self, and what
“Refutation” leads to.
In what follows, I will demonstrate that the doctrine of not-self is about
metaphysical identification, i.e. identification in terms of the nature of reality,
rather than existential differentiation, i.e. differentiation between existence
and non-existence. To that end, I elucidate the person-concept (individual-
concept) and the self-concept, two key concepts that have played important
roles in Buddhist traditions, and adjust perspectives on not-person and not-
self in a way that is conducive to gaining insight into the sentient world.
Moreover, I offer three critiques of Abhidharmic oriented perspectives formu-
lated from an understanding of the “not-dharmic” oriented perspectives, and
conclude by proposing a shift from “Abhidharmic” to “not-dharmic” perspec-
tives when engaging in philosophical inquiry into Buddhist teachings.
3 Duerlinger, 2003. See also Duerlinger, 2009.

398 Yao-ming Tsai
2 Not-Self Critically Reconsidered
Before Vasubandhu’s “Refutation” can be examined appropriately, it is neces-

sary to realize that one is not simply being confronted with intractable religious
and philosophical pluralism, but with the absence of an academic ground sub-
stantiating why the doctrine of not-self should be interpreted in a certain way
– because that is exactly what the Buddha has laid down in the sūtras. One way
of understanding the doctrine of not-self, however, is to examine what it
entails.
The issue of not-self is not about whether the self exists or not. Rather, it is
about whether a certain factor is to be identified as the self or not. In other
words, the issue of not-self is about metaphysical identification rather than
existential differentiation.
In many sūtras of the Śrāvakayāna,4 the Buddha frequently asks prospective
practitioners to attentively watch such factors as the five aggregates (pañcas-
kandha) in their state of impermanence, suffering, inclination to change, and
to critically consider if any such factor is a valid identification of the self (shi
wo 是我), as other than the self (shi yi wo 是異我) [in the sense of being owned
by the self (shi wosuo 是我所)], or as either the very factor being within the
self, or the self being within the very factor (shi xiang zai / se naizhi shi ji shi zai
wo zhong 是相在／色乃至識既是在我中、而我亦是在色乃至識中).5 Well-
taught disciples will unequivocally answer that none of the five aggregates are
qualified to be identified as the self (bu shi wo / fei shi wo 不是我／非是我), as
other than the self (bu yi wo / fei yi wo 不異我／非異我) [in the sense of being
owned by the self] (bu shi wosuo / fei shi wosuo 不是我所／非是我所), nor as
either the very aggregate being within the self or the self being within the very
aggregate (bu xiang zai / fei xiang zai / se naizhi shi ji fei zai wo zhong, er wo yi
fei zai se naizhi shi zhong 不相在／非相在／色乃至識既非在我中、而我亦非
在色乃至識中).6
4 See the Saṃyuktāgama nos. 23–24 at T.2.99: 5a11-b 27; nos. 30–34 at T.2.99: 6a24–8a4; no. 43
at T.2.99: 10c19–11a12; nos. 62–64 at T.2.99: 16a19–17a22; no. 76 at T.2.99: 19c12–24; nos. 82–87 at
T.2.99: 21b14–22b13; nos. 104–105 at T.2.99: 30c12–32c1; nos. 109–110 at T.2.99: 34a24–37c7; nos.
120–121 at T.2.99: 39b25–40a3; no. 125 at T.2.99: 40c6–27; Bhikkhu Bodhi, 2000: 616, 866, 888,
902, 927, 948, 969, 979, 992, 1003, 1135, 1166, 1389, …; Bhikkhu Ñāṇamoli and Bodhi, 2005: 123,
229–232, 279–282, 327–330, 527–529, 890, 1089–1091, 1122, etc.
5 See also Bhikkhu Anālayo, 2010: 49–50; Bodhesako, 2008: 79; Choong, 2000: 58–59; Wynne,
2007: 35, 140.
6 Example one: “rūpaṃ bhikṣavo ’nityaṃ. yad anityaṃ, tad duḥkhaṃ. yad duḥkhaṃ, tad anātmā.
yad anātmā, tan naitan mama, naiṣo ’ham asmi, naiṣa me ātmêty. evam etad yathābhūtaṃ
samyakprajñayā draṣṭavyam. vedanā saṃjñā saṃskārā vijñānam anityaṃ. …” (Chung, 2008:
The procedure for examining the issue of self/not-self can thus be analyzed
and differentiated into the following four steps.
First, one is required to observe the five aggregates in the flow of life. Without
the practice of observing the five aggregates, one is likely to be hindered from
attaining a firsthand understanding of the issue of self/not-self, and will, there-
fore, be confined within the sphere of beliefs, opinions, and ideologies.
However, to observe the five aggregates does not presuppose that there exists a
person or a self which is composed of the five aggregates. More to the point,
what is taken for granted as constituting the person or the self is to be rigor-
ously examined, or critically identified, but not as a given presupposition. The
five aggregates are brought on stage because they encompass existence in the
sentient world and are available to perception. What makes Buddhist observa-
tion special is that it does not fixate on the five aggregates but conforms to
their constant flux.
Second, actually observing the five aggregates in the flow of life affords a
vantage point from which the issue of the person or the self can be better
examined. Although the five aggregates can be perceived and observed, the
person or the self cannot be perceived or observed because the latter is located
at a level slightly remote from that of the perceptible and observable. Generally,
people assume the person or the self without pertinent observation or reason-
ing. In other words, the person or the self is unquestionably assumed or
projected in conventional usage. However, Buddhism offers a strategy for criti-
cally identifying the person or the self before any serious measures can be
taken beyond the level of conventional usage.
Third, after observing the five aggregates in the flow of life, it becomes clear
that none of the five aggregates can be identified as the person or the self.
Different philosophers or thinkers may have different ways in which they
define the person or the self. Various definitions come from different perspec-
tives and experiences, express dissimilar opinions, and then invite argument
and controversy. However, the task of identification, especially metaphysical
311–312, 323–326, 331–334). Example two: tathā hi rūpaṃ nâtmā, rūpavān nâpi câtmā, rūpe
nâtmā, nâtmani rūpam. evaṃ yāvat vijñānaṃ nâtmā, vijñānavān nâtmā, vijñāne nâtmā, nât-
mani vijñānam iti. tathā a[n]ātmānaḥ sarva-dharmā iti. (Yoshimizu, 2007: 20). (<http://www.
tulips.tsukuba.ac.jp/limedio/dlam/B26/B2674017/1.pdf>). Example three: see Harivarman’s
Satyasiddhiśāstra translated by Kumārajīva: T.32.1646: 314b16–18: “如長老差摩伽說。不說
色是我。不說受想行識是我。但五陰中有我慢、我欲、我使。未斷未盡是名我
慢”. Cf. yathā sthaviraḥ kṣemaka āha: ‘na khalv āyuṣman rūpam asmîti vadāmi, na vedanā, na
saṃjñā, na saṃskārā, na vijñānam, [nâpy anyatra vijñānād asmîti vadāmi]. api ca ma āyuṣman
pañcasûpādāna-skandheṣu anusahagato ’smîti mānaḥ, asmîti cchandaḥ, asmîty anuśayo,
’samuddhataḥ. [ity ādi]. ayam asmi-māna ity ucyate. (Sastri, 1975: 307; 1978: 285–286).
400 Yao-ming Tsai
identification, is prior to definition. If none of the perceptible and observable

five aggregates can be identified as the person or the self, various definitions of
such terms as ‘person’ or ‘self’ will at most be drifting expressions of opinions
rooted in linguistic convention without any referents in reality. Similarly, when
examining or looking into other concepts, given the range of potential issues,
it is always essential to prioritize issues by first addressing that of metaphysical
identification. For example, it would be illogical to take sides on the issue of
existential differentiation and assert either the existence or non-existence of
the self without antecedently understanding the non-identifiability (or non-
identity) of the self. How can one argue for or against the existence of the self
which in reality fails to be identifiable by any observable factors?7
Fourth, after proposing the thesis of non-identifiability (or non-identity) of
the self, i.e. the thesis of not-self, it is unnecessary, or even erroneous, to claim
either that the self exists or that the self does not exist. In other words, after
understanding the non-identifiability (or non-identity) of the self, one is freed
from any view attached to the self and the differentiation between existence
and non-existence.
The self-concept, or better put, the word “I,” is basically a multi-dimensional
and dynamic construct, active in verbal expression by recourse to the first-per-
son singular pronoun, “the content of which is influenced by the social situation
at a given time, in addition to being influenced by an individual’s current goals,
emotional and motivational state”.8 Although the self-concept is frequently
used in verbal communication, it does not mean that the self-concept actu-
ally refers to any observable factor that is exactly the self. Endowed with such
understanding, as is required for right understanding (samyagavagama; sam
yagavabodha), one may well use the self-concept in verbal communication
and teaching when necessary, but without entailing any substantialist views or
grasping any defilements associated with the self.9
Moreover, concerning the differentiation between existence and non-exis-
tence, it is preferable not to regard such differentiation as the nature of reality
or hold on to such differentiation. While contemplating the nature of exis-
tence or non-existence, it is better to explore either the diachronic sequence of
existence or the diachronic sequence of non-existence than to assert that a
certain entity simply exists or that a certain entity simply does not exist. A
conceptual model of such diachronic exploration occurs very often in the
sūtras of the Śrāvakayāna, and is formulated as follows: “When that exists (or is
7 Cf. Tsai, 2013: 338–339.

8 Cf. Winchester, 2009: 64.
9 See the Saṃyuktāgama nos. 23–24 at T.2.99: 5a11-b27; Bodhi, 2009: 947–948.
present), this comes to be; from the arising of that, this arises. When that does
not exist (or is not present), this does not come to be; from the cessation of
that, this ceases”.10 From such a perspective, synchronic differentiation
between existence and non-existence can be avoided by observing the dia-
chronic sequence of both dependent origination (pratītyasamutpāda) and
cessation (nirodha).11
In sum, before engaging in discourse on the self, one should examine
whether any observable factors can be identified as the self. Since none of the
observable factors are qualified to be identified as the self, the label of not-self
is used to indicate the non-identifiability (or non-identity) of the self. This
insightful understanding of not-self does not oppose the inclusion of the self-
concept in verbal communication or teaching as long as the self-concept
operates by and within linguistic convention. Furthermore, views that posit or
presuppose the existence and non-existence of the self are equally problem-
atic. The postulation of the existence (astitva) of the self is labeled as the view
of eternalism (śāśvatadṛṣṭi), implying that the self is regarded as eternal and
unchanging. The postulation of the non-existence (nāstitva) of the self is
labeled as the view of annihilationism (ucchedadṛṣṭi), which asserts that the
self is cut off and utterly destroyed at death. Both of these two extremes not
only misrepresent the reality of the dynamic continuity of mental process but
also shut the door to liberation, and therefore are to be discarded.12
3 The Two Concepts of Person (Individual) and Self Revisited: For

Better or for Worse
In Buddhist teachings, the person-concept (individual-concept) and the self-

concept are related to fundamental understandings of the sentient world, and
to the cultivation of liberation. For this reason, they have attracted the most
Buddhist thinkers.
10 asmin sati, idaṃ bhavaty; asyôtpādād, idam utpadyate. asminn asati, idaṃ na bhavaty;
asya nirodhād, idaṃ nirudhyate. imasmiṃ sati, idaṃ hoti; imass’ uppādā, idaṃ uppajjati.
imasmiṃ asati, idaṃ na hoti; imassa nirodhā, idaṃ nirujjhati. “此有故彼有，此生故彼
生；此無故彼無，此滅故彼滅”. See Vaidya, 1958: 221.
11 See the Saṃyuktāgama no. 262 at T.2.99: 66b6–67a21; Bodhi, 2009: 946–7.
12 See the Saṃyuktāgama no. 961 at T.2.99: 245b9–25; Bodhi, 2009: 1393–4. See also Vasu-
bandhu’s Apidamo jushe shilun 阿毘達摩俱舍釋論 translated by Paramārtha, T.29.1559:
307b; Vasubandhu’s Apidamo jushe lun 阿毘達摩俱舍論 translated by Xuanzang,
T.29.1558: 156a; Duerlinger, 2003: 90–91.
402 Yao-ming Tsai
However, different schools and authors have held divergent views about
these concepts, in large part because of an incomplete knowledge of the perti-
nent concepts and different perspectives on the sentient world. Consequently,
the objective of this section is to present a philosophical elucidation of the
person-concept (individual-concept) and the self-concept, rather than engag-
ing in a history of thoughts. Therefore, the focus of this section will be on
clarifying the connotations of these two concepts in explaining how the sen-
tient world works.
The person-concept (individual-concept) considered in this chapter is a
partial equivalent of the Sanskrit word pudgala (or puggala in Pāli). As the
term pudgala is difficult to translate into English, an exposition of its three
connotations follows below.
First, in its connotation as individual, the term “pudgala is traditionally
said to be derived from puṃ- (joining) plus -gala (breaking)”.13 Hence, etymo-
logically, the term pudgala means the individual or discrete individual that
undergoes modifications by combination and dissection. As for the rendering
as separate individual, the term pudgala may well be related to pṛthak (dif-
ferent; separate; individual), and could therefore be related to pṛthagjana
(ordinary beings; still vulnerable to being subverted by constantly-varying
affections and aversions).14 In short, etymologically speaking, the constant
process of integration and disintegration is characteristic of individuals in the
world.
Second, pudgala is rendered as a person. If the focus is on human beings
rather than on physical objects or other manifestations of sentient beings,
then the literal meaning of the term pudgala can shift from the individual to
the person or a single human. In Sanskrit, several other terms such as puruṣa,
manuṣya, nara, or jana can be used to mean human being in general and per-
son in particular. However, the term pudgala is probably more salient in
conveying the connotation of a person separate from other persons by provid-
ing an idea of being different or separate. Such a connotation can further
elevate the term pudgala’s individual subjectivity and capacity to be situated
as an objectified entity.
Third, pudgala is rendered as a continuous stream.15 If the focus is on the
process of integration and disintegration rather than on individual or person,
then the term pudgala is called upon to act as continuous stream, i.e. the bun-
dle of tendencies that keeps reincarnating as an individual or a person in a
13 Jaini, 1997: 101. See also Kattackal, 1999: 182.

14 See Tedesco, 1947: 174.
15 See Sutton, 1991: 100.
continuous flow of karma and rebirth. Of particular importance is that what is

going on in continuous stream may be inexplicable or beyond description, but
it is not necessarily mysterious in essence or substantialist in nature.
The foregoing exposition of the connotations of pudgala is not undertaken
simply for the sake of philological analysis or semantic clarification, but for the
purpose of philosophical argumentation. Taking the term pudgala as an exam-
ple, first, when an individual is meant, questions about individuality may be
brought forward: What is it that constitutes individuality? How is this individu-
ality to be understood? Is there a real difference between individuals? Second,
the connotation of a person raises perennial questions about personhood:
What makes an entity (or a being) count as a person? Is it possible to be a
human being but not a person? Third, when continuous stream is meant, fur-
ther questions relating to personal identity (or non-identity) may be raised: In
what sense can we maintain that a person at one time remains constant at
another time – despite the constant process of combination and dissociation?
What is the ontological status of the concept that refers to continuous stream?
Does the continuous stream ultimately exist as an entity?
4 Perspectives Reoriented to Not-Person and Not-Self
In response to bitter disputes in the Buddhist tradition throughout history, the

strategy of this chapter is to adjust perspectives in a way that is conducive to
gaining insight into the sentient world, that is to say, soteriological insight. In
other words, it is time to maneuver into reoriented perspectives to bring fresh
observation to life in Buddhist philosophy, rather than simply repeating or
challenging traditional partisan views.
Before the term pudgala can be safely used to support any doctrinal thesis,
it is crucial to examine what this term is. Two points in particular are worth
noting at this stage. First, no matter how broad the usage of the term pudgala
may become with its connotations from individual to person to continuous
stream, it is primarily a conceptual device to convey a set of perceptions. In
short, the term pudgala is a conceptual expression. Second, one of the most
overlooked and often neglected aspects of inquiry into this concept is asking
what can be identified as pudgala. If the constant flux of the five aggregates is
actually observed, there is no single factor identifiable as pudgala. Moreover,
apart from the five aggregates, none of the perceptible and observable phe-
nomena can be identified as pudgala either. Therefore, with or without the five
aggregates, nothing can be identified as pudgala, i.e. nothing can be identified
as individual or person or continuous stream. In light of non-identification, it
404 Yao-ming Tsai
can be proposed that the reality of individual is not-individual; the reality of

person is not-person; and the reality of continuous stream is not-continuous
stream. In short, the term pudgala is not-pudgala.
From the above analysis, it is certain that as a conceptual expression, the
term pudgala is in reality not-pudgala. This seemingly paradoxical discovery
comes mainly from a reorientation of perspectives. In other words, rather than
simply accepting traditional views or anxiously locating the referred object
from the received concept, this discovery sets the perceptible and observ-
able phenomena as a starting point for inquiry, and eventually rejects the
metaphysical legitimacy of such a concept as pudgala on account of its non-
identifiability (or non-identity).
Just as the term pudgala (person) is in reality not-pudgala (non-identity
of pudgala; not-person), so the term self in reality is not-self.16 This discovery
does not proceed from concepts to the objects they refer to, as most people
and philosophical traditions would assume. Rather, it reorients its perspec-
tive, and proceeds from the perceptible and observable phenomena to debunk
conceptualization.
The discovery and understanding of the thesis of not-pudgala (not-person)
and not-self can make a big difference for subsequent discourses and inquiry.
First, both concepts of pudgala and self may contain various connotations in
different contexts, but none of those connotations should be treated as exist-
ing in itself, separate from its usage. Second, both pudgala and self may be
used to roughly and remotely indicate some phenomena of the sentient world,
but the sentient world should not be treated as containing the exact entities to
which both concepts refer. Third, both pudgala and self may expand the range
of their connotations or indications from time to time, but as conceptual
expressions, both concepts should not be treated as endowed with any meta-
physical qualification in themselves. Fourth, when it comes to applying both
pudgala and self in a proposition, such an application should be confined to
the sphere of linguistic convention without entailing any unwarranted meta-
physical or ontological views.
Having shifted the perspective from pudgala to not-pudgala (not-person)
and from self to not-self, and keeping in mind the above four guidelines
when analyzing/enquiring into these concepts, it is time to move on to
16 Example one: na hi teṣāṃ subhūte bodhisatvānāṃm ātma-saṃjñā pravartsyate, na satva-

saṃjñā, na jīva-saṃjñā, na pudgala-saṃjñā pravartsyate. (Harrison, 2006: 21–22). Exam-
ple two: dharmo hi bhadanta maudgalyāyana, asatvaḥ satva-rajo-vigataḥ, nairātmyo
rāga-rajo-vigataḥ, nirjīvo jāti-cyuti-vigataḥ, niṣpudgalaḥ pūrvāntâparānta-paricchinnaḥ.
(Takahashi, 2006: 21).
examining perspectives on the person and the self in the ninth chapter of the
Abhidharmakośa.
5 Critiques of Abhidharmic Orientation of Perspectives
In the ninth chapter of the Abhidharmakośa, namely “Refutation of the

Doctrine of the Self” or “An Examination of the Person (or Individual),”
Vasubandhu presents his thoughts on the topics of pudgala and self. Although
Vasubandhu criticizes the Pudgalavādins and the Nyāya-Vaiśeṣika tradition,
they all share at least a very similar orientation of perspectives, which can be
labeled as “Abhidharmic”.
The most prominent characteristic of the Abhidharmic approach is
Abhidharma, i.e. dealing with (abhi-) phenomena (dharmas). If a perspective
is a way of seeing the world, then the Abhidharmic orientation takes concepts
as a starting point and sees the world based on how propositional thoughts are
constructed from these phenomena.
The Abhidharmic orientation is to some extent overshadowed by a reduc-
tion to concepts or conceptual thoughts which distance themselves from the
reality of the world and seemingly occupy a determinate position as concep-
tual entities. Such an orientation is very different from what is found in the
sūtras of the Śrāvakayāna and even in such Mahāyāna scriptures as the
Prajñāpāramitā sūtras, where the observation of what is really going on in the
sentient world is taken as a starting point, and phenomena are mostly regarded
as designations by provisional naming (prajñapti / paññatti) and can be lin-
guistically negated, or discarded, on account of observational insight into
reality and soteriological ascent on Buddhist paths of cultivation. As such, the
orientation of perspectives exemplified in the sūtras can be labeled as “not-
dharmic”. In a sense, the “Abhidharmic” orientation of perspectives is a
concept-based approach, while the “not-dharmic” orientation of perspectives
is a concept-transcending approach.
In the Buddhist tradition, it seems fairly straightforward to lead to the idea
of not-self from the impermanent five aggregates, but this is decidedly more
complicated when the concept of pudgala (person) is considered from the
Abhidharmic perspective. In that mode, different schools of Abhidharma tra-
ditions and their proponents often have little agreement over the issue of
combining the two concepts of pudgala and self.
406 Yao-ming Tsai
For example, both Vasubandhu and the Pudgalavādins can agree that pud-
gala is not the self.17 However, the ontological status of pudgala is itself
disputed. According to Vasubandhu, pudgala exists by reason of sharing a sim-
ilar existential status as collections of the five aggregates. By contrast, the
Pudgalavādins are portrayed as denying that pudgala shares a similar status of
existence as collections of the five aggregates. In particular, the Pudgalavādins
insist that pudgala exists by reason of being neither the same as nor different
from collections of the five aggregates, and therefore its existence is inexplica-
ble (avaktavya).
With the above overview in mind, three critiques can be formulated as to
the perspectives therein.
Critique one: Should pudgala be regarded as an entity possessing existence
as an intrinsic part of its nature? Just because the concept pudgala is a con-
cept, does not mean that pudgala is necessarily something like an entity right
over there. However, it seems that if one adopts the Abhidharmic orientation
of perspectives, taking phenomena as a starting point, and if one deems that
the referents of a great number of concepts – a mountain, a desk, a cat, etc. –
exist, since pudgala is also a concept, then it appears natural that the referent
of the concept pudgala should also exist. Surprisingly, it is in the same vein
that both Vasubandhu and the Pudgalavādins insist on the existence of pudga-
la.18 But let us briefly ponder this problematic view of existence. When one
says that a certain entity exists, one seems to assume that that very entity must
exist in spatio-temporal processes of becoming. If pudgala is taken in its con-
notation as continuous stream, and if pudgala exists, then how and in what
framework does pudgala exist? The existence of pudgala may be challenged
from three related levels. First, if the so-called existent happens to be an entity
and is framed by the spatio-temporal continuum, how can such a spatio-tem-
poral continuum as pudgala exist as an existent in a lower level? In other
words, as a continuous stream, pudgala appears as a continuous stream; but
this continuous stream cannot be downgraded to some level lower than con-
tinuous stream. If it were downgraded to an existent at a lower level it would
have lost its being as a continuous stream. Second, how can such a spatio-tem-
poral continuum as pudgala exist as spatio-temporal continuum at the same
level? In other words, as continuous stream, pudgala cannot be both continu-
ous stream and existent at the same level on account of the incompatibility of
continuous stream permeated by at least four-dimensional changes and exis-
tence remaining unchanged as permanent existence. Third, as continuous
17 Or translated as “a person is not a self”: see Duerlinger, 2003:124, 127, 199, 201–202, 280.
18 See Duerlinger, 2003: 2, 14, 24, 26, 35, 73, 131, 284.
stream, pudgala can only exist in a framework which is not only at a higher
level than pudgala but also encompasses pudgala. Since such a higher frame-
work by which pudgala can be accounted for is unknown to both Vasubandhu
and the Pudgalavādins, it will be pointless for them to say that pudgala exists.
Critique two: Should the assertion of existence/non-existence be applied to
a single item? Just like most of those who follow the Abhidharmic orientation
of perspectives, both Vasubandhu and the Pudgalavādins are inclined to apply
the assertion of existence/non-existence to a single item such as pudgala. Such
a convention is very common among ordinary people and is fully in accor-
dance with standard linguistic expressions. Nevertheless, it is one thing to say
that a single item exists, but quite another thing to examine if this type of
assertion is pertinent to more facets of reality. It is not that the assertion of
existence/non-existence is necessarily wrong, but the assertion is supposed to
reveal a more pertinent perspective on the reality it purports to represent,
rather than being confined within the walls of linguistic isolation and laby-
rinth. From the perspective of dependent origination (pratītyasamutpāda), an
item does not stand alone by itself, but arises from the condition of a priori and
surrounding context. If the existence/non-existence is asserted, it is better to
address the link between a certain item and its context.19 Obviously both
Vasubandhu and the Pudgalavādins fail to take dependent origination into
consideration when asserting the existence of pudgala, and, consequently, fail
to make it clear how pudgala can arise and cease. Without understanding the
mechanism of arising and cessation, the assertion of the existence of pudgala
begins with an attachment to a single conceptual item, and may end up with a
view of eternalism (śāśvatadṛṣṭi).
Critique three: How should the two concepts of pudgala and self be com-
bined to make assertions germane to the sentient world? Although both
Vasubandhu and the Pudgalavādins hold the same view that pudgala is not the
self, their divergent presuppositions, which on the surface appear to comply
with the thesis of not-self, definitely deserve further inspection. It is important
to know that both concepts of pudgala and self are not at the level that can be
perceived or observed. Rather, both concepts are convenient tools for verbal
communication and philosophical discourse. However, if philosophical dis-
courses merely remain at the level of conceptual items formulated in partisan
terms, such discussions will only continue to be entangled in seemingly unsolv-
able disputes. A feasible and better approach is to start by perceiving or
19 The general formula of this perspective runs thus: “When that exists (or is present), this
comes to be; from the arising of that, this arises. When that does not exist (or is not pres-
ent), this does not come to be; from the cessation of that, this ceases”.
408 Yao-ming Tsai
observing such factors as the five aggregates in the flow of life, and gaining
insight into the reality of the five aggregates, and explicating the relation
between the five aggregates, pudgala, and self.20
In order for the aforementioned approach to work effectively, the following
six steps should be implemented. First, perceiving or observing the constant
flux of the five aggregates may provide an insight into the reality that the five
aggregates do not exist as an independent or separate entity. In other words,
the five aggregates are empty (śūnya) of any independent or separate existence
or beingness (bhāvatā; vastutā) of their own. Second, due to fact of the imper-
manence and suffering, none of the five aggregates are qualified to be identified
as the self, whatever the self-concept may mean. In other words, the five aggre-
gates, either altogether or separately, are not the self. Third, if the phenomena
of the five aggregates are taken as individual, person or continuous stream,
then the term pudgala can be employed accordingly. Just as many other terms
can be used as media of indication, connotation or communication, so can the
term pudgala as long as it makes some sense for a given application. Fourth, no
matter what the term pudgala may mean in different contexts, as one of the
alternatives of the five aggregates in linguistic convention, the term pudgala
merely acts as a verbal designation. In reality, just as the five aggregates are
empty of the inherent existence of the five aggregates, so is the designated pud-
gala empty of the inherent existence of pudgala. Fifth, just as the five aggregates
are not the self, pudgala cannot be identified as the self either. Sixth, the rela-
tion between the five aggregates, pudgala, and self can therefore be explicated
simply and concisely as follows.
The five aggregates are at the level that can be and should be perceived or
observed. Although the five aggregates are impermanent and suffering, and in
reality are not qualified to be identified as the self, they are often individually
or collectively regarded as the self in linguistic convention and frequently
encountered in philosophical views as such. Although the five aggregates can
also be alternatively termed as pudgala to match such cognitive impressions as
individual, person or continuous stream, it is unnecessary, or even incorrect to
assert the independent existence of pudgala in itself. Therefore, both concepts
of pudgala and self can be switched to the five aggregates, which can be per-
ceived and understood as empty of such dharmas as the five aggregates. Since
the five aggregates are empty of said dharmas as the five aggregates, the so-
called five aggregates are in reality not the five aggregates. In light of this
20 Cf. bhagavān āha: tat kiṃ manyase subhūte atraiṣā saṃjñā samajñā-prajñapti-vyavahāraḥ
pañcasûpādāna-skandheṣu yad uta bodhisattva iti? (Wogihara, 1932: 71). See also Conze,
1975: 88.
insightful understanding of the emptiness of the five aggregates, the relation

between the five aggregates, pudgala, and self, on the one hand, is equally and
coherently empty without any substantialist or ontological distinction per se
– on the other, it is subject to cognitive differentiation and linguistic
variation.
6 Conclusion: From “Abhidharmic” to “Not-dharmic” Orientation of

Perspectives
The historical development of Buddhism has witnessed a considerable range

of schools, thinkers and texts, which has meant a proliferation of divergent
views and the introduction of more and more technical terms for the purpose
of validating and substantiating those views. Dharmas (phenomena), for
example, are such technical terms, and the scholarly pursuit of dealing with
(abhi-) these phenomena (dharmas) as its first and foremost task begets the
name of “Abhidharma”.
The Abhidharmic orientation of perspectives is inclined to organize, clas-
sify, analyze, and compare phenomena while paying little attention to the
question of firsthand observation in the treatment of the encountered con-
cepts. By neglecting to ascertain whether encountered concepts are eligible
for firsthand observation, and by overlooking the general tendency to regard
them as entities with preconceived expectations, the Abhidharmic orientation
consequently finds itself entangled in such concepts as the five aggregates,
pudgala, and self.
If we do not content ourselves with a mere knowledge of Buddhist history
about what has been said by whom and how it has been interpreted, but rather
relate these views to both the procedure through which they were formed and
to the reality of the sentient world, then, and only then, can we approach the
five aggregates not as phenomena, but as impermanent phenomena of the
sentient world. From the firsthand observation of the impermanence and
dependent origination of the five aggregates comes the insightful understand-
ing of the not-self and emptiness of the five aggregates. It is conventional
knowledge to speak of the five aggregates, pudgala, and self, but it is wise to
state that the so-called five aggregates are in reality not the five aggregates, and
that the five aggregates, pudgala, and self are designations of provisional nam-
ing without any substantialist or ontological distinction per se. This transition
from an “Abhidharmic” to a “not-dharmic” orientation of perspectives is a fun-
damental stance worth appreciating and appropriating within Buddhist
philosophy.
410 Yao-ming Tsai
Abbreviation

邊 and Ono 小野, 1924–1934)
Bibliography
Anālayo Bhikkhu (2010). “Saccaka’s Challenge: A Study of the Saṃyuktāgama Parallel

to the Cūḷasaccaka-sutta in Relation to the Notion of Merit Transfer”, Chung-Hwa
Buddhist Journal, pp. 39–70.
Bodhesako Sāmanera (2008). Beginnings: Collected Essays of S. Bodhesako. Kandy:
Buddhist Publication Society.
Bodhi Bhikkhu (2009). The Connected Discourses of the Buddha. Boston MA: Wisdom.
Burton, David (2004). Buddhism, Knowledge and Liberation: A Philosophical Study.
Aldershot: Ashgate.
Choong, Mun-keat (2000). The Fundamental Teachings of Early Buddhism. Wiesbaden:
Otto Harrassowitz.
Chung, Jin-il (2008). A Survey of the Sanskrit Fragments Corresponding to the Chinese
Samyuktāgama. Tokyo: Sankibo Busshorin.
Summary. Bolinas CA: Four Seasons Foundation.
Duerlinger, James (2003). Indian Buddhist Theories of Persons: Vasubandhu’s “Refutation
of the Theory of a Self”. Abingdon: Routledge Curzon.
Duerlinger, James (2009). “Vasubandhu’s Abhidharmakośa: The Critique of the
Pudgalavādins’ Theory of Persons”, In Buddhist Philosophy: Essential Readings. Edited
by William Edelglass and Jay Garfield. Oxford: Oxford University Press,
pp. 286–296.
Gowans, Christopher (2003). Philosophy of the Buddha. London: Routledge.
Hakamaya, Noriake (1979). “ <法印> 覺え書 (Notes on the Word “Fa-yin 法印)”, Journal
of the Faculty of Buddhism of the Komazawa University 37, pp. 60–81.
Hamilton, Sue (2000). Early Buddhism – A New Approach: The I of the Beholder. Richmond
VA: Curzon.
Harrison, Paul (2006). “Vajracchedikā Prajñāpāramitā: A New English Translation of the
Sanskrit Text Based on Two Manuscripts from Greater Gandhāra”, In Manuscripts in
the Schøyen Collection: Buddhist Manuscripts. Edited by Jens Braarvig. Oslo: Hermes
Publishing, pp. 89–132.
Jaini, Padmanabh S. (1975). The Jaina Path of Purification. Delhi: Motilal Banarsidass.
Kattackal, Jacob (1999). Indian Philosophy: Indian Religious Philosophies from Vedas to
Śankara-Rāmānuja-Śaivasiddhānta. Kottayam: Oriental Institute of Religious
Studies.
Ñāṇamoli Bhikkhu and Bodhi Bhikkhu (2005). The Middle Length Discourses of the
Buddha: A New Translation of the Majjhima Nikāya. Boston MA: Wisdom.
Sastri, N. Aiyaswami (1975). Satyasiddhiśāstra of Harivarman. Vol 1. Baroda: Oriental
Institute, Maharaja Sayajirao University.
Sastri, N. Aiyaswami (1978). Satyasiddhiśāstra of Harivarman. Vol 2. Baroda: Oriental
Institute, Maharaja Sayajirao University.
Sutton, Florin (1991) Existence and Enlightenment in the Laṅkāvatāra-sūtra: A Study in
the Ontology and Epistemology of the Yogācāra school of Mahāyāna Buddhism. Albany
NY: State University of New York Press.
Takahashi, Hisao et al. (ed.) (2006). Vimalakīrtinirdeśa: A Sanskrit Edition based upon
the Manuscript Newly Found at the Potala Palace. Tokyo: Taishō University Press.
- T.2.99: Saṃyuktāgama, Za ahan jing 雜阿含經, Guṇabhadra.
- T.29.1558: Vasubandhu, *Abhidharmakośabhāṣya, Apidamo jushe lun 阿毘達磨俱
舍論, Xuanzang 玄奘.
- T.29.1559: Vasubandhu, *Abhidharmakośabhāṣya, Apidamo jushe shilun 阿毘達磨
俱舍釋論, Paramārtha.
- T.32.1646: Harivarman, Tattvasiddhiśāstra, Chengshi lun 成實論, Kumārajīva.
Tedesco, Pablo A. (1947). “Sanskrit Pudgala: ‘Body; Soul’”, Journal of the American Oriental
Society 67/3, pp. 153–250.
Tsai, Yao-ming (2006). “Interpretative Approach to Navigating Concepts Concerning
the Doctrine of Nondual Middle Way”, National Taiwan University Philosophical
Review 32, pp. 115–166.
Tsai, Yao-ming (2013). “Human Life from the Perspective of Buddhist Medicine: A
Philosophical Inquiry Based on the Vimalakīrti-nirdeśa-sūtra”, In The Conception of
the Human Person in Medicine: Exploring Boundaries between Traditional Chinese and
Western Medicine. Edited by Lukas Kaelin, Tsai Yao-ming et al. Vienna: Verlag
Österreich, pp. 117–139.
Tsai, Yao-ming (2014). “Language As an Instrument of Soteriological Transformation
from the Madhyamaka Perspective”, Asian Philosophy: An International Journal of the
Philosophical Traditions of the East 24/4, pp. 330–345.
Gowans, Christopher (2003). Philosophy of the Buddha. London: Routledge.
Vaidya, Paraśurāma Lakṣmaṇa (ed.) (1958). Avadāna-śataka. (Buddhist Sanskrit Texts,
no. 19). Darbhanga: The Mithila Institute.
412 Yao-ming Tsai
Winchester, Jules (2009). “The Self Concept, Culture and Cultural Identity: An
Examination of the Verbal Expression of the Self Concept in an Intercultural Context”,
The Linguistics Journal: Special Edition – Language, Culture and Identity in Asia. British
Virgin Islands: Linguistics Journal Press, pp. 63–81.
Wogihara, Unrai. (ed.) (1932). Abhisamayālaṃkārālokā Prajñāpāramitāvyākhyā: The
Work of Haribhadra together with the Text Commented on. Tokyo: The Toyo Bunko.
Wynne, Alexander (2007). The Origin of Buddhist Meditation. New York NY: Routledge.
Yoshimizu, Chizuko (2007). Clarification of the Critic of Logic by the Madhyamaka School
in Indian and Tibetan Buddhism. Final Report of Thematic Project, University of
Tsukuba.
Index Index 413
Index
All terms are listed in Roman alphabetical order. When both the Sanskrit and the Pāli term oc-
cur, the Sanskrit term is listed first. Tibetan terms are listed under the first occurring Tibetan
radical letter. Chinese terms are listed according to the order of the pinyin transcription system.
English translations of Buddhist technical terms are indexed under the Sanskrit/Pāli term.
abhāva 354, 355, 356, 370 *Abhidharmahṛdaya (Apitan xin lun 阿毗曇
abhāvavādin 356, 360, 361, 367, 369, 370 心論 ) 93, 284, 302
Abhayagiri 176, 196 Abhidharmakośa (Apidamo jushe lun 阿毗達
abhedyaprasāda / avetyaprasāda / 磨俱舍論 ) 10, 13, 15, 19, 91, 93, 175, 176,
aveccapasāda (four) 7, 8 51, 54, 71, 77, 177,
78, 79, 80, 81, 82, 83, 84, 85, 87, 88, 89, 178, 180, 183, 194, 195, 284, 308, 315, 321, 322,
90, 91, 92, 93, 94, 95, 96, 98, 99 324, 327, 329, 334, 337, 338, 339, 343,
Abhidhamm’-atthasaṅgaha 191, 192, 196, 198, 345, 346, 364, 405
203, 205 Abhidharmakośabhāṣya 9, 15, 20, 95, 158,
Abhidhamm’-atthasaṅgaha Commentary 159, 175, 177, 183, 193, 215, 224, 225, 226,
191, 198, 203 227, 228, 231, 232, 236, 237, 238, 245, 253,
Abhidhammāvatāra 169, 172, 175, 178, 180, 257, 259, 353, 357, 370, 383, 396, 397
187, 190, 191, 199, 210, 218 Abhidharmakośaṭīkā Tattvārthā 227, 228,
Abhidhammāvatāraṭīkā 169, 170, 173, 174, 175, 239
178, 180, 187, 190, 192, 193, 199, 201, 206, Abhidharmamahāvibhāṣā (Apidamo da
207, 209, 213, 216, 218, 219 piposha lun 阿毘達磨大毘婆沙論)
Abhidharma / Abhidhamma ix, x, xii, xiii, 19, 223, 224, 231, 235, 236, 238, 240, 376,
xiv, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 386, 392
15, 16, 17, 18, 19, 20, 29, 30, 31, 32, 33, 34, Abhidharmapiṭaka / Abhidhammapiṭaka 1,
35, 36, 37, 38, 39, 41, 42, 43, 50, 51, 65, 85,
128, 140, 213
88, 90, 91, 94, 98, 99, 108, 119, 120, 128,
Abhidharmasamayapradīpikā 19, 225
129, 132, 135, 140, 141, 169, 170, 176, 177,
Abhidharmasamuccaya 70, 257, 270, 306,
178, 191, 202, 205, 208, 209, 212, 214, 223,
335, 345, 379
224, 225, 226, 227, 228, 229, 230, 231, 234,
Abhidharmasamuccayabhāṣyā 147, 153, 379
236, 237, 239, 240, 241, 243, 244, 245,
Abhidharmasamuccayavyākhyā 228, 306
248, 249, 269, 282, 283, 284, 285, 286,
abhijn͂ ā / abbhijhā 64, 62, 63, 72, 172, 217
287, 288, 289, 291, 298, 299, 300, 301,
302, 303, 304, 305, 306, 307, 308, 310, 311, abhisamaya 231, 386, 693
314, 315, 317, 318, 319, 321, 324, 326, 329, Abhisamayālaṃkāra 59, 60, 61, 70, 83, 84
343, 344, 345, 353, 354, 360, 361, 362, Abhisamayālaṃkārālokā 84
363, 370, 372, 376, 378, 379, 380, 381, 382, Abhisamayālaṃkāravŗtti 59, 60, 61
384, 385, 387, 391, 392, 397, 405, 409 ābhogatā 331
Abhidharma School 13, 282 284 absence of dispositions (see anabhisaṃskāra)
Abhidharmic ix, x, 1, 5, 7, 8, 9,12, 13, 15, 17, acknowledging the cognition of the principle
19, 20, 158, 160, 161, 162, 164, 224, 226, (see dharmajñānakṣānti)
288, 289, 298, 301, 305, 307, 309, 311, 315, acknowledging the cognition of the principle
345, 396, 397, 405, 406, 407, 409 in its subsequent realization (see
Ābhidharmika 15, 31, 38, 201, 223, 227, anvayadharmajñānakṣānti)
234, 291, 298, 299, 300, 301, 325, 356, acquisition (see prāpti)
364, 365, 392 activity (see kāritra)
Abhidharmadīpa 176, 178, 232, 353, 370 actual perception (see pravṛttivijñāna)
© Koninklijke
koninklijke Brill
brill NV,
nv, Leiden,
leiden, 2016 | doi 10.1163/9789004279278_012
10.1163/9789004318823_016
414 Index
adhipatipratyaya 160, 161, 324, 325 ālambanapratyaya 146, 150, 160, 239, 324,
aduḥkhāsukha 68, 334, 335 325, 330
Advaita Vedānta 37, 38 ālambanasaṃpramoṣa 331
aeon (see kalpa) ālayavijñāna 9, 19, 146, 147, 148, 149, 150, 151,
aeon of destruction (see saṃvartakalpa) 152, 154, 156, 157, 158, 160, 163, 164, 229,
aeon of formation (see vivartakalpa) 234, 242, 243
great aeon (see mahākalpa) Amituo jing 阿彌陀經 (see
incalculable aeon (see asaṃkhyeyakalpa) Sukhāvativyūhasūtra)
intermediate aeon (see antarakalpa) Amituo jing tongzang shu 阿彌陀經通贊疏
Afghanistan 29, 35, 48 309
Āgama 6 An Shigao 安世高 11, 248, 249, 250, 251,
aggregate (five) (see skandha) 252, 253, 254, 255, 256, 257, 258, 261,
agitation (see auddhatya) 262, 263, 264, 267, 268, 269, 270, 271,
Ahan koujie shi’er yinyuan jing 阿含口解十 272, 283, 285, 286, 301, 302, 303
二因緣經 251, 252,254, 255, 258, 259, An Xuan 安玄 285
260, 261, 266, 268 anabhisaṃskāra 331
ahrī 205 anābhoga 331
Ājīvika 3 anāgamya 332, 337, 362
ajjhatta- 113, 120, 125, 126, 127, 133 Anālayo Bhikkhu 136
Ānanda 3, 5, 126, 170, 211, 241
ajjhattā (dhammā) 120, 121, 122, 123, 125, 127,
ānantaryamārga 342
141
anapatrāpya / anapatrapa 205
ajjhattabahiddhā 7, 8, 17, 108, 109, 110, 111, 112,
anātmanaḥ 396
113, 114, 117, 118, 119, 120, 121, 123, 128, 129,
Anga-Magadha 36
130, 131, 132, 133, 134, 135, 136, 137, 138,
Aṅguttaranikāya 60, 71, 109, 110, 113, 115, 116,
139, 140, 141, 142
117, 126, 137, 261
ajjhattabahiddhā (dhammā) 114, 116, 120,
anityatā 223, 376, 378, 380, 393, 390, 391, 396,
122, 125, 127
398, 408, 409
ajjhattabahiddhārammaṇā (dhammā) An͂ n͂eyya-sattati-ṭīkā 169, 187, 189, 193
120, 122, 124, 125, 127, 128, 133, 135, 141 annihilationism (see ucchedadṛṣṭi)
ajjhattajjhatte 126 antarakalpa 326
ajjhattaṃ 7, 17, 108, 109, 110, 111, 112, 113, antonyms (pair ajjhattaṃ and bahiddhā) 111,
114, 115, 117, 118, 119, 121, 122, 123, 129, 131, 122, 139, 141
132, 134, 135, 136, 138, 139, 140, 141, 142 anumāṇa 136, 138, 377, 378, 380, 383, 384,
ajjhattārammaṇā (dhammā) 122, 124, 127, 388, 389, 391
128, 133 Anurādhapura 171
ajjhattika 115, 116, 120, 123, 126, 133, 140, Anuruddha 191, 192, 211, 219
217, 254 anuśaya 61, 76, 233, 341
ajjhattikabāhira 115, 117, 123, 134, 140 Anuṭīkā 170, 174, 213, 243
Akaniṣṭha 323 anvayadharmajñāna 342, 343
ākāra 173, 224, 240, 356, 380, 383, 384,385, anvayadharmajñānakṣānti 342
386, 392 ap 90, 176, 177, 183, 255, 287, 315
ākāśa 87 Apidamo da piposha lun 阿毘達磨大毘婆
ākāśānantayāyatana / ākāsānañcāyatana 沙論 (see Abhidharmamahāvibhāṣā)
124, 339 Apidamo xianzong lun 阿毗達磨顯宗論
akiṃcanyāyatana 339 (see *Abhidharmasamayapradīpikā)
akliṣṭājñāna 226, 241 Apidamo jushe lun 阿毗達磨俱舍論 (see
ālambana 239, 330, 384 Abhidharmakośa)
Index 415
Apitan ba jiandu lun 阿毗曇八犍度論 Atthasālinī 5, 119, 120, 121, 122, 123, 125, 126,
(see*Aṣṭagrantha / 127, 128, 134, 135, 141, 178, 185, 190, 196,
*Aṣṭaskandha(śāstra)) 201, 206, 209, 210, 212, 216, 244, 300
Apitan wufa xing jing 阿琵曇五法行經 283 Atthasamāsa 190, 191, 192
Apitan xin lun 阿毗曇心論 Atthuddhāra (see also Aṭṭhakathākaṇḍa) 123,
(see*Abhidharmahṛdaya) 127
apoha 388, 389, 391, 393 auddhatya 86, 211, 212, 213, 331
apramāṇa (four) 11, 52, 53, 54, 55, 57, 61, 62, aupacayika 242
64, 65, 67 69 avadāna 305
apratisaṃkhyānirodha 238, 341 āvajjana 153, 154, 270
artha 69, 379, 380, 388, 389 avaktavya 403, 406
Arthaviniścaya 65, 79 avasthānyathātva 235
ārūpya(samāpatti) four) 52, 53, 54, 55, 62, 64, avidyā 163, 198, 250, 255, 258, 259, 297, 306,
65, 69 341, 366, 371, 376, 377
ārūpyadhātu 324, 332, 339, 342, 383 awareness-holder dominating life 344
āryā meter 176, 197, 206, 207 āyatana 11, 65, 75, 76, 110, 115, 116, 117, 123,
Āryadeva 14, 307, 360, 361, 362, 370, 372 124, 126, 231, 245, 250, 253, 254, 256, 257,
[ārya]satya (four) 54, 55, 57, 62, 65, 68, 77, 259, 260, 264, 265, 266, 267, 268, 269,
85, 86, 98, 109, 110, 180, 193 224, 287, 299, 272 287, 300, 301, 308, 309 328, 329, 330,
342, 354, 363, 380, 382, 383, 384 364, 365, 376, 382
Āryasatyāvatāra 169, 176, 177, 178, 180, 181,
193, 196, 197, 214 Bactria 7, 29, 40, 41, 50, 93
asaṃhārya-dharma 66, 73, 74, 75, 98 bahiddha-
asaṃkhyeyakalpa 327, 328 bahiddhā (dhammā) 114, 116, 120, 121, 125,
127, 135, 141
asaṃskṛta dharma 31, 67, 286, 365
bahiddhārammaṇā (dhammā) 120, 122,
Asaṅga 59, 241, 243, 306, 307, 335, 345, 379
124, 127, 128, 133
asatkāryavāda 359, 360, 372
bahiddhā 7, 17, 108, 109, 110, 111, 112, 113,
Aśoka 5, 65, 90
117, 118, 119, 121, 121, 122, 123, 127, 128, 129,
*Aṣṭagrantha / *Aṣṭaskandha(śāstra) 40, 41,
130, 131, 132, 133, 134, 136, 138, 139, 140,
42, 43, 283, 302
141, 142
Aṣṭasāhasrikā Prajñāpāramitā 49, 50, 62, 63,
bāhira 115, 116, 117, 119, 120, 123, 133, 134,
66, 67, 73, 76, 77, 103
135, 140, 179, 217, 253
astitva 237, 401
Bahuśrutīya 14, 354
Asvabhāva 241
Bajaur Mahāyāna sutra 7, 8, 48, 51, 77, 78, 88,
Aśvaghoṣa 175 89, 94, 98, 99
Aśvakarṇa 320 Bamiyan 47
Atiyoga 319, 329 Bannihuan jing 般泥洹經 253
ātman 158, 359, 363, 365, 367 Bechert, Heinz 170, 176, 177, 184, 189
ātmavāda 38 Beckwith, Christopher I. 39, 40, 41, 42, 43
Ātmavāda-pratiṣedha 397 becoming (see bhava)
atom (see paramāṇu) beingness (see bhāvatā; dharmatā)
attachment (see rāga) benzhi xiangfen本質相分 389
attentiveness (see saṃprajanya) Bhāṣya 177, 214
aṭṭhakathā 113, 141, 175, 178 Bhāṣyaṭīkā 177, 215
Aṭṭhakathā 190 bhava 250, 257, 260
Aṭṭhakathākaṇḍa (see also Atthuddhāra) 123, bhāvanā 212, 381
135, 141 Bhāvanākrama 331, 333, 345, 346, 331
416 Index
bhāvanāmārga 224, 343, 380, 383, 384 buliaoyi 不了義 (see neyārtha)
bhāvānyathātva 235, 240 Butkara I 29
bhāvatā 408
Bhāvaviveka 361 caitasika / cetasika 122, 125, 205, 308, 309,
bīja 147, 203, 206, 207, 223, 234, 242, 359, 381 310
bimba 389, 390, 391 Cakravāḍa 320
birth (see jāti) calm abiding (see śamatha)
bliss (see sukha) Candrakīrti 359, 361, 364, 370, 371, 372
bliss of veritable refinement (see Caraka Saṃhitā 37
praśrabdhisukha) Cardona, George 42
Bodhgayā 170 Catuḥśatakaśāstra (Bai lun 百論 ) 307, 360,
bodhi 60, 73, 74, 286, 288, 321, 375, 376, 379, 371, 370
380, 381, 384, 389, 392 Cāturmahārājakāyika 322
bodhipakṣika-dharma / bodhipakṣya- causal condition (see hetupratyaya)
dharma 53, 54, 55, 63, 64, 65, 67, 75, cause (see hetu)
77, 98, 245, 302 cause of maturation (see vipākahetu)
bodhisattva xiv, 49, 59, 70, 73, 74, 76, 77, 80, Central Asia ix, x, 6, 9, 11, 39, 40, 41, 58, 223,
81, 94, 97, 306, 307, 320, 321, 322, 327, 279
328, 340 cessation (see nirodha)
bodhisattva path 7, 49, 321 cessation due to deliberation (see
body of form (see rūpakāya) pratisaṃkhyānirodha)
Bojjhaṅgasaṃyutta 112 cessation not due to deliberation (see
Brahmā 60, 321, 323 apratisaṃkhyānirodha)
Brahmajālasutta 113 cessation of perceptions and feelings (see
Brahmakāyika 321, 323 samjñāvedayitanirodha)
Brāhmaṇa 3, 86 cetanā 153, 154, 212, 261, 262, 263, 264, 331
Brahmanism 3, 5, 10 Ceylon (see also Sri Lanka) 17, 18, 169, 170, 171,
Brahmanic(al) 5, 17, 29, 32, 36, 37, 38, 175, 176, 202,210, 213
321 chanda 68, 112, 114, 131, 132, 331, 328, 339
Bronkhorst, Johannes 7, 17, 18, 50, 53, 99 chanshu 禪數 301
Buddha ix, x, 1, 3, 4, 5, 6, 8, 13, 30, 31, 49, 50, Chang Ahan shi bao fa jing 長阿含十報法經
51, 71, 73, 79, 80, 81, 82, 83, 88, 89, 91, 92, 59, 250
93, 94, 95, 96, 97, 98, 127, 176, 224, 233, Chang’an 長安 283, 302
240, 241, 248, 258, 266, 279, 282, 297, Cheng weishi lun 成唯識論 228, 229, 239, 241
298, 299, 300, 301, 304, 305, 306, 307, Cheng weishi lun liaoyi deng 成唯識論了義
309, 310, 315, 316, 321, 322, 362, 366, 370, 燈 390
375, 382, 391, 392, 396, 398 Cheng weishi lun shuji 成唯識論述記 376,
Buddhabhūmiśāstra 391 380, 388
Buddhabhumyūpadeśa 377, 378, 379, 382, China ix, x, 10, 11, 13, 15, 18, 223, 224, 227, 270,
392 271, 279, 280, 283, 285, 286, 290, 296,
Buddhadatta 211 301, 302, 305, 310, 393
Buddhaghosa 8, 17, 72, 79, 80, 86, 118, 119, Chos-’byung me-tog snying-po 345
122, 125, 126, 128, 130, 134, 135, 136, 137, Chu 楚 279
138, 140, 141, 174, 183, 185, 194, 195, 197, Chu sanzang ji ji 出三藏記集 279, 283, 289
198, 199, 205, 336 Chunqiu 春秋 281
Buddhamitta 213 cittaikāgratā 333, 334, 335
Buddhapālita 359, 361, 371, 372 cittānupassanā 7, 110
Buddhist logic 377 cittaviprayukta saṃskāra 308
Index 417
Cittuppādakaṇḍa 122, 127 cosmological 13, 280, 314, 318, 319, 320,
class (see jāti) 324, 338, 339, 345
cognition (see jn͂ āna) craving (see tṛṣṇā)
cognizing the principle (see dharmajñāna) Culla-Dhammapāla 211
cognizing the principle in its subsequent cumulative (see aupacayika)
realization (see anvayadharmajñāna) Cūlavaṃsa 169, 170, 171, 213
collocation (syntactic) 108, 109, 110, 111, 114, Cullaniddesa 109
116, 117, 119, 120, 122, 123, 134, 139, 140 cyclic existence (see saṃsāra)
Commentary on the Treasury of the Abhid-
harma (see Abhidharmakośabhāṣya) Da anban shouyi jing 大安般守意經 252
common knowledge (see saṃvṛti) Dao’an 道安 288, 289, 291, 302, 303, 304,
competence 334 305
concentration (see samādhi) Daodejing 道德經 254
concentration of cessation (see Dao di jing 道地經 250, 276
nirodhasamādhi) Daoism 12, 281, 282, 290
conceptual activity (see vikalpa) Da piposha lun 大毗婆沙論 301, 305, 357
conceptualism 376, 377, 392, 393 Dar-ma Khri-’u dum-btsan 346
conceptually existent (see prajñaptisat) darśana / dassana 69, 82, 96, 97, 111, 128, 177,
concomitant cause (see saṃprayuktakahetu) 182, 185, 187, 188, 194, 199, 212, 213, 214,
condition qua cognitive object (see 218, 219, 229, 366, 381
ālambanapratyaya) darśanamārga 224, 323, 342, 343, 383, 384
conditioned factor (see dharma / dhamma; Dārṣṭāntika 10, 177, 223, 225, 230, 231, 232,
saṃskāra) 233, 241, 242
conditioning (see saṃskāra) Dārṣṭāntika-Sautrāntika 174, 225, 230, 231,
Confucian(ism) 11, 12, 17, 18, 279, 280, 281, 233, 234, 239, 242, 243
282, 283, 286, 287, 289, 290, 291, 292 Daśabhūmikasūtra 80, 81
Confucius 12, 266 Dasheng fayuan yilin zhang 大乘法苑義林
consciousness (see vijn͂ āna) 章 298, 376
contamination (see kleśa) Dasheng fayuan yilin zhang buque 大乘法苑
contamination-hindrance (see kleśāvaraṇa) 義林章補闕 376, 389
contemplation (see dhyāna / jhāna) Dasheng yizhang 大乘義章 304
contemplation-as-cause (see Daśottarasūtra (see Chang Ahan shi bao fa
kāraṇadhyāna) jing)
contemplation-as-effect (see deed (see karma)
kāryadhyāna) delusion (see moha)
contemplation of feelings (see dependent origination (see
vedanānupassanā) pratītyasamutpāda)
contemplation of mental objects (see designation by provisional naming (see
dhammānupassanā) prajñapti / paññatti)
contemplation of mind (see desire (see kāma)
cittānupassanā) desire realm (see kāmadhātu)
contemplation of the body (see Dessein, Bart 87
kāyānupassanā) deterioration (see jarā)
conventional truth (see saṃvṛtisatya) Dhammajoti KL Bhikkhu 385
Conze, Edward 62, 63, 316 dhammānupassanā 7, 110, 114
co-occurring cause (see sahabhūhetu) Dhammapada 126
cosmology 318, 319, 320, 321, 322 Dhammapāla 116, 170, 171, 175, 186, 206, 211
Dhammarucikas 170
418 Index
Dhammasaṅgaṇi 120, 121, 122, 123, 124, 125, Ḍhṛtarāṣṭra 322

127, 128, 132, 133, 134, 135, 136, 140, 141, dhyāna / jhāna (four) 8, 53, 54, 55, 57, 60, 62,
210, 212, 244, 299 63, 64, 65, 68, 122, 124, 125, 126, 127, 134,
dharma / dhamma i, 7, 8, 14, 16, 30, 31, 32, 135, 172, 267, 314, 318, 321, 323, 331, 332,
33, 34, 35, 36, 49, 50, 51, 52, 53, 56, 57, 58, 333, 334, 335, 336, 337, 338, 339, 340, 345
59, 61, 64, 67, 68, 69, 70, 71, 73 76 77, 94, dhyānāntara 333, 334, 337, 338, 345
96, 97 ,98, 110, 112, 114, 115, 116, 120, 122, difference in mode of being (see
123, 124, 125, 126, 127, 131, 132, 133, 135, bhāvānyathātva)
138, 141, 152, 172, 173, 183, 184, 201, 202, difference in state (see avasthānyathātva)
206, 211, 213, 223, 230, 231, 234, 235, 236, Dīghanikāya 58, 71, 79, 88, 109, 110, 111, 112,
237, 242, 248, 262, 263, 268, 288, 298, 113, 114, 115, 116, 118, 121, 126, 129
299, 300, 301, 303, 305, 306, 307, 308, Dignāga 272, 388, 389, 391
309, 310, 314, 315, 316, 353, 354, 355, 356, Diṅnāgapāda 175
357, 358, 360, 361 362, 363,364, 365, 366, direct perception (see pratyakṣa)
367, 368, 369, 370, 371, 372, 378, 382, 383, direct realization (see abhisamaya)
385, 390, 405, 408, 409 Dīrghāgama 53, 54, 56, 71, 85, 111
(dhammas) having an external object (see discrimination (see pratisaṃvid /
bahiddhārammaṇā (dhammā))
paṭisambhidā)
(dhammas) having an internal object (see
disregard of consequence (see anapatrāpya)
ajjhattārammaṇā (dhammā))
distinctive contemplation (see dhyānāntara)
insuperable dharma (see asaṃhārya-
diversity 254, 255, 269, 323, 354, 355, 360
dharma)
rDo-rje bkod-pa 341
internal (dhammas) (see: ajjhattā
Dong Zhongshu 董仲舒 280
(dhammā)
dravyadharma 355, 357
internal and external (dhammas) (see
ajjhattabahiddhā (dhammā)) dravyasat 315
(dhammas) having an internal and dṛṣtadharmanirvāṇaprāpta 343
external object (see Duerlinger, James 397
ajjhattabahiddhārammaṇā (dhammā)) duḥkha 68, 70, 72, 76, 94, 197, 198, 199, 216,
Dharma / Dhamma 8, 36, 50, 60, 71, 77, 78, 255, 257, 306, 307, 316, 342, 376, 378,
81,82, 83, 88, 89, 91, 92, 93, 94, 95, 96, 97, 380, 383, 388, 391, 398, 408
98, 248, 297, 301, 310, 314, 362 *Dvādaśadvāraka (see Shi’er men lun 十二門
dharmacitta 368, 369 論)
Dharmaguptaka 10, 58, 65, 66, 98, 111 dvandva compound 8, 91, 115, 117, 139, 140
dharmajñāna 342, 343 dveṣa 233, 325
dharmajñānakṣānti 342 rDzogs-chen 343, 344, 345
dharmakāya 60, 340
dharmalakṣaṇa 365 earth (see pṛthivī)
dharma-nairātmya 396 East Asia x, 1, 16, 17, 18
Dharmarājikā 29 Eastern Jin 晉 dynasty 283-4, 284
Dharmasaṃgraha 59, 76 effect 82, 183, 191, 233, 234, 258, 315, 324, 325,
Dharmaskandha 54, 65, 90, 128, 245 358, 359, 361
Dharmaśreṣṭhin 87, 93, 284 effort (see ābhogatā; vyāyāma)
dharmatā 342, 344 eighteen elements (see dhātu)
Dharmatrāta 86, 87, 93, 94, 231, 235, 236, 284 emanational (see naiḥṣyandika)
dharmāyatanikāni rūpāni 308 emotion (see qing 情 )
dhātu 76, 159, 172, 217, 231, 245, 300, 301, 308, emptiness (see śūnyatā)
309, 328, 329, 330, 364, 365, 376, 382 empty (see śūnya)
Dhātukathā 54, 55, 120, 128 enduring (see sthiti)
Index 419
enlightenment (see bodhi) formless realm (see ārūpyadhātu)

envelopment (see paryavasthāna) four bases of mindfulness (see ṛddhipāda;
epistemology 388, 393 smṛti)
equanimity (see upekṣā) four great kings 321, 322
equanimity of conditionings (see four noble truths (see āryasatya)
saṃskāropekṣā) Frauwallner, Erich 90
equanimity of feelings (see vedanopekṣā) Fu Jian 符堅 283
essenceless (see niḥsvabhāva) Fu Yi 傅毅 279
eternalism (see śāśvatadṛṣṭi) further competence (see karmaṇyatā)
ethereal abode (see śuddhāvāsa)
ethereal class 323, 323 Gandhāra(n) 6, 7, 8, 10, 11, 15, 17, 29, 30, 31,
ethereal level (see śuddhāvāsakāyika) 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43,
examination (see vitarka) 47, 48, 49, 50, 51, 87, 93, 94, 99
examination of the person (see Pudgala- Greater Gandhāra 7, 10, 18, 29, 32, 38, 39,
viniścāya) 50
exegetical method 1, 2, 3, 4, 290, 292 Gāndhārī xiv, 41, 47, 48, 49, 53, 55, 56, 65, 70,
exegetical translation 254 75, 78, 80, 87, 88, 89, 95, 98
existence (see astitva; upapatti) gandharva 322
expedient means (see upāya) Gaoseng zhuan 高僧傳 288
external (see bahiddha-) geng 更 250, 252, 254, 255, 256, 261, 262, 264
external (see bāhira) Gethin, Rupert 79, 179, 195, 496, 199
external (dhammas) (see bahiddhā geyi 格義 12, 17, 18, 287, 288, 289, 290, 291,
(dhammā)) 292, 303, 304
externally (see bahiddhā) Gimello, Robert 296, 297
extinction (see nirvāṇa / nibbāna) Glang-dar-ma 346
goblin (see yakṣa)
Fabao 法寶 226, 227 gocarajjhatte 126
fachu suoshe se 法處所攝色 (see Godakumbura, Charles E. 178
dharmāyatanikāni rūpāni) grasping (see upādāna)
Faxiang school 法相宗 296, 391 great element (see mahābhūta)
factors conducive to awakening (see Griffiths, Paul 149, 151
bodhipakṣika-dharma) Guangwu 光武 279
faith (see śraddhā / saddhā; xin 信 ) Guenther, Herbert V. 316
faculty of faith (see śraddhendriya) *Guṇaprabha 226
fannao zhang 煩惱障 381 Guo Xiang 郭象 289
Fangguang bore jing 放光般若經 255
feeling of bliss (see sukhā vedanā) habitual residue (see vāsanā)
fire (see tejaḥ) hadayavatthu 153, 171
five aggregates (see pañcaskandha) Han 漢 dynasty 11, 12, 250, 253, 264, 265,
Fodijing lun 佛地經論 (see 266, 279, 280, 281, 282, 283, 285, 290
Buddhabhūmyupadeśa) Haribhadra 83, 84, 88
forgetfulness of imaging (see Harivarman 14, 353, 354, 355, 356, 360, 361,
ālambanasaṃpramoṣa) 362, 363, 364, 365, 367, 368, 369, 370,
form (see rūpa) 371, 372, 399
form realm (see rūpadhātu) Harrison, Paul 99
Former Qin 秦 dynasty 283 hatred (see dveṣa)
formless absorption (see ārūpya(samāpatti); heart-base (see hadayavatthu)Heidegger,
samāpatti) Martin 316, 317, 375
420 Index
Hellenistic 6, 7, 17, 39 internal (see ajjhatta-)

hetu 70, 71, 173, 177, 178, 179, 180, 184, 186, 189, internal (see ajjhattika)
190, 191, 193, 194, 195, 197, 198, 199, 200, internal and external (see
204, 209, 210, 211, 214, 215, 233, 234, 257, ajjhattikabāhira)
259, 260, 261, 270, 318 324, 325, 341, 353, internal in internal (see ajjhattajjhatte)
358, 359, 360, 361, 371, 383, 389 internal in oneself (see niyakajjhatte)
hetupratyaya 160, 236, 324, 325, 365, 371 internal in range / object (see visayajjhat-
hetuvidyā 377, 378 te)
higher knowledge (see abhijn͂ ā) internal in sphere (see gocarajjhatte)
homologous cause (see sabhāgahetu) internally (see ajjhattaṃ)
Hongming ji 弘明集 304 internally and externally (see
ajjhattabahiddhā)
hou de zhi 後得智 382
intrinsic nature (see svabhāva)
Hou Han shu 後漢書 282
īrṣyā / issā 207
huahu 化胡 282
Īṣādhara 320
Huayan 華嚴 297
Islamic world 39
Huaisu 懷素 226
Huan 桓 282, 283, 286 Jain(ism) 3, 4, 5, 7, 34, 35, 36, 76
Huan Wen 桓溫 283, 284 Jaini, Padmanabh S. 353
Huanglao 皇老 282 Jambudvīpa / Jambudīpa 83, 171, 320, 321
Huijiao 慧皎 288 Jambukhādakasutta 201
Huirui 慧叡 289 Janavasabhasutta 111, 113, 140
Huiyuan 慧遠 284, 308 jarā 201, 238
Huizhao 慧沼 15, 376, 380, 387, 388, 389, jarāmaraṇa 201, 250, 364
390, 391, 392, 393 jāti 201, 217, 218, 250, 364, 366, 368, 378, 379,
humaneness (see ren 仁 ) 380
Jayatilleke, Kulatissa Nanada 79
identification 15, 139, 228, 397, 398, 399, 400, jealousy (see īrṣyā)
403 Jetavana 176
identity 12, 15, 161, 162, 354, 355, 360, 387, Jizang 吉藏 289
396, 400, 401, 403, 404 jiaxing wei 加行位 381
ignorance (see avidyā) jiandao 見道 381
Imanishi, Junkichi 184 Jin wen jia 今文家 280
immediate condition (see samanantarapra- Jina 36, 366
tyaya) jiujing wei 究竟位 381
impermanence (see anityatā) jñāna 57, 59, 70, 74, 153, 234, 240, 330, 342,
359, 360, 371, 375, 376, 378, 379, 380, 381,
imprint (see vāsanā)
382, 384, 385, 387, 388, 389, 390, 391,
inclination (see chanda)
392, 393
Indo-Greek 32, 39, 50
jñānamūlaka-dharma (four) 53, 71
Indra 322
Jñānaprasthāna 40, 41, 283, 302
inexplicable (see avaktavya) Jn͂ eya-sampatti-ṭīkā 169, 176, 178, 214
inference (see anumāna) Jn͂ eya-saptati 181
initial thought (see vitarka) Jn͂ eya-saptati-bhāṣya 181
insight (see vipaśyanā / vipassanā) Jn͂ eya-saptati-ṭīkā 169, 176, 181, 182, 186, 189,
insight of skillful investigation (see 193, 214
pratyavekṣajñāna) jñeyāvaraṇa 61, 241, 381
instrument of valid cognition (see pramaṇa) Jotipāla 9, 10, 169, 170, 171, 172, 173, 174, 175,
instrumental cause (see kāraṇahetu) 176, 187, 188, 189, 190, 191, 192, 193, 194,
intention (see cetanā) 195, 196, 197, 199, 201, 202, 213, 214
Index 421
joy (see prīti) kṣānti 382, 383

Kučā 11
kalpa 323, 326, 327 Kuiji 窺基 4, 10, 13, 15, 17, 19, 228, 229, 238,
Kalpasūtra 35 239, 296, 297, 298, 300, 301, 305, 306,
kāma 11, 109, 112, 113, 114, 124, 131, 132, 187, 194, 307, 308, 309, 310, 311, 376, 377, 379, 380,
195, 203, 204, 211, 384, 387, 388, 389, 391, 392, 393
216, 256, 258, 264, 265, 266, 267, 268, 270, 271, Kumārajīva 81, 85, 137, 261, 289, 304, 399
272, 322, 323, 333 Kumāralāta 14
kāmachanda 112, 114, 131, 132 kumbhaṇḍa 322
kāmadhātu 125, 157, 321, 322, 323, 332, 336, Kuṇālā 36
342, 383 kuśala-caitta 310
Kamalaśīla 230, 333, 345 Kuṣāṇa 11, 36
Kāmasuttanideśa 109
Kang Senghui 康僧會 265, 286, 303 La Vallee Poussin, Louis de 91, 224, 225, 226,
kāraṇadhyāna 332 230, 231, 239
kāraṇahetu 324, 325 Lai, Whalen 285
kāritra 230, 235, 236, 238, 260, 325 Lamotte, Étienne 98, 226
karma 59, 70, 124, 147, 148, 172, 198, 205, 217, language 6, 8, 9, 29, 31, 47, 48, 69, 87, 88, 117,
257, 258, 259, 260, 261, 264, 270, 403, 170, 174, 245, 258, 260, 266, 268, 270, 280,
306, 343 282, 285, 336, 371, 372
karmaṇyatā 334 Laozi 老子 282, 289
Karunadasa, Yakupitiyage 244 latency (see anuśaya)
kāryadhyāna 332 laukikāgradharma 372, 383
Kashmir 7, 29, 32, 39, 40, 41 laya 331
Kassapa 200 laziness (see kausīdya)
Katsura, Shōryū 354, 372 Lewis, Mark E. 271
Kātyāyana 41 li 禮 287
kausīdya 331 Liao ben sheng si jing了本生死經 260
kāyānupassanā 7, 109, 129, 131 liberation 15, 156, 157, 158, 173, 234, 301, 376,
Khadiraka 320 381, 392, 401
Khandhavibhaṅga 133, 134, 135 liaoyi 了義 (see nītārtha)
Kharoṣṭhī 7, 41, 60 linguistic dharma 32
Khotan 11 sentence 32
Khuddakanikāya 109 speech sound 32, 33
kleśa 147, 150, 156, 157, 158, 162, 163, 164, 217, word 31, 32, 33, 36, 37
224, 228, 229, 230, 233, 234, 241, 257, linguistic existent (see prajñaptisat)
266, 317, 337, 341, 342, 376 377, 379, 381, Liu du ji jing 六度集經 265
392 Liu Jun 劉峻 287
kleśāvaraṇa 241, 381 liu ru 六入 (see āyatana)
kliṣṭamanas 19, 146, 152, 161, 162, 163, 164 Liu Yiqing 劉義慶 287, 303
knowledge (see zhi 智 ) Liu Ying 劉英 279, 282
Kǒryo edition 249, 251, 261, 263 Lokakṣema 62, 63, 285, 302
Kośa masters 227 lokamukha 363
Kośa School 13, 284 lokapāla 322
Kosambī 36 Lou fen bu jing 漏分布經 251
kṣaṇa / khana 33, 218, 224, 243, 244, 326, 330, Lu Xiujing 陸修靜 281
342, 343 Lunyu 論語 289
kṣaṇikatva 31, 34, 37, 223, 356, 357 Luoyang 洛陽 283, 285, 286, 301
422 Index
Lüshi chunqiu 呂氏春秋 285 main contemplation (see mauladhyāna)

Majjhimanikāya 60, 109, 110, 112, 113, 114, 115,
Madhyāntavibhāga 59, 152 116, 117, 118, 121, 126, 129, 152, 248, 258
Madhyamaka 14, 324, 353, 354, 360, 361, 365, manas 146, 152, 153, 155, 158, 159, 160, 161, 162,
367, 369, 372, 377 163, 164, 205, 216, 243, 261
Mādhyamika 14, 230, 354, 362, 364, 366, manovijñāna 138, 151, 152, 153, 154, 155, 158,
367, 369, 370, 371, 372 159, 160, 161, 162, 163, 164, 184, 357
Madhyamakaśāstra 355, 358, 359, 360, 367, Mantrayāna 319, 345
370, 371, 372 Mao shi 毛詩 281, 289
Madhyāntavibhāga-bhāṣya 70, 331 Māra 60, 73, 74, 76, 87, 98, 306, 322
madrasa 40 Martin, Dan 345
Magadha 3, 36, 173, 174 Mathurā 36
Māgadhī 173, 174 mātṛkā / mātikā 1, 5, 6, 8, 9, 12, 13, 14, 15, 17,
Mahābhāṣya 32, 33, 42, 43, 50 120, 121, 123, 127, 210, 286, 299, 305
mahābhūta 86, 90, 227, 287, 357 mātsarya / macchariya 207
Mahābrahmā 323, 338 matter (see rūpa)
Mahāhatthipadopamasutta 116 matter derived from the four great elements
Mahākalpa 327 (see upādāyarūpa)
Mahānidānasūtra (Ren ben yu sheng jing 人 maturation (see vipākaphala)
本欲生經 ) 250 mauladhyāna 337
Mahāniddesa 109, 113, 115 Mauryan empire 32, 39
Mahāprajñāpāramitāśāstra 69, 97, 137, 138 meaning (see artha)
meditative praxis 13, 314, 317, 345
Mahāpuṇṇamasutta 112
Menander / Menandros 6, 30
Mahāsaṃgha 171
mental perception (see manovijñāna)
Mahāsaṃghika 4, 5, 14, 19, 171, 185, 241,
metaphysics 316, 317, 318, 326
354
metaphysical 13, 15, 16, 290, 291, 314, 316,
Mahāṭīkā 186, 201, 202, 204, 206, 209, 211
317, 318, 319, 321, 345, 376, 377, 397, 398,
Mahāvaṃsaṭīkā 178
399, 400, 404
Mahāvastu 60
mi-lcogs med-pa 337, 345
Mahāvibhāṣā 41, 85, 87, 91, 94, 193, 305
Mian Kili 7, 48
*Mahāvibhāṣāśāstra (see Da piposha lun 大 miao guancha zhi 妙觀察智 382
毗婆沙論 ) Milinda 6, 31
Mahāvihāra 171, 176 Milindapañha 6, 30, 31, 335
Mahāvihāravāsin 9, 18, 169, 170, 176, 193, 196, mind (see manas)
202, 213 Ming 明 11, 279
Mahāyāna 7, 8, 9, 10, 12, 15, 16, 19, 29, 47, 48, miserliness (see mātsarya / macchariya)
51, 60, 64, 69, 96, 98, 99, 170, 171, 177, Miyazaki, Izumi 331
202, 223, 224, 225, 240, 241, 242, 299, mode (see ākāra)
300, 301, 304, 305, 306, 310, 370, 405 mofa 末法 297
Mahāyāna, early 7, 29, 47, 48,49, 50, 66, moha 199, 200, 201, 204, 233, 321 355
67, 98, 99, 369 Mohavicchedanī 200
Mahāyānasaṃgraha 19, 146, 147, 148, 149, moment (see kṣaṇa; samaya)
150, 151, 152, 156, 157, 158, 159, 161, 162, moment of completing an action
163, 164, 241, 242, 243 (bya-rdzogs-kyi skad-cig-ma) 326
Mahāyānasūtrālaṃkāra 59, 70, 154, 299 infinitesimal moment (dus-mtha’i
Mahāyoga 320, 327, 343, 344 skad-cig-ma) 326
Mahīśāsaka 228 momentariness (see kṣaṇikatva)
Index 423
morality (see śīla) nirodhasamāpatti 146, 147, 163, 229, 265

motanlijia 磨呾理迦 (see mātṛkā / mātikā) nirodhasatya 354, 363
Mt. Lu 廬 283, 284 nirupadhiśeṣanirvāṇa 343, 344
Mūlasarvāstivādavinaya 177 nirvāṇa / nibbāna 94, 123, 125, 180, 181, 182,
Mūlasarvāstivādin 171, 177 183, 184, 197, 198, 199, 200, 201, 214, 297,
Mūlaṭīkā 170, 174, 201, 211 299, 315, 316, 341, 346, 362, 364, 365, 381
mūrdhan 382, 383 nirvedha 381
Murti, T.R.V. 38 niṣṭhā 356, 381
nītārtha 4, 201, 232, 300, 301, 305, 311
nāga 176, 322, nityatā 223, 236
Nāgabodhithera 176, niyakajjhatte 121, 126
Nāgārjuna 14, 97, 98, 100, 103, 104, 106, 144, noema 389, 390
246, 292, 294, 306, 326, 348, 355, 358, noema qua basic stuff (see bimba)
359, 360, 361, 362, 369, 370, 372, 373, 374 noema qua image (see pratibimba)
Nāgasena 6, 23, 31, 335 noesis 390
naiḥṣyandika 242, Nolot, Édith 6
nairātmya 15, 223, 378, 396, nominalism 376, 377, 392, 393
naivasaṃjñānāsaṃjñāyatana / non-conditioned factor (see asaṃskṛta
nevasaññānāsaññāyatana 124, 125, dharma)
non-effort (see anābhoga)
135, 339, 341
non-existence (see nāstitva)
nāmarūpa 250, 252, 300
non-identifiability 400, 401, 404
name-and-form (see nāmarūpa)
non-identity 400, 401, 403, 404
Ñāṇamoli Bhikkhu 60, 69, 116, 134, 135, 144,
not-dharmic 15, 396, 397, 405, 409
194, 202, 215, 263, 398
not-self (see anātmanaḥ)
nāstika 356, 360, 361, 367, 368, 369, 370, 371,
not-self of factors (see dharma-nairātmya)
372,
gNubs-chen Sangs-rgyas ye-shes 20, 314, 319,
nāstitva 401 332, 338, 344, 345, 346
neither pleasure nor suffering (see gNubs-chen 331, 343, 345, 346,
aduḥkhāsukha) Sangs-rgyas ye-shes 320, 331, 338, 339,
Neo-Confucianism 290, 291 344, 345,
Neo-Daoism 282 numerical categories of dharmas in
Neo-Sarvāstivāda 10, 225, 229, 235, 238, 239, meditation (see chanshu)
240 Nyang-ral nyi-ma ’od-zer 345,
neyārtha 200, 232, 300, 301 Nyang-ral 345, 346
nihilism 367, 369, 370 Nyāya xiii, 37, 359, 367, 383, 397, 405
niḥsvabhāva 362, 371 Nyāya-Vaiśeṣika 359, 397, 405
Nikāya 6, 7, 17, 108, 109, 110, 111, 112, 113, 117, Nyāyamukha 377, 395
118, 119, 120, 121, 122, 123, 132, 133, 134, *Nyāyānusāra 19, 225, 226, 227, 228, 229,
136, 137, 139, 140, 141, 171, 198 230, 231, 232, 234, 238, 239, 240, 241, 242,
Nikāyasaṃgraha 171 245, 247, 353, 370,
Nikkhepakaṇḍa 121, 122, 123 Nyāyapraveśa 377
Nimiṃdhara 320
nimitta 97, 140, 189, 205, 336, 366, object of cognition (see ālambana)
Nirmāṇarati 323 old-age and death (see jarāmaraṇa)
nirodha 68, 70, 73, 81, 197, 198, 229, 306, 340, one-pointedness of mind (see cittaikāgratā)
341, 342, 358, 363, 365, 366, 367, 383, 401, ontology 30, 32, 33, 37, 39, 50, 362, 376, 380,
407 384, 387, 388, 389, 392, 393,
nirodhasamādhi 340 ontological difference 317
424 Index
Pakistan 29, 48 perspective 396, 397, 399, 403, 404, 405, 406,
pañcaskandha 9, 15, 110, 112, 121, 182, 193, 259, 409,
263, 287, 300, 301, 308, 328, 355, 362, Peṭaka 180, 222
363, 364, 365, 366, 367, 376, 382, 384, Peṭakopadesa 261, 263, 273,
390, 398, 399, 400, 403 405 406 408 409 phenomenological 13, 20, 258, 314, 319, 324,
Pañcaskandhabhāṣya 161 326, 345, 384, 385, 387
Pañcaskandhaka 9, 19, 146, 148, 149, 150, 152, phenomenon (see dharma)
153, 155, 156, 157, 158, 160, 161, 163, 164 Polonnaruva 213
Pañcaskandhakavibhāṣā 19, 146, 147, 148, Porāṇas 186, 187
149, 150, 152, 153, 154, 155, 156, 157, 158, post-enlightened cognition (see
160, 163, 164 pṛṣṭhalabdhajñāna)
*Pañcaskandhavivaraṇa 161 prabandha 227
pañcavastuka 300 pradeśa 34
Pañcaviṃśatisāhasrikā Prajñāpāramitā 59, prahāṇamārga 342
62, 63, 64, 65, 77, 100, 223, 276, 305 prajñā / paññā 69, 70, 73, 78, 79, 82, 86, 96,
Pāṇini 32, 33, 34, 113, 119, 226, 240, 291, 315, 329, 341, 383,
panjiao 判教 297, 311 385, 386, 387, 391
Papañcasūdanī 116, 118, 126, 128, 134, 135, 138, prajñāpāramitā 74, 77, 99,
142 Prajñāpāramitāsūtra 48, 101, 103, 305, 321
Parākramabāhu II 169, 170, 174, 220, prajñapti / paññatti 174, 355, 362, 363, 369,
paramāṇu 34, 228, 239, 240, 326 372, 389, 390, 393, 405, 408
paramārtha 300, 362, 363, 364, 365, 366, prajñapticitta 365,
367, 368, prajñaptisat 315, 316, 355, 362, 363, 364
paramārthamukha 363 pramāna 359, 377
paramārthasat 363, 366, Pramāṇasamuccaya 388
paramārthasatya 299-300, 362, 363, 364, prāpti 229, 264
365, 368 prasāda 79, 85, 86, 92, 93, 334
Param’-atthavinicchaya 191 prasaṅga 359, 369, 372
Paranirmitavaśavartin 322, 323 Prasannapadā 359, 373
pariṇāma 233, 235, 242 praśrabdhi 331, 334
Parthia ii, 283, 285, 301 praśrabdhisukha 336
particular characteristic (see svalakṣaṇa) pratibimba 371, 389, 389-390, 391, 393
particulars 376, 378, 380, 389, 394 pratimokṣa 321
parts and the whole 354, 356, 356, 358, 360, pratipad (four) 57, 65, 68, 72,
372, 387 pratisaṃkhyānirodha 341
paryavasthāna 233 pratisaṃvid / paṭisambhidā (four) 54, 61, 63,
passing away (see vyaya) 64, 65, 69, 70, 77, 173, 174, 212
Patan͂ jali 34, 42, 50, pratītyasamutpāda 34, 65, 75, 148, 229, 245,
path of cultivation (see bhāvanāmārga) 248, 249, 251, 252, 253, 254, 255, 256,
path of relinquishing (see prahāṇamārga) 257, 258, 259, 260, 262, 263, 264, 267,
path of seeing (see darśanamārga) 268, 269, 270, 272, 287, 299, 366, 367,
path of utter liberation (see vimuktimārga) 369, 370, 371, 372, 401, 407, 409,
path without obstacle (see ānantaryamārga) pratyakṣa 231, 234, 235, 377, 380, 384, 388,
Paṭṭhāna 120, 128, 189, 190, 191, 192 390
Pengcheng 彭城 279 pratyavekṣajñāna 382
perception (see saṃjn͂ ā; vijn͂ āna) pravṛttivijñāna 146, 148, 150, 151, 152, 153, 154,
permanence (see nityatā) 155, 156
person (see pudgala) prayoga 70, 381, 382
Index 425
precursory level (see sāmantaka) ruling condition (see adhipatipratyaya)

presentation (see samudācāra) rūpa 72, 73, 112, 115, 116, 119, 121, 122, 123, 124,
principial body (see dharmakāya) 125, 127, 128, 133, 134, 160, 172, 174, 178,
prīti 68, 69, 115, 207, 265, 332, 333, 334, 335, 179, 180, 182, 183, 184, 185, 186, 187, 188,
336, 338 193, 194, 211, 214, 216, 217, 239, 244, 256,
proclivity (see anuśaya) 258, 262, 263, 308, 309, 325, 328, 330,
progress (see pratipad) 340, 355, 357, 362, 363, 364, 365, 366,
protector of the world (see lokapāla) 367, 368, 369, 381, 398, 399
provisional naming (see prajñapti / paññatti) rūpadhātu 124, 125, 157, 323, 332, 329, 342,
pṛṣṭhalabdhajñāna 382 383
pṛthivī 90, 116, 122, 237, 255, 258, 287, 315 Rūpakaṇḍa 122, 123
Przyluski, Jean 2, 232 rūpakāya 340
pudgala 15, 35, 74, 171, 353, 396, 397, 402, 403,
404, 405, 406, 407, 408, 409, sabhāgahetu 324
pudgala-nairātmya 15, 396 Saccasaṅkhepa 169, 178, 186, 224, 225
Pudgalavādin 15, 35, 171, 396, 397, 405, 406, ṣaḍāyatana 250, 253
407, Saddhātissa, Hammalawa 196
Pudgala-viniścaya 397, 405 Sāgalikas 170
Puguang 普光 25, 226, 227, 247 sahabhūhetu 234, 235, 240, 241, 324
Puruṣapura 15 Śakra 83, 321, 322
Puṣyamitra 32 sākṣīkaraṇīya-dharma (four) 58, 73
pṛthagjana 365, 402 samādhi 68, 70, 82, 89, 96, 111, 115, 126, 127,
198, 307, 334, 335, 338
Qi chu san guan jing 七處三觀經 251, 276 samādhibhāvanā (four) 58
Qin Jing 秦景 11, 279 samanantarapratyaya 155, 159, 160, 325
qin suoyuanyuan 親所緣緣 390 sāmantaka 337, 345
qing 情 265, 266, 267, 268, 270 sāmānyalakṣaṇa 15, 300, 315, 375, 376, 377,
qingtan 清談 282, 287, 288, 290, 291 378, 379, 380, 381, 382, 383 384, 385, 386,
387, 388, 389, 390, 391, 392, 393
rāga 200, 201, 233, 258, 301, 306, 307, 404, 407 samāpatti 147, 148, 318, 332, 334, 339, 340
Rāhulapāda 175, 176, śamatha 331, 332, 340
rapture (see saumanasya) samaya 34, 117
rational inquiry 1, 2, 3, 9, 11, 14 Samayabhedoparacanacakra (Yibuzong lun
rational tradition 1, 2, 3, 7 lun 異部宗輪論) 4
ṛddhipāda (four) 8, 53, 54, 57, 62, 65, 68 sambhāra 186, 381
realism 367, 376, 377 Saṃdhinirmocanasūtra 154, 155, 327
receptivity (see kṣānti) Saṃgha 5, 8, 10, 60, 71, 81, 82, 83, 88, 91, 92
recollection (see smṛti) Saṃghabhadra 10, 19, 223, 225, 226, 227, 228,
recursive argument method 39, 40, 41, 42 229, 230, 231, 232, 233, 234, 235, 236,
release (see pratimokṣa) 237, 238, 239, 240, 241, 242, 302, 353, 356
ren 仁 281, 286, 287 Saṃghadeva 283, 284, 302
Ren ben yu sheng jing人本欲生經 (see saṃghārāma 40
Mahānidānasūtra) Saṃghavarman 87, 106, 284
retribution-born (see vipākaja) saṃgīti 5
right understanding (see samyag-avabodha; Saṃgītiparyāya 25, 58, 90, 106
samyag-avagama) Saṃgītisūtra 25, 53, 54, 55, 56, 57, 58, 65, 66,
righteousness (see yi 義 ) 67, 70, 71, 77, 89
ritual (see li 禮 ) saṃgraha 300, 307, 310
426 Index
bSam-gtan mig-sgron 13, 20, 314, 315, 319, sarvatraga-caitta 310

320, 321, 322, 324, 326, 327, 328, 329, 331, sarvatragahetu 325
332, 337, 338, 339, 341, 343, 344, 345, sāsrava 64, 76, 226, 242
346, Sastri, Aiyaswami 353
Sāmitīyas 171 śāśvatadṛṣṭi 401, 407
saṃjñā / saññā (four) 7, 55, 97, 121, 194, 195, Śatasāhasrikā 62, 63, 64
308, 328, 339, 364, 398, 399, 404, 408 satipaṭṭhāna 7, 8, 54, 108, 109, 110, 111, 112, 114,
samjñāvedayitanirodha 340, 341 115, 117, 118, 119, 120, 121, 129, 130, 132, 134,
Sāṃkhya 235, 359, 372 136, 137, 139, 140, 141, 142, 143, 144
Sammohavinodanī 120, 125, 128, 134, 135, 136, Satipaṭṭhānasutta 7, 108, 109, 110, 111, 112, 114,
141, 142 115, 118, 121, 129, 136, 137,
saṃprajanya 331, 333, 334, 335 Satipaṭṭhānavibhaṅga 121, 129, 130, 132, 134,
samprasāda 333, 334 141
saṃprayoga 300 satkāryavāda 359, 360, 372
saṃprayuktakahetu 324, 325 satya / sacca / saca 54, 55, 60, 65, 67, 91, 197,
saṃsāra 147, 156, 158, 194, 197, 198, 297, 319, 287, 300, 354, 362, 363, 364, 368
321, 325, 366 Satyāvatāra 169, 176, 177, 178, 180, 181, 193,
saṃskāra 233, 250, 251, 252, 253, 255, 258, 196, 197, 199, 214
259, 260, 261, 262, 263, 264, 269, 308, saumanasya 334
Saundarananda 175
328, 331, 263, 366, 398, 399
Sautrāntika 8, 10, 14, 17, 174, 177, 223, 225,
saṃskāropekṣā 335, 336
230, 231, 232, 233, 234, 239, 241, 242,
samudācāra 208, 209, 381
243,244, 299, 324, 340, 354, 356, 377
saṃvartakalpa 326
Sautrāntika Madhyamaka 324, 377
saṃvṛti 71
scent-eater (see gandharva)
saṃvṛtisat 363, 364
Scharfe, Hartmut 42
saṃvṛtisatya 300, 362, 363, 364, 365, 368
Schmithausen, Lambert 111, 130, 136, 137, 138,
samyag-avabodha 400 146, 147, 148, 151, 152, 153, 154, 155, 157,
samyag-avagama 400 159, 161, 162, 163, 164
samyakpradhāna (four) 53, 54, 65 scrutiny (see vicāra)
*Saṃyuktābhidharmahṛdaya (Za apitan xin seed (see bīja)
lun 雜阿毗曇心論) 86, 87, 94, 106, self (see ātman)
284, 294 self-confidence (see vaiśāradya)
Saṃyuktāgama 85, 90, 91, 398, 400, 401 self-identification 14, 280
Saṃyuttanikāya 90, 109, 112, 113, 126, 254, selflessness (see nairātmya)
257, 258 selflessness of the person (pudgala-
Śaṅkara 37, 38 nairātmya)
sanne 9, 10, 174, 175, 192 Sengrui 僧叡 289
santati 34, 233, 242 Sengyou 僧祐 25, 279, 283, 289, 294, 304,
Śāriputra / Sāriputta 7, 49, 50, 51, 52, 77, 95, 312,
96, 106, 299, 322 sensation (see vedanā)
*Śāriputrābhidharma (Shelifo apitan lun 舍利 sense contact (see sparśa)
佛阿毘曇論) 244-5, 245 sense-element (see dhātu)
Sarvāstivāda 90, 94 sense organ 11, 126, 158, 253, 254, 256, 257,
Sarvāstivādin 6, 7, 8, 9, 10, 14, 29, 41, 54, 260, 265, 266, 267, 268, 270, 271, 300
77, 90, 98, 99, 171, 176, 177, 194, 202, 223, sense perception 11, 149, 151, 152, 154, 155, 59,
224, 229, 231, 234, 235, 239, 241, 244, 245, 160, 161, 163, 249, 253, 256, 258, 264, 266,
274, 303, 305, 393 267, 269, 270, 271, 272
Sarvatathāgatādhiṣṭhānavyūhasūtra 82 sense-sphere (see āyatana)
Index 427
sensual desire (see kāmachanda) Southern India 10, 18, 170

seraphic mansion 323 space (see ākāśa)
serpent-spirit (see nāga) sparśa 11, 249, 250, 251, 252, 253, 254, 255,
seven mountains 320 256, 257, 258, 260, 261, 262, 264, 267,
shamelessness (see ahrī) 268, 269, 270, 271, 272
she 攝 (see saṃgraha) sphere of infinite consciousness (see
Shelifo apitan lun 舍利佛阿毘曇論 (see vijñānānantyāyatana)
*Śāriputrābhidharma) sphere of infinite space (see
Shentai 神泰 226 ākāśānañcāyatana /
Shi’er men jing 十二門經 253 ākāsānañcāyatana)
Shi’er men lun 十二門論 289, 294 sphere of neither perception nor non-percep-
Shi’er men lun shu 十二門論疏 289, 294 tion (see naivasaṃjñānāsaṃjñāyatana;
Shijing 詩經 281, 289 nevasaññānāsaññāyatana)
Shi shuo xin yu 世說新語 287, 303 sphere of nothing whatsoever (see
shou 受 237, 249, 250, 251, 253, 256, 257, 263, akiṃcanyāyatana)
267 Sphuṭārthā Abhidharmakośavyākhyā 224,
shu suoyuanyuan 疏所緣緣 390 232, 245
Shujing書經 281 spontaneous presence 344
Shun zhengli lun 順正理論 (See śraddhā 70, 71, 84, 85, 86, 92,
*Nyāyānusāra) śraddhendriya 344
Shuo wen jie zi 說文解字 280 sragdharā meter 196
sign (see nimitta) Śrāvakayāna 8, 10, 96, 223, 398, 400, 419
sil-bu’i dus 346 Sri Lanka (see also Ceylon) x, 6, 9, 30, 130, 206,
śīla 8, 71, 84, 85, 86, 88, 89, 90, 91, 92, 93, 94, 207, 243, 331, 356
95, 96, 97, 99,185, 199 Śrīlāta 228, 231, 232, 233, 239, 240, 242
Sinification (of Buddhism) 296 srotāpattyaṅga (four) 71, 88, 89, 90, 91
sinking (see laya) stability without inadequacy (see anāgamya)
six sense-spheres (see āyatana; ṣaḍāyatana) stage of application (see prayoga)
sixteen moments of wisdom on the path of stage of completion (see niṣṭhā)
seeing (see stage of cultivation (see bhāvanā)
ṣoḍaśadarśanamārgakṣaṇa) stage of equipment (see sambhāra)
skandha 9, 15, 31, 34, 35, 36, 37, 65, 72, 75, 76, stage of penetration (see nirvedha)
110, 112, 118, 119, 121, 135, 140, 161, 180, 181, Sthavira 170, 171, 176, 231, 232
182, 183, 184, 193, 194, 195, 201, 214, 231, Mahāyāna Sthavira 170
245, 252, 253, 258, 259, 261, 262, 263, Sthaviravādins 5, 171, 244, 245
287, 300, 301, 308, 309, 328, 329, 339, Sthiramati 146, 148, 150, 152, 153, 154, 155, 157,
343, 344, 345, 362, 363, 364, 365, 366, 160, 163, 165, 227, 228, 229, 239, 306, 379,
367, 376, 382, 384, 385, 386, 390, 398, 395
399, 400, 403, 405, 406, 408, 409 sthiti 49, 238
Smith, Helmer 189 store mind (see ālayavijn͂ āna)
smṛti / sati 68, 73, 82, 115, 331, 333, 335, 336 stūpa 29, 36
Smṛti 345 substantial existent (see dravyasat)
Smṛtijñānakīrti 345 substantial thing (see dravyadharma)
smṛtyupasthāna (four) 53, 56, 57, 62, 65, 253 Sudarśana 320
ṣoḍaśadarśanamārgakṣaṇa 342 śuddhāvāsa 321, 323
Sogdiana 11 śuddhāvāsakāyika 323
sopadhiśeṣanirvāṇa 343, 344 suffering (see duḥkha)
Southeast Asia 214 Sui 隋 dynasty 290
428 Index
sukha 68, 69, 114, 130, 131, 216, 309, 334, 335, tathatā 378, 379, 380
336, 337, 338, 339 *Tattvasiddhiśāstra 14, 19, 353, 354, 355, 356,
sukhā vedanā 130, 131, 336, 338 357, 358, 359, 360, 361, 362, 363, 364,
Sukhāvativyūhasūtra 309 365, 366, 367, 368, 369
Sumaṅgala 5, 169, 170, 173, 174, 175, 187, 189, Taxila 29
190, 191, 192, 193, 199, 205, 207, 209, 213 tejaḥ 90, 205, 207, 287, 315, 323, 330, 388
Sumaṅgalavilāsinī 118 Thaneshwar 35, 36
(Su)Meru 319, 320, 322 Theravāda 6, 8, 17, 30, 58, 65, 66, 118, 119, 128,
summit (see mūrdhan) 136, 140, 141, 153, 165, 170, 184, 211, 225,
śūnya 14, 97, 136, 197, 217, 306, 342, 362, 363, 243, 244, 245, 299, 308, 338
366, 367, 368, 369, 371, 372, 383, 408, Theravādin 9, 98, 169, 171, 175, 176, 178,
409 184, 191, 196, 244
śūnyatā 62, 74, 95, 97, 98, 99, 126, 307, 329, Theriya 10, 169, 170, 194, 196, 202, 203, 205,
353, 356, 360, 362, 364, 365, 366, 368, 206, 208, 210, 214
369, 370, 371, 372, 378, 381, 383, 409 third order chiliocosm (see
śūnyatācitta 365, 369 trisāhasramahāsāhasralokadhātu)
suo zhi zhang 所知障 381 thirst / thirsting (see tṛṣṇā)
sureness (see prasāda) thorough limpidity (see samprasāda)
sustained contemplation (see vicāra)
thought concomitant (see caitasika)
sūtra / sutta xiv, 1, 3, 4, 7, 8, 13, 47, 48, 49, 50,
Three Kingdoms 253, 265, 266
51, 54, 70, 77, 78, 86, 88, 89, 90, 92, 94,
three types of consciousness 365
95, 97, 98, 99, 128, 180, 183, 198, 201, 229,
threefold nature (see trisvabhāva)
230, 232, 233, 282, 288, 292, 299, 300,
Thūṇā (see Sthūṇā)
301, 304, 305, 310, 315, 342, 356, 366, 398,
thusness (see tathatā)
400, 405,
tianming 天命 280
Sūtrakṛtāṅga, 34
*Sūtrālaṃkāravṛttibhāṣya 154, 162 Tiantai 天臺 297
Sūtrayāna 345 tianzi 天子 280
Sūyagaḍa 34, 35 ṭīkā literature 169, 171, 203, 206, 213, 243
svabhāva 14, 92, 226, 235, 236, 237, 240, 300, time 17, 34, 35, 118, 138, 184, 236, 237, 318, 326,
326, 362, 327, 340, 343, 353, 403
svalakṣaṇa 15, 74, 300, 314, 355, 375, 376, 377, Tīrthika 15, 74, 76, 367, 396, 397
378, 379, 380, 381, 382, 383, 384, 385, 387, tong yang 痛癢 250, 252, 253
388, 389, 390 tongda 通達 (see nirvedha)
svarūpa 237, 371, 377, 380, 388 transcendence obtained visibly in this life
Śvetāmbara 34, 35 (see dṛṣtadharmanirvāṇaprāpta)
synod (see saṃgīti) transcendence with a remainder of the
syntagm 17, 108, 117, 139, 140 aggregates (see sopadhiśeṣanirvāṇa)
transcendence without remainder of the
Taiping dao 太平道 281 aggregates (see nirupadhiśeṣanirvāṇa)
Taizi ruiying benqi jing 太子瑞應本起經 transformation (see pariṇāma)
252, 264, 276 Trāyastriṃśa 321, 322, 323
Tamil country 9, 17, 170, 171 triad (ajjhattaṃ, bahiddhā, ajjhattabahiddhā)
Tang 唐 dynasty 13, 229, 284, 290, 298 109, 110, 112, 113, 117, 119, 133, 139, 140, 141
tantra xiii Triṃśikā 152, 154, 155, 321, 323
tantric 320, 321, 329, 341, 343, 344, Triṃśikāvijñaptibhāṣya 152, 155, 165, 229, 239,
tathāgata 51, 59, 64, 65, 67, 74, 77, 98 245
Tathāgata 7, 60, 63, 75, 95, 77, 95, 126, 127, Tripiṭaka / Tipiṭaka 1, 108, 111, 126, 130, 229
304, 306, 321, 365, 368 trisāhasramahāsāhasralokadhātu 319
Index 429
trisvabhāva 390 332, 342, 345, 353, 357, 383, 396, 397,
tṛṣṇā 53, 71, 116, 197, 198, 199, 205, 250, 254, 398, 405, 406, 407
255, 256, 257, 258, 260, 264, 266, 267, Vātsīputrīya 245, 353, 397
268, 270, 272, 316, 332, 333, 334 vāyu 183, 287, 315, 320, 323, 327
tṛṣṇotpāda / tṛṣṇāmūlaka-dharma (four) 57, Veda 117,
71 Vedic xii, 3, 4, 5, 8, 17, 117, 139
truth (see satya; tathatā) vedanā 66, 110, 121, 129, 130, 131, 133, 135, 182,
Tunnyun 229, 194, 195, 204, 235, 237, 250, 251, 252, 253,
Tuṣita 321, 322, 323 258, 260, 263, 268, 272, 308, 328, 333,
twelve sense-spheres (see āyatana) 336, 362, 367, 398, 399
two truths 362, 363, 364, 369, 370, 372, 384 vedanānupassanā 7, 110
vedanopekṣā 335, 336, 337
ubiquitous cause (see sarvatragahetu) veritable refinement (see praśrabdhisukha)
ucchedadṛṣṭi 401 Vibhaṅga 54, 55, 65, 69, 71, 100, 120, 121, 128,
Udāna 126 129, 130, 131, 132, 133, 134, 135, 139, 140,
ultimate truth (see paramārthasatya) 141, 174, 181, 212, 245, 268, 299
unbreakable confidence (see abhedyaprasāda vibhāṣā 10, 40, 41, 42, 43, 231
/ avetyaprasāda / aveccapasāda) Vibhāṣā 40, 41, 42, 42, 231
understanding (see prajñā / paññā)
vicāra 68, 218, 238, 332, 333, 334, 335, 336,
unity 240, 354, 355, 357, 360
337, 338, 340, 341
universal characteristics (see
Vidiśā 9
sāmānyalakṣaṇa)
Vigrahavyāvartanī 355, 358, 359, 362, 367
unlimited (see apramāṇa)
vihāra 40
upādāna 121, 198, 249, 250, 251, 253, 256, 257,
vijñāna 53, 57, 72, 124, 138, 148, 149, 150, 151,
258, 260, 261, 264, 267, 268, 356, 400
152, 155, 158, 159, 160, 161, 162, 164, 185,
upādāyarūpa 287
Upaniṣad 3 186, 187, 188, 195, 196, 202, 203, 204, 208,
Upaniṣadic 139 212, 213, 228, 229, 235, 239, 240, 250, 252,
upapatti 211, 332 255, 257, 259, 260, 263, 266, 270, 310,
upāya 72, 364, 378, 393 325, 328, 329, 330, 334, 339, 340, 356,
upekṣā / upekkhā 68, 69, 115, 204, 211, 212, 357, 398, 399
216, 331, 332, 333, 334, 335, 336, 337, 339 vijñānānantyāyatana / viññāṇañcāyatana
ūṣmagata 381 124, 125, 135, 339
Uttarakuru 321 Vijñānavādins 178
vijñaptimātra 152, 242
Vaibhāṣika 10, 11, 40, 178, 184, 224, 225, 226, Vijñaptimātra 389, 392, 393
227, 228, 229, 230, 231, 233, 234, 235, 236, vikalpa 153, 162, 355, 384
238, 240, 257, 326, 340, 353,356, 357 Vimalakīrtinirdeśa (Weimojie jing 維摩詰經 )
Vaidalyaprakaraṇa 358 81, 106
vaiśāradya (four) 60, 61, 63, 64, 65, 66, 70, vimāna 323
76, 77 vimuktimārga 342
Vaiśaravaṇa 322 Vimuktisena 83, 84
Vaiśeṣika 36, 37, 359, 372, 397,405 Vimuttimagga 196
vāsanā 150, 151, 224, 234, 241, 381 Vinataka 320
vaśitā (four) 59, 60, 61, 64, 66, 70 Vinaya 1, 4, 5, 13, 54, 77, 92, 93, 177, 304
Vasubandhu 9, 10, 15, 91, 92, 93, 94, 106, 146, vipākahetu 324, 325
148, 149, 150, 158, 163, 175, 177, 178, 223, vipākaja 242
224, 227, 228, 229, 230,231, 232, 233, 235, vipākaphala 315, 343, 344
236, 238, 241, 284, 315, 324, 328, 329, 331, vipaśyanā / vipassanā 118, 119, 340
430 Index
virtue (see śīla) Xiang Kai 襄楷 282

Virūḍhaka 322 xiangying 相應 (see saṃprayoga)
Virupākṣa 322 xin 信 82, 85, 86, 287
visayajjhatte 126, 127 Xiongnu 匈奴 11
vision (see darśana / dassana) xiuxi wei 修習位 381
Visuddhimagga 9, 10, 69, 118, 119, 122, 128, Xu Shen 許慎 280
136, 138, 153, 154, 174, 186, 197, 199, 200, xuanxue 玄學 282, 291
254, 257, 259, 336 Xuanzang 玄奘 10, 13, 19, 54, 58, 59, 81, 85,
Visuddhimagga with the Commentary 87, 90, 93, 170, 171, 224, 225, 226, 227,
written by King Parākramabāhu II 228, 229, 231, 237, 239, 284, 296, 297,
169, 171, 172, 173, 174, 175, 176, 177, 178, 298, 305, 388, 389, 391, 401
180, 181, 184, 185, 187, 193, 194, 195, 197,
198, 199, 200, 201, 202, 203, 204, 205, 206, yakṣa 322
207, 208, 210, 212, 214, 215, 216 217, 218, Yāma 322, 323
219 Yamaka 128
Vitaṇḍavādins 200 Yaśomitra 224, 227, 230, 232, 353,
vitarka 68, 238, 332, 333, 334, 335, 336, 337, yi 義 287, 380
338, 340, 341 Yibuzong lun lun 異部宗輪論 4
vivartakalpa 327 Yibuzong lun lun shuji 異部宗輪論述記 4, 32
volition (see cetanā) Yijing 義淨 171
Vyākhyāvyukti 175 Yid dang kun gzhi’i dka’ ba’i gnas rgya cher
vyaya 238 ’grel pa 147, 163
vyāyāma 331 Yin chi ru jing 陰持入經 250, 253, 256, 257,
261, 263, 265, 301, 302
Wang Mang 王莽 280, 281 Yin chi ru jing zhu 陰持入經註 253, 256, 257,
Wang Xun 王珣 283, 284 260, 263, 264, 265, 266, 267, 268, 269,
warmth (see ūṣmagata) 270, 272
water (see ap) Yin Shun 239, 242, 244, 245
Weimojie jing 維摩詰經 81, 106, 255 yingxiang xiangfen 影像相分 389
White, David G. 29 Yogācāra 8, 9, 10, 15, 17, 19, 20, 146, 147, 150,
Wijeratne, Ranjan 179, 195, 199 151, 152, 153, 154, 156, 157, 160, 162, 163,
wind (see vāyu) 164, 223, 225, 228, 229, 230, 234, 239, 241,
wisdom (see jn͂ āna / jhāna) 242, 243, 275, 296, 300, 305, 321, 375,
without inadequacy (see mi-lcogs med-pa) 376, 377, 380, 381, 382, 384, 389, 391, 392,
worldly supreme dharma (see 393
laukikāgradharma) Yogācārin 15, 19, 178, 224
Wu 吳 Kingdom 266 Yogācārabhūmi 19, 146, 147, 150, 151, 152, 153,
wu chang 五常 287 156, 157, 159, 160, 161, 162, 164, 242, 307
Wudoumi dao 五斗米道 281, 284 Yogācārabhūmiśāstra 305
wuwei 無為 286 Yujialun ji 瑜伽論記 229, 247
wuwei 五位 (see pañcavastuka) Yuezhi 月氏 285
wu xing 五行 283, 287 Yugaṃdhara 320
Yūki, Reimon 296
Xifang yaojue shiyi tonggui 西方要決釋疑通
規 309, 310, 312 Za apitan xin lun 雜阿毗曇心論
xiqi 習氣 242, 381 (see*Saṃyuktābhidharmahṛdaya)
xianxing 現行 381 Zajilun shuji 雜集論述記 306
xiangfen 相分 389, 390 Zacchetti, Stefano 251, 255
Index 431
Zhang Qian 張騫 11, 279 zhongzi 種子 381

zhi 智 52, 59, 70, 73, 82, 111, 287, 355, 362, Zhou 周 dynasty 280, 281
363, 366, 367, 368, 369 Zhu Faya 竺法雅 288, 291, 303
Zhi Qian 支謙 285, 286 Zhu Fonian 竺佛念 283, 303
Zhiyan 智儼 297 Zhuangzi 奘子 289
Zhiyi 智顗 297 ziliang wei 資糧位 381
Zhong benqi jing 中本起經 253, 266

Abhidharma Across Buddhist Scholastic Traditions

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Abhidharma Across Buddhist Scholastic Traditions

Transféré par

Droits d'auteur :

Formats disponibles

﻿ i

Text, History, and Philosophy

© Koninklijke Brill NV, Leiden, 2016 | doi 10.1163/9789004318823_001

Brill’s Indological Library

Richard Gombrich, Oskar von Hinüber,

The titles published in this series are listed at brill.com/bil

Text, History, and Philosophy

Library of Congress Cataloging-in-Publication Data

Names: Conference "From Abhidhamma To Abhidharma. Early Buddhist

Copyright 2016 by Koninklijke Brill nv, Leiden, The Netherlands.

This book is printed on acid-free paper and produced in a sustainable manner.

In memory of Lance S. Cousins (7 April 1942–14 March 2015)

1 Abhidharma and Indian thinking 29

2 Abhidharmic Elements in Gandhāran Mahāyāna Buddhism: Groups

3 Interpretations of the Terms ajjhattaṃ and bahiddhā: From the Pāli

4 Some Remarks on the Proofs of the “Store Mind” (Ālayavijñāna) and

5 Sanskrit Abhidharma Literature of the Mahāvihāravāsins 169

6 The Contribution of Saṃghabhadra to Our Understanding of

7 Pratītyasamutpāda in the Translations of An Shigao and the Writings

8 Abhidharma in China: Reflections on ‘Matching Meanings’ and

9 Kuiji’s Abhidharmic Recontextualization of Chinese Buddhism 296

10 Traces of Abhidharma in the bSam-gtan mig-sgron (Tibet, Tenth

11 Madhyamaka in Abhidharma Śāstras: The Case of Harivarman’s

12 Svalakṣaṇa (Particular) and Sāmānyalakṣaṇa (Universal) in

13 Perspectives on the Person and the Self in Vasubandhu’s

ments. The ‘World of Abhidharma’ proved to be a more interwoven complex

List of Figure and Tables

1 Extract from BajC2, part 1 (frame 2) 52

1 Development of lists based on the bodhipakṣya-dharmas after Bronkhorst,

Abhidharma as Rational Inquiry (Bart Dessein)

Although discussion exists on the precise meaning of the word Abhidhamma /

© Koninklijke Brill NV, Leiden, 2016 | doi 10.1163/9789004318823_002

‘revealing truth’.3 It is through rational thinking and argumentation that this

3 See Bronkhorst, 2001: 34.

10 See, e.g. T.1.12: 227b7–10; T.26.1536: 367b8–24.

16 Mahāparinibbāna suttanta, chap. 6, v.1.

mātṛkās were doctrinally interpreted, and became the argumentative material

philosophical system, as creating their peculiar system was to all likelihood

A study of the Theravāda Abhidhamma literature shows that Buddhaghosa’s

to the establishment of the Mahāyāna Yogācāra school is evident from the

38 See Frauwallner, 1956: 18.

41 See for the latter: Bronkhorst, 2011: 42–43, 51–61.

school that, in its Vaibhāṣika interpretation, flourished in the Kuṣāṇa empire.

knowledge and understanding of this interpretation. Traditional Chinese phi-

their time to Abhidharma studies. In this way, an Abhidharma School started

53 Cox, 1995: 30–31.

third-fourth century intellectual debate between adherents of the Ṥrāvakayāna

59 See Dessein, 2010: 67.

of ‘Buddhist philosophy’ and for the maturation of Mahāyāna Buddhism.60

Methodological Remarks (Weijen Teng)

60 See Willemen, Dessein & Cox, 1998: 270.

will try to signify some methodological advantage this volume as a whole

61 The enterprise of Abhidharma is universally treated as the Buddhist “Scholasticism”. To

dynamics are revealed. This is surely a legitimate and necessary approach.

Abhidharmic scholasticism, and shows how Kuiji attempted to construct a

found in the indigenous commentaries to the Abhidharma and Yogācāra texts.

DN Dīghanikāya (ed. Rhys Davids and Carpenter, [1890–1911] 1995–2007)

Anālayo (2014). The Dawn of Abhidharma. Hamburg Buddhist Studies 2. Hamburg:

Gernet, Jacques ([1972] 1990). Le monde chinois. Paris: Armand Colin.

Stache-Rosen, Valentina (1968). Dogmatische Begriffsreihen im älteren Buddhismus. Das

Abhidharma and Indian thinking

© Koninklijke Brill NV, Leiden, 2016 | doi 10.1163/9789004318823_003

In this chapter, I will concentrate on Abhidharma influence outside

i

16 Mahāparinibbāna suttanta, chap. 6, v.1.

Bronkhorst, Johannes (2013). “Reflections on the origins of Mahāyāna”, Septimo Cen

(e)[va vuto] bhag̱ava aï[śpa] (śa)[r](ip)u(tro edad oya sarvadharma)

(catvari ś̱paḏovaṭ́haṇa ·) (Four establishments of mindfulness),

balāni 怛薩阿竭十種力 daśatathāgata daśatathāgatabalāni daśatathāgatabalāni