UNIVERSITY OF CALIFORNIA

Santa Barbara

Buddhists, Brahmans, and Buddhist Brahmans: Negotiating Identities in Indian

Antiquity

A dissertation submitted in partial satisfaction of the

requirements for the degree Doctor of Philosophy

in Religious Studies

by

Nathan Michael McGovern

Committee in charge:

Professor Vesna A. Wallace, Chair

Professor David Gordon White

Professor William F. Powell

March 2013
 UMI Number: 3559822

All rights reserved

INFORMATION TO ALL USERS

The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.

UMI 3559822
Published by ProQuest LLC (2013). Copyright in the Dissertation held by the Author.
Microform Edition ProQuest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code

ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
The dissertation of Nathan Michael McGovern is approved.

____________________________________________
William F. Powell

____________________________________________
David Gordon White

____________________________________________
Vesna A. Wallace, Committee Chair

December 2012
Buddhists, Brahmans, and Buddhist Brahmans: Negotiating Identities in Indian
Antiquity

Copyright 2013

by

Nathan Michael McGovern

iii
 Acknowledgements

Although in the end I have written nearly exactly the dissertation that I said

that I would write when I applied to graduate school, the final product has been

shaped by the serendipitous series of events that have passed over the course of the

last nine and a half years. While teaching English in Thailand during the summer

after my first year in graduate school, I was able to immerse myself in a living

Buddhist culture and was immediately intrigued by the pervasiveness of Hindu

themes within it. In even the most casual tour of the capital Bangkok, one encounters

statues and reliefs of Hindu deities, Sanskritic references to Hindu mythology in the

names of prominent institutions, the painting of the Rmyaa along the inside wall

of the Temple of the Emerald Buddha, and even a temple for Brahmans employed by

the king just across the street from city hall. Based on the interest that was piqued

that first summeras well as the woman I met in Thailand who would later become

my wifeI decided to return the next summer to conduct research for my masters

thesis on one aspect of the Hindu influence on Thai Buddhist culture, namely, the

pervasive worship of Brahm by Thai Buddhists. When I ultimately decided to return

to the study of early Indian Buddhism for my dissertation research, I was profoundly

iv
influenced by what I had learned in Thailand. Seeing how problematic the distinction

between Hinduism and Buddhism is in modern Thai society, I became curious

about the dynamic between Buddhism and Brahmanism in early Indian society

and consequently obsessed with the category Brahman in the early Buddhist

literature. It is for this reason that I decided to focus my study of the early Buddhist

literature on the depiction of Brahmans and deployment of the category Brahman

within it.

Many people have helped me come to where I am today. First of all, thanks

must go to my three advisersVesna Wallace, David White, and Bill Powellwho

have each molded me in various ways to become the scholar that I am today. In

addition to these three, I have been particularly influenced in my training by Barbara

Holdredge and Jos Cabezn, both of whom have mentored me since my earliest days

at UCSB and have given useful advice and criticism in my latter days there as I was

working on this dissertation. Given the textual nature of my research, I also owe a

great debt of gratitude to my language teachers, who in my opinion are the great

unsung heroes of the university. First and foremost, I am eternally grateful to Greg

Hillis, who introduced me to both Sanskrit and Tibetan and also provided much

useful insight on matters language-related and un-related while I was working on my

dissertation. During my last few years in graduate school, Evelyn Wade taught me

German, and her brilliance in doing so will make me a better scholar and, through her

example, has I believe already made me a better teacher. In Chinese, I have been

blessed with too many wonderful teachers to count, much less name. At UCSB, I

v
wish in particular to thank Jennifer Hsu, whose pedagogical energy and tenacity

forced my brain, spoiled as it was by the study of Sanskrit and Thai, to contend with

the very different grammatical and written world of Chinese. At ICLP in Taipei, I

had many different excellent instructors, but I would like in particular to thank Liu

Xiaoping, who helped me to work through my very first encounter dialogs in

Chinese. In addition to those already mentioned, many people have provided useful

advice or criticism in one way or another over the course of the past few years as I

was thinking through and working on this dissertation. These include Michael

Jerryson, Justin McDaniel, Nathaniel Rich, Joel Gruber, Paul Harrison, Luis Gmez,

Bhikkhu Sujto, Emily Schmidt, Mark McLaughlin, and David McMahan. All of

these people have helped to make this a better dissertation in one way or another; I of

course take full responsibility for all weaknesses and errors that remain. Finally, I

would like to thank my colleagues at Franklin and Marshall College, who have been

gracious enough to welcome me back nearly a decade after my days here as an

undergraduatetheir support, professional guidance, and friendship has been

invaluable in allowing me to finish this dissertation while beginning my career as a

full-time instructor.

Of course, no one who earns a PhD is able to get there without the love and

support of those who are closest to them. In my case, I am increasingly, as I get older,

in awe of one particular gift that my parents gave to me when I was a childthe gift

of reading. Their tireless reading of bedtime stories (or, as I sometimes requested,

readings from a college-level world history textbook) gave me a love and ability for

vi
reading that truly must be understood as the most important condition for the writing

of this dissertation to have been possible. Thanks must also go to Helen, who

cleverly leveraged her role as the doting aunt to subsidize my personal library by

taking me to bookstores whenever we had the chance to visit with one another.

Finally, I owe an eternal debt of gratitude, one that can only be repaid over the course

of lifetimes, to my beloved wife Nanda, who among other things, has kept me sane

long enough to see this dissertation to completion.

vii
 Curriculum Vitae for Nathan Michael McGovern

Current Employment

Visiting Instructor in Religious Studies, Franklin and Marshall College (2012-2013)

Education

PhD (expected Winter 2013), University of California, Santa Barbara, Religious

Studies

C.Phil. (2009), University of California, Santa Barbara, Religious Studies

MA (2006), University of California, Santa Barbara, Religious Studies

BA (2003), Franklin and Marshall College, with majors in Physics and Religious
Studies

Dissertation

Buddhists, Brahmans, and Buddhist Brahmans: Negotiating Identities in Indian

Antiquity

Committee: Vesna Wallace (Chair), David Gordon White, William Powell

Research Interests

Early South Asian Religion

Southeast Asian Religion

Comparative Hindu/Buddhist Studies

Professional Publications

Buddhism and Hinduism in Thailand, Oxford Bibliographies Online (in press).

Brahm: An Early and Ultimately Doomed Attempt at a Brahmanical Synthesis,

Journal of Indian Philosophy 40, no. 1 (Feb. 2012): 1-23.

viii
Thailand, in Brill Encyclopedia of Hinduism, vol 1. Leiden: Brill, 2009.

G.S. Adkins and N.M. McGovern, Application of the effective interaction

method to the relativistic Coulomb problem, Am. J. Phys. {\bf 73}, 759
(2005).

G.S. Adkins, N.M. McGovern, R.N. Fell, and J. Sapirstein, Two-loop

corrections to the decay rate of parapositronium, Phys. Rev. A {\bf 68},
032512 (2003).

Teaching and Other Academic Work Experience

Visiting Instructor (Franklin and Marshall College):

Fall 2012, RST 122: Introduction to Asian Religions
Fall 2012, RST 372: Seminar in Theravda Buddhism
(scheduled) Spring 2013, RST 248: Buddhism (2 sections)
(scheduled) Spring 2013, RST 375: Seminar in Science and Religion

Teaching Associate (University of California, Santa Barbara):

Summer Session B, 2011, RG ST 4: Intro to Buddhism
Summer Session A, 2012, RG ST 4: Intro to Buddhism

ESL Tutor (Santa Barbara City College):

Spring 2011, Summer 2011, Fall 2011, Spring 2012
Tutoring ESL students at Santa Barbara City College.

Writing Teacher (University of California, Santa Barbara):

Taught Writing 2 to UCSB students during all 3 quarters of the 2008-2009
school year, as well as Fall 2009. Involved developing a syllabus and
teaching an entire course independently, under the supervision of a faculty
supervisor.

Writing Center Tutor (Santa Barbara City College):

Winter 2008
Worked as a writing tutor in the Cartwright Center at SBCC.

Reader (University of California, Santa Barbara):

Summer 2006, FR & IT 114X, Dante
Winter 2007, EALCS 181AA, Silk Road

Teaching Assistant (UCSB):

Fall 2004, RG ST 80A, Religion in the Western World (Ancient)
Winter 2005, GEOG 822, Nature, Science, and Religion

ix
 Spring 2005, RG ST 80C, Religion in the Western World (Modern)
Fall 2005, RG ST 80A, Religion in the Western World (Ancient)
Winter 2006, RG ST 4, Introduction to Buddhism
Spring 2006, RG ST 80C, Religion in the Western World (Modern)
Fall 2007, RG ST 4, Introduction to Buddhism
Winter 2008, Physics 4L, Introductory Physics Lab
Spring 2008, Physics 1, Introductory Physics

ESL Teacher (Mahidol University):

Summer 2004
Developed and taught two English courses (1st year and 2nd year) for Thai
students at the College of Religious Studies, Mahidol University, in Salaya,
Thailand.

Hackman Researcher (Franklin and Marshall College):

Summers of 2000, 2001, and 2002
Worked with Prof. Greg Adkins in the Physics Department at F&M during the
summers of 2000 and 2001 on the energy levels of hydrogen and during the
summer of 2002 on the decay rate of parapositronium. See Scholarships and
Awards.

Writing Center Tutor (Franklin and Marshall College):

Academic years 2001-2002 and 2002-2003
Tutoring fellow students in academic writing, with a particular emphasis on
structuring/organizing a paper, developing a thesis, and correct grammar and
mechanics.

Scholarships and Awards

OBO: Buddhism Graduate Student Award for 2012:

Awarded a contract to write an article (now accepted after peer review) for
Oxford Bibliographies Online on the topic Buddhism and Hinduism in
Thailand.

UCSB Graduate Dissertation Fellowship:

Provides one quarter of tuition/fee remission and a stipend to support
dissertation writing. Awarded for Fall 2011.

Taiwan Ministry of Education Huayu Enrichment Scholarship:

Awarded for Sept. 2009-Aug. 2010 (accepted for Jan. 2010-Aug. 2010) to
study Chinese at the International Chinese Language Program in Taipei.

x
Summer FLAS Scholarships for Chinese:
Awarded to study Chinese at National Taiwan Normal University (Taipei)
during Summer 2009 and at the International Chinese Language Program
(Taipei) during Summer 2010.

Academic-year FLAS Scholarships for Chinese:

Awarded for the 2008-2009 and 2009-2010 school years to study Mandarin
and classical Chinese at UCSB (Fall 2008-Fall 2009) and in Taipei (Winter-
Spring 2010). FLAS is a US government grant for the study of less-
commonly taught modern languages. Covers all tuition and fees and provides
a stipend to cover living expenses while studying the language.

Ninian Smart Memorial Award:

Awarded in spring 2007 by the Department of Religious Studies, UCSB, for
achievement in the comparative study of religion and philosophy.

Drew Award:
Awarded in spring 2007 by the Department of East Asian Languages and
Cultures, UCSB, for outstanding achievement in first year Chinese.

Pacific Rim Mini-Grant:

Awarded in November 2006 for preliminary dissertation research conducted
in Thailand in summer 2007.

Fellowships for Thai Language Study:

Awarded in 2005, 2006, and 2007 by the UCSB Department of Religious
Studies.

New Visions Paper-writing Fellowship:

Awarded in spring 2005 to fund the writing and publication of a paper on
Newtonian and Einsteinian Mechanics from a social constructivist perspective.

Humanities and Social Sciences Fellowship:

Awarded for research in Thailand in summer 2005.

Rowney Fellowship:
A 4-year research/teaching fellowship awarded to incoming graduate students
by the Department of Religious Studies, UCSB. Recipient in 2003.

Phi Beta Kappa:

Inducted in 2003.

xi
Williamson Medal:
Awarded to one graduating senior each year at F&M, on the basis of academic
excellence and school leadership. Recipient in 2003.

Hackman Scholars Program:

A summer stipend given to F&M students to work with a professor on a ten-
week research project. Recipient in 2000, 2001, and 2002.

Pi Mu Epsilon Mathematics Honor Society:

An honor society that admits undergraduates who have demonstrated high
achievement in mathematics, for the promotion of mathematics education and
awareness. Inducted in 2002.

Sigma Pi Sigma National Physics Honor Society:

An honor society that admits outstanding physics undergraduates for the
promotion of physics education both in and out of the classroom. Inducted in
2001.

Charles D. Spotts Religious Studies Award:

Awarded to one student in recognition for achievement in religious studies
over the course of the academic year. Recipient in 2002.

Kershner Scholar for Physics:

Awarded to several students in recognition for achievement in physics classes
over the course of the academic year. Recipient in 2000, 2001, and 2002.

Buchanan Scholarship:
A renewable scholarship awarded to incoming freshmen at F&M who have
demonstrated a commitment to community service during their high school
careers.

John Marshall Scholarship:

A four-year merit-based scholarship awarded to incoming freshmen at F&M.

Professional Conference Papers

Rethinking the Early Buddhist/Brahman Divide: A Comparative Study of

References to Brahmans in the Pali and Chinese Canons, presented at the
American Academy of Religion Conference, Atlanta, GA, Nov. 1, 2010

Revisioning the Buddhist/Brahman Divide: Brahmanical Orthodoxy around the

Time of the Buddha, paper presented at the North American Graduate
Student Conference in Buddhist Studies, Berkeley, CA, Apr. 18, 2009

xii
Brahma and Brahmas in Indian Cosmographies, paper presented at conference on
Map and World: Buddhist Notions of Cosmology and Geography, Riverside,
CA, Feb. 18, 2007

Hinduism in a Non-Hindu Context: A Study of Brahma Worship in Thailand,

paper presented at the American Academy of Religion Conference,
Washington, D.C., Nov. 19, 2006

Lectures

Thai Buddhism: Guest lecture in RG ST 4, Intro to Buddhism, UC Santa Barbara,

Feb. 10, 2006; Feb. 28, 2007; Nov. 16, 2007; Nov. 10, 2010; Nov. 9, 2011.

Thai Religiosity: What Buddhism Looks like outside the Books You Buy at
Borders: Community lecture presented at the Santa Barbara Public Library,
June 7, 2006.

Brahma in Thailand: Buddhists Worshipping a Hindu God?: Guest Lecture at the

Center for Hindu Traditions (CHiTra), University of Florida, Sept. 5, 2007.

Early Buddhist Ideas: Situating them in Context: Guest lecture in RG ST 4, Intro to

Buddhism, UC Santa Barbara, Jan. 13, 2009; Sept. 26, 2011.

Language Experience

Language Level Coursework

Mandarin Chinese Advanced Speaking and Reading 3 years regular; 10
months intensive
Classical Chinese Advanced Reading 1 year regular; 10 months
intensive
French Passed Graduate Competency 2 quarters
German Advanced Speaking and Reading 3 years regular
Pali/Sanskrit Advanced Reading 5 years regular
Thai Advanced Speaking and Reading 2 summers intensive
Tibetan Advanced Reading 3 years regular

xiii
 Abstract

Buddhists, Brahmans, and Buddhist Brahmans: Negotiating

Identities in Indian Antiquity
by

Nathan Michael McGovern

In this dissertation, I explore the way in which early Buddhists formed their

own self-identity vis--vis Vedic Brahmans. Interestingly, while today we understand

Buddhism and Hinduism as separate religions, and Brahmans as belonging to the

latter religion, early Buddhist texts frequently refer to the Buddha, his enlightened

disciples, or just the ideal person in the abstract as Brahmans. Many previous

scholars have argued or assumed that the category Brahman was a term that

belonged properly to the Brahmanical tradition, whose roots long pre-date

Buddhism, and thus concluded that early Buddhist uses of the word Brahman

(brhmaa) should be interpreted as borrowings from the Brahmanical tradition,

made for the purpose of polemic. I, however, argue that the word Brahman was a

contested category that was not owned by a monolithic Brahmanical tradition and

xiv
then borrowed by Buddhists to use polemically, but that it was a common honorific

term that was actively contested by both Buddhists and (Vedic) Brahmans and only

with time ceded by the former to the latter. In making this argument, I make use of a

wide variety of early South Asian sources, including Brahmanical texts, inscriptions,

and early Buddhist texts preserved in Pali, Sanskrit, Chinese, and Tibetan.

In the first part of my dissertation, I show that while there was a common

vocabulary for describing religious practitioners in ancient India, Buddhist and

Brahmanical texts deployed them differently in an effort to construct a unique and

superior identity for their own respective ideals. Then, in the second part, I argue that

the earliest Buddhist literature makes use of the word brhmaa (Brahman) simply

as an honorific for the ideal person, while later elaborations on this usage convert this

simple honorific into a polemic against the Vedic Brahmans, who are depicted as

unworthy of the name by which they call themselves. Finally, in the third part of my

dissertation, I examine texts that I refer to as encounter dialogstexts in which the

Buddha encounters an interlocutor who is identified as a Brahman, and argue that

through the implicit narrative structure of these texts, Buddhists effectively ceded the

category Brahman.

xv
 Table of Contents

Acknowledgements .................................................................................................. iv
Curriculum Vitae .................................................................................................... viii
Abstract .................................................................................................................. xiv
Abbreviations ...........................................................................................................xx
Part I: Introduction .........................................................................................................1
I.1 Received Narratives: Structuring Themes of Scholarly Thought .........................5
I.1.1 Burnouf and the Question of Chronological Priority .....................................6
I.1.2 Oldenberg, the Buddha as Reformer, and Historicism.................................12
I.1.3 The Early Twentieth Century: From Rhys Davids to Lamotte ....................17
I.2 Newer Scholarship on Buddhism and Brahmanism ...........................................22
I.2.1 Gombrich: Buddhism as Reaction to Brahmanism ......................................27
I.2.2 Bronkhorst: Greater Magadha and the Brahmanical Project .......................30
I.3 Developments in Indology ..................................................................................37
I.4 Methodology of this Study..................................................................................46
I.5 Outline of the Dissertation ..................................................................................57
Part II: Discursive Categories for Religious Identities in Ancient India ..................60
Introduction: Categories for Religion and Buddhism in Ancient India ............61
Chapter II.1: ramaas and Brhmaas ..................................................................71
II.1.1 ramaa and Brhmaa as Mutually Exclusive Categories.......................71

xvi
 II.1.2 Samaabrhma as a Unified Category Against Which to Construct
Buddhist Identity in the Early Buddhist Texts .....................................................76
II.1.3 Samaabrhma Used in a Positive Sense ...............................................85
II.1.4 Conclusion ..................................................................................................93
Chapter II.2: The rama System and Its Relationship to Brahmanism ..............98
II.2.1 Olivelles Characterization of the rama System as a Brahmanical
Theological Construct .........................................................................................98
II.2.2 The Rhetoric and Reality of Brahmanical Claims to Purity of Birth ........104
II.2.3 The Rhetoric of Mixed Castes ...............................................................110
II.2.4 The rama System as a New Brahmanical Systematization of Actual
Social Practice ....................................................................................................113
Chapter II.3: The Colloquial Categories that were Subsumed into the rama
System and their Treatment by the Buddhists ........................................................123
II.3.1 Greek Evidence for the Use of rama Categories in Colloquial Discourse
............................................................................................................................123
II.3.2 Ghastha and Gahapati.............................................................................128
II.3.3 Parivrjaka/Paribbjaka ..........................................................................138
II.3.4 Brahmacarya and the Brahmacrin..........................................................148
II.3.5 Vnaprastha and Jaila .............................................................................169
Chapter II.4: Conclusion to Part II .........................................................................193
Part III: Diachrony and the Use of Brahmanical Terms in Early Buddhist Texts..196
Introduction: Problematizing the Concept of Brahmanical Terms .....................197
Chapter III.1: Early and Classical Uses of the Word Brhmaa in Early Buddhist
Texts .......................................................................................................................204
III.1.1 Classical Uses of the Word Brhmaa in Early Buddhist Literature as
Polemic ...............................................................................................................204
III.1.2 Antiquity and Uniqueness of the Ahaka and Pryaa ........................215
III.1.3 The Non-Polemical Use of the Word Brhmaa in the Ahaka and
Pryaa ............................................................................................................232
Chapter III.2: Commentary and Framing as Hermeneutical Devices for Dealing
with Early Uses of the Word Brhmaa ................................................................247

xvii
 III.2.1 The Mahniddesa Commentary on the Ahaka Vagga...........................249
III.2.2 The Chinese Translation of the Arthavargya and the Hermeneutical Role
Played by Narrative Framing..............................................................................257
III.2.3 Framing in the Pryaa .........................................................................281
III.2.4 Brhmaa as a Term Taken from the Pan-Indo-Aryan Substratum ........297
Chapter III.3: Tevijja/Tisso Vijj............................................................................303
III.3.1 How Does One Attain Awakening? ........................................................305
III.3.2 Pre-canonical Buddhism and Substratal Influences on Buddhism ......327
III.3.3 The Three Knowledges Within the Context of the Development of
Buddhist Discourses on Awakening ...................................................................346
Chapter III.4: Conclusion to Part III ......................................................................355
Part IV: Encounter Dialogs in the Context of Oral Tradition ....................................358
Introduction: Encounter Dialogs ............................................................................359
Chapter IV.1: Orality and the Early Buddhist Texts ..........................................364
IV.1.1 The Oral Theory of Milman Parry and Albert Lord ................................364
IV.1.2 Applying the Oral Theory to the Early Buddhist Texts ..........................369
IV.1.3 Previous Applications of the Parry-Lord Theory to the Early Buddhist
Literature ............................................................................................................380
Chapter IV.2: Comparative Nikya/gama Studies and the Development of the
Nikya/gamas ......................................................................................................402
IV.2.1 Introduction .............................................................................................402
IV.2.2 Sayutta Nikya / Sayuktgama ..........................................................407
IV.2.3 Dgha Nikya / Drghgama ...................................................................420
IV.2.4 Majjhima Nikya / Madhyamgama .......................................................430
IV.2.5 Aguttara Nikya / Ekottarikgama........................................................454
IV.2.6 Conclusion ...............................................................................................461
Chapter IV.3: Collections of Encounter Dialogs ...................................................463
IV.3.1 Introduction .............................................................................................463
IV.3.2 The Brhmaa Sayutta/Sayukta of the Sayutta
Nikya/Sayuktgama .......................................................................................466

xviii
 IV.3.3 The Encounter Dialogs of the Slakkhandha Vagga/laskandha Nipta of
the Dgha Nikya/Drghgama ..........................................................................481
IV.3.4 The Brhmaa Vagga of the Majjhima Nikya and Brhmaa Varga of
the Madhyamgama ...........................................................................................520
IV.3.5 Encounter Dialogs in the Aguttara Nikya............................................531
IV.3.6 Encounter Dialogs in the Sutta Nipta ....................................................538
Chapter IV.4: Development of Formulae and Themes in the Encounter Dialog
Tradition .................................................................................................................546
IV.4.1 Introduction .............................................................................................546
IV.4.2 Brhmaagma and Brahmadeyya .........................................................548
IV.4.3 The Fame of Gotama Formula .............................................................558
IV.4.4 The Triple Veda Formula ....................................................................562
IV.4.5 The Theme of the Thirty-Two Marks of a Great Man ............................575
IV.4.6 The Recurring Characters Vseha and Bhradvja ...............................604
IV.4.7 The Stock Brahman Jussoi ................................................................612
Chapter IV.5: Conclusion to Part IV ......................................................................618
Part V: Conclusion .....................................................................................................628
Bibliography ...........................................................................................................645

xix
 Abbreviations

AB Aitareya Brhmaa
DhS pastamba Dharma Stra
AN Aguttara Nikya
Ap. - Apadna
AV Atharva Veda Sahit, aunaka Recension
BU Bhad rayaka Upaniad
BDhS Baudhyana Dharma Stra
BS Bhat Sahit
Bv. Buddhavasa
ChU Chndogya Upaniad
Dhp. - Dhammapada
DN Dgha Nikya
GDhS Gautama Dharma Stra
Iti. - Itivuttaka
Jt. - Jtaka
JB Jaiminya Brhmaa
MBh. - Mahbhrata
MDh Mnava Dharma stra
Mil. Milinda Paha
MN Majjhima Nikya
Mv. - Mahvagga
Mvu. - Mahvastu
Nidd. - Niddesa
Pais. Paisambhidmagga
Pe. Peakopadesa
Pv. Peta Vatthu
RV g Veda Sahit
RVKhil. g Veda Khilni
B atapatha Brhmaa
Sn. Sutta Nipta
SN Sayutta Nikya

xx
Ther. Theragth
Ther. - Thergth
TS Taittirya Sahit
Ud. - Udna
VDhS Vasiha Dharma Stra
Vin. - Vinaya

xxi
 Part I

Introduction

1
 In the Pryaa Vagga of the Sutta Nipta, generally considered one of the

oldest Buddhist texts to have been preserved to the present day, a man named

Dhotaka approaches the Buddha and beseeches him in the following terms:

I see in the world of gods and men a Brahman wandering about with nothing.
Therefore, I bow to that one who is all-seeing. Release me, kya, from my
doubts. 1

When the Buddha says that he cannot free one who has doubts, only offer him the

best doctrine, Dhotaka continues,

Having pity, Brahm, teach me the dhamma of detachment, that I may

understand,
(And) being unchanging just like space, calmed, I may wander right here
unattached. 2

Now fast forward several hundred years, to the third century CE. Zh Shxng

(), in the first recorded instance of a Chinese monk leaving China in search of

Buddhist scriptures, travelled to Khotan in search of the Pacaviatishasrik, since

at that time he and his Chinese compatriots only had access to the Aashasrik,

which they considered to be a mere extract from the former. Once he had made the

arduous journey and obtained this new and improved Perfection of Wisdom Stra,

Zh Shxng faced one more obstacle: the Hnaynists in the city didnt want him to

leave the city with what they considered a heretical text. According to Zh Shxngs

biography, they tried to convince the king not to let the Chinese monk take what they

1
Sn. 1063: passmaha devamanussaloke, akicana brhmaam iriyamna / ta ta
namassmi samantacakkhu, pamuca ma sakka kathakathhi //.
2
Sn. 1065: anussa brahme karuyamno, vivekadhamma yam aha vijaa / yathha
ksova abypajjamno, idheva santo asito careyya //.

2
called a Brahman book (in Chinese, ) back to China. Zh Shxng offered

to put the Mahyna text through a trial by fire, which it passed, and thus he was able

to leave the city with it over the objections of the Hnaynists. 3

How do we get between these two very different, indeed diametrically

opposed, Buddhist uses of the word Brahman? In the first example, Brahman is a

title of praise, used for no less than the Buddha himself. In the second, it is a sneer,

directed toward proponents of a new form of Buddhism, the enemy within. We are

all aware of the significance of Brahmans and Brahmanism 4 to early Buddhism

both as providing the context in which it arose and as providing a foil against which it

defined itself. Indeed, describing the Brahmanical context and the Buddhas putative

disagreements with his Brahmanical interlocutors is a given for any introductory class

on Buddhism. What I will argue in this dissertation, however, is that we cannot speak

about a monolithic early Buddhist response to Brahmanism; rather, we must

recognize that early Buddhists spoke about Brahmans and Brahmanical ideas as part

3
This story is quoted by E. Zrcher, The Buddhist Conquest of China: The Spread and
Adaptation of Buddhism in Early Medieval China, 3rd ed. (Leiden: Brill, 2007), 62-63.
4
By Brahmanism, I mean the religio-ideological system of the Brahmans that preceded
and to a large extent served as the backbone for classical Hinduism, which latter can be said to have
coalesced in the early to mid-first century CE. This is an admittedly vague definition, but one that I
think is sufficiently reflective of the use of the category as an etic term in Western scholarship,
particularly in contradistinction with Buddhism. As we will see in this dissertation, however, the term
Brahmanism is problematic because it typically presupposes a static institution of the Brahman
and a static ideology supporting that institutionboth of which we will find to have been fluid well
through the early centuries of Buddhism in India. For this reason, I refer to Brahmanism and the
category Brahmanical in this dissertation mostly for the purposes of interrogating and
problematizing these terms. There are two major exceptions. The first is the new Brahmanism, a
coinage that I borrow from Bronkhorst (see section I.2.2) to refer to a particular movement of
householder-supremacist Vedic specialists who arrogated to themselves on the basis of birth. The
second is Brahmanical to refer to a body of textsby which I mean the Vedic and post-Vedic texts
that are referred to as such in Western scholarship and became the basis of the Hindu tradition.

3
of their ongoing articulation of their own self-identity, that Buddhist self-identity did

not arise suddenly fully formed, and that the significance of the Brahman can only

be expected to have changed over the course of the process of Buddhist identity-

formation.

To this end, I will be examining the treatment of Brahmans in early Buddhist

stras and on that basis address two interrelated questions: (1) What do these early

Buddhist stras tell us about Brahmans and the state of development of classical

Brahmanism at the time that they were written that Brahmanical sources themselves

cannot and (2) how should we understand the project that Buddhists were engaged in

when they referred to Brahmans and Brahmanical concepts in their own stras?

Methodologically, I follow recent scholars who treat early Buddhism and

Brahmanism as extremely dynamic movements that must be viewed, not in isolation

from one another, but as occupying a common context in which they both shared a

common heritage and formed their own self-identities vis--vis one another over the

course of several centuries. Previous scholarship that has investigated these sorts of

issues, especially in the West, has tended to restrict itself to the early Buddhist texts

preserved in Pali, thus ignoring the wealth of comparative data that can be found in

versions of the early Buddhist stras that were passed down by sects other than the

Theravda and preserved in Chinese translation. In this dissertation, therefore, I will

make extensive use of these Chinese translations, in addition to the Pali suttas. What

I will argue is that the early Buddhist discourse on Brahmans and Brahmanical

doctrines and terminology can best be understood not only as polemic or marketing

4
against a rival religious sect, but as a positive participation in a discourse on

Brahmanhood that predates the rise of both Buddhism and its early orthodox

Brahmanical rivals. To this end, I hypothesize the existence of a pan-North-Indian

Aryan/Brahmanical substratum that served as a common source both for the early

Buddhists and for the Western Vedic schools that codified the rauta ritual and then

came into contact with the early Buddhists and other heterodox schools as they

gradually moved east. The emergence of Buddhism and Brahmanism (not to mention

Hinduism) from this substratum must be understood not simply in terms of

historical priority and causation, polemical reaction and marketing, or even

historical context, but rather as a dialectical process of identity-formation in which

key terms of great antiquityin particular brhmaabecame foci of repeated

contestation that left Buddhist and Brahmanical identities vis--vis one another fluid

and unstablecertainly in the late first millennium BCE, and perhaps throughout the

entire coexistence of Buddhism and Brahmanism/Hinduism in India.

I.1 Received Narratives: Structuring Themes of Scholarly Thought

The broad question of the relationship between Buddhism and Brahmanism is

of course not new; it is as old as the Western study of Buddhism itself. Although my

argument in this dissertation is in part a criticism of earlier theories, it is not my intent

here to construct a received narrative against which to pitch my own argument;

instead, I conceive my argument as a contribution to an ever-evolving set of scholarly

narratives on the relationship between Buddhism and Brahmanism. Indeed, I would

5
argue that there is no one particular received narrative for the birth of Buddhism

or Buddhisms relationship to Brahmanism (if in fact there is for any topic of

scholarly discourse); rather, there is a set of themessometimes interlocking,

sometimes conflictingbased on influential scholarly works of the past that have

come to structure the course of academic discourse on this topic. In the brief outline

that follows, I organize early scholarship on the origins of Buddhism, admittedly

somewhat arbitrarily, into three periods divided by three scholars that I consider to

represent particular turning points in scholarly discourse on Buddhist origins

Eugne Burnouf, Hermann Oldenberg, and tienne Lamotte. The last of these,

Lamotte, I chooseagain, somewhat arbitrarilyto mark the end of older

discourses on the relationship between Buddhism and Brahmanism because he is the

last scholar to have written a comprehensive history of Buddhism in India, even with

its somewhat limited chronological scope (From the Origins to the aka Era).

I.1.1 Burnouf and the Question of Chronological Priority

Although by the end of the eighteenth century Europeans had become aware

of the existence of a single religious tradition, dubbed Buddhism, unifying a number

of idolatrous (in the parlance of early European discourse) cults across Asia, it took

several decades for scholars to come to a consensus about the exact context in which

it had arisen. In large part this was due to the vicissitudes of history: Because

Buddhism had long-since died out in India by the time European colonialists came

there, the latter faced a lack of social and institutional knowledge to help them

6
contextualize Buddhism in its land of origin. As Donald Lopez notes in an essay on

the European discovery of the Buddha, scholars from the East India Company

learned from Puric texts that the Buddha was the ninth incarnation of Viu.

However, their Brahman informants appeared to be of two minds about the Buddha

some held him to be a venerable figure who condemned the killing of cattle, and

others condemned him as a heretic. In order to resolve this apparent contradiction,

Sir William Jones postulated that there were two Buddhas: an early one who was the

ninth incarnation of Viu and honored by Hindus as such, and a later Buddha who

was an enemy of the Brahmans. He argued further that the former Buddha was not

originally an Indian god, but rather was imported from elsewhere. In making this

argument, Jones pointed out that statuary depictions of the Buddha often gave him

crisp and wooly hair like Africansan argument that sparked decades of

scholarly debate over whether the Buddha was originally of the Negro race. 5

Similar confusion characterized early scholarly debate over the question of the

chronological relationship between Buddhism and Hinduism. The framework for this

debate was set up by Pierre de Joinville, who by 1801 recognized the similarity

between Buddhism and Brahmanism and the importance of determining their

chronological priority. 6 Joinville argued on logical grounds that Buddhism must

be prior since [t]he Boudhists eat animals; the Brahmins do not. The logic of this

5
Donald S. Lopez, Jr., Buddha, in Donald S. Lopez, Jr., ed., Critical Terms for the Study of
Buddhism (Chicago: University of Chicago Press, 2005), 16-17.
6
Philip C. Almond, The British Discovery of Buddhism (Cambridge, England: Cambridge
University Press, 1988), 29.

7
argument is that surely any religion that seeks to reform an older religion must

introduce some sort of reform; if the Buddhists were the reformers, then they would

be the ones who did not eat meat. 7 Brian Hodgson used a similar logic to argue the

exact opposite: He reasoned that since Buddhism was characterized by monastic

asceticism in morals; philosophical skepticism in religion, it surely must be the more

developed of the two religions. 8 Although Hodgson arguably represented the tide of

scholarly opinion, the theory of the priority of Buddhism remained quite alive, with

William Knighton arguing as late at 1845 in favor of that theory. 9 Thus the question

remained unresolved throughout the first half of the nineteenth century.

Resolution of these issues, though not absolute and certainly not immediate,

was nonetheless effectively achieved with the publication of Eugne Burnoufs

Introduction lhistoire du Buddhisme indien in 1844. This incredibly influential

book was based on a study of original Indian Buddhist texts discovered and sent to

Burnouf by Brian Hodgson, a novel development in the study of Buddhism that

Burnouf portrayed as central to his thesis and to his understanding of the history of

Buddhism. As Lopez and Buffetrille argue in the introduction to their recent

translation of Burnoufs Introduction, Perhaps the most important sentence in the

entire volume occurs on the first page of the foreword, where Burnouf declares that

the belief called Buddhism is completely Indian, literally a completely Indian fact

7
Ibid., 29-30.
8
Ibid., 31.
9
Ibid.

8
(un fait compltement indien). 10 For Burnouf, the discovery of Buddhist texts in the

Indian language of Sanskrit confirmed once and for all that Buddhism was born in

India, developed in India, and could be explained completely on Indian terms,

without reference to outside influences or dubious theories about the Buddhas

origins based on his hair.

Burnouf came to a similarly strong conclusion on the question of the

chronological relationship between Buddhism and Brahmanism. On the basis of his

reading of the Sanskrit manuscripts that had been sent to him by Hodgson, in which

he finds numerous references to Brahmanical gods, Burnouf concludes that at the

time when kyamuni traveled through India to teach his law, Brahmanical society

had reached its highest degree of development. 11 He finds this conclusion to be

secure regardless of the dating of the manuscripts at his disposal:

[T]he stras that attest to the existence of Brahmanical society were either
written around the time of kya, or a very long time after him. If they are
contemporary with kya, the society they describe existed then, because one
could not imagine why they would have spoken in such detail of a society that
was not the one in which kya appeared. If they were written a very long
time after kya, one does not understand any better how the Brahmanical
gods and personages occupy so vast a place there, because long after the
Buddha, Brahmanism was profoundly separated from Buddhism, and because
these two cults had but a single ground on which they could meet, that of
polemic and war. 12

10
Katia Buffetrille and Donald S. Lopez, Jr., Introduction to the Translation, in Eugne
Burnouf, Introduction to the History of Indian Buddhism, trans. Katia Buffetrille and Donald S. Lopez,
Jr. (Chicago: University of Chicago Press, 2010), 14.
11
Eugne Burnouf, Introduction to the History of Indian Buddhism, trans. Katia Buffetrille
and Donald S. Lopez, Jr. (Chicago: University of Chicago Press, 2010), 162.
12
Ibid.

9
We see in this argument an extension of Burnoufs broader argument that, among the

Sanskrit texts at his disposal, there are primitive stras (those in question here) that

are closest to the preaching of kya, [and] remain shielded from the double

influence that the system of celestial buddhas and bodhisattvas and the category of

tantras or most especially of dhrasexercised on the developed stras. 13 These

primitive stras were more historical, less fantastical, and thus more useful than the

later mahvaipulya (Mahyna) stras for understanding the historical context in

which Buddhism arose. 14

There are two major observations we should make about Burnoufs

conclusions regarding the chronological relationship between Buddhism and

Brahmanism. The first is that in retrospect we now see that Burnoufs reasoning was

flawed. The primitive stras Burnouf used to come to the conclusion that

Buddhism emerged in an already well-developed Brahmanical societyfrom the

Divyvadnawere in fact not particularly old 15; their position as primitive in the

19th century scholarly quest for the origins of Buddhism was quickly rendered

obsolete as the texts of the Pali Canon were more carefully examined after Burnoufs

death. Indeed, far from the presence of Brahmanical gods and personages being

inexplicable if the texts in question are late, we find now, with our more complete

13
Ibid., 156.
14
These are, I might add, all tropes that would lay the groundwork for later Orientalist
constructions of pristine Buddhism.
15
Andy Rotman dates the Divyvadna to the early centuries of the Common EraAndy
Rotman, trans., Divine Stories: Divyvadna, Part 1 (Boston: Wisdom Publications, 2008), 1.

10
understanding of the chronology of Indian Buddhist texts, that in some senses there is

an increase in references to Brahmanical culture in later textsa fact that is not

surprising given our increasing awareness of the lateness of certain elements of

classical Hinduism. 16

The second observation is that the flaws in Burnoufs argument hardly matter.

As Buffetrille and Lopez simply put it, Burnoufs Introduction was the most

influential work on Buddhism to be written during the nineteenth century. In

important ways, it set the course for the academic study of Buddhism, and especially

Indian Buddhism, for the next century. 17 One of the ways in which it did so was by

setting Buddhism firmly in not only an Indian, but a Brahmanical Indian context.

This legacy, in particular, is felt still to this day. I do not mean to imply, however,

that Burnoufs legacy in this respect is yet another Orientalist sin of our forebears to

be expunged from current scholarship. The flaws of Burnoufs argument

notwithstanding, his conclusion has withstood the test of time, insofar as the

discovery of even older texts in Pali has failed to turn up evidence for a pre-

16
For example, while the texts of the Pali Canon certainly refer to Brahmanical gods (Sakka,
Brahm) and personages (Brahmans themselves), they do not refer, generally speaking, to as wide a
variety of gods as Burnouf mentions finding in his textsNryaa, iva, Varua, Kuvera, Brahm
or Pitmah, akra or Vsava, Hari or Janrdana, akara, which is only another name for iva, and
Vivakarman (Burnouf, Introduction, 163)some of whom are clearly of a more classical
provenance. Recently Bronkhorst has made a convincing argument that, far from Buddhism and
Brahmanism becoming increasingly separated with time, the spread of Brahmanical ideology
throughout India led to an increasing incorporation of Brahmanical themes into Buddhist texts
Johannes Bronkhorst, Buddhism in the Shadow of Brahmanism (Leiden: Brill, 2011), 153-170.
17
Buffetrille and Lopez, Introduction, 1.

11
Brahmanical Buddhism. 18 More importantly, though, Burnoufs conclusion, when

understood within the framework of the scholarly discourse of his day, represents a

real advance in our understanding of the history of Indian religions. Prior to Burnouf,

scholars struggled to answer the simple question of which came first: Buddhists or

Brahmans. Although this is a problematic question, and more recent scholarship

(including this dissertation) has necessarily explored the theoretical issues implicated

within it, on a certain very basic level of chronology there is a simple answer, namely

that arrived at by Burnouf: Brahmans came first. Again, this is a problematic

statement whose implications need to (and will in this dissertation) be parsed, but at

the very least the oldest Vedic texts certainly pre-date the earliest Buddhist texts and

the probable date of the Buddha. Burnouf made an indispensible contribution in

situating early Buddhism in a Brahmanical Indian context; the task of subsequent

scholarship has been to determine how best to nuance and even problematize

Burnoufs insight.

I.1.2 Oldenberg, the Buddha as Reformer, and Historicism

One of the earliest ways in which Europeans elaborated upon the conclusion

that Brahmanism was prior to Buddhism was by superimposing a European religious

narrativenamely, Protestantisms narrative of its own originson the situation in

early India. In his British Discovery of Buddhism, Philip Almond argues that in

18
I believe that this statement holds in spite of the recent critique by Johannes Bronkhorst in
Greater Magadha: Studies in the Culture of Early India (Leiden, Brill: 2007). Nevertheless,
Bronkhorst provides several insights that must nuance our understanding of the priority of
Brahmanism, which I will explore below.

12
nineteenth century England [i]t was perhaps inevitable that the Buddha, qua

religious reformer, should be compared with Martin Luther, and that Buddhism

should be compared with the Protestant Reformation. 19 As we have seen above, the

rhetoric of reform had already influenced the scholarly discourse on the

relationship between Buddhism and Brahmanism prior to Burnouf; Joinville had

argued for the chronological priority of Buddhism on the basis that vegetarianism,

found among Brahmans, represented a reform of an earlier meat-eating practice, as

found among Buddhists. Once it became established that Brahmanism was

historically prior to Buddhism, however, it became easy to construct a narrative with

the Buddha serving as the Luther of India. The Brahmans were a priestly class; they

were associated both in ancient literature and in modern practice with ritual; and the

Buddha in some sense rejected the pretensions of the Brahmans. As Almond

argues, this narrative served as an anti-Catholic polemic at a time when anti-Catholic

sentiment in England was particularly high. 20 It also laid the basis for an Orientalist

discourse in which Hinduism was painted in a negative light as an analogue to the

popery of European Catholicism, and an essentialized portrait of Buddhism based

on a selective reading of early texts was used to criticize by comparison both Hindu

practice and the degenerate practices of contemporary Buddhists.

19
Almond, British Discovery of Buddhism, 73.
20
Ibid.

13
 In his Introduction, Burnouf had already warned against a simplistic portrayal

of the Buddha as a social reformer who abolished caste. 21 A real turning point in the

scholarly discourse on the Buddha as reformer 22 was not achieved, however, until

Hermann Oldenberg, in his Buddha: His Life, His Doctrine, His Order, argued

forcefully against the comparison of the Buddha to Martin Luther:

Above all it must be borne in mind that Buddha did not find himself
like other reformers face to face with a great, united power, capable of
resistance, and determined to resist, in which was embodied the old which he
attacked and desired to replace by the new.
People are accustomed to speak of Buddhism as opposed to
Brahmanism, somewhat in the way that it is allowable to speak of
Lutheranism as an opponent of the papacy. But if they mean, as they might be
inclined from this parallel to do, to picture to themselves a kind of
Brahmanical Church, which is assailed by Buddha, which opposed its
resistance to its operations like the resistance of the party in possession to an
upstart, they are mistaken. Buddha did not find himself in the presence of a
Brahmanical hierarchy, embracing the whole people, overshading the whole
popular life.
Thus Brahmanism was not to Buddha an enemy whose conquest he
would have been unable to effect. He may often have found the local
influence of respected Brahmans an obstacle in his path, but against this a
hundred other Brahmans stood by him as his disciples or had declared for him
as lay members. Here no struggle on a large scale has taken place. 23

21
Buffetrille and Lopez, Introduction, 17.
22
Within popular discourse, on the other hand, the narrative of the Buddha as reformer is of
course alive and well. In India, it takes a particular political valence; Richard Gombrich laments that
in lecturing at Indian universities he has found the view that the Buddha was born a Hindu and was
a Hindu reformer to be virtually universalRichard Gombrich, How Buddhism Began: The
Conditioned Genesis of the Early Teachings (New Delhi: Munshiram Manoharlal Publishers, 2007), 15.
I would argue that the reformer narrative continues to haunt Western scholarly discourse as well; it
remains an obvious comparison because Protestantisms myth of origins is so inextricably tied up in
the Western conception of religion on which the modern study of Buddhism is based. This is one
key reason that I think establishing a sophisticated theoretical understanding of the relationship
between Buddhism and Brahmanism is so important.
23
Hermann Oldenberg, Buddha: His Life, His Doctrine, His Order, trans. William Hoey
(London: Williams and Norgate, 1892), 170-172.

14
Almond notes that after Oldenbergs book was translated into English in 1882, the

scholarly literature records repeated references to his work as having conclusively

proven that the Buddha was not a social or political reformer. Almond argues that

this rather dramatic change in scholarly opinion was the result of an attempt to

protect the Victorian Buddha from being perceived as an early proponent of those

forms of socialism that were perceived by many as threatening the structure of

English society from the beginning of the 1880s especially. 24 Regardless of the

specific reasons why Oldenbergs argument became so widely accepted, on its own

merits it represents the addition of a significant level of nuance to the scholarly

understanding of Buddhisms Brahmanical context. Burnouf had perhaps

inadvertently laid the basis for the reformer narrative by emphasizing so

emphatically the thoroughly Brahmanical character of the world in which

Brahmanism arose. Oldenbergs contribution is to point out that, while there is ample

evidence that Brahmanism was present when Buddhism came into being, it hardly

represented a monolith. His argument is not so much that the Buddha was not a

reformer per se (and therefore he cannot be used as a champion of socialism), but

rather that Brahmanism did not represent the sort of widespread, hegemonic

institution that one could reform in the first place.

There is another, closely related respect in which Oldenbergs work represents

a major turning point in the scholarly discourse on the relationship between

Buddhism and Brahmanism. In his Buddha, Oldenberg rejected the theory,

24
Almond, British Discovery of Buddhism, 75.

15
propounded by mile Senart in the latters Essai sur la lgende du Buddha, that the

Buddhas life story was simply a compilation and transformation of ancient Indian

motifs centering on a solar deity. 25 In advancing this theory, Senart had effectively

subsumed the history of Buddhism under the history of Brahmanism by showing the

antecedents of the formers mythology within that of the latter and arguing that

Buddhisms foundational myth could be understood as little more than a

transformation of Brahmanical themes. The Dutch scholar Hendrik Kern took this

argument a step further, arguing not only that the story of the Buddha was derived

from ancient Indian solar mythology, but that it was wholly fabricated from ancient

Indian solar mythologythat is, that the historical Buddha never existed. Kerns

more radical theory never was taken very seriously by the broader scholarly

community, but Senarts was, and in rejecting it, Oldenberg, along with T. W. Rhys

Davids, with whom he is recognized for founding the Pali school of Buddhist

studies, set the tone for the next fifty years of scholarship on early Buddhism.

Oldenberg argued that the Pali texts, which by his time had come to be understood as

containing the oldest evidence for early Buddhism, were reliable and that they could

be used to reconstruct the life of the historical Buddha. 26 Although Oldenbergs

confidence in our ability to know much about the life of the Buddha himself would
25
In advancing this theory, Senart was employing a method that had been pioneered in large
part by Max Mllersee J. W. deJong, A Brief History of Buddhist Studies in Europe and America,
Eastern Buddhist 7 (1974): 79. Max Mllers comparative method, including the so-called sun myth,
was published in 1856, but was later mocked by the Rev. R. F. Littledale, who showed that, by using
Mllers own methods, one could prove that Mller himself is merely a solar myth. Both Mllers
original essay and Littledales rejoinder are published in F. Max Mller, Comparative Mythology: An
Essay, ed. Abram Smythe Palmer (London: George Routledge and Sons, n.d.).
26
For a more detailed overview of the scholarship of Senart, Kern, Oldenberg, and other
scholars of the late nineteenth century, see J. W. deJong, Brief History of Buddhist Studies, 78-81.

16
not be retained by later scholars, his confidence, more generally speaking, in our

ability to come to positive historical conclusions based on early (i.e., Pali) Buddhist

texts, and thus formulate an understanding of the history of Buddhism on its own

terms, would remain an important legacy of his work in the early twentieth century.

I.1.3 The Early Twentieth Century: From Rhys Davids to Lamotte

The tenor of Buddhist studies in the first half of the twentieth century was in a

key respect set by the publication, in 1903, of Rhys Davids cleverly titled Buddhist

India. The title, which appears at first glance to simply present the books topic

(Buddhism in India, or Buddhist aspects of India), quickly reveals itself upon reading

of the book to be a polemical expression of Rhys Davids thesis. For Rhys Davids

sets out in this book not simply to explore the Buddhist contributions to Indian culture,

but to argue that ancient India can be primarily be understood through a Buddhist lens,

that India in ancient times was very much a Buddhist, and not a Brahmanical, India.

Rhys Davids main criticism of previous scholarship in this book is that scholars have

relied too heavily on Brahmanical texts in reconstructing the history of early India; he

believes that the polemical and normative (to use a more modern term) character of

these texts gives the false impression that the only recognised, and in fact universally

prevalent, form of government was that of kings under the guidance and tutelage of

priests. But the Buddhist records, amply confirmed in these respects by the somewhat

17
later Jain ones, leave no doubt upon the point. 27 Rhys Davids argues that just as

much, and in many ways more, can be learned about early India from the Pali

Buddhist texts, which contain a wealth of information on ordinary customs and

religious practices, unencumbered by a totalizing Brahmanical ideology. 28

The conclusion that Rhys Davids comes to by constructing an understanding

of early Indian society on the basis of Pali Buddhist texts instead of Brahmanical texts

is that the viewpoint offered by the Brahmanical texts is little more than polemic. He

writes,

The fact is that the claim of the priests to social superiority had nowhere in
North India been then, as yet, accepted by the people. Even such books of the
priests themselves as are pre-Buddhistic imply this earlier, and not the later,
state of things with which we are so much familiar. They claim for the north-
western, as distinct from the easterly, provinces a most strict adherence to
ancient custom. The ideal land is, to them, that of the Kurus and Panchalas,
not that of the Kasis and Kosalas. But nowhere do they put forward in their
earlier books those arrogant claims, as against the Kshatriyas, which are a
distinctive feature of the later literature. The kings are their patrons to whom
27
T. W. Rhys Davids, Buddhist India (New York: G.P. Putnams Sons, 1903), 2. These
themes have been returned to quite recently by Johannes Bronkhorst, as I will discuss below.
28
This includes everything from coins, weights and measures, etc., to cults of tree spirits,
ngas, and the like. The obvious objection to Rhys Davids argument would be that even if
Brahmanical texts give a skewed picture of early India insofar as they are normative, Buddhist texts
are no better because they are normative as well. At the very least we can excuse Rhys Davids because
he was providing a corrective at the time he wrote his book, but I think we can go further and say that
Rhys Davids argument continues to have salience and relevance today. That is, I would agree with
Rhys Davids that the Brahmanical texts obscure the situation in early India in a way that the Buddhist
texts do not; that is, while the Buddhist texts are content to simply subordinate rival cults to the
Buddha (e.g., by having nga kings take refuge in the Buddha, having Brahm urge the Buddha to
preach, etc.), the Brahmanical texts weave a totalizing ideology that penetrates all aspects of (its
depiction of) Indian social life and thus gives a false impression that early India was, as Burnouf
thought, thoroughly Brahmanical. Although, as I will discuss below, I disagree with his main
conclusion, Brian K. Smith amply demonstrates the mechanism by which Vedic texts write
Brahmanical ideology into the fabric of the universe itself in his Classifying the Universe: The Ancient
Indian Vara System and the Origins of Caste (New York: Oxford University Press, 1994). More
recently, Johannes Bronkhorst has argued convincingly that the centerpiece of the Brahmanical
movements strategy was insinuating itself into the Indian social and historical discourse, in Buddhism
in the Shadow of Brahmanism.

18
 they look up, from whom they hope to receive approval and rewards. And it
was not till the time we are now discussing that they put forward claims,
which we find still vigorously disputed by all Kshatriyasand by no means
only by those of noble birth (a small minority of the whole) who happen also
to be Buddhists. 29

The widespread acceptance of Brahmanical ideology, its hegemony over Indian social

discourse, should not be read back into the early Indian context. This does not mean

that Brahmans or Brahmanical claims were not present in early India, only that they

were more claims than reflections of reality:

It is difficult to avoid being misunderstood. So I would repeat that the priests

were always there, were always militant, were always a power. Many of them
were learned. A few of them, seldom the learned ones, were wealthy. All of
them, even those neither learned nor wealthy, had a distinct prestige. But it
is a question of degree. Their own later books persistently exaggerate,
misstate, above all (that most successful method of suggestio falsi) omit the
other side. They have thus given a completely distorted view of Indian society,
and of the place, in it, of the priests. They were not the only learned, or the
only intellectual men, any more than they were the only wealthy ones. The
religion and the customs recorded in their books were not, at any period, the
sole religion, or the only customs, of the many peoples of India. 30

With Rhys Davids, therefore, we have come full circle from Burnouf. Where

Burnouf rightly established the chronological priority of Brahmanism, Rhys Davids

just as rightly pointed out that chronological priority does not imply hegemony. 31

Although I would argue that the most significant contribution made by Rhys

Davids Buddhist India (and that, I think, intended by Rhys Davids himself) is to our

29
Rhys Davids, Buddhist India, 61.
30
Ibid., 159, emphasis mine.
31
The recent publication of Bronkhorsts Buddhism in the Shadow of Brahmanism, which as I
have already noted in significant ways recapitulates (although obviously with more sophistication and
the benefit of a hundred years of additional scholarship) Rhys Davids argument about the misleading
picture of early Indian society provided by Brahmanical texts, demonstrates that Rhys Davids insight
is still as relevant today as it was in 1903.

19
understanding of ancient India as a whole, its most immediate legacy may have been

simply to open a space for Buddhist Studies as its own discipline, to be studied on its

own terms and apart from Brahmanism or Hinduism. Indeed, whereas nineteenth

century scholarship had been very much preoccupied with the question of the

relationship between Buddhism and Brahmanism, Monsignor tienne Lamottes 1958

Histoire du Bouddhisme Indienwhich, as I have already mentioned, represents the

last attempt to produce a comprehensive history of Indian Buddhismhas

comparatively little to say about the issue. In this monumental work of nearly 800

pages, a mere 7 are devoted at the beginning of the book to a discussion of Vedic

antecedents. The rest of the book is devoted to a discussion of the development of

Indian Buddhism on its own terms, with only infrequent reference to any non-

Buddhist context. The final paragraph of the section on Vedic antecedents, in which

Lamotte cites approvingly the words of his teacher Louis de La Valle Poussin,

provides a concise theoretical justification for such an approach:

This movement of ideas, a compromise between two civilizations, developed

during the seventh-sixth centuries in the region of the Middle Ganges. Being
situated more to the east, the lands which were to be the cradle of Buddhism
escaped it for the most part. This explains why the preoccupations of early
Buddhism are relatively remote from the speculations originated by the
Hinduized brhmins. It can be said, as did L. de La Valle Poussin, that the
brhmanism from which Buddhism sprang is not the brhmanism of the
Brhmaa and the Upaniad, but represents, even better than the latter, the
ancient Indian yoga. 32

This statement reflects an increased awareness of the geographical diversity of early

India (most importantly between an orthodox Brahmanical west and a heterodox

32
tienne Lamotte, History of Indian Buddhism: From the Origins to the aka Era, trans.
Sara Webb-Boin (Louvain-la-Neuve: Universit Catholique de Louvain, 1988[1958]), 7.

20
east), a more finely nuanced picture of early Indian chronology (in particular the slow

development and spread of Vedic texts and schools, with a gradual movement from

west to east), and a particular conclusion derived from this increased understanding of

the early Indian data that there is a fundamental separateness between the Buddhist

and Brahmanical traditions that allows the former to be understood largely

independently of the latter.

Although Buddhist Studies certainly came into its own in the early twentieth

century, and to a certain extent this sparked an increasing interest in studying

Buddhism on its own terms, scholarly work on the relationship between Buddhism

and Brahmanism by no means came to an end. Instead, it became more focused on

specific links, particularly in the realm of philosophical ideas. Oldenberg provided an

important impetus for this work by noting in his Buddha the interesting similarity

between Buddhist philosophical ideas and those found in the Upaniads. Debates

ensued over the relationship between early Buddhist philosophy, the Upaniads, and

the Skhya and Yoga philosophical systems. Oldenberg took a skeptical view

toward perceived relationships between Buddhist and Brahmanical philosophy,

arguing that while early Buddhism was probably influenced by Brahmanical thought,

they probably did not have direct knowledge of Brahmanical texts; in addition, he

denied that there was any relationship between Skhya and Buddhism. Hermann

Jacobi, on the other hand, in part based on Avaghoas portrayal of the Buddhas

21
teacher Ara Klma as a Skhyan, 33 argued that Buddhist philosophy was based

on a pre-classical Skhya system. Likewise, beginning with Kern there was interest

in a possible Yogic background to Buddhism, championed most importantly by La

Valle Poussin (as evinced by the quotation cited approvingly by Lamotte above), but

also by Hermann Beckh. Generally speaking, these issues concerning Buddhism,

Yoga, Skhya, and the Upaniads remain unresolved. 34

I.2 Newer Scholarship on Buddhism and Brahmanism

More recent scholarship on the relationship between Buddhism and

Brahmanism, in the latter half of the twentieth century and now continuing into the

twenty-first, has shown an increased and increasing interest in revisiting some of the

foundational questions about Buddhism and Brahmanism in ancient India that were

debated in older scholarship. In part, this has consisted of a continuation of the early

twentieth debates over the relationship between Buddhism, the Upaniads, Skhya,

and Yoga, but with a particular focus on the diversity of teachings on meditation

found in early Buddhist sources and what these suggest about the relationship of the

33
I should note that this piece of evidence, which at first seems quite convincing for a
Skhya influence on Buddhism, comes from a fairly late Buddhist text that has been shown by
Bronkhorst, convincingly in my mind, to be instead yet another example of the way in which
Brahmanical ideology came to colonize the past, even in Buddhist literatureBuddhism in the
Shadow of Brahmanism, 154-156. The articulation of Aras teaching using Skhyin categories in
canto 12 of the Buddhacarita (especially verses 17 and 18) is unique to Avaghoa and not found in,
say, the Pali account, found for example in MN 26, the Ariyapariyesan Sutta.
34
For a more detailed overview of the early twentieth century scholarship on the
philosophical relationships between Buddhist and Brahmanical schools, including references to
important bibliographies, see deJong, Buddhist Studies in the West, 84-87.

22
early Buddhist tradition to Yoga. 35 A key impetus for this line of scholarly inquiry

was provided by La Valle Poussins 1936 article Musla et Nrada: Le Chemin de

Nirva, 36 in which La Valle Poussin noted that there appear to be two rival

approaches to attaining liberation in the early Buddhist texts, one of which seems

closely allied to the Yogic traditions. Important work in dealing with this problem

has been contributed by several scholars, including Andr Bareau, Lambert

Schmidthausen, Geoffrey Samuel, Johannes Bronkhorst, Richard Gombrich,

Tillmann Vetter, Winston King, and Alexander Wynne. I will be discussing their

work in more detail in Chapter III.3, in which I address the appropriation of

Brahmanical terminology in the course of the development of Buddhist soteriological

theory.

Many important contributions to the study of Buddhism and Brahmanism in

recent scholarship have taken the form of close studies of specific references to

Brahmanism in the early Buddhist texts. In this genre, we should first make mention

of a very useful article by K. R. Norman that provides a concise catalog of

Brahmanical terms that are referred to in the Pali Canon. 37 Oliver Freiberger has

35
For a more comprehensive overview of the scholarship on this topic, see Stuart Ray
Sarbacker, The Debate over Dialogue: Classical Yoga and Buddhism in Comparison, ch. 4 in
Samdhi: The Numinous and Cessative in Indo-Tibetan Yoga (Albany, NY: SUNY Press, 2005), 75-
109.
36
Louis de La Valle Poussin, Musla et Nrada: Le Chemin de Nirva, Mlanges chinois
et bouddhiques 5 (1936-7): 189-222.
37
K. R. Norman, Theravda Buddhism and Brahmanical Hinduism: Brahmanical Terms in a
Buddhist Guise, The Buddhist Forum 51 (1991): 193-200.

23
done work on the Buddhist reinterpretation of the word yaja (Vedic sacrifice) 38 and

the comparison of Brahmans to dogs 39 in some early Buddhist stras. Joanna

Jurewicz has written a very interesting paper in which she argues that the Buddhist

doctrine of prattya-samutpda has antecedents in Vedic cosmogonic concepts. 40

And quite recently Brian Black published a paper in the Journal of the American

Academy of Religion in which he argues that there are interesting parallels between

the accounts of the Brahman students vetaketu in the Upaniads and Ambaha in the

Pali Canon. 41

Some more comprehensive works have taken a sociological approach to

early Buddhism that seeks to situate Brahmans and Brahmanism within the social

world in which Buddhism arose. In her Social Dimensions of Early Buddhism, Uma

Chakravarti writes that [t]he brhmaas firm entrenchment in the social world is

very evident in the Pali texts, 42 and remarks that [t]he opposition between the

brhmaas on the one hand, and the samaas as typified by the Buddhist bhikkhu on

38
Oliver Freiberger, The ideal sacrifice. Patterns of reinterpreting brahmin sacrifice in early
Buddhism, Bulletin dEtudes Indiennes 16 (1998): 39-49.
39
Oliver Freiberger, Negative Campaigning: Polemics Against Brahmins in a Buddhist
Sutta, Religions of South Asia 3, 1 (2009): 61-76.
40
Joanna Jurewicz, Playing with Fire: The prattyasamutpda from the Perspective of Vedic
Thought, Journal of the Pali Text Society 26 (2000): 77-103.
41
Brian Black, Ambaha and vetaketu: Literary Connections Between the Upaniads and
Early Buddhist Narratives, Journal of the American Academy of Religion 79, no. 1 (March 2011):
136-161.
42
Uma Chakravarti, The Social Dimensions of Early Buddhism (Delhi: Oxford University
Press, 1987), 43.

24
the other, is a constant feature of the Pali texts. 43 Nevertheless, after a

comprehensive analysis of all recorded social class statuses of people mentioned in

the Pali Canon, she finds that Brahmans make up the largest single social group

within both the sagha (i.e., monks and nuns) and the laity (i.e., those who take

refuge in the Buddha, but do not ordain), which leads her to conclude that many

Brahmans, due to the fact that they were already a religious group, were attracted to

Buddhism in spite of the general antagonism of Brahmans to asceticism. 44

A more satisfying answer to this problem is provided by Tsuchida Ryutaro,

who more openly suggests that the assumed opposition of Buddhism and

Brahmanism must be questioned. He suggests that there are in fact two distinct types

of Brahmans found in the early Buddhist textsrich householders

(brhmaagahapatika) and matted-hair ascetics (jaila)and that while the Buddha

was highly critical of the former, he was generally sympathetic to the latter.

Moreover, even rich Brahman householders were welcome to become followers of

the Buddha as long as they recognized the authority of his teaching. 45

Greg Bailey and Ian Mabbett, on the other hand, argue in their Sociology of

Early Buddhism that the relationship between the early Buddhists and Brahmans can

be understood in oppositional terms if done so with appropriate nuance. They suggest

the usefulness of the concept of marketing, which they define as the deliberate

43
Ibid., 41.
44
Ibid., 145-6.
45
Tsuchida Ryutaro, Two categories of Brahmins in the early Buddhist period, The
Memoirs of the Toyo Bunko 49 (1991): 51-95.

25
application of a panoply of techniques to parade, in an intentionally persuasive

manner, the ideology each group claimed to embody and its corresponding

lifestyle. 46 They argue that most of the way in which Brahmans are treated in the

early Buddhist texts, including the frequent mention of Brahmans entering the sagha

or becoming lay followers, can be understood as marketing to present the Buddhist

sagha as a better alternative in the competition with Brahmans for royal patronage.

Most importantly, they argue,

the Buddha identified himself as a protagonist of brhmaical ideals, revived

and purified. His teaching could be seen as wholly consistent with
brhmaical ideals understood in a special sensethe sense which in fact
corresponded, not to the values and institutions of brahmins in ordinary social
life, with their rituals and their memorized texts, or even to the priesthood of
the ancient Vedic pastoral society, but to the ascetic Brahmin programme, a
programme which itself has originated within the priestly fold probably in
large part as a response to the influence of the heterodox ramaas and could
be accommodated within brhmaical orthodoxy only at the cost of much
tension and ideological indigestion. 47

This argument is very much indebted to the earlier work of Richard Gombrich, who

together with Johannes Bronkhorst has produced some of the most comprehensive

scholarship on the issue of the relationship between Buddhism and Bramanism in

recent decades. Gombrichs and Bronkhorsts theories in many respects offer the two

major contrasting paradigms for understanding this relationship available in current

scholarship; therefore, I will now offer a more in-depth overview of their work so as

to contextualize my own argument.

46
Greg Bailey and Ian Mabbett, The Sociology of Early Buddhism (Cambridge, England:
Cambridge University Press, 2003), 109.
47
Ibid., 123.

26
I.2.1 Gombrich: Buddhism as Reaction to Brahmanism

Richard Gombrich is one of the most prolific of recent scholars studying the

relationship between Brahmanism and Buddhism, and he has highlighted in his work

many of the most concrete commonalities between the two traditions. In his most

extensive work on the subject, How Buddhism Began, Gombrich argues that

[t]o see the genesis of the Buddhas teaching as conditioned by the religious
milieu in which it arose is to adopt a truly Buddhist viewpoint which I also
believe to be good historiography. It is also to take a middle way between the
view that Buddhism is just a form of Hinduism and the view that it owes
nothing to its Indian background. 48

In particular, Gombrich argues in the same volume that to properly understand the

Buddhist teaching on antman, we must understand the Upaniadic doctrine of tman

that it was a response to. 49 Gombrich therefore comes down firmly in favor of the

view that the Buddha, or the authors of the early Buddhist texts, did have knowledge

of the Upaniads in some form. He laments the fact that as recently as 1927 no less

a scholar than Louis de La Valle Poussin was able to write that he believed that the

Upaniads were not known to the Buddhists. 50 Gombrich argues that the Buddhist

teaching of antman is only intelligible in light of the Upaniadic teaching of tman,

which he takes it to be an argument against. To this end, he cites approvingly K. R.

Normans article A note on Att in the Alagaddpama Sutta, which makes the

48
Gombrich, How Buddhism Began, 14. Gombrichs more recent bookWhat the Buddha
Thought (London: Oxford Center for Buddhist Studies, 2009)is largely a recapitulation of the
arguments in How Buddhism Began, but with more strident rhetoric on the coherence of the
Buddhas thought as discernible from the early Buddhist texts.
49
Gombrich, How Buddhism Began, 15-16.
50
Ibid., 14.

27
argument that a reference to the uniquely Upaniadic doctrine of the equivalence of

tman and brahman can be found in a Buddhist sutta. 51 Likewise, Gombrich argues

that the Buddhist doctrine of karma, which is identified with intention (as opposed

to its literal meaning of action), is the Buddhas answer to brahmin ritualism, 52

and he sees dependent origination as the Buddhas answer to Upaniadic

ontology. 53 Both the Buddha 54 and the Upaniads address the problem of the

changing, unsatisfactory world of phenomena, but while the Upaniads seek

liberation in an ontological state beyond it, the Buddha is uninterested in ontology

and seeks instead to understand the process of karma so as to bring it to an end. 55

Gombrich is perhaps best known, with respect to his work on Buddhism and

Brahmanism, for his argument that the Buddha made use of metaphor, allegory, satire,

and even humor in his teaching to cast his ideas in terms understandable to his

audience (i.e., Brahmanical terms) and cleverly recast, redefine, or at times simply

ridicule those terms in order to make his argument. In Recovering the Buddhas

51
K. R. Norman, A note on Att in the Alagaddpama Sutta, Studies in Indian Philosophy:
A Memorial Volume in Honour of Pandit Sukhlaji Sanghvi, LD series 84 (Ahmedabad, 1981), 19-29.
52
Gombrich, How Buddhism Began, 51.
53
Ibid., 48.
54
Gombrich states that he uses the word Buddha as shorthand for the authors of the early
texts (How Buddhism Began, 4), but he has been criticized, in particular by Schmithausen (see
Gombrichs account and apologia at How Buddhism Began, 11), for naively accepting that the early
Buddhist texts contain the actual teachings of the historical Buddha. Indeed, given Gombrichs
rhetoric, it is at times difficult to accept that his use of the word Buddha is just shorthand: He writes,
for example, that [t]he Buddha was the great communicator, the supreme master of skill in means,
and yet he correctly foresaw that even he would not be able to preserve his teaching from corruption
(How Buddhism Began, 26), and the very title of his latest book claims to reveal What the Buddha
Thought.
55
Gombrich, How Buddhism Began, 31-33.

28
Message, Gombrich notes that [a]gain and again we find that the Buddhas

references to brahmins and brahminism are humorous and satirical, and, in support

of his argument that much of what is found in the early texts may very well go back

to the Buddha himself, asks, Are jokes ever composed by committees?56 In How

Buddhism Began, Gombrich argues that the central metaphor of early Buddhism

that of fire, which is used by the Buddha in his second sermon to explain antman

and is intrinsic to the idea of nirva (lit., blowing out) itselfis a deliberate

response to the central role that (literal) fire plays in Brahmanical ritual. 57 Finally, in

The Buddhas Book of Genesis, Gombrich argues that the cosmogony found in the

Aggaa Sutta was not originally meant to be taken literally, as it was taken by the

later Buddhist tradition, but was intended as a satire of the Brahmanical cosmogony

found in the Bhadrayaka Upaniad. 58

Gombrich thus explains Buddhist references to Brahmanical concepts

primarily in terms of Buddhism being a reaction to Brahmanism. According to

Gombrichs account, the Buddha co-opts and redefines Brahmanical terminology,

devises philosophical rebuttals to Brahmanical doctrine, and even tells jokes to

satirize Brahmanical mythology. Gombrichs paradigm for conceptualizing the

56
Richard Gombrich, Recovering the Buddhas Message, in Buddhism: Critical Concepts
in Religious Studies, ed. Paul Williams, vol. 1 (New York: Routledge, 2005), 113-128. I find
Gombrichs argument here puzzling. Gombrich appears to be unaware of the existence of an entire
industry devoted to the institutional production, dissemination, and scientific study of what in many
cases amounts to humor. Madison Avenue is living proof that jokes (though not necessarily good ones)
are indeed composed by committees.
57
Gombrich, How Buddhism Began, 65-72.
58
Richard Gombrich, The Buddhas Book of Genesis? Indo-Iranian Journal 35 (1992):
159-178.

29
relationship between early Buddhism and Brahmanism, along with those that follow it

(such as Bailey and Mabbetts theory of marketing), therefore represents a

continuation of the strain of scholarly thought, present since Burnouf, that emphasizes

the historical priority of Brahmanism and its importance for understanding the

context in which Buddhism arose. Bronkhorst, whom I will discuss next, represents a

diametrically opposed strain of thought, championed in particular by Rhys Davids,

that seeks to deemphasize the role of Brahmanism in determining the context of

Buddhism and posits a large chasm between the worlds of Brahmanism and

Buddhism in early India.

I.2.2 Bronkhorst: Greater Magadha and the Brahmanical Project

Bronkhorsts Greater Magadha, based on but ultimately going far beyond his

much earlier work in The Two Sources of Indian Asceticism, 59 is a bold attempt to

question nearly all of our assumptions about the chronology of early Indian religion

and entirely rethink the way in which Brahmanism developed in conversation with

non-Brahmanical traditions. The core thesis of Bronkhorsts book is that in the late

first millennium BCE we can speak of a distinct geographical region in the eastern

Gangetic basin, which he calls Greater Magadha, that was characterized by a

distinct culture and set of religious traditions separate from those of the Vedic

Brahmans. This religious culture was characterized by the worship of round

sepulchral moundspreserved in Buddhism as the stpa cultbut more importantly

59
Johannes Bronkhorst, The Two Sources of Indian Asceticism (Bern: Peter Lang, 1993).

30
by the belief in rebirth, karmic retribution, and the existence of an immutable self

(tman) that is liberated through austeritiesbeliefs partially reflected in Buddhism,

but more clearly preserved in Jainism and, to the extent that we know about it,

jvikism. 60 Bronkhorst argues that some of the features of Greater Magadhan

religion, in particular the belief in an tman, karmic retribution, and rebirth, were

adopted later on by Brahmans as they entered into Greater Magadhaa region that

before the time of Manu was regarded as un-ryanand thus came into classical

Brahmanical doctrine from the outside, and not from within the Vedic tradition. 61

This is at odds with the prevailing scholarly viewpoint, which is that

Brahmanical texts that refer to tman, karma, and rebirth (i.e., the early Upaniads)

predate any non-Brahmanical texts that refer to those doctrines, and thus they

developed within the Vedic tradition and were then adopted or questioned by extra-

Vedic traditions. Bronkhorst counters this received narrative in two ways. First, he

presents a close study of the early Brahmanical texts in question and shows that not

only do they not articulate the aforementioned doctrines as clearly as they are made

out to; in addition, the same texts themselves admit that those doctrines came from

non-Brahmanical sources. 62 Second, in the last part of his book, Bronkhorst argues

that many late Vedic texts are not as old as they are generally held to be, and he

60
Johannes Bronkhorst, Greater Magadha, 1-74.
61
Ibid., 75-174.
62
Ibid.

31
points out that many of even the latest of them (e.g., the Ghya Stras) evince no

knowledge of the doctrines of karma and rebirth. 63

Bronkhorsts theory thus represents a sophisticated and innovative, if still

somewhat controversial, application of models of both substratal borrowing and

antagonistic confrontation directly to the pre-classical period in which Brahmans and

non-Brahmanical sectaries were still in the early process of formulating their self-

identities vis vis one another. Moreover, to the small extent that he addresses the

specific relationship between early Buddhism and Brahmanism, Bronkhorst

represents an interesting counter-narrative to that offered by Richard Gombrich.

While Gombrich seeks to explain the presence of Brahmanical doctrines in the early

Buddhist stras purely in terms of polemical appropriation from the Brahmanical

tradition, Bronkhorsts suggestion that two of the doctrines Gombrich focuses on

karmic retribution and tmanwere not even Brahmanical to begin with, but rather

came from the Greater Magadhan substratum, calls the application of a single model

of antagonistic appropriation into question. Indeed, as Bronkhorst notes, debates over

karma and tman in the early Buddhist stras are not clearly linked to Brahmanical

interlocutors, but rather to Greater Magadhan sects such as the Jains. 64 In addition,

Bronkhorst convincingly refutes Gombrichs assertion that the author(s) of the early

Buddhist stras must have known certain Brahmanical texts; in particular, the fact

that the Pali suttas repeatedly portray Brahmans as saying that they are descendents

63
Ibid., 175-264.
64
Ibid., 217.

32
of Brahm demonstrates, contra Gombrich, that they did not know the Purua Skta

directly, for in that text Brahmans are born from the mouth of Purua, not Brahm. 65

In his latest book, Buddhism in the Shadow of Brahmanism, Bronkhorst

addresses more directly the relationship between Brahmanism and Buddhism

specifically. He begins from the assumption, based on his earlier work in Greater

Magadha, that Buddhism was not a reaction against Buddhism, because, although he

admits that many Vedic texts already existed in oral form at the time of the Buddha,

the bearers of this tradition, the Brahmins, did not occupy a dominant position in the

area in which the Buddha preached his message, [so] this message was not, therefore,

a reaction against brahmanical thought and culture. 66 Bronkhorst argues that in its

early history, Buddhism faced little threat from Brahmanism; the kings who united

much of India shortly after the time of the Buddha, in particular Aoka, were hostile

to Brahmanism and favorably disposednot only in terms of sentiment, but in terms

of patronage and policytoward ramaic religions such as Buddhism. 67 Buddhism

at that time spread and evolved mostly in line with the Greater Magadhan

preoccupations (karma and rebirth, the debate over tman, the stpa cult) from which

it had sprung.

Bronkhorst argues that the references to Brahmans in the early Buddhist

stras (i.e., frequently cited suttas from the Pali Canon) are probably relatively late

65
Ibid., 213-214.
66
Bronkhorst, Buddhism in the Shadow of Brahmanism, 1.
67
Ibid., 12-25.

33
and in any case represent an early Buddhist response to what he calls the new

Brahmanism. 68 The rise of this new Brahmanism is, according to Bronkhorst,

closely linked with the process of Sanskritization, that is, the spread of the use of

Sanskrit as a prestige language of literature and political expression, beginning with

the well-known inscription of the Katrapa Rudradman dated to around 150 CE. 69

Bronkhorst thus disagrees strongly with Sheldon Pollocks thesis, expressed most

comprehensively in The Language of the Gods in the World of Men, 70 that the rise of

what he calls the Sanskrit cosmopolis was a political process unconnected with

religion. 71 The classical Sanskrit that was spread by the new Brahmanism was not

simply the same language preserved in the old Vedic texts (though it was portrayed as

if it were 72); nor was the ideological content of the new Brahmanism simply a

continuation of the old Vedic tradition:

The Brahmanism that succeeded in imposing itself, and its language, on

regions that had thus far never heard of it was a reinvented Brahmanism. It
was not a simple continuation of the vedic priesthood, but something new that
proposed far more than simply executing sacrifices for rulers who needed
them. Brahmanism had become a socio-political ideology, but one that
disposed of a number of tools in the service of the one ultimate goal:
establishing the superiority of the Brahmins in all domains that the ideology
claimed were theirs. 73

68
Ibid., 27-42.
69
Ibid., 50.
70
Sheldon Pollock, The Language of the Gods in the World of Men: Sanskrit, Culture, and
Power in Premodern India (Berkeley: University of California Press, 2006).
71
Bronkhorst, Buddhism in the Shadow of Brahmanism, 50-65.
72
Ibid., 147-148.
73
Ibid., 65.

34
Brahmanism reinvented itself in two ways. First, it wrote itself into the pasteven to

the point of giving itself a role in administration of the Mauryan, and most certainly

anti-Brahmanical, king Candraguptawith the intent of lending itself an air of trans-

historical importance. 74 Second, it borrowed heavily from the ascetic traditions of the

ramaic movements, in particular by developing the literary trope of the rama, or

hermitage, in order to better appeal to wealthy political leaders as a worthy

recipient of patronage. 75

Buddhism had to respond to the challenge posed by the new Brahmanism,

though Bronkhorst warns that the confrontation that ensued should not be conceived

of as a confrontation between two religions. Buddhism and Brahminism, he

writes, were not two religions in competition. Buddhism was rather a religion which

had to adjust itself to this particular social order. Buddhism may not always have

liked the brahmanical social order, but it could live with it where the latter was

imposed. 76 Bronkhorst argues that it was not the task of Brahmanism (unlike

Buddhism) to convert people, but simply to convince people to accept their vision of

society, with themselves at the top. They did this primarily by speaking of society in

terms of the vara system, which worked insidiously on Indian social discourse:

Although at first this system was rejected as inconsistent with social reality (as

74
Ibid., 65-74.
75
Ibid., 74-97.
76
Ibid., 111.

35
evinced by the early Buddhist stras), eventually the mere fact that people were

talking about it led to it becoming accepted as a normative vision of society. 77

Buddhists were unprepared to deal with this threat, Bronkhorst argues,

because they, throughout most of their history in India, were mostly content to focus

their academic energies on soteriological (lokottara) concerns, rather than the

mundane (laukika) concerns of social theory, political theory, even science and

mathematics. 78 Brahmans were happy to fill in this gap, which is why, according to

Bronkhorst, Sanskrit became ascendant throughout South and Southeast Asiakings

employed Brahmans as technocrats, and Brahmans, because of their ideological

agenda, insisted on using Sanskrit. The result was that the Buddhists, who had long

been heavily dependent on royal patronage, had to switch over to the use of Sanskrit

as well to be able to function in royal court settings. 79

More insidiously, though, as Brahmanical ideology became increasingly

normalized in Indian society, Buddhism itself became Brahmanized. Whereas

earlier Buddhist works portrayed the Buddha as living in a non-Brahmanical world,

with Brahmanical ideas criticized and marginalized, later works, most notably the

Buddhacarita of Avaghoa, portrayed him as living in a thoroughly Brahmanical

world, imbued with Brahmanical customs, philosophy, and ideas about society and

77
Ibid., 40-42.
78
Ibid., 99-122.
79
Ibid., 122-130.

36
kingship. 80 Bronkhorst therefore characterizes the new Brahmanical ideology as a

Trojan horse 81 that ultimately led to the downfall of Buddhism in India:

We may conclude that the brahmanical victory over Buddhism in the Indian
subcontinent has been complete. Either Buddhism disappeared altogether or,
as in the case of the Newar Buddhists, it survived in brahmanical shape.
Buddhism had come to think of itself as a deviation from Brahmanism, and of
Brahmanism as the default condition of Indian religion and society. 82

Between his work in Greater Magadha and Buddhism in the Shadow of Brahmanism,

Bronkhost therefore presents a comprehensive framework for understanding the

origins, development, and ultimate downfall of Buddhism in India without positing a

fundamental opposition to Brahmanism.

I.3 Developments in Indology

Much of the newer scholarship that deals with the issue of the relationship

between Buddhism and Brahmanism, not least of all Bronkhorsts work, has been

made possible by developments in Indology over the last few decades that have called

into question the antiquity of many of the elements of classical Hinduism and thus

further problematized the category of Brahmanism itself for the purposes of

comparison with Buddhism. The work of Michael Witzel can be mentioned first here

because it provides a chronological and theoretical framework for understanding the

80
Ibid., 153-170. Bronkhorst notes (p. 168, n. 232) that this phenomenon makes it
unsurprising that Burnouf, who was confined to late Sanskritic sources, would come to the conclusion
that the Buddha lived in a thoroughly Brahmanical world.
81
Ibid., 168.
82
Ibid., 170.

37
compilation, development, and dissemination of the earliest Brahmanical texts, the

Vedas. Witzels method is to use internal evidence in the Vedic texts (geographical,

climatic, and technological references, as well as linguistic features) to arrange them

according to geographical and historical provenance. He finds that the earliest of the

Vedic texts, that is, the oldest parts of the g Veda, appear to be focused on the

Punjab, while later sahit texts evince a shift to the Kuru-Pacla region, and still

later texts appear to show a shift even further to the east. 83 Witzel divides the Vedic

textual material into three chronological strata: Early Vedic (corresponding to the g

Veda), Middle Vedic (corresponding to the mantras, Yajur Vedic prose, and early

Brhmaas), and Late Vedic (corresponding to the late Brhmaas, rayakas,

Upaniads, and Stras). He argues that the early Vedic material corresponds to a time

of competing tribes in the Punjab who had a little-developed tradition of sacrifice and

little or no caste-organization. Later, as the Vedic tribes moved east, the Kurus,

whom Witzel calls the first Indian state, consolidated their power by exacting

tribute (bali), which they redistributed through a newly reorganized and expanded

rauta ritual system. This resulted in the formation of a highly specialized ritual

priesthood, who in turn produced the literature of the middle Vedic period and

developed the sophisticated institutional system of khs for preserving and

transmitting the growing corpus of Vedic texts. These Vedic khs spread further

east, where they produced the literature of the late Vedic period, under the patronage

83
Michael Witzel, On the Localization of Vedic Texts and Schools (Material on Vedic
khs, 7), in India and the Ancient world. History, Trade and Culture before A.D. 650. P.H.L.
Eggermont Jubilee Volume, ed. G. Pollet, Orientalia Lovaniensia Analecta, vol. 25 (Leuven, 1987),
173-213.

38
of kings there who sought their ritual services to legitimate their rulea process that

Witzel refers to as early Sanskritization. 84

A lot of fruitful work has also been done in recent years on the most important

Brahmanical text to have been composed after the end of the Vedic period, the

Mahbhrata. Much of the scholarly debate on this topic has centered on the

question of when, how, and for how long the Mahbhrata was composed, with,

most recently, Alf Hiltebeitel favoring a single, early date, possibly during the uga

period, for the composition of the MBh., 85 and James Fitzgerald favoring a model of

gradual composition, with a core text composed at around the same time postulated

by Hiltebeitel, followed by gradual additions that were effectively closed by the

dissemination of a version closely corresponding to the Critical Edition under the

Guptas. 86 This debate notwithstanding, there is an emerging consensus, albeit with

some disagreements over details, that the MBh. was composed in large part as a

Brahmanical response to the rise of the anti-Brahmanical empires of the late first

84
Michael Witzel, Early Sanskritization: Origins and Development of the Kuru State,
Electronic Journal of Vedic Studies 1, no. 4 (Dec. 1995). For more detailed examinations of the
linguistic considerations involved in the stratification of Vedic texts, see Michael Witzel, Tracing the
Vedic Dialects, in Dialects dans les litteratures indo-aryennes. Actes du Colloque International
organise par UA 1058 sous les auspices du C.N.R.S avec le soutien du College de France, de la
Fondation Hugot du College de France, de l'Universite de Paris III, du Ministre des Affaires
Etrangeres, Paris (Fondation Hugot) 16-18 Septembre 1986, ed. Collette Caillat (Paris: College de
France, Institut de Civilisation Indienne, 1989), 97-264, and Michael Witzel, The Development of the
Vedic Canon and its Schools: The Social and Political Milieu (Material on Vedic khs, 8), in Inside
the Texts, Beyond the Texts: New Approaches to the Study of the Vedas, ed. Michael Witzel, Opera
Minora, vol. 2 (Cambridge, Mass.: Harvard University Department of Sanskrit and Indian Studies,
1997), 257-345.
85
Alf Hiltebeitel, Rethinking the Mahbhrata: A Readers Guide to the Education of the
Dharma King (Chicago: University of Chicago Press, 2001), 17.
86
James Fitzgerald, Making Yudhihira the King: The Dialectics and the Politics of
Violence in the Mahbhrata, Rocznik Orientalistyczny 54, no. 1 (2001), 68.

39
millennium BCE, and in particular the rule of the Mauryan king Aoka. 87 This

suggests that many of the central themes of the MBh. that would become key

components of classical Hinduismthe supremacy of Brahmans and importance of

protecting and patronizing them, the overall theory of varrama dharma, and in

particular the importance of all varas avoiding mixture and faithfully executing their

assigned dutiesare in fact novel ideological strategies developed by Brahmans

around the turn of the era and even later, which is largely consistent with

Bronkhorsts theory of the new Brahmanism.

Indeed, one of these key components of classical Hindu doctrine has been

shown in detail to have been constructed piecemeal over the very period of time in

question. In The rama System, Patrick Olivelle argues convincingly that the

rama system was originally a set of different lifestyle options that later became a

sequential set of life stages. He argues that the ramas are primarily a theological

construct, 88 and he shows that in the earliest texts that speak of them, the Dharma

Stras, they are portrayed as four vocations one can choose from after completing

the period of Vedic studentship. 89 They were not, Olivelle argues, a defense of

Brahmanical orthodoxy, but rather an innovative legitimation of alternative lifestyles

87
Seminal arguments on this interpretation of the MBh. can be found in Madeleine Biardeau,
Le Mahbhrata: Un rcit fondateur du brahmanisme et son interprtation, 2 vols. (Paris: ditions du
Seuil, 2002); Alf Hiltebeitel, Buddhism and the Mahbhrata: Boundary dynamics in textual practice,
in Boundaries, dynamics, and construction of traditions in South Asia, ed. F. Squarcini (Florence:
Firenze University Press, 2005), 107-131; and Fitzgerald, Making Yudhihira the King.
88
Patrick Olivelle, The rama System: The History and Hermeneutics of a Religious
Institution (New Delhi: Munshiram Manoharlal Publishers, 1993), 7.
89
Ibid., 73-82.

40
that took time to gain acceptance; indeed, even the early Dharma Stras that mention

them are ambivalent as to whether they should be accepted as legitimate or not. 90

Only later, beginning with the Mnava Dharmastra, did the classical rama

system arise, in which the four ramas are interpreted as life-stages instead of

alternatives. 91 It was this latter system that sought not only to accommodate

alternative ascetic lifestyles, but to fully reconcile the values systems represented by

the married householder and the celibate ascetic in what would become a key theme

of classical Hinduism.

Scholarship has been somewhat more conflicted, however, on the question of

the development of ramas sibling in the fully developed system of classical

Brahmanismvara. Several decades ago in several articles including, most

importantly, Brahman, Ritual, and Renouncer, Johannes Heesterman provided

strong evidence that the classical rauta ritual was a relatively late ritualization of an

original agonistic sacrifice in which competing clan leaders traded rolls as

yajamna and officiating priests. 92 This would mean that although the term

brhmaa, for example, is found frequently even in the early g Vedic texts, it need

not be interpreted as a rigidly-conceived vara, determined by birth, as in the

classical theory. Indeed, Heesterman argues that originally the concept of vara did

not prescribe strict separation, but rather a system of connubial and other

90
Ibid., 83-100.
91
Ibid., 131-150.
92
J. C. Heesterman, The inner conflict of tradition: essays in Indian ritual, kingship, and
society (Chicago: University of Chicago Press, 1985).

41
exchanges. 93 Unfortunately, Heestermans theory has remained somewhat

controversial due to the fact that the book in which he most fully fleshed out this

theory, The Broken World of Sacrifice, 94 was widely criticized as relying too heavily

on late evidence (primarily the rauta Stras), to the neglect of the early texts such as

the g Veda that presumably would be more relevant to the time period Heesterman

was addressing.

As a result of this lack of consensus, some scholars continue to hold that the

key tenets of Brahmanical orthodoxy can be found in some of the earliest

Brahmanical texts and therefore the vara system was rigid at an early date. In his

Classifying the Universe, for example, Brian K. Smith argues that throughout the

Vedic texts, Brahmins wrote their class superiority into the universe by using the

vara system to classify the gods, space, time, flora, fauna, and even their own

scriptures. In making this argument, he presupposes that [a]lthough it has

sometimes been argued that originally the social classes we encounter in the most

ancient texts of India were fluid groups, [i]t is far more demonstrable that the three

or four social classes are in the Veda regarded as separate and hereditary. 95 The

evidence Smith adduces in his book, however, does not bear out this claim. Smith

does demonstrate that vara and vara-related categories are equated with a variety

of things throughout the universe, but as many scholars, including Smith himself,

93
Ibid., 199.
94
J. C. Heesterman, The broken world of sacrifice: an essay in ancient Indian ritual. Chicago:
University of Chicago Press, 1993.
95
Smith, Classifying the Universe, 28.

42
have shown, it was a characteristic of Brahmanical philosophical practice since at

least the time of the Brhmaas to seek bandhus between seemingly unrelated

objects. What remains to be demonstrated is that the early Brahmans considered

vara to be primary and constructed bandhus between vara categories and various

objects throughout the universe as an intellectual technology for social domination.

In point of fact, however, most of the examples Smith presents do not involve vara

categories directly, and they do not present the sort of one-to-one mapping of flora,

fauna, and the like onto the four varas that one would expect if the latter were the

dominant paradigm. Far more common than the actual four varas in Brahmanical

classification schemes are the three functionsbrahman, katra, and viwhich

are included in various combinations in schemes that include three, four, five, or even

six categories. As a result, in many cases the four varas or three functions cannot

account in a straightforward way for all of the categories in the classification scheme.

Indeed, the four varas proper, and references to brhmaas in particular, are

not very common across the Vedic literature; far more often one finds reference to the

three functions, especially brahman and katra. When the word brhmaa itself does

appear in the Vedic texts, it is generally not used in a way that unambiguously

indicates that it referred to a hereditary priesthood. What these references do make

clear is that Brahmans performed sacrifices, were seen as possessing special powers,

and were valued for their learning. 96 The earliest and most famous elaboration of the

96
See, e.g., RV X.71.8: hd taeu manaso javeu yad brhma sayajante sakhya /
atrha tva vi jahur vedybhir ohabrahmo vi caranty u tve //When the impulses of the mind are
fashioned in the heart, when Brahmans sacrifice together as friends / They leave one behind with
knowledge, while others possessing brahman ramble through. Note that while this passage indicates

43
four varas in toto, the Purua Skta (RV X.90), makes clear that Brahmans are the

highest vara, but nothing in the text need imply that this or any of the other varas

are determined rigidly by birth. Birth is not brought up frequently in the earliest

Vedic texts, and in one case where it is, a Brahman named Kavaa Aila, who is

spurned by the seers as the son of a slave-woman, a cheat, not a Brahman, is

ultimately welcomed by them when they realize that the gods know him. 97

Statements of the superiority of the Brahman are found, but again, in the earliest texts

they are not connected to birth. 98 Even in the comparatively late Jaiminya and

atapatha Brhmaas, however, Brahmanical supremacy is hardly unambiguous.

The B describes a ritual in which the Brahman makes his royal patron weaker than

himself, but it also says that the Brahman is an object of respect after the king. 99

Likewise, the JB openly acknowledges that both the Brahman and the vaiya are

subject to the katriya. 100 The earliest passage I have found that refers

unambiguously to one being born into a vara is in the Jaiminya-Upaniad

Brhmaa, which speaks of a persons desire to be reborn into a family of Brahmans

that Brahmins were valued for their erudition, it does not imply that there was such a thing as a
Brahmin who had no Vedic education at all. The Brahmin-by-birth (brahmabandhu) of later times was
characterized by a total negligence in the performance of Vedic sacrifices, not by a priestly vocation in
which he merely lacked expertise.
97
AB II.19: dsy putra kitavobrhmaavidur v ima dev. Cf. KB XII.3.
98
See, e.g., AB VII.15, in which Varua declares the Brahmin to be higher than the Katriya.
Even this passage is not particularly early, however; Witzel has shown on independent philological
grounds that the second half of the Aitareya Brhmaa is a late additionMichael Witzel, Tracing
the Vedic Dialects, 115.
99
B V.4.4.15; V.4.2.7.
100
JB I.285.

44
(brhmaakula)but of course the very reference to rebirth itself makes it clear that

this is an extremely late text. Another late passage, from the second half of the

Aitareya Brhmaa, seems to imply fixed varasbut its ultimate import is that a

katriya to whom evil happens will have Brahman-like offspring, a receiver [of

gifts], a drinker [of Soma], a seeker of livelihood, one to be sent away at will. 101

This implies not only that a family lineage can undergo a vara-transformation, but

that a transformation from katriya to Brahman is considered an unlucky fate!

In any case, I believe we can safely say that the work of Witzel has, at the

very least, vindicated Heestermans basic idea that there was a tradition of sacrifice,

whose traces remain in the g Veda, that preceded the codification of rauta ritual,

and that that earlier tradition of sacrifice was not tied to a rigid vara system. Witzel

places the codification of rauta ritual in the Kuru-Pacla region, that is, slightly to

the west of the Buddhist homeland, and somewhat before the time of the Buddha as

well. Although Witzel ties this Kuru-Pacla orthodoxy that emerged from the

middle Vedic literature to a successively stricter stratification into the 3 rya (twice-

born) and the additional dra (aboriginal) classes (vara), 102 Bronkhorsts theory of

the dissemination of the new Brahmanism shows that it is possible to construct a

viable model of early Indian culture in which the ideological content of, and culture

surrounding, the Vedic texts is irrelevant to most of India until the new Brahmanism

spread a reformulated version of those ideas to the rest of India slowly over the

101
AB VII.29: dyy pyy vasy yathkmapraypyo.
102
Witzel, Early Sanskritization, 5.

45
course of many centuries. That is, even if a rigid vara system was developed and in

some sense put into practice in conjunction with the composition of the middle Vedic

literature prior to the time of the Buddha, this development would have been limited

to the Kuru-Pacla region, and would have had no immediate bearing on the rest of

South Asia, including the Buddhist homeland just to the east.

But as I have suggested, there is good reason to doubt that there was a rigid

vara system even in a middle to late Vedic context. Timothy Lubin has recently

published interesting work on the Ghya Stras that shows that it was only in that

fairly late body of literature that we find the idea that all three rya-varas should

receive Vedic initiation and undergo brahmacarya, and concomitantly it is first in

that body of literature that the term dvija (twice-born) is applied to katriyas and

vaiyas. In earlier literature, the term dvija is applied only to Brahmans, and it is

assumed that anyone undergoing initiation and brahmacarya is a Brahman; in fact,

Lubin argues, it may be that undergoing this process is what made one a Brahman. 103

I.4 Methodology of this Study

As I have tried to show in this introduction, the major themes bequeathed to

scholarly discourse by early scholarship on Buddhism and Brahmanismthe

chronological priority of Brahmanism, the inapplicability of a reform model to the

rise of Buddhism, and the insight that Buddhism may have arisen from an extra-

103
Timothy Lubin, The Transmission, Patronage, and Prestige of Brahmanical Piety from the
Mauryas to the Guptas, in Boundaries, Dynamics and Construction of Traditions in South Asia, ed.
Federico Squarcini (Florence: Firenze University Press, 2005), 77-103.

46
Brahmanical worldhas led to a fundamental dichotomy in more recent scholarship.

On the one hand, there are those who, like Gombrich, emphasize the chronological

priority of Brahmanism and the need to understand the rise of Buddhism in a

Brahmanical context. They tend to speak in terms of opposition, reaction, and

even marketing. On the other hand, scholars like Bronkhorst focus on distance

between the Brahmanical and early Buddhist traditions and reject the idea that

Buddhism arose in opposition to Brahmanism. For Bronkhorst, following very much

in the footsteps of Rhys Davids, Buddhism emphatically did not arise in a

Brahmanical world, and Brahmanical texts must be read with a great deal of

skepticism toward their polemical intent, which is in fact to make the reader believe

that they are the product of a thoroughly Brahmanical world.

On the whole, I find Bronkhorsts critiques of oppositional models of the

relationship between Buddhism and Brahmanism to be quite convincing. In

particular, I am convinced by his argument that Buddhism operated in the first few

centuries of its existence in a world that was largely favorable to it and other religious

movements like it (jvakas, Jains) and that was largely hostile to Brahmans, and that

it was only after that period of time that the Brahmans reinvented themselves and

began to spread their ideology of vara throughout India on a large scale. I believe

that this theory of a new Brahmanism is supported by multiple trends in scholarship

on early India, including Witzels elucidation of the development of the Vedic

tradition, Heestermans theory of the agonistic sacrifice, the increasing realization

that the MBh. is a deliberate Brahmanical response to the dark ages of Nanda and

47
Mauryan rule, and other work that has shown the late development of elements of

classical Hindu doctrine (Olivelle on the rama system; Lubin on brahmacarya,

upanayana, and the term dvija).

What I find problematic about Bronkhorsts theory is the sharp dichotomy he

draws between Vedic and non-Vedic traditions, his desire to divorce Buddhism and

the other religions of what he calls Greater Magadha entirely from what at times

appears to be an essentialized Brahmanism. Bronkhorst convincingly argues that

there was a culture in the general area around Magadha that was discernible from the

Vedic culture in the east, and that this region only came to be recognized as ryavarta

in Brahmanical texts after the turn of the era, thus implying a late process of

Brahmanization. Less (if not entirely un-) convincing, however, is Bronkhorsts

attempt to show that the concepts of tman, karma, and sasra, and ascetic

technologies for effecting release from sasra, are wholly products of the Greater

Magadhan religious milieu that came to Brahmanism from outside. At times, in fact,

his argument appears to be somewhat circular. For example, he uses Kas

teaching of karmayoga in the Bhagavad Gta Brahmanical textas evidence of a

non-Brahmanical, Greater Magadhan doctrine of achieving liberation of the tman

through gnosis, thus begging the question of whether this doctrine should be

understood as originally belonging to a Greater Magadhan religious culture separate

from early Brahmanism. 104 Bronkhorst does, on the other hand, have a bit more

success in arguing that the Upaniads are not as good pieces of evidence for a

104
Bronkhorst, Greater Magadha, 35-38.

48
Brahmanical origin of these ideas as they are often thought to be. He convincingly

shows that the monism of the Upaniads did not play a major role in Brahmanical

thought until the rise of the Vednta in the late first millennium CE, 105 and he also

clearly demonstrates the relative simplicity and marginality of karma and rebirth in

the early Upaniads. 106 This, however, still leaves unexplained the fact that the

Upaniads, which are undeniably early (relatively speaking) and undeniably

Brahmanical (since they were preserved in Vedic khs), have, as Bronkhorst

himself admits, a great deal to say about the self, 107 a construct that he would like

to derive outside of Brahmanism, in the religious matrix of Greater Magadha.

Bronkhorsts explanation is that the Upaniads, which are of a generally

eastern provenance, openly admit that their novel teachings came from non-

Brahmans. 108 This argument is not new and is based on certain passages in which

Brahmans go to kings and receive teachings from them that Brahmans previously

were not aware of (CU 5.3.7, CU 5.11-24, BU 2.1). 109 But in a comprehensive study

of the theory of tman in the Upaniads published after Bronkhorsts Greater

Magadha, Brian Black shows that passages in the Upaniads that claim that certain

teachings came from katriyas are simply a literary device. He argues that the

105
Ibid., 140-141.
106
Ibid., 112-126.
107
Ibid., 126.
108
Ibid., 130.
109
These citations are taken from Patrick Olivelle, trans., The Early Upaniads (New York:
Oxford University Press, 1998), 11, which also lists Deussen and Horsch (n. 18) as proponents of the
theory of katriya origins. Olivelle himself does not agree with this theory.

49
claims made by katriya characters do not represent a true expression of a katriya

voice, but rather katriya characters embody brahmin idealizations about the position

of the king. 110 Indeed, we know that the claim that Upaniadic teachings were given

to Brahmans by katriyas cannot be taken literally because many of the teachings in

question can be traced to earlier Vedic texts. 111

Ultimately, what is at issue here is that Bronkhorst wants to posit a strict

dichotomy between the Vedic and the non-Vedic, the Brahmanical and the non-

Brahmanical, at the earliest recoverable stages of Indian history. This dichotomy is

central to one of his earliest works, The Two Sources of Indian Asceticism, the

conclusions of which he reproduces in Greater Magadha. 112 In these works,

Bronkhorst argues there is evidence in the ancient sources for three types of

asceticism in ancient India: one Vedic (a way of life that combines ritual activity and

asceticism) and two non-Vedic (the way of insight into the true nature of the self

and the way of inaction). 113 This is best illustrated in the Indian literature by the

pastamba Dharmastra, whose discussion of the four ramas combines two types

of ascetics, those of Vedic and those of Greater Magadhan extraction under the

rubric of vnaprastha, while parivrja refers only to a non-Vedic form of asceticism

110
Brian Black, The Character of the Self in Ancient India: Priests, Kings, and Women in the
Early Upaniads (Albany, NY: SUNY Press, 2007), 105.
111
Ibid., 101.
112
Bronkhorst, Greater Magadha, 79-93.
113
Ibid., 90. I quote from Greater Magadha since it is Bronkhorsts latest, and presumably
most up-to-date, work on the subject.

50
(the way of insight). 114 It is confirmed by a Greek source, the testimony of

Megasthenes, who reports finding Brachmanes who live in a grove in front of the

city, live in a simple style, and lie on beds of rushes or (deer) skins; and two

types of Sarmanesthe Hylobioi, who live in the woods, where they subsist on

leaves of trees and wild fruits, and wear garments made from the bark of trees, and

the physicians, who are engaged in the study of the nature of man and subsist on

alms. 115 Other sources (mostly Indian) that do not show this dichotomy between

Vedic and non-Vedic asceticism are simply examples of confused terminology and

distinctionsbecome blurred. 116

I believe that Bronkhorsts interpretation of these passages from pastamba

and Megasthenes is wrong. 117 pastamba does not describe two types of ascetics

under the rubric of the vnaprastha. Rather, he describes two alternative opinions on

how the life of a vnaprastha can be lived. The first involves simply living in the

forest and subsisting on roots, fruits, leaves, and grasses, followed ultimately by water,

air, and space (i.e., he starves to death) (DS 2.22.1-4). The second involves an

orderly sequence limited to the forest hermit (DS 2.22.6). This involves

undergoing Vedic studentship, marrying and setting up the Vedic fires, and only later

moving out into the forest, taking the fires with one (DS 2.22.7-24). Olivelle has
114
Ibid., 91.
115
Ibid., 92.
116
Ibid., 91.
117
For a similar critique of Bronkhorsts thesis, as it was originally presented in The Two
Traditions, see Patrick Olivelle, Review of The Two Sources of Indian Asceticism by Johannes
Bronkhorst, Journal of the American Oriental Society 115, no. 1 (1995): 162-4.

51
argued that this fairly unique option presented by pastamba is a way of providing a

space for ritual obligations within the vnaprastha lifestyle, and that this would later

provide a model for the classical rama theory of life-stages. 118 In any case, though,

this modified vnaprastha life-cycle still ends in exactly the same way as the

standard versionby subsisting on roots, fruits, leaves, grasses, and ultimately

water, air, and space (DS 2.23.2).

Likewise, I would argue that Megasthenes testimony attests not to three types

of ascetics, one Vedic and two not, but rather simply the three non-householder

ramas, with the distinction between Brahmans and ramaas imposed on top of

them. That is, the Brachmanes appear not to be ascetics so much as brahmacrins

living in a Brahmanical encampment. What is interesting here is that, as Bronkhorst

has noted elsewhere about Greek testimony, 119 the governing rubric is not a

Brahmanical theoretical construct (which in this case, if one followed the Dharma

Stras, would be the four-fold rama system without any reference to ramaas),

but rather what appears to have been popular categories (Brahmans and ramaas, as

attested in Buddhist, Jain, and epigraphical sources).

The key problem faced by Bronkhorsts strict dichotomy between

Brahmanical and non-Brahmanical, as it pertains to this dissertation, however, is the

Buddhist evidence. While I think that Bronkhorsts arguments for revising the

chronology of late Vedic and early Buddhist texts are quite plausible, they are not
118
Olivelle, The rama System, 113-114.
119
See Buddhism in the Shadow of Brahmanism, 73, where Bronkhorst argues that
Megasthenes description of Indian society according to a seven-fold system instead of the four varas
indicates that the vara system simply was not operative in Candraguptas empire.

52
sufficient to explain away all references to Brahmans and Brahmanical doctrines

found in the early Buddhist stras as late additions that reflect a late encounter

between Western Brahmans and Greater Magadhan Buddhists. Even if Bronkhorst is

right that the concepts of tman, karma, and sasra, and ascetic technologies for

escaping sasra, were indigenous to Greater Magadha and only at a relatively late

date accepted by the Brahmans, Brahmanical terminology is utterly pervasive

throughout the Buddhist texts, including in the earliest strata. If Brahmans were

simply loath to venture as far east as Greater Magadha prior to the time of Patajali,

and the encounter between Brahmans and Greater Magadhan religion occurred mostly

after that point, i.e., around the turn of the era, then we would expect to find Buddhist

polemic against Brahmans suddenly appearing in somewhat late stras such as those

discussed by Bronkhorst, reflecting the late entrance of Brahmans into Greater

Magadha. Instead, as I will demonstrate in this dissertation, what we find is a

progression from a virtual identification on the part of Buddhists with the

Brahmanical tradition to an increasingly antagonistic position in the later literature.

How are we to explain this without returning to a simplistic model of

Buddhism arising in opposition to Brahmanism? I find the theoretical structure

offered by David Seyford Ruegg to be helpful in this regard. Ruegg, in his recent

book on Buddhism and Brahmanism, argues that while the relationship between the

two can be described diachronically using a variety of models, which include models

of antagonism and confrontation, a substratum model serves to explain much

of the common terminology and ideas on the basis of common ancestry, rather than

53
deliberate borrowing, marketing, or the like. In addition, Ruegg argues that the entire

opposition between Buddhism and Hinduism, even synchronically speaking, may be

based on a category errorthat is, inappropriate comparison of the lokottara teaching

of Buddhism as being on the same level with the mostly laukika concerns of ordinary

Brahmanical Hinduism. 120

Ruegg thus provides two crucial insights that I believe can inform almost any

study of the development of Buddhism and Brahmanism vis vis one another in early

India: first, that Buddhism and Brahmanism might not always be best understood as

species of the same genus, and second, that common elements in later formulations of

the two traditions might need to be explained according to multiple models, including

both independent borrowing from a common substratum and direct borrowing in the

course of antagonistic confrontation. Rueggs sources, however, are generally later

than the period I am looking at: He is looking at medieval philosophical sources,

many even taken from the comparatively late Tibetan corpus of texts, and he is

interested in answering questions proper to that time period, such as how we are to

understand the presence of so-called Hindu gods such as Viu and iva in

Buddhist Tantras. In other words, he is mostly interested in looking at Buddhism and

Brahmanical Hinduism in their more fully developed, classical forms, for which

there is more clear evidence, and asking how we should theorize the processes that

led to them. Moreover, Ruegg focuses mostly on the emic Buddhist understanding of

120
David Seyfort Ruegg, The Symbiosis of Buddhism with Brahmanism/Hinduism in South
Asia and of Buddhism with "local cults" in Tibet and the Himalayan region (Wien: Verlag der
sterreichischen Akademie der Wissenschaften, 2008)

54
its own relationship to Brahmanical institutions, and less on constructing an etic

model of the same relationship on the basis of a variety of sources.

But a substratum model can be just as easily applied to the problem of

Buddhist and Brahmanical origins as it can to their latter-day history in medieval

India. The problems posed by Bronkhorsts theory of a Greater Magadha can be

resolved if we abandon the theoretical assumption that early India involved two

completely separate religious traditionsthat of the Brahmans in the West and that of

Greater Magadha in the Eastthat came into contact only in the late first millennium

BCE. I believe that just as much as the material Bronkhorst looks at requires the

postulation of a Greater Magadhan substratum (even if not all agree with Bronkhorst

on the exact content of that substratum), so the evidence from early Buddhist texts

requires the postulation of a pan-North-Indian Aryan/Brahmanical substratum. This

Aryan/Brahmanical substratum is to be distinguished from the orthodox

Brahmanical tradition that is directly observable to us in Vedic texts, and which

clearly had its origins in the West; rather, this substratum would have served as a

common source for both the orthodox Brahmanical tradition in the West and the

Buddhist tradition and other heterodox movements in Greater Magadha.

As I will argue in this dissertation, certain passages in which the Buddha

criticizes Brahmanical claims to superiority by birth seem to be not simply polemical

attacks on established orthodoxy, but rather reasoned criticisms, derived both from

experience and from a sense of common tradition, of at least partially novel claims

being made by a particular group of self-proclaimed orthodox elites. A substratal

55
hypothesis such as I am presenting here would explain why such an argument as this

would arise. If the concept of the Brahman, and indeed the existence of

Brahmans, was well known in Greater Magadha prior to the arrival of orthodox

Brahmans from the West (i.e., the situation described by Bronkhorst in Greater

Magadha), then presumably that pan-North-Indian, substratal understanding of the

Brahman would not fully conform to the self-conception that orthodox Brahmans

brought with them as they moved east and began to encounter Greater Magadhan

culture. If this substratal hypothesis is correct, then the early Buddhist rhetoric on

Brahmans must be read not simply as an attack on an established Brahmanical

orthodoxy, nor simply as a reaction to the new arrival of Brahmans from the West,

but as a reasoned debate over ideals that had pervaded North India for centuries,

between two groups that claimed equal rights to their legacy.

What I am seeking to do with this dissertation, then, is to abandon the idea

that there are essential Brahmanical and Buddhist traditions whose relationship we

must understand. Instead, I am interested in investigating how Buddhism formed, not

as an essence, but as an identity, and in constant dialectic with the simultaneously

emerging Brahmanical identity. Identities are not static, but fluid, so an investigation

of identity is inevitably diachronic, asking not about relationships between things, but

about interrelationships between processes. Likewise, identities are not descriptive,

but polemical constructs, so an investigation of identity asks not what Buddhists and

Brahmans were at any given time, but what they wanted to be. At the same time,

though, identities are greater than the people who articulate them; they are public, and

56
they are contestable. What follows, therefore, is a story of fantasies turning into

realities, and of hopes turning into burdens.

I.5 Outline of the Dissertation

This dissertation is divided into three major parts (Parts II-IV), each of which

tackles a major aspect of the problem of the category Brahman in the early

formation of Buddhist identity. In Part II, I explore the general field of categories for

religious identities that were available in ancient India, in order to situate brhmaa

within that field. After noting in the Introduction that religion and Buddhism

were not emic categories in ancient India, I begin in Chapter II.1 by exploring the use

of the compound samaabrhma in the early Buddhist texts. I show that, contrary

to what the oft-cited passage from Patajali would suggest, this compound is more

often than not not used as an oppositional compound. Then, in Chapter II.2, I turn to

the rama System of the Brahmanical Dharma Stras, and I argue that this system

was not a normative system of Brahmanical practice, but rather a New Brahmanical

description of actual practice, much of which it in turn rejected. Finally, in Chapter

II.3, I show that the categories used in the rama System were used in colloquial

discourse as well, and in fact were appropriated in different ways by the Buddhists.

In Part III, I turn to the category brhmaa itself, in order to make the

argument that this category was not originally owned by a monolithic

Brahmanical tradition, but rather was taken up by both Buddhists and proponents of

the New Brahmanism from a pan-Indo-Aryan substratum and deployed in different

57
ways by each. I begin in Chapter III.1 by reviewing the many cases in the early

Buddhist texts in which the word brhmaa is used in a metaphorical or polemical

sense to contrast the Buddhist ideal to literal Brahmans, i.e., Brahmans by birth. I

then contrast this classical usage to that found in the earliest Buddhist texts, in

which the word appears to be used as a simple honorific, with no polemical overtones.

In Chapter III.2, I then show how these earliest texts were interpreted by the later

tradition, through commentary and narrative framing, to transform the simple early

uses of the word brhmaa into polemics against proponents of the New Brahmanism.

Then, in Chapter III.3, I provide a counter-example, the word tevijja (Skt. traividya),

which I argue was borrowed by the Buddhist tradition from the New Brahmanical

tradition for the sole purpose of polemicizing against the latter. I thus argue that so-

called Brahmanical terms found in the early Buddhist texts cannot be treated

monolithically and unproblematically as such; rather, they must be treated on a case-

by-case and diachronic basis.

Finally, in Part IV, I examine the aspect of the early Buddhist tradition that

has served more than anything else to create a Buddhist identity as separate from

Brahmansa genre of stras that I call encounter dialogs. Because my

methodology depends on the comparative study of Pali and Chinese versions of the

early Buddhist stras as products of an oral tradition, I begin in Chapter IV.1 by

reviewing the Oral Theory of Parry and Lord and demonstrating its applicability to

the early Buddhist tradition. Then, in Chapter IV.2, I review scholarship on the

Nikya/gama traditions and make some arguments about their likely development

58
based on a comparative study of the Pali and Chinese versions. In Chapters IV.3 and

IV.4, I finally turn to the encounter dialogs themselves. In Chapter IV.3, I examine

collections of encounter dialogs in particular Nikya/gama traditions in order to

come to some conclusions about which examples of the genre most likely have the

most antiquity in the tradition. I then extend these conclusions by looking across

Nikya/gama boundaries at the use of particular oral formulae and themes within

the encounter dialog genre. In the end, I conclude that the encounter dialog genre

developed according to a clearly discernible pattern in the early Buddhist tradition,

and that although it was originally deployed in large part to combat the claims being

made by proponents of the New Brahmanism, in the end it served mostly to cede the

category brhmaa to them by depicting the Buddha interacting with Brahman

interlocutors who did not recognize him as a Brahman in return.

59
 Part II

Discursive Categories for Religious Identities

in Ancient India

60
 Introduction

Categories for Religion and Buddhism in Ancient India

Although it is certainly not a new insight, it bears repeating before embarking

upon a study of Buddhist identity formation in ancient India that neither of the

categories used in modern discourse, both emic and etic, about Buddhismthat is,

the genus religion and the species Buddhismcorrespond in any exact way to

the emic terms that were used in many pre-modern, and especially ancient, Indian

contexts. In modern Indian languages, the word dharma has largely taken on the role

of translating the Western category of religion, and in fact it serves as the official

word for religion in the Indian constitution. 121 The pre-modern use of the word,

however, is extremely complex, with a long and variegated history that differs

markedly from that of the Western category of religion, 122 with meanings including

121
Wilhelm Halbfass, Dharma in the Self-Understanding of Traditional Hinduism, in India
and Europe: An Essay in Understanding (Albany, NY: SUNY Press, 1988), 310.
122
For a brief history of the development of the category religion in the West, see Jonathan
Z. Smith, Religion, Religions, Religious, in Critical Terms for Religious Studies, ed. Mark C. Taylor
(Chicago: University of Chicago Press, 1998), 269-284. For a useful discussion of how Western
preconceptions about religion both shaped and were shaped by the colonial encounter with Eastern

61
foundation, rules of ritual, nature, norm, teaching, law, motion, and

even thing. 123 That the use of the word dharma to refer to religion in modern

Indian languages is primarily conventional, and not an inevitable evolution from the

words pre-modern uses, can be seen from comparison to what happened in other

Indic languages spoken outside of India. In Sihla, for example, the word that came

to be settled on to translate religion was gamaa word that in the Pali Canon

referred primarily to traditional or authoritative texts. 124 In Thaiwhich, although

not an Indo-Aryan language like Sihla, is just as Indic as say, Tamil (i.e., it uses

an Indic alphabet and has absorbed a great deal of Sanskrit vocabulary)the word for

religion is (san), a Sanskrit word that is derived from the verbal root s,

which has a semantic range that includes to punish, rule, and govern, but also

teach and instruct. The word dharma is present in Thai as (tham) or

religions, see Richard King, Orientalism and Religion: Postcolonial Theory, India and The Mystic
East (London: Routledge, 1999), esp. ch. 2, Disciplining Religion, 35-61.
123
Two classic treatments of the pre-modern history of the word dharma are, in English,
Halbfasss Dharma in the Self-Understanding of Traditional Hinduism, and, in German, Paul Horsch,
Vom Schpfungsmythos zum Weltgesetz, Asiatische Studien: Zeitschrift der Schweizerischen
Gesellschaft fr Asiankunde, vol. 21 (Francke: 1967): 3161. This work has been greatly expanded
upon recently by a variety of articles published together with a translation into English of Horschs
classic article in a special issue of the 32nd volume of the Journal of Indian Philosophy (2004). Even
more recently, a monumental book-length treatment of the word has appeared: Alf Hiltebeitel, Dharma:
Its Early History in Law, Religion, and Narrative (Oxford: Oxford University Press, 2011).
124
For a discussion of the history of how gama came to mean religion, and buddhgama
Buddhism, in Sihla, see John Ross Carter, On Understanding Buddhists: Essays on the Theravda
Tradition in Sri Lanka, ch. 1, Origin and Development of Buddhism and Religion in the Study of
the Theravda Buddhist Tradition (Albany, NY: SUNY Press, 1993), 9-25. It is interesting to note
that while gama came to mean religion in Sri Lanka, in Thailand, which as a Theravda Buddhist
country has close ties to Sri Lanka, gama () came to mean, in a sense, the opposite: magic or
magic spell.

62
(thamma), though it is used not to refer to the modern concept of religion, but in its

traditional semantic range, to refer, example, to the Buddhas teaching or the reality it

points to, and also to law ( or dharmastra refers to the academic study of

law and is the name of the countrys premier law school). While dharma came to be

assigned the role of translating religion for particular historically contingent reasons

connected with the projects of Christian missionaries and the proponents of Neo-

Hinduism, 125 the word sana most likely came to be chosen for that purpose in Thai

because of the particular importance that the word has in the Pali scriptures 126 that are

preserved by the Theravda school of Buddhism that is dominant in Thailand. 127

125
Wilhelm Halbfass, Reinterpretations of Dharma in Modern Hinduism, in India and
Europe, 334-348. Interestingly, though, the earliest missionaries chose the words mata and veda to
translate their concept of religion, and the use of the word dharma only began with the Baptist
missionaries in Bengal (p. 340).
126
The word ssana is used throughout the Pali Canon to refer to the Buddhas teaching or
instruction, but even more concretely to the (abstract) institution he set up for training monks. It is
especially common in later literature such as the Apadna, which refers repeatedly to buddhassa
ssana (the Buddhas dispensation). We even find the compound buddhassana a few times
scattered throughout the canon. In Thai, therefore, (phra phutthastsan) is used in
specifically Buddhist contexts to refer to Buddhism, while (san buddh) is used in more
secular contexts to refer to Buddhism, and by extension, we have also (san khris[t])
for Christianity, (san itsalm) for Islam, etc.

127
The problem of translating the concept of religion into local languages was by no means
confined to the Indic cultural sphere. In Tibet, the word chos lugs (dharma tradition) is used in
modern times to refer to religion, while in Mongolia, the word shashin (derived from the Sanskrit
sana, which entered into Mongolian via the Sogdians and Uighurs) is used now for the same purpose.
In East Asia, the Chinese word (Japanese: shky; Mandarin: zngjio) was used by the Japanese
in the 19th century to translate the Western term religion. It combines the Chinese characters for
ancestor or sect and teachingthus, teaching of the ancestors or teachings of a particular
sect, both of which reflect valences of the Western concept of religion. This use of the word was then
borrowed for use in modern Chinese. Individual religions are named by combining a character or
characters appropriate to that particular religion with the character for teachingthus, (fjio,
Buddha-teaching) for Buddhism, (dojio, Dao-teaching) for Daoism, (jdjio,

63
 The specific category of Buddhism or Buddhist, whether subsumed under

a broader category of religion or not, likewise lacks a precise counterpart in the

ancient Indian context. There is, however, in this case a specific, pre-modern

Sanskrit word that can more-or-less unproblematically be translated as Buddhist

namely, the word bauddha, which is simply the vddhi form of the word buddha.

Although the precise origins of the use of this word to refer to a sectarian identity are

not entirely clear, what is clear is that it arose, relatively speaking, quite late. Under

the definition ein Buddhist for the word bauddha, the Groes Petersburg

Wrterbuch of Bhtlingk and Roth lists the following Sanskrit sources: the

Trikaea, the Rjataragi, the Prabodhacandrodaya, the Viu Pura, and

the Vedntasra. 128 Of these, three have named authors Puruottama for the

Trikaea, 129 Kalhaa for the Rjataragi, 130 and Ka Mira for the

Prabodhacandrodaya 131who can all be dated with some reliability to the 11th and

12th centuries CE, and a fourth, the Vedntasra, can most likely be dated even later,

Christ-teaching) for (Protestant) Christianity, etc. See Daniel Overmyer and Joseph Adler, Chinese
Religion: An Overview, in Encyclopedia of Religion, edited by Lindsay Jones, 2nd ed., vol. 3 (Detroit:
Macmillan Reference USA, 2005), 1580.
128
Otto Bhtlingk and Rudolph Roth, Sanskrit-Wrterbuch, vol. 5 (St. Petersburg:
Kaiserlichen Akademie der Wissenschaften, 1858), 128.
129
First half of the 12th century: Ram Shankar Bhattacharya, ed., Trikaea of
Puruottamadeva (Varanasi: Ratna Publications, 1995), 1.
130
Written between 1148 and 1150: Ranjit Sitaram Pandit, Rjataragi (New Delhi: Sahitya
Akademi, 1968[1935]), xiii.
131
Middle of the 11th century: Kamira, The Rise of Wisdom Moon, trans. and ed. Matthew
T. Kapstein, The Clay Sanskrit Library, ed. Sheldon Pollock (New York University Press, 2009), xvii.

64
to the end of the 16th century. 132 The Viu Pura, like all Puras, does not have a

named author and is quite difficult to date. Some have dated it as late as the 9th or

even 11th century CE, but others have dated it much earlier, to the first half of the first

millennium CE. 133

An earlier date for the Viu Pura might find some corroboration in the use

of the word bauddha in certain Buddhist texts, although here the evidence is quite

mixed. The term is found in the early first-millennium 134 Saddharmapuarka Stra

(Lotus Sutra) nine times as a modifier of the word ja (Buddhist knowledge or

the knowledge of the Buddhas) 135 and once as a modifier of the word yna

(Buddha vehicle, i.e., the single vehicle proclaimed by the Lotus Sutra that leads to

full Buddhahood, as opposed to the three separate vehicles of the rvakas,

pratyekabuddhas, and bodhisattvas). 136 In the Mahvastu (2nd c. BCE-4th c. CE 137), it

is found once in the compound bauddhavaineyak, which Jones translates as those

ready to receive the Buddhas teaching, 138 and in the Gaavyha Stra (early first

132
J. R. Ballantyne, trans., The Vedanta-sara (London: The Christian Literature Society for
India, 1898), 22.
133
Ludo Rocher, The Puras, A History of Indian Literature, ed. Jan Gonda, vol. 3
(Wiesbaden: Otto Harrassowitz, 1986), 249.
134
Paul Williams, Mahyna Buddhism: The Doctrinal Foundations, 2nd ed. (London:
Routledge, 2009), 150.
135
Saddh. 2.47, 2.55, 2.99, 4.44, 6.2, 6.31, 7.84, 8.8, 15.2.
136
Saddh. 3.98.
137
J. J. Jones, trans., The Mahvastu (London: Luzac and Company, 1949), xi.
138
Mvu. 1.51: Jones, Mahvastu, 42 n. 3.

65
millennium CE139), it is found three times as an attributive adjective meaning of the

Buddhas. 140 None of these preceding references clearly uses bauddha as a category

of identity. In his Madhyamakahdayam, however, the 6th century philosopher

Bhavya refers twice to vaco bauddham in the context of a critique of Mms (9.16-

17). Although Lindtner translates this phrase simply as word of the Buddha,

Harmut Bscher argues (personal communication, 7 Oct. 2011) that it could instead

be interpreted as a doxographical designation. Although this evidence is quite scanty,

and certainly a comprehensive study of the genealogy of the word bauddha in both

Buddhist and non-Buddhist Sanskrit literature remains a desideratum that is outside

of the scope of this dissertation, it seems likely that the use of bauddha as a category

of sectarian identity emerged quite late, in the middle of the first millennium at the

earliest, perhaps out of a matrix of non-technical uses in the later (turn of the era and

after) Buddhist literature. 141

In any case, what is clear is that bauddha was not used as a Buddhist identity

during the early history of what we today call Buddhism. The Pali equivalent

boddha, indeed, is not found anywhere in the Canon. As discussed above (n. 119),

we quite often find the word ssana, or (especially in the later texts of the Canon) the

139
Williams, Mahyna Buddhism, 133, 331 n. 10.
140
Gv. 32: rutni bauddhni; Gv. 246: bauddho viayo; Gv. 290: bauddh vikurvitavyha. I
thank Vesna Wallace for these references and the previous references to the Lotus Stra and
Mahvastu.
141
That is, it is conceivable that uses of the word bauddha to mean of or pertaining to the
Buddha(s), such as bauddha yna and bauddha ja, could have led naturally to the sense of a
Buddhist vehicle or knowledge, especially with the increase in inter-sectarian polemics in the
medieval period.

66
compound buddhassana or even sammsambuddhassana, used to refer to the

teaching, instruction, and (more abstractly) dispensation of the Buddha. 142 In

addition, sagha is of course the standard word at all points in the history of

Buddhism, from the origins to the present, for the Buddhist community, especially

of monks. 143 It functions in much the same way as church does, in its abstract, but

not concrete, sense, in Christianity. 144

Most important for our purposes, however, is the term used attributively to

describe those whom we today would call Buddhists. Long before bauddha came

into use, this term was kya. In the Pali Canon, a standard phrase used to refer to

Buddhist monks, not simply as such, but as opposed to members of other sects, is

142
Interestingly, the compound buddhadhamma, which we would expect on the basis modern
Indian languages to refer to Buddhism, is found only quite rarely in the Canon, and in none of those
instances can it be interpreted as having such a meaning. Most of the instances of the compound are
found in late texts of the Khuddaka Nikya (Apadna, Buddhavasa, Niddesa, Paisambhidmagga,
Milindapaha, Nettippakaraa), and two are found in the Vinaya. Nearly all of these instances are
either plural or else a singular dvandva (which is trivial for our purposes). The two non-trivial cases of
buddhadhamma in the singular (Mil. 5.3.1 and Bv. Revatabuddhavasa v. 15) both refer to a
characteristic of a Buddha, as do the cases in the plural.
143
While the word sana may have been largely confined to the imaginary of early Buddhist
texts, sagha appears to have been an important colloquial category of Buddhist identity. This is made
evident by the fact that Aoka, lacking a term for Buddhism per se, deals with what we call
Buddhism primarily through the category sagha, which for him appears to refer exclusively to the
order of monks and nuns. This term is found in the Kaumb, Sc and Srnth Pillar Edicts, in
which punishment is ordered for those who foment schism in the sagha (from the Sc version: ye
sagha bh[]khati bhikhu v bhikhuni v odtni dus[n]i sana[dhpay]itu an[v]sasi
v[s]petaviy[e]); in the Rpnth and Bair Rock-Inscriptions, in which Aoka speaks of having
approached (up[e]te/[u]payte) the sagha; and the Calcutta-Bair Rock-Inscription, in which
Aoka speaks of saluting the sagha (sagha abhivde[t]na) and declares his respect for the
sagha as part of the triple gem (budhasi dhamasi saghas ti glave ca prasde ca). See E.
Hultzsch, Inscriptions of Aoka, Corpus Inscriptionum Indicarum, vol. 1 (Tokyo: Meicho-Fuky-kai,
1977), 159-163, 166-168, 171-173.
144
Church is the standard English translation of the Greek (lit., assembly), which,
in its earliest, Biblical uses, referred in an abstract sense to the Christian community or communities,
and only later came to refer to a type of building.

67
sama sakyaputtiy. 145 A sort of institution of this designation is given in the

Aggaa Sutta, in which the Buddha says, Vseha, coming from various births,

from various names, from various clans, from various families, you have gone forth

from home into homelessness. When asked, Who are you? answer, We are Sakyan

ascetics. 146 In spite of this illustrious story of institution by the Buddha himself, the

phrase sama sakyaputtiy is not common in the early stra literature; in the Pali

Canon, it is found in fewer than a dozen places in the Sutta Piaka. 147 For the most

part, the stras seem to be content to refer to Buddhist monks using the generic term

bhiku (Pali: bhikkhu). 148 The situation is completely different in the Vinaya,

however, where there is an explosion in the use of the phrase. This is to a certain

145
This expression may be a precursor to the expression kyabhiku that has been studied by
SchopenGregory Schopen, Mahyna in Indian Inscriptions, in Figments and Fragments of
Mahyna Buddhism in India (Honolulu: The University of Hawaii Press, 2005[1979]), 223-244. In
this article, Schopen argues that the term kyabhiku (for monks) and the term paramopsaka (for
laypeople) are closely associated in the epigraphic record with the Mahyna. Schopen notes in an
appendix to the republication of the article in Figments and Fragments (p. 244-6), however, that new
epigraphic evidence necessitates a revisitation of the conclusions he comes to in this article. Given
that, at least according to Schopens preliminary 1979 research, the term kyabhiku is associated
with the Mahyna and in any case does not appear in the epigraphic record until the 4th century CE, I
consider it to be outside of the scope of this work.
146
tumhe khvattha, vseha, nnjacc nnnm nngott nnkul agrasm
anagriya pabbajit. ke tumheti puh samn sama sakyaputtiymh ti paijntha.
147
DN 27, DN 29, MN 86, SN 3.7.10, SN 4.8.10, AN 5.209, AN 8.19, Ud. 4.8, Ud. 5.5.
148
By generic, I mean that the term bhiku (lit., beggar) does not explicitly mark the
person to whom it refers as Buddhist as opposed to a mendicant of another sect. Although it is
certainly true that bhiku was the standard term for a Buddhist monk (and bhiku for a Buddhist nun),
and that this was more or less unique to the Buddhists, it is certainly not the case that the Buddhists
owned the term. We find, for example, the word bhiku used in the Gautama Dharmastra to refer
to one of the four ramasnamely, what is called by Baudhyana and Vasiha, and now in
standardized modern treatments, as the parivrjaka (GDS 3.2, BDS 2.11.12, VDS 7.1). (pastamba
refers to this rama as mauna [DS 2.21.1].) We cannot explain this in terms of any sort of Buddhist
bias, since Gautama is not only a Brahmanical text, but does not recognize the validity of any rama
other than that of the householder (ghastha): ekramya tv cry pratyakavidhnd
grhasthyasya (GDS 3.36).

68
extent understandable given the nature of the Vinaya texts; one of the primary

concerns of the Vinaya is distinguishing the Buddhist sagha from other groups, in

the eyes of the laity who serve as potential donors, by promoting rigorous adherence

to distinctive rules of decorum for monks and nuns. This concern, although certainly

particularly prominent in the Vinaya, is hardly foreign to the stras, however, and

seems insufficient to explain the relative paucity of references to sama

sakyaputtiy (simply put, Buddhists) in the latter. Since the Vinaya, in its fully

compiled form, is generally regarded as being somewhat later than the stras, it

seems quite likely that the designation arose only after some time, when it was no

longer felt sufficient to refer to monks using the generic designation bhiku, and the

need arose to distinguish Buddhist monks from monks of other sects. 149

In Part II, I will explore the network of discursive categories relating to

identity out of which the phrase sama sakyaputtiy was able to arise. Although this

149
Although an extended history of the use of the word kya to refer to Buddhists or
Buddhist monks is beyond the scope of this dissertation, it is worth mentioning that that history
appears to have been quite long. Hultzsch, in his translation of the Aokan edicts, reads Aoka as
referring to himself as a kya in the Rpnth and Maski Rock-Inscriptions, although these readings
are somewhat uncertain, and parallel versions in the Sahasrm, Bair, and iddpura Rock-
Inscriptions have Aoka referring to himself as an upsaka (Hultzsch, Inscriptions of Aoka, 167 n. 18).
Clear references to Buddhist monks as kyas are found in the works of the 6th century Varha Mihira,
as well as Kauilyas Arthastra, which McClish dates in its current form to the 1st or 2nd century CE
(Mark McClish, Political Brahmanism and the State: A Compositional History of the Arthastra
[PhD diss., University of Texas at Austin, 2009], 315). In his Bhat Sahit, Varha Mihira says that
an astrologer can deduce that a person coming with a query is thinking of a thief if he should look at a
kya, the commander of the army if he looks at an updhyya, a merchant if he should look at a
nirgrantha, and so forth (51.21). Elsewhere in the same text, he says that the installation of images
should be performed by members of an order devoted to the deity, including kyas for a Buddha-
image (60.19). In the Bhajjtaka, he writes that if Mars is astrologically dominant at the time of birth,
the child is destined to become a kya, while the dominance of other planets signifies that the child
will become a member of various other orders (15.1). The Arthastra prescribes a fine for feeding
kyas, jvikas, or other proscribed groups in a rddha ritual (3.22). (Citations found in Hultzsch,
Inscriptions of Aoka, 167 n. 18 and under the definition ein Buddhistischer Bettelmnch for the
word kya in Bhtlingk and Roth, Wrterbuch, vol. 7, 131.)

69
designations ultimate specificity is contained in the word sakya, the name of the clan

into which the Buddha was born, the structure of the phrase of a whole, rather like in

modern biological taxonomies, reflects a more intricate hierarchy of categories, which,

I will argue, have much to teach us about the history of Buddhist identity-formation.

That is, we have not only a species sakyaputtiya, but also a genus samaa. The

classification of the Buddhist monk under this genus should not simply be taken for

granted, and an elucidation of how it came about is crucial for a study of Buddhist

identity-formation vis vis Brahmans because the Brahman and the Buddhist monk

came to be regarded in the emic system of categories in India not as two species of a

common genus, but as, respectively, an entire genus and a species of a completely

different genus. Therefore, I will begin by examining the way in which the discursive

complex ramaa-brhmaa is treated both in the early Buddhist stras and in other

sources, and I will show that there was a likely an evolution in the complex towards a

notion of antagonism and mutual exclusivity. Then, given the fact that the ramaa-

brhmaa distinction appears to be an element of colloquial discourse that was

adopted by the Buddhists but not by the Brahmanical sources, I will turn to an

examination of what appears to have been the Brahmans alternativenamely, the

rama systemand how it relates to colloquial categories that were shared with

the Buddhists. In the process, I will demonstrate that the ramaa-brhmaa

distinction, and thus the categorization of Buddhists as ramaas and not brhmaas,

was the result of a long process of identity-formation in which a variety of categories

were contested and deployed in various ways by different groups.

70
 Chapter II.1

ramaas and Brhmaas

II.1.1 ramaa and Brhmaa as Mutually Exclusive Categories

The impression that Buddhism arose in opposition to Brahmanism, which, as

discussed in the Introduction, has guided much of the scholarly work on the

relationship between these two religious traditions, is based largely on two related

factors. The first is the classification of the Buddha, and by extension his followers,

as ramaa. The second is a certain category of stras, which, following the

language used by Brian Black, 150 I will refer to as encounter dialogs, that

consistently, and somewhat uniquely, present the Buddha qua ramaa entering into

dialog with an interlocutor who is identified as a Brahman. According to the standard

ordering, the first of these encounter dialogs in the Pali Canon is the Ambaha Sutta

(DN 3), which can be taken as paradigmatic of the genre. In this sutta, immediately

after the nidna (i.e., the introduction of the setting), one of the Buddhas primary

150
Brian Black, Rivals and Benefactors: Encounters Between Buddhists and Brahmins in the
Nikyas, Religions of South Asia 3, no. 1 (2009): 25-43.

71
interlocutors is introduced as the Brahman Pokkarasti (brhmao pokkharasti).

This interlocutor then hears of the fame of the Buddha, which is expressed through a

particular formula that begins with the words, The ascetic Gotama, son of the

Sakyans (samao khalu bho gotamo sakyaputto). Thus, from the very

beginning, the Buddha and his interlocutor are identified using the categories

brhmaa and samaa. Throughout the sutta, Pokkharasti continues to refer to the

Buddha in his thoughts, as well as in speaking to other Brahmans, as samaa gotama,

and in addressing Pokkharasti, the Buddha repeatedly uses the vocative brhmaa.

Encounter dialogs such as the Ambaha Sutta therefore set up a discursive space in

which ramaa and brhmaa are clearly separate categories, with a certain amount

of opposition implied, if not made entirely explicit.

While, as I will argue further below, the uses to which discursive categories

are put in a normative text need not be read as representative of discourse outside of

that text, it is clear that for a certain period of time in ancient India, the categories

ramaa and brhmaa were used in colloquial discourse to refer to groups of people

with at least a degree of mutual exclusivity. In the Thirteenth Rock-Edit at Kl,

Aoka writes, There is no country where these (two) classes [niky], (viz.) the

Brhmaas and the ramaas, do not exist, except among the Ynas. 151 Even earlier

testimony is given by Megasthenes, who was the ambassador of Seleukos Nikator to

151
Holtzsch, Inscriptions of Aoka, 45, 47: n[a]thi c e jan[a]pade yat nathi ime niky
nat Y[o]ne[u] bahmane c[] amane c.

72
the court of Aokas grandfather Candragupta Maurya around 302 BCE. 152

According to Strabo, Megasthenes divided Indian society into seven social groups, 153

of which the first, the philosophers, are divided into two groups: the and

the (probably a corruption of ). 154 Strabo also records the

testimony of Kleitarchos, who accompanied Alexander the Great in his Asian

conquest and wrote a history of it. 155 According to Kleitarchos, The Pramnai

[: again, probably a corruption of ramaa] are philosophers opposed to the

Brachmanes, and are contentious and fond of argument. They ridicule the

Brachmanes who study physiology and astronomy as fools and imposters. 156 The

much later (appx. 3rd century CE) testimony of Bardesanes the Babylonian, so

called by Porphyry, whose record thereof is preserved as a fragment in the writings of

Stobaeus, also speaks of two such groups, and casts their distinction not in terms of

philosophical dispute, but in terms of recruitment:

The Indian Theosophs, whom the Greeks call Gymnosophists, are divided into
two sects, Brahmans and Shamans, Samanaeoi. The Brahmans are one family,
the descendents of one father and mother, and they inherit their theology as a
priesthood. The Shamans, on the other hand, are taken from all Indian sects

152
Patrick Olivelle, The Origin and Early Development of Buddhist Monachism (Colombo:
Gunasena, 1974), 25 n. 1.
153
Strabo, Geography, 15.39; cf. Arrian, Indika, 11.1. Much ink has been spilled over the fact
that Megasthenes speaks of seven castes () instead of the expected four. Bronkhorst has
recently argued that Megasthenes is not aware of the four varas because he visited Magadha at a time
when it had not yet been BrahmanizedBronkhorst, Buddhism in the Shadow of Brahmanism, 73.
154
Strabo, Geography, 15.59; on the form , see John W. McCrindle, Ancient India
as Described in Classical Literature (St. Leonards: Ad Orientem, 1971[1901]), 65 n. 1.
155
McCrindle, Ancient India in Classical Literature, 75 n. 5.
156
Strabo, Geography, 15.70: McCrindle, Ancient India in Classical Literature, 76.

73
 indifferently, from all who wish to give themselves up to the study of divine
things. 157

Of course, the most well-known and often-quoted piece of testimony regarding

ramaas and brhmaas comes from the second century BCE 158 Mahbhya, in

which Patajali uses the word ramaabrhmaa as an example of the rule that two

things which are in opposition (virodha) should be expressed in the singular

(ekavacana) when in a dvandva compound. 159 Taken together, this evidence

indicates that by the time of Candragupta Maurya (late 4th c. BCE) there was in

general colloquial discourse a conception of ramaas and brhmaas as separate,

perhaps even mutually antagonistic groups, and that this binary social division

persisted well after the turn of the era.

157
McCrindle, Ancient India in Classical Literature, 167. Interestingly, while this rather late
source has Bardesanes describing a clear distinction between Brahmans as being restricted by birth and
ramaas being open to people of all social classes, Arrian, based presumably on the testimony of the
much earlier Megasthenes (the ambassador to Candragupta Mauryas court), states without
qualification that to the philosopher alone [of the seven castes] is it permitted to be from any caste
whatever ( ), for no easy life is his, but the hardest of allArrian, Indika, 11.:
McCrindle, Ancient India in Classical Literature, 167 n. 2. This assertion follows a fairly standard
description of the caste system, to which the philosophers are said to be an exception: The custom of
the country prohibits intermarriage between castes:for instance, the husbandman cannot take a wife
from the artisan caste, nor the artisan a wife from the husbandman caste. Custom also prohibits any
one from exercising two trades, or changing from one caste to another. One cannot, for instance,
become a husbandman if he is a herdsman, or become a herdsman if he is an artisanArrian, Indika,
12: John W. McCrindle, Ancient India as Described by Megasthens and Arrian (Calcutta: Thacker,
Spink, and Co., 1877), 212-3. This particular statement in Arrian does not distinguish between the
Brahmans and ramaas, which distinction we know on the basis of Strabo that Megasthenes was
aware of. However, Strabo does not clearly indicate that Megasthenes was aware of Brahmans being
exclusive in membership on the basis of birth, although this is perhaps implied by the statement, From
the time of their conception in the womb they are under the care and guardianship of learned
menStrabo, Geography, 15.59: McCrindle, Ancient India in Classical Literature, 65.
158
Bronkhorst, Greater Magadha, 1.
159
Patajali, Mahbhya, 2.4.9-12.

74
 Although this evidence for a popular conception of Brahmans and ramaas

as separate, mutually antagonistic groups may seem quite early, and in fact can be

found in the some of the earliest non-oral historical Indian sources available to us, it

must be emphasized that even the earliest among them, the Greek sources dating

(ostensibly) to the reign of Candragupta Maurya in the late 4th century BCE, come

from a historical context significantly removed from that in which Buddhism arose.

Even if we accept, following the trend of current scholarly opinion, 160 the latest

possible date for the death of the Buddha, which is given by the so-called short

chronology that places the death of the Buddha at 100 years before the consecration

of Aoka (268 BCE), then the Buddha died in 368 BCE, or over 40 years before the

ascension of Candragupta Maurya in 324 BCE. 161 This may not seem like a very

long time, but given the fact that he is said to have lived 80 years, the Buddha would

have been born and flourished several generations prior to Candragupta, and in

particular would have attained enlightenment about a century before Candraguptas

ascension. In addition, this century was a century of massive changes in North Indian

social and political life, seeing a transition from small, lineage-based proto-states

(janapadas) to the expansive, dynastic empires of the Nandas and Mauryas.

160
The shift in scholarly opinion away from the long chronology preserved in the Theravdin
sources and toward the short chronology preserved in the Sanskrit and Chinese sources crystallized
with a symposium whose papers were published in Heinz Bechert, The Dating of the Historical
Buddha, 2 vols. (Gttingen: Vandenhoeck and Ruprecht, 1991-1992).
161
Lamotte, History of Indian Buddhism, 217.

75
II.1.2 Samaabrhma as a Unified Category Against Which to Construct Buddhist

Identity in the Early Buddhist Texts

Although there is reason to believe that much of what we find in the early

Buddhist stras represents an imperial context fairly removed from the time of the

Buddha, 162 these sources, which were not written all at once, but rather composed and

passed down orally over the course of many centuries, presumably do preserve

elements that date to an earlier time than any of our reliably datable written sources

(i.e., the Greek sources and inscriptions of Aoka cited above). Thus, it is worthwhile

to ask if they fully support the oppositional conception of ramaas and Brahmans

found in those later sources. I will argue that they do not. First, however, we must

acknowledge that the compound samaabrhmaa, which is fairly ubiquitous

throughout the Pali Canon, must generally speaking be treated grammatically as a

dvandva compound. The general rule appears to be that when the two elements,

samaa and brhmaa, are joined conjunctively, they are given in compound, but

when they are joined disjunctively, they are split up and joined by the conjunction v.

In addition, these conjunctive and disjunctive combinations of the two categories

correspond to the plural and the singular, respectively.

An example from the first sutta in the Canon, the Brahmajla Sutta (DN 1),

will suffice to illustrate. This sutta consists primarily of an enumeration of wrong

162
The recurring trope of the cakravartin found in many early Buddhist stras, especially of
the Drgha gama/Dgha Nikya, makes more sense in a context in which there actually are emperors
who control vast swathes of land beyond their native janapada. In addition, Bronkhorst has argued
that the Assalyana Sutta in particular (which, to fit his Greater Magadha hypothesis, must be late
because it narrates an encounter between the Buddha and a Brahman) must date after Alexanders
invasion because it refers to YonasBronkhorst, Greater Magadha, 353.

76
practices and views, which are ascribed to some ascetics and Brahmans (eke

samaabrhma). Generally speaking, these wrong practices and views are

ascribed to other ascetics and Brahmans in precisely this way, that is, conjunctively

and in the plural, but sometimes a more detailed enumeration of views is given, and

then there is a switch to the disjunctive, singular usage. Thus, at a certain point the

Buddha states, There are, monks, some ascetics and Brahmans who are eternalists,

who declare in four ways the self and the world to be eternal (santi, bhikkhave, eke

samaabrhma sassatavd, sassata attnaca lokaca paapenti cathi

vatthhi). In enumerating each of these four ways, however, the Buddha begins

with the stock phrase, Here, monks, a certain ascetic or Brahman (idha,

bhikkhave, ekacco samao v brhmao v). There thus is a rhetorical shift from

thinking about a group of people who share a certain type of wrong view, broadly

construed, to thinking about a particular individual within that group who holds a

particular wrong view within the broader category. Given that, grammatically, the

shift from plural to singular is accompanied by a shift from the conjunctive to the

disjunctive in relating the two categories samaa and brhmaa to one another, we

can see that, at least on a formal level, these two categories are understood to be

separate categories of individuals that can be treated together as an aggregate, but

such that any particular individual can only be a member of one or the other. 163

Although grammatically this seems to support an understanding of the

categories samaa and brhmaa similar to the oppositional, or at least discrete,

163
Otherwise, we would expect to see the compound samaabrhmao in the singular, as a
karmadhraya compound.

77
understandings found in the sources dating from the time of Candragupta and later

discussed above, the actual usage of the compound samaabrhmaa often militates

against such an interpretation. While encounter dialogs, as defined above, define

the Buddha as a samaa and his interlocutor as a brhmaa, the use of the compound

samaabrhmaa is mostly restricted to a different class of suttas, of which the

Brahmajla Sutta just cited is a prime example, in which the Buddha discusses wrong

views and practices. These suttas may or may not consist of a dialog between the

Buddha and an interlocutor, and when they do involve an interlocutor, more often

than not that interlocutor is identified as a paribbjaka, rather than as a Brahman. 164

In these suttas, which I will refer to as teachings on wrong views, the compound

samaabrhmaa is used in the plural to refer to a whole class of individuals against

whom as an aggregate the author of the sutta, using the voice of the Buddha, seeks to

construct a Buddhist identity. This is as true in suttas such as the Brahmajla Sutta,

in which the compound samaabrhma is sometimes split up into the disjunctive

samao v brhmao v for use in the singular, as it is in other suttas in which we

find only the compound. That is, regardless of whether the compound is technically a

dvandva or nominally refers to two separate groups of individuals, the compound is

used rhetorically to refer to a single class of individuals against which the Buddha, as

164
There are some exceptions, such as MN 4, in which the Buddha speaks about various types
of ascetics and Brahmans with the Brahman Jnussoi (a sort of stock Brahman found in many early
Buddhist stras), and MN 150, in which the Buddha speaks on a similar topic with a group of
Brahman householders (brhmaagahapatik) in the Brahman village (brhmana gmo) of
Nagaravinda in Kosala. One should also note that the fact that an interlocutor is identified as a
paribbjaka does not preclude his being a Brahman (about which more below); I merely mean to point
out that the more common way in which an interlocutor is identified in suttas involving discussions of
samaas and brhmaas is as a paribbjaka, rather than as a Brahman.

78
a literary character, seeks to define himself. What matters is not the distinction

between samaas and brhmaas, much less to which the Buddha and his followers

belong, but rather the distinction between the Buddha and his followers, on the one

hand, and the whole lot of samaabrhma, on the other. 165

The extent to which the distinction between samaa and brhmaa often does not

matter in the use of the compound samaabrhmaa in the Pali Canon can be seen in

the Sagrava Sutta (MN 100), in which the Buddha speaks to a Brahman student

(mavo) named Caalakappa who is incredulous about the faith of a certain

Brahman woman (brhma) in such a shaveling ascetic (muaka samaaka). 166

When Caalakappa meets the Buddha, he says,

There are, O Gotama, some samaabrhma who promise the fundamental

brahmacariya having attained accomplishment and perfection of super-
knowledge in this world. So which, O Gotama, of those samaabrhma
who promise the fundamental brahmacariya having attained accomplishment
and perfection of super-knowledge in this world is the Venerable Gotama? 167

Caalakappa is asking the Buddha where he places himself within the broad

category of ascetics and Brahmans who claim (to put it more simply) to have attained

165
In this respect, the compound samaabrhmaa appears to replace the category of samaa
alone as the foil against which to construct a Buddhist identity, as found in the Ahakavagga of the
Sutta Nipta. I will discuss this further in Part IV.
166
This is another exception to the tendency for teachings on wrong views not to involve an
interlocutor who is identified as a Brahman. In many ways, though, this sutta can be categorized as an
encounter dialog that happens to make use of the compound samaabrhmaa. This is particularly
interesting because, as I will demonstrate, the distinction between samaa and brhmaa is practically
inoperative in the use of this compound, even in the context of an encounter with a Brahman.
167
santi kho, bho gotama, eke samaabrhma dihadhammbhivosnapramippatt,
dibrahmacariya paijnanti. tatra, bho gotama, ye te samaabrhma
dihadhammbhivosnapramippatt, dibrahmacariya paijnanti, tesa bhava gotamo
katamo ti?

79
enlightenment. Now, this would be a perfect opportunity for the Buddha to self-

identify as a samaa, even if just cursorily to say, I am a samaa who has actually

attained enlightenment in this very life. But this is not what the Buddha does.

Instead, he begins by dividing ascetics and Brahmans into three groups whose

distinction practically ignores the distinction between ascetics and Brahmans. The

first group are traditionalists. They, having attained accomplishment and perfection

of super-knowledge in this world by means of tradition, promise the fundamental

brahmacariyajust like the tevijja Brahmans. 168 The second group are those who,

having attained accomplishment and perfection of super-knowledge in this world

only by mere faith, promise the fundamental brahmacariyajust like the logicians

and investigators. 169 The third group are those who, having attained

accomplishment and perfection of super-knowledge in this world by the super-

knowledge, completely by themselves, of a dhamma among dhammas that were not

passed down traditionally in the past, promise the fundamental brahmacariya. 170

The Buddha then identifies himself, as we would expect, as a member of the third

group, as someone who has actually attained enlightenment through super-knowledge,

rather than one who teaches about it merely on the basis of faith or oral tradition.

168
santi, bhradvja, eke samaabrhma anussavik. te anussavena
dihadhammbhivosnapramippatt, dibrahmacariya paijnanti; seyyathpi brhma
tevijj.
169
santi pana, bhradvja, eke samaabrhma kevala saddhmattakena
dihadhammbhivosnapramippatt, dibrahmacariya paijnanti; seyyathpi takk vmas.
170
santi, bhradvja, eke samaabrhma pubbe ananussutesu dhammesu smayeva
dhamma abhiya dihadhammbhivosnapramippatt, dibrahmacariya paijnanti.

80
 What is interesting here is that each group is invariantly referred to as some

ascetics and Brahmans (eke samaabrhma). This includes the first group, which

we find out at the end of the description, actually refers to Brahmans of the Triple

Veda (brhma tevijj). 171 In other words, even Brahmans as such are referred to

as some ascetics and Brahmans. But even more interestingly, the Buddha himself

fails to identify himself as either a samaa or a brhmaa; instead, he says, Now,

Bhradvja, I am among those samaabrhma who, having attained

accomplishment and perfection of super-knowledge in this world by the super-

knowledge, completely by themselves, of a dhamma among dhammas that were not

passed down traditionally in the past, promise the fundamental brahmacariya 172

upon which he recounts the story of how he left home and attained enlightenment. 173

171
This expression most likely refers to Brahmans who are identified as such in encounter
dialogs. The expression tevijj brhma is, in fact, used repeatedly in one such encounter dialog,
the aptly named Tevijja Sutta (DN 13). In this sutta, the Buddha encounters two Brahmans, Vseha
and Bhradvja, and discusses with them the merits (or rather lack thereof) of the teachings of
Brahmans of the Triple Veda. Although the expression is not used with such regularity in other
encounter dialogs that address the teachings of Brahmans or the definition of Brahmanhood, such
dialogs, including the Sagrava Sutta that we are discussing here, do often begin with a formula that
identifies the Brahman interlocutor as perfected in the Three Vedas (tia vedna prag).
172
tatra, bhradvja, ye te samaabrhma pubbe ananussutesu dhammesu smayeva
dhamma abhiya dihadhammbhivosnapramippatt, dibrahmacariya paijnanti,
teshamasmi.
173
This account of the Bodhisatta leaving home and attaining enlightenment is also found,
with only slight differences, in MN 36 and MN 85. The account begins with the Bodhisatta leaving
home against his parents wishes (they are described as weeping as he cuts off his black hair and puts
on the ochre robe) and going to study first under ra Klma and then under Uddaka Rmaputta.
(This first part of the story is also found in the Ariyapariyesana Sutta, MN 26.) Then, he tries several
different types of physical austerity, such as holding his breath and starving himself, but he ultimately
decides that austerity will get him nowhere, giving four similes to illustrate why it is impossible to
attain enlightenment when the body is weakened in such a way. Finally, he remembers a pleasant
experience he had as a child meditating under a tree in his fathers garden; he decides to replicate that
experience; and ultimately he attains enlightenment (which is expressed here as attainment of the three
knowledges, tevijj) after attaining each of the four rpa-jhnas in sequence.

81
Thus, the Buddha ultimately answers Caalakappas question regarding his place

within those samaas and brhmaas who teach about enlightenment not by

identifying himself, even implicitly, as a samaa, but simply by saying that he is one

of those samaas and brhmaas who have actually attained enlightenment.

A closely related example of how the compound samaabrhma is used in

such a way that the presumable distinction between the two elements of the

compound is effectively ignored is the treatment of what are often called in English

the six heretical teachers. Although often the early stras use ascetics and

Brahmans in the abstract as a foil against which to construct a Buddhist identity,

sometimes these six non-Buddhist teachers appear, together and more-or-less as stock

characters, to personalize the other against which Buddhist identity is constructed.

The most detailed explanation of the doctrines of these six teachers is given in the

Smaaphala Sutta (DN 2), in which King Ajtasattu of Magadha goes to the

Buddha and recounts to him conversations he had had previously with the six

teachers. Of the six, two are easily identifiable from the Jain scriptures: Nigaha

Ntaputta is Mahvra, revered by Jains as the last Trthakara, and Makkhali Gosla

is a student of Mahvra who parted ways with him over the issue of burning off

existing karma and became a leader of the jvikas. 174 Praa Kassapa, who in the

174
A. L. Basham, History and Doctrines of the jvikas: A Vanished Indian Religion (Delhi:
Motilal Banarsidass, 2002[1951]), 17. An excellent, and to my mind definitive, explanation of the
hitherto unclear niyativda ascribed to Gosla Makhaliputta as the point over which he and Mahvra
parted ways can be found in Bronkhorst, Greater Magadha, 38-51. Bronkhorst argues, in short, that
the difference between Jains and the jvikas on this point had to do with past karma: Mahvra and
the Jains taught that non-action had the power to both prevent new karma from accumulating and burn
off old karma, while Gosla and the jvikas held that it could only prevent new karma from

82
Smaaphala Sutta is presented as teaching a doctrine of non-action (i.e., that there

are no karmic results of good or bad actions), also appears to have been associated

with the jvikas. He is said to be chief of five hundred jvikas (pacamtrm

jvikaatnm pramukha) in the Saghabhedavastu of the Mlasarvstivdins, 175

and is mentioned by name in a Tamil text as having been held in high esteem by the

jvikas. 176 Ajita Kesakambal teaches a doctrine of materialism, which like that of

Praa Kassapa denies the efficacy of karma, but explains the non-efficacy of karma

in terms of the breakup of the body (and thus the final end of the human individual)

into its constituent elements at death. 177 Pakudha Kaccyana also appears to deny the

efficacy of karma, but does so on the basis of a theory of atomism; Basham argues

that this theory follows logically from Makkhali [Gosla]s determinism and notes

that it forms part of the teaching of the jvikas as described in Tamil texts. 178

accumulating; old karma must come to fruition on its own over the course of thousands of lifetimes
until it finally burns off.
175
Cited in Bronkhorst, Greater Magadha, 48 n. 86. The teachings attributed to Praa
Kyapa in this text, however, correspond to those of Ajita Kesakambal in the Smaaphala Sutta
although this latter teaching is, as we will see, quite similar.
176
Basham, History and Doctrines of the jvikas, 17, 80-81. The text is the Tamil poem
Nlakci, in which Lord Praa, without comparison in intelligence (praa ep puruvara-k-
kaava: v. 668) is presentented as the head of a monastery of jvika monks at Kukkuanagara at the
time of the Buddha.
177
Basham argues that Ajita, therefore, may be a forerunner of the later CrvkasBasham,
History and Doctrines of the jvikas, 17. Bronkhorst has a much different theory of the Crvkas, one
that identifies them with a non-Vedic group (of which Ajita must be one, since his doctrine explicitly
rejects the efficacy of sacrifice), but with the Brahmans themselves, who he argues largely rejected the
doctrines of karma and rebirthBronkhorst, Greater Magadha, 309-328.
178
Basham, History and Doctrines of the jvikas, 17, 236-239, 262-266. The texts in which
atomism is attributed to the jvikas are the Maimkalai, the Nlakci, and the Civaa-cittiyr. In
addition to atomism, the Nlakci attributes to the jvikas a doctrine, similar to that of Pakudha, that
the universe is static, referred to by the commentator on the text as avicalita-nityatvam.

83
Finally, Sajaya Belahaputta refuses to commit to any position on questions of

controversy, in words that elsewhere (DN 1) are attributed to the eel-wrigglers

(amarvikkhepik). 179

What is important to note about these six heretical teachers is that none of

them appears to be a Brahman in the conventional sensethat is, a person born

into a Brahmanical lineage who promotes the Brahmanical ideology of vara,

sacrifice, and Vedic knowledge, as do the Brahmanical interlocutors found in

encounter dialogs. Indeed, two of them (Nigaha Ntaputta and Makkhali Gosla)

are well-known leaders of the non-Brahmanical, ramaic Jain and jvika sects;

Praa Kassapa appears to have also been a leader of the jvikas; Pakudha

Kaccyanas teaching of atomism appears to be associated with the jvikas; Ajita

Kesakambals teaching of non-action explicitly rejects the efficacy of sacrifice; and

Sajaya Belahaputtas skepticism has no apparent connection to Brahmanical

ideology, or any other ideology, for that matter. What we appear to have here is six

rivals to the Buddha from within the ramaa movement. But this is not how they

are presented in the early Buddhist texts. In the Smaaphala Sutta, King Ajtasattu

describes his previous encounters with the six heretics after he poses a question to the

Buddha and is asked by the latter in return, Do you admit, Great King, that you have

asked other ascetics and Brahmans this question? 180 In four other suttas spread

179
Ironically, the much-reviled evasiveness of the eel-wrigglers is virtually
indistinguishable from the rhetorical tactics taken in certain Buddhist texts, including Ngrjuna and
his Mdhyamika followers, but also the Ahakavagga of the Sutta Nipta, which is likely one of the
earliest Buddhist texts still extant.
180
abhijnsi no tva, mahrja, ima paha ae samaabrhmae pucchit ti?

84
throughout the Canon (DN 16, MN 30, SN 1.3.1.1, Sn. 3.6), 181 the six heretics are

introduced using a stock formula that identifies them, again, not specifically as

samaas, but as samaabrhma:

O Gotama, these samaabrhma have saghas, have gaas, are teachers of

gaas, are known, famous ford-makers, highly respected by many people
namely, Praa Kassapa, Makkhali Gosla, Ajita Kesakambala, Pakudha
Kaccyana, Sacaya Belahaputta, Nigaha Naputta. 182

The six heretics, therefore, represent a clear example of how the compound

samaabrhma can be used in such a way that, not only does the distinction

between samaa and brhmaa not matter, but only one of the two categories (when

understood as distinct) appears to be present at all.

II.1.3 Samaabrhma Used in a Positive Sense

So far we have been looking at uses of the compound samaabrhma in a

negative sense, as a negatively portrayed other against which to construct Buddhist

identity. This, as already noted, is by far the most common way in which the

compound is used in the early Buddhist stras. There is, however, a subset of cases

in which the compound is used in a positive sense, to refer to a class of persons who

181
A passage in the Milindapaha (1.5) uses similar words to introduce the six heretics, but
omits the word samaabrhma: eva vutte pacasat yonak rjna milinda etadavocu: atthi,
mahrja, cha satthro prao kassapo makkhaligoslo nigaho naputto sajayo belahaputto
ajito kesakambalo pakudho kaccyano, te saghino gaino gacariyak t yasassino titthakar
sdhusammat bahujanassa, gaccha tva mahrja, te paha pucchassu, kakha paivinayass ti.
Two suttas (MN 77, SN 4.10.9) also use similar words to introduce the six heretics, but introduce them
individually and therefore omit the collective term samaabrhma. Another two suttas (MN 36, SN
1.2.3.10) refers to the six heretics in ways (including, in the case of the latter, through verse) that break
from the formula entirely.
182
yeme, bho gotama, samaabrhma saghino gaino gacariy t yasassino
titthakar sdhusammat bahujanassa, seyyathida prao kassapo, makkhali goslo, ajito
kesakambalo, pakudho kaccyano, sacayo belahaputto, nigaho naputto.

85
are worthy of honor, praise, and in particular gifts (dna). As we will see, this

positive use of samaabrhma, like the negative, refers to ascetics and Brahmans

as a single class and essentially ignores any distinction between them.

One way in which the compound is used in a positive sense is to refer to the fact (in

the eyes of the authors of the early Buddhist texts) that some people have indeed

attained enlightenment and thus put an end to the cycle of rebirth. Possibly the

earliest example of this is found in one of the earliest Buddhist texts to have come

down to us, the Pryaavagga of the Sutta Nipta. This text consists of a series of

short dialogs between the Buddha and various interlocutors; in one of these dialogs, a

certain Nanda asks the Buddha about ascetics and Brahmans who say purity is by

means of what is seen or heard, say purity is by means of morality and vows, say

purity is by means of various (things). 183 The Buddha responds by criticizing such

samaabrhmase, 184 saying that they have not crossed over birth and old age. 185

If it ended here, this would simply be another example of the use of the compound

samaabrhma(se) as a foil against which to construct Buddhist identity, but the

Buddha qualifies his answer by saying, Not all samaabrhmase do I say are

enveloped by birth and old age; rather, those who give up what is seen and heard and

183
Sn. 1079: ye kecime samaabrhmase, (iccyasm nando) / dihassutenpi vadanti
suddhi / slabbatenpi vadanti suddhi, anekarpena vadanti suddhi.
184
The nominative plural in se is found frequently in the Sutta Nipta; it is an eastern form
(Mgadhism) derived from the Vedic sasK. R. Norman, trans., The Group of Discourses (Sutta
Nipta), vol. 2 (Oxford: Pali Text Society, 1992), 134.
185
nrisu jtijaran ti. I follow Norman in emending nrisu to ntrisuNorman,
Group of Discourses, vol. 2, 377n.1080.

86
the like are without savas and have indeed crossed the flood. 186 Now, presumably

these good ascetics and Brahmans are the Buddha and his disciples, but regardless,

they are still referred to as samaabrhmase, as if that were the most basic

category by which to identify them.

Elsewhere in the Canon, 187 this conceptnamely, that some ascetics and

Brahmans have indeed escaped from sasrais generalized and made a part of

right view (sammdihi). This is expressed through the following formula:

There exists what is given; there exists what is offered; there exists what is
sacrificed; there exists the fruit, the result of meritorious and demeritorious
actions; there exists this world; there exists the other world; there exists
mother; there exists father; there exist spontaneously arisen beings; there exist
in the world samaabrhma, rightly gone, gone on the right path, who,
having realized by themselves through super-knowledge, declare this world
and the other world. 188

Wrong view (micchdihi), on the other hand, is defined as the precise converse:

There does not exist what is given; there does not exist what is offered; there
does not exist what is sacrificed; there does not exist the fruit, the result of
meritorious and demeritorious actions; there does not exist this world; there
does not exist the other world; there does exist mother; there does not exist
father; there do not exist spontaneously arisen beings; there do not exist in the
world samaabrhma, rightly gone, gone on the right path, who, having
realized by themselves through super-knowledge, declare this world and the
other world. 189

186
Sn. 1082: nha sabbe samaabrhmase, (nandti bhagav) / jtijarya nivutti brmi
/ ye sdha dihava suta muta v, slabbata vpi pahya sabba / anekarpampi pahya
sabba, taha pariya ansavse / te ve nar oghatiti brmi.
187
MN 41, 76, 110, 114, 117; SN 3.3.1.5, 4.8.13; AN 8.1.3.9, 10.4.2.10, 10.5.1.1.
188
atthi dinna atthi yiha atthi huta, atthi sukatadukkana kammna phala vipko,
atthi aya loko atthi paro loko, atthi mt atthi pit, atthi satt opaptik, atthi loke samaabrhma
sammaggat sammpaipann ye imaca loka paraca loka saya abhi sacchikatv pavedent
ti.
189
natthi dinna natthi yiha natthi huta, natthi sukatadukkana kammna phala
vipko, natthi aya loko natthi paro loko, natthi mt natthi pit, natthi satt opaptik, natthi loke

87
Again, presumably those ascetics and Brahmans who have realized by themselves

through super-knowledge are the Buddhists themselves, although the phrasing of

this formula leaves this ambiguous. In any case, those who have attained the highest

goal, whoever they are, are identified once again by the category samaabrhma,

and recognition of their attainment, in terms of this category and without regard to

any putative distinction between samaas and brhmaas, is written into the

definition of right view itself.

Closely related to this formulation of right view is a theme, found several

times in the Dgha Nikya (26, 30, 31), the Aguttara Nikya (3.1.4.6, 5.1.5.2, 5.5.3.7,

5.5.3.8, 8.1.4.8, 8.1.5.6, 8.1.5.8), and also in a Jtaka (457), that situates

samaabrhma within the context of lay dharma. These texts generally list

support (i.e., dna) for samaabrhma together with other types of support a

layperson would be expected to providesuch as to ones parents, children, slaves,

and other dependents. The Jtaka text puts this quite bluntly: Those by whom their

parents and samaabrhma are well-honored (susammnit) go to heaven, while

those who do not so honor them go to hell. Again, no attempt is made to distinguish

between samaas and brhmaas, much less to situate a particular worthy group of

religious specialists (i.e., Buddhist monks) within this broad category; rather, the

entire category is presented as worthy of respect.

samaabrhma sammaggat sammpaipann ye imaca loka paraca loka saya abhi

sacchikatv pavedent ti.

88
 This trope of encouraging support for ascetics and Brahmans as a group is

also taken up by Aoka. Repeatedly throughout his inscriptions, Aoka defines

dhamma in terms of generosity toward ascetics and Brahmans, along with other

dependent groups such as ones parents or the aged. In some inscriptions, Aoka

notes that [i]n times past, for many hundreds of years, there had ever been promoted

the killing of animals and the hurting of living beings, discourtesy to relatives, (and)

discourtesy to ramaas and Brhmaas, but that now, in consequence of the

practice of morality on the part of king Dvnpriya Priyadarin, the sound of

drums has become the sound of morality. 190 Elsewhere, Aoka defines this dhamma

directly, writing, Herein the following (are) comprised, (viz.) proper courtesy to

slaves and servants, reverence to elders, gentleness to animals, (and) liberality to

ramaas and Brhmaas; these and other such (virtues) are called the practice of

morality. 191 Aoka instructs his officials to go throughout his dominion and teach

the people about dhamma in similar terms that include giving to ascetics and

Brahmans, 192 and he even presents himself as an example by visiting ramaas and

190
Fourth Kls Rock-Edict, Fourth Jaugada and Dhauli Rock-Edicts. From the Kls version,
which is the most complete: atika[]ta a[]ta[la] bahuni vasa-satni v[adh]it[e] v
p[n]labhe vi[h]is c bhutna ntin asa[pa]ip[a]ti samana-b[a]bhanna asapaipati.
s[e] aj devnapiyas piyadasine ljine dham[a]-cal[an]en bheli-ghose aho dhama-ghose
(Hultzsch, Inscriptions of Aoka, 30; trans. by Hultzsch on p. 31).
191
Ninth Kls Rock-Edict, Eleventh Kls Rock-Edict, Ninth Dhauli and Jaugada Rock-
Edicts. The translation is taken from the Kls version: he[t] iya dsa-bhaakasi s[a]mypaip[a]ti
gulun apaciti [p][n]n[a] sayame s[a]man[a]-babhanna dne ese ane c heise
dhama-magale nm (Hultzsch, Inscriptions of Aoka, 38; trans. by Hultzsch on p. 38).
192
Third Kls Rock-Edict, Third Dhauli and Jaugada Rock-Edicts. From the Kls version:
sdhu mta-pitisu susus mita-sathuta-ntikyn[a] c babhana-sama[n]na [c] sdhu d[]ne
pnna anlabh[e] sdhu [a]pa-v[i]yt [a]pa-[bha][a]t[] sdhu (Hultzsch, Inscriptions of
Aoka, 29).

89
Brhmaas and making gifts to them in his dhamma-tours (dhama-yt). 193

Now, as we have already seen, Aoka certainly had a conception of ramaas and

brhmaas as two separate groups (nikya), 194 and he was also certainly aware of

individual groups that could be classified under ramaabrhma, including

Brahmans. 195 Indeed, this is to be expected given the evidence we have seen for a

conception of separate categories of ramaa and brhmaa as early as the reign of

Candragupta. Moreover, Bronkhorst has recently argued convincingly that Aokas

policies, in spite of the ecumenical tone of his edicts, were very much hostile to the

interests of Brahmans (i.e., Vedic ritualists). 196 Nevertheless, we still find in the

Aokan inscriptions, much more prevalently than any rhetoric of sectarian division,

repeated reference to ramaabrhma as a single class of persons who are worthy

of respect and donations, just as we find in the Aguttara Nikya. This shows that the

tendency to treat ramaabrhma rhetorically as a single category remained even

193
Eighth Kls Rock-Edict, Eighth Jaugada and Dhauli Rock-Edicts. From the Kls version:
[h]et iya hoti samana-babhanna dasane c dne ca vudh[]na dasa[n]e c[a]
hilana[pai[v]idhne c [j]napadas [ja]n[a]s das[a]ne dhamanusathi c dhama-palipuch c
tatopa[y] (Hultzsch, Inscriptions of Aoka, 36-7; trans. by Hultzsch on p. 37). In his Seventh Delhi-
Tpr Pillar-Edict, Aoka boasts that through imitation of him, his people have been made to progress
and will (be made to) progress in obedience to mother and father, in obedience to elders, in courtesy to
the aged, in courtesy to Brhmaas and ramaas, to the poor and distressed, (and) even to slaves and
servants: tena vahit ca vahisati ca mt-pit[i]su sususy gulusu sususy vayo-mahlakna
anupapatiy bbhana-samanesu kapana-valkesu va dsa-bhaakesu sapapatiy (Hultzsch,
Inscriptions of Aoka, 133; trans. by Hultzsch on p. 136).
194
See p. 58 n. 144 above.
195
The Seventh Delhi-Tpr Pillar edict refers to an order for mahmtras to look over the
affairs of sagha (referring to the Buddhists), the Brahmans, the jvikas, and the Nigahas (Jains):
saghahasi pi me kae ime viypa hohati ti hemeva bbhanesu [j]vikesu pi me kae ime viypa
hohati ti nagahesu pi me kae ime viypa hohati nn-psaesu pi me [ka]e ime viypa
hohati ti paivisiha pavisiha tesu tesu [te te mah]mt (Hultzsch, Inscriptions of Aoka, 132).
196
Bronkhorst, Buddhism in the Shadow of Brahmanism, 12-17.

90
after the two sub-categories are known to have been recognized as distinct and

perhaps even antagonistic.

The evidence presented so far shows that, in spite of an apparent grammatical

treatment of ramaabrhma as a dvandva, which would imply a conjunction of

two distinct categories, the compound is treated rhetorically throughout the early

Buddhist texts and even the inscriptions of Aoka as if it were a single category. This

is sometimes the case even when the compound is not present and the two elements

ramaa and brhmaa are given separately as distinct words. A recurrent formula

found throughout the Sayutta Nikya effectively equates the two categories in the

following terms:

And, monks, whatever samaas or brhmaas understand [here is inserted

whatever doctrinal elements are taken as the subject of the sutta], monks,
those sama or brhma are regarded by me as samaas among samaas
and brhmaas among brhmaas, and moreover those venerable ones dwell
having realized and attained in this world for themselves by super-knowledge
the advantage of being a samaa and the advantage of being a brhmaa.. 197

Within the context of this formula, the particular terminology one uses to refer to the

religious lifewhether it be samaa or brhmaais not important; in either case,

the goal is reached by understanding a particular point of Buddhist doctrine.

197
This formula is found at SN 2.1.2.3, 2.1.2.4, 2.1.3.9, 2.1.8.1, 2.3.4.8, 2.3.4.10, 2.6.3.5,
3.1.5.8, 3.2.1.5, 4.2.3.7, 5.4.1.6, 5.4.1.7, 5.4.3.9, 5.4.3.10, 5.4.4.4, 5.4.4.5, 5.12.3.2: ye ca kho keci,
bhikkhave, sama v brhma vpajnanti, , te kho me, bhikkhave, sama v brhma v
samaesu ceva samaasammat brhmaesu ca brhmaasammat; te ca panyasmanto
smaatthaca brahmaatthaca diheva dhamme saya abhi sacchikatv upasampajja
viharant ti. This usually follows a negative formula: ye hi keci, bhikkhave, sama v brhma
vnappajnanti, , na me te, bhikkhave, sama v brhma v samaesu v samaasammat
brhmaesu v brhmaasammat; na ca pana te yasmanto smaattha v brahmaattha v
diheva dhamme saya abhi sacchikatv upasampajja viharanti. The same formula, including
both the negative and the positive versions, is also found at Iti. 4.4.

91
 A similar theme is found in the Mahshanda Sutta of the Dgha Nikya (8),

in which the Buddha encounters a naked ascetic (acelo) named Kassapa and shows

him the worthlessness of the practice of going without clothes. As a refrain in his

sermon to Kassapa, the Buddha says the following three times:

When, Kassapa, a monk cultivates non-hatred, non-malevolence, a mind of

loving-kindness, he dwells, without savas due to the wasting away of the
savas, having realized and attained for himself in this world by super-
knowledge the liberation of mind, the liberation by wisdom. This monk,
Kassapa, is called both a samaa and a brhmaa. 198

By following this quintessentially Buddhist practice, therefore, one is called both a

samaa and a brhmaa. This practice is contrasted with that of the naked ascetic,

which is so easy that it would have been possible for a householder or a

householders soneven a slave-girl who brings waterto do this, 199 and it is for

this reason that this is a common sayingin the world: Being a samaa is hard;

being a Brahman is hard. 200 Likewise, in an udna (4.8), the people of Svatthi,

having been tricked by some jealous non-Buddhist wanderers into thinking that

Buddhist monks had had sex with and then murdered a woman, disparage the sama

sakyaputtiy in the following terms:

These sama sakyaputtiy are shameless, devoid of morality, of sinful

character, speakers of falsehood, not brahmacrins. They may claim to be
practicers of dhamma, practicers of peace, brahmacrins, speakers of truth,

198
yato kho, kassapa, bhikkhu avera abypajja mettacitta bhveti, savnaca khay
ansava cetovimutti pavimutti diheva dhamme saya abhi sacchikatv upasampajja
viharati. aya vuccati, kassapa, bhikkhu samao itipi brhmao itipi.
199
sakk ca paneta abhavissa ktu gahapatin v gahapatiputtena v antamaso
kumbhadsiypi.
200
pakati kho eslokasmi dukkara smaa dukkara brahmaan ti.

92
 endowed with morality, of good character, (but) they do not possess the
quality of being a samaa; they do not possess the quality of being a Brahman.
Perished is their quality of being a samaa; perished is their quality of being a
Brahman. Whence do they have the quality of being a samaa? Whence do
they have the quality of being a Brahman? These are those whose quality of
being a samaa is departed; these are those whose quality of being a Brahman
is departed. For how indeed could a man do the duty of a man and deprive a
woman of life? 201

Thus, we see, not only put into the Buddhas mouth as a pro-Buddhist polemic, but

also put into the mouths of ordinary laypeople as a colloquial expression, a rhetoric of

samaa and brhmaa that, while treating the two categories separately, treats that as

effectively synonymous, as two equivalent ways of referring to the ideal religious

practitioner.

II.1.4 Conclusion

To summarize, then, there is evidence that the categories ramaa and

brhmaa were not always conceived of as discrete and oppositional. There is, to be

sure, ample evidence that they were conceived as such from the time of Candragupta

onwards, based on Greek testimony, the inscriptions of Aoka, and of course

Patajalis oft-quoted commentary on Pinis rule about virodha-dvandva

compounds. In addition, the categories samaa and brhmaa are treated as separate

and opposed to one another in a certain category of suttas in the Pali Canon that I call

encounter dialogs, and in general the compound samaabrhma appears to be

201
alajjino ime sama sakyaputtiy dussl ppadhamm musvdino abrahmacrino. ime
hi nma dhammacrino samacrino brahmacrino saccavdino slavanto kalyadhamm
paijnissanti. natthi imesa smaa, natthi imesa brahmaa. naha imesa smaa,
naha imesa brahmaa. kuto imesa smaa, kuto imesa brahmaa? apagat ime
sma, apagat ime brahma. kathahi nma puriso purisakicca karitv itthi jvit
voropessat ti.

93
treated formally as a dvandva. Nevertheless, aside from in the encounter dialogs,

samaabrhma appears to be treated rhetorically as a single category, with the

distinction between samaa and brhmaa either passively ignored or even

completely obviated. In suttas that I refer to as discourses on wrong views,

samaabrhma is treated as a unified foil against which to construct Buddhist

identity. When the category samaabrhma qua people with wrong views is

subdivided, it is done so not on the basis of the samaa-brhmaa distinction, but on

the basis of completely different criteria pertaining to particular wrong views or

practices. Subcategories of the samaabrhma broadly construed are still referred

to as samaabrhma, even when the subcategory appears to fall specifically under

either the category of samaa or the category of brhmaa, according to the more

classical understanding. Sometimes the compound samaabrhma is used in a

positive sense, to refer to those who are worthy of honor and donations, and here

again it is treated as a single category, with little or no regard to the samaa-

brhmaa distinction. This is true even when it appears that Buddhists specifically

are being referred to, or when, as in the case of Aoka, a general hostility toward

Brahmans can be assumed.

Taken together, this evidence demonstrates that the often-quoted line from

Patajali that gives ramaabrhmaa as an example of Pinis oppositional

compound cannot be taken as representative of the full range of uses of that

compound, especially in times earlier than the second century BCE when Patajali

wrote, but perhaps to a certain extent even in Patajalis own day. In fact, none of the

94
examples we have looked at here even follow Pinis rule that Patajali was

commenting on. It calls for a virodha-dvandva compound to be written in the

singular 202and as we have seen, samaabrhma is consistently used in the plural.

This makes it problematic to argue that ramaas and brhmaas were, from the very

beginning, fundamentally separate, and raises the possibility that the idea that they

were rigidly separate and even opposed to one another only arose over timethat the

distinction was purposefully constructed.

This possibility is underscored by the fact that, as shown by Olivelle, the

oldest extant uses of the word ramaa are found in Brahmanical texts. The earliest

of these is the Taittirya rayaka, which refers to so-called vtaraana is as

ramaas and celibates (rdhvamanthina). 203 As Olivelle argues, this use of the

word ramaa probably does not refer to a category of persons, but rather is an

adjective (striving) that describes the vtaraana is. 204 The word ramaa is,

however, found in the Bhadrayaka Upaniad as referring to a class of persons. 205

The context of this passage is somewhat ambiguous: Since it is listed together with

both mother and father on the one hand and abortionists and Clas on the other, it

is difficult to say whether the ramaa is conceived of as good or bad, or, to put

202
Pini, Adhyy, 2.4.9.
203
T 2.7: vtaraan ha v aya rama rdhvamanthino. Cited in Olivelle, rama
System, 12 n. 20.
204
Olivelle, rama System, 14.
205
BU 4.3.22; cited in Olivelle, rama System, 14.

95
it differently, as Brahmanical or non-Brahmanical. 206 The use of the derivative

rmaaka in the Dharma Stras is equally ambiguous. It is used in three stras

(BDh 2.11.15, GDh 3.27, VDh 9.10) in the description of the rama of the forest

hermit to refer to, according to Olivelles interpretation, a special procedure for

establishing the sacred fire. 207 Two of these stras, Baudhyana (2.11.27) and

Gautama (3.36), reject all of the ramas except for the householder as unsupported

by the Vedaswhich would imply that they did not accept the rmaaka procedure

as Brahmanical. Vasiha, however, appears to accept all fourwhich would

imply that he did accept it as such.

So what then are we to make of Brahmanical references to ramaas? But

this question immediately implies another: What are we to make of the relative lack

of Brahmanical references to ramaas? Part of what makes the references we have

in Brahmanical texts to ramaas so ambiguous is that there are so few of them.

Brahmanical texts do not appear to have generally adopted the colloquial categories

of ramaa and brhmaa, and they did not construct Brahmanical identity against a

ramaic other. While a certain strain of the Buddhist tradition, represented by the

encounter dialogs, constructed Buddhist identity against a specifically Brahmanical

otherand again I emphasize that this represents only a small part of the suttas in the

Pali Canon, many others using samaabrhma for the same purposethe

Brahmanical tradition dealt with otherness much more indirectly, and with a different

206
atra pitpit bhavati mtmt lok alok dev adev ved aved. atra stenosteno
bhavati bhrahbhrah clocla paulkasopaulkasa ramaoramaas tpasotpasa.
207
Olivelle, rama System, 15 n. 31.

96
system of categories: the rama system. It is to this uniquely Brahmanical system of

taxonomizing religious praxis that we now turn.

97
 Chapter II.2

The rama System and Its Relationship to Brahmanism

II.2.1 Olivelles Characterization of the rama System as a Brahmanical

Theological Construct

As already mentioned in the Introduction, Patrick Olivelles work on the

rama system has provided an important contribution to our understanding of how

Brahmanical orthodoxy formed only slowly over time over the course of several

centuries around the turn of the era. In his seminal work on the subject, Olivelles

primary thesis is that the classical formulation of the rama system, as found

beginning in the Mnava Dharmastra, in which the ramas are a series of life

stages to be entered into in order over the course of a single life, was not the

original form that the system took. Instead, the rama system was originally, as

described in the four early Dharma Stras, a system of four choices of lifelong

vocations one could make after completing ones Vedic studentship. 208 This shows

that one of the central pillars of classical Hindu orthodoxy not only was not of great

208
Olivelle, The rama System, 4.

98
antiquity (since it is not found in any literature prior to the Dharma Stras), but also

changed quite a bit before taking its classical form.

There are other aspects of Olivelles argument, however, that are more

immediately relevant to the issue at hand, namely, the role that taxonomical

categories play in the formation of religious identity. Olivelle defines the rama

system as a theological construct 209 and argues that as such it should not be

confused with any social institutions. It is therefore improper to claim as many

scholars do, for example, that the first two ramas are known from early vedic times;

it is not these ramas but the institutions of vedic studentship and marriage that are

so known. 210 This important distinction allows Olivelle to dispense with debates

over the pre-history of the rama system that often do more to obscure than to

illuminate, and instead focus on the history of the very idea of an rama system.

Olivelles conclusion on this point is that the rama system was invented by

Brahmins living in urban centers [who] were influenced by and open to new ideas

and institutions than their village counterparts. 211 It was, in other words, an

invention of liberal Brahmans who were open to new ways of life and sought to

legitimate them within a Vedic framework. The rama system was then quoted,

209
Personally, I would prefer the term theoretical construct, which emphasizes the
ideational character of the system and avoids the unnecessary introduction of a Western religious
category. In any case, Olivelles point, which is well taken, is that the rama system qua system
should be understood as a normative description used by certain people beginning at a certain point in
history, and as such it should not be confused with any actual social institutions it purports to describe,
some of which may be older.
210
Olivelle, The rama System, 29.
211
Ibid., 98.

99
often disapprovingly, as the opinion of some by the more conservative authors of

the Dharma Stras. This explanation is in contradistinction with the theory advanced

by many earlier scholars, which was that conservative Brahmans invented the rama

system with the intention of resisting the new religious movements and of

safeguarding the Brhmaical religion by incorporating the renunciatory life style

into a scheme that would lessen its impact and reduce or eliminate the conflict

between it and the life of the householder. 212 This theory, Olivelle argues

convincingly, is based on the erroneous use of the classical formulation of the

rama system (which does seem to have been an orthodox attempt to incorporate

asceticism safely into a Brahmanical framework) to speculate on the origins of the

rama system. 213

Although Olivelle rejects the theory that the rama system was created by

conservative Brahmans, such as the ones who describe it in the Dharma Stras, 214 he

still places its creation firmly in the Brahmanical tradition. As he writes, The

authors of the rama system were without doubt Brahmins. 215 He concedes that the

fact that the authors of the Dharma Stras in which the rama system is first

212
Ibid., 94.
213
Ibid., 94-98.
214
Olivelle argues (ibid., 91-94) that pastamba and Vasiha accepted the rama system,
but he believes that they do not represent the liberal Brahmans against whom Baudhyana and
Gautama argued; rather, they probably lived during a time when the system had gained widespread
acceptance within the mainstream (ibid., 97). As we will see, none of the four Dharma Stras are
particularly enthusiastic about the rama system; pastamba, in particular, does not so much accept
it as fail to reject it outright as Gautama and Baudhyana do.
215
Ibid., 97.

100
presented are Brahmans may color the description. At the same time, however, he

argues that the viewpoint of the opponents who are cited by the Dharma Stra authors

shows all the marks of the Brhmaical way of thinking and arguing. 216

The most convincing evidence, I would argue, that Olivelle presents for a

Brahmanical identity for the authors of the rama system is the fact that

Baudhyana quotes two verses, both from the Taittirya Sahit, in support of the

rama system, and then refutes them with his own appeals to Vedic authority. The

first of these pro-rama citations appears to be the basis for the fourfold schema

itself:

Four paths leading to the gods cross between heaven and earth. Of them,
entrust us, all (you) gods, to that which brings us to the state of non-injury.
(TS 5.7.2.3, BDhS 2.11.11) 217

Baudhyana refutes this by saying that for lack of a Vedic text (adatvt) in

support of another interpretation, Four paths refers to rites of ii sacrifices,

animal sacrifices, and soma sacrifices, and ladle-oblations. 218 He then introduces

another verse in support of the ramas using the words Now this is cited (tad

ebhy ancyate):

216
Ibid., 98.
217
ye catvra pathayo devayn antar dyvpthiv viyanti / te yo ajynimajtimvaht
tasmai no dev pari datteha sarva iti.
218
BDhS 2.11.9-10, 29: ye catvra iti / karmavda aiikapukasaumikadrvihomm. The
fact that this statement is made twice is likely due to a transposition from verse 29 to verses 9-10:
Patrick Olivelle, trans., Dharmastras: The Law Codes of pastamba, Gautama, Baudhyana, and
Vasiha (Oxford: Oxford University Press, 1999), 387.

101
 This is the eternal greatness of a Brahman: he is not increased by karma, nor
is he decreased. Having known the realization of its own footprint, the self is
not smeared by sinful karma. (TB 2.12.7-8, BDhS 2.11.30) 219

Baudhyana immediately refutes this argument by presenting two other verses, one

from the Taittirya Sahit and the other from the g Veda, that emphasize the

importance of the Veda and the Vedic ritual. 220

Now, this argument does, to my mind, provide fairly convincing evidence that

Baudhyana, at least, was responding in part to a Brahmanical argument in favor of

the validity of the four ramas. I think it is worthwhile, however, to pause for a

moment and consider from a methodological standpoint what exactly this means.

Olivelle himself has argued, in a critique of Bronkhorsts work, that a pernicious

effect of thinking about Indian history in terms of Vedic and non-Vedic is that

these are often unconsciously assumed to be reified and static entities. 221 A similar

219
ea nityo mahim brhmaasya na karma vardhate no kanyn / tasyaivtm
padavitta viditv na karma lipyate ppakeneti. Interestingly, this very same verse is also quoted at
BDhS 2.17.7-8, but this time straightforwardly in support of (a classical form of) the rama system.
This portion of Baudhyana is part of what Olivelle calls Deutero-Baudhyanaa substantial
addition to an older Baudhyana text that, unlike the latter, accepted the rama system and even
provided detailed rules for renunciation. See Olivelle, The rama System, 86-87, and Olivelle,
Dharmastras, 127.
220
Olivelle presents some other arguments for his contention that the opponents against which
the Dharma Stra authors were arguing were Brahmans, but I find them less convincing. For example,
he writes that [t]he theory of option (vikalpa)comes from the Mms tradition (The rama
System, 98). In other words, he takes the reference to the ramas as options to be a technical term
from the Mms tradition (found, incidentally, only in Gautama)but it is not clear that there was a
well-established Mms tradition at the time the Dharma Stras were written; or that vikalpa
could not have been used as a non-technical, ordinary word; or that even if it were meant in a technical
sense, that it was not superimposed on the opponents argument by Gautama himself. Olivelle also
argues that [t]he system attempts to fit itself into the pre-existing framework of life-cycle rituals
(saskra), giving a new significance to the period of initiatory studentship as a preparatory school for
all the adult modes of life (The rama System, 98), but again, this could be due to the Dharma Stra
authors own agenda, rather than that of their opponents.
221
Olivelle, Review of The Two Sources, 163.

102
argument can be made for the categories Brahmanical and non-Brahmanical.

Olivelle notes that Vedic and Brahmanical are often used interchangeably, and

that this is problematic because it is not necessarily the case that all brahmins must

have subscribed to everything Vedic. 222 But I would argue that the methodological

problem runs even deeper than this. Not only are the terms Vedic and non-Vedic

problematic, but the terms Brahmanical and non-Brahmanical are as well. As

indicated in the Introduction, one of the central theses of this dissertation is that the

term brhmaa was a disputed category in the late first millennium BCE. This is

made evident, in particular, by the trope of the true Brahman found throughout the

early Buddhist stras. 223 This rhetoric in the Buddhist texts has typically been

interpreted as marketing or a co-opting of Brahmanical categories to promote a

Buddhist agenda, and to a certain extent this is undoubtedly true. 224 But from a

methodological standpoint, why must we assume that when the Buddha calls himself

a Brahman (It. 4.100), he is just using the term rhetorically, but when a real

Brahman does, it can be taken literally?

222
Ibid.
223
Sn. 3.9=MN 98; DN 4; AN 4.5, 7.15, 7.1.3; Dhp. 383-423; Ud. 1.2-10, 2.6, 3.6; It. 3.69,
3.99, 4.100, 4.103; Ther. 221, 828-833, 1169-1171; Mv. 1.1.3.
224
In Part III I will describe what I see is a shift in the early Buddhist literature from a
straightforward use of the word brhmaa as a self-referential category to a more metaphorical use
of the word, as a counterpoint to its use by Vedic Brahmans.

103
II.2.2 The Rhetoric and Reality of Brahmanical Claims to Purity of Birth

Now, one could argue that what makes real Brahmans really Brahmans, and

Buddhists who call themselves Brahmans just rhetorical copy-cats, is the existence of

lineages of Brahmans who transmitted the Vedas in the centuries prior the time of the

Buddha. While such lineages certainly did exist, there has been a host of scholarship,

already discussed in the Introduction, that has called into question the antiquity of the

rigid conception of Brahmanhood as defined by birth. Although it is quite possible

that a conception of Brahmanhood by birth predates the Buddha, 225 it is important to

remember that the codification of the rauta ritual took place in a very specific part of

India, to the west of where the Buddha lived and early Buddhism flourished, and that

Vedic lineages only very slowly migrated east. Moreover, as Timothy Lubin has

argued, there is evidence that prior to the Ghya Stras, the term dvija referred

exclusively to Brahmans, and the institution of brahmacarya was reserved for

Brahmans, and in fact may have been what made one a Brahman. 226 Thus, it is not

necessary to assume great antiquity for an institution of Brahmanhood by birth to

explain the transmission of the Vedas in the centuries before the time of the Buddha.

225
As discussed in the Introduction, early Vedic references to brhmaas are generally
ambiguous about what exactly makes a person a Brahman. The word brahmabandhu is found in the
Chndogya Upaniad (6.1.1) used in the classical sense of Brahman by birth only. rui tells his
son vetaketu to enter brahmacarya because no one in their family has not studied and therefore
become a brahmabandhu: vetaketur hrueya sa. ta ha pitovca vetaketo vasa brahmacaryam. na
vai somysmatkulnonancya brahmabandhur iva bhavatti. (A presumably earlier use of the word
brahmabandhu is found at AB 7.27, but there it is used simply as a slur and the precise meaning is
unclear.) Generally, the early Upaniads, including the Chndogya, are taken to be earlier than the
Buddha, but this has been called into question by Bronkhorst (Greater Magadha, 207-218).
226
Lubin, Transmission, Patronage, and Prestige, 84-9.

104
 Indeed, there is evidence that even when there definitely was a conception of

Brahmanhood by birth, that conception was often more rhetorical than real. In the

Ambaha Sutta (DN 3), the Buddha humiliates a young Brahman student (mnava)

named Ambaha 227 by revealing that he is descended, not from a pure Brahmanical

lineage as he claims, but from the black (kaha) baby of a slave-girl of King Okkka

(i.e., Ikvku), whom the Sakyans regard as their ancestor. Now, this could simply be

dismissed as a polemical story, but the discussion that follows indicates otherwise.

The Buddha asks Ambaha a series of questions about how children of mixed unions

would be treated, respectively, by the Brahmans and by the khattiyas. In each case,

Ambaha admits that the Brahmans would accept the child of the mixed union, while

the khattiyas would notthus demonstrating that khattiyas are in fact more strict in

preserving pure lineages than the Brahmans are. 228 Of course, this dialog is a

Buddhist polemic as well, but given that it is presented as a reasoned argument, it is

227
Although the one extant version of this stra in Sanskrit Sanskritizes the name as Amba,
Bronkhorst argues that the name could also be Sanskritized as Ambaha, which is the name of a
mixed caste derived from a Brahman father and a non-Brahman (usually vaiya) motherGreater
Magadha, 354. This would make Ambahas name, quite possibly, a joke, which, as we will see in
Part IV, appears to be a common trope in the portrayal of Brahmans in encounter dialogs in the early
Buddhist stras.
228
This statement in the Ambaha Sutta is, in a sense, contradicted by the story of Rma
Jmadagnya killing all of the katriyas of the world 21 times over, insofar as, after he does so, katriya
women go to Brahmans to bear children by them so that they can continue the katriya lineage (see,
e.g., MBh. 1.58.1-9). But this story is, of course, the product of a Brahmanical polemical agenda that
was critical of katriyas who did not pay proper respect to Brahmanson which see Robert Goldman,
Masters of the Earth: The Bhgus and the Katriyas, in Gods, Priests, and Warriors: The Bhgus of
the Mahbhrata (New York: Columbia University Press, 1977), 93-112. And indeed, even as such,
the account of the repopulation of the katriya lineage after Rma Jmadagnyas genocide clearly
belies the Brahmanical claim that varas are determined strictly by birth, insofar as it accomplishes
that repopulation of the katriya lineage through a mixture of varas.

105
difficult to imagine that it would have been very convincing if it did not accurately

reflect actual social practice.

Another Buddhist text pokes fun at the pretensions of Brahmans with a similar

appeal to a social reality of impure Brahman lineages. A sutta in the Aguttara

Nikya (5.191), which has been studied by Oliver Freiberger, 229 lists five

brhmaadhamm that are now found among dogs, but no longer among Brahmans.

The first of these is taking only females of ones own kind as a mate; while dogs only

take female dogs as mates, Brahmans, according to this sutta, now frequently go to

non-Brahman women. The polemic of this sutta depends on a concept of the purity

of the Brahmans of olda trope found in several other early Buddhist texts as

well 230but it also depends on the observation, presumably derived from actual

social practice, that Brahman lineages are not pure.

Interestingly, an example of the sort of acceptance of children of impure

lineage as Brahmans that is discussed in the Buddhist texts can also be found in a

Brahmanical text, the Chndogya Upaniad. Here we find a short anecdote about

how the Upaniadic teacher Satyakma Jbla first came to enter brahmacarya.

Satyakma, wanting to study with Hridrumata Gautama, asks his mother what his

gotra was. She replies,

I do not know what gotra you are, dear. I had you in my youth when I was
maid who got around a lot. So I do not know what gotra you are. But I am

229
Freiberger, Negative Campaigning.
230
Sn. 2.7; DN 3, 27; SN 4.35.132; AN 5.192.

106
 named Jabl. You are named Satyakma. So you should just call yourself
Satyakma Jbla. 231

Satyakma goes to Hridrumata, and when the latter asks him what his gotra is, he

does not simply give his name as Satyakma Jbla, but honestly admits that he had

asked his mother about it and discovered that he was a bastard. At this, Hridrumata

replies, A non-Brahman would not declare that! (ChU 4.4.5: naitad abrhmao

vivaktum arhati) and accepts Satyakma as a student. As Brian Black comments on

this passage, whether or not Satyakma is actually a Brahman is left wonderfully

ambiguous, but what is interesting is that he is accepted as a student in spite of not

having any known paternal lineage whatsoever. 232

Bronkhorsts recent work on the spread of Brahmanism provides a useful

framework for understanding such examples of dubious Brahmanical lineages. As

231
ChU 4.4.2: nham etad veda tta yad gotras tvam asi. bahv aha carant paricri
yauvane tvm alabhe. sham etan na veda yad gotras tvam asi. jabl tu nmham asmi. satyakmo
nma tvam asi. sa satyakma eva jblo bruvth iti.
232
Black, Character of the Self, 54. Black assumes that the issue in this passage is whether
Satyakma is a Brahman or not, and that this is left ambiguous. As much would seem to be implied by
Hridrumatas exclamation that a non-Brahman would not speak as Satyakma did, but the technical
issue around which the passage revolves is actually not Brahmanhood per se, but Satyakmas gotra.
It is for his gotra that Satyakma asks his mother, and which she is not able to tell him because she
does not know who his father is, and it is for his gotra that Hridrumata asks when Satyakma comes
asking to be his student. Although gotra has come to be associated specifically with Brahmans,
Brough has demonstrated that this was not always the case. See John Brough, The Early History of
the Gotras, Journal of the Royal Asiatic Society of Great Britain and Ireland, no. 1 (Apr. 1946): 32-
45, and John Brough, The Early History of the Gotras (Concluded), Journal of the Royal Asiatic
Society of Great Britain and Ireland, no. 1 (Apr. 1947): 76-90. What may be at stake is not so much
being a Brahman from birth as coming from an established Aryan lineage, as would be established by a
gotra. Indeed, even in the encounter dialogs found in the early Buddhist literature in which Brahman
interlocutors define Brahmanhood in part in terms of birth, what is emphasized about birth is not
Brahman parentage, but rather purity going back seven generations. (See, e.g., the Soadaa Sutta,
DN 4.)

107
already discussed in the Introduction, Bronkhorst argues in Buddhism in the Shadow

of Brahmanism that

[t]he primary task of the new Brahmanism was to impose its vision of society.
Imposing its vision of society meant speaking about society as hierarchically
ordered into Brahmins, katriyas, vaiyas, and dras. Our earliest non-
brahmanical sources do no such thing. The Aokan inscriptions do not use
this terminology, even though they acknowledge the presence of Brahmins.
The early Buddhist canon does not do so either, with the exception of some
passages that normally discuss the brahmanical claims. 233

Indeed, as Uma Chakravarti has noted,

One important feature of the term vaa [in the Pali Canon] is that it appears
only in the context of abstract divisions of society into various social
categories. We have no evidence of it being used in any concrete situation.
It seems to have remained a theoretical concept without any parallel in actual
practice. 234

Furthermore, when vara is referred to in the Pali Canon, it is done so generally in

the context of debates between the Buddha and Brahman interlocutors, 235 what I am

referring to here as encounter dialogs. Thus, it appears that at the time these

encounter dialogs were written, Vedic Brahmans were pushing their hierarchical

vara theory of society, and the encounter dialogs record pushback against that

theory, involving, as we have just seen above, an apparently empirical appeal to a

mismatch between the theory of vara and social reality.

233
Bronkhorst, Buddhism in the Shadow of Brahmanism, 40.
234
Chakravarti, Social Dimensions, 104; cited in Bronkhorst, Buddhism in the Shadow of
Brahmanism, 35.
235
Chakravarti, Social Dimensions, 98; cited in Bronkhorst, Buddhism in the Shadow of
Brahmanism, 35.

108
 Nevertheless, as Bronkhorst has argued, Brahmanism was so successful not

because it converted people to its position by logical or other means, but because it

changed the terms of social discourse throughout India. Bronkhorst calls such a

strategy framing the debate and compares it to the way in which Republicans

introduced the phrase tax reliefwhich implies that taxes are an afflictioninto

American political discourse: Over time the media picked up on the phrase, and

eventually Democrats were forced to use it as well by defaultthus putting them in

the awkward position of implying by their own language that Republican hostility

toward taxes was justified. 236 It is important to note that for such a strategy to work,

it is not necessary to make a logically coherent argument, or, to put it differently, to

make an argument that ones opponents would (or at least theoretically could) be

logically convinced by. Indeed, as we have seen from the example of the Ambaha

Sutta, early Buddhists were not convinced by the theory of vara and Brahmanical

superiority thereby, based on the simple empirical observation that Brahmans do not

have pure lineages. What is important, rather, for such a strategy to work is simply to

frame the debate by popularizing terms that, when ingrained into the popular

consciousness, force people to think about the world in a way that seems to lead

naturally to ones own position.

236
Bronkhorst, Buddhism in the Shadow of Brahmanism, 163-4.

109
II.2.3 The Rhetoric of Mixed Castes

One of the ways in which Brahmans did this for vara, I would argue, other

than simply speaking about society in terms of the four varas per se, was with the

concept of mixed castes. This concept is not addressed in pastamba, 237 which

may be the earliest of the Dharma Stras, 238 but it is addressed systematically in the

other three major Dharma Stras, as well as later Dharmastra literature such as

Manu. 239 Although these systematizations purport to indicate how the particular child

of a particular union of two persons of different varas would be classified, both the

fact that the different Dharmastric authors often differ on which mixed class derives

from which mixed union, and more importantly the names that are given to many of

these mixed classes, indicate that there was a different purpose to these mixed class

schemas. That is, many of the so-called mixed classes have names that appear to

be those of either certain occupations or certain ethnic or tribal groups. Within the

Dharma Stras we find the following occupation-like names for mixed classes:

dhvara (fisherman), 240 karaa (lit., doer, a helper or companion), 241 katt

237
pastamba does mention some classes of people that are presented as mixed classes in
other Dharma Stras, namely, Clas, Paulkasas (elsewhere Pulkasas), and Vaia, but he does not
present them as resulting from mixed unions. Rather, they result when a Brahman, Katriya, or vaiya
is reborn as a human after sinning heinously in a previous life and spending time in hell (DhS 2.2.6-7).
238
Olivelle, Dharmastras, xxxi. Olivelle cites the fact that pastamba does not deal with
mixed classes as one reason for believing that it is older than Gautama.
239
GDhS 4.16-28; BDhS 1.16.6-16, 1.17.1-15, 2.3.29-30, 1.2.13; VDhS 18.1-10; MDh 10.8-
73.
240
GDhS 4.19.
241
GDhS 4.21.

110
(attendant or doorkeeper), 242 mrdhvasikta (expert in dhanurveda), 243 rathakra

(chariot-maker), 244 sta (charioteer), 245 vapka (dog-cooker), 246 and vaia (maker of

bamboo-work). 247 Even more interestingly, we find the following names that appear

to refer to tribal groups or inhabitants of various countries: ambaha (the name of a

country and its inhabitants), 248 mgadha (person from Magadha), 249 nida (a non-

Aryan tribe), 250 praava (a people in the southwest), 251 vaideha/vaidehaka (person

from Videha), 252 and even yavana (a Greek!). 253

Now, it is of course possible that some of these names were used in different

senses, one relating to an ethnic group or occupation and the other to a mixed class,

and indeed the normative force of the Dharmastric texts may have led to this

242
GDhS 4.17; BDhS 1.17.1, 1.17.7, 1.17.10, 1.17.11; cf. MDh 10.12, 10.13, 10.16, 10.19,
10.26, 10.49.
243
GDhS 4.19.
244
BDhS 1.17.1, 1.17.6.
245
GDhS 4.17, 4.18; BDhS 1.17.1, 1.17.3; VDhS 18.6; cf. MDh 10.11, 10.17, 10.26, 10.47.
246
BDhS 1.16.9, 1.17.1, 1.17.11; cf. MDh 10.19.
247
BDhS 1.16.10, 1.17.1, 1.17.12, VDhS 18.2; cf. MDh 10.19, 10.49 (vea).
248
GDhS 4.16; BDhS 1.16.7, 1.17.1, 1.17.3, 1.17.9, 1.17.12; VDhS 18.8; cf. MDh 10.8, 10.13,
10.15, 10.19, 10.47.
249
GDhS 4.17, 4.18; BDhS 1.16.8, 1.17.1, 1.17.7; cf. MDh 10.11, 10.17, 10.26, 10.47.
250
GDhS 4.16; BDhS 1.16.7, 1.16.11, 1.16.13, 1.17.3, 1.17.13, 1.17.14, 2.3.29; VDhS 18.8; cf.
MDh 10.8, 10.18, 10.34, 10.36, 10.37, 10.39, 10.48.
251
GDhS 4.16, 4.21; BDhS 1.17.4, 2.3.30; VDhS 18.9; cf. MDh 10.8.
252
GDhS 3.10; BDhS 1.16.8, 1.17.1, 1.17.8, 1.17.10, 1.17.12; cf. MDh 10.11, 10.13, 10.17,
10.19, 10.26, 10.33, 10.36, 10.37, 10.47.
253
GDhS 4.21.

111
becoming the case. The overall preponderance of such names in the schemes of

mixed classes, however, is highly suggestive that their purpose was not so much to

define the status of the progeny of various un-dharmic sexual unions 254 as to

naturalize the vara system, which by itself likely bore no resemblance to social

reality. It did so by mapping the varas, through the concept of mixture, onto pre-

existing social groups. Some of these groups were likely caste-like groups within

(or on the margins of) Aryan societysuch as Clas 255but others may have

simply been ethnic groups that were not recognized as fully Aryan and whose

existence could be explained as resulting from intermarriage between hypothetically

pure varas in the past. 256 Such an interpretation appears to be confirmed by the fact

that Baudhyana, in a section of his Dharma Stra apart from the enumeration of

mixed classes, declares, The people of Avant; the people of Aga and Magadha; the

254
It is not clear how this would have worked on a social level anyway. Insofar as
Dharmastric prohibitions against mixing classes were recognized and accepted (which in itself is
questionable), illicit unions would presumably have been hidden and difficult to verify in an age before
paternity tests. Indeed, Manu seems to acknowledge as much when he says that Ignobility, cruelty,
ferocity, (and) neglect of rites / Make manifest here in the world a man born of a foul womb. // He
partakes in the paternal tendencies, or those of the mother, or even both. / In no wise does a person
from a bad womb restrain his own nature. // Even in an eminent family, a man of whose birth there be
a mixture of wombs / Necessarily joins together with those tendencies to a lesser or greater extent. //
(MDh 10.58-60: anryat nihurat krrat nikriytmat / purua vyajayantha loke
kaluayonijam // pitrya v bhajate la mtur vobhayam eva v / na kathacana duryoni prakti
sv niyacchati // kule mukhyepi jtasya yasya syd yonisakara / sarayaty eva tac chla
narolpam api v bahu//). There may be no DNA test to ferret out mixed-breed children, but they
inevitably reveal themselves by their acts.
255
Clas are, in fact, mentioned frequently in the Pali Canon as a despised social group.
In several texts, they are mentioned together with other groups that are listed in the Dharma Stras as
mixed classes (vena, nesda, rathakra, pukkusa): Vin. 2.5.1.2, 2.5.1.3; MN 93, 96, 129; SN 1.3.3.1;
AN 3.1.2.3, 4.2.4.5, 5.4.5.2, 6.2.6.3.
256
The early ascription of Mgadhas and Vaidehas to mixed class status may be related to
the fact that they are eastern peoples of Greater Magadha, which, as Bronkhorst has pointed out, was
not originally considered part of ryavarta (Bronkhorst, Greater Magadha, 1).

112
people of Surra; the people of the Southern Path; the people of Upvt, the Sindh,

and Suvrathese come from mixed wombs. 257 In Manu we find even more

references to ethnic groups as mixed classes: bhra, andhra, avantya, dravia, khasa,

licchivi, malla, stvata, vadhna. 258 In addition, a whole host of well-known non-

Aryan peoples are said to be katriyas who have become valas by neglecting rites

and disregard[ing] Brahmans 259: puraka, coa (Coa), dravia (Dravidian),

kmboja, yavana (Greek), aka, prada, pahlava, cna (Chinese), kirta, darada.

II.2.4 The rama System as a New Brahmanical Systematization of Actual Social

Practice

Summing up, then, if the vara system was pushed as a framing tactic by,

as Bronkhorst puts it, a new Brahmanism, and the rhetoric of vara did not

necessarily correspond fully, or in some cases at all, to social reality, then it becomes

problematic to speak of Brahmanism in ancient India as a well-defined entity to be

contrasted to a non-Brahmanical other. 260 To do so is to accept uncritically the very

257
BDhS 1.2.13: avantayogamagadh surr dakipath / upvtsindhusauvr ete
sakrayonaya.
258
MDh 10.15, 10.36, 10.21-23.
259
MDh 10.43: anakais tu kriylopdim katriyajtaya / valatva gat loke
brhmadaranena ca //.
260
Obviously, I am not denying that there was a coherent Vedic tradition from the time that
the rauta ritual was codified in the Kuru-Pcla state until the time period that we are interested in
here (late first millennium BCE). What I am saying is that we cannot uncritically accept Brahmanical
sources, especially late sources such as the Dharma Stras, where the vara system is for the first time
treated systematically, in developing our understanding of what provided coherence to that Vedic
tradition. Indeed, as Timothy Lubin has argued, brahmacarya and dvija status were, before the Ghya

113
vara rhetoric that the Brahmansthat is, the proponents of what Bronkhorst calls

the new Brahmanism, of the sort that appear as interlocutors in encounter dialogs

in the early Buddhist strasused to advance their vision of society. Put differently,

why must we assume that the Brahmans, the proponents of the new Brahmanism,

owned the word brhmaa? As we have already seen, Buddhists often referred to

themselves as Brahmans in their early texts. Although, as I will show in Part IV,

Buddhists eventually ceded this identity to the proponents of the new Brahmanism,

as a matter of methodology there is no good reason to assume a priori that the texts

we now call Brahmanical come from the real Brahmans, while those that we now

call Buddhist come from non-Brahmans who used the word brhmaa merely

as a metaphor, a polemic, or the like. The proponents of the new Brahmanism, then,

would be a particular group, arrogating to themselves the identity Brahman, in a

context in which the category was still contested and utilized by other groups such as

the Buddhists. 261

Returning now to the matter of the presentation of the rama system in the

Dharma Stras, we can address the question of whom Gautama and Baudhyana were

arguing against when they rejected the rama system. Clearly, all of the Dharma

Stras, used in reference only to Brahmans, and may have been so precisely because brahmacarya was
what made one a Brahman.
261
Although, as already indicated earlier in Part II, the encounter dialogs of the Pali Canon
already seem to accept the proponents of the new Brahmanism as the Brahmans insofar as they use
the category samaa to refer to the Buddha and the category brhmaa to refer to his interlocutors,
there are some references to proponents of the new Brahmanism in the Pali Canon that do seem to
indicate their particularity as Brahmans by using a qualifier when referring to them as such. For
example, in the Tevijja Sutta (DN 13), they are referred to repeatedly as tevijj brhmaBrahmans
of the Triple Veda. Likewise, SN 4.42.6 and AN 10.176 both refer to Brahmans of the western
region (brhma pacchbhmak).

114
Stras represent the new Brahmanism insofar as they advocate the vara system;

declare Brahmans superior to the other three classes by birth; and, like the Buddhas

Brahman interlocutors in the encounter dialogs, adhere to the householder lifestyle

as the best, or even the only dharmic, way of life. What then of the opponents they

allude to who support the rama system? It is true that, as already discussed, some

of these opponents appear to have made arguments based on the Vedas, but all this

definitely means is that there was someone whom Baudhyana (the only Dharma

Stra author who makes reference to such arguments) saw fit to refute, who was

willing and able to make an argument based on Vedic scripture in support of lifelong

celibate lifestyles. One can call such a person a liberal Brahman 262 if one wishes,

but more apropos for our purposes is that he is not within the tradition represented by

the Dharma Stras. He is not, in other words, an advocate of the new Brahmanism.

Indeed, the rhetoric used against the supporters of the rama system in the

Dharma Stras suggests not so much a point of disagreement within a self-conscious

tradition, but a boundary marker serving to distinguish the tradition out of which the

Dharma Stras come from those outside that tradition. Baudhyana, the only Dharma

Stra that presents Vedic quotes in support of the rama system, 263 also quotes the

262
Olivelle, The rama System, 96.
263
pastamba does present, in the mouths of his opponents, the following quote from a
Pura: The eighty thousand seers who desired offspring went along the suns southern course.
They obtained cremation grounds. The eighty thousand seers who did not desire offspring went along
the suns northern course. They, indeed, attained immortality (DhS 2.23.4-5: atisahasri ye
prajmir araya. dakienryama panthna te manni bhejire. atisahasri ye praj
neir araya. uttarerama panthna temtatva hi kalpate). The second part of the opponents
argument is that celibate religious practitioners have special powers. In reference to these two pieces
of evidencethe quote from the Pura and the special powers of renunciatespastamba has his
opponents summarize their argument by saying, Therefore, on the basis of ruti and visible results,

115
following saying in support of his position that there is only a single rama, the

householder rama: There was an asura named Kapila, the son of Prahlda. He made

this division, battling with the gods. A wise person should pay no attention to them. 264 So

even if Baudhyanas opponents are able to quote Vedic texts in support of their

position, they are not merely other Brahmans from Baudhyanas own tradition with

whom he respectfully disagrees over their interpretation of the Vedas; they are in fact

followers of a demonic enemy of the gods. Likewise, pastamba, after presenting the

view of his opponents that the celibate ramas are superior to that of the householder,

begins his rebuttal by stating simply, The position of the elders of the Triple Veda,

however, is that the Vedas are the standard. 265 Again, this does not seem like a

statement directed toward a member of pastambas own tradition; rather, it sounds

like a concise definition of pastambas own tradition, precisely in that aspect that

most distinguishes it from that of his opponents. pastamba, like the other Dharma

Stra authors, and like all of the Brahman interlocutors with whom the Buddha

speaks in the encounter dialogs of the Pali Canon, is a Brahman who defines

himself in terms of his knowledge of and adherence to the three Vedas.

some say that these ramas are superior. (DhS 2.23.9: tasmc chrutita pratyakaphalatvcca
viin ramn etn eke bruvate). The Pura quote is clearly referred to here as ruti, which
Olivelle (Dharmastras, 67) translates as vedic testimony, but it seems that the word is not being
used here in the technical sense it came to have later, as referring specifically to Vedic revelation. As
far as I know, the quote has never been traced, and it is in any case presented as coming from a
Pura, not the Vedas.
264
BDhS 2.11.28: prhldir ha vai kapilo nmsura sa. sa etn bhed cakra devai
saha spardhamna. tn man ndriyeta.
265
DhS 2.23.10: traividyavddhn tu ved pramam iti nih.

116
 Is it possible, then, that the Dharma Stra authors cited opinions not only of

fellow Vedic Brahmans who happened to have minor differences of opinion, but also

of those whose opinions were fundamentally opposed to their own brand of Veda-

centric Brahmanism? I believe that this is quite likely. A telling clue that this is what

is happening in the case of the rama system can be found in pastambas treatment

of the parivrjaka. pastamba writes that discarded clothing is prescribed for him,

but then adds, Some (say he should be) free of all (i.e., he should go naked). 266 The

option for a parivrjaka to go naked is presented, as with other options in the Dharma

Stras, as the opinion of some (eke), but it seems unlikely that pastamba is simply

citing here the opinion of certain other Vedic Brahmansthat is, other teachers in his

own tradition, with whom he would identify in some waythat a parivrjaka should

go naked. Rather, given what we know about Indian religious practice in this time

period, it seems likely that he is simply describing one particular parivrjaka practice,

the acelaka practiceas it was actually observed in his day.

Indeed, an often-overlooked aspect of the Dharma Stras is that they are not

simply normative texts prescribing what people should do; while they certainly have a

normative agenda, they also have a commitment to describing what people do do. As

Richard Lariviere has argued of dharma as it is conceived in the Dharma Stras,

There is much made of the Vedic source, but ultimately, the immediate source is

custom. The legal texts themselves tell us this in very clear terms. All custom is

266
DhS 2.21.11-12: tasya muktam cchdana vihitam. sarvata parimokam eke.

117
binding. 267 Lariviere argues that the fact that the dharma literature is a record of

custom is obfuscated by the fact that the idiom of all the dharma literature is one of

eternality and timelessness. It is further obfuscated by the fact that the dharma

literature clings to the claim that all of its provisions can be traced directly or

indirectly to the Veda, the very root of dharma. 268 The Dharma Stras do, in fact,

explicitly cite actual practice as a source of dharma, although they rank it below the

Vedas (DhS 1.1.1-3, GDhS 1.1-2, BDhS 1.1.1-4, VDhS 1.4-5). This privileging of

the Veda, Lariviere argues, was rhetorical, and had to be accommodated to actual

custom:

In general, however, the brhmaa dharmastra writers were constrained by

the burden placed on them as recorders and synthesizers of customary practice.
They were obliged by the interested constituencies, by the king, and by
considerations of social and political harmony to record the practice as they
found it. They were also obliged to explain how these customs fit with the
tradition, and it is in these explanations that we may find the most outr
flights of brhmaa imagination. 269

The task of the Dharma Stra authors, in other words, was not simply to determine

dharma through exegesis of the Vedic texts; it was also to justify established

customary law as dharma on the basis of the Vedas. 270

267
Richard Lariviere, Dharmastra, Custom, Real Law and Apocryphal Smtis, Journal
of Indian Philosophy 32 (2004), 618.
268
Ibid., 612.
269
Ibid., 616.
270
There is an interesting parallel, I would argue, between Larivieres understanding of the
Dharmastric project and that of the Sanskrit grammarians, especially when viewed through the lens
of Bronkhorsts recent comments on Sanskritization. There is, for both bodies of literature, a certain
scientific or empirical componentthe dharmastra authors must take into account actual human
custom in formulating their accounts of dharma, just as the grammarians must take into account actual
human speech in formulating their accounts of grammar. Both, however, are overlaid by an

118
 This is not to say that the Dharma Stra authors accepted all aspects of actual

social practice and found ways to justify them on the basis of the Veda. It does

appear, however, that they did attempt to address a wide variety of social practices in

one way or another. I believe that the rama system serves as a prime example of

this. rama as it is presented in the Dharma Stras, I would argue, served as a

means for advocates of the new Brahmanism to account for, taxonomize, and respond

to much of the same empirical social datai.e., actual social practicesthat were

referred to more loosely in colloquial discourse by the concept of ramaabrhma.

Now, this does not at all preclude the possibility that, as Olivelle has argued, the

original idea of a schema of four ramas was formulated by someone else who

favored and wanted to legitimate celibate, ascetic lifestyleswhether that may have

been a liberal Brahman, a person from an entirely different group willing to make

an argument from a Vedic text, or someone else. What I would like to emphasize,

however, is that there is no direct evidence for such a source; the only evidence we do

have is for a colloquial discourse on ramaabrhma on the one hand and an elite,

systematized discourse on roughly the same social phenomena, but conceived of as

four ramas, confined to the emerging orthodox Brahmanical tradition (i.e., the

new Brahmanism). The adoption of the language of rama would also explain

why there is a relative lack of reference to ramaas or ramaabrhmaas in the

ideological agenda that colors the presentation of actual custom or language and at times pushes
against it. While that ideological agenda in the dharmastra literature is obvious, so much so that,
Lariviere argues, it obscures its roots in customary law, the ideological agenda of Sanskrit grammar is
less obvious and has even been deniedsee Pollock, The Language of the Gods. Bronkhorst, however,
has argued convincingly that Sanskritization was intimately intertwined with the spread of the
Brahmanical ideological agendaBronkhorst, Buddhism in the Shadow of Brahmanism, 46-65.

119
Brahmanical literature. There would be no need to refer to these colloquial categories

when the tradition had at its disposal a set of categories that both were more

systematic, and thus more amenable to the taxonomizing tendency within

Brahmanical discourse, and also, whatever the original purpose of the system may

have been, ultimately proved amenable to advancing a particular Brahmanical agenda.

Seen in this light, the presentation of the rama system by the different

Dharma Stras need not be viewed primarily as a matter of acceptance or rejection.

All of the four Dharma Stras clearly fall within the aegis of a new Brahmanical

movement that privileged the householder lifestyle, and none of them show much

sympathy toward the radical celibate lifestyles that the three non-householder

ramas were meant to refer to. Although Olivelle refers to pastamba and Vasiha

as accepting the rama system, 271 pastamba delivers a lengthy argument against

those who see the celibate ramas as superior, and he maintains that immortality is

(having) offspring 272; likewise, Vasiha declares that of the four ramas, the

householder is superior. 273 I would argue, therefore, that pastamba and Vasiha

do not accept the rama system so much as Gautama and Baudhyana go an extra

rhetorical step in their critique of it by rejecting the non-householder ramas

altogether. All four authors are clearly uncomfortable with celibate lifestyles

lifestyles that moreover had a tendency to promote teachings that obviated the need

271
Olivelle, The rama System, 91.
272
DhS 2.24.1: prajtim amtam.
273
VDhS 8.14: caturm ram tu ghasthas tu viiyate.

120
for study of the Vedas 274; they simply differ in how exactly they respond to such

lifestyles, which they clearly felt obligated to comment upon since they formed a

significant part of actual social practice. Gautama and Baudhyana may have

actually been the more radical of the four insofar as they rejected the celibate ramas

outright, instead of adopting the more usual Dharmastric rhetorical practice of

justifying social customs within a Vedic framework. 275 They do not appear to

represent a diachronic movement either toward or away from rejection of the celibate

274
Although none of the Dharma Stras explicitly label the practitioners of the celibate
ramas as representing other or even rival groups, it is clear from their critiques of those ramas
that they were concerned with issues of Vedic learning and ritual practice, rather than only the issue of
whether it is necessary to produce offspring. In other words, it appears that the celibate ramas were
problematic not only because they failed to produce offspringsomething that could be understood as
a fully intra-Brahmanical disputebut because they did not acknowledge the need for studying the
Vedas and performing rituals. As already mentioned above, pastamba begins his rebuttal of the
opinion that the celibate ramas were superior by stating that the Vedas are to be taken as the
standarda point that hardly seems necessary to make if the opponents were simply liberal
Brahmans. Baudhyana attributes, as already mentioned, two Vedic quotes to his opponents, but he
ends his rebuttal of their position by stating, The three knowledges, brahmacarya, procreation, faith,
austerity, sacrifice, gift-giving: We are with only those who do these things; praising something else,
one becomes dust and perishes (BDhS 2.11.34: tray vidy brahmacarya prajti raddh
tapo yajam anupradnam. ya etni kurvate tair it saha smo rajo bhtv dhvasatenyat praasann
iti). This appears to quite sharply demarcate a line between a Vedic tradition and a non-Vedic other.
Finally, Vasiha ends his discussion of the ramas by, as already mentioned, declaring the
householder superior to the other three, followed by what I would characterize as a succinct description
of his vision of the Brahmanical tradition: Always performing ablutions, always wearing the sacred
thread, always reciting (the Veda), avoiding the food of outcastes, both going (to his wife) in season
and making offerings according to the rule, a Brahman does not fall from brahmaloka (VDhS 8.17:
nityodak nityayajopavt nityasvdhyy patitnnavarj. tau ca gacchan vidhivac ca juhvan na
brhmaa cyavate brahmalokd iti).
275
Again, justifying social customs within a Vedic framework need not imply that social
practices were always recorded with complete fidelity to the way in which they were actually practiced,
with sophistic arguments provided to legitimize them as such on the basis of the Vedas. pastamba,
for example, begins his description of the ramas by stating that all are required not to abandon
Vedic learning (DhS 2.21.4: sarvem antsargo vidyy). This need not be interpreted as
indicating that pastamba is only talking about Brahmans who pursue celibate ramas such as that of
the parivrjaka; in fact, as I have argued, his rebuttal of the claim that celibate ramas are superior
appears to presuppose that those opponents question or reject the authority of the Vedas. The assertion
that those who enter into various ramas must not abandon the Vedas is, in my opinion, better
interpreted as a normativizing fictionan attempt to reconcile diverse ways of life found in actual
practice with pastambas own Vedic ideal.

121
ramas 276; in fact, they appear to represent an idiosyncratic opinion that was

superseded by the accomodationist rhetoric and the shift to understanding ramas as

stages that became dominant in the later classical Dharmastric texts. In spite of

their differences, they are united with the other two Dharma Stras in using the

rama system as a vehicle to grapple with the empirical existence of lifestyles other

than their own preferred householder lifestyle. 277

276
If Olivelle is right in dating pastamba before Gautama, then the two Dharma Stras that
reject the celibate ramas outrightGautama and Baudhyanawould fall chronologically right in
the middle of the four.
277
The term rama is actually well suited to a discussion of lifestyles from a perspective
that favors a lifestyle that privileges producing progeny. As Olivelle explains in his study of the
rama system, the closely related word rama is used most significantly in the Vedas to refer to the
creative activities of Prajpati; it, moreover, is used as a synonym for sacrifice and has clear sexual
connotationsOlivelle, The rama System, 10.

122
 Chapter II.3

The Colloquial Categories that were Subsumed into the

rama System and their Treatment by the Buddhists

II.3.1 Greek Evidence for the Use of rama Categories in Colloquial Discourse

Although the rama system as a theoretical system is unique to Brahmanical

texts, there is ample evidence that the four ramas, however contrived they might

have been in their purport to cover the full range of lifestyles in ancient India, were

indeed derived from colloquial categories. As I argued in the Introduction, an often-

quoted citation from Megasthenes, a Greek ambassador to the court of Candragupta

Maurya, appears to superimpose the brhmaa-ramaa distinction on a set of

practitioners roughly coterminous with the three celibate ramas. The following

description, for example, is given of the :

The philosophers reside in a grove in front of the city within a moderate-sized

enclosure. They live in a simple style and lie on pallets of straw and (deer)
skins. They abstain from animal food and sexual pleasures; and occupy their
time in listening to serious discourse and in imparting knowledge to willing
ears. After living in this manner for seven-and-thirty years, each individual
retires to his own possessions, where he lives in security and under less
restraint, wearing robes of muslin and a few gold ornaments on his fingers and

123
 in his ears. They may marry as many wives as they please, with a view to
having many children, for from many wives greater advantages are derived. 278

Strictly speaking, this does not describe the brahmacrin rama, but rather a site for

practicing temporary brahmacarya and the normative life cycle, culminating in the

householder rama, prescribed by the Dharma Stras. One remains celibate while

dwelling in the grove in front of the city, and the time spent there is limited

(although the amount of time given is anomalousthe Dharma Stras prescribe a

maximum of 48 years, or 12 for each Veda 279). This is followed by what clearly is a

householders existence; notice not only the mention of possessions, but also the

emphasis on producing children.

The description of the [] attributed by Strabo to Megasthenes, on the

other hand, divides them into groups that clearly parallel the vnaprastha and the

parivrjaka of the rama system. First, Strabo describes the so-called ,

which means living in the woods and thus is a credible translation of the Sanskrit

vnaprastha. This is confirmed by the description of the that he gives:

Of the Sarmanes the most honourable, he [Megasthenes] says, are those called
the Hylobioi. They live in the forests, subsist on leaves and wild fruits, wear
garments made from the bark of trees, and abstain from wine and commerce
with women. They communicate with the kings who consult them by
messengers regarding the causes of things, and who through them worship and
supplicate the deity. 280

278
Strabo, Geography, 15.59; trans. McCrindle, Ancient India as Described in Classical
Literature, 65-6.
279
DhS 1.2.11-16; GDhS 2.45-47; BDhS 1.3.1-4.
280
Strabo, Geography, 15.60; trans. McCrindle, Ancient India as Described in Classical
Literature, 67.

124
This closely parallels the description of the vnaprastha found in the Dharma Stras;

for example, pastamba writes,

From that very state [temporary Vedic studentship], remaining celibate, he

goes forth. Clothes of the wilderness are prescribed for him. Then he
should wander subsisting on roots, fruits, leaves, and grass. 281

The similarities are unmistakable: Both the and the vnaprastha bear the

name forest-dweller, remain celibate, wear natural clothes, and live off of food that

they gather from the wild.

Megasthenes second category of [], the physicians (),

while lacking any discernible relationship to the final remaining rama category, the

parivrjaka, in nomenclature, nonetheless clearly parallel them in the substance of

their description. Strabo begins his description of them in the following terms:

Next in honour to the Hylobioi are the physicians, for they apply philosophy
to the study of the nature of man. They are frugal in their habits, but do not
live in the fields. Their food consists of rice and barley-meal, which every
one gives who is asked, as well as every one who receives them as a guest.
Women study philosophy with some of them, but they too abstain from sexual
intercourse. 282

This clearly refers, in addition of course to celibacy, to the practice of living off of

alms given by ordinary householders, and most likely refers as well to the path of

insight into the nature of the tman that a parivrjaka is said to follow (DhS

2.21.13-14; VDhS 10.14).

281
DhS 2.21.19, 2.22.1-2: ata eva brahmacaryavn pravrajati. tasyrayam cchdana
vihitam. tato mlai phalai parais tair iti vartaya caret.
282
Strabo, Geography, 15.60; trans. McCrindle, Ancient India as Described in Classical
Literature, 67.

125
 It appears, however, that much of Megasthenes description of the physicians

is based on their actual practices, rather than any normative ideal. A clear distinction

is made between them and the through the observation that, though frugal,

they do not live in the fields. One cannot help but be reminded here of the tension

in Buddhism between the ideal of wandering nine months of the year and the reality,

which may likely have become the norm by Candraguptas time, of living year-round

in settled monasteries. Actual practice, as opposed to normative ideals, may also

explain the appellation physician, which Strabo elaborates upon in some detail:

By their knowledge of medicine they can make persons have a numerous

offspring, and make also the children to be either male or female. They effect
cures rather by regulating diet than by the use of medicines. The remedies in
most repute are ointments and plasters. All others they suppose to partake
largely of a noxious nature. 283

Although such practices are not listed in the normative descriptions of the parivrjaka,

nor, to my knowledge, in the normative accounts of Jain or Buddhist monks, we

know that they must have been current among ascetic groups since they are listed

among the various bestial sciences practiced by some ascetics and Brahmans that

the Buddha can be praised for abstaining from in the Brahmajla Sutta (DN 1):

While some venerable samaas and brhmaas, having eaten food given in
faith, make a living by wrong livelihood, by such bestial sciences ascausing
virility, causing impotencean emetic, purgative, up-purge, down-purge,
head-purge, anointing of the ear, cleansing of the eye, treatment of the nose,
ointment, anointing, ophthalmology, surgery, pediatrics, the giving of a balm
as remedy for a medicine previously given, and so forth, the samaa Gotama
abstains from wrong livelihood, from such bestial sciences as these. 284
283
Ibid.
284
yath v paneke bhonto samaabrhma saddhdeyyni bhojanni bhujitv te
evarpya tiracchnavijjya micchjvena jvita kappenti, seyyathidavassakamma
vossakammavamana virecana uddhavirecana adhovirecana ssavirecana kaatela

126
We can feel confident that these medical arts, as well as the many other practices

that are criticized in the Brahmajla Sutta, most of which pertain to astrology and

divination, were indeed practiced by ancient Indian ascetics not only because it would

be pointless for the author of this Buddhist text to criticize them if they did not, but

also because many of them, including folk medicine, are practiced even in modern

times by Buddhist monks. 285

Thus, we have in the testimony of Megasthenes evidence that categories

similar to those found in the rama systemcertainly, at least, those of the

vnaprastha and parivrjakawere indeed in colloquial use in India around the time

of the writing of the Dharma Stras. This is confirmed by the fact that similar

categories are also found in the early Buddhist texts, though not in any wise

organized into a four-fold system. In fact, what is interesting about these categories

as they are found in the Buddhist texts is that they do not at all follow the same

pattern as the Brahmanical rama system, and in fact are never mentioned together

in one place, yet when considered together demonstrate an early Buddhist articulation

nettatappana natthukamma ajana paccajana slkiya sallakattiya drakatikicch

mlabhesajjna anuppadna osadhna paimokkho iti v iti evarpya tiracchnavijjya
micchjv paivirato samao gotamo ti.
285
My wife did a study of folk medicine still performed by a small number of doctor
monks () in the Thai province of Nonthaburi. The practices she observed included
exorcism; the use of medicaments and yantras to cure skin rashes; and the use of herbs to cure dog,
snake, and insect bitesNanda Raksakhom, [Medical
Treatment using Religious Methods] (BA Senior Thesis, College of Religious Studies, Mahidol
University, 2005). The performance of simple astrological and other divinological calculations,
which are also proscribed by the Brahmajla Sutta, are, on the other hand, a commonplace among
Thai monks.

127
of self-identity that offers a striking parallel to the new Brahmanical identity

constructed by the early Dharmastric presentation of the rama system.

II.3.2 Ghastha and Gahapati

The first (or, according to Gautama and Baudhyana, only) rama, the

ghastha or householder, is reflected in early Buddhist texts by the category of the

gahapati. Although the word is not exactly the same, at face value it would appear to

convey a similar meaning, though with a greater emphasis on such a persons lordship

or dominion over the house in which he lives. That the basic meaning of the word is

simply a person who lives in a housei.e., a householderis confirmed, however,

by the way it is defined, repeatedly, in the Suttavibhaga of the Vinaya, which is a

commentary on the Ptimokkha. For Ptimokkha rules in which the word gahapati

appears, the following gloss is given: A gahapati is whoever inhabits a house. 286

This is the same gloss that is given for other rules 287 that instead use the word agrika,

implying that the two words are interchangeable.

Uma Chakravarti has argued, however, that there is more to the gahapati than

simply living in a house. She writes, While the term gahapati in the sense of

householder or one who lives in a house or possesses a house is equivalent to other

words implying the samesuch as gihi, gahaha and ajjhvasatithese terms do

286
Vin. 1.4.1.6, 1.4.1.7, 1.4.1.8, 1.4.3.7, 3.2.1, 3.4.4.6: gahapati nma yo koci agra
ajjhvasati.
287
Vin. 3.4.3.8, 3.4.5.6.

128
not imply the range of characteristics that gahapati carries with it. 288 The category

gahapati is often found together with the categories khattiya and brhmaa, and as

such, Chakravarti argues that it represents the economy, as one of the seven

treasures (sattaratanni) of a cakkavatti (wheel-turning) monarch, 289 as a

controller of property, as a tax-payer, and as one who makes a living off of

agriculture. 290 The category appears to imply some status and to refer specifically to

the paterfamilias, rather than simply to any human being who happens to live in a

house. 291 Indeed, as Chakravarti notes, we have also the term gahapatiputta, which is

glossed twice in the Suttavibhaga as whoever are sons and brothers (i.e., of the

gahapati, who is mentioned just prior). 292 This would appear to refer to (male)

dependents of the gahapati, who would in turn be the head of the household.

Chakravarti is, therefore, certainly correct that what gahapati generally

referred to was not merely a householder in the sense of a person living in a house;

it was a title of status, referring to a true lord of a household, a person of means and

a controller of property. In fact, we do find in a few instances in the Pali Canon,

when a more generic sense of one living in a house is intended, the word gahaha

288
Chakravarti, Social Dimensions, 66.
289
DN 26: bhtapubba, bhikkhave, rj dahanemi nma ahosi cakkavatt dhammiko
dhammarj cturanto vijitv janapadatthvariyappatto sattaratanasamanngato. tassimni satta
ratanni ahesu seyyathida cakkaratanau hatthiratana assaratana mairatana itthiratana
gahapatiratana pariyakaratanameva sattama.
290
Chakravarti, Social Dimensions, 67-71.
291
Ibid., 80.
292
Vin. 3.2.1, 3.4.4.6: gahapatiputto nma ye keci puttabhtaro.

129
(which is the Pali equivalent of ghastha) itself. This word is found far less

frequently than the word gahapati, and it is often paired together with pabbajita to

refer, quite literally, to those who stay at home and those who go forth. 293 The

function here is clearly to refer to living in a house in general, rather than anything

more specific as in the case of gahapati. Given that this more general term was

known and available to the authors of the early Buddhist texts, then, we must ask

ourselves why they generally preferred gahapati as a category to speak about

householders. In addition, to what extent does the category of the gahapati as it is

found in the early Buddhist texts differ from that of the ghastha as it is found in the

Dharma Stras?

In response to the second question, I would argue not as much as it might

seem at first. It is true that the category gahapati is often found together with the

categories khattiya and brhmaa, which could be taken to imply a sort of three-fold

division of society, 294 perhaps even a colloquial equivalent to the Brahmans vara

system. Chakravarti adduces in support of this supposition another gloss in the

Suttavibhaga that defines a gahapatika as follows: Setting aside a king, a kings

servant, (or) a Brahman, what remains is a called a gahapatika. 295 This gloss should

not be taken too literally, however, as a comprehensive division of society into its

components. If we look back at the original rule that it is commenting upon, we find

293
Vin. 1.1.4, 1.1.6, 1.2.7, 1.4.2.6, 2.5.1.9; MN 11; AN 5.57, 5.88, 5.90, 5.223, 7.61.
294
Chakravarti calls it a triumvirate: Social Dimensions, 67.
295
Vin. 1.4.1.10: hapetv rja rjabhogga brhmaa avaseso gahapatiko nma.

130
that it is simply mimicking the categories found there: The rule begins, In case a

king or one in the service of a king or a brahmin or a householder should send a robe-

fund for a monk by a messenger. 296 The gloss on gahapatika, in other words,

simply indicates that in this rule the word gahapatika is being used as a catch-all term

for anyone other than a king, a kings servant, or a Brahman who happens to send a

robe-fund for a monk. No comprehensive division of society into khattiyas,

brhmaas, and gahapatis is implied. Moreover, although the category is often

found throughout the Canon together with the categories khattiya and brhmaa, they

are never found together in compound, nor, for that matter, necessarily found in

isolation as a group of three. 297 The more important lesson to be taken from the

frequent use of gahapati together with khattiya and brhmaa is not that they form a

comprehensive division of society into three parts, but rather that, as Bronkhorst has

noted, early Buddhist texts fail to use the vara system to divide society, even when

categories shared in common with the vara system are present. 298

296
Ibid.: bhikkhu paneva uddissa rj v rjabhoggo v brhmao v gahapatiko v dtena
cvaracetpanna pahieyya.
297
See, e.g., DN 1: yath v paneke bhonto samaabrhma saddhdeyyni bhojanni
bhujitv te evarpa dteyyapahiagamannuyoga anuyutt viharanti, seyyathida raa,
rjamahmattna, khattiyna, brhmana, gahapatikna, kumrna idha gaccha,
amutrgaccha, ida hara, amutra ida harti iti v iti evarp dteyyapahiagamannuyog
paivirato samao gotamo ti iti v hi, bhikkhave, puthujjano tathgatassa vaa vadamno
vadeyya.
298
Bronkhorst, Buddhism in the Shadow of Brahmanism, 34-5. Chakravarti notes that the
division of brhmaa, khattiya, vessa, and sudda, is associated most often with situations in which
the Buddha converses with a brhmaaSocial Dimensions, 98; cited by Bronkhorst, Buddhism in
the Shadow of Brahmanism, 35.

131
 Indeed, we know that the category gahapati cannot have been mutually

exclusive of khattiya and brhmaa because it is often used in the compound

brhmaagahapatika to refer to Brahman householders. 299 This compound is often

found in this sense in encounter dialogs in which the Buddha stays at a Brahman

village (brhmaagma) and is visited by the inhabitants of the village, who are

referred to as brhmaagahapatik. 300 As Chakravarti notes, this pairing of

brhmaa and gahapati in compound is unique; khattiya is not used in compound

with gahapati, that is, to refer to khattiya householders. She argues further that the

use of the term gahapati to describe Brahmans, given that it is found in the context of

Brahman villages, appears to [refer] to brhmaas based on land in villages which

were probably inhabited almost entirely by brhmaas. 301 In light of the argument I

have been presenting in Part II, I would go a step further and submit that the

Brahman householders of Brahman villages in fact represent the very ideal the

Dharma Stras construct and refer to as the ghastha-rama. As already noted,

brhmaagahapatik are found frequently in encounter dialogs, in which the

Buddha debates with Brahmans whose arguments clearly advance the tenets of the

new Brahmanism. In fact, in one particular encounter dialog, the debate in which

the Buddha engages with his Brahman interlocutor is precisely over the relative value

299
Bronkhorst cites Widmer as saying that this compound does not always mean Brahmans
and householders, but is sometimes simply a dvandvaBuddhism in the Shadow of Brahmanism, 34 n.
16.
300
DN 4, 5; MN 41, 42, 60, 95; SN 1.4.2.8, 5.11.1.7; AN 3.64, 5.30, 6.42, 8.86; Ud. 7.9.
301
Chakravarti, Social Dimensions, 72-3.

132
of the householder and the renunciate, with the Brahman asserting that only the way

of the householder (gahaha) is the skillful dharma (dhamma kusala). 302 It is

then no surprise that we would find these people, whose own texts advance a

householder ideal, clearly demarcated as householders.

But why would the Buddhist texts prefer the term gahapati, with its more

honorific implication of dominion over a household, rather than mere residence

therein, over gahaha? Bronkhorst argues that the exalted nature of the gahapati

may reflect a propagandistic tendency of the texts to depict the Buddha as being in

interaction with important people rather than with the proletariat, 303 and while there

may be some truth to this, I think there is a simpler explanation. The early Buddhist

texts refer to gahapatisto heads of households, rather than to people who dwell in

houses as suchmost frequently for the very same reason that they refer to other

prominent members of society, such as khattiyas, brhmaas, rjs, and even

courtesans (gaik)because they were the only people who could act as free agents

in ancient Indian society. Although the early Buddhist texts clearly evince a rejection

of the vara system, as Oldenberg argued over a century ago, 304 it is a mistake to see

the Buddha (or, for that matter, those who constructed him as a literary character in

the early Buddhist stras) as a great social reformer. There is in fact little evidence of

302
MN 99. Interestingly, the word used here for householder is gahaha, but this is to be
expected, since it is put into the mouth of the Buddhas Brahman interlocutor, a student (mnava)
named Subha. This sutta does not refer to brhmaagahapatik, but it is not set in a brhmaagma
and does not involve large numbers of Brahmans visiting the Buddha, only Subha.
303
Bronkhorst, Buddhism in the Shadow of Brahmanism, 35.
304
See p. 11f. above.

133
any substantive critique of prevailing social structures, with their concomitant

inequalities, in the early Buddhist texts, and in fact a great deal of evidence that social

hierarchy and inequality were accepted as ordinary parts of life. It is true that vara

was criticized, but this is a matter of rejecting a novel social schema being introduced

by the new Brahmanism, rather than overthrowing a preexisting system of social

inequality. Social inequality at the time when the early Buddhist texts were written

was encoded instead, it seems, in terms like gahapati that denoted a patriarchal

household structure in which a single adult male controlled not only societys

smallest units of land and wealth, but also most likely a fair number of other human

beings.

Insofar as the early Buddhists accepted such social inequality as normal, it is

unnecessary to appeal to a model of propaganda to explain the prevalence of

socially important individuals such as gahapatis in the early texts. Only people who

had the capacity to act as free agents in a society would be relevant in most social

interactions that would take place with the Buddha or his saghano matter whether

that interaction might involve joining the sagha, taking refuge and becoming an

upsaka, giving dna, or engaging in debate or even a simple conversation. In

addition, the Buddhist tradition has held consistently that the Buddha was born into a

khattiya lineageand although later tradition may have exaggerated or otherwise

distorted what exactly that would have meant, 305 we can be confident that it was

305
Compare, for example, the simple account of the Buddhas departure from home in the
Ariyapariyesana Sutta (MN 26) with the elaborate account of the Bodhisattvas resplendent life before
leaving the palace in the Buddhacarita (esp. ch. 2).

134
considered a high birth. Chakravarti has found through an exhaustive study of

references to the social origins of the members of the sagha and lay followers in the

Pali Canon that, at least as represented in the texts, the same was true of early

Buddhism as a whole. The ordained are dominated by Brahmans, khattiyas, and other

people of high birth, 306 while the laity are dominated by Brahmans, gahapatis,

khattiyas, and other people of high birth. 307 Members of lower classes (barbers,

potters, fishermen, etc.) are represented, but in far lower numbers. 308 Although it is

difficult to know for certain, it is very well possible that early Buddhism, at least by

the time that the bulk of the early stras were written, was an elitist movement that

established an economy of merit through the ritualized transfer of goods between

members of well-to-do families who retained ownership over the resources they

controlled and other members of the same families who formally renounced

ownership over those resources, but still continued to enjoy them. 309 In other words,

the exchange relationship between monks and laypeople in the society represented by

the early Buddhist texts may be even more restricted than is commonly thought. It
306
Chakravarti, Social Dimensions, 124.
307
Ibid., 132.
308
Ibid., 125.
309
An interesting parallel is offered by the early Franciscan movement in Italy. Kenneth Wolf
has argued that the poverty advocated and strived after by St. Francis of Assisi was a rich mans
poverty that both was qualitatively different from the poverty of the involuntarily poor, insofar as it
was created by a formerly rich person as an intentional disavowal of riches, and diminished the value
of involuntary poverty by valorizing a poverty that could effectively only be attained by the rich. As
Wolf writes, one could argue that it was his success in taking poverty as a virtue away from the
involuntary poor and giving it, in a newly spiritualized form, to the rich that secured for Francis the
respect and veneration of guilty burghers who had the resources and the influence to transform him
overnight into an alter Christus and his followers into a powerful orderKenneth Wolf, The Poverty
of Riches: St. Francis of Assisi Reconsidered (Oxford: Oxford University Press, 2003), 89.

135
would have consisted, I am suggesting, of a fairly narrow reciprocal relationship in

which certain members of well-to-do families agreed to renounce their birthright to

become lords over landed estates, so as to become a field of merit, while in

exchange for access to this field of merit other members of the same families

supported them with the resources drawn from their own landed estates. 310

In such a system, gahapatis would clearly represent a key social category, for

two reasons. First, they are the most basic free individuals with whom the sagha

could interact. Although any member of a household could technically interact with

the sagha on some level, the ultimate disposal of a households resources would

have been determined by the gahapati. 311 Second, as the most basic free individual

310
Richard Cohen has provided an insightful analysis of exchange relationships at Aja
using the categories restricted exchange and generalized exchange that were coined by Ivan
StrenskiRichard S. Cohen, Nga, Yaki, Buddha: Local Deities and Local Buddhism at Ajanta,
History of Religions 37, no. 4 (May 1998): 360-400. Cohen too emphasizes the importance of
restricted exchange in localizing Buddhism at Ajain part by declining to view merit as
somehow less real than material goods, which would imply that ordinary gifts of dna to the sagha
involve a deferral of reward, and thus fall under the category of generalized exchange. None of this
is to deny that generalized exchange is also involved hereclearly it is, insofar as monks renounce
ownership of their inheritance in exchange for the deferred repayment in the form of goods from
(possibly) another well-to-do family, and as such, this system of exchange would have buil[t] up
social solidarity (p. 366) between the well-to-do families of a particular local community. But at the
same time, insofar as the exchange is primarily between a finite group of well-to-do families and their
own sons who agree to renounce their right to ownership, it is a restricted exchange that localizes
Buddhism in a particular place by allowing the elite families of that place to gain access to the merit
generated through the institution of the sagha while still effectively retaining control of the resources
at their disposal, and thus dominance in the broader economy. This would represent a form of
localization that is in fact quite similar to that involved with the placation of ngas at Aja: Just as
Buddhists tend to nga shrines in exchange for the ngas benevolence and protection, and ngas
allow the existence of the monastery in exchange for Buddhists giving them offerings, so too do elite
families subsidize the sagha in exchange for effectively constituting the sagha (because they
populate it with their own sons), and the sons of elite families agree to join the sagha in exchange for
access to the property they renounce in the process, in the form of dna.
311
Needless to say, even in a strongly patriarchal society, women, children, and other
dependents are not magically deprived of their agency; they can and often do exert agency in various
under the radar ways, some of which act to subvert the patriarchal power structure. While that

136
of lay society, the gahapati represents the mirror-image of the bhikkhu. The Vinaya

rules for entrance into the sagha prevent the ordination of people who lack the social

authority to act on their own behalfwhether it be because they have not reached the

age of majority, lack their parents permission, are a slave, are under penalty of the

law, are in debt, or are in service to the king. 312 A bhikkhu and a gahapati are flip

sides of the same coinboth are free agents, one of which has used his agency to

take control of a household, the other of which has used his agency to renounce that

prerogative.

In this light, the ghastha of the rama systemor more properly speaking,

of the Dharma Stras writ largeis, in spite of the name that refers more generally to

being situated (stha) in a house, clearly the same thing as the gahapati of the

Buddhist textsthat is, a patriarchal head of household, not simply a human being

who happens to live in a house. The Dharma Stras are not directed toward

describing the proper life-cycle of every person or even most people in society; rather,

it is directed specifically toward males of high birth who, after undergoing a period of

study under the tutelage of Vedic Brahmans, will, upon attaining majority, take a wife,

kindle their own sacred fires, and become the heads of their own households. This is

clearly a vision for the proper life of a gahapati, and most likely a vision that was

carried out by a subset of gahapatis who are referred to as brhmaagahapatik in

certainly was the case in ancient India, most (though not all) early Buddhist texts do not seem
interested in recording such acts of subversion, much less encouraging them.
312
Vin. Mv. 1.27-34, 36, 41.

137
the early Buddhist stras. 313 The choice in the Dharma Stras to use the word

ghastha instead of ghapati, I would argue, reflects the categorys rhetorical role in

the rama. Just as the gahaha is used on occasion in the Pali Canon as a specific

counterpoint to one who has gone forth (pabbajita), so the use of the word ghastha

emphasizes the distinction between it and the other three ramasnamely, that it

involves living in a house. That the ghastha would not only live in a house, but rule

over it, is understood from the context.

II.3.3 Parivrjaka/Paribbjaka

The gahapati, it goes without saying, is constructed in the early Buddhist texts

as the total other of the bhikkhu. As already mentioned, the gahapati is precisely the

same thing as the bhikkhua socially free agentexcept that he has chosen to

exercise his agency by becoming the head of a household instead of going forth from

the home life into homelessness (agrasm anagriya pabbajita) as does the

bhikkhu. The category of the paribbjaka, howeverwhich is simply the Pali form

of parivrjaka, the term used by Baudhyana and Vasiha to refer to the fourth

rama 314occurs frequently in the Pali Canon and is constructed as a proximate

313
The Dharmastric vision, of course, calls on all gahapatis to live by the same life cycle,
with certain variations according to whether they could be categorized as brhmaa, katriya, or
vaiya. There is little evidence, however, that this vision of life was carried out by anyone outside of
the Vedic Brahman lineages, i.e., the proponents of the new Brahmanism, at the time the early
Buddhist texts were written, at least in northeastern India.
314
BDhS 2.11.12, VDhS 7.1. pastamba (2.21.1) refers to this rama as the mauna (though
later as parivrjaDhS 2.21.7), and Gautama (3.2) refers to it, interestingly enough, as the bhiku. I
believe that the latters use of the word bhiku lends credence to the argument that Buddhist bhikkhus
were, colloquially speaking, paribbjakas; the categories, in other words, appear to have been
interchangeable.

138
other to the bhikkhu. By this, I mean that we can infer that Buddhist monks would

have been considered, within colloquial discourse, as paribbjakas, but that the early

Buddhist texts seek rhetorically to distance the category bhikkhu from the category

paribbjaka; that is, they reserve the latter category almost exclusively for non-

Buddhist paribbjakas, even though Buddhist monks clearly would have been

considered paribbjakas as well.

That bhikkhus within the early Buddhist tradition would have been considered

paribbjakas/parivrjakas is clear from the way in which the category parivrjaka is

defined in the Dharma Stras. Gautama, who as already mentioned, refers to this

rama with the term bhiku instead of parivrjaka, defines the category as follows:

A beggar (bhiku) is without possessions, keeps his semen up (i.e., is celibate),

(and) stays in one place during the rains. He should go to the village for the
purpose of alms. He should go last, without turning back, without giving a
blessing. He is controlled with respect to speech, eyes, and actions. He
should wear a cloth to conceal the private parts. Some (say he should wear) a
cast-off (cloth) after washing it. He should not take a limb from a plant or tree
that has not fallen (on its own). He should not stay a second night in a village
out of season. He (should be) shaven or wear a tuft of hair. He should avoid
the destruction of seed. He (treats) beings equally, whether in violence or in
kindness. He is not enterprising. 315

Baudhyanas description, which refers to this rama as that of the parivrjaka, is

quite similar, making reference to wandering except for a rains retreat when the

person in question should stay in one place, giving up all possessions, shaving the

315
GDhS 3.11-25: anicayo bhiku. rdhvaret. dhruvalo varsu. bhikrth grmam iyt.
jaghanyam anivtta caret. nivtt. vk cakukarmasayata. kaupnc chdanrtha vso
bibhyt. praham eke nirijya. nviprayuktamoadhivanaspatnm agam upda dta. na dvitym
apartu rtri grme vaset. mua ikh v. varjayed bjavadham. samo bhteu hisnugrahayo.
anrambh.

139
head, and subsisting off of alms. 316 Vasiha also uses the term parivrjaka and

describes him as shaving his head, begging for food, and lacking a fixed residence. 317

Finally, pastamba, who uses the term parivrja, says that the person who chooses

this rama wanders, begs for food, and wears discarded clothes, although, like

Vasiha, he makes no mention of a rains retreat. 318 Clearly, the Buddhist bhikkhu, at

least as he was ideally constructed in the early Buddhist texts, falls into this category,

especially as it is described by Gautama and Baudhyana. 319

Although, as I will show, the early Buddhist texts tend to distance the bhikkhu

from the category of paribbjaka, it can nonetheless easily be inferred from the same

texts that bhikkhus would in common parlance have ordinarily been understood as

belonging to that category. One of the most common ways in which these early texts

refer to paribbjakas is through the stock phrase aatitthiy paribbjak

wanderers of other sects. 320 This phrase is used, much as the compound

samaabrhma often is, to refer to non-Buddhist vocational practitioners

collectively as a group. Indeed, on occasion it is even grouped together with samaas

316
BDhS 2.11.16-20. Baudhyana says that the parivrjaka shaves his head except for the
topknot, while Gautama says that he may either shave his head or have a topknot.
317
VDhS 10.6-13.
318
DhS 2.22.7-11.
319
For a more detailed comparison of sources on the characteristics of ancient Indian
parivrjakas, see Olivelle, Buddhist Monachism, 11-24.
320
MN 11, 13, 59, 150; SN 2.1.3.4, 2.1.7.10, 3.1.9.4, 4.1.8.8, 4.1.15.7, 4.2.2.9, 4.2.2.10,
5.1.1.5, 5.1.5.1-8, 5.2.6.2-4, 5.10.2.1-2; AN 2.33-42, 3.69, 3.94, 7.42-43, 8.83, 9.1, 10.58; Ud. 2.4, 4.8,
6.10. Ud. 4.8 is fairly unique in that it the wanderers of other sects are actual characters within the
plot of the sutta, who in fact conspire to frame a sagha of Buddhist bhikkhus for having sex with a
female wanderer and then murdering her!

140
and brhmaas in a single compound used to refer collectively to non-Buddhists. 321

Most often, aatitthiy paribbjak are referred to by the Buddha while teaching to

distinguish their views and practices from his own, or even more specifically in the

context of telling his monks how to respond when members of other sects ask them

about his teaching. Thus, in MN 11, which can be taken as exemplary, the Buddha

says to his monks,

It is possible, monks, that wanderers of other sects might say thus: But what
is your confidence, what the strength, by which you speak thus? Speaking
thus, monks, wanderers of other sects should be answered thus. 322

What follows then is the Buddhas teaching, which the monks can then use to answer

the questions of the wanderers of other sects. Presumably, texts such as this one were

intended to train monks in how to represent themselves, as members of the sagha

and followers of the Buddha, to members of other sects.

Clearly, much as in the case of references to samaabrhma studied above,

these references to aatitthiy paribbjak attempt to construct a Buddhist identity

by clearly demarcating distinctions in doctrine and practice between the Buddhists

themselves and the collective paribbjaka other. This demarcation is reinforced by

the fact that the category paribbjaka is reserved in the early Buddhist texts for that

other; to my knowledge, the Buddha and his monks are never referred to directly in

the Pali Canon as paribbjakas. Nevertheless, by the very phrasing wanderers of

321
atitthiyasamaabrhmaaparibbjakna (SN 2.2.11),
nntitthiyasamaabrhmaaparibbjak (Ud. 6.4-6), aatitthiy samaabrhmaaparibbjak (SN
4.8.7).
322
hna kho paneta, bhikkhave, vijjati ya aatitthiy paribbjak eva vadeyyu ko
panyasmantna assso, ki bala, yena tumhe yasmanto eva vadetha evavdino, bhikkhave,
aatitthiy paribbjak evam assu vacany.

141
other sects, it is clearly implied that this sect, that of the followers of the Buddha, is

itself a sect of wanderers. This conclusion is reinforced by the fact that, on a couple

of occasions (AN 3.144, 11.10), the Buddha is said to have stayed at a

paribbjakrmaa pleasure park for wanderers. Insofar, then, as paribbjaka is

a category to which the Buddhist bhikkhu should, according to ordinary parlance,

belong, but against which it constructed itself, it can be understood as serving as a

proximate other in the early Buddhist texts.

How does this proximate other relate to the category of the Brahman?

According to Bronkhorst, they are mutually exclusive. In explaining his theory of the

two sources of Indian asceticism in Greater Magadha, Bronkhorst writes that the

term paribbjaka in the Pli Buddhist canonrefers throughout to non-Vedic

ascetics. 323 This generalization is not entirely accurate. To begin with, there exists

in the Pali Canon, although only in a single instance, the compound

brhmaaparibbjaka. This compound is not a dvandva, since it refers to an

individual who approaches the Buddha and asks him about dhamma, and thus must be

translated as a Brahman who is a wanderer. 324 Similarly, there is a humorous

udna in which a husband and wife are identified both, on the one hand, as

paribbjaka and paribbjik, and, on the other, as brhmaa and mavik. 325 There

323
Bronkhorst, Greater Magadha, 91.
324
AN 3.55.
325
Ud. 6.2. The word used to describe the woman as a Brahman is actually
daharamavik, which could approapriately be translated as young (dahara) Brahman maiden
(mavik). The word mava is routinely used in the Pali Canon to refer to young Brahman males,
usually students, and the choice of the word mavik here instead of brhma may emphasize the

142
is even an encounter dialog in which the Brahman student Assalyana is specifically

chosen by his teachers to debate with the Buddha about the vara system because he

has practiced wandering (paribbjakan). 326

Finally, there is a sutta in the Aguttara Nikya (4.185) in which the Buddhas

interlocutors are referred to exclusively as paribbjakas, but appear from the context

to be Brahmans as well. This text is entitled the Brhmaasacca Sutta, after the topic

of the dialog, which is on four so-called Brahman truths. The Buddha approaches

these paribbjakas, asks them what they are talking about, and when they tell him

they are talking about these Brahman truths, he describes four truths that a

Brahman knowswhich are in fact his own teachings. The fact that the

paribbjakas in this sutta are discussing Brahman truths could be taken to imply

that they are themselves Brahmans, and this seems to be confirmed by the two

parallels to this sutta in the Chinese Canon. These parallels are both found in

translations of the Sayuktgama, the Z hn jng (, T. 99) and the Biy

z hn jng (, T. 100). In these versions, the Buddhas interlocutors

girls youth. As stated, this udna is clearly intended to be humorous. The girl is pregnant and about
to give birth, and she asks her husband (their relationship, besides being implied by the pregnancy, is
indicated by the use of the word pajpatichief wifeto describe the girl) to fetch oil to use in the
delivery. The Brahman does so by going to a storehouse where the king gives out free butter and oil to
samaas and brhmaas. He is only allowed to take as much as he can drink on the spot, so he drinks
an enormous amount, planning to regurgitate it after returning to his wife, but, predictably, gets a
terrible stomache ache as a result.
326
MN 93: carita kho pana bhot assalyanena paribbjakan ti.

143
not only are talking about Brahman truths ([]), but are specifically

referred to as Brahmans (). 327

Even if we ignore these Chinese parallels, however, the Pali version by itself

could be taken, since it involves a host of actors (including both the Buddha and the

paribbjakas) not otherwise identified as Brahmans discussing the category

Brahman, as further evidence of my broader thesis in this dissertation that the

category of Brahman was not simply appropriated by the Buddhists and redefined

as a reaction to or marketing strategy against the real Brahmans, but was an open

category freely contested not only by Buddhists, but also presumably by other groups

beside the proponents of the new Brahmanism. This interpretation depends in part

on the relationship of the Pali version to the Chinese versionsthat is, whether the

Chinese versions added the identification of the interlocutors as Brahmans to make

the Buddhas teaching to them more meaningful as a metaphorical appropriation of

the word Brahman, or the Pali version simply for some unknown reason omits an

original identification of the interlocutors as Brahmans. In either case, however, the

narrower point I am making here still stands: There is not always a strict distinction

between the categories paribbjaka and brhmaa in the early Buddhist texts.

327
The parallels in question are stras 206 of the Biy z hn jng (BZA) and 972 of the Z
hn jng (ZA). The BZA version refers to the interlocutors initially as (people of other paths,
i.e., non-Buddhists), but then subsequently simply as (plumn), which is the standard
Chinese transcription of brhmaa. The ZA version refers to them throughout as .
(lit., to exit the house) is ordinarily used (even in modern Chinese) as a verb meaning to go forth;
it is, in other words, a translation of the Pali verb pa-bbaj (Skt. pra-vraj). I suspect that here it is
serving as a translation for the very similar word parivrjaka, which would imply an original
compound brhmaaparivrjaka, just as in AN 3.55 cited above.

144
 Indeed, although he does not appear to be aware of the mixing of the

categories paribbjaka and brhmaa found in the early Buddhist texts, Bronkhorst

does acknowledge a mixing of the two categories in another sourcethe early Jain

texts. He notes that, [i]nterestingly, the Jaina canon also uses the term parivrjaka

(Ardha-Mgadh parivvyaga/-ya) to refer to Brahmins on some occasions, although

he explains this evidence away by arguing that [i]t is clear that this confused

terminology dates from a time when earlier distinctions had become blurred. 328 As I

believe the evidence presented here has shown, it is not entirely clear that the

distinction Bronkhorst appeals to ever existed in any strict way. The early Buddhist

texts, for one, do provide evidence of a concept of Brahman wanderers. In addition,

while two of the Dharma Stras (Gautama and Baudhyana) describe the parivrjaka

(or bhiku in the case of the former) in terms that seem to place him entirely outside

of the Vedic tradition, the other twoincluding pastamba, which might be the

oldest of the fouruse terms to describe the parivrja(ka) that seem to maintain a

link to the Vedas. 329

Nevertheless, Bronkhorst does point to an interesting dichotomy within the

early Buddhist texts that, while perhaps not as absolute as he implies, does reveal

328
Bronkhorst, Greater Magadha, 91.
329
The relevant citations are DhS 2.22.7-17, GDhS 3.11-25, BDhS 2.11.16-26, VDhS 10.1-
29. pastamba writes that the parivrja should [speak] only in recitation (of the Veda) (2.22.10:
svdhyya evotsjamno vca), and Vasiha writes of the parivrjaka, He should abandon all rites;
the Veda alone he should not abandon. / From the abandoning of the Veda (one becomes a) dra;
therefore, he should not abandon the Veda // (10.4: sanyaset sarvakarmi vedam eka na
sanyaset / vedasanyasanc chdras tasmd veda na sanyaset //). Gautama and Baudhyana,
who, as already discussed, reject the validity of the rama system, make no effort to tie the
parivrjaka/bhiku to the Vedic tradition, and in fact explicitly say that they do not perform ritual
(GDhS 3.25, BDhS 2.11.26).

145
something important about the role that Brahmans play in the imaginary created by

the authors of the early Buddhist texts. Many of the references to paribbjakas in the

Pali Canon are found in suttas in which the Buddha either approaches or is

approached by a paribbjaka and then answers the paribbjakas questions, which

usually pertain to the Buddhas teaching and how it differs from those of other

teachers. These suttas, 330 which I will call dialogs with wanderers, can be

considered to form a genre, much like the encounter dialogs in which the Buddha

encounters and debates with Brahmans. As implied by the evidence presented above,

this genre can be considered to overlap extensively with the teachings on wrong

views, insofar as the latter, as I pointed out, frequently are found in discussions

between the Buddha and paribbjakas, and only rarely in what can be considered

encounter dialogsthat is, debates between the Buddha and interlocutors identified

specifically as Brahmans. Discussions about a whole host of doctrinal issues,

including, in particular, the nature and existence of the self, are not generally

speaking found in encounter dialogs with Brahmans, which instead are usually

debates between the Buddha and his Brahmanical interlocutors on the specific

Brahmanical claim to superiority on the basis of vara and their knowledge of the

Veda and associated ritual.

330
DN 1, 7 (which is also embedded in 6), 9, 25; MN 51, 71-80, 136; SN 2.1.2.8, 3.1.9.4 (here
Anuruddha, rather than the Buddha, speaks to paribbjakas), 3.12.1-10, 3.12.16-20, 3.12.51-54,
4.2.3.1, 4.4.1 (here Sriputta speaks with a paribbjaka), 4.5.1 (again, with Sriputta), 4.10.7 (here
Mahmoggallna speaks with the paribbjaka Vacchagotta), 4.10.8-11, 5.1.1.10, 5.2.1.6; AN 3.55,
3.58, 3.65, 3.72, 4.30, 4.100, 4.185, 6.47, 9.7, 9.8, 10.65 (with Sriputta), 10.95-96, 10.116, 11.10; Sn.
3.6.

146
 In Greater Magadha, Bronkhorst uses this distinction as part of his argument

that a whole host of doctrinal concerns 331 belonged to the religious matrix of Greater

Magadha and were foreign to the Vedic tradition of the Brahmans, who only later

adopted these ideas as they moved east into Greater Magadha. While there can be no

doubt that the Brahmanical movement absorbed new ideas as it spread throughout

India, Bronkhorsts particular argument is predicated on a fundamental dichotomy

between Brahmanical and non-Brahmanical that is never clearly defined, except

implicitly by geography. Since, as we have seen, there are examples of Brahman

paribbjakas in the Pali Canon, I would argue that the lesson to be taken from the

general dissociation between encounter dialogs with Brahmans on the one hand,

and dialogs with wanderers and teachings on wrong views on the other, is not

that there is an absolute dichotomy between the Brahmans and a putative non-

Bramanical other (such as the Greater Magadha tradition identified by

Bronkhorst)a dichotomy that necessarily leads to appeals to confusion when it is

violated. Rather, I would argue, the lesson to be taken is that the encounter dialogs

form a distinct genre, separate from more generic, often almost encyclopedic

discussions of wrong views that often involve paribbjakas as interlocutors and

appeal to paribbjakas and/or samaabrhma as a foil against which to construct

Buddhist identity. The reason for this distinction is that the encounter dialogs were

designed to address a specific movement, what I am calling here, following

331
These doctrinal concerns included, most importantly, ideas about the self, karma, and
rebirth that in the early Buddhist texts are treated in teachings on wrong views, which, as I have
pointed out, overlap with dialogs with wanderers and not encounter dialogs involving Brahmans.

147
Bronkhorst, the new Brahmanism, that made specific claims and was qualitatively

different from the other movements against which the early Buddhist texts argued

(e.g., the Jains and jvikas) insofar as it was not a paribbjaka movement, but rather

a householder movement, and in fact a householder-supremacist movement! The

issue here is not a strict dichotomy between Brahmanical and non-Brahmanical

as I have argued, nobody owned the category of brhmaa, and there is evidence

for dual Brahman-parivrjaka identitiesthe issue is the existence of a specific

movement, evinced both by the Dharma Stras and by the encounter dialogs found

among the early Buddhist texts, that arrogated to itself the identity Brahman;

promoted Vedic learning and a vision of society with themselves at the top through

the ideology of vara; and valorized the householder over all other ways of life,

including the parivrjaka.

II.3.4 Brahmacarya and the Brahmacrin

While the category parivrjaka is one that we would expect the Buddhist

bhikkhu to fall under, but instead find it distancing itself from, the category of

brahmacrin is one that, given its association with Vedic studentship, we might

expect to find the early Buddhist texts distancing themselves from, but instead find

them embracing. While the categories gahapati and paribbjaka serve in the early

Buddhist texts as the total other and proximate other, respectively, to the bhikkhu,

the brahmacr serves, I will argue, as a rhetorical self. Moreover, I will argue that

the category brahmacrs role as a rhetorical self is readily explicable given the

148
historical usage of the word and the nature of the Buddhist bhikkhus life within the

sagha, without recourse to any theories of reaction or marketing.

Within a reaction model of Buddhisms relationship to Brahmanism,

brahmacariya would be, a priori, an example of a Brahmanical term that was

appropriated by early Buddhism and given a Buddhist sense; indeed, Norman lists

it as one of the [t]erms taken over by the Buddha but used with new senses. 332 The

new sense supposedly given to this particular term is celibacy. Brahmacariya

does, in fact, generally take this seemingly narrow and specialized meaning in the

Pali Canon, and accordingly, the negative form abrahmacariya takes the meaning of

non-celibacy. This narrow meaning is clear in DN 24, in which the Buddha quotes

a certain naked ascetic (acelo) embarking upon his religious career by making seven

vows, one of which is as follows: As long as I live, I will be a brahmacr; I will

not indulge in sexual intercourse. 333 The use of brahmacr to refer specifically to

celibacy is also expressed in the common formula, Having abandoned

abrahmacariya, he becomes a brahmacr, living far from, abstaining from the

village custom of sexual intercourse. 334 Elsewhere, the reference to sexual

intercourse is not explicit, but is clearly implied. For example, on several occasions,

people are criticized for being abrahmacri.e., not celibateeven though they

332
Norman, Theravda Buddhism and Brahmanical Hinduism, 194-195.
333
yvajva brahmacr assa, na methuna dhamma paiseveyya.
334
DN 2: abrahmacariya pahya brahmacr hoti rcr virato methun gmadhamm.
The same formula, or some variation thereof, is also found at DN 1; MN 27, 38, 51, 94, 101, 112; AN
3.71, 4.198, 5.180, 6.44, 8.41, 10.75, 10.99.

149
have vowed to be brahmacr. 335 Frequently, we also find the compound

accantabrahmacr, which means perpetual brahmacr and thus implies a lifelong

commitment to abstaining from sexual intercourse. 336 Numerous other instances of

the word brahmacr or abrahmacr can be understood from context to refer to

celibacy or non-celibacy, respectively. 337

It would be a mistake, however, to think that the words brahmacr and

brahmacariya in the Pali Canon can be reduced to a single, narrow meaning of

celibacy. By far, the most common use of the word brahmacr is in the compound

sabrahmacr, which, roughly speaking, can be translated as co-religionist. 338

Although this word certainly implies celibacy, this implication is somewhat trivial, as

the emphasis is on a relationship between human individuals based on a common

pursuit that has to do with much more than simple abstinence from sexual intercourse.

Moreover, the relationship implied is generally quite close, as of companions who

know one another personally; what we are dealing with here is not an imagined

community of ancient Indian Promise Keepers linked merely by putative celibacy,

but members of local, face-to-face communities, linked by common teachers,

335
SN 4.1.19.4; AN 3.13, 3.27, 3.114, 4.243, 8.19, 8.20; Ud. 5.5; It. 2.2.11: abrahmacr
brahmacripaio.
336
DN 21; MN 37; SN 3.1.1.4; AN 3.144-6, 7.61, 11.10.
337
DN 29; MN 8, 73, 81, 84, 99; SN 4.1.19.4; AN 3.13, 3.27, 3.62, 3.114, 4.243, 5.286, 5.293;
Ud. 4.8.
338
DN 16, 29, 33, 34; MN 2, 5, 6, 15, 16, 18, 24, 26, 31, 36, 48, 61, 65, 67, 69, 77, 85-6, 88,
90, 100, 104, 124, 128, 133, 138, 141; SN 1.4.3.2, 2.5.8, 3.1.9.3, 3.1.9.8, 4.1.12.3, 4.1.12.4, 5.3.2.3-4;
AN 2.33-42, 3.40, 4.87, 4.97, 4.111, 4.122, 5.21-2, 5.26, 5.31, 5.53, 5.65-6, 5.76, 5.81-7, 5.104-5,
5.135, 5.163-4, 5.166, 5.167, 5.205, 5.211, 5.232, 5.249, 6.11-2, 6.44-5, 6.54, 6.58, 6.60, 7.1-2, 7.23,
7.73, 8.2-4, 8.13, 8.62, 8.78, 9.4, 9.11, 9.71, 10.11, 10.14, 10.17-8, 10.23, 10.44, 10.48, 10.50, 10.71,
10.83, 10.87-8, 10.96, 10.115-6, 10.172, 11.6, 11.14; Ud. 1.10, 4.9; Sn. 4.16; Ther. 4.3, 6.3, 18.

150
practices, and systems of belief that go far beyond celibacyreal or imagined

although celibacy would have certainly played a part. An excellect example of the

use of the word sabrahmacr is with reference to Bharau Klma, who is

mentioned in a sutta (AN 3.127) as a former sabrahmacr of the Buddha himself

(bhagavato purasabrahmacr). The Buddha stays at Bharau Klmas

hermitage (assama) while in Kapilavatthu and (in an exchange that is somewhat

obscure) slights Klma in the presence of Mahnma, a prominent Sakyan, thus

forcing the former to leave the city. According to the commentary, Bharau Klma

is called a former sabrahmacr of the Buddha because, prior to the Buddhas

Awakening, they had both been disciples of ra Klma. 339 The use of this word in

such a context clearly implies something quite similar to the usage of brahmacarya in

Brahmanical texts: living a celibate life of study under the tutelage of a teacher.

Indeed, a passage from the Dgha Nikya emphasizes the importance of

having a good teacher to perfecting brahmacariya. After hearing about the disarray

among the Jain monks (Nigahas) after the death of their leader, the Nigaha

Ntaputta (Mahvra), the Buddha addresses a novice named Cunda about what will

keep the Buddhist sagha strong and in accord. In doing so, he says,

Moreover, Cunda, the brahmacariya is endowed with these qualities: The

teacher is not a recognized elder long-gone-forth, getting-on, attained to old-
age. Thus the brahmacariya is imperfect in this case. But since, Cunda, the
brahmacariya is endowed with these qualitiesthe teacher is a recognized

339
Cited by Alexander Wynne, The Origin of Buddhist Meditation (London: Routledge, 2007),
13. For an account of the Bodhisattvas discipleship under first ra Klma and then Uddaka
Rmaputta, see, e.g., the Ariyapariyesana Sutta (MN 26).

151
 elder long-gone-forth, getting-on, attained to old-agethus the brahmacariya
is perfect in this case. 340

The Buddha then goes on to explain, in the same terms, the importance of senior

monk disciples (ther bhikkh svak), monk disciples of middle standing (majjhim

bhikkh svak), junior monk disciples (nav bhikkh svak), senior nun disciples

(ther bhikkhuniyo svik), nun disciples of middle standing (majjhim bhikkhuniyo

svik), junior nun disciples (nav bhikkhuniyo svik), and various types of lay

disciples. 341 Finally, he says that brahmacariya itself must be prosperous, opulent,

extensive, public, well-spread, well-proclaimed by gods and mortals, and attained to

the highest gain and fame. 342 The brahmacariya being described in these terms is

clearly not only celibacy, but the entire enterprise the author of this text envisions the

Buddhist sagha being engaged in.

It is not entirely surprising that early Buddhists would articulate what it was

they were doing using the term brahmacariya (Skt. brahmacarya). The Vinaya

provides for a monk to be assigned both a preceptor (upajjhya) and a teacher

(cariya) upon ordination. 343 In addition, the Buddhist scriptures repeatedly cast the

340
DN 6: etehi cepi, cunda, agehi samanngata brahmacariya hoti, no ca kho satth hoti
thero ratta cirapabbajito addhagato vayoanuppatto. eva ta brahmacariya aparipra hoti
tenagena. yato ca kho, cunda, etehi ceva agehi samanngata brahmacariya hoti, satth ca hoti
thero ratta cirapabbajito addhagato vayoanuppatto. eva ta brahmacariya paripra hoti
tenagena.
341
Interestingly, the permutations for listing various types of lay disciples includes both those
who are brahmacr and those who are abrahmacrwhich terms must in this case refer to celibacy
or non-celibacy.
342
iddhaceva phtaca vitthrika bhujaa puthubhta yva devamanussehi
suppaksita, lbhaggayasaggappatta.
343
Vin. 4.1.25.6, 4.1.32.1.

152
Buddha in the role of teacher (P. satth, Skt. st) and, as already discussed above,

use the word ssana/sana (teaching or instruction) to describe in the abstract

what members of the sagha submit themselves to. Given this extensive rhetoric of

education, together with the requirement of celibacy, the similarities to the

Bramanical institution of brahmacarya are obvious.

Still, when viewed through the lens of the Brahmanical rama system, it

might seem natural to label the early Buddhists as parivrjakas, and not as

brahmacrins, because the similarities between the early Buddhists and the ramic

descriptions of the parivrjaka are clear, and brahmacrya appears to be rooted in a

Brahmanical institution of Vedic studentship. In this light, the tendency of the early

Buddhist texts to distance themselves from the category of paribbjaka on the one

hand, and embrace the category of brahmacr on the other, could be understood

simply as a strategic move to distinguish themselves by distancing themselves from

their most proximate otherthe hodge-podge of wanderer movements from which

Buddhism aroseand identify itself with a more distant otherBrahmanismthat it

would not be in danger of being confused with, and whose prestige it could co-opt.

This would, of course, be an example of marketing, as the term is used by Bailey

and Mabbett.

I believe that an uncritical acceptance of the Brahmanical rama system has

the potential to color our perception of non-Brahmanical uses of the categories found

therein, however, in three ways. First, the very taxonomization of ways of life into

four categories itself can give the false impression that these four categories are

153
mutually exclusive. The evidence already presented from the Pali Canon would

suggest that, colloquially, this may not have been the case. Although, as already

shown, the early Buddhist texts make an effort to distance the Buddhist bhikkhu from

the category paribbjaka, they also implicitly acknowledge that the Buddhist bhikkhu

is a paribbjaka, insofar as they refer to non-Buddhist practitioners as wanderers of

other sects. Indeed, it seems difficult to deny that Buddhist monks would have been

considered parivrjakas, given that they were celibate, shaved their heads, subsisted

off of alms obtained through begging, and wandered except for a retreat of fixed

residence during the rainy season. At the same time, however, it certainly seems

plausible to locate the Buddhist monks under the category of brahmacrin, insofar as

they were celibate, underwent a ritual of initiation, and studied under the tutelage of

more senior monks and, ultimately, the Buddha himself. There is no reason to

assume that the categories parivrjaka and brahmacrin were, in colloquial speech,

mutually exclusive. While the rama system does set them up as separate categories,

we must remember that that system is contrived to serve a particular agendanamely,

the Dharmastric agenda of taxonomizing social custom.

This brings me to the second way in which we must be careful not to be

misled by the rama system. The introduction of the concept of the rama system

itself, and the inclusion of life-long brahmacrya within it, represents a particular

ideological representation of a concept with roots far more ancient than the system of

rama. In his study of the rama system, Olivelle warns against confusing the two,

writing that [t]he four ramasare not coextensive with the respective social

154
institutions classified by the system and that [t]he history of the system, therefore,

is quite distinct from the history of these institutions taken individually or collectively,

and the study of the former should not be confused with that of the latter. 344 The

rama system, we should note, did two things to the concept of brahmacarya. First,

it bifurcated temporary Vedic studentship, which was enjoined, from brahmacarya

qua rama, which, as a vocation of lifelong celibacy, was, as we have seen, viewed

with hostility by all four of the early Dharma Stras, whose authors appear to have all

been proponents of the householder-supremacist new Brahmanism. Then, with the

advent of the classical rama system, 345 brahmacarya as rama was simply

subsumed into (temporary) Vedic studentship, thus totally effacing the concept of

brahmacarya as a lifetime vocation of celibate service to a teacher.

But what can we say about brahmacarya as it was understood, even within

Brahmanical traditions, prior to its incorporation into the rama system? The

earliest usage of the word brahmacrin is found in the tenth book of the g Veda,

where its purport is unfortunately rather obscure. It occurs in the context of a hymn

(RV 10.109) calling for the return of a Brahmans abducted wife. In the course of this

brief and rather cryptic narrative, the following verse is included:

A brahmacrin lives as a servant serving intently; he becomes one limb of the

gods.

344
Olivelle, rama System, 28-9.
345
For an overview of the classical rama system, in which, beginning with Manu, the
ramas are presented not as lifelong vocations, but as temporary stages of life to be entered into in
sequence over the course of a single life, see Olivelle, rama System, ch. 5-6, 131-182.

155
 In that way, Bhaspati found his wife (who had been) taken by Soma, like the
gods (found) the sacrificial ladle (when it was taken by Soma). 346

According to Wendy Doniger OFlaherty, a parallel is being drawn here between a

human Brahman appealing to the king to return his wife and a myth in which

Bhaspatis wife is abducted by Soma. 347 It is not entirely clear whether the

brahmacr who is the subject of the verb carati is a human Brahman, the god

Bhaspati, or perhaps both. What is somewhat clear is that being a brahmacrin

involves working or serving (veviat 348) as a servant (via).

Other slightly later Vedic sahit texts make equally brief and cryptic

references to the brahmacrin or brahmacarya. The Taittirya Sahit refers to a

son of Manu named Nbhnediha, who is deprived by his father of his inheritance,

as a brahmacrin, but no further background about this appellation is provided. 349

There is one other reference to brahmacarya and the brahmacrin in the Taittirya

Sahit, in the following verse:

Even while being born a Brahman is born with three debts: with [the debts of]
brahmacarya to the is, sacrifice to the gods, [and] offspring to the ancestors.

346
RV 10.109.5: brahmacr carati veviad via sa devnm bhavaty ekam agam / tena
jym anv avindad bhaspati somena nt juhva na dev //.
347
Wendy Doniger OFlaherty, trans., The Rig Veda (London: Penguin Books, 1981), 275.
348
This somewhat unusual form is a participle of the root vi, which is a reduplicating verb of
the third gaaArthur Anthony Macdonell, A Vedic Grammar for Students (Delhi: Motilal
Banarsidass, 1993), 419.
349
TS 3.1.9.4: mnu putrbhyo dy vy bhajat s n bhndiham brahmacrya
vsanta nr abhajat. The same story is also related at AB 22.9cited by Pandurang Vaman Kane,
History of Dharmastra, vol. 2, part 1 (Poona: Bhandarkar Oriental Research Institute, 1974), 271 n.
624.

156
 On the other hand, he is without debt who has sons, sacrifices, and dwells as a
brahmacrin. 350

This is a fairly well-known verse because it became the basis of the theology of

debts in later Brahmanical orthodoxy. 351 It does not, however, give any real sense of

what brahmacarya is or what it means to dwell as a brahmacrin. 352 The Atharva

Veda makes several references to the brahmacrin, most of which are equally cryptic.

One reference, included in a hymn asking for success in throwing dice, appeals to the

speakers having lived in brahmacarya (perhaps in fulfillment of the debt spoken of

in TS 3.1.9.4?) in asking the gods for help with the dice. 353 Two other references

associate the brahmacrin, unsurprisingly, with brahman. 354 Another verse refers to

the brahmacrin of Death, for whom he procures a man (purua)perhaps in

reference to the practice of a brahmacrin begging on behalf of his teacher. 355

350
TS 6.3.10.5: j yamno vi brhmas tribhr rav jyate brahmacryeribhyo
yajna devbhya prajy pitbhya e v anr y putr yjv brahmacrivs .
351
For a discussion of the history of the theology of debts in Brahmanical thought, see
Olivelle, rama System, 46-53. See also Charles Malamoud, Thologie de la dette dans les
Brhmaa, Pururtha : Science Sociales en Asie du Sud 4 (1980): 39-62; cited in Olivelle, rama
System, 47 n. 53.
352
This passage does seem to imply that Brahmans are born as such, but even this is unclear,
since the birth referred to could be the second birth of initiation that makes one a Brahman.
353
AV 7.109.7: dev n yn nthit huv brahmacrya yd im / ak n yd babhr n
lbhe t no mrantv dre //.
354
AV 6.108.2: medh m ah pratham brhmavat brhmajtm riutm / prpt
brahmacrbhir dev nm vase huve //. (Whitney translates prapt as drunk of, thus having the
brahmacrins drink brahman, but he notes that the form is doubtful, and that the Paippalda version
has instead praihit, which would mean something more like applied.) AV 19.19.8: brhma
brahmacrbhir d akrmat t pra pr aymi va / t m viata t pr viata s va rma ca
vrma ca yachatu //.
355
AV 6.133.3: mrtyr ah brahmacr yd smi niry can bht t prua yam ya / tm
ah brhma tpas rmenyaina mkhalay sinmi.

157
 The Atharva Veda also contains, however, an entire hymn extolling the virtues

of brahmacrin, constituting what may be the earliest text to address the brahmacrin

and brahmacarya at length. Given the significance of this hymn, I will quote it here

in full:

The brahmacrin wanders impelling both heaven and earth. The gods are of
one mind with respect to him.
He holds earth and heaven; he protects the teacher with tapas. (1)

The fathers, those born of the gods, the gods one-by-oneall go to the
brahmacrin.
The gandharvas came to himthirty-three, three hundred, six thousandhe
protects all the gods with tapas. (2)

Bringing the brahmacrin near, the teacher makes him an embryo within.
He bears him in his belly for three nights; the gods approach to see him born.
(3)

This piece of fuel is the earth; the second is the sky; and the atmosphere he
fills with fuel.
The brahmacrin protects the worlds with fuel, his girdle, toil, and tapas. (4)

First born of the brahman, the brahmacr, wearing sunshine, stood up by

means of tapas.
From him were born the Brahman, the most excellent brahman, and all the
gods, along with immortality. (5)

The brahmacrin goes kindled with fuel, wearing the skin of the black
antelope, consecrated, with a long beard.
He goes at once from the lower to the higher sea; he grasps the worlds and
brings them to himself over and over again in an instant. (6)

The brahmacrin, generating the brahman, the waters, the world, Prajpati,
Paramehin, Virj,
Became an embryo in the womb of immortality, and then having become
Indra, crushed the asuras. (7)

The teacher fashioned both clouds, the wide, deep earth and heaven.
The brahmacrin protects them with tapas; with respect to him the gods are
of one mind. (8)

158
The brahmacrin at first brought this land, the earth, as almsand heaven.
Having made them into fuel, he attends upon them; all beings are dependent
on them. (9)

The one here, the other beyondthe two treasures of the Brahman are
deposited secretly behind the sky.
The brahmacrin protects them with tapas; that alone the one with knowledge
makes into brahman. (10)

The one here, the other theretwo fires from the earth come together between
these two clouds.
Upon them are fixed firm rays; the brahmacrin ascends them using tapas.
(11)

Roaring, thundering, a large ruddy, whitish penis enters into the earth.
The brahmacrin sprinkles semen on the summit, on the earth; by means of
that the four quarters live. (12)

In fire, in the sun, in the moon, in the air, in the waters, the brahmacrin
deposits fuel.
Their flames go one-by-one in the cloud; their clarified butter is man, rain,
water. (13)

The teacher (was) death, Varua, Soma, plants, milk.

The clouds were attendants; by them this sun was brought. (14)

Varua, having become the teacher, makes his own ghee at home.
Whatever he desired from Prajpati, that the brahmacrin offered as a friend
from his own self. (15)

The teacher is the brahmacrin; the brahmacrin is Prajpati.

Prajpati rules; Virj became the ruler Indra. (16)

A king guards his kingdom with brahmacarya, with tapas.

A teacher seeks a brahmacrin with brahmacarya. (17)

By brahmacarya a maiden finds a young man (to be her) husband.

By brahmacarya an ox (or a) horse seeks to go to grass. (18)

By brahmacarya, by tapas, the gods destroyed death.

Indra then used brahmacarya to bring the sun to the gods. (19)

Plants, the past and the future, day and night, the tree,
The year with (its) seasonsthey were born of the brahmacrin. (20)

159
 Those animals that are of the earth, of the sky, wild and domesticated,
Without wings and with wingsthey were born of the brahmacrin. (21)

All individually bear the breaths of Prajpati in themselves.

The brahman brought in the brahmacrin protects them all. (22)

Elicited by the gods, this wanders unmounted, shining.

From it were born the Brahman, the most excellent brahman, and all the gods,
along with immortality. (23)

He bears the shining brahman (over that all the gods are woven together),
Generating the in- and out-breaths and then the circulating breath, speech,
mind, heart, brahman, (and) intelligence. (24)

Put eye, ear, fame in us, food, semen, blood, belly. (25)

Arranging those, the brahmacrin stood generating tapas on the surface of the
water in the ocean.
He, bathed, tawny, (and) ruddy, shines much on the earth. (26) 356
356
AV 11.5:
brahmacr carati rdas ubh tsmin dev smanaso bhavanti /
s ddhra prthiv dva ca s cry tpas piparti // (1)

brahmacra pitro devajan prthag dev anusyanti srve /

gandharv enam nv yan tryastriat triat asahasr srvnt s dev s tpas piparti // (2)

cry upanyamno brahmacra krute grbham ant /

t r trs tisr udre bibharti t jt drum abhisyanti dev // (3)

iy samt prthiv dyur dvit yot ntrika samdh prti /

brahmacr samdh mkhalay rmea lok s tpas piparti // (4)

p rvo jt brhmao brahmacr gharm vsnas tpasd atihat /

tsmj jt br hmaa brhma jyeh dev ca srve amrtena skm // (5)

brahmacry ti samdh smiddha k ra vsno dkit drghmaru /

s sady eti p rvasmd ttara samudr lok nt sagrbhya mhur crikrat // (6)

brahmacr janyan brhmp lok praj pati paramehna vir jam /

grbho bhtv mrtasya ynv ndro ha bhtv surs tatarha // (7)

crys tataka nbhas ubh im urv gambhr prthiv dva ca /

t rakati tpas brahmacr tsmin dev smanaso bhavanti // (8)

im bh mi prthiv brahmacr bhik m jabhra pratham dva ca /

t krtv samdhv pste tyor rpit bhvanni vv // (9)

160
arv g any par any divs prh d gh nidh nhitau br hmaasya /
tu rakati tpas brahmacr tt kvala krute brhma vidv n // (10)

arv g any it any prthivy agn samto nbhas antarm /

tyo rayante ramy 'dhi drh s t n tihati tpas brahmacr // (11)

abhikrndan stanyann aru itig brhc chp 'nu bh mau jabhra /

brahmacr sicati s nau rta prthivy tna jvanti prada ctasra // (12)

agnu s rye candrmasi mtarvan brahmacry ps samdham dadhti /

t sm arc i prthag abhr caranti t sm jya pruo varm pa // (13)

cry mrtyr vrua sma adhaya pya /

jm t sant stvnas tir id svr bhrtam // (14)

am ghrt krute kvalam cry bhtv vrua /

ydyad ichat praj patau td brahmacr pr yachat sv n mitr dhy tmna // (15)

cry brahmacr brahmacr praj pati /

praj patir v rjati vir ndro 'bhavad va // (16)

brahmacryea tpas r j rr v rakati /

cry brahmacryea brahmacram ichate // (17)

brahmacryea kany yvna vindate ptim /

anav n brahmacrye vo ghs jigati // (18)

brahmacryea tpas dev mrtym pghnata /

ndro ha brahmacryea devbhya svr bharat // (19)

adhayo bhtabhavym ahortr vnaspti /

savatsar sah rtbhis t jt brahmacra // (20)

p rthiv divy pava ray grmy ca y /

apak paka ca y t jt brahmacra // (21)

prthak srve prjpaty pr n tmsu bibhrati /

t nt srvn brhma rakati brahmacry bhrtam // (22)

dev nm ett paritm nabhyrha carati rcamnam /

tsmj jt br hmaa brhma jyeh dev ca srve amrtena skm // (23)

brhma bhr jad bibharti tsmin dev dhi vve samt /

prpnu janyann d vyn v ca mno hrdaya brhma medh m // (24)

cku rtra yo asm su dhehy nna rto lhitam udram // (25)

t ni klpan brahmacr salilsya prh tpo 'tihat tapymna samudr /

s snt babhr pigal prthivy bah rocate // (26)

161
There are several interesting features worth noting about the description of the

brahmacrin in this hymn. First, he has a teacher (crya). No mention is made of a

curriculum that the crya teaches the brahmacrin, however; instead, in v. 3, the

crya, bringing the brahmacrin near (upanyamno, which is derived from the

same root and prefix as upayana, the Brahmanical rite of initiation), makes him an

embryo within (grbham ant) and bears him in his belly three nights (t

r trs tisr udre bibharti), after which the brahmacrin is born (jt). What

exactly is born or what this birth entails is not entirely clear, although twice (v. 5,

23) the most excellent brahman (brhma jyeh) of the Brahman

(br hmaa) 357 is said to be born from that (tsmj jt). Although no mention

is made of the crya teaching his brahmacrin anything in the ordinary sense, the

brahmacrin clearly serves his crya, bringing in v. 9 alms (bhik), which are

then used as fuel, presumably in the sacrificial fire. Repeated mention is made of the

brahmacrins tapas, and of the seemingly immense power that the brahmacrin

possesses as a result of it. The brahmacrins tapas appears, in v. 17-19, to be

equated with brahmacarya itself as a power by which great things can be

accomplished. It allows a king to defend his kingdom, an crya to seek a

brahmacrin, a maiden to win a husband, an animal to obtain food, the gods to

destroy death, and Indra to bring heaven to the gods. Although it is not stated so

explicitly, it is likely that this power referred to as brahmacarya is derived from

celibacy, since it is equated with tapas (a concept associated throughout Indian

357
I am taking brhmaa here as an adjective meaning relating to the brhmaa (noun).

162
history with control over seminal emissions) and attributed to many types of people

other than just brahmacrinsin particular, a young woman, whom we can assume

wins herself a husband on the basis of her virginity.

When described in these terms, brahmacarya and the brahmacrin hardly

seem like inappropriate categories for describing the Buddhist bhikkhu. To begin

with, the bhikkhu, upon entering the sagha, has both an upajjhya and an cariya,

whom he serves. 358 The bhikkhu obtains sustenance by begging for alms; indeed, the

word bhikkhu (Skt. bhiku) itself is derived from the verb root meaning to beg (bhik).

The bhikkhu is celibate, and the word brahmacariya, as we have seen above, refers

both to celibacy per se and more broadly and abstractly to the endeavor that he is

engaged in. Finally, although the language about the crya birthing his

brahmacrin that we find in the Atharva passage is not, generally speaking, paralleled

in the early Buddhist literature, nonetheless the identity that the bhikkhu is expected

to assume as the result of his training is expressed using the same word found in this

and other Brahmanical textsbrhmaa. 359

Indeed, as Timothy Lubin has suggested, it is possible that undergoing

brahmacarya was originally what made one a Brahman, and that the idea of being

358
Vin. Mv. 1.25 and 1.32 describe, with great detail and in identical terms, the many ways
that a new monk (i.e., ordained for less than ten years) should behave toward his upajjhya and
cariya. His duties essentially entail serving as a personal assistant, even a servant, for the more senior
monk. On the use of these terms in the Brahmanical and Buddhist traditions, see Gish Nakano and
N.A. Jayawickrama, crya, in Encyclopedia of Buddhism, ed. G.P. Malalasekera, vol. 1
(Government of Ceylon, 1961), 163-8.
359
This will be discussed in more detail in Part III.

163
born (i.e., through natural birth) into a particular vara came later. 360 This appears to

at least be hinted at in AV 11.5. As already noted, the teacher makes his

brahmacrin into an embryo (garbha) for three nights, after which he gives birth to

him. Although no direct connection is made, the most excellent brahman of the

Brahman (br hmaa brhma jyeh) is also said to be born in the process. A

more direct connection is made in an extended discussion of the upanayana in the

atapatha Brhmaa. In this passage (B 11.5.4), it is said that a teacher (crya)

teaches his brahmacrin the svitr, but that in the past he did so only after a year, the

rationale being that children are indeed born the measure of a year. 361

Progressively shorter intervals of time are then given for waiting to teach the svitr

(B 11.5.4.7-10), each being correlated in some way with the normal gestational

period of one year for a human child. This culminates in an interval of three days,

which is said to correspond to a year because there are three seasons in the year (B

11.5.4.11). The brhmaa then continues,

Also concerning that, they sing a loka: The teacher becomes pregnant after
laying his right hand [on the brahmacrin]. / On the third [night] he is born
with the svitr as a Brahman. // 362

360
Lubin, Boundaries, 85.
361
B 11.5.4.6: savatsarasammit vai garbh prajyante.
362
B 11.5.4.12: tadapi loka gyanti cryo garbh bhavati hastam dhya dakiam /
ttyasy sa jyate svitry saha brhmaa // iti. The verse continues by saying that a Brahman
should nevertheless be taught the svitr at once because the Brahman is related to Agni and Agni is
born at once: sadyo ha tvva brhmaynubryd gneyo vai brhmaa sadyo v agnir jyate
tasmt sadya eva brhmaynubryt. The import of this admonition, which clearly contradicts the
loka, is unclear. Perhaps it reflects a later time when Brahmanhood is conceived of as deriving from
(natural) birth, and thus the gestational period after upanayana was considered unnecessary.

164
This passage clearly parallels the Atharva passage examined above, which also

speaks of the crya becoming pregnant with his brahmacrin for a period of three

nights. Here, however, it is clear that the birth that results is the birth of a Brahman.

The idea that an rya is born again through the upanayana is of course a

well-known element of classical Hindu doctrine, and the term dvija, twice born, is

used in reference to that idea. In classical Hinduism, of course, all three rya varas

(brhmaa, katriya, and vaiya) are eligible for the upanayana and thus dvija. The

idea that they are born twice is simply a perfunctory reference to the upanayana

metaphorically conceived of as a second birth; all it serves to do is distinguish them

from dras, who are ineligible for initiation. Lubin has noted, however, that the

word dvija was not used to refer to the three higher varas until rules of initiation for

all three of those varas were laid out in the Ghya Stras. Prior to that time, the

word dvija was used only to refer to Brahmans, and the person who undergoes

upanayana and brahmacarya is assumed to be a Brahman. In fact, even in the Ghya

Stras, Lubin notes, discussions of the upanayana sometimes seem to assume that the

initiate is a Brahman. On the basis of these observations, Lubin argues that, starting

with the Ghya Stras, Brahmans tried to inculcate Brahmanical piety in the broader

population by making Vedic initiation and study accessible to and incumbent upon

katriyas and vaiyas, as well as Brahmans. 363

Following Bronkhorst, however, I would argue that we must be wary of

assuming a pre-existing vara system in interpreting texts such as the Ghya Stras

363
Lubin, Boundaries, 84-88.

165
Lubin is looking at; indeed, as Bronkhorst has convincingly argued, the very idea that

society could be divided into four vara was as much a part of the agenda of the new

Brahmanism as any other new idea introduced in late-Vedic and early post-Vedic

textssuch as, say, that there are three dvija classes who are all eligible for

upanayana. If brahmacarya was indeed, as the texts we have looked at suggest,

originally the way in which one became a Brahman, and that process was conceived

of in terms of rebirth, then the term dvija would naturally have been applied to

Brahmans precisely because they had been reborn, as Brahmans. The innovation of

the Ghya Stras, then, would not simply have been an extension of the right to

upanayana, and thus dvija status, to two other classes, but rather a total reformulation

of the purpose of upanayana and brahmacarya. Whereas before it had been the

means by which a person was reborn and therefore became a Brahman, now all

people were born (naturally) into a particular vara, and upanayana became merely a

marker of elite status within that system, and brahmacarya and opportunity to

educate these elites in an increasingly sophisticated body of Vedic knowledge. The

metaphor of rebirth was retained in the term dvija, but merely as a relic of the old

system.

This brings me to the third and final way in the rama system can obscure

our understanding of the categories found therein. Insofar as the rama system was

used as an ideological tool of the new Brahmanism, it encourages us to see the

categories it brings into its orbit in terms dictated by the new Brahmanismnamely,

the vision of society as divided into four varas determined by birth, with the

166
Brahmans as custodians and executors of Vedic knowledge at the top. I have already

argued that, as a matter of methodology, we should not assume that the advocates of

the new Brahmanism owned the word brhmaa, and that all others who used it,

such as the early Buddhists, did so only derivatively as a reaction or marketing.

The same is true of the terms brahmacarya and brahmacrin. As I have shown, the

earliest references to brahmacarya in the Brahmanical literature make reference to

features that are shared by Buddhist monksincluding celibacy, collecting alms,

serving a teacher, and the goal of becoming a Brahman.

The only aspect of Buddhist brahmacarya that might possibly be considered

innovative, insofar as it makes it impossible to repay ones debt to the gods by

producing progeny (TS 6.3.10.5, cited above), is that it is (ideally) lifelong. Even

here, however, the evidence for ancient practice is ambiguous. Most of the passages

we have looked at, including the extended discussions of brahmacarya in the Atharva

Veda and the atapatha Brhmaa, say nothing about how long it is to be practiced.

Some other texts do clearly define it as temporary. There is another interesting

extended discussion of the brahmacrin in the atapatha Brhmaa, for example,

that at least alludes to the end of brahmacarya by saying that the brahmacrin should

no longer beg for alms once he has bathed. 364 Two passages in the Chndogya

Upaniad (4.10.1, 6.1.2) describe brahmacarya as lasting twelve years. Elsewhere,

however, there are stories of people or gods who undertake much longer periods of

brahmacarya. A story in the Taittirya Brhmaa (3.10.11) has Bharadvja

364
B 11.3.3.7: na ha vai sntv bhiketa apa ha vai sntv bhik jayaty apa.

167
remaining in brahmacarya until age 75. Likewise, according to Chndogya

Upaniad 8.11.3, Indra underwent a brahmacarya of 101 years. 365 One passage of

the Chndogya even describes a type of brahmacrin who settles himself

permanently in the family of a teacher 366which is precisely the practice described

as one of the four ramas in the earliest descriptions of the system in the Dharma

Stras.

While there are therefore clear lines of continuity between early conceptions

of brahmacarya and Buddhist brahmacariya, conversely there were, as we have seen,

clear and important innovations with regard to brahmacarya in the Brahmanical

tradition, innovations that were tied to the incorporation of brahmacarya into the

rama system and the subsequent development of that system. By incorporating the

problematic practice of brahmacarya as a lifelong vocation of celibacy into the

system, the proponents of the new Brahmanism sequestered it, thus creating space for

the construction of a normative brahmacarya that was dedicated to Vedic study and

most importantly temporary, which in turn allowed for the assumption of

householdership and the production of children. Then, with the development of the

classical system, lifelong brahmacarya was eliminated entirely, by replacing it with

the normative period of Vedic study. Now, this is not to say that brahmacarya as

practiced by the early Buddhists was actually more faithful to the primitive

concept of brahmacarya than brahmacarya as it came to be defined in the orthodox

365
Citations found in Kane, History of Dharmastra, vol. 2., part 1, 271-4.
366
ChU 2.23.1: atyantam tmnam cryakule vasdayan.

168
Brahmanical traditionneedless to say, the Buddhist formulation of brahmacarya

was just as much an innovative construction based on historical antecedents as was

the orthodox Brahmanical formulation. What I am arguing, however, is that we

should not privilege the orthodox Brahmanical understanding of brahmacarya over

the Buddhist simply because brahmacarya really belongs to the Brahmanical

tradition. Early Buddhism and the new Brahmanism were both equally innovative

and equally rooted in tradition; their unique understandings of brahmacarya are both

derived from this interplay between historical precedent and innovation.

II.3.5 Vnaprastha and Jaila

Of the four categories that were incorporated into the rama system by the

Dharma Stra authors, the vnaprastha 367 is the one that is the most difficult to

correlate with the early Buddhist texts. Indeed, as Bronkhorst points out, the word

vnaprastha is not found in the Pali Canon. 368 I would argue, however, that the same

type of religious practitioner referred to by the Dharmastric literature as

vnaprastha (and apparently known to Megasthenes as such 369) is referred to in the

367
This is the term used by pastamba, Baudhyana, and Vasiha. Gautama, however, uses
the term vaikhnasa.
368
Bronkhorst, Greater Magadha, 91.
369
As we have already seen, Megasthenes refers to one of the two types of as
, forest dwellers, which is an exact translation of vnaprastha. Megasthenes says that the
are the most honourable of the (Strabo, Geography, 15.60; trans. McCrindle,
Ancient India as Described in Classical Literature, 67), which would imply that his local informants
were in some way partial to the vnaprastha over and against the parivrjakas (which, as I have
argued, is what Megasthenes refers to as physicians). It is interesting to note that Megasthenes also
reports that [t]he Brachmnes are held in the higher estimation [over the Garmnesi.e., ramaas],
for they agree more exactly in their opinions (Strabo, Geography, 15.59; trans. McCrindle, Ancient

169
Pali Canon as jaila. The word jaila refers to the hair of these ascetics, which was

not shaved off, as in the case of Buddhist monks and perhaps most other parivrjaka

groups, but worn long and matted or braided in some fashion. Three of the Dharma

Stras (GDhS 3.34, BDhS 2.11.15, VDhS 9.1) use the same word jaila to describe the

hair of the vnaprastha. It is thus possible that the authors of the early Buddhist texts

used the word jaila to refer to vnaprastha-like ascetics as a sort of slang

perhaps derived from colloquial usage, although it is impossible to know for sure.

This conclusion would appear to be confirmed by a post-canonical text, the

commentary on the Jtaka. 370 Both the term jaila and the sorts of practices that are

associated with the vnaprastha in the Dharma Stras are found much more often that

in the canonical texts of the Pali Canon, and they appear to be correlated with one

another. The portrayal of jailas in the Jtakas is generally negative; in fact, most

often jaila characters appear as wily rogues referred to as kajaila or deceitful

jaila. 371 Nevertheless, reference to them is useful for our purposes, because

kajailas often are described as living in the same way as the Dharma Stras

describe of the vnaprastha. All four Dharma Stras state clearly that the

vnaprastha (or vaikhnasa) should not only live in the forest, but should live off of

India as Described in Classical Literature, 65). This would indicate that his informants were
Brahmans. Taken together, these data support the conclusion that the vnaprastha had a special link to
the (Vedic) Brahmanical tradition, which is paralleled by the link that the jailas have to the Brahmans
in the Pali Canon.
370
The actual Jtaka tales written in prose are found in the commentary and are not, strictly
speaking, canonical. The canonical text of the Jtaka consists solely of the verses that are embedded
in these stories.
371
Jt. 89, 138, 277, 283, 313, 325, 344, 438, 454, 492, 505, 522, 532.

170
it. Unlike the parivrjaka, who wears rags and eats alms from the village (both of

which involve a connection to human culture), the vnaprastha wears clothes taken

from the wild (such as bark or animal skins) and eats gleanings (such as roots and

fruits). 372

Little mention is made of the clothing of jailas, either in the Jtaka or

elsewhere in the early Buddhist texts, but in Jt. 138, the external characteristics of a

kajailanamely, his matted locks (jahi) and his garment made of skin

(ajinasiy)are criticized by the Bodhisatta as worthless if the inside is impure.

More often, we find reference to the practice of living off of gleanings in the forest.

In Jt. 313, the Bodhisatta is born as a Brahman, but goes off to live in the forest,

nourishing himself with various fruits (phalphalena ypento). Later in the story,

he encounters a king, who refers to him derisively as a kajaila. Likewise, in Jt.

344, a kajaila builds an assama (Skt. rama) near a river and lives off of ripe

mangoes that fall from the trees in a nearby mango grove. In Jt. 283, interestingly,

we find a story whose main characters are a tiger and a boar, but a kajaila is also

mentioned who eats the meat obtained (gahitamasakhdako) by the tiger. This

would appear to parallel a provision in Baudhyana, who says that a vnaprastha, in

addition to eating roots and fruits, may also make use of the flesh of an animal killed

by a predator. 373 Finally, in Jt. 505, a certain jaila practices gardening and sells

what he grows in the market. He is criticized by the Bodhisatta (in this story taking

372
DhS 2.22.1-3; GDhS 3.26, 34; BDhS 2.11.15; VDhS 9.1, 3.
373
BDhS 2.11.15: baikam apy upayujta.

171
the form of a prince), who says to himself, This kajaila, rather than doing the duty

of a samaa for himself, does the work of a green-grocer. 374 This would seem to

imply that the jaila should not be growing his own food; this is paralleled by

language found in three of the Dharma Stras, which specifiy that the vnaprastha

should not even set foot on plowed land, much less engage in gardening himself. 375

Sometimes, generic ascetics, usually termed tpasa, appear as characters in

Jtakas and and have some of the same characteristics as jailas in the early Buddhist

texts or vnaprasthas in the Dharma Stras. For example, in Jt. 532, a whole family,

including parents and children, go to live in an assama; they have matted hair, and the

children gather fruits for their parents to eat. Likewise, in Jt. 526, a certain tpaso

lives in an assama and offers fruits and berries, which he has presumably acquired

through gleaning, to a guest. 376 A similar situation is found in Jt. 444, in which two

Brahmans go to the Himalayas, build an assama, and nourish themselves with

various roots and fruits of the forest (vanaphalphalena ypent). Later, they

separate, and the one who remains, Maabya, is framed for a robbery when a robber

breaks into a nearby house, steals something, and then drops the thing he has stolen in

front of Maabyas hut. Interestingly, Maabya is introduced in this story as a

tpasa, but when the owners of the house discovers the stolen item outside of his hut,

374
Jt. 505: aya kajailo attano samaadhamma akatv paikakamma karot ti.
375
GDhS 3.32, BDhS 2.11.15, VDhS 9.3.
376
Note that all four of the Dharma Stras state specifically that the vnaprastha should
welcome guests: DhS 2.22.17, GDhS 3.30, BDhS 2.11.15, VDhS 9.7. Vasiha even says specifically
that he should honor a guest who has come to his rama with almsfood of roots and fruits:
mlaphalabhaikeramgatam atithim abhyarcayet.

172
they say, Eh, wicked jaila, you work as a robber by night, [but] by day you go about

with the form of a tpasa.

This statement, taken together with the other evidence we have accumulated

from the Jtakas, confirms my hypothesis that jaila is a slang term used to refer to

the type of ascetic referred to in the Dharma Stras as a vnaprastha. As we have

seen, ascetics referred to as jaila or kajaila, aside from the obvious fact that they

have matted hair, also find sustenance in the same way as vnaprasthasthat is, by

gleaning, rather than by growing their own food or begging. In other cases, ascetics

who live in this way are simply called tpasa, but the householders reaction in Jt.

526, when they discover that Maabya is the robber (or so they think), shows that

these are no different than jailas, and that in fact jaila was a term of derision. For in

Jt. 526, Maabya is generally referred to as a tpasa, and it is only when the

householders discover that he is the robber that they in anger (indicated by the rude

interjection are) refer to him as a wicked jaila (duhajaila).

Now that we have established, on the basis of this evidence from the Jtakas,

that jaila was a slang term used to refer the same sort of ascetic referred to by the

Dharma Stra authors as the vnaprastha, we can return to the canonical texts

themselves, where we find more extended stories about jailas and the Buddhas

interaction with them. By far, the most prominent practice with which jailas are

associated in the Pali Canon is the tending of a sacred fire. A rule in the Vinaya

exempting jailas from the ordinary probationary period for members of other sects

who wish to ordain in the sagha, which we will examine in more detail shortly,

173
refers to them as aggik jailak to indicate their association with fire. 377 Likewise, a

short anecdote in the Udna has the Buddha encounter jailas who bob up and bob

down in the Gay, perform upward and downward bobbings, sprinkle [themselves

with water], 378 and offer oblations to Agni 379 and criticize them for thinking that

purity comes from such actions. This again confirms our conclusion that jaila is

slang for a vnaprastha-like ascetic, because all four of the Dharma Stras describe

the vnaprastha, uniquely among the three celibate ramas, as maintaining a sacred

fire. 380

The few extended narratives that feature jailas as characters within them

associate them closely with fire as well. The shortest of these is a parable told by a

monk named Kumra-Kassapa to a certain Prince Pysi of Kosala, who held the

evil opinion (ppaka dihigata) that there is no other world, there are no

spontaneously born beings, there is no fruit or result of good and bad actions. 381 In

the parable, a person who is described as aggika jaila and as living in a leaf-hut

located in the wilderness (arayatane paakuiy) adopts a baby whom he finds

377
Vin. 4.1.25.
378
The performance of ritual bathing is corroborated by Baudhyana and Vasiha, both of
whom describe the vnaprastha as performing ablutions at dawn, noon, and dusk (BDhS 2.11.15:
savanedakam upaspa; VDhS 9.9: triavaam udakopaspar). Likewise, pastamba says that the
vnaprastha should bathe before offering oblations to his sacred fire (DhS 2.22.12-14).
379
Ud. 1.9: gayya ummujjantipi nimujjantipi, ummujjanimujjampi karonti osicantipi,
aggimpi juhanti; emphasis mine.
380
DhS 2.21.21, GDhS 3.27, BDhS 2.11.15, VDhS 9.10.
381
DN 23: itipi natthi paro loko, natthi satt opaptik, natthi sukatadukkana kammna
phala vipko ti.

174
abandoned and raises him as his own. When the child is older, the jaila feels

comfortable leaving him alone while going off on business, so he tells him to watch

over his fire and relight it if he should accidentally let it go out. Sure enough, the

child gets absorbed in his games and lets the fire go out. He tries to relight the fire,

but unfortunately has no idea how to do so. He tries to create fire by cutting wood, by

pounding it with a mortar, and by winnowing it. Needless to say, none of these

methods works, and when the jaila returns, he chastises the boy for his stupidity.

The point of the parable is that Prince Pysi is looking for the existence of the other

world and the workings of karma in all the wrong ways (he tried, for example, to see

a mans jva escape from his body by systematically stripping away his skin, flesh,

and so forth). For our purposes, however, this parable shows that a jaila, like a

vnaprastha, lives in a remote location and tends to a special fire that should not be

allowed to go out.

A humorous anecdote in the Vinaya similarly portrays a jaila as maintaining

a sacred fire. In this story, 382 the Buddha, while wandering in Ceti near Bhaddavatik,

is warned by people not to go to Ambatittha because there is a nga who lives there

in the rama of a jaila (jailassa assame). The Buddha himself does not go to

Ambatittha, but one of his monks, a certain Sgata, does go there. He enters the fire-

room (agygra) of the rama, sits in meditation, and then battles with and

ultimately defeats the nga by producing heat (tejo). After this, Sgata continues

wandering to Kosamb, where his reputation for having defeated a nga proceeds him,

382
Vin. 2.5.6.1.

175
andin a somewhat contrived twist in the plotsome laypeople in the city give him

a special type of liquor, which he drinks, causing him to pass out and sleep

disrespectfully with his feet pointed toward the Buddha. The ultimate point of this

story within the context of the Vinaya is that it leads to the Buddha pronouncing a

rule against monks drinking alcohol, but for our purposes, it once again shows the

association between a jaila and fire.

The best-known story about jailas found in the Pali Canon, however, is the

story of the conversion of the three Kassapas, found in the Mahvagga of the Vinaya.

This story begins with the introduction of Kassapa of Uruvela (uruvelakassapa), who

is the leader of 500 jailas; Kassapa of the River (nadkassapa), who is the leader of

300 jailas; and Kassapa of Gay (gaykassapa), who is the leader of 200 jailas.

The Buddha goes to the rama (assama) of Uruvelakassapa and asks to spend the

night in the fire-room (agygre). The Buddha is warned by Kassapa that there is a

nga who lives inside the room who could harm him, but the Buddha insists on

spending the night anyway. The Buddha encounters the nga in the fire-room,

defeats him in a great conflagration, and presents the snake to the owner of the

rama with the words, This is your nga, Kassapa; his heat has been overcome by

heat. 383 After defeating the nga, the Buddha performs a whole series of other

miracles (pihriya), for a total of 3,500. This finally leads the three Kassapas and

their combined total of 1,000 jaila disciples to ordain as Buddhist monks.

383
Vin. 4.1.12: aya te, kassapa, ngo pariydinno assa tejas tejo ti.

176
 After this mass-ordination, the Buddha preaches the famous Fire Sermon,

which is known in the Theravda tradition as the Buddhas third sermon because it is

the third teaching given by the Buddha in the Mahvagga narrative, after the

Dhammacakkappavattana and the sermon on anatt. In this Fire Sermon, the

Buddha announces, Everything, monks, is burning (sabba, bhikkhave, ditta).

He then goes through the six sense organs, six sense objects, and six sense

consciousnesses, declaring that each is burning with the fires of passion (rga), hatred

(dosa), and delusion (moha). Seeing thus (eva passa), the Buddha concludes,

the learned Aryan disciple (sutav ariyasvako) grows weary (nibbindati) of the

sense organs, objects, and forms of consciousness, as well as the feelings (vedayita)

that arise therefrom; by doing so, he is liberated. This concludes the account of the

conversion of the three Kassapas.

The framing of this story by the Buddhas battle of heat with the ngawhich

fits much more comfortably into the narrative here than in the analogous case of

Sgata at Ambatittha already recounted aboveat the beginning and the Fire

Sermon at the end serves to further a polemic against the jailas and in particular

their worship of fire. At the very beginning of the narrative, the Buddha demonstrates

his superiority in generating heat by doing battle with a fire-breathing monster in the

very room where the jaila keeps his sacred fire. The message the Buddha seems to

be sending when he presents the defeated snake to Kassapa is, You pride yourself on

tending your sacred fire; well, my fire is bigger and badder than yours. Then, after

finally converting the jailas, the Buddha completely inverts the value placed on fire

177
by transforming it from a sacred object into a metaphor for sasra. Fire is no

longer something to be tended and cared for, but something to grow weary of and

abandon. The sermon ends with the 1,000 former jailas attaining Awakening, so

apparently they do abandon these metaphorical fires in addition to their physical

fires; their conversion is thus complete. 384

As already mentioned above, while narratives about jailas are not common in

the canonical Pali texts, they are quite common in the post-canonical narrative

portions of the Jtakas. Two Jtakas (144 and 162) are particularly worthy of note

because of the way in which they portray the worship of fire in the forest. In both of

these Jtakas, the Bodhisatta is born into a northern Brahman family

384
Gombrich has provided an insightful analysis of this sermon in the context of other
metaphorical references to fire in the Pali Canon in HowBuddhism Began, 65-72. Gombrich argues
that the number of fires mentioned in this sermon is three (rga, dosa, moha) because the Buddha is
alluding to the three sacred fires of a Brahmanical householder. In support of this statement, Gombrich
refers to AN 6.44, in which these three fires (which are to be avoided) are juxtaposed with the three
fires of the householder (which are reinterpreted as referring to parents, dependents, and
samaabrhma and therefore worthy of veneration)on which see also Gombrich, Recovering the
Buddhas Message, 17-20. I would argue that Gombrichs hermeneutic of polemic does work well
for interpreting an isolated narrative such as the story of the conversion of the three Kassapas discussed
here. As I have argued, this story appears to polemicize against the jaila practice of fire worship both
by demonstrating the Buddhas superiority with respect to fire and by devaluing fire itself. I am less
convinced by Gombrichs attempt to tie together numerous apparently fire-related metaphors scattered
throughout the early Buddhist texts and attribute them to a singular polemic against Brahmanism
conceived of by the Buddha. The polemic found in the story of the conversion of the Kassapas is
directed against jailas, not Brahmanism broadly construed, and as such, it plays with the concept of
fire in general, not the havanya, grhapatya, and dakigni specifically, as does AN 6.44 in the
context of an encounter with a Brahman who is preparing a Vedic sacrifice. Absolutely nothing is
made of the fact that rga, dosa, and moha are three in number in the Fire Sermon, only of the fact that
they are fires; indeed, it would make little sense to compare rga, dosa, and moha to the three fires
of the householder in the context of a polemic against jailas/vnaprasthas, given that vnaprasthas
were not householders and, according to the Dharma Stras, tended to only a single fire. Likewise,
while Gombrich draws a connection between the negative portrayal of fire in the Fire Sermon and the
concept of nibbna (Skt. nirva), which literally means blowing out, no such connection is made
by the text itself. In fact, the word nibbna is not used at all in the sermon, even though the 1,000
former jailas attain Awakening at the end. Virtually every metaphor for Awakening except blowing
out (nibbna) is usedknowledge (a), liberation (vimutti), dispassion (virga), freedom
from the effluents (sava)but not nibbna itself.

178
(udiccabrhmaakule), and when he turns 16, his parents take his birth fire

(jtaggi) and give him two choices: Either he can learn the three Vedas and live in

a house, or he can take his birth fire into the woods and worship it there. In both

cases, the Bodhisatta decides to take his fire into the forest to worship it there, but in

both cases disaster convinces him of the futility of fire worship. In Jt. 144, he finds

a cow and wants to offer it to the fire, but when he goes off to get some salt to make

the meat tastier for Agni, robbers come, slaughter the cow, and eat it, leaving only the

tail, hide, and shanks. When the Bodhisatta returns and sees this, he realizes the fire

will not be able to protect him if it cannot even protect its own cow, so he feeds the

undesireable leftovers of the cow to the fire and then douses it with water. In Jt. 162,

the Bodhisatta accidentally sets his rama on fire by pouring a particularly fatty

oblation into the fire that causes it to flame up; realizing that the fire is not a good

companion, he waits for the rama to burn down and then douses the fire with water.

Now, in neither of these Jtakas is the Bodhisatta explicitly identified as a

jaila, although in Jt. 162, a critique of jailas is implied by the fact that the

recounting of the Jtaka is prompted when the Buddha and his disciples encounter a

group of jailas, whom the Buddha says are practicing a false asceticism

(micchtapa). 385 What is interesting about these narratives, however, is that the

Bodhisatta is identified as a Brahman, and his life follows a course similar to that

385
In Jt. 144, the situation that prompts the Buddha to recount his previous life is nearly
identical, except here the false ascetics he and his disciples encounter are jvikas, rather than jailas.
This makes for a disjointed narrative, however, since the Buddha continues, as in Jt. 162, by
criticizing the practice of making offerings to fire, which is a practice of the jailas and not of the
jvikas.

179
outlined by the Dharma Stras in their descriptions of the rama system; that is,

once he reaches a certain age, he is given a choice between living as a householder

and going out into the forest to live an alternative lifestyle of singular devotion to the

sacred fire. As in the Dharma Stras, and unlike the later classical texts, this is

presented as a lifelong decision, and in fact, even after abandoning the fire, the

Bodhisatta does not return home, but lives out the rest of his life as a homeless ascetic.

The fact that the Bodhisatta is portrayed in these two Jtakas as engaging in a

jaila-like practice of worshipping fire in the forest after being born as a Brahman is

paralleled by links with Brahmanical identity found in narratives about jailas. In the

story of the conversion of the Kassapa brothers already discussed above, for example,

Uruvelakassapa is at one point referred to as a Brahman. 386 A more substantive link

is found in the one major narrative about a jaila that we have not already discussed,

the story of the jaila Keiya. This story is found in two versions. We will focus on

the version in the Sela Sutta (MN 92) and reproduced at Sn. 3.7, since it contains the

most extensive links between jailas and Brahmans. 387 In this version of the story, a

jaila named Keiya hears of the fame of Gotama 388 and therefore goes to see him

386
Vin. 4.1.12.
387
The other version, found at Vin. 4.6.182, is essentially the same as the frame story for the
Sela Sutta, in which Keiya meets the Buddha and feeds him and his entourage of monks, but without
the embedded account about the Brahman Sela. Even though Sela makes no appearance in this version,
however, the Buddha still initially objects to Keiya feeding him on the grounds that he is devoted to
the Brahmans.
388
A particular formula is used here to describe the good reputation (kalyo kittisaddo) of
Gotama that is characteristic of encounter dialogs with Brahmans, of which the Sela Sutta is a
somewhat idiosyncratic example. The role of formulae in demarcating the encounter dialog genre
will be explored more deeply in Part IV.

180
and offers to feed both him and his entire entourage of 1,250 monks. The Buddha

responds to this offer by objecting that, on the one hand, his entourage is quite large

(and therefore would be difficult to feed) and, on the other, Keiya is devoted to the

Brahmans (brhmaesu abhippasanno). This devotion to the Brahmans is borne

out later in the story, when, after the Buddha finally accepts Keiyas invitation,

Keiya is busy preparing the meal and a certain Brahman named Sela comes by and

asks Keiya what he is doing. While Keiya is described simply as a jaila, Sela is

labeled explicitly as a Brahman and described with a formula exalting his Vedic

learning. 389 This encounter between Sela and Keiya is not by chance, however; the

text reports that at that time Keiya the jaila was devoted to Sela the Brahman. 390

The relationship between Keiya and Sela is unfortunately not discussed in any

further detail. Sela, impressed by what Keiya tells him about the Buddha, decides to

visit the Buddha and ultimately ordains in the sagha together with his entourage and

attains Awakening. Keiya, quite separately, continues with his preparations and

feeds the sagha, after which we hear nothing more about him.

One final characteristic of the portrayal of jailas in the Pali Canon that should

not go without mention is their association with ramas (P. assama). Of the four

Dharma Stras, only Vasiha mentions specifically that the vnaprastha lives in an

389
This is, again, a characteristic formula found in encounter dialogs, to be discussed in
more detail in Part IV.
390
tena kho pana samayena keiyo jailo sele brhmae abhippasanno hoti.

181
rama, 391 but jailas are repeatedly associated with assamas in the early Buddhist

texts. In each of the four major narratives about jailas found in the canonical Pali

textsthe parable told to Prince Pysi, the story of Sgatas encounter with the nga,

the story of the conversion of the Kassapa brothers, and the story of Keiyathe

jaila or jailas in the story are said to reside in an assama. Within the extended

narratives of the Vinaya and first four Nikyas, only two other people, not identified

as jailas, are said to reside in an assama. The first is Bharauklma, already

mentioned in the previous section, who is identified as a former co-religionist

(purasabrahmacr) of the Buddha and is said to reside in an assama at

Kapilavatthu (AN 3.127). The second is a certain Rammaka, who is identified as a

Brahman rather than a jaila, at whose assama the Buddha stays when he teaches the

Ariyapariyesana Sutta (MN 27). Although this latter sutta says nothing more about

Rammaka or his assama other than that the Buddha stayed there, the fact that the

owner of this assama is identified as a Brahman, when ordinarily assamas are

associated with jailas, reinforces the connection already outlined between jaila and

Brahman identity, or, to put it somewhat less rigorously, that the jaila is a

Brahmanical ascetic. 392

391
VDhS 9.7: mlaphalabhaikeramgatam atithim abhyarcayet (He should honor a
guest who has come to his rama with almsfood of roots and fruits.). pastamba, however, says that
the vnaprastha should build a house (gha) outside of the village (DhS 2.22.8), and both Gautama
and Baudhyana imply some sort of accommodations when they say that he should entertain guests
(GDhS 3.30, BDhS 2.11.15).
392
This is the same conclusion arrived at by Tsuchida Ryutaru, who argues that there are two
types of Brahmans present in the Pali textsthe brhmaagahapatikas or brhmaamahslas who
are wealthy landowners, and the jailas who are ascetics devoted to fire worship and austerity
Tsuchida Ryutaro, Two Categories of Brahman.

182
 Scholarship has been somewhat divided on the question of the origin of the

rama as hermitage and its relationship to other institutions. Olivelle, on the one

hand, has argued that the earliest meaning of the word rama, which is not found in

the Vedic Sahits or early Brahmaas or Upaniads and thus appears to be a

neologism, 393 was in fact a place of religious exertion, and by extension, referred also

to exceptional Brahmans who dedicated their lives in an extraordinary manner to

religious exercise (rama), living, in all likelihood, in areas somewhat removed from

villages and towns. 394 The authors of the rama system then coopted this term

to indicate that the ascetic life was as good as the ideal Brhmaical life expressed

by the term rama. 395 Bronkhorst, on the other hand, has emphasized the fictive,

literary aspect of rama in its meaning as a place of religious exertion. He notes that

ramas are mostly referred to in Brahmanical literature, and the descriptions of them

found therein tend to be quite idyllic, while the similar institution of the agrahra is

referred to more commonly in inscriptions commemorating land grants. He argues,

therefore, that

it is possible to consider that the references to agrahras which we find

mentioned primarily in inscriptions, and those to ramas which are so
frequent in classical Sanskrit literature, concern one and the same historical
institution, or better perhaps: two different institutions with considerable
overlap. Agrahras were donated to Brahmins because their donors expected
their occupants to live more less in accordance with life as it was presumably
lived in ramas, and Brahmins depicted ramas in this particular manner at

393
Olivelle, rama System, 8.
394
Ibid., 19.
395
Ibid., 20.

183
 least in part in order to entice their rulers to create such settlements, or more
of them. 396

Moreover, Bronkhorst argues that Brahmans developed the idea of the rama to

compete with other groups, such as the Buddhists, Jains, and jvikas, for grants of

land from kings. He notes that the earliest known royal grants, from the inscriptional

evidence, are to these non-Brahmanical groups, 397 and the earliest inscriptional

records of land grants to Brahmans are not found until the middle of the first

millennium BCE. 398

I believe that useful insights can be taken from both of these approaches to the

institution/concept of rama as a place of religious exertion. First, I find

Bronkhorsts argument linking the literary depiction of ramas to the institution of

agrahra land grants convincing, as well as his observation that the Brahmans, in this

process, appear to have been latecomers to the game of receiving royal donations.

This is consistent with Olivelles observation that the term rama is not found in

most of the Vedic literature, and appears to be a neologism that appears only in very

late Vedic and post-Vedic texts. It is also consistent with the distribution of

references to jailas and assamas in the early Buddhist literature. As already noted,

they appear only a few times in the narrative texts of the Vinaya and the first four

Nikyas; they appear with greater frequency, however, in the late canonical and post-

396
Bronkhorst, Buddhism in the Shadow of Brahmanism, 85.
397
Ibid., 96.
398
Ibid., 88-89.

184
canonical texts of the Apadna and Jtaka. 399 On the other hand, I am also

convinced by Olivelles argument that the earliest meaning of the word rama was

to a place of religious exertion. This is consistent with the fact that the use of the

word rama to refer to a system of four ways of life is unknown outside of the

Brahmanical tradition, while the use of the word to mean a place of religious exertion

is knownas, for example, in the Buddhist texts examined here. Furthermore, given

that jailas are depicted in the Pali Canon as actual interlocutors of the Buddha, and as

living in actual places called assamas, it seems unlikely that the rama was simply a

Brahmanical literary construct meant to attract land donations from kings. In fact,

while brahmadeyya land grants are frequently mentioned in association with wealthy

householder Brahmans (brhmaagahapatika), 400 the assamas of jailas are never

said to be the result of such grants. Although the Brahmanical literary depiction of

ramas may have come over time to serve the purpose of attracting royal donations

of land, the fact that they are also found in Buddhist textswhich presumably would

have had nothing to gain in attracting donations to Brahmansand disconnected

therein from land grants to Brahmans, indicates that this literary depiction was rooted

in actual practice that only later came to be coopted to a Brahmanical agenda for

attracting donations.
399
jaia: Ap. 1.1.3-1, 1.3.1, 1.3.4.9, 1.43.7, 1.50.8, 1.54.8, 2.4.2, 2.4.5; Jt. 377, 469, 487,
492, 526. assama: Ap. 1.1.3-1, 1.2.4, 1.3.1, 1.9.4, 1.13.2, 1.13.7, 1.13.10, 1.24.5, 1.31.5, 1.32.4, 1.32.5,
1.34.9, 1.40.4, 1.40.6, 1.40.10, 1.41.1, 1.41.5, 1.41.7, 1.41.8, 1.41.10, 1.42.1, 1.42.2, 1.42.3, 1.42.5,
1.43.3, 1.43.7, 1.45.2, 1.45.3, 1.45.4, 1.49.2, 1.49.3, 1.49.4, 1.49.7, 1.49.10, 1.50.8, 1.52.9, 1.54.8,
1.56.3, 1.56.8, 2.3.7, 2.3.8; Jt. 99, 253, 299, 346, 484, 490, 498, 503, 507, 523, 526, 532, 535, 540,
547. I have cited here only only those Jtakas in which jaila or assama is found in the canonical
verses; there are further examples in which one or the other term is found in the post-canonical
narrative.
400
DN 3, 4, 5, 12, 23; MN 95.

185
 Indeed, as always, imposing a binary between Brahmanical and non-

Brahmanical can obscure more than it reveals in trying to understand categories such

as jaila. On the one hand, jailas as they are depicted in the early Buddhist texts

appear to be closely related to Vedic Brahmansmost obviously, of course, their

practice is centered on making offerings to a sacred fire, but they are also, as we have

seen, depicted as being, as in the case of Uruvelakassapa, referred to as brhmaa,

and as being devoted, as in the case of Keiya, to a Brahman schooled in the three

Vedas. On the other hand, the Dharma Stras demonstrate that however close the

jailas might have been to the category Brahman or how close their relationship

might have been to Brahmans educated in the Vedas, their relationship to the

proponents of the new Brahmanism was tenuous at best. That is, at least insofar as

they practiced celibacy, they were branded vnaprasthas and either denied legitimacy

completely or frowned upon as inferior to the ghastha. As I have argued above, the

principle according to which the rama system is articulated in the Dharma Stras is

not a binary between Brahmanical and non-Brahmanical. Indeed, the modes of

practice circumscribed by all four of the ramas could be said to have antecedents in

the Vedic tradition; at the same time, the rama system appears designed to

encompass all forms of practice, and not simply those within a circumscribed

Brahmanical tradition, even a liberal one. 401 In other words, the principle

401
Indeed, this is why the rama system is presented as being so problematic in the Dharma
Stras. It strives to fulfill the two Dharmastric goals of representing actual custom and rooting that
custom in the Vedas, but in doing so it runs into the problem that several lifestyles found in actual
practice conflicted irreconcilably (in the eyes of the Dharma Stra authors, at least) with the Vedic
injunction to procreate. This is why Gautama and Baudhyana take the extraordinary step of declaring
entire swaths of actual custom invalid. The conflict, I argue, is not between conservative Brahmans

186
according to which the rama system is articulated in the Dharma Stras is one of

totality, irrespective of any imagined boundaries between Brahmanical and non-

Brahmanical, which is then judged as a whole against the strict interpretation of the

Veda adhered to by the proponents of the new Brahmanism. 402

For their own part, the early Buddhists seem to have been unconcerned with

any binary between Brahmanical and non-Brahmanical in situating themselves

vis--vis other samaabrhmaa groups. We have already seen that the early

Buddhist texts attempt to distance the Buddhist bhikkhu from the category

paribbjaka, even though, colloquially, he clearly would have fallen under that

category, but embrace the category brahmacr, not unreasonably, as reflecting his

lifestyle as a celibate student. We have also seen that the early texts are critical of a

wide variety of particular rival groups and seek to define the Buddhas teaching

against them; this is as true of the jailas as it is of the Nigahas (Jains) and jvikas.

Nevertheless, the Vinaya (4.1.25) singles out jailas as the one group of rival

sectarians who did not have to undergo a four-month probationary period (parivsa)

if they wanted to convertthat is, ordain as Buddhist monks. The ordinary rule for

members of other sects (aatitthiyapubb) who wanted to ordain was that they had

and liberal Brahmans, but between the authors of the Dharma Stras qua proponents of the new
Brahmanism and the social reality they sought to faithfully report.
402
Put differently, we could say that, according to the advocates of the new Brahmanism, only
the ghastha was Brahmanical, while the other three ramas are non-Brahmanical, simply on the
basis of the question of procreation, but this particular criterion for defining Brahmanical and non-
Brahmanical neither would have been shared by all people in late first millennium BCE India, nor is it
useful for the modern scholar, given, e.g., that jailas can reasonably be described as Brahmanical.
The more general point I am trying to make, though, is that any attempt to distinguish between
Brahmanical and non-Brahmanical in the time period we are looking at is ultimately arbitrary.

187
to wait four months; this was to ensure that new converts were committed to being

Buddhist monks and would not switch alliegences back and forth. The reason the

Buddha gives for exempting jailas from this rule is that they are kammavdino and

kiriyavdino.

What exactly does this mean? According to the commentary, it means that

they do not ward off causative action (kiriya); they are of the view that action

(kamma) and the result of action exist. 403 If this is taken to refer simply to the belief

in karma, then it is not clear why jailas would be singled out in this respect, and

other groups, such as the Jains, would be excluded, since the latter also believe in

karma. Nevertheless, this distinction on the basis of being kammavd and kiriyavd

is maintained consistently in other texts. For example, a sutta in the Aguttara

Nikya (6.57) reports that Praa Kassapa, one of the six heretical teachers already

discussed above, categorizes people in six groups, in which bhikkh are identified as

kammavd and kiriyavd, and other groups including the Jains are not:

Reverend sir, six species are made known by Praa Kassapathe

black species is made known, the indigo species is made known, the red
species is made known, the yellow species is made known, the white species
is made known, the supremely white species is made known.
With respect to that, reverend sir, this is the black species made known
by Praa Kassapa: sheep-butchers, fowlers, deer-hunters, hunters, fishermen,
robbers, killers of robbers, jailers, or whoever else follows a bloody trade.

403
ete kiriya na paibhanti, atthi kamma, atthi kammavipko ti evadihik. The word
paibhanti here is somewhat ambigious. Does it mean ward off, in the sense of avoiding engaging
in causative action, or does it mean reject, as in a philosophical position? The latter makes more
sense in the immediate context of this sentenceit would imply that the people in question reject the
theory of karmabut the former makes more sense as a description of what we know of Jain and
jvika soteriological theory. Perhaps this ambiguity is intentionalintended, that is, to blur the
distinction between the Jains and jvikas actual position and the accusation that they do not believe
in karma at all.

188
 With respect to that, reverend sir, this is the indigo species made
known by Praa Kassapa: bhikkhus of the thorn-practice and whoever else
propounds action (kammavd) and causative action (kriyavd).
With respect to that, reverend sir, this is the red species made known
by Praa Kassapa: one-cloth nigahas.
With respect to that, reverend sir, this is the yellow species made
known by Praa Kassapa: house-dwelling, white-clad disciples of naked
ascetics.
With respect to that, reverend sir, this is the white species made known
by Praa Kassapa: male and female jvikas.
With respect to that, reverend sir, this is the supremely white species
made known by Praa Kassapa: Nanda Vaccha, Kisa Sakicca and Makkhali
Gosla.
Reverend sir, these are the six species made known by Praa
Kassapa. 404

This taxonomy appears to present a hierarchy of human beings in which bhikkhus

presumably referring to Buddhistsare near the bottom, just a step above the

complete reprobates, and then are followed, in sequence, by Jains, lay disciples,

ordinary jvikas, and finally the fully realized jvika saints. What is important for

our purposes to note is that the bhikkhus are identified as kammavd and kriyavd,

404
praena, bhante, kassapena chaabhijtiyo paatt kahbhijti paatt, nlbhijti
paatt, lohitbhijti paatt, haliddbhijti paatt, sukkbhijti paatt, paramasukkbhijti
paatt.

tatrida, bhante, praena kassapena kahbhijti paatt, orabbhik skarik skuik mgavik
ludd macchaghtak cor coraghtak bandhangrik ye v panaepi keci kurrakammant.

tatrida, bhante, praena kassapena nlbhijti paatt, bhikkh kaakavuttik ye v panaepi

keci kammavd kriyavd.

tatrida, bhante, praena kassapena lohitbhijti paatt, nigah ekasak.

tatrida, bhante, praena kassapena haliddbhijti paatt, gih odtavasan acelakasvak.

tatrida, bhante, praena kassapena sukkbhijti paatt, jvak jvakiniyo.

tatrida, bhante, praena kassapena paramasukkbhijti paatt, nando vaccho kiso sakicco
makkhali goslo.

praena, bhante, kassapena im chaabhijtiyo paatt ti.

189
while the various other groups of people that can variously be identified as Jain or

jvika are not.

While Vin. 4.1.25 is alone in identifying jailas as kammavdino and

kiriyavdino, and the Aguttara passage just cited does not specify who, other than

bhikkhus, are kammavd and kriyavd, two other passages appear to link these

doctrines to the Brahmans. In two encounter dialogs, the Brahman interlocutor

praises the Buddha in a long formula that includes the following line: The samaa

Gotama is indeed a proponent of action, a proponent of causative action, devoted

without sin to the Brahmanical people. 405 Here, then, we have Brahmans not only

mirroring the attribution of kammavda and kiriyavda to jailas found in the Vinaya,

but explicitly linking those doctrines to Brahmanhood.

Now, it is not entirely clear why the authors of the early Buddhist texts see

kammavda and kiriyavda as doctrines tying themselves to the jailas and even

householder Brahmans and distinguishing them from Jains and jvikas, but I think

that we can hazard a guess. In Praa Kassapas taxonomy cited above, bhikkhus,

whom I take to refer to as Buddhists, are not only referred to as kammavd and

kriyavd, but also as kaakavuttik, an obscure term I translate as of the thorn-

practice. Presumably this refers to a perception that Buddhists, and others like them

who are left unspecified, are in some sense austere in their practices. This perception

(or rather perception of a perception, since the words are put into Praa Kassapas

405
DN 4, 5: samao khalu, bho, gotamo kammavd kiriyavd appapurekkhro
brahmaya pajya.

190
mouth by the Buddhist author of the text) is mirrored by depictions of jvika

teachings in the Pali suttas, which typically depict them as rejecting the efficacy of

karma. In the Smaaphala Sutta (DN 2), for example, Praa Kassapa himself is

said to proclaim a doctrine known simply as non-causative action (akiriya).

Makkhali Gosala declares that there is no condition for the defilement of beings;

without cause, without condition are beings defiled. 406 Ajita Kesakambal denies

that there are any fruits of actions or any other world in which to enjoy them, and

Pakudha Kaccyana declares a doctrine of atomism whereby killing entails not killing,

but a mere separation of the elements. Now, as Bronkhorst has convincingly shown,

these doctrines, although they misrepresent the actual doctrines held by those being

criticized, probably do nonetheless reflect an actual jvika belief that one could not

attain liberation through actions; one could only wait for ones karma to burn off

over countless lifetimes before being liberated. In this respect, they differed from the

Jains only insofar as the latter believed that one could undertake austerities to speed

up the process of burning off karma and thus attain liberation more quickly. 407 In

both cases, the ultimate goal was complete non-action; it was this, I argue, that the

Buddhistswho clearly valued the efficacy of good karma (i.e., merit, or puya) to,

among other things, transform a person into a Buddhaperceived as a rejection of

karma, and as distinguishing both the Jains and the jvikas from themselves, Vedic

Brahmans, and jailas.

406
natthi mahrja hetu natthi paccayo sattna sakilesya, ahet apaccay satt
sakilissanti.
407
Bronkhorst, Greater Magadha, 38-51.

191
 Regardless of why exactly the early Buddhists saw themselves as linked

uniquely to the jailas as kammavdino and kiriyavdino, what is important is that

that is how they understood themselves, irrespective of any imagined boundaries

between Brahmanical and non-Brahmanical. The early Buddhists certainly have

pointed critiques for householder Brahmans and jailas, just as much as they do for

Jains and jvikas, but on the matter of karma, they see themselves as more similar to

the former than to the latter. The key thesis of Bronkhorsts Greater Magadha, which

is that the Greater Magadha region out of which the non-Brahmanical movements,

including Buddhism, Jainism, and jvikism, arose, was distinguished from the

Brahmanical tradition by belief in karma and rebirth, is difficult to reconcile with this

self-perception on the part of the Buddhists. More generally, though, any

dichotomization of early Indian religion into Brahmanical and non-Brahmanical

obscures the way in which various groups made use of a common set of vocabulary

including ramaa, brhmaa, brahmacrin, parivrjaka, jaila, and othersto

taxonomize the socio-religious world they lived in and situate themselves within it.

192
 Chapter II.4

Conclusion to Part II

In Part II, we have examined a variety of discursive categories relating to what

in modern English we would refer to as religious practice in ancient India.

Although there was no word in ancient India that corresponded exactly to the modern

concept of religion, the categories we have examined here all have been related

either via the word ramaa in colloquial discourse or via the word rama in

Brahmanical discourseto the Sanskrit root ram. On the one hand, colloquial

discourse developed by the late first millennium BCE a concept of the ramaa as a

practitioner whose name was derived from the exertion he or she undertook and

who came to be conceived in opposition to the brhmaa. This category served as

the genus under which the early Buddhists categorized themselves as a species

(sama sakyaputtiy) for most likely hundreds of years prior to the adoption of the

term bauddha. As we have seen, however, the opposition that solidified between the

categories ramaa and brhmaa from the time of Candragupta Maurya onward

does not appear to have always been so absolute. As Olivelle has shown, the term

193
ramaa and other related terms are not foreign to the Brahmanical tradition and in

fact are found in Vedic texts. In addition, as I have shown here, the predominant

usage of the terms of samaa and brhmaa in the early Buddhist texts is, in spite of

what later sources such as Patajali would lead us to believe, not oppositional at all,

but treats the two together as a single group to be either praised as worthy of gifts and

honor or used as a foil against which to construct Buddhist identity.

The bifurcation of ramaa and brhmaa into separate, mutually exclusive,

and even oppositional categories was most likely driven by a movement that I call,

following Bronkhorst, the new Brahmanism. This movement, exemplified by the

Brahman interlocutors in the encounter dialogs found in the early Buddhist stras,

and by their own words in the Dharma Stras, was a householder-supremacist

movement that arrogated to themselves the category of brhmaa by promoting the

vara system as a scheme of social taxonomy and valorizing themselves as the

pinnacle of the social hierarchy on the basis of (presumed) birth and Vedic learning.

Instead of relying on the ambigious colloquial category ramaa, these proponents of

the new Brahmanism developed the more technically sophisticated rama system

in the Dharma Stras to incorporate and systematize several colloquial categories for

religious practice for the purpose of taxonomizing social custom in all its variety and

then subjecting that variety to a Vedicizing hermeneutic that valorized Vedic learning,

the vara system, and the householder lifestyle. The early Buddhists, on the other

hand, negotiated in less systematic ways with many of the same colloquial categories

that were incorporated into the rama system, as well as the categories ramaa and

194
brhmaa themselves, so as to situate themselves in the religious world of early India

and create an identity for themselves vis--vis other groups.

Methodologically speaking, the most important conclusion to be drawn from

the evidence presented in Part II is that social taxonomies and discursive categories

were sites of contestation within a fluid process of identity-formation in early India,

and therefore we should not bias our understanding of that period by assuming a

dichotomy between Brahmanical and non-Brahmanical. The distinction between

ramaa and brhmaa, and concomitantly the categorization of sakyaputtiy (i.e.,

Buddhists) as samaa and not brhmaa, was the result of this process of identity-

formation and not the cause. Although the Buddhists eventually came to abandon the

category brhmaa more-or-less completely to the proponents of the new

Brahmanism, their early texts provide ample evidence that they at first mounted a

significant struggle to retain it. It is this struggle that we will be examining in Part III.

195
 Part III

Diachrony and the Use of Brahmanical Terms

in Early Buddhist Texts

196
 Introduction

Problematizing the Concept of Brahmanical Terms

Any theory of the relationship between early Buddhism and Brahmanism must

of necessity contend with the use of seemingly Brahmanical terms, of which a

comprehensive list has been given by Norman, 408 in the early Buddhist texts.

Bronkhorsts theory of an original radical dichotomy between the religion and culture

of the Vedic Brahmans in the West and that of Greater Magadha, out of which

Buddhism arose, in the East fails to account for the pervasive presence of these

Brahmanical terms in the early Buddhist texts, except implicitly by suggesting that

certain Buddhist texts may be relatively late. 409 More commonly, scholars have

assumed a model of reaction or marketing for the rise of Buddhism vis--vis

408
Norman, Theravda Buddhism and Brahmanical Hinduism.
409
Bronkhorst, Greater Magadha, Appx. 6, Brahmins in the Buddhist Canon, 353-356. I
say implicitly because this short appendix actually deals specifically with certain long, narrative
suttas in which Brahmans feature as interlocutorsin other words, what I would call encounter
dialogsrather than the more general problem of Brahmanical terminology, some of which is
found more pervasively across a wide variety of early Buddhist texts.

197
Brahmanism, within the context of which Brahmanical terms in the early Buddhist

texts serve as a central example of that model at work. Norman, for example, in his

article in which he catalogs all such terms, labels them terms taken over by the

Buddha but used with new senses. 410 This wording is consistent with Gombrichs

overall approach to the issue of Buddhist origins, which emphasizes the Brahmanical

background against which Buddhism arose and the Buddhas creative use of

Brahmanical terminology and ideas to convey his own teachings. The Buddha,

according to Gombrich, employed skill in means by adopting the language of his

interlocutors to express his own position; often, this skill in means involved

allegoricizatione.g., the non-literal reference to himself as a Brahman and

punning on Brahmanical terms. 411 Like Norman and Gombrich, Tsuchida also

understands the presence of certain terms in Buddhist texts as an appropriation of

terms that are properly Brahmanical and giving them new Buddhist senses. He

writes, It is well known that in dealing with such Brahmanical concepts as tapa,

vijjcaraa, yaa, and brhmaa, the Buddha does not discard them as such but

gives them different connotations so as to incorporate them into his own scheme of

dhamma. 412 Citing Tsuchida, Bailey and Mabbett characterize the presence of

Brahmanical terms in Buddhist texts even more baldly as evidence of conscious

opposition: This almost obsessive mapping of Buddhist teachings upon the structure

410
Norman, Theravda Buddhism and Brahmanical Hinduism, 194.
411
Gombrich, How Buddhism Began, 17-20.
412
Tsuchida, Two Categories of Brahmins, 75.

198
of brhmaical tradition clearly demonstrates that the Buddhists saw themselves as

competitors with the Brahmins, in spite of the obvious differences in their

concerns. 413

Now, it is not my goal in Part III to reject a marketing explanation for the

presence of Brahmanical terms in the early Buddhist texts in toto. In certain,

indeed even many cases, a marketing model does seem to be the best explanation

for the use of the word or words in question. Oliver Freiberger has argued

convincingly, for example, that Vedic sacrifice (yaa) is reinterpreted in three

ways in the early Buddhist texts: by substituting vegetal for animal offerings, by

equating dna (giving, esp. to the sagha) with sacrifice, and by equating any

spiritual virtue such as going to refuge with sacrifice. These three reinterpretations

represented different ways of coping with the practice of and value placed on

sacrifice while maintaining the superiority of the Buddhist teaching. 414 Likewise, as

we will see shortly, many of the uses of the word brhmaa seem to clearly imply a

metaphorical or allegoricized meaning of Brahman as equivalent to the Buddhist

arhat, over and against a literal meaning of the same referring to a particular social

group that considered itself the highest of four varas. This is, indeed, to be expected

in a body of literature that contains, as we have already discussed and will investigate

in more detail in Part IV, encounter dialogs in which a clear dichotomy is set up

413
Bailey and Mabbett, Sociology of Early Buddhism, 123.
414
Freiberger, Ideal Sacrifice.

199
between the Buddha as samaa and proponents of the new Brahmanism as

brhma.

Instead, what I will argue in Part III is that we should not treat all so-called

Brahmanical terms found in the early Buddhist texts monolithically as examples of

a single phenomenon of marketing, allegory, or even appropriation. Indeed,

we should not assume uncritically that such words are Brahmanical to begin with.

As I argued in Part II, there are methodological advantages to be gained by not

assuming, as is typically done in scholarly literature, a dichotomy between

Brahmanical and non-Brahmanical in ancient India. Aside from the fact that it is

not strictly-speaking an emic distinction, it is difficult to imagine how the distinction

can be made unambiguously in an etic sense either. Social categories, including that

of brhmaa itself, were fiercely contested in ancient India, and sectarian identities

arose only slowly out of this process of contestation; they were not pre-given entities

that somehow metahistorically drove the process.

In particular, the rise of a cohesive (or what we might call classical)

Brahmanical identity was driven in particular by the agenda of what Bronkhorst has

called the new Brahmanism, whose proponents promoted the analysis of society

according to the vara system, valorized themselves as the Brahmans at the head of

that system on the basis of their (supposed) birth and Vedic knowledge, and rejected

celibate lifestyles as inconsistent with said Vedas. Although this group may have

ultimately been successful in arrogating to themselves the identity Brahman, there

is no methodologically good reason to assume that they represented an unchanging,

200
monolithic entity that originally owned a set of uniquely Brahmanical terms that

only could be borrowed by other groups through appropriation, marketing, or

the like. To borrow from the methodological language used by Ruegg, not all aspects

of commonality between Buddhism and Brahmanism need be explain by a

borrowing model; some can be explained in terms of a substratal model. 415

Indeed, there are good reasons to assume a substratum behind the classical

Brahmanical and Buddhist traditions since early Buddhist and Brahmanical texts

appear to derive, at least in part, from a common social, cultural, and linguistic world,

with a common vocabulary that goes far beyond any technical vocabulary that can be

reasonably assigned to one sect or another. 416 Geoffrey Samuel refers to this

substratum, this common world underlying the Brahmanical and Buddhist

traditions, as the generic Indo-Aryan cultural tradition 417quite aptly, in my

opinion, since it reflects both the common language (Indo-Aryan) in which the ideas

of both traditions are expressed and the term by which both traditions referred to the

cultured as opposed to the uncivilized (Aryan). If there was such a substratum out of

which the Buddhist and Brahmanical traditions arose through a period of contestation

of categories and identity-formation, then we should not assume that the early

Buddhists formulated their own identity vis--vis the emerging Brahmanical

415
Ruegg, Symbiosis of Buddhism with Brahmanical Hinduism, ch. 19, Borrowing and
substratum models for religious syncretism and/or symbiosis, 105-114.
416
Several of the terms listed by Norman as borrowed by the Buddha from the Brahmans
(e.g., kamma, pua, sla) would, I would argue, fall under this categoryfairly ordinary words within
the Indo-Aryan languages that came to be interpreted in different ways by the Buddhist and
Brahmanical traditions.
417
Geoffrey Samuel, The Origins of Yoga and Tantra: Indic Religions to the Thirteenth
Century (Cambridge, England: Cambridge University Press, 2008), 99.

201
movement monolithically in one fell swoop. We must, in other words, introduce

diachrony into our account and allow for multiple trajectories in the way that the early

Buddhists contested terms that were also used by the proponents of the new

Brahmanism.

In Part III, I will examine more closely two specific words often cited as

examples of Brahmanical terms that were appropriated and redefined by the early

Buddhistsnamely, brhmaa and tevijja. These two terms are particurly apt for

demonstrating my thesis in Part III because they appear to be among the most

quintessentially and unproblematically Brahmanical of terms used in the early

Buddhist textsand thus the most amenable to a simple model of appropriation and

redefinition of Brahmanical terms for the purposes of marketing. I will show that

the uses of these two words in the early Buddhists texts not only differ from one

another, but also demonstrate individually a diachronic development in the process of

Buddhist identity-formation in which they played a role. The term brhmaa, on the

one hand, is used in a wide variety of Buddhist texts to refer to ideal Buddhists,

including the Buddha. While in many cases this is presented as an intentionally

metaphorical usage of the word meant to polemicize against real Brahmans,

certain texts that are likely to be among the earliest Buddhist texts extant refer to the

ideal person quite straightforwardly as a Brahman, and only later were interpreted by

commentators as doing so metaphoricallythus suggesting a diachronic progression

from straightforward use of the term brhmaa as an honorific to polemical uses, as

the division between real Brahmans (i.e., proponents of the new Brahmanism) on

202
the one hand and Buddhists as ramaas on the other began to solidify. The term

tevijja, on the other hand, does appear to have been a technical term proper to the new

Brahmanism that was borrowed by the Buddhists and redefined for polemical

purposes, but it appears as such only late in a long process of development in

Buddhist conceptions of meditation and the Buddhas Awakeninga process that

appears to have roots in a substratum shared with the classical Brahmanical tradition.

Thus, by employing both models of direct borrowing and of a common substratum,

we can see how terms shared by the Brahmanical and Buddhist traditions were

contested differently, not only from one another, but also individually over the course

of time.

203
 Chapter III.1

Early and Classical Uses of the Word Brhmaa in Early

Buddhist Texts

III.1.1 Classical Uses of the Word Brhmaa in Early Buddhist Literature as Polemic

The curious use of the word brhmaa to refer to the Buddha or the ideal

Buddhist person in general is arguably the centerpiece of the theory that the authors

of the early Buddhist texts (or the Buddha himself) engaged in marketing by

appropriating Brahmanical terms and redefining them with new Buddhist meanings.

Brhmaa, after all, would appear on its face to be the most quintessential of all

Brahmanical terms, and thus any use of the term to refer to anyone other than a real

Brahmani.e., a Brahman by birthself-evidently ironical or metaphorical. Such a

reading of the Buddhist use of the word brhmaa of course assumes a clear

dichotomy between Brahman and non-Brahman, such that one can assume an

unambiguous distinction between literal and non-literal uses of the term. Gombrich

articulates his position in terms of this assumption quite succinctly:

The Buddha was not a brahmin in the literal sense, i.e. born as one, but the
Sutta Piaka contains several passages in which he argues that brahmin,

204
 properly understood, is not a social character but a moral one, referring to a
person who is wise and virtuous. 418

Although Gombrich concedes that the Buddha was engaging in a reasoned

argument over the meaning of the word brhmaathat, in other words, the term

was contestedthe characterization of the social understanding of the word as

literal priviliges that understanding over the one presented in the Buddhist texts and

implies that it is prior and even more legitimate. Bailey and Mabbett articulate this

assumption even more directly in reference to what they call the almost obsessive

mapping of Buddhist teachings upon the structure of brhmaical tradition, 419 of

which the term brhmaa is a central example:

A heterodox or minority tradition needs to relate itself to the orthodox or

mainstream practice; almost necessarily, Buddhism mapped itself upon a
structure supplied by the Brahmins, defining itself by reference to what it was
not; a series of systematic oppositions identified its relationship to the pre-
existing orthodoxy. 420

But as I have argued forcefully in Parts I and II, we should not assume that the

Brahmanism in whose context Buddhism arose was a pre-existing orthodoxy; rather,

Brahmanical orthodoxy grew slowly over a period of many centuries and was still in

a state of flux at the time that Buddhism arose. If we adopt the methodology

advocated in this dissertation and drop the assumption that there was a clearly-defined,

pre-existing Brahmanical orthodoxy prior to the rise of Buddhism, how does this

418
Gombrich, How Buddhist Began, 20.
419
Bailey and Mabbett, Sociology of Early Buddhism, 123.
420
Ibid., 121.

205
affect our understanding of the use of the word brhmaa to refer to the Buddhist

ideal in early Buddhist texts?

In order to answer this question, we must first acknowledge that there are

many cases in which the use of the word brhmaa in the early Buddhist texts does

clearly imply a distinction between the meaning understood by the advocates of the

new Brahmanism and that being presented as the word of the Buddha. The most

explicit examples are those in which the Buddha explicitly contrasts the

Brahmanical understanding of brhmaa with his own in narrative format. An

example of this is found in a sutta of the Sutta Nipta (1.7), in which the Buddha

encounters a Brahman named Aggikabhradvja who is in the midst of making a fire

offering. This sutta, at least in the form in which it comes down to us, with a short

prose narrative frame encompassing the verses that make up the bulk of the narrative,

can be considered a clear example of an encounter dialog, since it involves an

encounter between the Buddha and an interlocutor, with the Buddha identified as a

samaa and the interlocutor as a brhmaa. In this particular encounter dialog,

however, the very binary by which this genre of early Buddhist narrative is defined is

problematized by the content of the Buddhas teaching within it. When the Buddha

approaches the Brahmans house, the latter verbally abuses him, saying, Stop right

there, muaka! Stop right there, samaaka! Stop right there, vasalaka! 421 I have

refrained from translating the three terms of abuse since their connotation and

denotation are nearly impossible to simultaneously translate into English. All three

421
tatreva, muaka; tatreva, samaaka; tatreva, vasalaka tihh ti.

206
use the diminutive ka suffix to convey a dismissive attitude toward the Buddha. The

first term, mua, simply means shaved, and thus refers to the fact that the Buddha,

as a wandering mendicant, has shaved his hair. (Often muaka is therefore

translated as shaveling.) The second term, samaa, refers to the Buddhas identity

as a samaa, here clearly intended as an insult. Finally, the third term, vasala (Skt.

vala), is itself intrinsically a term of abuse; it refers to a little or contemptuous

manalthough Norman translates it as outcaste, which seems like an appropriate

translation given the way it is used later in the sutta. 422

The Buddha responds in such a way as to turn the the Brahmans insult on its

head by ethicizing the last of these three terms of abuse. Over the course of twenty

verses (Sn. 116-135), the Buddha describes various types of morally inferior people

thieves, murderers, liars, and the likeand equates them with the outcaste or

vasala with the words ta ja vasalo iti (He is to be known as a vasala). He

then concludes by saying,

Not by birth does one become a vasala; not by birth does one become a
brhmaa.
By action does one become a vasala; by action does one become a
brhmaa. 423

In order to illustrate this point, the Buddha concludes by recounting the case of a

certain cala named Mtaga who, in spite of his social status as a cala, was

422
Norman, Group of Discourses, 14.
423
Sn. 136: na jacc vasalo hoti, na jacc hoti brhmao / kammun vasalo hoti, kammun
hoti brhmao //.

207
reborn in brahmaloka (brahmalokpago) because of his meritorious deeds. 424 The

Buddha notes further that in spite of their high birth and knowledge of the Vedas,

some Brahmans perform evil deeds and receive the due recompense in their next life.

Thus, in this sutta the Buddha redefines brhmaa by first ethicizing vasala, the

socially-loaded term of abuse with which the Brahman reviled him, and then by

logical extension ethicizing the term brhmaa itself.

Another encounter dialog presents a similar example in which the Buddha

ethicizes the term brhmaa, except that here the Brahman interlocutor himself

participates in the ethicization. In the Soadaa Sutta (DN 4), a Brahman named

Soadaa (Dog-Stick 425) living Camp, which had been given to him by King

Bimbisra of Magadha as a brahmadeyya land grant, visits the Buddha after hearing

that he is in the area and is asked by the latter what it is that makes one a Brahman. 426

Soadaa answers by saying that there are five characteristics a Brahman must have:

424
A parallel to this story of Mtaga can be found at MBh. 13.27-29, which is cited in David
Gordon White, Myths of the Dog-Man (Chicago: University of Chicago Press, 1991), 250 n. 34. In
this version of the story, a person named Mtaga learns that he is a cala because he was born from
the union of a brhma and a dra barber. Indra appears to him and offers him a boon, so Mtaga
asks him to make him a Brahman. Indra says that this is impossible, but Mtaga persists in his desire
to become a Brahman and performs great austerities in the hopes that this will allow him to become
one. In the end, Mtaga gives up his quest to become a Brahman and instead asks for the ability to
assume any form, fly, and enjoy whatever pleasure he wishes. Indra grants his request by making him
a god. The ultimate point of this story is that Brahmanhoodwhich Mtaga never does attainis
even higher than this high state that Mtaga attains through his perseverance in austerity.
425
As will be discussed further in Part IV, Brahmans in encounter dialogs often have
humorous names.
426
Before the actual conversation between Soadaa and the Buddha begins, there is a long
narrative recounting how Soadaa came to hear of the Buddha and his fame, decided to go visit him,
and looked for and saw the 32 marks of a Great Man (mahpurisa), all of which are formulaic
features of a large subset of the encounter dialogs. These will be discussed in greater detail in Part
IV.

208
 1. He is well-born on both sides, on his mothers and his fathers, of pure
descent up to the seventh generation of ancestors, undisturbed, blameless
with respect to the matter of birth. 427
2. He is a scholar, a bearer of the mantras, perfected in the three Vedas
together with their vocabularies and rituals, with their phonology and
etymology, and the oral tradition (itihsa) as a fifthskilled in philology
and grammar, not lacking in the Lokyata and marks of a Great Man. 428
3. He is handsome, good-looking, pleasing (to look at), endowed with
supreme beauty of complexion, having the complexion of Brahm and the
body of Brahm, having an appearance that is not small to behold. 429
4. He is endowed with morality, of increasing morality, endowed with
increasing morality. 430
5. And he is wise, intelligent, the first or the second of those holding out the
sacrificial ladle. 431

Clearly, these characteristics, especially the first, second, and fifth of them, are

central to the definition of the Brahman advanced by the proponents of the new

Brahmanism. The narrative continues, however, with the Buddha asking if any of

these characteristics can be set aside as inessential. Soadaa, much to the dismay

of his fellow Brahmans who have come along to witness his meeting with the Buddha,

one-by-one admits that the first three characteristics are inessential. When the

Buddha asks him if either of the other two can be eliminated, Soadaa answers no,

427
ubhato sujto hoti mtito ca pitito ca, sasuddhagahaiko yva sattam pitmahayug
akkhitto anupakkuho jtivdena.
428
ajjhyako hoti mantadharo tia vedna prag sanighaukeubhna
skkharappabhedna itihsapacamna padako veyykarao lokyatamahpurisalakkhaesu
anavayo.
429
abhirpo hoti dassanyo psdiko paramya vaapokkharatya samanngato
brahmava brahmavacchas akhuddvakso dassanya.
430
slav hoti vuddhasl vuddhaslena samanngato.
431
paito ca hoti medhv pahamo v dutiyo v suja paggahantna.

209
for wisdom is purified by morality; morality is purified by wisdom. 432 The

Buddha then gives a long sermon that defines wisdom and morality in Buddhist terms.

Thus, the overall effect of the dialog is to take a Brahmanical understanding of the

word brhmaa and redefine it in Buddhist terms, first by eliminating those

characteristics deemed totally inessential (birth, knowledge of the Vedas, and

appearance), and then explicating the remaining two in terms proper to the dharma

advanced in the early Buddhist texts.

While the Soadaa Sutta rather subtly navigates from a definition of

brhmaa that derives from the agenda of the new Brahmanism to an ethicized one

based on Buddhist teachings, yet another encounter dialog, the Vseha Sutta (MN

98=Sn. 3.9), presents the contrast between these two definitions in much more stark

relief. This sutta begins with two Brahman students (mava) named Vseha and

Bhradvja debating over which of precisely these two definitions of the word

brhmaa is correct. Here the new Brahmanical position, attributed to Bhradvja,

is reduced to the single criterion of birth:

If he is well-born on both sides, on his mothers and his fathers, of pure

descent up to the seventh generation of ancestors, undisturbed, blameless with
respect to the matter of birth, then he is a Brahman. 433

Vseha, on the other hand, takes the position that morality defines the Brahman:

And if he is endowed with morality and endowed with vows, then he is a

432
slaparidhot hi pa; paparidhota sla.
433
yato kho, bho, ubhato sujto hoti mtito ca pitito ca sasuddhagahaiko yva sattam
pitmahayug akkhitto anupakkuho jtivdena, ettvat kho bho brhmao hot ti.

210
Brahman. 434 Neither being able to convince the other, they go to the Buddha to

settle the dispute. The Buddha does so in an extended sermon in verse that places

him firmly on the side of Vseha. He begins by noting that while various species of

plants and animals have marks (liga) to distinguish them from one another, human

beings do not (v. 600-611). Then he lists a variety of occupations, from farmer to

king, and states that each such person should be known by his occupation and not as a

Brahman (v. 612-619). The obvious implication is that people who call themselves

Brahmans often undertake occupations that are not proper to their vara. Therefore,

the Buddha does not acknowledge Brahmanhood simply on the basis of birth (v. 620).

With this, the Buddha launches into a 28-verse description (v. 620-647) of the

qualities that do characterize a Brahman. The virtues extolled in each of these verses

describe a person who has attained Awakening according to the generally-accepted

tenets of the early Buddhist textshe is free of sensual pleasures, is endowed with

wisdom, is non-violent, has put an end to rebirth, does not steal, is purified, is without

possessions, has destroyed the savas, and is even explicitly called an arhat.

Moreover, each of the verses follows a formula: They each end with the words tam

aha brmi brhmaa (Him I call a Brahman)with the emphatic nominative

pronoun aha included, even though it is grammatically unnecessary, to emphasize

the contrast between the Buddhas definition of a Brahman and the definition based

on birth. The sutta ends with the Buddha summarizing his argument (v. 648-656) that

434
yato kho, bho, slav ca hoti vatasampanno ca, ettvat kho, bho, brhmao hot ti.

211
designations based on birth are mere social conventions and that therefore one truly

becomes a Brahman by action and not by birth. 435

Aside from these three major discourses that present in explicit and extensive

detail the contrast between the Brahman by birth and the Buddhist conception of the

Brahman, casual references to the Brahman as an ideal are found scattered throughout

various texts of the Pali Canon, usually in verses. Many of these verses are collected

in the last vagga of the Dhammapada and the first vagga of the Udna. While some

of these cannot be unambiguously interpreted as contasting a Buddhist conception of

Brahman to a Brahmanical conception of Brahman, 436 at least one Dhammapada

verse explicitly contrasts these two conceptions of the brhmaa within the verse

itself:

Not by matted hair, not by gotra, not by birth does one become a Brahman.
In whom there is truth and dharma, he is pure and he is a Brahman. 437

Verses in the Udna are given prose frame stories, and these frame stories often serve

to make explicit a contrast between the Brahman by birth and the Buddhist

conception of the Brahman found in the verse. Thus, for example, in Ud. 1.4, a
435
Verse 650 in the Sn. version of the Vseha Sutta, in fact, states so in exactly the same
words as v. 136 in Sn. 1.7 quoted above.
436
This would include Dhp. 396-423, which are identical or nearly identical to v. 620-647 of
the Sn. version of the Vseha Sutta. These are the verses, already discussed above, that end with the
refrain tam aha brmi brhmaa (Him I call a Brahman), already discussed above. As I will
discuss in Part IV, there is good reason to believe that these verses were incorporated into the
Dhammapada tradition first, and only later borrowed from there for use in the Vseha Sutta, and then
only by the Theravda tradition. If this is true, then the verses that do not explicitly contrast the
Brahman as ideal with the Brahman as someone born that way were imputed with such a
contrastive meaning by the narrative framework of the Vseha Sutta, in the same manner as I will
discuss in Chapter III.2.
437
Dhp. 393: na jahi na gottena, na jacc hoti brhmao / yamhi sacca ca dhammo ca, so
suc so ca brhmao //.

212
Brahman of the huhuka type (huhukajtiko) 438 approaches the Buddha and

asks him what makes one a Brahman, and the Buddha answers with the following

verse:

He who is a Brahman, in the state of sin warded off,

Not a hu-hu-er, 439 without savas, self-controlled,
Gone to the end of the Veda, having lived brahmacariya,
Should rightly proclaim the brahma-doctrine,
Who has no distinguishing qualities anywhere in the world. 440

Likewise, in Ud. 1.9, which we already looked at in chapter II.3, the Buddha

encounters a group of jailas at Gay performing water ablutions and fire offerings,

and he comments on the scene in the following verse:

One is not purified by water, (yet) many people bathe here.

In whom there are truth and dhamma; he is pure and a he is a Brahman. 441

As we will see below, references to the word brhmaa are often situated in context

that draws out an explicity contrast between a literal Brahman by birth and a

Buddhist conception of the Brahman through the use of just such prose frame

narratives. 442 What is important to note for the moment, though, is that there do exist

texts within the Pali Canon that, either through extended narrative or through

438
This is apparently a reference, perhaps somewhat snide, to the practice of reciting Vedic
mantras.
439
This is presumably a reference to the Vedic Brahmans practice of reciting mantras.
440
yo brhmao bhitappadhammo / nihuhuko nikkasvo yatatto / vedantag
vsitabrahmacariyo / dhammena so brahmavda vadeyya / yassussad natthi kuhici loke // .
441
na udakena suc hot, bahvettha nhyat jano / yamhi saccaca dhammo ca, so suc so ca
brhmao.
442
This includes, ironically, Dhp. 396-423, which, in addition to their likely narrative framing
through incorporation into the Vseha Sutta (see Chapter III.2), are also given humorous (and
unrelated) frame stories involving silly Brahmans in the commentary on the Dhammapada itself.

213
contextual framing, do set up a clear contrast between the Brahman defined by birth

and the Brahman defined by deeds, and to that extent therefore are well-explained by

a model of polemical marketing.

I will argue in this and the next chapter, however, that a synchronic

interpretation of all references to the ideal person as a brhmaa in the early Buddhist

texts as an intentional punning on a literal meaning based on birth, such as we find

in the texts just cited, is insufficient. In other words, I will show that there is

evidence of a diachronic development in the treatment of the ideal person as

brhmaa. At first, brhmaa was applied to the Buddha or other ideal being simply

as an honorific drawn from the Indo-Aryan cultural substratum. Later on, as the

proponents of the new Brahmanism increasingly asserted their prerogative to the term

and the Buddhists developed a sense of identity as separate from them, the latter

increasingly came to interpret the old Buddha-as-Brahman or arhat-as-Brahman trope

in polemical contrast with the Brahman-by-birth. In order to make this argument, I

will begin by examining the use of the word brhmaa in what are considered by

many scholars to be (at least in part) the earliest Buddhist texts extantthe Ahaka

and Pryaa of the Sutta Niptaarguing that there is no intrinsic reason to

interpret them as a marketing ploy or anything other than simple honorifics. Then,

I will show how these early uses of the word brhmaa were brought into the

standard canonical narrative of the true Brahman through narrative framing and

commentarial glossing, and I will point to other places in the early Buddhist literature

where similar devices are found. Thus, I will argue that brhmaa was brought into

214
the early Buddhist vocabulary out of the common stock of Indo-Aryan linguistic

culture, and only over the course of time did the trope of the true Brahman in

contrast to the Brahman by birth develop.

III.1.2 Antiquity and Uniqueness of the Ahaka and Pryaa

Although the likely antiquity of the Ahaka and Pryaa, preserved as the

fourth and fifth vaggas of the Pali Sutta Nipta, respectively, is fairly widely

recognized, it is useful to review the reasons for believing that they are particularly

old since those reasons are not universally accepted. The reasons are fourfold:

metrical considerations; the fact that the two collections are commented upon in the

canonical Niddesa; the fact that they are referred to by name elsewhere in the Canon;

and the unique, frankly unorthodox doctrines found within them. None of these

reasons taken in isolation is sufficient to unequivocally establish the antiquity of these

two collections, but taken together, I believe that they make it difficult not to come to

the conclusion that they are among the oldest Buddhist texts extant.

Let us begin with meter since it is perhaps the most contentious of the reasons

for believing the Ahaka and Pryana to be old. Some scholars have argued that a

significant portion of the gths found either in isolation or embedded in prose

throughout the Canon (Sutta Nipta, Sagth Vagga of the Sayutta Nikya, certain

Jtakas, Itivuttaka, Udna) are particularly old and represent an early stage in the

215
development of Buddhism, 443 although the relationship between prose and verse

portions of the early Buddhist texts is complicated and thus difficult to generalize.

More detailed analyses can be and have been made, however, of the relative datings

of various gths on the basis of the types and and characteristics of meters used

therein; the most significant contribution to scholarship in this regard is A. K.

Warders monumental Pali Meter. Although he notes that the question is

complicated by the fact that there are likely two strata in the Ahaka and Pryana

(about which more below) that need to be distinguished for the purposes of metrical

analysis, 444 Warder argues that the vatta (Skt. vaktra, equivalent to the Sanskrit

anuubh and precursor to the classical loka) meter found in the Ahaka and

Pryana is more archaic than the vatta found in other texts such as the

Dhammapada, Thera/Thergth, and Jtaka. 445 Others have been critical of the

metrical analysis approach to dating, such as K. R. Norman, who states bluntly,

Dating by metre is not particularly helpful. 446 Nevertheless, Norman accepts the

relative antiquity of the Ahaka and Pryana on other grounds 447 and acknowledges

443
For a description of this view and the picture of early Buddhism it claims to disclose, see
Hajime Nakamura, Indian Buddhism: A Survey with Bibliographical Notes (Delhi: Motilal Banarsidass,
2007), 57-60. But see also Bhikkhu Sujto, A History of Mindfulness: How Insight Worsted
Tranquility in the Satipahna Sutta (unpublished manuscript), 18-20, for a critique of what he calls
the Gth Theory.
444
A. K. Warder, Pali Meter (London: Luzac, 1967), 199.
445
Ibid., 172-3. Warder bases his argument for the relative antiquity of the vatta in the
Ahaka and Pryana on the frequency with which the pathy form of the prior pda is used. The use
of the pathy form shows a clear increasing progression from 20% in Book 10 of the RV to 93% in the
Raghuvaa.
446
Norman, Group of Discourses, xxix.
447
Ibid., xxvii.

216
that this antiquity is supported at least in part by the fact that the original core of

verses in the Ahaka-v, is in the Triubh [Pali tuhubha] metre, which is generally a

sign of an early composition in Pli. 448

The second and most frequently cited reason for believing that the Ahaka

and Pryaa are quite old is that they are commented upon by a canonical text

called the Niddesa. Although the Niddesa is not strictly speaking the only

commentary found within the canonical texts of the Pali tradition, 449 it is fairly

unique as a self-contained, free-standing canonical text 450 that serves soley as a

technical commentary on other canonical texts. That the Niddesa is an early

commentary on even older texts is confirmed by the fact that the Ahaka and

Pryaa, as parts of the Sutta Nipta, are also commented upon by the ordinary

post-canonical commentary, Paramatthajotik II, which presupposes the existence of

the Niddesa and even quotes directly from the Mahniddesa. 451 The particular

antiquity of the Ahaka and Pryaatogether with the Khaggavisa Sutta (Sn.

1.3)is moreover confirmed by the fact that the Niddesa only comments upon those

448
Ibid., xxix.
449
This was pointed out to me at the American Academy of Religion Conference in Atlanta,
GA, 2010, by Bhikkhu Sujto, who is suspicious of arguments for the particular antiquity of the
Ahaka and Pryaa. He cited as other examples of canonical commentaries the Saccavibhaga
Sutta (MN 141), in which the Buddha gives a sermon expounding upon his first sermon that he had
given in the Deer Park at Isipatana after his Awakening (i.e., the Dhammacakkappavattana Sutta), and
the Suttavibhaga of the Vinaya, which is a commentary on the Pimokkha Sutta (which latter,
ironically, is not in and of itself canonical). See also Sujto, History of Mindfulness, 19, where he
criticizes Schopen on this point.
450
Actually, two texts: the Mahniddesa contains the commentary on the Ahaka-vagga and
the Caniddesa contains the commentaries on the Pryaa-vagga and the Khaggavisa Sutta.
451
Norman, Group of Discourses, xxxvii.

217
three texts, rather than the Sutta Nipta as a whole. This would suggest that the

Khaggavisa Sutta, Ahaka, and Pryaa were circulating as independent texts at

the time that the Niddesa was written, prior to their incorporation into the Sutta

Nipta. In the case of the Pryaa, however, the form that it took at that time was

most likely somewhat shorter than the form that it takes now in the Sutta Nipta,

since the introductory vatthu-gths found in the latter (Sn. 976-1031) are not

commented upon by the Caniddesa.

The third reason for believing in the antiquity of the Ahaka and Pryaa is

closely related to the second: Both of these collections are referred to by name

elsewhere. Within the Pali Canon there are six independent references that all treat

the collections as unified, titled works with recognized divisions within them.

According to Ud. 5.6, a certain Ven. Soa recited with melody 452 all sixteen

divisions of the Ahakavagga. 453 Five other references cite specific verses from

specific sections of one of the collections: v. 844 in the Mgandiyapaha of the

Ahakavagga (SN 3.1.22.3), v. 1038 in the Ajitapaha of the Pryaa (SN

2.1.12.31), v. 1048 in the Puakapaha of the Pryaa (AN 3.32, 4.41), v. 1106-7

in the Udayapaha of the Pryaa (AN 3.32), and v. 1042 in the Metteyapaha of

452
Here I follow Masefields translation of sarena (lit., with sound) as melodically
Peter Masefield, trans., The Udna (Oxford: Pali Text Society, 1994), 105.
453
yasm soo soasa ahakavaggikni sabbneva sarena abhai. This reference is
reproduced at Vin. 4.5.158 with the number 16 (soasa) omitted. Vetter has used this omission to
support his theory that the Ahaka is a composite textTillmann Vetter, Some Remarks on Older
Parts of the Suttanipta, in Earliest Buddhism and Madhyamaka: Panels of the VIIth World Sanskrit
Conference, ed. Schmidthausen and Ruegg (Leiden: Brill, 1990), 43.

218
the Pryaa (AN 6.61). 454 The recognition of the existence of the Ahaka and

Pryaa as independent collections is not limited to the Pali tradition, however;

aside from the canonical references in Pali just cited, we also find references to an

Arthavargya and Pryaa twice in the Sanskrit Divyvadna. 455 Now, the Ahaka

and Pryaa are certainly not unique as canonical texts that are cited elsewhere in

other early Buddhist texts, but the fact that they are cited several times and with

remarkable precision as to their title and internal divisions demonstrates that they

were definitely recognized and popular collections at a relatively early date. In

addition, as Norman has noted, while several suttas or portions thereof from the first

three vaggas of the Sutta Nipta are reproduced elsewhere in the Canon, none of the

texts from the Ahaka-vagga (Sn. 4) or Pryaa-vagga (Sn. 5) are. This would

seem to imply, he argues, that these two vaggas were regarded as a whole at the

very earliest period of Buddhism, and had already been given a status of original and

indivisible. 456

The fourth and final reason for regarding the Ahaka and Pryaa as

particularly old is the unique ideas found within them that do not accord well with the

454
Cited in Norman, Group of Discourses, xxxiv-xxxv.
455
Cited in Ibid., xxviii. The Divyvadna is somewhat later than most of the canonical Pali
texts, dating to the early centuries of the Common EraRotman, Divine Stories, 1. The relatively late
date of the Divyvadna may be taken as an indication of the enduring status of the Arthavargya and
Pryaa as independent collections; while Normans caution that the list in which the two are
mentioned may be older than the Divyvadna as a whole is well taken, we should keep in mind that
the Pali tradition is the only tradition known to have incorporated them into a larger work (i.e., the
Sutta Nipta). The Sutta Nipta has no known counterpart in any other tradition, and the Arthavargya
was, like several other texts of preserved in the Khuddaka Nikya by the Pali tradition, translated into
Chinese as an independent text (T. 198).
456
Ibid., xxxi-xxxii.

219
standard doctrines found throughout the more mainstream texts of the Pali Canon

(and corresponding Chinese gamas). The details lying behind this generalization

are fairly complex and directly relevant to the argument I will be making in Part III,

so it will behoove us to review them at length. The classic scholarly work addressing

the unusual doctrines of the older texts of the Sutta Nipta is Luis Gmezs 1976

article Proto-Mdhyamika in the Pli Canon. 457 In this article, Gmez argues that

the ideas found in the Ahaka, and to a certain extent the Pryaa, bear more in

common with Madhyamaka philosophy than anything found in the rest of the Pali

Canon. In particular, Gmez argues, the Ahaka does not proclaim a path based on

right view, but rather a much more radical path based on abandoning attachment to

any view whatsoever. The contrast with mainstream canonical doctrine is marked:

Stock phrases which in the Canon were used to indicate the highest knowledge, such

as jnmi passmi and a, were used here to indicate the false science of those

who were still attached to views. 458 Gmez rightly emphasizes that the Ahaka

advocates abandoning all views, not just some (i.e., wrong views), thus putting it in

opposition not only to standard canonical doctrine, but even to other parts of the Sutta

Nipta. 459

457
Luis O. Gmez, Proto-Mdhyamika in the Pli Canon, Philosophy East and West 26, no.
2 (Apr. 1976): 137-165. According to Gmez, it was his intention for there to be a question mark at
the end of the article title, but it was omitted inadvertently in the course of publishing.
458
Ibid., 139-40.
459
Ibid., 140-1. Gmez cites v. 800, 803, and 900 as particularly exemplifying the total
rejection of all views.

220
 Given the incompatibility of this radical rejection of all views with the

canonical distinction between right and wrong view, it appears that the Ahaka

posed a hermeneutical problem early on in the Buddhist tradition. Verse 790, for

example, reads as follows:

A Brahman does not say that purity is from something else, in what is seen, in
what is heard, in morality and vows, or in what is thought.
[He is] unsmeared with respect to virtue and sin, rejecting what has been taken
up, not doing [anything] in this connection. 460

Gmez interprets aato as I have translated it hereas referring to any other thing

that purity could rely on. This is not the interpretation of Mahniddesa, however; it

interprets aato as referring to another (i.e., non-Buddhist) dhammaan

interpretation that Gmez rejects:

The Mahniddesa fails to understand the true purport of this passage when it
glosses: If a man were made pure ... by another, impure path, by a false
path ... other than the Noble Eightfold Path . The very context of the
whole poem (788-795. Aha section iv), shows that the view under attack is
that of him who relies on knowledge (pacceti a) about things seen, heard
or thought. 461

Although Gmez does not refer to it, we can add that the Chinese version of the

Ahaka (T. 198), which we will be discussing in more detail shortly, appears to make

the same error in interpretation in translating the first pda as

Following another path, one does not obtain liberation. This tendency of the later

tradition to bring the Ahaka (and Pryaa) into line with more mainstream doctrine

460
na brhmao aato suddhimha, dihe sute slavate mute v / pue ca ppe ca
anpalitto, attajaho nayidha pakubbamno //.
461
Gmez, Proto-Mdhyamika, 141.

221
and usage, as evinced by both the Niddesa and the Chinese translation of the Ahaka,

will play an important role in my argument about the development in the use of the

word brhmaa.

The purpose of abandoning all views in the Ahaka is to end the proliferation

of sa, or, as Gmez translates, apperceptions, which lead to a false dualistic

view of the world, which in turn leads to passion and aversion, which in turn lead to

suffering. Thus, for example, v. 874 of the Kalahavivda Sutta in the Ahaka states

that, as Gmez renders it, form is made to cease by means of the control of

apperception, for dispersion with conception have apperception for their

cause. 462 Likewise, the Mgandiya Sutta of the same collection identifies

apperception as the crux of the problem of the human condition, which latter it

characterizes in terms of conflict:

For him who is detached from apperceptions there are no knots; for him who
is released by insight there are no delusions.
Those who hold on to apperceptions and views go around in the world
knocking about. 463

The Pryaa also speaks of the goal in terms of putting an end to apperception:

Free of passion with respect to all sensual desires,

Relying on nothingness, abandoning another,
Released in the ultimate release from apperception, would he remain there,
not following [other meditative objects]? 464

462
Gmez, Proto-Mdhyamika, 143. The verse in question reads as follows: na
saasa na visaasa, nopi asa na vibhtasa / evasametassa vibhoti rpa,
sanidn hi papacasakh //.
463
Sn. 847: savirattassa na santi ganth, pavimuttassa na santi moh / saa ca
dihi ca ye aggahesu, te ghaayant vicaranti loke //; cited by Gmez, Proto-Mdhyamika, 145.
464
Sn. 1071: sabbesu kmesu yo vtargo / kicaa nissito hitv maa /
savimokkhe parame vimutto, tihe nu so tattha annuyy //; cited by Gmez, Proto-Mdhyamika,

222
Gmez emphasizes that the nothingness (kicaa) referred to in this verse

cannot refer to the fourth arpa-jhna of the classical system, but is rather the very

core and apex of the path, 465 which is also conceived of in terms of release from

apperception. According to the classical system of arpa-jhnas, on the other hand,

the goal lies beyond the jhnas, in the cessation of consciousness (via-nirodha).

But as Gmez notes, the translation consciousness does not work well for via

in the Pryaahe prefers Hares translation of mind at workand the

Pryaa abstains from asserting the cessation of the via, and actually speaks of

a release from apperception (savimokhe). 466 At least parts of the Ahaka and the

Pryaa therefore appear to represent an early form of Buddhist doctrine quite

different from the classical form found in more mainstream canonical texts, and

whose radicalness was blunted as these more mainstream doctrines developed.

Scholars since Gmez have debated the exact implications of the unusual

doctrinal statements found in the Ahaka and Pryaa for the dating of the

collections. Tillmann Vetter, for one, has, while accepting Gmezs general

conclusions regarding the incompatibility of certain passages in the collections with

mainstream canonical doctrine, criticized him for coming to broad conclusions about

the collections based on those passages. He notes that there remain six out of the

sixteen suttas of the Ahaka that are not used by Gmez for his argument, and could

144. For a detailed discussion of the translation of this verse (which I have not followed exactly), see
Wynne, Origin of Buddhist Meditation, 78-84.
465
Gmez, Proto-Mdhyamika, 144.
466
Ibid., 145.

223
not be used because they do not advise overcoming apperception or abandoning

views and the like. 467 After a thorough analysis of each of the 16 suttas of the

Ahaka, Vetter concludes that only seven lack any approval of tenets known to

essentially belong to the Buddha's teaching from other parts of the canon and

therefore fit Gmezs argument: Suddhahaka (4), Paramahaka (5), Pasra (8),

Mgandiya (9), Kalahavivda (11), Cabyha (12), and Mahbyha (13). 468

Vetter notes of these suttas, which he refers to as the core of the Ahaka, do

not tend to conceive of the goal in terms of release from rebirth, unlike the Pryaa,

where the latter is a conspicuous theme. 469 The Purbheda Sutta (10), which is

not in the core, refers to becoming (bhava) and non-becoming (vibhava) only to reject

both. Vetter therefore hypothesizes that the core suttas of the Ahaka were composed

by a non-Buddhist mystical movement that was later incorporated into the Buddhist

sagha:

In the oldest days this mystical movement faced only groups of persons who
strove for a better existence after this life, not groups of persons who strove
for release from rebirth. At some moment this latter goal, not being too far
from renouncing the strife for better existences, was recognized by these
mystics, but nevertheless they rejected striving for it. 470

He continues,

[T]his group of mystics was integrated into the Gotama-sagha and tried to
adapt specific Buddhist methods and aims as is the case in other [i.e., non-core]

467
Vetter, Older Parts of the Suttanipta, 45.
468
Ibid., 46.
469
Ibid., 49.
470
Ibid., 50.

224
 suttas of the Ahaka. On the other hand, the Gotama-sagha attempted to
include (mainly in the sense of Paul Hacker's term inclusivism) the main
subject of this movement or to make use of it. 471

Vetter argues that the Pryaa was one of the first attempts to include the themes

taught by the mystical movement behind the core suttas of the Ahaka, insofar as it

refers to release from apperception and overcoming all views, but relates them to the

goal of escaping birth and old age, as well as such other Buddhist themes as

mindfulness and thirst. 472 As an early exercise in inclusivism, the Pryaa was,

according to Vetter,

a text which was composed by a person who wanted to mention as many

tenets or methods as he knew and as many as were necessary to complete the
solemn number of sixteen questions. All is related to the Buddha and, very
likely, to the aim of overcoming rebirth. 473

Thus, according to Vetters model, the Ahaka and Pryaa, while both quite early

relative to the rest of the Canon, are composite texts with complex internal histories

of their own.

This view has come under criticism from Alexander Wynne. Wynne

acknowledges that there may be different strata within the Pryaa, but he argues

that Vetter is too quick to identify incompatibilities between passages in different

sections of the Pryaa that allow him to reduce the text as a whole to an

inclusivistic exercise in compilation. Wynne argues that the Pryaa is, at least in

part, a very old text, and that as such it can be used to show that the Buddha taught a

471
Ibid., 51.
472
Ibid.
473
Ibid., 42.

225
form of meditation similar to what Vetter calls Dhyna-meditation, a meditative

practice based on the goal of ra Klma that was thought to lead to a non-

intellectual sort of insight. 474 The Dhyna-meditation referred to here is the full

classical system of four rpa- and four arpa-jhnas, which we will discuss in further

detail in Chapter III.3. For now, it suffices to note that since Gmezs classic article

on the disconnect between the doctrines of the Ahaka and Pryaa on the one hand

and those of other more mainstream canonical texts on the other, there has been

debate over the exact degree to which these unique doctrines pervade the two

collections and the exact way in which they are related to canonical doctrine.

A somewhat different approach to the problems addressed by Gmez, Vetter,

and Wynne has been taken by Grace Burford in her book-length study of the Ahaka-

vagga. Like the three former scholars, Burford accepts the antiquity of the Ahaka

and therefore its value for developing a better understanding of early Buddhism. 475

Like Gmez, she sees the doctrines found in the Ahaka as fundamentally different

from those found elsewhere in the Canon. In fact, she goes further and argues that the

orthodox Theravda doctrine drawn from the majority of canonical texts is not fully

coherent, but that the Ahaka on the other hand does present a coherent normative

value theory effectively and completely without the definition of kamma in terms

of death and rebirth, and without the consequent implication that the highest goal

474
Wynne, Origin of Buddhist Meditation, 75.
475
Grace G. Burford, Desire, Death and Goodness: The Conflict of Ultimate Values in
Theravda Buddhism (New York: Peter Lang, 1991), 8-9.

226
entails escape from the cycle of birth and death. 476 Finally, like Vetter, Burford sees

the Ahaka as a composite text. She divides the collection into two categories of

suttas: In the first category, ideal persons are constructed as those who are not

affected by desire, while the second category of suttas take as their theme the

overcoming of attachment to all views, as already discussed. 477 Burford notes that all

of the suttas in the anti-dihi category are in the triubh meter, lending credency to

the idea that they form a coherent category in spite of being interspersed with desire

suttas. 478 This interspersion can be explained, however, by the fact that the suttas of

the Ahaka are ordered perfectly according to increasing number of verses. Burford

therefore argues that this text in its present form appears to be a combination of two

sets of suttas that have been intermixed by an editor or editors who ordered them

strictly according to the number of verses they contain. 479

Burfords approach to the Ahaka goes in a different direction from that of

Vetter and Wynne, however, insofar as she does not focus on the composition of the

Ahaka itself to extrapolate backwards to a putative earliest Buddhism; rather, she

follows Gmezs lead in criticizing the Niddesas interpretation of an Ahaka passage

476
Ibid., 5.
477
Ibid., 60. Burford places suttas 3-5 and 8-9 in the anti-dihi category and suttas 1-2 and 6-
7 in the desire category. Sutta 10 ignores the issue of dihis until its final three verses, while sutta 11
does not reflect the distinctive anti-dihi outlook in spite of its title (Kalahavivda, Quarrels and
Disputes). The 12th and 13th suttas are given over entirely to discussion of the drawbacks of theories
and views, while suttas 14 and 15 contain certain elements of the anti-dihi argument. Suttas 14
and 15 reflect the sort of understanding of the goal that appears in the desire-oriented suttas, but
sutta 16 makes no mention of the advantages or disadvantages of views.
478
Ibid., 61.
479
Ibid., 62.

227
by analyzing comprehensively the way that the two commentaries on the Ahaka

the Mahniddesa and the Paramatthajotik IIinterpret it in different ways so as to

fit its unique doctrines into Theravdin 480 orthodoxy. These two commentaries use

similar strategies for dealing with the unorthodox aspects of the Ahaka, but they

bear subtle differences because they are written in different formatsthe

Mahniddesa is strictly a word-by-word commentary, while Buddhaghosas

Paramatthajotik employs a more discursive style.

In her analysis of the Mahniddesa commentary, Burford finds that while it

retains the Ahakas preference for defining the problem at the root of the human

condition in terms of desire, 481 it complexifies the ideas presented in the Ahaka by

defining words that are used interchangeably therein in distinct, hierarchical terms. 482

In addition, it adds to the ideas present in the Ahaka by making it clear that the goal

lies beyond death and rebirth; in other words, it inserts the orthodox concept of

nibbna as escape from rebirth into a text in which it is notably absent. 483 The most

problematic aspect of the Ahaka, of course, from the standpoint of orthodox doctrine,

is the rejection of attachment to any view whatsoever; Burford shows that the

Niddesa deals with this consistently in the same way as Gmez demonstrated for v.

480
I use scare quotes here because although Burford routinely uses the word Theravda to
refer to the orthodoxy that characterizes most canonical texts as opposed to the Ahaka, even as late
a text as the Niddesa is still, as a canonical text, relatively early and is not therefore strictly speaking
Theravdin.
481
Burford, Desire, Death and Goodness, 71.
482
Ibid., 81-2.
483
Ibid., 85-93.

228
790by glossing the views that are rejected in the root text as referring to wrong

views specifically. 484 Burford argues that this hermeneutic allows the Mahniddesa

to integrate the dualistic desire-oriented teaching with the nondualistic anti-dihi

approach and thus interpret the Ahakavagga as presenting one coherent teaching,

but in so doing it serves to undermine [the anti-dihi portions] originally

uncompromising nondualism. 485

According to Burfords reading, Buddhaghosas commentary on the Ahaka

in the Paramatthajotik II serves much the same purpose as that in the

Mahniddesato interpret the Ahaka as a text that is both inwardly coherent and

consistent with the main tenets of Theravda orthodoxyalthough it is able to do so

somewhat more comprehensively because of its narrative format that allows

Buddhaghosa to simply skip over words in the root text that are problematic. Like the

author of the Mahniddesa, Buddhaghosa makes it clear, where the Ahaka does not,

that the ultimate goal is liberation from rebirth. 486 It does so by speaking of liberation

in two stages: first the extinction of desire in this life, and then escape from rebirth at

death. Burford sees this interpretationand the orthodoxy for which it serves as the

basisas nonsensical since it strains the very idea of the concept of ultimate. 487

Likewise, Buddhaghosa deals with the problematic anti-dihi teachings of the

484
Ibid., 93-107.
485
Ibid., 108.
486
Ibid., 125-131.
487
Ibid., 156.

229
Ahaka by transforming them into anti-micchdihi teachingsi.e., teachings

against wrong view. 488 In so doing, Buford argues, Buddhaghosa, like the author of

the Niddesa, imposes unity on the texts of the Ahaka and blunts the radical non-

dualism of some of those texts so as to make them compatible with orthodox teaching.

As we have seen in this section, there are several reasons that taken together

make it highly likely that the Ahaka and Pryaa are among the oldest Buddhist

texts extant. The meters used within them display certain archaic features, and both

are referred to by name in several other early texts and formally commented upon as

independent collections in the Niddesaa text written early enough to itself be

considered canonicalthus proving that they circulated as independent collections

from an early date within Buddhist communities. In addition, both texts, but

especially the Ahaka, contain doctrines that are difficult to reconcile with, and at

times downright contradictory to, the doctrines found in most other early Buddhist

texts. Gombrich has noted rightly the remarkable consistency in the teachings found

in the early Buddhist texts 489a consistency that extends to the Chinese gamas as

much as to the texts of the Pali Canonalthough contra Gombrich I would attribute

this consistency to a normativizing tendency within whatever segment of the

Buddhist sagha (possibly the Sthavira branch 490) produced the mainstream texts

of the Nikyas/gamas, rather than the putative genius of the Buddha himself.

488
Ibid., 131-140.
489
Gombrich, What the Buddha Thought, 194.
490
As will be discussed in Part IV, the Pali Nikyas and Chinese gamas, with the possible
exception of the Ekottarikgama preserved in Chinese, all came from Sthavira schools.

230
Insofar as the Ahaka and Pryaa do not conform to this normativizing tendency,

they were likely composed and accepted in much the same form as we have them

today before those normative forces came into play.

For our purposes, though, it is not important that the Ahaka and Pryaa

are literally the oldest Buddhist texts extant, nor exactly how they were formed.

Given the evidence presented here, I do think that they are among the oldest Buddhist

texts we have, although it is quite likely that there are other equally old texts that

were incorporated into and elaborated upon by the more mainstream texts of the

Nikyas/gamasand that the Ahaka and Pryaa were spared this fate because

they were perceived from an early date as, as Norman puts it, original and

indivisible. The lack of commentary on the vatthu-gths of the Pryaa in the

Caniddesa make it almost certain that they were added to an originally shorter

Pryaa text, and I think that Vetter and Burford do present compelling reasons

based on internal evidence that the Ahaka is a composite text as well. Likewise,

internal evidencenamely the lack of acceptance of escape from rebirth as a goal in

the Ahaka as opposed to its prevalence in the Pryaawould seem to suggest

that the former is somewhat older than the latter. Ultimately, however, what is

important for our purposes is that the Ahaka and Pryaa are relatively old and

were preserved in a relatively archaic state up to the present. And I will argue that as

such they provide a useful window not only to early Buddhist doctrine, but also to

early Buddhist conceptions of identity.

231
III.1.3 The Non-Polemical Use of the Word Brhmaa in the Ahaka and Pryaa

Unorthodox doctrines such as the rejection of attachment to any view are not

the only unusual aspect of the Ahaka and Pryaa; another is a certain proclivity

for referring to the Buddha or the ideal person as a brhmaa. This tendency has

already been noticed by both Gmez and Vetter. In discussing the differences

between the Pryaa and the Ahaka, Gmez notes that the Pryaa prefers to use

the word bhikkhu to refer to the ideal person, whereas in the Ahaka the word

brhmaa is more common. 491 He does not elaborate on the significance of this,

though, and he consistently translates brhmaa as true Brahman, 492 indicating that

he reads the references to the ideal person as a Brahman as being somehow

metaphorical, as in other cases in the Canon already discussed above.

Vetter also notes the prevalence of references to the ideal person as a

brhmaa in the Ahaka, and he notes further a certain degree of correlation between

those suttas that refer to the ideal person as a Brahman and those that he includes in

his reckoning of the core of the Ahaka on the basis of their doctrinal content. 493

Unlike Gmez, however Vetter does not interpret these references to the ideal person

as a Brahman as metaphorical in the same sense as found elsewhere in the Canon:

Pasra (in 828) and Cabyha (in 883-4, 890) denounce the quarrelsome
ascetics (samaa) and (only Cabyha 891-2) the sectarians (?titthy). Here

491
Gmez, Proto-Mdhyamika, 146.
492
Ibid., 140, 141, 142, 147.
493
The correlation is close, but not exact. Of the seven suttas in Vetters core, four
(Suddhahaka, Paramahaka, Mgandiya, Mahbyha) refer to the ideal person as a Brahman. Only
one other sutta outside the core (Attadaa) does so.

232
 we get the impression that the Brhmaa as the person accomplishing the
mystical way is opposed to the Samaa and not to the Brhmaa by birth, the
latter opposition being a common theme in other parts of the canon, but not in
any way alluded to here 494

Vetter notes, however, that in one sutta (v. 866 and 868 of the Kalavivda), there are

positive references to a samaa in the singular, which he takes to refer to the Buddha:

Maybe we can take this as an indication that in this circle one now had
accepted the fact that there was one Samaa who was not a "sophist" and not
quarrelsome and could be adhered to also by followers of this "Brahmanical"
mystical way 495

In other words, Vetter recognizes that the word brhmaa is used in the Ahaka in an

unusual way that does not correlate well to the metaphorical way it is often used

throughout the rest of the Canon, i.e., in explicit contrast to Brahmans by birth, but he

uses this observation to further his theory that the core of the Ahaka was written by a

non-Buddhist group that was later incorporated into the Buddhist sagha.

Before commenting further on how we are to interpret the use of the word

brhmaa in the Ahaka, as well as the Pryaa, it will be useful to review all the

verses in which the word appears. Overall, the word brhmaa appears in 12 suttas

in the Ahaka and Pryaa, six in the Ahaka (4, 5, 9, 10, 13, 15) and six in the

Pryaa (3, 4, 5, 7, 11, 14). Out of these, only four use the word in ways that are

either consistent with the later usage of the term or are otherwise irrelevant to the

argument being made here, and three of these are found in the most likely later

Pryaa. One of these is the Puakamavapucch of the Pryaa (3), which

494
Vetter, Older Parts of the Suttanipta, 46.
495
Ibid., 47.

233
refers three times to khattiyas and Brahmans who each offered sacrifices to deities

here in the world. 496 Here the reference is clearly to Brahmans as a social category,

recognized along with khattiyas as prominent individuals who perform sacrifices

(yaa). The sutta as a whole is a criticism of the practice of sacrifice, which is said

not to lead to escape from rebirth, unlike the stilling desire, which does. In this sutta,

therefore, the word brhmaa is not used to refer to a Buddhist ideal at allinstead,

it refers to a social group that is criticized.

Another sutta of the Pryaa (11) is similarly irrelevant to the argument

being made here; it uses the word brhmaa only once (v. 1100), in the vocative, to

refer to the Buddhas interlocutor. This vocative is a nod to the narrative frame of the

Pryaa, at least as we have it today, in which the suttas of the text are said to be

short dialogs between the Buddha and each of 16 Brahmans who approached him to

ask him questions about how to go to the far shore of old age and death

(jarmaraassa pra) from which the Pryaa (Going to the Far Shore) gets its

name. This narrative frame is supplied both by the introductory vatthugths (v. 976-

1031) and by the first epilogue (v. 1124-1130 and the short prose section preceding

them), which both identify the Buddhas 16 interlocutors as Brahmans, but are

otherwise somewhat contradictory. As already mentioned above, the first of these

the introductory vatthugthsare not commented upon in the Caniddesa and

therefore are likely a late addition. This, however, still leaves the first epilogue, as

well as the vocative in the 11th dialog, which are commented upon therein. The issue

496
Sn. 1043-5: khattiy brhma devatna yaam akappayisu puthdha loke.

234
of the narrative frame of the Pryaa is somewhat complex, so we will return to it

later when we discuss the use of framing and commentary to (re)interpret older

material. Suffice it for now to note that brhmaa as a vocative to refer to the

Buddhas interlocutor occurs only once in the dialogs of the Pryaa and is only

made intelligible by the narrative framework.

This leaves two suttasone in the Pryaa and one in the Ahakaof the

four that I have mentioned as using brhmaa in ways that are consistent with later

usage. Both of these two suttas use the word in a manner that, as we saw in Chapter

II.1, is extremely common in the mainstream Pali Buddhist literatureas part of the

compound samaabrhmaa. 497 The Purbheda Sutta of the Ahaka (10) uses the

compound only once, to refer to people other than the ideal person being described by

the sutta:

That by which ordinary people, as well as samaas and Brahmans, might

speak of him
Is not esteemed by him; therefore, he does not shake in the midst of
doctrines. 498

Samaabrhmaa is therefore used here, much as we found elsewhere in the Canon

in Chapter II.1, as a foil against which to construct Buddhist identity. Any distinction

between the samaa and the brhmaa is irrelevant; they are grouped together as a

class of people to be opposed to the ideal person.

497
In most Pali texts, the nominative plural form of this compound is the standard
samaabrhma, but in the Sutta Nipta, the non-standard form samaabrhmase is used. See p.
86 n. 184.
498
Sn. 859: yena na vajju puthujjan, atho samaabrhma / ta tassa apurakkhata,
tasm vdesu nejati //.

235
 The use of the compound in the Nandamavapucch of the Pryaa (7) is

also quite similar to the standard canonical usage. The interlocutor of the sutta,

Nanda, asks the Buddha about whatever samaabrhmase say that purity is by

means of what is seen or heard, say that purity is by means of ceremonial observances,

say that purity is by various means. 499 The Buddha confirms that such people do not

escape from rebirth. Thus, as in the Purbheda Sutta, samaabrhmase serve

collectively as a foil against which to construct the Buddhist ideal. The Buddha

continues, however, by noting that not all samaabrhmase fail to escape from

rebirth:

I do not say that all samaabrhmase

Are enveloped in birth and old age.
Whosoever, having given up here what is seen or heard or felt, as well as
ceremonial observances,
Having given up all the various forms, having thoroughly known thirst, are
without savas.
I say that they indeed are men who have crossed the flood. 500

Thus, once again, samaabrhmase are treated collectively, as a group, some

members of whom have crossed the flood of rebirth. As in the many canonical uses

of the compound samaabrhma treated in Chapter II.1, the distinction here

between samaa and brhmaa is irrelevant; the only relevant distinction is between

those who have escaped from rebirth and those who havent.

499
Sn. 1079-81: ye kecime samaabrhmase / dihassutenpi vadanti suddhi /
slabbatenpi vadanti suddhi, anekarpena vadanti suddhi.
500
Sn. 1082: nha sabbe samaabrhmase / jtijarya nivutti brmi / ye sdha
dihava suta muta v, slabbata vpi pahya sabba / anekarpampi pahya sabba, taha
pariya ansavse / te ve nar oghatiti brmi //.

236
 Excepting, then, these four suttas that use the word brhmaa in one way or

another that is fairly similar to common canonical usages, we are left with eight

suttasfive in the Ahaka and three in the Pryaathat use the word in a way

that I will argue reflects an early usage that, like many of the doctrinal formulations

found in the same texts, is rather unique in comparison to what we ordinarily find in

most canonical texts. Some of these references we have already seen. At the

beginning of this dissertation, I quoted from the Dhotakamavapucch of the

Pryaa (5), in which a certain Dhotaka approaches the Buddha to ask him about

nibbna, and praises him in the following terms:

I see in the world of gods and men a Brahman wandering about with nothing.
Therefore, I bow to that one who is all-seeing. Release me, kya, from my
doubts. 501

The Buddha responds that he cannot release a person from doubts, only teach him the

best dhamma (dhamma ca seha). Dhotaka then asks him to do so:

Having pity, Brahm, teach me the dhamma of detachment, that I may

understand,
(And) being unchanging just like space, calmed, I may wander right here
unattached. 502

This sutta, therefore, unique among the suttas we will be examining here, has the

Buddhas interlocutor refer to the Buddha himself as a Brahman. Note that there is

no comparison, explicit or implicit, to a real or literal Brahmani.e., a Brahman

by birth. Dhotaka simply refers to the Buddha as a Brahman, as if that is a natural

501
Sn. 1063: passmaha devamanussaloke, akicana brhmaamiriyamna / ta ta
namassmi samantacakkhu, pamuca ma sakka kathakathhi //.
502
Sn. 1065: anussa brahme karuyamno, vivekadhamma yam aha vijaa /
yathha ksova abypajjamno, idheva santo asito careyya //.

237
word to describe him. As such, he refers to the Buddha using the vocative brahme

(Brahm), and this appears to be interchangeable with the other vocatives he uses to

describe himnamely bhagav (Blessed One) and mahesi (Great Seer). In

calling the Buddha a Brahman, Dhotaka does describe him as having nothing

(akicana), but this serves simply as an attributive describing the noun brhmaa.

What Dhotaka seems to be saying, therefore, is that he sees a Brahmannamely, the

Buddhaand is impressed by him because he has nothingand that is why he has

approached him to release me from my doubts.

This particular association between the Brahman and having nothing recalls

the final lines of the immediately preceding sutta in the Pryaa, the

Mettagmavapucch (4). This sutta, like all of the remaining suttas we will look at,

uses the term brhmaa abstractly to refer to the ideal person. In this sutta, the

Buddhas interlocutor, Mettag, asks him about how to cross the flood (ogha) of

birth and old age, and he responds by saying that one must not have attachment

(upadhi). After answering Mettags questions, the Buddha summarizes, saying,

Whomever one would recognize as a Brahman, having knowledge, without

possessions, unattached to sensual pleasure and existence,
He has certainly crossed this flood and, having crossed over to the far shore, is
without defect, without doubt. 503

The ideal Brahman spoken of here is said to be vedaga word with a problematic

etymology, but generally agreed to mean having knowledge. 504 In describing the

503
Sn. 1059: ya brhmaa vedagum bhija, akicana kmabhave asatta / addh hi
so ogham ima atri, tio ca pra akhilo akakho //.
504
Norman lists the word veda as one of the terms taken over by the Buddha but used with
new senses, and under that entry, he writes, The word vedagu [sic], which in its Brahmanical sense

238
Brahman in this way, the Buddha is effectively calling himself a Brahman, since at

the very beginning of their dialog, Mettag had announced that he thought the

Buddha was vedag (v. 1049).

We also find the Brahman associated with knowledge, in this case designated

by a, in yet another sutta of the Pryaa, the Poslamavapucch (14):

Having known the origin of nothingness, [he thinks,] Enjoyment is a fetter.

Understanding this thus, therefore he sees clearly with respect to that.
This is the true knowledge of that perfected Brahman. 505

This difficult verse has been commented upon by Wynne, who argues that the

Buddha is saying that one moves beyond the state of nothingness (kicaa) by

having the insight (vipassati) that it has its origin in the fetter delight (nand

sayojana iti). 506 Regardless of how exactly one interprets this verse and the short

dialog in which it is found, it is clear that the Buddha and his interlocutor Posla are

engaging in a rather technical discussion about meditative techniques. Within this

meant one who had gained competence in the Vedas, was interpreted as one who had gained
knowledge of release from sasraTheravda Buddhism and Brahmanical Hinduism, 198. It is
not clear, however, what exactly Norman means by its Brahmanical sense, since there is little
evidence that people learned in the Vedas were ever referred to as vedag. This form, in fact, with the
nominative singular in , is not even found in Sanskrit. The Paramatthajotik II (330.27; cited by
Norman, Group of Discourses, 208 n. 322) takes the suffix as coming from the root gam, but Norman
argues that the Pali form actually comes from an original Sanskrit vedaka, where the k in the suffix
changed to g and then the word as a whole transferred from the a to the declension, both of which
changes are otherwise attestedNorman, Group of Discourses, 208 n. 322, 208 n. 319, 181 n. 167. To
my knowledge, however, neither putative Sanskrit originalvedaga or vedakawas ever used to refer
to someone versed in the Vedas. (Vedaka is attested, but instead with the meaning of announcing or
proclaimingBhtlingk and Roth, Sanskrit-Wrterbuch, v. 6, 1358.) Since veda is a perfectly
ordinary Indo-Aryan word, derived from the root vid, for knowledge, I see no a priori reason to
assume that is was taken over from Brahmanism and given a new sense. The sense with which
it is used in the Ahaka and Pryaa is, in fact, its most literal and therefore presumably original
senseknowledge.
505
Sn. 1115: kicaasambhava atv, nand sayojana iti / evam eta abhiya, tato
tattha vipassati / eta a tatha tassa, brhmaassa vusmato //.
506
Wynne, Origin of Buddhist Meditation, 103-6.

239
context, the Buddha uses the word brhmaa, without any immediately apparent

comparative intent, simply to refer to the person who has attained the meditative goal

he is describing.

Thus, to summarize, there are three suttas in the Pryaa in which the word

brhmaa is used to describe the ideal person. In the Mettagmavapucch, the

interlocutor refers to the Buddha as vedag, and by the end of their dialog, the

Buddha implicitly calls himself a Brahman by defining the Brahman who is vedag

as one who has nothing. In the immediately following Dhotakamavapucch, the

identification of the Buddha as a Brahman is made explicit by the next interlocutor,

who refers the Buddha as a Brahman who has nothing, and thus uses the vocative

brahme together with the more common epithets mahesi and bhagav. Finally, in the

Poslamavapucch, the Buddha, in answering a fairly technical question on

meditation, uses the word brhmaa to refer to the person who has attained the

meditative goal. Together with these three suttas, there are three other suttas in

which the word brhmaa is not used to describe an ideal personin one to refer to a

social group that (along with the khattiyas) performs sacrifices, in another as a

vocative to refer to the Buddhas interlocutor, and in a third as part of the compound

samaabrhmase. Compared to each other directly, these uses of the word

brhmaa in the Pryaa are fairly heterogeneous. As the Pryaa comes down

to us, however, they are integrated into a framework in which the Buddhas

interlocutors are explicitly identified as Brahmans, and, according to the vatthugths,

as disciples of a Vedic master named Bvari.

240
 The use of the word brhmaa in the Ahaka, by contrast, is quite

homogeneous, in spite of a lack of an overarching frame narrative to make them

intelligible. With the exception of the single use of the word in the compound

samaabrhma already discussed (v. 859), all instances of the word brhmaa in

the Ahaka refer to the ideal person. As already discussed, most of these are found

within the suttas that Vetter identifies as the core of the Ahaka, and thus take as

their theme the non-dual teaching that Gmez described in his article. The brhmaa

is accordingly described in fairly negative terms, in terms, that is, of what does not

define him. We have already looked at, in the context of Gmezs discussion of the

apophatic teaching of the Ahaka, the following verse from the Suddhahaka Sutta

(4):

A Brahman does not say that purity is from something else, in what is seen, in
what is heard, in morality and vows, or in what is thought.
[He is] unsmeared with respect to merit and evil, rejecting what has been
taken up, not doing [anything] in this connection. 507

As already discussed, Gmez has convincingly argued that this verse advocates a

non-dual position, which we may now add is used to define the Brahman as an ideal.

This is confirmed by the final verse of the same sutta:

The Brahman is beyond boundaries. There is nothing cognized or seen that is

seized by him.
He is not empassioned by passion; he is not empassioned with dispassion.
Nothing else is taken up by him here. 508

507
Sn. 790: na brhmao aato suddhimha, dihe sute slavate mute v / pue ca ppe ca
anpalitto, attajaho nayidha pakubbamno //.
508
Sn. 795: smtigo brhmao tassa natthi, atv va disv va samuggahta / na rgarg
na virgaratto, tassdha natth paramuggahtanti //.

241
A similarly negative description pointing to a non-dual ideal is found in the

immediately following Paramahaka Sutta (5):

With respect to what is seen, heard, or felt here, there is not even the slightest
apperception formed by him.
Who here in the world could have doubts about that Brahman who has not
taken a view?

They do not form [apperceptions]; they do not honor [one over another]; and
dhammas are not accepted by them.
A Brahman is not to be inferred by morality or vows. Gone to the far shore,
such a one does not return. 509

In all three of these verses, the Brahman is described in terms of a non-dual ideal,

according to which ones excellence cannot be determined by any of the

characteristics that one might ordinarily determine excellencemorality (sla), vows

(vata), or merit (pua)and therefore, as v. 795 puts it quite bluntly, he is beyond

boundaries (smtigo).

Two other suttas in the Ahaka define the Brahman in similarly non-dualistic

terms, but focusing more specifically on his avoidance of disputes by eschewing all

views. Verse 907 of the Mahbyha Sutta (13) begins much as v. 790, saying,

There is for the Brahman nothing that needs to be taught by another, but then turns

to the question of views rather than morality and vows:

having considered that which is grasped among dhammas.

Therefore, he has gone beyond disputes, for he sees no other dhamma as
best. 510

509
Sn. 802-3: tassdha dihe va sute mute v, pakappit natthi api sa / ta brhmaa
dihim andiyna, kendha lokasmi vikappayeyya // na kappayanti na purekkharonti, dhammpi
tesa na paicchitse / na brhmao slavatena neyyo, pragato na pacceti tdti //.
510
Sn. 907: na brhmaassa paraneyyam atthi, dhammesu niccheyya samuggahta / tasm
vivdni uptivatto, na hi sehato passati dhammam aa //.

242
The sutta continues for a few verses with the Buddha describing the folly of those

who adhere to views; then, in v. 911, he once again defines the Brahman as the

antithesis of this folly:

The Brahman cannot properly be defined; he is not a follower of views, nor is

he attached to knowledge.
And having known common conventions, he is indifferent, (thinking), Surely
they (will) take it up. 511

A verse of the Mgandiya Sutta (9) puts the non-entry of the Brahman into disputes

into the context of his (ideal) inability to make distinctions:

Why would that Brahman say, True? Or with whom would he dispute,
[saying,] False?
For whom there is neither equal nor unequal, with whom would he enter into
an argument? 512

The Brahman, according to these two suttas, therefore, refuses to embrace any

particular view, and thus is not drawn into any disputessince obviously one cannot

take part in an argument unless one takes a side.

There remains only one more reference to the ideal person as a Brahman in

the Ahaka, which unlike the ones we have looked at so far, lies outside of the core

suttas identified by Vetter. As such, the Brahman in this verse is defined in almost

the opposite terms as the ones we have just looked at:

Not having deviated from the truth, the sage stands on dry ground, a Brahman.
Having given up everything, that one indeed is called calmed. 513

511
Sn. 911: na brhmao kappam upeti sakha, na dihisr napi abandhu / atv ca so
sammutiyo puthujj, upekkhat uggahaanti mae //. I am accepting the variant sakha here rather
than sakh, as found in some manuscripts.
512
Sn. 843: saccan ti so brhmao ki vadeyya, mus ti v so vivadetha kena / yasmi
sama visama vpi natthi, sa kena vda paisayujeyya //.

243
The Brahman here is called calmed because, following the desire theme

delineated by Burford, the Attadaa Sutta (15) in which this verse is found focuses

on the abandonment of desire as the means to attaining the goal. Unlike the anti-

dihi suttas we have just been looking atand more in concert with the rest of the

Canonthis sutta does not disparage knowledge, but uses it to define the ideal person.

The immediately following verse, in fact, can be taken as the mirror image of the

anti-dihi passages we have been looking at:

That one indeed is knowledgeable; he has attained knowledge. Having found

out the dhamma, he is independent.
Behaving properly in the world, he does not envy anyone. 514

Thus, the Brahman in this sutta is defined in terms more in keeping with what we

might find elsewhere in the Canonas an ideal person who is defined as such by his

attainment of knowledgerather than the non-dualistic ideal la Gmez that

disparages knowledge, as is found in the other Ahaka we have looked at.

Nevertheless, the pattern according to which the word brhmaa is used is the same

as in the other Ahaka passages. The ideal person is simply referred to as a Brahman,

without any apparent comparison or contrast to a social group going by the same

nameonly in contrast, rather, to those, whoever they may be, who do not fit the

ideal.

513
Sn. 946: sacc avokkamma muni, thale tihati brhmao / sabba so painissajja, sa ve
santoti vuccati //.
514
Sn. 947: sa ve vidv sa vedag, atv dhamma anissito / samm so loke iriyno, na
pihetdha kassaci //.

244
 Indeed, it is precisely this that is so striking about the use of the word

brhmaa to refer to the ideal person in the Ahaka. In the examples we looked at

from other parts of the Canon above, when reference to the Buddhist ideal person as a

Brahman is made, it is done so in contrasteither explicitly or else implicitly through

narrative framingto literal Brahmansthat is, members of a particular social

group that claim Brahmanhood by birth. In the Ahaka, no such comparison is found.

Brhmaa is simply used as a word for the ideal person, together with other words

that lack the historical baggage that the word brhmaa has acquired and thus appear

to us as more naturally Buddhistmost importantly, bhikkhu and muni. We can

even say that these three terms are used interchangeably with one anotherindeed,

this is demonstrated by v. 946, just cited, in which the words muni and brhmaa are

used in apposition to one another. Thus, there is nothing about the use of the word

brhmaa that seems to privilege it over or distinguish it from bhikkhu or muni; all

three are simply used to refer to the ideal person.

In addition, not only is the ideal personby whatever name he is callednot

contrasted with the literal Brahman by birth; he is actually contrasted, as Vetter has

noted, with the samaa. Verse 828 of the Pasra Sutta (8) warns,

These disputes arise among samaas. Among them there is victory and defeat.
Having seen this too, one should abstain from disputes, for there is no other aim than
the gain of praise. 515

Likewise, in the Cabyha Sutta (12), an (unnamed) interlocutor says to the Buddha,

515
Sn. 828: ete vivd samaesu jt, etesu ugghti nighti hoti / etam pi disv virame
kathojja, na haadatthatthi pasasalbh //.

245
 What some say is true, real, others say is vain, false.
And having thus gotten started, they dispute [with one another]. Why do the
samaas not speak as one? 516

To this, the Buddha answers enigmatically,

There is only one truth; there is no second about which a wise man might
dispute with a wise man.
Variously they proclaim their own truths; therefore, the samaas do not speak
as one. 517

As the Buddha continues to answer the interlocutors questions, it becomes clear that

what he is referring to is a non-dual, apophatic truth, beyond the positions that

people take in disputes. Therefore, the claims of all the various samaas to truth are

invalid, since they all equally claim that they are right and the others are wrong:

For if one is inferior by the word of another, the latter is of inferior

intelligence along with him.
And if he has himself attained to knowledge and is intelligent, there is no one
who is a fool among the samaas. 518

Samaas are therefore defined by their wont for disputes and concomitant lack of

access to truth. And as we have seen, to this is contrasted the ideal person who does

not enter into disputes, who does not cling to viewswhether he be called a

brhmaa, a muni, or a bhikkhu.

516
Sn. 882: yam hu sacca tathiyanti eke, tam hu ae tuccha mus ti. / evam pi vigayha
vivdayanti, kasm na eka sama vadanti.
517
Sn. 884: eka hi sacca na dutyam atthi, yasmi pajno vivade pajna / nn te
saccni saya thunanti, tasm na eka sama vadanti //. I follow Norman here in his reading of the
second padaNorman, The Group of Discourses, vol. 2, 331 n. 884.
518
Sn. 890: parassa ce hi vacas nihno, tumo sah hoti nihnapao / atha ce saya vedag
hoti dhro, na koci blo samaesu atthi //.

246
 Chapter III.2

Commentary and Framing as Hermeneutical Devices for

Dealing with Early Uses of the Word Brhmaa

Given that the use of the word brhmaa in the Ahakaand to a lesser

extent the Pryaaappears to be quite different from what we find elsewhere in

the Canon, how are we to construct a credible diachronic theory of the role of the

word brhmaa in the development of Buddhist identity? I believe that the approach

taken by Burford in her study of the Ahakathat is, comparing the root text to the

early commentary in the Mahniddesa and later commentary in the Paramatthajotik

IIprovides a useful method for discerning the way in which Buddhist

understandings of key concepts changed over time. This method allows one not only

to compare a term or concept in different literary contexts and in texts that were likely

written at different times, as we have been doing so far; it allows one to compare a

term or concept in the same literary context and in texts that were definitely written at

different times. In other words, we get to see directly how a later author grappled

directly with the problematic terms and concepts in an earlier text. While Burford

247
looked at the way the later commentators dealt with the problematic doctrinal aspects

of the root text, however, I will be looking at the way in which later authors dealt

with the problematic treatment of the word brhmaa.

I will begin by applying this method to the Ahaka, which based on the

evidence presented so far may be a bit older than the Pryaa and in any case

appears to be more uniformly at odds with mainstream canonical doctrine. There are

two significant texts that allow us to see how later authors grappled with the

original 519 verses of the Ahaka and the problematic ideas found therein. The first, of

course, is the Mahniddesa, which provides the earliest formal commentary on the

text and, as we have already seen, clearly seeks to interpret it within an orthodox

doctrinal framework. The second is another version of the Ahaka itselfnamely, a

version now only fully extant in Chinese translation that embeds the verses of the

Arthavargya within prose frame narratives. These two texts, I will show, represent

two different approaches to dealing with the figure of the Brahman in the

Ahaka/Arthavargya. Next, I will turn to the Pryaa, which presents a somewhat

more complicated case than the Ahaka because of the heterogeneous way in which it

treats the category Brahman. There the focus will be on the extant text of the

Pryaa itself and the way in which its internal composition served to frame and

thus make meaningful references to the ideal person as a Brahman within something

more akin to the mainstream canonical narratives on Brahmans and Brahmanhood.

519
By original, I simply mean the verses as they are preserved in Pali, with the possible
exception of v. 836, which is not commented upon by the Mahniddesa (Norman, Group of Discourses,
836), and/or their counterparts in the Chinese version. This does not necessarily imply that those
verses were composed together at one time by a single author; as we have discussed above, Vetter and
Burford have given compelling reasons for believing that the Ahaka is a composite text.

248
III.2.1 The Mahniddesa Commentary on the Ahaka Vagga

As a word-by-word commentary, the Mahniddesa is mostly focused on the

explication of problematic terms, or as we saw through Burfords analysis, on

transforming the poetic language of the Ahaka into a system of technical

terminology consistent with broader canonical usage. Needless to say, this does not

leave much room for extended narrative frameworks, such as we find elsewhere in the

Canon, that would display a clear conception of Buddhist identity vis--vis the

category of Brahman. Nevertheless, there are some clues scattered throughout the

commentary that do point to a somewhat different conception of the category than we

find in the Ahaka itself. To begin with, we find repeated referencemore than in

the suttas of the four main Nikyas, in factto the four varas (brhmaa, khattiya,

vessa, and sudda). 520 This would imply that the Niddesa author was well aware

perhaps even acceptingof the vara system. Moreover, on the commentary on v.

859, the one place in the Ahaka, as we have seen, that refers to samaabrhmase,

the Niddesa glosses the compound by splitting it into its constituents (i.e., as a

dvandva) and glossing brhma as whoever says Bho (ye keci bhovdik). Thus,

although the context of the compound samaabrhma in the original sutta renders

the distinction between samaa and brhmaa irrelevant (both being contrasted with

the ideal muni), the commentator imposes a distinction between them and makes clear

520
Nidd. 1.1.766-9, 1.3.782, 1.6.805, 1.6.808, 1.11.864-5, 1.14.915, 1.14.932, 1.15.935,
1.15.947, 1.15.948, 1.16.955, 1.16.957, 1.16. 970, 1.16.973. (I cite from the two Niddesas according
to whether it is the first [Mah] or second [Ca] Niddesa, followed by the number of the sutta, and
ending with the number of the verse according to the Sutta Nipta enumeration that is being
commented upon.) Interestingly, none of the references to the four varas are found in commentaries
on suttas that Vetter includes within the core.

249
that a particular social class is being referred to by brhmaa (bho being the term of

address they used with their perceived social inferiors).

The Mahniddesa also supplies some Brahman interlocutors for the Buddha

where none exist, explicitly at least, in the original Ahaka. On a couple of occasions,

this takes the form of short quotations from the Buddha himself that include the

vocative brhmaa to refer to the person whom the Buddha is addressing. 521 The

most significant case, however, is the Mgandiya Sutta (9), one of only a few suttas

in the Ahaka that has a named interlocutor. 522 This sutta is included within the

core identified by Vetter because, with the exception of its first two verses, it

consists of a dialog between the Buddha and a certain Mgandiya in which the

Buddha takes the strong position, typical of the core suttas of the Ahaka, that purity

does not come from holding a particular view, which can only lead to disputes. 523

The Ahaka does not identify the Buddhas interlocutor except by name, although, at

least in the context of his debate with the Buddha in v. 837-847, he could easily be

the Mgandiya of MN 75a paribbjaka who debates with the Buddha and in the

end is convertedor else someone like him. Tradition, however, appears to have

identified the Mgandiya of the Ahaka with another Mgandiya entirelyone rather

ill-suited for the heady philosophical debate that takes up the bulk of the Mgandiya

521
Nidd. 1.2.773, 1.15.939. I have not been able to trace the first quote, but the second is a
quote from MN 87.
522
The others are Tissa Metteya (7), Pasra (8), and Sriputta (16).
523
Burford has noted the irony of the fact that in this sutta the Buddha enters into a dispute
with Mgandiya in order to express his view that one should not hold any view, because holding views
leads to disputesBurford, Desire, Death and Goodness, 54-7.

250
Suttanamely, a certain rich Brahman who is impressed by the Buddha and offers

his daughters hand to him in marriage.

The connection is already indicated by the first verse of the canonical

Mgandiya Sutta, a verse often quoted as a prime example of Buddhist misogyny, in

which the Buddha rather contemptuously rejects Mgandiyas daughter:

Having seen Tah, Arati, and Rag [the daughters of Mra], there was not
even the desire for sexual intercourse.
What indeed is this, full of urine and excrement? I wouldnt want to touch it
even with my foot. 524

This verse appears to be rather clearly an addition to the Mgandiya Sutta (albeit a

relatively early one), intended to identify the Mgandiya of the sutta with the

Brahman Mgandiya who offered the Buddha his daughter. There are two reasons to

suspect this. First, the verse itself makes little sense as a lead-in to a philosophical

debate on abandoning all views. It makes perfect sense, however, in the story of the

Brahman Mgandiya offering his daughter to the Buddha, which is recounted in the

Pali tradition in the Ahakath commentary on the Dhammapada (2.1.5), where the

Buddha does indeed respond to Mgandiyas offer with this contemptuous response,

but does not follow up with a teaching on abandoning all views. (Instead, the girl

becomes rather understandably angry at the Buddhas insult and marries King Udena

of Kosamb.)

The second reason to suspect that the first verse of the canonical Mgandiya

Sutta was added is that the second canonical verse is even more suspect than the first.

524
Sn. 835: disvna taha arati ragaca, nhosi chando api methunasmi / kim evida
muttakarsapua, pdpi na samphusitu na icche //.

251
The second verse, intended as Mgandiyas response to the Buddhas rejection, rather

clumsily attempts to tie the Buddhas response to a marital offer to the debate on

views that follows:

If you do not want such a jewel, a woman desired by many kings,

What sort of view, morality, vows, way of life, and rebirth into existence, then,
do you profess? 525

This verse is almost certainly not original because, as already noted, it is the only

canonical verse of the Ahaka not commented upon by the Mahniddesa, and

moreover it has no parallel in the Chinese translation of the Arthavargya. The latter

has in its place the following two verses, which serve to provide an even clumsier

transition to the philosophical debate that follows:

I have said that I do not desire sexual intercourse. I do not observe within any
un-dharmic action.
Even though I hear something ugly, I do not bear hatred. As long as it does
not stop within, there are innumerable sufferings.

One sees the outside as beautiful; muscle is covered by skin. Why should the
sage accept this?
Be aware of and look at inner and outer actions. As for craftiness, I say it is a
foolish action. 526

The version preserved in Chinese also includes a prose narrative prior to the verses

that recounts the story of how Mgandiya offered his daughter to the Buddha; it is

similar to the account preserved in the Pali Dhammapada-ahakath, but by no

means identical. The story is also recounted, again with its own idiosyncratic details,

525
Sn. 836: etdisa ce ratana na icchasi, nri narindehi bahhi patthita / dihigata
slavata nu jvita, bhavpapatti ca vadesi kdisa //.
526
T. 198:

252
by Buddhaghosa in his commentary on the Mgandiya Sutta in the Paramatthajotik

II.

It appears, then, that an original series of verses outlining a debate over the

abandonment of all views between the Buddha and an otherwise unidentified

Mgandiya, when it was incorporated into the Ahaka/Arthavargya, had the urine

and excrement verse added at the beginning to associate it with the story of the

Brahman Mgandiya offering the Buddha his daughter. Different traditions sought to

incorporate this verse into the original sutta in different ways, as evidenced by the

lack of agreement between the intervening verses in the Pali and Chinese versions.

Finally, frame stories were added on in slightly different ways by different

traditionsinserted directly into the Arthavargya that is preserved in Chinese, but

delivered as a commentary by Buddhaghosa in the Paramatthajotik IIto place the

sutta in the context of the story of Mgandiya offering the Buddha his daughter.

The Mahniddesa, for its part, appears to represent an intermediate stage in

this development. On the one hand, it accepts the initial urine and excrement verse

as part of the sutta and comments upon it, and it appears to be aware of the story of

the Brahman Mgandiya that it alludes to insofar as it glosses the word mgandiya

when it appears throughout the sutta as that Brahman (ta brhmaa) whom the

Buddha addresses by name (nmena lapati). On the other hand, the Mahniddesa

does not comment upon (and thus presumably is not aware of) v. 836, which is the

only attempt within the canonical version to create a plot link between the story of

Mgandiyas daughter and Mgandiyas debate with the Buddha over views.

253
Moreover, it makes no attempt to flesh out that story through a prose narrative as the

version preserved in Chinese does and the later commentary by Buddhaghosa will do.

But even though it fails to provide a complete narrative, the Mahniddesa

accomplishes quite a lot just by identifying Mgandiya as a Brahman. Originally the

debate in the Mgandiya Sutta was consistent with the other core suttas of the

Ahaka insofar as it advanced a non-dualist position that rejected all views and

referred to the ideal person in this respect as a Brahman. By identifying Mgandiya,

who clearly is antagonistic to the Buddhas non-dualist teaching in this debate, as a

Brahmanthat is, one can only presume, a Brahman by birthhowever, the

Mahniddesa set up a dichotomy between the literal Brahman represented by

Mgandiya and the metaphorical or true Brahman represented by the ideal

presented in the Buddhas teaching.

The final, and most basic, way that the Mahniddesa transforms the use of the

word brhmaa in the Ahaka is the gloss that it gives for the word itself. The word

brhmaa appears eight times in the Ahaka (excepting its single appearance in the

compound samaabrhmaa), and each time the Mahniddesa glosses the word as

follows:

[One is] a Brahman due to the warding off of seven dhammas: the heresy of
individuality is warded off, uncertainty is warded off, attachment to
ceremonies is warded off, passion is warded off, aversion is warded off,
delusion is warded off, pride is warded off. He has warded off evil, unskillful
dhammas, which are corrupting, lead to further becoming, are troublesome,
cause suffering, and lead in the future to birth, old age, and death.

Having removed all evils (said the Blessed one to Sabhiya),

He abides without stain, well-concentrated.

254
 Having overcome sasra, he is alonethus unattached he is called
Brahm. 527

Although in some ways it follows the spirit of the Ahaka, such as in mentioning

freedom from attachment to ceremonies, this gloss clearly is attempting to fit the

Ahakas use of the word Brahman into the mainstream trope of the true Brahman

by making the word an explicit metaphor for a set of virtues that can be taken as

representative of a mainstream Buddhist saint. More importantly, however, in doing

so it sets up a dichotomy between purity and impurity, skilful and unskillful, that

contradicts many of the very verses it is commenting upon.

In addition, the verse quotation found at the end of the glosswhich is taken

from the Sabhiya Sutta (Sn. 3.6, v. 519)introduces an element of figurative

language where none existed in the original Ahaka. The verse in question, in its

original context in the Sabhiya Sutta, is the beginning of a four-verse response by the

Buddha to a question asked by his interlocutor Sabhiya, a paribbjaka:

Attaining what do they call him a Brahman? (asked Sabhiya)

Due to what [do they call him] a samaa, and why [is he] a bath-graduate
(nhtaka=Skt. sntaka)?
Why is he called a nga?
Asked by me, may the Blessed One explain. 528

527
The full gloss can be found at the first instance of the word brhmaa at v. 790
(subsequently the gloss is abbreviated using the word pe): sattanna dhammna bhitatt brhmao
sakkyadihi bhit hoti, vicikicch bhit hoti, slabbataparmso bhito hoti, rgo bhito hoti,
doso bhito hoti, moho bhito hoti, mno bhito hoti. bhitssa honti ppak akusal dhamm
sakilesik ponobhavik sadar dukkhavipk yati jtijarmaraiy. bhitv sabbappakni,
[sabhiyti bhagav] / vimalo sdhusamhito hitatto / sasramaticca keval so, asito tdi pavuccate
sa brahm //.
528
Sn. 518: ki pattinam hu brhmaa, (iti sabhiyo) / samaa kena katha ca nhtako ti
/ ngo ti katha pavuccati / puho me bhagav bykarohi //.

255
In the four verses that follow, the Buddha explains why a person is called each of four

epithets named by Sabhiya, in each case using a pun: One is a brhmaa because of

having removed (bhitv) all evils; one is a samaa by quieting oneself

(samitvi); one is a nhtaka by having washed away (ninhya) all evils; and one is

a nga because one does not commit offense (gu na karoti). 529 Contrast with

this the situation in the Ahaka. The epithets brhmaa and nga 530 are used

straightforwardly, without any punning or other figurative language, as honorifics to

refer to the ideal person who is beyond views. The word samaa, as we have seen, is

used in a negative sense to refer to those decidedly non-ideal persons who engage in

endless disputes. Finally, nhtaka, which literally must clearly refer to the bath-

graduate (Skt. sntaka) of the Dharmastric tradition, is not found at all. By

quoting from Sabhiya Sutta, the author of the Mahniddesa is placing the word

brhmaa into a different contextone in which samaa is recognized as positive,

technical terminology from the Dharmastric (i.e., new Brahmanical) tradition are

known and referred to, and problematic terms are dealt with by using punning.

Overall, then, the Mahniddesas treatment of the problematic category

Brahman bears an interesting parallel to its treatment of the problematic teachings

found in the Ahaka. As Gmez and Vetter have shown, the Ahaka, or at least

certain key parts of it, advances a non-dualistic teaching that rejects all views and

other markers of purity as inherently one-sided and incomplete, leading only to

529
Sn. 519-22.
530
The honorific nga is found only once in the Ahaka (Sn. 4.9.845), making it less common
than bhikkhu, muni, and brhmaa. It is used three times, however, in the Pryaa (Sn. 5.4.1058,
5.12.1101, and v. 1131 in the anugtigth) to refer to the Buddha himself.

256
conflict. And as Gmez and Burford have shown, the Mahniddesa breaks the

non-dualism of the Ahaka to make it conform to the dualistic teaching of the rest of

the Canon. I would argue that it does much the same with the category Brahman:

It introduces a dichotomy where none existed before. On the one hand, it refers to the

vara system and to individual Brahmans such as Mgandiya, thus reflecting a

consciousness of literal Brahmans in the form of a particular social class that

arrogated the category to themselves on the basis of birth. On the other hand, it

transforms the actual usage of the category Brahman found in the Ahaka into a

metaphor by introducing punning and defining it in terms of orthodox Buddhist ideals

that are foreign to the Ahaka itself.

III.2.2 The Chinese Translation of the Arthavargya and the Hermeneutical Role

Played by Narrative Framing

A useful counterpoint to the Mahniddesa is provided by another version of

the Ahaka entirely, namely the Yzjng (), 531 which is a Chinese translation

of an Indian text, which for the sake of convention I will refer to as the

Arthavargya, 532 that clearly is related to the Ahaka-vagga of the Pali Sutta Nipta

531
For a full English translation of this text, see P. V. Bapat, trans., Arthapada Sutra: Spoken
by the BuddhaTranslated by the Upsaka Che-Kien under the Wu Dynasty (222-280 A.D.), parts 1
and 2 (Calcutta: Visva-Bharati, 1951).
532
It is of course impossible to know what title exactly the text translated into Chinese
originally had. The choice of the Chinese translator to use the character (lit., foot) in the title
would suggest an original Sanskrit pada; accordingly, Bapat entitles his translation of the text the
Arthapada Stra. On the other hand, the extant Sanskrit references we have to the textthe two
references already mentioned above in the Divyvadna and a reference to one set of verses as an
artthavargya stra in the Sanskrit fragments of the text itself published by Hoernle (Fr. 1, obv., ln.

257
in spite of some important differences that we will discuss shortly. The translation is

attributedreliably, according to modern scholarsto Zh Qin (), a non-

ordained translator living during the first half of the third century CE, 533 and it is

named in the 6th century catalogue of Sengyou (, T. 2145) 534 as one of

thirty-six works translated by Zh Qin during the time of Emperor Wn of the

Wei, from the beginning of the Hungw period [=222-229 CE] of the W king

Sn Qun through the middle of the Jinxng period [253 CE] of Sn

Ling . 535 The verses corresponding to those of the Ahaka-vagga are rendered

in an unusual six-character style that was apparently pioneered by Zh Qin. 536 In Zh

Qins translation, the verses are separated into sixteen divisions referred to as stras

(); each stra contains a group of verses that has a direct parallel in one of the

sixteen suttas of the Ahaka-vagga. Although there are certain discrepancies

4)agree on the title Arthavargya, which is more consistent with the Pali title Ahaka-vagga, at least
with respect to the word vagga. The word ahaka in the Pali, however, most likely should be
Sanskritized as aaka, in reference to the fact that suttas 2-5 of the textall of which, incidentally,
have ahaka in their titlesare each composed of eight verses. This is reflected, however, neither in
the Sanskrit references that we have nor in the Chinese translation, where clearly refers to meaning
or artha and has nothing to do with the number eight. Given the fact that only four of the sixteen
suttas of the Ahaka-vagga/Arthavargya have eight verses, it is possible that early on in the tradition,
the reference to eight was forgotten and ahaka mistakenly Sanskritized as artha. In any case, I
refer to the non-Pali text or texts that correspond to the Pali Ahaka-vagga but include prose frame
narratives as Arthavargya, simply as a matter of convention since that title is attested in actual
Sanskrit sources.
533
Jan Nattier, A Guide to the Earliest Chinese Buddhist Translations (Tokyo: The
International Research Institute for Advanced Buddhology, Soka University, 2008), 116-7.
534
For a brief description of Sengyous catalogue, see ibid., 11-13.
535
Translated in ibid., 121; bracketed dates in the original, but tone-marks mine.
536
Ibid., 119, 134.

258
between the Chinese version and the Pali version (including, in some cases, entire

verses in one that are not found in the other), in each case the number of verses in a

Chinese stra 537 is the same as that found in the corresponding Pali sutta. The two

versions also largely agree on the order of the stras, with the exception of two (nos.

10 and 15 in the Pali) that appear to have been misplaced and simply tacked on the

end in the Chinese version (see Table 1). 538

Chinese stra Corresponding Pali sutta

1. (King Jitn Stra) 1. kmasutta (Sensual Desire Sutta)
2. (King Udayana Stra) 2. guhahakasutta (Cave Sutta)
3. (Sundar Stra) 3. duhahakasutta (Wicked Sutta)
4. (Makara the Brahman 4. suddhahakasutta (Pure Sutta)
Stra)
5. (King Mirror-Face Stra) 5. paramahakasutta (Highest Sutta)
6. (Old and Young All Die 6. jarsutta (Old Age Sutta)
Stra)
7. (Question of Maitreya 7. tissametteyyasutta (Tissa Metteyya
Stra) Sutta)
8. (Yngc the Brahman 8. pasrasutta (Pasra Sutta)
Stra)
9. (Daughter of Mgandika 9. mgandiyasutta (Mgandiya Sutta)
Stra)
10. (Flying Quarrels of 11. kalahavivdasutta (Quarrels and

537
That is, arthavargya verses. Several of the stras in the Chinese version also include
other verses, but these can be distinguished from the arthavargya verses by the fact that only the latter
are translated in the unique six-character style just mentioned. In many cases, the arthavargya verses
can also be distinguished by the following words introducing the verses: , [The Buddha]
said this arthavargya [or arthapada] stra.
538
It is of course possible that the change happened in the other direction, but I think it most
likely that the Pali order is more original because, as already mentioned, the order of the suttas in the
Ahaka-vagga is exact according to number of verses. In spite of minor differences in the verses, the
Chinese stras each have the exact same number of verses as their Pali counterparts, and with the
exception of the two just mentioned, its stras are in the same order as the Pali suttas, which indicates
that both come from a common ancestor with sixteen stras, each with a fixed number of verses,
ordered according to increasing length. The Chinese version deviates from this pattern only slightly
either because two of the stras were misplaced and added at the end at some point in the transmission
of the translation, or because it was translated from an original in an Indian language in which the
misplacement had already occurred.

259
Other Teachings Stra) Disputes Sutta)
11. (Suddenly-Sees the 12. cabyhasutta (Minor Arrangement
Brahman Stra) Sutta)
12. (Sees-Dharma the 13. mahbyhasutta (Great
Brahman Stra) Arrangement Sutta)
13. (Dul the Brahman 14. tuvaakasutta (Speedy Sutta)
Stra)
14. (Bhiku Lotus- 16. sriputtasutta (Sriputta Sutta)
Blossom-Color Stra)
15. (Meeting of Son and 10. purbhedasutta (Before Dissolution
Father Stra) Sutta)
16. (King Wilul Stra) 15. attadaasutta (Stick Taken Up
Sutta)
Table 1. Note that the order of the stras in the Chinese version appears to have come about by extracting suttas
10 and 15 from the Pali version and placing them at the end. Note also the general discrepancy between titles of
corresponding stras. (For names in Chinese stra titles, I have given a Sanskrit equivalent when the Chinese is a
transcription whose Sanskrit equivalent is clear, a direct English translation when the Chinese is itself a
meaningful translation of the name rather than a transcription, or simply pinyin when I am uncertain of the
translation or transcription.)

Generally speaking, however, the titles of the stras in the Yzjng do not

correspond to those of the suttas in Ahaka-vagga. This is related to most important

difference between the two versions: While the Ahaka contains only verses, the

stras of the Yzjng embed these verses within a narrative frame that is written in

prose. As is often the case with Buddhist texts in which verses are embedded into

prose narrativesincluding, as we will see later, in the Pali traditionthe verses

often have little other than a tenuous connection to the plot of the prose narrative in

which they are embedded. This explains the discrepancy between the stra titles in

the two versions; while the titles in the Pali version can of necessity refer only to the

content of the verses themselves, the titles in the Chinese version often refer to

characters in the prose narrative that are not found in the verses. While the exact

nature of the Indian text from which the Yzjng was translated (language, sectarian

affiliation, etc.) is unknown, we can feel confident that there was an Indian version

260
(or versions) of the Arthavargya that included prose frame narratives with the verses,

and that they were not simply added by Zh Qin. Fragments of a text from Eastern

Turkestan were published near the beginning of the 20th century that contain verses

corresponding to the Pali Ahaka together with prose similar to that found in the

Yzjng, but not found in the Pali text.539

The Yzjng therefore can be taken to represent a scriptural transmission of

the Arthavargya that, unlike the Pali transmission, embedded its verses within prose

narratives. The exact relationship between the prose-encrusted transmission(s) and

the Pali transmission is not entirely clear, but it seems that they were largely, if not

one-hundred percent, independent. We have already seen that the Mahniddesa

seems to be aware of a story behind the verses in the Mgandiya Sutta insofar as it

refers to Mgandiya as a Brahman, just as he is identified in the Chinese version.

This is somewhat understandable, however, given that the first verse, in which the

Buddha rejects a woman as full of urine and excrement, was apparently added to an

otherwise erudite discussion about abandoning all views at a very early date to

suggest a connection to the story of Mgandiya offering his daughter to the Buddha.

539
A. F. Rudolf Hoernle, The Sutta Nipata in a Sanskrit Version from Eastern Turkestan,
Journal of the Royal Asiatic Society 48, n. 4 (Oct. 1916): 709-732. The longest prose fragment (fr. 2,
rev. and fr. 3; p. 714-715) refers to a parivrjaka named Mgandika, and thus, according to Hoernle,
apparently corresponds to the Mgandiya Sutta of the Ahaka, which identification is what gave him
the first clue to the identification of the fragments (p. 714 n. 3). Hoernle was apparently unaware of
the existence of a Chinese translation of the Arthavargya that also included prose sections together
with the verses. Although the fragmentary nature of Hoernles Sanskrit version and the vicissitudes of
translation lying behind the Chinese version make a direct comparison of the two difficult, it seems
unlikely that the Chinese version was translated from a text that was exactly the same as the Sanskrit
text from which Hoernles fragments come; for example, while the Sanskrit fragments refer to
Mgandika as a parivrjaka, the Chinese version refers to him () as a Brahman ().

261
The Mahniddesa makes no attempt, regardless, to elaborate upon this story, nor does

it construct a narrative for any other sutta in the Ahaka.

We do find an attempt to narrativize the Ahaka in Buddhaghosas

Paramatthajotik II, but aside from being quite late, this narrativization neither is

complete, nor does it correspond in any exact way to the narrativization found in the

Yzjng. It is incomplete insofar as it does not provide individual background stories

for six of the suttas in the Ahaka (nos. 10-15); instead, it gives a perfunctory

introduction to these suttas saying that they were spoken on the same occasion as

already described earlier in the same commentary for the Sammparibbjanya Sutta

(Sn. 2.13), when the Buddha was asked by a nimmitabuddha to preach for the benefit

of a group of deities of lustful behavior (rgacaritadevatna). 540 Buddhaghosa

does provide individual background stories for the other eleven suttas of the Ahaka

(nos. 1-9 and 16), but at least some of these are completely different from those

provided in the Yzjng. Others bear a greater or lesser similarity to their

counterparts in the Yzjng, but none of them can be said to be exactly the same.

The most similar perhaps is the story Buddhaghosa provides for the

Mgandiya Sutta (9), which, as we would expect, fleshes out the details of how

Mgandiya offered his daughter to the Buddha, only to have her rejected as full of

540
See the Paramatthajotik II commentary on Sn. 4.10: k uppatti? imassa suttassa ito
paresa ca pacanna kalahavivdacabyhamahbyhatuvaakttadaasuttna
sammparibbjanyassa uppattiya vuttanayeneva smaato uppatti vutt. visesato pana yatheva
tasmi mahsamaye rgacaritadevatna sappyavasena dhamma desetu nimmitabuddhena
attna pucchpetv sammparibbjanyasuttam abhsi, eva tasmi yeva mahsamaye ki nu
kho pur sarrabhed kattabbanti uppannacittna devatna citta atv tsa anuggahattha
ahateasabhikkhusataparivra nimmitabuddha ksena netv tena attna pucchpetv ima
suttam abhsi.

262
urine and excrement. Also quite similar is the commentary on sutta 5, which like

the corresponding stra in the Yzjng, includes the famous parable of the blind men

and the elephant. The details are different, however. The Chinese version begins

with Brahmans arguing over their various viewpoints and has the Buddha tell a story

about how a king in the past had a group of blind men describe an elephant to him.

Buddhaghosas version, on the other hand, identifies the people who are arguing

simply as nntitthiy, rather than as Brahmans, and has a king in their own time

order the demonstration of blind men describing an elephant, rather than having the

Buddha recount it. Buddhaghosa also associates the same story with sutta 3 as does

the Chinese versionnamely, in which a group of jealous non-Buddhists sent a

woman named Sundar to the sagha and then killed her, so as to discredit the

Buddhist monks by making it look as though they had been consorting with this

woman and then killed her to cover it up. Buddhaghosa only recounts this story in

brief, however, and refers readers to the full account that is preserved in the Pali

tradition at Udna 6.4. The story that Buddhaghosa provides for the Kma Sutta (1)

is similar to the story found in the Yzjng, insofar as it involves the Buddha teaching

a certain Brahman after the latter loses his entire crop, but again there are great

variations in the details: In the Chinese version, the Brahman loses the crops to hail,

while in Buddhaghosas, he loses it to a flood; the content of the Buddhas

subsequent teaching is different in two versions; and overall the Chinese version is

considerably longer. 541 Likewise, Buddhaghosas story for sutta 2 is similar to the

541
For these reasons, I cannot agree with Bapats assessment that the Pali commentary

263
Chinese version in that it involves a king named Udena (=Skt. Udayana), but as Bapat

notes, the stories differ considerably in details. 542

Taken together, then, it appears that Buddhaghosa had access to a tradition or

traditions that had some relationship to the traditions that went into the composition

of the version of the Arthavargya from which the Chinese Yzjng was translated,

namely to the extent that certain stories, most likely well-known in general outline,

but not fixed in details, were associated with specific sets of versessuch as the story

of Mgandiya/Mgandika offering his daughter with sutta/stra 9, or the story of the

blind men and the elephant with sutta/stra 5. There is no evidence, however, that

Buddhaghosa had access to the Arthavargya from which the Yzjng was translated

itselfand, indeed, much evidence to suggest that he did not, since in most cases he

does not provide a story paralleling that found in the Yzjng. Regardless of this

evidence, however, it is unnecessary to assume that Buddhaghosa would have had

access to a prose-encrusted Arthavargya to be able to provide the back-stories for

his commentary, for there is abundant evidence from the early Buddhist tradition that

certain well-known stories were circulated in early Buddhist communitiesand in

many cases may originally not have had any sectarian affiliation at alland were

solidified at various times, in various sectarian communities, and at various stages

of the processes of canonical and commentarial formation by rendering them into

prose and associating them with verses.

provides an identical introductory story to that found in the YzjngBapat, Arthapada Sutra, 1 n. 2.
542
Bapat, Arthapada Sutra, 16 n. 1.

264
 Indeed, as I already hinted at, while the combination of prose narratives with

the verses found in the Yzjng is foreign to the Pali Ahaka tradition, the practice of

embedding verses within a prose narrative is not by any means foreign to the broader

Pali tradition. It is, in fact, found in many Pali texts, sometimes fully within the

canonical boundary, but in other cases straddling the canonical-commentarial divide.

The Sagth Vagga of the Sayutta Nikya, which we will be discussing in more

detail in Part IV, is an example in which canonical suttas are composed of verses

embedded in prose narratives (hence the name, with verses). Most examples,

however, are found in the Khuddaka Nikya, where the Ahaka itself is found. Many

of the canonical suttas of the Sutta Nipta in which the Ahaka is found include prose

narration along with their verses; the Ahaka is, in fact, unique insofar as it is the

only vagga of the Sutta Nipta that does not contain any prose. The Udna and

Itivuttaka are entire collections whose suttas consist of prose narrations that describe

the circumstance in which the Buddha said something in verse. Other collections in

the Khuddaka also associate prose narratives with verse, but in those cases, the

tradition considers only the verses to be canonical and the prose to be extra-canonical

commentary. The most well-known of these is of course the Jtaka, which is in fact

better known for its extra-canonical prose commentary than its canonical verses

since in most cases the actual story of the Buddhas former birth is only intelligible

from the former. The same is true of the Dhammapadathe canonical verses which

are so well-known in the West from repeated translation into Western languages are

supplemented in the commentary with extensive prose narratives that serve to

265
contextualize those verses; as we have already seen, this is one of the places in the

Pali tradition where the complete story of Mgandiya offering his daughter to the

Buddha is preserved.

Of these examples, the Udna provides an interesting counterpoint to the

Ahaka as an illustration of the process by which prose narratives and verses were

combined differently by different traditions. While the Pali Udna collection consists

entirely of prose narratives with embedded verses, most of the other extant Udna

collectionsincluding a Sanskrit version, a Tibetan translation, and one (T. 213) of

two translations found in the Chinese Canoninclude only verses. 543 The second

Chinese translation of an Udna collection (T. 212) does include prose narratives, 544

but it includes prose narratives only for some verses (providing in other cases only a

word-commentary), and even when a prose narrative is present, it more often than not

does not correspond to the narrative found in the Pali version. 545 Through a detailed

study and comparison of the Pali Udna and T. 212, Bhikkhu Anlayo has shown that

the udnas were most likely circulated without prose at first; that they grew as verse

collections differently in different traditions; and moreover that even when prose was

added, it was done so independently and in different ways by different traditions. 546

543
For citations of published editions of these texts, see Bhikkhu Anlayo, The Development
of the Pli Udna Collection, Buddhist Studies (Bukky Kenky) 37 (March 2009), 39-40 and 63 n. 2-
4.
544
Ibid., 40.
545
Ibid., 41-46. In fact, only in the case of the first three udnas is there an exact
correspondence between both the verses and the prose narrative in the Pali and T. 212.
546
Ibid., 55-57.

266
Anlayo notes in his conclusion that [t]he case of the Pli Udna collection thus

appears to be in some respects the reverse of the case of the Ahakavagga 547that

is, while the Pali tradition, unlike other traditions, kept the Ahaka verses free of

prose, it preserved the Udna verses completely embedded in prose, unlike most

other traditions, which did not. More importantly, though, Anlayos study shows

that even when two traditions do attach prose to a set of verses, they can do so

independently even while drawing upon a common tradition.

This conclusion is confirmed by the numerous correspondances found

between different traditions in which the same storyalbeit with certain variations in

detailis placed in different contexts, or associated with different verses, by different

traditions. Thus, even though the Pali tradition does not preserve prose narratives

together with the verses of the Ahaka, many of the stories found in the Yzjng are

found in other contexts within the Pali traditionin the Jtaka commentary, the

Dhammapada commentary, the Udna, and the Sutta Nipta commentary. 548 Thus, it

appears that different traditions had access to a common set of stories with which to

flesh out the Buddhas life, and they did so by anchoring those stories to verses that

were already established buddhavacana; each tradition, however, made its own, at

least semi-independent choices as to which verses to anchor those stories to. What is

interesting about the Yzjng, therefore, is not so much the stories per se, but rather

547
Ibid., 56.
548
For references to Pali and other parallels to the stories in the Yzjng, see the notes at the
beginning of each stra in Bapat, Arthapada Sutra.

267
the way in which they were used by the author(s) of the Arthavargya from which it

was translated to frame the verse, and thus implicitly interpret them.

Before we address the way in which the prose narratives of the Yzjng frame

and thus implicitly interpret the verses of the Arthavargya, though, it behooves us to

consider briefly some of the issues involved in using this Chinese translation to come

to conclusions about the original Indian text. Speaking generally, there is the issue of

the translation itselfits quality and the interpretive element it introduces to the text

merely as an act of translation. Vetter, for one, has questioned the quality of the

translation, and he has provided some credible examples that suggest that Zh Qin

may have made some outright mistakes. 549 On the other hand, as we have seen above,

the Chinese translation appears to interpret the text, at least in some cases, in a

manner more consistent with mainstream Buddhist doctrine, in which right view is

distinguished from wrong view, rather than maintaining the more radical rejection

of all views that Gmez and others have recognized as making the Ahaka unique. 550

I would argue, however, that it is not terribly surprising that a Chinese translation of

the Arthavargya, made many centuries after the verses were originally composed,

would be more consistent with orthodox doctrine than the much older Pali version. It

is possible, for example, that the version of the Arthavargya on which Zh Qin

based his translation had already introduced changes to lessen the radicalness of the

texts anti-views message. More likely, though, either Zh Qin himself or an Indian
549
Vetter, Some Remarks, 42: The translation is, however, of a doubtful quality; e.g. the
words stam and astam in 867, 869, and 870 are rendered with and .
550
See p. 220 above.

268
or Central Asian interlocutor who assisted him with the translation simply interpreted

the text in a manner consistent with mainstream orthodox doctrine.

More troubling for our purposes, however, is the issue of the word brhmaa,

which Zh Qin translates as fnzh (). 551 Although, as we will see, this word

appears frequently in his translation of the prose portions of the text, it appears only

once in his translation of the verses. This is in stark contrast to the Pali Ahaka

where it appears eight times (not including the compound samaabrhma in v. 859

of sutta 10 552). Now, it is possible that the reason for this is that the word brhmaa

only appeared once in verses of the text Zh Qin was translating. I think this is

unlikely, however; a more plausible explanation is that Zh Qin simply omitted the

word brhmaa for two reasonsat two characters, it takes up precious space within

the constraints of the six-character pda, and it is not necessary to produce a

meaningful translation in Chinese, which routinely omits the subject when it is

understood from context. 553 Indeed, given these factors, the fact that fnzh shows up

551
This word for brhmaa is a translation that contains a transcription within it. It is
based on a putative etymology of brhmaa as deriving from brahma and manas (i.e., one who has
the intellect of brahman [or Brahm]), with the character transcribing brahma and the character
translating manas. In later translations, fnzh came to be replaced as a translation for brhmaa by
plumn (), which is pure transcription.
552
Interestingly, Zh Qin does not include fnzh in his translation here either, but he does
include shmn () as a transcription for ramaa:
.
553
Strictly speaking, the same is true of Sanskrit/Pali, insofar as the subject is implicit in the
inflection of the verb; an explicit subject is usually included for emphasis. In this respect, the Ahaka
puts routine emphasis on the subject when it is the ideal person by making varied use of different
honorific terms, including bhikkhu, muni, and brhmaa. Zh Qins translation simply shows a
tendency to efface this variegated emphatic reference to the ideal person, and prefers to leave the
subject implicit.

269
even once in Zh Qins translation makes it likely, I would argue, that it was present

in the original text in other cases as well. Moreover, an examination of the Chinese

text shows that most of the verses in which brhmaa occurs in the Pali Ahaka are

preserved recognizably intact in the Chinese, and that the omission of the word

brhmaa in those cases is understandable because it does not appreciably affect the

meaning of the verse. 554 Therefore, in what follows, I will be operating under the

554
In sutta 4 of the Ahaka, brhmaa occurs twice, in v. 790 and 795. We have already
looked at the Chinese for the relevant pda in v. 790 (a) in the context of our discussion of the way in
which the Yzjng mirrors the Mahniddesas practice of reinterpreting the teaching against all views
as only referring to wrong views: , Following another path, one does not obtain
liberation. I have rendered the subject of obtain here as one because the subjectpresumably
brhmaa in the originalis omitted in the Chinese. The Chinese counterpart to v. 795, however, is
the one verse, already mentioned, where Zh Qin explicitly translates brhmaa in his translation.
The relevant pda (a) is , Without attachments, he is a Brahman.
The word brhmaa also occurs twice in sutta 5 of the Ahaka, in the final two verses, 802-3.
The Chinese equivalents to these verses are discernible, but obscured by what likely are
mistranslations. The relevant portion of v. 802 is the second hemistich, which reads, ta brhmaa
dihim andiyna, kendha lokasmi vikappayeyya, Who here in the world could have doubts
about that Brahman who has not taken a view? The Chinese, however, reads,
, which I interpret as With wisdom he looks at dharma and completely sees the meaning;
from this, he attains [Awakening?] and abandons the empty [things] of the world. The first sign that
there is a problem of translation here is the phrase (from this) for kena, when in fact kena is a
question word referring to a person (by whom?). The character (world) appears to reflect the
locative lokasmi, but it is used in a context that does have support in the Pali text. In the first pda,
Zh Qin uses two verbs meaning to see ( and ), although none are found in th Paliperhaps he
was misled by the noun dihim, which is derived from a verb meaning to see. The character
(wisdom) is difficult to explain, but may possibly be meant as a translation of ta brhmaa.
The Chinese counterpart to v. 803 is easier to reconcile with the Pali, but also appears to
involve mistranslation. The relevant pda in Pali reads, na brhmao slavatena neyyo, A Brahman
is not to be inferred by morality or vows. The Chinese, on the other hand, reads, ,
Only maintaining vows, he seeks for the truth. It is not clear to me where he seeks for the truth
comes from, but more importantly, Zh Qin says that the ideal person maintains vows, when the Pali
in fact says quite the opposite, that he is not to be inferred by morality or vows. In any case, as in v.
790 above, brhmaa appears to have been simply omitted from the translation as an implicit subject.
The Chinese counterpart to v. 843 in sutta 9 is of particular interest because here we find what
is almost certainly an interpretative translation of brhmaa that holds it to be metaphorical. The
relevant pda (a) reads, , What should the person who has truth say? The Pali,
however, reads, saccan ti so brhmao ki vadeyya, Why should that Brahman say, True? It
appears that Zh Qin has taken ki to mean what? when here it actually means why? This
renders the content of what the Brahman says (or rather does not say)namely, saccan ti (True)
superfluous, and so it appears that Zh Qin has somehow taken it as descriptive of that Brahman (so
brhmao), resulting in the phrase the person who has truth ().

270
assumption that the word brhmaa did appear in the Arthavargya in the verses of

the same stras as it appears in the Ahaka.

The first of the four suttas in the Ahaka that have the word brhmaa in

them is the Suddhahaka Sutta (4). The word appears twice in this sutta (v. 790 and

795), and the second of these instances happens to be the one case in which it is

translated explicitly as fnzh. The corresponding stra in the Yzjng, the Makara 555

the Brahman Stra, is therefore an excellent place to start with our investigation of

the way framing has been used to implicitly interpret the word brhmaa in the

There are two verses in sutta 13 of the Ahaka, 907 and 911, that contain the word brhmaa.
The first of these has no counterpart in the Chinese. There is a verse in the corresponding Chinese
stra (12) between the verses that correspond to v. 906 and v. 908 in the Pali, but as Bapat notes, it is
completely different from Pali v. 907Bapat, Arthapada Sutra, 110. The second of the two Pali
verses, however, does have a clear equivalent, the relevant pda (a) of which reads, ,
The one with wise thoughts attains anything without exception. The overall sense of the translation
is difficult to reconcile with the Palina brhmao kappam upeti sakh, The Brahman does not
attain a proper definitionalthough the sense of attaining may likely derive from the verb upeti.
More importantly, though, we find here once again what is probably an interpretive translation of
brhmaa. As already stated in n. 551, the standard translation Zh Qin uses for brhmaa is fnzh
(), which is based on an assumed etomylogical derivation of the word from brahma and manas,
and uses to transcribe brahma and to translate manas. In this verse, it appears that Zh Qin is
using to translate the supposed etomylogical roots of brhmaa even more explicitly
(intelligent) is for brahma, and (thought) is for manas. He is thus interpreting brhmaa here
as literally a person of intelligent thoughts.
The final instance of brhmaa in the Ahaka is in v. 946 of sutta 15, which has its parallel in
stra 16 of the Yzjng. The relevant portion is the first hemistich, which in Pali reads, sacc
avokkamma muni, thale tihati brhmao, Not having deviated from the truth, the sage stands on dry
ground, a Brahman. The Chinese translation, again, appears to be somewhat garbled:
, Riding the power of truth, he already is able to control his craftiness. He immediately
reaches the other [shore]; there is no concern about wisdom. Nevertheless, we can still see the
connection to the Pali in the characters (truth) for sacc and (which can also mean to stand)
for tihati. The word brhmao, together with its appositive muni, appears to have been omitted as an
implicit subject, or else incorporated in the sense of wisdom ().
555
The name given for this Brahman in the Chinese is Mji (), which is a standard
transcription for makara, which refers to a sea monster or monster-like sea-borne animal (such as a
crocodile). Bapat prefers to read Mji as a transcription for Mgadha, however, since it is also
attested as a transcription for Magadha (the janapada). I nonetheless prefer the reading makara, which
is consistent with the tendency of early Buddhist texts, as we will discuss further in Part IV, to give
Brahmans funny names.

271
Arthavargya, since it is the one stra for which there is absolutely no doubt that the

word brhmaa appeared in the verses of the Indian original. The prose narrative for

the Makara the Brahman Stra is fairly short compared to other prose narratives in

the Yzjng. It begins by introducing the character for whom the stra is named

Makara, a Brahman living in rvasti who died while the Buddha happened to be

staying in town at Jetavanrma. His fellow-students () put his body on

display and make the following announcement around town: Those who saw

Makara [when he was alive] all attained liberation; if you now see his corpse, you

will also be liberated; later on, those who hear his name will also be liberated. 556

The Buddhist monks hear this while they are on their alms round, and they report it to

the Buddha. In response, the Buddha utters verses corresponding to sutta 4 in the

Ahaka. The connection between the story and the verses appears to be based on the

first verse (v. 788), which in the Pali criticizes the idea that dihena sasuddhi

narassa hotiA man has purification by means of what is seen. With the

addition of the prose narrative, therefore, the sophisticated and doctrinally

problematic teaching on the independence of purity from anything at all (what is seen,

heard, or felt, as well as morality and vows) is reduced to a criticism of the more

specific claim that one can attain purity simply by looking at a holy mans corpse or

hearing his name.

The framing provided by the prose narrative not only serves to smooth over

doctrinal difficulties; it also serves to contextualize the use of the word brhmaa. In

556
T. 198: .

272
the original verses, brhmaa is simply used as an honorific term to refer to the ideal

person, who is beyond boundaries (v. 795) and who therefore does not say that

purity is from something else, in what is seen, in what is heard, in morality and vows,

or in what is thought (790). There is no contrast, explicit or implied, between this

Brahman and some other sort of Brahman; the only contrast is between the

Brahman who is beyond boundaries and the non-Brahman who is not. The prose

narrative of the Arthavargya provides the former sort of contrast and conflates it with

the latter. We now have literal Brahmansthat is, members of the social group

that call themselves Brahmansbeing contrasted with the Buddhist or true

Brahman of the versesprecisely on the basis that they, unlike the Brahman of the

verses, think that purity comes from seeing something, namely their dead comrade.

Other suttas of the Ahaka that have the word brhmaa in their verses are

paralleled by stras in the Yzjng that contain prose sections that contextualize the

term in much the same wayby introducing literal Brahmans to be contrasted with

the ideal Brahman of the verses. The fifth sutta of the Ahaka, the Paramahaka

Sutta, contains the word brhmaa in its final two verses (v. 786-7), and it is

paralleled by the fifth stra of the Yzjng, the King Mirror-Face Stra. We have

already discussed the prose section of the latter version briefly in the context of the

comparison to the prose narratives in Buddhaghosas commentaryboth the story in

the Yzjng and the story provided by Buddhaghosa incorporate the well-known tale

of the blind men and the elephant. This story is indeed well-suited to the

corresponding verses, which criticize people who cling to their own view and think

273
that it is better than anyone elses. There are several differences, as already

mentioned, between the Yzjng-version of the story and Buddhaghosas, but one in

particular is important for our purposes: the identity of the people fighting over who

has the highest view, whom the parable of the blind men and the elephant is intended

to criticize. In Buddhaghosas version, they are simply called various sectarians

(nntitthiy), and the king summons a bunch of blind men to describe an elephant so

as to demonstrate to them the ridiculousness of their sectarian squabbles. In the

Yzjng, however, they are specifically identified as Brahmans. Not only this, but

they are used by the Buddha to introduce the story of the blind men in the following

way: The Buddha said, These Brahmans are not stupid in [only this] one lifetime.

Long in the past 557 and then he recounts the story of the blind men and the

elephant, ending by saying, At that time I myself was King Mirror-Face, and these

Brahmans of the debate hall were the blind men. 558 Once again, then, the author of

the Arthavargya prose has gone out of his way to introduce literal Brahmans as

characters into his story to exemplify what is criticized by the verses, and thus serve

as a counterpoint to the ideal Brahman of the verses.

The Mgandiya Sutta, number 9 in the Ahaka, also features a Brahman

namely, Mgandiya himselfin the prose narrative found in the parallel version

(stra 9) in the Yzjng. Since we have already discussed this sutta and the story

associated with it more than once, it is unnecessary to discuss it in any great detail

557
T. 198: .
558
T. 198: .

274
here. I would only like to once again point out that Mgandika is presented as a

Brahman in the Chinese version, and that even the author of the Mahniddesa seems

to be aware of a story involving a Mgandiya who is a Brahman, since he identifies

him as such in his glosses of the name. As I have already argued, moreover, it is

likely that the story of the Brahman Mgandiya was implicitly attached to this set of

Ahaka/Arthavargya verses quite early in the compilation of the text, through the

addition of a verse at the beginning in which the Buddha rejects a woman as full of

urine and excrement. There is little that can be said about the framing effect that

the story has on the verses, however, since the connection of the story to the verses is

quite tenuous, and in fact seems to be based on little more than the fact that the

interlocutor in the verses is named Mgandiya. What we have here is less an attempt

to frame, contextualize, or interpret the verses than simply an attempt to explain who

Mgandiya is, by connecting it to a story that also has a character named Mgandiya

in it.

Sutta 13 in the Ahaka, which is paralleled by stra 12 in the Yzjng,

contains two verses (907 and 911) that use the word brhmaa, although the first of

these verses appears to be completely different in the Chinese translation. 559 The

prose narrative that precedes the verses in the Chinese version serves primarily to

introduce a certain Brahman named Sees-Dharma (), 560 after whom the stra is

559
See note 554, above.
560
Bapat suggests, plausibly, that this name may be a translation of dharmadarBapat,
Arthapada Sutra, 104.

275
entitled. Unlike in other stories, this Brahman does not appear to be a hostile or even

non-Buddhist interlocutor; on the contrary, he appears within a large crowd of

human and divine beings who have come to pay their respects to the Buddha. This

Brahman is nonetheless singled out, however, because he holds a pernicious view:

At that time, among the seated there was a Brahman named Sees-Dharma who
was also among the large crowd. Because of his karma, he held the view that
one who is liberated in nirva has a body; therefore, he gave rise to doubt on
account of his karma. The Buddha knew the doubt that the Brahman Sees-
Dharma had given rise to, so at that time he created a Buddha, well-built and
incomparable in form. All who saw him were pleased. He had the 32 marks
of a Great Man; he had a golden color and also shined; he wore the great
garment of dharma; and so forth as said above. 561

The purpose of this second, mind-made Buddha that the Buddha creates is to

pronounce the verses that are parallel to sutta 13 of the Ahaka, presumably in an

effort to free Sees-Dharma of his pernicious view that liberation is a purely corporeal

phenomenon.

Unfortunately, the effort to discern the connection between the frame

narrative and the verses is complicated somewhat by the fact that the Chinese version

of the verses is more substantially different from the Pali than in other stras,

differing both in order and in some cases in content. 562 In spite of this difficulty,

generally speaking both the Chinese version and the Pali version of the verses can be

described as being, as with many other suttas in the Ahaka, a criticism of those who

cling to their own view and enter into disputes over views, and concomitantly a praise

561
T. 198:

562
See Bapat, Arthapada Sutra, 106-14, for a full accounting of the differences.

276
of those who do not cling to views and thus are beyond disputes. It is not entirely

clear to me how this theme speaks to Sees-Dharmas pernicious view that the person

who is liberated has a body, although a criticism of those who see only name-and-

form (v. 909) may serve as a tenuous connection in this regard.

Regardless of the precise logical connection between the story and the verses,

however, the frame story clearly does introduce, once again, a literal Brahman who

can serve as a counterpoint to the ideal Brahman spoken of in the verses. Although

Sees-Dharma has a rather auspicious name and certainly cannot be understood as an

other in the same sense as some of the other Brahmans we have encountered, his

doubt, and the pernicious view on which it is based, serves as a counterpoint to the

ideal that is represented by the Brahman in the verses. His view evinces a

materialistic understanding of the Buddhist teaching that does not recognize that the

liberation that it points to transcends material form. Even more so than by the verses,

in fact, this view is successfully refuted by the Buddhas initial act of creating another

Buddha out of thin air. 563 This Buddhas ethereal body is marked by auspicious

signsincluding the 32 marks of a Great Man, which, as we will see in Part IV, are

particularly sought out by Brahmans in numerous encounter dialogs among the

early Buddhist textsbut these signs, as well as the body as a whole, are ultimately

not real. If the Brahman of the verses represents the ideal, then the literal Brahman

563
This would presumably be an early example of the Buddha being portrayed as producing a
nirmakya (lit., constructed body)not as a divine sambhogakya creating a phantasm of himself
to play out the life of the historical Buddha, as we find in the developed Mahyna doctrine of the
Three Bodies, but as a human (but fully Awakened) being making use of his special powers to
construct an ethereal body. See David Gordon White, Sinister Yogis (Chicago: University of Chicago
Press, 2009), 177-82.

277
represents the opposite of that ideala person who thinks in terms of material form

and is impressed, like other members of his vara, by the outward signs thereof.

This leaves one remaining sutta, number 15 in the Ahaka, that uses the word

brhmaa to refer to the ideal person. As I have already noted, this sutta is a bit of an

outlier among the Brahman-texts of the Ahaka insofar as it is not included among

the core suttas identified by Vetter; as such, the Brahman in this sutta represents an

ideal much closer to what one would expect according to mainstream early Buddhist

doctrine than what we find in the core suttas. Overall, the sutta takes up fairly

ordinary themes such as abandoning desire and becoming free from attachments, but

it begins with a brief reference to the fear that comes from those who have taken up

the stick (attada)from which the Pali title (attadaasutta) is derived. The

Arthavargya from which the Yzjng was translated appears to interpret the verses as

a whole as an exhortation against violence, and thus it introduces them with a long

prose narrative in which a king, because of a perceived slight, slaughters all of the

Sakyas, prompting the Buddha to predict that he will pay dearly for this sin in future

lifetimes. Although this narrative serves as another interesting example of how

framing can be used in general to interpret a set of versesin this case, by

emphasizing the theme of violenceit does not introduce Brahmans as characters in

any significant way, and thus it essentially ignores the use of the word brhmaa in

the verses.

With the exception of sutta 15, then, the parallel versions of all the Ahaka

suttas that use brhmaa to refer to the ideal person introduce, through the prose

278
narrative frame, a literal Brahman that serves as a counterpoint to the ideal

Brahman who is described in the verses. Insofar as the Brahman in the prose

narrative embodies the claims of the new Brahmanismnamely, that society is to

be divided into four varas, and that certain people are the highest of these four

varas, i.e., Brahmans, by virtue of their birthit transforms the word brhmaa in

the verses from a straightforward honorific for the ideal person into a polemical

rejoinder to the conception of the Brahman embodied by the character in the frame

narrative. This does not necessarily mean that at the time that a frame narrative was

added to a set of verses, the person who did so accepted that the proponenets of the

new Brahmanism are really Brahmansindeed, the whole point of counterposing

the ideal Brahman represented in the verses seems to be that they are notbut it does

demonstrate a degree of acceptance that there are real social agents in the world

called Brahmans. And as Bronkhorst has argued, this is exactly how Brahmanism

spread: By insinuating its socio-ideological scheme into the public discourse by

repeating their claims about vara and vara-superiority based on birth, proponents

of the new Brahmanism forced opponents to respond to them on their own terms.

The end result, as I argued in Part II, was the abandonment by the Buddhists of the

category brhmaa in favor of the category samaaas exemplified by the

encounter dialogs in which the Buddha is clearly identified as belonging to the latter

category and his interlocutors are clearly identified as belonging to the former.

Framing, such as we have seen here in the context of the Yzjng, was a step in this

process.

279
 Moreover, even though the Yzjng represents an effort to frame only a small

collection of verses out of the many that circulated in the early Buddhist tradition, and

even though it in and of itself was apparently unknown to the Pali tradition, its

technique of introducing literal Brahmans to serve as counterpoints to ideal

Brahmans in verses is neither unique nor foreign to the Pali tradition. I have already

mentioned that there are many collections in the Pali traditionwhether considered

fully canonical, or else straddling the canonical-commentarial dividethat

encapsulate verses in prose frame narratives in the same way that the Yzjng does.

Within these, the Brhmaa-sayutta of the Sagth-vagga of the Sayutta Nikya,

the commentary on the Brhmaa-vagga of the Dhammapada, and to a lesser extent

the Bodhi-vagga of the Udna all contain verses that refer to the ideal person as a

Brahman, which are then framed by prose narratives that introduce literal

Brahmansoften given funny names to emphasize their buffoonish qualitywho

serve as a counterpoint to the ideal Brahman described in the verses. At times the

verses already contain a contrast between the ideal Brahman and the literal

Brahman, but often they do not, and thus the prose narratives serve to either introduce

for the first time or at the very least accentuate the sense of dichotomy between the

real meaning of brhmaa as a normative ideal and the literal meaning of

brhmaa as a social reality. The use of the literal Brahman as a framing device

was therefore a commonplace in the early Buddhist tradition, clearly transcending

sectarian boundaries.

280
III.2.3 Framing in the Pryaa

As I have discussed above, the Pryaa is less intrinsically interesting as an

exemplar of an early, pre-orthodox tradition than the Ahaka because its overall

unifying theme is the rather orthodox goal of escaping from rebirth, and it places less

emphasis on such radical ideas as the rejection of all views. This is reflected as well

in the use of the word brhmaa, which include more instances of the canonically

common compound samaabrhmaa and even uses that clearly refer to a particular

social group. Nevertheless, there are certain dialogs in the Pryaa that do touch on

themes similar to those found in the Ahaka; likewise, there are certain uses of the

word brhmaaparticularly in the dialogs with Mettag (4, v. 1059), Dhotaka (5, v.

1063, 1065), and Posla (14, v. 1115)that are quite similar to the usage found in the

Ahaka. In the case of the dialog with Dhotaka, the usage of the word brhmaa

actually goes further than the Ahaka, not only using it as an honorific for the ideal

person in the abstract, but using it for the Buddha himself. And as we have already

seen, nothing in the immediate context of these three dialogs suggests, any more than

in the case of the Ahaka, that the use of brhmaa as an honorific is any more

unique or metaphorical than that of other honorifics such as muni. Therefore, it

will be useful to explore how framing has contextualized and thus given an implicit

interpretation to these otherwise straightforward uses of the word brhmaa to refer

to the ideal person.

Unlike in the case of the Ahaka, the most significant examples of framing are

not to be found in commentaries or other extracanonical texts; rather, they are to be

281
found in the canonical Pali text of the Pryaa itself. The oldest commentary on the

Pryaa, the Caniddesa, does provide a certain amount of framing, but not in such

a way as to really add anything not already found in the text of the Pryaa itself.

When the word brhmaa does appear in the latter in reference to the ideal person,

the Caniddesa provides the exact same gloss as does the Mahniddesa, defining the

brhmaa in terms of an ideal more consonant with orthodox canonical doctrine than

the immediate context of the Pryaa or Ahaka. 564 The Caniddesa also provides

a small amount of information, not included in the verses on which it comments,

about a figure named Bvar, but as we will see, this additional information is, for the

most part, already present in the canonical text of the Pryaa in the vatthu-gths.

The extracanonical Paramatthajotik II commentary also provides little additional

framing; although it does contain some prose narrative, this serves merely to

elaborate upon the narrative already found in the vatthu-gths. If we are to

understand how the dialogs of the Pryaa have been framed, then, we must look

within the canonical text of the Pryaa itself, and in particular to the vatthu-gths.

In order to understand how this framing process took place, it will be useful to

begin by looking at the finished productthe text of the Pryaa as it stands in the

Canon, with the vatthu-gths and all its other constituent parts. In this canonical

form, the Pryaa has a total of 19 parts: the introduction (vatthu-gths), 16

dialogs with interlocutors identified as Brahmans, and two epilogues (atthuti-gths

564
The gloss is word-for-word the same as found in the Mahniddesa, except that in the case
of v. 1063, in the dialog with Dhotaka, the word bhagav is added at the beginning to make clear that
the Brahman being referred to is the Buddha himself.

282
and anugti-gths). The text begins in the vatthu-gths with an extended narrative

explaining how a certain Brahman named Bvar came to send his 16 disciples to the

Buddha. Although he is not explicitly described as such, this Bvar clearly fits the

type of a jaila: Desiring the absence of possessions (v. 975: kicaa

patthayno), he lived on the bank of the Godhvar, off of gleanings and fruit (v.

977: vasi godhvarkle uchena ca phalena ca), where he performed a great

sacrifice (v. 978: mahyaam akappayi) and dwelt in an rama (v. 979). After

performing the great sacrifice, he is visited by another Brahman who is dirty, thirsty,

and hungry and demands five hundred pieces of money from him (v. 979-980).

When Bvar explains that he does not have any money to give him, the other

Brahman gets angry and declares that Bvars head will split open if he does not give

him the money within seven days (v. 982-983). Bvar is understandably distressed

by this threat, but he is put at ease by a deity (devat) who tells him that the other

Brahman is just a cheatdesiring money (v. 987: kuhakodhanatthiko) who

knows nothing about head-splitting. Bvar therefore asks the deity what he knows

about head-splitting, and the latter tells him he has to speak to the Buddha if he wants

to know more (v. 988-993). Bvar is excited to hear that there is a Buddha in the

world, and so he sends his 16 disciples to go find him (997-998). They ask him how

they will know the Buddha when they meet him, and Bvar replies that there are two

ways of knowing: First, the Buddha will have the thirty-two marks of a great man

(v. 1001: mahpurisalakkha), and second, he will be able to read their minds (v.

1004-1005). The Brahman students ultimately find the Buddha at the Psaka cetiya

283
in Magadha, where they find that he indeed has the full set of marks and is able to

read their mind (v. 1013-1024). Excited that they have actually found the Buddha,

they ask him about head-splitting, and the Buddha replies with a metaphorical

explanation in which the head stands for ignorance, which is in turn split by

knowledge, faith, mindfulness, concentration, resolution, and effort. 565

The 16 sections of the text that follow are short dialogs in which each of the

16 Brahmans asks the Buddha his own individual questions. The dialogs cover a

variety of topics that are generally united by the common theme of going to the far

shore (pra) of birth and old agethat is, escaping rebirth. They are followed by

two epilogues, the first of which contains a mixture of prose and verse, and the

second of which contains only verse. The first epilogue is, in a sense, the most

internally coherent of the two; it repeats the names of the 16 Brahmans who spoke to

the Buddha; explains why the Buddhas teaching to them is called the Pryaa

(because it leads to the far shore); and records that after hearing this teaching, the

16 Brahmans practiced brahmacariya in the presence of the one of excellent wisdom

(v. 1128: brahmacariya acarisu varapaassa santike).

The second epilogue, on the other hand, is less internally coherent, but in a

sense provides ultimate closure to the text. In this second epilogue, Pigiya, the last

of the 16 Brahmans who spoke to the Buddha, announces, I will sing Going to the

Far Shore (v. 1131: pryanam anugyissa), and then engages in an extended

565
Sn. 1026: avijj muddhti jnhi, vijj muddhdhiptin / saddhsatisamdhhi,
chandavriyena sayut //.

284
conversation with an unnamed interlocutor, whom he addresses as brhmaa, in

which he heaps praises on the Buddha, whom he just met, and says that he cannot go

back to see him because of his old age and therefore visits him in his mind (v. 1142-

1144). According to the Caniddesa, the Brahman with whom Pigiya is speaking

in this second epilogue is his teacher Bvar 566; this is confirmed by Buddhaghosa in

his commentary on the anugti-gth in the Paramatthajotik II. Indeed, this would

appear to be the logical interpretation of the second epilogue, given the context of the

canonical version of the Pryaa, including the introductory story about how the 16

Brahmans were sent by Bvar, in which it appears. The second epilogue brings

closure to this story by having one of Bvars students return (presumably) to him

and report to him the greatness of the Buddha.

While the canonical version of the Pryaa does appear to present a fairly

coherent narrative overallfrom Bvar dispatching his disciples to the Buddha at the

beginning to Pigiyas recounting of the greatness of the Buddha at the endthere

are nevertheless several inconsistencies that prevent a straightforward interpretation

of the text as a forming a cohesive narrative. To begin with, in the vatthu-gths,

when Bvar asks the deity who appears to him where he can find the Buddha, the

deity answers, The conqueror is in Svatth, palace of the Kosalas (v. 996:

svatthiya kosalamandire jino). 567 When Bvars 16 students actually go to find

the Buddha, however, they pass through many cities, including Svatth, but they
566
See the gloss on the vocative brhmaa in v. 1140: brhma ti gravena mtula lapati.
567
Interestingly, this and the immediately preceding verse, in which Bvar asks the question,
are in tuhubha, while the rest of the vatthu-gths are in vatta.

285
ultimately end up at the Psaka cetiya in Magadha (v. 1011-13), which is a

completely different janapada to the east of Kosala. Although Buddhaghosas

commentary explains that the Buddha left Svatth after the Brahmans began their

journey, and therefore the latter had to continue their journey past Svatth to find

him, 568 the text itself contains no such explanation; v. 1011-13 are simply consistent

with the prose section in the first epilogue, which reports that the Blessed One said

this while he was dwelling among the Magadhans at the Psaka cetiya, 569 and

inconsistent with the deitys earlier statement in v. 996.

Second, in spite of the rather lengthy introduction at the beginning of the text,

in which Bvar figures prominently as the main character, Bvars name is never

mentioned again in the entire Pryaa. This is the case in all of the 16 dialogs that

form the core of the Pryaa; when the 16 Brahmans approach the Buddha, none of

them make any mention of their teacher, the mission they have been sent on, or

indeed any of the themes dealt with in the vatthu-gth. It is also the case in the first

epilogue; although the names of the 16 Brahmans are listed, no mention is made of

them having been sent by Bvar. Most surprisingly perhaps, Bvar is not even

mentioned in the second epilogue, which appears from context to be, and indeed is

interpreted by the commentaries as, a dialog between Pigiya and his teacher Bvar.

When read by itself, without the benefit of the context provided by the vatthu-gth,

568
See Paramatthajotik II, commentary on the vatthu-gth of the Pryaa, v. 1013.
569
idam avoca bhagav magadhesu viharanto psake cetiye.

286
the Caniddesa, or the Paramatthajotik II, the anugti-gth is simply a dialog

between Pigiya and an unnamed Brahman.

This brings us to the third inconsistency, which relates to the character of

Pigiya himself. As already mentioned, Pigiya is the last Brahman to speak to the

Buddha, and as such his dialog with the Buddha is the last of the 16 dialogs that form

the core of the Pryaa (v. 1120-1123). When Pigiya speaks to the Buddha, the

first thing he says is, I am old, weak, my complexion gone (v. 1120: jioham

asmi abalo vtavao). This is confirmed by the second epilogue, in which Pigiya

says that he cannot return physically to the Buddha because I am old and of feeble

strength (v. 1144: jiassa me dubbalathmakassa). Yet the introductory story is to

have us believe that he, along with the other 15, is a student (sissa) of Bvara

narrative element reinforced by the use of the word mava in the titles of each of the

16 core dialogs. 570 At the same time, there is also an inconsistency in what Pigiya

did after meeting with the Buddha. As already mentioned, in the second epilogue,

Pigiya tells whomever he is talking towhether Bvar or someone elsethat he

cannot return to the Buddha in person because he is too old. But in the first epilogue,

it had already been said that all 16 Brahmans practiced brahmacariya in the

presence (v. 1128: santike) of the Buddha. The canonical text of the Pryaa,

therefore, cannot decide whether Pigiya is an old man or a young student, whether

570
Thus, Pigiyas dialog with the Buddha is entitled pigiya-mava-pucch. The
manuscript traditions of the Caniddesa are not, however, consistent about the titles of the individual
dialogs in the Pryaasee W. Stede, ed., Niddesa, vol. 3 (London: Pali Text Society, 1918), xx-xxi.

287
he eagerly practiced brahmacariya in the presence of the Buddha or left and could

only return to him in his mind due to the feebleness of his body.

Finally, there is a small but troublesome inconsistency introduced into the text

by v. 1146 of the second epilogue. In this verse, whomever Pigiya is speaking to,

after listening to Pigiyas effusive praises of the Buddha, says to him,

As Vakkali has declared his faith, and Bhadrvudha and avi Gotama,
Even so should you too declare your faith.
You will go, Pigiya, to the far shore of the realm of death. 571

Unfortunately, nowhere in the text of the Pryaa are the declarations of faith of

these three people mentioned. Not only this, but of the three, only Bhadrvudhas

name is mentioned elsewhere in the Pryaa, as one of the 16 students of Bvar.

The other two have no apparent place in the narrative of the Pryaa, and one of

themavi Gotamais not attested anywhere else other than this verse in the entire

Pali Canon! 572

These inconsistencies are not entirely surprising, however, if we assume that

the Pryaa is a composite text. As already discussed above, Vetter has come out in

favor of this position, and many scholars have assumed that at the very least the

vatthu-gths must be later than the rest of the Pryaa since they are not

commented upon by the Caniddesa. N. A. Jayawickrama, in his classic study of the

Sutta Nipta, has also adduced convincing arguments for the lateness of the vatthu-

571
yath ah vakkali muttasaddho, bhadrvudho avi gotamo ca / evam eva tvam pi
pamucassu saddha / gamissasi tva pigiya maccudheyyassa pra //. See Norman, Group of
Discourses, 389-90 n. 1146, for a discussion of the difficulties in translating muttasaddho.
572
A monk by the name of Vakkali is referred to in two other places: SN 3.1.9.5 and AN 1.208.

288
gth on the basis of internal evidence, including the use of late terminology such as

visaya (v. 977) and mandira (v. 996), the inclusion of the developed legend of the

Buddha as a Great Man possessing 32 marks, and a reference to the fairly late

doctrine of vsan (1009). 573 This does not mean that the story of Bvar was not

known in some form to the author of the Niddesaindeed, the references to Bvar in

the Niddesa, however oblique, make it certain that its author was aware of the story in

some form 574but it does make it unlikely that he was aware of the vatthu-gth as

we have them today, or at the very least, considered them to be authoritative. 575 The

supposition that there was a Bvar legend associated with the Pryaa prior to

the composition of the Pali vatthu-gth is strongly supported by fragments of a

Sanskrit version of the Pryaa, which refer to Bvar by name, but in a context that

cannot be directly correlated to the Pali text of the vatthu-gth. 576 Thus, it is likely

that the author of the Niddesa had access to a text of the Pryaa that did not

573
N. A. Jayawickrama, The Vaggas of the Sutta Nipta, University of Ceylon Review 6
(1948), 243-249.
574
Bvar is referred to in the Caniddesa in the commentaries on v. 1084, the prose section
at the beginning of the first epilogue, v. 1134, v. 1135, and v. 1138. In the last of these commentaries,
the Niddesa author tells us that Pigiya was Bvars nephew (pigiy ti bvar ta nattra nmena
lapati), which provides additional confirmation that the Niddesas information on Bvar was not
simply based on the vatthu-gth, since the latter makes no mention of Pigiyas relationship to Bvar.
575
Jayawickrama writes, The probability is that both the v.g. and Nd 2 were not separated
from each other by a long interval of time, and that the subject matter of the v.g. may have existed in
some form before Nd 2 was compiled, and that the latter was influenced by itJayawickrama,
Vaggas of the Sutta Nipta, 248.
576
See Norman, Group of Discourses, 359 n. 976-1031. The relevant fragment, which
Norman cites, is Turfan fragment no. 1582, published in Heinz Bechert and Klaus Wille, eds.,
Sanskrithandschriften aus den Turfanfunden, vol. 6 (Stuttgart: Franz Steiner Verlag Wiesbaden
GMBH, 1989), 199-200. Other fragments from the Pryaa (or, more properly speaking, fragments
that show parallels to certain parts of the Pryaa) are found in the same volume on p. 198 and in
Lore Sander and Ernst Waldschmidt, eds., Sanskrithandschriften aus den Turfanfunden 4 (Wiesbaden:
Franz Steiner Verlag GMBH, 1980), p. 236-38.

289
include the vatthu-gth, but was aware of some sort of frame story according to

which the 16 Brahmans who visited the Buddha were students of a Brahman named

Bvar. This story came to be recorded at different times in different waysthrough

occasional comments in the Caniddesa, through a more extended narrative in the

Paramatthajotik II, through the vatthu-gth of the canonical Pali version, and

through the story, now mostly lost, that was apparently included in a Sanskrit version

of the Pryaa. All of these, of course, are quite late, and represent attempts to

frame the Pryaa within a particular narrative.

What does the Pryaa look like without this particular frame narrative?

Quite different, as it turns out. To begin with, the first 92 verses of the Pryaa as it

was commented upon by the Caniddesathat is, the core 16 dialogs of the text

introduce a series of interlocutors who discuss with the Buddha a range of topics

concerning meditation, morality, and the path to deliverance from rebirth, without any

mention Bvar; head-splitting; the Buddhas 32 marks; or indeed even, for the most

part, anything about their own identity or why they are there. 577 Only in the first

epilogue do we finally get some information about the context of the 16 dialogs

where they happened and who the interlocutors were. The information provided here

577
I am, of course, ignoring the titles of the dialogs, which, as I have already discussed,
follow the story in the vatthu-gth by referring to the 16 interlocutors as mava, but, based on the
inconsistent evidence provided by Niddesa manuscripts, are almost certainly latesee n. 569 above.
The actual verses contained in these 16 dialogs provide almost no information about the interlocutors
identities other than their names; the one exception, which we already looked at earlier, is v. 1100 in
the Jatukai-mava-pucch (11), in which the Buddha refers to his interlocutor with the vocative
brhmaa. This verse (along with the two immediately preceding, which also present the Buddhas
words) is in the vatta meter, unlike v. 1096-7 (which present Jatukais question), which are in
tuhubha. Although we can of course only speculate, it is possible that additions, such as the vatta
verses of this dialog, were made by the same person who wrote the first epilogue, which is also in vatta
meter, to a pre-existing text or texts to fill out a desired scheme of 16 question-and-answer dialogs.

290
does not strictly speaking contradict what is found in the vatthu-gth, but it does

little to confirm it; on the one hand, it serves as an ill-fitting conclusion to the story

begun by the vatthu-gth insofar as it makes no mention of Bvar or the concerns

that supposedly led him to send his 16 student to the Buddha in the first place, and on

the other hand, it reads much as if it is explaining the circumstances of the 16 dialogs

for the first time, rather than concluding a story that had mostly been told already.

The first epilogue begins with the following statement, in prose, giving the

setting of the 92 verses that had preceded it: The Blessed One said this while

dwelling among the Magadhans at the Psaka cetiya. 578 No mention is made of

Svatth, where the deity in the vatthu-gth told Bvar the Buddha would be

dwelling, nor of any movement that led to him being found in Magadha. The prose

section then continues, Requested and asked in turn by the 16 attendant Brahmans,

he answered their questions. 579 This is the first time that we hear that the 16

interlocutors, as a group, are Brahmans. Nothing else is said about their identity as

Brahmans, however. They are not described as jailas; they are not said to be

students of another Brahman, named Bvar or otherwise; indeed, they are not said to

be related to one another in any particular way. Indeed, the way their names are listed

in the first three verses of this epilogue seems to suggest that they are not:

Ajita, Tissametteyya, Puaka and Mettag,

Dhotaka and Upasva, Nanda and also Hemaka, (v. 1124)

Both Todeyya and Kappa, and Jatuka the wise,

578
idam avoca bhagav magadhesu viharanto psake cetiye.
579
paricrakasoasna brhmana ajjhiho puho puho paha byksi.

291
 Bhadrvudha and Udaya, as well as Posla the Brahman,
Mogharja the intelligent, and Pigiya the great seer, (v. 1125)

These approached the Buddha, the seer of perfect conduct.

Asking subtle questions, they approached the best of Buddhas. (v. 1126) 580

As Vetter notes, the grand epithets used for the people in this list, such as mahisi for

Pigiya, suggests that they are prominent figures in their own right, rather than

students sent to do their masters bidding. 581 The same list, it must be noted, is given

in the vatthu-gth (v. 1006-8), but it is likely that it was simply copied from the first

epilogue, where it makes more sense in context, to produce a sense of consistency in

the resultant text when the vatthu-gth was added. 582

The second epilogue, when bereft of the context provided by the vatthu-gth,

is difficult to interpret. It is, to begin with, unexpected, since it serves to continue the

text after the first epilogue seemingly already brought it to a close. 583 In it, Pigiya

praises the Buddha to an interlocutor whom he refers to as Brahman, and that

interlocutor responds in a few verses (v. 1138-9, 1146), but it is not clear who that

interlocutor is. Insofar as it acknowledges that Pigiya is an old man (v. 1144), in

580
ajito tissametteyyo, puako atha mettag / dhotako upasvo ca, nando ca atha hemako //
todeyya-kapp dubhayo, jatuka ca paito / bhadrvudho udayo ca, poslo cpi brhmao /
mogharj ca medhv, pigiyo ca mahisi // ete buddha upgacchu, sampannacaraa isi /
pucchant nipue pahe, buddhaseha upgamu //.
581
Vetter, Some Remarks, 38.
582
Verse 1009, which immediately follows the list of the 16 Brahmans in the vatthu-gth,
nods to the grandiose titles given to them in the list by saying that they all had their separate groups,
[and were] famous throughout the world (paccekagaino sabbe sabbalokassa vissut). This seems
contrived, however; how can these 16 Brahmans be both mavas and famous teachers in their own
right?
583
The result is a jarring disjunction between v. 1130, the last verse of the first epilogue, and
v. 1131, the first verse of the second epilogue.

292
fact, it agrees better with the core dialog between Pigiya and the Buddha, where the

former is also said to be an old man, than with vatthu-gth, in which he is said to be

Bvars student. On the other hand, as already noted, it does not agree with the first

epilogue insofar as it says that the Pigiya is too old to return physically to the

Buddha, while the first epilogue says that he practiced brahmacariya in the Buddhas

presence. Even worse, as we have also noted, it makes a totally inexplicable

reference to declarations of faith made by Vakkali, Bhadrvudha, and avi Gotama

(v. 1146), even though no such declarations of faith are recounted anywhere in the

Pryaa, and two of the three people are not even mentioned in it eitherin spite of

the fact that an author familiar with the contents of the vatthu-gth, the first epilogue,

or even just the 16 core dialogs would have 16 (17 if we count Bvar) names to

choose from! It is hard to avoid the conclusion, then, that the second epilogue was

either taken as a whole, or else cobbled together from pieces 584 that were taken, from

a context or contexts now lost. In any case, given its redundancy after the closure

provided by the first epilogue, as well as its lack of consistency with it, it seems likely

that the second epilogue was, as Vetter suggests, an earlier addition. 585

584
Even the internal composition of the anugti-gth is somewhat suspect. For one thing, it
exhibits a remarkable heterogeneity of metersv. 1131-2, 1135-41, and 1147-8 are in vatta; v. 1133-4,
1142-4, and 1146 are in tuhubha; while v. 1145 and 1149 are mixed. This is not conclusive in and of
itself, but it is suggestive. In addition, while I have been emphasizing v. 1142-1144 (all in vatta), in
which Pigiya clearly states that he is not physically with the Buddha and can only return to him in his
mind because of his feebleness in old age, these are immediately preceded by verses in tuhubha in
which he tells his interlocutor, I am not absent from him [i.e., the Buddha] even for a moment,
Brahman (v. 1140: nha tamh vippavasmi muhuttam api brhmaa).
585
Vetter, Some Remarks, 38.

293
 If we look at just the 16 core dialogs and and the first epilogue by themselves,

what we see is a Pryaa with a very simple frame narrativewhat I will call for

convenience the inner-frame Pryaa. Sixteen Brahmans approach the Buddha at

the Psaka cetiya in Magadha with various questions about meditation, morality,

and the attainment of liberation, and the Buddha answers their questions and in the

process teaches them how to go to the far shore of birth and old age. They decide

to practice brahmacariya under him, and as a result, they attain liberation. Finally,

the reader or listener is told that he too shall attain liberation if he practices in

accordance with the teaching of the Pryaa.

But what of the category Brahman in this Pryaa? At first glance, the

Pryaa may appear much like an encounter dialog, in which the Buddha

encounters interlocutors clearly identified as Brahmans and debates with them. In

the inner-frame Pryaa, however, the Buddha does not really debate with the

Brahmans who approach him; rather, they approach him as a knowledgeable teacher

and he instructs them. Moreover, in an encounter dialog, not only are the

interlocutors clearly identified as Brahmans, but the Buddha is clearly identified as

belonging to a separate category, samaa. In the inner-frame Pryaa, this is not

the case; in fact, we can even say the opposite is true: The Buddha is portrayed as

categorically similar to his interlocutors, and he is even, in one case, referred to

explicitly as a Brahman by his interlocutor (v. 1063 and 1065 in the Dhotaka-

294
mava-pucch). More commonly, the Buddha is referred to as a Great Seer

(mahesi=mahisi), 586 an epithet shared by one of his interlocutors, Pigiya (v. 1125).

Finally, as we have seen, the 16 Brahmans who approach the Buddha are

referred to in the prose section of the first epilogue as 16 attendant Brahmans

(paricrakasoasna brhmana). Given the context provides by the vatthu-

gth, this can be interpreted as referring to the fact that they are students, and thus

attendants, of Bvar; indeed, this is one interpretation that the Caniddesa

provides for the phrase. 587 The Caniddesa also provides another interpretation for

the phrase, however, one that better fits the immediate context of the first epilogue

and the inner frame it provides for the Pryaa: Or, alternatively, those sixteen

Brahmans are attendants (paddh, paddhacar, paricrak) [and] students of the

Buddhaand thus [it says] 16 attendant Brahmans. 588 The inner-frame Pryaa,

therefore, portrays the Buddha as a Brahman amongst Brahmans. Their

Brahmanhood is neither portrayed in terms of the new Brahmanical criteria of birth

and knowledge of the Vedas, nor is it contrasted to such an understanding of

Brahmanhood 589; instead, it is used as an honorific, equally of the Buddha, his

interlocutors, and the ideal person in the abstract, much as in the Ahaka.

586
Sn. 5.4.1054, 5.4.1057, 5.5.1061, 5.5.1067, 5.7.1083.
587
paricrakasoasna brhmanan ti pigiyo brhmao bvarissa brhmaassa paddho
paddhacaro paricrako sisso. pigiyena te soas ti evam pi paricrakasoasna brhmana.
588
atha v, te soasa brhma buddhassa bhagavato paddh paddhacar paricrak siss
ti evam pi paricrakasoasna brhmana.
589
One could argue that a counterexample is provided by the Puaka-mava-pucch (3), in
which Puaka asks about sacrifice, and the Buddha explains that sacrifice does not lead to the far
shore. We have already encountered this dialog as one of the dialogs in which the word brhmaa

295
 The addition of what we can call the outer framethe story of Bvar

sending his 16 disciples to meet the Buddha, in whatever form, whether commentarial

or in the vatthu-gthchanges the role of the category Brahman completely. At

the very least, it decenters the Buddha as a Brahman teacher among Brahmans and

sets up a dichotomy between the Buddha, who converts the Brahmans, and the

Brahmans original Brahman teacher, Bvar. In the fully developed narrative form

of the vatthu-gth, the dichotomy between the Buddha as Brahman and Bvar as

a literal Brahman is made complete through the detailed description of Bvars

knowledge of the mantras (v. 976), practice of sacrifice (v. 978), and dwelling in an

rama (v. 979). Moreover, the vatthu-gth clearly imitates encounter dialogs by

incorporating what, as we will see in Part IV, is a key formulaic element: the

reference to the Buddhas possession of 32 marks that are supposedly predicted by

the Vedas as signs of a Great Man (mahpurisa), as well as the chief Brahman

interlocutors desire to discover whether the Buddha actually possesses all of them.

In addition, the vatthu-gth refers to the concept of head-splitting, which is found

elsewhere in Buddhist encounter dialogs, but also has parallels in the Brahmanical

literature. 590 Although there are limitations to the extent to which the vatthu-gth

can transform the Pryaa into a true encounter dialogit only serves as a frame,

appears, but used together with the word khattiya in apparent reference to a particular social group,
rather than the ideal person. There is no particular association between Brahmanhood and the practice
of sacrifice here, however; Brahmans are simply mentioned together with seers (isayo), human beings
(manuj) and khattiyas as among those who have sacrificed to gods in the past.
590
In the Ambaha Sutta (DN 3), which we will examine more closely in Part IV, the Buddha
declares that a Brahman student named Ambaha who has come to debate with him is actually
descended from the black baby of a slave girl. When the Buddha asks Ambaha to confirm that the
story is true, Vajirapi appears and threatens to split his head into pieces if he does not answer. On
the motif of head shattering in Brahmanical literature, see Black, Character of the Self, esp. 80-88.

296
and thus the core 16 dialogs are left intact and do not display the formulaic forms of

address that distinguish the interlocutor qua Brahman from the Buddha qua

samaait does serve quite effectively to place references to the ideal person,

including the Buddha as exemplar, in a context of comparison to a literal Brahman,

Bvar, and thus make the Buddhist use of the word brhmaa appear non-literal

and polemical.

III.2.4 Brhmaa as a Term Taken from the Pan-Indo-Aryan Substratum

As we have seen in this and the previous chapter, there is evidence that the use

of the word brhmaa to refer to the Buddhist ideal was not always metaphorical

that is, an explicit comparison to the counterexample provided by literal Brahmans,

i.e., the proponents of the new Brahmanism. The Pryaa-vagga and especially the

Ahaka-vagga of the Sutta Nipta, which are likely among the oldest Buddhist texts

still extant, make use of the term in such a way that it appears to have originally been

little more than an honorific, like muni or bhikkhu, deemed appropriate for referring

to the ideal person, however construed. But as we have seen, later expansions on

these textsincluding the vatthu-gth of the Pryaa, the Mah- and Ca-

niddesas, and Buddhaghosas commentary in the Paramatthajotik IIintroduced

comparisons to literal Brahmans, thus transforming straightforward uses of an

honorific term into ironic uses intended for polemic. In the case of the Niddesa, this

took the form of simply referring to literal Brahmans where they were not

mentioned in the root texts. More commonly, however, a dichotomy between the

297
ideal Brahman and literal Brahmans was introduced through frame narratives

stories that involved people who considered themselves Brahmans on the basis of

birththus setting up a contrast with the ideal Brahman constructed by the root text.

The word brhmaa is, needless to say, a very old word in the Indian tradition.

It is found several times in the g Veda, 591 and it is found quite frequently in the

Atharva Veda, 592 with increasing use in later Brahmanical literature. Now, even in

the oldest occurrences of the term that have come down to us, the Brahman is

associated with certain themes that in retrospect seem more Brahmanical than, say,

ramaicthe most important of which, of course, is sacrifice. As I argued in Part

II, however, we should not allow our understanding of ancient Indian religion to be

determined by a preconceived dichotomy between Brahmanical and non-

Brahmanical. Even if, once the Brahmanical and Buddhist traditions coalesced into

discrete identities around the turn of the era, the former incorporated the concept of

sacrifice into its identity more than the latter, 593 this does not mean that Brahmanism

and Buddhism were eternally discrete entities such that the former owned the

concept of the sacrifice and therefore the Brahman, while the latter did not. The work

591
The word brhmaa is found a total of 15 times in the g Veda and 6 times in the khilni:
RV 1.15.5, 1.164.45, 2.36.5, 6.75.10, 7.103.1, 7.103.7-8, 8.58.1, 10.16.6, 10.71.8-9, 10.88.19, 10.90.12,
10.97.22, 10.109.4; RVKhil. 1.4.10, 2.11.5, 3.10.1, 3.15.31, 3.18.2, 4.2.7.
592
The term brhmaa occurs over 100 times in the aunaka recension of the Atharva Veda;
for convenience, I list only the numbers of the hymns in which it occurs: AV 2.6, 4.6, 4.15, 4.19, 4.34,
5.17-19, 6.13, 6.38, 7.66-7, 9.1, 9.4-6, 9.10, 10.5, 10.7-8, 10.10, 11.1, 11.3, 11.5, 12.3-5, 13.3-4, 14.1,
17.1, 18.3, 19.6, 19.34-35, 19.59, 20.2, 20.67.
593
I avoid saying that Brahmanism accepted sacrifice while Buddhism rejected it because,
aside from the fact that the sacrifice that Brahmanism accepted varied widely across time and
space, the Buddhist tradition did not simply reject sacrifice tout courtsee Freiberger, Ideal
Sacrifice.

298
of Heesterman and Witzel, in particular, has shown that the practice of sacrifice

changed dramatically over the course of the Vedic period, from a loosely-organized

set of customs centering on (if we are to believe Heesterman) an agonistic sacrifice

involving the exchange of ritual duties to a highly codified ritual system formulated,

transmitted, and practiced by a professional priesthood. To be sure, the ideology of

the new Brahmanism that is evident in the encounter dialogs of the early

Buddhist texts and codified in the Dharma Stra literature arose out of this process,

but this does not mean that the new Brahmanical conception of Brahmanhood is

implicit in the Brahmans of the g Veda. The latter may indeed have been

involved in sacrifice, but the sacrifice they practiced was not the codified ritual of the

rauta Stras. Although, as their name shows, they were clearly associated with

sacred speech (brahman), these g Veda Brahmans did not memorize and transmit a

complex textual system known as the Vedas and their accompanying apparatus for

the simple reason that they had not been composed yet. Finally, there is no evidence

that the Brahmans of the g Veda could only be Brahmans because they were

physically born such; as Lubin has argued, they may have been called twice-born

because of a second birth that made them Brahmans. There were, in other words,

Brahmans before there was Brahmanism.

By the same token, the fact that early Buddhists did not see sacrifice as central

to the identity of the Brahman does not make their use of the term any less legitimate

on the one hand, or more innovative on the other, than that of the proponents of the

new Brahmanism. Early Brahmans were not associated only with sacrifice; they were

299
also associated with the taking of vows, 594 medical treatment, 595 and of course the

practice of debate (brahmodya), 596 all of which can be considered precursors of

ramaic practice. 597 Moreover, it would be unreasonable to judge the later use of a

word simply by the way in which it ignores aspects of the earlier use of the word, no

matter how important they might have been. One of the most central characteristics

of the Brahman as he is depicted in the g Veda and Atharva Veda is that he is a

drinker of soma 598; this is obviously not incorporated into early Buddhist

constructions of the Brahman, but it is also not central to the new Brahmanical

conception of the Brahman, if for no other reason than that no one any longer had

access to any soma to drink.

An instructive parallel example can be seen in the word muni, which, as we

have seen, is used in the Ahaka and Pryaa in much the same way as the word

brhmaa (interchangeably, in fact) as an honorific to refer to the ideal person. Muni

is, like brhmaa, a very old word in the Indian tradition, being found as far back as

the g Veda. The most descriptive text that refers to the muni there is RV 10.136, in

594
In the well known hymn RV 7.103, in which the chanting of Brahmans is compared to the
ribbiting of frogs, the Brahmans are described as practicing a vow (v.1: vratacria).
595
Treatment with plants is described in RV 10.97.22.
596
On the history of brahmodya, see Black, Character of the Self, 60-63.
597
Of course, as we have seen, vow-taking, medical practice, and debate have all been
rejected at one point or another in the Buddhist tradition, but this is beside the point, since all three are
characteristic of parivrjakas in ancient India, including, in many or even most cases, Buddhist
bhikus.
598
RV 6.75.10, 10.16.6, 10.71; AV 4.6, 9.1, 11.1, 14.1, 18.3, 19.59.

300
which munis girt by the wind, wearing reddish-brown, dirty (rags), 599 declare,

Excited with muni-hood, we mounted the wind, 600 and one of their number, named

Long-hair (kein), drinks a poison (via) with Rudra (v. 7). Elements of this

description can be taken as having inspired practices of various ascetic traditions,

including the long hair of the jailas, the nakedness of Jain and jvika ascetics (girt

by the wind), and the reddish rag-robes of the Buddhists. The use of the word muni

itself is also found in all of these traditions, although perhaps most prominently in the

Buddhist tradition, in which it endured, unlike the word brhmaa, in the later

tradition in the form of the common appellative for the historical Buddha, kyamuni.

In this case, I think it is obvious that muni was a very old term for a type of hermit

or ascetic, and that as such, it was picked up by various ascetic traditions that in turn

interpreted it in various ways. It was, in other words, part of the Indo-Aryan

substratum from which various traditions drew, rather than a uniquely Brahmanical

term that was then borrowed by non-Brahmanical groups for the purposes of

polemic. What I am arguing here is that the same is true of the term brhmaathat

it was an old term taken up and interpreted in various ways from the Indo-Aryan

substratum, rather than a term that was owned by a monolithic Brahmanical

tradition. What is different about the word brhmaa is that, over the course of time,

it became central to the identity of one particular groupthe proponents of the new

599
RV 10.136ab: munayo vtaraan piag vasate mal.
600
RV 10.136ab: unmadit mauneyena vt tasthim vayam.

301
Brahmanismso central, in fact, that as a result other groups, such as the Buddhists,

were eventually forced to abandon it.

302
 Chapter III.3

Tevijja/Tisso Vijj

In discerning the diachronic development of the way in which categories were

used by early Buddhists in articulating identity vis--vis the proponents of the new

Brahmanism, an instructive counterpoint to the word brhmaa is provided by the

term tevijja (=Skt. traividya) and the related phrase tisso vijj (three knowledges).

As I argued in the previous chapter, brhmaa is an extremely old term whose

simultaneous use by early Buddhists and the proponents of the new Brahmanism can

be explained by common borrowing from a pan-Indo-Aryan substratum, rather than

as a polemical borrowing by the Buddhists from the real Brahmans. In time, as the

proponents of the new Brahmanism became more assertive in arrogating the category

to themselves, Buddhists compared their own ideal qua Brahman favorably to the

Brahmans by birth, thus introducing an element of polemic to their own use of the

term that rendered it more figurative than it had been before. The terms tevijja and

tisso vijj are similar to the later use of the term brhmaa in the Buddhist literature,

insofar as they introduce a polemical contrast between Buddhist ideals and those of

303
the new Brahmanism. Tevijja is an adjective corresponding to the Sanskrit traividya

and is sometimes used in the Pali Canon, as in Sanskrit literature, as an honorific to

refer to Brahmans who are (ostensibly) learned in the three Vedas, but more

commonly in a polemical sense to refer to the Buddha or other arhat. Tisso vijj is a

noun used concomitantly to refer to the three knowledges attained by the Buddha at

his Awakening that allow him to be called tevijja: knowledge of recollection of

previous states of existence, knowledge of transmigration, knowledge of destruction

of the savas. 601 The contrast and its polemical intent are clear: Whereas the

Brahmans are tevijja because of their (from the Buddhist standpoint) mundane

knowledge of the Vedas, the Buddha is tevijja because of his possession of the

supramundane knowledges associated with Awakening. As I will argue in this

chapter, this polemical, contrastive usage of the term tevijja is the only one we find in

the early Buddhist texts; unlike the term brhmaa, we do not find evidence of an

earlier usage in which there was no intended comparison to the literal Brahmanical

usage.

In addition, I will argue that this term, as we would expect given our general

argument that the bifurcation between Buddhist and Brahmanical identities developed

only slowly over time, is not of early usage, but was introduced at a relatively late

stage of the tradition when polemicizing against the proponents of the new

Brahmanism had become a concern. In order to make this argument, I will be

examining the context in which the Buddhist conception of the three knowledges is

601
See, e.g., DN 33: tisso vijj pubbenivsnussatia vijj, sattna cutpaptea
vijj, savna khayea vijj.

304
foundnamely, discussions of the manner in which the Buddha attained his

Awakening. Therefore, I will begin by reviewing the scholarly literature on the

accounts of the Buddhas Awakening found in the early Buddhist literature and

theories of their chronological development. Then, I will discuss the issue of

precanonical Buddhism and the way in which early Buddhist accounts of

meditation and the attainment of Awakening can be understood, like the term

brhmaa, as arising out of a fluid, pre-sectarian substratum. Finally, I will argue

that the concept of tisso vijj was introduced at a relatively late stage in this process

in an effort to ground the polemical use of the word tevijja to describe the Buddha

and other Buddhist arhats in an actual meditative attainment of three knolwedges.

III.3.1 How Does One Attain Awakening?

It is of course well known that, with the development of Buddhist doctrine

over the centuries and the rise of the Mahyna and Tantric systems, conceptions of

Awakening and nirva, as well as methods of meditation, changed drastically, to the

point that many scholars have seen fit to speak about multiple Buddhisms rather

than a single, monolithic Buddhism. This diversity is not only found by comparing

the teachings of different traditions or different time periods in Buddhist history,

however; it is also found within the corpus of early Buddhist texts preserved in the

Nikyas/gamas. Indeed, if one turns to the early Buddhist stras with the question,

How does one attain Awakening? one quickly finds that the early tradition does not

305
answer this question with a single voice. De La Valle Poussin was one of the first to

notice this lack of unanimity in the early tradition, nearly 80 years ago:

We can, without negligence, discern in the Buddhist sources, whether ancient

or scholastic, two opposed theories, the same that the Bhagavadgt
distinguishes under the name of skhya and yoga: the theory that makes of
salvation a purely or above all intellectual labor; the theory that places
salvation at the end of the ascetic and ecstatic disciplines. 602

De La Valle Poussin named the article in which he articulated this observation after

two monks, Musla and Nrada, who appear in a Pali sutta (SN 2.12.7.68), and whom

he considers to be representative of the two opposed theories of meditationNrada

the path of praj and Musla the path of amatha. While this particular sutta comes

down on the side of amatha, de La Valle Poussin shows that various other texts,

both Pali and later, take various positions with respect to the two paths to Awakening

or in some cases struggle to reconcile them in various ways.

More recent scholarship has shown that this (as we will see, creative) tension

between intellectual and ascetic/ecstatic models of Awakening is paralleled in other

Indian religious traditions, and indeed appears to be reflective of a pan-Indian debate

over the nature of meditative experience. Much of this scholarship is summarized

and drawn upon by Stuart Ray Sarbacker in his comparative study of the Hindu Yoga

system and various Buddhist systematizations of meditative theory. 603 Sarbacker

602
La Valle Poussin, Musla et Nrada, 189-90: On peut, sans imprudence, discerner
dans les sources bouddhiques, anciennes ou scolastiques, deux thories opposes, celles mme que la
Bhagavadgt distingue sous le nom de skhya et de yoga: la thorie qui fait du salut une oeuvre
purement ou surtout intellectuelle; la thorie qui met le salut au bout des disciplines asctiques et
extatiques.
603
Stuart Ray Sarbacker, Samdhi: The Numinous and Cessative in Indo-Tibetan Yoga
(Albany, NY: SUNY Press, 2005).

306
argues that nearly all of Indian thought on meditation and yoga can be understood as

the playing out of a dynamic relationship between two dimensions of practice that

he terms numinous and cessative. 604 Since this binary and the dynamic

relationship between its two poles cuts across sectarian boundaries, Sarbacker argues

for a methodological approachone with which I am in broad agreementthat is

inherently comparative and understands Hinduism, Buddhism, and Jainism not as

isolated entities, but as existing in continuity within the greater scope of Indian

culture. 605 Although a full examination of this continuity is beyond the scope of this

work, we will see that a healthy cognizance of it fruitfully informs our understanding

of the Buddhist evidence that we do examine in this section.

How, then, does the debate between numinous and cessative models of

Awakening play out in the Buddhist tradition? In the Kosamb Sutta in which Musla

and Nrada appear, a certain monk named Saviha interrogates them and finds that

although they have both attained the same insights into dependent origination, only

Musla claims to be an arhat. The other monk, Nrada, compares his own situation to

a thirsty man who comes upon a well in the desert: He sees the water in the well, but

without the right equipment, he does not come into bodily contact with it. The

implication is that although he has seen the truth of nibbna, he has not personally

experienced itor, in other words, that insight alone is not enough.

604
Ibid., 1.
605
Ibid., 136.

307
 Gombrich has noted that the message of the Kosamb Sutta is directly

contradicted by another sutta that appears shortly after it, the Susma Sutta (SN

2.12.7.70). In this sutta, a monk named Susma enters the order and asks a group of

monks about their meditative attainments. He finds that they have nonethey have

not attained any supranormal powers, cannot recall their past lives, and have not

attained any of the arpa-jhnas. Nevertheless, they claim that they are liberated by

wisdom (pavimutt). This confuses Susma, so he goes to the Buddha, who

confirms that what the other monks have saidthat they have attained liberation

through wisdom alone, without meditative attainmentis indeed possible. We thus

have here a message in direct contradiction to that of the Kosamb Sutta: While the

Kosamb Sutta makes it clear that insight is not enough to attain liberation, that one

must experience it directly, the Susma Sutta makes the opposite claim, that an

intellectual understanding of the dhamma is itself constitutive of nibbna.

As Gombrich shows, however, the Pali version of the Susma Sutta cannot be

original. At the end of the sutta, Susma confesses to being a spy and is reprimanded

by the Buddha, but this turn of events is inexplicable in light of the greater context of

the sutta. In the Chinese version (T. 99, M 347), however, it is explained from the

beginning that Susma was a member of another sect who entered the Buddhist order

surreptitiously to learn the Buddhas secrets. He starts asking around about the

Buddhist dhamma, and a group of monks explain it to him and then claim, apparently

in jest, that they are already enlightened. For in this version, Susma interrogates

them further and realizes that they have not even eliminated greed and hatred and thus

308
must be lying about having attained liberation. The Buddha then takes interest in

Susma and begins to teach him the dhamma, which leads to his confession that he is

a spy. Gombrich argues convincingly that the Chinese preserves a more original

version, which was then redacted to make the claim, quite contrary to the original text,

that insight alone is sufficient to attain Awakening. 606 He argues further that the

changes made to the Susma Sutta that resulted in the Pali version were part of a

larger process whereby liberation by insight alone was delineated as a distinct

possibility. In particular, the phrase pavimutta (liberated by insight) used by the

group of monks in the Susma Sutta was, Gombrich argues, originally simply one

descriptor among others for liberation, but came to be understood, in the Puggala-

paatti of the Abhidhamma Piaka, as the name of a distinct kind of liberation. 607

Like Gombrich, Lambert Schmidthausen has also cast doubt on the antiquity

of the insight-alone path to liberation. In an important study, Schmidthausen

investigates the development of conceptions of liberating insight, working on the

basis of the assumption that the texts I make use of are to be taken seriously and

when there are instances of incoherence, they will have to be taken seriously and

will need to be explained (e.g. by reference to textual history). 608 He traces

conceptions of liberating insight to the latter part of the Buddhas discourse in the

606
Gombrich, How Buddhism Began, 123-7.
607
Ibid., 97-8.
608
Lambert Schmidthausen, On Some Aspects of Descriptions of Theories of Liberating
Insight and Enlightenment in Early Buddhism, Studien zum Jainismus und Buddhismus,
Gedenkschrift fr L. Alsdorf, ed. Klaus Bruhn and Albrecht Wezler (Wiesbaden: Franz Steiner, 1981),
200.

309
Dhammacakkappavattana Sutta, and certain parallel versions, 609 in which the Buddha

describes his liberation by enumerating the four noble truths, the need for their

realization, and their actual realization, and concludes by asserting, Since, monks,

knowledge and vision with respect to these four noble truths thus as they are with

their three rounds and twelve aspects have been well purified by me, I have

claimed to be fully awakened to the unexcelled full Awakening. 610 Schidthausen

concludes, It is not likely that this rather sophisticated and schematic account of the

Enlightenment of the Buddha is the original one. 611 And indeed, while I am not

certain that the idea of an original account of the Buddhas Awakening is a

meaningful concept, I think we can agree that this schematic treatment of the

Buddhas Awakening does not serve so much to provide a meaningful account of the

actual mechanics of how the Buddha attained Awakening as it does to emphasize, in a

rhetorical fashion appropriate for oral transmission, the importance of the four noble

truths to the Buddhas Awakening. As Buddhist philosophy developed further,

Schmidthausen argues, comprehension of the four noble truths came to be supplanted

609
See Schmidthausen, Some Aspects, 202, for references to studies of the the
compositional history of the Dhammacakkappavattana Sutta and its parallels. Schmidthausen notes
that the sutta appears to be a juxtaposition of at least three independent portions which are in fact
presented as three different discourses in the version of at least one school. He also notes (p. 202 n.
11) that, interestingly, the portion of the Dhammacakkappavattana Sutta in which the Buddha goes
through the 12 permutations of the four noble truths, the need to realize them, and the actual realization
of them is even preserved in the Pali tradition as an independent sutta at SN 5.12.2.11, immediately
after the complete Dhammacakkappavattana Sutta at SN 5.12.2.12.
610
SN 5.12.2.12: yato ca kho me, bhikkhave, imesu catsu ariyasaccesu eva tiparivaa
dvdaskra yathbhta adassana suvisuddha ahosi, anuttara sammsambodhi
abhisambuddho ti paccasi.
611
Schmidthausen, Some Aspects, 203.

310
by comprehension of dependent origination, which in turn was supplanted by

comprehension of not-self. 612 Thus, Buddhist conceptions of liberating insight

undertook a pattern of development that mirrors the development of Buddhist

philosophy.

Conceptions of liberating insight did not develop in a vacuum, however;

they oftenthough, as we shall see, not alwayswere tied up with descriptions of

meditative techniques that lead to liberation. Altogether, Schmidthausen has

delineated three basic templates for how Awakening comes about that are found in

the early Buddhist texts; two of these involve meditation, but are incompatible with

one another in the details, and the third involves only insight. 613 Since the history of

the introduction of the concept of tisso vijj is implicated in the historical relationship

between these three templates, I will describe them in some detail. The first of the

templates, 614 which I will refer to as four jhnas plus insight, involves the

successive cultivation of the four meditative absorptions, referred to in later

systematizations as rpa-jhnas, in which the factors (aga) of reflection and

examination (vitakka and vicra), delight (pti), happiness (sukha), one-pointedness

(ekaggat), and equanimity (upekkh) are cultivated and then successively abandoned

until only one-pointedness and equanimity remain. Once one attains the fourth jhna,

612
Ibid., 211-2. Not-self, I would add, was supplanted by emptiness (nyat) in the
Mahyna as the seminal truth to be comprehended.
613
For a concise summary of Schmidthausens arguments regarding these three templates, see
also Tilmann Vetter, The Ideas and Meditative Practices of Early Buddhism (Leiden: E.J. Brill, 1988),
xxiff.
614
For Schmidthausens discussion of this template, see Some Aspects, 203ff.

311
in which only these two factors remain, one attains insight and thus is liberated. This

liberating insight is described in one of two ways. In some cases, it is described in

terms of the tisso vijjrecollection of past lives, knowledge of the arising and

passing away of beings, and destruction of the savas. 615 In other cases, it is

described in terms of a single truthrealization of the four noble truths. 616 As we

shall see, these two forms are not unrelated, since the description of the destruction of

the savas is modeled upon the realization of the four noble truths.

The second template, 617 which I will refer to as eight jhnas plus cessation,

begins, as the previous template does, with the four (later so-called) rpa-jhnas, but

after the attainment of the fourth jhna, there is no discussion of insight; instead, the

fourth jhna is abandoned and one continues with a succession of four additional

meditative absorptions that later systematizers refer to as the arpa-jhnas. 618 These

formless absorptions are as follows:

615
Within the four main Nikyas, we find this 4 jhnas plus 3 knowledges template in the
following places: DN 2-13 (these all include the template as part of a larger teaching on the training
that they duplicate in near identical words), 28; MN 4, 19, 27, 36, 39, 51, 60, 65, 76, 79, 85, 94, 100,
101, 125; AN 3.59, 3.60, 4.198, 6.64, 8.11. It is also found once in the Vinaya, in several late texts of
the Khuddaka Nikya, and with great frequency in the Abhidhamma Piaka. Interestingly, as
Schmidthausen has noted (Some Aspects, 204 n. 15), it is found not even once in the Sayutta
Nikya.
616
The only complete version of the four jhnas plus insight template with only one truth
found in the Pali Canon is MN 112, but incomplete versions are found at AN 5.75-6. The Chinese
translation of the Madhyamgama (T. 26), however, prefers the version with only one truth (M 19,
102, 146, 182, 187, 203, 204), with the version with three truths found only at M 157 and 194.
Citations found in Schmidthausen, Some Aspects, 221 n. 75.
617
For Schmidthausens discussion, see Some Aspects, 214-9.
618
Within the four main Nikyas, this template is found in the following suttas: DN 15, 16, 33,
34; MN 25, 26, 30, 31, 43, 44, 50, 59, 66, 77, 111, 113, 137; SN 1.6.2.5, 2.3.2.1, 2.5.9, 3.7.9, 4.2.2.1,
4.2.2.5, 4.2.2.7, 4.2.2.9, 4.2.2.10, 4.7.6, 5.4.4.10, 4.10.1.8; AN 1.382-493-562, 5.166, 8.66, 8.117-147-

312
 1. sphere of infinite space (ksnacyatana)
2. sphere of infinite consciousness (vinacyatana)
3. sphere of nothingness (kicayatana)
4. sphere of neither-apperception-nor-non-apperception
(nevasansayatana)

After attaining and then surpassing each of these absorptions in turn, one attains a

final state, known as the cessation of apperception and feeling

(savedayitanirodha). In some cases, the following remark is made in

conclusion to the sequence to indicate that liberation is attained: And having seen

with wisdom, his savas are exhausted. 619 Nothing is said about the content of this

wisdom, however, and in general it appears that the attainment of the cessation of

apperception and feeling is itself constitutive of liberation.

Finally, the third template, which I will call insight alone, explains

liberation solely in terms of gaining some sort of insight, often into the three seals

(anicca, dukkha, anatt). 620 Many of the passages that follow this template seem to

indicate that meditation, or at least attainment of jhnas, is not required, and some

even appear to preclude the practice of meditation. In certain extreme cases,

liberation is attained simply by listening to the Buddha preach the dhamma and

626, 9.31-36, 9.38-47, 9.51-2, 9.61, 9.93-113-432, 10.72, 10.85, 10.99. It is also found in certain late
texts of the Khuddaka Nikya, and it is found quite often in the Abhidhamma Piaka.
619
MN 25, 26, 30, 70, 111, 113; AN 4.186, 9.34, 9.38-47, 9.51-2, 9.61, 10.99: paya cassa
disv sav parikkh honti.
620
For Schmidthausens discussion, see Some Aspects, 219-22. Citations are found at
p.219-20 n. 69.

313
thereby gaining insight into fundamental truths; Schmidthausen sees this narrative

trope as particularly characteristic of the Vinaya. 621

There are two basic reasons for believing that within this array of templates

for describing liberation, the role of liberating insight is a relatively late

development. The first is that there exists an account of the Buddhas Awakening,

which may be the oldest extant, in which no mention is made of liberating insightor

any other particular method for attaining liberation for that matter. This account is

found in the Ariyapariyesana Sutta (MN 26), 622 in which the Buddha recounts having

studied under two teachersra Klma and Uddaka Rmaputtawhose teachings

he was dissatisfied with because, he said, they led only to the sphere of nothingness

(kicayatana) and the sphere of neither-apperception-nor-non-apperception

(nevasansayatana), respectively. After leaving them, he went to

Sennigama near Uruvel in Magadha and there attained Awakening, which he

describes in the following terms:

I, bhikkhus, being myself subject to birth, realized the misery in what is

subject to birth, and seeking nibbna, which is unborn, unexcelled, secure
from bondage, I attained nibbna, which is unborn, unexcelled, secure from
bondage. Being myself subject to old age, I realized the misery in what is
subject to old age, and seeking nibbna, which does not grow old and is
unexcelled and secure from bondage, I attained nibbna, which does not grow
old and is unexcelled and secure from bondage. Being myself subject to
621
Ibid., 221.
622
It should be noted that in the text of the Ariyapariyesana Sutta as it comes down to us,
both in the Pali version and in a Chinese parallel (T. 26, M 204), the Buddhas account of his own
quest for Awakening is followed by a teaching to the monks he is addressing in which he does discuss
meditative techniques for attaining Awakening. Scholars who have argued for the antiquity of the
biographical story in this text have treated it separately from the teaching, which is incidental to the
contents of the story. It is also worth mentioning that the teachings about meditative techniques found
in the Pali and Chinese versions do not agree with one anotherin the former, they follow the eight
jhnas plus cessation template, while in the latter, they follow the four jhnas plus insight template.

314
 sickness, I realized the misery in what is subject to sickness, and seeking
nibbna, which does not get sick and is unexcelled and secure from bondage,
I attained nibbna, which does not get sick and is unexcelled and secure from
bondage. Being myself subject to death, I realized the misery in what is
subject to death, and seeking nibbna, which does not die and is unexcelled
and secure from bondage, I attained nibbna, which does not die and is
unexcelled and secure from bondage. Being myself subject to grief, I realized
the misery in what is subject to grief, and seeking nibbna, which is without
grief, unexcelled, and secure from bondage, I attained nibbna, which is
without grief, unexcelled, and secure from bondage. Being myself subject to
defilement, I realized the misery in what is subject to defilement, and seeking
nibbna, which is undefiled, unexcelled, and secure from bondage, I attained
nibbna, which is undefiled, unexcelled, and secure from bondage. Then
knowledge and vision arose in me: Unshakeable is my release; this is my
last birth; there is now no more coming into being. 623

Many scholars have noted the simplicity of this description of the Buddhas

Awakening: Although it describes nibbna at length, it says nothing about what

happened when the Buddha attained it other than that he attained it. The only trace of

any sort of insight in this accomplishment is the mere realization that this will be

his last birth.

Various scholars have argued for the antiquity of this account of the Buddhas

Awakening in different ways, though all have focused at least in part on the apparent

simplicity of the account. Alexander Wynne has argued that not only this particular

623
MN 26: so kho aha, bhikkhave, attan jtidhammo samno jtidhamme dnava viditv
ajta anuttara yogakkhema nibbna pariyesamno ajta anuttara yogakkhema nibbna
ajjhagama, attan jardhammo samno jardhamme dnava viditv ajara anuttara
yogakkhema nibbna pariyesamno ajara anuttara yogakkhema nibbna ajjhagama,
attan bydhidhammo samno bydhidhamme dnava viditv abydhi anuttara yogakkhema
nibbna pariyesamno abydhi anuttara yogakkhema nibbna ajjhagama, attan
maraadhammo samno maraadhamme dnava viditv amata anuttara yogakkhema
nibbna ajjhagama, attan sokadhammo samno sokadhamme dnava viditv asoka
anuttara yogakkhema nibbna ajjhagama, attan sakilesadhammo samno sakilesadhamme
dnava viditv asakiliha anuttara yogakkhema nibbna pariyesamno asakiliha
anuttara yogakkhema nibbna ajjhagama. a ca pana me dassana udapdi akupp me
vimutti, ayam antim jti, natthi dni punabbhavo ti.

315
pericope, but the entire story of the Buddhas training under ra Klma and

Uddaka Rmaputta, is extremely old. In addition to the simplicity of the account of

the Buddhas Awakening itself, Wynne points to certain unusual linguistic forms in

the text that could be Magadhisms; the fact that the Ariyapariyesana Sutta, alone out

of texts that refer to the Buddhas training under Klma and Rmaputta, seems to be

completely unaware of his performing austerities afterwards; the inclusion of the

potentially embarrassing episode in which Upaka the jvaka fails to take the Buddha

seriously when he claims to be Awakened; and the fact that the account of the

Buddhas training under Klma and Rmaputta does not seem to fit either a

polemical or an inclusivistic purpose. 624 Andr Bareau, on the other hand, does not

accept the authenticity of the story of the Buddha training under Klma and

Rmaputta, but nevertheless believes that the pericope in which the Buddha describes

his Awakening is the oldest preserved in the Canon because it is preserved in almost

exactly the same form in both the Pali and the Chinese version and because, given

their simplicity, it is quite difficult to consider their brief account as a summary of

the long narrative given by the other sources. 625 On the contrary, the

Ariyapariyesana Sutta pericope appears to have influenced later accounts; as

Schmidthausen notes, awareness of being liberated seems to have been regarded as

624
Wynne, Origin of Buddhism Meditation, 16-26.
625
Andr Bareau, Recherches sur la biographie du Buddha dans les Strapiaka et les
Vinayapiaka anciens: de la qute de lveil la conversion de riputra et de Maudgalyyana,
Publications de lEcole Franaise dExtrme-Orient, vol. 53 (Paris: Ecole Franaise dExtrme-Orient,
1963), 72-74. The quote is on p. 74: il est bien difficile de considrer leur bref rcit comme un
rsum de la longe narration donne par les autres sources.

316
an essential element, being found in all versions of the four jhnas plus insight

template, as well as the Dhammacakkappavattana Sutta. 626

The second reason for believing that insight is a late element within the

templates for liberation delineated by Schmidthausen is linguistic difficulties that

point to the late addition of elements pertaining to insight. Schmidthausen notes that

what I am calling the insight alone template includes the sentence, When it was

liberated, the knowledge arose, It is liberated (vimuttasmi vimuttam iti a

ahosi), which is also found in identical words in the four jhnas plus insight

template. Schmidthausen argues that the borrowing must have been by the former

from the latter, since vimuttam is in the neuter and has no appropriate neuter

antecedent in the insight alone template, but it doesnamely the word citta,

which is found slightly before itin the context of the four jhnas plus insight

template. 627 Thus, it appears that the insight alone template is indeed later than the

four jhnas plus insight template.

Within the four jhnas plus insight template itself there also appears to be

signs of development. As already mentioned, there are two forms of this template

one of which culminates with the attainment of three knowledges (the tisso vijj with

which we are concerned in this chapter) and the other of which culminates in only

one (realization of the four noble truths). The third of the tisso vijj is actually

knowledge of destruction of the savas (savna khayea), but in the four

626
Schmidthausen, Some Aspects, 207.
627
Ibid., 219 n. 69.

317
jhnas plus insight template, this is coupled with realization of the four noble truths

and then itself expressed in a form clearly modeled on the realization of the truths:

I knew as it is, These are the savas. I knew as it is, This is the origin of
the savas. I knew as it is, This is the cessation of the savas. I knew as it
is, This is the path leading to the cessation of the savas. 628

One could interpret from this that originally the insight of the four jhnas plus

insight template was expressed in terms of realization of the four noble truths, much

as we find in the Dhammacakkappavattana Sutta, and that later the concept of the

three knowledges was added by essentially grafting the third of these onto the form

provided by the four noble truths. This is confirmed, argues Schmidthausen, by a

linguistic peculiarity found in the version with three knowledges. Whereas for the

most part the description of the attainment of liberation uses the aorist tense,

including in the description of the destruction of the savas, the descriptions of the

first two knowledges stand out because they are in the present tense, which,

Schmidthausen convincingly argues, indicates that they were inserted into an earlier

description that included only one truth. 629

What, then, of the eight jhnas plus cessation template? Scholarship has

been nearly unanimous in regarding this as late. Many, though not all, have regarded

it as such in part because they believe, following Bareau, who was one of the first to

articulate the argument, that the account of the Buddhas training under ra Klma

628
See, e.g., MN 4: ime sav ti yathbhta abbhasi, aya savasamudayo ti
yathbhta abbhasi, aya savanirodho ti yathbhta abbhasi, aya
savanirodhagmin paipad ti yathbhta abbhasi.
629
Schmidthausen, Some Aspects, 221-2 n. 75.

318
and Uddaka Rmaputta, through which he learned about the two highest arpa-

jhnas, is a fabrication. Bareaus argument against the authenticity of the episode is

based primarily on the fact that it is not found within the Mahsaka Vinaya, even

though the latter is familiar with the two characters Klma and Rmaputtanamely,

as the two people whom the Buddha first thought of to teach to after he attained his

Awakening. Since this latter episode makes no reference to who Klma and

Rmaputta are, and the available accounts of their teaching the Buddha before his

Awakening describe his tutelage under them in nearly identical terms, Bareau argued

that the story of their teaching the Bodhisattva was invented to explain the cryptic

reference to them as people he could possibly teach after he attained Awakening. 630

Bareaus argument has been accepted by Bronkhorst 631 and Vetter, 632 both of whom

go on to reject the four so-called arpa-jhnas, of which the higher two correspond to

the states supposedly taught by Klma and Rmaputta, as later borrowings from a

non-Buddhist source.

Others, however, have questioned Bareaus dismissal of the authenticity of the

story of the Buddhas earlier training under ra Klma and Uddaka Rmaputta.

Ghiorgo Zafiropulo, for example, has systematically refuted the reasons offered by

Bareau for considering the episode inauthentic. In particular, he cites Frauwallners

observation that the Mahsaka Vinaya seems to represent the most corrupt Vinaya

630
Bareau, Recherches sur la biographie du Buddha, 16-22.
631
Johannes Bronkhorst, The Two Traditions of Meditation in Early India (Delhi: Motilal
Banarsidass, 1993), 85-6.
632
Vetter, Ideas and Meditative Practices of Early Buddhism, xxii.

319
tradition, with obvious lacunae that, Zafiropulo argues, could explain the absence of

the episode in question; likewise, while Bareau sees the similarity between the way

the Bodhisattvas tutelage under Klma and Rmaputta is described as a sign of its

artificial construction, Zafiropulo sees it as a trademark of oral tradition and thus a

sign of its antiquity. 633 Wynne also cites Frauwallner on the unreliability of the

Mahsaka tradition, and he also notes that, far from being unknown outside of the

story in question, Klma and Rmaputta are referred or alluded to several times

elsewhere in the Canon, such as in the case, already mentioned in Chapter II.3, in

which the Buddha encountered his former sabrahmacr Bharau Klma. 634 In

addition, on the issue of the stock phrasing used to described the Buddhas tutelage

under Klma and Rmaputta, Wynne notes that, contrary to what Bareau implies, the

two parts of the story are not identical: Whereas ra Klma himself claims to

have attained the highest goal and therefore offers to share leadership with the

Bodhisattva when he attains it himself, Uddaka Rmaputta is, as his name implies, a

disciple of a (presumably now deceased) teacher named Rma who claimed to have

attained the highest goal, and so when the Bodhisattva attains it himself, he offers to

hand over leadership to him completely. 635 For these reasons, as well as the ones

already discussed above regarding linguistic and narrative peculiarities of the text,

633
Ghiorgo Zafiropulo, LIllumination du Buddha: De la Qute a lAnnonce de lEveil
(Innsbruck, 1993), 22-29.
634
Wynne, Origin of Buddhist Meditation, 10-14.
635
Ibid., 14-16.

320
Wynne argues that the entire episode records actual historical events within the life of

the Buddha.

Irrespective of their views regarding the authenticity of the story about ra

Klma and Uddaka Rmaputta, however, nearly all scholars have agreed that the

formless spheres are either a late addition to Buddhist teachings about meditation or

at the very least somehow extra-Buddhist. Bronkhorst, whose conclusions are

followed by Vetter, 636 has argued that the formless spheres, unlike the four ordinary

(i.e., rpa-) jhnas, are not found in the earliest Abhidharma lists, and that therefore

they are a relatively late addition to Buddhist teachings on meditation. 637 A similar

argument has been made by Zafiropulo, who through an exhaustive examination of

references to the four rpa-jhnas alone and in conjunction with the four arpa-

jhnas shows that the group of those mentioning the four dhyna/jhna is not only

the most comprehensive group, the most extensive, but also that which cites the most

ancient texts or fragments of texts and therefore concludes that the arpa-jhnas are

later elaborations, in part probably determined by non- and pre-Buddhist influences,

and in any case relatively late. 638

636
Vetter, Ideas and Meditative Practices of Early Buddhism, xxii.
637
Johannes Bronkhorst, Dharma and Abhidharma, Bulletin of the School of Oriental and
African Studies 48, no. 2 (1985): 305-20.
638
Zafiropulo, LIllumination du Buddha, 32-70: le groupe de ceux mentionnant les
quatres dhyna/jhna est non seulement le groupe le plus comprhensif, le plus tendu, mais aussi
celui qui cite les textes ou fragments de textes les plus anciens (56); dlaborations ultrieurs,
en partie probablement dtermines par des influences non- et pr-bouddhiques, et en tous les cas
relativement tardives (57).

321
 In an earlier work, Bronkhorst also argued that the arpa-jhnas were late

additions to Buddhist meditation on the basis of the methodological principle that

Buddhist teachings that are similar to those found among non-Buddhist groups with

whom the Buddhists came into contact, and which are elsewhere in the early Buddhist

texts explicitly criticized, are likey to be late accretions. 639 Bronkhorst argues that the

arpa-jhnas fit these criteria because (1) he takes the story about the Buddha

training under ra Klma and Uddaka Rmaputta as a polemic against them and (2)

they fit in well with what what Bronkhorst identifies as the goal of main stream

meditation (i.e., that embraced by non-Buddhist groups such as Jains and the

jvikas)namely, the stopping of all activity, in this case, mental activity. 640 He

suggests, therefore, that non-authentic views and practices such as the arpa-jhnas

could have found their way into the early Buddhist texts through borrowing or

intrusion from other religious groups and ideals current at the time, 641 a

sentiment that is shared by Zafiropulo, who argues that they were introduced into

Buddhism by monks who had converted from other, non-Buddhist sects. 642

Given his strong support for the authenticity of the story of the Bodhisattva

training under ra Klma and Uddaka Rmaputta, Wynne has taken the four

formless absorptions perhaps most seriously as an old element of Buddhist teaching,

639
Bronkorst, Two Traditions of Meditation, vii-xiv.
640
Ibid., 78-93.
641
Ibid., ix.
642
Zafiropulo, LIllumination du Buddha, 57.

322
although for the same reason he too argues that they came from outside of Buddhism,

or as he in particular argues, from Brahmanism. He argues that the formless spheres

seem to have roots in the Brahmanical tradition, in two respects. First, the lower two

spheres (space and consciousness) are connected in the Buddhist tradition of kasia-

meditation to the four elements (earth, water, fire, and wind) and a certain sequence

of colors (indigo [nla], yellow, red, white). 643 Both the implicit sequence of six

elements (the ordinary four plus space and consciousness) and the sequence of four

colors, Wynne shows, have parallels in the Brahmanical tradition. The specific

sequence of indigo, yellow, red, white is found in Brahmanical sources (for example,

in reverse order, in association with the four yugas) and appears to represent a

progression from gross (indigo) to subtle (white). 644 More importantly, though,

certain Brahmanical cosmogonies begin with tman or brahman and then progress

through a sequence of five or six elements similar to the six found among the ten

Buddhist kasias. 645 Wynne also cites a portion of the Mokadharma Parvan in the

Mahbhrata (12:247-8) that teaches a form of meditation in which one takes a

sequence of elements as objects of meditation; although this sequence does not

correspond exactly to the Buddhist sequence, it does appear to follow an intended

643
Wynne, Origin of Buddhist Meditation, 31. See the following cited by Wynne: DN 33, 34;
MN 77; AN 10.25
644
Wynne, Origin of Buddhist Meditation, 35.
645
Ibid., 36. The closest parallel Wynne cites is MBh. 12.224, which has the following
sequence: brahman, manas, ka, vyus, jyotis, pas, bhmi).

323
progression from gross to subtle. 646 Wynne concludes, It seems that element

meditation in early Brahmanism, like element meditation in early Buddhism, was

based on Brahminical cosmogonies which were thought to provide meditative maps

of the path to liberation. Element meditation, so we must understand, was thought to

be the yogins way of reversing the creation of the cosmos and attaining

liberation. 647

This conclusion is supported by the second way in which the formless spheres

appear to have roots in the Brahmanical tradition, which pertains to the higher two

spheres (nothingness and neither-apperception-nor-non-apperception). Wynne argues

that in the Upaniads, the tman is conceptualized in similar terms. On the one hand,

the Mkya Upaniad describes the fourth quarter of brahman as not one with

awareness within (nntapraja), not one with awareness without (na

bahipraja), not one with awareness of both, not a mass of awareness

(prajnaghanam), not awareness nor non-awareness (na praja nprajam). 648

Likewise, in the Bhadrayaka Upaniad, Yjavalkya states that the tman is

neither conscious (saj) nor unconscious (moha) after death. 649 These descriptions

are reminiscent of Uddaka Rmaputtas goal, nevasansa. On the other hand,

Wynne sees parallels between ra Klmas goal of nothingness (kicaa) and

646
Wynne, Origins of Buddhist Meditation, 37-9. Here the progression is as follows: earth,
wind, space, water, fire, ahakra, buddhi, avyakta.
647
Ibid., 39.
648
MU 7; trans. by Wynne, Origins of Buddhist Meditation, 42.
649
BU 2.4.12; cited by Wynne, Origins of Buddhist Meditation, 43.

324
Uddlaka ruis likening of the nothing (na kicana) inside a banyan seed to the

tman (CU 6.12.1), as well Upaniadic cosmogonies that state that the world began as

non-existent (TU 2.7.1; CU 3.19.1). 650 Wynne therefore concludes that Uddlaka

Rmaputta and ra Klma belonged to a milieu that was probably a Brahminic

one in which the philosophical formulations of the early Upaniads were

accepted. 651

In tracing a Brahmanical origin for the formless spheres, Wynne does not

chart a course all that different from other scholars, such as Bronkhorst, who also

attribute them to an extra-Buddhist source; however, Wynne argues forcefully that the

formless spheres are not only extra-Buddhist, but pre-Buddhist. This is based in large

part upon his argument for the authenticity of the account of the Buddhas

studentship under Klma and Rmaputta, but it is also supported by what he sees as

references to formless-sphere-esque meditation in the Pryaa-vagga of the Sutta

Nipta, which, like myself and many other scholars, he regards as particularly old. 652

He presents a convincing argument that the discourse with Upasva (6) concerns a

question about meditating on nothingness (v. 1071: kicaa) in which the

Buddha surprises Upasva by saying that one should maintain that meditative state

650
Wynne, Origins of Buddhist Meditation, 44-5.
651
Ibid., 44.
652
Ibid., 73-5.

325
mindfully (v. 1070: satim). This prompts Upasva to become confused and ask if

one can maintain such a meditative state while also being mindful. 653

Wynne also argues that another discourse, that with Posla (14), also concerns

a meditative state focused on nothingness; Posla describes this state (v. 1113) and

asks if a person who has attained it can be led further. The Buddha answers with a

verse that we have already seen earlier in Chapter III.2, since it uses the word

brhmaa to refer to the ideal person:

Having known the origin of nothingness, [he thinks,] Enjoyment is a fetter.

Understanding this thus, therefore he sees clearly with respect to that.
This is the true knowledge of that perfected Brahman. 654

Wynne points out that the key here is not simply that one attain to a particular

meditative state, but that one understand itin terms of the distinction we have been

discussing here, that one have an insight experience. 655

Wynne therefore argues that this evidence from the Pryaa supports his

hypothesis that the Buddha had studied Brahmanical meditation techniques and

thus was conversant enough with them to discuss them with the Brahmans of the

Pryaa, but that he differed from his teachers and Brahmanical meditation

technique in general in regarding a liberating insight as necessary, in addition to the

mere attainment of meditative states. Given this understanding of the Buddha

actually taught, Wynne regards the first of Schmidthausens three template (four

653
Ibid., 75-100.
654
Sn. 5.1115: kicaasambhava atv, nand sayojana iti / evam eta abhiya,
tato tattha vipassati / eta a tatha tassa, brhmaassa vusmato //.
655
Wynne, Origins of Buddhist Meditation, 103-6.

326
jhnas plus insight) as in accordance with the teaching of the Buddha 656; the

second (eight jhnas plus cessation) as not surprisingsince the Buddhas

dialogues with Upasva and Posla seem to have allowed the cultivation of formless

meditations such as nothingness 657; and the third (insight alone) as certainly at

odds withthe sort of meditation taught by the Buddha, since it disposes of

meditation altogether. 658

III.3.2 Pre-canonical Buddhism and Substratal Influences on Buddhism

Wynnes reconstruction of a pre-Buddhist system of meditation, preserved in

what I am calling the eight jhnas plus cessation template, which is based on a

system of six progressively more subtle elements that one seeks to ascend in order to

reach some sort of absolute, bears a close resemblance to a fairly old idea in Buddhist

Studies known as pre-canonical Buddhism. This idea was promoted in the 1930s

by Stanislaw Schayer 659 and A. Berriedale Keith, 660 and despite an attempt by the

formers student Constantin Rgamey to revive it, 661 the concept never really caught

656
Ibid., 123.
657
Ibid., 119.
658
Ibid., 120.
659
Stanislaw Schayer, Precanonical Buddhism, Archiv Orientln 7 (1935): 121-32;
Stanislaw Schayer, New Contributions to the Problem of Pre-Hnaynistic Buddhism, Polish
Bulletin of Oriental Studies 1(1937): 8-17.
660
A. Berriedale Keith, Pre-Canonical Buddhism, The Indian Historical Quarterly 12
(1936): 1-20.
661
Constantin Rgamey, Le problme du bouddhisme primitif et les derniers travaux de
Stanislaw Schayer, Rocznik Orientalistyczny 21 (1957): 37-58.

327
on and for the most part died with Schayer and Keith during World War II. The idea,

as originally formulated by Schayer, was simple: that the basic concepts that appear

to be definitive of the earliest Buddhist scripturesnamely, the four noble truths, not-

self, impermanence, the five aggregates, dependent origination, etc.do not represent

the earliest Buddhist teaching, but represent a reaction against that earliest teaching,

which latter Schayer dubbed pre-canonical Buddhism. Schayer argued that this

pre-canonical Buddhism was different from canonical Buddhism in two major

respects. First, it postulated a series of six elements (dhtus) that progressed from

gross (earth) to subtle (consciousness). 662 The goal was to ascend these six elements

and then beyond them to nirva, which was conceived of as a place.

Second, Schayer argued that in pre-canonical Buddhism there was a sharp

dichotomy between rpa and dharma. The former corresponded to the lower four

elements; the latter, to space, consciousness, and nirva. Schayer argued that in the

beginning, anityat (impermanence) referred only to rpa and not to dharma. Thus,

the goal was to ascend from what is impermanent and thus causes suffering to what is

permanent and therefore does not. This system was broken and canonical Buddhism

born when the concept of impermanence came to be applied to all things, both

material and immaterial. This corresponded to the Abhidharmic extension of the

word dharma to refer to all phenomena, both material and immaterial, as well as

the development of the concepts of antman and dependent origination, which are

662
Schayer, Precanonical Buddhism, 125. Schayer based this hypothesis on a certain
adhtu Stra, now lost, but quoted in various extant sources, which divides the personality into six
elements (the four plus space and consciousness).

328
dependent on a radical view of impermanence. 663 Nevertheless, vestiges of the old

system in which an immaterial realm of dharma was conceived as permanent lived on

in, for example, the idea that space is permanent, 664 the conception of a permanent

dharmadhtu, 665 and conceptions of vija as transmigrating between bodies. 666

More recently, Christian Lindtner has revived the concept of pre-canonical

Buddhism and argued that such a concept can be used to explain how Ngrjuna

developed his Madhyamaka philosophy. Lindtner argues that Ngrjuna was

confronted, in the canonical texts, with contradictory descriptions of nirvasome,

vestiges of the pre-canonical period, conceiving it as a real place that one could go to,

and others, representing the more canonical conception, describing it merely in

negative terms as the end of craving, ignorance, suffering, etc. Ngrjuna felt

obligated, argues Lindtner, to respect both of these descriptions of nirva as

authentic buddhavacana, and so he developed his famous tetralemma to make them

equally correct (or rather, incorrect) descriptions of ultimate truth. 667

Although the concept of pre-canonical Buddhism was mostly abandoned

after World War II, Lindtner notes that the basic concept behind it has lived on

663
Schayer, Precanonical Buddhism, 125-130.
664
Wynne, Origin of Buddhist Meditation, 115.
665
Schayer, Precanonical Buddhism, 129.
666
Schayer, Precanonical Buddhism, 131, cites Buddhaghosa, Visuddhimagga 17.163,
which uses the metaphor of a man using a vine to swing over a ditch to explain the continuity between
lives. Schayer, Precanonical Buddhism, 130, also cites Vijavda as a later development based on
the precanonical understanding of the permanence of consciousness.
667
Christian Lindtner, The Problem of Precanonical Buddhism, Buddhist Studies Review 14,
no. 2 (1997): 109-139.

329
without the name in works since then that have questioned the antiquity of certain

canonical doctrines. In particular, Lindtner cites in this regard the work of

Noritoshi Aramaki and Kamaleswar Bhattacharya. Aramaki argues that the five

skandhas have a long and complex prehistory that can ultimately be traced back to

much more primitive and unsophisticated notions associated with the figure of

Yjavalkya. 668 Bhattacharya, on the other hand, argues that the Buddha, contrary

to orthodox Theravda Buddhism, did believe in a permanent self, and that the

doctrine of anatt was simply a tool for attaining it by first realizing that the body,

consciousness, etc. are not self. 669

Speaking more generally, we can say that the idea of pre-canonical

Buddhism lives on in all attempts to stratify the early Buddhist scriptures or

otherwise specify certain parts or teachings as earlier or later, such as in the case of

the scholars already discussed above. When scholars such as Gombrich and

Schmidthausen argue that the concept of liberating insight underwent a

development that can be traced in the early Buddhist texts; when scholars such as

Bareau, Vetter, and Bronkhorst argue that the account of the Bodhisattva studying

under Klma and Rmaputta is inauthentic; or when Bronkhorst and Zafiropulo

argue that the formless spheres are a late accretion from outside of Buddhism, they

668
Ibid., 111. The work that Lindtner cites is Noritoshi Aramaki, A Text-stratum-analytical
Interpretation of the Concept Pacaskandhas, Jinbun (1980): 1-36.
669
Kamaleswar Bhattacharya, Ltman-Brahman dans le Bouddhisme Ancien (Paris: cole
Franaise dExtrme Orient, 1973); cited by Lindtner, Problem of Precanonical Buddhism, 112. A
similar thesis is presented by Joaqun Prez-Remn, Self and Non-Self in Early Buddhism (The Hague:
Mouton Publishers, 1980).

330
are making a claim similar to Schayernamely, that much of what we find in the

early Buddhist scriptures is the product of a certain prehistory, and that this prehistory

can be traced to a certain extent within those same texts. They only differ from

Schayer in that they do not see this process as having resulted in a rupture so great

that one can postulate a pre-canonical Buddhismalthough even on this point,

Schayer himself admitted that the pre-canonical Buddhism he hypothesized cannot

have been wholly antithetical to the canonical doctrines that followed it. 670

Wynne, however, stands out among more recent scholars who explore

doctrinal development within the early Buddhist texts in that his reconstruction is

parallel in content to Schayers hypothesized pre-canonical Buddhism. The

similarity is quite obvious in the centrality he lends to the same six elements

discussed by Schayer; moreover, he goes beyond Schayer by tracing in detail the

Brahmanical roots of the six elements and tying them directly to the formless

spheres, a doctrinal concept never touched by Schayer in spite of the fact that he

clearly understood the six elements as something to be mystically ascended, as in

meditation. Wynne recognizes the similarity between his own theory and the pre-

canonical Buddhism of Schayer and later Lindtner, and he addresses it by essentially

championing his own theory as a better explanation for the same evidence. That is,

while he recognizes that some of the canonical passages cited by the proponents of a

pre-canonical Buddhism are indeed problematic, he does not believe that they

support the thesis that there was an earlier form of Buddhism that was substantially

670
Schayer, New Contributions, 13.

331
different from what the Canon teaches overall. On the contrary, he argues, they

should be seen as the literary product of some early Buddhists who were influenced

by early Brahmanism. 671 In this respect, Wynne is quite similar to scholars such as

Bronkhorst and Vetter, except that he understands the formless spheres as having

been present in Buddhism from the time of the Buddha himself, as a vestige of his

own pre-Awakening training, rather than being a late intrusion acquired through

simple proximity to non-Buddhist ideas and practices. According to Wynne, the

Buddha taught a path to liberation based on both meditation and insight, and later

Buddhists chose to emphasize one or the other in their formulations of the path.

I am in substantive agreement with these scholars, from de La Valle Poussin

to Wynne, all of whom have come to similar conclusions regarding, in particular, the

development of ideas about liberating insight in coordination with doctrinal

developments and the general lateness of the idea of liberation by insight alone. The

most recent contribution, made by Wynne, however, complicates the earlier work

done by other scholars that would dismiss not only the insight alone template for

liberation as a late literary development, but the eight jhnas plus cessation as well,

due to the fact that it contains elements that are extra-Buddhist. What Wynne has

shown is that there are good reasons to believe that the ideas, at least, behind the four

formless spheres are quite old and cannot be quickly dismissed as late additions.

How exactly the Buddhist tradition came to preserve two quite different

accounts of a meditative path to liberationnamely, the four jhnas plus insight

671
Wynne, Origin of Buddhist Meditation, 117.

332
and the eight jhnas plus cessation templatesis a problem that has not been

addressed as fully as the development of the insight alone template (or, as

Gombrich puts it, how insight worsted concentration in the Pali Canon 672), and

although a full response to the problem is beyond the scope of this work, it is

important enough to warrant some attention here. As we have seen, the earliest

reason for dismissing the four formless spheres as late was Bareaus argument that

the story about ra Klma and Uddaka Rmaputta teaching the Buddha prior to

his Awakening is a fabrication; both Zafiropulo and Wynne, however, have given

convincing refutations of Bareaus argument. 673 In The Two Sources of Buddhist

Meditation, Bronkhorst argued that the formless spheres were a late addition to

Buddhism using the methodological principle that teachings and practices that are

criticized in parts of the Buddhist Canon and bear similarity to those found among

non-Buddhists most likely intruded into Buddhism from the latter non-Buddhist

groups. While this methodological principle does provide a convincing argument that

the formless spheres are extra-Buddhist, it does not demonstrate that their

intrusion (already a loaded word) into Buddhism was late.

More pertinent in this latter respect is Bronkhorsts argument in Dharma and

Abhidharma, which is that the fact that the formless spheres do not appear in the

earliest Abhidharma lists indicates that they are a latecomer to the realm of Buddhist

672
Gombrich, How Buddhism Began, ch. 4, p. 96.
673
This is not to say that the Buddha actually studied under teachers named ra Klma
and Uddaka Rmaputtathis, I think, is ultimately unknowablebut that there are no convincing
reasons to dismiss the story out of hand; moreover, Wynne has presented convincing reasons to believe
that the story as it is presented in the Ariyapariyesaka Sutta is, in any case, quite old.

333
doctrine. Bronkhorsts argument here, however, is not entirely convincing.

Bronkhorst notes the importance of the early list of 37 Wings to Awakening

(bodhi-pakkhiy dhamm), which is found in various canonical texts, 674 but was

expanded upon in the mtk of the Abhidhamma text Dhtukath to include eight

more elements: the four (rpa-) jhnas and the four immeasurables (appama).

This Abhidhamma mtk list does not include the four arpa-jhnas, even though

they are found in numerous canonical suttas. Bronkhorst comes to two conclusions

on the basis of this discrepancy: (1) the arpa-jhnas were a late addition to the

Buddhist tradition and (2) the Abhidhamma texts began their formation before the

Sutta Piaka had itself become closed to changes and additions. Both conclusions, I

believe, clearly read too much into the evidence. Given that a list of 37

bodhipakkhiy dhamm is prevalent in the suttas, the fact that the Abhidhamma

mtk adds to that list in the first place appears to reflect a later provenance. In

addition, the term used for one of these additions, namely the four immeasurables

(appama), appears to reflect relatively late terminological usage. The equivalent,

and apparently earlier, term brahmavihra is found several times throughout the

Nikyas 675 and only once in the Abhidhamma. The term appamaa, on the other

hand, is found only once in the Theragth (6.2) and once in the Sutta Nipta (3.5);

but several times in various texts known to be late, such as the Sagti Sutta of the

674
For an extensive study of the bodhipakkhiy dhamm in the Canon and commentaries, see
R. M. L. Gethin, The Buddhist Path to Awakening (Oxford: Oneworld Publications, 2007).
675
DN 17; MN 83; SN 5.10.2.1-2; AN 5.192; Ther. 14.1; Pe. 6.

334
Dgha Nikya (33), the Apadna (1.9.4, 1.41.5, 1.55.9, 2.3.7), and the Niddesaand

it is clearly the preferred term in the Abhidhamma, where it appears 22 times.

If the earliest Abhidhamma mtk failed to include the four formless spheres,

then, it is difficult to attribute this to the antiquity of the list and the lateness of the

formless spheres; if anything, it may simply reflect the preference of the

Abhidharmists for the four jhnas plus insight template over the eight jhnas plus

cessation template. Schmidthausen has already noted that the transmitters of

different parts of the Canon (at least within the Theravda school) appear to have had

different predilections with respect to these two templates for the path to liberation.

In particular, the four jhnas plus insight template is found in the DN, MN, and AN,

but not in the SN 676; the insight alone template is found in the MN and SN, but not

in the DN and AN 677; and the eight jhnas plus cessation template is found in all

four of the main Nikyasalthough in the DN and SN it is presented simply as a set

of nine successive states, omitting the statement at the end that one sees with insight

and destroys the savas. 678 If the transmitters of individual Nikyas had preferences

for certain templates over others, it is not hard to imagine that the Abhidharmists had

their own preferencesnor, in particular, given their intellectualist bent, that they

would have prefered the simplified system of four jhnas culminating in insight to the

more elaborate system of eight jhnas that emphasized meditation over insight.

676
See, in particular, Schmidthausen, Some Aspects, 204 n. 15.
677
Ibid., 219 n. 69.
678
Ibid., 216 n. 55.

335
Conversely, the sheer pervasiveness of the eight jhnas plus cessation template

within the four main Nikyas, I would argue, could be a sign of its antiquity. Another

sign is the conspicuous absence of the four jhnas plus insight template from the

Sayutta Nikya, which, as we will see in Part IV, may be a sort of Ur-Nikya

from which the Dgha and Majjhima sprung.

This is not to say that there is no evidence that makes an argument for the

antiquity of the eight jhnas plus cessation template, and the meditative practices

that Wynne sees as lying behind it, problematic. Schmidthausen has noted that there

appear to have been different predilections for one template of liberation over another

not only within the Theravda tradition, but also between different sectarian traditions;

he notes that while the eight jhnas plus cessation template is common in the Pali

Majjhima Nikya, it is rare in the Chinese Madhyamgama. 679 And as we have

already noted, Zafiropulo has shown through an exhaustive examination of all extant

texts that the four jhnas plus insight template is more pervasive across the various

sectarian traditions than the eight jhnas plus cessation template and, in his opinion,

older. In addition, Wynne links the formless meditations to the system of six

elements in large part through kasia meditation, but the evidence for kasia

meditation appears to be somewhat late. The word kasia is found only a few times

in the four main Nikyas, 680 but it is found with much greater frequency in late

679
Ibid.
680
DN 33, 34; MN 49, 77, 102; AN 382-493-562, 10.25-26, 10.29-30. Note that DN 33
(Sagti Sutta) and 34 (Dasuttara Sutta) are late, Abhidhamma-like texts. Note also that the word
kasia is not found at all in the SN.

336
canonical texts such as the Niddesa, Paisambhidmagga, Nettippakaraa, and

especially the Abhidhamma Piaka. Indeed, the sort of meditative path that Wynne

envisions as lying behind the scheme of eight jhnasnamely, beginning with

kasia-meditation on the gross elements and then progressing to the more subtle

elements in the formless spheresfinds perhaps its fullest and most detailed

expression in the quite late (5th century CE), sectarian Theravda work of

Buddhaghosa, the Visuddhimagga. In this work, kasia-meditation is explicitly

taught as a means to attain the four lower (i.e., rpa-) jhnas (ch. 4), which is

followed later by attainment of the four formless spheres (ch. 10). 681

How, then are we to explain this divergence in the evidence? Is Wynne right

to see the four formless spheres as deriving from pre-Buddhist practices that were

actually taught to the Buddha himself, or are scholars like Bronkhorst right to see

them as a later, extra-Buddhist intrusion into the tradition? I believe that a

reexamination of some of the methodological assumptions made by Wynne helps to

shed some light on this problem. As already discussed, Wynne is critical of the

concept of pre-canonical Buddhism, and he argues that the evidence for different

beliefs is presented in the simplistic form of a dichotomy between precanonical

Buddhism vs. canonical Buddhism. 682 Wynnes point here is surely well taken:

681
This having been said, it should be remembered that the Visuddhimagga is a synthetic
work: kasia-meditation and the attainment of the lower jhnas is not followed immediately by the
four formless spheres, but by a discussion of various other meditative techniques, including foulness
(ch. 6), the six recollections (ch. 7), mindfulnesss (ch. 8), and the brahmavihras (ch. 9); moreover,
after the discussion of all of these various methods for attaining concentration in the first half of the
work, the second half of the work is dedicated to the cultivation of insight.
682
Wynne, Origins of Buddhist Meditation, 116.

337
The Canon was formed over many centuries and shows evidence of change and

contestation within it, so it is difficult to determine where the line between canonical

and pre-canonical should be drawn; indeed, many of the scholars whose work we

have looked at in this chapter take advantage of the conflicting teachings found

within the Canon as it comes down to us to come to historical conclusions without

recourse to a hypothetical pre-canonical Buddhism. Wynne, however, himself

takes recourse to a similarly false dichotomynamely between Buddhism and pre-

Buddhism. This dichotomy is problematic in two ways.

First, it is closely related to another problematic dichotomynamely, between

Brahmanism and non-Brahmanismin that Wynne clearly assumes that the pre-

Buddhism in which the Buddha was trained and from which he both developed his

ideas and distinguished himself as a teacher was Brahmanical. Wynne finds parallels

to Buddhist teachings on the six elements, kasia-meditation, and the formless

spheres in Brahmanical texts such as the Upaniads and the Mahbhrata; from this

he concludes that these elements of the Buddhist tradition were Brahmanical, 683

and as a corollary he concludes that ra Klma and Uddaka Rmaputta were

Brahmanical teachers from whom he learned these Brahmanical ideas. As I

argued in Part II, however, we should not read ancient Indian history through the lens

683
To be fair, the argument Wynne makes is that these ideas make more sense in the context
of the Brahmanical texts where they are found than in the Buddhist texts; in particular, Brahmanical
texts include cosmogonies in which brahman or tman devolves into ever more gross elements, which
appears to be the motivation behind meditative techniques that seek to ascend from gross to more
subtle elements. While I think that this does prove that these ideas are pre-Buddhist, since the
original context in which they were developed has been lost in the Buddhist tradition, there is no
reason to believe that the Brahmanical tradition owned the cosmogonical speculations in question.

338
of an unproblematized dichotomy between Brahmanical and non-Brahmanical.

This dichotomy, I have argued, is predicated on an uncritical acceptance of the new

Brahmanical claim that Brahmans, defined by birth and their knowledge of the

Vedas, are the uniquely legitimate custodians of a Brahmanical tradition that has been

transmitted without substantial change from the ancient pastwhat Bronkhorst has

called the brahmanical colonization of the past. 684 Uncritical acceptance of this

orthodox Brahmanical depiction of Indian historywhich dates within the Western

Buddhology, as I argued in the Introduction, to Eugne Burnoufhas led scholars to

assume that there was a clearly definable Brahmanical tradition that was the

primary agent in ancient Indian history, and that terms, teachings, ideas, and practices

that came to be associated primarily with the Brahmanical tradition were somehow

owned by that traditionthat Brahmanical uses of these terms, ideas, etc., were

primary, and all others secondary. What I have argued is that the Brahmanical

tradition should be seen as a product of centuries of contestation over various terms,

ideas, teachings, and practices, and not as a pre-existing agent that uniquely created

and owned them.

Parallel elements preserved by different traditions, I believe, are better

explained by substratal model such as proposed by Ruegg. Indeed, there is good

reason to believe that all three of the major traditions that emerged from Northern

India in the late first millennium BCEBuddhism, Brahmanism, and Jainismdrew

in part from a common substratum of ideas, teachings, and practices. Earlier, in

684
Bronkhorst, Buddhism in the Shadow of Brahmanism, 65.

339
Chapter III.2, of course, I made the argument that the word brhmaa was a

substratal term that came to be appropriated by both the Brahmanical and the

Buddhist traditions; a parallel example is the word sava (Skt. srava), which was

appropriated by both the Buddhist and the Jain traditions. 685 In addition, the work of

several scholars has shown that there are parallel verses shared by the Pali gth texts,

early Jain texts written in Ardhamgadh, and the Mahbhrata. 686 Even the literary

characters representing particular holy persons may have been to a certain extent

drawn from a common substratum, as is suggested by the Jain text Isibhsiyi

(Sayings of the Seers), which attributes verses to a multitude of famous is

including Vardhamna and Prva, but also Mahkyapa, Yjavalkya, Uddlaka,

and possibly even riputra and Maskarin Gola, all of whom the text treats on equal

terms. 687 And as Brian Black has recently shown through a comparison of the

685
The word srava etymologically means effluent, which doesnt make sense in the
context of the Buddhist understanding of the word (which is ethical), but does make sense in the
context of the Jain understanding (which has to do with the physical flow of karma, conceived of as
matter). This does not mean that the word is Jain and the Buddhists borrowed it from them, however;
it is likely that both borrowed the term from a common substratum, and that the Jains retained a
definition of the term more in line with its original meaning. For citations of relevant scholarship, see
J. W. de Jong, The Buddha and His Teachings, in Wisdom, Compassion, and the Search for
Understanding. The Buddhist Studies Legacy of Gadgin M. Nagao, ed. Jonathan A. Silk (Honolulu:
University of Hawaii Press, 2000), 172.
686
Work on parallels between Pali and Ardhamgadh verses has been done by Bolle, and on
parallels in the Mahbhrata by Rau and Murakami; for citations, see de Jong, The Buddha and His
Teachings, 173.
687
For an edition of the Ardhamgadh text, with an introduction and translation into Hindi
and English, see Mahopadhyay Vinaysagar, ed. and trans., Isibhasiyaim Suttaim (Rishibhashit Sutra),
trans. Kalanath Shastri and Dinesh Chandra Sharma (Jaipur: Prakrit Bharati Academy, 1988). In his
introduction, Vinaysagar attempts to identify as many of the seers in the text as possible, including the
ones mentioned above. In some cases the identifications are rather uncertain, as in the case of
Maskarin Gola (identified with a seer whose name is given as Makhaliputta in the Ardhamgadh)
and riputra (identified with a certain Stiputta Buddha). Although the the identification of the
latter is somewhat dubious given that there is no philological explanation for an r transforming into a t,
the identification of this seer with some sort of Buddhist figure is made plausible for two other

340
Upaniadic depiction of vetaketu and Buddhist depiction of Ambaha, narrative

tropes appear to have been shared across sectarian boundaries even when particular

characters were not shared by name. 688

Returning now to the question of the origin of the formless meditations in

Buddhism, we can see that, even if we accept the account of the Buddha studying

under ra Klma and Uddaka Rmaputta in the Ariyapariyesana Sutta as having

actually, historically happened, their influence on the Buddha is better explained as a

substratal influence on Buddhism, rather than a Brahmanical one. The earliest

version of the story, in spite of the conclusion that Wynne draws, never identifies

Klma and Rmaputta as Brahmans. Indeed, there is nothing about the story that

would suggest any sort of connection with Brahmans or Brahmanism at all, if it were

not for Wynnes detailed excavation of the parallels between Klmas and

Rmaputtas ideas, the formless spheres, and ideas found in certain Brahmanical

texts.689 But as we saw in P, non-Buddhist interlocutors are generally identified

specifically as Brahmans only in certain encounter dialogs in which the

interlocutors clearly are proponents of the new Brahmanism who debate the

Buddha specifically on the tenets thereofnamely, the vara system, the superiority

of Brahmans by birth, knowledge of the Vedas, ritual, etc.

reasons. First, of course, is the fact that he is referred to as Buddha, a title given to none of the other
seers in the text. Second, the actual content of the verses attributed to this seer has a vaguely
Buddhist ring to it, in particular Isi. 38.2-4, which seems to refer to the middle path.
688
Black, Ambaha and vetaketu.
689
This is in clear contradistinction with Buddhacarita ch. 12, which clearly depicts Ara
Klma and Udraka Rmaputra as Brahmans (see esp. v. 42) teaching about the tman. But as
Bronkhorst has argued, this later text by Avaghoa thoroughly Brahmanizes the Buddhas life story
Buddhism in the Shadow of Brahmanism, 154-5.

341
 The story about ra Klma and Uddaka Rmaputta is much more similar to

a different genre of early Buddhist texts, namely, what I have called teachings on

wrong views, in which rival teachers with wrong views (a category to which Klma

and Rmaputta must certainly belong) are generally referred to collectively as

samaabrhmaa category that, as we have seen, serves as a foil against which to

construct Buddhist identity and does not serve, in spite of being grammatically a

dvandva, in any meaningful way to distinguish samaas and brhmaas as separate

categories. In the same way, the earliest Buddhist tradition available to us did not see

fit to identify Klma and Rmaputta with the category Brahman, but simply with

the amorphous mass of teachers of wrong views, out of which also sprung the likes of

Makkhali Gosla and Nigaha Ntaputta. As evinced by the extensive genre of

teachings on wrong views, this population of various teachers clearly shared a

variety of views not only with one another, but with specific textual traditions

including the Buddhist Nikyas/gamas; the Jain Agas; and, yes, the Upaniads.

These commonalities are better explained by a common substratum, however, than by

privileging one particular sect that emerged from the debates and contestation of the

late first millennium BCE as the source.

The second way in which I find the dichotomy Wynne poses between

Buddhism and pre-Buddhism problematic is his insistence that we can determine

what the historical Buddha actually taught and from that deduce which teachings in

the Buddhist scriptures are authentic and which are inauthentic. In this respect,

Wynne follows Gombrich, who in his latest book claims to show not only what the

342
Buddha taught, but What the Buddha Thought. Wynne summarizes his method as

follows:

the Piakas suggest a person of a certain background and character; the

background and character of the Buddha that emerge from the Piakas allow
us to hypothesize the Buddhas manner of teaching; thus, particular instances
of the Buddhas teaching that conform to what we have deduced to be his
manner of teaching are likely to be historically authentic. 690

That such a method is problematic is easily seen by engaging in the following thought

experiment. Imagine that scholars centuries from now had nothing with which to

reconstruct the historical Barack Obama other than a large corpus of broadcasts

from Fox News. Using Wynnes method, these scholars would doubtless have to

dismiss as inauthentic Obamas ordering of a commando raid to kill Osama bin

Laden as inconsistent with the background and character of [Obama] that emerge

from [the corpus of Fox News broadcasts].

Like many scholars today, I am less interested in determining exactly what the

Buddha taught and didoften justly criticized as a misguided quest for origins

and more interested in studying portrayals of the Buddhas teachings and actions, and

how and why those portrayals changed over time. Indeed, I agree with Wynne, as

well as numerous other scholars such as the ones whose work we have explored in

this chapter, that we can fruitfully explore the development of depictions of the

Buddhas Awakening and teachings about how liberation comes about. One of the

most interesting findings Wynne presents in his work is that there are references to

teachings reminiscent of the formless spheres in the dialogs with Upasva and Posla

690
Wynne, Origin of Buddhist Meditation, 128.

343
in the Pryaa. Wynne has also shown that there are terms reminiscent of the four

rpa-jhnas in the dialog with Udaya in the same text. The key word here, however,

is reminiscent. What we do not find in the Pryaa is a clear enumeration of or

teaching about the four rpa-jhnas, the four formless spheres, or insight meditation

that comes anywhere close to the detailed templates, found in the four main Nikyas,

that have been identified by Schmidthausen. The likely antiquity of the Pryaa,

while lending support to the antiquity of numerous ideas found in the templates of the

four main Nikyas, including meditation on nothingness, certain factors of jhna, and

the combination of mindfulness with meditative concentration, lends no support to

any of the templates in their particular details. Worse still, if we turn to the Ahaka,

parts of which at least may be even older than the Pryaa, we find, as Burford has

shown, ideas at odds with the rest of the Canon, including the eschewing of all views

and an apparent lack of concern even with achieving liberation from rebirth. If the

practice of jhna, insight, or the formless spheres are confirmed by only vaguely

reminiscent terminology in the Pryaa because it is likely a very old text, what are

we to make of the equally old Ahaka, which seems to lack any interest in meditative

practices at all? 691

I believe that a healthy agnosticism about the origins of Buddhism allows

one to see that it is unlikely that a very clear, detailed teaching about how to meditate

691
This is not to say, however, that the teachings in the Ahaka have no connection with later
Buddhist doctrine, including on meditation and liberation. In particular, I think that the stated goal of
freedom from apperception (sa) found in several suttas of the Ahaka is reminiscent of, and may
prefigure, the state of cessation from apperception and feeling (savedayitanirodha) found in the
eight jhnas plus cessation template of later suttas.

344
or attain liberation was bequeathed to the early Buddhist tradition by the Buddhaor

in any case informed the development of the three templates delineated by

Schmidthausen, which do offer clear, detailed teachings on the subject. Indeed, this

appears to be confirmed by yet another text that Wynne has convincingly argued may

be quite oldnamely, the training of the Bodhisattva under ra Klma and

Uddaka Rmaputta and subsequent Awakening, as found in the Ariyapariyesana

Sutta. As we have seen, the Buddhas Awakening is described in quite simple terms

in this textlittle is said other than that he attained nibbna and became aware that

this would be his last birth. Even Wynne himself admits that a weakness of his

method is that the source of historical facts deduced, the Ariyapariyesana Sutta, has

nothing to say about what the Buddha taught. 692 I think a reasonable conclusion to

draw from this evidence is that the earliest Buddhist tradition was less concerned with

what precisely the Buddha taught about Awakening, or the precise mechanics of how

liberation comes about, than with the mere fact that the Buddha was in fact the

Buddha, i.e, Awakened. In this respect, the Buddhist tradition would parallel the

Christian tradition, for which the earliest extant literaturePauls lettersare less

concerned with what Jesus taught than with establishing that he is, in fact, the

Messiah.

If this is indeed the case, then all three of the templates delineated by

Schmidthausen would be early attempts to produce a systematic explanation of

liberationall of which have roots in earlier teachings such as found in the Pryaa,

692
Ibid., 125.

345
but which each emphasize certain ideas and techniques over others and thus led to the

development of differing schools of thought on the nature of the Buddhist path and of

liberation. Moreover, insofar as these templates take reflect varying positions in the

debate between intellectual and ascetic/ecstatic approaches to Awakening that is

evident in the early Buddhist texts, then they are reflective, la Sarbacker, of a wider

pan-Indian debate over the numinous and cessative dimensions of meditation and

yoga. They are, in other words, not better or worse candidates for the original

teaching of the Buddha, but a record of the early Buddhist traditions ongoing

engagement with its intellectual milieu.

III.3.3 The Three Knowledges Within the Context of the Development of Buddhist

Discourses on Awakening

We are now prepared to return to the question that has motivated this

discussion on the development of early Buddhist discourses on meditation, insight,

and attaining liberationnamely, the three knowledges (tisso vijj), consisting of

recollection of past lives, ability to see peoples rebirths, and destruction of the

savas. Based on what we have seen in this chapter, there are a couple of immediate

conclusions we can come to. To begin with, it is unlikely that the three knowledges

were a part of the earliest Buddhist descriptions of the Buddhas Awakening or the

path to liberation. The template in which the three knowledges are found, four

jhnas plus insight, is almost definitely older than the insight alone template,

which latter was most likely derived from it, and some scholars have argued that it is

346
more original than the eight jhnas plus cessation template as well. As we have

seen in this chapter, however, there are good arguments to be made for very early

influences from formless-sphere-like meditative practices on Buddhism; moreover, it

is likely that the earliest Buddhist tradition simply did not offer a very clear or

detailed account of the Buddhas Awakening or the exact mechanism by which

liberation is accomplished, and thus that all three of the templates for liberation

delineated by Schmidthausen were developed somewhat later, based in part on ideas

found in older texts such as the Pryaa, but each going in a different direction,

emphasizing different approaches to attaining liberationmostly insight in insight

alone, mostly meditation in eight jhnas plus cessation, and a more even mixture

of both in four jhnas plus insight.

Even within the context of the four jhnas plus insight template, however, it

is unlikely that the three knowledges were an original element. As already discussed

above, Schmidthausen has noted that in some cases the insight portion of this

template is presented as the realization of only a single truth, namely, the four noble

truths, similar to what we find in the Dhammacakkappavattana Sutta. Moreover, he

has argued persuasively that the one truth version is the original version, and that

the other two truths were added later, as evinced by the fact that they are cast

anomalously in a different tense. Likewise, the inclusion of the destruction of the

savas in the third truth belies its own lateness in the fact that it is cast in terms that

are clearly modeled on the four noble truthsthe savas themselves, their origin,

their destruction, and the path to their destruction. As Schmidthausen notes, this

347
realization, while mimicking the realization of the four noble truths perfectly, does

not make sense on its own terms, since the origin of the savas is said to be ignorance,

but ignorance itself is one of the savas! 693

A wider look at the distribution of references to the three knowledges in early

Buddhist texts also supports the supposition that it does not represent a particularly

old trope within the Buddhist tradition. Both the nominal phrase tisso vijj and the

adjective tevijja used to describe the ideal Buddhist person (as opposed to a traividya

Brahman) are found scattered throughout the Pali Canon, without any particularly

high frequency in most Nikyas and individual books. 694 Unlike the term brhmaa,

they are not found at all in the Pryaa and Ahaka. This is in contradistinction

with the term sava, which is found bereft of any three knowledges context in both

of these early texts 695a fact that we should not find surprising given that, as already

discussed, the term sava is almost certainly pre-Buddhist. The terms tevijja and

tisso vijj are found, however, with relatively high frequency in the Theragtha, 696

Thergth, 697 and Apadna. 698 The Apadna is definitely a late text, and much of

693
Schmidthausen, Some Aspects, 205.
694
tevijja: Vin. 1.1.4, 2.5.1.8, 5.6.2; MN 56, 71, 73, 91; SN 1.6.1.5, 1.7.1.8, 1.8.7, 1.8.9-10,
1.11.2.8; AN 3.59-60; Iti. 3.5.10; Sn. 3.9; also found in Jt., Nidd., and Mil. tisso vijj: Vin. 1.1.4,
2.5.1.8, 5.7.1; DN 33, 34; MN 86, 98, 145; SN 1.7.1.8, 4.1.9.5, 5.8.2.14; AN 3.59-60, 8.30, 10.102; Ud.
3.3; Pv. 4.1; Iti. 3.5.10; Sn. 3.9; also found in Pais. and Mil.
695
Sn. 4.13.913, 5.7.1082-3. The term ansavo is also found in the vatthugth of the
Pryaa (v. 996), although this is, as we have seen, certainly late.
696
tevijja: 1.12.2, 2.1.5, 3.1, 4.8, 5.5, 20.1, 21.1; tisso vijj: 1.3.4, 1.6.5, 1.7.6, 1.11.7, 1.11.8,
1.12.7, 3.1-2, 4.1-2, 4.5, 4.9, 4.12, 5.1, 5.7, 6.6, 7.3, 8.3, 10.4, 13.1, 16.8-9.
697
tevijja: 4.1, 5.11-12, 7.1, 12.1, 13.2, 13.4-5; tisso vijj: 2.4, 2.6, 4.1, 6.4, 7.2, 7.3, 8.1, 9.1,
13.3-4, 15.1.

348
the Theragth and Thergth are late as well 699; in addition, the three are related in

that all three deal with particular named monks, and the Apadna in fact draws much

of its material from the Theragth and Thergth. This would seem to indicate that

the trope of referring to the ideal person as tevijja, or as possessing the tisso vijj,

became popular within a particular, fairly late genre that focused on exalting the

accomplishments of the Buddhist communitys best and brightest.

An example from this genre demonstrates how the terms tevijja and tisso vijj

are used in explicit comparison with the lifestyle of the literal Brahman to cast the

Buddhist ideal in a favorable lightthat is, used polemically. In a short set of three

verses, a Buddhist monk contrasts his current state as a liberated arhat to his former

life as a Brahman, most likely a jaila:

Unclean, seeking purity, I attended the fire in the forest.

Not knowing the way to purity, I did tapas for the deathless.

This happiness is attained by happiness. See the good nature of the dhamma.
The three knowledges were attained, the instruction of the Buddha done.

In the past, I was a brahmabandhu; now indeed am I a Brahman.

I am tevijja (Skt. traividya) and a nhtaka (Skt. sntaka); I am sottiya (Skt.
rutriya) and Veda-gone (vedag). 700

698
tevijja: 1.1.3-1, 1.3.3, 1.9.4, 1.14.7, 1.40.3, 1.52.9, 2.3.7; tisso vijj: 1.1.3-1, 1.1.3-6, 1.1.3-
10, 1.14.1, 1.23.1, 1.38.4, 1.39.1, 1.40.1, 1.40.10, 1.41.1, 1.41.5, 1.42.1, 1.43.1, 1.44.1, 1.44.8, 1.45.1,
1.46.1, 1.47.1, 1.48.1, 1.49.1, 1.50.1, 1.51.1, 1.52.1, 1.53.1, 1.54.1, 1.55.1, 1.55.7, 1.56.1, 2.1.1, 2.2.7,
2.3.1, 2.3.10, 2.4.1, 2.4.10.
699
Dating material from the Theragth is complicated by the fact that the collection appears
to have undergone a long period of development; some parts may be early, but other parts can be
determined on metrical grounds to be relatively late. For a general discussion, see K. R. Norman,
trans., The Elders Verses I: Theragth (London: Luzac and Company, 1969), xxvii-xxix. For a more
technical discussion of the issues involved in dating the Theragth according to meter, see Warder,
Pali Meter, 95-6, 98, 115, 135, 225-6.
700
Ther. 3.1.219-221: ayoni suddhimanvesa, aggi paricari vane / suddhimagga
ajnanto, aksi amara tapa // ta sukhena sukha laddha, passa dhammasudhammata / tisso

349
Here we have a clear case in which terms associated with literal Brahmans,

including the word brhmaa, are used polemically to contrast the present state of a

Buddhist arhat to his former life as a (most likely) jaila Brahman. No attempt is

made here to explain what is meant by tevijja in the context of describing a Buddhist

arhatthat is, nothing is said about what three knowledges he possessesany

more than is an attempt made to explain in what sense an arhat is bathed (nhtaka)

or what exactly he has heard (sottiya). The only point being made is that only

through the Buddhas instruction has the person in question become worthy of these

honorifics.

Other texts, however, do attempt to specify exactly what is meant by words

such as tevijja in the Buddhist context. A good example is provided by the Tevijja

Sutta of the Itivuttaka (3.5.10), in which the Buddha directly addresses what he means

by the word tevijja. The opening line of this sutta makes clear that the definition of

the term being offered is a direct response to the Brahmanical meaning of the term:

According to dhamma, monks, do I declare [one] a tevijja Brahman, not another by

measure of muttered mutterings. 701 The Buddha then explains in great detail each of

the three knowledges according to which he declares one a tevijja Brahman:

recollection of past lives, knowledge of peoples rebirths, and destruction of the

vijj anuppatt, kata buddhassa ssana // brahmabandhu pure si, idni khomhi brhmao /
tevijjo nhtako camhi, sottiyo camhi vedag //.
701
dhammenha, bhikkhave, tevijja brhmaa papemi, na lapitalpanamattena.

350
savas. At the end of the description of the first knowledge, we find the following

formula:

This is the first knowledge (paham vijj) attained by him. Ignorance is

destroyed; knowledge is arisen; darkness is destroyed; light is arisenas it [is]
for one dwelling vigilant, ardent, resolute. 702

This formula is repeated as a sort of refrain for the second knowledge (dutiy vijj)

and the third knowledge (tatiy vijj).

In the context of an actual, extended description of the process of liberation

in this case, the four jhnas plus insight templatethis same refrain (with minor

variations) is sometimes used in the description of the knowledges one attains after

attaining the fourth jhna to enumerate them explicitly as the first, second, and third

of the tisso vijj. 703 More commonly, however, the three knowledges are simply

described as such, without this refrain. 704 In these latter cases, the knowledges not

only are not enumerated; the word vijj is not even used to describe them at allthe

word for knowledge used instead is a. Although it is difficult to know with

certainty the relationship between these two formulations of the three knowledges

within the context of the four jhnas plus insight template, I believe that the

latterthe version without the vijj-refrainis most likely later. It is more common

overall in the Canon; it is found in Dgha texts in which the four jhnas plus insight

702
ayam assa paham vijj adhigat hoti, avijj vihat, vijj uppann, tamo vihato, loko
uppanno, yath ta appamattassa tpino pahitattassa viharato.
703
Vin. 1 Verajakaa, MN 4, 19, 36, 85, 100; AN 3.59-60, 8.11.
704
DN 2, 10, 28; MN 27, 39, 51, 60, 65, 76, 79, 94, 101, 125; AN 4.198, 6.64. Other examples
can be found in later texts, including the Nidd., Pais., Mil., Nett., and esp. in the Abhidhamma Piaka.

351
template has been expanded to include the attainment of various insights and

supernormal powers; and it is found, unlike the version with the vijj-refrain, in later

texts of the Khuddaka Nikya and Abhidhamma Piaka. This would suggest that the

three knowledges were originally introduced into the four jhnas plus insight

template with the vijj-refrainmost likely, I would argue, to root an explanation of

the Buddhist meaning of tevijja, such as we find in Iti. 3.5.10, in an actual

description of the mechanics of liberation. Thus, the word tevijja was transformed

from a mere rhetorical tool with which to contrast the arhat with the literal

Brahman, such as we find in Ther. 3.1, into an actual set of three knowledges, such as

we find in Iti. 3.5.10, and finally into an integral part of the mechanics of liberation

itself, as described by the four jhnas plus insight template. 705 Later, as the four

jhnas plus insight template came to be formulaically reproduced in various contexts

as a succinct description of how one attains Awakening, the vijj-refrain that tied the

three knowledges attained at liberation to the polemically relevant term tevijja were

omitted as inessential in the immediate context of discussing liberation.

What we have found in this chapter, then, is that the word tevijja, unlike the

word brhmaa, does not appear to have been used in the very earliest Buddhist texts,

and appears to have been introduced at a somewhat later stage as a rhetorical device

for use in polemical comparisons of the ideal Buddhist person, the arhat, to literal

Brahmans. It is possible that at first the word tevijja was adopted merely

705
By this, I do not necessarily mean to imply a chronological progression from Ther. 3.1 to
Iti. 3.5.10 to the four jhnas plus insight template. I simply mean that there is a logical progression
in the uses (and non-use, in the final case) of the word tevijja in these texts that illustrates how the
three knowledges likely came to be incorporated into the description of liberation.

352
rhetoricallysuch as we find in Ther. 3.1in conjunction with other Brahmanical

terms such as nhtaka and sottiya, simply to assert that the Buddhist arhat is the only

one worthy of such honorifics, and that only later was an attempt made to explain

how exactly a Buddhist arhat can be called tevijja by specifying what three

knowledges he possesses. In any case, the conclusion I am assertingthat the

Buddhist use of the term tevijja, unlike the term brhmaa, was not taken from a pan-

Indo-Aryan substratum, but was, from the beginning, borrowed from the proponents

of the new Brahmanism in order to use polemically against themis supported by the

history of the equivalent Sanskrit terms, traividya and traividyaka, in the Brahmanical

literature. As discussed above, brhmaa is a very old term, being found several

times in the oldest extant Indian text, the g Veda Sahit and throughout the rest of

the Vedic literature. Because this Vedic literature developed slowly over many

centuries, and the ideological claims of the new Brahmanism only coalesced at the

end of that development, we can conclude, as I stated at the end of the last chapter,

that there were Brahmansor, in other words, the term brhmaalong before there

was a Brahmanism to arrogate ownership of the term to itself.

The exact opposite is the case for the terms traividya and traividyaka. These

terms are quite latein fact, we can justly call them neologismsthat did not first

appear until the post-Vedic stra literature. 706 Far from predating Brahmanism, these

terms appear to refer to the ideal person as constructed by the new Brahmanism

706
For the distribution of the terms traividya and traividyaka in the Vedic and post-Vedic
literature (in this case being confined to the latter), see Vishva Bandhu, A Vedic Word-Concordance,
vol. 5 (Index), part 1 (Hoshiarpur: V.V.R. Institute, 1964), 366.

353
movementa Brahman trained (ideally, at least) in all three Vedas, which are

conceived of as a fully developed corpus of knowledge. This traividyaka, in fact, is

precisely the sort of person that serves as the Buddhas interlocutor in the encounter

dialogs found among the early Buddhist texts, and in a few cases the equivalent Pali

terms tevijja and tevijjaka are used in encounter dialogs to refer to these

interlocutorsnot in a polemical sense, as found elsewhere in the Canon, but simply

in reflection of the terms these Brahmans used for themselves. 707 Therefore, whereas

brhmaa was a very old term in the Indian tradition that was inherited by both the

early Buddhists and the proponents of the new Brahmanism from a common

substratum, tevijja clearly was a term that was coined by the proponents of the new

Brahmanism and then borrowed by the early Buddhists to use polemically against

them.

707
DN 3, 4, 13; MN 95, 98, 100.

354
 Chapter III.4

Conclusion to Part III

In Part III, I have argued that we cannot use a single model of polemical

borrowing to interpret all cases in which seemingly Brahmanical terms are used in

the early Buddhist texts. Doing so involves the assumption that a certain set of terms

are unproblematically Brahmanicalthat, in other words, Brahmanism is an

unchanging, monolithic entity whose use of certain terms is somehow more original

or legitimate than that of other groups. I have argued that we should abandon

attempts to draw a hard and fast distinction between Brahmanism and non-

Brahmanism. That is, we should understand Brahmanism, like Buddhism and

Jainism, as emerging out of centuries of contestation over various terms, ideas, and

practices, rather than as a meta-historical agent that pre-dated, and participated in,

that period of contestation without change to its essence.

In particular, I argued that the term brhmaa should not be seen as an

exclusively Brahmanical term that was borrowed by the Buddhists for polemical

purposes. Instead, it should be understood as an old honorific, drawn from the pan-

355
Indo-Aryan substratum, that was contested by both Buddhists and proponents of the

new Brahmanism in the late first millennium BCE, but eventually was mostly

ceded to the latter as it became increasingly central to their articulation of their own

identity. In Chapter III.2, I supported this argument by showing that the earliest uses

of the word brhmaa appear to have been made without any polemical intentthat

is, without any comparison to the Brahmans of the new Brahmanism. Then, as I

showed in Chapter III.3, the Buddhist tradition transformed, through commentary and

narrative framing, these early uses of the word brhmaa into polemic by introducing

comparisons with literal Brahmans where none existed before.

To put this history of the Buddhist use of the word brhmaa into perspective,

I then contrasted it, in Chapter III.3, with the history of the Buddhist use of the word

tevijja. I argued that tevijja is not found in the earliest Buddhist texts, but that once it

was introduced, it was always used with a polemical intenti.e., to contrast the

Buddhist arhat as the true tevijja with proponents of the new Brahmanism who call

themselves tevijja. This contrast between the history of the use of the word

brhmaa in the Buddhist tradition and that of tevijja is consistent with the fact that

the former is an ancient term, and thus could have been taken by the Buddhists from a

pan-Indo-Aryan substratum, while the latter is a neologism created by proponents of

the new Brahmanism to exalt their own knowledge of the Vedas. At the same time, I

also made the argument that the various Buddhist templates for achieving

Awakening that are found in the early Buddhist texts probably developed slowly over

time by emphasizing and systematizing particular ideas found in the earliest tradition,

356
and that similarities between them and ideas found in Brahmanical texts are better

explained by a common substratum than by an assumption of Brahmanical priority.

At a certain point, the concept of tisso vijj was introduced into one of the Buddhist

templates (four jhnas plus insight) in order to ground the polemical use of the

word tevijja in an actual liberatory attainment of three knowledges.

What we have found so far, then, is that there appears to have been a

development in the early Buddhist traditionaway from simple utilization of terms,

ideals, and concepts from the pan-Indo-Aryan substratum, such as the concept of

brahmacariya and the honorific brhmaa, to an increasingly strident polemic

against the new Brahmanism. This polemic sought to demonstrate the superiority of

the ideal Buddhist person by representing him as the only one truly worthy of the

terms the proponents of the new Brahmanism used to describe themselves, including

old terms such as brhmaa and new ones such as tevijja. I have argued that this

took place as the new Brahmanism became an increasingly vocal force, around the

time of the composition of their signature texts, the Dharma Stras, in the centuries

just prior to the turn of the era. Buddhists responded to the threat in part by using the

terms favored by the proponents of the new Brahmanism polemically against them,

but more directly by refuting their claims logically in a new genre of texts that I have

been referring to as encounter dialogs. In Part IV, therefore, we will turn to an

examination of these texts, the rhetorical strategies they employ, and their

development as a literary genre.

357
 Part IV

Encounter Dialogs in the Context of Oral

Tradition

358
 Introduction

Encounter Dialogs

So far in this dissertation, I have focused on aspects of the early Buddhist

textssuch as the use of the compound samaabrhma to refer to a single class of

persons, rather than two antagonistic classes (Part II), and the use of the term

brhmaa as an honorific to refer to the Buddhist ideal person (Part III)that are

difficult to reconcile with a model of Buddhism as arising in antagonistic opposition

to Brahmanism. The latter model, if flawed, however, is needless to say not without

any basis whatsoever in the early Buddhist texts. Over the course of the preceding

pages, I have several times had occasion to refer to encounter dialogs, which I have

defined as stras in which the Buddha meets with an interlocutor who is identified

explicitly as a Brahman and thus set in opposition to the Buddha himself, who in turn

is often identified explicitly as a non-Brahman, i.e., as a ramaa. Although, as we

will see, encounter dialogs are not entirely uniform, 708 texts within this genre are

708
In choosing to refer to this genre of texts as encounter dialogs, I have attempted to use a
fairly neutral term so as to leave the category open to as wide a relevant variety of texts as possible.
Many encounter dialogs, indeed those that have most commonly caught scholars attention as

359
defined in distinction from other texts that refer to Brahmans by the fact that the

Buddha meets with and engages in a dialog with an interlocutor who is explicitly

identified as a Brahman, thus setting up an opposition between the identity Brahman,

which the interlocutor possesses, and the Buddha, who is not so identified. 709

As such, encounter dialogs serve, on the one hand, as the most powerful and

visible piece of evidence in favor of a model of Buddhism arising in antagonistic

opposition to Brahmanism, insofar as they are read synchronically as evidence of the

early Buddhist attitude toward Brahmanism. 710 On the other hand, they provide a

serious obstacle to Bronkhorsts theory that Buddhism arose within the religio-

purported evidence of the antagonism between early Buddhism and Brahmanism, are not simply
dialogs, but debates; that is, the Buddha encounters a Brahman and debates with him over the tenets
advanced by the new Brahmanism the vara system, Brahmanical superiority, the importance of
knowledge of the Vedas and ritual, etc. In other cases, however, the dialog is simply a dialogas, for
example, when a Brahman such as Jussoi approaches the Buddha and simply asks a question,
which the Buddha answers by giving a sermon. Even in these latter cases, however, the juxtaposition
of an interlocutor identified as a Brahman next to the Buddha who is not so identified serves to
establish a Buddhist identity as separate from that of the Brahmans.
709
There are some cases of encounter dialogs in which the Buddha refers to himself, or what
he considers the ideal person, as a Brahman, but the narrative structure of these texts qua encounter
dialogs sets up a dichotomy between the Brahman interlocutor on the one hand and the Buddha, and by
extension his followers, on the otherthus rendering any reference to the Buddhist ideal person as a
Brahman figurative and polemical. A good example of such a text is the Vseha Sutta (Sn. 3.9 and
MN 98), in which the Brahmans Vseha and Bhradvja ask the Buddha whether one becomes a
Brahman by birth or conduct and the Buddha responds with a long string of verses each ending with
the refrain tam aha brmi brhmaa that were likely borrowed from the Dhammapada (v. 396-423;
see Section IV.4.5). Although the Buddhas response in these verses defines a Brahman in terms of
conduct and not in terms of birth, the narrative frame identifies Vseha and Bhradvja, i.e. two
Brahmans by birth, as Brahmans, and not the Buddha as such, thus rendering his reply in the verses
reactive and polemical. In Part III we saw the exact same effect when narrative frames were added to
verses that referred to the ideal person as a Brahman; we can therefore say that in many cases the
narrative frames that were added to such verses transformed them into encounter dialogs.
710
Some recent scholarship has rightly taken notice of the fact that the early Buddhist attitude
toward Brahmans, even when read synchronically from the early stras, was not uniformly
antagonisticsee Freiberger, The Ideal Sacrifice and Black, Rivals and Benefactors. Nevertheless,
there has been little in the way of efforts to read the early Buddhist evidence diachronically, with the
partial exception of Bronkhorst, insofar as he implies that Buddhist texts referring to Brahmans must
be late. (See the next note.)

360
cultural complex of Greater Magadha, in isolation from the Vedic Brahmanical

culture found further to the west. 711 I have argued, in distinction to both these

theories, that the Buddhist articulation of its own self-identity, in particular vis--vis

the proponents of the new Brahmanism and in the context of the early Buddhist

texts, must be interpreted diachronically, and moreover that the earliest articulations

of Buddhist identity vis--vis the category brhmaa for which we have evidence

evince no sense of dichotomy between a Buddhist self and a Brahmanical other.

In Part IV, I will explore the development of the encounter dialog genre, which by

its very structure systematized, routinized, and normalized a Buddhist identity as

distinct from the Brahman. I will argue that through their implicit acceptance of the

arrogation of the category brhmaa by the proponents of the new Brahmanism to

themselves, the encounter dialogs represented a de facto cession of the category by

the Buddhists, even as the content of the encounter dialogs fought against the self-

proclaimed Brahmans claim to it.

The structure of the encounter dialog genre, and its distribution across the

various recensions of the Nikya/gama traditions as they have come down to us, is

711
Bronkhorst does attempt to deal with this problem to a certain extent in a short appendix to
his Greater Magadha (Appendix VI: Brahmins in the Buddhist Canon, 353-6), in which he points to
pieces of evidence that certain early Buddhist texts that refer to Brahmans (all of which I would
categorize as encounter dialogs) are relatively late. Unfortunately, Bronkhorst does not provide an
extensive enough treatment of Buddhist texts that refer to Brahmans to explain them away for the
sake of his theory in Greater Magadha. Nevertheless, my objection to Bronkhorsts theory has less to
with the arguments he presents in this appendixsince, as will become apparent in Part IV, I would
agree that the encounter dialog genre is the product of an extended period of developmentand more
to do with the fact that Bronkhorst has no explanation for the utter pervasiveness of Brahmanical
terminology, especially the category brhmaa, throughout the early Buddhist texts, including very
early texts such as the Ahaka, and the fact that he introduces a sharp dichotomy between
Brahmanism and the religion of Greater Magadha that, I believe, is both methodologically
problematic and not borne out by the evidence.

361
quite complicated, and questions concerning the development of the genre are

inextricably tied up with both the oral nature of the Nikya/gama traditions and

their internal structure. Part IV is therefore divided into five chapters, which progress

from discussions of general methodological issues concerning the study of early

Buddhist texts to more specific investigations relevant to the development of the

encounter dialog genre and the thesis I am advancing in Part IV. In the first and

second chapters, I address the general issues involved with the study of the

development of any textual theme or genre in the early Buddhist texts. Thus, in

Chapter IV.1, I begin with an exploration of issues involved with orality and the study

of early Buddhist texts as specifically oral texts. These early Buddhist texts, as they

are preserved both in the Pali Canon from the Theravda tradition and in Chinese

translations from various other sectarian traditions, were composed and transmitted

orally, and as such they are not amenable to the analytical tools of classical textual

criticism, but must be submitted to methods developed, beginning with Milman Parry

and Albert Lord in the 1920s, specifically for use with oral traditions. Then, in

Chapter IV.2, I turn to the still relatively nascent field of comparative Nikya/gama

studies, which seeks to trace the development of the early Buddhist oral tradition by

comparing the various snapshots of the early Buddhist oral tradition(s) that have

come down to us in the form of the Pali Nikyas and the Chinese gamas. I focus in

particular on the theory of the late Chinese scholar-monk Ynshn, which provides a

framework for understanding the development of the Nikya/gama tradition as

expansions out of the simple, early structure of the Sayuktgama.

362
 Next, in Chapters IV.3 and IV.4, I turn to a specific examination of the

encounter dialog genre within the context of the Nikya/gama tradition. First, in

Chapter IV.3, I examine the structure according to which encounter dialogs were

stored within the early Buddhist oral traditionthat is, the collections within which

encounter dialogs were gathered together and transmitted. I show that, although in

some cases these collections were post-hoc, developed by a specific tradition for the

sake of ease of transmission, at least two such collections are found across all

traditions and bear internal evidence that their component stras were composed

together, as a unit. Then, in Chapter IV.4, I apply the concepts of formula and

theme from the Parry-Lord oral theory to the encounter dialog genre within the

early Buddhist oral tradition in order to trace developments in that tradition that cross

the boundaries of Nikya/gamas, encounter dialog collections, and individual stras.

Finally, in Chapter IV.5, I bring together my findings from the first four

chapters in order to paint a broad picture of the development of the encounter dialog

genre. I argue that the genre had its origin in a collection of stras that were

constructed by adding short narratives, in which the Buddha encounters a Bhradvja

Brahman with a funny/meaningful name, to verses. This model was to a certain

extent imitated, but greatly expanded upon, by a collection of stras that would

become the basis of the Dgha Nikya/Drghgama, in which a more highly

developed use of formulae and themes developed. From there, I argue, many of those

formulae and themes became diffused across the early Buddhist tradition as

characteristic markers of the encounter dialog genre.

363
 Chapter IV.1

Orality and the Early Buddhist Texts

IV.1.1 The Oral Theory of Milman Parry and Albert Lord

Modern-day theory on oral traditions has its roots in the work, during the late

1920s and 1930s, of Milman Parry, a classicist who specialized in ancient Greek

epic. 712 Parrys work was concerned with the so-called Homeric Questionthe

observation, found occasionally in ancient and medieval works, but not pursued in

earnest until the 19th century, that Homers epics are qualitatively different from later

literature, insofar as they are characterized by the use of repetitive language and

clichd expressions. Those who, in the 19th century, sought to solve the Homeric

Question were looking for a way to explain how the Iliad and Odyssey came into

being, and in doing so, they generally fell into two schools of thoughtAnalysts

believed that the ancient Greek epics were composite texts, and so they attempted to

analyze them into strata in order to identify the original text, while Unitarians

712
A useful overview of Milman Parrys life and work can be found in the introduction to his
collected papers, edited and published posthumously by his son AdamAdam Parry, ed., The Making
of Homeric Verse: The Collected Papers of Milman Parry (Oxford: Clarendon Press, 1972), ix-lxii.

364
rejected this approach and argued that the epics were of unitary authorship. Both of

these camps, however, based their view on the assumption that one could speak of the

Iliad and Odyssey, in ancient times, as stable texts; they simply disagreed as to

whether the texts that have come down to us under those names represent the

original texts or not. Milman Parry rendered the entire Analyst-Unitarian debate

obsolete, and effectively solved the Homeric Question, by abandoning this

assumption, and arguing instead that the Homeric epics were the product of an oral

tradition, a process of development whereby an epic story was told over and over

again, without the benefit of the technologies of literacy, by bards who did not simply

recite the epics word-for-word from memory, but literally retold them each time, thus

introducing novel features with each new bard and each new performance. 713

The way in which the ancient bards were able to accomplish this impressive

feat (the Iliad, at least in the form of the performance that was recorded in writing, is

over 15,000 lines long, and the Odyssey, over 12,000), Parry argued, was through the

use of was he called formulaei.e., the very stereotyped expressions that had

instigated the Homeric Question in the first place. The use of formulae allowed the

ancient bards to propel the narrative forward while still maintaining meter. In order

to substantiate this theory, Parry set out to what was then Yugoslavia in order to

713
See ibid., x-xxi, and John Miles Foley, Introduction: The Oral Theory in Context, in
Oral Traditional Literature: A Festschrift For Albert Bates Lord, ed. John Miles Foley (Columbus,
Ohio: Slavica Publishers, 1981), 28-32. Foleys introduction to Oral Traditional Literature (p. 27-122)
provides a useful overview of the field of oral theory up to the date of publication, although it focuses
primarily on work on Old English poetry. A more comprehensive and slightly more up-to-date
overview of the field can be found the Introduction (p. 3-77) to John Miles Foley, Oral-Formulaic
Theory and Research: An Introduction and Annotated Bibliography (New York: Garland Publishing,
1985).

365
record a living epic tradition, in which, he found, bards indeed give extemporaneous

performances of verse epics, to the order of thousands of lines, in the process making

use of stock expressions like the formulae found in Homer. Unfortunately, however,

Parry was accidentally shot and killed shortly after returning from Yugoslavia in 1935,

when he was only 33 years old. 714 His comparative work, however, was continued by

his student and field assistant Albert Bates Lord, who finally published a full, book-

length treatment of the Oral Theory he and his late mentor had pioneered, The Singer

of Tales, in 1964. 715 The publication of this work instigated research into the oral

features of literature in not only Ancient Greek and modern Serbo-Croation, but also

Old English, Old French, and a host of other languagesincluding, as we shall see,

Middle Indic.

Since we will be referring to them repeatedly in Part IV in application to the

early Buddhist texts, it will be useful to define here the key technical terms that have

been used by Parry, Lord, and their followers in articulating and applying the Oral

Theory. The first and most fundamental of these, of course, is the concept of the

formula. Parry himself, in an oft-quoted passage, defined the formula, within the

Homeric context at least, as a group of words which is regularly employed under the

same metrical conditions to express a given essential idea. 716 Parry, as well as Lord

and their followers after him, studied verse epic, and so his definition of the formula

714
Parry, Making of Homeric Verse, ix, xli.
715
Albert B. Lord, The Singer of Tales (Cambridge, Mass.: Harvard University Press, 1964).
716
Parry, Making of Homeric Verse, 272, italics in the original. Albert Lord quotes this
definition at the beginning of his chapter on The Formula in The Singer of Tales, 30.

366
was conditioned by the particular restrictions imposed by extemporaneous metrical

composition. The work of Parry and Lord showed that oral epic poetry is possible

precisely because bards have memorized a whole set of stock expressions that they

can draw from at will when they want to express a particular idea at a particular point

in the meter. Oral poets of this type do not have the luxury of pondering what to say

next, reviewing what they have already said, or revising what they have already said

in the same way as do literate authors, and formulae enable them to retell a particular

story with a degree of consistency, with the speed demanded by oral performance,

and in the form demanded by meter.

The second key technical term associated with the Oral Theory is the theme.

Indeed, it is themes, together with meter, that determine what particular formulae can

be chosen from for use at a particular point in an oral epic performance. The theme,

as Lord defines it, is a group of ideas regularly used in telling a tale in the formulaic

style of traditional song. 717 As an example, Lord cites the theme of the council,

which he calls one of the most common and most useful themes in all epic

poetry. 718 In this particular example, at any relevant point in an epic performance, a

king may summon his councilors together for advice; since this can be a rather

frequent occurrence even in a single epic performance, the epic poet describes

councils each time they arise in a similar way, and thus the council becomes a

particular theme in his poetry. In this way, themes drive the epic performance on

717
Lord, Singer of Tales, 68.
718
Ibid.

367
the narrative level in a manner analogous to the way in which formulae drive it on the

level of the verse. Lord is at pains to point out, however, that unlike the formula,

[t]he theme, even though it is verbal, is not any fixed set of words, but a grouping of

ideas. He does concede, though, that [s]ome singers do not change their

wording much from one singing to another, especially if the song is one that they sing

often. 719

The third and final technical term that is associated with the Parry-Lord

Theory is tradition. Lord defines an oral tradition as the body of formulas, themes,

and songs that have existed in the repertories of singers or story tellers in a given area

over usually a long period of time. Thus, the Parry-Lord Theory argues that Homer

neither, on the one hand, composed the Iliad and Odyssey in toto from scratch by the

workings of creative genius, nor, on the other, did he compile them piecemeal by

borrowing from earlier works. Instead, the texts of the Iliad and Odyssey are

recordings of particular performances of those epics made by a particularly renowned

bard, presumably named Homer, who was trained and operated within an oral

tradition that likely went back many generations before him. The comparison to

Serbo-Croation epic poetry instigated by Parry and carried out by Lord was intended

to demonstrate the existence of such a living oral tradition and to show that the poetry

it produces has features in common with the Homeric epics. In a sense, the concept

of tradition is the most central concept of the Parry-Lord Oral Theory because it

revolutionized the study of Homeric epic and other literatures that came to be

719
Ibid., 69.

368
recognized as products of an oral tradition. No longer could such texts be studied

in the same way as written texts, as if they were either written down or compiled at a

single moment in history; instead, they must be understood as products of constant

telling and retelling over the course of many generations, conditioned by the

commonalities introduced by shared themes and formulae, but with a degree of

uniqueness to each performance. As Lord writes, Forcing traditional literature,

which is traditional by its origin and nature, into the straight-jacket of synchronic

observation is to distort it beyond recognition. 720

IV.1.2 Applying the Oral Theory to the Early Buddhist Texts

Although the study of the early Buddhist texts as products of an oral tradition

is still relatively young, scholarly opinion is nonetheless essentially unanimous that

these early Buddhist texts 721 were indeed originally oral. Mark Allon, the author of

the most comprehensive study to date of the early Buddhist texts as oral literature,

gives five reasons for this scholarly consensus:

1. The Pali Nikyas do not refer to writing, but do refer frequently to

learning and reciting suttas.

2. Although a few passages in the Pali Vinaya refer to writing, they do not

refer to the use of writing to create and preserve texts.

720
Albert B. Lord, Perspectives on Recent Work on the Oral Traditional Formula, Oral
Tradition 1, no. 3 (1986), 468.
721
By early Buddhist texts I mean the contents of the Pali Tipiaka and analogous texts that
have been preserved in Sanskrit, Chinese, and Tibetanthat is, most of the pre-Mahyna literature.
See further below.

369
 3. The Pali Vinaya has rules governing all possible monastic possessions, but

none for writing materials.

4. There is no archaeological evidence for the use of writing in India prior to

Aoka.

5. The early texts bear stylistic features characteristic of oral tradition. 722

This last reason, which refers to the repetitive and formulaic style of the early

Buddhist texts, is immediately obvious to anyone who has ever read one of them,

whether in an Indian dialect or in an ancient or modern translation; it is also the entry

point through which we can apply the framework of the Parry-Lord Oral Theory to

the early Buddhist texts. Allon has already provided a detailed study of the oral

stylistic features of Pali texts, using the Dgha Nikya, and the Udumbarikashanda

Sutta in particular, as a case study; therefore, it is unnecessary to provide a detailed

cataloging of all the oral stylistic features that have been observed in the Pali

literature. Instead, in this chapter I will simply show how the categories of the Parry-

Lord Oral Theory can be applied to the early Buddhist texts, with particular reference

to the encounter dialog genre, to provide a conceptual framework for the ideas I will

be presenting in Part IV. For the sake of brevity, I will limit myself to only a few

illustrative examples taken from the Pali Canon; a fuller treatment of the breadth of

formulae found within the encounter dialog genre, as preserved in various sectarian

traditions, will be given in Chapter IV.4.

722
Mark Allon, Style and Function: A study of the dominant stylistic features of the prose
portions of Pli canonical sutta texts and their mnemonic function (Tokyo: The International Institue
for Buddhist Studies of the International College for Advanced Buddhist Studies, 1997), 1.

370
 Let us begin with the most expansive concept within the Parry-Lord Theory,

that of the tradition. In the context of the early Buddhist texts, oral tradition would

naturally refer to the traditionor rather, traditionsof oral recitation of which the

early Buddhist texts that have come down to us in written form were particular

performances. This would include, of course, the texts of the Pali Canon, but also 723

the gamas, Vinayas, and Abhidharma texts that have been preserved in the Chinese

Tripiaka (), 724 as well as various individual (pre-Mahyna 725) stras and

723
By far the single most useful tool for comparing the different performances of early
Buddhist stras that is available today is the Online Sutta Correspondence Project
[www.suttacentral.net], which was created by Roderick Bucknell and Bhikkhus Anlayo and Sujto.
This website not only allows one to view, through simple hyperlink navigation, the cross-linked
parallels to any given stra found within any of the published collections of texts in eight languages
(Pali, Chinese, Tibetan, Sanskrit, Prakrit, Gndhr, Khotanese, and Uighur) that have been entered
into the websites database; it also allows one to instantaneously view any original source text or
translation into a modern European language that is available for free on the internet.
724
This is not to say that all or even most early Buddhist texts that were translated into
Chinese were translated directly from oral performances; indeed, Fxin, Xunzng, and Yjng all
travelled to India in search of written Buddhist scriptures to bring back for translation. When Chinese
translations of early Buddhist texts were made from written manuscripts, however, those written
manuscripts would have gone back, at some point, to oral performances.
725
In referring to pre-Mahyna stras, I am primarily referring to the scope of my study,
and do not mean to imply a hard-and-fast distinction between oral pre-Mahyna texts and written
Mahyna textsalthough certainly the Mahyna stras did arise at a time when literary technology
was on the rise in India and even the early Buddhist oral traditions were beginning to be written down.
The issue of orality vs. literacy with respect to Mahyna texts is complicated, and the degree to which
literacy played a role in the production and dissemination of Mahyna Buddhist texts is still a matter
of debate. Two of the more well-known arguments in favor of literacy are those of Gombrich, who
argued that writing is what allowed Mahyna stras to be preserved in spite of their utter novely and
incommensurability with the old oral traditional texts, and Schopen, who argued that the rise of the
Mahyna was associated with a cult of the booksee Richard Gombrich, How the Mahyna
Began, Journal of Pali and Buddhist Studies 1 (March 1988): 29-46, and Gregory Schopen, The
Phrase sa pthivpradea caityabhto bhavet in the Vajracchedik: Notes on the Cult of the Book in
Mahyna, Indo-Iranian Journal 17 (1975): 147-81. David Drewes, however, has argued that orality
likely continued to have an important role in the dissemination of Mahyna stras, particularly in the
form of preachers known as dharmabhakasDavid Donald Drewes, Mahyna Stras and Their
Preachers: Rethinking the Nature of a Religious Tradition (PhD diss., University of Virginia, 2006).
Even Drewes, however, speaks of orality with respect to the Mahyna stras as a literate orality
ibid., 17. This is quite different from the early Buddhist oral tradition, which appears to have been
illiterate.

371
fragments that have been preserved in Chinese, 726 Tibetan, 727 Sanskrit, 728

Gndhr, 729 and other languages. 730 Given the limited scope of this study, we are

726
The first two volumes of the Taisho (T. 1-151) are filled with texts that the Taisho editors
considered to be Hnayna stras, classified according to gama. Each of the complete (or nearly
complete) translations of the four gamas (in the case of the Sayuktgama, there are three, one
almost complete and two incomplete) is followed by translations of various individual stras that the
Taisho editors classified under that gama. Modern scholars do not always agree with these
classifications, however.
727
Interestingly, thirteen Pali texts from the Theravda tradition, most of them suttas, are
found in Tibetan translation in the Kanjura comprehensive study of these texts is found in Peter
Skilling, Theravdin Literature in Tibetan Translation, Journal of the Pali Text Society 19 (1993):
69-201. Most stras of the rvakayna preserved in the Tibetan Kanjur, however, are from the
(Mla)sarvstivda tradition. An edition and study of a particularly important anthology of such stras,
known as the Mahstras, is found in Peter Skilling, Mahstras: Great Discourses of the Buddha,
Sacred Books of the Buddhists, vols. 44 and 46 (Oxford: The Pali Text Society, 1994 and 1997).
Skilling provides a non-comprehensive list of rvaka texts of the Mlasarvstivda school found in
the Kanjur on p. 15-16 of vol. 46. For a list of all Sayuktgama stras found in the Kanjur, see
Andrew Glass, Four Gndhr Sayuktgama Stras: Senior Kharoh Fragment 5, Gandhran
Buddhist Texts 4 (Seattle: University of Washington Press, 2007), 32. These and other non-Mahyna
stras found in the Tibetan Canon are catalogued, with parallels, on the Online Sutta Correspondence
Project website at
http://www.suttacentral.net/disp_division.php?collection_id=3&collection_name=Tibetan.
728
There exists a plethora of fragments of early Buddhist stras written in Sanskrit, most of
them quite short, that have been discovered and published. One of the most extensive collections of
such fragments is the Turfan collection, assembled in Berlin by four German expeditions to Eastern
Turkestan. This collection of over 7000 fragments is still in the process of being published and now
stands at ten volumes: Ernst Waldschmidt et al., eds., Sanskrithandschriften aus den Turfanfunden,
Verzeichnis der Orientalischen Handschriften in Deutschland, vol. 10, parts 1-10 (Stuttgart: Franz
Steiner Verlag, 1965-2008). The most extensive, i.e., least fragmentary, Sanskrit manuscript of an
early Buddhist stra collection found to date is of the Sarvstivda Drghgama, which unfortunately
has not yet been edited and published. A study of this manuscript, which has also, unfortunately, not
been published and is thus not readily available, is found in Jens-Uwe Hartmann, Untersuchungen
zum Drghgama der Sarvstivdins (Habilitationsschrift, Georg-August-Universitt, Gttingen,
1991). Hartmann has, however, published three short articles on this new Drghgama manuscript:
Jens-Uwe Hartmann, Zu einer neuen Handschrift des Drghgama, in Vividharatnakaraaka:
Festschrift fr Adelheid Mette, ed. Christine Chojnacki et al., Indica et Tibetica, vol. 37 (Swisstal-
Odendorf: Indica et Tibetica Verlag, 2000), 359-67; Jens-Uwe Hartmann, Further Remarks on the
New Manuscript of the Drghgama, Journal of the International College for Advanced Buddhist
Studies 5 (2002): 133-50; and Jens-Uwe Hartmann, Contents and Structure of the Drghgama of the
(Mla)-Sarvstivdins, Annual Report of the International Research Institute for Advanced
Buddhology at Soka University 7 (2004): 119-37. A more comprehensive bibliography of the many
published editions of Sanskrit fragments of early Buddhist stras can be found on the Online Sutta
Correspondence Project website at
http://www.suttacentral.net/disp_sutta.php?division_id=24&collection_name=Sanskrit&division_name
=Other%20Sanskrit%20fragments&type=Division&division_acronym=Skt%20frgm.

372
mostly concerned with the texts of the Sutta Piaka in Pali; the four gamas

preserved in Chinese; and certain parallels found in other Pali, Chinese, Sanskrit, and

Tibetan texts.

Unfortunately, we know very little about how exactly the early Buddhist oral

traditions workedthat is, how and when they came into being in the first place 731

and how exactly they originally composed and then transmitted the texts that have

729
One of the most well-known early Buddhist texts preserved in Gndhr, and, until recently,
the only to have been published, is a version of the Dharmapada: John Brough, The Gndhr
Dharmapada, London Oriental Series, vol. 7 (London: Oxford University Press, 1962). A survery of
other early finds of Gndhr manuscripts written in Kharoh script is found in Richard Salomon,
Previous Discoveries of Kharoh Manuscripts, in Ancient Buddhist Scrolls from Gandhra: The
British Library Kharoh Fragments (Seattle: University of Washington Press, 1999), 57-68. This
latter volume also serves as a de facto introduction to the British Library/University of Washington
Early Buddhist Manuscripts Project, which is in the process of studying and publishing fragments of
Gandhran Buddhist texts in the British Library Kharoh collection, and now also the Senior
Kharoh collection as well. The findings of this project are being published in the Gandhran
Buddhist Texts series, edited by Richard Salomon. Volumes within this series of particular relevance
to the study of early Buddhist stras are Richard Salomon, A Gndhr Version of the Rhinoceros
Stra: British Library Kharoh Fragment 5B, Gandhran Buddhist Texts, vol. 1 (Seattle: University
of Washington Press, 2000); Mark Allon, Three Gndhr Ekottarikgama-Type Stras: British
Library Kharoh Fragments 12 and 14, Gandhran Buddhist Texts, vol. 2 (Seattle: University of
Washington Press, 2001); Timothy Lenz, A New Versiton of the Gndhr Dharmapada and a
Collection of Previous-Brith Stories: British Library Kharoh Fragments 16 + 25, Gandhran
Buddhist Texts, vol. 3 (Seattle: University of Washington Press, 2003); and Glass, Four Gndhr
Sayuktgama Stras.
730
See, e.g., Mauro Maggi, The Khotanese Karmavibhaga, Serie Orientale Roma, vol. 74
(Roma: Istituto Italiano per il Medio ed Estremo Oriente, 1995), which provides a partial Khotanese
parallel to the Cakammavibhaga Sutta (MN 135), and inasi Tekin, Maitrisimit nom bitig. Die
Uigurische bersetzung eines Werkes der Buddhistischen Vaibhika Schule. Berlin: Akademie
Verlag, 1980), which provides a partial Uighur parallel to the Dakkhivibhaga Sutta (MN 142).
Citations are from the Online Sutta Correspondence Project website.
731
Traditionally, of course, the Buddhist oral tradition began with the first sagti (lit.,
singing together) immediately after the Buddhas death. Modern scholarship, however, has been
skeptical of the historicity of this first sagti at Rjagha (less so, however, of the second, held at
Vail about 100 years later), and the Buddhist traditions themselves are inconsistent as to what
exactly was recited at that first communal recitation. For a translation of all five accounts of the first
two sagtis (from the Theravda, Mahsghika, Sarvstivda, Dharmaguptaka, and Mahsaka
Vinayas), as well as a summary of scholarly opinions on their historicity, see Kkkpalliye Anuruddha
Thera et al., trans., The First and Second Buddhist Councils: Five Versions (Hong Kong: Chi Lin
Nunnery, 2008).

373
come down to us. We do know, mostly from the Pali commentarial tradition, that

such an oral tradition existed, that the agents of that tradition were known as

bhakas (lit., speakers, from bha, to speak), and that these bhakas were

divided according to specialization in a certain collection of texts, such as the

Majjhima Nikya. 732 We also have epigraphical evidence for the existence of a fairly

developed bhaka tradition, apparently antecedent to the written Pali tradition of Sri

Lanka and Southeast Asia, in what is now Madhya Pradesh, dating to the uga

period (2nd century BCE). 733 As we would expect with an oral tradition, no two

performances are exactly the same, 734 as attested by the fact that, for example, a

particular sutta in the Pali Sutta Piaka will very often have a parallel in one of the

Chinese gamas, but that parallel will differ from the Pali, sometimes trivially,

sometimes more substantially. In addition, it would appear that the early Buddhist

oral tradition fragmented at least in part along sectarian lines, insofar as the Chinese

translators made an attempt to identify the sect (nikya) from which particular texts

came, but it is not entirely clear if or to what extent the fragmentation of the early

732
On the bhaka-tradition as described the Pali Ahakaths, see E. W. Adikaram, The
Bhanakas, in Early History of Buddhism in Ceylon, or State of Buddhism in Ceylon as Revealed by
the Pli Commentaries of the 5th Century A.D. (Colombo: M. D. Gunasena and Co., 1946), 24-32,
and Mori Sodo, The origin and the history of the Bhaka tradition, in nanda: Papers on Buddhism
and Indology. A felicitation volume presented to Ananda Weihena Palliya Guruge on his sixtieth
birthday, ed. Y. Karunadasa (Colombo: Felicitation Volume Editorial Committee), 123-9.
733
Lamotte, History of Indian Buddhism, 150[164-5], 414[455].
734
This is at least true for the different performances that have come down to usprimarily
in the form of the Pali texts and their parallels in the Chinese translationsbut these performance are
separated by sect, geography, and several hundred years. It is less clear to what extent performances
differed at a particular place, within a particular sectarian tradition, at a particular point in history.
This will be discussed further in Section IV.1.3.

374
Buddhist oral tradition only followed sectarian lines, as opposed to simply occurring

with time and geographical dispersionespecially given that the nature of the early

Buddhist sects is itself poorly understood.

Within the context of the early Buddhist oral tradition thus defined, Lords

concept of the theme can most often be usefully applied at the level of the stra, since

within this tradition, the stra is the most basic unit at which a group of ideas [is]

regularly used [to tell] a tale in [a] formulaic style. 735 In this respect, the early

Buddhist oral tradition is different from the oral traditions studied by Parry and Lord

insofar as it is not a tradition of oral epic, and therefore the texts performed by the

tradition are usually not long, extended narratives constructed out of a succession of

many themes, but rather short texts each governed by a single theme. 736 The most

important theme with respect to the research being presented in Part IV, of course, is

the theme of the encounter dialog itself. Conceived of as a theme, encounter

dialogs are characterized, most generally, by a meeting between the Buddha and a

Brahman, in which the Buddha addresses the Brahman with the vocative brhmaa

and the Brahman addresses the Buddha with the vocative bho 737 gotama.

735
To cite again from Lord, Singer of Tales, 68.
736
Joy Mann describes three categories of suttasermons, debates, and consultations
but indicates that more than one category can be found in a single sutta: Joy Mann, Categories of
Sutta in the Pli Nikyas and their Implications for our Appreciation of the Buddhist Teaching and
Literature, Journal of the Pali Text Society 15 (1990): 29-87. Thus, for example, a sutta can begin as
a debate and then end as a sermon once the Buddhas debate opponent is defeated.
737
According the PTS Pali-English Dictionary, the term bho is a familiar form of address,
used towards equals or inferiors; therefore, in using this form of address, the proponents of the new
Brahmanism are depicted as addressing the Buddha somewhat haughtily, or at least without much
deference. The Niddesa glosses the term brhmaa on several occasions (where it does not refer to the

375
 It should be noted, however, that themes within the early Buddhist oral

tradition are not mutually exclusive, and thus there are other themes that can be used

to characterize the structure of stras within particular subsets of the encounter dialog

genre, often with more specificity than the more general theme of an encounter

between the Buddha and a Brahman. Perhaps the two most important themes that

characterize subsets of the encounter dialog genre are debates and

consultations, 738 to borrow terms coined by Joy Mannwho did not, however, use

Lords category of theme to describe them. That they can indeed be considered

themes, however, is evident from the fact that, as Mann shows, debates and

consultations are defined, and can be distinguished from one another, by their

introductory and concluding formulas and by their internal structure. 739 Within the

encounter dialog genre, debates are found most often (though not exclusively) in the

Dgha and Majjhima Nikyas; a good example is the Assalyana Sutta (MN 93), in

which a Brahman student is sent by his teachers to debate the Buddha on the tenets of

the new Brahmanism (and loses). Most of the encounter dialogs in which Jussoi,

the friendly Brahman, serves as the Buddhas interlocutor, however, are

consultationsthat is, Jussoi approaches the Buddha to ask him a question about

dhamma, and the Buddha answers his question with a sermon. As themes,

ideal person, but to proponents of the new Brahmanism) as ye keci bhovdik (whoever says
bho)see its commentaries on Sn. 4.10.859, 5.3.1044, 5.7.1079-81.
738
I omit here the third category of sutta (as she puts it) that Mann describesnamely,
sermonssince, as Mann notes, Entire suttas which through their opening and concluding formulas
can be defined as Sermons are preached only to monksMann, Categories of Sutta, 32.
739
Ibid. For a full description of the formulas and internal structure of debates, see p.
44-61; for consultations, see p. 61-68.

376
consultations and debates can be said to run perpendicular to the encounter dialog

theme; that is, they bisect the encounter dialogs into two groups, but also characterize

numerous stras outside of the encounter dialog genre as well.

The third and final major conceptual element of the Parry-Lord Theory, the

formula, unfortunately poses some complications in its application to the case of the

early Buddhist oral tradition. The reason for this is that the early Buddhist tradition,

unlike the ancient Homeric and modern Serbo-Croatian traditions, created texts

primarily in prose. 740 As we have already seen, Parry defined the formula as a group

of words employed under the same metrical conditions 741a condition which

clearly does not apply to the case of the early Buddhist prose tradition. Lord also

appears to have believed that the formulae of his theory are only to be found in verse.

He writes:

one cannot have formulas outside of oral traditional verse, because it is the
function of formulas to make composition easier under the necessities of rapid
composition in performance, and if that necessity no longer exists, one no
longer has formulas. If one discovers repeated phrases in texts known not to
be oral traditional texts, then they should be called repeated phrases rather
than formulas. 742

Note that in this passage, Lord uses the phrases oral traditional verse and oral

traditional texts as if they are synonymous. But as we have already seen, there are

740
Of course, part of the early Buddhist oral tradition did consist of verseboth in
independent gths, such as are found in the Theragth, Thergth, Ahaka, and Pryaa, and in
stras of the geyya type. Unfortunately, little work has been done on the oral featureswhich might
include formulae as classically defined by Parry and Lordof this portion of the early Buddhist
tradition.
741
Parry, Making of Homeric Verse, 272.
742
Albert B. Lord, Perspectives on Recent Work on Oral Literature, in Oral Literature, ed.
Joseph Duggan (New York: Barnes and Noble, 1975), 18; italics in original.

377
several reasons, other than just the existence of repeated passages, that point to the

early Buddhist texts as having come from an oral tradition. Given that these texts are

in prose, must we then refer to the repeated passages simply as repetitions, or must

we acknowledge that the Parry-Lord conception of the formula can fruitfully be

applied to oral prose texts?

I would argue the latterand therefore, following Allon, 743 I do use the word

formula to refer to the repeated passages found in the early Buddhist tradition. In

defense of this choice, it is instructive to look at an example from the encounter

dialogs we are examining in Part IV. In the Ambaha Sutta (DN 3), we find the

following passage used to describe Ambaha, a Brahman student sent by his teacher

Pokkharasti to debate with the Buddha:

Now at that time Pokkharasti had a resident student named Ambaha, a

scholar, a bearer of the mantras, perfected in the three Vedastogether with
their vocabularies and rituals, with their phonology and etymology, and the
oral tradition (itihsa) as a fifthskilled in philology and grammar, not
lacking in the Lokyata and marks of a Great Man . 744

In the immediately following sutta of the Dgha Nikya (4), the Brahman Soadaa,

who, like Ambaha, meets with the Buddha and engages in a debate with him, is

described in almost identical terms:

743
See Allon, Style and Function, 10-14, for Allons argument in favor of the application of
the concept of formula in a non-verse context. Allons argument is somewhat different than my own
here, however, insofar as he questions Lords insistence that formulae are used only in oral texts that
are composed in the course of performance. That is, Allon believes, unlike myself, that texts of the
early Buddhist tradition were not necessarily composed in performance. See Section IV.1.3.
744
tena kho pana samayena brhmaassa pokkharastissa ambaho nma mavo antevs
hoti ajjhyako mantadharo tia vedna prag sanighaukeubhna skkharappabhedna
itihsapacamna padako veyykarao lokyatamahpurisalakkhaesu .

378
 For the Venerable Soadaa is a scholar, a bearer of the mantras, perfected
in the three Vedastogether with their vocabularies and rituals, with their
phonology and etymology, and the oral tradition (itihsa) as a fifthskilled in
philology and grammar, not lacking in the Lokyata and marks of a Great
Man. 745

We can see here that exactly the same words are used in both of these passages,

except at the beginning, where the appropriate name and a lead-in appropriate to the

context are inserted. In additional to these two suttas, we also find the same stock

expression used in the same way in several other texts of the encounter dialog genre,

by inserting an appropriate lead-in with the name of whichever Brahman is being

described. 746

Thus, according to the ordinary (non-technical in the Parry-Lord sense)

meaning of the word, it seems appropriate to use the word formula to describe a

repeated passage such as this one. Just as with a mathematical formula, what we have

here is an expression consisting both of constants (in this case, every word from

ajjhyako to anavayo) and a variable (the introductory phrase). Moreover, for this

very reason, we cannot dismiss this stock expression as merely a repetition. In

another article, Lord writes, The repetition is a phrase repeated to call attention

to a previous occurrence, for an aesthetic or other purpose. Formulas do not point to

other uses of themselves; they do not recall other occurrences. 747 But according to

745
bhava hi soadao ajjhyako, mantadharo, tia vedna prag
sanighaukeubhna skkharappabhedna itihsapacamna padako veyykarao,
lokyatamahpurisalakkhaesu anavayo.
746
DN 5; MN 91, 92, 93, 95, 100; AN 3.59, 3.60, 5.192, Sn. 3.7.
747
Lord, Perspectives on Recent Work on the Oral Traditional Formula, 492; italics in
original.

379
precisely this distinction, the stock expression for describing a Brahman learned in

the Vedas would have to be classified as a formula and not as a repetition, since it

does not call attention to other occurrences of itself. 748 Rather, it is used, in formulaic

fashion, whenever a Brahman is introduced in a text of the early Buddhist tradition, in

order to indicate that he is learned in the Vedas. In this way, it serves, along with

other formulas that we will investigate in more detail in Chapter IV.4, to convey in

concrete language the theme of an encounter between the Buddha and a Brahman.

IV.1.3 Previous Applications of the Parry-Lord Theory to the Early Buddhist

Literature

If formulae were used in the early Buddhist tradition, even though that

tradition produced texts that to a great extent were not in verse, the question then

becomes, what purpose did the formulae serve? Generally speaking, studies that have

taken note of the oral aspects of Pali literature have referred to the mnemonic aid

provided by repeated or stock passages, but this explanation is less precise that that of

the Parry-Lord Theory, which defines the role of the formula more specifically as

helping the bard to improvise versecomplete with a fixed meter. Formulae help the

bard to do this because they not only convey a particular stereotyped idea; they also

748
This is not to say that formulas in the early Buddhist tradition are never repeated for
dramatic effect. At times this happens, especially when conversations are repeated verbatim within a
single text, but even then, the passage in question still can be considered a formula insofar as it is used
in many different texts in order to help convey a particular theme. For a more detailed analysis of the
use of repetition in the Pali tradition, see Mark Allon, The Oral Composition and Transmission of
Early Buddhist Texts, in Indian insights: Buddhism, Brahmanism and Bhakti. Papers from the Annual
Spalding Symposium on Indian Religion, edited by Sute Hamilton and Peter Connolly (London: Luzac
Oriental, 1997), 50-3.

380
fill out a particular piece of the meter, such as, for example, a leading or a following

hemistich. Obviously, this particular problem does not exist when the texts being

performed are not constrained by meter, but there does of course remain the problem

of memory. But this alone does not explain the presence of formulae in the texts

derived from the early Buddhist tradition. On the one hand, one can have pure

improvisationtelling a story simply by telling a story, as every human being with

the capacity for speech does on an everyday basis. Although such storytelling does

involve a degree of formulaic structure, insofar as any particular storyteller follows

certain habitual speech patterns, these patterns in no wise exhibit the same degree of

uniformity as found in, for example, the formulaic description of a Brahman as a

learned Vedic scholar examined above. On the other hand, one can also have pure

memorizationreciting a text from memory. But again, this does not necessitate the

use of formulae, since any text, no matter how complex and unclichd its language,

can be memorized. The texts of the early Buddhist tradition clearly fall somewhere

between the extremes of informal storytelling and strict memorization. Much of the

still fairly small body of scholarly literature on the early Buddhist oral tradition has

either directly or indirectly been devoted to probing where exactly in this middle

ground it lies.

One of the first scholars to suggest applying the Parry-Lord Theory to the

early Buddhist texts was Lance Cousins, in a short article published in 1983.

Although he does not provide many concrete examples in this article, Cousins does

provide an argument for seeing the oral epic traditions studied by Parry and Lord and

381
the early Buddhist tradition as operating according to very similar principles.

Ignoring the issue of the presence or absence of meter, he characterizes the formulae

in oral epic poetry as being used both as an aid to actual performance and to

maintain the continuity and form of the epic tradition, adding, Both of these

features are certainly present in the sutta literature. 749 He then addresses the issue of

consistency across performances of an oral traditional text, noting that, as already

discussed above, no two performances of an oral epic such as those studied by Parry

and Lord are exactly the same; there is a degree of improvisation involved.

Nevertheless, when asked, a bard in such a tradition will say that he has not changed

anything of any consequence:

But, says the historian, only in the second version did the Sultan travel via
Dubrovnik. You have invented this and falsified history. Not so, says the
singer. It is normal for heroes to travel via Dubrovnik. Many songs tell of
this. It is easy to see that such an approach is un-historical. Nevertheless we
should note that it is an extremely traditional and conservative approach. The
important thing is to preserve the matter of tradition. The application of this in
a given situation may vary greatly and should do. The measure of the
experience, talent and versatility of the performer is his capacity so to adapt
his material. 750

Cousins points out that we have good evidence that exactly the same sort of approach

was taken by the purveyors of the early Buddhist oral tradition. To begin with,

It is quite evident that if we compare the Pali recension of the nikayas with
other surviving versions, the differences we find are exactly those we might
expect to discover between different performances of oral works. The titles
tend to change, the location may alter, material is abridged here, expanded
there. Even within the existing canon we find a great deal of this kind of thing.
749
Lance Cousins, Pali Oral Literature, in Buddhist Studies Ancient and Modern, ed. P.
Denwood and A. Piatigorsky (London: Curzon Press, 1983), 1.
750
Ibid., 2.

382
 Indeed the four great nikayas often read as if they were simply different
performances of the same material. Many of the episodes of a composition
such as the Mahparinibbna-sutta are to be found scattered over the other
three nikayas, often more than once. 751

In addition, Cousins cites an often-quoted passage from the Mahparinibbna Sutta

that indicates that early Buddhists thought consciously about what was important and

what not important in considering a particular text as buddhavacana. According to

this passage, authentic texts should come into sutta and agree with vinaya. 752 If

they meet these criteria, they are authentic; if they dont, they are not. Clearly,

incidental matters, such as where the Buddha was residing, to whom he was speaking,

and even in precisely what form he taught the dharma are not of central importance in

the transmission of the Buddhas word. As Cousins puts it,

What is envisaged for sutta is not then a set body of literature, but rather a
traditional pattern of teaching. Authenticity lies not in historical truth although
this is not doubted, but rather in whether something can accord with the
essential structure of the dhamma as a whole. If it cannot, it should be rejected.
If it can, then it is to be accepted as the utterance of the Buddha. 753

On the basis of these two pieces of evidencethe variations across different

recensions of parallel early Buddhist texts and the explicit lack of concern for exact

word-for-word accuracyCousins concludes that the early Buddhist tradition was, as

we would expect from the work of Parry and Lord, somewhat improvisatory, and that

751
Ibid.
752
DN 16: sutte osretabbni, vinaye sandassetabbni. It should be noted here that sutta here
must not refer to the fully developed texts as such, which in the Pali tradition are actually properly
called suttanta (Skt. strnta), or end of the suttas, but to the pithy doctrinal lists that are expounded
in fully developed suttantas.
753
Cousins, Pali Oral Literature, 3.

383
only later did texts become more fixed, most likely in conjunction with the

introduction of writing.

Other scholars, however, have questioned Cousins model of the early

Buddhist oral traditions development. Gombrich, for example, although he is overall

quite supportive of Cousins approach to the early Buddhist texts as oral texts, writes,

Where I slightly differ from Cousins, as will appear, is in his stress on the
probable improvisatory element in early recitations of the Buddha's
preachings. The whole purpose of the enterprise (as certainly Cousins would
agree) was to preserve the Buddha's words. I think the earliest Pali texts may
well be rather like the Rajasthani folk epic studied and described by John
Smith, in which the essential kernel is in fact preserved verbatim, but
variously wrapped up in a package of conventional verbiage which can
change with each performance. 754

Mark Allon, following Gombrich, has also expressed reservations about Cousins

emphasis on the improvisatory nature of the early Buddhist oral tradition, focusing in

particular on the existence of long repeated passages in the Pali texts. He argues that

it is difficult to see the gross forms of repetition just discussed the

repetition of whole passages, with or without modification, and the repetition
of structures with the replacement of various proportions of their wording
and the scale on which this is pursued, that is the proportion of the text
involved, as anything other than proof, or at least as a very strong indication,
that these texts were designed to be memorised and transmitted verbatim.

Likewise, Anlayo Bhikkhu, in his recent comparative study of the Majjhima Nikya,

argues that the oral characteristics of the Pli discourses testify to the importance of

verbatim repetition in the early Buddhist oral tradition. He sets this in contrast to

other oral traditions, such as oral epic, whose task is to present the main elements of

a tale in such a way as to best entertain the audience. This is different, he contends,

754
Gombrich, How the Mahyna Began, 22.

384
from the purpose of the early Buddhist tradition, which was the preservation of

sacred material, for which free improvisation is inappropriate. 755

To my mind, however, none of these three scholars, in spite of their

significant contributions otherwise to the study of orality in the early Buddhist

tradition, poses a coherent criticism of Cousins original argument that there was an

improvisatory stage in the early Buddhist tradition. In each case, in fact, they

contradict their own assertions of a lack of improvisation almost immediately after

making them. This is most obvious in the case of Gombrich, who in the passage just

quoted above, just after questioning Cousins emphasis on improvisation, compares

the early Buddhist tradition to another oral tradition in which the essential kernel is

in fact preserved verbatim, but variously wrapped up in a package of conventional

verbiage which can change with each performance. 756 It is not clear to me at all how

this differs from Cousins own argument, nor in fact how such a process cannot be

called improvisation. Likewise, shortly after making his own argument in favor of

the early Buddhist texts as having been designed to be memorized, Allon hedges and

admits the possibility that improvisation did in fact take place:

Although I have attempted to show that the early Buddhist sutta texts were, in
the words of R. Gombrich, deliberate compositions which were then
committed to memory, I would certainly agree that accounts of what the
Buddha is supposed to have said and discourses on his teaching would have
been given by the monks and nuns after the Buddhas death in an
improvisatory manner, at times drawing heavily on memorised material, or as
R. Gethin (1992) has argued, by using lists as a foundation. Such discourses
may then have become the basis of later fixed texts. But these accounts and
755
Anlayo, A Comparative Study of the Majjhima-nikya, vol. 1 (Taipei: Dharma Drum
Publishing, 2011), 17.
756
Gombrich, How the Mahyna Began, 22.

385
 discourses were fundamentally different from the essentially fixed, memorised
texts transmitted by the community, however imperfectly.

Again, I do not see how this is different from Gethins argument that there was an

improvisatory period in the early Buddhist tradition, followed by an increasing trend

toward memorization.

Perhaps most surprising is Bhikkhu Anlayos characterization of the early

Buddhist oral tradition as one of verbatim repetition, given that in the very same

study he documents extensively and in great detail the ways in which different

versions of Majjhima Nikya texts differ from one another in terms of wording, order

of presentation, etc. 757 Indeed, in the conclusion to that study, Anlayo contrasts the

method of early Buddhist oral transmission with that of the Vedas on the basis that

the purveyors of the former would have understood what they were memorizing,

while those of the latter in many (most?) cases would not. He notes that oftentimes

not knowing the meaning of what you are memorizing results in more exact recall

than when you do understand what you are memorizing. As examples of the sort of

changes that are introduced into an oral text when the latter is the case, Anlayo

provides numerous examples of parallel textsboth within the same tradition (Pali)

and across traditions (Pali and Chinese versions)in which a particular doctrinal list

757
To be fair to Anlayo, in the immediate context of his characterization of the early
Buddhist tradition as one of verbatim repetition, he points to a particular type of difference between
versions of a text that does point to memorization, namely, transmission errors, where in otherwise
closely similar Pli and Sanskrit passages the counterpart to a particular term shows close phonetic
similarity but has a considerably different meaning (Anlayo, Comparative Study of the Majjhima-
nikya, vol. 1, 17). Such errors could have arisen at a time when exact memorization was valued, but
the original meaning of the passage in question was lost, so simple errors were easily propogated. The
vast majority of the differences between versions that Anlayo documents point, as I will argue,
however, to mnemonic techniques intended to make verbatim memorization unnecessary.

386
is given in a different order in one than in the other. 758 Such variation is in fact seen

quite frequently, and it is not limited to simple lists; as I will show shortly, the entire

order of events in a particular text can vary from one version to the next. While I

certainly appreciate Anlayos (as well as Allons) emphasis on the role that formulae

played in the early Buddhist as a stabilizing force on the tradition, and the desire of

Buddhists from a very early date to preserve the Buddhas words faithfully, it strains

the definition of the word verbatim to use it in conjunction with a tradition that

demonstrates so much variation.

This is not simply a matter of semantics, however. One could assume, as

Anlayo seems to imply, that early Buddhist texts were memorized, and then, because

the monks who memorized them understood what they were memorizing, they

unconsciously, and quite accidentally, introduced changes into the text. This would

be akin to memorizing, for example, the line, The rain in Spain stays mainly in the

plain, but then recalling it as, The rain in Spain falls mainly on the plainwhich

means basically the same thing, but is still slightly different. While such a process

certainly played a role within the early Buddhist tradition (as it would, presumably, in

any oral tradition involving recall, including oral epic), what I am suggesting is

somewhat different. I am suggesting, following Cousins suggestion that we apply

the model developed by Parry and Lord to the early Buddhist materials, that the use

758
Anlayo, Comparative Study of the Majjhima-nikya, vol. 2, 867-77. Many other
examples of jumbled list-orders are given throughout the two volumes, in the commentaries on the
relevant texts of the Majjhima Nikya.

387
of formulae in the early Buddhist tradition, much like in oral epic traditions, made it

unnecessary to memorize much more beyond the formulae themselves.

It will be helpful to illustrate this with an example from the encounter dialog

genre. The Tevijja Sutta (DN 13) of the Theravda tradition has a parallel preserved

in the Chinese Canon (, D 26), which most likely comes from the

Dharmaguptaka tradition. Even though the Theravda and Dharmaguptaka sects are

quite closely related, both belonging to the Vibhajyavda branch of the

Sthaviravda, 759 there are key differences between their two representative

performances of this particular stra. The overall structure of both versions is the

same; each can be divided roughly into nine parts. First, two Brahmans named

Vseha and Bhradvja, as well as their disagreement over the correct path to

Brahm, are introduced. Second, these two Brahmans approach the Buddha to settle

their dispute. Third, the Buddha asks a series of questions about whether any

Brahman has ever seen Brahm face-to-face. Fourth, the Buddha presents a series of

similes to show how ridiculous it is to claim to know the way to Brahm when you

yourself have never seen him. Fifth, the Buddha asks a series of questions about the

characteristics of Brahmans and Brahm, showing that they are not commensurate

with one another. Sixth, the Brahmans ask if the Buddha knows the way to Brahm;

the Buddha uses a simile to say yes, and he agrees to teach the way. Seventh, the

759
In other words, the Dharmaguptakas are more closely related to the Theravdins than not
only the Mahsghikas, but even the Sarvstivdins. This close relationship is reflected by the fact
that the contents of the Chinese Drghgama are more similar to the contents of the Dgha Nikya than
any of the other Chinese gamastwo of which are Sarvstivdin (M and S) and one of which may
be Mahsghika (E)are to their corresponding Pali Nikyas. On the sectarian affiliations of the
Chinese gamas, see Chapter IV.2.

388
Buddha teaches the way to Brahm using a modified version of the long formulaic

teaching found in the Smaaphala Sutta. Eighth, the Buddha asks a series of

questions about the characteristics of a person who follows this path and of Brahm,

showing that they are indeed commensurate with one another. And finally, in the

ninth section, there is a conclusion describing the Brahmans reaction to the Buddhas

teaching.

Since both versions of this stra follow this basic structure, we can imagine an

early performer of it as having memorized the nine or so parts, according to a

mnemonic device now lost. The exact words used, formulae utilized, order of ideas

presented, and in some cases actual content within each of these nine parts differs

substantially between the two versions, however, indicating that there was

considerable freedom in the articulation of an early Buddhist stra at the local level

of the narrative. This freedom becomes immediately apparent at the very beginning

of the first section of the text. According to the nidna of the Pali version, the events

recorded in the text took place in the Mango Grove on the Aciravat River north of

Manaskaa in Kosala. According to the Chinese version, on the other hand, it took

place in a forest near Icchnakala in Kosala. In both versions, the Brahmans

Vseha and Bhradvja are then introduced, but they are introduced in different

ways. In the Chinese version, Vseha is first introduced together with his teacher

Pokkharasti, and then Bhradvja is introduced together with his teacher Trukkha;

in both cases, the formula used for describing a Brahman learned in the three Vedas,

already discussed above in Section IV.1.2, is used. This formula is not used in the

389
Pali version, however. Instead, a different formula, giving a list of famous Brahmans,

including Pokkharasti and Trukkha, is used, 760 and then Vseha and Bhradvja

are simply said, without any particular introduction to themselves personally, to be

walking together down a road. What follows is quite similar across the two versions.

Vseha says that the way to Brahm taught to him by Pokkharasti is right;

Bhradvja says the way taught to him by Trukkha is right; and Vseha uses a

formulaic description of the fame of Gotama 761 to suggest to Bhradvja that they

go to the Buddha to settle their dispute.

The second and third sections are similar in the two versions, but with some

key differences in presentation. In both versions of the second section, Vseha and

Bhradvja approach the Buddha and present their conundrum, but how they do so

differs in the two versions. In the Pali, Vseha and Bhradvja describe their

individual positions in turn; the Buddha repeats what they said; and then Vseha

suggests that perhaps all paths taught by Brahman teachers lead to Brahm, just as

multiple paths can lead to a single village. In the Chinese, however, Vseha and

Bhradvja do not present their positions at all; instead, the Buddha himself reads

their minds, tells them what they have been discussing, and then asks them to confirm

that he is correct. This is followed by Vsehas use of the comparison to paths

leading to a village, as in the Pali. In both versions of the third section, the Buddha

760
tena kho pana samayena sambahul abhit abhit brhmaamahsl manaskae
paivasanti, seyyathida cak brhmao trukkho brhmao pokkharasti brhmao jusoi
brhmao todeyyo brhmao ae ca abhit abhit brhmaamahsl. Versions of this
formula can also be found at MN 98=Sn. 3.9 and MN 99.
761
See Chapter IV.4.

390
then asks Vseha twice to confirm his suggestion that all paths could lead to Brahm.

He then asks him a series of questions about whether anyone has seen Brahm face-

to-face. In the Chinese version, this consists of asking whether any Brahman,

Brahmans teacher, or i has seen Brahm. In the Pali version, however, he also asks

(after the Brahmans teacher and before the is) whether any Brahmans teachers

pupil or any Brahman up to seven generations in the past has seen him. In both

versions, the conclusion is the same: The Buddha declares that since no Brahman has

ever seen Brahm face-to-face, their claims to know the path to him must not be true.

In the Pali version, however, he adds that the Brahmans are like a string of blind men

trying to lead one another to a certain destination.

Sections four and five differ mostly in the order in which ideas are presented

in the two versions. In Section 4, the Buddha uses a series of similes to illustrate the

ridiculousness of claiming to know the way to Brahm when you have never seen him

yourself. These similes are presented formulaically, with a similar structure of

questions and answers used for each simile. In the Pali version, the first simile is to

the sun and moon, which Brahmans cannot reach even though they worship them.

The second comparison is to a man who desires the most beautiful woman in the land,

even though he has never seen her and knows nothing about her. The third is to

building a staircase to a mansion that does not yet exist, and whose orientation is not

yet known. Finally, in the fourth simile, the Buddha compares the Brahmans to a

man trying to cross the Aciravat when it is flooded, which in turn he likens to the

five hindrances. In the Chinese, the order of presentation is different. The first simile

391
is to the beautiful woman, the second to the sun and moon, and the third to the

staircase and the mansion. The final simile is to crossing a flooded river, which is

likened to the five hindrances, as in the Pali, but this simile is expanded somewhat in

the Chinese version. First, the Buddha compares the Brahmans to someone trying to

cross the river by beckoning the other side to come to him; then he compares them to

one who does nothing to try to cross it; finally, he compares them to someone who

uses effort to cross it.

Section five is characterized by a similar difference in the order in which ideas

are represented in the two versions. In this section, the Buddha asks whether Brahm

and Brahmans possess any of five characteristics. The ultimate point is that what is

true of Brahm is not true of Brahmans and vice versa; therefore, they are not

commensurate with one another and Brahmans are not fit to attain union with Brahm.

In the Pali, the characterstics are asked in this order: Do they have possessions

(sapariggaha)? Are their minds full of anger (saveracitta)? Are their minds full of

ill-will (sabypajjacitta)? Are their minds tarnished (sakilihacitta)? Do they

exercise self-mastery (vasavatt)? In the Chinese, however, the order (as well as

possibly the characteristics themselves 762) is different: Do they have ill-will ()?

Do they have anger ()? Do they have resentment ()? Do they have family

and property ()? Do they exercise self-mastery ()? In addition to

this difference of order, the structure with which the Buddha makes the comparison
762
It is not entirely clear what Indic words the first three characteristics listed in the Chinese
version are translating, nor whether they can be exactly correlated to characteristics in the Pali version.
Even if they can, however, the order is clearly different since , which clearly correlates to
sapariggaha in the Pali, is found in the fourth instead of the first position.

392
between Brahmans and Brahm in the two versions is different. In the Pali, he first

asks about each characteristic with respect to Brahm; then he asks them again with

respect with Brahmans; and finally he asks if the two are commensurate with one

another with respect to each characteristic. In the Chinese, however, he simply asks

about each characteristic once, with respect to both Brahm and Brahmans at the

same time.

Sections six, seven, and eight are roughly the same in the two versions. In

Section 6, the Brahmans ask the Buddha if he knows the way to Brahm, and he

replies by comparing himself to a native of Manaskaa who will naturally know the

way to his own village. The Brahmans ask him to teach the way, and he agrees. The

Buddhas teaching of the way to Brahm is found in Section 7, which consists

primarily of a very long formula on the training, which is taken from the

Smaaphala Sutta and its parallel in the Chinese Drghgama, respectively. 763

This formula differs from its typical presentation only at the end, where instead of

describing the attainment of Awakening, the Buddha describes the four

brahmavihras, the attainment of which he equates with attaining union with Brahm.

Then, in Section 8, the Buddha again returns to the five characteristics asked in

Section 5 about Brahm and the Brahmans, except that now he asks them about the

person who attains the brahmavihras, and he shows that such a person is indeed

763
Although the Tevijja Sutta and the Snmng Jng use this formula in the same way, the
actual content of the formula differs in the two traditions. For a detailed comparison of this formula in
all available versions, see Konrad Meisig, Das rmayaphala-Stra: Synoptische bersetzung und
Glossar der chinesischen Fassungen verglichen mit dem Sanskrit und Pli (Wiesbaden: Otto
Harrassowitz, 1987), 39-52.

393
commensurate with Brahm and thus can be said to have attained union with him.

The order in which these five characteristics are addressed differs in the two versions,

following the orders found already in Section 5.

Finally, the conclusion to the stra found in Section 9 offers different

accounts of how Vseha and Bhradvja reacted to the Buddhas teaching. In the

Pali version, they respond with a long formula, found in many suttas of the Pali

Canon, in which they express their pleasure at the clarity of the Buddhas teaching,

which ends with them asking to become the Buddhas disciples. 764 In the Chinese

version, however, they give rise to the Dharma-eye () and are simply said to be

pleased with the Buddhas words ().

Tevijja Sutta (Pali DN 13) (Chinese D 26)

Section One
Nidna: Mango Grove on Aciravat River
north of Manaskaa in Kosala
Famous Brahmans formula
Conversation between Vseha and
Bhradvja on the true path
Vseha praises path of Pokkharasti
Bhradvja praises path of Trukkha
Vseha delivers Fame of Gotama
formula to Bhradvja and suggests
going to the Buddha
Introduction
Nidna: forest near Icchnakala in
Kosala
Introduction to Pokkharasti and
Vseha using Triple Veda formula
Introduction to Trukkha and Bhradvja
using Triple Veda formula
Vseha says his way is better
Bhradvja says his way is better
Vseha delivers Fame of Gotama
formula to Bhradvja and suggests
going to the Buddha

Section Two Introduction of Dispute to the Buddha

764
abhikkanta, bho gotama, abhikkanta, bho gotama! seyyathpi, bho gotama, nikkujjita
v ukkujjeyya, paicchanna v vivareyya, mhassa v magga cikkheyya, andhakre v
telapajjota dhreyya cakkhumanto rpni dakkhant ti. evam eva bhot gotamena anekapariyyena
dhammo paksito. ete maya bhavanta gotama saraa gacchma, dhamma ca bhikkhusagha
ca. upsake no bhava gotamo dhretu ajjatagge pupete saraa gate ti.

394
Vseha and Bhradvja approach
Buddha and exchange greetings
Vseha announces problem and gives
his opinion (pro-Pokkharasti)
Bhradvja gives his opinion (pro-
Trukkha)
Buddha repeats what they just said
Buddha asks what exactly the problem is
Vseha suggests that all paths lead to
Brahm, like different paths to a single
village
Section Three
Buddha confirms twice that Vseha is
suggesting that all paths could lead to
Brahm
Has any Brahman seen Brahm? No.
Has any Brahmans teacher seen
Brahm? No.
Has any Brahmans teachers pupil seen
Brahm? No.
Has any Brahman up to 7 generations in
the past seen Brahm? No.
Has any i (using i formula) seen
Brahm? No.
Buddha concludes that the Brahmans
claims to know the way to Brahm are
not true.
Buddha compares Brahmans to blind
leading blind.
Section Four
Comparison to worshipping sun and
moon
Comparison to desiring hypothetical
most beautiful woman in the land
Comparison to building staircase to
mansion that has not yet been planned or
built
Comparison to crossing flooded
Vseha and Bhradvja approach
Buddha and exchange greetings
Buddha recounts their debate and asks if
it truly happened that way
The Brahmans confirm the Buddhas
account and ask who is right
Buddha asks what exactly the problem is
Vseha suggests that all paths lead to
Brahm, like different paths to a single
village
Questions about Who Has Seen Brahm
Buddha confirms twice that Vseha is
suggesting that all paths could lead to
Brahm
Has any Brahman seen Brahm? No.
Has any Brahmans teacher seen
Brahm? No.
Has any i (using i formula) seen
Brahm? No.
Buddha concludes that the Brahmans
claims to know the way to Brahm are
not true.
Similes
Comparison to desiring hypothetical
most beautiful woman in the land
Comparison to worshipping sun and
moon
Comparison to building staircase to
mansion that has not yet been planned or
built
Teaching on five hindrances and
395
Aciravat, which is compared to five
hindrances
comparison to beckoning the other side
of a river
Comparison to doing nothing to cross a
river
Comparison to using effort to cross river

Section Five Questions about Brahm and Brahmans

Does Brahm have possessions? No.
Does Brahm have mind full of anger?
No.
Does Brahm have mind full of ill-will?
No.
Does Brahm have mind that is
tarnished? No.
Does Brahm have self-mastery? Yes.
Do Brahmans have possessions? Yes.
Do Brahmans have mind full of anger?
Yes.
Do Brahmans have mind full of ill-will?
Yes.
Do Brahmans have mind that is
tarnished? Yes.
Do Brahmans have self-mastery? No.
Is there a likeness between Brahm and Is there a likeness between Brahm and
Brahmans with respect to possessions? Brahmans with respect to ill-will? No.
No.
Is there a likeness between Brahm and Is there a likeness between Brahm and
Brahmans with respect to anger? No. Brahmans with respect to anger? No.
Is there a likeness between Brahm and Is there a likeness between Brahm and
Brahmans with respect to ill-will? No. Brahmans with respect to resentment?
No.
Is there a likeness between Brahm and Is there a likeness between Brahm and
Brahmans with respect to the mind being Brahmans with respect to family and
tarnished? No. property? No.
Is there a likeness between Brahm and Is there a likeness between Brahm and
Brahmans with respect to self-mastery? Brahmans with respect to self-mastery?
No. No.
Buddha disparages 3-fold knowledge of
Brahmans.
Buddha says that Brahmans cannot
answer difficult questions.

Section Six Request for Teaching

396
Vseha asks if Buddha knows way to
Brahm
Buddha answers by comparing himself to
a native of Manaskaa, who will
naturally know the way to the village
Vseha asks Buddha to teach the way to
Brahm and Buddha asks him to listen
Section Seven
Buddha delivers Tathgata arises
formula, but ends with the four
brahmavihras, each of which is
compared to a trumpeter
Section Eight
Does such a person have possessions?
No.
Does such a person have mind full of
anger? No.
Does such a person have mind full of ill-
will? No.
Does such a person have mind that is
tarnished? No.
Does such a person have self-mastery?
Yes.
Is there a likeness between Brahm and
such a person with respect to
possessions? Yes.
Is there a likeness between Brahm and
such a person with respect to anger? Yes.
Is there a likeness between Brahm and
such a person with respect to ill-will?
Yes.
Is there a likeness between Brahm and
such a person with respect to the mind
being tarnished? Yes.
Is there a likeness between Brahm and
such a person with respect to self-
mastery? Yes.
Section Nine
397
Vseha and Bhradvja asks Buddha to
teach way to Brahm
Buddha compares himself to a native of
Manaskaa, who will naturally know the
way to the village
Vseha and Bhradvja ask Buddha to
teach the way to Brahm and Buddha
agrees
The Way to Brahm
Buddha delivers Tathgata arises
formula, but ends with the four
brahmavihras
Questions about Ideal Person and
Brahm
Is there a likeness between Brahm and
such a person with respect to ill-will?
Yes.
Is there a likeness between Brahm and
such a person with respect to anger? Yes.
Is there a likeness between Brahm and
such a person with respect to resentment?
Yes.
Is there a likeness between Brahm and
such a person with respect to family and
property? Yes.
Is there a likeness between Brahm and
such a person with respect to self-
mastery? Yes.
Conclusion
Vseha and Bhradvja respond with
Excellent, excellent! formula and ask
to be accepted as lay disciples.
Vseha and Bhradvja give rise to
Dharma-eye and are pleased with the
Buddhas words.
Table 2. A comparison of the elements of the narratives of the Tevijja Sutta and its Chinese parallel, the Snmng
Jng.

The comparison between the Pali and Chinese versions of the Tevijja Sutta is

summarized in Table 2. What we have found is that, far from evincing minor

deviations that could stem from accidental misrecollections of a fixed, memorized

text, the two versions of the Tevijja Sutta are rather two tellings of a single story,

whose structure has been defined, but not wholly determined, by a narrative

framework, certain stock formulae borrowed from the broader tradition, and

additional groups of ideas peculiar to this particular story. In every single case where

ideas are presented in list-like form (the questions about various Brahmans having

seen Brahm in Section 3, the similes in Section 4, the questions about the five

characteristics in sections five and eight), the order in which the ideas are presented

differs between the two versions. In two of these instances, the lists are not even

exactly the same, with a more elaborate set of questions found in Section 3 of the Pali

and an elaboration of the simile of the flooded river in Section 4 of the Chinese.

Formulae appear to have been borrowed from the common tradition inconsistently by

the two versions as well. The Pali version ends with a stock formula in which the

Brahmans declare their amazement at the clarity of the Buddhas teaching and ask to

become lay disciples, while the Chinese ends with a simple statement that the

Brahmans attained the Dharma-eye. Conversely, in Section 1 the Chinese version

398
makes use of the very common formula for describing a Brahman learned in the

Vedas to introduce Vseha, Bhradvja, and their teachers, while the Pali version

does not make use of this formula.

Perhaps most obviously, however, the two versions of this stra do not even

agree on where the events recorded within it took place. According to the Pali

version, it took place in the Mango Grove on the Aciravat River north of Manaskaa

in Kosala, but according to the Chinese version, it took place in a forest near

Icchnakala in Kosala. 765 Such discrepancies between nidnas of parallel versions

of the same stra are extremely common throughout all four Nikyas/gamas.

Indeed, as Gregory Schopen has shown, there even exists a short account in the

Mlasarvstivda Vinaya in which the Buddha provides a list of stock places and

persons to use when one forgets the actual places or characters involved in a

particular stra. Accurate transmission of these minor contextual details appears to

have been of little concern to the early Buddhist tradition; indeed, at the end of his

instruction here, the Buddha even adds, In this there is no cause for remorse (di

la gyod par mi byao). 766 This sort of nonchalant attitude cannot be simply be

attributed to the idiosyncracies of one fairly late tradition, either; instructions to freely

765
Admittedly, the nidna for the Pali version makes more sense, since it is reflected in the
actual content of the sutta, insofar as the Buddha makes reference both to the Aciravat and
Manaskaa in the course of his discussion with Vseha and Bhradvja. While this nidna is
therefore arguably more original, the fact that an entirely different nidna is used in the Chinese
version demonstrates the fluidity with which nidnas could be assigned in any given performance of
an early Buddhist stra.
766
Gregory Schopen, If You Cant Remember, How to Make It Up: Some Monastic Rules
for Redacting Canonical Texts, in Bauddhavidysudhkara. Studies in Honour of Heniz Bechert on
the Occasion of His 65th Birthday, Indica et Tibetica 30, edited by P. Kieffer-Plz and J.-U. Hartmann
(Swisttal-Odendorf: Indica et Tibetica, 1997), 571-82.

399
make up forgotten details of a texts nidna are found in the Mahsghika Vinaya

and the Ekottarikgama preserved in Chinese. 767

In conclusion, then, I concur with Cousins that there must have been a stage

possibly quite long, in factduring which the early Buddhist texts were performed in

a somewhat improvisatory manner, as understood within the Oral Theory

developed by Parry and Lord. While scholars such as Gombrich, Allon, and Anlayo

are certainly correct in emphasizing that there was a certain conservative element

within the early Buddhist tradition from a fairly early date, this insight is in no wise

inconsistent with the Parry-Lord Oral Theory, and it certainly does not warrant the

use of such strong terms as memorization and verbatim repetition. Interestingly

enough, this sort of debate over improvisation and memorization is by no means

new to the field of Oral Theory. Even within the context of oral epic for which the

Oral Theory was first designed, there were attempts to minimize the role of

improvisation suggested by Parry and Lord. In responding to one such critique, Lord

wrote, Smith and I have different views of what is meant by improvisatory

technique and memorization. A more or less stable core and a fixed

memorizable text are not the same thing. 768 As I have shown in this chapter, the

texts of the early Buddhist tradition are consistent with the Parry-Lord Oral Theory

and thus can, properly speaking, be called a traditionprecisely because they have

a more or less stable core that was articulated in any given performance through

767
Anlayo Bhikkhu, Zeng-yi A-han, in Encyclopedia of Buddhism, ed. W. G. Weeraratne,
vol. 8, no. 3 (Sri Lanka: Department of Buddhist Affairs, 2009), 822.
768
Albert B. Lord, Characteristics of Orality, Oral Tradition 2, no. 1 (1987), 66-7.

400
improvisatory technique and not through memorization. This insight has a profound

impact on the way we approach the comparative study of stras as preserved by

different early Buddhist traditions, which we will be undertaking with respect to the

encounter dialog genre in Chapters IV.3 and IV.4. First, however, we must address

the question of how the early Buddhist tradition came to develop its structure and

content in the first place. This will be the topic of the next chapter.

401
 Chapter IV.2

Comparative Nikya/gama Studies and the Development

of the Nikya/gamas

IV.2.1 Introduction

As we have seen in the previous chapter, the application of the Parry-Lord

Oral Theory to the early Buddhist texts allows us to understand the differences seen

between different versions of a particular stra not only in terms of errors in

transmission or deliberate changes to a (written) text, but as a manifestation of the

dynamic oral tradition from which those texts are derived, particularly in its

performative aspect. The use of formulae and themes allowed early Buddhist

bhakas to perform texts over and over again, over the course of several generations

and across a wide geographical expanse, without memorizing anything longer than a

standard formula (usually not longer than a paragraph in modern editions) word-for-

word. Once the basic structure of a particular stra had been learned, a performance

could be given by drawing from appropriate formulae as needed. The evidence

clearly shows that while this method could produce performances with a more or

less stable core, various details, including the order of presentation and wording of

402
individual elements of the text, can vary quite widely from one performance to

another. We have also seen, from multiple sources, that the Buddhist tradition was

explicitly unconcerned with fixing these incidental details along any lines akin to

modern conceptions of historical accuracy.

Needless to say, an oral tradition of this sort represents, to a certain extent at

least, a creative process. But once we recognize that early Buddhists were repeatedly

creating and recreating stras with each successive performance, it does not take a

great leap to wonder what role this performative creativity played not only in the

transmission of learned texts, but in the production of new texts. Although the

Buddhist tradition traces the stras back to the Buddha himself, or to one of his

immediate disciples in the case of those set shortly after the Buddhas death, the

inconsistency of the traditions on the circumstantial details of particular stras, as

well as the injunction to accept a text merely on the basis that it conforms with stra

and vinaya, make it unlikely that the early Buddhist texts can be traced back to the

Buddha and his disciples in any literal sense. Instead, it appears that these texts were

accepted as going back to the Buddha, i.e., as buddhavacana, because they were

perceived to faithfully preserve the dharma as taught by the Buddha. 769

769
This is not to say, however, that doctrinal differences between, say, a Theravda version
and a Sarvstivda version of the same stra cannot be found. See, for example, Choong Mun-keats
work on the Chinese Sayuktgama, which he argues has been influenced in certain subtle ways by its
Sarvstivdin sectarian originChoong Mun-keat, The Fundamental Teachings of Early Buddhism: A
comparative study based on the Strga portion of the Pli Sayutta-Nikya and the Chinese
Sayuktgama (Wiesbaden: Harrassowitz Verlag, 2000). Overall, however, scholars have commented
on the relative doctrinal consistency across different versions of the early Buddhist texts.

403
 Indeed, there is reason to believe that it was precisely this emphasis on

preserving the dharma that served as the impetus for creating new Buddhist texts in

the first place. Rupert Gethin has suggested that doctrinal lists (e.g., the four noble

truths, the eightfold path, etc.), which found their most systematic expression in the

mtks of the Abhidharma tradition, but which also pervade the early stras, served

both as mnemonic devices in the early Buddhist oral tradition and as a structure

around which the tradition was built. As he puts it, It is apparent that much of the

scriptural sutta material preserved in the four primary Nikyas can be regarded as

exposition based around lists of one sort or another, and that very many suttas might

be resolved into and summed up in terms of their component lists. 770 Since these

lists encapsulated the dharma, which was what the early Buddhists were most

concerned with preserving faithfully, they could conceivably be expounded in

different ways and cast into different settings with different sets of characters to suit a

particular preaching purpose. Gethin writes, Given the model of interlinking lists,

one can easily see how there might be a version of a sutta mentioning the four

applications of mindfulness as a bare list, and another version mentioning them with a

brief exposition, and yet another version that goes on to give a very full

exposition.771

770
Rupert Gethin, The Mtiks: Memorization, Mindfulness, and the List, in In the Mirror
of Memory: Reflections on Mindfulness and Remembrance in Indian and Tibetan Buddhism, ed. Janet
Gyatso (Albany: SUNY Press, 1992), 149.
771
Ibid., 156.

404
 While many of the better known suttas of the Pali Canon, such as those found

in the Dgha Nikya, give the reader/listener a sense of a real historical occasion

during the life of the Buddha through close detail regarding the setting and cast of

characters, other suttas, especially in the Sayutta and Aguttara Nikyas, dispense

with these literary fictions to facilitate the efficient creation of new texts through

permutative application of doctrinal lists:

The peyylas or repetition sections of the Sayutta and Aguttara Nikayas

are particularly interesting in this respect. Here the texts, as we have them,
indicate an initial pattern or formula that is to be applied to various items in
succession. The result is a text with quite radical abbreviations. Indeed, it is
not always clear from the manuscripts and editions we have just how much we
are meant to expand the material to get the "full" text. Perhaps a certain
freedom is intended here; the peyyla sections of the Sayutta and Aguttara
Nikayas can appear to read more like guidelines for oral recitation and
composition than a fixed literary text. 772

Thus, we are confronted with the possibility that doctrinal lists were used to generate

ever-more complex texts, first through simple permutations of elements of a list

according to a pattern, and then by elaboration with explanations and narrative details.

Needless to say, the only means at our disposal today to study how the early

Buddhist tradition might have developed along such lines is to compare the different

versions of the early Buddhist stras that have come down to usprimarily in the

first four Pali Nikyas and in the four Chinese gamasin the hope that the

similarities and differences between them, and the principles according to which they

are collected, may illuminate the process according to which they developed in the

first place. Comparative Nikya/gama studies owes its existence to the publication

772
Ibid.

405
in 1908 of Anesaki Masaharus study of the four gamas in Chinese, in which for the

first time parallels were identified between the stras of the Chinese gamas and the

suttas of the Pali Nikyas. 773 Only in the past few decades, however, have scholars

begun to develop an understanding, on the basis of these comparative studies, of how

the Nikya/gama tradition developed in the first place. In this chapter, I will

synthesize some of the general findings of this scholarship in order to contextualize

my analysis of the encounter dialog genre in Chapters IV.3 and IV.4.

Central to my approach is the work of Ynshn, who argued that the

Sayuktgama was an early collection of Buddhist texts organized according to the

first three agasstra, geya, and vykaraa. I therefore begin in Section IV.2.2

with an explanation of Ynshns theory and its implications for our understanding of

the relationship between the Sayuktgama(s) in Chinese and Sayutta Nikya in

Pali. Then, in sections IV.2.3, IV.2.4, and IV.2.5, I turn, respectively, to the

Dgha/Drgha, Majjhima/Madhyama, and Aguttara/Ekottarika collections. In each

case, I examine the likely sectarian affiliations of the extant collections, as well as the

differences between the Pali and non-Pali versions, and I suggest possible ways that

each collection may have developed.

773
M. Anesaki, The Four Buddhist gamas in Chinese: A Concordance of their Parts and the
Corresponding Counterparts in the Pli Nikyas, Transactions of the Asiatic Society of Japan 35, part
3 (1908): 1-149. Anesakis list has since been refined; a more comprehensive guide to parallels that
was the standard work in the field for most of the 20th century was published in 1929 by Akanuma
ChizenAkanuma Chizen, The Comparative Catalogue of the Chinese gamas and the Pli Nikyas
(Nagoya: Hajinkaku-shobo, 1929). The need for revisions to Akanumas attributions of parallels has
been raised recently; for a revision to his list of parallels for the Majjhima Nikya, see Anlayo and
Roderick S. Bucknell, Correspondence Table for Parallels to the Discourses of the Majjhima Nikya:
Toward a Revision of Akanumas Comparative Catalogue, Journal of the Centre for Buddhist Studies,
Sri Lanka 4 (2006): 215-43. A much more comprehensive set of revised correspondence tables is
found at Anlayo and Bucknells Online Sutta Corresponce Project website [www.suttacentral.net].

406
IV.2.2 Sayutta Nikya / Sayuktgama

As already mentioned in Chapter IV.1, the only nearly complete or partial

versions of the Sayuktgama that have come down to us other than the Pali

Sayutta Nikya of the Theravdin school are in Chinese translation. The most

important of these is T. 99, entitled the Z hn Jng (), which is preserved

in the Taish version in 50 jun 774 and appears to be almost complete. This

translation is attributed to Guabhadra (Qinbtulu , 394-468 CE),

and it has been dated by Enomoto Fumio to 435-436 CE. 775 Although scholarly

opinion is not totally unanimous, most attribute it to the (Mla-)Sarvstivda

school. 776 As we will discuss in further detail shortly, this version of the

774
A jun () is a unit of division used in the Chinese Tripiaka that corresponds to the
physical volumes on which the Chinese translations of Buddhist texts were once written. Although
modern editions no longer physically divide a particular text into the same number of volumes, the
jun divisions are retained as a convenient system of division for reference purposes. It should be
noted that the division into jun is purely an artifact of the Chinese translation process and has nothing
to do with the original divisions of the Indian texts from which they are translated.
775
Glass, Four Gndhr Sayuktgama Stras, 28, 38-9.
776
Mayeda Egaku, Japanese Studies on the Schools of the Chinese gamas, in Genshi
bukky seiten no seiritsushi kenky [A History of the Formation of Original Buddhist Texts] (Tokyo:
Sankib Busshorin, 1964), 99-101. Although many scholars have distinguished between the
Mlasarvstivda and the Sarvstivda and debated whether particular texts (including the Z hn
Jng) should be attributed to one or the other, Enomoto Fumio has argued convincingly that the two
were not really separate sects; rather, Mlasarvstivda was simply a term used by (at least some)
Sarvstivdins as an affectation, in reference to their claim that the Sarvstivda was the root (mla)
of all the nikyas. Although this conclusion might seem to be precluded by the fact that the
Mlasarvstivda Vinaya preserved in Tibetan and the Sarvstivda Vinaya preserved in Chinese
are not the same, Enomoto argues that there was, for whatever reason, no single Canon for any
particular sect and that there may very well have been different versions of various texts used by, for
example, different groups of Sarvstivdins. See Enomoto Fumio, Mlasarvstivdin and
Sarvstivdin, in Vividharatnakaraaka: Festgabe fr Adelheid Mette, Indica et Tibetica:
Monographien zu den Sprachen und Literaturen des indo-tibetischen Kulturraumes 37, ed. Christine
Chojnacki, Jens-Uwe Hartmann, and Volker M. Tschannerl (Swisttal-Odendorf: Indica et Tibetica
Verlag, 2000), 239-50. In what follows, therefore, I do not generally distinguish between Sarvstivda
and Mlasarvstivda as sectarian attributions of works lacking traditional attributions and refer simply
to either Sarvstivda or (Mla-)Sarvstivda.

407
Sayuktgama, like many Chinese Buddhist translations, is in disarray in the version

that has come down to us. The disorder in the case of this text consists not only of the

jun being out of order; in addition, jun 23 and 25 are missing and parts of the

Aokvadna (similar but not not identical to chapters 26 and 27 of the extant

Divyvadna) have been inserted in their place. 777

Immediately following the Z hn Jng in the Taish are two other

translations from the Sayuktgama that are unfortunately not nearly as complete as

the former. The first of these, T. 100, is entitled Biy Z hn Jng (,

lit., Another Translation of the Sayuktgama Stra), which in the Taish is divided

into 16 jun. The identity of the translator of this text is unknown, but it was most

likely translated somewhat earlier than T. 99, between 385 and 431 CE, 778 and

Bingenheimer has noted a similarity in its translation style to that of texts known to

have been translated by Zh Fnin (), who was active at around the same

time (384-416 CE). 779 The Biy Z hn Jng appears to be simply an incomplete

translation of the Sayuktgama that for reasons unknown was never completed. It

contains parallels to most of the suttas in the Sagth Vagga of the Pali Sayutta

Nikya, as well as suttas from the Bhikkhu, Mahkassapa, Gmai, Mahnma,

777
Glass, Four Gndhr Sayuktgama Stras, 39.
778
This dating, determined by Mizuno Kgen, is cited by Marcus Bingenheimer, The Shorter
Chinese Sayukta gama: Preliminary Findings and Translation of Fascicle 1 of the Bieyi za ahan jing
(T. 100), Buddhist Studies Review 23, no. 1 (2006), 21.
779
Bingenheimer, Shorter Chinese Sayukta gama, 26.

408
Anamatagga, and Abykta Sayuttas 780; therefore, since it does not represent a broad

array of Sayuktgama texts, it is unlikely to be an anthology. In addition,

Bingenheimer has noted that inconsistencies in the translation suggest that the text

was never edited in any systematic way781; this would seem to corroborate the

supposition that the translation was simply halted and never completed.

The Biy Z hn Jng has been attributed to various schools, including the

Kyapyas, Dharmaguptakas, Mahsakas, and Mlasarvstivdins. Based on a

comparison of the wording of its gths to parallel gths in Vinayas of known

sectarian origin, however, the (Mla-)Sarvstivdin attribution appears to be the most

likely. 782 This attribution is also corroborated by the work of Roderick Bucknell,

who has shown that T. 100, which like T. 99 appears to be in disarray, can, when

rearranged, be shown to closely parallel the corresponding portions of the latter. In

fact, there exist two versions of the Biy Z hn Jngthe 16 jun version that is

more well known because it is preserved in the Taish, and a lesser-known 20 jun

version that is preserved in other versions of the Chinese Canon. Bucknell has shown

that the latter, 20 jun version is in fact much more orderly than the 16 jun version,

closely paralleling the order of the full Z hn Jng, and he has moreover shown the

process by which the 16 jun version likely became disarrayed. Unfortunately, by an

accident of history, the latter, defective version became the better known because it

780
Roderick Bucknell, The Historical Relationship Between the Two Chinese
Sayuktgama Translations, Chung-Hwa Buddhist Journal 24 (2011), 37.
781
Bingenheimer, Shorter Chinese Sayukta gama, 26.
782
See Ibid., 22, which cites in particular the work of Enomoto.

409
was, for reasons that are unclear, preferred by Sugi, the editor of the second edition of

the Korean Canon, which would become the precursor of the Japanese Taish. 783

The second incomplete translation of the Sayuktgama is even shorter than

the first and in fact appears to be simply a short anthology of Sayukta texts. It is

found immediately after the Z hn Jng and the Biy Z hn Jng at T. 101 and is

called the Z hn Jng, though I will refer to it exclusively as T. 101 in order to

distinguish it from the Z hn Jng that is a nearly complete translation of the

Sayuktgama at T. 99. Although no translator is assigned to T. 101 in the Chinese

catalogs, the Japanese scholar Hayashiya Tomojir argued on the basis of a

terminological comparison to known translations of n Shgo (), a Parthian

who inaugurated the earliest translations of Buddhist texts into Chinese, 784 that he

was in fact the translator of this anthology of Sayuktgama texts, an attribution that

has recently been supported by Paul Harrison. 785 Little work on T. 101 has been done

other than a short but important article by Harrison, in which he shows how the text

as it comes down to us may have become disarranged since the time when an early

Chinese catalog placed its stras in a different order, and argues that due to certain

similar readings to those found in T. 99 and T. 100, it may also be taken from a
783
Roderick Bucknell, The Other Translation of Sayuktgama, Chung-Hwa Buddhist
Journal 21 (2008): 24-54. For a more detailed argument specifically pertaining to the common (Mla-
)Sarvstivdin ancestry of the Z hn Jng and the Biy Z hn Jng, see Bucknell, The Historical
Relationship Between the Two Chinese Sayuktgama Translations.
784
For an overview of scholarship of n Shgo and his translations, see Nattier, Guide to the
Earliest Chinese Buddhist Translations, 38-72.
785
Paul Harrison, Another Addition to the An Shigao Corpus? Preliminary Notes on an Early
Chinese Sayuktgama Translation, in Studies in Honour of Dr. Sakurabe Hajime on the Occasion of
His Seventy-seventh Birthday, ed. Sakurabe Ronshu Committee (Kyoto: Heirakuji Shoten, 2002), 2.

410
Sarvstivdin version of Sayuktgama, though not conclusively, due to the

simultaneous existence of important differences. 786

Overall, however, the most important text for the study of the Sayuktgama

tradition, other than the complete Sayutta Nikya of the Theravdins preserved in

Pali, is T. 99, the Z hn Jng. As already mentioned above, this Chinese translation

almost certainly comes from a Sarvstivda tradition, thus providing a useful

counterpoint to the Theravda version in Pali, and with the exception of two jun that

were at some point lost and replaced with sections of the Aokvadna, it is complete.

The major obstacle in using it comparatively alongside the Sayutta Nikya is that it

is in disarray. Several solutions to restore the original order of the Sayuktgama has

been suggested over the years by various Japanese and Chinese scholars. The earliest

of these was offered by Anesaki in his already-mentioned 1908 comparative study of

the Nikyas and gamas. His reconstruction was adopted with only one slight

modification by the Japanese translator of the Sayuktgama, Shiio Benky, in 1935,

and then more substantively revised in 1954 by Hanayama Shd, who for the first

time produced a reconstructed order simply by rearranging the jun, without splitting

any of them up. This approach has been deemed superior, insofar as it is easier to

explain how the order of complete jun became disarranged than to explain how

individual jun could have been split up, and thus it has been followed ever since. 787

786
Harrison, Another Addition to the An Shigao Corpus?
787
Glass, Four Gndhr Sayuktgama Stras, 39-41.

411
 In the meantime, the Chinese scholar L Chng published in 1924 a revision

of Anesakis original reconstruction that took into account his discovery that the

Vastusagraha (T. 30, p. 772-882: Shsh Fn, ), a portion of the

Yogcrabhmi (T. 30: Yji Shd Ln ), contains a partial commentary

on the Sayuktgama, 788 which could be used to help reconstruct the original order

of the collection. L also took into account, for the first time, the placement of the

two missing jun in his reconstruction of T. 99s original sequence. In 1971,

followed by a slight revision in 1983, the Taiwanese scholar-monk Ynshn

combined the most important contributions of Hanayama and Lthe formers effort

to avoid splitting up individual jun and the latters effort to ascertain the position of

the two missing jun, as well as his recognition of the value of the commentary found

in the Vastusagrahato produce what is now the most up-to-date reconstruction

of the original order of the jun in T. 99, and thus the original order of the

Sarvstivda Sayuktgama. His reconstruction was accepted with only slight

modification in 1985 by the Japanese scholar Mukai Akira and has been reviewed

favorably by other Japanese, Chinese, and Western scholars. 789

The result of these reconstructions of T. 99 is a Sarvstivdin Sayuktgama

that is remarkably similar in structure to the Theravdin Sayutta Nikya. The latter

is divided into five sections, called vaggas in Pali (Skt. varga). They are, in order,

788
Ibid., 29; Roderick S. Bucknell, Sayuktgama, in Encyclopedia of Buddhism, vol. 7,
fasc. 4, ed. W. G. Weeraratne (Colombo: Government of Ceylon, 2006), 685.
789
Glass, Four Gndhr Sayuktgama Stras, 41-2. For a table comparing the
reconstructions of all these scholars, see ibid., 40.

412
the Sagth Vagga, the Nidna Vagga, the Khandha Vagga, the Sayatana Vagga,

and the Mah Vagga. Of these, the first stands out because, as its name suggests, all

of its constituent suttas include verses (gth). According to Ynshns

reconstruction, 790 the Sarvstivdin Sayuktgama also has five sections, referred to

as sng () in the Chinese, which corresponds to varga in Sanskrit, that, while found

in a different order, each correspond to one of the vaggas of the Pali version. First is

the Five Aggregates Section (W Yn Sng ), which corresponds to the

Khandha Vagga. Second is the Six Sense Spheres Section (Li Rch Sng

), which corresponds to the Sayatana Vagga. Third is the Causal Conditions

Section (Zyn Sng ), which corresponds to the Nidna Vagga. Fourth is

Path Section (Dopn Sng ), which corresponds to the Mah Vagga. Fifth

and finally is the Eight Assemblies Section (B Zhng Sng ), which

corresponds to the Sagth Vagga. As the name of the collection suggests, both the

reconstructed Sayuktgama 791 and the Sayutta Nikya are also divided into

smaller units called sayukta (P. sayutta), which literally means connected and

refers to the organizing principle of the collectionorganizing stras according to

common theme. Most sayuktas and sayuttas also have parallels in the other

790
Mukai, however, introduces two additional vargasthe rvakabhita Varga and the
Tathgatabhita Varga. See Glass, Four Gndhr Sayukta Stras, 42. The stras found in these
two vargas are for the most part distributed in the minor chapters addended in the second through
fifth Vaggas in the TheravdaSujto, History of Mindfulness, 42.
791
The Chinese text of T. 99 actually does not preserve sayukta (Ch. xingyng )
divisions and titles, but these can be inferred from the Vastusagraha and other texts.

413
tradition, but their distribution between the five vargas, and even more so their order,

differs somewhat. 792

Much of the success in putting together a credible reconstruction of the

Sarvstivdin Sayuktgama is due to the discovery and study of the commentary

thereupon in the Vastusagraha, first by L Chng and then by Ynshn. This

discovery is not only of relevance to the reconstruction of T. 99, however; it also is of

relevance to our understanding of the development of the entire early Buddhist

tradition. According to the Vastusagraha, the Sayuktgama is the foundation of

the four gamas, and it is divided into three components corresponding to the first

three of the nine or twelve agas referred to in multiple early Buddhist texts across

sectarian lines. 793 These three are stra (P. sutta), geya (P. geyya), and vykaraa (P.

veyykaraa). Based on the fact that the Vastusagraha records that the

Sayuktgama, composed of these three agas, is the foundation of all the gamas,

and the Mahsuat Sutta (MN 122=M 191) in fact refers to these same three

agas in isolation, Ynshn argues that they were the earliest of the agas, out of

which, as constituted in some early form of the Sayuktgama, the rest of the early

Buddhist oral tradition grew. 794

792
For a table illustrating the distribution of sayuktas in the Sayuktgama and sayuttas in
the Sayutta Nikya in parallel to one another, see Choong, Fundamental Teachings of Early
Buddhism, 19-22.
793
For a comprehensive overview of references to the nine or twelve agas, see Lamotte,
History of Indian Buddhism, 143-7. For a discussion of the agas that is sympathetic to Ynshns
interpretation, see Bhikkhu Sujto, History of Mindfulness, 52-69.
794
Choong, Fundamental Teachings of Early Buddhism, 9-11.

414
 The division of the Sayuktgama into agas is not a simple division of the

collection into three discrete parts, but rather a categorization of its stras into three

genres of texts. Of the three, geya is the easiest to explain and identify, since it is

identified in all sources as referring to texts that mix prose and verse (gth). Within

the Sayutta Nikya, most such texts, as already mentioned, are found in the Sagth

Vagga, which, as its name suggests, is composed of texts that include verse gths.

In the Sarvstivdin Sayuktgama, this division is paralleled by the Eight

Assemblies Division (B Zhng Sng ), which is also composed of stras

that combine prose and verse. In addition, the contents of the two collections are,

aside from differences in the ordering of the constituent sayuktas, quite similar, with

the major exception of the Bhiku Sayukta (), whose Pali parallel, the

Bhikkhu Sayutta, is found instead in the Nidna Vagga. But as Bucknell has

convincingly argued, this discrepancy is almost certainly due to an error at some point

in the transmission of the Pali Canon, in which the Bhikkhu Sayutta was transferred

from the Sagth Vagga to the Nidna Vagga. 795 Restoring the Bhikkhu Sayutta to

the Sagth Vagga yields a parallel to the B Zhng Sng that is remarkably similar

in content and structure, if not arrangement. 796

Interpretation and identification of the other two agas, stra and vykaraa,

is somewhat more complicated. What must first be understood is that stra as an

795
Roderick S. Bucknell, The Structure of the Sagtha-Vagga of the Sayutta Nikya,
Buddhist Studies Review 24, no. 1(2007), 13-18.
796
For a table giving the parallels between the xingyng (sayuktas) of the Sarvstivdin
Bzhng Sng and the sayuttas of the Sagth Vagga, see Choong, Fundamental Teachings of Early
Buddhism, 20.

415
aga is not coextensive with the category stra (or the Pali equivalent sutta) as it has

come to be used in the Buddhist tradition and modern Buddhist studiesthat is, to

refer to virtually any individual Buddhist text that records an event in the life of the

Buddha or one of his close disciples. This was in fact noticed over 150 years ago, by

none other than Eugne Burnouf:

The word stra is a quite well-known term in the literature of ancient India; it
designates these short and obscure sentences that contain the fundamental
rules of Brahmanical science, from grammar to philosophy. This signification
is not unknown to the Buddhists, but I must first hasten to remark that it is
not only in this way that the Buddhists understand the word stra, and that the
treatises to which this title is applied are of a very different character from
those that it designates in the orthodox literature of ancient India. 797

Within the broader context of Indian literature, in other words, stra should be

expected to refer to, quite literally, strings of short, aphoristic sayings. Moreover,

given the criterion, discussed above, that texts that conform with sutta and vinaya be

considered buddhavacana, and the fact that Buddhist texts typically incorporate

numerous string-like lists, the early Buddhists appear to have been aware of such an

understanding of the word stra. 798 What came to be the more common use of the

word stra in the Buddhist tradition was to refer to long, narrative texts, such as the

Mahparinirva Stra, which bear little in common with the broader Indian stra

797
Burnouf, Introduction to the History of Indian Buddhism, 85.
798
An excellent example of a Buddhist stra in the ordinary Indian sense of the term is the
Prtimoka Straindeed nothing more than a list of the rules to be followed by a monk or nun
which, although not per se canonical within the Theravda tradition, is imbedded within the Vinaya in
the Sutta Vibhaga, which in turn literally means Commentary on the Sutta. I do not agree with
Oskar von Hinber, however, that the strga referred originally only to the Prtimoka Strasee
Oskar von Hinber, A Handbook of Pli Literature (Berlin: Walter de Gruyter, 2000), 7, and Oskar
von Hinber, Die neun Agas: Ein frher Versuch zur Einteilung buddhistischer Texte, Wiener
Zeitschrift fr die Kunde Sdasiens 38 (1994): 121-35.

416
genre. But given that, as Gethin has argued (see Section IV.2.1), even the long,

narrative stras appear to be based largely on lists, it is not difficult to understand

how the Buddhists came to extend the use of the word stra in such a way. Again,

this was noted early on by Burnouf:

One should not believe, however, that these brief maxims, so appreciated by
antiquity, are entirely lacking in the teaching of kya; on the contrary, in the
stras one still finds several traces of this sententious exposition that
summarizes a long development in a few words or in a concise stanza. It is
permissible to believe that kya must not have abstained from using these
sentences, and that the memory of the use he made of them in his teaching
favored the entirely special application his disciples made of the term stra
through extending it to his moral and philosophical preaching. 799

Indeed, to this we can add that, at least within the context of the early Buddhist

tradition, what we commonly refer to as stras are, properly speaking, actually

referred to as strntas (P. suttantas)literally, end of the strawhich, in

parallel to the Upaniads qu Vednta in the Brahmanical tradition, suggests a

development and fulfillment of something prior known as stra.

The term stra as an aga, then, hypothetically at least, should refer to the

doctrinal lists on which stras as texts (i.e., strntas) are built. The term vykaraa

(lit., answer or declaration) should refer to expositions of those stra lists 800

which, needless to say, would in many cases correspond to the strnta texts

themselves. The Vastusagraha, however, classifies certain stras (i.e., strntas)

as belonging to the strga and others as belonging to the vykaraga. This is

not as illogical as it might seem at first, since many of the stras in the
799
Burnouf, Introduction to the History of Indian Buddhism, 86-7.
800
On the vykaraga, see Sujto, History of Mindfulness, 65-7.

417
Sayuktgama/Sayutta Nikya are, unlike the better-known narrative stras of the

Dgha and Majjhima Nikyas, very short, simple declarations of doctrinal points

made by the Buddha, with little or no narrative framework. Moreover, as has long

been recognized through study of the Pali version, the Sayutta Nikya as a whole is

clearly organized according to doctrinal lists. The overall structure of the Sayutta

Nikya appears to correspond to the four noble truths (Khandha and Sayatana

Vaggas to the first, Nidna Vagga to the second and third, and Mah Vagga to the

fourth), 801 and the sayuttas of the Mah Vagga reflect, or are reflected by, the 37

Wings of Awakening (bodhi-pakkhiy dhamm). 802 In any case, the

Vastusagraha classifies the stras outside of the Eight Assemblies Division (the

Sarvstivdin equivalent to the Sagth Vagga) as belonging to either the strga or

the vykaraga, and in a section suggestively called the *Stra-mtk (Qjng Sh

Xng Z Sh ), it provides a commentary on those parts of the

Sayuktgama that it classifies under the strga. 803

It should be noted, however, that the Vastusagrahas schema dividing the

stras of the Sayuktgama between the first three agas is not entirely

unproblematic. It makes this classification not at the level of individual stras, but

rather at the level of sayuktas, the most basic division of the Sayuktgamaand

that from which the collection gets its namein which stras connected by a

801
Ibid., 43.
802
Gethin, Buddhist Path to Awakening, 21-3.
803
Glass, Four Gndhr Sayuktgama Stras, 29.

418
common topic are grouped together. 804 That this is problematic is easily seen from

the fact that stras combining prose and verse, and thus presumably belonging to the

geyga, are not limited to the B Zhng Sng (or the parallel Sagth Vagga and

stray Bhikkhu Sayutta in the Pali), but are also found scattered throughout other

sayuktas that are variously classified as stra or vykaraa. Bhikkhu Sujto has

suggested that perhaps the Vastusagraha accurately preserves an institutional

memory that the first three agas were constituted in an early version of the

Sayuktgama, but that it artificially imposes a crude schema for classifying entire

sayuktas in the fully developed Sayuktgama as belonging to one of the three

agas. He then goes on to show how individual stras might be classified according

to aga within the Pali Sacca Sayutta and its Chinese equivalent, and he argues that

at one time vykaraa texts may have been grouped together separately and after

corresponding stra texts, and that this is reflected more clearly in the Sarvstivda

version of the Sayuktgama. 805 On the other hand, in defense of the

Vastusagraha, the sayuktas that it identifies as belonging to the strga do

appear to stand out, at least in terms of their names, as some of the most important

doctrinal lists of early Buddhism, which also closely parallel the Abhidharma mtks:

skandha, yatana, prattyasamutpda, satya, dhtu, vedan, smtyupasthna,

samyakpradhna, ddhipda, indriya, bala, bodhyaga, ryamrga, npna-smti,

804
For tables listing the sayuktas (Ch. ) of the reconstructed Sarvstivdin
Sayuktgama in parallel to the sayuttas of the Theravdin Sayutta Nikya, together with their
aga classification according to the Vastusagraha, see Choong, Fundamental Teachings of Early
Buddhism, 243-51.
805
Sujto, History of Mindfulness, 73-8.

419
ik, and avetyaprasda. 806 It is always possible, indeed quite likely, that the

individual stras that comprise these and the sayuktas identified as vykaraa have

been rearranged in both the Sarvstivdin and the Theravdin traditions. Clearly,

there is much work still to be done on studying the aga classifications of individual

stras in the Sayuktgama and their distribution through the various sayuktas of

the collection. Luckily, for our purposes, this is of little concern, since most of what

can be classified as encounter dialogs in the Sayuktgama are found in the

Brhmaa Sayukta, which is in the B Zhng Sng/Sagth Vagga, consists

entirely of stras combining prose and verse, and thus can unproblematically be

assigned to the geyga.

IV.2.3 Dgha Nikya / Drghgama

In the case of the Drghgama, we are quite lucky in that we have three

substantial texts (i.e., more than just fragments) representing three distinct sectarian

traditions. The first of these, of course, is the Pali Dgha Nikya, which represents

the Theravda tradition. It contains 34 suttas divided into three parts called vaggas.

Another complete version of the Drghgama is preserved in Chinese and is found at

the very beginning of the Taish, T. 1, entitled the Chng hn Jng (lit., Long

gama Stra). This version of the Drghgama was translated into Chinese by

Buddhayaas (Ftuysh ) and Zh Fnin () in 413 CE. It is

806
For a table comparing the strga sayuktas of the Sayuktgama and Sayutta Nikya,
as well as the mtks of three Abhidharma texts, see Choong, Fundamental Teachings of Early
Buddhism, 252.

420
virtually unanimously considered by scholars to be of the Dharmaguptaka sect, in part

because of similarities to the known Dharmaguptaka Vinaya, but also because

Buddhayaas was a Dharmaguptaka monk, and all of his other translations but one are

of the Dharmaguptaka sect. 807 The Dharmaguptaka is genealogically close to the

Theravda, insofar as both are branches of the Vibhajyavda, a subdivision of the

Sthaviravda alongside the Sarvstivda. This closeness is reflected in the contents

of the Chng hn Jng, which are quite similar to those of the Dgha Nikya. The

Chng hn Jng contains a total of 30 stras, of which all but three have parallels in

the Dgha Nikya. Conversely, all but seven of the suttas in the Dgha Nikya (6, 7,

10, 17, 22, 30, 32) have parallels in the Chinese. As is common even in cases when

the content of two versions of a Nikya/gama is similar, however, the order of the

stras is quite different in the Chng hn Jng than in the Dgha Nikya.

Until recently, the Dharmaguptaka Drghgama preserved in Chinese was the

only known parallel to the Pali Dgha Nikya other than individual stras and

fragments preserved in various languages. Since the late 1990s, however, a new set

of Sanskrit manuscripts preserving what was once presumably a complete

Drghgama collection has emerged from Afghanistan. These manuscripts have

unfortunately not yet been published, but they have been studied extensively by Jens-

Uwe Hartmann, who has published several articles giving preliminary findings from

his research. 808 Although this manuscript collection, as it stands, is not, and probably

807
Mayeda, Japanese Studies on the Schools of the Chinese gamas, 97.
808
See above, n. 728.

421
never will be, complete, it does include over half of the stras in the collection, and it

is possible to reconstruct (as Hartmann has) its remaining structure on the basis of

uddnas (tables of contents) found in the sections that have survived. A more

thorough study of the contents of this Drghgama will have to await the full

publication of the available manuscripts; nevertheless, what is already clear from

Hartmanns reconstruction of the original texts structure is that its contents are quite

different from that of either the Theravdin Dgha Nikya or the Dharmaguptaka

Chng hn Jng. The Sanskrit Drghgama is, like the Pali, divided into three

sections, but only the third of these, the laskandha Nipta, corresponds to one of the

divisions of the Pali Dgha Nikya (the Slakkhandha Vagga, which comes first in

that collection). The other two sections, the astraka Nipta and the Yuga Nipta,

while containing some stras that are parallel to suttas found in the Dgha Nikya, are

similar enough in either name or structure to be said to be parallel to either the Mah

Vagga or the Pthika Vagga of the Pali version. This is symptomatic of the fact that,

when viewed overall, the Sanskrit Drghgama is, at a total of 47 stras, much bigger

than both the Dgha Nikya and the Chng hn Jng, and many of those stras have

no parallel in the other D(r)gha traditions, in some cases lacking any parallel in any

other tradition whatsoever. 809 Based on a comparison to descriptions of the

Drghgama found in works of known sectarian affiliation, Hartmann has attributed

this Sanskrit version to the (Mla-)Sarvstivda, 810 which is consistent with the fact

809
Hartmann, Contents and Structure, 121-8.
810
Hartmann, Zu einer neuen Handschrift des Drghgama, 361.

422
that it is substantially different from the relatively similar Theravdin and

Dharmaguptaka versions insofar as the Sarvstivda is genealogically more removed

from the other two than they, as branches of the Vibhajyavda, are from each other.

A comparision of the stras, and their arrangement, found in the Theravdin,

Dharmaguptaka, and Sarvstivdin versions, as can be found in Table 3, demonstrates

the extent to which the three versions relate to and differ from each other. As already

stated, the Theravdin and Dharmaguptaka versions differ primarily in terms of

arrangement, while the Sarvstivdin version differs from the other two in that it is

much larger and includes many stras that have no parallel in the other versions.

Interestingly, ten of the stras in the Sarvstivdin version that are not found in either

of the other two versions are found in the Theravdin Majjhima Nikya. These same

ten stras are not, however, found in the Chinese translation of the Madhyamgama,

which, as we will see in Section IV.2.4, is attributed by scholars to the Sarvstivda.

Conversely, there are ten suttas in the Pali Dgha Nikya that are not found in the

Sanskrit Drghgama, but are found in the Chinese Madhyamgama. This

remarkable correspondence confirms that the Sanskrit Drghgama and the Chinese

Madhyamgama come from a common tradition and thus supports the attribution of

the Sanskrit Drghgama to the Sarvstivda. In addition, since it involves a

difference between the Sarvstivda and Theravda Drgha/Dgha and

Madhyama/Majjhima traditions of exactly ten stras in either direction, Sujto argues

that one of the two traditions most likely swapped an entire varga of texts (since

423
vargas often come in tens) between the Drgha and the Madhyama at some point in

its history. 811

Theravda (Pali Dgha Dharmaguptaka Sarvstivda (Sanskrit

Nikya) (Chinese Chng hn Drghgama)
Jng)
1. Brahmajla Sutta 21. 47. Brahmajla Stra
(Slakkhandha Vagga) (laskandha Nipta)
2. Smaaphala Sutta 27. 44. Rj Stra
(Slakkhandha Vagga) (laskandha Nipta)
3. Ambaha Sutta 20. 35. Ambha Stra
(Slakkhandha Vagga) (laskandha Nipta)
4. Soadaa Sutta 22. 33. roatya Stra
(Slakkhandha Vagga) (laskandha Nipta)
5. Kadanta Sutta 23. 34. Kadya Stra
(Slakkhandha Vagga) (laskandha Nipta)
6. Mahli Sutta 32. Mahallin Stra
(Slakkhandha Vagga) (laskandha Nipta)
7. Jliya Sutta 30. Maa Stra 1
(Slakkhandha Vagga) (laskandha Nipta)
31. Maa Stra 2
(laskandha Nipta)
8. Mahshanda Sutta 25. 46. Kyapa Stra
(Slakkhandha Vagga) (laskandha Nipta)
9. Pohapda Sutta 28. 36. Phapla Stra
(Slakkhandha Vagga) (laskandha Nipta)
10. Subha Sutta 42. uka Stra
(Slakkhandha Vagga) (laskandha Nipta)
11. Kevaddha Sutta 24. 29. Kaivartin Stra
(Slakkhandha Vagga) (laskandha Nipta)
12. Lohicca Sutta 29. 27. Lohitya 1
(Slakkhandha Vagga) (laskandha Nipta)
28. Lohitya 2
(laskandha Nipta)
13. Tevijja Sutta 26. 45. Vsiha (laskandha
(Slakkhandha Vagga) Nipta)
30.
25. Tridain Stra
(laskandha Nipta)
26. Piglatreya Stra

811
Sujto, History of Mindfulness, 83-4.

424
 (laskandha Nipta)
37. Kraavdin Stra
(laskandha Nipta)
(AN 4.198) 38. Pudgala Stra
(laskandha Nipta)
39. ruta Stra
(laskandha Nipta)
40. Mahlla Stra
(laskandha Nipta)
41. Anyatama Stra
(laskandha Nipta)
(MN 55, Jvaka Sutta) 43. Jvaka Stra
14. Mahpadna Sutta 1. 5. Mahpadna Stra
(Mah Vagga) (astraka Nipta)
15. Mahnidna Sutta 13. (T. 26, M 97)
(Mah Vagga)
16. Mahparinibbna Sutta 2. 6. Mahparinirva Stra
(Mah Vagga) (astraka Nipta)
17. Mahsudassana Sutta (T. 26, M 68)
(Mah Vagga)
18. Janavasabha Sutta 4. 13. Jinayabha Stra (Yuga
(Mah Vagga) Nipta)
19. Mahgovinda Sutta 3. 14. Govinda Stra (Yuga
(Mah Vagga) Nipta)
20. Mahsamaya Sutta 19. 24. Mahsamja Stra
(Mah Vagga) (Yuga Nipta)
21. Sakkapaha Sutta 14. (T. 26, M 134)
(Mah Vagga)
22. Mahsatipahna Sutta (T. 26, M 98)
(Mah Vagga)
23. Pysi Sutta (Mah 7. (T. 26, M 71)
Vagga)
24. Pika Sutta (Pika 15. [missing character] 9. Bhrgava Stra (Yuga
Vagga) Nipta)
25. Udumbarikashanda 8. (T. 26, M 104)
Sutta (Pika Vagga)
26. Cakkavattishanda 6. (T. 26, M 70)
Sutta (Pika Vagga)
27. Aggaa Sutta (Pika 5. (T. 26, M 154)
Vagga)
28. Sampasdanya Sutta 18. 15. Prsdika Stra (Yuga
(Pika Vagga) Nipta)
29. Psdika Sutta (Pika 17. 16. Prasdanya Stra

425
Vagga) (Yuga Nipta)
30. Lakkhaa Sutta (Pika (T. 26, M 59)
Vagga)
31. Siglaka Sutta (Pika 16. (T. 26, M 135)
Vagga)
32. niya Sutta (Pika 23. na Stra (Yuga
Vagga) Nipta)
33. Sagti Sutta (Pika 9. 3. Sagti Stra
Vagga) (astraka Nipta)
34. Dasuttara Sutta (Pika 10. 1. Daottara Stra
Vagga) (astraka Nipta)
11.
12.
2. Arthavistara Stra
(astraka Nipta)
(Vin. 4.1) 4. Catupariat Stra
(astraka Nipta)
(MN 60, Apaaka Sutta) 7. Apannaka Stra (Yuga
Nipta)
8. Sarveka Stra (Yuga
Nipta)
(MN 105, Sunakkhatta 10. alya Stra (Yuga
Sutta) Nipta)
(MN 4, Bhayabherava 11. Bhayabhairava Stra
Sutta) (Yuga Nipta)
(MN 12, Mahshanda 12. Romaharaa Stra
Sutta) (Yuga Nipta)
(MN 102, Pacattaya 17. Pacatraya Stra
Sutta) (Yuga Nipta)
18. Myjla Stra (Yuga
Nipta)
(MN 95, Cak Sutta) 19. Kmahika Stra
(Yuga Nipta)
(MN 36, Mahsaccaka 20. Kyabhvan Stra
Sutta) (Yuga Nipta)
(MN 85, Bodhirjakumra 21. Bodha Stra (Yuga
Sutta) Nipta)
(MN 100, Sagrava Sutta) 22. akara Stra (Yuga
Nipta)
Table 3. A comparative table of the stras found in the Theravdin, Dharmaguptaka, and Sarvstivdin versions
of the Drghgama, adapted from Lewis R. Lancaster, The Korean Buddhist Canon: A Descriptive Catalogue,
K647, and Sujto, History of Mindfulness, 81-4. Stras are arranged according to the order found in the Pali.
Parallels found outside of the respective Drgha tradition are given in parentheses.

426
 In spite of all their differences in terms of arrangement and content, the three

versions of the Drghgama at our disposal do bear one conspicuous similarity. As

already noted, both the Theravda and the Sarvstivda versions are divided into

three sections, but only one of these three bears the same name in the two:

Slakkhandha in Pali and laskandha in Sanskrit. This is not simply a similarity of

name, however; all of the suttas of the Slakkhandha Vagga in Pali have parallels

specifically in the laskandha Nipta in Sanskrit and not elsewhere. Now, even

though the laskandha Nipta is somewhat bigger than the Slakkhandha Vagga,

including eight stras that do not have parallels in the Dgha Nikya, this situation is

in stark contrast to the other two niptas, which cannot be correlated to the other two

vaggas of the Pali tradition in any way. In the case of the Dharmaguptaka

Drghgama, varga divisions have not been preserved in the Chinese translation, but

the parallels to ten of the thirteen suttas of the Pali Slakkhandha Vagga (the

exceptions are DN 6, 7, and 10, which as already noted, are not found in T. 1 at all)

can be found grouped together near the end as stra nos. 20-29. 812 This strongly

suggests that at one time these stras at the end of T. 1 formed their own division in

the Dharmaguptaka version as in the Theravdin, possibly called laskandha.

Moreover, this remarkable point of consistency between the three versionsthe

Slakkhandha Vagga of the Theravdin, the hypothetical laskandha division of the

Dharmaguptaka, and the laskandha Nipta of the Sarvstivdinmakes it quite

likely, as Sujto has suggested, that this group of stras formed the original core of

812
The 30th and final stra of T. 1, the Shj Jng, has no parallel in the Pali Canon.

427
the Drghgama, to which different traditions added various other long stras. 813

Given the relative similarities and dissimilarities of the three versions, it would

appear that the laskandha core was in existence before the split of the

Sarvstivda and Vibhajyavda, and that which other long stras to include in the

collection, but not their arrangement, was settled at some point between then and

whenever the Vibhajyavda split into different schools.

That the stras of the laskandha formed from early times, indeed most

likely from the time of their origin, a single unit is supported by the common internal

structure of those stras. Unfortunately, examination of the Sanskrit versions of these

stras must await publication of the Sarvstivda Drghgama manuscripts, but at

least in the case of the Pali and Chinese, all of the relevant stras include, in some

form or another, a particular formula that I refer to as the Tathgata arises formula.

This formula differs somewhat, as we would expect of oral traditions, between the

Dharmaguptaka and Theravda, 814 but in both cases it consists, in its fully developed

form, 815 of a long sermon detailing how a Tathgata arises in the world and teaches

the dharma, and then a person goes forth and trains under his dispensation,

813
Sujto, History of Mindfulness, 86.
814
For a comparative study of these two, as well as other known, versions of the formula, see
Meisig, Das rmayaphala-Stra, 39-80.
815
In most cases, the formula culminates with the monk attaining Awakening, although in the
Tevijja Sutta, it culminates in him attaining the four brahmavihras. More significantly, however,
only a small portion of the formula, specifically the Tathgatas teaching on morality (sla), is found in
the Brahmajla Sutta, which appears first in the Pali Dgha. The arrangement found in the Pali version
seems to suggest that this teaching on morality came first and then was borrowed by the next sutta, the
Smaaphala Sutta, to form a more extended sermon that then served as the prototype for the formula
as used in the remaining eleven suttas.

428
ultimatately attaining Awakening. We have already seen this formula at play above,

in our comparison of the Tevijja Sutta (DN 13) and its Chinese parallel (T. 1, no. 26),

where it is adapted, somewhat idiosyncratically, to explain how one attains union

with Brahm.

Because these laskandha stras share a common theme and structure, Sujto

has speculated that perhaps they had their origin in the Sayuktgama and were

moved out to form their own collection because of their excessive length. He

suggests that they may have come from the ik Sayukta, which still exists in the

Sarvstivdin Sayuktgama and retains short discourses on the three-fold training.

There is, however, no corresponding Sikkh Sayutta in the Pali Sayutta Nikya,

and most of the short stras found in the Sarvstivdin sayukta are instead found in

the Aguttara Nikya; this may indicate, Sujto argues, that while the Sarvstivdins

only moved the longest stras out of their ik Sayukta, the Theravdins simply

did away with their (hypothetical) Sikkh Sayutta altogether and moved all of its

constituent suttas elsewhere. 816 Although Sujtos suggestion is fairly speculative, at

the very least we can conclude that there is a discernible core to the Drghgama

tradition, formed around a common theme expressed through a common formulaa

technique that is strongly reminiscent of the Sayuktgama tradition, thus providing

preliminary support of the Vastusagrahas assertion that the Sayuktgama is the

basis for all the gamas.

816
Sujto, History of Mindfulness, 86.

429
IV.2.4 Majjhima Nikya / Madhyamgama

Within the Madhyamgama tradition, two complete versions have come down

to usthe first, of course, is the Pali Majjhima Nikya of the Theravda school, and

the second is a Chinese translation of a Madhyamgama preserved at T. 26, entitled

the Zhng hn Jng (, lit., Middle gama Stra). The Zhng hn Jng

was translated from 396 to 397 CE by a translation team led by the Kashmirian monk

Gautama Saghadeva (Qtn Sngjitp ); the Indian original was

read out loud by another Kashmirian monk named Sagharaka (Sngjiluch

) and transcribed by the monk Doc (), with the help of Lbo () and

Knghu (). 817 According to the Chinese records, this translation of the

Madhyamgama, which consists of 60 jun, replaced an earlier, incomplete

translation in 59 jun by Dharmanandin, which is now lost. Scholars generally

believe that the version of the Madhyamgama translated in T. 26 is of the

Sarvstivda school. 818

Consistent with what we have seen in the case of the Sarvstivda version of

the Drghgama, the contents and structure of the Zhng hn Jng are quite different

from those of its Theravda counterpart. The Pali Majjhima Nikya contains a total

of 152 suttas, which are divided into 15 vaggaseach vagga has exactly 10 suttas,

except for the Vibhaga Vagga, which has 12. These vaggas are in turn divided into

817
Anlayo, Comparative Notes on the Madhyama-gama, Fuyan Buddhist Studies 2
(2007), 33.
818
Mayeda, Japanese Studies on the Schools of the Chinese gamas, 98-9.

430
three divisions of five vaggas each called pasa (fifty)a reference to the fact

that each contains a total of 50 suttas, although this is a misnomer in the case of the

third pasa, which contains the aberrant Vibhaga Vagga with 12 suttas. The

result, therefore, is an extremely orderly collection (with the exception of course of

the two extraneous suttas of the Vibhaga Vagga) of three pasas, each containing

five vaggas, which in turn each contain 10 suttas. The Sarvstivda Madhyamgama,

as preserved in Chinese translation as the Zhng hn Jng, on the other hand, is

much larger and far less orderly. It contains a total of 222 stras divided into 18

vargas. While most of these vargas contain ten stras, seven of them contain more

than ten, in one case as many as 25. The Chinese translation does not organize these

vargas per se into larger divisions, but it does separately divide the stras of the

collection as a whole into five days of recital. These days of recital do not respect

varga divisions; for example, the first day of recital ends with stra number 64,

which is in the middle of the sixth varga. 819

There has been, relatively speaking, a fair amount of research devoted to

comparative studies of the Chinese Madhyamgama and Pali Majjhima Nikya, more

than any other Nikya/gama other than perhaps the Sayuktgama/Sayutta

Nikya. In the 1960s, the Vietnamese monk Thich Minh Chau published in Vietnam

a comparative study of these two versions of the Madhyamgama tradition, although

it did not become readily accessible internationally until it was reprinted by Motilal

819
Thich Minh Chau, The Chinese Madhyama gama and the Pli Majjhima Nikya: A
Comparative Study (Delhi: Motilal Banarsidass, 1991), 45-49.

431
Banarsidass in 1991. This was not only the first comparative study of the Majjhima

Nikya and its Chinese counterpart, but the first comparative study of any Nikya and

its Chinese counterpart in their entireties. 820 Anlayo Bhikkhu has emphasized the

monumental contribution Thich Minh Chaus work has made to comparative

Nikya/gama studies, but at the same time, he has criticized him for placing what he

sees as undue emphasis on sectarian differences between the Madhyamgama and

Majjhima Nikya, preferring to understand the differences between the two versions

as reflections of their orality, an aspect of the Nikya/gama tradition that was little

understood when Thich Minh Chau wrote his study. 821 Anlayo himself has recently

published his own mammoth two-volume study of the Majjhima Nikya in

comparison to Chinese parallels. 822 Although this study is limited by its Pali-centric

perspective (i.e., it does not cover stras found in the Chinese Madhyamgama that

lack parallels in the Majjhima Nikya), this is more than made up for by the

exhaustive comparative analysis of all 152 suttas found in the Majjhima Nikya; in

particular, Anlayo looks at all of the Chinese parallels to the suttas found in the

Majjhima Nikya, not only those that are found in the Madhyamgama.

Although most of Anlayos study focuses on individual suttas, and thus we

will refer to it in connection with the encounter dialogs found in the Majjhima Nikya

820
Thich Minh Chau, The Chinese Madhyama gama and The Pli Majjhima Nikya. On
this studys publication history, see Anlayo, The Chinese Madhyama-gama and the Pli Majjhima-
NikyaIn the Footsteps of Thich Minh Chau, The Indian International Journal of Buddhist Studies
9 (2008), 1.
821
Anlayo, The Chinese Madhyama-gama and the Pli Majjhima-Nikya.
822
Anlayo, Comparative Study of the Majjhima-Nikya.

432
later in Part IV, Anlayo does offer several significant insights into the structure of

the Pali version that are worth mentioning at this point because they are relevant to

our understanding of the development of the Nikya/gama tradition. As already

mentioned, the Pali Majjhima Nikya is quite orderly in the organization of its

(approximately) 150 suttas into (mostly) equally-lengthed vaggas and fifties.

Anlayo has noted that this latter division into three pasas is not simply

conventional, but appears to reflect a conscious organization of the suttas of the

collection according to their purpose. He notes that the first and third fifties consist

mostly of discourses delivered to groups of monks, while the second fifty consists

mostly of dialogs between the Buddha and a variety of interlocutors. The second

fifty also stands out because it incorporates more verse material than the other two

fifties, and because its vaggas are organized and named after the particular

audience to which the suttas therein are addressed (householders, monks, wanderers,

kings, and Brahmans, respectively). In addition, Anlayo sees an overall progression

over the course of the three pasas from basic to more detailed teachings. He

argues that this is consistent with the testimony of the commentary, which says that

monks specializing in the Majjhima began with the first pasa and then, according

to their abilities, proceeded to the other two in succession. 823 Taking a somewhat

different approach, Bhikkhu Sujto has argued, following Anlayos observations on

the differences between the pasas of the Majjhima, that the three fifties

represent the three core agassutta by the basic teachings in the first pasa,

823
Ibid., vol. 1, 1-5.

433
geyya by the verse-heavy dialogs in the second pasa (which also happen to be

organized according to audience as in the Sagth Vagga of the Sayutta Nikya),

and veyykaraa by the more detailed teachings of the third pasa. 824

With respect to the comparison of the structures of the Pali Majjhima Nikya

and Chinese Madhyamgama, Anlayo has mostly emphasized their dissimilarity.

According to his reckoning, there are only 96 stras that are shared by the two

collections, although most of the others in each collection have parallels in another

Nikya or gama. The organization of the stras in each version, moreover, bears

little evidence of any direct relationship. Anlayo finds only four cases of

vaggas/vargas that have similar names between the two versions, and with the

exception of the Gnpn Fnbi Pn (, Skt. Mlavibhaga Varga 825),

which shares a remarkable nine stras in common with its Pali parallel, the Vibhaga

Vagga, these parallel vaggas/vargas do not share the majority of their stras in

common. The Wng Xingyng Pn (, Skt. Rja Sayukta Varga) is similar

in name to the Rja Vagga, but shares only two stras with it. Likewise, the Fnzh

Pn (, Skt. Brhmaa Varga) and Shung Pn (, Skt. Yamaka Varga) are

similar in name to the Brhmaa Vagga and Mahyamaka Vagga, respectively, but

share only four stras in common with them in each case. 826

824
Sujto, History of Mindfulness, 78-9.
825
In giving Sanskrit equivalents of Chinese varga () titles, I follow the reconstructions
given by Minh Chau, Chinese Madhyama gama, 331-3.
826
Anlayo, Comparative Study of the Majjhima-Nikya, 7-8.

434
 There is, however, much more that we can say about the structures and

contents of the Theravda and Sarvstivda versions of the Madhyamgama in

comparative perspective. In an article, Anlayo provides a table in which he lists the

Majjhima Nikya parallels to the stras of the Sarvstivda Madhyamgama in the

order they are found in the vargas of the latterwith the intent of showing that the

organization of stras in the two collections is radically different. 827 While the

differences between the Madhyamgama and the Majjhima Nikya are obvious at

first glance, even with this simple chart certain patterns can be discerned; in particular,

stras with parallels in the Majjhima Nikya tend to be found in particular vargas of

the Madhyamgama, and most such vargas are found towards the end, rather than the

beginning, of the collection. Further patterns can be discerned by creating more

complete tables of all stra parallels between the Theravda and Sarvstivda

traditions (not just those found in the Madhyama/Majjhima), from the perspective of

both the organization of the Madhyamgama and the organization of the Majjhima

Nikya. In other words, since we have the four main Nikyas available to us in the

Theravda tradition and three of the four gamas (Sayukta, Drgha, and Madhyama)

available to us in the Sarvstivda tradition, we can chart the ways in which these two

broad traditions intersect within the specific context of the two traditions of the

Madhyama/Majjhima.

Let us begin by charting all Sarvstivda parallels to the suttas of the

Majjhima Nikya, within the organizational context thereof. This can be found in

827
Anlayo, Comparative Notes on the Madhyama-gama, 35-6.

435
Table 4. As can be easily seen, the vast majority of suttas in the Majjhima Nikya

have parallels in the Sarvstivdin Madhyamgama. In order to bring attention to

those that do not, suttas that have parallels in the Sayuktgama are highlighted in

blue, those that have parallels in the Drghgama are highlighted in green, and those

that have no known Sarvstivda parallel are highlighted in gray. Immediately one

vagga stands outthe Sayatana Vagga, whose suttas each have parallels, without

exception, in the Sarvstivdin Sayuktgama. This vagga, I would argue, is

suspicious to begin with because it has a conspicuously sayutta-like title,

Sayatana (the six sense bases), which in fact is a prominent theme even in the Pali

Sayutta Nikya and is the title of one of its vaggas. Not only does the Sarvstivda

tradition preserve all of these stras in the Sayuktgama, but it preserves all but the

first, second, and ninth among them in a single sayuktathe Rch Xingyng (

, Skt. yatana Sayukta). This strongly suggests that the Theravda tradition

borrowed these suttas from the Sayutta Nikya, perhaps to fill out a third set of

fifty in its remarkably symmetrical organizational scheme. Since most of these

stras are preserved by the Sarvstivda in a single sayukta, it is possible that the

Theravda tradition created the Sayatana Vagga of its Majjhima Nikya by

borrowing a single sayutta from the Sayutta Nikya, whose contents were close to,

but slightly different from, the Rch Xingyng of the Sarvstivdin Sayuktgama

as it has come down to us in the Z hn Jng.

Further patterns can be seen by highlighting stras of a single varga in the

Madhyamgama that appear close to one another in the Majjhima Nikya. The most

436
obvious pattern that emerges in this way is one that Anlayo has already brought

attention tothe parallel between the Vibhaga Vagga of the Pali and the Gnbn

Fnbi Pn (Mlavibhaga Varga) of the Chinese. In order to illustrate this, I have

highlighted all suttas that have parallels in the Gnbn Fnbi Pn in yellow. Nine of

these are found in the Vibhaga Vagga, and the tenth is found just two vaggas earlier,

in the Anupada Vagga. This certainly suggests, as Anlayo has argued, that the

Vibhaga Vagga of the Pali and the Gnbn Fnbi Pn of the Chinese have a

historical relationship to one another.

437
438
 While Anlayo restricted himself to single vaggas and vargas that have

similar titles, another interesting pattern can be found by ignoring the vagga/varga

titles and also looking across vagga/varga boundaries. In the first two vaggas of the

Majjhima Nikya, I have highlighted in orange those suttas that have parallels in the

Hu Pn ( 828) and in red those that have parallels in the Yn Pn (, Skt.

Nidna Varga) of the Sarvstivdin Madhyamgama, which are vargas eight and

nine of that collection, respectively. Immediately we see that eight of the ten stras

of the Yn Pn are clustered together in these first two vaggas of the Majjhima Nikya.

The other two, M 97 and 104, are not found in these two vaggas or anywhere else in

the Majjhima Nikya. In fact, these two stras are found in the Pali Dgha Nikya;

that is, they are among the twenty stras (i.e. ten in each direction), discussed in

Section IV.2.3 above, that appear to have been swapped by either the Theravda or

the Sarvstivda tradition between the Madhyama/Majjhima and Drgha/Dgha.

Indeed, in remarkable confirmation of this hypothesis, we find that although these

first two vaggas of the Majjhima Nikya lack parallels to two and only two stras of

the Yn Pn, which are instead found in the Dgha Nikya, we do find two parallels

here to stras in the Sarvstivdin Drghgama, which also happen to be among the

twenty stras swapped at some point between the Madhyama/Majjhima and

Drgha/Dgha, but in this case in the opposite direction.

Turning now to the suttas highlighted in orange, we find that in the first two

vaggas of the Majjhima Nikya there are parallels to five stras found in the Hu Pn

828
Minh Chau gives agaa (yard or courtyard) as the Sanskrit equivalent for (filth),
but I am uncertain of the reasoning behind this reconstruction.

439
of the Sarvstivdin Madhyamgama. The other five stras of the Hu Pn have no

parallel at all in the Majjhima Nikya, but instead are preserved by the Theravda

tradition in the Aguttara Nikya. Moreover, they are preserved there specifically in

the Book of the Tens (Dasaka Nipta). In sum, then, we have in the first two vaggas

of the Majjhima Nikya parallels to all of the stras in the Sarvstivdin Yn Pn,

except for two that were at some point by one or the other tradition traded for two

stras in the Drgha/Dgha, and we have parallels to half of the stras in the

Sarvstivdin Hu Pn, the other half of which are preserved by the Theravda in the

Dasaka Nipta of the Aguttara Nikya. This strongly suggests that there is a

historical relationship between the first and second vaggas (Mlapariyya and

Shanda) of the Theravdin Majjhima Nikya on the one hand and the eighth and

ninth vargas (Hu and Yn) of the Sarvstivdin Madhyamgama on the other. In an

earlier stage of the tradition, there existed two vargas, which were then rearranged

into two new vargas by a later tradition, with certain substitutions. It is not

immediately clear, however, which of the two versions that have come down to us,

the Theravdin or the Sarvstivdin, is the more original, and which represents the

rearrangement.

As for those vaggas/vargas that Anlayo flags as having a possible historical

relationship on the basis of similar name, aside from the clear relationship between

the Vibhaga Vagga and the Gnbn Fnbi Pn already discussed, the relationships

between these parallel vaggas/vargas are not necessarily well-supported by the

actual sutta/stra parallels within them. The most likely parallel is between the

440
Mahyamaka Vagga, the fourth vagga of the Majjhima Nikya, and the Shung Pn,

which is the 15th varga of the Madhyamgama. The Mahyamaka Vagga has only

four parallels to stras in the Shung Pn, which I have highlighted in brown, but it is

actually itself paired with the fifth vagga of the Majjhima Nikya, entitled

Cayamaka Vagga. Looking here, we find one more parallel to a stra in the

Shung Pn, also highlighted in brown. Four more parallels to stras in the Shung

Pn are found in the Majjhima Nikya, but much later in the Anupada and Suata

Vaggas of the third pasa (again, highlighted in brown). Now, it is possible that

the Shung Pn (Skt. Yamaka Varga) of the Sarvstivda version represents an early

varga that was then split up in the Theravda tradition into the Mahyamaka and

Cayamaka Vaggas. If this theory is correct, however, we must assume not only that

five extra pairs (yamaka) of suttas were added the fill in the additional space

provided by a second vagga, but that half of the stras in the original Yamaka Varga

were moved out of the yamaka vaggas altogether and replaced with others. As we

can see from Table 4, several of the suttas in the Mahyamaka and Cayamaka

Vaggas do not even have parallels in the Sarvstivdin Madhygama, but in other

Sarvstivdin collections, primarily the Sayuktgama. Another, perhaps simpler

explanation for this state of affairs is that the Theravda tradition developed its two

yamaka vaggas and the Sarvstivda tradition its one Yamaka Varga independently.

Stras often come in pairs in the early Buddhist tradition, and therefore it is a fairly

obvious choice to create a varga filled with pairs (yamaka) of stras. If they were

drawing from a similar corpus of stras, in which only a certain number of stras

441
came in pairs, two traditions could easily each develop their own Yamaka Varga(s)

independently that nonetheless have a certain amount of overlap.

Similar arguments can be made for the Rja and Brhmaa Vaggas (9 and 10),

which are ostensibly parallel to the similarly named Wng Xingyng Pn (Rja

Sayukta Varga) and Fnzh Pn (Brhmaa Varga) in the Sarvstivdin version. As

already mentioned, there are only two stras in parallel between the Rja Vagga and

Wng Xingyng Pn, so it is hardly even worth mentioning as a potential parallel on

this basis alone. Moreover, as we will see shortly, the use of the word xingyng

(sayukta) in the Chinese varga title suggests a completely different origin for the

Sarvstivda varga entirely. The Brhmaa Vagga and Fnzh Pn, on the other hand,

share exactly the same name, but they share only four stras in common. As we will

see in Chapter IV.3, it is very common throughout the early Buddhist tradition to

group stras that involve Brahmansi.e., encounter dialogstogether, and generally

speaking these groupings are entitled simply Brhmaa. As in the case of the

yamaka vargas examined above, it is entirely conceivable that the Theravda and

Sarvstivda traditions independently created vargas to collect stras addressed to

Brahmans, and if they did so, then drawing from a common corpus of stras, they

would have inevitably chosen some of the same stras to include therein.

One other observation about the context of the Brhmaa Vagga in the

Majjhima Nikya also suggests an independent origin for the Brhmaa Vargas of the

Theravda and Sarvstivda versions. The Brhmaa Vagga is the last vagga in the

second pasa (Majjhimapasa), and as already discussed above, this pasa

442
has certain unique featuresi.e., a preponderance of suttas with verses and a vagga-

organization scheme based on the audience to which suttas are addressed

(householders, monks, wanderers, kings, and Brahmans)that make it similar to the

Sagth Vagga of the Sayutta Nikya. On the other hand, looking again at Table 4,

the second pasa is, of the three, the one with the least parallels, particularly in the

Sarvstivdin Madhyamgama, but also in the three extant Sarvstivdin gamas in

general. I would argue that this indicates that the second pasa of the Majjhima

Nikya, as an organizational scheme, is unique to the Theravda tradition, and was

created in imitation of the organizational scheme of the Sagth Vagga in the

Sayutta Nikya. We have already commented that the Majjhima Nikya, especially

in comparison to the Sarvstivdin Madhyamgama, is exceptionally well organized,

and it appears that the organization of the second pasa into vaggas according to

audience was part of this purposeful organizational scheme. If this is the case, then

given that there is absolutely no sign of a similar organizational scheme in the

Sarvstivda version, it is likely that the Theravdins created their own Brhmaa

Vagga independently of the Sarvstivdin Brhmaa Varga, since Brahmans would

have represented one of the most important audiences to which a sutta could be

addressed.

Indeed, more compelling patterns can be found within the Majjhimapasa

if we once again ignore vagga titles and boundaries and just look for suttas in fairly

close proximity that come from the same Sarvstivdin varga. Although the Rja

Vagga (9) has only two parallels in the similarly named Wng Xingyng Pn, it has

443
four parallels in the Bldu Pn ( 829), which is the 17th and second-to-last

varga in the Sarvstivda version. These parallels are particularly conspicuous

because they appear together in this Pali vagga. In addition to these four, a fifth

parallel from the Bldu Pn is found fairly close by, in the Bhikkhu Vagga (7), and a

sixth in the Gahapati Vagga (6). I have highlighted all of these parallels in pink.

Likewise, five parallels to stras in the 18th and final varga of the Sarvstivdin

Madhyamgama, the L Pn ( 830), are found scattered through the first three

vaggas of the Majjhimapasa. Three of these are found in the Paribbjaka Vagga

(8) and one each in the Gahapati and Bhikkhu Vaggas. These are highlighted in

purple. Although these patterns are not among the most convincing of those I have

found in the Majjhima Nikya, they are certainly more convincing than parallels

based primarily on name similarities, and they may suggest some sort of historical

relationship between the last two vargas of the Sarvstivdin Madhyamgama and

the Majjhimapasa of the Majjhima Nikya.

Let us now turn to a comparison of the Theravda and Sarvstivda traditions

from the opposite perspectivethat is, a chart of all Theravda parallels to stras in

the Sarvstivdin Madhyamgama, shown within the latters organizational context.

This can be found in Table 5. Again, most parallels are to suttas in the Majjhima

Nikya, but to bring attention to those that are not, I have highlighted parallels to

829
Minh Chau gives potaliya as the Sanskrit equivalent of , but I am unable to find
attestation for such a Sanskrit word.
830
Minh Chau gives savidahana as the Sanskrit equivalent of , but I am unable to find
attestation for such a Sanskrit word.

444
suttas in the Dgha Nikya in green, those to suttas in the Sayutta Nikya in blue,

those to suttas in the Aguttara Nikya in orange, and those without any Theravda

parallel in gray. Immediately, this color scheme brings attention to an obvious

patternthe overall preponderance of stras that do not have parallels in the

Majjhima Nikya have them instead in the Aguttara Nikya. As already noted

earlier, there are fewer stras with Majjhima Nikya parallels in the earlier vargas

than in the later vargas; since Aguttara Nikya parallels tend to fill in these gaps, we

find a large number of Aguttara Nikya parallels especially in the earlier vargas, but

also in later vargas that lack many Majjhima Nikya parallels, such as the 11th and

largest varga, the D Pn (, Skt. Mah Varga).

It should be noted that the large number of Theravda Aguttara suttas found

in the Sarvstivdin Madhyamgama is not reciprocated by a large number of

Sarvstivda Ekottarika stras in the Theravdin Majjhima Nikya. Although we

cannot know this directly, since, as far as we know, no Sarvstivdin version of the

Ekottarikgama has come down to us, 831 we can infer as much because, looking back

at Table 4, we can see that most of the suttas of the Majjhima Nikya are already

accounted for by parallels in the three Sarvstivdin gamas that are available to us.

It is of course possible that some of the suttas highlighted in gray as having no known

Sarvstivda parallel in fact do have a parallel in the now-lost Sarvstivdin

Ekottarikgama, but the number of possible candidates is quite few. Moreover, it is

quite likely that several of these candidates truly do represent Theravda suttas

831
See Section IV.2.5.

445
without any Sarvstivda parallel whatsoever, just as the stras highlighted in gray in

Table 5 most definitely (because the Theravda Canon is complete) represent

Sarvstivda stras that are simply unique to the Sarvstivda tradition and do not

have any Theravda parallel whatsoever.

What, then, are we to make of the fact that there are so many stras

(approximately 80) that are preserved by the Theravda tradition in the Aguttara, but

by the Sarvstivda tradition in the Madhyama? Before we answer this question, it is

useful to consider a broader issue concerning the early vargas of the Sarvstivdin

Madhyamgama that are sparse in parallels to suttas in the Majjhima Nikya. Above,

in the context of the Theravdin Majjhima Nikya, I argued that the final vagga,

entitled Sayatana, was suspicious because of its very sayutta-like title, and that its

likely origin in the Sayutta Nikya is confirmed by the fact that all of its suttas have

parallels in the Sarvstivdin Sayuktgama, most in a single sayukta. All of the

first six vargas of the Sarvstivdin Madhyamgama can be said to be suspicious for

similar reasons.

446
447
 Consider first the very first varga, which is entiled Q F Pn (, Skt.

Saptadharma Varga). The title is at first glance perhaps more confusing than

suspicious, but its true significance becomes clear when one looks at the parallels to

its stras in the Theravda tradition. All but three of its stras have parallels, not

surprisingly, in the Aguttara Nikya, but more significantly, all of these seven

parallels are found specifically in the Sattaka Nipta, or Book of the Sevens. The

Aguttara Nikya and its parallels in other traditions organize stras according to the

number of dharmas discussed within them; clearly, then, the significance of the title

Saptadharma is that all of the stras in this varga deal with sets of seven dharmas,

just as one might find in an Aguttara or Ekottarika collectionand, in fact, the

Theravda does preserve most of these stras in its Aguttara collection! In addition,

all but one of these stras have parallels not only in the Theravdin Aguttara Nikya,

but also in the Ekottarikgama preserved in Chinese, 832 which may belong to the

Mahsghika school. As we will discuss further in Section IV.2.5, the Chinese

Ekottarikgama is extremely different from the Pali Aguttara Nikya, and so it is

clear that many of the suttas found in the Pali Aguttara Nikya are classified there

uniquely by the Theravda school (and possibly some other closely related schools

whose Aguttara/Ekottarika collections are now lost), and moreover that the early

common core lying behind the Aguttara/Ekottarika traditions must, especially if the

attribution of the Chinese Ekottarikgama is correct, be restricted to the rather small

832
M 1=AN 7.64=E 39.1, M 2=AN 7.65=E 39.2, M 3=AN 7.63=E 39.4, M 4=AN
7.15=E 39.3, M 5=AN 7.68=E 33.10, M 8=AN 7.62=E 40.1. In addition, the two stras of this
varga whose Theravda parallels happen to be in the Majjhima Nikya also have parallels in the
Chinese Ekottarikgama: M 9=MN 24=E 39.10, M 10=MN 2=E 40.6.

448
number of stras held in common between the Pali Aguttara Nikya and the Chinese

Ekottarikgama. It is remarkable, then, that most of the stras of the Q F Pn have

parallels in both the Pali Aguttara Nikya and the Chinese Ekottarikgama and

therefore likely go back to a very early Ekottarikgama tradition. It should be further

noted that while this is true for all of the Aguttara Nikya parallels in the Q F Pn,

it is not true for Aguttara Nikya parallels in the Sarvstivdin Madhyamgama in

general; that is, most of the stras of the Sarvstivdin Madhyamgama that have

parallels in the Aguttara Nikya do not have parallels in the Chinese Ekottarikgama

and therefore do not appear to go back to the early Ekottarikgama tradition. For

these reasons, I believe it is extremely likely that the Q F Pn was taken from the

Ekottarikgama, just as the Sayatana Vagga of the Majjhima Nikya was likely

taken from the Sayutta Nikya.

Turning now to vargas two through six, we find that four (2, 3, 5, and 6) are

immediately suspicious because their titles include the word xingyng (), which

is the Chinese translation for sayukta. Unfortunately, there is no way to prove that

these vargas came from an earlier version of the Sayuktgama, but the use of the

word sayukta in the title is highly suggestive, and the lack of many parallels to the

Majjhima Nikya makes it unlikely that these vargas have a particularly old lineage

within the Madhyama tradition. A few stras in these vargas in fact do have parallels

in the Pali Sayutta Nikya, but most have parallels in the Aguttara Nikya. This is

especially true of the X Xingyng (, Skt. Samudaya Sayukta Varga), all of

449
whose stras save two (which have no Theravda parallels at all) have parallels in the

Aguttara Nikya.

We may now, then, have a solution to the question we asked earlier, which is

why there are so many stras preserved by Sarvstivda in the Madhyamgama, but

by the Theravda in the Aguttara Nikya. Perhaps certain stras that had their

origin in the Sayuktgamawhich, as we have seen, the Vastusagraha reckons

as the root of all the gamaswere removed for some reason from that collection

(perhaps because of excessive size) during the early development of the Sthaviravda

tradition, but then reclassified differently by the different sub-branches of that

tradition. The Theravdins (or perhaps the Vibhajyavdins writ large) used them to

fill out their Aguttara Nikya, while the Sarvstivdins used them to fill out and add

vargas to their Madhyamgama. This would explain both why the Theravdin

Aguttara Nikya contains so many suttas that are found in the Chinese

Madhyamgama (see Section IV.2.5) and why the Sarvstivdin Madhyamgama is

so much bigger than the Theravdin Majjhima Nikya.

Of course, not all Sarvstivdin Madhyamgama stras with parallels in the

Theravdin Aguttara Nikya are found in vargas that include the word sayukta in

the title. But what I am suggesting is that the use of the word sayukta in varga titles

might represent entire sayuktas that were taken from the Sayuktgama and

transformed into Madhyamgama vargas. Other stras may have been taken

individually from the Sayuktgama and used to expand previously existing vargas.

In addition, another varga that includes a preponderance of parallels to the Aguttara

450
Nikya stands out conspicuously even though it does not include the word sayukta

in its title. This is the 11th varga, the D Pn (Mah Varga), which is unusual for

three reasons. First, it is extremely large, with a total of 25 stras. Second, in spite

of its large size, it only includes three stras with parallels in the Majjhima Nikya.

Finally, its name appears to be redundant, insofar as the 16th varga also is entitled

Mah Varga (rendered in Chinese as ), but has the standard number of ten

stras, almost all of which have parallels in the Majjhima Nikya. It is possible that

this latter, more normal Mah Varga is more original to the Madhyama tradition,

and that the Sarvstivdins created a second Mah Varga at varga 11 to serve as a

dumping ground for extra stras being introduced from the Sayuktgamastras

that the Theravdins, once again, instead mostly found places for in their Aguttara

Nikya.

There is one other possible obstacle for the theory that certain stras

originating in the Sayuktgama were placed in the Madhyamgama by the

Sarvstivdins and in the Aguttara Nikya by the Theravdins. This is the Wng

Xingyng Pn (6), which, despite having the word sayukta in the title, does not

contain any stras with parallels in the Theravdin Aguttara Nikya. Nevertheless,

hypothesizing an origin in the Sayuktgama is far from ridiculous, since it does

contain two stras with parallels in the Theravdin Sayutta Nikya itself. Moreover,

this varga is unique in that it contains a full four stras with parallels in the Dgha

Nikyanearly half of the ten that participated in the swap between the

Drgha/Dgha and Madhyama/Majjhima at some time during the development of the

451
Theravdin and Sarvstivdin traditions. It seems likely, therefore, that this varga is

somehow related to this event, which may, for all we know, have involved the

movement of two sayuktas out of the Sayuktgama, one into the long collection

and one into the middle collection, but oppositely so by the Sarvstivdins and

Theravdins. 833

Finally, returning to the question of the first six vargas of the Sarvstivdin

Madhyamgama, we are left with varga 4, which does not include the word sayukta

in its title, but is nonetheless unusual. It is entitled Wicngyuf Pn (),

which is a translation of Adbhutadharma Varga. Adbhuta is one of the nine (or

twelve) agas, and thus its use as the title of a varga is at the very least quite

interesting. That it is of further significance is confirmed by the fact that six of its

stras have parallels in the Aguttara Nikya, all of them in the Book of the Eights

(Ahaka Nipta). This suggests that this group of stras was considered as a unit

from an early date in the tradition. In addition, all of the Pali suttas that are parallel to

stras in this varga clearly have to do with abbhut dhamm, 834 i.e., marvelous

qualities, suggesting that this is indeed a collection of the adbhutga. It is not

immediately clear where this collection of adbhuta stras came from, but given that

most of the stras within it are not preserved by the Theravda in the Majjhima

Nikya, that they nevertheless form a logical and likely antique unit, and that adbhuta

833
This possibility has also been suggested by SujtoSujto, History of Mindfulness, 84.
834
All of the suttas in question, as far as I have been able to discern, actually use the word
abbhuta, except for AN 8.70=M 36, which nonetheless fits within this context because it narrates the
marvelous event of an earthquake that took place when the Buddha announced that he would pass
into parinibbna.

452
is recognized by all early Buddhist traditions as an aga and therefore an autonomous

genre, it seems likely that this varga had its origins outside of the Madhyamgama

as a set of texts either representative or constitutive of the adbhutga.

Summarizing what we have found through our comparative analysis of the

Theravdin Majjhima Nikya and Sarvstivdin Madhyamgama in this chapter, it is

difficult to discern a stable core to the Madhyama tradition as we were able to find

in the case of the Drgha tradition with the laskandha. There is one clear parallel

between the two versions in the form of the (Mla-)Vibhaga Vagga/Varga, and

close analysis suggests that there may be some other patterns of historical relationship

between the two versions that have been obscured by rearrangement in one or both

traditions. Overall, however, it appears that the Theravda and Sarvstivda

traditions managed their middle collections differently and for different purposes.

The Theravdins structured their Majjhima Nikya quite logicallyat least partially,

it seems, in imitation of the Sayutta Nikya (especially in the second pasa,

which is structured similarly to the Sagth Vagga)in order to use it as a

progressive training manual for monks. The Sarvstivdins, on the other hand,

appear to have used their Madhyamgama, in large part, as a dumping ground for

stras taken from elsewhere, perhaps largely from the Sayuktgama. Given that the

Theravda version appears to be unique in giving its Majjhima such a highly ordered

structure, it is possible that the Sarvstivdin practice of moving texts into the

Madhyamgama was a continuation of an earlier practice that predated the

divergence of the two traditions, and that the Theravdins ended the practice by

453
rearranging the suttas that had collected in its Majjhima so far into an ordered training

manual. In any case, the relative lack of correspondence between the two versions, as

well as the evidence of Sayukta/Sayutta influence on both, lends continued

credence to the claim of the Vastusagraha that the Sayuktgama is the root of

all the gamas.

IV.2.5 Aguttara Nikya / Ekottarikgama

In addition to the Theravdin Aguttara Nikya preserved in Pali, two

Ekottarikgama collections have survived in Chinese translation, one of which is

simply a short anthology, but the other of which is complete. The anthology, which is

preserved at T. 150a and is entitled the Q Ch Sn Gun Jng (), is

among the ealiest Buddhist texts translated into Chinese, as it was translated by

Chinas inaugural translator, n Shgo (), already discussed above in Section

IV.2.2, in the middle of the 2nd century CE. This short text, which includes 47 stras

and comprises only two jun in the Taish, has been preserved in considerable

disarray. This is not simply a matter of mixed-up jun (which would be impossible

since there are only two), but of stras being interrupted by other stras mid-sentence.

Paul Harrison has shown that by taping the text together in one continuous line and

then cutting at the three places where stras are interrupted in this way, as well as two

other locations, it is possible to reassemble the text in a logical way. The resulting

text consists of a core of 44 unnamed stras, flanked at the beginning by two, and

at the end by one, named stras. The first of these, at the very beginning of the text,

454
is entitled Q Ch Sn Gun Jng (Seven Places and Three Contemplations Stra)

and thus appears to have lent its name to the entire collection. This, and the third

named stra at the end of the text, have parallels in the Sayutta Nikya, and

therefore (along with the second named stra, which has no known parallels) may not

have come from an Ekottarikgama collection at all, but since they all do involve

numbers, they may have been included for this reason. The 44 core stras, which

do not bear titles, also all involve numbers, and moreover 36 of them have close

parallels in the Aguttara Nikya, which makes it likely that all 44 were likely taken

from an Ekottarikgama collection. Moreover, given that the majority have parallels

in the Aguttara Nikya and only 5 in the full Chinese translation of the

Ekottarikgama, to be addressed shortly, it is likely that this anthology was taken

from a version of the Ekottarikgama fairly close to the Aguttara Nikya of the

Theravdins. 835

The complete translation of an Ekottarikgama into Chinese, in 51 jun, is

located at T. 125 and is entitled Zngy hn Jng (). Zngy literally

means add one, and thus is a direct translation of ekottara, which (either in this

form or as ekottarika 836) appears to have been the standard Sanskrit title for the

gama corresponding to the Aguttara Nikya in Pali. This title refers to the fact that

835
Paul Harrison, The Ekottarikgama Translations of An Shigao, in
Bauddhavidysudhkara: Studies in Honour of Heinz Bechert on the Occasion of His 65th Birthday,
ed. Petra Kieffer-Plz and Jens-Uwe Hartmann, Indica et Tibetica, vol. 30 (Swisttal-Odendorf: Indica
et Tibetica Verlag, 1997), 261-84.
836
On the uncertainty between the forms ekottara and ekottarika, see Allon, Three Gndhr
Ekottarikgama-Type Stras, 11-12. I follow Allon in using the form Ekottarikgama because it is the
only one attested in Sanskrit.

455
this collection is organized according to a numeric principle; that is, stras within it

are organized into sections according to the number of dharmas discussed within

them. Thus, stras in the first section each discuss a single dharma; stras in the

second discuss two dharmas; those in the third, three; and so forth. According to a

preface to T. 125, the Ekottarikgama was translated into Chinese in 384-385 CE by

Zh Fnin (), based on an original read by the Tocharian monk

Dharmanandin. Some Chinese records, however, give Gautama Saghadeva as the

translator, which has led some to the theory that Gautama Saghadeva either revised

or replaced an earlier translation by Dharmanandin, as in the case of the

Madhyamgama. 837 A comparison of the translation style of T. 125 with that of T. 26

(the Madhyamgama), however, makes this unlikely. 838

Of the four complete gama translations into Chinese, the school attribution

of the Ekottarikgama remains the most controversial. Probably the most common

opinion is that it is a version of the Mahsghika. There are several reasons for this

theory, as summarized by Bhikkhu Anlayo, mostly centering on details found in an

introduction at the beginning of the text. According to this introduction, nanda

expresses hesitation before acceding to reciting the stras during the first council, a

detail found in the Mahsghika Vinaya, but not other Vinayas. After nanda

finishes reciting the stras, the earth shakes and flowers rain from the skya detail

also found in the Mahvastu of the Lokottaravda-Mahsghikas. The introduction

837
Mayeda, Japanese Studies on the Schools of the Chinese gamas, 102.
838
Anlayo, Zeng-yi A-han, 822.

456
also advises monks who have forgotten the setting of a stra to use rvast, an

allowance that is also found in the Mahsghika and Mlasarvstivda Vinayas.

Finally, the introduction includes a brief account of King Mahdeva that makes

reference to former Buddhas in such a way as in the tale of Ghakra that is found in

the Majjhima Nikya, Sarvstivdin Madhyamgama, and Mahsghika Mahvastu.

While the MN and M versions refer only to the former Buddha Kyapa, the

introduction refers to threeKyapa, Krakucchanda, and Kanakamunias does the

Mahvastu.

According to Anlayo, there are, however, some objections to the attribution

of the Chinese Ekottarikgama to the Mahsghika. First, it does not list the four

gamas in the same order as does the Mahsghika Vinaya. Second, it says that

Mahyna scriptures can be found in the Kudraka Piaka, a statement that appears to

be contradicted by the Vinaya. Third, it gives a different number of Prtimoka

precepts than are found in the Mahsghika Vinaya. Finally, the Ekottarikgama

lists twelve agas, while the Mahsghika lists only nine. Anlayo Bhikkhu has

disputed these objections, however, on the basis that similar discrepancies can be

found between other gamas and Vinayas that are known to belong to the same

school. It is very well possible that the Mahsghika Vinaya and the

Ekottarikgama date from different periods and therefore represent different stages

within the development of the Mahsghika tradition. Indeed, as Anlayo points

out, this is quite likely, since the Ekottarikgama, uniquely among all extant

457
gama/Nikya collections, incorporates clearly Mahyna elements and therefore

likely represents a fairly late stage of development. 839

Indeed, although there is a lack of clear scholarly consensus on the school

attribution of the Zngy hn Jng, an attribution of it to the Mahsghika, which

split from the Sthaviravda of which the Theravda is a branch 840 in the Great Schism

and therefore can be said to be genealogically more distant from the Theravda than

even the Sarvstivda, is consistent with the fact that its contents are markedly

different from those of the Aguttara Nikyamore so than any other extant gama

from its corresponding Pali Nikya. The difference here is not simply a matter of

accretionthe Ekottarikgama in Chinese is not simply a bigger version of the

Aguttara Nikya, produced over the centuries as more and more stras, including

eventually some with a Mahyna flavor, were addedrather, it is a matter of a

remarkable lack of parallels between the Pali and Chinese texts altogether. As I

suggested in Section IV.2.4, if the attribution of the Chinese Ekottarikgama to the

Mahsghika is correct, those stras it does hold in common with the Aguttara

Nikya may represent an early core to the Ekottarika tradition that dates to a very

early period in Buddhism, before the first Great Schism.

839
Ibid., 822-3.
840
Although the Pali word theravda used by the school of Buddhism now predominant in Sri
Lanka, Burma, Laos, Thailand, and Cambodia is completely equivalent to the Sanskrit word
sthaviravdaboth mean tradition of the eldersit is clear that the Theravdins use of this word to
refer to themselves is a conceit intended to express their embodiment of the true or original
Sthaviravda tradition. After all, there are many schools within the Sthaviravda lineage, including
even the Sarvstivdins. It appears that the term that was used by others to describe the self-styled
Theravdins was Tmraya. See Choong, Fundamental Teachings, 3. Also see Skilling,
Theravdin Literature in Tibetan Translation, 155-69, which provides an exhaustive study of attested
references to the Buddhists of Sri Lanka using Tmraya and related terms in Sanskrit and Tibetan.

458
 It is clear, however, that after that point, the two traditions expanded their

Ekottarika/Aguttara collections in very different ways. Although some of the stras

in the Chinese Ekottarikgama lack any known parallels in other traditions, most of

the stras that lack parallels in the Aguttara Nikya do have parallels in other Pali

Nikyas, especially the Sayutta and Majjhima. This would indicate that the

Mahsghika, or whichever school our Chinese text belongs to, expanded its

Ekottarikgama by moving in stras that were placed (or retained) by the

Theravdins, and most likely other Sthaviravda sects, in other collections.

Conversely, as we just saw in the context of our comparative analysis of the extant

Madhyama/Majjhima traditions in Section IV.2.4, it is clear that the Theravdins (or

their Vibhajyavda antecedents) moved a large number of stras into the Aguttara

Nikya that even the relatively closely related Sarvstivdins instead placed in their

Madhyamgama. That it was the Theravdins who moved them into their Aguttara,

and not the Sarvstivdins who moved them out of their Ekottarika, is supported by

the fact that, as already noted, most of them lack parallels in the Zngy hn Jng,

and therefore do not go back to the earliest discernible core of the

Ekottarika/Aguttara tradition.

In further support of the argument that the Theravdins actively worked to

flesh out their Aguttara Nikya, Bucknell has shown that entire vaggas in the

Aguttara Nikya were created by splitting up a sutta representing a higher number

into its constituent parts. In particular, he shows that the very first vagga of the

Book of the Ones (Ekata Nipta) was generated by taking a sutta from the Book

459
of the Fives (Pacata Nipta) and using its list of five dharmas (in this case, the five

senses) to create five different suttas, all identical except for which sense is referred

to. These five suttas were then duplicated by changing the person being talked about

from a man to a woman, resulting in ten suttas in totala complete vagga. The

hypothesis that the Theravdins simply generated the ten suttas in this way is

supported by the fact that they are paralleled by only two stras in the Chinese

Ekottarikgama (9.7, 9.8). 841 It is likely that the Theravdins felt the need to do this

to fill out the Book of the Ones, since one does not really constitute a list, and

therefore there are not many natural suttas that can be placed into this nipta to

bring it into balance with the other niptas of their collection.

Unfortunately, there are virtually no studies that have been done comparing

the Chinese Ekottarikgama and Pali Aguttara Nikya to one another as a whole. If

the attribution of the former to the Mahsghika is correct, then such comparative

work could provide valuable insights into the early stages of the Buddhist oral

tradition. What does seem clear from what we have presented here, however, is that

although the idea of organizing stras according to a numerical principle is likely

quite old, the actual content of the collections of oral texts based on this principle

developed quite slowly, and diversely across the various sectarian traditions. It

appears that various traditions chose different texts to fill out their

Ekottarika/Aguttara collections, and even generated stras formulaically to do so as

well. In doing so, they created versions of this gama that differed from one another

841
Bucknells argument here is as yet unpublished, but it has been cited and summarized in
Sujto, History of Mindfulness, 89-90.

460
most likely more than in the case of any other gama. This can be seen directly in

the case of the Theravda and the (likely) Mahsghika collections, but it can also

be inferred that the Sarvstivdin Ekottarikgama differed greatly from the Aguttara

Nikya on the basis that so many AN suttas have parallels in the Sarvstivdin

Madhyamgama. It therefore seems that the Ekottarikgama was, more so than any

other gama, less a body of particular stras than a principle for organizing new

stras or stras that otherwise lacked a home.

IV.2.6 Conclusion

Comparative gama/Nikya studies is still in its infancy, and there is still

much work that can be done to clarify the development of the Nikya/gama

traditions, which unfortunately falls outside the scope of this dissertation. From the

previous scholarship and additional evidence I have presented in this chapter,

however, I believe that there is a fair amount of support for the contention in the

Vastusagraha that the Sayuktgama is the root of all the gamas. We are lucky

to have available to us Sarvstivda versions of three of the four gamas, and of

these, the one with the greatest similarity to its Theravda counterpart in structure and

content is the Sayuktgama. The Drgha/Dgha tradition appears to have an early

core in the laskandha/Slakkhandhawhich itself may have had its origins in the

Sayuktgamabut the rest of the collection was in flux, apparently a dumping

ground for increasingly long stras. The Madhyama/Majjhima collection may

have had a core, at least within the Sthaviravda tradition, but this has been mostly

461
lost to us as the Theravdins apparently rearranged it for pedagogical purposes and

the Sthaviravdins used it as a dumping ground for stras that the Theravdins

instead dumped in their Aguttara Nikya. The latter, in fact, appears to have been

very much in flux among the various sectarian traditions, less a body of specific

stras than a principle for arranging them. Although these conclusions are somewhat

broad and tentative, they will at least be able to serve as guideposts as we turn

specifically to the encounter dialogs that were created and transmitted within these

gamic and sectarian branches of the early Buddhist oral tradition. It is to this genre

of stra that we turn in the next two chapters.

462
 Chapter IV.3

Collections of Encounter Dialogs

IV.3.1 Introduction

Now that we have addressed the matter of orality in the production and

dissemination of early Buddhist stras, as well as the utility of comparative

Nikya/gama studies in discerning the development of the early Buddhist oral

tradition, we can finally turn to the primary subject of Part IVthe encounter dialogs

in which the Buddha is depicted as encountering interlocutors who are identified as

Brahmans. As already indicated earlier, encounter dialogs are often found, at least in

the Nikya collections of the Theravda school, collected together, rather than

scattered randomly among other texts. This in itself is not entirely surprising, since,

as we have seen, the purveyors of the early Buddhist tradition used various techniques

to facilitate the memorization of the oral texts in their care, including grouping texts

together by common theme. Nevertheless, while this process could have happened,

and certainly did happen, over and over again at any and all points in the development

of the early Buddhist oral traditions, we have already seen earlier in Part IV hints that

463
some collections of stras could be of considerable antiquity. In this chapter, we will

explore the various collections of encounter dialogs found in the Pali Canon and

compare them to the way in which their constituent suttas are organized in the other

extant gama traditions in order to determine the extent to which those collections

are of antiquity within the broader early Buddhist tradition and may even have been

composed as units or, alternatively, were constituted peculiarly within the Theravda

tradition simply to facilitate transmission of the suttas within them.

We will begin, as in Chapter IV.2, with the Sayutta Nikya, in which suttas

that can be assigned to the encounter dialog genre are found almost exclusively in the

Brhmaa Sayutta of the Sagth Vagga. As we saw in Chapter IV.2, the fullest

extant parallel to the Sayutta Nikya, namely the Z hn Jng preserved in Chinese,

is most likely of the Sarvstivda school, and it is in many ways more similar to its

Theravda counterpart than either of the other extant Sarvstivdin collections are to

theirs. This is particularly true in the case of the Brhmaa Sayukta, which not only

exists in parallel to the Brhmaa Sayutta in the Pali version, but is quite similar in

content. This, I will argue, suggests that the Brhmaa Sayutta/Sayukta is likely

of great antiquity, having been collected and to a certain extent composed earlier than

perhaps any other collection of encounter dialogs that has come down to us.

Then we will turn to the Dgha Nikya, which, as we have already seen in

Chapter IV.2, includes a collection of strasthe Slakkhandha/laskandhathat is

likely of great antiquity due to the fact that it is found in all three versions of the

Dgha/Drgha that are extant, in spite of the fact that the three versions are otherwise

464
rather dissimilar. This collection is not per se a collection of encounter of dialogs,

since it also includes a number of stras that do not involve Brahmans as

interlocutors, but encounter dialogs nonetheless play a dominant role in the

collectionmaking up nearly half of its suttas in the Theravda version. On the basis

of the Fame of Gotama formula and other common features, I will argue that the

stras of this collection, although relatively early, forming as they clearly do the

core of the Dgha Nikya/Drghgama, most likely were composed together,

drawing from various themes and formulae developed earlier in the tradition.

Next, in sections IV.3.4 and IV.3.5, we will turn to the Brhmaa Vaggas of

the Majjhima and Aguttara Nikyas. These collections are not paralleled in any

close way in the non-Theravda counterparts to these two Nikyas, and thus appear to

have been created by the Theravda tradition (or its immediate forebears) out of

convenience, to group suttas together by common theme. In these sections, we will

explore the suttas of these collections in conjuction with their parallels in the extant

non-Theravda gamas (when they exist), and we will find that while some of the

suttas in question appear to be fairly young, in some cases even unique to the

Theravda tradition, others appear to be quite old, but for various reasons were

categorized differently by different traditions. Finally, in Section IV.3.6, we will look

at the encounter dialogs of the Sutta Nipta, which, although not exactly constituting

a collection, since they are to a certain extent scattered throughout the the vaggas

therein, are fruitfully considered together because of their close proximity in the

Theravda tradition and common use of verse to express the bulk of their narratives.

465
IV.3.2 The Brhmaa Sayutta/Sayukta of the Sayutta Nikya/Sayuktgama

Given its structure, the Sayutta Nikya is dominated by doctrinal teachings,

and therefore encounter dialogs do not feature prominently within it overall.

Excluding for the moment the Brhmaa Sayutta of the Sagth Vagga, only a

handful of suttas scattered elsewhere throughout the Nikya can be considered as

encounter dialogs. Of these, only seven have parallels in the Sarvstivdin

Sayuktgama preserved in Chinese, 842 and five have no parallel whatsoever in the

preserved Sarvstivda versions. 843 Most of these suttas, both those with parallels in

the Sarvstivda tradition and those without, are encounter dialogs in form only. That

is, they are primarily teachings by the Buddha on one subject or another, with the

occasion for the teaching provided by an interlocutor who happens to be a Brahman

asking the Buddha a question. Only a couple of suttas touch on more substantive

encounter dialog themes. 844

842
SN 1.4.3.1=S 1099, SN 2.1.5.6=S 300, SN 2.4.1.8=S 946=S2 339, SN 4.1.13.9=S
255, SN 4.1.13.10=S 253, SN 5.7.2.5=S 561, SN 5.11.1.7=S 1044. The first of these (SN
1.4.3.1=S 1099) is a somewhat problematic member of the encounter dialog genre, insofar as its
interlocutor is not really a Brahman, but Mra taking on the form of a Brahman to mock some monks.
It is, consistent therewith, found in the Mra Sayutta of the Sagth Vagga.
843
SN 2.1.5.7, SN 2.1.5.8, SN 5.2.6.5, SN 5.3.3.5, SN 5.4.5.2. SN 5.2.6.5, however, is
reproduced, in the Pali tradition, at AN 5.193. While SN 2.1.5.7 and SN 2.1.5.8 have no direct parallel
in the Sarvstivda versions, they are immediately preceded by SN 2.1.5.6, which does. Given their
similar structure (a Brahman approaches the Buddha and asks a doctrinal/philosophical question), it is
possible that 2.1.5.7 and 2.1.5.8 were produced in imitation of 2.1.5.6 within the Theravda tradition.
844
One of the exceptions is SN 1.4.3.1, already mentioned above, in which Mra takes the
form of a Brahman to mock some monks. Another is SN 4.1.13.9, in which some students of the
Brahman Lohicca mock monks outside the hut of Mahkaccna and are rebuked by the latter. This
prompts an angry visit from Lohicca, who is subsequently converted by Mahkaccna. Here we have
the rather classic encounter dialog theme of the arrogant Brahman and his conversion. SN 5.11.1.7 is
also somewhat exceptional insofar as it incorporates certain formulae characteristic of the encounter
dialog genrethe Buddha visiting a brhmaagma (Brahman village) and the fame of Gotama
formulae, which prompts a group of Brahmans to visit the Buddha. The content of this sutta, however,

466
 For the most part, suttas that can unequivocally be classified as belonging to

the encounter dialog genre, and which do draw on substantive encounter dialog

themes, are found in the Brhmaa Sayutta of the Sagth Vagga. Conversely, all

of the suttas in the Brahmaa Sayutta are quite clearly examples of the encounter

dialog genre. This is consistent with the structure of the Sayutta

Nikya/Sayuktgama as a whole, which, as we saw in Chapter IV.2, for the most

part places stras in which the Buddha addresses particular audiences in the Sagth

Vagga/B Zhng Sng, and organizes them into sayuktas therein according to which

audience is addressedin this case, Brahmans. In addition, as the name Sagth

Vagga within which it is found suggests, all of the suttas in the Brhmaa Sayutta

are of the geyyga; that is, they include verses. It is clear from the outset that the

verses found in the suttas of the Brhmaa Sayutta, in many cases, existed

independently and then were essentially commented upon by adding a prose

narrative, in much the same way as the verses of the Arthavargya were, as I argued

in Chapter III.2, effectively commented upon by adding prose narratives in the Yz

Jng. Many of the verses in the Brhmaa Sayutta, and indeed the Sagth Vagga

in general, have parallels in other Pali or Sanskrit texts, or even, in some cases,

simply elsewhere in the Sayutta Nikya itself. 845 This shows that these verses were

is, like most in the Sayutta Nikya, merely a teaching by the Buddha occasioned by a question asked
by these Brahmans.
845
For parallels, see Bhikkhu Bodhi, trans., The Connected Discourses of the Buddha: A New
Translation of the Sayutta Nikya (Boston: Wisdom Publications, 2000), 1967-82.

467
not composed uniquely for a particular stra, but were appropriated freely by the

early oral tradition in the composition of new texts.

We are quite fortunate in the case of the Brhmaa Sayutta/Sayukta to

have three complete versionsthe Pali version, of course, as well as a Sarvstivda

version preserved in the (almost) complete Chinese translation of the Sayuktgama

at T. 99 (Z hn Jng), but also a second Chinese version in the partial translation of

the Sayuktgama at T. 100 (Biy Z hn Jng). This second Chinese version does

not appear to provide a very independent sectarian alternative to the first Chinese

version, however, because, as discussed in Chapter IV.2 above, the Biy Z hn

Jng likely was translated from a version of the Sayuktgama that, while slightly

different from the corresponding parts of the Z hn Jng, nevertheless also came

from the Sarvstivda school. This conclusion is borne out by the particular contents

and structure of the Brhmaa Sayukta. The Brhmaa Sayukta of T. 100 is far

more similar to that of T. 99 than it is to the Brhmaa Sayutta of the Sayutta

Nikya; in particular, the two Chinese versions have nearly the same number of stras

(38 in T. 99 and 37 in T. 100) and, after restoration as discussed in Chapter IV.2, the

order of their stras is essentially identical. 846 This is in complete contradistinction

with the Pali version, which has far fewer suttas (only 22) and preserves them in an

order that bears no discernible relationship to that of the Sarvstivda versions.

846
The only discrepancies both in terms of number of stras and their order come at the end
of the Brhmaa Sayukta. The last two stras of the T. 99 version (nos. 1186-7) are paralleled by a
single stra in the T. 100 version, namely, the final stra, no. 100. In addition, the fourth and third to
last stras of the T. 99 version (nos. 1184-5) are reversed in order in the T. 100 version; they
correspond, that is, to stra nos. 99 and 98, respectively, in the latter version.

468
 In spite of these differences, it is clear that there is a close historical

relationship between the Theravda and Sarvstivda versions of the Brhmaa

Sayutta/Sayukta. Choong Mun-Keat has published a comparative study of the

three versions of this collection and found that [m]ost of the discourses are similar in

content in the three versions. 847 Indeed, although the Sarvstivda versions are

substantially larger than the Theravda version, this is primarily a matter of including

more material, rather than fundamentally different material; that is, Choong has been

able to identify parallels to all but two of the 22 suttas of the Pali Brhmaa Sayutta

in the Brhmaa Sayuktas of T. 99 and T. 100. It is not immediately clear why the

Sarvstivda versions are larger than the Theravda version, but it is clear that it

involved structural differences between the Sarvstivda and Theravda traditions,

rather than differences in contentof the 18 stras in T. 99 that do not have a parallel

in the Pali Brhmaa Sayutta, all but three have a parallel elsewhere in the Pali

Canon. 848 Most of these extra stras have parallels in the Aguttara Nikya, which

is not surprising given that, as we found in Chapter IV.2, the Theravda tradition

847
Choong Mun-keat, A comparison of the Pli and Chinese versions of the Brhmaa
Sayutta, a collection of early Buddhist discourses on the priestly Brhmaas, Journal of the Royal
Asiatic Society, Series 3, vol. 19, no. 3 (2009), 374.
848
Choong reports that only two stras in T. 99 and T. 100 lack a Pali parallel, but this
appears to be an error. In his table on p. 373, he reports that stra no. 1161 of T. 99 and stra no. 84
of T. 100 are parallel to SN 7.8. This, however, contradicts his table on p. 374, which reports that SN
7.8 has no Chinese parallel. Comparison of the actual contents of the three stras in question shows
that the latter table is correct. The two Chinese versions are quite similarthey both are extremely
short and involve an unnamed Brahman approaching the Buddha, asking in a single verse what makes
one a Brahman, and receiving a single-verse response from the Buddha. SN 7.8, on the other hand, is
substantially longer, and it involves a Brahman named Aggika Bhradvja who is about to perform a
fire sacrifice and engages in a short debate over the value of such sacrifices with the Buddha, for a total
of six verses. With the addition of this non-parallel to the two already identified by Choong in his
table on p. 373, there are a total of three stras in the Sarvstivda versions that lack any Pali parallel.

469
appears to have filled out its Aguttara Nikya with stras that were placed elsewhere

by other traditions. In addition, these extra stras are for the most clumped

together in the middle of the Sarvstivdin Brhmaa Sayukta, with stras that do

have parallels in the Pali Brhmaa Sayutta found at the beginning and end of the

Sarvstivda collections. This structural relationship between the two sectarian

versions can be seen in Table 6. It is not clear whether the Theravdins shortened

their Brhmaa Sayutta or the Sarvstivdins expanded their Brhmaa Sayukta,

but the structure of the parallels suggests that the extra stras were either added or

removed together as a bloc.

(T. 99) (T. 100) Pali

Number 1151 74 SN 7.3
Interlocutor (youth asurindakabhradvjo
xilu) (mava named brhmao
xiluyn)
Number 1152 75 SN 7.2
Interlocutor (mava akkosakabhradvjo
(youth Bnqji named Biy) brhmao
Brahman)
Number 1153 76 SN 7.2
Interlocutor akkosakabhradvjo
(Strongly scold (Plutlush) brhmao
Pludupzh
Brahman)
Number 1154 77 SN 7.4
Interlocutor bilagikabhradvjo
(Strongly scold (Plumntlush) brhmao
Pludupzh
Brahman)
Number 1155 78 SN 7.16
Interlocutor paccankasto nma
(Brahman named (Brahman named brhmao
Contrary Meaning) Returns Against)
Number 1156 79 SN 7.5
Interlocutor (No (No harm ahisakabhradvjo
harm Brahman) mava) brhmao
Number 1157 80 SN 7.12
Interlocutor (Fire- udayassa brhmaassa
giver Brahman) (Plutlush Great

470
 Brahman)
Number 1158 81 SN 7.1
Interlocutor aatarassa
(Pszh Brahman (Pszh Brahman lady) bhradvjagottassa
lady) and and brhmaassa dhanajn
(her husband, a nma brhma
(her husband Brahman with the
Pludupzh of the surname Plutlush)
Brahman class)
Number 1159 82 Sn. 3.5
Interlocutor (Mq (a mgho mavo
Brahman) Brahman named Mq)
Number 1160 83 SN 7.7
Interlocutor suddhikabhradvjo
(a Brahman (Shlblupkng brhmao
with a retinue carrying Brahman)
parasols and wearing
sluka [a type of
garment])
Number 1161 84 None
Interlocutor (another (a Brahman)
Brahman)
Number 1162 85 None
Interlocutor ??? ???
Number 1163 86-7 AN 3.51-2
Interlocutor ??? (one old dve brhma ji
Brahman)
Number 88 88 SN 7.19
Interlocutor (Wd mtuposako brhmao
(young Brahman mava)
named Ydulu)
Number 89 89 AN 4.39
Interlocutor (a ujjayo brhmao
(young Brahman mava named
named Yubji) Yubiji)
Number 90 90 AN 4.40
Interlocutor udy brhmao
(young Brahman (mava named
named Yubji) Yubiji)
Number 91 91 AN 8.55
Interlocutor (a ujjayo brhmao
(young Brahman mava named Fy)
named Yshji)
Number 92 258 SN 7.15
Interlocutor (Jiomn (mava mnatthaddho nma
Brahman) named Jmn) brhmao
Number 93 259 AN 7.44
Interlocutor (Long uggatasarrassa
body Brahman) (Yujitshl Brahman) brhmaassa
Number 94 260 AN 5.31

471
Interlocutor (mava suman rjakumr
(young Brahman named Sngjilu)
Sngjilu)
Number 95 261 AN 3.57
Interlocutor (Born- vacchagotto paribbjako
hear Brahman) (Brahman named
Born-hear)
Number 96 262 SN 7.14
Interlocutor (an old aataro
(another old Brahman) Brahman) brhmaamahslo
Number 97 263 SN 7.20
Interlocutor (an old bhikkhako brhmao
(another old Brahman) Brahman)
Number 98 264 SN 7.11
Interlocutor kasibhradvjassa
(Farmer (farming Brahman brhmaassa
Pludupzh named Dulush)
Brahman)
Number 99 265 SN 6.3
Interlocutor (a Aatariss brhmaiy
(Pshmn Heavenly bhiku named Fntin brahmadevo nma putto
King [Brahmadeva?]) [Brahmadeva])
Number 100 266 None
Interlocutor (another (a Brahman)
Brahman)
Number 101 267 AN 4.36
Interlocutor (a doo brhmao
(Dum of the Brahman with the
Brahman class) surname Ynsh)
Number 102 268 Sn. 1.7
Interlocutor aggikabhradvjassa
(Pludupzh (Fire-clan Dlish brhmaassa
Brahman) Brahman)
Number 1178 92 Ther. 133-8
Interlocutor Vseh ther (not
(Pszh Brhma) (Pszh Brahman lady) identified as a Brhma)
Number 1179 93 SN 7.10
Interlocutor aatarassa
(Ply (a Brahman bhradvjagottassa
Pludupzh named brhmaassa
Brahman) Yculutlush)
Number 1180 94 SN 7.22
Interlocutor brhmaagahapatik
(Brahman elders of a (Brahman elders in an sabhya
Brahman assembly) assembly)
Number 1181 95 SN 7.13
Interlocutor (Heaven devahitassa brhmaassa
Does Brahman) (Heaven Offers
Brahman)
Number 1182 96 SN 7.17

472
Interlocutor (a (a Brahman) navakammikabhradvjo
Brahman) brhmao
Number 1183 97 SN 7.18
Interlocutor (a (a Brahman) aatarassa
Brahman) bhradvjassa
brhmaassa
Number 1184 99 SN 7.9=Sn. 3.4
Interlocutor (Brahman) (Fire- sundarikabhradvjo
offering Brahman) brhmao
Number 1185 98 MN 7 (latter part)
Interlocutor (Brahman) (Brahman) sundarikabhradvjo
brhmao
Number 1186 100 SN 7.6
Interlocutor jabhradvjo brhmao
(wound-topknot (a topknot-hair
Ludupzh Plutpsh Brahman)
Brahman)
Number 1187 100 SN 7.6
Interlocutor jabhradvjo brhmao
(wound-topknot (a topknot-hair
Bludupzh Plutpsh Brahman)
Brahman)
Number None None SN 7.8
Interlocutor aggikabhradvjassa
brhmaassa
Number None None SN 7.21
Interlocutor sagravo nma
brhmao
Table 4. A comparison of the three versions of the Brhmaa Sayutta/Sayukta, found in T. 99, T. 100, and the
Pali Sayutta Nikya. Stra parallels are based on Choong, A comparison of the Pli and Chinese versions of
the Brhmaa Sayutta, 373. Stras in T. 99 and T. 100 are identified by their stra number according to their
order in the Taish, but are listed according to their order as reconstructed by Ynshn. Under the identifying
number for each stra can be found the name and title used for the Buddhas interlocutor in that stra. Parallel
identifications of a Brahman as a Bhradvja are highlighted in green. Parallel identifications of the Brahmans
personal name are highlighted in yellow.

In order to study the particular features of the stras in the Brhmaa

Sayutta/Sayukta as examples of the encounter dialog genre, it is useful to begin

with the Theravda version, which displays certain patterns more clearly than the

Sarvstivda versions. The 22 suttas of the Pali Brhmaa Sayutta are divided into

two vaggas. The first of these, entitled Arahanta Vagga, contains ten suttas, all of

which end with a formula in which the Buddhas Brahman interlocutor becomes an

arhat. The second, entitled Upsaka Vagga, contains twelve suttas, all of which end

473
with a formula in which the Buddhas Brahman interlocutor expresses amazement at

the Buddhas teaching and asks to become his lay disciple (upsaka). The Theravda

version of the Brhmaa Sayutta, therefore, is neatly divided into two groups of

suttas, according to what happens at the end of the suttaeither the Brahman whom

the Buddha encounters ordains, practices the dhamma, and becomes an arhat, or else

he simply takes refuge in the Buddha as an upsaka. The Sarvstivda versions,

however, provide little to no evidence that this division has any antiquity in the

broader early Buddhist tradition. Most of the stras in those versions end with a very

simple formula, in which the Brahman is said to be pleased by the Buddhas words

and leaves, or else the Brahman is not mentioned and the bhikus are said to be

pleased by the Buddhas words. 849 Out of the 37 or 38 stras in the Sarvstivda

versions, there are only seven exceptions to this simple formulaic ending, in which

the Brahman is instead said to enter the sagha and become an arhat, and these do

not correspond in any general way to the suttas of the Arahanta Vagga in the Pali

version. 850 Indeed, this finding is completely expected given that, as already

discussed, there is absolutely no correspondence between the order of the suttas in the

Pali Brhmaa Sayutta and that of the stras in the Sarvstivda versions of the

849
T. 99 shows preference for the former formula (in which the Brahman is said to be pleased
by the Buddhas words and leaves), while T. 100 shows a preference for the latter (in which the
bhikus are said to be pleased by the Buddhas words). The first stra in the two Sarvstivda versions
can be taken as paradigmatic. The T. 99 versions ends with the following formula:
. The T. 100 version, on the other hand, ends with this formula:
.
850
S 1157=S2 80, S 1158=S2 81, S 92=S2 258, S 98=S2 264, S 1178=S2 92, S
1179=S2 93, S2 99. The last of these has the Brahman become an arhat only in the T. 100 version;
in the T. 99 version, we find a simple departure formula.

474
Brhmaa Sayukta. It appears, then, that the strict division of this collection into

stras in which the Buddhas Brahman interlocutor becomes an arhat and those in

which he becomes an upsaka is unique to the Theravda tradition or its immediate

ancestors in the early Buddhist tradition. The Sarvstivda tradition, in many cases,

does not even record that the Brahman in question so much as took refuge in the

Buddha as a lay disciple.

A more interesting, and ultimately more significant, pattern in the stras of

the Brhmaa Sayutta/Sayukta can be found in the naming of the Brahmans who

serve as the Buddhas interlocutors in each. The stras in both the Theravda and the

Sarvstivda versions are, generally speaking, quite short, with the Buddha

encountering a Brahman and engaging in a short discussion with him, at least part of

which involves verse. In some cases only the Buddha speaks in verse; in other cases,

the Brahman does as well. The theme of their discussion in most cases revolves

around the hostility of the Brahman as a representative of the new Brahmanism to

the Buddha as a ramaa, and in this way differs from most of the small number of

encounter dialogs found elsewhere in the Sayutta Nikya, in which, as already

discussed above, the Brahman interlocutor serves primarily as a fortuitous occasion

for the Buddha to deliver a teaching. In this respect, therefore, the stras of the

Brhmaa Sayukta/Sayutta serve as substantive exemplars of the encounter dialog

genre, but on the other hand, they generally speaking lack the highly developed

formulaic idiom found in the longer encounter dialogs of the Drgha/Dgha and

Madhyama/Majjhima collections. What we have here is a collection of verses used to

475
address the claims of the proponents of the new Brahmanism by embedding them in

short prose narratives, largely bereft of a specialized formulaic structure, in which the

Buddha is depicted as a ramaa and his interlocutor as a Brahman other.

Although the stras of this collection are short and therefore mostly lack

extended formulae to describe the Brahmans involved and their interaction with the

Buddha, such as we will study in more detail in Chapter IV.4, many of them do

follow a certain overall formula of sorts. This is most clear in the Pali version. In the

Brhmaa Sayutta, every sutta of the first vagga (i.e., in which the Brahmans

ultimately become arhats) introduces the Buddhas interlocutor as a Brahman of the

gotra (P. gotta) Bhradvja. 851 In addition, all but the first of these ten suttas gives

the Brahman Bhradvja a personal name that is clearly intended to be humorous or

at least cleverly refer in some way to the dispute at the heart of the Brahmans

encounter with the Buddha. Thus, for example, in the second sutta, the Bhradvja

Brahman in question approaches the Buddha simply to heap abuse upon him, and

accordingly his name is Akkosaka. In the fifth sutta, the Bhradvja Brahman is

named Ahisaka, and the Buddha tells him that his name is a conceit because only a

person who does no harm at all is worthy of such a name. And in the eighth sutta, the

851
The first few suttas imply a sort of ongoing narrative behind the Brahmans Bhradvja
identity, in that, in the second, third, and fourth suttas, the Bhradvja Brahman is said to angrily
approach the Buddha because he has heard that the previous Bhradvja Brahman had ordained under
him. Indeed, according to the commentary, the Bhradvja Brahmans were brothers: samaena
gotamena mayha jehakabhtara pabbjentena jni kat, pakkho bhinno ti. This formulaic
rationale for the Brahman approaching the Buddha is discontinued after the fourth sutta, however, and
it is not paralleled in the Chinese versionswhich is not surprising since, in the latter, the Brahmans
who encounter the Buddha for the most part do not join the sagha.

476
Buddha encounters a Bhradvja Brahman who is about to perform a fire sacrifice,

and thus is appropriately named Aggika.

This pattern of having the Brahman interlocutor introduced as a member of

the Bhradvja gotra and as having a funny/clever name is in part discontinued in the

second vagga of the Pali vesion. The one conspicuous exception is the first sutta of

the second vagga (eleventh in the sayutta as a whole), in which the Buddhas

interlocutor is named Kasi Bhradvja. His given name refers to the fact that he is a

farmer and criticizes the Buddha for not growing his own food. After this sutta, the

interlocutors the Buddha encounters are no longer said to belong to the Bhradvja

gotra. Nevertheless, the basic pattern of the Buddha encountering a Brahman and

engaging in a short dialog with him, at least partially in verse, and in most cases

regarding the difference between Brahmans and samaas, continues. In addition, in a

few cases, the Brahman in question, though not said to be a Bhradvja, is given a

funny name. 852

The obvious question we must now ask ourselves is whether there is any

evidence of antiquity for these patterns in the naming of the Buddhas Brahman

interlocutors, or whether they are unique to the Theravda tradition. On internal

evidence alone, it seems unlikely that the use of funny/clever personal names is a late

introduction to the tradition, since in many cases such names appear to have been

deliberately chosen in the composition of the suttas in question for the sake of
852
In the 15th sutta, a Brahman named Mnatthaddha (Stiff with Pride) is criticized by the
Buddha for his pride. In the 16th sutta, a Brahman named Paccanka (Opposed) approaches the
Buddha with the intention of contradicting whatever he says. In the 17th sutta, a Brahman named
Navakammika (New-Worker) likes to work in the forest and asks the Buddha what work he does.

477
punning. This is most obvious in the case of the fifth sutta, in which the Brahman is

named Ahisaka, and the Buddha, referring directly to this name, says in verse, If

you were like your name, then you would be harmless. 853 But what of the

Sarvstivda evidence? In order to facilitate comparison of the Theravda version to

the Sarvstivda versions of the stras in the Brhmaa Sayukta, I have included

the names and titles given to the Brahman interlocutor in each of the stras in each of

the three versions in Table 6. Focusing first on the ten suttas of the Arahanta Vagga

and their Chinese parallels, it cannot be said that we have an exact correspondence

between the Theravda and Sarvstivda traditions. The first two stras of the

Sarvstivda versions, for example, which correspond to the third and second suttas

of the Theravda version, clearly make no reference to the Brahman interlocutor

being a Bhradvja, and in fact dont even refer to him as a Brahman, but as a

mava. Now, this latter difference is not of devastating significanceafter all,

mava is the standard term used in the early Buddhist tradition to refer to a Brahman

youthbut it does mark a stark contrast to the Brhmaa Sayutta in Pali, which

always refers to the Buddhas interlocutors as brhmaa, and never as mava.

Clearly, the vicissitudes of the oral tradition have introduced certain significant

variations in the naming of the Buddhas interlocutors in these parallel collections.

Nevertheless, if we expand our view to the three versions of the Brhmaa

Sayutta/Sayukta in their entirety, we can indeed find evidence to support the

supposition that the patterns in naming found in the Pali version have some antiquity

853
yath nma tath cassa, siy kho tva ahisako.

478
in the tradition. Although not all Pali suttas that identify the Brahman interlocutor as

a Bhradvja have a Chinese parallel that does so as well, a fairly large number of the

Sarvstivda straseight in the T. 99 version and seven in the T. 100 version

appear to refer to the interlocutor as a Bhradvja. 854 In addition, in many cases

where the Brahman interlocutor is given a funny/clever name in the Pali, this name is

also found in one or both Chinese versionseither as a direct translation or as a

phonetic rendering. Such correspondences can be found even when the Chinese does

not agree with the Pali on whether the Brahman was a Bhradvja or not, or even

whether he was an adult Brahman or simply a mava. Thus, the Asurindaka of SN

7.3 is supported by T. 99 and T. 100, which render his name phonetically as,

respectively, xilu () and xiluyn (). Stra no. 1153 of T. 99,

on the other hand, supports the name Akkosaka of SN 7.2 with a direct translation as

Strongly Scold (, Jinm). Altogether, I count at least thirteen likely

correspondences between the personal given by one or both of the Chinese versions

and that given by the Pali version. These correspondences are highlighted in Table 6

for reference.

What are we to make of this evidence? I believe that it shows that the

Theravda and Sarvstivda versions of the Brhmaa Sayutta/Sayukta go back to

a relatively old collection of encounter dialogs that were composed together

854
The Chinese translators appear to have been somewhat inconsistent in the way they
rendered the name Bhradvja in Chinese. The translators of T. 99 appear to have preferred
(Pludupzh), while those of T. 100 appear to have preferred (Plutlush), but
there are some variations even within a single translation. See T. 99 nos. 1186-7 and T. 100 nos. 77
and 264.

479
according to a fairly simple formula. This formula is unlike the formulae that we

have been discussing so far, which involve a specific set of words said at a specific

place in a narrative, and indeed the common Brahman-themed formulae found in

longer encounter dialogs are largely absent in these stras. Rather, the formula

here is a formula for producing a stra in its entirety. The formula is simple: Take a

verse or set of verses and use them as dialog in a short prose narrative in which the

Buddha encounters a Brahman with a funny namewho, in the earliest tradition,

appears, for whatever reason, to have always been a Bhradvjaand engages in a

short dialog with him over a theme that in some way puns upon his name. For

whatever reason, as the Brhmaa Sayutta/Sayukta was expanded upon by the

early tradition, this formula was not always followed exactly, with the naming

conventions for the Brahman interlocutor often being dropped, but the most basic

requirement that a verse be expanded upon with a prose narrative involving an

encounter between the Buddha and a Brahman was retained. In addition, as is

obvious from the differing sizes of the Sarvstivda and Theravda versions, different

traditions eventually made different decisions as to which stras to keep within this

collection and which to move elsewhere.

In spite of the way in which this process has made the stras of the different

versions of this collection heterogeneous, we can feel confident that the key features

of the formula defined here represent an early principle for the composition of

stras in this collection, rather than a later attempt to systematize the stras within it.

There are two reasons for this. First, internal evidence indicates that the personal

480
names for many of the Brahmans in the Brhmaa Sayutta/Sayukta were chosen

quite deliberately in the course of the original composition in order to be punned upon

within the dialog. Second, if the naming conventions for the Brahman interlocutors

were introduced late in the tradition, we would expect to find them only in the

Theravda version and not in the Sarvstivda versions. The Theravdin Brhmaa

Sayutta is clearly more orderly than the Sarvstivdin Brhmaa Sayukta, and in

fact it does include conventional elements that are not paralleled in the Sarvstivda

versions. Nevertheless, the naming conventions are found in the Sarvstivda

versionseven if they are somewhat inconsistent and are scattered randomly

throughout the collection. It appears that the Theravda tradition has better

preservedor, at least, maintained better organizedthe earliest core of the

Brhmaa Sayutta/Sayukta. In any case, what we clearly have here is an early

collection of short encounter dialogs, composed together according to a simple set of

narrative conventions.

IV.3.3 The Encounter Dialogs of the Slakkhandha Vagga/laskandha Nipta of the

Dgha Nikya/Drghgama

As already discussed in Chapter IV.2, we are fortunate to have three sectarian

versions of the Dgha Nikya/Drghgama that have come down to usa Theravda

version in Pali, a Dharmaguptaka version in Chinese, and a (partial, but substantive)

Sarvstivda version in Sanskrit. The first two of these, coming from sister schools

of the Vibhajyavda, are almost identical in terms of their contents, but the third,

481
coming from the more distant Sarvstivda school, is substantially larger, and

includes ten stras that the Theravda tradition includes in its Majjhima Nikya,

while omitting ten Dgha Nikya suttas that are instead found in the Sarvstivdas

Madhyamgama. In terms of their structure, the three versions are overall quite

different, but the Dharmaguptaka and Sarvstivda versions do appear to each have a

section that corresponds quite closely to the Slakkhandha Vagga of the Theravda

version. For this reason, I argued, following Sujto, that this

Slakkhandha/laskandha collection formed the original core of the Dgha/Drgha

tradition.

Because the Dharmaguptaka Drghgama is so similar to the Theravdin

Dgha Nikya, and the Sarvstivda Drghgama has yet to be published and is

therefore mostly inaccessible, it is convenient to begin our survey of the encounter

dialogs in the Dgha/Drgha tradition with the Theravda version. These can then be

compared to their parallels in the Dharmaguptaka version, and additional encounter

dialogs that are found in the Sarvstivda version can be fruitfully left until the

discussion of the Majjhima/Madhyama tradition since they are for the most part found

there in the Theravda tradition. Using the most basic criterion that an encounter

dialog involves the Buddha meeting an interlocutor who is identified as a Brahman,

there are a total of eight encounter dialogs in the Dgha Nikya, although two of these

are questionable. These two are the Mahli Sutta (6) and the Subha Sutta (10), both

of which are found in the Slakkhandha Vagga. The second of these, the Subha Sutta,

does not even technically fulfill the minimum criteria of an encounter dialog, insofar

482
as it does not involve either the Buddha or an interlocutor identified by the title

brhmaa. It is set after the Buddhas death, and it involves the Buddhas disciple

nanda meeting a mava named Subha. Although the word brhmaa is not

actually used, however, we know that Subha is a Brahman because, as already noted

above, the word mava is used in the early Buddhist tradition to refer to young

Brahmans; in addition, he is said to be Todeyyas son (todeyyaputto)which

probably means that he is a student of Todeyya, who is said elsewhere to be a famous

Brahman teacher. 855 This sutta has probably borrowed its interlocutor from MN 99,

an encounter dialog which we will be looking at in Section IV.3.4, but it itself lacks

any substantative characteristics of an encounter dialog beyond the mere fact that its

interlocutor is a Brahman. Subhas meeting with nanda serves as nothing more than

an occasion for the latter to teach the dhamma as taught by his recently deceased

teacher, the Buddha.

The Mahli Sutta does begin in a manner consistent with a typical encounter

dialog, insofar as it begins with Brahmans hearing of the fame of Gotama (a

formula, which we will discuss further shortly) and thereupon deciding to see the

Buddha. These Brahmans, however, turn out to be superfluous to the suttas narrative.

At the same time, a person named Ohaddha the Licchavi (who also has the gotra

name Mahli, from which the sutta gets its name) goes to see the Buddha, and it is he

who serves as the Buddhas primary interlocutor by asking the latter about divine

forms (dibbni rpni) and divine sounds (dibbni saddni). From this point on,

855
DN 13, MN 98, SN 4.1.13.10, AN 4.187.

483
the Brahmans who came to see the Buddha are never mentioned again. Not only this,

but much of the conversation between Ohaddha and the Buddha consists of the

Buddha recounting a conversation he once had with two paribbjakas named

Maissa and Jliya. This previous conversation is, in fact, preserved word-for-word

in the immediately following sutta, called Jliya Sutta (7). This latter sutta, by itself,

cannot be considered an encounter dialog at all, since it makes no reference

whatsoever to Brahmans; the interlocutors there are paribbjakas. It appears that the

Mahli Sutta was created simply by inserting the Jliya Sutta into a frame narrative in

which the Buddha has a conversation with a different interlocutor named Ohaddha

Mahli. This Mahli was not himself a Brahman, and it is frankly a mystery why

Brahmans are mentioned as coming to see the Buddha at the same time as him at all.

In any case, neither the Mahli Sutta nor the Jliya Sutta are found in the

Dharmagupta Drghgama preserved in Chinese. Hartmann has identified parallels

to them in the Sarvstivdin Drghgama (see Table 3 in Chapter IV.2 above), but

their contents are not yet available for study.

This leaves six suttas in the Dgha Nikya that serve as substantive examples

of the encounter dialog genre. Of these, I would like to focus on three in particular

here, because they clearly were created and transmitted together as a unit, and

because they share certain features that I will argue make them represent the height

of the encounter dialog genre. Before turning to these three, however, brief mention

should be made of the other three and why they are being given shorter shrift here.

These latter three are the Lohicca Sutta (12), the Tevijja Sutta (13), and the Aggaa

484
Sutta (27). The last of these, the Agaa Sutta, is of course the well-known sutta in

which the Buddha gives a polemical account of how the world and ultimately the four

varas came into being. It happens to be the only encounter dialog in the Dgha

Nikya that is not found in the Slakkhandha Vagga, and it is not found in the

Sarvstivdin Drghgama at allit is one of the ten Dgha Nikya suttas that the

Sarvstivda tradition instead places in its Madhyamgama. I prefer to save

discussion of both it and the Tevijja Sutta, which latter we already looked at briefly in

Chapter IV.2, until Chapter IV.4, because the two of them, together with a third sutta

found in the Majjhima Nikya (98) and the Sutta Nipta (3.9), share a pair of

Brahman interlocutors named Vseha and Bhradvja. I will therefore discuss them

together as examples of a common theme centering on these two characters.

The third encounter dialog that I will not devote much attention to here is the

Lohicca Sutta. This sutta is found in the Slakkhandha Vagga, and as such it is

dominated by the long Tathgata arises formula shared by all suttas in the

Slakkhandha Vagga, including the three encounter dialogs we will be looking at

shortly. It also includes certain formulae that are characteristic of the encounter

dialog genre, including a formulaic description of Lohiccas brahmadeyya land-grant

and of his decision to visit the Buddha after hearing of the fame of Gotama. Once

Lohicca meets with the Buddha, however, the actual content of their conversation

turns in the direction of the teachings on wrong views genre. Lohicca expresses the

wrong view that a samaa or Brahman who has discovered an auspicious (kusala)

485
dhamma should simply keep it to himself, and the Buddha uses this as an opportunity

to teach about good and bad teachers.

The three encounter dialogs that we will be focusing on here are the Ambaha

Sutta (3), Soadaa Sutta (4), and Kadanta Sutta (5). These three suttas appear to

go together as a unit. As already discussed in Chapter IV.2, all three versions of the

Dgha/Drgha collection that have come down to us contain a discernible

Slakkhandha/laskandha sub-collection with similar contents, but the stras of this

sub-collection are ordered completely differently in the three traditions. In spite of

this, the three stras that we are concerned with here are found together (or almost

together) in all three traditions. In the Theravda tradition, they appear as the third,

fourth, and fifth suttas of the Dgha Nikya, in the order just cited. In the

Sarvstivda tradition, they appear as the 33rd through 35th stras of the Drghgama,

which are the roatya Stra, Kadya Stra, and Ambha Stra,

respectively. In the Dharmaguptaka tradition, the three stras are separated from one

another, but only slightly. They appear in the same order as in the Theravda

tradition, but with the Fndng Jng (=Brahmajla Sutta) intervening between the

mzhu Jng (=Ambaha Sutta) and the Zhngd Jng (=Soadaa Sutta).

Although it is impossible to know why the Fndng Jng has been inserted in this

place in the Dharmaguptaka tradition, it is likely that its placement here is spurious.

As we have already seen in Chapter IV.2, the Brahmajla Sutta is a prototype for the

entire Slakkhandha Vagga, since it contains part of the Tathgata arises formula

that serves as the centerpiece of all of the suttas in this collection, and therefore its

486
placement at the head thereof in the Theravda tradition is most logical. Moreover, as

Sujto has pointed out, it is likely that the Brahmajla Stra was originally found first

in the Dharmaguptaka tradition as well, since the Dharmaguptaka Vinaya records that

this was the very first stra recited at the first sagti. 856

We will begin with the Soadaa and Kadanta Suttas and their parallels,

since these two texts are more similar to one another than to the Ambaha Sutta, and

the latter has a more complex structure. Indeed, from what we have just seen, we are

certainly justified in considering them together as a unit, since they are found, in all

three extant traditions, together and in precisely this order (Soadaa first,

Kadanta second), with the position of the Ambaha Sutta variable. Since the

Sarvstivdin version of the Drghgama has yet to be published, we are left with

only two versions of these two stras whose contents can be compared to one another:

the Theravdin in Pali and the Dharmaguptaka in Chinese. A side-by-side

comparison of the Pali and Chinese versions of the Soadaa and Kadanta Suttas

can be found in Tables 7 and 8, respectively. Looking at these tables, we

immediately see some important similarities between the introductions of the two

stras. Both begin by introducing a certain Brahman who lives in a village that was

given to him by a king as a brahmadeyya land-grant. Moreover, although it is only

clear in the Pali version, both Brahmans have funny names. Soadaa means

dog-stick, and Kadanta means either false tooth, sharp tooth, or ox tooth,

856
Sujto, History of Mindfulness, 61.

487
depending on how the word ka is taken. 857 In any case, these names appear to be

satiricialwhich is, as we will see, very much in line with the tenor of the narratives

in which they participateand not intended as actual names of real Brahmans.

Soadaa Sutta (Pali DN 4) (Chinese D 22)

Section One
Nidna: Gaggars lotus pond near
Camp in Aga
Camp given to Brahman Soadaa as
brahmadeyya by King Bimbisra of
Magadha
Brahman householders hear fame of
Gotama formula and decide to go see
Buddha
Soadaa sees them going, asks his
steward where, and then has his steward
tell them he will go with them
Five hundred visiting Brahmans hear of
Soadaas plans and ask if they are
true
Brahmans urge Soadaa not to go,
using well-born on both sides formula
Soadaa replies by applying modified
form of well-born on both sides
formula to the Buddha
Soadaa recounts the going forth of
Buddha, repeats the fame of Gotama
formula, and gives other praise
Brahmans agree to see the Buddha using
even a hundred yojanas formula
Introduction
Nidna: Gaggar pond near Camp in
Aga
Camp given to Brahman Zhngd as
brahmadeyya (Ch. fnfn) by King
Pasenadi (of Kosala)
Zhngd introduced using Triple Veda
formula
Brahmans hear fame of Gotama
formula and decide to go see Buddha
Zhngd sees them going, asks his
steward where, and then has his steward
tell them he will go with them
Brahmans ask where Zhngd is going
Zhngd answers using fame of
Gotama formula
Brahmans urge Zhngd not to go, using
well-born on both sides formula
Brahmans give additional praise of
Zhngd
Zhngd replies by applying modified
form of well-born on both sides
formula to the Buddha
Zhngd gives additional praise of the
Buddha
Brahmans agree to go see the Buddha

857
The PTS Dictionary prefers ox tooth, taking ka as meaning ox. Rhys Davids,
however, translated the name as pointed tooth, apparently extrapolating from the possible meaning
of ka as pinnacle or point. Since ka has a third meaning, that is, false, or a lie or deceit,
I suggest here yet another possible translationfalse toothalthough I leave it open as to whether
this would mean that the Brahman had dentures or simply was a liar.

488
Section Two Characteristics of a Brahman
They go to the Buddha, and Soadaa They go to the Buddha, and Zhngd
frets about what question to ask him hopes he will ask him about Brahman-
dharma
Exchange of pleasantries
Again Soadaa frets about what to ask
about and silently hopes the Buddha will
ask him about the Three Vedas
Buddha reads Soadaas mind and asks They all sit; the Buddha reads Zhngds
what qualities constitute a Brahman mind and asks what qualities constitute a
Brahman
Zhngd is impressed the Buddha could
read his mind
Soadaa responds with five qualities: Zhngd responds with five qualities:
1. well-born on both sides (ubhato 1. ancestors are Brahmans going
sujto) back seven generations (
2. a scholar versed in mantras )
(ajjhyako mantadharo) 2. versed in the Three Vedas (
3. handsome (abhirpo)
)
4. virtuous (slav)
5. learned and wise (paito 3. has the proper appearance (
medhv) )
4. virtuous ()
5. wise ()
Buddha asks if one can be left out, and Buddha asks if one can be left out, and
Soadaa says appearance Zhngd says birth
Buddha asks again, and Soadaa says Buddha asks again, and Zhngd says
mantras mantras
Buddha asks again, and Soadaa says Buddha asks again, and Zhngd says
birth appearance

Section Three Brahmans Protest

Brahmans urge Soadaa not to say this,
saying he is agreeing with the Buddhas
own words
Buddha tells Brahmans to be silent and
let Soadaa speak if they think he is
learned and worthy
Soadaa denies he is decrying
appearance, mantras, and birth
Soadaa points out his nephew Agaka
and says that while he has all five
Brahmans ask why these three qualities
dont matter
Buddha tells Brahmans to be silent and
let Zhngd speak if they think he is
learned and worthy
Zhngd says he will respond
Zhngd says his nephew has these three
qualities, but still would not be a

489
qualities of a Brahman, he could not call
himself a Brahman if he were immoral,
even if he is handsome, knows the
mantras, and has a good birth
Buddha asks if another quality can be
omitted
Soadaa says no, since wisdom and
morality purify one another
Brahman if he were immoral
Buddha asks if another quality can be
omitted
Zhngd says no, since wisdom and
morality purify one another
Buddha agrees and says that he calls such
a person a bhiku

Section Four Teaching on Morality and Wisdom

Buddha agrees and asks what morality
and wisdom are
Soadaa says he doesnt know and
asks Buddha to explain
Buddha delivers Tathgata arises
formula, up to four jhnas, and says this
is morality
Buddha delivers rest of Tathgata
arises formula, beginning with four
jhnas, and says this is wisdom
Zhngd asks what morality means
Buddha delivers Tathgata arises
formula, up to four dhynas, and says
this is morality
Zhngd asks what wisdom means
Buddha delivers rest of Tathgata
arises formula, beginning with four
dhynas, and says this is wisdom

Section Five Conclusion

Zhngd takes refuge and five precepts
Zhngd delights in Buddhas words
Soadaa replies with excellent
formula, takes refuge in the Buddha, and
invites him to a meal the next day before
leaving
The next day, Soadaa sends word that
the food is ready
Buddha and monks go to eat at
Soadaas residence
Soadaa explains that he cannot openly
show respect to the Buddha because he
would be despised by his fellow
Brahmans, but he tells the Buddha to
understand various insignificant gestures
as signs of great respect
Buddha instructs Soadaa and leaves

490
Table 5. A comparison of the Pali and Chinese versions of the Soadaa Sutta.

The two stras then continue with a standardized opening that, while varying

slightly between the two sectarian traditions, otherwise follows a set sequence of

formulaic introductions and narrative transitions. That is to say, the two stras open

according to the same formulaic pattern within each sectarian tradition, and that

formulaic pattern varies only slightly between the two traditions. In both traditions,

the nidna ends with a formulaic account of a particular town having been given by a

king to the Brahman interlocutor as a brahmadeyya land grant. The two traditions

disagree as to who this king was, however: According to the Theravda, he was, in

both cases, Bimbisra of Magadha, while according to the Dharmaguptaka, he was

Prasenajit (Pasenadi) of Koala. 858 The Dharmaguptaka version alone then

introduces the Brahman interlocutor further using the Triple Veda formulaa

formula that, though not found here, is not unknown to the Theravda tradition

which eulogizes a particular Brahman as a scholar well-trained in the three Vedas.

According to both sectarian traditions, this Brahman interlocutor then ends up

deciding to go see the Buddha according to the same pattern. First, other Brahmans

in his town hear that the Buddha is in town, expressed through what I call the fame

of Gotama formula, and decide to go see him. When the primary interlocutor sees

them going, he asks a servant to find out what they are doing, is told they are going to

see the Buddha, and has that servant go back and tell the other Brahmans to wait for

858
At least in the case of the Soadaa Sutta, the Theravdin reading is probably superior.
The Dharmaguptaka version agrees that the brahmadeyya land-grant in question was Camp, which is
in Aga, i.e., modern-day Bengal, even further to the east than Magadha. It makes more sense that the
king making the grant would have been the king of Magadha, rather than the king of Kosala to the west.

491
him, as he would like to visit the Buddha as well. This then prompts yet another

group of Brahmans to protest that he should not go to see the Buddha, as the Buddha

is of lower status than him and therefore should be the one to go visit him. This

results in a formulaic exchange in which the contrary Brahmans first praise the

primary Brahman interlocutor using what I call the well-born on both sides formula,

and then the primary Brahman interlocutor uses a modified version of the same

formula to in turn praise the Buddha. This convinces the other Brahmans, who then

decide that they would like to visit the Buddha as well. This introduction to the

Soadaa and Kadanta Suttas varies only slightly between the two sectarian

traditions, primarily in terms of the number of times formulae such as the fame of

Gotama are repeated.

Kadanta Sutta (Pali DN 5) (Chinese D 23)

Section One Introduction
Nidna: Ambalahik park in Brahman Nidna: Sissap-vana () near
village of Khnumata Khnumata ([/]) in Kosala
Khnumata given to Brahman Kadanta Khnumata given to Brahman Kadanta
as brahmadeyya by King Bimbisra of () as brahmadeyya by King
Magadha Pasenadi (of Kosala)
Kadanta introduced using Triple
Veda formula
Kadanta plans a great sacrifice Kadanta plans a great sacrifice
Brahman householders of Khnumata People of village hear fame of Gotama
hear fame of Gotama formula and go to formula and decide to visit Buddha
Ambalahik
Kadanta sees them going, asks his Kadanta sees them going, asks his
steward where, and is told the fame of steward where, and is told the fame of
Gotama formula Gotama formula
Kadanta thinks to himself that he has Kadanta asks steward to tell the people
heard the Buddha knows how to perform he will go too
the sacrifice and asks the crowd to wait
so he can go too
Several hundred Brahmans who are 500 Brahmans ask Kadanta where he is

492
visiting to take part in sacrifice ask if it is going
true Kadanta is going to see the Buddha
Kadanta replies with fame of Gotama
formula
Brahmans say Kadanta should not visit Brahmans say Kadanta should not visit
Buddha, using well-born on both sides Buddha, using well-born on both sides
formula formula
Kadanta replies by applying modified Kadanta replies by applying modified
form of well-born on both sides form of well-born on both sides
formula to Buddha formula to Buddha
Kadanta recounts the going forth of Kadanta gives additional praise to the
Buddha, repeats the fame of Gotama Buddha
formula, and gives other praise
Brahmans agree to see the Buddha using Brahmans agree to visit Buddha
even a hundred yojanas formula

Section Two Jtaka

All go to Buddha, and Kadanta asks All go to Buddha, and Kadanta asks
how to perform a sacrifice how to perform a sacrifice
Buddha says there was a king in the past Buddha says there was a king in the past
named Mahvijita who wanted to who wanted to perform a sacrifice and
perform a sacrifice and asked his told his Brahman ministers
purohita how
Purohita says kingdom is beset by
thieves and tells king to neither raise
taxes nor punish the wicked, but hand out
grain and capital and raise wages; king
agrees, and it works
Prominent people donate for the sacrifice
Mahvijita asks how to hold sacrifice, King tells ministers kingdom is strong
and purohita tells him to summon and asks how to hold sacrifice; they agree
khattiyas, advisers, Brahmans, and and tell him to tell the people in the
householders palace
Ministers say king should tell princes and
ministers, construct an enclosure to east
of city, pick animals, etc.
Purohita praises Mahvijita as possessing Eight qualities of a king, beginning with
eight qualities of a king, beginning with well-born on both sides formula
well-born on both sides formula
Purohita endowed with four qualities: Four qualities of ministers, incorporating
well-born, a scholar, virtuous, learned Triple Veda formula
Purohita tells Mahvijita not to regret
loss of wealth in sacrifice

493
Purohita describes 10 conditions for
recipient of sacrifice
Purohita gives 16 reasons someone
might criticize sacrifice, but cannot,
because they are not true
Buddha says no living beings were killed
for sacrifice
People try to offer king their wealth for
the sacrifice, but he turns them away
since he has his own wealth to give away
Four groups of people give away their
wealth anyway, at the four directions
Sacrifice performed
Brahmans exclaim this was a splendid
sacrifice
Kadanta remains silent
Kadanta asks Buddha if he was present
at sacrifice
Buddha says yes, he was the purohita
Section Three
Kadanta asks Buddha if there is a
simpler and more profitable sacrifice
Buddha says yes, giving gifts to virtuous
ascetics
Kadanta asks why; Buddha says arhats
wont attend sacrifice that involves
violence
More profitable? Providing shelter for
sagha
More profitable? Going for refuge
More profitable? Taking the precepts
More profitable? Buddha answers with
Tathgata arises formula
Section Four
16 reasons someone might criticize
sacrifice, but cannot, because they are not
true
10 conditions for recipient of sacrifice
Ministers tell king not to have 3 doubts
People try to offer king their wealth for
the sacrifice, but he turns them away
since he has his own wealth to give away
Buddha says no living beings were killed
for sacrifice
Sacrifice performed
King and others go forth and attain
brahmaloka
Buddha asks Brahmans what they think
of this sacrifice
Kadanta remains silent and Brahmans
ask why
Kadanta asks Buddha if he was the
king or the purohita
Buddha says he was the king
Better Sacrifices
Kadanta asks Buddha if there is a better
sacrifice
Buddha says yes, giving to sagha
Better still? Building buildings for
sagha
Better still? Going for refuge
Better still? Taking the precepts
Better still? Have heart of compassion
Better still? Buddha answers with
Tathgata arises formula
Conclusion
494
Kadanta replies with excellent
formula, takes refuge, and sets free all the
animals he was going to kill in the
sacrifice
Buddha delivers graduated teaching and
Kadanta attains Dhamma-eye
Kadanta invites Buddha to a meal
The next morning, Buddha and monks go
for meal at place of sacrifice; Buddha
instructs Kadanta and departs
Table 6. A comparison of the Pali and Chinese versions of the Kadanta Sutta.
Kadanta releases animals, takes
precepts, offers to feed sagha, and
departs
The next morning, Buddha and monks go
to Kadanta
Gth on highest things
Buddha gives graduated teaching;
Kadanta takes refuge and precepts
Kadanta feeds Buddha for seven days
Kadanta gets sick and dies
Monks ask Buddha his destination, and
Buddha says he is a non-returner
Monks pleased at Buddhas words

The central narratives that follow these formulaic introductions in the

Soadaa Sutta and the Kadanta Sutta are, with minor variations, presented in the

same way by the two available sectarian traditions, and each sutta deals in its own

way with the claims of the new Brahmanism. The Soadaa Sutta deals with

these claims more comprehensively, insofar as it addresses all of these claims at once.

When Soadaa (Zhngd in the Chinese version) meets the Buddha, he secretly

wishes that the Buddha will ask him a question that he, as a scholar of the Vedas, will

be able to answer, and the Buddha obliges by asking him what characteristics a

person must have to be a Brahman. Soadaa lists five characteristics, which are

essentially the same in the Pali and the Chinese: He must be of the proper birth,

going back seven generations; he must be learned in the Vedas; he must be handsome;

he must be virtuous; and he must be wise. The Buddha asks him repeatedly if any of

495
these characteristics can be omitted as not necessary for being called a Brahman, and

Soadaa admits, in turn, that the first threethat is, the three most connected with

the claims of the new Brahmanismare not necessary. This creates an uproar among

the Brahmans present, who accuse Soadaa (rightly) of abandoning the ideological

position of the new Brahmanism and taking up the position of the Buddha himself.

When the Buddha asks if the remaining two characteristicsvirtue and wisdomcan

be omitted, Soadaa says that they cannot. Not surprisingly, the Buddha agrees

with this assessment, and then proceeds to give an extended sermon on what virtue

and wisdom consist of. This sermon is based on the Tathgata arises formula

found common to all of the suttas found in the Slakkhandha Vagga (and its parallels

in the other traditions). Virtue consists of all of the training up to the acquisition of

jhna, and wisdom consists of the rest of the training, from the jhnas to final

realization.

The Kadanta Sutta also addresses the claims of the new Brahmanism, but

focuses more narrowly on one aspect thereofnamely, the Vedic sacrifice.

Kadanta is planning on holding a great sacrifice, and when he hears that the Buddha

is in town, he decides (implausibly) to ask the Buddha how to hold a sacrifice,

because he has supposedly heard that the Buddha is an expert in this. The Buddha

answers his question by telling a Jtaka, in which a king holds a grand sacrifice in

which no animals are killed. This story is a Jtaka because, as Kadanta guesses, the

Buddha was one of the characters within it, although the Pali and Chinese versions

disagree as to whether he was the king or his purohita. The ultimate point of the

496
story, of course, is that a truly meritorious sacrifice should not involve killing, but

after telling the story, the Buddha makes it clear that such a Vedic-style sacrifice,

even when completely vegetarian, is of relatively little benefit. When Kadanta asks

if there are any more beneficial sacrifices, the Buddha successively enumerates

various actions, such as giving dna and taking refuge (the list varies somewhat

between the two sectarian traditions), that are consistent with being an upsaka of the

Buddha. This list culminates, in both traditions, with the Tathgata arises formula,

tracing the training of a monk from his going forth to his attainment of Awakening

and this, of course, the Buddha deems the most beneficial sacrifice of all.

The Ambaha Sutta is longer and more complex, both in terms of narrative

and use of formulae, than the other two texts we have just looked at. Moreover, in

this case we have available to us versions from three sectarian traditions, rather than

just twothe Theravdin in Pali and Dharmaguptaka in Chinese as before, but also a

(Mla-)Sarvstivda version preserved in the Tibetan Mlasarvstivda Vinaya. 859 A

859
Actually, there are five versions of the Ambaha Sutta currently available, in addition of
course to the Sarvstivda version in Sanskrit that has not yet been published. The first two are, of
course, the Theravdin in Pali and the Dharmaguptaka in Chinese. In addition to the Dharmaguptaka
version in the full translation of the Drghgama at T. 1, however, there is a second Chinese translation,
of a version of the Ambaha Sutta by itself as a stand-alone text, at T. 20, entitled the F Kiji Fnzh
b Jng (). I have consulted the translation at T. 20, but do not discuss it here,
because it has certain stylistic peculiarities (i.e., as a Chinese translation) that obscure its presumably
original oral formulaic structure and warrant a close study by itself; moreover, its sectarian provenance
is, as far as I am aware, unknown. According to Chinese records, this stra was translated in the third
century by Zh Qin, but it is not listed as one of the authentic translations of Zh Qin by Nattier
Nattier, Guide to the Earliest Chinese Buddhist Translations, 121-2. Indeed, according to Xiao-Jung
Yu (personal communication), a specialist in pre-modern Chinese linguistics whom I asked to look at
the text, the translation bears certain linguistic features, including the use of and of , that appear
to belie a later date than that of Zh Qin.
In addition, the (Mla-)Sarvastivda version preserved in the Tibetan Mlasarvstivda
Vinaya is actually two separate versions; that is, the story of Amba appears twice in the
Mlasarvstivda VinayaDul ba gzhi, Lha sa bka gyur, Dul ba kha, fols. 144a-176a, and Dul ba
phren tshegs kyi gzhi, Lha sa bka gyur, Dul ba da, fols. 304a-337a. I have only consulted at length

497
comparison of these three versions can be found in Table 9. As we can see from this

table, the Ambaha Sutta begins much as the Soadaa and Kadanta Suttaswith

the Buddha coming to stay at a particular Brahman village (brhmaagma) and the

introduction of a prominent Brahman who has been given a brahmadeyya land-

grantin this case, the Pali and Chinese versions agree, by King Pasenadi of Kosala.

Here, however, the introduction of the location and the interlocutor is somewhat more

complicated than in the previous two texts, in two ways. First, the Brahman village

where the Buddha stays (Icchnakala) is not the same as the village that the

prominent Brahman received as brahmadeyya (Ukkhaha)although, from context,

we can tell that they are close to one another. 860 Second, the prominent Brahman

who is the recipient of this brahmadeyya (Pokkharasti) is not the only Brahman

interlocutor with whom the Buddha interacts; he sends his student Ambaha to the

Buddha on his behalf, and it is this Ambaha who debates with the Buddha

throughout most of the narrative and who lends his name to the title of the sutta.

Ambaha Sutta (Pali DN (Chinese D Mlasarvstivda Vinaya

3)
Section One
Nidna: Brahman village
of Icchnakala in Kosala
Brahman Pokkharasti
living at Ukkhaha, which
20)
Introduction
Nidna: Icchnakala ( Nidna: Village in Kosala
) in Kosala that was given to Brahman
Lotus Essence (Padmai
snying po) by King
Prasenajit
Brahman Pokkharasti
living at Ukkhaha (

the version in volume kha because, at the time I was conducting research for this dissertation, that was
the only version that had been edited. It is this version that I refer to here.
860
Bronkhorst has cited this text as evidence that brhmaagma and brahmadeyya are not
synonyms, even though in many cases (such as in the Soadaa and Kadanta Suttas), they do refer
to the same villageBronkhorst, Buddhism in the Shadow of Brahmanism, 92.

498
has been given to him as ), which has been given
brahmadeyya by King to him as a brahmadeyya
Pasenadi by King Pasenadi
Pokkharasti hears fame
of Gotama formula
Introduction to Introduction to Brahman Introduction to Lotus
Pokkharastis student Pokkharasti ( Essences student Ma
Ambaha using Triple ) and his student sdug
Veda formula Amba () using
Triple Veda formula
Pokkharasti repeats Pokkharasti hears fame Lotus Essence tells Ma
fame of Gotama of Gotama formula and sdug fame of Gotama
formula to Ambaha and tells Amba to see if he formula (twice) and tells
tells him to put Gotama to has all 32 marks of a Great him to see if he has all 32
the test Man marks of a Great Man
Ambaha asks how Amba asks how to tell
Pokkharasti explains 32 Pokkharasti explains
marks of a Great Man using two courses of a
using the two courses of Great Man formula
a Great Man formula

Section Two Ambaha Treats Buddha Rudely

Ambaha approaches
monks doing walking
meditation and asks them
where Gotama is; they
direct him to the Buddhas
dwelling
Ambaha is let in, and Amba goes to the Ma sdug goes with many
young men with him Buddha and acts rudely Brahmans and acts rudely
exchange pleasantries with toward him toward the Buddha
Buddha, but Ambaha
acts rudely and does not
sit
Buddha asks if Ambaha Buddha asks if Amba is Buddha asks if Ma sdug is
is so impolite with learned this rude with Brahmans this impolite with learned
Brahmans Brahmans
Amba asks what he
means and Buddha
explains
Ambaha says Amba says he respects Ma sdug says no, but the
shavelings (muak) his Brahman teacher, but Buddha deserves no better
are menials from not shavelings because he is a shaveling

499
Brahms foot and
deserve no better
Buddha calls Ambaha Buddha calls Amba
untrained untrained
Ambaha gets angry and Amba gets angry and Ma sdug says kyas
insults the Sakyas complains about the should respect Brahmans
kyas
Buddha asks how Sakyas Buddha asks how kyas Buddha asks how kyas
offended Ambaha offended Amba offended Ma sdug
Ambaha says that when Amba says they didnt Ma sdug says they showed
he visited Kapilavatthu greet him when he visited him disrespect when he
with Pokkharasti, they visited Kapilavastu with
joked around, acted like Pokkharasti
they were making fun of
him, and didnt offer him
a seat
Buddha says this is a Buddha says kyas were Buddha defends kyas by
trifle; compares Sakyas to in their own home, like comparing them to birds in
quail in his own nest birds in a tree their own nest

Section Three Ambahas Low Ancestry

Ambaha says three lower Amba lists four varas Ma sdug says that three
classes should pay and says Brahmans lower classes should
homage to Brahmans deserve most respect respect Brahmans, but
kyas dont
Buddha says kyas have
high lineage, but Ma sdug
is descended from slave
give of kyas
Buddha asks Ambaha his Buddha asks Amba his Buddha asks Ma sdug his
clan name, and Ambaha clan name, and Amba clan name, and Ma sdug
says it is Kahyan says it is Shngwng ( says it is Kyan
)
Buddha tells Ambaha Buddha tells Amba his Buddha repeats that Ma
that he is descended from ancestors were slaves of sdug is descended from a
a slave-girl of the Sakyas; the kyas slave-girl of the kyas
he describes at length the
origin of the Sakyas from
the banished sons of King
Okkka and the origin of
the Kahyans from
Okkkas slave-girl Dis
who gave birth to a black

500
(kaha) baby
Young men certify that Other Brahmans tell Brahmans tell Buddha not
Ambaha is well-born and Buddha he is wrong to say this
a scholar and ask Buddha
not to humiliate him
Buddha tells young men to Buddha tells them to be Buddha tells them to be
be quiet and let Ambaha quiet if they think their quiet if they think Ma sdug
speak for himself if they teacher is worthy is worthy to respond
think he is so worthy
Buddha says that Ikvku Buddha tells the story of
banished four sons; their Ikvku, the origin of the
mothers went to visit them kyas, and the origin of
and proposed that they the Kyans from a
have children through slave-girl
incest with their sisters;
this pleases Ikvku, and
he calls them kyas; their
slave girl Dis has sex
with a Brahman and gives
birth to a baby who can
talk (Ambas ancestor)
Buddha asks Ambaha Buddha asks Amba Buddha asks Ma sdug three
three times if he knows of three times if he knows times if he knows this story
his ancestry, warning that this story, warning that
his head will split into Vajrapi will split his
seven pieces if he doesnt head if he doesnt answer
answer the third time
Vajirapi the yakkha Vajrapi appears; Vajrapi appears; Ma
appears, holding a flaming Amba is terrified and sdug is terrified and finally
iron club above finally admits that he admits that he knows the
Ambahas head, waiting knows the story story
to strike if he doesnt
answer; Ambaha is
terrified and finally admits
that he knows the story
Young men shout out in Uproar among the Uproar among the
amazement Brahmans Brahmans
Buddha saves Ambaha Buddha tells Brahmans Buddha explains that Ka
from humiliation by not to call Amba a was a great sage and
explaining that Kaha, slave because his ancestor married the daughter of
though son of a slave, was became a great sage Ikvku
a great sage; tells story of
how Kaha sought
Okkkas daughters hand

501
in marriage
Section Four
Buddha asks Ambaha Buddha asks Amba
how Brahmans and
khattiyas, respectively,
would treat a man with a
khattiya father and
Brahman mother
Buddha asks Ambaha Buddha asks Amba
how Brahmans and
khattiyas, respectively,
would treat a man with a
Brahman father and
khattiya mother
Buddha concludes that
khattiya is senior in both
cases (because Brahmans
accept son of mixed
union, but khattiyas dont)
Buddha asks how an
exiled Brahman would be
treated by Brahmans
(poorly)
Buddha asks how an
exiled khattiya would be
treated by Brahmans
(well)
Buddha concludes that
khattiya is superior to
Brahmans even if he has
been humiliated
Buddha quotes verse
uttered by Brahm uttered by Brahm
Sanakumra
Buddha expresses
agreement with verse and
repeats it
Comparison of Katriyas Brahmans
and
Ma sdug is discouraged
with slumped shoulders, so
Buddha decides to ask him
more questions
Buddha asks Ma sdug how
how katriyas would treat the son of a katriya father
a man with a katriya and Brahman mother
mother and Brahman would be treated
father
Buddha asks Ma sdug how
how Brahmans would the son of a Brahman father
treat a man with a and katriya mother would
Brahman mother and be treated
katriya father
Buddha asks how an Buddha asks questions
exiled Brahman would be about exiled Brahman and
treated by katriyas katriya
(poorly)
Buddha asks how an
exiled katriya would be
treated by Brahmans
(well)
Buddha concludes that Buddha concludes that
katriyas are superior with katriyas are superior
respect to both women and
men
Buddha quotes verse Buddha quotes approvingly
verse uttered by Brahm
Sahampati
Buddha expresses
agreement with verse and
repeats it
502
 Buddha asks Ma sdug
questions about marriage
Buddha tells Ma sdug he
must abandon pride of birth

Section Five Teaching on Conduct and Knowledge

Ambaha asks about Amba asks about Ma sdug asks about this
conduct (caraa) and unexcelled wisdom () faith
knowledge (vijj) and practice ()
Buddha replies that the
conceit of birth and clan
does not come from
unexcelled knowledge and
conduct
Ambaha asks about this
conduct, this knowledge
Buddha responds with the Buddha responds with Buddha responds with
Tathgata arises Tathgata arises formula Tathgata arises formula,
formula, up to the four concluding, This is
jhnas, saying that this is knowledge and conduct
conduct
Buddha continues
Tathgata arises formula
to the end, saying that this
is knowledge

Section Six Questions about

Brahmans
Buddha describes four Buddha describes four
paths to failure: living on paths to failure
windfalls, living on tubers
and roots, tending the
flame at the edge of a
village, and honoring
guests in ones house
Buddha asks Ambaha if Buddha asks Ambaha if Buddha asks Ma sdug if he
he and his teacher fulfill his teacher fulfills each of and his teacher fulfill even
even these four paths (they the four paths to failure these four paths (they
dont); they do not have (he doesnt) dont)
unexcelled knowledge and
conduct either
Buddha repeats this Buddha asks how Ma sdug
fivefold failure of their can say that kyas are of

503
ands asks how Ambaha low birth
and Pokkharasti can call
others shavelings and
menials
Buddha points out that
Pokkharasti lives by the
grace of King Pasenadi,
but is not even allowed to
see him face-to-face
Buddha asks if a workman
would be speaking the
Kings words or be the
Kings equal just by
saying This is what King
Pasenadi of Kosala says!;
he compares this to
Brahmans saying they
speak for the sages of old
Buddha asks if the original Buddha asks if sages of Buddha cites sages of old
sages were well old had large homes like by name and asks if they
ornamented and indulged Brahmans of today, built lived, like Ma sdug, in
in sense pleasures, ate fine walls around their houses, cities; had armies and
foods, amused themselves slept on high and soft shows; had horses,
with women, rode in beds, or adorned chariots, and assemblies;
chariots, and guarded themselves (no in each got massages and
themselves in fortified case) ablutions; or were adorned
towns; since the sages (no in each case)
didnt and Ambaha and
his teacher do, the Buddha
concludes the latter are not
be sages
Buddha asks if one Buddha asks if one speaks
becomes equal to for the King just by saying
Prasenajit just by saying so, compares this to
so Brahmans claiming to
speak for the sages
Buddha asks if one
becomes Prasenajits
minister just by saying so
Buddha asks if one can
use donations according to
dharma (yes)

504
Section Seven The Thirty-Two Marks
Buddha tells Amba to
look for what he came for
Buddha and walk together, Amba sees all but two Ma sdug looks for the 32
and Ambaha looks for marks, finding all but two
the 32 marks, finding all
but two
Buddha shows the two Buddha shows the two Buddha shows the two
hidden marks hidden marks hidden marks
Ambaha leaves Amba leaves Ma sdug leaves

Section Eight Report to Pokkharasti

Ambaha returns to
Pokkharasti, who is
sitting in a park with a
large group of Brahmans
Pokkharasti asks
Ambaha if he saw the 32
marks, and Ambaha says
yes
Pokkharasti asks what
happened, and Ambaha
tells him
Pokkharasti becomes
enraged, yells at
Ambaha, and kicks him
over
Amba returns to Ma sdug returns to Lotus
Pokkharasti Essence and his assembly
Pokkharasti asks
Amba if he saw the 32
marks, and Amba says
yes
Pokkharasti asks what Lotus Essence asks if he
happened, and Amba saw the Buddha, and Ma
tells him sdug tells him what
happened
Pokkharasti chastises Lotus Essence gets angry
Amba and kicks Ma sdug

Section Nine Conclusion

Pokkharasti wants to see
Buddha, but Brahmans say
it is too late in the day; he
prepares food anyway and
sets out by torchlight
Pokkharasti wants to see Lotus Essence wants to
the Buddha, but waits see the Buddha, but waits
until the next morning to until the next morning to
do so do so
Pokkharasti sees all but
two marks, which Buddha
then shows him
Pokkharasti tells Buddha
his code for showing
respect to him
Lotus Essence offers
gifts to Buddha

505
 Pokkharasti takes refuge
and precepts and offers to
feed the sagha
Pokkharasti returns home
to prepare food
Buddha and monks go to Buddha and sagha are fed
Pokkharastis house for by Lotus Essence
meal
Buddha praises dna
Pokkharasti asks Buddha Pokkharasti asks Buddha Lotus Essence asks
about his exchange with three times to forgive Buddha about his exchange
Ambaha and asks him to Amba with Ma sdug and asks him
forgive him to forgive him
Buddha forgives Buddha says Amba will Buddha forgives Ma sdug
Ambaha get scabies, and he does
Pokkharasti looks for 32
marks and sees all but
two, which the Buddha
then shows him
Pokkharasti invites
Buddha to a meal
Buddha and monks go to
eat at Pokkharastis
residence
Pokkharasti tells Buddha
his code for showing
respect to him
Buddha delivers graduated Buddha delivers graduated Buddha delivers a Dharma
teaching to Pokkharasti, teaching to Pokkharasti, talk to rid Lotus Essence
who then attains the who then attains the of his pride; the latter
Dhamma-eye Dharma-eye understands
Pokkharasti serves
Buddha for seven days
Pokkharasti dies
Monks ask Buddha what
happened to Pokkharasti;
Buddha says he is a non-
returner; monks delight in
Buddhas words
Pokkharasti replies with Lotus Essence takes
excellent formula and refuge and leaves
takes refuge
Table 7. A comparison of the Pali, Chinese, and Tibetan versions of the Ambaha Sutta.

506
 The introduction section of this sutta, while formulaic and thus quite similar

to those of the previous two suttas we examined, takes a somewhat different narrative

course. As before, the Buddhas presence becomes known to the prominent Brahman

in question through the fame of Gotama formula, but unlike in the other two suttas,

there is no mass movement of people to go to see the Buddha that this Brahman

decides to join; in fact, Pokkharasti does not initially go to see the Buddha at all.

Instead, he sends his student Ambaha, and he sends him with the specific instruction

to see if the Buddha really does have all 32 marks of a Great Man (mahpurisa).

Although these 32 marks are mentioned in the fame of Gotama formula that, as we

already saw, played a prominent role in the Soadaa and Kadanta Suttas, and the

ability to read these marks is mentioned in the Triple Veda formula used to

introduce the Brahman interlocutors in the Chinese versions of those same two stras,

this particular interest in checking to see if the Buddha actually has the marks

represents a new narrative element. Thus, in the introductory section of the Ambaha

Sutta, we do not find an argument over whether the Brahman interlocutor should go

to see the Buddha or the Buddha should come to see him; instead, we find

Pokkharasti explaining to his student Ambaha what to look for when he meets the

Buddha. This consists, in part, of what I refer to as the two courses of a Great Man,

which recapitulates the prediction, well-known from the classic story of the Buddhas

birth, that a person born with the 32 marks will become either a Buddha or a

cakravartin.

507
 The primary narrative begins when Ambaha goes to visit the Buddha. Here

again we find a key difference from the Soadaa and Kadanta Suttas: Whereas

Soadaa and Kadanta were eager to see the Buddha and treat him respectfully,

Ambaha is clearly not eager to make the visit and acts quite rudely towards the

Buddhain the Pali version, even refusing to sit down while the Buddha is himself

seated, a serious breach of etiquette since it results in his head being higher than the

Buddhas. Interestingly, while this narrative element represents a break from the

more common narrative structure of encounter dialogs within the early Buddhist

tradition, as exemplified by the Soadaa and Kadanta Suttas, Brian Black has

argued convincingly that it draws upon a broader Indian trope of the humble teacher

and his arrogant student, which is also found outside the Buddhist tradition in the

Upaniadic story of Uddlaka rui and his student vetaketu. 861 The Buddha asks

the reason for Ambahas rude behavior, and the latter replies that he felt that he had

been shown disrespect by the kyas when he was visiting Kapilavastu with his

teacher Pokkharasti. The Buddha dismisses this as a trifle and says that the kyas

can act as they wish in their own home, using the specific comparison, found in all

three sectarian versions at our disposal, to birds in their own nest. The upshot of

Ambahas complain, however, is that he and his teacher are Brahmans, and the

kyas, being mere katriyas, should show them respect on that basis alone.

Ambahas explanation of the reason for his anger toward katriyas, and

kyas in particular, reveals the main theme of this text, which strikes right at the

861
Black, Ambaha and vetaketu.

508
heart of the claims made by the proponents of the new Brahmanism. Whereas the

previous two texts we have looked at address various sources of Brahman pride

birth, Vedic learning, and appearance in the Soadaa Sutta and the Vedic sacrifice

in the Kadanta Suttathis text confronts head-on the most radical claim that was

made by the new Brahmanismnamely, that society is divided into four varas, and

Brahmans are, by virtue of their birth, superior to the other three. The remainder of

the narrative in this sutta cannot but be read as a scathing, systematic, and

comprehensive attack on this claim.

This attack begins, in what I have labeled for convenience Section 3 in

Table 9 above, with an ad hominem attack on Ambahas personal ancestry. The

three versions we have differ somewhat in the order in which they present this attack,

but mostly agree on the details. Since Ambaha claims to be superior qua Brahman

on the basis of birth, we would expect him to indeed come from a high lineage of

Brahmans going back many generations. This, however, is not true: As the Buddha

reveals, Ambaha is in fact descended from the black (kaha/ka) baby of a slave-

girl of the kyas, and this is the reason that his clan name today is

Kahyan/Kyan. The kyas themselves, however, are descended from King

Ikvku (P. Okkka), founder of the great Solar Dynasty (sryavaa). The Buddha

demands that Ambaha confirm that he knows this story is true, and although he is

reluctant to do so at first, he finally admits to it when Vajrapiin likely his first

509
appearance in Buddhist literatureappears and threatens to split his head into seven

pieces if he does not answer. 862

This revelation creates an uproar among the other Brahmans who have come

with Ambaha and was clearly intended as a dramatic and turning point in the

narrative, akin to the point in a soap opera when a man and a woman who are

involved in a romantic relationship discover that they are long-lost siblings, or the

moment in Star Wars when Luke Skywalker discovers that Darth Vader is his father.

The particular details of this dramatic revelation, however, serve two purposes. First,

they render ridiculous Ambahas claim to any sort of superiority on the basis of birth.

Ambaha is not only not descended from the Vedic is; he is descended from the

lowest sort of person imaginablea slave. Second, by making this slave-ancestor of

Ambahas not just any slave, but a slave of the kyas, the narrative produces a

dramatic reversal of Ambahas earlier claim to superiority over the kyas by

making him clearlyand humiliatinglysubordinate to them. The Buddha does

then soften his attack on Ambahas ancestry by noting that, in spite of his birth, the

black baby Kaha/Ka grew up to become a great sage in his own right, but even

in making this concession, the narrative is preparing the audience for the ultimate

point of the text, which is that excellence depends on conduct and not on birth.

The Buddha does not, however, come to this conclusion directly. Contrary to

the modern perception that the Buddha taught the equality of varas, the Buddha, in

862
The concept of the head splitting if one cannot give an answer in debate appears to be pan-
Indian and trans-sectarian, as it is also found in Brahmanical sources. For a comparative study, see
Michael Witzel, The Case of the Shattered Head, Studien zur Indologie und Iranistik 13/14 (1987):
363-416.

510
what follows here in Section 4, actually drives home the point that not only are

Brahmans not superior to katriyas, but katriyas are in fact superior to Brahmans. I

have already made reference to the arguments here in Chapter II.2, where I noted that

they would only make sense if Brahmans were in fact known to not be fastidious in

maintaining pure lineages. The Buddha gets Ambaha to concede that, while

katriyas will not accept as a katriya a person of mixed birth, Brahmans will. In

particular, the mixed births in question are between Brahmans and katriyas. That

is, according to the argument made here, while a katriya will not accept as a katriya

someone who has a Brahman for a parent, Brahmans will accept as a Brahman

someone who has a katriya as a parent. The point being made here, then, is not only

that Brahmans come from impure lineages, but that they implicitly acknowledge the

superiority of katriya lineages by accepting Brahmans who have one katriya parent.

Interestingly, it is possible that Ambahas name, which does not on the surface

appear to be humorous in the same way as Soadaa and Kadanta, may in fact be

a punning reference to the argument being made here in Section 4. Bronkhorst has

noted that, although Ambaha appears to have been Sanskritized in the later Buddhist

tradition as Amba, another equally valid Sanskritization is Ambahawhich

happens to be the term used in much of the dharmastra literature for the product of

a union between a Brahman man and a non-Brahman woman. 863 It is possible, then,

that Ambahas name was chosen to make humorous reference to the propensity,

noted here in Section 4, of Brahmans to take non-Brahman wives.

863
Bronkhorst, Greater Magadha, 354.

511
 Section 4 continues, with some slight variation between the three traditions at

our disposal, by making further points that establish the superiority of katriyas to

Brahmans. These revolve around the apparent fact, conceded by Ambaha, that

katriyas will not welcome a disgraced and exiled Brahman, but that Brahmans will

welcome any katriya, even one who is disgraced and exiled. The Buddha concludes

his argument, in all three versions, by quoting a verse said to have been uttered by a

Brahm (the three versions disagree as to which Brahm). The Pali version of the

verse is as follows:

The katriya is best among those people who fall back on gotra.
He who is endowed with knowledge and [good] conduct is best among gods
and men. 864

Interestingly, we do indeed find this verse elsewhere, attributed, as claimed here, to a

Brahm (specifically Sanakumra in the Theravda tradition). In a sutta of the

Brahm Sayutta, in the Sagth Vagga of the Sayutta Nikya, Brahm

Sanakumra appears to the Buddha, utters this very verse, and the Buddha expresses

his approval. 865

Now, it is not immediately clear whether the story of Sanakumra uttering

this verse came first and then was incorporated into the Ambaha Sutta, or rather the

story was created for the sake of advancing the narrative in the Ambaha Sutta and

then was extracted as a separate sutta. There are clues, however, indicating that the

864
khattiyo seho janetasmi, ye gottapaisrino / vijjcaraasampanno, so seho
devamnuse //.
865
SN 1.6.2.1; cf. S 1190 and S2 103.

512
former is the case. To begin with, as Bhler pointed out over a hundred years ago, 866

there is a passage in the Mahbhrata that suggests that the Brahmanical tradition

was familiar with a story similar to this onenamely, that Sanakumra (Skt.

Sanatkumra) once spoke in praise of the preeminence of the katriyas. The context

of this story is a dispute between Atri and Gautama, in which Atri heaps praise on the

king Vainya, and Gautama criticizes him for this, saying that Indra is preeminent, and

accusing Atri of praising Vainya only in the hope of gaining gifts from him. To

resolve their dispute, the two go to Sanatkumra, who pronounces as follows:

Brahman is united with katra, and katra with brahman.

The king indeed is the first dharma and simply the lord of (his) subjects.
He indeed is akra and ukra; he is the Establisher; he is Bhaspati.

Prajpati, Virj, Sarj, katriya, Lord of the Earth, Lord of Men

Who that is lauded with these words is not worthy to be honored?

Of Ancient Origin and Conquering in Battle, Assailing, Joyful, Prosperous,

Guide to Heaven, Immediately Victorious, Tawnythus is a king called.

True in Wrath, Surviving in Battle, Setting in Motion the True Dharma.

The is, afraid of non-dharma, placed power in the katra.

The sun in the sky dispels darkness among the gods with its splendor.
Even thus does the Lord of Men mightily dispel non-dharma on the earth.

Hence the pre-eminence of the king (is established) by looking at the authority
of the stras.
The side by which King was said succeeds as superior. 867
866
Georg Bhler, Buddhas Quotation of a Gth by Sanatkumra, Journal of the Royal
Asiatic Society (July 1897): 585-8.
867
Mbh. 3.183.22-27: brahma katrea sahita katra ca brahma saha / rj vai
prathamo dharma prajn patir eva ca / sa eva akra ukra ca sa dht sa bhaspati //
prajpatir vir samr katriyo bhpatir npa / ya ebhi styate abdai kas ta nrcitum arhati //
puryonir yudhjic ca abhiy mudito bhava / svaret sahajid babhrur iti rjbhidhyate //
satyamanyur yudhjva satyadharmapravartaka / adharmd ayo bht bala katre samdadhan
// dityo divi deveu tamo nudati tejas / tathaiva npatir bhmv adharma nudate bham // ato
rja pradhnatva straprmyadarant / uttara sidhyate pako yena rjeti bhitam //.

513
This resolves the dispute in favor of Atri, and King Vainya, pleased with the praise

that has been bestowed on him, bestows riches on Atri in return. Bhler notes that

although the verse attributed to Brahm Sanakumra in the Buddhist texts certainly

does not correspond in any exact way to the extended eulogy found here in the

Mahbhrata, it does convey a similar idea and can easily be read as a summary of

the ideas found in the latter passage.

In making this argument, however, Bhler translates the verse attributed to

Sanakumra in the Pali texts differently than I have. My translation makes it clear

that two people are being spoken of. The katriya is referred to in the first pda as

best among those who fall back on gotra. Then, in the second pda, a person who

is endowed with knowledge and (good) conduct is said to be best among gods and

men. According to this reading, the two pdas refer to two entirely different types

of best persons, which are placed in parallel with one another for rhetorical effect.

I have chosen this translation because it fits best within the context of the Ambaha

Sutta. Within the latter context, the Buddhas citation of the verse serves as a turning

point, away from the argument over the superiority of katriyas in Section 4 to his

teaching on the best kind of person in general in Section 5. In Section 5, Ambaha

asks the Buddha what knowledge and conduct mean. The Buddha responds

using the Tathgata arises formula. With this, the debate over the relative

superiority of Brahmans and katriyas is dropped, and the argument shifts to the

absolute supremacy of a follower of the Buddha who puts the Buddhas teaching into

practice and ultimately attains Awakening. Overall, the point being made seems to be

514
that the claims of the new Brahmanism to vara supremacy are wrong on their face

because of the actual social supremacy of katriyas, but even more importantly

because they ignore what really makes a person great, which is not social rank, but

knowledge and conduct.

Bhler, however, gives a different translation of the verse in question.

According to his reading, both pdas of the verse refer to the same person, i.e., the

katriya: The katriya is best among those men who record their Gotras; endowed

with learning and virtue, he is best among gods and men. 868 This is an equally

acceptable translation of the Pali, and in fact can even be considered superior when

considering the verse in isolation, because there is nothing internal to the verse that

indicates that the so in the second pda refers to anyone other than the khattiyo of the

first. Moreover, Bhlers reading has the advantage of agreeing more closely with

Sanatkumras words in the Mahbhrata, which act exclusively as praise for the

king, and do not use the preeminence of the king as a mere springboard for discussing

a more abstract category of the ideal person. This, then, suggests another reason for

believing that the story of Sanakumra was older than the Ambaha Sutta and was

not simply created for the latters sake: It appears that it is being taken somewhat out

of context. That is, what Sanakumra said was simply a praise of katriyas as being

best among those interested in gotra and moreover, due to their knowledge and

conduct, best among gods and men. The Ambaha Sutta plays on the verse, however,

to make it not just about katriyas, but about an abstract (and particularly Buddhist)

868
Bhler, Buddhas Quotation, 587.

515
conception of the ideal person. Indeed, we find that in other places, this same verse is

taken even more out of context; in two other places in the Pali Canon, it is quoted in

contexts where the only point being made is to knowledge and conductthe

reference to katriyas and gotra being completely extraneous and irrelevant. 869

The quotation of the verse of Brahm therefore represents a turning point in

the Buddhas debate with Ambahafrom a direct assault on the new Brahmanical

claim to vara superiority based on birth to the formulaic teaching on the Buddhist

training that characterizes the entire Slakkhandha Vagga. In Section 6, the Buddha

solidifies Abahas defeat in the debate by listing four entrances to ruin

(apyamukhni) and asking if Ambaha and his teaching fulfill even these. These

entrances to ruin appear to describe the vnaprastha-rama, which we examined

in Chapter II.3, and which, I argued there, appears to correspond to the lifestyle of the

jailas that are held with a certain amount of esteem in the early Buddhist tradition.

The Buddha then continues by ridiculing the Brahmans claim to speak for the Vedic

is, saying that they do not speak for them simply by saying so any more than a

person would speak for King Pasenadi just by saying so. He then compares

Brahmans such as Ambaha and Pokkharastithat is, proponents of the new

Brahmanism who were householders and argued, as in the Dharma Stras, that non-

householder lifestyles were invalidunfavorably to the is, noting that the latter did

869
See MN 53 and AN 10.11. The verse is also quoted at the end of the Aggaa Sutta (DN
27), although there the question of vara superiority is clearly a relevant theme.

516
not live with all the luxuries of a householder, 870 while the former do. The point of

this line of argument is to establish a hierarchy of lifestyles and demonstrate that

Ambaha is, precisely the opposite of his original claim, at the dead bottom of that

hierarchy. At the top are those people who go forth as paribbjakas, train in the

Buddhas dharma, and attain Awakening. Next in line are jailas (or vnaprasthas, in

the language of the Dharma Stras), who do not ultimately accomplish Awakening

through their lifestyle, but at the very least lead lives in conformity with the Vedic

is of old. Finally, at the bottom are the proponents of the new Brahmanism, who

justify their luxurious lives by rejecting all lifestyles other than that of the

householder, and in so doing not only fail to make any progress toward Awakening,

but fail even to live up to the example of the is they claim falsely to speak for.

With this, we come to the end of the debate between the Buddha and

Ambaha, which, as we might expect, entails the utter and abject defeat of the latter.

In Section 7, Ambaha finally does what he came to dohe looks for the 32 marks

of a Great Man on the Buddhas body. This is treated in formulaic fashion: As with

every other Brahman who looks for the 32 marks on the Buddhas body, Ambaha is

able to see only 30, and the Buddha reveals the other twohis long tongue and

sheathed penisto him. In Section 8, Ambaha returns to his teacher, who is

enraged at him for embarrassing him in front of the Buddha. Finally, in Section 9,

Pokkharasti visits the Buddha personally, with a fair amount of variation among the

three sectarian traditions on the details of his meeting(s) with him. Thus, although the

870
On the theme of Brahmans in the Buddhas day not living up to the example of the Vedic
is, see also Sn. 2.7.

517
introduction of a Brahman student to the narrativeand one who is incredibly rude to

boot!entails a significant deviation from the pattern followed by the Soadaa

and Kadanta Suttas, the Ambaha Sutta returns to that pattern in the end by having

Pokkharastithe initially-introduced Brahman interlocutor who dwells in a

brahmadeyya land-grant given to him by the kingrespond positively to the

Buddhas teaching.

Taken together, these three encounter dialogs of the Slakkhanda/laskandha

collection in the Dgha/Drgha traditionthe Soadaa Sutta, the Kadanta Sutta,

and the Ambaha Suttashare a number of important characteristics. First and most

obviously, as members of the Slakkhanda/laskandha collectionwhich, as I

argued in Chapter IV.2, most likely formed the oldest core of the Dgha/Drgha

traditionthey all are built around the Tathgata arises formula that details the

training of a Buddhist monk from his initial going forth to his ultimate attainment of

Awakening. This indicates that these three texts were composed, most likely together,

according to a common formulaic pattern. More specifically, though, there are

various narrative and formulaic features that these three texts bear in common as

examples of the encounter dialog genre. First, all three of the Brahman interlocutors

involved have funny names. Soadaa is a dog-stick; Kadanta is sharp tooth,

ox tooth, or false teeth; and Ambaha isif Bronkhorst is correct in

Sanskritizing the name as Ambahaa miscegenistic half-breed. It is possible that

this narrative element was inherited from the Brhmaa Sayutta/Sayukta, where,

as we saw in Section IV.3.2, it is a defining feature.

518
 Certain oral formulae also serve to characterize these three texts as encounter

dialogs according to common patterns, especially at the beginning of the texts. All

three involve a Brahman interlocutor who resides in a village that was given to him

by a king as a brahmadeyya land-granta fact noted by a relatively short formula

used in these texts and many others throughout the early Buddhist tradition. In the

Chinese Dharmaguptaka tradition, this Brahman is also introduced by a common

Triple Veda formula that lauds his superior learning as a Vedic scholar. The

Brahman then becomes aware of the Buddhas presence in town (or nearby) through

the fame of Gotama formula. In the Soadaa and Kadanta Suttas, this leads to

a formulaic debate between the main interlocutor and a group of other Brahmans as to

whether he should go to see the Buddha or the Buddha should come to see him. The

Ambaha Sutta takes a different approach, using the fame of Gotama formula as a

springboard for having the Brahman interlocutor send his student to see if the Buddha

really does have the 32 marks of a Great Man. Although in this respect the Ambaha

Sutta deviates from the pattern set by the other two, the theme of the Brahman

seeking the 32 marks of the Buddha, and the formulae associated therewith, are, as

we will see in Chapter IV.4, common within the broader encounter dialog genre.

Finally, the actual narrative content of these three texts clearly marks them as

exemples of encounter dialogs. Each tackles, in some way, the claims being made by

proponents of the new Brahmanism. The Soadaa Sutta attacks the new

Brahmanical conception of Brahmanhood as being based on birth, Vedic learning,

and appearance. The Kadanta Sutta undermines the importance of the Vedic

519
sacrifice. And most damningly, the Ambaha Sutta attacks the primary claim of the

new Brahmanism, which is that society is divided into four varas with the Brahmans

at the top. Thus, with their narrative elements, extended use of oral formulae, and

incisive critiques of new Brahmanical ideology, these three texts represent highly

developed examples of the encounter dialog genre, especially when viewed in

comparison with the stras of the Brhmaa Sayutta/Sayukta. While the stras of

the latter served to convert verses into short, most likely humorous jabs at

Brahmanical pretensions, these long dialogs of the Dgha/Drgha tradition make use

of well-established formulaic patterns to mount extended and detailed rhetorical

attacks on new Brahmanical ideology. In so doing, they represent the epitome of the

encounter dialog genre, and most likely as a result have had the greatest influence on

later perceptions of antagonism between early Buddhism and Brahmanism.

IV.3.4 The Brhmaa Vagga of the Majjhima Nikya and Brhmaa Varga of the

Madhyamgama

With the Majjhima Nikya/Madhyamgama tradition, we begin to move

beyond the realm of clearly old collections of encounter dialogs and into the realm of

encounter dialogs that were only organized into groups relatively late in the

development of the early Buddhist tradition. As already noted in Chapter IV.2, both

the Pali Majjhima Nikya of the Theravdins and its only full parallel, the

Sarvstivdin Madhyamgama preserved in Chinese translation, contain a

vagga/varga named Brahman (P. brhmaa, Ch. ). These two collections, as

520
already shown, are drastically different. The Pali Brhmaa Vagga contains only ten

suttas, while the Chinese Brhmaa Varga contains 20 stras; moreover, only four

stras are held in common between the two collections, accounting for less than half

of the shorter collection. Although it is possible that these stras held in common

represent the remnants of an earlier Brahman collection in the

Majjhima/Madhyama tradition that then developed into the two sectarian versions we

have now, it is equally possible, as I argued above, that the two Brahman

collections were created completely independently by the Theravda and

Sarvstivda traditions. It appears to have been common in the early Buddhist

tradition to group together stras that involve Brahmans, and doing so from the

common stock of Majjhima/Majjhima stras would have inevitably led to similar

results, even if done independently.

Closer inspection of the contents of these two collections does not reveal

much in the way of further patterns between the two collections, although it does

reveal certain continuities between some of Majjhima/Madhyama encounter dialogs

and the broader encounter dialog genre. As before, let us begin with the Theravda

tradition. A list of all the encounter dialogs in the Majjhima Nikya and their

parallels in Chinese (mostly in the Madhyamgama) can be found in Table 10.

Nearly half of these encounter dialogs are found, as expected, in the Brhmaa

Vagga (MN 91-100). Another twelve suttas, however, are found outside of the

Brhmaa Vagga, scattered throughout various other vaggas. Looking at these, we

find a pattern similar to that observed earlier with respect to the Sayutta Nikya. In

521
Section IV.3.2, I noted that many of the encounter dialogs outside of the primary

encounter-dialog collection of that Nikya (i.e., the Brhmaa Sayutta) are

nominal encounter dialogsthat is, they are encounter dialogs only insofar as the

occasion for the Buddha providing a teaching is provided by an interlocutor who

happens to be a Brahman. Here in the context of the Majjhima Nikya, we find that

the same is true of those encounter dialogs that are found outside the Brhmaa

Vagga. These scattered suttas, for the most part, are just teachings that happen to

be given to Brahmans.

Pali Sutta (MN) Formulae Comments Parallels

4: Bhayabherava Jussoi asks E 31.1
Buddha how
monks deal with
fear in forest.
27: Cahatthipadoma Who am I to Paribbjaka M 146
praise Pilotika uses simile
Gotama? of elephants
Tathgata footprint to praise
arises (partial) Buddha to
Jussoi, who
then goes to
Buddha to recount
simile and is taught
about the training.
30: Casropama 6 heretical Pigalakoccha asks E 43.4
teachers Buddha if any of 6
8 jhnas heretical teachers
have direct
knowledge;
Buddha replies
with heartwood
simile.
41/2: fame of Brahman S 1042/3
Sleyyaka/Verajaka Gotama householders of
Sl/Verajaka ask
Buddha about
rebirth; Buddha

522
 teaches about
different types of
conduct.
60: Apaaka fame of Buddha teaches Skt. D 7
Gotama Brahman
Tathgata householders of
arises Sl about wrong
views and then
gives own teaching.
75: Mgandiya fame of Not really an M 153
Gotama encounter dialog:
(partial) Buddha stays in
fire chamber of a
Bhradvja
Brahman and
encounters
paribbjaka
Mgandiya, who
disapproves of him
staying there;
teaches him about
dangers of sensual
pleasures.
82: Rahapla fame of Brahman M 132, T. 68-9
Gotama householders of
Thullakohita go to
Buddha; Rahapla
decides to ordain
and has trouble
because of parents.
91: Brahmyu Triple Veda Brahmyu sends M 161, T. 76
fame of student Uttara to
Gotama look for 32 marks,
32 marks then goes to see
himself.
92: Sela fame of Jaila Keiya E 49.6, T.
Gotama offers Buddha 1428.42, Sn. 3.7
Triple Veda meal; Brahman
32 marks Sela sees this,
decides to visit
Buddha, and sees
32 marks; Sela
ordains and Keiya
gives meal.

523
93: Assalyana Triple Veda Student Assalyana M 151, T. 71
born of sent by Brahmans
Brahms to debate
mouth superiority of
Brahmans.
94: Ghoamukha Ghoamukha tells
Udena no
paribbjaka
lifestyle is in
accord with
Dhamma.
95: Cak Brahmadeyya Cak visits Skt. D 19
fame of Buddha and
Gotama encourages him to
Disagreement listen to student
over who Kpahika, who
should visit defends Brahmans
whom claim to
Triple Veda superiority.
96: Esukr Esukr debates 4 M 150
levels of service
and 4 types of
wealth (i.e., vara
system) with
Buddha.
97: Dhnajni Sriputta teaches M 27
negligent Brahman
Dhnajni how to
reach brahmaloka
before he dies.
98: Vseha famous Vseha and Sn. 3.9
Brahmans Bhradvja
fame of disagree on
Gotama whether
Brahmanhood
depends on conduct
or birth and ask the
Buddha.
99: Subha Todeyyas student M 152
Subha asks Buddha
about Brahmans
rejection of non-
householder
lifestyles.

524
100: Sagrava Triple Veda Sagrava Skt. D 22
persuaded by
Brahman lady to
visit Buddha; asks
Buddha if he
believes in gods;
Buddha answers
with story of quest
for Awakening.
107: Gaakamoggallna Gaaka M 144, T. 70
Moggallna asks
Buddha about
gradual training.
108: Gopakamoggallna Gopaka M 145
Moggallna asks
nanda if any
monk has all
qualities of
Buddha.
135: Todeyyas student M 170, T. 78-
Cakammavibhaga Subha asks Buddha 81, T. 755
about karma.
150: Nagaravindeyya fame of Brahman S 280
Gotama householders of
Nagaravinda go to
Buddha; Buddha
tells them which
samaa-brhmaas
to honor.
152: Indriyabhvan Student Uttara tells S 282
Buddha his teacher
Prsariya teaches
suppression of
sense faculties;
Buddha teaches
proper cultivation
of sense faculties.
Table 8. A list of the encounter dialogs in the Majjhima Nikya and their parallels. Aside from the Sn. and Skt.
D, all parallels are in Chinese. Also listed are formulae characteristic of the encounter dialog genre and a brief
synopsis (from the Pali version). Suttas that have a parallel in the Brhmaa Vagga of the Sarvstivdin
Madhyamgama are highlighted.

This is in stark contrast to the suttas of the Brhmaa Vagga. Nearly all of

these ten suttas involve a substantive encounter with a Brahman or Brahmans in

525
which the claims of the Brahmans, their hostility toward samaas, or at the very least

their otherness in comparison to the Buddha is somehow at stake. Most of the

suttas involve a debate over some aspect of the claims made by proponents of the

new Brahmanism. In the Assalyana Sutta (93), the student Assalyana is sent by his

teachers to debate with the Buddha over the superiority of Brahmans to other varas.

Similarly, in the Cak Sutta (95), the Brahman student Kpahika defends the

Brahmans claim to superiority to the Buddha. In both the Ghoamukha Sutta (94)

and the Subha Sutta (99), the topic addressed is the new Brahmanical claim,

championed by the Dharma Stras, that only the householder lifestyle is valid. In the

Esukr Sutta (96), a Brahman of the same name debates with the Buddha over the

claim that each vara, beginning with the Brahmans, is to be served by lower varas,

and that there are occupations proper to each vara. And in the Vseha Sutta (98),

the Brahmans Vseha and Bhradvja go to the Buddha to settle their dispute over

whether Brahmanhood is determined by birth or conduct.

The remaining suttas in the vagga do not directly address such Brahmanical

claims, but they do nonetheless touch upon themes relating to the otherness of the

Brahman interlocutors and, in most cases, the hostility of their broader community

towards the Buddha. Both the Brahmyu Sutta (91) and the Sela Sutta (92) involve

Brahmans going to see if the Buddha really has the 32 marks of a Great Mana skill

of prognostication said to be the province of Brahmansand in both cases the

propriety of Brahmans visiting or attending to the Buddha arises as a theme. In the

Sagrava Sutta (100), the Brahman Sagrava visits the Buddha, but only after

526
getting angry at a Brahman lady for her faith in the Buddha and being convinced by

her that he should give the latter a closer look. Finally, in the Dhnajni Sutta,

Sriputta visits the Brahman Dhnajni, who is an old friend and does not appear, in

contrast to most of the other Brahmans in this vagga, at all hostile to the Buddha or

his teaching. Interestingly, though Sriputta teaches him how to attain brahmaloka

instead of full Awakening, because these Brahmans are intent on the world of

Brahm 871for which he is criticized by the Buddha in the Pali version, but praised

by him in the Chinese. 872

Throughout the encounter dialogs of the Majjhima Nikya, including both

those within the Brhmaa Vagga and those without, we find formulae characteristic

of the genre such as we already saw in the encounter dialogs of the Dgha/Drgha

tradition. These formulae are, however, more diffusely scatteredthat is, we do not

find such a high concentration of Brahman-related formulae as in the Ambaha,

Soadaa, and Kadanta Suttas, and in fact some of the more memorable encounter

dialogs of the Majjhima (Ghoamukha, Esukr, Subha) largely do without these

characteristic formulae. Where we do find formulae, however, the most common by

far is the fame of Gotama formula, which serves to explain how a Brahman or

Brahmans come to hear of the Buddha and become interested in seeing him in nine

encounter dialogs of the Majjhima Nikya. Five of the ten suttas in the Brhmaa

Vagga use the Triple Veda to introduce the primary Brahman interlocutor. A few

871
ime kho brhma brahmalokdhimutt.
872
Anlayo, Comparative Study of the Majjhima-Nikya, 570-1.

527
of the suttas in this vagga also bear some other similarities to the encounter dialogs of

the Dgha: The Cak Sutta (95) includes the same formulaic argument over whether

a Brahman going to see the Buddha was beneath his dignity that we found in the

Soadaa and Kadanta Suttas. The Brahmyu (91) and Sela (92) Suttas include

the theme of the Brahman going to look for the 32 marks of a Great Man on the

Buddhas body, just as we saw in the Ambaha Sutta. Perhaps most interesting of all,

though, in two Majjhima encounter dialogs (27 and 60), both of which are outside of

the Brhmaa Vagga, we find included at least part of the Tathgata arises

formulathat is, the formula on the training of a monk found in all of the suttas of

the Slakkhandha Vagga in the Dgha Nikya, and their parallels in other sectarian

traditions.

These commonalities in use of formulae and themes show that there are

continuities between the encounter dialogs of the Dgha/Drgha and

Majjhima/Madhyama traditions, even though the Theravda and Sarvstivda

traditions largely agree on their classification of them between the two

Nikyas/gamas. As already shown in Chapter IV.2, there are fairly substantive

differences between the Theravdin Majjhima Nikya and the Sarvstivdin

Madhyamgama, including a block of ten stras in both directions that are switched

between the Dgha/Drgha and the Majjhima/Madhyama. As can be seen from Table

10, however, this switch affects only three encounter dialogs found in the Majjhima

Nikyathe Apaaka (60), Cak (95), and Sagrava (100) Suttasand most of

the remaining suttas have their parallels in the Sarvstivdinn Madhyamgama. Thus,

528
all of the formulae and themes noted here are attested, by both the Sarvstivda and

Theravda sectarian traditions, in both the Dgha/Drgha and Majjhima/Madhyama

Nikya/gama traditions, and thus can be considered as cross-tradition markers of the

encounter dialog genre. We will be investigating these formulae and themes further

from a cross-tradition perspective in Chapter IV.4.

Turning now to the Sarvstivdin Brhmaa Varga preserved in the Chinese

translation of that sects Madhyamgama, we find a much more expansive collection

than the Pali Brhmaa Vagga that is based more loosely on the common theme of

the Brahman than an adherence to heavy encounter dialog themes. A list of the

stras in this varga with their parallels can be found in Table 11. Four of the twenty

total, as already noted, are held in common with the Brhmaa Vagga of the

Majjhima Nikya, and as such, they address various new Brahmanical themes,

including the supremacy of the Brahmans over other varas, the vara system in

general, and the invalidity of non-householder lifestyle lifestyleas well as, in stra

no. 161, which corresponds to the Brahmyu Sutta, the theme of the Brahman seeking

the Buddhas 32 marks. Most of the other stras in the varga have their Pali parallel

in either other parts of the Majjhima Nikya or in the Aguttara Nikya. Some of

these do address themes relating to differences between the Buddhist position and that

of the new Brahmanism, including no. 143, which corresponds to a Pali sutta that

addresses the Vedic sacrifice; no. 157, which corresponds to a Pali sutta that

addresses the Buddhas lack of respect for Brahmans; and no. 158, which corresponds

to a Pali sutta in which the Buddha divides Brahmans into five types according to

529
ever-decreasing virtue with regards to marriage. Others, however, appear to be

included simply because they refer to Brahmans in some way. The Pali parallel to no.

155, for example, is not an encounter dialog at all, but does include a short Jtaka in

which the Buddha is a Brahman who gives away much of his wealth. It appears, then,

that whether there was any historical relationship between the Theravdas Brhmaa

Vagga and the Sarvstivdas Brhmaa Varga or not, the two traditions maintained

the boundaries of their Majjhima/Madhyama Brahman collection differently, with

the Sarvstivdins opening it to almost any text that referred to Brahmans and the

Theravdins restricting it to encounter dialogs that contributed significantly to the

construction of a Buddhist identity separate from that of the new Brahmanism.

Chinese Stra Parallels Comments

(M)
142: AN 7.20, E 40.2 Story of Magadhan minister Vassakra
going to Buddha for advice on attacking
Vajjians.
143: AN 3.60 Sagrava debates with the Buddha
over the relative value of Vedic
sacrifice and the life of the samaa.
144: MN 107, T. 70 Gaakamoggallna Sutta
145: MN 108 Gopakamoggallna Sutta
146: MN 27 Cahatthipadopama Sutta
147:
148: AN 5.31 Pali version involves no Brahmans, only
rjs daughter Suman.
149: AN 6.52, E 37.8 Jussoi asks the Buddha about
different classes of people (Brahmans,
katriyas, gahapatis, women, thieves,
samaas).
150: MN 96 Esukr Sutta
151: MN 93, T. 71 Assalyana Sutta
152: MN 99 Subha Sutta
153: MN 75 Mgandiya Sutta

530
154: DN 27, Ch. D 5, T. Aggaa Sutta
10, E 40.1
155: AN 9.20, T. 72-4, E Not really an encounter dialog. Buddha
27.3 has discussion with Anthapiika,
during which he tells the story of a
Brahman named Velma (Buddha in
past) who gave great gifts.
156:
157: AN 8.11, T. 75 Brahman of Veraj criticizes Buddha
for not showing respect to Brahmans.
158: AN 5.192 Buddha tells Brahman Doa about five
types of Brahmans.
159:
160: AN 7.69-70 Not really encounter dialogs. Buddha
tells stories about teachers in the distant
past who appear to have been Brahmans
or Brahman-like.
161: MN 91, T. 76 Brahmyu Sutta
Table 9. A list of the stras in the Brhmaa Varga of the Sarvstivdin Madhyamgama and their parallels,
with some identifying information.

IV.3.5 Encounter Dialogs in the Aguttara Nikya

As already discussed in Chapter IV.2, when it comes to the fourth and final of

the four major Nikyas, we are in the unfortunate position of not having, as with the

previous three, an extant Sarvstivda version with which to compare the Theravda

version in Pali. This renders any comparison of these two sectarian Nikya/gama

traditions incomplete. On the other hand, we do have the consolation in the case of

this Nikya of having a parallel version that is more different from the Pali than any

other extant non-Pali gama, and that may in fact come from the Mahsghika

tradition, a tradition which has remained fairly mysterious to modern scholarship due

to a lack of preserved material. This parallel to the Aguttara Nikya, which we refer

to conventionally as the Chinese Ekottarikgama, is radically different from the Pali,

531
holding only a minority of stras in common with the latter. Moreover, as we saw

from our investigation of the Majjhima/Madhyama tradition in Section IV.2.4, it

appears that the Sarvstivda and Theravda traditions differed greatly in their

classification of stras between the Majjhima/Madhyama and Aguttara/Ekottarika

collections; in particular, many suttas that the Theravdins placed in their Aguttara

Nikya, the Sarvstivdins instead placed in their Madhyamgama. This, together

with certain internal evidence in the Aguttara Nikya and the fact that the latter

differs so greatly from the only other complete Ekottarika collection to have been

preserved, suggests that the Theravdins moved a large number of suttas into their

Aguttara in order to fill it out. Comparison with the Chinese Ekottarikgama,

therefore, has the potential to reveal a core of Aguttara/Ekottarika stras that were

a part of that collection quite early in the tradition.

As it has come down to us in the Theravda tradition, the Aguttara Nikya

contains a fairly large number of encounter dialogsmore than any of the other three

main Nikyas, in fact. These can be found listed, along with their parallels, in Table

12. As can be seen from this table, the number of parallels to encounter dialogs in

this Nikya, especially in comparison to other Nikyas, is quite limited. More than

half of the encounter dialogs, in fact, have no known parallel whatsoever. 873 Of those

that do, most have their parallel in the Sarvstivdin Sayuktgama. Since we

already know that the Theravdins moved stras into their Aguttara that the

Sarvstivdins placed elsewhere, this may lend support to the theory that the

873
No complete parallels, that is. I have not included parallels found in fragments.

532
Sayuktgama is the basis of the other Nikya/gama traditions. That is, these

suttas may be encounter dialogs that originated in the Sayutta/Sayukta collection,

and that the Theravdins moved to their Aguttara, but that the Sarvstivdins did not

move (such as into their Madhyama). On the other hand, there are only four suttas

out of the total of 44 encounter dialogs in the Aguttara Nikya that have a parallel in

the Chinese Ekottarikgama. This suggests that the encounter dialog genre did not

play an important role in the earliest Aguttara/Ekottarikgama tradition. This is not

entirely surprising, given that the numerical principle on which the collection is based

is more conducive to doctrinal lists than to narratives, but it does strongly suggest that

whatever encounter dialogs did come to be found in Aguttara/Ekottara collections

were, in most cases, brought there from outside.

Pali Sutta Formulae Comments Parallels

(AN)
2.2.7 Jussoi asks Buddha why people
are reborn in hell.
2.4.3 Brahman asks Buddha what view he
holds; Buddha says he holds view of
action and inaction.
3.51 Brahmans Two elderly Brahmans seek
advanced in recognition from Buddha, but get
years none.
3.52 Brahmans Similar to 3.51. S 1163,
advanced in S2 86-87
years
3.53 Brahman asks how Dhamma is seen
in this life. (lust)
3.54 Similar to 3.53.
3.55 Jussoi asks how Dhamma is seen
in this life. (lust)
3.56 Mahsla Brahman asks reason for E2 14
changes in population. Buddha says
it is due to lust.

533
3.58 Triple Veda Tikaa praises tevijja Brahmans
and Buddha gives his interpretation
of tevijja.
3.59 Triple Veda Jussoi says people should give
only to tevijja Brahmans. Buddha
asks how they are tevijja and then
gives his own interpretation.
3.60 Sagrava says sacrifice is more M 143
meritorious than homeless life.
3.63 fame of Brahman householders of
Gotama Vengapura visit Buddha, and
Vacchagotta asks if Buddha uses
high seats. Buddha says he only
uses 3 metaphorical high seats
(meditative states).
4.22 Brahmans come to Buddha right
after his Awakening and asks if he
pays respect to elder Brahmans. He
says no and explains what makes a
true elder.
4.35 Vassakra discusses with Buddha
what makes a Great Man.
4.36 Doa sees wheel-marks in Buddhas S 101,
footprints and asks what type of S2 267,
being he is. E 38.3
4.39 Ujjaya asks Buddha about sacrifice. S 89,
S2 89
4.40 Same as 4.40, except Brahman is S 90,
named Udyi, and final verses are S2 90
different.
4.183 Vassakra asks Buddha about right
speech.
4.184 Jussoi says all fear death, but
Buddha says ideal person does not.
4.187 Vassakra asks Buddha if a bad man
can recognize a good/bad man;
Buddha says only a good man can.
Followed by story of company of
Todeyya ridiculing Eeyya and his
court for respecting Rmaputta.
4.233 Brahman Sikha Moggallna tells
Buddha that Brahman student
Soakyana said Buddha denies

534
 karma. Buddha denies this and says
he doesnt know the kid.
5.30 fame of Brahman householders of S 1250
Gotama Icchnagala try to feed Buddha, but
Buddha refuses.
5.192 Triple Veda Doa asks Buddha if it is true that M 158
seers of old he does not respect elder Brahmans.
Buddha describes 5 types of
Brahmans.
5.193 Sagrava asks Buddha why he SN 46.55
sometimes can remember mantras
and sometimes not.
5.194 Who am I to Pigiynin praises Buddha to
praise Gotama? Kraaplin.
5.195 Pigiynin praises Buddha in front
of Licchavis.
6.38 Brahman tells Buddha he believes in S 459
neither self- nor other-agency.
6.42 fame of Brahmans of Icchnagala try to S 1250-1
Gotama feed Buddha, but he refuses.
6.48 Brahman asks Buddha about
Dhamma in this life.
6.52 Jussoi asks about various types M 149,
of people (Brahman, katriya, E 37.8
gahapati, woman, samaa).
6.53 Brahman asks what is of benefit
now and hereafter.
7.20 Ajtasattu sends Vassakra to M 142,
Buddha for advice on attacking E 40.2
Vajjians.
7.44 Uggatasarra goes to Buddha for S 93,
advice on performing sacrifice; S2 259
Buddha talks about 3 fires.
7.47 Jussoi asks Buddha if he E 37.9
practices brahmacariya.
8.11 Brahman from Veraja asks if M 157,
Buddha pays respect to elder T. 75
Brahmans; Buddha says no and
explains how he attained
Awakening.
8.55 Ujjaya asks for teaching from S 91,
Buddha before going abroad. S2 91
8.86 Brahmans try to feed Buddha, but S 1250

535
 Buddha refuses.
9.38 Two Lokyata Brahmans ask
Buddha whether Praa Kassapa
(world is finite) or Nigaha
Ntaputta (world is infinite) is right.
Buddha dismissed question and
teaches about four jhnas.
10.117 Sagrava asks about near and far
shore.
10.119 Jussoi is about to perform
descent ceremony of Brahmans, and
Buddha gives his own version.
10.167 Same as 10.119, but with different S 1040
interpretation of descent ceremony.
10.169 Same as 10.119, but with different
interpretation of descent ceremony.
10.177 Jussoi asks if rddha rite is S 1041
efficacious; Buddha says it depends
on merit of deceased person.
19.209 Brahman asks why people are
reborn in hell.
Table 10. A list of the encounter dialogs found in the Aguttara Nikya, along with their parallels. Signifcant
formulae and a brief description of each sutta are also given.

As we can see from Table 12, encounter dialogs are scattered fairly diffusely

throughout all but the eleventh and final nipta of the Aguttara Nikya. There are

three vaggas entitled Brhmaa in the Aguttara Nikyaone each in the third,

fourth, and fifth niptasbut the title Brhmaa Vagga in these cases, unlike the

Brhmaa Vagga of the Majjhima Nikya, is mostly nominal. Although they do each

contain several suttas that pertain to Brahmans in some way, they also contain suttas

that do not; it appears, therefore, that the title Brhmaa Vagga was given merely as a

convenient label and not as a deep organizing principle. Again, this is not entirely

surprising; the primary organizing principle of the Aguttara/Ekottarika tradition is

numerical, and the division into vaggas of ten suttas each is clearly secondary.

536
 The encounter dialogs of the Aguttara Nikya are a mixture of a few

nominal encounter dialogsin which the encounter with a Brahman interlocutor

simply provides an occasion for the Buddha to teachand a fairly large number of

more substantive examples of the encounter genre, in which the tension between

Buddhist and Brahmanical identities is more evident. Themes include the Vedic

sacrifice, other Brahmanical rites (including rddha), the meaning of tevijja, and the

Buddhas lack of respect for elderly Brahmans. Interestingly, however, none of the

four encounter dialogs that are held in common with the Chinese Ekottarikgama

address topics such as these; instead, they can be classified as nominal encounter

dialogs in which the Braman interlocutor is simply an occasion for the Buddha to give

a teaching. This would suggest that the type of encounter dialog we are most

interested inin which the claims of the new Brahmanism are addressed and a

tension between Buddhist and Brahmanical identities is evidentwas not found in

the earliest Aguttara/Ekottarika tradition.

Formulae characteristic of the encounter dialog genre are found in the

Aguttara Nikya, but they are by no means common. The Triple Veda formula

used to introduce a Brahman interlocutor is found three times, as is the fame of

Gotama formula used to bring the Buddha to the attention of a Brahman interlocutor.

While such formulae are not frequently found here, we do find a different pattern not

encountered in other Nikyas of the Pali traditionnamely, individual Brahmans

who appear in multiple suttas as reoccurring characters. The most frequently

appearing and well known of these is Jussoi, who appears in a total of nine suttas,

537
but others such as Vassakra and Sagrava make repeat appearances as well. These

will be discussed further in Chapter IV.4.

IV.3.6 Encounter Dialogs in the Sutta Nipta

The only remaining place, within the Pali tradition, in which we find a

significant number of what can be considered encounter dialogs is the Sutta Nipta.

Seven suttas within this collection can be considered as belonging to this genre. All

but one of these seven has a parallel within the ordinary Nikya/gama traditions that

we have already examined. Nevertheless, there is enough variation between the

encounter dialogs found in the Sutta Nipta and related encounter dialogs found

elsewhere to warrant a brief survey of the former as a group. All seven of the

relevant suttas can be found listed in Table 13, together with a brief commentary on

parallels.

Sutta Nipta sutta Comments

1.4: Kasibhradvja Buddha criticized by Farmer Bhradvja for not
growing own food. Similar to SN 1.7.2.11 of the same
name, but with a different ending after the verses.
1.7: Vasala Buddha called outcaste (vasala) by Aggika
Bhradvja; gives lengthy response in verse. Although
Pali Brhmaa Sayutta has a sutta involving an Aggika
Bhradvja (SN 1.7.1.8), the content is not the same; a
true parallel is found in the Sarvstivdin Brhmaa
Sayukta (S 102, S2 268).
2.7: Brhmaadhammika Brahmans ask Buddha if contemporary Brahmans live up
to the standards of the Brahmans of old, and Buddha tells
them the many ways in which they do not. No known
parallel.
3.4: Sundarika Bhradvja wants to give Buddha donation
Sundarikabhradvja and asks if he is a Brahman. Prose introduction is
identical to SN 1.7.1.9 of the same name, but verses are
longer and completely different.

538
3.5: Mgha Mava Mgha visits Buddha and asks about making
merit by giving. Parallel to S 1159=S2 82, in the
Sarvstivdin Brhmaa Sayukta.
3.7: Sela Same as MN 92.
3.9: Vseha Same as MN 98.
Table 11. A list of encounter dialogs in the Sutta Nipta, with a brief commentary on parallels.

The last two of these encounter dialogs, the Sela Sutta (3.7) and Vseha

Sutta (3.9) in the Mah Vagga, are identical to two suttas of the same names in the

Brhmaa Vagga of the Majjhima Nikya (MN 92 and 98, respectively). None of the

remaining five encounter dialogs has an exact parallel within the Pali tradition, but

four of them do have clear links to the Brhmaa Sayukta/Sayutta collection of

the early Sayutta/Sayukta tradition. Three of these four, in fact, involve a

Brahman interlocutor who is identified by a given name and the gotra Bhradvja,

which, as we saw above in Section IV.3.2, appears to have been a characteristic

narrative feature of stras in the early Brhmaa Sayukta/Sayutta collection. Two

of these suttas have partial parallels within the Pali Brhmaa Sayutta, and the

other two, though lacking parallels in the Pali Sayutta, do have parallels in the

Sarvstivdin Brhmaa Sayukta.

The first (1.4) involves a certain Farmer (kasi) Bhradvja who criticizes

the Buddha for not growing his own food. The Buddha responds in verse, explaining

the sort of metaphorical farming he engages in. The Brahman then offers the

Buddha food, but the Buddha rejects it, speaking again in verse, because it has been

sung over with verses (gthbhigta). Up to this point, the Sutta Nipta version is

in complete agreement with the Sayutta version. After the Buddha rejects the food,

however, the Sutta Nipta version has the Brahman ask if he should give the food to

539
someone else, and the Buddha says that even though he has rejected it, he alone in the

whole world is capable of digesting it. The Brahman, at the Buddhas instruction,

throws the food away into a nearby river and is amazed to see that it makes the water

sizzle and boil. He then requests ordination and becomes an arhat. This ending

agrees with the parallel to the Kasibhradvja Sutta in the Sarvstivdin Brhmaa

Sayukta. In the Pali Brhmaa Sayutta, a similar ending, with the Buddha

ordering that food he has rejected be thrown out because only he can digest it, and the

water sizzling when the food is thrown into it, is found in another, otherwise

unrelated sutta, the Sundarikabhradvja Sutta (SN 1.7.1.9). It is not, however,

found in the Sayutta version of the Kasibhradvja Sutta. In the latter, after the

Buddha rejects the Brahmans food, the Brahman simply expresses his amazement at

the Buddhas words, takes refuge, and commits himself as a lay follower. There is no

discussion of what to do with the food, and the Brahman does not ordain or become

an arhat. 874

Interestingly, the Sundarikabhradvja Sutta, which, in the version preserved

in the Brhmaa Sayutta, has a similar ending to that of the Sutta Nipta version of

the Kasibhradvja Sutta, has its own parallel in the Sutta Nipta (3.4). The

874
This is consistent with the fact that the Kasibhradvja Sutta, within the context of the
Brhmaa Sayutta, in found within the second vagga, entitled Upsaka, and not in the first vagga,
entitled Arahanta. Recall from Section IV.3.2 that the Theravda tradition, unlike the Sarvstivda
tradition, organized its Brhmaa Sayutta into two vaggas, the first of which is comprised of suttas
in which the Brahman interlocutor ends up ordaining and becoming an arhat, and the second of which
is comprised of suttas in which the Brahman interlocutor merely becomes a lay disciple. It is
interesting that the Theravdins preserved this particular sutta with an ending in which Kasi
Bhradvja becomes a lay disciple and thus placed it in the second vagga even though, with its
interlocutor named Kasi Bhradvja, it clearly follows the pattern established in the first vagga and,
moreover, both the Sutta Nipta and the Sarvstivdin Brhmaa Sayukta versions of the same sutta
agree that Kasi Bhradvja did ordain and become an arhat.

540
Sayutta and Sutta Nipta versions of the Sundarikabhradvja Sutta, however,

agree only with respect to the short prose introduction. A Brahman by the name of

Sundarika Bhradvja, after performing a agnihotra (P. aggihotta), looks for

someone to eat the remnants of the offering. He sees the Buddha and is reluctant to

give it to him because he has a shaved head, but then he remembers that some

Brahmans have shaved heads (presumably as parivrjakas) and decides to ask him if

he is a Brahman. The Buddha answers in verses, but the verses are completely

different in the two versions. In the Sayutta version, the Buddha responds briefly

by saying that the question is improper since conduct, not birth, is what truly matters;

then, the Buddha rejects the food offering and orders it thrown away into the river,

resulting in the sizzling and boiling of the water as already described above. In the

Sutta Nipta version, however, the Buddha gives a much more extended response,

beginning with a denial that he belongs to any vara; in the end, he rejects the food as

in the Sayutta version, but there is no discussion of what to do with the food. In this

case, unlike the previous, the Sarvstivda version agrees with the Sayutta version

in Pali and not the Sutta Nipta version.

Two remaining encounter dialogs in the Sutta Nipta lack parallels in the Pali

Brhmaa Sayutta, but do have parallels in the Sarvstivdin Brhmaa Sayukta.

The first of these (1.7) involves an interlocutor named Aggika Bhradvja and thus,

on the surface, would appear to be parallel to the Aggikabhradvja Sutta in the Pali

Brhmaa Sayutta (SN 1.7.1.8). The content of these two suttas is entirely different,

however. In Sn. 1.7, entitled the Vasala Sutta, Aggika Bhradvja calls the Buddha

541
an outcaste (vasala), opening himself up to a long lecture from the Buddha, in verse,

on what the term vasala truly means. This story, while not agreeing with the story of

Aggika in the Pali Brhmaa Sayutta, does agree with that of a Brahman of the

same name in the Sarvstivdin Brhmaa Sayukta (S 102=S2 268). The Mgha

Sutta (Sn. 3.5), on the other hand, involves an interlocutor, namely the mava

Mgha, who does not appear in the Pali Brhmaa Sayutta at all, although he and

his story (essentially a conversation with the Buddha over making merit by giving)

are found in the Sarvstivdin Brhmaa Sayukta (S 1159=S2 82).

This leaves one remaining encounter dialog, the Brhmaadhammika Sutta

(2.7), that does not have any known parallel outside of the Sutta Nipta. In this sutta,

a group of aged mahsl Brahmans approach the Buddha and ask him if Brahmans

in their own day meet the standards of the Brahmans in the past. The Buddhas

response, entirely in verse, is quite interesting, and although it has no parallel

elsewhere in and of itself, it clearly expresses ideas that are found in the other, more

widely known encounter dialogs of the main Nikya/gama traditions that we have

already examined above. The Buddhas response, in short, is no, contemporary

Brahmans are not like the Brahmans of old. According to the Buddha, the Brahmans

of old did not accumulate wealth, did not indulge themselves in sensual pleasures, did

not kill living beings for their sacrifices, 875 and took wives properly (through mutual

consent) and had sex with them only at the proper time (not during the period).

875
Interestingly, this sutta puts great emphasis on the supposed fact that Brahmans of old,
unlike contemporary Brahmans, did not kill cows. Ironically, one of the oldest texts to speak of the
holy cow, therefore, would be this, a Buddhist text.

542
These themes are touched upon, as we have already seen, in the Ambaha Sutta, in

which the Buddha gets Ambaha to admit that he and his teacher, with their wealth

and indulgence in sensual pleasures, are nothing like the ancient is; in the

Kadanta Sutta, in which the Buddha describes in a Jtaka an ancient sacrifice that

did not involve the slaughter of any animals; and in the Doa Sutta (AN 5.192), in

which the Buddha criticizes Brahmans for the improper acquisition of and relations

with their wives. What the Brhmaadhammika Sutta adds to these themes, however,

is an explanation of how contemporary Brahmans came to be so degenerate, with

their wealth, indulgence in sensual pleasures, and violent sacrifices. According to the

Buddha, Brahmans became jealous of the wealth and women possessed by kings and

wanted to possess such things for themselves. They therefore made an arrangement

with King Okkka (Skt. Ikvku), such that they would perform animal sacrifices for

the king, and compose hymns for that purpose, and in return they would receive the

wealth sacrificed by the king. 876

As far as we can tell from the evidence available to us, the Sutta Nipta is a

purely sectarian collection, existing only in the Theravdin tradition. This does not

mean that individual texts found within it cannot possibly be oldindeed, as we saw

in Chapter III.1, the Ahaka, Pryaa, and Khaggavisa Sutta are all likely of

extraordinary antiquityonly that the collection as a whole is relatively young and

876
Interestingly, the process described here, though expressed in extremely polemical terms,
does bear a certain resemblance to the historical process by which the rauta ritual was codified for the
sake of state legitimation, as described by modern scholars such as Jan Heesterman and Michael
Witzel. I am of course not suggesting that the Buddhas account here in the Brhmaadhammika Sutta
is historical, but rather that it is offering a polemical account of an actual historical process
namely, the development of the rauta ritual and concomitant growth of a specialized class of ritual
specialists to administer it.

543
possibly unique to the Theravda tradition. Unfortunately, this both opens up the

possibility that particular texts within the collection are old and closes the possibility

of investigating their development through comparison with other sectarian traditions.

Nevertheless, as we have seen here, the encounter dialogs of the Sutta Nipta clearly

show continuities with the more mainstream early Buddhist oral traditions. In the

case of the two suttas with exact parallels in the Majjhima Nikya, it is likely that we

have cases of direct borrowing, but it is difficult to know in which direction. Since

the Sutta Nipta as a whole is a late, sectarian development, it would not be

unreasonable to suppose that suttas were borrowed from the four main Nikyas to fill

it out. In the case of the Sela Sutta, this in fact seems quite plausible, since the Sela

Sutta is a fairly ordinary encounter dialog in a style characteristic of the mainstream

tradition, and it is uncharacteristic of the Sutta Nipta insofar as it is composed

mostly of prose and includes a relatively small number of verses. The Vseha Sutta,

on the other hand, is composed almost entirely of verses and thus is very

characteristic of Sutta Nipta, but seems, conversely, out of place in the Majjhima

Nikya.

The five suttas that have parallels in the Brhmaa Sayukta/Sayutta

tradition demonstrate even more convincingly that there is no one-to-one

correspondence between the suttas of the Sutta Nipta and the mainstream tradition,

and that therefore the relationship between the two must be understood in terms of

oral tradition. Within the Theravda tradition, we see a mixing of beginnings and

endings between the Kasibhradvja and Sundarikabhradvja Suttas, with the

544
curious account of an undigestible food being thrown into the water and causing it to

sizzle being assigned to one sutta by the Sayutta-bhakas and to the other by the

compilers of the Sutta Nipta. The Sarvstivda tradition, in turn, agrees with

Sayutta version in one case and the Sutta Nipta version in another. In the case of

Aggika Bhradvja, the Theravda tradition preserves two completely different

stories, one each in the Brhmaa Sayutta and in the Sutta Nipta, but the

Sarvstivda tradition preserves only the latter in its own Brhmaa Sayukta.

Similarly, the story of Mgha was preserved by the Sarvstivda tradition in its

Brhmaa Sayukta, but the Theravda tradition preserved it in the Sutta Nipta and

not in the Brhmaa Sayutta.

We can thus summarize by saying that the Sutta Nipta has a complex

relationship with the mainstream early Buddhist tradition. It clearly shared material

in some way with the mainstream tradition, but preserved this material in its own way.

In the particular case of the four encounter dialogs with parallels in the

Sayutta/Sayukta traditions, the Sutta Nipta provides a unique glimpse at what

was clearly a fairly old tradition of constructing encounter dialogs by providing frame

stories to verses that incorporated a Brahman interlocutor, often with a funny/ironic

name and the gotra Bhradvja.

545
 Chapter IV.4

Development of Formulae and Themes in the Encounter

Dialog Tradition

IV.4.1 Introduction

In the last chapter, we examined the encounter dialog genre by comparing and

examining groups of collections of encounter dialogs across sectarian traditions. In

the case of the Dgha/Drgha and Sayutta/Sayutta traditions, this approach was

quite fruitful, since each contains a clearly discernible collection of encounter dialogs

that must be of relative antiquity because they are found in sectarian traditions as far

removed from one another as the Theravda and the Sarvstivda. Once we turned to

the Majjhima/Madhyama tradition, however, this approach began to break down;

even though both the Theravda and the Sarvstivda traditions include a collection

called Brhmaa Vagga/Varga, the two collections are quite different and could

easily have been formed independently of one another. Finally, with the

Aguttara/Ekottarika tradition and the Sutta Nipta, this approach broke down

altogether, since the two extant versions of the Aguttara/Ekottarika tradition lack

much in the way of discernible encounter dialog collections even in isolation, much

546
less in common given their radical difference from one another, and the Sutta Nipta

is a purely Theravda collection with no equivalent in any other extant tradition.

In the process of examining the encounter dialog collections (or lack of them)

in these traditions, however, we saw that there are certain narrative elements,

formulae, and themes that appear to serve as characteristic markers of the encounter

dialog. This is not to say that all encounter dialogs contain all or even any of these

characteristic elements, but rather that they are found often enough throughout the

entire range of encounter dialogs to be considered characteristic of them as a genre.

In addition, because these narrative elements, formulae, and themes appear across the

boundaries of not only sectarian, but also Nikya/gama, traditions, they offer us

another analytic framework through which to examine the development of the

encounter dialog genre. In this chapter, therefore, we will continue our examination

of the encounter dialog genre by tracing the use of a few of the more prominent

formulae and themes in encounter dialogs.

We will begin by looking at the more commonly used formulae characteristic

of the encounter dialog genre, that is, formulae found in encounter dialogs in all four

Nikya/gama traditions. We will begin in Section IV.4.2 with the extremely

common reference to brhmaagm, or Brahman villages, and the related but

slightly less common formulaic description of brahmadeyya land-grants. These shall

be seen as offering a possible clue to the dating of the stras in which they are found.

Then, in sections IV.4.3 and IV.4.4, we will turn to the fame of Gotama and Triple

Veda formulae, respectively, referred to repeatedly in Chapter IV.3. In the context

547
of the latter, we will discuss in particular the peculiar description of Triple Veda

Brahmans as learned in the Lokyata and what this likely meant at the time the

early Buddhist stras were composed.

After addressing these more common formulae, we will turn to some less

common, but nonetheless important, themes that are found in certain corners of the

encounter dialog tradition. In Section IV.4.5, we will look at the theme of the

Brahman looking for the 32 marks of a Great Man on the Buddhas body and discuss

some theories on where this concept of 32 marks came from. Then, in sections

IV.4.6 and IV.4.7 we will look at character themesthat is, themes whereby

certain stock characters (in our case, Brahmans) reappear in more than one stra. In

Section IV.4.6, we will look at the characters Vseha and Bhradvja, who appear in

only three stras, but are nonetheless significant because of the length, content, and

intrinsic interest of the stras in which they appear. Finally, in Section IV.4.7, we

will look at the most frequently reappearing and well-known Brahman stock

characterJussoi, the Brahman who serially asked to become a lay disciple of

the Buddha.

IV.4.2 Brhmaagma and Brahmadeyya

As we saw in Chapter IV.3, the nidna of an encounter dialog often begins by

stating that the Buddha was residing in a Brahman village (brhmaagma). The

stras that mention a brhmaagma are, in fact, not always encounter dialogs, but

for obvious reasons, stras that take place in or near a brhmaagma do in practice

548
often end up involving an encounter between the Buddha and a Brahman interlocutor.

In some cases, the brhmaagma in question is also referred to as a brahmadeyya

that is, a land grant made by a king to a Brahman, which probably in practice meant

that the recipient of the grant received tax revenue from the village in the kings stead.

It is important to note, as Bronkhorst and others before him have, that brhmaagma

and brahmadeyya are not two words for the same thing. The distinction is made clear

in particular in the Ambaha Sutta, in which the Buddha resides in a brhmaagma

named Icchnakala, but his Brahman interlocutor (Pokkharasti) comes from a

nearby place called Ukkhaha, which had been given to him as brahmadeyya. 877

Indeed, in this chapter I will argue that the distinctive use of these two terms may

offer clues to the historical development of the encounter dialog genre.

The term brhmaagma, to begin with, is much more common than the word

brahmadeyya. The former is found in a total of 22 independent instances (mostly

suttas) in the Pali Canon; these can be found listed in Table 14. In many cases a

Chinese equivalent meaning Brahman village is used in a Chinese parallel, and I

have indicated this as well where appropriate. In looking at the chart, it is

immediately obvious that there is a preponderance of references to brhmaagmas

in the Sayutta Nikya. This preponderance is reflected in the Chinese as well.

Several of the Sayutta Nikya suttas that refer to a brhmaagma have a parallel in

the Chinese Sayuktgama that does so as well; in addition, one of the Majjhima

Nikya suttas has a relevant parallel in the Sayuktgama and three Aguttara

877
Bronkhorst, Buddhism in the Shadow of Brahmanism, 92-3.

549
Nikya suttas that refer to the same brhmaagma are parallel to a single

Sayuktgama stra that also refers to it. If we accept Ynshns theory that the

Sayutta/Sayukta is the foundation of the entire Nikya/gama tradition, then this

aggregation of references to brhmaagmas around the Sayutta/Sayukta tradition

could be indicative of their relative antiquity within the broader context of the early

Buddhist oral tradition.

Pali Village Janapada Encounter Chinese

Sutta Name Dialog? Parallel?
DN 3 Icchnagala Kosala Yes D 20
DN 5 Khumata Magadha Yes D 23
DN 13 Manaskaa Kosala Yes D 26
DN 21 Ambasa Magadha No D 14
MN 41 Sl Kosala Yes S 1042-3
MN 60 Sl Kosala Yes
MN 95 Opsda Kosala Yes
SN Ekasl Kosala No S 1097
1.4.2.4 instead refers
to a kya
village
SN Pacasl Magadha No S 1095, E
1.4.2.8 45.4
SN Ekan Magadha/Dakkhigiri Yes S 98
1.7.2.1
SN Sl Kosala No S 621 has a
5.3.1.4 different
nidna
SN Sl Kosala No
5.4.6.1
SN Veudvra Kosala Yes S 1044 refers
5.11.1.7 to Veudvra
(),
but not
specifically as
a Brahman
village
AN 3.63 Vengapura Kosala Yes
AN 5.30 Icchnagala Kosala Yes S 1250 refers

550
 to
Icchnagala
(),
but not
specifically as
a Brahman
village
AN 6.42 Icchnagala Kosala Yes S 1250-1
(see above)
AN 8.86 Icchnagala Kosala Yes S 1250 (see
above)
Ud. 7.9 Tha Malla Yes
Sn. 1.4 Ekan Magadha/Dakkhigiri Yes
Mil. Intro Gajagala Himavant (?) N/A
Mil. 4.2.5 Pacasl ? N/A
Vin. Tha ? N/A
4.5.158
Table 12. A list of all suttas and other texts referring to a brhmaagma in the Pali Canon. Chinese parallels are
discussed as relevant; unless otherwise noted, the Chinese parallel also refers to a Brahman village.

Indeed, Oskar von Hinber has argued on the basis of independent evidence

that nidnas referring to the Buddha residing in a brhmaagma likely represent a

very ancient layer within the early Buddhist tradition. Hinbers argument is based

primarily on the phrasing of the nidna formula in which brhmaagmas (as well as

nigamas, i.e., market places) are referred to, which he regards as primitive. He

argues that the use of Brahman villages and market places as nidnas within the early

Buddhist tradition, which represents a relatively undeveloped stage of urbanization,

came to be replaced by the five great cities (nagara)especially Rjagaha,

Svatth, and Kosambwhich are, due to their prevalence in the tradition as it has

come down to us, much more familiar to readers of the Pali suttas. In addition, he

notes that (as can be seen in Table 14), there is a preponderance of Brahman villages

in Kosala (nine out of the fourteen total), which is in accord with the finding of

551
modern scholars that the Eastern frontier of Vedic culture in late Vedic times was

Kosala, the home of the Kva kh of the atapatha Brhmaa. 878

If the use of brhmaagmas as nidnas represents a fairly early practice in

the tradition that was largely abandoned in favor of large nagara such as Svatth and

Rjagaha, then it will be instructive to observe the intersection (and lack thereof)

between the use of the former and the encounter dialog genre. Several suttas in the

Slakkhandha Vagga refer to brhmaagmas; this is not entirely surprising, since

there is good cross-sectarian evidence that this collection centered on the Tathgata

arises formula goes back fairly far in the tradition. On the other hand, looking more

broadly at the chart, we find that many of the more memorable encounter dialogs we

looked at in Chapter IV.3 are not included in the list, and those that are in the list

often refer to less significant encounter dialogs, or even, in some cases, are not

encounter dialogs at all. Of the 22 encounter dialogs found in the Brhmaa

Sayutta of the Sayutta Nikya, for example, only one refers to a brhmaagma.

Likewise, in the Majjhima Nikya, in which, as we have seen, the most weighty

encounter dialogs (i.e., those that address significant points of tension with new

Brahmanical ideology) are found in the Brhmaa Vagga, only one sutta referring to

a brhmaagma is found in that vagga. Overall, it seems that, with the partial

exception of the Slakkhandha Vagga in the Dgha Nikya, there is not a close

correlation between encounter dialog collections or the encounter dialog genre as a

878
Oskar von Hinber, Hoary Past and Hazy Memory: On the History of Early Buddhist
Texts, Journal of the International Association of Buddhist Studies 29, no. 2 (2006[2008]), 200-2.

552
whole on the one hand, and the (likely old) use of brhmaagma nidnas on the

other.

Let us now turn to the term brahmadeyya. This word is used in a particular

formula, which can be exemplified by its first appearance in the Pali Canon, which is

in the Ambaha Sutta (DN 3):

Now at that time the Brahman Pokkharasti was living in Ukkhahafull of

living beings; with grass, timber, and water; with graina royal property
given by King Pasenadi of Kosala as a royal gift, a brahmadeyya.

This formula, and the word brahmadeyya itself, are found only in six suttas in the

entire Pali Canon. These are listed with relevant information and Chinese parallels in

Table 15. As can be seen from the table, all but one of these six suttas are found in

the Dgha Nikya, and all but one of those that are found there are found, even more

specifically, in the Slakkhandha Vagga. Interestingly, the one sutta that is found

outside the Dgha Nikya, the Cak Sutta, has, in spite of the fact that it is placed by

the Theravdins in the Majjhima Nikya (95), interesting links to the encounter

dialogs of the Slakkhandha Vagga. That is, its formulaic structure is quite similar to

that of the Soadaa and Kadanta Suttas. Like Soadaa and Kadanta, Cak

hears about the Buddhas presence in his domain when he notices that all of the

Brahman householders there, having heard the fame of Gotama formula, are going

to see him. After finding out where they are going, he decides to join them, but this

leads to an argument with other Brahmans, who argue that the Buddha should come

to see him instead of the other way around. Clearly, then, with such a closely related

formulaic structure, there must be a historical relationship between the Cak,

553
Soadaa, and Kadanta Suttas. Indeed, as we saw in Chapter IV.2, the

Sarvstivdin tradition placed the Kmahika Stra, which is parallel to the Cak

Sutta, in its Drghgama rather than its Madhyamgama. This, then, leaves only the

Pysi Sutta as a bit of an outlier in making use of the term brahmadeyya and the

associated formula. 879

Pali Place Recipient Donor Janapada Chinese

Sutta Parallels?
DN 3 Ukkhaha Pokkharasti Pasenadi Kosala D 20
DN 4 Camp Soadaa Bimbisra Magadha D 22
DN 5 Khumata Kadanta Bimbisra Magadha D 23
DN 12 Slavatik Lohicca Pasenadi Kosala D 29
DN 23 Setaby Pysi Pasenadi Kosala D 7, M 71
MN 95 Opsda Cak Pasenadi Kosala none
Table 13. A list of all suttas that make use of the brahmadeyya formula in the Pali Canon, together with the name
of the land-grant, the recipient, the donor, the janapada, and any Chinese parallels. All Chinese parallels also
make use of the formula. In the case of the D parallels, which belong to the Dharmaguptaka school, brahmadeya
is translated as .

What, then, are we to make of the particular correlation between the

brahmadeyya formula and the encounter dialogs of the Slakkhandha/laskandha

collection? Bronkhorst has argued that the term brahmadeyya, in contrast to the term

brhmaagma, may have entered the Buddhist oral tradition relatively late. 880 Such

a supposition is indeed advantageous to his hypothesis that Buddhism originally arose,

not in opposition to Brahmanism, but simply in a culture (i.e., that of Greater

879
The use of this formula in the Pysi Sutta is, in fact, particularly strange because the
interlocutor, Pysi, is never identified as a Brahman, but rather as a rjaa, and concomitantly the
sutta cannot even, presumably, be considered an encounter dialog. Tsuchida argues, however, that the
term rjaa probably refers in the Pali texts to a type of royal servant, who presumably could have
been a BrahmanTsuchida, Two Categories of Brahmans, 91 n. 16.
880
Bronkhorst, Buddhism in the Shadow of Brahmanism, 92.

554
Magadha) that was outside the sphere of Brahmanical culture at that time. 881

Encounter dialogs such as the ones listed in Table 15 are problematic for

Bronkhorsts thesis insofar as many of them, especially the three we examined in

detail in Section IV.3.3 above (DN 3-5), involve the Buddha engaged in rather

pointed polemic against the tenets of the new Brahmanism. Still, it is not

unreasonable to posit that the brahmadeyya formula entered the tradition late, given

that it is found clustered in a small group of closely-related texts, and not, like

brhmaagma, found diffused across the encounter dialog genre and beyond. On

the other hand, several suttas of the Slakkhandha Vagga, as can be seen from Table

13, also make reference to a brhmaagma, including two of the suttas listed in

Table 14 that make use of the brahmadeyya formula. Moreover, we have already

cited several times the evidence for the relative antiquity of the

Slakkhandha/laskandha collection based on the fact that it is found in all three

sectarian traditions of the Dgha Nikya/Drghgama. How then are we to interpret

the implications of the use of the brahmadeyya formula for the relative age of a stra

in the development of the early Buddhist tradition?

The most obvious way to go about answering this question is to look at the

evidence for the actual practice of kings giving brahmadeyya land-grants on the

ground. Unfortunately, in attempting to answer this question, we face a bit of an

impasse. As Tsuchida writes,

881
Ibid., 1: [The Brahmans] did not occupy a dominant position in the area in which the
Buddha preached his message, and this message was not, therefore, a reaction against brahmanical
thought and culture.

555
 Because of epigraphical and other textual evidence we are better informed
about the brahmadeya and similar institutions in later periods. In the period
with which we are here concerned, we lack sufficient historical materials to
investigate the exact legal status and actual conditions of lands granted to or
inhabited by Brahmins. 882

This, however, is perhaps an overly cautious way of saying that we do not have

strong evidence that brahmadeyya was practiced at all in the period that these stras

claim that it wasthat is, in the Buddhas lifetime, over one hundred years, at the

very least, before the ascension of Aoka. Bronkhorst provides a useful summary of

the evidence for land grants to Brahmans, and this evidence cannot be said to

demonstrate any great antiquity for the practice, at least within the time parameters

we are interested in. 883 Vedic texts as late as the atapatha Brhmaa (13.7.1.15)

and the Aitareya Brhmaa (39.8) refer to the giving away of land in negative

terms. 884 A story in the Chndogya Upaniad (4.2) refers to a donor giving a

Brahman a village, 885 but it is in the Dharma Stras that we finally find explicit calls

for kings to give gifts of land to Brahmans. 886 The use of the term brahmadeya to

refer to land grants, however, isaside of course from the Buddhist encounter

882
Tsuchida, Two Categories of Brahmans, 57.
883
Actually, Bronkhorst uses the evidence he presents to make an argument in the opposite
direction, that [o]ur texts confirm that land grants were known from an early time onward
Bronkhorst, Buddhism in the Shadow of Brahmanism, 86. However, Bronkhorst is referring to a much
later time referencenamely, the appearance of the word agrahra in the medieval period. For our
purposes, early must refer to a time prior to the time of the Buddha, not just before the medieval
period.
884
Cited in ibid.
885
Cited in ibid.
886
Ibid., 87.

556
dialogs we are looking at herenot, as far as I am aware, found until the

Arthastra, 887 which, although ostensibly written by Candragupta Mauryas purohita

Cakya (identified with Kauilya), is likely a composite work dating in large part to

the common era. 888 Finally, hard evidence that land grants were actually being made

to Brahmans first enters the epigraphical record in the first century BCE, 889 well after

the time Aoka, and even longer after the time of the Buddha. Indeed, the entire

trajectory of the epigraphical evidence in Indiawhich explodes into history, starting

with Aoka, with references to support for non-Brahmanical sects, and only slowly

over the course of the next several centuries shows evidence of a shift toward

Brahmanical groupssuggests that an institute such as that of brahmadeyya

described in the Buddhist stras under consideration here is more likely to be later

than it is to be earlier.

In the end, although the evidence is not entirely conclusive, I believe that the

reference to brahmadeyya land-grants in several encounter dialogs of the

Slakkhandha/laskandha collection should make us suspicious of the antiquity of

those stras, and thus, due to the close correlation between the two, of the collection

as a whole. Indeed, one of the stras in question makes use of the brahmadeyya

formula in a way that seems particularly likely to be anachronistic. According to the

Soadaa Sutta, the Brahman Soadaa resided in Camp, which was given to him

as brahmadeyya by Seniya Bimbisra of Magadha. Camp, however, is not just an

887
Ibid., 77.
888
Ibid., 67-73.
889
Ibid., 88-9.

557
insignificant village in Magadhait is, in fact, neither inisignificant nor in Magadha,

but is listed as one of the five great cities (nagara) and is the capital of an entirely

different janapada, namely Aga, which is to the east of Magadha in modern-day

Bengal. It is not clear what authority Bimbisra would have had to grant authority

over a city in another janapada, much less its capital, and one cannot help but wonder

if the statement that he did so is less an expression of the historical reality in the time

of the Buddha than of a later, more imperialistic age.

IV.4.3 The Fame of Gotama Formula

One of the most characteristic formulae of the encounter dialog genre, which I

have had occasion to mention repeatedly in Chapter IV.3, is the fame of Gotama.

This formula, which describes the fame (kittisadda) of the samaa Gotama, is

frequently used as a narrative device to introduce the Buddhas presence in the area to

an interlocutor and, by piquing the interlocutors interest, induce him to pay the

Buddha a visit. In Pali, the formula, in its fullest form, is as follows:

Now, regarding the Venerable Gotama a good reputation has gone forth thus:
That Blessed One is worthy, fully Awakened, endowed with knowledge and
conduct, well-gone, knower of the world, the unexcelled charioteer of people
who are to be trained, instructor of gods and men, the Awakened, the Blessed
One. Having realized for himself by higher knowledge, he declares this world
with the gods, with the Mras, with the Brahms; [he declares] the people
with the samaas and Brahmans, with the gods and men. He preaches the
dharma, which is good in the beginning, good in the middle, good in the end,
with meaning and articulation. He proclaims a wholly perfect, completely
pure brahmacariya. Well indeed is it to see such worthies. 890
890
See, e.g., DN 3: ta kho pana bhavanta gotama eva kalyo kittisaddo
abbhuggato iti pi so bhagav araha sammsambuddho vijjcaraasampanno sugato lokavid
anuttaro purisadammasrathi satth devamanussna buddho bhagav. so ima loka sadevaka
samraka sabrahmaka sassamaabrhmai paja sadevamanussa saya abhi sacchikatv

558
Sometimes, however, a shorter form of this formula, ending with the Awakened, the

Blessed One (buddho bhagav) is found. A list of all 891 instances of this formula, in

both the short and the long form, can be found in Table 16.

Pali Sutta or Other Text Full or Short Version? Encounter Dialog?

Vin. 1, Verajakaa Full N/A
Vin. 4.1.13 Full N/A
Vin. 4.6.180 Full N/A
Vin. 4.6.182 Full N/A
DN 2 Short No
DN 3 Full Yes
DN 4 Full Yes
DN 5 Full Yes
DN 6 Full Yes
DN 12 Full Yes
DN 13 Short Yes
MN 41 Full Yes
MN 42 Full Yes
MN 60 Full Yes
MN 75 Short Yes
MN 82 Full Yes
MN 91 Full Yes
MN 92 Full Yes
MN 95 Full Yes
MN 98 Short Yes
MN 140 Short No
MN 150 Full Yes
SN 5.11.1.7 Full Yes
AN 3.63 Full Yes
AN 3.65 Full No

pavedeti. so dhamma deseti dikalya majjhekalya pariyosnakalya, sttha

sabyajana. kevalaparipua parisuddha brahmacariya pakseti. sdhu kho pana
tathrpna arahata dassana hot ti.
891
Parts of this formula, in particular the list of epithets iti pi so buddho bhagav that
begins the description of the Buddha in the full versionwhich, incidently, is commonly memorized
and used as a mantra/prayer in Theravda Buddhist countries such as Thailandare found in a wider
variety of texts than I have listed in Table 15. I have only included, as instances of what I call the
fame of Gotama formula, those cases that begin with the introduction ta kho pana bhavanta
gotama eva kalyo kittisaddo abbhuggato, since only these cases make use of the description of
the Buddha that follows as a report of the reputation or fame of the samaa Gotama.

559
AN 5.30 Full Yes
AN 6.42 Full Yes
AN 8.86 Full Yes
Sn. 3.7 (=MN 92) Full Yes
Sn. 3.9 (=MN 98) Short Yes
Table 14. A list of all instances of the full or short form of the fame of Gotama formula in the Pali Canon.

Looking at this chart, we see some interesting patterns emerge. First, and

most obviously, almost all of the suttas that use the fame of Gotama formula are

encounter dialogs. Indeed, only three out of the total of 26 suttas (I am not including

the Vinaya passages here) are not. This observation may seem trivial at first, but

keep in mind that there is nothing particular about the wording of the fame of

Gotama formula that, hypothetically speaking, should restrict its use to encounter

dialogs. Encounter dialogs, as I define them, are characterized specifically by

Brahman interlocutors meeting with the Buddha, but the Buddha meets with many

people in the stras of the early Buddhist tradition, and there is no reason, again

speaking hypothetically, why non-Brahman interlocutors cannot become aware of the

Buddhas presence and fame, and thus decide to visit him, by means of the fame of

Gotama formula, just as Brahman interlocutors do in encounter dialogs. That is,

there is nothing in the content of the formula that would make its use inappropriate in

a context where the person learning of Gotamas fame is not a Brahman. Indeed, the

fact that there are three exceptions, i.e., three suttas using the formula that are not

encounter dialogs, proves this. Nevertheless, the vast majority of suttas making use

of the formula are encounter dialogs, which shows that the early Buddhist tradition,

for whatever reason, saw the formula primarily as one to be used in constructing

stras of that particular genre.

560
 Another pattern that can be seen from the chart is the distribution of suttas that

use the formula among the Nikyas. In particular, what stands out is that only one

sutta from the Sayutta Nikya makes use of the fame of Gotama formula. Not

only this, but the one Sayutta sutta that does use it is not part of the primary

encounter dialog collection in the Sautta Nikya, namely, the Brhmaa Sayutta.

This would indicate that the formula did not have its origins in the

Sayutta/Sayukta tradition, or that if it did, nearly all of those stras that used it

were pulled out to create or use in other Nikya/gamas. This is in clear contrast to

both the Dgha and the Majjhima Nikya, where the formula is used quite frequently.

In particular, the formula is used in every single 892 encounter dialog of the

Slakkhandha Vagga. Since we already know that the Slakkhandha/laskandha

collection is of a certain relative antiquity in the tradition, given the fact that it is

found in all three extant versions of the Dgha/Drgha, we can easily be led to suspect

that the fame of Gotama formula had its origins in this collection.

Although it is difficult to prove this hypothesis, there are some good reasons

for accepting it, at least tentatively. As we saw in Chapter IV.3, there are only two

collections of encounter dialogs that are of demonstrable antiquity in the early

Buddhist traditionthe Brhmaa Sayutta/Saukta and this

Slakkhandha/laskandha collection. The former does not make use of the fame of

Gotama formula at all, while the latter makes use of it in every single encounter

892
As in Chapter IV.3, I am not counting the Subha Sutta (DN 10) since its interlocutor is
identified not as a Brahman, but as a mava, and the interlocutor does not meet the Buddha, but
nanda.

561
dialog. Uses of the formula elsewhere are scattered. The lack of cohesion to the

Brhmaa Vagga in the Majjhima Nikya, which I argued for in Chapter IV.3, is

supported by the fact that, while several suttas in the vagga (91, 92, 95, 98) do make

use of the formula, the majority do not, and there is no discernible pattern as to which

do and which dont. 893 At the same time, the formula is used frequently in encounter

dialogs in the Majjhima Nikya outside of this vagga. Thus, it is possible that the

fame of Gotama formula was created for use in encounter dialogs in the

Slakkhandha/laskandha collection, and due to the popularity of this collection, it

came to be used, albeit with less consistency, in other encounter dialogs elsewhere.

Whether this particular hypothesis is correct or not, however, it is almost certain that

the formula was created specifically for use with encounter dialogs, and thus became

a characteristic marker of that genre.

IV.4.4 The Triple Veda Formula

Another formula characteristic of the encounter dialog genre that we have

encountered repeatedly already is the Triple Veda formula. This formula is

generally used to introduce a Brahman interlocutor, and it does so by asserting, in

some detail, that the Brahman in question is a master of the Triple Veda, along with a

list of certain ancillary sciences that, as we shall see, vary somewhat from one version

of the formula to another. In the Theravda tradition, the formula is mostly fixed, and
893
One slightly ancillary pattern is worth mentioning here, however. Both DN 13 and MN 98,
in which the Brahman interlocutors are Vaseha and Bhradvja, make use of the fame of Gotama,
and moreover make use of it in the relatively rare short version. This provides extra justification for
looking at the suttas that involve Vseha and Bhradvja together as a group, as we will do in Section
IV.4.6.

562
although there is some slight variation in the way the formula is incorporated into its

broader context in a particular narrative, the following kernel is found word-for-word

in all instances:

[The Brahman in question is] perfected in the three Vedastogether with

their vocabularies and rituals, with their phonology and etymology, and the
oral tradition (itihsa) as a fifthskilled in philology and grammar, not
lacking in the Lokyata and marks of a Great Man. 894

This formula is found in a total of 11 suttas 895 in the Pali Canon; these can be found

listed in Table 17.

Pali Sutta Context Chinese Parallel

DN 3: Ambaha Introduction to mava D 20: Dharmaguptaka
Ambaha version, used to introduce
Ambaha
DN 4: Soadaa Brahmans argument for D 22: Dharmaguptaka
why Soadaa is too version, used to introduce
good to go see Gotama Soadaa
DN 5: Kadanta Brahmans argument for D 23: Dharmaguptaka
why Kadanta is too good version, used to introduce
to go see Gotama Kadanta
MN 91: Brahmyu Introduction to Brahman M 161: Sarvstivda
Brahmyu version, used to introduce
Brahmyu
MN 92: Sela Introduction to Brahman E 49.6: Formula, if
Sela present at all, appears to be
abbreviated as He was
very learned ()
MN 93: Assalyana Introduction to mava M 151: Sarvstivda
Assalyana version, used to introduce
Assalyana
MN 95: Cak Brahmans argument for None, but there is a

894
tia vedna prag sanighaukeubhna skkharappabhedna
itihsapacamna, padako, veyykarao, lokyatamahpurisalakkhaesu anavayo.
895
Not counting Sn. 3.7, which is the same as MN 92, the Sela Sutta. The formula is also
found in some later texts, such as the Niddesa and the Milinda Paha, which I have not listed here.

563
 why Cak is too good to Sanskrit fragment
go see Gotama containing the formula,
close to the Sarvstivda
version
MN 100: Sagrava Introduction to mava None
Sagrava
AN 3.58: Tikaa Explanation of tevijja Partial parallel to S 884-5,
which does not include
formula
AN 3.59: Jussoi Explanation of tevijja Partial parallel to S 884-5,
which does not include
formula
AN 5.192: Doa Explanation of tevijja M 158: Sarvstivda
version, used by Doa to
explain traividya ()
Sn. 3.7: Sela Introduction to Brahman See MN 92
Sela
Table 15. List of all instances of the Triple Veda formula in Pali suttas, together with description of context
and Chinese parallels. Later Pali texts such as Niddesa and Milinda Paha are not included.

Looking at this chart, we immediately see a pattern similar to that of the

fame of Gotama formula: This formula can be found everywhere except the

Sayutta Nikya. Outside of the Sayutta Nikya, however, it is found scattered

across the various Nikya traditions, with many examples being found in the three

core encounter dialogs of the Slakkhandha Vagga in the Dgha Nikya and several

encounter dialogs of the Brhmaa Vagga in the Majjhima Nikya. This would

appear to suggest that, like the fame of Gotama formula, the Triple Veda formula

had its origins outside of the Sayutta/Sayukta tradition. Possible corroborating

evidence for this is provided by the Tikaa and Jussoi Suttas, which are

adjoining suttas in the Aguttara Nikya that include the formula. The only known

parallels to these suttas, however, are two adjoining stras in the Chinese

Sayuktgama, and these stras do not include the formula. In fact, in the Chinese

564
version, these stras are not encounter dialogs at all; whereas in the Pali versions, the

Buddha is prompted to give his own interpretation of tevijja only after a Brahman

interlocutor uses the Triple Veda formula to give a Brahmanical explanation, in

the Chinese version, the Buddha simply defines traividya () on his own, in a

brief sermons to his monks. It is very well possible, then, that the Pali versions of

these two suttas, in addition to being moved from the Sayutta to the Aguttara

collection, were embellished to convert them from simple sermons into encounter

dialogs, and in the process the Triple Veda formula was added.

The actual content of the Triple Veda formula is mostly straightforward.

The first and most important point made is that the Brahman in question (whether

actual or hypothetical) is a master of the Triple Veda. The reference here to only

three Vedas, instead of four, is standard in the Pali Canon and is supported by the

Dharmaguptaka version of the formula, which also refers to the three parts of the old

classics (). This version of the formula clearly dates to a time before the

Atharva was fully accepted as a fourth Veda. 896 The Sarvstivda version, however,

refers to four Vedasor rather, in Chinese, to four classics ()and thus

presumably, at least in the form it has come down to us, dates to a later time when the

896
The fact that early Indian texts, including the Pali Canon, refer to three rather than four
Vedas makes clear that the designation of the Atharva, which was originally known simply as the
Atharvgrasa, as a Veda is late. There is evidence that there was contestation over the Atharvas
Vedic status: Witzel (Development of the Vedic Canon, 278-9) has argued that the compilers of the
Atharva used archaicisms to bolster their collections claim to Vedic status, and Ronald Inden,
Changes in the Vedic Priesthood, in A. W. van den Hoek et al., eds., Ritual, state and history in
South Asia: Essays in honour of J.C. Heesterman, 556-77 (Leiden: Brill, 1992), argues that the
brahman priest was originally considered a master of the Triple Veda, and only came to be associated
with the Atharva when the latter became accepted as a fourth Veda.

565
Atharva had been fully accepted as the fourth Veda. We can, in any case, be

confident that the latter form is not original and rather represents an updating of the

formula in the latter-day oral tradition. Not only is the Pali Canon absolutely

unanimous in speaking of three Vedas, but, as we saw in Chapter III.3, playing off of

the concept of precisely three Vedas, or a threefold knowledge of the Brahmans,

was used to construct the idea of a threefold knowledgei.e., knowledge of past lives,

knowledge of the workings of karma, and knowledge of the destruction of the

savasthat is found frequently in the early Buddhist suttas. This theme must

clearly have been developed and diffused throughout the early Buddhist tradition

prior to the elevation of the Atharva to the status of fourth Veda.

The Pali version of the formula adds that not only is the Brahman in question

an expert in the three Vedas per se; he is also learned in a set of five auxiliary

sciences that are associated with the three Vedas. 897 It should be emphasized that

these five auxiliary sciences, though partially overlapping, at least in terms of general

meaning, do not correspond in any way to the traditional vedgas of the classical

Brahmanical tradition. The vedgas are six in number, and their namesik,

kalpa, vykaraa, nirukta, chandas, and jyotiado not correspond to any of the

names mentioned in the list of five sciences in the Pali triple Veda formula. 898 The

897
The association of these five sciences with the three Vedas is indicated by placing them in
compounds that begin with sa- (with) and end with genitive plural endings, in agreement with the
phrase tia vedna, thus indicating that they are attributes of the three Vedas, of which the
Brahman has gone to the far shore (prag), which I have translated more colloquially as perfected
in.
898
It is true that veyykaraa is mentioned in the formula, but it is not listed as part of the list
of sciencesthat is, as part of a compound that begins with sa- and agrees in case with the three

566
listing of these five auxiliary sciences is supported by the Sarvstivda version in

Chinese, which provides a translation of each of the five sciences and enumerates

them, as in the Pali, as being five in number: . It is

not clear, however, if these five sciences were listed in the Dharmaguptaka version.

In the Chinese translation of the Drghgama, it is simply said that the Brahman in

question is able to analyze the classics in every way (). It is

possible, however, that the original Indic version of the formula the translator was

working from did list the five sciences, but that the translator decided not to give an

exact translation because he either did not know how to translate the terms involved

or did not think they would be meaningful to his Chinese audience.

The most interesting part of the Triple Veda formula, at least in the Pali

version, is the very final phrase, not lacking in the Lokyata and marks of a Great

Man (lokyatamahpurisalakkhaesu anavayo). This phrase has instigated a

considerable amount of controversy, in particular over the meaning of the word

lokyata. It should first be mentioned, however, that not all sectarian versions of the

formula even include this phrase. In the Sarvstivda version, found in several stras

in the Chinese translation of the Madhyamgama, 899 the formula simply ends with the

Vedas. Instead, veyykaraa agrees in case with the Brahman himself, indicating that the latter, in
addition to being versed in the Vedas with its associated five-fold science, is versed in grammar. No
direct connection is drawn between the Vedas and grammar as is between the former and the five-fold
science. The six agas are, however, mentioned (though not by name) in a version of the fomula
found in the Mahvastu (vol. 1, p. 231). This version of the formula is most likely late and is also
corrupt, as it also lists the five auxiliary sciences found in the Pali version, but out of order, with oral
tradition as a fifth listed, inexplicably, in the middle of the other four: skaraprabhednam
itihsapacamn sanirgahakaihabhn.
899
M 63, 151, 158, 160, 161.

567
listing of the five auxiliary sciences. No mention is made thereafter of the marks of a

Great Man or anything else. Although it is possible that the Chinese translator simply

left this part of the formula out, this is unlikely since a version can be found in a

Sanskrit version of the Cak Sutta that also omits mention of the marks of a Great

Man or the Lokyata. In this version, after the five auxiliary sciences are mentioned,

the Brahman is called a padako (also found in the Pali, where I translated it as

skilled in philology) and said to be vykarae anapayya (not lacking in

grammar). 900 This is nearly the same as the ending of the Pali version of the formula,

except that the compound lokyatamahpurisalakkhaesu is omitted. On the other

hand, the Theravda tradition must not have been alone in including a reference at the

very least to the marks of a Great Man. According to the Dharmaguptaka version, the

Brahman is also capable with respect to the marks of a Great Man, the practice of

divination, and the sacrificial rituals. 901 It is not clear what exactly the original Indic

version said here, but it appears to have agreed with the Theravda version at least in

referring to the mahpurisalakkhani.

This leaves, in the Pali, the word lokyata, which appears in compound with

mahpurisalakkhaa, but must be taken separately as something in which the

Brahman is not lacking (anavayo). This word has long vexed scholars because, in

900
Jens-Uwe Hartmann, More Fragments of the Cagstra, in Buddhist Manuscripts, vol. 2,
edited by Jens Braarvig (Oslo: Hermes Publishers, 2002), 10.
901
. This is the version found in D 22, 23, and
29, which correspond to the Soadaa, Kadanta, and Lohicca Suttas, respectively. In D 20,
which corresponds to the Ambaha Sutta, the middle element of the list, , which refers to the
practice of divination, is omitted.

568
classical usage, lokyata is a synonym of crvka and refers to materialism, a

heterodox philosophical school that would be among the last things we would

expectagain, seen through the lens of classical categoriesto see in a eulogy of an

orthodox Brahman. Rhys Davids, in his translation of the Dgha Nikya, was one of

the first to comment on this problem, devoting an extended portion of his introduction

to the Kadanta Sutta to a discussion of this single word. 902 By comparing the Pali

commentarial glosses on lokyata to later uses of the word in Sanskrit literature, Rhys

Davids comes to the conclusion that around 500 BCE, lokyata meant something like

Nature-lorewise sayings, riddles, rhymes, and theories handed down by

tradition as to cosmogony, the elements, the stars, the weather, scraps of
astronomy, of elementary physics, even of anatomy, and knowledge of the
nature of precious stones, and of birds and beasts and plants. 903

Afterwards, Rhys Davids argues, lokyata came to be used as a term of abuse, and as

a result, it became a sort of catch-all category for whatever was viewed as heterodox

by the wielder of the term.

Although Rhys Davids interpretation does not provide much of an

explanation of how the term lokyata transformed from a generic or even positive

term to a term of abuse, it does help explain why the term may have been included in

the Triple Veda formula, in particular in conjunction with the marks of a Great Man.

As we will see in Section IV.4.5, it is likely that the Brahmans supposed ability to

read the 32 marks of a Great Man were related to the art of physiognomy, a branch

of divination that deals with the significance of various bodily characteristics. If

902
T. W. Rhys Davids, Dialogues of the Buddha, vol. 1 (London: Oxford University Press,
1899), 166-72.
903
Ibid., 171.

569
Rhys Davids is right in interpreting lokyata as nature-lore, then it may be the case

that what is being referred to by the word is the very practical sorts of knowledge,

including the ability to read signs, that would have allowed a Brahman to recognize

that the Buddha was indeed a Great Man (mahpurisa). Indeed, as we have already

seen, the Chinese translation of the Dharmaguptaka version of the formula says that

the Brahman in question has ability with regard to the marks of a Great Man, the

practice of divination, and sacrificial ritual. Unfortunately, without access to an Indic

original, it is impossible to know what exactly is being translated here, but the fact

that the translator lists divination in conjunction with the marks of a Great Man here

does seem suggestive.

More recent scholars, however, have been critical of Rhys Davids

interpretation of lokyata. Although it is beyond the scope of our discussion to

review the history of scholarship on the word lokyata here, a useful survey can be

found in an article about the Lokyata school by Eli Franco. 904 The most obvious

avenue for criticizing Rhys Davids interpretation is that there is no basis for

translating the word loka as nature, a fact that was pointed out by Giuseppe Tucci

in the 1920s. 905 Perhaps the most satisfying explanation of the use of the word

lokyata in the early Buddhist Triple Veda formula, as indicated by Franco, 906 was

904
Eli Franco, Lokyata, in Brill Encyclopedia of Hinduism, ed. Knut Jacobsen et al., vol. 3
(Leiden: Brill, 2011), 629-42.
905
Ibid., 631.
906
Ibid., 632: Perhaps the most important work on early Lokyata was accomplished by J.N.
Jayatilleke in his Early Buddhist Theory of Knowledge (1963).

570
given in the 1960s by K. N. Jayatilleke in his Early Buddhist Theory of Knowledge.

Jayatilleke notes in surprise that no scholar before him had drawn attention to two

additional references to the word lokyata in the Pali Canon outside of the context of

the Triple Veda formula in which it is usually discussed. The first of these is SN

2.1.5.8, in which a so-called lokyatiko brhmao approaches the Buddha and asks

him about the validity of four statements: All exists, All does not exist, All is

unity, and All is plurality. In response, the Buddha labels all four of these as

extreme views. The second Pali sutta that Jayatilleke cites is AN 9.38, in which two

lokyatik brhma approach the Buddha and ask him who is right: Praa Kassapa,

who asserts that the world is finite, or Nigaha Ntaputta, who asserts that the world

is infinite. Again, the Buddha responds by rejecting such questions as futile. 907

To these two suttas, Jayatilleke adds the testimony of a much later Buddhist

text, the Lakvatra Stra, in which the Buddha also encounters a lokyatika

Brahman. This text goes far beyond the two Pali texts just cited, mentioning a total of

32 Lokyata doctrines. Out of these 32, 24 have to do with either all (sarvam) or

the world (loko). In addition, all but three come in pairs of opposites, such as, for

example, the first two: All is created (sarva ktakam) and All is not created

(sarvam aktakam). 908 Jayatilleke comes to two conclusions based on this evidence.

First, the word loka in lokyata must mean simply world or cosmos, 909 which is,

907
K. N. Jayatilleke, Early Buddhist Theory of Knowledge (London: George Allen and Unwin
Ltd., 1963), 49-51.
908
Ibid., 51-3.
909
Ibid., 55.

571
of course, its ordinary meaning in Pali and Sanskrit. Second, the doctrines of the

Lokyataas attested by the SN sutta, the AN sutta, and the Lakvatra Stra

usually come in pairs of opposites, and thus appear to take the form of theses and

antitheses. Jayatilleke notes,

This dialectical opposition in these pairs of views reminds us of the Vedic

institution of the brahmodya, which found expression in the form of a
vkovkya, which was originally a dialogue and later a debate (dialectics) in
which one tried to outstrip the other by arguments, designed to disprove ones
opponents thesis and prove his own. 910

If this interpretation is correct, Jayatilleke adds,

then the later use of the term Lokyata to denote exclusively the Materialist
doctrines is a one-sided application and development of a term, which had a
wider coverage earlier, denoting as we see not only the Materialist doctrines
but their anti-theses, the Eternalist doctrines as well. 911

Finally, Jayatilleke explains the transformation of the term lokyata from denoting

the science of dialectics to denoting a particular school of thought that had been

cultivated therein by appealing to the need of brahman orthodoxyor what we are

referring to here, following Bronkhorst, as the new Brahmanismto defend the

authority of the Vedas against the free exercise of reason and speculation. As a

result, Lokyata qua dialectics became taboo within the Brahmanical community and

came to be used as a slur against those who continued to hold dialectically-nurtured

viewpoints that were, as Jayatilleke puts it, opposed to Vedic teachings. 912

910
Ibid., 50-1.
911
Ibid., 51.
912
Ibid., 56-7.

572
 Although I find Jayatillekes conclusions on the whole convincing, it should

not go without mention that his characterization of Lokyata qua classical

Materialism as opposed to Vedic teachings introduces the assumption that Vedic

teaching is an unproblematic category. Indeed, by arguing that the classical usage of

the word lokyata grew out of an earlier usage in which it referred to the practice of

dialectics that grew out of the brahmodya, Jayatilleke is implicitly conceding that

all dialectical positions that are mentioned in, e.g., the early Buddhist texts, had their

roots in a Brahmanical/Vedic context. In his Greater Magadha, Bronkhorst has in

fact argued that the classical school of the Lokyatas or Crvkas was not anti-

Vedic, but rather the Brahmanical reaction, still in classical times, against the new

doctrine of rebirth and karmic retribution that was slowly but certainly gaining

ground, and in fact held positions very close to those of the (Prva-)Mmsa. 913

When viewed from this perspective, it is difficult to accept Jayatillekes hypothesis

913
Bronkhorst, Greater Magadha, 158. It should be noted that it is possible to accept
Bronkhorsts narrow argument here regarding the Crvkas, which indeed I find convincing, without
accepting his broader argument in Greater Magadha that the concepts of karma and rebirth were
foreign to the Brahmanical tradition and were borrowed from the religion of Greater Magadha. As I
argued in the Introduction, Bronkhorsts argument is based on a problematic dichotomy between
Brahmanical and non-Brahmanical in ancient India. Indeed, while the early Buddhist tradition
clearly associates the term lokyata with interlocutors who are identified as Brahmans, the actual
thesis-antithesis pairs that make up the content of lokyata are most commonly discussed in the early
Buddhist tradition in the context of teachings about wrong views, in which the holders of such views
are most commonly not identified as Brahmans. Indeed, even one of the suttas involving lokyatika
Brahmans cited by Jayatilleke, AN 9.38, has the Brahmans in question attribute lokyata theses to
Praa Kassapa and Nigaha Ntaputta, who are not usually identified as Brahmansmost likely
because they were not proponents of new Brahmanical ideology. Whatever lokyata originally
referred to, it clearly was something in which proponents of the new Brahmanismi.e., the antagonists
of encounter dialogswere reputed to be skilled in, but that was practiced outside of their narrow
ideological community as well. In my view, it is much easier to understand the available data not on
the basis of a model of an original dichotomy between a western Brahmanical tradition (which is then
to be identified with the original Lokyatas/Crvkas) and an eastern non-Brahmanical tradition, but
on the basis of a model in which there is a pan-North Indian culture that included a dialectical tradition
known as lokyata, out of which a Brahmanical tradition arose as proponents of new Brahmanical
ideology fashioned an identity for themselves in opposition to others.

573
that lokyata qua dialectics became taboophilosophical debate within the

Brahmanical tradition continued well into the classical period and continues even to

this daymuch less that it then came to be used as a slur against Materialism because

it was opposed to Vedic teachings. At most, we might speculate that it came to be

used to refer to Materialists because in the eyes of their opponents they were

opposed to Vedic teachings. It is not entirely clear, however, why the word

lokyata was chosen for this purpose.

Before moving on, it is worth asking what, if anything, the use of the word

lokyata in the Triple Veda formula can tell us about the antiquity of the formula.

Unfortunately, given the fact that the formula itself appears to be the earliest

attestation of the term in Indian literature, the answer is very little. Nevertheless, we

can tentatively conclude from our cross-sectarian study of the use of the term within

the earliest Buddhist tradition that it is not of great antiquity within that tradition. As

already noted, the term does not appear at all in the Sarvstivda version of Triple

Veda formula, nor in versions of the formula found in the Mahsghika

Mahvastu and Sanskrit fragments of the Cag Stra. It may appear in the

Dharmaguptaka version, if we accept that the reference to the practice of divination

is a translation into Chinese of the term, but otherwise it seems to be isolated, within

the context of the Triple Veda formula, to the Theravda tradition. In addition,

both of the suttas cited by Jayatilleke in which lokyatika Brahmans appear as

interlocutors lack any known parallels in other sectarian traditions. This, combined

with the fact that the other major Buddhist reference to lokyatika Brahmans is found

574
in the relatively late Lakvatra Stra, suggests that the term lokyata may not be of

great antiquity in the early Buddhist traditionwhich is not to say that the Triple

Veda formula lacks antiquity, but that the word lokyata may not have been original

to it.

IV.4.5 The Theme of the Thirty-Two Marks of a Great Man

Although it features prominently in only a handful of stras, the theme of the

Brahman seeking the thirty-two marks of a Great Man (Skt. mahpurua, P.

mahpurisa) on the body of the Buddha is easily one of the most important, and

certainly one of the most interesting, themes associated with the encounter dialog

genre. I make use of the term theme instead of formula here because the

narrative trope in which a Brahman hears of the Buddha and decides to see if he

really has the thirty-two marks is not conveyed in a single formula, but rather through

a set of formulae that together can be said to comprise a theme that governs a

stras entire narrative structure. Within the Theravda tradition, a fair number of

suttas make reference to the mahpurisalakkhani, but only three make use of the

theme that we are interested in in this section. Of those that only make reference to

the 32 marks, most do so in the form of the Triple Veda formula examined above in

Section IV.4.4. As we saw there, this formula makes reference to a Brahmans

knowledge of the 32 marks in the Theravda and Dharmaguptaka traditions, but not

in the Sarvstivda tradition, and most likely not in the Mahsghika tradition either,

given its absence in a version of the formula used in the Mahvastu. The 32 marks

575
are also, incidentally, not mentioned at all in the fame of Gotama formula. This is

significant because, as we will see, the fame of Gotama formula is used within

suttas that employ the 32-marks theme to make the Brahman who decides to go look

for the 32 marks on the Buddhas body aware of the Buddhas presence in the area,

and thus would be an ideal place to make reference to the Buddhas possession of the

marks as part of his renown. These two pieces of evidencethe total absence of

reference to the 32 marks in the fame of Gotama formula and the likelihood that it

was added to the Triple Veda formula by the Vibhajyavda (i.e., antecedent to the

Theravda and Dharmaguptaka) traditionsuggest that the entire idea of the Buddha

possessing 32 marks, and thus the theme we are about to investigate in this section,

are, within the context of the early Buddhist tradition, younger than either the fame

of Gotama or the Triple Veda formula.

Two additional suttas in the Pali Canon, while not making use of the full

theme of a Brahman looking for the 32 marks, do make use of a particular formula

used in that theme in order to make reference to the fact that the Buddha (and

Buddhas in general) have the 32 marks. This formula, which I will refer to as the

two paths of a Great Man formula, explains in detail that a person possessing the 32

marks will either, if he remains a householder, become a cakkavatti (Skt. cakravartin)

monarch, or else, if he goes forth into homelessness, become a Buddha. This formula

is placed in the lips of fortune-telling Brahmans (nemitte brhmae) in the

Mahpadna Sutta (DN 14) who are called to prognosticate on the prince Vipass

before he becomes a Buddha, i.e., one of the Buddhas of past eras that are described

576
by the current Buddha (Sakyamuni) in this sutta. The formula is also found at the

beginning of the Lakkhaa Sutta (DN 30), in which the Buddha describes the 32

marks and (at least in the Theravda version) the karmic actions that lead to them to

his monks. 914 Of these two, the Mahpadna Sutta is almost certainly late, perhaps

even later than the theme of the Brahman searching for the 32 marks, because it refers

to the concept of past Buddhas, and moreover does so employing the fully developed

life story of the Buddha Skyamuni, which is ironically not found applied to

Skyamuni himself anywhere in the canonical Pali tradition, but only in Nidnakath,

which is the introduction to the Jtaka commentary. The Lakkhaa Sutta, on the

other hand, may be late in the form in which it appears in the Pali Canon, but this

stra is among the ten that were apparently traded between the Dgha/Drgha and

Majjhima/Madhyama collections by either the Theravda or the Sarvstivda

tradition, and the version preserved in the Sarvstivda Madhyamgama (M 59) is

much shorter than the Theravda version and could possibly be of some antiquity. 915

In the shorter, Chinese version, the Buddha simply overhears the monks talking about

the 32 marks using the words of the two paths of a Great Man formula and tells

914
Interestingly, the Lakkhaa Sutta is the only place in the early Buddhist tradition, at least
as far as I am aware, where knowledge of the 32 marks is not explicitly associated with Brahmans. In
the Pali version, the Buddha does concede, after naming the 32 marks, that they are known by
outsider is (bhirakpi isayo), but these non-Buddhists do not know the karmic roots of each of the
32 marks. This serves as a segue into the rest of the sutta, in which the Buddha describes in great
detail the meritorious deeds he performed to acquire each of the marks. Neither the concession that
outsider is know the marks nor the explanation of the meritorious deeds that lead to them are found
in the Sarvstivda version, which is found not in the Drghgama, but in the Madhyamgama (M
59).
915
Sujto has in fact suggested that it may have been the Theravdins who moved the
Lakkhaa Sutta to the Dgha Nikya, and added a considerable amount of material to it in the process
to make it appropriately long for its new homeSujto, History of Mindfulness, 59-60.

577
them what the 32 marks are; no mention is made of Brahmans at all as purveyors of

this tradition.

Let us now look more closely at the three suttas, within the Theravda

tradition, that do make use of the theme of the Brahman looking for the 32 marks.

The three are the Ambaha Sutta (DN 3), the Brahmyu Sutta (MN 91), and the Sela

Sutta (MN 92). Of these three, the Ambaha and Brahmyu Suttas follow a very

similar pattern, while the Sela Sutta makes use of the 32-marks theme in a slightly

different way. In both of the former two suttas, there are two Brahman interlocutors,

namely a Brahman teacher and his student (mava). The teacher (in the case of the

Ambaha Sutta, Pokkharasti, and in the case of the Brahmyu Sutta, Brahmyu)

hears of the presence of the Buddha through the fame of Gotama formula and then

asks his student (Ambaha and Uttara, respectively) to go and find out whether the

reputation Gotama has earned is true. The student then asks his teacher how he can

find this out, and the teacher explains, using the two paths of a Great Man formula,

that there are thirty-two marks of a Great Man, and that one who possesses these

marks can only become either a cakkavatti monarch or a Buddha. Note that this

aspect of the 32-marks theme is what allows the 32 marks to become a topic for

discussion in the first place. It is the fame of Gotama formula that makes the

Brahman teacher aware of the Buddhas presence and motivates him to send his

student to him, but there is no mention of the 32 marks in that formula. The students

request for further clarification on how to tell if the Buddhas reputation is well-

earned provides the narrative opportunity for the 32 marks to be introducedthey are

578
to be used, in the context of this narrative theme, as an outward sign that Gotama is

indeed a Buddha, as his reputation suggests.

At some point during the students meeting with the Buddha, the student sees

and is able to confirm that the Buddha has at least 30 of the 32 marks. Two, however,

are not immediately visible to hima long tongue and a sheathed penisand the

Buddha must demonstrate these specially to himthe former by licking every part of

his face, and the latter through an apparently magical feat that allows the Brahman

student to see his penis. The student then returns to his teacher and reports what he

has seen, and the teacher decides to go see for himself. The process is then repeated:

The Brahman sees 30 of the 32 marks unassisted, and the Buddha must demonstrate

the remaining two to him.

The Sela Sutta is slightly different from the other two suttas that make use of

this theme in that it involves only one Brahman interlocutor, Sela. There is another

interlocutor in the sutta, namely Keiya, a jaila who is apparently a follower of Sela,

but Keiya serves primarily to introduce Sela to the Buddha and shows no interest in

the Buddhas 32 marks himself. Instead, Sela becomes aware of the Buddhas

presence when he sees Keiya preparing a large meal for him. Keiya describes the

Buddha to Sela using the words of the fame of Gotama formula, and this prompts

Sela to simply think to himself, in the words of the two paths of a Great Man

formula, that if the fame of Gotama is true, then the Buddha must have the 32

marks. Sela does not send anyone on his behalf first, but instead immediately goes to

579
see for himself. As in the other two suttas, Sela at first only sees 30 marks and must

be shown the other two.

Scholars have long been fascinated, and frustrated, by the 32 marks of a Great

Man. With the exception of the Lakkhaa Sutta, and particularly its Chinese parallel,

the early Buddhist tradition is unanimous in associating the 32 marks with Brahmans,

and doing so (in most cases) in the context of encounter dialogs. That is to say, the

32 marks are almost always brought up in a context where proponents of the new

Brahmanism are introduced as interlocutors in order to serve as a foil against which to

construct Buddhist identity, and within that context, knowledge of the 32 marks

serves as a marker of the Brahmans identity as other to the Buddhists, while

simultaneously their recognition of the marks in the Buddha serves as a polemical

trope in which the greatness of ones own leader is certified by ones opponents. The

association of the 32 marks with Brahmans, it must be emphasized, is close to the

point of near exclusivity. As already discussed, most references to the 32 marks in

the Pali Canon are found within the Triple Veda formula, wherein knowledge of the

marks is named, along with knowledge of the Vedas, five auxiliary Vedic sciences,

and Lokyata, as a sign of the learnedness of a particular Brahman. As if this were

not clear enough, the two paths of a Great Man formula that is deployed in the 32-

marks theme and a couple other contexts begins with the explicit statement that there

are thirty-two characteristics in our mantras with which a Great Man [is]

580
endowed. 916 The use of the very specific term in our mantras (P. mantesu) here

would appear to imply not only that Brahmans prided themselves on knowing the 32

marks of a Great Man, but that they derived them from the Vedas.

The only problem is that no one has been able to identify a list of 32 marks

of a Great Man in the Vedas. This, it would appear, is a very old problem, dating

back to long before the advent of modern scholarship. In his commentary on the

Brahmyu Sutta in the Majjhima Nikya, Buddhaghosa writes,

Here, in the mantras means in the Vedas. Having heard that a Tathgata
will arise, the Suddhvsa gods in preparation put the characteristics into the
Vedas and, in the guise of Brahmans, teach in the Veda, These are called the
Buddha-mantras, thinking, Thus, influential beings will recognize the
Tathgata. In this way, the characteristics of a Great Man came into the
Vedas in the past. But when the Tathgata passed into parinibbna, they
gradually disappeared, and so now they arent there. 917

Thus, it appears that as early as the fifth century, even one of the greatest Buddhist

scholars in all of history was unable to find any justification for the claim that the

Brahmans had a list of 32 marks of a Great Man in the Vedas, and thus had to invent

a story about mantras being temporarily interpolated into the Vedas by gods

masquerading as Brahmans to explain the claim made by the two paths of a Great

Man formula.

916
Found at, e.g., MN 91: amhka mantesu dvattisamahpurisalakkhani, yehi
samanngatassa mahpurisassa.
917
tattha mantesti vedesu. tathgato kira uppajjissatti paikacceva suddhvs dev vedesu
lakkhani pakkhipitv buddhamant nma eteti brhmaavesena vede vcenti tadanusrena
mahesakkh satt tathgata jnissantti. tena pubbe vedesu mahpurisalakkhani gacchanti.
parinibbute pana tathgate anukkamena antaradhyanti, tena etarahi natthi. Cited by John Powers, A
Bull of a Man: Images of Masculinity, Sex, and the Body in Indian Buddhism (Cambridge, Mass.:
Harvard University Press, 2009), 18.

581
 Modern scholars have similarly grappled with the problem of the 32 marks

and have also been unable to identify an unambiguous source for the list of 32 marks

in the Vedas, or any other Brahmanical text for that matter. Rhys Davids commented

in a footnote in his translation of the Dgha Nikya that [n]o such list has been

found [a]nd the inference from both our passages is that the knowledge is scattered

through the Brahman texts. 918 Much of the modern scholarship since Rhys Davids

has, indeed, been focused on parallels to particular marks, or the idea of auspicious

bodily signs or of a Great Man in general in the Brahmanical literature, though not

with a great deal of success. A useful synopsis of this scholarship has been provided

recently by John Powers, who also adds some of his own findings from various Vedic

and non-Vedic Brahmanical texts. 919

Surprisingly, however, Powers does not mention in this synopsis one of the

earliest and most extensive discussions of the problem of the 32 marks, which was

provided by Emile Senart in his 1875 Essai sur la Lgende du Buddha. This

omission is understandable, however, given that Senarts Essai is associated primarily

with his theory of the origins of the Buddhas biography in an ancient solar myth, a

theory that fell into disfavor around the turn of the twentieth century with the

ascension of the historicist school of Hermann Oldenberg and T. W. Rhys

918
Rhys Davids, Dialogues of the Buddha, vol. 1, 110n.2. Cited by Powers, Bull of a Man, 17.
919
Powers, Bull of a Man, 16-19.

582
Davids, 920 thus banishing the Essai as a whole to relative obscurity. Senart discusses

the 32 marks in Chapter Two of his Essai, and although his findings overall are

certainly datedhis use of various Brahmanical and Buddhist sources is limited by

what was available to him and a rather limited sense of their chronological

relationship to one another, and the evidence he presents is marshaled in defense of

his broader thesis that the Buddhas biography has its origins in a solar mythSenart

nonetheless provides in this chapter an extensive and detailed discussion of what

evidence does exist for ideas relating to the 32 marks in the Brahmanical traditions,

which I believe has yet to be truly surpassed.

The most important Brahmanical source that Senart cites in this chapter is the

Bhat Sahit of Varha Mihira. That certain parallels to the 32 marks can be found

in this 6th century work on divination was noted briefly by Rhys Davids in his

introduction to the Lakkhaa Sutta, 921 and it has also been mentioned by more recent

scholars, including Powers, 922 but to my knowledge Senart was the first to recognize

this fact and one of the only to discuss the parallels to the 32 marks found in that text

in any detail. 923 Given that the parallels to the 32 marks found in the Bhat Sahit

920
Frank E. Reynolds and Charles Hallisey, The Buddha, in Buddhism and Asian History,
ed. Joseph M. Kitagawa and Mark D. Cummings (New York: Macmillan Publishing Company, 1989),
31.
921
Rhys Davids, Dialogues of the Buddha, vol. 3, 135.
922
Powers, Bull of a Man, 255n.77.
923
Emile Senart, Essai sur la Lgende du Buddha: Son Charactre et ses Origines (Paris:
Imprimerie Nationale, 1875), 134-5.

583
are frankly the most extensive that anyone has ever identified in any one place, it

behooves us to discuss them here in some detail.

Two chapters in the Bhat Sahit (68-69) are relevant to the search for

actual evidence of something like the 32 marks within the Brahmanical tradition. The

first of these, ch. 68, which is entitled Characteristics of Men (purualakaam), is

dedicated to physiognomy. As such, it does not deal with the concept of a Great

Man, much less does it give a list of 32 characteristics thereof; rather, it discusses in

general what the characteristics of various parts of a persons body portend about his

future. In most cases, a particular body part is addressed in a single verse, which

summarizes all the possible characteristics that body part can have and what kind of

person will have each of those possible characteristics. Thus, for example, the

following verse addresses the possible characteristics of a persons knees and what

they portend:

One with fleshless knees dies abroad; with small [knees people have] good
fortune; with monstrous [knees people are] poor.
Moreover, with low [knees people] are subdued by women, and with fleshy
[knees people attain] kingship; with large [knees people have] long
life. 924

As can be seen from this example, Varha Mihiras concern in this chapter on

divination is not with describing the characteristics of one particular kind of person,

great or otherwise, but with explaining how one can make a prediction about any

person based on the character of a particular body part, in this case the knees.

924
BS 68.3: nirmsajnur mriyate pravse saubhgyam alpair vikaair daridr /
strnirjit caiva bhavanti nimnai rjya samsai ca mahadbhir yu //.

584
 Although this particular verse does not contain any information that is parallel

in any way to the 32 marks listed in the early Buddhist tradition (none of which have

anything to do with the knees), I have identified 16 marks of a Great Man, as listed in

the early Buddhist tradition, that have possible parallels in chapter 68 of the Bhat

Sahit. These can be found listed in Table 18. Interestingly, most of these

parallels10 out of the 16involve kings. That is, when discussing a particular

body part, the attributes of that body part that parallel in some way one of the 32

marks of a Great Man listed in the early Buddhist tradition in most cases happen to be

those attributes that Varha Mihira claims portend kingship. In the few cases where

the attribute that is parallel to a mark of a Great Man is not associated with kingship,

it is instead associated with long life, auspiciousness, or in one case enjoyment. The

general association of attributes that parallel various marks of a Great Man with

kingship, however, is, I think, significant, and is easy to understand if there is indeed

a historical relationship between the 32 marks and Varha Mihiras chapter on

physiognomy. As we know from the two paths of a Great Man formula, the

Buddha was, because of his 32 marks, destined from birth to become, if not a Buddha,

then a world-conquering cakkavatti monarch. Indeed, because of this two-sided

destiny, the person of the Buddha has long been associated with kingship in numerous

contexts throughout history.

585
Mahpurisa- Charac- Verse Relevant Portion of Transla- Comments
lakkhaa teristic of a Num- Verse tion
Great Man ber
of
Paral
-lel in
BS
suppatihitapdo well-
established
foot
heh kho pana tassa Underneath 69.17 halamusalagadsi- feet and Refers
bhoto gotamassa the soles of akhacakradvipa- hands specifically
pdatalesu cakkni the feet of makarbjarathkitg marked by to the
jtni sahassrni the Ven. hri-hasta a plough, Mahpuru
sanemikni Gotama arise pestle, a of the
sanbhikni thousand- club, bhadra
sabbkrapariprni spoked sword, class.
wheels conch,
complete in wheel,
all aspects elephant,
with rim and crocodile,
hub. lotus, or
chariot.
yatapahi outstretched
heels
dghaguli long fingers 68.36 hastgulayo drgh Those with
ciryum long life
have long
fingers.
mudutalunahatthapd soft and 68.2 mdutalau caraau The feet of
o tender hands manujevarasya a Lord of
and feet Men are
soft-soled.
jlahatthapdo netted hands 68.2 ligul caraau The feet of The Pali for
and feet manujevarasya a Lord of this mark is
Men obscure,
have but has
connected traditionall
toes. y been
interpreted
as referring
to webbed
fingers and
toes. This
interpretati
on would
appear to
be
supported
by BS 68.2.
ussakhapdo over-shell 68.2 krmonnatau The feet of Both the
feet (??) (?) caraau a Lord of Pali for this

586
 manujevarasya Men are mark and
curved up the Sanskrit
like a for the
tortoise. possible
parallel are
obscure. It
is possible
that both
refer to
arched feet.
eijagho lower leg of
an antelope
hitako standing, 68.35 jnvavalambinau The arms
anonamanto ubhohi without bh pthivnm of Lords of
pitalehi jaukni bending, he the
parimasati rubs and Earth
parimajjati strokes his hang down
knees with to the
the palms of knees.
both his
hands
kosohitavatthaguyho that which is 68.8 koanighair bhp Those with The full
hidden by [a penis] context of
clothing (i.e., concealed this verse
the genitals) by a sheath makes it
is enclosed (i.e., clear that
by a sheath foreskin) koa
(i.e., the [become] (sheath)
foreskin) kings. refers to the
foreskin.
See also BS
69.16 on
the
Mahpuru
a of the
bhadra
class.
suvaavao golden- 68.10 dyutimn vara Kings have Unlike the
kacanasannibhattac colored 2 snigdha kitipn a bright, description
o with skin shining of the mark
resembling complexion in Pali, the
gold . BS parallel
does not
compare
the skin
specifically
to gold.
sukhumacchavi fine skin 68.10 vara uddha A clean
sukhumatt chaviy dirt and 2 ubhado complexion
rajojalla kye na perspiration is
upalimpati do not stick auspicious.
to the skin

587
 due to its
fineness
ekekalomo ekekni one body- 68.5 romaikaika kpake Kings have See also BS
lomni lomakpesu hair grows prthivn a single 69.16 on
jtni per pore body-hair the
per pore. Mahpuru
a of the
bhadra
class.
Uddhaggalomo with 68.26 pradakivartaromabh With body- See also BS
uddhaggni lomni bristling ir bhp hair curling 69.16 on
jtni nlni body- to the right the
ajanavani hairs the [people Mahpuru
kualvani bristling become] a of the
dakkhivaakajtn body-hairs kings. bhadra
i grow indigo, class.
the color of
collyrium,
turning like
rings, they
grow turning
to the right
brahmujugatto straight body
of Brahm
sattussado seven
protuberance
s (??)
shapubbaddhakyo the front half 68.18 sihakair One with The verse
of a lion manujendra the here is
hips/buttoc parallel
ks of a lion only in
[becomes] comparing
a Lord of a person to
Men. a lion; the
part of the
body so
compared
is different.
citantaraso heaped chest 68.27 hdaya samunnata Lords of This
(lit., pthu msala ca Men have a parallel is
between the npatnm heart that is valid only
shoulders) raised up, if
broad, antarasa
and in Pali
muscular. refers to the
chest and
not, as
sometimes
interpreted,
to the upper
back.
nigrodhaparimaalo round like a 69.13 bhujayugalapramita His height Refers

588
 yvatakvassa kyo Banyan samucchrayo sya is the specifically
tvatakvassa bymo, tree his measure of to the
yvatakvassa bymo body is as his two Mahpuru
tvatakvassa kyo tall as the arms. a of the
length of his bhadra
outstretched class.
arms; his
outstretched
arms are as
long as his
body is tall
samavaakkhandho back/shoulde
rs even and
round
rasaggasagg extremely
refined sense
of taste
shahanu jaw of a lion
cattlsadanto 40 teeth
samadanto even teeth 68.52 daan sam Even
ubh teeth are
auspicious.
aviraadanto teeth without 68.52 ghan daan Compact
gaps ubh teeth
are
auspicious.
susukkadho very white 68.52 sutkadar Very sharp This
(canine?) ubh canine parallel
teeth teeth are verse
auspicious. makes one
wonder if
the Pali
susukka
(very
white) is a
mistake.
Note that
the Pali
here says
dha
instead of
dantabut
why single
out the
canine
teeth as
white?
pahtajivho large tongue 68.53 jihv drgh Those with
bhogin jey long ...
tongues
should be
known as

589
 enjoyers.
brahmassaro with the
karavikabh voice of
Brahm
speaking like
a cuckoo
bird
abhinlanetto deep indigo BS 68.64-7
eyes discusses
eyes, but
lacks any
clear
parallels to
this mark.
gopakhumo eyelashes of
an ox
u kho panassa Ven. Gotama
bhoto gotamassa has white
bhamukantare jt hair growing
odt between the
mudutlasannibh eyebrows
resembling
soft cotton.
Uhsasso turban- BS 68.79-
headed (??) 80
discusses
head-
shapes, but
does not
mention an
ua
shape.
Kings have
heads like
parasols
(chatra)
and people
with long
life have
heads like
skulls
(kapoi).
Table 16. A list of all 32 marks of a Great Man (dvattisamahpurisalakkhani), along with parallels found in
ch. 68-9 of Varha Mihiras 6th century Bhat Sahit.

Some of the parallels I have identified are fairly clear and unproblematic. For

example, one of the 32 marks of a Great Man is long fingers, and according to BS

68.36, long fingers portend long life. Likewise, the marks of soft and tender

590
hands/feet, arms that stretch down to the knees, only one hair per pore, hair that curls

to the right, even teeth, and a long/large tongue all have relatively unambiguous

parallels in Chapter 68 of the Bhat Sahit. Other potential parallels, however, are

less clear. According to the early Buddhist tradition, a Great Man has golden skin

that is so fine, nothing sticks to it. Varha Mihira, on the other hand, speaks only of

skin that is bright (dyuti) and shining (snigdha), without comparing it to gold,

and says that a clean or pure (uddha) complexion is auspicious. In another case,

there is a significant parallel in that the early Buddhist tradition and Varha Mihira

both compare the body to a lion, but they do so in reference to different parts of the

body. One of the 32 marks of a Great Man is having the front half of the body

(pubbaddhakya)presumably referring to the torsoof a lion, while Varha Mihira

says that a person whose hips/buttocks (kai) are like a lions will become a king.

Other parallels are problematic mostly because the Pali of the early Buddhist

list of 32 marks and/or the Sanskrit of the Bhat Sahit is obscure. The exact

meaning of the Pali jlahatthapda (netted hands and feet) has long been debated,

but it typically is understood as referring to what we would refer to idiomatically in

English as webbed fingers and toes. This interpretation would appear to be

corroborated by BS 68.2, which refers to connected (lia) toes. The mark

immediately following this one describes feet using the obscure word ussakha

possibly derived from the upasarga ut- and the word sakha for conchand has

sometimes been taken to refer to an arch in the shape of the foot. This may also be

parallel to BS 68.2, which, using a slightly different metaphor, refers to feet that are

591
curved up (unnata) like a tortoise (krma). Another mark later in the list of 32

refers to a full between the shoulders region (antarasa). This is sometimes taken

to refer to the space between the shoulder blades, but it can also refer to the chest, in

which case it would be parallel to BS 68.27, which says that kings have a heart that is

raised up, broad, and muscular (samunnata pthu msala ca).

Teeth figure prominently as the subject of four marks of a Great Manthey

should be 40 in number and even, plus have two other attributes whose meaning is

less clear. Varha Mihira makes no mention of people with 40 teethwhich is not

surprising since it is eight more than the normal number in an adults mouthbut he

does refer to various attributes of teeth as auspicious in BS 68.52. One of these

attributes is even (sama), the same word used in the second of the four marks of a

Great Man that pertain to teeth. The third of these four marks says that the Great

Mans teeth are aviraa, which means not sparse or not thin. amoli and

Bodhi translate this as referring to teeth that are without gaps, 925 and indeed Varha

Mihira describes auspicious teeth as being ghana, or compact. The last of the four

marks relating to teeth is somewhat irregular in that, instead of using the word danta

for teeth, it uses the word dh, which properly speaking does not mean teeth in

general, but rather tusks, or in the context of a human mouth, the cuspids (canine

teeth). These are then described as susukka, which can only mean very white. This

would make sense if what was being described were the Great Mans teeth in general,

but it is somewhat odd that specifically his canine teeth are described as whiteafter

925
Bhikkhu amoli and Bhikkhu Bodhi, The Middle Length Discourses of the Buddha: A
new Translation of the Majjhima Nikya (Boston: Wisdom Publications, 1995), 746.

592
all, what is so great about having a mouth full of mostly yellow teeth, with only the

four canines being white? BS 68.52 suggests a solution, however: Near the end of

the verse, it comments specifically on the cuspids (dar), saying that it is

auspicious for them to be, not very white, but very sharp (sutka). This certainly

makes more sense in the contextsharpness is a quality that is associated, within the

context of the human mouth, specifically with the canine teeth to the exclusion of all

others, while whiteness is certainly a quality that one would desire in all teeth, or at

least as many as possible. It is difficult not to wonder if perhaps Varha Mihira has

preserved here an auspicious characteristic of teethsharp cuspidsthat has become

garbled in the early Buddhist tradition.

One final mark of a Great Man warrants a discussion herenamely, the well-

known sheath that covers the Buddhas genitals. Ordinarily, this is the word that is

used in translations, since this is the literal meaning of kosa, but this, I believe, gives

the mistaken impression that what is being referred to here is something unusual or

foreign to ordinary human anatomy. In fact, the entire compound used to describe

this mark is, not surprisingly, full of euphemisms, and kosa is just one of them. The

full compound (kosohitavatthaguyho) literally means that which is hidden by

clothing is enclosed by a sheath. That which is hidden by clothing (vatthaguyha)

is a euphemism for the penisthis much translations have made clear, sometimes

with the similarly eumphemistic male organbut sheath (kosa) is a euphemism

as well, namely for the foreskin. The parallel I have identified in the Bhat Sahit,

in fact, uses exactly the same euphemism, spelled koa in Sanskrit. The full context

593
of the verse, however, makes it abundantly clear that is the foreskin that is being

referred to:

Those with [a penis] concealed by the sheath [become] kings; those with a
long and split one [become] totally bereft of possessions;
those with a straight, round penis and those with a tail that has numerous
light veins become wealthy. 926

Within the context of this verse, the parallel to the mark of a Great Man is found, as

in so many cases, in the attributes said to portend kingship. Because this verse

discusses other types of penises, however, we can see what is distinctive about the

penis concealed by the sheath. The type of penis Varha Mihira discusses next

that of a person destined to povertyis said to be long and split. Clearly what must

be being referred to here is the quality of the foreskin. In the case of a person

destined to kingship, the foreskin completely covers the penis. In the case of a person

destined to indigency, however, the foreskin does not cover the penis wellit is

split. The Buddha, then, does not have some sort of inborn chastity belt or other

abnormal sheath, as is sometimes suggested; rather, he simply has a kingly

foreskinnamely, one that completely covers his penis.

The next chapter of the Bhat Sahit, Chapter 69, is entitled

Pacamahpurua-lakaa, or The characteristics of the five Great Men. Given

such a title, this chapter has attracted the attention of scholars studying the 32

Buddhist marks since Senart, but most have emphasized the differences between the

ideas presented in this chapter and the concept of the Great Man in the early Buddhist

926
BS 68.8: koanighair bhp drghair bhagnai ca vittaparihn / juvttaephaso
laghuirlain ca dhanavanta //. Quotation marks do not translate iti, but rather are used to
indicate euphemisms.

594
tradition. In this chapter, Varha Mihira discusses not one, but five different types of

Great Man, each of which is associated astrologically with a particular planet.

Dwelling on this admittedly not insignificant difference, however, scholars have

overlooked the fact that one of the five types of Great Man in particular stands out

from the other four in bearing striking similarities to the Great Man described in the

early Buddhist tradition, and even to the Buddha himself. This is the Great Man of

the bhadra type, which is associated with the planet Mercury (budha) and is

described in verses 13-19 of BS 69.

Within the description of the Great Man of the bhadra class, several parallels

to the marks of a Great Man found in the early Buddhist tradition are found, including

two that were not already found in Chapter 68. According to verse 16 of Chapter 69,

the Great Man has head-hairs that grow singly (i.e, one to a pore), are black, and are

curled, and that which is hidden (i.e., the penis) is completely concealed like that of

a horse or elephant. 927 Although this verse speaks of head-hairs being dark, curled,

and one to a pore, and the corresponding marks of a Great Man given in the Buddhist

tradition describe only body-hairs in these terms, it seems likely that the latter is a

result of the narrative context, in which the Brahman interlocutor could not possibly

see anything about the Buddhas head-hair because it would have been shaved off.

Indeed, there is nothing in the early Buddhist tradition, at least in the Theravda

version, to indicate that the Buddha had anything other than a completely shaven head

like his monks, and the 32 marks of a Great Man correspondingly have nothing to say

927
BS 69.16cd: iroruh caikajakakucits turagangopamaguhyaghat.

595
about the hairs of the head. Nevertheless, we can safely assume, I believe, that there

is a parallelism between the head-hairs and body-hairs. As we have already seen,

Chapter 68 of the Bhat Sahit describes the body-hairs in similar termsin

particular, as growing one to a pore and being curledand although the early

Buddhist oral tradition clearly regarded the Buddha as shaven-headed, and thus his

head-hairs as irrelevant, it appears that the later artistic tradition did not. In other

words, the parallelism between the body-hairs and head-hairs found in the Bhat

Sahit may explain why artists typically came to depict the Buddha as having a

head full of many tight, individual curls of hair. 928

As for the description of the bhadra-class Great Mans genitals, we find an

additional detail here that was not found in verse 8 of Chapter 68. That is, the

concealed penis is compared to that of a horse or elephant. This, again, reinforces

the conclusion that the sheath referred to both in the list of 32 marks and in BS 68

must be the foreskin, since horses and elephants have concealed penises only

insofar as they are well-covered by their foreskins. Interestingly, as Bhikkhu Anlayo

has pointed out, the Sarvstivda version of the list of 32 marks, preserved in the

Chinese translation of the Madhyamgama, also compares the Buddhas penis to that

of a horse. This gives additional support to our hypothesis that there is a connection

between the Great Man described in the early Buddhist tradition and Varha Mihiras

bhadra-class Great Man.

928
On the history of Western scholars confusion over the Buddhas hair, as well as a
traditional Buddhist explanation (that the Buddha never shaved, but simply cut his hair roughly with a
sword when he left the palace), see Lopez, Buddha, 13-36.

596
 The description of the Great Man of the bhadra class in BS 69 also contains

two additional parallels to marks of a Great Man as described in the early Buddhist

tradition that are not found at all in Chapter 68. According to verse 13, a Great Man

of the bhadra class has a height [that is] the measure of his two arms

(bhujayugalapramita samucchrayo). This clearly parallels the mark of a Great Man

in the Buddhist tradition that states that his body is as tall as the length of his

outstretched arms (yvatakvassa kyo tvatakvassa bymo). In addition, one of the

marks of a Great Man in the Buddhist tradition is that he has thousand-spoked wheels

on the soles of his feet. Nothing of this sort is mentioned in Chapter 68 of the Bhat

Sahit, but in the description of the bhadra-class Great Man in Chapter 69, we find

a parallel of sorts in that a Great Man of this type is said to have the palms of his

hands and the soles of his feet marked with any of a number of shapes, one of which

is a wheel (cakra). 929

In addition to these parallels to specific marks of a Great Man as described in

the early Buddhist tradition, this description of the bhadra-class Great Man has other

interesting details that parallel the story of the Buddha. In verses 14-15 of Chapter 69,

Varha Mihiri lists a number of general attributes of a Great Man of this class. While

most of these attributes refer to his physical perfection and acute intelligence, Varha

Mihira also mentions that this type of Great Man is a yog, which would obviously be

appropriate for the Buddha since he is also considered a yog by the Buddhist

929
BS 69.17: halamusalagadsiakhacakradvipamakarbjarathkitghrihasta. The
complete list of possible figures named here is plough, pestle, club, sword, conch, wheel, elephant,
crocodile, lotus, or chariot.

597
tradition. In addition, in verse 18, he writes that he becomes a king of the middle

country (madhyadeanpatir). This by itself is not of great significance because all

of the five types of Great Men described in this chapter are said to be destined to

become rulers of one part of the world or another. After saying that the bhadra-class

Great Man becomes a king of the middle country, however, Varha Mihira adds that

if he has the proper dimensions, 930 he becomes Lord of the whole earth

(sakalvanintha). This is, of course, parallel to the two paths of a Great Man

formula, which states that a Great Man, if he remains a householder, will become a

world-conquering cakkavatti monarch, and it is given as a possibility within the

context of ch. 69 of the Bhat Sahit only for a bhadra-class Great Man, and not for

any of the other four. Finally, in verse 19, Varha Mihira writes that the bhadra-class

Great Man lives for 80 years. 931 In this chapter, Varha Mihira assigns a different

life-span to each class of Great Man, and the bhadra-class alone has a life-span of

exactly 80 yearswhich of course is the same age the Buddhist tradition holds the

Buddha to have been when he died.

930
The Sanskrit here is obscure: yadi pus trydayo sya, which means something like, If
his dimensions are three , with the word di indicating an ellipsis. Presumably Varha Mihira is
referring back to verse 7, where he describes in detail the dimensions (i.e., height and span of two
outstretched arms) of each of the classes of Great Man as three digits greater than that of the previous.
Bhat therefore translates this portion of v. 18 as if his height and extent of outstretched arms be each
105 digits M. Ramakrishna Bhat, trans., Varhamihiras Bhat Sahit, vol. 2 (Delhi: Motilal
Banarsidass, 1982), 652.
931
BS 69.19: bhuktv samyag vasudh auryeoparjitm aty abda / trthe prs
tyaktv bhadro devlaya yti //, Having rightly enjoyed (i.e., ruled) the earth, which he has gained
with valor, for eighty years, the Bhadra abandons his breaths at a ford and goes to the abode of the
gods.

598
 Thus, to summarize, more than half of the 32 marks of a Great Man listed in

the early Buddhist tradition have parallels in Varha Mihiras discussion of

physiognomy in Chapter 68 of his Bhat Sahit, and most of those are associated by

the latter with kingship. In addition, in Chapter 69 of the Bhat Sahit, Varha

Mihira describes five different types of Great Men (mahpurua), and one of these

types in particular, the bhadra, has a number of attributes that are parallel to

particular marks of a Great Man as described in the Buddhist tradition, as well as

others that parallel aspects of the Buddhas own life. Given this large number of

correspondences, it seems quite clear that there is a historical correspondence

between the Buddhist list of 32 marks of a Great Man and the Brahmanical science of

physiognomy as described by Varha Mihira. Although it is possible that the

Buddhists themselves created the specific concept of 32 marks, it is clear that they did

not invent the concept of a Great Man, and they did not create at least most of the

individual marks in the list, but rather borrowed them from actual Brahmanical

physiognomy.

There is, of course, only one remaining problem with the parallelism between

the list of 32 marks and the Bhat Sahit: The latter was not written until the sixth

century CE. This is, at the very least, nine centuries after the death of the Buddha,

and certainly quite a bit later than the theme of the 32 marks of a Great Man in the

early Buddhist tradition. 932 This raises the obvious question: What were Varha

932
Even the most skeptical reconstruction of the chronology of early Buddhist literature must
accept that the contents of the Pali Canon were fixed by the time of Buddhaghosa, who lived a century
before Varha Mihira. In addition, the Sarvstivdin Madhyamgama, which contains, e.g., a version
of the Brahmyu Sutta, was translated into Chinese even earlier, at the very end of the fourth century.

599
Mihiras sources, and how far back can we trace back the physiognomical ideas we

have been examining here in those sources? Unfortunately, not much work has been

done on the history of physiognomy in India, and based on what little work has been

done, it does not appear that Varha Mihiras sources for the ideas he presents in

Chapters 68 and 69 can be traced at all. According to David Pingree in his volume on

Jyotistra, works on divination in India, including the Bhat Sahit, which

became the classic work of the genre, are for the most part derived from the Garga

Sahit, which was probably written in the 1st century BCE or the 1st century CE.

Unfortunately, this text, which is the earliest extant treatise on divination in India, has

not been published. According to Pingree, however, it clearly served as a blueprint

for Varha Mihiras much later work insofar as there is a direct correspondence

between the topics of many of the chapters of the latters Bhat Sahit and the

topics of the chapters of the Garga Sahit. Chapters 68 and 69 on physiognomy

and the the characteristics of the five types of Great Men, however, are not among

them: While the immediately preceding Chapter 67 and the immediately following

Chapter 70 correspond to chapters in the Garga Sahit, Chapters 68 and 69 do

not. 933

Clearly, much more work needs to be done in tracing the development of

divination in India, including a thorough investigation of the contents of the Garga

933
David Pingree, Jyotistra: Astral and Mathematical Literature, ed. Jan Gonda, A
History of Indian Literature, vol. 6, fasc. 4 (Wiesbaden: Otto Harrassowitz, 1981), 69-74. Note that
Pingree uses a different numbering scheme for the chapters of the Bhat Sahit than I have been
using here; thus, what I have been calling Chapters 68 and 69, he calls Chapters 67 and 68. This can
be seen from his description of the content of those chapters on p. 74.

600
Sahit and the identification of any other sources that were drawn upon by Varha

Mihira. Until then, we have hit a bit of a dead end in trying to trace the Brahmanical

parallels to the concept of the 32 marks of a Great Man back beyond the 6th century

and must resort mostly to speculation. According to the early Buddhist tradition, as

we have already seen, the Brahmans supposedly derived the entire concepts from

their mantrasi.e., from the Vedas, and modern scholars have attempted, with

varying success, to identify parallels to the marks in the Vedas. 934 One obvious

source for the concept of a Great Man (mahpurua) would be the Vedic myth of

the cosmic man (purua), and indeed Eugne Burnouf suggested that the purua of

the Purua Skta in the g Veda (10.90) was the source of the concept, but as Powers

has pointed out, the description of purua in this hymn bears no resemblance to the

physical attributes of the Buddha as described in the Pli canon. 935

Once again, Senart provides what is likely the most useful suggestion in this

regard. He points to another, less well-known Vedic hymn on purua, which is found

in the Atharva Veda (10.2). Although there are not many clear parallels to the 32

marks of a Great Man described in the Buddhist tradition in this hymn, the hymn does,

especially in its first half, discuss the purua in fairly minute anatomical detail, and so

it is conceivable that certain ideas on physiognomy, and in particular the anatomy of a

Great Man, could have been derived from a creative reading of this text. The most

obvious parallel I have found in the text is in verse 7, in which it is said that he has a
934
Again, see Powers, Bull of a Man, 17.
935
Ibid. Unfortunately, Powers does not provide a citation here, so I am not sure where
Burnouf suggested that the Purua Skta was the source of the concept of the Mahpurua.

601
full tongue (jihv m purc m). Likewise, verse 4 asks, Which and how many

gods were they who heaped up the breast (and) neck of man? 936 The verb used here

for heap up is -ci-, and one of the 32 marks in the Buddhist tradition uses the past

passive participle of this verb (cita) to describe the antarasa, which I have already

argued probably refers to the chest. Finally, verse 6 asks, Who bored out the seven

apertures (i.e., mouth, eyes, ears, and nostrils) in his head? 937 It is possible that this

bears some relationship to the obscure Pali phrase sattussada, although this has

usually been interpreted as referring to seven protuberances, rather than seven

apertures. In any case, further study of AV 10.2 in comparison with both the

Buddhist list of 32 marks and the Bhat Sahit may yield other possible examples of

ways in which this text may have been exegetized by later Brahman interpreters to

produce lists of auspicious anatomical features.

In the end, the state of our knowledge about the history of Indian divination

limits our ability to date the theme of the 32 marks of a Great Man on the basis of

external evidence. On the one hand, although the one source that has the most

convincing parallels to the list of 32 marks, the Bhat Sahit, is quite late, we can be

quite confident that the list of 32 marks is older, since both Buddhaghosas

commentary on the relevant Pali texts and the Chinese translations of parallels from

other Buddhist sectarian traditions predate Varha Mihira. On the other hand, the one

known source used by Varha Mihira, the Garga Sahit, is only about six centuries

936
AV 10.2.4ab: kti dev katam t san y ro grv ciky pruasya.
937
AV 10.2.6a: k sapt kh ni v tatarda ri.

602
older, bringing us back only to the turn of the era, and even it does not appear to

address the topics of physiognomy and the Great Man that are relevant here. Where

Varha Mihira derived the information for these topics is unclear, and even though

one possible candidate has been identified in AV 10.2, the Atharva Veda itself was, as

already mentioned, not even considered a Vedic text during the time that the

Theravda version of the early Buddhist tradition, at least, became solidified.

None of this inspires much confidence in the antiquity of the theme of the 32

marks of a Great Man. In addition, as we already saw earlier in this chapter, internal

Buddhist evidence surrounding this theme suggest that it was a fairly late

development in the early Buddhist tradition as well. One of the important formulae

deployed by the themenamely, the fame of Gotama formulamakes no mention

of the 32 marks. Likewise, the Sarvstivda version of another formula, the Triple

Veda formula, which may be older than the Theravda version, makes no mention of

the 32 marks. Indeed, in order to introduce the concept of the 32 marks, those stras

that do deploy the theme must resort to a narrative device to do so. That is, they

introduce a completely new formula, the two paths of a Great Man, which a

Brahman either uses to explain to his student how to know whether the Buddha is

worthy of his reputation (Ambaha and Brahmyu Suttas), or else thinks to himself

after hearing of the fame of Gotama (Sela Sutta). Taken together, this internal

evidence, along with the lack of evidence of antiquity to corresponding ideas in the

Brahmanical tradition, strongly suggests that the small handful of stras that refer to

the 32 marks are a relatively late development in the early Buddhist tradition.

603
IV.4.6 The Recurring Characters Vseha and Bhradvja

Stock characters and recurring characters are a common theme in the early

Buddhist tradition. Although the Buddha is, of course, the protagonist of most stras

within the early Buddhist tradition, there are certain other characters, most commonly

close disciples of the Buddha such as nanda and Sriputta, who also appear

frequently in early Buddhist stras and sometimes even become protagonists in their

own right. Interlocutors such as Brahmans also, although to a lesser extent, make

recurring appearances in the early Buddhist tradition. Because recurring characters

by their very nature bring a set of stock characteristics and motifs with them to each

stra they appear in, they can be considered to constitute themes in the oral

traditional sense, and as such, tracing their appearances throughout the early Buddhist

tradition can serve as another useful way of exploring the development of the latter.

One particularly interesting set of recurring Brahman characters in the early

Buddhist tradition is the pair of friends Vseha and Bhradvja. Vseha and

Bhradvja appear in only three suttas in the Pali Canon, all of them encounter

dialogs: the Tevijja Sutta (DN 13), the Vseha Sutta (MN 98=Sn. 3.9), and the

Aggaa Sutta (DN 27). Of these, the Tevijja Sutta has parallels in the

Dharmaguptaka Drghgama preserved in Chinese (Ch. D 26) and the Sarvstivda

Drghgama preserved in Sanskrit (Skt. D 45). It is part of the

Slakkhandha/laskandha collection that, as we saw in Chapter IV.3, appears to have

formed the oldest core of the Dgha Nikya/Drghgama. The Aggaa Sutta, on the

other hand, is one of the twenty stras that were exchanged between the

604
Dgha/Drgha and Majjhima/Madhyama collections by the Vibhajyavda and

Sarvstivda traditions. It has parallels in the Dharmaguptaka Drghgama (D 5)

and in the Sarvstivda Madhyamgama (M 154). 938 Finally, the Vseha Sutta has

no known parallels outside of the Theravda tradition. It is found within the Pali

Canon at both MN 98 and Sn. 3.9, but it has a distinctly Sutta Nipta stylenamely,

it is composed primarily of verses, with only a short prose introduction. Since the

Sutta Nipta is unique to the Theravda tradition, this may help to explain why the

Vseha Sutta has no non-Theravda parallels. 939

These three suttas as they are found in the Theravda tradition are preserved

in three fairly different contexts, but can be read together as forming a sort of

progression, from Tevijja to Vseha to Aggaa. We have already looked at the

Tevijja Sutta, in Section IV.4.1, as an example of the use of formulae in the early

Buddhist tradition. In this sutta, two Brahmans named Vseha and Bhradvja have

a dispute over whose teaching on how to attain union with Brahm is correct.

Vseha argues that his teacher Pokkharastis is correct, while Bhradvja argues

that his teacher Trukkhas is correct. Since they are unable to settle the dispute

between themselves, they go to the Buddha, who naturally tells them that neither of

them is correct, that only he (the Buddha) knows the way to Brahm. After ridiculing

the claims of the Brahmans through a number of similes, the Buddha teaches them the

938
Other parallels can be found at E 40.1 and in an independent Chinese translation at T. 10.
939
On the other hand, the Sela Sutta is found close to the Vseha Sutta in both the Majjhima
Nikya (92) and the Sutta Nipta (3.6), but it does have parallels outside of the Theravda tradition
(e.g., E 49.6). Unlike the Vseha Sutta, however, the Sela Sutta is composed of more prose than
verse, and thus feels less at home in the Sutta Nipta than the former.

605
way to Brahm himself using the Tathgata arises formula. Vseha and

Bhradvja then take refuge in the Buddha and become his lay followers.

In the Vseha Sutta, Vseha and Bhradvja are still ordinary Brahmans

quarreling with one another, but here there is a slight difference: Vseha has clearly

begun to embark on a way of thinking that puts him in line with the Buddha. The

nature of their argument here is this: Bhradvja argues that one becomes a Brahman

by birth, while Vseha argues that one becomes a Brahman because of ones

conduct. Again, they go to the Buddha to settle their dispute, but this time the

Buddha sides with Vseha. This is not suprising, since as we have seen, the

argument that Brahmanhood is based on conduct and not, as the proponents of the

new Brahmanism claimed, on birth is a recurrent theme of stras throughout the

encounter dialog genre. As already mentioned, however, most of the Buddhas

response is given in verses, rather than prose. The Buddha argues against

Bhradvjas position by first pointing out that there is nothing to distinguish one

vara from another as between different species of animals, then pointing out that

Brahmans take up a variety of trades and should be known by those trades rather than

the honorific Brahman, and finally delivering a long series of verses extolling the

virtues of the true Brahman. In the end, Vseha and Bhradvja once again take

refuge in the Buddha and become his lay disciples.

In the Aggaa Sutta, the situation is drastically different. Vseha and

Bhradvja are now living with the sagha, hoping to eventually ordain as monks,

and they are no longer quarreling with one another, but simply go to the Buddha to

606
receive a teaching from him. When they do, the Buddha asks them if their fellow

Brahmans ridicule them for associating with the Buddhist sagha and hoping to

become monks themselves, and they admit that this is indeed the case. The Buddha

then delivers yet another sermon on the absurdity of the vara system and the new

Brahmanical claim to superiority on the basis of birth, culminating in the well-known

cosmogony that was dubbed a Buddhist Book of Genesis by Rhys Davids, 940 but

has more recently been convincingly shown to be a parody of Brahmanical

cosmogonies by Richard Gombrich. 941

What conclusions, if any, can we come to about the historical relationship

between these three suttas? Although the Vseha Sutta is found only in the

Theravda tradition, it is useful to begin by looking at it in conjunction with the

Tevijja Sutta. Although there are important differences between these two suttas

most importantly, that the Vseha Sutta is a mostly verse composition that seems

quite at home in the Sutta Niptathere also are some telling similarities between the

two of them, in particular in their beginnings, which are in both cases composed in

prose. In both of these suttas, as we have already seen, the scenario is the same:

Vseha and Bhradvja get into an argument, and since they are unable to settle the

argument between themselves, they decide to go to the Buddha to arbitrate. There are

other, more formal formulaic similarities as well. In both cases, the nidna states that

the Buddha was residing at a Brahman village at the time that the two Brahmans
940
This oft-quoted comparison of the Aggaa Sutta to the first book of the Bible is found as
the title of Rhys Davids translation of the sutta in Rhys Davids, Dialogues of the Buddha, 77.
941
Gombrich, The Buddhas Book of Genesis?

607
came to see himManaskaa in the case of the Tevijja Sutta and Icchnakala in the

case of the Vseha Sutta. 942 It is then stated, using identical, formulaic language,

that several famous Brahmans were in town, including Cak, Trukkha, Pokkharasti,

Jussoi, and Todeyya. Vseha and Bhradvja and the nature of the dispute

between them are then introduced, and finally Vseha persuades Bhradvja that

they should go to the Buddha for arbitration, using the fame of Gotama formula to

introduce the Buddha to his friend.

Given these close similarities, it is hard not to suspect that one of these suttas

used the introduction of the other as a template. The question, of course, is which

sutta copied which? I believe that we can conclude with some confidence that the

Vseha Sutta was created in imitation of the Tevijja Sutta. To begin with, we know

that the Tevijja Sutta is of some antiquity in the early Buddhist tradition because it is

found in the Slakkhandha/laskandha collection, which is the only group of stras

found together in all three extant versions of the Dgha Nikya/Drghgama, and thus

likely formed the earliest core of that collection. The Vseha Sutta, on the other

hand, has no parallels outside of the Theravda tradition, and thus has no evidence of

having any great antiquity in the early Buddhist tradition at all.

This is not only, it should be added, an argument from silence. Although the

Vseha Sutta has no non-Theravda parallel as a whole, a substantial portion of its

verses do have parallels in other sectarian traditions. These verses are those in which

942
Strangely, however, in the case of the Vseha Sutta, it is not explicitly stated that
Icchnakala is a Brahman village, even though we know this from numerous other suttas (see Section
IV.4.2 above).

608
the Buddha extols the virtues of the true Brahman, using the common refrain (in Pali),

tam aha brmi brhmaaHim I call a Brahman. These verses are also found

in the Brhmaa Vagga of the Dhammapada, which in turn has parallels in all of the

other sectarian versions of the Dhammapada, including the Gndhr Dharmapada,

which likely belongs to the Dharmaguptaka school; the Patna Dhammapada, which

may have come from the Mahsghika school; and the Sanskrit Udna Varga (in its

several versions), which likely belongs to the (Mla-)Sarvstivda school. Although

all four of these versions have a section devoted to Brahmans, the contents of this

section are to varying degrees different in each of the versions. The Patna (possibly

Mahsghika) version is the most different of all; in particular, out of the 36 verses

ending in tam aha brmi brhmaa that are found in Pali Dhammapada and also

the Vseha Sutta, only four have a parallel in the Patna Dhammapada. The

Gndhr Dharmapada and Sanskrit Udna Varga, on the other hand, do contain

parallels to most of these 36 verses, but they preserve them in very different orders. 943

This, together with the fact that only the Theravdins have a Vseha Sutta, suggests

that these verses had their origin within the relatively free-floating verse tradition

lying behind the various versions of the Dhammapada/Dharmapada/Udna Varga,

and only much later were borrowed by the Theravdins, in exactly the same order in

which they had been preserved in the Theravda Dhammapada, to construct the

Vseha Sutta. To these verses, which because of their great number dominate even

943
nandajoti Bhikkhu, A Comparative Study of the Dhammapada (unpublished
manuscript), 20-21.

609
the fully completed sutta, were added a few more verses, as well as a prose

introduction modeled after the one found in the Tevijja Sutta.

Indeed, we can see internal signs that the prose introduction of the Vseha

Sutta is derivative of the introduction of the Tevijja Sutta. The nidna of the latter

sutta places the location in Manaskaa, a Brahman village that is completely unique

to this sutta and not mentioned anywhere else in the Pali Canon. The Vseha Sutta,

apparently in imitation of this nidna, also gives a Brahman village as the setting, but

in this case uses the more commonly recurring location of Icchnakala. Then

follows the formulaic statement that a number of famous Brahmans were staying in

the area at the time. This list of famous Brahmans is found only in three places in the

Pali Canonin these two suttas, and in the Subha Sutta (MN 99). In the context of

the Tevijja Sutta, there is a clear narrative purpose to giving this list of prominent

Brahmans staying in the area. As we find out immediately thereafter, Vseha is a

student of Pokkharasti, and Bhradvja is a student of Trukkha, and this is the

source of their conflict. In the Vseha Sutta, however, their conflict is totally

unrelated to who their teachers are (indeed, Vseha takes up a position in line with

the Buddha), and thus the mention of the fact that there are many famous Brahmans

in town is superfluous. As a result, we can suspect that the list of the famous

Brahmans was simply copied by the Vseha Sutta from the Tevijja Sutta, in spite of

the fact that it no longer had a narrative purpose in its new setting.

Although much of this evidence is circumstantial, taken together it strongly

suggests that the Vseha Sutta was created, either by the Theravdins or by their

610
immediate genealogical forebears, in imitation of the Tevijja Sutta. But what of the

Aggaa Sutta? Much of the scholarly debate over the history of this sutta has

centered on the question of whether or not the cosmogony found in this sutta is

original to it or was added to an earlier, much shorter version in which the Buddha

simply talked about the absurdity of new Brahmanical claims about vara. 944

Regardless of which side of this debate is correct (which is irrelevant since the

characters Vseha and Bhradvja only appear in the frame story), I believe that we

must conclude that the Aggaa Sutta is also derivative of the Tevijja Sutta, even

though it, unlike the Vseha Sutta, does have parallels in other, non-Theravda

sectarian traditions. To begin with, the Aggaa Sutta, unlike the Tevijja Sutta, is not

found in the stable Slakkhandha/laskandha collection; in fact, the Theravda and

Sarvstivda traditions do not even agree on whether it belongs to the Dgha/Drgha

or Majjhima/Madhyama collection. Moreover, the Aggaa Sutta appears to assume

that Vseha and Bhradvja are already known characters, and in fact appears to

continue their story by depicting them as now (having taken greater interest in the

Buddhas dharma after their earlier encounter with him) living with the sagha in

preparation to become monks themselves. That this is a sequel to the Tevijja Sutta,

rather than the Tevijja Sutta being a prequel to the Aggaa Sutta, is confirmed by the

fact that, while Vseha and Bhradvja both have a clear narrative purpose in the

944
The most extensive study, taking into account both Pali and Chinese versions, that argues
for the non-originality of the cosmogony in this sutta, is Konrad Meisig, Das Stra von den vier
Stnden: Das Aggaa-Sutta im Licht seiner chinesischen Parallelen (Wiesbaden: Otto Harrassowitz,
1988). Contra this method of dividing the Aggaa Sutta into strata, see Steven Collins, The
Discourse on What is Primary (Aggaa-Sutta): An Annotated Translation, Journal of Indian
Philosophy 21(1993): 301-393.

611
Tevijja Suttatheir argument at the beginning is what sparks the entire narrative

Bhradvja is completely superfluous in the Aggaa Sutta. Once Vseha and

Bhradvja approach the Buddha, Bhradvja completely drops out from the

narrative, and in fact the Buddha directs all of his comments directly to Vseha. It

appears that Bhradvja is only mentioned to begin with because he is known as the

companion of Vseha from the earlier Tevijja Sutta.

Taken together, then, these three suttas form a sort of mini-series featuring

the characters Vseha and Bhradvja in the early Buddhist tradition. The series

began in the relatively old Slakkhandha/laskandha collection, in which Vseha

and Bhradvja appeared as a unique pair of interlocutors to provide yet another

occasion for the Buddha to deliver his teaching on the training using the Tathgata

arises formula. Later, but still fairly early on in the early Buddhist tradition, these

characters were borrowed to construct a new encounter dialog, the Aggaa Sutta,

which continued the story of the two monks by showing them preparing to ordain as

monks. Finally, in a move that was apparently isolated within the Theravda tradition,

a group of verses on the qualities of a true Brahman was expanded and transformed

into an encounter dialog by adding a prose introduction modeled on the Tevijja Sutta,

and making use of its main characters Vseha and Bhradvja.

IV.4.7 The Stock Brahman Jussoi

Perhaps the best-known Brahman in the entirety of the early Buddhist

tradition is Jussoi. Jussoi appears as an interlocutor in so many suttas of the

612
Pali Canon that he can be regarded as a stock character. It is not entirely surprising,

however, that the name Jussoi could have come to be used this way in the early

Buddhist tradition. According to Buddhaghosa, Jussoi was not the name created

by his parents, but rather was received from another as a title; that is, the title

jussoi is the title purohita, which was given to him by the king, and therefore he is

called Jussoi. 945 It is therefore not clear whether Jussoi can even be regarded

as a recurring character, or if rather the stock character of Jussoi simply referred

to a generic royal official who may or may not be the same actual person from one

instance to the next. Indeed, unlike in the case of Vseha and Bhradvja, there is

not really even a semblance of continuity from one Jussoi sutta to the next;

instead, each sutta is for the most part a self-contained narrative, usually fairly simple

in structure, in which the Brahman Jussoi appears as a stock character.

Overall, a Brahman Jussoi appears in 14 suttas in the Pali Canon. 946

These can be found listed in Table 19. Looking at this table, we can see several

trends. First, within the Theravda tradition at least, most suttas involving Jussoi

are found in the Aguttara Nikya. Only three are found in the Majjhima Nikya and

two in the Sayutta Nikya. In addition, several of the Jussoi suttas have no

known parallel in a non-Theravda tradition. This is not entirely surprising given the

distribution of suttas within the Theravda tradition; as we saw in Chapters IV.2 and

945
See Buddhaghosas commentary on MN 4: jussoti neta tassa mtpithi katanma,
apica kho hnantarapailbhaladdha. jussoihna kira nmeta purohitahna, ta tassa
ra dinna, tasm jnussoti vuccati.
946
I am not including in this number MN 98=Sn. 3.9, in which the name Jussoi appears in
the list of famous Brahmans, already discussed above in Section IV.4.6, but the character makes no
actual appearance himself.

613
IV.3, the Ekottarikgama parallel to the Aguttara Nikya is quite different from the

latter, and it appears from comparison to the Sarvstivda Madhyamgama that the

Theravdins padded their Aguttara with material that other traditions placed

elsewhere at a relatively late stage in their tradition. Indeed, as we might expect

given these earlier conclusions, one of the Jussoi suttas that the Theravda

tradition places in its Aguttara is placed by the Sarvstivdins in their

Madhyamgama, and another is placed by the latter in their Sayuktgama. In

addition, there is fairly clear evidence that the appearance of Jussoi as an

interlocutor is in at least some cases a relatively late development in the tradition.

Three of our Jussoi suttas in the Theravda tradition do have parallels in other

sectarian traditions, but those parallels do not feature Jussoi as an interlocutor.

Pali Sutta Description Chinese Parallels

MN 4 Jussoi asks Buddha E 31.1
about living in the forest
MN 27 Vacchyana praises M 146
Buddha to Jussoi, and
Jussoi goes to Buddha
and hears about elephants
footprint (in form of
Tathgata arises formula)
MN 99 Jussoi appears at the The ending involving
very of the sutta, when Jussoi is not found in
Subha praises the Buddha the Chinese parallel, M
to him, using the same 152
formulaic language
employed by Vacchyana
in MN 27
SN 2.1.5.7 Jussoi asks Buddha
questions similar to those
of Lokyata Brahmans
SN 5.1.1.4 nanda sees Jussois S 769
splendid chariot and asks
Buddha about metaphorical

614
 chariot
AN 2.17 Jussoi asks Buddha
about why people go to hell
AN 3.55 Jussoi asks Buddha
about nibbna
AN 3.59 Jussoi says that gifts partial parallel to S 884-
should go to tevijja 5, which do not involve
Brahmans, and Buddha any Brahman interlocutor
refutes this
AN 4.184 Jussoi says that all fear
death, and Buddha refutes
this
AN 6.52 Jussoi asks about the M 149, E 37.8
aims, etc., of different
classes of people
AN 7.47 Jussoi asks Buddha if E 37.9
he practices brahmacariya
AN 10.119 Buddha sees Jussoi
about to perform the
descent ceremony and
gives his own version
AN 10.167 similar to AN 10.119 S 1040, in which the
interlocutor is a Brahman,
but not named Jussoi
AN 10.177 Jussoi asks about S 1041
rddha rite, and Buddha
gives his own version
Table 17. A list of all suttas in the Pali Canon in which Jussoi appears as a character, with descriptions and
Chinese parallels.

We cannot conclude on the basis of this evidence, however, that the entire

character of Jussoi is simply a late intrusion into the tradition. As can be seen

from Table 19, there are three Jussoi suttas that have parallels in the Chinese

Ekottarikgama, which, as discussed in Chapter IV.2, may belong to the rather distant

(i.e., from the Theravda) Mahsghika school, and in any case should be seen as a

significant comparative witness to the early Aguttara/Ekottarika tradition given its

great difference from the Pali Aguttara Nikya. That is to say, there are so few

615
stras held in common between the Aguttara Nikya and Ekottarikgama that the

few that are should be held as an important witness to the early tradition of that

collection. Of the three Jussoi suttas that do have parallels in the

Ekottarikgama, two are indeed placed by the Theravdins in their Aguttara Nikya,

while the third is placed by them in their Majjhimaalthough the fact that it is not

found in the Sarvstivda Madhyamgama suggests that the Majjhima classification

is not of great antiquity. All three of the suttas in question have a similar structure,

which is fairly simple and can be characterized as falling into the category of

nominal encounter dialog: The Brahman Jussoi goes to the Buddha, asks him a

question, and this serves as an opportunity for the Buddha to give a sermon. These

sermons are on fairly general topics (meditating in the forest, the aims of various

types of people, and brahmacariya) and do not serve to address the ideological claims

of the new Brahmanism or set up a dichotomy between the Buddha and his

interlocutor as his Brahman other. Instead, the Brahman Jussoi simply appears,

in his capacity as a royal minister, as an important person conversing with the Buddha,

much as the king himself does in various other contexts.

A few of the Jussoi suttas do take up more substantive encounter dialog

themes that highlight a dichotomy between Buddhist identity and Brahmanical

identity, addressing such topics as the superiority of tevijja Brahmans, the

Brahmanical descent ceremony, and the rddha rites. Only one of these suttas,

however, has any known parallel in a non-Theravda sectarian tradition. All of the

other Jussoi suttas, whether they have non-Theravda parallels or not, are

616
nominal encounter dialogs in the sense that the appearance of Jussoi (a

Brahman) makes them encounter dialogs, but in terms of their content they do not

address substantive encounter dialog themes. This, taken together with the fact that

the character (or characters) of Jussoi do appear to have a fair amount of antiquity

in the tradition, suggests that Jussoi was originally a stock character significant

not so much for his Brahmanhood, but rather his status as a government minister, thus

demonstrating the Buddhas familiarity with persons of high social status. Only later,

and only in a few instances, was this stock character adopted for use in the more

pointed endeavor of establishing a Buddhist identity vis--vis the Brahmanical

identity projected by the proponents of the new Brahmanism.

617
 Chapter IV.5

Conclusion to Part IV

In Part IV, we have looked at a vast swath of evidence from across the entire

extant early Buddhist tradition, including multiple Nikya/gama traditions and

multiple sectarian traditions, in order to better understand the development of the

encounter dialog genre. We began in Chapter IV.1 by affirming the applicability of

the Parry-Lord oral theory to the early Buddhist tradition, which enables, and requires,

us to understand multiple versions of a particular stra not as different recensions of a

single text, but rather as different performances of a particular story or teaching

within a living oral tradition. Then, in Chapter IV.2, we looked at the early Buddhist

tradition itself, comparing the structures of the Nikya/gamas between the

Theravda tradition preserved in Pali and other sectarian traditions, most notably the

Sarvstivda, preserved in Chinese.

This investigation yielded a couple of interesting findings. First, there appears

to be a correlation between the genealogical closeness of sects and the degree of

similarity in the structure of their Nikya/gama traditions. This can be seen most

618
clearly in the case of the Dgha Nikya/Drghgama, for which we have three

sectarian versions. As we saw, the versions of the Dharmaguptaka and Theravda,

which are closely related as branches of the Vibhajyavda, are quite similar to one

another in terms of content, but fairly different from the version belonging to the

Sarvstivda, which, being non-Vibhajyavda, is less closely related to either of the

former than they are to each other. Consistent with this finding, the Sarvstivdin

Madhyamgama has major differences in content from the Theravdin Majjhima

Nikya, while still holding a significant number of stras in common. On the other

hand, the Chinese Ekottarikgama, which may belong to the Mahsghikathat is,

the most genealogically removed sect from the Theravda possibleis radically

different from the corresponding Aguttara Nikya, bearing only a small percentage

of its stras in common.

Second, we considered the theory of Ynshn, based on the testimony of the

Vastusagraha, that the Sayutta Nikya/Sayuktgama served as the basis of all

of the Nikya/gamas, and we found some evidence to support this hypothesis.

Although, as just mentioned, there are substantial differences between the Theravda

and Sarvstivda traditions, the Sarvstivdin Sayuktgama is more structurally

similar to its Theravda counterpart than either the Drghgama or the

Madhyamgama is. At the same time, although the Sarvstivdin Drghgama is

overall not very similar in structure and content to its counterpart the Dgha Nikya, it

does share a similar component, namely the laskandha Nipta/Slakkhandha Vagga,

which, as Bhikkhu Sujto has suggested, has a sayutta-like theme (the training),

619
and may have been extracted from the Sayutta Nikya/Sayuktgama to become

the basis of a new collection of long stras. The Sarvstivdin Madhyamgama,

on the other hand, contains several vargas that actually contain the word sayukta

(Ch. ) in their titles, suggesting that they, as well, may have been borrowed from

an original Sayutta/Sayukta collection. Indeed, many of the stras in the

Sarvstivdin Madhyamgama, especially in those vargas with sayukta in the title,

have their Pali parallels not in the Majjhima Nikya, but in the Aguttara Nikya. I

have hypothesized that these stras may have originally been found in the

Sayutta/Sayukta collection, but were moved elsewhere, albeit differently, by the

Sarvstivdins and Theravdinsby the former primarily to expand their

Madhyamgama, but by the latter to fill out the various numerical categories of the

Aguttara Nikya.

In Chapter IV.3, we finally turned to the encounter dialog itself by looking for

examples of collections of encounter dialogs and tracing their development in the

context of the broader encounter dialog tradition. This approach proved quite fruitful

in the case of the Sayutta Nikya/Sayuktgama and the Dgha

Nikya/Drghgama, where we found collections of ecounter dialogs that appear to

have some antiquity in the tradition. The Brhmaa Sayutta/Sayukta, while

organized differently in the Theravda and Sarvstivda traditions, nonetheless shares

a large number of stras in common between the two traditions, and in particular

appears to have a core of stras that share the common characteristics of framing

verses with short prose narratives in which the Buddha encounters a Bhradvja

620
Brahman with a humorous, ironic, or otherwise meaningful personal name. The

Dgha Nikya/Drghgama, on the other hand, in spite of other differences in

structure and content, contains in all three sectarian versions that have come down to

us a collection of stras which all make use of the Tathgata arises formulamany

of which also happen to be encounter dialogs. These encounter dialogs, especially

the Ambaha, Soadaa, and Kadanta Suttas, also share a number of

characteristics, including funny/meaningful names for their Brahman interlocutors,

plot motifs, and the use of particular formulae. Given these commonalities and the

stability of the collection across sectarian traditions, the Slakkhandha/laskandha

collection would therefore appear to be of some antiquity in the early Buddhist

tradition as well.

In the case of the Majjhima Nikya/Madhyamgama and Aguttara

Nikya/Ekottarikgama traditions, however, this approach proved less fruitful. Both

the Theravda Majjhima Nikya and the Sarvstivda Madhyamgama contain a

Brhmaa Vagga/Varga, but these two collections share very few stras in common

and therefore, I argued, appear to have been created independently. The Aguttara

Nikya also contains a few Brhmaa Vaggas, but these vaggas appear to have been

so named primarily as a matter of convenience, as they do not exclusively contain

suttas dealing with Brahmans, and in any case they have no parallel in the Chinese

Ekottarikgama, which, as already mentioned, is radically different overall from its

Theravda counterpart. Therefore, in Chapter IV.4, we turned to a different approach

to analyzing the encounter dialog genre, namely by exploring the use of various

621
formulae and themes characteristic of the genre across the Nikya/gama traditions.

By tracing where these formulae are and are not found, we were able to uncover clues

to the development of the encounter dialog genre as a whole. The naming of

brhmaagmas as the nidnas of suttas, we found, is scattered across the early

Buddhist tradition, and in fact is not even restricted to the encounter dialog genre.

This finding is consistent with von Hinbers argument, on the basis of internal

linguistic evidence, that nidnas placing the Buddha in a brhmaagma are

relatively early and predate nidnas placing him in large cities such as Svatth and

Rjagaha, which represent a later, more urbanized stage in the tradition.

Other formulae and themes, however, display patterns that point to a later

stage in the development of the tradition. The fame of Gotama formula is found

only once in the entire Sayutta Nikya, but it is found in every single encounter

dialog in the Slakkhandha Vagga of the Dgha Nikya. For this reason, I argued that

this particular formula did not have its origin in the Sayutta Nikya, and may have

instead had its origin in the Slakkhandha Vagga. Likewise, the Triple Veda

formula is found scattered throughout the Dgha, Majjhima, and Aguttara Nikyas,

but it is not found even once in the Sayutta Nikya. This may also indicate, if

Ynshns hypothesis is correct, that the Triple Veda formula only developed in the

tradition after the other three Nikya/gamas separated out from the Sayutta

Nikya/Sayuktgama. Both of these formulae are relied upon heavily by the

encounter dialogs of the Slakkhandha/laskandha collection in the Dgha/Drgha

tradition, which, unlike the encounter dialogs in the Brhmaa Sayutta/Sayukta,

622
display a highly developed formulaic structure. And although the encounter dialogs

of the Slakkhandha/laskandha collection clearly are of a certain antiquity because

the collection as a whole is found in all three extant sectarian traditions, and they even

make use of the apparently antique brhmaagma nidna, they also refer to

brahmadeyya land grants, for which we do not have evidence in the epigraphical

record prior to the first century BCE. Although it is impossible to date these

encounter dialogs exactly on the basis of this evidence (or rather lack thereof), the

lack of firm independent evidence for early brahmadeyya land grants does not inspire

confidence in the antiquity of stras that make reference to them. Therefore, taking

this together with the highly developed use of formulae that for the most part do not

appear in the Sayutta Nikya, I have argued that the Slakkhandha encounter dialogs,

albeit of a certain antiquity in the tradition, most likely developed later than the early

collection of encounter dialogs that formed the basis of the Brhmaa

Sayutta/Sayukta.

Similarly, we found that what would become one of the most well-known and

influential themes relating to Brahmansnamely, the theme of searching for the 32

marks on the Buddhas bodyis actually found in only a small number of stras in

the early Buddhist tradition. Only three suttas in the Pali Canon, in fact, make use of

it, and there are good reasons on the basis of formula analysis, as well, to believe that

this theme was a relatively late development in the tradition. To begin with, the

fame of Gotama formula makes no mention of the 32 marks, which is surprising

within the context of suttas that use it to construct the theme of searching for the 32

623
marks, since it is by hearing this very formula that the Brahman in question is

inspired to visit the Buddha and see if he has the marks. Indeed, as we saw, a new

formulathe two paths of a Great Manmust be introduced so that the Brahman

in question can explain, either to his student or to himself, that it is by the presence or

absence of the 32 marks that one can know whether the report being spread about the

Buddha is earned. And while the learned Brahman, in the Theravda tradition,

makes reference to the 32 marks, both in the context of the theme of the 32 marks and

elsewhere, the Sarvstivda version of the same formula does not, suggesting that the

32 marks are not original to the formula. Finally, the entire premise of the theme of

the 32 marks, which is that discerning the 32 marks was a skill possessed and valued

by Brahmans, finds no substantive corroborating evidence in the actual Brahmanical

tradition until Varha Mihiras 6th century Bhat Sahit. Therefore, it seems likely

that this theme was a fairly late development in the early Buddhist tradition.

Finally, at the end of Chapter IV.4, we examined two examples of stock

charactersnamely, a set of three suttas referring to the Brahman friends Vseha

and Bhradvja, and the frequently recurring Brahman character Jussoi. We

found that the characters Vseha and Bhradvja likely had their origin in the

Tevijja Sutta, one of the encounter dialogs of the Slakkhandha Vagga. From there,

the Aggaa Sutta, found in the Dharmaguptaka, Theravda, and Sarvstivda

traditions, continued the story of Vseha and Bhradvja after they began living

with the Buddhist sagha, while the Theravda tradition alone created the Vseha

Sutta by adding a prose introduction to a set of verses borrowed from the

624
Dhammapada tradition. The character of Jussoi, on the other hand, appears to

have some antiquity in the early Buddhist tradition, insofar as he is found in a couple

stras shared by the Aguttara Nikya and the Ekottarikgama. Nevertheless,

several Pali suttas that feature Jussoi as a character have no parallel in any extant

non-Theravda tradition, indicating that some of the corpus of Jussoi suttas

represents a relatively late development within the narrow context of a particular

sectarian tradition. Moreover, only a few Jussoi suttas actually deal with

substantive encounter dialog themes; most simply use Jussoi as an occasion for

the Buddha to deliver a teaching. Given that Jussoi was apparently not

originally a given name, but a title for a government minister, it is possible that he

began as a stock character significant not so much as a Brahman, but as an occasion

for showing the Buddha hobnobbing with important members of society.

What is the overall picture of the development of the encounter dialog genre

that emerges from our examination in Part IV? It appears that the genre had its

origins in a collection of stras that were constructed by adding short prose narratives

involving Bhradvja Brahmans with funny/meaningful given names to verses and

that this collection became the basis of the Brhmaa Sayutta/Sayukta of the

Sayutta Nikya/Sayuktgama. Later, a collection of longer strasabout half of

them encounter dialogswas created on the basis of the Tathgata arises formula.

This collection may have had its origins in the Sayutta Nikya/Sayuktgama in a

sayutta/sayukta on the Buddhist training, but it ultimately became the basis for

the Dgha Nikya/Drghgama collection. The encounter dialogs of the

625
Slakkhanda/laskandha collection likely gave birth to the fame of Gotama

formula, and certain other formulae and themes, from which they were borrowed to

construct encounter dialogs elsewhere. Some of the developments within the

encounter dialogs of the Slakkhandha/laskandha collection are likely not of great

antiquity. Several encounter dialogs therein refer to brahmadeyya land grants, for

which we do not have independent evidence until the 1st century BCE, and two make

use of the theme of the 32 marks, for which we do not have clear independent

evidence until the 6th century CE. While it is not likely (and in the case of the 6th

century date for the Bhat Sahit, simply impossible) that the encounter dialogs in

question were created as late as the relevant independent evidence, the lateness of the

latter makes it unlikely that the former are of great antiquity. Other encounter dialogs

are scattered across the Nikya/gama traditions; since they are not organized

according to structures found consistently across the various sectarian traditions, it

appears that they belonged to an amorphous body of oral tradition that did not

become well organized until after the various sectarian traditions separated. In

particular, the Theravda tradition appears to have had a predilection for using such

suttas to fill out the numerical categories of its Aguttara Nikya, while the

Sarvstivda tradition preferred to simply expand its Madhyamgama to

accommodate these stras. Thus, the overall trend that our examination suggests is

from a very simple set of encounters between the Buddha and Bhradvja Brahmans

in the Brhmaa Sayutta/Sayukta of the Sayutta/Sayukta, to a collection of

more highly developed and innovative narratives in the Slakkhandha/laskandha

626
collection of the Dgha/Drgha, to a broad diffusion of Brahman-related formulae and

themes throughout the Dgha/Drgha, Majjhima/Madhyama, and

Aguttara/Ekottarikgama traditions.

627
 Part V

Conclusion

628
 At the beginning of this dissertation, I juxtaposed two Buddhist texts that both

refer to the category Brahman, but relate to it in very different ways. In the first,

the Pryaa, which, as we saw in Chapter III.1, is likely one of the earliest Buddhist

texts extant, a man named Dhotaka approaches the Buddha and asks him to teach him,

referring to the Buddha in the process as a Brahman. In the second, an account of

the Chinese monk Zh Shxng going to Khotan in search of the

Pacaviatishasrik, Zh Shxng is initially unable to leave the city with the

Mahyna text because the Hnaynists in the city, calling it a Brahman book, do

not want its heretical teachings to spread to China. In light of these two storiesin

one of which the word Brahman is an honorific, in the other of which it is a sneer

I asked, How do we get between these two very different, indeed diametrically

opposed, Buddhist uses of the word Brahman?

We are now in a position to answer this question. Early on in the Buddhist

tradition, many different honorifics were used to describe the Buddha and the ideal

that he exemplifiedbuddha itself, of course; but also arhat, nga, jina, muni, and

even brhmaa. These honorifics did not belong to any one particular sectarian

group, but rather to the common Indo-Aryan substratum from which many groups,

including the Buddhists, sprang. As these various sects coalesced and formed

identities in contradistinction to one another, it became inevitable that terms such as

brhmaa would become contested. In particular, the Buddhists were forced to

contest the meaning of the term brhmaa with the proponents of the new

BrahmanismVedic specialists who had a tradition of calling themselves

629
Brahmans due to their conflation of their own Vedic texts with the old concept of

sacred speech (brahman), but who increasingly arrogated to themselves exclusive

ownership of the term on the basis of their birth. Due to the latters increasing

belligerence in arrogating the category Brahman to themselves, the Buddhists

eventually were forced to cede the category to them, and thus the category Brahman

became one by which the Buddhists no longer defined themselves, but rather against

which they defined themselves. It became the quintessential other to the Buddhist

monk, such that, in the 3rd century CE, the Hnaynists of Khotan could denigrate a

Mahyna text as un-Buddhist simply by calling it Brahmanical.

In answering the question in this way, I am taking a different approach than

the one that has typically been taken in Buddhist Studies to describing the

relationship between Buddhism and Brahmanism. Most scholars have assumed that

there was a natural antagonism between the Buddhists and the Brahmans from

the very beginning of Buddhism, and they have interpreted all references to the

category Brahman in the Buddhist literature as reflective of this antagonism. Under

this paradigm, any seemingly positive use of the category Brahmansuch as

Dhotaka referring to the Buddha as a Brahmanmust in fact be polemical. As we

saw in Chapter III.1, this was definitely the case in some uses of the word Brahman

in the early Buddhist tradition, but I have argued that the earliest uses of the word

Brahman in that tradition are best understood as simple honorifics, and that only

through time did they come to be interpreted as polemics against the literal

Brahmansi.e., Brahmans who claimed to be so by birth.

630
 My approach also differs from that of Bronkhorst, who in contradistinction

with most scholars specifically understands Buddhism as having arisen in a non-

Brahmanical context, and thus all references to Brahmans and Brahmanism in the

early Buddhist texts as late. The reason Bronkhorst takes this approach is that he

posits a radical dichotomy in early India between two cultural spheresthe

Vedic/Brahmanical culture of the Western Kuru-Pacla region and the non-Vedic

culture of the Eastern Greater Magadha region. I do not find this argument entirely

convincing, in large part because of the utter pervasiveness of references to Brahmans

(whether literal, metaphorical, or otherwise) and Brahmanical ideas in the early

Buddhist texts. However we are to understand the origins of Buddhism, we certainly

do not have the luxury of sequestering it from Brahmanism broadly construed, nor of

investigating early Buddhism without formulating a clear methodological approach

for understanding its relationship to the Brahmanical tradition.

Nevertheless, I am sympathetic to Bronkhorsts argument that much of the

material dealing with Brahmans in the early Buddhist texts may be late, and that

Buddhists were eventually forced to contend with Brahmanism because (certain)

Brahmans sought to expand their base of patronage by belligerently asserting their

claims to superiority on the basis of birth and by rewriting history in order to

Brahmanize it. Bronkhorst refers to this particular agenda, and the movement

associated with it, as the new Brahmanism, language that I have adopted in this

dissertation as well. My methodological approach consists of adapting Bronkhorsts

insights to the theoretical framework offered by Ruegg for understanding the

631
relationship between Buddhism and Hinduism/Brahmanism, particularly through the

concept of substratal influences. Rather than assuming that common elements

shared between Buddhism and Brahmanism, such as the word Brahman, were

borrowed by one from the other, I argue that they could have been borrowed from a

common substratumwhat I call the pan-Indo-Aryan substratum that was shared

by all of the sectarian traditions that arose in ancient North India, including

Brahmanism and those of Greater Magadha. Concomitantly, I argue that these

sectarian traditions cannot be understood as essentialized, metahistorical agents, such

that one could arise purely in reaction to another. Rather, they must be

understandood as fluid, constantly interacting entities that emerged out of a common

substratum and only coalesced as discernible sects through a long process of identity-

formation, wherein terms such as Brahman were hotly contested between different

groupsin this case, the early Buddhists and the proponents of the new Brahmanism.

In Part II of this dissertation, I illustrated the fluidity of sectarian identities in

early India by exploring early Indian articulations of identity in Buddhist and

Brahmanical texts. I began in Chapter II.1 by showing that the the categories usually

cited as evidence of the antagonistic relationship toward Brahmanism in which

Buddhism arosenamely, ramaa and brhmaaare usually not used in the early

Buddhist texts in opposition to one another at all. Although, based on evidence from

Megasthenes and Patajali, the idea that the ramaa and the brhmaa are opposed

to one another clearly did arise by around the time of Candragupta Maurya, the early

Buddhist texts seem largely unaware of this opposition, and almost always treat

632
samaabrhma together as a group. In most cases, they treat them negatively as a

foil against which to construct Buddhist identity, but in some cases, they treat them

positively to refer to a unified class of individuals who are worthy of respect and

offerings. Although the early Buddhist texts do sometimes split samaabrhma

into various sub-groups, they do so according to criteria that are separate from, and

indeed cut across, any putative samaa-brhmaa divide.

The use of the terms ramaa and brhmaa together is common in the early

Buddhist texts, and it appears that this is a reflection of colloquial usage, given that it

is also found in the inscriptions of Aoka and Greek evidence. Interestingly, however,

it is never found in Brahmanical texts. I argued that this was because the authors of

the relevant Brahmanical literature, the Dharma Stras, eschewed these colloquial

categories and instead developed a more formal theoretical schemethe rama

systemto describe various lifestyles seen in India in the late first millennium BCE.

In Chapter II.2, I argued that the authors of the Dharma Stras were proponents of

what Bronkhorst calls the new Brahmanismthat is, householder-supremacist

Vedic scholars who claimed exclusive prerogative to the title Brahman on the basis

of birth. As I showed, however, their claims to purity of birth were more rhetorical

than real. There is ample evidence in both Buddhist and Brahmanical texts that Vedic

teachers were not consistent in accepting only students of pure Brahman lineages as

students. Moreover, the theory of mixed classes presented in the Dharma Stras

clearly is not meant to explain what actually happens when mixed marriages occur;

rather, it is designed to explain and justify a low social status for various marginal

633
social groups by declaring them to be products of the improper mixture of different

varas.

Since another major tenet of the new Brahmanism was that one must have a

son to fulfill the injunctions of the Veda, the authors of the Dharma Stras used the

theoretical framework of the rama system to systematically describe the actual

lifestyles practiced in India at the time that they wrote, and then without exception

they criticized those that involved celibacy as inconsistent with the Vedas. Although

the rama system is not found outside of Brahmanical texts, I showed in Chapter II.

3 that many of the categories adopted by the rama system are also found in the

early Buddhist texts, although they are used there in a very different way. I argued

that this was not due to the Buddhists borrowing from the Brahmanical traditions;

rather, the categories held in common by the Brahmanical rama system and early

Buddhism were colloquial categories that both groups adopted and deployed in

different ways in order to advance their own articulations of self-identity. The

Buddhists, on the one hand, did not identify with, but nonetheless sought an alliance

with, the gahapatisthe patriarchal house-lords of ancient Indian societywhom

they depended on for recruits and material support. They naturally identified with the

old concept of brahmacarya because of their practice of celibacy and studentship

under an crya, but they tried to distance themselves from the category paribbjaka

in order to distinguish themselves from other, similar wandering mendicant groups.

Finally, they treated the jailasascetics who tended to a sacred fire in the forest and

maintained close ties to the old Vedic traditionas sympathetic figures, with whom

634
they understood themselves to hold more similarities than the other paribbjaka

groups. The new Brahmanical authors of the Dharma Stras, on the other hand,

identified themselves strictly as ghasthasas those who stay in the home, i.e., to

have children. They were critical of the lifelong practice of brahmacarya, the

practice of wandering mendicancy (parivrjaka), and the practice of living as a

hermit in the forest (vnaprastha) because all of three of these lifestyles involved

celibacy and therefore made it impossible to have children. Thus, both proponents of

the new Brahmanism and the early Buddhists made use of a common set of colloquial

categories in order to advance their own agendas of identity articulation.

By showing that the Buddhists and proponents of the new Brahmanism were,

in general, making use of a common stock of vocabulary to articulate their own

visions of their own and others identities in this way, I thus lay the groundwork for

making the argument that the category of Brahman itself was a substratal term that

was contested between the Buddhists and the proponents of the new Brahmanism,

rather than a properly Brahmanical term that the Buddhists simply borrowed for

the purposes of polemic. It was to this task that I turned in Part III. I began in

Chapter III.1 by exploring the different ways that the word brhmaa is used in the

early Buddhist texts. I showed that while, in many cases, the word is used in a

polemical sensethat is, in explicit contrast to those who claimed to be Brahmans on

the basis of birthit is used in certain texts, in particular the Ahaka Vagga and

Pryaa Vagga of the Sutta Nipta, as what appears to be nothing more than a

straightforward honorific for the Buddha or the ideal person. Because these texts are

635
likely among the oldest Buddhist texts extant, I argued that there must have been a

transition from the use of the word brhmaa as an honorific to its use as a polemic

against a Brahmanical other.

In Chapter III.2, I then showed how the transition from the use of brhmaa

as an honorific to its use as a polemic was accomplished in the Buddhist tradition

through commentary and narrative framing. I looked at the Niddesa commentary on

the Ahaka and the Pryaa and showed that it introduces concepts that are foreign

to the root texts and that, in particular, tend to interpret references to the ideal person

as a Brahman in those texts as metaphorical. Even more importantly, though,

narrative frames into which the verses of these texts have been insertedin the

Chinese version of the Ahaka and the Pali version of the Pryaaintroduce

Brahmanical characters, i.e., people who claim to be Brahmans by birth, where

none existed in the original verses. This implicitly set up a contrast, I argued,

between the ideal person as Brahman in the verses and the literal Brahman in the

frame narrative, thus transforming the former from simple honorifics into polemics

against those who claimed to be Brahmans by birth. Given this progression in the

traditionfrom the use of brhmaa as an honorific to its use as a polemic against a

Brahmanical otherI argued that the word brhmaa was not borrowed by the

Buddhists from a monolithic Brahmanical movement, but rather that it was taken

by both groups from a common substratum and then contested between them.

In order to mark a contrast between this sort of substratal borrowing and an

actual case of polemical borrowing, I then in Chapter III.3 looked at another term,

636
tevijja, that I argued did orginate specifically within the new Brahmanical movement

and was borrowed by the Buddhists to use polemically against them. This term is

used in the early Buddhist texts to refer to proponents of the new Brahmanism in their

capacity as Vedic scholars (i.e., because they know the three Vedas), but it is also

used to refer to the Buddha as having attained three knowledgesthe ability to see

his own past lives, the ability to see the karmic destinies of others, and knowledge of

the end of the savas. By exploring the development of Buddhist formulae for

describing the path to Awakening, in which this latter, metaphorical use of the term

tevijja occurs, I showed that it was a relatively late development in the tradition and

was not used in the earliest texts. In addition, because the corresponding Sanskrit

word traividya is a post-Vedic neologism, as well as the fact that it intrinsically

presupposes a fully developed Vedic scholarly tradition, it is not possible that it was

borrowed by the Buddhists from a substratum held in common with the proponents of

the new Brahmanism. As a result, I argued that tevijja, unlike brhmaa, was

borrowed from Brahmanism for use as a polemic against it, and that we must

therefore treat all supposedly Brahmanical categories both diachronically and on a

case-by-case basis, rather than simply assuming that all such terms entered the

Buddhist tradition at the same time or in the same way.

Throughout Parts I-III, our discussion of all of these issues lay under the

specter of what I call encounter dialogsstras in which the Buddha is generally

not identified as a Brahman, but encounters an interlocutor who is. It is this genre of

texts, I have argued, that more than anything else has led to the impression that

637
Buddhism rose in opposition to a pre-existing, fully-formed, and coherent

Brahmanical tradition. Since these Buddhist stras pose the greatest threat to his

theory that Buddhism did not arise in a Brahmanical context, Bronkhorst has argued

that they developed only relatively late in the Buddhist tradition, as the Western

Brahmans slowly moved east into Greater Magadha. In Part II, I showed that, while I

do not agree with Bronkhorsts radical dichotomy between Vedic and Greater

Magadhan culture, encounter dialogs do dwell on themes that are clearly discernible

from those found in other texts that I call teachings on wrong views. Teachings on

wrong views deal with issues of karma, rebirth, the nature of the self (if any), and

liberation from rebirth, but these themes are almost completely absent from the

encounter dialogs. Instead, the latter deal mostly with the claims made by the new

Brahmanismthe vara system, Vedic learning and sacrifice, the superiority of

Brahmans on the basis of birth, and the superiority of the householder lifestyle to

celibate lifestyles. Since the teachings on wrong views, which do deal with karma,

the tman, and rebirth, do not generally refer to Brahman interlocutors, and the

encounter dialogs, which do refer to Brahman interlocutors, do not deal with karma,

the tman, and rebirth, Bronkhorst concludes that concepts surrounding karma, the

tman, and rebirth arose in a non-Brahmanical context, and that the Brahmans were

originally unconcerned with such concepts. 947

Because I reject, as a matter of methodology, any sharp dichotomy between

Brahmanical and non-Brahmanical in ancient India, I instead interpreted this

947
I am of course restating Bronkhorsts argument from Greater Magadha somewhat using
my own terminology of encounter dialogs and teachings on wrong views.

638
difference between the two genres of texts as reflective of two different projects in the

early Buddhists effort to articulate their own identity vis--vis others. In the

teachings on wrong views, the early Buddhists attempt to situate themselves within

the broader range of philosophical opinions by distinguishing their own view (right

view) from those of some (eke) samaabrhma, a category that I argued was

used as a foil against which to construct Buddhist identity. Some or even many or all

of these others with wrong views may have called themselves Brahmansjust as,

frankly, the Buddhists didand indeed, as much would seem to be implied by the

fact that they are referred to in these texts as samaas and brhmaas. These

various other groups are treated as an aggregate, however; whether any particular

group were really Brahmans is not of any concern to the Buddhist authors of these

texts. In encounter dialogs, on the other hand, this is the primary concern.

Encounter dialogs, I argued, are polemical texts directed toward a specific non-

Buddhist group, namely the proponents of the new Brahmanism, and they deal

specifically with refuting the claims of that group, first and foremost that they alone

are truly Brahmans, by virtue of their birth.

However, it was not until Part IV that I had the opportunity to focus

specifically on this latter genre of texts, in order to trace their development and

explore the tremendous effect they had on Buddhist identity. Because the

development of the encounter dialog genre can only be understood within the context

of the broader early Buddhist tradition, I began in Chapter IV.1 by discussing the

orality of the early Buddhist tradition and demonstrating the applicability of the Oral

639
Theory of Parry and Lord to its study. Then, in Chapter IV.2, I discussed the

development of the Nikya/gamas, and in Chapter IV.3, I traced groups of

encounter dialogs that appear to have been kept together in the early Buddhist

tradition, especially in the Sayutta Nikya/Sayuktgama and Dgha

Nikya/Drghgama traditions. Finally, in Chapter IV.4, I looked at specific oral

formulae and themes that are characteristic of the encounter dialog genre and cross

Nikya/gama boundaries.

On the basis of this detailed investigation of the encounter dialogs in the

broader context of the early Buddhist tradition, I was able to come to some

conclusions about their development as a genre. Two collections of encounter

dialogs appear to be of particular antiquity in the traditionthe Brhmaa

Sayutta/Sayukta in the Sayutta Nikya/Sayuktgama and the Silakkhandha

Vagga/laskandha Nipta of the Dgha Nikya/Drghgama. Of these, the former

would appear to be older because it is found in the Sayutta/Sayukta tradition,

which, according to the theory of Ynshn, is the oldest of all the Nikya/gama

traditions, and because it follows a very simple patternthe framing of verses by a

short prose narrative involving a Bhradvja Brahman with a humorous nameand

lacks the various more developed oral formulae found in the Dgha/Drgha tradition.

I therefore argued that the encounter dialog genre developed from these very simple

narratively framed verses such as found in the Sayutta/Sayukta to more

elaborate texts, depending on a complex network of oral formulae and themes, in the

Dgha/Drgha and throughout the rest of the early Buddhist tradition. In the process

640
of developing this genre to combat the claims of the new Brahmanism, however, the

early Buddhists effectively ceded the category Brahman to them. Even though the

logical purport of these texts was that the proponents of the new Brahmanism could

not legitimately claim to be Brahmans simply on the basis of birth, the narratives of

these texts, by their very structure, conceded the category to them by portraying them

as Brahmans in juxtaposition with the Buddha as non-Brahman. In so doing, the

encounter dialog genre solidified Buddhist identity in contradistinction to the

Brahmans and effectively brought an end to the use of the category Brahman by

Buddhists to articulate their own identity. It was this that made possible the existence

of distinct Buddhist and Brahmanical identities, the antagonism between those two

groups that came to a head in the medieval period, and of course the Hnaynists

slandering of the Pacaviatishasrik as a Brahman book when Zh Shxng

tried to take it from Khotan.

There are two major goals that I hope to have met in this dissertation. The

first is to establish the validity of a methodology that does not assume a sharp

dichotomy between Brahmanical and non-Brahmanical in ancient India, and that

in particular does not assume that Brahmanism has a metahistorical essence that

preceded Buddhism, and against which Buddhism arose in opposition. What I hope

to have shown is that Buddhism and Brahmanism, as clearly distinct and antagonistic

sectarian identities, developed only slowly over the course of time. Much of what

they share in common is due not to borrowing by one from the other, but to a

common substratum from which they both drew, and over which they contested with

641
one another, and it was out of that process of contestation that they emerged as

distinct and antagonistic sectarian identities.

The second major goal I hope to have met is to establish usefulness of

combining the insights of Oral Theory with the comparative study of the Pali Nikyas

and Chinese gamas to the end of developing a better understanding of the

development of the early Buddhist oral tradition. Because the Pali Canon and the

Chinese gamas were taken from different sectarian traditions and were written

down at different times, they can be regarded almost as tape recordings of

particular performances of the early Buddhist tradition, and by comparing them, we

can learn more about how that oral tradition worked and what changed over the

course of its development. As we saw in Part IV, such comparison can allow us to

come to conclusions about the relative antiquity of particular texts, themes, and

formulae in the tradition, and thus give us a sense of how the tradition developed.

Needless to say, even in a study of this length, there are certain limitations.

One limitation of this dissertation is that, while it addresses comprehensively the

relationship between early Buddhism and Brahmanism, it does so from a mostly

Buddhist perspective. This was somewhat inevitable, given my own strengths and

just the sheer amount of possible data, but it also opens up great opportunities for

future research, through investigating a broader variety of Brahmanical texts using

the methodological approach I have outlined in this dissertation. In particular, due in

large part to the importance of the encounter dialog genre in the early Buddhist

tradition, I have focused mostly on the proponents of the new Brahmanism and the

642
texts that are most closely reflective of their concerns, namely, the Dharma Stras.

Within the broad corpus of texts that we now understand to be Brahmanical,

however, it is clear that the new Brahmanismor, perhaps better put, the new

Brahmanism that is represented by both the Dharma Stras and the encounter

dialogsis not the only voice present. Even if the Buddhists were not successful,

through their encounter dialogs, in refuting the idea that Brahmanhood is determined

by birth, it does appear that they, together with their quasi-allies, the jailas, were

successful in gaining acknowledgement of the validity of celibate lifestyles. The

Dharma Stra authors, as we saw, were critical even of the vnaprasthas, who appear

to have been engaging in Vedic practices of a certain sort in the forest, because they

were celibate, but in Manu and the later Dharmaastric literature, this appears to have

become the model for an entire reformulation and normativization of the rama

system that allowed for such a lifestyle later in life. A lot more work needs to done

on the formulation of Brahmanical self-identity, and the role that pro-householder

and pro-ascetic/celibate groups played in this process. The Mahbhrata, which I

have only had the opporuntiy to treat in passing in this dissertation, obviously offers a

wealth of data that will provide insight into this process, and already important work

on it is being done that will lay the foundation for a better understanding of the

development of Brahmanical identity. 948

Another limitation of this dissertation is the limited scope of my comparative

study of the Nikya/gama traditions in Part IV. In part, this is due to the limited

948
See p. 39-40 for some references.

643
thematic scope of this dissertation as a whole; in part, it is also due to the fact that so

few Chinese gama texts have been studied or even translated in Western languages.

While I did my best in this dissertation to deal with the encounter dialog genre as

comprehensively as possible, this was most easily accomplished within the context of

Pali Nikya texts that have been translated into English, and it was not always

possible to find every Chinese gama text that can be considered to be an encounter

dialog or to investigate in depth every known Chinese parallel to a particular

encounter dialog identified in the Pali tradition. In addition, although the results of

our restricted study of the encounter dialog genre have been quite fruitful, a fuller

understanding of the development of the early Buddhist oral tradition can clearly only

be accomplished by tracing many different types of texts, with their associated oral

formulae and themes, in the early Buddhist tradition by comparison of the extant Pali,

Chinese, and even Sanskrit and Tibetan versions. This, however, must remain a

desideratum for future research.

644
 Bibliography

Adikaram, E. W. Early History of Buddhism in Ceylon, or State of Buddhism in

Ceylon as Revealed by the Pli Commentaries of the 5th Century A.D.
Colombo: M. D. Gunasena and Co., 1946.

Akanuma Chizen. The Comparative Catalogue of the Chinese gamas and the Pli
Nikyas. Nagoya: Hajinkaku-shobo, 1929.

Allon, Mark. The Oral Composition and Transmission of Early Buddhist Texts. In
Indian insights: Buddhism, Brahmanism and Bhakti. Papers from the Annual
Spalding Symposium on Indian Religion, edited by Sute Hamilton and Peter
Connolly, 39-61. London: Luzac Oriental, 1997.

________. Style and Function: A study of the dominant stylistic features of the prose
portions of Pli canonical sutta texts and their mnemonic function. Tokyo:
The International Institue for Buddhist Studies of the International College for
Advanced Buddhist Studies, 1997.

_________. Three Gndhr Ekottarikgama-Type Stras: British Library Kharoh

Fragments 12 and 14. Gandhran Buddhist Texts. Vol. 2. Seattle: University
of Washington Press, 2001.

Almond, Philip C. The British Discovery of Buddhism. Cambridge, England:

Cambridge University Press, 1988.

Anlayo Bhikkhu. The Chinese Madhyama-gama and the Pli Majjhima-Nikya

In the Footsteps of Thich Minh Chau. The Indian International Journal of
Buddhist Studies 9 (2008): 1-21.

________. Comparative Notes on the Madhyama-gama. Fuyan Buddhist Studies

2 (2007): 33-56.

645
________. A Comparative Study of the Majjhima-nikya. 2 vols. Taipei: Dharma
Drum Publishing, 2011.

________. The Development of the Pli Udna Collection. Buddhist Studies

(Bukky Kenky) 37 (March 2009): 39-72.

________. Zeng-yi A-han. In Encyclopedia of Buddhism. Edited by W. G.

Weeraratne. Vol. 8. No. 3. Sri Lanka: Department of Buddhist Affairs, 2009.

Anlayo and Roderick S. Bucknell. Correspondence Table for Parallels to the

Discourses of the Majjhima Nikya: Toward a Revision of Akanumas
Comparative Catalogue. Journal of the Centre for Buddhist Studies, Sri
Lanka 4 (2006): 215-43.

nandajoti Bhikkhu. A Comparative Study of the Dhammapada. Unpublished

manuscript.

Anesaki, M. The Four Buddhist gamas in Chinese: A Concordance of their Parts

and the Corresponding Counterparts in the Pli Nikyas. Transactions of the
Asiatic Society of Japan 35, part 3 (1908): 1-149.

Anuruddha Thera, Kkkpalliye, et al., trans. The First and Second Buddhist Councils:
Five Versions. Hong Kong: Chi Lin Nunnery, 2008.

Bailey, Greg, and Ian Mabbett. The Sociology of Early Buddhism. Cambridge,
England: Cambridge University Press, 2003.

Ballantyne, J. R., trans. The Vedanta-sara. London: The Christian Literature Society
for India, 1898.

Bandhu, Vishva. A Vedic Word-Concordance. Vol. 5 (Index). Part 1. Hoshiarpur:

V.V.R. Institute, 1964.

Bapat, P. V., trans. Arthapada Sutra: Spoken by the BuddhaTranslated by the

Upsaka Che-Kien under the Wu Dynasty (222-280 A.D.). 2 parts in 1 vol.
Calcutta: Visva-Bharati, 1951.

Bareau, Andr. Recherches sur la biographie du Buddha dans les Strapiaka et les
Vinayapiaka anciens: de la qute de lveil la conversion de riputra et de
Maudgalyyana, Publications de lEcole Franaise dExtrme-Orient, vol. 53.
Paris: Ecole Franaise dExtrme-Orient, 1963.

Basham, A. L. History and Doctrines of the jvikas: A Vanished Indian Religion.

Delhi: Motilal Banarsidass, 2002[1951].

646
Bechert, Heinz. The Dating of the Historical Buddha. 2 vols. Gttingen:
Vandenhoeck and Ruprecht, 1991-1992.

Bechert, Heinz, and Klaus Wille, eds. Sanskrithandschriften aus den Turfanfunden.
Vol. 6. Stuttgart: Franz Steiner Verlag Wiesbaden GMBH, 1989.

Bhat, M. Ramakrishna, trans. Varhamihiras Bhat Sahit. Vol. 2. Delhi: Motilal

Banarsidass, 1982.

Bhattacharya, Kamaleswar. Ltman-Brahman dans le Bouddhisme Ancien. Paris:

cole Franaise dExtrme Orient, 1973.

Bhattacharya, Ram Shankar, ed. Trikaea of Puruottamadeva. Varanasi: Ratna

Publications, 1995.

Biardeau, Madeleine. Le Mahbhrata: Un rcit fondateur du brahmanisme et son

interprtation. 2 vols. Paris: ditions du Seuil, 2002.

Bingenheimer, Marcus. The Shorter Chinese Sayukta gama: Preliminary

Findings and Translation of Fascicle 1 of the Bieyi za ahan jing
(T. 100). Buddhist Studies Review 23, no. 1 (2006): 21-60.

Black, Brian. Ambaha and vetaketu: Literary Connections Between the Upaniads
and Early Buddhist Narratives. Journal of the American Academy of Religion
79, no. 1 (March 2011): 136-161.

________. The Character of the Self in Ancient India: Priests, Kings, and Women in
the Early Upaniads. Albany, NY: SUNY Press, 2007.

________. Rivals and Benefactors: Encounters Between Buddhists and Brahmins in

the Nikyas. Religions of South Asia 3, no. 1 (2009): 25-43.

Bodhi, Bhikkhu, trans. The Connected Discourses of the Buddha: A New Translation
of the Sayutta Nikya. Boston: Wisdom Publications, 2000.

Bhtlingk, Otto, and Rudolph Roth. Sanskrit-Wrterbuch. Vol. 5. St. Petersburg:

Kaiserlichen Akademie der Wissenschaften, 1858.

Bronkhorst, Johannes. Buddhism in the Shadow of Brahmanism. Leiden: Brill, 2011.

________. Dharma and Abhidharma. Bulletin of the School of Oriental and African
Studies 48, no. 2 (1985): 305-20.

647
________. Greater Magadha: Studies in the Culture of Early India. Leiden: Brill,
2007.

________. The Two Sources of Indian Asceticism. Bern: Peter Lang, 1993.

________. The Two Traditions of Meditation in Early India. Delhi: Motilal

Banarsidass, 1993.

Brough, John. The Early History of the Gotras. Journal of the Royal Asiatic Society
of Great Britain and Ireland, no. 1 (Apr. 1946): 32-45.

________. The Early History of the Gotras (Concluded). Journal of the Royal
Asiatic Society of Great Britain and Ireland, no. 1 (Apr. 1947): 76-90.

________. The Gndhr Dharmapada. London Oriental Series. Vol. 7. London:

Oxford University Press, 1962.

Bucknell, Roderick. The Historical Relationship Between the Two Chinese

Sayuktgama Translations. Chung-Hwa Buddhist Journal 24 (2011): 35-70.

________. The Other Translation of Sayuktgama. Chung-Hwa Buddhist Journal

21 (2008): 24-54.

________. Sayuktgama. In Encyclopedia of Buddhism, edited by W. G.

Weeraratne, vol. 7, fasc. 4, 684-7. Colombo: Government of Ceylon, 2006.

________. The Structure of the Sagtha-Vagga of the Sayutta Nikya. Buddhist

Studies Review 24, no. 1(2007), 13-18.

Bhler, Georg. Buddhas Quotation of a Gth by Sanatkumra. Journal of the

Royal Asiatic Society (July 1897): 585-8.

Burford, Grace G. Desire, Death and Goodness: The Conflict of Ultimate Values in
Theravda Buddhism. New York: Peter Lang, 1991.

Burnouf, Eugne. Introduction to the History of Indian Buddhism. Translated by

Katia Buffetrille and Donald S. Lopez, Jr. (Chicago: University of Chicago
Press, 2010

Carter, John Ross. On Understanding Buddhists: Essays on the Theravda Tradition

in Sri Lanka. Albany, NY: SUNY Press, 1993.

Chakravarti, Uma. The Social Dimensions of Early Buddhism. Delhi: Oxford

University Press, 1987.

648
Choong Mun-keat. A comparison of the Pli and Chinese versions of the Brhmaa
Sayutta, a collection of early Buddhist discourses on the priestly
Brhmaas. Journal of the Royal Asiatic Society, Series 3, vol. 19, no. 3
(2009): 371-82.

________. The Fundamental Teachings of Early Buddhism: A comparative study

based on the Strga portion of the Pli Sayutta-Nikya and the Chinese
Sayuktgama. Wiesbaden: Harrassowitz Verlag, 2000.

Cohen, Richard S. Nga, Yaki, Buddha: Local Deities and Local Buddhism at
Ajanta. History of Religions 37, no. 4 (May 1998): 360-400.

Collins, Steven. The Discourse on What is Primary (Aggaa-Sutta): An Annotated

Translation. Journal of Indian Philosophy 21(1993): 301-393.

Cousins, Lance. Pali Oral Literature. In Buddhist Studies Ancient and Modern,
edited by P. Denwood and A. Piatigorsky, 1-11. London: Curzon Press, 1983.

Cowell, E. B. The Jtaka, or Stories of the Buddhas Former Births. Trans. Robert
Chalmers et al. 6 vols. in 3 physical vols. Delhi: Motilal Banarsidass
Publishers, 1994[1895].

de Jong, J. W. A Brief History of Buddhist Studies in Europe and America (Part I).
Eastern Buddhist 7, no. 1 (1974): 55-106.

________. A Brief History of Buddhist Studies in Europe and America (Part II).
Eastern Buddhist 7, no. 2 (1974): 49-82.

________. The Buddha and His Teachings. In Wisdom, Compassion, and the
Search for Understanding. The Buddhist Studies Legacy of Gadgin M. Nagao,
edited by Jonathan A. Silk, 171-81. Honolulu: University of Hawaii Press,
2000.

Drewes, David Donald. Mahyna Stras and Their Preachers: Rethinking the
Nature of a Religious Tradition. PhD diss., University of Virginia, 2006.

Dumont, Louis. Homo Hierarchicus: The Caste System and its Implications.
Translated by Mark Sainsbury, Louis Dumont, and Basia Gulati. Chicago:
University of Chicago Press, 1980[1966].

Enomoto Fumio. Mlasarvstivdin and Sarvstivdin. In

Vividharatnakaraaka: Festgabe fr Adelheid Mette, ed. Christine
Chojnacki, Jens-Uwe Hartmann, and Volker M. Tschannerl, Indica et Tibetica:
Monographien zu den Sprachen und Literaturen des indo-tibetischen

649
 Kulturraumes, vol. 37, 239-50. Swisttal-Odendorf: Indica et Tibetica Verlag,
2000.

Fitzgerald, James, trans. The Mahbhrata. Vol. 7. Chicago: University of Chicago

Press, 2004.

________. Making Yudhihira the King: The Dialectics and the Politics of Violence
in the Mahbhrata. Rocznik Orientalistyczny 54, no. 1 (2001): 63-92.

Foley, John Miles. Introduction: The Oral Theory in Context. In Oral Traditional
Literature: A Festschrift For Albert Bates Lord, edited by John Miles Foley,
27-122. Columbus, Ohio: Slavica Publishers, 1981.

________. Oral-Formulaic Theory and Research: An Introduction and Annotated

Bibliography. New York: Garland Publishing, 1985.

Franco, Eli. Lokyata. In Brill Encyclopedia of Hinduism, edited by Knut Jacobsen

et al., vol. 3, 629-42. Leiden: Brill, 2011.

Freiberger, Oliver. The ideal sacrifice. Patterns of reinterpreting brahmin sacrifice in

early Buddhism. Bulletin dEtudes Indiennes 16 (1998): 39-49.

________. Negative Campaigning: Polemics Against Brahmins in a Buddhist Sutta.

Religions of South Asia 3, 1 (2009): 61-76.

Gethin, Rupert. The Buddhist Path to Awakening. Oxford: Oneworld Publications,

2007.

________. The Mtiks: 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: SUNY Press, 1992.

Glass, Andrew. Four Gndhr Sayuktgama Stras: Senior Kharoh Fragment 5.

Gandhran Buddhist Texts. Vol. 4. Seattle: University of Washington Press,
2007.

Goldman, Robert. Gods, Priests, and Warriors: The Bhgus of the Mahbhrata. New
York: Columbia University Press, 1977.

Gombrich, Richard. The Buddhas Book of Genesis? Indo-Iranian Journal 35

(1992): 159-178.

________. How Buddhism Began: The Conditioned Genesis of the Early Teachings.
New Delhi: Munshiram Manoharlal Publishers, 2007.

650
________. How the Mahyna Began. Journal of Pali and Buddhist Studies 1
(March 1988): 29-46.

________. Recovering the Buddhas Message. In Buddhism: Critical Concepts in

Religious Studies, edited by Paul Williams, vol. 1, 113-128. New York:
Routledge, 2005.

________. What the Buddha Thought. London: Oxford Center for Buddhist Studies,
2009.

Gmez, Luis O. Proto-Mdhyamika in the Pli Canon. Philosophy East and West
26, no. 2 (Apr. 1976): 137-165.

Halbfass, Wilhelm. India and Europe: An Essay in Understanding. Albany, NY:

SUNY Press, 1988.

Harrison, Paul. Another Addition to the An Shigao Corpus? Preliminary Notes on an

Early Chinese Sayuktgama Translation. In Studies in Honour of Dr.
Sakurabe Hajime on the Occasion of His Seventy-seventh Birthday, ed.
Sakurabe Ronshu Committee, 1-32. Kyoto: Heirakuji Shoten, 2002.

________. The Ekottarikgama Translations of An Shigao. In

Bauddhavidysudhkara: Studies in Honour of Heinz Bechert on the
Occasion of His 65th Birthday, edited by Petra Kieffer-Plz and Jens-Uwe
Hartmann, Indica et Tibetica, vol. 30, 261-84. Swisttal-Odendorf: Indica et
Tibetica Verlag, 1997.

Hartmann, Jens-Uwe. Contents and Structure of the Drghgama of the (Mla)-

Sarvstivdins. Annual Report of the International Research Institute for
Advanced Buddhology at Soka University 7 (2004): 119-37.

________. Further Remarks on the New Manuscript of the Drghgama. Journal of

the International College for Advanced Buddhist Studies 5 (2002): 133-50.

________. More Fragments of the Cagstra. In Buddhist Manuscripts, edited by

Jens Braarvig, vol. 2, 1-16. Oslo: Hermes Publishers, 2002.

________. Untersuchungen zum Drghgama der Sarvstivdins.

Habilitationsschrift, Georg-August-Universitt, Gttingen, 1991.

________. Zu einer neuen Handschrift des Drghgama. In Vividharatnakaraaka:

Festschrift fr Adelheid Mette, edited by Christine Chojnacki et al., Indica et
Tibetica, vol. 37, 359-67. Swisstal-Odendorf: Indica et Tibetica Verlag, 2000.

651
Heesterman, J. C. The broken world of sacrifice: an essay in ancient Indian ritual.
Chicago: University of Chicago Press, 1993.

________. The inner conflict of tradition: essays in Indian ritual, kingship, and
society. Chicago: University of Chicago Press, 1985.

Hiltebeitel, Alf. Buddhism and the Mahbhrata: Boundary dynamics in textual

practice. In Boundaries, dynamics, and construction of traditions in South
Asia, edited by F. Squarcini, 107-131. Florence: Firenze University Press,
2005.

________. Dharma: Its Early History in Law, Religion, and Narrative. Oxford:
Oxford University Press, 2011.

________. Rethinking the Mahbhrata: A Readers Guide to the Education of the

Dharma King. Chicago: University of Chicago Press, 2001.

Hinber, Oskar von. Die neun Agas: Ein frher Versuch zur Einteilung
buddhistischer Texte. Wiener Zeitschrift fr die Kunde Sdasiens 38 (1994):
121-35.

________. A Handbook of Pli Literature. Berlin: Walter de Gruyter, 2000.

________. Hoary Past and Hazy Memory: On the History of Early Buddhist Texts.
Journal of the International Association of Buddhist Studies 29, no. 2
(2006[2008]): 193-210.

Hoernle, A. F. Rudolf. The Sutta Nipata in a Sanskrit Version from Eastern

Turkestan. Journal of the Royal Asiatic Society 48, n. 4 (Oct. 1916): 709-732.

Horner, I. B. The Book of the Discipline (Vinaya-Piaka). 6 vols. Oxford: The Pali
Text Society, 1982-1992[1938-1966].

Horsch, Paul. Vom Schpfungsmythos zum Weltgesetz. Asiatische Studien:

Zeitschrift der Schweizerischen Gesellschaft fr Asiankunde, vol. 21 (Francke:
1967): 3161.

Hultzsch, E. Inscriptions of Aoka. Corpus Inscriptionum Indicarum. Vol. 1. Tokyo:

Meicho-Fuky-kai, 1977.

Inden, Ronald. Changes in the Vedic Priesthood. In Ritual, state and history in
South Asia: Essays in honour of J.C. Heesterman, edited by A. W. van den
Hoek et al., 556-77. Leiden: Brill, 1992.

652
Ireland, John D., trans. The Udna and the Itivuttaka: Two Classics from the Pli
Canon. Kandy: Buddhist Publication Society, 1997.

Jayatilleke, K. N. Early Buddhist Theory of Knowledge. London: George Allen and

Unwin Ltd., 1963.

Jayawickrama, N. A. The Vaggas of the Sutta Nipta. University of Ceylon Review

6 (1948): 243-249.

Jones, J. J., trans. The Mahvastu. London: Luzac and Company, 1949.

Jurewicz, Joanna. Playing with Fire: The prattyasamutpda from the Perspective of
Vedic Thought. Journal of the Pali Text Society 26 (2000): 77-103.

Kane, Pandurang Vaman. History of Dharmastra. Vol. 2. Part 1. Poona:

Bhandarkar Oriental Research Institute, 1974.

Keith, A. Berriedale. Pre-Canonical Buddhism. The Indian Historical Quarterly 12

(1936): 1-20.

King, Richard. Orientalism and Religion: Postcolonial Theory, India and The Mystic
East. London: Routledge, 1999.

Kamira. The Rise of Wisdom Moon. Translated and edited by Matthew T.

Kapstein. The Clay Sanskrit Library. Edited by Sheldon Pollock. New York
University Press, 2009.

La Valle Poussin, Louis de. Musla et Nrada: Le Chemin de Nirva. Mlanges

chinois et bouddhiques 5 (1936-7): 189-222.

Lamotte, tienne. History of Indian Buddhism: From the Origins to the aka Era.
Translated by Sara Webb-Boin. Louvain-la-Neuve: Universit Catholique de
Louvain, 1988[1958].

Lenz, Timothy. A New Versiton of the Gndhr Dharmapada and a Collection of

Previous-Brith Stories: British Library Kharoh Fragments 16 + 25.
Gandhran Buddhist Texts. Vol. 3. Seattle: University of Washington Press,
2003.

Lindtner, Christian. The Problem of Precanonical Buddhism. Buddhist Studies

Review 14, no. 2 (1997): 109-139.

Lopez, Donald S., Jr. Buddha. In Critical Terms for the Study of Buddhism, edited
by Donald S. Lopez, Jr. Chicago: University of Chicago Press, 2005.

653
Lord, Albert B. Characteristics of Orality. Oral Tradition 2, no. 1 (1987), 66-7.

________. Perspectives on Recent Work on Oral Literature. In Oral Literature,

edited by Joseph Duggan, 1-24. New York: Barnes and Noble, 1975.

________. Perspectives on Recent Work on the Oral Traditional Formula. Oral

Tradition 1, no. 3 (1986): 467-503.

________. The Singer of Tales. Cambridge, Mass.: Harvard University Press, 1964.

Lubin, Timothy. The Transmission, Patronage, and Prestige of Brahmanical Piety

from the Mauryas to the Guptas. In Boundaries, Dynamics and Construction
of Traditions in South Asia, edited by Federico Squarcini, 77-103. Florence:
Firenze University Press, 2005.

Macdonell, Anthony. A Vedic Grammar for Students. Delhi: Motilal Banarsidass,

1993.

Maggi, Mauro. The Khotanese Karmavibhaga. Serie Orientale Roma. Vol. 74.
Roma: Istituto Italiano per il Medio ed Estremo Oriente, 1995.

Malamoud, Charles. Thologie de la dette dans les Brhmaa. Pururtha : Science

Sociales en Asie du Sud 4 (1980): 39-62.

Mann, Joy. Categories of Sutta in the Pli Nikyas and their Implications for our
Appreciation of the Buddhist Teaching and Literature. Journal of the Pali
Text Society 15 (1990): 29-87.

Masefield, Peter, trans. The Udna. Oxford: Pali Text Society, 1994.

Mayeda Egaku. Japanese Studies on the Schools of the Chinese gamas. In Genshi
bukky seiten no seiritsushi kenky [A History of the Formation of Original
Buddhist Texts], 94-103. Tokyo: Sankib Busshorin, 1964.

McClish, Mark. Political Brahmanism and the State: A Compositional History of the
Arthastra. PhD diss., University of Texas at Austin, 2009.

McCrindle, John W. Ancient India as Described in Classical Literature. St. Leonards:

Ad Orientem, 1971[1901].

________. Ancient India as Described by Megasthens and Arrian. Calcutta:

Thacker, Spink, and Co., 1877.

654
Meisig, Konrad. Das rmayaphala-Stra: Synoptische bersetzung und Glossar
der chinesischen Fassungen verglichen mit dem Sanskrit und Pli. Wiesbaden:
Otto Harrassowitz, 1987.

________. Das Stra von den vier Stnden: Das Aggaa-Sutta im Licht seiner
chinesischen Parallelen. Wiesbaden: Otto Harrassowitz, 1988.

Minh Chau, Thich. The Chinese Madhyama gama and the Pli Majjhima Nikya: A
Comparative Study. Delhi: Motilal Banarsidass, 1991.

Mori Sodo. The origin and the history of the Bhaka tradition. In nanda: Papers
on Buddhism and Indology. A felicitation volume presented to Ananda
Weihena Palliya Guruge on his sixtieth birthday, edited by Y. Karunadasa,
123-9. Colombo: Felicitation Volume Editorial Committee, 123-129.

Mller, F. Max. Comparative Mythology: An Essay. Edited by Abram Smythe Palmer.

London: George Routledge and Sons, n.d.

Nakamura Hajime. Indian Buddhism: A Survey with Bibliographical Notes. Delhi:

Motilal Banarsidass, 2007.

Nakano, Gish, and N.A. Jayawickrama. crya. In Encyclopedia of Buddhism,

edited by G.P. Malalasekera, vol. 1, 163-8. Government of Ceylon, 1961.

amoli, Bhikkhu, and Bhikkhu Bodhi. The Middle Length Discourses of the
Buddha: A new Translation of the Majjhima Nikya. Boston: Wisdom
Publications, 1995.

Nanda Raksakhom. [Medical Treatment using

Religious Methods]. BA Senior Thesis, College of Religious Studies,
Mahidol University, 2005.

Nattier, Jan. A Guide to the Earliest Chinese Buddhist Translations. Tokyo: The
International Research Institute for Advanced Buddhology, Soka University,
2008.

Norman, K. R., trans. The Elders Verses I: Theragth. London: Luzac and
Company, 1969.
________, trans. The Elders Verses II: Thergth. London: Luzac and Company,
1971.

________, trans. The Group of Discourses (Sutta Nipta). Vol. 2. Oxford: Pali Text
Society, 1992.

655
________. A note on Att in the Alagaddpama Sutta. In Studies in Indian
Philosophy: A Memorial Volume in Honour of Pandit Sukhlaji Sanghvi, LD
series 84, 19-29. Ahmedabad, 1981.

________. Theravda Buddhism and Brahmanical Hinduism: Brahmanical Terms in

a Buddhist Guise. The Buddhist Forum 51 (1991): 193-200.

OFlaherty, Wendy Doniger, trans. The Rig Veda. London: Penguin Books, 1981.

Oldenberg, Hermann. Buddha: His Life, His Doctrine, His Order. Translated by
William Hoey. London: Williams and Norgate, 1892.

Olivelle, Patrick. The rama System: The History and Hermeneutics of a Religious
Institution. New Delhi: Munshiram Manoharlal Publishers, 1993.

________, ed. and trans. Dharmastra Parallels: Containing the Dharmastras of

pastamba, Gautama, Baudhyana, and Vasiha. Delhi: Motilal Banarsidass
Publishers, 2005.

________, trans. Dharmastras: The Law Codes of pastamba, Gautama,

Baudhyana, and Vasiha. Oxford: Oxford University Press, 1999.

________, trans. The Early Upaniads. New York: Oxford University Press, 1998.

________, trans. Manus Code of Law: A Critical Edition and Translation of the
Mnava-Dharmastra. Oxford: Oxford University Press, 2005.

________. The Origin and Early Development of Buddhist Monachism. Colombo:

Gunasena, 1974.

________. Review of The Two Sources of Indian Asceticism by Johannes

Bronkhorst. Journal of the American Oriental Society 115, no. 1 (1995): 162-
4.

Overmyer, Daniel, and Joseph Adler. Chinese Religion: An Overview. In Lindsay

Jones, ed., Encyclopedia of Religion, edited by Lindsay Jones, 2nd ed., vol. 3,
1580-1613. Detroit: Macmillan Reference USA, 2005.

Pandit, Ranjit Sitaram. Rjataragi. New Delhi: Sahitya Akademi, 1968[1935].

Parry, Adam, ed. The Making of Homeric Verse: The Collected Papers of Milman
Parry. Oxford: Clarendon Press, 1972.

656
Prez-Remn, Joaqun. Self and Non-Self in Early Buddhism. The Hague: Mouton
Publishers, 1980.

Pingree, David. Jyotistra: Astral and Mathematical Literature. Edited by Jan

Gonda. A History of Indian Literature. Vol. 6. Fasc. 4. Wiesbaden: Otto
Harrassowitz, 1981.

Pollock, Sheldon. The Language of the Gods in the World of Men: Sanskrit, Culture,
and Power in Premodern India. Berkeley: University of California Press,
2006.

Powers, John. A Bull of a Man: Images of Masculinity, Sex, and the Body in Indian
Buddhism. Cambridge, Mass.: Harvard University Press, 2009.

Rgamey, Constantin. Le problme du bouddhisme primitif et les derniers travaux

de Stanislaw Schayer. Rocznik Orientalistyczny 21 (1957): 37-58.

Reynolds, Frank E., and Charles Hallisey. The Buddha. In Buddhism and Asian
History, edited by Joseph M. Kitagawa and Mark D. Cummings, 29-49. New
York: Macmillan Publishing Company, 1989.

Rhys Davids, T. W. Buddhist India. New York: G.P. Putnams Sons, 1903.

________. Dialogues of the Buddha. 3 vols. London: Oxford University Press, 1899.

Richard Lariviere, Dharmastra, Custom, Real Law and Apocryphal Smtis,

Journal of Indian Philosophy 32 (2004

Rocher, Ludo. The Puras. A History of Indian Literature. Edited by Jan Gonda.
Vol. 3. Wiesbaden: Otto Harrassowitz, 1986.

Rotman, Andy, trans. Divine Stories: Divyvadna. Part 1. Boston: Wisdom

Publications, 2008.

Ruegg, David Seyfort. The Symbiosis of Buddhism with Brahmanism/Hinduism in

South Asia and of Buddhism with "local cults" in Tibet and the Himalayan
region. Wien: Verlag der sterreichischen Akademie der Wissenschaften,
2008.

Salomon, Richard. Ancient Buddhist Scrolls from Gandhra: The British Library
Kharoh Fragments. Seattle: University of Washington Press, 1999.

657
________. A Gndhr Version of the Rhinoceros Stra: British Library Kharoh
Fragment 5B. Gandhran Buddhist Texts. Vol. 1. Seattle: University of
Washington Press, 2000.

Samuel, Geoffrey. The Origins of Yoga and Tantra: Indic Religions to the Thirteenth
Century. Cambridge, England: Cambridge University Press, 2008.

Sander, Lore, and Ernst Waldschmidt, eds. Sanskrithandschriften aus den

Turfanfunden. Vol. 4. Wiesbaden: Franz Steiner Verlag GMBH, 1980.

Sarbacker, Stuart Ray. Samdhi: The Numinous and Cessative in Indo-Tibetan Yoga.
Albany, NY: SUNY Press, 2005.

Schayer, Stanislaw. New Contributions to the Problem of Pre-Hnaynistic

Buddhism. Polish Bulletin of Oriental Studies 1(1937): 8-17.

________. Precanonical Buddhism. Archiv Orientln 7 (1935): 121-32.

Schmidthausen, Lambert. On Some Aspects of Descriptions of Theories of

Liberating Insight and Enlightenment in Early Buddhism. In Studien zum
Jainismus und Buddhismus, Gedenkschrift fr L. Alsdorf, edited by Klaus
Bruhn and Albrecht Wezler, 199-250. Wiesbaden: Franz Steiner, 1981.

Schopen, Gregory. If You Cant Remember, How to Make It Up: Some Monastic
Rules for Redacting Canonical Texts. In Bauddhavidysudhkara. Studies
in Honour of Heniz Bechert on the Occasion of His 65th Birthday, edited by P.
Kieffer-Plz and J.-U. Hartmann, Indica et Tibetica, vol. 30, 571-82. Swisttal-
Odendorf: Indica et Tibetica, 1997.

________. Mahyna in Indian Inscriptions. In Figments and Fragments of

Mahyna Buddhism in India. Honolulu: The University of Hawaii Press,
2005[1979].

________. The Phrase sa pthivpradea caityabhto bhavet in the Vajracchedik:

Notes on the Cult of the Book in Mahyna. Indo-Iranian Journal 17 (1975):
147-81.

Senart, Emile. Essai sur la Lgende du Buddha: Son Charactre et ses Origines. Paris:
Imprimerie Nationale, 1875.

Skilling, Peter. Theravdin Literature in Tibetan Translation. Journal of the Pali

Text Society 19 (1993): 69-201.

658
________. Mahstras: Great Discourses of the Buddha. Sacred Books of the
Buddhists, Vols. 44 and 46. Oxford: The Pali Text Society, 1994 and 1997.

Smith, Brian K. Classifying the Universe: The Ancient Indian Vara System and the
Origins of Caste. New York: Oxford University Press, 1994.

Smith, Jonathan Z. Religion, Religions, Religious. in Critical Terms for Religious

Studies, edited by Mark C. Taylor, 269-284. Chicago: University of Chicago
Press, 1998.

Stede, W., ed. Niddesa. Vol. 3. London: Pali Text Society, 1918.

Sujto, Bhikkhu. A History of Mindfulness: How Insight Worsted Tranquility in the

Satipahna Sutta. Unpublished manuscript.

Tekin, inasi. Maitrisimit nom bitig. Die Uigurische bersetzung eines Werkes der
Buddhistischen Vaibhika Schule. Berlin: Akademie Verlag, 1980.

Tsuchida Ryutaro. Two categories of Brahmins in the early Buddhist period. The
Memoirs of the Toyo Bunko 49 (1991): 51-95.

van Buitenen, J. A. B., trans. The Mahbhrata. Vols. 1-3. Chicago: Chicago
University Press, 1973-8.

van der Veer, Peter. The Concept of the Ideal Brahman as an Indological Construct.
In Hinduism Reconsidered, edited by Gnther-Dietz Sontheimer and Hermann
Kulke, 153-72. New Delhi: Manohar Publishers and Distributors, 2005[1989].

Vetter, Tilmann. The Ideas and Meditative Practices of Early Buddhism. Leiden: E.J.
Brill, 1988.

________. Some Remarks on Older Parts of the Suttanipta. In Earliest Buddhism

and Madhyamaka: Panels of the VIIth World Sanskrit Conference, edited by
Schmidthausen and Ruegg, 36-56. Leiden: Brill, 1990.

Waldschmidt, Ernst, et al., eds. Sanskrithandschriften aus den Turfanfunden.

Verzeichnis der Orientalischen Handschriften in Deutschland. Vol. 10. Parts
1-10. Stuttgart: Franz Steiner Verlag, 1965-2008.

Walshe, Maurice. The Long Discourses of the Buddha: A Translation of the Dgha
Nikya. Boston: Wisdom Publications, 1995.

Warder, A. K. Pali Meter. London: Luzac, 1967.

659
White, David Gordon. Myths of the Dog-Man. Chicago: University of Chicago Press,
1991.

________. Sinister Yogis. Chicago: University of Chicago Press, 2009.

Williams, Paul. Mahyna Buddhism: The Doctrinal Foundations. 2nd ed. London:
Routledge, 2009.

Witzel, Michael. The Case of the Shattered Head. Studien zur Indologie und
Iranistik 13/14 (1987): 363-416.

________. The Development of the Vedic Canon and its Schools: The Social and
Political Milieu (Material on Vedic khs, 8). In Inside the Texts, Beyond
the Texts: New Approaches to the Study of the Vedas, edited by Michael
Witzel, Opera Minora, vol. 2, 257-345. Cambridge, Mass.: Harvard University
Department of Sanskrit and Indian Studies, 1997.

________. Early Sanskritization: Origins and Development of the Kuru State.

Electronic Journal of Vedic Studies 1, no. 4 (Dec. 1995).

________. On the Localization of Vedic Texts and Schools (Material on Vedic

khs, 7). In India and the Ancient world. History, Trade and Culture
before A.D. 650. P.H.L. Eggermont Jubilee Volume, edited by G. Pollet,
Orientalia Lovaniensia Analecta, vol. 25, 173-213. Leuven, 1987.

________. Tracing the Vedic Dialects. In Dialects dans les litteratures indo-
aryennes. Actes du Colloque International organise par UA 1058 sous les
auspices du C.N.R.S avec le soutien du College de France, de la Fondation
Hugot du College de France, de l'Universite de Paris III, du Ministre des
Affaires Etrangeres, Paris (Fondation Hugot) 16-18 Septembre 1986, edited
by Collette Caillat, 97-264. Paris: College de France, Institut de Civilisation
Indienne, 1989.

Wolf, Kenneth. The Poverty of Riches: St. Francis of Assisi Reconsidered. Oxford:
Oxford University Press, 2003.

Woodward, F. L., and E. M. Hare, trans. The Book of the Gradual Sayings (Aguttara
Nikya), or More-Numbered Suttas. 5 vols. Oxford: The Pali Text Society,
1996-2001.

Wynne, Alexander. The Origin of Buddhist Meditation. London: Routledge, 2007.

Zafiropulo, Ghiorgo. LIllumination du Buddha: De la Qute a lAnnonce de lEveil.

Innsbruck, 1993.

660
Zrcher, E. The Buddhist Conquest of China: The Spread and Adaptation of
Buddhism in Early Medieval China. 3rd ed. Leiden: Brill, 2007.

661