Skip to main content

Full text of "A Buddha Within: The Tathgatagarbhasutra — The Earliest Exposition of the Buddha Nature Teachings in India"

See other formats

PDF Version: BPPB VI (2002) 

A Buddha Within: 

he Tathagatagarbhasutra 


The Earliest Exposition 
Q of the 
wadhenNature Teaching in India 

PDF Version: BPPB VI (2002) 

Bibliotheca Philologica et Philosophica Buddhica 

Editor-in-Chief : Hiroshi Kanno 

Volume VI 




The International Research Institute for Advanced Buddhology 
Soka University 
Tokyo 2002 

PDF Version: BPPB VI (2002) 





The International Research Institute for Advanced Buddhology 
Soka University 
Tokyo 2002 

PDF Version: BPPB VI (2002) 

Published by the International Research Institute for Advanced Buddhology 
(=IRIAB), Soka University: 
1-236 Tangi, Hachioji, Tokyo 192-8577, Japan 
T 192-8577 RMN EFT APART 1-236 
Phone: (+ 81-426) 91-2695 
Fax: (+81-426) 91-4814 

First published 2002 
Printed in Japan by Meiwa Printing Company, Tokyo 

© Michael ZimmMeRMANN 2002 

All rights reserved. 

Apart from any fair dealing for the purpose of private study, research, criticism or 
review, no part of this book may be reproduced, translated or utilised in any form, 
either by print, photoprint, microform, multimedia or any other means, now known 
or hereafter invented, without written permission of the copyright holder and 
publishers. Enquiries should be made to the publishers. 

ISBN 4-9980622-5-5 

Correspondence regarding all editorial matters, including manuscripts to be offered for 
publication, may be sent to the Editor-in-Chief of the Series “Bibliotheca Philologica et 
Philosophica Buddhica” at The International Research Institute for Advanced Buddhology, 
Soka University, Tokyo, Japan. 

PDF Version: BPPB VI (2002) 

Preface 7 
Acknowledgements 8 
Technical Note: My Use of the Asterisk 10 

I A Study and Annotated Translation of the Tathdgatagarbhasittra 11 

A General Study and Text-historical Considerations 12 
1 Textual History and Structure of the TGS 16 
1,1 Different Recensions of the TGS ..........ccescccssssssseesescesscesessessesseseseesesseseessenens 16 
1.2 The Representatives Of TGS2 .....sessssesesssesesesesscseeseesesensoeseseessessecseneseseeeses 24 
1.3 Similarities between the Chinese Translations .............ccccsssscsscssseesseesseceres 27 
1.4 Structure, Contents and Textual History of the TGS ...........cscsscsssssessssesesens 27 
1.5 The Structure, Nature and Contents of the Nine Similes .0.........cccc cesses 34 
2 The Meaning and Occurrences of the Term tathagatagarbha 39 
2.1 The Term tathGgatagardhd ....cccccccccccscccesscscessssssseesscsecsessessessceateneeacsasensensens 39 
2.2 The Textual Occurrences of the Terms tathagatagarbha and garbha ......... 46 
3 The Buddha-Nature Doctrine in the 7GS . 50 
3:1 The Buddha-Nature on. scccsseccisduasessssesscssinsseelescaugeconestesdondaceavecsusesdaaeasvescsuecveses 50 
3.2 Becoming a Buddha ............cccecescssssessssssscssseesessescssessensssesessenessesecaeseeseeaceeses 62 
3.3 How to Become a Buddha ............cceecesessssessesssessesssscsssscscsesteneesessesessesteaeeness 65 

4 The 7GS as a Part of Indian Buddhism: Its Sources, Motives and 

Reception 67 
4.1. The Titles of the TGS: i... sess. cassasdaccsacatensnis ade teaadh alee tad 68 
4.2 The Recorded Chinese Translations of the TGS ........ccscscssescesseessecseneeneens 69 
4.3 Possible Motives of the Authors of the TGS .......ccssssssscssseceeeseereteesseecsnseees 75 
4.4 The TGS in the History of Indian Buddhism ..............cccsscscsesseseeectetsesseeens 77 
4,5 The TGS in the Ratnagotravibhdga(vyakhya) and Other Indian Texts ........ 84 
4.6 The Twentieth-Century Reception of the TGS ........ceeeccecesececeeeseeesenseeneneees 90 

B An Annotated Translation of the Tathagatagarbhasitra into English 93 

II Critical and Diplomatic Editions of the Tathaégatagarbhasutra 163 

C The Textual Materials 164 

1 Information on the Tibetan Manuscripts and Xylographic Editions 
Utilized 164 
A — The Tabo Manuscript Fragment .............c:cscescsssssesessesceseesseseessesseesesneeneenens 164 
B—The Berlin Manuscript Kanjur ......... cs cessessseecsseesscsseseseseeseeseessecceeeeeeareees 165 
Bth — The Newark Manuscript Kanjur from Bathang ................:ccsssssssseerseneens 166 
Bu — The Citation in Bu ston Rin chen grub’s De bzhin gshegs pa’i snying po 
gsal zhing mdzes par byed pa’i ryan rsesscserscsserssseeenesercesesesesscseesssoeeneees 167 

D —The Derge Kanjur (Nyingma Edition) ............ccccsssscsseseesceseeecseceeseeseeatees 167 

PDF Version: BPPB VI (2002) 

J— The "Jang sa tham or Lithang Kamjur 00.0... ccesessssscesssesesceesesssesesesesseneess 
L —The Shel dkar Manuscript Kanjur (London) .............ccccscscsessseesscessvesseees 
N— The Narthang Kanjur 0.0... cccssssssscssescscssssssssesseceseesesessssscecsessceseeseeeseeaes 
P), P2, P3 — The Phug brag Ms Kanjur ............:cccsssssssssessssesesssssssssssssesssssssssepeses 
Q ~ The Peking Kanjur (Otani Reprint) ............ccccsssssssseererssesescsscssecscsssscasenses 
S— The Stog Palace Manuscript Kanjur ...........cccssessssssssesssssseesessessssscssseseesees 
T — The Tokyo Manuscript Kanjur ...........ccccsecssssssssscsescssssssesessescsesesecsecssssesenses 

2 The Stemmatic Relations among the Representatives of 7ib 
2.1 The Three Phug brag Versions .............scccsssssssssssssessesssessssssseseeseneseenenestecens 
2.2 The Kanjurs of the Tshal pa Lineage 00.0.0... ccscssssssecsscssessesessesessscceeseeess 
2.3 The Them spangs ma Kanjurs .............:sscsssessssessssersssssestssessesseseseseesseseersees 
ZA The Position Of Bu ...5ciscicikces decsiesi ssvastcketests cectinsvacsceicdsdehits aeieiag ceeds hie 
2:5 The POsition Of: A | c2csciscce.decuscecstiveesessessecevenetdtased cseste aivcbedsstaaneaoherstectevseas 
2.6 The Relation of the Main Transmissional Groups to Each Other ............... 
2.7 Possible Sternmas Of Tib ........cecescssssscsssensesescesceccsscasceccsessessssecssenesnecaeeeeses 

3 Characteristics of the Textual Witnesses of 7ib 
Sil Archaic Features .scisecscsnssessacustscesesoceucoesscessostasteosaiois vscdadesagts asthe obs ckedeeadeaets 
3.2 Irregular Verbal Forms .........ccccssssssscsseressccsncssasesccssesotosssnssessnsasecensensesnescees 
3.3 Colophons and Translators ..........scsscsescsceceseseeserersessessssseeeseeseeseeescensesesesess 

4 A Brief Evaluation of the Chinese Materials 

5 Remarks on the Various Editions 
5.1 Principles Governing the Critical Edition of Tib oc. cessssssessecsseeesseeonees 
5.2 Remarks on All Editions and Their Critical Apparatuses ................ccc0000 
5.3 The Editions of the Tibetan Translations ..............::cccccssscssssessecssesesrecseeseesee 
5.4 The Chinese Editions ..............ccsccsssscssscreesecseeeertseesesnes dsiheetiiipuateccoeeetene 

D The Editions 

The Critical and Diplomatic Editions 00.0.0... cssscesssssseseesseesscseasscnensseteneeeeeasees 
Apparatus of Secondary Variants 0.00... cesecssssssessssesesssessseorsssssseeesssesensasseseseoes 
Sigla, Symbols and Graphic Devices of the Tibetan Material... eeeesees 
Sigla and Graphic Devices of the Chinese Material 2.00... ssssssssssessseseessesoees 

EK Appendices 

Appendix A: Comparative Chart of the Bodhisattva Names in OE .............00006 
Appendix B: Comparative Chart of Pada Sequences .........ccccsssssssesssseeeseseeeens 
Appendix C: Comparative Table of Sections and Chapters of the 

Chinese and Tibetan Editions ..............cssccssssssesesscsscerssrsesensenseiees 


Primary Sources, with Abbreviations .............:cccsccscssesssesseneseersersenceseseeeseeceesaees 
Select Secondary SOUrCES .......ccccssssssescsssssscsesssssssessnsersessnsenseeseeseenesesesscssssessecnsnes 














PDF Version: BPPB VI (2002) 


The TYathagatagarbhasutra (TGS) is a relatively short text that represents the 
starting point of a number of works in Indian Mahayana Buddhism centering 
around the idea that all living beings have the buddha-nature. The genesis of the 
term tathagatagarbha (in Tibetan de bzhin gshegs pa’i snying po, in Chinese rulai 
zang UK), the key term of this strand of Buddhism and the title of the sutra, 
can be observed in the textual history of the 7GS. From there it set off to enjoy a 
phenomenal career in Central and East Asia, revolutionizing the understanding of 
Buddhist thinkers and leaving a deep impression on Buddhist philosophy and 
spirituality. This impact can be felt even today. Nowadays, the concept of the 
buddha-nature attracts attention also among Buddhist practitioners in the West. 
Academically, the study of this third large strand of Mahayana Buddhism, which 
never became a dogmatically systematized line of thought, has been, and still is, 
dominated by scholars from East Asia. This, and the historical fact that it was 
doctrinally absorbed by the two main schools of Mahayana Buddhism on Indian 
soil, the Madhyamaka and the Vijnanavada, could be the main reason why the 
existence of tathagatagarbha thought has hardly been noticed in many Western 
academic surveys. In recent decades this has changed for the better, and it is 
probably no exaggeration to call the study of the theory of the buddha-nature 
today one of the main vehicles of exchange between scholars of the East and the 
West in the domain of Buddhist studies. 

Convinced that a detailed study of the earliest expressions of buddha- 
nature thought would be a rewarding task, I was first confronted with the need to 
establish a reliable textual edition of the TGS which, in light of the fact that no 
Sanskrit manuscripts of the TGS have been transmitted to us, takes into account 
all available translations. The collation of the Tibetan and the Chinese texts of the 
TGS by Kydshun T6dod in 1959, though progressive for his time, cannot meet 
these basic needs. My decision to produce a more comprehensive edition, one 
making use of the Tibetan materials, was based on three reasons. First, the Tibetan 
translators in general followed, more so than the Chinese, a literal style of 
translation, so that concrete inferences regarding the Indian text can be drawn for 
large parts of it. Secondly, the hitherto accepted assumption that the oldest 
available translation of the TGS, by Buddhabhadra, by virtue of its age and 
uniqueness, reflects an Indian transmission which has not undergone the textual 
alterations of later centuries, is only partly true. Certain additional textual blocks 
in the Tibetan were doubtlessly interpolated at a later stage, but the source of the 
citations in the Ratnagotravibhaga(vyakhya), a sastra which was written at least 
fifty years before Buddhabhadra translated the siitra, has turned out to be the 
recension represented in the Tibetan tradition.' Finally, given that the Ratnagotra- 
vibhaga(vyakhya), the most influential treatise on the tathagatagarbha theory in 
India, based itself on the recension reflected in the Tibetan, this recension became, 
in terms of the impact it had (its wirkungsgeschichte), increasingly referred to in 
the following centuries. 

As luck would have it, in the process of my work a second, paracanonical 
Tibetan translation of the same stitra, from Bathang, came to my notice. I present 
it in a diplomatic edition, directly facing the standard Tibetan translation. The 

' Whether or not the later interpolations were at that point already a part of this recension of the 
TGS cannot be decided. 

PDF Version: BPPB VI (2002) 

second part of my work contains critical editions of the canonical Tibetan 
translation, the diplomatic edition of the translation from Bathang and the two 
edited Chinese translations. As my translation of the siitra is based on the Tibetan 
canonical version, I invested much time in the critical edition of this latter, 
collating all philologically relevant manuscripts and block prints, among them 
even some folios from the now famous Tabo monastery. A description of the 
characteristics of the fourteen manuscripts and block prints and the conclusions 
drawn about their stemmatic relations form another important element of the 
second part of my study. With this I hope that a contribution has been made to the 
Kanjur studies started in the last decades of the twentieth century by the pioneers 
in the field, Helmut Eimer and Pau! Harrison. 

For the Sinologist dealing with Buddhist canonical materials, the second 
part may be of particular interest for its results concerning the transmission of 
Buddhist siitras in the Taisho Tripitaka. Although the text of the older Chinese 
translation has been transmitted fairly well in the Taisho, that of the eighth- 
century translation of Amoghavajra found in the Zhonghua Dazangjing 
rhe A ieLeS represents a much better transmission. Based on this text, the Taisho 
version had to be emended in a fairly high number of cases. The Zhonghua 
Dazangjing 373 should from now on be regarded as an indispensable tool 
for any critical edition of Chinese Buddhist material contained in it. 

The first part of my study consists of an annotated translation of the 
canonical Tibetan version, preceded by an analysis of the textual history of the 
TGS, an interpretation of the term tathagatagarbha, a discussion of the authors’ 
ideas as reflected in the siitra, and the specification of the place of the TGS in 
Indian Buddhist history. In order not to go beyond the framework adequate for a 
study of this kind, I have had to strictly limit my discussions to the TGS. Only 
sporadically have I been able to take into account facts which reflect a later stage 
of development of the theory of the buddha-nature. Some of the reflections, 
particularly in section 4.4, have therefore had to remain abridged, and are 
doubtless in need of much more detailed and comprehensive argumentation. I 
hope that the future will provide me the possibility of treating such points with the 
detail they surely deserve. 

The decision to compose this study in English was taken for the sake of 
readers who do not know German. I apologize for any unidiomatic phraseology, 
but nevertheless hope that the formulations are clear enough. 


I am indebted to a great number of people and institutions who enabled me to 
produce this study. I owe the most to my teacher Professor Lambert Schmithausen, 
under whom I studied up to my M.A. and who also agreed to become my 
doktorvater for the present work, which was submitted in a slightly different form 
to the Asien-Afrika-Institut at Hamburg University. His inspiring suggestions and 
valuable corrections, and generosity with his time, made possible its successful 
conclusion. It has always been a great pleasure to learn from him, and there is no 
suitable way to express my gratefulness. 

There are other scholars in Hamburg whom I wish to thank too. I am 
grateful to Professor Albrecht Wezler whose stimulating lectures and seminars on 

PDF Version: BPPB VI (2002) 

Indian philosophy I greatly appreciated, and who also supported me in other ways. 
On the Sinological side I am indebted to Professor Michael Friedrich. Since his 
coming to Hamburg, there has developed a very fruitful dialogue between 
Buddhologists of the departments of Sinology and Indology. Thanks are also due 
to Professor David Jackson and to Dr. Felix Erb, the librarian in the Hamburg 
institute. Dr. Harunaga Isaacson read through my work and offered important 
advice and suggestions. Parts of the book have been put into a better English by 
my friend Keish6 VK Leary and by Philip Pierce. The responsibility is, of course, 
mine. Again, my thanks to all of them. 

I stayed in Japan during most of the time of my doctoral work and had the 
privilege to be financially supported by two institutions: the Bukkyd Dendo 
Kyokai (A@5j8 HZ (“Society for the Promotion of Buddhism,” Tokyo) and 
the only recently established International Research Institute for Advanced 
Buddhology in Hachioji (Tokyo). I am very grateful for their generous support. In 
particular I would like to express my thankfulness to the scholars working at the 
institute in Hachioji for accepting this work as a volume in their series. In Japan I 
had the chance to meet with many eminent scholars, foremost among whom is the 
father of modern tathagatagarbha studies, Professor Takasaki Jikidd. On several 
occasions I had the honor to receive his guidance in this field. Thanks are also due 
to my supervisors in Japan: Professor Mimaki Katsumi in Kyoto and learned 
persons at the institute in Tokyo: Professor Kajiyama Yuichi, Professor Yuyama 
Akira and Professor Karashima Seishi. They all looked after both my academic 
and practical needs. 

In Kyoto I was invited to join the weekly study group under the guidance 
of Professor Nagao Gadjin. I am particularly indebted to him and all of the 
members of this group. Kyoto is also associated for me with Professor Aramaki 
Noritoshi and Professor Florin Deleanu, with both of whom I frequently met to 
discuss academic questions and whose acquaintanceship was highly beneficial. 

Equally advantageous were my meetings with Professor Hara Minoru, 
whose vast knowledge of Indian literature I can only admire. He opened to me the 
doors of the superb library of the International Institute for Buddhist Studies in 
Tokyo. My thanks are also due to Professor Yotsuya K6d6 and the staff of that 
library. Further I feel indebted to Dr. Helmut Eimer, Professor Enomoto Fumio, 
Professor Eli Franco, Professor Funayama Toru, Professor Kagawa Takao, 
Professor Kanno Hiroshi, Mr. Werner Knobel, Dr. Carsten Krause, Dr. 
Mudagamuwe Maithrimurthi, Professor Matsuda Kazunobu, Professor Matsumoto 
Shird, Professor Moriyama Seitetsu, Professor David Seyfort Ruegg, Professor 
Shimoda Masahiro, Dr. Jan-Ulrich Sobisch, Professor Toda Hirofumi, Dr. Eva 
Wilden and Professor Stefano Zacchetti, and would like to thank them for their 
counsel and support. Last but not least, I wish to thank the fellow students who 
have also contributed to the completion of my work, in the first place my friends 
in Hamburg: Mrs. Oma Almogi, Mr. Achim Bayer, Mr. Martin Delhey, Mr. Peter 
Lutum, Miss Nakamura Ayako, Miss Ozawa Akiko, Mr. Johannes Vagt, Mr. 
Dorji Wangchuk and Mrs. Yang Mei. 

PDF Version: BPPB VI (2002) 

Technical Note: My Use of the Asterisk 

I have employed an asterisk (*) before Sanskrit terms throughout the study when 
the Tibetan translation did not allow an unambiguous reconstruction of the 
original. In many cases, however, parallel passages in the Saddharmapundarika- 
sutra and other texts serve as proof for the reconstructed form, for which I then 
did not use an asterisk. Similarly, I have refrained from employing an asterisk 
whenever Bth and the Chinese translations support a certain reconstruction (based 
on Jib) and exclude any other possibilities. One should, nevertheless, keep in 
mind that Tibetan translation rules and vocabulary are much less regularly and 
mechanically employed than is usually assumed.* To speak with absolute 
certainty concerning the reconstruction of the Sanskrit terminology is impossible. 

? On this subject, see Seyfort Ruegg 1992: 382ff. 


PDF Version: BPPB VI (2002) 

Part I 

A Study and Annotated Translation 
of the 


PDF Version: BPPB VI (2002) 

A General Study and Text-historical Considerations 

In comparison with the vast amount of work to be done in the field of Indian 
Mahayana it would seem to be an easy task to focus on a single sutra. 
Nevertheless, even such a limited undertaking has the potential to extend its reach 
significantly, inasmuch as research of this kind touches on several quite different 
topics of scholarly activity. One is the topic of the Tibetan textual transmission 
itself which will be dealt with extensively in part II. In part A I will focus on text- 
historical considerations and present a study of the fundamental terminology used 
in the text, the doctrine it expounds, and its place in Indian Buddhism. It goes 
without saying that, in order to cover these different aspects, I have had to restrict 
myself to what appeared most essential. In the following I would like to give a 
very short overview of the content of part A, and, at the same time, to draw 
attention to what I consider the most important outcome of my research in it. 

In chapter 1 I first deal with the interrelationship among the four oldest 
translations of the TGS (1.1). Three of them, it turned out, represent the same 
recension (TGS2), while the older Chinese translation is the single representative 
of another, shorter recension (TGS,). From a detailed analysis of the relation of 
the two recensions to the nine similes reproduced in the Ratnagotravibhaga, it 
was possible to prove that, though TGS, was the earlier translation (fifth century), 
the similes in the Ratnagotravibhaga, itself a work which probably came into 
existence as early as the fourth century, are based on TGS). Section 1.1 further 
contains a study of the main terminological differences between the two 
recensions, and 1.2 and 1.3 deal with the differences between the three 
representatives of TGS». Each translation of this latter has its own unique features, 
but no telling traits could be found that would establish their interdependence. 

Section 1.4 offers an analysis of the structure and textual history of the 
sutra. The main result of this chapter lies in localizing and proposing explanations 
for compositional irregularities in the first simile. I could thereby show that the 
textual history of the TGS can be divided into at least three steps. The term 
tathagatagarbha was only introduced into the siitra during the third step, as an 
insertion into the middle of the first simile. In the last eight similes, which one can 
consider the starting point and nucleus of the siitra, there are no occurrences of 
this term. 

Section 1.5 ends the first chapter. I deal there with the basic structure of 
the similes and discuss an adequate method of interpreting them. The content of 
each simile is summed up. By this means the richness and varying nuances of the 
authors’ message in them are brought to light. 

Chapter 2 is devoted to the terms tathdgatagarbha and garbha. In a 
comprehensive grammatical analysis in section 2.1 I discuss all intelligible ways 
of understanding the compound tathdgatagarbha. The most fitting interpretation 
in the context of the TGS is that of a bahuvrihi meaning “containing a tathagata,” 
in reference to living beings. The context of a growing process suggested by two 
of the similes and the parallelism with the term buddhadhatu of the 
Mahaparinirvanasitra may, however, have contributed to the prevailingly 

PDF Version: BPPB VI (2002) 

tatpurusa use of the compound in the later tradition. 

In section 2.2 I first analyze all passages in which the compound 
tathagatagarbha appears, and suggest an answer to the question why the Tibetan 
translators did not apply the particle can in their rendering of the compound. The 
most significant findings in this section center around two difficult passages of the 
sutra concerning the term fathagatagarbha. For one passage (verse 1.1), I could 
demonstrate with the help of a parallel in the Ratnagotravibhaga that the verse 
was understood wrongly by the Tibetan tradition. In the second case (5A) there is 
strong textual evidence from the Chinese and the alternative Tibetan translation 
from Bathang: instead of the correct tathagata-jfiana, we find in the canonical 
Tibetan translation a rendering of °garbha, a mistake which may even go back to 
the Sanskrit manuscript used by the translators. Another verse (10.10), however, 
clearly documents a non-bahuvrihi use of tathdgatagarbha. There are several 
indications, though, that this verse was not part of the oldest form of the TGS. 
Regarding the use of the term garbha, it seems that, contrary to the practice in the 
translation from Bathang, the canonical team of translators took care to 
differentiate between garbha as the space inside the calyx of a lotus 
(padmagarbha, pad ma’i snying po) and the petals forming its enclosure 
(*padmakosa, pad ma’i sbubs). With equal rigor, on the other side of the simile 
(upameya), snying po was used in the canonical Tibetan translation only in the 
meaning “embryonic essence” of living beings but not as indicating their “inside” 
or “womb.” 

At the beginning of section 3.1 I provide a chart with all the terms of the 
TGS used to designate the buddha-nature of living beings. Among them a variety 
of expressions ranging from concrete terms like jinakaya to more abstract ones is 
then analyzed. Yet all the terms are interchangeably employed by the authors of 
the siitra in combination with the concrete vocabulary of the upamana, a fact that 
implies that their main aim was to convey the idea of the inherent buddhahood of 
all living beings in easily comprehensible terms rather than to go into a detailed 
scholastic discussion of the exact properties of this hidden buddhahood. The terms 
(tathagata-)dharmata and tathagatadhatu are dealt with at greater length. I have 
tried to make clear why we can by no means be sure that the former term is meant 
to be understood as “absolute truth” similarly to tathata (it must rather mean 
“character/nature (of a tathagata)” without any monistic connotations), and why 
the term dhdtu in the passage of the TGS is bare of any notion of causality—a 
notion which does, however, characterize the Mahaparinirvanasitra and the later 
tathagatagarbha tradition. (The passage containing dhdatu in the TGS is cited in 
the Ratnagotravibhagavyakhya, and there used to prove that dhdatu should be 
understood as hetu.) From my analysis it thus becomes clear that all the terms for 
the buddha-nature employed in the TGS refer to the level of spiritual perfection. 
They indicate a state or entity which is already present and in no need of any 
further ripening or essential change, though two of the similes could at first glance 
provoke a different understanding. 

To show in detail why I think these two similes were not created in order 
to introduce a notion of ripening into the idea of the buddha-nature in all living 
beings is one concern of section 3.2. The idea the authors of the siitra had in mind 


PDF Version: BPPB VI (2002) 

can be called a “theory of revelation.” The perfect state of buddhahood is already 
present in all living beings. It is nevertheless covered with defilements and has to 
be revealed or manifested in order to become efficacious. In fact, the main 
category on which the authors’ ideas center is that of efficaciousness. 
Efficaciousness functions as the decisive criterion which distinguishes the buddha 
from the sattva, the authors’ focus on and illustration of this one category by far 
outweighing any concern for a neat and systematic conception of the two different 
soteriological states. Here the siitra reveals its true character: rather than pose the 
question of how the new message might be philosophically brought in line with 
other, traditional Buddhist doctrines, and formulate or discuss a possible 
metaphysical principle going beyond the concrete frame of an individual buddha- 
nature in the similes, the authors of the text seem to have pursued fairly pragmatic 
ends. This accords well with the fact that for them, becoming and being a buddha 
means in the first place to actively perform buddha-deeds. All these observations 
have led me to argue that the stitra’s origin has probably to be sought among 
circles of Buddhists characterized by an “attitude of worldly engagement 
predominating over mainly theoretical concerns” —a hypothesis which, of course, 
needs to be further substantiated and elaborated on the basis of other writings of 
the tathagatagarbha teachings. 

In section 3.3 I try to shed light on the process of purification from the 
defilements. While the * Tathdgatotpattisambhavanirdesa, which in view of one 
of its similes can be seen as the closest predecessor of the TGS, distinguishes 
clearly between the role of the Tathagata in the liberation process of living beings, 
on the one hand, and the active participation of living beings themselves on the 
other, the TGS is in this respect more vague, so that sometimes the part living 
beings play recedes completely behind what indeed appears to be the almighty 
figure of the Tathagata. Nevertheless, there are passages in the TGS which 
document that it was by no means its aim to exclude living beings from such 
participation. Rather, it was simply that the authors’ focus lay on other issues. 

Section 4.1 is an overview of the titles under which the siitra has been 
translated and cited. The oldest title of the stitra in Sanskrit most probably was 
Tathagatagarbha-nama-vaipulya(or: vaitulya)-sutra. 

In section 4.2 I deal with the oldest available catalogues of translated 
Buddhist literature in China with regard to the records on the 7GS. At the center 
of my analysis is the question of the plausibility of an entry in the Chu sanzang ji 
Ji HK =32Ce which claims the existence of a Chinese translation of the TGS for 
as early as the end of the third century CE by Faju }&¢8. Whereas the attribution 
of another translation of the TGS to Fazu ¥&4H (active ca 290-306), in the 
catalogue of Fei Changfang #/#, most probably does not correspond to 
historical fact, the situation regarding Faju is more complex. Much depends on the 
evaluation of the credibility of the entries in the Chu sanzang ji ji said to be 
largely based on the famous but lost catalogue of Daoan i. My study shows 
that there are good reasons to deal more critically with this part of the Chu 
sanzang ji ji. Its compiler Sengyou {44 has always enjoyed the highest esteem 
among scholars who work with the difficult material of Chinese Buddhist 
catalogues. This may have led to a certain lack of judiciousness when it came to 

PDF Version: BPPB VI (2002) 

the reconstructed entries of Daoan. Although I can show that the Chu sanzang ji ji 
probably knew of two conflicting traditions concerning the history of Chinese 
translations of the TGS, the claim that a translation of the sitra already existed in 
the third century is not in itself at all unreasonable. Concerning the translations by 
Buddhabhadra and Amoghavajra no further material for their exact dating could 
be found. 

As for the motives of the authors of the TGS (section 4.3) there are a 
variety of possibilities. The non-philosophical and non-scholastic style hints at the 
likelihood that they were writing primarily for non-specialists of the Buddhist 
doctrine. One of the motives could thus have been to attract to religious practice 
hitherto neglected segments of the Buddhist community, or even groups from 
outside it. The Saddharmapundarikasutra may have been another important factor 
leading to the composition of the TGS. The Saddharmapundarikasutra puts 
forward the ekayana theory, namely that all living beings can become buddhas. 
The TGS, from the mouth of the Buddha, provides a sound, if ideal, soteriological 
“proof” of this assertion. The fact, however, that nowhere in the siitra are there 
ethical conclusions drawn on the basis of this ekayanist theory of equality 
somehow comes as a surprise. 

In the first part of section 4.4 I try to establish a terminus ante quem for 
the TGS. Based on considerations of a doctrinal nature, the similar metrical 
structure of parts of the Ratnagotravibhaga and the Mahayanasutralamkara, the 
progress of philosophical thought from the Mahdayanasutralamkara to 
Vasubandhu’s Trimsika, and the fact that a citation from the Trimsika is found in 
the Lankavatarasutra, the middle of the fourth century CE is a probable date 
before which the TGS should have been composed. The 7GS, having clearly taken 
ideas from the *Tathdagatotpattisambhavanirdesa and the Saddharmapundarika- 
sutra, must be attributed a position within a strand in Buddhism which does not 
conceal its favor for the concept of a positive continuous subject—a strand which 
has its roots in the canonical scriptures. The notion of sunyata in its established 
meaning does not play any decisive role in the 7GS or in other early 
tathagatagarbha texts. According to the Ratnagotravibhaga, the tathagatagarbha 
teaching can be understood as complementing or even correcting ideas related to 
Sunyata. Though the TGS does not present a philosophically homogeneous 
framework, and even if its main impetus may indeed have been to encourage 
people to become active Buddhist followers, its soteriological conception of a 
buddha-nature present in all living beings cannot easily be discarded as mere 
upaya, that is, as pure means apart from any claim that the message embraces a 
true statement about the constitution of reality. Such a disclaimer would contradict 
the thrust of the authors’ work, even if their idea of living beings’ eternal 
buddhahood on the road to efficaciousness comes close to the “non-Buddhist” 
dtman doctrine, and can in the eyes of somebody aiming to doctrinally harmonize 
Mahayana provoke doubts about the Buddhist character of the tathagatagarbha 
teaching. This is exactly what has been done by the representatives of the “Critical 
Buddhism” group, with whom I shortly deal at the end. I feel that they are not 
willing to accept that Buddhism from early times on has been a heterogeneous 
phenomenon, one impossible to identify by two main tenets of purely doctrinal 

PDF Version: BPPB VI (2002) 

matter, in the way its critics do. I further believe that what they postulate to be 
“true Buddhism,” that is, “Critical Buddhism” in the sense they understand it, is 
another inadmissible restraint of Buddhist traditions. 

Section 4.5 tries to identify the texts in which citations from the TGS are 
contained. The Ratnagotravibhaga(vyakhya) is the most important among them. 
Though the Ratnagotravibhaga has rendered faithfully the content of the nine 
similes of the TGS, there is at least one instance where it has introduced a notion 
not found in the parallel passage of the TGS. This notion can be called 
“traditionally Buddhist” inasmuch as it emphasizes the necessity of several factors 
for the ripening of a sprout, which here illustrates the living being’s buddha- 
nature. In the vyakhya, however, one of the very few word-for-word citations from 
the TGS is interpreted differently from the Ratnagotravibhaga itself. Further, the 
vyakhya has tried to apply an unnatural scheme to the similes and has reduced 
their richness of descriptive nuance to a purely scholastic analysis focused on the 
main categories of buddhadhatu and klesa and their subcategories. 

Finally, in section 4.6 I describe the history of the twentieth-century 
reception of the TGS. The studies of the tathagatagarbha teaching in general have 
experienced a rapid acceleration since the 1960s. Both the Chinese and Tibetan 
translations of the TGS have been studied under various approaches aimed at 
different groups of readers. 

1 Textual History and Structure of the TGS 
1.1 Different Recensions of the TGS 

The 7GS can be said to exist in two recensions: 
* Recension TGS;, represented by the Chinese translation of Buddhabhadra 
dating from the beginning of the fifth century CE (Ch,), and 
* Recension TGS?, represented by the Tibetan canonical translation dating from 
around 800 CE (7ib),' the Tibetan paracanonical translation from Bathang 
(Bth), and the Chinese translation of Amoghavajra from about the middle of 
the eighth century (Chz). 
Differences between TGS; and TGS; are documented in the notes on every page of 
my translation. Among these differences are several major ones of a purely 
quantitative nature: 
* Ch, lacks the list of attributes of the arhats in OB. 
* The whole section 0C (the enumeration of participating monks) is missing in 
¢ Ch, lacks verse 0.5. 
* The repetitive passage at the end of 10B and the beginning of 10C is missing 
in Ch). 
¢ The introduction to the story of *Sadapramuktarasmi in 11A is Ch; 
(the passage in question has no counterpart in the verse portion of either 
* Sections 12A and 12B (the question of Ananda) are missing in Ch. 

1 On the authorship of the Tibetan canonical translation according to the colophons, see section 
3.3 in part Il. 


PDF Version: BPPB VI (2002) 

All these differences can be shown to be additions of TGS, rather than parts 
dropped by Buddhabhadra.’ In the following, I will deal with possible answers to 
the question why the above-mentioned passages were introduced into TGS). 

The list of attributes of the arhats in OB is a stock phrase description found 
in many other siitras as well. The addition shows that from a certain time on it was 
probably thought of as a requisite element.’ The same can be said for the addition 
OC, where 19 names of participating monks are listed. The best-known disciples 
of the Buddha are among the arhats. The mention of Ananda was probably felt to 
be especially important, since he later questions the Buddha in 12A and 12B, 
sections missing in Ch. 

Verse 0.5 goes beyond the rather rational question in the parallel prose 
section OJ. It is an emotional appeal to the Buddha, who is addressed by the 
epithets “Highest among Humans” (dvipadottama) and “Divine One” (deva), and 
requested to answer for the sake of all living beings. That the Buddha has to be 
requested (as many as three times) to teach is also a common element in many 
siitras." The convention goes back to the early sources, which report that the 
Buddha was asked three times to teach by the god Brahman before he decided to 
do so. It is surprising that the parallel prose section remained untouched by the 
redactors of TGS). 

In the case of the additional part in 10B/C, I cannot find any convincing 
reason why the redactors of TGS? felt the need to enlarge the text. The veneration 
of the tathagatas in the repetitive section is reduced to the strewing of flowers (as 
opposed to the donation of pavilions in the part common to TGS; and TGS2). In 
the interpolation, instead of the realization of tathagata-knowledge as the starting 
point, we find the search for the Dharma (dharmam paryesate); the “sons or 
daughters of good family” are replaced by “monks, nuns, updsakas or upasikas”; 
finally, instead of the internalization or arrangement into a book of as little as one 
simile, there is the veneration and joyful approval (anumodana) of the sitra. This 
last feature is typical of the Saddharmapundarikasutra (see SP; s.v. forms of the 
stem anu-mud) and other Mahayana sitras, and I suspect that this may have 
induced the redactors to interpolate the repetitive section into TGS). 

The introductory passage to the story of *Sadapramuktarasmi in 114A, 
stressing the beneficial effect of the TGS, resembles in its syntax the introduction 
to the story of Sadaparibhiita in the XIXth chapter of the Saddharmapundarika- 
sutra (see note in my translation). As it is not found in Ch,, it could well be an 
element, again based on a set phrase found in the Saddharmapundarikasutra, 
inserted later in order to smooth the transition from the sections on merit to the 
story of *Sadapramuktarasmi. 

Sections 12A and 12B are not found in Ch,, nor do they have a counterpart 
in the following verses of either recension. They contain Ananda’s question about 
the length of time till perfection and the Buddha’s answer. Explanations directed 
to Ananda are a common element in Mahayana sitras. This is understandable, 
since it is he alone who is generally said not to have attained arhatship among the 
Buddha’s main disciples. Besides, he is often entrusted with the preservation of 

? The additions are discussed in the notes to the corresponding passages in my translation. 

3 The same situation holds true for the Vimalakirtinirdesa. The list of attributes is found only in 
the Tibetan (see Lamotte 1962: 97). ; 

‘ See, for example, the second chapter of the SP where Sariputra requests the Buddha three 
times to expound the Dharma. 

PDF Version: BPPB VI (2002) 

the expounded discourse (at the end of the siitra), having been, according to the 
tradition, known for his excellent memory.° 

The position of the interpolation towards the end of the siitra is thus not 
surprising, though the passage in question does not directly deal with the 
preservation of the TGS. The redactors placed it within the originally monolithic 
verse section 11I and 12C. In order to join the interpolation neatly with the 
following verse triplet, they added a sentence at the end of the interpolation 
dealing with the worthiness of whoever preserves the TGS, which is the theme of 
the following three verses. The content of Ananda’s question, on the other hand, 
seems rather unique. One would have expected the stock question about the 
title(s) of the exposition® if the purpose had been to create a context for the 
appearance of Ananda. As a matter of fact, the relation of the question to the other 
parts of the TGS is not obvious, and I suspect that Ananda’s question was 
originally simply another more or less independent episodical unit before it 
became integrated into the TGS. 

The fact that Ch; covers just two thirds of the length of Ch; cannot simply be 
explained by these obvious interpolations, and the reasons must also be looked for 
elsewhere. When it comes to differences of style and content, the situation is 
much more complex. Although the same basic ideas can be found in both 
recensions, it seems that in the details, that is, terminology and syntactic relations, 
the recensions vary widely. The style of Ch; appears to be much more concise: the 
number of characters per pada is only five (Ch: seven), enumerations are 
generally abbreviated, and the line of narration seems to focus only on issues of 
main importance. My impression is that Ch, sticks less slavishly to an Indian 
original and, in contrast to the representatives of TGS2, also reproduces almost no 
inconsistencies which may already have been part of the Indian transmission. 

Besides these characteristics we find other terminological and doctrinal 
features which, more or less, seem to be peculiar to Ch,: 

a, Throughout the similes, the compound 402K}s% (or: (#:9%), usually rendering 
tathagatagarbha, is used 20 times in Ch,. In the Tibetan, the usage is restricted 
to the first simile, where it appears 4 times. In Ch, besides these 4 times in the 
first simile, it is found a further 5 times in other similes.’ In cases where Ch, 
employs 41kyg (or: (#548) in contrast to the Tibetan, the Tibetan is based on 
tathagata-dharmata (4 times), *jina-kaya (or: sugata’, tathagata’; 4 times), 
tathagata or buddha (3 times), buddhabhumi, buddhatva (sangs rgyas nyid) ot 
*avinaSadharmin (chud mi za ba’i chos can). In one passage of the first simile, 
Ch, reads #036 in contrast to tathagatagarbha in the Tibetan and Ch). 

b. In Ch; the term 402R%% (tathdgatagarbha) designates a separate entity found 
within living beings. ° Amoghavajra in Ch; seems to share such an 

5 See Lamotte 1962: 392, n. 40; Buddhist tradition has it that Ananda took a vow to become the 
first of “those who have heard a lot” (bahusruta) (MPPU, I 223). 

§ See Lamotte 1962: 392, n. 41. 

7 See the comparative list in Takasaki 1974: 48-53 (SHYBNSIMALH MH on p. 50 (III.2) is to be 
corrected to 407K) . 

* See OM: SHAM; 1.3: PERIRAI FAA AURA: 1.5: GRRE; 2A: —RAB 
ROAR; 2.2: FRA ANA (“the tathdgatagarbha [of] living beings”); 5B: #6384: A ADAIR; 6.2: 
#ei7ES. Only 8.4 and 1B could be interpreted differently: 745 403% (8.4): “Your body [is] 


PDF Version: BPPB VI (2002) 

c. Ch; tends to describe only the buddha-nature of living beings as wrapped in 
defilements.!° TGS , on the other hand, so describes both the buddha-nature 
and living beings.!! 

d. A translation of the term dharmata appears only once throughout the whole of 
Ch, while its usage is frequently attested for TGS, (chos nyid, 4\£). This 
single usage of dharmata in Ch; is found in the passage quoted in the 
Ratnagotravibhagavyakhya (see 1B): 4. The Tibetan employs the term chos 
nyid (for dharmata) altogether in 9 instances, for 7 of which there is a 
corresponding text in Ch,. In these 7 cases, Ch; uses the terms 424i (4) and 
KIB (3). Chz in these passages operates with 4 (4) and 34 (verse), Ei, 
and 428% (each once). In the two other passages Ch reads 4 and 

e. The term “bodhisattva” appears 34 times throughout 7ib.'? Ch, however, uses 
the term only 21 times. Among these are 3 cases where TGS? does not mention 
the term (7.5, 10B, 10.3). In 7 instances Ch; uses more general terms like #@4- 
(sattva) or fa instead. Five cases are without any correspondence in Ch). 

f. While Jib uses the stock phrase “Tathagata, Honorable One and Perfectly 
Awakened One” 14 times, it is only found 3 times in Ch), and in these once in 
a passage where all versions of TGS read only “tathagata” (6B). Otherwise, 
Ch, resorts to the terms 42% (3), #& (2), (#8 (2), or t+ (1), and when TGS, 
speaks of the attainment of the designations “Tathagata, Honorable One and 

the store of a tathagata!” However, such an understanding is unlikely in view of the preceding 
verse 8.3: SHRAIAGK. In 1B, FeRAM, Beet, BR, — OREM RE 
7x, corresponds to the citation in the RGVV (... sadaivaite sattvas tathagatagarbha iti /). The 
Chinese can be understood as “The store of a tathagata [in] all living beings is at all times present 
without change” or possibly “All living beings [are] stores of a tathagata; [this true nature of all 
things] is at all times present without change.” The latter translation is unlikely because the subject 
(48 for dharmata) and its predication (7{+7* #) appear separated from each other. 

” 402i as a separate entity appears in 1.1 (EAURARATYE), 3B CRAIRIEE), 5B 
following cases could at first sight be understood as “... [to see that] all sentient beings [are] stores 
of a tathagata”: OM: FI— a 1eaIRAR, 1A: ROA AIR, 1B: — A aOR 
7a, OB: F— 4a AAO. Only an examination of other passages, similar in syntax, shows 
that the grammatical function of 412% cannot be that of a predicate to @{# and that the 
expression must consequently be translated as: “tathagata-store [of/in] sentient beings.” That this is 
the right interpretation becomes clear from the following passages: OM: %—W#q BAK Meee & Se 
ABE 1S; 6A: FI— 4g QRH EBA EW. These are the only other examples with a syntactic 
structure equivalent to the one above. It is obvious that the defilements cannot be construed as the 
predicates of sentient beings. On the other hand, we find the following pada in 4.3: S0(F(R1SS 
3¢ (88: “knowing that those defilements [are] accidental.” From this we can infer that Amoghavajra 
would probably have employed the character £% in the instances above also, if he had intended to 
use the term 4[13R# as a predicate to sentient beings. 

10 Namely in OM, 2A, 2.2, 3.3, 4.3, 4.4, 6.3, 7.4, 7.5 and 9.3. The passages in 1B ({H{#384- 48 
TSE, AOR TH...) and in 7.3 CRERINA BRARAA ERAS) suggest that living 
beings are encased in defilements. 

" Passages where the buddha-nature is mentioned as wrapped in defilements are 2.2, 3.3, 7.4 
and 7.5. The defilements are said to cover living beings in 0M, 1B (Bth and Tib unclear), 2A, 4.3 
and 7.3. 

'2T am not counting the appearances in association with the names of bodhisattvas or with the 
four bodhisattvas who did not attain awakening in section 11. Among the 34 cases above, the term 
is 4 times part of an interpolation in TGS, not found in Ch,. 


PDF Version: BPPB VI (2002) 

Perfectly Awakened One,” Ch; paraphrases it as KM #B, BRIER or SF 
ned ey. 13 

g. The term sattvadhatu and the definition of sattva is not found in Ch, in 6B. In 
addition, the verses in chapter 4 of Ch, lack the term (citta)prakyti — 
agantukaklesa, a concept characteristic of the tathagatagarbha teaching. 

It is common in both Chinese translations that 41} is understood to be a 
separate entity found in living beings (b). Leaving aside the question why the 
word zang ¥& was chosen to render Skt. garbha," this is nevertheless surprising, 
since 4/3K% could easily be applied as a predicate of living beings, in the sense 
of living beings being the stores of a tathagata. It seems that already at the time of 
Buddhabhadra’s translation the general idea of tathagatagarbha as an entity 
within living beings was so prevalent that Buddhabhadra or his collaborators 
decided—even against the evidence of the siitra itself—to accept it for their 
translation into Chinese. This feature thus arose in the translational act and is not 
related to the Sanskrit text. 

It is impossible to say to what extent feature a, the continuous use of the 
compound 44kiX throughout Ch,, is also due to Buddhabhadra’s preference for 
this term.’° The fact that Amoghavajra also uses the term more often than the 
Tibetan translation does suggest that one reason for its prominence could be the 
wider range of its applicability once interpreted as a separate entity: whereas the 
term in the Sanskrit was only applied as a bahuvrihi relating to sentient beings, 
the Chinese now had the chance to replace other terms designating the buddha- 
nature in living beings with it. It could at the same time acknowledge the possible 
“popularity” of the term, which, after all, formed the sttra’s title. The Tibetans, 
favoring a literal translation, refrained from an abundant use of the compound 
and, in this instance, probably faithfully followed the Indian text. It is, however, 
true that I cannot completely exclude the possibility that the Indian TGS, had 
already taken the term tathdgatagarbha in the same sense as seen in Ch), and that 
it appeared more often in it than in TGS;. Whatever the case, a development from 
the usage of a variety of designations for the buddha-nature towards the 
standardized term 43K} is definitely a more plausible assumption than the other 
way round. Speaking in text-historical terms, Ch; in this instance is a recension 
further removed from a common Indian original than TGS). 

As I will show below, the term dharmatd in the TGS is used in at least two 
different meanings.'® In the passage where Buddhabhadra employs the translation 
YEE dharmata means “the rule to which all dharmas are subject.” It is the only 
passage in all of TGS, with dharmata in this meaning, and I suspect that 

3 Tn OB (the passage concerned is not found in Ch; in its entirety). 

4 Regarding this question, see Hirakawa 1990: 73ff. There it is left open whether 
Buddhabhadra adopted an earlier translation ((#3) or created one on his own (p. 76). One crucial 
impact on the rendering of tathdgatagarbha with rulai zang QOH (indicating an element in 
living beings) may have had the concept of “five internal organs” wu zang Flix in traditional 
Chinese medical theory. Medical works such as the Suwen 38] and the Lingshu SEAR are familiar 
with this concept, and it must therefore have been in use as early as the Han J dynasty (see 
‘Zhong Yi Dacidian’ Bianji Weiyuanhui «PEAR» MBAS (ed.), Zhong yi dacidian: 
Jichu lilun fence HER ARE: SE 4H, Peking: Renming Weisheng, 1982, pp. 42; 74ff.). 

5 Takasaki (1974: 53) even calls the use of the term £1%# by Buddhabhadra “excessive.” 

'6 The term is also employed in the parallel verses in RGV 1.100: vilokya tadvat sugatah 
svadharmatam avicisamsthesv api... 


PDF Version: BPPB VI (2002) 

Buddhabhadra in his translation chose the characters 4 in order to make clear 
that here dharmata should be understood in its “weak” meaning of “general rule.” 
For him, dharmata expressed with the characters 7¢'# likely implied the sense of 
“absolute truth” (see below). This may have led him to refrain from the use of 
YE in the other passages. The meaning of dharmata in TGS, except in the 
passage just mentioned and in another case (see below), is probably simply that of 
the “nature” of a tathagata. Hence Buddhabhadra probably felt that the characters 
YE went too far. He therefore avoided the translation } and employed other 
terms specific to the TGS, such as RAF, instead.!” 

Feature c arises from the fact that in most cases Ch; states that the buddha- 
nature (and not living beings) is wrapped in defilements. This is in line with the 
corresponding verses of the Ratnagotravibhaga, where living beings are only 
once described as covered by defilements.'* Of course, a “correct” allegorical 
interpretation would tend to require that the buddha-nature, as counterpart of the 
tathagata-bodies in the lotuses, be taken as the element covered by defilements 
rather than living beings themselves. However, no statement can be made in this 
regard about the degree of “correctness” of the Indian archetype of the TGS. 

In regard to the replacements for the term “bodhisattva” with less specific 
terms (e), further text-historical research needs to be done on Mahayana sitra 
literature in order to establish a comprehensive frame of possible general 
tendencies in the transmissional history of this genre.'? The less frequent use of 
the epithets “Tathagata, Honorable One and Perfectly Awakened One” in Ch, (f) 
probably has its origins in Buddhabhadra’s concise style, and cannot help in 
reaching conclusions about the relation to the Indian recensions, 

Finally, the occurrence of the terms sattvadhdtu and (citta)pra- 
krtilagantukaklesa” only in TGS (g) could point to a later addition. The terms are 
well known for their prominent position within the tathagatagarbha theory, and 
for this reason it is hardly imaginable that such key words would be dropped in 
the process of redaction or translation. On the other hand, as Ch, tends to 
standardize to a certain extent, this cannot be maintained without restriction. 

This discussion of the unique features of Ch, reveals the problems in 
assessing this translation and its relation to TGS2. We simply cannot know if it is 
the translator (and his team) or an earlier redaction of the text in India that was 
responsible for the differences. Moreover, it is very likely that the work of 
Buddhabhadra itself was exposed to a redaction by Chinese specialists 
immediately after its completion, in the process of which many passages could 
have been made to conform to basic rules of logic or to demands for a more 

" That 3 and other related terms suggesting aspects of absoluteness were in fact vividly 
discussed in the period in which Buddhabhadra was active we can see from the documentation of 
the different views of KumArajiva and Huiyuan #3 on this subject in Ren 1985: 694-701. 

18 See RGV 1.101: ragadvesamaladikosanivytam ... jagat. The buddha-nature as enclosed in 
defilements is found in I.106, 107, 110, 111, 116, 119 and 120. 

9 For one example of this kind, see Shimoda 1997: 171ff. For the Mahayanist 
Mahaparinirvanasiitra he shows that the word “bodhisattva” is consistently used only in certain 
sections of the text. Further, he observes that a “bodhisattvaization in terminology” appears at 
three levels of intensification depending on the different versions, i.e., from the first Chinese 
translation (416-418 CE) by Faxian 7£ #4 to the Tibetan translation, and finally the second Chinese 
translation by *Dharmaksema (also first quarter of the fifth century, but much longer than the 
translation by Faxian) with the highest number of appearances of the term “bodhisattva.” 

© See 1.125 for similar terminology in the verses of the TGS reproduced in the RGV. 


PDF Version: BPPB VI (2002) 

appealing Chinese style.”' Other translations bearing the name of Buddhabhadra, 
ones for which the Sanskrit is available, suggest that they have been done very 
carefully, more or less literally following the Sanskrit and rarely containing 
elements introduced through the translator’s peculiar way of understanding the 
text.” However, even if translations bear the name of Buddhabhadra, can we be 
sure that he cooperated always with the same Chinese scholar-monks who were in 
charge of the redaction of his work? Should we not expect that his coworkers 
were chosen according to the importance of the text and that the outcome, for a 
large part, exhibits their own personal style?” 

If we nevertheless assume that the redactional impact of the translation 
activities was only minimal, would it be possible to explain some of the 
differences over against TGS? as the features of an orally transmitted text? Indeed, 
a memorized text would naturally tend to be smoothed down in the process of 
unconsciously adapting content and structure of the various passages to the main 
line of thought and reoccurring patterns. From this we could explain why in 
several passages Ch; does not loose its thread, whereas the versions of TGS), in 
several passages, seem to be altogether inconsistent. Also, the reason for the 
frequent occurrences of the compound 413K could be found here, since one can 
assume that the central term of the stitra, and a constituent part of its title, crept 
into passages and replaced other designations of the buddha-nature. On the other 
hand, one could easily argue in the opposite direction, namely that a memorized 
text would hardly allow for alterations to the degree documented in TGS), since 
once memorized, the text would attain a more or less fixed form. Though I do not 
want to exclude the possibility that changes in the siitra were caused by an oral 
tradition, I rather tend to look for the reason for the origination of TGS; in the 
Chinese redaction of the text by Buddhabhadra and his coworkers. 

Regarding the relation of Ch; to TGS», it is obvious that Ch; represents an 
earlier stage of textual development of the TGS as far as the interpolated passages 
mentioned above are concerned. I am hesitant to claim that this kind of relation is 
valid for the parts contained in both recensions which are different in structure, 
terminology and length. These differences do not seem to be the result of a 
consistent redactional approach, and the wording in neither recension can be said 
with certainty to be earlier and more original than that in the other. It is thus 

1 An analysis of the Chu sanzang ji ji tH=i#2042 (compiled ca 515 CE) shows that, in 
general, the teams dealing with the translation were composed of three specialists: the main 
“translator” and reciter, the actual translator (if the reciter’s Chinese was not sufficient), and one or 
more redactors. See Zacchetti 1996: 350-52; and also Boucher 2001: 104ff. along with the works 
he provides in on. 35. Zacchetti further points out that, while in the case of the 
Mahdaparinirvanasutra (417 CE) Buddhabhadra was supported in his activity by the frequently 
mentioned translator Baoyun #38, only a short time later, when he translated the 
Buddhavatamsakasiutra, Baoyun was no longer mentioned (pp. 353-355). I am not sure if one 
should follow Zacchetti in his opinion that this must mean that Buddhabhadra had gained enough 
knowledge of Chinese to translate without help (p. 355). 

2 This holds true for Buddhabhadra’s translation of a part of the TUSN—the Sanskrit of which 
we know as a quotation in the RGVV (22.10-24.8). The date of the translation of the TGS 
(beginning of the fifth century) falls within the period of Buddhabhadra’s translation of the 
Buddhdavatamsakasitra, of which the TUSN is a part (418-420 CE; see DZJ sv. KATAIA 

33 Zacchetti notes that the central function of the main translator was to recite and interpret the 
text, whereas the translational work itself was mostly done by the actual translator of the team. 
From the Buddhist point of view, he argues, the translation would nevertheless be attributed to the 
reciter, because he was the one familiar with the tradition of Indian thought (pp. 358f.). 


PDF Version: BPPB VI (2002) 

impossible to plausibly determine the stemmatic relation between TGS; and TGS; 
we are forced to treat them as two recensions, each with a transmissional history 
of its own. While 7GS;, is the stylistically more concise and consistent recension 
of the TGS, TGS; has interpolated several textual additions typical of Mahayana 
sutra literature and, as might be suspected from the inconsistencies found in all 
three of its representatives, looks back on a manuscript tradition in India which, in 
several passages, harbored problematic readings. 

Judging from the date of translation as the terminus ante quem of the two 
recensions, there is a gap of more than 350 years between them. In the following I 
will show that TGS2, or a version very close to it, already existed before the 
middle of the fifth century, so that it can be placed much closer to the period in 
which TGS; was translated. We therefore need to focus on the reproduction of the 
nine similes of the TGS in the Ratnagotravibhaga. The following points deserve 
our attention: 

* In the second simile of the Ratnagotravibhaga it is twice said that the honey- 
hunter desires honey: purusas tadarthi (I. 102b), madhvarthi (1.104b). This 
characterization also appears in TGS? in 2A and 2.1, but is missing in Ch. 

* The fourth simile in recension 7GS2 deals with the gold nugget of a traveler 
which falls into putrid excrement. (The man is mentioned in 4A and 4.1.) The 
Ratnagotravibhaga also telis of this traveler: suvarnam vrajato narasya 
(1.108a). Ch;, however, just tells of gold fallen into an impure place without 
any mention of the traveler. 

In the sixth simile the Ratnagotravibhaga mentions the fruit of “mango trees, 
palmyra palms and other trees”: Gmrataladiphale (1.115a), talaphalamra- 
(I1.117b). A sequence of four kinds of fruit is found in 6A of TGS): that of a 
mango tree, rose apple tree, palmyra palm and cane (in the verses the mangos 
are missing). Ch;, on the other hand, mentions both in the prose and the verses 
only the fruit of mango trees. 
In 7B of TGS) it is explicitly stated that the buddha-nature is also found in 
animals. The same statement appears in [.119c (tiryaksv api) and [.120 
(tiryaksv api) of the Ratnagotravibhaga. Further, a quotation-like passage, 
which could well derive from 7GS, appears in Ratnagotravibhagavyakhya 
15.11f. (see the note in 7B of my translation). The passages regarding animals 
do not appear in Ch. 
In all three sections of the eighth simile of TGS, a poorhouse is mentioned as 
the residence of a depressed woman. This poorhouse also turns up in all three 
corresponding verses of the Ratnagotravibhaga: anathavasathe (I.121b), 
anathasale ([.122a), anathavesmani (1.123b). Ch; does not mention the 
woman’s residence. 
* The citation of TGS 8B.3-4 concerning tathdgatadhatu found in Ratnagotra- 
vibhagavyakhya 72.11—-12 agrees closely with the passages in Tib, Bth and Chp. 
(See the note in my translation for a more detailed comparison.) 

These examples show that the compilers of the Ratnagotravibhaga(vyakhya) must 
have known recension 7GS2 or a version similar to it. How otherwise could the 
Ratnagotravibhaga share passages with TGS) in its corresponding verse or prose 
sections? The date of. the translation of the Ratnagotravibhaga(vyakhya) into 


PDF Version: BPPB VI (2002) 

Chinese is 511 CE.” As Takasaki (1966) and Schmithausen (1971) have shown, 
the Ratnagotravibhadga(vyakhya) consists of at least two clearly different layers, 
namely the original karikas and a commentary written in verses and prose. The 
recomposition of the nine similes of the 7GS' in the Ratnagotravibhaga is part of 
the oldest layer of the Ratnagotravibhaga(vyakhya). The reproduced verses are 
written in complex meters, and it is hardly imaginable that alterations to the 
original wording of the Ratnagotravibhaga could have been introduced at a later 
stage. Now, we have to reckon with a complex history, from the time of the first 
appearance of recension TGS, the reformulation of its verses and their adaptation 
to the karikas in the Ratnagotravibhaga, the completion of one, if not two, 
commentarial layers to the Ratnagotravibhaga (part of which clearly follows a 
different interpretation) until, finally, the translation of the whole 
Ratnagotravibhaga(vyakhya) into Chinese in 511 CE. In view especially of the 
different exegetical approach to the Aarikds taken by the commentary, but also the 
various steps involved in the whole process, I suppose that we can assume a 
period of at least fifty to one hundred years between the first appearance of TGS 
and the final Chinese translation of the Ratnagotravibhaga(vyakhya). This, 
however, would place the terminus ante quem of TGS> more or less within the 
same period as JGS;. Based on the date of the translation of the 
Ratnagotravibhaga(vyakhya), therefore, we no longer have any reason to consider 
Ch, to be an older recension of the TGS.” 

1.2 The Representatives of TGS 

Bth, Tib and Ch; doubtlessly derive from the same manuscript tradition. They 
differ from each other only in some minor points, which, for the most part, can 
easily be explained as resulting from misreadings or misunderstandings of words 
or passages, redactional intervention or passages already unclear in the Indian 
manuscripts.“° As one might expect, each of the three translations has abundant 
variants not found in the two other texts. These “single variants” may have their 
origin in the respective Indian manuscript, faithfully translated into the new 
language, or in the translation process itself. 

As I have repeatedly stated in the notes to my translation, it seems that Tib 
was subject to some process of redaction at the time of its translation or after. I 
can show in many instances the reasons for such alterations.”” The most obvious 
documentation of this process is a comparison of the pada sequence in the verses 

Here I do not consider the relation of the RGV to other texts (such as the MSA) and the 
associated implications for the dating of the RGV. I shall deal with this matter in section 4.4. 

3 See section 4.4 where, based on other arguments, I arrive at a terminus ante quem of ca 350 
CE for the composition of TGS,. Whether the obvious interpolations found in TGS; were at that 
time already part of TGS, cannot be decided, since they do not form part of the central message of 
the sfitra and can thus hardly be expected to be cited in the RGV(V). 

6 Section 6B is a good example: each of the three versions offers its own way of reading the 
passage without being completely convincing. 

27 The fact that the translators or redactors of Tib deviated in some instances from the Indian 
text could, for example, be the reason for the different rendering of 8B.3-4, which has come down 
to us in Sanskrit from a quotation in RGVV 72.11-12 (see the note in my translation). There, Skt. 
utpanna is rendered as zhugs, and dhatu is rendered as rigs (usually: gotra or kula). The translation 
zhugs for utpanna is not a common one. Similarly, the preference for rigs over khams may reflect 
the choice of the translators/redactors. 


PDF Version: BPPB VI (2002) 

of the different versions (see Appendix B): in 26 of the 69 verses, Tib has a 
sequence different from the one common to Bth and Ch?. Two main reasons for 
the alterations in 7ib are evident. In 14 instances the pada containing the verb, 
which governs the whole verse, has been placed at the end of the verse. In 10 
cases the editors of Tib felt the need to position padas embracing relative clauses 
or other specifying elements before the element to be specified. In several 
instances, the different arrangement of the syntactical units led to a new 
understanding of the content. 

In the prose section, such alterations are not always so easy to determine. 
If there was reason to assume that redactional intervention has led to a reading 
only attested by Tib, I have always indicated this in the notes to my translation. 
However, Tib is a smooth, well-polished text with a more or less easy flow of 
sentences. Its terminology accords in almost all instances with the Mahdvyutpatti. 
Precisely because of these characteristics, we should be careful not to follow its 
argumentation uncritically in the effort to uncover the original content of TGS). 

Regarding Bth, we have reason to characterize it as a translation dating 
from a time when the standard Buddhist vocabulary was not yet established in 
Tibet, namely up to the eighth century, after which translation activities became 
more organized and standardized with the aid of compilations of compendiums 
like the Mahavyutpatti or the sGra sbyor bam po gnyis pa. As I have shown 
elsewhere,”* a comparison of the terminology chosen by the translators of Bth 
with the corresponding terms in Jib reveals that in most of the cases the 
terminology of Bth is not in line with the Mahavyutpatti. In Bth, a tendency to a 
more literal and less idiomatic style of translation seems to prevail. Whereas many 
of the verses in Jib have been rearranged according to the basic needs of the 
Tibetan syntax (see above), in Bth the pada sequence seems to reflect, if not 
wholly coincide with, the Sanskrit original, where the main verb of the sentence 
may not have been positioned at the end. Other refinements typical of the time 
after the appearance of compendiums like the Mahdvyutpatti or the sGra sbyor 
bam po gnyis pa are lacking in Bth: verbs in many cases do not indicate the 
hierarchic level of the subjects involved;” the particle dag expressing collectivity 
is less commonly employed; *? numbers appear which are not in their 
“Tibetanized” forms.’! Furthermore, the fact that some of the names of the arhats, 
translated in 7ib, are transliterated in Bth (cf. OC) also marks Bth as the earlier 
work, written when many of the (later) standard translations of arhat names had 
not yet been established.” 

For the prose, it is more difficult to judge how far Bth follows the syntax 
of the Indian original. Verbs are always placed at the end of the sentence. In the 
case of names, Bth clearly follows the Sanskrit, where the name usually proceeds 
titles, as in Bth: rDorje’i blo gros byang chub sems dpa’ sems dpa’ chen po 
against Tib: byang chub sems dpa’ sems dpa’ chen po rDo rje’i blo gros for Skt. 

_ 78 See Zimmermann 1998: 43-46. 

” Ibid., 47, n. 25. 

* Thid., 47, n. 26. 

3! Bth throughout uses brgya stong instead of ’bum for Satasahasra. 

32 The situation is similar to the employment of the auxiliary verb ‘os pa to express what are 
gerundives in Sanskrit: in all 18 cases of its appearance in Tib it is missing in Bth. See Simonsson 
1957: 156f. for an example where phyag ‘tshal in the old version of the SP is rendered as phyag 
byar ’os in the revised text (for Skt. vandaniya). The auxiliary verb ‘os pa seems at some point to 
have become the standard for rendering gerundives from Sanskrit. 


PDF Version: BPPB VI (2002) 

Vajramatir bodhisattvo mahasattvah. This principle is also found in the old 
translation of the Saddharmapundarikasutra (see Simonsson 1957: 173f.) and, in 
the case of the Aksayamatinirdesasutra, in the version closest to the Dunhuang 
fragments as well as in the Dunhuang fragments themselves (cf. Braarvig 1993a: 
ix). In most of the direct speeches, Bth places the verb of speech before the 
content of the speech, a feature common to the Sanskrit but not attested for 7ib, 
where this verb is always found at the end of the speech. 

Generally, the style of Bth is more prolix and in parts unintelligible, 
features whose reason may have been that the translators felt uneasy about 
departing from the original Indian syntax. In some passages this could have led 
them—faithfully following the (faulty?) Indian text—to place syntactical units 
one after the other without the least attempt to achieve the necessary consistency 
and continuity of discourse by introducing connective elements into the Tibetan. 

There can be no doubt that Amoghavajra also tried to keep as close as 
possible to the Indian text. The Sanskrit on which his translation is based must 
have been largely identical with the Sanskrit from which the Tibetan translations 
derived. It seems that he followed the Indian text more or less word for word, 
avoiding any kind of interpretative translations. Even in the syntax, he remains 
loyal to the Sanskrit: the pada order is for the most part untouched. The sequence 
of syntactical units in some passages, at odds with Chinese grammar, is most 
likely also modeled on the Sanskrit. Amoghavajra’s translation style thus again 
seems to be based on the idea that the single elements of the Indian text should 
remain discernible and be arranged in the order found in it. 

However, as already mentioned above, Amoghavajra takes the term 
tathagatagarbha (4(2i) as a designation for the buddha-nature itself, that is, as 
a separate entity within living beings. Further on throughout the text, we find a 
number of additional or substitutional elements of an explicative nature which 
could well be commentarial units of the translator that crept into the translation: 

* In 0G the flowers in the sky are described as “being united with each other as a 
whole” (44494 45%). The tathagatas are said to be endowed with the 
dvatrimsan-mahdapurusa-laksanas. Neither element is found in the other 

* 0M: Besides gods and humans in the list of living beings that pay reverence to 
the tathagatas in lotuses, mention is also made of dragons (for nagas), yaksas, 
gandharvas, asuras, garudas, kimnaras, mahoragas, amanusyas “and others.” 

* In the last pada of 1.5, instead of the purification of living beings as the motive 
for teaching the Dharma, it is stated that sentient beings may thereby attain the 
three bodies endowed with buddha-knowledge (4/8 = 4 B45). 

* In 3B, between the statements that buddhahood is found in all living beings and 
that the tathagata removes their defilements from them, the following insertion 
is found: 79382, Alpe, BAAR AZA. One possible 
translation is: “If [one] can = has the potential to?) awaken, [one] will become 
a perfectly awakened [one], settled firmly [and] peacefully in spontaneous 


PDF Version: BPPB VI (2002) 

1.3 Similarities between the Chinese Translations 

One surprising feature is the fact that in several cases Ch; lines up with Ch, 
against both Tibetan translations. The nature of the differences shows that it is not 
the wider Chinese context in which Ch; and Ch2 are embedded (or the Tibetan 
context for the Tibetan translations) that could be held responsible for this. Some 
of these common differences are: 

¢ Only the two Chinese recensions characterize the withering of the flowers in 
the introduction and the first simile as taking place “in an instant” (OH, OI, 0J, 
0M). Also, in 7A the death of the man in the wilderness is described as 

* Both Ch, and Ch; do not mention the “petals and stalks” of the flowers in the 
description in OJ. Later, in verse 0.3, the stalks are again not mentioned in 
either Chinese translation, while the petals are missing only in Ch). 

* In 0J, Ch; and Ch2 do not mention the putrid smell of the flowers. 

* In the enumeration of buddhadharmas in 5A, Ch; and Ch) lack the 

* 6A: The problematic expression jig rten (na) gnas pa (Tib) and ‘jig rten zhugs 
Shing gnas pa (Bth) in the Tibetan has no counterpart in the Chinese. 

* In verse 10.1 only the Tibetan translations mention explaining the TGS among 
several other methods of propagating it. 

* In 11C both Chinese recensions additionally mention the attainment of the five 
abhijnas (#4) by those being touched by the light of the bodhisattva 
*Sadapramuktarasmi. The Tibetan is restricted to the attainment of the 
incapability of turning back from supreme and perfect awakening (avaivartika 
anuttarayam samyaksambodhau). 

* Instead of the problematic ‘di ’dra (Tib) or de ‘dra (Bth) in verse 11.1, the 
Chinese has # (“continuously”). 

Up to now, I have no firm idea how to explain these common points in the 
Chinese. If we do not want to assume that Amoghavajra, to whom 
Buddhabhadra’s translation was certainly known and probably accessible, adopted 
some of the common points from the fifth-century translation, we need to reckon 
with the possibility that at least the cases where elements are missing in the 
Chinese came into existence coincidentally. 

1.4 Structure, Contents and Textual History of the TGS 

The 7GS is clearly dividable into seven main sections: 
I. (0) The setting in R&ajagrha, including the manifestation of supernatural 
II. (1-9) The nine similes: the actual TGS 
III. (10) The description of merit resulting from the propagation of the TGS 
IV. (11) The story of *Sadapramuktarasmi and *Anantarasmi 
V. (12A-B) The question of Ananda 
VI. (12C) Homage to the one who holds the TGS in one’s hands 


PDF Version: BPPB VI (2002) 

VII. (12D) Description of the delight and praise on the part of the onlookers 
(return to the scene of 0) 

Without a doubt, the nine similes form the essential part of the TGS. Not only is 
the title of the stitra derived from the description in the first simile, but the text 
itself also regards the nine similes as the actual TGS: in sections 0 and 10-12 the 
TGS is frequently referred to. This stitra within a sitra, the nine similes, is 
narrated by the Buddha himself. It is this section alone which is usually dealt with 
by later commentators. The nine similes contain the new message and the 
doctrinally relevant passages for which the TGS is known. They are enclosed by a 
frame (0 and 12D) which is common to most, if not all, stitras of the Mahayana. 
Between the similes and the end scene, three more units are found (10, 11, 12A— 
C). Whereas the description of the merit resulting from the propagation of the 
TGS is a common element in many other siitras, the story of *Sadapramuktarasmi 
and the (probably later interpolated) question of Ananda seem peculiar to the 
TGS. Later we will have to face the question what their relation to the similes 
might be. 

As mentioned above, the main speaker after the introductory description is 
the historical Buddha himself, who is interrupted only by the two questions of 
Vajramati and Ananda, his narration ending with a return to the setting. The text is 
in prose, with verse portions that repeat, usually at the end of a short unit, the 
main issues contained in prose before.*® This structure of verses ending a prose 
section is well attested in many other Mahayana siitras. 

At first glance, the introductory section 0 and the following exposition of 
all nine similes seem to be from one mold. The Buddha narrates in the third 
person, with his thoughts interspersed in the first person. The descriptions in the 
similes are all vivid and concrete. The similes appear to be constructed in a 
unified fashion. They all start with the description of a more or less well-known 
occurrence or process met with in daily life (with the exception of the lotus 
flowers in the sky), and then draw a comparison between it and the spiritual 
sphere. At the end, the prime role of the Tathagata is illustrated as that of teaching 
living beings about their unknown potential. From a doctrinal point of view, too, 
there seems to be no reason to assume that the similes are not the creation of a 
single author or group of authors put forward as a monolithic block. Nevertheless, 
an analytic examination of the first simile leads to several observations which 
clearly go against the assumption that the introduction and the nine similes had 
already been arranged in their present form at the very beginning. 

The most obvious manifestation of textual heterogeneity is the first simile 
itself. After his announcement that he would expound the 7GS, in 0M the Buddha 
sets about formulating the first simile, which, regarding the upamana, that is, the 
part of the simile used to illustrate the subject matter, is taken from the setting of 
the introductory section. Then, in 1A, he abruptly starts again to recount this first 
simile, this time with emphasis on different elements. A close look at both 
“versions” of this first simile, that is to say, version 0M and version 1A, reveals 
the following differences: 

(1) In 0M, the upamana is very close to the description in the introductory 
setting: The lotuses are mentioned as supernatural creations of the Tathagata; 

33 The fact that the section with the question of Ananda does not end with verses repeating his 
question is one sign that it is most probably not a part of the oldest Indian recension of the TGS. 


PDF Version: BPPB VI (2002) 

the tathagatas sitting in their center emit rays of light; the spectators recognize 
the tathdgatas and pay homage (without mentioning the divine vision); the 
flowers are, as stated in the introduction, open. 

This is all in contrast to 1A: there nothing is said about light effects, a 
central element of the introductory part; the flowers are not explicitly 
characterized as supernatural manifestations of the Buddha; no spectators and 
their act of homage are mentioned; instead, there appears a person endowed 
with divine vision who “desires to look at the forms of the tathagatas.” This 
last expression is reminiscent of simile 2, where the honey-hunter is described 
as “desirous of honey” (2A; 2.1), and of simile 3, where men and women want 
to remove the kernels from their husks because they desire to obtain edible 
food (3A). Further, the person in 1A is in need of a divine vision to see the 
tathagatas in the middle of the flowers because they are “not blooming and not 

(2) In contrast to the introduction and to 0M, in 1A the person with divine vision 
then removes the disgusting petals of the lotuses and cleans the tathagatas 
sitting inside. The process of purification in the upamana is a completely new 
element not found in the sections preceding 1A but common in the other 

(3) Regarding the upameya, that is, the subject matter to be illustrated in the 
simile, in both 0M and 1A the term tathagatagarbha appears. However, the 
sentence containing the term in 0M is found at the end of the chapter and is, 
furthermore, missing in Ch;. I consider it to be a later addition in TGS). This 
seems to be supported by the uncommon position of the sentence: the act of 
perceiving the buddhas within living beings had already been dealt with before. 
Meanwhile the Buddha had been stressing the equality between the tathagatas 
within and himself. Why should he again mention the act of perceiving? 
Further, the whole of the sentence, as it is found in the Tibetan, seems very 
uncommon and strange in the context of the TGS, the tathagatacaksus 
(“tathdgata-vision”) being characterized as *prasddika (mdzes pa; in my 
translation rendered as “admirable”), a term usually applied to persons. A 
plausible reason for interpolating the passage may have been the wish to use 
the central term tathdgatagarbha, which until this passage had not come up in 

(4) The simile in 1A ends with the statement that living beings realize 
buddhahood. How, then, are we to deal with chapter 1B? 1B starts with the 
paradigmatic statement that the Essential Law (dharmata) of all dharmas 
consists in the fact that all living beings are tathagatagarbha. It is this 
statement which has been cited word for word by the Ratnagotravibhaga- 
vyakhyd, Bu ston, in his mDzes rgyan, quotes from the beginning of 1A till the 
end of this passage. The text in the TGS, however, goes on to state that the 
Tathagata teaches the bodhisattvas, that the bodhisattvas thereupon follow his 
teachings, and that they finally realize buddhahood and perform the tasks of a 
buddha for other living beings. All this is repetitive since it is already stated in 
1A that, having accepted the Buddha’s teaching, the “tathagatas of living 
beings are established in perfection,” which can only be understood as the 

4 See BHSD s.v. prasadika. 


PDF Version: BPPB VI (2002) 

attainment of buddhahood. 

The structure of the passage extending from OM till 1B, however, is seen 
in a completely different and plausible way once we read through it without the 
last (interpolated) sentence in 0M, without the whole of 1A, and without the 
categorical statement at the beginning of 1B. The “repetitive part” in 1B thus 
becomes the natural continuation of the narration from OM: After the Tathagata 
has confirmed the equality of the buddhas within living beings with himself in 
OM, in 1B he then decides to teach the Dharma to the bodhisattvas. The 
bodhisattvas “apply themselves” to his teaching and finally become buddhas. 
The simile begun in 0M is thus expanded on logically and is rounded off at the 
end of 1B. The paradigmatic statement at the beginning of 1B, on the other 
hand, is not part of the simile. As a part of simile 1A, it may have served the 
function of a summarizing formula which repeats the foregoing in traditional 

(5) Finally, I shall deal with the five verses following the prose passage. Section 
1.1 shows a close affinity with 1A. The verse describes the flowers as “not 
opened” and mentions the person with divine vision. In the second verse the 
purifying activity takes place on the upamana level (missing in 0M). The 
following three verses, too, follow clearly the narrative in 1A. They close with 
the establishment of the buddha-bodies (in the perfection of the tathagatas), a 
parallel to the last sentence in 1A. There can be no doubt that the verses 
represent the simile as it is found in 1A. The categorical statement at the 
beginning of 1B is not reflected in them. 

The above analysis leads to the following results: In sections 0M to 1C two lotus 
similes have been combined. One is found in 1A (and possibly also the first 
sentence of 1B) and the verses of 1C; the other one comprises chapters 0M 
(without the last sentence containing the term tathagatagarbha, which is probably 
a later insertion) and iB (without the first sentence). The term fathdgatagarbha is 
thus only found in simile 1A. Simile 0M draws upon the introductory setting 
described from 0G to OK. Simile 1A, on the other hand, clearly follows the other 
eight illustrations of the TGS® and, with the introduction of a person with divine 
vision, adapts the upamdna to the upameya, where the Buddha recognizes the 
buddha-nature of living beings with his buddha-vision. In terms of the 
homogeneity of the two parts of the simile, 1A thus proves superior, since it aligns 
the lotus imagery smoothly with the idea of the hidden buddha-nature of living 
beings which needs to be set free. 

Concerning the conformity of simile 0M to the introduction, I have stated 
that they mesh: no supernatural vision is necessary since the flowers are open. 
There are, however, some phraseological inconsistencies between the two 
sections. In all the descriptions of the buddhas seated in the calyxes up to OJ, Tib 
employs the term sku (Skt. kaya, atmabhdava or the like) when referring to the 
buddhas in the lotuses. Then, from 0M onwards, the terminology in Tib shifts 
suddenly to gzugs (Skt. riipa?). Though Bth consistently employs sku, this change 
is confirmed by Ch, in which the character }{% (for rijpa?) appears here for the 
first time in the text to designate the tathagatas in the flowers. Parallel to this, 

35 Whilst simile 1A contains the introductory formula rigs kyi bu dag ‘di Ita ste dper na / (SKt. 
*tadyathapi nama kulaputrah), which is the common formulation applied in all other eight similes, 
in OM the introduction runs rigs kyi bu dag ji ltar. (In Bth, however, the introductory formula in 
OM is identical with the other similes: rigs kyi bu ‘di lta ste.) 


PDF Version: BPPB VI (2002) 

again from 0M onwards, there is a different term to denote a very high number. 
Whereas up to 0J we find kotiniyutaSatasahasra, from 0M onwards we find only 
kotiSatasahasra. Later, in the part following the similes, the former, longer 
expression is used again. Worth mentioning, too, is a change in the audience. In 
the chapters up to 0K the ones watching the supernatural manifestations are said 
to be “the whole multitude of bodhisattvas and the four assemblies” (0H; OI), “the 
world with [its] gods, humans and asuras, all the bodhisattvas, and the four 
assemblies” (0J), or “myriads of living beings” (0J). In OL, however, the 
Tathagata answers merely to the “whole multitude of bodhisattvas,” while in the 
simile 0M itself only “gods and humans” are mentioned. Though such differences 
between the two sections are definitely minor in view of their homogeneity, they 
nevertheless indicate some unevenness. I will come back to this point later. 

Concerning the textual history of the TGS, we are now confronted with a 
complex situation. The simile of the lotus occupies, no doubt, a special position 
among the nine similes. The fantastic description of lotus flowers rising to the sky 
marks a contrast with the realistic spirit of the other eight illustrations. I can 
hardly imagine a reason for its inclusion among the set of eight others than the 
need to build a bridge to the lotus scenery of the beginning which, for whatever 
reason, had been chosen to introduce the similes. 

We do not know how the Mahay4na siitras came into existence. Any 
discussion would soon become very speculative and go far beyond the limits of 
this work. However, it is reasonable to suppose two separate parts, namely the 
eight similes and the lotus scenery of the introduction, which originally existed 
independently of each other. To assume anything else would make it difficult to 
account for the differences in the basic conception between the introduction and 
the eight similes (supernatural vision, unopened flowers etc.). If the introductory 
scenery had been composed especially for the eight similes, it could have been 
adapted perfectly to the needs of the similes. But, as I have shown above, this is 
not the case. I therefore conjecture that when looking for a relatively fitting setting 
for the eight similes, the compiler decided to arrange the description of the fading 
flowers (taken from a different context) into an introduction to the similes. But 
then he was forced to compose parts 0M and 1B (without the sentences including 
tathagatagarbha) in order to explain plausibly the relation between the scenery 
and the eight similes, starting with the honey in the honeycomb. He tried to 
remain as faithful as possible to the lotus scenery but nevertheless unwillingly 
introduced the changes in phraseology discussed above (*kaya for *riipa etc.). 
Most likely, this newly created first simile did not include a verse portion. The 
reason for this could lie in the fact that at this stage the other eight similes, too, 
existed only in prose, or that the first illustration was felt to be but an introductory 
bridge between the lotus scenery and the similes of the honey and the rest, rather 
than an independent simile with a corresponding verse section. 

In a later step, simile 1A and the attendant verses would then have been 
interpolated. The compiler took great care to preserve the old lotus simile in its 
entirety and to insert the new simile in a position which would make it difficult to 
recognize any traces of such an insertion. The respect he paid to the old lotus 
simile by preserving all of OM/1B makes it more difficult to argue that the same 


PDF Version: BPPB VI (2002) 

compiler replaced a hypothetical old verse section with his new one. It is more 
likely, as suggested above, that the old simile had been without verses, or that it 
was at this stage that all similes became newly fitted out with a verse portion. This 
question cannot be answered wholly satisfactorily. 

The insertion of 1A could have been caused by the advent of the term 
tathagatagarbha, which till that insertion had not been used anywhere in the 
introduction and the similes. At the same time, the wish to transform the lotus 
simile into a “classically” structured one conforming to the style of the following 
similes (with homogeneous correspondences between upameya and upamana, and 
including verses), and doing justice to the tathagatagarbha concept, could have 
been one main factor at work. The compiler could thereby also compensate for the 
incompatibility between the introductory setting and the basic idea behind the 
eight similes: the missing person with divine vision representing the Tathagata, 
the fact that the tathagatas in the flowers are recognizable by everyone,”° and the 
lack of the purifying activity. 

The main reason, however, was probably the compiler’s wish to introduce 
the term tathagatagarbha. We cannot be sure what exactly led to the eminence of 
the term. As long as we have no other early text which could have coined the 
term, we should assume that it developed in fact from the lotus image in the first 
stage of the TGS. We can even imagine that the author(s) of the incomplete TGS 
(without 1A) searched for a designation for their new idea of the buddha-nature, 
and possibly also for a title of the sittra meant to expound their theory. Once they 
had discovered that the term tathagatagarbha (with the rich nuances described 
below) could serve their needs well, they strove to give the term a place in the 
Siitra itself. This may have been their motivation for inserting chapter 1A (and 
probably also the first sentence of 1B: ... dharmanam dharmata ...). 1A serves to 
provide a definition of the term tathagatagarbha. This can still be felt in the first 
sentence when it is said “... recognize that there are tathagatas sitting cross-legged 
in their center (*madhya), in the calyx of [each] lotus (padmagarbha),....” My 
impression is that the author(s) here deliberately used the term tathagata (and not 
*tathagatariipa as later on) and that specifically in this passage they glossed 
garbha as *madhya in order to draw readers’ attention to the meaning of the term 
garbha and its role in the compound tathagatagarbha. 

If my hypothesis is right, the textual history of the TGS can now be 
divided into at least three major steps: first, the composition of the eight similes; 
secondly, their combination with the introductory lotus scenery and the 
composition of the “old” lotus simile (0M; 1B); thirdly, the interpolation of the 
“new” lotus simile 1A with its verse section and the inclusion of the term 
tathagatagarbha (and probably also the selection of the title for the sutra). 

The upameya of the “old” lotus simile (0M; 1B) deserves special attention. As it 
came into existence after the compilation of the other eight similes, and thus 
represents the essence of the different metaphorical approaches to the idea of a 
buddha-nature in living beings, it reflects what the author at this stage of the text’s 
development thought to be the underlying pattern of the upameya. Furthermore, it 
has the advantage of not being constrained by a dominant upamdna: the fantastic 

6 It is definitely possible to use the term padmagarbha in the case of a fully blossoming lotus. 
Even when flowering, the petals will not open completely, so that a kind of interior space (garbha) 


PDF Version: BPPB VI (2002) 

introductory lotus setting is formulated loosely enough to allow different kinds of 
construction for the upameya. It does not force the author to adapt the upameya to 
the requirements and the elements of a realistic upamana. It is therefore probably 
no coincidence that we find here the most explicit formulation of the liberating 
path in the whole siitra. The passage asserts that the Buddha teaches the 
bodhisattvas, that he incites their faith, and that the bodhisattvas, by applying 
themselves to his teachings, attain buddhahood. Though the activity of the 
bodhisattvas (or living beings) in their own process of liberation is sporadically 
delineated in some of the other similes (but on the whole seems to remain in the 
background), it is clear that at least for the person(s) who creatively combined the 
similes with the introduction, be they the authors of the similes or not, this active 
participation of the bodhisattvas was seen as a very important element. 

If we now turn to the parts of the siitra following the similes, it is difficult 
to avoid the impression that we again are dealing with more or less originally 
independent elements added to the similes for various reasons. Let us first have a 
short look at the content of these sections. The description of merit arising from 
the propagation of the stitra (10) is a common theme in Mahayana literature. 
Though other methods of attaining merit, with which the siitra-related activities 
are contrasted, can differ according to the text,’’ the propagation of the siitra by 
preserving, reciting, copying and teaching it is very common. First, the TGS 
announces in a general way the positive result for one’s merit from applying 
oneself to the propagation of the siitra (10A). It goes on to illustrate this fact 
impressively by contrasting it with the meaningless production and donation of 
myriads of pavilions to myriads of buddhas and bodhisattvas or the bestowal of 
hundreds of thousands of flowers upon them for a hundred thousand kalpas (10B-— 
C). It is even said that the internalization of a single simile of the TGS or the 
simple joyful approval (anumodana) of what has been heard would by far 
outweigh the merit attained through such donations. This second idea of joyful 
approval leading to the accumulation of merit does not seem to be part of the 
common descriptions in other Mahayana siitras but does appear in the 
Saddharmapundarikasutra. Rather than enter into a discussion of the complex 
issue of the veneration and propagation of a text, and its development and social 
meaning, I shall simply propose that the underlying tendency here is to redirect 
emphasis to the internalization of religious values. If we keep in mind that the 
propagation of the siitra also encompasses its “understanding” and “explanation to 
others,” it is only natural that this process presupposes personal participation and 
reflection, elements which easily tend to be discarded during mainly external acts 
of veneration. The appearance of the term anumodayati, in the sense of cultivating 
an active and joyfully approving attitude towards the teaching, supports such a 

This section on merit might otherwise have been the end of the siitra, 
which is the position it usually occupies in other texts, were it not for the fact that 
it is followed in the TGS by two other parts, with which I will now deal. The story 
of *Sadapramuktarasmi and *Anantarasmi (11) was probably added to raise the 
authoritative status of the stitra by embedding it in a quasi-historical context, 

37 As mentioned in the notes to my translation, a passage in the KP resembles the TGS most in 
this respect. SP 332.9ff. contrasts the hearing of the SP with the practice of the five padramitas of 
dana, Sila, ksanti, virya and dhyana. For the RGV see V.3-6, where the comparison is with only 
the three virtues of dana, Sila and dhyana. 


PDF Version: BPPB VI (2002) 

which is at the same time the history of its transmission. The story itself may have 
been an independent entity before it became incorporated into the sitra. It 
contains two connecting links with the core siitra: one is the fact that the tathagata 
*Sadapramuktarasmi taught the TGS for five hundred great cosmic cycles, in 
consequence of which the listeners finally attained awakening; the second is the 
identification of *Anantarasmi, who had asked *Sadapramuktarasmi to teach, 
with the main representative of the audience, Vajramati. *Anantarasmi together 
with three other well-known bodhisattvas did not attain awakening. The sections 
before describe *Sadapramuktarasmi, from the time in the womb of his mother till 
his form as earthly remains, as emitting light that has a highly beneficial effect on 
all living beings. I can hardly see any connecting link between this story and the 
core sitra, aside from the (surely not central) fact that *Sadapramuktarasmi also 
emits light while in the womb (*garbha) of his mother. 

Seemingly even more unrelated is the question of Ananda (12A-B), which 
is missing in TGS, and is thus probably a later interpolation. Ananda asks the 
Tathagata how many buddhas one has to hear the Dharma from in order to attain 
perfection. In his answer the Tathagata does not restrict their number, but stresses 
the importance of generating the aspiration to awaken immediately. As stated 
above, one reason for the interpolation of this story with Ananda as the main 
figure could be his role as the one who is usually entrusted with the preservation 
of the teaching. However, this section does not contain any topical link with the 
other parts of the sutra. 

Another minor observation further highlights the heterogeneity between 
parts 10 to 12 and the similes: whereas throughout the similes the Tathagata 
addresses a large audience led by Vajramati (kulaputrah), in the following 
sections he only speaks to Vajramati. This strengthens the conclusion that we 
have a core sitra, consisting of the nine similes, embedded in an introduction and 
other essentially unrelated elements at the end. Particularly in regard to the story 
of *Sadapramuktarasmi, the reasons for its inclusion in the siitra remain unclear, if 
it was not simply the need to extend the sutra’s length. 

1.5 The Structure, Nature and Contents of the Nine Similes 

As mentioned above, all similes with the exception of the first one bear witness to 
a realistic, down-to-earth spirit. They are of a mundane nature; the scenes 
described can be easily imagined by readers and, for the most part, are taken from 
daily life. As I have argued above, the inclusion of the first image, though 
exceptionally fantastic in nature, was due to the fact that the compilers wanted to 
combine the eight similes with the introductory element centering on the lotuses, 
and therefore felt compelled to expand upon the setting of the introduction in the 
form of a simile. All the similes show the same specific structure:”* they first start 

38 See the entry in Meyer s.v. Gleichnis: “Sprachliches Gestaltungsmittel, bei dem eine 
Vorstellung, ein Vorgang oder Zustand (Sachsphare) zur Veranschaulichung und Intensivierung 
mit einem entsprechenden Sachverhalt aus einem anderen, meist sinnlich-konkreten Bereich 
(Bildsphare) verglichen wird. Bild- und Sachsphare sind im allgemeinen durch Vergleichspartikel 
(,,80 ... wie) ausdriickl. aufeinander bezogen, sie decken sich aber nicht wie in der Allegorie in 
mehreren Einzelziigen, vielmehr konzentrieren sich die einander entsprechenden Ziige beider 
Spharen in einem einzigen, fiir die Aussage wesentlichen Vergleichsmoment, dem Tertium 
comparationis, in dem die beiden Seiten sich beriihren. Das Gleichnis ist vom bloBen Vergleich 


PDF Version: BPPB VI (2002) 

with the description of a situation or process in the material sphere (upamdna; 
“Bildsphare”) to which the assumed factual situation (upameya; “Sachsphire”), 
namely the existence of buddhahood in all living beings, is then compared. In 
contrast to what we normally call “allegory,” a simile does not allow parallels to 
be drawn between a whole range of congruent factors featuring in the upamdna on 
the one hand and in the upameya on the other. Rather, the two spheres merely 
share a common fertium comparationis, the central point of comparison, which it 
is the aim of the simile to illustrate. Some of the similes do seem to call for an 
allegorical interpretation. In 4B the elements of the material sphere are explicitly 
compared to their counterparts in the spiritual realm. Nevertheless, what A. 
Jiilicher has convincingly demonstrated for the interpretation of similes in the 
New Testament, namely that they were undermined by the “Gefahr 
unsachgemafer allegorischer Auslegung, als man jeden Zug in ihnen tiefsinnig zu 
deuten versuchte,”*’ should serve as a warning for the interpretation of our 
similes. Let us keep in mind that it was not the intention of the author(s) to 
construct allegories with perfect proportional relations in regard to all single 

What, now, is the tertium comparationis of our similes? Generally 
speaking, it is the notion that a precious element, pure in nature, but hidden and 
unsuspected, truly exists, and while it is covered by impurities, its nature remains 
unaffected. The similes, though they each emphasize different aspects of this 
tertium comparationis, are meant to illustrate the same basic situation. As a 
further element common to most of the similes, and thus a second tertium 
comparationis of sorts, we find the description of the release of this precious 
element with the attending beneficial consequences for living beings. In the 
following, 1 will summarize the content of the similes and try to point out what 
each of them emphasizes.” 

As suggested above, the first simile was probably not part of the earliest 
set of eight. However, with the decision to adapt the introductory lotus scenery to 

durch die breitere Ausgestaltung und eine gewisse Selbstindigkeit des Bildbereichs unterschieden, 
wird des 6fteren auch gleichbedeutend mit Parabel verwendet; vielfach werden jedoch beide 
Begriffe in dem Sinne unterschieden, daB bei der Parabel die Sachseite nicht ausdriickl. genannt 
ist, sondern erschlossen werden mu8 (demnach setzt die Parabel das Bild statt der Sache, das 
Gleichnis setzt es neben sie). 

39 See Meyer s.v. allegorische Schrifideutung; further Adolf Jiilicher, Die Gleichnisreden Jesu, 
Freiburg, 1899. 

“° Though the similes shed light on the tertium comparationis from different angles, they are 
similar enough to convince one that (with the exception of the first simile, as argued above) they 
were composed and arranged by a single person or a group of persons who worked in very close 
cooperation. There are, nevertheless, some minor observations which suggest a kind of loose 
grouping of similes 2, 3, 4 and possibly 1A (= the “new” lotus simile) on the one hand and similes 
OM/1B and 5 to 9 on the other. That the newly introduced simile 1A mirrors partly similes 2 and 3 
(e.g., the person is desirous to see the tathagatas: ... to get the honey (2), ... to have food (3)) is 
natural since they were the ones immediately after 1A, and so the obvious prototypes. A common 
element of 2, 3 and 4 (verse 4.4 in Ch,) is the term buddhata/*tva /tathdgatata/*tva to designate the 
buddha-nature of living beings. Similes 0M/1B, 5, 6, 7, (8) and 9, on the other hand, share the term 
tathdgatajriana and, what may be of more significance, they state that the Buddha teaches 
bodhisattvas (and not common living beings) in order to remove their defilements. The awakened 
bodhisattvas are then expected to help again other living beings. However, I do not think that these 
differences are enough to support further text-historical conclusions, which could only be highly 


PDF Version: BPPB VI (2002) 

the eight similes, and the formulation of the lotus simile itself,*! it became the 
most arresting, impressive and comprehensive illustration. The compound 
tathagatagarbha itself, mentioned in close relation to the term padmagarbha, the 
“Jotus calyx,” probably also developed from it. The lotus, by far the commonest 
flower in Buddhist literature, and a symbol of the spotlessness of the bodhisattva 
acting in samsara, its flowering being compared with the manifestation of the 
virtues of buddhahood,” is worthy of the prominent position accorded to it in the 
first simile. The illustration draws a contrast between the ugly, withered flowers 
and the brightly shining tathagatas sitting in meditation in their calyxes. Just as 
these tathagatas are found inside the lotuses, completely unaffected by their 
abhorrent surroundings, so too do sentient beings contain fully developed 
tathagatas within themselves. In the upameya the author still presents the picture 
of a very concrete tathagata situated in living beings, an image borrowed from the 
material sphere of the same simile. The second main point of the illustration is the 
removal of the withered petals by a person with divine vision. Obviously only he 
can know about the tathagatas hidden in the petals. The act of removal is 
compared to the teaching activity by which the Tathagata induces or directly 
accomplishes the elimination of the defilements of living beings, which had 
hitherto veiled their buddha within. 

The simile of honey shielded by bees contains one irritating element, 
namely the collecting of honey, which is a highly destructive activity that robs 
bees of a source of life.’ Nevertheless, it was used by the author to illustrate how 
a person with skill in means would know about the honey contained in the comb 
and manage to expel the insects. This is compared to the removal of the 
defilements by the Tathagata. The honey is then used in an “appropriate” way. 

‘I To be exact, we need to deal in terms of an “old” and a “new” lotus simile (see above). 
However, here I will not differentiate between them but consider both as a unit. This is how the 
reader would understand things, and also follows the traditional exegetic approach. 

“ See RGVILS8. 

“Not much is known about honey-collecting activities in India. In a publication of the year 
1988 Eric Valli and Diane Summers describe how a group of “honey hunters” of the Gurung tribe 
in Central Nepal make their living by collecting and selling honey. They go into jungle areas and 
climb rocky cliffs where the bees, usually the Apis laboriosa, the largest honeybee in the world, 
have their combs. These nests may be as big as a human being, and contain between 50 and 60 kg 
of honey. A group of hunters, consisting of five to ten people, takes about one hundred nests a 
year. Amounts reported for the days of their grandfathers are about ten times as high, but yields are 
continually decreasing due to “the destruction of existing nests and the diminishing forest.” (This 
and all following quotations are taken from the introduction in Vaili and Summers 1988.) 

The work is very risky, since one of the hunters has to climb down a 50-m-long swaying rope- 
ladder from above the cliff and to cut out the comb. Before he can do that, he has to expel the bees 
with flaming bundles of leaves pushed under the comb. The bees react aggressively and “attack 
any moving creature” who does not protect itself with the smoke of leafy branches fed into a fire. 
Once the bees have left, the comb is partially cut out and lowered down to the ground in a basket. 
The authors remark that “unknowingly, the hunters have become terrible predators. Entire colonies 
of bees are destroyed when Mani Lal (the leader of the hunters] cuts down their nests. If the queen 
and enough worker bees escape and return, the colony can be quickly reconstructed and, within a 
month, the nest can be sufficiently pregnant with honey for another expedition to be launched.” 

All parts of the comb are exploited: the larva cells, rich in protein, are consumed as tonics; the 
melted and filtered wax “is sold in Katmandu to the artisans who use it in the lost-wax process of 
casting statues” (see simile 9); the precious honey, “valued as a universal remedy and tonic, is sold 
to villagers or exchanged for grain, yogurt, milk, a chicken, or even a day’s work.” The price in 
1988 for 454 g (1 pound) of honey was ten Nepalese rupees, a sum which must be seen in relation 
to the average two rupees a day on which about 60 percent of the rural population then lived. 


PDF Version: BPPB VI (2002) 

The beneficial result is likened to living beings who become tathagatas and can 
perform buddha-acts for the good of other sentient beings. 

In the third simile of kernels enclosed by husks, the main emphasis lies on 
the obviously aberrant belief, as Ch, puts it, that kernels in their husk have no use. 
The author is assured that every listener knows about the relation between kernels 
and husks in the case of cereals—on which the simile is based. It thus seems that 
in this illustration the author wanted to stress the commonness of the fact that 
buddhahood is found in all living beings (covered by defilements) and the 
normality of its manifestation. He compares the manifestation of buddhahood to 
the ordinary process of turning cereal to account by extracting edible grains. The 
usual tertium comparationis, namely the fact that something valuable is hidden in 
impurities, does not seem to be of primary significance here. 

A gold nugget in excrement is at the center of the fourth simile. The 
contrast between a nugget of pure gold and its surroundings, a “place of decaying 
substances and filth, a place full of putrid excrement,” could hardly be greater. 
The nugget remains hidden in this place for many hundreds of years. No one 
would expect such a place to be the depository of a gold nugget. The simile draws 
the reader’s attention to two points: first, the imperishable nature of the gold, 
which is later compared to the true nature of living beings. Its nature guarantees 
that the gold remains without any decay in spite of its surroundings. Secondly, the 
highly unexpected nature of the simile is meant to induce one to believe that the 
hidden buddha-nature will one day be revealed even in a sentient being who 
hardly anybody would think capable of becoming a buddha. The role of the 
divinity which in the upamana advises a person to look for the nugget and clean 
it, is filled in the upameya by the Dharma-teaching Tathagata. 

The topos of a hidden treasure beneath the house of some poor person for 
the unknown buddha-nature in living beings is, with some variations, also found 
in the Mahdparinirvanasitra.“ The simile in the TGS is replete with oppositions: 
The house owner (or nobleman) (grhapati) lives in poverty, though right below 
his house is a huge treasure of precious materials. He walks up and down lost in 
his thoughts, while the treasure, the solution to his problems, lies untouched. He is 
unaware of the treasure and nobody, not even the treasure itself, can inform him, 
even though it is a grhapati’s duty as one of the seven jewels of a cakravartin 
ruler to find hidden treasures. Indeed, he himself is obviously in need of a person 
with a supernatural eyesight to make him see that all he requires is beneath him. 
Such a person is not mentioned in the upamana of the simile, but appears as the 
Tathagata in the upameya. After the Tathagata has revealed the existence of such 
a treasure—living beings’ buddha-nature—to the bodhisattvas, they set about 
digging it out. It is clear that in this fifth illustration the author wanted to draw the 
attention of the readers to their unawareness of the “treasure” within themselves 
and, consequently, to the fact that there is no reason to think of themselves as 
spiritually poor. 

The sixth simile of a sprout in the seed and the eighth of the desperate 
woman with the cakravartin embryo in her womb apparently embody a different 
notion. Here, at first glance, it is the process of growing, the ripening of the sprout 
into a “great king of trees,” around which the comparison turns. This would at 

“ See MPNS S 184b1ff.; Q 105b6ff.; see also Takasaki 1974: 144f. 


PDF Version: BPPB VI (2002) 

least be the reader’s first impression. I will argue below that, though this notion. 
may be evoked, this is not the main intent of the authors, and that we should take 
care here not to fall into the trap of too allegorical an interpretation. A closer 
analysis of the wording will show rather that the essential sameness of sprout and 
full-grown tree, their alternate generation, and the fact that the tree is contained in 
its complete but not yet fully unfolded form already in the seed go to make up the 
focus of the simile. In contrast to most of the other similes no act of purification is 
needed. The ripening is described as a more or less automatic process once the 
seed has fallen on the soil. Section 6B, representing the upameya, can be said to 
be quite unusual for the TGS. The difficult and terminologically laden content 
differs in each translation, and follows a more abstract line: it seems to offer a 
definition of the term sattva, the term sattvadhatu appears, and describes a 
tathagata who discloses the Dharma as a “supreme, great accumulation of 

A dangerous wilderness, further on compared to samsara, is the setting of 
the seventh simile. The statuary image of a tathdgata wrapped in rotten rags was 
dropped and lies on the path. Travelers, unaware that the rags contain such an 
image, pass by in disgust. Only after a divinity orders a group of people to open 
the bundle does the image inside receive the veneration it deserves. In this simile 
two main features seem important to note: first, as with the gold nugget in 
excrement, the complete inappropriateness of where the object is deposited and 
thus the unexpectedness of the finding; secondly, the illustration plays up the 
contrast between contempt and veneration: nobody, however contemptible he or 
she may appear, should be looked upon as such, because everyone has a nature 
deserving of veneration. This is also the message found in the associated verse 
7.5, where the Tathagata wants the awakened bodhisattvas to know that all living 
beings have the buddha-nature and, one could add, therefore encourages them to 
apply themselves to the development of that spiritual capacity in living beings. 

A more direct call for concrete consequences based on the tathagata- 
garbha teaching is contained in the simile of the poor, depressed woman. An ugly, 
orphaned woman lives under desperate conditions in a poorhouse, despising 
herself. She is unaware of the fact that she has a future universal emperor 
(cakravartin)—the buddha-nature—in her womb. The illustration contrasts the 
apparently hopeless situation of the woman with the majesty and glory of the 
cakravartin. Of course, the idea of an embryonic cakravartin may, as in the sixth 
simile, embody the notion of a not yet fully developed element, the basis for 
awakening in the future. The main point in this simile, however, seems to be to 
show the Tathagata’s encouragement of living beings, who are again unaware of 
their buddha-nature, and his urging them to overcome defeatism: “Apply energy 
without giving in to despondency!” In this illustration, too, no activities to reveal 
the buddha-nature are necessary; the embryo is growing without help from 
outside, and his birth and future empowerment are inevitable. Section 8B contains 
the only passage in the TGS where the three-stage career sattva —> bodhisattva > 
tathagata is explicitly mentioned. 

In the last simile of golden figures within burned clay molds the process of 
casting golden statues according to the cire-perdue method is described. The 

45 See Takasaki 1974: 62ff. 


PDF Version: BPPB VI (2002) 

expertise of a smith is necessary in order to decide when the right time for striking 
the molds with a hammer has come so that—in an instant—the dirty black molds 
fall apart and the clean figures appear. For the Tathagata, too, “who uses the 
Dharma as a hammer,” the knowledge of the right time is necessary. The 
bodhisattvas he purifies have to “become calm and cool.” The illustration seems 
to stress the suddenness with which the golden figures come out. Nevertheless, 
this should probably not be taken as promoting the idea of a sudden awakening. 
Nothing that would substantiate such an idea occurs in the description of the 
spiritual sphere. This description serves rather to underscore the surprise and 
complete unexpectedness arising the moment the burned and ugly mold is 
replaced by a golden image from inside. The simile describes accurately and in 
detail this method of casting, which is still practiced among artisans in India and 

2 The Meaning and Occurrences of the Term tathagatagarbha 
2.1 The Term tathdgatagarbha 

As the summary of the similes above has revealed, the idea that all living beings 
have the buddha-nature is not illustrated in a completely homogeneous way. We 
should therefore try to grasp more clearly the conceptual elements which underlie 
the notion of this nature in the TGS and the relationship of the level of ordinary, 
still unrealized living beings to the level of a tathagata’s perfect realization. Here 
it may be useful to begin with the term tathagatagarbha, which gave its name to 
the siitra itself, and became well established in its Chinese translation rulai zang 
RAR as the main designation for the buddha-nature theory in East Asia. 

First of all, it should be noted that the term tathagatagarbha is not as 
pervasive in the TGS as one would expect, judging from the later tradition. Its use 
in the similes is restricted to the “new version” of the first simile.“® There, 
however, it occurs in the passages which seem to impart in concise form the main 
message of the sutra: In OM and 1A it is stated that the Tathagata perceives all 
living beings as tathagatagarbha.” Then, in 1B, a passage which is cited word for 
word in the Ratnagotravibhagavyakhya contains the main message of the TGS. 
The Sanskrit citation in the Ratnagotravibhagavyakhya reads as follows: 

es@ kulaputra dharmanam dharmata | utpadad va tathagatanam anutpadad va 
sadaivaite sattvas tathagatagarbha iti /(RGVV 73.11-12)”° 

“© Cf SP 11.48 (45.13f,). 

47 Reeves (1962: 118) reports that among smiths in Madras the hollow inside the mold is called 
karu (“womb” (= Skt. garbha); see Miron Winslow, Tamil-English-Dictionary, reprint of the 
edition from 1862, ed. K. L. Janert, Wiesbaden: Steiner, 1977, s.v. karu). 

48 See section 1.4. The “new version” comprises the last sentence in 0M, all of 1A, and the first 
sentence of 1B. 

“° OM (probably a later interpolation into TGS); see section 1.4): de bzhin gshegs pa’i mig des 
sems can thams cad de bzhin gshegs pa’i snying por mthong ngo //; 1A: rigs kyi bu dag de bzhin 
gshegs pas kyang sangs rgyas kyi mig gis sems can thams cad de bzhin gshegs pa’i snying por 
mthong nas / ... 

°° 1B: rigs kyi bu dag ‘di ni chos rnams kyi chos nyid de | de bzhin gshegs pa rnams byung 
yang rung ma byung yang rung /sems can ‘di dag ni rtag tu de bzhin gshegs pa’i snying po yin ... 


PDF Version: BPPB VI (2002) 

Son of good family, the True Nature (dharmata) of the dharmas is this: whether or 

not tathagatas appear in the world, all these sentient beings contain at all times a 

This passage was probably inserted together with the “new version” of the first 
simile. Placed at the end of the “new” simile, it was meant to summarize the main 
point of that simile, and should thus be considered the nucleus of the doctrinal 
message of the TGS. That the passage was felt to have such an eminent position is 
also attested by the fact that the Ratnagotravibhagavyakhya quotes it word for 
word, The author of this passage knowingly associates the claim that living beings 
are always tathagatagarbha with the well-known Buddhist formula dharmanam 
dharmata ..., which is usually employed to introduce the fundamental law of 
pratityasamutpada. This attests that he himself regarded the passage sadaivaite 
sattvas tathagatagarbhah as the essence of the sUtra, a standpoint which (if the 
authors of the “old” eight or nine similes and the “new lotus simile” are indeed 
different) would probably be shared by the composers or compilers of the other 

An interpretation of the meaning of the term ‘athagatagarbha must, as a 
matter of course, start from the context in the TGS in which it originates.°” The 
context is that of the withered lotuses with beautiful tathagatas sitting in the center 
of their calyxes (padmagarbha), In the same way that full-fledged tathagatas sit in 
the flowers, so also, according to the stitra, are buddhas contained in living 
beings.” If living beings are said to contain a tathagata, they should function as 
receptacles, and the compound tathagatagarbha must accordingly be understood 
either as a bahuvrihi in the sense of “containing a tathagata” or as a tatpurusa 
meaning “store of a tathagata.” However, in order to reach an adequate 
interpretation of the compound, I need to preface some remarks on the term 
garbha, and then give an overview of the range of possible interpretations of the 
whole compound, in part offered by the texts which succeeded the TGS.™ 

Concerning the term garbha, the Etymologisches Worterbuch des 
Altindoarischen * provides us only with the two meanings “Mutterleib; 
Leibesfrucht, Embryo, Neugeborenes.” It seems, however, that, starting from this 
biological background, garbha took on other, less specific senses, such as “the 
inside, middle, interior of anything, calyx (as of a lotus), ... any interior chamber, 
adytum or sanctuary of a temple &c.” (MW), or, as Hara has shown for epic 
literature, “germ, seed, infant, child” and, by analogy with the vocable putra as 
the last member of a compound, even simply “member (of a family lineage).”°° 
Also familiar is the function of -garbha at the end of a bahuvrihi compound, 
indicating that the prior member(s) of the compound is/are contained in the 

* See also Zimmermann 1999: 158. 

* For the occurrence of the term tathdgatagarbha in the Gandavyithasitra (GV 221.2-6), a 
controversial interpretation of it, and what this means for the relation between the 
Gandavythasitra and the TGS see Takasaki 2000: 75f. (with his references to Matsumoto 1994). 

>? Explicitly so stated in OM and verse 1.3. 

*4 An excellent discussion of this complex problem from different angles is found in Seyfort 
Ruegg 1969: 499ff.; 1976: 352f. 

°° Manfred Mayrhofer, Etymologisches Wérterbuch des Altindoarischen, 2 vols., Heidelberg: 
Carl Winter Universitatsverlag, 1992-1996, s.v. garbha. 

°° See Hara 1994, 


PDF Version: BPPB VI (2002) 

subject the compound refers to.°’ Especially in the last case it is difficult to judge 
how far a biological, embryonic shade of meaning was still felt, that is, to what 
degree -garbha had become a purely grammatical unit used at the time our sutra 
came into existence to express a relation of inclusion void of any strong lexical 
connotations. However, the original embryo-related meaning of garbha did not 
completely fade out in later centuries; indeed garbha even became associated with 
the semantic field of “offspring.” This suggests that the grammatical application 
of -garbha never became totally free of the underlying idea of an embryo still in . 
need of development in a nurturing, womb-like container, if the context in 
question was susceptible of such a nuance. 

But there is another secondary semantic field with which the term garbha 
seems to be closely associated, namely the “central part” or “essence” of 
something. This development is mirrored in the regular choice of Indo-Tibetan 
translators to render garbha into Tibetan as snying po: “the chief part, main 
substance, quintessence” (Jd s.v. snying po; see also snying: “the heart, ... the 
mind”). To be sure, this meaning is not supported by the Sanskrit dictionaries, nor 
by Pali. At best, the meanings “inner chamber” or “adytum” as the central and 
most holy part of a temple might suggest something in this direction. However, 
the corresponding forms of garbha in a number of modern Indian languages can 
also have the meaning of “core,” “heart,” “pith” and the like, and this thus further 
strengthens the case that Skt. garbha from a certain time on must have embraced 
this sense.” If this is so, we must also allow that the employment of -garbha at 
the end of a bahuvrihi compound could have such a nuance: “to contain 
essentially.” Whatever the case, from a certain time on, as Matsumoto has 
demonstrated, the use of the term garbha became widespread in Indian Buddhist 
scriptures, for example, in the Gandavyitha and the Dasabhimikasitra.” Names 
containing garbha as a member, such as the long list of bodhisattva names in the 
introduction of the Dasabhimikasutra (a list which is not yet in the earlier 
translation of Dharmaraksa), came into use. The term garbha proved highly 
popular, and one wonders if it was always possible to decide on an exact 
equivalent in a translation. Indeed it seems that once garbha became a fashionable 
term it was easily compounded with other words without itself exhibiting a very 
concrete meaning. 

Let us now turn to the compound itself. I have grouped the alternatives 
according to grammatical categories. 

57 See PW s.v. garbha: “Uebertragen am Ende von adj. compp. (f. 4): dieses als Leibesfrucht 
tragend, in seinem Innern bergend, enthaltend....” 

*8 See Turner no. 4055 (garbha-): “... Sindhi gabhu ... kernel, pith; ... Nepali ... gabho core, 
inside (e.g. of a fruit); ... Hindi ... gab pulp, pith; ... Marathi ... gabha heart, core; ... Konkani ... 
gabbo inner core of plaintain stem....” H. Isaacson kindly drew my attention to three relatively late 
passages: one in the Yogaratnamald, a commentary on the Hevajratantra, where garbha in the 
term vajragarbha is synonymized as hydaya (David L. Snellgrove, The Hevajra Tantra, A Critical 
Study, Part 2: Sanskrit and Tibetan Texts, London: Oxford Univ. Press, 1959, p. 136); the second, 
in Ramakarasanti’s as yet unpublished Muktavali, another commentary on the Hevajratantra, 
where the same gloss is given (National Archives, Kathmandu, MS 4-19 = Nepal-German 
Manuscript Preservation Project Reel No. A 994/6, f. 80rl; Tokyo University MS 513, f. 57r6); the 
third, in the commentary on the Tantraloka, the Tantralokaviveka by Jayaratha, where garbha is 
interpreted as sdra (Mukund Ram Shastri (ed.), The Tantraloka of Abhinava Gupta with 
Commentary by Rajadnaka Jayaratha, vol. 1, Kashmir Series of Texts and Studies 23, Allahabad: 
The Indian Press, 1918, commentary to 2.32). 

°° See Matsumoto 1994: 41 1ff., in particular pp. 421, 460, 471. 


PDF Version: BPPB VI (2002) 

@ Tathadgatagarbha as a tatpurusa compound. Here three cases need to be 

a. The prior member of the compound can be determined as having an elided 
genitive case ending: tathdgatasya or tathagatanam garbhah: “the embryo of 
a/the tathagata(s).”© This interpretation seems to be the most common one in 
the tradition following the TGS. The word can in this case be understood as 
referring to the “embryo of a/the tathagata(s)” within all living beings. In the 
TGS, however, tathagatagarbha cannot have this meaning for two reasons. 
First, in the statement sattvas tathagatagarbhah the compound cannot be 
understood as a tatpurusa designating the “embryo within” because 
grammatically it is the predicate of sattvah. Secondly, as mentioned above, 
rather than with the embryo of a/the tathagata(s) the TGS deals with full- 
fledged tathagatas in flowers (and in living beings). 

b. A living being as the “womb of a tathagata”: A living being as the womb of a 
tathagata seems to accord with the image of lotuses enclosing tathagatas within 
their womb-like calyxes in the upamana. Matsumoto adopts the translation 
“container of a tathagata” (IKOAMbD) for the TGS, rejecting 
Schmithausen’s and Ruegg’s categorical statement that the compound 
tathagatagarbha can never mean “matrix of the Tathagata” in Indian texts.°! In 
his argument why the compound within the TGS should be interpreted as a 
tatpurusa, Matsumoto obviously does not consider the alternative provided in 
@a below. He argues that the author of the TGS had first to “invent” the 
concept of padmagarbha, the “lotus calyx” where the tathagatas are seated, in 
order to construct the analogy between the lotuses and living beings. Once 
padmagarbha was adopted as the interior space of a lotus GE#EOWESZ2 )), 
the author of the TGS could develop the analogy of tathagatas seated in the 
interior space (garbha) of living beings. Matsumoto asserts that once living 
beings are the referent, they could accordingly be denoted as “containers of a 
tathagata.” He admits that the author of the TGS thus operates with two 
meanings of garbha, namely “interior space” and “container,” and that the 
author was well aware of this fact.” 

Now, I think Matsumoto’s argumentation is less than conclusive. First of 
all, the meaning of “container” for garbha seems not to be very common, if 
documented at all. On the basis of its meaning “womb,” garbha designates a 
space within rather than a container or a sheath around something. This is clear 
from the entry under garbha found in the PW and MW: “Mutterleib,” “interior 
chamber.” Regarding the second point, I agree with Matsumoto that the terms 
padmagarbha and tathagatagarbha are probably intentionally employed by the 
authors in some kind of analogy. But if this is the case, should we then not 
argue that -garbha expresses in both cases a comparable sense, namely the 
inside of a lotus and the inside of living beings respectively? This at least 
would make things less complicated. Alternative @a proceeds along these 
lines. Admittedly, from a purely grammatical point of view Matsumoto’s 
argumentation avoids the somewhat disturbing situation that the compounds, if 

See e.g. LAS 223.14: garbhas tathagatanam; LAS 357.15 (Sagathakam 346c): garbhas 

51 See Matsumoto 1994: 498ff., referring to Schmithausen 1971: 133, n. 44 (and further p. 156), 
who finds confirmation in Seyfort Ruegg 1969: 499ff. 

® See Matsumoto 1994: 495ff.; 526f. 


PDF Version: BPPB VI (2002) 

we follow @a, belong to two distinct categories, namely fatpurusa 
(padmagarbha) and bahuvrihi (tathagatagarbha). However, even if we follow 
his suggestion, the relation between the members within the two tafpurusa 
compounds would have to be analyzed differently as a genitivus subjectivus in 
the first case and a genitivus objectivus in the latter. 

One final remark: Matsumoto insists, on the one hand, on employing 
garbha in the sense of “receptacle,” while, on the other, he vehemently rejects 
the translation “womb” or “matrix,” the rendering adopted by Takasaki in his 
English translation of the Ratnagotravibhaga(vyakhya).” 1 do not find any 
fundamental difference between the two concepts, and do not think that the 
translation “womb of a tathagata” is impossible, at least any more so than 
Matsumoto’s own rendering. 

c. The living being as “embryo of the Tathagata”: The idea that all living beings 
are penetrated by the dharmakdya and thus are the embryos or, as Hara has 
shown, the children of the Tathagata is found in __ the 
Ratnagotravibhagavyakhya. 6 Nothing in the 7GS suggests that this 
understanding could be applied to the relation between living beings and the 
Tathagata. It is therefore not relevant for our discussion here. 

@® Tathagatagarbha as a bahuvrihi compound: 

a. The most “natural” analysis of the compound would be as a bahuvrihi. The 
PW (s.v. garbha) provides us with plenty of examples with garbha at the end 
of the compound meaning simply “to contain” (see above). The sentence 
sattvas tathagatagarbhah should thus be understood as “Living beings contain 
a tathagata.” (tathagato garbho yesadm te tathagatagarbhah).© In order, 
however, to understand garbha in a more characteristic way, we need to bring 
out its biological nuance or else the nuance, assumed above, of essence. The 
first case would result in a translation (1) “Living beings contain a tathagata as 
an embryo.”© In the second alternative we would have to translate (2) “... have 
a tathagata at their core,” or, more abstractly, “... have a tathadgata as their 
nature.” The image of the lotus renders both alternatives possible, since the 
tathagatas are seated in the center of their womb-like calyxes (garbha). They 
thus have become the core of the lotus, unaffected by the withered parts of the 
flowers, and “emitting hundreds of thousands of rays of light.” Nevertheless, in 
this first simile, as in most of the following illustrations, no growing process or 

6 See Matsumoto 1994: 498; Takasaki, in his later translation of the RGV(V), no longer 
employs this rendering (Takasaki 1989). 

& RGVV 70.16—18: tathégatadharmakayena niravaSesasattvadhatuparispharanartham 
adhikytya tathagatasyeme garbhah sarvasattva iti paridipitam /. For a different interpretation, cp. 
Matsumoto 1994: 498ff. 

§ See e.g. the parallel verse to the first simile in the RGV (1.101) where the following 
compound referring to living beings appears: ... sambuddhagarbham jagat: “‘... [having seen] 
living beings as containing a perfect buddha.” 

© This alternative would semantically be the equivalent of “.., have a tathagata in their womb” 
(tathdgato garbhe yesam te tathagatagarbhah), which is based on AiGr II, 1 § 109c, p. 279 
(“Bahuvrihis mit kasuellem Vorderglied”): “Sehr oft bezeichnet das Hinterglied einen Kérperteil 
und dann das Kompositum denjenigen, der am betr. Kérperteil etwas im Vorderglied genanntes 
halt oder haften hat z.B. ... vajra-ba@hu- der den Donnerkeil im Arme hat ... kilalodhni Kilala im 
Euter habend....” In contrast to the rendering above (“‘... contain a tathagata as an embryo”), the 
main thrust here should not be to characterize the tathigata as an embryo but to point out the fact 
that he is found in the womb. 


PDF Version: BPPB VI (2002) 

any other process of change is implied regarding the tathagatas. To speak of a 
tathagata-embryo can therefore be quite misleading, since the process of a 
gradual development and maturing is usually implied when speaking of an 
embryo in the womb. For this reason I have decided to adopt the simple 
rendering “... contain a tathagata” in the translation. That, as discussed in the 
second alternative, this tathagata represents the nature of all living beings need 
not be mentioned explicitly. It is the aim of the sitra itself to illustrate this fact 
by way of the nine similes. 

b. Although in @a. the term garbha was understood in the “weak” sense of 
“containing,” it can also be interpreted in its original meaning “womb” or 
“embryo.” Taking it as such, we would have to translate the whole compound 
in a possessive relation to living beings as: 

I. “Living beings have an embryo of a tathagata” (tathdgatasya garbho (’sti) 
yesam te tathagatagarbhah). 
II. (1) “... have a tathagata in their womb (garbha),”®’ or 
(2) “... have a tathagata as an embryo (garbha).” 
III. Or, assuming that garbha can have the meaning “core, essence,” as 
(1) “... have the tathigata-core” or 
(2) “... have a tathagata at their core.” 

Position III.(1) mirrors the understanding of the Tibetan when rendering the 
statement cited above as sems can ‘di dag ni rtag tu de bzhin gshegs pa’i snying 
po (can) yin... 

Semantically, there is no difference between having the embryo of a 
tathagata (@b.]) or containing a tathagata as an embryo (@a.1/@b.II (2)). In 
both cases the embryo still has to develop into a full-fledged tathagata,” a 
notion by no means expressed in the lotus image. One might argue that the 
situation in the upameya cannot be compared very smoothly to the upamdna. 
However, the other parts of the first simile clearly demonstrate that it is not an 
embryonic seed or such that is contained in living beings but a tathagata itself. 
Further, if a philological analysis like that provided in @a allows us to interpret 
the compound in a sense not contradictory to the setting of the material sphere, 
there can be no reason to adopt another alternative. 

Regarding the third translation in @b, namely that of III.(1) “... have the 
tathagata-core” or III.(2) “... have a tathagata at their core,” III.(2) is simply a 
grammatically different rendering of the translation given in @a.2 (“... have a 
tathagata as their nature.”). If we follow III.(1), the alternative that introduces 
the tathagata-core as a separate element, we reach a more abstract formulation 
of the same basic thought, more or less disengaged from the concreteness of 
the material sphere. This rendering manifests a stronger monistic inclination, 

$7 An analysis of this alternative has been discussed in the preceding note. 

§ For the inclusion of the particle can see below. 

® Prof. Seyfort Ruegg, in a personal communication, warns that to associate the word 
“embryo” with the ripening process ahead could lend itself to misunderstanding. If I interpret him 
correctly, he is implying that the term “embryo” (garbha) in India at that time primarily denoted a 
nucleus or core embedded in some shielding environment rather than a still completely mother- 
dependent being undergoing further growth. I agree with him insofar as the term garbha in its 
meaning “embryo” might certainly have evoked different associations according to the field in 
which it was applied (anatomy, genealogy, etc.). However, I do not think that the general 
readership of the stitra, be it in ancient India or in the modern world, would exclude the notion of 
ripening in connection with the term “embryo.” 


PDF Version: BPPB VI (2002) 

since it may be supposed that there is only one such buddha-nature. In contrast 
to this, the philosophically somewhat clumsy rendering of III.(2) has to a 
greater degree remained true to the imagery of a multiplicity of buddhas sitting 
each in his own calyx. 

c. A final theoretically possible interpretation is that of tathagatagarbha in the 
sense of “having a tathagata as a womb,” that is, “born from a tathagata.””° The 
compound would then refer to living beings themselves as “born from a 
tathagata/the Tathagata.” Whereas the simile of the cakravartin in the poor 
woman is based on the biological metaphor of an embryo found in the womb, 
none of the other similes involves such a context. Nor is the cakravartin said to 
be born from a tathagata. The interpretation “born from a tathagata/the 
Tathagata” would thus violate the thrust of the upamdnas and so cannot be 
applied in the TGS. 

The above analysis has revealed that, in light of the upamana, the grammatically 
adequate interpretation serving as the basic one of the compound tathagatagarbha 
is that of a bahuvrihi meaning “containing a tathagata.” However, it would be 
wrong to assume that the term fathagatagarbha, once the context of the first 
simile is left behind, would be restricted in the minds of readers to such a clear-cut 
definition. Having read the other eight similes, among them two employing the 
image of a seed and an embryo, the term garbha in its meaning “embryo” would 
automatically acquire greater prominence, and possibly overshadow the first 
interpretation based on the lotus imagery. The richness of the term garbha, which 
means “containing,” “born from,” “embryo,” “(embracing/concealing) womb,” 
“calyx,” “child,” “member of a clan” and even “core” would from the very 
beginning tend to keep its semantic range from being reduced to a single word. 

To judge by the textual history of the TGS, on the other hand, the 
introduction of the term tathagatagarbha with the “new lotus simile” (1A) took 
place after all the other similes had already been composed. We may thus assume 
that the author of this simile chose that term because it covered the whole variety 
of themes illustrated in the similes.’’ At the same time, garbha is also part of 
padmagarbha, a term expressing a central image in the introduction. The 
compound padmagarbha itself is also well known as a designation for the god 
Brahman, who is said to be “born from a lotus” (padmagarbha as bahuvrihi). This 
may have been another, highly venerable association for someone reading in the 
sutra of a tathagata seated in a padmagarbha. As attested by the prevailing use of 
the word tathagatagarbha in other writings, and in the Ratnagotravibhaga- 
vyakhya itself, as a tatpurusa compound (the “embryo of a tathagata” as a separate 
entity), a tendency towards a grammatically non-adjectival use of the term was 
inherent from earliest times. The title of the sitra further stimulated such an 
independent understanding, and with the appearance of the concept of 
buddhadhatu in the Mahayanist Mahaparinirvanasitra as more or less a synonym 
for tathagatagarbha, the temptation to interpret this latter as a tatpurusa probably 
proved irresistible. In the TGS this is documented by at least three of the four 
translations into Chinese and Tibetan, where the compound is regularly rendered 

™ See Seyfort Ruegg 1969: 511, n. 4. 
™ The vocable dhatu, too, has a wide range of meanings, the most notable being “primitive 
matter, constituent part of the body, mineral, relics.” 


PDF Version: BPPB VI (2002) 

as a separate entity found in living beings in the sense of “store of a tathagata” or 
“essence of a tathagata.””” 

2.2 The Textual Occurrences of the Terms tathdgatagarbha and garbha 

Besides the title of the siitra itself, the term tathagatagarbha is found in several 
passages. Let us, in the following, see to what extent the passages containing the 
whole compound or merely the word garbha fit the interpretations suggested 

1. In a series of nearly identical formulations, the TGS repeats three times that 
living beings are tathagatagarbha (0M, 1A, 1B).” The passage in 1B is cited 
in the Ratnagotravibhagavyakhya, and thus we know the Sanskrit text: 
sadaivaite sattvas tathagatagarbhah. Bth renders the compound as de bzhin 
gshegs pa’i snying po. The manuscript tradition in Tib, however, observes two 
ways of rendering it. One is identical with Bth. A second strand adds the particle 
can at the end of the compound: de bzhin gshegs pa’i snying po can. This second 
strand is represented by: 

* BDJNQ in 0M, 
* by no manuscript or print in 1A, and by 
¢ DESTP; in 1B. 

Now, the old translation Bth does not have the particle can. From this and the 
fact that no edition in 1A, and only some manuscripts and prints in the other 
passages, testify to the particle, we can be fairly sure that it was only inserted 
by Tibetan editors at some late revisional stage. The particle can, as Seyfort 
Ruegg points out,” may serve to indicate that the translated compound was 
understood as a bahuvrihi. The fact that it is not found in the earlier 
transmission could mean that at that time the rendering of a bahuvrihi 
compound by means of can was not yet common practice, if ever it was, or that 
they tried to retain the interpretational ambiguity inherent in the Sanskrit 
compound by resorting to a rigorously literal translation technique. A Tibetan 
reader with a background in Indian grammar would probably be aware of the 
range of possible interpretations of such a compound in Tibetan. For a less 
grammatically educated Tibetan, however, a sentence like sems can ‘di dag ni 
rtag tu de bzhin gshegs pa’i snying po yin.... would likely be understood as 
“All these sentient beings are at all times the essence of the Tathagata,” a 
statement which does not seem to be very precise. This disturbing fact may 
have been the reason why later redactors of Tib decided to add the particle can. 
They felt it necessary to differentiate explicitly between living beings and the 
essence contained in them, or, in the words of Seyfort Ruegg, to point out the 
difference between guna (“attribute”) and gunin (“bearer of the attribute”).”° 

” For the Tibetan translation from Bathang see below. 

73 Tib: ... sems can di dag ni rtag tu de bzhin gshegs pa’i snying po (can) yin ... 

4 See Seyfort Ruegg 1969: 510. 

5 For the particle can in the concerned compound in the texts of the tathagatagarbha theory 
see Seyfort Ruegg 1969: 509ff. On p. 511 Seyfort Ruegg argues that the TGS itself denies the 
identification of living beings (sattva) as tathagatagarbha. He supposes that the treasure in the 
fifth simile is compared to the tathagatagarbha and argues that there the treasure is said to be the 
cittasvabhava (‘l’étre propre de la Pensée”) rather than the sattva (5A). However, I cannot find 


PDF Version: BPPB VI (2002) 

2. Verse 1.1 contains the compound tathagatagarbha in pada c. 

Tib: Bth: 

ji ltar padma smad par ’os gyur pa // ci ltar pad mo smad' pa de ni 

de’i mdab ma sbubs gyur ma gyes la // de’ilo ma rnams kyi +++ ma gyes | 

de bzhin gshegs pa’i snying po ma gos te // de bzhin gshegs pai! snying po de 
yang ni // 

mi ’ga’ la las lha yi mig gis mthong // dri ma ma gos skyes’ pas ’phrul 
mig’ mthong // 
' for spyad;* for bskyes;? for mi. 

Ch I Ch2 



Parallel verse Ratnagotravibhaga 1.99 
yatha vivarnambujagarbhavestitam 
tathagatam diptasahasralaksanam /| 
narah samiksyamaladivyalocano 
vimocayed ambujapattrakoSsatah // 

Parallel verse Ratnagotravibhaga 1.101 

yadvat syad vijugupsitam jalaruham samminjitam divyadyk 
tadgarbhasthitam abhyudiksya sugatarh patrani samchedayet / 
ragadvesamaladikosanivytam sambuddhagarbham jagat 
karunyad avalokya tannivaranam nirhanti tadvan munih // 

A translation based on Tib would most naturally run: 
It is as if [there is] a disgusting lotus whose [unsightly] sheath-[like] petals are 
not opened out, yet the tathagatagarbha is unpolluted [by the petals], and a 
person with divine vision will perceive [it]. 

The fact that the term tathagatagarbha can here hardly be taken as a bahuvrihi 
teferring to living beings, the usage attested for OM, 1A and 1B, is cause for 
surprise, and sets this verse off from the prose section. The term 
tathagatagarbha is not expected in the upamana, it clearly being an element of 
the upameya. Had the Tibetan versions, as Ch; does (40465), employed Jus or 
gzugs in the upamdna instead of snying po (Skt. garbha), there would have 
been no irregularity. However, both Tibetan versions read snying po. 

A clue to how the pada should be understood can be obtained by reference 
to the parallel verses in the Ratnagotravibhaga. There, in the first pada of 1.99, 
the tathagata in the upamana is described as vivarnambujagarbhavestita: 
“enclosed in the calyx of a lotus of ugly color.” Similarly in 
Ratnagotravibhaga 1.101 the sugata is said to be found in “its (= the lotus’s) 
calyx” (tadgarbhasthita). In both cases the term garbha appears, and 
designates clearly the inside, the calyx, of the lotus. If we interpret the 
compound in 7GS 1.1 in this way, namely with garbha referring to the inside 
of the lotus, we end up with a reading perfectly in accordance with the situation 

any explicit identification of the treasure with the term tathagatagarbha, Accordingly, and in light 
of the version from Bathang, my understanding of the passage differs from Seyfort Ruegg’s 
translation (see the note in my translation). Further, in 5B the bodhisattvas are compared to the 
treasury of the qualities of a buddha. 


PDF Version: BPPB VI (2002) 


described in the prose before: “... yet the inside [of the lotus containing a] 
tathagata is unpolluted [by the petals],....”’° This reading makes it very clear 
that garbha can only be taken as the inside of the flower, in contrast to the 
sheath-like petals mentioned in the pada before. In the TGS these tend to be 
referred to by the term kosa. 

3. A translation of the compound fathdgatagarbha also appears in all 
manuscripts and prints of Tib in 5A, in an enumeration of buddha-qualities said 
to be present in living beings.’’ The only way of interpreting the compound 
there is as a separate entity found in living beings. However, a parallel in 5B, 
Bth and the Chinese versions leads one to believe that the reading should be 
tathagatajnana instead of °garbha: 

* For the parallel with ye shes instead of snying po in Tib see 5B.12f. 
* Bth: de bzhin gshegs pa’i yeshes dang : stobs dang : mi ‘jig pa dang : 
sangs rgyas kyi chos ma ’dres pa mdzod // 
Chr: HORA. JJ. RRA RRS, ... 
* Chr: MARIRER. 7. FAT Raa, ... 
I have accordingly concluded (see translation and note) that the reading in the 
Sanskrit must originally have been tathagatajfiana. 

4. Much more problematic is verse 10.10.”* Given that the content of this verse 
varies widely in the different translations, it would seem impossible to draw 
conclusions about the original Sanskrit text. However, the lotuses have 
withdrawn far into the background, so that an interpretation like the one 
suggested in my translation, with tathagatagarbha meaning “embryo of a 
tathagata,” seems all but unavoidable. 

The verse has no correspondence in the prose, which deals with the merit 
attained through propagating the TGS. The verse parallels to the prose come to 
a natural end in verse 10.9. Verse 10.10 causes wonder owing to its sudden 
return to a doctrinal message and the occurrence of dharmata, which is referred 
to with the pronoun ‘di, even though it has not been mentioned before. We 
have therefore good reason to assume that the verse 10.10 is a later 
interpolation into the common ancestor of TGS; and TGS). It could in fact be a 
citation from another work. If so, it would not come as a surprise that our 
compound turns out to be a tatpurusa—the prevailing usage in later times. 

The term garbha appears frequently in the tatpurusa compound padmagarbha and 
also as a single word designating the calyx of a lotus (in the sense of its interior). 
This usage is attested in 0G, OH (twice), OI, OM, 1A and, if my analysis above is 
right, also in 1.1 as part of the compound tathagatagarbha (tatpurusa; see above). 

In verse 1.3, Bth twice uses the term snying po, once to designate the 
inside of living beings and then to refer to the lotus whose snying po is said to be 

76 Alternatively, the Sanskrit construction may have been ftathdgato/*tam garbha 
anupaliptahl"tam or something similar (depending on the metrical structure; garbha for the 
locative case garbhe). If we assume that the verses were composed in Buddhist Hybrid Sanskrit, 
we might even expect simply tathagata garbha ..., with -ta replacing a classical Skt. -tah or tam. 
Later translators, already used to the compound tathagatagarbha, would then have automatically 
interpreted the phrases as a compound. 

™ The grammatical relation between the compound and the following qualities is not clear: 
whereas L, S and T place a genitive particle between de bzhin gshegs pa'i snying po and stobs 
dang /..., in all other texts of Tib no such particle is found. 

78 See also the note in the translation. 


PDF Version: BPPB VI (2002) 

disgusting. The word for the inside of living beings (Ch2: AY) in Tib is dkyil. If the 
Sanskrit for dkyil (Tib) or snying po (Bth) was in fact garbha, as the rendering of 
the passage in the Ratnagotravibhdga suggests (sambuddhagarbham jagat; 
1.101),” it would be hard to explain why the translators of Tib refrained from a 
rendering with snying po for the “inside” of living beings. I thus tend to assume 
that the Sanskrit had madhya or something similar instead. 

In the second case, Bth speaks of the disgusting calyx (snying po), which 
is compared to the coverings of defilements in the upameya. Tib does not have the 
term snying po and instead employs sbubs (so also Ch): {4¢#8), a term in Tib 
usually associated with the klesas, an element of the upameya. However, in 1.1 
Tib also compares the petals of the lotus to sheaths (de’i mdab ma sbubs gyur).*° 
In view of the meaning “interior of something” of garbha and the clear 
differentiation in verse 1.1 between snying po (garbha) as the inside of the calyx, 
on the one hand, and the sheath-like petals that constitute the enclosure, on the 
other, the use of snying po in this context is odd: in verse 1.1 above it was said 
that the snying po of the lotus remains unpolluted. To now call this same snying 
po disgusting is a clear contradiction, obviously not noticed by the translators of 
Bth. The term sbubs, probably a rendering of Skt. kosa, is not found once in Bth. 
To denote the sheaths of klesas, Bth resorts to mdzod, another translation for kosa. 
One can only wonder why, in the case of the lotus sheaths (*padmakoSa), the 
translators of Bth refrained from a translation with mdzod and applied the 
obviously unfitting snying po. 

In 6B the passage sbubs kyi nang na snying por gyur pa de bzhin gshegs 
pa’i chos nyid could be interpreted as “... the true nature of the tathagatas, having 
become the essence inside the sheaths...” or “... that true nature of a tathagata in 
an embryo-like state” (see note in the translation). The compound snying por gyur 
pa probably renders Skt. garbhastha or garbhagata, since Bth reads ... dbus* kyi 
snying por de bzhin gshegs pa’i chos nyid kyang gnaso // [* for dbul]. As an 
analysis of compounds of this form—though chiefly restricted to the 
Mahabharata—in Hara 1994: 38, n. 4(1) has proved, garbha- bears in this 
combination the meaning “womb,” and should consequently—if probably against 
the intention of the translators—also be understood in this way in our passage: “... 
the true nature of the tathagatas who are in the womb/who are in the interior/who 
are within.” The parallel verse in the Ratnagotravibhdga (1.117) reads 
phalatvagantaragatah sambuddhabijankurah. There °gata also refers to the place 
and not to the state in which the sprout is found (for which one would expect a 
formulation such as ... garbhatvam gata). However, the translators of Tib may 
have been thinking that snying po designates here the embryonic essence and not 
the womb. Besides this instance there is no other example of snying po denoting 
the inside or the womb of living beings in Tib. 

In Bth the term snying po gnas for, probably, garbhasthana or 
garbhavasati (“womb”) appears in 8.5. The equivalent in Tib is mngal gnas, 
which usually means “embryo,” but which I take as a literal rendering of the Skt. 
compound. (See the note in my translation for further arguments.) 

In a passage in 8B, Tib designates the interior of living beings with khong, 
while Bth has snying po. Fortunately, this passage has been preserved in Sanskrit 

The compound sartbuddhagarbha in 1.101 is a bahuvrihi. The grammatical structure of the 
verse in the TGS is different. 
8° Unfortunately the corresponding word is not readable in Bth. 


PDF Version: BPPB VI (2002) 

in a quotation of the Ratnagotravibhdgavyakhya (72.11-12), where we find 
garbhagata. The reason why Tib prefers khong could be a word play with khong 
in the part immediately following: ... khong na yod kyang sems can de dag gis 
khong du ma chud do //. On the other hand, in view of the fact that, in contrast to 
Bth, Tib avoids snying po as the term to designate the interior of living beings in 
8.5, and possibly also 1.3, one could argue that the translators of Tib refrained 
from using the term snying po when speaking about the inside of living beings and 
reserved it for exclusive use in the upameya, in cases where they meant it to imply 
“essence, embryo” (as in 6B). This would mean that they wanted to employ the 
vocable snying po in only one meaning in each part of the simile, namely lotus 
“calyx” in the upamdna and “essence, embryo” in the upameya, whilst the 
translators of Bth stuck to a rigorously literal translation, mechanically rendering 
garbha as snying po in all instances, 

Finally, garbha appears as a part of names. There are two bodhisattva 
names including garbha in OE: (20) Srigarbha, Tib: dPal gyi snying po (Bth: dPal 
kyis snying po) and (21) Suryagarbha, Tib/Bth: Nyi ma’i snying po. The name of 
the pavilion where the Buddha is staying is given as “Candanagarbha” (in the 
compound candanagarbhakutagara): tsan dan gyi snying po’i khang pa brtsegs 
pa (see my note in OB). Bodhisattva names with °garbha at the end of a 
compound are common in Buddhist literature. Garbha as part of the pavilion’s 
name was probably chosen or added*’ owing to the central role of the term garbha 
in the description of the introductory scene in the upamdna (viz., padmagarbha) 
and the upameya (viz., tathdgatagarbha). 

3 The Buddha-Nature Doctrine in the 7GS 

As has become clear in the summary of the similes above, the doctrine of the TGS 
describes living beings as bearing a full-fledged tathagata within themselves. 
Though this is the case, living beings are not yet buddhas, since they are not 
aware of their own precious content, the tathagatas within being covered by 
sheaths of klesas. Nevertheless, it is stated that the nature of living beings is not 
different from that of the Buddha and these living beings will become buddhas 
themselves once the sheaths of defilements have been removed. In the following I 
shall deal with three crucial questions: 

¢ What do the terms used by the authors to designate the buddha-nature of 

living beings imply? 
* How can the process of attaining buddhahood be described? 
* What actually leads to that realization of buddhahood? 

3.1 The Buddha-Nature 

Regarding the first query, we need to look at the terminology utilized in the 
upameyas when referring to the buddha-nature of living beings in the state of not 
yet being tathagatas. In the chart below I shall restrict myself to the Tibetan 

8! In Ch, a translation for °garbha is missing. 


PDF Version: BPPB VI (2002) 

wording and its inferred Sanskrit equivalents:*” 

OM: de bzhin gshegs pa; tathagata 
de bzhin gshegs pa’i chos nyid; tathagatadharmata 
de bzhin gshegs pa’i snying po (can); tathagatagarbha (“containing a tathagata’ 
1A: de bzhin gshegs pa’i snying po; tathagatagarbha (“containing a tathagata”) 
de bzhin gshegs pa; tathagata 
1B: de bzhin gshegs pa’i snying po (can); tathagatagarbha (“containing a tathagata”) 
de bzhin gshegs pa’i ye shes; tathagatajnana 
1C: rgyal ba rnams kyi lus; *jinakaya 
rgyal ba’i lus; *jinakaya 
2A: sangs rgyas nyid; buddhatval°ta 
Sangs rgyas nyid; buddhatval°ta 
2B: de bzhin gshegs pa’i ye shes mthong ba; _ tathagatajndnadarsana 
2C: de bzhin gshegs; tathagata 
Sangs rgyas; buddha 
3B: de bzhin gshegs pa nyid; tathagatatval°ta 
sangs rgyas nyid; buddhatval°ta 
rang byung nyid; svayambhutva 
de bzhin gshegs pa nyid; tathagatatva/°ta 
3C: sangs rgyas sa; buddhabhumi 
nga ‘dra’i chos nyid, *matsamadharmata 
4B: chud mi za ba’i chos can; *avinasadharmin (cp. RGV 108d) 
(Bth instead: chos nyid ma rung bar (avinasadharmata) 
mi ‘gyur ba) 
de bzhin gshegs pa’i chos nyid chud *tathadgatasya avinasadharmata 
mi za ba; 
4C: rang bzhin; prakyti 
5A: de bzhin gshegs pa’i snying po’i stobs *tathagatagarbhabalavaisaradyavenika- 
dang / mi ‘jigs pa dang / ma ‘dres buddhadharmakosamahanidhi 

pa dang / sangs rgyas kyi chos thams 
cad kyi mdzod kyi gter chen po, 

chos kyi gter chen po; *mahadharmanidhi 
5B: chos kyi gter chen po; *mahadharmanidhi 
chos kyi gter chen po; *mahadharmanidhi 
de bzhin gshegs pa’i ye shes dang / *tathagatajnanabalavaisaradyavenika- 
stobs dang | mi ‘jigs pa dang / sangs buddhadharmakosa 
rgyas kyi chos ma ’dres pa’i mdzod; 
5C: gter chen; *mahanidhi 
bde gshegs lus; *sugatakaya 
nga [= sangs rgyas] yi ye shes mdzod; *maj[= buddha]jfianakosa 
gter; *nidhi 
6B: snying por gyur pa de bzhin gshegs *garbhagatal°stha tathagatadharmata 
pa’ichos nyid; (“... Which is in the womb”, see. n. in transl.) 
de bzhin gshegs pa’i ye shes; tathagatajnana 
6C: bde gshegs lus; *sugatakaya 
7B: de bzhin gshegs pa’i lus; *tathagatakadya 
de bzhin gshegs pa’i ye shes mthong ba; _— tath@gatajnanadarsana 
7C: rgyal ba’i sku; *jinakaya 

®2 For the Chinese equivalents see Takasaki 1974: 48-53. 


PDF Version: BPPB VI (2002) 

rgyal ba; Jina 
bde bar gshegs kyi ye shes; sugatajnana 
8B: de bzhin gshegs pa’i rigs, but *tathagatagotra: but quoted in RGVV as 
Bth: de bzhin gshegs pa’i tathagatadhatu 
khams; Chz: HRA; 
de bzhin gshegs pa; tathagata 
8C: chos nyid; dharmata 
chos nyid; dharmata 
9B: sangs rgyas kyi chos; buddhadharmah 
zag pa med pa’i ye shes rin po che; anasravajfhanaraina 
de bzhin gshegs pa; tathagata 
de bzhin gshegs pa’i ye shes rin po che; tathagatajfanaratna 
9C: sangs rgyas ye shes; buddhajniana 
rgyal sras; *jinaputra 
stobs becu; dasa balani 

The chart shows that in one third of all cases the buddha-nature is referred to by 
the term buddha, tathagata, jina, *jinakaya, *sugatakaya etc., the denotatum of 
which is said to be found within living beings. It is surprising to see that the 
upameyas propound the idea of buddhas seated within living beings. To describe 
the buddha-nature in terms of a tathagata within is to have recourse to just another 
metaphor, a kind of upamana. Apparently the authors of the siitra were trying, 
even after the presentation of the upamana, to express their idea in words that 
were easy to understand. They wanted their readers to have a very concrete notion 
of the buddha-nature, one easy to grasp and difficult to forget. Their choice of 
figuration sheds light on the general problem of teaching an abstract idea, such as 
the concept of the buddha-nature, to a broader audience of non-specialists in the 
field of Buddhist abhidharma. For such readers it would be hard to associate the 
ideas of buddhatva or of tathagata-dharmata with themselves. The veneration of 
buddha statues was at the time of the TGS, as the seventh simile demonstrates, not 
unknown to Buddhist followers, and may even have featured significantly in their 
religious practice. To have such a “statue” within oneself and to know it to be 
present in all other beings could only increase respect and appreciation towards 
others and towards oneself. That the buddha-nature of living beings should be 
expressed by the image of a buddha seated within seems to be a rather unique 
conception, one showing that the authors of the TGS were not interested in 
presenting a definition of the buddha-nature in philosophically abstract terms. The 
Lankavatarasitra™ and the verses of the Ratnagotravibhaga that freely render 

83 Also intimately bound up with the idea of a buddha within living beings may be the practice 
of buddha visualization (buddhanusmryti). Techniques of this kind are described in early Mahayana 
sources, most eminent among them the PraS, the “Sutra of the Concentration of [the bodhisattva] 
who stands face-to-face with the Buddhas of the present” (Harrison 1978b: ix), a text already 
translated into Chinese by the second century CE (see Harrison 1978a; 1990; 1978b). The question 
if and how the ideas associated with buddhanusmyti are related with the genesis of the 
tathagatagarbha teaching is complex and cannot be dealt with here. 

* See LAS 77.14ff. (though there tathdgatagarbha is not a bahuvrihi): ... sa (=tathagata- 
garbho) ... dvatrimSallaksanadharah sarvasattvadehantargato mahardhamilyaratnamalina- 
vastra'parivestitam iva ... [! read vastra for vastu]: “... that [tathagatagarbha,] bearing the 32 
marks [of a Great Being, and] existing inside the bodies of all living beings like a jewel of great 
value and price that is enwrapped in a dirty garment,....” The passage clearly resembles passages 
of the TGS; see also the description of Ch; in 0G: fA——7EH, BAAR, MKS, 


PDF Version: BPPB VI (2002) 

parts of the 7GS are the only possible other sources that betray knowledge of such 
a concrete conception. In other texts, more abstract terms became preeminent, 
while at the same time the term tathagatagarbha was interpreted in a less concrete 

Now, when living beings later become buddhas they turn into what they 
already carry within themselves in miniature. This is different from the idea that 
they have tathagatajfiana or other typical qualities of a buddha which, once 
purified, will allow them to act as buddhas. In this latter case, it is an attribute 
common to living beings that has to be submitted to a process of purification in 
order for those beings to become a buddha. The former case has to do with the 
purification of the buddha within from the sheaths surrounding him, he himself 
being equipped with tathagatajfiana and the other qualities (cp. OM) which will 
allow living beings to turn into buddhas; that is, he is not an attribute but that into 
which living beings will turn. The siitra does not specify exactly how this process 
is to be imagined: does the sattva merge into the actualized state of the purified 
tathagata within, in a process of identification, so to speak, with what is within, or 
does the liberated tathagata within simply enable the sattva to act with the mind of 
a tathagata, as, figuratively, a buddha now bearing a purified tathagata within? 
This question obviously did not interest the authors of the TGS, and so they do not 
offer us any basis for arriving at a conclusion. It is possible that they adopted the 
idea of buddhas within living beings because of its high degree of immediate 
visualization, without hesitating at, or even being aware of, the somewhat irritating 
consequences for a lucid philosophical interpretation of the upameya. 

The 7GS also speaks of buddhahood (buddhatva/°ta etc.) in living beings. 
In contrast to the substantialist idea of a buddha(-body) discussed just now, 
“buddhahood” designates a state said to be found in living beings covered by 
defilements. Here the authors resort to abstract terms, adding the abstract suffix 
-tva or -ta to the words buddha and tathagata. When the sttra declares that this 
state of being a tathagata is wrapped in sheaths of defilements, we are reminded of 
the concrete terminology of the upamana, as if buddhatva was an entity 
comparable to the buddhas seated in living beings. I assume that the authors, 
captivated by the imagery of the upamana, set store by the terminology of 
“covering” and “wrapping,” and thus decided to apply it metaphorically even to 
the term “buddhahood.” But whereas the figure of buddhas wrapped in the 
defilements of living beings was a fitting one, it is odd to describe buddhahood in 
such terms. At best, we can think of the single qualities of buddhahood or the 
tathagatajfiana (see below) as being veiled, but this, I think, was not the intention 
of the authors. 

What our authors seem to be getting at, rather, is the notion that 
buddhahood is somehow already present within living beings, but not yet 
efficacious. They try to state this belief in terms expressive of awakening, among 
them abstract ones such as buddhahood, svayambhutva (3B), prakrti (4C) and 
buddhabhumi (3C), not feeling it necessary—or, possibly, not knowing a good, 
readily understandable way—to express the relation between the unawakened and 
awakened state of being in less metaphorical terms. 

Among the elements said to be wrapped in defilements, we frequently 


PDF Version: BPPB VI (2002) 

encounter tathagatajfiana (or: buddha’ etc.) and tathagatajfianadarsana, Both are 
common terms and designate the specific knowledge which traditionally confers 
upon a tathagata his extraordinary status. In the simile of the *Tathagatotpatti- 
sambhavanirdesa (TUSN),* which may have been the prototype for the authors of 
the TGS, the main element is tathagatajnana. It is said to pervade all living beings. 
The destruction of wrong conceptions harbored by living beings would make them 
aware of the fact that they are penetrated by tathdgatajfdna and lead them to a 
state of equality with the tathagatas.®*° 

However, the simile in the * Tathdgatotpattisambhavanirdesa differs from 
the 7GS in one important point: tathagatajfdna in the * Tathdgatotpattisambhava- 
nirdesa is not primarily perceived of as a separate element found within living 
beings. The set of similes in the *Tathagatotpattisambhavanirdesa of which the 
one discussed here is part is introduced with an analogy between ether and 
tathagatajnana (TUSN S 147a4—b4), namely that tathagatajfana, just like ether, is 
the resting place of all things but itself does not rest anywhere. Though not entirely 
clear, the following simile in the *Tathagatotpattisambhavanirdesa presents 
tathagatajfiana as an all-pervasive element which, when living beings become 
aware of its presence and when the fetters of the mind are removed, constitutes 

The descriptions in the TGS, on the other hand, suggest that tathagatajfiana 
is a separate element abiding steadily within living beings. If so, it would be 
possible to apply to it the metaphor of the sheaths of defilements. Of course, one 
could also argue that tathdgatajridana had to take this more individualized form in 
the TGS because the specific metaphor in the latter did not allow room for the idea 
of a single tathdgatajfana that pervades all living beings’ minds like ether, as 
seems to be the case in the *Tathdgatotpattisambhavanirdesa. In the Ratnagotra- 
vibhagavyakhya, the significance of the terms tathagatajfiana and buddhajnana is 
all but negligible. Tathagatajnana appears only once besides its occurrence in the 
quoted simile of the *Tathagatotpattisambhavanirdesa. In the crucial verse 1.27, 
where it is explained how living beings can be said to have the buddha-nature, 
buddhajfiana is mentioned as pervading them. In the following verse and in the 
prose commentary, however, it is replaced with the term dharmakaya. By contrast, 
both tathdgatajidna and tathdgatajnanadarsana are well-attested terms in the 

Another particular expression used to designate the buddha-nature of living 
beings is (tathdgata-)dharmata. This term has undergone considerable 
development in the history of Buddhism and can have at least three main 

(1) Dharmata as “natural and normal custom,” “habit,” “normal state” etc. 
Closely associated with these ideas, it can stand for the nature of dharmas in 

% Cited in RGVV 22.10-24.8. I have partly translated the simile at the beginning of section A 

% See RGVV 24.3-6. 

87 The upameya in the simile of the TUSN (see 3.3), it is true, may be meant to evoke another 
idea when it speaks of myriads of pieces of cloth with the whole universe painted on them. Here, 
tathagatajnana would naturally be imagined as a separate element in each living being. 

88 See BHSD s.v. dharmata; also Rahula 1974: 184; 187. 


PDF Version: BPPB VI (2002) 

the sense of the rule to which all dharmas are subject, primarily the law of 
pratityasamutpada. Both meanings are attested in the Pali Canon.” 

(2) Dharmata can simply mean the “character, nature or essence” of somebody 
or something.” 

(3) In a later development, restricted to Mahayana Buddhism, dharmatad came 
also to designate the true essence of all dharmas, here understood as 
metaphysical reality more or less synonymous with tathatd, which is said, like 
ether, to penetrate all phenomena. Dharmata represents here the absolute 

We encounter dharmata in the first meaning in the TGS in the well-known 
esa kulaputra dharmanam dharmata / utpadad va tathagatanam anutpadad 
va... (1B; cited in RGVV 73.11—12; cp. the note in my translation to 1B) 

The 7GS, however, does not introduce the law of pratityasamutpdda with this 
formula, as is usually done, but its own central message: sadaivaite sattvas 
tathagatagarbhah when referring to the buddha-nature of living beings. To speak 
of the dharmata of all dharmas in the sense of “things” does not make sense 
here.”” The meaning of dharma probably intended in this passage of the TGS is 
“teaching,” and consequently the dharmata should be understood as the essential 
law underlying all teachings (dharma), namely the fundamental thought that all 
living beings harbor a tathagata within. In the Chinese translation by 
Buddhabhadra, this is the only passage where the term dharmata is rendered with 
characters generally used for the translation of dharmata: 4&#. Buddhabhadra 
was obviously well aware of the different connotations of this important term and 
preferred in all other instances to render it differently. 

Dharmata in the meaning of “essential law” also seems to fit verse 7.5, 
where, according to my understanding, the dharmata of sentient beings is 
formulated as: “Here [within each sentient being] always dwells a victorious one, 
wrapped around [with defilements].” Buddhabhadra’s translation here does not 
mention anything like dharmata. 

The 7GS also uses dharmata in a different meaning, which, judging from 
the number of its appearances, is the prevailing one in the TGS. It is part of the 
upameya in the similes. It is said to be found within living beings, where it is 
wrapped with defilements. Whereas in the prose passages it appears regularly as 
tathagata-dharmata, in the verses it is simply spoken of as dharmazta. It is not 
possible here to view dharmata as the rule all dharmas necessarily follow. 
Restricted to the prose, dharmata could well be the one discussed under (2) above, 
in the sense of “essence/nature of the tathagatas” (tathagata-dharmata).”* 

% See Schmithausen 1969a: 105ff. (parallel to the term tathata); 116ff.; further see PTSD s.v. 
dhammata; Rahula 1974. 

» Tn this sense the term appears, for instance, in RGVV 10.4: dharmadhatum iti svadharmaté- 
prakrtinirvisistam tathdgatagarbham / (emendation °sisfa- to °Sistam (ms B) in Schmithausen 
1971: 135f.): “Dharmadhatu is the buddha embryo, which by its intrinsic nature is not different 
from their (= the tathagatas’) nature (dharmata).” 

*" See Schmithausen 1969a: 105f£.; 116ff. 

* Tt is possible that the famous formula was simply taken over without reflecting on its 
adequacy regarding the immediately following “law” of tathagatagarbha. 

3 The compound tathagata-dharmata could, however, also be understood as a karmadhdraya 
meaning “true nature in the form offlike a tathagata,” thus referring to the simile of the buddhas 
seated within living beings. 


PDF Version: BPPB VI (2002) 

There is no need to argue that the idea conveyed by dharmata under (3) 
was of any relevance for the authors of the TGS. The fact that dharmata is 
described as being encased in defilements does not mean that it was thought of as 
an actual entity. In the case of buddhatva, we encounter a similar figurative way of 
expressing its relation to living beings. On the other hand, as the verses 
demonstrate,” the term dharmata alone, even without the specification that it is 
the dharmata of the tathagatas, must contain a clearly positive nuance. In these 
cases it cannot simply be understood as “essence, nature” but must have a 
weightier meaning, such as “true nature.” Alternatively, we could also assume that 
the first part of the compound tathagata-dharmata was simply dropped due to 
metrical needs, as is often the case in verses. If we follow this assumption, namely 
that dharmata in the verses is to be understood as tathagata-dharmata, there will 
be no further need to argue that the meanings of dharmata in the TGS go beyond 
those mentioned in (2). 

If, on the other hand, we decide that the use of dharmata in the verses does 
not allow us not to attribute a positive character to it, we shall have to admit that it 
must mean “more” than the alternatives mentioned under (1) and (2). To identify 
this dharmata with the dharmata of (3), namely the one, all-pervading absolute 
truth, would nevertheless not be in line with the sitra itself. The contexts in which 
the term dharmata appears for the most part involve living beings in the plural. 
The purport is clearly the fact that all living beings have their dharmata rather than 
that a single dharmata pervades them. The philosophical background can thus 
hardly be called monistic; if there are any monistic overtones at all, they are not 
intended by the authors. Accordingly, I think that the semantic range of dharmata 
in the TGS occupies a position somewhere between (2) and (3) and is (not yet) 
thought of as a hypostatized unit. 

If we turn to the two sitras which, I believe, are most closely related to the 
TGS, namely the Saddharmapundarikasitra and the *Tathagatotpattisambhava- 
nirdesa, it becomes clear that the term dharmata in the meaning of “absolute 
truth” was likely not used in these texts.” Further, in the simile of the painted 

% See 8.4 and 8.5. In 3.4 the dharmatd is characterized as “like mine.” The use of dharmata 
there fits well into the semantic range of (2). 

% Throughout the SP the term dharmata appears only five times, among them once in the prose 
text (40.15), where it plainly means the “true universal law” in the second sense of (1) (see 
Zimmermann 1999: 157ff.). In K 57.13 (© 2.134), K 130.13 (= 5.38), and K 392.2 (= 20.1) 
dharmata stands for this true law as the most fundamental teaching, and is in some cases rendered 
as such in the Tibetan and the Chinese translation by Kumarajiva: K 57.13: #2 /(#x\; K 
130.13: chos, (982.4; K 392.2: chos. In K 294.14 (= 13.67) we find the pada bhavitva dharmam 
ca spysitva dharmatam. This part is missing in D; and D2; O and the Farhad-Bég ms read 
(bo)dhi(m) instead of dharmatam. The Tibetan follows, as usual, K (... chos nyid reg). 
Kumarajiva’s translation has 2439459, supporting the wording dharmata, whereas 
Dharmaraksa’s translation offers ]-#@ (bodhi)—with the exception of two relatively late, and 
probably (in view of Kumirajiva’s translation of this passage) “emended” manuscripts which read 
jE-#4 instead. This is the only passage in the SP where dharmatd can but need not (given that it is 
the older reading) be interpreted in the sense of “absolute truth.” However, the fact that the 
Kashgar manuscripts usually contain the older wording suggests that the reading dharmata is 
secondary (though it would be a /ectio difficilior!). We are thus left with one very meager example 
of dharmata in the meaning of an absolute entity. 

In the TUSN, as far as I can see, the term dharmata appears several times in the meaning “natural 
and normal custom,” “normal state” (1): Q 81a6 and 81b7: chos nyid rab tu thob pa’i phyir 
(*dharmata-pratilambhat: “in a normal, natural way,” “naturally, automatically,” see also BHSD 
s.v. dharmata); Q 100a8: chos nyid thob pas (*dharmata-praptena: “naturally, automatically”); Q 


PDF Version: BPPB VI (2002) 

cloth in the *7athdgatotpattisambhavanirdesa, the direct predecessor of the trope 
presented in the TGS, the term dharmata does not appear. Instead, we are told only 
of tathagatajfiana. If dharmata in the passages of the TGS designated in fact the 
hypostatized “absolute truth,” this would mean that, when it came to the usage of 
this basic philosophical term, the authors of the TGS did not agree with the 
Saddharmapundarikasitra and the *Tathagatotpattisambhavanirdesa but based 
themselves on a different terminological convention. 

Be that as it may, the mention of the (tathdgata-)dharmata serves to 
illustrate the buddha-nature of living beings and their essential equality with the 
tathagatas.”° Here also, as has already been observed in the case of tathdgatajfana, 
dharmata is described in a metaphorical setting: it is encased by defilements, and 
in 0M even said to be motionless. All living beings have this true nature. It 
constitutes their essence, thus relegating all differentiating aspects of living beings 
to a provisional, contingent level. 

Among other terms that render the idea of the buddha-nature in the TGS, 
we find the word tathadgatadhatu.”’ The term dhdatu is well known from the 
Mahayanist Mahdaparinirvanasutra, where it appears frequently in the compound 
buddhadhatu, synonymous with tathdgatagarbha, and obviously interpreted as a 
tatpurusa compound. Dhdtu can have the very concrete meanings “mine,” 
“primitive matter” (called also mahabhiita), “element of the body” (the six 
elements are prthivi, dpas, tejas, vayu, akasa, vijfiana), the “group of eighteen 
dhatus” forming the six sense organs with their respective sense objects and the 

140b1: ... chos nyid rab tu thob pa dang ... (*dharmatapratilambha- “[Bodhisattvas gathered]..., 
automatically,....”) etc.; dharmatd as “rule, teaching” (1) in Q 138a5f.: byang chub sems dpa’ 
thams cad kyi chos nyid rtogs par khong du chud pa (see BHSD s.v. gatimgata; *sarva- 
bodhisattva-dharmata-gatim-gata: “adept in all the rules for/teachings of the bodhisattvas”), in QO 
14lal; 7f.: de bzhin gshegs pa’i chos nyid mi ’phrogs (pa) (... yongs su) brjod and Q 141a5f.: de 
bzhin gshegs pa’i mnga’ mi ‘phrogs pa’i chos nyid ... brjod (*asamharya tathagatadharmata: “[to 
utter] the irresistible teaching(s) of the tathagatas”); probably in Q 138al: ... byang chub sems dpa’ 
de ni/ byang chub sems dpa’i chos nyid la yang dag par byung ba ma yin no // (see BHSD s.v. 
samudagacchati, ‘te and following entries; *bodhisattva-dharmatayam asamudagatah(?): “That 
bodhisattva (who does not know about or believe in the magnificence of the tathagata) has not 
accomplished the teaching for the bodhisattvas’(?); TUSNon: 2SAR RLS (630a1 6f.); 
TUSN ens: ASL ASAE EE (277c18f.)), in Q 14143: de bzhin gshegs pa’i chos nyid rab tu thob' 
pas [! Q: thos corrected to thob according to S (186a5) and TUSNer: 4 R0AGERL (278¢4-5)] and 
probably in Q 139b7f.: chos nyid rjes su bsnyegs pas bstan pa (“(you) explained according to 
(*samanubandhena) the teaching”; TUSNcx2: RAN YE (630011); TUSNcx3: BANAYEME (278b3)), 
though in the last case one could also approve the meaning “true, universal law” (1) or “absolute 
truth” (3); as “character,” “nature,” “disposition” (2) in Q 89b8-90al (verse); 90a2 (verse): chos 
nyid tshul (*dharmata-yogena: “(Rain water/The teaching of the Buddha becomes differentiated) 
according to the disposition (of the surface of the earth/of the disciples)”; unclear is the context 
and thus the meaning of Q 129b3: chos nyid khong du chud pa’i tshul gyis (*dharmatadhigama- 
yogena: “according to the realization of the teaching/universal law/absolute truth(?)”; TUSNcx2: 
BAER (627012); TUSNen3: SARE REM (275c23)) and O 138a5: byang chub sems 
dpa’ ... byang chub sems dpa’i dbang rang ‘byung gi chos nyid la gnas (*bodhisattva-vaSa- 
svayambhi-dharmata; the Chinese reads instead: TUSNom: FARRER (630a22f.); 
TUSN cus: FMB E (277024)). 

% See OM: “And [the Tathagata], having perceived inside those [sentient beings] defiled by all 
defilements the tathagatadharmata motionless and unaffected by any of the states of existence, 
then says: ‘Those tathagatas are just like me!’”” 

*’ The term tathagatadhatu appears once, in 8B (for details see my note in the translation). It is 
rendered in 7ib with the term rigs, which usually is applied as a translation for gotra or kula. On 
the term dhatu see also Seyfort Ruegg 1969: 261 ff. 


PDF Version: BPPB VI (2002) 

six forms of vijfidna, “relics” or (in medical theory) “essential bodily 
ingredient/humor.” Especially the two meanings “relics” and “bodily element” 
seem to have had some influence on the genesis of the buddhadhatu theory of the 
Mahaparinirvanasitra.?* However, the main thrust of the term buddhadhatu in the 
Mahaparinirvanasutra is to indicate the future possibility of all living beings 
becoming a buddha.” Though this understanding of dhdtu as a cause stands in 
contrast to the above-mentioned substantialist background from which the term 
dhatu is believed to have emerged in connection with the Mahdparinirvanasitra, 
it is not altogether surprising, since it can be an equivalent of hetu, as stated 
explicitly in, for example, Ratnagotravibhagavyakhya 72.10: tat(=tathdgatatva)- 
praptaye hetus tathagatadhatur iti / hetvartho ’tra dhatvarthah /.'° We further 
know that dhdtu can, like the term dharmata, also mean the “rule to which all 
dharmas are subject” and, again parallel to the development of the term dharmata, 
can even take on the meaning of a metaphysical entity, in the compound 

How are we to understand the term dhdtu in the passage in 8B? Let us first 
have a look at the quotation of the passage in the Ratnagotravibhagavyakhya and 
the context in which it is embedded there. The quotation appears after the last of 
three basic interpretations of the compound tathdgatagarbha. | cite the whole 
passage (72.7ff): 

paficabhir nidhitaruratnavigrahacakravartikanakabimbadystantais trividhabuddha- 
kayotpattigotrasadbhavartham' adhikrtya tathagatadhdtur esam garbhah 
sarvasattvanam iti paridipitam /  trividhabuddhakayaprabhavitatvam hi 
tathagatatvam / atas tatpraptaye hetus tathagatadhatur iti / hetvartho ‘tra dhatv- 
arthah | yata aha / tatra ca sattve sattve tathagatadhatur utpanno garbhagatah 
samvidyate na ca te sattva budhyanta iti / 
{' for the emendation °sadbhavartham for °svabhavartham see Schmithausen 1971: 157] 
By way of the five similes of the treasure, the tree, the precious image, the 
universal emperor and the golden figures—with reference to the fact that the 
disposition (gotra) of the threefold body of the buddhas to manifest is present [in 
all living beings]—it has been shown that the tathagata element (dhatu) is the 
embryo/germ (garbha) of all these sentient beings. [This can be said] because 
tathagatahood is constituted by the threefold body of the buddhas. Therefore the 
cause of the attainment of that [threefold body] is called the “element (dhatu) of 
a tathagata.” “Element” means here “cause” (Hetu). Therefore it is said [in the 
TGS]: “And though in each sentient being the element (dhatu) of a tathagata has 
arisen and is present as an embryo/germ (garbhagata), those sentient beings do 
not realize it.” 

The general context here is an explanation of the three svabhavas of the dhatu, 
namely dharmakaya, tathata and gotra, the three key terms of verses I.27-28. 
These verses constitute the heart of the Ratnagotravibhaga(vyakhya), being a 
commentary on the central phrase sarvasattvas tathdgatagarbhah. The passage 
translated above deals with the third aspect, namely the gotra, and contains the last 
of the three interpretations of the compound tathagatagarbha. Garbha is here 
identified with dhatu, which is immediately afterwards defined as the “cause” 

%8 See Seyfort Ruegg 1969: 516, n. 1; 1976: 352f.; Shimoda 1997: 272; 276; Habata 1989; Hob 
s.v. Byo. 

® See Shimoda 1997: 272. 

100 Yor further references for dhdtu in the sense of “cause” see Schmithausen 1969a: 114c; 145. 

1! See Schmithausen 1969a: 147. 


PDF Version: BPPB VI (2002) 

(hetu) of the attainment of buddhahood. Consequently garbha can here only have 
the meaning “embryo” or “germ,” which is the cause of living beings becoming 
tathagatas. In order to reinforce its analysis, the Ratnagotravibhdgavyakhya then 
cites the sentence from the TGS. The fact that this passage was cited by the authors 
of the Ratnagotravibhagavyakhya is probably not only due to the unique 
appearance of the term dhatu throughout the TGS, but also to the predicate 
garbhagata, which according to the Ratnagotravibhagavyakhya characterizes the 
dhatu “as an embryo” (see also RGVV,;: snying por gyur pa; the Chinese translation 
is here missing). 

The interpretation of garbhagata as indicating a state or condition is 
grammatically possible. Further, in view of the fact that the five similes this 
explanation in the Ratnagotravibhagavyakhya refers to include the sprout in the 
seed and the embryonic emperor in the womb of the poor woman, the 
understanding of garbha as a germ or embryo is in fact plausible. However, the 
context in the TGS imposes a different conclusion as to how garbhagata should be 
understood. Both Tibetan translations of the TGS support the notion that garbha 
means “womb” or “inside” of living beings: khong na yod (Tib); snying po la gnas 
pa (Bth).'* This analysis of garbhagata as indicating a local relation (“within 
(living beings)” or “in the womb”) is likewise possible from the point of view of 
Sanskrit grammar and, for the unbiased reader, even much more reasonable.!” But 
even leaving aside the two Tibetan translations that favor garbha as “inside,” there 
are three other arguments strongly supporting this understanding in contrast to the 

One argument is found in the Ratnagotravibhaga itself, in the verses that 
freely render the simile in question: though it is true that verse 1.122 compares the 
dhatu with the embryo of an emperor, in the following verse 1.123 we find two 
concessive formulations: ... garbhantarasthe nype (‘“... though the king is found 
within the womb [of sentient beings]”); ... sannathesu ca satsu 
svatmantarasthesv api (“... even though the good protectors are found within 
[living beings] themselves.”). In both cases the main accent lies on the fact that the 
emperor or “protectors” are present within sentient beings, not on the fact that they 
are still embryos. The presence of the word antara does not allow for any 
ambiguity. This nuance was duly expressed by the Tibetan translators (Tib), in a 
word-play with khong (“inside”): ... khong na yod kyang sems can de dag gis 
khong du ma chud do //. 

Right after the above sentence in question, the Tathagata repeats the same 
message a second time in direct speech to living beings (8B.7-8). Nothing is said 
on this occasion about an embryonic state. Instead, mention is made of a tathagata 
inside (the bodies) of sentient beings: nang na (Bth); BH (Ch); B-& (Chy); Tib 
has simply khyed la. 

Finally, we have the evidence displayed by Hara (1994: 38, n. 4), who, if 
mainly with regard to the epics, states that garbha bears the meaning “womb” 

102 Ch, takes garbha of garbhagata to be the buddha-store and not the womb: ELAN. Ch, 
offers no help here. 

13 The two interpretations perhaps do not lie very far apart, since speaking of the dhatu “in the 
state of an embryo” implies that it is present “in the womb” or “inside.” Nevertheless, if the dhatu 
is being described as just being “inside” living beings (a “neutral” formulation which the use of the 
term “embryo” would contravene), this does not mean that it has to be a hetu, as the RGVV would 
have it. 


PDF Version: BPPB VI (2002) 

“when construed with words expressive of staying in, dwelling at,” and cites 
examples for garbhastha, °gata, °vasa etc. 

There is little prospect of settling the question definitely. However, arguing 
from the perspective of the 7GS itself and the undoubtedly very old citations of the 
TGS in the verses of the Ratnagotravibhaga, we ought clearly to favor the latter 
alternative, though the (more recent) commentarial parts of the Ratnagotra- 
vibhagavyakhya and the general biological overtone of the cakravartin simile 
allow for a different opinion. 

Now, what does this all mean for an understanding of the term dhdtu in the 
TGS, the question that served us as the springboard for the above discussion? Once 
the reading of dhdatu as hetu in the Ratnagotravibhagavyakhya is seen to have no 
validity for the passage in the TGS, we are free to speculate on the meaning of 
dhatu in the contextual framework of the TGS. We should therefore note what 
other terminology in the upameya of the present simile is used to designate the 
buddha-nature. In the direct speech of the Tathagata in the second part of 8B, we 
find the term tathagata; in the verses below, dharmata twice appears. Neither of 
these words suggests the notion of “embryo.” 

On the other hand, the upamana of the same simile several times speaks of 
sattva (“life”) which enters the womb of the woman. The parallel to the passage 
under discussion is obvious: here also the tathagatadhatu enters into living beings. 
The direct speech of the Buddha that follows is similar in structure. The Buddha 
speaks of “the tathagata [who has] entered [and] is present within...” It thus seems 
that the upameya side of the simile has adopted the terminology of the upamana— 
a phenomenon similarly found in the descriptions of the tathagatas as encased in 
klesa sheaths in the upameya, In any case, given the biological motifs throughout 
the simile, should we not assume that the single employment of the term dhatu in 
the whole sitra is integrally related to this context? I think it likely that the author 
of the simile thought of the tathdgatadhdtu as a constituent element within the 
body of living beings, just as sativa is in the womb of the poor woman. He could 
have had in mind the six elements of the body (see above), with the vijfianadhatu 
as the most fundamental (in the sense of being constitutive of life) and central one 
(as underlying all other sense organs).'”’ In analogy to these six elements, and to 
the vijfianadhatu in particular, the tathagatadhatu would be understood as another, 
even more essential (metaphysical) element of living beings, constituting and 
guaranteeing their inherent buddhahood.'®° Designating as it does an independent 
element of the body, the term dhatu would fit well into the biological framework 
of the simile. At the same time, tathagatadhdatu recalls the term nirvanadhdatu, the 
sphere a person enters after liberation. The TGS, by replacing the term nirvana- 
with tathagata-, would continue the work of the Saddharmapundarikasutra of 
pointing out the insufficiency of nirvana as a goal and advancing the only 
destination really worth striving for in Mahayana, namely to become a tathagata. 

14 This argument loses some of its force when, with an eye on the occurrence of the term 
utpanna before the compound garbhagata, we read Hara’s following remark that the meaning 
“embryo” is prevalent “when construed with words of production, conception, development, and 

105 On vijfiana as the (cosmic) element see Langer 2001: 43ff. 

16 See also the terms citi-dhatu and dtma-dhatu in the terminology of the Vedanta. These are 
employed to designate the true self of living beings. Dhatu, in these cases, clearly undergoes an 
extension of metaphysical meaning (see Schmithausen 1969a: 83). 


PDF Version: BPPB VI (2002) 

We have seen that the buddha-nature of living beings is described with a great 
variety of terms in the TGS. The terminology includes markedly substantialist 
expressions, such as buddha and *jinakaya, and even terms which figure in the 
upamana—words like nidhi or kosa. Tathagatajhana, a term with strong 
ontological connotations, appears along with actual states of being, such as 
buddhatva and svayambhitva, With dharmata we reach the other end of the scale, 
a term similar to tathatd in being bare of any concrete nuances, and in designating 
the nature of a tathagata or, as some passages could imply, even absolute reality. 
Though the terms denoting the buddha-nature differ considerably in their degree of 
abstractness or substantialist connotations, the authors of the siitra did not hesitate 
to use them all in the standardized metaphorical phrase “wrapped with sheaths of 
klesas.” One characteristic of the TGS seems in fact to be this tendency to 
formulate the upameya with the help of vocabulary taken from the upamadna. The 
negligence in the choice of the basic terminology for designating sentient beings’ 
buddha-nature, on the other hand, shows, as will be demonstrated below, that the 
authors were more concermed to present this new idea effectively than to lay it out 
in clear philosophical terms. 

All the terms used to designate the buddha-nature of living beings can be 
characterized as being associated with the level of perfect spiritual realization.'”” 
The notion of potentiality, implied in terms like “germ” (gotra) or “embryo” 
(garbha), is inherent to none of them. This is even more surprising in view of the 
fact that in two of the similes the counterpart of the buddha-nature in the upamana 
seems to have causal force: the sprout and the cakravartin embryo. Though in 
these two illustrations they could have easily done so, obviously the authors were 
not interested in adding to their corpus of terms for the buddha-nature ones which 
carry the notion of potentiality. It appears to me that what the authors had in mind 
was, first of all, to stress the essential equality between living beings and the 
buddhas. They therefore had to choose a terminology which left no room for doubt 
about this basic fact. For them to speak in this context of an embryonic (tathagata-) 
garbha or (-)gotra of living beings would not have served their ends. 

If we compare all this with the conception of the buddha-nature in the 
simile of the *Tathdgatotpattisambhavanirdesa, we can observe another 
characteristic of the theory in the TGS. The main aim of the *Tathagatotpatti- 
sambhavanirdesa is to describe the wonderful and incomparable nature of the 
Buddha’s qualities. The greatness of tathdgatajnana, along with nine other 
characteristics, lies in the fact that it penetrates (pravisati) the mind of all living 
beings. It is in this context that the simile of the piece of cloth with the whole 
cosmos painted on it is expounded. The goal the TGS has set for itself, on the other 
hand, is not to describe and extol the limitlessness and greatness of 
tathagatajfana. The TGS seems to focus on the living being as the “container” or 
“owner” of tathagatajnana. Living beings are not simply perceived as the objects 
of the jriana of the Tathagata; they step into the foreground in their own right. 
Every single one of them is in possession of tathagatajnana. Seen from this 
perspective, tathagatajfana appears rather as an element segmented so as to be 
present in all living beings. The idea of a segmented tathagatajnana may not have 
been unknown to the author of the *Tathdgatotpattisambhavanirdesa. This is 

1°7 This is not so if we take tathagatadhatu in 8B as a cause, as suggested by the RGVV (see 


PDF Version: BPPB VI (2002) 

documented in the upamana of the simile where myriads of painted cloths appear 
as independent elements by analogy with tathagatajriana. Thus I do not wish to 
exclude the possibility that the philosophy of the *Zathdgatotpattisambhava- 
nirdesa in fact propounds the idea of a multitude of fragmented buddha- 
knowledge which, while embedded in different living beings, penetrates each part 
the other. Nevertheless, the *Tathagatotpattisambhavanirdesa seems rather to 
assume that the cosmos is penetrated by an all-pervasive ether-like jfidna in which 
living beings are embedded, whereas in the TGS the picture is clearly inverted; that 
is, jidna is contained in the minds of living beings. This latter perspective allows 
the bulk of attention to be directed to living beings as autonomous subjects. This 
difference in perspective is also mirrored in the core statements of the TGS and the 
corresponding passage in the *Tathdgatotpattisambhavanirdesa. In the TGS the 
agents are living beings containing a tathagata (sadaivaite sattvas tathdgata- 
garbhah), whereas in the * Tathagatotpattisambhavanirdesa the agent is tathagata- 
jnana in its diffusive power (tathagatajfanam ... sakalam anupravistam; quoted in 
RGVV 22.10—-11). Without this change in perspective the employment of terms for 
the buddha-nature with substantialist overtones would not have been possible in 
the 7GS. 

3.2 Becoming a Buddha 

The 7GS states that all sentient beings contain a tathagata and are essentially equal 
to the tathdgatas. The siitra, however, does not say that they themselves are 
therefore awakened and already tathagatas. The reason why they cannot be said to 
have attained awakening is the fact that their essence is not yet purified from 
klesas, which, in the metaphorical expression of the TGS, enwrap and pollute this 
essence. Only when these defilements (kleSa) have been removed will the essence 
of sentient beings become manifest to and efficacious for them. Thus the basic 
idea of the TGS is that sentient beings always partake of the nature of tathagatas. 
This nature cannot essentially be affected by the defilements, because it is 
unchangeable. The klesas, on the other hand, are said to be only dgantuka (4.3), 
that is, accidental in nature. They are removed in the process of purification so that 
finally the buddha-like nature will be revealed. The idea is best put into words in 
8B, where the Buddha in a direct speech states that “... it will happen that one day 
the tathagata [who has] entered [and] is present within you will become manifest” 
(*tathagatah pravistah samvidyamanah ... pradurbhavisyati).'© It is thus clear 
that becoming a buddha does not involve an essential transformation or generative 
causal act by which the gap between the state of being hindered by mental 
defilements and that of perfect awakening is bridged. 

However, two of the upamdnas appear to attest to a different view: the 
simile of the sprout in the seed and that of the future emperor in the womb of the 
poor woman. The sixth simile speaks of the sprout in a seed of a fruit. The seed is 
planted in the earth and becomes a big tree. Verse 6.4 explicitly compares the fact 
that the tree has grown from a seed to the awakening of sentient beings. Now, the 
sprout has to undergo different stages of growth and development before finally 

108 Tib: khyed la de bzhin gshegs pa zhugs pa yod pa dus shig na ‘byung bar ‘gyur te /. 


PDF Version: BPPB VI (2002) 

turning into a big tree. From this one might expect that the buddha-nature of living 
beings is first something comparable to a sprout which gradually, by the 
transformational process of ripening, develops into an entity possessing 
buddhahood. Such an immature buddha-nature would thus be the cause of the 
attainment of awakening, just as the sprout in the seed generates the tree. But 
while this interpretation, especially in light of the later understanding of the term 
tathagatagarbha as “buddha-embryo” or “buddha-germ,” could, prima facie, be 
felt to be what the authors of the 7GS wanted to suggest, a closer look at the simile 
reveals that this is not the case. 

What the authors have focused on in this simile is clearly found in verse 
6.2, which states that the body of a sugata exists similarly to a cane seed in all 
living beings. This verse is revealing, since the authors could have chosen to use a 
term other than *sugatakaya (such as, e.g., °gotra or °garbha). In other words, they 
could have expressed that the buddha-nature of sentient beings is still in need of 
further development.'” Instead, they continue to employ the same terminology as 
in the similes before. This terminology, chosen to describe the process which 
makes the buddha-nature of sentient beings efficacious for the world, does not 
connote any kind of growing or ripening. Rather, as in the earlier similes, it is 
expressive of the purification (6B, 6.4) from the sheaths of defilements—a fact 
which may indicate that the authors had in mind not mainly the growing process of 
a sprout but its being enclosed in seed coverings and the outer fruit from which it 
must be separated. 

Though the growing process of the seed is, of course, one important aspect 
of the simile, the main focus would seem to lie elsewhere. As in 6A and 6.1, the 
simile here affirms the imperishability of the seed, and the fact that the result 
(karya), namely the tree, is already contained in the seed. It emphasizes that seed 
and tree are of the same nature. The seed, as Ch has it, “realizes the quality of 
indestructibility as a result of the fact that the seed and sprout generate each other 
alternately....” The assertion by the authors of the TGS that the seed already bears 
the tree within itself certainly accords well with their “theory of revelation.” The 
figure offers correspondences to the full-fledged tathagatas in living beings. The 
manifestation of these tathagatas triggers the awakening of living beings. 
However, there is an important difference between this “tree” in the seed and the 
full-grown tree in terms of temporal development, even if it does not seem to be of 
relevance for the authors. What is decisive for them is the essential oneness of 
seed and tree. Nothing new has to be added; the complete tree is already found in 
the seed.'!° 

In the simile of the poor woman the situation is similar. Though the 
embryo of a cakravartin in her womb could easily be compared to the germ or 
embryo of a buddha in living beings, the author prefers to employ tathagatadhatu 
(8B), tathdgata (8B) and dharmata (8.4, 8.5). Here again, the fact that the 

109 Tt is true that in 6B dharmata is spoken of as garbhagatd or garbhasthd. | have argued in the 
note to my translation thet these latter should be understood as “in the womb.” However, it could 
also be understood as “in the state of an embryo,” and we cannot, as a matter of fact, exclude 
outright the possibility that the authors used this ambiguous formulation to allow for both 
interpretations (see also above, pp. 58ff.). 

"0 For orthodox Buddhist thinkers thée-idea of the sameness of nature between seed, sprout and 
tree, and the existence of the tree in the-seed, (satkdrya), would hardly hold water. Cp., for 

»dnstance, the Buddhist Salistambasiitra, where such ideas are strictly rejected as harboring the fault 
f of eternalism (Schoening 1995a; 285ff.). 


PDF Version: BPPB VI (2002) 

cakravartin is still an embryo does not seem to be crucial for an understanding of 
the simile. His nature, namely that of a cakravartin, will not change, his future rule 
is already preprogrammed, and the growth process the embryo is subject to does 
not figure in at all.''’ What the simile plays up, rather, is the striking contrast 
between the poor, ugly and pessimistic woman and the glory of the cakravartin 
inside herself. Still, this simile can easily give rise to a different view of the 
buddha-nature. Without the exact wording in mind, one is naturally tempted to 
compare the growth of the embryo to a continuous ripening process in sentient 
beings of their buddha-nature. To be sure, this would not be in line with the 
original intention of the authors, who favored a theory of manifestation. 

Now, the idea that a full-fledged tathagata is from the very beginning 
present in living beings raises serious problems for Buddhist philosophers. The 
TGS states that the tathagatas inside are covered by defilements and that only 
through a process of purification can living beings become awakened. 
Nevertheless, sentient beings are said to be always essentially the same as the 
tathagatas, and this immanence of the Absolute can easily lead one to downplay or 
even brush over the fact that between the state of being defiled and that of perfect 
awakening is a gap which has to be bridged. As seen above, the authors of the TGS 
do not seem overly cautious in this matter. They do not hesitate to attribute an 
obviously substantialist notion to the buddha-nature of living beings. More than 
the assumption of two qualitatively different stages of development, for them it is 
the gap between these two states that underlies the efficaciousness of sentient 
beings’ buddha-nature, which can be realized only through a process of 
revelation. '!” They therefore did not expend much effort in choosing similes with a 
unified conception of the buddha-nature. This may also be resulting from the 
authors’ very intention, which, as we shall see below, seems primarily to have 
been to present reasons why all living beings can become buddhas and, maybe 
even more importantly, to stimulate their readership, which had possibly hitherto 
not felt vitally drawn to Buddhist spirituality, to join the Mahayana community. 
The use of a philosophically and spiritually unsophisticated, imprecise and even 
ambiguous terminology, along with the substantialist and metaphorical 
formulation of the upameya, may have been considered necessary in order to attain 
their ends. It would therefore be wrong to expect elaborate explanations about the 
exact relation between the state of defilement and awakening. This painstaking 
work, whether intentionally or not, was left to later exegetes. 

Regarding the attainment of awakening, the authors of the TGS do not tire 
of emphasizing that this leads to the performance of the tasks of a buddha. They 
obviously consider this fact as an automatic consequence of the manifestation of 

‘Il It is, however, probably no coincidence that the only passages where the terms garbhagata 
or garbhastha appear are in the similes of the seed and the poor woman. Although the translations 
of the TGS itself and the parallels in the quotation of the similes in the RGV interpret garbha® as 
“inside,” the other possibility, i.e. “embryo,” cannot be ruled out. 

"2 This becomes clear in the similes of the seed which already contains the whole tree, the 
embryo of the cakravartin in the womb of the poor woman, and in the simile of the TUSN where a 
cloth on which the whole universe is painted is found compressed to a small particle and has to be 
opened in order to achieve its effect. Thus, the counterpart in the upamana to living beings’ 
buddha-nature is not yet in a fully developed state; for it to manifest fully requires a further stage 
of ripening or expansion. But given that the siitra stresses the essential sameness of the element in 
sentient beings before its revelation (when it is not yet efficacious) and afterwards (when it 
becomes so), this does not undermine the. basic concept. We should not expect or claim an 
absolute isomorphic relationship between the upamdna and the upameya for all similes. 


PDF Version: BPPB VI (2002) 

one’s buddha-nature, and in several passages it alone is stated to be a characteristic 
of buddhahood. This in itself demonstrates that efficaciousness was a main 
category in the earliest stage of tathdgatagarbha thought. The reason for 
describing a tathagata primarily in terms of dynamic activity may well lie in an 
attitude of worldly engagement predominating over mainly theoretical concerns. 

3.3 How to Become a Buddha 

The purification from the klesas leads to the manifestation of buddhahood. Now, 
what are the factors leading to this purification? Let us first note in what terms the 
*Tathagatotpattisambhavanirdesa deals with this question. There, when the 
Tathagata sees that living beings are not aware of their tathagatajfiana, he says 
(RGVV 24.4~7): 

yan nv aham esdm sattvandm aryamargopadesena' _ sarvasamjfakyta- 
bandhanapanayanam kuryam yathd svayam evadryamargabaladhanena mahatim 
samjnagranthim vinivartya tathagatajhanam pratyabhijaniran | tathagatasamatam 
canuprapnuyuh / te tathagatamargopadesena sarvasamjnakrtabandhanani 
vyapanayanti / [' Gryena margo° emended to aryamargo° according to Takasaki 1966: 397] 
Suppose now that by teaching [them] the Noble [Eightfold] Path, I remove all the 

fetters from these sentient beings that are caused by [their wrong] conceptions, so 

that through attaining power [by following]'’ the Noble Path they remove this big 

knot of [wrong] conceptions by themselves, recognize! tathagatajfana {which 

penetrates them] and attain equality with the tathagatas. Then, thanks to the 

tathagatas teaching [them] the Path, they will get rid of all the fetters caused by 

[wrong] conceptions. 

The passage is very clear in attributing the actual process of purification to sentient 
beings alone (svayam eva). The Tathagata’s role is limited to teaching them the 
Eightfold Path, thereby setting their purifying activities into motion. The text 
stresses that it is sentient beings themselves who remove their misconceptions and 
finally realize tathagatajnana. The verb used to describe this realization is 
pratyabhijanite. It is here used in the sense of “to become aware of,” “to 
recognize” something which living beings have always been carrying around with 
them, namely tathdgatajfiadna. This recognition is the result of living beings’ 
practice along the Noble Path and coincides with the full manifestation of their 
tathagatajnana, the realization of their buddhahood. The verb from 
pratyabhijanite implies, then, more than just the pure act of recognizing. 
According to the passage in the *Tathdgatotpattisambhavanirdesa, recognition 
can only take place when living beings follow the Holy Path. It is thus clear that 
the meaning of pratyabhijanite in this passage goes far beyond the process of 
abstractly remembering the fact that one has the buddha-nature. Here, the term 

13 My understanding of dryamargabaladhana follows from the two Chinese translations of the 
TUSN by Buddhabhadra ((%/\ 238; TUSNcw 624a20) and Siksinanda (22238; TUSNcu; 
273al), along with the Chinese translation of the passage in the RGVV ({% /\ 3234; in Nakamura 
1961: 46.4). 

‘4 The verb pratyabhijanite is not very common in Buddhist texts. Seyfort Ruegg (1973: 78) 
understands it as “la re-connaissance (ou anagnosis) de la réalité” in contrast to the Buddhist 
notion of “compréhension (adhi-gam-).” In the same passage he also provides us with several 
occurrences of the term in non-Buddhist schools. It is impossible to know if the verb appeared in 
the 7GS, since the Tibetan does not have a standardized translation for it. 


PDF Version: BPPB VI (2002) 

implies a fundamental change brought about by the adoption of certain rules that 
lead to the full manifestation of buddhahood. In the lines before the passage cited 
above, the Buddha had stated that, as long as living beings are defiled by wrong 
conceptions, they would not “be aware of” and not “realize” their tathdgatajfana. 
Four verbs are used in these lines, and we can see that they cover the meaning 
contained in the verb form pratyabhijanite. These verbs are jananti, prajananti, 
anubhavanti and saksatkurvanti (24.1).!° 

If we now tum to the TGS, we see that the process of purification is 
described in very similar terms. There, however, the distinction between the 
activity of the Tathagata, on the one hand, who teaches the Dharma and thereby 
makes sentient beings know that they themselves have the buddha-nature, and 
living beings’ own efforts, on the other, is less clear. It seems to me that the simile 
in the *Tathdagatotpattisambhavanirdesa already bears the seeds of such 
ambiguity. It is quite evident that the authors of the *Tathagatotpattisambhava- 
nirdesa stressed sentient beings’ contribution with the words svayam eva, while at 
the same time specifying the exact nature of the Buddha’s role in clearing away all 
doubt: the Buddha’s teaching activity, which leads to the realization of 
tathagatajnana; it is “attaining power [by following] the Noble Path” (Gryamarga- 
baladhanena) that capitalizes on this activity. This clarification seems necessary, 
since otherwise one may be inclined to read the situation according to the 
upamana, where it is the skilful person alone who makes the compressed cloth 
useful for the whole world. 

In the TGS such clarifications are only sporadically found. In most of its 
similes it is a single person who knows about the precious hidden substance and 
purifies it. In several passages this person is explicitly compared to the Tathagata. 
The most common formulation in the TGS is expressive of finality: the Tathagata 
teaches the Dharma in order to purify sentient beings. Of course, such a statement 
could mean that the Tathagata directly destroys the A/esas of sentient beings, but it 
could also be understood in the sense conveyed in the *Tathdgatotpattisambhava- 
nirdesa. Formulations in the simile of the golden figures in the molds are 
suggestive of the former alternative. There it is stated that the Tathagata with “the 
vajra[-like] hammer of the Dharma ... hews away all outer defilements in order to 
entirely purify the precious tathagata-knowledge ...” (9B) “so that their 
defilements are expelled without any remainder.” (9.4). However, it is hardly 
imaginable that this is to be taken as denying any kind of participation of sentient 
beings in the process of their own purification. It signifies rather that the authors 
tried to treat the two sides of the simile isomorphically, adopting the framework of 
the upamana when comparing the Tathagata to the smith. 

We do in fact find some passages in the sitra indicating clearly the active 
role sentient beings themselves play in the process of purification. The most 
representative among them is in 1B, where it is said that the Buddha teaches 
sentient beings the Dharma and leads them to have faith (mos) in order to destroy 
their defilements. Sentient beings then apply (abhiyujyate) themselves assiduously 

‘5 Tn the TGS the Tibetan does not allow the exact Sanskrit terms to be reconstructed where it 
is said that sentient beings “realize” (khong du chud pa) their buddha-nature after learning about it 
through the teaching of the Tathagata. It is quite possible that any such Sanskrit term could be 
understood in both ways, i.e., “to become aware” and “to accomplish.” The former alternative 
would entail that the knowledge of the presence of a buddha within was considered as leading to 


PDF Version: BPPB VI (2002) 

to these teachings and become free from their defilements.''® Again, in 5B and 5.5 
sentient beings or bodhisattvas are said to acquire faith in the teaching of the 
Buddha, which later, when they exert themselves (5.5), leads to their awakening. 
In 5.4 we find an exhortation in which the Tathagata calls upon the bodhisattvas to 
“take the treasury of knowledge” and to become protectors of the world. Finally, 
in the simile of the poor woman the Tathagata calls upon sentient beings to exert 
the energy which will lead them to realize buddhahood.!!” 

From the passages mentioned above it becomes obvious that living beings 
were expected to participate in the removal of their defilements. Nevertheless, 
under the sway of the august figures in the upamanas, the authors did not hesitate 
to use formulations which leave no or else very little room for the participation of 
living beings. Two consequences can be drawn from this observation. First, it 
demonstrates that the purificatory work of the Tathagata was conceived as the 
main factor in the process of manifesting living beings’ buddha-nature. It seems in 
fact to go far beyond his activity of teaching the Dharma. Secondly, the sutra 
touches only sporadically on the question what sentient beings have to do to attain 
awakening, when it refers to the very general notions of “faith in the Dharma” and 
its practice. This shows that the details of the process of purification were not an 
integral part of the message which the authors wanted to convey. Therefore they 
did not elaborate this point any more than they provided an exact definition of the 
buddha-nature—tasks they left for later systematizers. Other means of awakening 
(recitation etc.) suggested in the sections following the similes are a general 
feature of many Mahayana sutras, and thus not peculiar to the theory of the 
buddha-nature. This, of course, does not mean that the authors of the TGS did not 
consider it appropriate to practice them. 

4 The 7GS as a Part of Indian Buddhism: Its Sources, Motives and Reception 

The title of this section may raise expectations which cannot be fulfilled, among 
them successfully tracing the roots and defining the position of the 
tathagatagarbha teaching against the background of a comprehensive history of 
Indian philosophy in general, and in particular its relation to the upanisadic notion 
of atman and the Pasupata doctrine of abhivyakti, which latter concept, namely 
liberation as the manifestation of the innate Siva’s perfections through the removal 
of impurities (mala), is strikingly similar. I restrict myself to delineating only a 
few aspects of the sttra’s relation to other Buddhist texts. Likewise, I cannot here 
deal with the development of tathagatagarbha thought in Tibet and China. For the 
Indian region, my central ideas about post-TGS developments are based on 
Takasaki 1974, a work which has remained the point of departure for any serious 
historical study of the tathagatagarbha teaching in India. 

In Tibet, the teaching of tathdgatagarbha became a highly controversial 
issue in the well-known gzhan stong and rang stong debate among the Buddhist 
schools. The proponents of gzhan stong (“other-empty”) believe that the inherent 

6 4 similar conclusion is reached at the end of 5A. There it is said that as a result of not 
hearing of the buddha-nature sentient beings do not apply themselves to their purification. 

17 There seems to be no clear distinction between normal living beings (sattva) and 
bodhisattvas in the TGS, except for a passage in 8B where the threefold categorization sattva — 
bodhisattva — tathagata appears. 


PDF Version: BPPB VI (2002) 

buddha-nature of living beings is essentially empty of defilements and so, in this 
regard, come very close to the position of the TGS, which they take literally, while 
the rang stong (“self-empty”) position is that the texts of the tathagatagarbha 
teaching have to be interpreted on the basis of and harmonized with other major 
Buddhist tenets, such as the doctrine of emptiness (sunyata)—which finally led to 
the identification of the tathagatagarbha (“buddha-embryo”) with stnyata. 

In China, the tathagatagarbha teaching gained philosophical emponianes 
by way of the Ratnagotravibhaga and the Dacheng qi xin lun KEE. 
Opinions differ about how deeply it has in fact penetrated Chinese Buddhist 
thought, but without doubt it found fertile soil among Daoist teachings which were 
professing belief in an “original pure essence” before the arrival of Buddhism in 
China. With the rise of the concept of “original awakening” 7\# in the Dacheng 
qi xin lun and its fusion with tenets of Yogacara thought, however, the buddha- 
nature theory in China merged into a philosophy with a strongly monistic 
orientation—something the 7GS had not foreseen but nonetheless prepared the 
way for. This monistic position, which later took on more extreme forms, as in the 
attribution of the buddha-nature to the non-sentient realm, became the object of 
scholarly criticism at the end of the 20" century. The arguments of its critics, 
namely that this position, during parts of Japanese social and political history, was 
all too easily misused as a repressive tool to ensure the status quo, should certainly 
not simply be brushed aside. 

4.1 The Titles of the TGS 

The Sanskrit title transmitted to us at the beginning of Tib is Arya- 
tathagatagarbha-nama-mahayana-sitra.''* In the middle of the siitra, however, 
when the Buddha refers to the TGS, the title is given as Tathdgatagarbha-nama- 
vaipulya(or: vaitulya)-sutra (OL). This name is confirmed by both Chinese 
translations, which in the sttra itself as well as at the beginning of it speak of the 
KAS RORIAE (Chy; within the stitra: AHERN AR i) and 75 RERIAR 
ERS (Ch2). Both #72 and 7 are here renderings of vaipulya or vaitulya.'' 
(The word da X was probably added without having any direct correspondence in 
the Sanskrit.) We naturally should assume that the title as it appears at the 
beginning of a text is subject to alterations to a higher degree than a citation of it in 
the middle of one. This would mean that the title appearing in OL, namely 
Tathagatagarbha-nama-vaipulya(or: vaitulya)-sutra, was the name used in the 
earliest times, since it was faithfully transmitted by the Chinese versions, so that 
mahayana, which replaced vaipulya (or: vaitulya), as mirrored in both Tibetan 
versions, must be of later origin. The Ratnagotravibhagavyakhya refers to the TGS 
twice under the short title Tathagatagarbhasitra (25.9; 66.18). 

"8 Yq Bth the Sanskrit transliteration lacks arya and nama. Arya (‘phags pa) is also missing in 
the Tibetan title of Bth, and this could thus be a characteristic of pre-standardized texts (see part I], 
C 1: Bth), but the equivalent of ndma is found in the Tibetan title of Bth (zhes bya ba), and so was 
probably erroneously omitted in the Sanskrit transliteration. 

"9 See Karashima’s textual study on the SP, where he argues that Dharmaraksa chose the 
characters 77% in order to render Skt. vaitulya (in association with Skt. tulya, “equal to”), while 
the characters #7 have their counterpart in Skt. vaipulya. He also points out the problem in 
deciding which of the two forms (vaitulya or vaipulya) was the original one (Karashima 1992: 


PDF Version: BPPB VI (2002) 

4.2 The Recorded Chinese Translations of the 7GS 

Besides the two available Chinese translations by Buddhabhadra and 
Amoghavajra, the Chinese catalogues mention two other translations, which are 
now lost. The oldest record is contained in the Chu sanzang ji ji WKy— ieee 
(CSZJJ), a work published in about 515 CE by Sengyou (##4.'"° It mentions a 
sitra called Da fangdeng rulai zang jing KT FARIA (9c20) and states that in 
(an/the) old catalogue(s) it is called Fo zang fangdeng jing (tig. 7! 
Together with three other works, |? it attributes this translation to Faju 7&6, a 
translator active towards the end of the Western Jin 4 dynasty at the time of the 
emperor Hui FX (290-306) and Huai (8 (307-311). As to the translation of 
Buddhabhadra, the CSZJJ does not provide us with information concerning the 
exact date and place it was done. It does state that according to other sources it 
was called Rulai zang but that it was currently missing 
(11c15: RZ-FMRR. SH. ). According to the CSZJJ, the TGS thus appeared in 
two different translations, one each by Faju and Buddhabhadra (14613: 
= ARH). 

In the later Zhongjing mulu 2R#& BK (ZJM) by Fajing 4A and others 
dating from 593 CE, we come across some new information. Buddhabhadra’s 
translation of the TGS is there said to have been undertaken in the Yixi #§E8 
period (404-418), and Fali 4&7 is mentioned together with Faju as joint 
translators of the same text (117c16). I will come back to this below. 

The next catalogue, the Lidai sanbao ji FE = 8% (LSJ) by Fei 
Changfang 2, dating back to 597 CE, bears witness again to a different state 
of affairs. First of all, Fei Changfang mentions a further translation of the 7GS, the 
Da fangdeng rulai zang jing KARR (66b2) attributed to Fazu ¥4H 
(religious name: Bo Yuan §43%) (66b17f.). About Fazu, we know that he was 
active during the reign of Emperor Hui (290-306), and it seems that he had close 
relations with Dharmaraksa.'” Fei Changfang ascribes to him 23 siitra translations 
which, according to the Gaoseng zhuan —&\{@/#(GSZ), were lost in a time of 
troubles, and states that he himself had made a title list of 22 formerly 
unmentioned works,!™ among them the Da fangdeng rulai zang. About the other 
early translation of the TGS, attributed to Faju 7#£H by the CSZ//, the LSJ states 
that it was translated by Faju together with Fali jw (as we have already seen in 
the ZJM) and it restricts the period of translation to the reign of Emperor Hui 
(66b26-27). While we do not know where the information about the common 

12 See Ziircher 1972: 10f. 

12) See 9c20: BERTH EE; the question what exactly is meant by the expression #$¥% is 
complex, and has until now not been answered satisfactorily. It could be a part of Daoan’s old 
catalogue, the famous catalogue of Daoan itself, another catalogue, or even a group of other 
catalogues. See Tokiwa 1938: 36-41; 96; 683. 

'22 Namely the Lou tan jing BRE (T 23), the Faju ben mo jing i RJAARKE (T 211), and the 
Futian jing *§% (T 683) (9c19-22); the CSZJJ states that the latter two siitras were translated 
by Faju together with Fali #£ 17 (10al—3). See also CSZJJ 98a28-29 (without any reference to the 

13 Zarcher sees relations being maintained between Fazu and one of the collaborators of 
Dharmaraksa (1972: 76f.); Tsukamoto states that he was a disciple of Dharmaraksa (Tsukamoto 
1985: 730). 

124 SJ 66b17—20, based on GSZ 327b28~29 (=GSZps 40f.). In the CSZ/JJ (9c16-18) and the 
Zhongjing mulu 3R#8 B & (ZJM; 122c19-20 and 116b6), however, only the Wei dai pusa jing 
HES SEER is ascribed to Fazu. 


PDF Version: BPPB VI (2002) 

translation of the 7GS in the two catalogues comes from, the dating of it to the 
reign of Emperor Hui can be explained by another passage in the LSJ which states 
that Faju, after the death of Fali, continued to work alone on the translation.'*> We 
thus should conclude that, according to the LSJ, the period of activity of Fali was 
limited to the reign of Emperor Hui. 

Regarding Buddhabhadra’s translation, the LSJ provides us with the date 
420 CE (cER—+£) and gives the Daochang temple 442) in Nanking as the 
translation site (71a13). However, the LSJ considers this translation to be, 
chronologically, the second one, and characterizes it, strangely enough, as “little 
different” (/}\32) from Fali’s (71a13—14). It remains unclear why the translation by 
Fazu, which the same catalogue is the first one to mention, is not counted. In all 
following catalogues it is dealt with as the third one.'*° From then on newly 
compiled catalogues do not contain any further information, so that we are left to 
sort out the contradictions of the available catalogues by ourselves. 

Let me first deal with the translation attributed to Fazu. Whereas both the CSZ/JJ 
and the ZJM attribute only one stitra to Fazu, namely the Wei dai pusa jing 
HES =ETERE, Fei Changfang states that he consulted other catalogues and could 
identify 22 further texts translated by him.'”’ Now, Fei Changfang, the compiler of 
the LSJ is well known for his practice of attributing hitherto unknown or 
undetermined works to various translators without revealing his sources, basing 
himself on catalogues that he never could have seen or failing to provide any other 
convincing arguments.'”* This, though, is not sufficient grounds for discrediting all 
the new information he came up with. It is obvious that he had access to a 
different, and in several respects, more comprehensive set of catalogues,'”? and it 
would be wrong simply to disregard all his assertions. The attributions to Fazu do, 
however, seem not very convincing. There is a remark in the biography of Fazu in 
the Gaoseng zhuan {2S to the effect that besides three works whose titles are 
given, he also translated several short texts no longer found, and whose titles were 
even no longer known.” ° Fei Changfang toned up this remark, and so contributed 
to the glory and fame of Fazu and his era. Tokiwa, in an extensive analysis of the 
attribution of the 23 works, understandably laments the fact that Fei Changfang 
simply mentions “several miscellaneous catalogues” ( 4 #é ac #& ) without 
specifying his sources,’®’ Of the 23 works, only five are extant today.'?* That some 
of the other remaining 18 attributions to Fazu were already doubted at an early 
date is proved by the Kaiyuan shijiao lu Ei 7CRERAGER (KSL), which singles out 

Wirt ertise Rig SA. A reference to the death of Fali is also contained in the CSZ// 

28 See T6d6 1959: 1. 

127 TJ 66b17-20. The Gaoseng zhuan #54 attributes three works to Fazu (327627; GSZps 
40f.), among them the (£3227 "#4%, which is also included in the list of 23 works in the LS/. The 
titles of the other two works, however, are not found in the Fazu passage of the LSJ. 

128 See e.g. Hirakawa 1978: 28f.; 36; 37f£.; 44. 

2 See e.g. Hirakawa 1978: 39. 

30 See GSZ 327b27ff.=GSZps 40f. 

31 Tokiwa 1938: 704; his analysis is found on pp. 703-707. 

12 Namely T 5; 144; 330; 528; 777. These same five texts were already the only ones extant at 
the time another catalogue, the Kaiyuan shijiao lu BE7CRERGR (KSL; T 2154) of 730 CE, by 
Zhiseng “95, was being compiled (498b28-4). 


PDF Version: BPPB VI (2002) 

seven titles from the list of Fei Changfang as of questionable authorship. !°° 
However, the KASAI does not share such doubts. 

Tokiwa, for his part, makes clear that the translation of the TGS attributed 
to Fazu is very unlikely. He assumes that this attribution resulted from an entry 
which ascribes a translation of the TGS to Faju of the same period, whence it was 
erroneously attributed to Fazu. This is thus a classical case of “multiple 
attribution.”'34 We can, of course, not be absolutely sure that Tokiwa’s conclusion 
is right. However, the fact that nothing about the translations is mentioned in the 
older catalogues, the circumstances of the relatively late attribution to Fazu, and 
Fei Changfang’s silence about his sources should make us skeptical. Finally, we 
can also interpret the above-mentioned fact that the LSJ considers Buddhabhadra’s 
translation of the TGS to be the second one as an indication that its own new 
attribution of a third translation to Fazu lacked enough grounding to have any 
effect on the hitherto transmitted information about the TGS having been 
translated only twice. 

If we now turn to the translation of the TGS by Faju, or by Faju together 
with Fali, as the later sources have it, the basis for such an assumption appears 
much more solid. As stated above, the existence of such a translation is already 
attested in the oldest existing catalogue, and, furthermore, both Tokiwa and 
Hayashiya include the entry in their reconstructions of the catalogue of Daoan, on 
which the CSZJJ, to a large extent, is based.'*> In the concerned entry, the CSZI/ 
notes below the sutra title that in an/the old catalogue(s) the text was named Fo 
zang fangdeng jing (ii 17S (9020: BRR HI HT SX). It is not entirely clear 
to which catalogue(s) this remark refers. Tokiwa states that in this case it should 
undoubtedly be understood as referring to Daoan’s catalogue, °° but Hayashiya 
tends to believe that the words #3 #%, when appearing below the main entry, 
consistently mean the catalogue of Zhu Daozu ““3H#H of 419 CE.!°” As the more 
than one thousand pages each of Tokiwa and Hayashiya document, the study of 
the Chinese catalogues is a painstaking and complex endeavor. I myself have 
nothing to contribute to a final solution to the questions raised by the expression 
2%. The issue is of relevance for deciding whether the title AA SAIRRE 
appears as such in Daoan’s catalogue or whether it was introduced by Sengyou on 
the basis of, for example, the existence of the same title for Buddhabhadra’s 
translation. Further, even if we assume that the title (Hg 77%, below the main 
entry, derives from the later catalogue of Zhu Daozu (and not from Daoan’s 
catalogue), can we be absolutely sure that the title KA SAMRIA and the titles 
generally found in the CSZ//J appeared as such in Daoan’s catalogue? 

This question is not dealt with by the two great Japanese pioneers of 
catalogue research, and they seem to take it for granted that Sengyou’s avowal that 
he was following Daoan’s catalogue guarantees the authenticity of titles. It is only 

33 KST 499al1-19. 

34 Tokiwa 1938: 381; 683; 704. 

35 Thid., 168; 381; 583; 678; Hayashiya 1940: 406. 

6 Tokiwa 1938: 683: “KM, HHOMOBRL IL, WPS (He) ROSLEMMLS, ” Ina 
more general analysis of what could be meant by the words #&, he comes to the conclusion that 
they stand for Daoan’s.catalogue along with three other ones. He assumes that among these four 
catalogues it is the catalogue of Zhu Daozu “34H, completed in 419 CE, which is meant in most 
of the cases (Tokiwa 1938: 40f.). 

37 See Hayashiya 1940: 366f. Harrison (19907210) doubts Hayashiya’s conclusion, since the 
“Old Catalogue” is also cited for works translated after 419 CE. 


PDF Version: BPPB VI (2002) 

in a short passage that Tokiwa, when expounding the principles guiding his 
reconstruction of Daoan’s catalogue, touches on this issue.'?* In this passage he 
deals with a statement found at the beginning of Sengyou’s list of sutra titles 
which had not been attributed to an author by Daoan (#722273 KEG). ?? 
Sengyou there describes the extreme brevity of Daoan’s entries and the poor state 
of the copy of the catalogue: titles were abbreviated with only two characters, the 
number of juans # were not mentioned, the lines remained without a space in 
between, confusion caused by later scribes (4 A (433) in regard to titles and to the 
number of works was normal, and doubts arose because punctuation marks 
formerly written in red ink had vanished. Since this description is not found at the 
beginning of Sengyou’s catalogue, I am not sure if we should assume that it refers 
also to the part before, where the entry of the TGS is situated. Tokiwa notes that 
Sengyou’s work itself included the complementation of such abbreviated titles, 
and stresses that we can no longer know where the line between Daoan’s original 
entries and Sengyou’s additions should be drawn.'*° 

In any case, we can be sure that at the time of Daoan (314-385), whose 
catalogue is widely esteemed as a scholarly work of the highest standard,'*’ a text 

called (yt 7 EX or even AA SERRE was known and attributed to Faju. 

We are now faced with three basic possibilities regarding the title of the TGS 
attributed to Faju: 

(1) The title AAS AMAFAE or an appropriate two-character abbreviation 
(20%?) already appeared in Daoan’s catalogue. ##% in the annotation of 
Sengyou refers to the catalogue of Zhu Daozu “34H, where (ii 7 GEA is 
given as another name of the siitra. For the alleged identity between the 
{sha Hy SK and the PEK HE (the title of a siitra in the later catalogues) 
see below. 

(2) The title KA;SEANAGEHE was associated in other catalogues with Faju. 
Sengyou concluded that it was probably identical with the (i HER listed 
in the catalogue of Daoan. He was referring to Daoan’s entry when he wrote 

(3) Sengyou, concluding that the stra listed under the title (ii ER or (IK 
in Daoan’s catalogue was a different translation of the text translated by 
Buddhabhadra, “modernized” the title according to this latter translation. The 
title (ey SX in the annotation could, however, also have come from Zhu 
Daozu’s catalogue. If the title found in Daoan’s catalogue was (#48, Sengyou 
would have understood {# to be an abbreviation for 412K. 

These three possibilities indicate in themselves how much we are forced to work 
on the basis of assumptions when dealing with early Chinese catalogues. Thus, in 

38 See Tokiwa 1938: 94-96. 

3° See 16c8-17. 

140 See Tokiwa 1938: 95. Tokiwa obviously holds that when Sengyou states that the siitra titles 
in Daoan’s catalogue consisted only of two characters, he is referring to the titles at the top of the 
main entries. The passage in question runs as follows: FHM ie. (16c12). The 
expression 3+ A #%%, however, could, even if this is much less likely, also be interpreted as 
alluding exclusively to the cases where Daoan provides further information on the texts in the 
annotations below the main entries. 

‘1 See e.g. Hirakawa 1978: 31. 


PDF Version: BPPB VI (2002) 

contrast to his clearly noting that the translation of Buddhabhadra was lost,!*? we 
do not know if Sengyou had access to a copy of Faju’s translation. The 
identification of the TGS with the (Hi AZ is therefore especially troublesome, 
and can hardly be accepted without reservations. '” It is very likely that Sengyou 
only knew the title from a catalogue, namely the ##, and so had no chance to 
compare the siitra’s content with what he knew about the TGS. It is true, as already 
remarked, that (#637 4 could easily be a rendering of the title of the TGS as it 
appears in the text itself (OL), namely Tathdgatagarbha-nadma-vaipulya-sutra, 
assuming that tathagata was turned into {#. However, the character i seems to 
have been used from earliest times to render different Indian terms, so that (ix 
could also be a translation of buddhakosa, °gotra, °niddna, °pitaka and the like, 
while the addition 77} for vaipulya appears frequently in the CSZJJ. Further, we 
know of the existence of a siitra called (7 SX, which appears for the first 
time in the Zhongjing mulu FURS A RK (119c24), where it is classified as a different 
translation of a part of the 3E¢8 and where the titles FAJAABAKE and AXES, are 
given as alternatives. In the LSJ, in one of its typically not quite reliable entries, 
we learn that the same stitra is a translation of Daoyan {Hf made during the Song 
5€ period and that a different title for it is FJBASES.’ This information appeared 
from then on in the catalogues. The siitra is not extant, and again we can only 
speculate. Tokiwa, following a note in the KSL,'* tends to believe that the Nei 
zang da fangdeng jing WiktK HS was identical with the (iA AER and 
adds that it should have been the K77“EAIRIER of Faju.' 

All of this demonstrates that there cannot be any definitive answer to the 
question what Indian work lies behind the title of a Chinese translation which has 
not come down to us. Still, the information regarding the 7GS translation by Faju 
seems quite solid, based as it is on the catalogue of Daoan. The crucial point is the 
title attributed to it by Daoan, a problem which I cannot solve with precision. If we 
accept the first possibility above, there is no reason to doubt the early existence of 
Faju’s TGS. In both other alternatives, based on the reading (#64 instead of the 
ARI of Daoan’s catalogue, we are on less reliable ground. The annotation by 
Sengyou (Rk...) thus complicates matters, but at the same time it hints at where 
the source of a possibly mistaken identification of Faju’s translation with the TGS 
could be. For the time being we can only note that there is not too much reason to 
doubt the attribution of the 7GS translation to Faju. It is to be hoped that the future 
will bring a comprehensive study of the terminology and abbreviations involved in 
the early Chinese catalogues. 

There are some other irregularities concerning the TGS in the catalogues that 
deserve our attention. I have no concrete idea how to solve them, and thus shall 
restrict myself to simply mentioning them. First, there is the fact that the CSZ// in 
the 13" juan on the biographies of important monks (where, in an afterthought to 
Dharmaraksa, Faju and Fali are also treated) speaks of two translations done by 

42 See 11017: SHR. 

19 The title #EiI7°EKK is, however, except for the rendering #§ in regard to tathdgata, a 
perfect translation of the original Sanskrit title of the siltra: Tathdgatagarbha-nama-vaipulya(or: 
vaitulya)-siitra. The character % of the later Chinese versions has not been added to 773 in the 
translation of vaipulya (or: vaitulya). 

“4 O4a3ff. 

#5 See 502a12: NASR —E (SRERRAD SS). 

4 Tokiwa 1938: 768. 


PDF Version: BPPB VI (2002) 

Faju together with Fali, but of only one, the Lou tan jing HERE, done by Faju 
alone.'*’ This is clearly in contrast to the two other passages in the same catalogue 
where the AAERIARS of Faju is mentioned. Is this an unintentional 
omission, or should we assume that the biography comes from a different strand of 
transmission, one that does not know of a fourth translation by Faju? 

Equally surprising is the statement in the CSZJJ that the AFFAIRS 
of Buddhabhadra is lost.'** A survey of the 11 works attributed to Buddhabhadra 
in the CSZJJ shows that among them four are said to be lost: the KA SEAIARR, 
the Xin wei mi chi jing OUR RHS, the Ben ye jing ARS, and the Jing liu 
boluomi jing 7N 7 HEE. Out of these four, only the KA FAIA is extant 
today, whereas in subsequent catalogues the other three are either confirmed to be 
lost or said to be just different names for other works translated by Buddhabhadra 
which never had really been lost. It is also striking, then, that among the four 
translations attributed to Faju in the CSZJ/ it is the TGS alone which has not come 
down to us.'*? 

Regarding the dates of the translations of the TGS, I will not go into detail 
in the case of the first, ascribed to Faju. The different attributions, once to Faju 
alone, and later to Faju together with Fali, and even to Fali alone, need not overly 
concern us. For our inquiry the period from 290 till 311, namely the years of 
Faju’s activity, will be sufficiently precise. 

For the second translation, by Buddhabhadra, we find, as mentioned above, 
two sources containing conflicting information. The LSJ provides us with the date 
420 CE (7cER—‘) and mentions the Daochang temple GH#3#), while the 
Zhongjing mulu 348 EB # states that the translation took place in the Yixi ABE 
period (404418). No dates and places are given in the colophons. Nor does the 
oldest catalogue contain any information about the circumstances surrounding this 
translation. Obviously, the two relevant catalogues must be based on distinct 
sources. The following catalogues only repeat the one or other alternative. Neither 

7 9827-29. 


14° See KSL 505b20-c23 and Tokiwa 1938: 770-772. A tempting way to explain the reportedly 
missing translation of Buddhabhadra in the CSZ/J is to assume that Sengyou erroneously 
considered Buddhabhadra’s to be in reality Faju’s translation. This would also answer the question 
why, if the words #% in the CSZJJ do in fact refer to Daoan’s catalogue, the title AASan 
7H KK (which is also the title of Buddhabhadra’s translation) was adopted as the main entry. As 
to the translation attributed to Faju, which for the first time is mentioned as lost in the #R#8 EB & of 
the year 602 CE (7 2147, 175c16-17), this would explain why it has not come down to us, in 
contrast to the other three translations by Faju: with the appearance of other copies of the text 
indicating Buddhabhadra as the translator, it would have become evident that the translation had 
hitherto been falsely attributed. Consequently, the translation by Faju would from then on have 
been classified as lost. However, it is hardly possible that a scholar like Sengyou, who we know 
analyzed the style of works of undetermined translators mentioned in Daoan’s catalogue, made 
such a mistaken attribution, involving the styles of two very distinct translation periods. Further, it 
is difficult to think of a reason why a translation of the 7GS with no indication of its translator 
should be attributed to Faju rather than to Buddhabhadra, who at that time was without doubt far 
more well known. (On the work of Sengyou regarding texts with undetermined translator, the 
accuracy of his catalogue, and also some of his mistakes, see Okabe 1973.) Alternatively, one 
could assume that the translation of Faju was later erroneously attributed to Buddhabhadra or that 
Buddhabhadra revised Faju’s translation and published it under his own name, thus contributing to 
the “loss” of Faju’s work. However, the choice of terminology in the translation attributed to 
Buddhabhadra does not show any traces of “old” vocabulary belonging to the pre-Kumirajiva 
period, so that this assumption has little plausibility. 


PDF Version: BPPB VI (2002) 

of the periods is particularly associated with Buddhabhadra’s most important 
translation, the Avatamsakasitra, and we thus have no criterion for judging the 
authenticity of the two entries. It is not possible to specify the period of 
Buddhabhadra’s translation more precisely than within the period from 404 till 

Amoghavajra’s translation of the TGS is first mentioned in the Da tang 
zhen yuan xu kaiyuan shijiao lu KE ATCA ICH AGE of the year 794 CE under 
the title Da fangdeng rulai zang jing KA BYVRIE.'” The catalogues do not 
provide us with any information about the time and place of the translation. As 
neither catalogues nor the colophon'” of the translation itself sheds light on the 
date of Amoghavajra’s achievement, we can only assume that it was in the period 
after 741, the year of the death of his teacher Vajrabodhi and his journey to Sri 
Lanka, which marks the beginning of Amoghavajra’s steeply rising career as a 
translator and high-ranking priest. He died in 774.’ 

4.3 Possible Motives of the Authors of the TGS 

I have shown above how the idea of living beings endowed with the buddha-nature 
does not appear as a well-considered, completely consistent doctrine in the TGS. 
The exact nature of the inherent buddhahood of living beings remains hidden 
behind the vivid descriptions of the similes. Hence we may conclude that the TGS 
results less from careful abhidharmic considerations about the character of the 
buddha-nature than from the zeal of one or more persons motivated by other than 
doctrinal interests.’ The decision to expose their message exclusively in easily 
understandable similes probably indicates that the targeted readership were non- 
specialists in the field of Buddhist abhidharma. Only the use of such colorfui 
images made it possible to acquaint a broad audience with the abstract idea of 
buddhahood in all living beings. The emphasis lies on the message that all living 
beings have the buddha-nature rather than how this buddha-nature should be 
grasped in exact philosophical terms. 

Of course, we cannot know whether the idea of a buddha-nature in living 
beings resulted from a novel meditative experience or because the authors felt the 
need to assert its existence in order to improve an unsatisfactory worldly or 
philosophical state of affairs, or whether it was based on other experiences. All 
this is mere speculation. Further, it would certainly be inappropriate to assume a 
model which reduces the range of possible motives to a single one. The first 
appearance of the tathagatagarbha theory in India may well have been due to 
several different motivations, brought together by possibly more than a single 
author. It is thus natural to seek out several complementing motives as the ones 
that underlie the creation of the TGS. 

5° 72156, 753c11 and 768b17. 

‘51 The colophon consists nearly exclusively of the titles granted to Amoghavajra by officials. 
A translation of the colophon is included in the critical editions of part II. On the life of 
Amoghavajra, with references to the titles bestowed upon him, see Chou 1944/45: 284-307. 

'S2 For all dates and events regarding Amoghavajra’s life see Chou 1944/45. 

‘3 For a different opinion cp. Kariya 1979, esp. pp. 1136-1139. He considers the TGS, for 
reasons he does not reveal to us, a representative of “polished and intellectual, philosophical 
thought,” as consistent with the background of an “increase in monastic features.” 


PDF Version: BPPB VI (2002) 

One of the authors’ goals may have been to widen the circle of Mahayana 
followers by announcing that non-Mahayana Buddhists and even non-Buddhists 
were in possession of the buddha-nature. In this way, they could count on 
facilitating their conversion to Mahayana. The ekayana doctrine, suggesting that 
there is no other means to final emancipation than Mahayana, had already been 
earlier expounded in the Saddharmapundarikasutra, a text which, as I have been 
trying to show, must have had a strong impact on the authors of the TGS. 

Be that as it may, the doctrine put forward in the TGS would certainly 
encourage all segments of Buddhist society, and even non-Buddhists, to strive 
more earnestly for buddhahood, which, once reached, would result in their 
liberating activity on behalf of other living beings. If there is any concrete motive 
on the part of the authors which is formulated in the stitra itself, it is precisely this 
encouragement, as it appears in the simile of the desperate woman with the 
universal emperor in her womb (8A-C). There we find the following call by the 

Sons of good family, apply energy without giving in to despondency! It will 
happen that one day the tathigata [who has] entered [and] is present within you 
will become manifest. Then you will be designated ‘bodhisattva’ rather than 
‘[ordinary] sentient being (sattva).’ [And] again in the [next stage you] will be 
designated “buddha,” rather than ‘bodhisattva.’ 
In the TGS it is this message alone that directly urges the people to draw 
consequences from the fact that they possess the buddha-nature. The call is not 
very detailed, and we cannot know what exactly the authors had in mind when 
they put these words into the mouth of the Buddha. Nevertheless, besides the call 
in section 10 for the propagation of the TGS by various means (a standard feature 
in Mahayana sitras), the passage suggests that energy (virya) was considered a 
central element to be employed. Obviously it was not the purpose of the authors to 
deal with the question of how to realize one’s buddha-nature in detail.'** Nor did 
they draw any ethical conclusions. This is surprising for the modern reader, since 
the tathagatagarbha theory would seem to be an ideal ground for establishing an 
ethical system, namely one based on the principle that all living beings are equal 
by virtue of their buddha-nature. This absence of ethical implications indicates that 
the (early) buddha-nature theory centered on the importance of the individual’s 
inclusion in the “family of the buddhas” rather than on a doctrinal basis for ethical 
behavior. Even in later texts of this strand, direct ethical implications continue to 
be rather seldom, in contrast to the prevailing worldly orientation of some of the 
siitras propounding the tathagatagarbha theory.'* 

'S4 Kariya characterizes the doctrine put forward by the TGS as a system void of any active 
participation of living beings in purifying themselves from their defilements (Kariya 1979: 
1134ff.). Indeed, this can hardly be expected given the purpose of the siitra stated above. The fact 
that no such activities are described extensively should therefore not be interpreted to mean that 
the TGS suggests or even calls for the sole lordship of the Tathagata in this endeavor (see my 
discussion in section 3.3). 

55 Ethical implications in the texts propounding tathdgatagarbha thought are found, for 
example, in RGVV 1.157ff. which, on the basis of the buddha-nature doctrine, calls for the same 
respect for other living beings as for a teacher. In the Angulimdliyasutra, the doctrine of 
tathdgatagarbha is used to argue for a life of chastity and continence, against killing, and against 
the consumption of meat (see Seyfort Ruegg 1980: 236; T 120, vol. 2, 540a-541a). In the same 
article, Seyfort Ruegg supposes that tathdgatagarbha thought provided the decisive motive for the 
appearance of vegetarianism in Buddhism. The worldly orientation of some of the sitras of the 


PDF Version: BPPB VI (2002) 

As I have argued elsewhere (1999: 165-168), another possible driving 
force behind the creation of the TGS may have been the need to explain why all 
living beings can attain buddhahood—a central issue in Mahayana Buddhism, 
forcefully put forward in the Saddharmapundarikasitra. Chapter XIX of the 
Saddharmapundarikasitra, containing the story of the bodhisattva Sadaparibhita, 
is of special interest in this context. According to the Saddharmapundarikasutra, 
this bodhisattva acquired his name from his habit of declaring to all monks, nuns 
and lay devotees that they were not despised (aparibhiita),'** because in the future 
they would all attain buddhahood. For these sentiments, he had to endure harsh 
words and even bodily attacks from his interlocutors.'*’ Apparently, he could not 
provide reasons for his assertion that buddhahood is accessible to all. The TGS, on 
the other hand, deals precisely with this question. Its main intention is to awake 
living beings to the fundamental truth that they all already contain buddhahood 
within themselves and thus, sooner or later, will all become fully awakened 
tathagatas. Once this basic idea took hold as sound soteriology, there would be no 
reason anymore to attack Sadaparibhuta. From this perspective, the TGS can be 
understood as a text that subsequently provided a theory for the great assertion put 
forward in its popular predecessor, the Saddharmapundarikasutra, namely that all 
living beings will attain buddhahood. Thus while the Saddharmapundarikasutra, 
in a religious environment where it was probably not always heartily welcomed, 
announced that all living beings would eventually attain buddhahood, the TGS, in 
a successive step, adopted as its task that of working out a justification for this 
assertion. That the 7GS itself is introduced with impressive scenes of lotus 
flowers, and that the first simile has made use of this scenery to present the central 
message of the siitra, can therefore hardly be a coincidence (the lotus figures in the 
title under which the Saddharmapundarikasutra became known). 

4.4 The 7GS in the History of Indian Buddhism 

The analysis in section 4.2 has shown that there is not much reason to doubt that 
the first translation of the TGS was, as the Chinese catalogues have it, already 
executed around the end of the third century CE. I will now try to establish a 
second terminus ante quem for the composition of the TGS on the basis of its 

buddha-nature teaching is best documented by the Bodhisattva-gocaropaya-visaya-vikurvana- 
nirdeSa-sutra, which contains a long chapter on royal ethics (see Zimmermann 2000). 

56 The name Sadaparibhiita, of course, can also be analyzed as sada-paribhuta (“always 
despised”)—a possibility not dealt with in the SP (for the discussion of possible interpretations of 
the compound in the SP see Ueki 1998). The preacher of the “All will become buddhas”-doctrine 
in the SP, namely Sadaparibhita, in the beginning despised by his fellows (sada-paribhuta), would 
in the long run win appreciation (sada-aparibhuta). The name of the bodhisattva is thus 
preprogrammed to adapt to a future when the idea of the buddhahood of all living beings would 
become a common tenet. (For similar ideas see the end of the Bodhisattva-gocaropaya-visaya- 
vikurvana-nirdesa-sutra where it is stated that the stitra will be found again after fifty years when 
people would follow Mahayana and pay proper reverence to the text. A summary of the siitra is. 
found in Zimmermann 2000: 180ff.) One final point: assuming that the name of the bodhisattva 
was not chosen incidentally, it is surprising that the question of the potential attainment of 
buddhahood was linked to such characterizations as “despised” and “non-despised.” 

'57 For phraseological parallels between this chapter of the SP and the TGS see Zimmermann 
1999: 159-161. 

188 See Tola and Dragonetti 1996/97. 


PDF Version: BPPB VI (2002) 

relation to other texts. The focus of my analysis is the Ratnagotravibhdga- 
(vyakhya), in which all nine similes are found, albeit recomposed, along with 
quotations from the TGS in other passages. What can be said about the 
approximate date of origin of the Ratnagotravibhdga? Back in 1966 Takasaki 
showed that the Ratnagotravibhaga(vyakhya) consists of at least two different 
layers. Schmithausen elaborated on Takasaki’s theory in a review article and 
convincingly demonstrated that the recomposition of the similes in the TGS 
formed an essential part of the oldest stratum of the Ratnagotravibhaga.'” This 
part consists of nine verse triplets (with the exception of I.108—111 where the set 
consists of four verses), each dedicated to one simile. The structure of each triplet 
is identical: two tristubh or jagati verses followed by one verse in the elaborate 
meter sardilavikridita, which repeats the content of the two preceding verses. This 
structure was probably chosen with the esthetic expectations of an educated 
readership in mind. It is very rare in Indian Buddhist literature. There is, however, 
at least one notable parallel, in the chapter of a work whose content places it in 
close proximity to the Ratnagotravibhaga. 

The text I am referring to is the ninth chapter of the Mahayanasiitralam- 
kara. This chapter deals with the issue of awakening and discusses, among other 
things, the omniscience and nonduality of buddhahood, the way it is manifested 
and its profundity (gambhirya).'® Now, if we look at the first six verses of the 
ninth chapter, we find a very similar metrical structure: verses 1 and 2 are in 
anustubh meter, while verse 3, recapitulating what is said in verses 1-2, is a 
Sardulavikridita verse. The same holds true for verses 4-5 and 6. We could argue 
that this is pure coincidence, but the chapter has also other affinities with the 
Ratnagotravibhaga. I am not arguing that chapter 9 of the Mahayanasutralamkara 
subscribes to the tathagatagarbha theory. Nevertheless, there are elements which 
suggest a close relation between and common background to the Ratnagotra- 
vibhaga and this chapter. For example, in verse 37 there is the formulation 
tadgarbhah sarvadehinah which, as the commentary makes clear, means that all 
living beings contain a tathagata (sarvasattvas'®' tathagatagarbha ity ucyate). The 
formulation is a citation from the TGS (see 0M; 1A; 1B). In other verses, too, the 
idea of buddhahood already present within is prevalent. In verses 1-3 buddhahood 
is compared to a “casket of jewels” which only has to be opened; in the next verse 
triplet it is said to be a “mine of jewels.” 

A much more profound analysis would be necessary in order to define the 
exact relation between the ninth chapter of the Mahdydnasiitralamkara and the 
ideas found in the Ratnagotravibhaga. This cannot be done here. What can be 
said, though, I think, is that the texts have a common background and that, in view 
of their unique metrical structure, are very likely to have been composed in the 
same or in a closely similar milieu. For the date of the TGS this means that the 
period of the Mahayanasiutrdlamkara’s composition can serve as a terminus ante 

To establish a definite date for the composition of the Mahayanasitralam- 
kara is impossible. Speaking in terms of historical development, it is generally 
assumed that Vasubandhu’s Trimsika represents a later stage of Yogacara 

59 See Takasaki 1966: 10-19 and Schmithausen 1971: 123-130. 

160 A short overview of the contents of the ninth chapter is found in Griffiths 1990: 5Of. 

'6l The text in the edition reads sarve sattvas. This has to be emended on the basis of the 
manuscripts now available. 


PDF Version: BPPB VI (2002) 

philosophy than the Maha@ydnasitralamkara. In 1992 Schmithausen published an 
article about two close parallels in the Trimsika and the Lankavatarasitra. If we 
accept his conclusion, the Lankavatarasitra utilized the Trimsika as one of its 
sources. The citation of the Trimsika in the Lankavatarasiitra is already found in 
the Chinese translation of the latter by Gunabhadra, which dates back to 443 CE. 
We therefore know that by that point the Trimsika must already have been in 
existence for some time. The Lankavatarasutra was probably translated into 
Chinese not immediately after its completion, nor is it likely that the author of the 
Lankavatarasiitra would have cited the Trimsika by the then already well-known 
Vasubandhu as soon as it had appeared,’ presenting, as he has done, the citations 
as the words of the Buddha. If we consider these two chronological factors, it is 
very unlikely that the composition of the Trimsika took place later than in the first 
two decades of the fifth century. 

On the other hand, we may expect that a considerable period of time 
elapsed between the Mahayanasiutralamkara and the Trimsika—this in order to 
explain the maturing of Yogacara ideas that took place in the interval. It is, of 
course, a difficult task to estimate the time necessary for such fundamental 
developments within philosophy. However, it may not be too speculative to 
suggest that some decades must lie between the Mahayanasiitralamkara and the 
Trimsika, and thus that the Mahaydnasutralamkara could not easily have been 
composed after the ninth decennium of the fourth century. If we assume this 
period as the latest possible date, too, for the composition of the 
Ratnagotravibhaga, we arrive at approximately the middle of the fourth century as 
terminus ante quem for the TGS (more exactly: TGS). 

In my calculation of the span of time between the various milestones, I 
have tended not to assume too long periods. Frauwallner, in his Philosophie des 
Buddhismus, suggests 250 CE for the date of the composition of the 
Ratnagotravibhaga,'® and would consequently arrive at a much earlier date for 
the TGS than my terminus ante quem implies. In any case, the terminus ante quem 
of ca 350 CE established by my analysis and the information about the first 
translation of the TGS gained from the Chinese catalogues (ca 300 CE) do not lie 
far apart. The composition of the TGS in the second half of the third century is 
thus a fairly plausible assumption. 

We do not have any sources of information about how siitras came into existence 
in the third century AD and who the groups were that composed and propagated 
them. As proposed in section 4.3, the main motive for the creation of the TGS may 
have been the wish to encourage (potential) believers to engage in active Buddhist 
practice and to provide arguments why buddhahood is theoretically accessible to 
everyone—a position which was put forward in forceful terms in the 
Saddharmapundarikasutra. The formulation of the tathagatagarbha theory in the 
TGS is not the work of sophisticated philosophers. The whole siitra, rather, seems 
to exude an air of pragmatism, with the efficacy of the buddha-nature being 

182 See the biography of Vasubandhu (the author of the AK), which states that towards the end 
of his life he lived at the royal court “greatly honoured by all” (discussed in Frauwaliner 1951: 25). 
The Trimsika, it should be said, is believed to be one of his latest works (see Schmithausen 1992: 

"3 Frauwaliner 1994: 255. 


PDF Version: BPPB VI (2002) 

proclaimed again and again. In a previous article’ have tried to demonstrate 
how the TGS follows along the lines of the Saddharmapundarikasiitra which, even 
though the simile of the poor man with the jewel in his dress seems to call for it, 
shrinks from the assertion that living beings have an inherent potential for 
buddhahood, or even perfect buddhahood within. The story of the bodhisattva 
Sadaparibhita shows that the time for such an assertion may not have been ripe 
yet—a conclusion supported by descriptions in the Saddharmapundarikasitra of 
other harsh reactions towards its own propagators from the side of non-Mahayana 
followers. The TGS, in this point, was apparently not held back by any related 
considerations, zealously promulgating as it did the message of a universal 
buddhahood present in all living beings. A similar notion, that of tathagatajriana 
in all beings, had already been put forward in a simile in the *Tathdgatotpatti- 
sambhavanirdesa, and it seems that the authors of the TGS were aware of or even 
inspired by it when they composed their own illustrations. 

The underlying idea of the realization of a buddhahood which is always 
present in living beings was one most fit to convey to a readership of non- 
specialists in the field of abhidharma, and, in fact, as the following tradition 
shows, was extremely successfully conveyed. The main task of the TGS was thus 
fulfilled. Both the work of systematizing and elaborating the teaching and the task 
of rendering it compatible with orthodox Buddhist teachings were left to later 
exegetes. The problems they had to confront are well documented in the 
commentary on the Ratnagotravibhaga, the Ratnagotravibhagavyakhya, where 
several interpretations, including the three ways of viewing the compound 
tathagatagarbha, compete with each other without any final settlement. The 
authors of the 7GS may not have been aware of the openings their teaching 
allowed for questioning basic Buddhist concepts. Even if they were aware of them, 
their decision to compose primarily for non-specialists seemingly overshadowed 
any possible reservations otherwise. This does not mean that they thought of their 
own teaching as a mere upaya without any ontological reality to it. 

In the simile of the seed which turns into a fruit tree, the problematic aspect 
of the teaching is put into straightforward words: the result, namely the tree, is 
already perfectly contained in the seed. Seed and tree are of the same nature. The 
process of ripening is not the main consideration. The passage in question must 
have been provocative for Buddhist thinkers since it bears overtones of the so- 
called satkaryavada, the idea, central to the Sankhya system of philosophy and 
generally criticized by Buddhist philosophers, that there can be no production of 
something that does not exist before.’ In the Ratnagotravibhaga(vyakhya), as I 
will show in section 4.5, the main emphasis in the simile then shifts to the fact that 
several factors are necessary for making the seed become a tree; the process of 
growth thereby takes on more importance. Nevertheless, the problematic 
formulation harbors a potential danger for the whole tathdgatagarbha theory: 
given the emphasis on the essential sameness of living beings and tathagatas, the 
need to be purified from the accidental defilements (their presence being the only 
difference between living beings and buddhas) could easily be ignored. The notion 
that the effect is already contained in a perfect form in its cause, as opposed to 
being newly produced, could weaken the idea of “liberation,” and lead to the 

16 See Zimmermann 1999. 
16 For further examples of satkdryavada in Indian philosophy, see Seyfort Ruegg 1989: 138f. 


PDF Version: BPPB VI (2002) 

impression that serious religious practice is irrelevant, 

From a philosophical point of view, there is one factor favoring the 
introduction of a theory of manifestation of buddhahood over one of pure 
causation. The authors of the TGS, who were acquainted with and impressed by 
the Saddharmapundarikasitra, knew of the central place the teaching of the 
eternity of the Buddha had in that text. In two of their own similes they stress in 
fact the indestructible and imperishable character of gold (simile 4) and seed 
(simile 6), and claim this to be the case for the true nature of living beings as well. 
In absorbing this idea of eternal buddhahood from the Saddharmapundarikasitra, 
the authors of the TGS may have seen no other choice than to claim that 
buddhahood as realized by living beings must always have been present in them, 
albeit in a state of non-efficacy. What is eternal cannot have a beginning, and thus 
the causal production of buddhahood would violate the theory of the eternity of the 
Buddha, and buddhahood in general. However, we cannot know to what extent 
such philosophical reflections were in fact of import for the authors of the TGS. 

That the TGS is not an isolated case of a relatively early Mahayana work 
propagating a positive, enduring constituent in living beings is shown, for 
example, by the rewording of the verses of the Gaganagafijasiitra in the oldest 
stratum of the Ratnagotravibhaga.'© There it is the mind’s innate nature 
(cittaprakyti) which is compared to “space,” being described as “luminous,” 
“without cause and condition,” and “without origination and destruction.” }” 
Unlike the term tathagatagarbha, the concept “the luminous mind defiled by 
adventitious defilements” appears already in the sttras of the Hinayana schools 
and in other early texts of the Mahayana.’ It was during this same time that the 
Vatsiputriyas were active—a Buddhist school that believed in the existence of a 
“person” (pudgala) as a kind of continuous subject of rebirth and spiritual practice, 
and whose relation to the skandhas is stated to be indefinable.’ The fact that 
Vasubandhu dedicated his Pudgalapratisedhaprakarana mainly to the refutation 
of the theory of this school shows that the influence of its philosophy must have 
been felt far into the first millennium. Given this fact, the appearance and 
propagation of the TGS as a new, Mahayanist formulation of the old belief in 
terms of a positive continuous subject, namely the person or the mind, does not 
altogether come as a surprise.'”” 

Throughout the whole TGS the term sunyata does not even appear once, 
nor does the general drift of the TGS somehow imply the notion of sunyata as its 
hidden foundation. On the contrary, the siitra uses very positive and substantialist 
terms to describe the nature of living beings. In two verses in the oldest stratum of 

' Cf RGV 1.5562; see also Schmithausen 1971: 128f. 

167 Tn this study I cannot deal with the relation of the TGS to the MPNS, another fundamental 
work for the buddha-nature theory (for a detailed study see Shimoda 1997). The fact that the 
MPNS is not reflected in the oldest stratum of the RGV may mean that the parts dealing with the 
buddha-nature had not been composed by that time. The MPNS cites the TGS, the key term in the 
buddha-nature theory it propounds being buddhadhatu. The theory in the M/PNS may have been 
inspired by the TGS, but it developed from a different perspective, one of a veneration of relics 
(dhatu). For the key concept “luminous mind” see Seyfort Ruegg 1969: 41 1ff. 

168 Fg. the Dasabhimika or the Samadhirdja; see Seyfort Ruegg 1969: 41 1 ff. 

169 See Cousins 1994 (particularly n. 6 for further references to the Vatsiputriyas). 

1 See also the study on the various concepts of vijfidna in the Pali Canon in Langer 2001. 
Among the “Nebeneinander verschiedenster Positionen im Kanon” (p. 70), vijidna functions at 
times as the “tiberlebendes Prinzip,” taking its place in the womb of living beings and leaving 
them at the time of their death (pp. 1-8; 69). 


PDF Version: BPPB VI (2002) 

the Ratnagotravibhaga (I.156—-157) that mention the purpose of its instruction, we 
do find an explicit reference to sinyatd.'’! Why has the buddhadhdtu, it is asked 
there, been taught as existing in all living beings, given that all phenomena were 
previously said to be empty (sunya)? The following verse then mentions five 
deficiencies which cannot be sufficiently remedied (or are even reenforced?) by 
the latter notion, namely depression (linam cittam), contempt towards inferior 
living beings (hinasattvesv avajfia), clinging to the unreal (abhutagraha), denial of 
real phenomena (bhutadharmdpavada) and excessive self-love (dtmasneho 
‘dhikah). The question whether or not the propagation of the concept of 
tathagatagarbha by the TGS may also be grounded in such reasoning must remain 
open. That the Ratnagotravibhaga(vyakhya) understands itself to be the final and 
ultimate teaching of Mahayana is documented by its subtitle, Mahdyanottara- 
tantrasastra. Its intention, therefore, is to limit the validity of the corpus of 
Buddhist teachings that propagate sunyata thought to the realm of the accidental, 
world-constituting and unreal, and in so doing to complement this corpus. In this 
way the authors’ of the Ratnagotravibhaga(vyakhya) hoped to introduce their 
readership to a perfect understanding of the truth, and in the process, of course, 
avoid the deficiencies quoted above with which the former teachings were 
associated.’ 7” 

Some last few words remain to be said about the qualification of the TGS 
and later texts expounding tathagatagarbha thought as non-Buddhist. This 
criticism has been heard since the end of the 1980s mainly from two eminent 
Japanese scholars who created the name “Critical Buddhism” (Hihan Bukkyo #t#] 
{,#%) for their school of thought.'” Their main argument is that the tathagata- 
garbha teaching is not in line with the two factors which they consider to be the 
two basic constituents of Buddhism: the teaching of non-self (anatman) and the 
teaching of dependent origination (pratityasamutpdda). The law of dependent 
origination ought not to allow for the idea of an eternal self or an underlying basis 
or locus on which all phenomena depend. Their criticism of the TGS is partly 
valid: the existence of an eternal, imperishable self, that is, buddhahood, is 
definitely the basic point of the TGS, but the stitra does not express any 
philosophical view concerning the structure of the world. The task of drawing a 
conclusive picture of the relation between buddhahdod:.and the world of 
phenomena was left to succeeding texts, such as the Srimaladevisimhanadasutra'™ 
and the Ratnagotravibhaga(vyakhya). For some basic questions, unequivocal 

See Schmithausen 1973a: 134; for a different analysis of the passage see Seyfort Ruegg 
1973: 317, 345. 

'™ The open claim that the teaching of ‘athagatagarbha can serve as a therapy for the five 
problems mentioned above should not lead one to assume that the authors did not think of their 
teaching as revealing ultimate truth. The Buddhist teachings have always been seen as having a 
“therapeutic aspect” (on the discussion of the historical:relation between the Four Noble. Truths 
and the four systematic parts in the science of medicine and the association of the Bitddhist 
teaching with medical philosophy by the tradition itself, see Wezler 1984: 312-324). If the 
tathagatagarbha theory is understood to be an updya without any absolute truth, then so too can 
all other Buddhist doctrines. However, I doubt that this is how the Buddhist tradition perceived 

"3 See the articles and bibliography in Hubbard and Swanson 1997 for a well-nuanced 
treatment of what the discussion is about, 

14 Regarding the SMS, Richard King has demonstrated how the siltra failed to take a definite 
standpoint on several philosophical questions of fundamental importance. and so left them 
unanswered (see King 1995). 


PDF Version: BPPB VI (2002) 

answers were never given. A comprehensive dogmatic set of “classical” 
tathagatagarbha positions never did evolve in India. The reason for this arguably 
lay with the authors of the TGS themselves who, as stated above, were first and 
foremost interested in underpinning the Mahayanist assertion that all living beings 
can attain buddhahood. If they were aware of the wider doctrinal implications of 
their idea at all, they surely underestimated the difficulties that their assertions 
later presented for the establishment of a plausible philosophical framework. 
Pragmatically motivated, they accorded priority to getting their inspiring and 
positive message across over any such well-thought-out soteriology and 
metaphysics. But is there any reason why this philosophically naive birth of the 
tathagatagarbha teaching and its heritage should be labeled non-Buddhist? 

There can be no doubt that the authors of the TGS and those of following 
works were deeply rooted in a Buddhist religious environment. We can be sure 
that they thought of themselves as Buddhists, were inspired by Buddhist 
scriptures, and acted as Buddhists. One point should be made clear: if nowadays 
critics deny that the tathagatagarbha teaching is a part of Buddhism, it is because 
their attitude towards religion is mainly influenced by textual and doctrinal 
considerations. They have a clear-cut definition of what Buddhist philosophy is or 
should be about. To my mind, Buddhism is a phenomenon which cannot be 
satisfactorily described in terms of doctrines alone. To identify Buddhism solely 
with doctrines would mean to exclude from it many other equally important 
aspects such as Buddhist piety and religious practice, ritualism, ethics, art, 
archeology and rules of statesmanship. 

Further, the criteria which serve as critics’ measures for the inclusion of 
any particular philosophical doctrine in Buddhism are, of course, typical ideas 
within Buddhism. But can they really be generalized and said to be constitutive for 
all Buddhist doctrines? To my knowledge, no voices were raised in the tradition of 
Indian Buddhism that the tathadgatagarbha teaching is not Buddhist.’ And 
obviously, the criteria established by the critics were not wholly decisive for the 
philosophical climate of those early days. The Mahaparinirvanasutra and the 
Lankavatarasiitra characterize the tathagatagarbha explicitly as dtman.'"® As I 
have argued above, the TGS seems to continue a tradition in Indian Buddhism 
which was based on a positive use of the concept of “person” and was not 
committed to a one-sided denial of a self.!”’ From early times on, it seems, this 
positive strand co-existed with the majority of schools which categorically rejected 
the existence of a self. The criteria put forward by the proponents of Critical 
Buddhism are, to my mind, chosen arbitrarily, and cannot claim validity for all 
branches of the Buddhist tradition. In view of the fact, too, that we are still far 
from agreeing on the interpretation of what the founder of the religion himself had 
to say about the question of the self (we cannot even be sure what he said!), the 
critics cannot claim that their views represent the opinion of the historical Buddha. 

5 Tn the LAS the wish for clarification regarding the differences between “the doctrine of 
tathagatagarbha and the doctrine of Self of the non-Buddhist teachers” is, however, expressed by 
Mahamati towards the Buddha (LAS 78.1-4). 

6 Cf. the Mahdparinirvanasitra in the translation of *Dharmaksema (Tanwuchen S&4&3): T 
374, vol. 12, 407b9; LAS X7746 and X.760. The SMS identifies the dharmakdya with the 
atmaparamita (Tibetan: Q 760-48, vol. 24, dKon brtsegs, 'I 257b1-285a7: 280a8f.). 

” For balanced descriptions of positive and negative approaches to a self in the history of 
Indian Buddhism see Schmithausen 1969b and 1973b: 177f.; also Seyfort Ruegg 1989: chapter I, 
esp. 43f.; Bronkhorst 2000: 41-62, 91-94. 


PDF Version: BPPB VI (2002) 

The philosophical positions they hold to be tenable are thus not an appropriate line 
of approach—not even in consideration of the fact that they are dealing only with 
the limited domain of doctrinal teaching within the multifaceted phenomenon 
called Buddhism. 

4.5 The TGS in the Ratnagotravibhaga(vyakhya) and Other Indian Texts 

The TGS can be said to have been of utmost importance for the tathdgatagarbha 
theory, inasmuch as it coined its basic vocabulary and illustrated by use of vivid 
imagery the fundamental relation between the buddha-nature in living beings and 
the klesas. However, the richness and variety contained in the similes—which I 
have tried to describe above—tended to fade in later references to and 
rearrangements of the sutra. The main focus came to be the contrast between the 
pure essence on the one hand and the unpleasant material in which it is hidden on 
the other. The text in the Indian tradition which devotes most attention to the TGS 
is undoubtedly the Ratnagotravibhaga(vyakhya), the most influential sdstra within 
the corpus of tathagatagarbha works in India. It is thus necessary to deal briefly 
with the question how the Ratnagotravibhaga(vyakhya) treats the similes of the 
TGS (the only part of its contents extensively taken up by the Ratnagotravibhaga- 
(vyakhya)). However, before doing so, we need to address some other passages 
which seem to derive from the TGS. 

In TGS 7B, it is stressed that the body of a tathagata is present even in 
animals (... tha na dud ’gro’i skye gnas su song ba rnams kyang rung ste /). A 
very similar formulation is found in Ratnagotravibhagavyakhya 15.11f.: ... antasas 
tiryagyonigatesy api.... The passage could well be based on the TGS. 

Ratnagotravibhagavyakhya 72.11-12 is a word-for-word quotation from 
the TGS (8B.3-4), where it is stated that the tathagatadhdatu is existent in every 
living being without their being aware of it. This sentence and its citation in the 
Ratnagotravibhagavyakhya are discussed in section 3.1 and in the note on my 
translation. The passage deserves our particular attention in view of the fact that 
the Ratnagotravibhagavyakhya takes the term garbhagata in the sense of “in the 
state of an embryo.” I have argued above that within the TGS the term garbhagata 
means rather the location of the dhatu of living beings: “in the womb” or “within 
(living beings).”'”® An understanding of the passage in this sense is also attested 
for the parallel verses in the Ratnagotravibhaga in which the content of the TGS 
similes is rendered freely. The phraseology is a “king found within the womb” 
(1.123: garbhdntarasthe nrpe) and “good protectors within [living beings] 
themselves” (1.123: sannathesu ca satsu ... svatmGntarasthesv api). Obviously, as 
to this point, the TGS has been interpreted differently in the two citations within 
the Ratnagotravibhaga(vyakhya): verse 1.123 displays an understanding in line 
with the original meaning in the TGS, whereas the prose passage 72.11—12 
emphasizes the embryonic aspect of living beings’ buddha-nature. This aspect, as 
shown above, is inherent to the term tathagatagarbha. It seems, however, that the 
authors of the TGS, having embraced a theory of manifestation, did not stress this 
factor in any deliberate fashion. Only later exegetes like those responsible for the 

18 See pp. S8ff. 


PDF Version: BPPB VI (2002) 

prose section Ratnagotravibhagavyakhya 72.11-12 would have made use of the 
entire range of possible interpretations of the key term tathagatagarbha in order to 
reveal all its connotations. We can take the fact that the 7GS verses in the 
Ratnagotravibhaga exhibit a different understanding to be another indication that 
the author(s) of the prose and the rewriter of the TGS verses in the 
Ratnagotravibhaga(vyakhya) were not the same. 

The second word-for-word citation of the TGS is found in 
Ratnagotravibhagavyakhya 73.11—12. It contains the formulation of the essential 
law (dharmata) that all living beings always contain a tathagata (TGS 1B). The 
position of the passage in the Ratnagotravibhagavyakhya at the beginning of a 
section where the characteristics of the tathagatagarbha are discussed 
demonstrates that the author(s) of the Ratnagotravibhagavyakhya thought of it as 
one of the most fundamental expressions of the tathagatagarbha theory. The 
passage in the TGS summarizes the descriptions of the section before and was 
designed to function as a finalizing authoritative statement of general validity. It 
was understood and commented upon as such by the authors of the 

There are three other passages in the Ratnagotravibhagavyakhya where 
formulations of the TGS seem to have been taken over. They all focus on the 
relation of sentient beings to their buddha-nature: sarvasattvas tathagatagarbhah 
(RGVV 25.18; 68.13) and sada sarvasattvas tathagatagarbhah (RGVV 26.7). 
Similar wording is found in TGS 0M.19f., 1A.8f. and, as just seen, in 1B.2f. That 
RGVV 25.18 and 26.7 are inspired by the TGS is clearly demonstrated by the fact 
that only a few lines below, in 26.9f., it is stated that the three meanings of the 
formulation sada sarvasattvas tathagatagarbhah will be explained according to 
the TGS (tathagatagarbhasiitranusarena). 

We shall now turn to the section in the Ratnagotravibhaga(vyakhya) which 
deals with the nine similes of the TGS. It comprises pp. 59—72 (verses I.95—152). 
The content of the section is the following: 

(1) A summary and detailed exposition of the nine similes of the TGS (59.8-67.2; 
verses I.95—129) 

(2) The characteristics and a concordance of the defilements of the mind 
(cittasamklesa) mentioned in the nine similes (67.2-69.16; verses I.130—-143) 

(3) The characteristics and a concordance of the threefold nature of living beings’ 
buddha-nature mentioned in the nine similes (69.17—72.14; verses I. 144-152) 

As is well known, the Ratnagotravibhaga(vyakhya) itself has a complex textual 
history and consists of at least two different layers: first, a primary collection of 
verses for the most part congruent with the verses of the Chinese translation of the 
Ratnagotravibhaga found immediately before the complete Chinese translation; 
secondly, one or perhaps two commentarial layers consisting of verses and prose 
sections which must have been added at a later stage.'”” Of the sections (1) to (3), 
only the verses I.96—126 are part of the primary layer of the Ratnagotravibhaga. 
Verses 1.95, I.127—-152 and the prose belong to the commentarial level. 

The section 1.96-126 is made up of three summarizing verses (1.96-98) in 
which the upamdna elements corresponding to the klesas and to the buddha- 

1 On the textual history of the RGV(V) see Takasaki 1966: 10-19 and Schmithausen 1971: 
123-130; for a different result cp. Seyfort Ruegg 1976. 


PDF Version: BPPB VI (2002) 

essence of the nine similes are enumerated, and of nine verse triplets (with the 
exception of I.108-111 where the set consists of four verses), each dedicated to 
one simile.'®° The structure of each triplet is identical. They all start with two 
tristubh verses'*! and end with a verse in the elaborate sardilavikridita meter. The 
first verse of each triplet describes the upamdna, the second refers to the upameya, 
and the third and last verse repeats, in different words, the content of the two 
preceding verses. The verses I.96-126 thus retell all nine similes of the TGS in 
their own fashion. This free reproduction of the main part of the TGS is the longest 
citation of the TGS in Indian scriptures. Later works citing the similes of the TGS 
may have based themselves on this reproduction found in the Ratnagotravibhaga 
rather than directly extracting them from the TGS. 

As I have shown in section 1.1, the reproduction of the similes in the 
Ratnagotravibhaga is based on recension TGS. For my translation of Tib, the 
verses in the Ratnagotravibhaga functioned as an important guide to 
understanding certain passages which are unclear or ambiguous in the Tibetan 
translation. Though the content of the verses in the Ratnagotravibhaga diverges in 
some cases quite significantly from TGS2, we find many passages where the later 
writer employed similar grammatical structures and syntactic patterns to those of 
TGS, made use of its vocabulary, and faithfully transmitted its basic ideas. There 
is no indication that the verses in the TGS were composed in various meters (as 
there is in the Ratnagotravibhaga)—such things as different numbers of syllables 
in Bth or different numbers of characters in the Chinese translations. It is possible 
that the author of the Ratnagotravibhaga recomposed the section of the TGS, 
instead of citing it word for word owing to metrical considerations. We do not 
know the reason for the very rare metrical structure within the verse triplets of the 
Ratnagotravibhaga. This structure, requiring the use of a particular meter, may 
have forced the author to recompose the whole section. In any case, he did not 
slavishly follow the text of the TGS. He freely substitutes one word for another, 
introduces new formulations and features, emphasizes elements which in the TGS 
seem to belong to the background and, in general, illustrates the similes in a more 
refined and poetic way. The direct speech of the Buddha, including his questions 
and encouragement, which frequently interrupt his account of the similes in TGS), 
do not appear in the Ratnagotravibhdga verses. It is nevertheless clear that the 
author had an echo of the wording of the TGS in mind when composing the verses. 
This is documented by a number of passages in both texts which are in close 
accordance. '® 

The term tathdgatagarbha does not appear once in the reproduction. 
Instead we find the bahuvrihi compound sambuddhagarbha, relating to jagat 
(1.101), and the term garbha, in the meaning of “embryo” and “womb,” in the 
verses describing the poor woman in the poorhouse (I.121—123). In contrast to the 
TGS, the term (dharma-)dhatu is frequently used (1.103; 116; 122) in order to 
designate the buddha-nature of living beings. And whereas in general the 
concerned verses in the Ratnagotravibhaga are firmly based on the idea of a 
manifestation of an already perfect buddha-essence within living beings, the simile 
of the sprout in the seed has been adapted to Buddhist abhidharmic considerations. 

189 Here I recapitulate what was written in the first paragraphs of section 4.4. 

181 Onty the verses 1.99 and I.100 rendering the lotus simile are in jagati meter. 

1827 have noted similarities between the verses in the RGV and the 7GS in the notes to my 
translation. A good example is found and discussed at the end of 4A. 


PDF Version: BPPB VI (2002) 

The simile in the 7GS stresses the identity of nature of the seed and the fully 
grown tree. The statement that the result is already perfectly contained in the seed 
in fact amounts to satkaryavada, a theory normally attacked by Buddhist 
philosophers as leading to eternalism. This could be the reason why the author of 
the verses in the Ratnagotravibhaga emphasized other aspects of the simile, 
namely the growth of the sprout itself, its gradual development (kramat, 
kramena)'® and the need for the presence of factors (pratyaya) enabling the 
sprout’s growth, such as “water, sunlight, air, soil, time and space” (I.115—117). 
The perfect buddha (sambuddha) is even compared to a sprout in the state of being 
a seed (bijankura) (1.117). With this new formulation the simile was propelled in 
the direction of the classical Buddhist doctrine of pratityasamutpada. The possible 
reproach of maintaining a position of eternalism—based on the explicit wording in 
the TGS—could thereby be forestalled, though the implications arising from this 
doctrine of a revelation of the buddha-nature would still be open to the same kind 
of criticism. However, we can say that the rewriters of the verses paid more 
attention to abhidharmic needs than their predecessors, who formulated their 
message in a philosophically less vigorous way. 

The similes of the TGS figure also in the commentary of the Ratnagotravibhaga, 
the Ratnagotravibhagavyakhya. Verses 1.95, I.127-152 and some passages in 
prose in (1}13) reflect this point. An intent to systematize seems to have been 
prevalent in this section of the commentary. It is obvious that this systematization 
is forced on the similes, both in the classification of nine kinds of klesas (matching 
the nine similes) under four groups corresponding to four kinds of living beings 
(67.2-69.15) and in the threefold classification of the buddha-element (dhdtu) in 
relation to the nine similes (69.16—-72.14). Regarding this latter grouping of the 
similes according to the three concepts dharmakaya, tathata and gotra, terms used 
in verse 1.27'* to explain the phrase sarvasattvas tathagatagarbhah of the TGS, at 
least the similes of the sprout turning into a tree and the cakravartin embryo in the 
womb of the ugly woman seem to be appropriate to the chosen category. Together 
with three other similes, they are subsumed under the concept gotra which, in the 
Ratnagotravibhaga, is understood as the cause (hetu) of buddhahood. This concept 
is inherent in the upamana of the seed/sprout and the embryonic ruler, entities 
which still have to undergo a process of further ripening in order to arrive at 
perfection. But whereas the inclusion of the eighth simile (the embryonic ruler) in 
the gotra category may have occurred without great reflection, 185 the same 
classification of the simile of the seed which turns into a tree was probably well 
thought out. I have noted above that the authors of the Ratnagotravibhagavyakhya 
emphasize the connotations of development in this simile. The same tendency has 
already been observed in the commentator’s explication of the phrase garbhagata 
in the section on gotra in 72.11f—a citation from the same simile in the TGS (see 
above). The term garbha in this compound was not intended in the TGS to mean 
“embryo” but “womb” or “inside,” the accepted sense for the recomposed verses 

183 The commentary verse 1.139, where the simile is recalled in a reduced form, also 
emphasizes the gradual growing of the sprout (-kramodayat). 

184 Verse 1.27 is, be it noted, part of the oldest layer. The threefold classification itself must 
therefore be an old element. 

185 The authors of the RGV(V) make no mention that the embryo still must grow. Gotra as the 
last of the three terms covers the last similes, of which the cakravartin embryo simile is one. 


PDF Version: BPPB VI (2002) 

of the old layer in the Ratnagotravibhaga. Nevertheless, the commentators of the 
Ratnagotravibhaga cited it in order to document that the term dhatu can be used in 
the sense of hetu. 

The threefold characterization of the buddhadhatu can hardly have played 
any role in the composition and arrangement of the similes in the TGS. Verse 1.27, 
part of the oldest layer of the Ratnagotravibhaga, demonstrates that the threefold 
classification was already in use at the time the different blocks of text from which 
the Ratnagotravibhaga was formed '®° were merged together. The idea of 
illustrating these three categories by way of the nine similes, however, is probably 
that of the later commentator(s). 

The abhidharma-oriented interest in the similes displayed by the 
composer(s) of the vyakhya led them to focus primarily on isolated elements in the 
upamanas corresponding to the klesas and the buddha-essence. 8’? As a 
consequence, in the commentary the similes are portrayed as lifeless and as if 
lacking their main points of emphasis; the rich connotations of the tertium 
comparationis are reduced to a simple enumeration of corresponding elements. 

The Ratnagotravibhaga(vyakhya) contains the only known word-for-word 
citations of the TGS among Indian texts and can without doubt be viewed as a 
work which dealt with the 7GS with a maximum of care. The similes in the TGS 
and their paraphrases in the Ratnagotravibhaga largely agree with one another. 
Thus we cannot decide whether passages in other texts which draw on one or more 
of the nine similes had the TGS itself or the Ratnagotravibhaga as their source. A 
short recounting of all nine similes is found in the Foxing lun (p\t3@,'* a 
philosophical treatise closely mirroring the structure of the Ratnagotravibhaga- 
(vyakhya). According to tradition, the author of the text was Vasubandhu. It was 
translated into Chinese by Paramartha, who is now believed to be the true author 
of the text.'®? The wording of the similes in the (#34 does not allow a final 
determination on whether the author utilized the TGS (TGS, or TGS) directly or 
based himself on the similes quoted in the Ratnagotravibhaga (TGS2). (He could, 
of course, have been acquainted with both the recensions and the similes in the 

Other texts sporadically refer directly to the TGS or use analogies or 
similes which could have been influenced by the 7GS. In the following I will cite 
several such passages without, however, striving for completeness. 

The Mahdparinirvanasutra explicitly refers to the TGS in one passage, and 
attributes the statement that the buddhadhatu is present in all living beings to it.’ 
There is no word-for-word correspondence with the TGS. However, the passage 

186 The RGV is composed of several blocks, of which the similes of the TGS make up one. See 
Schmithausen 1971: 128ff. 

187 This is particularly evident in the case of the eighth simile: the commentary verse 1.141 
compares the impurities to the “coverings of the embryo/in the form of a womb” (garbhakosa-). In 
neither of the two descriptions in the TGS and the RGV is a comparison between the embryonic 
coverings and the defilements drawn. 

'8 T 1610, vol. 31; the similes are found in 807c9-808al 3. 

189 For a discussion of this issue see e.g. Takemura 1977: 6f. 

19 oZhan yang ‘di na dge slong la la de bzhin gshegs pa’i snying po’i mdo sde chen po ston par 
byed do! // sems can thams cad la ni sangs rgyas kyi khams yod la’ khams de rang rang gi lus la 
‘chang’ ste /.... [' S: de;? S: la I;? Q: tshang] (MPNS Q 99a6; S 17524f.; quotations of the Chinese in 
Takasaki 1974: 137.) As mentioned above on p. 78, the passage in the commentary to verse [X.37 
of the Mahayanasitralamkara (reading sarvasattvas tathagatagarbha ity ucyate) could also derive 
from the TGS (0M, 1A or 1B). 


PDF Version: BPPB VI (2002) 

could be a free rendering of 8B.3-4, the only one in the TGS containing the term 
dhatu.'*' The fact that the TGS was known to the authors of the Mahdparinirvana- 
sutra also gives rise to the assumption that the simile of the hidden treasure in the 
Mahaparinirvanasitra,'”* though different in details, may have been inspired by 
the corresponding simile of the TGS. 

In another well-known work, the Lankavatarasatra, we find a passage’”’ 
that refers to “the text of the stitras” where the tathdgatagarbha is said to be 
described as “luminous by nature, bearing the thirty-two marks [of a Great Being], 
being inside the bodies of all sentient beings, ... like a gem of great value and price 
that is enwrapped in a dirty garment....” Though the meaning attributed to the 
compound tathagatagarbha is not in line with the TGS, the concrete depiction of 
the tathagatas inside living beings as found in the first simile of the TGS,’ and the 
allusion to the simile of the buddha statue wrapped in rotten rags, leave little doubt 
that the Lankavatarasutra is here referring to the TGS. 

In the Dacheng fajie wu chabie lun KFRE FARE 5G which is based on 
the (#6 (4:34, the veiling Alesas are compared to “not yet opened lotuses” and “filth” 
in which gold is found.’ The comparisons, though of a general nature, remind 
one strongly of the TGS. 

The *Bodhisattva-gocaropaya-visaya-vikurvana-nirdesa-sutra describes 
the relation between the fathagatagarbha and the defilements in terms of ten 
likenesses, among them “a sprout in a seed,” “a jewel in a treasure,” “a statue in its 
mold,” and “an embryo in the womb.” Here again all four comparisons are 
undoubtedly inspired by the TGS. 

Even more indebted to the TGS is a series of similes in the Mahamegha- 
sttra.'”’ Two of these similes closely resemble the illustrations in the TGS. The 
similes compare the state of not living according to the teachings and of not 
entering into a certain samadhi to (1) winter rice etc. which have not yet fulfilled 
their nature of benefiting living beings and (2) the fruit of a palmyra palm, a 
mango tree, cane etc. which has not yet become a tree. The similes are here used to 
illustrate that one can talk of such fulfillment in something or somebody only if 
certain necessary processes leading to the unfoldment of that nature have occurred. 
Though the aim of the author of the similes in this sutra is different from that of 
the TGS, the illustrations clearly echo basic elements of the latter. In view of the 
occurrence of a “fruit of a palmyra palm, a mango tree and cane,” we can be sure 
that the, author based himself on recension TGS, for only there do we find these 

19! For further details see note 204 in my translation. 

12 MPNS S 184b1ff.; Q 105b6ff.; see Takasaki 1974: 144f. for the Chinese. 

193 14S 77.13-78.1; see note 84 above. 

194 Tn Ch (0G) we find a very similar passage that mentions the thirty-two marks; see note 84 
above for details. 

'5 T 1626, vol. 31; for the comparisons see 893b23ff. 

1% T 272, vol. 9; see 359b2f.; there is no correspondence in the Tibetan text. 

17 T 387, vol. 12; (Dacheng deng wuxiang jing KFC): 1104b13ff.; Tibetan: Q 898, 
vol. 35, mDo sna tshogs, Dzu 121a4—-237a6 (’Phags pa sprin chen po shes bya ba theg pa chen 
po’t mdo): 219b7ff.; see also Takasaki 1974: 290f. 

188 Recension TGS, has only “a mango tree.” The rewording in the RGV speaks of “fruits of the 
palmyra palm and mangos” (RGV 1.116, 117). 


PDF Version: BPPB VI (2002) 

Two other siitras mention the TGS: The Mahayanist Angulimdliyasatra,! 
in asserting that all living beings have the tathagatagarbha, draws attention to the 
significance of the TGS and the need to appreciate it properly. The Da fa gu jing 
KEBLE (*Mahdabherihadrakasttra) states that it is only the bodhisattva- 
mahasattvas who understand the eternal character of the Tathagata and that, 
among other positive activities, they preserve the TGS.” Finally, in the Adhy- 
ardhasatika Prajfiaparamita the phrase sarvasattvas tathagatagarbhah appears 
without the mention of a source.7” 

From these references to the TGS, the great importance of this siitra for the genesis 
and development of the tathdgatagarbha strand of Buddhist thought becomes 
clear. The Ratnagotravibhaga(vyakhya), the main representative of the Indian 
tathagatagarbha theory, contains a large section dedicated to the illustrations 
found in the TGS. The 7GS thus affected the philosophical formulation of the key 
concept of this theory in a decisive way. Its title became synonymous with the 
basic doctrine that all living beings have the buddha-nature; its similes, as shown 
above, remained powerfully alive and were utilized, with certain modifications, 
not only to describe the relation between the buddha-essence and the sheaths of 
defilements but also to illustrate other relevant aspects of the tathagatagarbha 

4.6 The Twentieth-Century Reception of the TGS 

The modern history of scholarly works dealing with the TGS has been dominated 
by Japanese Buddhologists. In 1933 Tokiwa Daijd published a Japanese 
translation in the kundoku #)ll%é style of the recension of Buddhabhadra in the 
Kokuyaku Issaikyo. Probably inspired by the publication of the Sanskrit text of the 
Ratnagotravibhaga(vyakhya) in 1950 by Edward H. Johnston, which revealed the 
huge meaning the 7GS had for the development of the tathagatagarbha theory, 
Tod6 Kyoshun in 1959 collated the Tibetan texts found in the Kanjurs from Derge, 
Narthang and Peking.” Facing the Tibetan text, he arranged the two Chinese 
translations so as to make a close comparison of the versions possible. In an 
introduction, he dealt with the contents of the siitra, the doctrinal environment 
from which the TGS had presumably evolved, and entries in Chinese catalogues 
regarding Chinese translations of the TGS which are reported to be lost. 

The first translation from the Tibetan (Derge) was accomplished by 
Kagawa Takao in 1962. Many useful hints on related passages in other Mahayana 
texts are found in his annotations. In 1974, Takasaki Jikido’s monumental work on 
the formation of the tathagatagarbha theory appeared.” Thirty-two pages are 

199 T 120, vol. 2 (Yangjuemoluo jing HEPES): 539c15; Tibetan: Q 879, vol. 34, mDo sna 
tshogs, Tsu 133b2—215al (’Phags pa sor mo’i phreng ba la phan pa shes bya ba theg pa chen po’i 
mdo): 201b4., 

2® T 270, vol. 9: 295a6; see also Takasaki 1974: 235. 

20 See the reprinted text of the “Togand Edition” of the Ardhyardhasatika Prajfaparamita in 
Yukio Hatta (ed.), Index to the Arya-prajfid-paramita-naya-Satapaficasatika, Kyoto: Heirakuji, 
1971, p. 6.18 (198.18). 

202 The apparatus, however, is not always correct, and in some passages is ambiguous. 

203 Seven years before, in 1966, Takasaki had published an English translation of the RGV(V), 
which contains the reproduction of the nine similes of the TGS in the RGV. 


PDF Version: BPPB VI (2002) 

dedicated exclusively to the TGS, of which he translated several passages, 
characterizing it as one of the three basic receptacles of tathagatagarbha thought. 
He dealt extensively with the earliest stage of the buddha-nature theory as 
contained in the TGS, and concluded that the compound tathdgatagarbha in the 
siitra must be understood as a bahuvrihi. Eight years later, in 1981, Takasaki 
translated the whole TGS from the Tibetan version (Peking) into modern Japanese 
as part of the Daijd Butten series. His work also contains relevant observations 
concerning the textual history of the Tibetan version and the Chinese translation 
by Buddhabhadra, along with detailed annotations on the translated text. The 
translation deserves deep respect for having introduced the siitra to a broad 
Japanese public while still maintaining philologically high standards. 

In the West, David Seyfort Ruegg’s study and translation of the De bzhin 
gshegs pa’i snying po gsal zhing mdzes par byed pa’i rgyan by Bu ston made 
some parts of the original Tibetan text of the TGS known to the Western reader in 
French.*™ His rendering is based on the citations of the TGS by Bu ston and the 
canonical version in the Lhasa Kanjur. The focus of Seyfort Ruegg’s interest was 
the question how the Tibetan traditions dealt with the tathdgatagarbha doctrine, 
and in particular how they interpreted the formulation of it in the TGS. 

Only in 1995 was the first complete translation of the TGS directed at a 
more general public published in a Western language. It is contained in the volume 
called Buddhism in Practice. The translator, William Grosnick, based himself on 
the Chinese rendering by Buddhabhadra, and in an introduction described the main 
tenets of the siitra) Among modern studies of the TGS, the translation by 
Buddhabhadra has received notably more attention than the one by Amoghavajra. 
This may be due to the (erroneous) belief that the recension by Buddhabhadra was 
the earlier and more influential one.”°° Also worth mentioning is the rendering into 
Spanish of a passage from the Chinese translation of Amoghavajra (OF—-OH) and 
Buddhabhadra (OM-1B). This translation, by Fernando Tola and Carmen 
Dragonetti, also appeared in 1995, in the Revista de Estudios Budistas. 

Further, the TGS has been the subject of variously lengthy treatments, of 
which I will mention the most important ones. Among them, Matsumoto Shiro’s 
Zen shisd no hihanteki kenkyu (*A Critical Study of Zen Thought) BAO ROA 
AYHTIE (1994) is by far the most extensive and profound. Pages 411-543 contain a 
study of the TGS and TGS-related passages in other sitras. From an ample 
philological basis, he argues in a very transparent way that the introduction of the 
term tathagatagarbha into the TGS was closely connected with the word 
padmagarbha, and that the former should be interpreted as a tatpurusa compound. 
Moreover, he demonstrates that the Saddharmapundarikasutra had a crucial 
impact on the origin of the TGS. He also translates core passages of the TGS and 
deals with the different strands of the siitra’s transmission. 

Nakamura Zuiryt, who two years earlier had published a new edition of 
the Sanskrit text of the Ratnagotravibhaga(vyakhya) facing the Chinese 
translation, had already in 1963 pointed out some of the differences between the 
two available Chinese versions of the TGS and dealt with catalogue entries bearing 
on the lost translations. His conclusion that the verses of the TGS reproduced in 
the Ratnagotravibhaga must be based on the recension represented by the 

24 By ston cites passages from 1A—1B and 8B. 
205 See section 1.1, where I argue against this view. 


PDF Version: BPPB VI (2002) 

translation of Amoghavajra and the Tibetan (= TGS2) can now be fully confirmed. 
Another interesting, though somewhat biased contribution is that of Kariya 
Sadahiko (1979), who deals with the doctrinal relation between the Lotus Sutra 
and the TGS. I (1999) have briefly responded to his criticism of the TGS, which in 
his article he characterizes as being beholden to an intellectual approach. The 
“fundamental difference” which he claims to exist between the Lotus Sutra and the 
TGS (and which he associates with a step towards monastic dominance) requires 
further illustration and textual evidence in order to gain plausibility. Finally, there 
is a recent contribution by Kaneko Yoshio (1998) centering on the TGS. The 
author creatively analyzes the nine similes from a Jungian perspective and 
describes how they can be interpreted as documents of a “process of 


PDF Version: BPPB VI (2002) 

B_ An Annotated Translation of the Tathagatagarbhasiutra into English 

My translation of the Tathagatagarbhasutra is intended to mirror the Tibetan 
syntax closely and so to render the text as literally as possible. In many instances, 
however, where the English wording would have become incomprehensible or the 
sentences too long, I have taken the liberty of adapting the wording to a less 
Buddhist-Hybrid style. My own additions to the translations are placed in 
brackets, and I have usually added the Sanskrit underlying the Tibetan in 
parentheses whenever crucial technical terms appear or the Tibetan translation 
deviates considerably from the meaning of the conjectured Sanskrit equivalent. 
The reader should bear in mind that the main goal of this study is to throw light on 
the stitra in the form it existed in India. Therefore, though my translation follows 
the Tibetan text, I render many Tibetan terms based on an understanding of the 
background of their Sanskrit counterparts; that is, where doubts arise, I give 
preference to the TGS as it was possibly understood in India. To translate the stitra 
according to how it may have been read by Tibetans would, of course, be a 
different approach, one fruitful in its own right. 

This same guideline holds true for the translation of Tibetan verbal 
prefixes such as kun tu or rab tu into English. I have only tried to translate them 
explicitly into English when they do not simply function as formal elements (to 
render a Sanskrit prefix literally) but entail a meaningful modification of the 
verbal expression.’ 

English phrases in italics indicate that the phrases are embedded in a 
different way in the Tibetan sentence. However, I only refrained from following 
the- syntax of the Tibetan sentence if the Tibetan did not make any sense or 
obviously contradicted the main line of thought of the siitra. All passages of this 
kind are extensively discussed in the notes. 

Underlined terms are not found as such in the Tibetan text, but are the 
result of an emendation from my side not based on any of the textual 
representatives of Tib. 

' See the passage in the sGra sbyor concerning the translation of verbal prefixes into Tibetan in 
Simonsson 1957: 255f. 


PDF Version: BPPB VI (2002) 

[0 Frame story] 

[0A Title and invocation} 
In Indian: Arya-tathdgatagarbha-ndma-mahdydna-sitra. 
In Tibetan: The Holy Mahayana stitra called Tathdgatagarbha. 
Homage to all buddhas and bodhisattvas (namah sarvabuddhabodhisattvebhyah)!' 

[OB Setting} 
At one time I heard the following:” In the hot months? ten years’ after [his] 
complete awakening (abhisambuddha), the Exalted One (bhagavat) was staying 
on the mountain Vulture Peak (Grdhraktta) near Réjagrha, in the 
Candanagarbha pavilion (Atagara) of Ratnacchattra palace (prdsdda),’ together 
with a great community of monks (bhiksusamgha), fully a hundred thousand [in 
number]. The monks [were both] s7dvakas under training and [those] no [longer 
in] need of training;® almost all [of them were] honorable ones (arhat) [whose] 
contaminations were stopped (ksinasrava), [who were] free of defilements 
(nisklesa), [who] had attained mastery (vasibhita), [with] completely liberated 
minds and insight (suvimuktacitta, suvimuktaprajfia), of noble race (ajaneya), 
[powerful like] great elephants (mahdndga), [whose] duties were done 

' Bth lacks the invocation. It adds bam po dang po instead. Ch, start with OB after the 

2 Skt. evam mayd srutam ekasmin samaye bhagavan...: It seems that already among Indian 
commentators such as Buddhaghosa the two alternatives of combining ekasmin samaye (dus gcig 
na) either with evam maya Srutam or with bhagavan ... viharati sma were current (see Brough 
1950: 420). As all of the collated Tibetan texts of the TGS divide the sentence after dus gcig na /,1 
based my translation on the Tibetan punctuation. Fora comprehensive bibliography on this issue 
see Schoening 1995a: 200 and also Tola and Dragonetti 1999. 

3 Bth and Ch, do not mention the hot months (shin tu tsha ba’i dus kyi tshe). 

“ Bth (lo bcu bdun): “seventeen years...” 

> (1) The Vulture Peak near Rajagrha appears in many siitras as the place where the Buddha and 
his disciples gather, whereas the Candanagarbha pavilion seems to be found only in the TGS. I 
understand the compound *ratna-cchattra-prasdda-candana-garbha-kutdgara partly as an iti- 
compound (parallel to e.g. GV 434.14: dharmadhdtu-dik-samavasarana-garbham nama 
kiutagaram). The fact that -garbha is not found in Ch, does not necessarily qualify it as a later 
addition. Cf. e.g. DbhS A.7f.: maniratnagarbhe ... prasade where a rendering of garbha is lacking 
in Buddhabhadra’s translation (/#/E #4 |) but is found in the earlier translation by Dharmaraksa 
(QO RLY RL). The choice of garbha as part of the name of the pavilion in the TGS might 
have been influenced by the title of the stitra itself. But also the imagery of the lotuses rising into 
the sky from the pavilion (section 0G) qualifies the pavilion as a kind of womb (garbha) which 
gives birth to the flowers. Matsumoto (1994: 413, n. 5) stresses the analogy to the attribute 
candana-gandaha of the rising stiipa in the SP (239.1 ff.) by which the name candana-garbha could 
have been inspired (cf. Zimmermann 1999: 163, n. 44). 

(2) For kiitagara as “gabled mansion” or “roofed pavilion on any story of a palace” see Chandra 
1950: 177f., Coomaraswamy 1928: 262f., and Bollée 1986, 1989. 

(3) Ch, shows candra (FA) instead of chattra. Ch; lacks a rendering of prasdda, Apart from the 
location Rajagrha, Ch; also mentions Mygaramaty (£8. Rt). However, the Mygaramaty palace is not 
located in Rajagrha but in Sravasti (see MPPU, I 181, n. 3). 

§ (1) See the parallel SP 1.5-2.9: ... mahata bhiksusamghena sardham dvadasabhir bhiksuSataih 
... anyabhyam ca dvabhyam bhiksusahasrabhyam Ssaiksasaiksabhyam /: “... was accompanied by a 
great community of monks, 1200 monks [in number] ... and the two other [groups of] a thousand 
monks [each, i.e.,] the [monks] under training and [those] no [longer in] need of training.” Further 
also Kpun 1.5—2.14. Bth and Ch; only have a thousand instead of a hundred thousand of monks. 

(2) Ch; mentions only the great community of a hundred thousand monks without further 
differentiation and lacks the list of attributes which follows in TGS). 


PDF Version: BPPB VI (2002) 

(krtakrtya), [whose] tasks were performed (Aytakaraniya), [who had] laid down 
[their] burden (apahytabhara), [who had] reached their own goal 
(anupraptasvakartha), [in whom all] the fetters to existence were eliminated 
(pariksinabhavasamyojana), [whose] minds were completely liberated by 
perfect knowledge (samyagajfidsuvimuktacitta), and [who had] attained 
excellent supremacy in the control over the whole mind (sarvacetovasiparama- 

[OC Enumeration of the monks]® 

[Among] the fully hundred thousand monks were the venerable (ayusmat) 
Mahakasyapa, the venerable Uruvilvakasyapa, the venerable Nadikasyapa, the 
venerable Gayakasyapa, the venerable Mahakatyayana, the venerable Maha- 
kausthila, the venerable Vakula, the venerable Revata, the venerable Subhiiti, the 
venerable Pirnamaitrayaniputra, the venerable Vagisa, the venerable Sariputra, 
the venerable Mahamaudgalyayana, the venerable Ajiatakaundinya, the 
venerable Udayin, the venerable Rahula, the venerable Nanda, the venerable 
Upananda, the venerable Ananda and others.’ 

[OD Description of the bodhisattvas] 
Also accompanying him were bodhisattva-mahdsattvas'’® who had come together 
from various buddha-fields (buddhaksetra}—as many as the sands of sixty 

7 (1) This list of attributes of arhats appears with a few variations also in MVy 1075-1088; cf. 
also SP 1.6-1.9, Kpun 1.6-2.1 and Lamotte 1962: 98, n. 2; for interpretations of most of the 
attributes see MPPU,, I 203-219. Unusual in the TGS is phal cher. The above-mentioned sources 
show sarva instead (sarva also in Bth: kun). The use of phal cher is in line with the differentiation 
into Saiksas and asaiksas in the passage above. 

(2) Ch; lacks kytakaraniya (so does Bth) and samyagajnasuvimuktacitta. In some instances the 
Chinese shows an understanding different from the Skt.: ##/$]-) (“attaining of the right 
knowledge”) for ajaneya agrees with the Tibetan (cang shes pa) and is interpreted in this way in 
the *Mahdaprajfiaparamitasastra (MPPU, 1 211f.). See also BHSD s.v. djanya; #440 KEE (“like a 
great dragon”) for mahdanaga: also this interpretation is found in the *Mahdprajfidparamitasastra 
(MPPU, I 212f.; so also Bth: klu chen po); EU}A(k = (“arrived at the other shore”) seems to be an 
abbreviated rendering of sarvacetovasSiparamaparamiprapta. For the analysis of the term 
sarvacetovasiparamapdramiprapta as a dvandva see AAA 10.10—12 and also Bth. 

(3) According to my understanding, sha stag in 0B.11 and OD.7 emphasizes the 
comprehensiveness of the group concerned (“all and not less than all”), Sha stag can function as an 
equivalent for eva (see SpS 131: ... sarvany eva ... mahitani ca |! = de dag kun kyang ... bkur ba 
sha stag lags /; the following commentary is misleading), but is in most cases introduced by 
Tibetan translators without any Skt. correspondence: e.g. Sukh: sangs rgyas kyi zhing de’i byang 
chub sems dpa’ rnams ni ‘di lta bu sha stag ste | for idrsa ... tasmin buddhaksetre bodhisattva ... 
(p. 298); further Sukh pp. 274, 276. The reconstructed Skt. sarva for sha stag in the 
Aksayamatinirdesasutra (Braarvig 1993a: 12.8; reconstructed in Braarvig 1993b: 47.7) is not 
confirmed by Skt. sources. Further see Tshig mdzod 2823: sha stag: ‘ba’ zhig gam kho na /.... In 
combination with the term phal cher, I thus translated sha stag as “almost all [of them].” 

8 OC is not found in Chy. 

° Ch;-has Kausthila instead of Mahakausthila, lacks Pirnamaitrayaniputra and djfata- of 
Ajiiatakaundinya. Bth lacks Mahakatyayana and Upananda. Ananda is usually thought not to have 
attained arhatship (see e.g. SP 2.8 where he is‘attributed the title of a Saiksa whereas the arhats are 
called mahasravakas; ASPyy 8.22f.: ...ekam pudgalam sthapayitva yad ut’ Gyusmantam anandam |/; 
for the reasons why he is not regarded as an arhat see MPPU, I 222-225). 

© Mahasattva is a common epithet of bodhisattvas. As Kajiyama 1982 showed, the canonical 
meaning of the term “bodhisattva” can be identified as “[person] who clings to bodhi” (PA. satta: 
p.p. of the root sa7ij). Later the meaning of safij became mainly associated with the negative 
“clinging (to sensual desires)” and the understanding of sattva as “energy” became prevailing. The 


PDF Version: BPPB VI (2002) 

Ganges Rivers (sastiganganadivalukasama). [They were] all of them!! [only] 
one lifetime away [from perfect awakening] (ekajatipratibaddha)'? and had 
attained the [five] great supernatural faculties,’* the [ten] powers (bala) and the 
{four kinds of] self-assurance (vaisdradya), [had] venerated many myriads 
(kotiniyutaSatasahasra) of buddhas and had set in motion the wheel of the 
Dharma [which] never regresses (avaivartyadharmacakrapravartaka). It 
happened that sentient beings of immeasurable, innumerable world systems 
(lokadhatu)'* attained non-regression in [their striving after] supreme and perfect 
awakening (anuttarasamyaksambodhi) from hearing their names only.!° 

[OE Enumeration of the bodhisattvas]'° 

[Among them were] the bodhisattva-mahdsattvas Dharmamati, Simhamati, 
Vyaghramati, Arthamati, Ratnamati, *Pravaramati, Candraprabha, *Ratna- 
candraprabha, *Pirnacandraprabha, Mahavikramin, *Aprameyavikramin, 
Anantavikramin, Trailokyavikramin, *Acalapadavikramin, Mahasthamaprapta, 
Avalokitesvara, Gandhahastin, *Gandharati, *Gandharatisri, Srigarbha, Siirya- 
garbha, Ketu, Mahaketu, *Vimalaketu, *Anantaratnayasti, *Tyaktaratnayasti, 
*Vimalaratnayasti, Pramodyaraja, *Sadapramudita, Ratnapani, Gaganagafija, 
Meru, Sumeru, Mahameru, *Gunaratnaloka, Dharanisvararaja, Dharanimdhara, 
*Sarvasattvaroganivartana, *Pramodyamanas, *Khinnamanas, *Akhinna, 
*Jyotiskara, Candana, *Jhavivarta(na), *Aprameyabhigarjitasvara, *Bodhi- 
samutthapana, Amoghadarsin, Sarvadharmavasavartin, the bodhisattva- 
mahdasattva Maitreya, Manjusri as a young man (kumarabhita) and other 
bodhisattva-mahdsattvas, as many as the sands of sixty Ganges Rivers. 

[OF Other participants] 
Also accompanying [him] were an immeasurable [number of] divinities (deva), 
snake-gods, (tree) spirits, celestial musicians, demons, man-birds, man-horses, 

term bodhisattva was accordingly analyzed as “[person] whose energy [is directed towards] 
bodhi.” Parallel to this, the compound mahdsattva could be understood as “[person] of great 

a Yang in 0D.2 (thams cad kyang) expresses comprehensiveness (Skt. api): “all without 
exception.” This aspect is emphasized by sha stag in 0D.7 (see n. 7 (3) above). 

|, Literally: “hindered [from awakening by only] one birth.” 

13 For the five or six abhijfids see MPPU, IV 1809ff. 

4 Tshad med grangs med pa (apramanasamkhyeya) could also refer to sems can. 

5 (1) Differences in the Chinese versions: Ch,;: Instead of “various buddha-fields”: 
“innumerable buddha-fields” (#8 (#541); missing are the characterizations of the bodhisattvas as 
ekajatipratibaddha and vaisaradyaprapta; the elements abhijfd and bala are rendered by Hx) 
(viryabala), one of the pafica balani (see MVy 983-987; MPPU, III 1127; 1200); missing are also 
the “immeasurable, innumerable world systems”; for anuttarasamyaksambodhi: #& +38. 

According to Ch), sentient beings attained non-regression just by praising their names (#87844). 
However, the same matter of attaining non-regression by hearing names is also attested for the 
Tibetan version of the RCaP (40-42): ... rigs kyi bu ‘am rigs kyi bu mo yang rung de bzhin gshegs 
pa dgra bcom pa yang dag par rdzogs pa’i sang rgyas ... de dag gi mtshan thos na bla na med pa 
yang dag par rdzogs pa’i byang chub las phyir mi ldog par ‘gyur ro //. 

(2) The term avaivartyadharmacakrapravartaka or avaivartika® is typical for the SP (see BHSD 
s.V. avaivartya, °tika and SP)). Its interpretation, however, is not at all clear. Does avaivartya refer 
to dharma, to cakra, or to both of them? Or should we assume a “wheel of the Dharma for the 
avaivartyas,” or even a “wheel of the Dharma by which one becomes avaivartya’”? 

'6 For a detailed analysis of the bodhisattva names in the four translations and the meaning of 
the asterisk see appendix A. In Ch,; the designation bodhisattva is inserted after each name. 


PDF Version: BPPB VI (2002) 

serpent-beings,.” human beings (manusya) and [further] non-human beings 

Then, after the Exalted One had been surrounded (parivrta) and honored 
(puraskyta) by many hundreds of thousands of assemblies,'® [he] was honored, 
venerated, worshiped and revered by kings, chief ministers (mahdmdtya), guild 
leaders (sresthin), noblemen (grhapati), ministers (amdtya), citizens (naigama) 
and country folk (janapada)."° 

[0G The manifestation of supernatural phenomena] 

At that time, after having been served food, the Exalted One withdrew for 
meditation in that same Candanagarbha pavilion,” whereupon through the 
power of the Buddha appeared myriads of lotuses [coming out] from the 
Candanagarbha pavilion, with myriads of petals, as large as the wheels of carts 
(Sakatacakrapramana), colorful and not [yet] open.”! The [lotuses] then rose 
into the sky,” covered this whole (sarvavat) buddha-field, and remained” 
[there] like a jewel canopy (ratnavitana). 

The legendary beings mentioned are very common in Indian literature: ndga, yaksa, 
gandharva, asura, garuda, kimnara, mahoraga. 

'8 1 iterally: “... was then surrounded (parivyta) and placed in front (purashyta)....” 

'? (1) My translation “noblemen” for grhapati (khyim bdag) is inferred from the common 
combination of grhapati with Sresthin (see e.g. RGVV 47.17). Further see BHSD: “capitalist, guild- 
leader” and PTSD: “wealthy Noble” for grhapati. 

(2) bsTi stang du byas / bla mar byas / rim gror byas shing mchod do represent more or less 
synonymous expressions. Cf. e.g. SP 5.7f.: satkyto gurukyto manitah pijito 'rcito. 

(3) Ch, lacks the enumeration of the worshipers in the second part of OF. Among the first group, 
human beings and amanusyas are missing. Chz adds before the first enumeration: fEH tH FL: “in 
the immeasurable [many] world systems”; for “citizens and country-people”: 24 \ 3 (Bth: grong 
‘dal’ dang : sems can rnams [' for ‘dab]). 

0 For nang du yang dag (par) ‘jog, Skt. pratisarlayana, see MVy 1488 s.v. dhyanani, MVy 
1642 s.v. yogangani and BHSD s.v. pratisarhlayana as well as the entries before and after; for 
nang du yang dag ‘jog la zhugs pa (for p.p. pratisamlina) see SP 182.1—-4: ... bhagavan ... viharam 
pravistah pratisarhlayanaya tatha pratisamlinas ca ... viharasthita evasit' // [O 175b4-7: ' asthasit 
sakydbaddhena paryamkena for evasit] = SP, 80a3-a5S: ... bcom Idan ‘das ... nang du yang dag 
"jog la "jug pa’i phyir gtsug lag khang du zhugs so //... ‘di ltar/... gtsug lag khang du gnas te 
nang du yang dag ‘jog la zhugs so //. 1 assume that none of the representatives of Tib has 
transmitted the original wording as they have not taken nang du as part of pratisamlayana. 
Whereas BJNQP,2; have replaced ... de nyid du nang du ... by ... de nyid kyi nang du... and thus 
understood nang du as “inside” of the pavilion, LST have avoided the sequence du nang du yang 
(thinking of it as a slightly emended dittography?) by dropping du nang. Ch,: REFILLS 
BATFAS=BR, Mee tHat., : “At that time the Exalted One was sitting on the Candana pavilion 
(kuta@gara), upright in absorption, and brought forth supernatural phenomena.” Ch): HFS iA 
AEH, BZ. APD. BOR REARS TSE. : (lit.:) 
“At that time, after the meal, the Exalted One on the great Candanagarbha pavilion entered the 
[state of having the] supernatural powers of a Buddha. And because of this, suddenly myriads of 
lotuses streamed out from the Candanagarbha [pavilion].“ Alone Bth does not render 
pratisamlayana or have something similar. 

2) Chy and Bth also mention the scent of the flowers. 

22 See BHSD s.v. abhyudgacchati; SP 241.15: ... abhyudgamyopary antarikse vaihayasam ... = 
SP, 104a2: ... mngon par byung ste / steng gi nam mkha’ bar snang la.... 

23 Kun tu gnas par gyur to might be a translation of (samantdt) (pari)samsthita abhivan, 
(samantat) (pari)samtisthante sma or something similar; cf. SP 390.10: ... vaihayase ‘ntarikse 
Samantan mahdpuspavitanam parisamsthitam abhut // = SP, 166a5f.: ... steng gi nam mkha’ bar 
snang la me tog gi bla re chen pos kun du khebs par gyur to // [see also Bth: kun tu g.yog' ba...;' 
for g.yo]. Kun tu gnas pa could also represent a form of the root samava-stha (in this sense 
Matsumoto 1994: n. (8) to p. 413). The perfect passive participle (PPP) in combination with a form 


PDF Version: BPPB VI (2002) 

In each calyx of the lotuses (padmagarbha) was seated, cross-legged 
(paryanka), the body of a tathagata, emitting hundreds of thousands of rays of 
light and visible everywhere.“ And all the lotuses were opened up in blossom.” 

[OH The fading of the lotuses] 
Then, by the supernatural power (adhisthana) of the Buddha all the petals of the 
lotuses, without exception, became dark, deep-black, putrid (durgandha) and 
disgusting, and no [longer] pleasing. But in the calyxes of the lotuses the bodies 
of the tathagatas sitting cross-legged and emitting hundreds of thousands of rays 
of light were [still] visible everywhere. 

Further,” this whole buddha-field became filled with the [rays of light 
from] the bodies of the tathagatas sitting in the calyxes of the lotuses. This 
buddha-field became extremely beautiful [during] that time. Then at that time 
the whole?’ multitude of bodhisattvas and the four assemblies were extremely 
astonished (aScaryaprapta) and filled with pleasurable excitement (audbilya- 

of the verb bhavati indicates that the action expressed by the PPP has occured and that its result 

4 The Skt. for kun tu snang ba could have been sardpsyate. Almost identical parallels appear 
further down in 0H.3—5, 01.4—6 and 0J.10-11. By replacing snang ba by gda’ ba (Ja: “‘to be, to be 
there ...”) in 0J.11 (Bth here again snang ba!) Tib supports the understanding of snang ba as in my 
translation and not in the sense of “to illuminate” (cp. Matsumoto 1994: n. (10) to p. 413). In the 
passage here Bth employs /dang instead of snang. If this does not simply reflect a misreading of an 
eventual dbu med spelling, it could also represent Skt. pradurbhavati. 

5 (1) In order to describe the arising of circumstances in the past which last continuously, Tib 
employs a verb (preferably in its perfect form) and the perfect stem of the auxiliary verb ‘gyur: 
0G.2 (zhugs par gyur), 0G.6 (byung bar gyur), 0G.7f. (khebs par gyur), OG.8f. (kun tu gnas par 
gyur), 0G.11 (kha bye bar gyur), 0H.3 (mngon par dga’ bar ’gyur ba ma yin par gyur), OH.8 (shin 
tu mdzes par gyur), OH.9f. (shin tu ngo mtshar du gyur cing dga’ bar gyur), etc. This is probably 
influenced by the Skt. structure of a perfect passive participle together with an aorist of bhavati 
(see n. 23). 

(2) In Ch, it is not explicitly stated that the lotuses come out from the pavilion. The flowers are 
only said to have a thousand petals (so also SP 261.2: sahasrapattre padme 
Sakatacakrapramdnamatre ... and GV 408.9: ... sahasrapattram ...-padmam ...); instead of 
kotiniyutasatasahasra (Tib, Ch2) for the number of flowers: 8x#€#4; instead of “buddha-fields”: 
{tF; the tathagatas are not described as being in the paryanka-position; for the hundreds of 
thousands (of rays of light): 4&4; the lotuses are said to blossom at the same time ({AJEF S748). 

Ch: in the beginning the lotuses are not characterized as not yet open. The flowers do not cover 
the whole but ail buddha-fields. The flowers in the sky are further described as “being united with 
each other as a whole” (#84AX). Chz describes the tathagatas in the flowers after mentioning 
the opening of the lotuses. This order seems more natural then the one given in the Tibetan. 
Further, the tathigatas are said to be endowed with the dvatrimsanmahdpurusalaksanas (for the 
dvatrimsanmahapurusalaksanas see MVy 235-267; MPPU, 1.271ff.; RGVV III.17-25). 

Instead of kha ma bye ba (lit: “[the lotuses with their] mouths not opened”) Bth reads: snying po 
la ‘dus: “[with petals] contracted inside (garbha)”). Instead of ratnavitana Ch; and Bth are based 
on ratnavimana (= RR; rin po che'i gzhal med khang): “jewel palace.” The reading -vitdna 
(Ch,, Tib) is probably original and is attested in SP 112.10. 

*6 The particle kyang could correspond to Skt. punar, which led to my rendering with “further.” 

27 Just as the confusion between sarva and sarvavat documented at the end of the following n. 
suggests, it might also here be more plausible to assume an original sarvavan bodhisattvaganah: 
“the whole multitude of bodhisattvas.” See SP; s.v. bodhisattva-gana. 

28 (1) The four assemblies (catasrah parsadah), as is well known, comprise bhiksus, bhiksunis, 
upasakas and upasikas (see MPPU, I 232ff.). 

(2) Bth (ngo mtshar thob cing rmad du gyur pa thob par gyur nas :) suggests adbhuta® instead of 
audbilya°. Since the SP only has the combination ascaryaprapta (or °bhuta) — adbhutaprapta 


PDF Version: BPPB VI (2002) 

[01 Doubts of the onlookers] 
[But] after seeing that supernatural display (rddhyabhisamskara) of the Exalted 
One, [they] became uncertain (samsayaprapta) [and questioned themselves]: 
“What is the reason”’ that the petals of all these myriads of lotuses became so 

unsightly (durvarna), and that the[ir] stalks (ndla) too became unsightly, 
disgusting and not pleasing, whereas in the calyxes of the lotuses each body 
of the tathagatas is [still] sitting cross-legged, and in that [they] emit hundreds 
of thousands of rays of light are visible everywhere as [something] extremely 

Thereupon, [the Exalted One] motioned (nimitta) to the entire multitude of 

bodhisattvas and the four assemblies who had become uncertain to come 

closer? At that time there was a [certain] bodhisattva-mahdsattva named 

Vajramati [who had also] gathered [with the others] in the Candanagarbha 


[OJ Questions of the bodhisattva Vajramati] 
Then the Exalted One said to the bodhisattva-mahdsattva Vajramati:*” 

“Son of good family (kulaputra), venture to question the Tathagata, the 
Honorable One and Perfectly Awakened One,*? with reference to an 
exposition on the Dharma!”** 

At the Exalted One’s permission, the bodhisattva-mahdsattva Vajramati, 
realizing that the world with [its] gods, humans and demons and all the 

(instead of audbilya°) when the formula contains only two links, Bth should here represent the 
original reading (see SP; s.v. aScaryaprapta, °bhuta). 

(3) Ch, and Ch, are speaking of the withering in an instant. Ch; attributes the withering to the 
flowers; the petals remain unmentioned (cf. Matsumoto 1994: 417). Ch, also does not mention that 
the buddha-fields become filled with the rays of light. The bodhisattvas and the four assemblies 
are expressed by —t) A3R. Ch; continuously speaks of “all” buddha-fields. 

?° The terms rgyu (hetu) and rkyen (pratyaya) function here as quasi-synonyms (see BHSD s.v. 
pratyaya (1)). Therefore my simplified rendering as “reason.” 

3° The addition “the Exalted One” is based on Bth. ‘Dong ba (JNLSTP;) or rather nye bar ‘ong 
ba (Bth) could translate upa(sam)kramana or upasthana. The sentence is missing in Ch, . 

5! Further discrepancies with Ch, and Ch: Ch;: It is not mentioned that the supernatural display 
is seen, The withering, as in Ch , takes place in an instant. Ch, does not mention petals and stalks 
(so also Chz) and does not have the second part of the question (“..., whereas in the calyxes ...”). 
The last sentence is completely missing. In the question Ch, adds that the emitted light made the 
people rejoice (45 A). At the end, Ch; describes Vajramati and the great multitudes as sitting 
in the pavilion with feelings of respect. 

32 Ch, adds: RRR tH ASF EASRATEE....: “At that time the Exalted One understood what 
the great multitude of bodhisattvas had become uncertain about...” 

33 De bzhin gshegs pa dgra bcom pa yang dag par rdzogs pa’i sangs rgyas are the first three 
members (Skt. tathdgato ‘rhan samyaksambuddhah) of a standard set of names applied to 
buddhas. A traditional analysis of the whole formula can be found in MPPU,, 1.126-144. The 
Tibetan rendering of arhat with dgra bcom pa (“the one who has overcome the enemies (of the 
mind)” interprets arhat in the sense of ari-han (cf. MPPU, 1.127). 

34 The Skt. on which chos can gyi gtam las rtsoms te / ... yongs su zhu bar spobs par gyis shig // 
is based, could be *dharmydm/dharmim katham arabhya ... pariprcchda pratibhatu. In the Skt. 
pratibhatu expresses a request like “Think of a question with reference to...!” Ch): PABA 
SR ATRERRTZ Ati. _: “[Put forward] without reserve what you [want] to ask [with reference to] 
all doubts concerning the Dharma of the Buddha!” For dharmya see BHSD s.v. dharma 3 and 
SWTF s.v. kathd 2. For dharmya katha: Chy: @6¥&; Ch?: BEREB. 


PDF Version: BPPB VI (2002) 

bodhisattvas and the four assemblies, were anxious with doubts, then asked him 
the following:** 

“Exalted One, what is the reason this entire world system is covered with these 
myriads of such unsightly and putrid lotuses, yet in their centers sit cross- 
legged bodies of tathdagatas emitting hundreds of thousands of rays of light 
and visible everywhere, and now myriads of living beings,”° seeing the bodies 
of the tathagatas, raise their joined palms (kytdfijali) in homage?”>*” 

Then at that time the bodhisattva Vajramati uttered these verses:° 
[0.1] “Myriads*® of buddhas are seated motionless in the center of lotuses: 
[with] such supernatural powers (rddhi) you display [them]. Never 
before have I seen [anything like] this!”*° 
[0.2] “The [sight] of the leaders (nayaka) emitting thousands of rays of light, 
covering this entire buddha-field [with their splendor, and] wonderfully 
displaying a facile mastery of the dharmas, [is] constantly beautiful.”*' 


> Instead of enumerating the single groups by name, Ch, speaks simply of “all the great 
multitudes”: 247A 3%/st. Ch, does not mention the asuras; in Bth the bodhisattvas are missing. A 
parallel to zug rngu “to be anxious (with doubts)” does not appear in Ch,7. Also the mentioning of 
the Exalted One’s permission (gnang pas) does not appear in Ch, and Bth, but is made explicit by 
Chy: ... FREES: “... [and Vajramati] was granted the holy order of the Buddha.” 

36 Pranikotiniyutasatasahasra; Ch,: 3¢€77&; Ch) erroneously attributes kotiniyutasata- 
sahasra to the tathagatas in the flowers, characterizing them as sitting in afjali-position “exalted” 
({3X) and motionless. That in the Skt. text the subject should be the living beings and not the 
tathagatas, is proved by the parallel in 0M where Ch, is in accordance with the other translations. 

7 Further divergences with Ch,, Ch; and Bth: Ch, repeats in the question that the flowers rose 
to the sky and describes them as “innumerable.” Like Ch; also Ch, speaks of their withering “in 
one moment.” Ch; do not mention the putrid smell of the flowers. In Ch, the paryanka-position of 
the tathagatas is missing. 

°8 Ch, and Bth add the attribute mahdsattva to Vajramati. 

3° With “myriads” I not only render kofiniyutasatasahasra but also smaller numbers in the 
verses like kofisahasra in this case and in 10.4, 11.1, 8.6, kotisatasahasra (OM, 2A) or even kofi in 
2.2 and 2.3. See also n. 57. 

“ Ch, does not describe the buddhas as “motionless.” For the “center of the lotuses” (dbus): 
37698. Bth uses stobs instead of ston (rdzu ‘phrul stobs; rddhibala). However, bala is not found 
in any of the other versions. 

“l (1) 1 understand mdzad in padas a and b as emphasizing the autonomous aspect of the acts of 
the “leaders.” If mdzad is understood as a causative auxiliary in the sense of “the leaders causing 
[the tathagatas in the flowers] to emit...” the leaders would represent the buddhas generally and 
among them the Buddha on the Vulture Peak. However, this second alternative is contrary to the 
Chinese versions. 

(2) Ngo mtshar chos kyi rnams la rol mdzad pa: The adverbial rendering of ngo mtshar (aScarya) 
follows Ch, (see below). With the dharmas most probably supernatural qualities like the 
absorptions, the yddhis, abhijnas, etc. are meant. For similar expressions with rol (Bth: rnam par 
rol; vikridati) see BHSD s.v. -vikridana(-ta) and vikridita. 

(3) The text in 0.2d is problematic. As a point of reference for the genitive, the subject “sight” 
had to be supplied. Variant OK, n. 16 (kyis) allows the following translation similar to Ch,: 
“Through the leaders ... [the whole world?] is constantly beautiful.” However, a text in accordance 
with Ch (see below) is found in Bth: ‘dren pa rnams ni rtag tu +++! po’? // [! probably bzang; ” 
from po yin > po yi > po'i? Or parallel to the genitive Ayi in Tib?]. 

(4) The second half of the verse in Ch;: Beis soy OE PSE HER: “Flawless! [are] the leaders, 
decorated [are] the worlds.” {' Probably nirangana or nirafjana (fits) instead of nirantara (bar 
chad ma mchis)}. Chz SSPE M RGR AAAS RR: “Those buddhas, showing in a 
wonderful way mastery in the dharmas, are beautiful.” 


PDF Version: BPPB VI (2002) 

[0.3] “There, in the center of unsightly lotuses [with] disgusting petals and 
stalks, sit [tathdgatas], as if they [had] the nature of a jewel 
(*ratnasvabhava).” Why have [you with your] supernatural powers 
(rddhi) created these [manifestations]?””* 

[0.4] “I see buddhas equal in number to the sands of the Ganges River, [and] I 
see the exquisite (*visista) [manifestations] of the [Tathagata’s] 
supernatural powers. Never before have [I] witnessed such a miracle 
(vikurvita) like this one existing right now.”“ 

[0.5] “[I] implore the Highest among Humans (dvipadottama), the Divine 
(deva), to teach. [I] implore fhim] to explain the reason [for this 
miraculous display]. [I] implore [him] to speak [with] solicitude 
(anukampa) [in order to] benefit the world. [I] implore [him] to remove 
the doubts of all embodied [beings] (dehin).”” 

[OL Introduction to the first simile] 
Then the Exalted One said to the whole multitude of bodhisattvas, [including] 
the bodhisattva-mahdasattva Vajramati and others: 

“Sons of good family, there is a sutra of great extent (vaipulya) called 
Tathagatagarbha.”* In order to teach it the Tathagata has produced these 
signs’’ [which] appeared [to you]. Listen therefore closely, be attentive, and 
[I] will teach [you].”“8 

42 None of the other translations confirm 7ib in this passage. In Ch, the comparison does not 
appear at all. Ch, compares the tathdgatas in the lotuses with the brilliant manifestation of a 
wonderful jewel (410) TMABER). Bth also deviates: de dag ams kyang de la dga’ ba! ste : ['\dga’ 
ba for Skt. rata?: rata is probably a variant reading of ratna]. 

“ Ch, speaks of the sudden withering of the flowers. Petals and stalks (stalks are also missing 
in Ch2) are not mentioned. 

“ (1) The second ‘di of pada d is not rendered in my translation. Could it result from ni as it is 
found in Bth (pada b)? But see the similar formulation ‘di lta bu ‘di in OL.5 (Bth: de Ita bu)! 

(2) In Ch, visista does not appear; pidas c and d (of 7ib) are abbreviated to a single pada, and 
verse 0.5 (of 7ib) is shortly expressed in pada d (FRER275/3R): “[I] ask for a discerning 
explanation for [me/us] (f§ short for F3F&?)!” Ch, with 388%, does not translate vikurvana or 
vikurvita but vikridita. Their meanings are synonymous (see BHSD s.v.). In Bth pada b (of Tib) is 

4 (1) Tib in 0.5c could also be translated as: “[I] implore [the one who] benefits the world to 
speak [from] solicitude...!” Thugs brtse (Bth: thug brtse phyi ru) probably reflects a form of Skt. 
anukampa as found in e.g. SP 41.1, 41.11, 147.4 (© V1I.10c) and in the ASP (see ASP; s.v. 
anukampayai). For the meaning “solicitude” of anukampd see Maithrimurthi 1999: 118 ff. 

(2) Ch: “[I] kindly ask the One Worshiped by the Gods (Xr) to expose the reason for [this] 
manifestation! [I] kindly ask to practice compassion [for] the benefit of the world in order to 
remove the doubts of all [beings]!” 

(3) Bth corresponds in syntax with Ch; in the first half of the verse, but does not mention the 
gods. With phyi ru (metri causa for phyir) in Bth 0.5Sc the reasons for teaching are stated to be: “for 
the benefit of the world and from solicitude.” In pada d Bth does not mention the removal of 
doubts but instead speaks of the establishing (‘god) of all living beings (in the bodhi?). 

© Instead of “siitra” Ch, reads #LYEYKES (“very profound Dharma discourse (dharma- 

‘7 (1) The term nimitta (mtshan ma; Bth: Itas [emended for lhas]) is discussed in Schmithausen 
1969a: n. 67. Apart from “appearance” it can also mean “symptom, omen, sign” (clearer: 
piirvanimitta), as is the case in SP 233.1ff.: while digging for water on a dry piece of land, the 
discovery of wet earth and the spots of mud on the body of the digging people function as 
pirvanimittas for the immediate appearance of water. The appearance of rays of lights as (purva-) 
nimittas before the teaching of a siitra are common in Buddhist literature (see e.g. SP 6.6f£.; TUSN 
S 105b1ff. = Q 77b5ff.). See also DN I 220 where rays of light are a sign for the appearance of the 


PDF Version: BPPB VI (2002) 

“Just so!” replied the bodhisattva-mahdsattva Vajramati and the whole multitude 

of bodhisattvas to the Exalted One, thereby acquiescing, and the Exalted One 


“Sons of good family, just as these unsightly, putrid, disgusting and no 
[longer] pleasing lotuses, supernaturally created by the Tathagata,’ and the 
pleasing and beautiful form”! of a tathagata sitting cross-legged in [each of] 

god Brahman. The emitting of light through the Buddha is discussed in MPPU, 1.431ff. For 
occurrences of other kinds of piirvanimittas see Lamotte 1962: 335. 

(2) sNang ba’i mtshan ma in the TGS could also represent avabhasanimitta, “light [emitting] 
signs.” However, neither Ch, (AAMT) nor Ch, GERI BIE 48) seem to support such an 
interpretation. In the passage above, the nimittas are the supernatural manifestations of the 
beautiful and then withered lotuses with the tathagatas in their centers. 

“8 In Ch, the Tathagata does not announce that he is now going to teach. 

(1) For ltar nyan pa, Skt. pratispnoti, the pW gives, next to others, the two following 
meanings: ‘‘1) hinhdren, sein Ohr leihen ... — Jmd. (Gen. ...) ... — 2) bejahen, zusagen, 
versprechen; mit Acc. der Sache und Dat. oder Gen. der Person.” In this passage the second 
meaning (which is the only one found in P4.) fits the context. The identical formulation appears in 
e.g. Kpun, Q 151a6f. (rin po che rnam par snang byed kyis bcom Idan ‘das de bzhin no zhes gsol te 
bcom Idan ’das kyi ltar nyan pa dang / bcom Idan ‘das kyis byang chub sems dpa’ rin po che 
rnam par snong byed la ‘di skad ces bka’ stsal to //) with its corresponding Skt.: sadhu sadhu 
bhagavann iti Ratnavairocano bhagavatah pratyasrausit / bhagavan Ratnavairocanam 
bodhisattvam mahasattvam etad avocat | (Kpun 7.13ff.; for pratisynoti with the genitive case see 
also RGV 1.90a: ... tasya pratisrutya). It was probably not at all easy for later Tibetans to 
understand the genitive, based on the Skt., as well as /tar as part of the verbal compound and the 
special meaning of Itar nyan. This is documented by the redactional variants of P;. The above- 
cited passage of the Kpun is introduced with a wording nearly identical to the TGS. The Skt. runs: 
Synu kulaputra sadhu ca susthu ca manasikuru bhasisye ’ham te / (Kpun 7.12f.); see also Kpun 

(2) The passage beom Idan ‘das kyi Itar nyan pa dang is not found in Bth. However, the position 
of the syntactical units in Bth agrees (in contrast to Tib) closely with the Skt. Bth reflects also the 
vocative bhagavan (dge’o : beom Idan ‘das shes ...) which Tib constructs as a prepositional object 
(bcom Idan ‘das la legs so zhes ...). 

(3) In the last sentence Bth and Chz add the epithet mahdsattva to the bodhisattvas. Ch, simplifies 
the last passage: KB. ak, FRSA. HS. ....: “All said: ‘Just so! [We] want to listen!’ 
The Buddha said:....” So also Chz: 24%, tt#&®, FREQ. HE. .... 

°° Ch, shortens: SO(FAT(LARBSIE AR SEM....: “Just as innumerable lotuses supernaturally 
manifested by the Buddha, withered in an instant,....” 

Ch, also speaks of the withering in an instant (9472 fi). 

5! (1) In contrast to the chapters before, 7ib speaks here of the forms (gzugs; *rijpa?) of the 
tathagatas (but Bth: sku as before; Ch, ({#7%) could also stand for riijpa; Ch,: {{#). Unlike 
alternative expressions for “tbody” like aziga, kaya, vapus or Sarira, riipa stresses the phenomenal 
aspect of the body, i.e., its shape and physical appearance. 

(2) The triplet mdzes pa / gzugs bzang ba / blta na sdug pa can also be found in Pa. canonical 
literature (abhirupa dassaniya pasadika). It expresses pleasing and beautiful appearance. See 
PTSD s.v. abhirtipa, dassaniya: e.g. referring to the night (DN 1 47), to a brahman (DNI 114), toa 
woman (AN II 203); SP: prasddiko darganiyo ’bhirupah (425.13) = mdzes pa bita na sdug pa 
gzugs bzang ba (Q 181b6); abhiripah prasadiko darsaniyah (441.14) = gzugs bzang ba mdzes pa 
blita na sdug pa (Q 189a1). In the translation I have restricted myself to the rendering of only two 
main aspects of this triplet. 

Ch, mentions innumerable: and’ supernaturally manifested buddhas in the lotus calyxes with 
beautiful, splendid appearance (ARA(CORTESTEAN), BATTER). Cho: METEOR, fem 
BRA‘, PS POCHS, Fae, ARTSSEL. ....: “[and] just as in the flowers forms of buddhas 
are manifested..., with beautiful and dignified appearance, [which is] what humans desire to 
see/take joy in seeing...” 


PDF Version: BPPB VI (2002) 

the calyxes of these lotuses, emitting hundreds of thousands of rays of 
light,” [are such that when they are] recognized by gods and humans, [these 
latter] then pay homage and also show reverence [to them], in the same 
way, sons of good family, also the Tathagata, the Honorable One and 
Perfectly Awakened One,” [perceives] with his insight (prajfid), knowledge 
(jfidna) and tathagata-vision® that all the various sentient beings”® are 

52 (1) Ch, has simply: KIA. 

(2) There are two alternative ways to explain the reading ‘khod pa in 0M.5.’Khod pa could 
function as an auxiliary verb stressing the aspect of duration of ‘gyed pa in the sense of “emitting 
continuously....” The second, and in my eyes more plausible, alternative is that ’khod pa was first 
associated with the triplet mdzes pa / gzugs bzang ba / bita na sdug pa (“sitting [there with] a 
pleasing and beautiful shape”). This is how ‘khod pa, which is rendered as bzhugs pa in Bth, 
appears in Bth (bzang zhing sdug la gzugs dang Idan par bzhugs par). See also the parallel syntax 
in Tib 0M.13f. where the adverbial mi g.yo bar is associated with the following ’khod pa. It is 
therefore possible, that the original position of the triple adjectives before “khod pa was only 
altered in a second step of revision whereas ’khod pa remained there. That the emission of light 
was originally mentioned before the description of the beauty of the buddha forms is also 
confirmed by Ch. 

53 (1) Apart from gods and humans Ch, also mentions dragons (for ndgas), yaksas, gandharvas, 
asuras, garudas, kimnaras, mahoragas, amanusyas “and others” (see n. 17). Ch; abbreviates to 3%. 
(2) mChod pa’i las ... byed pa could mirror a figura etymologica in the Skt.: piijakriyam karoti or 
°karam karoti (for kdram karoti see BHSD s.v. kara’). The counterpart of phyag ‘tshal zhing / 
mchod pa’i las kyang byed pa in Ch;: #4; Ch: FEF. The last part (from “[are such when 
they are]....”) in Ch,: RBA, BEARAAL: “..., [are such that when] the multitudes see [them, 
they] are astonished (dscarya, adbhuta) [and] there is none who does not pay homage.” 

** Instead of “Tathagata, Honourable One and Perfectly Awakened One” Ch, reads #8. 

5 The Skt. probably had a compound like *svaprajfajfanatathagatacaksus (so also Bth) and 
could also be analyzed as a karmadharaya in the sense of “his tathagata vision [consisting in] 
insight and knowledge.” See RGVV 59.2: asangapratihataprajfdjidnadarsanam ... (RGVV, 
without -prajrid-). The reading DA#6 C7 3S-C45ER ... of Chz is probably based on Skt. -prabha- 
(65) instead of -prajfia-. The appearance of prajfa here is surprising, as this term is only found 
within the TGS in one other passage as one of the attributes of the arhats at the beginning (shes 
rab shin tu rnam par grol ba, suvimuktaprajfta; 0B.7). However, also the assumption that the 
correct reading was -prabha- is difficult to prove, as I have not found any such compound in other 
texts. Ch, simply reads (#8. Similar sequences are found in the context of awakening when the 
Buddha realizes the four Noble Truths (e.g. in Vin I 11: cakkhum ... Aanam ... panna ... vijja ... 
aloko ... udapddi). The terms in the TGS could well be inspired by such canonical passages or their 
counterparts in the Sanskrit. A classification of the different visions (caksus) is given in MPPU, 
V.2260ff. In the TGS itself only the divya-, buddha- and tathagata-caksus are mentioned. These 
supernatural eyes enable to perceive the buddha-nature within sentient beings or, in case of the 
divinity and the person with divine vision, it leads to the perception of the precious substance 
hidden in inferior materials. 

56 Srog chags su gyur pa, Skt. *pranaka-jdta (MVy 4908), *prana-bhita (MVy 4917), *prani- 
jata (cf. also n. 87): for sata at the end of a compound the pW shows: “Alles was — heisst, irgend 
ein —, allerhand —.” The whole term sems can srog chags su gyur pa, Skt. *sattvah pranijdtah 
could thus be translated as “sentient beings, [i-e.,] all that lives” or “living beings of all kinds of 
life.” Pranijatah emphasizes the inclusion of all different kinds of life similar to OM.16f. (... srid 
pa’i ’gro ba thams cad) which designates all different states of existences. For the discussion of 
Pali pana see Schmithausen 1991 where the specific meaning of pana “animals” is noted (p. 19). 
Animals are explicitly said to have buddha-nature. in 7B.3-6. Supposing, as another possibility, 
sattva pranabhitah as the underlying Skt., Schmithausen 1991: 58f. points out: “ ... ‘pana’ is, both 
in Buddhist and Jaina sources, also used in a way which suggests comprehensiveness, e.g. in 
phrases like ‘all sattas, panas, bhitas....’ ... originally in these phrases the terms are used as quasi- 
synonyms, with a tendency towards co-extensiveness,....” Also this analysis would fit well with the 
supposed focus on animals. For p(r)anabhiita see Norman 1987: 39f. Ch): —t05R4E; Chz: 
— DAB. 


PDF Version: BPPB VI (2002) 

encased in myriads” of defilements,** [such as] desire (rdga), anger (dvesa), 
misguidedness (moha), longing (trsna) and ignorance (avidya).°° 

And, sons of good family, [he] perceives that inside sentient beings 
encased in defilements sit many tathagatas, cross-legged and motionless, 
endowed like myself with a [tathagata’s] knowledge and vision.© And [the 
Tathagata], having perceived inside those [sentient beings] defiled by all 
defilements®! the true nature of a tathagata (tathagatadharmata)” 

*” Bye ba phrag ‘bum, kotisatasahasra are literally “a hundred thousand of ten-millions.” The 
terminology shifts here from kofiniyutasatasahasra (OD till OK) to kofisatasahasra. 

°8 (1) Nyon mongs, Skt. klesa, Pa. kilesa: My rendering “defilements” corresponds with the 
semantic of kilesa in the Pali Canon (there mostly upakkilesa). Originating from this, kilesa can 
also have the meaning “wickedness, desire, etc.” Furthermore, based on the theories of the 
Abhidharma and on the influence of standard Skt., klesa is understood as “suffering, paining” 
(Schmithausen 1987b: 246f.). As the enumeration shows, klesas comprise “wrong intellectual as 
well as affective attitudes.” 

(2) In the TGS enumerations of klesas appear in six passages, among them four times the same 
enumeration of raga, dvesa, moha, trsna and avidya (0M.9f.; 1A.9-10; 6A.5—6; 6B.1-2). 6B.4 is 
reduced to avidya and only in 2B.2—-5 we find a different, longer list: “raga, dvesa, moha, mana, 
mada, mraksa, krodha, vyapada, irsya, matsarya etc.” (see n. 92 (2)). The first members of the list, 
namely raga (or lobha), dvesa (or pratigha) and moha, constitute the common group of the three 
akusalamila. In YBha 168.12ff. they are subsumed among the group of the three bandhanas and 
the three upaklesas which are said to be synonymous terms (paryaya) for klesas (YBhu 166.24ff). 
Trsnd and avidyd maintain also a central position among the kleSas. Avidya forms the departure 
point in the pratityasamutpada and is the basis of the four viparydsas. Also trsnd is a member of 
the pratityasamutpdda. Being divided into kamatrsna, bhavatrsna and vibhavatrsna, trsna 
constitutes the main factor for being born in the sarmsara. Ch, shows only the first three Alesas: 
raga (&2k), dvesa (=) and moha (FA). It does not mention that sentient beings are encased, nor 
does it say anything about the number of the klesas. 

°° The construction nyon mongs pa ... gyi sbubs su gyur pa (0M.10 and similar 0M. 12) deserves 
attention: The phrase nyon mongs pa’i sbubs (kleSakoSa) should be interpreted as a karmadharaya: 
“sheaths that are defilements.” The Skt. in 0M.8-11 could have been: *radga-dvesa-moha- 
trsnavidya-klesakosa-kofisatasahasra-gatan sarvasattvan pranabhutan ... vilokya. Gyur pa could 
here represent Skt. -gata in the sense of “gone to [and thus being in].” 

6 Bth does not mention that the tathagatas are also endowed with the Tathagata’s vision. Ch, 
translates the two sentences above as follows: "In the same way, sons of good [family], I see with 
my buddha-vision that inside the defilements ... of all living beings is the knowledge, vision and 
body of a tathagata sitting cross-legged, dignified and motionless.” Ch, avoids the repeated 
mentioning of the fact that living beings are encased in defilements (also Ch; avoids this). The 
address kulaputrah that is separating the two sentences in the Tibetan appears in Ch, only at the 
beginning of the following section. In its contents Ch, differs from the other versions in stating 
that not living beings themselves, but the tathagatas within are encased in defilements. This 
relation between living beings, buddha within and defilements seems to correspond correctly to the 
analogy of the tathagatas inside the lotuses (= living beings) hidden by their petals (= defilements). 
Ch;: “In the very same way, sons of good [family], the Tathagata, Honourable One and Perfectly 
Awakened One perceives with the knowledge-radiating vision peculiar to the Buddha, the 
defilements ... of all sentient beings. [And he perceives that] in the wombs of those sons and 
daughters of good {family} who have sunk deeply into defilements, there are myriads of buddhas, 
all like myself. [He perceives with] a tathagata’s knowledge [and] vision that they all have the 
nature of the buddha-qualities (446389), [the buddhas within] sitting cross-legged, quiet and 
motionless.” Ch, is not entirely clear, but seems to interpret the last part of the passage as an 
independent statement. The reading of Tib, Bth and Ch, which attribute the knowledge and vision 
of a tathagata in this last part to the tathagatas within, is more convincing as it avoids the otherwise 
repetitive pattern. 

‘! The underlying Skt. is probably a bahuvrihi like *sarvaklesaklistesu to a not mentioned 
sattvesu. With gyur pa Tib stresses the perfective aspect of klista. 

© Chy: HIRI Cho: SORE. 


PDF Version: BPPB VI (2002) 

motionless and unaffected by any of the states of existence,” then says: 
‘Those tathagatas are just like me!’™ 

Sons of good family, in this way a tathagata’s vision is admirable, 
[because] with it [he] perceives that all sentient beings contain a tathagata 

[1 The simile of a tathagata in a lotus] 

“Sons of good family, it is like the example of a person endowed with divine 
vision [who] would [use this] divine vision to look at such unsightly and 
putrid lotuses, not blooming and not open, and would [owing to his vision] 
recognize that there are tathagatas sitting cross-legged in their center, in the 
calyx of [each] lotus, and [knowing that, he] would then desire to look at the 
forms of the tathagatas;°° [he would] then peel away and remove the 

8 Srid pa’i ‘gro ba, Skt. bhavagati: generally (1) the hells, (2) animals, (3) pretas (manes or 
hungry ghosts), (4) humans and (5) gods; see AKBh IIl.4ab and commentary  114.6ff.). In some 
texts also (6) asuras (“demons”) are included. 

* Instead of nga dang ‘dra’o (0M.17f.) Bth has: de bzhin gshegs pa de dang ‘dra’o (also Chy): 
HNFX). The passage in Ch,: “Sons of good [family], though staying in all states of existence (#2), 
all living beings have inside [their] defiled bodies the store of a tathagata (413%) [which is] 
constant, unspoiled and endowed with [all] qualities—just like me, not different.” A translation for 
tathagatagarbha appears in the Tibetan versions and in Ch, only in the following sentence that is 
missing in Ch). Chp: ... ROREWAEHE EH, MAR AATAERE. : “... [and he perceives] the 
store of the qualities of a tathagata, [which is] originally motionless [and] cannot be spoilt by any 
state of existence...” A passive construction with FAT is not common. One would rather expect 
the sequence fj i. Based on the text as it is, the phrase 28 #X AAT AEX should be understood 
as “... cannot be spoilt by wrong views (dysti) [about/in] any of the states of existence” or (as a 
dvandva) “... cannot be spoilt by any of the states of existence and wrong views.” 

§ The last sentence is missing in Ch,. Ch; approximately corresponds with the second part of 
the sentence (BT, MARRERO ASLO. ): “Sons of good [family], in the 
same way the Tathagata perceives with [his] vision of buddha-knowledge that all sentient beings 
are stores of a tathdgata.” The Tibetan text gives rise to certain doubts. The adjective mdzes pa 
(Bth: bzang po) for Skt. *prasadika in order to describe the Tathagata’s vision is not common at all 
and does not appear in Ch;. Also the phrase sems can thams cad (sarvasattvah) without any 
demonstrative pronoun is surprising. In all other cases of summarizing statements which repeat the 
before-said in simple words (see 1B.1-3; 1.5; 7.3), the demonstrative pronoun ‘di accompanies 
“living beings.” 

6 (1) With 1A the first of the nine similes starts with the introductory formula common to all 
the similes. In contrast to the chapters before, in 1A a supernatural vision is necessary in order to 
recognize the tathagatas in the lotuses. The flowers are described as being closed. These two points 
form an inconsistency to the chapters before. In 0G (last passage) the flowers are said to have 
opened and there is no remark in the following chapters suggesting that they became closed again. 
The buddhas within the flowers are described as being seen everywhere. No reference to a 
supernatural vision is made. For the meaning of these inconsistencies regarding the textual history 
of the TGS see section A 1.4 in part I. 

(2) In Bth and Ch, the supernatural vision only appears once; Ch; combines its second 
appearance with the recognizing of the tathadgatas in the flowers. Ch, describes the lotuses as not 
yet blooming (BX7E; 1.1b: F= BAB), and does not speak of the desire to look at the buddhas. 
Instead, it is said that the person directly sees them. Ch, describes the lotuses vividly as “wrapping 
and bothering [the tathagatas]” with their petals (...383H 76320828, 3258. ). Instead of the tathagata 
forms Ch, speaks of the “true essence/true composed body of a buddha” (#13 BAS) (see Unger 
2000: 114, s.v. ¢’ 4). Bth continues here and in the following passage to speak of the “bodies” 
(sku for kaya?) of the tathagatas and omits the paryanka position (... bzhugs na ni : de bzhin nyid 
mthong zhing shes ... for ... skyil mo krung bcas shing ‘dug par rig....). | am not sure whether IB 


PDF Version: BPPB VI (2002) 

unsightly, putrid and disgusting lotus petals®”’ in order to thoroughly clean 
the forms of the tathagatas.™ 

In the same way, sons of good family, with the vision of a buddha, the 
Tathagata also perceives that all sentient beings contain a tathagata 
(tathdgatagarbha), and [therefore] teaches the Dharma [to them] in order to 
peel away the sheaths of those sentient beings [encased in such] defilements 
[as] desire, anger, misguidedness, longing and ignorance. And afier [those 
sentient beings] have realized the [Dharma, their] tathagatas [inside] are 
established in the perfection [of the tathagatas].””° 

“Sons of good family, the essential law (dharmata) of the dharmas is this: 
whether or not tathagatas appear in the world, all these sentient beings at all 
times contain a tathagata (tathagatagarbha).”! 

in Ch) and de bzhin nyid (*tathata) in Bth share a common counterpart in the Indian original which 
has not been transmitted or has been dropped in 7ib. De bzhin nyid could, of course, also render 

§7 (1) The verb ‘byed pa, Skt. *bhinatti (“‘to split, to break”), reminds one of the simile of the 
great cloth of the TUSN (Q 117a5—118b1; S 154b4—156a6) cited in RGVV 22.10-24.9. This cloth, 
on which the whole universe is painted, has the size of an atom. When this atom is split, it is said: 

... tat paramanurajo bhittva .... (RGVV 23.12); 
... rdul phra mo ‘di! kha phye ste / [' S: ‘di dag kha.] (TUSN: Q 118a1; S 1554). 

(2) Ch, only mentions the removal of the (whole) withered flowers (so also in 1.2a) without 
attributing any other negative characteristics to them REBAO). Ch; .. BARBER 
2, ...: ““... [and] would necessarily remove the putrid, ugly leaves.” 

% Ch, is not expressive of finality: .. RR: “... [and he] then attains the manifestation et 

the tathagatas].” Ch): ... BS SAREE. “Sin Sade to allow the buddha forms to appear.” 

® The second half of the sentence and the following passage in Ch: ... ARSBHR, 
Sie, RBIS, BREA GSTE. : “.. [the Buddha] wants them to disclose [their true nature and 
therefore] teaches the Dharma for [them in the form of] siitras, destroys [their] defilements and 
manifests [their] buddha-nature.” In the first part of the passage the supernatural vision is not 
mentioned.Ch, more or less agrees with the Tibetan versions. Instead of “to peel away” it has: 
KRSE (‘to dispel”); the construction with 4 ... 4X to express finality is not common and instead of 
4 one would expect BB. 

7 The text as it is preserved in Tib does not make sense. Bth seems to show the better reading 
(gang nan tan byed pa ni de bzhin gshegs pa yongsu dag par gnas so :): “[He,] who practices [this 
Dharma, his] tathdgata [within] is established in purity” or “[He,] who practices [this Dharma] is 
established in the purity of the tathagatas.” The difference in the readings yang dag pa nyid du 
(Tib), (*samyaktva), and yongsu dag par (Bth), *parisuddha or “ddhi, is partly to explain as a 
graphic variant (yang dag pa > yongs su dag pa) in the Tibetan. The addition of nyid could well be 
the result of a revision not based on the Indian text. That the reading *parisuddha/°ddhi should 
definitely be considered as the original one, is proved by Ch (j#¥#). The irritating genitive pa? in 
Tib 1A.11 (Bth: pa ni) could result from a mistake in the Skt. manuscript used by the translators of 
Tib which read *(sarvasattvah) yat-pratipannas tathagatah parisuddhau pratisthitah instead of ... 
tathagata-parisuddhau ... The subject of the sentence is also missing in Ch: FAA, AIIE 
(ETT, BMG eS toe : “Because [sentient beings] listen to the Dharma [they] accordingly 
practice in the right way and then gain the pure real essence of a tathagata.” In Ch, the last sentence 
is condensed to the statement that (the Tathagata) manifests the buddha-nature (of living beings): 

7| This is one of the two passages of the TGS cited word by word in the RGVV (73.11-12): 

esa kulaputra dharmanam dharmatd | utpadad vd tathagatandm anutpadad va 
sadaivaite sattvas tathagatagarbha iti / 
The Skt. is in nearly perfect accordance with Tib. I was not able to check the Skt. manuscripts in 
order to verify the vocative singular reading kulaputra. The singular rigs kyi bu is only found in P; 
and Bth. Compared to the passage in Bth there are several points suggesting that Bth is the less 


PDF Version: BPPB VI (2002) 

Sons of good family, in view of [this fact and] because [sentient 
beings] are encased in the disgusting sheaths of defilements, the Tathagata, 
the Honorable One and Perfectly Awakened One, teaches the Dharma to 
bodhisattvas and also leads [them] to put faith in this [revelatory] activity in 
order to destroy their sheaths of defilements and [thereby] also completely 
purify the tathagata-knowledge (tathagatajfana) [contained within].’” When 
(vada) in this [connection] the bodhisattva-mahdasattvas [who] assiduously 
apply [themselves] (abhiyujyate) to these Dharmas have completely become 
free from all defilements and impurities,” then (tada) [they] will be 

faithful translation: rigs kyi bu chos rnaths kyi chos nyid de bzhin te : de bzhin gshegs pa byung 
ngam : de bzhin gshegs pa ma byung yang sems can de dag thams cad ni de bzhin gshegs pa’i 
snying po’o :. Concerning Bth, at the beginning there is no equivalent for Skt. esd. Instead, Bth 
seems to have read evam (de bzhin). Regarding sadaivaite sattvas, Bth reads sems can de dag 
tharhs cad which corresponds to Skt. *sarve te sattvas. Ch; BAF, HHA, Beit, 
Ath, FREAD ARATE. : “Sons of good [family], the dharmata (A&E; 
“essential law”) of all buddhas [is this]: whether or not buddhas appear in the world, [in] all living 
beings the store of a tathagata is at all times present without change.” Ch;: BB, 40364 
tH, HTH, APES, OA AI BTE. : “Sons of good [family], whether or 
not the tathagatas appear in the world, [this is] the dharmata (%:'), the dharmadhatu (4-52) 
(“essential law”): the store of a tathigata [in] all sentient beings remains eternally without change.” 
For dharmadhatu as a synonym for dharmatda see Schmithausen 1969a: 145ff. The first part of the 
passage is a well-known formula usually associated with the law of pratityasamutpada and already 
attested for the Pali Canon (for further references and passages in Mahayana sutras, see Seyfort 
Ruegg 1969: 330f.): 

jatipaccaya bhikkhave jaramaranam /' uppadd va tathdgatanam anuppadda va 

tathagatanam” thita va sa dhatu dhammatthitata dhammaniyamata idappaccayata // 

[' edition without danda; ? edition has //] (SN II.25) 
The term dhatu is in later texts replaced by the term dharmatd (see Schmithausen 1969a: 146f.). 
The 7GS, however, associates this traditional introduction with the message that all living beings 
have buddha-nature and thus characterizes this doctrine as most central. See also Zimmermann 
1999: 157f. and 164 for what I would call an indirect reference to the formula within the SP. 

? (1) The passage in Ch, (EGEREIRISBN, HR, BARA, RRS, 
+. : “But because those living beings are covered by defilements, the Tathagata appears in the 
world and extensively teaches the Dharma for [them. He] destroys [their] impurities and purifies 
[their] knowledge of all [matters].” The statement pertaining to faith is connected with the 
following passage and translated below. In Bth the part concerning faith is not found at all. 

(2) I try to render adhimucyate (mos pa) with “to put faith.” For an extensive discussion of this 
term see Schmithausen 1982: 408f. where the meanings “intellektuelles Festhalten oder 
Uberzeugtsein ... Gefallenhaben ... Neigung ... Wtinschen oder Wollen” are given. The 7GS uses 
mos pa (*adhimucyate) in three further passages: 5B.3 ... chos kyi gter chen po de la mos nas 
(5B.3); ... ngas bstan pa la mos gyur pa // (5.5a) as well as in the corresponding verse 5.5c. In all 
cases mos pa refers to faith in the teaching. Also the Chinese versions make it hard to believe that 
Tib which speaks of faith in the activity, could be based on the Skt. text. In Ch, adhimukti (33) 
clearly refers to the Dharma. Ch», on the other hand, expresses a differentidea: 403% ... (ERIE 
Bs, SiRHRR. : “The Tathagata ... performs such tasks {as to] lead them ( sentient beings) to 
faith.” Bth does not have any correspondence. 

™ Nye ba’i nyon mongs pa, Skt. upaklesa: there is no clear distinction between klesa and 
upaklesa in early Mahayana texts (see e.g. YBhu 166.24ff.). Also in the 7GS the mentioning of 
both, klesa and upaklesa, has a purely extending character. In the verse section considerations 
metri causa could have been decisive. Consequently I have chosen the more general translation of 
“impurities” instead of adopting an abhidharmic technical-analytical rendering such as e.g. “side- 
defilements” or “sub-defilements” (see AKVy 493.24ff. (with my own punctuation): ta upaklesa 
eveti klesa-samipa-rupah // kleso va samipa-varty esdm iti klesa-pravytty-anuvytter // aparipurna- 
klesa-laksanatvac copaklesah //). 


PDF Version: BPPB VI (2002) 

designated ‘tathagata, honorable one and perfectly awakened one,’ and 
[they] will also perform all the tasks of a tathagata.”” 


Then at that time the Exalted One uttered these verses: 

[1.1] “It is as if [there were] a disgusting lotus whose [unsightly] sheath-[like] 
petals were not opened out, yet [whose] inside [containing a] 
tathagata,’° were unpolluted [by the petals], and a person with divine 
vision perceived [this].””” 

[1.2] “If this [person] peeled away its petals, in the center the body of a 
victorious one’® would appear, and no impurity would then arise any 
longer [from this] victorious one. He would appear as a victorious one 
[in] the whole world.”” 

74 Grangs su gro ba, Skt. samkhyam gacchati/upeti, has the meaning of “being designated as” 
as well as “being counted as” (see Schmithausen 1969a: 134, n. 96). Cho: FHB. 

”> The last passage in Ch,: “Sons of good [family], if bodhisattvas have faith and interest in this 
Dharma and practice with a mind exclusively [dedicated to it], then, attaining liberation, [they] 
will become perfectly awakened, granting widely the acts of a buddha to the world.” Ch): “[When 
these bodhisattvas,] after having gained faith, rigidly stay with the Dharma, [they] will gain 
liberation from all defilements and impurities. At exactly that time [they] will attain their destiny 
[as] tathagata, honourable one and perfectly awakened one in [each of] their worlds. They will 
[then] be able to perform the buddha-acts of a tathagata.” In contrast to the Tibetan versions, the 
Chinese translations both mention the world(s) ({tf#]). The attribute mahdsattva appears alone in 

76 The verse is extensively discussed in A 2.2 of part I. 

7 In Ch, the sheaths are not mentioned. Instead of the not opened petals of pada b, Ch speaks 
of the disgusting lotus “with its womb[-like] petals and stamen” ( ## HAG#E R385). 

78 «Victorious one” is my translation for rgyal ba, Skt. jina, another epithet of a buddha. As is 
the case in the SP, the term appears exclusively in the verses. Ch; speaks of the “unhindered 
leader” (EBB, *andvrtandyaka; see the parallel verse RGV 1.100: ... sugatah ... anavrto ...). 
Ch): WIARG (tathagatakaya). 

” (1) The use of the term upaklesa here surprises as in the TGS it is a typical element of the 
upameya, which describes the situation for living beings. However, it can also be employed in a 
material sense as the CPD documents (s.v. upa-kkilesa). 

(2) The Skt. on which the Tibetan in 1.2c is based could be ... na sambhaveyuh (see Bth: ... 
‘byung ste). In pada d the Skt. could have been sambhavati. A Tibetan equivalent for this is found 
in Bth: kun jduy ’byung. In Tib, however, only ‘gyur is found. Kun tu appears before, separated 
from it by some syllables, and has to be integrated differently into the sentence. I assume that an 
original reading kun tu ’byung turned erroneously into kun tu ‘gyur (both are graphically similar 
and often variae lectiones for each other). Then, in a redactional step that bypassed the Sanskrit, 
the uncommon kun tu ’gyur was separated into kun tu and ‘gyur. 

(3) The second half of the verse in Ch,: FOBRHAIKEAC ExRRHA TER: “In order to cut away the 
defilements the Victorious One appears [in] the world.” or “Because the defilements have been cut 
away the victorious one [in the lotus]....” The second alternative is in accordance with the Tibetan. 
In the first alternative the Victorious One cannot be in the lotus but manifests himself in the world 
from another sphere. Chp: (87 +i34/0 MZ AA MEEISIE®: “[The body of the tathagata] would 
not anymore become spoilt by defilements and [that tathagata] would then awake perfectly in the 


PDF Version: BPPB VI (2002) 

[1.3] “In the same way, I also see bodies of victorious ones placed in the midst 
of all living beings, encased in myriads of defilements [that are] just 
like the disgusting sheaths of a lotus.”® 

[1.4] ‘“[And] because I also desire to remove [the defilements] of those 
[sentient beings, I] am continually teaching the Dharma to the wise,*! 

‘May these sentient beings become awakened!’ 
[And | purify [their] defilements, so that [they may become] victorious 

[1.5] “My buddha-vision is like that [person’s divine vision]: with the [vision 
of a buddha I] see that in all these sentient beings the body of a 
victorious one is established,® and in order to purify them [I] preach 
the Dharma.”** 

80 Bth has snying po instead of dkyil in Tib; Chz: AY. The bodies of the jinas: ROZR% (Ch,); 
ROARED ES (Ch): “wonderful essence of a tathagata.” Instead of kofisahasra (“myriads”): Chy: 
4; Ch, does not mention the encasing but speaks of the removing (f) of the defilements. The 
analogy in d of Ch, © c in 7ib) runs: FJ/RREHIZESHZE: “disgusting just as the withered lotus 

81 mKhas pa rnams refers to the bodhisattvas, as becomes evident through its parallel in the 
prose (1B.5—7). “Wise” does not appear in Ch,. 

82 Im the second half of the verse Tib and Ch, have a thought of the Buddha (3% Bhat 
#(§ 78 SRR HIDR#g): “The Buddha continually reflects: ‘All those sentient beings desire to 
accomplish [their] essence of a tathagata.’” Bth reads ci Itar instead of zhes and constructs the 
statement as the consequence of the purifying activity: “... so that (ci /tar) these living beings 
become awakened.” Padas c and d in Ch;: #ASRRIEK SURAM APA: “[I] widely teach the right 
Dharma for [them] so that [they] quickly accomplish the path of a buddha.” Pada d of Tib (= c in 
Bth) is missing in Ch. 

8 (1) For the introductory formula in pada a see the parallel in verse 7.3, pada a. 

(2) The text of Tib is dubious. The idea that “all these sentient beings are established in the 
body of the Victorious One” is known as a later step of philosophical development. So, for 
example, RGV I.27 states that living beings are enshrouded and penetrated by the dharmakaya. 
However, neither the context of the first simile nor the following illustrations of the TGS allow for 
such an understanding as expressed in Jib. This becomes even more plausible if we compare the 
two padas with Bth, which seems to have preserved the meaning as intended in the Skt. original: 

nga’i' sems can ‘di dag kun mthong ba // 
de ru rgyal ba’i sku yang rab tu gnas // 

[With] my [supernatural vision I]' perceive all these sentient beings: inside them the body of a 

victorious one is established (*pratisthita). 

[! originally probably nga yis (“I [perceive] ...”) which became nga yi>nga’i; yis would constitute the 

ninth syllable for the correct meter.] 
My translation above, in italics, is based on this. A possible alternative to the suggested translation 
would be: “... that all these sentient beings are established as a body of a victorious one (i.e., as a 
body containing a buddha).” See Takasaki 1981: 16:.. -WOOREBILLECL<, BIRO 
HLEUCHYIT4AOX®.... Ch: —ORES PRAM: “... that [in] the body of all living 
beings the store of the buddha is established, quiet [and] hidden.” In Ch , the same statement is 
abbreviated to only a single pada (—' WBE): “... [that] all sentient beings are established 
[in] the position of a buddha.” 

* The last pada in Ch,: 22¥%£SBHER: “{I] teach the Dharma so that [the store of a buddha 
within] is disclosed and manifested.” The last two padas of Ch;: “Because of this [establishment of 
all sentient beings in the position of a buddha] I continuously teach the wonderful Dharma so that 
[these sentient beings] may attain the three bodies endowed with buddha-knowledge.” 


PDF Version: BPPB VI (2002) 

[2 The simile of honey shielded by bees] 
“Sons of good family, again it is as if there were, for example, a round 
honeycomb hanging from the branch of a tree, shielded on all sides® by a 
hundred thousand bees and filled with honey. And a person desiring honey, 

[and knowing of the honey within,] would then with skill [in the application 

of appropriate] means® expel all the living beings, the bees,*’ and then use 

the honey [in the way] honey is to be used.*® 

In the same way, sons of good family, all sentient beings without 
exception are like a honeycomb: with a tathagata’s mental vision®® [I] 
realize that [their] buddhahood within is ‘shielded on all sides’ by myriads 
of defilements and impurities.” 

8 Kun tu srung ba for *araksati or *samraksati, which describes the activity of the bees in the 
TGS, has an overall positive connotation. In opposite to this, RGV I.102-104, taking up this 
illustration, shows upagidha, avrti and avyta. The choice of terminology in the RGV could be 
influenced by its convenient applicability to the second level of the simile, designating the 
encasing function of the k/esas. In this sense the choice of the terminology is more standardized, 
whereas the 7GS still uses the vocabulary relating to the function of the bees, namely the 
protection of the collected honey. The TGS thereby vividly stresses the risks involved for the 
honey robber. 

86 Thabs (la) mkhas pa, Skt. upayakausalya, in the soteriological sphere designates the ability 
of buddhas and bodhisattvas to know and to apply stratagems appropriate to the character of living 
beings and their situation, in order to lead them to awakening. Here, the choice of this 
characteristic term on the first level of the illustration already paves the way for the description of 
the Buddha’s activity, as is stated explicitly in 2.3d. Ch: 1599 (8; Chp: 5A (EB. 

°7 The term srog chags kyi rnam pa bung ba (Bth: bung! ba srog chagsu gyur pa [' for bus]), Skt. 
*ksudrapranakajata, appears partly in the corresponding section RGVV (61.3): ksudrapranaka- ...; 
srog chags sbrang ma ... (RGVV, 119.12): “minute living beings.” The Tibetan has translated 
ksudra with sbrang ma (in the TGS: bung ba) and rendered jata with rnam pa (Tib) and su gyur pa 
(Bth). However, the more common meaning of -/ata is “all (different kinds of)” (cf. n. 56). Chj2 
translate only #%. 

88 (1) Instead of a round honeycomb hanging from a branch, Ch, speaks of pure honey (}@%) 
in a tall tree (7ER#tIH) guarded by innumerable (£8) bees. It is not said in Ch, that the person 
desires honey (but cp. RGV 1.102b: purusas tadarthi and 1.104: naro madhvarthi). According to 
Ch, the honeycomb hangs from a big tree. It is not explicitly stated that the comb is filled with 

(2) The last part of the sentence in Ch,: ... BEB, BRT. : “... [and] would, just as [he] 
likes, use [it] as food, [and its] benefit would reach [beings] far and near.” (For fit %& cf. the parallel 
verse RGV I.104.b: yathakamatah). 

8 My translation of “mental vision” for jfidnadarsana, ye shes mthong ba, departs from a 
grammatical analysis as a karmadharaya: “vision, i.e., knowledge” (in this sense also e.g. Vin III 
91). Citing this illustration, the RGV is operating with sarvajnacaksus instead of 
tathagatajfidnadarsana (RGV 1.103a); Ch,: (AR. 

” Rig go in 2A.8 could also refer back to the very beginning of the sentence (including rigs kyi 
bu ...) and thus reign the complete sentence. The position of the agent de bzhin gshegs pa’i ye shes 
mthong bas after the object in 7ib is unusual and could be based on the Skt. text, as the concerned 
part appears also in this position in Ch, (but not in Bth!). Ch;: “In the same way, sons of good 
[family, in] all living beings there is the store of a tathdgata just as pure honey is in a tall tree. 
[They] are covered by all defilements just as this honey is shielded by a multitude of bees. With 
[my] buddha-vision I see this in accordance with reality.” Ch): “In the very same way, sons of 
good [family], all sentient beings are like a honey comb: encased and shielded by myriads of 
defilements and impurities. When with [my] tathagata’s mental vision [I] had been able to 
comprehend this [fact, I] achieved perfect awakening.” or “... this [fact], then [they will be able] to 
achieve...” The last part could also be understood generically as “When ... [one] has been able to 
comprehend this [fact], then [one] will...” Instead of “buddhahood” in Tib and Bth, Ch; shows 
QUaKRHE. In Ch no equivalent exists. 


PDF Version: BPPB VI (2002) 

Sons of good family, just as a skillful person by [his] knowledge 
realizes that there is honey inside a honeycomb shielded on all sides by 
myriads of bees, in the same way [I] realize with [my] tathagata’s mental 
vision that buddhahood is without exception ‘shielded on all sides’ in all 
sentient beings by myriads of defilements and impurities.””? 


_ “And, sons of good family, just like the [person who] removed the bees, also 
the Tathagata, with skill in [the application of appropriate] means 
(upayakusala), removes sentient beings’ defilements and impurities [from 
their buddhahood] within, [such as] desire (raga), anger (dvesa), 
misguidedness (moha), pride (mdna), insolence (mada), jealous 
disparagement (mraksa), rage (krodha), malice (vydpada), envy (irsyda), 
avarice (mdtsarya), and so on,” [He] then teaches the Dharma in such a way 
so that those sentient beings will not again become polluted and harmed by 
the defilements and impurities.” 

*! This last passage does not give any new information and is missing in Ch,. The fact that the 
wording is widely identical with the passage right above could mean that it resulted from a 
redactionally amplified dittography or that it is a later interpolation. Also in Bth a part of the 
passage is missing. This, however, is probably due to a scribe omitting the part between bzhin du / 
(2A.10) and thabs la (2B.1). The passage is also found in Ch,, with a syntax which probably 
imitates the Skt. word order: (E—WAALUIRE WAC, PME, RRM ET ORIK. Be 
TES ATS: “*..., in the same way, [as for] all sentient beings, when I recognized [it]/understood 
[them] with [my] tathagata’s mental vision, [I realized that their disposition of] becoming a buddha 
({#8) inside them is covered by myriads of defilements and impurities.” 

* (1) sBrang rtsi byed pa (Bth) instead of bung ba in 2B.1 is also found in the corresponding 
verse I.104b of the RGV: madhukaran. 

(2) For ’chab pa, Skt. mraksa, see BHSD s.v.; Chz: ##. Nearly identical enumerations are found 
in e.g. Lal 52.12ff., 411.15ff. and Siky 198.8f. Bth mentions dregs pa [emended for drag pa] 
(usually Skt. darpa) instead of rgyags pa (mada) and ngan sems (usually dustacitia) instead of 
phrag dog (irsya). Probably in order to match the high number of bees the list of Alesas has been 

(3) In Ch; 2B is reduced to: “With skill [in the application of appropriate] means [I] teach the 
Dharma according [to the situation] (BA/, *anukula), destroy and remove the defilements [of 
living beings and] disclose [their] mental vision of a buddha [so that they will] widely perform the 
tasks of a buddha for the world.” As is the parallel in Ch, 2.4, it is not clear if the performers of the 
acts of a buddha are intended to be living beings or the Buddha himself. If, in contrast to the 
Tibetan (Jib has ’byed and not mdzad!), the second is the case, the final part in Ch, should be 
understood as “... disclose [their] mental vision of a buddha. [Therefore I widely....” 

% 1 iterally: “... so that among those sentient beings there will not again appear [those] polluted 
by the defilements and impurities and [so that they are not again] harmed.” It is not clear how to 
understand the construction sems can de dag la ... nyon mongs pa can du mi 'gyur.... That sems 
can de dag la reflects a genitive construction in the SKt. text is only possible if we assume that the 
Tibetan translators have erroneously separated it from *dharmam desayati (chos ston to). 
However, there is no evidence for that in the other translations. If, on the other hand, we 
understand /a in a partitive sense, the whole passage seems more plausible, as it is probably 
thought that only a part of the whole number of living beings is endangered of becoming polluted 
again. The particle Ja does not appear in Bth (see below). As is the case in the compound de bzhin 
gshegs pa’i snying po above, where in some versions of 7ib the particle can has been affixed in 
order to differentiate between, on the one hand, living beings themselves as the essence of a 
tathigata (without can this would be the most natural way of understanding) and, on the other 
hand, the essence of a tathagata as something possessed by living beings, the appearance of the 
particle can in nyon mongs pa can may be caused by similar considerations, namely in order to 
differentiate clearly between the defilements and living beings. The use of can is not found in Bh. 
The passage in Brth can be understood in two ways: ... nye bar nyon mongs pa med par bya ba’i 


PDF Version: BPPB VI (2002) 

[When their] tathagata’s mental vision has become purified, [they] 
will perform the tasks of a tathagata in the world. Sons of good family, this 
is how I see all sentient beings with my completely pure vision of a 

Then at that time the Exalted One uttered these verses: 
[2.1] “It is as if there were a honeycomb here, shielded on all sides and hidden 
by bees, [but] a person desiring honey would perceive the [honey 
within] and expel the bees.” 

phyir : de bzhin du chos ‘chad do : ci ltar ci lta bu yang sers can de dag ... nye bar nyon mong, 
par mi 'gyur ba dang : gdungs par mi ’gyur ba dang : de bzhin du de bzhin gshegs pa’i yeshes 
mthong ba de rnam par sbyangs te : ... byed do :: (a) “In order to remove the impurities ... [the 
Tathagata] teaches the Dharma in such a way that those sentient beings do not again become 
polluted and harmed ... and, after the[ir] tathagata’s mental vision has become purified in such a 
way, [they] perform...,” or (b) “In order to remove the impurities ... [the Tathagata] teaches the 
Dharma accordingly. In such way that those sentient beings do not again become polluted and 
harmed ... [ignore dang!] in that way [the Tathagata] purifies the[ir] tathagata’s mental vision and 

Analysis: a. Contrary to Tib but with Ch2, Bth considers the removal of the impurities as the 
intention of the teaching. The different understanding could result from a mistake in the Skt. 
msnuscript which instead of *°klesanam vinasaya could have read °klesan vindsya (so Tib). The 
reading in Ch, and Bth seems much more plausible than the temporal sequence of removing the 
klesas and then teaching the Dharma as suggested by 7ib. Also in the parallel verse 2.3 of Tib we 
find the relation of finality as in the prose of Ch) and Bth. b. The position of chos ‘chad do 
between the description of the removal of the klesas and the mention that living beings will not 
again be harmed is parallel to Ch, (and contrary to Tib). Bth and Ch; probably reflect the Skt. 
syntax with the verb in the middle of the sentence. 7ib, on the other hand, has arranged the 
sentence with the tatha ... passage at the end. The original could have been: *... tatha (tathd) 
dharmam desayati yatha te sattva na bhiiyas taih klesopaklesair upaklisyante na (ca).... A similar 
syntactic structure is found in the RGVV (50.2; citation from the Sagaramatipariprccha(sutra)): te 
‘smabhis tathd pratyaveksitavya yatha na bhiiyah Slisyeyuh / (= RGVV, 97.7: de ni kho bos ci nas 
kyang 'brel bar mi 'gyur ba de Itar so sor brtag par bya’o /) “Those [klesas] we have to consider 
thoroughly so that [they] might not again enclose [living beings].” The particle yang for Skt. punar 
or bhuyah in 2B.5 has its counterpart in Ch, with 7. 

This passage and the following in Ch2: “In order to instruct sentient beings [and/how] to dispel 
the impurities..., [in the same position] as the person harming the bees, by the power of skill in the 
application of appropriate means (updyakausalya), the Tathagata teaches the Dharma in such a 
way so that sentient beings are not polluted by defilements, [that] there is never again annoyance 
[for them] and [that they] are not enclosed! [again by defilements]. Sons of good [family], how 
[should] I with [my] tathdgata’s mental vision perform the acts of a Buddha in all worlds [for] 
these sentient beings in order to purify [them]?” [' {tif could represent a translation of a form of the 
root slis; see RGV 103.4cd: ... aslesam ... ddadhati //) 

The text of Ch; in this passage is not very clear. The corresponding verse 2.4 of Ch; relates the 
whole question to living beings asking how they could become buddhas to perform buddha-acts. 
Apart from this, Amoghavajra seems to have positioned tathagatajfidnadarsana (h07K*9 A) in a 
way different from 7ib. The interrogative Z.{®] which is also found in the corresponding verse 2.4 
of Ch, could be based on the reading kena instead of yena (ji Itar). 

4 Ch) describes sentient beings as pure (R284 (8 41288). Again, this final sentence 
which somehow tries to essentialize the simile is not found in Ch;. Since it is also not reflected in 
the verses, it should be regarded as a later interpolation. 

°° Ch, keeps to its own prose and accordingly speaks of “honey in a tall tree”; the bees are 
innumerable (4%) and the desire for honey is not mentioned. 


PDF Version: BPPB VI (2002) 

[2.2] “Also here” in the same way, all sentient beings [in] the triple states of 
existence” are like the honeycomb. The many myriads of defilements 
are [like] the [bees, yet I] see that inside the defilements there exists a 

[2.3] “Also, in order to clean [this] buddha,” I remove the defilements, even as 
[the person desiring honey] expels the bees. [Using appropriate] means, 
[I] teach here dharmas so that myriads of defilements will be 

[2.4] “[I do this] in order to induce those [sentient beings], after becoming 
tathagatas, to continually perform the tasks [of a tathagata] throughout 
the world, and with readiness in speech'”! to teach the Dharma, [which 
is] like a pot of honey from bees.”'° 

% ‘Dir, Skt. iha or atra, in this case designates the non-fictive level of the simile, i.e., the world 
of sarnsara. 

” Srid pa gsum, Skt. tribhava comprise (1) kama-, (2) répa- and (3) arupyadhatu (see e.g. 
AKBh 111.2). 

*8 Logically the analogy should include the verb mthong: “In the same way [as the person 
perceived the honey, I] see that...” Such a translation is made impossible in Tib due to the particle 
do at the end of pada b. However, neither Bth nor Ch; support the reading suggested by Tib. The 
text of Tib could be improved by altering do into de. In Ch; no particle of comparison appears to 
take up the first half of the analogy in verse 2.1. Instead, the applicability of the analogy between 
sentient beings and the honeycomb is intensified by adding #444: “without difference.” The 
verse in Ch,: “The store of a tathagata [in] living beings is like that honey [in] a tall tree: [evil] 
propensities and impurities encase [this store] just as the multitude of bees shields the honey.” 

” T translate sangs rgyas in 2.3a as the object of the purifying process. This understanding is 
parallel to the prose in 2A.6 and 2A.11 where sangs rgyas appears as sangs rgyas nyid (for the 
same abbreviation in verse and prose cf. MSA IX.22 and MSABh 36.21 and 22). Taking sangs 
rgyas to be the subject of the sentence would result in the formulation “I, the Buddha, remove....” 
Ch; seems to support this understanding (Fx(#)'# FAYPER MD; see also SP XIII.51: tathaiva buddho 
ahu dharmardja.... 

100 (1) Bth, Ch, and Ch, confirm that nyon mongs bye ba has to be taken as object of the verb 
gnod in 2.3c. Translating gnod pa, reigned by the agent nyon mongs pa in 2B.7, J decided for its 
meaning in the Tibetan “to harm,” whereas here the meaning of its possible Skt. original *(upa- /vi-) 
hanti, “to eradicate,” is stressed. 

(2) Ch,: “[Using appropriate] means I teach the perfect Dharma for all living beings. [I] destroy 
and remove [their] defilements [which are like the] bees [and] disclose the store of a tathagata 
[within them].” 

11 For spobs, Skt. pratibhana, in the Pali canon and early Mahayana siitras see McQueen 1981 
and 1982. In the context of the TGS, pratibhana appears as a quality attributed only to buddhas. 
“Readiness in speech” indicates the spontaneous and transcendentally inspired correct teaching of 
the Dharma in order to lead other living beings to buddhahood. 

12 (1) An analogy between the preacher of the sweet Dharma and a honey-giving bee is found 
in Udr, 11.25ff. For the Dharma compared with sweet honey see Av-s 187.6ff., 242.11ff., 249.15ff. 
Tib allows also the understanding of chos ston in pada d as “teacher of the Dharma,” who would 
thus become the point of reference for the analogy with the honey pot. However, this possibility is 
excluded by Bth where chos and bshad (for ston) are separated by kyang. 

(2) There are essentially two ways of interpreting the text of 7ib. The one suggested in my 
translation takes bya in 2.4d as expressing the inducing activity of the Buddha. Consequently ji 
ltar (for yathd or yena) in a and c grammatically depend on bya. Choosing the second way of 
reading, one would have to understand the construction ji ltar ... phyir ro as an expression 
indicating that a result is aimed at (For the usage of yathd in such a way see Speijer § 471). In that 
case bya would just indicate the aspect of potentiality and both ji /tars would be govemed by ston 
in 2.3d: “{I teach the Dharma] in order that [sentient beings} ... might perform....” I have no 
material to show that such a construction is common in Tibetan translations. Nothing parallel to 
bya phyir ro appears in Bth. (Should we assume a change in Bth 2.4d from ... bshad phyir ro to ... 
bshad par ro?) Another interesting way to understand the construction in 2.4 is found in Takasaki 


PDF Version: BPPB VI (2002) 

[3 The simile of kernels enclosed by husks] 

“Sons of good family, again it is like the example of winter rice’, barley, 
millet or monsoon rice’™ [whose] kernel (sara) is shielded all around by a 
husk (tusa): as long as the [kernel] has not come out of its husk, [it can]not 
serve the function of solid, soft and delicious food.’ But, sons of good 
family, [it can serve this function very well once] some men or women, 
desiring that [these grains serve their] function as food and drink in hard, 
soft or other [forms],’™ after having it reaped and threshed, remove the 
[coarse] sheath of the husk and the [fine] outer skin.”’” 

1981: 18 (from pada 2.3¢ on): (PPA... MIR H< OD, [2.4] ROMBMRL Roc 
IZ, RRB< REEDG CHS, : “Why do [I] teach the dharmas? [2.4] Because when they 
(= living beings) finally have become tathagatas, [they] are expected to teach the Dharma.” 
However, this alternative does not take into consideration ji /tar in 2.4 a and c. 

Parallel to the prose Ch, has the interrogative z-{a]. The particle of comparison #7411 in c could 
represent a misunderstood yatha. Ch: “How may [sentient beings] awake [so that they can] 
perform buddha-acts, [so that they can be] continually like honey pots in the world just as if [they] 
would teach [with] readiness in speech the excellent honey, [and] so that [they] would verify 
(=realize) the pure Dharma body of a tathagata?” Ch,: “Endowed with unhindered readiness in 
speech [I] preach the Dharma of the sweet dew so that [living beings] attain perfect awakening 
everywhere. [On account of (my)] great compassion [I/they then] save living beings.” 

13 For (‘bras) sa lu, Skt. Sali: “winter rice,” see Vogel 1972 and Achaya 1994: 185, 279. 
Winter rice is thought to be an exquisite cereal (see Prakash 1987: 96) and might therefore be 
mentioned first in 7ib. Kumar 1988: 27ff. gives numerous examples in Indian Buddhist literature 
which show the important role of rice. Instead of ’bras sa lu, Bth reads so pa which should be 
emended to so ba, a kind of barley. 

14 For ‘bru, Skt. vrihi: “monsoon rice,” see Vogel 1972 and Achaya 1994: 283. The four 
enumerated cereals are part of the seven pubbanna (“primary foods”), which are commonly 
mentioned in Pali literature (sali, vihi, yava, godhima, katigu, varaka, kudrisaka) (see Vogel 
1972). Ch) in the prose: #4, B, 3%, G:.“Rice, wheat, millet [and] beans”; in the 
verse: BFANFARE HERE AV|\RERPG: “Rice, millet, barley, wheat, other [cereals] and even 

105 (1) For bza’ ba dang bca’ ba dang myang ba see ...-khadyabhojyannapana-... in SP 339.3 (= 
Q 146alf.: ... bza’ ba dang bea’ ba dang zas dang skom dang ...); Pa. annapana-khadaniya: 
“boiled rice, drink and solid food” (CPD s.v. anna-pana); BHSD (s.v. asvadaniya, khddaniya, 
bhojaniya) mentions the combination of khadaniya = khddya, “solid food”) with bhojaniya (“soft 
food”) and also the compound khdadya-bhojya-svadaniya (s.v. svadaniya); also ASP 249.13f.: 
khadaniyam bhojaniyam svadaniyam ca (Tib. according to Derge: bza’ ba dang bca’ ba dang 
myang ba) and MVu 1.38.7£.: khadaniya-bhojaniya-asvadaniyena. The Tibetan of the TGS 
corresponds to the latter two passages. A classification of food into the four categories kKhadaniya, 
bhojaniya, pana, svadaniya is found in Buddhist and Jainist texts according to Prakash 1987: 125f. 
Ch; simply has 73H: “... is not fit to be eaten.” 

(2) bZa’ ba dang / bea’ ba la sogs pa zas skom in Tib 3A.4f. could be a translation of the 
compound found in SP 339.3 (mentioned in (1)) with the addition /a sogs pa. Bth is obviously 
based on another Skt. wording: bca’ ba dang : bzod' pa dang::, btung bar {'in the passage above: 
bzang ba]. Chz: REEZ_A: “people, [who] want to eat.” 

1 (1) The Skt. on which brngas shing brdungs te is based is probably two absolutives of the 
roots Jd and mrd (for rdung as an equivalent of mardana (to be emended from mardana) see DTS, 
folio 89b). Pairs of the roots /Z and mrd in the context of proceeding grain can be found in Ja 
1.215: ... layitva madditva ...; Vin 11.180: ... lavapetva ... maddapetva; Mil 360: ... lavana- 
maddanena bahudhafifiako.... These and other examples from Jaina sources are quoted in Balbir 
1996: 334ff. The reaping is not mentioned in the Chinese. Instead of reaping Bth reads phyir, 
which probably should be emended to phyar, the perfect form of ‘phyar ba (Ja: “ ... to lift up the 
grain in a shovel, hence: to fan, to sift, to winnow...”). Chz: .. LHPABKAR, MICKA. : 
“_.. threshing with their pestles {in the] mortars [they] remove their (= the grains’) husks so that 
[the grains] become eatable.” 


PDF Version: BPPB VI (2002) 


“Sons of good family, in the same way [that people are aware of the precious 
kernel within the husk, so] too the Tathagata 1°’ perceives with [his] 
tathdgata-vision that tathagatahood, buddhahood, svayambhiitva \°° — 
wrapped in the skin of the sheaths of defilements—is [always] present in 
every sentient being.’” Sons of good family, the Tathagata also removes the 
skin of the sheaths of defilements, purifies the tathagatahood in them and 
teaches the Dharma to sentient beings, thinking: 

‘How [can] these sentient beings become free from all the skins of the 
sheaths of defilements [so that they] will be designated in the world as 
‘tathagata, honorable one and perfectly awakened one’?””!!° 

Then at that time the Exalted One uttered these verses: 
[3.1] “[It is] like monsoon rice or winter rice, or [like] millet or barley, [which,] 
as long as they are in the husk, [can]not serve [their] function.”’"! 

(2) 3A in Ch;: “[It is] as if a poor, stupid [man] would disdain nonglutenous rice [and] millet [as 
long as it were] not separated from [its] husks, and would regard it as something to be thrown 
away. [But] as soon as, after removing [the husk, the kemel] has been purified, [this same grain] 
continually is of imperial use.” The character /] should here have the same meaning as is 
expressed by = in the parallel verse 3.2 of Ch,: “imperial.” 

17 Instead of “Tathagata”: #8 (Ch,); WK, FEE. IESBAD (Ch.); Bth is without subject. 

18 Rang byung nyid, Skt. svayambhitva, is here used as a synonym for buddhahood. The 
adjective svayambhu counts as an epithet of the Buddha as well as of the central divinities of 
Hinduism: Brahman, Siva and Visnu. Originally referring to the myth of the cosmogonic egg out 
of which Brahman is bom meaning “arisen out of himself” (Hacker 1978: 490f.), in the case of the 
Buddha, svayambhi should be understood as “become [awakened] by his own [power].” The 
enumeration is a particularity of the two Tibetan translations. Ch, reads only 412kfE S01, 
whereas Chp is restricted to #2688. Cf. the corresponding verse RGV 1.106: ... sattvesv api 
klesamalopasrstam ... jinatvam /, For a similar enumeration see e.g. ASPyy 570.2f.; 16f.; 
(571.13ff.): tathagatatvam buddhatvam svayambhitvam sarvajnatvam. 

109 After this passage Ch, has the following insertion which could be a very free rendering of 
the terms left out immediately before: BgetTEHE, AREER, BEAEBRZS. A possible 
translation could be: “If [they] are able to understand [this fact, namely their innate buddha-nature, 
they] will become perfectly awakened [ones], settled firmly [and] peacefully in spontaneous 

110 The Tibetan connects the three syntactical units of the sentence with dang: ... bsal ba dang / 
... Sbyang ba dang / ... ston to. Bth replaces the first dang by phyir. sNyam nas for iti (krtva) in 
3B.8 appears in Bth as phyir (for phyir rendering iti cf. AK s.v. iti). As also the Chinese versions 
show a construction of finality, we should assume that this was the original structure of the 
sentence. Replacing the second dang in Tib with phyir, the future form of the verb sbyang, in the 
present construction impossible to explain, would be justified. Similar confusion in the structure of 
the text of 7ib occur in verses 3.3 and 3.4. Ch, drops the appellation at the beginning of the 
passage and then runs: “Therefore, with [appropriate] means [I] teach the Dharma according [to 
circumstances] (4M, *yathdyogam), so that [living beings] remove [their] defilements, purify the 
knowledge of all [matters] and become perfectly awakened in all worlds.” Ch2: “Sons of good 
[family], that store of a tathagata resides amidst all defilements. The Tathagata, for the sake of 
[making] those living beings remove the skin of defilements, so that they will be purified and 
become buddhas, teaches the Dharma for [them], continually thinking: ‘When [should I teach them 
the Dharma so that] sentient beings strip off all skins of the sheaths of defilements [and] become 
tathagatas appearing in the world?’”” The temporal interrogative (kada instead of katham?) is also 
found in 3.4 of Ch). Since there it is combined with 3 in the same pada, the question must be 
interpreted as in my translation above. 


PDF Version: BPPB VI (2002) 

[3.2] “[But] having been pounded [and their] husks having been removed, they 
[can] serve all [their] various functions. [However,] the kernels in the 
husks [can]not serve [any] function for sentient beings.”!!? 

[3.3] “In the same way [that people are aware of the precious kernel within the 
husk,] I see that the ground of buddha[hood’” of] all sentient beings is 
covered by defilements. And then I teach the Dharma in order to purify 
them and let [them] attain buddha[hood] quickly.”!" 

[3.4] “In order that [they] may quickly become victorious ones, [I] teach the 
Dharma so that, like mine, [their] true nature (dharmata), which, 
[though] wrapped in hundreds of defilements, is in all sentient beings, 
becomes purified [in] all [of them)].”!!° 

'" Ch, (parallel to its prose): “[It is] as if some poor person would still disdain all nonglutenous 
rice (and] millet [whose] husks have not yet been removed and would regard it as something to be 
thrown away.” 

‘2 In Bth the padas comprise only seven syllables. The second half of the verse shows a 
separate understanding: de dag snying po phub mar Idan : bsal nas sems can don kyang byed : [the 
position of de dag may be determined by the syntax of the Skt.]: “After the kemels with the husks have 
been cleaned, [they] bring about benefit [for] sentient beings.” Pada c and d of both Tibetan 
versions seem redundant as they do not add any new information compared to the first two padas. 
Ch,: “Though [from] outside it looks as if without use, the kernel within [remains] undamaged. 
[As soon as] the husk is removed [this same grain] functions as food [for] a king.” In the light of 
two emendations, the second two padas of Ch; show similarity to the statement that the kernel 
remains undamaged in Ch;: RRM TS PEAR /EFI: “The kernel is placed [in] the 
husk and [remains] undamaged. [This] undamaged [kernel has the function of] bringing about 
benefit [for] living beings.” The idea of an undamaged kemel appears in both Chinese versions 
and constitutes a new aspect in 3.2 missing in the Tibetan. It may have been part of the original 
Skt. text but was, however, not rendered (or misunderstood?) by the Tibetan translators. 

13 The use of the term sangs rgyas sa, Skt. buddhabhiumi, might be inspired by the agricultural 
context of the illustration (Ch,: (RK; Ch: AN3RHH). 

"4 Th the second half of the verse Tib differs from all other versions because the construction of 
finality with phyir includes both statements, namely the purifying activity along with the 
attainment of buddhahood. In all the other three translations the attainment of buddhahood appears 
separated in pada d as the consequence (Ch;: 4 ...) of the teaching or as a wishful exclamation 
(Ch2: BR ...; Bth is ambiguous). The translators or revisers decided to position the second half of 
former pada c (bya phyir chos ston), where it is still found in Bth, to the end of the verse (in 
correspondence with basic rules of Tibetan grammar). Thus the statement referring to the 
attainment of buddhahood became dragged into the constructional frame of the motive. Instead of 
“buddha[hood],” in pada d, Ch, employs again —t)}; Ch): #4. 

"5 (1) Instead of dharmata: A038 (Ch,); Chz has %k. 

(2) Just as in verse 3.3, also here the last two padas of Tib show a different structure. Bth and Ch, 
(for Ch; see below) present the last pada that living beings will become jinas (Ch: (#5) as the 
thought of the Tathagata, parallel to the prose. In 7ib, again, thams cad ji Ita (bu)r, originally part 
of the question as found in the parallel prose section, has been exchanged with bya phyir chos ston 
(to). As the result of this process thams cad became impossible to construe with the rest of pada c. 
Ch, parallel to its prose, formulates a question with the interrogative {@JR¥ different from the 
Tibetan (Bth: ci nas) (see above). Also in the case of this verse, 7ib definitely is a revised text. The 
original position of the question is in the last pada just as found in Bth: “How (ci nas) may [they] 
all quickly become victorious ones?” 

(3) The verse in Ch,: “Just as my tathagata-nature, so also are living beings. [I] disclose [this 
nature of living beings and] induce [its] purification [so that they] quickly attain the supreme 
awakening (f% [38).” 

"6 (1) For “narrow path” see Tshig mdzod (s.v. gseb lam): lam chung ngam lam gu dog po /. 
Ch; has: ... RAK, BRE, WASTE, BRST RH... “.. somebody carrying a 


PDF Version: BPPB VI (2002) 

[4 The simile of a gold nugget in excrement] 
“Sons of good family, again it is like the example of a round nugget (pinda) 

of gold [belonging to] someone (purusdntara) [who] had walked [along] a 
narrow path,''® [and whose nugget] had fallen into a place of decaying 
substances and filth, [a place] full of putrid excrement. In that place of 
decaying substances and filth full of putrid excrement, the [gold nugget], 
having been ‘overpowered’ by various impure substances,’’’ would have 
become invisible,'"® [and would have remained] there for ten, twenty, thirty, 
forty, fifty, a hundred or a thousand years,'! [but it would, though 
surrounded] by impure substances, [never be affected by them, owing to] its 
imperishable nature (avindsadharmin). [Because of the covering of impure 
substances, however, it could] not be of use to any sentient being.”””° 

gold nugget would pass by the side [of the place of excrement when this gold nugget} would by 
inadvertence suddenly fall down into filth...” 

(2) Bth: ... yongsu dag pa’i lam de’i dbusu ... should be emended to ... yongsu gang ba’i lam de’i 
dbusu ... (see line 4). The person, according to Brh, is walking in the middle of the path. However, 
this text part of Bth is grammatically impossible to construe. Instead of the “place of decaying 
substances and filth” (rul pa dang nyal nyil gyi gnas) Bth mentions a pit (but in 4A.4 a heap: 
phung) of “decaying filth” (rul ba’i lud kyi khung). See KP §49: ... sarnkarakiitam ... (phung could 
render Auta); cf. SP; s.v. samkaradhana: phung po for dhana. Bth is confirmed by Ch): 
2 3e ETRE: “bad accumulations of putrid filth.” Ch,;: A‘ (see below). For a similar 
expression see SP 113.13 (1V.22): samka@radhdnam ... putikam uccaraprasravavinasitam ca /. 

(3) Lam nas could reflect an instrumental of space in the Skt. For purusdntara see CPD s.v. 
antara?: “a certain one, someone.” Throughout 4A-C neither Bth nor the Chinese versions 
mention that the nugget is round. 

7 gZhan dang gzhan is probably a translation of anyonya: “various, different (see BHSD 
s.v.). The terminology (g)non pa is used to express the activity of the impure substances. In 4.3 the 
same word characterizes the Alesas. If it renders abhibhiita, it means “to drown, to superpose.” 

8 b/Tar mi snang bar gyur: lit. “having become [something which] does not appear [anymore 
so that it could] be looked at.” Ch,: A\#4; missing in Ch. 

"9 Instead of “fifty years” Bth reads “sixty years” and does not mention “thousand years” at all. 
Ch: $8 EER (see below); Ch;: “... ten, twenty, fifty or a hundred thousand years.” 

120 (1) Whereas Ch) at least shows a contrasting element with 7f (“... yet [it] cannot benefit the 
people”), Tib is hardly understandable without extensive elaborations. Like Ch; also Bth 
terminates the first statement with a verb (‘dug pa; §%) which is missing in 7% but is found in the 
verses: gnas (4.1d). Misplaced and without correspondence in any of the other translations is Bth: 
... brtson par byed pa de dag gi don kyang ... (“also the benefit of those [who] practice...” or “also 
for those who...”) which appears instead of (lo stong du?) mi gtsang bas in Tib. Ch; in this 
passage: ... HR. BHAT, £4, DR....: “... lies in putrid feces. And though its 
nature is not perishing [and] not polluted, yet....” Chz: 4\# could be the counterpart for (mi) 
gtsang bas in Tib 4A.7, possibly misunderstood and not attributed to the gold in Tib. 

(2) The whole of 4A in Ch;: : “It is as if pure gold would fall into an impure place, [lying there] 
hidden [and] immersed [so that it would] not appear [anymore and] years would pass by. Though 
the pure gold would not perish, nobody would be able to know [its existence in that place].” 

(3) In order to throw light on the original form of this statement we should have a look on the 
parallel verse RGV 1.108: yatha suvarnam vrajato narasya cyutam bhavet samkaraputidhane / 

bahuni tad varsaSatdni tasmin tathaiva tisthed avindsadharmi // 

Considering content and structure of the verse we find great similarity to the passage in 4A. Only 
the last aspect, i.e., that the gold cannot be of use for sentient beings, is not expressed in the RGV 
passage (nor does it appear in Ch,). In the RGV there is a verb of existence (tisthet) and, what is 
most remarkable, avinaSadharmi, the counterpart of Tib: chud mi za ba’i chos can, is positioned at 
the end of the whole verse. It does not function as the predicate of the sentence but as an attribute 
to suvarnam and tad above. The verse RGV 1.108 could be translated as follows: 


PDF Version: BPPB VI (2002) 

“Sons of good family, [if] then a divinity with divine vision looked at that 
round gold nugget, [the divinity] would direct a person:'”! 
‘O man, go and clean that gold of excellent value [t]here, [which is only 
externally] covered with all sorts (-jdta) of decaying substances and 
filth, and use the gold [in the way] gold is to be used!!”* 

In [this simile], sons of good family, [what] is called ‘all sorts of decaying 
substances and filth’ '* is a designation for the different kinds of 
defilements.’”° [What] is called ‘gold nugget’ is a designation for [what] is 
not subject to perishability (avindsadharmin) [, i.e., the true nature of living 
beings].'*° [What] is called ‘divinity [with] divine vision’ is a designation 
for the Tathagata, the Honorable One and Perfectly Awakened One.'?’ 

Sons of good family, in the same way also the Tathagata, the 
Honorable One and Perfectly Awakened One, teaches the Dharma to 
sentient beings in order to remove the defilements—[which are like] all 
sorts of decaying substances and mud—'® [from] the imperishable true 
nature (dharmatd) of a tathagata found in all sentient beings.”!”° 

Just as gold of a traveler might fall into a receptacle of dirt and putrid material [and] that [gold], having 

the nature of unperishability, would lie there in exactly the same way for many hundreds of years, .... 

121 (1) Ch, does not mention that the divinity is looking at the gold (this is said in 4.1c of Ch)). 
The divinity speaks to a group of people (cf. RGV'I.109b: narasya; but RGVI.111b: nynam). 

(2) bsGo should be the future form of the verb sgo ba. The RGV has present optative forms in the 
parallel verses: pravadet, upadiset (1.109, 111). With bsgo, the so-called “future” form, the Tibetan 
could render an optative in the Sanskrit. At the end of 7A we face the same situation (bsgo) along 
with darsayet and vadet in the corresponding verses of the RGV (1.118, 120). Cf. the section on 
irregular verb forms II C 3.2. 

12 The parallel in RGV 1.109 runs as follows: suvarnam asminn idam' agraratnam visodhya 
ratnena kurusva karyam // [‘idam emended for navam according to Schmithausen 1971: 154]. All 
elements can also be found in Tib. Ch; understands the order to clean and to make use of the gold 
already as the following action: HARMIG, FIBER. SC, xt, BRSATA. : “Having 
heard [the order], this person then takes it out. After having attained [it, he] cleans [it and uses it] 
in accordance with what gold is used [for].” In Ch, there is no directive to go. Instead of the 
purification it is said: 77 +4 7: “Take it out!” 

30. bsgo na rigs kyi bu ... the particle na refers to the whole illustration from the beginning 
in the sense of “... if a gold nugget had fallen down..., and a divinity had directed a person..., 
[then], sons of good family, [the following analogy could be drawn)...” 

4 Only Tib does not include “the place” in the explanation of the analogy. Bth: rul pa’i lud 
phung' {' for khang]; Ch: AYR; Chy: RRBOR. 

°° Chy: BIBS: “innumerable klesas”; Chy: TASER THR BEAM: “different kinds of klesas 
and upaklesas.” 

126 Bth reads here and also in 4.1 avinasadharmatd (chos nyid ma rung bar mi ’gyur ba; see 
also Chz: emended to 7\#8%t':. In Ch,, as usual, dharmata does not appear, therefore: 417RiK. 
Tib has most likely been adapted to conform to the expression in 4A.7, where it has chud mi za 
ba’i chos can. Bth and Ch; are here probably based on the Sanskrit original. 

27 Ch, reads RIBK. 

128 Instead of nyon mongs pa rul pa dang / ‘dam rdzab lta bu rnams, Bth reads nyon mongs pa’i 
rul ba’i ‘dam du gyur pa. This makes clear that in the Skt. there was a compound ending with ~jata 
or -bhita (lta bu for jata see MVy 5390; for bhiita see MVy 5391). -Jata or -bhitta could here of 
course also be understood as showing a comparison. However, as in 4B.3, where -/ata (there: rnam 
pa) had the meaning: “all sorts,” this meaning should also be adopted here. In the Chinese too a 
comparison does not appear explicitly. Similar compounds as here appear also in the parallel 
verses of the RGV: tatklesapanka- (1.110c); klesamahasuci- (1.111c). 

'29 This last passage is different in all translations. The shortest version has Ch,: “Therefore the 
Tathagata widely teaches the Dharma for [living beings] so that they all remove and destroy [their] 


PDF Version: BPPB VI (2002) 

Then at that time the Exalted One uttered these verses: 

[4.1] “It is just like [the example of some] man’s nugget of gold [that] has 
fallen into all sorts of filth: though it remained there in such a state for 
not a few years, [yet it would never be affected owing to its] 
imperishable nature.”'*° 

[4.2] “[And] a divinity perceiving it with divine vision, in order to clean [it], 
would tell somebody: 

‘Here is gold of excellent value! Clean [it] and use it [in the way gold] 
is to be used!?”"?! 

[4.3] “In the same way I [can] see that also all sentient beings have for a long 
time been constantly overpowered by defilements, [but] knowing that 
their defilements [are only] accidental (a@gantuka), [I] teach the Dharma 
with [appropriate] means in order to purify [their] intrinsic nature 

defilements, [they] all attain perfect awakening [and] perform the acts of a buddha.” Ch;: “Sons of 
good [family], [in] all sentient beings [there is] the true nature (74) of the tathagatas, [just like] a 
truly excellent jewel, immersed in defilements [just like] putrid filth. Therefore, in order to remove 
all defilements from sentient beings [just like from the nugget of gold the person removed] putrid 
filth [and] impurities, the Tathagata, Honourable One and Perfectly Awakened One teaches the 
wonderful Dharma so that [sentient beings] become buddhas, appear in the world [and] perform 
the acts of a buddha.” In the Chinese the following four basic elements can be separated: 

a. The existence of the dharmata within sentient beings (Ch;). 

b. The teaching activity of the Tathagata. In Ch; the exact relation between teaching and the 

purification process cannot be determined unambiguously. In the Tibetan, on the other hand, the 

purification is clearly the aim of the teaching (7ib: bsal ba’i phyir; Bth: med par bya ba’i don kyi 

Phyir). In this respect the Tibetan is in accordance with RGV 1.110: ... tatklesapankavyavadana- 

hetor dharmambuvarsam vyasyjat ... and RGV 1.111: ... dharmam adisat tacchuddhaye.... 

c. The removal of the defilements. 

d. The subsequent activity of living beings after the removal of their defilements, i.e., bestowing 

the buddha-acts on others. This last element is not found in the Tibetan. 

The particle of comparison: de ltar appears only in 7ib. Bth reads ‘di ni instead. 

39 Instead of avindsadharmin Bth reads avinasadharmata; Ch, simply has 7<##. Contrary to 
the verse RGV 1.108 (quoted above), 7ib and Ch, (against its own prose) show a concessive 
construction in pada d (kyang; Hf). Alone Bth keeps to the same structure as in its prose. Neither 
Ch,; nor Bth seem to have a counterpart for Tib: de Itar in 4.1d. But cp. RGV I.108d: ... tathaiva 
tisthet.... As is the case in the prose of Ch), also here the falling of the nugget is described as being 
sudden (4%) and by inadvertence ({%) in Ch). Ch, 4.1 anticipates verse 4.2ab of Tib: “As if gold 
would be placed in impure materials, hidden [and] immersed [so that] nobody would be able to see 
[it, but] somebody with divine vision would perceive [it] and then inform a group of people about 
it:” Also this verse, like the prose, shows strong similarity with its parallel verse RGV'1.108. 

131 Yn pada b, Ch; combines padas b and c of Tib. Pada c and d of Ch) correspond to pada d and 
4.3a of Tib. Ch2 does not mention the motive of giving the directive (rnam par sbyang phyir). For 
“gold of excellent value” it simply has: <=. Ch,: “O you, if [you] take out this [gold and] wash [it] 
so that [it] becomes pure, [you will be able to] use [it] as you like [so that] all [your] relatives 
would attain blessing.” The last pada of Ch, is not found in any of the other translations. Also 
verse 4.2 of Tib is nearly identical with the parallel verse RGV 1.109. 

132 Titerally: “... knowing their accidental defilements,....” My translation above is in 
accordance with Ch, (FRG iM SSE): “... knowing that their defilements are accidental,....” Bth 
does not have a counterpart for rtag (pada b) and chos (pada d) which are both missing also in Ch). 
sMras in Bth instead of ston is the perfect form in line with RGV I.111d: ... dharmam adisat. Ch; 
continues with what corresponds to 7ib bed and adds in pada d: S238yA}PONACE: “... so that 
[sentient beings] testify (= realize) the pure tathagata-knowledge.” Like Bth, Ch; does not mark the 
cleaning of the intrinsic nature as aim: B MESA (ER: “... [and knowing that their] intrinsic 
nature [is] pure, [he] teaches with [appropriate] means....” Also here the translators or revisers of 


PDF Version: BPPB VI (2002) 

[5 The simile of a hidden treasure beneath the house of some poor person] 

“Sons of good family, again it is as if in the earth beneath a storeroom in the 
house of some poor person, under a covering of earth seven fathoms 
(purusa) deep there were a great treasure, full of money and gold, [of the 
same] volume as the storeroom.'*? But the great treasure—not being, of 
course, a sentient being, given [its lack of] a mental essence—'“[could] not 
say to the poor [man]: 

‘O man, I am a great treasure, but [I am] buried [here], covered under earth.’ 
[In his] mind the poor man, the owner of the house,'*> would consider 
[himself] poor,'*° and even though [he] walked up and down directly above 

Tib have probably added phyir in trying to structure the verse clearly. Only Ch, comprises four 
verses in 4C. Verses 4.3 and 4.4: “[4.3] [With] the vision of a sugata (#$3i) [the Tathagata] 
perceives all kinds of living beings similarly: in the dirt of defilements [there is] buddhahood, not 
[subject to] decay. [4.4] [He] teaches the Dharma according [to the situation] (f8F@; anukila) so 
that [they awake and] perform all acts [of a buddha]. The covering defilements of buddhahood are 
quickly removed so that [they all become] purified.” Ch, does not use the characteristic 
terminology of (citta)prakyti — dgantukaklesa. The thread of the last verse of Ch, is not very clear. 
As it is in the prose, the performing of buddha-acts would be expected to be mentioned at the end 
of the verse. The buddha-acts appear in Ch, and Ch; in the prose and in the verse of Ch,. As Ch, 
contains four verses in 4C it is very likely that the buddha-acts were also part of a fourth verse in 
the Skt. original which was not transmitted by TGS. However, I cannot explain why the buddha- 
acts are not found in the prose of Tib and Bth. 

33-1) For gter chen po (Skt. mahdnidhi, mahdanidhana): Ch;: “treasure of rare 
jewels” G2 87%); Chz: “great hidden treasure” (A{AiR). 

(2) A counterpart for the first kosa does not appear in Ch, (Bth: re Ito?). In the second case Bth 
(rgyang grags) and Ch; ({&4) read kroga, an Indian unit of measurement of several hundred 
meters. Bth (‘khor rgyang grags) should be understood as “with a circumference of one krosa”; 
Chy: “in length and width equally one krosa” (iti iFF=—{B BS). 

(3) Mi bdun srid: “seven purusas deep”; one purusa corresponds to the armspan of a human (so 
according to Balbir 1993: 29, citing S. Srinivasan; cp. MW s.v.: “the height or measure of a man”). 

(4) The passage in Ch,: “It is as if there were a treasure of rare jewels [in] a poor household.” 

34 The last part of the sentence from ‘di /ta ste till yin pa’o // gives the reason why the treasure 
cannot speak. The construction in Skt. may be introduced with yathapi (see Bth: kyang ‘di Ita ste : 
dper) which can have the meaning of “... because of course, because obviously, in giving a (more 
or less evident) reason for what has just been said....” (BHSD s.v. yathapi (1)). Bth, not combinig 
ngo bo nyid with sems, seems to be smoother: kyang ‘di lta ste : dper gter chen po de la yid med 
de : ngobo nyid kyis sems can ma yin pa na :: “Because of course the great treasure does not have 
a mind and is by nature no sentient being,....” Ch gives another reason: |\b7# ex: “Because [the 
treasure] is covered by earth,....” This statement (not as a reason) is part of the following direct 
speech in the Tibetan translations. That the idea of a treasure with intelligence is not at all absurd 
in the classical Indian context is shown by a discussion of Katyayana in which he states: sarvam 
cetanavat (see Thieme 1984: 135ff.). 

35 Tn contrast to OF where Skt. grhapati is rendered with khyim bdag meaning “nobleman,” 
with khyim gyi bdag po here the translators express a different understanding as implied in my 
rendering. I am not sure in which sense the Sanskrit should be understood. Accepting also here the 
meaning “nobleman” the “storeroom” in 5A.1f. could thus well be an empty treasury of a 
(formerly) rich nobleman. Also of interest is the function of the grhapati as one of the seven 
jewels of a cakravartin: it is the grhapati’s function to discover hidden treasures with his divine 
vision (which the grhapati in the TGS obviously lacks) and to make them part of the ruler’s 
possession given that there is no other proprietor. 

36 Cf SP 108.1ff. where it is said about a poor man working for a rich person without knowing 
that this person is his own father: 

atha khalu bhagavan sa daridrapuruso paryayena tac ca tasya grhapateh prabhutam 

hiranyasuvarnadhanadhdnyakosakosthagaram samjaniyad atmand ca tato nihsprho 


PDF Version: BPPB VI (2002) 

the [treasure], he [could] not hear of, know of, or perceive the existence of 
the great treasure beneath the earth. 2’ 

Sons of good family, in the same way, [in] all sentient beings, beneath 
the[ir] thinking, [which is based on] clinging (abhinivesamanasikara)— 
[and] analogously to the house’**—1there is [also] a great treasure, [namely] 
the treasury'*’ of the essence of a tathagata (tathdgatagarbha), [including 
the ten] powers (/dasa] baldani), [the four kinds of] self-assurance ({catvari] 
vaisaradyani), [the eighteen] specific [qualities of a buddha] (/astddasa-] 
a@venika[-buddhadharmah/]), and all [other] qualities of a buddha.'*° 

bhaven na tasmat kirhcit prarthayed antasah saktuprasthamilyamatram api tatraiva ca 

katapalikuficikayam vasam kalpayet tam eva daridracintam anuvicintayamanah // 

In this passage daridracinta should be interpreted as “thoughts [characteristic for] poor [people]” 
or as an iti-compound in the sense of “thoughts [like ‘I am] poor.” In any case, the thoughts of the 
son are unjustifoed as he in fact is the legitimate son of the rich person. The terminology of the 
TGS: *daridracittena anu(vi)cintayet should be understood in the same way as the context is of 
similar nature. See also SP 108.6; 115.2. Chz: -()SEBSIR. RE. BIS, A Bt. : “[His] 
mind embraces poverty, anxiety, grief [and] pain, thinking day and night.” 

137 (1) The above two passages in Ch,: “The treasure cannot say: ‘I am here!’ And since [the 
nobleman] himself does not know [about its existence] and nobody would tell [him about it, he] 
could not disclose this rare treasure.” 

(2) See Chandogya Upanisad 8.3.2 for a similar description: “But just as one who does not know 
the spot, does not find the gold-treasure [hiranyanidhi] concealed under it, although he again and 
again goes over [upary upari samcarantah] that spot, so also all these creatures do not find the 
world of Brahman, although they enter into it every day (in deep sleep); ...” (translation in Paul 
Deussen, Sixty Upanisads of the Veda, vol.1, transl. from German by V. M. Bedekar and G. B. 
Palsule, Delhi: Motilal, p. 193f.; My thanks are due to Mr. Martin Delhey and Dr. Ulrike Roesler 
who called my attention to and located this passage.) 

88 The Sanskrit for khyim Ita bur gyur pa could be *nivesanasthdniya. The sequence of 
niveSana and abhinivesa before could have been intended in the Skt. text to underline, by a play 
upon words, the parallel between the nobleman bound to the house, and living beings caught in 
their negative way of thinking. The analogy is missing in Bth and Ch;. Ch, shows a understanding 
different from Tib: —WAiA(E ... SH, MA ... ER, ....: “All sentient beings live in the 
house of ... and yet they have the store of all buddha-qualities....” Obviously Ch, has interpreted 
sthaniya in the sense of “to be (in).” Amoghavajra probably had in mind the simile of the buming 
house of the SP where the house is compared with sarnsara (see SP 72.1ff.). 

139 mDzod kyi gter chen po appears in Bth as mdzod // gter chen po bzhin gyur te :: “a treasury 
similar to [that] great treasure.” 

4° The groups of qualities mentioned here characterize a buddha. They are dealt with in e.g. 
RGVV 91.14-97.16; see also MVy 119-153 and BHSD s.v. avenika, dasabala, vaisaradya. In the 
above enumeration, 7ib has analyzed the compound avenikabuddhadharma as a dvandva: ma 
‘dres pa dang / sangs rgyas kyi chos. However, in 5B.13 Tib interprets it correctly as sangs rgyas 
kyi chos ma 'dres pa. Bth analyzes the compound in both cases as in Tib 5B.13. As Bth (de bzhin 
gshegs pa’i yeshes dang : stobs dang :), Ch, (R028 514), and also Chy (A024 %*) demonstrate, 
the original wording can hardly have been tathagatagarbha (Tib: de bzhin gshegs pa'i snying po). 
All versions except Tib are based on tathagatajfana. Tib itself has ye shes instead of snying po in 
the parallel in 5B. We must thus conclude that already the Indian manuscript on which 7Tib was 
based, had the reading -garbha instead of -jfiana, or that the newly introduced reading garbha is 
due to a mistake or even a deliberate alteration by the translators of Tib. However, the words 
tathagatajnana and tathagatagarbha are metrically equal in Skt. and, especially if we assume an 
oral tradition, confusion between the two terms could easily be explained. A deliberate alteration 
by the translators of Tib is also not very likely because this would probably have included the 
alteration of the parallel passage in 5B. Following the supposed original Skt. wording with -/fana, 
as it appears in 5B, we should understand the “tathagata-knowledge” as another member in the 
enumeration of buddha-qualities. Ch;2 do not mention the dvenikabuddhadharmas. Instead of 
tathagatajfidna, Ch, has 40385115 (*tathagatajndnadarsana). 


PDF Version: BPPB VI (2002) 

And yet sentient beings cling to color and shape (rupa), sound 
(sabda), odor (gandha), flavor (rasa) and tangible objects (sprastavya),'! 
and therefore wander in samsara, [caught in] suffering (duhkhena). And as a 
result of not having heard of that great treasure of [buddha] qualities [within 
themselves, they] in no way apply [themselves] to taking possession [of it] 
and to purifying [it].”! 

“Sons of good family, then the Tathagata appears in the world and manifests 
(samprakasayati) a great treasure of such [buddha] qualities among the 
bodhisattvas.'“? The [bodhisattvas] then acquire confidence in that great 
treasure of [buddha] qualities and dig [it] out. Therefore in the world [they] 
are known as ‘tathagatas, honorable ones and perfectly awakened ones,’ 
because having become [themselves] like'“* a great treasure of [buddha] 
qualities, [they] teach sentient beings the aspects of [this] unprecedented 
argument [of buddhahood in all of them] (*apurvahetvakdra), similes 
[illustrating this matter], reasons for actions, and [tasks] to fulfill.’ [They] 

141 Tn the Buddhist enumeration of the five sense objects Sprastavya is usually translated as reg 
bya (cf. MVy 1863). However, the Tibetan version of the LAS has also reg pa (cf. LAS, s.v. 
sprastavya) just as is the case in Bth. 

182 Only Tib arranges the purification in a finality clause (yongs su sbyang ba’i phyir). The last 
two passages in Ch,: “[They] do not hear nor recognize [the treasure, they] take pleasure in and are 
mislead by the five desires, wander around in samnsara [and] experience measureless pain.” In Ch, 
the “wandering in samnsara” is omitted. The reason for not hearing of the treasure is the clinging to 
the five desires and the pain involved (... Sf. B. BB. GR. ARSE, VE. ). The final part 
in Ch, runs as follows: “Therefore they do not hear of the great treasure of [buddha] qualities, let 
alone that it could be attained. If [they] eradicate those five desires [they] will attain purity.” 

3 Ch, speaks of “several buddhas” (38/8) and does not mention the bodhisattvas. Instead of “a 
great treasure of such qualities” (alternatively: “such a great treasure of qualities”) Ch, reads 
AiG BM: “great store of various jewels[, i.e., the buddha] qualities (dharmas) (?).” It goes 
without saying that the twofold meaning of Skt. dharma can be applied throughout 5.A-C. 
Whereas Ch, continuously understands dharma in the sense of “qualities” of a buddha, the Tibetan 
and Ch, allow an ambiguous interpretation. See, for example, the use of the verb samprakasayati 
in this passage which could mean the concrete opening of the treasure (Ch,: 64) but also the 
disclosing of the teaching. This room for interpretative ambiguity may already have been part of 
the Skt. original. 

14 (1) For ita bur in Tib there is no parallel in the other versions. Instead, Bth has gdon' 
miza bar [' for gnod; metathesis] for *avasyam; Ch: F.... 

(2) According to Matsumoto 1994: 515ff., this passage proves that the original intention of the 
illustration is to identify the living being with the treasure (sattva = nidhi). All other passages 
mentioning the treasure within living beings (sattva # nidhi) display, according to Matsumoto, a 
later development in which nidhi as the counterpart for tathagatagarbha is already understood as a 
separate entity. I find Matsumoto’s position philologically difficult to maintain and not necessary. 
If we consider as the counterpart for nidhi in the upameya the tathagata within living beings (as 
expressed in 5.3) the compound tathagatagarbha could still be interpreted as a bahuvrihi referring 
to sattva. 

145 What the terms of the enumeration in detail stand for, is unclear. Similar passages in the SP 
point to the context of teaching living beings with different dispositions under the topic of 
updyakauSalya. For example: 

drstantahetun bahu darsayanti bahukaranan jianabalena nayakah / 

nanadhimuktams ca viditva sattvan nanabhinirhar’ upadarsayanti // 54.1f. = 11.107, 
Further: .... nanabhinirhara-nirdeSa-vividha-hetu-karana-nidarsanarambana-nirukty-upaya- 
kauSalyair nanaddhimuktanam sattvanam ... tathagato ... dharmam desayati / (71.7ff.; very similar 
also 41.2ff.; 41.12ff.; 42.14f.; further 39.11f. and 45.5f. = II.44). 


PDF Version: BPPB VI (2002) 

are donors (danapati) [who give from] the storeroom of the great treasure, 
and having unhindered readiness of speech (asangapratibhanavat), [they 
are] a treasury of the many qualities of a buddha, including the [ten] powers 
and the [four kinds of] self-assurance.'“° 

Sons of good family, in this way, with the completely pure vision of a 
tathagata, the Tathagata, the Honorable One and Perfectly Awakened One, 
also perceives that all sentient beings are like the [poor owner of the house 
with the hidden treasure] and then teaches the Dharma to the bodhisattvas in 
order to clean the treasury [in all sentient beings, which contains such 
qualities as] the tathagata-knowledge, the [ten] powers, the [four kinds of] 
self-assurance and the [eighteen] specific qualities of a buddha.”'*’ 

Then at that time the Exalted One uttered these verses: 
[5.1] “[It is] as if beneath the house of a poor [man] there was a treasure full of 
gold and money, in which neither motion nor thinking'** was existent 
and [which could] not say: ‘I am yours!” 

(1) sNgon ma byung ba’i gtan tshigs kyi rnam pa, Skt. *apurvahetvakara, appears also in Ch; 
(REE [Al4R). Bth: rim kyi thog mar rgyu dang : rnam pa could render the same Skt. compound 
(with *anu-purva® instead of a-piirva®) interpreted as a dvandva. Similar expressions occur in the 
MPNS and the Suv. There, with hetu they also designate the “arguments,” or less strictly, the 
“contents” of the siltra: Suv 126.4: ... ndndvidhdni sitrantahetani (= Suv, 97.5-6: mdo sde’i gtan 
tshigs rnam pa sna tshogs); MPNS cited in Shimoda 1997: 271, n. 65 (p. 597): ... sems can thams 
cad la de bzhin gshegs pa’i snying po yod do zhes bya ba’i gtan tshigs dang rtags rnams kyi sgo 
nas... (Chinese (Faxian £88): ... [Ki@e+H5K....). For Tibetan gtan tshigs (in contrast to rgyu) for 
hetu cf. Ja .v. gtan tshigs: “1, argument...” 

(2) Byed pa’i gtan tshigs (5B.6), Skt. *kdranahetu, remains difficult (see MVy 2260: byed pa’i 
rgyu for karanahetu). The term appears in a classification of hetus in LAS 83.9, 14; see also 
AK(Bh) IL.50. The SP only knows the compound -hetukarana-, Bth reads dgongs pa instead. 

146 Instead of stobs dang / mi ‘jigs pa dang / sangs rgyas kyi chos Bth reads simply shes rab, 
Skt. prajfia. Ch2 already here enumerates all the elements mentioned only at the end of 5B in the 
Tibetan. After mentioning that the bodhisattvas have acquired confidence Ch, continues: “... then 
[they] dig [for the treasure, i.e.,] enter the bodhisattva stages. [Finally as] tathagatas, honourable 
ones and perfectly awakened ones [they] function as stores of the Dharma [in] the world. [They] 
perceive [in] all sentient beings the aspect of the unprecedented cause [for buddhahood] 
(KEE [K48; *apirvahetvakara). Therefore [they] teach the great store of the Dharma in similes, 
[they] become great donors, storerooms of unhindered readiness in speech, of measureless 
knowledge, the [ten] powers, the [four kinds of] self-assurance [and] the [eighteen] specific 
qualities of a buddha.” The passage in Ch,: “After those [living beings?] have full of confidence 
attained the pure knowledge of all [matters], they widely disclose the store of a tathagata for living 
beings; [they have] unhindered readiness in speech [and] act as great lords of the giving [of 
donations] (mahadanapati).” 

47 The enumeration of buddha-qualities is missing in Ch, (instead: 412K#). In Bth and Ch, 
there is no equivalent for de Ita bur in 5B.11. The aim of the purification (sbyang ba’i phyir) is not 
expressed in Bth (scribal mistake?). Therefore: ... mdzod yongsu sbyangs pa’i byangchub sems 
dpa’ rnams: “... to the bodhisattvas [whose] treasury of ... has become pure.” Ch, as well as Ch; 
show the following logical relation: because the Tathagata perceives all living beings as having the 
store of a tathagata (412K¢#R{), he teaches the Dharma to the bodhisattvas. 

48 The word pair g.yo ba — rlom sems, Skt. ifijand — manyand, points to the second level of the 
simile: the negated terms can characterize the mind of somebody who has entered the state of 
absorption (e.g. SP 5.10f.: anifijana) or the mind of a tathagata (see RGVV 9.10ff). The word pair 
is missing in Ch,. For manyand the BHSD shows the following explanation: “conceit in the sense 
of vain, illusory imagining....” However, Tibetan rlom sems does not comprise this meaning (cf. Jd 
s.v. rlom pa: “rlom sems pride, arrogance.”) Here, however, the term *amanyand probably simply 
refers to the fact that the treasure is without “a mental essence,” as mentioned in the prose. 


PDF Version: BPPB VI (2002) 

[5.2] “At the same time a sentient being, the owner of the house, had become 
poor. But because [he] would not know [of the treasure] and [because] 
there was nobody [who] had informed him, the poor [man] would 
remain above the [treasure without digging it out].”!°° 

[5.3] “In the same way, with the vision of a buddha I see that in all sentient 
beings—though [from the outside] they resemble poor [men] —there is 
a great treasure; and [I see this treasure] as the motionless body of a 

[5.4] “I see that [treasure and] teach! [the following] to the bodhisattvas: 

‘O you [bodhisattvas], take the treasury of my knowledge! Act [so that 
you may] become treasures of the supreme Dharma, being free of 
poverty and becoming protectors of the world (lokanatha)\’”'? 

4° Tn Chy) and Bth it is not said that the treasure is beneath the house. However, only Ch, 
speaks of a treasure in the house also in the prose. Instead of a treasure (gter) Bth mentions a 
vessel (snod). Ch,, parallel to its own prose, does not describe the treasure as filled with gold and 
money. 5.lcd in Ch, run as follows: EEE S15, MARES: “Since the master [of the house] 
would not know about [nor] perceive [it], and the jewel [treasure] also could not speak, [5.2]....” In 
Ch the treasure cannot say the following: “This [here] is a certain thing!” (23£4). 

150 Bh differs from the other translations in pada b: mi shes pas na dbus gyur pa : (dbus several 
times for dbul): “As [the nobleman] did not know [about the treasure, he] became poor.” Padas a 
and b in Chy: RARBRES BAZ MAA: “Though this person further is the master [of 
the house, he] experiences poverty and does not know [about the treasure].” In pada d, Ch, repeats 
its on statement of pada b: ZAR (EAIE: “... and [he] experiences poverty [and] dwells in 
sorrow.” In pada c it seems that subject and object have been confused: {7K S2WIAJER A: “That 
[master?] also does not inform anybody.” The whole verse in Ch,: “Poverty-stricken years [his 
mind] embraces dullness (= ignorance about the treasure). There is nobody telling [him the facts. 
Though] there is a treasure, yet [he] does not know [about it.] Therefore [he] continues to be in 
poverty and sorrow.” The statement of Ch, is also found in the parallel verse of the RGV (1.113cd) 
referring to living beings in the upameya: abudhyamananubhavaty ajasram daridryaduhkham 
bahudha prajeyam //: “[In the same way, because living beings] do not recognize [the treasure 
within themselves, they] perpetually experience suffering in form of poverty (or: the suffering [of 
samsara] similar to poverty) in manifold ways.” Cf. in particular the compound daridryaduhkha 
with Ch: @#. 

'S! (1) Both Chinese translations differ in pada b. Ch,: S¢7{48 HH: “... [see that living beings], 
though circulating [in] the five paths [of existence].,....” Ch2: —W@lH RISE: “... [see] that all 
sentient beings dwell in poverty.” The comparison between living beings and the poor is only 
stated in Jib (Bth has gyur pa instead of ‘dra; cf. n. 128). See RGV I.114c: ... sattva 

(2) Sugata is another of the standard set of epithets of a buddha. For traditional interpretations of 
sugata see MPPU,, 1.131f. Both understandings, namely “[one who] has well spoken [the 
Dharma]” as well as “[one who] has perfectly entered [nirvana],” are discussed there. See also 
Griffiths 1994: 105. 

(3) 5.3d in Jib is grammatically difficult; literally: “... [and I] see [this treasure] as the motionless 
and sugata[-like] body(?).” Pada d as found in Ch; (see below; the reading gains support by Bth) 
seems more appropriate. Ch, does not mention the body of a sugata. Instead it is said that the 
treasure is always in the body (of living beings), without alterations (KBESW BETS). 
Padas cd of Ch, (SHAK Ki FERGIE B®): “[I see that] in the bodies [of sentient 
beings] there are great hidden treasures: there is the essence of the buddhas, motionless.” 

82 The perfect form of ston pa (Bth: smras) might have been employed due to the temporal 
relation between the teaching activity and the following acquisition of confidence (described in 
5.5) in this teaching. 

153 Ty pada c a great treasury of knowledge is mentioned in Bth (yeshes kyi ni mdzod chen) and 
Ch, (K#®), whereas Tib adds the personal pronoun of the first person (mama instead of maha- 
?). A further parallel between Bth and Ch; is the verb “to dig out” (thon, imperative of ‘don pa; =) 
instead of “to take” (zung). Pada d in Ch2: BERRAE LZIER: “.. [so that you] may grant the 
supreme wealth of the Dharma/the wealth of the supreme Dharma.” Also, Ch, and Bth mention 


PDF Version: BPPB VI (2002) 

[5.5] “Whoever acquires confidence in [this] my teaching—in each of those 
sentient beings is a treasure. Whoever, having acquired confidence, 
exerts himself will quickly attain excellent awakening.”!* 

[6 The simile of a sprout in the seed] 
“Sons of good family, again it is like the example of a fruit of a mango tree, a 
rose apple tree, a palmyra palm or of cane:! inside the sheaths of the outer 

“wealth” (XH; nor) and the element should thus be considered as original. Pada d of Bth is 
similar to Ch,: bla na med pa chos kyang nor yang sbyin : : “... [so that you may] grant the 
supreme Dharma, as well as wealth.” The difference between Ch and Bth might be explained by 
supposing a different interpretation of a compound *anuttaradharmavasu or *anuttaradharma- 
dhana. The complete verse in Ch,: “Having observed thus [the Tathagata] teaches for living beings 
so that [they] may attain the treasury of knowledge {leading to] great wealth and wide profit.” The 
bodhisattvas remain unmentioned in Ch. 

‘54 (1) It is very unlikely that Tib has here preserved the original intention of the Skt text. Brh 
does not show any demonstrative pronoun in pada b that takes up the relative pronoun of pada a, 
nor is there a particle ending the sentence at the end of pada b. The demonstrative particle appears 
only in pada d of Bth which accounts for the following translation: “[Those] who acquire 
confidence in that [what] I have taught, {namely that] in sentient beings there is even a treasure, 
who acquire confidence and apply themselves [to its excavation], they will quickly attain even 
excellent awakening.” 

(gang cig nga’is' bshad pa de la mos // 

sems can la ni gter yan, yod : 

gang zhig shin tu mos shing brtson byed pa /I 

de ni byangchub mchog kyang myur rnyed do // 

[' nga’is for nga’i sa in the ms]) 
Also the Chinese versions make such a rendering most plausible as the original intention of the 
verse. Pada c in Chy: “..., [and] practices [with adequate] means (upaya) faithfully and 

(2) bDag nyid in Tib 5.5c remains without any counterpart in the other three translations. 

155 (1) All of the plants mentioned are widely known in India and often mentioned in the 

a. (Shing) a mra, Skt. dmra: the mango tree (mangifera indica) is a huge, evergreen tree, whose 
fruits belong to the best known in India. Inside the fruit there is a massive kernel which produces 
the sprout (see McCann 1959: no.25). 

b. ‘Dzam bu or dzam bu, Skt. jambu: also the rose apple tree (eugenia jambolana) is very well 
known in India and is considered the center of the continent jambudvipa, one of the continents 
believed to constitute the world. The rose apple tree can reach more than 25 meters in height. Its 
fruits are longish and about 2.5 cm in size. When the fruits become mature they look similar to 
black olives and are eaten by humans and animals or processed into juice. In relation to the fruit 
pulp around, the kernel inside the fruit is relatively big. See Syed 1990: 288-302; McCann 1959: 

c. Ta la, Skt. tala: also the palmyra palm (borassus flabellifer) can reach more than 25 meters in 
height. In the old texts it serves as a measure unit for heroes and other imposing figures. The 
palmyra palm grows slowly and flowers for the first time after 12 till 15 years. Stem and fruits 
are hard. the fruits are eatable but not tasty. Above all, the tree’s stem and fan-shaped leaf crown 
are used as materials. The fruits are found grouped at stems with a diameter of max. 15 cm each. 
Shielded by a peel like leather, inside the fruit there are two or three seeds again covered by 
fibrous material. The seeds are about 5 cm in diameter. See Syed 1990: 308-325; McCann 1959: 

d. sPa, Skt. *vetra: Jé gives the meaning “cane” for spa and mentions examples where spa in 
compounds means “bamboo.” In MVy 4217 the Skt. equivalent vetra is found for sba (acc. to Ja 
another spelling of spa). Bth reads ‘deb ‘dre (prose) and beb/ bab tra (verse) instead of spa 
which could be a transliteration of vetra. (The transliterations ’ba or dba for Skt. va in Bth are 


PDF Version: BPPB VI (2002) 

peel there is a seed of imperishable nature (*avinasadharmin) [containing] a 
sprout, [a seed] which, thrown on soil, will become a great king of trees,)*° 
Sons of good family, in the same way also the Tathagata perceives 
that [sentient beings who are] dwelling in the world'*’ are completely 
wrapped in the sheaths of the outer peel of [such] defilements [as] desire 

not uncommon: cf. ’Ba’ gi ya sha for *Vagisa, and u ru dbyil ba for uruvilva. Difficult to 

explain is in fact -b at the end of the first syllable.) The pW has “eine grdéssere Art Calamus, etwa 

fasciculatus, zu Stécken gebraucht” for vetra and knows the compound vetraphala. Ch; confirms 
vetra (ff and i (“wisteria”) seem to be not strictly distinguished): “reed, rattan, cane.” Takasaki 

(1981: 23) translates spa as “pine tree” (#{X). The fact that cane or bamboo is conceived of as a 

“great king of trees” is not surprising as it can reach a height of 40 meters and its fruits, in case 

of the melocanna bambusoides, a kind of bamboo, are described as similar to apples (Meyer s.v. 

Bambusgewachse). Also worth mentioning is the enormous speed of growing bamboo of up to 

one meter on a single day. 

(2) Bth mentions another fruit: ba sa na’i ‘bras bu. Its identification is not clear. Ch, here and in 
6.1 only speaks of a mango fruit (#47252). Ch, enumerates the plants in the opposite order. In case 
of the mango and the rose apple tree Ch, speaks of “the seeds of the fruits” (42+) whereas for the 
palmyra palm and cane it only reads “seed” (+). This differentiation is not found in the Tibetan. 

156 (1) Chy; do not mention the “sheaths of the outer peel.” Bth does not describe the peels as 
“outer”; shun phrag should probably be emended to shun pags (cf. Jd: id. [with shun pa)). 
However, in 6.1b Bth also speaks of “outer peel”: phyirol shun{i] phrag. 

(2) The compound bijankura (“‘sprout [in the state] of a seed”) of RGVV 63.20, RGV I.115b and 
1.117.c is not found in 7ib; instead: myu gu’i sa bon, Skt. ankurabija: “seed [whose characteristic 
is an encapsulated] sprout”; Ch): =-2F for bijankura? Bth reads sa bon dang : myu gu and adds 
‘byung bar ‘gyur ba yang (in Ch2: $-2F FR #8404:; missing in 71b). The compound in Bth as well 
as in Ch (for a translation see below) must be understood as a dvandva. Bth: “In spite [of the fact 
that the tree appears different,] emerging [alternatingly as] seed and sprout, [it is] of unperishing 
nature(?).” Takasaki’s rendering of ankurabija in Tib with F & 77: AX fH (1981: 23f.): “seed 
which will become a sprout” suggests that the seed turns into a sprout only in a later step. This is 
in contradiction to verse 6.1 where the sprout is clearly described as already within the seed (... spa 
yi ‘bras bu...// nang na spa yi myu gu yod pa ste //). The section in Ch;: “Just as the kernel inside a 
mango does not perish, [and when one] sows it into the earth, [it] becomes a great king of trees,....” 
Ch; ... HEF FRBME, RMR, ab, BME, AMABILE. : “[ust as the 
seed of ... a mango] realizes the quality of indestructibility as a result of the fact that the seed and 
sprout generate each other alternately (£48; *paramparyena), [and the seed,} when meeting with 
the conditional factor earth (44), sown into it, after a long time becomes a great king of trees,....” 
(For # see RGV I.116c: ... tat tat kusalam pratitya ... and 1.117d ... subhapratyayaih //.) The 
content of 7ib differs from Bth and Ch; in that it omits the passage that seed and sprout appear in 
alternation. Due to the fact that this statement appears in the verse section (6.1d) of Tib, though in 
a different form, it can be assumed that it was in fact also part of the prose of TGS;. However, the 
statement is also missing in both the prose and verse section of Ch,. In light of the statement of the 
mutual alternating generation of seed and sprout it seems that Bth and Ch, put their main accent on 
the eternal sameness of the tree’s nature in spite of manifesting itself in different shapes. This 
aspect appears to carry more weight than the future development into a king of trees. 

57 Probably due to the significant position of the term sattva in the following section, Tib 
paraphrases sentient beings here with ‘jig rten na gnas pa; Bth: ... mdzod du ‘jig rten zhugs shing 
gnas pa': “.., [perceives] that living beings (/oka) have entered and [now] dwell in the sheaths...” 
or (parallel to 71b): “... [perceives sentient beings who] have entered and are [now] dwelling [in] 
the world, [covered with] sheaths....” The verb mthong has probably been omitted by the scribe as 
the line ends with gnas pa. However, Ch, mention sentient beings and do not have an equivalent 
for jig rten, Skt. loka. It is also remarkable that only 7ib does not mention the Tathagata’s vision. 
This could mean that the intention of the translators or editors of Tib was to read the text as: “... in 
the same way, also the tathagata residing in living beings (loka) is perceived as wrapped...” 
However, in the 7GS, this would be a rather uncommon formulation and can hardly be based on 
the Sanskrit original. 


PDF Version: BPPB VI (2002) 

(raga), anger (dvesa), misguidedness (moha), longing (trsnd) and ignorance 


“In this [connection] the true nature (dharmatd) of a tathagata, being in the 
womb (garbha)'*’ inside the sheaths of [such] defilements [as] desire, 
anger, misguidedness, longing and ignorance, is designated’ ‘sattva.’ 
When it has become cool, it is extinct (nirvrta). And because [it is then] 
completely purified [from] the sheaths of defilements of ignorance,'*' [it] 
becomes a great accumulation of knowledge [in the] realm of sentient 
beings (sattvadhatu).' The world with [its] gods (sadevako lokah), having 
perceived that supreme, great accumulation of knowledge [in the] realm of 
sentient beings speaking like a tathdgata, recognizes ' [him] as a 

158 Ch): “In the same way, sons of good [family], with the vision of a buddha I see that [in] all 
living beings the store of qualities of a tathagata is found [in] the peels of ignorance, just as the 
seeds of a fruit are found inside the kernel.” As does RGV 1.116, Ch, does not consider living 
beings but the buddha-qualities (RGV: dharmadhatu) to be wrapped. Ch): 
RPRELRIARIR AHO AS .. SERS. : “.. the Tathagata, with the vision of a 
tathagata, perceives all sentient beings [with their] defilements ... even to the end of the peels.” 
The last part of Ch; is not intelligible. The parallel verse RGV 1.117c in this passage reads -klesa- 
phalatvagantaragatah. Instead of antaragata (Bth: zhugs shing gnas pa; Tib: kun tu dkris pa?) Ch) 
has possibly read (pary-)antagata G&S). 

9 Tib alone suggests the following understanding: “In this [connection] the true nature of the 
tathagatas, having become the essence inside the sheaths....” or even (Takasaki 1991: 24): fg VEO 
TLERBIL HS POMHOAPE: “ ... that real nature of a tathagata in an embryo-like state.” 
sNying por gyur pa is probably a rendering of Skt. garbhastha or garbhagata (see Bth: ... dbus’ kyi 
snying por de bzhin gshegs pa’i chos nyid kyang gnaso // (' for dbul]). By the analysis in Hara 
1994: n. 4 (1) (though referring for the most part to the Mahabharata) it becomes clear that 
garbha- in these compounds must be understood as “womb.” This meaning should therefore also 
be accepted for the passage under discussion. See the parallel RGV I.117c (-phalatvagantaragatah 
sambuddhabijankurah), where -gata is also used to point out the place (and not the state) in which 
the sprout can be found. See the passages in 8B.4 and 8.5 and also Matsumoto 1994: 503f. 

‘© The construction zhes bya ba'i ming du chags pa is probably a translation of Skt. 
samjnotpadita (see Suv, s.v. ming). In this way also Bth’s ‘du shes pa could be. explained by 
supposing that the verb remained without translation. 

'S! Ma rig pa, Skt. avidya, is considered the root of other defilements (see e.g. Ybhi 166.16: 
tatra viparyasamilam avidya /). 

1° The Skt.,equivalent for ... tshogs su gyur pa gang yin pa de ni rnyed pa’o // and Bth: ... 
phung po nyid rnyed pa’o // could be *yo -sambharabhdvas sa praptah, rendered lit. as: “... that 
which is the (state of] being an accumulation..., is attained.” For -prdpta at the end of a compound 
with an abstract noun before cf. BHSD s.v. -prdpta. The construction here reminds one of a similar 
passage in the rendering of the TGS verses in RGV I.116 concerning the same simile (... 
Subhadharmadhatuh upaiti ... munirajabhavam //): “... the pure dharmadhatu becomes king of the 
munis.” Ch differs here (see below). 

13 For ‘du shes, Skt..samjfid, in this meaning see BHSD s.v. samjfia (5). Samjid comprises 
mental acts like imagination, identification, ideation or even interpretation. In ‘du shes su byed 
(*samjnam kurvanti) I do not understand the function’ of su. (8th without su: ‘du shes byed do //). 

14 The versions differ widely and leave extensive rooni for interpretation. That the original 
intention has not been preserved is shown by the great discrepancies between prose and verses. In 
the following I will first deal with the key terms which appear, before translating and commenting 
on the varying translations. Finally I will try to sum up the contents. 

(1) The definition of sattva is only found in the Tibetan and Ch. In orthodox Buddhist 
philosophy. the term sattva is mentioned along with atman. The existence of a separate entity 
called saétva is not admitted. See e.g. ... Sinyah samskarapunjo ‘yam na hi sattvo ’tra vidyate /! ... 


PDF Version: BPPB VI (2002) 

evam skandhan upddaya samvrtya sattva ucyate // iti / (AKBh 466.2, 4 cit.); ... ndstiha sattva atma 
va dharmas tv ete sahetukah // (AKBh 466.9 cit.); ... dtmadystir bhavati sattvadystih / nirviseso 
bhavati tirthikaih sardham // (AKBh 466.14 cit.). The AAA even derives sat- of sattva from the 
root sad meaning “to perish”: sidan ’atmakatvat sattvah (AAA 81.5 with note | (the form sidana is 
attested in Pali); referring to ASPyyw 80.25f.: ... mahatyd Gtmadrstyah sattvadystyah jivadystyah...). 
The definition found in the TGS gives a different explanation, by stating that the term sattva is 
used as long as the sheaths (of klesas) have not been destroyed. Sattva is also considered one of the 
three gunas which constitute the world in Sankhya philosophical terminology. 

(2) The equation “becoming cool” or “having become cool” (sitibhdva) = nirvana is common 
(see BHSD s.v. sitibhava, Ssitibhiuta). 

(3) The compound ye shes chen po’i tshogs could be a translation of Skt. *mahdajidnasambhara. 
Mahat, as the first member of the compound, could of course also go with jadna. 

(4) Apart from the meaning “the whole of living being,” the RGV and the AAN know sattvadhatu 
also as a specific term for the dharmakaya in samsara, as synonyms for tathagatagarbha: 
RGVV 40.16ff. (cit. AAN): ayam eva Sariputra dharmakayo 'paryantaklesakosakotigidhah 
samsarasrotasa uhyamdno ’navardgrasamsaragaticyutyupapattisu samcaran sattvadhatur ity 
ucyate | 

Sariputra, that very dharmakaya, [as long as it is] covered by myriads of sheaths of 
defilements, carried away by the stream of sathsara, [and] wandering in dying and arising 
without beginning and end in the existences of samsara, is called sattvadhatu. 
The jinagarbha in an inpure state is described as sattvadhatu in verse RGV 1.47, which is part of 
the oldest stratum of the text: 

asuddho ’Suddhasuddho ’tha suvisuddho yathakramam / 

sattvadhatur iti prokto bodhisattvas tathagatah // 47 // 

[Depending on the jinagarbha being] impure, [partly] impure and [partly] pure, and 

completely pure, [it] is called respectively sattvadhatu, ‘bodhisattva’ and ‘tathagata.’ 
That the term tathagatagarbha is used identically is shown by RGVV 21.8ff.: tatra samald tathata 
yo dhatur avinirmuktaklesakosas tathagatagarbha ity ucyate /: “In this [connection] the defiled 
True Reality is what is called tathdgatagarbha [as] the essence not freed from the sheaths of 

(5) The passage in the other versions: Ch,: “Sons of good [family], that store of a tathagata is 
cool, without heat. [It is] a great accumulation of knowledge, the wonderful nirvana [and, when 
purified, it] is designated as ‘tathagata, honourable one and perfectly awakened one.”” As usual 
Ch, is considerably shorter than the other versions. The definition of sattva, the term sattvadhatu 
and the process of purification do not appear. (The corresponding verse 6.3 of Ch, differs from the 
prose in this respect.) The repetitive enumeration of klesas and the statement that the dharmata is 
wrapped in sheaths is also missing. It could constitute an element inserted later by the redactors of 
TGS). Ch;: “... in the sheaths of defilements ... there is the nature of the store of a tathagata.' As 
long as [one] does not know’ this, the designation sattva [applies]. When [one] is able to bring [the 
defilements] to rest, [this is] designated as ‘cool’ and named nirvana. When [one] is able to 
remove the defilements of ignorance, then that element of the sattva (4187); sattvadhatu) is 
designated ‘essence of the great accumulation of knowledge’ [and] such a sattva is named ‘great 
accumulation of knowledge’. When [this being called great accumulation of knowledge then] 
appears [as] a buddha in the world [with the] gods, [it] teaches the subtle Dharma [and] when [the 
world with the gods] see this, [they] call [that being] ‘tathagata.’” ['4N2K#¢M: in light of the Tibetan 
(tathagatadharmata) an emendation to 403K should be worth considering; ? Should 7{& be emended 
according to verse 6.3a (4°#8): “As long as [one] does not destroy these [sheaths],...”? The phrase 7<{% has 
no equivalent in the Tibetan prose.] 

A possible translation of Ch, has to deal with many factors of uncertainty. As does Ch;, also Ch2 
has 40783) (contrary to dharmatd of the Tibetan). Leaving aside an emendation to 4RE1, 
one reason for adding to the compound might be that Amoghavajra tried to give a quasi- 
etymology for the term sat-tva by defining it as “where the nature (‘# for Skt. -tva) of the store of 
a tathagata can be found (4 for Skt. sat from the root as).” In this case, the variant #{# rather 
than 47/8 should be considered as original (in contrast to the verse) when defining the term sattva. 
Ch; seems to use the term sattvadhdtu as a designation for the unpurified true nature of living 
beings. Once purified, this same element is named “essence of the great accumulation of 
knowledge” (K)3% #8). The characterization of sattva (Tib: sattvadhatu) as “supreme” (dam pa; 


PDF Version: BPPB VI (2002) 

Sons of good family, in this [connection] the Tathagata perceives that 
[all sentient beings]'® are like the [seed containing a sprout], and then 
propounds the matter to the bodhisattva-mahdsattvas in order that [they] 
might realize the tathigata-knowledge [within themselves].”!© 

mchog) could correspond to {£2 #4. Ch, has probably interpreted para in the sense of “another” 
or even misread para instead of vara. The phraseology =: HI|4453 does not correspond to the 
Tibetan rnyed pa’o; ji lta ba de bzhin du (Tib) or ci ltar khong du chud pa de bzhin du (Bth) 
remain without counterpart in Chp. 

(7) Bth: ... nyon mongs pa’i mdzod kyi' dbus’ kyi snying por de bzhin gshegs pa’i chos nyid kyang 
gnaso // de la gang sems can du ‘du shes pa zhi bar gyur na : ma rig pa’i nyon mongs pa’i mdzod : 
yongsu sbyangs pa’i sems can kyi khams de yeshes chen po’i phung po nyid rnyed pa’o // de sems 
can mchog ste : yeshes kyi phung po chen po’o // ci ltar khong du chud pa de bzhin du smra ste : 
lha dang : ‘jig rten du bcas pas mthong nas / de bzhin gshegs pa zhes ‘du shes byed do // 

[' for kyis; ? for dbul]. 

Now, the true nature of a tathagata is found as the essence, [being in] the center of the 

sheaths of defilements.... When in this [connection a being] which is called sattva has 

calmed down, the essence of the sattva (sattvadhatu), completely purified [from] the 

sheaths of defilements of ignorance, becomes a great accumulation of knowledge. Such [a 

being] is a supreme sattva, [as the] accumulation of knowledge is great. [This supreme 

sattva] speaks as having understood [the Dharma], and after the world with the gods has 

seen [this, they therefore] recognize [this being] as ‘tathagata.’ 
Bth combines the definition of sattva with the following sentence and does not leave room for an 
identification of the dharmata with sattva. The term nirvana does not appear. The unusual ji /ta ba 
de bzhin du (6B.7) of Tib could represent a shortened form of ci /tar khong du chud pa de bzhin 
du as it is found in Bth. The same wording appears e.g. in KP §2, where it is said about the 
bodhisattva: yathasrutams ca dharman yathaparyaptan parebhyo vistarena samprakasayati / = ... 
ji ltar khong du chud pa... 

(8) In light of the other versions of TGS, Tib shows the following characteristics: The equation 
“defiled dharmatd = sattva” is only found in the prose of Tib. Even verse 6.3 of Tib does not 
confirm this explicitly. Ch, seems to see different categories in the sattva itself as an ordinary 
living being and in the element or nature of this living being (sattvadhatu), and thus avoids a direct 
equation. The genitive relation between sattvadhatu and the “great accumulation of knowledge” in 
Tib is not confirmed by the other two versions where it is said that the sattvadhdatu becomes (the 
essence of) a great accumulation of knowledge. However, in 7ib it is hard to understand 
sattvadhatu as anything else than “realm of sentient being.” The interpretation of sattvadhatu as 
“element, essence” is supported by RGV I.116ab, where the term subhadharmadhdtu is used 
instead (sattvesv ... Subhadharmadhatuh). Within the reproduced verses of the TGS in the RGV the 
term dharmadhdtu appears only once. The second mention of sattvadhdatu in 6B.6 is probably not 
based on the Skt. text since Bth (sems can) and Ch) (4) simply read sattva. 

(9) The content of the section in TGS, thus focuses on the following four main points: 

a. The definition of the term sattva as characterizing the true nature of living being enclosed by 


b. The statement that this nature has become cool and extinct. 

c. Though already extinct, the external purification from the defilements leads to the nomination 

“great accumulation of knowledge.” This stage is associated with the attribute “supreme” (dam 

pa; mchog). 

d. Only when teaching the Dharma to other sentient beings, is such a “supreme being” granted 

the nomination “tathagata” by them. 

15 The passage in Ch; seems corrupt: “Sons of good [family], when there the Tathagata, 
Honourable One and Perfectly Awakened One is seen, [this same Tathagata] makes all 
bodhisattva-mahdsattvas aware of [their own] tathagata-knowledge so that [they] may disclose [it 
to sentient beings].” The interpretation of the Tathagata as the object of the act of seeing differs 
from all other versions. In the last part, artha, being the “matter” to be exhibited according to Ch, 
and the Tibetan, has been understood by Amoghavajra as indicating finality (45 BAER 0). 


PDF Version: BPPB VI (2002) 

Then at that time the Exalted One uttered these verses: 

[6.1] “Just as all the fruits of cane have a cane sprout inside [their seeds], and 
[just as @ sprout] is also in all the [fruits of] palmyra palms and rose 
apple trees: when the result,’ [Which is already perfectly] contained 
within [the fruit’s outer peel,] is made to germinate, [a great tree will] 

[6.2] “In the same way also the Master of the Dharma, the Leader,'® perceives 
with the supreme, uncontaminated vision of a buddha’ that in all 
sentient beings without exception—similarly to the cane seed—there is 
the body of a sugata.”!” 

[6.3] “The [being in the state when] the sheaths [of defilements]'”! have not 
been destroyed is called ‘sattva.’ Even though [the essence of this 
sattva, namely the body of a sugata,] dwells [hidden] in ignorance, 
there is no illusory imagining (manyana). [It] dwells in absorption 
(samadhi), is completely calm and there is no motion whatsoever.” 

166 Tib: ’bras bu corresponds in Bth with don. The Skt. was probably karya (cf. MVy 4630, 
6977/78, 7198). The compound yod pa’i ‘bras bu in Tib thus represents Skt. satkarya, a central 
term in the teaching of the Sankhya. 

‘67 The translations show considerable differences in this verse: Ch, follows widely the first 
section of its own prose (see above). Ch2: “Just as in a cane seed the tree—[manifesting itself as] a 
cane sprout [etc.]—in its entirety can be found; [and just as] a rose apple tree [which already 
completely] exists in the root (!) grows again once its seed is planted [in the earth],....” Ch, differs 
from its prose and argues in perfect accordance with the satkarya doctrine. The character #, Skt. 
* mula, does not have any equivalent in the other translations (variant reading of ta/a?). Bth differs 
particularly in the last two padas: ‘dzam bu kun kyi' dbus na yang ni yod : de btab* bas ni don 
kyang ‘byung bar 'gyur : [' for kyis; ? for btang, see prose]: “Also in all rose apples [it] exists. By 
planting it, the result{, which is already completely contained in the seed,] emerges.” 

8 The terms dharmesvara and (vi)ndyaka also designate the Buddha. Ch,, are without these 
designations. Ch, instead: AAR: “I see all [sentient beings] without exception (*asesa).” 
Cf. asesa ~ dharmesvara. 

'© The compound could be *andsravabuddhacaksurvara, which was rendered by Tib without 
altering the position of its members. In Bth the elements -buddha- and -vara are missing. Ch 
corresponds to Tib (Rix iS GEHR EA). For andsrava cf. BHSD (s.v. asrava), where lokottara is 
quoted as an equivalent. 

1 Instead of “body of a sugata” Ch, reads QIK; Ch: 202688. The verse in Ch,: “The 
Tathagata, [with his] uncontaminated vision, perceives the store of a tathagata in the body of all 
living beings—just like the kernel in all fruits.” 

1 With sbubs (Bth: mdzod), koSa, Tib most probably does not mean the unperishable “store” of 
a Tathagata as Takasaki (7.7? 4 RR; 1981: 25) and also Kagawa (7.8872 FRE [BE is to be 
emended to #8]; 1962: 13) understand it. 7ib uses the term sbubs throughout the TGS another 17 
times. In all of those cases it designates the sheaths of defilements covering living beings. Tib 
renders kosa with mdzod when used in a positive meaning as “treasure” in the fifth illustration. 
Ch: “[When] the imperishable is concealed {in the sheaths of defilements, it] is called ‘sattva’” 

172 (1) Pada a of Bth should probably be emended to ... thadad mdzod ma yin (instead of pa 
yin). Tha dad corresponds to Tib: bshig and could render a form of the Skt. root bhid. The gter 
tsheg following sems can takes the position of brjod in Tib and could symbolize an abbreviation 
for Skt. iti, uktva or something similar. 

(2) Instead of amanyand (rlom sems med), Bth speaks in pada b of mkhyen, probably a form of 
the Skt. root jfia. Although the combination rlom sems med — g.yo ba ma yin (aninjand) is 
common (BHSD s.v. ifijana; manyand, °na) and appears also in 5.1, in this case Bth has probably 
preserved a wording closer to the original, for also Ch;, show # (*jfidna) instead of amanyand. 
With the term jridna, this verse reminds of *mahdjrdnasambhara in the prose of 6B. The pada 


PDF Version: BPPB VI (2002) 

[6.4] “Thinking: 

‘How may these sentient beings become awakened—just as a great tree 
has grown from a seed—and [thus become] refuges (Sarana) for the 
world with [its] gods?’ 

I speak the Dharma in order to completely purify [sentient beings].”!” 

[7 The simile of a tathagata image wrapped in rotten rags] 
“Sons of good family, again it is like the example of a poor man’ [who] has 
a tathagata image the size of the palm of a hand [and] made of seven kinds 
of jewels.'” It then so happened that the poor man wished to cross a 
[dangerous] wilderness’”® carrying the tathagata image [with him]. And in 
order that it might not be discovered by anybody else, or stolen by robbers, 
he then wrapped it in some rotten, putrid rags.'”” 
Then the man died'”® owing to some calamity in that same wilderness, 
and his tathagata image, made of jewels [and] wrapped in rotten rags, then 

according to Bth (mi shes pa’i dbusu gnas par mkhyen //): “Knowing that [this body of a sugata] 
dwells in ignorance,....” In Ch the term avidya, testified in all other translations, is not mentioned. 
Instead of it we find 7\34, Skt. *ananya. This variant could result from the graphic similarity of 
ananya and avidya. Chy: FA PAE MAH: “Within [the sheaths?] is the knowledge, not different 
[from myself?].” 

(3) Padas c and d in Ch): “[This knowledge] peacefully dwells [in] absorption (32; samadhi), is 
placed [in] tranquility. Further [it is] without motion (anifijand), there is nothing to be attained.” 
For a similar passage cf. SP 161.10f. = VII.10: cittam pi te santagatam susamsthitam 
anifiyabhutam sada aprakampyam | viksepu naivasti kadacit tava' atyantasantasthitu tvam 
anasravah // [' tava metrically impossible]. 

The whole verse in Ch,: “Ignorance covers the store of a buddha. You should trustfully know, 
that [it] is furnished with absorption and knowledge [and that] there is nothing [by which it] could 
be destroyed!” As usual, Ch, deviates also here considerably from the other translations. Just as in 
the prose section, the term sattva does not appear. 

"3 The pada concerning the refuge runs as follows in Ch): BEStKANZATHK: “... and then 
become supports [for] the world.” The gods are not mentioned; AT+K represents *asraya rather 
than sarana. Ch,: “Therefore I teach the Dharma [and] disclose that store of a tathagata [for living 
beings so that they may] quickly accomplish the unsurpassable path, just as a fruit becomes a king 
of trees.” 

"4 Though the man possesses a precious tathagata image, he is described as poor. This 
characterization seems irrelevant for the illustration and is not found in Ch. 

5 Rin po che sna bdun, Skt. saptaratina, appears often in Buddhist literature (cf. BHSD s.v. 
ratna (2); Suv; s.v. rin po che). It comprises the materials suvarna, rupya, mukta, vaidurya, 
sphatika, musGragalva and lohitika (acc. to BHSD). Ch; reads sarvaratna (—+*)§). Ch, speaks of 
an image of pure gold (8 (8) and gives no measure. 

% For 'brog dgon, MVy 2992 gives the equivalent kantara: “... wilderness, ... a difficult road 
through a forest,....” (MW). The Chinese translation (f@#%) focuses on the second meaning. 

'” Chy: . FTA HE, BEB, HAR, BEI, BRB. : “... [the man], going 
to another country, would pass a risky path. Afraid to encounter robbery [he] would wrap [the 
image] with ragged materials so that nobody [could] recognize [it].” The last section in Ch;: 
“Fearing the robbery of the [image he] would then take rotten, ragged silk and wrap his image with 
[it] so that nobody would suspect [it there].” Ch, then continues with the appellation “Sons of good 

18 (1) Tib translates ‘chi ba’i dus byas par ma gyur (but Bth: shi bar gyur nas). Should Tib be 
interpreted as “not to reach [his predetermined] time of dying”? Cf. Skt. kdlam karoti for “dying.” 
The Chinese versions speak of a sudden death: 2(#Ap#% (Ch); BIRAGKE (CA). 

(2) The meaning of dar nad or dor nod in Bth instead of nyes pa is not clear. However, nad 
implies an illness leading to the man’s death. 


PDF Version: BPPB VI (2002) 

lay around on the footpath.'” [But] travelers,!®° unaware [of the precious 

tathagata image in the rags], repeatedly stepped over [it]'®! and passed by. 
And [they would] even point [at it] as something disgusting [and question]: 
‘Where has the wind brought this wrapped bundle of rotten, putrid rags 
from?’!** And a divinity dwelling in the wilderness, having looked [at the 
situation] with divine vision, would show [it to] some people and direct 
‘O men, [here] inside this bundle of rags is a tathigata image made of 
jewels, worthy to be paid homage by all worlds. So [you] should open 


“Sons of good family, in the same way also the Tathagata perceives that all 
sentient beings are wrapped in the wrappings of defilements and that [they 
are like something] disgusting, wandering around for ages throughout the 
wilderness of sarnsara. '** And, sons of good family, [the Tathagata] 

1 (1) I rendered rdog lam simply as “footpath” without having been able to locate other 

occurrences of this compound. Bth has just lam; see Kagawa 1962: 14 (#§(#) and Takasaki 1981: 
25 GEig): “wayside.” 

(2) “To lay around” for ‘phyan pa, Skt. bhramati; in 7B.3 appears another form of the root bhram 
which refers to samnsara: kun tu ’khyam pa (cf. MVy 5108). 

(3) Ch, (including a part of the following passage): PMESRRAGY. TABI, BB 
AS : “There the golden image, [which] has been abandoned in the wilderness, is considered as 
unpure by all the wanderers trampling upon [it].” Ch): RUREREV RE, BBR, Bie 
Fuh, HiGRWRB. : “The precious image of the Tathagata remains in the rotten, ragged silk 
thrown on the ground [and] wanders around [in] the wilderness.” 

189 "Dron po or ‘gron po (P,S): “traveler.” My translation is in accordance with Jd, who 
conceives of ’gron pa as a synonym for ‘grod pa, having the meaning “to go, to travel.” BCA;s.v. 
‘dron pa gives as Skt. equivalents “1) pathi vartin (N[arthang]: ‘gron pa). 2) sartha [‘caravan’].” 
Only the first meaning is confirmed by the Chinese (7 A; (TESA), by Bth (mi long ba, which 
should be emended to mi song ba, mi ‘dong ba or even mi ‘ong ba), and the parallel verse RGV 
1.118d: adhvaga. 

1811 try to render the repetition ‘goms shing ’goms by adding “repeatedly.” Bth reads ’gong 
mchong. Tshig mdzod denotes for ‘gong ba: (rnying) ... (4) smad pa ... [“to despise”]. ‘Gong 
mchong could accordingly be translated as “to jump over contemptuously.” 

182 Ty Bth (ral ma rul pa dam pos dkris pa rlung gi ded ‘khyam' pa ‘di ga las byung zhes [' for 
‘khyab]), a second verb byung and the position of ‘di imply the following translation: “From where 
appeared (byung) this bundle wrapped ... [which] is driven by the wind?“ I understand ded as the 
perfect form of ‘ded pa. The passage in Ch;: “[But] people [who] are walking along the path come 
and go, pass, trample upon, jump over {the bundle and] do not know that there is a tathagata image 
inside. And because the [image] is wrapped in rotten silk having been thrown on the ground, all 
[consider it something] disgusting. How could [they] have the idea that [it is] a buddha?” 

183 In Ch, the divinity is not mentioned directly (#4 K8R%). This agrees with the parallel RGV 
1.120b: divyanayana. In the prose of Ch,, contrary to its own verse 7.2, the being with divine 
vision is not directing others to open the bundle: SARA ARMPEBSR, Heuwz, —w 
#24. : “Somebody with divine vision sees that in the ragged materials there is an image of pure 
gold and immediately takes {the image] out for [other living beings so that they] all pay homage 
[to it].” Ch, speaks of several divinities $7). The directive runs as follows: “O men, in this 
rotten silk here is a tathagata image. Unwrap [it] quickly! All worlds shall pay homage [to it]!” 

184 (1) The analogies wilderness/sarnsara and wrappings/defilements are missing in Bth and 
Ch,, but appear in RGV 1.120: samsdravartmo-; ... klesavipiitivastra-. Cf. Chz: SIR 52 BR. HL 
ER IATS: “defilements like that rotten, ragged silk.” 

(2) As in Chy, the verb “to perceive” appears in Bth just once in the first two sections of 7B. The 
second mthong in Tib was probably inserted by the translators of Tib in order to make the syntax 
more transparent. 


PDF Version: BPPB VI (2002) 

perceives that also within sentient beings [who] are wrapped in the 
wrappings of various defilements—and even though [they] may have come 
into existence as animals—'*there is the body of a tathagata of the same 
[kind] as my own.'®° 

Sons of good family, 

‘How does the mental vision of a tathagata (tathdgatajndnadarsana) [in all 
sentient beings] become free and completely purified from impurities so 
that [sentient beings] become worthy of the homage of all worlds, as I am 

Thus thinking, the Tathagata teaches in this [connection] the Dharma to all 
bodhisattvas in order to cause [such beings] to become free from the 
wrappings of defilements [in which they] are wrapped.”!°’ 

Then at that time the Exalted One uttered these verses: 

[7.1] “It is as if a sugata image were wrapped in putrid, disgusting 
[materials]—made of jewels [and yet] wrapped in rags—[and] had been 
left!® on a path and lay around [there].”!® 

[7.2] “And the divinity, having perceived it with divine vision, had said to 
some [people]: 

‘Here is a tathagata [image made of] jewels. Open quickly this bundle 
of rags!””!° 

[7.3] “My [vision] is like this divinity’s vision:'?' with that [vision I] see that 
without exception all these sentient beings, wrapped in the wrappings of 

'85 Cf, RGVV 15.11f., which is probably based on this passage of the TGS. The passage here, its 
parallel verses RGV 1.119, 120, and RGVV 15.11f. are, in the early tathagatagarbha teaching, the 
only passages which explicitly state that animals have buddha-nature: ... lokottarayd prajriaya 
sarvasattvesv antaSas tiryagyonigatesv api tathagatagarbhastitvadarsanat.... In Ch, this passage 
does not appear. Chz: ... ShRE+ ZG. : “... [perceives that all sentient beings] attain 
innumerable bodies as animals.” 

'8 Ci in nga ci ‘dra seems redundant but is also found in Bth. It appears again in Bth 7B.7 and 
could well be an archaic formulation. 7B in Ch, from the beginning: “In the same way, sons of 
good [family], I see that living beings [due to their] various defilements circulate for a long time 
(£% &; “dirgharatra), in innumerable [re]births and deaths. [And I see that yet] the wonderful 
store of a tathagata dwells within their bodies, exalted [and] pure, not different from myself.” 

'87 Both Bth and Ch; place the part of the sentence expressing finality before the direct speech. 
Neither the Tibetan nor Ch; shows clearly if the purification of the bodhisattvas or of living beings 
is meant. In RGV I.120d the process of liberation (... vimuktyai) refers to buddhadhatu. In Ch, the 
bodhisattvas do not appear. The last part is different from all other translations and is not found in 
the verse section of Ch,: “Therefore the Buddha teaches the Dharma for living beings, removes 
[their] defilements, purifies [their] tathagata-knowledge [and] increasingly (#848) guides all 

18 For bor te see the parallel verses RGV 1.118, 120: ujjhita. 

18° Ch, follows its own prose: “It is as if somebody carrying a golden image went to another 
country and would [therefore] wrap [the image] with ragged, rotten materials, and [the image] 
would [then] have been abandoned and lay in the wilderness.” In Ch; the image is left at a “risky 
[and] bad place [in] a wilderness” (Fi 56 ft). 

'® (1) Su in 7.2a functions as a metric expletive (see Hahn 1985: 206). 

(2) As in its prose, Ch; does not mention a divinity. The direct speech runs: ith, 
Ee —t0AXRE: “Removing the rotten [materials, you will] reveal a pure [golden] image [so 
that] everybody [will] greatly rejoice!” Ch, mentions several divinities and adds that the image 
should be venerated (444%). Both Tibetan translations have /ha des though the divinity has not 
been mentioned in the verse before. 


PDF Version: BPPB VI (2002) 

defilements, are suffering severely [and] are continuously tormented by 
[this] suffering of satnsira,”!*? 

[7.4] “I perceive that inside the wrappings [consisting of] defilements the body 
of a victorious one is firmly established,” that that [body] is without 
motion and change, and that yet there is nobody setting that [body] 

[7.5] “Having seen [this], I then urged [the bodhisattvas]:!*° 
‘[O you] who have entered excellent awakening, listen! Thus [is] the 

essential law (dharmata) [regarding] sentient beings: here [within 
each sentient being] always dwells a victorious one, wrapped around 
[with defilements].””!*° 

[7.6] “‘When the sugata-knowledge [within] has been set free and all 

defilements are pacified, then this [sentient being] is called 

'9l Tib has two particles: de bzhin and ‘di ‘dra ba. The fact that the three other versions only 
show one such particle, and instead Bth and Ch; mention the act of seeing a second time (mthong 
ba; 54), could be explained by assuming that 7ib has misread a form of the root dys “to see” as 
idysa or (e)tadysa. For an alternative explanation cp. n. 196 (1). 

The versions could be understood in the following way: Bth: “In the same way, with my 
supernatural vision I see living beings. I always see in the most excellent way [that they] are 
concealed by masses! of defilements [and that they] are also troubled by the suffering of sarnsara.” 
[ gong bur to be emended to gong bus; cf. Bth 7B.3; 5.] Ch;: “My divine vision is also like this: 
[with it I] see these living beings wrapped in defilements [and] bad deeds, [and I see how they 
become] born and die accompanied by lots of suffering.” Ch»: “With [my] divine vision I see like 
this: I see [how] all living beings are subject to the wrappings of the silk of defilements [and how 
they] excessively experience sorrow and the suffering of sathsara.” 

'93 Instead of mnyam par bzhag gyur (Skt. *samahita, *samadhigata), Bth simply has gnas. Ch; 
mentions in the same pada the paryanka position of the tathagata-body (402K #8). 

4 (1) The act of perceiving comprises the whole content of the verse. It is not mentioned in 

(2) Bth shows the phrase mi spyod instead of mi ’gyur. This could be a rendering of Skt. acara or 
avicara; for acara the BHSD gives the meaning “unchanging, constant.” 

(3) The verse in Ch,: “Further [I see that with] living beings, in the middle of the dirt of 
ignorance [there is] tathagatahood without motion, [and that yet] there is nobody who could 
remove [the dirt].” 

15 The perfect form of ‘debs is probably chosen because what is described in the following 
verse 7.6 is thought to happen after the urging. For skul ma btab, probably a causative form of the 
root (sam-)cud, Ch, employs ¥¥€1#;, “to alarm, to shake up”; Ch, has simply 3. Different and 
obviously incomplete is Bth with ’grol bar bya instead: “[in order] to cause [the buddha bodies] to 
become free [he says]....” This statement is, however, also part of the corresponding passage in the 
prose of 7B. As is said explicitly in Ch, the urging is directed towards the bodhisattvas. 

1% (1) My translation follows Bth and Ch, (#) in attaching rtag of pada c to pada d, which 
seems to have been its original position. The translation of Takasaki 1981: 27 includes rtag in the 
statement of pada c. This must have been the intention of the Tibetan translator(s) or reviser(s): 
RARE ORMISBIZINEFCHS. : “The true nature of living beings is always thus.” 

(2) For a phraseologically and syntactically similar construction cf. the Samadhirajasutra 
(Régamey 1990: 53, XXIII.17): evam sambhava buddhanam / lokandthana idrSah / na jatu 
kenacic chakyam / pasyitum ... (Tib.: de ltar sangs rgyas 'byung 'gyur te / ‘jig rten mgon po’i de 
’dra bas / sus kyang ... nam yang mthong bar mi nus so //); Régamey’s translation (p. 90): “Such is 
the nature of the Buddhas, Rulers of the World. Nobody can perceive them....” 

(3) Instead of dharmata Ch; has just %#. Ch,’s statement that the buddha is found within fears 
(i for bhaya, trasa or so on) must be based on an erroneous variant reading of the Skt., on 
which (vongs su) dkris pa and ‘di na (Bth: de na) were based. 

(4) Ch,;: “As the Buddha has seen thus, [he] teaches for the bodhisattvas: ‘Masses of 
defilements—the bad deeds—cover the body of a victorious one [within living beings].”” 


PDF Version: BPPB VI (2002) 

‘awakened’ (buddha) and the hearts of gods and humans are full of 
. 999197 

[8 The simile of the future universal emperor in the womb of a poor, 
depressed woman] 

“Sons of good family, again it is like the example of a woman without a 
protector (anathabhuta), of unsightly complexion, having a bad smell, 
disgusting, frightening, ugly and like a demoness (pisaci), [and this woman] 
had taken up residence in a poorhouse.'* While staying there she had 
become pregnant. And though the life that had entered into her womb’” was 
such as to be destined to reign as a world emperor,’” the woman would 
neither question herself with reference to the sentient being existing in her 
womb ‘Of what kind is this life [that] has entered my womb?’, nor would 
she [even] question herself in that [situation]: ‘Has [some life] entered my 
womb or not?’””! Rather, thinking herself poor, [she would be] depressed, 

'7 (1) The syntactical construction could derive from Skt. yadd ... tadd...; see also Ch (GBF ... 

(2) Pada a of Ch): BURR (RE ER: “When this [buddha within] has been set free and [one has] 
manifested the buddha knowledge,....” In the last pada Ch; mentions aprt from the joy also the 
venerating activity (#£%) of gods and humans. Ch,: “Purify diligently [the buddha within and] 
remove [the defilements]; bring forth the tathagata-knowledge! [Then you will] be venerated by all 
gods, humans, dragons (naga), hungry ghosts (preta) and snake-gods (yaksa)!” 

198 Bud med mgon med par gyur pa, Skt. *andthabhita stri, indicates a female without any 
male family members (cf. Ch,: “orphan woman”; #\3§2c,\.) who could be in charge of her 
“protection.” Characteristic for the Indian concept of women in need of protection is Manu IX.3: 

pita raksati kaumare bharta raksati yauvane / 
raksanti sthavire putrd na stri svatantryam arhati // (cf. also V.147ff.) 

What exactly mgon med pa’i khang pa, Skt. *anathasala/-vasatha, stands for, is not very clear. In 
Patis | 266.30ff. compassionate people are said to bring bandages to an andthasald so that the 
inhabitants could bandage their open wounds (probably inflicted through punishments). An 
anathasala could be an assembly point for people at the edge of the society without any family 
ties. Ch, renders it as a (socially) low, run-down house (F353, $£587.3<). The meaning “run- 
down” of andatha is also prevalent in the description of a vihdra (see Couvreur 1957: 317). My 
translation “poorhouse” might therefore not cover the complete range of functions and notions of 
such a place in India. 

'° The term sattva seems common in ancient Indian literature to designate the entity entering 
the womb and becoming the embryo (cf. e.g. the quotations in Hillebrandt 1906: 183; also Divy 
98.20ff.: ... sattvo ... kuksim avakrantah /). Instead of “to enter” (zhugs for *avakramati?) Bth has 
here and in 8A.6 skyes from skye: “to arise.” 

200 rGyal srid byed pa is probably based on Skt. rdjyam karoti or karayati (see Udr; s.v. rgyal 
srid), Bth has simply rgyal po. For the characteristics of a “world emperor” (cakrevartin) see 
EncBuddh s.v. cakravartin. 

201 (1) In Bth the question which the woman does not ask runs as follows: “What kind of life 
has arisen in my womb, has entered my womb?” Instead of the phrase yid la mi byed which 
appears after each of the two questions in Tib, Bth states: “‘... [this she] does not think, [this she] 
does not consider, [this] does not enter [into her] consciousness.” (... zhes mi sems : yid la mi byed 
: der ’du shes mi ‘jug go //). Whereas Bth obviously combines the second jug pa with ‘du shes 
(samjfia), which is not attested in 7ib, Tib includes it in the second question (sam / ma zhugs), 
referring to the womb. Mi sems in Bth should have its equivalent in one of the two yid la mi byed 
of Tib. The second question, pertaining to whether life has entered the woman’s womb or not is 
only attested for Tib. However, in verse 8.3 it comes up in both Jib and Bth. 

(2) The passage in Ch;: “By chance [she then] had intercourse [and] would become pregnant in 
[her] belly; [and] this [would be] an embryo [destined] to necessarily function as a world emperor- 


PDF Version: BPPB VI (2002) 

[and] would think thoughts [like] ‘[I am] inferior and weak,’ and would pass 
the time staying in the poorhouse as somebody of unsightly complexion and 
bad smell.” 


“Sons of good family, in the same way also all sentient beings [think of 
themselves as] unprotected and are tormented by the suffering of sathsara. 
[They, too,] stay in a poorhouse: the places of [re]birth in the states of 
being.” Then, though the element of a tathagata has entered into sentient 
beings and is present within, those sentient beings do not realize [it].2™ 

king. And yet, even though this woman were pregnant, [she] would not at all have thoughts such 

as [appropriate for the mother of an emperor].....” or “...were pregnant, [she] would not at all have 

itself, which in the Tibetan follows directly after, does only appear at the end of 8A: “And [she] 
would not know for certain: ‘What kind of human is it that has arisen in my belly?’” (RAE 
ACG: Caece ee 

202 (1) The Skt. on which Tib is based could be ... (na) anyatra sa daridracintam linahina-...- 
cintam ca anu(vi)cintayet durvarnadurgandhatayd ca... (cf. n. 136). Instead of the triplet zhum pa 
dang / dman pa dang | kho ru chung ba the text in Bth shows ngan pa dang : nyon mongs pa dang 
. mi dge ba; Chz: HE. FAZ)...R4, ....dMan pa should correspond with ngan pa (hina; cf. 
8.5d; Chz: 4). Zhum pa (Skt. dina, nica , lina, viséda, samkoca or something similar) does not 
correspond with Bth: nyon mongs pa (probably for vyasana;, Chz: [&) nor with Bth: mi dge ba. 
Difficult is also kho ru chung ba, for which the dictionaries do not have entries. It could 
correspond with Chz: 33 (*durbala = stobs chung ba or nyam(s) chung ba). However, also in 
this case Bth (mi dge ba) is different. Two compounds in a row with the same final member to 
which the prior members of the compounds are referring, in the form ax bcdx, are not uncommon 
in Buddhist Hybrid Sanskrit. The first compound “daridracinta I have taken to be an iti- 
compound, similar to the one in n. 136. The second compound */ina-hina-durbala-cintad cannot be 
analyzed in this sense, as Jina usually describes the temper rather than the content of a thought (cf. 
RGV1.157a: linam cittam; BHSD s.v. lina); for hina cf. 8.5d. 

(2) The Skt. for dri mi zhim pa nyid kyis could have been something like durgandhatvena. 
Therefore my translation: “as somebody of ... bad smell.” Bth simply has dri mi zhim pa. 

(3) The passage in Ch: “[She] would only think [herself] poor, deficient [and] inferior. 
Following the weakness of her mind [she] would continuously have this thought: ‘I am ill- 
looking.’ [She] would pass [her] time dwelling in a low, run-down house. And [she] would not 
know for certain... (see above).” All of 8A in Ch;: “And again, sons of good [family], it is like a 
poor, low [and] ill-looking woman, hated by everyone, and yet [she] would carry a precious child 
[in her belly] who later would be a noble king, ruling the four worlds. [Yet] this person would not 
know [about it and] pass the time thinking continuously that [also] the child [will be] inferior [and] 
of low birth.” 

203 (1) “Unprotected,” andtha, is rendered by Ch as “without master, without support” (4% 
=, #0. 

(2) Srid par skye ba’i gnas possibly for Skt. bhavopapatty-ayatana; bhava is synonymous with 
samsara. Ch): 4:=74rp: “[re]borm in the three states of being”; cf. the corresponding passage 
RGV1.122a: anathasaleva bhavopapattih. 

(3) The passage in Ch,: “In the same way, sons of good [family], the Tathagata observes [how] 
all living beings wander around in samsara [and] experience the poison of all [kinds] of suffering.” 

24 (1) Instead of gotra or hula (Tib: rigs) Bth and Ch, have dhatu (khams; 5#). The passage is 
cited in RGVV 72.11-12 and reads dhdtu: tatra ca sattve sattve tathdgatadhatur utpanno 
garbhagatah samvidyate na ca te sattva budhyanta iti /. For a discussion of the meaning of dhatu 
here see section A 3.1. Dhatu is also attested for the parallel pada RGV I.121c. I have no definite 
idea if and why the translators or revisers of Tib preferred rigs to khams, or if and why already one 
strand of the Indian tradition replaced originally dhdtu with gotra. One would also expect the 
character PE rather than 5} (Chz) in the common translation 412 ( for tathagatadhdtu, to express 
the buddha essence in living beings. A possible reason for a deliberate alteration from dhatu to 
gotra (or khams to rigs in the Tibetan tradition), however, could be the classification system 


PDF Version: BPPB VI (2002) 

Sons of good family, in order that sentient beings do not despise 
themselves, the Tathagata in this [connection] teaches the Dharma with the 
[following] words: 

‘Sons of good family, apply energy without giving in to despondency!”™ It 
will happen that one day the tathagata [who has] entered [and] is present 
within you will become manifest.”°° Then you will be designated 
‘bodhisattva,’ rather than ‘[ordinary] sentient being (sattva).’ [And] again 

established later in the RGYV (cf. section A 4.5) where this eigth simile is subsumed under the 
gotra aspect. In the first part of the translation of Bth (de nas sems can dag de bzhin gshegs pa’i 
khamsu skyes pa yang yod : snying po la gnas pa yang na : sems can de dag gi khong du mi chud 
te ;) the terminative particle su after khams is to be deleted, if we want Bth to agree with the 
Sanskrit quotation. Otherwise it could be understood as “Then living beings come into existence 
and stay in the dhatu of the tathdgatas.” The two Tibetan translations of this passage are 
characteristic as Bth shows the more literal but grammatically also more problematic version, 
whereas Tib has introduced some changes (that are not based on the Sanskrit) and thus becomes a 
more smoothly readable text. 

(2) The expression that the dhatu of a tathagata has “arisen” (utpanna) in living beings is 
probably influenced by the upamana, where it is said that life has entered (zhugs) the womb of the 
woman. The translators of Jib probably decided to render utpanna with zhugs, a verb not common 
in translating forms of ut-pad (cf. Bth: skyes pa), since the same verb is used in the the upamana. 

(3) Khong (na) in 8B.4 renders garbha (cf. Bth: snying po; parallel verse RGV 1.123b: 
garbhantarastha). Ch; obviously interprets this garbha not as the inside or the womb of living 
beings, but as the “store[, i.e., element] of a tathagata” (402%#) parallel to dhadtu. The passage in 
Chy K-—FAls ARO, SUR. BRABAR, AH. : “And yet each living being 
has the element of a tathagata (tathdgatadhdtu), has the store of a tathagata. [But] this [is 
something what] living beings are not aware [of, what they] do not know.” Ch, constitutes the 
version with the widest gap to the transmitted Sanskrit: HS 4Au0RwRM, LOK 
ZA, Aw. : “In all their bodies there is the precious store of a tathdgata—just as in case of 
the woman—yet [they] do not know [Iit].” 

(4) A construction expressive of concession (... samvidyate na ca te...) can also be found in the 
correspondent verse I.123 of the RGV: sannathesu ... svatmantarasthesv api //. 

(5) The first part of the passage cited in the RGVV could also have been the basis for a quotation 
of the TGS found in the MPNS, Of course, it should be titled a free rendering rather than a word- 
for-word quotation. The first part of the quotation in the MPNS runs as follows (including the 
introduction): gzhan yang ‘di na dge slong la la de bzhin gshegs pa’i snying po’i mdo sde chen po 
ston par byed do! // sems can thams cad la ni sangs rgyas kyi khams yod la” khams de rang rang gi 
lus la ‘chang’ ste /.... [! S: de; * S: la |; Q: tshang.] (MPNS Q 99a6; S 175a4f.; quotations of the 
Chinese in Takasaki 1974: 137). 

25 (1) brTson ‘grus brtan par gyis shig could be the equivalent to *viryam dydham 
arabhadhvam, .,. utthapayata (with drdham understood adverbially) or *drdhaviryam arabhadvam 
(with drdha as an attribute of virya; in this sense Ch; BABHREZL; cf. BHSD sv. 
drdhaviryata; KP §153: -drdhavirya-). The character # probably renders a form of the root a- 
rabh or ut-tha (see BGDJ 1255d). 

(2) A detailed analyses of the meaning of virya is found in Pagel 1995: 201-216. Concerning the 
passage here, some verses in the BCA (VII.2, 16, 17) are of particular interest because they 
mention dlasya (“indolence”) and visaddtmavamanyana (“dispondency and self-contempt’) as the 
antipodes of virya. There it is also said that one should not deny the possibility of awakening for 
oneself (VII.17: naivavasaddah kartavyah...). Also in the BCA, as the commentator Prajiakaramati 
makes clear (cty. to II.17), virva is taken as the main force to overcome a negative self-assessment. 

(3) For sro shi bar ma byed pa the Tshig mdzod (s.v. sro shi ba) has the following entry: (rnying) 
dpa’ zhum pa |. In Tib the auxiliary verb byed pa, rendering the expression transitive, is added so 
that bdag nyid can be taken as the object. Bth has here and also in 8B.5f. (Jib: khyad du mi gsad 
par bya) (bdag cag la) gtses par ma byed par. Ch2, too, uses the same terminology 4S... #kZE in 
both passages, so that we can assume that also in the Skt. the same formulation appeared twice. 

206 (1) Chy 7S Shea Hees. PEAIESERIE®, : “Within all of your bodies there is the 
essence of a tathagata! Later [you] will all (or: finally; 3¢) realize perfect awakening!” For Ch, see 


PDF Version: BPPB VI (2002) 

in the [next stage you] will be designated ‘buddha,’ rather than 

Then at that time the Exalted One uttered these verses: 

[8.1] “It is as if a simple (bala) woman without a protector, of bad complexion 
and ugly disposition would [go to] stay in a poorhouse, [and] after a 
time would there have become pregnant,” 

[8.2] “[and yet] the [life] that had entered her womb would be such that the 
[embryo] was destined [to become] a world emperor king, elevated by 
[his] magnanimity (mahdtmata) [and his seven] jewels, and ruling [over 
all] four continents.”?”° 

[8.3] “[But] that simple woman would [behave] like this: [she] would not know 
if [some life] had entered [her] womb, rather [she] would [continue to] 
stay”'’ in the poorhouse and pass the time in the belief that [she] was 

207 (1) Probably, Ch) uses prthagjana (J) instead of sattva to separate the three categories 
sharper. The same motive might explain the statement that entering into the third category (that of 
being a tathagata) will only happen after a Jong time (A {&). This contradicts the parallel in 8.6b: 

(2) The latter half of 8B in Ch;: “Therefore the Tathagata teaches widely the Dharma for [living 
beings] saying: ‘Sons of good [family], do not look down on yourself (for &@f see BGDJ 238a: 
garhita)! [In] your own body you all have the buddha-nature ((#{)! If [you] practice diligently 
[and] diminish ali evil, then [you] will attain the designations ‘bodhisattva’ and ‘exalted one.’ 
[You will] guide [and] save innumerable living beings!’” The last part is not found in TGS), but it 
appears in the last verse of 8C in all four translations. 

208 In Ch , only five verses are found. 

209 There is no equivalent for bala in Bth and Chj. Ch2 mentions the pregnancy with the 
imperial embryo already in pada d: @=Ef4Z. Also Ch, integrates parts of 8.2 in the first verse: “It 
is as if a poor woman of extremely ill-looking appearance would yet carry a child of precious 
characteristics [in her womb], who later would be the cakravartin king.,....” 

210 (1)The use of ‘byung in Bth instead of ‘phags pa indicates *udgata, *(sam)udita or 
something similar in the Skt. original; Ch,: [E]. For rin chen rnams, Skt. (sapta)ratna, see BHSD 
s.V. ratna. 

(2) The understanding of Bth is different: “... [and yet] she would thus have attained such [a life] 
in her womb [that] would be destined to become the world emperor king, through [whose] 
magnanimity all [seven] jewels would arise [to him, who] would become the ruler of all four 
directions.” Ch, corresponds with Bth in the first half of the verse; the second half could be 
rendered: “This king would [have] magnanimity (mahatmata) [and] would be surrounded by the 
seven jewels. [He] would rule over the four continents and be the sovereign.” In Ch, the first half 
of the verse corresponds to padas c and d of Ch; and Bth; the second half already covers parts of 
verse 8.3 in TGS): “Seven jewels would complete {his] merit, [and as a] king [he] would have the 
four continents [under his control]. But this [woman] could not know [it and would rather] 
continuously think [that she and her child were] inferior.” 

211 gNas byed could be a translation of Skt. vasam karoti. The auxiliary verb byed (in contrast 
to ’gyur) further seems to stress that staying in the poorhouse is an autonomous act based alone on 
the decision of the woman herself without being inevitably forced to do so by other circumstances. 

212 There is no particle of comparison in Bth (contrary to Tib: ‘di Ita bur). Ch; uses 4ifz as the 
object of 4: “This simple [and] ugly woman would not yet know that in her belly exists such [a 
king].” Instead of the passing of time in pada d Ch; mentions the negative thoughts of the woman: 
“... [and she] would cherish poverty [and] pain, [with her] heart [full of] distress.” Ch, corresponds 
roughly with 8.4 in TGS;: “I see that all living beings suffer like infants. [Yet their] body embraces 
the store of a tathagata, but [they] do not realize [it].” 


PDF Version: BPPB VI (2002) 

[8.4] “In the same way, I see that all sentient beings also [think of themselves 
as] unprotected and are distressed by dharmas, [which lead to] 
suffering, remaining [caught up] in the lesser pleasures of the three 
spheres, [even though] inside [them] there is the true nature 
(dharmata)—like [the world emperor in] the womb [of the woman].””” 

[8.5] “Having seen thus [I] taught the bodhisattvas:”"4 
‘As all sentient beings do not know about the true nature (dharmata) 

within [their own] wombs [which] grants benefit [to] the world, [take 
care of them and let them] not consider themselves inferior!’””!° 

[8.6] “‘Apply energy (virya) firmly! Soon you yourselves 7! will become 

victorious ones. At some point [you will] attain the essence of 

213 (1) BDJNQ in pada d have mngal (la) gnas (“embryo”). The analogy is thus between the 
dharmata and the embryo: “... there is the dharmata—like the embryo [in the womb of the 
woman].” Bth reads stong pa instead of mngal and could by that mean the hollow of the womb. 
The same pada in Ch): “The store of [buddha] qualities in the body [of sentient beings] is like the 
store of [the woman’s] womb.” Bth and Ch, both lack a verb of existence. 

(2) Chz does not speak of distressing dharmas: “... [they are] experiencing poverty [which] drives 
[them] to suffering.” (42335074); instead of “remaining in the lesser pleasures of the three 
spheres”: “being in the three spheres, indulging [in] pleasures.” 

(3) The fourth verse in Ch, corresponds rather with verse 8.5 of TGS and will be translated 

214 (1) In the prose, the speech is directed to living beings. 

(2) The perfect form bstan (Bth: smras; see also RGV 1.110d: ... vyasyjat) is probably based on 
the Skt., and is intended to allude to the teaching activity of the historical Buddha. 

215 (1) Though the meaning of mngal gnas (Bth: snying po gnas) is usually “embryo,” I take it 
as a literal rendering of Skt. garbhasthadna or garbhavasati: “womb.” Such an analysis is 
confirmed by Ch;: “... [the dharmata] in the womb, [being] the benefit of the world, full of 
light” (B§4> tt Fil). The Sanskrit may also have been a predicative construction joining that 
in 8B.4, where the quotation in the RGVV read garbhagata and the parallel verse in the RGV 
1.123b had garbhantarastha. The variants in fn. 18 and 19 of P;2 in the Tibetan edition document 
the modification from mngal gnas na to mngal na gnas. This can be considered a redactional 
attempt to give more sense to the Tibetan wording, which (with mngal gnas na: “in the embryo”) 
is, from the standpoint of the Tibetan reader, not understandable. 

(2) Tib does not associate the dharmata with light. However, this qualification is found in all 
other translations: ‘od byed (Bth); AIHA (Ch2); AWHEAY (Ch, in 8.4d); see also the parallel 
verse RGVI.121c: ... garbhena rajasriyam udyahanii...: “with [her] womb carrying the /uster of a 

(3) Whereas Tib only gives sense when the last pada is understood in a causative sense “[Take 
care that living beings]...,” the basic understanding of Bth and Ch; is probably that the 
bodhisattvas should not look down on living beings. Both Bth and Ch, do not have a personal 
pronoun of the first person (bdag of bdag dman ‘du shes), but use a pronoun of the second person 
(Bth: khyod rnams ngan pa’i ‘du shes), or do not mention a pronoun at all (Ch2). The causative 
sense is only found in 7ib, and thus can hardly be an original element of the Skt. text. The last 
three padas in Brth: “‘Knowing about the dharmata of all sentient beings [which], found [in their 
own] wombs, benefits the world and emits light, you should not have a bad idea [about them]!”” 
Contrary to Bth which has no negation of “to know,” Ch has no equivalent at all: “‘All sentient 
beings have the dharmatd [which], in [their own] wombs, [grants] benefit [to] the world [and] 
emits light. [Therefore] respect [all sentient beings and] do not deceive [them]!’”” 

(4) The parallel verse 8.4 in Ch,: “[All living beings do not realize their buddha-nature], therefore 
I disclose to the bodhisattvas: ‘Take care not to look down on yourself! Your body [has/is] the 
store of a tathagata! [In your body] there is always the light to liberate the world!’” 

2167 understand rang lus (Chz: § §) for Skt. *svakdya, *svadeha or *svanga here as simply 
reflexive, similar to the reflexive use of Gtman. It serves to stress the identification of the 
individual practitioner with the prophecy made in the verse. 


PDF Version: BPPB VI (2002) 

awakening.”!” Then [you] will proceed to liberate myriads of living 

[9 The simile of golden figures within burned clay molds]””” 


“Sons of good family, again it is like the example of figures of horses, 
elephants, women or men being fashioned out of wax, then encased in 
clay” [so that they are completely] covered [with it and finally, after the 
clay has dried,] melted [in fire]; and after [the wax] has been made to drip 
out, gold is melted. And when [the cavity inside the mold] is filled with the 
melted [gold],””’ even though all the figures, having cooled down step by 
step (kramena) [and] arrived at a uniform state,””” are [covered with] black 

217 For byang chub snying po, Skt. bodhimanda, see BHSD and Lamotte 1962: 198, n. 105. Bth 
reads erroneously bodhimandala (byangchub [sems] kyi dkyil ’khor) instead. 

218 Dada a in Cho: Se ERS LLSEF: “[If you] bring forth firm energy [and] keep [it] by [your] 
practice,...”; for dus zhig in pada c: A\A”’H: “then, not after a long time.” Ch, ends with verse 8.5, 
which corresponds with verse 8.6 in TGS): “If [you] apply diligently energy, [you] will, after not a 
long time, attain the essence of awakening (bodhimanda), realize the path of perfect awakening 
[and] liberate innumerable [living beings].” 

219 (1) The here described technique for metal casting is the so-called lost-wax process or cire- 
perdue method. Common on all continents except Australia, it dates from the 3™ millennium BC 
(see EncBrit s.v. lost-wax process; Bol 1985: 19f., 119ff., 125ff.). In India this method is known 
by the name madhucchistavidhdna (see Banerjea 1956: 214f.). The descriptions of the process in 
the 21 verses of the Devatabhakti chapter in the Abhilasitarthacintamani of thel 1/12" century CE 
(see Saraswati 1936), along with the monograph of Reeves (1962), who describes the practice of 
this technique in India and Nepal till the modern times, appear to be practically identical with the 
process presented in the 7GS. For a comprehensive description of the lost-wax process as found in 
Nepal see also Michaels 1988. As is documented by the Pratyutpanna-buddha-sammukhavasthita- 
samadhi-sutra, figures of buddhas were already known at the beginning of the second century CE 
(see Harrison 1978a: 38; Hob s.v. butsuzo). A chart in Reeves 1962: 84 documents that even in the 
second half of our century in certain parts of Madhya Pradesh, figures of elephants and horses are 
the ones most frequently cast. The prominent position of these animals is confirmed by the simile 
of the TGS. 

(2) As it is the case in the parallel verses of the RGV, in the TGS the expression for the mold 
containing the figure on the one hand, and the figure itself on the other hand is the same: gzugs 
(RGV 1.124; 126: bimba). In order to render the process more distinctly I had to supply a number 
of additional annotations in brackets. 

2207 it.: “... [the figures are} then put into clay,....” Bth instead mentions that the figures are 
besmeared (bskus) (before: ... lus kyi lugs sgoms sam :?). Chz: .. JEEH EL, ....: “.. [with] clay 

[one] covers them,....” Instead of the figures of women Bth has lion figures. 

221 I Ch, the melting and dripping out of the wax is not mentioned. Instead it is said that 
heating with fire is employed: FAA. The passage in Ch,: “And further, o sons of good [family], 
just as a smith casts figures of pure gold and, as the casting is completed, puts [them] on the 

222 (1) Instead of the gradual cooling down of the figures Ch, says that (the artisan) has waited 
for the (figures’) cooling down (fRHUGB). 

(2) The Chinese versions are without an equivalent to mnyam par gnas par gyur pa 
(*samavastham prapya; Bth: *samatdvasthdm prdpya). The text of Ch, (LICHGRAGE. : 
“fhaving waited for the figures’ cooling down,] their (= the figure’s) artisan brings [them] back 
home”) could easily result from a minor variant in the Skt. text: whilst Jib is based on something 
like samdvastham prapya, Ch, may have read svam avasatham prapya. I am not sure whether the 
statement that the figures arrive at a uniform level is intended to point out the importance of an 
equal heating and cooling down of the figures. In the process of heating this is definitely a decicive 
factor, as is shown by a remark regarding antique bronze casting: “Unless it is baked slowly and 
uniformly, it is quite likely to crack.” (Lechtman and Steinberg 1970: 5; emphasis by me). 


PDF Version: BPPB VI (2002) 

clay and unsightly outside, [their] insides are made of gold.” 

Then, when a smith or a smith’s apprentice [uses] a hammer [to] 
remove” from the [figures] the outer [layer of] clay [around] those figures 
which he sees have cooled down, then in that moment the golden figures 
lying inside become completely clean.”**° 


“Sons of good family, likewise also the Tathagata perceives with the vision of 
a tathagata that all sentient beings are like figures [in] clay: the cavity inside 
the sheaths of outer defilements and impurities is filled with the qualities of 
a buddha /and with] the precious uncontaminated” knowledge (andsrava- 
jiidna); inside, a tathagata exists in [all] magnificence.” 

Sons of good family, having then perceived that all sentient beings are 
like this, the Tathagata goes among the bodhisattvas and perfectly teaches 
[them] these [nine] Dharma discourses of that kind[, i.e., on the tathagata- 
knowledge within all sentient beings].””° [Using] the vajra[-like] hammer of 
the Dharma,” the Tathagata then hews away all outer defilements in order 

223 Ch: SMAEHESE, PARANA, : “Though [the mold] outside was burned black, the figure 
inside is unchanged.” 

224 @ technical term for the removal could be a causative form of the root sphut: “... to burst or 
rend suddenly, break, split, divide...” (MW s.v.). See, for example, the above mentioned text of the 
11/12" century, where the removal process is described: sphotayen myttikam dagdham ... 
(Saraswati 1936: 142, line 14). However, Tib employs continuously ’gogs (or bkogs), which does 
not seem to be a very appropriate term to describe the treatment. A Tibetan equivalent for 
sphotayati could be ‘ges pa (Ja: “to split, cleave, divide”). Its intransitive form ’gas pa is given in 
MV y 6492 for sphutam. Bth, on the other hand, has bong bas bsnun te : bcom nas : phyirol kyi sa 
med par byed do :, which should be translated as: “... [he] pricks into with a small stone (?), 
destroys [the mold] and removes the outer [layer of] clay.” Cf. the parallel verses RGV I.124d, 
126b and d, where samchedayet and samchedayati are used. For the emendation from samcodayet 
and samcodayati cf. Schmithausen 1971: 155. Ch2 does not mention any tool. For Ch, see below. 

225 Before stating that the figures become clean, Ch, adds after the mention that the mold has 
been destroyed: “As [he] has polished [the figures],...” (EZ}##¢C; the compound verb ##} is 
uncommon). Ch,: BA, With, @f5HZ. : “{The caster] opens the mold [and] takes out the 
figure. The appearance of the gold[en figure] is dazzling.” 

226 For anasrava cf. n. 169 above. 

227 Tih does not make sense in this last passage, and I have therefore in syntactical matters 
relied on the other two translations of TGS). In Bth the verb rab tu gang governs the 
buddhadharmas as well as *andsravajfanaratna. Nang na in 9B.4, which in Tib is related to 
*anasravajnanaratna with a genitive particle, appears in Bth (dbus na) as part of the following 
unit as becomes clear by my translation: “...— inside, a tathagata...”; Ch2: PARE YP Ra PRA. 
SIS. The phrase mdzes par ‘dug pa I understand in the sense of Pa. subhatthdyin (see 
PTSD s.v.). Bth (legs par gnas pa) instead suggests a form of the root su(sam-)stha or something 
similar. There is no equivalent for this in Ch2. The whole section in Ch,: “Likewise, o sons of good 
[family], the Tathagata observes that [in] all living beings there exists the store of a tathagata {in 
their] body [and that they] are endowed with the characteristics [of a buddha].” 

228 For chos kyi rnam grangs (Bth: chos kyi gzhung), Skt. dharmaparyaya, see BHSD s.v. 
paryaya (2); RGV has, in the parallel verse I.126d, dharmakhydnanaya- instead, Bth remains 
without the collective particle dag in connection with the Dharma; Ch): 2%. Ch; is much shorter: 
“Having observed thus, [the Tathagata] widely discloses [the Dharma] for [all].” Go2#C, B 
SRAZR. ). Except Tib, all versions seem to have the general Dharma in mind, as in the other 
illustrations before. 

229 A vajra is a very hard stone or diamond. The compound *dharma-vajra-mudgara is 
probably a rijpaka-karmadharaya: “a vajra[-like] hammer which actually is the Dharma” (dharma 
eva vajramudgarah). Ch just mentions a vajra-pounder (4H #8 4x). 


PDF Version: BPPB VI (2002) 

to entirely purify the precious tathagata-knowledge of those bodhisattva- 
mahdsattvas who have become calm and cool.”° 

Sons of good family, what is called ‘smith’ is a designation for the 
Tathagata.”*' Sons of good family, after the Tathagata, the Honorable One 
and Perfectly Awakened One, has perceived with [his] buddha-vision that 
all sentient beings are like this, [he] teaches the Dharma in order to establish 
[them] in buddha-knowledge, having let [them] become free from the 

30 Tn the Tibetan, the cooling down is limited to the bodhisattvas (but in verse 9.4 Bth mentions 
also the sattvas). Ch, seems to comprise both groups (see below). Ch, does not mention the 
bodhisattvas in the prose. In the parallel verse RGV 1.126 the bodhisattvas are not explicitly 
mentioned. The verse before, with the formulation jagad ... visodhayati, rather implies that santam 
avetya ... manah in I.126d should also be the manas of all living beings. That the Tibetan here only 
mentions the bodhisattvas is also surprising, because verse 9.6 later states explicitly that the 
Tathagata perceives both groups. 

731 The last two passages have been understood differently in each version. It seems that Bth 
and Ch; give a more detailed explanation why ‘smith’ functions as a designation for the Tathagata. 
Ch: “(1a) Whether bodhisattva-mahasattvas or [other sentient beings who] have become calm 
[and] cool, the Tathagata for those sentient beings, with a vajra[-like] pounder purifies their 
[supernatural] vision of the Dharma [and] removes their defilements and impurities (1b) in order 
to purify the store of the precious tathagata-knowledge [within them]. (2) Sons of good [family], 
the Tathagata is like somebody carrying a precious figure (dhdraka for kdraka?). (3) [And], sons 
of good [family], in light of the similarity between [the artisan who] destroys the [figures’ outer] 
forms (?) and [the Tathagata who destroys living beings’] impurities and lets [them] attain 
liberation, [‘artisan’] is a designation for the Tathagata.” Without my additions the text of Ch; in 
(2) and (3) would hardly be understandable. My interpretation is based on Bth which, however, has 
to be emended. 

The first passage of Bth in brackets [ ... ] appears twice and I take its repetition to be caused by an 
inattentive scribe who, as a classical case of aberratio oculi, jumped into the following passage 
also ending with ... bar byed pa (9B.13). dKon mchog gi ye shes in 9B.11—12 makes no sense too 
and I have to ignore it. The text of Bth: (1a) de nas byangchub sems dpa’ sems dpa’ chen po zhi 
bar ’gyur grangs bar gyur te : de bzhin gshegs pas chos kyi rdorje'i mtshon kyis de dag gi phyirol 
kyi nyon mongs pa mang po thams cad ma rung bar byed pa [de ni de bzhin gshegs pa’i tshigs bla dag so : 
rigs kyi bu ‘di bzhin du de bzhin gshegs pas : sangs rgyas mig gis] (1b) ye shes dkon mchog rab tu sbyangs pa’i 
Phyir ro : [dkon mchog gi ye shes] (2) las byed pas : (3) gzugs de dag bcom ste : (2) dkon mchog gi 
gzugs (3) nyon mongs pa la thar bar byed pa ‘di ni de bzhin gshegs pa’i tshigs bla dag so : 

It is likely that Bth has kept the assumed original text in (2) and (3): “That the smith destroys the 
[outer] forms [means] that [the Tathagata] frees the figure[-like] jewel[, the tathagata-knowledge 
within,] from defilements. In light of this [fact, ‘smith’] is a designation for the Tathagata.” 

The last two passages in Ch, until the end of 9B: “[When] those living beings become calm and 
cool, with [his] vajra[-like] insight, [the Tathagata] smashes the defilements [and thus] discloses 
and purifies the buddha-body [within], just as [an artisan] takes a golden figure out [of its burned 
mold of clay].” 

782 (1) The Tibetan leaves open if the liberation from the defilements is part of the construction 
of finality ending with ... phyir (9B.15), or if the teaching takes place after the liberation. The 
before mentioned alternative considers the teaching as the basis for any emancipative process, 
whereas the latter alternative reminds us of the statement in 2B (cf. n. 93), which suggests that the 
teaching follows the liberation from the defilements. This is less probable. 

(2) With rab tu dgod pa Tib could translate a causative form of the root *prati-stha; cf. e.g. ... 
hinayanasmi pratisthapeyam ekam pi sattvam na mametu sa@dhu // (SP 47.4 = 11.57, Tib.: ... theg 
pa dman la ... bkod na // (Q 23a6). The idea that one can become established or fixed in 
knowledge is further attested in e.g. SP 64.4f. = II.21 where Sariputra says about himself: tada mi 
vidhvamsita sarvasamSaya vicikitsa nasta ca sthito ’smi jriane //. 

(3) Ch) is quite different in this passage: “Sons of good [family], the Tathagata, Honorable One 
and Perfectly Awakened One, perceives that the store of a tathagata of all sentient beings has sunk 
into the sheaths of an unlimited [number] of myriads of defilements. For those sentient beings, [he] 


PDF Version: BPPB VI (2002) 

Then at that time the Exalted One uttered these verses: 

[9.1] “It is like [the example of casting golden images: first, wax] figures are 
covered outside with clay; [then, after the wax has been melted and so 
drips out,] the inside [of the clay] has a cavity and is empty; [finally,] 
when [these cavities] are filled with precious melted materials, [they] 
turn into many hundreds of thousands [of golden figures].” 7° 

[9.2] “[Then] a smith, realizing that [the figures in the clay] have thoroughly 
cooled down, hews away the coverings of clay around the figures, 

“What can [I] do so that (vatha) these [black molds with their insides] 
made of precious materials may turn into clean figures?”””** 

[9.3] “In the same way I see that all sentient beings, without exception, are like 
golden figures covered with clay: [their] outside crusts are the sheaths 
of defilements, but inside there is the buddha-knowledge 

[9.4] “[Using] the tool of the Dharma, [the Tathagata] then hews away the 
[defilements of those] bodhisattvas who have become calm and cool, so 
that (yena) their defilements are expelled without any remainder.””° 

breaks the sheaths of defilements [and leads them], established in the mental vision of a buddha, 
[to] supreme and perfect awakening (#% | 1 #£##).” Since the subject agrabodhi (“[the one who 
has reached] the foremost awakening,” i.e., the Buddha) of the parallel verse RGV 1.125 appears 
only in the second half of the verse, it is possible that in Ch, the position of the term 4 LiF 
S832 $2 at the end could result from the original Skt. syntax. Such an analysis assumes that 
Amoghavajra tried to stick as closely as possible to the word order of the Skt. text. 

233 As in the prose, Ch, uses different characters for the wax models (#8, Tib: gzugs), on the one 
hand, and the golden figures ({, also gzugs), on the other. The interpretation of the number 
(*Satasahasra) varies: Bth: brgya dang stong' [' for stang]: “hundreds and thousands”, Ch): 
Hee X—T: “... their number is a hundred or a thousand.” Ch, is different from its own 
prose: “It is like innumerable figures of pure gold [in case of] a large [scale] casting: a stupid 
person would look from outside [and] only see burned black clay [molds].” 

234 Ty all other translations pada d of Tib already appears as pada b. The verse in Ch): “When 
the artisan has realized that [the figures in the clay mold] have cooled down, [he] breaks their 
clay[-coverings and] manifests the figures. When the clay has been removed, [he] purifies those 
precious figures. The artisan thinks: ‘[By this] refining process all [figures] will reach 
completion!”” As Bth is hardly intelligible in the last two padas, one cannot know which of the two 
versions, 7ib or Ch, comes closer to the original idea of TGS2. Ch,: “The caster measures the time 
of the cooling down [of the figures and] then opens the mold so that the substance [inside] 
becomes manifest. As the dirt is removed, [the figures’] beautiful appearance becomes clearly 

235 Ch: “With the vision of a buddha I see [all] kinds of living beings like this: within the mud 
of defilements all [living beings] have tathagatahood.” 

236 Ch differs only in the first two padas from 7ib: “Whether [sentient beings who] have 
become calm and cool, [or] bodhisattvas [whose] knowledge has become pure [already] earlier,....” 
In assuming that the two padas mention two different groups I follow the prose of Ch;. As a matter 
of fact, the verse padas could also be understood as indicating the same group of bodhisattvas: 
“When the bodhisattvas, having become calm and cool, with their wisdom [already] earlier 
purified,....”” Bth, apart from the bodhisattvas, also has sems can for sattva: ses can rnams ni 
byangchub sems dpar gyur : [' for na]. As the preceding pada ends with na, the translation should 
be: “[When they are calm...], sattvas have become bodhisattvas.” Further in pada c: “To them [I] 
give the cutting tool [which is] the Dharma.” In Ch, the cooling down is lacking, though it appears 
in its prose. The verse in Ch,: “Bestowing with vajra[-like] insight the smashing of the mold of 


PDF Version: BPPB VI (2002) 

[9.5] “[Living beings’ internal] child of a victorious one who has become clean 
in this [world are] just like the beautiful precious figures: [their] bodies 
are filled with the ten powers [of a buddha], and [they] are venerated 
here [by] the world with [its] gods.””” 

[9.6] “Thus I see all living beings; thus I see also the bodhisattvas. Thus 
purified [by the Tathagata they] become sugatas. [Having become] pure 
sugatas, [they] then teach the rule (netri) of the buddhas.””** 

[10A Merit from propagating the 7GS] 
Then the Exalted One said to the bodhisattva-mahdsattva Vajramati: 

“Vajramati, sons and daughters of good family, whoever—whether a layman 
(grhin) or ordained (pravrajita)— **? learns (*udgrhnati) this Dharma 
discourse (dharmaparyaya) [called] Tathdgatagarbha, preserves [it] 
(*dhdrayati), recites [it] (*vdacayati), understands [it] (*paryavapnoti), 
arranges [it] into a book, explains [it] (*desayati) also to others in detail and 
teaches [it] (*samprakasayati),”*” will produce much”! merit (punyam 

defilements, [the Tathagata] discloses the store of a tathagata [within living beings], just as the 
precious golden [figure] became manifest.” 

37 (1) The content of verse 9.5 is not found in the prose. Verse 9.6 in TGS, perfectly 
complements verse 9.4 because it states that living beings are purified “in such a way” (Tib: de ltar 
dag pa; Bth: de ltar [b]dag; Chz: {05278 }P). “In such a way” could refer to the purifying activity 
of the Tathagata with the “tool of the Dharma” mentioned in 9.4. In this context, verse 9.5 of TGS; 
maintains an extraneous position and interrupts the natural flow of the verses. The verse 9.5 may 
be a later interpolation into TGS,. Verse 9.5 of Ch; corresponds rather with verse 9.6 of TGS. In 
Ch, there is no counterpart to the content of 9.5 of TGS}. 

(2) Ch; lacks the predicate “clean” in the first two padas. This can be explained by assuming a 
simple variant reading: buddha instead of Suddha. This could explain why two terms meaning 
“buddha/tathagata” appear in pada a of Ch (202K, (#8). 

238 (1) Tshul (netri or naya) is used here parallel to dharmaparyaya, which appears in the 
corresponding prose section 9B.5—7. Instead of “rule,” Ch, and Bth speak of “buddha-vision” 
(buddhanayana or buddhanetra). The originality of Tib is supported by the parallel in RGV 1.126d: 
dharmakhyananayaprahara-: “strokes [in form of different] methods (aya) of declaring the 

(2) The verse in Chz: “I see that all sentient beings, thus purified [by the Tathagata], become 
sugatas and, having become sugatas, [they] realize the buddha-vision. [They are] perfectly 
endowed with the supreme knowledge of all [matters] (for sarvajria, sarvajfiata see BGDJ 453d 
s.v. (22%; cf. also the parallel verse RGV I.126c, where the Buddha is called sarvavid).” Pada b 
of the Tibetan versions concerning the bodhisattvas is missing in Ch;. Ch; instead adds with pada 
d a new element also found in the prose of Ch;. However, Ch;, too, speaks of bodhisattvas and 
thus makes Ch; as the original wording of TGS; less probable. Ch,: “[I will] tell the bodhisattvas 
what I have observed in this way (?): ‘O you, keep well [to what I tell you]: Convert all living 
beings!’” It is impossible to integrate pada a organically into the verse. 

239 Tt is not clear if the here mentioned laymen and ordained followers are conceived of as a 
separate group or if they are part of “sons and daughters of good family.” 

240 (1) For dhdrayati (’chang ba) with the two possible meanings of “to preserve” and “to carry 
[on the body]” see Schopen 1977: 142 and Lalou 1957: 328f.; cf. ASPyw 582.5f.: ... prajfia- 
paramita kayagata va bhavisyati pustakagata va, which the cty. (582.9) explains as: kayagata veti 
urogati karanat. pustakagata veti pustakadharanat. 

(2) For glegs bam du byas te ‘jog pa cf. the intransitive equivalents with the siitra as subject in SP 
232.1: pustakagatas tisthet (= Q 100a3: glegs bam du byas te gzhag pa) and KP § 160. Ch, lacks 
this element. 

(3) The enumeration of meritorious acts is found with variations in many other siitras. See for 
example SP 226.3ff: imam dharmaparydyam (1) sakalasamaptam udgrhniyad va' (2) dharayed va 


PDF Version: BPPB VI (2002) 

[10B Devaluation of external worship! 
“Vajramati, a certain bodhisattva~? might apply [himself] to realize the 
tathagata-knowledge, and for the purpose of venerating all buddhas without 
exception in every single world system (lokadhdtu),’” [he] would, after 
achieving supernatural powers (¢ddhi), attain (samapadyate) such an 
absorption (samadhi) that through the power created by [this] absorption 
[he could] day by day present (nirydtayati) pavilions to every single 
existing > tathagata among the buddhas, the exalted ones, even more 
[numerous] than the sands of the Ganges River, in myriads 
(kotiniyutasatasahasra) of buddha-fields even more [numerous] than the 
sands of the Ganges River, together with [their] bodhisattvas and the 
communities of srdvakas.”° [To reside in these pavilions would be] pleasant 

(3) vacayed’ va (4) paryavapnuyad’ va (5) prakasayed’ va (6) likhed va (7) likhdpayed va’ (8) 
likhitva canusmaret.... 
[! va inserted acc. to O; ?O: desayisyamti for vacayet; * O: paryapsyamti; ‘O: pravartayisyamti; *(7) missing 
in O and the Tibetan translation.; ° (8) in O: likhitarm va ‘nusmarisyamti] 
= Q 98al: ... chos kyi rnam grangs ‘di (1) mtha’ dag chub par ‘dzin tam (2) ‘chang ba’am / (4) 
kun chub par byed pa ‘am (5) rab tu ston pa ‘am (6) yi ger ‘dri ba ‘am (8) bris nas rjes su dran 
par byed pa’am.... 

Further see SP 375.5f.: ... cemam evamripam sutrantam (1) dharayisyanti (2) vacayisyanti (3) 
deSayisyanti (4) paryavapsyanti (5) parebhyas ca vistarena samprakaSayisyanti .. 
= Q 159b2f.: ... mdo sde ‘di lta bu ‘di (1) ‘dzin pa dang (2) klog pa dang (4) kun chub par byed pa 
dang (3) ’chad' ba dang (5) gzhan dag la yang rgya cher yang dag par rab tu ston pa... 
[' for ‘chang]. 

Ch, adds the veneration ((3) of the siitra. 

241 Chi: ANBTE+S and Chy: #4: “measureless [merit].” 

2 Only Ch; adds in this passage “sons and daughters of good [family]” (AB BT. & 
2A). Ina similar passage KP § 158, which will be discussed below in n. 246, only Aulaputra and 
kuladuhitr appear. 

3 Tib shows, once again, a unique arrangement of the elements in the first section of the 
sentence. Bth states that the bodhisattva applies himself (1) to the realization of tathagata- 
knowledge as well as to the veneration or, as a second alternative, (2) to the veneration in order to 
realize... (de bzhin gshegs pa'i yeshes bsgrub pa’i phyir ; ++) sangs rgyas thams cad la mchod 
pa't phyir : nan tan du byed cing : ...). Chz only speaks of the accumulation of tathagata- 
knowledge as the purpose of the striving and veneration (#44), (£3). Ch, does not mention the 
veneration but states as the purpose “the [realization] of the buddha path” (3(#isdD). It is 
difficult to judge which version of TGS, should be considered as coming closest to the Skt. 

4 Tib and Bth both mention the entering into samadhi after the attainment of supernatural 
powers. The reverse order would be more appropriate. In Ch, the attainment of the rddhis does not 
appear. Instead of the “power created by this absorption” Ch, merely reads “the rddhis of this 
samadhi” (jth = bik 7I). However, Ch, seems to support the Tibetan: (£74 7H38, AZB=RK 
(translation below). The syntax of the Skt. on which 7ib and Ch are based, may be similar to SP 
406.6: (sa tasyam veldyam) tathariupam samddhim samapanno yasya samadheh... (cf. also SP 

245 The fact that two verbs considering the existence of the buddhas appear in the Tibetan (ib: 
bzhugs shing gnas pa in 10B.8; Bth: bzhugs shing spyod pa in 10B.6) reminds one of formulations 
as found in e.g. SP 42.2: ... tathagatd ... tisthanti dhriyante yapayaniti...; cf. also BHSD s.v. 
yapayati (2). Both Chinese versions operate instead with *pratyutpanna (EAZE): “existing at the 
present moment.” 

246 (1) The passage would remain difficult, if we did not know a similar passage in KP §§ 158- 
159. Through the passage of the KP it becomes clear that the relation between sangs rgyas bcom 
Idan ‘das in 10B.7 and de bzhin gshegs pa re re in 10B.8f. is that of a partitive genitive: 
gamganadivalukasamanam ca buddha[nam] bhagavantanam ekekasya ca tathagatasya.... 
(§158.5f.); gamganadiva[lukasa]manam ca buddhanam bhagavatam ekekasya ca tathagatasya.... 
(§159.1f.) [letters in brackets in case of mutilated passages supplied from Weller 1965: §159, n. 4 and 6; for 


PDF Version: BPPB VI (2002) 

in [every] season.”*” 

[their] height ten yojanas. [They] would be made of all [kinds of] jewels, 
[and would be] heavenly fragrant, being strewn with a variety of fallen 
blossoms 74? and furnished with all immaculate (anavadya) objects of 
enjoyment (bhoga). [The number of pavilions would be] as many as fifty 

[Their] width [and length would each be] one yojana, 

the conjecture ekekasya for ekekam see his n. 7]. In 10C.2 the same partitive relation is clearly 
expressed with Jas: ... sangs rgyas ... de dag las de bzhin gshegs pa re re.... Further, Ch, seems to 
split the passage into two separate sentences and starts the second sentence with 4112/5338 
HE RTE ... Hie... HBBA, ....: “In this way (= venerating each single buddha as said 
before: fS——(#pF GE 7 SH), [he] would each present ... pavilions ... [so that the whole number 
of] buddhas, exalted ones would result in more than fifty [times] the sands of the Ganges River.” 
This is a somehow forced translation, but I believe that it is close to what had been originally 
focused on by the Indian authors. My interpretation in this point follows Ch; (see below). In the 
Tibetan translations there is no equivalent found for the passage which states that the number of 
buddhas would gradually accumulate. Neither does the passage of the KP prove such a thought. 

(2) Ch, and also Bth attribute kofiniyutasatasahasra not to the buddha-fields but to the tathagatas 
themselves being present in the buddha-fields. Also, in contrast to Tib, Ch) only speaks of a 
hundred thousand of pavilions (in the next passage) and instead attributes the number of “as many 
as fifty [times] the sands of the Ganges River” to the buddhas, exalted ones. Bth is ambivalent in 
this last point and could well represent the position of the phrases in the Skt.: de bzhin gshegs pa 
re re la chu bo gang ga Inga bcu’i bye ma snyed kyi khang bu brisog pa brgya phrag stong.... The 
position of chu bo gang ga Inga bcu’i bye ma snyed for *paficasadganganadivalikopamah 
between “tathagatas” and “pavilions” could easily lead to the different attribution as attested in Tib 
and Ch). 

(3) The parallel passage in the KP is too long to be quoted. It is the answer by the Buddha to the 
question by Kasyapa as to how much merit would be produced by someone who teaches the 
Ratnakuta (of which the KP is a part) (§157). Structurally the answer is very similar to the passage 
in the TGS. The elements, however, vary in stating that sons and daughters of good family would 
fill up all the worlds with the seven jewels and present them to the tathdgatas. They would 
venerate the tathagatas for many cosmic cycles and finally erect seven-jeweled stipas for those 
tathagatas who have entered into nirvana (§§158-159). It is interesting that in the KP the Skt. has 
several repetitions and is much longer than its Tibetan translation. This seems also to hold true for 
the Tibetan translations of this passage in the 7GS in relation to the more repetitive Chinese 
versions. Besides the general structure of opposing external worship on the one hand, and 
preservation of just a single verse of the siitra on the other hand, a close relation can be seen 
between the 7GS and the KP in this passage in terms of single elements: respect to the buddhas in 
two ways (veneration and giving of presents), the creation of stiipas or kutagaras, the mentioning 
of the sra@vaka community which does not appear in other passages of the TGS, the use of the 
expression ganganadivalukasama (seven times in KP §§158-159), and finally a fondness for the 
attribute saptaratna. For the second passage in this section see below. 

47 For dus du bde ba cf. the similar expression in the Subhdsitaratnakarandakakatha 
(Zimmermann: 1975, verse 102): prasade ... sarvarturamye...., “im Palaste ... zu allen Jahreszeiten 
behaglich....” Ch, (SHS #0KB. 298) does not attribute it to the pavilions but specifies here 
the act of presenting (38): “[to present] at a very pleasant, peaceful time.” 

248 Bth and Ch, read saptaratna instead of sarvaratna. 

a (1) The Skt. for me tog sil ma ... bkram pa should be muktakusumabhikirna (SP 103.4; = Q 
46b5: me tog sil ma gtor ba) or muktapuspdni sthdpayati (Suv 105.9; = Suv, 81.6: me tog sil ma 
dgram par bgyi). Skt. mukta characterizes the blossom as “loose, fallen (from its stalk)” (BHSD 
s.v. mukta). It is translated into Tibetan as sil ma in the sense of “little piece, a fragment” (Ja s.v. 
sil bu). Chz has #48) 7: “(different) blossoms, being dispersed.” 

(2) Ch, and Bth do not mention the blossoms at all. Instead, Ch, speaks of palanquins spread with 
heavenly silk (AXA #8). As, on the one hand, the blossoms do not appear in the verses of any of 
the versions, and, on the other hand, in 10.5 and 10.6 of all versions silk-spread thrones are 
mentioned, it is quite likely that the blossoms are a later element, interpolated into the prose of 
TGS, (but not in Bth!). 


PDF Version: BPPB VI (2002) 

[times] the sands of a hundred thousand Ganges Rivers. For fully a hundred 
thousand?” cosmic cycles (kalpas) [he] would show reverence in this way.”*! 
If [on the contrary] a certain son or daughter of good family” should 
forge the resolution (citta) to [strive for] awakening (bodhi), and internalize” 
or arrange into a book (pustaka) only one simile from this Dharma discourse 
[called] Tathdgatagarbha, [then], Vajramati, **~the previously [described 
bodhisattva’s] accumulation of merit™* does not come near by even a 
hundredth, a thousandth, a hundred thousandth—any number, any tiny part 
(kala), any calculation or any resemblance—to his accumulation of merit 
(punyabhisamskara); nor does [it] bear any comparison.”*°° 

250 Ch, has “up to a thousand kalpas” (J528--$h); Ch: 5246 BT BH (see below). 

5! (1) Both Tibetan translations have pas at the end of this passage. This is surprising as one 
would rather expect a particle indicating a contrast such as Ja (cf. Ch, which has 740; 740 can 
function as an equivalent for anyatha (cf. BCSD s.v.): “on the contrary”). This is in fact the case in 
the parallel 10C.4. However, the employment of pas is also attested for the parallel passage KP 
§159 without any direct counterpart in the Skt. Also there both Chinese translations read 4\ 40. Pas 
was probably added here as a particle in order to indicate the comparative relation the passage is 
dealing with. Grammatically, however, this comparative relation is already expressed with ... la 
(10B.18: “in [comparison to] ...”), and the whole first part of 10B till ... byed na (10B.17) can 
only be understood as an introduction which leads to the actual comparison and which describes 
the two competing ways of gaining merit (which should, as mentioned above, be separated with a 
contrasting element rather than with pas). 

(2) The section in Ch,: “Vajramati, if a bodhisattva, in order [to realize] the path of a buddha, 
would practice hard, progress skillfully, cultivate supernatural powers, enter into the samddhis 
[and], desiring to increase the roots of [his] merit, venerate all the present buddhas, more than the 
sands of the Ganges River, [and if he] would create more pavilions than the sands of the Ganges 
River, [made of] the seven jewels, [their] height ten yojanas and equally one yojana in width and 
length, furnished with palanquins [made of] the seven jewels [and] spread with heavenly silk, [and 
if he] would day by day create more pavilions than the sands of the Ganges River, [made of] the 
seven jewels, for every single buddha, [and if he would] present (LA FH#@X for *niryatayati?) 
[them] to each single tathagata and also the bodhisattvas [and] the community of sravakas, by 
these acts [he] would thus widely for (?) all present buddhas, more than the sands of the Ganges 
River, successively arrive at [a number of] pavilions more than fifty [times] the sands of the 
Ganges River, [made of] all jewels, [which he] would present to the present buddhas more than 
fifty [times] the sands of the Ganges River and also the bodhisattvas [and] the community of 
Sravakas, up to innumerable hundreds, thousands, tens of thousands of kalpas long.” 

The text of Ch;, too, is in several instances problematic. The fact that the bodhisattvas and the 
Sravakasamgha appear a second time towards the end of the passage can be explained as erroneous 

252 Ch, has bhiksus, bhiksunis, updsakas and upasikas. The passage in KP is like Tib. 

253 My translation “ to internalize” is for /us la “chang, Skt. kayagata; Ch: 7E}*§. For the pair 
kayagata and pustakagata cf. e.g. SP 282.11 and the quotation of ASPyy in n. 240 (1) where 
kayagata (variant: kanthagata) is explained as urogatikarandt. 

254 The text part between “4°” and “*?*” is lacking in Ch, and has no counterpart in the 
verses of all versions. As it is of repetitive nature (cf. 10C) it could well be a later interpolation in 

255 In Ch2 403R# 17 is added as in the parallel 10C : “[the previous accumulation of beneficial 
acts which] was established [in] the tathagatas [as field of merits (punyaksetra)] (?).” The 
characters #137 could alternatively mean that the Tathagata “stipulated” the then following 
comparison (cf. n. 260). 

2°6 The basic structure of the whole of 10B becomes clear due to parallels, again with different 
elements, in other siitras. See e.g. SP 332.9ff.: 

kascid eva kulaputro va kuladuhita va ... pancasu paramitasy ... caret / ... yena cdjita 

kulaputrena va kuladuhitra vemam ... dharmaparyayam Srutvaikacittotpadikapy adhimuktir 

utpaditabhisraddadhanata va kta / asya punyabhisamskarasya kusalabhisamskarasyasau 
paurvakah punyabhisamskarah kusalabhisamskarah ... Satatamim api kalam nopayati 


PDF Version: BPPB VI (2002) 

[10C Superiority of joyful approval of the TGS] 

“Then, Vajramati, [suppose] a bodhisattva, in searching (paryesate) for the 
Dharma of the buddhas,”*’ strewed four hundred thousand triple-bushels of 
flowers of the coral (mandara) tree for every single tathagata among the 
buddhas, the exalted ones, for fully a hundred thousand cosmic cycles.?°8 

Vajramati, if [on the contrary] any monk, nun, updsaka or updsika 
should decide [to strive] for awakening, and after listening to this Dharma 
discourse [called] Tathagatagarbha raised [their] joined palms and said [just] 
the single phrase ‘I joyfully approve [what I have heard]!’ (anumodayami),’” 
[then] Vajramati,“*™* the previously [described bodhisattva’s] accumulations 
of merit and benefit, connected with the offering of flowers and flower 
garlands, planted among the tathdgatas [as fields of merit (punyaksetra)],° 

Sahasratamim api Satasahasratamim api kotisatasahasratamim api kotinayuta- 
sahasratamim api kotinayutasatasahasratamim api kalam nopayati samkhyam api kalam 
api gananam apy upamam apy upanisam api na ksamate / 

= Q 143b4-144al: 

rigs kyi bu’am rigs kyi bu mo la la ... pha rol tu phyin pa Inga po ... la spyod de ... na / ma 

pham pa rigs kyi bu ’am rigsi?? bu mo gang gis ... chos kyi rnam grangs ‘di thos nas / sems 

bskyed pa gcig tsam zhig mos pa bskyed dam mngon par dad na de’i bsod nams mngon par 
‘dus byas pa ‘di la/... yongs su rdzogs par bsod nams mngon par byas pa ... snga ma des 
brgya’i char yang nye bar mi ’gro / stong gi cha dang brgya stong gi cha dang ... yang nye 
bar ni ’gro ste / grangs su yang char yang bgrang bar yang dper yang rgyur yang mi bzod 
do I/ 
Cf. also the passage in KP §§158-159: 

kulaputro va kuladuhita va ... stipa kardpayet % yas ca kulaputro va ... ekam api gatha 

udgrhneya dhara[yet] asya punyaskandhasya sa pirvakapunyaskandhah satimam api / 

kal@<m> nopaiti / ... upanisam api / na ksa[mate].... 
It is clear that in this passage Bth follows much closer the Skt. as Tib and the Tibetan translation of 
the SP. Ch; has an enumeration even longer than the one in the SP. With bye ba Bth, too, shows a 
part of kotiniyutasatasahasra which obviously was not translated in Tib. The first half of the 
passage in Ch,: “Vajramati, [if] on the contrary (75411) somebody would take pleasure in the 
awakening [and] would learn ... and venerate the TGS, even if it would just be a single simile,....” 

257 Along with KP §158 Ch has kulaputra and kuladuhity instead of bodhisattva; instead of the 
Dharma of the buddhas: “supreme awakening” (4 | 2 #2). 

58 (1) The uncommon construction ... de dag Jas in 10C.2 for a genitive in the Skt. is also 
found in the KP: sangs rgyas bcom Idan ’das ... rnams las de bzhin gshegs pa re re’i... (§§158.5f., 
159.1f.) for... buddhdnadm bhagavantandm ekekasya ca tathagatasya.... However, the construction 
with /as is clearer than the formulation in 10B8f. with a genitive in Tib (cf. note 246 (1)). Chz does 
not show a partitive relation and speaks as in 10B of “buddhas, exalted ones together with 
bodhisattvas and the great communities of sravakas” (#- R# fe. BlHIA5R) to whom they would 
make offerings with (AX) the flowers. 

(2) Instead of a hundred thousand kalpas, Ch; and Bth have only a thousand kalpas. Instead of 
four hundred thousand kharas, Bth has kotisatasahasra, whereas Ch; (Fi---ff}) reads “a hundred 

259 The important function of “joyful approval” (anumodand) in order to aquire merit becomes 
clear through a verse of the SP (93.1f. = 111.106), in which the Tathagata says to Sariputra: 

yas capi te bhasati kasci sattvo anumodayamiti vadeta vacam / 
murdhnena cedam pratigrhya sutram avivartikam tam naru dharayes tvam // 106 // 
And if any sentient being would say to you ‘I joyfully approve [this teaching]!’ and 
would humbly accept this sfitra, you should remember that person as [somebody who 
will] never [again] turn back [from supreme awakening]! 
For further passages concerning anumodand resulting from the “rejoicing at the merit of others” 
see e.g. PraS 23A-23G, chapter 6 in the Asfasadhasrika Prajndparamita and MPPU, IV.1879f. 

260 T am not sure whether I understand Tib here in the right way. The same passage in Bth reads 

de bzhin gshegs pa metog phreng' ba dang : bcas pa phul zhing : gtor ba [' for phring]. Instead of 


PDF Version: BPPB VI (2002) 

do not come near by even a hundredth, a thousandth, a hundred thousandth— 
any number, any tiny part, any calculation or any resemblance—to those 
accumulations *°! of merit and benefit; 7 nor do [they] bear any 

Then at that time the Exalted One uttered these verses: 

[10.1] “[Suppose], having brought forth the wish for awakening,” a certain 
sentient being listened to this [discourse and] learned [it], copied [it] or 
arranged [it] into a book*® and explained [even just] a single verse with 
appreciation (sagaurava);””®° 

[10.2] “or if after listening to this Tathagatagarbha[sutra] somebody searched 
for this excellent awakening: listen to [my description of] the benefit 
(anusamsa) [accruing] to him in these cases—[a description of] what 
amount of merit is produced!”?°” 

bzhag pa (from ‘jog pa) Bth uses gtor ba, which is the same verb as used above, with regard to the 
flowers (“to strew”). The verb ‘jog pa in Tib, however, cannot have the meaning “to strew.” It is 
hardly possible to bring sense into the wording of Bth. My iranslation of Tib leads to a similar 
understanding as expressed by the Chinese versions. Ch2: “Vajramati, as for these best benefits 
and wholesome roots [in comparison] with the previous wholesome roots, [i.e.] the merit [resulting 
from] offering the flowers, established [in] the tathagatas [as field of merits (punyaksetra): if one] 
compares [these best benefits and wholesome roots with] the previous merit, a hundred parts [of 
the previous merit} ... do not equal one part [of it].” The formulation 41362237, which is also 
found in 10B, may, as already suggested in n. 255, alternatively be understood in the following 
way: “... the merit [resulting from] offering the flowers, the Tathagata has stipulated: [If one] 
compares...” But cf. Ch; which has #24 (##AT&: “[the wholesome roots ...] planted among the 

261 Both forms mngon par ‘du byed pa in 10B and mngon par ‘du bya ba here are most 
probably translations of the same Skt. abhisamskara. Present, perfect and future forms of ‘byed pa 
are also found in the translations of abhisamskara in the SP (cf. SP;s.v.). 

2 Tih employs here the pair punydbhisamskara — kuSalabhisamskara; Bth uses once 
kuSalamilabhisamskara instead of kusalabhisamskara; Chz speaks, in the case of the joyful 
approval, of “these best benefits [and] wholesome roots” (LRH. #¢#8), and compares them 
with “the previous wholesome roots, [i.e.] the merit [resulting from] offering the flowers” 
(UR, REDE). 

263 The whole section in Ch: “Vajramati, even if the wholesome roots [and] the benefit of this 
son of good [family], planted among the buddhas [as the fields of merit], are measureless, 
compared with the merit attained by the son and daughter of good [family who propagated the 
TGS}, a hundred parts [of the former merit] do not come near one [part of this merit],....” 

264 The Skt. on which Tib: mos pa bskyed nas is based, could be chandam janayati. Bth has 
dad' bskyed pa [' for ‘dad] instead, Ch, reads 3%; Chz: #5. Cf. e.g. SP 47.13 = I1.62b: bodhaya 
janetha chandam /. The translations of Bth and Ch, are not unusual. For mos pa as an equivalent 
for chanda see BCA;8.v. mos (pa) 3. 

26 7 cannot explain why the majority of the collated Tibetan texts has the future form gzhag 
instead of the present form 'jog, which appears in the prose of 10A. P;2; read bzhag. 

266 Ch,» do not mention that the verse would be explained (to others). The verb bshad (perfect 
of ‘chad) could (though not in the passage in 10A) also represent a form of the root path: “to 
recite” (see BCA, s.v. bshad pa). Kagawa (1962: 19) translates according to the variant of P;3 in 
10D fn. 7 (pa’i for pas): ... FAO—1h & Mi < : “... and explains a single verse of veneration.” But 
cp. Bth: sti stang byas te :. 

287 Bth and Ch) bear a similar message as Tib. Bth has bsngag pa, Skt. prasamsd or varna 
(“praise, glory”), instead of phan yon for anusamsd (Ch): #4 probably for anusamsa). Ch, differs 
considerably: “The subtle store of a tathagata in an instant brings forth joyful approval (aE; 
anumodand). Listen [therefore] to this correct doctrine [and your] merit will be measureless!” The 
search for awakening in pada b of 7ib appears in Ch, only at the beginning of the following verse 


PDF Version: BPPB VI (2002) 

[10.3] “[Suppose] a hero (vira) abiding in these excellent supernatural 
powers *® worshiped for a thousand cosmic cycles the highest of 
humans (narottama) and [their] srdvakas in the ten directions.” 

[10.4] “[He] would present to each teacher (acarya) of the world without 
exception excellent palaces (vimdnasrestha) made of jewels—in 
number several myriad [times the sands of] the Ganges [and] more, 
unimaginably (acintya) many.”””° 

[10.5] “The [palaces] would be ten yojanas high and one yojana*’' wide and 
long, be excellently furnished with fragrances and incense, and inside 
be provided with thrones made of jewels.”?”” 

{10.6] “[These] thrones, and palanquins too, [would be] spread with silk and 
calico a hundred [times],”” as innumerable as [the sands of the] Ganges 

10.3, and should be the counterpart to the realization of tathagata-knowledge at the beginning of 
section 10B. That Ch,, in relating the search for awakening (Zib 10.2b) to 10.3, is probably the 
more authentic version, is supported by Bth and Ch; where the same pada appears at the end of 
verse 10.2. If we relate the first verse half in Bth and Ch; to 10.1 and combine its second half with 
10.3, we arrive at the same basic structure as found in the prose, and should consequently consider 
it as the original pada sequence. 7ib, on the other hand, has altered this authentic pada sequence 
and thus has to be translated in a completely different way. None of the other versions, 
furthermore, show an equivalent to the particle ji tsam for *kiyat as found in Tib. 

268 Ch>: “... by [use] of supernatural powers [he} resides in the upper vehicle” (J1##i J0{E 


o dPa’ bo, vira, is only attested for Tib. It must designate the person dedicated to external 
worship. Bth (rdzogs) and Ch, (#2) have purna instead: “for fully a thousand kalpas.” In Tib, 
dPa’ bo appearas again in 10.9. Ch; mentions the period only in 10.7, adds “bodhisattvas” to 
“Srdvakas” and reads {§ instead of narottama. 

210 Instead on dcarya both Bth (spyod) and Ch, (#7) are based on dcara; Ch,: {#§. I have found 
the compound vimanasrestha only in the SP. For the meaning of vimana see PTSD: “palace 
chariot.” Ch, attributes the first two padas to the kalpas. Amoghavajra translated kalpa (4) 
instead of koti. 

27! Ch, gives width and length of the palaces with [-+-&, which corresponds to about 23 km. 

27 (1) Fragrances (gandha) and incense (dhipa(na)) do not appear in Ch, (as in the prose of 
Ch,). Instead, Ch, characterizes the palaces as “majestically decorated [and] furnished with many 
wonderful [things]” (RBH ZRH2). Instead of “excellently furnished with” (rab Idan), Ch; reads 
{it3@, meaning “to venerate” (rendering a form of Skt. upa-stha?). 

(2) bShams in pada d is a translation of prajfapta (cf. Ch,: ff). The verb is missing in Bth and 
Ch). But cp. 10.6b where Ch, has Bxa for prajfiapta. 

23 (1) The Skt. on which dar dang bcos bu’i ras (Bth: dar dang gos; Ch;: (K)#2#9) is based, 
may be pat(t)adiusya (Kpun 56.4f.: nandpattadusya- ... dsanani prajhaptani /). More common is 
the compound disyapat(tja, which is analyzed as a karmadharaya: BHSD s.v. pata (1): “some 
kind of fine cloth.” This is how Ch, understands the compound (484k; cf. BWDJ s.v. patta: $84 
for patta) 

(2) Instead of brgya in pada a, Bth reads brgyan. “Hundred” is also lacking in the Chinese 
versions. brGyan could be a translation of alamkyta. That would explain BR&ffiffg of Ch, in the 
preceding pada. bTing in pada a of Bth should probably, as is indicated by the “colon” before it, be 
constructed with the following pada, in the sense of “furnished” (*prajfiapta) with thrones and 
palanquins. Its counterpart appears in Ch, only in pada b: $24. With the characters BH + 
“spread over [the thrones],” on the other hand, Ch, would use a more specific term than alamkyta. 
It thus is very likely that what appears as brgya in Tid results from brgyan, and that bting, which 
can be understood in two ways, namely “furnished with” or “spread with,” was consequently 
interpreted in the latter way. 

(3) I understand khri as “throne” (*dsana) and khri stan (Bth: stan) as “palanquin” (*samstara). 
The palanquins are not mentioned in the verses of the Chinese versions, but appear in the prose of 


PDF Version: BPPB VI (2002) 

River; [he] would present [these palaces with thrones] to each 
victorious one.” 

[10.7] “Upon the victorious ones who reside in world systems—those 
victorious ones more [numerous] than the sands of the Ganges River— 
[he] would thus bestow [these palaces, and] would venerate [them] all 
with appreciation.””” 

[10.8] “If [on the contrary] some wise [person], having listened to this sitra, 
learns only one single simile correctly, or having learnt [it], explains [it] 
to somebody [else, then] he will [produce] a greater amount of merit 
thereby [than the previous person].”?”° 

[10.9] “Regarding the [former] merit seized by the hero [who worshiped the 
tathagatas, it] does not come near by any tiny part or resemblance [to 
the merit of this wise person]. [The wise person therefore] becomes a 
refuge (Sarana) for all living beings, and he quickly attains excellent 

[10.10] “The wise bodhisattva who reflects upon [the following]: 

‘The tathagatagarbha exists in the same way [in all beings]! This is 
the true nature (dharmata) of all sentient beings,’ 

will quickly become an awakened one [through] his own power 

Ch,. gZhan rnams kyang (Bth: gzhan yang) is probably a translation of anyac capi (Chz: MER). 
The ‘pada in Ch,: “... according to the throne each [cloth] would be very different” (2S AS). 

2 Ch, instead of “each victorious one” has “the buddhas and [their] great communities [of 
bodhisattvas and sravakas]” (B® RAR). 

275 (1) Instead of lokadhdtu Ch, has ksetra (Ail). 

(2) Ch, is quite different: “[He}] would venerate/bestow all with it, day and night without rest, 
fully a thousand millions of kalpas long. [Regarding] the merit [he thereby} gained, [the situation] 
would be such:....” Pada a of Ch, corresponds with pada d of all the other versions. Pada b and d of 
Ch, have no counterpart in TGS;. Pada c, comprising the number of kalpas, appears in a similar 
form already in verse 10.3 of TGS). Since padas b and c of Tib (= c and a in Ch and Bth) repeat 
just what had been said in the preceding verse 10.6 and the position of the specification of the 
number of kalpas in Ch, (at the end of the description of the external worship) is identical with 
that in the prose of all versions, it is most likely that Ch, shows more resemblance to the original 
Skt. wording. In order to bridge over missing padas TGS; might in fact have merely filled up the 
lacuna by repeating the padas of the preceding verse. On the other hand, pada d of Ch;, when 
compared with the prose, appears too early and seems better placed in Tib and Bth, where it is part 
of verse 10.9 (pada a in Tib; b in Brh). 

276 (1) Instead of “learning” the siitra, Bth speaks of its “preservation” (‘chang for *dharayati). 
Ch; introduces a second verb in pada b: “to take hold of one simile and practice correctly” 
(AX TE 77). 

(2) My translation takes into account the particle pas at the end of verse 10.7, and therefore uses 
the comparative form “greater.” 

277 (1) In Ch padas a and b are missing. Vira (dpa’ bo) of Tib, as in 10.3, is not found in any of 
the other translations. Ch, identifies the refuge with the TGS: “Sentient beings take refuge in this 
stra...” (AB RTKPAULES). In Ch, sentient beings (and not the wise person) attain awakening. In 
the Tibetan, however, it is not clear whether it is the wise person or living beings who are meant to 
attain awakening. 

(2) After the first half, the verse seems to enter abruptly into a new theme. As this second half of 
10.9 and the whole of verse 10.10 are not contained in the prose, it is quite likely that they are a 
later interpolation into the common ancestor of TGS, and TGS). 

78 The verse in Ch,: “The bodhisattva reflects about the very profound store of a tathagata 
{and} knows that [it] is found [in] all living beings[. He/... and that they] will quickly realize 
supreme awakening (f&_/38).” Ch2: “If a wise bodhisattva reflects about the qualities (74) 
connected (*9E; *samprayukta, *sahabhu or something similar) with this store of a tathagata 


PDF Version: BPPB VI (2002) 

[11 The story of *Sadapramuktarasmi and *Anantarasmi] 
[11A The appearance of *Sadapramuktarasmi] 

“Vajramati, again, by way of this [following] kind [of exposition] (parydaya), 
[one] should know thus: namely, that this Dharma discourse is extremely 
beneficial (bahukara) for bodhisattva-mahdasattvas [because it] will lead to 
the realization of the knowledge of an omniscient one (sarvajfiajfiana)?” 

Vajramati, formerly, in the past, innumerable, vast, measureless, 
unimaginable, unparalleled and [quantitatively] inexpressible cosmic cycles 
[ago], [and] even more beyond the other side of that [time]—then, at that 
time—there appeared in the world the tathagata, the honorable one and 
perfectly awakened one, named *Sadapramuktarasmi, realized in wisdom and 
conduct, a sugata, a world-knowing one, a charioteer of human beings to be 
tamed, unsurpassable, a teacher of gods and men, a buddha, an exalted 

[thinking]: ‘All sentient beings [have?] the highest dharmata,’ [then he] will quickly awake to 
spontaneous knowledge ({ #4; svayambhijriana).” The same pada sequence is also attested for 
Bth: “‘T have the tathagatagarbha.’ Bodhisattvas who reflect [like that and the following]: ‘The 
dharmata {belongs to] all sentient beings,’ [they] all without exception will become awakened 
ones [through] their own power.” Bth has thams cad instead of myur du in Tib (pada d). I have no 
idea how to explain most of the differences which occur in pada a. The readings ‘dra (Tib) and 
bdag la (Bth), however, could derive from the variants sadysa/mdadrsa or sama/mama. 

Similar to verse 1.1, we may also here assume that the first pada, composed in Buddhist Hybrid 
Sanskrit, originally had something like *sama(s) tathagata garbha asti, with garbha representing a 
locative form: “A tathagata exists in the same way within [all living beings].” The translators, 
however, would have understood automatically tathagata garbha as a compound and translated it 

The verse 10.10 (and the second verse half of 10.9) have no counterpart in the prose and they 
may well be a later interpolation into the common ancestor of TGS, and TGS>. If we assume that at 
the time of the interpolation the term tathagatagarbha in the meaning “tathagata-embryo” was 
already prevailing, we could also understand the pasage as rendered in Takasaki 1981: 34: ... 
TTOREOAIZ, CORMRO BROT EXRHEOHS HL... “.. the fact that in all 
living beings exists this real nature like a tathagata-embryo....” The verse may well be an 
interpolated quotation from another siitra. 

Verse 10.10 summarizes the central message of the text and may have functioned as a verse at the 
end of the siitra before the final description of the effect on the onlookers (12D). This would lead 
to the highly speculative assumption that sections 11A till 12C were interpolated into the main text 
at a later point. 

279 The passage is neither contained in Ch, nor is there any counterpart in the following verses. 
Cf. the similar introduction to the story of Sadaparibhitta in the XIX" chapter of the SP: anenapi 
tavan' Mahasthamaprapta paryayenaivam veditavyam yatha ya imam evamrijpam dharma- 
paryayam ... (375.1—2) [' tavat is missing in D,, D,, O and Q]: mthu chen thob rnam grangs ‘dis kyang ‘di 
ltar rig par bya ste....}. Similar constructions are also found in SP 43.2f. and 82.9f. For further 
discussions of similarities between the story of *Sadapramuktarasmi and the one in the SP see 
Zimmermann 1999: 159ff. 

780 (1) Apart from the words printed in boldface and the additional mtshungs pa med pa 
(*asama) and brjod du med pa (*an-/nir-abhilapya) in the TGS, the passage is in accordance with 
SP 375.9ff.: bhiitapirvam ... atite ‘dhvany asamkhyeyaih kalpair asamkhyeyatarair vipulair 
aprameyair acintyais tebhyah parena paratarena yad asit tena kalena tena samayena 
Bhismagarjitasvarardjo nama tathdgato ’rhan samyaksambuddho loke udapadi vidydcarana- 
sampannah sugato lokavid anuttarah purusadamyasarathih Sasta devanam ca manusydnam ca 
buddho bhagavan.... SP 156.1ff is also similar; for the list of epithets see MVy 1-11 and MPPU, 
1.115ff. The positioning of the name of the tathagata in my translation corresponds with the Skt. of 
the SP, of Bth and Ch,2. 

(2) Ch, adds -rdjan to the name: (3% #(¢fA 4). For another possible reconstruction of the name 
as Sadarasmimukta see Zimmermann 1999: 159, n. 37. 


PDF Version: BPPB VI (2002) 

[11B The light from the bodhisattva fills the world systems] 

“Vajramati, why is that tathagata called *Sadapramuktarasmi? Vajramati, 
immediately after the exalted one, the tathagata *Sadapramuktarasmi, then a 
bodhisattva,”®’ had entered the womb of [his] mother,” light was emitted 
from [his] body, while”*? [he was still] within the womb of his mother, [so 
that] in the east hundreds of thousands of world systems, as many as the 
atomic-sized dust of ten buddha-fields,”* came to be constantly filled with 
brightness.”* In the same way, in the [other nine of the] ten directions, 
[namely] the south, the west, the north, the southeast, southwest, northwest 
and northeast, along with the nadir and the zenith, hundreds of thousands of 
world systems, as many as the atomic-sized dust of ten buddha-fields, came 
to be constantly filled with brightness. And owing to the pleasant and 
beautiful light from the body of that bodhisattva, [which] caused [sentient 
beings] to rejoice [and] led [them] to delight,7®° as many as hundreds of 
thousands of world systems constantly came to be filled with brightness.” 

[11C The light improves sentient beings] 

“Vajramati, all sentient beings in the hundreds of thousands of world systems 
who were touched by the light from that bodhisattva within the womb of [his] 
mother attained strength (ojas), beauty (varna), mindfulness (smrti), 
comprehension (mati), understanding (gati) and readiness of speech 

281 Ch, runs AX{F356ESHEY: “... when [he] practiced the path of a bodhisattva,....” Cf. e.g. Kpun 
46.17f.: ... tathagatena piirvam bodhisattvacaryam caratd ... = Kpunc, (174413): Af. 

2 CE, Kpuncy: BEFRELAG (268b19; = Ch,) for avataranam mdtur garbhe (249.5): “descending 
into the womb of the mother.” 

28 bZhin du does not appear in any of the other versions and I can only understand it in its 
function of forming a present participle (cf. Jd s.v. bzhin 2.): “while....” The parallel verse 11.1 
runs ... lus las ‘di ’dra’i ‘od ... and ‘di ‘dra must be interpreted differently; cf. the position of 
tatharipa in the quotation of the Lalitavistara below in n. 285! There it has to be construed with 
prabha (‘od). 

284 The Skt. was probably *dasa-buddha-ksetra-paramGnu-rajah-sama lokadhatavah (cf. e.g. 
Kpun 11,8; 80.11: sahasra-buddha-ksetra°). The number of ten buddha-fields is also attested for 
DbhS C3. Ch; mentions only the illumination of [buddha-]fields, as many as the atomic-sized dust 
of a thousand buddha world systems in the ten directions (+7y--(@BtE FRR ESA). Ch, 
continues then with 11C, which incorporates the last issue of 11B regarding the pleasure of living 
beings at its beginning. Since the long-winded description dealing with the other directions is 
missing in Ch, and is not mentioned in the verses of TGS), it is plausible that it was only added at 
a later date to the prose of TGS). 

285 The basic structure of the second half of this passage is found in the Lalitavistara (cf. Lalyx 
374.14-17): ... bodhisattvena tathariipa kayat prabha pramuktabhid, yaya prabhayayam ... 
lokadhatur evam ... avabhdsena parisphuto ’bhit. bZhin du in 11B.4 could be a translation of 
tatha after the second part (riépa) had erroneously been read as kaya (in many Indian scripts ri is 
similar to ka; pd is similar to ya) and dropped by a scribe who thought of it as a haplography as the 
following word was kayat. 

785 Ch, has ... HAS ASMA, FE BCH. : “... the light makes humans glad [and] full of 
pleasure.” In Bth the reason for the joy is not the light itself but the fact that the world systems 
were constantly filled with it. Accordingly, not the light but this fact is said to be beautiful. 

287 (1) For gati in a set with smyti and mati cf. BHSD s.v. gati (3) and gatima(nt). Ch, has 
instead of this triplet 2, Sats. I am not sure if # renders mati or gati. Ch, shows an 
additional, fourth element beside gati (£47): 8). Ch, adds, as a result, that all (living beings) 
became full of pleasure (—)@x=¥) and their klesas vanished (A {§7RY). The first of these two 
additional element appears in TGS, in a more elaborated form at the end of 11B. 


PDF Version: BPPB VI (2002) 

All sentient beings in the hundreds of thousands of world systems who 
had been born in hells, in animal existences, in the world of Yama and [as] 
demons (naraka-tiryag-yoni-yama-lokasuropapatti)*®= immediately [could] 
abandon (cyavate) their birth by virtue of being touched by the light from that 
bodhisattva, and were [re]born among gods and men. 

Those [born as gods and men] by virtue of being touched [by the 
light], immediately became incapable of turning back from supreme and 
perfect awakening (avaivartika anuttarayam samyaksambodhau).” 

In addition, all those incapable of turning back who had been touched 
by the light immediately, when touched by the light, attained intellectual 
receptivity [to the truth that] dharmas [have] no origination (anutpattika- 
dharmaksanti).”° [They] also obtained the efficacious formulas (dharani) 
called ‘Chapter of the Five Hundred Qualities.””?! 

[11D The light makes the world systems delightful] 

“All those hundreds of thousands of world systems which had been touched by 
the light from the body of the bodhisattva within the womb of [his] mother 
came to be established as made of beryl,” laid out in [the form of] a 
chessboard with golden threads, [with] jewel trees coming out of each square, 
[the trees] having blossoms, fruits, fragrances and colors.” 

When the jewel trees were shaken and moved by the wind, such 
pleasant [and] charming [sounds] came up as there are: the sound ‘Buddha,’ 
the sound ‘Dharma,’ the sound ‘religious community’ (samgha) [and] the 
sound ‘bodhisattva,’ along with the sounds ‘[five] powers (bala) of a 
bodhisattva,’ ‘[five spiritual] faculties’ (indriya), ‘[seven] branches of 
awakening’ (bodhyanga), ‘liberation’ (vimoksa), ‘absorption’ (samadhi) and 
‘attainment’ (samdpatti).°* Because of those sounds of the jewel trees, 

(2) Instead of the touch by the light, Ch, continuously speaks of the “seeing” of the light. This is 
also the case in the following passage 11C.6 in Bth (mthong ba). That both possibilities are 
reasonable is documented by a passage in the Kpun dealing with the same context: ... sarve tam 
avabhasam pasyeyuh sprseyuh sarmjaneyuh... (239.1f.). 

788 Ch, adds hungry ghosts (preta; (ff 5%) and speaks of the “[world of] Yamardja” (RFE). 

8° Both Chinese versions mention further the attainment of the five abhijfds (H+). 

2° For anutpattikadharmaksanti cf. BHSD s.v. anutpattika-dharma-ksanti and ksanti. 

| Ror *pancasatagunaparivarta dharanyah (Ch;: paficadasa-...; i+ Tie HEBE ESE) see e.g. 
SP 327.8: ... koftinayutasatasahasraparivartaya dharanyah pratilambho ’bhit /. That not only 
ordinal numbers but, despite the grammatical standard, also cardinal numbers sometimes have pa 
can be proved by the index to the SP (SP;8s.v., e.g., trimSat, pafica-Sata, parica-Sat, etc.). 

2% For ... las grub pa (Skt. -maya) see MVy 6477; yongs su gnas par gyur te for *parisamsthita 
abhiivan (cf. Bth: g.yog par gyur to): the translation kun tu khebs pa for parisamsthita is found in 
the SP (see SP; s.v. parisamsthita); also BWDJ s.v. pari-sam-Stha: X& = Ch. For similar 
descriptions cf. SP 65.8ff., 244.3ff. etc. 

oor (1) In Chy the fact that the trees come out from the squares is not mentioned. Ch, has 
/\{74H4|| instead, which I cannot understand. Cf. the similar passage SP 65.10f.: ... 
vaiduiryamayam suvarnasutrdstapadanibaddham / tesu castapadesu ratnavrksa bhavisyanti ... 
ratnanam puspaphalaih samarpitah //, cf. also BHSD s.v. nibaddha. For astdpada in the meaning 
of “chess-board” see Liiders 1940: 171f. 

(2) Acc. to Jé the meaning of Ijon shing is “a tree from paradise.” However, in translations from 
Skt. both shing and Ijon shing simply render druma, vrksa, etc.: “tree.” For the choice of one of the 
two phrases metri causa see Simonsson 1957: 192.4f. 

24 The enumeration in Chy;: =38. SETH. #8. HO. WK. FASE. AHR. Bth and Ch, 
have bodhi instead of bodhisattva. Tib takes samadhi and samdpatti as two different parts. In the 
Skt. the compound samadhisamapatti may also be understood as a tatpurusa. 


PDF Version: BPPB VI (2002) 

sentient beings in all of the hundreds of thousands of world systems became 
and remained satisfied and joyful. In all the buddha-fields the hells, animal 
existences, the world of Yama, and the world of the demons (asurakdya) 

The bodhisattva within the womb of {his] mother emitted light like the 
disk of the moon for all those sentient beings. Three times a day and three 
times a night [they] raised their joined palms [to pay homage] while [he] was 
still in the womb.”2° 

[11E Why the bodhisattva is named ‘*Sadapramuktarasmi’] 

“Vajramati, when that bodhisattva had been born, had set out for ascetic life 
(abhiniskramana), and [finally] completely awakened to buddhahood 
(bodhim abhisambuddhah),’”’ light continued to be emitted in such a way 
from the body of that bodhisattva. Even after [his] complete awakening, light 
continued to be emitted from the body of that exalted one. Even when the 
exalted one entered into parinirvana, that light [from] his body continued to 
be emitted in the same way.?” [And] even after that tathagata had entered 
parinirvana and {his body] remained [as] relics (dhatu) in a stipa, the light 
[from his] body [still] continued to be emitted. For this reason, Vajramati, 
that exalted one is named the ‘Always Light-Emitting One’ (*Sadapramukta- 
rasmi) by gods and men.” 

[11F The bodhisattva *Anantarasmi questions the buddha *Sadapramukta- 

“Vajramati, under the rule (sdsana) of that exalted one, the tathagata, the 

honorable one and perfectly awakened one, *Sadapramuktarasmi—tight after 

{he had become] completely awakened—there appeared a certain bodhisattva 

295 (1) Ch, states that listening to these sounds, all attained “pleasure in the Dharma, that [their] 
faith became solid [and that they] got forever rid of bad paths (= existences).” (... #E, {F# 
EXE], 7q<MEREGN. ). Cho: ... MARAE. ith. : “[They all] attained pleasure in the Dharma 
[and] delight in the absorptions.” For the common list of bad births cf. e.g. SP 151.9f: apagata- 
niraya-tiryagyoni-yamalokasura-kayam. For asurakaya cf. CPD s.v. asura-kaya: “the asura world 

(2) I have refrained from translating yang in 11D.11, 13, 16 and 11F.1. It is probably rendering 
Skt. api and indicates a change of subject. 

296 Ch, makes very clear that living beings are the ones who pay homage to the bodhisattva, and 
not the bodhisattva himself (... —W2R4- SEAM, HB. ANGERS, Ae. ). The comparison 
with the moon is missing. In Ch; it is the bodhisattva who is said to stay in the womb in the ajijali 
position (directed to the buddhas?) GE{K#FiE, .. GRIER. QA HBOLM BREE, ). 
“Six times” is connected with the emission of the light. The mentioning of the bodhisattva’s birth 
appears in the other versions at the begin of the next section. To divide 24 hours into six periods 
seems to be common Indian usage. See e.g. MPPU,1.416 and MSA XX.56 (ahoratram satkytvah) 
(=MS X.23). 

°7T am not sure how to translate the verb btsas pa. Bth has sems bskyed' pa [' for bskyes] 
instead. In Ch; the phrase is, as said above, incorporated into the last sentence of 11D (75=al4). 
Ch; then continues: ... IFC RIER. : “... and immediately after being born, [he] realized 
perfect awakening.” Ch, corresponds to Tib: ... RABE JU A. SRERYEVE, .... Both Chinese 
versions are lacking a counterpart to abhiniskramana. 

2°8 Bar du could be a translation of yavat, antaSah or antatah (“even, indeed”), which also has 
an equivalent in Ch; (7522). Bth in 11E.5 reads gzhung, which should probably be emended to 
chung. Chung ngu appears in 7B.3 in Bth instead of tha na ... kyang rung ste in Tib and probably 
renders antasah. 


PDF Version: BPPB VI (2002) 

named *Anantarasmi.”” [He was accompanied] by a retinue (parivdra) of a 

thousand [bodhisattvas].° And, Vajramati, that bodhisattva *Anantarasmi 
questioned the exalted one, the tathagata, the honorable one and perfectly 
awakened one, *Sadapramuktarasmi, with reference to this Dharma discourse 
[called] Tathdgatagarbha.” 

[11G *Sadapramuktarasmi teaches the TGS] 

“In order to benefit (anugraha) the bodhisattvas and to win [them] over [to his 
side] (parigraha),°*”' the exalted one, the tathagata, the honorable one and 
perfectly awakened one, *Sadapramuktarasmi thereupon perfectly explained 
this Dharma discourse [called] Tathdgatagarbha for five hundred*” great 
cosmic cycles, remaining*” in the same seat (ekdsane nisannah). And 
because he perfectly explained to the bodhisattvas this Dharma discourse 
[called] Tathdgatagarbha *™ in intelligible (*vijfeya) words [and by 
employing] various means (*(abhi)nirhara) [with regard to the] Dharma, 
explanations (nirukti) and hundreds of thousands of similes, * [the 
bodhisattvas] in all world systems in the ten directions, as many as the 

289 (1) Dang po in 11F.2 (Bth: thog mar) is a formal translation of a Skt. construction with 

prathama as adverb. See PW s.v. prathama (2): “... eben, alsbald, sogleich”; e.g. Vin I 1 (line 2f.): 
... buddho ... viharati ... pathamabhisambuddho. 

(2) The name Anantarasmi (’Od zer mtha’ yas; Bth: Med pa’i ‘od gzer; Chy: #83856; Chz: 
4e-Y¢), “[the one whose] light [is] endless,” appears in a list of names of tathdgatas in Kpun 
158.16 (Q 221a2f: ’Od zer mtha’ yas; Kpuncn12: #£CAA). 

300 (1) Instead of a retinue of a thousand, Bth speaks of only twenty, whereas the Chinese 
versions have twenty Xotis of bodhisattvas. 

(2) Similar constructions are found in e.g. SP 20.15: tena khalu punar Ajita samayena tasya 
bhagavatah sdsane varaprabho nama bodhisattvo ‘bhut or SP 377.9ff. 

3°l (1) Bth and Ch, document that this part (*tesdm bodhisattvandm anugrahdayalpari°) was 
found at the end of the first sentence. It could therefore be combined with the following unit (so 
did Bth; Ch even associates the first sentence with parts of the following unit). 

(2) For anugraha and parigraha Ch, has simply #7: “to protect [them under his shelter and] to 
think [of them](?).” 

3% Ch, reads “for fifty great kalpas” (@F+X#H). 

3° For ‘dug bzhin du in 11G.3 cf. n. 283. 

304 Ch: “Mahayana siitra [called] Tathdgatagarbha” (RUA CH). 

395 As in 5B, also here the lists vary according to the versions, and it is impossible to decide on 
the original form of the Skt. As usual, Ch, is very short: “numberless causes [and] hundreds of 
thousands of similes” (#€BXAk. FA--EF Igy). I do not know what is meant with “causes” (cf. also 
the citations from the SP below: hetu-karana). Bth lacks a counterpart for dharma, which appears 
in both Tib and Ch;. I am not sure if Bth: rtsa ba or brjod pa (for abhildpa?) should be taken as a 
translation of nirukti. Ch: “... with various phrases, discernment concerning the Dharma, 
unhindered readiness in speech (asangapratibhana), and hundreds of thousands of similes” (L1H 
fam. PATE, Se BS. ASE). As in 5B, I suggest that also this list is closely 
connected with the concept of updyakausalya. Similar enumerations can be found in the SP, e.g., 
29.8f: [O adds: nana-|vividhopdyakausalya-jfanadarsana-hetu-karana-nirdesandrambana- 
nirukti-prajfapti (O: -vijiapti for prajfiapti}; 39.11: ndna@-nirukti-nirdesabhilapa-nirdesana (0: -ni- 
darsana}; 41.2f: nadnabhinirhara-[O adds: nana-|nirdesa-vividha-hetu-kdrana-nidarsandramba-na- 
[O adds: nana-\nirukty-updyakausalya. Cf. also TUSN Q 141a5-6 (=T 278: 631a7-9 and T 279: 
278c2—3) where a group of bodhisattvas is said to teach the same law (dharmata) of the Tathagata 
in the following terms: ... thams cad na ‘ang de bzhin gshegs pa’i mnga’ mi 'phrogs pa’i chos nyid 
tshig ‘di dag nyid dang yi ge ‘di dag nyid dang / nges pa’i tshig mngon par bsgrub pa ‘di dag 
nyid kyis don 'di nyid brjod de lhag tshad med par brjod do //. 


PDF Version: BPPB VI (2002) 

atomic-sized dust of ten buddha-fields, understood [this Dharma] easily 

[11H Four bodhisattvas do not attain awakening] 

“Vajramati, in this [connection] the roots of virtue (kusalamiila) [of] all the 
bodhisattvas who heard this Dharma discourse [called] Tathagatagarbha, [or] 
even. (antasah) only the title Tathdgatagarbha, successively (anupurvena) 
came to maturity. Then, in such a way that the marvelous manifestation of the 
excellent qualities [of their buddha-fields] conformed to their [roots of virtue, 
these bodhisattvas] attained supreme and perfect awakening—apart from four 

Vajramati, if you think that then, at that time, the bodhisattva 
*Anantarasmi was somebody other [than yourself, you] should not see it this 
way! Vajramati, you yourself were then, at that time, the bodhisattva 
*Anantarasmi!*°* Who are the four bodhisattvas who under the rule of that 
exalted one *” have not attained supreme and perfect awakening to 
buddhahood up until today? The four are the bodhisattvas Manjustri, 
Mahasthamaprapta, Avalokitesvara and you yourself, Vajramati! 

Vajramati, this Dharma discourse [called] Tathagatagarbha is thus of 
great benefit, since listening [to it leads] immediately to the realization of 
buddha-knowledge (buddhajndna) for bodhisattva-mahdsattvas.”>"° 

Then at that time the Exalted One uttered these verses: 

3% Instead of ten buddha-fields, Ch, has a hundred thousand buddha-fields, in which the 
tathagata’s voice is heard (HBWH+7...). Chz speaks of hundreds of thousands of kotis of 
world systems. Ch, does not mention that the bodhisattvas attain understanding. 

3°7 For the expression yon tan bkod pa de ‘dra ba nyid du (Bth: ‘di ‘dra ba’i yon tan mang pos), 
Skt. gunavyittha, cf. BHSD s.v. vyitha (2). 1 take nyid du to be the rendering of the instrumental 
case of an abstract noun (-tayd or -tvena): “in such a way that....” Ch, reads: “... each attained 
perfect awakening in a different [buddha-]field” (... S#A32E0 MARIE, ...). Ch, does not 
mention the roots of virtue, has no equivalent for gunavyiha or the different buddha-fields, and 
does not add the hearing of only the title of the siitra. Instead, the propagation and preservation of 
the siitra along with the practice according to its teaching are mentioned (... fa] UL, Be, FH, 
aR, am, AORRET. ). 

38 For a similar passage see SP 22.8ff: syat khalu punas te ... kanksd ... / anyah sa tena kalena 
tena samayena Varaprabho nama bodhisattvo mahasattvo ’bhiid ... / na khalu punar evam 
drastavyam | tat kasya hetoh / ... tvam eva ... nama bodhisattvo ’bhut.... 

3 The specification “under the rule of ...” is missing in Ch). “Rule” for sasana could in these 
cases also be replaced by “teaching.” 

310 (1) For this last sentence cf. SP 383.3-6, cited in Zimmermann 1999: 160. Immediately 
before this passage, which emphasizes the benefit of the TGS, it is spoken about the four 
bodhisattvas who could not attain awakening. In that position, this final sentence is thus somehow 
contradictory. Also in the passage in the SP the passage appears at the end of a prose section. 
There, however, it follows upon the description of the attainment of awakening by a large group of 
persons. The inconclusive arrangement of the passage in the TGS could document the strategy of 
the authors to follow closely the main structure of the parallel in the SP without paying sufficient 
attention to differences in content. 

(2) Ch, does not refer to the bodhisattvas. Instead of buddha-knowledge, it mentions the 
awakening to buddhahood ((##3E). 

(3) Only the Tibetan versions stress the immediate result of listening to the siitra, Whereas Tib 
could also be interpreted as “just by listening” (thos pa tsam gyis), Bth uses ma thag tu in a clear 
temporal sense. The Skt. probably had a construction with °mdtrena. 


PDF Version: BPPB VI (2002) 

[11.1] “In the past, endless cosmic cycles ago, the exalted one 
*[Sada]pramuktarasmi appeared. From such*"' light being emitted from 
his body, myriads of [buddha-]fields*’” came to be illuminated.” 

[11.2] “At that time, right after [that] victorious one had attained complete 
awakening, > the bodhisattva *Anantaragmi asked that sugata, 
victorious one and master*" [for this discourse, and the latter then] 
perfectly [and] without pause explained this sutra.” 

[11.3] “All those who happened to hear this siitra personally from the leader, 
under the rule (Sdsana) of that [same] victorious one,’ quickly attained 
noble awakening—apart from four bodhisattvas:” 

[11.4] “Mahasthamaprapta, Avalokitesvara [and], third, the bodhisattva 
Mafnjusri. And you yourself, Vajramati, are the fourth! At that time they 
[all] heard this sutra.” 

[11.5] “The bodhisattva *Anantarasmi, who at that time had questioned the 
victorious one [about the Tathdgatagarbhasitra and who] had been 
tamed by that [same victorious one] *’°—{this] son of a sugata, 
Vajramati, was at that time yourself!”?!’ 

[11.6] “I, too, when [I] formerly practiced the path [of a bodhisattva],*!® 
happened to hear the title of this sitra from the sugata Simhadhvaja. 
[And] having heard [it] with appreciation [I] raised my joined 

31! As already in the prose of 11B (cf. my n. there), the translation of ‘di ‘dra in this passage is 
problematic. Both Chinese versions have 3% instead. 

312 Ch,: “innumerable fields” (#44 +); Chp: “worlds, hundred thousands of kotis [in number]” 

313 T take rig to be a short form for abhisambuddha (cf. Chj2: B%; Bth: chud). Alternatively, rig 
could also be a mistake for reg (Skt.: sprsati) in the common formulation bodhim sprsati (cf. e.g. 
SP, s.v. Vsprs). Thog mar probably renders prathamam (cf. n. 299 above): “right after...” 

314 GBang is here probably short for dbang po (*indra or something similar); Chz: 4:=. 

315 (1) With the variant /as instead of /a (111, fn. 20), the text should be understood as follows: 
“All those who happened to hear this siitra from the teachings of....” That this can hardly be the 
original sense of the pada is attested by Chz: B(K(#HRF ... (“At the time of that buddha....”) and the 
parallel in the prose 11H.10: bcom Idan ‘das de’i bstan pa la (Ch): (RGB). 

(2) Only in Tib the leaders are mentioned (Bth: dran pa, to be emended?). Instead of mngon sum 
(*pratyaksa) Bth has mngon du (ba); Chz: (EHS GR) seems to render simply *sakasat, antikat or 
something similar; Ch,: #(44 35444 (“Those [who] encounter the victorious one [and hear]....”). 

316 T take dbang byas to be a short form for dbang du byas pa rendering vasikyta or 
adhipateya/°patya (cf. Bth: dbang po). In 11F *Anantarasmi is said to appear under the rule 
(Sasana) of the tathagata “Sadapramuktarasmi. The expression with adhipatya could, with other 
words, underline this superior position of that tathagata. Takasaki (1981: 39) offers another 
interpretation of dbang byas: BAISICRALUT. £0 Ge) BECLERMOF ..: 
“.. the son of a victorious one, [who] asked the sugata and [by this way] empowered that 
(sitra),....” However, Takasaki’s interpretation does not correspond to Ch, which with $3— refers 
to the bodhisattva. Ch, is without any equivalent. 

317 Chy: “[You,] Vajramati, being him, the foremost [among the bodhisattvas], the son [with] 
supernatural powers, were at that time called ‘*Anantarasmi,’ who had already questioned for this 
siitra!” Ch, consists only of three padas, omitting the pada which states that the bodhisattva had 
questioned with reference to the siitra. 

318 SPyad pa spyod pa may refer to the practice of a bodhisattva, directed to supreme 
awakening, as is the case in SP 382.6: bodhisattvacaryam carantu or SP 383.13 (= IXX.3d): 
yiiyam hi caryam carathdgrabodhaye // ; cf. also BHSD s.v. cari. Ch,: 334: “to search the path”; 
Chz: 47847: “to practice the highest path.” 

° Ch, mentions the siitra itself (and not its title), which the narrator had attained from the 
buddha Simnhadhvaja. As in its prose (11H), Ch, then states that it was taught and practiced 


PDF Version: BPPB VI (2002) 

[11.7] “By those well-done deeds*”° I quickly attained noble awakening. 
Therefore wise bodhisattvas should always learn this excellent 

[12A Ananda questions the Buddha about the length of time till 

“Vajramati, [when] sons and daughters of good family who are restricted by 
obstacles [caused by their] deeds (karmdvarana) listen to this Dharma 
discourse [called] Tathagatagarbha, [and when they] show (uddiSati), recite 
(svadhydyati) or teach [it, then]—with regard to listening to this Dharma 
discourse, showing [it], reciting [it], explaining [it] and copying [it]—they 
will all, seeing the Dharma before their eyes (*pratyaksa), easily 
(alpakrcchra) become purified [from] the obstacles [caused by their] 

Then the venerable Ananda asked the Exalted One the following: 

“Exalted One, [as for] those sons and daughters of good family who are not 
restricted by obstacles [caused by their] deeds and [who] apply [themselves] 
(*prayujyate) to this Dharma discourse, from how many buddhas, exalted 
ones, [do they preserve] expositions of the Dharma as [persons of] great 
learning (*bdhusrutyena) so that [they] become perfected (niryata)?”>™ 

according to the siitra. The verse 11.6 differs from the correspondent prose, as the narrator, namely 
the Buddha, speaks about himself and not about the bodhisattvas. 

320 Chy: BE4 and Chz: 24813 are based on the term kusala-mila. KuSala-miila appears also in 
the parallel prose 11H of Tib, Bth and Ch. 

321 As in the preceding verse, Ch, mentions the necessity to teach the siitra in addition to its 
preservation ($#). The bodhisattvas are not characterized as “wise.” Ch2, however, speaks only of 
the wise (4°) and does not mention the bodhisattvas. In order to attain an even number of padas, 
Ch, had to fit c and d into one pada, which probably led to the omission of “bodhisattvas.” 

322 Sections 12A and 12B are missing in Ch,. 

323 (1) The basic structure of this passage remains obscure in the Tibetan. I therefore follow 
Ch», which clearly marks the second enumeration of propagandistic measures as the reason for the 
purification from the obstacles: FA fa Lee2.... (For a detailed comparison of the passage 
transmitted in the Tabo manuscript (A) and the other representatives of Tib cf. my section C 2.6.) 
Ch seems to be based on udgrhnati (424) instead of uddisati. However, combinations of forms 
of ud-dis and svadhya are found frequently in e.g. the SP (cf. SP; s.v. uddesa-svadhyaya etc.); for 
the translation “## for uddesa cf. YBhu, s.v. FF (for uddesa) and PRA (for uddesa- 
svadhyaya-desana---Sabda). The translations of uddisati as lung nod (“to receive instructions”) 
and 34% (“to receive and keep”) stress the receipt of the teaching. This would fit very well into 
the sequence ‘receiving (uddisati) ~ recitation — teaching’ in the text above. In CPD s.v. uddisati 
the following meaning is mentioned: “... 2.b esp. to point at (somebody as one’s teacher), 
acknowledge somebody as spiritual master...” Should we thus translate uddisati as “to 
accept/adopt [the siitra as an authority]”? 

(2) The position of de dag kyang (12A.5) between chos and mngon sum du ’gyur is disturbing. 
Bth has no equivalent for it. 

(3) Pratyaksa in this context could mean the direct experience or vision of the tathagata in all 
living beings; Ch): #RAFRAT. 

(4) Whilst Tib is speaking of the purification (byang ba) from the obstacles, Bth (zad pa) and also 
Ch, ($8) have their destruction. 

324 (1) The negation in 12A.8 (bsgribs par mi ‘gyur) does not make sense. In Ch; a negation is 
not found. Bth has an insertion (— ... <-) which renders a negation reasonable and could well be 
an original element: — las kyi bsgrib pa bsgribs pas : de bzhin gshegs pa’i snying po chos gzhung 
‘di nyan pa dang : ston par dang : khadon du bgyis pas : < las bsgrib pas : bsgrib pa ma mchis 
par gyur to :: “... those ... restricted by obstacles ... become unrestricted by obstacles ... through 
listening to this....” Bth then continues: “The one who becomes perfected (‘byung ba) through this 


PDF Version: BPPB VI (2002) 

The Exalted One said: 
“Ananda, there are sons and daughters of good family as well who are 
[already] perfected on account of having preserved expositions of the Dharma 
from a hundred buddhas.”*”> 

[12B Perfection through preserving expositions of the Dharma from myriads 
of buddhas] 

“Ananda, [but] there are also sons and daughters of good family who are [only] 
perfected on account of having preserved expositions of the Dharma from two 
hundred buddhas, three hundred, four hundred, five hundred, one thousand, 
two thousand, three thousand, four thousand, five thousand, six thousand, 
seven thousand, eight thousand, nine thousand, ten thousand or a hundred 
thousand buddhas, [or] even myriads of buddhas.°”° 

Ananda, a bodhisattva who preserves this Dharma discourse, recites 
[it], perfectly teaches [it] in detail also to others and preserves [it] as a book?”’ 
should bring forth this thought: ‘I [wish to] attain supreme and perfect 
awakening already now!’ He is worthy of the homage and veneration of the 
world with [its] gods, humans and demons (asura), as I [am] now.”?”® 

[12C Homage to the one who holds the 7GS in his hands] 

Then at that time the Exalted One uttered these verses:>”” 

exposition of the Dharma will definitely (*niyatam) be a [person] of great learning among the 
buddhas, exalted ones and teach the Dharma.” A counterpart for *niyatam does not appear in the 
other versions. It could be a mistake for the particle kiyat since there is no question expressed in 
Bth. This part of the text in Bth can hardly represent the Skt. original, as the following answer by 
the Buddha would remain without question. 

(2) It seems that Ananda, who usually in Mahayana sitras is “only” a saiksa, thinks mainly of 
himself when inquiring about the number of buddhas from whom one should collect Dharma 
expositions, since the term bdhusrutya is usually connected with him (see MPPU, I 22). Besides, 
he is often entrusted with the preservation of an exposition at the end of it. On the term bahusrutya 
expressing the “passive aspect” of the Mahayana initiate see Harrison 1978b: xx-xxi. 

(3) I take nges par ‘byung ba to be a translation for niryata (cf. BHSD s.v.). Bth has ’byung ba in 
all instances. Here and in all following passages, Ch2 does not have any equivalent for it. 

(4) The passage in Ch): “Exalted One, those sons and daughters of good [family] who are 
restricted by ..., [under] how many buddhas, exalted ones can they increasingly preserve (A[+#) 
expositions of the Dharma, attain great learning (44fx]), and become bound (4H; *prayoga) to 
this Dharma discourse?” 

325 (1) Kun tu bzung ba in Tib (Bth: ’chang ba; emended for ‘chad pa) could be a translation of 
samgraha (“keeping, bringing together, collecting”), samdharana (“holding together, 
maintaining”) or Gdhdrana (“bearing, holding”). 

(2) Ch: “Ananda, [there are] sons and daughters of good [family] who can increasingly preserve 
expositions of the Dharma under a hundred buddhas.,....” 

326 The number of years differs in the versions. Both Ch, and Bth speak of 200,000 instead of 
100,000 and, again, insert between the numbers the standard phrase of the attaiment of perfection 
due to preserving expositions (so Brh; in Ch, simply AIFF). In Ch; the end of the passage is 
different from the Tibetan and also from the standard pattern in Ch before: “... or there are 
myriads of buddhas under [whom they] received expositions of the Dharma, [to which they] 
listened [and which they] preserved.” 

227 The sequence in Ch;: “Ananda, if a bodhisattva obtains this Dharma [exposition called] 
Tathagatagarbha, writes [it] down, recites [it], preserves [it], reflects about its meaning [and] 

328 T acking the comparison “as I am now,” Ch, ends the section in the following way: “After 
the Buddha had said this, [the onlookers] were pleased, [shouting] ‘Bravo!’” 

329 The introduction to the verses in Ch; differs from the formulas before, by adding {f. 


PDF Version: BPPB VI (2002) 

[12.1] “When a bodhisattva has heard this stitra, [he] thinks: 
‘I [wish to] attain the noble awakening!’ 
He in whose hands this stitra is found is worthy of the homage of the 
world, as I [am now].”22° 

[12.2] “[Being himself] a protector of the world (Jokandtha) [and] training 
[sentient beings], he is worthy of the praise (prasamsya) [of] the leaders 
(nayaka) [and] trainers (vindyaka). Thus he in whose hands this sutra is 
found should be called ‘king of the Dharma.””??! 

[12.3] “He in whose hands this siitra is found is worthy of being looked upon 
[as] the best of men (purusarsabha), [as] bearer of the lamp of the 
Dharma (*dharmolkadharin) [and] like the full moon. Like a protector 
of the world, [he is] a foundation [worthy] of being paid homage.”?*” 

[12D Delight and praise of the onlookers] 
After the Exalted One had spoken thus, the bodhisattva**? Vajramati, the entire 
multitude of bodhisattvas, the great sravakas, the four assemblies and the world 
with [its] gods, humans, demons and celestial musicians were delighted, and 
praised [what] the Exalted One had said.>** 

[Here] ends the Holy Mahayana sutra called Tathagatagarbha. 

[13 Tibetan colophon] 
The Indian master Sakyaprabha and the Venerable Great Reviser and Translator 
Ye shes sde have executed [this] translation and revised and established [it] 

330 (1) In Ch, the comparison with the Buddha himself is missing. 

(2) The first verse half in Ch, (continuing the verses 11.1 till 11.7): “Having heard [this siitra] and 
practicing as [it] was taught, [the bodhisattvas] become buddhas just as I [am] now.” 

(3) For a repetitive construction similar to the formulation “He in whose hands this sitra is 
found,....” (12.1, 12.2, 12.3) in a relative clause cf. SP 292.11ff. @XX.5, 6, 9, 10, 11) quoted in 
Zimmermann 1999: n. 44. There the act of preservation of the siitra is expressed with the verb 

331 The versions differ considerably. Tib: rnam par in pada a could result from a revised rnams 
as it is still found in Bth. Ch: “The one in whose hands the stra comes: the buddhas, exalted ones 
[and] great leaders praise that human as the highest among humans and call [him] ‘most eminent 
king of the Dharma.’” Ch,: “The one who attains this siitra is called ‘king of the Dharma of the 
buddhas.’ Being then a protector of the world, [he] is praised by the buddhas.” 

332 (1) Instead of mi yi khyu mchog for *purusarsabha (Chy: ttt) Bth has bsrung (deriving 
from the verb raksati?) Ch, reads {H[JHR (*lokaksan) for the whole expression in pada b of Tib. 

(2) gNas in pada d has no equivalent in any of the other three versions and can hardly be an 
original element. 

(3) Instead of “worthy of being looked upon like the full moon” Ch, has “shining like the full 
moon” (HARE iia A). Ch, repeats the same statement as in the preceding verse instead, stating that 
the person is called a dharmardjan: 2 \ iE. 

333 Ch, adds mahasattva. 

334 The sentence appears in slightly modified forms at the end of many Mahayana sitras (cf. 
e.g. SP 487.1-5; Kpun 420.46; Suv 250.8ff; Sukh 66.23ff.). The reconstructed Skt. of the TGS 
may be: *idam avocad bhagavan / attamana Vajramatir bodhisattvah sa ca sarvavan 
bodhisattvaganas te ca mahasrdvakas catasras ca parsadah sadevamanusasuragandharvas ca 
loko bhagavato bhasitam abhyanandann iti //. Chz does not mention the four assemblies and the 
gandharvas. Instead it has $$ at the end of the enumeration. In Ch, the mahasravakas remain 

335 The colophons are discussed in section C 3.3. 


PDF Version: BPPB VI (2002) 

PDF Version: BPPB VI (2002) 

Part II 

Critical and Diplomatic Editions 
of the 


PDF Version: BPPB VI (2002) 

C. The Textual Materials 

The appearance in the last years of several studies dealing with Kanjur 
manuscripts and prints makes it unnecessary in the following to repeat what 
already has been stated about the history of the various editions. I will concentrate 
my remarks on the legibility, appearance and orthography of the texts. For studies 
dealing with the history of the Kanjurs, and for particular case studies, see the 
publications of Braarvig, de Rossi Filibeck, Eimer, Hahn 1988, Harrison, Imaeda, 
Schneider, Schoening, Silk, Skilling and Zimmermann 1998 cited in my 

1 Information on the Tibetan Manuscripts and Xylographic Editions 

* A— The Tabo Manuscript Fragments 

The fragments of the TGS among the Tabo manuscripts cover about 40 percent 
of the whole siitra. The fragments are very likely to be the oldest among the 
Tibetan materials utilized in this study. Steinkellner states that “a considerable 
portion” of the 35,500 folios assembled were written in the eleventh and twelfth 
centuries.’ The manuscript of the TGS is carefully written in dbu can hand, with 
only very few mistakes, eleven lines to a folio, and is nearly free of damage of 
the type affecting legibility. Every folio has two string holes near the center of 
its left and right halves, each with a diameter of about five letters. 

The orthography exhibits particularities well known from Dunhuang and 
other Tabo manuscripts: 

— the ya btags in all words beginning with m- followed by the vowel i or e 
(e.g. myed, myi etc.), 

— the da drag (e.g. bskald, mkhyend, gyurd, stond, bstand, ‘drend, stsald, 
zhend, shard etc.), 

— the mtha’ rten ’a(mdo’, ’dra’, dpe’), 

— occasionally a reversed gi gu, 

— the spelling /as stsogs pa for la sogs pa, 

— in some cases cen instead of chen, 

— ba(r) instead of the regular pa(r) when the preceding syllable ends with -n: 
mngon bar, Idan ba, brtson ba, yin ba, but never in rkun po, mgon po, snyan 
pa, ’thon pa, ’dron po, chen po, Ijon pa or ’brog dgon pa, 

— ba instead of the regular pa when the preceding syllable ends with -m: bcom 
ba, zhim ba, gsum ba but never in dam pa or rnam pa(r)- 

| Steinkellner 2000: 319. Further, see the contributions in Scherrer-Schaub and Steinkellner 
1999; Steinkellner 1994. 

? These and other archaic features are attested for Dunhuang materials; see e.g. Terjék 1969, 
Harrison 1992a: xxi, n. 43 and Schoening 199Sa: 187f. As Schoening has shown in his study (esp. 
147ff., 188), there is no homogeneity in the Dunhuang materials regarding these characteristics. 
We therefore have to be careful not to hastily conclude that ail features appearing in those 
materials are archaic. The two questions why and when in old materials the regular pa after final - 
n becomes ba, and also why we often find ba after final -m, definitely deserve further 
investigation. That there is nothing like a standard for spellings among these old manuscripts and 
that even in one and the same text we find different spellings is attested by A, which has bskald pa 


PDF Version: BPPB VI (2002) 

The accuracy of the scribe of A is further confirmed by the fact that at the end of 
a sentence immediately before the beginning of a new verse he has employed a 
triple shad instead of the usual nyis shad. Tsheg is always set before the shad. 
The gap between the shad and the following letter is of about one letter. The 
copyist made sporadic use of the anusvara-like abbreviation for m in the words 
thams, rnam, rnams, sems and ’am. Further worth mentioning is the contraction 
of the cluster st- (#) to the horizontal ligature ™, of (§) spy- to &, and of rts- (8) 

to *6 in the handwriting whenever these combinations appears. 

B — The Berlin Manuscript Kanjur 

The manuscript Kanjur kept in the Staatsbibliothek in Berlin dates from 1680.° It 
is said to be a copy from a lithograph of the Yongle 7 # (1410) or Wanli A/F 
edition (1606), and thus predates the widely used “Otani reprint,” which is a 
photolithographical reprint of the Kangxi S52 edition of 1717-20.* 

The manuscript is beautifully written in dbu can and is in excellent 
condition. Each folio has eight lines. From the middle of the fifth line on folio 
327b till the folio end the script is larger, and instead of rkyang shad the 
obviously different (and seemingly inexperienced) copyist repeatedly used the 
nyis shad. 

As has also been confirmed by other scholars, the number of mistakes is 
fairly high. We find 
— frequent omissions, 

— aberrationes oculi, 

— confusion concerning vowels, 

— orthographic confusion between -ng- and -d- (though -ng- and -d- are 
clearly distinguishable in the script), and 

— a large number of obvious misreadings of letters (e.g. Jig for jim, nyid for 
nyil, stod for ston, thams for thabs, dam for dri ma, mdod for mdog, ’dun 
for ‘dren, dpan for dper, ‘bum for dbul, shes for shing etc.). 

The last two categories in particular suggest that the text was not read aloud 
when written down by the copyist but just copied letter by letter, thus allowing 
for many senseless variants. 

There are just a few instances of later corrections to the writing. At the end 
of a line we frequently find words with subscribed letters. The tsheg is never set 
before shad. The gap between the two shads of a nyis shad and between rkyang 
shad and the following word is of about one letter. 

in 11A and 111 but dskal ba in 11G (see also Harrison 1992a: xxi, where for the DKP the forms 
snyan and snyand are confirmed). For paleographic, orthographic and phonetic peculiarities in the 
Tabo manuscripts see also Steinkellner 1994: 124f., de Rossi Filibeck 1994: 139, Tauscher 1994: 
175ff., Eimer 1989/91, Scherrer-Schaub 1999b, and the contributions in Scherrer-Schaub and 
Steinkellner 1999; for other materials with archaic features see Taube 1980 with an extensive 

3 See Haarh 1954: 540. 

4 tmaeda (1977: 27, n. 18) states that it is more likely that the Berlin manuscript is based on a 
Wanli lithograph than on a lithograph of the Yongle edition. He does not provide us with any 
reason for his assertion. Eimer mentions both possibilities (2000: 36, n. 24). 


PDF Version: BPPB VI (2002) 

+ Bth — The Newark Manuscript Kanjur from Bathang” 

Bth is the only known representative of a separate, paracanonical translation of 
the TGS. Judging from its terminology and syntax, it must have been executed 
before translations became more standardized following the compilation of 
compendiums like the MVy and the sGra sbyor bam po gnyis pa, that is, before 
the early ninth century. The assumption of the sixteenth century as the one in 
which Bth was copied is based on art-historical considerations’ and does not 
conflict with the orthographical facts, given that the employed orthography 
exhibits very few archaic characteristics (ci for ji, stsogs, in some rare cases the 
use of the mtha’ rten 'a (dpe’ for dpe), ’og du, yan cad). 

The folios were probably kept together by two strings up the middle of the 
left and right halves of the book resulting in black stripes extending from the top 
to the bottom on the photos of the folios, and rendering the affected letters 
unreadable. With the exception of two eight-lined pages (one recto and one 
verso), both sides of all folios contain nine lines. The handwriting (dbu can) is 
sloppy and irregular and has some lacunas. These lacunas were partly caused by 
the erasure of letters. In most cases, however, they are found where the bottom 
parts of letters from the line above consume much space, so that leaving the 
position in the line below empty made for greater legibility.’ 

A tsheg before shad is found in several cases. The gap after shad and 
“colon” (see below) is irregular but tends to correspond to the width of one or 
two letters. Between or after the nyis shad, however, there is almost no gap. The 
text is rich in contractions:® 

skyabs(s)u, skyes(s)o, khams(s)u, grangs(s)u, ‘gyur(r)o, chags(s)u, rjes(s)u, 
btags(s)o, stag(g)o, bdag(g)is, ‘das(s)u, nam(m)kha’, gnas(s)o, rnams(s)u, 
phyogs(s)u, dbus(s)u, tshigs(s)u, rdzogs(s)o, bzhugs(s)o, yongs(s)u, lags(s)o, 
Shun(n)i, bshad(d)o, gsungs(s)o. 

The use of the anusvara-like abbreviation for -m- is extremely common, and 
subscripts also appear. Well-established terms are usually written without a 
separating tsheg between the syllables: skyes-bu, kha-tog, kha-lo, khrig-khrag, 
ngo-bo, ngo-mtshar, gti-mug, sti-stang, tha-dad, the-tsom, rdo-rje, rnam-par, 
pha-rol, phyi-rol, phra-mo, phrag-stong, byang-chub, blo-gros, ma-rig, me-tog, 
rdzu-'phrul, zhe-sdang, ’od-zer, yang-dag, ye-shes, re-re, sems-dpa’. 
Abbreviations are found, such as thad (for thams cad), sbra-rtsi (for 
sbrang rtsi), gshyes (for gshegs), saryas (sic!; for sangs rgyas) and semdpa’ (for 
sems dpa’). The text contains repetitions and omissions of passages, partly 
corrected later by adding the missing words beneath or above the line. 

> I have elsewhere described the characteristics of the TGS manuscript from this Kanjur in 
detail (Zimmermann 1998); see especially pp. 38-42. 

° Olson 1971: 114. 

7 In some cases a lacuna is found in the line below words first omitted by the scribe and later 
added between the lines. This documents that the newly completed line was at least partially 
checked before starting with the next line. 

® Only those contractions are cited for which obvious reasons, such as shortage of space at the 
end of the line, cannot be found. The letter supplied in brackets is omitted in the contraction. 

° Some less well-established combinations also appear: bsten-te, byas-te (with superscribed s-), 
mi-za and mig-gis. The dash between the syllables merely serves to make the words more readable. 

The manuscript has e.g. SAA (mig-gis). 


PDF Version: BPPB VI (2002) 

Misspellings are not very common but clearly exceed the number found in Q or 
S. It is to be noted that the adding or omitting of final -s in the use of the particle 
kyi(s) seems unreflected. Other irregularities are: 

— the use of the particles Ayi and kyis after final -n, -m, -r and -/ (the proportion 
of the particle Ayi(s) (against the regular gyi(s)) after the finals: -n 83% (against 
gyi(s) with 17%) ; -m 71%; -r 62%; -1 77%), 

— confusion between final -/ and -s or -r: byol for byos; dbus for dbul; rus for rul; 
rgyal for rgyar; gsel for gser, 

— confusion between pa and ba, 

— omitting/adding of subscribed r: skod for skrod; bskad for bskrad; phrag for 
pags; smrad for smad, 

— various spellings for the same word: khung and phung; grags and drags (for 
grangs); dkyil mo grung and dkyil mo drung (for skyil mo krung); ci tse and ci 
rtse; the tsom and the rtsom; sman, smod, smon and smrad (for smad); ril and 
rul, and finally 

— variant spellings possibly caused by same or similar pronunciation: rgya rgod 
for bya rgod; rgyun po for rkun po; bsgrub for sbubs; ngam for ngan; gtong for 
mthong; biten for bstan; spyod for bskyod; 'phags for pags; dbus for dus; shin 
for shing. 

Very common throughout the text is the employment of two vertically aligned 
dots resembling a colon where in other Kanjurs a shad is used. The rkyang shad 
appears only twelve times throughout the whole text. The nyis shad, too, appears 
less often than in other Kanjurs.’” 

Instead of pa’i at the end of the line we sometimes find pa with an ‘a 
chung beneath. Above the pa there is a mark resembling a parallel double ’greng 
bu. The same feature is attested for Dunhuang materials (see Schoening 1995b: 

¢ Bu — The Citation in Bu ston Rin chen grub’s De bzhin gshegs pa’i snying po 
gsal zhing mdzes par byed pa’irgyan 

Two short parts of the Tibetan Canonical translation of the TGS are quoted in the 
De bzhin gshegs pa’i snying po gsal zhing mdzes par byed pa’i rgyan, a work by 
the famous Tibetan scholar Bu ston Rin chen grub (1290-1364) finished in 
1359.'' The text I utilized is the reproduction of a lithograph from blocks 
prepared in the years 1917-19 in Lhasa. The writing is in dbu can; the lithograph 
is in excellent physical condition and has seven lines per folio. 

¢ D— The Derge Kanjur (Nyingma Edition) 

The Nyingma reprint dates from the beginning of the 1980s. It is said to be a 
reproduction of the so-called Karmapa edition.’” In the case of the TGS, the 

10 The statistics testify the usage of the “colon” for 84% of all punctuation marks, the simple 
shad for only one percent, and the nyis shad for about 15%. The usage of the nyis shad is low 
compared with its 39% rate in the TGS of the London Kanjur. The simple shad in Bth is in every 
second case preceded by na or nas. 

"| See Seyfort Ruegg 1973: 1. 

? See Skilling 1994: xxxviii and Silk 1994: 63. 


PDF Version: BPPB VI (2002) 

reprinted folios are identical to the folios of the Taipei edition,” with the only 
differences being that in the Taipei edition the Tibetan pagination at the left 
border of the folio is not framed and that at the right border of the folio no 
additional frame (in the Nyingma reprint containing a number added by the 
editors) is found. In general, the Taipei edition is more legible than D’s weak 
print. It has been pointed out by Silk (1994: 63f.) and Skilling (1994: xxxviii) 
that the Nyingma edition cannot be called a pure reproduction of the Karmapa 
edition, inasmuch as it has been retouched in several passages and damaged 
leaves have been replaced with the aid of photographs made from the Derge 
edition owned by Harvard University (Harvard-Yenching Library) and by the 
United States Library of Congress in Washington, D.C. 

The Karmapa edition (1976-79) itself is a reproduction of the Derge 
Kanjur deposited in Rumtek (Sikkim) which, according to Eimer (1980), differs 
from the copies held in the University Library in Cambridge and in the Academy 
of Sciences of the Czech Republic in Prague only in terms of its superior 
graphical quality but not in its wording. The Derge Kanjur dates back to 1733 
(Imaeda 1981: 229). 

D is very readable with seven lines per folio. It has an extremely low 
number of single variants and no abbreviations or subscripts. The syllables pa 
and ba are hard to differentiate. Within the nyis shad and between a shad and the 
following syllable there is a gap of about two letters. A tsheg is placed before 
shad only after final -ng. 

e J— The ’Jang sa tham or Lithang Kanjur 

The date of the carving of this Kanjur has not been established with certainty. 
Imaeda (1982b: 12; 14), based on the Chinese colophon, suggests the period 
from 1608 to 1621 as the time of its production, while Samten (1987: 17f.) 
argues in favor of the years 1609 till 1614. 

The quality of the microfilm I obtained is very poor. The letters are 
sometimes hardly recognizable against the dark background; in other cases both 
background and letters are too light. In general, the letters are not very clear, pa 
and ba along with nga and da being for the most part not differentiable. In many 
cases it is also impossible to decide where the tsheg is placed, and hence where 
the syllables should be split. 

The folios have eight lines. The gap within the nyis shad or between a 
rkyang shad and the following syllable is two to three letters. A tsheg before a 
Shad is extremely exceptional. The Lithang TGS is very carefully carved and 
contains just a few mistakes, mostly the omission of vowel marks. Gaps are 
frequent; the reversed gi gu appears twice throughout the text. A particularity of 
J is the use of the particle du instead of tu after final -n. In 90% of all 
appearances of kun tu J reads kun du, in 93% of all cases J has shin du for shin 
tu. The form rol du, too, appears. 

The Cone Kanjur from 1721-1731 is based on this Lithang Kanjur despite 
some differences in the order of the sections. Eimer therefore calls the Cone 
print “a true copy of the Lithang Kanjur” (1992: 181). I can confirm his 

3 In The Tibetan Tripitaka, Taipei Edition, ed. A. W. Barber, SMC Publishing: Taipei, 1991, 
vol. 14, no. 258, mDo sde, Za 245a2—259b4. 


PDF Version: BPPB VI (2002) 

statement based on a comparison of some parts of the TGS in the Cone Kanjur 
with J.'* However, most of the mistakes’* in J have been emended by the carvers 
of Cone, the lacunas in J are of course not longer found in Cone, and the 
reversed gi gus have been standardized. We still find kun du, shin du and rol du, 
with some exceptions of tu where J mostly also reads tu (in some cases Cone has 
du where J reads tu!). Of particular interest is the variant dkrung in J against the 
usual krung, which appears altogether six times. In all these cases Cone reads 
krung with a lacuna of one letter before krung, resulting most probably from a 
first copied but later erased d- (still found in J), In the only case where J also has 
krung, no lacuna is found in Cone. We face a similar situation in SA, endnote e, 
where only J reads dbyigs instead of dbyig. In Cone we find dbyig followed by a 
lacuna of one letter. 

¢ L— The Shel dkar Manuscript Kanjur (London) 

The London copy of the Shel dkar manuscript Kanjur was completed in the year 
1712.'° The script is of excellent legibility with eight lines per folio.’ There is a 
string hole in the center of the left and right halves, each with a diameter of 
about three letters. A tsheg appears before a shad only after a final -ng; there is 
virtually no gap after a shad or within a nyis shad. 

Of the collated materials, ZL contains the highest number of contractions 
(yongs(s)u, etc.); frequent, too, are subscripts and abbreviations, which occur for 
the most part at the end of a line and before string holes, but are not restricted to 
these positions: becom, thamd for thams cad, rnamr for rnam par, tshot for 
tshogs, bzhut for bzhugs, g.yot for g.yogs, longspyod (*°§5'), gshye for gshegs, 
semn for sems can. Besides these particularities, there are single variants, 
restricted to a number of omissions of mostly single syllables and, in one case, a 
dittography consisting of several words. Lacunas resulting from later erasures of 
letters appear sporadically. 

¢ N— The Narthang Kanjur 

According to Tucci and Petech,'® the blocks of this Kanjur were carved from 
1730 to 1732. The photocopies I utilized are of poor quality. The letters in some 

‘4 My thanks are due Professor K. Mimaki (University of Kyoto) who provided me with the 
relevant photocopies of the TGS in the Cone print—which he acquired from the Toy6 Bunko in 
Tokyo: mDo mang, Za 274a8-289a4. 

5 Yn OE J omits ‘i gnas as part of the name of the bodhisattva Mi g.yo ba’i gnas rnam par gnon 
(*Acalapadavikramin). In C, however, the name appears as Mi g.yo ba gnas rnam par gnon 
(275a4). I am not sure what conclusion should be drawn from this fact, considering the likely 
source for the emendation in C. Skilling (1994: xxxvi) states that in the case of Mahasutra 3 the 
editors of C consulted an edition belonging to the Peking lineage. In our case, however, B and Q 
read ba’i gnas and not ba gnas. 

16 On the date see Pagel and Gaffney 1996: x, and the article by Jampa Samten and Peter 
Skilling contained in Pagel and Gaffney 1996 (pp. 1-11). 

7 All three colophons of the Kanjur state that the calligraphers came from “sNye mo, birthplace 
of Thon mi Sambho fa, a place in Central Tibet renowned for its tradition in calligraphy” (Samten 
and Skilling (p. 9) in Pagel and Gaffney 1996). 

'8 The statements by Tucci 1949: 479f. and Petech 1972: 160f. are based on the dkar chag of 
the Kanjur. 


PDF Version: BPPB VI (2002) 

passages look like meaningless black spots. In general the appearance of the 
script is blurred. The difference between pa and ba is sometimes impossible to 
make out. Similarly difficult in many instances where the tsheg is hardly 
recognizable is how to determine syllables. The folios have only seven lines. A 
tsheg before a shad is only found after syllables ending in -ng; the gap within a 
nyis shad or between a rkyang shad and the following syllable is normally about 
one letter. 

One particularity of N is the contraction of yang to ‘ang following a 
syllable ending in a vowel (e.g. bu’ang). This feature is found throughout the 
whole text. There are not many misspellings; most of them are due to a missing 
’greng bu or na ro or to confusion between these vowels. Abbreviations by 
means of subscripts are rare. Abbreviations in general are found at the end of a 
line. Interesting among them are gshet pi yais for gshegs pa’i ye shes and yton 
for yon tan.|° 

P;, P2, P; —- The Phug brag Manuscript Kanjur 

This manuscript Kanjur was copied between 1696 and 1706 at the Phug brag 
monastery in Western Tibet.” The TGS is not the only text which is contained 
several times in the Phug brag Kanjur (other texts appear even in different 
translations). The three manuscripts of the 7GS found in the Phug brag Kanjur 
share many features. They all have eight lines per folio”’ and are written in a 
clear dbu can hand. This and the fact that the gap within a nyis shad or a rkyang 
shad and the following syllable is of about two letters (in P; closer to three 
letters) render the manuscripts easy to read. The script in P; is bigger than in the 
two other versions, and there is more space between its letters. Unfortunately, 
some parts of P; were overexposed when photographed and are too faint to be 
made out. 

The pagination written in the left margin of the leaves of P23 is 
interesting: instead of nyis brgya the margin has just two crosses (in P; it has 
three crosses). P2 further bears the word mngon at the top of the left margin. 

The tsheg before shad appears regularly after syllables ending in -ng. P2 is 
particular in the way it handles the nyis shad between the verse padas: all other 
manuscripts and prints strictly use the nyis shad to separate the padas, but P2 
prefers the rkyang shad, employing the nyis shad regularly only when the pada 
has a sentence-closing particle (... go etc.) at its end (in one case even after ... 

The manuscripts of the TGS' in the Phug brag Kanjur can be said along 
with B and T to contain the highest number of mistakes. Among the three 
versions, P; is the one with the most single variants (427), followed by P2 (329) 
and P; (239), Frequent errors include the omission of single words or whole 
phrases and the adding or dropping of final -s in syllables like Ayi(s), stag(s), 
byam(s) and bzang(s). 

'? The abbreviations yais for ye shes and yton for yon tan (33) are quoted in Bacot 1912: 600 

and 604. 
? Bimer 1993: v. 
2! Only the folio (verso) where:the stra starts and the following recto of P; have seven lines. 


PDF Version: BPPB VI (2002) 

P; adds the mtha’ rten ’a in five cases: ‘dra’ (4 times), g.yo’. Another 
archaic element is the reading stsogs for sogs in P;2 (four times each). Some 
further peculiarities of spelling in P;: the inconsistent use of ‘a sngon jug where 
the majority of versions have m- (’gar, ’thil, ’thon, ‘dab, ’dun), the omission of 
prefixed g- and b- ((b)skyod, (g)ci, (g)cig, (b)snyen, (b)stan, (b)rlabs, (b)shig, 
(b)sgrubs), and du instead of tu after kun and syllables ending in -g. 

Furthermore, P; clearly shows a number of redactional variants (see 
below) and, in comparison with P;2, is more orthographically standardized, as 
the section on archaic features documents, 

* QO — The Peking Kanjur (Otani Reprint) . 

The Otani reprint is a photolithographical reproduction of the third Kangxi EFEE 
edition from 1717-1720. As is well known, missing parts have been supplied 
from the Qianlong #7 edition of 1737, which derives from the same blocks. 
One lithograph of this Qianlong edition is kept in the Bibliothéque Nationale in 

Q has eight lines per folio with some passages of rather unclear 
appearance, Pa and ba along with da and nga are not always differentiable. The 
print shows a number of lacunas, most not longer than one or two letters. A 
tsheg appears regularly before a shad after final -ng and occasionally even after 
other letters. The gap within a nyis shad or between a shad and the following 
syllable is of about one letter. 

Q is written quite carefully without hardly any omissions. Most mistakes 
occurred owing to confusion between vowels, nga and da, and ba and pa, and 
finally from a wrong segmentation of groups of letters (e.g. dag ’bar for dga’ 
bar, rnam sa for rnams, rdza pa for rdzab, so so for sas). 

* S — The Stog Palace Manuscript Kanjur 

The manuscript used is an offset reprint of the handwritten dbu can Kanjur 
preserved in the Royal Palace in Stog. The Kanjur itself was copied from a 
Bhutanese original “some time during the first half of the 18th century.”” The 
manuscript was reprinted between 1975 and 1980 in Leh (Ladakh).”* 

S is a very carefully and clearly written manuscript, with the lowest 
number of single variants among the collated materials. It has seven lines per 
folio; the gap within a nyis shad is of one to two letters, while the gap before and 
after a rkyang shad is usually of one letter each. A tsheg is found before a shad 
only after -ng. 

It seems that S follows a very unique punctuation system, omitting the 
rkyang shad in many passages where it is found in all other versions. The 
manuscript has a number of abbreviating subscripts. 

22 taeda 1977: 32. 
23 Skorupski 1985: xii. 
24 Skorupski 1985: xiii. 


PDF Version: BPPB VI (2002) 

¢ T- The Tokyo Manuscript Kanjur 

The Tokyo manuscript Kanjur was brought to Japan from Gyantse. It is said to 
have been copied between 1858 and 1878.”° The photocopies I obtained are 
rather difficult—in some passages impossible—to read, since usually features 
show through from the other side of each folio. It is hard to differentiate pa and 
ba, and the position of the tsheg is in many cases unclear. 

The folios have eight lines each and are written in dbu can. A tsheg before 
shad is rare and does not seem to be limited to syllables ending in a certain 
letter. As a rule there is no gap within a nyis shad or between a rkyang shad and 
the following word. In many instances, however, gaps are found, especially 
when the copyist did not make use of the remaining space before the end of a 
line. In such cases the second shad of the nyis shad is positioned at the end of 
the line, so that the gap can extend to as many as ten letters. Gaps which bear no 
traces of erasure of syllables are not mentioned in the present critical 

T contains a very high number of single variants, Among them are 
frequent omissions of single syllables, very often rkyang shad where all other 
versions have nyis shad (sometimes also vice versa), a host of wrong spellings 
(e.g. cad for can, Jol for jog la, snyas for snyam, bstas for bltas, dpen for dper, 
btsegs for brtsegs, yeng for yang etc.), corrections which have been inserted 
later, or lacunas resulting from deleted syllables, words or whole passages. In 
all, it is obvious that the copyist did not work very accurately. 

2 The Stemmatic Relations among the Representatives of Tib 

In the following I shall discuss the relations among the versions of Tib collated in 
my edition. The basis for my considerations will be the variants shown by the 
different manuscripts and prints. To be sure, drawing conclusions from variants is 
a complex undertaking, and I shall begin with some comments on the nature of the 
variants and their meaning for the establishment of a Kanjur stemma. 

First of all, we need to keep in mind that there are two main kinds of 
variants, namely ones which Harrison (1992a: xxv) has called “recensional” and 
“transmissional” variants. Recensional variants “reveal either extensive and 
deliberate editorial changes to the text, or the adoption of a different text 
altogether, rather than errors resulting from scribal lapses or casual attempts to 
improve or modernise the text (which are indeed usually deliberate, but generally 
rather trivial in scope).””° As for the TGS, we hardly find any variants of this type 
except for some few examples in the Tabo fragments. The absolute majority of 
variants in the TGS are of the second type, namely of a transmissional character. It 
is clear that if two manuscripts bear some identical redactional feature, they likely 
derive from a common ancestor, may have influenced each other, or else been 
influenced by a third manuscript which has the reading in question, but in cases of 
transmissional variants the situation is more complex. 

5 See Saitd 1977: 6 (401). His assessment is based on the years mentioned on the first page of 
the volumes devoted to different prayer texts. 
6 See Harrison 1992a: xxv. 


PDF Version: BPPB VI (2002) 

The decisive criterion for the evaluation of a transmissional variant is less 
the question if it was introduced deliberately or not, and more its degree of 
probability. That is to say, even if we have to deal with a trivial difference in the 
spelling of a word, such as sngan cad for what is commonly spelled sngan chad, 
the nature of the variant is rather indicative. This spelling is very rare in non- 
archaic manuscripts, and its occurrence in the same position of the text in two 
manuscripts can hardly be a coincidence (e.g. OK, n. 9). Similarly indicative are 
omissions, additions or changes of parts of the text which cannot be explained as 
caused by deliberate intervention on the part of an editor or scribe who intended to 
emend an “obviously faulty” text. At best, such readings can, in the light of other 
translations, clearly be shown to be mistakes. 

The direct textual surroundings in which the passage is embedded may 
have caused variant readings, as in the case of aberrationes oculi, where the scribe 
intends to continue to write from the word he stopped at but, owing to the multiple 
occurrence of the same word, jumps to the wrong position in the text. Such 
peculiarities, if found in two manuscripts, will naturally be less indicative in terms 
of their stemmatic relation. Generally speaking, the stronger text-internal reasons 
there are which could have led to the occurrence of the same transmissional 
variant in more than one manuscript, the less valuable the variant becomes for 
proving any stemmatic relation between the manuscripts. As the main principle in 
my analysis, each variant shared by several manuscripts underwent a thorough 
check in order to assess its probability. Only if the probability of its independently 
occurring was low could it be used as a variant indicative of stemmatic relations.”’ 

Also, the punctuation of the manuscripts and prints plays an insignificant 
role in my analysis. Whereas the punctuation clearly confirms the close relation 
within the three groups BDJNQ, LST and P;23 separately, it allows no conclusion 
regarding the relations of the groups to each other in terms of the degree of 
originality exhibited by the transmitted texts. 

2.1 The Three Phug brag Versions 

As mentioned above, three versions of the TGS are contained in the Phug brag 
manuscript Kanjur. The fact that the three versions do not, as one would expect, 
appear one after the other but instead in different volumes, may mean that they 
were not separated from the group of texts in which they were most likely 
transmitted or stored from earlier times. It may well be that such groups were 
brought together, or copied and then brought together, from different regions in 
order to collect them for a larger project. 

What, then, is the stemmatic relation between P;, P, and P3? First of all, it 
is clear that the three versions share a sufficient number of variants to prove that 
they derive from a common ancestor (Po). Just a few characteristic variants shared 
exclusively by P23 may be enough to illustrate this fact: 

(1) — OE, n. 20 Py3: gzugs, P2: gzug for gzungs. 

— OK, n. 15 P23: tshogs for chos. 

— OM, n. 25 Pj23: om. nyid. 

— 2A, n. 13 Py23: om. sbrang rtsi des (aberratio oculi). 

—2C, n. 8 Pigs: skyod for skrod. 

271 have virtually neglected variants such as kyis for kyi, pa for ba and ba ‘am for ba’am. 


PDF Version: BPPB VI (2002) 

— 2C, n. 22 Py23: spos for spobs. 

— 4A, n. 3 Pj23: ‘am for dang. 

— 6A, n. 9 Py23: om. de bzhin du. 

— 6B, n. 17 Py23: om. zhes. 

— 7A, n. 41 Py23: om. bu’i thum. 

~ 9A, n. 22, 23 P23: om. gang dang. 

-9C, n. 12 P23: om. shin tu. 

- 10C, n. 26 Py: ba dang / dge ba mngon par ’du bya ba snga; P;: ba dang / dge (ga) 
bas mngon par ’du bya ba snga [ga marked with three dots in the form of a triangle 
for deletion]; (P}23: aberratio oculi). 

—11C, n. 13 Py23: om. reg pa. 

— 11], n. 16 P13: ba for dbang. 

Each of these three versions, however, has essential variants”* of its own, not 
shared by any of the remaining two versions, which leads to the conclusion that 
none of the three could have served as the master copy for any of the other 

manuscripts. Such essential variants include:”” 
In P;: (2) — 0B" P;: om. / dbang. 
— 0H? P,: om. gi mdab ma de dag (aberratio oculi). 
- oH* P,: ‘bar ‘ong ba for par ’os pa. 
—4A/P,: om. gser gyi. 
— 5B'P,: dga’ for rko. 
— 10B *” P,: om. de la / bsod nams mngon par ‘du byed pa. 

In P;: (3) — OE ° P2: om. spos dga’i dpal dang /. 
—OE”” P2: om. dbyu gu here and inserts it some words later. 
— 0H * P,: sku pad ma’i for de dag gi mdab ma. 
~1B *P: om. ma byung yang rung (aberratio oculi). 
—9C"P,: inserts de ltar. 
—11C? P,: om. dran pa dang Idan pa dang | (aberratio oculi). 
—11H™ P,: om. rdo rje’i blo gros khyod nyid de bzhi’o // (aberratio oculi). 

In P;: (4) —0B™ P;: om. shes rab shin tu rnam par grol ba/. 

~ OD? P;: sngon bdag for gang dag. 

— OE ® P;: om. sa ‘dzin dang / (name of one of the bodhisattvas). 

—OL” P3: om. /tar nyan pa dang /bcom Idan ‘das kyis ‘di skad ces. 

~1C'P;: phye for gyes. 

—2A™ P: rig par mthong ngo for mthong bas rig go. 

—4C"P,: nyi shu pa for mi nyung ba. 

~ 5B ™P;;: ins. mig shig du yongs su dag pas sems can thams cad de Ita bur 

mthong nas | de bzhin gshegs pa’i (aberratio oculi). 

—6A"P,: om. ‘dzam bu’i ‘bras bu ‘am / ta la’i ’bras bu ‘am (aberratio oculi). 

—11LE? P,: ¥ig for mchod. 
Knowing that all three versions derive independently from a common ancestor, 
we should further inquire whether there are essential variants shared by two of the 
manuscripts. If this is the case, we would be obliged to assume that there was 
another step in the copying between Py and P23, shared by two of the manuscripts. 
Regarding this analysis, the facts are very clear: 

There are virtually no essential variants common to P; and P;. The few 

readings worth mentioning are: 

28 By “essential variants,” I mean variants which consist of more than just a different way of 
spelling or an erroneous omission or change of a single letter within a word. 

9 Theoretically, it is, of course, possible that one of the manuscripts was checked with the help 
of Py while being copied from one of the other two manuscripts, so that major mistakes found in 
the master copy could be corrected. While we can never be sure how exactly the copy of any given 
text came into existence, in our case the evidence supplied by the variants is so overwhelming that 
there is no need to assume a more complicated situation, at least as long as the appearance or non- 
appearance of variants in the different manuscripts can be explained satisfactorily. 


PDF Version: BPPB VI (2002) 

(5) —9C, n. 5: Pi3: g.yog for gsog. 

— 10C, n. 30: P,3: om. grangs dang /; P2: grangs dang / in small letters 
above the line (the position in the text where grangs dang / is to be 
inserted is marked with a cross).°° 

— 11H, n. 20: P;3: om. de. 

The situation is similar in the case of P2 and P3: 
(6) ~ 0G, n. 11: P23: om. bye ba?! 

~ OI, n. 4: P23: ni for gi. 

— 7A, n. 7: P23: om. des. 

— 11] (verse), n, 23-25 P,: de’i tshe de dag mdo ‘di thos //, 

P): de ‘i tshe de dag mdo ‘di thos gyur nas /, 
P;: de tshe de dag mdo ‘di thos gyur nas // 
for de tshe de dag gis ni mdo ‘di thos // ( gyur nas in P,; was 
probably the reading of P, and omitted by P,) 
The common readings of P; and P,, on the other hand, do not need any further 

(7) - OC, n. 4 and n. 7: P}2: insert chen po. 
— OE, n. 4: Pj2: blo gros for zla ’od. 
—0J, n. 19: Py: dga’ for gda’ (metathesis). 
— OK, n. 28: Pj2: om. complete pada 0.4b. 
— 1A, n. 35: P)2: insert rnams ni yang dag pa (dittography). 
— 2B, n. 5: Py: om. gti mug dang /. 
—TA, n. 36: P12: insert dngos por yang ston pa dang* / ’brog na gnas 
pa’i [*P,: da nga for dang] (aberratio oculi). 
- 10C, n. 10: P12: om. dge slong ngam / dge slong. 
— 12B, n.18: P}2: om. tha dang / mi dang. 

It is now clear that we must assume that P; and P; have a common ancestor Py, 

from which the essential variants found in P;2 but not in P; derive. The stemmatic 
relation between P72; should then be: 


Px Py (?) 

P, Pp P3 

One more word concerning the single variants of P3: while it is clear that the bulk 
of variants found in P; result from a scribe copying the master copy without 
careful attention and sticking to his own particular spellings (see above), it is 
equally evident that several passages contain redactional variants, that is, that 
certain passages or words have been altered consciously by the scribe or had 
already been altered in a hypothetical copy Py between Pp and P3. These 
alterations reflect the attempt on the part of the redactor to “emend” or to “make 
sense of” passages which were hardly understandable for a Tibetan reader but 

30 The words are part of a well-known enumeration of merit resulting from the propagation of a 
siitra. This list appears twice in the TGS. Obviously P; at first had the same omission, probably a 
mistake of Po, but it was later emended. Thus this variant is useless for trying to establish a close 
relation between P; and P3. 

31 The text in this passage reads ... kha ma bye ba bye ba khrag khrig.... A shared variant like 
this cannot be used to prove a close relation between two manuscripts. The text seems to contain a 
dittography, and the copyists could independently have thought it necessary to “emend” it. 


PDF Version: BPPB VI (2002) 

which could be rendered so by the use of stylistically more common terms. The 

variants in question are: 

(8) — OE ! P;: mi skyo ba for yid skyo: The name Yid skyo (*Khinnamanas) in the 
enumeration of the participating bodhisattvas (no. 40) seems in fact not very adequate as 
the name for a bodhisattva. Thus the editor of P; changed it to mi skyo; the syllable ba is 
found in all three versions P;);. The name mi skyo ba in P; means “Unweary.”** 

— 0H‘ P5: om. mog: The text in this passage runs ... mdab ma de dag thams cad mog mog! 
po dang / nog nog po dang.... It is thus clear that the double mog mog is parallel to the 
following nog nog and here is no dittography. Most probably the redactor of P; tried to 
“emend” the second mog by taking it as dittographic. 

—OL ’ P;: om. the passage from “v—»“ till “<—v.“ The text runs: bcom Idan ‘das kyi ‘ Itar 
nyan pa dang / bcom Idan ‘das kyis ‘di skad ces*” bka’ stsal to /!, This is part of the 
common introductory formula in many siitras (see n. 49 in the translation). Though 
bcom Idan ‘das kyi ltar nyan pa is a standard expression in the Tibetan (for Skt. ... 
bhagavatah pratyaSsrausit /), the genitive ... kyi seems impossible to construe. The 
redactor of P; therefore dropped the part between the genitive and the following bka’ so 
that the genitive can be constructed with bka’. 

— 2A ™ P: rig par mthong ngo for mthong bas rig go. The text runs ... de bzhin gshegs 
pa'i ye shes mthong bas rig go” //: “[I} realize with the Tathagata’s mental vision 
that....” The term de bzhin gshegs pa’i ye shes mthong ba, Skt. tathagatajfa@nadarsana, 
was obviously unknown to the editor of P;, which led him to change the sequence of the 
words. He thus probably understood: “Realizing the Tathagata-knowledge [I] see 

— 2B ° P;: yang skyes bu thabs. The text here reads ... de bzhin gshegs pa yang* thabs la 
mkhas pas bung ba bsal ba de bzhin du....: “... just like the [person who] removed the 

bees, also the Tathagata, with skill in [the application of appropriate] means 
(upayakuSala)....” | have supplied the subject of comparison in brackets (“person who”), 
just as the editor of P; felt it necessary to mention it. 

-~4C"P,: Jo ni nyi shu pa for lo ni mi nyung ba. The context is the nugget of gold 
remaining in the filth “for not [just] a few years”. The redactor of P; has changed the 
text to “twenty years”. 

— 6A! Ps: shing shun for phyi shun. It is in fact puzzling why the text here speaks of a 
seed within the “outer” peels. In an attempt to make sense out of the text, P; was altered 
to “the peels of a tree,” given that the larger context is the essential identity of seed and 
tree. (The seed, however, is of course not within the peels of the tree, that is the bark, 
but within the fruit.) 

— 8B ” P;: dgon pa’i khang pa for mgon med pa’i khang pa. The poorhouse in the text 
becomes in P; a “house [in the] wilderness.” As the house is compared to the places of 
rebirth, the variant of P; works well, even though the poor woman lodges in a poorhouse 
in 8A (also in P)). 

The variants neither of P;, P2 nor P; suggest an affinity with any of the other 
collated versions which is stronger than what is “natural” when dozens of variants 
are compared.” 

As a result of this analysis, we can now assume that the reading of the 
archetype of the TGS in the Phug brag lineage (Po) can normally (but not 
automatically, as has been shown above) be posited if either P; or Pz sides with 
P;. If, however, P; and P2 stand against P;, no such conclusion as to the original 
reading of Po is possible since P; and P2 derive from the same master copy Py. 

32 Skt. *Khinnamanas (“[The One with} a Grieved Mind”) as a name for a bodhisattva is 
probably meant to convey the sense that the bodhisattva is worried about the suffering of all living 

33 P, usually shares the reading cig instead of gcig with BNQ, ’a sngon ’jug instead of m- with 
BJNQS, and in some instances the omission of a prefixed b- with BJNQ. This can, of course, be 
interpreted as no more than a similar mannerism in certain spellings, and certainly not as a genetic 


PDF Version: BPPB VI (2002) 

In the following I will deal with the position of Po in relation to the other 
Kanjur lineages. A statistical analysis reveals that the Kanjurs of the Tshal pa 
lineage do not use the rkyang shad in 110 cases where L, on which the 
punctuation of the critical edition is based, has it. Pp lacks the rkyang shad in sixty 
cases where L has it and adds rkyang shad in thirty positions where it is not found 
in any of the other Kanjurs. Fifty of the sixty positions where Pp lacks the rkyang 
shad are identical with the position among the 110 cases where the Tshal pa 
tradition does not add a rkyang shad. 

Based on punctuation alone, the result thus confirms the preceding studies 
in Harrison 1992a, Schoening 1995, Silk 1994, and Skilling 1994, which 
characterized the Phug brag manuscripts as an independent tradition in the sense 
described above. We should further inquire about the nature of the variants found 
only in P;23: can they all be characterized as transmissional variants, or are there 
also readings which suggest a deliberate editorial intervention? On the whole, P23 
seem to have no major editorial variants; in view of the high number of faulty 
transmissional variants, I am not even sure if the following few cases, the only 
such I could find in the whole text of P23, should be labeled recensional. We will 
come back to this question later: 

(9) —4A, n. 3 Pj23: ’am for dang. 

— 7B, n. 1 P13: pa for cing in ‘os su(s) gyur cing.... 

—7C, n, 23 P23: thar pa for thar byed, 

— 8B, n. 11 Py23: byed cig / for byed par: P;23 use the imperative particle cig without the 

imperative form of byed pa: byos. 

— 111, n. 16 P23: rgyal ba for rgyal dbang. 

— 111 (verse), n. 23~25 P|: de’i tshe de dag mdo ‘di thos // (also A: de’i for de; gyur nas, 
which is found in P2;, was probably the reading of Pp and omitted 
by Py), 

P,: de’i tshe de dag mdo ‘di thos gyur nas /, 
P;: de tshe de dag mdo ‘di thos gyur nas // 
for de tshe de dag gis ni mdo ‘di thos // 
—12C,n. 15, 16 P23 zla nya for nyi zla; P): zla ba nya. (P23 in accordance with Ch, and 
— Py3: a colophon. 

2.2 The Kanjurs of the Tshal pa Lineage 

The so-called Tshal pa edition is named after the monastery where it was 
produced, namely Tshal Gung thang monastery in east Central Tibet (dBus). The 
colophon of the siitra section of the Lithang Kanjur™ gives the years 1347 to 1349 
as the time of origin of the Tshal pa Kanjur and states that it was based on a 
number of source materials that formed part of the prestigious Old Narthang 
Kanjur, and on a copy of the Old Narthang edition itself.°* This Tshal pa 
manuscript or one of its copies®® later served as the basis for the Lithang Kanjur at 

4 The colophons of the Lithang Kanjur are transliterated and translated in Samten 1987. 

35 Harrison (1994: 297f.) doubts that the Old Narthang was more than a “collection, in some 
cases of multiple copies, providing the raw materials for an edition proper,” and consequently he is 
hesitant to call it an edition in the strict sense of the word. He further stresses that the colophon 
quoted in Samten 1987: 30f. is not to be understood as implying that a number of canonical 
collections in addition to the Old Narthang was used (Harrison 1992b: 80, n.14). 

36 Samten (1987: 17), basing himself on Tibetan sources, states that the original Tshal pa 
Kanjur was brought from ’Phying ba sTag rtse, where it was later kept, to "Jang sa tham on the 


PDF Version: BPPB VI (2002) 

the beginning of the seventeenth century after several famous Tibetan scholars 
had retouched the manuscript in the past centuries?’ As the manuscript which 
served as master copy for the Lithang wood blocks was at that time kept in 
’Phying ba sTag rtse, the group within the Tshal pa lineage which is based on this 
manuscript is called the ’Phying ba sTag rtse group. 

A copy of the Tshal pa Kanjur had been utilized two hundred years earlier 
for the production of the earliest wood-block print edition: the Yongle 7 
edition of 1410. This Yongle wood block became the master for all following 
editions in the Peking line: the Wanli 77 edition of 1606 (the same blocks were 
used), the three Kangxi EER editions (1684-1720; with new blocks) and the 
Qianlong edition of 1737 (using the Kangxi blocks). Though the blocks partially 
suffered from the effects of time, as is now well known, even the new wood 
blocks were produced on the basis of impressions from the old ones—a 
phenomenon that led Eimer to coin the term “technical identity” for all the Peking 
block prints.°* For my edition I collated B and Q as two representatives of this 
Peking line. I will come back to the relation between these two later. 

For the ’Phying ba sTag rtse-group, I collated J along with N and (for the 
main chapters 0L-9C) D. Concerning N, Tucci (1949: 479) states that according to 
Tibetan sources it was based on the Tshal pa Kanjur. This could well be so, given 
that in the studies of the Pratyutpannabuddhasammukhavasthitasamdadhisutra, the 
Lokanuvartanasitra, the Vimalakirtinirdesa, the Drumakinnararajaparipycchasitra, 
the Udanavarga, the Mahasiitras in vol. mdo Ra (25), the Abhiniskramanasiutra in 
vol. mdo La (26), the Sandhinirmocanasitra in vol. mdo Ca (5),°? and the recent 
editions of the Aksayamatinirdesasiitra”’ and the Sdlistambasitra"', N shows a 
close affinity with the Kanjurs of the Tshal pa lineage. However, in other cases, N 
sides with the Them spangs ma line. This seems the rule for texts of the Vinaya 
section and also for the Punyabalavadana, the Sukarikavadana, the Jnanakasiitra- 
buddhavadana, the Brahmajalasitra,” and, as dealt with in recent studies, the 
Heart Siitra in the Prajfapdramita and the Tantra section,’? and the Sekoddesa.“ 

Finally, I also consulted D, which is known to have drawn mainly from the 
Lithang Kanjur but also to have incorporated readings from a copy based on the 
Them spangs ma, and which must therefore be categorized as conflated.*> For any 
edition which does not primarily aim at illuminating the relations between the 
different Kanjurs but is in search of the text from which all versions may have 

orders of the king of "Jang sa tham, and there copied in order to establish a new edition. Harrison, 
however, states that only a copy of this Tshal pa Kanjur was used (Harrison 1992b: 80). 

3” See Imaeda 1982a: 179. 

38 Eimer 1992: 180. 

3° For information on the texts mentioned up to here see, Eimer 1992: xvf. 

4° See Braarvig 1993a, 1993b. Following an analysis of the variants, he states that “there is 
reason to assert that N is a direct descendent of J” (1993a: x) and that the cases where N does not 
follow the single variants of J “may be explained as recensional changes, and even as conflations 
with other versions.” 

41 See Schoening 1995a, 1995b. According to Schoening, concerning N in the Salistambasitra 
“the wording mostly agrees with the Lithang Kanjur as opposed to the Peking tradition or West 
[=Them spangs ma] tradition.” (1995a: 140). 

*2 For the information on the texts mentioned till here see Eimer 1992: xv. 

See Silk 1994, 

“ See Orofino 1994. 

See Samten 1987: 18f. The copy based on the Them spangs ma Kanjur was the IHo rdzong 
Kanjur, compiled from 1595 to 1658 on the advice of the Fifth Dalai Lama. 


PDF Version: BPPB VI (2002) 

derived, the value of conflated material is low. When the nature of the variants of 
D became clear, after the collation of about half of the text of the TGS, I decided 
instead to invest my time in other, more promising versions. 

Let us now turn to the results of my study of the variants of the Tshal pa Kanjurs. 
In the Peking group we have B and Q. The high number of common variants not 
shared by any other of the collated versions makes their close relation obvious. I 
will give just a few examples: 
(10) — OE, n. 26 BQ: par sgrags for bsgrags. 

- OI, n. 16 BQ: rtsom for tshom; JP 2ST: tsom; 0K tg BQ: rtsom; JP 2: tsom. 

— OJ, n. 15 BQP3: gda’ for bda’; P2: mnga’ for bda’: see Bth: zhim. 

~ 1A, n. 4 BQ: na/ tha skyes bu lha’i mig can for na / skyes bu lha’i mig can. 

— 4B, n. 13 BQ: bgo for bsgo. 

— 6A, n. 6 BQ: om. na. 

— 7A, n. 25 BQ: la mi for lam. 

— 8A, n. 18 BQ: sems can dang for sems dang. 

— 8C, n. 11 BQ: ji lta for byis pa. 

—8C, n. 12 BQ: nyid for ba. 

— 9B, n. 10 BQ: Jug for ‘dug. 

— 10B, n. 25 BQ: bzhin for zhing. 

— 10B, n. 29 BQ: om. gzhan. 

— 10E, n. 20 BQ: sems dpa’ for sems pa. 

— 11E, n. 3 BQ: om. Jas. 

— 11H, n. 4 BQ: rtogs so for gtogs so. 

~-12C, n.1 BQ: insert ’di ni. 

There are some particular variants of B worth mentioning: 
(11) — 0C, n. 20 B: nyas dang for nya dang; Q: lacuna of one letter between nya and dang. 
—0C, n. 25 BP,23: dang / thabs for dang thabs; Q: lacuna of one letter between dang and 
— OK, n. 24; 1C, n. 6, 11, 12; 5C, n. 9, 11; 8C, n. 15; LIT, n. 9 (all verses): B reads 77 
instead of yi (in B 7 can count as a syllable), Q has a compressed yi. 
— 1A” B: ston pa te for ston te; Q: lacuna of one letter between ston and to [Q: to for te]. 
—2A’B: sbrang brtsi for sbrang rtsi; Q: lacuna of one letter between sbrang and rtsi. 
~ 2A ° BP: pas / bskrad; P;: pas // bskad; Q: lacuna of one letter between pas (without 
tsheg) and bskrad. 
— 6A” B: rgyal chen por ’gyur for rgyal po chen por ’gyur; Q: rgyal po chen por 
—10B™ B: dpen gcig for dpe gcig; Q: lacuna of one letter between dpe and gcig. 
—10C, n. 18 BJP3: yid rang for yi rang; Q: lacuna of one letter between yi and rang filled 
with tshegs. 
—12C, n. 9 BP»: sngags for bsngags; Q: b- of bsngags with very small letter. 
— 12D, n. 7 ABP}23: yid rangs for yi rangs; Q: lacuna of one letter between yi and rangs. 

In all the cases under (11) it is obvious that the archaic, uncommon, or even 
wrong reading found in B must have also been part of Q or its predecessors before 
letters were erased or replaced, or before missing letters were inserted on the 
block prints by some redactor. This resulted in the lacunas or in letters which 
appear much smaller or compressed on the lithograph in Q.*° The examples given 

46 The physical process of such alterations is described, with several examples, in Eimer 1992: 
1-16; 191-202. Eimer further states that the date of a larger revision of the Peking wood blocks 
may have been during the production of the Mongolian Kanjur, which falls within the same period 
as the 1717-20 reprint (1983a: 59f.; 1992: 12). The sources for the alterations, however, are not 
entirely clear. In any case we must treat the Peking Kanjurs from a certain date on as conflated. 
(See also Eimer’s study on the Hevajratantra II, v. 1-2, in which he shows that a new wood block 
containing two more slokas was carved between 1684/92 and 1717/20 (Eimer 1992: 165—174).) 


PDF Version: BPPB VI (2002) 

above thus show that B copied faithfully many of the peculiar or wrong readings 
which must already have been part of the Yongle or Wanli block print on which B 
was based. 

On the other hand we find examples (12) where both Q and B have the 
same reading but where Q shows traces of emendation (resulting in lacunas etc.) 
which led to the reading. We could assume that these emendation were already 
part of the master copy of B, namely the Yongle or Wanli block print, and that 
they were adopted by the scribe of B. In Q, however, due to the “technical 
identity” of the wood blocks, traces of the corrections can still be found. Under 
(13) I have collected cases where B does not follow the text of Q which, in these 
passages, differs from the other Kanjurs. Examples include: 

(12) — OD, n. 3, en. d JP 123: drug beu’i for drug cu’i; Q: lacuna of one letter between drug and 
cu’i (same situation in OE, n. 31, en. vv). B: drug cu’i. 

— 2C, n.15 BQ: sol for sel [B: g- of (g-)sol marked with three dots in the form of a triangle 

above for deletion]; Q: lacuna of one letter between mongs and sol. 

-— 4A, n. 26 JP\3: sum beu for sum cu; Q: lacuna of one letter between sum and cu filled 

with tshegs. (B: sum cu). 

~ 10D, n. 18: Q: yi for i (verse) [yi compressed, most probably altered from ’i]; B: yi. 

~11H"Q: ma rgyas in sangs ma rgyas compressed; P23T: om. ma: Obviously Q at first 

also omitted ma and read the common sangs rgyas. Later ma was inserted by carving 
the whole phrase sangs ma rgyas with narrowed letters. B reads sangs ma rgyas. 
(13) — OJ‘ Q: dbugs for dbus. 

—1A™"Q: ‘khor for ’khod; B: mtho’ for ‘khod. 

—4C*Q: des for ngas. 

~5A'Q:so so for sas. 

— 5A” Q: stong for steng. 

—6C' Q: yang bar for yod par. 

—7A, n. 5 Q: la mthil for lag mthil; B: la gar thil for lag mthil. 

—7C, n. 30 Py,: mi for ming; Q: med for ming. 

~8A‘%Q: om. byed of par byed 'gyur [par ’gyur with larger spaces between the letters 

than usual]. 

— 10D, n. 21 B: snying for stong; Q: steng for stong. 

—12B™" Q: inserts yod pa in bla na med pa yod pa yang dag par rdzogs pa’i byang chub. 
It is, of course, very likely that the scribe of B polished archaic spellings which 
seemed inadequate to him and that he more or less unconsciously did away with 
obvious mistakes in the master copy. As stated above, most of the cases under 
(12) can be accounted for if we assume that the wrong or uncommon readings had 
already been emended in the Yongle or Wanli wood blocks. In 2C, n. 15 we have 
to assume that the prefix g- was still found in the master copy of B, copied by the 
scribe of B, and then deleted. Later the prefix was also deleted in the wood blocks, 
leading to the lacuna found in Q. 

The majority of readings under (13) can be explained by assuming that the 
scribe of B corrected obviously faulty readings when copying from the 
Yongle/Wanli lithograph (0J, 5A, 5A, 12B), or that one of the latter two was not 
clear in the concerned passage and was consequently read differently by the scribe 
of B and the carvers of the Wanli/Kangxi wood blocks, for which the 
Yongle/Wanli wood blocks served as technical models (1A, 4C, 6C, 7C, 10D).”” 

Eimer assumes the Lithang Kanjur to be a possible source for it (p. 172). Eimer (1992: 143; 172), 
Harrison (1992a: xxxi), Schoening (1995a: 140; 172ff.) and Skilling (1994: xxxivf.) show that 
some readings of the Peking line were at a certain stage altered on the basis of a comparison with 
J. No such examples are found for the TGS, and two examples quoted under (12) show that those 
alterations cannot derive from J, since J bears the pre-emended readings. 

‘7 For this assumption see Eimer 2000: 32f. 


PDF Version: BPPB VI (2002) 

The omission of byed in 8A may not have been a feature of the master copy which 
served the scribe of B. The fact that the spaces between the letters in that passage 
appear to be larger than usual means that byed was still in the first wood block, 
and only later erased, and accordingly more space became available for the single 
letters—space made use of by the carvers of the Kangxi edition. Regarding 7A, 
the term Jag mthil may have appeared in an abbreviated form in Yongle/Wanli, 
and then its complete spelling may have been reconstructed differently by the 
scribe of B and the carvers of the Wanli/Kangxi block prints. As a result of the 
above analysis, we can say that there is no need to consider B as having 
undergone any contamination from outside the Peking tradition. 

Let us now turn to the second group on the Tshal pa side. We should first focus on 
the relation between J and N in order to judge if it is plausible that N derives from 
the Tshal pa Kanjur, as maintained by Tucci (see above). There are several 
variants shared solely by J and N (and partly by D) whose existence would be 
hard to explain if J and N did not derive from the same master copy: 

(14) — 0G, n. 9 IN: phang for '‘phang. 

— 5A, n. 28 JN: ’khod for ‘khor. 

— 6C, n. 12 DIN: thob for mthong. 

— 8C, n. 24 JN: sman for dman. 

— 9A, n. 5 DIN: om. cig. 

— 9A, n. 29 DIN: bskogs for bkogs. 

—11A, n. 14 JN: ’khor for kha [N: ’a sngon ‘jug of ’khor in a small letter inserted later]. 

— 12C, n. 3 JN: sems dpas for sems dpa’. 

The variants 6C, 8C, 9A (n. 29) and 11A are of special importance, since it can 
hardly be imagined that they appear coincidentally in two manuscripts. Section 
11A is revealing inasmuch as it documents that the reading was probably first the 
correct kha (in skyes bu ’dul ba’i kha lo sgyur ba) and was only later altered to 
’khor by adding the ‘a sngon ‘jug within a very limited space. This fact indicates 
that N derives rather from a master copy shared with J (the ’Phying ba stag rtse 
manuscript) than from a manuscript on which also the Peking Kanjurs are based. 

I must admit that the number of variants common to JN which I have 
given is not very high. The reason for this is that both J and N in general do not 
have a lot of variants, and many of the shared readings are only minor (Kyi for kyis 
etc.), so that listing them here would not produce any more clarity. However, we 
should in any case check the relation of N to the Peking line (against J). The only 
serious variant of N common to BQ is: 

(15) — 12B, n. 16 BNQ: de for nga. 

There is thus no reason to assume that N derived from a manuscript close to the 
Peking tradition. This makes the assertion that N, like J, is based on the ’Phying 
ba stag rtse manuscript more plausible. 

Does N follow the Tshal pa transmission through all the text or do we find 
variants shared with representatives of the Them spangs ma lineage? Let us 
examine the nature of the variants shared between the Them spangs ma and N: 

(16) — OB, n.7, 8 BJQPi23): bzangs can for bzangs tsan dan; N: tsandan (®%3') for tsan dan. 

— OB, n. 21 BJQP)23T: sems can thams for sems thams. 

8 That N is based on the ’Phying ba stag rtse manuscript is posited in the studies of Harrison 
(1992a: xxx), Schoening (1995a: 131) and Skilling (1994: xxxix). Braarvig, however, states that N 
“is a direct descendent of J” (1993a: x) and explains the fact that several single readings of J do 
not appear in N as “recensional changes, and even conflations with other versions” (p. 131, n. 3). 


PDF Version: BPPB VI (2002) 

— 0H n. 7 BJQP)2;: om. kyang. 

— OI, n. 2 BJQ: tsom for tshom. 

— OF, n. 16 BQ: rtsom for tshom; JP 12ST: tsom. 

— OL, n. 6 BJQ: om. yid la. 

— 2B, n. 4 BIQP 23: om. de. 

— 3B, n. 1 BJQP2T: om. de bzhin du (aberratio oculi). 

— 4A, n. 16 BJQ: om. gzhan dang (aberratio oculi). 

— 4B, n. 2 BJQ: om. Ja. 

—5C,n. 14 LN: ’thob for thob; S: mthob. 

—6C, n.17 BDI: spang for sbyang; Q: spang or sbang (?). 

— 11D, n. 5 BJQ: om. nas. 

— 11G, n. 19 BJQP,23: om. phra rab kyi rdul. 

— Throughout the text: BJQP12:T: pad ma for padma (LSN: 43); only in 0G, n. 20 BJQ: 
pad mo for padma; N: padmo; P1237: pad ma. 

— Throughout the text: NS: gangga (If); L: gangga; BIQT: gang ga (B in two passages 
with P23); Pi23: gang ga. 

The number of major variants where N sides with the Them spangs ma Kanjurs is 
surprisingly high. Among them are, it is true, several stock phrases and terms 
appearing in other passages of the TGS, which could easily have been emended 
without drawing on other versions (e.g. the list of attributes of arhats in 0B 
appearing regularly in sutra literature). In other instances it is impossible to 
explain in this way why N displays the same reading as (all) representatives of 
Them spangs ma against the Tshal pa family (e.g. the non-omission of (the 
aberratio oculi:) gzhan dang in 4A), so that we must assume that the editors of N 
accessed an unknown text of the Them spangs ma family which they sometimes 
preferred to follow. The situation is thus very similar to the Salistambasitra, 
concerning which Schoening states: “The position of N is curious because it 
usually agrees with the Eastern tradition, in particular with CJ, and yet in a 
handful of cases it has Western readings. In each of these instances, N was 
probably emended either by contamination with the Western tradition or 
independently by scribes or editors.””” 

The number of common variants with the Them spangs ma group in the 
TGS is too low to allow an opinion where the version used for the emendations 
could have come from. One realistic possibility would be the Shel dkar rdzong 
manuscript, which we know served as the basis for a number of edited works in 

It remains to inquire about the position of D within the stemma of Kanjurs as it 
pertains to the TGS. The Derge Kanjur is well known to be a conflation of J and 
the so-called IHo rdzong Kanjur, a Kanjur which was produced on the basis of the 
Them spangs ma, with or without an intermediate copy.” In the case of the 
Aksayamatinirdesasutra, however, the Derge version turned out to be a copy of 

* Schoening 1995a: 170; further see Pagel 1999: 204. 

5° Cf Eimer 1992: xv. The Narthang Kanjur was carved near Shel dkar rdzong, the place where 
L was also copied from the manuscript Kanjur kept at the monastery of Shel dkar (see Harrison 
1992b: 80, Pagel and Gaffney 1996: ix). Braarvig (1993a: x) reports that N, which in the case of 
the Aksayamatinirdesasutra is based on J, supplied part of the text (one whole folio obviously 
missing in the early Tshal pa tradition) by drawing on the Them spangs ma version. It would be 
interesting to see which of the representatives of the Them spangs ma line comes closest to N in 
this passage. 

>! Harrison 1992b: 79. 


PDF Version: BPPB VI (2002) 

“the first Tshal pa manuscript, or an early copy of it” and could thus “be 
considered a better transmission of the original Old sNar than than JNQ, though 
correct readings may of course have been preserved in JQ, readings lost in the 
transmission to D[erge (Prague)].” (both Braarvig 1993a: x-xi). 

Are there any variants which suggest that in the case of the TGS, too, D 
may be the copy of an earlier, better representative of the Tshal pa lineage? If so, 
this would mean that D has unique archaic readings and variants which produced 
a better transmission of the Tshal pa line than BJ(N)Q. What about the single 
variants of D? They include: 

(17)  —-1A"D: ’gyed for ’byed; B: byed; Q: ‘phyed. 

—4C'D: bzhig for ba zhig. 

—5A "D; inserts byed (dittography). 

— 5B *D: slobs for spobs. 

—5C"D: shig for shes. 

—7A™D: ‘dong bar for ‘dod par. 

—7C™ D: mi for ni. 

—8C'D: des for dus. 

-~9A“D: gyis for gyi. 

It is obvious that none of the variants can be said to represent a better reading than 
the other versions do. They are all explainable as mistakes made by the copyist. 
There is thus no reason to conceive of D as a representative of the early Tshal pa 

The affiliation of D with both main transmissional lines is also obvious. It 
shares enough variants with BJNQ (in particular, the punctuation: it drops the 
shad together with BJNQ not less than 83 times) to allow it to be based on the 
Tshal pa tradition. On the other hand, the apparatus documents numerous cases in 
which D sides with LST—a fact that leaves no room to doubt that the editors also 
drew upon the IHo rdzong Kanjur to establish the text of D. 

That D is part of the ’Phying ba sTag rtse group in the Tshal pa family is 
clear from the common omission of shads in 15 cases and other variants shared 
with JN but not with other representatives of Tshal pa: 

(18) —1A,n. 30 DJNL: srid for sred. 

-—1C,n.9 DIN: gos de for gos te. 

—1C, n. 20 DIN: gis for gi. 

—6C, n. 12 DIN: thob for mthong. 

~ 8C, n. 9 DINP)23: Api for kyis. 

~—9A,n. 5 DIN: om. cig. 

—9A, n. 29 DIN: bskogs for bkogs. 

Given that D is based on J, is it possible to draw any conclusions on the question 
whether N is a direct descendant of J or if it is a sister of J? Let us examine the 
nature of the variants shared by DJ, DN and JN. The two competing models for 
the relation of DJN are: 


PDF Version: BPPB VI (2002) 

(a) (b) 
*Phying ba sTag rtse manuscript *Phying ba sTag rtse mansucript 
Lithang Lithang Narthang 
Narthang Derge Derge 

If we turn to the variants shared by D./ (but not N), we find the following 
(19) — Throughout the text: BDJQP)2;T: pad ma; (LSN: 4a’). [General preference of N for 
transliterations written without intersection] 
—3A,n.7 DJ: om. de. 
— 4B, n. 8 DJ: om. /. 
-4C,n. 11 BDI: ‘di ni for ‘di na. 
— 6C, n. 17 BDJ: spang for sbyang; Q: spang or sbang (7). [NV emended to sbyang ba “‘to 
purify,” common in the TGS] 
—7C, n. 28 BNQ: kyis for kyi DDILSTP 23: Ay). 
Variants common to JN (but not D): 
(20) —2A, n. 4 BINQLSP3: ba // for ba /. 
— 2A, n. 18 JN: gis for gi. [Gis is obviously erroneous in sbrang tshang gis nang na.] 
~—4A, n. 6 INQP;: ci for gci; n. 11 INQP),;: ci for gci. [The Derge editors emended to the 
MVy standard: bshang gci (MVy 2626-2628).] 
— 4B, n. 11 BINQLP3: om. gser gyis [The omission contradicts the same construction in 
2A.4f.; probably, therefore, gser gyis was inserted by the Derge editors. ] 
— 4B, n. 20 JN: Apis for kyi. [Kyis is obviously erroneous in ... sangs rgyas_kyis tshig bla 
dags so /I.] 
— 5A, n. 28, 29 JN: ’khod de for ‘khor te; BQ: ’khor de. [The Derge editors selected the 
well-known formulation ...’khor ba na ’khor te /: “to wander around in samsara.”] 
—6C, n. 13 JN: bshigs for bshig; P;Q: pa shig for bshig. [The Derge editors probably felt 
bshig adequate as the perfect form of ‘jig pa, with gzhig as the future form.] 
~6C, n. 15 BINQP)2;: gang for nang. [The pada according to ADLST seems to have a 
clearer meaning: mi shes nang na gnas kyang rlom sems med //, That this is the 
original sense is confirmed by Bth (dbus) and Ch; (fS"4).] 
—7C, n. 9 JN: bor de for bor te. [The Derge editors emended to the standard.] 
— 8C, n. 22 JN: ma shes gyis for ma shes kyis; BQ: bgyis. [The imperative form gyis from 
bgyid does not make sense here.] 
~ 8C, n. 24 JN: sman for dman. [The reading sman “medicine” here is meaningless and 
was therefore emended by the Derge editors.] 

Variants common to DN (but not J): 
(21) — In three passages: DNLST: de bzhin gshegs pa'i sku gzugs for de bzhin gshegs pa’i 
gzugs. [DN is made honorific.] 
— OL, n. 6 BJQ: om. yid la. [DN adopt the stock phrase.] 
— 1B, n. 9 DNP3S: -s zhes for —s shes.; 5B, n. 1 DN: -s zhes for —s shes. [DN adopt the 
standard; see Tshig mdzod s.v. zhes.] 
— 2B, n. 4 BIQPi23: om. de of de bzhin du. [The term de bzhin du appears often in the 
TGS, and de was therefore probably thought to have been inserted in line with Them 
spangs ma.]} 
— 3B, n. 1 BJQP,T: om. de bzhin du (aberratio oculi). 
- 4A, n. 16 BJQ: om. gzhan dang (aberratio oculi). 
— 4B, n. 2 BJQ: om. Ja of la bitas. 
~4C, n. 16 BJQP,2T: glo bur for blo bur. 
— 5A, n. 17 BJQ: ‘dug go // zhes for ‘dug go zhes. 


PDF Version: BPPB VI (2002) 

—7A , 0. 33 BJQP;: kun du for kun tu. [DN adopt the MVy standard.] 

—8C, n. 4 DNP);: re zhig for re shig. [DN adopt the MVy standard (8132).] 

In any case, we should keep in mind that both D and N are conflated with the 
Them spangs ma line, so that there is always the possibility that one or both of 
them did not follow the text of the master copy but adopted a Them spangs ma 
reading. If D or N thus does not share a reading with JN or JD respectively, the 
variant is only significant in those cases where D or N is not easily accounted for 
as an obvious mistake of the scribe (single variant reading) and if it is not in 
accordance with the Them spangs ma versions, Otherwise, we cannot exclude the 
possibility that D or N was altered on the basis of the Them spangs ma, and so it 
would be inappropriate to make statements about the stemmatic relation of DJN. 
We should investigate the nature of the variants in order to see if there are 
convincing reasons which could have led the editors of D or N to give preference 
to the Them spangs ma reading. There are only two variants which fulfill both 
criteria stated above: 

(19) - 7C, n. 28 BNQ: Ayis for kyi (DILSTP 23: Avi), where N sides with BQ, and 
DJ could well represent a better Tshal pa reading in accordance with LST. That Ayi 
in bde bar gshegs kyi ye shes yongs bkrol nas // is the right reading in this case is 
attested by the parallel in Bth: yang grol bder gshegs yeshes te :. 

(20) — 2A, n. 4 BINQLSP3: ba // for ba /, an uncommon case where the majority 
of witnesses use nyis shad after the description of the honeycomb hanging 
(‘phyang ba //) from the branch of a tree. 

Neither case is revealing, and hence we need to investigate the nature of 
the other variants. In most cases I have, after the variant, given a plausible reason 
for the editors’ alteration according to the Them spangs ma. I cannot, however, 
find any good reason why the editors of N should have emended the variants 3A, 
4B and 4C in (19) against the reading attested for J and Derge. One solution may 
be to assume that N is in fact a sister copy of J and the three readings in question 
are variants found in J and continued by D, but not found in the ’Phying ba sTag 
rtse manuscript. The emendations of 3B, 4A, 4C and 5A in (21) are not easy to 
explain either. We can generally assume that editors who had the choice between 
two alternatives tended to prefer the textually more extensive alternative, 
whenever there was no notable difference in content. For them to drop some 
words of the Buddha would have seemed worse than not doing so. This argument 
applies to 3B and 4A; for 4B, grammatical reasons must have been decisive. The 
passages in 4C and SA may simply have involved the personal taste of the editors. 

Some questions remain: Why did the editors of both Derge and N follow 
the Them spangs ma in emending gzugs to sku gzugs in the case of the Buddha 
and leave mig unchanged, whereas LST, again concerning the Buddha, 
continuously employ spyan instead? What were the guidelines for the editors of 
the two Kanjur projects that led them to decide the same questions differently? 

Through the analysis above it has become clear that the number of readings shared 
by D with the Them spangs ma line is higher than those shared by N. Both D and 
N seem to be based mainly on manuscripts related to ’Phying ba sTag rtse. 
Though this is far from being proven, three of the readings cited in (19) indicate 
that, as far as the TGS is concerned, N does not descend from J but rather is a 
sister copy of it. As to the position of D, it can only be asserted that it is not 


PDF Version: BPPB VI (2002) 

affiliated with the early Tshal pa transmission, as is otherwise the case for the 

2.3 The Them spangs ma Kanjurs 

Given the high number of variants shared only by LST, it is clear that, regarding 
the TGS, we should view these three versions as descendants of a common 
ancestor—the Them spangs ma Kanjur. Tibetan tradition has it that the Them 
spangs ma Kanjur was copied in 1431 from the Old Narthang manuscript and then 
deposited in the dPal ’khor chos at Gyantse.°* Whereas the Tokyo manuscript 
Kanjur from Gyantse is said to be a direct copy of the Them spangs ma Kanjur™ 
(which for the moment remains the only one accessible to Western scholars), we 
know that there were intermediaries in the case of the Shel dkar manuscript 
Kanjur (London) and the Stog Palace manuscript Kanjur. The Shel dkar 
manuscript Kanjur was based on the Shel dkar chos sde, a manuscript kept at Shel 
dkar in southern Tibet. That the Shel dkar chos sde was a copy of the Them 
spangs ma, however, is confirmed by a remark in the colophon of the brGyad 
stong pa volume of the Shel dkar manuscript Kanjur.* The Stog Palace 
manuscript Kanjur derives from a Bhutanese original, which could well be the 
Bhutanese copy of the Them spangs ma.» 

Most of the text-critical studies have so far proved that the Shel dkar 
manuscript Kanjur, the Stog Palace manuscript Kanjur and the Tokyo manuscript 
Kanjur agree closely with each other, and differ significantly from the Tshal pa 
Kanjurs. A further argument for the close relation between the three Kanjurs and 
the Them spangs ma is the resemblance in the arrangement of texts, which is in 
accordance with the description of that of the Them spangs ma provided in the 
gsan yig of two Tibetan masters.”° 

What, then, is to be said about the relation between L, S and T in the case 
of the TGS? Indicative common variants of S$ and T include: 

(22) — OK, n. 9 ST: sngan cad for sngan chad. 

— 4B, n. 5 BDJINQLP 1: ni for na. 

— 8C, n. 10 ST: bzhugs for zhugs. 

— 9B, n. 7, 8 ST: om. kyi chos in sangs rgyas kyi chos kyis (aberratio ocult). 

— 11H, n. 15 Py23ST: dung du yang for dung yang [P3: rung for dung]. 

—12C, n. 17 BQST: ita for bita. 

— 12D, n. 11 JST: rdzogs so; N: rdzogso; (ABQLP 23: EqN'5'), 

52 Bimer 1983a: 95, 102; Pagel and Gaffney 1996: xi. Skilling (1994: xlii-xliii) doubts that the 
Old Narthang functioned as a master copy, given the evidence from all the texts used in his study 
and from the biography (fifteenth century) of Situ Rab brtan kun bzang ’phags pa, the sponsor of the 
Them spangs ma—in all of which no such association is made. I will deal with this question later. 

3 See Eimer 1992: 179. As Skilling (1994: xliii-xlv) convincingly shows, we probably have a 
number of Them spangs ma Kanjurs that were later revised editions. The phrase “a direct copy” 
thus must be used cautiously. Eimer states that manuscripis of the Them spangs ma Kanjur were 
copied in Gyantse for more than four centuries, starting as early as 1439 and running up to the time 
of the Tokyo manuscript Kanjur, which was copied between 1858 and 1878 (Eimer 1992: 178f.; 
Saité 1977: 6 (401)). 

%4 Pagel and Gaffney 1996: ix. 

°° Skorupski 1985: xi-xii. 

°° Samten.and Skilling in Pagel and Gaffney 1996: 1. 


PDF Version: BPPB VI (2002) 

Variants shared by Z and T: 
(23)  -9A,n. 18 LT: de na for de nas. 
—10E, n.1 LT: om. % in bcos bu’i ras (verse). 
— 11B, n. 15 BINQP}23S: bar for ba in dga’ ba byed pa. 
-11C,n. 11 LT: na for du in ha dang mi rnams kyi nang du skyes so. 
— 11H, n.7 Py23T: om. pa in yin pa snyam du (L: pa inserted later in a small dbu med 

hand above the line). 
— 12B, n. 16 BNQ: de for nga; LT: om. nga in nga bzhin du (see Bth: da ltar nga). 
Variants shared by Z and S: 

(24) — OB, n. 21 BJQPj23T: sems can for sems. 

— 3B, 0. 1 BJQP2T: om. de bzhin du in de bzhin du de bzhin gshegs pa (aberratio oculi). 
The number of pairwise variants within LST is rather low, and among them 
several are clearly trivial: both variants in (24) are due to mistakes by T. In the 
first case can is erroneously added to sems, the subject being the mind (sems) of 
an arhat. The second case is a simple aberratio oculi documented also in other 
versions and reappearing in sections further down. Thus no indicative variants 
common to Z and S remain. 

In (23) the first three common variants may be called trivial and so are not 
very indicative, while the following two have most probably not arisen 
independently. Meaningful are the variants OK, 9B and 11H of (22). The reading 
rdzogs so instead of rdzogs s,o in 12D is but a standardized modem spelling in S 
and T, both manuscripts of relatively late origin. 

The number of indicative variants is too few to allow definite conclusions 
to be drawn on the relation among L, S and T. Nevertheless, the fact that LS never 
share mistakes vis-a-vis T, or, in other words, that T shares some variants with L 
and other variants with 5S, requires explanation. One possible solution is the 
assumption that 7, in fact, is the version among LST closest to the Them spangs 
ma Kanjur. The high number of single variants found in T probably has its roots in 
part in the copyist of 7; another cause may have been an erroneous master copy— 
the Them spangs ma manuscript or a (reedited) copy of it. When the Shel dkar 
chos sde, the master copy of Z, and the Bhutanese manuscript, that is, the master 
copy of the Stog Kanjur, were copied from a Them spangs ma manuscript, we 
may assume that they were at the same time checked against another version so as 
to discover possible mistakes in the Them spangs ma text. Variants shared alone 
by ST should thus constitute a reading found in the Them spangs ma, perpetuated 
by S but altered by the Shel dkar chos sde editors or the editors of L on the basis 
of another non—Them spangs ma manuscript. The same process is imaginable for 
variants of LT: in this case it could be either the editors of the Bhutanese Kanjur 
or of the Stog Palace Kanjur who collated another manuscript in order to correct 
supposed mistakes found in the text. The nature of some of the variants given 
above makes it very unlikely that such mistakes in the master copy, be it the Them 
spangs ma manuscript or a later copy of it, could have been recognized without 
the help of other manuscripts, most probably ones from outside the Them spangs 
ma tradition.*” 

57 The assumption that the editors of the Shel dkar chos sde or the Bhutanese Kanjur could have 
compared their master copy, probably a copy of the Them spangs ma manuscript, with a much less 
error-laden earlier version or even the original of the Them spangs ma when producing their new 
Kanjur is not quite plausible. Why should they then have made use of the inferior later version at 


PDF Version: BPPB VI (2002) 

Such an assumption needs to accommodate the three points below. 
Granting that the new editions basically followed the Them spangs ma 
manuscript, it should be possible to state what led the editors to adopt or reject the 
reading of a different text. 

1) The Them spangs ma reading of T is in any case clearly inferior when 
compared with the new reading adopted by the editors of both Z and S.° 

2) In the case of the combinations LT against S on the one hand and ST against L 
on the other, each side of the combination cannot afford to be obviously faulty, 
since otherwise it would be difficult to explain why the editor of S or L 
respectively did not adopt the correct reading. 

3) All readings shared by LST cannot afford to be easily recognizable as mistakes 
when compared with an alternative non—Them spangs ma variant, because it 
would otherwise be implausible that the editors of both Z and S refrained from 
emending it. 

As to the first point, concerning the variants shared by Z and S under (24), it is 
obvious that in OB LS represent the correct reading, since the subject is the mind 
(sems) of an arhat. Concerning 3B, T presents an obvious aberratio oculi which, 
if not caused later by the scribe of T, was detected by the editors of ZL and S and 

Concerning the second point, we have to account for the variants noted in 
(22) and (23). Among them we find ones which tell us more about the individual 
tastes of the scribes than about the correctness of a reading: snang cad (cad is an 
archaic spelling of ST) instead of the standard spelling snang chad (L; surprising, 
given that LZ is the older manuscript!), bzhugs for zhugs (in the case of the 
cakravartin embryo in the womb), dung du yang for dung yang, Ita ba’i ’os for 
blta ba’i ’os, rdzogs s,o for rdzogs so, de na for de nas, bcos bu ras for bcos bu’i 
ras in the verse section, dga’ bar byed pa for dga’ ba byed pa, and Iha dang mi 
rnams kyi nang na skyes so for [ha dang mi rnams kyi nang du skyes so. 

In the case of 4B (‘di ni or ‘di na), the context allows both readings 
equally well. The editors had no criteria for deciding on the correctness of the 
variants. Concerning (22) 9B, if we read the passage according to ST (which omit 
kyi chos), namely as ... nvon mongs pa’i sbubs kyi nang gi sbu gu sangs rgyas (kyi 
chos) kyis gang ste /, it is clear that the text of ST makes just as much sense and is 
in accordance with the main line of the TGS in stating that “‘... the cavity inside the 
... defilements is filled with (the qualities of) a buddha.” Thus, again, this passage 
does not violate the principle that both readings could be applied by the editors 
with equal justification. 

From the standpoint of classical Tibetan grammar, in (23) 11H, only S 
reads correctly ... yin pa snyam du, whereas T omits pa; in L pa is inserted by a 
later corrector in an dbu med hand. However, to argue that the reading without pa 
is a major mistake which an editor would immediately emend is not possible since 
we find several cases in different editions of the TGS in which the particle pa/ba 
or a particle of finality/question has not been used before snyam (see 01.7; 3B.8; 
7B.9; 8A.8). 

In 12B, however, S together with the majority of manuscripts and prints 
reads the correct nga bzhin du (also Bth: nga la bzhin no :), whereas LT simply 

58 We could alternatively assume that the inferior variant of T was introduced by the scribe of T 
after the master copy of T had already been used to establish the predecessors of Z and S. 


PDF Version: BPPB VI (2002) 

have bzhin du; BNQ contain de bzhin du. The passage without nga or de reads de 
ni ji ltar da ltar bzhin du.... (I cannot find any coherent sense in this wording). 
Here, the question why the editor of Z has not emended the text remains open. I 
have no satisfactory answer to it, 

Regarding the combination LST in relation to non—Them spangs ma readings 
(3)), we are confronted with a number of inconvenient variants:” 

(25) — 0G, n.1,2 LST: om. du nang in ... nyid du nang du yang.... 

- OL, n. 8 BDINQ: dang ngas bshad; P;: dang / ngas bshad (stock phrase!). 

— OM, n. 34 LST: de bzhin gshegs pa spyan for de bzhin gshegs pa’i mig. 

— OM, n. 37 BDINQ: de bzhin gshegs pa’i snying po can du for ... por. 

— 1B, n.2 DLSTP;: de bzhin gshegs pa’i snying po can yin. 

— 6A, n. 10 ALSTP).;: om. na in ‘jig rten na gnas pa. 

— 10D, n. 26 LSTP;: mchu dang rgyar for chu dang rgyar. 

~ 11D, n. 18 LST: om. yang in na /' yang [' ABJNQP)2;: om. /.]. 

— 12A, n. 29 LST: om. chos in chos bstan pa’i.... 

— 13, n. 4 (colophon) LST: om. skad gsar bcad kyis kyang bcos nas. 

As I have already stated above, I assume that the general tendency of the editorial 
process should be to maintain rather than to omit words when forced to select one 
of two alternatives. In the case of LST this principle does not hold true in OL, 6A, 
11D, 12A, to a certain extent in 0G, or in the colophon which, to be sure, enjoys a 
special status. I have not cited the examples where LST clearly follow this 
principle. In almost all of the six passages where LST are lacking words I cannot 
find a reason which could have kept the editors from adopting the more 
comprehensive variant. In the case of 11D, we could argue that yang was not 
adopted because the shad, which is missing in all other editions, needs to remain 
after na. The appearance of yang immediately after the shad would, of course, be 
rather uncommon. 0G, however, is a particular case, for neither of the two main 
traditions has preserved the supposed original wording there. 

The passage in 6A is rather dubious, and the fact that the variant is shared 
with A and P23 indicates that the omission of na is in fact the original reading. 
Nevertheless, it is difficult to make sense of the passage without na, so that it is 
surprising that the editors of S and L did not adopt it. It is difficult to imagine that 
they had a chance to judge the genuineness of the missing na on a basis 
comparable to ours, and so to be led to refrain from inserting it. 

The omission of the genitive particle in OM, however, could be an archaic 
element which escaped the notice of the editors since the master copy, in any case, 
already read spyan instead of mig. The spelling mchu instead of chu in 10D 
could be meant to differentiate it from chu meaning “water,” though no other 
examples of this spelling are known to me. 

I have no answer to the question why the particle can is not added by the 
editors of Z and S to de bzhin gshegs pa’i snying po at the end of 0M, whereas it 
appears in 1B, where living beings are similarly characterized as tathagatagarbha. 
However, we cannot exclude the possibility that the manuscripts used as the 
sources for the emendations introduced into ZL and S were not identical with any 
surviving manuscripts. They could well have been texts which, to a certain extent, 

°° Here I have extracted only the problematic cases from a high number of LST variants. 

© Panglung observes that the Tanjur version of the sGra sbyor bam po gnyis pa, in comparison 
with its Dunhuang fragments, inserts “particles of the genitive case, or dang, du” and thus 
“contributes to clarify the meaning of the sentences and to render the concise style of the Ta pho 
version more smoothly” (Panglung 1994: 172). 


PDF Version: BPPB VI (2002) 

were conflated with readings from the Them spangs ma line, which itself would 
not contain some of the words likewise not found in LST. 

As stated before, the number of variants regarding the relation of LST is too low to 
allow definite settlement of the question of their stemma within the TGS family of 
manuscripts. As to the stemmatic relation among LST which I have hypothesized, 
there remain the above problematic variants. Nevertheless, none of the readings in 
(25) essentially affect the content of the passages in question, nor are they easily 
recognizable as mistakes without the help of a more refined text-critical 
instrumentarium. Graphically, the imagined stemma of LST looks approximately 
like this: 

“contamination Them spangs ma 
Ne | (several 
; intermediat 
Shel dkar chos wees | pene : contamination 
——— ; 
Bhutanese Kanjur 

Stog Palace Kanjur 

r | 

Reprint of the 
Stog Palace Kanjur (5) 

The combination LST should thus represent the original Them spangs ma reading. 
If only Z and S share the same reading, then there are two ways to account for this. 
What occurs in T could be a mistake introduced by the copyist of T or some 
manuscript in the line of T after the copying of Z and S or, as the second 
alternative, the reading of T could in fact be the original reading of the Them 
spangs ma which was altered to a superior reading independently of each other by 
the editors of the Shel dkar chos.sde or Z, on the one hand, and by the editors of 
the Bhutanese Kanjur, the Stog Palace Kanjur or S, on the other. In most of the 
cases, as the nature of the variants show, the copyist of T is very likely to be the 
main source for the single readings of T. As stated above, there are many single 
variants in T resulting from inaccuracy. 

The combinations LT and ST can also be explained in two ways. If the 
reading of S against LT, and L against ST is not an erroneous single variant, it 
could well be a revised reading introduced by the editors of Z or S or one of its 
master copies on the basis of a comparison with a manuscript from outside the 
Them spangs ma tradition. When this contaminative factor was introduced into 
the separate transmission lines of L and S is impossible to say. The question with 
which line the manuscripts used for the emendation were associated is not 
answerable either. The Kanjurs of the Peking line, however, are not possible 
candidates, as the variant 11C, n. 3 proves.”! The suggestion that S is 

§! In 11C, 0.3 AJNLSTP 2; read gzi brjid against BQ: gzi brjid chen po. If, in the description of 
the beneficial result of the light of a bodhisattva touching living beings, the revisers of the master 


PDF Version: BPPB VI (2002) 

contaminated by Derge was originally brought up by Eimer (1982: 129) and 
discussed again by Orofino (1994: 34). [f this is so, contamination of S by D in the 
particular case of the TGS cannot be ruled out. But for Z or its ancestor, this 
possibility does not arise, since L came into existence about twenty years before 
the Derge Kanjur. 

My suggestion for the possible relation among LST contradicts in several 
aspects the findings of previous studies. S is usually assigned a position within the 
Them spangs ma group apart from LT. Harrison, in his study of the Drumakinnara- 
rajapariprcchasutra, rules out any contamination of S (1992a: xxv-xxvi) and 
explains the close similarity of LT as due to “the use of an intermediary copy to 
produce L and 7” (p. xxvii), The Drumakinnarardjaparipycechdsutra does not show 
any indicative variants common to S and 7 which are not also shared by L. In the 
case of the Mahasutras the situation seems to be different. Skilling characterizes S 
as “a copy of a descendent of a more perfect edition, while L(N)T are copies of a 
less perfect different edition” from a number of various Them spangs ma Kanjurs 
(1994: xliv-xlv). Finally, as mentioned above, there are the studies of Eimer and 
Orofino, which qualify S as most probably contaminated by a manuscript from the 
Tshal pa tradition.” 

2.4 The Position of Bu 

Not much can be said about the affiliation of the parts of the TGS cited by Bu ston 
in his De bzhin gshegs pa’i snying po gsal zhing mdzes par byed pa’i rgyan. The 
length of the quotations is obviously. not enough to decide whether Bu is more 
closely affiliated with one of the main groups.” Its punctuation sides sometimes 
with LST, and at other times with BDJNQ. Bu seems to have preserved a text free 
from the particular readings of the P23, Tshal pa and Them spangs ma tradition: it 
does not share the erroneous variant of BDJNQ in 1A, n. 22: ba for Ja, the 
mistakes of Pj23, and the replacement of mig with spyan and the probably non- 
original adding of can to the term de bzhin gshegs pa’i snying po (can) in 1A by 
LST.“ On the other hand, Bu introduces several unique readings which can clearly 
be characterized as recensional. They document that Bu ston had his own idea of 
how certain passages should be read in order to obtain a smoother and more 
elegant Tibetan text. The main textual “improvements” are: 
(26) — 1A Bu: skyes bu lha’i mig can zhig gis / ‘di ltar / kha dog ngan cing... 

for skyes bu lha’i mig can Ia Ia zhig gis / lha’i mig gis ‘di ltar kha dog ngan cing /....: 

Bu omits Ja Ja, which is already expressed by the following zhig “someone”; Bu 

further omits the somewhat redundant /ha’i mig gis. 

— 8B Bu: khyed bdag nyid la spro shi bar ma byed par for khyed bdag nyid sro shi bar ma 
byed par: The archaic expression sro shi ba (see Tshig mdzod s.v. sro shi ba: (rnying) 

copies of ZL and S were confronted with a choice between the two alternatives, they would most 
probably have decided for the addition of chen po. 

Silk and Skilling do not take a firm stand on the question of the position of S. Silk remarks 
that he cannot adequately explain the fact that Z, T and N sometimes disagree with S (1994: 25f.). 
Braarvig restricts his text-critical edition to a collation of S and T within the Them spangs ma 
group. : 

Further, for the passages of the TGS cited in Bu, the folios of A are missing. 

For the hypothesis that Bu ston (1290-1364) and other scholars from Zha lu may have been 
the main persons responsible for the editing process of thé Them spangs ma Kanjur (1431) see 
Harrison 1994. 


PDF Version: BPPB VI (2002) 

dpa’ zhum pa /: “despondent”), probably no longer common at that time, was changed 
to spro shi, lit. “cessation of joy”. Furthermore, bdag nyid was clearly marked as the 
object with the preposition /a. 

— Bu continuously uses rkyang shad after the address rigs kyi bu (dag). 

2.5 The Position of A 

Several studies have included the Tabo manuscripts in critical editions.” In a 
further study two manuscript folios of the sGra sbyor bam po gnyis pa have been 
interlinearly arranged within a critical edition. The evaluated material has so far 
been very limited, so that generalizations cannot readily be formulated from the 
results of these studies. In the case of the Sambandhapariksa (Tanjur), Tauscher 
could convincingly show that the Tabo version “is independent of the four major 
canonical editions and thus reflects a version of the texts prior to the oldest 
common source of these editions,” that “in many cases the T[abo] readings are 
better than the canonical ones” and thus “closer to the original Tibetan 
translation.”©’ The Tabo fragments of the sGra sbyor bam po gnyis pa, in 
comparison with its corresponding Tanjur and Dunhuang versions, turned out to 
be a less elaborate version issued some twenty or thirty years earlier. The later 
Tanjur and Dunhuang versions, on the other hand, can be said to have resulted 
from increased experience in translating Indian texts into Tibetan and thus to 
contain more refined and enlarged guidelines for translational activities than the 
earlier Tabo version. 

In the case of the TGS, is it possible to assert an independent position for 
A? Does A represent a version of the TGS closer to the original Tibetan 
translation? The second question will be answered in the following section by way 
of an evaluative study of the variants in comparison with the other versions.™ 

The unique position of A becomes obvious from a number of major 
recensional single readings listed in (31) below. Neither in punctuation™ nor in 
variants can A be said to side exclusively with one of the three other groups. A 
shares readings with each of them. Some of these readings are: 
(27) — 5B, n. 3 A: la sngon ma byung ba sngon ma byung ba’i (dittography); LST: la sngon 

ma byung ba ’am //* sngon ma byung ba’i [*L: /]. 

— 6A, n. 10 ALSTP,2;: in Jig rten na gnas pa. 

— 10D, n. 24 BINQ: pa // for pa’i // (verse). 

— 11H, n. 21 BINQ: bzhi po’o for bzhi pa’o; L: bzhi’o; ST: bzhi ’o [T: bzhi at end of line]. 

§ For instance, a part of the Paficavimsatika (de Rossi Filibeck 1994) and, for the Tanjur, a part 
of the Sambandhapariksa (Tauscher 1994). In regard to the relation between the Tabo 
Paficavimsatika and the canonical version, de Rossi Filibeck only remarks that there are no 
substantial variations. The versions just differ in regard to absence or inversion of words (p. 138). 
Further see the studies in Scherrer-Schaub and Steinkellner 1999; also Pagel 1999. 

§ See Panglung 1994. 

*’ Tauscher 1994: 181. 

6 See also the forthcoming article by the present author dealing alone with the question of the 
position of A in the stemma of the TGS. 

® A together with the Tshal pa group omits the rkyang shad against Them spangs ma and Pp in 
14 cases, together with Tshal pa and Pp against the Them spangs ma in 8 cases and together with 
Pp» against Tshal pa and Them spangs ma in only 4 cases. Tshal pa omits the rkyang shad in the 
same textual unit 10 times against A and Them spangs ma, and is joined by Po in 2 cases. 
However, A alone against all other versions omits the rkyang shad in 14 cases and exhibits a 
rkyang shad where all other versions have none in 5 cases. 


PDF Version: BPPB VI (2002) 

—12A, n. 9 LST: om. du in d-/ton du byed. 

—12A,n. 18 APi2: chu ngus for chung ngus (wrong decomposition of the abridgment 

~ 12D, n. 7 ABP}2;: yid rangs for yi rangs; Q: lacuna of one letter between yi and rangs. 

- 12D, n. 11 JST: rdzogs so for rdzogs s,0; N: rdzogso; (ABQLP)23: Eqs), 

These readings will be discussed and supplemented with more variants (including 
A) in the next section. The specification of the independent status of A is closely 
linked with the question of the originality of its transmission. 

2.6 The Relation of the Main Transmissional Groups to Each Other 

It now seems that we have four main lines of transmission for Tib: A, Po, the Tshal 

pa and the Them spangs ma lines. In order to determine how these main groups 

are related to each other, we need to take a closer look at the combinations in 

which these groups share variants. I will not list obviously erroneous variants 

represented by only one group. In parentheses behind the variant, an evaluation of 

the reading according to Ch;, Ch2, Bth or common sense will be provided 

whenever possible. 

A and Them spangs ma versus Py, and Tshal pa: 

(28)  -5B,n.3 A: la sngon ma byung ba sngon ma byung ba’i for la sngon ma byung ba’i 
(dittography); LST: la sngon ma byung ba ’am //* sngon ma byung ba’i [*L: /]. 
(ALST against Ch, and Bth (sngon ma byung ba’i gtan tshigs kyi rnam pa, Skt. 
*apirvahetvakara, appears also in Chz: #47] #5). Bth: rim kyi thog mar rgyu 
dang : rnam pa could render the same Skt. compound interpreted as a dvandva and 
with *anu-purva’ instead of a-piirva’. 

— 5C, n. 13 BDINQP 123: gyur la for gyur pa. 

—6C,n. 15 BINQP).;: gang for nang. (The pada according to ADLST seems to have 
more meaning: mi shes nang na gnas kyang rlom sems med //. That this is the original 
understanding is confirmed by Bth (dbus) and Ch; (}/S).) 

— 11B, n. 15 BINQP),3S: bar for ba in dga’ ba byed pa. 

— 11D, n. 1 BINQP)2;: bris for bres in gser gyi skud pas mig mangs ris su bres par gyur. 

—11G, n. 19 BJQP}23: om. phra rab kyi rdul (BJQP 123 is clearly an aberratio oculi). 

— 111, n. 32 BINQP)23: sngon gyi spyad for sngon ni spyad (BINQP)); less probable; Bth 
also adverbial: sngon kyang spyad....) 

—12A, n. 22 BINQP)23: om. yang. 

A and Tshal pa versus Py and Them spangs ma: 
(29) — 11H, n. 15 P).ST: dung du yang for dung yang; P3: rung du yang. 

A and Pg versus Tshal pa and Them spangs ma: 
(30) (-11H,n. 21 BJNQ: bzhi po’o for bzhi pa'o; L: bzhi’o; ST: bzhi ’o.) 
— 12D, n. 7 ABP}23: yid rangs for yi rangs; Q: lacuna of one letter between yi and rangs. 

Combinations of variants represented by a single group against all other three 
groups (variants are only those not clearly discernible as faulty): 

A versus all other three groups: 

(31) -—11C%,*” A: rab tu thob for thob. 

—11E, n. 1 A: khyim nas byung ba for mngon par byung ba. 

—11I,n. 15 AP): de’i tshe for de tshe (verse). 

—12A A: ... gang dag gis / de bzhin gshegs pa’i snying po’i chos kyi rnam grangs ’di thos 
pa dang | lung ‘bog pa dang | kha ton du byed pa dang / stond pa de dag chos kyi 
rnam grangs ‘di thos pa dang / lung ‘bog pa dang | kha ton byes pa dang / rab tu 
bshad pa dang / yi ger bris pas tshegs chu ngus de dag la chos kyang mngon sum du 
gyurd | 



PDF Version: BPPB VI (2002) 

... gang dag de bzhin gshegs pa’i snying po’i chos kyirnam grangs ‘di nyan tam / lung 
nod dam | kha ton byed dam / ston kyang rung chos kyi rnam grangs ‘di nyan pa 
dang / lung nod pa dang / kha ton byed pa dang / rab tu 'chad pa dang / yi ger ‘dri ba 
de dag la tshegs chung ngus chos de dag kyang mngon sum du ’gyur / 

— 12D, n.3 A: om. de. 

— Throughout the whole text when introducing the verses: gsungs so for bka’ stsal to. 

— Throughout the whole text (with one exception): brgya stong for ‘bum. 

— For the unique colophon of A see section 3.3. 

Po versus all other three groups: 
(32) — 6B, n. 14 Piy3: ji Ita bu for ji lta ba. 

— 6C, n. 14 BP,y3: na for ni. 

— 7A, n, 32 P43: thum po for thum bu (see Bth: dam po). 

— 10D, n. 23 Pj23: pa’i bsam for pa bsam (verse). 

— 10D, n. 25 Pj23: pa for Ja. 

— 11D, n. 11 P)23: om. des, 

— 11D, n. 12 Pj93: par for pa. 

~ 11H, n. 7 Py23T: om. pa. AJ: ba for pa. 

~— 11], n. 4 Py23: pa for na; T: ma for na. 

—11I, n. 16 Pj23: rgyal ba for rgyal dbang (P12; against Bth: rgyal dbang). 

— 111 (verse), n. 23-25 P,: de’i tshe de dag mdo ‘di thos // (also A: de’i for de; gyur nas, 
which is found in P23, was probably the reading of Pp and omitted 
by P)), 
P.: de’i tshe de dag mdo ‘di thos gyur nas /, 
P;: de tshe de dag mdo ‘di thos gyur nas // 
for de tshe de dag gis ni mdo ‘di thos // 

— 12B, n. 20 P13: pa bya ba’i for pai. 

-—12C, n. 15, 16 P23 zla nya for nyi zla; P,: zla ba nya. (P23 in accordance with Ch, and 


Tshal pa versus all other three groups: 
(33) — 10D, n. 24 BINQ: pa // for pa’i // (verse). 
— 6A, n. 10 ALSTPy)3: om. na in Jig rten na gnas pa. 
—6C, n. 11 BDJNQ: ni for gi; P,: gis for gi; Pz: migis for mig gi (BDINQ against Chp: ... 
Ste SLRS HRERER; also P1. against Ch2). 
—6C, n. 17 BDJ: spang for sbyang; Q: spang or sbang (7) (BDJ(Q) against Ch): }). 
—7A, n. 20 BDINQ: de ’i tshe de for de’i de. 
— 10D, n. 24 BINQ: pa // for pa’i // (verse). 
—11A,n. 2 ALSTP);: bsgrub for sgrub; B: sgyub. 
-—11F,n. 3 BJNQ: om. dang po (BINQ against Bth: de thog mar mngon). 
- 111, n. 13 BJNQ: por for po. 
—12A, n. 34 BJNQ: de dag for dag. 
—12B, n. 16 BNQ: de for nga; LT: om. nga (BNQLT against Bth: nga). 
—12C, n. 7 BNQP.: da for nga; J: nga or da (7) (BO)NQP? against Bth: rang). 
—12C, n. 3 BINQP;: de dag dang for de dang. 

Them spangs ma versus all other three groups: 
(34) ~11C,n.11 LT: na for du. 
—11D,n. 14 LST: ‘bum po de dag for ‘bum de dag; A: brgya stong de dag. 
—11D,n. 18 LST: om. yang 
— 10E, n. 6 LST: zhugs for bzhugs. 
—11E, n. 11 LST: /a for na in mchod rten la. 
—12A, n. 9 LST: om. du in d-/ton du byed. 
—12A, n. 29 LST: om. chos. 
— 12B, n. 16 LT: om. nga; BNQ: de for nga (BNQLT against Bth: nga). 

(35)37) show the combinations of the three groups for the portions of text 
missing in A. Again, obviously faulty variants of a single group are not provided: 

Them spangs ma versus Po and Tshal pa: 
(35) —OB, n.1 BINQP23: thos pa’i dus for thos pa dus. 


PDF Version: BPPB VI (2002) 

~ OB, n.7, 8 BIQP 23): bzangs can for bzangs tsan dan (BJQP}93 against the parallel in 

— OB, n. 21 BJQP)23T: sems can thams for sems thams (BJQP}23T wrong: “mind (sems) of 
the arhat’). 

—0C, n. 5 BQP3: chu rlung ’od srung for chu bo ’od srung; JNP12: klung for bo 
(Nadikasyapa, MVy 1050: chu klung ’od srungs; Bth: chu bo ’od srung). 

—0G, n. 1,2 LST: om. du nang in ... nyid du nang du yang...; BINQP 23: kyi nang for du 

— 0H, n. 7 BIQP}23: om. kyang. 

— OI, n. 6 LST: chung ba for chu ba; P3: cung; (LST against Bth: sbu gu for chu ba). 

— OK, n. 8 BINQP 23: sngon for sngan. 

— 1B, n.2 DLSTP3: po can yin for po yin. 

— 1C, n. 28 LST: nga for de. 

— 2A, n. 14 BDINQP a3: kun tu for shin tu. 

— 2B, n. 4 BJQP 123: om. de. 

—2C, n. 19 LST: om. rtag tu (LST against Ch): 34). 

— 4A, n. 17 BDINQP)23: om. dag. 

— 4A, n. 19 BDINLP);: brnan for mnan (also Bth: brnan): against Ch;: 24 Wi RE: 
P.: gnan for mnan; Q: pa rnan for mnan. 

~5A, n. 18 BDINQP}23: om. zi. 

— 6B, n. 6 DLST: dsal for bsil (DLST against Chj2: ¥H YA) Ps: gsil. 

—7C, n. 3 BDINQP 123: pa 7 for par. 

— 8A, n. 16 BDINQP 3: du for pa. 

— 8C, n. 16, 17 (irrelevant single variants are not shown): 
BDJNQ: nang na chos nyid mngal gnas ‘dra yod mthong // 
Pi23: nang na chos nyid mngal ba [P>2: du for ba] ‘dra yod mthong // 
LST: nang na chos nyid mngal ‘dra yod par mthong // 

— 9B, n. 7, 8 ST: om. kyi chos in sangs rgyas kyi chos kyis (aberratio oculi; ST is against 
Bth: saryas! kyi chos). 

— 10B, n. 3 BINQP)23: om. yang. 

Combination Tshal pa versus Py and Them spangs ma: 
(36) -OK,n. 6 BJNQ: gang for dag. 

— OK, n. 13 BINQ: khengs for khebs (BINQ against Ch,: $2, “cover completely”; cf. 
0G.7f. and 0J.9). 

— OK, n. 17 BJNQ: mdzod for mdzes (BINQ against Ch,: #£ 3% and Ch: SRR). 

- OL, n. 8 BDINQP;: dang ngas bshad for dang bshad. 

— 0M, n. 37 BDJNQ: snying po can du mthong for snying por mthong. 

—1A, n. 22 BDJNQ: ba for la. 

— 1B, n. 3; 5A, n. 10; 5A, n. 16: BDINQ: g.yogs for yog (P2: once g.yog; T: once yogs). 

—1C,n. 18 BDJNQ: par for pa. 

—1C,n. 22 BDJNP3S: to for te. 

— 2A, n. 11 BDJNQ: na for nas. 

— 2A, n. 23 BDINQ: par for pas (BDINQ against the parallel constructions in 2A.8 and 

—2C, n. 17 BDINQ: bas for ba (BDINQ against Bth and the Chinese; see translation). 

— 3A, n. 8 BDJNQ: nas for las (Bth: las). 

— 3A, n. 9 BDJNQ: phyung (BDINQ impossible since no agent is given). 

~3A, n. 11 BDJNQ: dang for ‘am. 

— 4A, n. 27 BDJNQ: lo Inga for Inga. 

— 4A, n. 30 BDINQP.: par for pa. 

~4C, n. 15 BDJNQ: gis (BDINQ, which mark sentient beings as the agential knowers of 
the nature of their defilements, contradict RGV 1.110). 

—7A, n. 40 BDJQ: dum bu ’di yi thum bu for dum bu thum bu; N: dum bu ’di’i thum bu. 

— 7B, n. 3 BDJNQ: rtag tu for kun tu. 

— 8A, n. 6 BDJNQ: sbrum par for sbrum mar. 

— 8A, n. 11 BDJQP;: de i for de; N: da’i for de. 

— 9A, n. 13, 14 BDINQ: gyur pa na for gyur nas. 

-10C, n. 8 BINQ: om. pa in bzhi pa ’bum. 


PDF Version: BPPB VI (2002) 

Po versus Tshal pa and Them spangs ma: 
(37)  -O0K,n. 30 BNP),3: tas for lhas (BNP}2; against Ch2: FR). 
—1C, n. 27 P23: ‘di for de. 
— 2B, n. 2 Pj23T: om. Ja in thabs la mkhas pas. 
—AC, n. 6 P13: lhas ni for dha yis. 
— 7B, n. 1 P23: pa for cing in ’os su(s) gyur cing. 
—7C, n. 14 LP 23Q: ni for na. 
— 7C, n. 23 Pi3: thar pa for thar byed (P12; faulty; construction needs transitive verb). 
— 8B, n. 11 P23: byed cig / for byed par (P23 use imperative particle cig without the 
imperative form of byed pa: byos). 
What conclusions can we draw from the material presented above? Let us first 
turn to the part of the TGS where all four groups are represented ((28)-(34)). 
There, we have a remarkable disproportion in the variants shared by two groups: 
(a) A and Them spangs ma versus Py and Tshal pa: 8 
(b) A and Tshal pa versus Py and Them spangs ma: 1 
(c) A and Po versus Tshal pa and Them spangs ma: | (and in one 
other case only accordance between A and Po) 

The high number of variants in combination (a) suggests that there may be a 
closer relation between Them spangs ma and 4 on the one side, and Py and Tshal 
pa on the other. The reason why there are so few cases in (b) and (c) can only be 
that both sides of the combination are not represented together. In other words, A 
and Them spangs ma are close, whereas A/Tshal pa and A/Phug brag have 
essentially no co-occurring variants, and Tshal pa and Phug brag are close, 
whereas Tshal pa/A and Tshal pa/Them spangs ma have not much in common. We 
should further inquire which pair of the two in (a) may be “responsible” for the 
relatively high number, that is, in which cases under (28) the reading is a shared 
mistake and can be called “unoriginal” when compared with its counterpart. 

5B looks like plain dittography in A without any parallel in Bth or Chp. 
LST have obviously tried to soften the mistake by adding the particle ‘am. If my 
above hypothesis regarding LST is true, neither the editors of S (or one of its 
predecessors) nor the editors of LZ (or one of its predecessors) have altered the 

In 6C the reading nang is confirmed by Bth and Ch2; BJNQP)23 can be 
said to be the unoriginal reading; the situation in 11G is similar (a clear aberratio 
oculi on the part of BJNQP,23), and probably again in 111, where the adverbial 
construction of sngon is attested for Bth. Nothing can be said about the other 
variants. As a result, we note that in three of the four cases where it is possible to 
decide the original reading it is the reading of the pair Tshal pa/Phug brag which 
is not original. 

If we turn to the portion of the text without the testimony of A, is there a 
similarly high ratio of non-original Tshal pa/Phug brag readings? An analysis of 
the twenty-one variants in (35) yields eight cases in which it is possible to decide 
upon the originality of the reading. Of them, Tshal pa/Phug brag can be said to be 
secondary in four cases. In the other four cases the Them spangs ma reading is 
definitely secondary. The weight of this purely statistical result becomes clear 
when we see in contrast that in (36) (Tshal pa versus Phug brag/Them spangs ma) 
Tshal pa is alone faulty in all six verifiable readings, and in the case of the 
combination Phug brag versus Tshal pa/Them spangs ma (37) Phug brag is 
secondary for all three verifiable variants. In all of these cases it is thus the single 
reading which is faulty against the reading shared by the two other groups. This 


PDF Version: BPPB VI (2002) 

demonstrates on the one hand the fact that the pairs Phug brag/Them spangs ma 
and Tshal pa/Them spangs ma are not “love marriages”; most of their common 
variants are simply due to the erroneous third group, namely Tshal pa or Phug 
brag. On the other hand, it indicates a close relation between Tshal pa and Phug 
brag, which share four of eight analyzed unoriginal readings—a fact that thus 
confirms the result of the analysis of (28)-(34). 

Based on the results above, we can now assume two principal groups of 
transmission, each sharing common variants: A/Them spangs ma and Phug 
brag/Tshal pa. But there are a few cases in which other combinations of the four 
groups are found. In (29) P)23; together with ST insert du after dung. Both readings 
are of course possible. No definite statement about the original formulation can be 
made.” In (30) the first variant suggests that the original reading was bzhi pa’o 
attested by A and Phug brag. In the Tshal pa tradition this pa turned into po, whilst 
in Them spangs ma pa was completely dropped. Both traditions thus adapted the 
cardinal number to the classical standard in accordance with their understanding 
of the passage. This could have happened independently of each other. In the 
second case of (30) we find yi rangs instead of yid rangs. I suggest a similar 
process: the original yid rangs became uncommon in classical times, and the 
editors of LST and those of Tshal pa both adapted it to the new standard. The 
Peking line still reflects this transformation, in that it has preserved the lacuna of 
the erased -d and, in B, the complete yid. 

However, there is one combination among the readings in the following 
sections (31)~(37) which calls for a closer analysis. It is the variant 
(38) 12C, n. 15, 16 P23 zla nya for nyi zla; P): zla ba nya (verse) 
in (32). 

The corresponding passages in Ch2 (3 Al) and Bth (zla ba nya) are both in 
accordance with P;23. There can be no doubt that in this case only Phug brag has 
preserved the original wording zla nya for *(pari-)purnacandra, whereas all other 
groups show the secondary variant nyi zla. The term nyi zla, Skt. candrasurya, is, 
as a matter of fact, much more appealing as an epithet for somebody who 
preserves the TGS as described in the text. This fact becomes evident when 
studying the index of, for example, the SP, where a number of compounds which 
include the term candrasirya (Tib. nyi zla) are found: Candrasiryapradipa, 
Candrastryapradiparaja, candrasuryaprabha and Candrasiryavimalaprabhasasri. 
The term (Pari-)Piirmacandra, on the contrary, appears only once in the whole SP. 
It is the name of a bodhisattva, and rendered in Tibetan as Zla gang. The situation 
is similar for the Lalitavistara, where the Tibetan version in the Derge Kanjur 
attests the combination nyi zla twenty-seven times, while zla nya or zla gang is 
found only once throughout the text. 

If we are dealing with a “closed” tradition, that is, if there are no cross- 
contaminations between the main Kanjur lines, we would have to argue that the 
widespread usage of the compound nyi zla must have been what led the editors of 
A/Them spangs ma and the Tshal pa line to change—independently of each 


Note that Pagel (1999: 199f.), for the Bodhisattvapitaka, mentions several cases where L 
shares readings with A and Tshal pa against ST. 

™ 1 use the terms “closed” and “open” recension in the sense introduced by Giorgio Pasquali. 
Closed recensions suggest that readings move only vertically, from the master copy to the copy 
itself, whereas in an open recension readings can also circulate horizontally, leading to 


PDF Version: BPPB VI (2002) 

other—the original z/a nya (including reversing the word order) to nyi zla, the 
form abundantly attested in the SP (see above). Po, however, has preserved the 
original uncommon wording, and deserves respect as a more conservative line of 
transmission, one which, according to the results of my analysis above, introduces 
more mistakes because of inattentive copying than such redactional intervention 
as the alteration of z/a nya to nyi zla in the other Kanjurs. However, it is clear that 
one can by no means rule out cross-contamination even at an early stage of 
transmission.”” The more convincing explanation of the nyi zla variant shared by 
three groups which generally share no common variants in the TGS, rather, is to 
assume contamination. The assumption of such a contaminative process in an 
open tradition can account for why only Phug brag has preserved the original 
reading zla nya. We will come back to this question after dealing with the 
recensional single variants of A quoted under (31). Only such an evaluation can 
help to shed light on the question whether A’s single variants can be treated as 

In 11C “ and “”, A reads rab tu thob instead of thob in the context of the 
attainment of the anutpattikadharmaksanti and the attainment of efficacious 
formulas (dharani). Usually forms or nominalizations of the verb pratilabhate are 
used in this context.” Though rab tu cannot completely be objected to, the far 
more usual rendering of the prefix prati in Tibetan is yongs su. If we assume that 
in the Sanskrit of the passage in the TGS we originally had a form of pratilabhate, 
as attested in, for example, the Saddharmapundarika, we are faced with the 
question how the reading rab tu thob came into existence. Not irrelevant is 
Braarvig’s (1993a: ix) noting the reading rab tu thob instead of thob for the Stog 
and the Tokyo manuscripts of the Aksayamatinirdesasutra. These are the 
manuscripts closest to the Dunhuang version. 

In a unique case (11E, n. 1) A reads khyim nas byung ba instead of mngon 
par byung ba in all other canonical versions. Bth simply has byung ba. The 
passage describes a bodhisattva from whose body rays of light are emitted from 
the time he “had been born, set out for ascetic life (abhiniskramana) and [finally] 
completely awakened to buddhahood.” Though the translation of abhiniskramana 
in A is clear enough, the usual rendering of abhiniskramana in Tibetan translation 
literature is doubtlessly mngon par byung ba, the reading attested in the canonical 
versions.“ I have no parallels of this rendering (khyim nas byung ba) in other old 

A major variant is found in passage 12A.” The Buddha explains to 
Vajramati that “sons and daughters of good family who are restricted by obstacles 
[caused by their] deeds (karmdvarana) will become purified” owing to the merit 
attained from propagating the siitra. Apart from the divergent translations of 
individual words, A offers an overall different understanding, in that it ends the 

” For a similar case, see Braarvig’s edition of the Bodhisattvacaryanirdesa (1994), where he 
assumes the two Phug brag versions to be “direct copies of the Old Narthang, notwithstanding all 
errors” (p. 139). The main reason for his assumption is the correct reading rab instead of the rang 
of all other versions in § 15, n. 24. In this case it can be stated with almost one hundred percent 
certainty that the alteration of rab to rang did not take place independently in the other two main 
lines of transmission. 

73 See e.g. SP 327.8: ... kofinayutasatasahasraparivartaya dharanyah pratilambho ‘bhut /. 

74 See e.g. SP;s.v. abhinis- vkram. 

% This variant is discussed in detail in my forthcoming publication. 


PDF Version: BPPB VI (2002) 

second unit of the enumeration of activities with ... pas, indicating that this 
enumeration contains the means by which to reach the desired result. One 
problematic feature in A is that gang dag gis, functioning as a relative pronoun, 
seems to be taken up twice by de dag and by de dag la. If we decide that the 
relative clause refers to de dag la as it does in Tib, de dag cannot be included in 
the structure of the sentence. One possible translation of A is: 
Vajramati, for those sons and daughters of good family restricted by obstacles 
[caused by their] deeds who listen to this Dharma discourse [called] 
Tathagatagarbha and expose, recite or teach [it]—by_virtue of listening to this 
Dharma discourse and exposing, reciting, explaining and copying [it, it will 
happen that they] will easily see the Dharma before their eyes [and they will] 
become purified [from] the obstacles [caused by their] deeds. 

A is more explicit in stating that the activities performed are the means of getting 
rid of the obstacles—an idea not openly expressed in 7ib. Further, the particle de 
dag, redundant in A, in Tib refers to chos. In Bth, de dag (gis) is found at the end 
of the first enumeration (and not, as in Tib, after chos), so that, grammatically, 
there is no relative clause. De dag la does not appear. However, bris nas should 
most probably be emended to bris pas, since it would be grammatically impossible 
to combine the former with the nominalized stems connected by dang before. Bth 
offers no indication that the employment of the particle ’am, as seen in Tib, might 
be based on a Sanskrit original. It would be hard to explain why a reviser should 
alter the use of dang and ’am in favor of the solution seen in A (dang throughout). 
Regarding Ch2, Amoghavajra, by using the particle you fH (syntactically 
governing the sentence up to ... = #34 #8), marks the second enumeration clearly 
as the necessary condition for the result mentioned at the end of the sentence: “... 
in consequence of listening ...” and thus seems to support the reading found in A. 
As for originality, it is hardly possible to settle with certainty the question whether 
Tib in this passage is a revised version derived from the text as found in A, though 
there is some evidence which seems to suggest this. The explicit characterization 
of the activities as the means in A—which is also found in the Chinese translation 
and (with the emendation of nas to pas) in Bth—would favor such an explanation. 
The redactors of Tib would afterwards have changed the position of the irritating 
de dag (still found in A), combining it with chos. Bth for its part agrees with A in 
uniformly using the particle dang in both enumerations instead of ‘am, which is 
found in Tib. 

Throughout the whole text (with one exception) 4 has brgya stong instead 
of the ‘bum of the canonical versions. The Sanskrit in all these cases was most 
probably satasahasra. brGya stong sticks literally to the Sanskrit. In §19 of the 
sGra sbyor bam po gnyis pa it is laid down that in order to write “good” Tibetan, 
numbers should be translated in a Tibetan way.’° brGya stong for ’bum is attested 
for Bth throughout the text, and in Braarvig 1993a: vi for the Stog and the Tokyo 
manuscripts (which are the manuscripts closest to the Dunhuang material). 

When introducing the verses spoken by the Buddha, A continually 
employs gsungs so instead of bka’ stsal to. Here again, in all these cases Bth reads 
gsungs so too. 

As indicated above, an evaluation of these single variants found in the Tabo 
fragments is vital for establishing a hypothetical stemma of Tib. Basically, there 

78 See Simonsson 1957: 254f. 


PDF Version: BPPB VI (2002) 

are two ways to explain the unique readings of A. They could (a) represent the 
original phrases—ideally, chosen by the translation team itself or otherwise 
mirroring a very old stage of transmission——whereas all other versions must have 
undergone a process of revision in these cases. Or they could (b) result from later 
intervention by a reviser based on his own judgment or on the evidence of 
available Sanskrit manuscripts. That Sanskrit manuscripts were used to revise 
Tibetan translations in Tabo is known to be true of the Paficakrama.” Harrison in 
his analysis of several Tabo fragments of the mdo mang section does not seem to 
exclude this possibility either.” However, I cannot imagine that a later editor 
would have had any good reason to alter the readings ’bum, bka’ stsal to, and 
mngon par byung ba of Tib into the much less commonly used phrases found in A. 
In addition, what would have led him to replace thob rendering a form of 
pratilabhate with rab tu thob? If he indeed had had access to a Sanskrit version of 
the TGS, should we not rather expect that he would have supplied the particle 
yongs su as the “regular” counterpart to prati? Finally, why should he have left 4’s 
dittography shared with Them spangs ma in 5B (28) uncorrected? If, on the other 
hand, we assume that a reviser changed the text in A without resorting to a Sanskrit 
manuscript, how could he have known about the construction of the passage of 
major variants in 12A which, after all, seem to be confirmed by Bth and the 
Chinese?” I think that the evidence should lead us to assume that the single 
readings of A discussed above in fact document an older stage of transmission 
which escaped later revisional intervention. That A too suffered alteration, 
however, is clear from the mistake nyi z/a instead of zla nya (32), which A shares 
with the Them spangs ma and Tshal pa block prints and manuscripts. 

Let us now come back to the question where we can reckon with major 
contaminations without which, as seen above, it seems impossible to establish an 
appropriate stemma for Jib. There are basically two alternatives how major 
contaminations could have taken place: 

a. Tshal pa and Phug brag derive from a common hyparchetype different from 
the hyparcheype of A/Them spangs ma. This explains the common variants 
between Tshal pa and Phug brag. Later, the common ancestor of A and the 
Them spangs ma line was contaminated by a manuscript of which the Tshal 
pa manuscript is a descendant. This led to the introduction of readings found 
in the Tshal pa line into the A/Them spangs ma tradition.°? Phug brag 
remained uninfluenced by these Tshal pa readings, and so preserved parts of 
the original text such as zla nya discussed above. For editors within the 

7” See Tomabechi 1999: 88. 

78 See Harrison 1999: 53. 

” For a discussion and evaluation of the colophon of A see section 3.3. 

8 Also the reverse order is thinkable, i.e., a contamination of the Tshal pa tradition by a 
manuscript leading to the Them spangs ma manuscript. The model suggested in the main text sees 
the unoriginal nyi zla variant as having originated in one of the manuscripts before the Tshal pa 
manuscript, after the Phug brag branch had already become established. This apparently fitting, 
though unoriginal, variant was “exported” to Them spangs ma/A by contamination. The omission 
of phra rab kyi rdul ((28): 11G, n. 19), shared by Phug brag and Tshal pa, on the other hand, was 
not adopted by the editors of the archetype of Them spangs ma/A because it could easily be 
recognized as a mistake on the basis of their own manuscripts. The model with the reverse order, 
on the other hand, would imply that the myi zla variant appeared first in a manuscript leading to the 
predecessor of A and the Them spangs ma, and from there was introduced into a pre~-Tshal pa 


PDF Version: BPPB VI (2002) 

A/Them spangs ma tradition who were confronted with the choice between 
zla nya and nyi zla, the common nyi zla was the variant they preferred. A 
second major contamination has to be assumed in order to explain why A 
has preserved many old readings, whereas the Them spangs ma 
representatives, while sharing the same predecessor with A, do not. I 
suppose that at a time when A or its predecessor(s) had already become 
“separated” from the pre-Them spangs ma line another manuscript of the 
Tshal pa line or a pre-Tshal pa manuscript was used in editorial activities 
that eventually led to establishing the Them spangs ma copy (and thus was 
able to contaminate that transmission in a second “thrust”). The 
corresponding stemma is shown as “Alternative 1” (in 2.7). 

b. The second alternative is based on the assumption that Phug brag forms an 
independent line of transmission, as shown in the stemma of “Alternative 2” 
(of 2.7). Thus it preserved original readings which have been altered in the 
second line of transmission, to which all other collated versions of Tib 
would then have to belong. The readings shared with the Tshal pa line, 
however, would be the result of Phug brag’s contamination with a 
manuscript of the Tshal pa line.*’ Considering the numerous common 
readings between Phug brag and Tshal pa among which, as documented 
above, a high percentage of unoriginal variants can be found, this 
contamination must have been rather strong. It is difficult to answer, under 
this second alternative (b.), why, in view of such a strong contamination, the 
editors of the Phug brag archetype decided to preserve their reading zla nya 
against the more common reading nyi zla contained in the contaminator, and 
why, on the other hand, they opted for the unjustified omission of phra rab 
kyi rdul in 11G. Under this alternative, too, the (pre—)Them spangs ma line 
must have been contaminated by a manuscript of the Tshal pa or pre—-Tshal 
pa line. Only then can the fact that A has preserved original readings found 
in no other manuscript or block print be reasonably accounted for. 

There are some other variants which are in need of explanation. 

(39) BQ: brjid chen po dang for brjid dang. 

Only BQ seem to have preserved the reading corresponding to Ch; (AEX) for 
Skt. mahaujas (gzi brjid chen po). But there are several reasons which make such 
an interpretation hardly allowable. First of all, it is very unlikely that all the other 
three groups (A, Them spangs ma and ’Phying ba sTag rtse) have omitted chen po 
independently of each other. Secondly, Bth also only reads mdangs, another 
equivalent for ojas according to the MVy (6409). Finally, Ch2, in this passage, 
clearly seeks to preserve a four-syllable rhythm AR, tHE, Bo, Bo 
#@ Bf, BS, B87. ), which may have been the reason why X was 
added to Vi. 

(40) 11], n. 20 ABJNQLSTP),: bstan las for bstan la (verse). 

The corresponding prose bcom Idan ‘das de’i bstan pa la in 11H.10: “under the 
rule of that exalted one” documents that the pada should be read according to P3: 

8! Nothing can be said about when the contamination took place. The variants shared by Tshal 
pa and Phug brag in the TGS do not show any prevalence on the part of one of the two branches of 
Tshal pa, viz., the "Phying ba sTag rtse or the Peking line, so that either of them could have 
contaminated Phug brag. 


PDF Version: BPPB VI (2002) 

rgyal de’i bstan la. This is also in accordance with Ch) which shows the same 
understanding of the verse (E(k (#sHF), and in the prose ({R#S{HEF). It is, of 
course, possible that only P; has preserved the original Ja, especially as the 
conservative character of Phug brag in (38) has already been demonstrated. 
However, an alternative is to assume that P; has secondarily adopted the reading 
la according to the parallel in the prose portion, only after the mistake in the early 
transmission was made of associating bstan la with the following mdo sde ’di and 
hence of “emending” Ja to las. 

With this variant we enter the last category where a reading shared by all versions 

is suspected of contradicting the original wording of the text. 

(41) (— 5A, n. 25 ABDJINQP).;: snying po; LST: snying po’i. Though the context makes it 
clear that the Indian original must have read jfidna (ye shes; see the note in the 
translation), all versions of Tib have snying po(’i) (Skt. garbha). We cannot be sure if 
garbha was already part of the Indian manuscript on which 7ib was based or if we are 
confronted with a mistake, or even a deliberate alteration, on the part of the 

—9A, n. 27, 28 BDINQLSTP:: phyir rim for phyi rol; P23: phyi rim, see Bth: phyirol. 

— 10E, n. 2: All versions: brgya; probably originally brgyan. 

In 5A we have no reason to assume that the Tibetan transmission is guilty of the 

apparently wrong reading, thus rendering the passage irrelevant for an analysis of 

the relation among the Tibetan versions, but in 9A the graphical similarity 

between rol and rim is striking, and the assumption that the mistake is due to a 

very early mistake in the transmission of Tib is most plausible. The emendation to 

phyi rol is supported by the frequent occurrences of the term in the last simile 

(9A.5, 9B.2, 9B.10, 9.1, 9.3) and by Bth, which reads phyirol. Concerning the 

content of the simile itself, it is clear that the reading (phyir) rim gyi(s) 

(“gradually”) would conflict with the following passage, which states that the 

figures become clean “in that moment.” 

I am not sure whether brgyan was the original reading in 10E against 
brgya of all attested versions of Tib. The Chinese versions do not mention 
“hundred”; Bth, too, reads brgyan. For the consequences of such an emendation 
for one’s understanding of the passage, see the note in my translation. However, 
the change from brgyan to brgya could be explained as a mistake which was 
clearly part of the Tibetan transmission. We find further the following unique 
readings in Tib: 

(42) — The sequence of the padas within a verse is changed and arranged according to the 

principle that the governing verb is usually placed at the end of the section governed. 
Padas which contain relative clauses or other specifying elements are placed before the 
element specified. On the one hand, this leads to a reduced scope of interpretative 
ambiguity and follows basic rules of Tibetan grammar. On the other hand, the idea 
expressed in the original obviously became distorted In some instances: 
¢ In 3.3 the main verb ston to and its object chos are shifted from pada c to the end of 
pada d. The particle of finality phyir right before chos was in the process, as a part of 
the second half of the pada, also moved to d. The last pada, stating that living beings 
quickly awake to buddhahood, thus becomes part of the finality clause with phyir, 
which it was not in the original, as documented by Bth and Ch). 
* Regarding 10.2/10.3, see my comment in the translation. 
— At the end of 1A Jib reads yang dag pa nyid du, *samyaktve. That the original wording 
must have been a form of parisuddha is attested to by Bth (yongsu dag par) and Ch; 
(YB). The transformation from yongsu dag pa to yang dag pa carreasily be explained 
as a result of the graphical similarity between yang and yongsu, but the addition of nyid 
is clearly a recensional element of the Tibetan translators/editors. 


PDF Version: BPPB VI (2002) 

— In 1.2d the Skt. could have been sambhavati. A Tibetan equivalent for this is found in 
Bth: kun du ’byung. Tib, however, has only ‘gyur. Kun tu appears before, separated 
from it by some syllables, and has to be integrated differently into the sentence. I 
assume that an original reading kun tu ’byung was turned erroneously into kun tu ‘gyur 
(both are often variae lectiones for each other). Then, in a redactional step that bypassed 
the Sanskrit, the uncommon kun tu ’gyur was separated into kun tu and ’gyur. 

— In 8B.4 only 7%b features rigs for Skt. gotra or kula. The other versions of TGS and the 
quotation in the RGVV have dhatu (see the note in the translation). Here the reason is 
probably an underlying faulty Indian manuscript (gotra and dhatu are metrically 
equivalent), or an either deliberate or accidental alteration by the translators of Tib. 

2.7 Possible Stemmas of Tib 

It is evident that Tshal pa and Phug brag share an intermediary copy or that one of 
the two must have been subject to contamination by the other, given the fact that 
many of their shared readings are not original. But in the case of A and Them 
spangs ma it is virtually just once that a reading shared only by both can be said 
not to be original, namely the variant 5B, n. 3 (28). Other readings in (28) are 
mistakes shared by Tshal pa and Phug brag (3 cases) and thus do not bear on the 
relation between A and Them spangs ma, or are readings of a rather trivial nature 
(pa versus la, ba versus bar, bres versus bris, yang versus omission of yang). 
Variant 5B n. 3 is, no doubt, a characteristic one and cannot be downplayed as 
mere coincidence. In fact, if this variant did not exist, I would be reluctant to call 
A and the Them spangs ma descendants of a common intermediary copy. 

As mentioned above, a crucial point in reconstructing the stemma is 
precisely the evaluation of this variant 5B n. 3. I have assumed that the 
dittography crept into the hyparchetype of A and Them spangs ma. The two 
proposed stemmas are based on this supposition. Alternatively, the dittography 
could have been part of an earlier transmission before A became separated from 
Tshal pa and Them spangs ma. A would then not necessarily derive from the same 
hyparchetype as Them spangs ma, and stemmatic relations could again be 
arranged differently.* 

In the stemmas on the next pages, I will try to summarize the results of the 
analysis above. As has become clear, we are dealing with highly hypothetical 
configurations. This is due to, generally, our imperfect knowledge of how editors 
and compilers of new manuscripts worked and what degree of horizontal 
transmission we need to assume. For the 7GS in particular, the variant situation is 
rather restricted, since we are dealing with a relatively homogeneous transmission 
with mostly transmissional variants. The stemmas are therefore tentative, and it is 
quite possible, if other old 7ib material appears, that they may have to be 
readjusted, The siitra is too short, and so does not allow for definite conclusions as 
to the stemmatic structure in all instances. For the sake of clarity, I have decided 
to give two alternative stemmas, of which, based on the analysis of the variants, 
Alternative 1 is slightly more likely to conform to the actual development. The 
two alternatives vary in the relation shown between Phug brag and Tshal pa. 

®2 The assumption that an easily committed mistake shared by certain lines of transmission does 
not mean that the lines in question also share the same hyparchetype but that the mistake could be 
a remnant of an earlier transmissional state which was not emended in the lines in question 
violates, however, the classical rules of stemmatological reconstruction and would render the 
situation hopeless. I therefore refrain from suggesting possible stemmas for such a scenario. 


PDF Version: BPPB VI (2002) 

Accordingly, the process of contamination, which surely must have already been 
taking place in that early phase, is indicated differently. Again depending on the 
alternative basic structure, the position of the Old Narthang manuscript of the TGS 
may be that of the hyparchetype of Phug brag and Tshal pa (Alternative 1)—a 
possibility which has hitherto been ruled out in almost all text-critical studies. In 
neither of the alternatives could Them spangs ma directly derive from the Old 
Narthang manuscript, given that A is in fact of earlier origin than the beginning of 
the fourteenth century, the period when the Old Narthang collection is said to 
have come into existence.™ The exact date of A, however, cannot be known with 
certainty, but it is to be hoped that further progress in the field of codicology® 
will allow for a more detailed periodization of Kanjur material. Concerning the 
recensional variants in A, they may in fact be of archaic origin, and so reflect the 
original wording of the Tibetan translation. 

Other, more significant variants are definitely needed to prove that N 
descends directly from the ’Phying ba sTag rtse and not from Lithang. Still open, 
of course, is the question whether B is based on the Yongle or the Wanli block 

Finally, something should be said about the high number of 
contaminations underlying my hypothetical stemmas. As a rule, I have tried to 
apply the classical text-critical methods associated with scholars like Karl 
Lachmann or his twentieth-century successor Paul Maas, wherever possible. 
These scholars had mainly non-contaminated traditions in mind when they 
formulated their basic tools of analysis. Their methods alone, however, in the case 
of the TGS, obviously cannot lead to a satisfactory explanation of the variants. 
Transmissional reality may in many regards have been even more complex than 
my purely hypothetical stemmas can display. Contaminations seem to be an 
integral part of this reality, and this apparently applies to the history of 
transmission of 7ib from earliest times on. At least for the history of the canonical 
translation of the TGS, it is now clear that we have an utterly open transmission 
before us. It cannot be excluded that the validity of this result may be upheld by 
further Kanjur studies in the future.*° 

83 Yarrison mentions the possibility that Phug brag manuscripts could be based on one of the 
two “edited versions of the Old Snar thang,” i.e., Tshal pa and Them spangs ma, copies of which 
are known to have been distributed throughout Tibet (1994: 308). 

* This does not rule out that Them spangs ma may have undergone contamination by Old 
Narthang—a realistic possibility under Alternative 1. 

8 See Scherrer-Schaub 1999b. 

86 For the Greek and Roman manuscript traditions, Pasquali has “demonstrated that 
contaminated traditions are more the norm than the exception” (Tarrant 1995: 109). 


PDF Version: BPPB VI (2002) 

Stemma of 7ib (Alternative 1) 

Translation from Sanskrit (ca 800) 

Tabo (A) 
(11/12 century?) 

v.. contamination a. 
Them spangs ma 

contamination”, (1431) ms at contamination w Po 
“ (several : Le 
intermediate a Old 
Shel dkar chossde copies?) ™., contamination Narthang? / 
contamination . Py ? 
Nae /\ 
Bhutanese Tshal pa P, P, P; 
London copy (Z) Kanjur (1347-49) Phug brag 
(1712) | (1696-1706) 
Tokyo Stog Palace Kanjur 
Ms Kanjur (7) (ca 1700-50) 
Reprint of the 
Stog Palace Kanjur (5) 
Yongle 1410 
Wanli 1606 Be 
*Phying ba sTag rtse | Berlin 
contamination with Kangxi 1684-92 Ms Kanjur (8) 
4°"Them spangs ma ..p | (1680) 
Lithang (VJ) \ contamination Kangxi 1700 
contamination (1608-21) with J? 

with [Ho rdzong™... Kangxi 1717-20 (Q) 
(Them spangs ma line) “@ ¥ Narthang (4) | 
a (1730-32) Qianlong 1737 
Derge (D) 
(1733) Cone (1721-31) 


PDF Version: BPPB VI (2002) 

Stemma of Tib (Alternative 2) 

Translation from Sanskrit (ca 800) 

Tabo (A) 
(11/12" century?) 

Them spangs ma wy. contamination 

contamination. (1431) - (Po) 
4 (several # 
intermediate contamination .“ 
Shel dkar chds sde copies?) # i 

contamination Py ? 
Bhutanese Tshal pa P, P, P; 
London copy (Z) Kanjur (1347-49) Phug brag 
(1712) | (1696-1706) 
Tokyo Stog Palace Kanjur 
Ms Kanjur (T) (ca 1700-50) 
Reprint of the 
Stog Palace Kanjur (5) 
Yongle 1410 
Wanli te. 
’*Phying ba sTag rtse | Berlin 
contamination with Kangxi 1684-92 Ms Kanjur (B) 
“Them spangs ma ...y | (1680) 
Lithang (J). \ nn contamination Kangxi 1700 
contamination (1608-21) with J? 

with [Ho rdzong ~.._ Kangxi 1717-20 (Q) 
(Them spangs ma line) “@ ” Narthang (V) | 
ee (1730-32) Qianlong 1737 
Derge (D) 
(1733) Cone (1721-31) 


PDF Version: BPPB VI (2002) 

3 Characteristics of the Textual Witnesses of 7ib 
3.1 Archaic Features 

In the following I list the archaic features of the editions of Tib, though I am not 
sure that all of them can strictly be called so. They may just reflect preferences of 
a copyist, regional characteristics or peculiarities of a certain (but not necessarily 
ancient) period. This field of research deserves systematic elaboration in the 
a. The syllable-forming genitive particle yi after syllables ending in vowels 
appears in several cases as i in the verses: 

P;: 19 times; P;P2: 17 times; B: 9 times; A, L: 4 times; S: 2 times; 

J, T: once; Q: yi compressed, most probably altered from ’i: 9 times. 
Simonsson characterizes this feature as archaic, and states that it is especially 
common in the old translation of the SP (1957: 22). 

b. The complete omission of the genitive particle is also said to be characteristic 
of the archaic style (Panglung 1994: 172). The particle is omitted in 

P;, Pz: 17 times; P3: 12 times; L: 7 times; T: 6 times; J: 2 times; 

ABJNQS: once. 

c. Reversed gi gu: Besides A only 2 times in J. 
d. The mtha’ rten ‘a (mdo’, ‘dra’ dpe’, g.yo’): A: 8 times; P;: 5 times; B: 2 times; 

JP} P): once. 

e. sTsogs for sogs and bstsal for bsal: P;P2: 5 times; A: once. 
ff Ci nas for ji nas: LST: once. 
g. bCu instead of cu after drug or sum: JP }23: 3 times; Q: 3 times lacuna of one 
h. Shes instead of zhes after syllables ending in -s: 
LT: 8 times; P;P2S: 7 times; BJ: 6 times; Q: 5 times; A: 4 times; 
P3: 2 times; N: once, *® 
i. Du instead of tu after kun, shin and rol: J: 14 times; Q: 2 times; BLS: once. 
j. Du instead of tu after syllables ending in -g: P3: 6 times; ZL: once. 
k. The spelling sngan cad instead of sngan chad: ST : once. 
l. The spelling gzer instead of zer: P;: 15 times; A: 14 times; P2: 13 times; 

S: 4 times; L: 2 times. 

m. Among the few transliterations a long vowel is usually not marked by P)2(3), L 
or A (and in some cases by NST). P; seems to have been checked and corrected 
after copying. O’u for au appears twice in LSTP 23." 

P;, P; and—though it shows traces of a standardizing revision—-P; contain 
especially abundant traits which are known to be archaic. The figures for A are 
misleading because it covers only 40 percent of the siitra and happens not to 
provide testimony for most of the verse sections. This diminishes, in particular, 
the number of cases of A among a., e. and g. However, the paucity of omissions of 
the genitive particle (b.) and the non-existence of the forms kun du and shin du (i.) 
in A are surprising. I am rather sceptical that the forms kun du and shin du 

87 T shall not repeat the archaic features of A described in section C.1. There, in fn. 2, references to 
some important previous studies of archaic features are provided. 

88 According to the Tshig mdzod the modern standard is zhes after syllables ending in -s (s.v. 

® See Skilling (1994: lii). 


PDF Version: BPPB VI (2002) 

represent an archaic spelling. It is quite possible that both forms, namely du and 
tu, were used far into the eighteenth century, when tu became the standard. In J/ it 
may well have been the individual taste of the copyist or a local dialect that led to 
the preference for kun du and shin du. 

D, however, exemplifies modern standard orthography in every respect, 
followed by N, where the features listed above are very rare. By contrast, non- 
standardized elements are frequent in S—which documents the scribe’s high 
degree of fidelity towards his master copy. 

It is, of course, true that a large number of ancient features in a text does 
not necessarily guarantee that the text itself has been transmitted faithfully. An 
early redactor could have easily “emended” the wording significantly to suit his 
ideas of what the Buddha’s word should be, while at the same time leaving the 
orthography untouched. It is a well-known fact in the realm of textual criticism 
that the manuscript orthographically revised to the latest standard can nevertheless 
be the document with the highest degree of fidelity to the original wording. The 
TGS is one of the cases where the stemma does not correspond to the level of 
orthographic standardization in all regards. A is definitely a very important piece 
of testimony for the oldest stage of transmission. For P;2; the situation is different: 
together with (or contaminated by?) Tshal pa, they underwent heavy revision and 
further introduced a number of mistakes on their own. The Phug brag versions 
nevertheless display the most archaic features alongside A. This persistent 
tendency towards preservation, once the major revision was completed, has 
already been noted in regard to the variant 12C, n. 15, 16 (see (38)), where P12; 
has preserved the correct reading in the face of all other versions, including A. 

The representatives of the Them spangs ma line also exhibit a moderate 
number of archaic features. This corresponds to their position within the stemma 
as versions which have undergone a less intensive revision than the Tshal pa 
manuscripts and prints. 

3.2 Irregular Verbal Forms 

Current knowledge of the ancient Tibetan verb system is still far from perfect. 
Many of my attempts to explain the irregularities found in the texts must therefore 
remain hypothetical. Verbal irregularities include:”” 

¢ bKris instead of dkris: The perfect form of the verb dkri ba is several times 
spelled bkris (B: 7; Pz: 5; Py: 2; LNQP3: 1). In view of the verb pair dkri 
(transitive; “to wind, to wrap”) — ‘khri (intransitive), the form bkris could be 
old. The prefix d- in d-kri could be explained as the prefix of the present form. 
Alternatively, the spelling bkris could be an erroneous adoption of the common 
prefix b- in the perfect form. 

* bZag (BQ in 9A, n. 8) instead of zags: Zags is the perfect form of the verb 
‘dzag pa, “to drop, to drip”. bZag looks like a transitive perfect form from 
‘dzag. No other occurrences of this form are known to me. 

*In 7A, n. 28 and ™ the forms ’goms and (AP)2Q:) bgoms appear twice. The verb 
*gom pa with the perfect form bgoms is noted in the Tshig mdzod as meaning 
“to pass over, to go by taking a step downwards” (steng nas gom pa brgyab ste 

® T have only quoted verb forms which cannot easily be explained as mistakes caused by a 
careless copyist. 


PDF Version: BPPB VI (2002) 

ba’am ’gro ba /). This meaning fits the context of the section. However, a 
perfect form bgoms from ’goms in the present is uncommon, ’g- at the 
beginning of the present form turning normally into bk- in the perfect form. Ja 
notes the verb bgom pa, “to walk, to step,” which I take as a denominative 
formation from gom pa (“a pace, step”). The form bgoms would thus be the 
regular perfect form of the verb bgom pa. Should we then assume a secondary 
form ’gom/’goms supplementing b-gom/b-goms? 

bsGo in 4B and 7A should be the “future” form of the verb sgo ba. The RGV 
has present optative forms in the parallel verses. Tibetan could render an 
optative in Sanskrit with bsgo, the so-called “future” form. bsGo could, 
however, also be the present form generally applied at the end of the similes. 
We may thus assume that sgo, the form under which the verb is usually given 
in dictionaries, is just a secondary spelling or an abstraction. bsGo can also be 
found as the perfect form of sgo ba in several glossaries and dictionaries (for 
references see my note to 4B in the translation). 

bsKyed instead of skyed: In 8C, n. 25 the imperative form of the verb skyed pa, 
“to generate,” is spelled bskyed in LST. I am not aware of any historical 
imperative form with the prefix b- and so tend to regard this form as a 

In 9A n.29, 9B “, 9C ™ and in 9.4 the verb ’gogs pa (in L and J once ’gog pa) 
or its perfect form bkogs (DJN: once bskogs; P3: once bgogs) appears. ’Gogs 
could be a variant spelling of ’gog (perfect form: bkog), for which Jd gives the 
meaning “to tear away, to peel....” In the TGS ’gogs must have the meaning “to 
hack off [the outer layer of clay].” It can therefore hardly correspond to Ja: 

’gogs pa, “another form for ’gegs pa, to prevent, to avert...” bGogs of P; can 
only be a mistake (see the discussion of bgoms above), but the form bskogs of 
DJN seems to derive from another verb: Ja gives the meaning “to splinter off” 
for the verb kog pa (unlikely) for which he also mentions a secondary form 
skog pa. bsKogs could thus well be the perfect form of skog. 

Besides the regular thob pa, “to attain,” there appear the forms ’thob (LN) and 
mthob (S) in 5C, n. 14. The form ’thob may be an adaptation to the verb classes 
which have the ‘a sngon ‘jug only in the present form. 

In 4B and 4C all versions feature the verbal form non pa, “to cover, to 
suppress.” In 4A, n. 19, on the other hand, besides the faulty brnan (against 
Ch2 but with Bth), ST read mnan and P2 reads gnan. Assuming a regular 
present form gnon pa with its perfect form mnan(d), the form non(d) in 4B and 
4C could be an old perfect form possibly derived from the same form non(d) 
used as an imperative. gNan(d) (P2), on the other hand, is probably a secondary 
formation of the perfect form from a time when the g- of gnon was no longer 
thought of as a prefix, but as part of the verb stem. 

The verb g.yogs, “to cover,” appears 7 times throughout the siitra in 
syntactically similar constructions (1B, n.3; 5A, n. 10; 5A, n. 16; 9A K-9.1; 9.2; 
9.3). In none of these cases do we have the testimony of A. In the first three 
instances LSTP}23; read yog instead; thrice we find the form g.yog (P;: twice; 
P2: once); T reads yogs in 1B, n. 3; Bth displays the two forms yog and g.yog. 
The fact that Phug brag sides with Them spangs ma is in most cases the result 
of BDJNQ bearing a reading of secondary origin. This would mean that 
originally the verb had the form yog and was only in a later revisional step in 
some cases altered to g.yogs. Then in another revision, probably the one which 


PDF Version: BPPB VI (2002) 

imprinted the characteristics of the Tshal pa versions on them, all other cases of 
yog which had remained unchanged, such as those in chapter 9, were also 
adapted to the form g.yogs.”! The reason why an alteration from yog to g.yogs 
was felt necessary may have been the following: the originally intransitive verb 
yog (perfect form: yogs?) was now judged to have transitive character, and 
accordingly made transitive by adding g-. 

* Confusion between the future (gzung) and the perfect form (bzung) of ’dzin ba, 
“to grasp,” in several cases; for the verb ‘jog pa (“to place”) this is the case in 
7C, n. 21 (BQ: gzhag for bzhag, perfect form) and in 10D, n. 6, where only P;23 
read bzhag instead of gzhag. I am not sure which of the two readings should be 
given preference in the latter case. 

* bSad instead of gsad (8B, n. 6): the future form of the verb gsod pa, “to kill, to 
extinguish,” is spelled bsad by NP12T. Ja considers this form “usual” and both 
forms are attested in other texts as well. 

3.3 Colophons and Translators 

The earliest mention of a translation of the TGS into Tibetan is found in the 
catalogue of [Han dkar, compiled in 812 or 824 CE.” The name of the TGS 
appears in the section Theg pa chen po’i mdo sde sna tshogs la | bam po bcu man 
chad la, “Various Mahayana siitras shorter than 10 bam pos:””? 

’Phags pa de bzhin gshegs pa snying po | 310 slo ka/ 1 bam po/ 10 slo ka //"* 

We have good reason to assume that the translation mentioned is in fact Tib.°> The 
mention of the existence of the translation of Tib already at the beginning of the 
ninth century is in line with the colophons found in some of the versions of Tid, 
where the famous Ye shes sde and others are credited with the execution and the 
revision of the translation. Concerning Ye shes sde, we know that he was active 
during the time of the composition of the three manuals regulating the “New 
Terminology” (skad gsar). In one of these manuals, the 814 version of the sGra 
sbyor bam po gnyis pa,” he is mentioned among the Tibetan scholars involved in 
the compilation of the Mahavyutpatti from about the same period.”’ Thus his time 
of activity must have been the early years of the ninth century. The vocabulary 

9! My assumption is supported by the evidence of the Li shi gur khang, completed in 1536, 
which states that texts which had not yet undergone the Normative Prescription of New 
Terminology read yog pa instead of g.yogs pa (see Taube 1978: 173). This Normative Prescription 
should be the one formulated in the sGra sbyor bam po gnyis pa, which we can now date to the 
year 795 (see Panglung 1994: 167). 

% See references in Mimaki 1982: 9, n. 21. 

% For the term bam po see Scherrer-Schaub 1992: 218-220. 

See Lalou 1953: 323. 

* The Tshig mdzod (s.v. sho lo ka) gives a length of 30 to 32 syllables per sloka, which would 
amount to 9,300 to 9,920 syllables for the 7GS according to the length mentioned in the IHan dkar 
catalogue. Given that the verses in Tib consist of four verse padas with nine syllables each, a sloka 
is likely to comprise 36 syllables. Thus, for 310 slokas, we arrive at 11,160 syllables. The number 
of syllables in 7ib is in fact about 10,650 and in Bth about 10,570. Both translations, Tib and Bth, 
lie within close range of the number of s/okas mentioned in the IHan dkar catalogue. Only Tib, 
however, has the term ‘phags for arya at the beginning of the title—the same term that appears in 
the [Han dkar catalogue entry. 

% There exists an older version issued in the year 795 or 783 (see Panglung 1994). 

°7 See Simonsson 1957: 241f. 


PDF Version: BPPB VI (2002) 

used in 7ib is for the most part in accordance with the terminology decided on in 
the MVy.® Thus knowing, on the one hand, that 7ib reflects the New 
Terminology,” and that, on the other, a translation of the TGS is already 
mentioned in the [Han dkar catalogue, we have no reason to doubt that Jib is a 
product of the translation activities at the beginning of the ninth century. 

The analysis of the variants in Tib has revealed the existence of four main 
transmissional lines. This result is confirmed by the colophons of Tib, which 
differ according to the particular group: 

(a) Pi23: no colophon 

(b) LST: rgya gar gyi mkhan po sha kya pra bha dang | zhu chen gyi lo 
tstsha ba ban de ye shes sdes bsgyur cing zhus te gtan la phab 
“The Indian master Sakyaprabha and the Venerable Great Reviser and 
Translator Ye shes sde have executed [this] translation and revised and 
established [it] definitively.” 

(c) BDJNQ: rgya gar gyi mkhan po sha kya pra bha dang / zhu chen gyi lo 
tstsha ba ban de ye shes sdes bsgyur cing zhus te skad (g)sar 
chad(!) kyis kyang bcos nas gtan la phab pa // 

“*... have executed [this] translation and revised [it], and having emended [it] 
according to the Normative Prescription of New Terminology as well,’ 
established [it] definitively.” 

(d) A: rgya gar gyi mkhan po 'dzi na myi tra dang / da na shi la dang | 
zhu chen gyi lo tstsha ba ban de ye shes sde las _stsogs pas 
bsgyurd cing zhus te! skad' gsar cad kyis kyang bcos nas gtan la 
phab te chos kyi phyag* rgyas btab pa // [' ms reads skar; * ms reads 
“The Indian masters Jinamitra, Danasila, the Venerable Great Reviser and 
Translator Ye shes sde and others have executed [this] translation and revised 
[it], and having emended [it] according to the Normative Prescription of New 

Terminology as well, established [it] definitively, and confirmed [it] with the 
Dharma seal.” 

No conclusions can be drawn from the colophons as to the stemmatic relations 
among the four groups. The wording of colophons is too standardized, and they 
can easily be added to texts without giving rise to suspicion. Tib demonstrates this 
clearly. Two different groups of translators/revisers are mentioned: Tshal pa and 
Them spangs ma bear the names Sakyaprabha and Ye shes sde, while 4 has “Ye 
shes sde, Jinamitra, Danasila and others.” The TGS is hardly the only instance of 
such confusion in regard to translators,’°’ and we should be careful about 

% See Zimmermann 1998: 42-46, 

% Whether Tib is the revised version of an earlier translation of the TGS is an open question. 
Bth, however, is too different to be a possible candidate for such a pre-standardized version of Tib. 

1 This must be the revision of translations and the execution of new translations according to 
the recently compiled manuals such as the MVy at the beginning of the ninth century. 
, 1°1 Two other cases come immediately to my mind: the DKP, where the Phug brag version has 
Silendrabodhi and Ye shes sde (no. 294) against dPal gyi lhun po and dPal brtsegs in all other 
versions; and the Siramgamasamadhisitra, where the Phug brag versions have no colophon at all 


PDF Version: BPPB VI (2002) 

definitive attributions. We do not know when colophons were added to canonical 
translations and what exactly certain translator names conjured up at the time. 

Generally speaking, one assumes in colophons a tendency to substitute 
famous names for less well-known ones rather than vice versa. In the case of the 
TGS, this would mean that Sakyaprabha is the name of the original translator. In 
the Peking Kanjur (Otani), he appears as translator for only ten texts.'°? Nothing is 
heard about him as a collaborator in the team of scholars that compiled the rules 
and the vocabulary of the New Terminology. Both Jinamitra and Danasila, on the 
other hand, are mentioned as having participated in that project.’ In the Peking 
Kanjur (Otani), Jinamitra is credited with 160 translations, while Danasila is 
mentioned 107 times. Together with Ye shes sde, they functioned as an important 
translation team at the beginning of the ninth century.'“ The fact that the 
attribution of Tib to Jinamitra and Danasila in A is likely to be secondary does not 
accord with the above evaluation of the recensional variants of A as original.'© 
However, as stated above, we should be cautious and not overplay the possibility 
that A’s colophon could be secondary. We should not, that is, let this possibility 
outweigh the “hard facts” of the variant analysis. The issue, in any case, is without 
relevance for determining the period in which Tib originated, for Ye shes sde, one 
of the main exponents of this period, appears in all of the available colophons. ° 

In view of the fact that 7ib, in its choice of vocabulary, follows closely the 
standards of the New Terminology, it is impossible to explain why the colophon 
in LST does not mention that the text has been “emended according to the 
Normative Prescription of New Terminology,” as expressed in the Tshal pa group 
and A. This inconsistency is also found in other texts, such as the Drumakinnara- 
rdjapariprcchasitra'”’ or the Saddharmapundarika,'® even though, as with Tib, 
there are no traces in the particular versions of anything less standardized than the 
New Terminology. I refrain from a speculative discussion of this problematic 
point. A great deal more comprehensive research in regard to the differences of 
colophons of canonical translations and their appearance in the various Kanjurs 
needs to be done in order to arrive at satisfactory conclusions. 

(no. 207 and 387), while the Stog Kanjur version has Jinamitra, Silendrabodhi and Ye shes sde 
(no. 67) versus the Peking Kanjur’s (no. 800) Sakyaprabha and Ratmaraksita. 

12 Of these ten translations, the TGS is the only text he translated together with Ye shes sde 
alone. ; 

1 See Simonsson 1957: 241. 

1% Bu ston (1290-1364), in his “History of Buddhism,” mentions Jinamitra and Ye shes sde 
among others (see Obermiller 1932: 196f.). 

15 A biography of Rin chen bzang po believed to be written by one of his immediate disciples 
states that he worked together with the scholars Jinamitra, Silendrabodhi and Ye shes sde 
(Snellgrove and Skorupski 1980: 105.15f. (90.7)). As Rin chen bzang po, who had close links with 
Tabo monastery, lived in the tenth/eleventh century, this can hardly be possible. What is 
demonstrated, however, is that from early times on Jinamitra, who appears in the colophon of A, 
was somehow associated in people’s mind with the activities of Rin chen bzang po. This simple 
reason may explain why in the colophon of the Tabo manuscript he and his colleague Danasila 
could oust the little-known Sakyaprabha. 

106 Tn Bu ston’s “History of Buddhism,” too, the translation of the TGS is associated with Ye 
shes sde (see Nishioka 1980: 75). 

©” See Harrison 1992a: xlv f. 

"8 See Simonsson 1957: 219ff. 


PDF Version: BPPB VI (2002) 

Similarly problematic is the meaning of the colophons. We do not know 
what exactly is meant by the different operations mentioned in the colophon’™ or 
whether all scholars mentioned participated in all of the operations.'’° Further, the 
real history of a text from its first rough translation to the canonical version may 
have been more complex than the standardized colophon can indicate. This has 
been demonstrated by Simonsson in the case of the SP, for which he assumes 
several revisions which finally led to a/the canonical version (1957: 114). For the 
time being, in order to determine the relation between the versions of a text one 
can only compare these different versions by noting their variants—not by rely on 
colophons which, given their highly conventional form, could easily have been 
transferred from one version to another without necessarily applying to the text 
version in question. 

4 A Brief Evaluation of the Chinese Materials 

The Taish6 edition of the translation by Amoghavajra (Ch2) contains several 
readings which, in view of the Jin edition and the other translations, must be 
called faulty. In 44 instances the text had to be emended. This is an average of 
about 9 emendations per Taisho page. Of these emendations, the Taisho edition 
deviates from the Korean and Jin edition in 12 cases. In these instances the 
readings of the Taishd are inaccurate, introduced by its editors. I emended the 
Taish6 edition 24 times according to the Jin edition; 7 additional times I had to 
introduce a reading not found in any of the materials utilized for Ch2. The high 
number of corrections following the Jin edition underscores the great value of this 
edition for working with Ch2. Of the 50 differences in the text between T and the 
Jin edition,'’’ only 2 readings of the Jin edition can be said to be definitely 
erroneous, while in 11 cases both alternatives, namely the readings of the Taisho 
and of the Jin edition, seem equally possible. 

The result may be randomly extreme and true only for Ch2. Nevertheless, 
it demonstrates that in any further critical edition of Chinese Buddhist texts, the 
Jin edition should be checked, for it could turn out to provide indispensable 

In the case of Ch,, the translation of Buddhabhadra, the situation is different. 
Altogether, 16 different versions are reflected in my edition, the majority of them 
collated by the editors of the Jin and Taisho editions. I have further included in 
my edition the only two Dunhuang fragments (Dh,, Dh) of the TGS. According 

1 The last phrase in the colophon of A rarely occurs: ... chos kyi phyag rgyas btab (“... 
confirmed [it] with the Dharma seal.”). It remains to be seen whether this formulation appears 
regularly in the colophons of the Tabo manuscripts and what the presence of this addition may 
mean for the history of the version in question. 

® For an extensive discussion on the operations and the relation of the operations mentioned in 
the colophons see, among others, Simonsson 1957: 210ff. Simonsson reflects on the possibility 
that the verb bsgyur in the colophon may not imply that the text to which the colophon belongs is 
necessarily a “direct translation from a foreign language” (p. 223) but could also indicate a 
revision of the text. 

"I The technical apparatus of Ji; indicates differences between it and Gaz. Unfortunately, it 
cannot be said to comprise all of them. I consequently have had to check the Korean edition on my 


PDF Version: BPPB VI (2002) 

to the colophon, Dh, can be dated to the year 639. The characters are written very 
accurately and artistically. This cannot be said about Dh,, the second Dunhuang 
fragment. It does not have a colophon, but judging from the style the characters 
are written in, it should be younger than Dh,.!!* 

There are plenty of variants throughout the text. As the combinations of 
manuscripts and xylographs which share readings vary, combinations under which 
they can be grouped together are difficult to establish. In general the combinations 
MiYu, KuMiSoYu, and JsNaPuQiQs(Zi) seem to be common. Any statement about 
their possible stemma will have to await a more detailed analysis, which is beyond 
the aim of this study. 

The position of the Jin edition in Ch; is not particularly noteworthy. 
Among the 21 emendations, this edition shows the correct reading in only 5 cases. 
A much more faithful transmission, however, is Dhz: 14 of the 21 emendations of 
the Taisho edition are based on, among others, Dk,. Nothing can be said about the 
transmissional quality of Dh,, because the portion of text for which it has been 
preserved is too short. 

5 Remarks on the Various Editions 
5.1 Principles Governing the Critical Edition of Tib 

The primary aim of this edition is to establish Tib in a form as close as possible to 
what came from the hands of the translators. In fact, the materials collated are just 
one part of many more different versions of Tib, which are either difficult to 
inspect or have not been made accessible at all yet. To say “as close as possible as 
to what came from the hands of the translators” means going as far back as the 
present materials allow. Discoveries of new material may well contain more 
archaic versions of Tib, and these would certainly enable us to draw a fuller 
picture of what the earliest version of Tib looked like. For the time being, the 
main principle of the edition has been to give preference to those variants which 
accord with Bth and/or the Chinese translations, which derive from an Indian text 
very close to the text Tib is based on. I have refrained from emending the Tibetan 
text when no variants justifying such an emendation could be found and when it 
was not plausible to assume a mistake within the Tibetan transmission caused by 
graphical similarity or the like. Whether any particular questionable reading of Tib 
has its source in an already faulty Indian manuscript, in a decision on the part of 
the translators not to render the passage in accordance with the Indian text or ina 
mistake pure and simple in the translational process is hence of no further 
meaning for the process of emendation. 

In cases where the other translations are not a sufficient basis upon which 
to evaluate variants of Tib, I have adopted the following guideline: No case can be 
decided mechanically; that is, no combination of manuscripts or prints can 
automatically be accorded preference without analyzing at the same time any 
other variants. Decisions should be made on a case-by-case basis, taking into 
account the content and weighing the probability that individual variants could 
arise. In many instances, it is thus possible to explain a variant by arguing on the 

"2 T owe this assessment to Professor Ikeda On from Soka University (Hachioji, Tokyo). I 
would like to thank him for his help regarding the Dunhuang materials. 


PDF Version: BPPB VI (2002) 

basis of such things as aberratio oculi, dittography, haplography and number of 
syllables in the verse padas, A more rigid method has been chosen only if such 
criteria could not be convincingly applied: As the analysis above shows, Tshal pa 
and Phug brag are closely affiliated. The relation between Them spangs ma and A 
is not absolutely clear. In (28) above it is evident that in three cases out of four it 
was the pair Tshal pa/Phug brag which introduced the mistake, probably owing to 
the heavy revision the transmissions have undergone. For variants conforming to 
this particular combination where we have no means of deciding upon their 
originality, I have accordingly adopted the A/Them spangs ma reading for the 
edition, since the probability of Tshal pa/Phug brag being non-original may here 
be greater. (A sides with Tshal pa and L in only one instance (see (29)). In this 
case I have arbitrarily adopted the reading shared by A, L and Tshal pa.) 

For the portions of text lacking the testimony of A, we are confronted with 
just two main lines of transmission: Tshal pa/Phug brag and Them spangs ma. As 
stated above, Tshal pa/Phug brag seem to have undergone a heavy revision— 
heavier than Them spangs ma was exposed to, as is suggested by the low number 
of single variants of Them spangs ma shown under (34). LST can thus be said to 
have preserved a version closer to the original Tibetan, and I have therefore 
decided to adopt the readings of the Them spangs ma for the edition, even when 
both Phug brag and Tshal pa show a different reading. According to my 
hypothesis put forward above, I give priority to the Them spangs ma reading 
established by LST—the best case—or by LT and ST, wherever only two 
representatives of the Them spangs ma share a reading. 

Regarding punctuation, transliterations and orthography, I decided to 
follow the oldest manuscript within the tradition of the Them spangs ma, namely 
the London copy of the Shel dkar chos sde from 1712. However, I did not adopt 
the numerous contractions of ZL for the edition and the omission of genitive 
particles, though the latter feature could be explained as an archaic characteristic 
(see section 3.1). Concerning orthography, the borderline between orthographic 
idiosyncrasies and faulty spellings is in some instances not clear (e.g. LSTP3: 
mchu for chu in 10D, n. 26). Decisions in this matter must be made on a more or 
less subjective basis until more ancient Tibetan material has been systematically 
analyzed. The critical edition of Tib is consequently intended to establish: 

(1) the urtext of Tib for the portions of text shared with A,''° 

(2) the Them spangs ma archetype for the portions of text where we have 

no testimony of A and where the reading of the Them spangs ma 
archetype does not violate the urtext reconstructed on the basis of the 
other translations, and 

(3) the diplomatic punctuation, transliterations and orthographic 

idiosyncrasies (but not obvious misspellings, contractions and 
abbreviations) of Z throughout the whole text. 

5.2 Remarks on All Editions and Their Critical Apparatuses 

In order to allow an immediate comparison of all versions of any given passage, I 
have decided to arrange the corresponding sections on facing pages. In general, 

"3 Though we have had reason to argue that the major single variants of A may well represent 
an archaic state of transmission, I have refrained from adopting these readings into the edition. 


PDF Version: BPPB VI (2002) 

one double page comprises one section. In some cases, the section is too long to 
fit on facing pages, and consequently I have had to place the second half of the 
section on the following two pages. The numbering within the critical apparatuses 
is not affected by this. 

I have always arranged sigla of versions which belong to the same family 
(e.g. L, S and T, or Ku, Mi, So and Yu) as a cluster unaffected by the rule of 
alphabetic precedence (e.g. LSTN and not LNST, or KuMiSoYuNa and not 

For added clarity, I have refrained from using italics for the sigla of the 
collated texts of Tib when mentioning them in the apparatuses. 

5.3 The Editions of the Tibetan Translations 

Following the examples in Eimer 1983b and Harrison 1992a, I have decided to 
resort to a divided critical apparatus in the presentation of 7ib. The main 
apparatus, comprising all variants shared by more than one manuscript or 
xylographic edition, is found with superscribed Arabic numerals right below the 
Tibetan text of each section. Single variants, marked with superscribed lower case 
roman letters, appear at the end of the whole edition. Some few cases of variants 
shared by two manuscripts or xylographic editions are given in the critical 
apparatus at the end because of their very trivial and coincidental nature (e.g. the 
use of a spungs shad at the beginning of a line). On the other hand, some 
significant single variants which imply a different understanding of the text have 
been adopted in the main apparatus. 

I have tried to keep the apparatuses as easy to read as possible so that, even 
without studying intensively the guide to the edition, one should be able to deduce 
the reading of each manuscript. For the sake of maximum clarity, I have therefore 
noted what the variant reading reported in the apparatus replaces in the main text 
(e.g. yong; for yongs su; ‘di for cig). In many cases the situation was clear and 
deserved no further specification: for example, when the main text reads “... dam /P” 
with the variant “° BQ: /.” it is evident that BQ have the rkyang shad instead of the 
nyis shad, or in the case of the variant “" P3: kyi.” corresponding to the main text 
“... tsan dan gyi™ snying po’i....” that P; reads kyi instead of gyi. In the case of 
additional words in a manuscript or edition, I have typed the additional part in 
boldface, adding the syllable right before and behind it: for instance, the single 
variant “* P): mongs pa rnams.” indicates that the text “... nyon mongs’ rnams 
hyis....” appears as “... nyon mongs pa rnams kyis....” in P2. 

The letters and punctuation in the various manuscripts and xylographic 
editions are not always unambiguously readable. This is especially true of nga and 
da, pa and ba, and of the tsheg between syllables (the last particularly in J and 7). 
I have documented in the apparatuses all cases where letters are completely 
illegible. However, if single letters are doubtful but could graphically be 
interpreted in accordance with all other collated manuscripts and xylographic 
editions of Tib in that passage, I have adopted the reading in question as probably 
the same as in the other materials, 

In the transliteration of the Tibetan, the tsheg is rendered with a blank 
space: so, for example, %9% is transliterated as Ja’am whereas Y9* becomes la 


PDF Version: BPPB VI (2002) 

‘am. Exceptions are the blank spaces before and after shad. These spaces have 
simply been inserted for the sake of more lucidity in the transliteration. 

I have indicated in the apparatuses all of the lacunas occurring in the 
materials belonging to Tib. In the case of Bth, however, lacunas are extremely 
frequent, so that I have mentioned only cases where partially erased letters are still 
found in them. 

If go, ‘gro or a word ending in -g is placed before a nyis shad, the first 
shad of the nyis shad is always omitted by the scribes of BJNQLSP; and also Bth. 
The scribes of D and T do not always follow this rule. However, in all these cases 
my transliteration is go //, ‘gro // or -g //. A places the nyis shad and even a triple 
shad after go or -g. In the prose, P2 omits the first shad after go or -g like the 
majority of manuscripts and xylographs. In the verse section, where P, generally 
separates the padas with a rkyang shad followed by a gap right after the end of the 
preceding pada (e.g. ... 485] -*a4R’,,..), the rkyang shad appears instead at the 

end of the lacuna, when this pada ends with go or -g: e.g. ...959  A9%.... 

Of minor interest but clearly indicating the different affiliations of the texts 
are the forms of the rin chen spungs shad or kyog shad. They are usually 
employed instead of a shad at the beginning of a line after one of the first 
syllables, and in some cases also towards the end of a line. I have always noted 
the position of this spungs shad within the line and, if there was any, mentioned 
the probable reason for its employment. The forms vary and look approximately 
like this: 

AJNP2BuBth: no spungs shad; 

B: | , jj, from 324b onwards: j , jf; 

D:{ ort; 


5:7, 775 


P. iP. 325 z) i; 

Q: 1,11. 

Because subscripts (e.g. mong, rdzog, etc.), contractions (e.g. skyes(s)o, 
yongs(s)u etc.), the use of the anusvara-like abbreviation for m and abbreviations 
in the scripts are often due to such things as the scarceness of space towards the 
end of a line, string holes and emendations, I have, whenever obvious, tried to 
report these features in the apparatuses. In Bth I have therefore indicated the end 
of a line by the symbol “!”, In this way one can differentiate between what 
follows from the particular arrangement of the text on a folio and what may go 
back to the manuscript which the scribe was copying. It is hoped that this 
apparently superfluous information will be of some value in the event that further 
Kanjur materials are discovered in the future. 

In several instances the transliterations pi (Q) and bi (@) appear in the 
apparatuses. They are short for pa’i (8%) and ba’i (9%) respectively. Similarly, for 
pa’i in Bth at the end of the line, we sometimes have a pa with an ‘a chung 
beneath. Above the pa is a mark resembling a parallel double ‘greng bu. The 
character appears as pai in the diplomatic edition. 

Concerning A, I have not noted the use of the ya btags in any words 
beginning with m- in combination with the vowel i or e (e.g. myed, myi etc.). 


PDF Version: BPPB VI (2002) 

Another particularity of A throughout the text which I have not mentioned in the 
edition is the contraction of the cluster st- (¥) to the horizontal ligature ™, of spy- 

(4) to &, or of rts- (8) to. 

5.4 The Chinese Editions 

For the edition of the Chinese texts, I have consulted T 666, Dhg, Dhp, Fs, Ga; and 
Ji; in the case of Ch;, and T 667, Ga and Ji in the case of Chp. I have further 
incorporated the variants noted in the editions of T and Ji;/Jiz (see below). In the 
case of Fs, which is one of the texts consulted by the editors of Ji;, I verified the 
readings on my own since in some cases, unfortunately, variants of Fs are not 
mentioned in the apparatus of Ji;. I could not verify whether the variants of the 
other texts collated in the apparatus of Ji; are completely and correctly reported. 

The basis for my critical editions in all points but punctuation has been the 
two texts found in 7. In the case of variants, I have always adopted the reading of 
T if there were no reasons for thinking that a variant found in another version 
reflects a more authentic stage of textual transmission. When the reading x in T 
was emended to such a variant y (marked by “y[versions where y is 
found]<-x[versions where the reading of T is found] (reason for alteration)”, I 
have usually supplied a reason why I thought this necessary in parentheses at the 
end of the corresponding footnote in the apparatuses. In most instances my 
arguments are based on parallels in the other translations, on other similar 
passages in the same Chinese translation, or on the 4-syllable rhythm in the prose 
and the 2—3 (Ch,) or 4-3 (Ch2) rhythm in the verses, though the argument from 
rhythm should be treated with care: it cannot be employed mechanically 
throughout the whole text. Especially when the Sanskrit is translated by means of 
well-established technical terms, the considerations of rhythm are of merely 
subordinate significance. 

In contrast to the critical edition of Tib, I have restricted myself here to 
mentioning only significant variants in the critical apparatus below the text. 
Variants consisting of characters written without any semantic variation in 
abbreviated or uncommon forms are usually not mentioned whenever the 
characters were identified as equivalent according to the DK, HD or NJJ (e.g. in 
Ch, 0.3b: FR, JER, 3, BK). In cases where I could not be sure whether the variant 
expressed a slightly different notion (and this accounts for most cases), I decided 
to give the variant in a footnote (e.g. in Ch; 5.3d: 3% 77/752 for FH TED). 
Not mentioned, however, are variants resulting from obvious mistakes by the 
copyist, and invariably hindering a meaningful understanding of the context (e.g. 
Ch, 11D: Ji;: —# for =). 

I have dealt with the complex questions of whether to punctuate the Chinese 
editions--and if so, how—in the following way. Opinion among Sinologists 
experienced with Chinese translations of Indian Buddhist texts varies greatly in 
this matter. The view that the text should not be punctuated at all and should be 
rendered as it was produced by the translator, namely without any indicators, 
forms one extreme of the range of alternatives. For the reader, this offers the 
advantage that he can approach the text without being biased towards the 
punctuation of an editor. To question the understanding suggested by an editor 


PDF Version: BPPB VI (2002) 

when translating the text oneself is not that easy, and one can quickly end up 
unconsciously following a given interpretation without taking pains to examine all 

On the other hand, I think that the activity of an Indologist specializing in 
Buddhist studies should not be limited to a purely descriptive collation of the 
Chinese materials. It is equally a part of his work to show the reader how he 
himself understands the text, even at the cost of sometimes overshadowing other 
possible readings. Especially when taking just a quick look at a passage in a 
certain text I, for one part, have found it quite helpful and timesaving to seek some 
guidance from an editor’s punctuation, and so avoid having to consult some 
translation. This standpoint, in fact, probably represents the view of anybody who 
cannot call himself a Sinologist and whose main field of Buddhological studies 
lies—as is still the case for many Western Buddhologists—in India. The situation 
of a Sinologist with a good acquaintance of the genre of Buddhist translations is 
likely to be different. My punctuation will be disturbing to such a scholar, and he 
is advised to consult only the emendations suggested in my apparatuses when 
reading the text in the Taisho. 

At the other extreme of the range of possible punctuations is a system 
which can be called currently dominant on mainland China, where it is widely 
employed when dealing with old as well as contemporary literature. It 
incorporates virtually all of the Western punctuation marks, including even 
question and exclamation marks, inverted commas and semicolons. And though it 
may be the freely chosen standard of punctuation in China, it reminds a Western 
scholar immediately of Europe’s Orientalist custom of imposing inappropriate 
Western standards upon local indigenous cultures. I cannot judge whether 
Western punctuation fits the structure of the Chinese language. This question can 
only be answered by specialists in that field. In the editions, I restrict myself to the 
use of three punctuation marks, namely the juhao GH “, ”, the douhao 325 
“~ ” and the dunhao ty “. ”, two of which (juhao and dunhao) can already be 
found in Chinese texts from the Han }& dynasty. The role of the dunhao in my 
editions is the easiest to define: it functions as an enumerative comma separating 
the single members in a list of coordinated items. Long enumerations, especially 
in OE, which contains the names of 50 bodhisattvas, thus become much easier to 
deal with. The decision to use it, however, forced me in many instances to 
determine how to segment compounds consisting of several characters. If the 
Tibetan parallel did not help and no other occurrences of parts of the combinations 
could be found in the same text, choices had in some instances to be made on an 
admittedly arbitrary basis. 

The decision whether to choose the juhao 43% “, ” or the douhao ia5R 
“ ” may also have been arbitrary in many cases. Whilst the juhao is here in 
general intended to indicate where a group of closely connected syntactic units 
which we may call a “sentence” should be divided from another such group, the 
douhao serves to show component structures within this group. No douhao is set 
between a subject, predicate, object and any adverbial specifications. The douhao 
is, however, regularly employed to mark embedded subordinate clauses, to 
separate predicative elements from each other and to set off direct speech (e.g. 

I am, of course, aware that the borderline between the juhao and the 
douhao is often impossible to define; that is, there is not always a definite 


PDF Version: BPPB VI (2002) 

sentence break. Whenever the basic rules of Chinese grammar allow, I have tried 
to tailor my decisions to the Tibetan versions (especially when dealing with Chz) 
and the possible Indian structure on which the translations are based. 

It is clear that this kind of rule cannot provide more than a rough 
indication of why I felt these Chinese translations ought to be so punctuated. For 
texts from a different genre, the guidelines might well be quite different. I do not 
consider the textual understanding implied by my punctuation to be the only one 
conceivable—rather a suggestion that clearly mirrors the Indian background of the 


PDF Version: BPPB VI (2002) 

D The Editions 

The Critical and Diplomatic Editions 

PDF Version: BPPB VI (2002) 


// rgya gar skad du* /° arya! ta° tha’ ga ta” garbha’ na‘ ma® ma ha? ya° na si’ tra /‘ bod 
skad du / ’phags pa de bzhin gshegs pa’i snying po zhes bya ba® theg pa” chen po’i 
mdo //* sangs rgyas dang’ byang chub sems dpa’ thams’ cad la phyag ’tshal lo’ / 

Tp: arya; P23: ’a chung of compressed letters. T Pio: su; P3: 'a chung of su 
arya with a very small letter 4 Pio: na; P3: ’a chung of nd with a very small letter 
(inserted later?). with a very small letter (inserted later?). 

2 Piss ta [P2: ‘a chung of ta (inserted later?). 8 JNQT: /; P3: double spungs 
with a very small letter 5 Di: ha; Po3: maha for ma ha; shad (after first syllable); S: 
(inserted later?)]. P2: ’a chung of had with a gap of about nine letters 

3 P,: gha rbha for garbha; Ps: very small letter (inserted between the shads of the 
garbha for garbha; S: ga ta later?). nyis shad, 
garbha with small, § Piast ya. ° ST: dang /. 



' MiYu: IER for RK. 
? KuMiSoYu: = ie Ai for = HK. 

“Translated during the Eastern Jin by the Tripitaka [specialist]* Buddhabhadra 
[from] India”.” 

* KuMiSoYu: “... by the Dharma Master of the Tripitaka....” 
> MiYu: “... [from] northern India.” 


PDF Version: BPPB VI (2002) 

rgya gar skad du ta tha ga ta gar ba ma ha ya na su tra / bod skad du : de bzhin 

gshyes! snying po zhes bya ba theg pa chen po’i mdo : bam po dang po : 





* All the titles mentioned in the colophon are found in the biographies of Amoghavajra contained in T 

2056 CARSACA TERE Ze APE IE RS Te = TIAA) and T 2061 OR rae 1S). A translation of the 
part concerning Amoghavajra in the #/=)(@( was done by Chou (1944/45: 284-307). 

' Ht¢-BM (see the title of the biography of Amoghavajra by his pupil (according to BSKD vol.7, p. 
372c) (T 2056): AREA A ERB Ze AEE Re = RTA; the same titles are also mentioned 
in the text itself (294a28); see also T 2061, 713c6-7: HALA RES IES = 3M, which should probably 
be emended to MEIAH ERS =m.) 

? B& could also refer to AMIE. 

“Tn compliance with an imperial decree translated by the Tripitaka [Specialist] of the 
temple Da Xingshan,* with the title “One of Great and Wide Knowledge”, the 
Sramana Amoghavajra, Commander Unequalled in Honor,’ Specially Promoted 
Probationary Chief of the Court of State Ceremonial,’ Duke of Su [with] a Fief of 
3,000 Households,® endowed with a purple [robe],’ given [the title of a] Minister of 
Works, posthumously called ‘One Who Greatly Reveals Righteousness”".” 

* The temple Da Xingshan was situated in the capital, Chang’an. 

> The title A/R*} appears in connection with = 7 (T 2056, 293b22: He ARE = HK; T 2061, 713al1- 
12: HOHE =F) but also as part of Amoghavajra’s posthumous name (T 2061, 713c6-7: see 
above fn. 1; 72056, 294a28: BA AHI BETH. 

° See Hucker 1985: 3105; T 2056, 293c21-22; T 2061, 713b21-22. 

4 See Hucker 1985: 6335; 5204; 2905; T2056, 293b22; 7 2061, 713al1. 

* See Hucker 1985: 3258; 5258; T2056, 293c22; T2061, 713b22. 

The purple robe was bestowed on Amoghavajra by the Emperor after Amoghavajra had managed “to 
make it rain” in a very dry summer. See T 2056, 293a24; T2061, 712c16. 

® See Hucker 1985: 5687; T2056, 294a28; T 2061, 713c6. 

"See T 2056, 294a28; T2061, 713c6-7: BEULARHE IES — is, which probably should be emended to 
ait A HELE JS = 


PDF Version: BPPB VI (2002) 


di skad bdag gis thos pa! dus gcig* na / bcom Idan ’das mngon par rdzogs par? sangs 
rgyas nas /” lo beu bzhes pa dang’ /’ shin tu’ tsha ba’i* dus kyi tshe / rgyal po’i' khab 
na /° bya rgod® kyi® phung! po’? ri la® rin po che’i gdugs* kyi khang bzangs! tsan’ 
dan® gyi™ snying po’i khang pa” brtsegs pa na° /” slob’ pa dang /° mi slob pa’i nyan 

°i!° dge slong gi dge ’dun chen po phal cher* 

thos kyi dge slong ’bum du tshang" ba 
dgra' bcom pa / zag pa zad pa / nyon mongs pa med pa / dbang” dang Idan par’ gyur 
pa'' / sems shin tu’* mam par” grol ba / shes rab shin tu* mam par’ grol’ ba / cang 
shes pa /’? glang”® po chen” po /'* bya ba byas pa / byed pa byas pa /’* khur bor ba / 
bdag gi“ don rjes su!® thob® pa / srid par'” kun tu™ sbyor ba!® yongs su’? zad pa /®8 

yang dag pa’i shes pas’ sems” shin tu!’ mam par grol ba / sems”! thams cad kyi”” 

dbang gi*? dam pa’i pha rol tw! son pa“ sha stag la /** 
TBINQP)3: pa’i dus. here and in the preceding 77 Pio3: pa. 
2 BINQP,;: om. /. note is contrary to the iY BQP 123: ba /. 
3 JN: om. /. parallel in 0G). STP: yongsu; P3: yong, for 
“BIQ: om. /; T: spungs shad ° BINQP}: om. /. yongs su (at the end of the 
(after second syllable).  BOP;: pai. line). 
> BINQP,,3: om. /. LT: ‘gyur ba for gyur pa. 0 P,.: pas /. 
§ pias: la /. 2 7Q: du. 21 BYQP12:T: sems can thams. 
y BJQP 2: can for tsan; P3: Bp Lom. /. 22 IN: kyis. 
om. tsan. 4p: om. /. 33 Di: om. gi. 
8 BIQPi3: om. dan; N: 5 BQ: om. /[Q at the end of 74 BINQPi23: om. /. 
tsandan (883') for tsan dan the line]. 
1 LN: rjesu. 

(The variant of BIQP 2:3) 


eke PEELS RS ELS ASB, pcb, 


PDF Version: aoe VI (2002) 


*di skad bdag gis thos pa’i dus gcig na / bcom Idan ’das rgyal po’i khab rgya rgod 
spungs pa’i ri rin po che’i gdugs kyi khang pa tsan gyis snying pos brtseg pa’i khang 
bu na mngon par! rdzog, pa, sangs rgyas nas : lo beu bdun kyi dus na dge slong chen 
po dge ’dun dge slong lo stong tsam pa dang : nyan thos bslab pa dang : mi slob pa 
dang : thabs gcig tu bzhugs pa kun ++ degra bcorn pa zag pa zad pa sha stag go: nyon 
mongs pa, med) dbang du gyur + sha stago : semns shin tu mampar grol + sha stag go: 
shes rab shin tu mam par grol ba sha stag go : cad shes pa sha stag go : klu chen po 
sha stago : bya ba byas pa sha stago : khur bor ba shasteg go : bdag gi don thob pa 
sha stago : ’byung bai! sbyor ba zad + sha stag go // yangdag pa’i dkas serns mar 
pat grol ba sha stogo : sems can thams cad kyi dbang dang : pharol tu phyin pa 
mchog thob pa shastago // 


mmo, HGR, SHE, LER, BERR, BEES, @ 
 BlGavigl Be. 



PDF Version: BPPB VI (2002) 


di Ita ste / tshe dang Idan pa ’od srung! chen po dang / tshe dang Idan pa’ lteng” 
rgyas ’od srung’’ dang / tshe dang Idan pa chu bo” ’od srung®” dang / tshe dang Idan 
pa ga® ya’ ’od srung’”" dang / tshe dang Idan pa ka!” tya’i!? bu chen po dang /° tshe 
dang Idan pa gsus po che dang / tshe dang Idan pa ba ku la’* dang / tshe dang Idan pa 
nam gru dang / tshe dang Idan pa rab ’byor dang /* tshe dang Idan pa byams ma’i® bu 
gang po dang / tshe dang Idan pa ngag dbang dang / tshe dang Idan pa sha’ ri’i bu 
dang /* tshe dang Idan pa mo’w!® dgal!” gyi bu chen po! dang / tshe dang Idan pa cang 
shes ko’u’® di’? nya” dang / tshe dang Idan pa ’char ka dang / tshe dang Idan pa 
sgra” gcan™ zin® dang / tshe dang Idan pa dga’ bo dang / tshe dang Idan pa nye 
dga’® dang / tshe dang Idan pa” kun dga’ bo dang / de dag! la sogs”* pa dge slong 
*bum du tshang ba dang”> thabs gcig”° tu bzhugs so // 

' DS: srungs. 3B: katya’i for ka tya’i; JQ: '9 J: di for di; NP3: ti for di; 
5 JNQP)2: steng. tya’i for tya’i; Pz: ta’i for P,2: ‘di for di; T: de for di. 
P3S: srungs. tya’i (with partially erased 0 B: nyas; Q: lacuna of one 

‘P.>: srung chen po dang. 

2 BQP;: rlung for bo; JNP}2: 
klung for bo (MVy 1050: 
chu klung ’od srungs; Bth: 
chu bo ’od srung). 

® Ps: srungs. 

7 Pio: srung chen po dang. 

5 INQ: ga [JQ: ’a chung of ga 
probably inserted later]; P: 
ka or ga (?); Pz: om. ga; P3: 

° NP: ya for ya; J: ld or ya 
(?); Q: la for ya. 

ps: srungs. 

" p.): srung chen po dang. 

2 INQ: ka [Q: ’a chung of ka 
probably inserted later]. 

Ch, - 

zhabs kyu). 

4B: bakku la for ba ku la; J: 
ba kkula, ba kku la, bakkula 
or bakku la (?); Q: bakkula 
for ba ku la. 

3 D,: sha; Pz: a’ chung of sha 
with a very small letter 
(inserted later?). 

16 BNQ: mau [BQ: additional 
to double na ro of mau: 
subscribed ‘a chung]; J: 
maud; P,: mu’u for mo’u. 

" BIQ: gal: P18: ‘gal; T: 
dga’,N: maudgal(415’) for 
mau dgal. 

8 BINQ: kau for ko’u; Py2: 


letter between nya and 

7 Q: sgrag; JN: sgra or sgrag 

ee BQLP 23: can for gcan [P): 
can with small letters 
beneath the line]; JN: sgrag 
can or sgra gcan (?). 

Ge Py23: ‘dzin. 

4B. lasogs (without tsheg 
between); P,Q: las sogs for 
la sogs. 

2 BP},3: dang /; Q: lacuna of 
one letter between dang and 

6 BNOQP;: cig. 

PDF Version: wee VI (2002) 


*di lta ste : tshe dang Idan pa ’od srung chen po dang : tshe dang Idan pa u ru dbyil ba 
"od srung dang : tshe dang Idan pa! chu bo ’od srung dang : tshe dang Idan pa ’ga’ 
yang ’od srung dang : tshe dang Idan pa kosti la chen po dang : tshe dang Idan pa ba 
ku la dang : tshe dang Idan pa red pa dang : tshe dang Idan pa rab ’byor dang : tshe 
dang Idan pa gang po me tre’i bu dang : tshe dang Idan pa ’ba’ giya sha dang! tshe 
dang Idan pa sha ra dva to’i bu dang : tshe dang Idan pa mau ’a gal kyi bu chen po 
dang : tshe dang Idan pa ag nya da ko’u ’di dang : tshe dang Idan pa ’char pa dang : 
tshe dang Idan pa sgra gcan dang : tshe dang Idan pa dga’ bo dang : tshe dang Idan pa 
mtha’ yas! la sogs pa dge slong stong stong tshangs pa dang : 

mee, AAU. RRM. ARR. RaeRS. AS 
(ARRAN, SSR. RSI. Rec. Raat eile, RS he 

" $6[GayJiz] <8. 


PDF Version: BPPB VI (2002) 


byang chub sems dpa’ sems dpa’ chen po’ sangs rgyas kyi zhing tha dad pa nas ’dus 
pa’ /' gangga’i? klung® drug “cu’i? bye® ma snyed thams* cad kyang /* skye ba gcig 
gis® thogs pa /° mngon par shes pa chen po dang / stobs dang /* mi ’jigs" pa thob pa’ / 
sangs! rgyas bye ba khrag khrig ’bum phrag mang po la bsnyen’ bkur byas pa  phyir 
mi! Idog™ pa’i chos kyi ’khor lo rab tu" bskor pa® / gang dag? gi ming thos pa tsam 
gyis "jig rten gyi khams tshad med grangs med pa’i sems can bla na med pa yang dag 

par rdzogs* pa’i byang chub las phyir mi Idog par ’gyur ba’ sha stag la /"° 
 BINQPia3: om. /. gang ga’i P3Q: snyen. 
2 BP}23: ang ga i for ; JP 23: bcu’i. : BINQLSP,;: ba. 
angga’i (IS); JT: gan BJNQ: om. /. JNP3: gyur pa for ’gyur ba. 
gare tN) Ti eae 5 BINQ: om. /. 10 BINQP,93: om. /. 

gai; NS: gangga i ag); Q: 6.5: 0m. /. 

Se Bae HAY SSA REN, CRRRA THeabh the 


PDF Version: Ae VI (2002) 


chu bo gang ga drug beu’i bye ma snyed kyis saryas kyis zhing sna tshogs nas : *dus 
pa’i byangchub sems dpa’ serns dpa’ chen po rnams thas cad kyi tshe gcig gi thog 
pa sha stag go : mngon! par shes pa chen po dang : mi ’jig pa myed pa shastago // 
sangs rgyal mang po bye ba khrag khrig brgyad stong la bsnyen bkur byas pa sha 
stag go // phyir mi ldog pa’i chos kyi ’khor lo bskor ba sha stag go // ’di dag ni ming 

(24611 svis tshad med grangs meg pa’i ’jig rten kyi kharns kyi seths ca, 

thos pa tsam 
rnams bla na med pa yang dag par rdzogs pa’i byangchub las phyir mi ldog par ’gyur 



SAN DRI Rae. ale, Ageia, MRS. Fe— 

| 4s «eh (—4: 46 GR for Skt. ekajatipratibaddha). 


PDF Version: BPPB VI (2002) 


’di Ita ste / byang chub sems dpa” sems dpa’? chen po chos kyi° blo gros dang /’ 
sengge’i” blo gros dang / stag gi blo gros dang / don gyi blo gros dang / rin po che’i 
blo gros dang / rab mchog blo gros dang / zla ’od? dang / rin® chen zla ’od* dang / zla 
ba‘ nya ba’i ’od dang /° mam par gnon pa chen po dang /* mam par gnon pa dpag® 

med’ dang / mam par" gnon pa mtha’ yas® dang / ’jig rten gsum rnam par gnon’ dang / 

mi g.yo! ba’i!? gnas’! mam par gnon! dang / mthu chen thob dang * spyan ras gzigs 
dbang phyug dang / spos' kyi glang po dang / spos™ dga’ dang / spos" dga’ ba’i!” dpal 

dang /° dpal gyi snying po dang / nyi ma’i’ snying po dang /"° tog" dang / tog chen 

po dang / tog dri ma‘ med pa dang / rin chen mtha’ yas dbyu’* gu" dang /” 

: P12: ‘od blo gros dang. 10: om, ‘i; P12: g.yo’i for g.yo 
4 P.,: blo gros for zla ‘od. ba'i. 
> PT: spungs shad [P,: after "| Jp5: om. gnas. 
first syllable in top line of ? BINQP 2: dga’i for dga’ 
the folio; T: after first ba’i; P3: dag gi for dga’ 

'P.: dang / ala ’od dang rin 
chen dla ’od dang / seng 
ge'i; Po: dang zla ’od dang / 
rin chen zla ’od dang / 
se(or: si)ng ge’i; T: spungs 

shad (after first syllable). syllable]. 

2 BINQP,3ST: seng ge’i for 6 Po: dpag tu med [P>: -p- of 
sengge 'i(SR2); Po: seng dpag with a small letter 
above the line]. 

ge’i or sing ge'i (?) ['a of "BINQP 2: med pa dang 


3 LT: spungs shad (after first 

4 INP) 3: dbyug for dbyu; (Pz: 
om. dbyu). 

ge’i with a very small letter 
beneath ge (inserted 

® BQ: yas pa dang. 
° NPy23: gnon pa dang. 

Se. Smee. AS. ME. ANSE. BIS. 
' MiYu: SUllSS SES UES for Smee HE. 


PDF Version: BPPB VI (2002) 


*di Ita ste : byangchub sems dpa’ serns dpa’ chen po chos kyi blo gros dang : sege’i 
blo gros dang : stag gi blo! gros dang : don gyis blogros dang : dkon mchog gi blo 
gros dang : mchog rab, gyi blo gros dang : zla ’od dang : rin po che’i zla ’od dang : 
zla ba gang ba’i ’od dang : dpa’ ba chen po dang : tshad med par dpa’ ba dang : 
mtha’ med par dpa’ ba dang : kharns gsurn na dpa’ ba dang : rkang pa! mi g.yo bar 
dpa’ ba dang : mthu chen po rnyed pa dang : ram par spyan ras gzigs kyi dbang po 
dang : spos kyi bal glang dang : dris rol pa dang : dri kun dga’ ba dang : dpal kyis 
snying po dang : nyi ma’i snying po dang : dkon mchog gi dpal dang : dpal! chen po 
dang : dri med pa’i dpal dang : mtha’ yas pa’i chen po che’i khar ba dang : 


Teche. SRS. RES. RSE. Pee Ae 

2. SRotee Armee. tse. AMS. PSE. 
fe Chee 



PDF Version: BPPB VI (2002) 

rin chen dbyu’® gu” ’dor dang /* rin chen dri med dbyu’® gu dang / mchog tu dga’ 
ba’i rgyal po dang /’ rtag tu rab” dga’!” dang / lag na rin po che™ dang / nam mkha’i!® 
mdzod dang /** ri bo dang /'” ri rab dang / ri bo chen po dang / yon tan rin chen“ 
snang” dang / gzungs”’ kyi dbang phyug gi" rgyal po dang / sa ’dzin dang /®° sems 
can thams cad kyi™ nad! sel dang / rab tu yid dga’ dang / yid! skyo”! dang / skyo” 
med dang / ’od byed dang** / tsan* dan! dang / g.yo ba”* zlog”® dang / dpag med™” 
mngon bsgrags”® dbyangs™ dang / byang chub kun nas bslang® dang /*? mthong ba 
don yod dang / chos thams cad la dbang”’ sgyur®® ba dang / byang chub sems“ dpa’ 
sems dpa’ chen po™ byams® pa dang / ’jam dpal gzhon nur" gyur pa dang / de dag la 
sogs” pa™ byang chub sems dpa’ sems dpa’ chen po gangga’i’? klung drug ““cu’i?! 

bye ma snyed dang yang™ thabs gcig*? go™ //” 

'S INP, 93: dbyug for dbyu. for deletion]; P,.: om. /. 27 Dio: om. dbang. 
16 INP 23: dbyug for dbyu. ce Pj3: gzugs; P2: gzug. a8 NP3: bsgyur. 
'7p..: om. dga’; P3: dad for 21 D103: skyo ba dang; S: skyob 2° B: lasogs for la sogs; Q: las 
dga’. for skyo. sogs for la sogs; P\2: stsogs 
18 BJQP.3: nam mkha’ for nam 22 D,: skye for skyo; P2: skyed for sogs. 
mkha’i, L: namkha’i for skyo. ® BP.;: gang ga’i for 
(AR); NP: namkha’ for . JQ: can for tsan; T: tsam. gangga’i; J QT: gang ga & 
nam mkha’i. is P.S: om. ba. iy NS: ganged i; Py: gang gi’. 
19 B. dang | ri bo dang / ri rab BJNQ: om. zlog; P2: bzlog - JP 123: bcu i 
(dittography);T: dang /(ri 9g £08 7/08. BNQPs: cig. 
bo) / ri rab [ri bo marked Ps: sgrags for bsgrags; BQ: 
par sgrags for bsgrags. 

with dots above and beneath 


OCS. HDC. Sie. BeBe. FS. Tees 
fps, NERS. VERS. Meee. SCORE. HM 
ESN Se Bae, RRS A, 



PDF Version: BPPB VI (2002) 


dri ma med pa’i chen po che’i khar ba dang : ’bar ba’i rgyal po dang : rtag tu dga’ ba 
dang : lag na rin po che dang : namkha’i mdzod dang : rab lhun po dang : shin tu rab! 
Thun po dang : rab lhun po chen po dang : yon tan rin po che’i ’od dang : gzungs kyi 
dbang po’i rgyal po dang : sa ’dzin dang : serhs can tharhs cad kyi nad sel ba dang : 
rab tu dga’ ba’i yid dang : skyo ba dang : skyo ba med pa dang : ’od byed dang! tsan 
dhan dang : dris las mi ldog pa dang : tshad med par bsgrags pa’i dbyangs dang : 
byangchub sems dpa’ kun Idang bar byid pa dang : snang bas don med par mi ’gyur 
ba dang : chos tharhs cag la dbang gyur ba dang : byams pa la! byangchub serhs dpa’ 
setns dpa’ chen ++ la stsogs pa : chu bo gang ga bye ma drug bcu snyed kyi 

byangchub sets dpa’ setns dpa’i chen po mams dang : 



| Fiz: BS for $ (cf. similarity to 3 of the graphic variant 33 for $# (see NIJ p. 188)). 

” Rl Gagi2] Bk. 


PDF Version: BPPB VI (2002) 


lha dang /' klu dang /” gnod sbyin dang / dri za dang / Iha ma yin dang /* nam mkha”’ 
Iding dang /° mi ’am‘ ci dang / Ito *phye chen po dang / mi dang mi° ma yin pa dpag 
tu med pa dang yang’ thabs® gcig® go‘ //* de nas bcom Idan ’das® ’khor ’bum phrag du 
mas yongs su” bskor cing mdun! gyis’ bitas! te /“ rgyal po dang / blon po! chen po 
dang / tshong dpon dang /” khyim bdag dang / blon po dang / grong rdal” ba® dang /° 
yul gyi mi mams kyis bsti’ stang du” byas / bla mar byas / rim gror'° byas shing® 
mchod do // 

' BOP): om. /. 4 JN: mi ‘am or mi’am (2). P3: om. ba. 
2N: shad crossed out?; > BNP;: cig. * BOP 23: Sti. 
P,2: om. /. § p.,: ‘das la ‘khor. 0 NP ,2: gro; P;: 'gror. 
3 LNP 1»: namkha’ for nam 7 Pigs: gyi. 
mkha’. 8 JT: pa; N: pa or ba (2); 

fil, ASR, SBE. 


PDF Version: BPPB VI (2002) 


tshad med grangs med pa’i lha dang : klu dang : gnod sbyin dang : dri za dang : tha 
ma yin! dang : namkha’ Idi dang : mi ’am ci dang : Ito’phye chen po dang : mi dang 
mi ma yin pa thabs gcig go // de nas bcom Idan ’das ’khor du ma brgya phrag stong 
gi bskor + mdun du byas nas : rgyal po dang rgyal! pos blon po dang : khyim bdag 
tshong dpon dang : grong khyer dang : grong ’dab dang : setns can rams kyi bla mar 

byas : rim gror byas : mchod par byas so : 



Jin: GE for HB. 
? [Gayiz]<H. 



PDF Version: BPPB VI (2002) 

de’i tshe becom Idan ’das bshos? gsol ba’i ’og tu? tsan dan’ gyi snying po’i’ khang pa 
brtsegs® pa de nyid du! nang’ du‘ yang dag ’jog la® zhugs® par gyur to" // de nas 
sangs rgyas kyi! mthus / tsan dan“ gyi snying po’i! khang pa™ brtsegs* pa de” las 
padma* mdab° ma® bye ba khrag khrig phrag’ *bum yod pa / tshad® shing rta’i 
*phang’ lo tsam pa / kha dog dang? Idan pa /'° kha ma bye" ba bye ba!! khrag khrig 

phrag* ’bum' byung" bar gyur’? te /'? de dag steng gi nam mkha’i’* bar snang la’® 

mngon par ’phags nas /’° thams’ cad” dang Idan pa’i sangs rgyas kyi zhing ’di khebs 

/’ tin po che’i bla re” bzhin du kun tu’? gnas par 


par gyur te ” ’di Ita ste /'” dper’ na 
gyur to™ //*> padma’i*”’ snying po re re la yang” de bzhin gshegs pa’i sku skyi 
mo krung” beas te” bzhugs shing /** ’od zer® ’bum dag rab tu” ’gyed par®* kun 
tu”® snang la /” padma”® de dag thams cad kyang shin® tu™ kha bye bar gyur to //" 

 BINQP 193: yi for du; chad for tshad. 7 BINQP);: om. /. 
LST: om. du ; see note 20 in i JN: phang. 18 INTP,: om. /. 
my translation. e BINQP 123: om. /. Bs JP 23: du. 
; LST: om. nang. Ny: Ja for ba; P23: om. bye bad BJQ: pad mo’i; N: padmo ’i; 
P12: bzhugs. ba. Pi231: pad ma'i. 
BOP». rises 2p, ie 21 BOP: dlyil. 
4: rtsegs. 123: “‘gyur. QP 123: dkyil. 
> BJQPT: pad ma for padma 8 BQ: to // for te /. ?2 By: dkrung; T: klung(?). 
(8H); Po: pad for padma. “ BJQ: nam mkha’ for nam 3 Pio: te /. 
6 INP.S: ‘dab kha’i; LTP}: namkha’i for ae BINQP 23ST: om. /. 
3b: * ns 2 
is ‘i nam mkha’i; Pi3: pa. 
a iste a is N: namkha’ for namkha’i; 26 JP3: du for tu; P2: om. ’gyed 
hr Re P3: om. nam mkha'i. par kun tu (aberratio oculi). 
8h. ise for tshad; NQ: B st: Ja/. 27 Diag: om. /. 
: ne. ee 16 BINQP,2: om. /. 8 JQP 231: pad ma. 

fre, 4 SAC, MRR. —OTEN, BAH, LAMEZ, Tee 
THA, Jean, —SEE RE. — SEITE 



PDF van Eee VI (2002) 


de nas de’i tshe beorh Idan ’das bshos gsol 494) gyis *og du tsan dhan kyis snying po 
khang bu brtseg pa na bzhugs te : de nas sangs rgyas kyis mthus tsan dhan kyi snying 
po khang bu brtsegs pa : de las pad mo bye ba khrikhrig phrag stong byung ba shing 
rta’i ’khor lo tshad tsam lo ma bye ba khrikhrag stong pa! khatog dang Idan pa : dri 
dang Idan pa snying po Ja ’dus te : +s pa de dag gi namkha’i bar snang gi gnas las 
*phags de : des sangs rgyas kyi zhing thams cad g.yog par gyur pa na : “di Ita ste: rin 
po che’i gzhal med khang bzhin du kun tu g.yo ba pad ma’i! snying po re re la de 
bzhin gshegs pa’i sku dkyil mo drung bcas de bzhugs na : ’od gzer brgya stong rab tu 
gtong bar kun tu kun tu Idang ste : pad mo de dag las thas cad kyang rab tu +s par 
gyur to // 



& BRE. ema laAme, ma -OaAt, ARSRIRS 

' inserted according to GayJi, (four-syllable rhythm). 


PDF Version: BPPB VI (2002) 


de nas sangs rgyas kyi byin gyi’ rlabs” kyis’ /* padma” de dag gi mdab® ma® de dag” 
thams® cad kyang’ mog mog‘ po® dang /‘ nog® nog po dang / dri nga’ ba® dang / smad 
par ’os pa” dang / mngon par dga’ bar ’gyur ba ma yin par gyur to //' ’on kyang 
padma’i’® snying po de dag la /"! de bzhin gshegs pa’i sku mams) skyil’? mo“ krung” 
beas te /“* bzhugs! shing /’* ’od zer™ bum” dag rab tu° ’gyed par kun tu!® snang ngo!” / 
de bzhin gshegs! pa’i sku'® padma’i”” snying po la bzhugs”’ pa de dag gis”! kyang’ 
thams cad dang Idan’ pa’i sangs rgyas kyi zhing ’di” khyab par gyur te /”? de’i tshe 
sangs rgyas kyi zhing ’di shin tu! mdzes par gyur to // de nas de’i tshe byang chub 
sems dpa’i tshogs thams" cad dang / ’khor bzhi yang” shin tu” ngo mtshar* du” gyur 

cing” dga”” bar gyur to //* 
' Pos: gyis. beneath the line; Q: dring of the line]. 
? BINQP;: brlabs. for dri nga. 87 Pp): om. sku. 
: BQ: gyis; P3: Avi; T: lacuna - BJQP23T: pad mai. 2 BJQP)3T: pad ma’i; P2: om. 
of about five letters between u BINQP,23: om. /. padma ‘i. 
kyis and shad. 2 Dio: dkyil; Ps: skyil [ s- with 2° BQ: zhugs. 
4 BINQP,23S: om. /. a very small letter (inserted 21 D.,: om. gis. 
5 BJQPST: pad ma; P\2: pad later?)]. 22 D.>: ‘di dag khyab. 
ma’i for padma. 3 JP,: dkrung. 3 BQ: om. / [Q at the end of 
° JNP3S: ‘dab. 4 BINQP 23S: om. /. the line]. 
T BJQP123: om. kyang. 5 BINQP p33: om. /. 4 IT: /, 
° BQ: nag. 16 JP3: du. oe Py: tu. 
ops: nga with a small letter INP: snango [P; at the end 

Pei, ARC, BR, HEA, — OREO. Bik 
' Dh Fs: BM for HI. 


PDF Mee eee VI (2002) 


de nas sangs rgyas kyi byin kyi rlab kyis! pad mo de dag gi lo ma tharhs cad ryes 
shing zhum + tog ngan pa dang : dri mi zhim pa dang : smrad pa dang : mi bzang bar 
gyur to : de nas yang pad mo de dag la de bzhin gshegs pa’i sku dkyil mo grung bcas 
de bzhugs nas ’od zer brgya! phrag ++ rab tu gtor ba kun tu snang ste : pad mo de’i 
snying po la : de bzhin gshegs pa’i sku bzhugs pa sangs rgyas kyi zhing thams cad 
dang Idan par khyab par gyur nas : de’i dus na sangs rgyas kyi zhing ’di mchog tu 
bzang! bar gyur+ to : de nas de’i dus na byang chub sets dpa’ mang po : thams cad 
dang : ’khor bzhi bcom Idan ’das kyi rdzu’phrul mngon par ’dus byas pas : mthong 
nas : ngo mtshar thob :cing) rmad du gyur pa thob par gyur nas : 




” Jin: BE for BRR. 



PDF Version: BPPB VI (2002) 


bcom Idan ’das kyi’ rdzu ’phrul mngon par ’du mdzad pa de mthong nas the tshom” 
du gyur te * gang padma’ bye ba khrag khrig’ phrag *bum® ’di dag gi’ ’dab° ma 
rams‘ ’di ltar kha dog ngan cing chu® ba rams kyang kha dog ngan la smad pa’i 
’os® sué gyur te fe mngon par" dga’ bar ’ gyur' ba ma yin pa dang A padma’i® snying 
po de dag la yang de bzhin' gshegs pa’i sku re re skyil’ mo krung!? bcas te’! bzhugs 
shing’” ’od zer’? *bum™ dag rab tu ’gyed" pas’ shin tu° mdzes par? kun tu’ snang" ba’ 
*di’i rgyu ni gang yin / rkyen ni gang yin snyam" mo’ //” de nas thams cad dang Idan 
pa’i® byang chub sems dpa’i tshogs dang /’° ’khor bzhi po the tshom’® du gyur pa 
rnams ’dong!” bar!® bya ba’i mtshan ma byas so’” //” de’i tshe” tsan dan” gyi snying 
po’i”? khang pa brtsegs*’ pa der /° byang chub sems dpa’ sems dpa’ chen po rdo 
rje’i”! blo gros shes” bya ba ’dus par gyur” te“! *dug® go” // 

' BQ: Apis. oculi). '6 BQ: rtsom; JP;2ST: tsom. 

2 BJQ: tsom. 8 BIQPya3T: pad mai. BOP): ‘dod for ‘dong; P3: 

3 BJQP2;T: pad ma; N: pad > Pros: dhyil. ‘dor for ‘dong; (see Bth: nye 
for padma. 10 5QP;: dkrung. bar ong). 

4p,,: ni for gi.  p.o3: te /; Q: lacuna of one i‘ BQT: par. 

5 PT: mdab. letter between te and Tp: byaso. 

° LST: chung for chu; P3: bzhugs. 20 BINQ: om. ‘i. 
cung; (see Bth: sbu gu for Poo: shing |. 27 Dios: om. ‘i. 
chu ba). 3 pS: gzer. 2 NP 193: zhes. 

7 P: om. passage from bcom 4 D3: pas /. Pio: ‘gyur. 

(01.1) till gyur te / (aberratio 5 BQ: om. /. 


1 JsNaPuQiOsMiYu: FREE for FEAR. 


PDF Version: §PPB VI (2002) 


ci’i rgyu ci’i rkyen kyis! pad mi! lo ma’i bye ba khrag khrig brgya phrag stong pa ’di 
khatog ngan par gyur : sbu gu khatog ngan cing smad pa dang : mi bzang bar gyur la 
pad ma’i snying po “di las : de bzhin gshegs pa’i sku rere dkyil mo grung beas de 
bzhugs nas : od! gzer brgya phrag stong rab tu gtong zhing shin tu bzang bar kun tu 
snang zhes the rtsom du gyur to : de nas beom Idan ’das kyis byangchub setns dpa’ 
mang po tharns cad dang : *khor thetsom du gyur pa nye bar ’ong ba’i phyir Itas 
byaso! de nas de’i dus na yang tsan dhan kyi snying po khang bu brtseg pa der 
rdorje’i blogros zhes bya ba byangchub sets dpa’ serns dpa’ chen po "dus pa "dug 
go // 

' pad mi at the beginning of the line, small. 




1 )) inserted according to Ji; (cf. parallel in 0J). 


PDF Version: BPPB VI (2002) 


de nas beom Idan ’das kyis’ byang chub sems dpa’ sems dpa”* chen po” rdo rje’i? blo 
gros la bka’ stsal pa°/ rigs kyi bu khyod kyis‘ chos can gyi gtam’ las rtsoms* te /* de 
bzhin gshegs pa dgra becom pa yang dag par rdzogs pa’i sangs rgyas la yongs su® zhu 
bar spobs® par gyis' shig //® de nas byang chub sems” dpa’ sems dpa’ chen po rdo 
rje’i’ blo gros /* bcom Idan ’das kyis gnang bas” Iha dang /’° mi dang /"! Iha ma yin 
dang beas pa’i ’jig rten dang / byang chub sems dpa’ thams cad dang! / ’khor bzhi po 
dag gi the tshom’? gyi zug) mgu* rig! nas /? beom Idan ’das la ’di skad ces gsol to // 
bcom Idan ’das jig rten gyi khams ’di™ thams cad padma’* kha dog ngan cing dri 
mi" bda’’* ba di Ita bu bye® ba khrag khrig phrag bum ’di dag gis? khebs pa dang (4 
de dag gi dbus’ na yang’ de bzhin gshegs pa’i sku’® skyil!’ mo krung' bcas shing” 
bzhugs’ te /” ’od zer'® *bum* dag rab tu ’gyed par kun tu’ gda’”’ ba dang / de bzhin 
gshegs pa’i sku”’ de dag mthong nas kyang /”' srog chags bye ba khrag khrig” phrag 
"bum thal” mo sbyar te’” phyag ’tshal ba’i rgyu ni gang lags / rkyen™ ni gang lags / 

PL: kyl. shad; T: /I. 16D: sku’i dhyil. 

? Pio3: om. ‘i; N: rda rja’i for > Pins: bas /. 7 P,): dkyil; P3: s- of skyil 
rdo rje’i (due to lack of 10NP 13: om. /. with a very small letter 
space above the letters). " BOP: om. /. superscribed later. 

3 Pj): gtan. 2 BIQP 12: tsom [P2: -m of 18D: gzer for zer; P2: gzer 

4 BQ: brtsoms; P;: stsogs for tsom with a small letter pDhrag ‘bum. 
rtsoms; P2: stsoms; T: risogs beneath the line]. ® B: gang for gda’; Pip: dga’ 
for rtsoms. BD: om. /; Q: spungs shad for gda’ (metathesis). 

> BP»: om, te /; J: om. /; Q: de (after first syllable). 20 Pi: sku’i de. 
for te /; T: // for /. 4 BIQP aT: pad ma. a1 BJNQ: om. /. 

®NQ: sbobs. 5 BOP 3: gda’ for bda’; P2: 22 Dios: te /. 

a LP 23 om. 7. mnga’ for bda’: see Bth: 

8 BINQP,:: om. /; P3: la for zhim. 



| £& between 4H and 4 omitted according to Ji; (four-syllable rhythm). 


PDF Version: BPPB VI (2002) 


de nas beorn Idan ’das kyi rdorje’i blo gros byang chub serhs dpa’ °*7*") dpa’chen po 
la bka’ stsal pa : rigs kyi bu de bzhin gshegs pa dgra beom pa yang dag par rdzogs 
pa’i sangs rgyas la : khyod kyis rtsob par bgyi la chos dang Idan pa’i gtam rtsom 
lagso : de beoth Idan ’das kyis bka’ rdorje’i! blogros byang chub serns dpa’ sets 
dpa’ chen pos lha dang mi dang : lha ma yin dang ’jig rten du beas pa dang : ’khor 
bzhi the tsom tu gyur pa ’i zug rngu khong du chud nas : bcorn Idan ’das la ’di skad 
ces gsol to : beorn Idan ’das! ci’i rgyu ci’i rkyen kyis ’di’i ’jig rten kyi khaths tharns 
cad ’di ’dra bar pad mo bye ba khrag khrig brgya phrag stong gis kun tu g.yog cing : 
"di Ita bu’i khatog ngan cing dri mi zhim pa’i dbus na : de bzhin gshegs pa’i dkyil! 
mo grung bcas de bzhugs nas : ’od gzer brgya phrag stong rab gtong bar kun tu snang : 
de bzhin gshegs pa’i sku ’am : ’di mthong nas : srog chags bye ba khrag khrig brgya 
stong thal mo sbyar de phyags ’tshal : 

HA, Se. SERUM, AS. the, LMTAR—D 


PDF Version: BPPB VI (2002) 

de nas de’i tshe byang chub sems dpa’’ rdo rje’i? blo gros kyis tshigs su’ bcad* pa "di 
dag” gsol to //° 
[0.1] sangs‘ rgyas stong phrag bye ba mi g.yo bar® //** 
padma’ dag® gi® dbus na” bzhugs par! ni //” 
5 khyod! kyis ’di ’dra’i rdzu‘ ’phrul ston! mdzad pa //™ 
bdag gis” sngan® chad’ nam° yang ’di'? ma mthong!! //” 
[0.2] °od zer* stong” mams rab tu ’gyed® mdzad cing //'* 
sangs rgyas' zhing" ’di’ thams cad khebs!? par mdzad //"* 

ngo mtshar chos!° kyi mams la” rol mdzad pa’i //* 

10 *dren’ pa mams kyi!® bar chad ma mchis mdzes!” //? 

' JN: dpa’irdo. bdag for dag. “cover completely”; see 

2 LPo5: om. ‘7. 7 BP\3: / (due to following 0G.7f. and 0J.9). 

I LT: tshigsu. kh-?). “BT: /. 

* P, generally inserts only a . BINQP 123: sngon. 15 D..5: tshogs for chos. 
single shad between the ° ST: cad. 1s INLSTP)23: Ayis. 
padas of the verses; Ie Pi2: om. ’di. \ BJINQ: mdzod for mdzes 
deviations from this rule are 1 p.,: mthong ngo //. (contrary to Ch,: +£#% and 
noted in the apparatus. ? BQ:/. Chy: Sagi). 

> BIQP 237: pad ma. 3 BINQ: khengs for khebs 

® BINQ: gang for dag; P}: (contrary to Ch,: 3H, 



[0.1] KREAGR tee SA 

[0.2] SRGRBOt Hatz 
Beason Hee TR 


PDF Version: BPPB VI (2002) 

de nas rdorje’i! blogros byangchub sems dpa’ sems dpa’ chen po de’i dus na : tshigsu 
bead pa ’di bshad do: 
[0.1] bdag gi sngon ni di Itar yang ma mthong : 
de yis ci ’dra rdzu ’phrul stobs byas pa: 
sangs rgyas bye ba stong ni "dir gnas pa: 
pad mo rnarns! kyi dbusu mi g.yo bar // 
[0.2] ’od gzer stong gi mam par rab tu gtong // 
sangs rgyas kyi zhing kun rab tu g.yo bar byed : 
ngomtshar chos |rnams, kyi ni mam par rol // 

*dren pa mars ni rtag tu +++ po’i // 


mir chleShe. Bale itteel. 


[0.2] MICHAMPER Ae Al 

\[=Jig]<3K (cf. parallel in 0.4c). 




PDF Version: BPPB VI (2002) 

[0.3] mdab!*® ma dag dang™ chu ba smad”® ’os® la” //' 
kha dog ngan pa’i padma”’ mams dbus der //”! 
de dag rin chen rang” bzhin ’drar® bzhugs”’ pa // 
ci’i* slad du rdzu ’phrul®® ’di dag sprul //** 

°i°° bye snyed mthong”” // 

[0.4] bdag gis™ sangs' rgyas! gangga 
de yi! rdzu ’phrul khyad ’phags bdag™" mthong ste™ //”8 
de ring gda’” ba’i®° mam sprul ci ’dra ba //°P 
sngon chad“? nam” yang "di ’dra® di” ma™ mthong //” 

{0.5] rkang gnyis gtso™ bo lhas”” ni bstan du gsol // 
tgyu’* gang rkyen gang lags pa bshad du” gsol // 

’jig rten don mdzad thugs brtse” gsung™ du”? gsol //?! 
lus can kun gyi“ the tshom?? dgum?? du gsol //** 

'8 BINQS: ‘dab. 4 INQST: ci yi for ci’t [Q: yi erased te; P3: om. mthong. 

2 BP;: / (due to following compressed]. a Pi2: om. pada 0.4b; QT: /. 
kh-?). 5 NT: /; S: double spungs °N: dga’ for gda’ 

2° BJQP 13T: pad ma [P,: p- of shad (after first syllable). (metathesis). 
pad written with a small 6 B: gang gi’i for gangga ‘i; 3° BNP 93: /tas for has 
letter beneath -d]. JQT: gang ga’i; NS: (contrary to Ch): A). 

71 BQ: /. gangga i; P23: gang gai. 31 QT: /. 

2 D.: ngar for rang 27D: mthong (ste) // [ste >2 BO: rtsom; JP 12: tsom. 
(metathesis); P2: dar for marked with dots above for 3 BO: ’gum; Py: gtum for 
rang. deletion; P2: between dgum; P2: gcad for dgum; 

3N: bzhut; P23: zhugs. mthong and shad partially P;: dkum. 


[0.3] HiCZES BARR 
[0.4] Fai A He 

ARE ARIS BI allah 

1 JsNaPuQiOsMiYu: ERRE for FER. 
2 Dh FsKuMiSo Yu: 8 for {. 




PDF Version: BPPB VI (2002) 


[0.3] de dag maths kyang de la dga’ ba ste :) 
ci ltar ngan tog ngan pa pad mo’i d++s na bzhugs // 
lo ma mi sbu’u smrad pa ste // 
ci’i slad du rdzu ’phrul mam par ’phrul : 

[0.4] bdag mthong sangs rgyas chu bo gang ga byed // 
bd++ ’di Ita bu ni sngon ma mthong : 
mam par ’phrul! pa ci dra bdag la snang : 

[0.5] rkang gi m+g gi bdag la bshad du gsol // 
ci’i rgyu ci’i rkyen kyi snang : 
jig rten don dang thug brtse phyi ru gsung // 
srog chags tharns cad dgod pa bshad du gsol // 


[0.3] J@xO eR PRET AE 

[0.4] KARINE Ais 


2 Jin: HE for HE. 



PDF Version: BPPB VI (2002) 

de nas bcom Idan ’das kyis' byang chub sems dpa’ sems dpa’ chen po rdo rje’i" blo 

gros la sogs” pa thams cad dang Idan pa’i byang chub sems dpa’i tshogs la bka’ stsal 
pa /* rigs kyi bu dag” de bzhin gshegs pa’i snying® po zhes bya ba’i® mdo* shin tu 

rgyas pa® yod de* / de rab" tu bstan pa’i phyir /° de’ bzhin gshegs pas snang ba’i 

gros dang" / thams cad dang Idan pa’i byang chub sems dpa’i tshogs* des /? bcom 

Idan ’das kyis ’di skad ces“ bka’ stsal to //"! 

ST: hyis /. 

mtshan ma ’di Ita bu ’di byas so* //' de’i phyir™ legs par rab tu nyon la yid la® zung 
Idan ’das la legs so! zhes gsol te’ /" bcom Idan ’das kyi ’*Itar nyan” pa dang / bcom 
LPj3: om. 1; Q: I 

e . 
shig’ dang” bshad" do° /” byang chub sems dpa’ sems dpa’ chen po rdo rje’i? blo 

compressed, inserted later 
3B: lasogs for la sogs; Q: las 

yode; T: yod do 

5 BDINQP23: om. /; T 
for sogs 

sogs for la sogs; P,2: stsogs 
P12: yin te for yod de 

spungs shad (after first 

° BDJNQ: om. /; L: spungs 
shad (after first syllable) 
 LTP,3: legso [P; originally 
legs so altered to legso 
$ BJQ: om. yid la 
TNS: zhig 
i § BDINQ: dang ngas bshad: 
contrary to Ch): 4; L P,: dang / ngas bshad. 

through erasion of s(o)]; N. 
legs so legs so zhes. 
" Bp: pa / for to // 

ae, CTE 0 

il S 

we, ate, 

apd mo aeigieta aig BAT, KAS RAR. AUR 

BBR o KS o ERK, RSE AK To {8 


PDF Version: BPPB VI (2002) 

de nas bcom Idan ’da,! kyi rdorje blogros byangchub serndpa’ serns dpa’ chen po 
dang : snga ma’i byangchub sems dpa’ mang po thams cad ’tshog pa bka’ stsal pa : 

[2471] mdo 

rigs kyi bu de bzhin gshegs pa’i snying po zhes bya ba rab tu rgyas pa’i 
sde de bshad pa’i de Ita bu’i snang ba’i lhas byas kyi de shin yid la byol la : leg pa 
nyon cig dang bshad to : zhes nas dge’o : bcorn Idan ’das shes de rdorje’i blogros 
byangchub sens dpa’ setns dpa’ chen po dang : byangchub serns dpa”! seths dpa’ 

chen po : mang po de dag thams cad la : bcorn Idan ’das kyi di skad ces bka’ stsal to : 

&, BERKE. he. 

| inserted according to Ji, (four-syllable rhythm). 


PDF Version: BPPB VI (2002) 


rigs kyi bu’ dag ji Itar' de bzhin gshegs pas” sprul pa’i padma” kha dog ngan pa /? dri 
nga ba / smad par ’os pa /* mngon par dga’ bar® ’gyur ba ma‘ yin pa’ ’di dag dang® / 
de bzhin gshegs pa’i” gzugs mdzes pa / gzugs' bzang ba’ / blta* na sdug pa dag 
padma’i® snying po ’di dag la skyil’ mo krung’ beas shing ’khod de /™ ’od zer® bum 
dag rab tu ’gyed cing ’khod" pa de dag kyang rig® nas / lha dang’? mi mams? phyag 
tshal zhing? /!" mchod pa’i las kyang byed pa de bzhin du’ /'* rigs kyi bu dag!’ de 
bzhin gshegs pa dgra bcom pa yang dag par rdzogs pa’i' sangs rgyas kyis kyang /"“ 
rang gi shes rab dang /" ye shes dang / de bzhin gshegs pa’i mig’ gis'° sems can srog 
chags su’ gyur pa” / *dod* chags’ dang /'” zhe sdang dang / gti mug dang /“ sred™ pa 
dang /*® ma rig® pa’i nyon mongs™ pa bye ba phrag "bum gyi sbubs su® gyur pa 

thams cad dang fe 

D>: Ha. deletion]. S17 st: spyan for mig (that 
? BDIQP123T: pad ma; (LSN:  Pios: dhyil. spyan is here original is most 
NF), 8 p,S: gzer. unlikely: no revisor would 
3 4 BDJNQP\23: om. /. correct the honorific form 
Py23: om. /. 10 i 
4D om. / NP;: dang /. spyan in the case of the 
5 PT: pa a 1 DINP 23: om. /. Buddha’s vision to mig. 
qi: a 12 Ps 6 ° io 7 
6 BDIOQP,25: pad mai; T: om. . BDINQE zs: om. I ae gyis; S: gyi. 
es : Py2: om. rigs kyi bu dag; Py: om. /. 
pad: ma’i (ma) snying for Teoma 

padma’i snying [ma marked 

14 . 
with a dot triangle above for BDINQP 12s: om. /- 



" DhoFs: Hil for HI. 


PDF Version: BPPB VI (2002) 


rigs kyi bu di Ita ste : dper na : de bzhin gshegs pas sprul pa’i pad mo ‘di dag la 
khatog ngan zhing : dri mi zhim la slad! cing mi bzang bar gyur la pad moi! de dag 
thams cad la snying po ’di dag la yang : de bzhin gshegs pa’i sku dkyil mo grung 
beas te bzhugs nas ’od gzer brgya phrag stong rab ++ tong zhing bzang zhing sdug la 
gzugs dang Idan par! bzhugs par Iha dang mi maths kyi? de dag shes nas phyag ’tshal 
zhing mchod pa’i las byed do : rigs kyi bu de Itar de bzhin gshegs pa dgra bcoth pa 
yang dag par rdzogs pa’i sangs rgyas rang gi shes rab kyi ye shes dang : de bzhin! 
gshegs pa mig gis semns can srog chagsu gyur te’ : "dod chags dang : zhe sdang dang : 
gtimug dang : sred pa dang ma rig pa dang : nyon mongs pa bye ba brgya phrag 
stong gi mdzod du mthong ngo : 

’ mo with a very small ’a chung beneath and a symbol similar to a double ‘greng bu over the na ro; 
parallel to pai (pa with ‘a chung beneath and ‘double ’greng bu’) representing pa’i at the end of the 

? Between kyi and de dag excised ston. 

3 Not clear: fe or fo? 


fe, MUP, A, FEA, RR. Wen, BAT. WR, 





PDF Version: BPPB VI (2002) 

rigs kyi bu dag sems can nyon mongs pa’i sbubs® su™ gyur pa de dag gi nang na /'® 
nga’? dra bar ye shes”’ dang Idan pa / mig dang Idan pa’i de bzhin gshegs! pa mang 
po ’khod cing”! skyil”? mo krung! beas nas /*? mi“ g.yo bar" ’khod™™ pa™ mthong 
ste / nyon mongs pa thams™ cad kyis nyon mongs pa?? can“ du gyur pa de dag gi 
nang na /** de bzhin gshegs™ pa’i chos nyid” mi g.yo zhing /”° srid® pa’i ’gro ba 
thams cad kyis ma gos pa dag" mthong nas /”’ de bzhin gshegs pa”* de dag ni nga” 
dang ’dra’o*° zhes smra’o*' //™ rigs kyi’’ bu dag” de bzhin gshegs pa’i mig” ni® de 
Itar®? mdzes pa yin te /Y de bzhin gshegs” pa’i’* mig®® des** sems can™* thams cad 

de bzhin gshegs pa’i’”’ snying por*’ mthong ngo //°* 

18 BDJNQ: om. /; P2: na / ma 
gos pa dag nga (in the ms. 
which served P, as original, 
ma gos pa dag could already 
have been placed between 
the lines and became thus 
wrongly inserted by the 
copyist of P2; cf. endnote tt). 

ue BQ: de for nga. 

© BQ: shes de dang. 

21 Dias: cing /. 


2 D.o5: digil. 

? BDINOQP;: om. /. 

24 BDINQ: om. /. 

5 pi3: om. nyid. 

26 BDINQP 23: om. /, 
27 BDINQ: om. /. 

8 BQ: pa’i de. 

2° BQ: de for nga. 

3° BQ: ‘dra‘o // zhes. 
31 BQ: smras so for smra’o. 
32 LST: spyan for mig. 

3 LST: Itas (see Bth: de Itar). 

47ST: om ‘i; Pz: om. pa’i 
(see Bth: pa’i). 

> LST: spyan for mig. 

a Py23: des |. 

37 BDINQ: snying pocandu 
mthong (see Bth: snying por 
mthong; see the disussion of 
the term tathagatagarbha in 
1A 2; cf. 1A.8f and 1B, fn. 




PDF Version: BPPB VI (2002) 


rigs kyi bu serns can de dag nyon! mongs pa’i mdzod du gyur pa’i dbus na : nga dang 
*dra ba ye shes dang Idan pa de bzhin gshegs pa mang po’i dkyil mo grung bcas te : 
mi g.yo bar bzhugso : nyon mongs pa thams cad kyis nyon mongs pa de dag gis dbus 
na :) de bzhin gshegs pa’i chos nyid mi g.yo zhing : byung ba tharhs cad kyi rgyud 
kyi ma gos par mthong bas na : de bzhin gshegs pa de bzhin gshegs pa de dang 
*dra’o : zhes smra’o : rigs kyi bu de Itar de bzhin gshegs pa mig! bzang po de bzhin 
gshegs pa’i mig des serns can thams cad de bzhin gshegs pa’i snying por mthong ngo : 


iB 0 



PDF Version: BPPB VI (2002) 


(Bu,igs* kyi bu® dag! di Ita ste” dper® na /** skyes bu‘ Iha’i mig can la la° zhig gis /° 
lha’i mig gis® ’di Itar’ kha dog ngan cing /” ’di Itar® dri nga ba’i padma®? kha ma 
bye!® zhing /"! ma® gyes! pa dag la bltas! te / de’i dbus na“ padma’i!” snying po la de 
bzhin gshegs pa skyil’? mo krung! beas shing ’dug par rig nas /'4 de bzhin gshegs 
pa’i™ gzugs '**blta!® bar" ’dod de /'” de bzhin gshegs pa’i gzugs*! de yang dag par® 

22 723 omad 

sbyang ba’i!® phyir /” padma’i?? mdab*!? ma kha dog ngan cing dri nga la 
par ’os pa de dag ’byed! cing sel ba de bzhin du /** rigs kyi bu dag”* de bzhin gshegs 
pas kyang”® sangs rgyas kyi mig gis”’ sems can thams cad de bzhin gshegs" pa’i 

snying por mthong nas /® sems can de dag gi’ ’dod chags dang / zhe sdang dang /”” 

gti mug dang /” sred? pa dang / ma rig’ pa’i nyon mongs pa’i sbubs dbye”! ba 
phyir chos ston” te* /’ de sgrub® pa’i 

nyid du gnas so”° // 



de bzhin gshegs pa mams ni” yang dag pa 

(Bu. Fom here till 1B.3 (°") included in the textcritical edition: the first of two quotations from the TGS 
in Bu ston’s mDzes rgyan f. 3a1-3a5. 

'P5Bu: om. dag. 

? P,Bu: ste /. 

> DIN: om. /. 

“BQ: / tha skyes. 

> BDJNQP3: om. /; S: spungs 
shad (after first syllable). 

§ pT: om. gis; Bu: om. lha’i 
mig gis. 

7 BDINQP}23Bu: om. /. 

® BDJQPyasT: pad ma. 

° P.: pad ma / [P): spungs 
shad (after first syllable]. 

 QBu: phye for bye. 

" BDINQP\23Bu: om. /. 

? BDIQP aT: pad ma’i. 

BP P23: dkyil. 

™ BDINQBu: om. /. 

5 p..: om. the passage from 
“15>” till “<-15” (aberratio 

1 D.Bu: Ita. 

7 BDINQ: om. /; L: spungs 
shad (after first syllable). 


18 BP: sbyang pa’i; L: 
sbyangs pa’i; Pi: sbyar ba’i; 
Q: sbyad ba’i. 

° BDINQBu: om. /. 

2° BDIQP i237: pad ma’i. 

71 BDJNQSBu: ‘dab. 

22 BDINQ: ba for la: BDINQ 
assimilated to 0M.1f: ... ngan 
pa! dri nga ba / smad par 

‘os pa | mngon par dga’ 
bar.... In OM.1f the three 
attributes are set one after the 
other. Here, the coordinating 
particle cing, by establishing 
a close relation between the 
two first members, requires 
the governing function of 
smad par ’os pa with Ia. 

2 DINP}23Bu: om. /. 

4 DINQ: om. / [Q: lacuna of 
about seven letters between 
du and rigs]. 

5 Di3: om. dag; Bu: dag /. 

26 BT: om. Ayang; P3: kyang /. 
27 LST: spyan gyis for mig gis. 
28 BDINQ: om. /. 

?° BQ: om. /; S: shad crossed 
out with a wriggled line? 

30 DINL: srid. 

31 BQ: 'byed for dbye. 

32 BQ: pa’i; Pz: om. ba’i. 

33 BQ: bsgrubs [B: bsgrub, at 
the end of bottom line of the 
folio]; T: bsgrub. 

4 D>: sbubs dbye ba’i for 
sgrub pa’i; contrary to Ch2: 
IE{E17; S: ba’i for pa’i. 

35 Pio: pa rnams ni yang dag 
parnams ni yang dag pa 
nyid (dittography). 

36 LP: gnaso [P; at the end of 
the line]. 


| Dh Fs: fil for Of. 


PDF Version: BPPB VI (2002) 


rigs kyi bu ’di lta ste : dper na skyes bu zhig gi ’phral kyi mig gi de Ita’i khatog ngan 
pa dri mi zhim pa pad mo’i snying po de la! ’das pa pad mo’i snying po de’i nang na : 
de bzhin gshegs pa bzhugs na ni : de bzhin nyid mthong zhing shes na : de bzhin 
gshegs pa’i sku mthong bar ’dod de : khatog ngan pa dang : dri mi zhim par smod 
pa’i pad moi’ “*"! Jo ma hag par byed cing sel ba ni de bzhin gshegs pa de bzhin 
gshegs pa de’i sku sbyangs ba’i phyir ro : rigs kyi bu de Itar de bzhin gshegs pas : 

! serns can tharns cad de bzhin gshegs pa’i snying 

sangs rgyas kyis mig gis serns can 
por mthong ste : serns can de dag ’dod chags dang : zhesdang dang gtimug dang : 
sred pa dang : ma rig pa dang : nyon mongs pa’i mdzod sbyangs pa’i phyir : chos 
bshad de gang nan tan! byed pa ni de bzhin gshegs pa yongsu dag par gnas so : 

' mo immediately followed by ‘a with a symbol similar to a double ‘greng bu above. 





? 88 [Gani] #8. 


PDF Version: BPPB VI (2002) 


rigs kyi bu dag*® ’di ni° chos mams kyi chos nyid de’ / de bzhin gshegs pa rnams 
byung yang rung’ ma byung yang rung’ / sems can ’di dag ni’ rtag tu de bzhin 
gshegs pa’i snying po” yin na’8" / rigs kyi bu dag smad par’ ’os pa’iX nyon mongs 
pa’i sbubs' mams™ kyis" yog’ pas /° de dag gi? nyon mongs pa’i? sbubs gzhig" pa 
dang / de bzhin gshegs pa’i® ye shes kyang* yongs su‘ sbyang ba’i" phyir /° de bzhin 
gshegs pa dgra bcom pa yang dag par rdzogs pa’i sangs rgyas byang chub sems dpa’” 
rnams la chos ston te /~ bya ba ’di la yang* mos par’ byed do //* de la byang chub 
sems dpa’ sems dpa’ chen po chos de dag la’ mngon par brtson par gnas pa de dag 


gang’ gi tshe /” nyon mongs pa dang / nye ba’i nyon mongs pa thams cad las yongs 

su“ grol bar gyur pa® de’i tshe /* de bzhin gshegs” pa dgra becom pa yang dag par 

10 > >jhh 

rdzogs pa’i sangs rgyas shes” bya ba’i grangs su’° ’gro ste /®® de bzhin gshegs pa 

bya ba thams cad kyang byed do! /# 

' BDINQP;Bu: rung /; LT: T: yogs. 7 BDINQ: om. /; S: spungs 
rung //. * Pi: om. kyang. shad (after second syllable). 

2 DLSTP;: po can yin (see s BDJNQ: om. / [J: lacuna of : BDJNQ: om. /; L: spungs 
Bth: snying po’o; cf. also one letter between phyir and Shad (after first syllable). 
1A.8f. and OM, fn. 37). de}. ° DNP;S: zhes. 

3 BDINQ: g.yogs; Po: g.yog; 6 Pi): om. gang. 0) P): grangsu. 



’ Dha: fa SHEE for (SMA. 


PDF Version: BPPB VI (2002) 


rigs kyi bu chos mams kyi chos nyid de bzhin te : de bzhin gshegs pa byung ngam : 
de bzhin gshegs pa ma byung yang serns can de dag thas cad ni! de bzhin gshegs 
pa’i snying po’o : rigs kyi bu de nas yang smon pa nyon mongs pa’i mdzod kyi 
bsgrub pa : nyon mongs pa de dag gis mdzod gzhom pa’i phyir de bzhin gshegs pa’i 
ye shes yongsu sbyangs! pa’i phyir : de bzhin gshegs pa dgra bcorh pa yang dag par 
rdzogs pa’i sangs rgyas byangchub sems dpa’ rnams la chos ’chad do : de la 
byangchub seths dpa’ chos + dag la nan tan du byed cing gnas te] nam nyon mongs 
pa tharns jcad, dang : nye bar nyon mongs pa thams cad las yongsu grol bar gyur pa 
de’i tshe : de bzhin gshegs pa dgra bcom pa yang dag par rdzogs pa’i sangs rgyas kyi 
grangsu ’gro’o : tharns cad kyang! de bzhin gshegs pa’i bya + byed do : 





PDF Version: BPPB VI (2002) 


de nas de’i tshe* bcom Idan ’das kyis tshigs su’ bead pa ’di dag bka’ stsal to // 

[1.1] ji ltar padma”” smad par’ ’os gyur’ pa? //*° 
de’i® mdab’ ma sbubs gyur ma* gyes' la //8 

de bzhin gshegs® pa’i snying po ma gos" te’ //"° 

mi ’ga” la las lha yi'' mig gis mthong // 
[1.2] de ni de’i'** mdab!? ma ’byed pa na // 

dbus na rgyal ba’i lus’* ni’ mthong gyur nas //™ 

rgyal ba" nye ba’i nyon mongs phyir’> mi° ’gyur® //1 
de ni ’jig rten kun tu’ rgyal bar ’gyur’ // 

LN: tshigsu. 

2 BDJQP);T: pad ma; P»: dam 
pa ma for padma. 

> Pio: om. par; P3: pa’i for 

“P.: cf. OK, fn. 4. 

> BQ: /. 

6 DINQST: de yi for de’i [Q: 
yi compressed]. 

"BDINQS: ‘dab. 


® BQ: /; T: spungs shad (at the 
end of the line), 

* DIN: de for te. 

BOT: /; P3: double spungs 
shad (after first syllable). 

" BP 37: Lha i; Q: yi 

” DINQT: de yi for de’i [Q: 
yi compressed]; P3: nga’i for 
de’i; P2: padma’i for de’ 

[1.1] Seaneeeeye He 


with small letters beneath the 

® BDINQS: ‘dab; T: ’bad for 

4 LST: sku for lus; parallel to 
“spyan for mig” (OM, fn. 15). 

5B: phyes; DINQP}2: phyis. 

PDF Version: BPPB VI (2002) 

de nas bcom Idan ’das de’i dus na tshigsu bead pa ’di dag bshad do // 
{1.1] ci ltar pad mo spyad pa de ni 
de’i lo ma mams kyi +++ ma gyes / 
de bzhin gshegs pai! snying po de yang ni // 
dri ma ma gos +skyes pas ’phrul mi mthong // 
[1.2] de’i lo ma de rnarns sel bar byed : 
de’i nang na sku yang mthong : 
phyis ni rgyal la nyon mongs ’byung ste + 
rgyal ba ’di ni ’jig rten kun |duy ’byung! ste : 


Ra Ear tA 



1 Jip: All for BU. 



PDF Version: BPPB VI (2002) 


[1.3] de bzhin ngas’® kyang srog chags thams cad kyi' /* 
dkyil der gnas pa’i’ rgyal ba” mams* kyi’ lus // 
ji ltar smad ’os” padma’i’’ sbubs ’dra™ ba’i //” 
nyon mongs” stong phrag bye bas khebs“ pa’® mthong //* 
[1.4] nga’? yang de dag gi” ni bsal”! ba’i phyir //* 
mkhas pa rnams la rtag tu chos ston® te”? //” 
sems can ’di dag sangs rgyas ’gyur bya zhes // 
rgyal ba’i phyir ni nyon mongs rnam par sbyong”* //™ 
[1.5] nga yi’ sangs rgyas mig” ni!’ de”’ "dra! ste //* 
de”® yis”® rgyal ba’i lus su’? gnas pa yi?! //" 
sems can ’di dag thams cad mthong gyur” te // 

de dag rnam par sbyang™” phyir chos smra’o //™ 

16 p.,: ngan for ngas. 2 BDINP3S: to. line]. 

7 BDIQPy3T: pad ma’i; 3 OT: /, 8 LST: nga for de (see Bth: 
S: padmo ’i. 24 p.,: spyod for sbyong. de ru). 

'8 BDINQ: par. > B: ni for yi; JP3: nga’i for 2°-N: yi for yis; P13: de’i for 

® BQ: de for nga. nga yi. nga yis. 

2° DIN: gis. 26 1 ST: spyan for mig. 30 JL: lusu; P»: la for su. 

21 DINQP;: gsal; P,: ga la for 7 Diy: ‘di for de [P2: ‘d(i) 31 pio: yin for yi. 
bsal. with small letters beneath the 32 Do: ‘gyur. 


eee LC 
[1.4] HRA PURI 
Ratha SRE 
[1.5] FxLi RAR FL —UsRAS 

1 D\LDh,FsKuMiSoYu]<-&. 


PDF Version: PPB VI (2002) 

10 [1.3] bdagis srog chags : kun mthong ba’i snying po 
de la rgyal ba’i sku mams mthong ba’i gnas : 
nyon mongs pa bye ba stong gis kun tu khebs : 
de bzhin pad mo’i 48°") snying po smad pa ’o : 
[1.4] bdagis de dag gsal ba’i phyir : 
15 rtag tu mkhas pa rams la chos kyang bshad : 

nyon mongs pa ni mams rgyal ba irgyaly ’i jthamns, phyir : 

ci ltar setns can ’di dag sangs rgyas par ’gyur : 
[1.5] di ltar ’di ’dra ’i sangs rgyas! mig : 
nga’i serns can ’di dag kun mthong ba // 
20 de ru rgyal ba’i sku yang rab tu gnas // 
de dag ram par sbyang ba’i phyir : chos kyang 





2 ay <4 (cf. parallel in 1.1a etc. ( @)ER3E) as equivalent for smad par ’os pa). 



PDF Version: BPPB VI (2002) 


rigs kyi bu dag gzhan yang ’di’ Ita ste’ dper na /* sbrang tshang zlum po shing gi yal” 
ga la’ *phyang® ba" //* bung ba® ’bum gyis' kun tu® bsrungs" shing! //° sbrang rtsis 
yongs sw’ gang ba zhigp* yod la //* de nas sbrang ‘rtsi ’dod pa’i mi zhig gis® srog chags 
kyi mam pa bung ba! de dag thabs mkhas pas” bskrad'® nas!! /’? sbrang rtsi des” 
sbrang rtsi’i” bya ba byed do” //° rigs kyi bu dag de bzhin? du sems can thams cad 
kyang sbrang tshang dang "dra ste / de la’ sangs rgyas nyid nyon’ mongs pa dang / 
nye ba’i nyon mongs pa bye ba phrag ’bum gyis® shin'* tu' bsrungs pa! /'° de bzhin 
gshegs pa’i ye shes mthong" bas rig’ go’”™ // rigs kyi bu dag ji Itar sbrang tshang gi'® 
nang na /’° sbrang rtsi bung ba* bye ba phrag’ ’bum’ gyis”” kun™ tu” bsrungs pa”! 

/* sems can“ 

yod® par skyes bu mkhas pa zhig gis” shes pas” rig pa de bzhin du 
thams cad la yang® sangs rgyas nyid nyon mongs pa™ dang /*8 nye ba’i nyon mongs 

pa bye ba phrag ’bum gyis kun tu™ bsrungs' pa yod par de bzhin gshegs pa’i* ye 

shes! mthong bas rig ste“ / 

1B: ste //; DINQP»: ste /. bskrad. '? BDINQP)23: om. /. 

? BDINQ: om. /; L: //. 0 Di: bskad; P3S: skrad. 20 Dio: gyis /. 

3 P,: om. Ja; Pz: om. ga la. Mu BDJNQ: na for nas. 21 D4: bsrung ba; P2: bsrungs 

“DPT: /. ? BDINQP ys: om. /. ba. 

> BDINQ: om. //; Pi23T: /. 3 D135: om. sbrang rtsi des 2 Dios: gis /. 

° BDINQSTP;;: /. (aberratio oculi). 33 BDINQ: par for pas: 

7B: brtsi for rtsi; Q: lacuna of  BDINQP}23: kun for shin. against par cf. the parallel 
one letter between sbrang 5 BQ: srungs de for bsrungs constructions in 2A.8 and 
and rtsi. pa. 2A.12f. 

® Pi: gis //; Pas: gis /. 1 BDINQP)3: om. /. 4 BDINQP 123: om. /. 

? BP: pas /; P,: pas //; Q: 177 P.: rigo [P, at the end of SYP: pi [L at the end of the 
lacuna of one letter between the line]. line}; T: pa’ for pa 7. 

pas (without tsheg) and 18 ON: gis. 


ax, £57, PSE, RRHRBR. TH. RA 
A, S78, FRR, DRS, MERA, Be. Me, 


PDF Version: BPPB VI (2002) 


rigs gyi bu gzhan yang di Ita ste : dper na : sbrang rtsi’i stod zlum por gyur pa shing 
gis ya,! ga ’phyang ste : bung ba brgyastong gi srungs sbrang rtsi yongsu gang ba 
sbrang rtsi ’dod pa’i skyes bus ba srog chagsu gyur pa dag thams cad mkhas pas 
bskrad nas sbrang rtsi sbrang rtsis ’i bya ba byed do :] rigs kyi bu de Ita bas na di 
bzhin sems can thats cad kyang sbra rtsi ’dra ste : de bzhin gshegs pa’i ye shes kyi 
mthong bas : sangs rgyas su ’gyur -+nyon mongs pa dang : nye bar nyon mongs pa! 
bye ba stong phrag brgyas srung bar shes so : rigs kyi bu ci Ita bu sbrang rtsi’i snod 
kyi dbus na bar shes pa’i skyesbus sbrang rtsi bye ba phrag stong gis srung bar ye 
shes kyi shes so : de ltar 





1 Tin: HS for #. 



PDF Version: BPPB VI (2002) 

rigs kyi bu dag de’ la de bzhin gshegs pa yang" thabs” la” mkhas pas’ bung ba* bsal! 
ba de* bzhin du sems can® de dag gi ’dod' chags dang / zhe sdang dang / gti mug 

dang /° nga® rgyal dang /° rgyags pa dang /" ’chab pa dang / khro ba dang / gnod sems 

dang / phrag dog’ dang / ser sna la sogs® pa’i nyon mongs pa’ dang / nye ba’? nyon 

mongs! pa* mams bsal’? nas /" ji ltar sems can de dag la yang! nyon mongs” pa dang /* 

nye ba’i nyon mongs pa de’* dag gis° nye ba’i nyon mongs pa? can du mi ’gyur ba 

dang /* gnod par mi ’gyur ba de Ita de Itar’? chos ston to // de bzhin gshegs pa’i ye 


shes mthong ba de'* mam par sbyangs nas /> ’jig rten na de bzhin gshegs pa’i bya 


ba’ byed de‘ /'® rigs kyi bu dag’ nga’i" de bzhin gshegs pa’i mig'*’? yongs su” dag 

pa des /”’ ngas sems can thams’ cad de Itar mthong ngo” // 

'Di3: de dag for dag de. 

2 Pyo37: om. Ja. 

a P1293: pas /. 

4 BIJQP 123: om. de. 

> P..: om. gti mug dang /; 
BQ: om. /. 

§ BQ: om. /[Q at the end of 
the line]. 

Dp: phrag dogs; P23: phra 

dog; wrong decomposition of 
the abbreviation phradog 

(A551; see Bacot 1912: 415)? 

5 B: lasogs; Q: las sogs; Pip: 


stsogs for sogs. 

° B: nyon mong, pa dang / nye 
ba’i with small letters 
beneath the line (= correction 
of an aberratio oculi); P,2: 
om. nye ba’i. 

10 p..: bstsal; P3Q: gsal. 

u BDINQ: om. /. 

12 pS: om. de. 

31 P.: de ltar for de Ita de 
ltar, P;: de ltar da ltar for de 
Ita de Itar. 

n P23: de/. 

'S BDINQ: om. /; T: spungs 
shad (at the end of the line). 

16 NP,: de // [N: second shad 
of the nyis shad only half]. 

17D, T: om. dag; Q: lacuna of 
about three letters between 
dag and nga ’i. 

18 LST: spyan for mig. 

bd Pip: mig gis yongs. 

0 LT: yongsu. 

71 BDINQ: om. /; T: //. 



PDF Version: BPPB VI (2002) 


thabs pa, serhs can thams cad la sbrang rtsi byed pa bskad pa bzhin du sems can de 
dag gis ’dod chags dang : zhe sdang dang : gtimug dang : nga rgyal dang : drag pa 
dang : ’chag pa dang : khro ba dang : gnod sems dang :! ngan seths ’jungs pa "di la 
stsog, pa’i nyon mongs pa dang : nye bar nyon mongs pa med par bya ba’i phyir : de 
bzhin du chos ’chad do : ci Itar ci lta bu yang sems can de dag nye ba’i nyon mongs 
pas nye bar nyon mong,! par mi ’gyur ba dang : gdungs par mi ’gyur ba dang : de 
bzhin du de bzhin gshegs pa’i yeshes mthong ba de mam par sbyangs te : de bzhin 
gshegs pa’i byed pa ’jig rten du byed do : rigs kyi bu nga’i de bzhin! gshegs pa’i mig 

yongsu dag pas : ngas sems can ’di dag tharns cad mthong ngo : 


i. 1. OR 22). OR IRI, URI, A, BAIT 
} ee (cf. the sequence in the Tib.: ... mraksa (GH), krodha (3), vyapada (@), irsya 


PDF Version: BPPB VI (2002) 

de nas de’i tshe beom Idan ’das kyis tshigs* su’ bead pa "di dag bka’ stsal to //* 
[2.1] ji ltar ’di na sbrang tshang” yod pa la //*4 
bung bas* kun tu bsrungs” shing“ sbas° gyur pa‘ // 
mi gang sbrang® rtsi ’dod pas de® mthong nas’ //" 
de! ni bung ba! rab tu skrod® par byed //' 
[2.2] de bzhin ’dir yang sbrang tshang™ Ita bu’ ni //'° 
srid pa gsum gyi" sems can thams cad do° / 
de dag nyon mongs bye ba mang ba ste //4 
nyon mongs dbus na de bzhin gshegs”"’ ’dug’* mthong’ // 

'LT: tshigsu. ; BQ: srungs; P3: bsrangs; i Py2: bur. 
2 BT: /; N: spungs shad (after T: bsrung. id P;QT: /. 

third syllable). ° P.>: pa des for pas de. 1 pi: gshegs pa ‘dug. 
3p): ef. OK, fin. 4. 7 LP: na for nas. 2 p,,: ‘dug pa (m)thong. 
4 BT: /. , BQ: bskrod; P23: skyod. 

[2.1] SaaS 

| 4 Dh,KuMiSo YuJsNaPuQiQsZi (Fs illegible)}<—40] (see parallel in 2A: R(k¥e Be Me ...). 


PDF Version, RPPB VI (2002) 


de nas bcorn Idan ’das kyis de’i tshe de’i dus na : tshigsu bead pa ’di bshad do // 

[2.1] ci Itar sbrang rtsi’i sdod du gyur pa ni // 4981) 
bu ba marns kyi bsrungs shing sbas pa de : 
sbrartsi ’dod pa’i mi yis mthong nas su // 
bung ba mars ni skod par byed do : 

[2.2] ci Itar sbrang rtsi sdod ni ’di bzhin du // 
setns can thams cad ’byung ba gsurn na mthong : 
bung ba "de dag"! nyon mongs bye ba mang na : 
nyon mongs dbus na : de bzhin gshegs pa gnas 


PRS ER Hf Al. 

[2.1] SKE sees Pee 


1 At[=Jiz (E)]1BE (as usual in introducing the verses). 




PDF Version: BPPB VI (2002) 

[2.3] nga yang’ sangs rgyas rnam par sbyang”? don du //* 
bung ba” skrod”* pa Ita bur nyon mongs "sel // 
gang “gis'® nyon mongs bye ba’’ gnod ’gyur ba* //Y 
chos rnams” ’dir ni thabs™ kyis’® rab ston te // 

[2.4] de dag ji ltar de bzhin gshegs gyur la // 

*jig rten kun tu’ rtag tu’? bya”’ byed”! cing” //4 


spobs”” Idan ji ltar sbrang ma’i® sbrang™ rtsi yi //*® 

snod™ *dra’i" chos ston ’gyur bar” bya phyir ro! //** 

[2.3] Free A ieatikas 


[2.4] Rebs RSH HK 



3 LST: spyad for sbyang. 8D: kyi. the 9-syllable rhythm. 
BQ: bskrod. 1° LST: om. rtag tu contrary 2 B: sbobs; J: -bs of spobs 
'S BQ: sol for sel [B: g- of to Chy: # (see fn. 21). compressed; P13: spos for 
(g-)sol marked with a dot 0 LST: bya ba byed. spobs. 
triangle above for deletion]; 1 LST: byed ’gyur c-/zh-ing: 3p, T: ba for bar. 
Q: lacuna of one letter by the insertions of fn. 20 
between mongs and sol. and here LST compensated 
6 pi: gi. for the omission of rtag tu 
7 BDINQ: bas for ba. (see fn. 19) in order to keep 

PDF Version: BPPB VI (2002) 

10 [2.X] ngas ni: saryas : sangs rgyas ram par sbyangs pa’i phyir : 
bung ba ci Itar’ rtsi ’i sdod ni ’di bzhin du // 
seths can tharns cad ’byung ba! gsum na mthong : 
bung ba de dag nyon mongs pa bye ba mang : 
nyon mongs pa dbus na : de bzhin gshegs pa gnas // 
15 [2.3] ngas ni sangs rgyas rnampar sbyangs pa’i phyir : 
but++ ci Itar skrod pa de bzhin nyon mongs! pa gsal : 
thabs kyi chos mams kyang ni smra bar bya: 
nyon mongs pa bye ba de mams med par bya : 
[2.4] de bzhin gshegs pa ci Itar ’gyur ba de: 
20 rtag tu ’jig rten thams cad don kyang byed : 
ci Itar sbrang ma’i sbrang rtsis! sdod gyur pa // 
de Itar spobs shing chos kyang bshad par ro // 

 Aberratio oculi: copyist repeats from 2.2a: rtsi sdod.... 



[2.4] ZAHEER MMH aNe As 



PDF Version: BPPB VI (2002) 


rigs kyi bu dag gzhan yang ’di Ita ste* dper® na /! ’bras sa” lu° ’am / nas sam /‘ ci tse? 

’am’ / ’bru* mams* ni snying po phub mas? yongs su° bsrungs® pa" yin te /' de’ rang 

gi phub ma las* ma byung’ gi bar du’® bza’ ba dang / bca” ba“ dang / myang! ba’i bya 
ba mi byed kyi™ /" rigs® kyi bu dag skyes pa ’am!! / bud med gang dag? bza’ ba dang /1 
bea” ba la sogs” pa‘ zas skom gyi bya ba ’dod pa de dag gis brngas!? shing' brdungs” 
te /" phub ma’i sbubs” dang™ phyi* shun sel to! //'° 

'BDINQ: om. /. 

? BDINQST: sd for sa. 

3B: tsi rtse for ci tse; P\2Q: 
rtse for tse; T: tsi for tse. 

4 Py: ‘bru’i for bru: the 
genitive could be a survival 
of preclassical times, when 
the nominal character of the 
following rnams was still 
felt. Or should the genitive 
be explained in relation to 
the variant rnam pa in en. f? 

5 BDINQ: ma nas for mas. 

° LN: yongsu. 

TDI: om. de; P23: de’i for de. 


® BDINQ: nas for las (Bth: 

° BDINQ: phyung (required 
agent is not given); P,: 

° B: bar du bar du bza’ 

(dittography); P,: du //; 
P.: du/. 

" BDINQ: dang for ‘am 
(There is no reason for the 
preference of ‘am contrary to 
dang in the edition besides 
the general priority granted 
to L(ST). 

2 BQ: lasogs for la sogs; L: 


#8 Feel Ao 

| W[=KuMiSoYuJsNaPuQiQsZi]<¥s. 


la sog, (at the end of the 
line); P12: Jas stsogs for la 

3 DNQT: brngas or brdas (?). 

i P,2: dang /. 

> B: sol to for sel to; PT: te 
for to; P3: selo for sel to 
(before nyis shad at the end 
of the line). 

16D: spungs shad and rkyang 
shad (after first syllable); 
PyoT: /; S: double spungs 
shad (after first syllable). 

an ro") 3e. BRESELI, 

PDF Westone age VI (2002) 


rigs kyi bu gzhan yang ’di Ita ste : dper na ’bru ’am : nas sam : ci rtse ’am : so pa 
mams kyi snying po phub mas ++ par gyur te : nam phub ma las! yongsu ma byed 
pa’i bar du mi bca’ pa mi bzang : myong bar yang mi ’gyuro : rigs kyi bu bea’ ba 
dang : bzod pa dang : btung ba, bya ba’i bud med dam : skyes pas phyir zhing bat++ 

nas : phub ma’i nyon mongs pa! gsal te : phyirol ’dir ro : 

(gx, SRT, Bs, BR. SABRE. BPA, F 
BRA, BET, RAZA, 48, 4%, DAKAR, Mm 


PDF Version: BPPB VI (2002) 


rigs kyi bu dag de bzhin du’ de bzhin gshegs pa yang*” de bzhin gshegs” pa’i spyan 
gyis’ sems can thams cad° la de bzhin gshegs pa nyid /* sangs rgyas nyid° rang 
byung’ nyid / nyon mongs pa’i sbubs kyi® shun pas dkris’ shing’ gnas par® mthong 
ngo // rigs kyi bu dag de la" de bzhin gshegs' pa yang! nyon mongs pa’i sbubs kyi® 
shun pa‘ bsal? ba dang / de bzhin gshegs pa nyid yongs su’ sbyang™ ba dang /" sems 
can ’di dag ji ltar nyon mongs pa’i sbubs kyi shun’ pa thams cad las grol te° /"! ’jig 
rten du de bzhin gshegs pa dgra becom? pa yang dag’ par rdzogs pa’i sangs rgyas 

13 > 

shes’ bya ba’i’ grangs su’? ’gro bar® ’gyur snyam! nas" /'* sems can rams” la chos 

ston to // 

'BIQP.T: om. de bzhin du § Pio: ni for Ayi. original gsal) with its right 
(aberratio ocult). 7B: bkris; P,: bkri,, P2: dkas; half deleted. 

2 PS: om. de bzhin gshegs pa Q: dkrigs. 10 DAT: shin. 
yang (aberratio oculi). ® BQ: kyis. yr: //, 

7 LP3: gyi. ° PL: gsal; Q: b- of bsal 2 BDINOQP;: zhes. 

: Py3: om. /; S: //. smaller; lacuna of a half Bip). grangsu. 

5 BDINQTP 2: nyid /; letter between b- and -sal: b- 4 BDINQ: om. /; P;T: // [T at 
P;: nyid //, is the left half of a g- (of the end of the line]. 


1 Fis: HERR for Ze ttt (contrary to four-syllable rhythm: #428 tH). 


PDF Version: BPPB VI (2002) 


rigs kyi bu de ltar ’dir de bzhin gshegs mig gis sems thams cad nyon mongs pa’i 
mdzod kyis phub mas dkris pa de bzhin gshegs pas : ’gyur ba : sangs rgyasu’ ’gyur 
ba : rang ’byung ba don dang bcas te : gnas pa : mthong ngo : rigs kyi bu de bzhin 
gshegs pas kyang nyon mongs pa thams cad kyi phub ma bsal ba’i phyir de bzhin 


gshegs pa nyid yongsu sbyangs pa dang : ci nas’ sems can "di dag nyon mongs pa 

tharns cad kyi mdzod kyi phub ma las grol te : ’jig rten du de bzhin gshegs pa dgra 

bcom pa yang dag par rdzogs pa’i sangs rgyas su ’gro ba’i phyir : °°" serns can 

mams la chos ’chad do : 


tek, BAT, WOR. Reg Tee ILRI AUR 



P, Mini, BK. ARES. RA MADRE, ¢ 


1 Jin: @l, for 4. 

? FAB [=Ji.]}C- ERS (Ji shows the lectio difficilior; the reading BItHNE is to be expected. Though 
the particle 4 is also used in Ch, to mark the object of an action, it never appears throughout the text 
in case of the exposition of the Dharma (e.g. ... 2¢7/& ). On the other hand, the combination Bf*> 
appears several times in order to mark the subjects benefiting from the Buddha’s activity: (1B) 203% 
remarkable that Ch, adds # (which is without any counterpart in the Tib. versions) to %& rather than 

to employ #/* to characterize 3% as the object. (Cf. 5B quoted above; also: (7B) Fish MRw A). 
The emended text should accordingly be understood as “... teaches the Dharma for [them] (f3#4).”) 


PDF Version: BPPB VI (2002) 


de nas de’i tshe bcom Idan ’das kyis* tshigs su’ bead pa "di dag bka’ stsal to // 

[3.1] ji ltar® ’bru ’am sa! 1u’i” bras kyang rung‘ //?* 
ci tse* ’am ni ’on te nas kyang rung //° 
ji srid bar du de dag phub* beas® pa /* 
de srid' bar! du bya ba mi byed de* // 

[3.2] de dag gis ni brdungs™ nas phub bsal" na® /” 
bya ba ram? pa" mang* po dag kyang byed // 
snying po phub ma Idan pa de dag ni //" 

sems can rnams la bya ba mi byed do” //” 

' BDJNQST: sd for sa. 3D: ef. OK, fn. 4. 
? Pio: om. 7. “BQ: rise. 


MPa Lez 

[3.1] BD KOK 
BABES ais ey 
RAO Rta 

"2 [=KuMiSo YuJsNaPuQiQsZi}<h. 


5B: /; Py2: om. pada 3.1b. 

PDF Version: BPPB VI (2002) 

de nas bcom Idan ’das kyis de’i tshe de’i dus na : tshigsu bead pa ’di dag bshad do // 
[3.1] ci ltar "bru dang : 
ci Ita nas dang : 
ci tse rams phub ma beas gyur pa : 
don du bya ba 
[3.2] de dag phrad te bsal nas su // 
tmarn pa mang! po’i don kyang byed : 
de dag snying po phub mar Idan : 

bsal nas serns can don kyang byed : 



[3.1] SRO A RERHS 


1 Fe[=Ga,]<fK (possible is also an emendation to 28[=Ji,] “Chinese millet”). 

? $3 [=Ji,|<H% (see Tib, Bth (3.2c) and Ch, (3.2b): AEB). 
> $41 =Ji.]<—13 (I understand 7.48 as the “undamaged [kernel]”). 


PDF Version: BPPB VI (2002) 

10 [3.3] de bzhin sems can kun gyi® sangs rgyas* sa’ //” 
nyon mongs” mams kyis™ khebs”® par® ngas mthong nas //“@ 
ngas” ni de® dag mam" par sbyang®® ba" dang? //'° 
sangs rgyas myur thob bya phyir chos ston to //" 
[3.4] sems! can kun la!’ nga ’dra’i’? chos nyid ni /* 
15 nyon mongs brgya’’ yis" dkris'* nas™™ gang yod pa //™ 
de ni!> mam°° sbyangs’° thams?? cad ji Ita% bur //* 

rgyal bar!’ myur ’gyur® bya phyir chos ston to //* 

° B: kyi; LP): gyis. vertically arranged dots after nga ‘dra’i. 
7 Pi: so for sa. dang; QST: /. 3 BQ: rgya (contrary to Ch: 
® P,: de with small letter " L: between /a and nga: Ba). 
beneath the line; S: om. de. deleted letter (not clear); it if BQ: dDkris. 
* P.: yang for dang; P3: bya could be ’a from originally 1S p13: na for ni; P2: nas for ni. 
for dang. la’ang (see the following 16 DINP;: sbyang. 
'©L: om. // (at the end of the fa.). 17,03: ba for bar. 
line): the copyist marked the 2 p.): yang ‘dra ba’i for nga 
end of the the line with four ‘dra'i; P;: ‘ang ‘dra ba’i for 

[3.3] RARER Me ea 

2 JsMi: $&% for 2. 


PDF Version: BPPB VI (2002) 

10 [3.3] de Itar bdagis serns can tharns cad mthong : 
nyon mongs pas ni sangs rgyas sa yang khebs // 
de dag bsal ba’i phyir chos kyang bshad :! 
myur tu tharns cad sangs rgyas pa yang ’gyur : 
[3.4] semns can tharns cad chos nyid nga ’dra ste : 
15 nyon mongs pa ni brgya’i bskor nas gnas : 
de dag marh par dag phyir chos kyang smra J//) 
ci nas tharns cad ymyur du, rgyal ’gyur ba // 



[3.4] Sea Als HEA ATI RCEE 
Fe DRE PaaS 

4 th[=Ji2]<— (cf. sangs rgyas sa in Tib and Bth). 


PDF Version: BPPB VI (2002) 

rigs kyi bu dag gzhan* yang "di Ita ste! dper® na / rul pa” dang’ / nyal nyil® gyi* gnas 

bshang> gci° dri mi zhim pas® kun tu' gang ba zhig® tu" /’ mi gzhan gseb' lam nas 

song ba zhig gi® gser gyi gar bu zlum po zhig“ Ihung’ bar gyur! la /™ bshang”® gei!!" 

dri nga bas kun tu° gang ba /” rul? pa!? dang /'* nyal nyil gyi? gnas' der!> mi gtsang* 

ba gzhan dang’® gzhan dag’’ gis'* mnan’® pas” /”! bitar mi snang bar' gyur cing” der 

de” lo beu ’am /™ nyi shu ’am 7° sum **cu ’am / bzhi beu ’am /*” Inga beu ’am /” 

lo brgya ’am / lo stong du mi gtsang”* bas”” chud* mi za ba’i’ chos can de sems can 

gang la yang” phan pa’? mi byed do™ // 

d Pi2! ste /. 
? DNQS: ba; J: pa or ba (?). 
3 P1253: ’am for dang. 

‘ BP;: kyi; P;: gi; P2: om. gyi. 

> J: gshad(?); L: bshad; 
NQP)3: gshang; P2: om. 
_ Dshang. 
§ INQP;: ci for gci. 

7 BDINQP 1: om. /. 

® BOLSTP3: gis. 

P)2: /tung; N: not clear. 

Wy: gshad; NP12: gshang; T: 

Hi JNQP 123: ci for gel, 

2 BDINQ: om. /. 

3 Q: ba; S: pa or ba (2). 

4 BOP,: om. /; T: spungs 

shad (at the end of the line). 

I P23: der /. 


16 BJQ: om. gzhan dang 
(aberratio oculi);, T: dang / 

 BDINQPj2;: om. dag. 

i BDINQP 23: gyis for gis; S: 

1° BDINLP,3: brnan for mnan 
(see Bth: brnan): contrary to 
Chy: BRAT 
P,: gnan for mnan; Q: pa 
rnan for mnan; S: lacuna of 
about one third of a letter 
between m- and -nan of 
mnan (caused through 
deletion of tsheg?); first -n- 
of mnan very tall (caused 
through deletion of 
superscibed r- from originally 

BuRSerSe, iz, TH, WE 

' MiYuJsNaPuQiQs (Fs not readable): §. 


20 DN: bas; J: pas or bas (?); 
P;: ba. 

21 BDINQ: om. /; S: /. 

2 DINP23: cing /. 

3 DJ: des for de; N: des or de 
la (?) for de. 

4 NP): om. / [N: gap 
between ’am and nyi of the 
same size as the gap 
between the following ’am / 
and the numbers; T: //. 

3 Di: om. /. 

fa JP,3: bcu for cu; Q: lacuna 
of one letter between sum 
and cu filled with tshegs. 

27 BDINQ: / lo Inga. 

8B: om. gtsang; Py3: btsang. 

si Py2: bas /; Q: ba’i for bas. 

3° BDINQP;: par for pa. 

aR, RET 

PDF Version: BPPB VI (2002) 

Bth — 

rigs kyi bu gzhan Yang! "di Ita ste : dper na : rul ba’i lud kyi khung slad stng : gcin 
dang : dri mi zhim pas : yongsu dag pa’i lam de’i dbusu skyes bu gzhan zhig ’gro ba 
na : gser kyi ’chin pa zhig stong bar gyur nas : der slad sa dang : gein! gein dri mi 
zhim par yongsu gang : ril ba’i lud phung mi gtsang ba gzhan maths kyis brnan pas : 
lam kyi snang bar ’gyur te : de na de lo beu ’am : nyi shu ’am : surn cu ’am : bzhi 
beu ’am : drug beu ’am : brgyar brtson par! byed pa de dag gi don kyang ma rung bar 
mi ’gyur ba "dug ste : des de na serms can gang la yang phan par mi byed de: 




PDF Version: BPPB VI (2002) 


rigs kyi bu dag de nas lhas lha’i mig gis gser* gyi gar bu’ zlum® po! de la? bitas¢ 
nas /*° mi zhig' la kye® mi khyod song" la‘ ’di na’ /* rin po che’i mchog' gser rul pa’ 
dang /* nyal nyil gyi? mam pas non pa de’ byi™ dor gyis la®’° gser gyis'! gser gyi® 
bya ba gyis? shig!? ces bsgo”? na" rigs kyi bu dag’> rul pa‘ dang /'° nyal nyil gyi 
mam pa zhes bya ba de ni’ nyon mongs pa mam pa‘ sna tshogs kyi tshig’ bla” dags 

so!’ //” gser” gyi gar bu zhes bya ba de ni chud* mi’ za ba’i chos can gyi tshig” bla 

18 1/8 /9 

dags so lhas Iha’i mig” ces bya ba de ni /’ de bzhin gshegs” pa dgra becom pa 
yang dag par rdzogs pa’i sangs rgyas kyi’” tshig bla dags so” //“ rigs” kyi bu dag” 
de” Itar de bzhin gshegs® pa dgra bcom pa yang dag par rdzogs pa’i"" sangs rgyas 
kyang /”? sems can thams cad" la! de bzhin gshegs pa’i chos nyid““ chud mi za ba 
yod pa’i" nyon mongs pa” rul pa™™” dang / ’dam™ rdzab®™ Ita bu rams’? bsal% ba’i 

phyir /*> sems can™ mams* la chos ston to // 

' BQ: pos for po. ° P,: kyi for gyi; BQ: gyis for dags,o [dagso emended to 
2 BJQ: om. la. gyi. dags ,o through subscription 
3 BDINQ: om. /; L: spungs  BDINQ: la /. of a second -s-]. 

shad (after first syllable). 1! BINQLP3: om. gser gyis '8 TNP>3: dagso [N: dagso // 
4 Ps: lal. (contrary to same at the end of the line]. 
5 BDINQLP),: ni for na (see construction in 2A.4f). '? BDINQ: om. /. 

the parallel verse RGV'1.109 ? p,: om. shig; P2: zhig. 20 IN: kyis. 

cited in the note to my ® BQ: bgo; Pi: bsgu. IT ps: dagso. 

translation: ... asmin ...). 4 Pig: na /. 2 p.,: da for de. 
§ BDINQP)23: om. /. 'S BINQP:: dag /. 3 BDINQP ys: om. /. 
7 Q: ba; DINS: pa or ba (7). '6 BOP,: om. /. 4 D3: pa/. 
8 DJ: om. /. 1 LP: dagso [L: dagso // 5 BDQPi23: om. /. 

before string hole]; T: 



| Dh, FsZi: 3% for 38. 


PDF Version: ie VI (2002) 


de nas rigs kyi bu gser kyi ’chin pa de lhas ’phrul gyis mig gi mthong na : skyes bu 
gzhan! zhig ’di smras so : kye skyes bu song la ’jigs na : rin po che mchog gi gsel rus 
pa ’i lud kyi non cing dug pa de la yongsu sbyongs dang : gser kyi gser kyi don byed 
do : rig, gyi bu rul pa’i ud khang zhes bya ba! ’di na’ nyon mongs pa sna tshogs : 
mam pa du ma’i tshigs bla dag so // gser gyis ’chin pa zhes bya ba ni chos nyid ma 
rung bar mi ’gyur ba’i tshigs bla dag so : lha’i ’phrul gyis mig ces bya ba ni : de 
bzhin! gshegs pa dgra bcorn pa yang dag par rdzogs pa’i sangs rgyas kyi tshigs bla dag 
so // rigs gyis bu ’di ni de bzhin gshegs pa dgra bcom pa yang dag par rdzogs pa’i 
sangs rgyas kyis ma rung bar mi ’gyur ba’i 7°”! chos nyid yod pa la nyon mongs 
pa’i rul ba’i ’dam du gyur pa med par bya ba’i don kyi phyir serns can rams la chos 
*chad do // 

' Not clear: na or ni? 


, Maw, Bowe, Hs, mes. 

! Ga,T> omit #E (but contrary to four-syllable rhythm: TBE TE; cf. Bth: chos nyid ma rung bar mi 
‘gyur ba for *avinasadharmata). 



PDF Version: BPPB VI (2002) 


de nas de’i tshe bcom Idan ’das kyis* tshigs su’ bead pa ’di dag? bka’ stsal to //* 

[4.1] ji ltar mi yi? gser gyi gar bu‘ ni //** 

nyal nyil ram pa’i‘ nang du hung gyur la’ //* 

der de lo ni® mi nyung ba” zhig' tu // 

de* Itar gnas kyang mi ’jig® chos can no! //™ 

[4.2] tha yis® lha yi’ mig gis® de mthong nas /” 

ram par sbyang phyir gzhan® la smras” pa’? ni // 

di na!! rin chen? mchog gi‘ gser’ yod pa‘ //"* 

mam par sbyongs' la des! ni” bya ba gyis’ //” 

[4.3] de bzhin ngas* ni sems can thams cad kyang //” 

nyon mongs mams kyis'* yun ring rtag non” mthong // 

de dag gi!’ ni blo’® bur™ nyon mongs shes // 

rang bzhin sbyang”™ phyir thabs kyis” chos ston to®? // 

LT: tshigsu. 

2 ST: yis for yi. 

> Po: of. OK, fn. 4. 

‘ST: /[S at the end of the 

> BQ: ‘jigs for ‘jig. 

° Pio3: thas ni for tha yis. 

” ios: dha’'i for tha yi [P2: WY: 

-s crossed out (?); gi gu 

above and ‘a chung beneath 

-s not clear]. 

: Py3: gi. 
P42: smra. 

19 Pip: ba. 

" BDI: ni for na; P2: na with 
a deleted gi gu above. 

2 po: insert rkyang shad 
beneath the line for nyis 
shad; end of pada had 
probably been overlooked 
[P2: rkyang shad four times 
as long as usual (in bottom 

[4.2] 72S YR BS 
[4.3] SRI BRA 
[4.4] SRE TBA SH} OS 

| Jip: 50 for 5 (but cf. Tib 4.2a: mthong). 

BN: deg; P3: de for des. 

4 BOL: Ayi [Q at the end of 
the line]. 

5 BDINQ: gis (contrary to 
BDJNQ, which marks 
sentient beings as the agent 
of knowing about the nature 
of defilements; cf. RGV 

6 BIQP LT: glo. 

2 Dh, FsKuMiSo YuJsNaPuQiQsZi: §R for we. 

3 DhoFsZi: & for fe (cf. BARE for *anukila which fits into the context of teaching; the variant EX, 
*yathakama, might be evoked by 4.2c: f&#ii...). 


PDF Version: BPPB VI (2002) 

de nas bcorn Idan ’das kyi de’i tshe de’i dus na : tshigsu bead pa ’di bshado //! 

[4.1] ci ltar skyes bu’i gser min 
lud kyi khong du stor ygyur, pa : 
de’i de la lo mi nyung 
chos nyid ma rung mi ’gyur gnas // 
[4.2] lha’i ’phrul mig de mthong nas : 
de sbyangs phyir ni gzhan la smras : 
gser ’di ni rin chen mchog! 
ram par de sbyangs don kyang byed // 
[4.3] ngas ni de ltar sems can thams cad mthong // 
yun ring du nyon mongs pa thams cad non // 
de dag nyon mongs : gro’ur shes pa na / 

rang bzhin gyis ram par thabs kyang smras //! 





[4.3] (IRR ARRIBA 



PDF Version: BPPB VI (2002) 


rigs kyi bu dag gzhan yang ’di Ita ste’ dper na /” dbul’ po zhig gi? khyim gyi nang® 
gi? mdzod kyi ’og gi’ sa‘ la gter chen po dbyig® dang’ gser gyis rab tu gang ba® /” 
mdzod kyi tshad tsam zhig mi bdun srid kyi® sas’ yog™® pa’i ’og na!! yod la‘ //’? gter® 

chen po" de mi dbul po de la ’di skad du! /"? kye mi nga” ni gter chen‘ po ste /’° sas! 


yog’® cing ’dug go™!” zhes ni'® mi smra” ste /"”  ’di Ita ste /*° gter chen po ni sems 

kyi ngo bo nyid kyis° sems can ma? yin pa’o‘ // mi dbul po khyim gyi bdag po’ de ni 

dbul ba’i2’ sems kyis® rjes su’ sems shing" de nyid kyi steng’ na” mam par rgyu 

yang” /” sa’i ’og na gter chen po yod” pa de ma thos mi shes ma mthong ngo’ // rigs 
kyi bu dag” de bzhin du sems can thams cad” kyi® mngon par“! zhen® pa’i yid la 
byed® pa®® khyim Ita" bur gyur" pa’i ’og na /*“ de bzhin gshegs pa’i snying po’i”® 
stobs dang /! mi ’jigs pa dang “ ma ’dres pa dang /" sangs rgyas kyi chos thams cad 
kyi mdzod kyi™ gter chen po yod kyang /*° sems™ can de dag’’ gzugs dang / sgra 
dang /? dri dang /“4 ro dang / reg pa la” chags pas™ sdug bsngal bas’” ’khor ba 

na” ’khor® te”? /* chos kyi gter” chen po” de™* ma thos pas thob par ma gyur?”? 

cing’ yongs su>! sbyang™ ba’i*? phyir brtson par” yang™ mi byed do® //™ 
'DINP?: ste /. 4 po: nga’i ni. 4 AD: om. /. 
2B: //; DIN: om. /. 5 BQ: om. /. 5 ABDJNQP};: om. ‘i (see 
3p: gyi; P3: om. gi. 1S BDINQ: g.yogs. note in my translation). 
“ BP»: gis. " BIQ: go //. 6 ADP;: om. /. 
7 BDINQ: dang /. vs BDINQP)23: om. ni. at Pjy3: om. de dag. 
§ BQP): bar. 19 DIN: om. /. 28 IN: ’khod. 
T BDINQPy23: om. /; L: //. 20 ABQ: om. /; T: om. ‘di lta 2° BINQ: de for te. 
. BQ: kyis; Py2: om. yi. ste /. 30 APio3: cing /. 
° p,: pas for sas; BQ: sa for "11S: po for ba'i; T: pa for 31 UT: yongsu. 
sas. ba’i; P3: pot for bai. 32 A: bar; LST: pa. 
10 BDINQ: g.yogs for yog. 22 DT: om. na. 3 Di: om. yang. 
Npinal. 3 ABDINQ: om. /; L: 
BR BDINQST: /; P13: om. //. vertically arranged dots for 
3 BDINQ: om. /. shad (at the end of the line). 



PDF Version: BPPB VI (2002) 

rigs kyi bu gzhan yang ’di Ita ste : dper na : dbus pa’i khyim kyis nang re lto’i ’og gis 
sa gser chen po ’khor rgyang grags kyi tshad nor dang : gser gyis yongsu gang ba 

_ skye ba bdun srid kyi ’og na : sa g.yo cing! ’dug ces smra’o : kyang "di Ita ste ; dper 

gter chen po de la yid med de : ngobo nyid kyis sems can ma yin pa na : gter chen po 
de skyes bu dbus po de la : kye skyes bu nga gter chen po ’og na : sas yog cing! ’dug 
kyang ces smra’o : skyes bu dbus po khyim kyi bdag po de sems can dbus po rjesu 
sems shing : nga’i stang na ’gro yang gter chen po sa la: ’dug pa des : ma thos mi 
shes mi gtong ngo : rigs! kyi bu de Itar ’di bzhin serns can tharns cad de bzhin gshegs 
pa’i yeshes dang : stobs dang : mi ‘jig pa dang : sangs rgyas kyi chos ma ’dres pa 
mdzod // gter chen po bzhin gyur te : gnas so : setns! de dag gzugs dang : sgra dang : 
dri dang : ro dang : reg pa la sa chags pa’i phyir : sdug bsngal bas ’khor ba ’khor 
yang ma thos pa’i phyir : de dag gi chos kyi gter chen po myed par mi ’gyuro : 
yongsu sbyangs pa’ dang :! nan tan du byed par mi °gyuro : 

' Not clear: pa or ba? 




aaa, Pt. B. A. OR. ee, SE. AULA AER, Oh 

"SCR Ji] KK (for purusa). 


PDF Version: BPPB VI (2002) 


rigs kyi bu dag de nas* de bzhin gshegs” pa ’jig rten’ du byung ste / byang chub sems 
dpa’i nang du ’di Ita bu’i’ chos kyi gter chen po yang dag par rab® tu ston’ to // de dag 
kyang chos kyi gter chen po® de la mos nas" rko! ste / de’i phyir ’jig“ rten na! de 
bzhin gshegs pa dgra bcom pa™ yang dag par rdzogs pa’i sangs" rgyas mams° shes’ 
bya ste /? chos kyi? gter chen po Ita bur gyur? nas /" sems can mams' la sngon' ma 
byung? ba’i gtan tshigs kyi" mam pa dang / dpe‘ dang /° byed pa’i” gtan tshigs® dang /” 
bya ba mams’ ston™ pa’ gter chen po’i mdzod kyi sbyin bdag® chags pa med pa’i 
spobs” pa dang” Idan zhing /’° stobs dang /'! mi ’jigs pa dang / sangs rgyas kyi°® 
chos mang po’i® mdzod du gyur™ pa yin no® //* rigs kyi bu dag de Itar de bzhin 
gshegs®® pa dgra bcom pa" yang dag par rdzogs pa’i sangs rgyas kyang’* de bzhin 
gshegs" pa’i mig shin! tu“ yongs su dag" pas™™" /’? sems can thams cad™ de Ita bur 
mthong nas /"* de bzhin gshegs’? pa’i*’ ye shes™ dang /** stobs dang /'* mi ’jigs pa 
dang / sangs rgyas kyi" chos ma ’dres pa’i mdzod yongs su'® sbyang ba’i phyir /”” 

byang chub sems dpa” mams la chos ston” to // 

' DN: zhes. > ABDJNQP,,: om. /; L: ™ BQ: om. /. 

2DT: //. spungs shad (at the end of 2 Py23: Ayang /. 

3 AS la sngon ma byung ba the line). a BDINQP}23: om. /. 
sngon ma byung ba’i 6 Pi4: tshig. 4 BDINQ: om. /; T: /. 
(dittography); LST: la sngon 7 Pas: rnam. 5 BOP): om. /. 
ma byung ba ’am |/* sngon 8 Pio3: bdag / [P13 at the end of 1ST: yongsu. 
ma byung ba’i[*L: /]. the line]. 7 ABDINQ: om. /; T: //. 

4 A: dpe’; P12: dbye ba for ° BP): dang / [B: spungs shad 
dpe. (after second syllable)]. 

0 BDINQP}2: om. /. 



‘BUS S1E=Dh Fs] BBB (four-syllable rhythm). 


PDF ves VI (2002) 


de nas rigs kyi bu de bzhin gshegs pa’i ’jig rten du byung nas / byang chub sems dpa’ 
dbus na bzhugs te : ’di Ita bu’i ches kyi gter chen po 7%""! kun ston to : de dag gi de 
las dad nas chos chen po’i gter te bton pas : gnod miza bar de bzhin gshegs pa dgra 
bcorm pa yang dag par rdzogs pa’i sangs rgyas shes bya ste : jig rten du chos kyi gter 
du gyur nas /! sems can maths Ja rim kyi thog mar rgyu dang : mam pa dang : dpe 
dang : dgongs pa dang : bya ba smra zhing chen : chen po’i mdzod kyi bdag go // 
spobs pa ma thogs pa ’o : shes rab mang po’i mdzod du gyur pa’o : rigs! kyi bu di 
Itar de bzhin gshegs pa dgra beorn pa yang dag par rdzogs pa’i sangs rgyas shin tu 
yongsu dag pa’i de bzhin gshegs pa’i mig gi sems can thams cad mthong nas : de 
bzhin gshegs pa’i yeshes! dang : stobs dang : mi ’jigs pa dang : sangs rgyas kyi chos 
ma ’dres pa’i mdzod yongsu sbyangs pa’i byangchub sems,dpa’ mams chos ’chad 
do // 




PDF Version: BPPB VI (2002) 


de nas de’i tshe beom Idan das kyis’ tshigs su” bead’ pa ’di dag bka’ stsal to* // 
[5.1] ji ltar dbul po’i® khyim gyi® ’og logs na //*4 
dbyig dang gser gyis gang ba’i gter yod pa //* 
de la g.yo' ba’am’ rlom sems yod min te // 
nga ni khyod kyi® yin zhes de mi smra //" 
[5.2] de yi® tshe na sems can khyim bdag de // 
dbul bar gyur’ la“ ram par mi shes shing //' 
sus kyang de la bsnyad’ pa™ med pas na" //° 

dbul po de ni® de yi’ steng na” "dug //1 

: P3S: kyi. *P.: cf. OK, fn. 4. P,: bsnyam; P3: snyed. 
2NLT: tshigsu. > AB: ba ‘am. 8 po: na for ni. 
3 NP; gcad; P2: bca(s)a 6 AP 93: de’i for de yi; B: yin ° BP123S: de’i for de yi; Q: yi 
[-s- marked with dots for yi. compressed. 
above for deletion, -d TL: bsnyed; T: brnyed 
inserted beneath]. (contrary to Chy: 22); 

[5.1] BUBAR ANAL Bix 
[5.2] BFE RAB 

' KuMiSo YuJsNaPuQiQsZi (but not Fs as variant in Ji, erroneously indicates): 4% instead of 7X 
(contrary to parallel in 5A: ELA. B #0). 


PDF See VI (2002) 

de nas bcorn Idan ’das kyi de’i tshe de’i dus na : tshigsu bead pa ’di bshado :! 
[5.1] ci Itar dbus po’i khyim du gter 
nor kyi gang bar snod gyur pa : 
mi g.yo mi sems de mi smras : 
[5.2] sems can de ni khyim bdag ste : 
mi shes pas na dbus gyur pa : 
gzhan ni su yang de la mi bstente : 

de’i stang na dbus po! shin tu gnas / 


many Ea ITA 




PDF Version: BPPB VI (2002) 

10 [5.3] de ltar ngas ni sangs rgyas mig gis sul? // 
sems can’ de dag thams® cad dbul ’dra‘ la” // 
de dag” rams la gter chen yod pa dang // 
g.yo ba med cing bde gshegs lus su” mthong //* 
[5.4] ngas ni de mthong’ byang” chub sems dpa’ la //°° 
15 khyod kyis® nga yi!! ye shes mdzod zung la //*4 
dbul ba med cing ’jig rten mgon® gyur” dang® //** 
bla med chos gter ’gyur bar gyis shes" bstan! //** 
[5.5] gang dag ngas" bstan™” pa la’? mos gyur™ pa”? // 
sems” can de dag kun la gter yod do //”? 

20 gang dag mos nas bdag nyid rtsol byed pa% // 
de dag myur du byang chub mchog thob'*” ’gyur // 
LP,,: gisu [P; at the end of yi compressed. '4 LN: ‘thob; S: mthob. 
the line]. 2p, om. la; P3: om. pala, 
| BP. ,5: nga’i for nga yi; Q: 8 BDINQP a3: Ja for pa. 



[5.5] AeA —t0A ie 

2 Dh,FsKuMiSo YuJsNaPuQiQsZi: 5 for $A. 

3 Dh F'sKuMiSoYuJsNaPuQiQsZi: F for 7. 

* Dhg: hil for 49; FsKuMiSo YuJsNaPuQiQs: ¥\ for 3. 

5 Dh Fs: Fit€ Bh instead of 4)77(& (variant contrary to the prevailing rhythm of 2 and 3 syllables in 
the first and third padas of the verses of Ch,;: 77{€ must be a compound). 


PDF Version: BEPB VI (2002) 

[5.3] de ltar ngas ni sangs rgyas mig gis mthong : 
10 sems can ’di kun rab tu dbus gyur pa // 
de la gter chen po shin tu yod pa ni: 
bde bar gshegs pa kyis lus ni mi g.yo ba: 
[5.4] ngas mthong :’ byangchub serhs dpa’ la yang smras // 
yeshes kyi ni mdzod chen kun kyang thon // 
15 mi dbul ’jig rten kyis mgon por gyur // 
bla na med pa chos kyang nor yang sbyin! : 
[5.5] gang cig nga’i sa bshad pa de la mos // 
sems can la ni gter yang yod J 
gang zhig shin tu mos shing brtson byed pa // 
20 de ni byangchub mchog kyang myur rnyed do // 

'Not clear: sbyin or spyin? 


[5.3] ERLE — ARSE 
SMAAK fei hE 

[5.4] Rik See ESAS 
SHA (FS Bemba 



PDF Version: BPPB VI (2002) 


rigs kyi bu dag gzhan yang" ’di Ita ste /’ dper® na ’ shing® a mra’i ’bras* bu ’am® / 
’dzam bu’i' ’bras bu ’am® /* ta la’i> ’bras bu ’am" /' spa’i’ ’bras bu yang* rung ste / 
phyi! shun gyi™ sbubs" kyi° nang na° //’ myu gu’i? sa bon? chud mi za ba’i’ chos can 
yod pa gang gis sa la btab‘ na /* shing' gi rgyal po" chen por” ’gyur* ro’ // rigs kyi 
bu dag de bzhin du’ de bzhin gshegs pa yang” ’jig rten na’ gnas pa'! ’dod chags 
dang /** zhe®® sdang dang / gti mug dang / sred pa dang /** ma rig pa’i nyon mongs 
pa’i phyi!” shun gyi“ sbubs kyis’® kun tu® dkris’* par mthong ngo™ //* 

' ABQSTP,;: om. /. ° BQ: om. na.  ALSTPi23: om. na (see my 
2 B: doube spungs shad (at the T ABDINQP23: om. //; ST: /. translation). 

very end of the line); P3Q: //; § ABDJNQ: om. /; S: spungs "| BDINQP,;: pal. 

DJN: om. /. shad (after first syllable in 2 Dp,T: Phyir, contrary to: 
> BQ: //. the top line of the folio). phyi shun in 6A.3. 
“N: shad: only upper half: * Pins: om. de bzhin du;T:om. — BQP»3: Kyi. 

PyT: //; Po: om. /. du. 4D): bkris. 

> Pi: ma’i for la’i. 


1 Ji, JsNaPuQiQsZiKuMiSoYu: %& for & (but see 6.1b: ABA SEB). 
2 DhaglsQiKuMiSoYu: #8; NaQsZi: 8; Fs: unclear. 


PDF Version: EREB VI (2002) 

rigs kyi bu gzhan yang "di Ita ste : dper na : am bra’i "bras bu ’am : ba sa na’i bras 

bu ’am : ’jam bu’i ’bras bu ’am #! tala’i "bras bu ’am : ’deb ’dre’i bras bu ’am : shun 

phrag kyis mdzod kyis dbus nas : sa bon dang : myu gu ’byung bar ’gyur ba yang mi 
jig pa’i chos nyid de : sa la btab dang : shing gi rgyal po chen por ?°"*" byed par 
gyuro : rigs kyi bu de ltar de bzhin gshegs pas : de bzhin gshegs pa’i mig gis ’dod 
chags dang : zhesdang dang : gtimug dang : sred pa dang : marig pa’i mdzod du ’jig 
rten zhugs shing gnas pa! 




PDF Version: BPPB VI (2002) 


de la ’dod chags* dang /” zhe sdang dang /* gti mug dang / sred’ pa dang /” ma rig 
pa’i‘ nyon mongs pa’i® sbubs‘ kyi nang na snying por gyur® pa’ de bzhin" gshegs pa’i 
chos nyid de ni* sems can zhes bya ba’i ming du! chags so° // de la gang bsil® bar 
gyur’ pa“ de ni mya ngan las das pa ste /' ma rig™ pa’i nyon mongs pa’i sbubs" yongs 

/© sems can gyi® khams? kyi’ ye shes chen po’i tshogs su!! 

su’ sbyangs® pa’i? phyir 
gyur’ pa® gang yin pa‘ de ni myed pa’o // sems can gyi khams kyi” ye shes chen 
po’i” tshogs dam* pa’ de ni /'? de bzhin gshegs pa’? ji Ita” ba!* de bzhin du smra™ bar 
lha dang beas pa’i ’jig rten gyis'”>® mthong nas /““ de® bzhin gshegs pa!® zhes'” bya 
ba’i® *du shes su®® byed do™ //’® rigs kyi bu dag de" la! de bzhin gshegs pas de Itar 
mthong nas /' byang chub sems dpa’ sems“ dpa’ chen po mams! Ja /° de bzhin 
gshegs pa’i ye shes khong”’ du chud par bya ba’i phyir™ don de™ nye bar ston” 

to /PP 

' BT: srid. TLT: yongsu. 6 5: pa’i for pa. 

? AQ: om. /; L: spungs shad 8 BQ: sbyang. 17 D3: om. zhes. 
(after first syllable in the top BOQP;: ba’i; S: bai or pa’i 18 AP,: / [A at the end of the 
line of the folio). (?). line]. 

3 AP 123: pa /.  BDINQ: om. /. ? ABQ: om. /; T: //. 

§ APio3: ni /. 1 LT: tshogsu. 20 AD: om. /; T: //. 

5 LNPi23: chagso [P: chagso //  BDINOQP;: om. /. aL BQ: khongs. 
at the end of the line]. 3 BOP: pa’ ji. 

6 DLST: bsal (contrary to 14 D3: bu for ba. 

Chip: FRR; Ps: geil. 5 BLP3: gyi. 



' Zi: 3 for J (contrary to Chz: #F¥ and Tib: bsil bar gyur pa: “cool”). 
2 JsNaPuQiQsMiYu: REERBA for Ke. 


PDF Version: BPPB VI (2002) 


de na ’dod chags dang : zhesdang dang : gtimug dang : sred pa dang : ma rig pa dang : 
nyon mongs pa’i mdzod kyis dbul kyi snying por de bzhin gshegs pa’i chos nyid 
kyang gnaso // de 1a gang serns can du ’du shes pa zhi bar! gyur na : ma rig pa’i nyon 
mongs pa’i mdzod : yongsu sbyangs pa’i serhs can kyi kharhs de yeshes chen po’i 
phung po nyid myed pa’o // de serns can mchog ste : yeshes kyi phung po chen po’o // 
ci Itar khong du chud! pa de bzhin du smra ste : Iha dang : ’jig rten du beas pas 
mthong nas / de bzhin gshegs pa zhes ’du'shes byed do // rigs kyi bu de dag de bzhin 
gshegs pa de Itar ++ zhing : byang chub serns dpa’ sems dpa’! chen po maths de 

bzhin gshegs pa’i yeshes khong du chud par bya ba’i phyir : don de ston to : 


' See the note in my translation: j& for 38? 
? DAA [Jiz]<—D (see the note in my translation: #8 for 37). 
3 4 (Jiz]<—ME (see the note in my translation). 


PDF Version: BPPB VI (2002) 


de nas de’i tshe beom Idan ’das kyis tshigs su’ bead pa di dag bka”* stsal” to® // 

[6.1] ji ltar spa‘ yi? *bras bu thams cad ni //° 
nang na spa’ yi myu gu yod pa ste° //® 
ta la" dang® ni ’dzam! bu’ kun la’’ang® yod //* 
nang na yod pa’i bras bu bskyed' na™ skye" //° 

[6.2] de bzhin chos kyi dbang phyug "dren? pa yang //4 

sems can thams cad spa yi’ sa bon ’dra’° // 

de kun nang na bde gshegs lus yod par' //" 

zag med sangs rgyas mig gi’! dam pas mthong”? //” 

7 LT: ishigsu. ™N: jambu (PE); Py: jam 
2 . " i+ Pas 
AP)\3: spa’i for spa yi; Po: bu 

a ’i for spa yi. , , 
rh ios Be ha 8 AB: la ang for la’ang; Pj»: 
44% eeehe fate Sos yang for ‘ang; P;: om. ‘ang. 
APiy3: spat for spa yi. AP. spa’ f : 
sy 123: Spa'i for spa yi. 
P}2: de for ste. 10 AP. "drag? 
5 A: nang for dang. ai : 


FEL ez 0 

[6.1] BIS ARTS 


' 21 Dh, Ji;|<4E (all variants mentioned in Ji; read #£). 


H BDJNQ: ni for gi, contrary 
to Chy: SRG EER ER; 
P,: gis for gi; Pz: migis for 

mig gi. 
"DIN: thob for mthong. 

PDF Version: BPPB VI (2002) 

de nas bcom Idan ’das kyis tshigsu bead pa ’di gsungs so // 
[6.1] ci ltar ’beb’ tra’i bras bu myu gu ba // 
de dang ta! la’i phyirol shuni phrag yod : 
*dzam bu kun kyis dbus na yang ni yod : 
de btang bas ni don kyang ’byung bar ’gyur : 
[6.2] de Ita ’dren pa chos kyi dobos” kyang : 
+++’i sa bon de bzhin sems can kun :! 
de’i zag med mig : gis mthong ba ni : 
de’i dbus na : bder gshegs lus kyang yod // 

1 Not clear: ‘bab, bab, ’beb or beb? 
? Not clear: dobos or ngobos? 


Sa ith. 

[6.1] SRP cope ROMs 

[6.2] MeKneRR —OA RT 


PDF Version: BPPB VI (2002) 

10 [6.3] sbubs bshig'? ma gyur” de* ni'* sems can brjod // 
mi shes nang’ na gnas kyang rlom sems med //” 
ting ’dzin” thob ste gnas nas“ rab” zhi ste //° 
de la g.yo ba ci yang“? yod ma yin //* 

[6.4] dper na sdong” chen® sa™ bon’ las byung Itar /# 

15 sems can ’di dag ji Itar ’tshang rgya zhing // 
lha dang beas pa’i ’jig rten skyabs“ ’gyur zhes // 

yongs su!® sbyang”” ba’i!® don tu’? chos gtam smra //" 

BIN: bshigs for bshig; P12: BINQP 23: gang for nang, UV BDI: spang for sbyang; Q: 
shig for bshig; P3Q: pa shig contrary to Bth (dbus) and spang or sbang (?). 
for bshig. Chz (FS 4). 'S BP;: pai. 

4 BP.03: na for ni. 161 TP: yongsu. 'SBDINQTPi23: du. 


[6.3] RAB bi KORRES Al 
[6.4] ATA PACHA 


PDF Vee VI (2002) 

[6.3] serns cano | ’di dag thadad mdzod pa yin // 
mi shes pa’i dbusu gnas par mkhyen // 
ting nge ’dzin thob cing =! zhi la gnas pa ni 
*di la nam yang g.yo ba med par ’gyur // 
[6.4] de sbyangs pa’i phyir ni chos kyang smras : 
ci nas ’di kun sangs rgyas par gyur to // 
ci ltar sa bon shin chen ’byung ba bzhin // 
lha dang beas pa’i ’jig rten! skyab ’gyur ba: 
' Small circle in the bottom half of the line; could be a symbol for Skt. iti. 

[6.4] Bia TARE 



PDF Version: BPPB VI (2002) 


rigs kyi bu dag gzhan yang® ’di Ita ste’ dper na /’ skyes® bu dbul po zhig la /* de bzhin 
gshegs‘ pa’i> gzugs rin po che sna bdun las byas pa’ //* lag* mthil’ tsam zhig yod la /° 
de® nas skyes" bu dbul po des’ /* de bzhin gshegs pa’i? gzugs de khyer te /'° ’brog 
dgon' pa las! shin“ tu’ ’da’ bar’! ’dod par™ gyur” nas //"? ci’? nas kyang de gzhan™ 


gyis’ mi° tshor zhing /'® rkun pos mi khyer bar bya ba’i phyir /'” des? de® ras rul pa’ 

dri mi zhim pa’i’ dum bu du mas‘ dkris" te’ /'? de nas skyes bu de ’brog dgon pa de 

nyid du” nyes’? pa gang* gis’ kyang ’chi ba’i dus byas par” ma gyur™ la /*° de’i?° de 

> ji42 1 hh p ’ 723 

bzhin gshegs™ pa gzugs” rin po che" las® byas pa’ ras rul pa’i dum bus 

dkris™ pa yang! rdog lam™ de na ’phyar! cing”® *dug pa dang  ’dron”’ po mams 
kyis!' ma™™ shes nas / ’goms”® shing ’goms™ shing”’ dong” 1a //*° ras rul pa®! dri mi 
zhim pa’i?? dum bu’i thum bu” kun tu®® dril ba** ’di*? gang nas rlung gis™ bdas® 
shes®> smad pa’i dngos“! por" yang ston pa dang /" ’brog na gnas pa’i°® thas” lha’i 
mig gis mam par bltas*” nas” /°° mi gzhan dag cig” la bstan™ te ”Y kye” skyes bu 

i” thum"’ bu ’di’i nang na /? de bzhin gshegs pa’i*? gzugs™ rin 

/% >ii9 rten thams cad° kyis“* phyag bya bar™ ’os pa yod 

dag ras kyi dum bu 
po che las byas pa 
kyis®®? phye™ shigt ces! bsgo’o // 

' ADIN: om. /; BT: //.  ABDJNQ: om. /; L: double 31D: pa or ba (2); Q: ba; Po: 

? ABDJNQ: om. /. spungs shad (at the end of pa with small letters beneath 
3 DLST: pa’i sku gzugs (see