An Annotated Translation of the Tattvasawgraha

(Part I) with an Explanation of the Role of

the Tattvasafgraha Lineage in the
Teachings of Kukai

Dale Allen Todaro

Submitted in partial fulfillment of the

requirements for the degree of
Doctor of Philosophy
in the Graduate School of Arts and Sciences

Columbia University
1985
 (£) 1985
Dale Allen Todaro

ALL RIGHTS RESERVED

 ABSTRACT

An Annotated Translation of the Tattvasawgraha

(Part 1) with an Explanation of the Role of
the Tattvasawgraha Lineage in the
Teachings of Kukai

Dale Allen Todaro

This is the first annotated, English translation of Part 1

of the Buddhist tantric text known in Sanskrit as the Sarva-tathagata-

tattva-sa~braha-nama-mahayana-sntra (The Mahayana sutra known as the

Compendium of T~~th of all the Tathagatas). The English translation

is based on the 1983 romanized edition of the complete Sanskrit text

in five parts by Kanjin Horiuchi. Horiuchi based his edition on the

two complete Sanskrit manuscripts of this sutra discovered in Nepal

in 1932 and 1956 by G. Tucci and David L. Snellgrove respectively.

This sutra is studied in the context of its role in the teachings

of Knkai (774-835 A.D.), founder of the Japanese Shingon sect of tan-

tric Buddhism. Kukai says that the esoteric Buddhist doctrines he

transmitted from China are contained in both the Tattvasawgraha-sutra

and the Mahavairocana-sutra. The Tattvasaigraha is also the major

Yoga tantra in Tibet and Tibetan and Indian exegeses are used.

Part 1 discusses the commentaries used, the structure. date,

place of origin of the text and gives a synopsis of the text. In

Japan the major commentaries are by Ennin and Donjaku. Part 2 ex-

plains the role of the sutra in Kukai's teachings. Two points are

made. First, it is well known that this sutra is the source for the

Vajradhatu map9ala first introduced from China to Japan by Kukai.

Second, the Tattvasawgraha provided the framework for the Vajradhatu

Recitation Manual introduced by Knkai. This was a ritual meditation

manual soon incorporated into the series of ~oga practices stand-

ardized by the Shingon school. This manual is the major example of

Tattvasawgraha lineage practices in the Shingon school. It explains

a praxis leading to enlightenment, the content of which is symbol-

ically depicted in the Vajradhatu ma~~ala. Part 3 contains the

annotated translation of Part 1 of the Tattvasawgraha.

 TABLE OF CONTENTS

INTRODUCTION . •. . •••. . ••••. •. •. 1

PART 1 THE TATTVAS.ArjtGRAHA

CHAPTER I The Text

Sanskrit, Chinese and Tibetan Editions,

Commentaries and Important Reference Works •••••••••• 8

Final Comments on the Translation •••••••••••••••••••. 18

The Title of the Tattvasawgraha •••••••••••••••••••• 19

The Structure of the Tattvasa~graha • • • • • • • • • • • • • • • • 23

The Date and .Place of Origin of the

TattvasaJ}lgraha • • • • • • • • • • • • • • • • • • • • • • • • • • 28

The Eighteen Assembly Vajra6ekhara .

• • • • • • • • • • • • • • • • • • 36

CHAPTER II Synopsis of Part 1 • • • • • • • • • • • • • • • • • • • 40

PART 2

1~E ROLE OF THE TATTVAS~GRAHA LINEAGE

IN THE TEACHINGS OF KUKAI

Chapter I The Role of the Tattvasa~graha Lineage in

the Practices transm:J.tted by Kiikai ••••••••• 60

A. The Vajradhatu Recitation Manual

The Relationship Between the Vajradhatu Ma~~ala,

the Vajradhatu Recitation Manual and the
Tattvasawgraha in China and Japan ••••••••••••• 62

The Vajradhatu Recitation Manual ••••••••••• '!' • 71

Key to Understanding the Relationship Between

the Vajradhatu Recitation Manuals and the
Vairadhatu ManrlalR: The Practice of th~
Vajradhatu Re~itation (~ r.tll$ ~ i~ ;~ JiiJ ) •••• 75
i
 B. Other Practices Transmitted by Knkai and
Activities of Kukai connected with the
Tattvasawgraha lineage ••.•••••••••• 100

CHAPTER II The Role of the Tattvasawgraha Lineage

in the Teachings transmitted by Kukai ••.•••• 113

PART 3
THE ANNOTATED, ENGLISH TRANSLATION OF PART 1
OF THE TATTVAS~GRAHA SUTRA

Table of Contents 128

Annotated, English translation 134

BIBLIOGRAPHY .. ........................• ............ . 369

~

APPENDIX I Tattvasa~graha Lineage Texts in Kukai's

Sangakuroku •••••.••• 387

APPENDIX II Kanjin Horiuchi's Romanized, Sanskrit

Edition of the Tattvasawgraha, Part 1 ••••.•• 392

APPENDIX III Amoghavajra's Translation of Chapter

1 of the TattvasaQ1graha .................... 470

APPENDIX IV Sego's Translation of Part 1 of the

Tattvasawgraha ••••••••••. 488

APPENDIX V Ennin's view of the Kong~chokyo and his

Synthesis of Tendai and Shingon: A
Summary ••.....•...... 518

ii
 DIAGRAMS

1 Line Drawing of Vajradhatu Mandala from Hasedera ••.•••. iv

••
2 The Nine Assembly (Genzu) Vajrad~atu Ma9<}ala •••••••••• 69

3 Lineage of Shingon Sect Patriarchs (abbreviated) ••••••• 80

iii
 ~·····"'·
DIAGRAM 1
ii. JJ. *
Line Drawing of Vajradhatu Map~ala from Hasedera 1

1
shinko Mochizuki, ed., Bukkyo Daijiten, Vol. 2 (Kyoto: Bukkyo
Daijiten Hakkosho, 1932), opposite p. 1316. (This reduction is approxi-
mately one quarter the size of the drawing in the text.)
iv
 1

INTRODUCTION

This thesis has two principle foci. First, the first annotated,

English translation of Part 1 of the s~ripture called in Sanskrit

Sarla-tathagata-tattva-sawsraha-nama-mahayana-sutra (hereafter abbre-

viated as Tattvasawgraha). In Japanese this text is usually referred

to in abbreviation as the Kcngochokyo. Japanese scholars also use the

Sanskrit title Vajra~ekhara-sutra for this work. This designation is

based on the title Kukai gives for this sutra written in the Siddham
1
script. Second, an explanation of the role of the Tattvasawgraha

lineage of texts in the teachings of Kijkai.

It has been written often that Kukai (774-835), founder of the

Japanese Shingon school of Tantric or esoteric Buddhism, relied prin-

cipally on two sutras in formulating the doctrines of the Shingon

school. 2 These were the Mahavairocana-abhisawbodhi-vikurvita-adhi~tha­

4
na-vaipulya-sutra3 and the Tattvasawgraha-sutra.

1see Shunkyo Katsumata, ed., Kobo Daishi Chosak~ Zenshu (here-

after abbreviated as KCZ), Vol. II (Toky~: Sankibo busshori~ 19q5),
p. 264. KO:kai bases his title in part on T. 18, No. 865 (~ ~y !~~) .
2Bokusho Kanayama, Shingon Mikkyo no Kyogaku (Ky~to: Rinsen sho-
ten, 1973), pp. 232-73; Yukei Matsunaga, Mikkyo no Rekishi (Kyoto: Heira-
kuji shoten, 1974), p. 189; Sir Charles Eliot, Japan~se Buddhism (Lon-
don: Edward Arnold & Co., 1935), p. 338; Daigan Matsunaga and Alicia
Matsunaga, Foundations of Japanese Buddhism, Vol. I, The Aristocratic
Age (Los Angeles and Tokyo: Buddhist :Books International, 1978), p. 181.
3Taisho Shinshu Daizokyo (hereafter abbreviated as T.), ed. and
comp., Junjiro Takakusu and Kaigyoku Watanabe, Vol. 18, No. 848 (Tokyo:
Taisho Issaikyo Kankokai, 192~1 ). This sutra was orally translated by
Subhakarasi~ha and recorded.in Chinese by I-hsing in 725 A.D. It
was brought to Japan during the Nara period (710-94) by the Japanese
Genbo <'( B$' ) who was in China from 716 to 735 A.D. See Mikkyo Jiten,
(hereafter abbreviated as MJ) cd. and,comp.:_Ryuken Sf!wa (Kyoto: Hoz~­
kan, 1975), pp. 473-74. <"i.%lJ!.iJ1317);1!:fV~t/Q~~)

4T. 18, No. 865. Translated by Arnoghavajra in 753 into Chinese.

It was first Kukai in 806 A.D. and then Ennin (794-884, Tendai sect),
 2

"Though Kiikai makes reference to a variety of Buddhist

literature representing a lvide range of schools of thought,
his ideas are fundamentally based upon two sutras and ~~stras.
These are the Mahavairocana and Tettvasawgraha siitras and the
Bodhicitta sastral and the Commentary on the Awakening of
Mah:Iy:Ina Faith2.n3 · - -- --

KOkai himself writes:

"The esoteric Shingon doctrine, the secret treasury given in

the two sutras, is unfolded by the Dharmak:Iya Mahavairocana
Buddha for the sake of his own enjoyment."4

Due to the importance of these two sutras it would seem natural

that there would exist translations, numerous commentaries and histo-

rical studies on both these scriptures. However, only the Mahavairo-

cana5 has been studied. Why has the Tattvasawgraha been neglected?

Several reasons can be given for this. First, Kukai obtained on-

ly Chapter 1 of the text and he, as founder of the Shingon school, left

no major commentary. Moreover, the only two major commentaries on the

text in Japan by Ennin (794-884) and Donjaku (1674-1742) often do not

interpret the text but instead give more information about it from

other texts. Ennin, a Tendai monk, wrote the first major commentary

(T. 61, No. 2223) but this only treats part of Chapter 1. Donjaku, a

Engyo (799-852, Shingon), Eun (798-869, Shingon) and Shiiei (809-884,

Shingon) who introduced Amoghavajra's three chuan translation of Part
1, chapter 1 of tbe Tattvasawgraha. MJ, p. 241.

1T. 32, No. 1665

2T. 32, No. 1666. Yoshito S. Hakeda, The Awakening of Faith attri-
buted to A~vaghosha (New York and London: Columbia University Press,
1967).

3:t.anoru Kiyota, Shingon Buddl~ism: Theory and _!'ractice (Los Angeles

and Tokyo: Buddhist Books International, 1978), p. 28.

4yoshito S. Hakeda, Kukai: Major Works Translated, with an Account

of his Life ~nd ~ Study of his Thought (New York and London: Columbia
University Press, 1972), p. 224.

sE.g., see the bibliography in Etude sur le Mah~vairocana-sutra

 3

Shingi-Shingon monk, analyses (T. 61, No. 2225) only Chapter 1. In_~on­

trast, ~ubhakarasi~ha's (637-735) 1 commentary on the entire Mahavairo-

~ significantly interprets this sutra which has greater litera~y and

intrinsic value and his interpretation is revered aud relied upon for

clarifying praxis and orthodox Shingon doctrine. In addition, while it

cannot be denied the Tattvas~wgraha played an important role in provid-

ing a source for the doctrines of the Shingon school, it must not be over-

looked that an entire lineage of Tattvasawgraha texts has exerted a sig-

nificant influence on the Shingon school. This is clearly evident in Ku-

kai's writings when he refers to or quotes the "Vajra~ekhara," this a

general term for any number of texts in the Tattvasawgraha lineage.2 In

other words, regardless of the specific Tattvasawgraha lineage text Kukai

quotes, for him Lltey all belong to one school of thought and practice.

Consequently, a study of the Tattvasa~graha can involve the arduous task

of a study of over sixty texts which has not yet been undertaken.

To be explicit Kukai gives in his Shingonshu shogaku kyoritsuron

mokuroku3 a list of the sutras, commentaries, etc. he required his disci-

ples to study. This list contains 424 fascicles of texts. Kukai lists

sixty-two as Tattvasawgraha lineage texts. Amoghavajra (705-774) trans-

lated all of these texts except ten. (See Appendix I for the

by Ryujun Tajima (Paris: Maisonneuve, 1936), pp. 142-46 and Yukei Matsu-
naga, ed., "Nihonjin no Mikkyo Kenkyu, Mikkyo Bunken Mokuroku," in Gendai
Mikkyo Koza, Vol. 8 (Tokyo: Daito shuppansha, 1976), pp. 270-76.
1T. 39, No. 1796. See Wilhelm Muller, Shingon Mysticism, ~ubha­
karasiwha and !~sing's Commentary to the Mahavairocana-sutra, Chapter
One, an annotated translation (Ann Arbor, Michigan: University Microfilms
International, 1976), order number BLK76-22210.

2Hakeda, Kukai: Major Works, p. 222 (T. 18, No. 865), p. 223 (T.
19, No. 957), p. 241 (T. 19, No. 1003), p. 242 (T. 18, No. 880) •
..,
JKCZ, Vol. II, pp. 168-82. Abbreviated Sangakuroku. Dated 823 A.D.
 4

list of the Tattvasawgraha lineage texts.)l For this reason Y.ukai is

said to have principally inherited the esoteric Buddhism transmitted by

Amoghavajra. 2 This list is clear evidence of the importance Kukai

attached to the Tattvasawgraha lineage as a whole in contrast to that

of the Mahavairocana lineage. Thus, at once, it should be recognized

that saying Kukai relied on the Tattvasawgraha-sutra in establishing

the doctri,es of the Shingon school is to be imprecise and invites

misunderstanding, It is the Tattvasawgraha lineage "f texts along

with the Mahavairocana-sutra and its commentary which Kukai relied

upon and transmitted in establishing the Shingon school. Moreover,

without a study of this lineage as a whole it can be difficult to

understand the Tattvasawgraha. This is evident in the fact that both

Ennin and Donjaku relied on T. 18, Nos. 870 and 871, both Tattvasawgra-

ha lineage texts, in "e~plaining" the Tattvasawgraha. 3

In addition to the above ~osition of the Tattvasa~Sf~~~ within

the Shingon school, this scripture belongs to a Buddhist tantric

tradition which, as it developed in India, Tibet, China and Japan,

placed emphasis on ri~ual meditation. In particular, Buddhist tantras

like ~he Tattvasawgr~ha articulate in some manner the practice of the

three mysteries of the body (mudras), speech (mantras) and mind (visual-

ization). 4 The Tattvasawgraha is not concerned primarily with explain-

1see T. 52, No. 2120, p. 839ff. for one list of Amoghavajra's

translation.:;.

lvukei Matsunaga, Mikkyo ~ Sojosha Sono ~ado to Shiso (Tokyo:

Hyoronsha, 1973), p. 187.
3 T. No. 870: &Jfi-~~fi!l m.J$.1hllt)')1Ur1':t'(f~;.tr'l; No. 871: i:~m~
ma-~a-t -t f'u·f.
4Alex Wayman, The Buddhist Tantras Light on Indo-Tibetan Eso-
tericism (New York: Samuel Weiser, 1973), pp. 36-42.
 5

ing Buddhist tantric doctrine but in describing in outline how to

draw various ma~~alas, initiation into these ma~2alas and the powers

resulting from these initiations. Such matters traditionally have

baen transmitted orally and only then.can be understood in detail.

The practice of the three mysteries, initiation into mandalas, etc. 1 ..

ideally have as their ultimate goal the enlightenment of the practi-

tioner who then in turn works out of compassion for the benefit of

mankind. Even though the Tattvasawgraha does not typically develop

Buddhist tantric doctrine,it does, however, give evidence of an assumed

philosophical position in its structure and content. This becomes

evident in the commen.~aries which make this philosophical position

clear.

Bearing in mind all the above factors this thesis has as its

first goal the first annotated translation of Part 1 of the Tattvasaw-

·,

graha for the following reasons: 1) The Tattvasawgraha is known to be

the basis for the Vajradhatu ma~9a1a which often is so prominently dis-

played in Shingon temples. In particular Chapters 1 through 5 of Part

1 and Chapters 6 and 7 of Part 2 are the basis for eight out of the

nine assemblies in the Vajradhatu mandala. .. For this thesis only Part

1 will be translated because this is a single unit in the scripture

corresponding to the Tathagata family. Moreover, because the ma~~alas

described in Chapters 6 and 7 have the same structure as the ma~~ala

described in Chapter 1 for the purposes of this thesis it is not neces-

sary to translate Chapters 6 and 7. If this were an art history thesis,

of course,translation of these chapters would be necessary to clarify

•.
the names of the deities, their places in the mandala, etc. For our pur-

1Ibid., "Preparation of Disciples; the Meaning of Initiation," pp.

54-70.
 6

poses it is enough to know that the first 7 Chapters are the major source

for the Vajradhatu ma~9a1a. The latter is an iconographic representa-

tion of the teachings and practices of the Tattvas~graha lineage trans-

mitted in ~he Shingon tradition and these will be made clear below. 2)

The Tattvasawgraha will be shown below to be the source for the frame-

work of the Vajradhatu Recitation Manuals transmitted in the Shingon

school. Although the Tattvasawgraha was not used specifically in the

meditation practices in the Shingon school it provided the framework

fur the Vajradhatu Recitation Manuals which are the major example of

Tattvasawgraha lineage practices transmitted by Kukai. (For 1) and 2)

see Part 2 below.) These Vajradhatu Recitation Manuals are based spe-

cifically on the Chin kang ting ching lien hua .£.!:! hsin nien sung .! kuei
(T. 18, No. 873) translated by Amoghavajra and which is listed by Kukai

in his Sangakuroku. Although this text has been thoroughly studied 1 it

will be necessary in the future to research in depth the role of all

Tattvasawgraha lineage texts Kukai lists in his Sangakuroku. Most of

these are ritual manuals which provided a basis for various rites in

the Shingon school. 2 3) The Tattvasawgraha is the first text Kukai

listsin his Sangakuroku, and regarding it as the basic text of the Tattva-

sawgraha lineage, Kukai takes the opportunity in his brief commentary

on the titl~ of the sutra to present a summary of the fundamental

teachings of the Tattvasawgraha lineage ·Which he was initiated into.

In the Tibetan tradition as well the Tattvasawgraha is considered

1E.g., Keiko Yoshida, Kontai Ryobu Shingon Geki (Kyoto: Heirakuji

shoten, 1978).

2ror some of these rites see Shojaku and Shoe, Himitsu Giki Zui-
monki, Vol. I of the Shingonshu Zensho(hereafter abbreviated SZ)(T~kyo:
Shingonsi1u Zensho Kankokai, 1933-37); Shinjo, Himitsu Giki Denju Ku-
ketsu, Vol. II of SZ.
 7

to be the fundamental Yoga tantra.1

The second goal of this thesis will be to explain the role of

the Tattvasaigraha lineage of texts in the teachings of Kukai. Because

not all texts of the Tattvasawgraha lineage have been examined for this

thesis the role of all of these in Kukai's teachings and practices can-

not bP. completely explained. This would require many years of research.

Rather the role of this lineage will be partially explained by an exami-

nation of Kukai's major writings. Attention will be focused on Kukai's

references to, quotations from and interpretations of known Tattvasamgra-

•-
ha lineage texts. Of course the major teachings of Kukai are well known.2

The purpose of the proposed researcl. is to clarify wh~ch practices and

teachings of Kukai can be traced to the Tattvasawgraha lineage on the

basis of textual evidence and how Kukai may or may not have changed spe-

cific teachings of this lineage.

Part 1 below discusses the editions of the Tattvasa~graha and the

commentaries used in this thesis, the date of the sutra, etc. Part 2

discusses the role of the Tattvasawgraha and its lineage in both the

teachings and practices established by Kukai. In this ?art the relation-

ship b~tween the Tattvasawgraha, the Vajradhatu ma~~ala and the Vajra-

dhatu Recitation Manuals will be summarized as now understood by Japa-

nese scholars. Moreover, the key to understanding the relationship be-

tween the Vajradhatu ma~~ala and the Vajradhatu Recitation Manuals will

be explained. Part 3 is devoted to the first annotated, English trans-

latio:t of Part 1 of the Tattvasamgraha.

1F. D. Le:;sing and Alex Hayman, Introduction to the Buddhist Tan-

tric Systems, translated from Mkas Grub Rje's Rgyud sde spyihi _!nam }Jar
g~ag .E.!~ par brjod (Delhi: Motilal Banarsidass, 1978), p. (215).

2Hakeda, Kukai: ~ Wo=ks, pp. 61-100.

 8

PART 1
The TATTVAS~GRAHA

Chapter I The Text

Sanskrit, Chinese and Tibetan Editions,

Commentaries and Important Reference Works

The following editions of the Tattvasawgraha and commentaries

were consulted for the English translation. Complete Sanskrit edi-

tions of the Tattvasawgraha were first discovered only within the last

fifty-three years. Giuseppe Tucci first discovered a complete San-

skrit manuscript written in Devanagari in Nepal in 1932. David L.

Snellgrove also discovered another complete manuscript written in

Brahmi in Nepal in 1956 which was published in an edited and roma-

1
nized edition by Isshi Yamada in 1981 as noted below. It should be

noted that Amoghavajra's translation of the Tattvasawgraha which Kukai

introduced to Japan corresponds to just Chapter 1 of Part 1 of the

complete text. The first complete Chinese translation of the Tattva-

sawgraha was made by Sego (active in the late tenth and early eleventh
2
centuries) by 1015 A.D. or possibly the late tenth century but this

translation never influenced the Shingon school. Kukai never obtained

any part of the original Sanskrit text or a copy thereof nor did he in-

traduce to Japan the entire text. Because Part 1, Cha~ters 1 through 5

1Kanjin Horiuchi has argued that Tucci's manuscript is based on

the manuscript in Brahmi (dated ca. 1000 A.D. by Snellgrove) discovered
by Snellgrove. See "Shit tan no Kongochokyl5 Bon.pon ni tsuite, Snellgrove
hon (Hodgeshi Shashinhon) o miru," Mikkyo Bunka, No. 106, 1973, pp. 96-
68. The first photogra~hic reproduction of Snellgrove's manuscript
made public was published by Sakai Shiuten, Bon.p:mshoe ~ Kongochokyo ~
hon (Koyasan: Henjo Koin Rekisei Zensho Kankokai, 1979).

2T. 18, p. 445; Cf. Buddhist Text Informativn, No. 35, 1983, p. 5
and KZ>zen Takeuchi, "Sodai Honyaku Kyoten no Tokuchu ni tsuite," Mikkvl5
Bunka, No. 113, 1975, pp. 27-94 (Fu Sodai Honyaku Kyoten Hennen Moku-
roku). Jap. Sego:~f~! ; Sanskrit? (Danapala ?). Born in 09~iyana, India.
 9

is a single unit in the Sanskrit text corresponding to the

Tathagata family (Skt. Kula) I will translate all of Part 1, found in

Chinese only in Sese's translation and which has no complete Japa-

nese translation. The manuscripts of the Tattvasawgraha are written

in Classical Sanskrit.

SANSKRIT EDITIONS

A) Kanjin Horiuchi, ed., Shoe Kongochok~ no Kenkyu, Bonp~n

Koteihen Bo3zokan Taisho, Vol. 1, Part 1, Chapters 1 through 5

(Koyasan University: Mikkyo Bunka Kenkyujo, 1983), pp. 1-314. The

complete text is edited in two volumes. This work is based on a

comparison of both manuscripts of the Tattvasaigraha and is a revision

of Horiuchi's earlier edition published in Mikkyo Bunka (Vol. 90ff.)

and Koyasan Daigaku Ronso (Vol. III) which was based only on Tucci's

manuscript. A modern Japanese rendition of Amoghavajra's three chUan

translation, based on Horiuchi's revised work, was recently published. 1

It should be noted that Horiuchi follows the commentaries of Anandagar-

,
bha and Sakyamitra noted below in his analysis of the text. With this

work Horiuchi has now established an authoritative edition of the

Tattvasawgraha and it is his edition which served as the basis for the

present English transla~ion. The romanized edition of Part 1 is given

in Appendix II.

B) Isshi Yamada, ed., Sarva-tathagata-tattva-sangraha-nama-maha-

,
yana-sutra, Part 1, Vol. 262 of the Sata-pifaka Series, Indo-Asian Li-

teratures (New Delhi, 1981), pp. 3-151. This edited and romanized ver-

lRyobu paikyo, Vol. II (Tokyo: Shingonshu Buzanha Shumujo, Kobo

Daishi Sengohyakugojunen Goonki Kinen Jigyoiinkai,1983), pp. 273-360.
Amoghavajra's translation was remarkably faithful to the original Sanskrit
text.
 10
sian of the Tattvasawgraha is based only on the manuscript dis-

covered by Snellgrove in Nepal. 1 This manuscript and that dis-

covered by Tucci are identical in content. Yamada's edition is

valuable in giving in the footnotes every variation from the San-

skrit in Amoghavajra's, Sego's and the Tibetan translations. Both

Yamada and Horiuchi note the equivalent sections in the Tibetan and

Chinese translations of the original Sanskrit text. 2

CHINESE TRANSLATIONS

A) _ghin kang ting i ch'ien ~ lai chen shih she ta hsien cheng

ta chiao wang ching, T. 18, No. 865, pp. 207a-223b, three chGan, trans-

lated by Amoghavajra in 754 A.D. 3 This corresponds faithfully to

Chapter 1 of Part 1 of both Sanskrit editions. 4

B) !. ch' ieh ~ lai shen shih she ta ch' eng hsien cheng san mei

ta chiao wang ching, T. 18, No. 882, translated by Sego (Danapala?) by

the late tenth or early eleventh century in thirty chuan. This is also

a faithful and complete translation of the Sanskrit text and was intro-

duced to Japan by the Japanese Tendai monk Jojin (b""£~n in 1081 A.D.
Part 1 in Sego's translation corresponds to pages 341-369 (according

to Sego's division, chuan no. one to chuan no. eight out of a thirty

chuan text, pp. 341-445).

1see David L. Snellgrove, Sarva-tathagata-tattva-sangraha Facsi-

mile Reproduction of ~ lOth Century Manuscript from Nepal, No. 269 in
the Sata-pi!aka Series (New Delhi, 1981).
2Motilal Banarsidass plans to publish in late 1984 another edition
of the Tattvasawgraha. in Devanagari by Lokesh Chandra. Only if neces-
·sary will reference be made to this. new edition if it becomes available
before this thesis is completed.

3T. 55, No. 2157, p. 881.

4Translated into Japanese by Kojun Tomita in Kokuyaku Issaikyo,

(hereafter abbreviated KIK), Mikkyobu II (Tokyo: Daito shuppansha, 1931),
 11

C) Chin kang ting x!! ch'ieh chung lia ch'u nien sung ching, T.

18, No. 866, translated by Vajrabodhi (A.D. 671-741) in 723 in four

chuan. 1 This text was probably brought· to Japan in the Nara period

(710-94) by the Japanese Genbo. 2 It is not properly speaking a tr~ns­

lation of the Tattvasawgraha but a somewhat unorganized and partial ~ut­

line of major practices in the Tattvasawgraha lineage. In contrast to

the Tattvasawgraha translated by Amoghavajra this text explains in

greater detail how to make and enter the central assembly of the Vajra-

dhatu ma~~ala, how to perform a homa or burnt offering, etc. This text

and Amoghavajra's translation are the first two texts Kukai lists in

his Sangakuroku and so this indicates the importance he attached to

both of them. This text begins by saying "I. .. will briefly explain

the teachings of the highest secrecy of the Sarva-tathagata-tattva-

sawgraha." Because there are many sections with the identical word-

ing of the Sanskrit text, Part 1, Chapter 1, and Amoghavajra's

translation, and because Ennin and Donjaku quote these sections along

with their quotations from Amoghavajra in their section by section

analysis of the first chapter, this text cannot be ignored (e.g., the

long description of the thirty-seven deities in the ceutral assembly

of the Vajradhatu ma~~ala). Horiuchi has identified all passages in

this text corresponding to specific passages in the Sanskrit.

In his commentary listed below (beginning on p. 38a, 1.7-20) on

the thirty-seven deities of the central assembly of the Vajradhatu ma~-

9ala Ennin gives two variant translations of thi~ section of the Tattva-

pp. 152-206.

1T. 55, No. 2157, p. 875a, 1.6.

2Ryoko Kushida, "Naracho no Himikkyo Ko," Mikkyo Ronso, No. s,

 12
samgraha in addition to Amoghavajra's which he refers to as the old

sutra and the different text. Donjaku tells us in his commentary

given below (p. 213b, 1.6ff.) the old sutra refers to the four ch~an

translation (i.e., T. 18, No.866) while the different text refers to

the six volume work. The latter (not found in the Taisho canon)· is.

the six volume translation of T. 18, No. 866 which, according to Annen,1
was revised to form the present day four chuan text. A copy of the six

volume work was only recently discovered in Japan by Shujo Shikii in

September, 1981. The exact relationship between the four and six chuan

texts and the Tattvasawgraha is currently being investigated in Japan.2

TIBETAN TRANSLATION

Because this writer cannot read Tibetan the following Tibetan

translation was not consulted for this thesis.

A) De bshin g~egs .E.§!. thams cad kyi de kho na nid bsdus .E.§!. shes

bya ba theg .E.§!. chen ~ mdo, translated by Sraddhakaravarman and Rin

chen bzan po (A.D. 958-1055) in the early eleventh century. Peking

-
edition, Bka9 Vgyur Mdo Na (VIII), 1h1-162ba: Tibetan Tripifaka, No.

112, Vol. 4, pp. 217-83; Snar-than edition, Bka~ ~gyur Mdo Ja (VII) ,

1936, pp. 41-94; Text translated into Japanese by Ryujo Kamibayashi,

KIK, Mikkyobu I (1937), p. 213ff.
1Mikkyo Daijiten (hereafter abbreviated MDJ) (Kyoto: Hozokan,
1983)' p. 713.
2see Ryoshii Misaki, "Kongocho yuga churyaku shutsu nenjukyo ni
tsuite no Ikkosatsu," Tendai Gakuho, No. 24, 1982, pp. 29-42; Jakuun
K1yoda, "Kongoc:ho ryaku shutsu nenjukyo ni tsuite, Rokkanhon to s'!:likan-
hon to no Hikaku," Indo BukkyiSgaku Kenkyii, Vol 30, No. 1, 1981,
pp. 87-92; Ibid., "Kongocho ryaku shutsu kyo no Bonkan Taishc ni tsuite,"
Eizangkuin Kenkyu Kiyo, No. 5, 1982, pp. 59-86; Ibid., "Kongochokyosho
no Inyosho ni tsuite," Tendai Gakuho, No. 23, 1981, pp. 24-29; Yukio
Hatta, "Mandara no Sekai," Gendai Mikkyc Koza, Vol. 5 (Tokyo: Daito,
1977), pp. 137-46. The six volume work will soon be published.
 13
AI
213a4-440. Sde dge Edition, Bka~ ~gyur Mdo Na (VIII), lbl-142a7:

Tohoku Catalogue (Sendai: 1934), No. 479. Chapter 1 of this transla-

tion was translated into Japanese.!

J~~ANESE COMMENTARIES ON THE TATTVAS~GRAHA

1) Kukai, Kongochogyo .kaidai, T. 61, No. 2221, pp. 1-5; Kobo

Daishi Zenshu (hereafter abbreviated KBZ), Vol. I (VIII volumes; Koyasan:

Mikkyo Bunka Kenkyujo, 1965-67), pp. 690-712; KCZ, Vol. 2, pp. 255-278.

2) Ibid., Kong~chogyo £kyo kaidai, T. 61, No. 2222, pp. 5-7;

KBZ, Vol. I, pp. 713-721; KCZ, Vol. 2, pp. 279-290.

Bo~h these short commentaries discuss only the title of the Tattva-

sa~graha as translated by Amoghavajra (T. 18, No. 865).

SUBCOMMENTARIES on Kukai's first commentary

3) Gi::!hi::! (1306-1362), Kongochogyo .kaidai konsha hizoyobunshu, sz,

Vol. 22, No. 77, ten chuan.

4) Yukai (1345-1416), Kongochi::!gyt! ·kaidaisho, Vol. 16 of the Nihon

Daizokyt! (hereafter abbreviated NDZ) (51 Vols., Tokyo: Nihon Daizokyo

Hensankai, 1914-21), pp. 6-128.

Yukai says (p. 95a) there are sixteen different ways of interpret-

' title in the Shingon sect.

ing a scriptures Actually there are eight

which are applied in pairs. Kukai uses many of these in his commentary.

These are as follows: 1) eliminating passions and manifesting virtues;

2) the shallow and profound; 3) a letter's attributes and a letter's

meaning; 4) one letter uniting many and many letters returning to one;

5) one letter explaining many and many letters explaining one; 6) one

letter completing many and many letters completing one; 7) one letter

destroying many and many letters destroying one; 8) sequential order and

1Eihan Sakano, Kongochokyo ni Kansuru Kenkyu (Tokyo: Kokusho

Kankokai, 1976), pp. 107-213. This translation was read.
 14

reverse order. (See MDJ, p. 897.)

5) Emdn (793-864), _!{ongocho daikyo okyoso, T. 61, No. 2223,

pp. 7-114. This has been translated into Japanese by Kyojun Shimizu-

tani in KIK! Kyoshobu, Vol. 16 (1938), pp. 35-350 (Vol. 17 in the 1981

edtion by Daito shuppansha). This commentary on AmoghAvajra's trans-

lation by the Tendai monk Ennin was the only sizable commentary

available in Japan un.til Donjaku's (1674-1742) commentary listed be.low

and thus traditionally has been consulted by both Shingon and Tendai

scholars. This commentary, however, only covers part of the trans-

lation (in Amoghavajra's translation, up top. 217b, 1.15; in Horiuchi's

edition, p. 115, section 209). Between 839 and 847 Ennin studied and

traveled in China. His commentary probably reflects either the oral

tradition transmitted to him by Yuan ching ~£Jl) of the Ta hsing chan

( ~~) temple who initiated Ennin into the Vajradhatu ma~9a1a or

the oral tradition of Ch' uan ya ~ fi) of the K' ai yuan ( )i.ft'Cij jG ) tern-

ple who initiated Ennin into the rites of the various deities shown in

the Vajradhatu ma~Qala. 1 (See also Appendix V, p. 518ff.)

6) R&ison (active in the eleventh century;itZ ~), Kongochokyo

~...:a-
kesshaku, T. 61, No. 2224, pp. 114-117. This is a brief commentary on

the enigmatic list of terms at the beginning of the Tattvasawgraha

(Horiuchi's editon, section 11) 'li'hich basically repeats what Ennin says.

7) Donjaku (1674-1742;~ ~ ), Kongocho daikyo £kyo shiki, T.

61, No. 2225. This is the most complete Japanese commentary covering

all of Chapter 1 , Part 1, or all of Amoghavajra's translation (in

Horiuchi's editon, up top. 215, sections 317-8). Donjaku copies

lEdwin 0 Reischauer, Ennin's Diary The Record of a Pilgrimage

to China in Search of the Law (New York: Ronald Press-co., 1955), pp.
79-80, 294; MDJ, appendix entitled Mikkychc ryu keifu, p. 2.
 15

verbatim Ennin's commentary and then adds further commentary of his

own. There are no commentaries available in Chinese or Japanese on

Chapters 2 through 5 of. Part 1.

8) Jiun (1718-1804; G. i!it),

II.;.;J ~
Kyo £_ kyo shaku, Jiun Sonja Zenshu,

Vol. VIII, compiled by Hoshu Rase (Osaka: Kokiji, 1925), pp. 54-61.

Jiun attempted to restore part of the Sanskrit text on the basis of

Amoghavajra's translation (Jiun Sonja Zenshu, Vol •. IX, Pt. II, p~

364ff.)

INDIAN COMMENTARIES IN TIBETAN!

1) Buddhaguhya (middle eighth century A.D.), Tantrarthavatara,

Tohoku Catalogue No. 2501. This is not a varse by verse analysis of

the Tattvasawgraha but rather a general commentary on the key doctrines,

terms and practices associated with the sutra.

2) Padmavajra (second half of the eighth century), Tantrartha-

avatara vyakyara, Tohoku Catalogue No. 2502. This is an extensive sub-

commentary on Buddhaguhya's commentary. Both the above are extensively

2
analysed and translated into Japanese by Takeo (Taido) Kitamura.

3) Sarvarahasya Tantra, Tohoku Catalogue No. 481. Translated

into English by Alex Wayman. 3 Wayman writes "The history of this tantra

lsee F. D. Lessing and Alex Wayman, Buddhist Tantric Systems,

pp. 24-25, n. 13, pp. (214)-217. lfuenever commentaries in Tibetan
are available in English and Japanese translations these have been con-
sulted. There are numerous works preserved in Tibetan classified as
Yoga tantras. See Alex Wayman, The Buddhist Tantras, p. 236.

~"Tantrarthavatara o Chushin to shita Kongochokyo no Kenkyu,"

Mikkyogaku, Part I, No. 7, 1970, pp. 1-22; Part II, No. 8, 1971, pp. 1-
34, Part III, No. 9, 1972, pp. 1-41; Part IV, No. 10, 1973, pp. 15-69,
Part V, No. 11, 1974, pp. 1-86; Part VI, No. 12, 1975, pp. 1-66.

3Indo Katen Kenkyu Acta Indologica VI Shimpishisoronshu Mysticism

(Naritasan: Shinshoji, 1984), pp. 521-569. This was also early trans-
lated into Japanese by Shoun Toganoo in Mikkyo Bunka, Nos. 24-25, 1953,
pp. 3-67.
 16
is obscure; it may well have been composed along with the Tattva-

sawgraha in the eighth century A.D. i f not earlier." This is an

"explanatory tantra" on the side of means of the Yogatantra.

,
4) Sakyamitra, Kosalalamkara, Tohoku Catalogue No. 2503.

5)
-Anandagarbha, Tattvaloka, Tohoku Catalogue No. 2510.

The two above authors may be tentatively placed in the tenth

century. These two works are often referred to by Horiuchi in his re-

vised edition of the Tattvasawgraha and by Kitamura in his articles

mentioned above.

TIBETAN COMMENTARY

A) Introduction to the Buddhist Tantric Systems, translated from

Mkas Grub Rje' s Rgyud sde spyi}Ji rnam par g{ag .F.! rgyas par brjad with

original text and annotation, F. D. Lessing and Alex Wayman (Delhi:

Motilal Banarsidass, 1978); Tohoku Catalogue No. 5489. Mkas grub rje

lived from 1384 to 1438. Chapter six concerns the Yoga tantras of which

the Tattvasawgraha is considered the fundamental text in the Tibetan

tradition.

SOME IMPORTANT, SECONDARY REFERENCE WORKS

1) Chin kang ting ching ta _y!! ch' ieh .E.! mi hsia ti fa ~.! chueh,

T. 39, No. 1798. This is supposedly Vajrabodhi's oral commentary on his

translation above (T. 18, No. 866) as recorded by Amoghavajra. Only the

first chuan of an original three chUan work remains. This was brought

from China to Japan by both Kukai and Enchin (814-891).

2) Chin kang ting ching ~ ch' ieh shi.h ~ hue chih kuei, T. 18,

No. 869, translated by Amoghavajra. This text gives the first outline

of the supposed eighteen assemblies of the Tattvasawgraha.

3) Tu £!!. t'o lo ni mu, T. 18, No. 903, translated by Amoghavajra

 17

(KIK, Mikkyobu, II). This gives a summary of key doctrines and

practices in the Tattvasawgraha as well as an outline of the map~alas

described in the sutra.

4) Liao shu chin kang y!!_ ch'ieh fen pieh wei hsiu cheng fa men,

T. 18, No. 870, translated by Amoghavajra (KIK, Mikkyobu~ II).

5) Chin kang ting yg_ ch'ieh lian shu san shih ch'i tsun hsin ~·

T. 18, No. 871, translated by Amoghavajra (KIK, Mikkyobu, IV).

18, No. 872, translated by Amoghavajra (KIK, Mikkyobu, III).

7) Goho ( 1306-62; l f ), Kongocho shukogai, T. 77, No. 2451.

Several excellent introductions to the Tattvasawgraha have recently

been published and these are considered essential reading for this the-

sis. David L. Snellgrove 1 has given a thorough, chapter by chapter

description of the entire Tattvasawgraha. His discussion includes trans-

lated excerpts, several charts classifying by family the numerous dei-

ties found in the different ma~~alas described in the text and a compari-

son and clarification of the different part and chapter divisions in the
2
Tibetan and Sanskrit editions. Matsunaga has given fine introductions

to basically Chapter One and several important, related topics from the

perspective of the Chinese and Japanese traditions of the Tattvasa~graha.

Matsunaga's explanations are largely based on the authoritative studiesof

1snellgrove, Facsimile Reproduction, pp. 5-67. Snellgroves con-

tention (p. 38) that sexual rites are involved in Chapter Five is not
supported by any of the commentaries.

2yrrkei Matsunaga, Mikkyo Kyoten Seiritsu Shiron (Kyoto: ~oz~kan,

1981), pp. 187-198; "Mikkyo Kyoten Kaisetsu," in Gendai Mikkyo Koza,
Vol. II (Tokyo: Daito shuppansha, 1981), pp. 123-165. See also Shoun
Toganoo, "Kongl5cht:lkyt:l Gaisetsu," in Nihon Shukyo Daikoza, Vol. VII
(Tokyo: Tohoshoin, 1930), pp. 1-50 and Ryujun Tajima, Les Deux Grands
MatJ.s;lalas et la Doctrine ~~- !~~.E-~~!:.~.!1~. _Shil!.SE.!!. (Presses Universitaires
de France, 1959).
 18

Munetada Suzuki. 1 The pioneering work in English on Kukai and the

Shingon school of Hakeda and the more recent work by Minoru Kiyota

are also essential background texts for this thesis. 2

FINAL COMMENTS ON THE TRANSLATION

Because the Japanese commentato~s never obtained a Sanskrit

manuscript of the Tattvasawgraha their commentaries refer only to the

Chinese translations. Kukai and Ennin refer to Amoghavajra's trans-

lation while Donjaku, having available to him Sego's later translation

as well, refers to the two translations by Amoghavajra and Sego. By

and large the Chinese translations are faithful to the Sanskrit text

although on occasion certain words, names or phrases are not always

translated. Because Yamada's edition of the Tattvasa~graha lists in

the footnotes every instance where either the Chinese or Tibetan trans-

lations differ from the Sanskrit the reader is referred to his work if

a comparison between the various editions is sought. For the benefit

of the reader of Chinese both translations by Amoghavajra and Sego

are included in Appendix III and IV respectively.

In the body of the annotated translation only when the Japanese

commentators refer to some word(s) or passage(s) in either Chinese

translation which is not found in the Sanskrit will the Chinese be

supplied. Thus, for example, because Kukai comments on Amoghavajra's

Chinese translation of the title which contains some words not in the

Sanskrit, the Chinese will be provided. If thete is no difference be-

tween the Sanskrit and either of the Chinese translations annotated

1suzuki Munetada Shosakushu, Vol. y, Kihon Daij~ Himitsu Bukkyo

(Tokyo: Gannando shoten, 1978); Munetada Suzuki, "Kongochokyo no Jitsu-
zai Keitai," Shtikyo Kenkyu, Nos. 112-13, 1942, pp. 227-253.

2Hakeda, Kukai: Major Works; Kiyota, Shingon Buddhism.

 19

neither the Sanskrit nor the Chinese is supplied in the translation

and the reader in these instances if referred to the Appendixes for

the original Sanskrit and Chinese translations. Major differences

between the two Chinese translations will only be referred to. For

the benefit of the reader in the Appendixes the corresponding pages

of the English translation in the body of the thesis will be given.

In the translation a hyphenated line separates the commentaries

from the translation. Then a solid line separates the end of the

commentaries from the beginning of the. next section translated.

The Title of the Tattvasawgraha

1) Horiuchi's edition: Sarva-tathagata-tattva-safgrahaw nama

mahayana-sutram; Part 1: Sarva-tath~gata-mahayanabhisamayo nama maha-

kalpa-rajab; Chapter 1: Vajradhatu-maha-mandala-vidhi-vistarab.

2) Yamada's edition: Sarva-tathagata-tattva-sa~graha; Part 1:

Sarva-tathagata-mahayanabhisamaya nama maha-kalpa-raja; Chapter 1: Sar-

va-tathagata-mahayanabhisamayan maha-kalpa-rajad Vajradhatu-maha-ma~qala­

vidhi-vistarab.

3) Amoghavajra's translation, T. No. 865: Chin kang ting! ch'ieh

~ lai chen shih she ta ch'eng hsien cheng ta chieh wang ching; Chapter

1: Chin kang E_hiai taman t'u lo k'uang ta.! kuei ~ (Sanskrit: Vajra-

~ekhara-sarva-tath~gata-tattva-sawgraha-mahayanabhisamaya-maha-kalpa-ra­

~-sutr~; Vajradhatu-maha-map9ala-vidhi-vistarab.)

4) T. No. 869, p. 284c, translated by Amoghavajra: ! ch'ieh ~

lai shen shih che wang chieh (Skt.: Sarva-tathagata-tattva-sa@graha-

raja-kalpa.)

5) Sego' s translatictt"; T. No. 882: Fo shou ! ch' ieh i£ lai shen

 20
shih she ta ch'eng hsien cheng san mei ta chieh wang ching; Chapter

1: Ching kang chiai ta ~ t 'u lo kiuang ta .! kuei. (Skt. : Buddha-

bhasita-sarva-tathagata-tattva-sawgraha-mahayanabhisamaya-samaya-maha-

6) Kukai, T. 61, No. 2221, p. 2b, 1.11. Kukai gives the title

in Siddham 1 which is rendered by Katsumata as (KCZ, Vol. II, pp. 264-5)

Vajra-u§~I~a-sarva-tathagata-samaye mahavawdabhisafbodhi-mahahaw-raja-

sutrem. On the basis of T. 99, p. 288a, Katsumata corrects this as

follows: Vajra-~ekhara-~-tathagata-satya-sawgraha-mahayana-pratyu-

panna-abhisambuddha-maha-tantra-raja-sutram.

7) Ibid., T. No. 2221, p. la, 1.23: Title: Kongocho ~ issai

nyorai shinjitsu sho daijo gensho daikyo ~kyo. Part 1 (Ibid., 1.27ff.):

yuga-sarva-tathagata-tattva-sawgraha-mahayanabhisamaya-maha-kalpa-raja-

sutra; Part 1: Sarva-tathagata-tattva-sawgraha-mahayanabhisamaya-maha-

8) Ibid., T. 61, No. 2222, p. 6a, 1.28ff. Same as Amoghavajra's

translation.

9) T. 18, No. 866, translated by Vajrabodhi, p. 223c, 1.23: Chin

kang ting 1.! y!!, ch'ieh wang chieh (Skt.: Vajra~ekhara-maha-~-raja-

kalpa).

10) Sraddhakaravarman and Rin chen bzan po, Tohoku Catalogue

No. 479: De-bshin-g~egs-pa thams-cad-kyi de-kho-na-nid bsdus-pa shes-

bya-ba theg-pa chen-pohi mdo (Skt.: Sarva-tathagata-tattva-sawgraha-

nama-mahayana-sutra).

11) English translation of the title given by ".ioriuchi: The

1see also Yukai's con®entary for the transliteration of the S .f.:l-

dham. Kukai's Chinese translation of the Siddham is as follows:
--
 21

Mahayana sutra known as the Compendium of Truth of all the Tatha-

gatas; Part 1: The King of Great Practices known as the Mahayana

Realization of all the Tathagatas; Chapter 1: Description of the

Rites of the Vajradhatu Maha-ma~9ala.

The titles of both manuscripts of the TattvasaiPgraha are found

at the end of Part 1, Chapter 5, at the end of Part 2, Chapter 14

and elsewhere. 1 The titles of Part 1 and Chapter 1 are found at the

end of Chapter 1 and the titles of the remaining chapters (Nos. 2-5)

of Part 1 are also found at the end of each chapter in both manu-

scripts. (The title of Part 1 is repeated at the beginning of each

chapter heading of Part 1 as in Yamada's edition.)

Sakano notes that in the Siddham title Kukai gives in his Kongo-

chogyo kaidai he leaves out the two words tattva and sawgraha. 2 Kukai

also adds samaye. In T. 61, No. 2222, he also adds the word~·

Kukai (T. 61, p. 1a, 1.24ff.), Donjaku (T. 61, No. 2225, p. 128a)

and Yukai in his commentary (p. 25) all state that Chin kang ting rl_

ch'ieh, based on T. 18, No. 869, .was a· title used to indicate the
entire supposed eighteen assembly text, notwithstanding each chapter

ha~ its own distinct title. The word~ (~ ch'ieh) 'given by Kukai

in T. No. 61 is also found in the title of each of the eighteen sections

given by Amoghavajra in T. 18, No. 869. Many works translated by

Amoghavajra and Vajrabodhi belonging to the Tattvasa,graha lineage

lThe following discussion is based in part on Jakuun Kiyoda,

"Kongt!chDkyD no bonmyZ> ni tsuite," Indo Bukkyogaku Kenkyu, Vol. 2,
No. 2, 1954, pp. 630-32.

2Eihan Sakano, Kongochokyo ni Kansuru Kenkyu, p. 362. Yukai

argues in his commentary (p. 56) Kukai deleted tattva because its
meaning was incorporated in the word Tathagata. See below, p. 135 •
 22
contain t'ne word~ (T. Nos. 866, 871, 872, 876, 879, etc.) and so

Kukai follows this usage. This word also indicates works dealing

with ritual meditation. 1

From the above the following observations can be made: 1) Kongo-

chokyo (Vajra~ekhara sutra), the common Japanese, abbreviated designa-

tion for this sutra, obviously comes from Amoghavajra's translation

of the title although Kukai mistakenly has u~Fi9a for ~ekhara (See #6). 2)

The words in the title Chin kang ting and Chin kang ting ~ ch' ieh are

found only in tha translations by Amoghavajra and Vajrabodhi. Accord-

ing to Kukai Chin kang (Vajra) refers to the thirty-seven Vajrasattvas

discussed in the sutra. 3) Only Sego gives Fo shou (Buddha bhasita)

in his title, probably in accord with the old Chinese tradition of

ascribing all satras to Buddha. 4) By deleting Vajra~ekhara ~ and

Buddha bhasita from the titles given by Amoghavajra and Sego we arrive

at the Sanskrit title, i.e., Sarva-tathagata-tattva-satpgraha-nama-maha-

y~na-sutra, except for the words abhisamaya-maha-kalpa-rajag. These

obviously refer to Part 1. That is, the title and part headings have

been combined in the Chinese translations. Kukai also combines these

headings (compare the above, 1, 2, 3, 5, 6). The Chapter headings given

by Sego and Amoghavajra are accurate translations of the Sanskrit head-

ings. 5) The Tibetan translation gives the correct title of the San-

skrit text. The additional heading mula-tantra given by Horiuchi is

based on the Tibetan classification of this sutra. The Tattvasawgraha

was also considered an "outer" Yoga tantra.

" ••• Candrakirti in his commentary on Chapter XVII cites

the 'MUla-tantra,' which Tson-kha-pa identifies as the

lyukei Matsunaga, "Mikkyo Kyoten Kaisetsu," Gendai Mikkyo Koza,

Vol. 2, pp. 155-57.
 23

Mula-tantra of the Yoga-tantra, namely the Tattvasawgraha •••

According to The Blue Annals, Book VII ('The Preaching of the
Tantras'), there was-a distinction of 'Outer' Yoga-tantra and
'Inner' Yoga-tantra, with the 'inner' variety becoming sep-
arately called 'Anuttarayoga-tantra.' One can therefore un-
derstand Candrakirti's citation as indicating his adherence
to this terminology, since the Tattvasawgraha is the Mula-
tantra of the 'outer' Yoga-tantra."l

The Structure of the Tattvasawgraha

Both manuscripts of the Tattvas~graha as discovered by Tucci

and Snellgrove contain twenty-six chapters which, although unnumbered,

are named at the end of each chapter. The complete Chinese translation

of the Tattvas~graha made by Sego (T. 18, No. 882) also has twenty-six

chapters. In addition, the manuscripts of the Tattvasa~graha divide all

twenty-six chapters into five parts which are clearly named. The titles

of these five parts and the chapters contained in each part are as
')
follows: ..

1 Sarva-tathagata-mahayanabhisamayo n~a maha-kalpa-rajah (The

King of Great Practices known as the Mahayana Realization of all the

Tathagatas); Chapters 1-5.

2 Sarva-tathagata-vajra-samayo nama mah~-kalpa-rajaij (The King

of Great Practices known as the Thunderbolt Pledge of all the Tathaga-

tas); Chapters 6-14.

3 Sarva-tathagata-dharma-samayo nama mah~-kalpa-rajab (The King

of Great Practices known as the Law Pledge of all the Tathagatas) Chap-

ters 15-18.

1Alex Wayman, Yoga of the Guhyasamajatantra The Arcane Lore of

Forty Verses ! Buddhist Tantra Commentary (Delhi: Motilal Banarsidass,
1977), p. 87.

2see Snellgrove, Facsimile Reproduction, pp. 6-14, for an analysis

of the section divisions of the Tibetan translation and a comparison of
this with the Sanskrit divisions.
 24

4 Sarva-tathagata-karma-samayo nama maha-kalpa-rajap (The

King of Great Practices known as the Action Pledge of all the Tatha-

gatas); Chapters 19-22.

5 Sarva-tathagata-tattva-sawgraha nama mahayana-sutra (The Maha-

yana sutra known as th~ Compendium of Truth of all the Tathagatas),

Chapters 23-26. This title is the same as the title for the entire

stitra as well.

No. 869) translated by Amoghavajra states this sutra is divided into

just four parts and this analysis is also accepted by Kukai ('.C. 61,

No. 2221, p. lb, 1.3). In this regard part five is the only part to

be named siltra in distinction to the first four parts named kalpa-

rajap (in Sego's translation as well) and is clearly described as a

continuation (uttaratantra) in the Tibetan translation. 1 The first

four parts, as suggested by the term kalpa, prescribe ritual and give

rules for various ceremonial acts. These initial four parts are corre-

lated with the four families (kulas) known as Tathagata, Thunderbolt

(Vajra), Law/Lotus (Dharma/Padma) and Action/Gem (Karma/Ratna) re-

spectively.2 The last part known as a siltra

" ••• assumes a knowledge of the actual rituals and repre-

sents a 'thread of discourse' conducted by Buddha (still
understood as Vairocana) and the Bodhisattvas who lead
the various Buddha families. They each in turn pronounce
on the benefits to be won by the many different perfor-
mances as performed within the family circle of each."3

Although Amoghavajra does not describe a fifth part it appears certain

lF. D. Lessing and A. Wayman, Buddhist Tantric Systems, p. 216.

2Taido Ki.tamura, "Kobo Daishi no Shiman ono ono furisets'.l ni

tsuite," MikkyDgaku Kenkyu, No. 6, 1975, pp. 1-10.

3snellgrove, Facsimile Reproduction, p. 9.

 25

this was composed by the eighth century because Buddhaguhya quotes

from this part in his commentary entitled the Tantrarthavatara.

.'
In each of the first four parts four mandalas are described in

accordan~e with the above family division. At the end of each part,

in addition, a catur-mudra (four seal) and eka-mudra (single seal)

1
ma~~ala are described giving six ma~9alas in all. This six ma~~ala

structure is described by Amoghavajra in the above work (T. 18, No.

869). In each of the first four ma~~alas (except for Part 22) the

same basic set of thirty-seven divinities is given, namely, Vairocana

and the four Buddhas of the four directions (East - Ak~obhya; South -

Ratnasa~bhava; West- Loke~vararaja; North- Amoghasiddhi), four pa-

ramita goddesses, sixteen Maha-bodhisattvas or Vajrasattvas, 3 eight

goddesses of offering (puja; four inner and four outer), and four

attracting (sawgraha) deities, i.e., doorkeepers. Although four

Buddha families are discussed throughout the text, these are set

within a five Buddha pattern as evident by the five Buddhas given in

the ma~9alas (the five families are based in parL on the five skandhas).

"The fact that the Fundamental Tantra only describes

four, although there are five Families, is explained by
Buddhaguhya as a merger of action (karma) and agent (ka-
raka) since it has the Ratna Family in the sense of the
agent that accomplishes the wishes of the living beings

1see Minoru Kiyota, Shingon Buddhism, pp. 93-104, for a descrip-

tion of these six ma~~alas.
2snellgrove, Facsimile Reproduction, p. 11 and p. 47ff.

3Although Tattvasawgraha lineage texts refer to these sixteen

Vajrasattvas indiscriminately as both Bodhisattvas and Mah~-bodhi­
sattvas (E.g., T. 18, No. 871, p. 292b, c), and the Japanese tradi-
tion refers to these regularly as Bodhisattvas, the Tattvasawgraha
itself never calls them Bodhisattvas but Mahl:l-bodhisattvas. More-
over, these are knowledge beings (j~na-sattva; T. 18, No. 869, p.
284c, 1.22) which in this case, according to the Tibetan commentaries,
are considered corporeal manifestations of Vairocana. See F. D.
 26
and has the Kanna Family in the sense of the action that
does so."1

The four Buddha families are correlated with four types of map9a1as

(Maha - Great; Samaya - Symbolic; Dharma - Law; ~arma - Action)2 and

four knowledge seals (j~na-mudr~) similarly named. 3 The above

thirty-seven deities each are endowed with these ma99alas and know-

ledge seals which are mutually interpenetrating 4 and the practitioner

who realizes these also realizes the four types of Dharmak~yas, i.e.,

Svabhava, SaUJ.bhogakaya, Ninnagak~l! and Ni§yandakaya(See below, p. 114.).

Both Isshi Yamada and Kanjin Horiuchi follow the above five part,

twenty-six chapter division of the Tattvasawgraha in their editions of

the sutra. Horiuchi also gives a detailed sub-division breakdown of

the entire text. In the translation of Part 1 in this thesis generally

it is Horiuchi's sub-divisions which will be followed because of their

practical advantage. Although these sub-divisions are not given in

the manuscripts they are obvious from the explanations of the

commentaries and the different topics discussed in each section.

The commentaries on the TattvasaVlgraha by Ennin (p. 17c, 1.20ff.)

and Donjaku (p. 162a-c) distinguish an introduction (Jap. Josetsubun -

\~ ~~..:'"'
,..
IJ ~IL n ),
L'
the body of the text (Jap. Shosetsubun -if.~ n) and the

Lessing and A. Wayman, Buddhist Tantric Systems, pp. 162-63.

1F. D. Lessing and A. Wayman, Buddhist Tantric Systems, p. 217.

Cf. T. 18, No. 864, p. 287c and T. 18, No. 903, p. 898c. Snellgrove
thinks this explanation is a subsequent justification. See Snellgrove,
Facsimile Reproduction, p. 12. See also Yukei Matsunaga, Mikkyo KyO-
ten Seiritsushiron, pp. 121-131 and Kanjin Horiuchi, "Shichi shibutsu
ni tsuite," Mikkyo Bunka, Vol. 144, 1983, pp. 40-63.

2on the four ma~9alas see Hakeda, Kokai: Major Works, pp. 90-91.

3an the four knowledge seals see F. D. Lessing and A. Wayman,

Buddhist Tantric Systems, pp. 229-249.

4Kcz, Vol. 2, p. 266.

 27

conclusion (Jap. Ruzubun - ~~ ti. . ) . They both agree Amoghavaj ra' s

translation lacks the conclusion. They further divide the introduc-

tion into two parts: a "preface" (Jap. Tsiijo -M) and "formal intro-
duction" (Jap. Betsujo - KlJfl ). 1
Ennin maintains the "preface" runs from the beginning of the text

down to "evaw pramukhair ~-navatibhir bodhisattva-kotibhi\1. n2 The

"formal introduction" then continues down to "sarvagryo Bhunefvarah. " 3

Then from the word "Bhagavat"4 the body of the text begins.s Donjaku

agrees with this division. Jiun says only that the introduction runs

down to "Sarva-tath~gata-htdaye!?U vijahara,"S just one sentence more

than where Ennin says it ends.

The Indian commentaries also distinguish an introduction divided

into a "preface" and "formal introduction." 6 According to Sl!kyamitra7

the formal introduction begins wfth the words "Atha Bhagavan Mahavai-

rocanap."8 and ends with the same passage as indicated by Jiun above.

Buddhaguhya and Anandagarbha also follow this interpretation.9

1sakano notes this analysis of a sutra was in vogue among Buddhist

coln..-nentators by the sixth century A.D. in China. Eihan Sakano, Kongo-
chokyo ni Kansuru Kenkyii, p. 68ff.
2
Hoiuchi's edition (hereafter abbreviated H.) p. 2, section 5;
Amoghavajra's translation (hereafter abbreviated A.) p. 207a, 1.22.
3H., p. 15, section 16; A. p. 207c, 1. 7 0

4H., p. 22, section 17; A. p. 207c, 1.8.

sH., p. 22, section 17; Kyo .£ kyo shaku, in Jiun Sonja Zenshu,
Vol. 8, p. 54.

6Horiuchi's edition follows the Tibetan interpretation.

7Kanjin Horiuchi, "Kongochl5kyl5 no Betsujo ni tsuite," Indo Bukky_C-

gaku Kenkyn, Vol. 15, No. 1, 1966, p. 44.

8s. p. 9, section 7; A. p. 207a, 1.26.

9Takeo Kitamura, "Tantrarthavatara o Chushin to shita Kongl5chl5kyo

 28

The Indian commentaries make this division because they re-

cognize a distine:t.iuu between ~he f'laha-~ Vairocana appearing in

the "preface" and the Mahavairo~ana appearing in the "formal intro-

du<.:tion." In the formal introduction the intrinsic nature (svabhava)

of Mahavairocana is manifested as a Knowledge body (jn~nakaya) whereas

Vairocana, because he i3 described as existing in'the Akani~tha heaven,

is a Sawbhogakaya.

Although Kukai was aware of and used the above threefold analysis

of a sutra 1 he does not make use of it in his commentaries on the title

of the Tattvasafgraha.

A full summary of the Chapters and subdivisions of Part 1 of the

Tattvasawgraha is given below (p. 128) just before the translation

as a Table of Contents.

The Date and Place of Origin of the Tattvasawgraha

Jrr~namitra's Arya-praj~aparamita-naya-latapa~ca~atika-tika states

that the Tattvasawgraha appeared first in the reign of King Indrabhuti 2

who is dated from the end of the seventh to the first half of the eighth

century. 3 Buddhaguhya, who belongs to the middle of the eighth cen-

tury A.o.,4 was aware of all five parts of the Tattvasawgraha in his

no Kenkyu, I," Mikkyogaku, No.7, 1971, pp. 11-12.

1T. 61, No. 2237, p. 612c, 1.14ff.

2shoun Toganoo, "Himitsu Bukkyoshi," in Gendai Bukkyo Meicho Zen-

shu, Vol. 9, Bukkyo Ippan, I., ed. by Hajime Nakamura, et. al. (Tokyo:
Ryubunkan, 1962), pp. 36, 40.

3Alex Wayman, Yoga of the Guhyasamaiatan~ra, p. 96.

4F. D. Lessing and A. ~ayman, Buddhist Tantric Svstems, p. 25.

 29

Tantrarthavatara and he quotes from part five. 1 The Chin kang ting

ching .lli ch'ieh shih e hue chih kuei (T. 18, No. 869) translated by

Amoghavajra states the Tattvasawgraha is divided into just four parts

however. Amoghavajra's faithful translation of Part 1, Chapter 1, of

the Tattvasawgraha, which generally d~scribes the central assembly of

the nine assembly Vajradhatu ma~9ala, is dated 754 A.D.2 Moreover,

the major part of the texts transmitted and translated by Amoghavajra,

which are mostly Tattvasaqtgraha lineage texts, are thought to have been

composed in India between 713 and 742 A.D. Amoghavajra traveled to

South India and Ceylon in 741 A.D. and returned to China in 746 A.D.,
3
bringing these with him. The Chen yuan hsin· ting ~ ~.J!lli. lu4

says that when Vajrabodhi (671-741) was thirty-one he went to South

India and studied the Tattvasawgraha for seven years. This would have
5
been in 701 A.D. The Chin kang ting ~ ch'ieh chung lioh chu nien

sung ching (T. 18, No. 866) was translated by Vajrabodhi in 723 A.D. 6

Although this is not a translation of the Tattvasawgraha it includes

an almost identical discussion of the thirty-seven deities of Chapter

1 as found in Amoghavajra's translation and the original Sanskrit text.

1Takeo Kitamura, "Tantrarthavatara o Chushin to shita Kongochokyo

no Kenkyii, I," Mikkyogak~, No. 7, 1970, pp. 6, 14-15; Ibid., II, Mikkyo-
z.aku, No. 8, 1971, pp. 3, 6, 11, 19. etc.

2T. 55, No. 2157, p. 881; MDJ, p. 1897. ~}t.ltf1:jl~§~.

3Yukei Matsunaga, Mikkyo no Sojosha Sono ~Kodo to Shiso, Toyojin

no~ to Shiso, Vol. 3 (Tokyo: Hyoronsha, 1973), pp. 178-79.

4T. 55, No. 2157, p. 875b.

5ynkei Matsunaga, Mikkyo no Sojosha, pp. 146-47.

6T. 55, No. 2157, p. 875a.

 30
It also discusses the second and third assemblies based on Chapters

2 and 3 (T. 18, pp. 240-43) and describes in greater dt:!tail Lhau

Amoghavajra's translation how to draw the central assembly.

The central assembly in basic to all Vajradhatu ma~~alas. Its

structure is repeated in six out of the nine assemblies in the Vajra-

dhatu ma~~alas in Japan. The ninth century Vajradhatu

- ..
mandalas in

Japan showing nine assemblies are known to be based on Part 1, Chap-

ters 4 to 5 (the basis for assemblies 1 to 6) and Part 2, Chapters

6 to 7 (the basis for assemblies 8 and 9), of the Tattvasawgraha

(see below, p. 66).

The weight of the above evidence points to the late seventh cen-

tury1 or possibly the beginning of the eighth as the date for the

composition of Part 1. On the basis of Buddhaguhya's quotations

from Part 5 and Amoghavajra's mention of four parts of the Tattvasaw-

graha it is apparent the entire text was completed by the middle of

the eighth century as well.2 Toganoo quotes a letter from Amoghavajra

to the Emperor Tai-tsung in which he laments .that· he has not. translated

the entire Tattvasawgraha due to his old age.3

All five parts of the Tattvasa~graha were translated into Chinese

for the first time by Sego possibly by the late tenth century A.D. and

thes~we should be reminde~were translated into Tibetan in the early

eleventh century by Sraddhakaravarman and Rin chen bzan po.

1cf. Gemmyo Ono, "Kongocht!kyo no Seiritsu ni tsuite," Ono Gemmyo

Bukkyo Geijutsu Chosakushu, Vol. 4 (Tokyo: Kaimyo Shain, 1977), pp.
488-506.

2Yukei Matsunaga, Mikkyo no Sojosha, p. 143.

3shoun Toganoo, "Kongochokyo Gaisetsu," Nihon Shiikyo Daikoza,

Vol. 7 (Tokyo: Toho Shain, 1930), p. 7; Cf. KCZ, Vol. II, p. 72.
 31

That Part 1 was completed by the early eighth century is

corroborated further by the iconographical work entitled Ritasogyara-

~bushinkan 1 (or simply Gobushinkan) • Japanese scholars agree the Go-

.
bushinkanwas drawn by ~ubhakarasimha 2 (637-735 A.D.) who entered

Chang-an in 716 A.D. This depicts a six assembly Vajradhatu ma~9ala

whose structure is outlined in all of Part 1 of the Tattvasawgraha.

In contrast to the ~ypical Japanese Vajradhatu ma~~alas this agrees

with the Indian commentaries in designating animal mounts3 for the

five central Tathagatas.4 Animal mounts are also prescribed in

Vajrabodhi's translation.s

Explicit instructions for drawing the various assemblies of the

Vajradhatu mav~alas are not found in the original Sanskrit text of

the Tattvasawgraha. However, the structure it describes is followed

in the ma~9a1as. This makes it possible to date the text on the basis

of Vajrabodhi's translation and the Gobushinkan. As is the case with

the majority of tantras, the author of the Tattvasawgraha is unknown.

Based on the Chinese accounts of the activities of Vajrabodhi and

Amoghavajra6 most Japanese scholars who have studied the Tattvasawgraha

1T. 87, No. 2975. See Hatta Yukio, Gobushinkan no KenkyU.

2Hisatoyo Ishida, Mandara no Kenkyu, (Tokyo: Tokyo Bijutsu, 1975),

p. 244.
3Yukei Matsunaga, "Kongokai Mandara ni tsuite, Re Shuhen Jiin no
Chosha Hokoku," Mikkyogaku Kenkyii, No. 10, 1978, pp. 83-100.

4The position of the five Tathagatas in the center of the ma~Qala

can be traced back toT. 20, No. 1092 (p. 346b), translated by Bodhi-
ruci in the first half of the sixth century in China. See Yukei Matsu-
naga, "Ry~bu Mandara no Shifu, Gobutsu o Chtlshin to shite," Mikkyo Bun-
ka, No. 87, 1969, pp. 27-36.

5T. 18, No. 866, p. 227b.

6T. 50, No. 2061, pp. 711b-714; T. 55, No. 2157, pp. 875-76, 879.
See Chou Yi-liang, "Tantrism in China," HJAS, VIII, 1945, pp. 241ff.
 32
are of the opinion that it was composed in South or Southeast India. 1

Vajrabodhi may have been born in South India and

studied the Tattvasawgraha there. It is particularly significant that

Amoghavajra, perhaps fulfilling the wishes of his master Vajrabodhi,

traveled to Souti1 India and Ceylon ca. 741 A.D. in order to obtain

more texts of the Tattvasawgraha lineage. 2 Whether or not Amoghavajra

brought back to China 500 texts as it is recorded it is clear fr~m the

list of works thought to have been translated by him one of the main

goals of his going to India was to transmit Buddhist tantras, most of

them in the Tattvasawgraha lineage. 3 The Chin kang ting ching ta .1!!.
ch'ieh .E.! me hsin t i fa men_! chueh4 also states that the Tattvasaill-

graha was composed in South India and that Vajrabodhi brought to China

an abbreviated version, i.e., this is presumably T. 18, No. 866.

Another text of the Tattvasawgraha lineage is said also to have been

taught at Dhanyaka~aka, an old name of Amaravati. 5 Kukai, quoting

from the LaAkavatara sutra, also claims the secret doctrine came from

South India. 6

On the basis of the above evidence it seems reasonably certain

that the Tattvasawgraha was composed in South India. For lack of

1shoun Toganoo, "Himitsu Bukkyoshi," p. 36; Ryujo Kanabayashi,

"Kongi5chi5 yuga churyakushutsu nenju kyo kaidai," Kokuyak~ Issaikyo
Mikkyobu, I (Tokyo: Daito shuppansha, 1937), pp. 213-15; Yukei Matsu-
naga, Mikkyo Kyoten Seiritsu, p. 196; Kyojun Shimizudani, "Dainichi-
Kongccho Ryobu Daikyo no Seisaku Chiho narabi sono Sakusha ko," in
Bukkyogaku Shomondai (Tokyo: Iwanami shoten, 1935), pp. 713-45; Cf.
Chou Yi-liang, "Tantrism in China," Harvard Journal of Asiatic Studies,
Vol. 8, 1944-45, pp. 272-307.

2Yukei Matsunaga, ~£no ?ojosha, pp. 169-79; T. 50, p. 713a.

3T. 52, No. 2120, p. 839ff. 4T. 39, No. 1798, p. 808a,b.
5shoun Toganoo, "Himitsu Bukkyoshi," p. 36.

6KCZ, Vol. I, p. 28, Vol. II, pp. 91-92.

 33

greater detail in the above records, however, it would be imprudent

to speculate further where it may have been composed in South India.

Although our knowledge of Buddhism in South India during the

period of the composiiton of the Tattvasawgraha is sketchy, it should

be pointed out that political conditions were not adverse towards

Buddhism in the seventh and eighth centuries A.D. The Pallava Kings

who had real hegemony over Southeast India from the mid sixth until

the late eighth centuries were known patronizers of the arts and men

of learning. Vajrabodhi himself was preceptor to the Pallava King

Narasi~havarman Rajasiwha II of the late seventh and early eighth

centuries whose capital at Kanchi (near present day Madras) was known

earlier as a center of Buddhism from Hsuan-tsang's travelogue. 1 HsUan-

tsang, traveling in India from 629 to 645 A.D., recorded that in Maha-

kosala (the territory later ruled by the Pallava Kings) Mahayana Bud-

dhism was flourishing. 2 The Bhaumakara Dynasty which ruled in present

day Orissa in the eighth century is known to have patronized Buddhism.3

Speculation has long been made also on the affiliation of the Tattva-
, ...
sa,graha lineage with the Tantric tradition of the eighth century Sr1-

vijaya Kingdom in the Malay archipelego and the mid-eighth century monu-

ment Barabu~ur in Java. 4 The ~ailendra rulers who established their

1R. C. Majumdar, ed., The History and Culture of the Indian Peo-
ple The Classical Age (Bombay: Bharatiya Vidya Bhavan, 1962), pp. 621,
258, 261. T. 51, No. 2087, ChUan #10.
2R. D Banerjee, History ,,f Orissa from the Earliest Times to the
British Period, Vol. I (Delhi: Bharatiya Publishing House, 1980), p. 142.

3Biswarup Das, The Bhaumakaras Buddhist Kings of Orissa and Their

Times (Ne~·r Delhi: Oriental Publishers aml Distributors, 1978), pp. 144-
145.
4Benjamin Rowland, The Art and Architecture of Ifidia Buddhist/Hin-
du/Jain, The Pelican History_ of Art (New York: Penguin Books~ 1977)-,-
 34

empire in the Malay peninsula, Sumatra and Java in the mid-eighth

century seem to have migrated from Kalinga in Southeast India and

were connected with the ~ailodbhavas who ruled Oris~a before the

Karas from the late sixth until the mid-eighth century. Inscriptions
,
of both the Sailodbhavas and Pallavas have been discovered in the Ma-

lay peninsula. 1

It is known that a strong Bhakti movement did develop in South

India from the sixth century until the end of the eighth century.

Even so, the strength of Buddhism in the south is apparent because

the Bhakti literature from this period 11 is replete with the sense of a

great danger to the orthodox paths from the spread of the heretical

creeds of Buddhism and Jainism. 112

Because of Jnanamitra's reference above to the appearance of the

Tattvasawgraha in the reign of King Indrabhuti who ruled in 09~iyana

in extreme northwest India Toganoo early speculated that this text

soon found its way from south India to that region. 3 In this regard,

because Vajradhatu map~alas are based on the Tattvasawgraha, reference

should be made to those known to exist in Tibet and elsewhere. Toganoo4

pp. 449-460; Alex Wayman, 11 Reflections on the Theory of Barabu4ur as

a Matl~ala," in Luis 0. Gomez and Hiram W. Woodward, Jr., editors, Bara-
budur: History and Significance of ~ Buddhist Monument (Berkeley: Asian
Humanities Press, 1982), pp. 139-149.
1R. D. Banerjee, History of Orissa, p. 131; K. A. Nilakanta Sastri,
South Indian and Southeast Asia Studies in their History and Culture
(Mysore: Geetha Book House, 1978), p. 231.

2K. A. Nilakanta Sastri, Development of Religion in South India,

(Calcutta: Orient Longmans, 1963), p. 39.

3shoun Toganoo, 11 Kongi:ichtskyD no Gaisetsu," pp. 5-6.

4shoun Toganoo, Mandarc; r•.o Kenkvii (Koyasan: Mikkyo Bunka Kenkylijo,

1958), p. 201 and plate no--.76~
 35

detected a picture of a Vajradhatu lineage ma~9a1a in A. H. Francke's

Antiquities of Indian Tibet, I. 1 This map~ala is in the Alchi temple

in Tibet which, according to Tibetan chronicles, was built during the

life of the famous Rin chen bzan po (958-1055). A r~cent expedition

to this temple has confirmed Toganoo's find and over-all ten mandalas
••
of the Tattvasa,graha lineage were discovered in this temple. Differ-

ing from the Japanese Vajradhatu ma~9alas, these include four types
,.,. . ..c.. ~··

of Jojinne <f1X9fza ) or central assembly map9a1as, five types re-

sembling the Catur-mudra ma~9alas and one type resembling the Rishue

( !! ~ ~ ) or seventh assembly ma'P~ala in the nine assembly matt~alas

of Japan. 2 G. Tucci earlier mentioned some of these mandalas and
••
others in Tibet in this lineage.3 K. W. Lim has also described a

complex of bronzes at Nganjuk in Java which conform to the Vajradhatu

map~ala described by Anandagarbha in his Tattvaloka. 4 Yoritomo re-

cently discovered Vajradhatu lineage map9a1as similar to those shown

in the Gobushinkan and eighty-one deity Vajradhatu map~alas (see be-

low, p. 63 ) in Orissa. These are sculptured in stone and, like

the Nganjuk bronzes, are known as rittai (jt1~) or~ ma~9a1as,

1Archaeological Survey of India, Vol. 38, New Imperial Series,

Part I (Calcutta: Superintendent, Government printing, 1914), plate
No. 39 opposite p. 92.

2Yukei Matsunaga, Mikkyo Kyoten Seiritsu, pp. 176-77; Yukei

Matsunaga and Kei Kate, photographter, Mandara Nishi Chibetto no
Bukkyo Bijutsu (Ma~9ala Buddhist Art of West Tibet), Two vols.-zMai-
nichi shinbunsha, 1981), color photograph vol., pp. 016-042, 047-049;
commentary vol. printed in blue and yellow, pp. 043-045, 055, 057-
078, 097-098, 117-124.

3Indo-Tibetica, Vol. III, Part I, Roma, 1935, p. 39ff. and Vol.

IV, Part I, Roma, 1941, pp. 100, 110ff., 119 (charts opposite pg.
120 entitled Chapter 32 and 33).

4"studies in Later Buddhist Iconography, I, The Vajradhatu-ma~­

~ala of Nganjuk," Bydragen Tot de Taal Land-en Volkenkunde, Deel 120,
 36

representing the actions of the deities.1

The Eighteen Assembly Vajratekhara

Based on the Tu ~ t 'o lo ni mu2 and the Chin kang ting ching ~

ch'ieh shih E.! hue chih kuei 3 , both translated by Amoghavajra, the

Tattvasawgraha is described as being the first scripture of the 100,000

:loka Ching kang ting ~ ch'ieh (Vajra~ekhara-yoga), a title for a col-

lection of texts consisting of sermons delivered at eighteen different

assemblies. The Tattvasa~graha is said to contain four parts, these

corresponding to the Great (Maha) map~ala, Symbolic (Samaya) mar.~ala,

Law (Dharma) map9ala and Action (Karma) ma~9a1a, respectively. T. 18,

No. 869 (pp. 284-87) is the only text briefly describing the contents

of these eighteen assemblies. Kukai in his Kong~chDgyD kaidai (T. 61,

No. 2221, p. 1aff.) and Kyo£ kyo kaidai (T. 61, No. 2222, p. 6c)

accepts the existence of this eighteen assembly scripture in four parts

and copies verbatim Amoghavajra's description of it. A Tibetan source

also mentions an eighteen assembly scripture which appeared for the

first time during the reign of King Indrabhuti. 4

According to T. 18, No. 869, the sixth assembly of the Vajra-

:ekhara is the Naya-sutra (T. 8, No. 244; Jap. Rishukvo) while the

fifteenth assembly is the Guhyasamaja tantra (T. 18, No. 885). How-

ever, based on the description ofT. 18, No. 869, it is clear the pre-

sent day Naya-sutra and Guhyasamaja tantra have been expanded. The

('S-Gravenhage: Martinus NYboff, 1964), pp. (327)-341.

lMotohiro Yoritomo, "Indo ni Genzai suru Ryckaikei Mikkyo

Bijutsu," Ars Buddhica, No. 150, 1983, p. 139ff.

2T. 18, No. 903, p. 898c. 3T. 18, No. 869, p. 284c, 1.16ff.
4shoun Toganoo, "Himitsu Bukkyoshi," p. 36.
 37

description of T. No. 869 hints at earlier and shorter versions of

these scriptures. Texts corresponding to the second, third, fourth,

thirteenth and sixteenth assemblies have also been identified, 1 yet

their ~ontents again are not identical with the description of T. No.

869. 2

Notwithstanding the above references, an eighteen assembly text

has never been found. Because all available texts apparently belong-

ing to it have been expanded or altered, and based on the usage of the

term Vajra6ekhara-yoga in T. No. 869 and T. No. 903, it is clear the

latter referred collectively to a whole lineage of texts essentially

dealing with ritual meditation. 3 Based on this usage Donjaku lists

texts belonging to this lineage. 4 Rather than referring to a single

text, the eighteen assembly scripture more likely referred to a

single tradition which was incorporating a wide range of ritual medi-

tation manuals which were being written and/or expanded. That this

collection comprised specifically 100,000 ~lokas, however, is no longer

accepted. 5

IT. 18, Nos. 908 and 909; Tohoku Catalogue No. 480; T. 18, Nos.
883 and 887, respectively. Shinten Sakai, "Kongochoyuga Goma Giki no
Issetsu ni tsuite," Mikkyo Bunka, Vol. 19, 1952, pp. 1-12; Ibid.,
"Kongochokyo no Daisane ni tsuite," Mikkyo Kenkyu, Vol. 71, 1939, pp.
117-147; Ibid., "Kongocho Gozanze Kaigiki Ro-o-kyo Chu Kanjizai Bo-
satsu Shinshingon Issai Nyorai Renge Daimandarahin ni tsuite," Mikkyo
Bunka, Vol. 12, 1950, pp. 16-22; Ibid., "Kongochokyo no Daijusane ni
tsuite," Mikkyo Bunka, Vol. 32, 1956, pp. 34-41.

2yrrkei Matsunaga, Mikkyo Kyoten Seiritsu Shiron, p. 187ff.

3This is evident from the appendage of the term yoga to the titles
of the texts in T. 18, No. 869.
4T. 61, No. 2225, pp. 124c-125a. He identifies, T. 18, Nos. 866,
867, 873, 876; T. 20, Nos. 1122, 1125, 1145, etc.

5Yukei Matsunaga, Mikkyo no Rekishi, pp. 63-68. Ryuzui Nakai had

asserted a single text existed:"Kongochokyo no Seiritsu Nendai,"
 38
The first mention of a full version of the Vajra~ekhara com-

prising 100,000 :lokas and an abbreviated version is found in Chin

kang ting ching ta .Yli ch' ieh £!. me hsin ti fa ~.! chiieh. 1 Kaiun

of the ninth century first records that the shorter version consisted

of 4,000 ~lokas. 2 Again, the Tai tsung ••• chih chi 3 mentions the
• '9

100,000 ~loka text of the Vajra~ekhara and Vajrabodhi also mentions

a 100,000 ~loka text.4 Also, in a memorial Amoghavajra laments he

hasn't translated the 100,000 ~loka text due to his old age.S

Regardless of these numerous references the designation 100,000

,
sloka text is not believed to be an accurate count of the number of

llokas in the "eighteen assembly" text but, instead, as with the Maha-

vairocana sutra, is thought to symbolize the high value attached to

the scripture.6 This vier-~point is in line with Kukai' s insistence

that the full text of the Vajra~ekhara is the Dharmakaya's perpetual

Law Map~ala, i.e., constant teaching, whereas the transmitted text

comprises eighteen assemblies. 7

Shukyogaku Kiyo, 1931, pp. 176-183.

lT. 39, No. 1798, p. 808a-b.

2T. 51, No. 2001, p. 784b. Sego tells us after his translation
of the entire Sanskrit text that it consists of 4,000 6lokas; T. 18,
No. 882, p. 44Sb, 1.2. As T. 18, No. 869, p. 287b,c, says this sutra
in eighteen assemblies has 4,000 6lokas or 5,000 or 7,000 or 100,000,
this would lend credence to the fact that texts were being assembled
along with the Tattvasa~graha and referred ~c collectively as the
Vajra~ekhara-~.

3T. 52, No. 2120, p. 846b. 4T. 18, No. 866, p. 233c, 1.12.

Sshoun Toganoo, "Kongcchckyc Gaisetsu," p. 7. Cf. KCZ, Vol. 2, p.

72.
6Eihan Sakano, Kongochokyo ni Kansuru Kenkyu, p. 15; Munetada Su-
zuki, Himitsu Bukkyoshi, p. 116; Seigai Omura, Mikkyo Hattatsushi (Tc-
kyc: Kokusho kankokai, 1973), p. 486.

7Rishukyo kaidai, KCZ, Vol. 2, p. 294 or T. 61, No. 2236, p. 6lla,

 39
.J- . - 1 2
In this regard, Ryokai ( 1.t ~ , 1698-1755) says the legend

that Nagarjuna discovered the Mahavairocana and Tattvasawgraha sutras

in a small iron stupa in southern India means that he discovered the

source of his own mind at which time the Shingon teachings appeared.

The legendary full editions of both texts in 100,000 flokas he inter-

prets to mean the constant preaching of the Dharmakaya while the known

"abbreviated" versions are what is explained at one place and time.

1.22ff.; Cf. Kukai's Dainichikyo kaidai, KCZ, Vol. 2, p. 186; T. 61,

p. 123b, 1.13ff. and T. 39, No. 1798, p. 808a-b.
1
Jujrishinron Kanchu, Chizan Zensho, Vol. 7 (Tokyo: Chizan Zensho
Kankokai, 1967), p. 339bff.

2T. 39, No. 1798, p. 808a-b.

 40

Chapter II Synopsis of Part 1

David L. Snellgrove has written a general synopsis of the entire

Tattvasawgraha which the reader is referrec to. 1 The present synopsis

is limited to Part 1 translated for this thesis and includes references

to Kukai's analysis of its contents as found in his writings. The

topics introduced here will anticipate the account of the role of

the Tattvasawgraha lineage in the teachings of Kukai discussed in Part

2.

Preface

The preface begins in the conventional manner of Buddhist Mahayana

sutras with a description of the presiding Buddha, here identified as

Maha-krpo-Vairocana. The listener of this sutra is identified as

Vajrasattva by the connnentaries which in the Shingon tradition. is

another name for bodhicitta or the mind or thought of enlightenment.

Vairocana is described as consisting of the four knowledges and Buddhas

and he is seated in his palace in the Akani~fha heaven surrounded by

myriads of Bodhisattvas.

Formal Introduction

Now begins a description of Mahavairocana in the Vajradhatu ma~~a­

la. Mahavairocana is identified as a knowledge being (j~nasattva).

1snellgrove, Facsimile Reproduction, pp. 15-38. See pp. 26-29

for lists of the names of the deities in the various mandalas in the
sutra. ••
2MDJ, p. 686 •
 41
The initial verses are equated with the four paramitas immediately

surrounding Mahavairocana. Kukai calls these the four Buddha mothers!

because they are Bodhisattvas who nurture all Buddhas and raise the

Tathagatas of the three times • He also says they arise from the

Dharmadhatu knowledge. There follows a description of the sixteen

Maha-bodhisattvas of the Law ma~~ala surrounding the four Buddhas

Akjobhya (East), Ratnasawbhava (South), Lokelvararaja (West) and

Amoghasiddhi (North). Four bodhisattvas surround each of the four

Buddhas in the four cardinal directions.

From the viewpoint of traditional Shingon the0ry these four para-

mitas are offered to Mahavairocana by the four Buddhas. 2 Prominent

among texts of the Tattvasawgraha lineage which formulate this position

I
is the Bodhicitta Sast~allegedly composed by the tantric master

Nagarjuna. The Bodhicitta Sastra is found in the Jikkanjo 3 along with

six other works by Kukai. Since the Tokugawa period (1600-1867) this

compilation has been required reading for all Shingon monks. The

following quotation shows that the paramitas emerge from the Buddhas.
"The moon disc has sixteen divisions. These are
symbolized by the sixteen }iaha-bodhisattvas in the Yuga
(i.e., the Tattvasawgraha sutra) from Vajrasattva down
to Vajrasandhi. Among the thirty-seven deities the
Buddha ranks in the five directions each express one
knowledge. Akiobhya Buddha in t.he eastern quarter, by
perfecting the great, perfect mirror-like knowledge, is
also called thunderbolt knowledge. Ratnasa~bhava Buddha
in the southern quarter, by perfecting the equality
knowledge, is also called consecration knowledge.
Amitabha Buddha in the western quarter, by perfecting
the discriminative knowledge, is also called lotus
knowledge and also, knowledge that turns the wheel of
the teaching. Amoghasiddhi Buddha in the northern

1KCZ, Vol. 2, pp. 431, 614. 2

M. Kiyota, Shingon Buddhism, p. 96.
3(Sth ed., revised; Koyasan University Press, 1978), pp. 205-19.
The earliest edition is dated 1732. T. 32, No. 1665. MJ, p. 307. Note
in the quotation (and the Shingon school) Amitabha is used interchange-
ably with Loke~vararaja.
 42

quarter, by perfecting procedure-of-duty knowledge, is

also called action knowledge. Vairocana Buddha in the
central quarter, by perfecting the DharmadhDtu knowledge,
is the foundation. From the above four Buddha knowledges
emerge the four paramita bodhisattvas. These four bodhi-
sattvas are Vajra, Ratna, Dharma and Karma. These are the
mothers which raise and nurture all noble ones in the
three times."1

The sixteen Maha-bodhisattvas are the iconographic representation

of the practice of the four knowledges. According to Kukai, the

samadhis of each are to be cultivated. 2 Cor:tinuing with the explana-

tion o f the Bodhicitta Sastra,

'- these sixteen are as ind icated.

" ••• Now, from the intrinsic nature of the Dharmadhatu which
has been sealed, there arise the four Buddhas. Each of the
Tathagatas in the four quarters is united with four Bodhi-
sattvas. Ak~obhya Buddha in the eastern quarter is united
with four Bodhisattvas: 1) Vajrasattva; 2) Vajraraja; 3)
Vajraraga and 4) Vajrasadhu ••• Ratnasa~bhava Buddha in the
southern quarter is united with four Bodhisattvas: 5) Vajra-
ratna; 6) Vajrateja; 7) Vajraketu; 8) and Vajrah~sa •••
Amitabha Buddha in the western quarter is united with four
Bodhisattvas: 9) Vajradharma; 10) Vajratik~~a; 11) Vajrahetu
and 12) Vajrabhasa ••• Amoghasiddhi Buddha in the northern
quarter is united with four Bodhisattvas: 13) Vajrakarma;
14) Vajrarak~a; 15) Vajrayakfia and 16) Vajrasandhi)j3

After the above sixteen Maha-bodhisattvas of the Law ma~~ala are de-

scribed the self-nature of Mahavairocana is described as the sixteen

Maha-bodhisattvas of the Great ma~~ala and those of the Symbolic

ma~~ala. The introduction closes with an often cryptic description

of Mahavairocana as the self-nature of the Dharmadhatu.

Kukai says the bodhicitta is the pure Dharmakaya of all Buddhas

as well as the impure mind of living beings. 4 This Dharmakaya is re-

presented in the Vajradh~tu mav~ala. Identifying this Dharmakaya with

--------·-··-----
1JikkanjD, p. 213; T. 32, No. 1665, p. 573c, 1.2ff.
2Hakeda, Kukai: Major Works, p. 226. 3T. 32, p. 574a, 1.4ff.

4Kcz, Vol. 2, p. 149.

 43
Vairocana, Kukai says it consists of five knowledges or the thirty-

seven deities.1 The four map~alas 2 , representing respectively

Vairocana's physical extension (Great), his unity with all living

beings (Symbolic), the constant teaching of Vairocana (Law) and his

action (Karma), are all different aspects of the Dharmakaya. By the

practice of the Vajradhatu recitation it is taught in the Shingon

tradition that the devotee awakens the latent bodhicitta having these

aspects. That is, by the conscious attempt to act in harmony with

this dormant bodhicitta through the practice of the three mysteries3

and the Vajradhatu recitation enlightenment is experienced.

Because the deities in the Vajradhatu ma~~ala arise from Vairo-

cana, even meditation on just one of these deities can lead to en-

lightenment because any deity is ultimately an extension of Vairocana.4

Chapter 1 Description of the Rite of the Vajradh~tu Mah~­

Mandala

Section 1
1) The Five Stages of Realization (pa~cabhisawbodhi).

Now that the character of Mahavairocana has baen described in

the introduction, Chapter 1 begins with a discussion of how to

attain the enlightenment represented by Mahav~irocana. This is the

practice of the five stages of realization.5

The leading Bodhisattva Samantabhadra defined as the mind of en-

lightenment resides in the hearts of all Tathagatas in the Akani~)ha

heaven. The latter approach the Bodhisattva Sarvarthasiddhi with

Safbhoga type bodies to teach the fjve stages. Having achieved en-

1Kcz, Vol. 2, pp. 98, 119, 540. 2rbid., pp. 438, 439, 622.

3rbid., PP· 632-33, 656. 4Hakeda, Kukai: Major Works, p. 99.

5This practice in an altered form is an integral part of the
Vajradh~tu recitation system transmitted by Kukai.
 44

lightenment through the five stages of realization the Bodhisattva

Sarv~rthasiddhi is consecrated with the name Vajradh~tu. His reali-

zation is then described in terms of the four transformed categories

of perception (vij~ana). 1 Having been empowered by all the Tathaga-

tas the Vajradhatu Tathagata then proceeds to Mount Sumeru where as

a four-faced Vairocana he reveals the T~ttvasawgraha.

Section 2 Samadhi of Vairocana Tathagata

1) Manifestation of the sixteen Maha-bodhisattvas

After the samadhi of Vairocana is des~ribed as consisting of the

five Tathagatas and knowledges,the epiphany of the sixteen Vajra-

sattvas, etc. of the Vajradhatu Maha-ma~9ala begins. Vairocana enters

a particular samadhi based on the pledge of a Maha-bodhisattva and then

issues from his heart the deity in question as a mantra. These mantra

names as given above for the sixteen Vajrasattvas are regularly used

in the Shingon school to identify these deities. All these deities

represent different aspects of the enlightenment of Vairocana.

The pattern of the manifestation of all sixteen Vajrasattvas is

the same. First Vairocana enters a samadhi which has its origin in

the pledge of a certain Maha-bodhisattva. Onc.e in samadhi Vairocana

issues from his heart a mantra described as the heart of all Tathaga-

tas. Then either Bhagavat Vajrapani, Vajradhara or Samantabhadra issue

•
from the hearts of all Tathagatas and further enter the heart of Vairo-

cana. Then the Bhagavat (who is really the manifested form of the

mind of enlightenment) issues the Vajrasattva's symbolic object (a

vajra, hook, etc.) from the hearts of all Tathagatas and places it in

Vairocana's hand. From the object there are manifested countless Ta-

thagatas who engage in supernatural activities. These Tathagatas

lHakeda, Kukai: Major Works, pp. 84-85.

 45

coalesce into the above Maha-bodhisattva whose pledge is the basis

for the samadhi. This Vajrasattva then utters a verse in praise

of his nature. Immediately afterwords this Yajrasattva leaves the

heart of Vairocana, takes his place in the mav9ala and requests

instruction. Vairocana then enters a samadhi for the sake of

teaching the Vajrasattva a specific function which is an aspect of

Vairocana's enlightenment and the Vajrasattvds intrinsic nature.

/airocana then consecrates him and confers an initiate\ name.

1

It should be remembered that all deities manifested in the

map~ala are conceived of as manifestations of either Samantabhadra,

Vajradhara or Vajrapa~i, the latter predominating in Chapter 1 •

This is one way the sutra has of reminding the reader that, despite

the various deities manifested, they all have their origin in the

mind of enlightenment (bodhicitta) which is often indicated by the

above names.1

2) Manifestation of the four paramitas

In this section four paramitas of feminine gender are created

by the four principle Tathagatas to make all Bodhisattvas successful

in their implementation of the knowledges of the Tathagatas. Ak~obhya

creates Vajra paramita, Ratnasa~bhava creates Ratna paramita, Lake~-

vararaja creates Dharma paramita and Amoghasiddhi creates Karma para-

mita. These take their place around Vairocana in the very ce~ter

1MJ, pp. 234, 236, 239, 593. In his Hokekyoshaku Kukai refers
to the Kongochokyo (T. 18, No. 866, p. 288a) which suggests that
"Samantabhadra Bodhisattva is the enlightened mind of all Tathagatas.
Also, Samantabhadra is the absolute Dharmakaya." KCZ, Vol. 2, p. 389.
In his Kongo hannya haramitsu kyo _kaidai Kukai also quotes the same
sutra (T. 18, p. 250b) which says "Samantabhadra Dharmakaya pervades
everywhere and he is the Lord of this world's beings. He is without
beginning or end and is neither born nor dies." KCZ, Vol. 2, p. 433.
 46

of the mat;~ala.

The pattern of the mantfestation of all four paramitas is ideo-

tical. Each Tathagata enters a sam~dhi which originates in the pledge

of the paramita under discussion and issues a mantra from his heart.

There immediately issue objects symbolic of the paramitas. From these

the Bhagavat Vajradhara creates countless Tathagatas. He then seals

the paramita and creates another object symbolic of the deity which

becomes the single body of the universe. The paramita then takes

,
her place in the mandala .
and finally utters a verse descriptive of her

nature.

3) Manifestation of the Eight Offering (Poja) Deities

Four inner pujas: Vairocana now creates four pujas who are

offered to the four Tathagatas.

Four outer pnj~s: In order to return the honour of Vairocana's

offerings the four Tathagatas present four outer offerings who due

honour to Vairocana. These pujas are called inner and outer because

the former are within the large central circle while the latter are

in the four corners of the second or middle square enclosure of the

mandala.
•. Each is conceived of as an aspect of Vairocana's enlighten-

ment as described in the commentaries.

Kukai says Vairocana emanates light everywhere and that inwardly

he changes into the four inner pujas who externally are offered to

the four directions. These four inner pujas are known as the four

realized knowledges. The four outer pnj~s are changed forms of the

four Tathagatas which attend to Vairocana in the center of the ma~~ar

la. 1

lKCZ, Vol. 2, pp. 588-89.

 47

4) Manifestation of the four Attracting (Sawgraha) Deities

This last group of deities of the thirty-seven in the ma~9ala are

the four guardians of the portals. Their names are Vajr~nku~a, Vajra-

pa~a, Vajrasphofa and Vajrave~a. Vairocana creates these four guard-

ians who have, among other functions, those of attracting, drawing in,

tying and delighting all Buddhas in the mandala. They also protect
••
the various virtues of the enlightened mind.

Kokai says Vajranku~a is the Tathagata knowledge which helps

the practitioner enter samadhi when the thought of enlightenment first

arises. Vajrapa~a binds all living beings and draws them to the seat

of enlightenment so they do not depart. VajrasphoFa is the Tathagata

knowledge which enables the practitioner to always have good thoughts,

which surp=esses evil thoughts and enables Bodhisattvas always to be

resolute in the sea of sa~s~ra. Vajravesa signifies insight(praj~a).

He makes living beings happy by removing ignorance and passion and is

the knowledge of joy of all the Tathagatas. 1

Section 3 Samadhi of the Supreme Action Lord

In this section Vairocana Tathagata summons all Tathagatas by a

snap of his fingers. Both Ennin and Donjaku say all are assembled in

order to be taught the rites of the ma~9a1a which lead to Buddhahood.

heart of Vairocana in the

.
They pay homage to the mandala with all its deities and then enter the
'
cen~er of the mandala. Each Tathagata in
••
turn manifests a ma~9a1a consisting of hosts of Bodhisattvas. Then

they utter a verse proclaiming the common origin of all Buddhas from

the mandala •
••

lKCZ, Vol. 2, pp. 591-92.

 Section 4 Method of Consecration 48
1) 108 Invocations

For the empowerment nf the Vajradhatu

- ..
mandala the great Vajra-

dhara is acclaimed with the names of the sixteen great Vajrasattvas

extended to 108 variations. That is, each of the sixteen is invoked

with their alternative names. It should be noted that similar invo-

cations occur at the beginning of Parts 2, 3 and 4 of the Tattvasai-

graha.

2) Vajradhatu ma~~ala: the central assembly

Now the way to make a Vairadhatu Maha-ma~~ala is explained in outline

by Vajradhara. This is usually an oral tradition and is not committed

to writing in de~ail. All the main deities of the mandala have been
••
presented above.

3) Preparations of the Master (acarya) for entering

disciples into the map~ala.

This section explains what both the acarya and disciple to be

initiated should do when entering the map~ala. The mudras to be

formed, the mantras to be recited and the visualizations to be made

are outlined.

4) Way of Entering a Devotee into the Mandala

••
Part 4 describes the way to perform a Kechien consecration accord-

ing to Donjaku and Horiuchi, although Ennin says the acarya consecra-

tion is described. The former is permitted those of all walks of life

as long as they seek to attain enlightenment and help others to do the

same. First one prostrates to the four Tathagatas. Then a blindfold

is tied on the devotee who enters the mavpala. An oath of secrecy is

administered and the consecrations are bestowed. The ac§rya then em-

powers the disciple, the latter tosses a flower onto the mandala and
••
 49
after having the blindfold removed, is given the secret flask

consecration.

Section 5
1) Now the various sets of powers attained by means of

the consecrations are explained. Each power is sealed by a mantra.

2) This part closes this section with a description of

the oath of secrecy required after the initiation, the effects of

the four thunderbolt mudras and the general way of achieving union

with Vairocana.

In his work entitled Sokushin jobutsugi or Principle of Attain-

ing Buddhahood with this Very Body, Kukai comments on T. 20, No.

1125, in reference to initiation into the Vajradhatu ma~~ala. This

important and pertinent analysis will be quoted in full as translated

by Hisao Inagaki. 1 Kukai's comments are in brackets.

"Again it is said, 'If (a practitioner) avails himself

of the teaching arising out of the inwardly realized wisdom
of self-enlightenment expounded by the Self-Enjoyment Body
of Mahavairocana Buddha and also avails himself of the wis-
dom of the Enjoyment Body for Others' Sake of Vajrasattva
in the state of the great Samantabhadra, he will meet a
Mav~ala Ac~rya and be able to enter the Mappala. That is
to say, he will acquire the Karma (for abiding by the pre-
cepts) and, as (the Acarya) conjures up Vajrasattva in
Samantabhadra Samadhi, Vajrasattva will enter his body.
Owing to the divine power of empowerment, he will instantly
attain immeasurable Samayas and Dhara~I-gates. (The Ac~rya)
transforms with the wonderful Dharma his disciple's seeds
of innate self-attachment. The disciple will immediately
acquire in his body the merit and wisdom to be accumulated
during the period of one great asa~khya kalpa, whereat he
will be considered to have been born into the Buddha's
family. He has been born from the mind of all the Tatha-
gatas, from the Buddha's mouth, from the Buddhas' Dharma,
and from the teaching of Dharma, and has acquired the trea-
sure of Dharma. The treasure of Dharma refers to the teach-
ing of (awakening) Bodhi-Mind through the Three Mystic
Practices.' (This shows the benefit which a practitioner

1Kukai's Principle of Attaining Buddhahood with the Present

Body (Kyoto: Ryukoku University Translation Center, 1975), pp. 26-7.