18,15), and sittram mdtrkd ca,
equivalent
to dgamacatustayam (p.
Abhidharmakosabhasyam-Vol-1-Vasubandhu-Poussin-Pruden-1991
.
.
"The four srdmanyaphalas are called pratisamkhydnirodha.
"
"What is srotadpannaphala? If the three klesas to be cut off by Seeing are cut off; if satkayadrsti, vicikitsd and silavrata are exhausted, this is called srotadpanna-
12
phala" Sariputra takes up the question again, "What is srotadpannaphala} The
three klesas to be cut off by Seeing being cut off, satkdyadrsti-vicikitsd-silavrata being exhausted: if one obtains amrta, this is what is called srotadpannaphala"
It appears that what we are encountering here is a terminology alien to the Abhidharma and to the Sarvastivada.
73 viii. The Abhidharmasdra.
1. Before Vasubandhu, many masters undertook to summarize the doctrines of the Abhidharma. We possess notably three works: 1. The Abhidharmasdra of
74
Dharmas'rl in ten chapters, made up of kdrikds (probably in dryan stroph)
and a
commentary; 2. a second edition of this same Sara by Upasanta, to which the
Chinese give the name of Abhidharmasdra-ching [=sutra]: the same kdrikds with
a more developed commentary; and 3. a third edition of the Sara, the Tsa
[=Miscellaneous] Abhidharma-ching, by Dharmatrata, which is in fact a new
75 work, containing a new chapter and many new kdrikds.
76
2. The preface to the Vibhdsd (Taisho 1546)
Dharmasri before the Jndnaprasthdna: "After the nirodha of the Buddha, the bhiksu Dharmas'ri composed the four volumes of the Abhidharmasdra. Then Katyayanlputra composed the Abhidharma in eight books . . . "
3. The work of Dharmasri contains ten chapters: Dhdtu, Samskara, Anusaya, Arya, Jndna, Samddhi, Sutra, Tsa and Sdstravarga or Vddavarga.
Between the ninth and the tenth chapters of Dharmasri, Dharmatrata places a new chapter, the Pravicayavarga, which indeed appears to constitute an independent work.
There is a stanza of introduction: "even though many dharmas have been spoken of, their meaning remains confused . . . " and four concluding stanzas: "The author has composed this book based on the book of Dharmasri, not through pride or in order to acquire a reputation . . . ".
It begins with the dharmacakra, the Wheel of the Dharma: "The Muni said that the darsanamdrga is called dharmacakra, either because it goes into the mind
by Tao-yen places the work of
? of others . . . (Kosa, vi. 54).
There then comes the brahmacakra (vi. 54, vii. 31), the updsaka (iv. 69 ), the
four parts of sila (iv. 29), the prdtimoksa . . . Later (p. 959b), cosmology: the periods of loss, etc. (iii. 99), destruction by fire, etc. (iii. 102); and then there follows the theory of the three "fallings away" (p. 960c; Kosa, vi. 59) and the definition of the Bodhisattva (iv. 108).
Suddenly (p. 96lc): "How many types of Sarvastivada are there? " Presenta-
tion of the four doctrines (Kosa, v. 25-26) without mentioning the name of the
four masters. The second and the fourth are bad because they confuse the time
periods. The first (difference in bhdva, translated fen): "One should know that
77 this is the parindma-sarvastivada. "
There is a diversity of opinion as to whether the Truths are seen at the same time (Kosa, vi. 27), Sarvastivadins and Vatslputrlyas on the one hand, Dharma- gupta on the other; antardbhava (iii. 34); then the Sarvastivadin proof. And at the end of the paragraph, the discussion "whether the Buddha is part of the Samgha. " Finally, the concluding stanzas.
4. The Samskdravarga treats of the simultaneous arising of the citta-caittas and of atoms (Kosa, ii. 22), of the four laksanas of "conditioned things" (ii. 45), of the hetus and the pratyayas (ii. 48, 61).
The Sutravarga is a collection of notes on the three Dhatus and a calculation of the places that they contain: sixteen in Rupadhatu, but, according to some, seventeen (Kosa, iii. 2): the sattvdvdsas (iii. 6), the vijndnasthitis (iii. 5-6); the three vartman of pratityasamutpdda (iii. 20, 25), the twelve limbs; the mahdbhutas, the Truths, the fruits of the Aryans, etc.
The Tsa-varga defines the mind-mental states as samprayukta, sdsraya, etc. (ii. 34); it enumerates the viprayuktas: dsamjnika, two non-conscious absorptions, sabhdgatd, ndmakdyddayas, jivitendriya, dharmaprdpti, prthagjanatva, four laksanas (compare ii. 35-36); it concludes with half a karika on the four bhavas (in. 13) and a karika on "disgust" amd "detachment" (vi. 79).
The Sdstravaraga (or Vddavarga) is made up of ten questions in verse, followed by answers in prose, relative to samvara (iv. 13), to the results, etc. Dharmatrata adds sixteen questions.
5. In order to appreciate the character of the treatises of DharmaSrT, Upasanta, and Dharmatrata, and Vasubandhu's debt with respect to Dharmatrata, which appears to be notable, we may see how two dharmaparydyas, the chapter of the three obstacles (dvaranas) and that of avijnapti, are treated by the different masters.
a. Obstacles, Kola, iv. 95-102, Vibhdsd, p. 599.
Poussin 45
? 46 Introduction
Dharmasri, p. 815: "The Bhagavat says that there are three avaranas: karman, kiefa, and vipdka. What is their definition?
"Anantarya actions which are without remedy, developed defilements, bad aaion experienced in the painful realms of rebirth, are the dvaranas.
"These three form an obstacle to the Dharma; they hinder the grasping of the Aryan dharmas\ they are thus called 'obstacles/ Which is the worst aaion?
"The action which divides the Sarhgha is said to be the worst.
"This action is the worst. One guilty of this remains a kalpa in Avici hell. Which is the best aaion?
"The cetand or 'volition' of Bhavagra is the greatest. u
Naivasamjndndsamjndyatana is Bhavagra. The volition which belongs to the realms of this sphere is the greatest and finest: its result is a life of some 80,000 kalpas in length. "
Upasanta, p. 843b-c, has the same two stanzas, but a less meager commentary:
"That which hinders the Path of the Aryans and the means (updya) to this path is said to be an obstacle. The obstacles to aaion are the five dnantaryas, namely, the killing of one's father, etc . . . He who commits such an aaion is immediately and necessarily reborn in Avici: thus the aaion is dnatarya. The killing of one's father and mother destroys goodness, hence it is Avici hell. Those guilty of the other three are led to injure a field of merit. The obstacles of the defilements are 'agitated' and 'sharp' defilements: the first is habitual defilement; the second is the overriding defilement. This refers to the 'present' defilements, not to the defilements that one 'possesses' (that one has as potential), for all beings 'possess' all the defilements . . . "
There is a variant to the second stanza: "Lying which divides the Samgha . . . ; volition in Bhavagra, among good aaions, has the greatest result," which is better. The commentary notes the differences in the two schisms (cakrabheda, karmabheda).
Dharmatrata is longer (p. 898b-899c) and very close to Vasubandhu:
According to the first stanza of Dharmasri: 1. the dvarana of klesa is the worst; the dvarana of aaion, mediocre; and the dvarana of retribution, the least; 2. the division of the Samgha, by nature, is non-concord; this is a viprayukta samskdra of the anivrta-avydkrta class; 3. the division is a thing of the Samgha; the transgression is of him who divides the Samgha; he experiences, in Avici, a retribution of kalpa\ 4. the bhiksus are divided in their opinion of who is the Master, of what is the Path: this is the division of the Samgha which was united, and he who breaks it is 'one who possesses views' {drspicarita)\ 5. in three continents, a minimum of eight persons is required for karmabheda\ in
? JambudvTpa, a minimum of nine persons is required for cakrabheda; 6. cakrabheda is impossible in six time periods: when a boundary is not delimited; at first; following; when the Muni has passed into Nirvana; when the running sore has not been produced; when the pair of chief disciples has not been established (six pddas); 7. lying which divides the Sarhgha is the worst of aaions; the volition of Bhavagra is said to bear the greatest result.
b. Avijnapti, Kosa, iv. 2, 3.
DharmasrT (p, 812c):
"Bodily action is vijnapti and avijnapti. The vijnapti of the body is the movement of the body, good, bad, or neutral: good when it arises from a good mind . . . For avijnapti: when one does an action in a firm manner, the mind can change, but the seed remains. If, for example, a person undertakes the precepts, his mind can then be bad or neutral: nevertheless the precepts continue . . . Action of the manas is solely avijnapti. . . because this action is not visible . . . vijnapti is good, bad, or neutral; the same for the avijnapti which belongs to the manas. The other vijnaptis are never neutral. "
Upasanta (p. 840) adds a bit. The hunter is regarded as free from bodily avijnapti. Mental action is called avijnapti because it does not inform others. Some say that it is called vijnapti because it is discourse (jalpa? ).
Dharmatrata (p. 888b) replaces the terms vijnapti and avijnapti with "doing" and "not doing" (karana, akarana) (Kos'ajv. 14):
"Bodily action is of two types: karanasvabhava, or akaranasvabhava. 'Doing'
78
(karana): movement of the body, exercise
'Not doing' (akarana, wu-tso): when the movement of the body has ended, the nature (good or bad) of this movement, of this action which is the movement, continues to arise, simultaneous with minds of a different nature, even as the good precepts produced by their undertaking (kusalasamddanasila) continue to arise even when bad or neutral minds are present. Like the immoral person (dauhsilya- purusa): even when good or neutral minds are present, his immorality continues to arise. "
". . . Action of the manas is cetand, volition, by nature . . .
"'Not doing' (akarana) is also called nirati (? Vyut. 21, 114), virati; upeksa, akriyd (pu-tso). Because it does not do, it is called 'not doing. ' If one says that this is not an action (karman), this is wrong, because it does. Good does not do evil, evil does not do good: this is also an action. As the upeksa part of Bodhi is not upeksa by reason of what is called upeksa; but the practice of the Path, the arresting of things, is called upeksa. The same here. Furthermore, in doing the
of the body is the 'doing of the body. '
Poussin 47
? 48 Introduction
cause one does the result: . . . 'not doing' is not rupa, but the doing of it (which is the cause of the 'not doing' or avijnapti) is rupa\ 'not doing' is thus called riipa. In
79 this same way then, 'not doing' is action.
? 1. Originally printed as a Foreword to de La Vallee Poussin's Cosmologie bouddhique: 1913, and published 1919 in the four-part Memoires of l'Acadmie royale de Belgique (Luzac, London). This contained the restoration of the kdrikds of the third chapter of the Ko/a, the Tibetan kdrikds, the Bhdsya, and the text of the Vydkhyd\ in the appendix, a summary of the Lokaprajnapti and the Karanaprajnapti.
2. Cullavagga, xi. 1. 8. [For a more exact presentation,J. Przyluski, Concile de Rdjagrha, p. 311,345, 349].
3. Oldenberg, Buddha . . . , 6th edition, p. 202; Fr. trans. Foucher, 2nd edition, p. 177. [Psychology, yes; but ontology is doubtful].
4. In the Divya, a Sarvastivadin work, where we encounter the expressions sutrasya vinayasya mdt- rkdydh: "The monks ask with respect to the Sutra, the Vinaya, and the Mdtrkd" (p.
18,15), and sittram mdtrkd ca, equivalent to dgamacatustayam (p. 333,7), Kern {Manual, p. 3) thinks that the term mdtrkd is employed "as synonymous with abhidharma. " It cannot in any case designate the Abhidharmas of the Sarvastivadins of which we are speaking below (p. 3) which are treatises; it fits the Abhidhammas a little less poorly, but without being satisfying. Does it designate some lists "omitting all the explanations and other details" (Childers, 243), lists of items which form part of the dgama and which are not specifically Abhidharma! The Sautrantikas, who deny the existence of an Abhidharma Pitaka distinct from the Sutra, certainly had such an "index," exactly like the Sarvastivadins of that period, and earlier than the Abhidharmas to which the expressions
of the Divya refer. Does it designate some presentations, in the manner of the sutras, like those that constituted the mdtikds of the Vibhanga? , In this book, which is the property of the Abhidharma, it is often a type of commentary in the form of glosses.
5. See the article of Rhys Davids in Hastings' Encyclopedia of Religion and Ethics.
6. Developpement de VAbhidharma, Dogme et philosophies 122; J. Przyluski, Concile, third
chapter and 179, 353; Acoka, 322; ? un4railles, 49; Levi, Seize Arhats, 20, 39.
7. In Hastings' Encyclopedia, 1. 19-20; Winternitz, Geschichte, 134. Scholastic definitions of the
Abhidharma: AttasMini, 48-50 and following; Satrdlamkdra, XI. 3.
8. It knows, however, that the author of the Kathdvatthu foresaw and refuted in advance the heresies to come; see Atthakathd, pp. 6-7. The remark is by Minayeff, and the observations by H. Oldenberg (Buddh. Studien, p. 633, 676) do not demonstrate that the Kathdvatthu has not been amplified in the course of time.
9. This is incorrect. The Samgitiparydya is only the Samgitisuttanta. The second part of the Dhdtukdya has a close relationship with the Dhdtukathdprakarana. A careful study will show other points of contaa, and one can see that the Sarvastivadins simply enriched by their inventions (theory of the viprayuktas, of the mahdbhdmikas, etc. ) the earlier material of the Abhidhamma.
10. The account of Buddhaghosa, Kathdvatthu-Atthakathd, p. 6, holds that because of this, at least the Vibhajjavadins are the orthodox party.
11. This is one of the aspects of the problem of kiriyavdda.
12. This definition of the two schools is borrowed from the Abhidharmako/a, v. 9; see Kathdvatthu, 1. 8 (which does not entirely confirm our interpretation). The controversy of time and the pudgala in the Vijndnakdya, Etudes Asiatiques, 1925.
13. Geography of the Sarvastivadin sect, J. Przyluski, Apoka. I know that Sinologists, notably Takakusu, are not settled on the language in which the first of the Abhidharmas of the
Poussin 49
? 50 Introduction
Sarvastivadins, the Jndnaprathdna, was written: "In what language, however, the original text was composed we have no means of ascertaining. All we can say is that the text brought by Sarhghadeva and Dhammapiya [Dharmapriya] from Kacmira [383 A. D. ] seems to have been in
a dialect akin to Pali, whereas the text used by Hiuen-tsang [657 A. D. ], as in other cases, seems to have been in Sanskrit. But this supposition rests solely on the phonetic value of Qiinese ideographs employed in these translations, and is not corroborated by any other evidence . . . It seems to me more than probable that the JHdnaprasthdna at least was written in some dialect: one thinks naturally of the dialect of Kacmira, but we really have no certainty that the Jndnaprasthdna was not composed in Kosala (JPTS; 1905, p. 84,86). "
We possess a fragmentary quotation from the Sarvastivadin Pratimoksa which proves that some earlier forms, Paji or dialect, remained in use: "When, in the Posadha ceremony, the Vinayadhara asks, 'Are you pure? ' (bhiksuposadhe hi kacci ttha parisuddhd iti vinayadharend- nusrdvite), if any bhiksu does not confess his transgression . . . " (Abhidharmakofavydkhydad iv. 72; compare the introduction of the Pdtimokkha and the remarks of Rhys Davids, Dialogues, II, p. 257). (See L Finot, "Pratimoksasutra des Sarvastivadins," JA, 1913,2. 177-9).
But we possess a fragment of the Jndnaprasthdna, quoted in the Abhidharmakosavydkhyd (ad i. 49): katamad buddhavacanam tathdgatasya yd vdg vacanam vyavahdro gir niruktir vdkpatho vdgghoso vdkkarma vdgvijnaptih // buddhavacanam kusalam vaktavyam athdvydkrtam vak- tavyam / sydt kusalam sydd avydkrtam // katarat kusalam / kusalacittasya tathdgatasya vdcarh bhdsamdnasya yd [vdgjvijnaptih // katarad avydkrtam / . . . purvavat // punas tatraivdnantaram uktam / buddhavacanam ka esa dharmah / ndmakdyapadakdyavyanjanakdydndm ydnupur- vavacand anupurvasthdpandanupurvasamayoga iti /
14. Invention of the prdptis, of sabhdgatd, of the existence of the past and the future, of diverse types of cause, of apratisamkhyanirodha, not to mention the nirvedhabhagiyas, etc.
15. The Abhidharmako/avydkhyd speaks of satpdddbhidharma-mdtrapdthina Abhidharmikas, Abhidharmikas "who read only the six-legged Abhidharma" which we understand to mean "who do not read the Vibhdsd. " These are Sarvastivadins; but all Sarvastivadins are not "followers of the Vibhdsa' (Vaibhasikas). We know, for example, that there were four ways of understanding "all exists," those of the Sarvastivadins Dharmatrata, Ghosaka, Vasumitra, and Buddhadeva: the Vaibhasikas of Kas'mTr condemn the first, the second and the fourth; and the first for the serious reason that it is confused with the non-Buddhist teaching of the Saihkhyas.
16. Better, "after the reign of Kaniska," Inde sous les Mauryas . . . , p. 328.
17. See this Introduction, Darstantikas, and Index, Sautrantikas.
18. I omit here the rather long note where the bibliography on the "dating" of Vasubandhu is summarized, and where the texts proving the existence of an "earlier Vasubandhu" were brought together; see below.
19. All the opinions, or almost all the opinions, marked in the Ko? a or in the Bhdsyam by the adverb kila ("certain," "it is said," grags so), are erroneous opinions of the Vaibhasikas. A correct translation would be: "The School says, wrongly, that. . . "
20. N. Peri, "A propos de la date de Vasubandhu," BEFEO, 1911, p. 374. The Tibetan Siddhantas also take a great deal from the Kosa. Note that it was translated into Chinese only in 563, and the Tibetan version, by Jinamitra and Srikutaraksita, during the period of Ral-pa-can (816-838).
21. This does not exist in extenso in Chinese (JPTS, 1905, p. 77). This is the treatise the first two parts of which are analyzed in the Appendix of Cosmologie bouddhique.
22. The Vydkhyd, the commentary on the Bhdsyam by Yasomitra, adds many details.
? 23. It is from this point of view that Oldenberg recommends the study of the Abhidharmakoiay in Bttddhistische Studien, ZDMG; III; p. 644 (1898).
24. According to P. Demieville, BEFEO, 1924, p. 463.
25. With reference to Gunamati, see H. Ui, Studies in Indian Philosophy {^Indo-tetsugaku
Kenkyu], 5th volume, pp. 136-140.
26. Missing in the two treatises of Sarhghabhadra, the Aryan quoted in Vydkhyd i. 31, which is a criticism of Kosa i, kdrika 11.
In the two treatises of Sarhghabhadra, the first chapter is entitled MiUavastunirdesa, the second Vitesanirdefa, the third Pratityasamutpddanirdes'a. As is proper, the Pudgalapratisedha- prakarana, an appendix to the Kofa, is ignored.
27. It has been preceded by Susumu Yamaguchi [September 1931].
28. See Ko? a i. l. Obermiller, in the preface to his translation of the Uttaratantra {Asia Majory 1931), digresses from the thesis of Ui.
29. Perhaps because the work of Dharmatrata enjoyed, for a long time, a great reputation; because, in the eyes of the Sarvastivadins, the Kola passed, with good reason, for heretical and tendentious.
30. Quoted iii. 59, on the explanation of the word utsada.
31. See below.
32. Satpdddbhidharmamdtrapdfhittas, a good reading for the Kosa, v. 22, note 80.
33. An account of the council in Ta-chih-tu-lun. Przyluski, Concile de Rdjagrha, p. 72.
34. Translated in 383 by Gotama Samghadeva of Kasmir, and by Hsiian-tsang.
35. Watters, i. 294; S. Levi, Catalogue geographique des Yaksas, 55; J. Przyluski, Acoka, 263.
36. On the language in which the Jndnaprasthdna was written, see Takakusu, p. 82, 84, 86. See above p. 3.
37. On the laukikdgradharmasy Kofa, vi. 19c, and "Parayana quoted in the Jndnaprasthdna" Melanges Unossier (where we see that the Jndnaprasthdna poorly presents the problem of the nirvedhabhdgiyas).
38. Same text, Small Vibhdsd, p. 5b.
39. Compare Ko/a. i. 3, and Documents d'Abhidharma; Vibhdsd, p. 236b.
40. The controversy of time and of the pudgala in the Vijndnakdya, in Etudes Asiatiques, 1925, i. 343-376; Inde sous les Mauryas, 1930,138; Note in Bouddhique, ii. AC Belgique, Nov. 1922.
The fourth volume of theJapanese translation reached me in September, 1931. It contains the Vijndnakdya. The translator, Bun'yu Watanabe, in a short preface, treats of the philosophic import of the book, of its compilation, and its relation with the Abhidhammas.
41. One must be more precise with respect to the remarks made by Barth (ii. 355): in truth the editors of the Dipavamsa are alone in knowing a Tissa Moggaliputta "who must have presided over the council of Ashoka and composed the Kathdvatthu. " But the Sarvastivadin tradition knows of a Mu-lien to whom it attributes, in the controversy of the past and the future, the position that the Dipavamsa assigns to Tissa. There is certainly much legend in Singhalese hagiography.
42. The enigmatic Gopala of Hsiian-tsang? Our sources are in agreement in attributing to the
Pomsin 51
? 52 Introduction
Sammltlyas, and to the VatsiputrTyas, the doarine of the pudgala. See Madhyamakavrtti, pp. 275-276.
43. Kathavatthu: saccikapphaparamapphena puggalo upalabbhati. The Sanskrit formula is not restored with any certainty. We have tattvdrthatah (satydrthatah? ) paramdrthatah pudgala upalabhyate sdksikriyate sampratividyate (? ) sathvidyate. The edition of Devasarman, more developed than that of Tissa, appears to be later.
44. Sanskrit sources like the expression purusapudgala; for example, the Sanskrit edition of Majjhima, iii. 239 (chadhdturo ayam puriso) has saddhdtur ayarh purusapudgalah. See Madhya- makavrtti, pp. 129,180, etc
45. The expression that I translate as "Recognize the contradiction into which you fall! ",/# t'ing tuofu, corresponds to a Sanskrit original ajdnihi nigraham. Tuofu is in fact translated by nigraha sthdna in a word list {Tetsugaku Daijisho, Tokyo, 1912) abstracted by Rosenberg ("Introduction to the study of Buddhism," i, Vocabulary, Tokyo, 1916). Thus we have here the exact equivalent of the formula djdndhi niggaham of the Kathavatthu and the MUinda.
"What is srotadpannaphala? If the three klesas to be cut off by Seeing are cut off; if satkayadrsti, vicikitsd and silavrata are exhausted, this is called srotadpanna-
12
phala" Sariputra takes up the question again, "What is srotadpannaphala} The
three klesas to be cut off by Seeing being cut off, satkdyadrsti-vicikitsd-silavrata being exhausted: if one obtains amrta, this is what is called srotadpannaphala"
It appears that what we are encountering here is a terminology alien to the Abhidharma and to the Sarvastivada.
73 viii. The Abhidharmasdra.
1. Before Vasubandhu, many masters undertook to summarize the doctrines of the Abhidharma. We possess notably three works: 1. The Abhidharmasdra of
74
Dharmas'rl in ten chapters, made up of kdrikds (probably in dryan stroph)
and a
commentary; 2. a second edition of this same Sara by Upasanta, to which the
Chinese give the name of Abhidharmasdra-ching [=sutra]: the same kdrikds with
a more developed commentary; and 3. a third edition of the Sara, the Tsa
[=Miscellaneous] Abhidharma-ching, by Dharmatrata, which is in fact a new
75 work, containing a new chapter and many new kdrikds.
76
2. The preface to the Vibhdsd (Taisho 1546)
Dharmasri before the Jndnaprasthdna: "After the nirodha of the Buddha, the bhiksu Dharmas'ri composed the four volumes of the Abhidharmasdra. Then Katyayanlputra composed the Abhidharma in eight books . . . "
3. The work of Dharmasri contains ten chapters: Dhdtu, Samskara, Anusaya, Arya, Jndna, Samddhi, Sutra, Tsa and Sdstravarga or Vddavarga.
Between the ninth and the tenth chapters of Dharmasri, Dharmatrata places a new chapter, the Pravicayavarga, which indeed appears to constitute an independent work.
There is a stanza of introduction: "even though many dharmas have been spoken of, their meaning remains confused . . . " and four concluding stanzas: "The author has composed this book based on the book of Dharmasri, not through pride or in order to acquire a reputation . . . ".
It begins with the dharmacakra, the Wheel of the Dharma: "The Muni said that the darsanamdrga is called dharmacakra, either because it goes into the mind
by Tao-yen places the work of
? of others . . . (Kosa, vi. 54).
There then comes the brahmacakra (vi. 54, vii. 31), the updsaka (iv. 69 ), the
four parts of sila (iv. 29), the prdtimoksa . . . Later (p. 959b), cosmology: the periods of loss, etc. (iii. 99), destruction by fire, etc. (iii. 102); and then there follows the theory of the three "fallings away" (p. 960c; Kosa, vi. 59) and the definition of the Bodhisattva (iv. 108).
Suddenly (p. 96lc): "How many types of Sarvastivada are there? " Presenta-
tion of the four doctrines (Kosa, v. 25-26) without mentioning the name of the
four masters. The second and the fourth are bad because they confuse the time
periods. The first (difference in bhdva, translated fen): "One should know that
77 this is the parindma-sarvastivada. "
There is a diversity of opinion as to whether the Truths are seen at the same time (Kosa, vi. 27), Sarvastivadins and Vatslputrlyas on the one hand, Dharma- gupta on the other; antardbhava (iii. 34); then the Sarvastivadin proof. And at the end of the paragraph, the discussion "whether the Buddha is part of the Samgha. " Finally, the concluding stanzas.
4. The Samskdravarga treats of the simultaneous arising of the citta-caittas and of atoms (Kosa, ii. 22), of the four laksanas of "conditioned things" (ii. 45), of the hetus and the pratyayas (ii. 48, 61).
The Sutravarga is a collection of notes on the three Dhatus and a calculation of the places that they contain: sixteen in Rupadhatu, but, according to some, seventeen (Kosa, iii. 2): the sattvdvdsas (iii. 6), the vijndnasthitis (iii. 5-6); the three vartman of pratityasamutpdda (iii. 20, 25), the twelve limbs; the mahdbhutas, the Truths, the fruits of the Aryans, etc.
The Tsa-varga defines the mind-mental states as samprayukta, sdsraya, etc. (ii. 34); it enumerates the viprayuktas: dsamjnika, two non-conscious absorptions, sabhdgatd, ndmakdyddayas, jivitendriya, dharmaprdpti, prthagjanatva, four laksanas (compare ii. 35-36); it concludes with half a karika on the four bhavas (in. 13) and a karika on "disgust" amd "detachment" (vi. 79).
The Sdstravaraga (or Vddavarga) is made up of ten questions in verse, followed by answers in prose, relative to samvara (iv. 13), to the results, etc. Dharmatrata adds sixteen questions.
5. In order to appreciate the character of the treatises of DharmaSrT, Upasanta, and Dharmatrata, and Vasubandhu's debt with respect to Dharmatrata, which appears to be notable, we may see how two dharmaparydyas, the chapter of the three obstacles (dvaranas) and that of avijnapti, are treated by the different masters.
a. Obstacles, Kola, iv. 95-102, Vibhdsd, p. 599.
Poussin 45
? 46 Introduction
Dharmasri, p. 815: "The Bhagavat says that there are three avaranas: karman, kiefa, and vipdka. What is their definition?
"Anantarya actions which are without remedy, developed defilements, bad aaion experienced in the painful realms of rebirth, are the dvaranas.
"These three form an obstacle to the Dharma; they hinder the grasping of the Aryan dharmas\ they are thus called 'obstacles/ Which is the worst aaion?
"The action which divides the Sarhgha is said to be the worst.
"This action is the worst. One guilty of this remains a kalpa in Avici hell. Which is the best aaion?
"The cetand or 'volition' of Bhavagra is the greatest. u
Naivasamjndndsamjndyatana is Bhavagra. The volition which belongs to the realms of this sphere is the greatest and finest: its result is a life of some 80,000 kalpas in length. "
Upasanta, p. 843b-c, has the same two stanzas, but a less meager commentary:
"That which hinders the Path of the Aryans and the means (updya) to this path is said to be an obstacle. The obstacles to aaion are the five dnantaryas, namely, the killing of one's father, etc . . . He who commits such an aaion is immediately and necessarily reborn in Avici: thus the aaion is dnatarya. The killing of one's father and mother destroys goodness, hence it is Avici hell. Those guilty of the other three are led to injure a field of merit. The obstacles of the defilements are 'agitated' and 'sharp' defilements: the first is habitual defilement; the second is the overriding defilement. This refers to the 'present' defilements, not to the defilements that one 'possesses' (that one has as potential), for all beings 'possess' all the defilements . . . "
There is a variant to the second stanza: "Lying which divides the Samgha . . . ; volition in Bhavagra, among good aaions, has the greatest result," which is better. The commentary notes the differences in the two schisms (cakrabheda, karmabheda).
Dharmatrata is longer (p. 898b-899c) and very close to Vasubandhu:
According to the first stanza of Dharmasri: 1. the dvarana of klesa is the worst; the dvarana of aaion, mediocre; and the dvarana of retribution, the least; 2. the division of the Samgha, by nature, is non-concord; this is a viprayukta samskdra of the anivrta-avydkrta class; 3. the division is a thing of the Samgha; the transgression is of him who divides the Samgha; he experiences, in Avici, a retribution of kalpa\ 4. the bhiksus are divided in their opinion of who is the Master, of what is the Path: this is the division of the Samgha which was united, and he who breaks it is 'one who possesses views' {drspicarita)\ 5. in three continents, a minimum of eight persons is required for karmabheda\ in
? JambudvTpa, a minimum of nine persons is required for cakrabheda; 6. cakrabheda is impossible in six time periods: when a boundary is not delimited; at first; following; when the Muni has passed into Nirvana; when the running sore has not been produced; when the pair of chief disciples has not been established (six pddas); 7. lying which divides the Sarhgha is the worst of aaions; the volition of Bhavagra is said to bear the greatest result.
b. Avijnapti, Kosa, iv. 2, 3.
DharmasrT (p, 812c):
"Bodily action is vijnapti and avijnapti. The vijnapti of the body is the movement of the body, good, bad, or neutral: good when it arises from a good mind . . . For avijnapti: when one does an action in a firm manner, the mind can change, but the seed remains. If, for example, a person undertakes the precepts, his mind can then be bad or neutral: nevertheless the precepts continue . . . Action of the manas is solely avijnapti. . . because this action is not visible . . . vijnapti is good, bad, or neutral; the same for the avijnapti which belongs to the manas. The other vijnaptis are never neutral. "
Upasanta (p. 840) adds a bit. The hunter is regarded as free from bodily avijnapti. Mental action is called avijnapti because it does not inform others. Some say that it is called vijnapti because it is discourse (jalpa? ).
Dharmatrata (p. 888b) replaces the terms vijnapti and avijnapti with "doing" and "not doing" (karana, akarana) (Kos'ajv. 14):
"Bodily action is of two types: karanasvabhava, or akaranasvabhava. 'Doing'
78
(karana): movement of the body, exercise
'Not doing' (akarana, wu-tso): when the movement of the body has ended, the nature (good or bad) of this movement, of this action which is the movement, continues to arise, simultaneous with minds of a different nature, even as the good precepts produced by their undertaking (kusalasamddanasila) continue to arise even when bad or neutral minds are present. Like the immoral person (dauhsilya- purusa): even when good or neutral minds are present, his immorality continues to arise. "
". . . Action of the manas is cetand, volition, by nature . . .
"'Not doing' (akarana) is also called nirati (? Vyut. 21, 114), virati; upeksa, akriyd (pu-tso). Because it does not do, it is called 'not doing. ' If one says that this is not an action (karman), this is wrong, because it does. Good does not do evil, evil does not do good: this is also an action. As the upeksa part of Bodhi is not upeksa by reason of what is called upeksa; but the practice of the Path, the arresting of things, is called upeksa. The same here. Furthermore, in doing the
of the body is the 'doing of the body. '
Poussin 47
? 48 Introduction
cause one does the result: . . . 'not doing' is not rupa, but the doing of it (which is the cause of the 'not doing' or avijnapti) is rupa\ 'not doing' is thus called riipa. In
79 this same way then, 'not doing' is action.
? 1. Originally printed as a Foreword to de La Vallee Poussin's Cosmologie bouddhique: 1913, and published 1919 in the four-part Memoires of l'Acadmie royale de Belgique (Luzac, London). This contained the restoration of the kdrikds of the third chapter of the Ko/a, the Tibetan kdrikds, the Bhdsya, and the text of the Vydkhyd\ in the appendix, a summary of the Lokaprajnapti and the Karanaprajnapti.
2. Cullavagga, xi. 1. 8. [For a more exact presentation,J. Przyluski, Concile de Rdjagrha, p. 311,345, 349].
3. Oldenberg, Buddha . . . , 6th edition, p. 202; Fr. trans. Foucher, 2nd edition, p. 177. [Psychology, yes; but ontology is doubtful].
4. In the Divya, a Sarvastivadin work, where we encounter the expressions sutrasya vinayasya mdt- rkdydh: "The monks ask with respect to the Sutra, the Vinaya, and the Mdtrkd" (p.
18,15), and sittram mdtrkd ca, equivalent to dgamacatustayam (p. 333,7), Kern {Manual, p. 3) thinks that the term mdtrkd is employed "as synonymous with abhidharma. " It cannot in any case designate the Abhidharmas of the Sarvastivadins of which we are speaking below (p. 3) which are treatises; it fits the Abhidhammas a little less poorly, but without being satisfying. Does it designate some lists "omitting all the explanations and other details" (Childers, 243), lists of items which form part of the dgama and which are not specifically Abhidharma! The Sautrantikas, who deny the existence of an Abhidharma Pitaka distinct from the Sutra, certainly had such an "index," exactly like the Sarvastivadins of that period, and earlier than the Abhidharmas to which the expressions
of the Divya refer. Does it designate some presentations, in the manner of the sutras, like those that constituted the mdtikds of the Vibhanga? , In this book, which is the property of the Abhidharma, it is often a type of commentary in the form of glosses.
5. See the article of Rhys Davids in Hastings' Encyclopedia of Religion and Ethics.
6. Developpement de VAbhidharma, Dogme et philosophies 122; J. Przyluski, Concile, third
chapter and 179, 353; Acoka, 322; ? un4railles, 49; Levi, Seize Arhats, 20, 39.
7. In Hastings' Encyclopedia, 1. 19-20; Winternitz, Geschichte, 134. Scholastic definitions of the
Abhidharma: AttasMini, 48-50 and following; Satrdlamkdra, XI. 3.
8. It knows, however, that the author of the Kathdvatthu foresaw and refuted in advance the heresies to come; see Atthakathd, pp. 6-7. The remark is by Minayeff, and the observations by H. Oldenberg (Buddh. Studien, p. 633, 676) do not demonstrate that the Kathdvatthu has not been amplified in the course of time.
9. This is incorrect. The Samgitiparydya is only the Samgitisuttanta. The second part of the Dhdtukdya has a close relationship with the Dhdtukathdprakarana. A careful study will show other points of contaa, and one can see that the Sarvastivadins simply enriched by their inventions (theory of the viprayuktas, of the mahdbhdmikas, etc. ) the earlier material of the Abhidhamma.
10. The account of Buddhaghosa, Kathdvatthu-Atthakathd, p. 6, holds that because of this, at least the Vibhajjavadins are the orthodox party.
11. This is one of the aspects of the problem of kiriyavdda.
12. This definition of the two schools is borrowed from the Abhidharmako/a, v. 9; see Kathdvatthu, 1. 8 (which does not entirely confirm our interpretation). The controversy of time and the pudgala in the Vijndnakdya, Etudes Asiatiques, 1925.
13. Geography of the Sarvastivadin sect, J. Przyluski, Apoka. I know that Sinologists, notably Takakusu, are not settled on the language in which the first of the Abhidharmas of the
Poussin 49
? 50 Introduction
Sarvastivadins, the Jndnaprathdna, was written: "In what language, however, the original text was composed we have no means of ascertaining. All we can say is that the text brought by Sarhghadeva and Dhammapiya [Dharmapriya] from Kacmira [383 A. D. ] seems to have been in
a dialect akin to Pali, whereas the text used by Hiuen-tsang [657 A. D. ], as in other cases, seems to have been in Sanskrit. But this supposition rests solely on the phonetic value of Qiinese ideographs employed in these translations, and is not corroborated by any other evidence . . . It seems to me more than probable that the JHdnaprasthdna at least was written in some dialect: one thinks naturally of the dialect of Kacmira, but we really have no certainty that the Jndnaprasthdna was not composed in Kosala (JPTS; 1905, p. 84,86). "
We possess a fragmentary quotation from the Sarvastivadin Pratimoksa which proves that some earlier forms, Paji or dialect, remained in use: "When, in the Posadha ceremony, the Vinayadhara asks, 'Are you pure? ' (bhiksuposadhe hi kacci ttha parisuddhd iti vinayadharend- nusrdvite), if any bhiksu does not confess his transgression . . . " (Abhidharmakofavydkhydad iv. 72; compare the introduction of the Pdtimokkha and the remarks of Rhys Davids, Dialogues, II, p. 257). (See L Finot, "Pratimoksasutra des Sarvastivadins," JA, 1913,2. 177-9).
But we possess a fragment of the Jndnaprasthdna, quoted in the Abhidharmakosavydkhyd (ad i. 49): katamad buddhavacanam tathdgatasya yd vdg vacanam vyavahdro gir niruktir vdkpatho vdgghoso vdkkarma vdgvijnaptih // buddhavacanam kusalam vaktavyam athdvydkrtam vak- tavyam / sydt kusalam sydd avydkrtam // katarat kusalam / kusalacittasya tathdgatasya vdcarh bhdsamdnasya yd [vdgjvijnaptih // katarad avydkrtam / . . . purvavat // punas tatraivdnantaram uktam / buddhavacanam ka esa dharmah / ndmakdyapadakdyavyanjanakdydndm ydnupur- vavacand anupurvasthdpandanupurvasamayoga iti /
14. Invention of the prdptis, of sabhdgatd, of the existence of the past and the future, of diverse types of cause, of apratisamkhyanirodha, not to mention the nirvedhabhagiyas, etc.
15. The Abhidharmako/avydkhyd speaks of satpdddbhidharma-mdtrapdthina Abhidharmikas, Abhidharmikas "who read only the six-legged Abhidharma" which we understand to mean "who do not read the Vibhdsd. " These are Sarvastivadins; but all Sarvastivadins are not "followers of the Vibhdsa' (Vaibhasikas). We know, for example, that there were four ways of understanding "all exists," those of the Sarvastivadins Dharmatrata, Ghosaka, Vasumitra, and Buddhadeva: the Vaibhasikas of Kas'mTr condemn the first, the second and the fourth; and the first for the serious reason that it is confused with the non-Buddhist teaching of the Saihkhyas.
16. Better, "after the reign of Kaniska," Inde sous les Mauryas . . . , p. 328.
17. See this Introduction, Darstantikas, and Index, Sautrantikas.
18. I omit here the rather long note where the bibliography on the "dating" of Vasubandhu is summarized, and where the texts proving the existence of an "earlier Vasubandhu" were brought together; see below.
19. All the opinions, or almost all the opinions, marked in the Ko? a or in the Bhdsyam by the adverb kila ("certain," "it is said," grags so), are erroneous opinions of the Vaibhasikas. A correct translation would be: "The School says, wrongly, that. . . "
20. N. Peri, "A propos de la date de Vasubandhu," BEFEO, 1911, p. 374. The Tibetan Siddhantas also take a great deal from the Kosa. Note that it was translated into Chinese only in 563, and the Tibetan version, by Jinamitra and Srikutaraksita, during the period of Ral-pa-can (816-838).
21. This does not exist in extenso in Chinese (JPTS, 1905, p. 77). This is the treatise the first two parts of which are analyzed in the Appendix of Cosmologie bouddhique.
22. The Vydkhyd, the commentary on the Bhdsyam by Yasomitra, adds many details.
? 23. It is from this point of view that Oldenberg recommends the study of the Abhidharmakoiay in Bttddhistische Studien, ZDMG; III; p. 644 (1898).
24. According to P. Demieville, BEFEO, 1924, p. 463.
25. With reference to Gunamati, see H. Ui, Studies in Indian Philosophy {^Indo-tetsugaku
Kenkyu], 5th volume, pp. 136-140.
26. Missing in the two treatises of Sarhghabhadra, the Aryan quoted in Vydkhyd i. 31, which is a criticism of Kosa i, kdrika 11.
In the two treatises of Sarhghabhadra, the first chapter is entitled MiUavastunirdesa, the second Vitesanirdefa, the third Pratityasamutpddanirdes'a. As is proper, the Pudgalapratisedha- prakarana, an appendix to the Kofa, is ignored.
27. It has been preceded by Susumu Yamaguchi [September 1931].
28. See Ko? a i. l. Obermiller, in the preface to his translation of the Uttaratantra {Asia Majory 1931), digresses from the thesis of Ui.
29. Perhaps because the work of Dharmatrata enjoyed, for a long time, a great reputation; because, in the eyes of the Sarvastivadins, the Kola passed, with good reason, for heretical and tendentious.
30. Quoted iii. 59, on the explanation of the word utsada.
31. See below.
32. Satpdddbhidharmamdtrapdfhittas, a good reading for the Kosa, v. 22, note 80.
33. An account of the council in Ta-chih-tu-lun. Przyluski, Concile de Rdjagrha, p. 72.
34. Translated in 383 by Gotama Samghadeva of Kasmir, and by Hsiian-tsang.
35. Watters, i. 294; S. Levi, Catalogue geographique des Yaksas, 55; J. Przyluski, Acoka, 263.
36. On the language in which the Jndnaprasthdna was written, see Takakusu, p. 82, 84, 86. See above p. 3.
37. On the laukikdgradharmasy Kofa, vi. 19c, and "Parayana quoted in the Jndnaprasthdna" Melanges Unossier (where we see that the Jndnaprasthdna poorly presents the problem of the nirvedhabhdgiyas).
38. Same text, Small Vibhdsd, p. 5b.
39. Compare Ko/a. i. 3, and Documents d'Abhidharma; Vibhdsd, p. 236b.
40. The controversy of time and of the pudgala in the Vijndnakdya, in Etudes Asiatiques, 1925, i. 343-376; Inde sous les Mauryas, 1930,138; Note in Bouddhique, ii. AC Belgique, Nov. 1922.
The fourth volume of theJapanese translation reached me in September, 1931. It contains the Vijndnakdya. The translator, Bun'yu Watanabe, in a short preface, treats of the philosophic import of the book, of its compilation, and its relation with the Abhidhammas.
41. One must be more precise with respect to the remarks made by Barth (ii. 355): in truth the editors of the Dipavamsa are alone in knowing a Tissa Moggaliputta "who must have presided over the council of Ashoka and composed the Kathdvatthu. " But the Sarvastivadin tradition knows of a Mu-lien to whom it attributes, in the controversy of the past and the future, the position that the Dipavamsa assigns to Tissa. There is certainly much legend in Singhalese hagiography.
42. The enigmatic Gopala of Hsiian-tsang? Our sources are in agreement in attributing to the
Pomsin 51
? 52 Introduction
Sammltlyas, and to the VatsiputrTyas, the doarine of the pudgala. See Madhyamakavrtti, pp. 275-276.
43. Kathavatthu: saccikapphaparamapphena puggalo upalabbhati. The Sanskrit formula is not restored with any certainty. We have tattvdrthatah (satydrthatah? ) paramdrthatah pudgala upalabhyate sdksikriyate sampratividyate (? ) sathvidyate. The edition of Devasarman, more developed than that of Tissa, appears to be later.
44. Sanskrit sources like the expression purusapudgala; for example, the Sanskrit edition of Majjhima, iii. 239 (chadhdturo ayam puriso) has saddhdtur ayarh purusapudgalah. See Madhya- makavrtti, pp. 129,180, etc
45. The expression that I translate as "Recognize the contradiction into which you fall! ",/# t'ing tuofu, corresponds to a Sanskrit original ajdnihi nigraham. Tuofu is in fact translated by nigraha sthdna in a word list {Tetsugaku Daijisho, Tokyo, 1912) abstracted by Rosenberg ("Introduction to the study of Buddhism," i, Vocabulary, Tokyo, 1916). Thus we have here the exact equivalent of the formula djdndhi niggaham of the Kathavatthu and the MUinda.
