31, which is a
criticism
of Kosa i, kdrika 11.
Abhidharmakosabhasyam-Vol-1-Vasubandhu-Poussin-Pruden-1991
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.
46. The doarine of the Bhagavat is a path between two extremes. It avoids the extreme theory of permanence by saying that he who eats the fruits of the aaion is not the same person who carried out the aaion (sa karoti so'nubhavati? ): it avoids the theory of annihilation by denying that he who eats the fruit is anyone other than he who carried out the aaion. Compare Samyutta, ii. p. 23.
47. The edition of the Dirgha has only four oaades: vimoksa, abhibhu, lokadharma and samyag- mdrga (which recalls the Pali samattas).
48. On the different Vasumitras, see Watters, i. 274-5; the Introduction to the Traite surles Sectes, Masuda, Asia Major, ii. p. 7; T&ranatha, 174.
49. Who (Vibhdsd, p. 152a) declares that all citta-caitta disappears in nirodhasamdpatti?
50. Perhaps Taisho 1556, anonymous: Sarvdstivddasamaya-pancavastuka, notable for its enumera-
tion of the viprayuktas; 997c: prdpti, asamjnisamdpaUi, nirodhasamdpatti, dsamjnika, fivitendriya, sabhdgatd, desaprdpti (? ), dyatanaprdpti, jdti'jard, sthiti, anityatd, ndma-pdda-vyanjanakdyas.
51. It quotes the Abhidharma-acaryas. It admits the mahdbhumikas, p. 994b3, and also avijnapti, p. 992c, which it explains, along with the samvaras, exaaly as the Kosa does. But the formula, "There are two gates to amrta, the contemplation of the loathsome and the regulation of the breath," is to be noted, p. 989b.
52. On the andgdmin, Vibhdsd, p. 879b; on the meaning of bhava, p. 960b; on the absorptions which follow a good mind of Kamadhatu, p. 961c.
53. One gloss says that the word signifies "Bodhi-taking," which would give Bodhilata.
54. The Darstantikas deny the caittas; the Sautrantikas admit the caittas, but differ on their number.
55. See above, p. 30.
56. Sautrdntikd Bhadantddayah; sautrdntikadarsanavalambin.
57. According to Wassiliew, 279, Samgharaksita differs a little. We do not know this master, nor
the Bhumisena of p. 280.
58. The opinion on seeing through the visual consciousness, on the number of the caittas, on the non-existence of avijnapti, are clearly non-Vaibhasika. Moreover, the Vibhdsd carefully notes the opinions of the Bhadanta.
? 59. See also Vibhdsd, p. 219.
60. However Dharmatrata, in Abhidharmasdra, p. 885, explains the causes of the viprayuktas.
61. See the references to the Mahasaihghikas, the VatsTputrlyas, the Mahisasakas, and the Dharmaguptas in the index.
62. Ch'uan-yil\ Ch'uan translates agama or avavdda.
63. Vasubandhu, Kofa, iv. 56, as well as Samghabhadra, p. 572 (which is surprising enough), do
not take into account the second alternative and follow the doctrine of the ch'mn-yil.
64. A note translated imprecisely in Kosa, v. 9d, note.
65. The truth is that Paramartha wrote one word for another.
66. This is obscure; the Samgraha quotes the Agama of the Mahisasakas and ignores the Vibhajyavadins; the Vibhdsd, it appears, ignores the Sammitlyas.
67. This is the meaning oiyogdcdra in the Saundarananda, and in the Mahdvastu.
68. Or vaiydprtya, Avaddnafataka, ii. 235.
69. "Some say: 'When the Buddha was in this world, Sariputra, with an end to explaining the words of the Buddha, compiled the Abhidharma. Later, the monk VatsTputrTya recited [this work]. Up to the present day, this is what is called the Abhidharma of Sariputra,'" J. Przyluski, Concile, p. 73.
The only book of the Abhidharma which teaches the doctrine of the pudgala appears to be Taisho vol. 32, no. 1649, the Sammittya-nikdya-sdstra, Kofa, ix.
70. "The past is that which has arisen and is destroyed; the future is that which has not arisen, not appeared," p. 543b.
71. For the last terms, compare the variant p. 526c: the Sanskrit reading is doubtful:
dkds'dyatanajndna . . . and dkdsdyatanapratyayatfndna].
72. Compare the doctrine of the Dhammasangani, Kola v. 4, note.
73. Taisho volume 28, numbers 1550, 1551,1552, Abhidharmahrdaya; see above p. 16.
74. We have the Sanskrit text of one of the kdrikds, Kosa, v, note 14.
75. For example, the ninth chapter: Dharmatrata takes up twenty kdrikds of Dharmasn and
interpolates six new kdrikds; he continues with twenty-two new kdrikds. 16. Taisho 1546; mentioned by Takakusu, p. 128.
77. The third doctrine, difference in avasthd, (trans, fen-fen).
78. We have fang-pien, which should translate vydydna (see Demieville, Milinda) more often than updya.
79. See above p. 33.
Poussin 53
? CHAPTER ONE
The Dhdtus
. Homage to the Buddha.
1. He has, in an absolute manner, destroyed all blindness; He has drawn out the world from the mire of transmigration: I fender homage to Him, to this teacher of truth, before composing the treatise called the Abhidharmakosa.
Desiring to compose a treatise, with the intention of making known the greatness of his master, the author undertakes to render him homage and to first present his qualities.
? 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.
46. The doarine of the Bhagavat is a path between two extremes. It avoids the extreme theory of permanence by saying that he who eats the fruits of the aaion is not the same person who carried out the aaion (sa karoti so'nubhavati? ): it avoids the theory of annihilation by denying that he who eats the fruit is anyone other than he who carried out the aaion. Compare Samyutta, ii. p. 23.
47. The edition of the Dirgha has only four oaades: vimoksa, abhibhu, lokadharma and samyag- mdrga (which recalls the Pali samattas).
48. On the different Vasumitras, see Watters, i. 274-5; the Introduction to the Traite surles Sectes, Masuda, Asia Major, ii. p. 7; T&ranatha, 174.
49. Who (Vibhdsd, p. 152a) declares that all citta-caitta disappears in nirodhasamdpatti?
50. Perhaps Taisho 1556, anonymous: Sarvdstivddasamaya-pancavastuka, notable for its enumera-
tion of the viprayuktas; 997c: prdpti, asamjnisamdpaUi, nirodhasamdpatti, dsamjnika, fivitendriya, sabhdgatd, desaprdpti (? ), dyatanaprdpti, jdti'jard, sthiti, anityatd, ndma-pdda-vyanjanakdyas.
51. It quotes the Abhidharma-acaryas. It admits the mahdbhumikas, p. 994b3, and also avijnapti, p. 992c, which it explains, along with the samvaras, exaaly as the Kosa does. But the formula, "There are two gates to amrta, the contemplation of the loathsome and the regulation of the breath," is to be noted, p. 989b.
52. On the andgdmin, Vibhdsd, p. 879b; on the meaning of bhava, p. 960b; on the absorptions which follow a good mind of Kamadhatu, p. 961c.
53. One gloss says that the word signifies "Bodhi-taking," which would give Bodhilata.
54. The Darstantikas deny the caittas; the Sautrantikas admit the caittas, but differ on their number.
55. See above, p. 30.
56. Sautrdntikd Bhadantddayah; sautrdntikadarsanavalambin.
57. According to Wassiliew, 279, Samgharaksita differs a little. We do not know this master, nor
the Bhumisena of p. 280.
58. The opinion on seeing through the visual consciousness, on the number of the caittas, on the non-existence of avijnapti, are clearly non-Vaibhasika. Moreover, the Vibhdsd carefully notes the opinions of the Bhadanta.
? 59. See also Vibhdsd, p. 219.
60. However Dharmatrata, in Abhidharmasdra, p. 885, explains the causes of the viprayuktas.
61. See the references to the Mahasaihghikas, the VatsTputrlyas, the Mahisasakas, and the Dharmaguptas in the index.
62. Ch'uan-yil\ Ch'uan translates agama or avavdda.
63. Vasubandhu, Kofa, iv. 56, as well as Samghabhadra, p. 572 (which is surprising enough), do
not take into account the second alternative and follow the doctrine of the ch'mn-yil.
64. A note translated imprecisely in Kosa, v. 9d, note.
65. The truth is that Paramartha wrote one word for another.
66. This is obscure; the Samgraha quotes the Agama of the Mahisasakas and ignores the Vibhajyavadins; the Vibhdsd, it appears, ignores the Sammitlyas.
67. This is the meaning oiyogdcdra in the Saundarananda, and in the Mahdvastu.
68. Or vaiydprtya, Avaddnafataka, ii. 235.
69. "Some say: 'When the Buddha was in this world, Sariputra, with an end to explaining the words of the Buddha, compiled the Abhidharma. Later, the monk VatsTputrTya recited [this work]. Up to the present day, this is what is called the Abhidharma of Sariputra,'" J. Przyluski, Concile, p. 73.
The only book of the Abhidharma which teaches the doctrine of the pudgala appears to be Taisho vol. 32, no. 1649, the Sammittya-nikdya-sdstra, Kofa, ix.
70. "The past is that which has arisen and is destroyed; the future is that which has not arisen, not appeared," p. 543b.
71. For the last terms, compare the variant p. 526c: the Sanskrit reading is doubtful:
dkds'dyatanajndna . . . and dkdsdyatanapratyayatfndna].
72. Compare the doctrine of the Dhammasangani, Kola v. 4, note.
73. Taisho volume 28, numbers 1550, 1551,1552, Abhidharmahrdaya; see above p. 16.
74. We have the Sanskrit text of one of the kdrikds, Kosa, v, note 14.
75. For example, the ninth chapter: Dharmatrata takes up twenty kdrikds of Dharmasn and
interpolates six new kdrikds; he continues with twenty-two new kdrikds. 16. Taisho 1546; mentioned by Takakusu, p. 128.
77. The third doctrine, difference in avasthd, (trans, fen-fen).
78. We have fang-pien, which should translate vydydna (see Demieville, Milinda) more often than updya.
79. See above p. 33.
Poussin 53
? CHAPTER ONE
The Dhdtus
. Homage to the Buddha.
1. He has, in an absolute manner, destroyed all blindness; He has drawn out the world from the mire of transmigration: I fender homage to Him, to this teacher of truth, before composing the treatise called the Abhidharmakosa.
Desiring to compose a treatise, with the intention of making known the greatness of his master, the author undertakes to render him homage and to first present his qualities.
