The English version has not yet been audited. Please use it for reference only.

Chapter 1: One Book, Two Identities

The Abhidharmakośa is two books at once: the best digest the Sarvāstivāda ever received, and the most exact criticism it ever received. The two identities interleave paragraph by paragraph, and nothing between them announces that the scene has changed. This chapter builds the basic skill of the whole part: how to tell which of the two identities the sentence in one's hand is serving. Only when the rules have been laid down do the seven chapters that follow have common ground to stand on.

Format and composition. Verse and prose alternate. The words "the verse says" stand five hundred twenty-seven times in the book and "the treatise says" five hundred twenty-two, one verse to one exposition, with a regularity that is almost mechanical.1 The root verses are arranged in the main on Vaibhāṣika doctrine, gathering the whole stock of the Mahāvibhāṣā's two hundred fascicles into a frame of six hundred verses, the tally for each of the eight chapters printed beneath its own chapter title and adding up to exactly that; the prose expounds the verses one by one, and in the expounding sets up a view of its own again and again. The order of the eight chapters has a logic of its own. The chapters on the Elements and on the Faculties draw up the list of dharmas; the three on the World, on Action, and on the Latent Defilements treat the causes and effects of defilement; the three on the Noble Ones, on Knowledges, and on the Attainments treat the causes and effects of purity. The stock is laid out first, then the round of transmigration, then the return to cessation. That sequence is itself a Sarvāstivāda sequence, the frame taken over entire from the other side; and that is the first identity.

The verse of self-declaration at the end of the book. The second identity the book confesses itself, at the close. At the end of fascicle twenty-nine Vasubandhu puts a question to himself: this treatise takes up and summarizes the Abhidharma—on what principle, then, does it expound the Abhidharma? The verse answers: the deliberation of Kaśmīra has established the principle; I have for the most part followed them in expounding the Abhidharma; what little detraction there is, is my own fault; the judging of dharma by right principle rests with the Sage. His own prose note follows: the Vaibhāṣika masters of the land of Kaśmīra have well established the principle of the Abhidharma by their deliberation; I have for the most part followed them in expounding the doctrine of the Abhidharma; what little detraction there is, is my own fault; the judging of dharma by right principle rests with the Sage.1 Three things are pressed into four lines. I have for the most part followed them concedes that the frame belongs to somebody else. What little detraction there is, is my own fault concedes that there is detraction in the book: the posture is lowered, and the pen is not withdrawn. The judging of dharma by right principle rests with the Sage places the final power of adjudication not with the adjudicator of Kaśmīra but with the Buddha. Later tradition compressed the position into four characters, the stronger argument governs. The four never stand together in either of the two Tang commentaries: Puguang's Record has "he takes principle as his tenet," Fabao's Commentary has "he takes the better principle as his tenet" and "here the stronger argument is what is right," and of the two halves it is Fabao who first supplies the words for stronger argument.2 The compression is not wrong, only the verse is sharper than the compression, since it amounts to a declaration that the school's own rulings are not the last court.

The rules of discrimination: three markers. The two identities are told apart by markers the original carries in itself. First, "it is said." In Xuanzang's translation practice, wherever Vasubandhu restates a Vaibhāṣika doctrine while keeping a reservation of his own, the two characters for "it is said" are set upon it; they stand thirty-three times in the book. The verse line met in Part 16, Chapter 8 ("transmitted and granted, it is stated in terms of positions") belongs to the same family. One occurrence of "it is said" is one place where Vasubandhu declined to sign. Second, the signatures of the kind "the Vaibhāṣika masters say thus"; "the Vaibhāṣika masters" stands sixty-six times in the book. What is restated has an owner, and responsibility for the words is entered to a separate account. Third, the testimony from the other side. Saṅghabhadra's Nyāyānusāra calls Vasubandhu the sūtra-master throughout, two hundred fifty times;3 the greatest of his opponents takes sūtra for his surname as well, from which it appears that in the eyes of the opposition the book's second identity was the truer of the two. With the three markers used together the two identities can for the most part be read apart paragraph by paragraph. The stretches where the criticism runs densest are the passages on the dharmas dissociated from mind in the chapter on the Faculties and the debate over the three times in the chapter on the Latent Defilements, and the chapters that follow take their material by exactly this distribution.

The two Chinese translations. This part takes Xuanzang's thirty fascicles for its base text; the vernacular translation and the classical text are both in the library, and citations are claimed back against the original. Paramārtha's older translation, the Commentary on the Abhidharmakośa in twenty-two fascicles, is cited by way of comparison only where the doctrine diverges, and is not treated as a second base text; read side by side with Xuanzang, it diverges most at the verse of self-declaration at the end of the book.4 Differences of format between the two translations are handled in place, without a section of their own.

One Vasubandhu or two. Only the conclusion of the scholarly case is stated here. The old account makes the author of the Kośa one man with the Vasubandhu who turned to Yogācāra in later life; modern scholarship has argued for two Vasubandhus, and the question remains open to this day. This book takes the minimal commitment: the man who arranged the Kośa, who drew Saṅghabhadra's rebuttal, and who left this verse of self-declaration is one and the same treatise-master; his relation to the Yogācāra works is left to the part on Yogācāra to give an account of (a scholarly case; the present state of the scholarship to be tracked). Where this part says Vasubandhu, it means without exception the author of the Kośa.

To Chapter 2. The rules are laid down. The ground Vasubandhu stands on when he criticizes the Sarvāstivāda is called the Sautrāntika; what that ground is made of must be seen clearly first.

References

References

  1. Abhidharmakośabhāṣya, 30 fascicles, composed by Vasubandhu, trans. Xuanzang, Taishō vol. 29 (T1558). "The verse says," 527 occurrences; "the treatise says," 522; "it is said," 33; "the Vaibhāṣika masters," 66—all obtained by searching the whole classical Chinese text in the CBETA edition. The passage of self-declaration at the end of the book reads: "this treatise takes up and summarizes the Abhidharma: on what principle does it expound the Abhidharma? The verse says: the deliberation of Kaśmīra has established the principle; I have for the most part followed them in expounding the Abhidharma; what little detraction there is, is my own fault; the judging of dharma by right principle rests with the Sage. The treatise says: the Vaibhāṣika masters of the land of Kaśmīra have well established the principle of the Abhidharma by their deliberation; I have for the most part followed them in expounding the doctrine of the Abhidharma; what little detraction there is, is my own fault. The judging of dharma by right principle rests with the Sage." The exact number of the verses is settled by the chapter titles of Xuanzang's separate edition of the root verses, the Abhidharmakośakārikā in one fascicle (T1560): forty-four for the Elements, seventy-four for the Faculties, ninety-nine for the World, one hundred thirty-one for Action, sixty-nine for the Latent Defilements, eighty-three for the Noble Ones, sixty-one for Knowledges, thirty-nine for the Attainments, six hundred in all. A line count agrees: the eight chapters run to two thousand four hundred and four lines of verse, and all but two verses in the chapter on the World are of four lines. Puguang's Record of the Abhidharmakośa, fascicle one, has "he composed the Abhidharmakośa in six hundred lines of verse," the prose commentary running to eight thousand ślokas; Fabao's Commentary on the Abhidharmakośa, fascicle one, reckons more finely: "this treatise has six hundred and seven verses in all, divided broadly into three. The three verses at the front are the commentator's preface; the six hundred that follow are the root verses, which are the main body; the four at the end, together with the Refutation of the Self, are the commentator's conclusion"—three of preface and four of conclusion counted apart from the six hundred. On Paramārtha's side the biographical tradition gives the same figure: the Life of Vasubandhu says that he "composed in this order more than six hundred verses summing up the doctrine of the Vibhāṣā." For the modern Chinese version see this site's Abhidharmakośa section (rendered from the CBETA text). 

  2. Puguang, Record of the Abhidharmakośa, 30 fascicles, Taishō vol. 41 (T1821); Fabao, Commentary on the Abhidharmakośa, 30 fascicles, same volume (T1822). The four characters rendered "the stronger argument governs" nowhere stand together in either work; a search of both returns zero. Puguang supplies the second half and not the first. Fascicle one, on the occasion of the treatise, has "the treatise-master takes principle as his tenet and holds to no party"; fascicles eight and ten both have "the treatise-master takes principle as his tenet and does not take the adjudicators of the Vibhāṣā for his measure"—three places in all. Fabao supplies both halves. Fascicle one has "in the prose he takes only the better principle as his tenet and leans to no one school, though through the body of it he mostly takes the Sautrāntika sense for correct," and again "so it is clear that here the stronger argument is what is right, and no one school is settled upon"; fascicle ten has "the treatise-master does not hold fast to one school; where the argument is stronger, he takes it." Puguang studied under Xuanzang and his Record is the earlier of the two, Fabao's Commentary the later; the words for stronger argument, however, appear first in Fabao. The fixed four-character form is in neither work; its source lies elsewhere, and this book has not settled it. 

  3. Abhidharma Nyāyānusāra Śāstra, 80 fascicles, composed by Saṅghabhadra, trans. Xuanzang, Taishō vol. 29 (T1562). The appellation "the sūtra-master" stands 250 times in the whole work, on a search of the classical Chinese in the CBETA edition. 

  4. Commentary on the Abhidharmakośa, 22 fascicles, composed by Vasubandhu, trans. Paramārtha, Taishō vol. 29 (T1559). The fascicle count is given by Huikai's preface at the head of the work: "the text of the treatise, twenty-two fascicles; the verses of the treatise, one fascicle; the exposition of the sense, fifty-three fascicles." The one fascicle of verses is lost, so the old translation affords no separate verse text to count. The work of translating began in the first month of the fourth year of Tianjia, and the revised text was settled in the twelfth month of the first year of Guangda. Of the passages this part cites, possession not being real stands in fascicle three of the old translation, the unconditioned not being real in fascicle five, and the opening of the Refutation of the Self in fascicle twenty-two; in doctrine these agree with Xuanzang and differ in fascicle number and in wording. The substantive divergence is at the verse of self-declaration, which in the old translation stands at the end of fascicle twenty-one: "the deliberation of the Kaśmīra Vaibhāṣikas has established the principle; I have for the most part followed them in stating this treatise; where I have held to a party in the true dharma, that is my fault; for the judging of dharma by right principle the Buddha is the measure." The third line says partisanship where Xuanzang says detraction. The prose diverges further. The old translation has "the Blessed Buddha alone is the highest measure… and if the Buddha's noble disciples judge the true dharma apart from the Āgamas and from reasoning, they too are no fit measure," where Xuanzang has "the judging of dharma by right principle rests with the Blessed One alone, and with the great noble disciples of the Tathāgatas." Whether the noble disciples carry authority the two versions answer in opposite ways, the old translation drawing the power of adjudication in tighter. The passage on awareness arising with respect to an object belongs to the Nyāyānusāra and has no old translation to set beside it.