Chapter 1: What Kind of Book the Two Hundred Fascicles Are¶

For sheer bulk the Mahāvibhāṣā has few rivals among the Buddhist texts translated into Chinese: two hundred fascicles, one million five hundred and forty thousand characters of classical Chinese, and a vernacular translation running to well over two million. Faced with a book of that size, the first question is a question about size: what entitles it to be so long. The question is not rhetorical, and the answer settles how this whole part will read the book. It is long because it multiplied out everything there was to multiply; it is not complex because the multipliers are few. This chapter turns that sentence into a finding that can be checked.

Look first at what it says it is. The very first question of fascicle one asks who composed this treatise. The answer is the Buddha, the World-Honoured One, and the reason given is that the nature of what is knowable in all its modes is profound and subtle, and that apart from the Buddha, the World-Honoured One, the possessor of omniscience, who could fully awaken to it and set it forth. As for whose mouths the questions and answers in the book come from, some say that the venerable Śāriputra asks and the Buddha answers.¹ That opening has already given away the book's understanding of itself: the truth has been stated in full by the Buddha, and the work of the treatise-masters is not discovery but exposition and arrangement. The formulas that state a motive for composition match this. Whenever the book sets up a topic it commonly opens by asking why this treatise is made, and two set phrases recur in the answers. One is that it is for the sake of stopping other schools and displaying its own thesis, which together with its variants (displaying the orthodoxy, displaying the right principle) comes to more than a hundred places. The other is that the sūtra in question is the root on which this treatise rests, and that what the sūtra left unsaid should now be said: fifteen places word for word, and seventy-five for the family of phrases built on "the root on which it rests."¹ Fascicle eight is where it is first used, and the case is the Sūtra of the Simile of the Pool: the sūtra analysed the twenty statements of the view of the existing body only in brief and did not say how many of them are views of a self and how many views of what belongs to a self; that sūtra is the root on which this treatise rests, and what it left unsaid the treatise now wishes to say, and therefore this discussion is made.¹ Put the two set phrases together and one has the charter the book drew up for itself: refute other people, and complete the text of the scriptures. It takes itself to be conducting a debate, and it takes itself to be saying on the scripture's behalf what the scripture did not finish saying. To measure it by the standards of debate is not to demand too much of it; it is to take it at its own word.

A few sentence-forms, one root treatise. The frame of the main text is exposition of the Jñānaprasthāna. Kātyāyanīputra's root treatise sets down its topics one by one, and the Mahāvibhāṣā passes judgment on them one by one, working through every adjacent question in the process—expounding the Jñānaprasthāna and ranging wide, as the phrase goes. Besides the Jñānaprasthāna there are six further treatises known together as the six "feet," and for this chapter it is enough to say this much of them: what is expounded is the root treatise, and the six "feet" are the household stock of the same school. What stretches the bulk out is a handful of formulas, and their division of labour can be read off the counts. Question and answer are the foundation, and "why is this treatise made" is the standing opening question. "Further," 3,625 occurrences, is a further answer to the same question set alongside the first; the answers are not ranked in sequence and do not displace one another. "Some say," 3,313 occurrences, is the door through which another position comes on. "Likewise," 1,626 occurrences, is the book's real multiplier: the same structure is run again under another name, and the bulk multiplies with the number of runs. Its densest fascicle is fascicle ninety-two, with fifty occurrences in one fascicle. That fascicle works out item by item how many latent defilements are severed and how many fetters exhausted at the time thorough knowledge is attained of the consciousness that takes the eye faculty as its object, and then stamps its way down a roster: the ear, nose, tongue and body faculties likewise; the male faculty and the faculties of suffering and displeasure likewise; the mind faculty, the faculties of pleasure, joy and equanimity, the five faculties beginning with faith and the three undefiled faculties likewise.¹ One case worked through, one roster stamped: that is the commonest page in the two hundred fascicles. "Why is that?", 785 occurrences, is the self-posed demand for a reason. "…And so on at length," 696 occurrences, writes the omission out in the open, which amounts to a note on the page saying: expand here according to the formula. "A tetralemma should be made," 312 occurrences, draws a two-by-two grid first and fills dharmas into it after. Fascicle five asks whether there is anyone who possesses the highest worldly dharma and does not possess the possession of disjunction from it, and answers that a tetralemma should be made: possessing this and not that, possessing that and not this, possessing both, possessing neither—a kind of person assigned to each of the four cells.¹ Multiply the grid by the stock of dharmas and there is another section. "The evaluation says," 289 occurrences, closes off the several positions; Chapters 2 and 3 take it up head-on. The survey table in Chapter 0 lists precisely this set of operators, and the sentence this chapter has to add is that apart from these the book has almost no other machinery.

An equal-length comparison splits the bulk into two variables. Length can come from either of two sources: many positions, or a format doing multiplication. The way to tell them apart is to measure repetition. This book cut two vernacular translations made by one and the same translator to equal length and compared them item by item.²³ Repetition rate in a twelve-character window: the Vibhāṣā 15.9 percent, the Kośa 1.8 percent. In a twenty-character window: 7.4 percent against 0.4 percent. General compression ratio: 0.270 against 0.346. And the Vibhāṣā in fact draws on fewer distinct Chinese characters, 1,984 against 2,109. At equal length, the density with which the Vibhāṣā repeats a twelve-character string is about nine times the Kośa's, and for twenty-character strings the factor is around eighteen. The classical side runs the same way: 1,540,000 characters over two hundred fascicles, 2,872 distinct characters, a twenty-character-window repetition rate of 8.4 percent. The Kośa hardly ever says a sentence twice. A digest and a great commentary are different genres and are entitled to differ; but a difference of nine times and eighteen times is more than genre can carry. Only one explanation is left: the format is doing multiplication. The number of positions will not hold up bulk on this scale. Only a stock of dharmas multiplied out to the last cell will hold it up.

The multiplication is deliberate. Three pieces of evidence. "…And so on at length" is not a vague ellipsis but an explicit instruction to expand: where multiplication is called for, the treatise knows that it is multiplying. "A tetralemma should be made" draws the grid first and fills in the dharmas after, and the grid is drawn upon the stock of dharmas—as many dharmas as there are, so many cells. The heaviest piece of evidence is the charter sentence itself: that sūtra is the root on which this treatise rests, and what it left unsaid should now be said. Whatever the scripture did not say in full, the treatise supplies; and there is always something the scripture did not say in full, so the treatise can never be finished. For a book that has defined itself as a procedure for completing another book, running to two hundred fascicles is not a loss of control. It is the charter operating normally.

The identity of the Vaibhāṣika is written into the signature. Beneath the title at the head of the first fascicle stands a line: composed by five hundred great arhats and others.¹ That line compresses the legend of a council into a signature—the tradition being that in the time of King Kaniṣka five hundred arhats assembled this treatise together. The record of the legend and the assessment of its historicity belong to Chapter 2 of Part 3, in Part I; this part does not retell the history and takes one thing from it only, namely that this line of signature sets the tone of every ruling in the book. "The evaluation says" is not the opinion of some one master; it is a collective adjudication. From here on "Vaibhāṣika" becomes an identity: whoever holds the orthodoxy of this treatise is a Vaibhāṣika. That a treatise should manufacture a school's name for itself is a rare thing in the sectarian literature. Chapter 3 will ask after the relation between adjudication and argument, and the root of that question is buried here: a ruling spoken in the tone of a resolution requires, from the outset, a kind of support different from argument.

It is handed on to Chapter 2. Where the bulk comes from is now clear: a small set of formats multiplied over an enormous stock of dharmas. The next chapter looks at the actual state of the argument beneath those formats. Two rules of reading are fixed from here on. Where there is repetition, first identify which operator is running. Where there is a ruling, first ask what holds it up. The eight chapters that follow cannot do without either.

References¶