Chapter 7: Entities and regress¶

This chapter settles the account. The first six chapters recorded, place by place, what Sarvāstivāda puts in. This chapter considers the entries together. The table in Essay 14, Chapter 7 has lacked one column throughout: the problems each addition itself creates. That column is filled only now, because it takes all seven chapters to see whether the additions are compatible with one another. Three tasks remain afterward: gather the deficiencies recorded in the first six chapters into a chain and explain why it cannot stop; bring Sarvāstivāda's several ways of handling regress together to see whether they use the same procedure; and finally compare the verdict in Essay 14, giving this book's judgment on the Abhidharma route of analysis. Saṅghabhadra's response comes after the chain has been drawn: we must first see the shape of the problem to know whether he is answering this problem at all.

First fill the third column.

Not one row of the third column is empty. More tellingly, not one problem in it was devised for Sarvāstivāda by an outsider. Acquisition of acquisition appears in the questions and answers of fascicle 158; arising of arising, in those of fascicle 39; whether activity and substance are identical or different, in those of fascicle 76; the boundary of intermediate existence, in the Kośa's restatement. Sarvāstivāda knows that each addition brings a new question. It records the questions unchanged in its textbook, then supplies an account of how to stop at each point. This chapter asks what it uses each time it stops, not whether it saw the problem.

The final row deserves a sentence of its own, because its shape differs from the others. The three unconditioned dharmas do not fall within the three times, and so have no activity. Yet Sarvāstivāda's main grounds for real existence are two: not dissolving under analysis and having activity. The unconditioned does not dissolve under analysis, but the ground of activity is wholly inapplicable. Sarvāstivāda classifies the unconditioned as not included in time, thereby admitting that this compartment obtains its real existence on another set of grounds. Chapter 2 noted that the Southern tradition has only one unconditioned dhamma, nibbāna. Establishing two fewer unconditioned dharmas means giving two fewer formal things the status of real dharmas. This chapter adds that the extra two occupy not only two more compartments, but bring the criterion another exception.

One more row belongs in the table, and it does not come from Essay 14. The Southern tradition has not left the lower limit of matter empty. It has material groups, the pure octad group, and an account of material groups arising from the four causes of karma, mind, temperature, and nutriment. In the same place, Sarvāstivāda puts atoms, eight coarising constituents, and aggregation. This row must be added because Chapter 6 supplies the cleanest sample. Atoms are invisible, intangible, unable to exist alone, and do not touch one another but are merely adjacent without an interval. All four are stipulations; none is inferred from elsewhere. And each added stipulation requires another to fill the hole left by the last. If atoms do not touch, they should scatter when struck, so holding together by the wind element is added. If they do not touch, striking should not produce sound, so arising without an interval as the very cause of sound is added. Stipulation follows stipulation, patched all the way to wind and sound. The other rows have the same shape, only less clearly visible.

These deficiencies come together in three chains, and all three begin with the same act.

The first begins with "all exists." Whatever is a dharma exists in all three times. If it exists in all three, yet the three must differ, activity alone can draw the distinction. Activity thus becomes the load-bearing wall of the whole system. The wall's own status is immediately questioned: is it a dharma? If so, it too must have three times, raising the question of the activity of activity. If not, something absent from the list draws the distinction among the three times. Chapter 1 followed this step to the treatise author's statement: it cannot definitely be said to be identical or different, and therefore one should not press the objection.

The second begins with not-self. The person does not really exist: Sarvāstivāda guards this rigorously. Yet discourse about persons must continue. This dharma belongs to this person; this person is still alive; this person belongs to this kind; this person still counts as keeping the precepts when asleep after receiving them; this person has passed from this life to the next. Five statements, five dharmas: acquisition, the life faculty, group homogeneity, nonmanifest material form, and intermediate existence. Chapter 4 already pointed out that of the six additions, only the four characteristics concern dharmas alone; the other five all point to persons. Once acquisition is established, it too must be possessed, giving rise to acquisition of acquisition.

The third begins with momentary cessation. Dharmas arise and cease moment by moment; these two ends, arising and ceasing, need bearers, so the four characteristics are established. The four characteristics are conditioned dharmas, and each conditioned dharma has four characteristics, so the four secondary characteristics are established. Once the second level exists, one must still answer why it does not produce a third.

Figure terminology and omitted wording. The opening statement means “Something must be made intelligible.” “All three times exist” means that dharmas exist in all three times. The four secondary characteristics are arising of arising (生生), abiding of abiding (住住), alteration of alteration (异异), and cessation of cessation (灭灭). “Current possession” means acquisition simultaneous with the dharma (法俱得); “prior possession” means acquisition before the dharma (法前得). “Original nature” means inherent nature (本性). “Function” means activity (作用), and “critics” refers to the treatise’s adjudicators (评家). The note concerns the four Chinese characters 三世实有, “real existence in the three times.”

The three chains seem to run separately, but share their final stretch. The four secondary characteristics must act; acquisition must act; every newly added dharma, as long as it is still conditioned, must act. And the distinctions among the three times are established precisely through activity. Each added dharma thus hangs another object on the same load-bearing wall. Chapter 3 said that the opening lies at the center of the system; Chapter 4 saw it leaking down through each newly added dharma. Here too is the answer to why the category of formations dissociated from mind is the hardest to handle. All fourteen dharmas in it are formal entries, and formal entries have the greatest difficulty explaining what they do. Yet this system requires every dharma to have activity. Earlier scholarship has examined one by one the reasons for establishing the dharmas of this category and the formation of judgments of their real existence.¹

Four things in this chain were not made by Sarvāstivāda itself. Chapter 2 recorded that the table of five categories and seventy-five dharmas does not appear in the Mahāvibhāṣā; Chapter 3, that the four Chinese characters 三世实有, "real existence in the three times," never occur in its 200 fascicles; Chapter 4, that the terms "acquisition preceding the dharma" and "acquisition accompanying the dharma" never occur in the Vibhāṣā, while "acquisition following the dharma" occurs only once. The fourth requires another step back: even "five categories" (五位) is not the Kośa's own wording. In fascicle 4, Vasubandhu says "five classes" (五品): "All dharmas, in brief, have five classes: first, matter; second, mind; third, mental factors; fourth, formations dissociated from mind; fifth, the unconditioned." The two Chinese characters 五位 occur just once in thirty fascicles, in fascicle 15, referring to the position of intermediate existence and the positions within and outside the womb, unrelated to this list. The number seventy-five is never written out. On the Vibhāṣā side, the four Chinese characters 七十五法, "seventy-five dharmas," likewise have a count of zero.^{2 3} Here, then, is a peculiar fact: "five categories and seventy-five dharmas," the designation nearly every Chinese introduction to Buddhism uses for Sarvāstivāda, is stated in neither of the two primary works. The four examples together make a rather embarrassing picture. Much of the vocabulary for which this school is best known was organized for it by its critic, while the most famous item of all was not even stated by that critic. Vasubandhu arranged the table, named the claim, supplied the three designations, and then criticized them one by one. This directly concerns the present chapter. Every link in the chain drawn here has a clear source, but the chain itself was straightened out by later hands. The Vibhāṣā is a book that presents four masters' views side by side and raises objections everywhere. It does not arrange its own claims into a line running from beginning to end. This book holds that such a line can be drawn, because both ends of every link have original passages in the Vibhāṣā's questions and answers. But there is insufficient evidence to say that the Vibhāṣā was consciously following that line. This must be stated at the outset.

All three chains begin with the same sentence: something must make sense. Chapter 4 read a second ground for Sarvāstivāda's judgments of real existence out of its practice. If an entry cannot be omitted from a sentence that must hold true, it is established as a really existing dharma. Acquisition cannot be omitted, because otherwise the two Chinese characters 成就, "possession," in the sūtras have nowhere to land. Group homogeneity cannot be omitted, because otherwise nothing makes a class of beings one class. Nonmanifest material form cannot be omitted, because otherwise a sleeping person would not count as keeping the precepts. This ground converts linguistic indispensability directly into ontological real existence, at an exchange rate of one to one.

Sarvāstivāda never writes this criterion down as a criterion. Once written, it would have to be tested, and it cannot withstand two tests. First, it has no stopping condition. There can be as many dharmas as sentences, and Abhidharma sentences only multiply. Second, it does not apply to itself. Dissolving or not dissolving under analysis is the first criterion; being dispensable or indispensable in a sentence is the second. The two do not judge the same thing. Chapter 2 already pointed out that what is called one criterion bars nineteen entries from material dharmas yet invites fourteen into formations dissociated from mind; the two sides are actually using different reasons. Here lies the greatest convenience of an unwritten criterion: it need not answer for its two uses.

The chain ends in a word, and the word is empty. Chapter 1 recorded a statement in fascicle 76. Asked what constitutes the intrinsic nature of the three times, it answers: all conditioned dharmas. It then adds: "What is said of intrinsic nature applies likewise to self, entity, own-substance, characteristic, part, and inherent nature." Seven terms are declared interchangeable. If a word used as the final explanatory term can be freely exchanged for six others, it marks not the endpoint of explanation but the place where explanation stops. Chapter 3 adds another layer. In fascicle 76's own checks by aggregate, past and future matter has no alteration or resistance, feeling does not experience, and consciousness does not cognize. The substance that always exists is what remains of a dharma after everything by which it is defined has been subtracted. At the chain's end, what supports it is neither a more fundamental fact nor a more basic principle. It is the same word renamed six times.

Regress appears three times in this system. Sarvāstivāda stops it each time, but not in the same way.

The first is acquisition of acquisition. Fascicle 158 first records one answer: one should say that acquisition has a further acquisition; even infinity would involve no fault, and this is why saṃsāra is difficult to end and cross. The answer adopted is another. In one moment there are only three dharmas: the dharma in question, acquisition, and acquisition of acquisition. Through acquisition, one possesses the dharma and acquisition of acquisition; through acquisition of acquisition, one possesses acquisition. The two acquire each other.³ The chain folds into a loop. Instead of producing a third level, the second turns back and governs the first.

The second is the characteristic of a characteristic. Fascicle 39 likewise first records the statement that admitting infinity involves no fault: the three times are vast, so could there be nowhere for it to abide? It then settles on nine dharmas arising together. Arising can produce the other eight dharmas; arising of arising produces only one, primary arising.³ The stopping method differs from the first. Here the chain does not fold back; the rule changes at the second level. The question follows: if both are called arising, why can one produce eight things and the other only one? The answer is that such is the nature of dharmas, and one should not raise an objection.

The third is the activity of activity, and this time the Vibhāṣā does not stop it; it refuses to answer. Pressed in fascicle 76 on whether activity and substance are identical or different, it answers that they cannot definitely be said to be identical or different. Its reason is an analogy, ending with "therefore one should not press the objection."³ It gives no stopping method, but declares that the question should not be asked. Of the three cases, the first two limit losses; the third shuts the mouth.

The three stopping methods can be weighed side by side. Folding into a loop is the most economical. Its price is that the first and second levels become conditions of each other's establishment, and neither can stand alone. Changing the rule at the second level comes next. Its price is the need for another reason for the asymmetry, and the reason found is "such is the nature of dharmas." Refusing to answer takes the least trouble and costs the most, because the rejected question falls precisely on the whole system's load-bearing wall. Chapter 2 recorded the shape of one closure: an analogy followed by a statement that no further question is needed. Chapter 4 recorded its third appearance: the three successive analogies in fascicle 16's discussion of the associated cause (two bundles of reeds standing by leaning on each other, many ropes pulling a large log together, and many people joining hands to cross a river), followed by the statement that conditioned dharmas are weak by nature and can accomplish their work only by depending on one another. The shape is identical all three times: give an everyday analogy, then declare a halt. An analogy can make a claim acceptable; it cannot establish it.

Here is a number that can be checked. "Infinity" (無窮) occurs 47 times in the 200 fascicles of the Mahāvibhāṣā. Twenty-four are concentrated in two fascicles: seven in fascicle 39 and seventeen in fascicle 158, precisely the discussions of the four characteristics and acquisition.³ The remaining 23 are scattered through other fascicles, almost always used in reverse: if some claim holds, such-and-such a succession follows to infinity, making liberation and escape impossible; hence the claim fails. In the Vibhāṣā's hands, infinite regress is chiefly a weapon for refuting others. Only at these two points in its own system is it turned around into something acceptable, both times with an added reason why it may be accepted: saṃsāra is difficult to end and cross; the three times are vast, so could there be nowhere to abide? A tool for refuting others fails in two places, both of them the school's own roof beams. This is no coincidence.

There is another ready-made set of formulas for ending questions, and their distribution too can be checked. The three Chinese characters 不应责, "one should not press the objection," occur 72 times in the Vibhāṣā, not once in the thirty fascicles of the Kośa, and seven times in the eighty fascicles of the Nyāyānusāra. The four Chinese characters 意欲尔故, "because that is what was intended," occur 76 times in the Vibhāṣā and zero times in the other two works.^{3 2 4} One layer must first be subtracted. "Because that is what was intended" mostly answers why the root treatise's entries are arranged as they are. The answer is that this was the author's intention, and as long as it does not violate the characteristics of dharmas, one should not press the objection. Such language belongs to the commentary genre. The Kośa is not a commentary and has no use for it, so this cannot count as a difference. What remains after subtracting this layer is still considerable. To judge a question one that should not be raised, the Vibhāṣā has a habitual phrase that Vasubandhu never uses. His not using it does not mean he has no stopping line. He either draws the line elsewhere or simply leaves this question open. Chapter 3 noted that Vasubandhu closes with another sentence: "The nature of dharmas is profound, beyond the domain of discursive thought."² Both shut the mouth: one says one should not ask, the other that questioning cannot reach there.

There is a third stopping method that need not even use the three Chinese characters 不该问, "one should not ask." The first two are that one should not ask and that questioning cannot reach. The third uses sectarian allegiance itself as a reason. This is hard to count in the Vibhāṣā, which is a work of school doctrine and speaks of its own school throughout. But it leaves clear traces in the Kośa: whenever Vasubandhu restates a Vaibhāṣika ruling, he copies its sign-off unchanged. A search of all thirty fascicles yields seven such places.²

Set in a column, the steps are clear. Fascicle 4 gives two reasons side by side: "because it does not contradict reason" and "because it is what our school upholds." Chapter 4 already pointed out that they carry different kinds of weight, yet are used there as though the weight were the same. In the two places in fascicle 5, the first reason disappears, leaving only the four Chinese characters 违我宗故, "because it contradicts our school." A claim is rejected, and the entire reason given is its failure to agree with the school. The passage on the four characteristics in fascicle 5 rises to a declaration, elegantly put: should one abandon one's own school because many people raise objections? Should one not sow barley for fear that deer will ruin it, or not eat a fine ball of food for fear of flies? Therefore, when meeting objections, one should diligently reconcile and explain them, and follow and practice the meaning of one's own school. Fascicle 13 first calls a halt, then declares the school. Just before "let these many disputes cease for now," Vasubandhu has spent an entire passage laying out the difficulties of nonmanifest material form.

The summit is fascicle 28: "Although there is this reasoning, it is not what our school upholds." The opponent's account makes sense; this is admitted. Yet even after the admission, it is not adopted, because it does not belong to the school. The earlier places can still be read as omitting the argument. This one cannot. It explicitly separates making sense from being accepted, then says the two need not coincide.

The matter must not be told from only one side. A work of school doctrine is constituted by establishing a position; marking out one's own school and others is part of its job. Reading every "because it is what our school upholds" as argumentative defeat would be unfair. Moreover, several of these seven places do give arguments beforehand; the sentence is merely a sign-off. This book reads only one thing from the seven places: the sign-off's position is highly regular. It always falls where the argument has just been blocked: after "both paths are well stated," after an account of nominal designation has just run smoothly, after "many disputes," after "although there is this reasoning." The sign-off bears no weight of its own, yet each time it stamps exactly the compartment bearing the greatest weight.

Chapter 3 stopped on the Vibhāṣā's last page, and that page asked only half the question. Fascicle 200 assigns concealment and manifestation to the eternalist views of non-Buddhists. Its criterion is that characteristic and substance are one: if characteristics differ, substance should differ too. If a dharma is said to be permanent, with only its phases concealed or manifest, its characteristic must be admitted to remain unchanged throughout. This criterion works accurately against non-Buddhists. The unasked half is what happens when it is applied to the school itself. Sarvāstivāda wants precisely an unchanged substance with different characteristics: dharma-substance always exists in the three times, while activity arises and ceases. The arising and ceasing of activity is nothing other than a difference in characteristic. The Vibhāṣā never raises this question itself. Vasubandhu raises it for the school.

Saṅghabhadra takes up the question. In the chapter on latent tendencies in fascicle 52 of the Nyāyānusāra, an exchange has Vasubandhu press harder at each level and Saṅghabhadra answer level by level. It is the only passage of direct confrontation on this problem in the three treatises. Vasubandhu first asks: if a dharma is established as past, future, or present through activity, what distinguishes the three times of activity itself? Must we say that activity has another activity?⁴ This is precisely the dilemma inferred in Chapter 1, only pressed more urgently by Vasubandhu.

Saṅghabhadra's first answer dismantles the question: when did Abhidharma masters ever establish activity as past, future, or present? Asking now whether it has three times does not constitute a valid question. I have never said that activity also exists in the past and future; how can you turn around and demand whether activity has a past and future?⁴ This answer does not answer the question; it denies its legitimacy. The thing being questioned was never established as an entry that needed a place.

Vasubandhu immediately tightens the dilemma. If there is no activity in the past and future, one must admit that activity "did not exist before, now exists; having existed, no longer exists." Since activity is a distinction within a dharma, one must choose between difference and nondifference. If different, it has a separate own-substance, and formations should be like it. If not different, one must explain how something can be nondifferent from dharma-substance yet have a distinction.⁴ Here Saṅghabhadra concedes half a step: I admit that activity is a distinction within a dharma, but it cannot be said to differ from dharma-substance. Choosing nondifference invalidates the statement from Chapter 1 that it cannot definitely be said to be identical or different. The question of identity or difference that the Vibhāṣā refused to answer has been answered by Saṅghabhadra. He chooses the side of nondifference.

Vasubandhu follows up: how can there be a distinction without difference? Saṅghabhadra answers with a counterquestion, and it strikes accurately. In your own school, a wholesome mind contains distinct seed-functions induced by dharmas of other kinds, such as the unwholesome. These functions are not different from the wholesome mind, yet there is indeed a distinction. How do you explain that?⁴ He then gives several examples: feeling has the characteristic of experiencing, yet includes the distinction between pain and pleasure; in a single continuum, the dharma-characteristic of earlier and later moments is the same, yet earlier and later differ; an object burned by fire has an unchanged coarse characteristic, yet subtle differences from moment to moment. This counterquestion is effective. Every account that puts change upon an unchanging bearer owes the same thing. Sautrāntika seeds owe no less than Sarvāstivāda activity. Chapter 4 said that Vasubandhu's side entrusts persistence to functions and distinctive transformations of a continuum, all nominally designated. Saṅghabhadra's point is that nominal designation does not remove the need for an account.

Saṅghabhadra's positive answer is a triple negation. The distinctive activity of the present is not different from the dharma, because it has no separate substance; nor is it identical with the dharma, because dharma-substance can exist while activity is absent; nor can it be said not to exist, because once activity arises it can project a fruit. He accordingly establishes the verse: "A continuum has no separate substance, yet is allowed distinct work. The principle of activity is the same; hence the meaning of the times is established."⁴ These three statements place activity where it neither has a separate substance nor is nothing at all. Regress is indeed stopped, because the chain's first link fails to hold: something without a separate substance need not be assigned three times. The preceding question is answered too. A difference in characteristic without a difference in substance makes sense because that difference carries no substance, whereas the criterion in fascicle 200 of the Vibhāṣā judges characteristics with substance and can go no further here. This is a clean answer. Its cost lies in the model he uses.

Notice what he takes as his analogy. The passage before the verse states it plainly: distinctive activity cannot be said to differ from the substance to which it is attached, just as with a continuum. Conditioned dharmas arising moment by moment without an interval are called a continuum. A continuum is not different from the dharmas, because it has no separate substance; nor is it identical with the dharmas, because a single moment cannot be said to have a continuum; nor can it be said not to exist, because things really are accomplished in a continuum.⁴ The three sentences have exactly the same form as the three about activity. The continuum comes first, then "the principle of activity is the same." Yet for Sarvāstivāda, a continuum is a nominal dharma. To secure real existence in the three times, Saṅghabhadra selects as his model something judged nominal.

Chapter 6 reached a statement of the same shape at the material end. An atom never acts alone; the five consciousnesses depend on aggregates and take aggregates as objects; and the distinctions among the three times are established through activity. Thus an atom can borrow its present only from the aggregate in which it lies: the real must borrow time from the nominal. The two inferences differ. One proceeds from the supports and objects of the five consciousnesses; the other examines Saṅghabhadra's own chosen analogy. Their conclusions land in the same place. This book holds that one gap is exposed at two ends, rather than two separate gaps.

This exchange has been discussed repeatedly in modern scholarship. Saṅghabhadra's responses to Vasubandhu's three objections have been fully analyzed; this chapter does not repeat that work.⁵ It adds only one point. Earlier discussions mostly stop at whether Saṅghabhadra successfully blocks the regress. This book is more concerned with the model he uses to block it. That he blocks it is a fact; that he uses a continuum as his model is also a fact. The two must be seen together to show how much the answer actually costs.

This must be put fairly. Saṅghabhadra does what the Vibhāṣā dared not do. Where the Vibhāṣā judged the question of identity or difference one that should not be pressed, he answers it. Where it left activity without a status, he gives activity a place. He also points out that the opponent owes the same step, and Vasubandhu's side has no adequate answer to this. Yet the stopping line has only moved down one compartment; it has not been removed. The triple negation itself accepts no further questions. What "not different, not identical, and cannot be said not to exist" jointly amounts to receives no continuation. Nor has Saṅghabhadra himself abandoned the formula. In answering objections to the coexistent cause in fascicle 15, he places the three Chinese characters 不应责, "one should not press the objection," directly where the question lands. In an aggregate, he says, all things are causes and all are fruits in relation to one another, "therefore one should not demand a decisive reason fixing cause and fruit."⁴ This is one of the seven Nyāyānusāra occurrences counted in the previous section. Chapter 3 called the substance of this exchange a struggle over the stopping line. This chapter can finish the statement: the struggle moves the line one compartment, while neither side changes its way of struggling.

Chapter 4 handed this chapter a question: what does Sarvāstivāda preserve and what does it relinquish along the line of not-self? First, what it preserves. Sarvāstivāda does not admit the real existence of a person, and this remains firm from beginning to end. The five dharmas it establishes for persons (acquisition, the life faculty, group homogeneity, nonmanifest material form, and intermediate existence) also arise and cease moment by moment. None is a permanent, unitary, governing own-substance. Reading these dharmas as inviting the self back would be unfair; this book rejects that reading.

What it relinquishes is something else. Not-self originally sought to abolish a position: a holder outside dharmas who bears them. Sarvāstivāda does not establish this holder, but establishes holding itself as a real dharma. Acquisition is not the person who holds; it is the relation of holding. Once the relation is made an entity, the position remains empty, yet every task once assigned to it has been taken over. Chapter 4 counted clearly: five of the six additions point to persons. What is preserved is the letter of not-self; what is relinquished is the expense that not-self originally sought to save.

Fairness requires saying this of the Southern tradition too. Essay 14, Chapter 6 reached a conclusion: continuity is not a property of mental dhammas, and no amount of detail in the list of eighty-nine consciousnesses can determine it. The Southern tradition does not establish acquisition, but continuity likewise does work in its system without a status. One school establishes it as a dharma and pays in regress; the other does not establish it and pays in what cannot be said. Neither has actually saved the expense.

The verdict begins with a small difference. Chapter 1 compared the two schools' nominal existence. Sarvāstivāda's is reached through decomposition: where decomposition can go no further, there is the real. The Southern tradition's is what cannot gain entry, meaning that it cannot be taken separately as an object of contemplation.⁶ One Chinese character, 假 ("nominal"), and two reasons, presupposing different activities. Decomposition is a speculative act that can be completed in a study; entry on the list is an act of practice to be completed on the meditation seat. Across seven chapters, this small difference has worked throughout. Wherever reality is judged through decomposition, reasoning decides where decomposition stops, yet reasoning has no built-in stopping condition. The atom therefore needs a stipulation to seal the bottom, and acquisition needs mutual acquisition to close the loop. Wherever reality is judged through entry, the stopping point is decided by contemplation's reach. However useful, what contemplation cannot reach cannot enter the table. The Southern list is therefore short; the price of shortness is that several things at work have no status. Both schools' defects come from their respective acts, not their respective intelligence.

One part of the verdict must be positive, and it must come first. Sarvāstivāda takes these three problems head-on and writes every subsequent step in words that can be quoted and refuted. Essay 14 said that the Southern system's ontologization occurs in no declared doctrine; it occurs in cataloguing procedures and the pressure of format. Sarvāstivāda is the reverse: every step is declared. This difference is decisive for the whole subsequent history of Buddhist philosophy. Only clearly stated claims can be systematically refuted, and a large part of the argumentative forms of Madhyamaka and Yogācāra is elicited by Sarvāstivāda's clearly stated claims. Without an opponent who fully states his position, Nāgārjuna's way of dismantling things one by one would have nowhere to exert its force. This book's first judgment of Sarvāstivāda is therefore: it is worth refuting. In intellectual history, that is no low appraisal.

The first six chapters have already supplied the other three points.

The first comes from Chapter 1: this verdict will not say that the Southern tradition is clever and Sarvāstivāda muddled. The Southern tradition's caution has its own cost. It says nothing about the group of problems concerning time, leaving them there. Sarvāstivāda takes them up and answers directly. It answers poorly, but at least the answers can be tested. Testability is itself a contribution.

The second comes from Chapter 3: Sarvāstivāda's stopping line is forced out by questioning, not chosen. The four Chinese characters 故不应责, "therefore one should not press the objection," say that the question should not be asked, not that it has already been answered. The tone is impatient, not composed. This chapter adds that the distributions of the 47 occurrences of "infinity" and 72 occurrences of "one should not press the objection," considered together, give this judgment numerical support. It is more than a feel for the wording.

The third belongs to this chapter: the three empty places left in Essay 14 mark the boundary of the analytical route itself, not omissions peculiar to the Southern tradition. The reason lies in the complementary shapes of the two schools' failures. To use a list, there must first be relations; to order moments, there must first be time; to establish a list, there must first be a boundary. These three are conditions that make analysis possible, not objects of analysis. The Southern tradition leaves them outside its list, so the list cannot claim to enumerate all that exists. Sarvāstivāda brings them inside, and the entries brought in turn around and demand another set of entries for themselves. The chain begins. One leaves them outside, the other takes them inside: these are two sides of one boundary, not two kinds of error.

At the end of Chapter 3, a question was asked: can one obtain both a bearer of persistence and a bearer of difference? This book's answer is that within this method, one cannot. Confine reality to the present, and connections across time can only be relations, yet there is no place to put relations. Spread reality across the three times, and the present's distinctiveness can only be activity, yet activity can be given no status. Each route loses one end, and the end it loses is precisely the end the other holds. There is no third route between them unless what changes is not the list, but how analysis itself is to be done.

Changing analysis itself is the business of later essays. The Yogācāra tradition brings the bearer inside consciousness: seeds are stored in store consciousness, and persistence is borne by the continuum of consciousness. The work of acquisition and nonmanifest material form thereby finds a place. This targets the group of difficulties in Chapter 4 of this essay. The Madhyamaka tradition removes the premise of the question. If intrinsic nature cannot be established at all, an explanatory chain ending in intrinsic nature need not be drawn to the end, because it points in the wrong direction from the beginning. This targets the group of difficulties in Chapter 1, namely the seven mutually glossing terms. Each route faces one place in this essay. Recording that before reading the two later essays will make them easier to read.

Finally, draw up a list for the later essays to use.

One class consists of targets for refutation: intrinsic nature, real existence in the three times, the division between real and nominal existence, atoms and aggregation, and the criterion that what cannot be omitted from a sentence is established as a real dharma. Madhyamaka targets the first and fifth; Yogācāra, the fourth. Both use the second as a counterexample.

Another class is retained after being reworked. Seeds and perfuming take over the work of acquisition and nonmanifest material form. The framework of six causes and four conditions survives almost unchanged in later treatises. The distinction between coexistent and homogeneous causes, and the pairing of maturation cause and fruit of maturation, continue into Yogācāra treatises. Opposing a system and retaining its tools have never been in conflict.

Another class is neither refuted nor reworked, but sinks beneath language: dharma (法), substance (体), activity (用), association (相应), object (所缘), immediate precedence (等无间), predominance (增上), coexistence (俱有), maturation (异熟), real existence (实有), and nominal existence (假有). These terms are still used every day in Chinese Buddhist discourse. Most who use them no longer remember the particular problem each was made to solve. From beginning to end, this essay has in fact done only one thing: put these familiar words back into the disputes in which they were made, and make them suspect once again.

Over to the next essay. The next essay turns to the Mahāvibhāṣā itself. This essay has tested Sarvāstivāda's claims as philosophical propositions; its conclusions are set out above. The next asks something else: how those propositions were first adjudicated into orthodoxy from a heap of parallel alternatives. Across 200 fascicles, four masters' views are presented side by side, followed by the adjudicators' rulings; the grounds for those rulings are not all arguments. First seeing how an orthodoxy was made, then how insiders dismantled some of it in the Kośa, gives a precise object for understanding what Madhyamaka later dismantles.

References¶