Chapter 2: Argument, fine classification and coarse reasoning¶

This chapter does not evaluate the doctrines. It examines how the book reaches them. The central claim can be stated in one sentence. Its powers of classification and argument are on different scales, and the gap has a regular cause. The same distinctions that serve as precision instruments in classification never enter the work of argument. This conclusion does not mean that the book cannot think. The passage on memory in fascicle 12, the discussion of omniscience and free will in fascicle 11, and the dilemma concerning atoms in fascicle 132 show it at its best. Its weakness lies in argument, not intelligence. The following sixteen sections sample sixteen ways of proceeding. Each identifies its examples by fascicle so that they can be checked.

Section 1: First, the criteria. We take the work's own criteria. Chapter 1 established that it understands itself to be conducting debate. Its declaration that it aims to refute other schools and clarify its own position occurs more than a hundred times. To measure it by the standards of argument is to take it at its word. We begin with its most unambiguous declaration, in fascicle 16. "These six causes, however, are not taught in the sūtras. The sūtras teach only four conditions. We now wish to distinguish the conditions by means of causes, and therefore teach these six causes."¹ Read this before the seventy-five occurrences of the charter formula from Chapter 1. The charter says that the sūtras left things incomplete and the treatise will supply them. This sentence says that the scheme is absent from the sūtras altogether. Having set a high standard, we need two positive examples to calibrate it, lest the next fifteen sections read as complaints. In fascicle 11, the question and answer on omniscience and free will both reason. In fascicle 17, the refutation of an analogy addresses the relation of support on which the analogy depends, and does so precisely.¹ The main set of positive examples belongs to Section 14. These simply establish the standard.

Section 2: Two levels of reductio. Reductio works at two levels. One rests on common sense, which everyone must accept. The other rests on the school's own doctrine and is effective only for those who believe that doctrine. The arguments in fascicle 10 mix the two. If the opponent stands outside the teaching, the second level cannot take a single step.¹ Mixing them is not itself a fault. Failing to distinguish them is. Nowhere does the book pause to ask for whom the conclusion of a particular reductio is absurd.

Section 3: The inertia of the table. Once a grid is drawn, it must be filled. Fascicle 13 assigns an example to each of six kinds of "increase through such-and-such a cause," and all six examples are isolated cases. In the same fascicle, one "why" receives eight successive "what" answers. The female and male faculties each occupy a cell in the table. The cells exist before any reason is given.¹ The leading example is in fascicle 30. Nārāyaṇa's strength is arranged on a ladder of nine successively multiplied levels. Every level has a name; none has a source.¹ The three lists of existence in fascicle 9, set beside one another without adjudication and discussed in detail in Chapter 1 of Essay 15, belong to the same family. Tables proliferate by themselves.

Section 4: False dilemmas and circular presuppositions. Fascicle 9 asks whether an intrinsic nature can know itself. Starting from the school's implicit premises, it proves the school's presuppositions and circles back to its starting point.¹ The firmest figure is this. The work contains twenty-seven cases in which it declares all the proposed positions faulty and then issues a verdict. Each has been read. In seventeen, the verdict identifies the faults and then takes back one of the branches. In five, it abandons both branches for a third option. In three, it changes course and says that each branch has its reasons. In one, it declares the opponent's sūtra text wrongly copied. In one, it makes a partial concession. Most of these verdicts are for show.

Section 5: Examples in place of analysis, attacking similes without explaining them. The gushing spring simile in fascicle 10 is never analyzed from beginning to end. The four characters 引世現喻, "citing a manifest worldly analogy," perform three contrary tasks in three fascicles. The single simile subjected to two challenges in fascicle 30 makes a fourth. Fascicle 73 explains how the sphere of mental objects gets its name by citing eight precedents in succession in place of a single analysis.¹ Examples in this book are often the argument itself, rather than aids to argument.

Section 6: The slippery slope. The expression 何妨 ("what would prevent it?") occurs eleven times across the work, with seven crowded into one passage of fascicle 10.¹ Each step of the slope matches a number to a doctrinal set. Once it reaches six, it stops. The four characters 百千無數, "hundreds, thousands, countless," stand bare at the sentence's end, with no further step of argument. The passage in fascicle 17 is a counterexample, already used for calibration in Section 1.

Section 7: Reasoning from incomplete lists and the final claim that there could be no liberation. Several possibilities are listed and eliminated one by one; the remainder is the answer. But the completeness of the list is never explained. This kind of inference recurs throughout the work. Its most forceful form is a final pronouncement. "If that were so, there could be no liberation." It is as though an unacceptable consequence automatically established the premise.

Section 8: Double standards in four stages. First, the rule is laid down. Fascicle 9 explicitly says that manifest worldly analogies cannot be used to challenge the dharmas of the noble ones, because noble and worldly dharmas differ. The same passage rejects the Mahāsāṃghikas' analogy outright. "There is no need to account for it; it is not included in the Tripiṭaka."¹ Second, the instruments are prepared. Fascicle 14 divides this prohibition into four pairs of categories, refining the tools further. Third, the rule is broken. Fascicle 12 uses manifest worldly analogies to explain the school's own position, and fascicle 13 uses worldly ridicule to settle the number of sites of the material faculties. The fourth stage is the most revealing. 依世俗說不依勝義, "stated conventionally, not ultimately," occurs only twice, in fascicles 19 and 73. Both use the two truths to get the text out of difficulty, not to analyze the problem.¹ The tools are there, but they come out only when an escape is needed.

Section 9: Three answers stacked together. The main example is in fascicle 13. Three answers stand side by side, their mutual incompatibility unnoticed, with the best answer buried in the middle. There are two supplementary examples. Fascicle 10 stacks three answers to an objection about sequence, two of which cancel each other out. Fascicle 12 gives three answers to one question about Devadatta and Yajñadatta.¹ The "Furthermore" format queues answers without ranking them. The longer the queue, the less anyone asks whether they fit together.

Section 10: "One should not demand an account," a statement that does not count. 作論者意欲爾, "the author of the treatise intended it so," occurs sixty-seven times; 不應責, "one should not demand an account," seventy-two times; and 不應難, "one should not object," thirteen times.¹ Literally, all three say to stop here. In practice, the text says them and then answers what needs answering anyway. At one point in fascicle 13, it says "One should not demand an account" and then gives more reasons than it does when it omits the phrase. This forms a precise pair with the next section. There, it really stops. Here, the mouth stops while the hand keeps working.

Section 11: Three forms of the stopping point of argument. Whenever a question touches the Buddha's perfection, the method shifts from reasoning to something else. All three forms can be found. One cites scripture as the standard, "because the sūtra is authoritative," in fascicle 12. Another declares outright that "this is unreasonable," in fascicle 11. The third appeals to how things naturally are, "such is the nature of dharmas," in fascicle 13, accompanied by "it does not violate the nature of dharmas" and "otherwise doubt would arise about everything, so one should not demand an account."¹ All three stopping points have one feature in common. They explain nothing and merely announce that the question ends here.

Section 12: "To conform to the wording." At one point in fascicle 12, a metaphysical challenge receives two parallel answers. One appeals to the two truths. The other says 為順文故, "to conform to the wording."¹ A stylistic reason and a philosophical reason occupy the same cell and are treated as having the same weight. It is the shortest section, but its stance is the most revealing.

Section 13: Elasticity between names and substances. When does a name designate a new dharma, and when is it merely another name for an existing dharma? The criterion is never stated. One passage in fascicle 12 lists twenty names but allows only fifteen substances. Yet in the case of acquisition and nonacquisition, an absence is sufficient to establish a real dharma.¹ The exchange rate between names and substances fluctuates as needed, and the fluctuation itself is never explained.

Section 14: Proportion, and what the book does well. Here is the main set of positive examples. The memory passage in fascicle 12 classifies finely and observes accurately. Forgetting is not an absence of mindfulness. It handles the respective elasticity of names and substances with a sure sense of proportion. The dilemma concerning atoms in fascicle 132 is a real dilemma, accompanied by three further challenges. The three difficulties concerning the sphere of tangibles in fascicle 73 offer another version of the same problem. Read the two fascicles together, and one can see where its judgment is sound.¹ Alongside these positive examples belongs the mechanism of making instruments and leaving them unused. The work creates a number of excellent tools for distinction. Those already named include the four kinds of answers, the two levels of discourse, four kinds of remoteness, the five kinds of existence in fascicle 9, and knowledge through instruction. The term 教智, "knowledge through instruction," occurs only once in the entire work, then is discarded.¹ This belongs here rather than in Section 8 because the two are different. Double standards mean laying down a rule and then breaking it oneself. Making instruments without using them means creating good tools that never come into play again. The latter is the mechanism behind this chapter's title. The same distinctions that serve as precision instruments in classification never enter the work of argument.

Section 15: Placing evidence beyond reach. There are two routes. One is a lost sūtra. Fascicle 16 says that the Ekottarika Āgama originally proceeded from one dharma up to a hundred, but now only one through ten remain. Through the power of knowledge born of resolve, the venerable Kātyāyanīputra saw the passage in the sūtra that taught the six causes and composed the treatise on that basis. This is the work's sole use of such knowledge to authenticate the source of a lost sūtra passage.¹ The other route is a past life. Fascicle 18 offers three past-life stories in succession as evidence.¹ Both make refutation impossible in principle through the very form of the evidence offered. The evidence therefore supplies no support.

Section 16: Ontology first, argument afterward. This brings the chapter to a close. Fascicle 30 supplies the axiom. "Because the intrinsic natures of all dharmas are fixed." This is the sole explicit statement of that premise in the work. The same fascicle treats superiority and inferiority as fixed qualities that can be extended uniformly, rewriting relations as components.¹ Fascicle 158 shows what the axiom looks like when worked through to the end. Three questions about infinite regress in abandoning defilements receive one answer. Acquisition and nonacquisition are given eleven differences. Finally, "acquisition of acquisition" is established. "In a single moment there are only three dharmas: that dharma, acquisition, and acquisition of acquisition. Because they acquire one another, there is no infinite regress."¹ The structure has four stages: an axiom, an inference, a cost, and the cost renamed as doctrine. The three difficulties concerning the sphere of tangibles in fascicle 73 and Śāriputra's symmetrical six-times-two derivation belong here too. First arrange the grid evenly, then deal with the problems that emerge inside it. This section is the first beam supporting the verdict in Chapter 9.

Drawing the threads together. The sixteen sections make one point. The book's apparatus of reasoning is far weaker than its apparatus of classification, and the weakness has a recognizable shape. It does not lack discernment. It lacks a passage through which discernment can enter argument. What blocks that passage? Chapter 3 gives half the answer. Wherever argument cannot reach the end, something easier is available.

On to Chapter 3. With the actual workings of argument laid out, the question naturally turns to verdicts. When argument cannot settle a matter, what does?

References¶