Chapter 2　The warrant for the definitions¶

The previous chapter stated the definitions. This chapter checks where they came from. The third class in Chapter 1's boundary clause says that a reductio is licensed by using the opponent's words as the opponent uses them. For terms already defined, this becomes a procedure: each definition must return, word for word, to a text the opponent uses. This chapter supplies the warrant in three stages and adds a comparison with a modern account. It then states a formal conditional fallback for opponents who refuse the definitions, and finally records the cost of the last retreat, refusing logic itself.

Warrant 1: the three conditions for intrinsic nature. The "Examination of Existence and Nonexistence" opens with the opponent's question.¹ "Dharmas each have their own nature, since they have power and function. A pot, for example, has a pot's nature, and cloth has cloth's nature. This nature emerges when the various conditions combine." The question brings two pieces of information with it. The opponent identifies nature by power and function and claims that nature can emerge from a combination of conditions. Nāgārjuna's answer immediately fixes the third condition: "That nature should exist within the various conditions cannot be so. A nature that emerges from conditions is thereby called a made dharma." If nature can be made, it fails to qualify as nature. Qingmu's commentary adds a diagnostic analogy and a set of contrasts: "Gold mixed with copper is not true gold"; "If nature were determinate, it should not depend on something else to emerge. It is unlike long and short, this and that, which have no determinate nature and therefore exist in dependence on another." The sources for all three conditions are now in place. "Not produced by conditions" comes from "A nature that emerges from conditions is thereby called a made dharma." "Not constituted in dependence on other dharmas" occurs word for word in the verse. "Not changing with conditions" comes from the verse beginning "If a dharma really had intrinsic nature, it should not subsequently differ." The contrasts deserve a separate entry. The commentator classifies mutually dependent properties such as long and short, this and that, under "having no determinate nature, they exist in dependence on another." The opponent's own inventory likewise lists them separately, as we will see below. That relational properties do not rank as ontological entities is one of the rare points of agreement between the two sides. Chapter 4's relational premise will begin from this agreement.

Warrant 2: the Sarvāstivāda criterion for real existence. The settled formulation appears in fascicle 22 of the Abhidharmakośabhāṣya.² "If cognition of it disappears when it is broken, or likewise when its other constituents are removed by discerning analysis, it is conventional, like a pot or water. What differs from this is called ultimate." Essay 15, Chapter 1 analysed the two procedures, breaking and analysing, and how their results are read. Here we take only the two points needed for warrant. First, this is the opponent's established statement of the criterion, not a formulation drafted on the opponent's behalf by this book. Second, the list of five kinds of existence in fascicle 9 of the Mahāvibhāṣā defines the entry for real existence in these words: "Real existence means that all dharmas each abide in their own intrinsic nature."³ Real existence is defined through intrinsic nature. This is the most important warrant in the essay. It is the opponent's dictionary, not our D3, that ties E3 to D2. The combination D3 addresses was tied together by the opponent's own hands. The same inventory separately lists relative existence: "such things as this shore and the other shore, long and short." It agrees with the category in Qingmu's commentary discussed above.

Warrant 3: Saṃghabhadra's criterion for real existence. In fascicle 50 of the Nyāyānusāra, Saṃghabhadra introduces another identifying mark.⁴ "Giving rise to cognition as its object is the mark of true existence." Where cognition of it arises without dependence on anything else, there is real existence; where cognition of it arises in dependence on something else, there is nominal existence. Two points provide warrant. This mark comes from the opposing camp's final systematic defence of itself. It was established to preserve the inventory of dharmas of the three times, acquisition, and the unconditioned; Essay 17, Chapter 7 explains its history. Its purpose differs from that of the test of decomposition and analysis, and their extensions need not coincide, confirming the entry in Chapter 1. The proof here needs only one thing to hold: that the opponent acknowledge some x as really existent by any criterion the opponent accepts.

Comparison with a modern account. Westerhoff's chapter on intrinsic nature divides modern interpretations into two dimensions, the ontological (pp. 20–45) and the cognitive (pp. 46–52).⁵ All three conditions in D2 belong to the ontological dimension. Saṃghabhadra's identifying mark, which appeals to cognition, lies alongside the cognitive dimension. Each of our definitions has a place in this modern classification. The comparison has just one purpose: to establish that these definitions are not private inventions within Buddhism. Modern research describing the same concept cuts at the same joints. Detailed comparisons will be supplied in later chapters as each item is used.

Conditional weakening, formally stated. If an opponent refuses these definitions, this essay does not debate definitions with that opponent. The proposition as a whole becomes conditional. For any combination of predicates R and N, provided R serves as real existence and N satisfies the three-condition schema, all three versions of T continue to apply to every proposition of the form ∃x[R(x) ∧ N(x)]. Someone who refuses has only two ways out. Offer a weaker N, allowing dependence on another and change with conditions, and it is outside D3's scope. We have no dispute with it; the cost is that what the opponent holds is no longer an ontological entity in the traditional sense. Offer an equally strong N, and the schema and proof still apply. This closes the dispute over definitions. The essay addresses the schema of three conditions. Anyone who accepts that schema falls within the scope, whatever new name is given to it.

The last retreat: refusing logic. One route remains, to reject the law of noncontradiction and reductio together. This essay calls it overturning the table. Once the table is overturned, nothing on it remains to be lost. We do not debate this route either. Every step in arguing with someone who refuses argument uses something that person rejects. There is nothing available for refutation; we merely record the cost. First, someone who refuses the law of noncontradiction can no longer say what they maintain. If an assertion no longer excludes its negation, a creed loses its content. The most solemn affirmation becomes indistinguishable from its desecration. Second, the law of noncontradiction is the skeleton of every doctrinal system. Distinguishing orthodoxy from heresy and doctrine from error, and deriving consequences from doctrine, all use contradiction as their boundary. Remove the boundary, and the first thing to fail is doctrinal inquiry itself, not this essay's theorem. Third, abandoning logic invalidates all one's own proofs at the same time. Every argument offered for an ontological entity or a deity relies on logic. One cannot abolish logic when challenged and reinstate it when offering evidence. To switch it on at selected times would require a meta-rule determining when logic is on duty, and that meta-rule would perform precisely the job logic performed before. Fourth, mature theological traditions themselves refuse this route. Omnipotence has always been interpreted as excluding the production of contradictions, because a contradiction is not something that could be done but has not been done. What this retreat buys is withdrawal from the case, not victory. It preserves the wording of the sentence at the cost of its no longer being a claim. This differs from the dialetheist reading discussed in Chapter 7, which is a logical proposal still governed by norms and remains a participant in reasoned debate. Theology is precisely the field most often expected to take this door. Chapter 8 will return to it when addressing eternity and creation.

What this chapter delivers. Five documents: the original sources for the three conditions; the opponent's dictionary, in which real existence and intrinsic nature define one another; a record of the two criteria's distinct, nonoverlapping roles; conditional weakening in schematic form; and an inventory of the costs of abandoning logic. From Chapter 3 onward, the tools enter the scene.

What follows is a plain-language restatement of this chapter. The rule is the same as in Chapter 1: change only the language and add no content. Where the two diverge, the upper half governs.

This chapter's work resembles checking a delivery. The three conditions for intrinsic nature and two criteria for real existence listed in the previous chapter are returned to their sources one by one, to see whether they really belong to the opponent.

First, the chapter in the Mūlamadhyamakakārikā. The opponent speaks first: all things have their own nature; a pot has a pot's nature and cloth has cloth's nature, which emerges when the conditions come together. Nāgārjuna's first reply closes that road. What emerges from conditions is made, and what is made does not qualify as nature. The commentator adds an analogy: gold mixed with copper is not true gold. Then comes a set of contrasts. Long and short, this and that, are things everyone accepts as established by comparison, existing only in contrast with one another. Nobody grants them the status of ontological entities. The three conditions for intrinsic nature come from here: not produced, not constituted in dependence on something else, and not changing with conditions. The first two are almost the exact words of the verse and commentary; the third comes from the same chapter's line "If a dharma really had intrinsic nature, it should not subsequently differ."

Now the criterion for real existence. This is the test the Sarvāstivāda wrote into the Abhidharmakośabhāṣya. Smash something; if cognition of it disappears too, its existence is merely conventional. Leave it unbroken but mentally remove its constituents one by one; if the cognition disappears, it is likewise only conventionally existent. Only what survives both procedures counts as ultimately existent. Still more important is the inventory in the Mahāvibhāṣā. Its definition of real existence is there in black and white: all dharmas each abide in their own intrinsic nature. What is solid existence? Each thing's holding to its own nature. In other words, the opponent tied real existence and intrinsic nature together. The previous chapter addresses that very combination, not a pair we forced together to make up the numbers.

Saṃghabhadra's identifying mark takes another route. Can the thing independently serve as the object of cognition? The two criteria do different jobs and need not return the same judgment. That is why the previous chapter says we need only the opponent's acknowledgment of such a thing by a criterion the opponent accepts.

Modern research makes the same division. Westerhoff distinguishes an existence side and a cognition side in readings of intrinsic nature. Our three conditions fall on the former side; Saṃghabhadra's mark adjoins the latter. These definitions are no private Buddhist merchandise. The concept takes this shape in anyone's hands.

What if the opponent refuses these definitions? The reply is that we will not quarrel. Make your definition of an ontological entity softer, allowing it to depend on others and to change, and it is outside our scope. We have nothing to say against it. The price is that what you hold is no longer an ontological entity in the traditional sense. Make the definition just as hard, and renaming it does not take it outside the scope. We address the structure of those three conditions. Anyone who accepts that structure is included.

One last road remains: overturn the table and refuse logic itself. We do not take this up either. In reasoning with someone who refuses reason, every step appeals to something that person rejects. Refutation is impossible and unnecessary. But the price must be calculated. Someone who refuses the law of noncontradiction can no longer say clearly what they believe. Saying that something exists no longer excludes saying that it does not, so the most solemn affirmation becomes indistinguishable from its desecration. What holds a system of doctrine together? The possibility of distinguishing orthodoxy from heresy and deriving consequences from a creed. All of it rests on the law of noncontradiction. Overturn the table, and it is doctrine that collapses first, not our theorem. Anyone who overturns the table invalidates their own proofs at the same time. One cannot declare logic invalid under attack and invite it back when offering evidence. This road buys withdrawal; victory is out of the question. Besides, seasoned theologians themselves do not take it. Omnipotence has always excluded making a contradiction, because a contradiction is not something to be made at all. When Chapter 8 turns to eternity and creation, we will stand at this door again.

The next chapter prepares the tools: two kinds of negation and a modern reading of the tetralemma.

References¶