Does the argument undermine itself?

What you will work out

Reconstruct the dilemma the opponent of the Vigrahavyāvartanī presses against the claim that everything is empty.
Explain Nāgārjuna’s replies: that his own statement is empty and still does its work, and that debate proceeds within conventional truth.
Compare three readings of “I do not have any thesis” and say what each makes of the dilemma.
Say what the reply establishes and what it leaves open, and apply the study’s arguments and readings to a new case.

A notice about notices

Imagine a notice pinned to a community board: “All notices on this board will be taken down on Friday.” Does that include this notice? Of course it does, and nothing is wrong with that. It does its job until Friday and then comes down with the rest. A statement can include itself without contradicting itself.

Nāgārjuna’s claim that everything is empty includes itself too, and lesson 7 showed that he accepts this: emptiness is empty. His opponents saw a far more troubling consequence. If his own claim lacks a nature of its own, what gives it any power to refute someone who disagrees? This lesson follows that objection in its strongest form, in a work written to answer it.

The Dispeller of Disputes

The Vigrahavyāvartanī, or Dispeller of Disputes, is a short work of seventy verses with Nāgārjuna’s own prose commentary. Its first twenty verses state objections to his position; the rest reply to them. Westerhoff, whose translation this lesson quotes, thinks it was written after the Mūlamadhyamakakārikā, for readers who already knew his arguments and wanted their difficulties addressed.[1] Verses are cited by number, and Nāgārjuna’s commentary on a verse shares its number.

One point of translation matters from the start. Westerhoff renders svabhāva as “substance”, meaning something that does not depend on anything else. It is his word, in this translation, for what the study has called intrinsic nature. Wherever the quotations below say “substance”, read “intrinsic nature”.

The dilemma

The opponent opens with the objection:

If the substance of all things is not to be found anywhere, your assertion which is devoid of substance is not able to refute substance.

The commentary spells it out by turning one of Nāgārjuna’s own arguments against him. He had argued that a sprout is not found in its seed, in the earth and water that condition it or in their combination, and so it lacks substance. The same, the opponent says, is true of his assertion. It is produced by the chest, the throat, the lips, the tongue and the rest, and is found in none of them, so it too lacks substance. And then: “a non-existent fire cannot burn, a non-existent knife cannot cut, a non-existent water cannot moisten. In the same way, a non-existent assertion cannot negate the substance of all things.”

The second verse gives the other horn:

Moreover, if that statement exists substantially, your earlier thesis is refuted. There is an inequality to be explained, and the specific reason for this should be given.

Suppose Nāgārjuna says instead that his own statement is the exception, the one thing with substance. Then not everything is empty, his thesis is false, and he owes an explanation of why this one statement should be treated differently from everything else. The opponent then works through six variations on the two horns. Their upshot is simple. If the claim is empty, it is powerless; if it is not empty, it is false.

Westerhoff points out what the opponent is assuming. The opponent reads “empty” as “non-existent”: an empty statement, on his reading, is one that is not there at all.[2] Keep that assumption in mind. Nāgārjuna’s first reply turns on it.

The opponent’s sharper point

The opponent also anticipates a comeback. Nāgārjuna might reply that the opponent’s objection is just as empty as his own claim, and so just as powerless. The opponent rejects this in advance:

If you thought that the negation’s negation is also like this, that would indeed not be correct. Thus your thesis, not mine, is corrupted by the specific characteristic.

The “specific characteristic” is the defining claim of Nāgārjuna’s position, that everything is empty, and the opponent’s point is that it belongs to Nāgārjuna’s thesis, not to his. As the commentary puts it: “In this context you say ‘all things are empty,’ I do not.” The opponent’s words are empty only if Nāgārjuna is right, and whether he is right is what the debate is about. To answer the objection by calling it empty would assume the conclusion.

Westerhoff finds a sharper version of this point beneath the surface. Nāgārjuna’s opponents in this work often argue from the assumptions of the Nyāya school, for whom words gain their meaning by connecting with things and properties in a world that exists independently of us. A statement that lacks substance, then, would be one whose meaning comes only from human conventions. Westerhoff offers his own illustration of how such a statement looks to the opponent: it is like money earned playing Monopoly, which works perfectly well inside the game but cannot buy anything outside it.[2] On this reading the objection is not that Nāgārjuna’s words do not exist. It is that they never leave the game of convention, while the opponent’s words connect with reality.

Empty, and still at work

Nāgārjuna begins by accepting the first horn. His own speech, he says, is neither in its causes and conditions nor separate from them, and so it is empty, like everything else (VV 21). Then comes the reply on which the rest depends:

The dependent existence of things is said to be emptiness, for what is dependently existent is lacking substance.

To be empty is not to be nothing; it is to exist in dependence. And dependent things work. The commentary gives the examples: “a chariot, pot, cloth, and so forth, which are empty of substance because they are dependently originated, perform in their respective ways by removing wood, grass, earth, by containing honey, water, or milk, and by bringing forth protection against cold, wind, or heat.” Nāgārjuna’s speech, empty in the same way, can play its part in showing that things lack substance.

The next verse offers an image that lesson 5 met in chapter 17 of the Mūlamadhyamakakārikā:

Suppose one artificial being were to hinder another artificial being, or an illusory man would hinder one brought about by his own illusionistic power. This negation would be just like that.

As Westerhoff explains the image, a figure produced by a conjurer can really stop another conjured figure, because both belong to the same performance. In the same way, an empty statement can really refute an empty claim. Nāgārjuna then refuses the second horn outright:

This speech does not exist substantially, therefore there is no destruction of my position. There is no inequality, and no particular reason to be mentioned.

There is no exception to explain, because there is no exception. As Westerhoff puts it, “‘all’ means ‘all’”. He adds in a note that “empty” does not mean “false” either; if it did, the claim that everything is empty would turn into a version of the liar paradox.[2] So the first reply keeps the levels apart. Ultimately, the statement that everything is empty has no substance. Conventionally, it is spoken, understood and does its work, as a pot holds water.

“Do not make a sound”

A second exchange shows where Nāgārjuna thinks the debate takes place. The opponent had considered, on Nāgārjuna’s behalf, a comparison: perhaps the empty statement works like someone who says “Do not make a sound”, using a sound to stop other sounds. He rejected it at once:

If you were of the opinion that it is like “do not make a sound,” this would not be adequate, for in this case there is the prevention of a future sound by an existing one.

Nāgārjuna agrees that the comparison fails, and disowns it: “This is not our example” (VV 25). He offers a better one: an artificial man, created by the Buddha or one of his disciples, who frees another man from the belief that an artificial woman is real (VV 27). Illusion corrects illusion. Then he turns back to the rejected example:

It is rather that the example is of the same nature as what we want to establish, for there is no existence of sound. We do not speak without assenting to the conventional truth.

The opponent’s objection to the example depended on contrasting an existing sound with a non-existent statement. But sounds, too, lack substance, so the contrast assumed what was in dispute. And Nāgārjuna adds the principle behind his whole reply: “For it is not without having had recourse to the conventional truth that the nature of things can be explained.” He then quotes a verse, without naming its source, which Westerhoff identifies as 24.10 of the Mūlamadhyamakakārikā, the verse lesson 6 read in Siderits and Katsura’s translation:

Not having had recourse to the conventional, the absolute is not taught. Without having approached the absolute, liberation is not reached.

Westerhoff draws the lesson: “the level of absolute truth is not the level at which philosophical debates are conducted.” Arguments about emptiness, like all arguments, happen within conventional truth, where words have shared meanings and statements can succeed or fail.

This is not quite the reply we met at the end of lesson 7. There the commentator Bhāviveka used a similar comparison in Nāgārjuna’s defense: saying “Quiet!” to stop noise is not inconsistent. Siderits and Katsura point readers to verse 28 for Nāgārjuna’s own response to the comparison, and that response, as we have seen, begins by setting it aside as “not our example” before turning it around.[3]

“I do not have any thesis”

Nāgārjuna answers the opponent’s fourth objection in what Westerhoff calls the most famous verse in the work:

If I had any thesis, that fault would apply to me. But I do not have any thesis, so there is indeed no fault for me.

The commentary adds: “while all things are empty, completely pacified, and by nature free from substance, from where could a thesis come?”

Read literally, this is puzzling. The Dispeller of Disputes is full of claims, and so is the Mūlamadhyamakakārikā. In his 2009 book Westerhoff distinguishes three ways of reading the verse, each with a history in the Buddhist commentarial tradition.[4] None of them speaks for all of Madhyamaka.

The semantic reading, which is Westerhoff’s own. In his words, “Nāgārjuna does not claim to hold no thesis whatsoever, but claims only to accept no statements that are taken to have a particular semantics”, that is, a particular account of how words get their meaning and statements their truth. In his commentary on the translation, Westerhoff explains what this does to the dilemma. The charge of powerlessness assumed that statements come in two kinds: some, like the opponent’s, get their meaning and truth from a world that exists independently of us, and others, like Nāgārjuna’s, only from human conventions. If there were both kinds, the second could never overrule the first. Nāgārjuna has no thesis of the first kind, and he denies that there are any. If all statements, the opponent’s included, are given the same convention-based account, then, to borrow Westerhoff’s image, there is no money outside the game, and an empty claim meets its rivals on equal terms.

The argumentational reading. Here, in Westerhoff’s summary, Nāgārjuna “makes a claim about how one should proceed in debates, namely by always refuting opponents via reductio arguments, without ever adopting any thesis oneself.” A reductio shows that an opponent’s own premises lead to an absurd conclusion. On this reading the dilemma finds no target: it attacks a thesis, and the Mādhyamika only draws out the consequences of other people’s theses. What it needs is a way to square this with how much Nāgārjuna seems to assert; Westerhoff, for his part, holds that the works are plainly full of philosophical claims.

The transcendent reading. This one, in Westerhoff’s words, “reads Nāgārjuna’s statement as the assertion of the existence of an inexpressible reality beyond concepts and language.” On it the dilemma touches only statements, and what the arguments point to lies beyond anything statements can say. It needs such a reality, and it inherits the objection from lesson 6: a reality with a nature of its own, however inexpressible, is what Westerhoff thinks the arguments deny.

Westerhoff traces the semantic and argumentational readings to Tibetan writers including Tsongkhapa, the scholar lesson 6 mentioned, and transcendent readings to other Tibetan scholars. He then sets the second and third aside, not because they are worthless in their own settings, but because, he judges, they tend to use the verse as “a textual peg” for arguments about other matters. That judgment is his. This study has read only his brief descriptions of those two readings, not the works in which they were developed.

What the reply settles, and what it leaves open

On the semantic reading, the reply does real work. The claim that everything is empty is not inconsistent, since it applies to itself without exception. It is not powerless merely because it is empty, since empty things, from pots to statements, do what they do. And it does not need to be exempt from the conventions through which every statement is made.

But the opponent still has a move, and Westerhoff says so. The opponent can refuse Nāgārjuna’s account of meaning altogether.[4] Recall the two kinds of negation from lesson 7. The opponent need not assert the opposite of Nāgārjuna’s claim; he can reject the claim together with the presupposition that all meaning rests on convention, as one rejects “the number 2 is red” without implying that it is some other color (the example is Westerhoff’s). At that point the dispute is no longer about emptiness alone. It is about which account of meaning and truth is right: one on which words reach a world that exists independently of us, or one on which meaning is convention all the way down. Westerhoff notes that modern critics of global relativism have pressed a dilemma of the same shape, and, without equating the two positions, he suggests they might be defended in similar ways. He does not claim that the larger question is settled, and this study does not settle it either.

An optional connection

The Dispeller of Disputes has a second cluster of objections, beginning at verse 5: if everything is empty, then perception, inference, testimony and likeness, the means by which anything is known, are empty too, so how could anyone know that things are empty?[1] Westerhoff reads the opponent here as arguing from Nyāya assumptions, and in his 2009 book he reconstructs Nāgārjuna’s answer: nothing is intrinsically a means of knowledge, yet procedures can still count, conventionally, as means of knowledge within a particular inquiry.[4] The first lesson of the study “Inference under challenge”, “Nyāya: what warrants an inference?”, sets out the Nyāya account of warrant this objection presupposes; the comparison is optional, and this study does not pursue it.

Final checkpoint: the orchestra’s old debt

Imagine a community orchestra founded sixty years ago. No founding member still plays, and none of its present players joined before the last decade. Twenty years ago it merged with a local wind ensemble and took a new name. Forty years ago, it borrowed money from a music trust to buy timpani, and the loan has not been fully repaid. At a committee meeting, the treasurer, Winifred, who has been reading about Nāgārjuna, puts an argument: “The orchestra that borrowed that money is gone. Every player has changed, and so has the name. And if Nāgārjuna is right, an orchestra has no nature of its own anyway. An empty orchestra can’t owe anything.”

You can work from the case alone; no further facts are needed. Writing an answer is optional.

Task 1: reconstruct

Set out a Madhyamaka argument that the orchestra lacks intrinsic nature, using one pattern from this study: neither identical nor distinct, the three times, or parts and whole. Then identify one premise an opponent could reasonably reject.

Task 2: separate

Explain which facts about the orchestra still hold conventionally, such as its debt, its rehearsal times and whether a performance was good, and why the orchestra’s emptiness does not cancel the debt. Say what your answer leaves unsettled.

Task 3: compare

Describe how two readings from this study would state the conclusion differently: for example, the reading on which there is no ultimate truth and the reading on which a reality beyond concepts remains, or two readings of “I do not have any thesis”. Explain why neither can speak for all of Madhyamaka, and name the kind of evidence this study has read.

A full explanatory response

Task 1: Reconstructing the argument

The pattern of “neither identical nor distinct”, which lesson 4 applied to the self and to fire and fuel, fits the orchestra well.

Suppose the orchestra had a nature of its own. Then it would have to be either identical with what it is made of (its players, instruments, music and ways of rehearsing) or something distinct from them. If it were identical with them, it would change whenever they changed. A new oboist would make a new orchestra, and the orchestra that borrowed the money would indeed be long gone, replaced by a succession of orchestras as numerous as the changes of membership. If it were distinct from them, it would have none of their marks. It could not play, rehearse or sign a loan, since players do those things, and it could go on existing with no players at all. Neither option describes anything an orchestra could be. So, ultimately, the orchestra has no nature of its own. It is conceived, and named, in dependence on its players, its music, its records and its practices, as the chariot is conceived in dependence on its parts.

An opponent could reject the premise that identity and distinctness exhaust the options for a real thing: perhaps a whole can be something real that its parts make up, without being either just those parts or a separate thing that could exist without them. The Personalists of lesson 4 tried something like this for the person, and Nāgārjuna’s reply, through fire and fuel, was that such a relation leaves neither side a nature of its own. An Abhidharma thinker, by contrast, would reject nothing here at all. He would happily agree that an orchestra is a conceptual fiction, and insist that the players, sounds and moments of consciousness from which it is built, analyzed down to their simplest elements, are ultimately real. The argument against the orchestra does not reach those elements. For that, Nāgārjuna needs the arguments of lessons 1 to 3, and their premises about causes and time.

Task 2: Separating the levels

Winifred’s argument moves from “the orchestra has no nature of its own” to “the orchestra cannot owe anything”. That is the step this study has examined again and again, and it fails for the same reason each time. Lacking intrinsic nature is not the same as not existing. The Dispeller of Disputes makes the point with chariots and pots, which are empty and still carry wood and hold water; an empty orchestra still rehearses and still performs.

Her other point, that everything has changed, is the one Dagny made in lesson 5. It assumes that if the present orchestra is not strictly identical with the one that borrowed, it must be strictly a different one. Ultimately, the two are neither identical nor distinct, so the ultimate level delivers neither verdict. Conventionally, the orchestra is one continuing body. Its members, its records and the trust all treat it that way, and that practice is what lets an orchestra borrow money in the first place. Within those conventions, the debt is a fact, and so is the rehearsal schedule. Whether last season’s concert was good is a conventional judgment too, and like Fergus’s call in lesson 6, it can be right or wrong by the standards musicians and listeners share.

So emptiness does not cancel the debt, because emptiness is not a fact about the orchestra that competes with its conventional facts. As lesson 6 found in chapter 25, even “no distinction whatsoever” at the ultimate level says nothing against distinctions that hold conventionally.

This leaves questions open. Whether this particular debt passed to the merged orchestra depends on the terms of the merger and the loan, which the case does not give; those are conventional questions that only the relevant agreements and practices can settle. And the answer shows only that emptiness is no reason to disown the debt. It does not show what makes one set of conventions binding rather than another, the question left open at the end of lesson 5.

Task 3: Comparing two readings

On the reading Westerhoff develops, on which there is no ultimate truth, the conclusion is short. There is no fact about the orchestra beyond its players, practices, records and agreements, because there is no way things are apart from all such conventions. Conventional truth is the only truth there is, and it includes the debt. Nothing deeper has been discovered that the debt could fail to match.

On the reading that a reality beyond concepts remains, as Westerhoff reports it from Stcherbatsky and Murti and, in another form, from Dharmapāla, the conclusion is stated differently. The orchestra, its debt and its concerts belong to the world of appearances. The arguments show that this world is not ultimately real, and they point beyond it to a reality that no description, including a description of debts, can capture. The debt holds within the world of appearances; it simply is not the last word about what there is.

The two readings agree that Winifred is wrong. They disagree about what “ultimately” means when we say the orchestra is ultimately empty: on the first, that there is no deeper fact at all; on the second, that the deeper fact cannot be put into words. Neither can speak for all of Madhyamaka. Each has had serious defenders, each needs something the other denies, and the verses do not settle the choice by themselves. The same holds for the three readings of “I do not have any thesis”.

This study has read Nāgārjuna’s verses in translation from the Sanskrit, guided by the translators’ commentary and the classical commentators they report; the Dispeller of Disputes in Westerhoff’s translation; and two books by Westerhoff, one modern philosopher reconstructing the arguments. Other voices, from Abhidharma thinkers to Tibetan interpreters and modern logicians, have reached us only through those books.

End of the study

The third stage began by asking what was left of the two truths. Lesson 6 found that, once no ultimately real elements remain, the ultimate can be read in at least three ways: as a reality beyond concepts, as the absence of all entities or as the recognition that there is no ultimate truth at all. It also found that the study’s own habit of saying “ultimately” is itself a reading. Lesson 7 followed Nāgārjuna’s insistence that emptiness is empty too: a medicine that must leave with the illness, a correction rather than a hidden substance, and a verse that denies all four alternatives without abandoning logic, on the reading followed there. This lesson followed the objection that such a claim undermines itself, and two replies: that empty things, statements included, still do their work, and that all argument proceeds within conventional truth.

The study began with a question: can things exist, change and matter without possessing a nature of their own? Nāgārjuna’s answer, as we have read it, is that they can and must. “All is possible when emptiness is possible” (24.14): the causes, motions, selves, actions and truths we rely on work only because they are dependent. At every step the conclusions had two levels. Ultimately, nothing has intrinsic nature, not even emptiness. Conventionally, seeds sprout, walkers reach the top, promises bind and umpires can be wrong. And at every step the argument depended on premises an opponent could question: about time and momentariness, about whether the line between a thing and its features is our own, about whether a convention can bind if nothing grounds it, and now about whether meaning can rest on convention all the way down. That last question is the one this lesson leaves most clearly open. If an opponent refuses a convention-based account of meaning, does anything remain that could decide between the two accounts?

What this study has read, as the end of the checkpoint’s response set out, is a place to begin, not a final account.

If you came to this study from the guided course, you can return to its short lesson on Nāgārjuna, “Does emptiness mean that nothing exists?”, and read its account of chapter 24 again with this study’s readings in mind. The guided course continues with lesson 13, “Learning to become humane”.

Sources and limits

[1] Nāgārjuna, The Dispeller of Disputes: Nāgārjuna’s Vigrahavyāvartanī, translated with commentary by Jan Westerhoff (Oxford University Press, 2010), for the verses and Nāgārjuna’s own commentary.

The objections: VV 1–4 with the autocommentary, including the six points after VV 2.
The replies: VV 21–29 with the autocommentary, including the quotation of Mūlamadhyamakakārikā 24.10 in the commentary on VV 28.
The second cluster of objections: VV 5–6, read only to locate it.
The text and its translation: introduction, on the single Sanskrit manuscript, the basis of the translation and the rendering of svabhāva as “substance” (with notes 25–27).

The ebook has no printed page numbers, so verses are cited by number and Westerhoff’s commentary by section. The translation follows a transliteration of the single surviving Sanskrit manuscript, found in Tibet in 1936, and turns to the Tibetan translation where the Sanskrit is defective. Westerhoff’s commentary rearranges the verses so that each objection is followed by its reply; the verse numbers keep the original order. VV 7–19 and 31–70 were not read; VV 20 and the opening of VV 30 were read only for context. No Sanskrit was consulted.

[2] Jan Westerhoff, commentary in The Dispeller of Disputes (Oxford University Press, 2010), section 3.1, “The Status of the Theory of Emptiness”, for his reading of the dilemma and the replies. Sections 3.1.1–3.1.3 and their endnotes 1–30 were read in full.

The opponent’s reading of “empty” as “non-existent”; the conjured figures of VV 23; “‘all’ means ‘all’”; “empty” is not “false”: §3.1.1, with endnote 5.
The sound example and the level of debate: §3.1.2.
The two readings of VV 4; Monopoly money; the semantic reading of VV 29 and the “two-flavor” picture; the number 2: §3.1.3.

Monopoly money and the number 2 are Westerhoff’s illustrations, not Nāgārjuna’s. So are several others in this commentary that the lesson does not use, among them characters in a film, a banknote, the Ames room and the rules of the road. His comparison with the modern dilemma for global relativism is his own.

[3] Nāgārjuna, Nāgārjuna’s Middle Way: Mūlamadhyamakakārikā, translated with commentary by Mark Siderits and Shōryū Katsura (Wisdom Publications, 2013), for Bhāviveka’s “Quiet!” (commentary on 22.11, p. 248) and the translators’ footnote pointing to Vigrahavyāvartanī 28, for 24.14 (p. 276) and for 24.10 in their translation (p. 273). Their version of 24.10 reads: “The ultimate truth is not taught independently of customary ways of talking and thinking. / Not having acquired the ultimate truth, nirvāṇa is not attained.” The two translations differ in vocabulary (“the conventional” and “the absolute” against “customary ways of talking and thinking” and “the ultimate truth”; “liberation” against “nirvāṇa”), not in sense.

[4] Jan Westerhoff, Nāgārjuna’s Madhyamaka: A Philosophical Introduction (Oxford University Press, 2009), for the three readings of VV 29, the opponent’s refusal of the semantics and the means of knowledge. These chapters and their endnotes were read in full.

VV 29 in its context; the three readings and their historical holders: §§9.1–9.2, pp. 183–187, with the three definitions on pp. 186–187.
The semantic reading; the opponent’s refusal of the non-standard semantics: §§9.3–9.4, pp. 188–198, with the refusal on pp. 193–194.
Means of knowledge and their conventional status: chapter 8, pp. 165–182, with the conclusion in §8.4, pp. 181–182.

The 2009 book translates VV 29 slightly differently: “If I had some thesis the defect [just mentioned] would as a consequence attach to me. But I have no thesis, so this defect is not applicable to me.” The lesson quotes the 2010 translation. The 2009 book also criticizes the Tibetan reading on which Nāgārjuna has no thesis that exists substantially, while the 2010 commentary uses a similar formulation within its own semantic reading; the lesson follows the 2010 wording for his view. The argumentational and transcendent readings are known here only through Westerhoff’s brief descriptions; the Tibetan works he cites were not read.

The notice board, the orchestra and Winifred’s argument are original teaching examples.