Mabbett- Devadatta in House Nagarjuna VV Liar

7/29/2019 Mabbett- Devadatta in House Nagarjuna VV Liar

1/26

IAN MABBE-IT

IS THERE! A DEVADATIA IN THE HOUSE?Nligkirjunas Ugrahavytivartani and the Liar Paradox

I. JNIRODUCTIONIn the Vigruhavyiivurtanl Nagarjuna claims that all things are void, andthat he is asserting no proposition. He acknowledges that the statement(vacana) of his teaching counts as some kind of thing, and that thisstatement is thus void. To be void is to lack intrinsic reality. Whatlacks intrinsic reality is not real. Therefore, it seems to follow - andNagarjunas words appear to confirm it - that when he says All thingsare void, he is not expressing any real view. If he were, then he wouldbe expressing the view that his own expression of a view is unreal -that is, the view that he is not really expressing any view.

Is it really as simple, and absurd, as this? Many interpretations havebeen, and are, advanced about the meaning of Nagtijunas claims. Someconsider that he offers a logically impregnable argument; some, thathe offers no argument. Nagarjuna himself certainly appears to say thathe is saying nothing, especially in the Vigruhavy~vurtunl; I have noproposition,* he says. Because there exist no things to be apprehendedhe neither affirms nor denies.3 I negate nothing, and there is nothingto be negated.4 When he denies that things have any intrinsic reality,he says, his denial does not in fact establish any absence of intrinsicreality.5 It may therefore seem to be only a ghost of a denial.Nagarjunas teaching that all things are void has seemed to many,in his own time and since, to involve him in something like the LiarParadox (the sort of paradox involved in denying what one is saying).This interpretation will seem natural (although it will be contestedbelow) most especially if void is taken to imply false when it isapplied to a proposition. If all things are void, then all propositions arefalse, including the proposition that all things are void. A comparableresult follows if void is taken to imply meaningless when appliedto a proposition: what Nagarjuna means is meaningless.Much of the modem literature on Madhyamaka proceeds from theassumption that Madhyamaka does indeed declare all propositions falseJournal of Indian Philosophy 24: 295-320, 1996.@ 1996 Kluwer Academic Publishers. Printed in the Netherlands.


2/26

296 IAN MABBE-ITor meaningless, and seeks to rescue Nagarjuna from self-refutation orself-stultification by suggesting how he can be offering such a doctrinewithout advancing any proposition (which is what some of his statementsappear to be doing). Some modem defenders of Madhyamaka havesupposed that, when he says that he has no proposition, he means quiteliterally that he has no argument to advance, and have been inclinedto see his move as a desperate way of escaping the Liar Paradox.If he were unable to advance a proposition without falling foul ofhis own claim that all things are void, he might simply renounce allclaim to be asserting anything, justifying his renunciation by the appealto the disappearance of all separate things in the corrosive glare ofultimate truth. This would destroy all his opponents arguments, andhis own as well. Nothing that he said could be on any better footingthan anything that anybody else said. Like Samson pulling down thetemple, or Sherlock Holmes hurtling in a fatal embrace with Moriartyto his death in the Reichenbach Falls, Nagarjuna would be demolishinghis opponents at Pyrrhic cost.This article is addressed to the limited and specific purpose ofshowing that such interpretations are unnecessary; the Vigrahavyb.zrtanican be read without making the assumption that Gzya means falseor meaningless when applied to propositions. This purpose can beachieved without examination of the complex issues inhering in theinterpretation of Nagarjunas major work, the MtilamadhyamakakfirikZis,and without necessarily proving correct one out of all the possibleinterpretations of the real meaning of Madhyamaka philosophy. Themethod will be to set up the hypothesis that, whatever the real meaningof Madhyamaka philosophy, it is one that does not read void (SUnya) asimplying false or meaningless, and thus does not commit Nagarjunato the Liar Paradox; after discussion of the context in modem scholarlyliterature, this hypothesis will be set to work by examining the passagesin the Vigrahavyc%zrtanl which most appear to support the contestedinterpretation. It must be emphasized, and the point will bear repetition,that the exercise is not intended, and within this scope cannot beexpected, to demonstrate the correctness of just one interpretation ofthe real meaning of Madhyamaka philosophy, and it is not a study ofthe I&r&is.

There is no doubt that the Liar Paradox has loomed over the studyof Madhyamaka and in some ways shaped the purposes of much of thescholarship in the field. It is therefore an appropriate starting-point here;it would be well to be clear what it is from which NagCirjunas modemexegetes often wish to rescue him. If we decide that what Nagarjuna


3/26

IS THERE A DEVADATTA IN THE HOUSE? 297asserts is a denial of itself, what follows? The proposition in question isa case of the Eubulides form of the paradox: This sentence is false,6which appears on examination to be true if it is false and false if it istrue. If a sentence denies itself, what can we say about it?The problem is complex and has been much discusseda One wayof taking the Liar Sentence leads to the conclusion that it has no truth-value. A sentence which declares some proposition to be true or falsecannot be assigned a truth-value if it does not adequately identify theproposition. For example, the sentence The proposition is true (orfalse) does not, by itself, identify what is referred to by its subjectand therefore does not by itself have any truth-value, any more thandoes a nonsense sentence. To give it one, we must be able to supplythe proposition. The proposition p is true (or false) can be given atruth-value if we know what p is.The Liar Sentence (This sentence is false) declares a propositionto be false, Let the proposition thus declared to be false be p. Thus theLiar Sentence says: p is false.The truth-value of this cannot be judged until we supply a specificationof p in a determinate form capable of being judged true or false. Weknow what p means. It means p is false. We can therefore (as a firststep in an obviously doomed quest) substitute this for the symbol p,and we obtain: p is false is false.No truth-value can yet be assigned to this, however. It is necessaryto make a further substitution for p, and we obtain: p is false isfalse is false.To put it another way, Fp is equivalent to F(Fp), which is equivalentto F{F(Fp)}, which is equivalent to F[F{F(Fp)}].

A specification of p, capable of being assigned a truth-value, thuslies at the latter end of an infinite series. There exists no term whichis at the latter end of an infinite series (or the series would not beinfinite). Therefore the Liar Sentence does not allow the specificationof a proposition in a form in which it can have a truth-value, and theLiar Sentence cannot be said to be true, or false; no truth-value canbe assigned to it. Therefore, if Nagarjuna actually means his claimthat All things are void in a sense which commits him to the LiarSentence, he is not saying anything profound, or self-guaranteeing (ashis modem supporters sometimes maintain). He is simply uttering asentence to which no truth-value can be assigned.On the other hand, it can be argued that the Liar Sentence implies acontradiction.* One way of making the contradiction explicit is to addto the Liar Sentence the otherwise unspoken illocutionary element


4/26

298 IAN MABBETFof assertion; it can be turned into: I assert that p is true and that itis false. This is contradictory, whether or not p means anything. Anyphilosophy which commits itself to the Liar Paradox therefore producesan undesirable result.Whether or not he could deploy the technical concepts of logic,Nagigsujunawas no doubt as much aware of this undesirability as aremodem scholars. It is not plausible to claim that Nagarjuna deliber-ately refutes himself or argues for a contradiction; on the contrary, herejects both the conjunction and the bi-negation of p and -p. Manyhave supposed that in one way or another Nagarjuna sought to offerthe doctrine that all propositions are false (or meaningless) withoutactually asserting any proposition. The view to be defended here isthat it is possible to make sense of Madhyamaka (at least, in theVigruha~dvartant - the KErikcis must be put on one side) withoutresorting to such a supposition.

II

At the risk of repetition, it is necessary to emphasize the limits of whatcan be attempted here. Nagarjuna has attracted a remarkable quantity ofscholarly interest, and numerous interpretations of Madhyamaka ideashave been offered. There is no consensus, and it is probably fair todescribe the present state of opinion about the meaning of his philosophyas in disarray. Now, to isolate for study a particular problem in theinterpretation of Vigraha+ivartun~, which is in view here, requires thatmany issues in Madhyamaka studies of engrossing interest to manyscholars must be set on one side, and at least some assumptions must bearbitrarily made; such assumptions may lack conviction for a majorityof those who have studied the Klir&s closely, for no one interpretationcommands majority allegiance. All that can be done within the presentlimits is to make as few assumptions as possible about the Madhyamakacontext, and to appeal for a willing suspension of disbelief.Here are some ways of understanding Madhyamakas Siinya doctrine,selected not to map the whole field of interpretation but to providecontext for the present discussion:1. &nya~ entails that all propositions are false or meaningless. Therefore, to avoidthe Liar Paradox, when Nagarjuna negates all propositions, his act of negation mustnot be understood to commit him to the assertion of any proposition. (It is the chiefpurpose of what follows to show that this interpretation can be avoided.)2. There is no reality whatsoever. The universe contains no things, no events, andno propositions. Nothing ever happens. (This interpretation is just as disastrous forNagajuna as is No. 1. It commits him to saying that he is not saying anything,


5/26

IS THERE A DEVADATTA IN THE HOUSE? 299that he is not delusively appearing to say anything, and that nobody is making themistake of believing him to be saying anything. Quite simply, nothing is happening.This interpretation must be shown to be implausible as well as No. 1 if Nagarjunais to be rescued from paradox.)3. Absolute reality transcends language and concepts, so that no propositions cancapture ultimate truth.4. On the level o f conventional truth, things exist. On the level of ultimate truth,nothing whatsoever exists.5. On the level of conventional truth, phenomena can be treated as manifestationsof immutable essences. On the level of ultimate truth, immutable essences do notexist. Phenomena are merely manifestations of other phenomena. (There are problemsin deciding exactly what this means, but they are no greater than the problems indeciding exactly what many philosophies mean.)

Whatever assumptions may be made here about the character ofMadhyamaka doctrine, they must be assumptions which obviate thefirst two interpretations, which would commit Nagarjuna to refutinghimself (This sentence is false or meaningless; This sentence doesnot exist). This object is secured by adopting as a working assumptionthe proposition that voidness, for Nagarjuna, means the same thing asthe Buddhist principle of conditional origination, prutttyasamutp~%&and that this principle entails at least that phenomena exist.The first part of this proposition is scarcely debatable; we are obligedby Nagarjunas own repeated and explicit claims to recognize theclose identification of voidness with pratifyasamutpiida, conditionalorigination. As Nagarjuna says,yah pratiiasamutpiidah S@nyat@z t@ pracak?ahe /(MMK 24.18ab)We declare that conditional origination is Snnyata.

The significance of this important verse has been much discussed,but the identification of voidness with the original Buddhist doctrineof conditional origination is here quite clear. Whatever is meant bythe concept of conditional origination, it is meant by Nagarjuna whenhe speaks of voidness. This identification is a cardinal principle of theK&-ii& and it recurs throughout the Vigrahavyihzrtan~. It is not ametaphor or a faGon de parler.To say that all things are void is thus not a statement about theconditions of knowledge or about the truth-value of propositions.It is the ontological claim that all things originate conditionally, independence. Let us stipulate, for the purpose of this exercise, that theorigination of things is equivalent to the manifestation of phenomena.This has the effect of making unnecessary the first two interpretationslisted above. In saying that all things are void, Nagarjuna is saying


6/26

300 IAN MABBETTthat phenomena are manifested, not that propositions are false ormeaningless. Further, if phenomena are manifested, however relativelyor conditionally, at least something is happening in the universe, andthat something may in principle serve as a trace or sign of a trueproposition. (Interpretation No. 2 is thereby avoided.)As far as possible, it is desirable to avoid commitment to any oneof interpretations 3, 4 or 5, which cannot be explored without tacklingissues of contention. The recent interesting contribution by T. Woodseems to stop short of the extreme nihilism of interpretation 2,11 but itwould not be easy without detailed analysis to assess t in relation toother interpretations such as those of C. Oetke or D. Seyfort Ruegg,noted below.My own preference is for No. 5, but there is no space here toattempt a justification or full explanation. It offers, at least, a way ofunderstanding the often-overlooked identification of Madhyamaka withthe Middle Way, madhyama pratipad.* The KG-&is can be read as anattempted demonstration that, for any phenomenon whose manifestationcan be affirmed on the level of conventional truth, rigorous analysisproves both that such a manifestation requires the existence of a specificimmutable essence, and that such an immutable essence cannot exist.The conclusion from this is not that phenomena are non-existent, butthat their real substrata can never be found (they are absent or infinitelydeferred, so to speak). Phenomena are dependent upon each other ratherthan upon real substrata. Dependent means unreal, and unreal doesnot mean either existent or non-existent, although of course eithermay be affirmed as a provisional way of negating the other - Nagarjunafrequently claims that things, such as his own propositions, do not exist,but at appropriate points he also denies that things do not exist.Obviously these observations are inadequate as a demonstration ofthe correctness of interpretation No. 5 in the list above; they raise manyproblems. For the purpose of what follows, it is enough to stipulatethat, whatever void means in Madhyamaka, it does not mean false,meaningless or (without qualification) non-existent.Let us turn to the concepts employed in the VigrahavyavartanL. Thereare various terms designating statements or expressions of belief whichneed to be kept distinct in order to avoid confusion. In the first place,the term view or philosophical view will be used here to refer tophilosophical statements or doctrines (dariana, vtida) in general. It is abroad category. It is needed in order to discuss the interpretation notedabove, according to which &nya means false and Nagarjuna rejected


7/26

IS THERE A DEVADAPA IN THE HOUSE? 301any sort of philosophical claim. View will here mean philosophicalclaims in general.

Within this broad category, it is necessary to identify the narrowerone of speculative views (&fi), which Madhyamaka consistentlyrejected, whether or not it rejected all views whatsoever.13 (These canbe considered to include views which hypostatize concepts, treatingthem as independently real things.)Again, proposition or thesis (pruti~~) designates a proposition- -.advanced in argument. In the VigruhavyQvurtunl; Nagarjuna denies thathe has any prutijCii (n&i cu mama prutijiiii). D. Seyfort Ruegg hasargued that in Madhyamaka, this term sometimes refers to propositionsin general but, in the relevant parts of Nagarjunas discussion as wellas in some later Madhyamaka contexts, is limited to propositionspostulating the existence of real things.14The concept of a proposition is very complex when we seek toanalyse the argument in the Vigruhuqtiurtuni. We must distinguish(a) between prutijti in the sense of proposition in general and prutijein the sense of a proposition which postulates the existence of anyindependently real entities; (b) between proposition as an event onthe empirical plane, a concrete utterance, and proposition as abstractcontent; and (c) between proposition as we understand it and whatevernearest equivalent might exist in Nagarjunas thought. Here, the wordproposition will be used to refer to the abstract content of a statement,the normal usage, as distinct from a concrete utterance; it will also beused to translate Nagarjunas prutijti. Context should indicate which iswhich. But we need to be aware that Nagarjuna did not work with anyconsistent clear distinction between the abstract and concrete meanings,and his usage of prutij&i, for reasons which will be discussed below,often tends to make of it a concrete entity on the same plane as objectsof sense. It nevertheless seems desirable to use the word propositionfor prutijti because it is the normal translation, and it best capturesthe appropriate range of connotations.An important topic in the Vigruhuvytivurtuni is negation. Is Nagarjunaemploying a special sort of negation distinct from ordinary denying?Since Nagarjunas time, there has been much discussion of the meaningof negation in Madhyamaka teaching, among madhyamikas as well asamong modern scholars. For the purpose of understanding Madhyamaka,the subject is obviously important and deserves close attention. Itspervasiveness in the scholarly literature demands that it should bedocumented here. (It must be recalled, though, that, as is claimedhere, for the particular purpose of resolving the superficially apparent


8/26

302 IAN MABBETTcontradictions in the Vigruhavy&x-tuni, the exploration of Madhyamakavarieties of negation is a red herring. There is in fact no need to appealto any unusual sense of negation.)One way of justifying the claim that Nagarjuna negates all viewswithout asserting one of his own (even the view that some particularproposition is false) would be to associate it with the claim that thetruth, according to Buddhism, is ineffable. Some modern commentatorson Madhyamaka have seen it as a way of pointing towards an absolutewhich transcends verbal constructions rather than as a philosophicalteaching. l5 The notion of Siinyutki as an absolute will not be acceptedhere. On the other hand, it is true enough that the eirenic tradition inBuddhism, which disapproved of polemical debate as an obstruction onthe path to salvation, deserves to be recognized as a possible influenceupon the style of Madhyamaka; D. Seyfort Ruegg has emphasized thecontribution of this tradition. *6However, these considerations cannot diminish the need to applythe normal standards of logical analysis to the Vigruhavy~vurtuntsclaim that Madhyamaka asserts the universality of voidness apparentlywithout the use of any proposition. Most of the scholarly discussionhas been directed to the logical character of Nagarjunas rejection ofall speculative views (d&) or his denial that he offers any proposition(prutijiiii).From the time of Bhavaviveka onwards, madhyamikas claimedfor themselves a special type of negation, prusujyupruti+zdhu, whichdenied a proposition without implying the affirmation of any alter-native proposition whatsoever, even its contradictory, in contrast topuryudkzpruti~edhu, negation that implied the truth of some alter-native. Some Tibetan scholars argued that Nagarjuna, in applying hisnegation to all views, had propounded no doctrinal system of his own(svumutu), advanced no propositional thesis (prutijii&; others, notablyTson kha pa and his pupil mKhas grub rje, argued that this could notbe true.17Many modem scholars have been ready to accept as correct the claimthat Nagatjuna asserted no philosophical claims. La VallCe Pousssin, forexample, wrote: Le Madhyamika ne parle jamais en son nom propre;il na pas de systeme. T. R. V. Murtis study of Madhyamaka gaveimpetus to the interpretation of Nagarjuna that holds him to negate allviews without offering any of his own. Every thesis is self-convicted;Nagarjuna has no tenet of his own and does not maintain any point ofview in polemic. Negative judgment is the negation ofjudgment, andnot one more judgment. It is on a higher plane of self-consciousness.


9/26

IS THERE A DEVADATTA IN THE HOUSE? 303Negation of positions is not one more position. Hence in Murtisview Naggarjuna negates all the views of other schools, committinghimself to none of his own.B. K. Matilals writings offer several comments on the issue. In Logic,Language and Reality he argued that Nagarjuna is not philosophicallyuncommitted; he sees metaphysical questions as only pseudo-questionsabout pseudo-concepts; so they are empty of essence. In his laterwork, Perception, Matilal interprets Nagarjunas repeated denial ofintrinsic reality (svubhavu) to all things (bhiivus) as entailing the denialof certainty or meaning to all theses.21 Madhyamakas rejection of allstatements is a form of commitmentless denial.22 Nagarjuna howeveris not guilty of asserting propositions which would thus fall victim tothemselves, for the rejection of a position need not always amount toa counter-position.23 In a late article, Matilal compared Madhyamakanegation to deconstruction, and suggested that Madhyamaka assertspropositions, deletes them, and lets the deleted propositions and thedeletions stand together.24Also relevant is a debate between Stafford Betty and David Loy.The former claimed that Nagatjuna made an attempt to evade the LiarParadox by claiming to have no view, but his denial of all assertions, if itmeans anything, amounts to an assertion itself and is thus contradictory;25the latter defended Nagarjuna, claiming that the voidness which is trueof all things is not a matter of true or false, right or wrong, but acondition that transcends the polarity or rightness and wrongness.26D. Seyfort Ruegg does not subscribe to the view that Nagarjunaadvanced no propositions, but he has contributed to the discussionof negation in Madhyamaka doctrine; he suggests that the conceptof a neustic component in a sentence, the illocutionary element Ibelieve that . . . which is usually understood as part of the meaningof an utterance, offers a guide to the character of prusujya negation:it is a denial of the neustic I believe that . . . , so that the speakerdenies that he asserts the content of the proposition, without imply-ing the assertion of its falsity. 27 This interpretation of madhyamakanegation is rejected by Kartikeya Patel, who argues that Nagarjunasnegation of all propositions makes sense if it is seen as belonging tothe dialogical/conversational universe of discourse, as opposed to theargumentative/systematic; in the former, propositions are entertainedbut not asserted argumentatively.28To treat Madhyamaka negation as a performative negation of theneustic content of a statement is not to accept the conclusion, espoused- -.by so many writers, that Nagarjuna asserts no propositions at all. This


10/26

304 IAN MABBE-ITconclusion has been contested by D. Seyfort Ruegg in two importantarticles appearing in 1983 and 1986F9 Though sometimes neglected,they demand attention.Ruegg maintains that Madhyamaka does not refrain from offeringany thesis. In claiming to offer no prat+ Nagarjuna was referringonly to those propositions (i.e., propositions presupposing the realityof entities) which his doctrine showed to have no truth-value. In itscontext, I have no prutijtimeans that he asserts no thesis or proposition positing the existence of a bhnva(positive, negative, both positive and negative, or indescribable). It does not imply,however, that he has nothing of philosophical significance to say himself, nor thathe denies all content to the stitra teachings which he proposes to explicate.3

Here we find a clear distinction between pratijiiiis, which Nagarjunadoes not have, and other sorts of proposition, which he does have.The madhyamika view about prutij~s asserts that they cannot beapplied to entities:What the Madhyamika achieves, then, by means of his prasafiga-type reasoning is thedissolving or deconstruction of all propositional theses postulating substantial entities(bhava), rather than their refutation (involving the setting up of a counter-thesis andthe holding of a counter-position within the framework of binary altematives)?i

So statements about entities are dissolved by Madhyamaka reasoning,but other sorts of statements may stand.32 Madhyamaka was not a schoolof thought without any teachings (durrianu, viidu, etc.)Rueggs interpretation of Madhyamaka principles provides one sortof reason for denying that Nagarjuna refrains from asserting anything.A different sort of reason is offered by nihilist interpretations. ForRuegg, the teaching of SUnyu dissolves all dichotomies, so that there isno thing (bhavu) which can be described as existing or not existing, andNagarjuna asserts nothing that implies the existence or non-existenceof any concrete thing. For the proponent of a nihilist interpretation, onthe other hand, things ultimately do not exist, and it is for this reasonthat no assertion can be made about anything.An objection might be that NagZirjuna often enough rejects negativestatements as well as positive - for any entity x, x neither exists nordoes not exist. However, the nihilist interpretation can be defended ifit is assumed that, in the view of Madhyamaka, to attach any predicatewhatsoever to x (even does not exist) is to say something about agiven entity whose existence is presupposed. Thus for a subject x anda predicate F, it is possible to claim that -(Fx) and also -(-(Fx) andalso -(Fx.-Fx) and also -(-{Fx}.-{-Fx}) if there is absolutely nox. All claims about the present King of France, positive or negative (to


11/26

IS THERE A DEVADATTA IN THE HOUSE? 305the effect that he is bald, or not bald, or anything else) can in a sensebe consistently denied. It has proved tempting to see the Madhyamakaconclusion as the product of this logic. It was perhaps Richard Robinsonwho first identified the interpretation as a coherent logical argument, andthe point is familiar in the scholarly literature.33 T. Wood has recentlydeployed the argument ~for a nihilist interpretation at length,34 and invarious places C. Oetke has identified Nagarjunas voidness doctrines asa denial of the existence of all things in ultimate reality (paramtirthata~);he raises the question whether this implies that Nagarjuna can have nointelligible argument.35

IIIAt this point, let us take stock. Nagarjuna asserted that all things(which, in the Vigrahavytivartanl, he acknowledged to include his ownstatement) are void, and he denied that he had any proposition (pratijiiii).This, right from the beginning, appeared disconcertingly close to theLiar Paradox, and critics condemned it on that score as a self-defeatingargument. Down the centuries, students of Madhyamaka have examinedthe nature of Nagarjunas rejection of speculative views and of the latermadhyamika concept of prasajya negation in quest of an understandingof his meaning which might make better sense.On the epistemological interpretation of the voidness doctrine, number1 in the list, all propositions must be negated because they are void(and hence false); Nagarjuna negates them, but his negation is of aspecial sort, such that he does not offer any counter-proposition of hisown which would thus fall victim to the Liar Paradox.There are alternative interpretations, which treat the voidness doctrineas a teaching about the ontological status of entities rather than about theconditions of knowledge. In one type of alternative, Stinyatti implies thatnothing can be stated about things, which are neither existent nor non-existent, because all dichotomies are dissolved. In another, Nagarjunaadvances no proposition about any thing (even the proposition thatit does not exist) because there are no things to serve as subjects forstatements.

This, very broadly, is the context of current thought about the problemsof negation raised by the Vigrahavytivartunt. Here the purpose is tolook at this text (not at early madhyamaka in general), with a viewchiefly to show that the concepts of Cnyatti and pratijki do not entailthat Nagarjuna was really trying to say that he was not saying anything.


12/26

306 IAN MABBEITIV

In applying this interpretation, it is useful to apply a distinction betweentwo senses of any word designating a statement or sentence, whichmay oscillate in actual usage according to context. In the first sense, asentence belongs to the conceptual plane; it is not an entity or eventspecifiable by place and time; it is abstract. This is the usual sense ofproposition. As such it may be said to be true or false. In the second,a sentence is a particular event, an utterance, a concrete occurrenceinvolving the use of a writing instrument or the lungs, throat, tongue etc.(as stated in Vigruhavyavurtuntla).36 It can be described as happeningor not, but not as true or false. (This sense, utterance, appears tocorrespond substantially to the Vigruhuvytivurtunts use of vucunu orvucus.) Keeping the distinction in mind will make it easier to recognizethat, according to Nagarjuna, utterance-events are unreal, but this doesnot compromise their capacity to convey real meanings. Unreality isnot simple non-existence. It is, we assume here, a sort of provisional,relative or dependent existence.This distinction between abstract propositions and utterance-eventsis one that may be useful to us in analysing Nagarjunas thought,specifically in allowing us to see how he can be making an unreal(but not totally non-existent) utterance which successfully asserts thereal content of a proposition, but this does not mean that he had itclearly in mind himself. What he is more likely to have had in mind,and what needs to be understood if we are to make sense of his claimthat he has no proposition, is not the notion of a proposition in theabstract, a meaning, but the notion of a mind-object. For Buddhism,mind is a sixth sense and its objects are things just as are objects ofthe other five senses. It is just for this reason that he is obliged to treathis own conceptions as unreal things, on the same plane as the rest ofthe furniture of the universe. This fact has important implications forhis attitude to negation, as we shall see. It complicates his reply to acritic who objects to his doctrine of voidness on the assumption thatwhat is negated must exist.Such a critic may well be a naiyayika; it may be that Nyayateachings supplied a major source of the objections refuted in theVigruhuvy&w-tunf, and that Nagarjuna was involved in polemicalexchanges with proponents of this schoo1.37It is reasonable to presume that the Vigruhuvyavurtunt was writtenlater than the Mtilumudhyumuk&ik~s, which is the locus classicus ofMadhyamaka doctrine, and that in the Vigruhuvy&zrtunl (whose titlemeans dispelling of strife) Nagarjuna sought to dispose of objections


13/26

IS THERE A DEVADATTA IN THE HOUSE? 307which had been provoked by, or which could be foreseen to, the messageof the MiAamadhyamatiril. The Vigrahavytivartantconsists of seventyverses, each followed by the authors own commentary (indicated hereby the notation a). The first twenty verses (with their commentaries)consist entirely of words put into a rhetorical opponents mouth, voicinga series of objections to the Madhyamaka teaching that all things arevoid. From the twenty-first onwards, Nagarjuna rebuts these objectionsone at a time. It is important to recognize this structure. Each verse ofNagigsirjunas wn argument is to be seen, not necessarily as a developmentof the thought in the preceding verse, but often as a reply to a newobjection which has been advanced earlier, without reference to thepreceding verse. It may be that the objection addressed by a givenverse has already been essentially refuted, but in turning to each newobjection Nagarjuna seeks to make a fresh rebuttal in order to administerthe coup de grbce.The first two verses taken together accuse Nagarjuna of committinghimself to a paradox.

I.

II.

If there exists nowhere any intrinsic reality (svabhava) of anythings (bhSiva) whatsoever, then your statement is withoutintrinsic reality, and is in no position to attack3* intrinsic- -reality (sarve.$i~ bhavanaqz sarvatra na vidyate svabhava-&et I tvadvacanam asvabhava? na nivartayituqz svabhavamalum II).But if this statement does have intrinsic reality, then yourformer proposition39 is falsified. There is a discordance, andyou must state a special justification for it (atha sasvabhZivametad Czkya~ pi&vii hata pratij% te / vai~amikatva~ tasminviiesahetus ca vaktavyah //.4o

Madhyamaka teaches that all things (bhava) are void (Gnya); thatis, they lack intrinsic reality (svabhava). Nagarjunas statement to thiseffect is included in all things, because it depends upon events takingplace in the chest, throat, lips, tongue and so forth (Ia). (Thus, at theoutset, it is clear that Nagarjuna is discussing the concrete sense ofhis statements as utterances, vacana; Nagatjunas own statement is ab&vu.) So, in the eyes of this rhetorical opponent, it condemns itself.Lacking intrinsic reality does not mean false, but the opponent istaking it to mean non-existent, which is just as good for the purpose ofconvicting Nagarjuna of self-contradiction: just as a non-existent firecannot bum and non-existent water cannot moisten (VigrahavyavartantIa), Nagarjunas statement cannot produce any effects. On the other hand,


14/26

308 IAN MABBETTif Nagarjunas statement is taken as existent (and therefore capable ofdenying the reality of things), it contradicts itself; there is a discordance.By this is meant (as will later be confirmed) that the statement of thecharacter of all things does not itself accord with that character; astatement to the effect that all things have a character C must itselfhave the character C since it is a thing, but this condition is not fulfilled,and the discordance must be specially justified.These objections go straight to the heart of the paradox, and it isimportant to see how Nagarjuna rebuts them. We might expect that,if a void statement is non-existent and therefore presumably incapableof achieving anything, NagFrrjuna would be anxious to deny that hisown statement is void. However, in verse XXI he contentedly acceptsthat his statement is void, like everything else. In verse XXII hereplies to the complaint that his statement would need to have intrinsicreality, but would then contradict itself, and it is in this reply that wemust expect to find the key to Nagarjunas defence. What he sayshere is that voidness is the name for the dependent nature of things(yaSca pratityabhavo bhtiv&@ii)z stinyateti sa proktti); what is dependentby nature Qratttyabhavo) lacks intrinsic reality (svabhavatva). Hethus identifies voidness squarely with the primal Buddhist doctrine ofdependent origination (pratityasamutpda).In the commentary to this verse, he first accuses the opponent ofnot clearly distinguishing the meaning of voidness (Siinyatiirtha? cabhavdn bhaviimim anavasaya). This accusation must be heeded; ifNagtijunas contemporaries misunderstood, we must make sure thatwe understand the misunderstanding. Nagarjuna continues: because allthings are dependent (a teaching which no Buddhist could deny), theylack intrinsic reality. If they had intrinsic reality they would not bedependent. So they are void, and Nagarjunas statement is void. But itis not thereby invalidated, and this is the capital point for our purpose:Just as things like carts, cloth or pots, in spite of being dependent and devoid ofintrinsic reality, can still carry out their functions of transporting wood, grass orearth, containing honey, water or milk, and protecting from cold, wind or heat, evenso this statement of mine, in spite of being dependent and void of intrinsic reality,can carry out its function of demonstrating that things are devoid of intrinsic reality.

This should be clear enough. The madhyarnika dictum All thingsare void is a statement about the ontological status of things. It doesnot mean that all propositions are false, or meaningless; nor does itmean that NagFtrjuna is refraining from advancing, though in principlehe could advance, any proposition, As Nagarjuna explains, to say that astatement is void and lacks intrinsic reality is to say that it is dependent.Nagarjunas utterance is devoid of intrinsic reality, but his words still


15/26

IS I-HEW! A DEVADATTA IN THE HOUSE? 309carry out their functions. We are surely entitled to conclude that thevoidness of Nagarjunas utterance does not deprive him of the right toassert any philosophical views.It is also important to notice that what Nagarjuna says here isinconsistent with extreme radical nihilism, of the sort which denies thatanything whatsoever is happening in the universe. There are indeedphenomena, even though they lack real basis. Therefore Nagarjuna doesnot incur the difficulties of one who says Nothing is happening, so Iam not speaking and this sentence does not exist.What the Vigrahavyikzrtunl tells us, then, is that the dependenceof a statement upon the material conditions of its creation, its lack ofreal substance, does not prevent it from doing its job and expressing aproposition. Therefore Nagarjunas statement all things are void doesnot contradict itself.Verse XXIII introduces the image of an unreal man, produced byartifice or magic (nirmituku; miiyi@wu~u), successfully suppressing(p-uti@zuyetu) another unreal man. Even though both are unreal,the suppression of one by the other actually occurs. Even so, a voidstatement can produce a rebuttal (pruti~edhu) of the intrinsic realityof things. (Here we observe again the assimilation of propositions,mind-objects, to things in the world: one argument quasi-physicallyassails another just as one man assails another.) Nagarjuna therebyinvites us to think of the status of things in the world, which are voidand dependent, as delusive (like a magical conjuration), not as totallynon-existent. (An apparition in a dream is not real but can really expressa proposition, with real consequences.)Verse XXIV is addressed to the complaint in verse II that Nagarjunasstatement, to be effective, would have to be non-void, and thereforefalse. This can now be simply rejected. The statement does not haveto be non-void; it can be, and is, void like everything else, and there isno discordance. A statement declaring the character of all things sharesthat character (i.e., the character of being void).At this point we need to turn to the next objection. Verse III, followingup the accusation that Nagarjuna is trapped in a paradox, deals with thepossibility that Nagarjuna might wish to evade his dilemma by claimingthat his statement can be compared to somebody producing silence bysaying Let there be no sound (ma ~uabduvuditi). By producing a sound,one might prevent subsequent sound from being produced. Similarly,says the opponent, Nagarjuna might think that his statement can establishthe voidness of all things. But this would not work, because if all thingsare void Nagarjunas statement is void; so that statement, being non-


16/26

310 IAN MABBEITexistent (asat), cannot be compared to the (existent) injunction: Letthere be no sound. This objection clearly is made without reference tothe replies by Nagarjuna to the previous objections, in which he madeclear that void does not mean totally non-existent.In his reply to this, in verses XXV-XXVIII, Nagarjuna does not simplyrepeat himself. He takes the opportunity to sort out the relationshipbetween the ontological status of denial and denied. The example of asound preventing other sounds is anyway inappropriate, because it is acase of something with a given character (that of sound) denying theexistence of anything with that same character (which indeed wouldbe self-refuting), whereas Nagarjunas statement, which is void, deniesthe existence of anything with the (different) characteristic of intrinsicreality (XXV).If a void statement denied things being void (i.e., being itself void,asserted that all things are not void), it would tend to establish theirhaving intrinsic reality; that would be contradictory (XXVI).A statement with a character C is quite consistent when it deniesthat reality has character not-C; an unreal man (a magical apparition)may consistently deny that a certain unreal woman is real; the unrealman denies reality, tending to establish unreality (XXVII).The opponents argument claiming the example (hetu) of the injunc-tion Let there be no sound for the conclusion that Nagarjunas thesisat least must be non-void is actually guilty of begging the question(sadhyasamu). Th is is because, if everything is void, the opponentcannot assume as premise that Let there be no sound is real in orderto prove that real things exist; Nagarjunas proposition can carry out itsfunction but it is still void, as everything is.4* All things are void, andthe words Let there be no sound are void. The ultimate truth cannotbe taught without reference to the conventional truth - that is, by usingvoid statements (XXVIII).Nagarjuna then turns to the objection voiced in verse IV. This againis an objection to an imagined defence which we will learn Nagarjunadoes not actually want to make.The imagined defence is that Nagarjunas denial of the reality ofthings cannot be denied. If all things are void, then it might be urgedagainst this claim that it destroys itself, because it makes itself void,unreal. But, on this same supposition, the counter-argument is itselfvoid. In order to discredit Nagarjunas statement, it has to declare itvoid, thereby accepting its truth, at the same time it declares itself voidand discredits itself. The opponent cannot reject Nagarjunas statementwithout accepting it.42


17/26

IS THERE A DEVADATTA IN THE HOUSE? 311Against this imagined defence, the opponent tersely retorts: It is yourproposition which is rendered defective by virtue of its own specific

character, not mine (evam tava pratij& la@a+o deyate na mama).That is, it is Nagarjuna who actually claims the voidness of his ownand everybodys proposition. The opponent does not claim it; he merelypoints out that one who claims it must accept that his own propositionis void.This objection is again made without reference to Nagarjunas expla-nation that voidness does not prevent things or statements from carryingout their functions. So far as the objector is concerned, voidness is afatal flaw, and his purpose is to prevent Nagiujuna from imputing thisflaw to him.It is in the light of this specific attack on the Madhyamaka thesisthat we must assess Nagatjunas defence in verse XXIX, the one inwhich he claims to have no proposition:

If I had any proposition, then I would be liable to this defect.But I have no proposition, so the defect does not affect me(yadi ti cana pratij%i syan me tata e;a me bhaved doTah In&i ca mama pratij% tasman naivasti me doSub //).43

A proper understanding of this claim to have no proposition requiresattention to the context of the opponents objection. Nagarjuna is dealingwith the charge that his voidness dictum is defective by virtue of its ownspecific property (Zaksa+a&). A Laksav is the specific characteristicby which a thing can be distinguished. The alleged defect is that ofdeclaring all things to be void, while itself having the specific characterof being a pratij&. The opponent accuses Nagtijuna of advancing aconcrete thesis, while he himself does not; he merely points out theself-contradiction resulting from Nagarjunas proposition.The commentary, XXIXa, shows how the reply to this objection shouldbe understood. Nagarjuna acknowledges that, if he were advancing a realproposition in order to attack the reality of all things, his argument wouldbe defective. However, what he is asserting does not in ultimate truthhave the specific character of being a proposition, because ultimatelythere are no real specific characters of anything. When all things arevoid, brought to perfect resolution, perfect purity of nature, how can therebe any proposition? How can anything find purchase upon the specificcharacter of a proposition? (sarvabhavesu siinye~vatyantopa&nte~uprakrtivivikte!u kutah pratij&i I kutah pratijCilak~a~apr?ipti~).44 Thatis, seen from the point of view of ultimate truth, param~rthata~, thereexist no distinct and specific entities. On the level of conventional truth,


18/26

312 IAN MABBETTsa~vyavuhtira, one has to resort to the use of terms presupposing theseparate reality of propositions and other things, but ultimately thereis no question of one thing affecting another, one argument attackinganother. There is, ultimately, no discrete entity with the Zuk+znu of apruti$&. The identification of entities as if they were independentlyreal belongs to the plane of conventional truth.Proponents of the view that the SGnyudoctrine entails that all views arefalse might suppose that Nagarjuna is claiming that other people advancepropositions, which are all defective, while he somehow demonstratesthis fact to us without advancing any proposition of his own. On thecontrary, he is pointing to the unreality of all propositions.45Verses V and VI introduce the objection that the entities denied byNagarjuna in his statement that all things are void must presumablybe first apprehended by one of the means or criteria of knowledgebefore they can be denied; but if all things are void there can be nosuch thing as apprehension. In Verse XXX Nagarjuna replies that, ifthere were apprehension of something, it would indeed be appropriateto affirm or deny it; but in fact there is no actual apprehension, sothere is neither affirmation nor denial (tusm&r nu pruvurtuy~%~Gnunivurtuy~%~i).~~ Following verses refute the validity of the criteria ofknowledge (prumii~s). Again, we must recognize that Nagarjunasargument proceeds from the absence of intrinsic reality in things.Apprehension of something is a concrete event and as such is void,so one cannot make any statement about it, affirming or denying itsexistence. This does not mean that void events, such as utterances,cannot serve as vehicles for true propositions.Verse IX introduces another objection: it argues, in effect: if allthings are void, then the name absence of intrinsic reality is void. Ifabsence of intrinsic reality is void, there is no such thing as absenceof intrinsic reality. Therefore there is such a thing as intrinsic reality.

If there were no intrinsic reality of phenomena, then thevery name absence of intrinsic reality itself would notexist, for there is no name without an object. yadi cu nubhuvet svubhZivo dhurm@Ciqz ni&wubh2ivu ityevu I narntipibhuven nuivuqa namu hi nirvustuku~ nasti Il.Though this argument looks trivially sophistical, it has legitimacyfrom the principle that a mind-object has the same sort of reality orunreality as a physical thing, being an object of sense. NagarJunasreply, in verse LIX, is that since all things are void, the label absenceof intrinsic reality, being a thing, is itself void; it is not a real entity.47


19/26

IS THERE A DEVADATTA IN THE HOUSE? 313So the criticism fails because it simply mistakes what Nagarjuna says:he has already admitted that the name absence of intrinsic reality isnot real; but this admission, for reasons already explained, in no waydisables Nagarjunas argument: a meaning can be successfully conveyedby utterances which are void. So Nagarjunas statement that he has noproposition is not a doomed attempt to escape self-refutation (the viewthat all views are false is false); it means that no concrete expressionsof views are real. But these expressions nevertheless do their job.It is now easier to deal with certain passages (in verses LX111 andLXIVa) which depend upon unfamiliar theoretical assumptions.

Here we confront a principle which was espoused by Nyayaphilosophy and not rejected by NagSirjuna. It appears in the objec-tion voiced in verse XI: one cannot negate something that does notexist; NagZirjuna negates intrinsic reality; therefore intrinsic realityexists. By virtue of the act of negation itself, the intrinsic reality of allthings is not negated @ratisedhasa~bhiIv&d eva sarvabhfivasvabhavopratiqiddhab).The Naiyayikas maintained the existence, not only of those thingswhose reality one affirms, but also of those things whose reality onedenies. Whatever one negates must exist in order to be negated, tobe available for negation. It is not possible to negate any non-existentthing.48It is perhaps not surprising that Nagarjuna accepted this assumption.As was noticed before, Madhyamaka thought envisaged affirmationand denial as quasi-physical relationships between entities; a disputantcould not assail, repel or dispose of an object that was not there. Torefute svabhava (refute being a transitive verb with a substantive asobject) was to admit its substantial existence. It was perhaps not easy forhim, in his philosophical environment, to distinguish between I denysvabhtiva (implying its existence) and I deny that svabhava exists;his attempt to formulate his sense of the distinction was a pioneeringeffort not taking a form familiar to us.49In verse LXIII, Nagarjuna turns to the objection that, since thereis no name without a thing (that is, if something can be referred toit must exist), then if Nagarjuna negates the application of the nameintrinsic reality to entities, it must belong to something nevertheless.This objection raises no real problem for Madhyamaka, he claims,for a reason that is very simply stated: I negate nothing, for thereis nothing to negate.50 That is, no act of rebutting or assailing takesplace involving a disputant and an object of attack, because there is


20/26

314 IAN MABBETTno negating of an object; there is only the denial that something is thecase. No real substantive is required as the object of the verb.

It is verse LXIV and its commentary, though, which have perhapsmost contributed to the impression that in some sense Nagarjuna isseeking to divest himself of the responsibility for saying that all thingsare without intrinsic reality, while at the same time attempting to claimthe credit due to one who says this. Such an acrobatic feat, if it isactually performed, looks suspiciously like the Liar Paradox.The objection which is here rebutted is the one advanced in verseXII, which is to the effect that, if intrinsic reality does not exist, itis without words that it is to be negated. Here reappears the doctrinethat what is negated must be existent. No significant negation can beexpressed that is not of something that exists. The rebuttal of whatdoes not exist requires no utterance: rte vacantit pratisedhah sidhyatehy asatah.Nagarjunas reply is that the words (with which he denies intrinsicreality) make known that it is non-existent, but do not attack it (atrajiiiipayate vtig asad iti tan na pratinihanti).Here, for the reader convinced that Stinya, applied to propositions,mans false or meaningless, it might appear that Nagarjuna informsus of the non-existence of intrinsic reality without actually asserting anargumentative denial of it, since the assertion of argumentative denialswould contradict the doctrine of voidness. On this reading, he seeksto evade the Charybdis of contradiction by embracing the Scylla ofsilence. Of course, a silence which says out loud that it is silence isself-contradictory; Nagarjuna has already been over this. What thendoes he mean?

The answer, by now, is not difficult. He means that he is not a realdisputant attacking a real object; rather, there is no disputant and noobject to attack (pratinihan). Both are unreal.In the commentary to this verse Nagarjuna makes a distinction. Insaying that things lack intrinsic reality, Nagatjunas statement does nothave the effect of depriving all things of intrinsic reality (lit. doesnot make all things to be necessarily devoid of intrinsic reality: na- - - -nikvabhavan eva sarvabhavan karoti); what it does, rather, is (giventhat there is no intrinsic reality) to make known that things lack intrinsicreality (kimtvasati svabhtive bhava nibvabhava ltl Jnapayate).This is further explained by the example of a response to a personwho claims, erroneously, that Devadatta is in the house. Somebodytells him in reply: He is not (M&i). This statement He is not doesnot have the effect of depriving Devadatta of existence (lit. does not


21/26

IS THERE A DEVADATTA IN THE HOUSE? 315make Devadatta non-existent, na . . . devadattasytisadbhavam karoti).52Nagarjunas statement that things lack intrinsic reality is like thisresponse.The distinction between making something so and making knownthat something is so may wrongly suggest to some interpreters that (forwhatever reason) Nagarjuna refrains from asserting a true propositionp, but, although he does not assert that p, his words somehow make pknown nevertheless.This sort of interpretation might seem appropriate if one were influ-enced by the belief that Ciinya means false when applied to statements.One might imagine that Nagarjuna has in mind an action comparableto nudging somebodys elbow and pointing through the window tomake known that it is raining, without incurring the responsibility ofasserting that it is raining. But, if it actually is raining, what is wrongwith asserting the fact? A temptation may then arise to suppose that,according to Madhyamaka, all assertions are false or meaningless, andthat to avoid the Liar Paradox Nagarjuna must contrive to produceeffects in the minds of his hearers by a special technique which doesnot involve asserting anything.

Now, it is quite true that the nudge and the pointing finger canindeed be compared to the technique of reductio ad absurdum, deployedrepeatedly in Nagarjunas Miilamadhyamakakiirikiis - the protagonistdoes not advance an independent argument of his own but points outsomething which his opponent can see without changing his point ofview. However, this similarity does not license us to jump to conclusions.The employer of reductio ad absurdum is not refraining from makingany assertions out of a belief that all assertions whatsoever are falseor meaningless.The temptation to see the Liar Paradox looming like Banquos ghostbehind the verses of the Vigrahavyavartant should be resisted. VerseLXIV and its commentary do not mean that Nagarjuna forgoes the rightto make any assertion.The use of karoti clearly suggests the assumption that subject andobject of a verb must be on the same plane of reality. The statementN attacks M has the effect, if true, of making M an object of attack.Similarly, A denies the svabhava of B entails that As statementsomehow deprives a real B of its pre-existent svabhava.Given the tradition in which he was working, the terms available toNagat-juna to express the notion of denial tended to carry with themthe implication of a relationship between real things (A attacks B).By contrast, Nagatjuna wishes to say that intrinsic reality is of itself


22/26

316 IAN MABBETTnon-existent; it does not require any attack to make it so. He is thereforemerely registering a fact, not depriving something of reality.

Given this interpretation of verse LXIV, it is clear that Nagarjunais not claiming to be saying nothing yet somehow gaining credit forletting truths be known by a sort of content-free nudging and pointing,which would hardly be convincing. He means that the opponents attackfails because his, Nagarjunas, denial of intrinsic reality to all thingsdoes not imply an assertion of the positive existence of any intrinsicreality anywhere.At this point, it is legitimate to conclude that Nagarjunas repudiationof propositions does not commit him to the Liar Paradox, or to an evasionof it by resort to an unusual sense of negating (that is, to interpretationNo. 1 as listed above). The passages so far considered are those whichappear most like self-denying statements; there are others, particularlyin the MzZamadhyamakakiiriks, but they can be dealt with on the linesdeveloped here. In the MziZamadhyamakakiirik2s, Nagarjuna asserts thatall speculative views (dygi) are void, and that even those whose viewis that all is void are hopelessly lost;53 he also says that the Buddhasnever really preached any doctrine. 54 This means that doctrines whichinsist upon any concepts (even voidness itself) as names of separatelyinstantiated real things necessarily miss the point, for things are unreal.So Nagarjuna does not deny that an assertion can capture a truth.On the contrary, some propositions are true, and can be conveyedby utterances. These utterances, however, are not real because, likeall supposedly concrete things and events, they lack intrinsic reality.Whatever is meant by this lack of intrinsic reality, it does not meanthat they are incapable of carrying out the function of asserting thecontent of propositions. (As interpretation No. 5 in the list above wouldclaim, they have the shadow-existence of all void things, which areprovisional, relative, dependent, pratttyasamutpanna.)

ACKNOWLEDGEMENTSThough nobody but the author can be held responsible for anything inthis article, it has benefited considerably from valuable comments andcriticism from D. Seyfort Ruegg and C. Oetke.


23/26

IS THERE A DEVADATIA IN THE HOUSE? 317NOTES

Vigrahavy~arranr XXIa; And just as this statement of mine is void because ofits being devoid of an intrinsic nature, so also are all things void because of theiryg devoid of an intrinsic nature.Vigrahavy&artani verSe XXIX: nasti ca mama pratijhii; see E. H. Johnston and A.Kunst, eds., The Vigruhavy~artani of Nagarjuna, with the authors commentary, inK. Bhattacharya, E. H. Johnston and A. Kunst, The Dialectical Method of Nagarjuna,plhi (Banarsidass) 1978 (reprint o f text with translation of Bhattacharya), p. 29.Vigrahavyiivartani; XXX.4 Vigrahavycivartani LXIII: pratisedhaytii nnham kimcit pratisedhyam asti na cakimcit. Vigrahavymartanr LXIV, LXIVa. The denial does not render entities void of intrinsicreality; it somehow makes known that they are void. Nagarjuna has appeared tosome scholars to be trying to avoid the responsibility of advancing a philosophicalproposition (for propositions do not exist) and yet to produce the same effect as ifhe had done so - eating his cake and having it. The actual meaning of this verseis discussed further below.6 The Eubulides form must be distinguished from the more familiar Epimenidesform (Everything a Cretan says is false), which introduces the type of statementyhose paradoxical character is contingent upon facts external to the sentence.See for example the various papers in the special issue of the Journal ofPhilosophical Logic, vol. 13 No. 2 (1984) on Truth-value Gaps, Truth-value Gluts,and the Paradoxes.* See particularly T. Parsons, Assertion, Denial and the Liar Paradox, Ibid.,rp. 137-152.J. Lyons, Semantics, vol. II, Cambridge (C.U.P.); 1977, pp. 749f., 768f., 802f.citing R. M. Hare, Meaning and Speech Acts, Philosophical Review vol. 79 (1970):to negate the assertion of a proposition is performative negation of its neusticforce.i For brief discussion of Madhyamaka arguments see I. W. Mabbett, An AnnotatedTranslation of Chapter XVI of Candrakirtis Prasannapada, Journal of Ancient IndianHistory, Vol. 15 Pts l-2 (198485), 47-84, and An Annotated Translation of ChaptersXII and XIV of Candrakirtis Prasannapadci, Journal of the Department of Pali,I$iversity of Calcutta, Vol. 4 (1987-88), 113-146.Thomas E. Wood, Niigarjunian Disputations: a philosophical journey through anEdian looking-glass, Honolulu (University of Hawaii Press), 1994.See MMK 24. 18cd; cf. MMK 15.7.I3 What the Madhyamika has disowned, then, is any thesis, assertion or view(drsti) that posits the existence of some kind of bhnva or dharma possessing asvabhziva, and not all philosophical statements, doctrines and theories (dursana)without distinction: D. Seyfort Ruegg, Does the Madhyamika have a Thesis andPhilosophical Position?, Buddhist Logic and Epistemology: Studies in the BuddhistAnalysis of Inference and Language, ed. B. K. Matilal and R. D. Evans, Dordrecht{ydel), 1986, pp. 229-237 at p. 233. Ibid., pp. 232f.See E. Conze, Buddhism, its Essence and Development, Oxford (Cassirer), 1957,p. 125; Jaspers, cited by Oetke, Rationalismus, pp. 2f.; S. Betty, NagarjunasMasterpiece - Logical, Mystical, Both or Neither ?, Philosophy East and West, vol.33 NO. 2 (1983), pp. 123-138 at pp. 133-135; K. Bhattacharya, ed. and tr., TheDialectical Method of Nagiirjuna (Vigrahavy~artani), Delhi (Motilal Banarsidass)1978, p. 24, n. 3.I6 D. Ruegg, On the Thesis and Assertion in the Madhyamaka/dBu ma, in


24/26

318 IAN MABBETTContributions on i?betan and Buddhist Religion and Philosophy, ed. E. Steinkellnerand H. Tauscher, Wien (Arbeitskreis fur Tibetische und Buddhistische Studien,Universitiit Wien), 1983, pp. 205-241 at pp. 211f., n. 17.I7 Ruegg, Thesis and Assertion, esp. p. 216, and idem, Does the Madhyamikahave a Thesis and Philosophical Position?, in Buddhist Logic and Epistemology:Studies in the Buddhist Analysis of Inference and Language, ed. B. K. Matilal andRobert D. Evans, Dordrecht/Boston/Lancaster/Tokyo (Reidel), 1986, pp. 229-237,yssp. at p. 233.L. de La Vallee Poussin, Le Bouddhisme, 1909, p. 197; cited by Oetke, Rational-ismus, p. 2.I9 T. R. V. Murti, The Central Philosophy of Buddhism, London (Allen and Unwin),1955, pp. 136, 145, 155, 161.i Language, Logic and Reality, p. 304.

B. K. Matilal, Perception: An Essay on Classical Indian Theories of Knowledge,Oxford (Clarendon), 1986, p. 48.** Ibid., p. 66. Matilal cites J. Searle here for the concept of illocutionary negation:gs rejection of views is illocutionary, not a self-invalidating assertion.Ibid., p. 90.24 Idem, Is prasanga a Form of Deconstruction ?, Journal of Indian Philosophy,;501. 20 (1922), pp. 345-362 at p. 361.Stafford Betty, Nagarjunas Masterpiece - Logical, Mystical, Both or Neither?,ghilosophy East and West, vol. 33 No. 2 (1983) pp. 123-138, esp. p. 128f.David Loy, How Not to Criticize Naggarjuna: a Response to L. Stafford, Ibid.,r701.34 No. 4 (1984), pp. 437-450 at p. 439.Ruegg, Thesis and Assertion, p. 238; Ruegg, Does the madhyamika have at$esis?, pp. 235f.Kartikeya Patel, The Paradox of Negation, Asian Philosophy vol. 4 No. 1 (1994),p. 17-32.Ruegg, Does the madhyamika have a thesis? and On the Thesis and Assertion.

3o Ruegg, Does the madhyamika have a thesis?, at p. 232. See also Ruegg, Onthe Thesis and Assertion, and Galloway, Some Logical Issues, pp. 9, 26f. n. 5,where the attribution of a distinction to Ruegg does not seem exact.ti Ruegg, On the Thesis and Assertion, p. 212 n. 17.Rueggs argument is addressed to the history of madhyamika thought as a whole,not specifically to the Vigrahavyavartani; The evidence cited for the currency ofa narrow usage of pratijiia consists of two sets of passages in the Prasannapadaof Candrakirti, one set described as very clearly referring to a propositional thesispostulating an entity, and the other using the term in the more general or neutralsense (Ruegg, On the Thesis and Assertion, p. 213).It will be argued here that, in the Vigrahavyavartanl; Naggarjunas claims tobe advancing no pratijria can be understood by making the distinction betweenpropositions as things, bhava (which have no svabhava and are ultimately unreal)and as abstract views or teachings, darsana, vada (which Madhyamaka advanced).Nagarjuna could claim to have no pratijfia in the first sense, without prejudice tohis claim to teach truths in the second sense.33 See R. Robinson, Some Logical Aspects of Nagatjunas System, PhilosophyEast and West, vol. VI No. 4 (1957) pp. 291-308, especially p. 302; cf. B. Galloway,Some Logical Issues in Madhyamaka thought, Journal of Indian Philosophy, vol.17 (1989), pp. l-35.34 T. E. Wood, Nagarjunian Disputations: a philosophical journey through an Indian~&g-glass, Honolulu (University of Hawaii Press), 1994.C. Oetke, Die Metaphysische Lehre Nagarjunas, Conceptus, vol. 56 (1988),pp. 47-64; idem, Remarks on the Interpretation of Nagarjunas Philosophy, Journal


25/26

IS THEXE A DEVADATTA IN THE HOUSE? 319ofhdiun Philosophy, vol. 19 (1991), pp. 315-323 at p. 317: there is much to besaid for the interpretation of Nagarjunas teaching according to which on the levelgf the highest truth there is nothing of any kind.urah-kanthausfha-jihva-dantamiila-t~lu-nnsik~-m~rdha-prabh~i~u yatnesu.37 On the complex relative chronology of the Mulamadhyamakakib-ikiis and theNyaya Sun-as see J. Bronkhorst, Nagarjuna and the Naiyayikas, Journal of IndianPhilosophy, vol. 13 (1985), pp. 107-132. Nagarjuna knew an early version of theNyaya Sutras; a later version of the latter knew the Mtilamadhyamakakatikiis. Themutual accusations of self-contradiction advanced by Nagarjuna and his Naiyayikaopponents are discussed by Roy W. Perrett in Self-Refutation in Indian Philosophy,{turnal of Indian Philosophy, vol. 12 NO. 3 (1984), pp. 237-263; see pp. 250-255.nivartuyati can be translated as deny or repel. As noted above, Nagarjunaregards ideas as objects on the same plane of reality or unreality as physical objects.Therefore the objection that the voidness thesis merely voids itself has to be per-ceived in quasi-physical terms: the voidness thesis is alleged to deprive itself ofthe power to void things. Hence repel catches the sense better, though not usuallyapplied in English to a refuted argument. (This point is owed to C. Oetke, personal;~mmunication.)Proposition hem translates pratijria, on the meanings of which vide supra. Inverse I, the Madhyamaka dictum is appropriately called a vacana, utterance, becausethe opponent is pointing to its lack of reality as a specific and independently realthing. In verse II, the argument is that, if the utterance is not void, it contradicts themeaning or content of an abstract proposition (hence prutijfia), but Nagarjuna doesnot think of pratijiia purely as an abstract meaning; he thinks of it as a weapon orvictim of attack, like a physical object.a Betty, lot. cit., p. 128, summarizes verses I-II and Nagarjunas reply. Cf. Loy,How not to criticize Nagarjuna, p. 439, claiming that in fact the opponent cannotcriticize Nagarjunas claim without accepting it.4 Outside the Vigrahavyavartani; or at least outside Madhyamaka, sadhyasamaclearly has a meaning very like begging the question, and this interpretation isfollowed here. Nagarjunas use of the expression is however often taken in anothersense, and the issue is not entirely clear. See D. Seyfort Ruegg, Thesis and Asser-tion, op. cit., p. 210; B. K. Matilal, Logic, Language and Reality, Delhi (MotilalBanarsidass), 1985, pp. 47f.; K. Bhattacharya, Note on the interpretation of theterm sadhyasama in Madhyamaka texts, Journal of Indian Philosophy vol. 2 Nos.314 (1974), pp. 225-30. At Miilamadhyamakiirikiis 4.8-9 the expression can beunderstood to mean something like begging the question, and Candrakirti takes itin this way. However, the interpretation of these verses is problematic.42 Cf. Loy, lot. cit. p. 439; Mtilamadhyamakarikas IV.8-9.43 For the gloss on this verse by mKhas grub rje, see Ruegg, On the Thesisand Assertion, p. 219. On this verse see also Oetke, Rationalismus, pp. 22-26(Nagarjuna makes no proposition, not as a matter of contingent fact, but becausein ultimate truth all things are non-existent and therefore there are no such thingsas utterances of propositions by anybody); idem, Remarks, p. 320, Matilal, Isprasariga a Form of Deconstruction? at p. 355.44 The terms used here, which are scarcely susceptible of a literal translation, haveclear echoes from PrajiZipatamita literature; see K. Bhattacharya et. al., The DialecticalMethod of Nagarjuna, op. cit., pp. 23f (note). The rather monistic language used atthis point lends itself somewhat to a Vedanta-type concept of an absolute (as indeed$hattacharya himself takes it), but this is not the only possible interpretation.C. Oetke has emphasized that the reason why Nagarjuna has no proposition is thesame as the reason why nobody has: Remarks on the Interpretation of NagarjunasPhilosophy, Journal of Indian Philosophy, vol. 19 (1991), pp. 315-323 at p. 320.


26/26

320 IAN MABBEIT46 On verse XXX see Ruegg, On the Thesis and Assertion, pp. 207, 220, wherediscrepant constructions appear; the commentary to the text shows that in factpratyaksiidibhir arthais belongs in the protasis.47 LIXa: na hi vayam nama sadbhtitam iti brumah; We do not say that it is a realt$ng.On Nyaya negation, see B. K. Matilal, Reference and Existence in Nyaya andBuddhist Logic, Journal of Indian Philosophy, vol. 1 (1970) pp. 83-110; idem,Logic, Language and Reality, Delhi (Motilal Banarsidass), 1985, pp. 77-112; DineshChandra Guha, Navya Nyaya System of Logic (Some Basic Theories and Techniques),Vamnasi, 1968, pp. 112f.; K. Bhattacharya, The Dialectical Method of Niigarjuna,g. 13.. . . Thus pomt is substantially owed to comments by C. Oetke, personal communi-cation.So Vigrahavyavartani LXIII: pratisedhayami naham kimcit pratisedhyam asti naca kimcit. On the gloss on this verse by mKhas grub rje, see Ruegg, Thesis andAssertion, p. 221 (read 63ab for 64ab): there is no negandum established byself-nature . . . an unreal . . . may&like negandum and negator are accepted. Sarikrtyayanas reading is: kintv asatsvabhavo bhavanam asatsvabhavanam iti:.IIhnston and Kunst, p. 48 n. 12.For the var. let. see Johnston and Kunst, p. 49 n. 1.53 Mtilamadhyamakak~ikas X11.8: Stinyata sarvadrstinam prokta nihsaranam jinaih1 yesam tu Sunyatadrsfis tan asadhyan babhasire.54 Mulamadhyamakakarikas XXV.24 na kva tit kasya tit kascid dharmo buddhenadesitah. See also ibid. IV.%9; XVIII.6, 8; XXIV.13.

Department of HistoryMonash University, Clayton, VictoriaAustralia

Mabbett- Devadatta in House Nagarjuna VV Liar

Documents