Jump to ContentJump to Main Navigation
Nagarjuna's MadhyamakaA Philosophical Introduction$

Jan Westerhoff

Print publication date: 2009

Print ISBN-13: 9780195375213

Published to Oxford Scholarship Online: May 2009

DOI: 10.1093/acprof:oso/9780195375213.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 21 June 2018

The Role of Negation in Nāgārjuna’s Arguments

The Role of Negation in Nāgārjuna’s Arguments

(p.53) 3 The Role of Negation in Nāgārjuna’s Arguments
Nagarjuna's Madhyamaka

Jan Westerhoff (Contributor Webpage)

Oxford University Press

Abstract and Keywords

This chapter deals with the theory of negation described in the non-Buddhist Nāgārjuna school of thought. Since many of Nāgārjuna’s arguments are formulated against the background of this theory and set out to refute it it is important to gain a clear understanding of the Nāgārjuna account in order to grasp the point of Nāgārjuna’s arguments. The chapter deals in particular with the problem of negating non-denoting terms (such as ‘unicorn’) and the problem of the temporal relation between the negation and the negated object.

Keywords:   negation, Nāgārjuna, Nāgārjunaa, non-denoting terms, temporal relations

Chapter 2 acquainted the reader with the main objective of Madhyamaka thought, that is the rejection of svabhāva. Before we can discuss the further ramifications of this idea in Nāgārjuna’s philosophy, it is necessary to discuss some formal aspects of his arguments which those acquainted primarily with Western philosophical literature might find puzzling. They arise mainly from specific logical and methodological considerations connected with the concept of negation which were widespread in Indian philosophy but are not always shared by the Western notion of negation, which is derived primarily from formal logic. What makes the issue particularly intricate is the fact that there exists a tension between some presuppositions of the traditional Indian account of negation and the contents of Nāgārjuna’s philosophical views, so that Nāgārjuna sometimes sees himself challenged to adapt these presuppositions in order to formulate his philosophical position.

In this chapter I will discuss Nāgārjuna’s view of the standard conception of negation as presented in the Nyāya system. chapter 4 will discuss a specific form of argument, the catuṣkoṭi or tetralemma, which is frequently employed in Nāgārjuna’s writings and essentially involves single and iterated negations.

Nāgārjuna’s central argumentative aim is to develop a philosophical theory which does not have recourse to the notion of substance or svabhāva. His main strategy is to examine all the possible ways in (p.54) which particular phenomena (such as physical objects, causation, the self, language, etc.) could be thought to exist with svabhāva, and to conclude that on close inspection none of these are satisfactory. It then remains to conclude that the phenomenon in question does not exist with svabhāva. Since many of Nāgārjuna’s conclusions are therefore negative ones, it is essential to gain a clear understanding of the role of negation in his philosophical system.

Doing so is more difficult than it may sound initially, especially because Nāgārjuna’s discussion of these matters, the greatest part of which is to be found in the VV and, to a lesser extent, in the VP, is formulated against the background of the Nyāya theory of negation. This differs significantly from accounts of negation with which those acquainted primarily with Western philosophical discussions are likely to be familiar.

3.1. Nyāya Theory of Negation

The philosophical system known as “Nyāya” incorporates a comprehensive theory of logic and epistemology which proved to be extremely important in Indian intellectual history, influencing not only different kinds of philosophical inquiry, but also such disciplines as linguistics, poetics, rhetoric, and law.1 The system is based on the Nyāyasūtra attributed to Gautama (also known as Akṣapāda). There is little agreement on when the sūtra was composed; the dates proposed range from the sixth century B.C. up to the second century A.D.2 It is, however, relatively clear that it achieved the form in which it has been transmitted to us around the time of Nāgārjuna and might even have in parts been composed as a reply to Nāgārjuna’s arguments.3 We must therefore keep in mind that when we refer to the relations between Nāgārjuna’s Madhyamaka and Nyāya we are dealing with a very early phase of the latter. At Nāgārjuna’s time none of the long sequence of works on Nyāya,4 including Vātsyāyana’s Bhāṣya, had been written yet.5

(p.55) In order to understand the Nyāya theory of negation, we have to note that in the underlying ontological system, properties are seen as separate entities over and above the substrata in which they inhere.6 Now the Naiyāyika regards the absence (abhāva) of a property as a category in its own right (padārtha),7 as something that can equally be possessed by a substratum.8 The referent of a negative statement such as “there is no pot in the house” is therefore regarded as the qualification of the house by an absence, namely the absence of a pot.9 Judgments, whether they are affirmative (“There is a pot in the house”) or negative (“There is no pot in the house”) involve a qualification or an attribution, which can be either an attribution of a presence or an attribution of an absence. As such the attribution can be construed as either affirmative or negative. Whether it is expressed by an affirmative or a negative statement is then a question about how the judgment of that attribution is formulated in ordinary language, not a fact about the attribution itself.10

According to the Naiyāyika account, a judgment is correct if it combines some parts of reality in the way in which they are indeed combined (such as the house and the absence of the pot, if there is no pot in the house), and it is erroneous if it combines elements from reality in a way in which they are not combined (such as the house and the absence of the pot if there is a pot in the house).11 Error will therefore always arise from the way elements are combined in judgments, never from simple perceptions. For the Naiyāyika “a simple, noncomplex property can never be empty.”12

(p.56) 3.2. Negation and Nondenoting Terms

This concept leaves us with a substantial difficulty if we want to deny the existence of certain entities, for the Naiyāyika faces the very problem Quine observes at the beginning of “On what there is”:13

When I try to formulate our difference of opinion, I seem to be in a predicament. I cannot admit that there are some things which [my opponent] McX countenances and I do not, for in admitting that there are such things I should be contradicting my own rejection of them.

Suppose we want to say that Pegasus does not exist (or that every place is characterized by an absence of Pegasus). Such a judgment would intuitively be regarded as true, but for the Naiyāyika a true judgment has to combine elements of reality in the right way. But Pegasus is not an element of reality, since he does not exist. Matilal observes:14

The property of Pegasus-ness thus arrived at would be, according to Nyāya, unexampled or fictitious because it has no locus to occur in, that is, no locus possesses this property. In such cases, Nyāya asserts that we cannot even say that such a property is absent or does not occur somewhere… . Thus, if a sentence is said to express an absence of such an unexampled property, it becomes nonsensical.

Now of course this does not mean that in adopting the Nyāya semantics we have no way of saying that a possible entity (like Pegasus) or an impossible one (like the round square) does not exist. If this was indeed the case it would not be a very satisfactory theory to begin with. What we have to do is to rephrase the statement to make evident that the source of the vacuity of such properties is the combination of more basic properties each of which exists in reality but which are not thus combined in the world. So the Naiyāyika can make the meaningful assertion that wingedness and horseness are never combined in the same animal, and that roundness and squareness are never combined in the same figure.15 All the properties referred to now have referents in reality, so the Naiyāyika’s semantics is taken care of.16 There is an obvious similarity between this approach and the Russellian procedure of replacing a non-denoting term by a definite description.17

(p.57) This view of non-denoting terms is just a reflection of the fact that for the Naiyāyika, language must hook up with the world at some fundamental level via a denotation relation. Even if there are non-denoting terms in our language, they can exist only parasitic on denoting terms. Simple designators are therefore guaranteed to refer, while complex designators may or may not do so. This of course means that according to Nyāya theory, negative statements involving simple designators (statements of absence of some entity) can only ever be statements of a local absence and will always entail the presence of that entity somewhere else.18 In his Vārttika on Vātsyāyana’s Bhāṣya on 2.1.12 of the NS Uddyotakara notes:19

[W]hen the word “jar” is coordinated with the term “does not exist” it does not convey the non-existence of the jar; all that it does is to deny, either the [spatial] connection of the jar with the house or its specification [as located at] a particular point in time.

Since the statement “There is no pot in the house” or “There is an absence of a pot in the house” is meaningful only if the pot referred to does indeed exist, it must be present somewhere else. It would be nonsensical if there were no pots at all, at least if “pot” is regarded as a simple designator. Vātsyāyana raises this objection in his commentary on the passage from the NS 2.1.11, where he argues against the opponent’s attempt to deny the existence of means of knowledge:20

If you want to deny the existence [of the means of knowledge, this denial] implies their existence, and the refutation of [means of knowledge like] perception and so forth is not accomplished.

Phaṇibhūṣana’s subcommentary elucidates this point by adding:21

The very attempt to deny their existence presupposes the admission of their existence inasmuch as there is no sense in demolishing the possible existence of something which has no existence at all, just as it is impossible to smash with a stick the jar which does not exist.

(p.58) The idea that simple designators have to have a denotation seems to possess some intuitive plausibility. Suppose I present you with some simple, non-denoting name such as “Hopzik.” You could not be taught the meaning of “Hopzik” by ostension (since there are none), nor could I give you an analysis in terms of other properties (since it is a simple designator). But then the problem occurs of how you could make sense of any sentence containing the term “Hopzik,” including statements such as “Hopzik does not exist.” We do not have any idea what this negation means, since we do not have any positive notion of the entity being negated.

Having taken account of this background, it is understandable that the issue of non-denoting terms is raised by the Naiyāyika opponent against the central philosophical thesis of Madhyamaka that there is no svabhāva. He argues that if the Madhyamaka claim was true and if there was indeed no svabhāva, then the claim would be nonsensical. For if a negative statement about svabhāva was interpreted along the same lines as a negative statement about a pot in the house,22 we would have to hold that the existent svabhāva stood in an absence relation to the world in the same way as the pot stood in an absence relation to the house. But if we understand the statement in this way, svabhāva exists after all and so the Madhyamaka thesis must be false. The statement “there is no svabhāva” has to be either false or nonsensical, since “there is no name without referent.”23

A later manifestation of the same difficulty can be found in certain problems connected with formal reasoning.24 According to the Nyāya theory, a formally set out “inference for oneself” (svārthānumāna) establishing that the subject (pakṣa) has the qualifying property (sādhya) must provide both an agreeing and a disagreeing example (udāharaṇa).25 Thus, in order to establish the thesis that all white things are colored we need both an “agreeing” example of a subject having the qualifying property (such as a conch shell, which is both white and colored), as well as a “disagreeing” example of the contraposed version (“whatever is not colored is not white,” space, which is neither, being a case in point).

But we realize that this reasoning leads to a problem if we assert a universal thesis such as “all things lack svabhāva” or “all things are momentary.” In this case the disagreeing example would have to be an instance of some object that (p.59) has svabhāva or is not momentary. Since we want to establish that there are no such things, the term has to be empty. But if the occurrence of an empty term renders a statement meaningless, as the Naiyāyika asserts, it immediately follows that the two theses cannot be established. If, on the other hand, the two theses are meaningful (as they appear to be), then they must assert absences of existent things, and svabhāva or the property of permanence must exist. So once again we are faced with the unenviable choice between falsity and nonsense.

Nāgārjuna suggests a variety of possible replies to the Nyāya difficulty of non-denoting terms. First of all he remarks:26

To one who says that the name is sadbhūta you would have to reply: “There is svabhāva.”… [However,] since things have no svabhāva, that name also lacks svabhāva. Because of this it is empty, and, being empty, it is not sadbhūta.

Bhattacharya27 here translates sadbhūta as “existent.” This does not strike me as a very fortunate rendering, primarily because Nāgārjuna does not want to claim that emptiness entails a lack of existence. It is evident that what Nāgārjuna wants to say in the first sentence is that if the Nyāya account was indeed correct (and each simple term in a negative statement had to denote an existent object), svabhāva would exist. It therefore seems plausible that saying that a name is sadbhūta is supposed to mean not that the name exists (something that neither Nāgārjuna nor his opponent denies) but that it functions in accordance with Nyāya theory: that each non-complex term is hooked up with a designated object in the real world.

But this is exactly what the proponent of emptiness denies. For the Mādhyamika, a satisfactory semantics cannot consist in an objectively existent reference relation which links the terms of our language to an objectively existent world. He will argue that both the question of how the world is sliced up into individual entities and the question of how these entities link up with the parts of language are to be settled by convention. There is no “ready-made world” of simples out there which could provide the semantic foundation for the simple terms of our language.28

Nāgārjuna therefore argues in this passage that the Naiyāyika criticism is justified only if one is antecedently convinced of the Nyāya picture of semantics. This, however, is something that the Mādhyamika does not want to share.

(p.60) We might wonder at this point why a relatively obvious reply to the Naiyāyika worries about statements like “there are no objects existing with svabhāva” is not made by the Mādhyamika. This reply consists in arguing that svabhāva is not a simple designator. As we saw earlier it is possible to assert that there are no unicorns even on Nyāya terms. This is done by rephrasing the statement as an assertion about real entities (such as hornedness and horseness) that do not occur together. We could now similarly break up the property “existing with svabhāva” into its simpler components (such as not depending causally on other phenomena, not depending notionally on other phenomena, and so forth) and argue that since all these dependence relations exist, all we are asserting by a statement denying svabhāva is that there is no object that is qualified by the absence of all these dependence relations at the same time.

This reply is certainly adequate for answering the Naiyāyika worry, but it is hardly a position a Mādhyamika would want to adopt without further qualification. For, according to the Naiyāyika’s interpretation of this answer, there is still a world of objectively existent simple properties which the simple terms in our language refer to in a way that is independent of linguistic conventions. While the statement “there is no svabhāva” is thus at least rendered comprehensible to the Naiyāyika, it is done so at the price of accepting a view of semantics fundamentally at odds with the one the Mādhyamika is arguing for. For this reason trying to establish that svabhāva is a complex designator is not a very satisfactory response by Madhyamaka standards.

Within the context of discussing the problem of negating a non-denoting term, Nāgārjuna also discusses a somewhat curious objection raised in VV 12. The opponent asks about the point of negating a non-existent object since “the negation of a non-existent, such as the coolness of fire or the burning of water, is established without words.”29 This passage assumes that there are two kinds of negation for the opponent: those established without words, and those established with words. If one looks at the choice of examples used, it seems that members of the former group include negations of an essential property of an object, such as the heat of fire. What kinds of negation are included in the latter group is not quite clear. A reasonable assumption is to include negative contingent statements in here (such as “There is no pot in the house”). Now the opponent does not want to say that the absence of the pot in the house is brought about by the assertion “there is no pot in the house”: just saying it is so does not make it so.

There are two different ways of understanding what the opponent could mean here. First, note that the Naiyāyika does not draw any fundamental (p.61) distinction between the judgments expressed by affirmative and negative statements. As Matilal observes:30

All determinate cognitions or judgmental cognitions (savikalpa jñāna), which can be very well regarded as the counterpart of statements, involve a qualification or attribution, and such an attribution cannot be construed as either affirmative or negative.

The affirmative–negative distinction is not one the Naiyāyika regards as ontologically fundamental. Whether a particular statement is affirmative or negative just mirrors the way the particular qualification or attribution is expressed in ordinary language. What is meant therefore by saying that a negation such as “There is no pot in the house” is established with words is that it is only by the force of language that a negative meaning is expressed. The world itself contains no negations, only presences and absences of different kinds. It is only the words that bring negations into existence.

The absence of a non-existent entity (such as the absence of the opposite of an essential quality, like the absence of coolness in fire) can be regarded as a “negation established without words” (vacanād vinā siddhaḥ pratiṣedhaḥ), since it cannot rely on language for its expression as a negation, because the Nyāya restrictions on non-denoting terms render the phrase “absence of coolness of fire” meaningless. It is of course the case that there is no coolness in fire, but, for the sake of argument assuming that “coolness of fire” is a simple designator, this is not something that can be expressed in language, nor is it something that indeed needs to be expressed. We might, after all, think there is a pot in the house when there is none, but who would think there is coolness to be found in fire?

Second, we can understand the claim that some negations are established with words while some are established without not as referring to the negation itself, but rather as referring to the corresponding cognition of the negation created in our mind.31 To establish a cognition like the one that there is no pot in the house, we generally need language (since we are not able to inspect all the parts of a house at once). Other cognitions, however, like those that fire is not cool and water is not burning, are established without linguistic mediation. Our acquaintance with fire or water directly acquaints us with these essential properties of them. There is therefore no need to assert the absence of properties that are opposed to the essential nature of things in language. Because the (p.62) claim concerning the absence of svabhāva is of the very same nature, it is therefore not obvious what the point of asserting this negation is supposed to be.

Nāgārjuna replies by saying that the purpose of a statement such as “there is no svabhāva” is to point out the absence of svabhāva, not to create the absence of something which is in fact there.32 He uses the example of saying “Devadatta is not in the house” when he is in fact not there—this statement obviously just reports the absence of Devadatta and does not cause him not to be in the house.33 We noted above that it is quite unlikely that the opponent should hold such a peculiar position, unless what is created is supposed to be the cognition of the absence, rather than the absence itself. In this case Nāgārjuna could be interpreted as saying that both kinds of negation, those involving essential and nonessential properties, have to be linguistically mediated to be cognized by us.

What would be replied to our first interpretation of the opponent’s worry, that is, the position that the absence of something necessarily non-existent (such as the coolness of fire, or the svabhāva of things) cannot and need not be expressed? We have already seen that the impossibility of expressing such negations depends on the peculiar nature of the Nyāya semantics which the Madhyamaka will not want to accept. And concerning the need to express this negation, the Madhyamaka will argue that while nobody in his right mind will think fire to be cool (and could therefore benefit from having this pointed out to him), the belief in the svabhāva of things is extremely widespread, and realizing its falsity is one of the essential preconditions of liberation. Thinking that there is svabhāva in things is like thinking there to be real water in a mirage. Deceived people who are likely to run toward the mirage to quench their thirst will benefit from getting to know that there is no real water there, just as ordinary people will benefit from learning that things exist without svabhāva.34

Nevertheless we might think that when we negate water perceived in a mirage, there is still the person perceiving the mirage, its perception, and the perceived object (i.e., the mirage), as well as the person doing the negating (us), the negation, and the object negated (namely real water in the mirage).35 But (p.63) if this is the case, the Naiyāyika will argue we are facing the familiar problem again: if the thing we negate (the mirage, svabhāva) does exist after all, then the statement asserting its negation is obviously a falsity. If, on the other hand none of them exist, if there is no perceiver, perception, and perceived object, no negator, negation, and object to be negated,36 then the Madhyamaka argument vanishes altogether and the existence of svabhāva is established by default.37

What we have to say here (and what Nāgārjuna in fact does say)38 is that all these things can exist without existing in quite the way the Naiyāyika supposes. As we have seen for the Naiyāyika, each of the different entities referred to by the simple designators in a negative statement (a statement of absence) has to be real for the statement to be meaningful. The Mādhyamika, however, thinks that an unreal entity, such as the water in a mirage or the appearance of svabhāva, can very well be the object of an (erroneous) cognitive state and also be able to be referred to in a true sentence. From the fact that “something has become the ‘object’ (viṣaya) of a cognitive state, it does not follow that it must have been causally related to the production of that cognitive state.”39 This is due to the fact that for the Mādhyamika the source of error is not located exclusively in the erroneous combination of individually existing properties, as the Naiyāyika assumes.40 The example of the mirage presents us with the case of a simple yet erroneous perception. As long as we assume that the object of perception and the object of negation are all dependently arisen objects rather than entities existing in their own right, we can deny their existence without antecedently having to regard them as real.41

Even though the term “the water in the mirage” is non-denoting, since there is no water in the mirage, there is still something created by the interplay of our senses, light, and heat on which the presence of water is superimposed, which we can subsequently deny. Similarly our language and general cognitive habits can, the Mādhyamika argues, create the unreal superimposition (samāropa) of svabhāva which Nāgārjuna’s arguments set out to refute.

(p.64) 3.3. Negation and Temporal Relations

Apart from worries about negative statements involving non-denoting terms, the second main difficulty to do with negation raised by Nāgārjuna’s opponent concerns the possible temporal relation between a negation and the object negated.42 This is a form of argument which we encounter frequently in Nāgārjuna’s works, the most prominent examples being in the discussion of the relation between means of knowledge and object known, and between cause and effect.

The worry of the opponent is that there is a general difficulty with negative statements (including the Mādhyamika’s assertion that there is no object with svabhāva). He argues that the negation can exist neither before, after, nor at the same time as the object of negation, and therefore cannot possibly exist. Now this might strike us as a strange position to maintain. Since we usually regard negation as a logical relation, temporal considerations seem to be wholly irrelevant, in the same way as there is no point in asking which numbers in a mathematical equation are there first.

In order to see the point at issue here, we have to note that for the Naiyāyika the negation is the instrument making known a particular absence of a quality in some substratum. This making known is obviously a causal process,43 so that it is clear that the Naiyāyika worries here just stem from an application of Nāgārjuna’s criticism of causation44 to epistemology. The Naiyāyika will argue that if causes and effects cannot exist standing in any of the three temporal relations (as the Mādhyamika sets out to show), then this must also apply to epistemic causes and effects,45 and thus also to negations, which constitute one particular kind of epistemic cause. Therefore, if we take the Madhyamaka view of causation seriously, we have a problem with establishing the negative thesis that there are no objects with svabhāva.

The argument itself proceeds in the expected manner. The opponent argues that the negation cannot exist before the object to be negated, because then there would be a negation without anything negated. More worryingly, if the negated object does not (yet) exist, what is the point of negating it? Nor could the negation exist after the object of negation, for what is the point of negating something existing? We also have to note that it is only the causal interpretation (p.65) of negation in this context which allows us to make sense of the opponent’s criticism of the simultaneous existence of the negation and its object. We might think that this was indeed a satisfactory way of thinking of the two (in the same way in which we might think that all the numbers in a mathematical equation exist at the same time). But in considering negation in causal terms we face the problem that “the negation is not the cause of the object known by negation, nor is the object known by negation the cause of the negation.”46 As is illustrated by the familiar analogy of the two horns of a cow47 which do not cause each other, in the case of simultaneously existing cause and effect we would have a problem in establishing which is which, since the conceptual distinction between cause and effect is drawn in terms of temporal priority.

There are various ways in which one can respond to this problem. In VV 69 Nāgārjuna tries to turn the tables on his opponent.48 As we saw, for the Naiyāyika the existence of a negation is equivalent to the existence of an object of negation, that is, of an object whose absence in a particular substratum could be asserted. But if there is no negation, as the Naiyāyika has just been trying to argue, there is also no object whose absence can be asserted, and therefore it follows on the Naiyāyika’s own terms that there is no svabhāva.49

A more general way of replying is to point out that in the same way as we can still talk about causal relations even if the realist’s picture of causality turns out to be unsatisfactory, the fact that some epistemic process cannot be made sense of in Nyāya terms does not mean it could not be made sense of at all. After all, what Nāgārjuna criticizes in his analysis of causation is the conception of causes and effects as mutually independent, objectively existing entities. Similarly, in his treatment of epistemology he sets out to refute the conception that being a means of knowledge is an essential property of some cognitive processes. If we do not make this presupposition, however, there is nothing intrinsically problematic with the existence of causes and effects in general, and also not with causes and effects in epistemic processes. (p.66)


(1.) Matilal (1968: 21); Potter (1970–2003: II, 1–3); Guha (1979: 1–2). A concise summary of the assumptions underlying the Nyāya system is given in Matilal (1986: 5–6).

(2.) Potter (1970–2003: II, 4). Jha (1939: viii) even suggests a date of composition as late as the sixth century A.D. A comprehensive account of the history of the composition of the NS is given in Meuthrath (1996).

(3.) Bronkhorst (1985) argues the greatest part of the NS existed before Nāgārjuna and was known to him, and that some parts were added later in response to Madhyamaka objections. Bronkhorst’s conclusions have been severely criticized by Oetke in his (1991) and (1997). For a reply see Bronkhorst (1993).

(4.) Potter (1970–2003: II, 9–12).

(5.) For more details on the relation between the VV and the NS see Meuthrath (1999).

(6.) Matilal (1968: 16). The Nyāya system of logic and epistemology usually relies on the Vaiśeṣika ontology. The association between the two systems is so close that one often refers to them jointly as Nyāya-Vaiśeṣika. Opinions differ on how the two systems came to be associated, and even on whether we are to speak of two systems rather than one. See Potter (1970–2003: II, 12–13); Bronkhorst (1985: 123–124).

(7.) Chatterjee (1939: 166–168); Sharma (1970: 11–12).

(8.) We will not go into the reasons offered in defense of this assumption. Sharma (1970: 3–11) argues that the conception of liberation (mokṣa or apavarga) espoused by the Nyāya-Vaiśeṣikas made it “imperative for this system to posit Absence as an ultimately real entity (padārtha)” (6), so that this ontological point also acquired soteriological significance.

(9.) Matilal (1968: 3).

(10.) Matilal (1968: 92–93).

(11.) Matilal (1970: 95). As such there is a close similarity between this view and the familiar correspondence account of truth we find, for example, in a semantics based on states of affairs. There the sentence “there is a pot in the house” is regarded as true iff the referents of the constituents of the sentence (the pot, the house, the “inside of” relation) are arranged in structurally the same way in the state of affairs in which the constituents are arranged in the sentence. The main difference is, of course, that this view does not accord any ontological status to absences, as the Nyāya theory does.

(12.) Matilal (1970: 96).

(13.) Quine (1953: 1).

(14.) Matilal (1968: 154–155).

(15.) See Uddyotakara’s commentary on the Nyāyasūtrabhāṣya 3.1.1 Chakravarti (1982: 232–233).

(16.) Matilal (1968: 9, 23).

(17.) Matilal (1970: 85), Chakravarti (1982: 211–212).

(18.) Phaṇibhūṣaṇa (Chattopadhyaya and Gangopadhyaya 1968: II:28) remarks that “there is no sense in denying the absolutely non-existent like a flower imagined to blossom in the sky. The denial of something can have sense only when its existence is admitted somewhere else, e.g., to say that there is no jar in the room means that it exists elsewhere.”

(19.) na hy ayaṃ nāsti nāsamānādhikaraṇo ghaṭādiśabdo ghaṭābhāvaṃ pratipādayati api tu gehaghaṭasaṃyogaṃ vā kālaviśeṣaṃ vā […] pratiṣedhati. Gautama (1887: 192: 20–22); Jha (1984: II: 623).

(20.) tad yadi sambhavo nivartyate sati sambhave pratyakṣādīnāṃ pratiṣedhānupapatiḥ. Nyaya-Tarkatirtha and Tarkatirtha (1985: 425: 2–3).

(21.) Chattopadhyaya and Gangopadhyaya (1968: 2: 26).

(22.) VV 11.

(23.) nāma hi nirvastukaṃ nāsti. VV 9.

(24.) This is discussed in Matilal’s analysis of part of the Ātmatattvaviveka by Udayana (10–11th century A.D.) in (1970).

(25.) For a concise summary of the Nyāya theory of inference see, Potter (1970–2003: II, 179–208).

(26.) yo nāmātra sadbhūtaṃ brūyāt sasvabhāva iti evam bhavatā prativaktavyaḥ syāt | […] tad api hi bhāvasvabhāvasya abhāvān nāma niḥsvabhāvam tasmāc chhūnyam śunyatvād asadbhūtam. VV(S) 76:16–77:2.

(27.) Bhattacharya et al. (1978: 128).

(28.) See chapter 9 for further discussion of this point.

(29.) asato hi vacanādvinā siddhaḥ pratiṣedhaḥ tadyathāgneḥ śaityasya apāmauṣṇyasya. VV(S) 51:6–7.

(30.) Matilal (1968: 92).

(31.) As argued below on page 64 this is also the most satisfactory way to understand the opponent’s worry about the temporal relations between negation and negated object.

(32.) VV 64.

(33.) A similar point is made by Uddyotakara in his Vārttika on Vātsyāyana’s Bhāṣya on NS 2, 1, 11: “negation does not have the power to make an existing thing otherwise [i.e., non-existent]. Because it makes something known, it does not cause the existence of something to cease; [therefore] this negation [too] makes something known and does not cause the existence of something to cease.” na ca pratiṣedhasyaitat sāmarthyaṃ yad vidyamānaṃ padārtham anyathā kuryat jñāpakatvāc ca na saṃbhavanivṛttiḥ jñāpako ’yaṃ pratiṣedho na saṃbhavanivartaka iti. Gautama (1887: 191: 13–15); Jha 1984: (II: 619).

(34.) VV 13.

(35.) VV 14.

(36.) Of course what the opponent must mean here is that the object of negation (i.e. svabhāva) does not exist as an object of negation (because there is no negation), not that it does not exist at all.

(37.) VV 15–16.

(38.) VV 65–67.

(39.) Matilal (1970: 94).

(40.) Matilal (1970: 96).

(41.) VP 16. This point is also underlined by Candrakīrti in commenting on MMK 15:11: “A healthy person does not perceive any of the hairs which appear to one afflicted by vitreous floaters. When he says ‘these [hairs] are not,’ he does not say that they are an existing entity the existence of which was denied because the object of negation is not real. In the same sense we say that ‘all things are not,’ in order to remove clinging to an error in those who see things in the wrong way like those afflicted by vitreous floaters.” yas tu taimirikopalabdha keśeṣviva vitaimiriko na kiṃcid upalabhate sa nāsti iti brūvan kiṃcin nāsti iti brūyāt pratiṣedhyābhāvāt | viparyastānāṃ tu mithyābhiniveśanivṛttyartham ataimirikā iva vayaṃ brūmo na santi sarvbhāvā iti. PP 273:14–274:3.

(42.) VP 20, VP 13–15. The same problem is raised in NS 2.1.12 (NS 425–426).

(43.) See Jha (1984: 621).

(44.) See the discussion in chapter 5.

(45.) VP 12.

(46.) na pratiṣedhaḥ pratiṣedhasyārthasya kāraṇam pratiṣedhyo na pratiṣedhasya ca. VV(S). 54:13–14. This translation appears to me more satisfactory than that of Bhattacharya’s, who just has “object of negation” for prati-ṣedhasya ārtha (Bhattacharya et al. 1978: 106). It would seem very peculiar to ascribe to Nāgārjuna’s Naiyāyika opponent the view that the object of negation is causally brought about by the negation (or the other way round).

(47.) See Bhattacharya et al. (1978: 106, n. 1) for a list of references to this example.

(48.) “By virtue of your statement a negation is not possible in the three times, and, like the negation, the thing to be negated also [does not exist].” tathā hi tvadvacanena pratiṣedhastraikālye ’nupapanna pratiṣedhavat sa pratiṣedhyo ’pi. VV(S) 83:17–18.

(49.) In VP 14–15 we find a different reply. Here the opponent argues that once the existence of the means and objects of knowledge is denied in the three times, this denial then also infects the possibility of negation of the means and objects of knowledge, which also has to exist in the three times. Nāgārjuna then replies that it is not feasible to first accept the negation of some object and then use this very acceptance to argue for its existence. See the commentary in Tola and Dragonetti (1995b: 108–110).