23

Introduction

The Vaidalyasūtra and the Vaidalyaprakaraṇa

WHEN CONSIDERING the Vaidalyaprakaraṇa we are in fact looking at two texts. The Tibetan collection of translations of Indian commentarial works, the Tengyur (bstan ‘gyur), contains the shorter Vaidalyasūtranāma (VS, zhib mo rnam par ‘thag pa zhes bya ba’i mdo), as well as the longer Vaidalyanāmaprakaraṇa (VP, zhib mo rnam par ‘thag pa zhes bya ba’i rab tu byed pa) that embeds the VS in a commentary.³ Unfortunately the relation between these two texts cannot be captured by the simple formula VP = VS + commentary.

First of all, the VS and the VP were translated into Tibetan by different teams of translators. The VS was translated by the Kashmiri pandit Ānanda (kha che’i pandi ta ā na nta) and the Buddhist monk and translator Drakjor Sherab, while the VP was translated by the Kashmiri pandit Jayānanda and the monk Kudo Dewar.⁴ Kajiyama⁵ notes that the texts of the VS and its version embedded in the longer VP “often deviates in verbal form, though not in the sense, from the text of the separately extant Vaidalya-sūtra” (which is just what we would expect of two separate translations of the same text).

Secondly, even though the VP is structured according to the familiar model of a concise root text that is elaborated and commented upon in a 4commentary, distinguishing root text and commentary in a precise way is often difficult since the VP does not follow the convention of contrasting a metrically structured root text with a prose commentary. Rather, both the root text and the commentary are written in prose.

It is tempting to speculate on what motivated this unusual form, since the usual reason, that the metrical root text can be more easily memorized, does not apply in this case. According to the colophons of the two texts, the VS was received orally from Nāgārjuna⁶ while the VP is said to be composed by him.⁷ This suggests the possibility that the VS is effectively a set of notes based on a lecture by Nāgārjuna (which would explain its lack of metrical structure) and the VP was then subsequently composed by him in order to elucidate and explain this abbreviated summary of his lecture.⁸

While it is frequently the case that (despite their extremely compressed form) root texts can be understood without referring to a commentary, several passages in the VS would be incomprehensible without the commentary.

For a particularly clear example consider VS 04 and 05. Reading them without a commentary one would understand VS 05 as contradicting VS 04. Yet this is not the case, as Nāgārjuna’s “This is not so” (ma yin te) at the beginning of VS 05 is not directed at the position put forward in VS 04, but rather against a suggestion made by the opponent before VS 05, which is only found in the commentary. Without the commentary no amount of guesswork would allow us to infer what position VS 05 is reacting against.

For this reason it seems somewhat unlikely that Nāgārjuna first composed the VS as a concise treatment of the topic, and subsequently explained and elaborated it in the autocommentary we now find in the VP. If the combination of root text and commentary expresses what Nāgārjuna wanted to say, examples like the one just mentioned suggest that the VS could not have been conceived as a concise “stand-alone” exposition of the same meaning.⁹ Root text and commentary rather seem to form an integral unit and give 5the impression of having been composed in dependence on one another. The commentary obviously depends on the root text (because that is what it is explaining), but as we have just seen, the root text also depends on the commentary for preventing the reader from misunderstanding the way the different parts of the root text relate to one another.

The distinction between the VS and the VP is therefore somewhat artificial. Not all sections of the VS are able to stand on their own as an independent text; moreover, it would not be too difficult to conceive of the VP as a continuous prose text with no root text/commentary distinction at all.

In addition, the boundary between the root text and the commentary is often somewhat blurred. The term gzhan yang (api ca, “moreover”) is usually regarded as part of the prose commentary, introducing a new section of the root text. In some cases (e.g., VS 30 and VS 40), however, this constitutes part of the root text without there being an obvious distinction in meaning between these instances of gzhan yang and all the other instances that did not get included in the root text. It is not very implausible to argue that these instances of gzhan yang were included by mistake and should not be considered part of the root text, even though they are included in the VS.

The not entirely perspicuous relation between root text and commentary also makes it difficult to number the parts of the root text and associated commentarial sections.¹⁰ As the root text is in prose, we cannot just count metrical units to individuate the items. In some cases VP and VS are only distinguished by the fact that the VP contains some additional short phrase in the middle of a passage that is missing in the VS. Are we then to count these as two distinct sections of the root text, separated by the short phrase, or as one continuous section interrupted by a piece of commentary?

For example, sūtra 68 from the VS reads as follows:

khyod kyi thams cad tshig don du brjod pa yin gyi | don dam pa ni ma yin no zhe na | ma yin te | lan thams cad la thal bar ‘gyur ba’i phyir ro

The version we find in the VP is very similar, but it adds the phrases set in boldface:

smras pa | khyod kyi thams cad tshig don du brjod pa yin gyi don dam par ma yin no zhe na | brjod par bya ste | ma yin te | lan thams cad la thal bar ‘gyur ba’i phyir ro

As this makes clear, we are dealing here with an objection by the opponent (introduced by the phrase smras pa [“Objection”]), followed by Nāgārjuna’s reply (following the phrase brjod par bya ste [“it is to be replied”]).

We appear to be equally justified in saying that “sūtra 68” from the VS is actually not one sūtra, but two, an objection and a reply, or that it is one sūtra, and that the VP interpolates its commentary into the sūtra.

While not much hangs on the way of numbering we choose for understanding the contents of the text, problems arise if different authors group different units together in different ways and thereby end up referring to different passages when they talk about “verse x.” Kajiyama includes what Tola and Dragonetti regard as the second sūtra as part of the first.¹¹ Thus, according to Kajiyama, the root texts consist of 72 sūtras, whereas for Tola and Dragonetti there are 73. Pema Dorje’s Sanskrit reconstruction also counts only 72 sūtras, but he does not fuse what Tola and Dragonetti count as sūtras 1 and 2 into sūtra 1, but what they count as sūtras 2 and 3 into sūtra 3. In the present discussion we will follow the numbering of the sūtras adopted by Tola and Dragonetti, although we regard what they count as sūtra 73 as being two separate sūtras. According to our counting, the VP thus contains 74 sūtras.

The Question of Authenticity

With the exception of the MMK, which is regarded as a work of Nāgārjuna by definition, the authenticity of all other works traditionally ascribed to him has been questioned. The VP is no exception.

The VP belongs to a set of six texts, sometimes referred to as the Yukti corpus, which the Tibetan tradition has canonized as Nāgārjuna’s “six works on reasoning” (rigs pa’i tshogs drug).¹²

The VP differs from other works traditionally included in the Yukti corpus, the Mūlamadhyamakakārikā, Śūnyatāsaptati, Yuktiṣaṣṭikā, and Ratnāvalī, in not having any Indian or Tibetan commentaries associated with it.¹³ It shares this trait with the Vigrahavyāvartanī (the only extant commentary on both texts is what is assumed to be Nāgārjuna’s autocommentary). As opposed to the Vigrahavyāvartanī, which is quoted relatively often in the Buddhist commentarial literature¹⁴ however, the VP appears not to have been quoted much by later authors.¹⁵

The Sanskrit originals of both the VS and the VP are lost; both texts are preserved in Tibetan translation. The colophons of both texts identify Nāgārjuna as the text’s author. Among Indian Madhyamaka authors, Bhavya¹⁶ and Chandrakīrti¹⁷ mention the VP as a work of Nāgārjuna’s.

Modern scholarship is divided over the question whether the VS and VP are indeed the work of Nāgārjuna. Yūichi Kajiyama and Christian Lindtner explicitly argue in favor of this attribution,¹⁸ while Fernando Tola and Carmen Dragonetti, as well as Ole Pind, raise arguments against it.¹⁹

Kajiyama bases his positive verdict on its ascription to Nāgārjuna in the colophon, together with the authority of later Tibetan historians such as Butön. Moreover, he argues that “its contents nowhere suggest any other author to whom the work might be better ascribed.”²⁰

Lindtner considers the VS and VP as genuine because they agree with the MMK in style, scope, and doctrine, and are additionally ascribed to Nāgārjuna by trustworthy witnesses such as Bhavya and Chandrakīrti:

Judging solely from the text itself, the style and tenets would indicate the same author as for VV, the work where the parallels are closest out of all those ascribed to Nāgārjuna.²¹

Tola and Dragonetti²² mention various reasons that might raise doubts regarding Nāgārjuna as the author of the VP, even though they admit that none “by themselves alone, discard in an absolute sense the authenticity of the work.”²³ The most interesting of these refer to specific topics or assumptions Tola and Dragonetti identify in the VP, arguing that insofar as these topics or assumptions can be found in other texts, such texts are too late to allow Nāgārjuna to be the author of the VP.²⁴ Yet this approach has too many methodological problems to allow us to form any reliable view of 9whether or not Nāgārjuna can be regarded as the author of the VP. First of all it is often not clear whether the respective topics or assumptions are indeed the ones the VP talks about (regarding the parallels with Bhartṛhari and Yogāchāra that Tola and Dragonetti identify, I believe this is in fact somewhat doubtful). But even if we accept that the very same idea is expressed in the VP and in a text that probably was not around until some centuries after the generally accepted date of Nāgārjuna in the first or second century CE, what does this show? In order to draw any firm conclusion regarding the VP’s authorship from this we would have to exclude the possibilities that (a) the later text borrowed the idea from the VP, (b) both texts derive the idea from a common predecessor, and (c) that two texts can discuss identical or similar ideas without one drawing on the other. As it is impossible to exclude any of these possibilities, the parallels mentioned cannot provide evidence that allows us to decide the question of the authenticity of the VP.

Pind’s argument focuses on the idea that some passages of the VP are incompatible with Madhyamaka views, a fact he considers “to settle, once and for all, the debate about the authenticity of the VP.”²⁵ He remarks with respect to VP 58 that:

In any case, it seems certain that the paragraph of the VP is not consistent with basic Madhyamaka assumptions. It is therefore highly unlikely that the VP was written by a Mādhyamika in the original sense of the word…. VP cannot under any circumstances be attributed to Nāgārjuna.²⁶

Pind’s argument presupposes an emendation of the text; neither this emendation nor the conclusions he draws from it appear wholly cogent to me.²⁷ Pind’s conclusion that they refute Nāgārjuna’s authorship “once and for all” and that “we can now safely exclude the VP from the corpus of Nāgārjunian writings”²⁸ thus appears somewhat premature.

It seems to me that in the absence of more decisive considerations than those that have at present been brought forward against the authenticity of the VP, ascriptions from the Indian and Tibetan tradition (such as those found in the colophons and the works of later Indian and Tibetan authors) should be taken very seriously. There is no denying that the VP occupies a unique position in the Yukti corpus, differing from the other members of that set with respect to its formal features, and with respect to the degree of its embedding in the scholastic tradition via commentaries and quotations. But unless some harder evidence emerges that would allow us to exclude the VP from the collection of Nāgārjuna’s key Madhyamaka works I suggest that it should be considered as authentic.

The Aim of the Vaidalyaprakaraṇa

Nāgārjuna’s aim in the VP seems very clear: it is to refute the sixteen Nyāya categories.²⁹ Yet we need some more explanation in order to understand what precisely this refutation amounts to,³⁰ and what its role is in the larger Madhyamaka enterprise.

A first question to ask is whether Nāgārjuna wants to refute the Nyāya categories from an ultimate or from a conventional perspective. Tola and Dragonetti³¹ suggest that we distinguish two different aims Nāgārjuna might want to accomplish in the VP:

1. “[To] examine and analyze the padārthas from this metaphysical point of view [of emptiness], in the metaphysical level, in order to reach the conclusion that they are śūnya, that they do not exist paramārthataḥ….”

2. To show “that the padārthas are affected by manifold logical defects,” to criticize “logic from the point of view of empirical reality, although not from a metaphysical point of view.”

Even though I am not quite sure what is meant by criticizing logic “from the point of view of empirical reality,” this is the underlying idea suggested by Tola and Dragonetti: One criticism Nāgārjuna can bring forward against the Naiyāyika’s set of logical categories (and against several other parts of the Nyāya system) is that it conflicts with the Mādhyamika’s anti-substantialist metaphysics.

There can be little doubt that according to the Naiyāyika’s own view the sixteen categories exist at the level of ultimate truth. Not only do they belong to the ultimate furniture of the world, knowledge of them is also instrumental for obtaining liberation (niḥśreyasa).³² It is evident that the anti-foundationalist standpoint of Nāgārjuna’s Madhyamaka does not cohere well with this understanding of the categories as ultimately real.³³

A second line of criticism is purely internal. Here the problem is not that the Nyāya theory fails to measure up to the demands of another metaphysical system, but that it encounters problems on its own terms. Its 12contradictions are internal, not just external and arising when confronted with the theory of emptiness.

It is not entirely clear, however, that we can really draw a sharp distinction between these two aims.

Consider, for example, VP 47, which criticizes the idea that something can be a reason for something else. The argument is that either the reason requires another reason, in which case we have an infinite regress, or it does not, in which case we have to justify why we are comfortable with letting this reason stand without a reason, but not the conclusion of the original argument. We might wonder whether this actually is a logical defect of the notion of a reason. If it were then the opponent could not bite the bullet and remain a rational agent. But there is nothing irrational in asserting that there is no principled way of deciding which statements need reasons and which do not. All we can appeal to is the fact that we feel a need for certain things to be justified but not for others. The point is not that such a position would be logically deficient, but that it would be incompatible with a foundationalist understanding of reasons such as the one the Naiyāyika is supposed to espouse. The problematic consequence that the notion of a reason either leads to a regress or begs the question (as Nāgārjuna argues) only follows if we assume that there is a way of deciding which statements need reasons and which do not that is somehow objectively grounded in the nature of things — that is, if we make substantialist assumptions about reasons.

In order to do justice to such passages, it is essential to be aware of two points when trying to make sense of Nāgārjuna’s aims in the VP:

1. In keeping with the overall Madhyamaka outlook Nāgārjuna wants to refute the ultimate existence of all of the sixteen categories.

2. He does not want to refute the conventional reality of the categories, since if they were taken to be lacking any functional role even at the transactional level, there would be no possibility for a Mādhyamika (or indeed for anybody else) to employ any of them in an argument.

Tola and Dragonetti are right when they point out that Nāgārjuna wants to show “that the padārthas are affected by manifold logical defects,” but this statement might easily be mistunderstood. Nāgārjuna does not want 13to say that because all of the Nyāya categories lead to contradictions they should all be rejected at the conventional level. Rather, he wants to point out that they lead to contradictions if understood against the background of the Nyāya framework. Nāgārjuna’s main aim is to show that the way the categories are interpreted by the opponent makes certain substantialist assumptions about ultimately existent entities, and that these assumptions have to be rejected, without, however, giving up the conception of the categories tout court. For each category we therefore have to investigate how (a) following the Nyāya understanding it brings with it such substantialist assumptions, and (b) how these assumptions can be removed and the conception of the category be reinterpreted without losing its functionality. As a result, we will then be able to show that we can still formulate a Madhyamaka-compatible form of these categories that does not connect them with the problematic metaphysical assumptions that the Naiyāyika wants to make. A suitably desubstantialized variant of the Nyāya categories, Nāgārjuna argues, may still be retained.

It is useful at this point to consider what the Tibetan tradition understood the purpose of the VP to be.

In his commentary on the MMK, Maja Changchub Tsöndru points out that:

Some may object that if things have no intrinsic nature, it would contradict the establishment of such natures by intrinsically qualified means of knowledge [rang bzhin tshad ma]. [However,] to establish that things exist intrinsically, using epistemic means, it is necessary that when the epistemic means are investigated, there is no contradiction, yet this is not the case. In this way, the epistemic means and objects, etc., the sixteen categories of the Naiyāyikas are contrary to reason [mi ‘thad pa]:

So that those intent on dispute,

Priding themselves in their expertise in logic,

May give up this pride,

I shall explain this detailed examination.

It is pointed out that, having investigated the epistemic means and objects, etc., establishing the intrinsic nature of things in detail, they are rejected.³⁴

Tsong Khapa is a bit more concise but makes essentially the same point:

These treatises [i.e., the Yukti corpus] are of two primary kinds: those which demonstrate the way things really are — free from two extremes of existence and nonexistence — and those which demonstrate that the path free from these extreme views leads to liberation. The former is demonstrated in two ways: Mūlamadhyamakakārikā refutes the thesis concocted by the reificationists, viz., that persons and phenomena have essences; the Vaidalya refutes the Naiyāyika’s use of their sixteen categories,³⁵ such as epistemic instruments, to prove that they do.³⁶

Maja Changchub Tsöndru and Tsong Khapa therefore understand the VP as a reply to an opponent who claims that if things are established by reliable epistemic instruments, they have intrinsic nature or svabhāva.³⁷ 15Why would one think this? It is perhaps easier to look at the reason for the contrapositive: if things have no intrinsic nature then they are not established by reliable epistemic instruments. The reason for this is that if the Mādhyamika’s thesis of universal emptiness is accepted, all epistemic instruments are empty too. They are thus not knowledge-producers by their intrinsic nature; in fact, ultimately speaking there are no epistemic instruments. How could the theory of emptiness then be established as the kind of view of the world required for attaining liberation?³⁸ The opponent thus argues that because we have access to the world by means of epistemic instruments, the Madhyamaka thesis of universal emptiness cannot be correct. One way for Nāgārjuna to reply to this charge is to point out that the epistemic instruments (together with the remainder of the Nyāya categories) are in fact not reliably established themselves. If they are rejected (‘gog pa) as unreasonable (mi ‘thad pa) the opponent’s argument based on their authoritative role loses much of its force.

It therefore appears clear that both Tibetan authors understand Nāgārjuna as not simply rejecting the ultimate establishment of the Nyāya categories. According to them, Nāgārjuna’s aim is not just to point out that the sixteen categories, together with everything else, are empty, but to refute their probative force (sgrub byed…’gog pa) by showing that they are not “without contradiction” (mi ‘gal ba). At the same time it seems unsatisfactory to assume that these contradictions rule out any employment of the categories at the conventional level, for in this case any talk of epistemic instruments or objects, inferences, examples, and so forth, would have to be given up, thereby effectively robbing oneself of the ability to put forward any structured philosophical arguments. When “refuting the Naiyāyika’s use of their sixteen categories” the aim is not to refute any use of the categories whatsoever, but to reject those kinds of usage that make unacceptable substantialist presuppositions, such as that behind the discussion in VP 47, namely that each justification has to be traceable to an unjustified justifier. The Nyāya categories are not contradictory as such, but only when coupled with certain svabhāva-entailing assumptions. It is the refutation of these assumptions, not of the categories themselves, that is the aim of Nāgārjuna’s criticism in the VP.

When Nāgārjuna begins his discussion of the epistemic instruments and objects he spends a considerable amount of time refuting the idea that they are self-established, that is, that their ability to act as sources of knowledge or as objects known flows from their intrinsic nature, from being the very kind of things they are. Nāgārjuna rejects this idea, but this does not mean that he denies the existence of epistemic instruments or objects tout court. If this were the case, the Mādhyamika would have no way of accounting for the fact that we know things, and that some ways of getting knowledge are better than others. His aim is rather to underline the necessity of developing an understanding of these two crucial epistemological notions that does not presuppose their existing by intrinsic nature. Such an understanding is achieved, for example, by conceptualizing the epistemic instruments and objects as mutually existentially dependent. If each could not exist if the other did not, the existence and properties of either could not flow solely from itself.

Similarly, when examining the notion of examples, Nāgārjuna’s arguments in verses 28–31 set out to refute the notion of a similarity relation that exists “out there” in the objects, a relation that depends on nothing else but the intrinsic properties so related — in other words, a similarity relation existing by svabhāva. Again, Nāgārjuna rejects this idea, but not in order to make the point that no object is similar to any other one, but in order to justify the introduction of another, svabhāva-free similarity relation, one that depends crucially on the conceptualizing mind that regards objects as similar or dissimilar.

I would want to propose that understanding the purpose of the VP as a justification for a desubstantialized understanding of the logical categories discussed by the Naiyāyika is one that allows us to make sense of the aim of the text in the context of Nāgārjuna’s other philosophical works. It is certainly superior to considering it as a mere display of sophistical fireworks (which would set it apart from other texts by the same author), an attempt to show that the Nyāya categories are as empty as everything else (failing to explain why Nāgārjuna devoted an entire book to the topic, given that this follows simply from the thesis of universal emptiness he defends elsewhere), or as aiming at demonstrating the inner inconsistency, and thus conventional deficiency of the sixteen categories (making it impossible for the Mādhyamika ever to engage in a philosophical debate using these categories).

At this place it is useful to point out the relation of the VP with the methodology usually associated with Prāsaṅgika Madhyamaka. The Prāsaṅgika, it is argued, does not endorse any philosophical conclusions of his own, but simply deduces deficiencies from the theses put forward by his opponent.

Instead of formal probative proof statements, Mādhyamikas should restrict themselves to arguments that proceed from the opponent’s own premises, either through a consequence (prasaṅga, thal ‘gyur) or through an inference whose elements are accepted by the opponent alone (gzhan grags kyi rjes dpag, *paraprasiddhānumāna).³⁹

In the Mādhyamika’s deductions, some logical concepts and techniques have to be used. Is the Prāsaṅgika approach now not vitiated if these logical techniques presuppose philosophical propositions that the Mādhyamika is implicitly taking on board by employing them in his refutation? Put more simply, if these concepts and techniques themselves presuppose the existence of svabhāva somewhere, is the Prāsaṅgika not committed to accepting these, thereby contradicting the theory of universal emptiness?

One potential reply would be to say that these concepts and techniques are simply paraprasiddhānumāna, that is, something that the Prāsaṅgika uses because his opponent accepts them, not because he is himself committed to them. But this understanding does not cohere well with the arguments we find in the VP. If Nāgārjuna had argued that the entire logical machinery of the sixteen Nyāya categories is to be used on the understanding that it is accepted by the opponent alone, once again there would have been no reason to critique the categories one by one, and therefore no reason to compose the VP. It would have been sufficient to bracket them all collectively, in the same way in which other assumptions made by other opponents are bracketed.

It is therefore plausible to assume that Nāgārjuna was aiming to show something about the categories that necessitated the examination of each category individually. This, I have argued, is that each of the categories as understood by the Naiyāyika presupposes the existence of svabhāva in 18some form, that these presuppositions lead to contradictions and that it is possible to develop versions of these logical concepts and techniques that do not make such problematic presuppositions. Nāgārjuna’s aim in the VP is to show that the Naiyāyika’s understanding of the sixteen categories is not svabhāva-free, but that it can be made to be, and that the Nyāya categories can then be employed by Mādhyamikas in their reasoning.

Synopsis of the Text

The purpose of this synopsis is to make navigating the VP easier by providing brief summaries of the key points and arguments expressed in the individual sūtras. The compressed nature of this survey necessitates some simplification. The individual summaries are therefore not to be understood as expressing the entire meaning of each passage in a comprehensive manner, but merely serve as pointers to facilitate following the flow of the argument.

Introduction

01 Rejection of the sixteen Nyāya categories for the proponent of emptiness.

Pramāṇa

02 Mutual dependence of epistemic instruments and epistemic objects.

03 Epistemic instruments and objects are not self-established.

04 Dependent existence rules out existence by svabhāva.

05 Assuming epistemic instruments and objects leads to an infinite regress.

06 Opponent: There is no regress, since the epistemic instruments are self-illuminating.

07 The light does not illuminate whether or not it is connected with darkness.

08 Contactless illumination is unsatisfactory.

09 Darkness is a mere absence, so light cannot be cleared away.

10 Self-illumination fails, since there is no darkness in the light.

11 Self-illumination fails, as otherwise darkness would be self-concealing.

12 Epistemic instruments and objects cannot have temporal relations.

13 Opponent: In this case the epistemic instruments and objects and their negation cannot have temporal relations either.

14 It is not the case that the negation of the epistemic instruments and objects establishes their existence, as the Nyāya understanding of negation would presuppose.

15 Once the rejection of epistemic instruments and objects has been established in VP 12 there is no room for any further moves such as those described in VP 13 and VP 14.

16 The negation of epistemic instruments and objects is possible because objects that are erroneously assumed to exist can be negated too.

17 Opponent: Epistemic instruments and objects exist because there is correct knowledge.

18 Even if there are epistemic instruments, there might be no epistemic objects.

19 A pot is not an epistemic object.

20 Mental cognitions are no epistemic instruments.

Saṁśaya

21 There is no doubt, since it can neither be directed at conceived nor at non-conceived objects.

22 Opponent: There is doubt because there are indistinct perceptions.

23 There is no doubt because every perception is determinate.

Prayojana

24 There is no purpose, because it can be neither existent nor nonexistent.

Dṛṣṭānta

25 Opponent: Examples and purposes exist, because there are examples of things with purposes.

26 Distinguishing between actions with purpose and those without does not avoid the difficulty mentioned in VP 24.

27 The end (anta) is not seen (dṛṣṭa), because beginning and middle are not seen.

28 There are no concordant examples.

29 There are no discordant examples.

30 There is no slight similarity between example and exemplified.

31 There is no major similarity either.

Siddhānta

32 The end (anta) is not established (siddha) because beginning and middle are not established.

Avayava

33 There are no parts of a syllogism because no whole exists in the parts taken together.

34 There is also no whole in the parts taken individually.

35 The whole is not the same as its parts because the parts could have inconsistent properties, and because the whole would exist in addition to the parts.

36 The whole does not exist in the three times.

37 Opponent: The whole exists in the aggregate of the parts.

38 The whole cannot be taken to be an aggregative effect of the parts because this would require a separate argument.

39 The whole cannot be taken to be an aggregative effect of the parts because they are not simultaneously present.

40 The parts cannot be established by something else, nor can they be self-established.

41 Thesis and reason are neither identical nor different.

42 Reasons generate an infinite regress.

43 If the first three members of a syllogism cannot exist, the last two cannot exist either.

44 The relation between example, application, and reason is unclear.

45 All other parts apart from the reason are meaningless, as it has exclusive probative force.

46 If another part had such force, the other members would be meaningless.

47 Successful inferences are not sufficient to establish the individual members of a syllogism.

48 Only one part of a syllogism can exist at a given time.

49 Inferences can be successful at the conventional level despite momentariness, as the case of polysyllabic words shows.

Tarka

50 Tarka is like doubt, which exists neither for known nor for unknown things.

Nirṇaya

51 There is no determination since neither a thesis nor its negation can be regarded as ultimately true.

Vāda

52 There is no debate, because there are no expressions and referents.

53 Claiming that expressions and their referents are linked by convention does not help.

54 One thing can be denoted by many words.

55 There is no essential connection between a word and what it denotes.

56 Expressions and referents are neither identical nor distinct.

Jalpa and Vitaṇḍā

57 Both are refuted in the same way as vāda.

Hetvābhāsa

58 A pseudo-reason is neither similar nor dissimilar to a reason.

59 A pseudo-reason cannot be what is a reason by nature nor what is not a reason by nature.

60 Opponent: Yet there are examples of pseudo-reasons.

61 The examples indicated are not pseudo-reasons.

62 There are no pseudo-reasons because of momentariness.

Viruddha

63 Opponent: There are contradictions because of the disagreement about the existence of pseudo-reasons.

64 The contradictory items do not exist at the same time, so there is no contradiction.

Kālātīta

65 Opponent: Momentariness entails the existence of mistimed reasons, so there are pseudo-reasons.

66 Opponent: Mistimed reasons exist because language commonly refers to time.

67 Past reasons cannot exist in the present.

Chala

68 Opponent: There is an equivocation, because Nāgārjuna only has ultimate existence in mind.

Jāti

69 Because there is no production there is no production of false rejoinders.

Nigrahastāna

70 There is no repetition, because it is neither identical with nor distinct from the original.

71 Grounds of defeat cannot be produced from what is the same or different.

72 There is no defeat in grounds of defeat, as there is no binding in a place of binding.

Conclusion

73 Rejecting the categories does not make it impossible to assert negations.

74 Expressions and referents do not exist.

_______________

  3.   VP has been translated into English by Tola and Dragonetti (1995) and into Japanese by Yamaguchi (1944) and Kajiyama (1974). Sempa Dorje (1974) offers a reconstruction of the Sanskrit text of the VS. At least three further English translations of VP have been announced (an “annotated English translation” by Yūichi Kajiyama [1965, 131]; one by Christian Lindtner [1986, 358n110]; and one by Ole Pind), though none of these have been published so far.

  4.   See p. 286.

  5.   Kajiyama 1965, 131.

  6.   slob dpon ’phags pa klu sgrub kyi zhal snga nas mdzad pa.

  7.   slob dpon ’phags pa klu sgrub kyis mdzad pa.

  8.   Venkata Ramanan’s (1966, 37) list of Nāgārjuna’s works classifies the VP under “Commentaries or/and Records of Oral Instruction (upadeśa).”

  9.   See Pind 2001, 155. He points out that the sūtras of the VS “could not