“Nāgārjuna’s Ways of Argumentation: Some Logical Issues” — Gokhale, Pradeep P. Paper presented at the international seminar “Buddhist Culture in Asia: Unity and Diversity”, K. J. Somaiya Centre for Buddhist Studies, Mumbai, 10–13 January 2008. Affiliation given as Central University of Tibetan Studies, Sarnath.
Bibliographic gap. The extraction carries the seminar details but no publication venue, volume or pagination. Page references cannot be given for this source, and the citation cannot be completed without asking Gokhale directly.
Thesis / main argument
Gokhale sorts the logical problems Nāgārjuna’s argumentation raises into four: the catuṣkoṭi and its apparent transgression of classical logic; the nature of the Madhyamaka negation; the charge that Nāgārjuna commits outright fallacies; and the choice between prasaṅga and autonomous method. His answer runs through all four from one idea. Nāgārjuna argues inside classical logic for as long as a common presupposition is available to be denied, and steps outside it at the limiting cases where none is left — the denial of existence and non-existence, and the denial of holding any thesis at all. At those limits the negation stops being truth-functional and becomes what Ruegg calls a performative negation: an expression of a negative attitude towards a statement rather than the making of a further statement. So it is not subject to double negation or the excluded middle, and no proposition of non-existence results from it.
The transgression is real, and Gokhale says so plainly rather than explaining it away. What drives it is not an interest in constructing a deviant logic but a critical attitude to language: the units of language and thought are vikalpa, they present themselves as fixed and essential, and they are in fact open and relative, so paradox is what one gets by pressing them. On this reading the Mūlamadhyamakakārikā is an implicit criticism of language, and its logical oddities are symptoms of that criticism rather than errors.
The paper closes on a problem Gokhale leaves open. Nāgārjuna both attacks language and uses it, and approves the Buddha’s use of it. Gokhale holds that the Two Truths cannot settle this, since Nāgārjuna ranks the two hierarchically and treats conventional truth as delusive, so the double attitude “is in need of a better explanation.” He supplies the Tractatus say/show distinction and the ladder as the better explanation, while marking the differences from Wittgenstein.
Key claims
- The four issues. (A) Catuṣkoṭi and the transgression of classical logic; (B) the nature of the Madhyamaka negation, paryudāsa versus prasajya-pratiṣedha; (C) whether Nāgārjuna commits fallacies (Robinson 1957 on denying the antecedent, Hayes 1994 on equivocation); (D) whether prasaṅga or autonomous method fits the project. Gokhale treats them as interconnected and answers them from the negation question.
- Galloway’s common-presupposition defence works locally and fails globally. Galloway (galloway-logical-issues-1989) blocks the excluded-middle objection by pointing out that Nāgārjuna denies all alternatives because he denies a presupposition they share. Gokhale accepts this for any single tetralemma but shows it does not survive iteration. Deny the three modes of going in the Gatāgataparīkṣā and the shared presupposition is that there is an act of going; deny going and staying (MMK 2.15–17) and the presupposition is existence; deny existence and non-existence (MMK 15) and there is nothing left behind them. His conclusion is stated without hedging: “one cannot deny existence and non-existence without transgressing the laws of logic such as excluded middle and double negation.”
- Two distinct uses of negation, and only the second is anomalous. (a) Within a given trilemma or tetralemma the negation is truth-functional, assertible in two-valued logic and subject to the excluded middle, because the shared presupposition supplies its ground. (b) At the limiting cases the negation of one alternative delivers the other, and the second negation cannot return to the first, which has already gone. That denial is not truth-functional.
- Performative negation, after Ruegg (1986, p. 235). A performative negation expresses a negative attitude towards a statement rather than making one, so it is not itself a proposition and cannot be truth-functionally negated in turn. The law of double negation therefore does not apply to it. Gokhale’s test case is the VV 29 denial nāsti ca mama pratijñā: “I am not making any proposition” cannot be a proposition without paradox, and the paradox dissolves once the utterance is read as a pseudo-proposition carrying a non-truth-functional negation.
- Robinson’s fallacy charge answered by Galloway. The charge of denying the antecedent (Robinson 1957), which Galloway restates as violation of contraposition, is met by reading Nāgārjuna’s yadi as “only if” (galloway-logical-issues-1989 pp. 22–24). Gokhale endorses the answer without adding to it.
- Hayes’s equivocation charge answered by the open-concept reading — this wiki’s most transferable move. Hayes (hayes-nagarjunas-appeal-1994) holds that Nāgārjuna’s central argument turns on using svabhāva in more than one sense. Gokhale’s reply is that this is better described not as equivocation but as demonstration: Nāgārjuna is showing that svabhāva is an open concept whose diverse implications generate the paradoxes. He gives the same treatment to arising (“coming into being” against “coming into being out of something”), destruction (“annihilation” against “merger of effect into cause”), causation (satkāryavāda against asatkāryavāda), identity (absolute against virtual), difference (distinguishability against separability), and svabhāva itself (“an identity mark” against “an ontologically eternal characteristic”). The relativeness of these core concepts is the finding, not a defect in the arguing.
- The methodological consequence (issue D). If the negation were truth-functional, neither school’s method would serve. Autonomous inference for śūnyatā would make emptiness a view, which MMK 27.30 (sarvadṛṣṭiprahāṇa) forbids; prasaṅga used to reject all views would be self-inconsistent, since the negation of a view is itself a view. The prasaṅga method becomes intelligible only on the non-truth-functional reading. Gokhale therefore reaches a pro-prasaṅga conclusion from a purely logical premise, with no appeal to commentarial authority.
- Nāgārjuna generalises the Buddha’s silence rather than repeating it. The opening frames Theravāda’s treatment of the avyākata (Cūḷamālunkyasutta, Aggivacchagottasutta) as pragmatic wisdom about a bounded set of transcendent questions, and Nāgārjuna as extending the same response to all questions: not only whether the Tathāgata exists after death but whether he existed in his lifetime (MMK 25.18), not only the eternality of the world but whether phenomena exist at all (MMK 25.22).
- The Two Truths are declared insufficient. Gokhale raises Nāgārjuna’s double approach to language, notes that the standard answer is the lokasaṃvṛtisatya / paramārthataḥ satya distinction, and rejects it as an answer: “This two truth theory, however, cannot put a full stop to the issue because Nāgārjuna himself arranges the two truths in a hierarchical order.” Conventional truth is delusive or corrupt (saṃvṛtatva), and Nāgārjuna’s whole enterprise aims at showing it so; a truth he is engaged in discrediting cannot licence his own use of language. The Tractatus say/show distinction is offered in its place.
- The Dharmakīrti contrast. Both hold reality ineffable, for different reasons. Dharmakīrti’s sāmānyalakṣaṇa form an internally coherent order through which reality is reached by anumāna, so language has an explained constructive role. Nāgārjuna acknowledges a pragmatic and illuminating role for language but, on Gokhale’s reading, “does not try to explain how such a role becomes possible.”
Methodology
Analytic philosophy of logic, working directly on the Sanskrit text and on the modern secondary literature about it. The interlocutors are Matilal (1971), Robinson (1957), Galloway (1989), Hayes (1994) and Ruegg (1986); the comparative frame is Wittgenstein’s Tractatus. The commentarial tradition is almost entirely absent. Bhāviveka and the Prāsaṅgika–Svātantrika split appear only as a disagreement about the logic of proof, with no reference to Buddhapālita’s or Candrakīrti’s actual arguments; there is no Tibetan material, no neyārtha / nītārtha, no Three Turnings. Gokhale states his stance explicitly: “My approach in this paper is not that of justification or criticism of Nāgārjuna but that of sympathetic understanding.”
Notable quotes
- “one cannot deny existence and non-existence without transgressing the laws of logic such as excluded middle and double negation.”
Connections
- Wood / wood-nagarjunian-disputations-1994 — the substantive opposition, and the reason this source matters. Both engage the paryudāsa / prasajya distinction directly and both ask whether the prasajya negation obeys the excluded middle. Wood says it does and terminates in non-existence; Gokhale says the limiting-case negation is not truth-functional and terminates in nothing propositional at all. Neither cites the other, which makes Gokhale an independent voice at the exact point of dispute rather than a partisan of the tradition.
- Ruegg / ruegg-svat-pras-2006 — the performative-negation reading is Ruegg’s (“Does the Mādhyamika Have a Thesis and Philosophical Position?”, in Matilal and Evans 1986, p. 235); Gokhale adopts and applies it to VV 29. The wiki’s Ruegg material is on the Svātantrika–Prāsaṅgika criteria rather than on the 1986 essay, so this is a new Ruegg strand.
- Oetke / oetke-remarks-interpretation-1991 — both are analytic reconstructions with no commentarial engagement, and both take VV 29 as the test case. They land in opposite places. Oetke universalises the denial by instantiation from “on the level of highest truth there is nothing of any kind” and reaches a nihilist reading; Gokhale reads the denial as performative and reaches an explicitly anti-nihilist one. Same method, no framework in either case, contrary verdicts — which bears directly on how tightly may claim framework-removal determines the outcome.
- No-thesis / vv-29-three-readings — Gokhale’s performative reading is a further live reading of VV 29, distinguished from the others by denying that the verse states a proposition at all.
- catushkoti-must-negate-all-four-extremes — the second route to the same anti-Wood conclusion.
- gokhale-essentialism-eternalism-1996 — the external-negation account this paper’s performative reading grows out of, and the source of the third possibility at note 8. It also carries the “neither moral nor practical” statement of the objection, which is stronger than the version in the 1992 chapter.
- gokhale-nagarjuna-skepticism — VV 24 and 29 handled there without the performative account, with the gap acknowledged: Nāgārjuna’s statements are said to differ in logical structure from his opponents’, and what that structure is Gokhale says Nāgārjuna does not make clear. This paper is where the blank is filled.
- Kalupahana — opposed on the culmination question, though Gokhale never names him.
- Galloway / galloway-logical-issues-1989 — added 2026-08-07 on the strength of this paper. Gokhale adopts his answer to Robinson and rejects his catuṣkoṭi defence as failing under iteration; whether the rejection reaches the text is unresolved (note 8). Note also that the two land on the same side against Wood by different routes — Gokhale by non-propositionality, Galloway by subject-failure — and that Galloway supplies a further reading of VV 29 alongside Gokhale’s.
- Hayes / hayes-nagarjunas-appeal-1994 — added 2026-08-07 on the strength of this paper. The equivocation charge is Hayes’s, and Gokhale’s open-concept reply is one of two incompatible answers to it now on file (the other is Siderits’s; see Svabhāva). Hayes also grants what Gokhale’s whole essay is designed to explain away: that Nāgārjuna’s logic is standard. Where Gokhale concedes a transgression at the limiting cases and argues it is harmless, Hayes denies there is any transgression at all and locates the failure in the inferences instead.
- Richard Robinson, B. K. Matilal (no wiki pages) — the rest of the logic cohort this paper brings into view. Robinson, “Some Logical Aspects of Nāgārjuna’s System”, PEW 6:4 (1957), 291–308; Matilal, Epistemology, Logic and Grammar in Indian Philosophical Analysis (Mouton 1971), ch. 5.