Madhyamaka Wiki

No-thesis

16 min read

Definition

The “no-thesis view” is the Madhyamaka — and especially Prāsaṅgika — claim that the Mādhyamika does not hold a thesis (pratijñā) of his or her own. Its locus classicus is verse 29 of Nāgārjuna’s Vigrahavyāvartanī: “If I had any thesis, that fault would apply to me. But I do not have any thesis, so there is indeed no fault for me.” The claim is the formal counterpart of the methodological commitment to prasaṅga (consequence) over svatantra-anumāna (autonomous syllogism), and it is the doctrinal ground on which Buddhapālita, Candrakīrti, Atiśa, the Ninth Karmapa, and others build the Prāsaṅgika identity.

This must be distinguished from prasaṅga as a method. The question “does the Mādhyamika reason only by reductio?” is methodological; the question “does the Mādhyamika hold a position of his own?” is doctrinal. The two are connected, since without an autonomous thesis there is nothing for an autonomous syllogism to establish. But the doctrinal question is logically prior, and it is what splits the tradition four ways at VV v. 29 (see below). For the school distinction itself see Prāsaṅgika-Svātantrika; for the technique see the prasaṅga discussion there.

Indian primary anchor

VV v. 29 is the only place in Nāgārjuna’s surviving corpus where the no-thesis view is asserted in propria persona. The autocommentary glosses it by appeal to the Two Truths (v. 28 quotes Mūlamadhyamakakārikā 24:10 verbatim): the Madhyamaka argument operates at the conventional level; at the ultimate level the question of how an empty thesis refutes a substantial one does not arise. This means the no-thesis claim is not a free-standing piece of negative logic — it is parasitic on the framework that makes the conventional/ultimate distinction available.

A sūtra-stratum analogue, offered as an analogue and not as a source. The Aṣṭasāhasrikā Prajñāpāramitā has no doctrine of theses and no vocabulary of pratijñā, but at 1.8–1.12 it enacts the corresponding refusal. Śāriputra tries to convert Subhūti’s report that no bodhisattva and no perfection of wisdom are found into a claim about what exists — “if that mindset is free of a mindset, does that mindset exist?” — and Subhūti will not let the conversion happen. He does not assert non-existence, does not assert existence, and does not concede the question; he asks whether existence or non-existence is found there, and then whether the question was well formed. Ratnākaraśānti glosses the negation as one that “perfectly transcends both existence (bhāva) and non-existence.” This is the same shape as the v. 29 position — a practice that produces no proposition to be defended — appearing in the literature before there is any technical apparatus for describing it. It is worth noting for provenance and should not be cited as an Indian anchor for the no-thesis view proper: the sūtra makes no claim about theses, and reading one into it is exactly the move Galloway and others rightly resist elsewhere. See astasahasrika-84000-2026.

Buddhapālita develops the position into commentarial practice (every chapter of his Vṛtti on the Mūlamadhyamakakārikā proceeds by drawing out unwanted consequences from the opponent’s premises rather than asserting Buddhapālita’s own position). Bhāviveka objects: refutation that does not commit to an alternative position is logically incomplete (an anumāna requires a pakṣa). Candrakīrti in the Prasannapadā (Ch 1) defends Buddhapālita by re-anchoring the no-thesis line in v. 29 itself, and the [[Madhyamakāvatāra]] takes it as load-bearing for the Prāsaṅgika identity (MA 6.173). The Buddhapālita–Bhāviveka–Candrakīrti exchange on MMK 1.1 is the textual hinge.

The split on what v. 29 actually denies

The verse is verbally simple (“I have no thesis”) and substantively contested. Eight substantively different readings are tabulated here, and Mabja’s, discussed below, is a ninth. Six of the tabulated readings restrict what is denied — by content, kind of thesis, semantic register, truth-level, perspective, or occasion and assertion-context. The last two restrict nothing: one denies that the verse asserts anything at all, the other makes the denial a report of what the Madhyamaka negation always yields:

ReadingWhat v. 29 deniesSourceRange of application
TsongkhapaTheses qualified by svabhāva (theses asserting things to be intrinsically real)tsongkhapa-illuminating-intent-1418; cf. Jinpa on VV 26bNarrow — conventional theses are perfectly admissible; the Mādhyamika holds plenty of theses, including the thesis that things lack svabhāva
GorampaAutonomous theses. Both schools accept the four theses (no arising from self, other and so on) as mere theses (དམ་བཅའ་ཙམ་); Prāsaṅgikas hold reductio theses, asserted provisionally against the opponent, not autonomous ones proved on a commonly established subjectgorampa-distinguishing-views-1469 ch. 3 .5.1 (after Rongtön’s typology, Cabezón n. 346); against the equipoise reading, ch. 2Dialectical — the denial is indexed to the kind of thesis. Gorampa rejects confining v. 29 to meditative equipoise (མཉམ་བཞག་) when a defender of Dolpopa does so to keep “this truly exists” for the aftermath
Ninth KarmapaTheses of one’s own: the Mādhyamika has no position to affirm, and speaks of existence, emptiness and the rest from the perspective of otherskarmapa-feast-fortunate-1578 pp. 155–157, 173Perspectival — not indexed to equipoise; in Buddhapālita’s voice, no thesis of valid cognition is held in either truth (p. 173)
WesterhoffTheses to be supplied with realist (two-flavour) semanticswesterhoff-vigrahavyavartani-2010Semantic — Nāgārjuna asserts plenty of theses but rejects a two-tier system in which substantial theses overrule conventional ones; all Madhyamaka theses receive uniform convention-based semantics
OetkeAll declarative utterances at the paramārtha leveloetke-remarks-interpretation-1991Universalist — v. 29 follows by universal instantiation from the central tenet “on the level of highest truth there is nothing of any kind”; this is not a peculiarity of Nāgārjuna’s theses
Gendün ChöpelTheses as expressions of one’s own commitment in the context of ultimate analysis (vs assertions-for-others made out of conventional necessity)lopez-madmans-middle-way-2006 (Klu sgrub dgongs rgyan ¶51, ¶77–¶84)Equipoise-and-context — the no-thesis claim is indexed both to the meditative state (¶51) and to the assertion-context (¶80); operates across the neyārtha / nītārtha register, not just within Prāsaṅgika
GokhaleNothing — the verse denies no class of theses, because it makes no assertion. The negation in nāsti ca mama pratijñā is non-truth-functional (“performative”, after Ruegg 1986, p. 235) and the utterance a pseudo-propositiongokhale-ways-of-argumentation-2008; earlier and without the performative account at gokhale-nagarjuna-skepticismFormal — “I am not making any proposition” cannot be a proposition without paradox; so the law of double negation and the excluded middle do not reach it
GallowayStatements with content — v. 29 reports that Nāgārjuna makes only “zero statements”, the contentless output of a properly Madhyamaka prasajya negation. The verbal formulations he does produce are advanced procedurally, to refute othersgalloway-logical-issues-1989 n. 5 (after Ruegg 1983)Consequential — the no-thesis claim is not a separate doctrine but what the negation-theory delivers; a pratijñā in sense 1 (a statement with content) he never makes, in sense 2 (any verbal formulation) he makes constantly

These do not collapse onto one another. Tsongkhapa’s reading restricts the content of the denied class (only theses-with-svabhāva); Gorampa’s restricts the kind (autonomous theses only); Westerhoff’s restricts the semantic register (realist semantics only); the Ninth Karmapa’s restricts the perspective (one’s own, not others’); Gendün Chöpel’s restricts both occasion and assertion-context; Oetke’s generalises the denial to all declarative utterances at paramārtha. They differ in what makes the verse non-trivially true and in what burden it places on the rest of the Madhyamaka project. Gokhale’s stands apart from all six: it rejects the question rather than answering it, on the ground that the verse is not in the business of denying anything. Galloway’s stands apart from all seven in a different way: it answers the question but makes the answer fall out of something else, since if every Madhyamaka negation yields a statement with no content then “I have no thesis” is a description of the method rather than a claim about a restricted class.

Galloway’s objection to the Tsongkhapa reading, from outside the tradition. Quoting Lopez 1987 (p. 51) — on Tsongkhapa, Nāgārjuna “has no inherently existent position” — galloway-logical-issues-1989 objects that on those terms “one could hold any position (pratijñā) at all”, asserting even that things arise from what is other than themselves, provided one covered oneself with the qualifier; and Nāgārjuna does not seem to give himself that freedom (n. 5, p. 27). It is a complaint a Sakya-side reader could take up, though it is not Gorampa’s own: Gorampa allows the Prāsaṅgika mere and reductio theses, and his recorded objection to the qualifier is that negating all views and all four extremes is idle if each is negated only as truly existent (Distinguishing the Views ch. 3). Galloway reaches it as a logician with no Tibetan commitments, working only from Ruegg’s distinction between two senses of pratijñā and apparently unaware that there is a Tibetan controversy to land in. Two qualifications keep it honest. His target is Lopez’s summary in a 1987 study of Svātantrika, not Tsongkhapa’s own argument in Ocean of Reasoning, which is more careful: the verbal qualifier is optional for Prāsaṅgikas and what is non-negotiable is the substantive mode-of-existence distinction (see Non-affirming Negation on Ocean Homage .1.1.2). And in the same note Galloway refuses Lopez’s further step from having no position to nihilism, on the ground that holding no philosophical view paramārthavat is not nihilism, and that the Candrakīrti passage adduced concerns asvabhāvatā rather than apratijñātva — which is the same distinction the Tibetan tradition draws, again reached independently.

(See vv-29-three-readings for the case that the Tsongkhapa / Gorampa / Westerhoff disagreement is a load-bearing inter-school contradiction. Oetke’s universalist reading is a further live option there, but it sits structurally closer to a deflationary-nihilist gloss than to the traditional Tibetan readings. Oetke and Gokhale make the same verse the test case by the same method and reach opposite verdicts, which is the sharpest evidence that method alone does not settle the reading.)

The Gendün Chöpel articulation of the equipoise reading

The cleanest single-paragraph statement of the yogin-in-equipoise reading is Klu sgrub dgongs rgyan ¶51 (see lopez-madmans-middle-way-2006): “When one opponent who has assertions debates using scripture and reasoning with an opponent without assertions who abides in a state of meditative equipoise, free from verbalization, whatever answers the latter gives all become mere assertions. Thus, there is no place to fit this view of having no assertions within words, sounds, and, particularly, the reasoning of logicians.”

Gendün Chöpel (1903–51) supplies three further structural moves:

  1. The assertions-for-oneself / assertions-for-others distinction (¶80–¶84), illustrated with the cakravartin analogy (¶81): saying to Bu long ma “you are a cakravartin king” out of fear is an assertion in one’s manner of speaking but not in one’s own system. The necessary conventional concessions of unenlightened beings (being burned by fire, cooled by water, moved by wind) are similarly assertions-for-others. GC is explicit that “this way of understanding is not limited to the Prāsaṅgikas. It is not different in the other tenet systems that assert a presentation of the provisional and definitive and the two truths.” The distinction operates across the neyārtha / nītārtha register, not just within the Prāsaṅgika position. This anchors the no-thesis reading at the Mahāyāna hermeneutical framework level rather than at the level of school-internal commitment.

  2. The vow-of-silence-cranes analogy (¶77). The instruction “Don’t say anything, I’m sleeping”, together with the cranes who “have been very famous from long ago for the sound ‘mi dang mi dang’” because their leader said “Don’t talk”, blocks the standard Geluk objection that “I have no assertion in the context of ultimate analysis” is itself an assertion. The genre of the utterance matters: saying “Don’t talk” while flying is not the same genre as talking.

  3. The Buddha’s silence as paradigm, not exception (¶78–¶79): the Buddha’s silence at the fourteen unindicated views, the Buddha’s silent acceptance of Anāthapiṇḍada’s mealtime invitation, and Nāgārjuna’s Ratnāvalī I.74 reference to “the discipline of not speaking” are paradigms of the no-thesis position, not edge cases. This generalises the equipoise reading from a yogin’s meditative state to a pedagogically warranted method — what makes the no-thesis stance available to the Mādhyamika is the same structure that makes the Buddha’s silence pedagogically functional.

GC’s articulation shares the own-system / for-others distinction with the Ninth Karmapa, for whom the Mādhyamika speaks “from the perspectives of others” (karmapa-feast-fortunate-1578 p. 156); GC does not cite the Karmapa, and he adds what the Karmapa does not use, the index to meditative equipoise. GC’s reading is not on a “Gorampa side”: Gorampa argues against the equipoise reading (next section), and GC’s one mention of Gorampa is a criticism on another point (Adornment ). Within the Geluk, Lopez tentatively finds the same referent in Tsongkhapa’s answers to Rendawa, with a difference in what follows from it (Lopez 2006, commentary on §–85). GC’s independence was itself disputed by his critics: the editor of the 1997 Geluk refutations calls the Adornment an imitation of Gorampa, Shakya Chokden and Taktsang (Lopez 2006, “The Critics”).

Gorampa against the equipoise reading

In Distinguishing the Views ch. 2 (gorampa-distinguishing-views-1469) a defender of Dolpopa confines MMK 7:33, VV 29 and MMK 22:11 to the context of equipoise, so that in the aftermath (རྗེས་ཐོབ་) one may still distinguish what truly exists from what does not. Gorampa rejects this outright. If the non-establishment of the noncomposite applied only in equipoise, so would that of the composite, and the analysis of arising, abiding and ceasing would be pointless; and if the Vigrahavyāvartanī passage applied only in equipoise, the Mādhyamika’s refutation of the realists would belong to equipoise rather than to the aftermath, where it is in fact made. Gorampa’s own reading of v. 29 is the kind-of-thesis reading in the table above (ch. 3). Cabezón’s note to the verse (translation n. 151) numbers it v. 24; the verse quoted is the one standardly numbered VV 29.

The earliest Tibetan witness: Mabja’s twelfth-century VV self-citation

Mabja Jangchub Tsöndrü’s Ornament of Reason (mabja-ornament-of-reason Preliminary Discussion → “The Sixfold Collection of Reasoning → Rebuttal of Objections”) is an early Tibetan commentary on MMK and articulates the no-thesis stance by direct VV self-citation. Mabja first stages the Nyāya self-refutation objection in propria voce (paraphrasing VV vv. 1–2: “if all of these things did not have a nature, your words would also lack nature, and thus be incapable of refuting nature …”) and then quotes Nāgārjuna’s reply by citing VV v. 63 (“As there is nothing whatsoever to negate, I myself negate nothing at all. Therefore, when accusing me of negation, it is you who are guilty of denigration”) and VV v. 23 (“An emanation may refute an emanation; an illusory being may refute an illusion”) directly. Mabja’s gloss makes the no-thesis position explicit within the Two Truths frame: “Because they lack nature, my words cannot, in reality, refute the position of others. Hence, since both the object and the agent of negation are not established in reality, I do not claim that a negation of the positions of others takes place either. Nevertheless, although no nature exists, this does not mean that another’s position cannot be refuted in terms of mere convention, just as an illusory person may defeat another illusory person, and the water in a dream may seem to extinguish fire.”

The reading is pre-sectarian — it pre-dates both the Tsongkhapa narrow reading (theses-with-svabhāva admissible at the conventional level) and Gorampa’s kind-of-thesis reading (mere and reductio theses admissible, autonomous ones not). Mabja’s gloss has affinities with more than one later reading: the Two Truths fold (no thesis at the ultimate level; refutation operates conventionally) is structurally on the Tsongkhapa side, while the “no negation takes place either” formulation is close to the Ninth Karmapa’s, for whom the Mādhyamika’s reasonings refute nothing that truly exists (Feast pp. 156–157). The position is best read as a common pre-sectarian Tibetan inheritance that the later school readings differentiate themselves from — pre-sectarian, not pre-Prāsaṅgika–Svātantrika, since Mabja already rejects the Autonomists’ particular characteristics even conventionally. Mabja’s reading places the no-thesis stance squarely within the VV’s own argumentative context, the Nyāya self-refutation challenge, rather than treating it as a free-standing piece of negative logic. This is the structural feature on which the framework-necessity argument at vv-29-three-readings turns.

What the no-thesis view does not mean

Several historically tempting readings are rejected by every named participant in the tradition:

  • It does not mean “the Mādhyamika asserts nothing” in any literal sense — the Mūlamadhyamakakārikā, the Vigrahavyāvartanī, and the Prasannapadā are filled with assertions, refutations, and commitments. (Even Burton, who reads Madhyamaka as nihilist, takes v. 29 as something other than a literal performative claim of asserting nothing.)
  • It does not mean “the Mādhyamika has no philosophical position” in the sense of methodological quietism. Prasaṅga itself is a positive method, and the conclusion that things lack svabhāva is a substantive metaphysical-epistemological commitment.
  • It does not mean “the Mādhyamika cannot be refuted because she has nothing to refute” — that is the opponent’s misreading at VV v. 1–2.

What the six restricting readings share is that the denial is qualified in some way (by content, by kind of thesis, by semantic register, by truth-level, by perspective, or by occasion and assertion-context), and their disagreement is over which qualification. The two framework-external readings are the exceptions that show what the six have in common. Gokhale’s shows that they all take the verse to assert something, and he does not. Galloway’s shows that they all take it to be a claim about theses, where he takes it to be a report of what the negation always produces.

Linked pages

Graph View

Backlinks