Madhyamaka Wiki

Prāsaṅgika-Svātantrika

17 min read

Definition

The Prāsaṅgika-Svātantrika distinction divides Indian Madhyamaka into two sub-schools based on their method of argumentation. Svātantrikas (associated with Bhāviveka) use independent syllogisms (svatantra-anumāna) with terms whose referents are accepted by both parties. Prāsaṅgikas (associated with Buddhapālita and Candrakīrti) use prasaṅga — consequentialist reasoning that draws out the absurd implications of the opponent’s own premises without advancing a counter-thesis.

The distinction originates in the Indian debate between Buddhapālita and Bhāviveka over the correct method for establishing emptiness, and is codified by Candrakīrti’s defence of Buddhapālita in the Prasannapadā. In Tibet, the distinction became a major axis of philosophical classification.

Four early Tibetan misreadings (per Tsongkhapa): In jinpa-tsongkhapa-qualms-1999, Jinpa reconstructs four positions in the Lam rim chen mo that Tsongkhapa subjects to detailed critique, all relating to misunderstanding Prāsaṅgika methodology: (1) Jayānanda’s epistemological scepticism about tri-modal logic; (2) universal scepticism rejecting all pramāṇa, attributed to Tibetan translators studying under Jayānanda (possibly Khu Lotsawa); (3) “present-day Prāsaṅgikas” who claim no thesis in either conventional or ultimate sense; (4) followers of Candrakīrti (possibly Mabja Jangchub Tsöndrü and followers of Patshap Lotsawa) who mistake the critique of autonomous reasoning for a wholesale rejection of logical argumentation. These represent the earliest layer of Tibetan Prāsaṅgika interpretation — the very positions Tsongkhapa sought to reform.

Historical note: Ruegg demonstrates that the terms svātantrika and prāsaṅgika as school-designations are unattested in Indian sources. They were codified in Tibet by Pa tshab Nyi ma grags and his disciple Jayānanda at the end of the eleventh century. Candrakīrti himself was largely ignored in Indian Buddhist thought for several centuries after his time; real recognition came only around the tenth century, possibly through Atiśa’s influence. Tsongkhapa justified the Tibetan terminology by arguing that the philosophical distinction is consistent with Candrakīrti’s own treatment, even though Candrakīrti did not use the formal nomenclature. (From ruegg-svat-pras-2006)

Convergence across registers — traditional-side restatement of the founder question: Ian Coghlan, working within the traditional Geluk frame as translator of the Buddhapālitavṛtti (coghlan-buddhapalita-2021 Introduction), reaches the same conclusion as Ruegg: “though Nāgārjuna and Buddhapālita clearly taught the Prāsaṅgika system, neither is the founder of the Prāsaṅgika school because historically neither clearly set forth this view in contradistinction to the Svātantrika position.” Coghlan’s phrasing frames Bhāviveka as founder of Svātantrika because he was first to articulate the distinction, and Candrakīrti as founder of Prāsaṅgika for the same reason. The convergence across registers (academic historical-critical Ruegg + Geluk-trained traditional translator Coghlan) strengthens sp-hierarchy-is-tibetan-construction: two independent hermeneutic starting-points reach the same historical conclusion about founder-retrospectivity.

Indian primary-text basis: the Prasannapadā MMK 1.1 commentary

The Tibetan distinction is grounded in a single Indian text: Candrakīrti’s commentary on MMK 1.1 in the Prasannapadā, added as candrakirti-prasannapada-mmk1-excerpt (corresponding to ACIP TD3860 folios 05B–11B). This text quotes Buddhapālita’s argument verbatim, then Bhāviveka’s three-pronged critique from the Prajñāpradīpa, then defends Buddhapālita and rebuts Bhāviveka at length. It is the locus classicus and the textual root of every later development of the distinction.

Four moves in this text are load-bearing for the Tibetan literature:

  1. The “no thesis” defence (section IV) — citing Āryadeva’s Catuḥśataka and Nāgārjuna’s Vigrahavyāvartanī 29 (“If I had a thesis, I would have a fault. / Since I have no thesis, I am strictly faultless”), Candrakīrti grounds the prohibition on autonomous inferences in Madhyamaka’s own non-positionality. This becomes the central Prāsaṅgika dictum, read in subtly different ways by Tsongkhapa (no autonomous thesis, conventional positions preserved), Gorampa (stronger no-thesis position), and the Ninth Karmapa (no thesis at all — the foundation of the three stages of analysis in karmapa-feast-fortunate-1578).

  2. The implicit five-part probative argument (section VI) — Candrakīrti reconstructs Buddhapālita’s argument as containing reason, application, example (the clearly-manifest vase), and conclusion, all using inferences accepted by the Sāṃkhya opponent. This shows that prasaṅga is not formally deficient; it just uses opponent-accepted reasons rather than jointly-established reasons.

  3. The paramārthatas critique (sections IX–X) — Candrakīrti rejects Bhāviveka’s qualifier “ultimately” on multiple grounds: arising-from-self is rejected at both truths (citing MMK 28.10 and the Lalitavistara); the qualifier is communicatively idle for worldly opponents; and the relative phenomena Bhāviveka invokes as subjects are seen by mistaken cognition and so cannot be bases for non-mistaken affirmation. This is the Indian textual root of Ruegg’s criterion (2) in ruegg-svat-pras-2006 — and it is stronger than Ruegg’s neutral phrasing suggests, since Candrakīrti is not merely noting the qualifier but arguing it is meaningless.

  4. The boomerang argument (section XI) — Bhāviveka had himself argued, against the śrāvakas, that a reason like “because the Tathāgata said so” is either non-established (relative) or contradictory (ultimate). Candrakīrti shows this same dilemma applies to Bhāviveka’s own autonomous arguments: their subjects (“the inner sense sources”) and reasons (“they exist”) are equally non-established once Madhyamaka analysis is applied. This is the most rigorously internal critique of autonomous inference in the entire literature.

The defence concludes (section XII) with a meta-argument about logical method: refutation operates through reasons accepted by the counterpart, not through reasons jointly established — “as it is in the world, so it should be in logic, for only the conventions of the world can apply in the treatises of logic.”

The fact that this single passage generated four early-Tibetan misreadings catalogued in jinpa-tsongkhapa-qualms-1999 (Jayānanda; Khu Lotsawa circle; “present-day Prāsaṅgikas”; followers of Patshap Lotsawa), the divergent Tsongkhapa, Gorampa, and Karmapa readings, and Mipham’s pedagogical-convergence reading — is itself the strongest single piece of evidence for the framework-internal-debate-is-productive claim.

Interpretations

Tsongkhapa (in his own words): In Illuminating the Intent, Tsongkhapa presents the Svātantrika identification of the object of negation first (through Kamalaśīla’s Light of the Middle Way), then the subtler Prāsaṅgika identification. The Svātantrika defines true existence as “existence through its own objective mode of being, not posited in dependence on being perceived by cognitions”; the Prāsaṅgika extends the negation further — all phenomena are posited through mere conceptualisation, like a snake imputed on a rope. The Svātantrika accepts “existence by virtue of essential nature,” “existence through intrinsic characteristic,” and “intrinsic existence” on the conventional level; the Prāsaṅgika negates all six synonyms. Tsongkhapa presents the Svātantrika view as “great skilful means to help guide those who are, for the time being, not capable of easily realising” the subtler Prāsaṅgika view — a graduated pedagogical approach. (From tsongkhapa-illuminating-intent-1418, Ch 9)

Tsongkhapa’s substantive-not-methodological framing of the distinction (now primary-grounded): Jinpa’s introduction reconstructs Tsongkhapa’s central thesis on this question (element 9 of the eleven-point programme, pp. 33–34): Candrakīrti’s critique of Bhāviveka’s autonomous inference is not merely a methodological dispute about how best to establish emptiness, but a substantive philosophical disagreement about whether a Mādhyamika can subscribe to any notion of objective intrinsic existence on the conventional level. Tsongkhapa cites three textually-grounded contexts in which Bhāviveka’s residual realism is exposed: (a) Bhāviveka’s assumption that the three elements of a syllogism (subject, reason, example) can be commonly established by both parties — Candrakīrti rejects commonly-established factors precisely because there are no objectively existing referents to ground the commonality; (b) Bhāviveka’s nihilism-charge against the Cittamātra three-natures theory, which assumes the paratantra and pariniṣpanna must exist by intrinsic characteristic if they are to exist at all; (c) Bhāviveka’s distinction between veridical conventional truths (water, faces, real elephants) and distorted conventional truths (mirages, mirror reflections), which Candrakīrti rejects as objective categories — for him, distinctions within conventional truth can be drawn only from the worldly perspective, not in terms of objectively real intrinsic existence. The substantive reading converges with Ruegg’s six criteria (especially criteria 4 and 6) and supplies the strongest single primary-text basis for treating the distinction as ontological as well as methodological. Critics across schools (Karmapa, Mipham, Shakya Chokden, Atiśa) reject this substantive framing — but it must now be engaged in its own primary-text formulation rather than only as a target of polemic. (From tsongkhapa-illuminating-intent-1418, Introduction by Thupten Jinpa; primary text base at Ch 9 on the six synonyms and at Ch 12 on the three reductios against intrinsic-characteristic, MA 6.34–36)

Tsongkhapa (via Gorampa’s critique): Gorampa sees Tsongkhapa as elevating the distinction to a philosophically fundamental divide, not merely methodological but ontological. (From gorampa-distinguishing-views-1469)

Gorampa: While Gorampa accepts the distinction, gorampa-distinguishing-views-1469 situates it within his broader threefold taxonomy. For Gorampa, both Prāsaṅgika and Svātantrika are rangtong positions, and the more fundamental divide is between the genuine “freedom from extremes” view and the two erroneous extremes (Dolpopa’s eternalism and Tsongkhapa’s grasping at emptiness).

Primary-text basis: Gorampa .5 sub-chapter. Gorampa devotes a whole sub-chapter of his refutation of Tsongkhapa to the Prāsaṅgika-Svātantrika distinction, with three further sub-subdivisions: (1) .5.1 — the difference between Prāsaṅgikas and Svātantrikas with respect to theses (dam bca’i khyad par); (2) .5.2 — the difference as regards adequate argumentation (‘thad pa’i khyad par); (3) .5.3 — the basis on which the two-truths division is drawn (bden gnyis gang gi steng du ‘byed pa). This is the Tibetan-side analogue, in load-bearing weight, to Candrakīrti’s MMK 1.1 commentary in the Indian register (cf. candrakirti-prasannapada-mmk1-excerpt and ames-prajnapradipa-ch2-1995). Gorampa’s own .1.2.1 enumerates the five great reasonings (Five Great Reasonings) and notes the Svātantrika–Prāsaṅgika methodological split applies to all five: “Svātantrikas accept these as autonomous syllogisms, while Prāsaṅgikas accept them as ‘reasons acceptable to others’ (gzhan la grags kyi rtags).” That Gorampa is reading the Indian debate one specific way is itself a datum for — the Tibetan reception of the distinction is plural, and Gorampa’s prasaṅga-only reading is one substantive option among others (Tsongkhapa’s hierarchy, Mipham’s pedagogical convergence, the Karmapa’s debunking, Atiśa’s pre-distinction “Great Madhyamaka”).

Ninth Karmapa — methodological, with Autonomists defended as genuine: The Ninth Karmapa’s primary interest in the Svātantrika-Prāsaṅgika distinction is “debunking myths” — particularly the claim, attributed to Tsongkhapa’s tradition, that the Autonomists either misapprehend or cannot fully understand emptiness. For the Karmapa, “the Consequentialists and Autonomists differ in regard to the words they use to communicate, but their intentions are the same.” The Autonomists do not believe conventional phenomena are real on the relative level; they speak of them “from the perspective of their performing functions.” The distinction is primarily one of method for refuting arising from the four extremes, not of beliefs about what constitutes reality. The Karmapa is “often more interested in debunking myths” than in presenting a clearly defined boundary. He notably elevates the Autonomists’ standing, “rescuing their reputation from the status of inauthentic Followers of the Middle Way” assigned by Tsongkhapa. However, he does acknowledge that the Consequentialists find the Autonomist approach problematic — calling anything “valid cognition” ascribes “undeceiving” quality, which entails “a certain level of clinging to conventional terms.” Despite this, the Autonomist approach does lead to progressive realisation of the Middle Way. (From karmapa-feast-fortunate-1578)

Śāntarakṣita / Mipham — pedagogical convergence: Mipham’s commentary on Śāntarakṣita’s Madhyamakālaṅkāra offers a third position: the Svātantrika-Prāsaṅgika distinction is one of pedagogical emphasis, not philosophical rank. Svātantrika emphasises the approximate ultimate (rnam grangs pa’i don dam) — emptiness as conceptually cognised through reasoning — while Prāsaṅgika emphasises the actual ultimate (rnam grangs ma yin pa’i don dam) — emptiness beyond all conceptual elaboration. Both converge on the same actual ultimate truth. This dissolves the hierarchy that Tsongkhapa constructs (Prāsaṅgika as subtler) without dismissing either approach. Śāntarakṣita himself is classified as Svātantrika (specifically Yogācāra-Svātantrika) by the Tibetan tradition, but Mipham argues this classification does not entail a lesser view. (From shantarakshita-madhyamakalankara)

Four further points from the MA commentary sharpen this reading and bear directly on whether the distinction is a rank at all:

  1. The “neither one nor many” argument is formulable as either a consequence or an autonomous inference (MA commentary, “A Prāsaṅgika or a Svātantrika argument?”). The same root reasoning runs as a prasaṅga (drawing the consequence from the opponent’s true-existence assertion) or as a svatantra inference (on the merely mind-posited subject, via “other-elimination”); citing Kamalaśīla’s Madhyamakāloka, Mipham holds the empirical/imputed distinction “unnecessary” and that both methods “refute… equally.” The method is not fixed by the content of the argument — direct support for the reading that the S–P difference is presentational. (Bears on the open question below: “Is the ‘neither one nor many’ argument genuinely svātantra…?” — Mipham’s answer is either, as suits the disciple.)
  2. The genuine no-assertion stance belongs to ārya meditative equipoise. The actual ultimate — on which “the Svātantrikas, like the Prāsaṅgikas, make no assertion” — is “what noble beings on the Bodhisattva grounds see with the utterly stainless primordial wisdom of meditative equipoise”; even Āryas, in post-meditation, “remain within the scope of thought and word, assertion or denial” in order to teach and debate. The Prāsaṅgikas “emphasise the primordial wisdom of the union… in meditative equipoise, while the Svātantrikas emphasise the wisdom that distinguishes the two truths in the post-meditation period.”
  3. An assertion-making “Prāsaṅgika” is, by Mipham’s own criterion, a Svātantrika: “inasmuch as certain ‘Prāsaṅgikas’ remain on the level of the approximate ultimate truth, making assertions about the distinction of the two truths, there is no distinguishing them from Svātantrikas.” Discursive, thesis-bearing Madhyamaka is Svātantrika practice whatever its label.
  4. Mipham’s own diagnostic observation: “in Tibet, even the explanation of the Prāsaṅgika view reverts to that of the Svātantrikas” — i.e. the operative Tibetan “Prāsaṅgika” of the textbooks proceeds by Svātantrika-style distinction-making. This is the textual seed of the wiki author’s working hypothesis (see Open questions and madhyamaka-is-method-not-standalone-system §F): that the four tenets are accepted as a graded conventional ontology, so the Tibetan schools operate as de-facto Svātantrikas and “Prāsaṅgika” names the ārya-equipoise terminus rather than a discursive standpoint.

This MA reading converges with the criterion-clusters already on this page — Ruegg’s six, Dzongsar Khyentse’s six, Phya pa’s six questions — all of which “converge on the status of saṃvṛti”: Mipham’s reduction of the difference to approximate-vs-actual ultimate is itself a saṃvṛti-indexed reading, since what the Svātantrika adds (and the Prāsaṅgika withholds) is exactly an assertion-bearing handling of the conventional/approximate register.

Shakya Chokden — the distinction is philosophically insignificant: Shakya Chokden goes further than any other Tibetan thinker in dissolving the Prāsaṅgika-Svātantrika distinction. He argues that: (1) Prāsaṅgika and Svātantrika share the same identification of the final ultimate truth — “what is beyond objects of sounds and concepts, divested of all collections of proliferations”; (2) they share the same account of conventional phenomena posited by worldly minds; (3) their only difference lies in the type of reasoning they use to determine the ultimate (autonomous reasons vs consequences) and the pedagogical path toward it (direct plunge vs gradual ascent through lower tenet systems). Because these are merely methodological differences, the Prāsaṅgika/Svātantrika distinction cannot serve as a valid division of Madhyamaka. The genuine division is Niḥsvabhāvavāda / Alīkākāravāda (= self-emptiness / other-emptiness). This effectively replaces one axis of classification with another. (From komarovski-visions-unity-2011, pp. 117, 137–139)

Dzongsar Khyentse — six criteria mapped to the syllogism components, derived from Gorampa’s sa bcad: A contemporary teaching transmission (dzongsar-khyentse-madhyamakavatara-2003 on MA 6.8) sets out the distinction as six differences corresponding to the six elements of a Buddhist syllogism: (1) subject (chos can) — Svātantrika requires it specific and mutually accepted, Prāsaṅgika allows “etc.” and only opponent-acceptance; (2) predicate (bsgrub bya’i chos) — Svātantrika adds “in the ultimate truth,” Prāsaṅgika does not (because for them things are not arisen even conventionally); (3) thesis (bsgrub bya) — Svātantrika has one, Prāsaṅgika has none; (4) reasoning (rtags) — Svātantrika requires mutual acceptance, Prāsaṅgika requires only opponent-acceptance; (5) example (dpe) — same asymmetry; (6) syllogism (sbyor ba) — Prāsaṅgika rejects autonomous inference altogether and uses the opponent’s logic to defeat the opponent. Smoke-and-fire example: Svātantrika “there is fire because there is smoke” (forward inference); Prāsaṅgika “there should not be smoke if there is no fire” (reductio against the opponent’s commitment). This is the cleanest pedagogical articulation of the distinction available in English. It converges precisely with Ruegg’s six criteria despite being derived independently from Gorampa’s sa bcad rather than from Ruegg’s textual-historical analysis. The convergence across registers (Khyentse from inside the tradition + Ruegg from academic philology + coghlan-buddhapalita-2021’s traditional-Geluk translator-frame above) is itself evidence that the criteria track something substantive in the Indian texts.

Phya pa Chos kyi seng ge’s twelfth-century six questions (via Westerhoff 2024): westerhoff-candrakirti-2024 (Introduction b, pp. 22–24) reconstructs the clash between Phya pa Chos kyi seng ge (1109–1169, sixth abbot of Gsang phu ne’u thog) and the Indian commentator Jayānanda. Phya pa held that Candrakīrti’s positions undermined Buddhist soteriology; Jayānanda defended them. The six questions: (1) Can the Indian-logic toolkit be used for emptiness, or only prasaṅga? (2) Do Mādhyamikas hold a positive thesis? (3) Can the unenlightened mind be an epistemic instrument (admitting Diṅnāga–Dharmakīrti)? (4) Is the ultimate accessible to language and thought? (5) Do conventional truths have philosophical grounding, or only everyday-practice grounding? (6) Does a Buddha have any cognitive events at all? Westerhoff notes (p. 23) that Tsongkhapa’s response was not Phya pa’s (reject Candrakīrti) but a constructive synthesis: he kept Candrakīrti while folding in the Diṅnāga–Dharmakīrti epistemology side-by-side, generating “substantial criticism within the Tibetan scholastic tradition” (citing Jinpa 2019 and The Yakherds 2022). This is the historically earliest Tibetan articulation of the criteria that later codify into Prāsaṅgika–Svātantrika — pre-Tsongkhapa by two centuries — and is convergent in conceptual cluster with Ruegg’s six and Dzongsar Khyentse’s six. The convergence is now three-layered (Phya pa twelfth-century, Ruegg twentieth-century academic, DKR contemporary oral) and supports the reading that the criteria track something substantive in the Indian sources rather than projecting back from Tsongkhapa. Phya pa’s question 5 (grounding of conventional truth) and question 6 (Buddha cognitions) are recognisable as the same worries Burton raises in burton-emptiness-appraised-1999 — evidence that the nihilism worry is a recurring framework-internal pressure, not a modern Western imposition.

Ruegg — historical-critical, multi-criterial analysis: Ruegg argues that the distinction involves at least six interrelated criteria, all converging on the status of saṃvṛti/vyavahāra (conventional reality): (1) use of autonomous inference (svatantra-anumāna) vs prasaṅga; (2) whether the qualifier paramārthatas (“ultimately”) is affixed to arguments; (3) status of external objects on the conventional level; (4) acceptance of svalakṣaṇa (self-characteristic) on the saṃvṛti level; (5a) the paramārthānukūla (“concordant ultimate”); (5b) the paryāya-paramārtha (discursive/figurative ultimate); (6) the tathya/mithyā-saṃvṛti distinction (veridical vs distorted conventional). These criteria are interrelated, not independent tests. The distinction does not define two frozen, diametrically opposed doctrinal positions but tracks a complex set of philosophical and methodological issues that have been dynamically elaborated in both Indian and Tibetan thought. Ruegg proposes an indeterminacy/complementarity model for understanding the apophatic-cataphatic duality in Madhyamaka. He concludes that the distinction retains “descriptive, taxonomic, analytical, and heuristic usefulness” when handled with historical awareness. (From ruegg-svat-pras-2006)

Textual loci

  • sp-hierarchy-is-tibetan-construction — the Prāsaṅgika-over-Svātantrika hierarchy is a Tibetan codification, not an Indian given; the Indian sources support diversity of method, not a ranked doctrinal divide. Synthesises Ruegg’s terminology evidence, Apple’s Atiśa manuscripts, and the convergent dissolution by Mipham, the Ninth Karmapa, and Shakya Chokden.
  • framework-internal-debate-is-productive — the Buddhapālita-Bhāviveka-Candrakīrti exchange is the earliest Indian instance of the productive-disagreement pattern this concept exemplifies.
  • madhyamaka-is-method-not-standalone-system — the Prāsaṅgika negation of all six synonyms even conventionally (against Svātantrika acceptance of svalakṣaṇa) is the load-bearing evidence that Madhyamaka declines a positive conventional ontology; the Svātantrikas’ borrowing of Sautrāntika/Yogācāra conventional systems is the in-tradition proof that the conventional layer is a swappable input.

Graph View

Backlinks