Madhyamaka Wiki

Framework-Internal Debate Is Productive

9 min read

The claim

Indian and Tibetan Madhyamikas disagree profoundly with one another. They divide over the object of negation, the status of conventional reality, the existence of buddha-wisdom, the meaning of MA 6.23, the soteriological scope of arhat realisation, even the classification of Madhyamaka sub-schools. Yet these disagreements are philosophically productive. They widen the conceptual repertoire, refine the questions, and produce sophisticated alternative resolutions to genuine problems. They do not produce the regress paradoxes (Burton) or the hermeneutical self-undermining (Kalupahana) that mark framework-absent modern readings (cf. framework-absence-yields-nihilism). The framework does not censor disagreement; it is the shared structure within which disagreement becomes intelligible and cumulative.

Evidence for

  • MA 1.8 — the paradigm case (see tenpa-tibetan-battleground-notes). Tsongkhapa and Gorampa agree on the conclusion that arhats realise dharma-nairātmya, yet reach it by opposed methods. Tsongkhapa argues from the seventh-bhūmi alternation; Gorampa works through a programme that refuses to conflate theoretical registers. The Eighth Karmapa introduces a third architecture (“complete in kind, limited in scope”) that dissolves part of the dispute by reframing the question. Mipham dissents on completeness via the seawater analogy and his Two Truths diagnostic. Four positions, sharply distinct, all generated by engagement with the same verse and the same framework. The disagreement between them is itself philosophically generative: the Karmapa’s reframing only becomes possible because Tsongkhapa and Gorampa have first articulated the apparent dilemma.
  • MA 6.23 — three-way disagreement on a single verse. Drawing on Madhyamakāvatāra and karmapa-feast-fortunate-1578: Tsongkhapa reads the “dual natures” as identical in nature but distinct in conceptual identity; Gorampa reads them as quasi-ultimate against real ultimate; the Karmapa reads them as neither same nor different, beyond elaborations. Three philosophically substantive readings, all framework-internal, none collapsing into incoherence, each clarifying what the others assume. That the three non-Geluk readings converge against the Geluk one invites a deflation: three rivals of Tsongkhapa might align against him for sectarian reasons alone. Two things rule that out. They reach the verdict by argumentatively distinct routes (Gorampa’s catuṣkoṭi-based critique; the Karmapa’s phenomenology of clinging; Mipham’s two-ultimates), and each anchors it to the shared root verse — MA 6.23 read against MMK 22:11 and the four-extremes corpus — rather than to school-proprietary premises. The convergence runs toward the primary text, which marks principled rather than polemical agreement. It is the same independence test the VV 29 case passes, where Gendün Chöpel has no school-internal motive to defend the equipoise reading.
  • The Buddhapālita–Bhāviveka–Candrakīrti exchange. The earliest Indian instance. Buddhapālita reads MMK 1.1 by prasaṅga alone; Bhāviveka critiques this in Prajñāpradīpa Ch 2 as methodologically incomplete; Candrakīrti defends Buddhapālita in the Prasannapadā. The exchange did not produce confusion. It produced the entire Indian and Tibetan literature on autonomous inference against consequentialism, including Ruegg’s six-criteria analysis (ruegg-svat-pras-2006) of how the methodological dispute unfolds across at least six interrelated philosophical issues, all converging on the status of saṃvṛti. A single early disagreement seeded centuries of refinement.
  • The four-axis split on buddha-wisdom (see apple-jewels-middle-way-2018). Atiśa holds there is no wisdom-continuum at buddhahood; Tsongkhapa holds omniscient wisdom; Gorampa holds it ineffable; Dolpopa holds a truly existent pristine wisdom. Four incompatible positions, all framework-internal, all generated by sustained engagement with the same Mahāyāna sources. The disagreement clarifies what is at stake in the question, drawing out what would have stayed latent in any single position.
  • Convergent dissolution of the S–P hierarchy (see sp-hierarchy-is-tibetan-construction). Mipham, Ninth Karmapa, Shakya Chokden and Atiśa each arrive independently at minimising or dissolving the hierarchy, and each by a different argument. That convergence is itself a productive outcome of framework-internal debate. Disagreement at one level — the Tsongkhapa hierarchy against its critics — generates new philosophical positions that would not have been articulated otherwise: Mipham’s pedagogical convergence, Shakya Chokden’s Niḥsvabhāvavāda/Alīkākāravāda reclassification.
  • Mipham’s synthesis on the Madhyamakālaṅkāra — productivity as construction, not just disagreement (see shantarakshita-madhyamakalankara and Madhyamakālaṅkāra). This is the clearest case that framework-internal engagement builds rather than merely contests. Śāntarakṣita unifies Cittamātra (the conventional account) with Madhyamaka (the ultimate) in a single system, and Mipham then unifies the Svātantrika and Prāsaṅgika methods on top of it. He reads the S–P split as the approximate-ultimate / actual-ultimate distinction (MA vv. 70–72), so that both methods converge on one actual ultimate: the equipoise object on which “the Svātantrikas, like the Prāsaṅgikas, make no assertion.” This is a third structural shape in the Tibetan debate, distinct from both Tsongkhapa’s graduated hierarchy and Gorampa’s dismissal of that systematisation, and it is generated by engagement with the same framework apparatus (two ultimates, neyārtha/nītārtha, Three Turnings) the disputants share. Synthesis of this kind is the strongest available rebuttal to the Burton-style charge that the tradition only elaborates its assumptions: the framework here produces a genuinely new, coherent integration of two systems and two methods, not a refinement-in-place.
  • Tsongkhapa’s own diagnosis as productive (see jinpa-tsongkhapa-qualms-1999). Tsongkhapa’s three qualms about early Tibetan misreadings, and the four specific positions he critiques in the Lam rim chen mo, are themselves productive framework-internal debate. He identifies real philosophical errors in his Tibetan predecessors and offers principled corrections that even his strongest critics, Gorampa and the Karmapas, engage with on their own terms.
  • The debates converge on a single locus — the status of conventional truth (the strongest structural evidence). What looks from outside like scattered disagreement is, on inspection, repeated triangulation of one hinge: how a Mādhyamika may handle saṃvṛti, the conventional, the join between the two truths. Ruegg’s six criteria for the Prāsaṅgika–Svātantrika distinction “all converge on the status of saṃvṛti / vyavahāra” (ruegg-svat-pras-2006). Tsongkhapa’s substantive reading recasts the entire Buddhapālita–Bhāviveka dispute as one question: whether a Mādhyamika can subscribe to any objective intrinsic existence on the conventional level (jinpa-tsongkhapa-qualms-1999, via the three contexts where Bhāviveka’s residual realism surfaces — commonly-established factors, the nihilism-charge against Cittamātra, and veridical-versus-distorted conventional truths). Gorampa gives over a dedicated sub-subdivision (.5.3, bden gnyis gang gi steng du ‘byed pa) to the basis on which the two-truths division is drawn (gorampa-distinguishing-views-1469). The later object-side / subject-side split is, once more, a disagreement about where the conventional sits relative to the cognising subject: Tsongkhapa’s “two natures found by two cognitive processes” stands against the Sakya–Kagyü–Nyingma coalition, with Dzongsar Khyentse sharpest (the division must be drawn subjectively, never objectively; dzongsar-khyentse-madhyamakavatara-2003). Tellingly, the locus is no Tsongkhapa-era invention. The pre-Tsongkhapa twelfth-century Phya pa Chos kyi seng ge / Jayānanda exchange already poses it as a fixed question-set, most pointedly Q4 (is the ultimate accessible to language and thought?) and Q5 (do conventional truths have philosophical grounding, or only everyday-practice grounding?) (westerhoff-candrakirti-2024 Introduction b). That a millennium of framework-internal disagreement keeps re-deriving the same load-bearing question is the signature of productive, cumulative debate rather than chaos: the framework supplies not merely a shared vocabulary but a shared problem. And the locus is exactly the one the external nihilism charge attacks. Phya pa’s Q5 is recognisably Burton’s worry (per Prāsaṅgika-Svātantrika), and Wood rejects the Two Truths precisely as “a literal or philosophical distinction” (wood-nagarjunian-disputations-1994). The internal debate is therefore not idle refinement. It is the centuries-long forging of the very account of conventional truth that nihilism-charge-refuted then deploys to defeat the charge.
  • The two-level signature of the disagreement (divergence at saṃvṛti, convergence at paramārtha) — sharper than the shared-locus point. The schools do not merely keep re-deriving the same question. They answer it with a characteristic shape. They diverge at the conventional level, each supplying a different conventional (Sautrāntika external realism, Yogācāra mind-only, Tsongkhapa’s pramāṇa-warranted conventional, Gorampa’s snang tsam), while converging at the ultimate level: freedom from the four extremes, the non-affirming negation, no intrinsic nature (རང་བཞིན་, svabhāva). That shape is the fingerprint of a method operating on a received, interchangeable conventional rather than a system carrying its own. It is positive structural evidence for madhyamaka-is-method-not-standalone-system, and (see the “not unique to Madhyamaka” reply below) the decisive disanalogy with generic commentarial traditions, which share a canon and a vocabulary but need exhibit no such systematic level-split.

Evidence against / objections

  • Productive does not mean converging on truth. Sophisticated disagreement can refine the conceptual repertoire indefinitely without ever resolving the underlying question. The claim should not be that framework-internal debate eventually converges on a correct answer, only that it does not collapse into incoherence. The Burton-style charge that the tradition merely reproduces its own assumptions in ever more elaborate ways is not directly answered by demonstrating productivity.
  • Productive disagreement is not unique to Madhyamaka. Aristotelian, Thomist and Confucian commentarial traditions also produce sophisticated disagreement within shared frameworks. So the observation that framework-internal debate is productive may track a generic property of mature commentarial traditions, not anything distinctive about the Mahāyāna framework. This weakens the framework-necessity inference unless it is paired with the negative argument from framework-absence-yields-nihilism. Partial reply: the convergence-on-conventional-truth finding (above) is more framework-specific than generic commentarial disagreement, since the debates triangulate one structural feature of the two-truths apparatus itself rather than merely a shared canon. But this sharpens the response rather than settling it, since other traditions plausibly have their own signature loci too. Stronger reply (the two-level signature): the decisive disanalogy is not that the debates share a question but that their answers share a shape — divergence at saṃvṛti, convergence at paramārtha (see the signature bullet above). A Thomist or Confucian tradition can disagree productively within a shared canon without systematically fixing the ultimate while varying the conventional. That level-split is specific to a framework whose ultimate analysis is constant precisely because it is a method run over a received conventional. The framework-necessity inference therefore does not rest on productivity per se (which is indeed generic) but on this signature, which points to what kind of thing the Mahāyāna framework is. That is why the objection is answered here rather than merely deflected onto the negative argument.
  • Selection effect. The paradigm cases of productive disagreement chosen here — Tsongkhapa, Gorampa, the Karmapa, Mipham — are the ones that worked. There may be Tibetan disputes that did produce incoherence and that have not been surveyed. As stated, the claim is vulnerable to the reply, “but you have only looked at the successful cases.”
  • Framework-internal debate is itself sometimes destructive. The marginalisation of Shakya Chokden — Geluk suppression of his woodblocks, Sakya preference for Gorampa — noted in komarovski-visions-unity-2011, shows that working within the framework generates real institutional risk. “Productive” must mean philosophically productive, not socially harmless.

Linked pages

Graph View

Backlinks