Madhyamaka Wiki

Gzhan Stong Is Refined Cittamātra, Not Madhyamaka

7 min read

The claim

Dolpopa’s gzhan stong is structurally Cittamātra, not Madhyamaka. The point is not that it is Yogācāra in vocabulary — it draws heavily on Maitreya’s last three works, on Asaṅga, on Vasubandhu — but that it satisfies four diagnostic criteria that classify a position as Cittamātra in the grub mtha’ literature. The argument originates with Red mda’ ba (1349–1412) and is transmitted via Rongtonpa (Shes bya kun rig, 1367–1449), reaching us through Gorampa’s verbatim restatement in of Lta ba’i shan ‘byed. Its earlier expression in Sakya is Drakpa Gyaltsen’s (1147–1216): “you accept that Yogācāras are Mādhyamikas, and a Buddha’s gnosis is free from all proliferations — internally contradictory”. That places the Sakya rejection of zhentong-style positions a century before Dolpopa’s mature work, and three centuries before Tsongkhapa.

Evidence for

  • **Gorampa ** (gorampa-distinguishing-views-1469). The four diagnostic criteria, transmitted from Rongtonpa:
    1. Residue of true-grasping at the pariniṣpanna after only the paratantra is negated. Dolpopa’s zhentong declares the paratantra truthless but retains the pariniṣpanna as truly existent — Yogācāra’s defining structural commitment.
    2. Acceptance of the Saṃdhinirmocana as definitive (nītārtha). Identical to Cittamātra hermeneutics.
    3. Asaṅga, Vasubandhu, Dignāga, Dharmakīrti called “Great Mādhyamikas”. All four classical Cittamātras imported into the Madhyamaka tradition by terminological reclassification, with their textual lineage taken as Madhyamaka lineage.
    4. Three-natures hermeneutic as master interpretive key. The parikalpita / paratantra / pariniṣpanna triple, not the two-truths / prasaṅga analytic, is the principal exegetical framework.
  • Yogācāra-internal supplement at Gorampa : even on Yogācāra’s own terms Dolpopa is wrong, because for Maitreya/Asaṅga/Vasubandhu the paratantra is supposed to “really exist as a substance, like the basis of an illusory manifestation” — not to be empty. So gzhan stong is not even faithful to its declared Yogācāra base; it is sui generis.
  • Earlier Sakya attestation in the Drakpa Gyaltsen boomerang (cited by Gorampa ): the charge that one cannot consistently call the Yogācāras Mādhyamikas while also holding a Buddha’s gnosis free from all proliferations (quoted in full under “The claim” above). Drakpa Gyaltsen is twelfth-century, writing before Dolpopa and before Tsongkhapa. His testimony shows that the Sakya rejection of zhentong-structured positions is not a fifteenth-century Sakya/Geluk alliance against Jonang but a long-standing position internal to Sakya.
  • Structural Sakya/Geluk convergence: Tsongkhapa takes Dolpopa as his principal opponent in Drang nges legs bshad snying po (per Hopkins’s annotations to taranatha-essence-other-emptiness-2007), reaching the same conclusion as Gorampa via different reasoning. The two systems agree on excluding zhentong from Madhyamaka; they disagree about everything else.
  • Thuken’s external structural reduction (thuken-crystal-mirror-1802 chapter 9) — independent of Gorampa-via-Red-mda’-ba and operating at a different level. Where Gorampa’s reduction is internal (zhentong is refined Cittamātra, falling short of Madhyamaka, but still inside Buddhism), Thuken’s is external (zhentong has structural Hindu analogues at three load-bearing points). Three matched parallels:
    1. Śabda-Brahman (Brahman-as-sound) — permanent, beginningless reality transforming into all entities, paralleling the Jonangpa permanent fixed pervader of animate and inanimate.
    2. Sāṃkhya — liberation as dissolving transformations back into a permanent isolated conscious self, paralleling the Jonang account of buddhahood as the appearance of a permanent solitary reality after conventionalities are reduced to nonexistence.
    3. Vedānta + Mīmāṃsaka — permanent all-pervading non-dual knower behind mistaken appearances; the intrinsic nature of mind held to be stained. The two reductions converge on excluding zhentong from Madhyamaka, but they come at it from different sides: Gorampa from inside the Buddhist grub mtha’ hierarchy, Thuken from inter-religious structural homology. When two methodologically independent reductions agree, the shared conclusion carries more weight than either argument alone.
  • Śāntarakṣita as the in-tradition control case (Madhyamakālaṅkāra / shantarakshita-madhyamakalankara). The cleanest demonstration that the zhentong handling of Yogācāra material is the aberrant one is a Mādhyamika who uses the same Yogācāra register correctly. Śāntarakṣita accepts Cittamātra — mind-only appearances, reflexive awareness — as conventional truth. He then turns the “neither one nor many” analysis on consciousness itself: the true existence of consciousness is the Cittamātra’s “weak point” (MA vv. 44–60). What he arrives at is an actual ultimate that is beyond all elaboration, explicitly not truly existent (“the actual ultimate is free from constructs and elaborations”, MA v. 70). This is exactly the move zhentong refuses. Where Śāntarakṣita empties the Yogācāra register ultimately, Dolpopa retains the pariniṣpanna as truly existent. So the first diagnostic criterion (residue of true-grasping at the pariniṣpanna after only the paratantra is negated) sharpens into a contrast within the tradition: the Yogācāra-Madhyamaka founder shows that the correct Madhyamaka use of mind-only is conventional-and-then-emptied, which makes zhentong’s truly-existent ultimate the deviation rather than a rival Madhyamaka reading. One caveat against over-reach. This presupposes the rangtong verdict that the ultimate is a non-affirming negation; the Tāranātha and Shakya Chokden responses below contest exactly that presupposition, so Śāntarakṣita is a control case for the rangtong cohort, not a neutral arbiter.

Evidence against / objections

  • Dolpopa’s own response (see dolpopa-mountain-doctrine-1333): the Mountain Doctrine both confirms and resists the reduction. It confirms the textual markers — the three-natures hermeneutic is the master key, and Asaṅga, Vasubandhu and Maitreya are read as covert Great-Middle proponents (MD 235, 245). But it resists the verdict on two grounds in the primary text that the reduction does not address. First, the ultimate is an affirming negative (ma yin dgag) and “a third category… free from existence and non-existence” (MD 470, 309.7) — not the Yogācāra paratantra-as-substance the first diagnostic criterion targets, and not a flat first-koṭi posit. Second, Dolpopa explicitly handles the MMK and the Collections of Reasonings (e.g. MMK 25’s two nirvāṇas, MD ~447) inside the zhentong reading, so the claim that he abandons the Madhyamaka analytic for the three-natures alone is not quite right: he subordinates the analytic, but he engages it. The “refined Cittamātra” charge is the strongest argument against him, and it is no longer answerable only through Tāranātha’s synthesis.
  • Tāranātha’s response (taranatha-essence-other-emptiness-2007): the four diagnostic criteria assume that the three-natures and the Saṃdhinirmocana are Cittamātra property, but the zhentong claim is that the late Maitreya works and the Uttaratantra — which use the same vocabulary — are Great Madhyamaka (dbu ma chen po). On this counter-reading, calling Asaṅga and Vasubandhu “Great Mādhyamikas” is not category confusion but accurate doxography; the three-natures analysis at the Mahāyāna’s highest register is itself Madhyamaka. The reduction succeeds only if one presupposes that the doxographic boundary excludes zhentong — i.e. it is question-begging.
  • Shakya Chokden’s “grand unity” (komarovski-visions-unity-2011): Shakya Chokden, also Sakya, embraces precisely the move Gorampa rejects — calling Yogācāra (specifically Alīkākāravāda) a form of Madhyamaka on equal footing with Niḥsvabhāvavāda. This is a productive complication: a Sakya master, contemporary with Gorampa, reaches the opposite conclusion via a different reasoning. Shakya Chokden’s reading suggests that the four diagnostic criteria are not as decisive as Gorampa presents them.
  • Selection effect inside Yogācāra: the Cittamātra is a heterogeneous category. Rnam bden pa and rnam rdzun pa schools differ profoundly. Calling all four classical Cittamātras “Great Mādhyamikas” is at least not category confusion at the level of exact identity — Dignāga/Dharmakīrti are pramāṇa scholars, not Cittamātra metaphysicians in the same sense. The third diagnostic criterion overstates its case.
  • Geluk-side critique of the first diagnostic criterion: the Geluk would say that the residue of true-grasping at the pariniṣpanna is exactly the disease that the Geluk med dgag of bden grub cures. So Gorampa’s first criterion is also his own ground for rejecting Tsongkhapa — but Tsongkhapa, on Gorampa’s reading, fails by reifying a negative ultimate, while Dolpopa fails by reifying a positive one. The two diagnoses are structurally parallel; whether the parallelism is a strength (Gorampa’s catuṣkoṭi-based diagnostic is unified) or a weakness (it proves too much) is itself a philosophical question.

Linked pages

Graph View

Backlinks