Definition

The Yogācāra–Madhyamaka boundary is the doctrinal line separating two of Indian Mahāyāna’s three principal philosophical schools: Yogācāra (also Cittamātra, vijñaptimātratā, “consciousness only”) and Madhyamaka (the “middle way” school of Nāgārjuna). Both schools share Mahāyāna’s prajñāpāramitā base and the doctrine of universal niḥsvabhāva; they diverge over (i) whether dependent nature (paratantra) and consciousness (vijñapti, ālayavijñāna) are exempted from the negation; (ii) whether the absolute is best characterised positively as luminous awareness or negatively as the freedom of phenomena from intrinsic nature; and (iii) whether the tathāgatagarbha tradition is read as a Yogācāra-internal positive metaphysics or as a Madhyamaka-compatible neyārtha teaching pointing to śūnyatā.

The boundary matters in this wiki because three live interpretive disputes turn on it: the contradiction over Candrakīrti’s alleged Vedānta-tendency (Kalupahana vs Glasenapp + Westerhoff); the comparative-structural reading of zhentong as Yogācāra-recursive (Red mda’ ba’s reduction in zhentong-is-refined-cittamatra-not-madhyamaka); and Śāntarakṣita’s two-step method (Madhyamakālaṅkāra) that synthesises the schools without collapsing the boundary.

Indian textual loci

  • Candrakīrti, Madhyamakāvatāra 6.45–97 — the systematic Madhyamaka refutation of vijñaptimātratā and the ālayavijñāna. Candrakīrti reads Yogācāra as illegitimately exempting consciousness from universal niḥsvabhāva; if everything else is empty of svabhāva, consciousness must be too. This is the locus classicus for the Madhyamaka side of the boundary.
  • Bhāviveka, Madhyamakahṛdaya / Tarkajvālā (Ch 5) — pre-Candrakīrti Madhyamaka critique of Yogācāra. Distinct from Candrakīrti’s in being more conciliatory but unambiguously refusing to exempt consciousness.
  • Śāntarakṣita, Madhyamakālaṅkāra — a synthesis in which Yogācāra serves the conventional level (mind-only as the right account of the relative) and Madhyamaka the ultimate (mind itself, pursued, dissolves into emptiness). This is not the dissolution of the boundary but its repositioning: the schools are placed sequentially, not on a par.
  • The eight Indian MMK commentators (Coghlan, citing Saito 1984) — Tibetan tradition classifies four (Akutobhayā, Buddhapālita, Bhāviveka, Candrakīrti) as Madhyamaka and four (Devasharman, Guṇaśrī, Guṇamati, Sthiramati) as Yogācāra. The boundary is therefore not an abstract typology imposed in retrospect; it cuts through MMK’s own commentarial reception.
  • The third turning sūtras (Saṃdhinirmocana, Laṅkāvatāra, the tathāgatagarbha sūtras) supply the scriptural base for Yogācāra-leaning readings; Madhyamaka treats them as neyārtha (provisional) per the standard prajñāpāramitā-as-definitive hermeneutic.

Tibetan reception

Tibetan tradition systematises the boundary via the grub mtha’ (siddhānta) doxographies. Standard four-school taxonomy: Vaibhāṣika / Sautrāntika / Cittamātra / Madhyamaka, with Cittamātra further split into True Aspectarian (Satyākāravāda, rnam bden pa) and False Aspectarian (Alīkākāravāda, rnam brdzun pa). Madhyamaka in turn splits into Svātantrika and Prāsaṅgika (Prāsaṅgika-Svātantrika). Within the Yogācāra-Madhyamaka relation specifically, four positions are visible across the wiki:

PositionHoldersReading of the boundary
Sharp boundary, Madhyamaka criticalCandrakīrti, Tsongkhapa, Gorampa (mostly), Mipham (mostly)Yogācāra is neyārtha; vijñaptimātratā and ālayavijñāna must be subjected to Madhyamaka analysis
Synthesis (sequential)Śāntarakṣita, Mipham (in his Madhyamakālaṅkāra gloss)Yogācāra at the conventional level; Madhyamaka at the ultimate. Boundary preserved but ranked
Unification (substantive)Shakya Chokden (via komarovski-visions-unity-2011)Alīkākāravāda Yogācāra and Niḥsvabhāvavāda Madhyamaka teach the same ultimate — both are valid paths and both end at the same view
Yogācāra-recursive zhentongDolpopa, TāranāthaWhat Yogācāra calls paratantra / luminous awareness is what zhentong calls the gzhan stong ultimate; the rangtong reading is incomplete. Per zhentong-is-refined-cittamatra-not-madhyamaka, this is structurally a Yogācāra position dressed in Madhyamaka vocabulary

These four positions are the live alternatives in Tibet; later doxography (Tsongkhapa, Thuken, Geluk lineage broadly) standardises position 1 as authoritative and reads positions 3 and 4 as Madhyamaka-deficient.

Modern scholarly relevance

Two contradictions in the wiki turn directly on which side of the boundary is the locus of Mahāyāna’s drift toward substantialism:

  • The Candrakīrti / Vedānta charge. Kalupahana in kalupahana-mmk-1986 alleges that Candrakīrti led Madhyamaka toward a “Vedāntic interpretation” that contributed to Buddhism’s disappearance from India. glasenapp-vedanta-buddhism-1950 and westerhoff-candrakirti-2024 independently locate the Vedānta-approaching tendency in Yogācāra (Asaṅga, Vasubandhu) — the school of which Candrakīrti is the principal Indian critic. The boundary matters because Kalupahana’s charge fails if Glasenapp and Westerhoff are right about the location of the Vedānta-approaching tendency. See glasenapp-westerhoff-vedanta-unification.
  • Yogācāra-recursive zhentong. Red mda’ ba’s four-point reduction (transmitted via Rongtonpa and Gorampa) reads Dolpopa’s zhentong as “refined Cittamātra, not Madhyamaka”: the gzhan stong ultimate has the structural shape of Yogācāra’s paratantra exempted from the universal niḥsvabhāva negation. The boundary debate is therefore the doxographic instrument by which Sakya/Geluk reject zhentong as Madhyamaka. With the Mountain Doctrine now added (dolpopa-mountain-doctrine-1333), both sides are primary-checkable: the three-natures-as-master-hermeneutic and Dolpopa’s reclassification of Asaṅga and Vasubandhu as covert Great-Middle proponents (MD 235, 245) are right there in the text — corroborating the “refined Cittamātra” diagnosis — but so is Dolpopa’s reply that the matrix is “free from existence and non-existence” and “a third category,” not a Yogācāra substance (MD 309.7, ~13264). See zhentong-is-refined-cittamatra-not-madhyamaka.

A third structural use: the boundary is one principal axis along which the Madhyamakāvatāra and the Madhyamakālaṅkāra are read as complementary or competing texts. Tibetan doxography uses MA for the sharp-boundary reading and MAL for the synthesis reading.

Linked pages