Chapter 4 of Jan Westerhoff’s The Golden Age of Indian Buddhist Philosophy (OUP, 2018), pp. 217–281. Page numbers below are printed-book pages, taken from the page breaks recorded in the extract; a claim that spans a break is cited to the page where the discussion sits.
Main argument
Westerhoff traces the logico-epistemological school (pramāṇavāda) from Diṅnāga to the end of Buddhist philosophy in India, and reads it throughout as a school shaped by the people it argued with. The biographies already make debate with non-Buddhists the main arena of philosophical work (p. 219), and the doctrinal sections are organised by what Dharmakīrti had to concede or defend to hold a position against Mīmāṃsā and Nyāya.
Two of his readings do the most work for us. The first is the “sliding scales of analysis” (pp. 250–259): Dharmakīrti argues mainly from the particularist level, which he knows to be false, because it is the lowest common denominator, “the best balance between conceptual sophistication and widespread acceptance” (p. 255). The second is the treatment of scriptural authority as a practical necessity (agatyā) rather than an epistemic instrument (p. 243). Dharmakīrti defends it as a fallible choice, not as knowledge, since the qualities that would make the Buddha authoritative are themselves supersensible (pp. 242–243).
Westerhoff also rejects the often-repeated reading on which Diṅnāga’s foundationalism answers Nāgārjuna’s attack on the pramāṇas in the Vigrahavyāvartanī. His reason is narrow and worth keeping: what Nāgārjuna denies is that the epistemic instruments deliver authoritative knowledge by their own nature (svabhāvatas), and an argument that knowledge cannot be grounded in that way is not an argument that there is no knowledge. On Westerhoff’s account there was no sceptical position for Diṅnāga to be answering (pp. 224–225).
Chapter breakdown
1. The Lives of Diṅnāga and Dharmakīrti (pp. 217–220)
Reads the hagiographies for what they show about the discipline rather than for biography: the arena shifts decisively towards debate with non-Buddhists.
- The composition story of the pramanasamuccaya — the cave, the chalk erased by Kṛṣṇamunirāja, and Mañjuśrī catching the chalk in mid-air (p. 217)
- The lineage Vasubandhu → Diṅnāga → Īśvarasena → Dharmakīrti; Dharmakīrti’s understanding surpassed Īśvarasena’s, which is what prompted the pramanavartika (p. 218)
- Dharmakīrti infiltrates Kumārila Bhaṭṭa’s household to learn his doctrines, then defeats him (p. 218); defeats Śaṅkara, who drowns himself in the Ganges (p. 219)
- The biographies place increasing weight on debate with non-Buddhists as the arena of philosophical activity (p. 219)
2. Epistemology (pp. 220–225)
Sets out the two-object, two-instrument scheme and then presses the hard case: how a perception restricted to particulars can reach anything at all.
- Diṅnāga opens with perception and the distinction between specifically-characterised (svalakṣaṇa) and generally-characterised (sāmānyalakṣaṇa) (p. 221)
- The Buddha addressed as pramāṇabhūta: authoritative not by status but because his awakening follows from correct use of the epistemic instruments (p. 220)
- Perception as nonconceptual and indescribable — a “total but unconceptualized, prelinguistic image” (p. 221)
- The problem of aggregates: if perception reaches only aggregates of infinitesimal particles, and aggregates are conceptual superimpositions, how does perception give access to the nonconceptual? Later authors answer with singularity of effect rather than singularity of entity (p. 222)
- This difficulty is what moves Dharmakīrti to drop realism about external objects in favour of Yogācāra idealism (p. 223)
- Perception is by its very nature immune from error; all error is relegated to the mind (p. 223)
- Three kinds of perception in Diṅnāga: sensory, mental, and yogic (p. 224)
3. Inference (pp. 225–231)
Argues that anumāna is not formal logic, and that the two kinds of inference differ structurally rather than in point of view.
- Inference covers much more than the Western notion does: for Diṅnāga even a perceptual judgement (“this tomato is red”) is inferential (p. 225)
- Triple mark (trairūpya): pakṣadharmatā, anvaya, vyatireka; the relation between reason and property to be established is pervasion (vyāpti) (p. 226)
- Inference for oneself is a sequence of psychological states; inference for others is its linguistic expression. Not a difference of perspective but of structure (p. 227)
- Dharmakīrti requires only two parts of an inference for others — the assertion of pakṣadharmatā and the pervasion — with no need to state the thesis (p. 228)
- Both kinds are tokens, not types: concrete events rather than abstract structures, yet not psychologistic, since worldly facts determine validity (p. 228)
- Role of the examples: the positive example assures us the properties are real, the negative example shows due diligence (p. 229)
- Indian inference is fallibilist and externalist; a “good reason” (saddhetu) needs soundness, relevant doubt, a desire to know, and strategic timing (p. 230)
4. Metaphysics (pp. 231–235)
Locates Dharmakīrti’s departure from Diṅnāga in a single move — causal efficacy as the mark of the real — and follows what it costs.
- Only svalakṣaṇa entities have causal efficacy (arthakriyāsamartha), which is the mark of the real; Pramāṇavārttika 3:3 is cited with Manorathanandin’s additions (p. 231)
- momentariness and causal efficacy coincide: trans-temporal objects cannot be efficacious (p. 232)
- Two arguments for momentariness: the argument from cessation (vināśitvānumāna), on which things perish spontaneously, and the argument from existence (sattvānumāna), on which real things constantly produce effects and so constantly change (pp. 232–233)
- Consequences of rejecting object-types in general: arguments against the ātman, against a creator god, and against the brahmanical concept of caste (jāti = object-type = caste) (p. 233)
- Perceived and inferred objects are radically distinct, so Diṅnāga rejects pramāṇasamplava, the mixing of epistemic instruments (p. 234)
5. Language — Apoha (pp. 235–238)
- apoha as a substitute for object-types that incurs no realist commitment (p. 235)
- Two kinds of negation in sequence, non-implicative (prasajya) then implicative (paryudāsa); they do not cancel because they operate differently (pp. 236–237)
- Dharmakīrti’s innovation is to relativise causal power to human desires: all particulars are unique, but they may satisfy a common need. The example is antipyretic drugs, which lower fever by causally different routes (p. 237)
6. Scriptural Authority and Yogic Perception (pp. 238–250)
The longest section and the one that most resists a tidy conclusion: Dharmakīrti proves the Buddha’s authority and then withholds epistemic status from the proof.
- Scripture is subsumed under inference, so no third instrument (śabda, testimony) is needed (p. 241)
- The proof of the Buddha’s authority runs through the Pramāṇasamuccaya’s opening verse: compassion → practice → teaching → embodiment of the epistemic instruments. “Compassion is the proof” (p. 239)
- The threefold analysis (དཔྱད་པ་གསུམ་, dpyad pa gsum): checking scripture against perception, against inference, and for internal consistency (p. 239)
- Three epistemic classes: manifest (pratyakṣa), imperceptible but inferable (parokṣa), radically inaccessible (atyantaparokṣa) (p. 239)
- Dharmakīrti is sceptical about scriptural inference: we cannot establish that the Buddha has the qualities that would make him authoritative, since mental states are supersensible (pp. 242–243)
- Scriptural authority as practical necessity (agatyā), a rationally defensible fallible choice rather than an objective instrument (p. 243)
- Yogic perception (yogipratyakṣa) is not an epistemic super-power. It is directed at concepts rather than at momentary particulars, and Dharmakīrti compares it to vivid dreams and fantasies; it is validated by its soteriological goal, not by disclosing ultimate reality (pp. 247–250)
7. How to Classify Diṅnāga’s and Dharmakīrti’s Philosophy (pp. 250–259)
The section this page is most used for.
- Four levels: ordinary beings, Abhidharma reductionism, particularism (Sautrāntika), idealism (Yogācāra) (p. 253)
- Dharmakīrti argues mainly from level 3, which he knows to be false, because it is the lowest common denominator (p. 255)
- The neither-one-nor-many argument is what drives the movement through the levels, and Westerhoff raises the question whether a fifth stage could be reached by applying it again (p. 255)
- The causal relation is itself unreal by Dharmakīrti’s own particularist criteria, a puzzle the sliding-scales reading is meant to dissolve (pp. 256–258)
- Jitāri and Mokṣākaragupta read Dharmakīrti as a Mādhyamika; Westerhoff frames this as “doing philosophy on the basis of texts regarded as authoritative” rather than as doxography (pp. 258–259)
8. Relation to Mīmāṃsā (pp. 259–270)
- Epistemology: Mīmāṃsā’s intrinsic authoritativeness of the pramāṇas (svataḥ prāmāṇya) against the Buddhist externalist and fallibilist account (pp. 261–263)
- Philosophy of language: eternal authorless Vedas with a primordial word–world connection, against conventional speaker-dependent language and apoha (pp. 263–266)
- Epistemic optimism against pessimism: Mīmāṃsā trusts untrained awareness; Buddhism treats it as shot through with ignorance (pp. 267–268)
- A social dimension underneath the doctrinal one: Brahmins as defenders of the status quo, Buddhists as sceptics detached from brahmanical power structures (pp. 268–270)
9. The End of Buddhist Philosophy in India (pp. 270–281)
- Śāntideva and Atiśa Dīpaṃkaraśrījñāna treated briefly
- Tāranātha judges that nothing after the “six ornaments” (Nāgārjuna, Āryadeva, Asaṅga, Vasubandhu, Diṅnāga, Dharmakīrti) matched them; Westerhoff calls the judgement unreasonably harsh while granting part of it (p. 270)
- Atiśa arrives in Tibet in 1042, the conventional start of the later dissemination (ཕྱི་དར་) (p. 278)
Key claims
- Causal efficacy (arthakriyāsamartha) is Dharmakīrti’s criterion of the real, and it is his innovation rather than Diṅnāga’s (p. 231, citing PV 3:3).
- Perception is by nature immune from error; every error belongs to the mind (p. 223). Together with the aggregates problem, this is what pushes Dharmakīrti from external realism to idealism (p. 223).
- The aggregates problem is answered by later authors with a singularity of effect rather than a singularity of entity (p. 222).
- Inference for oneself and inference for others differ structurally, not perspectivally, and both are tokens rather than types (pp. 227–228).
- Inference is fallibilist and externalist, and a good reason is defined partly by pragmatic conditions — relevant doubt, a desire to know, timing (p. 230).
- Diṅnāga rejects pramāṇasamplava: because perceived and inferred objects are of different kinds, the instruments cannot converge on one object (p. 234).
- Apoha needs both negations, non-implicative then implicative, and they do not cancel because they work differently (pp. 236–237).
- Dharmakīrti grounds the common reference of a term in a common human need rather than a common property — the antipyretic-drugs example (p. 237).
- Scripture is not a third epistemic instrument; it falls under inference (p. 241), and reliance on it is a practical necessity (agatyā) rather than a source of knowledge (p. 243).
- Yogic perception takes concepts as its objects and is validated soteriologically, not by access to ultimate reality (pp. 247–250).
- Dharmakīrti argues mainly from a particularist level he knows to be false, chosen as the lowest common denominator for inter-school debate (p. 255).
- By Dharmakīrti’s own criteria the causal relation is unreal, which the sliding-scales reading is invoked to resolve (pp. 256–258).
- Nāgārjuna’s Vigrahavyāvartanī is not a sceptical text, so Diṅnāga’s foundationalism cannot be explained as a reply to it (pp. 224–225).
Notable quotes
- “Having bowed down to Him, who embodies the epistemic instruments, who seeks the benefit [of] the world, the teacher, the well-gone, the protector.” — the pramanasamuccaya homage verse, quoted p. 217
- “Compassion is the proof” of the Buddha embodying an epistemic instrument. — p. 239
- Someone who aspires to liberation “cannot proceed without relying on the validity of scripture”, though there is no objective fact making scripture valid. — p. 243
- Reliance on scripture rests on “the absence of any other way (agatyā) leading to the aim of liberation.” — p. 243
- Antipyretic drugs “all answer to the human need for a medicine that lowers fever, but they do so in causally different ways.” — p. 237
- Dharmakīrti argues “from a perspective he knows to be false”, since only one chapter of the Pramāṇavārttika consistently takes the idealist view. — p. 255
- He considered the particularist perspective “the lowest common denominator, the best balance between conceptual sophistication and widespread acceptance.” — p. 255
- Diṅnāga’s project as an “exasperated attempt to secure the possibility of valid cognition” — the reading of the Nāgārjuna connection that Westerhoff quotes and then rejects, p. 224
- “There was practically none equal to the older ācāryas.” — Tāranātha on the period after the six ornaments, p. 270
Ligature note: the extraction drops fi/ffi/fl, so “benefit” and “different” above are restored from “benet” and “di|erent” in raw/extracted/Dignaga-and-Darmakirti.md.
Relevance to the three main texts
- pramanasamuccaya: substantial. The opening verse is quoted and made to carry the proof of the Buddha’s authority (pp. 217, 239); Diṅnāga’s two-object scheme, the triple mark, apoha, the three kinds of perception, and the rejection of pramāṇasamplava are each set out (pp. 221–235).
- pramanavartika: the main object of the doctrinal sections. Causal efficacy with PV 3:3 (p. 231), the two momentariness arguments (pp. 232–233), the sliding scales (pp. 250–259), scriptural authority (pp. 238–244) and yogic perception (pp. 247–250). Nothing at verse level beyond PV 3:3.
- pramanayuktanidhi: not discussed, and Westerhoff has nothing on the Tibetan reception at all. The nearest contact is his account of apoha (pp. 235–238) and of the split between perception and inference (p. 234), the two issues Sa-paṇ takes up in Treasury chapters 3 and 4.
Critical assessment
An orientation, not a resource for verse work. The passages quoted from the three primary texts reach Westerhoff through secondary translations (the Pramāṇasamuccaya opening verse, p. 217; PV 3:3 with Manorathanandin, p. 231), and no Tibetan is cited. What the chapter supplies that our other sources do not is the opponent side: the Mīmāṃsā material (pp. 259–270) states positions the Tibetan commentaries usually assume rather than argue.
One tension is left standing. At p. 223 Dharmakīrti drops realism about external objects in favour of Yogācāra idealism; at p. 255 he argues mainly from the particularist level. The sliding-scales framework is meant to reconcile the two, but Westerhoff sets it out in section 7 without returning to the p. 223 claim, so how much idealism the argument actually needs is not settled here. Compare Dunne’s version of the same framework before relying on it.
Translation-relevant material — FLAG
No verse translations usable as a base text. The chapter quotes the pramanasamuccaya opening verse in English (p. 217) and PV 3:3 with Manorathanandin’s commentary (p. 231), both from secondary sources rather than from Tibetan or Sanskrit editions. PV 3:3–4 and a further PV line appear in Sanskrit in the endnotes (p. 281), which are worth consulting when that chapter is translated.
Sources
- dreyfus-recognizing-reality — covers much of the same ground from the Tibetan reception side