Valid Cognition (ཚད་མ་, pramāṇa)
The central concept of the Buddhist logico-epistemological tradition. The term pramāṇa can mean either the means (instrument) of valid cognition or valid cognition itself. Buddhist epistemologists (and the Prabhākara Mīmāṃsakas) choose the second meaning: pramāṇa is valid cognition itself, not a separate instrument that produces it (dreyfus-recognizing-reality, p. 289).
Definitions
Dharmakīrti (Pramāṇavārttika)
མཚན་གཞི་ (Definiendum): ཚད་མ་ (pramāṇa) — valid cognition
མཚན་ཉིད་ (Definiens): མི་བསླུ་བའི་ཤེས་པ་ (avisaṃvādi-jñāna) — nondeceptive cognition
མཚོན་བྱེད་ (Illustration): Perception of fire that enables one to deal with the fire appropriately
dharmakirti provides two complementary characterisations:
Account 1 — Nondeceptiveness:
“Valid cognition is that cognition which is nondeceptive (མི་བསླུ་བ་, avisaṃvādi). Nondeceptiveness consists in the readiness [for the object] to perform a function.”
English: A cognition is valid if, and only if, it is nondeceptive — meaning the object it represents can actually perform the function attributed to it.
Source: pramanavartika, Commentary II; dreyfus-recognizing-reality, p. 289
Account 2 — Revealing the unknown:
“Or, [pramāṇa] is the revealing of a [yet] unknown thing.”
English: A cognition is valid if, and only if, it reveals some hitherto unknown aspect of reality — combining truth and novelty.
Source: pramanavartika, Commentary II; dreyfus-recognizing-reality, p. 290
Account 3 — Unified (from the Pramāṇaviniścaya):
“[Perception and inference are valid cognitions] because they are nondeceptive with respect to the purpose [of the action] in the application [toward an object] after having determined it.”
English: A cognition is valid when it is both practically nondeceptive (leading to successful action) and normatively correct (the object was correctly determined beforehand).
Source: Pramāṇaviniścaya (Ascertainment); dreyfus-recognizing-reality, p. 291
The tension between the first two accounts has generated extensive commentary. Account 1 is practical/pragmatic: validity is a matter of causal connection and successful outcomes. Account 2 is normative/intentional: validity requires that the cognition be directed toward an object in accordance with the object’s nature, and that the content be new. sakya-chok-den explains that nondeceptiveness is primarily of the object (its disposition to perform a function) but can be extended to consciousness (its apprehension of the object in accordance with the object’s causal dispositions) (dreyfus-recognizing-reality, p. 289). The unified account from the Ascertainment combines both elements: (a) practical nondeceptiveness, (b) correct determination of the object, and (c) novelty.
Indian commentators disagreed on how to reconcile the two accounts:
- Devendrabuddhi: either definition will do (a choice)
- Prajñākaragupta: the Commentary defines valid cognition by combining both statements
- dharmottara: emphasises the practical element but adds a normative corrective (the mirage example)
Novelty Requirement
An important consequence: memory (སྨྲ་པ་, smṛti) is not valid cognition. Even though a memory can be true, it does not bring anything new to the cognitive process and has no demonstrable link to reality to ensure its validity (dreyfus-recognizing-reality, p. 290).
Two Types
Dharmakīrti recognises exactly two types of valid cognition:
- perception (མངོན་སུམ་, pratyakṣa) — nonconceptual, direct apprehension of the specifically-characterised
- inference (རྗེས་དཔག་, anumāna) — conceptual, mediated cognition of objects through reasoning
Dharmottara’s Mirage Example
dharmottara illustrates the insufficiency of practical success alone: imagine seeking water on a hot day; you see what appears to be water (actually a mirage) and proceed; by luck, you find water under a rock at that spot. This is not genuine knowledge because the success does not correspond to the previous determination of the object (dreyfus-recognizing-reality, p. 292).
Gorampa’s Three-Feature Reading of “Nondeceptive”
gorampa glosses Dharmakīrti’s definition at PV 2.1 (“Validity is a nondeceptive consciousness / existing through having the capacity / to perform a function; it is undeceiving”) with three aspects of nondeceptiveness — calling this “the exclusive Sakya way of explaining nondeceptive” (gorampa-pramanavartika, §33):
- Object — valid cognition is undeceiving toward objects that are to be engaged or isolated
- Agent — a cognition that embraces its phenomenon exactly as it exists is an undeceiving cognition
- Action — when engaged in an intrinsically existing phenomenon, it is nondeceptive toward it as existing through the capacity to perform a function; by isolating a non-intrinsically existing phenomenon, it is undeceiving toward it as being unable to perform a function
Gorampa treats “nondeceptive consciousness” and “revealing unknown phenomena” as synonymous formulations of a single definition, against Devendrabuddhi (two separate definitions) and the Ornament commentaries of Prajñākaragupta and Yamāri (separate definition and separate definiendum). Both alternative positions are refuted in sakya-pandita’s Treasury of Reasoning.
Wholly Designated Valid Cognition vs. Innate Characteristic
Gorampa distinguishes two levels in Dharmakīrti’s treatment of how validity is ascertained (PV 2.4–5):
- The “consciousness” aspect of pramāṇa is an innate characteristic (རང་གི་མཚན་ཉིད་) — ascertained directly by self-awareness (“from itself, it knows its own nature”)
- Nondeceptiveness is determined by a subsequent ascertaining cognition (Gorampa: “subsequent ascertaining consciousness”, ངེས་ཤེས་) as a cognitive designation (“from designation, it is a valid cognition”)
The famous scheme of “ascertainment from self” vs. “ascertainment from others” (two object-knowers, two self-knowers, inference — from self; initial cognition, unengaged cognition, cognition with cause of error — from others) is read as operating exclusively at the level of wholly designated valid cognitions. Gorampa credits this reading to the “unique commentarial tradition of Jamyang sakya-pandita” and charges that “most Tibetans” collapse the two levels and therefore cannot consistently ascertain either kind (gorampa-pramanavartika, §33).
Sapan’s Account
sakya-pandita defines valid cognition as consciousness that is non-deceptive in relation to action, agent, and object. He and gorampa take Dharmakīrti’s two formulations (nondeceptive cognition and revealing the unknown) as a single complete definition expressed in two ways (sonam-introduction-treasury).
Sapan identifies only three classes of non-valid cognition: non-apprehending cognition (མ་རྟོགས་པ་), doubt (ཐེ་ཚོམ་), and wrong cognition (Sönam: “distorted cognition”, ལོག་ཤེས་). He rejects the earlier Tibetans’ five-type scheme — particularly investigative thought (Sönam: “correctly assuming consciousness”, ཡིད་དཔྱོད་), which he regards as a pure fabrication, and appearance-without-ascertainment (Sönam: “inattentive cognition”, སྣང་ལ་མ་ངེས་པ་), a pseudo-classification (sonam-introduction-treasury). For the full Sa-gya system — definitions, sub-divisions, the dilemma-refutation of investigative thought, and bounding pervasions — see non-valid-cognition and subsequent-cognition.
He also refutes the earlier Tibetans’ distinction between explicit and implicit realisation (དངོས་རྟོགས་དང་ཤུགས་རྟོགས་), arguing that affirmation and negation mean gaining certainty at the epistemological level, not affirming or negating things ontologically (sonam-introduction-treasury).
Ge-luk vs. Sa-gya Interpretations
The two traditions diverge significantly on the implications of valid cognition. A key issue is whether conceptual cognition (inference) can be genuinely valid given Dharmakīrti’s antirealist ontology, which restricts reality to individuals. See can-inference-be-valid for this debate.
Mīmāṃsā Challenge: Intrinsic Authoritativeness
Kumārila argues that the Buddhist approach — settling the trustworthiness of a perception by appeal to a second perception (e.g., confirming the water’s causal efficacy) — leads to an infinite regress, since the second perception’s status is no more secure than the first. He concludes that epistemic instruments are intrinsically authoritative (svataḥ prāmāṇya): the causes of a cognition’s veridicality are among the causes of the cognition’s arising itself. A cognition is authoritative by default, unless other epistemic instruments undermine it (westerhoff-dignaga-dharmakirti, pp. 261–263).
This contrast — Mīmāṃsā epistemic optimism (trust in untrained awareness) vs. Buddhist epistemic pessimism (untrained awareness is shot through with ignorance) — marks one of the deepest fault lines between the two traditions (westerhoff-dignaga-dharmakirti, pp. 267–268).
Etymology: Pramāṇa as “Measuring”
The Sanskrit word pramāṇa literally means “the measuring of something.” Valid cognition, then, is something that has the cognitive measurement of something as its use (don-byed nus-pa, Skt. savyāpāra). This etymology is explicit in Śaṅkarasvāmin’s Nyāyapraveśa (berzin-science-of-mind, Commentary on “A Manual for Engaging in Logic”):
“The apprehending (yang-dag-pa’i shes-pa, Skt. adhigama) of something is both what the valid cognizing of it does as well as the result of the valid cognizing.”
Because valid cognition’s essential nature is the apprehension of its object, what it does and its result are not different — valid cognition both does the apprehending and is the result of that apprehending.
Two Pairs: For Oneself and For Others
The Nyāyapraveśa opens with a fundamental division of valid cognition (berzin-science-of-mind, Commentary on “A Manual for Engaging in Logic”):
- For bringing understanding to oneself: perception (Berzin: “bare cognition”; མངོན་སུམ་) and inference (རྗེས་དཔག་)
- For bringing understanding to others: logical proof (sgrub-pa, Skt. sādhana) and refutation (sun-‘byin, Skt. dūṣaṇa)
Together with their semblances (ltar-snang, Skt. ābhāsa):
- A semblance of valid cognition appears valid to the unknowledgeable but is flawed
- A semblance of direct perception (Berzin: “semblance of bare perception”): conceptual cognition of an external object — it lacks the individually characterised phenomenon as its object of experience
- A semblance of inference (Berzin: “semblance of inferential cognition”): cognition that arose from a semblance (fallacious) reason
The “Valid Cognition” Division from the Point of View of the Name
The Geluk bsdus grwa tradition gives a three-fold division of things called by the term “valid cognition” (Perdue: “valid cognizer”, ཚད་མ་) (perdue-buddhist-reasoning-debate, p. 317; cf. PV Ch. 2):
- Persons such as the Buddha (called “valid persons” / “valid teachers”)
- Scriptures such as the teaching of the four noble truths (called “valid scriptures”)
- Minds — direct perception (Perdue: “direct perceivers”, མངོན་སུམ་) and inference (Perdue: “inferential cognizers”, རྗེས་དཔག་) — the actual valid cognitions
This is explicitly classed as a “division from the point of view of the name” (མིང་གི་སྒོ་ནས་དབྱེ་བ་) — not a true partition. Only consciousnesses are actual valid cognitions; persons and scriptures are merely called by that name in dependence on the validity of the minds that uttered or composed them. The division has soteriological force for buddha-as-pramana and for the status of scriptural-authority — Perdue notes that even reading the Buddha’s most definitive statement “is not a valid cognition until readers realize it on their own” (p. 315). Buddhists rely on scripture, but the reliance is mediated through the receiving consciousness, which alone can be a valid cognition.
The argument that exactly two pramāṇas exhaust valid cognition is recorded by Perdue (p. 314) as derived from Dharmakīrti’s PV Ch. 3: “Because objects of comprehension are two, valid cognizers are two” (Perdue’s wording; our preferred rendering: “…valid cognitions are two”). Objects of comprehension partition exhaustively into specifically- and generally-characterised phenomena (see specifically-characterised, generally-characterised); a consciousness suited to the first is direct perception, to the second is inference. Other Indian proposals — knowledge by context, by analogy (upamāna), by verbal authority (śabda) — are absorbed into inference or, in the case of textual authority, reclassified as investigative thought (Perdue: “correctly assuming consciousness”, ཡིད་དཔྱོད་) “because it is induced by a mere term and does not realize its object” (Perdue p. 314, citing Geluk Collected Topics).
Sources
- dreyfus-recognizing-reality — Chapter 16 (defining pramāṇa), Chapter 17 (pragmatism), Chapter 18 (validity of inference)
- westerhoff-dignaga-dharmakirti — the Buddha as pramāṇabhūta, the Mīmāṃsā challenge on intrinsic authoritativeness (pp. 261–263)
- sonam-introduction-treasury — Sapan’s definition (non-deceptive in relation to action, agent, object), three classes of non-valid cognition, rejection of explicit/implicit realisation
- berzin-science-of-mind — etymology of pramāṇa as “measuring”; two pairs (for oneself vs. for others); semblances of valid cognition
- gorampa-pramanavartika — three-feature gloss of “nondeceptive” (object/agent/action); wholly designated valid cognition vs. innate characteristic; ascertainment from self/from others as operating at the designation level
- perdue-buddhist-reasoning-debate — Kensur Yeshi Thupten’s “valid cognition is the measure of what exists and what does not exist” (p. 60); Sautrāntika Geluk bsdus grwa presentation of the two pramāṇas (pp. 313–317); “division from the point of view of the name” (person, scripture, mind); reclassification of śabda as investigative thought (Perdue: “correctly assuming consciousness”, ཡིད་དཔྱོད་)