Introduction to Sakya Pandita's Treasury of Valid Cognition

Provenance: based on an essay by Sonam Jamtsho (Dzongsar), first published March 2018; substantially revised, corrected and rewritten by Tenpa Bhikshu (Nicolas Pettican), 2026. Judgements below are the revised essay’s unless attributed otherwise.

Terminology note: where the essay uses a rendering our raw/glossary.tsv lists as a variant, the preferred term is used here and the essay’s wording is parenthesised at first occurrence. The recurring substitutions are: investigative thought (Sönam: “correctly assuming consciousness”, ཡིད་དཔྱོད་) · appearance-without-ascertainment (Sönam: “inattentive cognition”, སྣང་ལ་མ་ངེས་པ་) · wrong cognition (Sönam: “distorted cognition”, ལོག་ཤེས་) · conceptual cognition (Sönam: “conceptual thought”, རྟོག་པ་) · self-awareness (Sönam: “self-cognising”, རང་རིག་) · conceptual generality (Sönam: “meaning universal”, དོན་སྤྱི་) · apprehended object (Sönam: “held object”, གཟུང་ཡུལ་).

Main argument

An introduction to sakya-pandita’s pramanayuktanidhi (Treasury of Valid Cognition, ཚད་མ་རིགས་གཏེར་) organised around the Treasury’s own division: ignorance is mistaken perception, so removing it requires knowing the object (ཡུལ་), the cognition that comprehends it (ཡུལ་ཅན་), and the relation between them (ཡུལ་ཡུལ་ཅན་གྱི་འབྲེལ་བ་). Chapters 1–7 ascertain the object, chapters 8–11 the cognition. The essay takes that division from Sa-paṇ rather than imposing it, and uses it to place each of the eleven chapters.

The controlling claim is about what Sa-paṇ was arguing against. On the essay’s reading his target is not error in general but one specific error: an antirealist philosophy read as a moderate realism. Dharmakīrti demotes universals to constructs with no ontological weight; Ngok Lotsawa and cha-ba, while claiming to follow the Indian pramāṇa system, accepted a moderate form of the reification that Dharmakīrti’s Nyāya and Sāṃkhya opponents held outright. Sa-paṇ, with Sanskrit at his command and Śākyaśrībhadra’s circle to check his readings against, concluded that the texts had been “polluted by the innovations of earlier Tibetans” and that the mistake was the norm, not the exception. Hence the Treasury’s two purposes: positive, to propagate a correct reading of Dharmakīrti; negative, to refute Ngok’s and Cha-ba’s.

The essay is candid that this leaves a hard problem in place rather than solving it. If universals are merely conceptual, and inference reaches reality only through universals, how is inference valid at all? The essay calls this “the greatest predicament within the Dharmakirtian tradition” and suggests it is why the earlier Tibetans moderated Dharmakīrti’s rejection in the first place — a repair it judges indefensible as a reading of Dignāga and Dharmakīrti, without claiming Sa-paṇ has a better answer to the underlying difficulty.

Chapter breakdown

The Indian Intellectual Milieu; Dignāga and Dharmakīrti — sets the schools Dharmakīrti argues with (Sāṃkhya’s unchanging conscious Self, Vaiśeṣika atomism, theistic creation, Cārvāka denial of rebirth and karma) and reads his project as the search for an inter-traditional standard of validation: one that proves Buddhist doctrine without presupposing Buddhist metaphysics, using only principles the opponents already grant.

Transmission to Tibet — Ngok Lotsawa Loden Sherab (1059–1109) as the first native Tibetan to compose original works on Indian logic, with the suggestion (attributed to “many scholars”, unnamed) that non-Buddhist tenets entered his reading through his Kashmiri teacher Kalden Gyalpo; then cha-ba (1109–1169) as the most innovative of Ngok’s successors and the founder of the monastic debate tradition.

Sakya Pandita: Life and Significance; Sa-paṇ’s Critique — Sa-paṇ (1182–1251) as the first ruler of a unified Tibet after 842 CE under Mongol suzerainty, and the reason pramanavartika rather than any other of the seven treatises became the primary text of Tibetan logico-epistemological study. His method is the point: he quotes Dharmakīrti directly instead of working through the commentarial traditions other Tibetans preferred.

Structure and Scope — eleven chapters of unequal length with preamble and colophon; root text in verse, terse and mnemonic; an auto-commentary whose authenticity “has been questioned by some”, recorded and not pursued.

Chapter 1, Objects of Comprehension (ཡུལ་བརྟག་པ་) — objects defined as possible objects of valid cognition (TR 1.1); two objects, two criteria of existence (apprehensibility by valid cognition, and causal capacity), which do not coincide, since generally characterised phenomena satisfy the first and fail the second. Sa-paṇ’s target is the earlier Tibetan division of the apprehended object (Sönam: “held object”, གཟུང་ཡུལ་) into three types, extending it to conceptual cognition and to mistaken non-conceptual cognition. Two counter-arguments: those cognitions would then be non-mistaken, contradicting Dignāga and Dharmakīrti outright; and their objects would be available to any bystander, which they plainly are not. Then his own claim, TR 1.11, and the hermeneutical frame of three worldviews, with the pramāṇa masters’ final position identified as Cittamātra.

Chapter 2, Consciousness (བློ་བརྟག་པ་) — cognition as clarity and knowing (TR 2.1); non-valid cognition as that in which non-deceptiveness is not established (TR 2.18). Sa-paṇ attacks the first three of the earlier Tibetans’ five non-valid types and keeps the last two, doubt and wrong cognition, which reappear in his own threefold list. (The essay glosses “the last two” as perception and inference, which does not fit its own five-item list; taken as doubt and wrong cognition it matches what follows. [unverified] which was meant.) Investigative thought does not exist at all: if it depends on valid reasoning it is inference, and if it does not it falls under non-realisation, wrong cognition or doubt. Appearance-without-ascertainment and subsequent cognition go the same way, both resting on the assumption that valid cognition determines its object. Three classes remain: non-realisation, doubt, wrong cognition.

Chapters 3–7, How Consciousness Apprehends Its Objects — the essay groups these as a block resting on the two ontological types and their two epistemological counterparts. Ch. 3 (universals and particulars): the rejection of real universals as a consequence of universal flux and the critique of wholes, against the Naiyāyikas; perception as “complete engager” (སྒྲུབ་འཇུག་), conceptual cognition as mistaking the objective aspect (གཟུང་རྣམ་) for the world; Gorampa’s three universals, and Gyaltsab’s warning that denying reality to duration will undermine karma and its fruits. Ch. 4 (apoha): language and conceptual cognition engage negatively, by exclusion; perception neither affirms nor negates but holds its object and induces the conceptual cognition that determines it. Ch. 5 (signified and signifier): naming is convention, and a mistake all unenlightened beings share; it survives only from the practical standpoint (འཇུག་པའི་ཚེ་), not the theoretical (འཆད་པའི་ཚེ་). Ch. 6 (relations): identity and causality are the only two admitted, and Sa-paṇ refutes both — identity because a relation needs two terms, causality because cause and effect are never simultaneous. Ch. 7 (contradiction): the earlier Tibetans’ definition would make a cause contradict its effect; two types are admitted, exclusion of not being together and mutual exclusion.

Chapter 8, Definition and Valid Cognition (མཚན་ཉིད་བརྟག་པ་) — whether definition itself has a definition, answered with Cha-ba’s solution (the name has one, the meaning-distinguisher need not); the three fallacies of definition; definition as “a meaning-distinguisher established by the exclusion of direct contradiction” (auto-commentary on TR 9.15). Dharmakīrti’s two formulations of valid cognition (PV 2.1, PV 2.5) are read by Sa-paṇ and Gorampa as one definition stated twice, and the earlier Tibetans’ explicit/implicit realisation distinction is refuted: affirmation and negation are gaining certainty, not ontological operations.

Chapter 9, Valid Perception (མངོན་སུམ་བརྟག་པ་) — the longest chapter and the one where the essay makes its own judgements. Sa-paṇ rejects the distinction between perception and valid perception: attributing cognitive content to perception makes it constructive and commits one to real universals, where Dharmakīrti’s perception is passive. Definition: non-mistaken cognition free from conception; four types (sense, mental, self-awareness, yogic). The yogic-perception section carries the argument for the possibility of enlightenment, which needs the continuum of mind to be stable — hence the rebirth arguments, and the claim that nothing but a preceding instant of consciousness can be the substantial cause (ཉེར་ལེན་) of consciousness. The essay flags as surprising Sa-paṇ’s further claim that the sense organs likewise carry over as substantial causes of future sense organs.

Chapters 10–11, Inference — correct reason as endowed with the three modes (ཚུལ་གསུམ་), in three types (same nature, effect, non-cognition), the first two affirming and the third negating. Ch. 11 defines the three parties to formal debate, criticises the Naiyāyika five-limb proof, and admits four responses where the Tibetans before Sa-paṇ had three: contradictory, uncertain, unestablished, and acceptance.

Key claims

1. The Treasury’s architecture follows from a soteriological premise: ignorance is mistaken perception, so object, cognition and their relation must each be analysed (Structure and Scope).
2. Sa-paṇ’s charge against Ngok and cha-ba is specific — moderate realism about universals — not general incompetence (Sa-paṇ’s Critique).
3. Existence has two criteria in Dharmakīrti that do not pick out the same things: apprehensibility by valid cognition, and causal capacity. Generally characterised phenomena satisfy the first only, and Dharmakīrti still lists them among objects (Ch. 1).
4. Only external material objects of sense perception count as apprehended objects for the Indian masters; the earlier Tibetan extension to conceptual and mistaken cognition is a misreading (Ch. 1).
5. Objects cast their aspects (རྣམ་པ་གཏད་པ་) towards perception, so object and subject stand in a causal and therefore sequential relation. Cha-ba’s coexistence of object and subject is closer to common intuition and is denied by Dharmakīrti (Ch. 1).
6. “Specifically characterised phenomena alone are the objects of cognition” (TR 1.11). Bodong Panchen called Sa-paṇ an ontological nihilist for it; neither Gyaltsab nor Gorampa reads it literally, Gyaltsab adding “ultimately” and Gorampa distinguishing objects of the mode of apprehension from objects of engagement and invoking conventional valid cognition (ཀུན་ཏུ་ཐ་སྙད་པའི་ཚད་མ་).
7. Sa-paṇ identifies the final position of the Indian pramāṇa masters as Cittamātra, arguing it by the sign of clarity (གསལ་རིག་གི་རྟགས་) and the sign of simultaneous apprehension (ལྷན་ཅིག་དམིགས་ངེས་ཀྱི་རྟགས་) (Ch. 1).
8. Investigative thought (Sönam: “correctly assuming consciousness”, ཡིད་དཔྱོད་) does not exist: it collapses into inference on one horn and into non-realisation, wrong cognition or doubt on the other (Ch. 2).
9. No perception determines its object; it holds the object in the perceptual field and induces the conceptual cognition that determines it. Subsequent cognition is therefore always conceptual and always invalid (Chs. 2 and 4).
10. Sa-paṇ’s own list of non-valid cognitions is threefold: non-realisation, doubt, wrong cognition. Non-realisation is absent from his opponents’ fivefold list (Ch. 2).
11. Both relations Buddhists admit are refuted: identity needs two terms to relate, and cause and effect never co-exist (Ch. 6).
12. Dharmakīrti’s two formulations of validity, non-deceptiveness (PV 2.1) and revealing the unknown (PV 2.5), are one definition stated twice — Sa-paṇ and Gorampa against the readings that make them two (Ch. 8).
13. The earlier Tibetans defined a correct reason as one that is the three modes; Sa-paṇ and Gorampa as one endowed with them (Ch. 10, n. 42). The essay calls the distinction subtle and important without spelling out what turns on it.
14. The disagreement over which doxographical school Dharmakīrti belongs to (Madhyamaka or Cittamātra) points to different philosophical strata inside his writings rather than to a single answer (Ch. 1).

Notable quotes

• “Polluted by the innovations of earlier Tibetans.” — Sa-paṇ’s verdict on the pramāṇa material he inherited, quoted in Sa-paṇ’s Critique.
• “Regarding valid cognition, even those who claim to be correct expositors are seen to be in error.” — Sa-paṇ, auto-commentary preface, quoted at n. 9.
• ཡུལ་གྱི་མཚན་ཉིད་བློས་རིག་བྱ། — “The definition of an object is: that which is known by the mind.” (TR 1.1; rendering from chapter-1-objects)
• གཞལ་བྱ་རང་མཚན་གཅིག་ཁོ་ན། — “The object of valid cognition (prameya) is the single specifically characterised phenomenon (svalakṣaṇa) alone.” (TR 1.11, same file — the verse the essay calls Sa-paṇ’s most startling claim)
• བློ་ཡི་མཚན་ཉིད་རིག་པ་ཡིན། — “The defining characteristic of mind (buddhi) is awareness.” (TR 2.1; rendering from chapter-2-cognition)
• མི་བསླུ་མ་གྲུབ་ཚད་མ་མིན། — “That in which non-deceptiveness is not established is non-valid cognition.” (TR 2.18, same file)
• “A meaning-distinguisher established by the exclusion of direct contradiction.” — Sa-paṇ’s definition of definition, auto-commentary on TR 9.15.
• “A superimposition which appears as this object through the elimination of its direct opposites.” — Gorampa’s definition of exclusion, Ch. 4.
• “When theoretically considering, it is wise to make the distinction. When engaging [practically], it is achieved through mistaken [cognition].” — TR 5.14, quoted at n. 24.
• “It remains the finest gateway into the mind of Dharmakīrti.” — the essay’s closing verdict.

Relevance to the three main texts

pramanayuktanidhi

The whole essay is about this text, and it is the most useful short orientation in the vault: eleven chapters placed, the polemical target named, and the loci given for the four root verses that recur in every commentary (TR 1.1, 1.11, 2.1, 2.18). Its distinctive contribution is on the chapters other summaries skip. Chapter 6’s two refutations of relation, chapter 7’s two types of contradiction, and chapter 11’s four admissible responses (against the earlier Tibetans’ three) are each stated with enough structure to check against the root text. What it does not give is the argument at verse level; for that the auto-commentary and gorampa-pramanavartika are needed.

pramanavartika

Engaged as the text Sa-paṇ promoted rather than as an object of commentary. The essay’s claim here is historical and strong: the centrality of the Commentary in Tibetan logico-epistemological study is largely Sa-paṇ’s doing, and he is very likely the source of all the Tibetan pramāṇa lineages of it now extant. Two verse-level contacts matter for translation work — PV 2.1 and PV 2.5 as the two formulations of valid cognition (Ch. 8), and the second chapter of the Commentary as the model for Sa-paṇ’s rebirth arguments and for “naturally luminous” (Ch. 9, nn. 38–41), where the essay notes that Sa-paṇ does not explain how the term relates to the Tathāgatagarbha tradition.

pramanasamuccaya

Present only as a source Sa-paṇ draws on: chapter 1 of the Treasury is said to rest principally on the theory of objects in the Compendium and the Commentary, and chapter 2 on their inference-for-oneself chapters together with the Nyāyabindu. Dignāga’s own arguments are not examined.

Critical assessment

The essay is a map, not an argument. It reports Sa-paṇ’s positions and the Sakya commentarial consensus without adjudicating between Gyaltsab’s and Gorampa’s non-literal readings of TR 1.11, both of which it presents as legitimate. Three places do carry judgement: the “surprising” sense-organ claim in chapter 9, the note that Sa-paṇ leaves “naturally luminous” unconnected to the Tathāgatagarbha material, and the admission that the auto-commentary’s authenticity has been questioned — the last recorded without follow-up, though the dispute is documented in dreyfus-recognizing-reality (pp. 26–27).

Two attributions are left vague. “Many scholars believe” non-Buddhist tenets reached Ngok through Kalden Gyalpo, and “some” question the auto-commentary; neither is named, and the second is answerable from Dreyfus. The claim at n. 42 — that “is the three modes” versus “endowed with the three modes” is a subtle but important distinction — is asserted and not cashed out, and it is exactly the sort of point a reader would want the essay to finish. [unverified] as to what Sa-paṇ takes to follow from it.

Coverage follows the essay’s weighting rather than the Treasury’s: chapters 1, 2 and 9 run to several pages, chapters 5 to 7 to a paragraph each.

Translation-relevant material — FLAG

No sustained translation. The essay quotes short root verses with English paraphrases at TR 1.1, 1.11, 2.1, 2.18, TR 5.14 and the auto-commentary on TR 9.15, plus PV 2.1 and PV 2.5 — useful as a check on wording, not as a base text.

Do not lift Tibetan from the source file. Its docx conversion has mangled subjoined letters throughout (བལོ for བློ, སདེ for སྡེ, གྟེར for གཏེར), so every string needs checking against a clean witness. The verses quoted above were taken from translations/pramanayuktanidhi/, with the renderings already settled there. One spelling variant to be aware of at TR 2.18: our translation file and the essay read མི་བསླུ་, while raw/extracted/ཚད་མ་རིགས་གཏེར་རང་འགྲེལ།.md reads མི་སླུ་. Both are attested; nothing turns on it here, but a collation of the chapter should record which witness is being followed. [unverified]

Scholars mentioned

dignaga, dharmakirti, sakya-pandita, gorampa, sakya-chok-den, gyel-tsap, cha-ba, ngok-lotsawa, Bodong Panchen, Śākyaśrībhadra, Kalden Gyalpo, Candrakīrti, Mipham, Go Lotsawa, Khenpo Sonam Gyatsho.

Bibliography (as given in the essay)

Tibetan sources

• Sakya Paṇḍita, Treasury of Valid Cognition: Root Text and Auto-Commentary (ཚད་མ་རིགས་གཏེར་རྩ་འགྲེལ་)
• Gorampa, Clarification of the Seven Treatises (སྡེ་བདུན་གསལ་བྱེད་)
• Gorampa, Explanation of the Treasury (ཚད་མ་རིགས་གཏེར་གྱི་རྣམ་བཤད་)
• Gorampa, Rays of Samantabhadra (ཀུན་བཟང་འོད་ཟེར་)
• Gorampa, Sun of Samantabhadra (ཀུན་བཟང་ཉི་མ་)
• Śākya Chokden, The Defeater (དགག་ལན་)
• Gyaltsab Dharma Rinchen, Commentary on the Treasury
• Go Lotsawa, Blue Annals (དེབ་ཐེར་སྔོན་པོ་)

Western sources

• Dreyfus 1997, Recognising Reality — see dreyfus-recognizing-reality
• Dunne 2004, Foundations of Dharmakīrti’s Philosophy