Pramana Wiki

Pramāṇasamuccaya

23 min read

Pramāṇasamuccaya (ཚད་མ་ཀུན་བཏུས་)

Compendium of Valid Cognition — the foundational treatise of the Buddhist logico-epistemological tradition, by dignaga.

Overview

This text established the systematic framework for Buddhist epistemology that dharmakirti would later elaborate and defend. It covers the entire range of Buddhist logic and epistemology in six chapters.

Opening Verse

The famous verse of homage addresses the Buddha as pramāṇabhūta (“embodying the epistemic instruments”) — not authoritative by enlightened status but because his enlightenment flows from the correct application of epistemic instruments (westerhoff-dignaga-dharmakirti, p. 221):

Having bowed down to Him, who embodies the epistemic instruments, who seeks the benefit of the world, the teacher, the well-gone, the protector, I here make a single compendium of my various scattered [writings], to establish epistemic instruments.

dharmakirti later unpacked these epithets into a formal argument for the Buddha’s reliability in Ch.2 of the pramanavartika (westerhoff-dignaga-dharmakirti, p. 240).

Integration of Two Traditions

As Katsura explains, the Pramāṇasamuccaya is the first systematic integration of two previously separate traditions of Indian logic (katsura-dignaga-lectures):

  • The debate tradition (vāda) — from Brahmanical brahmodya debates to the Points of Defeat (nigrahasthāna) of the Carakasaṃhitā and Nyāyasūtra, and Buddhist vāda manuals by Nāgārjuna and Vasubandhu
  • The epistemology tradition (pramāṇa) — from the Carakasaṃhitā’s five pramāṇas through the various orthodox schools

Chapters I, II, and V belong to the pramāṇa tradition; Chapters III, IV, and VI to the vāda tradition. Dignāga integrated them through the terms svārthānumāna (inference for oneself) and parārthānumāna (inference for others), showing via the trairūpya theory that inference and proof are essentially the same — the former a mental process, the latter its linguistic expression (katsura-dignaga-lectures).

Chapter IV (Example) should properly be included in Chapter III (Proof), since Dignāga’s proof consists of three propositions — Thesis (pakṣa), Reason (hetu), and Example (dṛṣṭānta) — and Chapter III only deals with the first two (katsura-dignaga-lectures).

Chapter VI examines Fallacious Objections (jāti), rejecting reductio ad absurdum (prasaṅga) as proof while admitting it as a means of refutation. Points of Defeat receive no special treatment from Dignāga; it is dharmakirti who treats them fully in his Vādanyāya (katsura-dignaga-lectures).

The text chapter by chapter

The Sanskrit is lost and Yijing’s Chinese translation with it, so there is no easy way into the treatise itself. The walk-through below follows Fazun’s Abridged Exposition (集量論略解), the modern Chinese translation of the root verses and auto-commentary made from the Tibetan with a running gloss; each chapter’s full summary is a page of its own, linked at the head of its section. Tibetan verses are quoted from the Tengyur text as set out in the working translation files under translations/pramanasamuccaya/ — Fazun worked from that same Tibetan, using the Derge translation collated against the Peking one.

Every chapter divides into establishment of Dignāga’s own position and refutation of rivals: the Vādavidhi (a logic manual Dignāga opens by denying is Vasubandhu’s own work), the Nyāya, the Vaiśeṣika, the Sāṃkhya and the Mīmāṃsā. The refutations are usually the longer half, and they are why the treatise matters as a doxographical source — it preserves opponents’ formulations that survive nowhere else. Fazun finished the work in his eightieth year with failing eyesight, says in his preface that he had no oral transmission on the treatise and few commentaries to hand, and offers it as a draft. Where he leaves a reading open, the passage has been cross-checked against Khenchen Kunga Wangchuk’s annotation commentary on the treatise (ཚད་མ་ཀུན་བཏུས་ཀྱི་མཆན་འགྲེལ་ཚིག་དོན་གསལ་བྱེད་), and this page gives the reading that survives the check.

Chapter 1: Perception (མངོན་སུམ་, pratyakṣa)

Full summary: chapter-1-perception.

The chapter argues from the object, not from the knower. There are two valid cognitions because the knowable has two aspects and no third: the specifically-characterised phenomenon (རང་མཚན་, svalakṣaṇa), which perception takes, and the generally-characterised phenomenon (སྤྱི་མཚན་, sāmānyalakṣaṇa), which inference takes. Scripture and analogy are valid cognitions in name only. The same argument disposes of two candidate extras: a cognition that combines perception with a judgement (“this form is impermanent”) reaches no object the two do not already reach, and re-cognition of what one has already cognised would multiply valid cognitions without end, “as with memory and the like”. Novelty is the criterion.

མིང་དང་རུས་སོགས་སུ་སྦྱོར་བའི། །རྟོག་པ་དང་བྲལ་མངོན་སུམ་མོ། །

That which is free from conceptual construction — the application of names, class and so forth — is perception.

The implication runs one way only. All perception is non-conceptual; not every non-conceptual cognition is perception, since a cataract-sufferer’s cognition of a double moon is non-conceptual and still pseudo-perception (མངོན་སུམ་ལྟར་སྣང་). Fazun sorts the excluded constructions by the kind of word fused to the bare object: arbitrary names, class-words, quality-words, action-words, and words naming a property-bearer. Perception is fourfold — sense, mental, self-aware, yogic — and sense cognition is named after the faculty rather than the object because the faculty is the uncommon cause (ཐུན་མོང་མིན་པའི་རྒྱུ་): the patch of blue can produce another person’s eye-consciousness and my own mental cognition, but my eye produces only my eye-consciousness. Pseudo-perception is sevenfold: the mirage, the “jar” and “cloth” of conventional cognition, inference and post-inferential cognition, memory, expectation, and the cataract case, the first six conceptual and the last not.

The heaviest passage is the result of valid cognition (ཚད་མའི་འབྲས་བུ་, pramāṇa-phala), where Dignāga gives three successive accounts, each refining the last. The result is not a fourth thing standing outside the act, any more than a tree’s being cut stands outside the felling. First the cognition is the instrument and its confirming function the result; then self-awareness (རང་རིག་) is the result, the object-image the measured and the object-mode the instrument; finally the object-aspect, the grasping aspect and the knowing are one entity posited three ways, which is the source of the later threefold division into object-aspect, subject-aspect and self-awareness-aspect. Two arguments carry it: one remembers not only a past object but one’s cognition of it, and what was never experienced cannot be remembered; and if a cognition were experienced by a further cognition, that one would need a third, with no term ever reached. The move that makes self-awareness universal without collapsing thought into perception is stated flatly — conceptual cognition is self-aware perception with respect to its own nature and conceptual with respect to its object, and the two do not compete because they concern different things.

The five refutations are uneven in interest. Against the Nyāya definition — cognition born of faculty–object contact, free of verbal designation, non-erroneous, and determinate in nature — Dignāga shows the first two qualifiers idle (no sense cognition designates verbally, and “non-erroneous” says only that a cognition takes its own object), the fourth fatal (determination is the result, so a perception that determines leaves no result to produce), and the contact clause false to the facts, since mountains are seen at a distance and the eye takes objects larger than itself. He then presses the Naiyāyika on pleasure and pain, which arise from no faculty–object contact and rest on no reason: neither perception nor inference on their account, so the prameya they grasp would be no object of knowledge at all.

Chapter 2: Inference for Oneself (རང་དོན་རྗེས་དཔག་, svārthānumāna)

Full summary: chapter-2-inference-for-oneself.

རྗེས་དཔག་རྣམ་གཉིས་རང་དོན་ནི། །ཚུལ་གསུམ་རྟགས་ལས་དོན་མཐོང་བའོ། །

Inference is of two kinds. Inference for oneself is seeing the object from a reason that possesses the three modes.

Inference divides in two where perception does not, and again the reason is the object: the specifically characterised cannot be displayed, so it admits no division into a form kept to oneself and a form communicated. The chapter’s governing claim is that inference never reaches the particular. Even the apparent counter-case, inferring the tangible from the visible, moves from the universal of the colour to the universal of the tangible on the strength of a previously observed co-presence; the tangible as actually touched is never made manifest. The Vaiśeṣika presses with wind — from tangibility one infers a substance not visible to sight — and the reply states a principle the rest of the chapter runs on: the narrower may prove the wider, never the wider the narrower. Touch inheres in many kinds of substance, so it cannot single out wind. What settles wind is a trustworthy statement (ཡིད་ཆེས་ཀྱི་ལུང་), and Dignāga admits such statements as inference not because they denote their object but because acting on them is found non-deceiving. The same restriction blocks the Sāṃkhya inference to pradhāna.

On the members, the valid reason is present in the subject and in the similar class (མཐུན་ཕྱོགས་) and unconditionally absent in the dissimilar class (མི་མཐུན་ཕྱོགས་), where the dissimilar class is only what lacks the property to be proved, not the contrary of it and not everything else. Fazun sets out six fallacious reasons against the standard subject, sound, three lacking two modes and three lacking one — “sound is permanent, because it is produced” has the property of the position and nothing else; “because it is audible” is present in the subject and in the dissimilar class alike.

Two disputes carry the weight. First, what an inference proves. Not the bare property, since smoke would then prove a fire nobody doubted; not the connection, since “where there is smoke there is fire” would have to be cast in the genitive — “the fire-connection of this place” — which is not what anyone says. What is proved is the property-bearer as qualified by the property: not “fire”, but “fire here”. Second, how the reason reaches it. By exclusion, not by positive display: fire is obtained only as turned back from non-fire. The blue lotus supplies the ladder — being a quality excludes non-substance, being a smell excludes the non-earthen, the fragrance excludes bad smell, and the particular fragrance excludes everything but the lotus. Were the reason to display positively as perception does, one would know either nothing or everything, since one has not seen every fire. The chapter closes by separating the Buddhist connection from Vaiśeṣika inherence (samavāya): the connection is asymmetric (no smoke without fire, but fire without smoke), and it is posited nominally on the inseparability of the two rather than being a real entity alongside them. Pervasion by itself is never what does the inferring — “having horns” pervades cow, yet nobody infers cow from horns.

Chapter 3: Inference for Others (གཞན་དོན་རྗེས་དཔག་, parārthānumāna)

Full summary: chapter-3-inference-for-others.

གཞན་དོན་རྗེས་སུ་དཔག་པ་ནི། །རང་གིས་མཐོང་དོན་རབ་གསལ་བྱེད། །

Inference for the sake of others clearly reveals to another the object that one has oneself seen.

The statement is called an inference because the cause takes its effect’s name: words are not a cognition, they produce one. Three qualifications fix the thesis. It states the proponent’s own nature, meaning the property still to be established rather than one already agreed; it is asserted as one wishes, relative to one’s own system and selectively even there; and it is un-negated by perception, by inference, by one’s own scripture, or by common renown — so “sound is inaudible”, “a pot is permanent”, and, for the fourth, “the hare-bearer is not the moon”, which fails on nothing but settled usage. The self-refuting case is “all verbal meanings are false”: include the utterance and both the speech-nature and the falsity go at once; exempt it and it shows other speech to be not-false.

The chapter’s core is the nine-fold analysis of the property of the position (ཕྱོགས་ཆོས་), the hetucakra: three states in the similar class — present, absent, both — crossed with the same three in the dissimilar. Two are valid (the middle of the first and last triads), two contradictory (the middle of the second), and the remaining five inconclusive. Fazun lists the standard formulae for each cell, all against sound. Before that, four ways the property of the position fails outright: unestablished for both parties (“because grasped by the eye”), unestablished for one (urging “because produced” on a theorist for whom sound is manifested rather than produced), unestablished through doubt (arguing for fire where one is unsure the haze is smoke), and unestablished as to its locus (proving a pervading self).

Two distinctions repay attention. The uncommon reason (ཐུན་མོང་མིན་པ་) — “sound is permanent, because it is audible” — is doubtful precisely because nothing but sound has that nature, so the reason is excluded from everything else one-sidedly and settles nothing. And Fazun separates two things Tibetan and Chinese sources often blur: contradiction-determination (相違決定, viruddhāvyabhicārin) takes two different reasons establishing opposite theses and establishes neither, so it is a doubt-reason; the property-difference contradictory reason uses one reason and definitely establishes the opposite of what was wanted. Underlying all of it, the reason works “through the gateway of exclusion, not affirmation”.

A consequence (ཐལ་འགྱུར་, prasaṅga) is not a proof. Formulae such as “sound is impermanent, because otherwise action and the rest would be permanent” run through the opponent’s own gateway and expose a fault; they have no property of the position of their own. On the same ground Dignāga rejects the Sāṃkhya’s avīta as a separate kind of inference: apart from the correspondence-reason there is no other, and if what survives a round of one-by-one refutations counted as proof, the opponents’ answers being endless, everything would collapse into refutation. On one line the Chinese and the Tibetan run in opposite directions, and the Tibetan annotation commentary settles it in favour of the Tibetan: a reason whose property of the position is unestablished — eye-grasped-ness for the impermanence of sound, pleasure and the rest for a pervading self — is unfit even as a refutation, not merely unfit as a proof. Only a reason established as the property of the position for both parties does both work, refuting and proving; the reasons that fail through doubt become usable if some further argument establishes them.

Chapter 4: Example and Pseudo-example (དཔེ་, dṛṣṭānta)

Full summary: chapter-4-examples.

The shortest of the six in Fazun’s exposition, and the one he is least satisfied with. Its work is structural: the reason-formula states only the property of the position — “because sound is produced” — so the other two modes have to be supplied, and the example is what supplies them.

བསྒྲུབ་བྱའི་རྗེས་སུ་འགྲོ་བའི་རྟགས། །བསྒྲུབ་བྱ་མེད་ན་མེད་ཉིད་གང་། །

The reason’s following the probandum, and its very absence where the probandum is absent — these two are to be fully shown in the examples.

The similar example runs “whatever is produced by effort is impermanent, like a pot”; the dissimilar, “whatever is permanent is not produced by effort, like space”. The negations differ in kind, and the difference does cross-school work: the similar example negates while affirming something in its place (implicative negation, མ་ཡིན་དགག་), the dissimilar merely denies (non-implicative, མེད་དགག་), so a Sautrāntika who refuses space the status of a substance can still cite “like space”, nothing being asserted to exist. Order matters too. Inverting the similar form into “all that is non-produced is permanent” proves permanence by non-production, which nobody set out to prove; inverting the dissimilar makes lightning come out both permanent and produced by effort.

Both examples are needed, and Dignāga states why by naming what goes wrong when one is dropped: state only the similar and the reason may still occur in the dissimilar class, giving the common-inconclusive fallacy; state only the dissimilar and the reason may be absent from the probandum’s own class, giving the uncommon; state neither and the reason operates in the dissimilar class alone, which is contradiction. Even both examples leave two of the nine reasons in error: the last of the first triad, “conch-sound is produced by effort, because it is impermanent”, and the last of the last, “sound is permanent, because it is not tangible”. What each example does is show one mode directly and the other by implication — the similar example the positive concomitance directly and the negative pervasion by implication, the dissimilar the reverse — so both are still stated, unless one of them already produces the understanding on its own. Against the Naiyāyika five-member proof, only three members survive — reason, example, thesis — the application and the conclusion being dropped, the conclusion because it merely restates.

The pseudo-examples are set out against the frame “sound is permanent, because it is not tangible”: lacking the reason (like an atom), lacking the probandum (like a mental action), lacking both (like a pot). Three more, run against the converse thesis, fail to establish the negative pervasion; two invert the concomitance; and one has no positive concomitance at all — “whoever speaks is passionate, like the man on the highway”. Eighteen faults of the example are taught elsewhere. The worldly logicians’ dissimilar formula, “like space, what is permanent is not produced”, falls short of all of this: it shows only that permanence and non-production are found together, never the relation of the reason to what it proves.

Fazun’s closing note on this chapter is unusually blunt, and it is why the section above was checked line by line against the Tibetan: the exposition is overly brief, diverges considerably from the other Tibetan translation, and should be collated against a Tibetan commentary should one be obtained. That collation has now been made, and the points he left open — the ordering of the two modes, the force of the two negations, and which reasons still err with both examples stated — resolve as given here.

Chapter 5: Exclusion of Other (གཞན་སེལ་, apoha)

Full summary: chapter-5-apoha.

The longest stretch of Fazun’s exposition, and one argument throughout: verbal cognition is not a third valid cognition. The Mīmāṃsakas and Naiyāyikas make the uttered word a means of knowing in its own right. Dignāga subsumes it under inference. This is the chapter behind the wiki’s apoha page and behind everything Tibetan authors say about the problem-of-universals.

སྒྲ་ལས་བྱུང་བ་རྗེས་དཔག་ལས། །ཚད་མ་གཞན་མིན་འདི་ལྟ་སྟེ། །རང་གི་དོན་ལ་བྱས་སོགས་བཞིན། །གཞན་སེལ་སྒོ་ནས་སྟོན་པར་བྱེད། །

Verbal cognition is not a valid cognition other than inference. It is like this: just as with “produced” and the like applied to its own object, it reveals its meaning through the gateway of exclusion of other.

A word reaches its object by the same “not arising without it” relation as a logical reason, and by ruling out what it is not. The realist alternative is dismantled in stages. A kind-word cannot express the particulars, on two counts: they are limitless, so no word can be connected with each of them, and “existent” spreads indifferently across substance, quality and activity, which yields doubt rather than determination. Nor can the word express the possessor-of-the-kind (tadvat), since that theory needs a metaphorical transfer of cognition from universal to particular, and no such transfer occurs — calling a servant “king” does not produce the cognition of a king, and the cognitive images simply differ.

What the theory has to explain, and does, is co-reference: how “blue” and “lotus” pick out one thing and stand as qualifier and qualificand. Each word excludes a different range, and because the exclusions converge on one substratum they form a common locus; the substratum is neither blue alone nor lotus alone. Fazun’s illustration is the upright shape and the bird perching on it, two marks neither of which decides the matter alone. The chapter then works out positive and negative concomitance for words on the ladder tree → earth-born → substance → existent: the particular determines the universal and never the reverse, so a birch is necessarily a tree but a tree need not be a birch, and it is the negative side that makes the inference go through. Words exclude each other because each seeks mastery over its own range, “like princes vying for one throne”, and one word cuts off the particulars of unrelated universals as one rejoices at the defeat of a friend’s enemy.

The realist is then put to a dilemma about the compound: are the member and the possessor-of-members, with their substratum, distinct or not? Take them as non-distinct and the blue-meaning and the lotus-meaning collapse into each other, since within one meaning there is no other — or else the aggregate, not being distinct from the many, comes out many. Take them as distinct and there is no common locus either, since a separate aggregate-ness would have to operate on the aggregate-dharmas or they on it, and neither is found. Two smaller objections get short answers. If a word has no real referent, how did anyone ever have a first cognition of one? There is no first: the causal series of conventional designation runs back beginninglessly. And what of names imposed at will, which no convention explains? An arbitrary name is a loosely stated universal, and states its object by the differentiation of that universal like any other word — between kind-words and aggregate-words, Dignāga says, there is not the slightest difference.

Two moves close the system. Where a term’s application is settled by usage, Dignāga refuses to look behind it, citing the Bhagavān: “one should not be attached to the names of beings, nor should one discard worldly names” — the world applies “colour” to blue and not to flavour, and that is what determines the exclusion. And conceptual cognition, though conceptual towards its object, is self-aware, hence perception with respect to itself; the point from chapter 1 returns to keep verbal and conceptual cognition from becoming a further valid cognition. Analogy (upamāna) goes the same way: learn from another that the wild ox resembles the tawny ox and it is verbal cognition, work it out oneself and the similarity is inferred. Neither is a separate instrument.

Chapter 6: Sham Rebuttals (ལྟག་ཆོད་, jāti)

Full summary: chapter-6-defeat-conditions.

The chapter treats the jāti, which Fazun glosses as 似破, the pseudo-refutation: an opponent who cannot find a fault in a sound proof charges it with a merely apparent parity. Every item runs against one formula, “sound is impermanent, because it is produced immediately after effort”, so the chapter reads as a catalogue of ways to answer badly.

ཚད་སྣང་བརྗོད་པར་འདོད་རྣམས་ལ། །རིགས་པའི་ཆ་ཤས་སྨྲ་པོ་བཞིན། །ཕྱོགས་གྲུབ་མཚུངས་པ་མ་ཡིན་ཞེས། །བསྟན་ཕྱིར་མ་ཚང་ལ་སོགས་བརྗོད། །

To those who wish to declare their pseudo-pramāṇa a genuine proof, Dignāga states the charges of “deficiency” and the rest in order to show that their thesis is not equally established.

The specimens include the conjunction/non-conjunction parity (if the reason reaches the probandum it is no longer distinct from it, and if it does not reach it, it is on a par with every non-reason) and its three-times variant, on which a reason prior to the thesis has nothing yet to be the reason of, a posterior one comes too late, and a simultaneous one stands to the thesis as the left horn of an ox to the right. Then come the permanence parity (impermanence, being a nature the thing does not abandon, makes sound permanent); the not-yet-stated and not-yet-arisen parities; the effect-parity; the homology and heterology parities, both run through space and the pot; the differentiation parity (the pot is combustible and visible, sound is neither, so sound alone should be permanent); the apprehensible parity, which points out that lightning’s impermanence is known without the reason; and the doubt parity.

The replies fall into a few shapes, and the general one is the strongest: these charges would dismiss every reason that has an example whatever, so they prove too much, and the opponent’s own thesis falls to the objection he raises — the fault of self-injury (自害). Where the parity turns on a supposed production of the thesis by the reason, it misses the relation involved, which is knower and known rather than producer and produced. Where it turns on some particular feature of the example, the answer is that inference proceeds on the general property and not the particular; were each object determined separately, no inference could occur at all. Dignāga also rejects the Vādavidhi’s classification of these faults into inversion, unreality and contradiction, on the ground that the three cannot be kept apart — the same sham rebuttal falls under all of them, so the division does not hold.

Notable for what it omits: Points of Defeat (nigrahasthāna), the standing furniture of the debate manuals, receive no separate treatment here; it is dharmakirti who takes them up in the Vādanyāya (katsura-dignaga-lectures). The chapter ends with three verses turning on conceptual construction as the root of the opponents’ error, a warning against reinterpreting the Tathāgata’s dharma-characteristics, and the dedication of the whole treatise.

Key Innovations

Tibetan Reception

Tibetan thinkers primarily understood Dignāga’s ideas through dharmakirti’s interpretations. The Pramāṇasamuccaya was translated into Tibetan but received less direct commentarial attention than the pramanavartika (dreyfus-recognizing-reality, p. 1).

Dreyfus argues that the Sakya reading of apoha — particularly in gorampa’s and sakya-chok-den’s signifier/signified analysis — ends up closer to Dignāga than to Dharmakīrti, which is surprising given that Sa-paṇ’s stated method is to read Dharmakīrti literally. Dignāga and Sa-paṇ share a polemical situation (refuting realism — Hindu for the former, Tibetan for the latter), so both treat language as primarily an internal articulation of the conceptual realm and minimise the causal-experiential bridge to reality. Dharmakīrti, on the defensive against Hindu charges of arbitrariness, inserts the causal element. Sakya thinkers, locked in their own polemic against Cha-ba, recover Dignāga’s emphasis (probably unbeknownst to themselves) (dreyfus-recognizing-reality, pp. 274–75).

Dreyfus also gives the most extensive English treatment of Candrakīrti’s critique of Dignāga in Prasannapadā (Ch. 1), which targets Dignāga directly rather than Dharmakīrti: the strict criteria of validity, the rejection of pramāṇasamplava, and the limitation of perception to bare individuals all “do violence to common usage” and exemplify the foundationalist project that Madhyamaka must dismantle (pp. 451–56).

Sources

  • index — Fazun’s Abridged Exposition (集量論略解, CBETA 大藏經補編 vol. 9, no. 40), the source of the chapter walk-through above; six chapter summaries, one page each. It carries the parts of the argument the secondary literature reports rather than reproduces: the sevenfold division of pseudo-perception, the three accounts of the result of valid cognition, and the refutations of the Vādavidhi and the Brahmanical schools that fill the second half of every chapter
  • Khenchen Kunga Wangchuk’s annotation commentary ཚད་མ་ཀུན་བཏུས་ཀྱི་མཆན་འགྲེལ་ཚིག་དོན་གསལ་བྱེད་, covering the Pramāṇasamuccaya together with the Nyāyabindu and the Pramāṇaviniścaya, held under raw/commentaries-tb/. Used to settle the readings Fazun leaves open in chapters 1, 3 and 4
  • dreyfus-recognizing-reality — Ch. 11 introduces Dignāga’s apoha; Ch. 15 argues Sakya readings of language are closer to Dignāga than Dharmakīrti; Conclusion (pp. 451–56) presents Candrakīrti’s critique of Dignāga’s epistemology in detail
  • westerhoff-dignaga-dharmakirti — extensive treatment of epistemology, inference, and apoha
  • katsura-dignaga-lectures — detailed account of the integration of vāda and pramāṇa traditions, chapter structure, perception theory, and self-awareness

Graph View

Backlinks