Pramana Wiki

Chittamatra

35 min read

Chittamatra (སེམས་ཙམ་)

Chittamatra, Yogācāra, cognition-only

Yogācāra and Chittamatra are two names for one school, and what separates them is reference, not doctrine. Yogācāra names the philosophical and meditative tradition; Chittamatra describes its specific ontological view, that there is mind only. The etymologies say as much. Yogācāra means the practice of yoga, so the school is named after the practitioners it grew out of, the yogins; cittamātra means mere mind, so it is named after what it holds. Vijñaptimātra, cognition-only, and Vijñānavāda, the doctrine of consciousness, do the same work with other words (Wikipedia). No author surveyed here treats the names as marking two positions to be told apart, and a writer who switches between them is varying the emphasis rather than the school.

Sakya Paṇḍita’s text shows the switching happening. The root verses of the Pramāṇayuktanidhi never say སེམས་ཙམ་: they say cognisance (རྣམ་རིག་, vijñapti) and call the school’s adherents the proponents of cognisance (རྣམ་རིག་པ་). His Auto-commentary then uses all three names on one folio of its chapter 9. It heads the section རྣམ་རིག་པའི་ལུགས་, “the system of the proponents of cognisance”; states the opponent’s objection in the opponent’s terms, that on such a view there can be no valid cognition and result for the effect-reason of the Cittamātrins (སེམས་ཙམ་); quotes the verse about the trace of fire; and closes by declaring that the inference of the Yogācāras (རྣལ་འབྱོར་སྤྱོད་པ་) is established (Auto-commentary on chapter 9, p. 403 of the Tibetan). One position, three names, fifteen lines.

The commentators inherit the habit, each keeping to a favourite. Khenpo Sodargye’s lectures say Cittamātra; Gorampa’s translator says Yogācāra and, at one point, Vijñaptika; Dreyfus says Yogācāra and Mind-Only. Sakya Paṇḍita’s own auto-commentary pairs the school with the Sautrāntika as the two ways Dharmakīrti settled his system: the system of the object knower (དོན་རིག་) and the system of cognisance (auto-commentary on Pramāṇayuktanidhi chapter 1). Alexander Berzin states the Geluk textbook equation flatly: “Other names for Chittamatra are Yogachara (rNal-‘byor spyod-pa) and Vijnanavada” (studybuddhism.com, a Geluk-oriented presentation). What modern scholars dispute, Wikipedia reports, is not the names but whether the position they name is an idealism at all.

One author does drive a wedge between them, and it is a doctrinal wedge rather than a quibble about words. Sakya Chokden holds the school’s final view to be the False Aspectarian one, awareness with no duality of subject and object left in it, not even the internal duality of a real aspect, and he describes that view “as Yogācāra but not as Mind-Only” (dreyfus-recognizing-reality, p. 433). Yogācāra is then the wider term, and Mind-Only names only the True Aspectarian wing, which still keeps an internal object for the mind to have. Nobody else here follows him. Gorampa, describing the same view, calls it “the best of Mind-Only”; Khedrup, in Dreyfus’s words, “does not make the distinction between Yogācāra and Mind-Only” at all (nn. 14–15, p. 557). The disagreement is over how much idealism a single name may cover, and it is the only place in these sources where the choice of name decides anything.

This page writes Chittamatra, and the translations of the root texts write Cittamātra; quotations keep whatever their author wrote.

Where the three texts stand

The three texts do not refute this school; they climb to it. In Pramāṇayuktanidhi chapter 1 sakya-pandita determines the Sautrāntika (མདོ་སྡེ་པ་) position by refuting the Vaibhāṣika (བྱེ་བྲག་ཏུ་སྨྲ་བ་) and the early Tibetans, and then determines the Yogācāra position by refuting the Sautrāntika in turn. In pramanavartika chapter 3 dharmakirti gives three successive accounts of valid cognition (ཚད་མ་) and its result, and Gedun Drub marks the second as “the Yogācāra school’s distinctive establishment”, the first having been “the Sautrāntika-Yogācāra common establishment” (Gedun Drub’s commentary on PV III). The pramanasamuccaya is counted among the eleven foundational treatises of the Yogācāra school by the Chinese edition the commentary here follows (Fazun’s exposition of pramanasamuccaya I, publisher’s preface).

What makes the page difficult is that the ascent is almost never announced. Dharmakīrti argues on the assumption that external objects exist through nearly all of his work and withdraws that assumption in a handful of passages; whether those passages state his considered view, and what the earlier assumption then amounts to, is the question every commentator surveyed here is answering, and they do not agree.

Overview

The movement has been traced to the first centuries of the common era, and appears to have grown out of yogis of the Sarvāstivāda and Sautrāntika traditions in north India who adopted the Mahāyāna (Wikipedia). Buddhist tradition attributes it to Maitreya-nātha; the systematisers are Asaṅga and Vasubandhu, both placed in the fourth to fifth century. Its root scripture is the Saṃdhinirmocanasūtra; among its treatises are Asaṅga’s Mahāyānasaṃgraha and Vasubandhu’s Viṃśatikā and Triṃśikā. Śāntarakṣita carried a synthesis of it with Madhyamaka to Tibet in the eighth century, and Xuanzang carried it to China in the seventh. It has no separate institutional life today: its ideas and texts remain subjects of study in Tibetan and East Asian Buddhism (Wikipedia).

Three doctrines from the wider school stand behind the arguments these texts use, though only the first is argued for in the material here. First, the eight bodies of consciousness: the five sense consciousnesses, mentation, the defiled self-consciousness, and the foundation consciousness (ཀུན་གཞི་, ālayavijñāna), which holds the karmic seeds until they ripen. Second, the three natures: the imagined, the dependent, and the perfected. Third, a division within the school between True Aspectarians (རྣམ་པར་བདེན་པར་སྨྲ་བ་, satyākāravādin) and False Aspectarians (རྣམ་པར་རྫུན་པ་སྨྲ་བ་, alīkākāravādin) over whether the aspect (རྣམ་པ་, ākāra) through which an object is represented in cognition is itself real. That last division is doxographically later than Dignāga and Dharmakīrti, and the commentaries here apply it to them anyway; the strain this produces is one of the things this page has to report rather than resolve.

Berzin supplies the standard terms for the second of those, and an argument for the school’s central thesis that these texts never use. The three natures are the imagined or totally conceptional (parikalpita), the dependent (paratantra), and the perfected or thoroughly established (pariniṣpanna); the dependent is the basis on which the other two are imputed, and all the dependent phenomena in one moment of cognition share “the same natal source”. That phrase carries the argument. An external phenomenon is defined as one arising “from different natal sources (rdzas) than the valid cognitions of them”; a visual form and the consciousness that sees it in fact arise together from a single karmic tendency, so that “it is not that first the sight exists and then the consciousness sees it”; and nothing then answers to the definition (studybuddhism.com, a Geluk-oriented presentation). Sakya Paṇḍita gets to the same conclusion by two other routes entirely, the destruction of the partless atom and the co-apprehension reason, and the karmic trace enters his chapter 1 only afterwards, to explain how a true cognition still differs from a false one.

Whether the school is an idealism at all is disputed in modern scholarship: Wikipedia sets Garfield and Hopkins, who read it as idealist, against Anacker, Lusthaus and Wayman, who hold it definitely is not. The material here does not enter that argument. Sodargye’s lectures and Gorampa’s commentary both take the position to be that no external object exists and that what appears is cognition, and the reasonings they reconstruct are arguments for exactly that.

Core doctrines

That there is no external object

Sakya Paṇḍita’s determination of the position occupies the closing movement of Pramāṇayuktanidhi chapter 1: nine verse-blocks as the translation here sets them out, running from the refutation of the atom to the summary verse. They divide into a demolition and a construction.

The demolition takes down the partless particle (རྡུལ་ཕྲན་, paramāṇu), which is what both realist Buddhist schools rest on:

།གཅིག་དང་དུ་མ་བྲལ་བའི་ཕྱིར། །དོན་མེད་དེས་ན་རྣམ་པའང་མེད།

Being neither one nor many, [external] objects do not exist, and so neither do [their] aspects. (Pramāṇayuktanidhi chapter 1)

The argument is Vasubandhu’s six-atoms case, which Sodargye states as a dilemma about the atom in the middle of a cluster. Ask whether the central atom has spatial parts. If it does, it is not partless. If it does not, then no accumulation of such atoms can ever produce spatial extension: aggregate as many as you like and they collapse to a single point, Mount Meru included. Since “many” is an accumulation of “ones”, the failure of the one takes the many with it, and with the object goes the aspect the object was supposed to cast on cognition (sodargye-pramanayuktanidhi, on chapter 1). Sodargye is careful about the scope: this is the reasoning of being free from unity and multiplicity (ekānekaviyogahetu), but deployed here from the Yogācāra standpoint against partless atoms, not as the Madhyamaka reasoning that establishes ultimate emptiness, and it is not turned on the mind the Yogācāra accepts.

The construction is two proofs in one verse, and this is the densest half-page in the Pramāṇayuktanidhi material:

།གསལ་དང་ལྷན་ཅིག་དམིགས་པས་འགྲུབ། །ལྷན་ཅིག་དམིགས་ལ་གཞན་མི་སྲིད། །རྫས་གཞན་ཡིན་ན་རྟགས་མི་སྲིད། །ཟླ་གཉིས་སྣང་བཞིན་སྔོ་འཛིན་ཡང༌། །གཞན་ལ་མཐའ་གཞན་བཀག་པས་ཁེགས།

[That all knowables are mind] is established by clarity and by co-apprehension. For what is co-apprehended, being other is impossible; were they different substances, the reason would be impossible. As with the appearance of two moons, so too with blue and its apprehender. [The reason] is excluded from what is other, since the alternatives are negated. (Pramāṇayuktanidhi chapter 1)

The first reason is clarity (གསལ་བ་). Its formal statement: the knowable external object is one entity with the cognition apprehending it, because it is a manifestation of clear awareness and bears the defining characteristic of cognition; as fire and its heat are one entity. Two steps get it there, and they are not equally strong. The first is that the only way anyone ever establishes that an object exists is that it appears before cognition, which a realist can grant, since it is a claim about access, not about the object. The second is that whatever appears before my cognition is not separable from the clarity that is my cognition, and this is where a realist will stop, because it is close to the conclusion. What follows the second step is only a rule about definitions: whatever bears an entity’s defining characteristic is classified as that entity, dewlap and hump making a cow and clear awareness making a cognition (sodargye-pramanayuktanidhi, on chapter 1).

The second reason is co-apprehension (ལྷན་ཅིག་དམིགས་པ་, sahopalambha): the seen object and the seeing cognition are not different entities, because they are invariably apprehended together, as with a double moon. The positive concomitance is that whatever is seen is one entity with the cognition seeing it, the double moon being the instance: the second moon is inseparable from the cognition that sees it and never occurs without it. The negative concomitance is where the verse’s last line does its work. Anything genuinely other than cognition ought to be findable at some time when its cognition is absent; nothing of the sort has ever been found; so the supposed “other” meets an invalidating reason (གནོད་པ་ཅན་, bādhakahetu), since nothing is both co-apprehended and genuinely other (sodargye-pramanayuktanidhi, on chapter 1).

The auto-commentary sets out the alternatives the last line rules out. Things that differ in time are apprehended at different times, so they are not co-apprehended. A superimposition at the same moment is known to be a superimposition, so its object is nothing but the cognition. And things that are different substances at the same moment are excluded by an invalidating reason: what is unrelated is not invariably apprehended together, like blue and yellow (translation notes to Pramāṇayuktanidhi chapter 1). Gorampa adds that the co-apprehension argument is aimed at opponents who hold that cognition has no aspect (Gorampa on Pramāṇayuktanidhi chapter 1).

The weight of the second reason rests on that third case, and it was attacked from inside Sakya Paṇḍita’s own lineage. His teacher Tsurtön Zhönnu Sengé judged both reasons invalid. The clarity reason is either unestablished or inconclusive: examined, it is not established that the object is the knower, and being known or being mere awareness does not settle the matter. The co-apprehension reason fails because nothing excludes simultaneous, distinct substances that are invariably apprehended together; the supposed invalidating reason does not reach them (Tsurtön Zhönnu Sengé, Lamp of Wisdom, p. 8). Sakya Paṇḍita answers these objections in the auto-commentary, together with those of the Indian critic Subhagupta, who had attacked the co-apprehension argument in his Proof of External Objects (auto-commentary on Pramāṇayuktanidhi chapter 1). Our material does not give his answers in detail, so how far the reasoning is meant to carry (proof, or a stalemate that favours the idealist) is not settled here.

The thesis then gets its flat statement:

།དོན་དུ་སྣང་བ་བློ་ཉིད་ཡིན། །སྣང་བ་དེ་ཉིད་ཕྱི་ན་མེད།

What appears as an object is mind itself; that very appearance does not exist outside. (Pramāṇayuktanidhi chapter 1)

True and false, with nothing outside to check against

The obvious objection to the thesis is that it abolishes the difference between a valid cognition and a mistaken one. If neither the white conch nor the yellow conch is out there, if one moon and two moons are equally not in the sky, what is left to make one of each pair correct? Sodargye reports the objection as pressed against both wings of the school, True Aspectarian and False Aspectarian alike (sodargye-pramanayuktanidhi, on chapter 1).

The reply is one line, and it relocates the criterion:

།བག་ཆགས་བརྟན་དང་མི་བརྟན་ལས། །བདེན་དང་རྫུན་པའི་རྣམ་གཞག་བྱེད།

True and false are presented on the basis of stable and unstable traces. (Pramāṇayuktanidhi chapter 1)

Real and illusory alike are manifestations of traces (བག་ཆགས་) on the foundation consciousness; what separates them is whether the trace is stable. Thousands of people with the stable trace see the conch white, and go on seeing it white; jaundice deposits an unstable trace and the conch looks yellow for a while. Sodargye pushes the same point across the boundary between rebirths: our stable human trace makes water drinkable, and when it is exhausted and the hungry-ghost trace ripens the same river runs with pus and blood (sodargye-pramanayuktanidhi, on chapter 1). Stability and duration, not correspondence, are what “true” now means. The view was not new in Tibet: Cha-ba had already reported the False Aspectarians as calling cognitions with stable traces (one moon) unmistaken and those with unstable traces (two moons) mistaken (Cha-ba, Clearing of Mental Obscuration, fol. 2a).

The same machinery answers a harder objection about inference, which is treated at the end of Pramāṇayuktanidhi chapter 9. If only what appears to cognition exists, inference should be impossible: the fire behind the mountain either exists while unseen, in which case something exists that does not appear and mind-only fails, or it does not exist and there is nothing to infer; and if it comes into existence only once the smoke has appeared, then the cause follows its effect. The reply again runs on traces:

།མེ་སྣང་བག་ཆགས་སྙིང་པོ་ལས། །དུ་བར་སྣང་བའི་བག་ཆགས་སྐྱེ། །དེས་ན་སྐྱེས་བུ་མི་སླུ་བ། །རྗེས་དཔག་ཡིན་ཞེས་མཁས་རྣམས་བཞེད།

[For the Cittamātrins (proponents of cognition-only, རྣམ་རིག་པ):] from the essence of the trace of appearing-fire, the trace appearing as smoke arises; therefore, being non-deceptive to persons, the wise hold it to be inference. (Pramāṇayuktanidhi chapter 9)

Sodargye’s illustration is a dream: there is no fire in the external object at all, yet seeing smoke rise on the dreamt mountain one infers fire, runs towards it, and finds fire, the fire found being the particular fire within the dream. And he notes that the difficulty was never peculiar to this school. On the Sautrāntika account too, by the time the smoke appears the fire that produced it has ceased, so that inference also rests on a trace of the producer-produced relation carried in one’s own continuum rather than on any present cause (sodargye-pramanayuktanidhi, on chapter 9).

Two perceptions, or four

The Pramāṇayuktanidhi sorts the Buddhist schools by how many kinds of perception (མངོན་སུམ་) each can accept, and the Chittamatra has the shortest list:

།མངོན་སུམ་བཞི་པོ་མདོ་སྡེའི་ལུགས། །བྱེ་བྲག་སྨྲ་གསུམ་རྣམ་རིག་གཉིས།

The four perceptions [accord with] the Sautrāntika system; the Vaibhāṣikas [accept] three, the Cittamātrins two. (Pramāṇayuktanidhi chapter 9)

Without a real external object there is nothing for sense perception and mental perception to take, so only self-aware perception (རང་རིག་) and yogic perception remain. Sodargye immediately qualifies the count, and the qualification matters more than the count. The two-perception figure is the school’s ultimate view. Provisionally the school can accept all four, and on his reading it must: Dignāga’s Ālambanaparīkṣā already teaches the four perceptions, and Dignāga is reckoned a master of the Chittamatra following reasoning; Asaṅga expounds the form aggregate, and whoever accepts the form aggregate accepts the form faculties (sodargye-pramanayuktanidhi, on chapter 9). The doxographical label and the working position come apart here in a way the roster of “two, three, four” conceals.

The sense faculty

One further position comes from Sodargye rather than from the root text, and it splits the school internally.

On what the sense faculty (དབང་པོ་) is: the Pramāṇayuktanidhi’s chapter 9 is set on Sautrāntika ground, where the faculty is an insentient capacity, and the Auto-commentary says so. On the Chittamatra view the faculty is posited upon cognition instead. The masters following reasoning, Dignāga and Dharmakīrti, take it to be an uncommon capacity to illuminate form residing on the preceding moment of eye consciousness. The masters following scripture (Asaṅga, Vasubandhu, and apparently Jñānaśrī) take it to be a particular latent trace of apprehending objects residing on the foundation consciousness (sodargye-pramanayuktanidhi, on chapter 9).

True and False Aspectarians

Take the blue of the co-apprehension argument above. Both wings agree there is no blue outside; they divide over the blue that is left inside. For the True Aspectarians (རྣམ་པར་བདེན་པར་སྨྲ་བ་, satyākāravādin) that blue is real: it is consciousness itself in the shape of an object, and whatever is consciousness is real. For the False Aspectarians (རྣམ་པར་རྫུན་པ་སྨྲ་བ་, alīkākāravādin) it is not: the mind’s own nature is clear and carries no blue, and the blue lies over it falsely, as the hairs a diseased eye sees lie over empty air. Stated technically, the question is whether the aspect (རྣམ་པ་, ākāra) through which an object is represented in cognition is itself real, and the two answers give the two wings their names.

So both wings hold ordinary perception mistaken, and they differ over how far the mistake goes. The True Aspectarian confines it to externality: the objective aspect a mental state holds is substantially identical with consciousness, so what perception gets wrong is where the blue is, not the blue. The False Aspectarian carries the mistake further in, and makes the representation itself a deluded construct, superimposed on a nature that is “ineffable, utterly beyond the duality of subject and object” (dreyfus-recognizing-reality, p. 433). Dreyfus’s compressed version: “True Aspectarians understand this object as consciousness ‘impersonating’ external objects, False Aspectarians maintain this appearance does not rest on anything real” (p. 434).

The entailment between the wings and the tenet systems runs one way only, which is what makes the labels useful. A Sautrāntika has to be a True Aspectarian: if external objects exist and perception is what validates them, then perception is undistorted, and so are the aspects it hands on to conceptual thought. A True Aspectarian need not be a Sautrāntika, since one can hold aspects real while denying that anything outside answers to them, which is exactly the True Aspectarian Chittamatra position (p. 433).

One of the names is a trap. False Aspectarians are also called proponents of no-aspect (རྣམ་པ་མེད་པར་སྨྲ་བ་, nirākāravāda), and Sakya Paṇḍita uses the short form rnam med of them. In the wider Indian debate that same name belongs to the direct realists, Nyāya-Vaiśeṣika, Vaibhāṣika, Mīmāṃsaka and Jain, who deny aspects for the opposite reason: consciousness is amorphous, like a light that reveals an object without itself changing, so perception needs no intermediary at all (dreyfus-recognizing-reality, p. 335). The True and False Aspectarian quarrel is internal to the other camp, the sākāravāda, which holds that cognition apprehends its object not nakedly but through an aspect. Dreyfus flags the ambiguity: “since the view of False Aspectarians differs from direct realism, the former should be clearly differentiated from the latter” (n. 13, p. 557).

The criterion, and a Geluk dissent

Most Tibetan authors draw the line where Dreyfus does, over the reality of the aspect; he calls this well established in the later Indian and Tibetan sources and notes that even Khedrup and Tsongkhapa assume it (p. 433). Gyeltsap and most later Geluk writers reject it, on a short argument: False Aspectarians are Yogācārin, so they must grant that consciousness is real even in its deluded states, and cannot then deny that aspects are real; the dispute must therefore be about something else. Gyeltsap makes it a dispute about commonsense objects, over whether the appearance of blue as extended and covering its many parts is “not distorted by ignorance” (ma rig pa’i bslad pa ma zhugs) (p. 434, with n. 17, p. 557). Dreyfus is unpersuaded and says so: this way of distinguishing the two “is hard to understand. It excludes from the debate any ontological issue. But what then is the debate about?” (p. 434).

Berzin, as a Geluk source, works with Gyeltsap’s criterion rather than the ontological one. True aspectarians there “assert that whatever appears grossly to valid sensory cognition is established as existing as how it appears”; false aspectarians deny it and take a commonsense object to be “just a conceptual interpolation on sensibilia”; Geluk sides with the first and the non-Geluk Tibetan traditions with the second (studybuddhism.com). That criterion still decides something substantial: on the True Aspectarian Chittamatra view “the appearance of colored shapes pervading areas of pixels (particles) and the appearance of commonsense objects exist as they appear” (studybuddhism.com).

The sub-wings, and where the commentators stand

Both wings subdivide, and one of the subdivisions is endorsed in the material here. Khedrup, as Dreyfus reports him, splits the True Aspectarians three ways by how many aspects one cognition can take. The Proponents of Perceptual Duality, literally “the proponents of the equal number of held [objects] and holding [subject]”, hold that several cognitions are needed to apprehend several aspects. The Proponents of Perceptual Non-Duality hold that several aspects can appear to one cognition. The Half-Eggists, in between, hold that only a single aspect can appear to a single consciousness (n. 24, p. 558). Sakya Chokden names only the first two and prefers the second, since for him the aspect is a unity and the diversity of external objects is cognitively irrelevant however important it is causally (p. 435). Mipham Rinpoche takes the first, and Sodargye follows him: a fabric with a hundred differently coloured parts is answered by a hundred apprehending parts of the eye cognition, and that exact correspondence is what makes True-Aspectarian Mind-Only the right account of conventional truth (sodargye-pramanayuktanidhi, on chapter 4). The False Aspectarians divide in turn into those with stains (dri bcas) and those without (dri med), over whether any false conventional appearance is left in a Buddha’s continuum (n. 19, p. 558).

The two wings also give different schemes of valid cognition and its result, and Sodargye rejects one of them. For the True Aspectarians the self-aware part that makes the apprehended object (གཟུང་ཡུལ་) appear is the object of comprehension, the apprehending part is the valid cognition, and true self-awareness is the result, all three simultaneous and of one essence, as the experiencer and the experienced in a felt pleasure are one entity. For the False Aspectarians the superimposed part that makes the object appear is the object of comprehension (outer mountains and rivers being neither one’s own mind nor truly existent, appearing as hairs in the sky do), and what results is only an imputed self-awareness, since they posit it in dependence on an object they hold insubstantial. Sodargye judges this untenable for the business of establishing conventions, and settles with the True Aspectarians, citing the Commentary on the Madhyamakālaṃkāra (sodargye-pramanayuktanidhi, on chapter 9).

That scheme is the top of a four-rung ladder, and Sakya Paṇḍita states all four rungs in one passage of the Auto-commentary. On the Vaibhāṣika account the object, the valid cognition and the result are substantially distinct and simultaneous; on the Sautrāntika account substantially distinct and successive, the object having already ceased; on the True Aspectarian account they are of one substance and simultaneous; and on the False Aspectarian account the object of comprehension and the result are held to be precluded from being one, as with the appearance of hairs. He then rules on Dharmakīrti: of these four systems, the Master proceeds as a Sautrāntika when he accepts external objects, and accepts rnam med, the False Aspectarian position, when the knowable is drawn inward (auto-commentary on Pramāṇayuktanidhi chapter 9). The second half of that ruling is the closing verse of chapter 1, restated in prose.

Which wing was Dharmakīrti’s?

Disputed all the way down, and the disagreement is old. Devendrabuddhi and his disciple Śākyabuddhi read him as a True Aspectarian and Khedrup follows them, granting that the evidence conflicts. Dharmottara and Prajñākaragupta read him as a False Aspectarian and Gyeltsap follows them, asking how, if every sense perception an ordinary being has is mistaken, one could mark off a part polluted by ignorance from a part that is not. Sakya Paṇḍita and Gorampa reach the False Aspectarian conclusion too, though they disagree with Gyeltsap about how the two wings are to be told apart: Dharmakīrti “is a Sautrāntika when he accepts the existence of external objects and a False Aspectarian when he denies it”, and since the second view is the higher one, it is his (dreyfus-recognizing-reality, p. 434, citing the Auto-commentary 151.b.5). Chodrak Gyatso also makes him a False Aspectarian, but on a reading of Yogācāra as a Madhyamaka philosophy of the tathāgatagarbha tradition, which is a long way from Sakya Paṇḍita’s (pp. 434–435). Dreyfus’s own answer, given in the Dharmakīrti section below, is that this was not a question Dharmakīrti cared to settle.

Two things here are left standing against each other. The two reasonings of Pramāṇayuktanidhi chapter 1 are worded differently: the clarity reason has a positive predicate (all knowables are established as mind), the co-apprehension reason a negative one (blue and its cognition are not different) (auto-commentary on Pramāṇayuktanidhi chapter 1). Whether that difference marks a shift between the wings, the first leaning True Aspectarian and the second False Aspectarian, is a question neither the root text nor Sodargye raises. And Sodargye himself rejects the False Aspectarian scheme as unworkable for establishing conventions and settles on the True Aspectarian wing, while Sakya Paṇḍita, on the auto-commentary’s ruling and Dreyfus’s reading, takes Dharmakīrti’s considered view to be the False Aspectarian one. The two claims answer different questions and are not formally in conflict, but nothing surveyed here sets them side by side. [unverified] whether the Auto-commentary anywhere states a preference of Sakya Paṇḍita’s own between the wings; the folio Dreyfus cites for the ruling on Dharmakīrti (151.b.5) has not been checked against the Tibetan here.

Where the epistemologists stand

Dignāga

The pramanasamuccaya material never argues for mind-only in so many words. What it does is build the apparatus the argument later runs on, and in one place take a step that only makes sense if the external object has already been let go.

The apparatus is the analysis of the result of valid cognition, given in three successive accounts at the end of chapter 1. The third is the one Fazun’s exposition singles out as Dignāga’s advance over earlier Buddhist epistemology:

གང་ཚེ་སྣང་བ་དེ་གཞལ་བྱ། །ཚད་མ་དང་ནི་དེའི་འབྲས་བུ། །འཛིན་དང་རྣམ་རིག་དེ་ཡི་ཕྱིར། །དེ་གསུམ་ཐ་དད་དུ་མ་བྱས། །

When the appearance is the object to be known (prameya), the valid cognition and its result are the apprehension and the cognition [respectively]—therefore these three are not established as distinct. (pramanasamuccaya I)

Object-aspect (གཟུང་རྣམ་), subject-aspect (འཛིན་རྣམ་) and self-awareness-aspect are one entity, posited as three only from different points of view (Fazun’s exposition of pramanasamuccaya I). Nothing in the verse says there is no external object. But the second of the three accounts already says something stronger than a realist can comfortably say: since the object is not apart from cognition, the cognition determines the object-sense by its own essence, and one and the same appearance is called both the object measured and the instrument of measuring (same source). The step is small and it is not flagged.

The step that is not small is in the Ālambanaparīkṣā, which is not among the texts covered here; Dreyfus reconstructs its argument. Realism about external objects rests on the intuition that perception is accurate, so that objects exist as we perceive them, extended and unitary. No individual atom carries the aspect of extension. If the atoms are a mere collection (saṃghāta), like a forest, the collection is a conceptual overlay and does not exist substantially, so it cannot cause a cognition of extension. If they are an aggregate (saṃcita) with a collective structure of its own, that structure is attributable to the cognition and not to the object, on pain of readmitting the Nyāya whole standing over its parts. Both accounts therefore fail, and it is inconsistent to hold both that external objects exist and that we perceive them correctly (dreyfus-recognizing-reality, pp. 101–102).

Dharmakīrti

The pramanavartika chapter that carries the idealism is chapter 3, which the translations here do not yet cover. What follows is reconstructed from Gedun Drub’s commentary, which numbers no verses, and from Dreyfus, so this section quotes no root verses.

Gedun Drub’s chapter 3 works the argument as a sequence of refusals. Asked what “cognising the object” amounts to, Dharmakīrti rejects the answer that it is the cognition’s resembling the object: the atoms cannot make the coarse aspect appear, so the coarse aspect is not a real object; and if resembling-and-arising-from were sufficient for experiencing, the later moment of a blue cognition would experience the earlier moment of blue cognition, which it does not. Asked what is the fault in saying the cognition experiences an external object, he answers that there is none, but that the reason for saying so has never been given: the aspect might come from the object or from elsewhere, and examined, the external object is not established. The manifested blue is the body of seeing blue, and cannot be apprehended apart from the blue-cognising consciousness. Faced with the objection that without an external object the occasional arising of a cognition becomes inexplicable, he answers that a previous consciousness awakens a later habituation, which does the work the object was doing (Gedun Drub’s commentary on PV III). Grasper and grasped are then set up “only at the level of appearance”, their apparent difference traced to a defective mind, as in the cognition that manifests hair-floaters; and self-awareness is the result “since cognising the object has no cognising of any other thing—because there is no external object” (same source).

Dreyfus’s reading of why this material is there, and why there is so little of it, is the most developed account we have, and it is worth having in his terms. Dharmakīrti’s problem is that perception is supposed to be an undistorted reflection of reality while extension exists in no atom; his solution is to root out the assumption that generated the problem, namely that objects exist externally as a result of atomic aggregation (dreyfus-recognizing-reality, p. 102). This produces two flatly conflicting metaphysical positions inside one work, and Dreyfus refuses the comfortable escape: “we have to accept that Dharmakīrti does introduce conflicting views within a single work” (p. 103). Several Tibetan sources read the conflict as a hierarchy (external objects accepted provisionally, refuted at a deeper level), which is what Dreyfus calls the strategy of ascending scales of analysis, lower positions introduced as needed and yielding to higher ones as analysis is pushed (p. 104). He takes the hierarchy to be genuine rather than merely convenient, on the strength of the verse that closes one of the few Yogācāra passages, about the Buddha acting “in a spirit of strict [agreement with] the world” while setting the meaning of ultimate reality aside (p. 104). The scarcity of such passages he explains by audience: Dharmakīrti argues before a mostly non-Buddhist Indian philosophical public, to whom Yogācāra idealism would be unacceptable, so he “provisionally adopts a view, which later doxographers will describe as Sautrantika, according to which external objects exist, while maintaining the Yogacara denial of external reality in final analysis” (p. 104).

Two limits on this reading are recorded in the material itself. Dreyfus takes Dharmakīrti’s essential preoccupation to be epistemology, not metaphysics, so that “the exact nature of real entities (whether, for example, they are ultimately mind dependent or not) is a lesser concern” (p. 105). And on the True/False Aspectarian question he argues that Dharmakīrti gives a False Aspectarian answer at Commentary III:209–219 and the opposed True Aspectarian answer at III:220–222, in the same passage, and concludes that “Dharmakīrti does not care very much whether aspects are real or not as long as the answer supports his main concern”, which is answering the Nyāya about unitary objects (dreyfus-recognizing-reality, p. 437). If that is right, the doxographical question the Tibetan commentaries argue over is one Dharmakīrti declined to settle.

The Tibetan tradition nonetheless settles it, in more than one direction. Gorampa’s translator reports the range: Tibetan sources make Dharmakīrti either a Sautrāntika or a Chittamatra; chapter 1 has been taken by Tibetan epistemologists to illustrate Sautrāntika realism; chapter 3 has been understood to portray Yogācāra idealism; and on Tillemans’s assessment his final position, with Dignāga’s, was Yogācāra idealism (gorampa-pramanavartika, translator’s introduction, citing Tillemans 2020). Some Tibetan masters place him in Madhyamaka instead (same source, citing van der Kuijp 1983).

Sakya Paṇḍita

The Pramāṇayuktanidhi’s treatment is structurally the clearest of the three, because the chapter announces its own ladder. Chapter 1 lists the competing systems by what each takes to be basic:

།ངོ་བོ་ཉིད་དང་དབང་ཕྱུག་དང༌། །གཙོ་བོ་རྡུལ་དང་རྣམ་རིག་དང༌། །རྟེན་འབྲེལ་ཡིན་ཞེས་གྲུབ་པའི་མཐའ། །ཐ་དད་རང་གཞན་སྡེ་པ་འདོད།

That [the cause of things] is intrinsic nature, Īśvara, primal substance, atoms, cognisance or dependent arising: these are the different tenets held by schools of our own and others. (Pramāṇayuktanidhi chapter 1)

Then it works upwards through the last three. The Vaibhāṣika falls, the early Tibetans who had followed it fall with it, the Sautrāntika position is established, and then the Sautrāntika falls in turn. The chapter’s closing verse marks the join:

།ཇི་སྲིད་ཕྱི་རོལ་ཁས་ལེན་པ། །དེ་སྲིད་རྒྱུ་ལ་གཟུང་ཡུལ་ཟེར། །ཤེས་བྱ་ནང་གིར་ཞུགས་པ་ན། །ཡུལ་དང་ཡུལ་ཅན་གྲུབ་པ་མེད།

So long as external [objects] are accepted, the cause is called the apprehended object; once objects of knowledge are drawn inward, object and subject are not established. (Pramāṇayuktanidhi chapter 1)

Sodargye ranks the three positions the chapter has passed through: on the Sautrāntika view external objects exist, and this is the less definitive position; on the Yogācāra view external objects do not exist while mind does, and this is the provisionally definitive position; on the ultimate Madhyamaka view object and mind alike merge into the dharmadhātu, and this is the fully definitive position (sodargye-pramanayuktanidhi, on chapter 1). Chapter 9 states the same order for the theory of valid cognition and result, in Mipham Rinpoche’s formulation of Dharmakīrti’s three modes: without examination, according with worldly people; with slight examination, according with the Sautrāntika; with extremely subtle examination, according with the Chittamatra (sodargye-pramanayuktanidhi, on chapter 9).

Where Sakya Paṇḍita himself stands is a question the chapter does not answer, and the commentary is careful not to answer it either. It says he replies to the Madhyamaka objection “provisionally standing within the Yogācāra position”, which concedes that the standing is provisional (sodargye-pramanayuktanidhi, on chapter 1). What the chapter records about his own allegiance is narrower and sharper, and it is about the Sautrāntika rung rather than this one: in the Auto-commentary he says he establishes his position following the two Masters of Reasoning while the early epistemologists follow the Vaibhāṣika, and Sodargye reports the remark as made in a displeased tone (sodargye-pramanayuktanidhi, on chapter 9).

Intra-Buddhist criticism

Three objections gathered here are aimed at this school, and they come from three different directions.

The first is the Madhyamaka objection in Pramāṇayuktanidhi chapter 1, and it is the school’s own weapon returned. If the reasoning of freedom from unity and multiplicity destroys the partless atom, it destroys the partless moment: let the mental continuum play the part of the whole and the moments the part of the atoms, and every present moment turns out to have a portion bordering the past and a portion bordering the future, so it is three and not one. The exchange runs to three verses:

།སེམས་རྒྱུན་ཡན་ལག་ཅན་དང་མཚུངས། །སྐད་ཅིག་རྡུལ་གྱི་ཆ་ཤས་བཞིན། །དེས་ན་སྐད་ཅིག་གསུམ་བདག་ཕྱིར། །གཅིག་དང་དུ་མ་བྲལ་ཞེས་ཟེར།

།སྐད་ཅིག་གསུམ་གྱི་བདག་ཉིད་ཕྱིར། །སྐད་ཅིག་གཅིག་པ་མི་སྲིད་འགྱུར། །སྐད་ཅིག་པ་གཅིག་མི་སྲིད་ན། །གསུམ་གྱི་བདག་ཉིད་ཡིན་པ་ཉམས།

།རགས་པ་ཅིག་ཆར་བསྐོར་བས་ན། །དབུས་ཀྱི་རྡུལ་ཕྲན་ཆ་བཅས་འགྱུར། །དུས་གསུམ་ཅིག་ཆར་མི་སྐྱེ་བས། །ད་ལྟའི་སྐད་ཅིག་ཆ་མེད་ཡིན།

“The mental continuum is like a whole, and its moments are like the parts of an atom; so, since [a moment] has the nature of three moments, it is neither one nor many”, so [the Mādhyamikas] say. If it has the nature of three moments, a single moment becomes impossible; and if a single moment is impossible, having the nature of three is lost. In a coarse [object], because [the atoms] surround it all at once, the central atom comes to have parts; but since the three times do not arise all at once, the present moment is partless. (Pramāṇayuktanidhi chapter 1)

Two replies, of unequal weight. The first (your three collapses with our one, since three is built from one) Sodargye calls a clever riposte and concedes that it works at the verbal level only; the Madhyamaka reasoning, in Mipham Rinpoche’s image, is fire meeting dry grass, and reduces everything it touches to emptiness (sodargye-pramanayuktanidhi, on chapter 1). The second is substantive: atoms and moments are disanalogous, because six atoms genuinely do surround a central one at the same time, whereas the past has ceased and the future has not arisen, so there is nothing coexistent for the parts-analysis to grip. Even this Sodargye holds only provisionally tenable: the partless moment is time’s smallest unit conventionally, and pushed further it dissolves as the partless atom does (same source). The school’s own commentator, that is, does not think the school survives the objection. Sakya Paṇḍita himself is firmer. In the auto-commentary he notes that some Tibetans equate the analysis of partless moments with the analysis of partless atoms and conclude that no partless moment of cognition exists, and he rejects that conclusion (auto-commentary on Pramāṇayuktanidhi chapter 1).

The second objection is that mind-only cannot do the work pramāṇa needs done. It appears in chapter 1 as the second horn of an attack on Sakya Paṇḍita’s thesis that the object of comprehension is solely the specifically characterised phenomenon: the realist schools have been refuted by Vasubandhu and Dharmakīrti, but a merely cognition-internal apprehended object will not serve either, because the Yogācāra grants no apprehended object built from partless atoms, while pramāṇa is framed to accept external objects and Dharmakīrti says as much (sodargye-pramanayuktanidhi, on chapter 1). Sakya Paṇḍita’s answer is the whole ladder that follows, which is a change of level rather than a rebuttal on the objector’s own terms.

The third is not directed at the Indian school at all but at the ordering itself, and it comes from Dreyfus against the Tibetan commentators who make the ascent soteriological. Sakya Chokden treats the True Aspectarian positions as steps towards a final view, “the final False Aspectarian view, which Sakya Chok-den describes as Madhyamaka and attributes to Prajñākaragupta” (dreyfus-recognizing-reality, pp. 435–436). On this reading, Dreyfus says, “the scheme of ascending scales becomes soteriologically oriented”, the steps leading “the thinker from a lower view of reality to a higher realization of nonduality”; he adds that this “seems to be Sakya Chok-den’s understanding” of Dharmakīrti, so the attribution is his inference from Sakya Chokden’s method rather than a quotation of him (p. 437). Dreyfus accepts the model of ascending scales and rejects the soteriological reading of it: Dharmakīrti nowhere relates his Yogācāra views to soteriology, and where he does discuss soteriology, in Commentary II, he refers to the selflessness of persons and not to emptiness understood as the absence of duality between subject and object, so “it is doubtful that the scheme of ascending scale is primarily soteriological” (p. 438).

What the debate turns on

Every rung of the ladder set out here is entered by refuting the rung below, and none of them is entered by observation. That is what makes the school hard to place. The Sautrāntika position is a working ontology one can argue in front of Naiyāyikas; the Chittamatra position is where the arguments actually end up when the atom is pressed, and the whole apparatus of Pramāṇayuktanidhi chapter 9 (faculty, aspect, valid cognition and result) is then re-posited on the new ground without a new argument for any of it. Sodargye’s own summary makes the pattern explicit when he says that on the Chittamatra view what is hidden in an inference is beings’ traces, and that “positing it by beings’ traces, all reasoning is entirely reasonable, with nothing unreasonable whatever” (sodargye-pramanayuktanidhi, on chapter 9). Every problem the external object was solving is solved again by the trace.

What that leaves unsettled is whether the ascent is a doctrine or a technique. Dreyfus’s answer is that it is a technique with a real hierarchy behind it, and that Dharmakīrti’s own commitments beyond the epistemology are thinner than either his Tibetan interpreters or his modern readers would like. Sakya Chokden’s answer is that the hierarchy is a path. The Pramāṇayuktanidhi material declines both by never saying which rung Sakya Paṇḍita himself is standing on when he is not arguing: chapter 9 is set on Sautrāntika ground by the Auto-commentary’s own statement, chapter 1 ends above it, and no verse of the root text tells the reader how to hold the two together. [unverified] whether the Auto-commentary settles this; the English sources here report its ruling only for the faculty in chapter 9, and the Tibetan of the Auto-commentary has not been checked on the point.

Graph View

Backlinks