Pramana Wiki

Sautrāntika

34 min read

Sautrāntika (མདོ་སྡེ་པ་)

The Sautrāntika is the school Dignāga, Dharmakīrti and Sakya Paṇḍita write from rather than against. Its name is rare in them. It occurs once across the six chapters of the pramanasamuccaya, and nowhere in the pramanavartika verses themselves. But its positions are everywhere: that only partless particles and partless instants of cognition are real, that a cognition arises from an object that has already ceased and carries its aspect (རྣམ་པ་, ākāra), that the object is therefore never seen bare. sakya-pandita builds the position up over the second half of Pramāṇayuktanidhi chapter 1, sets his whole treatment of perception (མངོན་སུམ་) in chapter 9 upon it, and then, in the closing verses of chapter 1, takes it apart.

That double movement is the shape of the page. A page on the Sāṃkhya can be organised around refutations; this one has to be organised around a position that is adopted, used, and finally withdrawn.

One caution before anything else. In these three texts “Sautrāntika” is a doxographic label, not a report about an Indian sect. Dreyfus is explicit that Dignāga and Dharmakīrti “adopt a so-called Sautrāntika standpoint” (dreyfus-recognizing-reality, p. 49), and that the view Dharmakīrti provisionally works within is one “which later doxographers will describe as Sautrāntika” (p. 104). Whether the historical school held everything the label carries is a separate question, and an open one.

Overview

The historical outline in this paragraph follows the Wikipedia article on the school; none of the scholarly works cited below treats the sect as a sect. Its origin is traced to the elder Kumāralāta (c. 3rd century CE), author of the Kalpanāmaṇḍitīkā, and Charles Willemen identifies its adherents as a western branch of the Sarvāstivādins active in Gandhāra who separated before about 200 CE. They kept the Sarvāstivāda Vinaya lineage, which is why no independent ordination line survives: the split was doctrinal, not institutional. The Sarvāstivādins sometimes called them Dārṣṭāntika, those who use the method of examples, possibly as a slight. Śrīlāta, Vasubandhu’s teacher, wrote a Sautrāntikavibhāṣā, now lost; Yaśomitra characterises the group as those who hold to the sūtras but not to the Abhidharma commentaries. The name says the same thing: the sūtras are authoritative scripture, the Abhidharma treatises are not.

What the epistemologists actually draw on is the version of that critique in Vasubandhu’s commentary on his own Abhidharmakośa. Dreyfus sets out the problem it solves. The Vaibhāṣika (བྱེ་བྲག་ཏུ་སྨྲ་བ་) list of ultimately real phenomena combines two incompatible typologies: an ontological one, in which only the four elements are substances, and a phenomenological one, in which the five kinds of sensible object (colour, sound, odour, taste, touch) are also substances. So colours and tastes are counted real while being aggregates of atoms, and the criterion of the real (resistance to analysis) contradicts itself. Vasubandhu’s reduction removes the sensibilia from the list: ontologically only the four elements are substances, and what we perceive is not real in the full sense but is the causal work of its constituent atoms. Dreyfus quotes Hattori: the Sautrāntikas “did not, like the Sarvāstivādins, consider as real what is seen by the organ of vision such as the blue color or the round shape; instead, by understanding as real that which has the efficacy to produce visual cognitions, they sought to resolve the difficulty inherent in the Sarvāstivādin atom theory.” On this account “only infinitesimal atoms and moments of consciousness are real”; everything else is real only as an object of conventional practice, a view Dreyfus compares to Wilfrid Sellars’s claim that tables, ice cubes and colours do not really exist (dreyfus-recognizing-reality, pp. 84–85).

Who exactly the school was is a live scholarly problem, and the disagreement is on the record. Dreyfus records three positions: André Bareau holds it a small and relatively unknown Hīnayāna school defending the positions of Vasubandhu’s auto-commentary; S. Mookerjee holds the Sautrāntikas an independent school marking a transition between Hīnayāna and Mahāyāna; Hattori sees the system as a bridge between Sarvāstivāda and Yogācāra. The Geluk doxographical account, Dreyfus adds, is in effect a compromise between these (dreyfus-recognizing-reality, p. 466 n. 14). Nobuyoshi Yamabe, again by way of the same article, calls the identification of the Sautrāntika one of the largest open problems in current Buddhist scholarship. None of the three texts settles it, and none of them needs to: what they use is a position, and they attach it to a name they inherited.

Following scripture and following reasoning

Tibetan doxography cuts the school in two: Sautrāntikas following scripture (ལུང་གི་རྗེས་འབྲང་གི་མདོ་སྡེ་པ་) and Sautrāntikas following reasoning (རིགས་པའི་རྗེས་འབྲང་གི་མདོ་སྡེ་པ་). Dreyfus records the pair as a Geluk distinction and refers the reader to Anne Klein for it (dreyfus-recognizing-reality, p. 466 n. 15). None of the three root texts uses the pair as a school label. It is a later repair to a classification that would not otherwise hold the epistemologists, and the repair is worth understanding before reading anything below, because everything this page calls the Sautrāntika position belongs to one half of the pair.

What separates the halves is which treatise each takes as its authority. Jan Westerhoff puts it in a clause: the scripture-followers are “based on works by Vasubandhu”, the reasoning-followers “on works by Dharmakīrti” (westerhoff-dignaga-dharmakirti, n. 109 to the discussion at pp. 251–252, citing Sopa and Hopkins, Klein, and Jackson). Both halves are Sautrāntika for the reason the name gives: both hold external objects real, and both deny the Vaibhāṣika’s Abhidharma treatises the authority of the sūtras. The subdivision then sorts them again by which treatise they do rely on: the Abhidharmakośa with its auto-commentary on one side, Dharmakīrti’s seven treatises on the other. A school named for refusing the treatises is subdivided by its choice of treatise, which is the first sign that the label is doing classificatory work rather than historical work.

Doctrinally the line falls on perception. The Padmakara translators state the classification and the reason for it together: “Sautrantikas following reasoning” identifies Dignāga and Dharmakīrti as realists who accept an extramental world, but distinguishes them from the Vaibhāṣika and from the Sautrāntikas following scripture “owing to, among other things, their complex and sophisticated theory of perception”. The Vaibhāṣika holds that nonmental objects are known directly by the sense organs. The reasoning-followers take the difference between mind and matter as axiomatic, so that “the mind cannot enter into direct contact with physical entities but detects them indirectly via the aspects, or mental images, that these same entities are said to cast upon it, in the same way that things cause their reflections to appear in a mirror”. Mipham’s verdict on that account is that for anyone who posits an external world, no epistemology is “more coherent than this, and more tenable” (shantarakshita-madhyamakalankara-intro, pp. 25–26). The aspect (རྣམ་པ་) doctrine set out in the sections below is therefore not the Sautrāntika theory of perception but the reasoning-follower’s.

The one place the distinction earns its keep in commentary on the three texts is a small point about causation. The Abhidharmakośa counts mutually dependent simultaneous phenomena (long and short) as a co-existent cause. Khenpo Sodargye notes that the reasoning-following Sautrāntika refuse to treat that as a genuine cause-and-effect relation, taking it as mutual positing only, so that the Abhidharmakośa’s co-existent cause and the logicians’ cooperative cause come apart (sodargye-pramanayuktanidhi, on chapter 6). The split does exactly what its definition predicts: the half that reads the Abhidharmakośa keeps the sixfold causal scheme, and the half that reads Dharmakīrti keeps only as much of it as a causal argument will carry.

Sodargye applies the same pair one storey up, to the Chittamatra, and there he names the parties. Asked what the sense faculty is, the masters following reasoning (Dignāga and Dharmakīrti) make it an uncommon capacity to illuminate form residing on the preceding moment of eye consciousness, while the masters following scripture (Asaṅga, Vasubandhu, and apparently Jñānaśrī) make it a particular latent trace of apprehending objects residing on the foundation consciousness (ཀུན་གཞི་) (sodargye-pramanayuktanidhi, on chapter 9). Vasubandhu comes out a scripture-follower on both storeys, and Dignāga and Dharmakīrti reasoning-followers on both. Whatever the pair is sorting, it is not two communities of monks; it is two kinds of warrant, and the same two authors fall on the same side of it whichever tenet system they are being filed under.

For Geluk pedagogy the reasoning-follower is where the beginner starts. Daniel Perdue’s Course in Buddhist Reasoning and Debate sets out its whole ontology from that standpoint, and notes that this system alone equates ultimate truths with impermanent phenomena, an ultimate truth being defined as “a phenomenon that is ultimately able to perform a function” (perdue-buddhist-reasoning-debate, pp. 65, 187, 213). Perdue is candid about the reason: the Tibetans do not hold the Sūtra School to be the final and highest system, “it is the place to begin, for it is easy to see that it describes the way the world appears to be” (p. 65).

Dreyfus does not think the repair works. The view defended through most of Dignāga’s and Dharmakīrti’s texts “is different from the view usually ascribed to Sautrāntika”; Tibetan scholars see the discrepancy clearly, and “attempt to explain away this discrepancy by describing Dharmakīrti as a ‘Sautrāntika following reasoning’“. His verdict is that the term is oxymoronic and shows “the limitations of doxographical considerations” (dreyfus-recognizing-reality, pp. 20–21). Westerhoff adds a complaint of John Dunne’s: later Tibetan commentators can be found subsuming any of several quite distinct positions under the single label “Sautrāntika” (n. 120, citing Dunne 2004: 71). The asymmetry between the two halves tells in their favour. No doctrine is stated on the scripture-follower’s behalf anywhere in these three texts or the commentaries on them; that half appears only as the thing the reasoning-follower is not.

Core doctrines

Only particles and instants are real

Gorampa states the ontology as he reads it out of Dharmakīrti’s Seven Treatises, and states it as a correction of his contemporaries, who take the verse before him to mean simply that subtle particles are not objects of the sense consciousnesses: this “shows that they have not considered the tenets of the Sautrāntika as explained by the Seven Treatises (of Dharmakīrti). The Sautrāntika school states that subtle particles with no directional parts, and instants of consciousness that contain no former and succeeding parts of time, exist as realities and that the other-knowing direct perception and self-knowing direct perception that apprehend them are unmistaken consciousnesses. Coarseness and all continuums are provisional realities, and the conceptual cognition that apprehends them is mistaken consciousness” (gorampa-pramanavartika, on PV II.87).

The verse he is glossing is the reply to a Vaiśeṣika objection, and it is worth having in full because it fixes how much the school may say the atoms are perceived:

།དུ་མ་ཉིད་ན་སྔ་མ་བཞིན། །ཁྱད་མེད་ཕྱིར་དང་ཕྲ་རབ་ཕྱིར། །རྟོགས་མིན་ཞེ་ན་ཁྱད་མེད་པར། །མ་གྲུབ་ཁྱད་པར་ཅན་རྣམས་ནི། །དབང་ཡུལ་དེ་ཕྱིར་རྡུལ་ཕྲན་མིན།

You [Vaiśeṣikas] may say, “If [the body is] many [atoms], then, like the earlier [, separate atoms], they would be no different and subtle, and so would not be apprehended”; [however,] their being no different is not established. Possessing distinctive features, they are objects of the senses, and so [they are] not [called] atoms. (pramanavartika II.87)

Atoms in contact with other atoms acquire distinguishing features and become objects of the senses; individual uncontacted atoms do not. Many particles do not compose one substantial coarse form, so there is no coarse form apart from the particles, and what the sense consciousness apprehends is many particles in contact presenting an aspect (gorampa-pramanavartika, on PV II.87).

Gorampa’s target here is not the Vaiśeṣika. He quotes Jamyang Gekpo against Tibetan logicians who concede that subtle particles are imperceptible and unobstructive, and then posit a coarse form built from those particles, a separate entity from them, intrinsically existent, obstructive, perceptible, and heavy with a weight of its own rather than the weight of its particles: “Therefore they accept form as the possessor of parts and of this extra weight, as asserted by the Vaiśeṣika” (quoted at gorampa-pramanavartika, on PV II.87). The charge is that a Buddhist realism about everyday objects is a non-Buddhist realism wearing Buddhist vocabulary. Sakya Chokden makes the historical version of the same point: no Indian proponent of either the Sautrāntika or the Vaibhāṣika view accepted commonsense objects as real at the ontological level, and the Tibetan insistence that they are “does not correspond to what is found in most Indian texts” (dreyfus-recognizing-reality, p. 97).

One qualification from Dreyfus cuts against the neatness of this. He argues that the reductionist account is present in Dharmakīrti but is not the whole of him: there is “a tension within Dharmakīrti’s ontology between an atomistic reductionism, which is in accordance with his overall ontological parsimony, and a less reductionistic delineation of reality, which allows for the reality of extended objects”, produced by the fact that perception is supposed to be an undistorted reflection of reality and we do seem to perceive extended things (dreyfus-recognizing-reality, pp. 85–86). On momentariness (སྐད་ཅིག་མ་, kṣaṇika) he finds no such tension: “Dharmakīrti’s view of momentariness does not differ from the Sautrāntika view advocated by Vasubandhu” (p. 63).

Cognition arises from the object and takes its form

Sakya Paṇḍita states his own position in chapter 1 in two verses, the first giving the causal conditions and the second the consequence:

།ཡུལ་དང་དབང་པོ་ཡིད་ལ་བྱེད། །དེ་ལས་སྐྱེས་པ་རྣམ་ཤེས་ཡིན། །ཟས་སོགས་བུ་ཡི་རྒྱུ་ཡིན་ཡང༌། །ཕ་མའི་གཟུགས་བཞིན་ཡུལ་འདྲར་སྐྱེ། །དེས་ན་ཡུལ་ལ་འབྲས་བུ་གཉིས། །བློ་ཡང་ཚུལ་གཉིས་ཡིན་པར་འདོད།

That which arises from object, sense faculty and mental engagement is consciousness. Although food and the like are causes of a child, it is born resembling its parents’ form; [so too consciousness] arises resembling its object. Hence the object has two results, and the mind too is held to have two aspects. (Pramāṇayuktanidhi chapter 1)

Sodargye reads the three as the standard triad of conditions: the observed condition (དམིགས་རྐྱེན་, ālambanapratyaya), the external object; the immediately preceding condition (དེ་མ་ཐག་པའི་རྐྱེན་, samanantarapratyaya), mental engagement; and the ruling condition (བདག་རྐྱེན་, adhipatipratyaya), the sense faculty (དབང་པོ་). They gather in the first moment and cognition arises in the second (sodargye-pramanayuktanidhi, on chapter 1). The second verse answers the obvious follow-up: if three conditions produce the cognition, why does the cognition resemble the object rather than the faculty or the attention? Because that is how causation runs elsewhere too: food, warmth and time are all causes of a child, and the child resembles its parents.

The last line is the load-bearing one. “The object has two results” means that an entity produces both the next moment of its own kind, for which it is the substantial cause, and a sense cognition bearing its form, for which it is a cooperating condition alongside faculty and attention (auto-commentary on Pramāṇayuktanidhi chapter 1; sodargye-pramanayuktanidhi, on chapter 1). “The mind too is held to have two aspects” is read in two ways. In the auto-commentary the two aspects are the cognition that knows the object and the self-awareness that knows that cognition; Gorampa takes them as the objective aspect and the subjective aspect (translation notes to Pramāṇayuktanidhi chapter 1). Sodargye takes object-apprehension to be subsumed in self-awareness (རང་རིག་, svasaṃvedana) in the end (sodargye-pramanayuktanidhi, on chapter 1). The phrase is Dignāga’s. The Tibetan of the Pramāṇasamuccaya has བློ་ཡི་ཚུལ་གཉིས་ where the Pramāṇayuktanidhi has བློ་ཡང་ཚུལ་གཉིས་, and the argument in Dignāga is the same one:

ཡུལ་ཤེས་པ་དང་དེ་ཤེས་པའི། །དབྱེ་བས་བློ་ཡི་ཚུལ་གཉིས་ཉིད། །

By the distinction between cognising the object and cognising that [cognition itself], the mind (buddhi) has precisely two modes. (pramanasamuccaya I)

Dignāga’s first account of valid cognition and its result carries the family simile as well (an effect bears the form of its cause, as a son resembles his father), and it is from these accounts that he derives the threefold division into object-aspect, subject-aspect and self-awareness-aspect (pramanasamuccaya I, the three accounts of the result of valid cognition). So the verse in which Sakya Paṇḍita states the Sautrāntika position is, in its wording, a restatement of Dignāga.

The aspect, and an object that is never seen bare

What follows from a cause that has ceased is that the object is not what appears. Sodargye puts the school’s half of this bluntly against the schools above it: “The Sautrāntikas hold that we do not see the pillar’s particular itself but only its aspect; Cittamātra and above hold that the aspect of the external object appears before the mind” (sodargye-pramanayuktanidhi, on chapter 4). Gorampa spells out the mechanism as a three-part sequence: first, the particular particles in contact, which bestow the aspect; second, the coarse aspect appearing at the time of the sense consciousness, which is not coarse form but the entity of the sense consciousness itself; third, the aspect appearing as one to a subsequent conceptual cognition, projected outwards and engaged by way of exclusion (gorampa-pramanavartika, on PV II.87). Only the first is an external phenomenon. The other two are cognition and conception.

In his commentary on the Pramāṇayuktanidhi, Gorampa reduces the school to five theses. First, what appears to a sense perception of blue as something out there at a distance is the cognition itself, not an external object. Second, the hidden external object is the cause, and the cognition bearing its aspect is the result. Third, conceptual cognitions and mistaken non-conceptual cognitions have no entity as their appearing object and are mistaken, while non-conceptual perceptions have entities as their appearing objects and are not. Fourth, “external object” names two things: the coarse, extended object grasped by the conceptual thought that follows a perception, and the subtle external object that actually causes the sense cognition. Fifth, calling an external object and the perception of it “apprehended” and “apprehender” is only a convention, since they are cause and effect and so never exist at the same moment (Gorampa on Pramāṇayuktanidhi chapter 9).

This is the doctrine that gives the school its standing difficulty, and the difficulty is best stated in the form Dreyfus gives it, because the Tibetan commentators argue about that form rather than about the doctrine in the abstract. The Sautrāntika theory of perception rests on a view of causation in which causes cease when their effects arise, which is the direct consequence of radical momentariness. Applied to perception, it “makes it very difficult to explain how perception directly perceives external objects. Since objects have ceased when perceptions of them arise, how can they be perceived? The Sautrāntika response grants that external objects do not appear directly (dngos su mi snang) but are represented by aspects which are similar to them” (dreyfus-recognizing-reality, pp. 380–381).

Gorampa’s word for this is “hidden”, and Dreyfus notes that he uses it in a way peculiar to himself: the object is hidden in that it is delivered to consciousness through its representation, not in that it is inaccessible in principle. Consciousness “does not directly cognize (dngos su mi rig) the object”; perception arises only as bearing a resemblance to it. For Gorampa the resemblance is the decisive element, because it is what keeps perception from being shut up inside its own appearance the way conception is (dreyfus-recognizing-reality, p. 419). He is a consistent representationalist and still refuses to say that external objects are imperceptible, on the ground that in Dharmakīrti’s system being real and being an object of perception come to the same thing, so an object that could only be inferred could not be real (p. 420).

The time-gap objection presses him hardest: cause and effect cannot coexist, both last an instant, so there is no moment at which consciousness perceives the object. Gorampa’s answer is that the objection analyses perception in terms of agent and action, and for a Buddhist agent, action and object of action are conventions rather than real things, so of course nothing is found when perception is analysed that way: “external objects are perceived in reality”, though not when so analysed (dreyfus-recognizing-reality, pp. 420–421). Dreyfus does not let this pass: if agent and action are only conventionally real, saying objects are “perceived in reality” reduces to a convenient description of a momentary causal process, and “our phenomenological intuitions are hardly satisfied” (p. 421).

Four perceptions, and a faculty that is not conscious

Sakya Paṇḍita’s chapter on perception is set on Sautrāntika ground and says so, in a line that sorts the Buddhist schools by how many kinds of perception each can accept:

།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)

The count follows from two commitments. Because the Sautrāntika accepts external objects, sense perception and mental perception stand; because it accepts self-awareness, self-aware perception stands; yogic perception no Buddhist school denies. The Vaibhāṣika loses one by rejecting self-awareness, and the Chittamatra (སེམས་ཙམ་) on its ultimate view keeps only self-aware and yogic perception, since without a real external object there is nothing for sense and mental perception to take (sodargye-pramanayuktanidhi, on chapter 9).

The faculty gets an argument of its own, and it is an inference rather than an observation:

།དབང་པོ་ཡུལ་ལ་འཛིན་ནུས་ཏེ། །རྗེས་སུ་འགྲོ་ལྡོག་ངེས་པས་འགྲུབ།

The sense faculty (དབང་པོ) is the capacity to apprehend the object; this is established by the ascertainment of positive and negative concomitance (རྗེས་སུ་འགྲོ་ལྡོག). (Pramāṇayuktanidhi chapter 9)

The faculty is not the eyeball or the flap of the ear (Sodargye disposes of that with the observation that large-eared beings would then hear louder sounds) but a capacity, established by the fact that with an object present and attention directed, damaging the eye stops eye-consciousness. Whether that capacity is itself a cognition or an insentient thing divides the schools, and Sodargye reports that the Auto-commentary settles it here on the Sautrāntika side: the faculty is insentient, and it is on that view that chapter 9 is set forth. On the Chittamatra side the faculty is posited upon cognition, either upon the preceding moment of eye-consciousness (the reasoning-followers, Dignāga and Dharmakīrti) or as a latent trace upon the foundation consciousness (the scripture-followers) (sodargye-pramanayuktanidhi, on chapter 9).

Valid cognition and its result

Chapter 9 ends with a survey of how each school divides object of comprehension (གཞལ་བྱ་, prameya), valid cognition and result of valid cognition (ཚད་འབྲས་, pramāṇaphala). In the auto-commentary the Sautrāntika scheme runs: the object of comprehension is the hidden object; the cognition that arises resembling it is the valid cognition; and the convention “the object is cognised” is the result. Object and valid cognition-with-result therefore differ both in time and in nature (auto-commentary on Pramāṇayuktanidhi chapter 9). Sodargye gives the same scheme in terms of moments: the first-moment external object, the second-moment cognition bearing its aspect, and that cognition’s discriminating of the object (sodargye-pramanayuktanidhi, on chapter 9). What matters is the qualification Sodargye attaches, following Mipham’s commentary on the Pramāṇavārttika: Dharmakīrti posits valid cognition and result three times over: without examination, in agreement with ordinary people; under slight examination, in agreement with the Sautrāntika; under extremely subtle examination, in agreement with the Chittamatra. The Sautrāntika account is the middle one, and it is “correct” only in that middle sense (same source).

The same chapter’s inference section then supplies a reason for not treating the Sautrāntika scheme as safer than the idealist one. Pressed with the charge that on a mind-only view one cannot infer fire from smoke, since the fire does not exist before it appears, Sodargye replies that the Sautrāntika is in the same position: when smoke appears, the fire that produced it has already ceased, so the Sautrāntika too is inferring the continuum of a cause from an effect by means of a trace (བག་ཆགས་, vāsanā) of the producer-produced relation carried in one’s own continuum (sodargye-pramanayuktanidhi, on chapter 9). The realist gains nothing by his realism at the moment the inference is actually made.

Where the epistemologists stand

Dignāga

The Pramāṇasamuccaya names the school once. In chapter 4, on examples, the point of the dissimilar example is that it is a non-implicative negation (མེད་དགག་, prasajya) and therefore asserts nothing, so that even a party who admits no eternal entities “such as the Sautrāntika who denies that space is a substance” can still use “like space” as a dissimilar example (pramanasamuccaya IV; the naming of the school sits inside Fazun’s running exposition, which incorporates the auto-commentary together with his own glosses, so whether the identification is Dignāga’s or Fazun’s cannot be decided from the exposition itself). The school appears there as an ontological minimalist whose scruples the proof formula has to survive, not as a party to the argument.

The positions later called Sautrāntika are in chapter 1 without the label: the two modes of the mind, self-awareness as the result of valid cognition, the object-aspect as the instrument, the son-resembles-his-father simile for the effect bearing its cause’s form. Dreyfus’s summary is that ontology in both Dignāga and Dharmakīrti “remains subordinated to epistemology”, which is why they feel free to alternate between conflicting metaphysical standpoints: “in most parts of their works, Dignāga and Dharmakīrti adopt a so-called Sautrāntika standpoint, presupposing the existence of external objects”, and elsewhere shift to a Yogācāra rejection of them (dreyfus-recognizing-reality, p. 49).

Dharmakīrti

Gorampa’s translator reports the Tibetan tradition as divided over Dharmakīrti’s own tenets, between Sautrāntika and Chittamatra, and reports the standard division of labour within the Pramāṇavārttika: “The first chapter on logic and inference for oneself has been adopted by Tibetan epistemological thinkers as illustrating the realism of Sautrāntika tenets”, while chapter 3 on perception has been understood as Yogācāra idealism (translator’s introduction to gorampa-pramanavartika). Gorampa’s own annotations to chapter 1 assume it throughout: “the tenets of the Sautrāntika system as presented in this work”; “In the Sautrāntika philosophy, which is the tenet of the first chapter at least, intrinsic existence (rang mtshan, svalakṣaṇa) is accepted”; the Sautrāntika are realists, and it is in contrast to the reality of specifically characterised phenomena (རང་མཚན་) that isolates are non-real (gorampa-pramanavartika, notes 2, 33 and 101).

Chapter 3 then treats the school as an audience to be led rather than a position to be held. Gedun Drub, expounding the objection that a mind-only view has no use for the doctrine of two aspects, gives Dharmakīrti’s reason for teaching it in terms of external objects anyway: “Saying two aspects based on external object has purpose: to guide the speakers of external object (the Vaibhāṣikas and Sautrāntikas) into true meaning. The two aspects of consciousness are themselves established, since they are established by determinate co-experiencing” (Gedun Drub’s commentary on PV III). The same chapter argues that inference of fire from smoke survives the loss of external objects, being carried by traces instead. Chapters 3 and 4 are followed here through Gedun Drub, who numbers no verses, so arguments from them are given in prose rather than quoted; where a verse reference to chapter 3 appears below, it is Dreyfus’s.

Dreyfus’s account of why Dharmakīrti proceeds this way is the fullest available, and it is an account of a strategy rather than of a confusion. Lower and more commonsensical positions are introduced as needed and withdrawn when analysis is pushed further; the Sautrāntika view “offers a provisional basis for discussion”, the Yogācāra view “describes reality” (dreyfus-recognizing-reality, pp. 104–105). Two things support this. Dharmakīrti demonstrably states conflicting views within a single work: Dreyfus points to PV III.209–222, where the question whether aspects are real receives both answers with no preference marked (p. 103). And the reason for the reticence is audience: Dharmakīrti is arguing with non-Buddhists, and an idealism would be unacceptable to almost all of them, so he “relies on the lowest common denominator between competing Buddhist views” and uses vocabulary with broad currency outside Buddhism (p. 104).

Sakya Paṇḍita

Chapter 1 of the Pramāṇayuktanidhi is the clearest statement of where the tradition stands, because it is built as a ladder and climbs it in front of the reader. Sodargye’s section headings track it: Determining the Sautrāntika position (refuting the Vaibhāṣika, refuting the Snow Landers, establishing one’s own position), then Determining the Yogācāra position, refuting the Vaibhāṣika and Sautrāntika together (sodargye-pramanayuktanidhi, on chapter 1).

The two rungs below are cleared first. Against the Vaibhāṣika claim that the physical faculty rather than consciousness is what sees, the reply is that matter cannot see and that simultaneous things cannot be related. Against the position the root text calls the Snow Landers (གངས་ཅན་པ་), on which object and consciousness are simultaneous and so apprehended and apprehender, the reply is that simultaneity leaves no relation, and that a consciousness with no prior cause is a contradiction; when the opponent splits the roles, making the previous moment’s object the cause and the simultaneous one the apprehended object, the answer is that the cognition has already arisen from the prior object and the simultaneous one has nothing left to do. Sodargye draws the practical consequence: seeing a pillar, one should not think “this present pillar is my concurrent object”, because the object that generated the cognition ceased in the preceding moment (sodargye-pramanayuktanidhi, on chapter 1). Both verse exchanges are quoted in full on the vaibhasika page.

The commentary identifies the Snow Landers as early Tibetan logicians and does not name one. Dreyfus documents the same position in Chapa Chökyi Sengé, who insisted that the held object and the appearing object are the same and that object and consciousness coexist, and quotes Sakya Chokden’s paraphrase of him: “The master Cha-ba asserts that what is called held object is the appearing object. [This object] only exists simultaneously with the consciousness that takes it as its object because a nonsimultaneous subject and object [relation] is not tenable” (dreyfus-recognizing-reality, p. 381). The attribution is Dreyfus’s and Sakya Chokden’s, not the commentary’s.

Sakya Paṇḍita says plainly elsewhere where the Sautrāntika belongs. In his Entrance Gate for the Wise (མཁས་པ་རྣམས་འཇུག་པའི་སྒོ་), when external things are being examined, the Sautrāntika tenets are the reasoning that runs by the force of fact; when external things are being refuted, the Mind-Only’s are; and when reality itself is being settled, the Madhyamaka’s are. The auto-commentary on chapter 9 says the same of Dharmakīrti: when he accepts external objects he proceeds as the Sautrāntika does, and when the knowable is drawn inward he accepts the no-aspect [Yogācāra] (auto-commentary on Pramāṇayuktanidhi chapter 9). Even yogic perception is set out on this ground. Dharmakīrti, whom Sakya Paṇḍita calls the crown jewel of the logicians, follows the Sautrāntika in defining yogic perception as the seeing of the four truths, and only then extends it, by the force of fact, to selflessness (auto-commentary on Pramāṇayuktanidhi chapter 9).

Then comes the own-position verse quoted above, and with it chapter 9’s whole apparatus. Sodargye reports Sakya Paṇḍita’s own statement of allegiance from the Auto-commentary, and reports its tone: he establishes his position following the two Masters of Reasoning, Dharmakīrti and Dignāga, whereas the early epistemologists establish theirs following the Vaibhāṣika, “in a rather displeased tone”, and unwilling to do the same (sodargye-pramanayuktanidhi, on chapter 9). Elsewhere the same commentary notes that in analysing objects the Pramāṇaviniścaya and its like stand on the Sautrāntika standpoint (on chapter 2), and that chapter 3’s fourfold division of universal and particular presupposes the Sautrāntika doctrine that the present is real and past and future are not (on chapter 3).

Having climbed to that rung, chapter 1 kicks it away. The argument is the co-apprehension (ལྷན་ཅིག་དམིགས་པ་, sahopalambha) reason, preceded by the unity-and-multiplicity refutation of the atom:

།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 chapter’s closing verse states the whole movement as a conditional, which is why it can be read either as a defence of the Sautrāntika or as its dismissal:

།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)

Gorampa closes his commentary on the chapter with a verse of his own that says the same: Dignāga teaches two objects of comprehension by way of how cognition engages them, but what is engaged is only the specifically characterised; the Sautrāntika has external objects as apprehended objects, and for the proponent of mind-only there is no such thing as an apprehended object at all (Gorampa on Pramāṇayuktanidhi chapter 1).

Dreyfus reports the doxographical consequence, and reports that it is Sakya Paṇḍita’s doing. Sakya Paṇḍita describes Dharmakīrti, in Dreyfus’s words, as “advocating a system in which Sautrāntika and Yogācāra ideas are mixed”, a description Gyeltsap renders as a text common to both (mdo sems thun mong ba, as Dreyfus gives it) and which later Sakya and Geluk writers accept (dreyfus-recognizing-reality, p. 33). Summing up the field, Dreyfus places Sakya Paṇḍita, Gorampa, Gyeltsap and Kaydrup together as having described Dharmakīrti “as a Yogācārin who defends a Sautrāntika position for essentially strategic reasons throughout most of his work”, which lets them use his epistemology without accepting his ontology; against the earlier Tibetan reading of Ngok and Chapa, Sakya Paṇḍita “insists on the first view, presenting Dharmakīrti as a Yogācārin who uses Sautrāntika ideas as a bridge to his system” (pp. 428–429).

Intra-Buddhist criticism

The atom does not survive analysis. The refutation inside chapter 1 is the reasoning of freedom from unity and multiplicity, applied to the partless particle:

།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)

Sodargye runs it through Vasubandhu’s six-atoms-around-a-centre analysis: if the central atom has spatial parts it is not partless, and if it has none, no aggregation of them makes a coarse thing: even Mount Meru would collapse to a point. He is careful about the argument’s scope: here it is used from the Yogācāra standpoint to refute Vaibhāṣika and Sautrāntika atomism, not as a Madhyamaka reasoning establishing ultimate emptiness, and it does not extend to refuting the mind the Yogācāra accepts (sodargye-pramanayuktanidhi, on chapter 1). Sakya Paṇḍita gives the same warning himself in the Entrance Gate for the Wise: to read “neither one nor many” here as a Madhyamaka argument mistakes the context (translation notes to Pramāṇayuktanidhi chapter 1). The aspect falls with the atom, which is why the second clause of the verse matters: an aspect is an aspect of something.

The indestructible particle is a permanence. Chapter 7, on contradiction, surveys three accounts of what happens when a fire-particle meets a water-particle, and the third is the Buddhist one:

།འགའ་ཞིག་ཕྲད་པ་མེད་པར་འདོད། །རྟག་སྨྲ་དེ་དག་འཁྲུལ་པ་ཡིན།

some [Hearer schools] hold that there is no contact at all. These proponents of permanence are all mistaken. (Pramāṇayuktanidhi chapter 7)

Sodargye identifies the third party as the Vaibhāṣika and Sautrāntika masters, who hold that the particles never touch but that the stronger occupies the weaker’s place and blocks it from forming into a coarse thing; the two differ slightly, the Vaibhāṣika holding that particles genuinely touch and the Sautrāntika that an interval always remains, as with one stone set on another. He also explains why Sakya Paṇḍita’s verdict of “proponents of permanence” (རྟག་སྨྲ་) covers Buddhists: although they do not call the particle permanent, they hold that under analysis its nature is indestructible, since the smallest unit out of which coarse things are composed cannot be destroyed. The refutation is that whether or not the particles touch, if both persist together the relation is coexistence, and coexistence is exactly what a contradiction of non-coexistence excludes (sodargye-pramanayuktanidhi, on chapter 7).

A causal theory cannot show that its causes are there. Chapa’s objection is the sharpest thing said against the school anywhere in this material, and Dreyfus notes that it is borrowed from Yogācāra polemic and turned to a different end. If we have no access to objects independent of perceptions, how do we determine that the two resemble each other, or that they are cause and effect? And an object that has ceased when its perception arises is hidden, and cannot be the object of that subject. Then the sha za, the flesh eater: the Sautrāntika posits an entity that is in principle unobservable, so why not say a flesh eater casts the aspect on consciousness? If the reply is that no flesh eater is established by perception or inference, the same holds of the hidden object (dreyfus-recognizing-reality, pp. 380–381). Chapa’s own conclusion is not idealist but conventionalist: epistemology should follow common convention, and the held object should be the appearing object, coexisting with the cognition. What Sakya Paṇḍita answers is the coexistence half, in chapter 1; the flesh-eater argument itself is nowhere taken up in the Tibetan commentaries.

Representationalism taken to its end. Sakya Chokden pushes the analysis further than Gorampa will. He rejects similarity even within the Sautrāntika framework, asking how an aspect, which is imprinted consciousness, could resemble an aggregation of atoms; he holds that external objects are not perceptible but only inferred, which Gorampa rejects as incompatible with their being real (dreyfus-recognizing-reality, pp. 419–420). Dreyfus reads the result as a near-solipsism that Sakya Chokden is content to accept, because for him “Dharmakīrti’s so-called Sautrāntika theories, which often are the primary focus in his work, seem to be a way to build a bridge with the Yogācāra idealism”: accepting the Sautrāntika theory of perception is a way of coming to see how little evidence we have for an external world, and that realisation opens the door to Mind-Only (p. 427).

What the debate turns on

Whether the external world in this system is something shown or something postulated. Gorampa’s answer is that it is shown, and that the appearance of inference is an artefact of the idealist’s own move. Dharmakīrti’s argument that perception must depend on external objects, since it does not arise where those are unobserved, is not offered to explain how we infer an unobservable world; it is offered because the Mind-Only opponent has already discounted appearance as irrelevant by pointing out that the appearance is part of the cognition. Once appearances are brushed aside, the external world can only be argued for. Dreyfus endorses the reading and draws the conclusion sharply: “for the Sautrāntika, the external world is not beyond the veil of appearance; on the contrary, it is so obvious that idealist philosophers miss it”, and interpreters who treat it as a mere logical presupposition or a transcendental principle forget that Dharmakīrti says external objects cause their perceptions, so they are effective, hence real, hence perceptible (dreyfus-recognizing-reality, p. 422).

Sakya Chokden’s answer is that it is postulated, and deliberately so, as a bridge. Sodargye’s answer is a third: the Sautrāntika account is what survives slight examination and not subtle examination, so the question of whether it is finally true has been declined rather than answered.

What none of this settles is what Dharmakīrti himself thought he was doing. Dreyfus states the residue as a question rather than a result: “If Dharmakīrti really holds this philosophy, why does he not say so and why does he focus instead on a contradictory view? Why does he insist on providing us with incompatible ontologies?” He suggests that the answer may be that defending idealism was not his foremost concern (dreyfus-recognizing-reality, p. 427). The Tibetan consensus these texts belong to answers the question with the ascending scales, but the scales are a reading, and Gorampa’s “hidden” (an object that is hidden without being inaccessible, perceived without being directly perceived) is the point at which the reading is under most strain.

The school’s afterlife in the curriculum explains why the position outlived the sect. Dreyfus observes that the Geluk student meets the distinction between nominal and real existence first in the Sautrāntika system, where nominal existence is restricted to a few phenomena such as object universals and is posited on the basis of the real existence of other things; that makes it an easier version of a distinction the student will later need in a far subtler form in Madhyamaka (dreyfus-recognizing-reality, p. 458). A position kept as a teaching step is a position no one has to defend, which may be part of why it is so rarely argued for in these three texts and so constantly used.

Graph View

Backlinks