Pramana Wiki

Sāṃkhya

15 min read

Sāṃkhya (གྲངས་ཅན་པ་)

The Sāṃkhya (“enumeration”) school is the opponent Dharmakīrti works hardest against on the question of universals, and its doctrine that all things share one nature is the standing target of the single longest polemic in pramanavartika chapter 1. All three of our primary texts engage it: dignaga refutes its theories of perception and inference in the pramanasamuccaya, dharmakirti attacks its metaphysics of the single nature and its account of liberation in the pramanavartika, and sakya-pandita treats its one-substance universal in the pramanayuktanidhi—where he also charges the early Tibetan logicians with holding the same view without admitting it.

Overview

Sāṃkhya is one of the six orthodox (āstika) schools of Indian philosophy. Its traditional founder is the sage Kapila, of whom no writings survive; the earliest surviving text is Īśvarakṛṣṇa’s Sāṃkhyakārikā (c. 350 CE, in any case before its Chinese translation of 569 CE), whose author presents himself as Kapila’s successor through Āsuri and Pañcaśikha (Wikipedia). Dreyfus, following Frauwallner, notes an early Sāṃkhya epistemologist, Vārṣagaṇya (early fourth century CE), reconstructable from fragments in Dignāga and Jinendrabuddhi—the school may have been among the first in India to develop a logico-epistemological system (dreyfus-recognizing-reality, ch. 2). As an independent tradition it declined around the tenth or eleventh century and was absorbed into the orthodox fold; Vijñānabhikṣu attempted a theistic revival in the sixteenth century, and today its cosmology survives chiefly inside classical Yoga, which shares its metaphysics, rather than as a school with its own adherents (Wikipedia; general web references—the precise dating of the decline is secondary reporting, not from our material).

The school’s name comes from its enumeration of twenty-five realities (tattvas), of which two matter most. The first is prakṛti (universal principle, སྤྱི་གཙོ་བོ་; also called the innate nature, རང་བཞིན་, and, in the contexts where it is being proved rather than described, the primal substance, pradhāna, གཙོ་བོ་): a hidden, eternal, unchanging stratum that pervades all existence, a cause but never an effect, resting in an equilibrium of three qualities (guṇa, ཡོན་ཏན་)—sattva (སྙིང་སྟོབས་), rajas (རྡུལ་) and tamas (མུན་པ་)—and manifesting as the entire world of experience. The second is puruṣa (the person or self, སྐྱེས་བུ་): pure consciousness, neither cause nor effect. Everything except these two is a manifestation of prakṛti. Liberation, for the Sāṃkhya, is a matter of knowledge: the bondage of beings is the failure to see that experience is only prakṛti’s display, and when meditative concentration comes to perceive prakṛti directly, its manifestations withdraw and the self stands apart (translator’s introduction to Gorampa’s commentary, gorampa-pramanavartika). The school was mainly nontheistic, and it accepted three means of valid cognition (ཚད་མ་): perception, inference and scriptural testimony (same source).

For the Buddhist epistemologists the school matters because this metaphysics is a theory of universals. Prakṛti is the one real universal (སྤྱི་); every particular (བྱེ་བྲག་) is a manifestation (pariṇāma) of it and therefore one in nature with it and with every other particular. Dreyfus calls the position a “partly monist” realism, distinct from the Nyāya’s plural real universals: there is only one real universal, and the diversity of predicates must somehow be constructed from it—which, he observes, leaves the school epistemologically weak, unable to explain predication (dreyfus-recognizing-reality, ch. 4).

Core doctrines

The single nature (རང་བཞིན་གཅིག་པ་)

The claim, as Gorampa reports Kapila arguing at PV I.163: unless something pervaded entities over and above their mere exclusion-isolates (ལྡོག་པ་), a sprout could not be the result of seed and water and fertiliser together, for the seed would be totally distinct from the water and fertiliser. Many causes produce one result, so the many must be of a single nature (pramanavartika I.163 with gorampa-pramanavartika ad loc.). This is the Sāṃkhya’s positive argument, and it is worth pausing on, because it is a real argument: cooperation among distinct causes needs explaining, and the single nature is offered as the explanation. Everything Dharmakīrti says against the school in this stretch of the text is addressed to it.

The result already exists in its cause (satkāryavāda)

Since manifestation is all that ever happens—prakṛti never creates anything new—the result exists, unmanifest, in its cause. Gorampa attaches the view to PV I.22: the Sāṃkhya “asserts that the result exists in the cause even though it is not apprehended” (gorampa-pramanavartika on PV I.22). In chapter 2 the same doctrine appears in a stronger form: all entities possess the potential for consciousness (pramanavartika II.165).

The universal is one entity with its particulars

Applied to language and cognition, the single nature yields a distinctive semantics: the direct referent of a word is a real universal that is one in entity with its particulars—against the Vaiśeṣika, for whom the universal is a really existing but separate substance (pramanavartika I.93–94; I.150–151). Sakya Paṇḍita reports the same position in the pramanayuktanidhi: the twenty-three principles apart from puruṣa and prakṛti are manifestations of prakṛti, so universal and particular are one substance, related as fire and its heat—one blue-universal standing whole on the blue cloth, the blue mirror and the blue eye alike (pramanayuktanidhi chapter 3, with the auto-commentary).

Perception and inference

The Sāṃkhya account of direct perception (མངོན་སུམ་), as Dignāga reports it: the five sense faculties (དབང་པོ་), empowered by the mind (manas, ཡིད་), operate on their objects in the present—and objects differ in kind because the three guṇas stand in different proportions in them (pramanasamuccaya I, refutation of rival views). In inference the school divided proof into two kinds, which our translation renders “possessing-the-aspects” (vīta, རྣམ་པ་ལྡན་པ་) and “dismiss-and-return” (avīta, གསལ་ཏེ་འོངས་པ་, where the text carries an editorial emendation of གསལ་ to བསལ་)—the second establishing a thesis by eliminating the alternatives—and its standard proof of the primal substance ran: pradhāna exists, because individual things are seen to have positive concomitance (རྗེས་འགྲོ་)—that is, because distinct things converge on common effects, they must share one underlying cause (pramanasamuccaya III, refutation of rival views).

The refutations

Dignāga

The pramanasamuccaya refutes the Sāṃkhya twice, once on perception and once on inference.

Against the perception theory (PS I), Dignāga opens with a dilemma on the guṇa-proportion account of objects. If what a faculty grasps is individuated by the proportion of the three guṇas, then a single sound admits endless variation in those proportions and requires endlessly many faculties; if instead the guṇas are taken as undifferentiated, one faculty would grasp sound and tangible alike, and the five faculties collapse into one—for on this theory there is no sound-class over and above the three guṇas. The root verse presses the same fork on whether the guṇas are other than what they compose:

གནས་སྐབས་དེར་འགྱུར་སྙིང་སྟོབས་སོགས། །མ་ཡིན་གཞན་མ་ཡིན་པའི་ཕྱིར། །གཞན་མིན་ཞེ་ན་འབྲས་མིན་པའམ། །གཞན་ཉིད་མིན་ན་གཟུང་བྱ་མིན། །

In that case, sattva and the like [the guṇas] would not [be the nature of sound and the like], because of not being other. If [you say] they are not other—then either [sound] is not [their] result, or—if they are not of a different nature—they would not be separately apprehensible. (pramanasamuccaya I)

A sub-school (the commentary calls it “ancient Kapila”) tried to escape by giving each own-nature its own supreme atom, with effects manifested by combination; Dignāga replies that combination of disparate kinds yields no transformation, so three natures cannot produce one result. The refutation ends on ground that matters for his own system: if the faculty’s operation is experienced by a later mental consciousness, that consciousness would be memory without prior experience, which is impossible; if the two arise simultaneously, neither is a means of valid cognition. Only admitting self-awareness (རང་རིག་) avoids the fault (pramanasamuccaya I, refutation of the Sāṃkhya position).

Against the inference theory (PS III), Dignāga takes the pradhāna proof apart fault by fault: pradhāna is not known by any valid cognition, so its existence cannot serve as a probandum at all; there is no genuine positive concomitance among individual things (the effects cited, such as joy, would prove the three guṇas as readily upon the puruṣa as upon external objects); the example offered (tree-bark and the like) has many causes, contradicting the thesis of one cause; and the example cannot even be stated, tree-bark being itself among the individual things to be proved. He then dismantles the vīta/avīta division itself: mere refutation of rivals never establishes one’s own position (ruling out “snow-melt” does not prove “rainfall”—the flood’s own character does that), and since opponents’ answers are endless, a proof defined as “what survives when the rivals are refuted” never terminates (pramanasamuccaya III, refutation of rival views; Fazun flags one stretch of this passage as still requiring study, so part of the detail here rests on a text this translator found difficult).

Dharmakīrti

Dharmakīrti’s main assault is on the single nature, in pramanavartika I.150–192, and it has two prongs.

The first is about universals in general and hits Sāṃkhya and Vaiśeṣika together. If the class (རིགས་) is one real thing—whether other than its instances (གསལ་བ་) or the same—then when a new instance arises, the universal that was seen in the old instances cannot be cognised in the new one. Two verses carry the whole dilemma:

།གང་དག་རིགས་ནི་གསལ་བ་ལས། །གཞན་ནམ་གཞན་མིན་ཡོད་ཉིད་དེའི། །སྔོན་མེད་གསལ་བ་རྣམས་ལ་ནི། །སྤྱི་ཡི་བློ་གྲོས་ཇི་ལྟར་ཡིན།

།གཅིག་ལ་ཡོད་པ་མཐོང་བ་ནི། །གཞན་ལ་མཐོང་བ་མི་སྲིད་ཕྱིར། །གཞན་ཉིད་མིན་ན་རྗེས་འགྲོ་མེད། །གཞན་ཉིད་ན་ཡང་རྟེན་མིན་ཕྱིར།

If the class exists as one—whether as other than the instances [as the Vaiśeṣika hold] or as the same [as the Sāṃkhya hold]—how could there be a cognition of the universal within newly arisen instances that did not exist before? For what is seen in the one [earlier instance] cannot [then] be seen in another [new one]. If [the universal] is not other [than the instances], it does not pervade [the new ones]; and even if other, it would not depend [on them]. (pramanavartika I.150–151)

Gorampa’s gloss presses the point into a dilemma over persistence: the universal does not abandon its earlier support, has no parts to distribute between old and new instances, and so—“alas, what a cause of great loss!”—cannot be in the new particulars at all (gorampa-pramanavartika ad loc.).

The second prong answers Kapila’s cooperation argument directly. If the many produce a single result because they share one nature, then the cooperating conditions are pointless—the seed alone, where that nature already resides, should suffice (pramanavartika, the reply immediately following I.163). And the single nature proves too much: if all entities are of one nature, everything should produce everything, whereas in fact only some things (rice seeds) create some results (rice sprouts), which is exactly the distinctness the Sāṃkhya denies (pramanavartika I.173–174 with Gorampa’s gloss). Dharmakīrti’s own answer to the cooperation problem is that the distinct attributes themselves are the producers—no shared nature needed (pramanavartika, within the exchange at I.163–180). At I.181–184 he adds that the same reasoning disposes of the Jains’ curd-and-camel doctrine, with the taunt that anyone holding curd and camels to be one nature should “run off and drink a camel when looking for curd” (Gorampa on PV I.182). A final exchange (I.187–192) blocks the school’s attempt to make bare “existence” the probandum (བསྒྲུབ་བྱ་) that proves the prime universal: existence established for both parties proves nothing, and its alleged features (permanence, oneness) have no concomitance with any example.

Chapter 2 carries three further engagements, sharper in tone. Against the satkāryavāda in its strongest form—all entities possess the potential for consciousness—the reply runs over three verses:

།གལ་ཏེ་དངོས་པོ་ཐམས་ཅད་ནི། །རྣམ་ཤེས་ནུས་ལྡན་ཕྱིར་འདོད་ན། །རྩ་སོགས་རྩེ་ན་གླང་ཆེན་བརྒྱ། །སྔོན་ཚེ་མ་མཐོང་ཡོད་དོ་ཞེས།

།དེ་སྐད་གྲངས་ཅན་ཕྱུགས་ལས་གཞན། །ངོ་ཚར་བཅས་པ་སུ་བརྗོད་རྩོལ། །རྒྱུ་ནི་ལན་བརྒྱར་རྣམ་ཕྱེ་ཡང་། །ངོ་བོ་གང་ཞིག་སྣང་གྱུར་པའི།

།ངོ་བོ་དེ་ནི་སྔར་མི་སྣང་། །ཇི་ལྟར་དེ་ནི་ཡོད་པར་འགྱུར།

[Objection—Sāṃkhya:] “[We hold that] all entities possess the potential for consciousness.” [Reply:] Then a hundred elephants would [already] exist, unseen, on the tip of a blade of grass—who, apart from the cattle-like Sāṃkhya, would labour to make so embarrassing a claim? Though you take the cause apart a hundred times, the nature that [supposedly] would become manifest does not appear at such [earlier] times; how, then, does it [already] exist? (pramanavartika II.165–167)

At II.17 the Sāṃkhya proof that the intellect and pleasure are mindless because impermanent is dismissed for lacking pervasion—the stated reason has no necessary connection with mindlessness (pramanavartika II.17, where Gorampa names Kapila). And in the antidote-to-self section, the Sāṃkhya path to liberation is turned against itself: the school says attachment arises from taking the self and prakṛti’s manifestations as one, and freedom from seeing them as distinct; Dharmakīrti replies that beings already know the self as distinct from the faculties—they strive to obtain other, superior things, and they cognise arising and ceasing—yet attachment persists, so seeing distinctness cannot be what liberates (pramanavartika II, the exchange following II.248, where Gorampa names the Sāṃkhya on the self and its objects).

There is also a methodological engagement in chapter 4, preserved in Gedun Drub’s commentary: the qualifier “self-seen” in the definition of inference-for-others is there specifically to cut off a Sāṃkhya manoeuvre—citing the Buddhist’s own scripture as a reason against the Buddhist (“intellect-pleasure is mindless, because it has origination and perishing, like the five subtle elements”) while holding no valid cognition of the reason oneself. Dharmakīrti’s reply distinguishes this from the legitimate use of an opponent’s admissions to expose a fault: the Buddhist who derives a contradiction from the Vaiśeṣika’s permanent universal aims only to break a wrong grasping, whereas the Sāṃkhya uses the borrowed reason to establish his own thesis, which an unestablished reason cannot do. The same chapter records the “crafty” Sāṃkhya inference that the eye and the rest are “for the use of another” (intending: for the use of the puruṣa), refuted because the intended probandum fails on any example (Gedun Drub’s commentary on PV IV).

Sakya Paṇḍita

The pramanayuktanidhi engages the Sāṃkhya principally in chapter 3, on universals—and the engagement is double-edged, because Sakya Paṇḍita’s real target stands nearer home.

Against the one-substance universal the root-text exchange is two lines long:

།ཁ་ཅིག་རྫས་གཅིག་ཡིན་ཞེས་ཟེར།

།རྫས་གཅིག་ན་ཡང་སྣང་རུང་མིན།

[Opponents—the Sāṃkhyas and the early Tibetans:] Others say they are one substance. [Reply:] Even were they one substance, the universal that ought to be perceptible [when the particular appears] is not observed. (pramanayuktanidhi chapter 3)

If a real universal stood on its particulars in a one-substance relation, the eye that sees the particular would see it; apart from the particulars not the slightest trace is found. This is nonapprehension of what should appear (སྣང་རུང་མ་དམིགས་), and the same reason disposes of the Vaiśeṣika’s separate-substance universal in the line before, so the two classic realisms fall together. The chapter then charges three consequences against any real universal held to be one substance with its particulars, each in a single line:

།སྤྱི་དང་བྱེ་བྲག་རྫས་གཅིག་ན། །ཡན་ལག་ཅན་ཡང་ཇི་ལྟར་དགག

།དབང་པོའང་རྟོག་བཅས་ཉིད་དུ་ཐལ།

།འགལ་བ་ཀུན་ཀྱང་རྫས་གཅིག་འགྱུར།

If universal and particular are one substance, on what grounds could [you] refute the [Vaiśeṣika] whole-with-parts (avayavin)? It would also follow that the sense [consciousnesses] are conceptual. And all contradictories whatsoever would become one substance. (pramanayuktanidhi chapter 3)

Refute their whole-with-parts (avayavin, ཡན་ལག་ཅན་) and your universal falls with it; if the eye sees the universal, eye-cognition has taken conception’s object, and conceptual cognition (རྟོག་པ་) and perception collapse into one another.

The concluding verse of the chapter places the school in a lineage Sakya Paṇḍita constructs deliberately, ending on exclusion (སེལ་བ་) as the position of the wise:

།ཐ་དད་སྤྱི་འདོད་འུག་ཕྲུག་པ། །རྫས་གཅིག་ཡིན་ཞེས་གྲངས་ཅན་པ། །གངས་ཅན་པ་རྣམས་དེ་རྗེས་འབྲང༌། །སེལ་བར་འདོད་པ་མཁས་པའི་ལུགས།

The Owl School (འུག་ཕྲུག་པ་, the Vaiśeṣikas) hold the universal to be a distinct substance; the Sāṃkhyas hold it to be one substance; the Snow Landers (early Tibetans) follow them; to hold it as exclusion is the tradition of the wise. (pramanayuktanidhi chapter 3, concluding verse)

Gorampa’s commentary on PY (rigs gter) explains the charge: Chapa Chökyi Sengé and the early Tibetan school never openly assert a real universal, but they accept that the universal stands in a same-substance relation with real entities, “which agrees substantially with Sāṃkhya and makes them followers of it”. The commentary is candid that the three-consequences attack only bites against Buddhists—a Sāṃkhya would simply reply that he never refuted the whole-with-parts and accepts it happily. The Indian school here is partly a stalking-horse for an intra-Tibetan debate (see realism-vs-antirealism).

Two smaller engagements: in chapter 6, the Sāṃkhya (holding the self sentient) and the Vaiśeṣika (holding it insentient) posit an “agent-connection” between the permanent self and its activities, which Sakya Paṇḍita dismisses in a line—a permanent self is not established even by conventional valid cognition, so there is nothing for the connection to relate (pramanayuktanidhi chapter 6). In chapter 7, on contradiction (འགལ་བ་), the Sāṃkhya answer to how fire expels water is that the weaker particle becomes hidden within the stronger, as stars still exist by day but cannot be seen. Sakya Paṇḍita takes it together with the two rival answers in one verse:

།ལ་ལ་ཉེ་བར་ཞི་བར་འདོད། །ཁ་ཅིག་ཕྲད་ནས་བསྒྱུར་བར་འདོད། །འགའ་ཞིག་ཕྲད་པ་མེད་པར་འདོད། །རྟག་སྨྲ་དེ་དག་འཁྲུལ་པ་ཡིན།

Some [Sāṃkhya] hold that [the weaker] subsides into latency; some [Vaiśeṣika] hold that on contact it is transformed; some [Hearer schools] hold that there is no contact at all. These proponents of permanence are all mistaken. (pramanayuktanidhi chapter 7)

Subsiding into latency fails because a really existing though unmanifest water-particle (རྡུལ་ཕྲན་) coexists with the fire, and coexistence is exactly what contradiction excludes: a thief hidden in a house still lives there with the owner (pramanayuktanidhi chapter 7).

What the debate turns on

Every thread above pulls on the same knot: whether “being of one nature” can do causal or cognitive work. The Sāṃkhya posits the single nature to explain real convergence—many causes, one result; many particulars, one word—and each refutation argues that the explanation destroys the difference it was meant to organise (everything produces everything; the eye sees what only thought can take; the hidden coexists with what should expel it). The Buddhist alternative is to explain convergence without a shared real nature, through exclusion: things utterly distinct in themselves are grouped by the negation of what they are not (apoha, problem-of-universals). One limit of the Buddhist position is recorded in the material itself: Dharmakīrti concedes at pramanavartika I.22 that mere nonapprehension cannot establish the nonexistence of the unmanifest result in the cause—prakṛti’s hiddenness is real dialectical shelter, and the question of when nonapprehension does refute the prime universal returns at I.205–211 without a fully settled answer.

Graph View

Backlinks