Pramana Wiki

Mīmāṃsā

30 min read

Mīmāṃsā (རྒྱལ་དཔྱོད་)

The Mīmāṃsā holds that the Vedas are true because nobody wrote them, and this single claim draws the longest unbroken polemic anywhere in our material: pramanavartika I.224–I.339, 116 of chapter 1’s 340 verses, which fill the last five of the numbered sections in our translation (§–32). (Verse numbers throughout follow Gorampa’s enumeration, which our translation of chapter 1 adopts.) dignaga had already given the school its own refutation section on direct perception and another on the number of means of valid cognition; sakya-pandita answers Jaimini by name on exclusion and lists the school among the seven opponents his treatise was written to crush. The question under all of it is whether authority can be manufactured by subtracting the author.

Overview

The school’s name means “critical examination”, and its business is the interpretation of the Vedas. Its founder is Jaimini, whom Gorampa’s translator places in the fourth century BCE (gorampa-pramanavartika, translator’s introduction); Wikipedia dates the Mīmāṃsā Sūtras attributed to him to c. 3rd–2nd century BCE, the one dating our material and the reference sources disagree over, and the disagreement is about the man rather than the text. The first surviving commentary is Śabara’s; the school’s two great sub-traditions were founded by Kumārila Bhaṭṭa in the seventh century (author of the Ślokavārttika) and Prabhākara around 700 CE, with Maṇḍana Miśra following Kumārila (Wikipedia). Pūrva Mīmāṃsā takes the ritual portions of the Veda as its subject; Uttara Mīmāṃsā, which is Vedānta, takes the Upaniṣads. The school held near-dominant influence over learned Hindu thought in the early medieval period, declined later, and today is eclipsed by Vedānta, surviving as a technical discipline inside Sanskrit hermeneutics and ritual and legal scholarship rather than as a school with its own adherents (Wikipedia; general web references). One oddity in our own material: a translator’s footnote to Gorampa describes the Mīmāṃsakas as “later Vedānta philosophers” who bring out the Vedas’ obscure meanings (gorampa-pramanavartika, note 143), which reverses the standard Pūrva/Uttara division just given.

Gorampa’s translator summarises the position Dharmakīrti attacks: the followers of Jaimini were orthodox adherents of the Vedas, took Vedic sounds to be eternal rather than words produced by humans, held that higher rebirth is gained by sacrificial rites, and denied that omniscience is possible, because mental afflictions belong to the nature of mind and cannot be removed. Their concern was epistemological; they paid little attention to deity worship or to a creator and were essentially nontheistic. Counting the sub-schools, six means of valid cognition can be ascribed to them: direct perception (མངོན་སུམ་), inference (རྗེས་དཔག་), scriptural testimony (ལུང་), postulation, negation and comparison (upamāna, ཉེར་འཇལ་) (gorampa-pramanavartika, translator’s introduction). Dreyfus gives the same list split by sub-school: “The Prabhakaras accept five means, whereas the [Bhāṭṭas] and the Vedantins accept six” (dreyfus-recognizing-reality, p. 294; the extraction’s OCR mangles this school’s name on almost every page, so square brackets inside the Dreyfus quotations here and below are editorial restorations), the sixth being non-perception, which the Prābhākaras reject (Wikipedia). Two of these extra means, postulation and negation, are the ones Dignāga takes apart by name.

The school’s ontology, on Dreyfus’s account, is realist and pluralist, built on the distinction between substance and properties: of the six or seven categories the Nyāya formulates, the Mīmāṃsā adopts five (substance, quality, motion, universal and nonexistence), dropping individuations and inherence, the latter “too obviously ad hoc” (dreyfus-recognizing-reality, p. 213). On universals its realism is milder than the Nyāya’s: “Kumarila’s view is an attempt to retain a realist view of universals without going to the extreme of radical separation between individuals and their universals” (p. 139). Dreyfus’s own first pass over the school treats it as marginal, noting that its views “will not require special attention” because they are so close to the Nyāya’s (p. 52). That judgement is confined to ontology. On language the two schools part company, and there the Mīmāṃsā takes “a more extreme but also a more consistent position” (p. 213), which is why Dreyfus later corrects the ranking: on apoha, “Dignaga’s main opponent here was not the Nyaya but the more orthodox [Mīmāṃsā]. Its leading proponent, Kumarila, formulated the most incisive critique of the apoha theory” (p. 213).

Our sources name the school in five different ways, and it is worth keeping them apart. རྒྱལ་དཔྱོད་ is the glossary’s headword. སྤྱོད་པ་བ་ is what the Tibetan of the pramanasamuccaya uses, and Fazun’s Chinese renders the same school 觀行派, glossed as the Aupavarṣa (pramanasamuccaya I). རྒྱལ་དཔོག་པ་, “followers of Jaimini”, is the auto-commentary’s term in the pramanayuktanidhi. རིག་བྱེད་པ་, “followers of the Veda”, appears in the same auto-commentary where the point at issue is Vedic rather than specifically Jaiminīya. A fifth, ཐོས་སྒྲོག་པ་ (“proclaimers of what is heard/recited”), appears in the Pramāṇayuktanidhi colophon and is rendered “Vedic reciters (Mīmāṃsā)” but marked [translation uncertain]: the Chinese edition reads it as 声闻, Śrāvaka, which would make it a Buddhist school in a list of non-Buddhist opponents (translations/pramanayuktanidhi/translation-notes.md, chapter 11).

Core doctrines

The Vedas have no author

The thesis in Dharmakīrti’s statement of it:

།ཚིག་རྣམས་ལོག་པ་ཉིད་ཀྱི་རྒྱུ། །ཉེས་རྣམས་སྐྱེས་བུ་ལ་བརྟེན་ཕྱིར། །སྐྱེས་བུས་མ་བྱས་བདེན་དོན་ཅན། །ཞེས་ནི་ལ་ལ་རྗོད་པར་བྱེད།

[Objection—Mīmāṃsā:] Since the causes of erroneous words are faults [such as desire] that depend on a person, [the Vedas,] not being produced by a person, have the reality of truth—so some assert. (pramanavartika I.224)

Gorampa sets it out as a formal proof: the Vedas have the reality of truth, because they teach something not produced by a person; the pervasion (ཁྱབ་པ་) holds because the causes of erroneous words are faults such as desire, and these arise from persons (gorampa-pramanavartika, on PV I.224). The point of authorlessness (apauruṣeyatva, སྐྱེས་བུས་མ་བྱས་པ་—literally “not made by a person”, སྐྱེས་བུ་) is not modesty about origins. It is a guarantee: a text with no author has no author’s faults, and there is no other way to certify a body of teaching whose subject matter lies beyond anyone’s power to check.

Dreyfus sets out what rides on this. The Vedas are the ultimate source of knowledge about dharma, their topics lie beyond human reasoning, and testing them against other texts would open an infinite regress. Either humans cannot know dharma at all, or they know it through revelation, in which case the Vedas must be intrinsically valid (dreyfus-recognizing-reality, p. 214). The Nyāya secures the same conclusion more cheaply by giving the Vedas a divine author; the Mīmāṃsā refuses even that, and Dreyfus notes the Buddhists reject the Nyāya move too and “deny the validity of the Vedas all together” (p. 215).

If the Vedas had to be construed by human convention, human fallibility would re-enter by that door. So the Mīmāṃsā denies that language is conventional at all: the relationship between Vedic words and their referents is natural, not made by any person (pramanavartika I.227 with gorampa-pramanavartika ad loc.). Dreyfus fills in the doctrine: words are not their audible realisations but permanent ideal articulations of meaning, so that the word cow is a constant articulation of the universal cowness whose phonic realisation may vary; the link between word and universal is essential (svabhāvika), and words “have an inherent capacity (sakti) to signify the universals to which they are internally related”. Hence “language is naturally meaningful” and fully transparent when framed in Sanskrit (dreyfus-recognizing-reality, p. 214). This is the doctrine that made Kumārila the leading opponent of apoha: a theory on which words work by excluding what they do not mean is a direct denial that words mean anything by nature.

Sound and speech are permanent

The corollary the Buddhists attack hardest. Vedic sounds are eternal, not produced; a recitation makes an already existing speech manifest rather than creating it. Gorampa gives the reason, and the terms at issue are sound (སྒྲ་) and verbal expression (བརྗོད་བྱེད་): “The Mīmāṃsaka do not assert mere sound to be verbal expression because it has this and that fault of having former and later parts, and those former and later parts cannot possibly become a whole. Therefore they say that it is the partless words and speech made manifest by the sequence of mere sounds that is the verbal expression” (on PV I.258). This partless-speech doctrine resembles the Grammarians’ sphoṭa, and the chapter-1 attribution map records the decision to follow Gorampa here rather than the doxographic expectation: the position is Mīmāṃsā, not Vaiyākaraṇa (translations/pramanavartika/translation-notes.md, chapter 1 opponent map at 1.247–1.268). The same commitment to permanence generates the school’s theory of destruction, on which a thing’s perishing is a separate entity produced by an external cause such as fire, rather than the thing’s own momentary nature (gorampa-pramanavartika, on PV I.280), and it survives into Sakya Paṇḍita’s material as the subject of the standard proof of momentariness.

No person sees beyond the senses

Because dharma is knowable only through revelation, the Mīmāṃsā must deny that anyone could know it directly. The reasons Dharmakīrti reports them giving: a supposed exceptional person has cognition, senses, speech and is a person, and everyone with those features is ordinary (pramanavartika I.311). Gorampa gives the argument its polemical form: “If you assert someone is a special person, then from the reason of that person having cognition and so on, it can be proven that they are not special” (gorampa-pramanavartika, on PV I.311). The same denial reappears in pramanavartika chapter 2 as an objection to the Buddha’s validity—“[To be] valid [a being] would have to know hidden objects (parokṣa, ལྐོག་གྱུར་); but since there is no means of accomplishing that, no one strives [for it]” (II.29)—and again as the claim that the Buddha’s faults are not exhausted, since he speaks and is a person (II.143). Gorampa’s note explains the connection between the two: for the Mīmāṃsaka, speech from a person arises from desire, so a speaking person has not abandoned all faults, and “the pure and true speech is that of the Vedas, which were not authored by a person” (gorampa-pramanavartika, note 244).

Validity from within

Dreyfus records one further commitment, which does not surface in our root-text material but shapes the dispute over valid-cognition: “Kumarila claims that cognitions are internally valid. They are presumed valid in normal circumstances”, and “Dharmakirti rejects the [Mīmāṃsā] idea of intrinsic or internal validity” (dreyfus-recognizing-reality, p. 340). A text that certifies itself needs a theory on which cognitions certify themselves.

The refutations

Dignāga

The pramanasamuccaya gives the school a titled refutation section in two of its six chapters, on direct perception and on the means of valid cognition. In both, Dignāga fastens on a definition.

The Mīmāṃsā definition of direct perception, in our translation of the Tibetan:

སྤྱོད་པ་བ་རྣམས་ནི་ཡོད་པ་དང་ཡང་དག་པར་སྦྱར་ནས་སྐྱེས་བུའི་དབང་པོའི་བློ་སྐྱེས་པ་དེ་ནི་མངོན་སུམ་མོ་ཞེས་ཟེར་རོ།

The Mīmāṃsakas assert: The cognition of a person’s sense faculties that arises from genuine contact with an existent—that is direct perception. (pramanasamuccaya I)

Fazun’s Chinese renders the same definition: “The sense-cognition of a person, arising from a correct conjunction with the existent, is perception.” Dignāga works through the definition word by word. “Existent” does no work, since “conjunction” already implies that something is there; and if it was inserted to exclude mirages, that fails too, because the sun-heat and the elements that produce a mirage are themselves existent and do conjoin with the eye. “Conjunction” over-generates: atoms conjoin with the faculty as well as objects do, and collyrium and foot-ointment “bless” or clarify the eye as much as any object, so every one of those would count as perception. If the faculty must be thoroughly conjoined with its object to grasp it, then sound and colour refute the account, since the eye and ear demonstrably take objects at a distance and objects larger than themselves. And the appeal to a cognition-producing aggregate of object, faculty, mind, self, conjunction and activity is idle: if the object alone is the instrument, the clause about cognition arising is superfluous, and if the self must transform in order to produce cognition then the self is impermanent, which the school does not want, while if it does not transform it cannot be an instrument at all (pramanasamuccaya I, refutation of the Mīmāṃsā position on perception). One step of the exposition rests on a reading Fazun himself hedges, glossing the Tibetan by “excellence” and “fine display” without settling it [unverified].

In chapter 2 the target is the school’s count of the means of valid cognition. The Mīmāṃsakas say that inference and the rest are not separately stated in their sūtras because they are all preceded by perception. Dignāga’s reply: then memory, desire, anger and striving are also preceded by perception and should be counted as means of valid cognition too. He then refuses the two candidates the school adds beyond perception and inference. Postulation (“by-object cognition”) divides into the determinate case, which is simply inference—knowing from one boiled grain that the rest in the pot are boiled—and the indeterminate case, which is a cause of doubt and no valid cognition at all. Negation (“non-fact cognition”) is illustrated by “the dark-coloured one is not in the house”, said to establish that he is outside; Dignāga answers that the inference to “outside” runs through the implication of the word house, not through any bare cognition of absence, since if there were neither house nor outside one could say only that he is not (pramanasamuccaya II, refutation of the Mīmāṃsā). This is the same reduction he applies to the Nyāya’s four means, and he says so: the earlier refutations of inference and the rest dispose of this school as well, since what little it states beyond them is Vaiśeṣika.

The apoha chapter opens against those who make verbal cognition (śabda-pramāṇa) a third means, and answers that a word reaches its object by the same “not arising without [it]” relation that a logical reason does, so verbal cognition falls under inference (pramanasamuccaya V). Fazun’s exposition does not name the proponents at that point; the wiki commentary identifies them as the Mīmāṃsakas and Naiyāyikas (PS V commentary), and Dreyfus independently makes Kumārila the principal opponent of the theory (dreyfus-recognizing-reality, p. 213). Dreyfus also reports the three counterarguments Kumārila brought against apoha in the Ślokavārttika, which are not preserved in our own material: that the theory is “(1) counterintuitive, (2) circular, and (3) useless”—counterintuitive because nothing is negative about this is a cow; circular because understanding x by eliminating non-x presupposes knowing non-x, whose elimination of non-non-x is just x again; and useless because Dignāga’s own theory already has an intuitional idea (pratibhā) doing the semantic work, so the negative element is idle (p. 215).

Dharmakīrti

Dharmakīrti’s answer is the longest sustained refutation of any single school in our material—longer than the Cārvāka thread that occupies PV II.35–II.119—and Gorampa organises it under three heads: refuting the pervasion of the root reason, refuting the root reason itself (from I.239), and refuting the thesis (from I.317) (gorampa-pramanavartika, on PV I.224 ff.).

The inference run backwards. The first move is symmetry, and the reply verse is built line for line on the thesis verse it answers:

།ཚིག་རྣམས་བདེན་པ་ཉིད་ཀྱི་རྒྱུ། །ཡོན་ཏན་སྐྱེས་བུ་ལ་བརྟེན་ཕྱིར། །སྐྱེས་བུས་མ་བྱས་ལོག་པའི་དོན། །ཅིས་མིན་ཞེས་ནི་གཞན་དག་བརྗོད།

[Reply:] Since the causes of true words are qualities [compassion, wisdom] that depend on a person, why would [the Vedas,] not produced by a person, not have the reality of error?—so others [the Buddhists] reply. (pramanavartika I.225)

Nothing about the argument’s form favours one conclusion over the other, and the reply costs Dharmakīrti nothing, since it is his own view that true speech has a personal source. The second move is that the school’s premise defeats its own conclusion: understanding meaning requires convention, convention depends on persons, “then even if the words were not produced by a person, the possibility of [their] error remains” (I.226).

The natural relation. The school escapes the convention argument by denying that convention is involved at all: word and meaning are related by nature. Everything turns on this, since it is the one Mīmāṃsā thesis that contradicts Dharmakīrti’s own account of how words work. Dharmakīrti’s consequence is immediate: then meaning would be understood by anyone who had learned no convention at all; and if the natural relation only becomes effective once convention makes it clear, the natural relation is doing nothing, since where convention is applied there is understanding and where it is not there is none, natural relation or no (I.227). Then the dilemma. If a Vedic word is naturally fixed to one meaning, no second meaning could ever be got from it. If it is naturally connected with many, it is connected with all, so “If you desire higher realms, give burnt offerings to the fire” would as readily reveal contradictory things, “eat dog-meat” among them (pramanavartika I.228 with gorampa-pramanavartika ad loc., which supplies the burnt-offering example). Either way the point of asserting authorlessness collapses, since it was asserted to guarantee that the text yields only what is wanted (I.229).

Then the memorable image:

།སྐྱེས་བུས་ལེགས་པར་མ་བྱས་པ། །རྣམ་པ་ཀུན་ཏུ་དོན་མེད་འགྱུར། །ལེགས་བྱས་ཁས་ལེན་ན་འདི་ནི། །དངོས་སུ་གླང་ཆེན་ཁྲུས་འདྲར་འགྱུར།

Not being well-composed by a person, [the Vedas] would have no meaning at all; and if you concede they were [well-]composed [by a person], this is just like washing an elephant [—trading one stain for another]. (pramanavartika I.230)

Gorampa spells the analogy out both ways: to remove the dirt of error you assert the Vedas were not produced by a person, and “they become smeared by the wetness of not being able to reveal any meaning”, and to remove that you assert composition, and they are smeared with the dirt of error again. The translator adds what an Indian reader would have supplied: an elephant washing itself covers itself in water to get the dust off, then rolls in dust to dry (gorampa-pramanavartika, on PV I.230 and note 145).

The relation itself then fails on analysis. Its bearers are impermanent, so no permanent relation can hold of them (I.231). If the relation is an entity distinct from sound and meaning, cognising it would be a third cognition; if not distinct, it just is the sound and the meaning, and there is no third option (pramanavartika I.236). And a really existing relation cannot be a relation of anything else, since what is substantially existent does not depend on another (I.237).

The evidence. From I.239 Dharmakīrti turns on the school’s grounds for believing the Vedas authorless. The first is that nobody remembers who wrote them, and it draws the sharpest line in the section:

།བྱེད་པ་པོ་རྣམས་མི་དྲན་ཕྱིར། །སྐྱེས་བུས་མ་བྱས་ཀྱང་འདོད་ལོ། །འདི་ལའང་རྗེས་སུ་བརྗོད་པ་ཡོད། །དེས་ན་ངན་པའི་མུན་པས་ཁྱབ།

[Mīmāṃsā argue:] because the authors are not remembered, they assert [the Vedas] were not made by a person—and even to such [a non-reason] they hold so! Therefore [their world] is pervaded by a wretched darkness. (pramanavartika I.239)

Gorampa gives two answers behind the insult: Buddhists do remember authors, naming Nirbhartsanā, Vāmanadeva and Viṣvamitra from Dharmakīrti’s own auto-commentary as authors of Vedic mantras, and the reason has no pervasion in any case, being the “old well” pattern of presumption—there was water in the well our ancestors used, so there must be water now (gorampa-pramanavartika, on PV I.239 with notes 149 and 150). The second ground is that reciters presuppose earlier reciters without beginning. Dharmakīrti replies that authored texts are transmitted the same way and nobody infers their beginninglessness (I.240–I.241); that even a beginningless text is not thereby free of persons, or children’s play with stones would count as uncreated (I.244); and, in the argument that does the most damage, that the customs of barbarians and the doctrines of the Cārvāka materialists are as beginningless as anything else and would come out authorless and therefore true on the same reasoning (I.245). Truth is not conferred by not being able to find the author.

Syllables, speech, and destruction. From I.247 the target is what exactly is supposed to be uncreated. If it is the syllables, they are no different from ordinary syllables and proving them uncreated achieves nothing; if it is a speech distinct from the syllables, no such thing is apprehended (I.247). A speech with parts either has meaningless parts, in which case the whole is a mere imputation, or meaningful ones, in which case a single part would convey the whole meaning; and a partless speech could not be understood in sequence at all, having no parts to be taken in order (I.248–I.250). If speech is impermanent it is effort-produced and so made by persons; if permanent it would be heard perpetually, and the plea that a cooperating cause is missing concedes dependence, which contradicts permanence (I.251–I.252). The section on destruction (I.269–I.283) attacks the same permanence from the side of ontology: if fire produces destruction as a separate entity, the wood should still be visible, since it has not itself disintegrated, and the destruction, being itself produced, would have its own destruction, after which the wood should reappear (I.269–I.270, I.273). The school’s fallback, that existence is not pervaded by disintegration because some existents are uncreated, is answered by the observation that distinct natures follow from distinct causes, and there are no distinct causes here (I.281–I.282). The last of these arguments is the school’s “flawless” proof—created speech is seen to be untrue, therefore uncreated words are true—which Dharmakīrti rejects for leaving the reason’s absence in the dissimilar class unshown, so that doubt remains whether uncreated words consort with untruth after all (I.287–I.288).

Mantras, caste, and exceptional persons. The mantra section (I.292–I.307) is where the argument reaches practice. If mantras work because persons apply the convention, they depend on persons; if they work by the power of a permanent entity, then since a permanent thing admits no differences, any arrangement of the syllables would work as well (I.292). The results would arrive perpetually, or, if they wait on ritual, the mantra is affected by conditions and is not permanent (I.293). The consequence Dharmakīrti draws next is the one with social teeth: a permanent mantra’s power “would produce results for everyone” and cannot depend on differences of agent, so a brahman’s recitation and an outcaste’s must work alike—and in fact “[a mantra’s] power is not determined by the features of agents… for those [caste-distinctions] arise merely from convention” (I.294–I.295). Mental recitation and whispered recitation would also be purposeless on the school’s account, since audible sound is the domain of the ear (I.298). His own account is the one he needs anyway: syllables arise from the speaker’s volitional consciousness, each bearing the feature of the one before, so sequence is a causal sequence in persons, which is why rasa and sara have different effects (I.302–I.305).

The same material then establishes what the school most wants to deny. Whoever has the knowledge and power to form the sequence of a mantra “is established as exceptional, since those [the requisite knowledge and power] do not exist in others” (I.308), and the reasons offered against exceptional persons—having cognition, senses, speech, being a person—are reasons with a remainder and prove nothing (I.311). Gorampa states the purpose of the mantra section flatly: it exists “to repudiate all the Mīmāṃsaka expressions of proof that persons have no ability to see things beyond the senses” (gorampa-pramanavartika, on PV I.311). The thread runs straight into chapter 2. There the Mīmāṃsaka objection that a valid being would have to know hidden objects, and that there is no means of becoming such, is answered by narrowing what validity requires: “Hence it is the wisdom to be accomplished… that should be examined. Knowing the number of insects [in the world] is of no use to us whatsoever” (pramanavartika II.31). The two verses that follow fix the standard and then mock the alternative:

།བླང་དང་དོར་བྱའི་དེ་ཉིད་ནི། །ཐབས་དང་བཅས་པ་རིག་མཛད་པ། །གང་དེ་ཚད་མ་ཉིད་འདོད་ཀྱི། །ཐམས་ཅད་རིག་མཛད་མ་ཡིན་ནོ།

།རིང་པོ་མཐོང་བའམ་མིན་ཡང་རུང་། །འདོད་པའི་དེ་ཉིད་མཐོང་བ་ཡིན། །གལ་ཏེ་རིང་མཐོང་ཚད་ཡིན་ན། །ཚུར་ཤོག་བྱ་རྒོད་བསྟེན་པར་གྱིས།

One who knows the reality of what is to be adopted and what discarded, together with the means [to each], is the one we assert to be valid—not one who knows all things. Whether or not he sees far off, he is one who sees the reality that is sought. If seeing far made one valid—come, then, and take to worshipping vultures! (pramanavartika II.32–33)

At II.143 the school argues that the Buddha’s faults are not exhausted because he speaks; Dharmakīrti answers that the reason founders on doubt about the counter-pervasion, and then that the three possible grounds—the faults are permanent, there is no method, or the method is unknown—all fail, since the faults have causes and cease when their antidotes are trained (II.143–II.145). The whole positive argument for the Buddha as a valid being (see buddha-as-pramana) is built on ground cleared from this school.

Jaimini’s access. The founder is the next target. “The sounds [of the Vedas] do not declare, ‘I [mean] this is the meaning, and this is not’; so this meaning is [necessarily] something construed by persons—and they too are possessed of desire and the rest” (I.312). If construal is unavoidable, the claim to know the definitive meaning is a personal claim, and then: “[Do] that one [Jaimini] alone know that reality, and not others? Who drew this distinction? Likewise, why do you not accept that some [others], being equally persons, [too] possess [such] knowledge?” (I.313). The school cannot have it both ways—it denies exceptional persons and then makes an exception of its own founder. Gorampa’s third and last branch opens on the same point at I.317: being possessed of desire yourself, you do not learn the Veda’s meaning from others who are equally so possessed, and the Vedas do not declare it, so from what do you know it? To the reply that worldly usage settles the meanings, Dharmakīrti answers that Vedic expressions are not used in worldly senses, citing “high rebirth” and “goddess” as terms with special referents (I.319–I.320); and to the reply that the Vedas disclosed their meaning to Jaimini, that one deceiver says “‘This is the path,’ the tree-stump declared” and another says it of himself, and there is no difference between them (I.325).

The last defence is inductive: since one part of the Vedas is true, as “fire dispels cold” is true, the rest is true as well, and so “the others—that is, the elders of the Mīmāṃsaka—have said” (pramanavartika I.330 with gorampa-pramanavartika ad loc.). Dharmakīrti’s answer lists what the Vedas actually propound—permanent persons as agents, permanent entities, the supersensory apprehended by the senses, things at odds with their own cause, duration and disintegration—and then draws the conclusion:

།འགལ་བ་ཀུན་ཏུ་མ་བསལ་ཞིང་། །བསྟན་བཅོས་དོན་ཡང་མ་བསྟན་ལ། །དེ་བདེན་དོན་བཅས་ཁས་ལེན་པ། །གཡེམ་མ་མུ་ཅོར་གྱིས་རྒྱལ་ལོ།

Without dispelling those contradictions at all, and without [even] showing [a treatise’s required] meaning, to hold [the Vedas] possessed of truth is [a victory only] like a wanton woman’s foolish boast of conquest. (pramanavartika I.334)

Gorampa: “The Mīmāṃsaka are like a lustful woman foolishly claiming victory over her master”, because they dispel none of the contradictions and never show the three features a treatise must have (gorampa-pramanavartika, on PV I.334). The translator records that the Sanskrit and Tibetan diverge on the verb. The Sanskrit jayet means “surpass”, so the analogy has the school’s claims “surpass in absurdity even that of an unchaste woman”; the Tibetan རྒྱལ་བ་ means “victory”, the woman “foolishly believing she has succeeded in convincing her husband” (gorampa-pramanavartika, note 162) [translation uncertain].

The apoha objection. Kumārila appears once by name outside the Vedic sections, at pramanavartika I.114, with the circularity objection Dreyfus reports from the Ślokavārttika: if “tree” is apprehended by excluding non-tree, and non-tree by excluding tree, the two are mutually dependent, and where neither is prior neither is grasped, so no convention can be framed by exclusion (pramanavartika I.114, where the translation names “Kumārila and others (Mīmāṃsā)”). Dharmakīrti’s first move is to hand the dilemma back: in framing your own convention, is non-tree precluded or not? If it is, then on your own account, with no prior apprehension of tree, you could not have understood non-tree either (I.115). His positive answer follows at I.119–I.121: a cognition that takes things as one is produced by some objects and not by others, and by that very fact sorts its causes from its non-causes before any convention is applied; the convention is then applied to that already-sorted difference, and the cognition is mistaken only in taking earlier and later as a single entity.

A Vedic example without a school. Gedun Drub’s commentary on pramanavartika chapter 4 works through the inference that “the ear-cognition cognizing Veda-sound is preceded by an impermanent self-cause-sound”, meets the defence that Veda-sound is permanent, and answers that a permanent sound would be heard before any effort was made and that effort could have no work to do (Gedun Drub’s commentary on PV IV). The commentary does not name the opponent; the thesis that Vedic sound is permanent is the Mīmāṃsā’s throughout our chapter-1 material, but the identification is not made in that file.

Sakya Paṇḍita

The pramanayuktanidhi names the school in four of its eleven chapters, and the engagements are shorter and more surgical than Dharmakīrti’s.

In chapter 4, on proof and exclusion, Jaimini’s circularity objection returns almost word for word, seven centuries after Kumārila framed it, and Sakya Paṇḍita’s reply—the same move as Dharmakīrti’s—is compressed into the answering stanza, which turns on direct contradiction (དངོས་འགལ་) and on what would become of convention (ཐ་སྙད་) (pramanayuktanidhi chapter 4, where the auto-commentary names རྒྱལ་དཔོག་པ་):

།དངོས་འགལ་གཅོད་པ་གཞན་སེལ་སྒྲ། །དེ་ནི་ཕན་ཚུན་བརྟེན་པའི་ཕྱིར། །གཅིག་མ་གྲུབ་པས་གཉིས་ཀ་མེད། །དེས་ན་གཞན་སེལ་མི་སྲིད་ལོ།

།ཁྱེད་ཀྱང་དངོས་འགལ་མི་གཅོད་ན། །དེ་ཡི་ངོ་བོ་རྟོགས་མི་སྲིད། །གཅོད་ན་འགལ་ཟླ་མི་རྟོགས་པས། །ཐ་སྙད་ཐམས་ཅད་མེད་པར་འགྱུར།

[Objection—Jaimini (Mīmāṃsā):] The exclusion-of-other word that cuts off the directly contradictory mutually depends [on what it excludes]; if one [side] is unestablished, neither is—hence exclusion-of-other is impossible. [Reply:] You yourselves: if you do not cut off the directly contradictory, you cannot cognise the [thing’s] own nature; if you do, then since the contradictory counterpart is [equally] not cognised, all conventions become void. (pramanayuktanidhi chapter 4)

The positive answer comes a few verses later, and is the plainest statement of apoha in the chapter—the agreed-upon convention (བརྡ་) attaches to a kind (རིགས་), and the kind is exclusion-of-other (གཞན་སེལ་):

།ཡལ་ག་ལོ་མ་ལྡན་མཐོང་ནས། །དེ་ལ་ཤིང་གི་བརྡ་བྱེད་པས། །བརྡ་དེ་རིགས་ལ་སྦྱོར་བར་བྱེད། །རིགས་ནི་གཞན་སེལ་ལས་གཞན་མིན།

[Self position:] On seeing something with branches and leaves, we make for it the convention “tree”; that convention is then applied to the kind, and the kind is nothing other than exclusion-of-other. (pramanayuktanidhi chapter 4)

In chapter 6, on logical connection, the school appears twice. The subject of the standard proof of momentariness—“Whatever exists perishes, as with a vase; sound too exists”—is the Vedic sound the Mīmāṃsā holds permanent, so the nature-reason (རང་བཞིན་གྱི་རྟགས་) is aimed at them (translations/pramanayuktanidhi/translation-notes.md, chapter 6, citing the auto-commentary’s རིག་བྱེད་ཀྱི་སྒྲ་). And the three consequences pressed against Sakya Paṇḍita’s causeless-perishing thesis—that directly contradictory properties would become one substance, that perishing is untenable as something other, that the obstructive and non-obstructive would be on a par—are assigned by the auto-commentary to Vedic and Jaiminīya opponents, and dismissed in a half-line: “By these three consequences [our position] is not refuted here” (pramanayuktanidhi chapter 6).

In chapter 9, on direct perception, the school heads a survey of four rival accounts of valid cognition and its result (ཚད་འབྲས་), all four packed into one verse:

།ལྟ་དང་ངེས་པ་ཚད་འབྲས་དང༌། །ཕྲད་དང་དོན་རྟོགས་ཚད་འབྲས་འདོད། །ཁྱད་པར་ཁྱད་པར་ཅན་དེར་སྨྲ། །ལ་ལ་དབང་ཤེས་ཚད་འབྲས་འདོད།

[Some—the Mīmāṃsakas—] hold seeing and ascertaining to be valid cognition and result; [some—the Naiyāyikas—] hold contact and cognizing the object; [some—the Vaiśeṣikas—] speak of the distinction and the distinction-bearer (ཁྱད་པར / ཁྱད་པར་ཅན); and some—[the Vaibhāṣikas]—hold the sense faculty and sense cognition to be valid cognition and result. (pramanayuktanidhi chapter 9)

Khenpo Sodargye’s exposition gives the position its content—the non-conceptual perception of a pillar is the valid cognition, the subsequent conceptual ascertainment that the pillar is red or impermanent is the result—and the objection: a non-conceptual cause producing a conceptual result is an arising from a different kind. The deeper fault is common to all four accounts, the Mīmāṃsā’s included: each makes valid cognition and result a producer and a produced, and then when the valid cognition is present the result is absent and when the result is present the valid cognition is gone (sodargye-pramanayuktanidhi, on PYN chapter 9).

The colophon lists the school among the seven opponents the treatise was written against, alongside Kapila, Akṣapāda, the Owl’s Son, the Sky-Clad Jains, the Cārvākas and the Snowland masters (pramanayuktanidhi chapter 11)—with the caveat about ཐོས་སྒྲོག་པ་ noted in the Overview above.

Where the sources disagree. Two attributions in the Pramāṇayuktanidhi material are contested between the auto-commentary and Khenpo Sodargye, and the translation follows the auto-commentary in both. At chapter 5, the thesis that the specifically characterised phenomenon of the external object is what a word expresses is assigned by the auto-commentary to the Grammarians and the Vātsīputrīyas (བརྡ་སྤྲོད་པ་དང་གནས་མ་བུའི་སྡེ་པ་), while Sodargye names the Mīmāṃsakas, together with the Vātsīputrīyas within Buddhism (sodargye-pramanayuktanidhi, on PYN chapter 5; translations/pramanayuktanidhi/translation-notes.md, chapter 5, which records that an earlier draft of the note followed Sodargye and was corrected). At chapter 6, the agency-connection between a permanent self and its activities is assigned by the auto-commentary to Sāṃkhya and Vaiśeṣika, while Sodargye says “Mīmāṃsaka et al.”; conversely the three consequences reported above are Vedic and Jaiminīya for the auto-commentary and Naiyāyika or Vaiśeṣika for Sodargye (same notes, chapter 6). Nothing in our material settles either disagreement, and this page follows the translation’s choice rather than resolving it [unverified].

What the debate turns on

Every argument above bears on one question: whether removing the author from a text can put truth into it. The Mīmāṃsā answer depends on a chain—no author, so no author’s faults; no convention, so no human construal; permanent sound, so no production—and each Buddhist refutation attacks the chain at a different link while granting the others for the sake of argument. What makes the debate more than a demolition is that Dharmakīrti needs an answer of his own, because he also holds a scripture. His answer is to relocate scriptural authority rather than abolish it: trustworthy words are so called not because they are scripture but because they have three examinable features, being coherent throughout, setting out methods that accord with the aim, and conveying a purpose for beings; being nondeceptive about the four truths, which reasoning can check, they are inferentially valid about what reasoning cannot reach, and only there, “because there is no opportunity” for entity-based valid cognition in an ordinary mind (pramanavartika I.214–I.217 with gorampa-pramanavartika ad loc.; see scriptural-authority). Authority is earned by passing a test on the checkable part, which is exactly what the Mīmāṃsā claims is unavailable in principle.

The section does not close cleanly, and the last verses say so. Because scripture and its objects stand in no invariable relation, nothing is ascertained from scripture—and Dharmakīrti immediately draws the cost: “even reversing that [scripture as an authority], the non-existence of an entity is not established [by it]; and so a resulting, definite non-existence is not gained by [mere] non-observation” (pramanavartika I.338–I.339, with Gorampa’s gloss that if a doubtful entity did exist there would be no certainty an ordinary valid cognition would catch it). The tool that dismantles the Vedas cannot then be used to prove that what the Vedas describe is not there. The chapter’s final verse then turns aside from the Mīmāṃsā altogether to belittle the Naiyāyika, who holds the Vedas authoritative on the different ground that God composed them: holding the Vedas valid, propounding that permanent beings can act, seeking Dharma by washing, proclaiming caste, and mortifying the body “are the five signs of the foolish, of corrupted discernment” (I.340, where Gorampa names the Naiyāyika, not the Mīmāṃsā).

Graph View

Backlinks