Vaiyākaraṇa (བརྡ་སྤྲོད་པ་)
The Grammarians are the narrowest of the non-Buddhist opponents in these texts. Nine verses of pramanavartika chapter 1 and three chapters of the Pramāṇayuktanidhi answer them, almost always on one question: is the connection between a word and the thing it names real, or set by whoever is speaking? No self, creator, ritual or path to liberation is at stake. What is at stake is a claim about language that the Buddhist theory of convention was built to replace, and Sakya Paṇḍita’s account of common renown (གྲགས་པ་, prasiddha) depends on having replaced it. There is a twist, taken up at the end of the page: on Dreyfus’s reading, the positive theory of meaning Dignāga put in its place was itself partly borrowed from Bhartṛhari, so this school is both opponent and source.
Overview
Vyākaraṇa is grammar, the analysis of Sanskrit. It is one of the six Vedāṅgas, the auxiliary sciences of Vedic study, where its task was to keep the transmission of the Veda from decaying (Wikipedia). Its foundation is Pāṇini’s Aṣṭādhyāyī, some four thousand sūtras in eight chapters, usually dated to the fifth century BCE; Dreyfus places Pāṇini more loosely “between sixth and third centuries B.C.E.” (dreyfus-recognizing-reality, p. 207). Grammar became philosophy with Kātyāyana’s Vārttika and Patañjali’s Mahābhāṣya in the first century BCE or CE. Dreyfus singles out the question Patañjali asked there, what the object of a word (padārtha) is, as the question from which the apoha theory eventually grew (p. 207). The line of philosophical grammarians runs on to Bhartṛhari, whom Dreyfus dates to the fifth century CE (p. 207), and whose Vākyapadīya is the school’s main philosophical work (Wikipedia).
Two answers to Patañjali’s question divided the tradition, associated with Vyāḍi and Vājapyāyana. Vyāḍi’s denotationism holds that a term picks out individual substances. The word cow refers to an individual cow, and does so “not directly or through the intermediary of a universal but by eliminating the contradictory of cow; that is, noncow”, so that white in white cow “does not indicate a connection with whiteness but denies all other colors to the cow” (dreyfus-recognizing-reality, p. 207). Dreyfus is careful to add that this is linguistics, not yet philosophy: “Vyadi’s object is not the status of universals. He is a linguist, not a philosopher” (p. 207). Vājapyāyana’s connotationism holds that language signifies first through universal attributes and only secondarily picks out things.
Bhartṛhari takes the connotationist side and goes further. The sentence, not the word, is the primary bearer of meaning; a word on its own has meaning only by abstraction from sentences; and what a hearer grasps is a single ideational meaning, pratibhā (Dreyfus gives the Tibetan as སྣང་བ་), taken in at a stroke rather than assembled from parts. “[T]he object of a word can be only a universal, for that is the object we have in mind when we speak” (p. 208). This is the sphoṭa doctrine: real language is an inner articulation of meaning that spoken sounds reveal rather than create. Dreyfus stresses that this does not make meaning private. The sphoṭa is “an intersubjectively available sphere of meaning that is present in all those who participate in a linguistic community” (p. 212). The school also differs from its neighbours on what knowledge is most basic: for most Indian philosophers it is sense-contact, “with the exception of the Grammarians, who hold communicative knowledge to be more basic” (p. 345).
The Tibetan name is བརྡ་སྤྲོད་པ་. Khenpo Sodargye’s lectures on the Pramāṇayuktanidhi call them the Śabda school, after śabda (སྒྲ་), word or sound. How marginal they are here can be measured. Gorampa’s translator, surveying “the non-Buddhist schools that Dharmakīrti concentrates on”, lists Cārvāka, Nirgrantha, Mīmāṃsā, Sāṃkhya, Vaiśeṣika and Nyāya, but not the Grammarians, though Gorampa names them three times in the commentary that follows (translator’s introduction to gorampa-pramanavartika). Dreyfus agrees: Dharmakīrti sometimes refutes the Sāṃkhyas, the Grammarians and others, but “the bulk of his argument is directed against the Nyaya positions” (dreyfus-recognizing-reality, p. 52). No grammarian is named anywhere in our root texts and commentaries (not Bhartṛhari, Pāṇini or Patañjali), and no title of theirs. What they refute is a position, never a book.
As a body of Sanskrit linguistic scholarship the school continues, now mainly as a historical and academic field; as a philosophy with adherents defending the sphoṭa, it does not (Wikipedia).
Core doctrines
A real connection between word and object (སྒྲ་དོན་)
Everything else follows from this. Sakya Paṇḍita states the thesis in half a verse: the grammarians hold a real connection (དངོས་འབྲེལ་) between word and object (སྒྲ་དོན་) (Pramāṇayuktanidhi chapter 10). Sodargye makes it concrete: between the word “pot” and the particular external pot, the Śabda school holds that “there is a particular connecting relation” (sodargye-pramanayuktanidhi, on chapter 10). Shākya Chokden reports how the grammarians themselves divided names. A real name (as opposed to an imputed one) has three features: it has an etymological explanation, it is fit to be applied to that thing, and a real relation holds between the name and the thing (Shākya Chokden on Pramāṇayuktanidhi chapter 5).
The Buddhist counter-thesis has its own name, the wish-to-express (བརྗོད་འདོད་, vivakṣā): words are put to work by what a speaker means to convey, so they do not answer to their objects at all.
A word expresses the specifically characterised phenomenon
The thesis in its Pramāṇayuktanidhi form opens chapter 5. The root says only “some”; the auto-commentary names the Grammarians, the Vātsīputrīyas and others, and Sodargye the Mīmāṃsakas and the Vātsīputrīyas:
༈ །།བུམ་པ་ལ་སོགས་དོན་རང་མཚན། །བརྗོད་བྱ་ཡིན་ཞེས་འགའ་ཞིག་སྨྲ།
“Objects that are specifically characterised phenomena, such as the vase, are objects of expression”, some declare. (Pramāṇayuktanidhi chapter 5)
The argument behind it is the ordinary one, and it is not weak. The auto-commentary gives it: when someone says “bring the vase”, what is fetched is the very thing that holds water (auto-commentary on Pramāṇayuktanidhi chapter 5). Sodargye fills it out. Nobody supposes the universal was meant, and “one could hardly pull the concept of a vase out of one’s head and hand it over”; since actual vases get brought, the particular is what was named (sodargye-pramanayuktanidhi, on chapter 5).
Words express by the power of the entity
Dharmakīrti states the position as a conditional whose antecedent he rejects:
།གང་དག་གི་ཚིག་དངོས་པོའི་དབང་། །ལྷུར་བརྗོད་བརྗོད་འདོད་མི་བརྟེན་པ། །དེ་དག་ལ་ནི་དྲུག་པ་དང་། །ཚིག་དབྱེ་ལ་སོགས་ཀླན་ཀ་རིགས།
།ཕྱི་རོལ་དོན་ལ་མི་ལྟོས་པའི། །ཚིག་གང་སྨྲ་བོས་ཇི་ལྟ་བུར། །རྗོད་བྱེད་ཉིད་དུ་ངེས་བྱེད་པ། །དེ་དེ་བཞིན་དུ་རྗོད་བྱེད་ཡིན།
For those [Vaiśeṣikas and Vaiyākaraṇas] whose words are uttered principally by the power of entities, not relying on the wish-to-express, objections [based on] the sixth case, number inflections and so on are fitting. [For us, however,] whatever word the speakers, without depending on an external object, settle on as the expression [of something], in whatever way, is the expression [of it] in that way. (pramanavartika I.65–66)
The objection Dharmakīrti is answering is a grammarian’s, and it turns on Sanskrit morphology. If a thing and its properties are not really distinct, why does language mark them apart? The sixth case, the genitive, separates a possessor from what it possesses; number separates one from many. Gorampa gives the examples: the genitive, the declensions, and phrases such as “the entity of the cow, the cow itself” (gorampa-pramanavartika on PV I.65). For the Vaiśeṣika and the Grammarian, who make words track the entity, grammatical form is evidence about the world.
Property and property-bearer are really two
Sakya Paṇḍita answers a compressed form of the same objection in chapter 4, though there the auto-commentary leaves the speaker unnamed:
།དེ་ལྟ་ཡིན་ན་ཆོས་ཆོས་ཅན། །བརྗོད་པ་དོན་མེད་འགྱུར་ཞེ་ན། །དོན་གྱི་ངོ་བོ་གཅིག་ཉིད་ལ། །བརྗོད་འདོད་སྟོབས་ཀྱི་སྒྲ་གཉིས་སྦྱར།
You may say, “If that is so, the expressions of property and property-possessor would become meaningless”; [however,] to one and the same nature of the object, two words are applied by the power of the wish-to-express. (Pramāṇayuktanidhi chapter 4)
Sodargye’s example is a conch. “White conch” expresses the bearer, the whole conch being white; “the white of the conch” expresses the property, leaving aside the conch’s being conditioned, impermanent and so on. If words bear no necessary relation to objects, the objection runs, the two phrases should collapse into each other (sodargye-pramanayuktanidhi, on chapter 4).
Knowledge is necessarily linguistic
The school’s most far-reaching commitment does not appear in the root texts. Dreyfus reports it while identifying the target of pramanavartika chapter 3: the view “that perceptual knowledge depends for its production on the object as well as on the recollection of a linguistic sign”, a convention (བརྡ་, samaya). “This view is held by the Grammarians, for whom knowledge is necessarily linguistic” (dreyfus-recognizing-reality, p. 351). If that is right, the school has a larger stake in the definition of perception than our texts give it credit for.
Effects can occur without their causes
One grammarian doctrine here is not about language. Gorampa reports it at pramanavartika I.34: “The Vaiyākaraṇa (Grammarian) school says that results, in terms of concomitant and counter pervasions, are fallible with regard to their causes, because smoke, which is the result of fire, arises from a termite hill even though there is no fire” (gorampa-pramanavartika on PV I.34). The claim strikes at the effect-reason itself: if smoke can occur without fire, seeing smoke never licenses inferring fire. Why this belongs to the Grammarians, neither Gorampa nor Dreyfus says; the attribution rests on Gorampa alone, and its rationale is [unverified].
The refutations
Dignāga
The pramanasamuccaya material never names this school. Searched across the translation and Fazun’s exposition, the count is zero: no Vaiyākaraṇa, no Śabda school, no Bhartṛhari, no sphoṭa.
What it does contain is the example over which the later dispute is fought. Among the four things that must not undermine a thesis, Dignāga lists common renown:
རང་གི་ངོ་བོ་ཉིད་དང་ལྡན། །རང་དང་འདོད་འགྱུར་མ་གསལ་པ། །མངོན་སུམ་དོན་དང་རྗེས་དཔག་དང་། །ཡིད་འཆེས་གྲགས་པས་རང་བརྟན་ལའོ། །
[The probandum] possesses [the proponent’s] own nature [the thesis-property still to be established, not one already established for both parties]; it is as one oneself wishes [admitted within one’s own system], and is what is [explicitly] wished [to be proven]. And it is un-negated—by perceptible fact, by inference, by trusted scripture, and by common renown—with respect to its own subject (the dharma-bearer). [The root reads ma gsal pa; the commentary glosses ma bsal ba, “un-negated,” and that sense is followed here.] (pramanasamuccaya III)
Fazun’s illustration of the fourth is the case that runs through everything below. Where a term is uncommon and no inference is available, a thesis may still be “negated by a meaning that contradicts commonly-established speech”, for example by saying “the hare-bearer is not the moon”, since “hare-bearer” is an ordinary name for the moon (pramanasamuccaya III, with Fazun’s exposition). Dignāga names no opponent. Sakya Paṇḍita supplies one, and it is this school.
The other side of Dignāga’s relation to the Grammarians is not a refutation, and it is not in our material either. On Dreyfus’s account Dignāga took the negative half of his theory of meaning from Vyāḍi’s denotationism and the positive half from Bhartṛhari: he “holds a connotationist view, which he finds in Bhartrhari’s idea that the sense of a sentence is the intuitional idea associated with it”, and applies pratibhā to words as well as sentences (dreyfus-recognizing-reality, pp. 211–212). “Dignaga and the Grammarians hold a holistic view of meaning, according to which the sentence is the primary linguistic entity” (p. 266). Dreyfus thinks the combination unstable, and calls the relation between apoha and pratibhā an obvious difficulty in Dignāga’s system (p. 212). None of this can be checked against our sources. pramanasamuccaya chapter 5 gives the doctrine of the term-universal (སྒྲ་སྤྱི་) and exclusion (“As the [object-]meaning is exclusion-of-others, so too the universal”, in Fazun’s exposition), but the word sentence does not occur in our translation of that chapter or in the commentary behind it.
Dharmakīrti
Nine verses of pramanavartika chapter 1 deal with the school, in three places, and three of the nine are shared with the Vaiśeṣika. The chapter 1 attribution map assigns I.34–38 to the Grammarian, I.65–67 to the Vaiśeṣika and the Grammarian together, and I.92 to the Grammarian. Gorampa names the school at each of the three places and nowhere else in the two chapters he comments on.
The anthill (I.34–38). Against the claim that smoke can come from a termite hill without fire, Dharmakīrti draws a consequence and then makes a distinction:
།རྒྱུ་མེད་གཞན་ལ་མི་ལྟོས་ཕྱིར། །རྟག་ཏུ་ཡོད་པའམ་མེད་པར་འགྱུར། །དངོས་པོ་རྣམས་ནི་རེས་འགའ་ཞིག །འབྱུང་བ་ལྟོས་པ་ལས་ཡིན་ནོ།
།གལ་ཏེ་བརྒྱ་བྱིན་སྤྱི་བོ་མེའི། །རང་བཞིན་ཡིན་ན་དེ་ཉིད་མེ། །འོན་ཏེ་དེ་མེའི་རང་བཞིན་མིན། །ཇི་ལྟར་དེ་ལས་དུ་བ་འབྱུང་།
Since the causeless does not depend on anything else, it would always exist or never exist; entities arise occasionally because they depend [on conditions]. If Indra’s crown [, the termite mound,] were of the nature of fire, it would itself be fire. If it is not of the nature of fire, how could smoke arise from it? (pramanavartika I.35–36)
Smoke without a cause would depend on nothing, so it would either always be there or never be, instead of coming and going as smoke does. And either the anthill is of the nature of fire, in which case there is fire there and the counter-example fails, or it is not, in which case the smoke did not come from it. Gorampa works through both and adds the general principle at I.37: fire has a specific capacity to produce smoke, and what lacks that capacity cannot produce it (gorampa-pramanavartika on PV I.35–37). The rendering “anthill” for བརྒྱ་བྱིན་སྤྱི་བོ་ (literally “Indra’s crest”) follows Gorampa’s “termite hill”; the Tibetan alone is opaque.
The wish-to-express (I.65–67). Here the school’s real position is met, and these are the verses the rest of the tradition builds on. Dharmakīrti grants the objection against the opponent and refuses it for himself. If words expressed by the power of the entity, the genitive, the declensions and the number-forms would be evidence of real distinctions, and a Buddhist who denies the distinctions would owe an account of the grammar. But speakers settle what words express without consulting the object, so a language that distinguishes the cow from the cow’s own entity distinguishes nothing in the cow. He presses the point with cases where nobody thinks grammar tracks anything: “wives”, “six villages”, and “the nature of space, space itself”. Gorampa explains: a single woman can be referred to by the plural “wives”, and a district of six villages by a singular, and space has no parts for a possessive to mark (pramanavartika I.67 with gorampa-pramanavartika ad loc.). The price is that grammatical structure loses all evidential weight in metaphysics. That cuts both ways: a defender of real distinctions can no longer be refuted from language either.
Fetch a pot (I.92). The objection is the everyday one, and Dharmakīrti’s reply separates the moment a convention is made from the moment it is used:
།སྒྲ་རྣམས་ཀྱིས་ནི་བརྡར་བཏགས་སྟོན། །དེ་ནི་ཐ་སྙད་ཆེད་དུ་བྱས། །དེ་ཚེ་རང་གི་མཚན་ཉིད་མེད། །དེས་ན་དེ་ལ་བརྡར་མ་ཡིན།
Words show what has been designated by agreed-upon convention, and that [designation] is made for the sake of [later] convention; at that time the specifically characterised phenomenon [present when it was designated] does not exist; therefore the agreed-upon convention is not [made] for it. (pramanavartika I.92)
The pot in front of whoever first fixed the word is not the pot anyone later fetches. A convention is made to be used again, and a specifically characterised phenomenon (རང་མཚན་) cannot be used again, because by the time of the next use it has gone. So whatever the word was attached to, it was not that.
Two further engagements survive only in commentary, without verse numbers, since our translations of chapters 3 and 4 are not yet available. In chapter 3, the target of the argument that perception is non-conceptual is, on Dreyfus’s identification, the grammarian thesis that perceiving requires recalling a linguistic sign. Dharmakīrti replies that eye-consciousness “arises directly from the [object] independent of the linguistic sign”, and Dreyfus notes frankly that the supporting argument “is convincing only if one presupposes the acceptance of the Buddhist theories of causality and impermanence” (dreyfus-recognizing-reality, pp. 351–352). In chapter 4, Gedun Drub spends several pages on whether “the hare-marked is callable ‘moon’”, Dignāga’s case, and on why inference from common usage differs from inference grounded in fact, since “common-usage inference does not proceed by the existence-and-non-existence of facts, but is from wish-arisen conception”. He names no school (Gedun Drub’s commentary on PV IV). Dreyfus, reporting Steinkellner and Tillemans, records that the related notion of fitness (yogyatā) is a grammarians’ term rather than a Mīmāṃsā one, that Dharmakīrti uses it against Bhartṛhari in particular, and that Karṇakagomin lists the Grammarians among the targets (p. 514 n. 24).
Sakya Paṇḍita
The Pramāṇayuktanidhi gives this school more room than either Indian text, in three chapters that build on one another. Chapter 5 refutes the real word-object connection; chapter 4 disposes of a grammarian objection to that refutation; chapter 10 uses the refutation as a premise.
Chapter 5. The thesis quoted above is answered in the next verse:
།རང་མཚན་སོ་སོར་ངེས་པ་ནི། །མཐའ་ཡས་ཕྱིར་ན་བརྡ་མི་ནུས། །ཐ་སྙད་བྱེད་པའི་ཚེ་ན་ཡང༌། །དང་པོའི་རང་མཚན་རྙེད་པར་དཀའ།
[However,] specifically characterised phenomena, each determined separately, are limitless, so an agreed-upon convention cannot [be made for them]; and even at the time of conventional usage, the specifically characterised phenomenon [named] at first is hard to find. (Pramāṇayuktanidhi chapter 5)
There are two arguments here. The first concerns assignment: particulars are each distinct and limitless in place, time and form, so no act of naming could reach them all. The second is Dharmakīrti’s point from I.92, which Sodargye makes vivid. Suppose I named a child and the child died the next instant: if the name had been fixed to that particular, it could not pass to another child (sodargye-pramanayuktanidhi, on chapter 5). Three more consequences follow in the root text. Past and future things could not be spoken of, having no present particular, yet we speak of Kāśyapa and of Maitreya. The eye and the other faculties would be idle, since if saying “pillar” already delivered the pillar’s colour, sound, smell, taste and feel, no sense would have anything left to do. And since there is no connection in the object, the senses cannot grasp word and object as connected at all; joining them as “this is that” is the work of thought, which is to say elimination of other (གཞན་སེལ་, anyāpoha) (Pramāṇayuktanidhi chapter 5).
The consequence the commentary returns to is blunter. If word and object really were connected, Sodargye says, “the moment anyone utters the sound ‘fire’ with the mouth, flames would appear in his mouth”, and saying “water” would set water flowing there (sodargye-pramanayuktanidhi, on chapter 10). It is a crude argument, and it does real work: it fixes what a “real connection” would have to mean if it meant anything, and leaves no weaker reading of the thesis standing.
Chapter 4. Having denied that words depend on objects, Sakya Paṇḍita owes an account of why language separates property from property-bearer. The verse quoted above gives it: one entity, two wishes to express, two words with different jobs. From the side of the thing, “white conch” and “the conch’s whiteness” are the same conch; what differs is whether the speaker wants the bearer whole or one attribute picked out (sodargye-pramanayuktanidhi, on chapter 4).
Chapter 10. Here the refutation is used rather than argued, and what it secures is common renown as a kind of valid reasoning:
།བརྡ་སྤྲོད་པ་དག་སྒྲ་དོན་ལ། །དངོས་འབྲེལ་འདོད་པ་སྔར་བཀག་ཅིང༌། །བརྗོད་འདོད་དབང་གིས་འཇུག་པ་ཡང༌། །བསྒྲུབས་ཟིན་དེས་ན་གྲགས་པའང་འགྲུབ།
The grammarians’ (བརྡ་སྤྲོད་པ, vaiyākaraṇa) claim of a real connection between word and object (སྒྲ་དོན) was refuted before; and that [words] operate by the force of the wish to express (བརྗོད་འདོད) has [already] been established—therefore common renown too is established. (Pramāṇayuktanidhi chapter 10)
Because names are given at will, any name could in principle attach to anything, so what a community has actually settled on is a matter of usage, not nature. That is common renown. Sodargye distinguishes what is already renowned from what can become renowned. To call an ox “the moon” now is a joke, because the established usage says otherwise; but from the side of what could become usage, nothing prevents it (sodargye-pramanayuktanidhi, on chapter 10).
The Grammarians push back with Dignāga’s own example:
།ཡོད་པའི་རྟགས་ཀྱིས་རི་བོང་ཅན། །ཟླ་བ་ཡིན་པ་འགོག་པ་ལ། །འདི་ལ་དཔེ་མེད་ཡོད་རྣམས་ལ། །སྒྲ་ཀུན་བརྗོད་དུ་རུང་ཕྱིར་རོ།
།སྒྲ་ཡི་དོན་གྱི་ཡོད་པ་ལ། །ཟླ་བར་རུང་ཕྱིར་འགལ་བ་ཡིན།
[Objection (Grammarians):] By the reason “because it exists” [the opponent] refutes the hare-bearer’s (རི་བོང་ཅན) being [called] “moon.” [Reply:] For this there is no example, since of existents all words may be spoken. The word-referent existing, it is fit [to be called] “moon”—and so [the reason] is [in fact] contradictory. (Pramāṇayuktanidhi chapter 10)
The opponent’s reason is inconclusive and contradictory at once. Inconclusive, because it has no similar example: nothing exists that, by existing, cannot bear a given name. Contradictory, because existing is exactly what makes a thing nameable. The root text sums up: to claim that an all-pervading renown “does not enter this object” is an uncommon reason, “like ‘audible’”, the textbook case of a reason found only in the subject (Pramāṇayuktanidhi chapter 10). Sodargye then turns the school’s own premise against it. They hold name and particular to be directly related, so if existing things cannot take the name, the hare-bearer cannot take it either, “because the existence of any other dharma is not the existence of the hare-bearer” (sodargye-pramanayuktanidhi, on chapter 10).
The commentary records a qualification. Mipham Rinpoche and other masters hold that the Śabda school does in fact allow the hare-bearer to be called “the moon”; what it denies is moving the name to something else. Their stated thesis, taken strictly, denies even the first, and that is what the refutation attacks (sodargye-pramanayuktanidhi, on chapter 10). The reply may therefore answer a position stated more sharply than its holders meant it.
What the debate turns on
Every engagement above comes down to where the meaning of a word is fixed: in the thing, or in the speaker. The Grammarians put it in the thing, and can then read metaphysics off grammar: the genitive marks a real possessor, the plural a real many, the sentence a real unified content. The Buddhists put it in the wish-to-express, and gain three things at once: language about past and future, language about what does not exist, and permission to treat grammatical form as evidence of nothing. What they give up is any hold on the world through language. That is why Sakya Paṇḍita has to rebuild the reliability of ordinary speech from the other end, as common renown, and why the rebuilding takes a chapter.
Dreyfus points out an irony in the Tibetan reception that our material cannot see. On his account Dharmakīrti moved away from Dignāga precisely where Dignāga followed Bhartṛhari: away from the sentence as the primary unit and from language as inner articulation, towards language as a series of voice sounds that express meaning “in dependence on the conventions within a linguistic community” (p. 267). Yet the Sakya interpreters who refute the Grammarians hardest (Sakya Paṇḍita, Gorampa, Shākya Chokden) describe language in terms of a direct signifier that is the term-universal and a direct signified that is the conceptual generality (དོན་སྤྱི་). Dreyfus judges that they thereby end up, “probably unbeknownst to them”, closer to Dignāga than to Dharmakīrti, with a view “similar to the sphota view of the Grammarians” (p. 274). If he is right, the tradition that produced the longest refutation of the real word-object connection also kept the grammarians’ picture of where language lives. The judgement is Dreyfus’s alone; the question is never raised in our other sources.
Three things this page cannot settle. First, a sphoṭa-like doctrine is refuted at length in pramanavartika I.247–268, where partless words and speech are said to be made manifest by a sequence of sounds. Gorampa assigns that stretch to the Mīmāṃsā, not the Grammarians (“The Mīmāṃsaka do not assert mere sound to be verbal expression”, gorampa-pramanavartika on PV I.258), while Dreyfus treats sphoṭa as Bhartṛhari’s doctrine (dreyfus-recognizing-reality, p. 212). Dreyfus himself notes the overlap: Kumārila and Bhartṛhari agree that a universal is the meaning of a general term (p. 137), and both reject the conventionality of language (p. 218). Second, the two commentaries name the opponent of Pramāṇayuktanidhi chapter 5 differently. The auto-commentary says the Grammarians and others, as our translation has it; Sodargye says “the Mīmāṃsakas (and within Buddhism, the Vātsīputrīyas)” throughout the same passage (sodargye-pramanayuktanidhi, on chapter 5). Third, why the anthill counter-example to the effect-reason belongs to the Grammarians is nowhere explained.
Related pages
- apoha: the theory that replaces the real word-object connection, and, on Dreyfus’s account, the answer to a question the grammarians asked first
- problem-of-universals: Bhartṛhari’s claim that the object of a word can only be a universal
- conceptual-cognition and perception: the grammarian thesis that knowing requires a linguistic sign, and Dharmakīrti’s denial of it
- specifically-characterised: what a convention cannot be assigned to
- inference: common renown as a kind of inference distinct from fact-grounded reasoning
- vaisesika: the co-defendant at PV I.65–67
- mimamsa: the other school of the permanent word, and the rival attribution for the partless-speech doctrine
- dignaga, dharmakirti, sakya-pandita, gorampa: the refuters, and in Dignāga’s case the borrower