Chapter 5: Exclusion of the Other (Apoha)
The Mīmāṃsakas and Naiyāyikas count verbal cognition as a third means of valid cognition alongside perception and inference. Dignāga’s answer is that a word reaches its object by the same “not arising without it” relation that a logical reason does, displaying its meaning by excluding other meanings, so verbal cognition falls under inference and is no separate pramāṇa. The rest of the chapter dismantles the realist alternatives in turn: the kind-word, the possessor-of-the-kind, and the relation between quality and quality-possessor.
Condensed from Fazun’s Chinese translation-and-exposition of the Pramāṇasamuccaya and its auto-commentary, and rendered into English. Root verses are given as Fazun gives them; the prose beneath summarises his exposition. On the provenance and the limits of this rendering, see the series introduction.
Chapters: 1. Perception · 2. Inference for Oneself · 3. Inference for Others · 4. Example and Pseudo-example · 5. Apoha · 6. Sham Rebuttals
Volume Five
Chapter on the Examination of Exclusion-of-the-Other (Apoha)
The chapter opens from the two-pramāṇa position established earlier. Some schools (the Mīmāṃsakas and Naiyāyikas among them) add verbal cognition (śabda-pramāṇa) as a third: taking the uttered sound as cause, one comprehends the meaning stated. Dignāga’s whole chapter is the denial of that third pramāṇa.
“Verbal utterance is not apart from inference, [as if it] were some other pramāṇa; through the gateway of exclusion-of-others it displays its own meaning, like ‘being produced.‘”
A word reaches its object by the same “not arising without it” connection (avinābhāva) that a logical reason does. Just as “being produced” establishes impermanence by ruling out the non-produced, a word displays its meaning by excluding other meanings. Verbal cognition is therefore subsumed under inference and is no separate pramāṇa. This thesis governs everything that follows; the rest of the chapter is the refutation of the alternatives.
[Objection, realist] The kind-word is constituted out of each individual particular, and in order to yield certainty about what is stated it is a particular-word.
“The various particulars have no kind-word, because [they are] limitless and because [it would be] erroneous.”
“Word” here means the expresser. Take the kind-word “existent”: it cannot be the expresser of substance and the rest, on two counts. First, limitlessness. The particular dharmas cannot each be connected with the word, and a word with no connection cannot express a meaning at all; it can only announce its own sound. Fazun’s analogy: hearing the speech of a foreign region, one hears the noise and grasps nothing. Second, error. “Existent” operates on substance, but equally on quality and activity, since all are existent. A word that spreads indifferently produces doubt, not determination, so it is not an expresser.
[Objection] Then let only the kind-word, or only the word for the possessor-of-that (tadvat), be the expresser, since it is easily connected and non-erroneous.
“The kind[-word] or the possessor-[of-that] word does not separately express the various meanings.”
We say “existent substance,” “existent quality,” “existent activity,” treating substance and the rest as sharing one locus with “existent.” But if “existent” or “possessor-of-that” were what the word expresses, it could not be present in substance or in quality; one would have to say “the existence of substance” in the genitive, which is precisely not what the co-referential construction says.
“Quality and quality-possessor (有德, guṇin), because stating [their] difference [implies they are] determinately distinct; the substance [serving as] common locus, and the word, are both fully established [for the disputants]; the connection is stated as drawn out by a connection-bearing dharma. Likewise, by ‘being-produced’ one states the thing, not by anything else.”
A summary verse. Merely having a connection does not make two things inter-connected, since obstructions may intervene. Things are said to be connected only by a connection-bearing dharma; a thing’s own dharma cannot generate the connecting word. “Being produced” states the thing by that route and by no other.
“Therefore the kind-word’s expressing [the particulars] is unreasonable.”
[Objection] Then state only the possessor-of-the-kind (具種類), which shares a substratum with the particular-words, is easily connected, and is non-erroneous.
“Having no autonomy, [it is] not the possessor.”
If that held, the self-nature of substance would have to be reached through the differentiation of “existent” rather than by saying “substance” directly. But “existent” does not draw in pot and the other particulars, so it shares no common locus with them, exactly as “non-existent” and “non-produced” share none, and as “white” and “word” share none. The consequence is that “existent” and “sweet” and the like fall apart into distinct things.
“Because the metaphorical positing (假設, upacāra) does not exist.”
The tadvat theory works by having the word express its own essence or the kind, then metaphorically positing that it operates on the particular and possesses it. Whatever is metaphorically imputed onto a basis is a non-real meaning, and even the characteristic of such positing is not found. The theory requires that cognition transfer from universal to particular by sameness of quality, or by sameness of the benefit the quality confers. Neither sameness exists.
“As in calling a subject [a commoner] ‘king’ — because the cognitive images are each distinct.”
Why is there no transfer of cognition? Because the images differ. One who says “that person is my very self,” applying the word “self” to a servant, does not thereby produce the same cognition of king and of servant.
“Those stated in sequence — like the whiteness of flower, conch, and the like.”
Where cognition really is the same, the instances are stated in sequence: the white colour of the kunda flower, of the kumuda flower, of the conch. The tadvat kind-word is not like this; it is meant to apply all at once, and there no notion of cognition-transfer arises.
“With respect to a thing benefited by a quality, [the cognition] would operate even without [a separate] awareness [of the quality].”
If the referent were the thing possessing the quality, as crystal is benefited by the red it holds, then cognition of the substance would not depend on cognition of the quality. In the cognition of red crystal there is no separate cognition of the substratum, and those unfamiliar with it have no distinct cognition of the quality apart from the thing.
“The mixed image, with respect to everything, would all become erroneous cognition.”
Extending the crystal case: every word-cognition, being mixed with the object-meaning, would be mistaken, and no knowledge would accord with its object.
“The universal and the rest, [its] plurality, and those [cognitions] that apprehend [many] at once — the benefit [conferred] would be contradictory.”
Consider cognitions that take many at once, as when one says pot, earthen, substance, bitter, sweet, fragrant. The benefits conferred by these qualities contradict one another. The universal cannot become one with the substance through its qualitative image, since no differentiation would remain; nor can many qualities be apprehended on a single object-field through one qualitative image.
“Seeing the many colours of an agate.”
[Objection] All the qualities belong to the pot, so they may confer benefit simultaneously. Reply: yet nothing is thereby grasped of the thing “pot.” On that account, seeing all the mixed colours of an agate would count as obtaining all the things at once, which is not how word-cognition works.
“Those various particular dharmas — the fault of ‘possessor-of-that’ is the same [for them]; the kind-ness or the possessor-ness — by what [reasoning] would they not incur [it too]?”
The same must be said of distinct entities: stating the kind as a distinct entity, then metaphorically positing distinctness with respect to the thing-kind, carries the fault already stated for the possessor-of-that. One can superimpose the word’s own nature on the stated kind and call it substance, which is how “possessor-of-the-kind” and “kind-ness” arise. They inherit the fault.
“Stating the possessor-of-that as distinct — that has already been refuted as before.”
“Already refuted” points back to “the various particulars have no kind-word” and to the extension of that fault to the possessor-of-the-kind.
[Objection] The kind is stated only by reference to the possessor-of-the-kind, not by the self-nature of the particular dharmas.
“Only the possessor-of-that should [then] be examined — as [a matter of] connection or as the existent.”
Then the possessor-of-that must itself become either a thing-condition or a connecting quality. What counts as connection is subsumption, function, benefiting-another, and common designation by others where distinct things are not confused. Either way the earlier verse applies: “the kind[-word] or the possessor-[of-that] word does not separately express the various meanings.”
“The pot and the like as the possessor-of-that meaning — if cloth and the like do not operate [under it], what is its universal-meaning?”
A universal is what operates on many dharmas together. If pot possesses the universal but cloth does not fall under it, the two have no universal at all. Saying “one states the possessor-of-the-universal” only pushes the question back: what is that universal’s superimposed meaning? And if “existent” is the term expressing the universal, there is no further “existent” above it, so a determinately admitted “existent” universal-meaning is not found in the object.
“Only the word-meaning is the same.”
Pot and the rest, as possessors-of-that, are what “existent” expresses. That the possessor-of-that combines with “existent” has never been established.
“[We] admit it has no causal condition.”
No non-distinct causal condition is seen among the particular-words, so this cannot be stated as a universal either, and the various qualificands drop away with it.
[Objection] “Existent” operates on a substance possessing a non-distinct quality, as “deep blue” is admitted to operate on pot, and thereby becomes qualifier to that qualificand.
“If [it were] one, like blue and the rest — not so; and even if admitted, because [there is] kind without kind, [it is] not [so].”
It would have to become a connected quality, and it does not. “Blue” has no “deep blue” and the like beneath it through a possessor; explaining it by another possessor-of-that is unreasonable; and “the connection of blue” is not the word’s meaning. Granting the universal anyway, a kind has no further kind above it, so it is not a universal. Granting even that, there are no blue qualities divided into three classes by a blue-universal. Since those who admit the possessor-of-that already admit the differentiation of the kinds of pot and the rest, the pot-as-tadvat has no linkage left; at most the particulars are drawn in by the force of the object-meaning.
“The drawing-in by the [object-]meaning is not certain.”
To be drawn in by the object-meaning is to arise determinately by the force of that meaning, as in “if a devotee does not eat during the day, he determinately eats at night.” But “existent” is not restricted to pot and the rest, so the reason is inconclusive and there is no valid drawing-in. The realist thesis that the kind-word is the universal connected with the particulars, expressing them through the possessor-of-the-kind, is therefore no answer.
“Therefore the word too [operates by] exclusion-of-others.”
The verse returns to the opening thesis: like “being produced,” the word displays its own meaning through the gateway of exclusion-of-others.
“Although the expressed [referents] are many, the word cannot make all [of them] known; [what it does is] the result of excluding [others, namely] the meaning connected with itself by concomitance. The word also cannot operate on the meanings of many dharmas; [it operates] only on what is combined [with it], not by the word’s quality and the rest. — These are summary verses.”
[Objection] If the word-meaning is only exclusion-of-others, how can “blue lotus” form a common locus and a qualifier-qualificand relation? Reply: the universals are distinct because each is the exclusion-of-others effected by its own differentia-word, so no fault arises.
“Because the exclusion of the [respective] other meanings differs, the comprehension of its own meaning differs from confusion; because of the single, non-distinct result, [there are] qualifier and qualificand.”
In “blue lotus,” each word excludes a different range, and each excludes the differentiation of its own meaning. Fazun’s comparison is with “standing erect” and “a bird alighting”: several exclusions gather on one thing and so form a common locus. Taken singly, each word remains inconclusive with respect to its own differentiation; it is not that one word becomes the excluder in company with the other. Because they converge on one result, they also stand as qualifier and qualificand. One substratum is neither blue alone nor lotus alone, yet is both.
“That is not only blue, nor only lotus, because the aggregate [the compound’s joint reference] is what is expressed.”
The words “blue” and “lotus” combine with the whole of that substratum, not with a single kind.
“Like ‘colour’ [the bare syllables], [it] would be meaningless. (Xinhui’s text reads: ‘like a [single] letter, it should be meaningless.’)”
Were the reference a single kind, the case would be like reciting “ni” and “la” as bare syllables: the sounds are there, but “nīla” carries no meaning as separate parts.
“Colour has not the slightest meaning [from its parts]; yet by those two words [together] one can comprehend what they express, and that too comprehends the single meaning.”
[Objection] Saying “colour,” the component sounds “ni” and “la” give no fraction of understanding. Reply: the syllables are emptied of separate sense by being absorbed into “nīla,” and in the same way “blue” and “lotus” are emptied of separate reference by the aggregate-word; the compound is understood as a single meaning. Where distinct words are used instead, the force of the object-meaning makes them state distinct meanings. So on the exclusion-of-others account both the common locus and the qualifier-qualificand relation come out reasonable, and on the rival accounts the word-meaning does not.
The rival is now put to a dilemma. Are the member (支) and the possessor-of-members (有支), together with their substratum, distinct or non-distinct? Take non-distinct first.
“Because [they] would mutually become one; the aggregate has no single-ness.”
If the aggregate were one, the lotus-meaning and the blue-meaning could not differ, since within one meaning there is no other, and the two would collapse into each other.
“The aggregate would become many.”
Conversely, since the aggregate is not distinct from the many dharmas, it becomes many. On either reading it is not one “existent.”
“The word does not relinquish its own [particular] meaning.”
Even granting the aggregate exists, it and the blue-lotus meaning share no common locus, since each word operates on a single thing. If a single separate meaning were the meaning of both “blue” and “lotus,” nothing within the word or within the aggregate could licence saying “those are this.”
[Objection] No fault: the universal subsumes the particulars, the two particular meanings are of one nature by the universal-meaning, and in order to make the meaning known along with its differentiations one combines and states both; hence the two form a common locus by possessing the aggregate-meaning. Reply: “blue” expresses an undifferentiated meaning by exclusion, not a kind. If the universal of “blue” expresses neither substance nor kind, there is no ground for examining it as subsumed within differentiation or within substance and kind. And if the substance too were expressed by the universal (Xin-hui: “the substance possessing blue is also what is expressed by the universal”), then “lāba” (partridge) and “utpala” (lotus) and other unrelated things would be drawn in as well. This was already settled at “having no autonomy, [it is] not the possessor.”
“Because of doubt, [it is] not exclusion-of-others.”
The universal-word raises doubt about the differentia-word, precisely because its expressing the particular is unreasonable. [Objection] Even so, the differentia-word makes one comprehend the universal without error. Reply: then the universal subsumed within the differentiation is not validly expressed either, since it is not by exclusion-of-others. If “blue” expressed the universal directly, as “utpala” does, “blue” would not operate by exclusion, and one would have to state the universal as subsumed within the particular. Subsumed, it is subsumed within the expressed meaning; and if both universal and particular are the differentia-meaning, the words become synonyms (異門, paryāya).
“[The] one [thing] would become [denoted by] synonyms.”
A synonym is stated for someone who has not understood, or where several meanings are in play, to convey the earlier meaning by another route. Synonyms still express the former meaning and are not stated all at once. On this view “blue” alone would already convey the lotus-meaning, and the qualifier-qualificand relation would vanish. That completes the faults on the non-distinct horn.
The distinct horn fares no better. One who reckons them distinct posits, apart from the various aggregate-dharmas, a separate single aggregate-ness, and must then say whether that aggregate-ness operates upon the aggregate-dharmas or they upon it, wholly or in part. Neither is found.
“If distinct, then [it is] not the two either.”
There too there is no common locus and no qualifier-qualificand. Is the aggregate-ness one in object-meaning or one in word? If one in meaning, is it one as quality-and-substance or one as possessor-of-that? The same fork applies to being one in word.
“The meaning’s substance and quality — if [they] operate on one [thing], then universal and particular, the two, do not exist.”
If the blue-quality and the lotus-kind operate on one substance, both the common locus and the qualifier-qualificand relation vanish: the blue-quality does not possess the lotus-kind, nor the lotus-kind the blue-quality.
“The possessor-of-those-two also does not exist.”
If instead the blue-quality and the lotus-kind possess the universal-meaning, there is again no common locus. In a substance on which the other operates, neither the qualifier-qualificand relation nor the two-in-one is found, because universal and particular do not combine. So much for the object-meaning side.
“Likewise the word should be only the stating of kind and quality.”
On the word side: if the word expresses only kind and quality, then since these are utterly distinct, that too fails. They have no common locus, since no connection is stated, and the word has no qualifier-ness. (Fazun: this seems to read “has no qualifier-qualificand.“)
“In stating the possessor-of-that, [it] would become [both] same and not-same.”
If the words for kind and quality express one differentiated substance, their substratum is the same; if they express distinct qualifier-substances, it is not the same. The indeterminacy leaves no common locus.
“Because [they are] the same, [one] does not care to state [it].”
[Objection] We accept the common locus, and since they are the same we do not care to express the difference. (Xin-hui: “one does not care to state the not-same.“) Reply: then there is no differentiation of connection. Wherever there is a cause of the same cognition one calls it the common locus, and other differentiations, such as “the flower has not yet opened,” are simply not what one cares to state; but this defence generates endless attacks. Where both are held together and a shared meaning persists, and the word has its function there, one ought to care to state it. If there is a connection that one does not care to state as a differentiation, how is the existent thing stated at all?
“Ox, horse, and the like would also be the same.”
Take ox and horse as having no differentiation at all by the connection of ox-ness, existence and the rest, and suppose one does not care to state the universal but merely states the thing. Then ox and horse would form a common locus just as “lotus” and “blue” do, which nobody accepts.
“Those are not [yet] differentiated. (Xinhui’s text: ‘those are differentiated in aspect.’)”
[Objection] Ox and horse differ in shape or aspect, and the cause by which each excludes its kind differs, so they are stated separately; the charge of “not caring to state” does not stick. Reply: then they are not like “lotus” and “blue,” and equally should not form a common locus in a single referent.
“[There] would become many aggregations; and what is admitted also would not exist.”
Since many universals can be displayed on a single thing, there would have to be an aggregation formed of many shapes, and no such thing is observed. As for “the same”: with ox and horse it has not been stated whether the substratum is same or not-same, and ox-ness and existence are not displayers; call them the same and they form a common locus. (Xin-hui reads: ox and horse would become substratum same-and-not-same, because they are not expressed and because existence and ox-ness are displayers; there too, not caring to state the not-same, they would form a common locus.) And “what is admitted also does not exist” means that those who do admit a common locus, stating “lotus” and “substance” as existent, would not obtain it. If the universal’s being a displayer of the particular is what denies ox and horse a common locus, then the existence of lotus and substance would likewise have to be displayer-natured.
“Quality and substance should not be displayers.”
Quality and substance are not shapes, so the universal connected with them cannot be a displayer. [Objection] The universal displays particular and common characteristics by the differentiation of causal capacity (功能, śakti), so no difference of shape is needed among substances. Reply: not acceptable, since we too admit a differentiation of capacity among the particular things themselves, as with the lotus. And if the universal is the cause that displays the particulars, one must say what the differentiation of ox-ness and horse-ness consists in. The answer is that the words expressing ox and the rest follow the speaker’s wish, and what each word expresses is just that differentiation.
“The difference is by the difference of the expresser; the word itself has no [intrinsic] differentiation.”
So universal and particular are distinct only by the word, not in their own nature. [Objection] An undifferentiated thing is metaphorically posited as distinct by differentiated words. Reply: the realist cannot say this, since on his view the differentiations are the cause displaying the universal, so the differentiations of ox and the rest cannot be merely verbal. [Objection] It is like “kara” and other bare phonemes, distinct by difference of expresser. Reply: that needs analysis, since one dharma is expressed by many synonyms too. The point being made is that the word, like substance, has no intrinsic differentiation of its own; ox and the rest have none either, and are made into differentiations by the universal, which is exactly the realist’s own posit turned back on him. [Objection] Is not the manifest thing itself the differentiation of the word “ox”? Reply: the manifest thing has differentiation, but it is not what is expressed, since it is not seen together with the conceptual meaning and since ordinary users of conventional terms like “cart” cannot state that particular. So it is by reference to the universal that the differentiations among words are divided.
[Objection, Vaiśeṣika] The universals have differentiations by their own nature, because a thing depending on its respective universal is the displayer, as one says head, back, belly, hand.
“Displayer and displayed being distinct, [you] would obtain mutual dependence.”
Ask what the self-nature of substance is, and the answer is that it displays its own universal; ask what the self-nature of the universal is, and it is what is displayed by its own substance. Each is defined by the other, so neither can display its own essence.
“Self-nature being many and one, mutual display would [make them] many.”
A single substance displays many universals such as existence, so it has many natures; and many substances, the earthen and the rest, display existence, so the existent has many natures. Since in every case the connection of qualities, the capacity, and the expresser all differ, one thing would have to be admitted as having many natures.
“Because quality, capacity, and expresser differ, [the thing] is posited as distinct; because one and many are mutually connected, [it] would become many. — This is an interstitial verse.”
The conclusion of the whole stretch: on the realist account there is no common locus.
“Qualifier and qualificand have no autonomy [on that view] — as stated before.”
Nor, as already shown under the possessor-of-that, is there any qualifier-qualificand relation. Both therefore come out entirely reasonable on the exclusion-of-others account of word-meaning, and on no other.
“[For] the non-other universal and particular, [the words] should be stated as synonyms.”
Though all are the same in being exclusion-of-others words, the meanings of the synonyms, the universal and the qualifier-qualificand of non-other nature, are still exclusion-words: there is no contradiction, and synonyms do not combine all at once. Being equally exclusion-of-others does not mean each abandons its own meaning. Universal-words exclude others with respect to their own differentiations only by repeatedly taking up the differentia-words, since that is what is sought: a birch is “birch, not palāśa,” and equally “not pot.” This answers the objection that the universal is not the word-meaning, and shows why the universal-word depends on the qualifier-qualificand words when it fixes its object.
“Because [there is] no contradiction, the aggregation can also express the other meaning.”
The word-meaning of aggregate and universal is still exclusion-of-others, and nothing here is unreasonable. Because each universal-and-particular word operates on the universal of its own meaning, it can also express the other meaning as differentiated by two or many words.
“Admitting [it] thus, and because admitting [thus they are] distinct, [the words] do not abandon their own universals; not grasping a single meaning, these [individually] equally give rise to doubt; though many grasp the universal, by the differences [they are] non-erroneous. Qualifier and qualificand — those are not [merely] the same. — These are summary verses.”
Question: by what cause does a word exclude the other particular word-meanings?
“A particular, with the other particulars, being mutually contrary, can exclude [them].”
Particular-meaning words are mutually contrary because each seeks mastery over the universal-meaning for itself, like princes vying for one throne; towards meanings they are not competing for, they coexist. Saying “this is a birch,” the word “birch” carries the image co-present with “tree” and thereby excludes the dāmbu and other trees, establishing its own object.
“The other universal’s various particular meanings are contrary to one’s own universal and the rest.”
For unconnected universals such as pot: the word “tree,” having gained mastery over what is earthen-made, is contrary to pot and the rest under that heading. What it cuts off is cut off as one rejoices at the defeat of a friend’s enemy, and the meaning too knows what has been cut off. So the particulars of other universals, quality and colour among them, whether connected or unconnected, are excluded and set aside on the friend-and-friend’s-enemy model.
“That [word] does not directly exclude that [unconnected thing].”
The word “birch” does not directly exclude pot and the like.
“Because the universal is not [their] common nature.”
If it could exclude them directly, pot would have to share the meaning of the word “tree.”
“Likewise, [as] distinct from the other [particulars], that [word] cannot exclude [them directly].”
“Tree” does not directly exclude palāśa, a kind of tree, and by parity “birch” cannot either. [Objection] No fault, since the amount each can exclude differs. Reply: unreasonable. “Tree” and “birch” each express the distinct things tree-ness and birch-ness, expressing utterly distinct meanings, so the meanings come out unmixed and the same holds for direct exclusion.
[Objection] If a word can exclude the other differentia-meanings, how does one state one quality and another as sharing a locus, as when moistness, coolness and heaviness are stated in sweet flavour?
“If [they have] the same substratum, a quality and another quality, because [they] operate on one substance, the substratum is not contradictory.”
Where they exist in one substance, qualities such as moistness can be posited in sweet flavour, since they combine in one object-meaning. No contradiction follows.
“Because [it is] not seen, [the word] also excludes [it].”
Each particular-word also excludes by its referent’s not being seen in the other particular-meanings. “Not seen” is stated to refute those who grasp the thing-meaning with self-conceit and take that as sufficient.
“The universal would [then have to] exclude its own particulars.”
If mere not-being-seen sufficed for exclusion, the universal-word, whose referent is likewise not seen in its own particular dharmas, would have to exclude them, which the realist does not want.
“[No,] because [the universal] is otherwise seen [to be present]. [translation uncertain]”
“Seen” means that by universal-words such as “time” one does see the particular.
“Thus that doubt appears.”
Then the universal-word, appearing with respect to the differentia, would reasonably give rise to doubt.
“If [you] say doubt is reasonable. (Xinhui’s text: ‘if [you] say doubt is unreasonable.’)”
[Objection] If the particular is obtained by words like “time” and by the concurrent force of the object-meaning, how does doubt arise?
“[It is] the isolated [particular] that [remains to be] ascertained.”
One should not say so. Doubt about the possessor-of-those means that, once a thing is determined as possessing the meaning, doubt afterwards arises by reference to the isolated particular.
“If the particular has no isolated [presence].”
[Objection] The universal-word does not show the particular in isolation among the differentia, as with “the kind-word is not the expresser of the various particular dharmas.”
“Depending on the hearer, [it] is seen.”
Hearers, having excluded others by way of the universal, then go looking for the isolated particular. The exchange runs: “Is this tree born from the earth-element, or from the five elements?” — “Born from the earth-element.” Once determined, “born from the earth-element” is stated of that tree, and the doubt about the isolated particular is what the determination answers. Merely hearing the phrase yields an isolated inference-formula, so no fault follows. As the object-meaning is exclusion-of-others, so is the universal: the meaning excludes other meanings, which is the meaning-universal, and the word excludes other words, which is the word-universal.
“Having excluded the other words, [the word] expresses [its meaning] by the word-universal.”
Just as “being produced” excludes “non-produced” and yields impermanence by the universal, a word excludes the other words and expresses its meaning by the word-universal, not with respect to the external object.
“Where [there is] doubt about one meaning, [there is] doubt about many meanings too.”
Doubt about a word-meaning being “cart-part” and the like spreads: one word raises doubt about many meanings, as “bhāva” is also doubted as “existence-ness.” Yet the hearer, starting from that universal-word, also views it as a particular-word.
“Because [the word], not seeing the other word-meanings, displays the portion of its own meaning, the word’s connection-nature is easy [to establish], and there is no error either.”
This is the account of positive and negative concomitance (anvaya, vyatireka). A word operates where things resemble its referent and not where they differ. In the similar locus one does not claim it operates throughout every instance, since among limitless meanings some class may go unstated. In the dissimilar locus, even if the referent were present somewhere unstated, the word cannot operate across limitless loci; its being “not stated” rests only on its not being seen. So apart from what is connected with itself, the rest is not seen, and the inference that excludes that rest is precisely what expresses the word’s own meaning. Were one to infer by positive concomitance alone, then just as “tree” leaves doubt whether the manifest image is birch or another, “born from the earth-element” would leave doubt about non-substance and the rest. Since “tree” is not seen in what is not born from the earth-element, it is by its negative side that the inference goes through.
“Tree, earth-born, substance, existent — the object of measure in reverse order: four, three, two, one [are] doubtful, the rest is the determining reason. — This is an interstitial verse (see the appended table).”
Apart from what carries the connection, other words and their reason-nature do not express the meaning, since one word cannot express all of many meanings. Yet no word is erroneous about its own meaning, since each expresses distinctly, and the alleged fault does not arise.
Appendix: Table of the Universal–Particular Relation [Figure: B09p0441_01.gif]
The particular determines the universal; the universal does not determine the particular. A birch is necessarily a tree, a tree necessarily of the earth-element, earth necessarily a substance, substance necessarily existent. Read downwards it fails: the existent need not be substance (quality and activity are existent too), substance need not be earth (water, fire), earth need not be tree (pot, cloth), a tree need not be a birch (elm, willow). Hence the counts in the verse.
“Because [it] refutes the other pervasion, that particular-meaning is not [excluded as merely] ‘other.‘”
The universal-word excludes other meanings without drawing in its own differentiations, so it can share a substratum with the particular-word, and there is no fault of its being too distinct from the particular-word’s meaning. On that view the same-word and the exclusion-of-others meaning are both what the word expresses. The word is not erroneous about its own meaning, since the isolated particular does not operate on the other.
“Because [the word’s] direct operation has no distinct [other referent].”
A word grasping the other meaning does not operate on each thing’s other meanings, so there is no fault of failing, under another’s mastery, to draw in the appropriate particular dharma.
“Because its particulars are limitless.”
Unable to pervade limitless particulars, the word is not their direct expresser, so no universal-fault arises either. The meaning is only exclusion-of-others, being neither each distinct thing nor the substance. One should therefore not look for a universal possessing other differentiations, since the word directly obstructs the other meanings. Free of the stated faults, exclusion-of-others alone, taken as the word-meaning, is the most elegant account.
“The various class-dharmas abide [stably].”
The class-dharmas are to be regarded as one, permanent, and pervading all particular dharmas. The word abides in just this nature, being non-distinct, its substratum uninterrupted, and comprehending all the meanings. Having neither the stated faults nor the lack of these virtues, the word for things excludes other meanings and so serves as their differentiation.
“By what is the exclusion-of-others determined?”
[Objection] The word “colour” (rūpa) excludes flavour and the rest, but it does not additionally state visible-colour specifically; how does it exclude something utterly distinct, which needs no excluding within colour? Reply: admit that blue and the rest are not distinct from colour and the fault disappears: blue is colour, flavour is not.
“Worldly acceptance — [one] should not seek [beyond it].”
The Bhagavān said: “One should not be attached to the names of beings, nor should one discard worldly names.” Where a term has a conventional cause and is commonly proclaimed, we follow worldly convention and do not hunt for a real referent. The world admits that “colour” is accomplished with respect to blue and the rest, not with respect to flavour.
“[Otherwise] colour would [apply] equally [to all].”
If everything is utterly distinct, why does colour-ness operate on blue and not on flavour? Refuse it to flavour on grounds of utter distinctness, and by parity it must be refused to yellow. [Objection] There is a reason: being apprehensible-by-the-eye is non-distinct from blue and absent in flavour, since their self-natures differ.
“[Then it is by] the function [of being] apprehensible-by-the-eye.”
What is grasped as object by the eye is called apprehensible-by-the-eye. So the word “colour” is applied to blue by a function-cause, not by a kind-cause. If that nature were non-distinct from colour, there would be no point in speaking of colour separately. Say instead that a thing is called apprehensible-by-the-eye through connection with colour-ness, and then apprehensible-by-the-eye either combines with the function or manifests colour-ness; by what, then, is the universal “apprehensible-by-the-eye-ness” itself determined? Its determination has to fall back on the thing’s own self-nature.
“Substance and the like would also be so.”
Substance, number and magnitude are also apprehensible-by-the-eye, so they too would have colour-ness.
“[Then] there would be no distinguishing of white and the rest.”
Being apprehensible-by-the-eye admits no differentiation, so blue, yellow, great-blue and deep-blue could not be distinguished. Drop the appeal to it: the world commonly applies “colour” to blue, yellow and the rest, and not to flavour, and that is what determines the exclusion.
[Objection] A word can express its meaning without depending on excluding the other meaning.
“[Then it] would be accomplished by positive concomitance alone.”
If a word expressed its meaning positively and directly, the analysis into positive and negative concomitance would have to be abandoned, whereas that analysis works precisely because grasping the other differentiations determinately and stating both together each have their result. Expressing meaning by way of the counter-class is as when one says “the excellent doing is activity (karma).” [Objection] If the word-meaning is only exclusion-of-others, then negative concomitance alone should be the expresser.
“[We] do not admit a direct pervasion.”
Were positive concomitance not admitted, that would follow; but among things there is no kind either distinct or non-distinct, as already shown. When it is said that the word has not seen the other word-meanings, there being no real kind, the word’s positive and negative concomitance operate on the meaning differentiated by exclusion-of-others, with no reference to a real other meaning.
[Objection] Ox and the rest are said to be excluded; but since the bare non-existent does not transform at all and yet becomes the basis of all natures, things would have to come to have substantial existence out of non-existence.
“If [someone] admits everything is connected with the ‘nothing’ [non-existent], by what does one state which dharma, [given the] existent has many natures?”
If everything is connected with the “nothing,” then setting existence-ness aside, the remaining ox-negations superimposed as “ox-existence” would make everything operate by positive concomitance with the bare existent, and the objector cannot then reply that everything would be nothing.
[Objection] “Nothing” and “existent” are known not to be distinct, just as bowl and the rest are known to belong to clay-nature, and the transformation of self-nature is likewise known.
“If [you] admit that, by the difference of clay, [there is] a distinct cognition of bowl and the like — [then since] the ‘nothing,’ apart from difference, is not other [than the existent], why [do you] refute the distinct cognition [of existents]?”
If the bowl is not other than the clay, there should be only clay-cognition. How then, by affirmative expression (表詮, vidhi), do you admit the cognition of bowl as distinct? And where there is a distinct cognition of existents against the “nothing,” why not admit it by the force of the word’s own difference? (For “by the difference of clay,” Xin-hui reads “clay has no difference”: clay and bowl are non-distinct, yet by difference of cognition one posits bowl as distinct; likewise “nothing” and “existent” are non-distinct, so why not posit existent dharmas as distinct by difference of cognition?)
“The qualities and the supreme self-nature (prakṛti) do not belong to the path of the seen; whatever belongs to the visible path is all like an illusion, utterly empty; though there are mutual differences, yet by seeing the hump [of a humped ox] and the like — that cognition contradicts [your] example, and this is according to your own tenet.” (“Hump” (犎) — Xinhui’s and other texts all read “dewlap” (胡).)
Turned against the Sāṃkhya on their own tenet. Turning back from non-ox, the cognition “ox” comes by seeing the hump and the rest; “ox” and “hump” are distinct as words, and though they differ and mutually exclude, what the Sāṃkhya admits is stated determinately even while it contradicts their own example.
“This [cognition] has no [object it] depends on as known; [it is] only the fabrication of one’s own conceptualisation. For a word that does not [operate by] exclusion expresses no nature at all.”
With hump and dewlap there is no further possessor-of-the-universal nature left over, and anything non-dependent is determinately non-existent, as shown earlier. Since a thing’s self-nature is not what is expressed, things cannot be made distinct by that route.
[Objection] There would then be no first cognition.
“Because [it is] beginningless, [it is] happily established.”
A “first cognition” applies only to non-existence, because the cause-and-effect series of conventional designation has been handed down from beginningless time. Were there a first cognition, it could not fully possess the kind, could neither turn back nor fail to turn back from the possessor-of-the-kind, and could operate on neither one nor many.
[Objection] Cognition has no real operation.
“[It operates] by excluding through the universal-gateway.”
No real entity does the excluding, as a real kind would; it is only the universal-dharma operating on the excluded object, which is what is called the excluder. This merely displays that the referent is not seen in the dissimilar locus, and serves as an inference. The objector’s own tenet incurs the same charge, and if the universal operated throughout the similar class there would be no limit to what is pervaded. Hence “because it has horns, it is not a horse”: whatever has horns is not seen to be a horse, so one infers in accordance with what is excluded, rather than separately excluding yellow-horse and separately operating towards ox. The Sāṃkhya’s own account of the receding and following of cognition works the same way.
“If [you] admit there is a separate self, examining [its] oneness and manyness, that is unreasonable for the thing.”
Grant the self real existence and both limbs of the examination fail.
“Your [Sāṃkhya] ‘manifestation’ and ‘capacity’ would be likewise [unreasonable].”
If the examination is reasonable only where there is no real existence, then Sāṃkhya “manifestation” (顯) and “capacity” (功能) fall under the same objection. Were the manifest and non-manifest states of the capacities of one nature, nothing would mark off the manifest; were they of many natures, then since the capacities are not separable from the manifestation, calling prakṛti one becomes contradictory.
“Examining the meaning expressed by the word: this is not a sense-faculty’s object-field; if this word and the like operate [on it], [it is] not the various individual sense-faculties [that do so].”
Objects of perception cannot be verbally displayed. To say “there is no separate self” is already to point at the supposed separate self, which shows that this is not something perception operates on.
“If what is stated by speech is metaphorically posited as the word-meaning, [then] the differentiated word-meaning that appears [in expression] arises initially from that [convention].”
What speech states has no real meaning of its own, since scattered phonemes have no capacity; it is established by instruction, that is, by convention, as its self-nature and condition are. Convention has to be invoked because the rival doctrines cannot make it work. So the word-meaning is upheld as the manifestation of the word-meaning connection to a hearer not yet familiar with the referent, and since word-nature and that meaning have no difference within the word-meaning itself, it is afterwards posited by the conceit of apprehending the word-meaning, unrestrained as by a hook. Even on the opponent’s own showing, discarding the individually appearing external meanings and comprehending the word-meaning is only conceptual. [translation uncertain]
“Just as, by a familiarised word, even without [an external] meaning, a cognition arises, by the connection with its own apprehended[-aspect], [it] comprehends many kinds of images severally.”
A word backed by the latent trace (習氣, vāsanā) of past familiarisation with an external meaning produces understanding as though it could deliver the meaning, as with “permanent” and things non-distinct from it. There is no external meaning, yet by connection with its own apprehended-aspect the word generates various conceptions. The example: on hearing love-poetry, those with desire generate a desirous mind, those free of desire a mind of revulsion, from the same words.
“That [word], by [excluding] the other word-meanings, is admitted [to express] the excluded meaning; knowing the connected word-meaning is not different from inference.”
As things appear separately in accordance with reality, so the word-meaning is distinct from other words and arises through comprehending the connection. Verbal cognition is therefore subsumed under inference.
“Whatever various conceptual cognitions arise from whatever [cause] — those too are self-awareness, not different from perception.”
Conception is self-awareness and so is itself perception; it is not to be split off as a further pramāṇa.
[Objection] Words with a conventional cause express their meaning by excluding other meanings; what of arbitrary words imposed at will (隨欲, yadṛcchā-śabda)?
“The arbitrary [words operate] by the difference of meaning.”
An arbitrary word such as “tittibha-fruit” (帝答塔果) is a loosely stated universal, and is stated precisely by the differentiation of that universal. Asked what separates kind-words from universal-aggregate words, the answer is: not the slightest difference. The kind-word, by force of common agreement, is in some cases metaphorically imputed onto individual parts, as calling pork “pig” in “one should not eat the settlement’s pig.”
“Some [words] express [the whole] directly, like ‘existent,’ without depending [on parts]; [some] operate [via] number, magnitude, and shape, like the many drops of water and the like, expressing [them] as combined; some operate not only on the parts; and some, like [those operating] by the differentiation of shape and visible[-colour] parts, express [via that differentiation] — [yet] the word is not seen to express [merely by] operating on those parts.”
A taxonomy with Fazun’s examples. Difference of shape: wheel, round, long, square, and likewise fist, knot, garland, ear-ornament, none stated merely by their parts. Difference of visible colour: variegated colour, flower-cushion. Difference of the whole (有支, possessor-of-parts): hundred, thousand, the measures sheng and dou, month, year, liang, none of which operate on the parts. Aggregation-words sometimes work as complete words on individual things, as with “going” and “coming.” Some are metaphorically posited, as bodily action is called “my action.” Some do not operate on the parts at all: water, cloud, forest.
[Objection] Some words display no meaning by making a connection, as with “panasa-fruit” (Xin-hui reads “this panasa,” and “this” should be present); how is anything inferred there?
“Because [the demonstrative] displays the meaning [thereby] reached.”
One knows the meaning through the word “this,” joined to a pointing finger or an already established name, not through “panasa-fruit.” The meaning that the name displays is the common locus of the name-meaning and the word “this”; both express the thing, and the connection is not manufactured into a name-meaning for its sake.
[Objection] Then the connection with “panasa-fruit” should be the object measured by the word.
“No, because the connection is [merely] metaphorically posited.”
Meaning and word are first apprehended by another pramāṇa; the judgement “this is this,” this fruit is this name, is what yields the connection with “panasa-fruit.” Being connected with the object measured, it is a case of inference. So verbal cognition is not a separate pramāṇa.
Question: how are the remaining pramāṇas, analogy (比喻量, upamāna) and the rest, not separate either?
“This [method] also accounts for the rest.”
Take analogy: understanding that a tawny ox and a wild ox (gavaya) are similar. Learn it by hearing from another and it is verbal cognition; work it out oneself and one first comprehends the two word-meanings by other pramāṇas and then their similarity by the mind. Either way the similarity is inferred, so analogy is no separate pramāṇa. Being nothing but inferential conception, it too is exclusion-of-others.
End of Volume Five of the Jiliang lun lüejie.
← 4. Example and Pseudo-example | All chapters | 6. Sham Rebuttals (jāti) →