Reverberations of Dharmakirti’s Philosophy

by Birgit Kellner | 2020 | 264,305 words

This page relates ‘Dignaga on the View of a Generic Term as Denoting a Relation’ of the study on the philosophy of Dharmakirti (6th century) and his predecessor Dignaga (5th century). This collection of articles reflects philosophical currents in India, China and Tibet during their time and investigates the Buddhist theories of Pramana (“instruments of trustworthy awareness”).

Dignāga on the View of a Generic Term as Denoting a Relation

(By Hideyo Ogawa)

I deeply respect Dr. Ole Holten Pind, whose epoch-making achievement is represented by Pind (2009, 2015). The present paper owes much to the work, without which I could not even have stood on the threshold of Dignāga’s apoha theory. Editor’s note: This contribution contains changes that the author has left to the editor’s discretion.


Like grammarians, Dignāga bases his linguistic theory, which is well known as the apoha theory, on actual usage of language. His observation of the fact that the word go “cow” is not found to be used in actual usage to convey the meaning of the word aśva “horse” leads him to build up the apoha theory, which is constructed out of three factors: a word’s own meaning (svārtha), the meaning to be conveyed by another word (arthāntara, anyaśabdārtha), and the exclusion of the latter (apoha, nivṛtti, vyāvṛtti, vyavaccheda). These three factors are connected with one another in such a way that the word go brings about the understanding of its meaning as qualified by the exclusion of the meaning of another word, say, the word aśva (arthāntaranivṛttiviśiṣṭabhāva). Dignāga assumes that from the word go one understands nothing else but the cow, and not simply the cow. When he says that anyāpoha “exclusion of others” is the meaning of a word, he wishes to imply that anyāpoha is a meaning of a word as a qualifier (viśeṣaṇa) of the word’s own meaning, or as the occasioning ground for the use of the word (pravṛttinimitta).

In the Apoha chapter of his Pramāṇasamuccaya Dignāga argues that a generic term (jātiśabda) like sat “being, existent” cannot denote an individual (bheda), a generic property (jāti), a relation between the two (yoga, sambandha), or an individual qualified by a generic property (tadvat).

In the latter half of the second kārikā, he, relying on the observation of utterances such as sad dravyam “A substance is existent,” in which the item sat is supposed to be co-referential with a co-occurring item for an individual, states his view that the generic term cannot denote a relation.

PS V.2cd:

[na jātiśabo] vācako yogajātyor vā bhedārthair apṛthakśruteḥ /

Nor does the generic term denote a relation or a generic property, since it refers to the same entity (apṛthakśruti) as [words] signifying individuals.[1]

The third kārikā of the Apoha chapter is devoted to adducing the conclusive reason for the view. The kārikā goes as follows:

PS V.3:

sambandhaś cātra sambandhidharmeṇa vācya ucyate /
tathā hi bhāvaḥ kṛtvoktaḥ, bhāvaś cānyena yujyate //

Pind 2015: II 18–19 gives the following translation of the kārikā:

And in this context it is explained that the connection is denotable through the property of the relatum (sambandhidharmeṇa vācya ucyate). That is, it [viz. the connection] is denoted on the assumption that it is a state of action (bhāvaḥ kṛtvoktaḥ); and a state of action is connected with the other [relatum] (bhāvaś cānyena yujyate).

Pind’s interpretation of the kārikā is well grounded. But I have to say that, resting even on Pind’s interpretation, it is almost impossible to fathom what Dignāga intends to say in the kārikā.

Dignāga here simply brings out the point Bhartṛhari makes about the denotation of a relation. According to Bhartṛhari, there is no nominal (nāman) that denotes a relation qua relation; a nominal such as sambandha “relation” cannot denote a relation in its own property (svadharmeṇa) but simply as a substance (dravya), because an act (bhāva) denoted by an action noun (bhāvasādhana) is treated like a substance (dravyavat).

The aim of this paper is to give a plausible interpretation of the kārikā by taking into consideration arguments Bhartṛhari brings forward about the denotation of a relation in his Vākyapadīya and thereby to show clearly the reason for Dignāga’s argument that a generic term cannot denote a relation. It will be shown that the same approach to everyday speech (lokavyavahāra) as Dignāga takes for the purpose of establishing anyāpoha as a word meaning (padārtha) is found in his arguments about the denotation of a relation.

Part 1.

Before turning to a closer examination of the question at issue, a few remarks should be made concerning Dignāga’s apoha theory. It is extremely important to note that Dignāga ac-cepts the abstraction (apoddhāra) theory which is known as forming the pivot of Bhartṛhari’s linguistic theory.

Consider PS V.46, where Dignāga introduces the concept of the abstraction of words from a sentence.

[A] PS V.46:

apoddhāre padasyāyaṃ vākyād artho vikalpitaḥ /
vākyārthaḥ pratibhākhyo’yaṃ tenādāv upajanyate //

When a word is abstracted from a sentence, this [i.e., anyāpha] is conceptually posited (vikalpita) as the meaning of the word. This sentence meaning called pratibhā “a flash of knowledge” is brought about by means of [the grasping of] that [word meaning] at the outset [when one has not familiarized oneself with the sentence].[2]

Dignāga comments on the first half of the kārikā as follows:

[B] PSV on PS V.46 (Pind 2015: I 55–56):

(a) padasyāsato 〈’pivākyād apoddhṛtasya yathāgamaṃ utprekṣayārtho vyavasthāpyate kevalasyāprayogāt prakṛtipratyayavat.[3]

(b) sā cotprekṣānyeṣv āgameṣv ayuktārthagrahaṇī . tasmād idam arthāntaram utkṣiptam,[4]

(a-1) A word or a syntactic unit (pada) is abstracted from a sentence (vākya).

(a-2) The word is unreal (asat) because it is not used in isolation [in the realm of actual communication], just like a base (prakṛti) and an affix (pratyaya) that are abstracted from a word. (a-3) Nonetheless, [a certain entity is] posited as its meaning through invention (utprekṣā) in accordance with tradition (ya-thāgamam). (b-1) And the invention based on other traditions is that through which a [word] meaning that is logically invalid (ayuktārtha) is conceived of. (b-2) Therefore, this [word] meaning [called anyāpoha], which is different from entities posited as [word] meanings by other traditions, has been brought forward here.[5]

As Pind (2015: II Appendix 14) has pointed out, [B](a) presents a striking parallel to Bhartṛhari’s Vṛtti.[6] Let us take note of (a-3): A word meaning is an invented entity (pa-dasya … utprekṣayā … artho vyavasthāpyate). The invention is conditioned by tradition (yathāgamam). There are different inventions according to different traditions. As men-tioned earlier, regarding the question of what a generic term denotes Dignāga tries to reject four views: an individual, a generic property, a relation between the two, and an individual qualified by a generic property. Entities posited as the word meaning in these views are inventions according to traditions.[7]

It is important to note [B](b-1) and [B](b-2), which clearly show that anyāpoha is posited as a word meaning through invention in accordance with the Buddhist tradition. According to Dignāga, the assumption that anyāpoha is a word meaning is more reasonable than the others. That is, this assumption is not only based on the Buddhist tradition (āgama) but also is well grounded on logical reasoning (yukti).

But in terms of what is the assumption reasonable? Consider the following Vṛtti.

[C] PSV on PS V.37b (Pind 2015: I 46):

(a) tasmād asmābhir apiloka-vyavahārā naimittikā vāpāribhāṣikābhūtārthatvena na mṛśyante,

(b) lokavad evānugamyante.

(c) siddhaś ca rūpaśabdo loke nīlādiṣv eva, na rasādiṣu.

(a) Therefore, we, too, do not cling (mṛśyante) to everyday speech (lokavyava-hāra),[8] whether it has a basis for application or comes from a scientific field, as related to a real entity (bhūtārtha). (b) Just as the world follows (anugamyante) everyday speech, so do we. (c) And in everyday usage it has been established that the word rūpa “color” denotes only (eva) blue and the like and not taste and others.

Let us take note of [C](b): Dignāga makes the point that he follows everyday speech, just as does the world. Clearly what this implies is that he holds that his apoha theory can most reasonably account for everyday speech. Moreover, when he states [C](c), he intends to imply that the apoha theory is constructed on the basis of the fact found in everyday speech such that the word rūpa denotes only (eva) blue and the like and not taste and others. It may be said without much exaggeration that Dignāga’s theory of apoha is meant for reasonably explaining the fact of speech that a certain linguistic item occurs only in the domain of a certain meaning and not in the domain of another meaning.

In [C](a) it is stated that Dignāga does not suppose that everyday speech must be related to a real entity. Naturally this suggests that he stands far apart from realism according to which language reflects reality.

Interestingly and importantly, Dignāga’s approach to established everyday usage, which is shown in [C](b), is not different from Bhartṛhari’s. Bhartṛhari asserts that he accepts things as they are spoken of, without making an ontological commitment to them.[9]

Relevant to this point is Bhartṛhari’s following remark:

[D] VP 3.3.88:

vyavahāraś ca lokasya padārthaiḥ parikalpitaiḥ /
śāstre padārthaḥ kāryārthaṃ laukikaḥ pravibhajyate //

The world (loka) achieves verbal communication by resorting to conceptual word meanings. In the science [of grammar] the word meanings commonly known to the world (laukika) are divided [into actions (kriyā), substances (dravya), qualities (guṇa), and others][10] (pravibhajyate) for the sake of [gram-matical] operations.[11]

Thus, if we accept that in constructing his linguistic theory Dignāga takes the same position as Bhartṛhari, namely that the usage common in the world is taken as the standard that is to be followed, we may say the following: Dignāga conceptually posits anyāpoha as the word meaning in order to establish a linguistic theory that holds up in the universe as the world speaks of it and believes it to be real. The universe as such is not spoken of, it is only a conceptual construct.

Part 2.

Now let us consider how Dignāga argues against the assumption that a relation between a generic property and its bearer is denoted by a generic term. Here I will consider the Vṛtti on PS V.3.

To begin with, I will give the text of the Vṛtti and its translation by Pind.

[E] PSV on PS V.3:

sambandhanaṃ hi sambandhaḥ. 〈so’nyena yujyate rāgādivat〉. tasmāt sambandhidharmeṇa sambandho vācya iti kṛtvāśaṅkitaṃ svadharmeṇa tu nāsti sambandhasya vācakaḥ śabda ity idaṃ tat prati nāsti. ato naivasya (sic) jātiśabdena vācyatvam upapadyate.

For connection means “state of connecting;” it [namely the state of connecting] is connected to the other [relatum] in the same way as the state of colouring (rāgādivat), etc. Therefore, assuming (iti kṛtvā) that the connection is denotable through the property of the relatum, doubt (āśaṅkitam) about (prati) the claim (idan tat) that a word, on the other hand, does not denote the connection by its own property (svadharmeṇa) is meaningless (nāsti). Consequently (ataḥ) its [viz. the connection’s] denotability (vācyatvam) by a general term is not at all (naiva) justified (upapadyate). (Pind 2015: II 20)

2.1. Let us first look at the following introductory sentence.

[F] PSV on PS V.3:

(a) sambandhanaṃ hi sambandhaḥ /

2.1.1. This statement is meant to give a vigraha, i.e., “an utterance which paraphrases the meaning of an integrated form (vṛtti),”[12] of the term sambandha, which shows that this term is bhāvasādhana, that is, the item that is formed from a verb with a kṛt affix signifying bhāva.[13] The term sambandha is derived from the verb bandh preceded by the preverb (upasarga) sam with the kṛt affix ghañ,[14] meaning “connecting, connection, relation.” The term sambandhana is derived from the same string with the kṛt affix lyu.[15] Jinendrabuddhi properly notes that the term sambandha is an item that denotes bhāva, or what is called an action noun.[16]

In Pāṇini’s grammar, bhāva as signified by a kṛt affix is to be distinguished from bhāva as signified by an L-affix (as in āsyate devadattenaDevadatta is sitting:” A 3.4.69 laḥ karmaṇi ca bhāve cākarmakebhyaḥ): the former is an act in the abstract, something that has been brought to accomplishment (siddha), termed bāhyabhāva “external bhāva,” while the latter is an act in process, which is the significand of a verbal base itself (prakṛtyartha) and hence something that is to be brought to accomplishment (sādhya), termed ābhyantarabhāva “internal bhāva.”[17]

2.1.2. The external bhāva is denoted by an item ending in a kṛt affix (kṛdanta), while the internal bhāva is denoted by an item ending in a verbal ending (tiṅanta). Both bhāvas have in common the property of being an action.

They have the following distinguishing characteristics:[18]

1) kṛdabhihito bhāvo dravyavad bhavati /

2) The external bhāva behaves like a substance (dravya). It functions as what brings an action to accomplishment (sādhana, kāraka).

3) tiṅabhihitena bhāvena kālapuruṣopagrahā abhivyajyante /

4) Time (kāla), person (puruṣa), and aspect (upagraha) are manifested by the internal bhāva.

5) tiṅabhihito bhāvaḥ kartrā samprayujyate /

6) The internal bhāva is connected with an agent.

7) ayam api viśeṣaḥ syāl liṅgakṛtaḥ saṅkhyākṛtaś ceti /

8) The external bhāva is capable of being connected with gender and number.

Thus what Dignāga means by the statement in question ([F]) is clear: The sambandha in question is a relation as signified by the action noun sambandha.

2.2. Next let us take up the second sentence.

[G] PSV on PS V.3: (b)

so’nyena yujyate rāgādivat〉.

2.2.2. In mentioning the term rāga, Dignāga might have had the following rule in mind.

[H] A 6. 4.27

ghañi ca bhāvakaraṇayoḥ //[19]

By this rule we have rāga “coloring, dye” as follows: ranj-ghañ raj-arāj-a (A 7.2.116) → rāga (A 7.3.52).[20] Rule A 3.3.121 halaś ca accounts for raṅga “paint, place of public amusement,” in which ghañ signifies an instrument (karaṇa) or a locus (adhikaraṇa) and which is used as an appellation (saṃjñā) to denote a particular thing. Thus we may say that the item rāga is an apt example of an action noun.

2.2.2. The expression so’nyena yujyate must be separated from the phrase rāgādivat, which has to be construed with the previous sentence. The expression corresponds to PS V.3d: bhāvaś cānyena yujyate, which Jinendrabuddhi glosses as bhāvaś cānyena samba-ndhinā yujyate.[21] The act of connecting in the abstract, which is denoted by the nominal sambandha that is an item ending in the kṛt affix ghañ, is to be treated like a substance. Thus what Dignāga means by the expression is that the entity relation denoted by the nominal sambandha is related to the other relatum through another relation, which implies that a relation (sambandha) itself becomes a relatum (sambandhin): to be precise, a relation behaves like a relatum.

2.2.3. In the situation in which a relation behaves like a relatum, the term which denotes such a relation is a sambandhiśabda “relative term.” At this point, it is important to note what a sambandhiśabda is. According to Patañjali, the item called sambandhiśabda implies (ākṣipati) a relatum. One uses the utterance mātari vartitavyam “One should obey a mother” instead of svasyāṃ mātari vartitavyam “One should obey one’s own mother.” The word mātṛ “mother” is a sambandhiśabda. By virtue of a relation, the mother is understood to be the mother of a certain person.[22] Similarly, when the term sambandha is used, a relation signified by this term behaves like a relatum, so that it implies a relatum on the basis of another relation.[23]

2.3. The next step is to consider the remaining passage.

[I] PSV on PS V.3:

(c) tasmāt sambandhidharmeṇa sambandho vācya iti kṛtvāśaṅkitaṃ svadharmeṇa tu nāsti sambandhasya vācakaḥ śabda iti idaṃ tat prati nāsti. ato naivasya jātiśabdena vācyatvam upapadyate.

2.3.1. Let us note the statements sambandhidharmeṇa sambandho vācyaḥ and svadharme-ṇa nāsti sambandhasya vācakaḥ śabdaḥ. Strangely enough, Dignāga and Jinendrabuddhi do not specify what the terms sambandhidharma and svadharma mean, which seems to be self-explanatory for them.[24]

In this connection it will be useful to consider what Bhartṛhari says in the Sambandhasamuddeśa.

[J] VP 3.3.4:

nābhidhānaṃ svadharmeṇa saṃbandhasyāsti vācakam /
atyantaparatantratvād rūpaṃ nāsyāpadiśyate //

There is no term which denotes a relation qua relation [lit. “a relation as characterized (upalakṣita) by its own property”].[25] Since a relation is absolutely dependent, its form is not spoken of [by any particular nominal term].

A relation’s own property (svadharma) is that of being absolutely dependent (atyantapara-tantratva). According to Bhartṛhari, a power (śakti) and a quality (guṇa) are of a dependent nature; a relation is a power that even powers have and a quality that even qualities have.[26]

2.3.2. It is proper to say that an entity cannot at the same time both be a relation (sam-bandha) and not be a relation, or be a relatum (sambandhin). We must say accordingly that in a certain context a relation has a property which is like that of a relatum.[27] In addition, we have seen that a relation’s own property is that of being absolutely dependent. What is meant by the term sambandhidharma is just opposite to this property. In order to make these points clear, it is useful to consider the Vṛtti on VP 2.439.

On the assumption that contact (saṃyoga) and inherence (samavāya), which are relational concepts of the Vaiśeṣika school, are a relation, Bhartṛhari says:

[K] VP 2.439:

sambandhidharmā saṃyogaḥ svaśabdenābhidhīyate /
sambandhaḥ samavāyaś ca sambandhitvena gamyate //

Contact, even if it is a relation, is conveyed as something whose property (dharman) is like that of a relatum by its own word [i.e., the word saṃyoga]. Inherence, even if it is a relation, is also understood as a relatum [from its own word samavāya].[28]

In his Vṛtti Bhartṛhari comments as follows:

[L] Vṛtti on VP 2.439 (312.17–22):

(a) aṅgulyoḥ saṃyoga iti sambandho’pi saṃyogaḥ saṃvijñānapadena svatantro’bhidhīyamānaḥ puruṣādivat samba-ndhidharmābhidhīyate /

(b) sambandhisambandhas tu ṣaṣṭhyā nimittatvāya kalpate /

(c) yadi śāstrāntaradarśanam avaśyam abhyupagantavyam etad āśrīyate saṃyogasamavāyaṣaṣṭhīviṣaya (read: saṃyogasamavāyaḥ ṣaṣṭhīviṣaya) iti /

(d) samavāye’pi ca saṃyogasaṃyoginoḥ samavāya iti sati vyapadeśe samba-ndhitvena samavāye’bhidhīyamāne bhavitavyaṃ sambandhāntareṇa /

(e) yadi tu śāstravyapadeśa evāyaṃ lokavyavahārānupātī tatra pratipādanā-rtham upacāreṇa saṃvyavahāraḥ kriyate //

(a) In the utterance aṅgulyoḥ saṃyogaḥ “There is contact related to two fingers,” contact (saṃyoga), even if it is a relation, is conveyed as something independent (svatantra) by a saṃvijñānapada “a conveying word” like saṃyoga,[29] just like the servant in the utterance rājñaḥ puruṣaḥ “the king’s servant.” The contact is conveyed here as that which has the property that is like the property of a relatum (sambandhidharman).

(b) In the same utterance, on the other hand, the relation with the contact as a relatum becomes the cause for the occurrence of a sixth-triplet ending.

(c) If another philosophical doctrine is necessarily to be accepted, this is admitted: [in the utterance in question] inherence-relation with the contact enters the domain of [A 2.3.50 ṣaṣṭhī śeṣe that provides for] the occurrence of a sixth-triplet ending.[30]

(d) And in the case in which the expression saṃyogasaṃyoginoḥ samavāyaḥ “There is inherence related to the contact and its bearer” is used of the inherence, too, there must be another relation since the inherence is conveyed as a relatum.

(e) If, however, this very expression used in a scientific field follows everyday speech, the inherence is metaphorically spoken of as a relatum in order to afford an understanding of it.[31]

In [L](a) it has been shown that contact, if it is denoted by the term saṃyoga, is conveyed as something independent (svatantra). The property of a relatum (sambandhidharma) is that of being independent. In addition, in [L](e) it is stated that in the utterance saṃyogasaṃyoginoḥ samavāyaḥ the inherence denoted by the term samavāya is metaphorically (upacāreṇa) spoken of as a relatum.[32]

Now we are in a good position to give an interpretation of statement A: sambandhidha-rmeṇa sambandho vācyaḥ and statement B: svadharmeṇa nāsti sambandhasya vācakaḥ. Recall that the third-triplet ending both in sambandhidharmeṇa and in svadharmeṇa can be taken as itthambhūtalakṣaṇa-tṛtīyā (A 2.3.21).[33]

Thus, A and B respectively mean:

A: “A relation as characterized by the property of a relatum is to be denoted [by a generic term];”

B: “There is no nominal which denotes a relation as characterized by its own property.”

2.3.3. In PSV on PS V.3 (c) ([I]) as it is given by Pind, we find the expression idaṃ tat prati, which can be deleted from the text.[34]

We must note the following comments by Jinendrabuddhi:

[M] PSṬ Ms B 195a5–195b1 (Pind 2015: I 4, n. 15): [(a)]

[sambandhanaṃ hi sambandha iti … rāgavat … tasmāt sambandhidharmeṇa sambandho vācya iti kṛtvāśaṅkitam iti … svadharmeṇa tv iti. svarūpeṇa nāsti sambandhasya vācakaḥ śabda iti.]

[(b)] idaṃ tad iti pratyavamarśāyogyarūpatvenāsattvabhūtatvāt svarūpābhi-dhānaṃ praty āśaṅkaiva nāsti. ato naivāsya jātiśabdena vācyatvam upapa-dyata iti.

((a) omitted.) (b) A relation is not a substance (sattva) since it is incapable of being reflexively grasped as “this” or “that;” therefore, there never arises a suspicion that [a generic term] denotes a relation per se.

The terms dravya, sattva, and vastu are synonymous with one another in the context of grammar.[35]

According to Helārāja, a substance (dravya) has the following properties (dravyadharma):

1) the property of being capable of being reflexively grasped as “this” or “that” (idaṃ tad iti pratyavamarśayogyatvam);

2) the property of having been brought to completion (pariniṣpannatā);

3) the property of being independent (svātantrya);

4) the property of taking on a fixed gender and number (liṅgasaṃkhyāyoga).[36]

The first property is what Bhartṛhari explains in the Bhūyodravyasamuddeśa.

[N] VP 3.4.3:

vastūpalakṣaṇaṃ yatra sarvanāma prayujyate /
dravyam ity ucyate so’rtho bhedyatvena vivakṣitaḥ //

That object, with reference to which a pronominal that refers to an entity (vastu) is used, is called dravya when it is intended to be conveyed as something to be differentiated (bhedyatvena vivakṣitaḥ).[37]

The point is that whatever the pronominals idam or tad, grammatically assigned the name sarvanāman,[38] can refer to is called a substance (dravya).[39]

Part 3.

Now let us turn to PSV on PS V.3. I will propose the following reading of the text.

[O] PSV on PS V.3:

(a) sambandhanaṃ hi sambandhaḥ / 〈rāgādivat / so’nyena yujyate〉 /

(b) tasmāt sambandhidharmeṇa sambandho vācya iti kṛtvāśaṅkitam /

(d) svadharmeṇa tu nāsti sambandhasya vācakaḥ śabda iti nāsti /

(e) ato naivāsya jātiśabdena vācyatvam upapadyate //

(a) Indeed, the word sambandha is a word derived in the sense of an action in the abstract, such as the word rāga “coloring” and the like, meaning “relating, relation.”

(b) The [relation as denoted by the word sambandha, forming one relatum], is connected with the other [relatum].

(c) Therefore, assuming that a relation as characterized by the property of a relatum can be denoted [by the word sambandha], the proponent has a suspicion that a generic term denotes a relation.

(d) However, considering there is no nominal which denotes a relation as characterized by its own property, such a suspicion does not arise.

(e) Accordingly, it is absolutely improper to say that the relation is to be denoted by a generic term.

On the basis of this text of the Vṛtti and its interpretation, I will also propose the following reading of the third kārikā.

[P] PS V.3:

sambandhaś cātra sambandhidharmeṇa vācya ucyate /
tathā hi bhāva ity ukto bhāvaś cānyena yujyate //[40]

The question arises: what does the word atra “here” in pāda a refer to? Consider the immediately preceding passage.

[Q] PSV on PS V.2 (Pind 2015: I 3–4):

(a) tathā hisad dravyam, san guṇaḥ, sat karmeti bhedārthair dravyādiśabdaiḥsāmānādhikaraṇyaṃ na syāt. tac ca dṛṣṭam.

(b) na hi sattātadyogo vādravyaṃ guṇo vā bhavati, kiṃ tarhi, dravyasya guṇasya vā.

(c) āha ca:

vibhaktibhedo niyamād guṇaguṇyabhidhāyinoḥ
sāmānādhikaraṇyasya prasiddhir dravyaśabdayoḥ

(a) To explain. [If a generic term denoted a relation or a generic property,] the generic term sat could not be co-referential with words like dravya “sub-stance” which signify individuals in the utterances sad dravyam “A substance is existent,” san guṇaḥ “A quality is existent,” and sat karma “An action is existent;” but, in reality, this is observed.

(b) Indeed, the generic property being or the relation to it is not a substance or a quality. [Question] Then what? [Answer] They are something related to a substance or a quality.[41]

(c) Thus he [i.e., Bhartṛhari] says:

“The items that denote respectively something principal (guṇin) and something subsidiary (guṇa) take different nominal vibhaktis[42] because of a restriction (niyama) [such that a sixth-triplet ending occurs only after an item whose significand serves as something subsidiary].[43] Two words, each of which denotes a substance, are established to be co-referential with each other[, so that the principal-subsidiary relation is not known from an utterance consisting of the two words].”[44]

Note the statement “They are something related to a substance or a quality” (dravyasya guṇasya vā) [Q(b)]. What is meant by this statement is that when the term sat denotes the generic property being or a relation, it becomes a sambandhiśabda because a generic property implies its bearer (jātimat) and because a relation implies its relatum. This is how one has to have the utterance *sad dravyasya which is paraphrased as sattā dravyasya or sambandhaḥ (yogaḥ) dravyasya.

It is important to note that Dignāga quotes VP 3.14.8. In rājñaḥ puruṣaḥ “the king’s servant,” the significand of the nominal base rājan “king” is a qualifier (viśeṣaṇa) with respect to the significand of the nominal base puruṣa: what is a qualifier is subsidiary to what is a qualificand (viśeṣya). In the utterance sambandhaḥ dravyasya “the relation of the substance” the relation behaves like a substance and serves as a qualificand with respect to the substance, just as the quality white does in the utterance paṭasya śuklaḥ “the white of the cloth.”

Thus in the given context the word atra refers to the utterance *sad dravyasya that is equivalent to sambandhaḥ dravyasya. On this assumption, therefore, we can interpret PS V.3 as follows:

In addition (ca), it is said (ucyate) that, in this utterance [*sad dravyasya as paraphrased by sambandhaḥ (yogaḥ) dravyasya] (atra), a relation as character-ized by the property of a relatum (sambandhidharmeṇa sambandhaḥ) is to be denoted (vācya) [by the generic term]. That is to say (tathā hi), a relation, on the assumption that it is bhāva [i.e., what is denoted by an action noun such as sambandha],

is said [to be something to be denoted by a generic term] (bhāva ity uktaḥ); but (ca), [a relation as] bhāva [i.e., a relation as denoted by such an action noun, forming one relatum,] is connected (yujyate) with the other [relatum] (anyena).

What is crucial for understanding the present kārikā is that a generic term, insofar as it is a nominal (nāman), cannot denote a relation qua relation and that a relation which is denoted by a nominal cannot claim to be a relation per se.

4. Conclusion

A relation is posited as a word meaning through invention in accordance with some tradition. For Dignāga, who accepts the apoddhāra theory, a relation, which is a conceptual construct, is posited as the meaning of a nominal such as sambandha on the one hand and as a meaning of a sixth-triplet ending on the other. As is clear from [O], he accepts that a relation is a meaning of the sixth-triplet ending. In the everyday world one not only uses the sixth-triplet ending to convey a relation but also speaks of a relation by using the nominal. He is aware that one has expressions such as aṅgulyoḥ saṃyogaḥ and saṃyogasaṃyoginoḥ samavāyaḥ as well as those such as rājñaḥ puruṣaḥ. According to Dignāga, however, the same observation of everyday usage reveals that the nominal cannot denote a relation qua relation. Thus a generic term, being a nominal, has no possibility of denoting the relation per se. We must pay deep attention to the fact that the basis for his arguments about a relation is everyday speech, just as it is the basis for his arguments for positing anyāpoha as the word meaning. It is no accident that he enters into the arguments about relations by grammatically analyzing the term sambandha, which is commonly used in everyday speech. Dignāga knows that grammar is grounded on everyday speech.

References and abbreviations

A Pāṇini’s Aṣṭādhyāyī. Appendix III (Aṣṭādhyāyīsūtrapāṭha). In: Cardona 1997.

Cardona 1997 G. Cardona, Pāṇini, His Work and Its Traditions. Volume I: Background and Introduction. Delhi 1997.

Cardona 1999 G. Cardona, Approaching the Vākyapadīya. Journal of the American Ori-ental Society 119.1 (1999) 88–125.

Hideyo Ogawa 329

Houben 1995 J. E. M. Houben, The Saṃbandha-samuddeśa (Chapter on Relation) and Bhartṛharis Philosophy of Language: A Study of Bhartṛhari Saṁbandha-samuddeśa in the context of the Vākyapadīya with a translation of Helārājas commentary Prakīrṇa-prakāśa. Groningen 1995.

Iyer 1971 K. A. Subramania Iyer, The Vākyapadīya of Bhartṛhari: Chapter III. pt. i; English Translation. Poona 1971.

Iyer 1974 K. A. Subramania Iyer, The Vākyapadīya of Bhartṛhari: Chapter III. pt. ii; English Translation with Exegetical Notes. Delhi 1974.

Iyer 1977 K. A. Subramania Iyer, The Vākyapadīya of Bhartṛhari: Kāṇḍa II; English Translation with Exegetical Notes. Delhi 1977.

MBh Patañjali’s Vyākaraṇamahābhāṣya: (a) Śrībhagavatpatañjaliviracitaṃ Vyākaraṇama-hābhāṣyam (Śrīkaiyaṭakṛtapradīpena nāgojībhaṭṭa-kṛtenabhāṣyapradīpoddyotena ca vibhūṣitam), ed. Vedavrata. 5 vols. Gurukul Jhajjar (Rohatak) 1962–63. (b) The Vyākaraṇa-mahābhāṣya of Patañjali, edited by F. Kielhorn, third edition, revised and furnished with additional readings, references and select critical notes by K. V. Abhyankar. 3 vols. Poona 1962–72. (1: 1962; 2: 1965; 3: 1972.) [References of the text of the Mahābhāṣya are to volumes, pages, and lines of MBh (b).]

NC Siṃhasūri’s Nyāyāgamānusāriṇī nayacakravālavṛtti: Dvādaśāraṃ Nayacakraṃ of Ācāryā Śrī Māllavādi Kṣamāśramaṇ; with the Commentary Nyāyāgamānusāriṇī of Śrī Siṃhasūri Gaṇi Vādi Kṣamāśramaṇa, Part II (5–8 Aras), ed. Muni Jambūvijayajī. Bhāvnagar 1976.

Ogawa 2009 H. Ogawa, On Bhartṛhari’s notion of ‘power’ (śakti). In: Bhartṛhari: Lan-guage, Thought and Reality (Proceedings of the International Seminar, Delhi, De-cember 12–14, 2003), ed. M. Chaturvedi. Delhi 2009, 225–252.

Ogawa 2010 H. Ogawa, Bhartṛhari on unnamable things. In: Logic and Belief in Indian Philosophy, ed. P. Balcerowicz. Delhi 2010, 403–418.

Ogawa 2017 H. Ogawa, The qualifier-qualificand relation and coreferentiality in Dignāga’s apoha theory. In: Reading Bhaṭṭa Jayanta on Buddhist Nominalism, ed. P. McAllister. Vienna 2017, 83–152.

Ogawa forthcoming H. Ogawa, On a bias for doxographical accounts in later commen-taries on the Vākyapadīya of Bhartṛhari: With special reference to pratibhā. (Forth-coming.)

Paddhati Vṛṣabhadeva’s Paddhati: See VP (c).

Pind 2009 O. H. Pind, Dignāga’s philosophy of language: Dignāga on anyāpoha; Pramā-ṇasamuccaya V, texts, translation, and annotation. PhD diss., Universität Wien. Wien 2009., last visited 16-05-2016.

Pind 2015 O. H. Pind, Dignāgas Philosophy of Language. Dignāga. Pramāṇasamucca-yavṛtti V on anyāpoha. Part I: Text. Part II: Translation and Annotation, ed. E. Steinkellner. Vienna 2015.

Pradīpa Kaiyaṭa’s Pradīpa: See MBh (a). Prakāśa Helārāja’s Prakāśa: See VP (b) and VP (d). PS Dignāga’s Pramāṇasamuccaya: See Pind 2015.

PSṬ Jinendrabuddhi’s Pramāṇasamuccayaṭīkā: See Pind 2009.

PSV Dignāga’s Pramāṇasamuccayavṛtti: See Pind 2009.

PWT See Cardona 1997.

SK Bhaṭṭojī Dikṣita’s Vaiyākaraṇasiddhāntakaumudī: Śrī-bhaṭṭojī-dīkṣita-viracitā vaiyā-karaṇa-siddhānta-kaumudī śrīmadvāsudeva-dīkṣita-praṇītayā bālamanoramākhya-vyākhyayā śrīmaj-jñānendra-sarasvatī-viracitayā tattva-bodhiny-ākhya-vyākhyayā ca sanāthitā, ed. Giridhara Śarmā Caturveda and Parameśvarānanda Śarmā Bhāskara. 4 vols. Varanasi 1958–61.

Ṭīkā Punyarāja’s Ṭīkā. See VP (e).[References are to pages and lines.]

VP Bhartṛhari’s Vākyapadīya: (a) Bhartṛharis Vākyapadīya: Die Mūlakārikās nach den Handschriften herausgegeben und mit einem Pāda-Index versehen, ed. W. Rau. Wies-baden 1977. [References to kārikās of the Vākyapadīya are made according to Rau 1977, so that kārikā numbers are given according to his edition.] (b) Vākyapadīya of Bhartṛhari with the Commentary of Helārāja, Kāṇḍa III, Part 1, ed. K. A. Su-bramania Iyer. Poona 1963. (c) Vākyapadīya of Bhartṛhari with the Commentaries Vṛtti and Paddhati of Vṛṣabhadeva, ed. K. A. Subramania Iyer. Poona 1966. (d) Vākyapadīya of Bhartṛhari with the Commentary of Helārāja, Kāṇḍa III, Part II, ed. K. A. Subramania Iyer. Poona 1973. (e) The Vākyapadīya of Bhartṛhari (An Ancient Treatise on the Philosophy of Sanskrit Grammar), Containing the Ṭīkā of Puṇyarāja and the Ancient Vṛtti, Kāṇḍa II, with a Foreword by Ashok Aklujkar. Delhi 1983.

Vṛtti Vākyapadīya Vṛtti. See VP (c) and VP (e). [References to kārikās, pages, and lines.]

Footnotes and references:


Pind 2015: II 14 (§ 4): “(vācakaḥ) neither the general property nor the inherence relation, because it is not ‘heard apart’ (apṛthakśruteḥ) from [words] whose referents (bhedārthaiḥ) are particular [general properties].”


Pind 2015: II 166 (§ 61): “the referent of the syntactical word (padasya) is imagined (vikalpitaḥ) when abstracted (apoddhāre) from the sentence (vākyāt). Yet the referent of the sentence which is called intuition (pratibhā) is in the beginning (ādau) produced by that [namely the syntactic word].” According to Jinendrabuddhi, the pronominal ayam “this” in pāda a refers to anyāpoha. PSṬ Ms B 236b1 (Pind 2015: II 166, n. 554): ayam ity anyāpohaḥ. I shall waste no words on Dignāga’s view that pratibhā is the sentence meaning. This view, needless to say, is borrowed from Bhartṛhari. The details of Bhartṛhari’s concept of pratibhā are to be left to Ogawa (forthcoming).


Pind 2015: II 167–168 (§ 61): “Even though the syntactical word is unreal (asat) as abstracted from the sentence, its referent is determined by invention (utprekṣayā) according to the [grammatical] tradition, because it is not used in isolation (kevalasyāprayogāt) in the same way as a stem and an affix (prakṛtipratyayavat) [are not used in isolation].”


Pind 2015: II 169–170 (§ 61): “And this invention apprehends a referent that is not justified (ayuktārthagrahaṇī ) in other traditions. Therefore this different referent (arthāntaram) has been brought forward (utkṣiptam).”


The phrase arthāntara (lit. “another meaning”) refers to a word meaning that is reasonable and that is called anyāpoha. PSṬ Ms B 237a7: yuktimad … anyāpohākhyam. See Pind 2015: II 169, n. 566.


Vṛtti on VP 1.24 (65.1–4): tatrāpoddhārapadārtho nāmātyantasaṃsṛṣṭaḥ saṃsargād anumeyena pari-kalpitena rūpeṇa prakṛtavivekaḥ sann apoddhriyate / pravivaktasya hi tasya vastuno vyavahārātītaṃ rūpam / tat tu svapratyayānukāreṇa yathāgamaṃ bhāvanābhyāsavaśād utprekṣayā prāyeṇa vyavasthā-pyate / In this passage Bhartṛhari makes the following points. Of the eight topics to be dealt with in the Vākyapadīya (aṣṭaka), what is called apoddhārapadārtha “abstracted word meaning” is absolutely fused in a single whole meaning (atyantasaṃsṛṣṭa). The word meaning is abstracted from a putative complex (saṃsarga) on the basis of a form that is to be inferred and that is conceptually assumed, so that it is distinguished from other abstracted word meanings. Such a partial word meaning which is distinguished from other partial word meanings has a form that is outside the realm of actual communication. Gram-marians posit (vyavasthāpyate) such an assumed form (parikalpita) through invention (utprekṣayā), in accordance with what they have understood (svapratyayānukāreṇa) following traditions (yathāgamam) and exposing themselves to them so repeatedly as to have their latent traces (bhāvanābhyāsa). See Cardona 1999: 96–98 for a full account of the points made by Bhartṛhari here.


Jinendrabuddhi says: PSṬ Ms B 237a1–2 (Pind 2015: II 168, n. 561): yasya (scil. padasya) ya āgamaḥ: keṣāṃ cid bhedā vācyāḥ, pareṣāṃ jātir, anyeṣāṃ sambandhaḥ, keṣāṃ cit tadvad iti padasyārtha ity āgamaḥ. (“That which is the tradition as to a word is the tradition such that this is the meaning of a word: some have a tradition that individuals (bheda) are to be denoted by words; some have a tradition that a generic property is to be denoted by a word; others have a tradition that a relation between a generic property and its bearer is to be denoted by a word; some have a tradition that an individual qualified by a generic property is to be denoted by a word.”)


On the meaning of mṛśyante, see Pind 2015: II 142, n. 468.


In the Sādhanasamuddeśa Bhartṛhari says of the power of functioning as agent (kartṛtva) the following: VP 3.7.38: tattve vā vyatireke vā vyatiriktaṃ tad ucyate / śabdapramāṇako lokaḥ sa śāstreṇānugamyate // (“The property of being an agent, whether it be identical with its bearer or different from the latter, we declare that it is different from its bearer. The world holds usage as their standard. The science of grammar follows them.”) Iyer 1971: “As to whether it (power) is identical with or different from (its substratum), it is held that it is different. For the world, the word is the authority and the śāstra (that is, Grammar) follows it too.” The point made here is that, without examining the ontological question of whether the agent-power is identical with or different from its bearer, grammarians favor the view commonly accepted by the world that the agent-power is different from its bearer. I need not elaborate on Bhartṛhari’s concept of power here. See Ogawa 2009.


Prakāśa on VP 3.3.88 (181.11–12): -śāstrīyakāryaprasiddhyarthaṃ laukika eva kriyādravyaguṇādilakṣaṇo’poddhārapadārthaḥ pravibhajyata iti … /


Iyer 1971: 119: “Verbal communication in the world is done with meanings of words fashioned by the mind and in the science of grammar, it is the meanings adopted in the world on the basis of which the work (of explaining the forms) is done.” Houben 1995: 321; 424: “But verbal usage of everyday life [takes place] by means of conceptualized meanings of words. In grammar (śāstra) the word meaning of everyday usage is separated [from the sentence meaning] for the sake of grammatical operations.” (Houben’s interpretation of pravibhajyate “… is separated [from the sentence meaning]” is far-fetched.) According to Helārāja, that which is framed out of word meanings (padārthaprakriyā) and which is well established in the world is to be accepted as it is in the science of grammar. Prakāśa on VP 3.3.88 (181.9–10): tad atra lokaprasiddhaiva padārthaprakriyā samāśrayaṇīyā … / Note that the word laukika is derived from loka with the taddhita ṭhañ by A 5.1.44 lokasarvalokāṭ ṭhañ, signifying “known in the world” (“known in X” [tatra viditaḥ]).


SK, sarvasamāsaśeṣaprakaraṇa (II.215): kṛttaddhitasamāsaikaśeṣasanādyantadhāturūpāḥ pañca vṛtta-yaḥ / parārthābhidhānaṃ vṛttiḥ / vṛttyarthāvabodhakaṃ vākyaṃ vigrahaḥ / MBh on A 1.3.1 (I.256.20): kaḥ punar bhāvaḥ / bhavateḥ svapadārtho bhavanaṃ bhāva iti / Pradīpa on MBh to A 1.3.1 (II.186): bhavanam bhāva iti vigrahaḥ kartṛsādhananivṛttyarthaḥ /


To adduce a typical example of a vigraha for a bhāvasādhana term, I will cite from the Bhāṣya. MBh on A 3.1.22 (II.28.13–14): samabhihāra iti ko’yaṃ śabdaḥ / samabhipūrvād dharater bhāvasādhano ghañ / samabhiharaṇaṃ samabhihāraḥ / The term samabhihāra, used in A 3.1.22, signifies the act of performing repeatedly, intensively, or repeated, intensive perfomance.


Here and in the following, bold letters are used to signify the anubandhas.


sam-bandh-ghañ (A 3.3.18); sam-bandh-lyu(A 3.3.115) → sam-bandh-ana (A 7.1.1).


PSṬ Ms B 195a5 (Pind 2015: II 20, n. 44): bhāvatvam asya darśayati. Grammatical rules we have to take into consideration here are the following: A 3.3.18 bhāve and A 3.3.115 lyuṭ ca. A 3.3.18 and A 3.3.115 provide respectively for introducing ghañ and lyuṭ after a verb on condition that an act in the abstract (bhāva) is to be signified.


MBh on A 3.3.19 (II.145.16–18): nanu coktaṃ vihitaḥ pratyayaḥ svārthe bhāve ghañ itīti / anyaḥ sa bhāvo bāhyaḥ prakṛtyarthāt / anenedānīm ābhyantare bhāve syāt / kaḥ punar etayor bhāvayor viśeṣaḥ / ukto bhāvabhedo bhāṣye // See PWT 297. In the Kriyāsamuddeśa Bhartṛhari states: VP 3.8.47: ākhyātaśabde bhāgābhyāṃ sādhyasādhanavartitā / prakalpitā yathā śāstre sa ghañādiṣv api kramaḥ // (“In grammar, just as, with reference to a finite verb form, its divisions are fictitiously assumed respectively to denote something to be brought to accomplishment and something to bring it to accomplishment, so, with reference to items ending in affixes like ghañ also, the same analysis is given”); VP 3.8.48: sādhyatvena kriyā tatra dhāturūpanibandhanā / sattvabhāvas tu yas tasyāḥ sa ghañādinibandhanaḥ // (“In those items ending in affixes like ghañ, an action is conveyed as something to be brought to accomplishment on the basis of a verb, while the action’s status of being a substance is conveyed on the basis of affixes such as ghañ”); Iyer 1974: 27–28: “Just as Grammar divides a verb into two parts, one expressive of a process (sādhya) and the other of a thing (sādhana), the same can be done to a word ending in a primary suffix like ghañ;” “The expression of action as a process depends upon the root and its aspect as a thing depends for its expression on suffixes like ghañ.”


In the Bhāṣya on A 3.1.67 Patañjali states: MBh on A 3.1.67 (II.57.7–17): asti khalv api viśeṣaḥ kṛdabhihitasya bhāvasya tiṅabhihitasya ca / kṛdabhihito bhāvo dravyavad bhavati / kim idaṃ dravyavad iti / dravyaṃ kriyayā samavāyaṃ gacchati / kaṃ samavāyam / dravyaṃ kriyābhinirvṛttau sādhanatvam upaiti / tadvac cāsya bhāvasya kṛdabhihitasya bhavati / pāko vartata iti / kriyāvan na bhavati / kim idaṃ kriyāvad iti / kriyā kriyayā samavāyaṃ na gacchati / pacati paṭhatīti / tadvac ca asya kṛtabhihitasya na bhavati / pāko vartata iti / asti khalv api viśeṣaḥ kṛdabhihitasya bhāvasya tiṅabhihitasya ca / tiṅabhihitena bhāvena kālapuruṣopagrahā abhivyajyante / kṛdabhihitena punar na vyajyante / asti khalv api viśeṣaḥ kṛdabhihitasya bhāvasya tiṅabhihitasya ca / tiṅabhihito bhāvaḥ kartrā samprayujyate/ kṛdabhihitaḥ punar na samprayujyate / yāvatā kiṃcit sāmānyaṃ kaścit viśeṣo yuktaṃ yad ayam api viśeṣaḥ syāl liṅgakṛtaḥ saṅkhyākṛtaś ceti /


This rule provides that the penultimate n of ranj “dye, color” is deleted in a stem followed by ghañ introduced on condition that an action in the abstract (bhāva) or an instrument (karaṇa) is to be signified (A 3.3.18–19).


See PWT 446.


PSṬ Ms B 68b7 (Pind 2015: II 18, n. 41): sambandhanaṃ hi sambandha iti bhāvarūpeṇa sambandha-śabdenābhidhānāt. bhāvaś cānyena sambandhinā yujyata iti. sambandhyantarākāṅkṣā jāyate: kasya sambandha iti. tataś ca sambandhy eva sa bhavatīti na svarūpeṇa sambandho’bhidhīyate.


MBh on A 3.1.19 (II.27.14–16): sambandhiśabdāś ca punar evamātmakāḥ yaduta sambandhinam ākṣipanti / tadyathā / mātari vartitavyaṃ pitari śuśrūṣitavyam iti / na cocyate svasyāṃ mātari svasmin vā pitarīti / sambandhāc caitad gamyate yā yasya mātā yaś ca yasya piteti /


Jinendrabuddhi illustrates this point by taking as an example the term samuccaya “connection” which is supposed to convey a meaning of the particle (nipāta, A 1.4.57) ca “and.” PSṬ Ms B 195a1–3 (Pind 2015: II 19, n. 43): iha kaścic chabdārthaḥ kenacic chabdenābhidhīyamānaḥ sambandhyantarākā-ṅkṣopajanahetuḥ. tathā hi samuccayaḥ samuccayaśabdābhidhāne parākāṅkṣopajanahetur bhavati, sa eva cādyupādānaḥ kasyeti ākāṅkṣāṃ nopajanayati (em: ākāṅkṣām upajanayati Pind). bhāvaś ca bhāvasādhanena śabdenābhidhīyamāno niyatam anyaviṣayām ākāṅkṣāṃ janayatīti: (“In this everyday speech, a certain meaning of an item, when being conveyed by a certain other item, becomes the cause of awakening expectations of the other relatum. To explain. Connection (samuccaya), when being conveyed by the term samuccaya, is the cause of awakening expectations of something other than it. But the very same entity connection, when being referred to by a particle such as ca ‘and,’ does not awaken expectations as ‘connection of what?’ The act [of connecting], when being conveyed by an action noun, necessarily awaken expectations regarding something different from it.”) Pind 2015: II 19–20, n. 43: “In this case when a certain word referent is being denoted by a certain word it is the cause of generating expectation [of the complementation of] the other relatum. Like, for instance, an accumulation is the cause of generating expectation [of the complementation] of the correlate when being denoted by a word denoting accumulation. The [word denoting accumulation] comprising [the word] ‘and,’ and so on, with certainty generates expectation [of complementation] at the thought ‘[accumulation] of what’? And a state that is denoted by a word having a state [of action] as its means of realisation necessarily generates expectation concerning the other relatum.” Suffice it to give the utterances ghaṭaś ca paṭaś ca “a pot and a piece of cloth” and ghaṭasya paṭena samuccayaḥ “the connection of a pot with a piece of cloth.”


As to the term svadharma Pind (2015: II 20, n. 48) says: “The term applies to any given noun that is not subject to the grammatical operation of introducing the sixth triplet that denotes the relation …” I cannot understand his point here.


Helārāja takes the third-triplet ending of svadharmeṇa to denote a characteristic that makes known something as having acquired a certain aspect (A 2.3.21 itthaṃbhūtalakṣaṇe). Prakāśa on VP 3.3.4 (128.10–11): tatra svena asādhāraṇena dharmeṇa svabhāvenopalakṣitasya vācakaṃ pratyāyakam, abhidhānaṃ ṣaṣṭhīvyatiriktaṃ nāsti, idaṃtayā svarūpānavadhāraṇāt /


Bhartṛhari states the following kārikā. VP 3.3.5: upakārāt sa yatrāsti dharmas tatrānugamyate / śaktīnām api sā śaktir guṇānām apy asau guṇaḥ // (“Where, by virtue of one thing rendering service (upakāra) to another, there is the relation, there the property of being absolutely dependent is recognized. It [i.e., the relation] is a power even of powers; it is a quality even of qualities”); Iyer 1971: 81: “Where there is service rendered, there an attribute (that is, relation) is understood. It is the power of powers, it is the attribute of attributes”; Houben 1995: 170; 341: “Where this [relation] is, because some service is rendered [from one thing to another, or: from signifier to signified and vice versa], there one arrives at a property (viz. dependence). Even for capacities [dependent on that which has the capacity] it is a capacity, even for qualities [dependent on that which has the quality] it is a quality [so it is extremely dependent].” According to Bhartṛhari, when a certain entity is seen to render service to, i.e., help (upakāra), another, there obtains a relation between them. A power and a quality are what renders service to an entity: a power helps the entity to bring about a certain result and a quality helps the entity to be differentiated from other entities. A power is also to be served by a certain thing in order to be determined both to abide in a locus and to bring about a specific result. Such a thing is a relation that the power has with respect to its locus (āśrayāśrayibhāva) and to its result (kāryakāraṇabhāva), and the relation can consequently be viewed as a power in that it helps the power to abide in a locus and to bring about a specific result. The reason that, with regard to qualities, a relation is called a quality is that it is viewed as meant for a quality (parārtha) because it serves to determine the quality abiding in a specific substance. On these points, see Ogawa 2009.


The following kārikā is useful in considering what it means to say that an entity has a property which is like that of another entity. VP 3.3.6: taddharmaṇos tu tācchabdyaṃ saṃyogasamavāyayoḥ / tayor apy upakārārthā niyatās tadupādhayaḥ // (“But, contact (saṃyoga) and inherence (samavāya) are termed sambandha because they have the property which is like the property of that [i.e. the relation]. Even those two have specific factors which delimit that [i.e., the relation] and which serve to render service to the relation”); Iyer 1971: 81: “Conjunction and inherence are called relations because they have the attribute thereof. They have definite conditions the purpose of which is to render service;” Houben 1995: 176; 341: “As regards saṃyoga (connection) and samavāya (inherence), they (are tacchabdāḥ) are called by that word (sc. ‘relation’), because they have (as it were) that property (sc. dependence). Even these two have restricted functions of rendering service, with this [relation] as limiting factor.” A relation that is delimited by a specific factor, such as the property of residing only in a substance (dravyaikaniyatatva) or its being related to being as a quality (guṇatvena sattāsambandhaḥ), is termed saṃyoga; a relation that is delimited by a specific factor, such as the property of its relata being inseparable (ayutasiddhi), is termed samavāya. Prakāśa on VP 3.3.7 (130.5–6): tathā ca dravyaikaniyatatvaṃ guṇatvena sattāsambandhaḥ saṃyogasyopādhir iti tatropakāraḥ sambandhena tasyāvacchidyate /; Prakāśa on VP 3.3.8–11 (131.11): anena cāyutasiddhiḥ samavāyasyopādhir uktaḥ / According to Bhartṛhari, contact and inherence, which the Vaiśeṣikas posit as relations, do not fit the definition of a relation given by him. Suppose that the defining feature of a relation is to be dependent; contact and inherence have this feature with respect to substances and qualities, so that the term sambandha is used for them. The point is that contact and inherence are dependent entities but not absolutely dependent entities. They are metaphorically spoken of as relations. An important thing to note is Bhartṛhari’s use of the phrase taddharmaṇoḥ “those two whose property is like the property of that [i.e., the relation],” which is an uṣṭramukha “camel-faced” type of bahuvrīhi. MBh on A 1.1.70 (I.180.18–19): uṣṭramukham iva mukham asya so’yam uṣṭramukhaḥ / The bahuvrīhi uṣṭramukha signifies someone who has a face (mukha) like (iva) that of a camel (uṣṭra).


Iyer 1977: 187: “The relation called contact is expressed as that which is related (sambandhin) by its own word (namely, saṃyoga); similarly, inherence is also expressed as that which has inherence.”


On the concept of saṃvijñānapada, see Ogawa 2010.


On A 2. 3.50, see PWT 251.


See n. 28. [L](e) suggests that there is a twofold relatum: primary and metaphorical.


Puṇyarāja says that in the utterance saṃyogo dravyayoḥ “There is contact between two substances” contact is understood from the term saṃyoga as something that has been brought to accomplishment (pariniṣṭhitasvabhāva), that is, as a substance. Ṭīkā on VP 2.439 (173.5–10): saṃyogaśabdāt saṃyogo dravyayor iti yathāvat pariniṣṭhasvarūpaḥ [read: pariniṣṭhitasvarūpaḥ] tatra sambandhisvabhāvaḥ pratīyate, na ca tasya tadrūpam / evaṃ samavāyaśabdāt samavāyaḥ pratīyata ity āha–saṃbandhaḥ samavāya ityādi / sambandhitvena sambandhasvarūpaviraheṇa samavāya ity asmāt pratīyate / athavā yadā saṃyogaśabdaś cāsau sambandhaḥ svatantro’bhidhīyate tadā tatra samavāyaḥ sambandhatvena gamyata iti yadā sambandhatvena iti pāṭhas tadaivaṃ yojanīyam // Puṇyarāja tells us that there is a variant reading of pāda d: sambandhatvena gamyate. Given this reading, the kārikā is interpreted as arguing that in the utterance saṃyogo dravyayoḥ the inherence-relation is understood as a relation to be denoted by the sixth-triplet endings.


See n. 26.


Pind 2015: I 5, n. 15: “(3)KV do not reproduce the expression idaṃ tad prati that is to be construed with āśaṅkitam.”


Paddhati on Vṛtti to VP 1.13 (45.26–46.7): vastu iti / nāmapadavācyam / tyadādīnām iti sarvanāmopalakṣaṇam / sattvalakṣaṇānām iti / sattvaṃ dravyaṃ tal lakṣyata ebhir iti / vastūpalakṣaṇaṃ yatra sarvanāma iti dravyalakṣaṇam /


Prakāśa on VP 3.1.13 (26.14–27.1): tatra dravyadharmā idaṃ tad iti pratyavamarśayogyatvam, pariniṣpannatā, svātantryam, liṅgasaṅkhyāyogaś cetyevamādayaḥ / From the point of view of those who hold that a word denotes a substance (dravyapadārthapakṣa) Bhartṛhari says: VP 3.1.13: dravyadharmā padārthe tu dravye sarvo’rtha ucyate / dravyadharmāśrayād dravyam ataḥ sarvo’rtha iṣyate // (“In the view that a substance is a word meaning, it is stated that any word meaning has the property of a substance, by resorting to the property of a substance. Hence it is accepted that any word meaning is a substance”); Iyer 1971: 16: “In the view that substance is the meaning of words, the meaning of all words can have the characteristic of substance. As the meanings of words are susceptible of assuming the properties of substance, the latter is said to be the meaning of all words.” According to Helārāja, in the present view a substance is twofold: primary and metaphorical. Prakāśa on VP 3.1.13 (26.12–14; 27.2): yathā tava dravyavādinaḥ kecid eva śabdā mukhyadravyābhidhāyinas tadanye tūpacaritadravyābhidhāyinaḥ tathā mamāpi jātivādino mukhyāṃ jātim abhidadhati kecit, upacaritām anya iti matadvaye’pi sāmyam /… teṣāṃ śuklādiguṇeṣv apy adhyāropāt te’pi dravyasyeva dharmā yeṣām iti dravyadharmāṇaḥ /


Iyer 1971: 123: “That in reference to which a pronoun can be used is substance, presented as something to be differentiated.” Bhartṛhari introduces the concept of a speaker’s intention (vivakṣā) here. It depends on the speaker’s intention what dravya is in a given situation. The status of being guṇa and that of being dravya are not ontologically fixed; they depend on how things are spoken of.


A 1. 1.27 sarvādīni sarvanāmāni // See PWT 53.


This concept of a substance (dravya) forms a pair with that of a quality (guṇa), which Bhartṛhari defines as follows: VP 3.5.1: saṃsargi bhedakaṃ yad yat savyāpāraṃ pratīyate / guṇatvaṃ paratantratvāt tasya śāstra udāhṛtam // (“Whatever is related [to a certain thing] and differentiates the thing [from others] is regarded as guṇa when it activates the function [of differentiating], because it is something dependent. This is what has been illustrated in grammar”); Iyer 1971: 126: “Whatever rests on something else (saṃsargi) differentiates it (bhedaka) and is understood in that function (savyāpāra), is, being dependent, called ‘quality’ in the śāstra.” A quality is defined as something dependent (paratantra), from which it naturally follows that a substance related to the quality must be something independent.


I have emended bhāvaḥ kṛtvoktaḥ to bhāva ity ukto. Jambūvijayajī reconstructs this portion as bhāvīkṛ-tyocyate. NC 607: sambandho’py atra sambandhidharmavācyo’bhidhīyate / tathā bhāvīkṛtyocyate bhāvo’py anyena yujyate // Pind reports: Ms 195a5: bhāvaḥ kṛtvokta iti (Pind 2015: I 4, n. 14); Ms B 195a5: bhāva ity abhiprāyeṇoktaḥ (Pind 2015: II 19, n. 42). It is hard to justify bhāvaḥ kṛtvoktaḥ gram-matically. To be sure, in view of the pratīka given by Jinendrabuddhi, we can assume bhāvīkṛtyoktaḥ “[the relation is] said to be [a significand of the generic term] after treating it as bhāva.” But, in this case, we have a vipulā “irregular”-anuṣṭubh. With bhāva ity ukto, on the other hand, we have a pathyā “regular”-anuṣṭubh (5th, 6th, 7th syllables of pāda c: short, long, long). I gratefully acknowledge helpful discussions with Dr. Yūto Kawamura on this point.


Pind 2015: II 14–16: “That is (tathā hi), there would be no co-reference (sāmānādhikaraṇyaṃ na syāt) with words like ‘substance’ (*dravyādiśabdaiḥ) whose referents are particular [general properties] (*bhedārthaiḥ), like, for instance, ‘existent substance’ (*sad dravyam), ‘existent quality’ (*san guṇaḥ), and ‘existent action’ (*sat karma); but this is observed (tac ca dṛṣṭam). For existence (sattā) or its inherence relation (tadyogaḥ) is neither a substance (dravyam) nor a quality (guṇaḥ), but is rather (kiṃ tarhi) of a substance (dravyasya) or of a quality (guṇasya).”


Here the term vibhakti signifies a nominal ending (sup: A 1.4.104 vibhaktiś ca).


Of this restriction Bhartṛhari says the following. VP 3.7.157: dviṣṭho’py asau parārthatvād guṇeṣu vyatiricyate / tatrābhidhīyamānaḥ san pradhāne’py *upayujyate // (“Even if the relation resides in two entities (dviṣṭha), it becomes something additional in entities that are subsidiary because of being intended for others. The relation which resides in the subsidiary entity, when it is denoted [by a sixth-triplet ending], is, as something residing in the principal entity also, of use to the latter.”) Iyer: upabhujyate. I have followed Rau’s reading. Iyer 1971: 237: “Even though it (the śeṣa relation) rests on both it brings about a distinction in what are secondary because of their being subordinate to something else. Being expressed there (that is, in what are secondary), it touches what is primary also.”


VP 3.14.8. Iyer 1974: 125: “The two words expressive of the secondary and the primary have necessarily different case-endings. Where they have the same case-endings, both express substance;” Pind 2015: II 16–17: “It is, moreover, explained that (āha ca) [a word] denoting a quality and one denoting the bearer of that quality (guṇaguṇyabhidhāyinoḥ) have different case affixes (vibhaktibhedaḥ) because of a restrictive rule. However, for two words that denote a substance (dravyaśabdayoḥ) co-reference is acknowledged (sāmānādhikaraṇyasya prasiddhiḥ).” Iyer takes the terms guṇa and guṇin respectively as meaning “the secondary” and “the primary,” which I think is suited to the given context. Prakāśa on VP 3.14.8 (154.1–8): paṭasya śukla iti dravyaguṇābhidhāyipadaprayoge śābdo guṇapradhānabhā-vaḥ / vīraḥ puruṣa ityādau tu samānādhikaraṇe viṣaye dvāv api dravyaśabdau svaniṣṭhaṃ svārtham ācakṣāte / tathā ca prathamaiva / sāmarthyanibandhanas tu guṇapradhānabhāva ukto viśeṣyaṃ syād anirjñātam ityādinā / evaṃ ca saty api guṇapradhānabhāve śabdāśabdatvakṛto viśeṣaḥ sāmānādhi-karaṇyavaiyadhikaraṇyayor … // On the relation of the subsidiary and the principal between word meanings, see Ogawa 2017.

Like what you read? Consider supporting this website: