The semantics of general noun of the Mīmāṃsā is based on this theory of injunction. Understanding of the meaning of a word is also injunctive; there is something ought to be understood. Gaṅgeśa defines it by our actual understanding from the word that refers to particulars, but the Mīmāṃsā does not accept it. For the word “cow” to be universally applicable to any cow that has cowness, it cannot refer to the endless series of all particular cows.
* I would like to express my gratitude to Prof. Arindam Chakrabarti for his comments on manuscripts of this paper.
Notes
1）E.g., the beginning of Gaṅgeśa’s Jātiśaktivāda. TC, p. 558: “na ca govyaktimātram arthaḥ . . . .” 2）Cf. Mahābhāṣya on Aṣṭādhyāyī sūtra 1.2.58 and 1.2.64; Nyāyasūtras 2.2.58–69. 3）PP, p. 411. 4）Gaṅgeśa’s conclusion is stated in TC, pp. 588–589. Following the position of Nyāyasūtras, he accepts the shared physical form (ākṛti) as also a part of meaning, but I do not consider that in this paper. 5）It is not certain that this “Maṇḍana” refers to Maṇḍana Miśra, the author of Sphoṭasiddhi, etc. Gaṅgeśa cites a verse in support of Maṇḍana’s viewpoint, but that is, in fact, Kumārila’s verse in Ślokavārttika. 6）TC, pp. 578, 586–587. 7）TC, pp. 564, 578. 8）TC, p. 578. 9）Three explanations are given in TC, pp. 562–565. 10）TC, p. 567. 11）TC, pp. 572–573.
Abbreviations
TC The Tattvacintāmaṇi of Gaṅgeśa Upādhyāya. Ed. Kamakhyanath Tarkavagish. Vol. 4, pt 2. Calcutta: Asiatic Society. Reprint, Delhi: Chaukhamba Sanskrit Pratishthan, 1990.
PP Prakaraṇapañcikā. Ed. Subrahmanya Shastri. Benares: Banaras Hindu University, 1961. Bibliography
Raja, Kunjunni. 1969 (2000). Indian Theories of Meaning. 2nd ed. Adyar Library Series 91. Madras: Adyar Library and Research Centre.
Scharf, Peter M. 1996. The Denotation of Generic Terms in Ancient Indian Philosophy: Grammar, Nyāya, and Mīmāṃsā. Philadelphia: American Philosophical Society.
(This research was supported in part by JSPS KAKENHI Grant Number 15J04441) Key words meaning, artha, śakti, semantics, Nyāya, Prābhākara
（JSPS Research Fellow）
On the Theory of Phonemes Conveying
the Sentence Meaning
Saitō Akane
0. Introduction
Kumārilabhaṭṭa (6c.), a foundational philosopher of the Mīmāṃsā school, who had a huge influence on the later development of the linguistic philosophy of the sentence, claimed in the Vākyādhikaraṇa (ŚVv) of his Slokavārttika (ŚV) that the sentence meaning (vākyārtha) is conveyed by the meanings of the words that compose it. In his argument from v. 110 to v. 117, he refers to and briefly rejects the idea that it is the individual phonemes (varṇāḥ) that are responsible for conveying the sentence meaning. The doctrine that phonemes as the conveyors of sentence meaning is also criticized by Śālikanātha of the Prābhākara school of Mīmāṃsā in the Vākyārthamātṛkā of his
Prakaraṇapañcikā (PrPv), and by Vācaspati too in his Tattvabindu (TB), in which he often
uses very similar texts to those used by Śālikanātha. On the other hand, Jayantabhaṭṭa criticizes Kumārila’s doctrine from the standpoint of the Naiyāyika school in his
Nyāyamañjarī (NM) 6.2, formulating his unique argument on the relationship between
phonemes and sentence meaning.
In this paper, I shall first discuss how the Mīmāṃsakas rejected the theory that it is the phonemes that convey the meaning of sentences (the varṇa theory). I shall then compare their criticism with Jayanta’s argument in favour of the theory.
1. Kumārila’s Refutation of the Varṇa-Vākyārtha Theory
In Kumārila’s doctrine, phonemes play an important role in conveying the meaning of the word (padārtha). For him, it is not the word itself but the component phonemes that make us understood the word’s meaning. We might imagine that the same strategy would be applied to the sentence meaning, but Kumārila refutes that idea in the ŚVv, vv.
112–117 on several grounds. “Phonemes” (varṇāḥ) in his discussion in vv. 112–114 are, as his commentator Pārthasārathimiśra explains, entities that disappear immediately after their utterance and survive only in the memory of the hearer.1) The varṇa theory is always vulnerable to the objection that individual phonemes disappear before the manifestation of the word’s or sentence’s meaning. But in case of the word’s meaning, of which Kumārila and the later Mīmāṃsakas accept phonemes to be the cause, the latent impressions (saṃskāra) left by the perception of the phonemes as well as the memory (smṛti) produced by them are postulated as mediating between the phonemes and the understanding of the word’s meaning. In case of the sentence’s meaning, however, according to Kumārila, it is much more difficult to claim this role for phonemes since sentences are longer and more complex entities than individual words. Kumārila also says that phonemes exhaust their role by conveying the meaning of the word.2)
What then if it is claimed that the last phoneme (antyavarṇa) supported by all the memories of the previous phonemes is what conveys the meaning of the sentence?3) This idea too is refuted by Kumārila: the latent impressions of phonemes cannot work both on the cognition of word’s meaning and on that of the sentence’s meaning.4) Furthermore, the cognition of the sentence’s meaning inevitably comes after the cognition of the word’s meaning, which blocks any direct function on the part of the phonemes. Thus in Kumārila’s view, phonemes cannot be accepted as the conveyors of the sentence’s meaning.
2. Śālikanātha’s and Vācaspati’s Follow-up of Kumārila
Kumārila showed us two types of the varṇa theory, in which (1) the last phoneme supported by the memory of the previous phonemes or (2) the series of phonemes in the form of memory conveys the sentence meaning. Following his argument, Śālikanātha and Vācaspati criticize the same two views. There is a high possibility that Vācaspati knew Śālikanātha’s PrPv and used his text in the TB in a similar manner. For the discussion and its wording in the two texts are quite close.5) Here let us have a look at Vācaspati’s criticism of the varṇa theory. In his argument, there are two types of the opponent: (1) the first claims that it is the last phoneme (antyavarṇa) together with the latent impression of the previously perceived words and their meanings that conveys
the sentence meaning; and (2) the second holds that it is the series of phonemes (varṇamālā) located in the memory that makes us understood the sentence meaning.6) How does Vācaspati refute these two opinions?
(1) Against the antyavarṇa theory Vācaspati asks the following question. If perception of the last phoneme is the cause of our understanding the sentence meaning, then when is it that the hearer recollects the linguistic convention (saṅketa) of the last word? If he recollects it after hearing the last phoneme, and the last phoneme conveys the meaning of the sentence thereafter, there would be the undesirable consequence that the last phoneme, which will already have disappeared at the time that one recollects the linguistic convention, will be the cause of the understanding of the meaning of the sentence. If, on the other hand, the hearer does not recollect the linguistic convention that defines the meaning of the last word, then there would be no point in any linguistic convention intervening between the word and the meaning, and anyone who hears any sentence will understand its.7)
Thus, the antyavarṇa theory entails the problem of how to recollect the linguistic convention. However, the second theory, that of the varṇamālā theory, can remove this problem, because in that case, since the whole series of phonemes is grasped as a single sentence in the form of memory, the hearer recollects its relation with the meaning. Here there is no need to recollect the meaning of the last word.8)
(2) Against the varṇamālā theory, Vācaspati raises three problems. [1:] If the series of phonemes, namely the sentence, is taken to be a single entity with its own meaning, we have to postulate a new capacity for every sentence. [2:] if the word’s meaning is not cognized in the sentence, the sentence meaning would be never manifest because the latter depends on the former. [3:] if the sentence meaning is understood after cognizing the words’ meanings, there would be no use for phonemes as the conveyor of the sentence meaning, because the words’ meanings would have the same role.9)
In this way, Vācaspati [and Śālikanātha] criticize the varṇa theory in almost the same way as that of Kumārila, but since their criticism is much more detailed, it suggests that there might be real people who asserted that theory. Moreover, we can find a similar but positive argument in the NM of Jayanta.
112–117 on several grounds. “Phonemes” (varṇāḥ) in his discussion in vv. 112–114 are, as his commentator Pārthasārathimiśra explains, entities that disappear immediately after their utterance and survive only in the memory of the hearer.1) The varṇa theory is always vulnerable to the objection that individual phonemes disappear before the manifestation of the word’s or sentence’s meaning. But in case of the word’s meaning, of which Kumārila and the later Mīmāṃsakas accept phonemes to be the cause, the latent impressions (saṃskāra) left by the perception of the phonemes as well as the memory (smṛti) produced by them are postulated as mediating between the phonemes and the understanding of the word’s meaning. In case of the sentence’s meaning, however, according to Kumārila, it is much more difficult to claim this role for phonemes since sentences are longer and more complex entities than individual words. Kumārila also says that phonemes exhaust their role by conveying the meaning of the word.2)
What then if it is claimed that the last phoneme (antyavarṇa) supported by all the memories of the previous phonemes is what conveys the meaning of the sentence?3) This idea too is refuted by Kumārila: the latent impressions of phonemes cannot work both on the cognition of word’s meaning and on that of the sentence’s meaning.4) Furthermore, the cognition of the sentence’s meaning inevitably comes after the cognition of the word’s meaning, which blocks any direct function on the part of the phonemes. Thus in Kumārila’s view, phonemes cannot be accepted as the conveyors of the sentence’s meaning.
2. Śālikanātha’s and Vācaspati’s Follow-up of Kumārila
Kumārila showed us two types of the varṇa theory, in which (1) the last phoneme supported by the memory of the previous phonemes or (2) the series of phonemes in the form of memory conveys the sentence meaning. Following his argument, Śālikanātha and Vācaspati criticize the same two views. There is a high possibility that Vācaspati knew Śālikanātha’s PrPv and used his text in the TB in a similar manner. For the discussion and its wording in the two texts are quite close.5) Here let us have a look at Vācaspati’s criticism of the varṇa theory. In his argument, there are two types of the opponent: (1) the first claims that it is the last phoneme (antyavarṇa) together with the latent impression of the previously perceived words and their meanings that conveys
the sentence meaning; and (2) the second holds that it is the series of phonemes (varṇamālā) located in the memory that makes us understood the sentence meaning.6) How does Vācaspati refute these two opinions?
(1) Against the antyavarṇa theory Vācaspati asks the following question. If perception of the last phoneme is the cause of our understanding the sentence meaning, then when is it that the hearer recollects the linguistic convention (saṅketa) of the last word? If he recollects it after hearing the last phoneme, and the last phoneme conveys the meaning of the sentence thereafter, there would be the undesirable consequence that the last phoneme, which will already have disappeared at the time that one recollects the linguistic convention, will be the cause of the understanding of the meaning of the sentence. If, on the other hand, the hearer does not recollect the linguistic convention that defines the meaning of the last word, then there would be no point in any linguistic convention intervening between the word and the meaning, and anyone who hears any sentence will understand its.7)
Thus, the antyavarṇa theory entails the problem of how to recollect the linguistic convention. However, the second theory, that of the varṇamālā theory, can remove this problem, because in that case, since the whole series of phonemes is grasped as a single sentence in the form of memory, the hearer recollects its relation with the meaning. Here there is no need to recollect the meaning of the last word.8)
(2) Against the varṇamālā theory, Vācaspati raises three problems. [1:] If the series of phonemes, namely the sentence, is taken to be a single entity with its own meaning, we have to postulate a new capacity for every sentence. [2:] if the word’s meaning is not cognized in the sentence, the sentence meaning would be never manifest because the latter depends on the former. [3:] if the sentence meaning is understood after cognizing the words’ meanings, there would be no use for phonemes as the conveyor of the sentence meaning, because the words’ meanings would have the same role.9)
In this way, Vācaspati [and Śālikanātha] criticize the varṇa theory in almost the same way as that of Kumārila, but since their criticism is much more detailed, it suggests that there might be real people who asserted that theory. Moreover, we can find a similar but positive argument in the NM of Jayanta.
3. Jayanta’s Argument on the Varṇa Theory
In the long discussion of the varṇa-vākyārtha idea, Jayanta refers to the views of the senior Naiyāyika philosophers: Ācāryāḥ (“teachers”), Vyākhyātāraḥ (“commentators”), and Śaṅkarasvāmin. Among them, the Ācāryas’ view is the simplest, and it is close to that which we have seen as the antyavarṇa theory. And therefore this view has the problem that it cannot explain when the linguistic convention of each word is recollected.10)
The Vyākhyātṛs’ view is a revised version of this: After cognizing the first word (W1), one recollects its linguistic convention, and through it there arises the cognition of W1’s meaning (W1a) demarcated by the cognition of W1. The cognition of W1a creates a latent impression. The same process is repeated until the cognition of the last word. Immediately after the cognition of the last phoneme of the last word, there arises a single memory of all the preceding phonemes and words, a memory which has all the previous words and the meanings demarcated by them as its object. In this single memory together with the cognition of the meaning of the last word, there arise the aggregate of the words and that of the words’ meanings, namely the sentence and its meaning.11)
This view, however, is also rejected by Jayanta. How then does Jayanta explain the
varṇa theory? In Jayanta’s view, the cognition of W1 brings about its latent impression
as well as the recollection of the relation with W1a; and the cognition of W1a, which is produced by the recollection of its linguistic convention, also causes its latent impression to arise.12) Thus, three kinds of latent impressions, namely of phonemes, words, and words’ meanings, are stored separately in the mind. Kumārila claimed that the latent impressions of phonemes exhaust their role by conveying the word’s meaning. But in Jayanta’s opinion both processes, those of phonemes’ conveying the word’s meaning and of phonemes’ conveying the sentence meaning share the following process: [1] Cognition of phonemes → [2] Latent impression → [3] Memory (= Cognition of the word) → [4] Recollection of its convention → [5] Cognition of the meaning. In case of the sentence meaning [3] and [5] produce the latent impressions and those of the subsequent words are separately stored until the cognition of the last word’s meaning.
4. Conclusion
Though the discussion in the ŚVv is too brief to reveal the existence of real opponent, it is possible that the situation is different in case of the PrPv and TB. It has been more or less agreed in recent scholarship that Jayanta is earlier than Vācaspati.13) Furthermore, the flow of the argument in the TB is very close to that of the sixth chapter of the NM. From these facts, I propose the possibility that the opponents holding the
varṇa-vākyārtha theory in the TB are the Naiyāyikas. To be more precise, the antyavarṇa theory
is close to the theories of the Ācāryas and Vyākhyātṛs, while the varṇamālā theory is close to Jayanta’s. The editors’ notes in the PrPv and TB support half of my view: they identify the holders of the antyavarṇa theory as the Naiyāyikas, but do not identify those who hold the varṇamālā theory.
Notes
1）See the commentary on the ŚVv, v. 112. 2）ŚVv, vv. 112cd–115ab.
3）The similar argument is seen in the Sphoṭavāda of the ŚV when Kumārila claims phonemes as the conveyor of the word’s meaning.
4）ŚVv, vv. 115cd–117.
5）See Moriguchi [2014] for the further information on the TB and PrPv. 6）TB, p. 164, l. 3–p. 165, l. 2. 7）TB, p. 231, l. 1–p. 234, l. 2. 8）TB, p. 237, l. 1–p. 238, l. 2. 9）TB, p. 238, l. 5–p. 243, l. 2. 10）NM, p. 192, l. 7–p. 193, l. 2. 11）NM, p. 194, l. 17–p. 195, l. 14. 12）NM, p. 199, l. 12–p. 200, l. 8. 13）See Marui [2014], pp. 231–299. Abbreviations and Primary Sources
NM Nyāyamañjarī of Jayantabhaṭṭa with Ṭippaṇī—Nyāyasaurabha by the Editor. Edited by K. S. Varadācārya. Vol. 2. Mysore: Oriental Research Institute, 1983.
TB Tattvabindu by Vācaspatimiśra with Tattvavibhāvanā by Ṛṣiputraparameśvara. Rev. ed. Edited by V. A. Ramaswami Sastri. Delhi: Bharatiya Kala Prakashan, 2009.
ŚV/ŚVv Ślokavārttika of Śrī Kumārila Bhaṭṭa with the Commentary Nyāyaratnākara of Śrī Pārthasārathi Miśra. Edited and revised by Ganga Sagar Rai. Varanasi: Ratna Publications, 1993.
3. Jayanta’s Argument on the Varṇa Theory
In the long discussion of the varṇa-vākyārtha idea, Jayanta refers to the views of the senior Naiyāyika philosophers: Ācāryāḥ (“teachers”), Vyākhyātāraḥ (“commentators”), and Śaṅkarasvāmin. Among them, the Ācāryas’ view is the simplest, and it is close to that which we have seen as the antyavarṇa theory. And therefore this view has the problem that it cannot explain when the linguistic convention of each word is recollected.10)
The Vyākhyātṛs’ view is a revised version of this: After cognizing the first word (W1), one recollects its linguistic convention, and through it there arises the cognition of W1’s meaning (W1a) demarcated by the cognition of W1. The cognition of W1a creates a latent impression. The same process is repeated until the cognition of the last word. Immediately after the cognition of the last phoneme of the last word, there arises a single memory of all the preceding phonemes and words, a memory which has all the previous words and the meanings demarcated by them as its object. In this single memory together with the cognition of the meaning of the last word, there arise the aggregate of the words and that of the words’ meanings, namely the sentence and its meaning.11)
This view, however, is also rejected by Jayanta. How then does Jayanta explain the
varṇa theory? In Jayanta’s view, the cognition of W1 brings about its latent impression
as well as the recollection of the relation with W1a; and the cognition of W1a, which is produced by the recollection of its linguistic convention, also causes its latent impression to arise.12) Thus, three kinds of latent impressions, namely of phonemes, words, and words’ meanings, are stored separately in the mind. Kumārila claimed that the latent impressions of phonemes exhaust their role by conveying the word’s meaning. But in Jayanta’s opinion both processes, those of phonemes’ conveying the word’s meaning and of phonemes’ conveying the sentence meaning share the following process: [1] Cognition of phonemes → [2] Latent impression → [3] Memory (= Cognition of the word) → [4] Recollection of its convention → [5] Cognition of the meaning. In case of the sentence meaning [3] and [5] produce the latent impressions and those of the subsequent words are separately stored until the cognition of the last word’s meaning.
4. Conclusion
Though the discussion in the ŚVv is too brief to reveal the existence of real opponent, it is possible that the situation is different in case of the PrPv and TB. It has been more or less agreed in recent scholarship that Jayanta is earlier than Vācaspati.13) Furthermore, the flow of the argument in the TB is very close to that of the sixth chapter of the NM. From these facts, I propose the possibility that the opponents holding the
varṇa-vākyārtha theory in the TB are the Naiyāyikas. To be more precise, the antyavarṇa theory
is close to the theories of the Ācāryas and Vyākhyātṛs, while the varṇamālā theory is close to Jayanta’s. The editors’ notes in the PrPv and TB support half of my view: they identify the holders of the antyavarṇa theory as the Naiyāyikas, but do not identify those who hold the varṇamālā theory.
Notes
1）See the commentary on the ŚVv, v. 112. 2）ŚVv, vv. 112cd–115ab.
3）The similar argument is seen in the Sphoṭavāda of the ŚV when Kumārila claims phonemes as the conveyor of the word’s meaning.
4）ŚVv, vv. 115cd–117.
5）See Moriguchi [2014] for the further information on the TB and PrPv. 6）TB, p. 164, l. 3–p. 165, l. 2. 7）TB, p. 231, l. 1–p. 234, l. 2. 8）TB, p. 237, l. 1–p. 238, l. 2. 9）TB, p. 238, l. 5–p. 243, l. 2. 10）NM, p. 192, l. 7–p. 193, l. 2. 11）NM, p. 194, l. 17–p. 195, l. 14. 12）NM, p. 199, l. 12–p. 200, l. 8. 13）See Marui [2014], pp. 231–299. Abbreviations and Primary Sources
NM Nyāyamañjarī of Jayantabhaṭṭa with Ṭippaṇī—Nyāyasaurabha by the Editor. Edited by K. S. Varadācārya. Vol. 2. Mysore: Oriental Research Institute, 1983.
TB Tattvabindu by Vācaspatimiśra with Tattvavibhāvanā by Ṛṣiputraparameśvara. Rev. ed. Edited by V. A. Ramaswami Sastri. Delhi: Bharatiya Kala Prakashan, 2009.
ŚV/ŚVv Ślokavārttika of Śrī Kumārila Bhaṭṭa with the Commentary Nyāyaratnākara of Śrī Pārthasārathi Miśra. Edited and revised by Ganga Sagar Rai. Varanasi: Ratna Publications, 1993.
Secondary Sources
Marui Hiroshi 丸井浩. 2014. Jayanta kenkyū: Chūsei Kashimīru no bunjin ga kataru Niyāya tetsugaku ジャヤンタ研究: 中世カシミールの文人が語るニヤーヤ哲学. Tokyo: Sankibō Busshorin. Moriguchi Mai 森口眞衣. 2004. “Gakusetsu no hensen: Vāchasupati ga shōkai shita bun’i ninchi
riron no haikei” 学説の変遷: ヴァーチャスパティが紹介した文意認知理論の背景. Indo tetsugaku Bukkyōgaku 印度哲学仏教学 19: 117–131.
Key words vākyārtha, Ślokavārttika, Kumārilabhaṭṭa, Prakaraṇapañcikā, Śālikanātha, Tattvabindu, Vācaspatimiśra, Nyāyamañjarī, Jayantabhaṭṭa
（JSPS Research Fellow）
The Proof for the Existence of God in Nyāyalīlāvatī
Hibi Mayumi
1. Introduction
The question about God (īśvara) has been one of the central topics in Indian traditions of thought. It is discussed by most philosophical schools, some believing, some disbelieving in God, and others believing in but rejecting any rational proof for God’s existence. Naiyāyikas and Vaiśeṣikas establish God’s existence by inference, while the atheists such as Mīmāṃsakas, Jainas, and Buddhists find such inferences fallacious.
In this paper, I analyze the proof for the existence of God in Nyāyalīlāvatī (= NL) composed by Vallabha (12c).1) The NL is, as Sjödin (2006: 43) says, “an independent text with the same overall structure as a classical Vaiśeṣika text.” On the other hand, the complex terminology of logic and the sophisticated debating techniques used in this work show its substantial affinity with later texts of the Navya-Nyāya tradition.2) Analyzing the discussion on God’s existence in NL should help us track the development of the rational theistic argument in the post-Udayana and pre-Gaṅgeśa period.
2. The Outline of the Discussion on God’s Existence in NL
In NL, as well as in the commentaries on the Padārthadharmasaṅgraha, the problem of God is discussed just after the discussion on the creation and destruction of the material world, both of which are said to be overseen by God in the Vaiśeṣika doctrine.
Vallabha declares that the existence of God is established through the following inference: The things under dispute (= the earth, etc.) have a maker, because they are effects, just as a pot is an effect and has a maker (NL(Ch) 239.2: vivādādhyāsitaṃ
sakartr̥kam, kāryatvāt, ghaṭavat).3) Then he proceeds to define the target property of
having a maker (sakartr̥katva) as the property of being produced by the direct knowledge of the raw materials of the object to be made, as well as by the desire to
