Page #1
--------------------------------------------------------------------------
________________ THE APOHA THEORY OF DIGNAGA Shoryu Katsura In the Fifth Chapter of the P(ramana)(amuccaya) Dignaga presents the theory of anyapoha (exclusion of others) which may be regarded as his most original contribution to Indian logic!). Dignaga declares in the introductory verse (K. 1) that verbal cognition (sabda) is nothing but inference (anumana) because both of them function on the principle of anyapoha. He refutes four possible candidates for the meaning of a word (sabdartha), viz. (i) an individual (bheda), (ii) a universal (jati), (iii) a relation (sambandha) between them and (iv) a thing possessing a universal (jatimat / tadvat), and concludes that the meaning of a word is anyapoha (KK. 2-11). Here he inserts two summarizing verses (KK. 12 & 13)2). He then deals with the relation of co-reference (samanadhikaranya) and that of qualifier-qualificand (visesanavises yabhava) with regard to compound expressions such as 'nilot pala' (a blue lotus) (KK. 14-25ab). Next he gives a detailed account of the apoha theory, with special emphasis on the scope of 'exclusion' (KK. 25cd-38). He also criticizes the Samkhya position (KK. 39-45) and discusses the meaning of a sentence (vakyartha) (KK. 46-49). In the final verse (K. 50) he takes up a few additional topics and concludes that the other so-called pramanas, i. e., upamana and so on, should be regarded as inference. The aim of this paper is to present the essence of Dignaga's theory of anyapoha discussed in PS, V, KK. 25cd-38. Dignaga's discussion centers upon the question of what is excluded by a word, that is, what is meant by the word anya of anyapoha. In this connection, he seems to be presupposing the following Vaisesika-like hierarchy of words or universals3). Before discussing what is excluded, Dignaga explains three cases where there is no exclusion. (i) Two synonyms (parya yasabda), say 'urksa' and 'taru' (both meaning 'tree'), do not exclude each other's meaning, for they by definition have the same meaning4). (ii) The word 'vrksa' does not exclude the meaning of the universal word (saman yasabda / jatisabda)5), 'parthiva' (made of the earth ele - 493
Page #2
--------------------------------------------------------------------------
________________ (17) THE APOHA THEORY OF DIGNAGA (S. Katsura) jneya sat asat dravya guna karman parthiva apya, etc. rupa, tec. utksepana, etc. veksa ghata, etc. simsapa palasa, etc. puspita phalita, etc. ment), to which it belongs, for the two words are not incompatible (avirodhin) with each other. The word 'vrksa' rather encompasses and engenders definite understanding (niscaya) ot the meanings of 'parthiva,' 'dravya' (substance), 'sat' (existent), etc.; if something is called 'tree,' it must be made of the earth element a substance, existent, and so on. Generally speaking, A UNIVERSAL WORD OF LOWER ORDER DOES NOT EXCLUDE BUT ENCOMPASSES THE MEANINGS OF UNIVERSAL WORDS OF HIGHER ORDER WITH WHICH THE FORMER HAS AN INVARIABLE RELATION (avyabhicarin)6)......Rule 17). (iii) The word 'sat' neither excludes nor encompasses the meanings of 'dravya,' 'parthiva,' 'vyksa, etc., for it awakens expectation (akanksana) of determining which particular (bheda) it actually refers to whether something called 'existent is, for instance, a substance or a quality (guna) or an action (karman) - or it raises doubt (sardeha) as to which particular it refers to. Generally speaking, A UNIVERSAL WORD OF HIGHER ORDER NEITHER EXCLUDES NOR ENCOMPASSES THE MEANINGS OF UNIVERSAL WORDS OF LOWER ORDER ......Rule 118). Now let us see how Dignaga explains what is excluded by a word. It is to be noted that Dignaga recognizes two types of apoha, viz. direct and indirect apoha. (i) The word 'vrksa' directly excludes the meanings of the word 'ghata' (a pot), etc., which share the same universal of 'being made of the earth - 492 --
Page #3
--------------------------------------------------------------------------
________________ (18) THE APOHA THEORY OF DIGNAGA (S. Katsura) element' (parthivatva) with 'vrksa,' for 'vrksa' is not compatible (virodhin) with them in respect of that common universal- once something made of the earth element is called 'tree', it cannot be a pot, etc. Generally speaking, TWO UNIVERSAL WORDS OF THE SAME ORDER EXCLUDE EACH OTHER'S MEANING......Rule III9). (ii) The word 'vrksa' indirectly excludes the meaning of the word 'apya' (made of the water element), for 'vrksa' encompasses the meaning of 'parthiva' (Rule 1) which in turn excludes the meaning of 'apya' (Rule III). Generally speaking, A UNIVERSAL WORD OF LOWER ORDER EXCLUDES WHATEVER IS DIRECTLY EXCLUDED BY UNIVERSAL WORDS OF HIGHER ORDER WITH WHICH THE FORMER HAS AN INVARIABLE RELATION ......Rule IV10). In order to exhaust all the possible relationships between two universal words, Dignaga mentions a third case: (iii) the word 'parthiva' disregards the meaning of the word 'rapa' (color), for 'parthiva' indirectly excludes the meaning. of 'guna' (Rule IV) which in turn neither excludes nor encompasses the meaning of 'rapa' (Rule II). Generally speaking, A UNIVERSAL WORD OF LOWER ORDER DISREGARDS THAT WHICH BELONGS TO WHATEVER IS DIRECTLY EXCLUDED BY UNIVERSAL WORDS OF HIGHER ORDER WITH WHICH THE FORMER HAS AN INVARIABLE RELATION......Rule V11). In brief, a given word excludes the meanings of words which share the same universal with it and it indirectly excludes whatever is directly excluded by words of higher order, thus engendering definite understanding of their meanings, and it awakens expectation or raises doubt about the meanings of words of lower order. In the remaining space, I would like to mention a few interesting points found in this section of PS, V, and not explicitly discussed in previous studies of Dignaga's theory of anyapoha, (i) The expression 'samanadhikaranya' has two distinct meanings, viz. the relation of referring to the same thing and the relation of having the same locus12). Although Dignaga normally employs the expression in the first meaning, he does recognize the second meaning in PS, V, K. 30. 491
Page #4
--------------------------------------------------------------------------
________________ THE APOHA THEORY OF DIGNAGA (S. Katsura) (19) (ii) G. Cardona has shown that the Indian grammarians determine the constant co-occurrence of a word- -he calls it a linguistic item-and a meaning by anvaya and vyatireka13). He quotes Kaiyata's formulation: "[Anvaya :) the understanding of a meaning when there is an item; [vyatireka:] the non-understanding of a meaning when an item is absent'. (pp. 337-8) A similar account of anvaya and vyatireka is given by Dignaga in PS. V, K. 34 and Vrtti. Dignaga considers that anvaya and vyatireka are the two means of a word to express its meaning, and he defines them respectively as 'employment in similar cases' (tulye vrttih) and 'non-employment in dissimilar cases' (atulye 'vrttih)14). In other words, anvaya is the employment of a word when a meaning is intended, and vyatireka is the non-employment of a word when a meaning is not intended. The difference between Dignaga's formulation and the grammarians' is due to the fact that the former is made from the point of view of the speaker, while the latter is made from that of the listener. (iii) A universal, say rapatva (colorness), is generally believed to be a ground for applying (pravrttinimitta) a word, say 'rapa', to those which are supposed to possess that universal, e. g. blue, yellow, etc. Dignaga rejects this view and concludes that we employ a certain word for a certain group of objects only because we follow the linguistic convention generally accepted by ordinary people15) 20. 8. 1979 NOTES 1) The following previous contributions to the study of Dignaga's theory of anyapoha have been consulted: E. Frauwallner, "Dignaga, sein Werk und seine Entwicklung," WZKSO III (1959), pp. 100-105; M. Hattori, "A Study of the Chapter on Apoha of the Mimamsaslokavarttika (I & II)" (in Japanese), Memoirs of the Faculty of Letters, Kyoto University, Nos. 14 & 15 (1973-75); Do., "The Sautrantika Background of the Apoha Theory," Buddhist Thought and Asian Civilization, California (1977); Do., "The Apoha Theory and the Sautrantika Doctrine" (in Japanese), Tri-pitaka 140 (1977); Muni Jambuvijaya, Dvadasaram Nayacakram, pt II, Bhavnagar (1976)...... pp. 607-8, 629-33, 638-40, 650-1, 728-9 contain the Sanskrit fragments and reconstructions of PS, Chap. V. 2) Frauwallner in the above-mentioned article has pointed out that the two verses correspond to PS, II, KK. 13 & 17= Nyayamukha, KK. 17 & 18. 3) The table has been reconstructed from the information given in PS, V, K. 25 -490
Page #5
--------------------------------------------------------------------------
________________ (20) THE APOHA THEORY OF DIGNAGA (S. Katsura) cd, Vitti & Tika, K. 28, Vitti & Tika, and K. 35, Vstti & Tika. It should not be regarded as complete, but as subject to more specifications. 4) PS, V, K. 25cd and Vstti. 5) Although Dignaga mentions the name 'particular word' (bhedasabda), I do not use it because the difference between a universal word and a particular word is only relative......the word 'dravya' is a particular word with regard to 'sat' but it is a universal word with regard to 'parthiva'. 6) In this connection x has an invariable relation with y, if x is absent whenever y is absent. 7) PS, V, K, 25cd and Vstti, and KK. 27 & 35. 8) Ibid., K. 25cd and Vstti, and KK. 26 & 35. 9) Ibid., K. 28ab and Vstti. 10) Ibid., K. 28cd and Vitti. 11) Ibid., K. 28cd and Vitti. 12) G. Cardona, "Panini's Karakas," J. of Ind. Philos., Vol. III (1974), pp. 289-291. 13) "Anvaya and V yatireka in Indian Grammar," The Ad yar Library Bulletin, Vol. 31-32 (1967-68). 14) Cf. PS, II, K. 5cd: anume ye 'tha tattulye sadbhavo nastitasati!! 15) PS, V, KK. 37-38ab and Vstti.