Dignagas Theory Of Direct Knowledge

Author(s): Massaki Hattori
Publisher: Massaki Hattori
Catalog link: https://jainqq.org/explore/269375/1

Page #1
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE - An Annoted Translation of Pramanasamuccaya, Chapter 1- Masaaki HATTORI INTRODUCTORY REMARKS It is just half a century ago that the late Mahamahopadhyaya S. C. Vidyabhusana published his monumental work, History of the Mediaeval School of Indian Logic, in which he first introduced Dignaga's theory of knowledge to the scholars of Indian philosophy. Notwithstanding this incentive, the furtherance of the study on this great figure was not a task that could be carried out without toil. Pramanasamuccaya (PS), a comprehensive and systematical work on epistemology and logic, though being known as a reformative work in the histroy of Indian Logic, is unfortunately not preserved in the original Sanskrit, and the Tibetan version, the only available material of the study on this text, cannot be considered to be quite a readable one, perhaps mainly due to the unfitness of the Tibetan language to this kind of strict and subtle arguments and also sometimes to the translator's lack of understanding. Attempts have been made by Randle and other scholars to collect passages of PS scattered in the Nyaya texts and elsewhere, and these attempts have proved to be very helpful for the study of Dignaga's theory. Controversial points of his theory have been thus gradually made clear. An epoch was marked by Stcherbatsky when he published an elaborative study on the Nyayabindu. While annexing precise notes to this concise treatise of Dharmakirti, he made frequent references to PS, and moreover, translated the portion wherein Dignaga discussed the problem of self-cognition with Jinendrabuddhi's commentary. Dignaga's theory was brought forth under the brighter light and the preliminary course to the textual study of PS was well set up by his effort. Owing to the successful result of Dr. Rahula Sankstyayana's second expedition to Tibet, we are now provided with the Sanskrit text of Pramanavarttika (PV), an extensive critical commentary on PS, along with some commentaries on it. These are precious materials to further the study of the doctrines of Dignaga as well as of Dharmakirti, for PV treats with the same problems as discussed in PS, and, fortunately enough, the passages of the latter are found often cited in the commentaries of the former. Recently some portions of PS have been rendered by Japanese scholars into their own language from the Tibetan version. The portion translated and annoted here was put into Japanese some years back by Mr. Takemura, Asst. Prof. of Ryukoku Univ., Kyoto (Ryukoku Daigaku Ronshu, No. 351, 1956), but, according to the opinion of the present writer, his translation can hardly be recognized as a scholary work,

Page #2
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE misreading of the text being found almost in every line. The present translation is based on, in principle, Tshad-ma kun-las btus-pahi hgrel-pa (Pramanasamuccaya-vrtti), Kanakavarman's version, Peking Ed., Mdo-hgrel XCV (Ce), fol. 93 bff. Verses of PS are properly inserted between lines by the present writer. Constant references have been made to Vasudhararaksita's version, Peking Ed., Mdo-hgrel XCV, fol. 13 aff and Jinendrabuddhi's Visalamalavati-nama Pramanasamuccaya-vrtti, Sde-dge & Peking Eds. The writer wishes to acknowledge his gratitude to Prof. Hatano, Tohoku Univ., Asst. Prof. Miyasaka, Koyasan Univ., Asst. Prof. Ihara, Kyushu Univ. and Mr. Hasuba, Otani Univ., through whose courtesy he could obtain the photographic copies of the abovementioned texts. The writer has also to express his indebtedness to Jain Muni Jam buvijaya who was kind enough to send the writer a proof of his Sanskrit reconstruction of PS, Chap. I, which is expected to be out soon as an appendix to his edition of Nayacakra-vrtti. Owing to this excellent Sanskrit reconstruction, the writer could clarify some ambiguous points. Muni Jambuvijaya collected many original passages of PS thus far left unnoticed from various sources, which, however, have not been noted in this paper because the writer does not like to refer to them before the publication of the said work. (Fol. 93b, 5) $1. SALUTATION Verse 1. I salute Him who is the personification of valid knowledge, who pursues the benefit of the living beings, who is the teacher, the sugata, the protector. And, in order to establish the means of valid knowledge, I shall unite here under one head the scattered fragments from all my other treatises!). BL Abbreviations. AKV Sphutartha Abhidharmakosavyakhya, Ed. by U. Wogihara, Tokyo, 1932-36. Th. Stcherbatsky, Buddhist Logic, 2 Vols, Bibliotheca Buddhica XXVI, Leningrad, 1930-32. DP Dharmottarapradipa, Ed. by D. Malvania, Patna, 1955. NB Nyayabindu of Dharmakirti. NBh Nyayabhasya of Vatsyayana. NBT Nyayabindu-tika of Dharmottara, Ed. with DP. NM Nyayamukha, Chinese Version, Taisho Tripitaka, XXXII, pp. 1-6. NV Nyayavarttika of Uddyotakara, Varanasi Ed. NVTT Nyayavarttika-tatparya-tika of Vacaspatimisra, Kashi Skt. Ser. Pramanasamuccaya. PSV Pramanasamuccaya-vstti. PSVT Visalamalavati-nama Pramanasamuccaya-tika, Tibetan Version, Tohoku, No. 4268. TS Tattvasamgraha of Santaraksita, Gaekwad Oriental Series, Nos. XXX, XXXI, Baroda 1926. TSP Tattvasamgraha-panjika of Kamalasila, Ed. with TS. Pramanavartikabhasyam or Vartikalamkarah of Prajnakaragupta, Ed. by R. Sankstyayana, Patna, 1953. VPM Vibhuticandra annexed to Dharmakirti's Pramanavarttika with a Commentary by Ma norathanandin, Ed. by R. Sankstyayana, Patna, 1937. VA

Page #3
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE (2... Here (in the verse of salutation), in order to cause in the heart of the people, at the beginning of the treatise, the faith in the Buddha), the praise is expressed in honour of the Buddha, the Adorate, who, because of his perfectness of the cause and the effect, is to be regarded as the personification of the valid knowledge (pramanabhuta)). There, the cause is the completion of the intention (asaya) and the practice (prayoga). (The completion of) the intention means that the Buddha is the One pursuing the benefit of the living beings. (The completion of) the practice signifies that the Buddha is the teacher because he teaches the people. The effect is the 1) Dignaga is generally regarded as the founder of the so-called Logician Vijnanavadins (nyayanusarino vijnanavadinah), the younger branch of the Yogacara-Vijnanavadins; the elder school of it represented by Asanga and Vasubandhu is known by the name of Vijnanavadins basing upon Scripture (agamanusarino vijnanavadinah). The names given to these two subdivisions of the Yogacara show clearly the difference of attitude between the two towards the Buddhist truth. Dignaga does not recognize, as the masters of the elder school do, the authority of the Scripture. According to him, the words of the Buddha must be subjected to the critical test before they are accepted as valid, and we find that his works are consistently permeated by the critical and rationalistic spirit. This spirit, however, is not introduced newly by him into Buddhism, but he inherited it from the Buddha, who used to exhort his disciples not to accept any of his words merely out of reverence but to test it critically, just as people test the purity of gold by burning it in fire, by cutting it and by examining it on a touchstone , cp. Anguttara Nikaya, III, 65,3, Majjhima N., sutta 38, TSP, p. 12, Jnanasarasamuccaya, k. 31 (transl. by S. Yamaguchi, Chukan Bukkyo Ronko, p. 327), Th. Stcherbatsky, BL, Vol. I, p. 77, S. Radhakrishnan, Indian Philosophy, Vol. I, p. 611, do., The Dhammapada, pp. 10-11, S. Mookerjee, The Buddhist Philosophy of Universal Flux, p. xl. Dignaga is convinced of his following the teaching of the Buddha by establishing the theory of knowledge, and basing upon this conviction, he gives, at the beginning of the treatise, a salutation to the Buddha, who, according to his expression, is to be reckoned as the personification of the valid knowledge (pramana-bhuta). It is reported by Bu. ston that Dignaga wrote this verse on the side of a rock in the cavern known by his name, and that as he wrote down this salutation and his determination to establish the true theory of knowledge, various ominous signs appeared, cp. Obermiller, History of Buddhism (Chos-hbyun) by Bu-ston, Vol. II, p. 150. Dharmakirti attaches importance to this verse, by which, he thinks, the fundamental standpoint of the Buddhist Logicians is declared. He makes thorough discussion on this point in his extensive critical commentary on Pramanasamuccaya, namely Pramana. varttika, and, in consequence, the chapter on the establishment of the means of valid knowledge (pramana-siddhi-pariccheda) is, in this work, separated from the chapter on the direct knowledge and is treated as independent. The full verse is cited at VPM, p. 518, and the first half at VA, p. 3, and AKV, p. 7, 5.6: pramana-bhutaya jagad-dhitai sine pranamya sastre sugataya tayine / pramana-siddhyai svamatat samuccayah karisyate viprasitad ihaikatah // 2) cp. VA, p. 3, 12ff, 115, 31-32, 116, 5.6. 3) PSVT, 2a, 5-2b, 2: 'pramana-bhuta' implies that the Buddha is the authority (=the valid knowledge) and that He has come into the world (thad ma yan hdi yin la gyur pahan yin pa= idam paramanam ca bhutan ca). 'bhuta' means 'to be born' (skyes-pa=jata) or 'to come into the world' (byun-ba=utpanna). What is meant by this term is that the Buddha, in whose per sonality the valid knowledge is embodied, is distinguished from the supposed authority which has no actual relation with the world, such as isvara who remains in eternal and non-self-revealing state and the like.

Page #4
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE (Fol. 94a, 1) attainment of the objectives of his own and of others. "The attainment of the objective of his own' is (evidenced) by the fact that the Buddha is titled sugata (one that has attained bliss); that title (can be explained as) indicating three meanings"), i.e., (i) praiseworthiness (prasastartha), just like the handsome (su-rupa)5), (ii) non-retrogression (=final emancipation, apunaravrtty-artha), just like the radical cure from the fever (su-nasta-jvara) and (iii) wholeness (nihsesartha), just like the jar filled with water (su-purna-ghata). These three meanings (of the title 'sugata' are grounds of) distinguishing the Buddha's attainment of his own objective from the attainment of the outsiders (of Buddhism) who has subdued passions (vita-raga) or of those who is undergoing religious training (saiksa) or who no longer need religious training (asaiksa)6). "The attainment of the objective of others' is to be taken as implying that the Buddha is possessed of the 'protector-hood' (tayitva) in the sense that he is, the protection for others. Saluting the teacher who is endowed with such merits, the author will compose, in order to establish the means of valid knowledge, Pramanasamuccaya or Collection of the Theory of Knowledge by collecting (passages) from Nyayamukha and other treatises of the author himself. (The purpose of composing this work is) to reject the others' theory of knowledge and to elucidate the characteristics of the author's ...2) 4) 'Sugata', the word primarily meaning 'well (su)-gone (gata)', is counted among ten titles of the Buddha, cp. Mahavyutpatti, Nos. 1-10, in the sense that He has well attained the enlightenment. This title of the Buddha is explained here as implying His three merits, prasastata, apunaravyttitva and nihsesata, which are respectively the attribute of surupa, sunastajvara and supurnaghata, each of which is prefixed to with 'su' like 'sugata'. It seems that this dogmatical etymology of 'sugata' was prevalent among the Buddhist scholars of this school, for we find the same in Durvekamisra's commentary on NBT, cp. DP, p. 3, 11ff, see also PV, III, kk. 141-144. 5) Tib. skes-bu gzugs-legs-pa: a graceful, handsome person. Surupa might also be taken as the name of a legendary king who gave son, wife and himself to be eaten by an ogre in exchange for religious instruction, cp. Edgerton, Buddhist Hybrid Skt. Dictionary. But this had better be taken here as a common noun, because DP, p. 3, says, when explaining prasastartha, that those who make living by beauty of their form are 'surupa' (surupi rupajivah). 6) Among those Buddhist disciples (sravaka) who has reached the stage of the sage (arya. pudgala), the arhat is called 'asaiksa', because he, extinguishing the influence of passions (asrava. ksaya), no longer needs religious training, and other seven (from srota patti-prati pannaka to arhatprati pannaka), who are to study more in order to attain the arhat-hood, are called 'saiksa'. cp. Abhidharmakosa, VI, pudgala-marga-nirdesah. 7) Dignaga's works preserved in the Tibetan Tripitaka are fourteen in number, while the Chinese Tripitaka enumerates six, of which four are missed in the Tibetan works now available, we find the same theory as expounded in PS(V) being stated, at least, in Hetucakradamaru, Tohoku, No. 4209, cp. PSV, III, 19, Alambanapariksavrtti, Tohoku, No. 4206, XXXI, pp. 888-889, cp. PSV, ad. I, 4, 5, 9, Abhidharmakosavytti-Marmapradipa, Tohoku, No. 4095, cp. PSV, ad. I, 4, and Nyayamukha. Especially many verses and passages of NM are seen rearranged in PS(V), cp, G. Tucci, The Nyayamukha of Dignaga.

Page #5
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE own theory of knowledge); for there is a divergence of opinion with regard to the means of knowledge), on which the apprenhension of the object to be known depends10). 82. TWO SOURCES OF VALID KNOWLEDGE V. 2-3ab: There are two sources of valid knowledge; for the object to be known has two aspects. As regards (cognitions of those objects which are) related to these two aspects, no other independent source of valid knowledge is to be recognized. Nor is there other source in the case of recognition, because (if the other source be admitted to be required in this case), the fallacy of infinitude would be logically concluded, just as in the case of recollected knowledgell). 8) In each chapter of this treatise, Dignaga, after elucidating his own theory, refutes the views of Vadavidhi, Naiyayika, Vaisesika, Samkhya and Mimarsaka. 9) PSVT, llb: vipratipatti=viruddha-pratipatti (hgal baui rtogs pa ni log par rtogs pa rnams te). Theories maintained by others are contradicting each other in four points of view, i.e., the result (hbras-bu=phala), the nature (ran-gi no-bo=svarupa), the object (yul=visaya) and the number (grans=samkhya) of the means of valid knowledge, cp. NBT, p. 35, 1: catur. vidha catra vipratipattih samkhya-laksana-gocara-phala-visaya. TSP, p. 366, 14: tatra pramane svarupa-phala-gocara-samkhyasu paresam vipratipattis catur-vidha. Dignaga's theory is quite unique on each of these four points as seen below. 10) pramanadhino hi prame yadhigamah. cp. the opening statement of NBh: pramanato'rthapratipattau pravrtti-samarthyad arthavat pramanam. In spite of the apparent affinity between Dignaga and the Naiyayika in respect of the theory that parameya (artha) is apprehended by means of pramana, both differ totally in the understanding of the nature of pramana and prameya. While the Naiyayika holds, basing upon the Nyaya realism, that pramana and prameya are independent entities, Dignaga emphasizes their ideated character, cp. below ad. I, 9-10. Nagarjuna denies the possibility of apprehending prameya by means of pramana being mutually involved, have no independent substantiality, cp. Vigrahavyavartani, kk. 31-33, Vaidal yaprakarana, ad. sutra 1-2, and Candrakirti, who lays stress on the Madhyamic transcenden. tal viewpoint, does not agree to the Dignaga's proposition 'pramanadhino prameyadhigamah', for there is nothing to be apprehended or asserted from his point of view, cp. Prasannapada, p. 55ff. However, it is recognized by the Madhyamika too that the transcendental truth does not stand aloof from the empirical world, but reveals itself in the world. Thus the transcendental intuition, admitting that it is essentially inexpressible, must also be transformed into the practical knowledge by virture of which the worldly things are apprehended. That empirical knowledge in which the transcendental truth is revealed can alone be valid and be the criterion in criticizing the illusiveness of merely empirical apprehension. Dignaga's aim of discussing the means of knowledge consists in making clear the structure of this knowledge and its relation to the common knowledge. Accordingly, his discussion does not overstep the boundary of cognitive phenomena, and the transcendental truth is referred to within the scope of its being reflected in empirical knowledge. And the knowledge, whether it be merely empirical or mediated through the transcendental intuition, so far as it is the act of knowing, there must be something to be known by it. In this sense, it can righteously maintained that where there is a fact of knowing, pramana neya are supposed to be there, although they are not to be regarded as real entities. 11) VPM. p. 140, cp. VA, p. 242, 29: pratyaksam anumanam ca pramane laksana-dvayam / prameyam tatra sandhane na pramanantaram na ca // 2 // punah punar-abhijnane 'nistha-a sakteh smrt'adi-vat / 3a-b

Page #6
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE (94a, 4) There are only two sources of valid knowledge!2), direct knowledge (pratyaksa) and indirect knowledge (anumana)13); for the object to be known has two aspects. There is no other object to be known than the particular (svalaksana) and the universal ber (samkhya) of the source of valid knowledge recognized by the different schools of Indian philosophy is as follows: Carvaka, one---perception (pratyaksa); Vaisesika, two-perception and inference (anumana); Sarkhya & one school of Nyaya, three-verbal testimony (sabda) besides the said two; orthodox Nyaya, four-comparison (upamana) together with above three; Prabhakara Mimarsa, five-implication Carthapatti) together with above four; Bhatta Mimamsa & Vedanta, six-negation (abhava) together with above five; Pauranika, eight-inclusion (sambhava) and tradition (aitihya) together with above six, cp. Randle, Indian Logic in the Early Schools, p. 305. The doctrine recognizing aitihya, arthapatti, sambhava and abhava as independent source of valid knowledge is refuted by the Naiyayika, cp. "NS, II, ii, 1ff, and it is supposed that the Nyaya doctrine of four sources of valid knowledge was most authoritative at the time of Dignaga. Among the Buddhists, the author of the Upayahrdaya, a Hinayanist before Nagarjuna, recognizes four sources of valid knowledge as agreed by the Naiyayika, and the elder the Yogacara recognizes three, excluding upamana, but without mentioning reason, from the abovesaid four. Dignaga maintains that sabda is not a different source of knowledge from anumana, because the knowledge derived from sabda indicates its own object through the 'exclusion of other objects' (anya-apoha), which is nothing but the function of anumana, PS, V, k. 1, cited at TSP, ad. k. 1515, p. 441, 6-7: na pramanantaram sabdam anumanat tatha hi sah (=tat)/ krtakatv'adi-vat svartham anyapohena bhasate //, transl. by Stcherbatsky, BL, Vol. I, p. 459, cp. Ihara, Jinna ni okeru Gengo to Sonzai no Mondai, Kyushu Daigaku Tetsugaku Nenpo, Vol. XIV, p. 114, my paper, Fragments of Pramanasamuccaya, Journal of Indian and Buddhist Studies, Vol. VII, 1, p. 326. Upamana, by means of which the similarity of two things is apprehended, is also rejected by Dignaga. If the apprehension of an object by its name, say gavaya, be derived from the words 'gavaya is similar to cow', then it is the same in structure as sabda. If the similarity between cow and gavaya be thought to be apprehended by the cognizant himself, then the apprehension is resulted through the operation of the mind (anumana) which relates two things separately perceived. Hence upamana also is not an independent source of valid knowledge, cp. PSV, V, 169b, 4-5: re sig ne bar hjal ba ni ba lan dan ba min dag hdra bar rtogs pahi don can yin na // de la gsan las thos nas rtogs na sgra las byun ba yin la/ ran nid kyis yin na ni don gnis tshad ma gsan gyis rtogs na / yid kyis hdra bar rtog par byed pa yin la / de yan ishad ma gsan ma yin te. That pratyaksa and anumana are only two sources of valid knowledge (pramane dve eva) is thus proved by Dignaga. cp. NM, p. 3b, 10-11: HORI . $ E t. In respect of the number of pramana, the Vaisesika is in accord with Dignaga, but it is worth while noticing that the inconsistency of the Vaisesika doctrine was the incentive which gave rise to the Nyaya theory of four praminas, cp. Ui, Indo Tetsugaku Kenkyu, Vol. I, p. 304. The Vaisesika includes in the concept of pratyaksa the determinate perception or the perceptual judgement (savikalpaka-pratyaksa in later terminology), the characteristic of which consists in associating name or word with sense-datum, while sabda, which also the apprehension of the object by dint of words, is regarded as anumana. It is to make good for this defect of the Vaisesika doctrine that the Naiyayika treats sabda and upamana as separate sources of valid knowledge from pratyaksa and anumana. Accordingly, the theory of two sources of valid knowledge elucidated by Dignaga on his consistently critical ground is to be clearly distinguished from that of the Vaisesika. 13) Dignaga's etymological explanation of pratyaksa is : aksam aksam prati vartata iti pratyaksam (that kind of knowledge which exists in close connection with each sense faculty is pratyaksa), cp. NM, p. 3b, 17, cited at TSP, p. 373, 26, DP, p. 38, 26. Candrakirti, who adopts the Vedanta definition-pratyaksam aparoksam (that which is not beyond our ken is pratyaksa) ,

Page #7
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE (samanya-laksana), and we should recognize that direct knowledge takes the particular as its object and indirect knowledge the universal14). How then are those cognitions which cognize colour etc. in the form of a judgement 'this is non-eternal' 15) or which cognize one and the same object in assails the above explanation on the ground that it could also absurdly mean that a knowledge about the sense or a knowledge whose object is the sense is pratyaksa, cp. Prasannapada, p. 72, Stcherbatsky, The Conception of Buddhist Nirvana, p. 159. Stcherbatsky is erroneous in noting that the citation here is attributed to Prasastapada ; whose definition, however, slightly differs from the above: aksain aksam pratityotpadyata iti pratyaksam, Prasastapadabhasya, Chowkhamba Ed., p. 552, 28. To answer to the Candrakirti's hypercriticism, the Nyaya etymology-aksasyaksasya prati visayam vrttih pratyaksam, NBh, ad. I, i, 3 may be better. Dharmottara, whose interpretation pratyaksam iti prati-gatam asritam aksam. pratyaksa means that the sense-organ is approached, reposed upon)-does not differ materially from Dignaga's, makes distinction between the etymology and the real meaning, cp. NBT, ad. I, 3, p. 38, 3.4. Anumana (anu+matana) literally means such knowledge as preceded by some other knowledge. What precedes anumana is, according to the Naiyayikas, the preception of a mark and of the universal relation between this mark and the possessor of it, cp. NBh, ad. I, i, 5: linga-linginoh sambandha-darsanam lingadarsanam ca. However, Dignaga interpretes differently the implication of the preffix anu-, which is righteously taken by the Naiyayika as meaning "pascat' (afterwards) or '-purvaka' (preceded by), cp. NBh, ad. I, I, 3: mitena lingenarthasya pascan manam anumanam, NS, I, i, 5: tat-purvakam trividham anumanam. Dignaga's definition of svartha-anumana or anumana for one's own self is : tshul gsum paui rtags las rjes su dpag pani don (V's transl. rjes su dpag par bya bahi don) mthon ba gan yin pa de ni ran gi don gyi rjes su dpag paho (That apprehension of an object which is grounded on the triple-conditioned logical mark (tri-rupa-linga) is svartha anumana), PSV, II, 109a, 2-3, cp. NB, II, 3: tatra svartham trirupal lingad yad anume ye jnanam tad-anumanam. The prefix anu-- is thus replaced by ablative case-ending and is taken as implying logical ground. As the determinate perception is regarded by Dignaga as a kind of anumana, cp. below note 15), pratyaksa and anumana may adequately be rendered respectively as direct knowledge (or awareness)' and 'indirect knowledge (awareness) 14) According to the Vaisesika-Nyaya realism, every individual existence (vyakti), excepting the extreme universal (para-samanya) and the extreme individual (antya-visesa), is in possession of the universal (jati=samanya, visesa). Hence we perceive a thing, at the first moment, in indifferentiated obscurity (nirvikalpaka-pratyaksa), but later on determinately (savi kalpuka. pratyaksa), conjoining differentiated vyakti and jati (jati-visista-vyakti). Dignaga does not assent to this realist view. He makes an essential distinction between svalaksana and simanya. laksana; the former is the particular individuality which can never be generalized or replaced by a concept, and the latter is the universal which, being constructed through the generalizing faculty of our thought, is lacking reality. In correspondence with this essential distinction between two kinds of prameya, the theory of a radical distinction between two sources of knowledge (pramana-vyavastha) is definitely stated here. Evidently this theroy is set up in opposition to the Nyaya view of the coalescence of different sources of knowledge (pramana. samplava), i. e., the view that the same object can be cognized by any of four kinds of pramana, cp. NBh, ad. I, i. 3. Elaborative arguments made by Uddyotakara and Vacaspatimisra on this point are precisely traced by Stcherbatsky, and we need no further remark, cp. BL, Vol. II, Appendix II, p. 301ff. 15) Savikalpaka-pratyaksa or the determinate perception, which can be formulated in the judgement this is red', is recognized by most of the philosophical schools of India as a kind of pratyaksa. But Dignaga, who is in the opinion that pratyaksa apprehends only the particular stripped of the universal, is to be asked to explain by which pramana the determinate perception is caused. See postscript of this paper.

Page #8
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE (18.. (94b,12) repetition (as (17... Certainly there is such cognition as can be expressed in the judgement 'this is non-eternal', but this cognition, being related to the said two aspects of the object, is not another independent source of valid knowledge. In the case of such cognition, one cognizes at first colour etc. from two aspects, i. e., the inexpressible particularity (avya padesya=svalaksana) on the one hand and the universal colour-ness (varnatva= samanya-laksana) on the other, then, relating through the operation of the mind (manas) the universal colourness to the universal 'non-eternity', forms the judgement 'colour etc. are non-eternal'. Hence (such cognition) is not other source of valid knowledge (than indirect knowledge). Nor is there an independent source of valid knowledge in the case of recognition (pratyabhijna). (Indeed) one and the same object can be cognized in repetition, but this recognition also is not other independent source of valid knowledge. Why? Because, (if it be regarded as an independent source of knowledge,) then the fallacy of infinitude (anistha) would be logically concluded. In case, indeed, all kinds of cognition be deemed to be valid knowledge, there would be infinite number of the source of valid knowledge, and, as for instance, a recollected knowledge (which is commonly accepted as non-valid would also be regarded as valid). The term 'recollected knowledge' (smsta) in the verse stands for recollection (smsti). Such mental faculties as recollection, desire (iccha), anger (dvesa) and the like, operating on the object once cognized before, are not independent source of valid knowledge : 19) likewise (the recognition also should not be deemed to be an indpendent source of valid knowledge). 16) Recognition (pratyabhijna) is generally thought to be of the same kind as recollection (smrti), both being produced by dint of the impression (samskara) of previous cognition. What distinguishes it from recollection is the direct awareness of the object, which operate along with the impression. We are not acquainted with any particular school reckoning recogni. tion as an independent source of knowledge, but Sastradi pika, Chowkhamba Ed., p. 115, 1ff indicates that there are some (probably Vedantins) who hold that recognition is a separate means of apprehending an object from other five kinds of thought-construction. ion of spaikalpaka-prat yaksa as an independent source of valid knowledge. 18) Refutation of recognition as an independent source of valid knowledge. 19) The Mimasa definition of pramana-anadhigata-artha-gant, pramanam (the function of the source of knowledge consists in cognizing an object which is not yet cognized -is accepted by the Buddhist, cp. NBT, p. 19, 2: ata eva canadhigata-visayar pramanam. Vacaspatimisra pleaded against this definition on the ground that it cannot include the case of an object which has stability being cognized by a series of perception (dharavahika-vijnana), cp. NVTT, p. 21, 6-8. This criticism may be taken as being directed to the Mimarsakas, because, according to the Buddhists, there is no such object that has stability and duration.

Page #9
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE $ 3. DEFINITION OF DIRECT KNOWLEDGE V. 3cd : Direct knowledge is free from thought-construction. (Thought-construction) implies to associate name, genus etc. (with the object immediately perceived)20). [94b, 1) There, (i. e., among two kinds of valid knowledge,) direct knowledge is free from thought-construction (kalpana-apodha)21). The knowledge wherein there is no sign of thought-construction is direct knowledge. What, then, is this thought-construction? It implies to associate name (nama), genus (jati) etc. (, which are considered to be the efficient cause of the verbal designation ($abda-pravytti-nimitta), with the thing immediately perceived). In the case of spontaneous words (yadrccha-sabda, proper (22 20) pratyaksam kalpanapodham, nama-jaty-adi-yojana / 3c-d * cp. NV, p. 41, 19ff : apare tu manyante pratyaksam kalpanapodham iti. (NVTT, p. 153, 20: samprati Digndsya laksanam upanyasyati, apare iti.) atha keyam kalpana ? namajatyadi yojaneti. TSP, ad. k. 1221, p. 368, 23 : katham laksana-karenoktam namajatyadi yojana kalpaneti. 21) This definition of pratyaksa which, we find, is almost discussed in every Sanskrit work on epistemology and logic is framed on the basis of the essential distinction between svalaksana and samanya-laksana, cp. above note 14). Notwithstanding its simplicity, it is supposed to be so exhaustive that any other qualifier is possible to be added without yielding to a superfluity, cp. note 35). It is perhaps under the influence of this definition that Vacaspati interpreted the word avyapadesya in the defnition of pratyaksa in NS, I, I, 4 as referring to nirvikalpaka-pratyaksa and the word vyavasaya-atmaka as indicating savikdpaka-pratyaksa. 22) TSP, ad. k. 1224, p. 369, 22ff, NVTT, p. 153, 23ff. Dignaga is in affinity with the Grammarian in holding that the thought-construction is inseparable from the verbal expression, and, in this respect, differs in opinion from Vatsyayana, cp. Randle, Indian Logic in the Early Schools, pp. 119-120. Santaraksita and Kamalasila lay stress on the expression (ucyate> '(a thing...) is expressed in word' in these passages, and consider it to be the evidence of Dignaga's understanding of kalpana as being inseparably related with word (nama=sabda), and not with genus" etc. (jaty-adi), cp. TS & TSP, 1233. According to their interpretation, 'nama' in Dignaga's definition of kalpana as 'nama-jaty-adi-yojana' should be essentially distinguished from `jaty-adi'. Jati etc. being not recognized by Dignaga as real entities, jaty-adi-yojana is a heretical theory to be discarded ; Dignaga's own interpretation of kalpana is nama-yojana, cp. ibid., 1219--1221. Or, even when the existence of jati etc. be provisionally admitted, it must be noticed that these are related to a thing only through the medium of nama, cp. ibid., 12241225. After elaborating these arguments, they conclude that the coalescence with word (nama) is the characteristic feature of Dignaga's definition of kalpana. These arguments, admitting that they are not off the point in conclusion, seems to be not faithful to the original thought of the passages here, because it is evident that yadrccha-sabda is meant by the term 'nama': yadrccha-sabdesu namna visi sto'rtha ucyate 'dittha' iti. Classification of sabda into five categories is perhaps taken from the Grammarians, cp. Mahabhasya ad. Sivasutra 2, Vart., 1: catustayi sabdanam pravrttih, jati-sabdah, guna-sabdah, kriya-sabdah, yadrccha-sabdas ca caturthah. As regards dravya-sabda, the origin is obscure, but Prasastapada also adopts this category, cp. Randle, op. cit., p. 107ff. Dharmakirti is more prudent than Dignaga in defining kalpana as 'a distinct cognition of mental reflex which is capable of coalescing with a verbal designation' in order to include the thought-construction of infant and dumb person who have the potentiality of verbal designation but do not utter an actual word, cp. NB, 1, 5: abhilapa-samsarga-yogya. pratibhasa-pratitih kalpana.

Page #10
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE noun), a thing qualified by (or associated with) a name (as an efficient cause of verbal expression, e. g., Pittha-tva) 23) is expressed in the word 'Dittha'. In the case of genus-words (jati-sabda, common noun), a thing qualified by a genus (e. g., go-tva (cow-ness) is expressed in the word 'go' (cow). In the case of quality-words (guna-sabda, adjective), a thing qualified by a quality (e. g., sukla-tva ) is expressed in the word 'pacaka' (a cook). In the case of substance-words (dravya-sabda), a thing qualified by a substance (e. g., dandi-tva (stick-holder-ness) or visani-tva (horned-ness>) is expressed in the word 'dandin' (a stick-holder) or 'visanin' (the ...22) horned). (The characteristic feature of the thought-construction thus consists in the verbal designation of a thing through the association of name, genus etc. with it). * Here, (with regard to the last two cases of the above-mentioned five,) some (are in the different opinion that the efficient cause of verbal expression is nothing but the relation between an action or a substance and its possessor, and) maintain that (a thing) qualified by this relation (is expressed in the words pacaka', dandin' and so on)24). On the other hand, some others hold that a thing qualified by the word which essentially can denote no real entity (artha-Sunya) is expressed (in all five cases mentioned above) 25). 23) cp. Balamanorama ad. Siddhantakaumdi, 1781=MBH, V, i, 119: (Page #11
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE Anyway, that which is devoid of the thought-construction thus characterised is direct knowledge. (94b, 6) For what reason, then, the direct knowledge is called 'praty-aksa' (lit. being direct to sense-organ) and not 'prati-visaya' (lit. being direct to object), despite the fact that it is caused in dependence upon both the sense-organ and the object ? V. 4ab: It is signified by the sense-organ, for the sense-organ is the specific cause of it26). It is not signified by the object, such as colour and the like. The reason is that the object is a factor common (sadharana) to other kinds of knowledge because of its being a cause of the mind (mano-vijnana) or of the knowledge of other persons (anya-samtana) too. And it is generally known that (a thing or a fact) is designated by the name of its specific component (asadharana), as for instance, we use verbal expressions 'the sound of a drum' or 'the sprout of berley' to indicate a certain sound or a sprout, instead of naming it 'the sound of hand' or 'the sprout of earth', although the hand or the earth is also a cause in each case27). 95a, 1) That direct knowledge is free from thought-construction is established (thus on the ground that it is direct to the sense-organ). It is stated in an Abhidharma treatise too that "one whose visual sense is normal perceives blue (nilam vijanati) but does not apprehend that 'this is blue' (nilam iti vijanati). Facing to an object, he perceives the object in itself (artha-sahjna) but does not form an idea (dharma-samjna) of it" 28). $ 4. VARIETIES OF DIRECT KNOWLEDGE (95a, 2) If direct knowledge be absolutely devoid of thought-construction, then whence is this (stated in the Abhidharma treatise) :-"Five kinds of perceptive function take exclusion of non-cow' (anya-apoha, -vyavrtti). It is with the mind to make his own interpretation of genus etc. explicit that Dignaga interpolated this sentence here. Some others' (anye), accordingly, are the Buddhists, cp. TSP, ad. k. 1229, p. 371, PSVT, 19a, 1. 26) VPM, p. 177: asadharana-hetutvad aksais tad vyapadisyate /4a-b. 27) cp. Abhidharmakosa, Taisho, XXIX, p. 12b, 18-C, 2: Lit*iu prit ER ...... p # (ato'sadharanatvac ca tair eva nirdisyate)...... XXX LUX (anya-caksur-vijnanasyapi) RO H .... ZB R . (yatha bheri-sabdo yavankurah). Sanskrt is quoted from AKV, p. 87, 13ff. Candrakirti refers to this discussion after criticizing Dignaga's etymological expalanation of pratyaksa, cp. Prasannapada, p. 72. 28) cp. Prasannapada, p. 74. This quotation, says Stcherbatsky, could prove that Dignaga's theory of pratyaksa is foreshadowed in previous Sautrantika works, cp. The Conception of Buddhist Nirvana, p. 162, note 3.

Page #12
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE the congregation (of atoms) as their respective object"? 29) (The object of the sense being a single atom (paramanu), the congregation (samcita) should be apprehended by means of the thought-construction which unites together the perception of each constituent atom. It seems, therefore, to be incongruous to hold that the direct knowledge is free from thought-construction and yet it cognizes the congregation of atoms.) Again it is mentioned (in the Abhidharmakosa) that "these (mental functions) make the particular their respective object, so far as it is the particularity of outer seats (ayatana-svalaksana) and not of substance (=atom, dravya-svalaksana)" 30). How is this statement to be understood ? V. 4cd: There (in the Abhidharmakosa), (the perception, being caused by (the sense-organ through its contact with) many objects (=congregation of many atoms), makes the whole (samanya) its object when cognizing its own object31). Since it (=perception) is caused by the sense-organ through its contact with (congregated) many substances (=atoms), it is said to take the undivided whole (=the congregation itself) as its object on cognizing its own (outer) seats32). It, however, does not apprehend the whole through the thought-construction which, after perceiving split substances (=atoms), unites them together. (Therefore, there is no incongruity with the statement of the Abhidharma in asserting that the direct knowledge is free from thought-construction.) However, the author does not completely agree the above Abhidharma statement in respect of the object of the sense. The author's own view) is stated as follows:-33) 29) VPM, p. 176: samcitalambanah panca vijnanakayah. A similar passage is found in Abhidharmakosa, p. 12a, 26-28 : ...... Fitt i parit (sancitasrayalambanatvat, AKV, p. 86, 9-10), and this is attributed to the Vaibhasika. According to Vimsatika, k. 11 & Com., Alambanapariksa, kk. 1-5 & Com. and Trimsikabhasya, ad. k. 1, realists are divided into three groups in view of their theory concerning the object of cognition (alambana). The first group maintains that the object of cognition is an individual atom (paramanu), the second the congregation or the gathering of many atoms (samcita) and the third the union of atoms (samghata). It is obvious that the theory here referred to is that of the second group, which is reported by Kuei-chi to be the Vaibhasika, cp. Yamaguchi, Seshin Yuishiki no Genten Kaimei, p. 78, note 2. Vimsatikatika of Vinitadeva quotes exactly the same sentence explainig the theory of the second group. 30) Abhidharmakosa, p. 3a, 9-11 : *UHL FL IXH , H (ayatana-svalaksanam praty ete svalaksana-visaya na dravya-svalaksanam prati, cp. AKV p. 28, 10-16). Cp. VPM, p. 176. 31) VPM, p. 176, VA, p. 279, 10: tatranekartha-janyatvat svarthe samanya-gacaram // 4c-d // 32) ayatana=bahya-ayatana ; form, sound etc. When the perception operates, what is perceived is a form in its wholeness (ayatana-svalaksana), and is not the individual atom (dravyasvalaksana) which is the constituent of the form. 33) PSVT, 225, 2-33: Rejecting a divergent view in regard to the object to be cognized (spyod-yul las log par rtogs pa=gocara-vipratipatti), the author states definitely the impossiblity of the object being conceptually cognized (rnam par rtog pa med pa nid=avikalpatva).

Page #13
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE 13 V. 5: The substance constituted by many elements (=the congregation of atoms) can never be cognized by the sense-organ. What is perceptible by the sense-organ is the self-cognizable, inexpressible form34). Anyhow, (it is established) in this way that direct knowledge which is caused by five kinds of sense-organ is devoid of thought-construction. According to the doctrine maintained by others, some other qualifications (of direct knowledge) are to be added here35). However, 'being free from thought-construction' is enough (to define direct knowledge). V. 6ab: The mind, so far as it perceives object or it perceives internally desire etc., is also free from thought-construction36). The mind (manas), which, (though) leaning upon the object such as form and the like, operates in the form of an immediate awareness, is also free from thoughtconstruction37). Desire, anger, ignorance, pleasure, pain etc. (which are functions of 34) VA, p. 298, 1 (VPM, p. 189): dharmino 'neka-rupasya nendriyat sarvatha gatih / svasarvedyam anirdesyam rupam indriya-gocarah // 5 // cp. NM, p. 3b, 18-19: fi 3-F IVE IS HET EER. 35) NS, I, I, 4 defines pratyaksa as that knowledge which is produ ed by the contact of sanse-organ with object, and which is inexpressible, non-erroneous and determinate (indriyarthasamnikarsotpannan jnanam avyapadesyam avyabhicari vyavasayatmakam pratyaksam). Dignaga attacks this definition saying that the cognition caused by the sense-object contact is incapable of being expressible', 'being erroneous' or 'being determinate', PS, I, k. 19a-b, cited at VA, p. 338, 17, cp. my paper, Fragments of Pramanasamuccaya, p. 330. The qualification being inexpressible' is superfluous, because the expressible is cognized only by inference and can never be perceived through sense-organ. The error being attributed to the mind (manas) which is not sense-organ, the qualification 'non-erroneous' is also unnecessary. The nature of determination consisting in connecting the sense-datum with the universal which cannot be cognized by the sense, the third qualification does not hold good, cp. PSV, ad. I, 19, 97b, 8-98a, 8, cp. Kitagawa, Shori-gakuha no Genryo-setsu ni tai suru Jinna no Hihan, Nagoya Daigaku Bungaku-bu Kenkyu Ronshu, Tetsugaku, XXI, pp. 58-63. The second qualification is adopted also by the elder school of the Yogacara, cp. Yogacarabhumi, Taisho, XXX, 357a, 15-16: Wit (pratyaksa) # M ET. JEL (aparoksa). 2 (parikalpita-parikalpya-abhava). JELLET (abhranta). Dharmakirti adds again this qualification to the Dignanga's definition, cp. NB, I, 4: tatra pratyaksam kalpanapodham abhrantam, and the meaning of this addition is variously interpreted by his commentators, cp. my paper,Bukkyo-ronri-gakuha no Genryo-setsu ni kansuru Ichiko satsu, Indo-gaku Bukkyo-gaku Kenkyu, Vol. II, 1, pp. 123-124. Perhaps Dharmakirti's adoption of this qualification is meant for answering to the objection againt Dignaga that the cause of the erroneous cognition is not always the operation of the mind but the defect of sense-organ also is to be taken into account, cp. TS & TSP, 1313, 1314. 36) VPM, p. 191, VA, p. 303 : manasam cartha-ragadi-svasamvittir akalpika / 6a-b According to Jinendrabuddhi, the compound artha-ragadi-sva-sarvitti should be interpreted as implying artha-samvitti and ragadi-sva-samvitti, PSVT, 246, 4: don gyi sgra hdi ni ses byahi rnam grans so // hdod chags la sogs pa rnams kyi ran ni chags la sogs / ran no // 37) NBT, ad. I, 9 attempts to make clear the structure of mental sensation (manasapratyaksa). cp. BL, Vol. II, p. 311 ff.

Page #14
14 DIGNAGA'S THEORY OF DIRECT KNOWLEDGE the mind) are also direct knowledge in the sense of internal perception, because those have nothing to do with sense-organ. Likewise : V. 6cd: The intuition of the object itself which is exercised by yogins without relying upon the teacher's instruction (is also a direct knowledge, being free from thought-construction)38). The intuitive perception of yogins which is not mingled with thought-construction (95b, 1) pertaining to the reliable words of the teachers and which relates to a bare object is also a direct knowledge. (One may assert) that the idea (kal pana-jnana) also would be deemed to be a direct knowledge, (in case this is defined simply as being free from thought-construction). This is right. V. 7ab: The idea also is regarded (as a direct knowledge) in the case of self-cognition. However, when the idea is formed with regard to an object, it is not a direct knowledge, because of thought-construction being there39). The idea is not a direct knowledge similar to desire etc. in so far as it is formed with regard to an object, but in the case of self-cognition the idea is not non-directo). Hence there is no harm in defining direct knowledge as being free from thoughtconstruction. Such is indeed direct knowledge. 85. APPARENT DIRECT KNOWLEDGE V. 7cd-8ab: Illusion, cognition of empirical reality, inference, its result, recollection and affection are apparent direct knowledge and are accompanied by obscurity (sa-taimira)"). 38) VPM, p. 191 : yoginam guru-nirdesavyatibhinnartha-matra-drk // 6c-d // 39) VA, p. 331, VPM, p. 204: kalpanapi svasamvittav ista narthe vikalpanat / 7a-b 40) cp. below ad. I, 9-10. 1) VA, p. 332, 20, TSP, ad. k. 1324 : bhranti-samurti saj-jnanam anumananumanikam // 7c-d // smartabhilasikar ceti pratyaksabhasai sataimiram / 8a-b PSVT, 27b, 2ff says that four sorts of apparent direct knowledge are enumerated here, namely (1) illusion (bhranti), (2) cognition of empirical reality (sarvrti-saj-jnana), (3) inference (anumana), its result (anumanika), recollection (smarta) and affection (abhilasika) and (4) sa-taimira. Sa-taimira is, according to ibid., 286, 2, an obscure knowledge caused by the defect of the sense-organ (dban-po la ne-bar gnod-pa-las skyes-pa=indriya-upaghata-ja), such as timira (eye-disease) and the like. This interpretation seems not to be faithful to the original thought expressed in this verse. The last word 'sataimiram' is to be regarded as being in apposition

Page #15
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE 15 (95b, 3) Among these, illusive cognition, being caused through the conceptual understanding which takes, for instance, vapour floating over sands as real water, is an apparent direct-knowledge. Cognition of empirical reality, functioning as a conceptual apprehension which, superimposing an extraneous element (e.g., name) upon a thing itself, takes the latter as having the form of the former, is an apparent direct-knowledge. Inference and cognition resulted by it etc., comprehending conceptually what has been immediately perceived before, are not direct knowledge. $ 6. IDENTITY OF RESULT AND PROCESS OF COGNITION (95b, 5) Here also, (according to our opinion), V. Scd : Resulted content of the cognition being supposed to possess the act of cogaizing, it is simaltaneously the result and the process of cognition42). We do not admit here, as the realists do, that the result of cognition differs from the cognitive process*3). That cognition which has been resulted (is indeed to be regarded as the result of cognitive process in so far as it is the apprehension of the object, but from another point of view, it) is known as possessing with it the act of cognizing (sa-vyapara), because it arises in conformity with the form of the cognized object. In this latter sense, it is metaphorically called the cognitive process, i. e., the with the preceding word 'pratyaksabhasam'; otherwise 'ca' is necessary to be added after 'sataimiram'. Our opinion is supported by the fact that Dignaga explains in his own commen this verse three sorts of apparent direct knowledge only and not the forth. It is obvious that PSVT modified the original thought of this verse basing upon the theory of Dharmakirti, who, thinking it to be necessary to remove the erroneous cognition due to the defect of senseorgan from pratyaksa, adopted the qualification 'abhranta' in his definition of pratyaksa, cp. above note 35), Hasuba, Jinendrabuddhi ni yoru Jigenryo Kaishaku ni tsuite, Yamaguchi Hakushi Kanreki Kinen Indo-gaku Bukkyo-gaku Ronso, p. 205ff. 42) VA, p. 349, 5, VPM, p. 221, Nyayamanjari, Kashi Skt. Ser., p. 66, 20: sa-vyapara-pratitatvat pramanam phalam eva sat // 8c-d // 43) It is generally admitted by the Naiyayikas that the operation or the process of cognition (pramana) is distinguished from the resulted content (pramiti=pramana.phala), cp. NBh, ad. I, i, 3. Dignaga, when criticizing the Nyaya definition of pratyaksa, takes up this problem for discussion. If it be held, he says, that the determinate cognition is pramana in accordance with the Nyaya definition, then it would be unnecessry that the result differing from it should be produced, because the cognition is already determined. If again, he continues, it be assumed that pratyaksa-pramana is the perception of the universal (samanya) which is the qualifying adjunct of the individual, and that pramana-phala is the cognition of the individual qualified by the universal, then this assumption would lead us to the absurd conclusion that the object cognized in the process of perception differs from that actually cognized as the result of that process: this Nyaya doctrine is as ridicurous as the statement that the axe struck at the khadiratree cuts in the result the palasa-tree, PSV, ad. I, 22, 99a, 2ff, cp. Kitagawa, op. cit., pp. 13-14, TSP, ad. 1345, p. 399.

Page #16
16 DIGNAGA'S THEORY OF DIRECT KNOWLEDGE means of knowledge"). (We conclude thus that the result and the process of cognition are not different each other.) In case, however, the resulted cognition be debarred of activity, it cannot be called the process. As for instance, when the effect produced (e.g., the corn) agrees in kind with its cause (=the seed), they say that it has 'taken' the shape of its cause, but it does not agree with the common notion to say that this expression is made even when the effect is devoid of activity. Similar is the case with the discussion here. (That is to say, when the resulted content of the cognition possesses the act of cognizing, it is rightly supposed to take the form of the object.) 87. SELF-COGNITION V. 9: It can also be maintained here that the self-cognition is the result of cognitive process, because its nature consists in determining the object. And, the mental image bearing a resemblance to the object is the means of knowledge, through which the object is cognized45). (95b, 7) It can also be maintained here (in examining direct knowledge) that the selfcognition (svasamvitti) is the result (of cognitive process). When the cognition is caused, it has two sides, appearance of the subject (sva-abhasa) and appearance of the object (visaya-abhasa)46). The self-cognition (which takes place) between these 44) Kumarila raises objection against Dignaga's this theory assented to by Dharmakirti in his criticism of the Nyaya theory, cp. above note 43), but for different purpose. The axe, the instrument of cutting is distinct from the cut, the result,--this distinction is as generally recognized as the fact that the axe struck at the khadira-tree does not cut the palasa-tree in effect. Likewise, Kumarila concludes, the distinction between the instrument of cognition (pramana=prama-karana) and the cognition resulted through it (pramiti=pramana-phala) must not be neglected, cp. Slokavarttika, IV, 75, TS & TSP, 1345. The Buddhist answer to this objection is as follows: --The mere invariable concomitance (avinabhava) between the instrument of cognition and the object to be cognized does not provide sufficient ground for considering that instrument to be pramana. Through the instrument may the cognition of the object be produced, but it is not the actual act of cognizing. Moreover, absolutely speaking, all things being instantenous, the relation of producer and produced (utpadya-utpadaka-bhava) cannot be established between the instrument and the result. The relation between pramana and prameya must, therefore, be that of what determines and what is determined (vyavasthapya-vyavasthapakabhava), and what determines the object is the very act of cognizing, i. e., the appearance of the object in the mental reflex, cp. TSP, ad. k. 1346, NB & NBT, I, 18-19. 45) VPM, p. 215, 221, (cp. VA, p. 349, 7, p. 393, 28) : svasamvittih phalam va'tra tad-rupo hy artha-niscayah / visay'akarataivasya pramanan tena mi yate // 46) The theory that alayavijnana or store-consciousness appears or manifests itself (abhati, pratibhati, avabhati, khyati) as both the apparent subject (sva-abhasa=grahaka) and the apparent object (artha-abhasa=grahya) is met everywhere in Vijnanavada treatises, and we need no explanation of it here. It is evident from the context of this passage that Dignaga established

Page #17
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE two sides is the resulted content of the act of cognizing. Why? Because the nature of this self-cognition consists in determining the object (artha-niscaya)47). When the idea, in which the form of the external object is represented, is taken as the object of cognition, the self-cognition which is corresponding to that idea determines the object either as something desirable or undesirable. But when the mere external thing is thought to be the object of cognition, then the appearance of the object in the mental reflex of the cognizant is the means of cognizing this object. Although the cognition should be self-cognizable in this case too, the appearance of the subject (in the mental state of the cognizant) is disregarded, and to the mental reflex resembling to the object is attributed the role of the means of cognizing this object, for the object is cognized through that mental reflex. Whatsoever be the reflex of the object which has appeared in the cognition, whether it be the reflex of something white or nonwhite or of any other colour, this mental reflex which possesses the object within itself has the function of determining the object. (96a, 3) Thus, in accordance with the variety of (the nature of) cognition, the role of the cognizing agency or of the object to be cognized is hypothetically attributed (upacaryate) to respective factor in each case, because (absolutely speaking,) all elements of existence, (being instantenous) are devoid of any efficiency48) (and, in consequence, there is no element which is to be defined invariably as the cognizing agency or as the object to be cognized). The same content is stated (in the following verse). V. 10: That which appears in the image (in comformity with the form of the external thing) is the object of cognition, and the cognizing agency and the resulted content of cognition are respectively the apparent cognizant, (i.e., the his theory of knowledge on the ground of the Vijnanavada philosophy, although he does not use the term 'alayavijnana'. The theory of self-cognition (svasamvit) which marks the specific feature of his theory of direct knowledge is understandable only on the basis of the Vijnanavada doctrine. It deserves to be noticed that the verse 10 following the arguments here is cited in Dharmapala's commentary on Vijnaptimatratasiddhi as an evidence of Dignaga's theory of the triple division of vijnana (grahya-akara, grahaka-akara, svasamvit), which is criticized from the viewpoint of the Dharmapala's fourfold division theory (HTE E sva-svasamvit ? besides the above three), cp. Cheng-wei-shih-lun, Taisho, XXXI, p. 10b, 13-16: #LA . DET ER Hal EHHERR E EI, de la Vallee Poussin Vijnaptimatratasiddhi, La Siddhi de Hiuantsang, traduite et annotee, Tome 1, p. 139. 47) Niscaya is synonimous with vikalpa, kalpana, adhyavasaya etc., meaning thought-construction, cp. BL, II, Indices. Here, this term is used in different sense as is explained in the passages that follow. 48) cp. TSP, ad. k. 1346, p. 399, 12-13: yasman na paramarthikah karty-karan'adi-bhavo na nirvyaparatvat sarva-dharmanam, ibid., ad. 1222, p. 369, 11-12: naiva tu kascit karcid yojayati, nirvyaparatvat sarva-dharmanam, see above not 44).

Page #18
18 DIGNAGA'S THEORY OF DIRECT KNOWLEDGE appearance of the subject) and the self-cognition (procuced through the relation between apparent object and apparent subject, which actually are two sides of the same cognition). Therefore, these three factors of cognition are not different each other9). 88. TWO KINDS OF COGNITION AND SELF-COGNITION (96a, 5) How, then, is it recognized that cognition is of two kinds ? V. 11ab: Cognition is of two kinds, because of the constitutional difference between the cognition of the object and the cognition of that (cognition)50). That cognition by virture of which the (external) object such as form and the like is cognized consists of two sides, i.e.,) the image of the object and the apparent subject. But that cognition which introspects the above said cognition of the external object consists of the idea, which reflects the external object, and the corresponding apparent subject51). Otherwise, if the (first) cognition is in conformity with the (external) object only (and is lacking subjective side), or if it consists of the apparent subject only (and is devoid of the objective side), then the cognition of that (first) cognition would be of the same nature as (the first one, i. e.,) the cognition of the external object. Further, (if the difference between these two kinds of cognition be not admitted,) then there would be no possibility of subsequent cognitions appearing in conformity with the (form of) the past, remote object. Why? Because the past object, (being not there at the time of subsequent congnitions)52), cannot be their object. Hence, (there must be the cognition which is other in constitution than the cognition of the external thing and by virture of which the form of the past, remote thing is cognized. That kind of cognition is none other than the cognition which cognized the previous cognition possessed of the image of external thing. Thus) it is proved that the cognition is of two kinds. 49) VPM, p. 221, 229, Nyayamanjari, p. 67, 30-31 : yad abhasam prameyam tat pramana-phalate punah / grahak'akara-svasamvitti trayan natah prthak-kytam // 10 cp. above note 46). 50) VPM, p. 232, 244, VA, p. 425, 12 : visaya-jnana-taj-jnanavisesat tu dvi-rupata / 11a-b 51) Cp. note 46). 52) As all things are momentary, the object of the previous cognition does not continue to exist till the subsequent cognition arises. There is, accordingly, no possiblility of the same object being cognized by a series of perception, cp. note 48).

Page #19
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE 19 V. 11cd : (That there are two kinds of cognition is evident) also from (the fact that we have) recollection afterwards. The recollection (of something) does not occur to the mind without having experienced (it before) 53). The words 'also from the fact that we have recollection afterwards' in the verse (96b, 1) relate to the above-stated words 'cognition is of two kinds' (in the stanza c). Some time after we perceived something, the recollection of our cognition (that we have perceived it) occurs to our mind together with the recollection of the object. Therefore, the cognition is of two kinds, (i. e., the cognition of the external object and that of the cognition or of the idea,) and in consequence it should be maintained that it is self-cognizable. Why is this maintained)? Because, the recollection (of something) does not occur to the mind without having experienced it before. If a thing has not been perceived before, we have no recollection of it; as for instance, the recollection of the colour and like (occurs to our mind only when we have perceived it before. Likewise, the recollection of a previous cognition is possible on the assumption that the cognition was self-cognized before.) V. 12: If it be maintained that the recollected cognition is the apprehension of previous cognition by means of another cognition, then an infinite sequence would badly be resulted. Because, with regard to it (=second cognition) again will there be a recollection. Further, if it be so, the cognition of one object would not transit to that of another object. But, (in fact,) the transition of the cognition is commonly noticed54). (96b, 2) Some may hold this :- As colour and the like (are perceived by means of the other factor than themselves, i. e., by sense-organs), the cognition also is apprehended by means of another cognition. This is not right. Because, if the recollected cognition be assumed to be the apprehension (of the previous cognition) by means of another cognition, then the absurdity of infinite sequence (anistha), i. e., the endless series (anavastha) of cognition would be logically concluded55). (In order to be consistent, we have to admit that) this second cognition will also be apprehended by another 53) VA, p. 425, 5 : smyter uttarakalam ca na hy asav avibhavite // 11c-d // 54) Slokavarttikavyakhya, sunyavada-sthanam, ad. k. 27, Ed. Ramanatha Sastri, p. 247, 23-34 : jnanantaranubhave' nistha, tatrapi hi smrtih / visayantara-sancaras tatha na syat sa cesyate // 12 // 55) The Naiyayikas, who hold that a cognition must be proved by avoid to face this difficulty of infinite sequence, cp. NBI, II, i, 17-19. It is to overcome this difficulty that the later Naiyayikas establish the theory of anuvyavasaya or self-consciousness.

Page #20
DIGNAGA'S THEORY OF DIRECT KNOWLEDGE cognition. Why? Because, with regard to that (second cognition) again will there be a recollection, (the occurance of which is justifiable only on the assumption of the third cognition, and ad infinitum). In the case of that (second) cognition by virture of which the (first) cognition is apprehended, we have, as generally known, the recollection of it afterwards. Hence, if it be held that the cognition is apprehended by a different subsequent cognition, then the third, the fourth and thus) the endless series of cognitions would be logically concluded.56) Further, (if it be so, then the first cognition would be continuously apprehended by subsequent cognitions, and consequently), the cognition of one thing would not transit to that of another. But (in fact,) the transition of the cognition is the fact generally admitted. Therefore, it must necessarily be recognized that the cognition is selfcognizable. And that very (self-cognition) is the resulted content of the act of cognizing. In this way it is well proved that the direct knowledge is free from thoughtconstruction. 56) Candrakirti makes reference, in Madhyamakavatara, VI, 72, to the theory of self-cognition established through the argument on the recollection as is unfolded here, but he ascribes this theory to the Sautrantika, cp. S. Yamaguchi, Bukkyo ni okeru Mu to U to no Tairon, pp. 284285. Cp. also Pramanavarttika, II, k. 485. Postscript : Among two questions raised after the theory of radical distinction between two pramanas was expounded, the first one has been understood by the present writer as concerning savikalpaka-pratyaksa, see notes 15) & 17). This understanding, however, should be corrected so as to be in accord with Jinendrabuddhi, whose interpretation is as follows:- The proposition "this (colour) is non-eternal" (pratijna) is inferred through the minor premise "the colour is a product" (hetu) and the major premise "whatsoever is a product is non-eternal, e.g., a pot" (drstanta). If the theory of radical distinction between two pramanas should be maintained, then the above inference could not avoid committing fallacy. While the colour which is the subject of hetu is samanya-laksana, the colour which is the subject of pratijna is 'this' (svalaksana) immediately perceived. The subjects of hetu and pratijna being thus radically different each other, this hetu is incompatible with paksa-dharmatva, the first aspect of tri-rupa-linga. Dignaga, therefore, is asked to explain the ground of this proposition being valid, cp. PSVT, 146, 5-7.