On Dharmakirtisunderstanding Of Pramanabhuta And His Definition Of Pramana

Author(s): Helmut Krasser
Publisher: Helmut Krasser
Catalog link: https://jainqq.org/explore/269451/1

Page #1
Helmut Krasser On Dharmakirti's Understanding of pramanabhuta and His Definition of pramana* At the outset of the Pramanasiddhi chapter of his Pramanavarttika which Dharmakirti devotes to the explanation of the mangalasloka of Dignaga's Pramanasamuccaya, he comments on the term pramana bhuta' which, among other terms, is assigned by Dignaga as an epithet to the Buddha. The aim of his undertaking is to prove that the Bhaga. vat, in contrast to other concepts such as God (isvara) or the Veda conceived by opposing traditions, is the highest authority. For this purpose he first (PV 2.1-6) defines pramana and explains some problems involved. Then he equates the Bhagavat (PV 2.7a) with a pramana Thereupon he interprets the term bhuta as having been applied by Dignaga in order to exclude eternal pramanas such as God, etc.,' and shows (up to k. 28) that such concepts, even under the condition that they were assumed to be noneternal, cannot be accounted for." The problem with the notion of pramana in the compound pramanabhula which has been dealt with extensively by modern scholars and which also did not escape Dignaga's and Dharmakirti's commentators is, how can the Bhagavat be said to be a pramana, and pramana, in its . I would like to thank Prof. Karin Preisendana for reading the manuscript of this paper so meticulously, and for her many valuable comments and suggestions - This paper is a revised version of the one presented at the XXXVIth International Congress of Asian and North African Studies, Montreal, Aug. 27 - Sept. 2, 2000. The meaning of the term promanabhuta has been discussed quite frequently cf.e.g., Seyfort Ruegg 1994, 1995 and recently Hakamaya 2000 (without knowing Seyfort Ruegg's articles, however). "I do not subscribe to Prajnakaragupta's interpretation of bhula as being used to dispel error (bhrantiniurtyartham PVA, 84,5). Its interpretation of most of the Indian commentators as jata, panna and the like seems to be more probable, as I consider this the starting point for Dharmakirti's digression on the refutation of eternal pramanas in PV 2.8-28 For an analysis of the chapter, see, e... the table of contents and introduce tion in Vetter 1990: 5 and 13-35, Inami - Tillemans 1986, or the first chapter in Franco 1997: 15-43 WZKN 45 (2001) 173-199

Page #2
174 H. Krasser turn, to be cognition? It is obvious that the Bhagavat is not knowledge or cognition, at least on a conventional level, and there is no room for doubt that pramana is considered to be cognition, for Dignaga himself states in his Pramanasamuccaya that there are only two kinds of cognition that can be considered as pramana (pratyaksam anumanam ca pramane - de gnis kho na ste [PS 1.2ab and Vrtti]), whereas Dharmakirti defines it as being reliable cognition (pramanam avisamvadi jnanam [PV 2.lab]). Moreover, both consider pramana to be the resultant cognition, i.e. pramanaphala. In order to account for this problem, roughly two kinds of solution have been proposed: 1) pramana should be understood either in an extended sense, as authority, or metaphorically or 2) bhuta should be understood in the sense of a comparison. Thus, pramanabhuta is translated either as "[he] who has become authoritative", or as "[the Bhagavant] who is (like ?) a (means of) correct knowledge". The first interpretation of pramana as being employed in an extended sense has been strongly argued for by Tillemans (1993: 5-9), whereas Dunne suggests a metaphorical interpretation (1999: 254262), and the second has been argued for by Seyfort Ruegg in his papers of 1994 and 1995. Seyfort Ruegg (1994: 315), of course being aware of Dharmakirti's comment on bhula, relates his rendering of This question is explicitly raised in a purvapaksa in Sakyabuddhi's PVT (quoted and translated below p. 179f.): mnon sum dan rjes su dpag pa ni tshad ma yin pa de bas na ji llar na boom Idan 'das de'i ran bein can ma yin pa la de skad du ce na. On an absolute level, however, he is considered to consist of perception. Sakyabuddhi, for example, presents the idea that the Bhagavat, on account of his being stainless, non-conceptual and unerring wisdom, actually is pratyaksapramana in a purvapaksa and agrees with the Bhagavat's description (cf. below, p. 179f; for the Sanskrit as preserved by Vibhuticandra cf. below n. 22: bhavanabalanispannanirmalavikalpabhrantajnanatmakatvad bhagavatah pratyaksapramanasvabhavala saksad asty eva). This concept is also shared by Prajnakaragupta (pratyaksarupa eva bhagavan pramanam; cf. below, n. 31). Tillemans 1993: 3. Seyfort Ruegg 1995: 819. References to and discussion of earlier "literal" interpretations such as Steinkellner's "[he] who has become a means of valid cognition" are to be found here. Dunne leaves the term pramanabhuta untranslated. abhutavinivrllaye/bhuloktih (PV 2.7bc) "The mention of bhala serves to avert [the wrong supposition of pramanas] that did not come into existence". On Dharmakirti's Understanding of pramanabhuta pramanabhuta as "pramana-like" to Dignaga's understanding only. although he is of the opinion that "the use of 'bhuta in the meaning 'like'... is found even in some of the post-Dharmakirti commentaries of the Buddhist Pramana-school known to us either from the surviving Sanskrit texts or from their Tibetan translations"." Both Tillemans and Dunne derive their understanding from Devendrabuddhi's commentary (and similar passages in other texts) on the phrase tadvat pramanam bhagavan (PV 2.7a) presuming that the metaphorical meaning is conveyed by the word tadvat. Seyfort Ruegg, on the other hand, bases his statement that the use of 'bhuta in the meaning "like" is found even in post-Dharmakirtian texts on passages such as the aforementioned explanation by Devendrabuddhi. 175 As an explanation in Jinendrabuddhi's Pramanasamuccayatika (PST)" dealing with this very topic sheds new light on the problem, I first will have a closer look at this explanation, then reconsider the translations of the passages at stake in Devendrabuddhi's Pramanavarttikapanjika (PVP), then propose a new interpretation of PV 2.7a and finally reconsider the meaning of the two qualifications of pramana, avisamradi jnanam and ajnalarthaprakasa, in PV 2.1ab and 5c respectively. The passage where Jinendrabuddhi explains pramanabhula of PS 1.la runs as follows (1b4-2a1): pramanabhulayelyadi. yatha Ivam merus tvam samudro 'si natha tvam kalpapadapah | Ivam suvaidyah pradipas tvam tvam eval paramah plavah || ily atrevasabda prayogam antarenapi ladartho gamyate, tathehapiti bhaga van pramanam iva pramanam. "yatha hi pralyaksadipramanam purusarthopayogino nadhigatasyarthasya prakasakam avisamvadakam ca. tatha bhagavan api yatra parah purusarthah pratibaddhas caturaryasalyalaksane lattve tadvisayam jnanam asadya nihsreyasarthinam tasyanadhi galasya prakasako visamvadakas ca. lasmat pramanasadharmyat pramanam. bhuto jatah, utpanna ity arthah, bhutavacanam abhutasya nitya syesvaradeh pramanasya paraparikalpitasya pratisedhartham. pramanam casau bhulas ceti pramanabhutah, tasmai pramanabhulaya Seyfort Ruegg 1994: 311 (references to the texts are given here); the same idea is expressed in Seyfort Ruegg 1995: 821. Until now only the Tibetan translation of dPan lotsava Blo gros brtan pa (1276-1342) was available. eva has no equivalent in Tibetan. 13 Cf. Vibhu 518,29-519.4 (cf. Steinkellner 1980: 100). "The text is rendered in Tibetan in the following way (PST, D2a1-2b1 = Q2b-7): Ishad mar gyur pa ies pa la sogs pa ste ji ltar | khyod ni ri rah rgya mtsho

Page #3
176 H Kramer On Dharmakirti's 'rulerstanding of prema nakala On "pramanabhutaya". etc. Just as in this (praise You are (mount] Meru: You are the ocean: oh Lord, You are the wishing-tree (kalpapadape): You are the perfect healer: You are the light: You are the best boat (plara)! the meaning of the word "like") is understood, even without the employment of the word "like" (ira), it is also the case here in the compound praminahital. Thus, the Bhagavat is valid cognition inasmuch as he is like i.e. similar to a valid cognition. Just as a valid cognition, namely perception and so on, illuminates an unknown object (artha) that is useful for a human purpose (artha) and is reliable (with regard to that object), in the same way the Bhagavat, too, after he has gained a cognition of that to which the highest human purpose is bound. [i.e. the reality characterized by the four noble truths, illuminates that reality which is not known to those striving for the highest good (nifreyasa) and is reliable (with regard to that). Therefore she is said to be a pramda due to this similarity to valid cognitions the word iva being self understood). bhuta (means) "he has come into existence in the sense of "he has arisen". The mention of bhula werves to negate pramanas that have not come into existence (abhila), i.e.) that are eternal, such as God (isvara), inasmuch as they are assumed by others (i.e. non Buddhists) The compound) pramanabhula is a karmadharaya mean ing) "he is a pramana, and he has come into existence"." (Having paid obeisance) to him who is valid cognition (pramanabhala) and has come into existence According to this explanation of Jinendrabuddhi, the word pramana is applied to the Buddha not in an extended sense but metaphorically and does not require any qualification in order to be understood as metaphor. Thus, at least if one follows his interpretation, there is no need to assume that the metaphorical meaning is conveyed by tadual, and it is also clear that his explanation does not support the interpretation of bhuta as "like" "6 Jinendrabuddhi's exposition provides us with a key that opens the door to a better understanding of the Tibetan translation of Devendrabud dhi's remarks. It not only enables us to look with another eye upon these materials, but, as the words pramanam iva are rendered by Ishad ma dan 'dra bas (PST.. D2a3; cf. above, n. 14), it also draws our attention to the fact that in other texts, too, the phrase tshad ma dan dra bas may represent a Sanskrit pramanam iva. If we now take the Tibetan Ishad ma dan 'dra bas na in Devendrabuddhi's PVP as equivalent to pramanam iva, his introductory remarks on PV 2.lab where pramana is defined, namely (PVP DIb2-3 = Q2a1-3): Ishad mar gyur pa ze bya ba ni ishad mar 'khruns pao || Ishad ma dan 'dra bas na tshad ma ste bcom Idan 'das (PV 2.7a) 80 | tshad mar no bo de ci zig yin na langis de dan 'dra bar dam bca' bar byed ce na Ishad ma slu med can ses pa || (PV 2.lab) zes bya ba smos le can be translated as [The compound) pramanabhata sin PS LI means "the Bhagavat) is a pramana, and he has come into existence" (shad mar 'khruns pa khyod moon khyod dpag bramkan thusleiman pa baan khyod aron ma khyod 11 khyod ni agrol byed dam pao || ses pa 'di 'dra bar sgrasbyar ba med par yar der don rroge pa deltar 'dir yan no se pas ishad ma dat 'drabas boom Idan 'daslahad ma' Il gangi phyir ji liar non sum la soge pa i Lahad ma ni skyes buy don la fe bar mkho besar ma rogs pai don gial bar byed in mi baba yin pa de llar bcom Idan das kyar gan la skyes bu'don gyi mchograg las pa 'phage pa i bden pa bai Y mtshan Rid kyi de kho na id la de yucan gyises pa bskyed the par legs padon du gier ba rams lakhon du ma ckud pa de goal bar byed pa dan mi balu ba de'i phyir tahad ma dan chos mishuis pa riid kyi phyir tahad mao Il gyur pa ni skyes pa ste byun ba ses pa i don to il gyur pa'i shig si ma byut berlag pa dban phyug la soge pali tshad ma gian gyis yons briage pa dgag pa i don duo || tahad mo yon di yin la gyur pa'an yin pos Ishad mar gyur pa ste Lahad mar gyur pa de la o ll. That the analysis of the compound is to be understood as that of a karma dharaya and not as that of a duandos as suggested in Steinkellner 1989: 181, was expressed to me by Steinkellner himself in the meantime and is also stated in Franco 1997: 16. n. 3. Instances of karmadharaya compounds are collected in Hakamaya 2001 16 The last point is made clear by an interesting note of Vibhaticandra who, when rephrasing the last part of the above-mentioned passage of Jinendrabuddhi's PST. explicitly declares that bila does not have the meaning of iva. This is inter esting inasmuch as it is the only remark I am aware of that indicates a possible understanding of bhuta in pramanabhuta as meaning "like" (iva) and thus supports Seyfort Ruegg's supposition. The statement is meaningful only when it is understood as being directed against someone who holds the view that bhala means in C. Vibhu 519,2-4: bhata wpannah, bhitaracanam aprajatasyedarddeh paraparikal pitasya pratisedhartham. ivarthas tu aimarthyagala ili na tadartham elad ili rakyale "bhuta (means) has arisen. The mention of bhita serves to negate pramanas) such as God (isvara) that have not come into existence (aprajata), inasmuch as they are imagined by others li.e. non-Buddhists). The meaning 'like' (na) is understood due to the adequacy (of just this meaning in the present context). Thus it will be explained that the mention of brita) does not have that meaning". - This remark of Vibhaticandra and a part of the Tibetan translation of PST quoted above (n. 14) have been referred to by Seyfort Ruegg (1994: 311, n. 39) in corroboration of his interpretation

Page #4
178 H Kramer On Dharmakirtis inderstanding of pribble 179 pramanajalah)." The Bhagavat is valid cognition (inasmuch as he is similar to a valid cognition. What is this nature of pramana so that it is asserted that the Bhagavat) is similar to it? (Reply: Valid cognition is reliable cognition." On the interpretation of Ishad mar 'khruns pa as a karmadharaya see Sakyabuddhi's explanation and the Sanskrit text of Sakyabuddhi's comments on this passage preserved in Vibhuticandra's notes (reproduced below, n. 20). The sentence Ishad ma dan 'dra bas na tshad ma se bcom Idan 'das so probably corresponds to a Sanskrit text as available in Jinendrabuddhi's PST: bhagavan pramanam iva pramanam = lshad ma dan dra bas bcom ldan das Ishad ma'o (PST. D2a3). This is also corro borated by Sakyabuddhi's explanation of this sentence according to which the word pramana implies a similarity (PVT D71b5-72a2 = Q86b4-87a2): de la 'grel pa mdaad pa bdag flid kyis tshigs su bcad pagrel pa mdaad cin shad mar gyur pa zes bya ba de 'chad pa lalahad mar 'khruris pages bya bar gyur palagra ni 'khruris pai don to c had ma yan de yin la gyur pa yan de yin pas na ishad mar gyur pa'o de llar na ishod me rlag par ring pa bral (D:gsal) ba yin no morsum den rjes su dpag pa ni Lahad ma yin pa de bas na jiltar na bcom Idan das dei ran bin can ma yin pa la de skad du ce na Ishad ma dari 'dra bas na Ishad masle ses bya ba smos te tshad ma'i agradpe nan du dus pa can yin no sex bya ba i don to || gal te begoms pa i stobs las grub pa dri ma med param par mi rlog pa khrul pa med pa'i ye tes kyi bdag find can yin pa'i phyir bcom Idan das ni mhon sum gyi Ishad mairan bin can find du dros subeuge pa nid yin na fie bar blags pa la brier pas cu zig bya senadila skyon yod pa ma yin le rlog pa dan boas par des pa'ignas skabe la dgons nas brjod pair phyir rozes bya ba ni gian dag gi yin no ll dir gal te bcom Idan das ni ji skad du bead paitshad ma'i bdag niid can yin pa della na yan de llar rab tu grage pa ma yin no de bas na tha sfiad du byas pa' lshad max dper mdad payin no ie8 bya ba 'di ni rige pa yin no By relating (pitanikam kurvan) this i.e. PV 2.lab, etc.] to the stoka [i.e. PS 1.1] the Vrttikara [i.e. Devendrabuddhi] himself explains the [expression) pramanabhutaya (PS 11c) (with the phrase) pramunajata, (which means that the word bhula has the meaning of "coming into existence" (The compound pramanabhuta is to be understood as a kar. madharaya in the sense of: He fi.e. the Bhagavat) is a pramana, and he has come into existence. In this way the assumption (rlog pa) of an eternal pramana is rejected. (Opponent:J" pramanas are direct percep. tion and inference. Thus, how can the Bhagavat who is not of their nature be said to be this (ie. pramana? [Therefore Devendrabuddhi) says: "[The Bhagavat is) valid cognition inasmuch as he is similar to a valid cognition". This means that the word pramana implies a similar ity. (Opponent:]" "The Bhagavat essentially is stainless, non-concep * bhaparalah praya belpaj inlandvasthanay "This interpretation of bhila is also known to Ravigupta (PVV 294a1), de la bur gan gyur pa de ni tshad mar 'khruns ses bya'o ll. IN This passage is partly referred to in Tillemans 1993: 61, n. 7; the whole passage is quoted and translated in Seyfort Ruegg 1994: 312 and Dunne 1999: 255, n. 7. As a further piece of evidence preserved in the Sanskrit original that this line of commentators used the wording pramdam iva in this context, another note of Vibhaticandra referred to in Dunne 1999: 2551, n. 7 may be adduced (Vibhu 521.28-30): aara ca sabdo 'nyatra prayujamno narbhutopamartha prayujyala ily aha - pramanam iveli praminasabdo jane mukhyah taratra tu kena sidhar. myenopamanopameyatam ily ala-avisamuddili. "And any word inasmuch as it is employed for something different [from what it primarily denotea) is employed referring to an implied similarity. Therefore he says: 'like premagal. The word pramana primarily (denotes cognition Grana). By which similarity now is in the other case (when it is metaphorically applied to the Bhagavat] (the cognition that to which the Bhagavat) is represented as similar and the Bhagavat that which is similar to it? In response to this question, he says: 'reliable. I have not been able to identify the source of this note and it is possible that Vibhaticandra put together bits and pieces from different passages in the manuscripts available to him, combining them with his own words. C. Vibha 521.26-27: lad atraurttikarah ilokapitanika u n pramanabhuMayely (PS Lla) elal sayam yicale - pramanajala ili bhulasabdah produrbha varthah (* Vibho 521.27 reads pramanam jata ili... which seems very unlikely. The emendation is based on the Tibetan translation Lahad mar thru pa): cf. Steinkellner 1981: 290 (frag. 4). # C. PST 2.8: pramdam odsou bhutas ceti promdiabetas cf. Steinkellner 1980: 100 bhagavan, lathapy # Cr Vibha 522,1-5: nanu bhavanabalanispannanirmaldvikalpaberantaj final makalvad" bhagaralah pratyakpapramanasabhavala wikaid asty era, kim upaca rdirayeneti cel, adoro yam, savikalpojlandasthi ayenabhidhanddily eke. idam te aira yuklam-yady api yathoklapramanamakan salam bhagans, tathapyaamua vaharibo advanasthabhedal, tatal simoyavaharikapramenopamiyale (This is the reading proposed in Steinkellner 1981: 290 (frag 5]: cf. also Dunne 1999: 256, n. 9. Vibhu 522. I reads: Mandbalanispannaniskalp wikarpd ikalpabhrantajnima kated. The emendation of asrayen Vibho 522,3] to adayend' is based on the Tibetan dgn mas-sada has no equivalent in Tibetan and may be an addition of Vibhuticandra). atra de la refers to PV 2.lab just mentioned by Sakyabuddhi in the preceding sentence (PVT D71b5): Lahad ma to bya ba la sogs pas ses bya ba ni Ishad ma slu med can ses pa sex bya ba la sogs pas 801 # The remaining section of the text is also referred to and translated in Inami 1994: 29., n. 2. # The remaining section of the text is also referred to and translated in Dunne 1999: 256 and 258, n. 10

Page #5
INO H. Kramer On Dharmakirtis inderstanding of promanable 181 The Blessed One is a pramana like that Dunne 1999: 261. n. 15 (aforementioned definition) (Free rendering:) ... like the two preceding items Oetke 1999a: 249 also the Buddha is a pramina so no yo ni ... (= Like this...) Hakama ya 2000: 316 tual and unerring wisdom that has been accomplished by force of his meditational practice. Hence his being by nature a pramina, namely perceptionapplies (only eva) directly. so why does one need to rely on a metaphor!" Some say that this is not a fault because she is said to be pramana intending a state of conceptual cognition. But here the correct response is as follows: Even if the Bhagavat has the nature of the aforementioned pramana, he is nevertheless not commonly known as such. Therefore he is represented as similar to a conventional muya taharika) pramana. The material presented so far clearly shows that both Devendrabuddhi as well as Sakyabuddhi interpret pramana in pramanabhula in the same way we have seen with Jinendrabuddhi, as implying a similarity, and that its metaphorical use is thus justified. Then, how can we understand Dharmakirti's remark ladvat pramanam bhagavan (PV 2.7a) meaningfully? Until now, scholars have unanimously understood tadvat in a comparative sense. However, their interpretations of lad differ: the anaphoric pronoun is either specified as referring to PV 2.1-6 or specified as relating to the defining characteristics in PV 2.lab and 5c, or it remains unspecified. A list of translations I am aware of, in chronological order, displays the following variety: Before I proceed to look at Devendrabuddhi's commentary, it should be noted that -val in tadual is usually understood by the Tibetan transla tors in a possessive sense, so that PV 2.7a is rendered into Tibetan corresponding to Sa skya pandita's and Sakyasribhadra's translation of the Pramanavarttika itself as de Idan bcom Idan Ishad ma nid.* An exception to this is the rendering by Subhutisrisanti and dGe ba'i blo gros who were responsible for the first translation of the Pramanavarttika (not preserved in the bsTan 'gyur) together with Devendrabuddhi's Panjika (included in the bsTan 'gyur) where ladval is rendered as de bein." It also may be mentioned that Prajnakaragupta favors a possessive meaning of -vat in tadvat and a literal interpretation of pramana in pramanabhuta. Such a valid knowledge instrument is the Blessed Nagatomi 1967: 15 One. In such a way is the Blessed One an authority. Jackson 1988: 343 The Lord (Buddha) is a means of valid cognition van Bijlert 1989: 158 such as that which we have discussed in PV II.1.6) sore lo doyo ni ... ("Similar to that ...") Inami 1994: 17 (Translating from Sanskrit:) Like that, the Tillemans 1993: 6 Illustrious One is a pramana. (Translating PV 2.7a as rendered into Tibetan in PVP (s. below, p. 182: de bzin bcom Idan Ishad ma niid):) Similarly the Illustrious One is a pramana. So sin view of what is stated in verses 1-6), the Seyfort Ruegg 1995: 820 Lord [being] (a means of) correct/efficacious knowledge ... The Exalted One is such a means of knowledge. Franco 1997:56 17 For an overview on the various translations of the PV and its commentaries into Tibetan cf. Mejor 1991: 179-193. PVA 27b6; PVAT 39a1, 39b2, 10263, 4023 (only de Idan); PVV 305a6; cf. also Inami 1994:30, n. 2. As Sakyabuddhi does not comment on tadual we have no translation. >> Cf. Mejor 1991: 180f This fact was already noticed by Gyal tahab rje; see Tillemans 1993: 7. n. 7: "rNam grel thar lam gral byed Vol. I. p. 236,10-11: lha dbang blo'i 'grel a gyurlar na de bzhin zhes dpe dang sbyar roll 'Following the translation of Devendrabuddhi's commentary. ladvat like that') is the connection with an example.' royal tshab then proceeds to argue against Devendrabuddhi". More precisely, Gyal tshab argues against the interpretation as proposed by these translators. In Jackson's translation (1993: 188) "according to Devendrabuddhi" gyur is not consid ered. 11 Cr his comment on PV 2.7a (PVA, 84,1-2): Ladeal pramanam bhagavan tathagalo hi bhagavams tadvan ili krivi pratyaksarupa eva bhagavan pramanam. "The Bhagavat is a pramana which possesses those two defining characteristics For having in mind (krted) that the Tathagata, [here called the Bhagavat, possesses those two defining characteristics) (Dharmakirti said that the Bhagavat, indeed being of the nature of perception, is pramana". This passage is also referred to by Tillemans 1993: 7, n. 7, Inami 1994:30, n. 2 and Franco 1997: 17. n. 3. Based on this formulation Franco unnecessarily considers the possibility that Prajnakara. gupta read ladvan instead of ladval in PV 2.7a. I agree with Franco that in Seyfort Ruegg's translation of pratyaksarupa ewa bhagavan pramanam as "the Bhagavat being similar to (or: of the nature of) pratyaksa is pramana" (1994:311) the variant * This pirvapaksa displays the position which is normally attributed to Pra. jNakaragupta cf. below, n. 31

Page #6
182 H. Kranser On Dharmakirtis muerstanding of proaminahala pramana is that whose nature is indicated by the twofold defining char acteristic as explained before." Keeping this in mind I will now look at Devendrabuddhi's explanation. leaving, for the time being, PV 2.7a untranslated (PVP D6b4-7 = Q7b1-5): 139i skad du biad pa'i mishan riid rnam pa griis kyix" bulan paino bo can gan yin pa de ni tshad ma yin no | *** de bain boom Idan tahad ma nid * (PV 2.7a) 'di tahad ma dan yan ci sig mishuns na gan gis na tahad ma dan dra bas Lahad ma nid yin shad ma i mishan vid ram pa gilis riid dan mishun pa yin no || jillar mon par 'dod pa bzin du bugrub par bya bai don la mislu ba'i phyir dan mises pa'i don geal bar byed pa'i phyirishad ma riid yin noll de llar na bcom Idan 'das kyar mes par legs pa la soge par mishan niid can gyi skyes buli don la mislu bar mdzad pa dan rloge pa por" skyen buli don bagrub par bya ba mi ses pa slon par mdsad pal phyir lshad ma rid yin no l shad ma flid kyi chos mihun perlid can 'chad par gyur bayan grub par byas nasi dper byas pa yin no 11" proposed in parentheses is to be preferred. The rendering of pratyakarupa as "being similar to pratyaksa" does not take into account that Prajnakaragupta considers the Bhagavat to be paramarthikapramana. a Cf. PVV 9,15: yathokladvividhalaksanam uktam yal pramanam, ladvad bhagavan pramanam. "A pramana has been taught by Dharmakirti as possessing the twofold defining characteristic as explained in PV 2.lab and 50). The Bhagavat is a pramana ljust like this (tadual) pramana as just determined, and not in an other sense". If this text as transmitted is accepted, one has to assume that Manorathanandin is the only commentator who, without giving any further explanation, understood al as having a comparative meaning. On the other hand, Manorathanandin normally follows the interpretation of Devendrabuddhi. Thus, a corruption of the text cannot be excluded. However, as there is no evidence for this being the case and because it is possible that Manorathanandin here deviates from Devendrabuddhi, there is no basis for a correction of the text. The reading as suggested by Inami (1994: 29, n. 1) and Dunne (1999: 280, n. 62) which is also found in the edition of the PVV prepared by Ram Chandra Pandeya (Delhi 1999. p. 5,24). namely yathoklad vividhalaksanayuktam instead of 'lakaawam welam, does not contribute to a possible solution of the problem 1 Cr Vibha 9, n. 7: dvividhena yathoktena lakanena nirdistam yal, elal pramdnam: identified in Frauwallner 1960: 121. # C. dvividhena Vibhu 9. n. 7: DQ: mam pa griis 15 Cr nirdislam Vibha 9, n. 7: D: brlen. Qom. n. 12 C. PVV 9.12: yathabhihitasya salyacatuslagasyviamvadan tasyaiva parair ajnalasya prakasanac ca. D: sgrub. * D: poi # C Vibhu 9, n. 7pramanaddharmyam tuddhayisyemam siddham kru dahrtam: identified in Frauwallner 1960: 121. +1 D: 'chad par 'gyur bar byos mas In what way is he similar to a pramana so that, inasmuch as he is similar to a pramana (Ishad ma dan dra bas, pramanam ira), he is said to be pramana? He is similar in having the twofold defining characteristic. Just as (a cognition is said to be a pramana because it is reliable (mi slu ba = avisamuddin) with regard to a (human) purpose to be accom plished as intended, and because it illuminates a (previously, unknown state of affairs (mi des pa'i don gsal bar byed pa = ajnalarthaprakisana), in the same way the Bhagavat too is said to be pramana because he is reliable with regard to the human purpose defined as the highest good (nes par lego pa, nihareyasa) and so on, and because he makes known [these previously unknown human purposes to be accomplished by the (person] who (now) realizes these purposes) The (Bhagavat's similarity to a pramana, however, which will be es. tablished in the remaining part of the Pramanasiddhi chapter). Was applied (udahpla) [by Dharmakirti in PV 2.7a), taking (krtc) it for granted (siddha). This passage confirms the results gained from Devendrabuddhi's introductory remarks on PV 2.lab treated above (p. 1777.), namely that the metaphoric use of pramana is self-understood without any further re. quirements. It also stresses the fact that the Bhagavat is similar to a pramana because he is avisamvadin and because he makes known a previously unknown state of affairs, in other words, he possesses these defining characteristics of a pramana. If we now look at Devendrabud dhi's introductory sentence to PV 2.7a, namely ji skad du bead pai mishan nid rnam pa gilis kyis batan pa'i no bo can gan yin pa de ni Ishad ma yin no, it is quite clear that he understood -val in ladval not in a comparative but in a possessive sense. The rendering of tadval into de bzin by Subhatisrisanti and dGe ba'i blo gros thus is obviously based on their misunderstanding of Devendrabuddhi's intention, and this mis. take has been corrected later on by Sa skya pandita and Sakyasribha dra. Thus, following Devendrabuddhi's interpretation of tadual pramanam bhagavan (PV 2.7a), which also has been accepted by Jinendrabuddhi This sentence is also referred to in Dunne (1999: 280, n. 62) who, like Inami (1994:29, n. 1), prefers the reading rnam pa gris brien po' no bo can instead of ram pagilis balan pano bo can and translates "A pramana is that whose nature depends upon the aforementioned two kinds of characteristics".

Page #7
18+ H Kramer On Dharmakirtis nderstanding of pramanabkula 185 and the Tibetan tradition at least from Sa skya pandita onwards. I would propose the following translation: "The Bhagavat is a pramana that has the twofold defining characteristic]":" for pramanabhula I suggest the rendering one who has come into existence being a pramana". In both cases I understand pramana as "valid cognition in a metaphorical sense If we now consider Devendrabuddhi's concluding remark in the statement treated above, namely that the Buddha's similarity to a pramana will be established later in the Pramanasiddhi chapter (tshad ma nid kyi chos mthun pa vid can 'chad par 'gyur ba grub par byas nas dper byas pa yin no || PVP D6b7 (see above, p. 182] = pramanasadharmyam tu sadhayisyamanam siddham kytvodahrtam Vibha 9, n. 7), this can only refer to the section which Dharmakirti devotes to the proof that the Bhagavat is pramana-like, ie, the section beginning with PV 2.7b and continuing to the end of the chapter. Thus, the interpretation arrived at by Inami and Tillemans (1986: 128) and based upon dGe 'dun grub pa's topical outlines (sa bead) of his commentary on PV 2 could equally be based upon Devendrabuddhi's understanding of the Pramanasiddhi chapter and should not be considered as "curious" at all: We are thus led to the rather curious result that in Dharmakirti's chapter devoted to Pramapasiddhi it is only k. 1-6 which speak about actual pramana; the main subject of PV II is pramana taken in the metaphorical sense of pramanabhila or Ishad mai skyes bu. If this result is accepted, and I think this is the reading of PV 2 to be preferred, one also has to assume that Dharmakirti himself considered the two defining characteristics provided by him in PV 2.lab (pramanam avisamvadi jnanam) and 5c (ajnalarthaprakaso va) as being necessary and sufficient conditions for a cognition's being a pramana. Otherwise the whole enterprise of proving that the Bhagavat due to his similarity to a pramana can be subsumed in a metaphorical sense under the concept of pramana would be ill-founded. This also implies, at least to my understanding, that both characteristics refer to a conventional (samvyavaharika) pramana, for absolutely (paramarthika) speaking the Buddha is considered to be pramana. Such an interpretation is attested to by Sakyabuddhi." In the following section I would like to present some statements of Dharmakirti either from his Pramanasiddhi chapter or scattered in his other works that can be adduced as evidence in order to substantiate the results presented by me above. For the sake of convenience let me recall Dharmakirti's definition of pramana: pramanam avisam vidi jnanam (PV 2.lab) ... ajratarthaprakso (PV 2.5c] pramana is a reliable cognition ... or the illumination of an unappre hended object. The first remarks I would like to adduce are taken from the Pramana siddhi chapter and refer to the Bhagavat: heyopadeyalattvasya subhyupayasya vedakah / yah pramanam asav isto na lu sarvasya vedakah // (PV 2.32) The one who makes known the nature of what is to be avoided and what must be appropriated as well as the means thereof), is considered as pramina, not the one who makes known everything." layah svadratamargoktih (PV 2.145a) ... tatah pramanam tayo vd caluhsatyaprakasanam / (PV 2.146ab) [The Bhagavat's) protection consists in stating the way to liberation which he has seen himself.... Because of this he is a pramana. Or, protection consists in) making known the four noble truths. dayayi dreya dcaste jnanad bhutam sasadhanam lac cabhiyogawin waktum yalas lasmal pramanata // (PV 2.282) By subsuming the Buddha under the concept of pramana on account of his similarity to pramana qua cognition endowed with these two defining characteristics, Dharmakirti is inspired by Dignaga who states that authoritative words/ words of a credible person (aplavada) can be called inference because they are similar in being reliable (avisampida): yid ches Ishig kyani mi slu bar | mishuns phyir rjes su daug pa riid || (PS 2.5ab) yid ches pa'i fshig tid brun nas kyan mi balu bar mahune pa'i phyir de yan rjes su dpag pa niid du brjod do I (PS-Vrtti (D 4204] 29020); for the Sanskrit wording of PS 2.5ab laplavidavisamuddasd manyad ann. manala), cf. below, n. 51. - It goes without saying that my earlier rendering of Ishad mar gyur pa as "ferl. der zum Erkenntnismittel geworden ist" (Krasser 1991: 19) has to be corrected. 15 For a discussion of the issue as to whether Dharmakirti provides a definition by these two characteristics, and if so, whether they are necessary and/or sufficient conditions, in connection with the problem of the meaning of the disjunctive par ticle wi in PV 2.50, see the dispute between Oetke (1998, 1999b) and France (1999); cf. also Katsura 1984, Dreyfus 1991 and Dunne 1999: 279-281 and 343f - CE. PVT D720 (quoted above, p. 179): de bas na tha wilad du byas i Ishad mas dper mdaad pa yin no zes bya ba 'di ni rigs pa yin no 11. 11 Translated in Nagatomi 1957: 40. *PV 2.145-146 is translated in Nagatomi 1957: 148, Vetter 1990: 52 and Franco 1997: 26.

Page #8
186 H. Kraker On Dharmakirtis 'nderstanding of pramanashifu 187 Due to compassion he proclaims salvation, because of knowledge she proclaims truth, and because he exerts himself to impart that" to gether with the means thereof). [the Bhagavat is pramana." In these passages it is clear that the Buddha is considered to be a pramana because he makes known what is to be avoided, etc., in other words, because he is teaching the four noble truths. This amounts to a correspondence to the second definition of a pramana qua cognition, namely ajatarthaprakasa. That is to say, the Bhagavat can be called a pramana because he possesses this defining characteristic. It remains to be noted that nowhere in the Pramanasiddhi chapter is the Buddha explicitly said to be avisamvadin. However, already in the first chapter, when explaining why Dignaga considers agama to be a pramana," he stated that this is due to the fact that authoritative words / words of a credible person are said to be anumana also with regard to imperceptible (paroksa) objects because they are similar in being reliable (aptavadavisamadasamanyad anuma. nala / PV 1.218ab) to those authoritative words whose truth can be checked by means of perception and inference. It should be men. tioned, however, that cognition based on agama, although it is grounded in words, is different from normal cognition based on words because it does not make known just the speaker's intention, but a real state of affairs. Thus it is reliable with regard to this real state of affairs. So far I have treated two instances of pramana, namely the Buddha and his teaching, which are special inasmuch as they are not cognitions. Let me now turn to the regular conventional pramanas, perception and inference. The first passage is from the beginning of the Pratyaksa chapter of his Pramanaviniscaya where Dharmakirti grosso modo states that there are two kinds of pramana for which he uses the synonym samyagilana, namely perception and inference, because only thene kinds of cognition are reliable: yan dag pa de pa de ni rnam pa gris temton sum dan nirjan su dpagi ces bys'o il ne ky abhyam arthane paricchidya pravartamd northe kriyayam visamadyale. (PVin 1.30,15-18) Correct cognition (samyagjina) is of two kinds, "perception and inference", because one who acts having ascertained the object by means of these two kinds of cognition) is not betrayed with regard to the fulfillment of Chis) purpose. minon sum yan don la mislu barid las tahad ma yini no (PVin 1.38,10) Perception, too, is a pramana because it is reliable with regard to the object. de las dros poi rart buin la misluba siid ni 'di'i yan tahad ma yin no (PVin 2.3,21-23) Therefore, the validity (Inhad ma, "pramanya) of it sie anuminaalis the reliability with regard to the nature of the real thing. abhipriyavisamnidad api bhranteh pramanala (PV3.56ab) [Inference which is an erroneous cognition is nevertheless a promana because it is reliable with regard to the intention These passages may suffice to show that both conventional promanas are assumed by Dharmakirti to be avisamvadin. With regard to the second defining characteristic, ajatarthaprakaso ni (PV 2.5e), Dharmakirti himself, in a digression in his Hetubindu, supplies us with the necessary material as to how it should be understood. The passage of PV which I want to read side by side with the digression in the Hetubindu reads: ajratarthaprakaso vd saripadhigateh param praplan samanyarijanam avijte malakxane yaj janam ily abhiprayat svalakrasavicaratah (PV 2.5e-6) * Truth or salvation or knowledge! Cf. Franco 1997: 22, n. 16. Translated in Nagatomi 1957: 256, Vetter 1900: 171 and Franeo 1997: 211 # C, PVSV 108,1-2 yal tarkidam aplaudarimuddasimanyad animali ily agamasya pramanyam anumanayam wktam lat katham (* = PS 2.5ab [= PV 1.218ab); cf. above, n. 43). For a detailed treatment of these passages, see, eg.. Yaita 1987. Tillemans 1993: 10-11 or the section "The Question of Credibility in Dunne 1999: 263-277. C. PV 2.10-2 (avisamwadanam sabde 'py abhiprayanivedanal // wakirudd ravisayo yo rtho buddhau prakdale il pramanyamelala faldasye rarthalattuaniban. dhanam ) where knowledge derived from words is said to be reliable, because it indicates the speaker's intention. The validity of words is, however, not based on reality. 4 Cr PVSV 109,10-11: talah sabdaprabhavapisali na iddaad abhiprayam ni vedayaly cely artha Visa mindad anumanam api In another passage perception is said to be a reliable non conceptual cogni tion: des gan alu ba can min parlog med de ni mion rummo (PVin 1.28ed) The Sanskrit text as available in NBho 171,13 has pramana for min sm (pralyka: yaj janam avisamudi, tai pramdam akalpakam. In his comments on 28ed Dhar. makirti again states that pratyaka is a pramana because of its reliability...al pratyakam tac crime raditrdil promasam. Both Sanskrit passages are identified in Steinkellner 1972: 203

Page #9
188 Hkrasser On Dharmakirtis inderstanding of the 189 Or pramana is illumination of an unapprehended object. Objection: According to this definition the cognition of a universal (arising sub sequent to the cognition of the particular (svarupa) would be pramana) Answer: No. because it is (our) intention that pramana is a cogni tion with regard to an unapprehended particular, for the particular is examined (here The digression in the Hetubindu (HB 2.18-4,2 = HB, 35-37) shows why the conceptual cognitions which arise immediately after the first moment of seeing a particular have to be excluded from the realm of pramana. Otherwise there would be an over-extension (ativyapti) of the definition. Dharmakirti states (HB 2,18-19): latra yad adyam asidharanavisayam darsanam, tad era pramdam There [ie, of the first moment of seeing and the subsequent affirmative and negating conceptual cognitions only the first moment of seeing which relates to the particular (asadharana) is a pramana Here it is worthwhile to point out that, according to Arcata, with the phrase adyam asidharanavipayam darsanam Dharmakirti is referring to Kumarila's alocanajranam prathamam; if alocanajnanam is understood as asadharanavisayam darsanam, the over-extension mentioned above could be avoided. In this way, Dharmakirti goes on, the conceptual cognitions subsequent to the first moment of seeing the particular are excluded from being a pramana, because they do not cognize a new object (apurvarthadhigamabhaval), as the real entity that is capable of fulfilling a purpose (arthakriya) has already been cognized through this first moment of seeing. Moreover, contrary to inference, the subsequent conceptual cognitions do not cognize an entity that has not already been seen and which fulfills such a purpose. As Arcata explicates, inference, although its direct object is the universal (saminya), is never. theless indirectly connected with the real entity and thus a pramana. In the process of discussion Dharmakirti rephrases the definition of PV 2.50 and explains how the over-extension can be avoided (HB 3.8-9).** tasmad anadhigalarthavisayam pramanam ily apy anadhigale salakana ili widepaniyam. Therefore, the definition] "pramana is a cognition which relates to an unapprehended object", too, is to be qualified by the words with regard to an unapprehended particular" If one keeps in mind that the aim of this digression in HB is to ensure that only the first moment of seeing (adyam darsanam) is pramana and to exclude the conceptual cognitions subsequent to this moment from the realm of pramana, it becomes clear that Dharmakirti here raises the very same problem which is addressed in the purvapaksa in PV 2.5d-6a (svar padhigateh param praptam samanyavijnanam; cf. above, p. 1871.). The solution of the problem provided by Dharmakirti in HB, namely to specify the definition (anadhigate svalaksana ili visesaniyam), corre sponds to that of PV 2.6be (avijnate svalaksane / ya jnanam ity abhi prayat). Thus it is obvious that the qualification of a pramana in the HB-passage (anadhigatarthavisayam praminam) corresponds to that of PV 2.5c (jflatarthoprakdso vu). If this correspondence between the digression in the HB and PV 2.5-6 is accepted, and I cannot see any reason why this should not be done, it can be concluded safely that this second qualification ajnatarthapra kasa, when understood as qualified by avijfate svalakpane, is a necessary II, as suggested by Kimura (1997: 252), ajfiatarthaprakabo is not under stood as a defining characteristic of pramana, but of kistra in PV 2.5b (astram mahanivartanam /) and artha in ajralartha' as referring to the meaning explained in the Pramanasiddhi chapter starting from k. 8, then k. 60-6 cannot be understood at all The central part of this digression (HB 2,18-3,16) is also translated in Dunne 1999: 472-475. CE HBT 25.11.12: tatra tesu darsananidhipratifedhavkalper * CT. HBT 25.12-17: tad adyam yad etal-asti hy alocandjanam prathamam SV Pratyaksa 112ab) ily adau vikalpapravytter bhavam iti adyam akhyalam andhara. narimayam salaksanavisayam darsanam, tad era pramanam, na vidhipratisedhani kalau api, tasyair prom nalak onayogid ilarayoi ca tadasambhaval WHB 2.24-3.1: ... apurarthadhigamabhavad arthakriyasadhanasya daranal. adrasya punas tatsadhananabhavanya vikalpenapratipaltes can manaval "C HBT 27,28-28,5: arumanavad ili waidharmyadrolantah, yatha pratyakse. narthakriyaaddhane pradesakhye dharminy adhigale 'py anadhigatasyagner artha kriyaadhanasydaamanyabarena parokrasya salaksanakarena pralipatum asakya tudi pratipattih, naivam vidhivikalpena simanya karena nadhigalam arthakriyasadha nam adhigam yale, tasyilocanajfianenaividhigamal; cf. further HBT 28.28-29.4: mar wameua tu samanyam wa karicid arthakriyam pakalpayali yallu samanyaman manavikalpagrahyam, tal baranavyapakasambaddhalinganiscayadedrayalam um baddhasambandhad anadhigalarthakriyasadhanarimayim arthakriyim upakalpaya tili tadvigayo vikalpah pramanam. PV 2.6d (salaksanavicaralak) is explicated in HB 3,13-14: arthakriyayogya vipayatolladarthind proutten, arthakriyayogyalaksanam hitu.... This is also how Steinkellner understood it; cf. HB, 36: "Deshalb mullman Auch die Bestimmung] Massgebliche Erkenntnis ist, was eine nichterkannte Sache zum Objekt hat, durch den Zusatz'sofern ein Individuelles nicht erkannt ist differenzieren". In his n. 36, Steinkellner refers to PV 2.50.

Page #10
190 HA On Dharmakirti's 'ndlerstanding of prowwabibila 191 condition for a conventional cognition's being a pramana. From this it follows that Prajnakaragupta's interpretation of artha in ajnalarthapra kaso va, as referring to absolute reality consisting in the cognition's nonduality, and of PV 2. lab and 5c as presenting conventional and absolute definitions respectively," is not in accordance with Dharmakirti's own intention. So far we have seen that the Bhagavat is proven to be a pramana in terms of the second definition, while aplavada, pratyaksa and anumana are shown to be pramanas in terms of the first definition. Moreover, from among conventional cognitions only pratyaksa and anumana refer to an unapprehended (avijriata) particular (svalaksana). Thus it is clear that both pramanas, perception and inference, are considered by Dharmakirti as being both avisamvadin and arialarthaprakasa. This is also how Devendrabuddhi and Jinendrabuddhi understand it when they explicate that the Bhagavat is said to be a pramana because he, like pramanas qua cognition, is reliable and because he makes known an unapprehended state of affairs. From this it can be concluded that the two qualifications, although not conceptually identical, are equivalent in the sense that exactly the same ranges of objects are subsumable under them, that they constitute both necessary and sufficient condi tions for a cognition's being a pramana and that it is sufficient to prove and formulate only one of them in a specific context. That Dharmakirti considers it sufficient to prove one of them can be seen from the passages referred to above (p. 185-187) where either the one or the other is applied." This also implies that the particle vi in PV 2.5e which connects the two defining characteristics can only be understood as carrying a disjunctive meaning ("or"), not a conjunctive meaning ("and"), be cause otherwise, if pramana were defined as both X and Y. it would by no means be sufficient to prove only X or only Y for a specific claimed That anadhigalarthavisaya qualifies & conventional cognition is also con firmed by another statement of Dharmakirti's with regard to mental cognition (manovijnana): 'di tahad ma na ma yul 'dzin pa yan ma yin lelahad ma nima rloge pa yul can yin pa riid kyi phyer te gaan du na ha canthal ba'i phyir ro | (PVin 1.60,17-19) "This does not grasp the object of the previous pramona either because a pramina relates to something unapprehended; if not, this would lead to absurd consequences". CY PVA. 79,15-19: atha varthaabdendira paramitha woyale ajridlarthapra kada ili paramarthaprakasa ily arthah. paramarthas cadwaitarald alprakdanam era pramanam, latha ca pratyapadi - sarupaya salo galir ili, wktam ca-prama. nyam pyara hareneli tatra piramarthikapramanalaksanam elal, puram tu imoyavahari kasya. Thus, of the three possibilities of combining PV 2.lab and 5c 88 proposed in Oetke 1999a: 250, the second one can be excluded: "... Thus the two specifications might not be extensionally equivalent, and Dharmakirti's definitional enterprise could be paraphrased by expressions of the form: 'Pramdna, in so far as it relates to empirical reality, is equivalent to F. and pramana, in so far as it relates (also) to absolute reality, is equivalent to G" * C above, p. 182: ji liar... mislu ba'i phyir dari muides poi dongsal bar byed pai phyir Lahad ma rid yin no || de llar na bcom Idan das kyan ... mislu bar mdzad pa dari...don..mi des pa ston par mdaad a'i phyir tshaid ma nid yin no II. and p. 175: yatha hi pratyaksidipramanam ... anadhigalasyarthasya prakadakam avisontvadakam ca, latha bhagawin api... lasyanadhigalanya prakdako visamvadakasca This corresponds to Oetke's first proposal (1999a: 250): "First, even if the specifications represented by avisamvadi and ajnalathaprakso were equivalent in the sense that exactly the same ranges of objects are subsumable under them, they are not conceptually identical and their equivalence is not trivial or self evident In view of the material presented so far, the claim that Dharmakirti does not provide a general definition in the initial verses of PV 2 as formulated in Franco 1997: 45f., in his second chapter "The Initial Verses: On the Absence of a Definition of Means of Knowledge in Dharmakirti's Writings", has to be reconsidered * I do not agree with van Bijlert's interpretation (1989: 151) that PV. "complements the definition of a pramana given in PV II lab" and "only refers to direct perception, for only through direct perception it is possible to directly cognize new objects, previously unknown particulars". If the qualification in or did not constitute a sufficient condition but only complemented the first one, the disjunction ud would be difficult to account for. Besides, on account of his corre spondence to the specification in 5c, the Bhagavat is subsumed under the concept of pramana which - as in the case of a platada - can only mean that he represents an instance of anumana. Moreover, if anumana did not relate to and indicate a pre viously unknown particular (cf. PV 3.81-82), even though indirectly, it would he excluded from the realm of pramana, like the conceptual cognitions subsequent to the first moment of seeing. # The fact that in his Pramanaviniscaya Dharmakirti distinguished valid from non valid cognitions by their reliability only, without referring to ajalorthaira kaia, led Go rams pa to the conclusion that PV 2.lab (pramanam arinardi irlanam is a complete definition. This is in accordance with Sakya Pandita who considers the two defining characteristics as having the same intention (dgn pa gcig). For references and also for the differing views of the die lugs pas of Kimura 1997: 257-256 (also Kimura 1995: 176-175). 1 Based on such statements of Devendrabuddhi as the one treated above (el p. 182: ji lar ... mislu ba'i phyir dari mi se pai don gaal har byrd pe phyir ishad ma nid yin no de llar na bcom ldan 'dar kyan...mislu bar mdzad pa dan ...don. mi des paslon par mdaad pay phyir (shad ma nid yin wo) Dunne (1999: 3438. n. 185) strongly argues against a disjunctive and for a conjunctive meaning of vi All the passagen he adduces for his argument. however, smoothly can be real assuming a disjunctive meaning of ex.

Page #11
192 H. Kranser On Dharmakirtis Understanding of pramanabhala 193 instance of pramana. This is sufficient only under the condition that X and Y are different concepts that have the same impact with regard to a cognition's being a pramana. That is to say, when a cognition is pramana it not only is reliable but reveals an unknown object. However, to define pramana as being X and Y would not be appropriate if Y is thus already implied in X and vice versa. This leads us to the next question, namely as to why Dharmakirti provides us with these two defining qualifications at all. Would it not suffice to give only one definition? An answer to this question can be found in Arcata's commentary on the digression in the HB I have discussed above (p. 188-189), in his explanation of HB 3,8-9: tamad anadhigalarthavi ayam pramanam ily apy anadhigale salakyana ili visesaniyam. Therefore, the definition] "pramana is a cognition which relates to an unapprehended object", too, is to be qualified by the words) "with regard to an unapprehended particular". Arcata comments on this passage as follows (HBT 33,8-21): tad esam nilam drafva nilam id jane pratibhasamanam samanyam na karicid arthakriyam upakalpayatiti prasidhyanarthakriyakarivisaya - sypi vikalpasya pratyaksaprathabhavinah pramanyaprasangid alivyapbir ili latrapurvartha vijanam ili pramanalaksane mimamuakair videga nam upadeyam ili darsayann aha -lasmal. gala evam anarthakriyded dhanaviqayalaya dardanaprothabhavino vikalpasya pramanyam ayuktam, tasmad a.madabhimatam pramanam avisamuddi jrianam (PV 2.lab) ili pramanalaksanam vyudasya, anadhigatarthariqayap pramanam, talrd. purvarthavifanam pramanam ily api, elasminn apy ahopuruskayanyas min pramanalaksane kriyamane 'livyaptipariharaya videpaniyar vienanam upadeyam. katham visepaniyam anadhigale salakpara ili anena hi videranenanumanavikalpasya ca pramanyam sidhyati, alocanajfiana prsthabhavinas ca vikalparya pramanyam vyudasyala ili saruam mustham. Having thus demonstrated that a universal which appears after one has seen blue in a conceptual) cognition in the form "this is blue" does not fulfill any purpose, [Dharmakirti says "therefore" in order to show that the Mimamsakas must employ a qualification in their definition of pramana (namely there [i.e of them (pramana) is a cognition with regard to a new object", as iti) otherwise) there would be an overextension of the definition), because there obtains the (undesired) con sequence that a conceptual cognition that is subsequent (prethabhavin) to perception, although it has an object which does not fulfill any pur pose, would be a pramudna. Because in this manner, inasmuch as it has an object which does not fulfill any purpose, it is not appropriate that a conceptual cognition subsequent to the first moment of seeing is a pramana, therefore(Dharmakirtil, putting aside (uyudasya) the defini tion of pramana maintained by us, [i.e. "pramana is a reliable cogni tion". (said that the definition) "pramana is (a cognition) that relates to an unapprehended object" (which corresponds to "there i.e. of them pramana is a cognition with regard to a new object", too, is to be qual ified, li.e. also in this other definition of pramana when being made due to great self-confidence, a qualification is to be employed in order to exelude this over-extension. Why is it to be qualified by the words "with regard to an unapprehended particular"? Because by this qualification on the one hand the validity of the conceptual cognition consist ing in inference is established, on the other hand the validity of the conceptual cognition that is subsequent to the perceptive awareness is excluded. Thus, everything stands well. The major points of Arcata's position that can be derived from this explanation are: The Buddhist definition of pramana is avisamvadi janam. The definition of the opponent addressed by Dharmakirti with ana dhigatarthavisayam pramanam is that of the Mimamsakas formu. lated as apurvarthavijnanam pramanam. > This definition is correct when it is understood in a modified way. If we now, basing ourselves on this information provided by Arcata, combine the two pertinent passages lasmad anadhigatarthavisayam pramanam ily apy anadhigate svalaksana iti visesaniyam. (HB 3,8-9) and ajrialarthaprakaso v svarupadhigaleh param praptam samanyavijnanam avijfate salakxane yaj jianamity abhiprayat malaksanavicaralah II (PV 2.5c-6) I feel justified to suggest that already Dharmakirti's second definition of pramana as ajnatarthaprakasa in PV 2.5e corresponds to that of the Mimamsakas, namely apurarthavijnanam pramanam, in a modified form because Dharmakirti explicitly understands his definition here as including the qualification (avijnale svalaksane ... abhiprayat) which, according to his statement in the later HB, must be added in the defini tion of the opponent. If the scenario thus derived from Dharmakirti's The aim of the digression in the HB therefore is not, as suggested by Stein kellner (cf. above, n. 63), to show that Dharmakirti's own definition of PV 2.50 should be specified, but to show that the definition of the Mimamsakas is faulty

Page #12
194 H. Krasser On Dharmakirti l'nverstanding of priminalhin own statements and presupposed by Arcata is correct, then the oppo. nent whose definition is at stake here would be, as in many other cases in Dharmakirti's work, Kumarila. The formulation of the definition which Arcata equates with Dharmakirti's in the HB is, as we can gather from a quote by Ratnakirti, most probably taken from Kumarila's Brhattika (BT): tatha brhalfikapi latripurvarthavijnanam niscitam badhavarjitam adustakaranarabdham pramanan lokasammalam (R 113,10-12) There i.e. of them) (only) a decisive cognition with regard to A new object, that is free of sublation (and) that is brought about by faultless causes, is assumed in the world to be a pramana." Now, if PV 2.5-6 were related to this definition in the BT or in some other Mimamsa work of the time, we would expect to find some echo of this in the commentaries on the pertinent passage in the PV. Devendrabuddhi and Sakyabuddhi are silent on this point. Prajnakaragupta quotes the BT-verse, however not in connection with PV 2.5-6, but in his introduction to PV 2.3 (cf. PVA, 53,4-5). The reason for this may be that he favors the interpretation of ajratarthaprakasa as representing a definition of pramana in the absolute sense (paramarthikapramanalakaana; cf. above, p. 190 with n. 65), an interpretation which is obviously not supported by the mention of lokasamata in the BT. However, a possible relationship between the two definitions is corroborated by Ravigupta's commentary which quotes the definition of the BT and states that PV 2.6d is directed against it." Before drawing conclusions I would like to consider whether Arcata's remark that avisarvadi jfianam is the Buddhist definition of pramana can be substantiated. If we look at the passages of Dharmakirti's work referred to above (p. 185-187) it is obvious that only the Bhagavat is said to be a promana on account of his revealing the four noble truths, etc., which amounts to a correspondence to ajralarthaprakasa, while aplaudda, pratyaksa and anumana are pramanas on account of their being reliable (avisamvadin). I have not yet been able to locate a passage in his works where the validity of perception or inference is derived in terms of ajratarthaprakaka. Moreover, Dignaga, too, considers a plavida to be a pramana in terms of its reliability (cf. above, n. 51). Therefore we do not have any reason to doubt Arcata's attribution. Thus, coming back to the question as to why Dharmakirti provides us with a second alternative definition, the only possible answer I can se is that he adopted the modified definition of the BT in order to prove to the Mimamsakas that, even according to their own definition when understood properly, not the Veda, but the Buddha is to be regarded as a pramana." and has to be modified. Dunne, too, does not relate this digression to the Mimam. sakas, for he introduces his translation of HB 2.13-3.16 with the remark: "... This leads Dharmakirti to more general considerations concerning the characteristics of a pramana" (1999: 472). # Without mention of its source the verse is also quoted in PVA, 53,47., TBV 13.24f., 318,25f., 394.161.: TR 126.21; cf. Mimaki 1976: 88f, and 284f. " If Dharmakirti in PV 2,5-6 is referring to this passage from the Brhatika and if Frauwallner's assumption that the Brhattika is a remake of the Slokavart. tika under the influence of Dharmakirti's first work which Frauwallner calls Hetuprakarana is correct (1962: 332-335), then the Brhatika must have been accomplished between the composition of the Hetuprakarana and PV 2. As Frauwallner (1954: 148) assumes that PV 2-4 have been written directly after the Hetuprakar. ana which then had been incorporated as chapter into the Pramanavarttika, we are thus led to the very unlikely scenario that Dharmakirti wrote his Hetupra karana which he gave to Kumarila who rewrote his Slokavirttika and presented the results to Dharmakirti who then started to compose his second work PV 24. Thus, one of the above assumptions cannot be correct and has to be reconsidered This question, however, is not our concern here. I would further like to point out that although the idea that promia should be related to a new object can be found in the Slokavirttika in such statements as sorosan palabdhe the pramanyam mpiranyatha (SV Autpattika Ilab, referred to in Franco 1997: 62, n. 38), it is not reflected in the definition of pramana in the Slokavarttika (lasmad drdham yod wipannam napi samvadam vili na visamvadam rechali jrlanantarena vijnanam lal pramanam pratiyalam SV Codana 80). NPVVK 304b1f.: delta na mishan vid ni de la shon med don des palmes legnad papan dan | skyon med pa las skyen pani || Ishad mar jig rien pa Wod do sex ser la de dir blog par byabo'phyir rangi mlahan Rid dpyad phyir rol (PV 2.6) zes by ba la yod pa riid dan med pa lid kyi gial bya mthon bar ses pa silahad mer brjod doll Thus, Oetke's third proposed possibility (1999a: 250) does not differ greatly from my results: "Thirdly, one could suppose that both alternative specifications are equivalent in Dharmakirti's eyes, but that nevertheless the fact of their equiva lence depends on certain theoretical assumptions, in particular epistemological theorems. More specifically, the first definition could be considered as tailored to Dharmakirti's own theoretical assumptions where the second alternative repre sents a more general explication".

Page #13
196 BT D Dreyfus 1991 DTUTP Dunne 1999 Franco 1997 Franco 1999 Frauwallner 1954 Abbreviations and Bibliography Frauwallner 1960 Frauwallner 1962 Hakamaya 2000 .... Hakamaya 2001 HB HB,, HBT H. Krasser Brhattika (Kumarila). De dge Tibetan Tripitaka bsTan hgyur Preserved at the Faculty of Letters, University of Tokyo. Tahad ma 1-21, ed. J. Takasaki - Z. Yamaguchi Y. Ejima. Tokyo 1981ff. Georges Dreyfus, Dharmakirti's Definition of pramana and its Interpreters. In: SBET 19-38. Shoryu Katsura (ed.). Dharmakirti's Thought and Its Im pact on Indian and Tibetan Philosophy. Proceedings of the Third International Dharmakirti Conference, Hiroshima, November 4-6, 1997. Wien 1999. John Dowling Dunne, Foundations of Dharmakirti's Philo sophy. A Study of the Central Issues in his Ontology, Logic and Epistemology with Particular Attention to the Svopajnavrtti. Unpublished thesis presented to the Harvard University. Cambridge, Massachusetts 1999. Eli Franco, Dharmakirti on Compassion and Rebirth. Wien 1997. Id., Response to Claus Oetke's paper. In: DTIITP 253-259. Erich Frauwallner. Die Reihenfolge und Entstehung der Werke Dharmakirti's. In: Asiatica, Festschrift Friedrich Weller. Leipzig 1954, p. 142-154 (= Kleine Schriften, p. 677-689). Id., Devendrabuddhi. WZKSO 4 (1960) 119-123 (= Kleine Schriften, p. 842-846). Id., Kumarila's Brhattika. WZKSO 6 (1962) 78-90 (= Kleine Schriften, p. 323-335). Noriaki Hakamaya, pramana-bhata to kumara-bhuta no gogi - bhuta no yoho wo chushin toshite [The Meaning of the Words pramana-bhula and kumara-bhuta - Focussed on the usage of bhata). Komazawa tanki daigaku bukkyo ronshi 6 (2000) 328-299. Id., Pramana-bhuta hoki (A Supplementary Note on pramana bhula). Komazawa tanki daigaku kenkyu kiyo 29 (2001) 433-448. Hetubindu (Dharmakirti): Ernst Steinkellner, Dharmakir ti's Hetubinduh Teil I: Tibetischer Text und rekonstruierter Sanskrit-Text. Wien 1967. Ernst Steinkellner, Dharmakirti's Hetubinduh. Teil II: Ubersetzung und Anmerkungen. Wien 1967. Hetubindutika (Arcata): Sukhlalji Sanghavi and Jinavijayaji (eds.), Hetubindulika of Bhatta Arcata with the subcommentary entitled Aloka of Durveka Misra. Baroda 1949. Inami 1994 Inami Tillemans. 1986 Jackson 1988 Jackson 1993 Katsura 1984 Kimura 1995 Kimura 1997 Krasser 1991 Mejor 1991 Mimaki 1976 Nagatomi 1957 NBhu On Dharmakirti's Understanding of pramanabhuta Oetke 1999a Oetke 1999b PS 1 197 Masahiro Inami, Pramanavarttika Pramanasiddhi sho no kenkyu (3) [A Study of the Pramanasiddhi Chapter of Pramanavarttika (3)]. Shimane kenritsu kukusai tanki daigaku kiyo (1994) 17-53. Masahiro Inami Tom Tillemans, Another. Look at the Framework of the Pramanasiddhi Chapter of Pramana varttika. WZK8 30 (1986) 123-142. Roger R. Jackson, The Buddha as pramanabhata: Epithets and Arguments in the Buddhist Logical Tradition. JIP'h 16 (1988) 335-365. Id., Is Enlightenment possible? Dharmakirti and royal tshab rje on Knowledge, Rebirth, No-Self and Liberation. Ithaca. 1993. Shoryu Katsura, Dharmakirti's Theory of Truth. J/Ph 12 (1984) 215-235. Seiji Kimura, pramana teigi ni tsuite [On the Definition of pramana]. Komazawa tanki daigaku bukky ronshu I (1995) 180-169. Id., Teigi to pramana teigi ni tsuite [On Definition and the Definition of pramana]. Komazawa tanki daigaku bukky ronshu 3 (1997) 260-244. Helmut Krasser, Dharmottaras kurze Untersuchung der Gultigkeit einer Erkenntnis, Laghupramanya pariksa. Teil II - Ubersetzung. Wien 1991. Marek Mejor, On the Date of the Tibetan Translations of the Pramanasamuccaya and the Pramanavarttika. In: SBET 175-197. Katsumi Mimaki, La refutation bouddhique de la permanence des choses (Sthirasiddhidusana) et la preuve de la momentaneite des choses (Ksanabhangasiddhi). Paris 1976. Masatoshi Nagatomi, A Study of Dharmakirti's Pramanavarttika. An English Translation and Annotation of the Pramanavarttika. Book I. Unpublished thesis presented to the Harvard University. Cambridge, Massachusetts 1957. Nyayabhusana (Bhasarvajna): Svami Yogindrananda (ed.). Srimadacarya Bhasarvajna-pranitasya Nyayasarasya sepa jnam vyakhyanam Nyayabhusanam. Varanasi 1968. Claus Oetke, The Disjunction in the Pramanasiddhi. In: DTITP 243-251. Id., Clarifications. In: DTIITP 261-266. Pramanasamuccaya. Chapter 1 (Dignaga): Masaaki Hattori, Dignaga, On Perception, being the Pratyaksapariccheda of Dignaga's Pramanasamuccaya. From the Sanskrit Frag

Page #14
198 H. Krasser On Dharinakirti's Inderstanding of pramanabhula 199 PST. Seyfort Ruegg 1995 PV1 PV 2-4 Steinkellner 1972 Steinkellner 1980 PVA, Steinkellner 1981 PVAcib PVAT Steinkellner 1989 ments and the Tibetan Versions Translated and Annotated. Cambridge, Massachusetts 1968. Pramanasamucca yatika (Jinendrabuddhi): D Tshad ma Vol. 20, No. 4268, Ye 1b1-314a7 = Q Vol. 139. No. 5766, Re 1-354b. Pramanavarttika, Chapter 1 (Dharmakirti) in PVSV. Pramanavarttika, Chapters 2-4 (Dharmakirti): Yusho Miyasaka (ed.), Pramanavarttika-Karika (Sanskrit and Tibetan). Acta Indologica 2 (1971/72) 1-206 (PV 2 = Pramanasiddhi; PV 3 = Pratyaksa; PV 4 = Pararthanumana). Pramanavarttikalankara (Prajnakaragupta): Motoi Ono (ed.), Prajnakaraguptas Erklarung der Definition gultiger Erkenntnis (Pramanavarttikalamkara zu Pramanavarttika II 1-7). Teil 1: Sanskrit-Text und Materialien. Wien 2000. Pramanavarttikalankara (Tibetan text): D Tshad ma Vol. 5, No. 4221, Te 1b1-308a7. Pramanavarttikalankaratika (Jayanta): D Tshad ma Vol. 7, No. 4222, De 1b1-365a7. Pramanaviniscaya, Chapter 1 (Dharmakirti): Tilmann Vetter, Dharmakirti's Pramanaviniscayah, 1. Kapitel: Pratyaksam. Einleitung, Text der tibetischen Ubersetzung, Sanskritfragmente, deutsche Ubersetzung. Wien 1966. Pramanavarttikapanjika (Devendrabuddhi): D Tshad ma Vol. 2, No. 4217, Che lb-326b = Q Vol. 130, No. 5717(b), Che 1-390a. Pramanavarttikasvavrtti (Dharmakirti): Raniero Gnoli (ed.), The Pramanavaritikam of Dharmakirti. The First Chapter with the Autocommentary. Roma 1960. Pramanavarttikatika (Sakyabuddhi): D Tshad ma Vol. 4, No. 4220, Ne 1-282a = Q Vol. 131, No. 5718, Ne 1-348a. Pramanavarttikavrtti (Manorathanandin): Rahula San krtyayana (ed.), Dharmakirti's Pramanavartlika with a Commentary by Manorathanandin. Patna 1938-1940. Pramanavarttikavrtti (Ravigupta): D Tshad ma Vol. 9, No. 4224, Pe 293b1-398a7. PVin 1 SV of the rsi, acarya and tathagata in Grammatical, Epistemo). logical and Madhyamaka Texts. BSOAS 57 (1994) 303-320. Id.. Validity and Authority or Cognitive Rightness and Pragmatic Efficacy? On the Concepts of pramana, pramanabhula and pramana(bhuta)purusa. Asiatische Studien 49 (1995) 817-827. Ernst Steinkellner, New Sanskrit-fragments of Pramana. viniscayah, First Chapter. WZKS 16 (1972) 199-206. Id., Some Sanskrit-fragments of Jinendrabuddhi's Visala. malavati. In: A.L. Basham et al. (eds.), A Corpus of Indian Studies. Essays in Honour of Professor Gaurinath Sastri. Calcutta 1980, p. 96-105. Id., Philological Remarks on Sakyamati's Pramanavarttikatika. In: K. Bruhn - A. Wezler (eds.), Studien zum Jai. nismus und Buddhismus. Gedenkschrift fur Ludwig Alsdorf. Wiesbaden 1981, p. 283-295. Id., Miszellen zur erkenntnistheoretisch-logischen Schule des Buddhismus: VI. A New Approach towards Improving the Textual Basis for an Understanding of Dignaga's Pramanasamucca yavrtti. VII. Another Note on pramanabhula. WZKS 33 (1989) 177-181. Slokavarttika (Kumarila): Dvarikadasa Sastri (ed.), Slokavarttika of Sri Kumarila Bhatta with the Commentary Vyayaratnakara of Sri Parlhasarathi Misra. Varanasi 1978. Tattvabodhavidhayini (Abhayadevasuri): S. Sanghavi and B. Dosi (eds.), acarya-sri-Siddhasena-Divakarapranitam Sam. matilarkaprakaranam ... srimad-Abhayadevasuri-nirmitaya Tattvabodhavidhayinya vyakhyaya vibhusitam. 5 Vols. Ahmedabad 1924-1931. Tom J.F. Tillemans, Persons of Authority. The ston pa tshad ma'i skyes bur sgrub pa'i gtam of A lag sha Ngag dbang bstan dar. A Tibetan Work on the Central Religious Questions in Buddhist Epistemology. Stuttgart 1993. Tarkarahasya, ed. Acharya Paramanandan Shastri. Patna 1979. Vittorio A. van Bijlert, Epistemology and Spiritual Author. ity. Wien 1989. Tilmann Vetter, Der Buddha und seine Lehre in Dharmakirtis Pramanavarttika. Wien ?1990. Vibhuticandra's notes in PVV. Hideomi Yaita, Dharmakirti on the Authority of Scriptures (agama). Nanlo Bukkyo 58 (1987) 1-17. PVP PVSV TBV PVT PVV Tillemans 1993 PVVR TR The Tibetan Tripitaka, Peking Edition, ed. D.T. Suzuki. Tokyo - Kyoto 1955-1961. van Bijlert 1989 Vetter 1990 Ratnakirti-Nibandhavalih (Buddhist Nyaya Works of Ratnakirti), ed. Anantalal Thakur. Patna 21975. Vibhu SBET Yaita 1987 Ernst Steinkellner (ed.), Studies in the Buddhist Epistemological Tradition. Proceedings of the Second International Dharmakirti Conference Vienna, June 11-16, 1989, Wien 1991. D. Seyfort Ruegg, pramanabhuta. *pramana(bhuta)-purusa, pratyaksadharman and saksatkrtadharman as Epithets Seyfort Ruegg 1994