Logic Of Svabhavahetu In Dhharmakirtis Vadanyaya

Author(s): Ernat Steinkellner
Publisher: Ernat Steinkellner
Catalog link: https://jainqq.org/explore/269411/1

Page #1
THE LOGIC OF THE SVABHAVAHETU IN DHARMAKIRTI'S VADANYAYA by Ernst Steinkellner, Vienna One of the signs of an extraordinarily creative person is continuous work and, what is more important, continuity in this work. Philosophy is - among other things - work. And work results in change: changing problems and changing answers. History and scholarly culture have a tendency to petrify, to create solid forms, structures and systems that can be taken home, stored, studied and transmitted. Difficulties, imbalances, unsolved problems, developments and differences - in short: the qualities of real life in an author's work are usually neglected and, unnoticed, soon disappear from the transmitting consciousness of the heirs to any given tradition. Dharmakirti's fascinating mind has suffered this fate. His elaborate efforts established Dignaga's tradition of epistemology and logic in a form that has lasted-within the Tibetan Buddhist culture - up to our times. However, its intrinsic fascination was soon lost in the reception of a highly interesting philosophical system which leaves a great deal of room for later debate. To understand Dharmakirti's significance for the subsequent tradition he must be seen inde. pendently of it, must be read in his own words, and must be studied within his own context of questions, efforts and solutions. What Dharmakirti himself achieved is one thing that must be investigated; what his pupils and the even later branches of his tradition contributed is another; what they thought he had achieved is yet another field for investigation. As soon as we start reading Dharmakirti on his own terms we find ourselves participating in his philosophical workshop. And the philological situation in his case is luckily such that we can literally observe him at work, taking up a theme again and again, adapting it, fitting it together with other themes he has taken up again, and welding them together so that they secm never to have been separate. Professor Frauwallner concluded his paper of 1954 on the origin and sequence of Dharmakirti's works with a statement that points the way to one of the lines of future research on Dharmakirti: "It will be a fascinating task to trace the origin and gradual development of his thought in detail."! Nothing much has happened until now, more than 30 years later, but there is an increasing number of scholars today who have demonstrated in their research that these words were not written in vain. At the First International Dharmakirti Conference in Kyoto 1982 I offered a paper with a rough working hypothesis concerning a development in Dharmakirti's works with regard to a theorem that I consider as central to his logical thought: the theorem of the ascertainment (niscaya) of the logical Dexus (vyapti) in the case of an essential property as logical reason (svabhavahetu). I presented an idea of this theorem's development in Dharmakirti's major works, starting from the first chapter of the Pramanavarttika together with its so called "autocommentary" through the Pramanaviniscaya and Nyayabindu to its final form in the Hetubindu and the Vadanyaya. I also said that this development and the reasons for it can be discerned in connection with the development of corollary theorems, and E. Frauwallner, "Die Reihenfolge und Entstehung der Werke Dharmakirtis", in Asiatica, Festschrift F. Weller, Leipzig 1954 (142-154): 154. 2 Cf. c.g. the papers of T. Tani on the development of Dharmakirti's prasanga-concept, of T. Iwata on his work on the svabhava- and karyahetu-concepts, and of M. Inami on the treatment of paksabhasa in this volume.

Page #2
312 E. Steinkellner that there is a veritable "causal complex" of theories and theorems which finally merge into a complete and consistent logical theory. This I would like to refer to as the "final form of Dharmakirti's logic". I regret to have to admit today that I have not yet been able after all these years to fulfil my promise and present a study of this development. Under these circumstances I would like to offer you today not another working hypothesis, but what I would like to refer to as an experiment in interpretation. Irrespective of my working hypothesis that the "final form of Dharmakirti's logic" took shape in the formulation of his late works, the Hetubindu and the Vadanyaya, an investigation of the logic in these works as such, without reference to its connections with earlier forms of the same theory is a necessary task in itself. If I call it an experiment I mean that in dealing with Dharmakirti's final formulations it remains to be seen whether they contain a logical theorem which obviates the need for information from his earlier works and allows us actually to consider it as the nucleus of a coherent logical theory. For, if this hypothetical presumption of conceptual coherency is not accepted with regard to the limited corpus of his late work, the specific works under consideration will have to be seen as being in danger of losing their literary and systematical consistency. Accordingly I shall base my explanations entirely on the formulations of the Hetubindu, where important terms are used for the first time, and above all of the Vadanyaya, where their meaning and methodical application is explained to some extent. I shall refer to his other works only in order to clarify his conceptual usage where there is no reasonable doubt of any change in this. In order to demonstrate the generally binding force of the main statements in the Vadanyaya regarding the proof of the logical ncxus (vyaptisadhana) in the case of an essential property as logical reason (svabhavahetu) that I would like to investigate, a glance at the context of these statements will be sufficiently revealing. The subject of Dharmakirti's last major work, the Vadanyaya," is the definition of the so-called "points of defeat" (nigrahasthanani) in disputations (vada) and the refutation of other definitions in fact those proposed by the early Nyaya school). Within the frame of definition of the "points of defeat we find in the Vadanyaya a complete formulation of Dharmakirti's logic, i.e. his theory of the logical reasons (hetu), when for the major part of the explanation of the term defining the points of defeat for the propounder", Dharmakirti considers a "point of defeat" to be any deficient usage Textual references are given to the editio princeps of Rahula Sanktyayana (Patna 1935-36). Swami Dwarikadas Shastri's edition (Varanasi 1972) is no improvement on the first edition mainly for methodological reasons. A new and critical edition of the first part of the Vadanyaya together with a German translation was submitted by Michael Torsten Much as a PhD dissertation at the University of Vienna in 1983. Meanwhile the second part of the text has been prepared and the complete work will be published by 1991. For a survey of the system of the kinds of "points of defeat' and their definition as implied in the fastrafartraka verse of the Vadanydya cf. Michael T. Much, "Dharmakdrti's Definition of Points of Defeat' (nigrahasthana), in Buddhist Logic and Epistemology, ed. Bimal K. Matilal and Robert D. Evans, Dordrecht etc. 1986, 133-142 (with a synopsis of the definitions on p. 138). 5 VN 3,1-60,3. Dharmakdrti expounds the definitory term asadhanangavacanam ("non-means-of-proof-formulation") which entails the definitions of the possible "points of defeat" on the side of the propounder (vadin). When the negation is constructed with the second member of the compound (sadhanangasya avacanam VN 3,11.), i.e. the non-formulation of a means of proof). it is further taken to mean "the non-formulation of such mcans (anga) of inferential cognition (sadhana = siddhi) as are available in form of the only threefold indicator (trividham eva lingam)" (VN 3.3f.). He finally says that "non-formulation" (avacanam), i.e. "non-presentation" (anuccarana) of this indicator, can have two reasons: "silence" (simbhava) or "nonjustification" (asamarthana) of this means (VN 3,2f.).

Page #3
The Logic of the svabhavahetu in Dharmakirti's Vadanyaya 313 of logical reasons or indicators, and that such deficient usage consists in non-justified usage. Here "non-justification" (asamarthana) is nothing but the fact that the propounder does not indicate a definite, ascertaining cognition (*niscaya) with regard to all the three forms or aspects (rupa) of the logical reason. This is the context which requires a description of what a proper, i.e. justified (samarthita) logical reason is like. And what follows is - apart from various digressions - a succinct formulation of Dharmakirti's logic. He first states the three well-known kinds of possible logical reasons as a means for proving something not perceived: essential property (svabhava), effect (karya) and non-perception (anupalambha).8 Such a reason has to be justified, else its propounder is defeated. Justification (samarthana) of the reason means proving the reason's presence in the problematic locus and proving the reason's pervasion by the argued property." There follow detailed descriptions of what exactly constitutes a proof (sadhana) of the pervasion (vyapti) by the argued property (sadhya) for each of the three kinds of logical reasons, in other words a theory of the ascertainment of a logical nexus.10 In this paper I would like to limit the discussion to the prescriptions concerning the mode of . establishing a pervasion in the case of an essential property as reason (svabhavahetu), because it is in this area that we find a manifest shape of Dharmakirti's final logical thought.11 Before we go into the details of this theory it is of considerable relevance for our further remarks to acknowledge again the fact that the context that I described briefly above which demands a justification of the logical reason is of a purely general kind. Thus this context will not content itself with an answer provided within the limited frame of the logical operations of a particular inference only, e.g. the sattvanumana. The justification of a svabhavahetu as proposed below must be the result of a method applicable in all cases of essential properties used as logical reasons. The definition of the proof of pervasion (vyaptisadhana) for a logical reason of this kind in the Vadanyaya is this: "In the case of this (essential property as logical reason) a proof of pervasion is the demonstration of a valid cognition which negates (the logical reason) in the contradictory opposite (of the argued property)."12 This definition is already extant in a fuller linguistic form in the Hetubindu, when the ascertainment of positive concomitance (anvayaniscaya) is defined,13 but the complementation of the two terms - namely viparyaye and badhaka-- is also supported by a later repetition in negative form from the Vadanyaya.14 7 VN 5.1f.: tasya samarthanam sadhyena vyaptim prasadhya dharmini bhavasadhanam, and VNT 3,26: sadhanangasyasamarthanam tris api rapena niscayapradarsanam. 8 VN 3,3f.: trividham eva hi lingam apratyaksasya siddher angam svabhavah karyam anupalambhas ca. 9 VN 5,1f.: tasya samarthanam sadhyena vyaptim prasadhya dharmini bhavasadhanam. 10 In the case of a svabhavahetu: VN 6,5-13,2, of a karyahetu: VN 13,3-18,2, and of an anupalambhahetu: VN 18,3-60,3. 11 The extent to which the results gained here are also meaningful for an interpretation of the karyahetu remains to be investigated. 12 VN 6,5f.: atra vyaptisadhanam viparyaye badhapramanopadarsanam. 13 HB 4,5: sa sadhyaviparyaye hetor badhakapramanattih. The proposition that Arcata relates sa to vastutas tadbhavata (HBT 44,3f.) is a mistake, for only... anubandhasiddhih of the previous sentence can be referred to meaningfully and grammatically. 14 VN 8,2: sadhanasya sadhyaviparyaye badhakapramananupadarsane...

Page #4
314 E. Steinkellner Dharmakirti follows this definition with a more detailed explanation than that given in the Hetubindu. And with the help of these explanations we can now answer several vital questions: What is the sadhyaviparyaya, tentatively translated as "the contradictory opposite of the argued property"? What is the meaning of badhaka? And what is the cognition called badhakapramana? And in general, aside from his having to explain how a logical reason is justificd, is there an additional reason for explaining the method of ascertainment in yet more detail? To answer the last question first, an additional reason can be found in Dharmakirti's life-long concern with the ideas of his teacher Isvarasena. The basic motif for defining the pervasion's proof in this way goes back to the lack of certainty with regard to the logical relation between reason and argued property that seems to have been discovered by lsvarasena." When Dignaga formulated the third characteristic of a logical reason, i.e. its absence in cases where the argued property is absent (vipakse 'sattvam), he did not provide any means of control for this "induction domain"18 thereby opening a door to "the demons of doubt" with regard to the realm beyond a non-omniscient ordinary being's powers of cognition. The impossibility of ascertaining the absence of the reason in the absence of the argued property thus became one of the causes for the fact that Dignaga's formulation of these three characteristics of the logical reason can be considered as a statement of the necessary conditions of certainty", but not as a statement of "the sufficient conditions of certainty".20 Isvarasena, as I have previously shown, not only discovered this problem but also tried to solve it in two ways: by evolving a theory of non-perception (anupalambha) in order to provide a means of control with regard to the absence of something (abhava), and by introducing a fourth characteristic 15 VN 6,5-11,1 The proof of a logical nexus (vyaptisadhana) in the case of an essential property as logical reason (svabhavahetu) 6,5-8,2 1) Definition: 6,5-6 a) Definition: demonstration of a valid cognition which negates the reason in the contradictory of the argued property (viparyaye badhakapramanopadarsana) 6,6-8.2 b) Example in form of a prasanga in the case of the sandnumana 8,2-11,1 2) Explanation: 8.2-9,2 a) The necessity of a sadhyaviparyaye badhakapramana: without it the contradiction between the reason and the contradictory of the argued property cannot be established, the negative concomitance (vyatireka) would therefore be doubtful and the reason indeterminate (anaikantika). 9,1-2 a) The absence of the reason cannot be established merely by the non-perception (adarsanamdtra) of a non-omniscient person. 9,3-6 b) The function of a badhakapramdna (in case of the samdnumana): hypothetical establishment of the contradictory of the reason through negation of the reason's pervading property in a hypothetical locus. 9,7-11,1 c) Refutation of the objection: infinite progress (anavastha) is unavoidable if non-perception is not considered as valid. * = 9,7-12 + 8,5-6 + 10,1-11,1 10 Earlier interpretations of these terms deviating in part from mine can be found in Yuichi Kajiyama, An Introduction to Buddhist Philosophy, Kyoto 1966: 97 and Katsumi Mimaki, La refutation bouddhique de la permanence des choses (sthira. siddhidasana) et la preuve de la momentaneite des choses (ksanabhangasiddhi), Paris 1976: 55 and 59ff. 17 cr. Steinkellner 1966: 82. 18 The term was proposed by Richard P. Hayes, "An Interpretation of anyapoha in Dinnaga's General Theory of Inference", in Buddhist Logic and Epistemology (cf. above note 4) (31-57): 32. 19 cr. Richard P. Hayes, "Dinnaga's Views on Reasoning (svarthanumana)". Journal of Indian Philosophy 8/3, 1980 1219277): note 33. As the second cause can be considered as the fact that in Dignaga's theorem of the three characteristics of a logical reason (trairupya) the problematic case (paksa) is not part of the induction domain". 20 Cf. note 18.

Page #5
The Logic of the svabhavahetu in Dharmakirti's Vadanyaya 315 of the reason (abadhitavisayatva) especially related to this problem.21 For both of these attempts he was continuously criticized by his pupil Dharmakirti.22 Dharmakirti made Isvarasena's problems his own from the very beginning of his logical work, but rejected his solutions, trying rather to find an answer of his own. On another occasion I have tried to show how Dharmakirti was motivated by this task and how he claborated his own solutions.23 Here, all I want to point out is the fact that it is still this same problem he himself declares to be the motif for the last of his formulations concerning the certainty of a logical relation. With reference to the example of the sattvanumana, he states that this proof of pervasion is the necessary condition for dispelling any insecurity with regard to the logical relation of reason and argued property: "Since there is no incompatibility (virodha) [between the proving property and the contradictory of the argued property] if a valid cognition is not demonstrated in this way, such that it negates the proving (property) in the contradictory of the argued property, the suspicion (sanka) that it might be existent or produced and still eternal will never wane, even if an occurrence [of the proving property] in [a locus of the occurrence of] the contradictory of the [argued property] is not perceived."24 And he continues to express himself on this most important basic and original motif of post-Dignagean logical research by stating: "Since the negative concomitance (vyatireka) would then be doubtful, this would be a case of the indeterminate apparent reason (anaikantikah hetvabhaJah).-25 Finally he repeats for the last time his central objection to Isvarasena's attempt to avoid this consequence by providing non-perception (anupalambha) as an adequate means of control:26 "For27 absence (vyavrtti) [of the reason in the absence of the argued property] is not (established) by mere non-perception (adarsanamatra), since the non-perception of someone who does not see everything does not prove absence in the case of matters distant (in terms of place, time and condition] for somebody with a vision of our (ordinary) kind (simply) does not see certain things even though they exist."28 So much for the problem that here is expressly stated to be the motif for formulating the definition of the proof of the pervasion in the manner proposed and not in any other. To repeat: the logical relation between reason and argued property is uncertain and the reason therefore is unjustified as long as the negative concomitance remains doubtful. Now we can return to the terms of the new method proposed and to our previous questions in order to see whether these terms create the necessary conditions for avoiding this problem, thereby offering a means of providing logical certainty. 21 Cf. Steinkellner 1966: 82f. 22 Cf. Steinkellner 1966: 75ff. and the refutation of the sadlaksano hetu in the Hetubindu (cf. E. Steinkellner, Dharmakirti's Hetubinduh, Teil II, Ubersetzung und Anmerkungen, Wien 1967: chapter VI). My hypothetical assumption that this hetu-theory was Isvarasena's has meanwhile been corroborated by external evidence as well (cf. Steinkellner 1988: note 47). 23 Cf. Steinkellner 1988. 24 VN 8,2-4: sadhanasya sadhyaviparyaye badhakapramananupadarsane virodhabhavad asya viparyaye vrtter adarsane <'pi> san kako va syan nityas cety anivrttir eva sankayah. 25 VN 8,4-9,1*: tato vyatirekasya sandehad anaikantikah syad dhetvabhasah. The lines VN 8,5-6 are misplaced and belong to the end of the next page, after 9,12. 26 First formulated in PVSV 12,4ff,; cf. Steinkellner 1966: 75f. 27 Following Santaraksita's explanation (VNT 12,7): apisabdo yasmadarthe. 28 VN 9,1-2: napy adarsanamatrad vyavrttih, viprakrstes asarvadarsino 'darsanasyabhavasadhanat, arvagdarsanena salam api kesancid arthanam adarsanat.

Page #6
316 E. Steinkellner What is the sadhyaviparyaya and - as a corollary to this question - why does Dharmakirti use the new term viparyaya? The synonyms available, vipaksa 29 and sadhyabhava 30 explain only the logical function of sadhyaviparyaya but not the concept of the term itsell, i.e. the particular kind of "absence of the argued property" needed. First of all we may ignore for the time being and for our purposes the ambiguity of the term sadhyabhava and its use which consists in the fact that it either refers to a locus of such nature that it is a case of the occurrence of the discussed property, or to the property itself. When Dharmakirti in the above mentioned sentence says that a suspicion regarding the "induction domain" cannot be ruled out, when a badhakapramana is not demonstrated," the reason is given to consist in the absence of the relation of incompatibility (virodhabhava), i.e. an incompatibility between the proving property (sadhana) and the sadhyaviparyaya. The result of the function of the badhaka is therefore indicated to consist in the cognition of an incompatibility (virodha) which makes all doubt impossible. And the sadhyaviparyaya in this case is clearly understood as a property, since only as such can it be incompatible with the proving property. Such an incompatibility then * necessarily has to cover the whole logical field and cannot be understood as an "incompatibility of contrariety" (sahanavasthana), but only as an "incompatibility of contradiction" (parasparaparihara) between two contradictories, where a third term is impossible. The requirement of stringency applied to the incompatibility between the proving property and the sadhyaviparyaya in order to remove all suspicion entails a like requirement for the definition of the terms that exclude each other, or else the incompatibility would be meaningless for a statement of negative concomitance (vyatireka). 29 Cr. HBT 44,4; sadryasya viparyayo vipaksah tatra. SU C. VN 9,11: "For in this way the reason would be proven to be absent in the absence of the argued property (sadhyabhave 'sau), ..." (avam sa heruh sadhyabhave 'sau sidhyer, ...). And VN 9,96: yad adarsanam viparyayam sadhayari hetoh sadhyaviparyaye ..., where Santara ksita glosses viparyayam abhavam (VNT 11,12), although this refers only to the absence of the reason. 31 Ch. VN 8,2-4 (note 23). 32 Dharmakini works with two kinds of incompatibility (virodha) (cf. PVSV 5.13-16; 104,15-17; PVin II 13,2-6; NB III 7275 and F. Th. Sicherbatsky, Buddhist Logic II, Leningrad 1930: 187ff.). The first is defined as sahanavasthana (NBT 199,3; cf. also PVSVT 36,16 - Karakagomin's Pramanavdrttikasvattisika, ed. Rahula Sankrtyayana, repr. Kyoto 1982]). exemplified by the case of cold and warm" (cf. PVSV 6,11; NB III 74), to be taken as an incompatibility between two opposed facts, and is translated best when used in the logical context as 'contrariety". This kind of incompatibility, when talking of terms, is an incompatibility of contrarics, where a third term is always possible. The other kind of incompatibility is defined as parasparaparihara (NB III 75, or anyonyopalabdhiparihara PVSV 5,141.), i.e. as "the mutual exclusion of two terms, and exemplified by the cases of being and non-being" or "eternal and non-eternal" (NB III 75: PVSV 5.15). This is an incompatibility between contradictories, a third tcrm being impossible. it as 'contradiction". Clear and most useful observations on possible translations of the term virodha and the types of opposition it may refer to can be found in a recent paper by Nandita Bandyopadhyay ("The Concept of Contradiction in Indian Logic and Epistemology", Journal of Indian Philosophy 16, 1988 (225-246): note 1). Following her considerations we should not translate virodha by "opposition" but use the term "incompatibility". She also proposes the term "absolute contradiction" and "relative contradiction for the relation of contradiction in the strict sense and for the relation of contrariety respectively, and suggests that "contradiction as a term in the wider sense of incompatibility may be used as an equivalent of virodha. In the case of Dharmakirti's usage of the term virodha it will not be necessary to adopt these latter generalizations because it is clear in most cases which type of virodha is intended. I therefore translate virodha by "incompatibility, and interpret it as the relation of contrariety on the logical plane when defined as sahanavasthana (meaning "actual or physical contradiction", cf. ibid., 230-232), and as the relation of contradiction when defined as parasparaparihara ("mutual exclusion"). 33 cr. Santaraksita's explanation: yadi sadhanasya sadhyavipar yasya ca parasparam virodhah syat, bhaved adarsanamarrena(:-se) Sankaya vyavretih (VNT 8,26f. + 11,281.)

Page #7
The Logic of the swabhavahetu in Dharmakirti's Vadanyaya 317 Now, if the sadhyaviparyaya were considered to be nothing but sadhyabhava, i.e. the argued property's absence, the insufficiency of induction cannot be accounted for and a contradiction of the reason to it would not yield the required stringency. This is only the case when sadhyaviparyaya is understood as "the property which is in contradiction to the argued property", for only the property contradictory to the argucd property is not only a certain casc of the absence of the argued property but also excludes a third possibility that could be taken as a further area for the occurrence of the reason, thereby causing the latter to be indeterminate. Accordingly the vipaksa would then have to be taken as "that locus where the contradictory of the argued property occurs." And this amounts to no less than a new way of expressing the vipaksa or sadhyabhava. I would consider the new clcment as being that the locus of reference for a formulation of the negative concomitancc (vyatireka) is determined by its contradictory character. If the sadhyabhava is thus contextually defined as that which is a property in contradiction to the argued property or a locus of its occurrence, this new concept itself may have been sufficient reason for Dharmakirti to use a new term, viparyaya, "the reverse" or "contradictory", instead of the traditional abhava.34 The term viparyaya would then be understood as actually defining the absence (abhava) of either the proving or the argued property as their respective contradictories." In giving such a concrete meaning to the term viparyaya instead of merely taking it as a synonym of abhava we are supported by another passage, where Dharmakirti explains why he does not deny in general that non-perception can prove an absence:36 "The non-perception (adarsana) which proves the viparyaya of the logical reason in the viparyaya of the argued property is called a valid cognition which negates this (logical reason because it conceptually establishes (pratyupasthapanat) [in the viparyaya of the argued property137 a (property which is) contradictory (viruddha) (to this reason) 38-39 And we are further supported by the subsequent argument which points to the establishment of the negative concomitance for the sake of certainty as the task of the cognition under discussion: "For in this way the reason would be proven to be absent (asan) in the absence of the argued property An additional reason might be found in an attempt finally to clarify (also terminologically) the meaning of vipaksa when we think of the three possible interpretations Dignaga had alrcady dealt with in the Pramanasamuccaya (cf. note II, 53 in my translation of the Hetubindu and the sub-note for these interpretations; cf. also Kajiyama (cited in note 16) note 181 and Steinkellner 1979: note 81). Dharmakirti's usage of the term viparyaya is, of course, not limited to the meaning "contradictory opposite". He does, however, use the term when he refers to the incompatibility of mutual exclusion" (parasparaparihara), which Stcherbatsky (cited in note 32. p. 187 note 3) refers to as the "logical opposition' or Contradiction. This is the case not only in the Nyayabindu (cf. NB III 85: sadyan iparyaya, cf. also III 81, 84, 88) but can already be found in his earlier works (cf. PVSV 5.2; 121,25; 174.22). Besides this strict logical usage as 'contradictory opposite'. a meaning of merely "the opposite is also attested, c.g. in PVSV 5,4; 78.23 (v. 156c158c]): 79.5: 112.19: PV 11 226d: PV III 8Sd: PV IV 195b). The term also functions in a spiritual context where it serves as a synonym of pratipaksa (cf. PV 1221b - 223b) and PVSV 111,7; 163,8) and can also connote "change" (cf. PV I 232b - 234b) and PVSV 116,13; 135,13). The only evident usage as a synonym of abhava is PVSV 117,22 (cf. PVSVT (cited in note 32J 425,12) and possibly also PV III 85d (cf. PWV 143.3). Under these circumstances it seems evident that in Dharmakirti's language viparyaya cannot be taken simply as a synonym for abhava. 36 VN 9,9: na, bhavasadhanasyddarsanasyapratisedhat. 37 C. VNT 11,176. 38 C. VNT 11,14: asyeti vartale. 39 VN 9,9-11: yad adarsanam viparyayam sadhayati hetoh sadhyaviparyaye, tad asya viruddhapratyupasthapanad badhakapramanam ucyate.

Page #8
318 E. Steinkellner (sadhyabhave), if there it were opposed (badhyeta) by a (property) contradictory to itself (svaviruddha) for which a valid cognition is available (pramanavat)." 40 From the second statement we can understand clearly, that the absence (abhava) of the reason is proven, if its contradictory has been established in the absence of the argued property (sadhyabhave). That means that the negative concomitance (vyatireka) is proven by a cognition that positively establishes the contradictory of the reason. And it is as a statement of this positive establishment of the contradictory of the reason in the occurrence of the contradictory of the argued property that we have to understand the beginning of the first statement therefore, and not as a statement of the negative concomitance (vyatireka). I would therefore propose to translate this beginning as "that non-perception which proves the contradictory (viparyaya) of the logical reason in the contradictory of the argued property (sadhyaviparyaye)." For Dharmakirti says that this cognition is called negating the reason because it establishes its contradictory. And he continues to say that this is the only way to prove the negative concomitance. The absence (abhava) of the reason is then a certain consequence of the presence of its contradictory (viparyaya), but not the viparyaya itself." If the phrase viparyayam sadhayati hetoh referred not to the contradictory of the reason, but to its absence (abhava) we would have no meaningful area left in this sentence for the contradictory of the reason (asya viruddha-) which is said to be conceptually established." What is the cognition called badhakapramana and what is the meaning of badhaka? Dharmakirti explains this cognition in two ways. These can be distinguished as referring to its cognitional and its logical function respectively. Its cognitional function is explained when he says that "it conceptually establishes a (property which is) contradictory" (vinuddhapratyupasthapanat VN 9,10).44 By the act of pratyupasthapana this cognition "provides" (akaryati VN 9,4.) the desired property. This cognitional function has been explained by Dharmakirti in PV IV 228-23646 and can be connected with the verb prakalpayati (PV IV 233d - PVin II 25d) which refers to the conceptual cognition of establishment of a "usage" (vyavahara) without a real objective basis (asatyartha) when it affirms (vidhi) or negates (nisedha). Here in the Vadanyaya this function is conveyed by the term pratyupasthapana, which can therefore be translated as "conceptual establishment". The logical function of the badhakapramana is explained when Dharmakirti identifies it as vyapakadharmanupalabdhi (VN 8,6),47 which already occurs in HB 4,18f. as vyapakanupalabdh148 in the 40 VN 9,11f: evam sa heruh sadhyabhave 'san sidhyet, yadi tatra pramanavara svaviruddhena badhyeta. 41 VN 9,91.: yad adarsanam viparyayam sadhayari hetoh sadhiyaviparyaye, .. 42 Here I do not follow Santaraksita who glosses yad adarsanam viparyayam abhavam (!) sadhayati (VNT 11.12.) * In addition it may be noted that the two formulations viparyayam sadhayati hetoh sadhyaviparyaye (VN 9.91.) and sa heruh sadhyabhave 'san sidhyer (VN 9,11) are too close not to be questioned with regard to the reason for their linguistic difference if viparyaya and abhava/asat are taken as synonyms. 44 Arcata paraphrases tad vipartiadharmapraryavasthapakam (HBT 44,5), where the adjective pratyavasthapaka evidently bears the same meaning as our substantive pratyupasthapana. 45 Cf. also HBT 44,23f.: tad viparyayarapasydsarvakarsandt. 40 These verses were incorporated into PVin II as w. 20-28 (for the new counting of the verses in PVin II cf. Steinkellner 1988: 1434) and are translated in Steinkellner 1979: 42ff. 47 The two lines VN 8.5-6 are misplaced and should follow 9,12. 48 This kind of non-perception (anupalabdhi) is explained in PV I, 29a-c' (=31a-c") = PVin II 74a-c' (cf. Steinkellner 1979:

Page #9
The Logic of the svabhavahetu in Dharmakirti's Vadanyaya 319 function. This "non-perception of the pervading property" serves as the reason for a negation of its pervaded property, in our case the original logical reason. Its resultant inferential cognition is the nonexistence of the pervaded property, its negation. This cognition therefore results in a "conceptual establishment" (pratyupasthapana) "cancelling" or "negating" (badhaka) this property. badhaka thus means strictly "negating" here, and a "badhakapramana" is a "negating valid cognition" which in the case of the svabhavahetu has the form of an inference from the non-perception of the pervading property (vyapakanupalabdhi). After these clarifications we can summarize in the words of Dharmakirti that the proof of pervasion (vyaptisadhana) in the case of an essential property as logical reason (svabhavahetu) consists in the demonstration of a valid cognition which negates (badhaka) the logical reason in the contradictory opposite of the argued property, and that this negation is inferred from the non-perception of the reason's pervading property (vyapakanupalabdhi). A final question remains to be asked, however: whether this method of proving the pervasion is to be applied in all cases of essential properties as reason, or whether different methods are still conceivable for different kinds of such reasons as proposed, e.g., in Dharmakirti's earliest work?49 I said in the beginning that the context of the Vadanyaya, but also of the Hetubindu, is of general kind and requires a proposition for the justification (samarthana) of a svabhavahetu valid in all cases. Since Dharmakirti does not in fact offer alternative methods we have to see how the method proposed is put to work not only in case of the sattvanumana, where we can rely on the example provided by the Vadanyaya, but also in case of the simsapatvanumana which is not mentioned at all in the Vadanyaya or the Hetubindu. Dharmakirti uses the sattvanumana to exemplify the various points of the logical structure proposed. He begins with the proof as such as an example for the justification (samarthana) of a reason which consists in proving its occurrence in the problematic locus (dharmin) and proving the pervasion by the argued property:50 "E.g.: 'What is existent or produced, all that is impermanent; like a pot etc.; (and) sound is existent or produced."51 He then exemplifies the demonstration of the valid cognition which negates (badhakapramanopadarsana) the reason in the contradictory (viparyaya) of the argued property as proof for the pervasion 136f. for a translation). On Dharmakirti's theory of anupalabdhi as a logical reason for a cognition of non-existence cf. PVin II 11,12ff. as well as HB SS 4.3 with its extensive digression on the nature of negative cognition and my translation and notes (cited in note 22: 60ff., 154ff.) 49 In my paper delivered at the First International Dharmakirti Conference in Kyoto 1982 I said in conclusion with regard to Dharmakirti's first work: "While a method to ascertain the causal relation has been already developed here, a likewise generally applicable method of ascertaining the relation of identity is not given. The required valid cognition is said to be demonstrated by the example, but only in the case of the ksanikaranumana is an additional inference developed for proving the pervasion (vyapti)." At that time I was under the impression that Dharmakirti wanted to say that the demonstration by means of an example is in fact a way of ascertaining the pervasion wherever an example can be provided for essential properties. But I now think that this is not the case at all. The task of the presentation of an example is only to indicate an already established valid cognition regarding the real identity and the logical relation of two essential properties based upon it. And this necessary relation between words, concepts, properties as based upon the undivided identity of the instance of reality referred to does not have to be established or ascertained in a particular way for the simple reason that within the culturally given system of linguistic conventions it is already known (prasiddhi cf. PVSV 16,30f. and below 321). 50 VN 5,1: tasya samarthanam sadhyena vyaptim prasadhya dharmini bhavasadhanam. The sequence implied in this formulation is said to be logically irrelevant. (VN 6,2-5). 51 VN 6,1f.: yatha yat sat krtakam va tat sarvam anityam; yatha ghatadi; san krako va sabda iti.

Page #10
320 E. Steinkellner (vyaptisadhana): "If everything existent or produced were not perishing at every moment, it would be only non-existent because it would be excluded from what is defined by capability for causal efficiency, since for a non-momentary (thing) causal efficiency is neither possible successively nor simultancously, For that which is defined such that no capability can be stated (of it), is a non-existent. 52 The cognitive function of the badhakapramana which negates the reason thereby "conceptually establishing" its contradictory is exemplified with regard to the reason "existence" (sattva) as establishing its contradictory "incapability as the defining characteristic of a non-existent":53 "Where (causal efficiency) is not possible successively or simultaneously, that is incapable for every (effect); and this (impossibility) is extant in a non-momentary (thing)." And finally Dharmakirti exemplifies the argument for the particular pervading property (vyapakadhama): "In this case the capability (for causal efficiency), is proven as pervaded by the possibility to produce an effort) successively or simultaneously, because there is no other way (of producing)." From these exemplifications the following structure results in case of the sattvanumana: the logical pervasion (vyapti) between the essential properties (svabhava) "existence" (sattva) as logical reason (hetu) and "momentariness" (ksanikatva) as argued property (sadhya) is proven by a non-perception of the reason's pervading property "possibility of successive or simultaneous (efficiency) (kramayaugapadyayoga) in the case of "non-momentariness" (aksanikatva) as the contradictory of the argued property, because this non-perception negates (badhaka) the reason thereby conceptually establishing "non-existence" (asattva) as its contradictory property (svaviruddha). In this way the non-perception of the pervading property (vyapakanupalabdhi) proves the contradictory of the reason (h) in the contradictory of the argued property (-s) and thereby establishes the pervasion between reason (h) and argued property (s): This is the schematic model that can be drawn for the sattvanumana on the basis of the information available in the Vadanyaya. For the Simsapatvanumana or comparable inferences we are not given any specific indications as to what the schematic model should look like: there are no trees or simsapas in either the Vadanyaya or the Hetubindu. So we are forced to construct a schematic model on the basis of whatever information is available in Dharmakirti's work as a whole, in order to find out whether the newly developed method for the ascertainment of the logical nexus (vyapti) can be applied at all in this case. This is, of course, an experiment whose result will either prove or disprove the assumption that the proposed method is valid for every essential property used as logical reason (svabhavahetu). 52 VN 6,6-8.2: yadi na sarvam sat korakam va pratiksanavinasi syat, aksanikasya kramayaugapadyabhyam anhakriyayogad arthakriydsamarthyalaksanato nisam iry asad eva syat. sarvasamarthyopakhydvirahalaksanam ni ninupakhyam iti. > VN 9,3-5: badhakam punah pramanam ...... iti pravartamanam asamarthyam asallaksanam akarsari. S4 VN 9,3-4: yatra kramayaugapadyayogah, na tasya kvacit samarthyam, asti caksanike sa iti. > VN 8,5f.: tatra samarthyam kramakramayogena vydpiam siddham, prakarantarabhavdi.

Page #11
The Logic of the svabhavahetu in Dharmakirti's Vadanyaya 321 That we are allowed or rather forced to construct a model for the case of the simsapatvanumana in accordance with that schema given for the special case of the sattvanumana by Dharmakirti himself, can be justified by two interrelated arguments. Firstly, the assumption that the method prescribed by Dharmakirti in the Vadanyaya was meant to be valid only for the sattvanumana but not for all other cases of possible inferences using an essential property as reason (svabhavahetu) would imply that Dharmakirti had presented an incomplete theory of logic in both of his last works, an assumption that one can hardly defend in the face of the general meaning of these works, particularly the Hetubindu, as formulations of a theory of logical reason (hetu). Secondly, we would have to assume that instead of demonstrating the sattvanumana as a crucial inference of especial Buddhist concern by means of, and on the basis of a logical method developed for undisputable ordinary cases of logical reasons, Dharmakirti had expounded a logical theory developed only for the sattvanumana. An assumption of this kind would be against all historical and systematic reason. Nevertheless, I must again emphasize that what follows is an experiment of interpretation. If we take the famous example from Dharmakirti's first work: vrkso 'yam simsapatvat,56 our sadhyaviparyaye badhakapramanam could be taken to work in the following way: Since the pramana which negates the logical reason in the field of this property is a non-perception of its pervading property (vyapakanupalabdhi), we have to answer the question as to what could serve as this pervading property in the case of the property simsapatva. Here a brief line of Dharmakirti's is of assistance: "Since only a certain particular possessor of branches etc. is known in this way (i.e. as 'simsapa")."57 What can we deduce from this sentence that is of relevance for our question? "Tree" and "simsapa" are both designations (vyavahara) which refer to general properties that can be understood as "the capability for the designation 'tree" (*vrksavyavaharayogyatva) and "the capability for the designation 'Simsapa"" (simsapavyavaharayogyatva) respectively according to Dharmottara's explanation.58 In the sentence quoted above Dharmakirti means that the property "possessing branches etc." (sakhadimattva) as extant in the particular thing which is designated as "simsapa" is the reason (nimitta) for its designation as "tree". If this essential property, when absent, may force the property of a "capability for the designation as 'simsapa" to be absent, as stated in the main clause preceding, 59 it is evident that it is conceived as a pervading property (vyapakadharma) of the latter. 56 PVSV 2,16; NB II 16. 57 PVSV 16,30f.: sakadimadvisesasyaiva kasyacit tathaprasiddheh. 58 According to Dharmottara the above inference has the following meaning: "This (thing) can be called 'tree', because it can be called 'simsapa" (vrksavyavaharayogyo yam simsapavyavaharayogyanvar, NBT 106,11: Cf. also the formulation of this proof in DhPr 107,9f.). And Dharmottara goes on to explain: "In this case a stupid person in an area rich in simsapas unversed in the usage of (the word) simsapa, when somebody shows him a tall simsapa and says 'this is a tree' then out of stupidity determines the simsapa's tallness too as a reason (nimitta) for the usage of (the word) 'tree', (and) then (further) determines the small simsapa which he sees as a non-tree. This block is introduced into the usage of (the word) 'tree' as having no other reason than simsapaness. Tallness etc. here [in this simsapa, or in this area?] do not constitute further reasons for the usage of (the word) 'tree', only simsapaness is the reason; that is: possessing branches etc. (sakhadimattva) as extant in a simsapa is the reason (nimitta)." (yatra pracurasimsape dese viditasimsapavyavaharo jado yada kenacid uccam simsapam upadaryocyate yam vrksa' iti tad asau jadyac chimsapaya uccatvam api vyksavyavaharasya nimittam avasyati tada yam evanuccam pasyati Simsapam tam evavrksam avasyati. sa madhah simsaparvamatranimitte vrksavyavahare pravartyate. noccatvadi nimittantaram iha viksavyaharasya, api tu simsapatvamatram nimittam -simsapagatasakhadimattvam nimittam ity arthah. NBT 106,11-107,2). Cf. also DhPr 107,22-28. 59 Therefore either an essential property (svabhava) which is connected with that [real existence of the reason] as such may cause the very essence (bhava) [which is propounded as a reason] to be absent (nivartayet). [PV 1 23a-c'(=25a-c')] - e.g. the tree a simsapa." (tasmat tanmatrasambandhah svabhavo bhavam eva va / nivartayet-yatha vrksah simsapam. PVSV 16.2730). Cf. the prayoga formulated by Kamalasila: yo yadvyapakadharmarahitah sa tadvyaptadharmavikalah, yatha vrksatvadharma

Page #12
322 E. Steinkellner Therefore we can assume that this property would have been Dharmakirti's candidate for the position of the vyapakadharma of the vyapakanupalabdhi had he cared to explain the simsapatvanumana as well. We can therefore consider as a property contradictory to the argued property (sadhyaviparyaya) in the case of the simsapatvanumana thc property "non-capability for the designation 'tree" (*vsk$avyavaharayogyatva) and the property "possessing branches etc." (fakhadimattva) as the pervading property (vyapakadhanna) of the logical reason "capability for the designation 'simsapa" (simsapavyavaharayogyatva). The logical nexus between the properties "treeness" and "Simsapaness" is then clearly ascertainable by means of the viparyaye badhakapramanam: in the case of non-capability for the designation 'tree' a capability for the designation 'Simsapa' is denied because of the non-perception of its pervading property "possessing branches etc." Thus we would arrive at the same schematic model as in the case of the sattvanumana with the only difference being that the argued property (sadhya) and the proving property (sadhana) are not coextensive:60 In this way the difference between these inferences from two kinds of essential properties as reasons would not be constituted by different methods in ascertaining their logical nexus with the respective argued properties. In both cases the logical nexus (vyapti) of the reason and the argued property would be ascertained by an additional inference, the vyapakanupalabdhi-argument, which proves the absence of the first logical reason (sattva or simsapatva) in the contradictory of the argued property (aksanikatva or vyksavyavahardyogyatva) by a non-perception of the first reason's pervading property (kramayaugapadyayoga or sakhadimattva) as a second logical reason. This ends our experiment, I feel successfully, and we are now able to draw the following conclusion: the method proposed in the Hetubindu and explained in more detail in the Vadanyaya for an ascertainment of the logical nexus (vyaptiniscaya) in the case of an essential property as logical reason (svabhavahetu) is in fact, as should be expected, prescriptive for every logical reason. Towards fanyo ghafadis tadwydptasimsapan adharmavikalah. (Tattvasangrahapanjika, ed. D. Shastri, Varanasi 1968: 1025, 17-19). 60 It is not of logical relevance here that a part of the argued property's loci, i.e. trees other than simsapds would also be loci of the absence of the reason, because the logical nexus is established only with regard to the absence of the argued property proper. Moreover, Dharmakirti defined the pervasion (vydpri) in the Hetubindu as an asymmetric relation: "Pervasion is the necessary existence of the pervading (property) where the pervaded property exists) or the existence of the pervaded (property) only when the pervading property exists)." (HB 2.7f.: vydptir vyapakasya tatra bhava ala vydpyasya vd tatrana bhavah.) These two definitions can be written as (x) (hx sx) meaning "For all x is valid: if (is) h. then x (is) s. and as (x) (-5X -hx) meaning "For all x is valid: if (is) not s, then x (is) not h'.

Page #13
The Logic of the svabhavahetu in Dharmakirti's Vadanyaya 323 the end of his work Dharmakirti proposed a new and generally valid method, one that was no longer flawed by a different treatment of the same kind of reason. What is still to be investigated is the question of whether the different treatments of the svabhavahetu and the karyahetu in this respect were not also resolved in a certain sense in order to design a homogeneous logical system, or at least, whether there are not indications to be found in Dharmakirti's work that he was aiming in this direction. At the beginning of this paper I referred to Frauwallner, who paved the path towards a historical interpretation of Dharmakirti's work. Let me now end this investigation by referring to another great scholar who has, in many important ways, promoted our knowledge of Dharmakirti's thought and tradition with his critical analysis of the theories and later polemics: Satkari Mookerjee. Satkari Mookerjee long ago recognized with reference to the sattvanumana that its treatment amounts to an acceptance of a theory of "internal concomitance" (antarvyapti), although he knew that it was not accepted in the Buddhist epistemological school except for the late Ratnakarasanti.1 Latero Mookerjee saw that this theory was a consequence of Dharmakirtis concept of the svabhavapratibandha as the real fundament of a logical relation: "The relation of antarvyapti is then a deduction from Dharmakirti's conception of natural concomitance (svabhavapratibandha)."63 Mookerjee also felt Dharmakirti's importance for the beginnings of the Jaina tradition of the antarvyapti-theory with Siddhasena Divakara.64 In the Nyayavatara this theory and the term for it is to be found - according to our present knowledge - in total isolation and without any Jaina background, but in Dharmakirti we can now say that this theory seems to be the final product of a life-long occupation with the problem of an ascertainment of the logical nexus at least in the case of the svabhavahetu. And although Dharmakirti did not himself refer to his new theory by the term antarvyapti, he can definitely be considered its creator.65 That his own tradition did not choose to follow these new lines of thought in a straightforward way but chose rather to interpret Dharmakirti with an emphasis on the Dignagean heritage, is another matter. But we can fully support the late Buddhist antarvyaptivadin Ratnakarasanti, who insists on Dharmakirti as the propounder of this theory, when he says that the acarya - whom I consider to be Dharmakirti - regards the example in the formulation of the saltvanumana merely as a concession to slowminded people, but not as logically necessary. 01 Cf. The Buddhist Philosophy of Universal Flux, (repr.) Delhi 1975: 380ff. (the first edition of this PhD-thesis 1932 appeared Calcutta 1935). 04 Cf. "A Critical and Comparative Study of Jaina Logic and Epistemology on the Basis of the Nyayavatara of Siddhasena Divakara", Vaishali Institute Research Bulletin 1, 1971 (1-143): 4-9. I would like to thank Prof. E. Mikogami of Ryokoku University, Kyoto, who called my attention to this work and sent me a copy. 63 ibid., 7. 64 ibid., 83f. W On Dharmakirti's authorship of this theory and on the somewhat enigmatic treatment of the sattvanumana in its logical structure by the later Buddhist logicians cf. the valuable observations and materials collected in Kamaleswar Bhattacharya, "Some Thoughts on Antanyapti, Bahinyapti, and Trairupya" (in Buddhist Logic and Epistemology (cf. above note 4) 89-105), which is in parts a reworking of his paper "Ratnakarasanti and Ratnakinti" (in Surabhi, Sreekrishna Sarma Felicitation Volume, Tirupati 1983, 131-140). Antanyaptisamarthana, in Sir Buddhist Nydya Tracts in Sanskrit, ed. Haraprasad Shastri, Calcutta 1910, 112,4-9. Cf. Mimaki (cited in note 16) p. 52.