Svabhavarat Bandha Again

Author(s): Ernat Steinkellner
Publisher: Ernat Steinkellner
Catalog link: https://jainqq.org/explore/269412/1

Page #1
SVABHAVAPRATIBANDHA AGAIN By E. Steinkellner, Vienna Substantial improvement in our understanding of Dharmakirti's thought depends on a considerable increase in research on various details of his theories and theorems, mainly on details of his linguistic and conceptual usage-not to speak, of course, of the philological toil still being owed to a respectable body of unedited and uninterpreted texts. A gratifying effort of this sort was recently made by Matsumoto Shiro with respect to the term svabhavapratibandha. In a stimulating paper he examined the meaning of this term which is of significant importance in the context of Dharmakirti's logical theory. Although his minute observations decidedly enlarged the scope for an interpretation of the term, his results are still not convincing and, moreover, methodologically problematic. I would, therefore, like to try once more to clarify the meaning of the term. 1 L 2 To begin with, the following is a summary of the main points of Matsumoto's article. In its first part (4981.) M. shows that Dharmottara's interpretation of the compound svabhavapratibandha as an instrumental-tatpurusa is not the only interpretation to be found within the commentators, but that we also find an interpretation as a locative-tatpurusa with Sakyamati (M.: Sakyabuddhi), the oldest commentator of the relevant text, I wish to thank Prof. Albrecht Wexler, Hamburg, who in a very helpful letter kindly explained to me the meaning of Durvekamisra's remarks in connection with sadhanam rteti (DhPr 110, 14-19). This paper was written in the spring of 1982 while I stayed in Kyoto as a guest of the Uni versity of Kyoto, invited by the Japanese Ministry of Education I wish to thank the Japanese authorities, my colleagues, students and friends in Japan, and above all my host, Professor Kaji yama Yuichi, for a most interesting, motivating and fruitful period of life. Abbreviations used are the same as in Matsumoto's article (cf. note 1) and in my edition and translation of the Pramanaviniscaya, 2nd chapter. 1 Svabhavapratibandha. Indogaku Bukkyogaku Kenky 30, 1981, 498-494. 2 This is the interpretation on which I based my translation of the term in previous publications as refered to to by M.

Page #2
E. Steinkeller and the commentator generally considered as authoritative in the tradition of the school M. however, considers both interpretations as "inadequate" in the light of his attempt to de luce the correct interpretation of the compound directly from Dharmakirti's own contextual statements. The result of M's deductions in this second part (497-495) is a new interpretation: The term has two meanings according to the two kinds of reason" (svabhavahetu, karya heru) In case of a stabharaheru the compound must be taken as a locative tatpurusa, in case of the kar yaheru as a genetive-Lalpurusa. But, beyond that, the terms svabhava and pratibandha which constitute the compound connote two dillerent concepts respectively, too, as shown in M.'s conclusion by means of the formula "karyasya sabhavasya pratibandhal" and "bhavasya sabhave prati bandhal" to which he adds the translations "limitation of properties" and "counter.com nection with its essence". Belore discussing M's derivations in detail, and re-examining the philological possibili ties of a clear decision on the question, I would like to state in short the main system atical reasons for not being able to accept M's new proposals, so as not to obscure the basis for further discussions, It seems to be generally accepted that Dharmakinti offers a logical theory that is new in the Buddhist tradition of logic as founded by Dignaga. What exactly is new theory? It is Dignaga's merit in the development of Indian logic to have clarified the formal possibilities of the nexus (vya pri) between logical reason (her) and result (sadhya). And it is Dharmakini's merit to have answered the question for the reason of this logical nexus. The core of his answer consists in his explanation that and how-this logical nexus is based on a relation in reality, and in which concepts such a basis in reality can be demonstrated. Thus, his answer is characterised mainly by two closely interwoven theorems: the theorem of a suabhava pratibandha as the real basis of the logical nexus (avinabhava, miyama, vydpei), and the theorem of the three kinds of logical reasons (trividha ketu) for which such a real basis of their nexus with a result can be assumed. The function of the term svabhava pratibandha in Dharmakirti's theory, therefore, is to indicate that relational character of reality which can be considered as the source and guarantee of logical necessity, too. Since the word praribandha has only a formal meaning, the word swabhava is responsible for connoting the reality needed. The conclu. Svabhavapratibandha again 159 sion to be drawn is: The meaning of the word svabhava in the compound svabhavapratibandha--when used as a term to indicate the reason of the logical nexus--can only be "essence". Dharmak Irti's denotation for the real being as a totality of vanous causal possibilities. An interpretation of the term like M's "limitation of properties" in case of the karya helu seems therefore inacceptable. For the term understood in this way would lose its function within the theory. A limitation of properties" is no indication of the real basis of the logical nexus, and we would have to continue asking for the reason of such limitation Secondly: Although Dharmakirti uses the term svabhava in texts of relevance here with two meanings according to a difference resulting from usage either in ontological or logical statements, it can have only one meaning in case of the term svabhava prati. bandha, and that is the ontological one. A differentiation of the meaning of the word svabhava in the case of this term according to whether it is used to indicate the reason for the nexus of the karyahetu or the sabhavahetu would, in fact, render the term meaningless, since it could no longer serve to indicate the sufficient reason of logical necessity. Then, of course, a totally new interpretation of the term's function, and, beyond that, of the character of Dharmakirui's logic as such, and particularly with respect to its historical originality, would be necessary. From his paper I cannot see whether M. intend ed to propose such a new interpretation, rather I have the impression that he was not aware of these consequences when he tried to solve the riddle of the meaning of the compound Finally: No doubt, the best way to interprete the meaning of a term is by means of observing its immediate context, ie, by deduction from related statements by Dharma kirti himself. I the result of such a deduction, however, differs from an interpretation extant in the exegetical tradition, the latter cannot be discredited simply with reference to one's own interpretation of the dharmak Irtian context. Rather one's own interpretation of Dharmak Irti's terms and statements can be secured and supported only when a reason able explanation is given for (a) the cause for the difference of an exegetical opinion in the commentators from the opinion supposed to be Dharmakirti's by the investigating scholar, and (b.) for the development of different opinions in the commentaries--if there are such differences. To wit: The "wrong reading has to be explained with regard to its causes, too actly is new in this o ch my paper : Philological remarks on Sakyamati's Pramanavar tikaika. In: Studien zum Juinismus und Buddhismus. Gedenkschrift fir Ludwig Alder). Wiesbaden 1980, 283f. 4 Cl. my paper : Wirklichkeit und Begriff bei Dharmakini. WzKS 15, 1971, 179-211.

Page #3
E Steinkeller Is it possible to deduce Dharmakini's intentions with the term sabhava pratibandha, i. e. the correct" interpretation of the compound, directly from any of his statements ? As far as I can see, Dharmakinti expresses himself nowhere in a way that such a deduction is possible. But before we try to find help elsewhere, we have to scrutinize Matsumoto's attempt to deduce the correct interpretation from the respective definitions of the two kinds of logical reasons, which is the major point of his article. M. thinks that it can be assumed that the two modes of wabhava pratibandha will no doubt be indicated" (my spacing) in these definitions "because the two kinds of Teason do not deviate from the results only when there is sabhava pratibandha" (497). This assumption would be appropriate, I admit, if Dharmak Irti really gave a definition of his reason. But, in fact, what Dharmakiri does in PV I 2-4) is only to give a list of the kinds of reasons: he states karya and swabhave in this verse, anupalabdhi follows. Dharmakirti only says that karya is a logical reason, and specifies to what extent with regard to its properties. M., too, does not find the sabhava prati bandha to be indicated in this "definition", for, to satisfy his quest for it, he quotes PVSV 17, 6. And from this he concludes, that sabhava pratibandha in the case of a karyahetu must be "construed as 'rvabhavarya pratibandhah' (limitation of sabhava)", and he further interpretes this as a "limitation of properties" (496). The crucial mistake in his derivation is that he thinks, it is clear that the phrase vabhavam niya mayati' explains the meaning of the word swabhava pratibandha" (496). In fact, the two statements are only very indirectly related. Dharmakinti, after showing that only because of a rvabhava pratibandha there is necessity in the absence of a reason that is an essential property (svabhava), continues to show the same for the absence of a reason that is an effect (karya): "...or a cause (causes the absence) of an effect, because the latter) does not deviate (from the former). (That means:] The absent cause causes the absence of the effect. Otherwise that (which is assumed to Svabhavapratibandha gain be the effect) would not be just the effect of this (cause). An established relation of cause and effect, however, restricts the essence of the effect). The absence of the reason), therefore, is caused only by the rabhava pratibandha in the two ways (as just explained above)." Here the meaning of the word abhava pratibandha is not explained by the phrase Isabhavam niya mayati, but by Aaryakaragabhavab as one of the kinds of a suabhava pratibandha which are refered to by the word wbhayatha. This relation of causality is the cause of the absence of an effect as logical reason, because it has restrictive force (niya mayati) on the essence of the effect. In other words, the restrictive determination (niyama) of the rabhava is the consequence, the effect of an extant rabhaveprati bandha, not this relation itself M's derivation, that the term sabhava pratibandha must be analysed as a genitive tatpurus, i.e. svabhavasya pratibandhab ("limitation of properties") in case of the Aaryaherm is, therefore, unlounded. Moreover, his interpretation of the term praribandha as "limitation" seems to have no other reason than the need for such a "limitational connotation on account of his own assumption that this pratibandha was indicated in the karyahetu"definition" of PV I 2. Yet, such a meaning is questionable at least, and would have to be established by a separate argument. In order to show how the sabhava pratibandha is indicated in the definition of the svabhavahetu M. quotes PV I 23 a-c and the explaining sentences of PVSV 17, 1-3, In his interpretation he differentiates the relation (M. : connection) in question according to whether it is "svabhavabadva connection", which he identifies as "anwbandha 5 loc. cit. 497-495. 6. It is difficult to follow M. when he speaks of two modes of nulla prati hand , because one, naturally, thinks of wadainya and adult the "we made of this relation. But this des mot em to be the meaning of M's expresion Rather the two meaning of the term subhanapatibandha is what he means by the two modes," because these two meanings are the issue of his arguments. 7 siddha tu karyakararathah haan miyarmally why he rabbraprati handhad ubhyatha I understand as indicating the two modes of the wabharatibandha that have been explained before the iw (PVSV 16,28-17,7). since mula translation of adverbs in the is to be prefered to a local one. M translates it locally: "Therefore in both cases [ie sabharu kra and karyahetu)......" (497). That it must be taken modally, and that it means the two ways" just explained above is also the opinion of Sakyamati who says: de bos bied they pistal Eyi kigu ini bye Mwi (PVT 4710). The model meaning further corrob orated by the following paraphru di'i dag tid dan de weby Ni r d bis (PVT 47b2) = adamyen and party w (PVSVT 75, 231.). 9 kar karyam wyobhiarata (PV 1 230- (-25c-d)) tarassa stara karyop sa anuha taf large baryam N yd aiddhas ta karyerasabhare the bham kiyamaally where therapatihand had w i PVSV 17, 4-7. 10 dl the famous verse PV 1 31(33): karyakarusellend harada makatlarin Mhavariyana derinden derdi il on the meaning of n i ya d my paper "Wirklichkeit und Begriff (note ) 1881; and on the restriction of the properties of an effect by the properties of its uk HB 421 and 5 b 12214, and my note in HB IL 125-129.

Page #4
463 E. Steinkeller (connection, following)", or a "bhavanabhapa connection" which he identifies as a "pratibandha (counter-connection, followedness)". His solution for the interpretation of svabhava pratibandha is, that in case of a svabhavahetu it must be understood "as svabhave prati bandhah (connection with svabhava)" (496). But what is the meaning of Dharmakirti's "definition" of a svabhavahetu? The "de. finition" sabhave bhavo 'pi bhavamatranurodhini (PV 1 21c-d (= 41c-d]) means: "[And] also a property (bhava) [is a logical reason) for another) essential property (svabhava) which follows only the first property's) real existence (bhava)." The svabhava praribandha indicated here and shortly mentioned in support of the delinition is made more definite in the commentary on the repeated definition when Dharmakirti says: yo ni bhavamatranurodhl svabhavas tatravinabidvo bidvasya isyate-tadabhave svayam bhavasyabhavah sydd abhedatab (PV 139d (= Weva bhavo bhavamatranurodhi sabhavaity weyate, 14 eva svaya vastu to bhavah. sa carmanam parityajya katham bhavet. ="A property (bhava) is assumed to have a necessary connection (avinabhava) with that other) essential properly (svabhava) which follows only (its) real existence (bhava). -[For] if this (property) was absent, the other property (bhava) itself would also be absent, since [the two are factually the same (abheda). -The very property (bhava) which fo. llows only the existence (bhavamatra) (of the other property) and is called 'essential property (svabhava)', just this is in reality (vastutah) the (other) property. And how could this (other property) exist without itself [i. e, without that property which itfactually-is itself]?" The same description of the svabhava pratibandha is given when Dharmak Irti suppo rts the necessity of the concomitance in difference (wyatireka), the text that is releed n1uu Svabhivapratibandhe Again to by Matsumoto; tasmat tanmatrasambandhah sabhavo bhavaneva va niva. riayet (PV 1 23 -c [=25 a-c 1]) ..... suam ca suabhavam parityajya katha bhavo bhavet, svabhavas yaiva bhavarude (PVSV 17, 11.). "Therefore, either an essential property (svabhava) which is connected only with the existence of the other property] would cause the absence of this very property (bhava). - .. And how could the property (bhava) exist without its essence (svabhava), since that very esse nce (svabhava) is the property (bhava)?" The terms involved may be exemplified in accordance with Sakyamadi's explanations: The designation or concept "tree" as property to be proven (sadhya) is connected, be. ing an essential property (svabhava), only with the existence of the reason, the designa. tion or concept "Simsapa". From its absence, therefore, follows the absence of the latter necessarily. Since it is inconceivable, that the property (bhava) with the designation Simsapl" should exist without its own essence which is designated as "tree", "for it is only a particular (reality) with branches etc. that is known (by) such a name, i, e. Simsapa')." The reason for this impossibility, then, is that the very essence which is call. ed "tree" is the property (bhava) that is called "Simsapa". Or, in other words, that the two designations or concepts are essential properties (svabhava) of the same reality or essence (svabhava). The limiting eva, thus, serves to emphasize that the designationproperty "Simsapa" is limited to the reality "tree": There is no "Simsapa" that is not a "tree". M.'s idea, further, that the connection (sambandha) between bhdua and swabhava has two directions in agreement with which an anubandha and a pratibandha have to be differentiated, cannot be supported by any statements from Dharmakirli or the com mentaries. The word anubandha occurs in our context only in attributive positions / 11 The "definition of PV I 21-d (=4c-d) is repeated (cf..... brrer Mad balas te (PV T 6643] = ...... parrokiam anwedati (PVSVT 107, 24]) in PV 1 39ab (=41ab), just before the digression on apola. In addition we have the description of the concomitance in difference (wyatirka) of PV 1 23e-c|(25a-c). Since M does not translate the crucial terms bhava and rabhawe, both of which are used in their capacity to denote different concepts, I do not really know, however, how to understand his translation 12 Cl. PVT 1508 yod som den brel pa cak (D12b4:am P) yi gian thig god pa tam dan brel pa con gyi modo mi gran ishiga te/ - PVSVT 29, 136. : bhavandir nurodhini hatuadhamu matranurodhimi bharo Artur cf. my "Wirklichkeit und Begriff" (note 4). 205 and note 97 for a translation of this definition's repetition in PV I 39-b (-418-b). 13 PVSV 4, 2. madamyum ky artharya lanmdir rodiny v ...... 14 PVSV 24,11-15. 15 The example was introduced PVSV 16, 301 16 Cl. PVT 4722-5(= PVSVT 74.27-78, 14; cf. also PVinT 34306-16). 17 PVSV 16, 3011 lakhadi madriejasyaiva karyani tathaprasiddhed. 18 Cf. Matsumoto 495 1 9 loc. cit.4961, as explained above (dl. pf ). 20 And it synonyms which are either substantives in second position of a bahuerthi, or possessive adjectives, as is clear from the following examples: blauemadurodhin (PV 12d (= 4d). PV 1 395 (-41b). PVSV 24, 11, 14), Lanmadawrodhim (PVSV 4. 2) Landtrawhandhin (PVSV 6. 26), Lanmarasambandhu (PV 1930 [=25a]), tadbhatumdirnurodhin (PV in Il 58ab), studsatta Aidiraldvin (NB II 15), shunadharnabhavamdir amubandho- (HB4, 4). Sadhanadharmabhara maranayi (HB 5, 10), sadhundharmawdiranya (VN 9, 51.). Matsumoto's reference to PVSV 6,261. (note 11) is, in addition, useless for his purpose, because rabhave bhatarya there means the wence of the causal complex (hava - herundmagt), d. "Wirklichkeit und Begriff" 185.

Page #5
E Sebeller Svabhavaratibandha again . IIL and indicates, as such, the concomitance of the property to be proven (sadhyadharma). The word pratibandha, however, is not a term for that same relation, only with another direction, but for that relation which is the real basis of any concomitance. The preposition prati-, here, has the meaning towards, tear_to" and can be considered as being translated by "con-" in M's rendering connection". The use of prati-has the purpose only to emphasize the closeness of the connection, if it has any pratibandha in this function, i. e, meaning the real basis of the logical nexus, is synonymous with sambandha. Because of the originality and methodological importance of Matsumoto's attempt to deduce the correct meaning of the term from contextual evidence it was necessary to discuss his results in detail. Since his attempt was not successful, I think, and since I do not see another possibility to deduce Dharmak Irti's intention with the term in a similar way, there is only one resort left: the old commentators. Dharmakirti's concept of the term vabhava pratibandha has not changed, in my opin ion, since he propounded a theory of the logical nexus and its basis in his first work In this work the term occurs in the following places: PVSV 2,191; 3,4; 10,24; 17, 21; 17,7; 17,12; 53,27; 147, 11. "Old"commentaries on these texts are the PramAnavarttika tika (PVT) of Sikyamati (ca. 660-720 A.D.), the Pramkoavarttika (sva) vilka (PVSVT) by Karpakagomin (around 800 A. D), and for the "parallel texts" of the Pramanaviniscaya- the Pramimaviniscayafika (PVin T) of Dharmottara (ca. 750-810 A. D.). The following table gives the relevant passages in the commentaries. The sign of equa tion means that Karpakagomin copied his text from Sakyamati's; il negated in means that he did not copy from Sakyamati. In brackets I add the Sanskrit ending for the first part of the compound when either available from the grammatical analysis in the commentary, or to be assumed salely in the case of the Tibetan translations PVSV PVT PVSVT PVint PV in 11 1: 2. 191. 12a8 23, 18 235b3(ena) 10,14 12481 () 23, 181("ena) -(1326) --(24, 25) 237b8 11,6 2: 10,34 33a81. -57, 171.(+'ena) 324b5 40,10 3: 17,21. 47a6l.(*) 75,141.(.) 34365(e) 45,326. 17,7 472 75,24 344a2 46,6 4: 17.12 47b51.(e) =76,9(e) 344a7(sya) 46,17 5: 53,27 142a1 = 218, 26('sya) 6: 147, 101 400a41. (sya) =528, 171 21 Ch., . , HB 41, 8: anulandho 'nugamanam vyrui 22 W taken with the meaning against counter". Wat handla has to be translated as "obstacle" since it connotes a binding, fixation, which is against". Thus, M's "counterconnection" (496) cannot be the meaning of the word, because this translation translates these two concepts that can alternatively be meant by the one word "all" at the same time 23 The Tibetans, & , translate pratibandha just by 'bral pa 24 The Tibetan translation of the compound is of me help for a decision on the kind of at purus intended, because a) it has been made in all probability by Subhotinanti) and dGe b'i bo gros, the team that also translated the Karikh and Devendrabuddhi's commentary on the other chapters in the 11th century. Thus, whatever their interpretation was, it would be in accordance with a late exegetical tradition b) The majority of instances (PVSV 20, 24 PVSV, 411al: 17, 2-414be; 17, 7 = 415a1; 17, 12-41543; 53, 2744122) shows a literal translation by compound Gran lain 'bel pa) which I am tempted to call a compound of embarrassment.i. e, not knowing-and not needing to know-how to analyse the compound cctly, it is taken--and certainly with good right toe well known to need an exact translation with particle determinating the relation between the compound's members Sincerele is constructed with the particle den, naturally this particle would have to be supplemented in the first place c) In fact, there are two instances (PVSV 2, 191PVSV, 40564; 3, 31, 40567) that show translation with this writive particle danie rar in der Orel The translators, thus have chosen the particle naturally ruled by the soun d to this choice, however, does not necessarily indicate that the compound has been interpreted an instrumental-tatpurusa strictly peaking, kative a r e being psible as well. It only ways that an interpretation as an instrumental a pure cannot be ruled out for certain d) As to the genitive particle to be found in PVSV, 5103(PVSV 14.11). of below . H 25 This is the original form of the present Pramagavartike, first chapter, together with its "commentary"; d. E. Fruwallner: Die Reihenfolge und Entstehung der Werke Dharmakirti's Asiatica, Fritschrift Friedrich Weller. Leipeg 1954, 142ff. 26 Although the Tibetan translation of patibond he chosen here, vitrag u pe shows that the interpreters emphasised the meaning of dependence", which is definitely the meaning of the word in PVSV 147, 3 where this pagal related digree begins, the meaning of the compound is the mess in the other places wherebral is the usual translation 27 On his date and the relationship of his commentary to that of Sakyamatid my peper Mielien sur arkenntnistheoretischlagschen Schule des Baddhismus 1: Zur Datierung Karnaka gamine WZKS 23, 1979, 141-150. cl. o A. Alamatan : Kamakagomin and Santara kita Indological Review 3, 1981, 53-58. a l

Page #6
466 E. Steinkellner From among these texts the explanations given on PVSV 2,19f. are most suitable for an attempt to understand the variations extant in the analysis of the compound. The wavy line in the following texts indicates the words from PVSV, and a straight line. indicates the explanatory analytical phrasing of the first part of the compound (sva bhava-) as far as of relevance to our problem. 1: PVSV 2,19f.: svabhavapratibandhe hi saty artho 'rtham na vyabhicarati. la: PVT 12a8f.: yan ci ste [D: cis te P) dnos po bsgrub pa ghis kho na yin te naran bin dan 'brel po zes bya ba la sogs pa smos te ran biin bsgrub par bya ba'i dnos po dan 'brel pa ni de la rag las pa Aid de de yod na'o 28 1b: PVSVT 23, 181.: kim punar dvav eva vastusadhanav ity aha (svabhavaprati bandha iti svabhavena pratibandhah, sadhanam krteti samasah. svabhavena pratibaddhatvam, pratibaddhasvabhavam iti yavat. tasmin saty...... 1c: PVinT 235b3: ran bzin 'brel pa ni ran biin gyis 'brel pa stel 'brel pa'i ran btin zes bya ba'i don to // de yod na..... 31 Id: NBT 110, 1f.: svabhavapratibandha iti. svabhavena pratibandhah, sadhanam kyteti samasah, svabhavapratibaddhatvam pratibaddhasvabhavatvam ity arth ab. le: NBT, 62a2f.: ran bein 'brel pa tes smos te ran biin gyis 'brel pa yin te Tai bzin 'brel pa fid ni 'brel pa'i ran bein Rid ces bya ba'i don te While Sakyamati (text 1a) interprets the compound clearly as a locative-tatpurusa, Karnakagomin (text 1b) does not follow this interpretation. Instead he interpretes it as an instrumentel-tatpurusa. There is no doubt that Karpakagomin has copied Sakyamati's explanation in the context; thus the way the text of the PVT was used by Karnakagomin for this particular piece of explanation is quite telling: While the context and the introduction are faithfully copied from PVT, the analysis of the compound itself is taken 30 2 28 The introduction of Karnakagomin's explanation was copied from Sakyamati, for the texts immediatly preceding and following have also been copied (PVSVT 22, 10-22; 236=PVT 11b7-128; 12b1); but the pratika was lost. 29 Katyayana's Varttika 2 on Pan 2. 1. 33; it also occurs as a sutra in Sakatiyana 2. 1. 37 (ed. Bombay 1907, 121). 30 The following vabhavapratibandhah of the extant editions must be deleted (as in Ms. C of Malvania's edition). 31 This explanation of the compound with reference to fatyayana's Varttika has no correspondence in the Tibetan translation (NBT.) but must be considered as being supported by the parallel in PVSVT. 32 As indicated first by Matsumoto, loc. cit. 498. Svabhavapratibandha again from another source whose interpretation was evidently favoured by Karpakagomin. This interpretation is Dharmottara's analysis as an instrumental-tatpurusa. The rest of Karpa kagomin's explanation has been taken, in fact, from Dharmottara's Nyayabindutika (Text 1d). Dharmottara's commentary (Text 1c) on the parallel-text in PVin II is not likely to be the source for two reasons: Although Dharmottara analyses the compound as an instrumental-tatpurusa in the PVinT, too, the remaining paraphrase is still limited, less redundant, and above all-it is only in NBT 110, 2ff. that Dharmottara gives a wellformulated reason for his analysis as an instrumental-tatpurusa, i. e. that it has the advantage of expressing the connection in both kinds of logical reasons by means of a single compound. To summarize: Sakyamati explains the compound as a locative-tatpurusa, svabhava to be the sadhyavastu, and "connection with to mean "dependance upon.". Karnakagomin substitutes for this Dharmottara's interpretation as an instrumental-tatpurusa, which Dharmottara stated first-possibly in his PVinT, and then repeated with the addition of a clear argument for the value of this new interpretation in the NBT. The decisive difference between the two kinds of explanation seems to be indicated by the fact that only Sakyamati identifies svabhava as sadhya. This account for the material situation can be corroborated by the limited evidence coming from the explanations of PVSV 10,7: tayoh kascit svabhavapratibandho 'py estavyah. 2: PVSV 10, 2a: PVT 33a8f.: bigrub par bya ba dan sgrub pa de gnis de'i bdag Aid dan/ de las byun ba'i mtshan flid kyis 'brel pa 'ga' tig kyan 'dod par bya dgos te 2b: PVSVT 57, 171.: tayoh sadhyasadhanayoh kalcit svabhavena pratibandhas tadatmyatadut pattilaksano 'py estavyah. 2c: PVinT 324b5f. de ghis ran bein 'brel pa'i 'brel pa nid 'dod par bya dgos te/ 467 While Sakyamati (text 2a) gives no analysis of the compound, not repeating the first member at all, but instead inserting the terms for the two kinds of the connection, Karpaka " " 33 Explaining NB II 19: svabhavapratibandhe hi saty artho 'rtham gamayet 34 NBT 110, 2-4 karane svabhave ca sadhye svabhavena pratibandhah karyasvabhavayor avisista ity ekena samasena dvayor api samgrahah. Cf. also DhPr 110, 23f. where in continuation of DhPr 110, 17ff. the originality of this explanation is distinctly accentuated. Cf. also below. pTI 35 For Karnakagomin's using Dharmottara's PVinT cf. my paper of note 27, MESB 1, note 23. Karnakagomin's dependance on the NBT in case of the text under examination is, to may knowledge, the only instance noted so far to establish the relation between these two texts 36 Cf. below. p

Page #7
E Steinkellner gomin (text 2b)-who copies the PVT here - adds rabhauena. He, thus, emphasises again his preference for the interpretation as an instrumental taperusa; and it is probably his own addition, since PVinT (text 2e) on the parallel text of PVin Il gives no analysis of the compound at all. It must be underlined that rahduena of text 2b explain the connection between the two, Idd Aya and sadhana, and cannot mean one of these properties themselves. And this calls our attention to an important implication of this formulation of Dharmakiri: Since it is logically impossible to say that "between two items there is a connection of/with one of these two", rabhas must be taken as an attributive explanation of pratibandha. In the light of this, Stkyamati's substitution of vabhava by the term for the two kinds of the connection in an instrumental (ta. datmyatadut parrilaksana) must be considered as being motivated by the same te sons. A look at the explanations of PVSV 17,21. brings further clarity: 3. PVBY 17, 2f.: iti tasya sabhava pratiband had avyabhicara 30 : PVT 47a6f.: gran tskigt di'i phyir igru pa din sa pa la sops po bdag Mid du gyur pa de ni rant btin 'bral pa kho na'i phyir terar din begreb par bya bar 'dod pa si la sog: pa dant i skad blad pa'i ruam pas 'brel pa tho nas khrul pa med doll 3b: PVSVT 75,141.: ili hatos tasyai mabadtasya sadhanaiya lima padeh sabhave iddiyabhimate urajadaw yatholtena prakarana pratibandhad rudoyab Nicarab. 3c: PVinT 34315: des na si fa pe Aid kyi no bo ldog pur dyed pa (?) bigrub par bya ba'i no bo la 'brel pa yin la/ 'brel pal payir na "hrul ba med doll These texts analyse the compound a locative-latpuruse, and they all explain the lini member, svabhava, to mean the addhya. This, evidently, is the reason why Dharmo tiara and Karpakagomin, too, do not deviate from Sakyamati here: the first member of the compound does not explain the pratibandha. 4: Pvsy 17,12: tasmat svabhavapratibandhad eva hetuh sdhyamamayati. 4a: PVT (751.: gan gi payir de Itayin pa de'i phyir rai bin 'bral pa Aho nas bagrub par bya bar 'dod pa'i dilos po dal 'brel pa kho nas 4b. PVSVT 76,9: yata ruan farmde rabhasa pratibandhad evo sadhyabhimate Dostui pratibeddharudd .. 4c: PVinT 3467: de'i phyir rar blin 'brel pa kho nas iu bya ba ni rantingyi 'bral pe yin pa'i phyir. The explanation of the compound locative-latpurusa is coupled again with that of mathavaas meaning addhya in Saikyamati's (text 4) and Karpakagomin's paraph Tame (text (b) which is copied from the former. The analysis as genitive-latpurusa by Svabhavaratibandha again Dharmottara (text (c)- if it is not a corruption (gyi for gyis) - could mean nothing other than Dharmottara takes svabhava to mean the sadhana instead of the sad Ayo This analysis as a genitive-tatpurusa is also to be found in the explanations of PVSV 53, 26. and 147, 101. (cl. texts Sb, 6a). 5: Pay 53, 261: tadu pairidharma bhavah svabhava pratibandhad apekate ndma... Sa: PVT 142a1: ran bein 'brel cixi rag les pa'i phyar phan par byed pa blios pa tes by ste sb: PVSVT 218, 26: sabhavasya pratibandhad dyartarudd apekate nama wa karinam. 6: PVSV 147, 10f.: na ca tajjanmalaksapat svabhava pratiband had anyab prati bandho nama. 6a: PVT 400941, skye ba'mishan Hid can te skye ba'i dba po can kyi Tak bein gyi rag las pa las rag lus pa tes bya ba gian god pa ma yin noll 6b: PVSVT 528, 171, no hi janmalaksandi janmasvabhdudi sabhava prati. bandhad anyab pratibandho nama Both passages differ from those discussed so far, in that they do not talk about sa bhava pratibandha in the context of logic, as the basis for the necessary nexus between two concepts. Rather, more generally, they refer to the seabhavapraribandha-idea as the only reason for dependance. In the first statement, which serves a discussion of the sd mdnya (text 5), dependance (apeksa) is said to be based on the sabhava prati bandha; and the second statement (text 6), serving a discussion of the pudgala, rules out all connections except for the sabhava pratibandha. In both cases only the causal connection is intended. Here, the first member of the compound, svabhava, does not mean the sadAyo, of coure, but the actual essence of something. Since the term, as used here, means the whole entity, without any abstract properties being taken into considera tion in the context, the analysis as a tatpurusa with an objective genitive connection of the essence" is a natural explanation. But this genitive-tatpurusa is not the same as that of text 4c because of the different connotation of the term Jabhasa. Thus, strictly speaking because the intention of the term rabhasa pratibandha is limited here to causality only the analysis as a genitive-tatpurusa as in the case of these two oscurances does not contribute directly to an understanding of the term's meaning as the real basis of the necessary logical relation The conclusion to be drawn from this material is evident: The different analyses of the compound as instrumental, genitive, and locative-tatpurusa result from the respect ive interpretation of the meaning of the compound's first member. If sabhava was

Page #8
E Sizinkiler Svabhavapratibandha sain modal. Modern interpreters show some fluctuation between these two possibilities. I have found no satement of Dharmakirti so far which allows a safe exclusion of one of them. Considering the "logie" of the expression, however, it is difficult to find a clear line of Meparation between the two connotations. After all, if something is connected "by its essence", it is connected essentially" because this kind of cause" is contemporaneous Since even from a text looking as modally as Durvekamista's yah arepeya kvacid ayattas tanya rabhavan tatra pratibaddha dyatta iti (DhPr 110,211) we cannot deduce definite exclusion of a causal translation, I think that both translations are acceptable, while the model one, with an adverb, is usually more practicable. IV. understood to mean radhya, the property to be proven, the compound is analysed locative tatpurusa; this is the case in texts 1a, 3a, 3b, 3c, 4a, 4b. And if taken to mean sadhana, the proving property (possibly in text 4e), it would be analysed as a genitive tatpurus. In texts lb, le, id and 2b, however, wabhave does not mean one of the properties, but the essence, nature, thing itself. In these cases the compound is analysed as an instrumental tatpurusa. Here it is particularly instructive to note that Sakyamati's interpretation as a locative tatpurusa (text la) which is combined with an interpretation of sabhava as sadaya has been substituted by Karpakagomin with Dharmottara's inter pretation as an instrumental tatpurusa while at the same time dropping the interpretation of swabhava as sadhya from the copied text. Since the material used in this investigation was strictly limited to the few securances of the term rabhava pratibandha in the PVSV and the early exegetical tradition, the proposal of the following working-hypothesis may be useful to enlarge the meaning of our conclusions with a few to further research: Before decision is made on the analy. sis of the compound, the meaning of rabhava must be determined. It either means one of the logical properties (dharma) or the essence, the real thing (muarepa, diman). In the fini care mainly the sadhyadharma will be meant, but generally speaking there is an analytical alternative. The compound can be explained, then, either as a genitive, or as a locative-tatpurusa with reference to the fact that the two cases represent the linguistic way to express the two different ends of the relation in question: The connection pratibandha) is one of that which is connected in our case usually the logical reason (het) - with that with which it is connected. In other words, we may assume that an analysis of the compound as a genitive tatpurua is given with the meaning "connection of the abhave, ie the hetu", and an analysis as a locativetatpurusa with the meaning "connection with the suabhava, i.e. the sadhya". Naturally we can find statements where both ends of the connection are indicated as, e. & in PVSV 17.21. (d. texts 3 and 3a, 3b). I wablada means the essence, an analysis as instrumental-atpurusa is the correct one. That it is taken with this meaning in the compound already by Dharmakirti is clear because of the implications of his words PVSV 10,24 (text 2). It seems to be Dharmottar, however, who takes this interpretation to be the correct one also in at least one other crucial instance of the term (cl. texts 1c, id) where it was not applied previously (cf. text la). One question remains to be answered with regard to this analysis an instrumental Latpurusa : What kind of instrumental? In our case it can only be taken as causal or as It has been assumed above that Dharmottara found it better to interprete the term svabhava pratibandha as an instrumental tatpurusa (texts 1c, id) at its surance most consequential for an understanding of the theory of the logical nexus, Le PVin II 10, 14 and NB II 19. In NBT 110, 2-4 Dharmottara adds a sentence to his analysis of the compound with a view to showing the value of this interpretation : karane madhave ca sadhye vabhavana pratibandhah karyasvabhavayor avidisaity skena sam asena dvayor api samgrahah. This sentence is important for two reasons: It gives clear survey of Dharmottara's interpretation of the case relations involved, and it allows the consideration of possible historical implications when he concludes "Thus [the connections of both of these (reasons) are summarily stated by one and the same compound." For, if we follow Durvekamisra's comments (DhP: 110, 17-19 and 231.), the phrasing kena ramdiena means some emphasis on this interpretation as bringing an advantage Carilaya) over another, older one which explained the compound as a genetive-tatpurusa in case of the karyaher, and as locative tatpurus in case of the wabhavahetu, 38 E Th Stcherbatsky in his Buddhist Logic 1. 69 gives a model translation dependent in its own existence", and 69, note 7 causal one: "being tied up by one's own existence. Mookerje Negaki in their translation of 1964 translate by various attributes (15: "real and natural relation". 16: "natural relation". 44, 67: "ecary concomitance", 66: "natural concomitance") which is based on model interpretation. I myself and to translate causally Verknupfung durch das Wer", but now peeler the medial translation "Wesentliche Verknupf ung. sential connection". 39 d. p . 40 Le connection by eeeeentially (rabharna) of proving cantial property naha mary) with an ential property to be proven (nawe Aadhye). H467 37 Cl. above

Page #9
472 E. Steinkellner 73 thus refering actually to two different compounds. Moreover Durvekamisra's phrasing (especially in DhP: 110,238.) clearly reveals that he considered this new interpretation to be Dharmottara's achievement. It must be noted, however, that Durvekamisra is the only one known so far to know of this interpretation which relates the compound to the two kinds of reasons respectively, Dharmollara's statement as such does not need to be taken other than as an explanation that the spabhava pratibandha is the same for both kinds of reasons. Beyond that, since the various explanations of the compound discussed above all agree on the fact that this would be the same in each case for both kinds of reasons, and since I tried to explain why Matsumoto's proposal for a solution similar to the one possibly indicated here cannot be accepted, I have to admit that I still do not know whose interpretation could be considered as having been improved by Dharmottara. We, also, cannot exclude the possibility that Durvekamisra offers a piece of over interpreta. tion here. Svabhavaraibandha again [connection) follows from the fact that the reason) is the self / essence of the resultJ"), he introduces the objection: tadatmatne sadhyasadhanabhedabhava iti cet (PVSV 2,21) (="If [the reason is the self essence of the result). there is no difference be. tween reason (sadhana] and result [sadhya]"), and he answers: na, dharmabheda pari. kalpanad iti vaksydmah (=[This) is not the consequence), because of the assump. tion of different properties, as we shall explain.") We have to conclude, then, that beyond the explanation of the difference between the Teason and the result as properties, Dharmakirti's expression tadatmatvat must be taken to mean that the reason is in reality the same as the result. And this is what I translate as "real identity" VI. That the terms tadatmya (PVSV 4.2) or tadatmata (PVSV 2,21) are synonyms of tadbhavata - and not terms for M.'s "two directions of the connection in logical context (495), is evident from PVSV 17, 13: sa ca tadbhavalaksanas tadut patti laksayo vd. The term denotes that real connection, connection in reality, which provides the basis of the logical nexus between an essential property as reason (svabhavahetu) and the respective essential property to be proven (sadhya). The best support for the translation of tadatmya etc. by "real identity" or, short, identity" are those sentences where Dharmakirti lays the foundation for the need to deal with the apo ha theory. The beginning of the first passage (PVSV 2.2/-3.3) is suffi. / cient for our purpose here. When Dharmakirti says: sa ca tadatmarval ("And this In note 12 of his paper Matsumoto, finally, does not accept my interpretation of the word tadarman as a tatpurusa, "but rather Dharmottara's interpretation that the word is a bahuvrihi-compound" (494). My interpretation is not based on NBT 106, 61. and HE. T 57, 61. as proposed by M., since these passages are, indeed, useless in this connection. Lacking any passage where a decision could be derived from Dharmak Irti himself, I had to base such an interpretation on the commentaries on passages like the three mentioned above, where the compound tada man or its synonym tadbhava. occurs. The questions remaining, then, are: What are the explanations of the commentators? Why do we find two different explanations And why did Dharmakirti not give a clarification of the exact meaning of the compound? Explaining radaimatvat of PVSV 2.2 the Tibetan translation of Sakyamati's PVT takes it as a genitive-latpurus. The same is the case when tadatmatve of this line is explained, and to this second explanation Sakyamati adds: bsgrup par bya ba gan 41 It does not seem to be Vinitadeva's, whose paraphrase on the beginning of NB II 19 I do not really understand, however : ran bin dan 'brel pe dan Iran din eyi de hos 'hrel Murant bin gyi 'brel Auste (NBTI 64, 61.). The instrumental mobos (if it is not a corrupath for me boi) could indicate that the interpretation as a modal instrumental at purusa was already thought of belore Dharmottare, perhaps without particular argument for it such as the one given by Larmottare. 42 The second page introduces the apoka digression: ya ebu tarhi krtelah seranitgo bheda kaudt. pretij Marthaikadele helw sya misa dosah. yama..... (PVSV 24, 361.). And the third concludea this digression Irna rabhavasyai na Mdhyasadhanabhawe 'n addhyaradhanalaisargah fax nafrati Marthaikadele hatur iti (PVSV 93, 3-5). 43 For Dharmakirti's concept of bheda c. PVSV 20, 211. 44 I have to admit that in writing note 2 of my paper : On the Interpretation of the sabhavahetub. WZKS 13. 1974. 117-129. I should have refefed to Dharmottara's explanation. As far as I can remember, I did not think of it, then Rather I thought only of the "wal way of translating it, not making myself sufficiently aware of the fact that this translation must be traced to Sicherbatsky's knowledge of Dharmotara's interpretation 45 PVT 1262 3: de sites byu ba ni raxi brin dar 'brel pa'oll de'i bdag Mid yin pa'i byir te bya ba ni bsgrub per bya ba'i rak bin yin pa'i phyir // PVSVT 23, 211: sa ca trabhasaprat bandhak...adatmatudd iti sadhyaswabhavalvde. 46 PVT 1213-4: de'i bdag nid yin na de bya ba ni gul te igrul A begreb par by he's bdag Kid frin na=PVSVT 32, 22f.: adatmalde adhyavabhaudematee sadhanaya.

Page #10
474 yin pa de nid sgrub pa yin pas (PVT 12b3f.) yad eva sadhyan tad eva sadhanam iti (PVSVT 23, 221.). In the case of the commentaries on tadbhavalaksana- from PVSV 17, 13 it is again only the Tibetan translation of Sakyamati's PVT which supports the E. Steinkellner interpretation as a genitive-tatpurusa. 50 The same result can be gained from a survey of the relevant passages of the Pramanaviniscaya. It is particularly interesting to note that Dharmottara in his explanation of these passages analyses tadatmatva (PVin II 10, 15) as a tatpurusa, as well as tadbhava (PVin II 7,25; 46, 18f., 33), while he seems to analyse tadatmya (PVin II 24, 12) as a bahuvrlhi. In NBT on NB 11 22 he also explains it as a bahuvrihi: sa sadhyo 'rtha atma svabhavo yasya tat tadaima. tasya bhavas tadatmyam (NBT 113, 3.), as pointed out by Matsumoto (note 12). Vinitadeva explains it as a tatpurusa The crucial statement in the Hetubindu is HB 4, 31. sadhyadharmasya vastu tas tadbhavataya sadhanadharmabhavamatranubandhasiddhi ("is the proof that the property to be proven follows only the presence of the proving property because in reality [the property to be proven] is the essence [bhava] of that [proving property]."). The compound tadbhava-, here, is explained as a tatpurusa by Arcata. 54 47 The explanation of radarnya from PVSV 4, 2 is too short (cf. PVT 15bif., PVSVT 29, 15). 48 PVT 47b6f: dei no bo'i misham id ces bya ba la sogs pa'i bagrub par bya ba'i ran beingxi no bo'i mishan Aid=PVSVT 76, 11f.: Ladbhavalaksana iti sadhyasvabhavalaksanah---- Additional support is given by the short paraphrases of tadbhava- from PV 1 27a (=29) and a number of passages in the Vrtti on vv. 27-28 (-29-30). 49 Cf. PVin II 24, 12 (tadaimya) and 10, 15 (tadatmatva) 7, 25; 46, 18f. and 33 (tadbhava). 50 Cf. PVinT 235b6: ga gi phyir sgrub par byed pa de bigrub par bya ba'i ran bain Aid yin pa des na***** 51 Cf. PVinT 224b6f., 344a7f., 345a7. 52 Cl. PVinT 282b4: bigrub par bya ba'i chos de'i bdag Rid gan yin pa'i sgrub par byed pai chos de l'oll 53 NBT, 63a2 bigrub par bya ba'i don de'i bdag Aid gak yin pa de'i dos po ni de nid de l Cf. also the formulation of NBT 162, 12 which corroborates an interpretation as bahuvrihi: yo hi sadhyadharmah sadhanadharmamatranubandhavan, sa eva tasya sadhamadharmasya sabharo nanya And of NBT 162, 15f.: tasmat sa eva sadhyah kartavyah (sic!) yuh sadhanarya subhavah sya 54 NBTi 65, 1f.: rtags 'ga tig bagrub par bya ba'i don gyi ran bein yin pa de'i phyir 55 HBT 41, 15-17: tadbhavataya sa sadhanadharme bhavah svabhave yarya tasya bhavalaya tadbhavataya yo hi sadhanadharmah sadhyadharmasya sabhavah so hatham tam namu badhatyat ("This, [ie] the proving property; the existence [bhava.] [. e.] the essence [rabhava] of which, because it is the existence-essence [bhava] of that. For how could the proving property, being the essence [rvabhava] of the property to be proven, not follow this [property to be proven] ?"") Svabhavapratibandha again But on another comparable occasion Arcata prefers to interprete tadatman of HB 8, 13 as a bahuvrihi. 475 Thus we find two interpretations of the compounds tadatman-, tadbhava-: as a tatpurusa by Sakyamati, Karpakagomin, Arcata and Dharmottara (PVinT), and as a bahuvrihi by Arcata and Dharmottara (NBT and PVinT). I have to admit that I could not find a statement in these commentaries that would provide some information as to a distinct reason for such a twofold explanation of the compound, nor can I think of one. We have to keep in mind, however, that the meaning of the compound within the context of Dharmakirti's theorem of the svabhavapratibandha remains the same, whether it is taken as a tatpurusa or as a bahuvrihi. For, interpreted as a tatpurusa, it means that the concept of the proving property (sadhanadharma, hetu) is, in reality, the essence (atman, bhava) of the property to be proven (sadhyadharma). And taken as a bahuvrlhi, it means that the concept of the proving property has the property to be proven, in reality, as its essence(atman, bhava). The difference between the two solutions is, that in the case of the tatpurusa-solution the concept of the hetu is introduced as being reduced to its reality, thus serving as the real essence of the concept of the sadhya; and in the case of the bahuvrthi-solution these positions are reversed. Since the function of the compound within the contextual theorem remains unchanged, it is possible that there was no awareness in the commentators, e. g. Arcata and Dharmottara, that it needed to be dissolved in one particular way only. It seems that this is one of those cases where the modern scholar has a problem that was none to the tradition under examination, and where, therefore, he looks for an answer that was never given. With regard to the term tadatmya we can observe that the earlier explanations of the compound fadatman take it as a tatpurusa, while-possibly starting with Arcata-it is then also understood as a bahuvrihi. In both cases of explanation, however, the term tadatmya indicates the fact, that one property is in reality the same as the other pro 56 Cf. HBT 83, 5: tadamano yo yasya svabhavah tatsvabhavaya..... 57 HB 5, 12 does not support, I think, either interpretation. I do not accept, however, Matsumoto's argument against my construction of the expression lingivabhava. Arcata (HBT 57, 11ff.) does not express himself regarding the nature of the compound lingisvabhava, from HB 5, 11f we can only understand, that Dharmakirti wants to emphasise that the reason is vastuto lingine bhava, where rvabharu definitely has the meaning "essence" (cf. HBT 57, 13f.). rabhava of this sentence also cannot be connected with HB 5, 14f, because there Dharmakirti turns against the assumption that even a property that is conditioned by something else, and non-concomitant can be considered as an essential property (rabhava) (cf. HB II 102f., note 7).

Page #11
476 E. Steinkellner perty, and the rendering of the term as "real identity" or just "identity" is appropriate. Methodically speaking, it is a legitimate and desirable procedure to try to understand a theorem by basing oneself on Dharmakirti's work alone. It is sometimes- e. g. where we suspect a development of a theory in Dharmakirti's works-the only approach possible, as I shall show on another occasion. But when the commentators differ from the results deduced out of Dharmakirti's statements alone, we have to give, in addition, a convincing explanation for that difference/ If we do not, or cannot account for this difference of interpretation in a historically reasonable way, our interpretation of Dharmakirti's statements may only be a mis-interpretation and certainly needs to be re-examined.