Professional Documents
Culture Documents
An Annotated Translation of Prasannapada 55.11 To 75.13
An Annotated Translation of Prasannapada 55.11 To 75.13
13
DAN ARNOLD
Introduction1 The Prasannapada is an important but sometimes frustrating text. Nagarjunas Mulamadhyamakakarika (MMK) is available to us today in the original Sanskrit only as embedded in this commentary by Candrakirti (fl. 600 CE), which is the only commentary on Nagarjunas text known to have survived in the original Sanskrit2. But Candrakirti himself seems to have had little influence on the subsequent course of Indian philosophy; it was, rather, almost invariably the works of Dharmakirti and his philosophical heirs that were taken up by later Indian philosophers (Brahmanical and Buddhist alike). Despite that fact, Candrakirtis influence on the larger reception of Indian Madhyamaka has been considerable, owing to his having been judged by most Tibetan traditions of interpretation to represent the definitive interpretation of Madhyamaka philosophy, which is almost unanimously claimed by Tibetans to represent the pinnacle of Buddhist thought. This fact itself is striking, not only because of Candrakirtis negligible influence in the Indian context, but because the scholastic traditions of Buddhist philosophy were directly introduced to Tibet by Santarakita (725-788) and Kamalasila (740-795) whose
My work on this text has benefited, over the years, from conversations and study with several people. I would like to thank, in particular, Larry McCrea (with whom I read through the whole text) and Shelly Pollock for discussions concerning various aspects of the Sanskrit; and, for their generously detailed and thoughtful readings of one of the most recent drafts of this article, Rick Nance and Ulrich T. Kragh. I also benefited from the helpful comments of an anonymous reviewer. It should be noted, in addition, that my work on this text would not have been possible without the fine work that has been done on this and related passages by David Seyfort Ruegg, Mark Siderits, and Tom Tillemans. 2 Ruegg 1981:1, n.3. Journal of the International Association of Buddhist Studies Volume 28 Number 2 2005
1
412
DAN ARNOLD
thought, though surely affiliated with the Madhyamaka tradition in which Candrakirti stands, reflects the predominance of thinkers (like Dharmakirti) whose approach is (on Candrakirtis own view, at least) generally antithetical to Candrakirtis. I suspect that this historical fact at least partly explains the extent to which Tibetan interpreters claim Candrakirti as normative, while yet retaining much of the epistemological discourse that he so clearly rejected3. Not only, though, is the Prasannapada thus historically puzzling, it is also discursively rather odd. In places, Candrakirtis Sanskrit is extremely lucid, almost conversational; there are sections that the intermediate Sanskrit student can pick up and read with some confidence. He displays a great familiarity with the grammatical traditions of Brahmanical learning4 and also, as Karen Lang (2003) has emphasized particularly with respect to the first four chapters of the Catusatakaika, with the literature of Sanskrit stories and dharmasastras. These facts are as befits someone who claims, as Candrakirti characteristically does, always to defer to and exemplify conventional usage (lokavyavahara). And yet, as no less a scholar than the estimable J. W. de Jong observed, the first chapter, in particular, is difficult5. It seems to me that it is not always clear whether this is so chiefly because of Candrakirtis Sanskrit, or because the logic of the arguments is hard to follow. Like those of Nagarjuna, Candrakirtis arguments can seem at once pregnant with import, and maddeningly elusive and paradoxical. Candrakirtis examples of ordinary reasoning that is familiar in the world, for instance, are often counterintuitive, perhaps as much to his Indian readers as to the modern interpreter. The elusive character of the arguments is often reflected in some puzzles concerning the deep structure of the dialectic that is, even more than is typically the case with Sanskrit philosophical works, it is often a difficult question which voice, as it were, speaks each part of the argument. Needless to say, it makes a good deal of difference to ones sense of the argument whether one takes a particular point to be made in supOn the introduction of Candrakirtis thought to Tibet, see Lang 1992. See, e.g., Bhattacharya 1980, 1980-81. 5 Cf., de Jongs review (1981) of Sprung 1979, where he laments that Sprungs knowledge of Sanskrit was insufficient to the task of translating such a difficult text.
3 4
413
port of the position that Candrakirti is working to defend or instead to be precisely the point he means to attack. Alas, it is not always compellingly self-evident simply from the formal features of text which of these is the case, and some tricky questions bedevil the would-be translator of Candrakirtis work. While my familiarity with Tibetan traditions of interpretation is not great enough that I can comment authoritatively, it seems clear to me that the influence of these can be said particularly to inform much modern interpretation of Candrakirtis texts. A case in point is the svatantrikaprasanika division of Madhyamaka philosophy which, although not without basis in the antecedent Indian texts, represents a particularly doxographical lens imposed by Tibetans6. I would consider it an uncontroversial remark to say that David Seyfort Ruegg long one of the leading historians and interpreters of Indo-Tibetan Madhyamika literature has been influenced over the years by his significant engagement with certain Tibetan traditions of interpretation7. Recently, Ruegg has made another signal contribution to the study of Indo-Tibetan Madhyamaka; his modestly titled Two Prolegomena to Madhyamaka Philosophy (2002) comprises a nearly complete translation of chapter one of the Prasannapada, along with commentaries thereon by the seminal dGe-lugs-pa thinkers Tsong-kha-pa and rGyal-tshab-rje8. The first chapter of Candrakirtis text has long been recognized to be of par6 See Dreyfus and McClintock 2003. Cf., Rueggs comment (1981:51, ff.) on the likely Tibetan origins of these doxographical terms. 7 See, among his many works, Ruegg 1981 (which remains the definitive survey of the topic), 2000, and 2002 (which will be much engaged in the present article). For a critical assessment of some of the characteristically Tibetan interpretations generally upheld by Ruegg, see Oetke 2003a. 8 More precisely, Ruegg has translated that portion of Candrakirtis chapter that is framed as commenting specifically on the first verse of Nagarjunas root text; in the standard edition of Candrakirtis text (La Valle Poussin 1970a, as supplemented by de Jong 1978), that means pp.1-75, in a chapter of 91 pages. (The edition of Vaidya [1960] largely reproduces La Valle Poussins edition, and gives the pagination thereof.) The first chapter of the Prasannapada was translated into English by Stcherbatsky (1927), whose work, though dated and eccentric, remains useful. The partial translation of Sprung (1979) is the closest there is to a complete translation of the Prasannapada into English, but should be used with caution; cf., the reviews by de Jong (1981) and Steinkellner (1982). Other Western-language translations from the Prasannapada (e.g., May 1959, Schayer 1931) do not include the first chapter.
414
DAN ARNOLD
ticular importance, comprising as it does some of Candrakirtis most extensive and systematic engagement with what he took to be alternative understandings of Madhyamaka, and of Buddhist thought more generally. The first chapter of the Prasannapada has become a locus classicus, in particular, for what Tibetan traditions emphasized as the split between the Svatantrika and Prasangika schools of Madhyamaka. Perhaps following the emphasis of the Tibetan tradition, most contemporary scholars have been principally concerned to understand this aspect of Candrakirtis opening chapter9. What has less often been appreciated is that the first chapter of the Prasannapada also comprises a lengthy engagement with an unnamed interlocutor whose thought looks very much like that of Dignaga. In the standard edition of the Prasannapada, this section spans some twenty pages10. Despite its thus constituting fully a fifth of the chapter, this section has been little studied. This scholarly neglect perhaps owes something to the fact that some influential Tibetan discussions of at least parts of this section take Candrakirti to have been continuing his attack on Bhavaviveka, so that what can very well be read as an engagement with Dignagas epistemology gets subsumed in the svatantrika-prasangika discussion that has instead preoccupied most scholars.11 I would argue, though, that understanding Candrakirtis arguments here as an engagement specifically with the epistemology of Dignaga affords us an unusually good opportunity for appreciating the logically distinctive character of Candrakirtis Madhyamaka. It should be noted that I am here making a chiefly philosophical point, and that the historical question of Candrakirtis target is perhaps more complex. As a matter of intellectual history, the texts of Dignaga were variously circulated and appropriated, and it may be difficult (if not imposIn addition to the recent work of Dreyfus and McClintock, see Yotsuya 1999, which provides a useful text-critical analysis of the relevant passages from the original Indian sources of Candrakirti, Buddhapalita, and Bhavaviveka (or Bhaviveka though Candrakirti uses the former name). 10 La Valle Poussins 1970a:55.11-75.13. 11 For the view that Candrakirti is still addressing Bhavaviveka in at least part of this section of the text, cf., Thurman 1991:292-295, which translates a section of Tsong-khapas Legs bshad snying po based on a discussion occurring at Prasannapada 66.1-68.4. Cf., also, Eckel 1978, Huntington 2003, Yoshimizu 1996:49-94.
9
415
sible) to determine whether Candrakirti finally had Dignaga himself chiefly in mind, or Dignaga as appropriated by, say, Bhavaviveka. (It is, though, interesting in this regard that although Candrakirti recurrently names Bhavaviveka as the target of his critique and Buddhapalita as the thinker he defends in the sections of the text that constitute the locus classicus for the svatantrika-prasangika debate, his interlocutor in the section here translated goes unnamed.) From a philosophical perspective, there is a sense in which it may not finally matter whether it is particularly Dignaga whom Candrakirti had in mind, or whether he here targets that part of Bhavavivekas project that is informed by Dignaga; for in either case, Candrakirti can be said to have philosophical problems with Dignagas project. The logically distinctive character of Candrakirtis arguments, then, can in either case be appreciated by considering the arguments here developed vis--vis the arguments of Dignaga. There is, I am suggesting, some specifically philosophical value in appreciating particularly what it is about Dignagas approach that Candrakirti would refuse12. It is in the hope of facilitating the appreciation of this interesting exchange that I here offer a translation of this section of Candrakirtis Prasannapada, and that I will occasionally identify Candrakirtis interlocutor as Dignaga. The most detailed study of the passage here translated is the illuminating work of Mark Siderits (1981), who evinces an insightful grasp of the conceptual significance of the passages, even where his translations are problematic. Siderits in turn gets some help from Satkari Mookerjee (1957), whose work basically paraphrases Candrakirtis text; and from Masaaki Hattori, whose extensively annotated translation of the first chapter of Dignagas Pramaasamuccaya (1968) makes frequent reference to Candrakirti13. Since the publication of Sideritss article, our understanding of Candrakirtis principled objections to the foundationalist trajectory of Buddhist philosophy have been much advanced
12 Nevertheless, there are several points at which Candrakirti seems clearly to have had Dignagas text before him (many of which were noted by Hattori 1968), and these will be noted as we proceed. 13 Hattoris annotations thus represent a useful source for appreciating the likelihood that Candrakirtis interlocutor is Dignaga, with Hattori often pointing out where Candrakirtis engagement closely tracks Dignagas text.
416
DAN ARNOLD
by the work of Tom Tillemans, whose study and translation of Candrakirtis commentary on chapters 12 and 13 of Aryadevas Catusataka represent a philosophically sophisticated engagement with texts that closely parallel the arguments in our section of the Prasannapada14. The characterization of Candrakirtis critique of Dignagas foundationalism that is ventured by Georges Dreyfus (1997:451-60) chiefly follows the work of Siderits and Tillemans. While the aforementioned works are enormously helpful in philosophically situating Candrakirtis Madhyamaka vis--vis the foundationalism of Dignaga, it can still be said that there is a relative paucity of studies of this important text, and that close readings of Candrakirtis critique of Dignaga as that is developed in the Prasannapada remain a desideratum. I have developed a philosophical interpretation of this section elsewhere15. It is partly owing to the rather more speculative character of my other works (which, though interpreting his texts, amount to rational reconstructions of Candrakirtis arguments) that I here want to venture a more literal interpretation in the form of a translation. I also venture this, however, in light of the extent of occasionally significant divergence between my translation and that of Ruegg. To be sure, Rueggs translation which is informed by Rueggs particular appreciation for the Tibetan reception of Candrakirti is likely to become a standard reference for this portion of the Prasannapada. This is as it should be, since Rueggs translation is (as expected) generally quite reliable. It is also, however, not likely to be very accessible to non-Sanskritists. More significantly, there is an important sense in which Rueggs translation which occasionally deploys the kinds of locutions that Paul Griffiths memorably characterized as Buddhist hybrid English16 may undermine Candrakirtis own points; for insofar as Candrakirti finds it in principle important to defer to conventional usage, it becomes important to capture the naturalness of his Sanskrit. This is not achieved when, for example, an important expression like lokavyavahara (ordinary usage, though the
14 See especially Tillemans 1990: vol. 1, pp.41-53. Among more recent works by Tillemans, one can also usefully consult, inter alia, that of 2003. 15 Arnold (2005), Chapters 5-7; see also Arnold 2001, 2003. 16 Griffiths 1981; cf., also, Rueggs own perceptive remarks on translation (Ruegg 1992).
417
term also has an eminently mercantile connotation that I think is nicely captured by business as usual)17 is rendered, as it is by Ruegg, as transactional-pragmatic usage. Moreover, given the difficulty of some of the exchanges in the first chapter of the Prasannapada, it is to be expected that different judgments can be made about how to understand the text. Having carefully considered Rueggs translation of this section, I have judged that in the places where our readings diverge, mine are at least defensible, and in some cases (particularly towards the end of the passage) significant. Thus, in the hope of further advancing our understanding of a surprisingly neglected (and very interesting) philosophical exchange, I propose the following translation, which could, I think, quite profitably be used in consultation with that of Ruegg as well as with the work of Siderits, and with my own, more speculative interpretation. The following translation is intended to stick closely to Candrakirtis Sanskrit (though I have taken the liberty of inserting material in brackets where that is required to make the sense of the English more plain) though it is hoped that it will also be experienced as being in English. In my annotations to the translation, I have given the Sanskrit text (from the edition of La Valle Poussin), noting de Jongs proposed revisions as well as a few emendations of my own. Where I have found it useful to consult the Tibetan translation by sPa-tshab nyi-ma-grags (as available in the sDe-dge edition of the bsTan-gyur), I give the Tibetan as well. In addition, I have also provided in the annotations something of a commentary, briefly explaining what I take to be the salient points particularly of more complex passages. I have also noted significant indications of the interlocutors likely identity as Dignaga with some interesting clues to be found, in this regard, in Yoshiyasu Yonezawas recent edition of the *Lakaaika, a very brief commentary (really more like a students personal annotations) on the Prasannapada dating to probably the 12th century18.
See Apte 1992:1514: Affair, business, work profession, occupation dealing, transaction commerce, trade, traffic. See also Rhys Davids and Stede 1995, s.v. vohara. 18 So Yonezawa (2001:27). See also Yonezawa 1999, 2004.
17
418
DAN ARNOLD
In the following translation, numbers given in square brackets represent the page and line numbers of La Valle Poussins edition. For ease of use, the dialectical flow of the argument has been signaled by indicating the main changes of voice in bold type. Translation [55.11] At this point, some object: Is this certainty19 that existents are not produced20 based on a reliable warrant (pramaa),21 or is it not based on a reliable warrant? In this regard, if its accepted that its based on a reliable warrant, then you have to explain: which warrants, having what characteristics and what objects? Are [these reliable warrants] produced from themselves, or from something else, or both, or altogether without cause? On the other hand, if [your certainty] is not based on a reliable warrant, this doesnt make sense, since comprehension of a warrantable
Or conviction; see n.27, below. Candrakirtis interlocutor here refers to MMK 1.1 (which is the verse being commented on for most of chapter one), according to which There do not exist, anywhere at all, any existents whatsoever, produced either from themselves or from something else, either from both or altogether without cause (La Valle Poussin 1970a:12.13-14: na svato napi parato na dvabhya napy ahetuta / utpanna jatu vidyante bhava kvacana kecana). 21 In translating pramaa as warrant (cf., Apte 1992:1101, meaning 6), I have in mind the sense of the latter word as meaning justification for an action or a belief; grounds (American Heritage College Dictionary). There is a systematic ambiguity in the word pramaa in the Indian philosophical tradition, this word alternately referring to a reliable means of knowing (sa yenartha pramioti, tat pramaa), and to an episode of veridical cognition such as results from the exercise thereof (pramiyate iti pramaam). This ambiguity is preserved in the translation of pramaa as reliable warrant: warrant can refer to the outcome of a cognitive episode, to what one has (justification) in virtue of having formed a belief in a reliable way (so Plantinga 1993:3: that, whatever precisely it is, which together with truth makes the difference between knowledge and mere true belief); but it also conventionally denotes justification in the sense of the criterion or grounds of belief (What is your warrant for thinking there was a fire?; I saw it, or I saw smoke). A good translation of pramaa in the latter sense might be (following Alston 1989) doxastic practice but this fails to capture the other sense. The idea of warrant (and of being warranted), I think, also captures (without begging any important questions) the complex relation between pramaa and knowledge or truth, and is in this sense to be preferred to standard translations like valid cognition (Dreyfus 1997:570) or (translating this passage) valid means of right knowledge (Ruegg 2002:95). See, however, the next note on one problem with my translation.
20 19
419
object (prameya)22 depends on reliable warrants23 for an uncomprehended object cant be comprehended without reliable warrants. Hence, if theres no understanding of an object because there is no reliable warrant, how is [yours] a justifiable certainty (samyagniscaya)? So it doesn't make sense to say, [as in MMK 1.1, that] existents are unproduced24. Or again: It will be my [certainty] precisely that all existents exist, and that based upon the same thing as your certainty that existents are unproduced! And just as your certainty is that all things are unproduced, in exactly the same way, [56] mine will be that there is production of all things25.
22 The translation of prameya as warrantable is not unproblematic; for it is beliefs that are warranted, not (what is typically characterized in Sanskrit as prameya) objects. It is difficult, however, to find translation equivalents for this pair of words (pramaa and prameya) that avoid this problem while yet reflecting the fact that they are permutations of the same verbal root. Knowable seems better to capture the sense of prameya as describing any possible objects of (warranted) cognition; but as indicated in n.21, if we then translate pramaa as means of knowledge, we risk begging important questions about the relations between justification and truth. It is, then, my translation of pramaa as reliable warrant that informs the rendering of prameya as warrantable but the latter should, in this context, be understood as short for the more cumbersome object regarding which one could have a warranted belief. 23 With this point (pramaadhinatvat prameyadhigamasya; Tib., gzhal bya rtogs pa ni tshad ma la rag las pai phyir te), Candrakirti seems to allude to Dignaga (as noted by Ruegg 2002:95, n.155), whose Pramaasamuccaya begins with the claim that understanding of a warrantable object depends upon reliable warrants (in Kanakavarmans Tibetan, as given at Hattori p.175, gan gi phyir gal bya rtogs pa ni tshad ma la rag las pa yin). Cf., Hattoris n.1.10, p.76; and n.47, below, for another citation of Dignagas point. 24 55.11-16: Atra kecit paricodayanti: Anutpanna bhava iti kim aya pramaajo niscaya uta-apramaaja? Tatra, yadi pramaaja iyate, tada-ida vaktavya: kati pramaani, kilakaani, kiviayai, ki svata utpannani, ki parata ubhayato hetuto va-iti? Atha-apramaaja sa na yukta, pramaadhinatvat prameyadhigamasya. Anadhigato hy artho na vina pramaair adhigantu sakyata iti, pramaabhavad arthadhigamabhave sati, kuto yam samyagniscaya iti? Na yuktam etad anutpanna [de Jong] bhava iti. Clearly, the challenge with which this section thus begins is very much like challenges anticipated in Nagarjuna's Vigrahavyavartani (VV). Interestingly, though, Candrakirti does not note the VV's argument against this challenge until several pages into the present section (cf., n.47, below). 25 55.16-56.1: Yato va-aya niscayo bhavato nutpanna bhava iti bhaviyati tata eva mama-api sarvabhava santi-iti! Yatha ca-aya te niscayo nutpanna sarvadharma iti, tatha-eva [p.56] mama-api sarvabhavotpattir bhaviyati. That is, if Candrakirti is willing to give up on thinking his own beliefs to be demonstrably warranted, then he cannot think there are any grounds for preferring his beliefs to those of his interlocutor.
420
DAN ARNOLD
Or [perhaps you will say] you have no certainty [to the effect that] all existents are unproduced. In that case, since theres no persuading another of something not ascertained for oneself, its pointless to begin this treatise, and all existents stand unrefuted26. We reply: If we had anything at all like certainty, it could [be said to be] based on a reliable warrant, or not based on a reliable warrant. But we dont! How so? If there were the possibility of doubt (aniscaya) in regard to this, there could be a certainty opposed to that and dependent upon it. But when we have no doubt in the first place, then how could there be certainty opposed to it?27 For [such certainty] would be independent of anything else sharing the relation, as in the case of the longness or shortness of a donkeys horn. And when, in this way, there is no certainty, [p.57] then we will imagine reliable warrants for the sake of proving what? How, then, will they [i.e., reliable warrants] have number, characteristic, or object? Whether [their] production is intrinsic, dependent, both, or causeless none of this has to be explained by us28.
26 56.1-3: Atha te na-asti niscayo nutpanna sarvabhava iti, tada svayamaniscitasya parapratyayanasabhavac chastrarambhavaiyarthyam eva-iti, santy apratiiddha sarvabhava iti. 27 It is chiefly having in mind these first sentences of Candrakirtis response that I have translated niscaya, here and in the preceding paragraph, as certainty. One might more appropriately render this as conviction or, more weakly, opinion but this seems to me not as sharply to capture the contrast between niscaya and aniscaya, or (more significantly) the force of Candrakirtis claim here that there can be no question of the latter. It is not clear what it would mean for Candrakirti to claim that there is no possibility of non-conviction (aniscaya) in regard to the issue in question. If, in contrast, he is saying there is no room for (what is the opposite of certainty) doubt in the matter, it becomes possible to understand him as suggesting (what I take him to be arguing) that constitutively Madhyamika claims cannot coherently be thought to require the same kind of justification here demanded by his interlocutor, insofar as the truth of MMK 1.1 is a condition of the possibility even of expressing any doubt. Such is the line of interpretation I have more generally advanced in Arnold (2005). 28 56.4-57.3: Ucyate: Yadi kascinniscayo nama-asmaka syat, sa pramaajo va syad apramaajo va. Na tv asti. Ki karaa? Iha-aniscayasabhave sati, syat tatpratipakas tadapeko niscaya. Yada tv aniscaya eva tavad asmaka na-asti, tada kutas tadviruddho [de Jong] niscaya syat? Sabandhyantaranirapekatvat, kharaviaasya hrasvadirghatavat. Yada ca-eva niscayasya-[p.57]-bhava, tada kasya prasiddhyartha pramaani parikalpayiyama? Kuto va ea sakhya lakaa viayo va bhaviyati? Svata parata ubhayato ['hetuto] va samutpattir iti sarvam etan na vaktavyam asmabhi. This passage is translated by Huntington (2003:77-78), who identifies Candrakirtis interlocutor as Bhavaviveka (despite the proximate allusion to Dignaga; cf., n.23, above). Can-
421
[Objection:] If, in this way, [you have] no certainty at all, then how is your own statement which has the form of something ascertained, to wit not from themselves or from something else, nor from both nor altogether without cause, do existents exist understood?29 We reply: This statement is ascertained by reasoning that is just familiar for ordinary people, not for the venerable (arya). Does this mean the venerable have no reasoning? Who can say whether or not they do? For ultimate truth is a matter of venerable silence. So how could there be, in regard to it, any possibility of [the sort of] conceptual elaboration that is reasoning or non-reasoning?30 [Objection:] Well, if the venerable do not expound reasoning, then how, here and now, will they awaken the world to ultimate truth?31 [Response:] The venerable surely do not expound reasoning according to ordinary usage.32 Rather, granting, for the sake of awakening others, reasoning that is familiar only in the world in just that way they awaken the world33. For example, those in the throes of passion, comdrakirtis response in terms of niscaya represents a point conceptually similar to Nagarjunas claim, in the Vigrahavyavartani (v.29), to have no thesis (pratija): If I had any thesis, then the fault would be mine; but I do not have a thesis, so I have no fault at all. (Bhattacharya 1990:14: yadi kacana pratija syan me tata ea me bhaved doa / nasti ca mama pratija tasman naivasti me doa.) 29 57.4-5: Yady eva niscayo na-asti sarvata, katha punar ida niscitarupa vakyam upalabhyate bhavata? Na svato napi parato na dvabhya napy ahetuto bhava bhavantiti. 30 57.5-8: Yady eva niscayo na-asti sarvata, katha punar ida niscitarupa vakyam upalabhyate bhavata? Na svato napi parato na dvabhya napy ahetuto bhava bhavantiti. Ucyate: Niscitam ida vakya lokasya svaprasiddhayaivopapattya, naaryaa. Ki khalv aryaam upapattir na-asti? Kena-etad uktam asti va nasti va-iti? Paramartho hy aryas tuibhava [de Jong]. Tata kutas tatra prapacasabhavo yad upapattir anupapattir va syat? 31 57.9: Yadi hy arya upapatti na varayanti kena khalv idani paramartha loka bodhayiyanti? 32 Ruegg (2002:99) translates: The Aryas do not propound any justified ground in virtue of the transactional-pragmatic usage of ordinary folk in the world. There is a sense, however, in which such a technical rendering of this makes Candrakirtis own statement of his argument performatively incoherent; his use of ordinary language is best understood as itself exemplifying his deference to such. In this regard, we might also render lokasavyavaharea as according to business as usual, capturing the eminently conventional, mercantile sense of the word vyavahara (see n.17) but my taste for this translation has met with such howls of protest that I here defer to cooler heads. 33 There seems to be a tension here; why is the latter (i.e., the venerables causing the
422
DAN ARNOLD
mitted to a mistake, do not apprehend even the actual (vidyamanam api) impurity of the body and having imputed an unreal aspect of beauty, [they] suffer. For the sake of [cultivating] their dispassion, a manifestation of the Tathagata or a god could describe in detail the defects of the body, which were previously concealed by the idea of beauty. [They will describe these, for example,] by saying things such as that there are hairs on the body34. And those [who had been passionate], by abandoning that idea of beauty, could attain dispassion. [p.58] So, too, in this context: by virtue of being [ones] the eye of whose mind is impaired by the cataracts of ignorance35, ordinary people imputing to existents an essence (and in some cases, some particular qualification) whose nature is not at all being perceived by the venerable suffer excessively36. Now the venerable awaken them [to all this] through reasoning that is familiar to them. For example, its [generally] granted that there is no
world to understand having accepted reasoning that is familiar in the world) not a case of their propounding something according to ordinary usage? Siderits (1981: 125-126) comments: Here we must note the extreme care which Candrakirti takes to avoid the suggestion that the aryas seek to prove the ultimate truth. When they set out to instruct the world through the manipulation of the conventionally accepted epistemic practices, what they construct is not a proof but rather what would be considered by the world to be a well-established proof. The qualification is crucial, for if the Madhyamika is said simply to prove the ultimate truth, there is the implication that he is in possession of ultimate means of proof, that is, that he is in possession of a theory of pramaas which he knows to be unconditionally valid. Sideritss point is recommended by the contrast here signaled only by Candrakirtis emphasis: the venerable do not, as it were, themselves depend for their knowledge on ordinary usage; rather, they provisionally adopt that reasoning which is familiar only in the world (lokata eva ya prasiddhopapattis ta). 34 As La Valle Poussin notes (57: n.5), Candrakirti here alludes to the practice of smtyupasthanabhavana, as described, for example, in the Sikasamuccaya; cf., Bendall 1970:235. 35 This would seem to be a favorite expression, for Candrakirti uses it repeatedly. Cf., inter alia, p.261.4, where precisely the same expression is used. 36 57.10-58.3: Na khalv arya lokasavyavahareopapatti varayanti. Ki tu lokata eva ya prasiddhopapattis ta paravabodhartham abhyupetya tathaiva loka bodhayanti. Yathaiva hi vidyamanam api sarirasucita viparyasanugata ragio nopalabhante subhakara ca abhutam adhyaropya pariklisyante. Tea vairagyartha tathagatanirmito devo va subhasajaya prak pracchaditan kayadoan upavarayet. Santy asmin kaye kesa [itya]dina. Te ca tasya subhasajaya vigamad [de Jong] vairagyam asada-[p.58]-yeyu. Evam ihapy aryai sarvathapy anupalabhyamanatmaka bhavanam avidyatimiropahatamatinayanataya viparita svabhavam adhyaropya kvacic ca kacid viseam atitara pariklisyanti pthagjana.
423
production of an [already] existent jar from the clay and so forth; in this way, it should be determined that there is no production, since what exists prior to production already exists37. Or, for example, its accepted that a sprout is not produced from the coals of a fire, which are other than it; likewise, it should be ascertained that [production] is not from the seeds and so forth, even though they are intended [as the cause of sprouts]38. [Objection:] Then one could [rejoin that] this is our experience39.
37 That is, the causation of something from itself would entail that the thing in question already exists in which case, its coming-into-being would no longer require explanation. This basically reproduces Buddhapalitas argument regarding the na svato part of the tetralemma presented at MMK 1.1; Buddhapalitas Sanskrit is cited by Candrakirti at p.14.13. (The Tibetan translation of Buddhapalitas entire commentary on MMK 1.1 can be found in Walleser 1970:11.8, ff. See also Saito 1984.) The argument is traditionally understood as directed against the Sakhya proponent of the doctrine of satkaryavada i.e., of the view that effects pre-exist in their causes (insofar as there is, for the Sakhya, properly speaking no causation whatsoever, but only the transformation [pariama] of prakti). 38 58.3-6: Tan idanim aryas tatprasiddhayaivopapattya paribodhayanti. Yatha vidyamanasya ghaasya na mdadibhya utpada ity abhyupetam, evam utpadat purva vidyamanasya vidyamanatvan, na asty utpada ity avasiyata. Yatha ca parabhutebhyo jvalangaradibhyo nkurasyotpattir na astity abhyupetam, eva vivakitebebhyo pi bijadibhyo na astity avasiyata. In other words, seeds are what the proponent of this account of causation intended to allow to stand; but these cannot be allowed, either, because they are just as different from the sprout as coals are. Particularly here, Candrakirtis argument seems not to have a very strong claim to represent reasoning that is familiar. It can, though, be so understood, if it is appreciated that the argument here is a basically a priori analysis of concepts, and not an a posteriori analysis of the phenomena putatively explained thereby. Specifically, the argument turns simply on the definition of other; the point is that the general concept of otherness leaves us with no principled way to know which other things are relevantly connected to the thing whose arising we seek to explain, and we are thus left to suppose that anything that is other than the latter (even, e.g., the coals of a fire) could give rise to it. Candrakirtis argument again repeats that of Buddhapalita, who had similarly argued only by reducing to absurdity the opponents account of arising from another, without offering his own, alternative account of causal production. Thus, Buddhapalita (Walleser 1970: 11) says: Existents do not arise from something other. Why? Because it would follow that anything [can] arise from anything else (gzhan las kyang skye ba med do / cii phyir zhe na / thams cad las thams cad skye bar thal bar gyur bai phyir ro). Cf., Prasannapada 36.11-12, where Candrakirti approvingly quotes Buddhapalitas Sanskrit (na parata utpadyante bhava, sarvata sarvasabhavaprasangat). 39 58.7: Athapi syad anubhava eo smakam iti. The point of the objection, it seems, is that surely we all just see that things are produced from other things. We can follow the lead of the Madhyamakavatara in finding here a discussion of the status of what is, for Dignaga and his foundationalist heirs, the privileged faculty of perception (with the issues raised by following this avenue being issues that Can-
424
DAN ARNOLD
[Response:] This doesnt make sense, either, since this experience is false, [simply] because of its being experience like the experience of two moons on the part of someone with cataracts. Therefore, by virtue of the fact that experience similarly requires proof, this objection doesnt make sense40. Therefore, existents are not produced in this way, the first chapter [of Nagarjunas Mulamadhyamakakarika] begins first of all by countering the imputation of a false nature. Now, the remainder of the treatise is undertaken for the sake of refuting some qualifications that are imputed in particular cases. Dependent origination does not have any single qualification, not even such as being the agent, the locus, or the action of motion [this treatise is undertaken] for the sake of showing [that]41. [Objection:] It is only ordinary usage (vyavahara) regarding warrants and warrantable objects that we have explained with [our] treatise42.
drakirti will go on to elaborate in the section of the Prasannapada that is presently unfolding). Thus, in considering the same objection, Candrakirtis Madhyamakavatarabhaya specifically introduces perception (pratyaka) as what chiefly informs our experience (anubhava). (On the basic equivalence, for Dignaga, of anubhava and pratyaka, see Dignagas commentary on Pramaasamuccaya, 1.6ab, in Hattori 1968:27.) It is, the interlocutor there argues, evident simply on the basis of perception that existents are produced from other existents; and appeal to reasoned argument is appropriate only with respect to things that are not perceptible, and not with respect to what is perceptible. Therefore, even without any argument, it must still be true that existents are produced from other [existents]. La Valle Poussin 1970b:101: rigs pa nye bar god pa yang dngos po mngon sum ma yin pa kho na la os kyi mngon sum la ni ma yin te / dei phyir thad pa med par yang dngos po rnams gzhan las skye ba yod pa kho nao. 40 58.7-9: Etad apy ayukta, yasmad anubhava ea ma, anubhavatvat. Taimirikadvicandradyanubhavavad iti. Tatas ca anubhavasyapi sadhyasamatvat tena pratyavasthana na yuktam iti. 41 58.10-13: Tasmad anutpanna bhava ity; eva tavad viparitasvarupadhyaropapratipakea prathamaprakaraarambha. Idani kva cid ya kascid viseo dhyaropitas tadviseapakaraartha seaprakaraarambha. Gantgantavyagamanadiko pi niravaseo viseo nasti pratityasamutpadasyeti pratipadanartha. Ruegg translates: the remaining chapters [of the MK] have been taken up [by Nagarjuna] in order to exclude some particularity [mistakenly] imputed in some place, and this with the purpose of conveying that also no particularity at all exists [as a hypostatic entity] for origination in dependence. (2002:102; as in all of my references to Rueggs translation, the ellipses here represent places where Ruegg has supplied the Sanskrit or Tibetan terms) 42 58.14-15: Atha syad ea eva pramaaprameyavyavaharo laukiko smabhi sastreanuvarita iti. If we take sastrea here in the sense of treatise, then perhaps the reference is specifi-
425
[Response:] Then it should be explained what the fruit of [your] explanation of this [ordinary usage] is43. [Objector continues:] It [i.e., ordinary usage] has been destroyed by sophists (kutarkikai), through their predication of a mistaken definition. [p.59] We have stated its correct definition44. [Response:] This doesnt make sense, either. For if, based on the composition of a mistaken definition by sophists, everyone were mistaken regarding whats being defined (kta lakyavaiparitya lokasya syat), [then] the point of this [proposed re-description of our epistemic practices] would be one whose effort was fruitful. But its not so, and this effort is pointless.45
cally to Dignagas Pramaasamuccaya. Ruegg (2002:102) instead translates by means of the [philosophical] science. Following that lead, we might better render this adverbially: which has been explained by us philosophically. That reading particularly underscores that though he claims to offer an account of our conventions, Dignaga does so by way of a peculiarly technical re-description thereof. This is, in any case, a crucial juncture in the argument; for having thus anticipated his interlocutors claim only to be offering a conventionally valid account, Candrakirti will be concerned from here on to argue only that Dignaga cannot coherently claim this. What is conventionally true is just our conventions, and it is therefore self-contradictory to elaborate a project that purports to be conventionally valid, while yet deploying words in something other than their conventional sense. From this point on, then, Candrakirti will argue only that Dignagas use of the key terms svalakaa and pratyaka cannot accommodate ordinary usage of these words. 43 58.15: Tadanuvarasya tarhi phala vacya. 44 58.15-59.1: Kutarkikai sa nasito viparitalakaa-[p.59]-bhidhanena. Tasya asmabhi samyaglakaam uktam iti cet. 45 59.1-3: Etad apy ayukta. Yadi hi kutarkikair viparitalakaapraayana [according to the Tibetan available to La Valle Poussin, brjod pas, =Skt. praayanat; adopted by Vaidya (1960: 20), whose reading I follow] kta lakyavaiparitya lokasya syat. Tadartha prayatnasaphalya syat. Na ca etad evam iti vyartha evaya prayatna iti. It is with respect to this passage that the anonymous author of the *Lakaaika specifically identifies Dignaga as Candrakirtis interlocutor: He says that on this view, it makes sense only [to speak of] the worldly convention regarding warrants and warrantable objects, not [what is] ultimate[ly the case]. [This is what is said in the passage] beginning Atha. [Its correct characteristics have been explained] by us means by Dignaga, et al. It is the master [i.e., Candrakirti] who says, at this point, the fruit of this intention should be explained, and it is Dignaga who rejoins, [It has been destroyed] by sophists. It [here] means convention. (Lakaaika 2b4; Yonezawa 2004: 142: laukika eva pramaaprameyavyavaharo yukto na paramarthika ity asmin pake aha / athetyadi / asmabhir Dignagadibhi / tadanubandhanasya phala vacyam ityatrarya, kutarkikair iti Dignaga, sa iti vyavahara).
426
DAN ARNOLD
Moreover, if comprehension of warrantable objects is dependent upon reliable warrants46, [then] by what are these reliable warrants [themselves] ascertained? This fault was pointed out in [Nagarjunas] Vigrahavyavartani. Since you still havent answered this, theres no illumination of the correct definition [by you].47 Moreover, if you say there are [only] two reliable warrants, corresponding respectively to the two [kinds of warrantable objects, i.e.,] unique particulars and abstractions48, [then we are entitled to ask,] does the subject (lakya) which has these two characteristics exist?49 Or does it not
46 Candrakirti here again alludes to Dignagas claim that pramaadina prameyadhigama (with the Tibetan here matching the Tibetan translation of Dignaga: gzhal bya rtogs pa tshad ma la rag las pa yin; cf., n.23, above). 47 59.4-6: Api ca, yadi pramaadhina prameyadhigamas tani pramaani kena paricchidyanta ityadina Vigrahavyavartanya vihito doa. Tadapariharat samyaglakaadyotakatvam api nasti. Candrakirti here finally refers to the main argument against the main objection from the Vigrahavyavartani specifically, the argument at Vigrahavyavartani 31-33: If your establishment of all these points is based on pramaas, we say: how is there establishment of these pramaas of yours? If the establishment of pramaas is by other pramaas, there would be an infinite regress (Bhattacharya 1990:15-16: yadi pramaatas te tea tea prasiddhir arthanam, tea puna prasiddhi bruhi katha te pramaanam. Anyair yadi pramaai pramaasiddhir bhavet tadanavastha) It is interesting, though, that while Candrakirti clearly endorses the argument, he does not elaborate on it, instead merely noting that it has not yet been met by his opponent. Here, he has other fish to fry specifically, relating to the ordinary use of conventional terms. 48 I render svalakaa as unique particular when it is Dignagas usage that is in play (though Candrakirtis point will be that the word cannot coherently be thought to mean this); I render samanyalakaa as abstraction (rather than more customarily as universal). The category of samayalakaa would, to be sure, include such examples of universals as sets and (if such were ever explicitly discussed in the Indian context) propositions. It is also meant, however, to include items such as satanas, mental continua cases, that is, such as later exponents like Mokakaragupta will characterize as vertical, as contra horizontal, samanyalakaas. 49 This could also be rendered: is that which has these two characteristics a lakya, or not?; or, taking lakya more literally as a gerundive, is that which has these to be characterized, or not? On any of these readings, though, the effect of Candrakirtis point remains substantially the same: svalakaa, which etymologically refers to an act of characterizing (lakaa), must involve the characterizing of something. Dignaga cannot allow this to the extent that his use of the term involves a fairly radical commitment to the idea that unique particulars (which is what svalakaa denotes for him), if they are really to count as unique, can neither be nor have any properties; for any reference to properties is, ipso facto, the kind of discursive activity that trades in things (namely, the referents of words) that are constitutively not unique.
427
exist? If it exists, then there is an additional warrantable object50; how, then, are there [only] two reliable warrants?51 Or perhaps [you will say] the subject [that is characterized by these characteristics] does not exist. In that case, the characteristic, being without a locus, doesnt exist either; how, [in that case,] are there [as many as] two reliable warrants? As [Nagarjuna] will say [in MMK 5.4]: When a characteristic is not operating, a subject to be characterized doesnt stand to reason; and given
Namely, the subject of which these different lakaas are characteristics. That is, insofar as the number of pramaas, for Dignaga, tracks the number of kinds of existents, the need to introduce an additional kind of existent would undermine his epistemology. Ruegg seems to understand this passage a little differently, translating as follows: Furthermore, if [Dignaga] has stated [the existence of] a pair of pramaas in conformity with the pair [comprised] of the own [i.e. particular characteristic] and the generic characteristic, does there exist this characterized definiendum [i.e. the twofold pramaa] for which there is this pair of defining characteristics? (2002:104; insertions original; emphasis added) As reflected in his final insertion, Ruegg here takes the point to concern the definition (lakaa) hence, the existence of the two pramaas (which are thus taken as what is lakya, being defined). But the point here does not, I think, concern the definitions of the pramaas themselves; rather, the point is simply that the words sva- and samanya-lakaa, insofar as they are forms of the word lakaa (which denotes an act of characterizing), constitutively involve some relationship specifically, between a characteristic (lakaa), and the thing characterized thereby (lakya). And Candrakirtis point is that Dignaga cannot concede this, insofar as he understands svalakaa as a unique or bare particular that is, as neither being nor having any properties or characteristics at all. This point is further obscured by translating svalakaa, on what Candrakirti (at least) takes to be Dignagas use thereof, as particular characteristic (as Ruegg does; consider, as well, Dreyfuss rendering of this [1997:580, et passim] as specifically characterized phenomenon). Shoryu Katsura, in explaining a critique of Hattoris translation of Pramaasamuccaya 1.2, makes a point that cuts as well against these renderings of svalakaa (again, on Dignagas understanding thereof); specifically, they may suggest that the object to be cognized is a possessor of the two lakaas and [is to that extent] something different from them. [But] I do not think that Dignaga admitted any bearer of the two lakaas. (Katsura 1991:136; cf., Arnold 2003) Similarly, Candrakirti thinks Dignaga cannot admit that svalakaa are (as Ruegg says) particular characteristics (or as I will translate what Candrakirti takes to be the conventional sense of the word, defining characteristics) at all, since that would compromise his commitment to the view that there are only two types of existents; for on the conventional sense of the word, svalakaas would thus have to be the properties (or characteristics) of some additional kind of existent. Candrakirti is not, then, here talking about pramaas as the characterized definiendum (lakya); he is simply starting to make his point that Dignagas use of the word svalakaa is incoherent.
50 51
428
DAN ARNOLD
the unreasonableness of a subject to be characterized, there is no possibility of a characteristic, either.52 [Objection:] [p.60] It is not that lakaa means that by which [something] is characterized. Rather, [according to the rule that] the ana affix is variously applicable,53 taking the affix in the sense of an object (karmai), lakaa means what is characterized.54 [Response:] Even so, the same problem [still obtains], because of the impossibility of somethings being characterized by itself; for that instrument by means of which a thing is characterized is something different from the object [that is characterized thereby]55.
52 59.7-11: Ki ca yadi svasamanyalakaadvayanurodhena pramaadvayam ukta, yasya tallakaadvaya ki tal lakyam [de Jong; so, too, Vaidya] asti? Atha nasti? Yady asti, tada tadapara prameyam astiti, katha pramaadvaya? Atha nasti lakya, tada lakaam api nirasraya nastiti katha pramaadvaya? Vakyati hi: lakaasapravttau ca na lakyam upapadyate, lakyasya anupapattau ca lakaasyapy asabhava, iti. Candrakirti's commentary on MMK 5.4 verse is on pp.131-2 of La Valle Poussin's edition, but given the prominence of his discussion in chapter 1, there is surprisingly little there. There are, however, many other passages relevant to the critique of svalakaas in Candrakirtis Madhyamakavatara. Cf., e.g., 6.22-36, which ends with a point specifically contra Dignagas view of svalakaas: gal te rang gi mtshan nyid brten gyur na / de la skur pas dngos po jig pa'i phyir / stong nyid dngos po jig pa'i rgyur gyur na / de ni rigs med de phyir dngos yod min (6.34; La Valle Poussin 1970a:117: If [an entity exists] in dependence on a svalakaa, then through negation of that the entity would be destroyed, and emptiness would be the cause of its destruction [i.e., if emptiness were taken as negating really existent svalakaas, then it would be a nihilistic doctrine]. This is not the case, however, because entities do not [intrinsically] exist. And 6.36 (p.123): dei phyir rang gi mtshan nyid kyi skye ba ni bden pa gnyis char du yang yod pa ma yin no (Therefore, from the point of view of either of the two truths, there is no production of particulars). The latter point (i.e., that this sense of svalakaas does not obtain from the point of view of either of the two truths) neatly expresses Candrakirtis contention that Dignagas account of our epistemic practices is not only not ultimately, but not even conventionally valid. 53 Citing Paini's Aadhyayi III.3.113 (ktyalyuo bahulam). See Katre 1987:303. 54 60.1-2: Atha syan na lakyate neneti lakaa. Ki tarhi ktyalyuo bahulam iti karmai lyua ktva lakyate tad iti lakaa. The same rule from Paini can be invoked to explain the different senses of the word pramaa (cf., n.21, above), which too is formed by affixing the -ana suffix to a verbal root. 55 60.2-3: Evam api tenaiva tasya [de Jong] lakyamaatvasabhavad yena tal lakyate [de Jong] tasya karaasya karmao rthantaratvat sa eva doa. I thus take the force of eva (tenaiva) to be by that very same thing; hence, I translate, by itself. Candrakirtis point is the eminently grammatical one that the instrument by which something is effected (in this case, by which something is characterized) is, by virtue of its being an instrument, something that cannot at the same time be an object. Thus, just as a
429
[Objection:] Well, perhaps this could be said: Because of cognitions being an instrument, and because of the inclusion of this in [our concept of] the unique particular (svalakaa)56, this is not the problem [you have said it is]57. [Response:] In this connection, that which is the unique, intrinsic nature (svarupa) of existents is [what is conventionally referred to as] their defining characteristic (svalakaa)58. For example, earths [defining characteristic] is resistance, [that] of feeling is experience, [that] of perceptual cognition is the specific representation of an object.59 Therefore, taking [svalakaa] in the sense of what is characterized,60 and [thus] disregarding the etymology that follows the familiar sense, [our interlocutor] takes it as denoting an object (karmasadhanam)61. And by
semantically complete verbal construction requires reference to various karakas, so, too, the act of characterizing constitutively involves reference to the discrete components of that action. 56 Here, I again translate svalakaa to reflect Dignagas use of the word. 57 60.4: Atha syat: Janasya karaatvat, tasya ca svalakaantarbhavad, ayam adoa iti. 58 In this crucial passage, Candrakirti explicitly states what he takes to be the conventional sense of the word svalakaa which, as made clear by the examples he gives, is to be translated as defining characteristic when it is his favored sense of the word that is in play. 59 Note that Candrakirtis examples can be found in Vasubandhus Abhidharmakosa; thus, the svalakaa of vijana is adduced at Abhidharmakosa 1.16a (Pradhan 1975:11), and that of pthivi at Abhidharmakosa 1.12 (Ibid.: 8). My rendering of vijana as perceptual cognition reflects my agreement with the observation of Bruce Hall (1983: 84n), who notes, with the Abhidharmakosas definitions of the terms in mind, that vijana in the Abhidharmakosa roughly corresponds to the sense of pratyaka (perception) recommended by Dignaga and Dharmakirti, and that the Abhidharmakosas usage of saja (conception) corresponds to their sense of anumana (inference). Cf., also, Madhyamakavatara 6.202-3 (La Valle Poussin 1970a: 316), where Candrakirti trots out a similarly Abhidharmika list of defining characteristics (svalakaas) of all of the skandhas: Form has the defining property (svalakaa) of color and shape (rupaa); vedana has the nature of experience; saja grasps characteristics; saskaras fashion [things]; the defining property of perceptual cognition is a conception regarding any object (gzugs ni gzugs rung mtshan nyid can / tshor ba myong ba'i bdag nyid can / du shes mtshan mar dzin pa ste / du byed mngon par du byed pa'o // yul la so sor rnam rig pa / rnam shes rang gi mtshan nyid do /). 60 See n.54, above. 61 The Tibetan translation renders this as las su sgrub pa, which suggests established as an object. But the sense of -sadhana as denoting or expressive of (cf., Apte, p.1666, meaning #4) comes from its being a synonym for karaka the Sanskrit gram-
430
DAN ARNOLD
positing [at the same time] the instrumental nature of perceptual cognition, it is said [in effect] that one unique particular has the quality of being an object, and another unique particular has the quality of being an instrument. [p.61] In that case, if the svalakaa62 of perceptual cognition is an instrument, then it must have a separate object (tasya vyatiriktena karmaa bhavitavyam)63. This is the fault (in your position)64.
marians category for designating the various components of an action. Cf., in this regard, not only Abhyankar 1977:423 (s.v. sadhana; cited by Ruegg, p.106, n.188), but also Bhattacharya 1980 (especially pp.87-89), who cites similar uses by Candrakirti of the term sadhana in the sense of karaka. See also Bhattacharya 1980-81. (There is a precisely similar usage of the term -sadhana in Dharmottaras Nyayabinduika: karaasadhanena manasabdena sarupyalakaa pramaam abhidhiyate [by virtue of the word mana, which denotes an instrument, pramaa is defined as characterized by conformity]; Malvania 1971: 39.) In characteristically Sanskritic fashion, then, the argument here is advanced entirely in grammatical terms. 62 Particularly in this section, it is often difficult to translate this word one way or the other, without begging precisely the question at issue (viz., what the word should mean). Cf., nn.69, 98, below. 63 Here, note the use of the gerund bhavitavyam to indicate something like the mode of necessity. This bhave prayoga construction is missed by Siderits, who instead translates But then if the consciousness svalakaa is the instrumental, it should be by means of a distinct accusative of that, just this is the defect (p.134; my emphasis). This translation leads him to suppose there is a problem understanding the antecedent of the final pronoun (of that), which he then spends a couple of pages explaining. On my reading, though, the point is straightforward. 64 60.4-61.2: Ucyate: Iha bhavanam anyasadharaam atmiya yat svarupa, tat svalakaa. Tadyatha pthivya kainya, vedanaya anubhavo [de Jong], vijanasya viayaprativijapti. Tena hi tal lakyata iti ktva, prasiddhyanugata [de Jong] ca vyutpattim avadhuya karmasadhanam abhyupagacchati. Vijanasya ca karaabhava pratipadyamanena-idam ukta [de Jong] bhavati, svalakaasyaiva karmata, svalakaantarasya karaabha-[p.61]-vasceti. Tatra yadi vijanasvalakaa karaa, tasya vyatiriktena karmaa bhavitavyam iti sa eva doa. This has seemed to me to be a difficult passage, though I have been persuaded by one of my readers that the problems are resolved by translating the passage as I have here. The problems start with the passage tena hi tal lakyata iti ktva (60.6). (This is La Valle Poussins conjecture. His manuscripts read tena hi tad va na lakyate; see his note 6, p.60.) The particle hi suggests that the phrase qualifies what immediately precedes it, explaining why the defining characteristics just adduced should be reckoned as the svalakaa of the things in question; thus, we could (as I had originally wanted to) read, for [in each of these cases,] by that [quality the thing in question] is characterized. So, too, Ruegg (2002:106): this [particular entity] is characterized by this [its specific characteristic] (tena hi tal lakyata). The Tibetan translation, too, seems to take the passage this way, moving tena tal lakyata up to the beginning of Candrakirtis own definition, thereby clearly including this as part of what Candrakirti commends: bshad par bya ste
431
[Objection:] Then it could be this way: What is apprehendable (gamya) by perceptual cognition, such as the resistance and so forth that are comprised by things like earth that just is the direct object65 of
/ re zhig dir ji ltar des de mtshon par byed di sai sra ba dang / tshor bai myongs ba dang / rnam par shes pai yul so sor rnam par rig pa ltar bdag nyid kyi rang gi ngo bo gzhan dang thun mong ma yin pa gang yin pa de ni rang gi mtshan nyid yin na. (Cf., La Valle Poussins reconstruction of the Sanskrit from this, his note 6.) Although this is a conceptually possible reading, the placement of the phrase in the Sanskrit text together with its echo of the alternative interpretation of lakaa proposed above (n.54) recommends instead taking tena in the transitional sense of therefore, and tal lakyate as a quotation of the earlier attempt to salvage the interlocutors favored sense of svalakaa. Not only does this yield a more forceful indictment of the interlocutor, but it makes better sense of the two continuative ca particles that follow: thus it is the same person who takes the word this way (tal lakyate iti ktva); who thus disregards the familiar sense (prasiddhyanugata ca vyutpattim avadhuya); and who also wants to allow that cognition is an instrument (vijanasya ca karaabhava pratipadyamanena). The passage might be rendered a little clearer by emending abhyupagacchati (at 60.7) to abhyupagacchata; this would complement pratipadyamanena, giving two instrumental present participles to construe with idam ukta bhavati. Consider, by contrast, Rueggs translation, which rather obscures the fact that Candrakirti is here indicating that two contradictory things are said by the same person (which is why he can conclude by convicting his interlocutor of incoherence): having put aside the derivation [of the term lakaa] that has been generally acknowledged, one takes it to have an objective realization. Apprehending the cognition to be an instrument, one states that the specific defining characteristic itself has the condition of being the object and that another specific defining characteristic has the nature of an instrument. Here the fault lies precisely in the fact that if the svalakaa of a vijana is an instrument, there has to exist for it an object separate [from it] (2002: 106). Rendering the passage literally, with the emendation here suggested, we get instead: By one who is taking [svalakaa] as denoting an object taking [svalakaa] in the sense of what is characterized, and disregarding the etymology that follows the familiar sense and who is [also] positing the instrumental nature of perceptual cognition, it is said. (This is clear in the Tibetan, which subordinates the first two clauses to the third, which is then the subject of idam ukta bhavati: rab tu grags pa dang rjes su brel pai bye brag tu bshad pa bor nas / las su sgrub pa khas len zhing rnam par shes pa byed pai ngo bor rtogs pas ni / rang gi mtshan nyid kho na las nyid yin zhing rang gi mtshan nyid gzhan ni byed pai ngo bo yin no zhes bya ba di smras par gyur ro.) I am indebted to an anonymous reviewer of this article for some of the foregoing suggestions. The proposed emendation of abhyupagacchati accomplishes (though perhaps more straightforwardly) the same thing as an emendation suggested to me by Sheldon Pollock (personal communication): if abhyupagacchati were read as a locative present participle, pratipadyamanena could be emended to pratipadyamane, giving two locative absolute constructions: When, disregarding the usage which follows the familiar sense, one accepts, and when one [at the same time] accepts. 65 That is, in the grammatical sense that still governs the discussion. Throughout this section, the expression direct object will here render karma in this grammatical sense.
432
DAN ARNOLD
that [perceptual cognition], and it is not distinct from the unique particular (svalakaa) [that is really being perceived]66. [Response:] Even so, then because the defining characteristic (svalakaa) of perceptual cognition is not [itself] a direct object [of cognition], it could not be a warrantable object (prameya)67, since only a svalakaa in the form of a direct object can be a warrantable object68. And thus, since you have specified (ity etad viseya) that two kinds of things unique particulars and abstractions can be the objects of reliable warrants, youre now forced to say: one unique particular is the
66 61.3-4: Atha syat: Yat pthivyadigata kahinyadika vijanagamya, tat tasya karmasty eva, tac ca svalakaavyatiriktam iti. Here, I think, two points are being made. In the grammatical key that is here predominant, the interlocutors point is that vijana which conventionally occurs (as at n.59, above) in the expression the svalakaa [defining characteristic] of vijana is, as designating an act of cognizing, a grammatical instrument (that by means of which some agent apprehends something); and as such, it must itself have some direct object. By pointing this out, he hopes to meet Candrakirti's immediately preceding point that his position entails an infinite regress. His point is that with vijana, we do have an instance of something that is both itself a svalakaa in the sense of a grammatical instrument, and that yet has a svalakaa (unique particular) as its direct object. This proposal further amounts, I think, to a second point: that svalakaa as defining characteristic is the same thing as svalakaa as unique particular, so that, e.g., if we speak of the earth's resistance as its defining characteristic, we can do so because there is a corresponding unique particular that we perceive. Thus the thing commonly adduced as earths svalakaa (i.e., hardness) in fact has an ontological correlate, in the form of Dignagas unique particular. We can, then, speak of perceiving somethings defining characteristic just insofar as there invariably corresponds to this some unique and concrete particular. Again, this is proposed as a way for Dignaga to retain his commitment to the view that svalakaa means unique particular, while yet explaining common expressions like resistance is the svalakaa of earth. 67 That is, it could not itself be the object of a pramaa which would, for Dignaga, be tantamount to saying that it does not exist. Again, the point is here made in grammatical terms: an act of pramaa (warranting) must have a prameya, a direct object and to the extent that vijana is instead thought to be (grammatically) an instrument, it therefore could never fulfill this role. 68 61.4-5: Eva tarhi vijanasvalakaasya karmatvabhavat, prameyatva na syat, karmarupasyaiva svalakaasya prameyatvat. Here, Candrakirti has effectively rejoined that the interlocutors previous move will no longer allow us to accommodate the conventional usage, according to which there is a defining characteristic (svalakaa) of vijana; for if, instead, we read this conventional expression as meaning the unique particular which is vijana, then this is tantamount to saying that the subjective cognitive act of awareness is really the object of some other cognitive act, insofar as Dignagas usage takes the word as karmasadhanam (denoting an object).
433
object of a reliable warrant the one thus pointed out as what is characterized; and one is not the object of a reliable warrant the one by which something is characterized.69 [Perhaps you will rejoin that] that one, too, denotes an object (karmasadhana)70. Then that one must, [in turn,] have some other instrument71. And given this conception of the status of an instrument on the part of another [moment of] cognition72, an infinite regress ensues73.
69 Ruegg translates: Some thing, a svalakaa which is the prameya, is designated as what is definingly characterized; and some thing [other], which is not the object of a right cognition, is designated as definingly characterized by that (2002:107). But this misses, I think, the order of predication; that is, the salient point of the sentence is that prameyatvam and aprameyatvam are here predicated of these two kinds of subjects, so that the interlocutor is forced to admit that one of them is aprameya. Consider, as well, Rueggs annotation of the passage (107, n.190): In his discussion here Candrakirti seems to conjoin lakaa, or svalakaa1, (specific) defining characteristic of a thing (in the Abhidharma example), or of a term, and svalakaa2 particular characteristic which (in the Pramaa-school) is the cognitive object of pratyaka. (Cf., n.98, below, for a similar point by Siderits.) While Ruegg is surely right to note that this passage (like this entire section of the text) crucially involves some alternation in meaning, the point to be made is not, I think, that Candrakirti thus conflates these; rather, Candrakirti is offering defining characteristic as the conventional sense of the word, and unique particular as the sense that Dignaga presupposes and he is saying that Dignaga ends up with a contradiction, unable both to remain true to his spartan epistemology, and to explain familiar uses of the word. 70 That is, that the svalakaa of perceptual cognition (vijana), too, is a (perceptible) object. The upshot of this is that if a cognition, in order to count as such, must be not only an instrument but also an object, then each instance of cognition must be accompanied by a further cognition for which it is such an object. Candrakirti is here driving towards a consideration of Dignagas idea of svasavitti which, however, will chiefly be considered, in this context, as the unique example of something that is at the same time an instrument and an object. The point in Dignagas introducing the idea at this juncture, in other words, will chiefly be to salvage the possibility that his svalakaas might be (as Candrakirti thinks he must say) simply self-characterizing. 71 tada tasyanyena karaena bhavitavya. Siderits again misses the sense of the bhave prayoga construction, instead giving If the means of action [the cognition] is just that [the svalakaa], then it should come to be by means of another instrumental of that [cognition] (p.136; my emphasis) But the third case here (karaena) indicates not an instrument, but the subject of the verb bhavitavyam, i.e., the thing which must (despite the interlocutors view to the contrary) exist on the part of that (tasya). Such constructions are clearly expressed in the Tibetan, which handles them without use of the gerundive: dei tshe de la byed pa gzhan zhig yod par bya dgos la. 72 I.e., given that it, too, would have to be the object of a further cognition if it is to count as an instance of cognition that yet counts as a svalakaa (where, of course, that is understood as karmasadhana). 73 61.5-9: Tatas ca dvividha prameya svalakaa samanyalakaa ca. Ity etad viseya vaktavya: kicit svalakaa prameya yal lakyata ity eva vyapadisyate,
434
DAN ARNOLD
Perhaps you think there exists [the faculty of] apperception (svasavitti). Based on that, [you maintain that], given that [cognitions] being an object obtains due to [its] apprehension by apperception, [cognition] is included among warrantable objects74. To this we respond, based on an extensive refutation of apperception in the Madhyamakavatara75: it doesnt make sense to say a svalakaa [p.62] is characterized by another svalakaa, and that one by apperception. Moreover, this latter cognition doesnt exist at all, since given that theres no subject to be characterized (lakya), owing to the impossibility of [its] establishment by a separate svalakaa there is no possibility of the operation of a characteristic without a locus76. And thus [it says] in the Venerable Questions of Ratnacua [Sutra]77:
Not seeing thought, he [the bodhisattva] investigates the stream of thought [as to] whence it has its arising78. Its [arising] is thus: Thought arises when
kicid aprameya yal lakyate neneti vyapadisyata iti. Atha tad api karmasadhana; tada tasyanyena karaena bhavitavya. Janantarasya karaabhavaparikalpanayam ania doas [de Jong] ca-apadyate. Ruegg, citing Pramasasamuccaya i.12, notes (p.108, n.191): The term ania is here equivalent to anavastha. La Valle Poussins mss. read anityadoa. The Tibetan (thug pa med pa) recommends La Valle Poussins emendation to anavasthadoa. The point, in any case, is clear. 74 That is, perhaps svasavitti could be proposed as that in virtue of which vijana could count simultaneously as the instrument and the object in acts of cognizing. 75 Cf., especially, Madhyamakavatara 6.72-78 (La Valle Poussin 1970a:166-174). 76 61.10-62.3: Atha manyase svasavittir asti. Tata svasavittya grahaat karmataya satyam asty eva prameyantarbhava iti. Ucyate: vistarea Madhyamakavatare svasavittiniedhat, svalakaa [p.62] svalakaantarea lakyate tad api svasavittya iti na yujyate. Api ca, tad api nama jana svalakaavyatirekeasiddher asabhaval lakyabhave nirasrayalakaapravttyasabhavat sarvatha nastiti kuta svasavitti? Ruegg translates: Furthermore, this putative knowledge also does not exist at all: there is indeed no existence [of it] because [this jana] is unestablished separately from the svalakaa; [and] because, when there exists nothing characterized, a lakaa, lacking [such] a ground will not come into operation (2002:109). This suggests that there are two separate reasons given here; I take it, rather, that the locative absolute here subordinates the reason given in the ablative. 77 Tohoku 91. 78 Stcherbatsky (1927:153, n.6) who follows Burnouf's reading, noted by La Valle Poussin, n.4 is puzzled by asamanupasyan, and emends to cittam samanupasyan. But the Tibetan (yang dag par rjes su ma mthong bas) suggests that La Valle Poussins reading is correct. Siderits (who follows Vaidya's edition) translates: How does the arising of consciousness, not perceiving what possesses consciousness, investigate the stream
435
there is an intentional object (alambana). Is it, then, [the case that] the intentional object is one thing, and the thought another? Or is that which is the intentional object precisely the [same as] the thought? If, first of all, the intentional object is one thing and the thought another, then there will obtain [its] being two thoughts (dvicittata). Or if the intentional object itself is the thought, then how does thought perceive thought? For thought does not perceive thought. Just as a sword-edge cannot be cut by that same sword-edge, [p.63] and a finger-tip cannot be touched by that same finger-tip, in just the same way, a [moment of] thought cannot be seen by that same thought. For one who is thus properly disciplined79, thought has the quality of not abiding (anavasthanata), the quality of being neither interrupted nor eternal (ucchedasasvatata), of not being the paramount self (na kuasthata), of not being causeless, nor of being negated (viruddha) by conditions80, neither from this nor from that, neither this nor that [the bodhisattva] thus knows that stream of thought which [has all of these qualities], that creeping vine of thought (cittalatam), that reality (dharmata) of thought, that unlocatedness of thought, that immovability of thought, that unseen-ness of thought, [the fact of] being the defining characteristic (svalakaatam) of thought; thus does [he] see [this] as suchness (tathata), and [he] does not obstruct it. Thus does [he] realize this analysis of thought, thus does [he] see. This, son of noble family the bodhisattvas consideration of thought with respect to thought is the foundation of mindfulness.
Thus, there is no [faculty of] apperception; [and] since it is non-existent, what is characterized by what?81 And would it be a characteristic by virtue of difference from the subject to be characterized, or by virtue of non-difference? In this regard, if,
of consciousness? (p.137, my emphasis) Thus, he correctly reads asamanupasyan, but mistakes the subject of the sentence. The problem vanishes if we consult the Sikasamuccaya, which quotes precisely this passage (in the Bibliotheca Buddhica edition of Bendall, p.235; cited by de Jong). That the text given by La Valle Poussin is correct is suggested by what there precedes the present quote: Examining thought he [i.e., the bodhisattva] does not see it as internal, he sees it not outside him, nor in the conformations, nor in the elements, nor in the organs of sense. Not seeing thought, he follows the course of thought, asking, Whence does thought arise?. The point, then, is that, after prior investigations, he (the bodhisattva) has failed to find anything answering to the designation thought, and it is this failure which impels the present search into the nature of the stream of thought. 79 tasya eva yonisa prayuktasya; Tib., de di ltar tshul bzhin rab tu sbyor ba la. 80 Ruegg (110): contrary to condition. 81 62.4-63.8: Tatha ca-uktam aryaratnacuaparipcchaya: Sa cittam asamanupasyan, cittadhara paryeate: kutas cittasya utpattir iti? Tasyaiva bhavati. Alambane sati, cittam utpadyate. Tat kim anyad alambanam anyac citta, atha yad evalambana tad eva citta? Yadi tavad anyad alambanam anyac citta, tada dvicittata bhaviyati. Atha yad
436
DAN ARNOLD
on one hand, its by virtue of difference, then because of being different82 from the subject characterized, the characteristic wouldnt be a characteristic, either, as though it were a non-characteristic. And because of [its] being different from the characteristic, the characterized subject wouldnt be a characterized subject, either, as though it were a non-subject. [p.64] In this way, because of being different from the subject to be characterized, the characteristic would have a subject-to-be-characterized with no need for a characteristic and hence, because of being without need of a characteristic, it could not be a subject to be characterized! [It would, then, be] just like a sky-flower83. On the other hand, if subject-to-be-characterized and characteristic were not distinct [from one another], then, because of [its] not being distinct from the characteristic, the subjects being a subject (lakyata) is forevalambana tad eva citta, tat katha citta [de Jong; cf., Bendall 1970:235] citta samanupasyati? Na ca citta citta samanupasyati. Tadyathapi nama taya-evasidharaya saivasidhara na [p.63] sakyate chettu. Na tenaivangulyagrea tadevangulyagra sakyate sprau. Evam eva na tenaiva cittena tad eva citta sakya drau. Tasyaiva yonisa prayuktasya ya cittasyanavasthanata-anucchedasasvatata na kuasthata nahetuki na pratyayaviruddha na tato nanyato na saiva nanya, ta cittadhara cittalata cittadharmata cittanavasthitata cittapracarata cittadsyata cittasvalakaata, tatha janati tatha pasyati yatha tathata na ca virodhayati [de Jong]. Ta ca cittavivekata tatha prajanati tatha pasyati. Iya kulaputra [bodhisattvasya; per Tibetan] citte cittanupasyana smtyupasthanam iti. Tad eva nasti svasavittis, tadabhavat ki kena lakyate? Candrakirtis conclusion here (what is characterized by what?) makes clear that the discussion of svasavitti has in this context been chiefly meant to address the possibility of there being something essentially self-characterizing of there being, that is, at least some example of a characteristic (lakaa) that is not the characteristic of anything (which is how Dignaga must understand svalakaas). The critique of svasavitti, like that of svalakaa, thus chiefly turns (like many characteristically Sanskritic arguments) on eminently grammatical presuppositions. In both cases, the point that Dignaga wants to salvage is shown to require that there be some verbally expressible action (characterizing, cognizing) that is not the characterizing or cognizing of anything which, on the karaka-analysis of actions, is incoherent. 82 In characteristically Buddhist (not to say Madhyamika) fashion, Candrakirti here makes an argument that depends on taking different to mean altogether unrelated. 83 63.9-2: Ki ca, bhedena va tal lakaa lakyat syad, abhedena va. Tatra yadi tavad bhedena, tada lakyad bhinnatvad, alakaaval lakaam api na lakaa. Lakaac ca bhinnatvad, alakyaval lakyam api na [p.64] lakya. Tatha lakyad bhinnatval, lakaasya lakaanirapeka lakya syat; tatas ca na tal lakya, lakaanirapekatvat, khapupavat. Something like a sky-flower would, of course, be altogether non-existent, which therefore simply could not have any properties or characteristics.
437
feited, as though [the subject] were itself the characteristic84. And because of [its] not being distinct from the subject to be characterized, the characteristic would not be one whose essence was that of a characteristic, either, as though it were itself the subject85. As it is said [in Nagarjunas Lokatitastava]: If the characteristic were other than the subject to be characterized, then the subject to be characterized would be without characteristic; [and] it is clearly admitted by you that if there is no difference [between them], then neither one exists86. And with respect to establishment of subject and characteristic, there is no other way than as being the same or different. Thus, [Nagarjuna] will say [in MMK 2.21]: How can there be [any] establishment of these two when their establishment is neither as being the same or different?87 Alternatively, if it is said that there will be establishment [of lakya and lakaa] as being ineffable (avacyata), [we respond that] it is not so88. For
Ruegg (2002:112): in the same way as the nature of the lakaa. Ruegg (ibid.): just like the nature of a lakya. 86 Lokatitastava, verse 11; see Lindtner 1987:132. 87 64.2-9: Athabhinne lakyalakae, tada lakaad avyatiriktatval lakaasvatmavad vihiyate lakyasya lakyata. Lakyac ca avyatiriktatval lakyasvatmaval lakaam api na lakaasvabhava. Yatha cokta: Lakyal lakaam anyac cet, syat tal lakyam alakaa; tayor abhavo nanyatve vipaa kathita tvaya iti. Na ca vina tattvanyatvena lakyalakaasiddhav anya gatir asti. Tatha ca vakyati: Ekibhavena va siddhir nanabhavena va yayo, na vidyate; tayo siddhi katha nu khalu vidyate iti. This passage, it seems to me, represents what is not only a characteristically Madhyamika display of linguistic pyrotechnics, but one that is characteristically Sanskritic, in general. (For insightful reflections on the extent to which Sanskritic philosophy is motivated by grammatical and linguistic categories, see Ingalls 1954.) While this type of argument is likely to strike the non-Sanskritic reader as rather underwhelming, it should be remembered that Candrakirtis overriding concern here is with how words are conventionally used, and that this all represents an eminently conventional sort of discourse. It seems to me that the conceptual force of this particular passage is much the same as that of his opening rejoinder (i.e., at 59.7-9; n.52, above). The characteristically Madhyamika deployment of such an argument is similarly on display in Nagarjunas Vigrahavyavartani; thus, with respect to the mutually reciprocal terms pramaa and prameya, Nagarjuna is there concerned to argue (as Oetke puts it) that means of knowledge cannot be what they are, namely means of knowledge, without the existence of that for which they are means, whereas the objects of knowledge cannot be what they are, i.e. prameyas, if there are no pramaas. (Oetke 2003b: 144n) Indeed, all that Madhyamikas are finally concerned to show, in a sense, is that any proposed explanatory terms turn out to be constitutively relational in which case, none can be thought to provide any ultimate explanatory purchase on the phenomena putatively explained thereby. 88 Here, Candrakirti may have in mind Dignagas characterization of svalakaas as
84 85
438
DAN ARNOLD
ineffability, by definition (nama), obtains [only] when there is no recognition of the mutual classification of terms; and where there is no recognition of [such mutual] classification, there is complete absence of these two, as well for there is no possibility of specifying, according to the difference [between them] (viseatas), this is the characteristic, this is the subject. Therefore, there is no establishment as being ineffable, either89. Moreover, if cognition is the instrument with respect to the determination of an object, what is the agent? For without an agent, there is no [p.65] possibility of instruments and so forth, just as in [the case of] the action of cutting [wood]90. Then [perhaps] it is imagined that in this case, thought (citta) has the quality of agency. But this doesnt make sense, either, since, [on your own theory,] the function of thought is apprehension of a bare object (arthamatradarsana); apprehension of the qualifications (visea) of an object [is the function] of [other] derivative mental operations (caitasa) this based on [your] acceptance [of the authoritative text which says that] in this regard, apprehension of an object is perceptual cognition (vijana), while derivative mental operations concern its qualifications91.
indefinable (avyapadesya), etc. (cf., e.g., Pramaasamuccayavtti ad 1.17). Dignagas point in so characterizing (!) svalakaas seems to be simply (but in the end, radically) that unique particulars cannot themselves be the referents of words. Candrakirti, however, introduces this move as specifically meant to explain how there could, in fact, be some way, other than by identity and non-identity (ekibhavena or nanabhavena), of establishing how the lakaa and lakya of his svalakaa might be related. Cf., Ruegg, 2002:113, n.202. 89 64.10-13: Atha-avacyataya [de Jong] siddhir bhaviyatiti cen, na etad eva. Avacyata hi nama parasparavibhagaparijanabhave sati bhavati. Yatra ca vibhagaparijana na-asti, tatra ida lakaam, ida lakyam iti viseata paricchedasabhave sati dvayor apy abhava eva-iti. Tasmad avacyataya-api na-asti siddhi. 90 64.14-65.1: Api ca, yadi jana karaa viayasya paricchede, ka karta? Na ca kartaram antareasti karaadina sabhava chidikriyayam iva. Candrakirtis point here again, an eminently Sanskritic one follows the standard analyses of the Sanskrit grammarians, for whom any action can be analyzed into the terms required to express it as a semantically complete verbal construction. Candrakirti thus has in mind the karaka analysis of sentences on which, cf., e.g., Matilal 1990:40-48. 91 65.1-4: Atha cittasya tatra karttva parikalpyate; tad api na yukta, yasmad arthamatradarsana cittasya vyaparo, rthavisea[darsana] caitasana. Tatraarthadir vijana, tadvisee tu caitasa ity abhyupagmat. Candrakirti here quotes Madhyantavibhaga 1.8 (Pandeya 1999:27). (Ruegg [p.113] erro-
439
For when one governing action (pradhanakriya) is to be effected, instruments and so forth92 have their instrumentality and so forth because of the acceptance of their being subordinate, by virtue of [their] respectively (yathasvam) performing subordinate actions (guakriya)93. But in this case, cognition (jana) and perceptual cognition (vijana) do not have one principal function [in common]. Rather, the principal function of perceptual cognition is determination of a bare object (arthamatraparicchitti), while that of cognition is determination of its qualifications (visea); hence, cognition does not have any instrumentality, nor does thought have any agency. This, then, is the problem94. [Objection:] Well, perhaps it could be that, since scripture says all dharmas are without self, there is no existence of any agent whatsoever
neously attributes this text to Sthiramati; while Sthiramati wrote a ika on Vasubandhus bhaya on the Madhyantavibhaga, the root text here quoted is traditionally attributed to Maitreya.) Candrakirti clearly alludes to Dignagas recurrent point that the distinguishing of separate visea and viseya (qualification and thing qualified) is a constitutively conceptual operation in which case, perception can never itself register such a distinction; cf., e.g., Pramaasamuccaya 1.23. In regard to the way the categories in play here all line up, La Valle Poussin (p.65, n.3) offers many useful textual citations (starting with the Nyayabinduika), summarizing thus: Soient les quivalences: cittam = vijanam = nirvikalpakajanam = arthamatragrahi; caitasa = janam = savikalpakajanam = arthaviseagrahi. See also n.59, above, citing Bruce Cameron Halls observation that the Abhidharmikas vijana basically corresponds to what Dignaga calls pratyaka, while the Abhidharmikas saja corresponds to anumana. Candrakirtis text here seems to recommend Halls observation, effectively attributing to his interlocutor a usage according to which vijana=pratyaka. This is further clarified in the immediately ensuing passage. Regarding Candrakirtis textual citation, Ruegg (2002:113, n.203) comments: Candrakirtis reference here to a major source for the Vijanavada in support of his own view is noteworthy; this might suggest that here his opponent was a Vijanavadin. Indeed, the Madhyantavibhaga seems a perfectly natural text for Dignaga to defer to (and thus, for Candrakirti to cite as something that Dignaga ought not to contradict). Cf., Hattori 1968: 101102 (n.I.61). 92 That is, the various karakas required to express any action in the form of a semantically complete verbal expression. 93 Rueggs reading of yathasvam is unintelligible to me: in virtue of assuming subsidiarity through effecting a subordinate activity in accord with the possessed (2002:114). 94 65.4-8: Ekasya hi pradhanakriyaya sadhyaya yathasva guakriyanirvttidvarea-angibhavopagamat [here, I decline to accept de Jongs emendation: angabhavopagamat] karaadina karaaditva. Na ca-iha janavijanayor eka pradhanakriya, ki tarhy arthamatraparicchittir vijanasya pradhanakriya, janasya tv arthaviseapariccheda; iti nasti janasya karaatva, napi cittasya karttva. Tatas ca sa eva doa.
440
DAN ARNOLD
hence, even without an agent, ordinary discourse, involving verbs and other [parts of speech,] does transpire95. [Response:] This isnt right, either, for you have incorrectly ascertained the sense of scripture. This is explained in the Madhyamakavatara.96 [Objection:] [p.66] Even when there is no possibility of qualifiers (viseaa) that are separate from a body or a head as [in the expressions] the body of a statue (silaputraka)97 or the head of Rahu
95 65.9-:10: Atha syat, anatmana sarvadharma ity agamat, kartu sarvathabhavat, kartaram antareapi vidyata eva kriyadivyavahara iti. Ruegg: Therefore, there is indeed found to exist a transactional-pragmatic usage that relates to an act, etc., even without an agent (2002:114). 96 65.10-11: Etad api nasti, agamasya samyagarthanavadharaat. Etac cokta [de Jong] Madhyamkavatare. Ruegg counter-intuitively reads artha here as though it meant something like alambana: For no correct intentional object has been specified by the [cited] scriptural testimony (2002: 114). Stcherbatsky (2002:157, n.8) cites Madhyamakavatara 6.68, ff., as what is likely referred to, although I dont see how the passages he cites relate to the discussion at hand. Ruegg (2002:114, n.205) more helpfully cites, instead, the bhaya on Madhyamakavatara 6.76, which argues that Dignagas notion of svasavitti is incoherent given typical Sanskritic analyses of the verbal expression of actions though the relevance of this to the present point is not altogether clear to me. I confess, though, to being unable to venture an alternative suggestion. 97 The primary sense of the word silaputraka is millstone or pestle, which is reflected in the Tibetan translation (mchi gu). If (with Ruegg [2002:115]) we take it that way, the point of this example would differ slightly from that of the other and indeed, would not be altogether clear. The point would perhaps be that, insofar as the word involves a semantic unit that ordinarily refers to persons (i.e., putra, such that the words nirukti makes it mean something like stone boy), one might be inclined to suppose that the body of such is, like the body of a person, animate hence, the force of the subsequent part where were told that the akanka that goes with this word is buddhi, intellect. Thus, the reason a silaputraka just is a body is that it is inanimate (whereas a statue would only be a body if it happened to be a headless statue). However, it seems to me preferable to follow Stcherbatsky (1927:158) in reading this to mean statue in which case, the point of the example is exactly the same as that of the Rahus head example (where the mythological Rahu in question is a celestial being who, having been beheaded, now exists only as a head); viz., both cases involve a genitive relation between two referents when there is in fact only one thing (i.e., a statue just is a body). The example of the expression Rahus head is common in Indian philosophy; cf., e.g., the usage attributed, in the Sarvadarsanasagraha, to the (materialist) Carvakas, who point out that expressions like my body ought not to be taken as evidence of a really existing subject of the genitive; rather, such expressions are, like Rahus head, merely figurative (Sarvadarsanasagraha, p.2: mama sariram iti vyavaharo raho sira ityadivad aupacarika).
441
there is [nevertheless] a relation of qualifier and qualified; just as [in that case], here, too, there will be [a relation] even when there is no possibility of any earth apart from its svalakaa, [so that we are, after all, in a position to make sense of the familiar expression,] earth's svalakaa.98 [Response:] This isnt so, because [these cases] are not the same. For the use of words like body and head depends on other associated categories, such as, [in the case of heads,] intellect, etc., and, [in the case of bodies,] hands, etc. That being the case99, the production of an idea based only on the words body or head creates a semantic expectation regarding the other associated categories100, [such that one expects to know] whose body? whose head? Another [person], with a desire to preclude connection with any other qualifiers101, removes an interlocutors semantic expectation by suggesting the qualifications that are statues and Rahu102 [a suggestion that] is in conformity with mundane convention (saketa). This makes sense. But in the present case, where there is no
98 66.1-3: Athapi syat: Yatha silaputrakasya sarira raho sira iti, sarirasirovyatiriktaviseaasabhave pi, viseaaviseyabhavo sti. Eva pthivya svalakaam iti, svalakaavyatiriktapthivyasabhave pi, bhaviyatiti. I have, once again, left these occurrences of the word svalakaa untranslated in order to reflect the fact that Dignaga is here simply reporting the attested example, while remaining neutral with respect to how we understand the word. Siderits observes that it should be pointed out that here the opponent has reverted to the traditional usage of svalakaa, as meaning own defining characteristic; this is made clear in his reference to hardness as the svalakaa of earth (1981:142). But I think we should understand Candrakirtis interlocutor as simply reporting the example that Candrakirti has challenged him to accommodate; thus, Dignagas task here is to show that the word can mean what he takes it to mean (unique particular), and yet make sense of this attested usage. Naturally, it favors Candrakirtis point that the examples he adduces can only be translated using the expression defining characteristic. 99 The first phrase is all a locative absolute, with the subject of the phrase here translated (pravtti) actually occurring in the locative. 100 More literally, the arising of this idea only functions along with a semantic expectation regarding the other associated categories (buddhyupajanana sahacaripadarthantarasakanka eva vartate). On akaka (semantic expectation), see, inter alia, Matilal (1990:50, 109-10), Abhyankar (1977:53). 101 That is, one strives to eliminate his interlocutors syntactic expectation out of a desire to render the reference more precise, to preclude connection with some other possible referent. 102 That is, these terms, when related to them in the genitive case, qualify (respectively) body and head.
442
DAN ARNOLD
possibility of earth and so forth apart from [defining characteristics] such as resistance103, the relation of qualifier and qualified doesnt make sense104.
103 Candrakirtis point is that the conventional understanding of a defining characteristic (svalakaa) is not that it qualifies some particular example of the kind in question (as, e.g., red earth), but rather, that it makes something an example of that kind in the first place. Thus, adjectival qualification (viseaa) is called for only when there is some syntactic expectation (akanka), such that we need to know more in order to know precisely which token of some type is being picked out. In contrast, since there cannot meaningfully be any earth which is not earth by definition which is not, that is, possessed of the characteristic that makes it an instance of earth we do not, when encountering some instance of resistance, wonder what it belongs to; for when one encounters an instance of earth, one just is encountering an instance of resistance. This is just what it means for the latter to be a defining characteristic of the former. This point can be understood as counting against Dignagas contention that perceptual cognition affords access to uninterpreted data; for Candrakirtis argument here advances the point that we invariably encounter things as they are defined. That is, tokens of the type earth are invariably encountered under a description (viz., as hard or resistant). 104 Prasannapada 66.3-8: Naitad evam, atulyatvat. Sarirasirasabdayor hi buddhyadipayadivatsahabhavipadarthantarasapekatapravttau, sarirasirasabdamatralambano buddhyupajanana sahacaripadarthantarasakanka eva vartate. Kasya sarira, kasya sira iti? Itaro pi viseaantarasabandhaniracikiraya silaputrakarahuviseaadhvanina laukikasaketanuvidyayina pratipattu kankam upahantiti yukta. Iha tu kahinyadivyatiriktapthivyadyasabhave sati na yukto viseaaviseyabhava. This is the passage discussed by Thurman and Eckel, both of whom follow Tsong-kha-pa in taking Candrakirtis target here to be Bhavaviveka; cf., n.11, above. We can understand why Tsong-kha-pa sees here an engagement with Bhavaviveka if we appreciate that on Tsong-kha-pas view, the distinction between the Svatantrika Madhyamaka of Bhavaviveka and Candrakirtis Prasangika Madhyamaka centrally involves the concept of svalakaa. Specifically, Tsong-kha-pa thinks that Bhavaviveka must accept the existents posited by an opponent as being established by virtue of svalakaa (rang gi mtshan nyid kyis grub pa = svalakaena siddha). But Tsong-kha-pas understanding of this issue involves a sense of svalakaa that is, I think, not present in Candrakirti. Thus, we can note that Thurman (translating Tsong-kha-pa) is right to see the present discussion as turning on different understandings of the word svalakaa: the intrinsic identity (svalakaa) involved in (this sort of) intrinsically identifiable status is altogether quite different from the ultimate particular (svalakaa) explained precisely as functional capacity in the logicians treatises, and from the defining characteristic (svalakaa) explained as that which characterizes (something as) different from everything else, such as heat in the case of fire, in the Abhidharma Scripture, etc. (Thurman 1991: 292) But it is really only the latter two senses of svalakaa that are in play in our text from the Prasannapada, with Tsong-kha-pa himself having introduced (in the first occurrence reflected in Thurmans translation) an additional sense. On this point, cf., Ruegg 2004:338-9.
443
[Objection:] Because of the acceptance by non-Buddhists of distinct subjects (lakya), in accordance with that, [our] definition of characteristic is without fault105. [Response:] [p.67]: This isnt so; for it is not suitable to accept, with regard to your own occasion (svasamaye)106, the categories imagined by non-Buddhists, which are devoid of arguments (yuktividhura); for you would have to admit, as well, [their] additional [list of what count as] reliable warrants, and so forth107. Moreover, because of the real existence108 of the qualifier, familiar without analysis, which is a statue109 [conventionally described as] an appropriator (upadat) whose appropriated basis (upadana) is a body, [a relation] that is included in ordinary discourse and because of the real existence of the [qualifier, familiar without analysis]110, which is Rahu, [conventionally described as an] appropriator whose appropriated basis is a head [because of the real existence of these,] just as [in the case of] derivatively [existent entities] like the person, this example doesnt make sense111.
105 66.9: Tirthikair vyatiriktalakyabhyupagamat, tadanurodhena viseaabhidhanam aduam iti cet. 106 Ruegg (2002:116): into ones own doctrine, reading svamata per the Tibetan rang gi gzhung lugs. 107 67.1-2: Na etad eva; na hi tirthikaparikalpita yuktividhura padartha svasamaye bhyupagantu nyayya, pramaantarader apy abhyupagamaprasangat. Of course, this is an unwanted consequence only for Candrakirtis interlocutor, since Candrakirti himself will, in fact, end by endorsing (as conventionally valid, at least) the Naiyayika list of pramaas; cf., 75.6-8 (n.182) below. 108 Ruegg (2002:116): given the [designational] existence. Rueggs insertion is understandable, given the unexpected nature of this claim from a Buddhist; but the Sanskrit (sadbhavat) is clearly stronger. Candrakirti qualifies this point, as expected, presently. 109 Again, the statue is here a qualifier because, in the genitive case, it qualifies the word body, removing our semantic expectation to know whose body is being referred to. 110 I take the expression viseaasya-avicaraprasiddhasya to govern both examples. 111 67.3-5: Api ca pudgaladiprajaptivat, sasariropadanasya silaputrakasyopadatur laukikavyavaharangabhutasya viseaasya-avicaraprasiddhasya sadbhavat, sira-upadanasya ca rahor upadatu sadbhavad, ayuktam etan nidarsana. Note that the Tibetan makes the last sentence more clear, first translating all of the genitives, and then concluding: gzhan yang lus kyi rten can khyad par byed pa jig rten pai tha snyad kyi yan lag tu gyur pa ma brtags na grub pa rten pa po mchi gu dang / mgoi rten can brten pa po sgra gcan ni gang zag la sogs par brtags pa ltar yod pai phyir dpe di rigs pa ma yin no. This makes clear that, among other things, sadbhava is supposed to construe with pudgaladiprajapti, too. I have developed my interpretation of the issues relating to this passage at length in Arnold (2005, Chapter 6).
444
DAN ARNOLD
[Objection:] In fact, the example is established, since, because of the non-establishment of any other object apart from the body and the head, there is apprehension only of these [a body and a head]112. [Response:] This isnt so, because such critical analysis doesnt operate in ordinary discourse, and because the existence of ordinary categories is not based on such critical analysis. Just as a self, critically considered, is impossible as [something] distinct from form and so forth113, but nonetheless, relative to the aggregates (skandhan upadaya)114, conventionally has existence115 so, too, in the case of Rahu and the statue. Hence, there is no establishment of the example116. In the same way, even if, on the part of things like earth, there is no subject [when] being considered apart from [defining characteristics] like resistance, and [even if the] characteristic, when separate from the subject, is without a locus
112 67.6-7: Sarirasirovyatiriktasya-arthantarasya-asiddhes, tanmatrasya-upalambhat, siddham eva nidarsanam iti cet. That is, the interlocutor here suggests that the various properties (or qualifiers) of any unique particular are not among the things perceived, and hence cannot be thought real in which case, he may after all salvage his understanding of svalakaa as the unique particulars that alone are perceived (and therefore still say that an expression like earths svalakaa is to be understood as having only one real referent). 113 That is, apart from such analytic categories as the skandhas. 114 This phrase, I have argued in Chapter 6 of Arnold (2005), is central to understanding Nagarjunas constitutively Madhyamika category of upadaya prajapti. 115 Ruegg (2002:117): it exists in supported-dependence. It might be thought counter-intuitive that the selfs existence (qualified as skandhan upadaya) is allowed as conventional; for if the whole Buddhist critique of a self is to have any purchase, it would seem that the convention in the matter would really be that the self exists atmana or svabhavena (that is, that it exists in itself or essentially). Perhaps it was this thought that led Siderits to translate this passage thus: but by worldly convention there is the reality of that, not depending on the skandhas (1981:144; my emphasis) as though, presumably (but impossibly), skandhan upadaya were to be construed as a compound (skandha-anupadaya). But Candrakirtis point here is not that the selfs existing relative to the aggregates is the content of the convention; rather, his point is just that, given the aggregates as a basis of imputation, there can arise the convention that the self exists. 116 That is, the interlocutor had invoked these examples (a statues body, Rahus head) as meant to show that there are cases where we speak as though there were two things, even though we all know there is only one real referent; and Candrakirti has responded that the salient point about examples like Rahus head is not that there are two terms but only one real referent, but simply that there are two terms only insofar as convention requires it. The fact that these examples are in this respect like selves (or pudgaladiprajapti) should, Candrakirti thinks, have most Buddhists assent that such expressions therefore ought not to be thought ultimately to involve any real referents.
445
nevertheless, this is the convention117. The teacher [Nagarjuna]118 settled the matter by establishment [of all these categories] as simply being mutually interdependent (parasparapekamatrataya)119. And this is necessarily to be accepted in this way; for otherwise, the conventional [p.68] would not be [characterized by its being] free from reasoning this would be reality, not the conventional120. And it is not [the case that] there is the impossibility only of things like statues when they are investigated by reasoning121. Rather, according to the argument that is going to be set forth, there is no possibility of form and feelings and so forth122, either; hence, their existence, too, like that of the statue, would have to be accepted as conventional. And this is not how [you accept them]; hence, [your position is] false.123 This presentation of rel117 That is, the reference of words is explicable with reference only to what convention requires, and cannot itself be explained by anything more real than such conventions which, on Candrakirtis reading, is precisely what Buddhists like Dignaga are after. 118 Ruegg (118): Teachers have propounded. But it seems clear that Candrakirti is here (in eminently conventional fashion) using the honorific plural to refer to Nagarjuna. 119 67.7-12: Na etad eva. Laukike vyavahara itthavicarapravtter avicaratas ca laukikapadarthanam astitvat. Yathaiva hi rupadivyatirekea vicaryamaa atma na sabhavati, api ca lokasavtya skandhan upadaya-asya-astitva, eva rahusilaputrakayor apiti nasti nidarsanasiddhi. Eva pthivyadina yady api kahinyadivyatirikta vicaryamaa lakya nasti, lakyavyatirekea ca lakaa nirasraya, tathapi savtir eeti, parasparapekamatrataya [de Jong] siddhya siddhi vyavasthapayababhuvur acarya. 120 Ruegg (2002:118): otherwise, savti would not be deprived of a justified ground; and it would then be reality itself, and not savti. Candrakirtis point seems simply to be that critical analysis is, ipso facto, constitutive not of the conventional but of (a search for) ultimate truth; thus, if it were thought that conventionally admitted existents could withstand this kind of ultimate analysis, then they could not be described as conventionally admitted, and would instead have to be said to exist ultimately. Of course, Candrakirtis characteristically Madhyamika point is that there is nothing that can thus withstand analysis. 121 Ruegg: Nor is there [simple] non-existence of a silaputraka and the like that are in fact being analytically investigated as to [their having a] justified ground (2002:118). This misconstrues the eva (in fact) as though it qualified vidyamana (rather than, as I take it, silaputraka). Candrakirti clearly means to emphasize that once we open up a critical analysis, it is not only conventionally admitted existents that are undermined, but the putatively ultimately existent (paramarthasat) analytic categories of the Abhidharma tradition, as well. 122 That is, the skandhas (of which these are the first two), which are among the subjects to be treated in the remainder of the MMK. 123 67.12-68.4: Avasya caitad evam abhyupeya, anyatha hi savtir upa- [p.68]
446
DAN ARNOLD
ative indication (upadaya prajapti) is also extensively taught in the Madhyamakavatara, so that should be consulted, too124.
pattya na viyujyeta, tada-iyam [de Jong] tattvam eva syan, na savti. Na ca upapattya vicaryamaana silaputrakadinam eva-asabhava, ki tarhi vakyamaaya yuktya rupavedanadinam api nasti sabhava iti; team api savtya silaputraka iva-astitvam astheyam syat. The last underlined portion here reflects a possible textual problem. I have made what seems to me the best sense of this passage by refusing an emendation proposed by La Valle Poussin, who follows some versions of the Tibetan (de dag kyang mchi gu la sogs pa bzhin du kun rdzob tu yod pa ma yin pa nyid du khas blangs par gyur na) in suggesting the reading: team api savtya silaputrakadivan nastitvam astheya syat... (Vaidya [1960:23] adopts La Valle Poussins emendation. De Jong [1978] does not comment.) This gives the opposite of my sentence: They, too, like statues and so forth, would have to be accepted as not existing conventionally. This is a conceptually possible reading, according to which Candrakirtis point would have to be that even the conventional existence of such things would have to be disallowed if it were thought (counterfactually) that the conventional could be characterized by critical examination. It would, then, be the latter that Candrakirti here means to deny; for what cannot be doubted, in any case, is that the skandhas fail to survive critical examination. The reading I prefer, though, seems more straightforwardly to follow what precedes it, as Candrakirtis point is instead that the merely conventional existence of the skandhas is precisely what we have to accept. I take this as stated counterfactually, then, insofar as it is a conclusion that Candrakirti thinks his interlocutor wishes to avoid (though of course Dignagas generally Abhidharmika idea that there is an enumerable set of ultimately existent entities involves only svalakaas, not the skandhas). (For a conceptually similar passage, see n.141, below.) I would venture that it is the optative here that gives pause; for this makes the sentence counterfactual, but it is not immediately clear (given the characteristically laconic na caitad evam ity asad etat that follows) what is counterfactual about it. My reading is warranted, though, by all of the manuscripts available to La Valle Poussin (cf., his n.3, p.68). Ruegg (2002:118, with n.217) reads the Sanskrit as I do, noting some divergence between different editions of the Tibetan canon, with the sDe-dge edition not warranting La Valle Poussins emendation. Ruegg translates, however, very differently (although plausibly): Hence, as in the case of the silaputraka, etc., on the surface-level their existence is to be accepted. But since it is not [really] so, [in ultimate reality] it is non-existent. (Ruegg 2002:118; my emphasis) 124 This sentence is not in the available Sanskrit manuscripts, but is preserved in the Tibetan (and quoted by Tsong-kha-pa as occurring in Candrakirtis text; cf., Thurman 1991: 295): brten nas brtags par rnam par bzhag pa di yang dbu ma la jug ba las rgyas par bstan pas de nyid las yongs su btsal bar bya'o. Candrakirti refers us to the Madhyamakavatara throughout the Prasannapada, so that the reference given here in the Tibetan is not at all out of place though it is not immediately clear how much of the foregoing discussion is to be included as having been concerned with a presentation of upadaya prajapti. Presumably, though, Candrakirti refers back to where he first exemplifies what he takes as the conventional usage of svalakaa (p.60.5), and it is quite possible that he means to characterize the entire discussion of svalakaas as concerning upadaya prajapti. Ruegg (2002:119, n.218) cites Madhyamakavatara 6.120, ff., for other discussions of upadaya prajapti. In a footnote to Tsong-kha-pas reiteration of Candrakirtis con-
447
[Objection:] Whats the use of this hair-splitting? For we do not say that all discourse involving warrants and warrantable objects is true; rather, what is familiar in the world is [all that is] established by this argument125. We respond: We, too, say, Whats the use of this hair-splitting, which delves into ordinary discourse? Let it be! Until there is understanding of reality, the conventional its existence (sattaka) come into being (atmabhava) as projected by nothing but error is, for those who desire liberation, the cause of the accumulation of the roots of merit that convey [one] to liberation126. [p.69] But having introduced reasoning at some point, you incoherently (anyayato) destroy it127, because of being one whose intellect is ignorant of the distinction between conventional and ultimate truth. I am the one who, based on skill in settling conventional truth, situate myself in the ordinary perspective. Like a respected elder (lokavddha), I overturn one argument dedicated to the refutation of one part of the conventional by another argument and in so doing, I refute only you, who are deviating from the conduct of the world. But [I do] not [refute] the conventional128.
cluding sentence, Thurman (1991: 295, n.19) refers us instead to Madhyamakavatara 6.32, ff. I take 6.158-165 as the section to which Candrakirti here alludes though of course it is difficult to be certain. I have developed my analysis of the expression upadaya prajapti in Arnold (2005), Chapter 6. 125 68.5-6: Atha syat: Kim anaya sukmekiya? Naiva hi vaya sarvapramaaprameyavyavahara satyam ity acakmahe, ki tu lokaprasiddhir eamuna nyayena vyavasthapyata iti. 126 Ruegg: What, indeed, is the use of this subtle [investigation] that introduces into [sic] the transactional-pragmatic usage of ordinary folk in the world? To begin with, let there be this savti wherein the existence of an entity is acquired through mere misapprehension, [but which may none the less be] the motivating cause so long as there is [yet] no knowledge of reality. (2002:119-20) 127 Thus I take eta to pick up upapatti (reasoning), not (from the preceding sentence) savti. If, as it would be possible to do, we read eta as thus referring back to savti, Candrakirtis point would be the similar one that his interlocutor undermines the conventional, simply by introducing a putatively probative argument with the conventional, Candrakirti has already said, being constitutively lacking in critical analysis. On my reading, by contrast, the point is that the problem is with Dignagas replacing the conventional with something else that is not itself conventional (i.e., with a peculiarly technical account thereof, a putatively probative upapatti), while at the same time claiming that his own account is conventional. On this reading, Candrakirti is basically charging Dignaga with self-referential incoherence. 128 68.7-69.5: Ucyate: vayam apy eva bruma: Kim anaya sukmekiya laukikavyavahare vatarikaya? Tihatu tavad ea viparyasamatrasaditatmabhavasattaka savtir
448
DAN ARNOLD
Therefore, if it is ordinary discourse, then there must also be a subject that possesses a characteristic (lakaavallakyeapi bhavitavya)129. And therefore just this is the problem [with your conception]. But in terms of ultimate truth, since there [ultimately] are no subjects (lakyabhavat), this pair of characteristics [i.e., sva- and samanyalakaa] does not exist, either; whence, then, [your] two reliable warrants?130 Now perhaps it is not accepted [by you] that the derivation (vyutpatti) of words thus depends on a connection between action and agent131. This is extremely problematic. You engage in discourse (vyavaharati)132 using those very words whose use (pravtta) depends on a connection between action and agent and yet you do not acknowledge the sense of words as involving things like actions and instruments. You fool! Your sense is bound to a mere fancy133.
mumukua mokavahakakusala-[p.69]-mulopacayahetur, yavan na tattvadhigama iti. Bhavas tv eta savtiparamarthasatyavibhagadurvidagdhabuddhitaya kva cid upapattim avatarya-anyayato nasyati. So ham savtisatyavyavasthavaicakayal laukika eva pake sthitva, savtyekadesanirakaraopakiptopapattyantarantaram [emend to upapattyantaram] upapattyantarea vinivartayan loka vddha [emend to lokavddha] iva lokacarat paribhrasymana bhavantam eva nivartayami, na tu savti. 129 I thus read lakaavallakyea as a compound, and the -vat suffix in the sense of possession and not (as the Tibetan translation construes it) as the indeclinable lakaavat, which would read as in the case of a characteristic, there must also be a subject. (So Ruegg: Therefore, if [this be] worldly transactional-pragmatic usage, then, necessarily, there must exist a lakya in the same way as a lakaa [2002:120]. The Tibetan reads: dei phyir gal te jig rten pai tha snyad du yin na ni, dei tshe mtshan nyid bzhin du mtshan gzhir yang.) Given that Candrakirti has wanted all along to show that Dignagas conception of svalakaa founders on the necessity of admitting that there must be some lakya in which it is instantiated, the reading I have chosen seems to make more sense. Thanks to Rick Nance for pointing this out to me. 130 69.5-7: Tasmad yadi laukiko vyavaharas, tada-avasya lakaavallakyeapi bhavitavya; tatas ca sa eva doa. Atha paramarthas, tada lakyabhaval lakaadvayam api nastiti, kuta pramaadvaya? 131 That is, perhaps Dignaga will deny that the karaka analysis of verbal expressions (conventionally normative in the Sanskritic world) should hold sway. Candrakirti takes this concession as an occasion to restate the extent to which his whole critique of Dignagas account of svalakaas (according to which, svalakaas constitutively neither having any properties, nor being the properties of anything are simply self-characterizing) has been informed chiefly by the characteristically Sanskritic analyses of language that were conventionally normative for his context. 132 Ruegg (121): Your honour engages in transactional-pragmatic activity. Here in particular, I might have liked, You conduct your business; see nn.17, 32, above. 133 69.8-10: Atha sabdanam eva kriyakarakasabandhapurvika vyutpattir nangikriy-
449
And when, in this way, it has not been shown that there are [only] two [kinds of] warrantable objects, then, by virtue of their not having as objects either unique particulars or abstractions, [it follows that] tradition and so forth [can also] have the status of additional reliable warrants134. Moreover, because it doesnt include instances of ordinary discourse like a jar is perceptible,135 and because of the acceptance of the discourse of ordinary people (anarya), [your] definition (lakaa) has insufficient extension136 it doesnt make sense.137 [Objection:] [p.70] Things like color, which are what is appropriated (upadana) as jars, are [said to be] perceptible [simply] because of [their]
ate. Tad idam atikaa. Tair eva kriyakarakasabandhapravttai sabdair bhavan vyavaharati, sabdartha kriyakaraadika ca na-icchatiti. Aho bata-icchamatrapratibaddhapravttita [de Jong] bhavata. Here again, then, a charge of self-referential incoherence: Candrakirti thus urges that his interlocutors own use of language has as a condition of its possibility the very features that his view of svalakaas commits him to denying. 134 69.11-12: Yada ca-eva prameyadvayam avyavasthita, tada [sva]samanyalaka[a]viayatvena-agamadina pramaantaratva. Of course, this is only an unwanted consequence for Dignaga; Candrakirti here approaches his concluding endorsement of the Naiyayika pramaas as representing the best account of our conventional epistemic practices; cf., 75.6-8 (n.182), below. 135 Here, Candrakirti turns to a consideration specifically of Dignagas account of perception (pratyaka). He begins by stating the main point he will be concerned to make: that on the conventional use of the word pratyaka, it is the adjectival sense (perceptible) that is primary. One might also (with Siderits 1981:148, ff.) take the salient point of the example a jar is perceptible to be that it is wholes like jars that are perceptible, and not the foundationalists fleeting sense data. (See also Arnold 2001a:259, where it is taken the same way.) This is, to be sure, as Candrakirti would wish to argue, and it is clearly one upshot of this conventional usage; but Candrakirtis way of making the point is, in characteristically Sanskritic fashion, to emphasize the grammatical point. In fact, pratyaka must be an adjective in the example adduced by Candrakirti; the noun form of the word is neuter, and in Candrakirtis example it has taken the masculine gender of the word (ghaa) that it modifies. (The rule that explains this is cited by Dharmottara, who also criticizes Dignagas etymology; see n.148, below, for the reference.) That Candrakirti has a good claim to expressing the primary sense of the word is reflected in Apte (1992:1085), who gives several adjectival senses first. 136 Ruegg: for the [postulated] lakaa, there exists no (logical-epistemological) pervasion (2002:122) But Candrakirti would not, I think, have in mind this peculiarly technical, dialectical use of the word, which has a quite common sense in grammatical literature; cf., Abhyankar 1977:48 (s.v., avyapti). 137 69.13-14: Ki ca ghaa pratyaka ity evam adikasya laukikavyavaharasyaasagrahad, anaryavyavaharabhyupagamac ca, avyapita lakaasya-iti na yuktam etat.
450
DAN ARNOLD
determinability by the reliable warrant which is perception138. Hence, just as it is taught that the birth of buddhas is bliss [an expression that is understood as] involving figurative reference to the effect with respect to [what is really] the cause in the same way, a jar, even though occasioned by perceptibles like its color, is designated as perceptible, making a figurative reference to the cause with respect to [what is really] the effect139. [Response:] [Appeal to] figurative usage does not make sense with respect to a cognitive object of this kind. For in the world, birth is apprehended as separate from happiness. Indeed, because of its being the cause of many hundreds of evils which is because of its having as its nature the characteristic of [being] compounded it [i.e., birth] is precisely non-bliss. With respect to the sort of object where what is being taught it [i.e., birth] is happiness is incoherent (asabaddha eva), figurative usage makes sense. But in the present case a jar is perceptible there is nothing at all called a jar which is imperceptible, [noth138 That is, Dignaga once again challenged to show how his peculiarly technical usage of words can be reconciled with examples of their ordinary use here suggests that the adjectival sense of the word pratyaka is derivative, and that the word primarily functions to pick out an epistemic faculty. 139 70.1-3: Atha syat: ghaopadananiladaya pratyaka, pratyakapramaaparicchedyatvat (Tib., mngon sum gyi tshad mas yongs su gcad par bya ba yin pai phyir mngon sum yin te). Tatas ca yathaiva karae karyopacara ktva, buddhana sukha utpada iti vyapadisyate, eva pratyakaniladinimittako pi ghaa karye karaopacara ktva pratyaka iti vyapadisyate. The same example (for which, cf., Dhammapada 14.16: sukho buddhana uppado sukha saddhammadesana / sukha sanghassa samaggi samaggana tapo sukho //) is cited and discussed in Vasubandhu's Abhidharmakosabhayam (ad Abhidharmakosa 1.10; Pradhan 1975:7), where it is also an example of karae karyopacara. Cf., also, Nyayabindu 3.2 (and Dharmottaras commentary thereon; Malvania 1971: 120-21), where the case of pararthanumana is similarly considered a figurative usage of the word anumana. Dignagas appeal to upacara in his account of pratyaka is much as Candrakirti here represents it: The word pratyaka is used with respect to three things: the reliable warrant, the awareness [that results from the exercise thereof], and the object [of this awareness]. With respect to these, [the usage designating] the reliable warrant is primary, and the others are secondary (nye bar btags =Skt. aupacarika). In this regard an object is [figuratively] characterized as pratyaka since it is cognized by [the reliable warrant called] pratyaka. (Pramaasamuccayavtti ad 1.41c-d; Tibetan from Hattori, pp.233: mngon sum gyi sgra ni tshad ma dang shes pa dang yul gsum la jug go. de la tshad ma la ni gtso bo yin la, gzhan dag la ni nye bar btags pa yin te: de la yul la ni mngon sum gyi gzhal bya yin pai phyir mngon sum du btags pa yin no.)
451
ing at all] separately apprehended (pthag upalabdha) to which perceptibility could figuratively belong140. If it is said that perceptibility is figurative because of the non-existence of a jar apart from [perceptible qualities] like color, then [appeal to] figurative usage makes even less sense, since there is no basis which is being figuratively described; for the sharpness of a donkeys horn is not [even] figuratively asserted (upacaryate). Moreover, if it is imagined that a jar, which is part of ordinary discourse, has [only] figurative perceptibility since (iti ktva) it doesnt exist apart from things like its color, then since things like color dont exist apart from things like earth, either, the [merely] figurative perceptibility of that color and so forth would also have to be posited. As it is said [in Aryadeva's Catusataka]: [p.71] Just as a pot does not exist as separate from things like its form, so, too, form does not exist as separate from [basic elements] like air, etc141.
140 70.3-7: Na evavidhe viaya upacaro yukta; utpado hi loke sukhavyatirekeopalabdha. Sa ca sasktalakaasvabhavatvad anekadukarasatahetutvad, asukha eva. Sa sukha iti vyapadisyamano sabaddha evety; evavidhe viaye [de Jong] yukta upacara. Ghaa pratyaka ity atra tu, na hi ghao nama kascid yo pratyaka pthagupalabdho yasya-upacarat pratyakatva syat. The point is that recourse to upacara requires two terms: the thing figuratively described, and the thing appealed to so to describe it. One is entitled to (indeed, one must) seek a secondary or figurative meaning, then, only when the two terms in play are such that, given their primary meanings, their association produces an incoherent (or otherwise unexpected) sentence. Thus, if we are to say that a jar is only figuratively perceptible, we must already know that there is, in fact, such a thing as a jar, and that it is simply not really perceptible, such that these two terms (jar and perceptible) were (like utpada and sukha) asabaddha. But in fact, conventional usage reflects the belief that jars are perceptible, so there is no obvious contradiction such as would require recourse to figurative explanation. This point is obscured, I think, in Rueggs translation, which I do not find clear: no such thing as a pot not directly perceptible is separately apprehended which might become directly perceptible through transfer (2002:122). 141 70.8-71.2: Niladivyatiriktasya ghaasya-abhavad aupacarika pratyakatvam iti cet, evam api sutaram upacaro na yukta, upacaryamaasya-asrayasya-abhavat; na hi kharaviae [de Jong] taikyam upacaryate. Api ca, lokavyavaharangabhuto ghao yadi niladivyatirikto nastiti ktva tasya-aupacarika pratyakatva parikalpyate, nanv eva sati pthivyadivyatirekea niladikam api nastiti, nilader asya-aupacarika pratyakatva kalpyata. Yathokta: Rupadivyatirekea yatha kumbho na vidyate, vayvadivyatirekea tatha rupa na vidyate iti. Candrakirti thus argues that an appeal to figurative usage based on its really being the parts of a jar (or jar-appearing sense data) that are perceived is even less promising for Dignaga, since that only opens the way for Candrakirtis characteristic point against the reductionist version of Buddhism. That is, Candrakirti will gladly concede that medium-
452
DAN ARNOLD
Therefore, since [your] definition does not accommodate these examples of ordinary discourse, [your] definition has insufficient extension142. For143 the perceptibility of things like jars and colors and so forth is not accepted from the point of view of one who knows reality (tattvavid apekaya); but according to worldly convention, the perceptibility of jars and so forth is precisely to be accepted! As it is said in the [Catu]sataka: The whole jar, unseen, is present even when only its color is seen; but what knower of reality would say that a jar is [ultimately] perceptible? By this very same reasoning, sweet fragrance, melodious sound, softness all [of these] are [similarly] to be denied by one possessed of supreme intellect144. Moreover, because of the fact that the word perception is expressive of the meaning visible (aparoka), an object that is plainly before us is
sized objects like jars are analytically reducible but once having opened the way for this kind of critical analysis, he will then press the point and urge that there is no irreducible remainder, so that analytic categories like sense data must themselves be understood as dependent. This passage, it seems to me, provides some warrant for my reading of the problematic passage at 68.2-4; cf., n.123, above. The verse cited is Catusataka 14.15; cf., Lang 1986:130-31. 142 That is, it doesnt cover all the usages that a successful definition would have to cover; cf., n.136, above, and Ruegg (2002:123). 143 Ruegg (2002:123) here begins a new paragraph, and indicates that a different point is being addressed. I follow the paragraphs of La Valle Poussins edition, though, and take Candrakirti here to be amplifying the same point. 144 71.3-9: Tasmad evamadikasya lokavyavaharasya lakaena-asagrahad, avyapitaiva lakaasyeti. Tattvavidapekaya hi pratyakatva ghaadina niladina ca naiyate; lokasavtya tv abhyupagantavyam eva pratyakatva ghaadina. Yathokta Satake: Sarva eva ghao do rupe te hi jayate, bruyat kas tattvavinnama ghaa pratyaka ity api? Etenaiva vicarea sugandhi madhura mdu, pratiedhayitavyani sarvay uttamabuddhina iti. The last passage quoted is Catusataka 13.1-2 (Lang 1986:118-9; Tillemans 1990: vol. 1, 175). I take the point to be that, conventionally, one speaks of the whole jar as perceptible, even though it is of course technically true that one only really perceives various aspects of it. It is, nevertheless, said to be perceptible from the conventional perspective, and the contrast is thus with the perspective of a knower of reality (tattvavid) that is, the perspective of a fully realized Buddha, who of course realizes that there is nothing that is ultimately perceptible. This quotation furthers Candrakirti's critique of pratyaka as a privileged pramaa. It is fitting that Candrakirti should begin this section by quoting the Catusataka; for arguments like those advanced here are also developed in chapter 13 of Candrakirti's Catusatakavtti, which is a key source for this argument against pratyaka; in Tillemans 1990, vol.1, pp.175-199 (trans.), and vol.2, pp.60-127 (Tibetan text, Sanskrit fragments); and especially p.277n, 287n.
453
[conventionally said to be] perceptible.145 By virtue of the fact that (iti ktva) the sense organ (akam) is directed towards it, the perceptibility of visible things like jars and color and so forth is established. [p.72] Since a cognition that discerns these [jars, color, etc.] has a perceptible [object] as its cause, [it] is designated [as] being a perception, as [in the case of] a straw- or chaff-fire146. But147 the etymology of one who etymologizes (vyutpadayati) the word perception as [what] is directed towards each sense faculty (akam aka prati vartate)148 doesnt make sense, because of the cognitions not having the sense faculty as its object rather, its object is an object. [Following the etymology of Dignaga,] we should
145 If I rightly understand the section that begins here, Ruegg (2002:124, ff.) has significantly mistaken the dialectical flow of the argument here; for Ruegg introduces this passage as having been spoken by Candrakirtis imagined interlocutor whereas on my reading, this passage advances Candrakirtis preferred account of the word pratyaka. That is, Candrakirti is here elaborating what he takes to be the conventional the primarily adjectival use of the word, according to which cognitions are so called only in the derivative sense that they arise in connection with something perceptible. 146 That is, such fires are similarly so-called because of having straw or chaff as their causes. This is Candrakirtis preferred way of explaining the (nominal) sense of the word pratyaka as denoting an epistemic faculty. Of this move, Tillemans (1990, vol.1: 44) aptly says: By shifting etymologies Candrakirti tries to make perception banal: any consciousness, conceptual or not, caused by a perceptible (pratyaka) object will be termed pratyaka. As we will see below (at p.75.2-4; cf., n.179, below), one of the upshots of this is that, given the range of objects conventionally described as perceptible, it becomes appropriate to say (contra Dignaga) that what Dignaga would consider to be abstractions are among the objects of perception. 147 Ruegg (2002:125) takes this to begin Candrakirtis response to the preceding; on my reading, though, Candrakirti is here building on the same point. More particularly, Candrakirti now proceeds to show that Dignagas preferred etymological explanation of the word cannot be made to fit with the preceding observations which, again, describe Candrakirtis sense of the conventional usage. It would, then, be incoherent to attribute the preceding passage to Dignaga, since Candrakirtis point now is to show that the preceding is precisely what Dignagas usage cannot explain. 148 Hattori notes (1968: 76-77, n.1.11) that Candrakirti here critiques the etymology given by Dignaga in his *Nyayamukha; cf., Ruegg (2002:125, n.233). As Ruegg there notes, Dharmottaras Nyayabinduika also discusses Dignagas etymology, against which Dharmottara proposes his own account of pratyaka interestingly, one that does explain the adjectival sense of the word (though the main objective of Dharmottara and his commentator Durvekamisra is to argue for an etymology that makes it possible for manasa-, yogi-, and svasavedana-pratyaka to count as instances of pratyaka, whereas given the etymology of Dignaga, it only makes sense to think of indriya-pratyaka as properly an example of the genus.) See Malvania 1971:38-39 (where Durvekamisra specifically names Dignaga as the one whose view is there under critique).
454
DAN ARNOLD
[counterfactually characterize the faculty that picks out perceptible objects like jars as] occurring in connection with an object (prati-viayam) or occurring in connection with a thing (praty-artham)149. [Objection:] Even given that the functioning of perceptual cognition (vijana) is dependent upon both [the sense faculty and an object], it is based on conformity with the acuity150 of the basis (asraya) i.e., because perceptual cognitions have the quality of changing as that [basis] changes151 that there is designation [of the epistemic faculty] precisely in terms of the basis [thereof]. [Thus, for example, we speak of] ocular cognition (cakurvijana), [which is named for the eye]. In just the same way, even if it proceeds always towards objects, nevertheless, proceeding always in reliance upon the senses, the cognition is designated in terms of the basis [upon which it thus relies]; hence, it will be [called] perception (praty-akam, with respect to the senses). For it is [com149 71.10-72.3: Api ca-aparokarthavacitvat pratyakasabdasya, sakad abhimukho rtha pratyaka. Pratigatam akam asminn iti ktva, ghaaniladinam aparokaa pratyakatva siddha bhavati. Tatpariccheda-[p.72]-kasya janasya tatuagnivat pratyakakaraatvat (Tib.: mngon sum gyi rgyu can yin pai phyir), pratyakatva vyapadisyate. Yas tv akam aka prati vartata iti pratyakasabda vyutpadayati, tasya janasya-indriyaviayatva[d viayaviayatva]c ca na yukta vyutpatti. Prativiaya tu syat pratyartham iti va. With this last point, Candrakirti effectively charges that Dignagas own etymology doesnt serve his purposes; for insofar as Dignaga wants primarily to designate the privileged epistemic faculty that operates with respect to (prati vartate) perceptibles, he would be better off etymologizing viaya viaya prati vartate in which case, though, the faculty would be called prativiaya (or pratyartha), not pratyaka. Ruegg, thinking that Candrakirti is here giving his preferred account, misses the fact that Candrakirti is adducing what is an unwanted consequence for Dignaga. Thus, Ruegg seems not to appreciate that the optative verb at the conclusion is meant to indicate something counterfactual, and instead translates the last sentence thus: But let there stand [the expression] prativiayam, or [the expression] pratyartham. (Ruegg 2002:125-6) 150 asrayasya paumandata; literally, the basiss being sharp or dull. 151 Literally, because of cognitions being possessors of change when there is change of that The reference here (noted by Ruegg [2002:126, n.235]) is clearly to Abhidharmakosabhayam 1.45 (Pradhan 1975:34), which begins tadvikaravikaritvad asrayas cakuradaya (the loci [of the senses are] the eyes, etc., because of [cognitions] changing when there is change in those). The bhaya (Ibid.) explains: cakuradina hi vikarea tadvijanana bhavaty anugrahopaghatapaumandatanuvidhanan na tu rupadina vikarea tadvikara (for [change] arises on the part of cognitions arising from [the senses], inasmuch as they are functioning or destroyed, sharp or dull, etc., according to change on the part of the eyes and so forth; but change in that [i.e., cognition] is not according to change on the part of the [cognized] forms, etc).
455
monly] seen [that there is] designation [of a thing] in terms of its specific cause (asadharaa), [as, e.g., we speak of] the sound of a drum, a sprout of barley, [even though there are also other causes operating to produce these effects]152. [Response:] This [case of pratyaka] is not the same as the foregoing153; for in that case, if perceptual cognition (vijana) were being defined in terms of its object as, for example, perceptual cognition of form, etc. then the difference[s that obtain] on the part of the sixfold perceptual cognition could not be made clear, since mental cognition (manovijana) proceeds with respect to the very same object as visual cognition154. [p.73] That is, if the sixfold cognition of color, etc.155, were [merely] called perceptual cognition (vijanam), [simpliciter,] there [would] arise a conception accompanied by an expectation, [to wit:] is this a perceptual cognition produced by the senses that possess form156, or is it a mental [cognition]? But when the specification is in terms of the basis [of the sense], even given the possibility that mental cognition
152 72.4-7: Atha syat, yathobhayadhinayam api vijanapravttav, asrayasya paumandata-anuvidhanad vijanana tadvikaravikaritvad, asrayeaiva vyapadeso bhavati, cakurvijanam iti; eva yady apy artham artha prati vartate, tathapy akam akam asritya vartamana vijanam, asrayea vyapadesat; pratyakam iti bhaviyati. Do hy asadharaena vyapadeso bherisabdo yavankura iti. Here, it is a discussion in Dignaga's Pramaasamuccaya (1.4a-b) that is likely referred to. Hattori gives the Sanskrit of Dignagas karika at p.87, n.1.32: asadharaahetutvad akais tad vyapadisyate. The vtti on this passage then proceeds to discuss precisely the examples here adduced by Candrakirti (viz., those of the sound of a drum and sprout of barley). Hattoris translation of the vtti here is at 1968:26, and the Tibetan is at pp.179-181; cf., also, Hattoris n.1.33, p.87; and Rueggs comment (2002:127, n.237). 153 That is, the example of cakurivijanas being so called is not relevant to the case of pratyakas being so called, for reasons to be made clear presently. 154 That is, the only reason for specifying the different vijanas in terms of the various indriyas from which they arise is to distinguish them from manovijana (mental cognition); for according to standard Abhidharmika analysis, the proper object of mental cognition just is the outputs of (the cognitions yielded by) the other five senses. Manovijana, then, cannot be distinguished in terms of its object insofar as it has as its object the outputs of the other vijanas and hence, their objects. The point that mental cognition thus bears, in a sense, on the same object as the various sensory cognitions is, it seems to me, not clear from Rueggs translation: for a mental cognition bears on a single object, along with eye-cognition, etc (2002:127). 155 That is, of all of the various things that respectively constitute the proper objects of the six kinds of sensory cognition. 156 That is, by one of the five bodily senses.
456
DAN ARNOLD
functions with respect to the objects of the ocular and other [sensory] cognitions, the difference between them is [nonetheless] established157. But in this case [i.e., that of pratyaka]158, if, with a desire to explain the definition of reliable warrants, its accepted [by you] that the fact of being a perception belongs only to what is devoid of conception159, [then] because its desired [by you] that we distinguish that [i.e., perception] only from conception, no benefit whatsoever is seen in designation according to the special cause160. And given that the function and number of reliable warrants are dependent on warrantable objects, and because of the presentation of the nature (svarupa) of [your] two reliable warrants whose reality is gained by virtue of the fact simply of [their] following the forms of [the two kinds of] warrantable objects specification in terms of the sense [faculties] does not help at all; hence, designation precisely by the object is in every way suitable161.
157 72.8-73.3: Na-etat purvea tulya. Tatra hi viayea vijane vyapadisyamane, rupavijanam ityevamadina, vijanaakasya bhedo na-upadarsita syat, manovijanasya cakuradivijanai sahaika-[p.73]-viayapravttatvat. Tatha hi niladivijanaake vijanam ity ukte, sakaka eva pratyayo jayate [de Jong]: kim etad rupindriyaja vijanam, ahosvin manasam? iti. Asrayea tu vyapadese, manovijanacakuradivijanaviayapravttisabhave pi, parasparabheda siddho bhavati. 158 Here, the iha still relates to the previous paragraphs na-etat purvea tulyam (this is not the same as the preceding). Thus, the interlocutor had wanted to say that pratyaka should be named in terms of its asraya (i.e., the senses), just as the various vijanas are. Candrakirti has responded that the cases arent comparable, and has just explained why the vijanas are designated as they are. Now, he explains why pratyaka is designated as it is or rather, as Dignagas project would require. 159 Here, Candrakirti brings in Dignagas definition of pratyaka as constitutively devoid of conceptual elaboration (kalpanapoha). 160 73.4-6: Iha tu pramaalakaavivakaya kalpanapohamatrasya pratyakatvabhyupagame sati, vikalpakad eva tadviseatvabhimatatvad asadharaakaraena vyapadese sati [strike this], na ki cit prayojanam upalakyate. Here, I read according to La Valle Poussins n.8, p.73, which indicates that the second sati in this sentence (i.e., p.73.5) is lacking in the Tibetan, and should be struck, so that the locative can be taken as a viaye saptami. The point is just that, given Dignagas definition of pratyaka, all he should be interested in doing is being sure to advance a nirukti that excludes kalpana just as cakurvijana is so called only in order to distinguish it from manovijana. But taking praty-akam to refer to the asraya does not advance that cause in any way. 161 73.6-8: Prameyaparatantraya ca pramaasakhyapravttau, prameyakaranukaritamatrataya ca samasaditatmabhavasattakayo pramaayo svarupasya vyavasthapanan,
457
[Objection:] Since the word perception, in the sense intended, is well known in the world, and since the word with respect to an object (pratyartha) is not well known, the etymology just in terms of the basis [of the sense faculty] is followed [by us]162. We respond: [p.74] This word perceptible is indeed well-known in the world163; but it is explained by us [and not by you] precisely as it is in the world164. But if, with disregard for ordinary categories as they are established, your etymology is being offered, [then] there would also be disregard for the expression well-known!165 And based on that [disregard], what is [commonly] called perception would not be such166.
na-indriyea vyapadesa ki cid upakarotiti, sarvatha viayeaiva vyapadeso nyayya. Again, Candrakirti here accepts, ex hypothesi, Dignagas goals, noting that according to these one ought to want a nirukti that etymologizes pratyaka in terms of its object, since the whole point of Dignagas account is that pramaa follows/corresponds to prameya. But of course, if Candrakirti wins this concession, then hes well on the way to advancing the trivialization of Dignagas privileged epistemic faculty. 162 73.9-74.1: Loke pratyakasabdasya prasiddhatvad, vivakite rthe pratyarthasabdasya-apratisiddhatvad, asrayeaiva vyutpattir asriyata iti cet. Here, Candrakirtis interlocutor turns the tables, rejoining that, on Candrakirtis etymological principles, the epistemic faculty that picks out perceptibles ought to be called pratyartham and since such is clearly not the case, it cannot be that the adjectival sense is rightly thought to be primary. 163 Ruegg (taking asti in an existential sense, and not as a copula): There is this word pratyaka which is current among ordinary folk in the world (2002:129). 164 Ruegg misses the disjunctive sense of tu here, taking it instead as vacuous (indeed): We have indeed expressed this just as it is in the world (2002:129). This misses the contrast that Candrakirti here urges between his own deference to conventional usage, and (what he takes to be) Dignagas inability to accommodate such usage. 165 The compound prasiddhasabda could also be rendered thus: there would also be disregard for the well-known word [i.e., pratyaka]. 166 74.1-3: Ucyate: asty aya pratyakasabdo lokaprasiddha [de Jong]. Sa tu yatha loke, tathasmabhir ucyata eva. Yathasthitalaukikapadarthatiraskarea tu tadvyutpade kriyamae, prasiddhasabdatiraskaro pi syat [de Jong], tatas ca pratyakam ityeva [na] syat. Again, Ruegg reads the optative (syat) in an existential rather than a copulative sense (cf., n.163, above): Thus there would not exist the term pratyaka. (2002:130) But Candrakirtis point, I think, is one that is comparable to a point made by many twentieth-century ordinary language philosophers: viz., that one cannot use a well-known, ordinary word, and yet substitute for it a peculiarly technical sense for in that case its then no longer the same thing that is under discussion. What Candrakirti is saying, then, is not that the word pratyaka would not exist, but simply that the things conventionally designated by the word would not be rightly so called in which case, most people would be wrong in their use of what is supposed to be a conventional word (a supposition that contradicts its being conventional).
458
DAN ARNOLD
And there could not be, on the part of one visual cognition, whose basis is a single moment of sense faculty, the quality of being a perception, since there would be no point in repetition (vipsarthabhavat)167; and if there is absence of the quality of being a perception on the part of one [such moment], there would be [such absence] on the part of many [instances of cognition, i.e., a continuum], too168. And because you accept that only cognition that is devoid of conception is perception169; and since nobodys discourse is by way of that [kind of cognition]170; and because of the desirability of explaining171 worldly
167 Here, Candrakirti begins a new tack, one that again accepts, ex hypothesi, Dignagas commitments here, presumably, the idea of kaikatva (momentariness), according to which pratyaka, like any cognitive event, would be constitutively episodic would, that is, have to consist in atomic moments of sense-function. This, then, is the context for Candrakirtis reference to the grammarians device of vipsa (Tib., zlos pa), repetition specifically (according to Apte, p.1487, s.v., meaning #2), the notion of Repetition of words to imply continuous or successive action. (Apte gives the example vka vka sicati.) This is the device that is invoked in etymologizing pratyaka as akam aka prati vartate, and Candrakirti is pointing out that repetition (akam akam) implies a continuity or successiveness such as could not obtain given the idea of radical kaikatva. Thus, a single, atomic moment of (say) ocular perception could not warrant the grammarians device of vipsa. 168 74.3-5: Ekasya ca cakurvijanasya-ekendriyakaasrayasya pratyakatva na syad vipsarthabhavat (Tib., zlos pai don med pai phyir) ekaikasya ca pratyakatvabhave, bahunam api na syat. This last point is then reminiscent of one of the fundamental points of Vasubandhus critique of atoms in the Visatika (and of Dignagas similar arguments from the Alambanaparika): if a single, atomic moment of perceptual cognition does not make sense (here, insofar as it renders the repetition akam akam meaningless), then there is nothing out of which to build up a succession (or continuum, satana) of sensory cognitions, either. This is all put very elliptically, and the point, again, is simply to argue that Dignagas proposed account does not advance (because it is not coherent with) his own goals. 169 Ruegg (2002:130) here, as throughout this section, renders pratyaka as perceptible; but I think that if we are to make sense of these passages, we must take Candrakirti to be provisionally adopting Dignagas preferred sense of the word. As with the earlier discussion of svalakaa, then, the debate concerning pratyaka similarly involves some alternation in meaning, insofar as it is precisely what the word should mean that is most basically in question. 170 The word tena here will, I think, admit of two readings: it can mean something like thus or therefore, in which case what follows (lokasya savyavaharabhavat) is intended counterfactually (because there would be no meaningful discourse on the part of the world); or it can refer back to kalpanapohajana, in which case (as in my translation) it simply says there is no meaningful discourse in the world of the sort that makes use of this conception. The latter is probably the more Sanskritically straightforward read-
459
discourse with respect to reliable warrants and warrantable objects [your] conception of the reliable warrant that is perception becomes quite senseless172. [You have cited, in support of your claim that perception is constitutively devoid of conceptual elaboration, a familiar Abhidharma text that says,] A man endowed [only] with visual cognition senses173 blue, but [he does] not [know] that it is blue.174 The point of this authoritative text (agama) is not to state a definition of pratyaka175, but [is instead simply
ing in the present context, but Candrakirti can (and, I have argued elsewhere, should) be understood as well to be making the stronger point viz., that Dignagas whole project is self-referentially incoherent, insofar as his use of ordinary words in peculiarly technical senses would absurdly entail the impossibility even of the discourse in which he is engaged; for it is a condition of the possibility of meaningful discourse that people understand words in more or less the same way. Ruegg renders tena here in the way that could recommend this stronger point: and because there is, therefore, [according to your doctrine on this matter] no [simple] transactional-pragmatic usage. (2002:130; my emphasis) 171 vyakhyatum iatvat. This could be read in a couple of ways it could be taken normatively (i.e., as reflected in the translation I have given here), in which case we might prefer here to see a gerundive; or it could simply be stating Dignagas own avowed interests (in which case, we might render, since you desire to explain). 172 74.6-8: Kalpanapohasyaiva ca janasya pratyakatvabhyupagamat, tena ca lokasya savyavaharabhavat, laukikasya ca pramaaprameyavyavaharasya vyakhyatum iatvat, vyarthaiva pratyakapramaakalpana sajayate. 173 Here, I read (following a quotation of this sentence by Yasomitra) vijanati (rather than with Candrakirtis janati). See n.174. 174 Cakurvijanasamangi nila janati no tu nilam iti. Dignaga cites this quotation in his vtti to Pramaasamuccaya 1.4 (translated at Hattori, p.26; Tibetan at Hattori, p.179), where he reports that it is said in an Abhidharma treatise (chos mngon pa las). The passage can be found in, e.g., the Abhidharmakosavyakhya of Yasomitra, whose text reads vijanati (in contrast with Candrakirtis janati); see Shastri 1998: 72. (According to Ruegg [2002:131, n.254],Yasomitra is quoting the Vijanakaya.) Hattori succinctly summarizes the motivation behind Dignagas citation of this: The expression nila vijanati implies that one has an immediate awareness of the object itself. On the other hand, nilam iti vijanati implies that one forms a perceptual judgement by associating a name with the object perceived. Thus, the above Abhidharma passage expresses the thought that perception is free from conceptual construction (kalpanapoha). (1968: 88, n.1.36) On my rendering nilam iti with a that-clause, see Arnold (2005), Chapter 7. 175 agamasya pratyakalakaabhidhanarthasya-aprastutatvat. Though a Sanskritically natural way to express Candrakirtis point, this phrase is particularly difficult to render in a syntactically literal way into English. Such a rendering would go something like this: because the point (artha) of stating a definition of pratyaka is not the subject (a-prastuta) of this agama or more precisely (disclosing the fact that this predication is stated in the form of a genitive-plus-tva construction), because of [this] points not being the
460
DAN ARNOLD
that of] [p.75] demonstrating (pratipadaka) the insentience (jaatva) of the five [bodily] senses176. Hence177, not on the basis of authoritative texts, either, [can it be said that] the quality of being perception belongs only to cognition that is devoid of conception; hence, this [characterization of pratyaka as devoid of conceptual elaboration] does not make sense178. Therefore, in the world, if any (sarvam eva) subject of characterization (lakya) whether it be a svalakaa or a samanyalakaa is visible, because of being directly apprehended, then it is established as perceptible, along with the cognition that has it as its object [which is derivsubject of this agama. Such constructions are, I think, generally best rendered in the latter way (Xs being Y), which discloses that such constructions generally state simple identities (X is Y) but do so in such a way as to make it possible to show the inferential consequences of such identities (so that, putting the Y term in the ablative, we get, because of Xs being Y). But when the term in the genitive (arthasya) is, as in this case, the final member of a lengthy tatpurua compound, this becomes impossible. It is, however, important to see that the construction can be read this way, since, in the present case, another alternative is to take the second genitival compound as a bahuvrihi: because of the irrelevance of this agama, which has as its purpose the expression of a definition of perception. So Ruegg: there is no relevance [here] of the Agama which has the sense expressing the defining characteristic of direct perception. (2002:131; my emphasis) The problem is that on this construal, the claim does not effectively counter Dignagas appeal to it. Ruegg tries to salvage Candrakirtis point by taking the sentence to mean that this passage, though defining perception, is not relevant here. On my reading, in contrast, Candrakirti more basically contests Dignagas understanding of the passage. In fact, the context for Yasomitras citation of the passage (see n.174) recommends Candrakirtis point; for Yasomitra adduces the quotation in commenting on the part of Vasubandhus text that treats the cognitive outputs of the five non-mental senses and the point of the passage is (as Candrakirti goes on to say) thus to urge simply that the outputs of the five sense faculties are not meaningful until they have become the objects, as well, of the manovijana. This quotation, as deployed by these Abhidharmikas, therefore indeed does not state a definition of perception, but instead makes a characteristically Abhidharmika point about the relationship between the five bodily vijanas and the manovijana. 176 That is, their being non-epistemic until they have become the objects of the manovijana; see the preceding note. Rueggs rendering of the first reason (preceding note) obscures the recognition that this second reason represents a contrasting interpretation of the same text (i.e., of the quotation adduced by Dignaga); thus, Ruegg translates: and (ii) Agama makes known that the five cognitions belonging to the sense faculties are [in themselves] insentient. (2002:131) 177 Both of the foregoing phrases occur in the ablative, giving two reasons for the conclusion now stated. 178 74.8-75.2 Cakurvijanasamangi nila janati no tu nilam iti ity ca-agamasya pratyakalakaabhidhanarthasya-aprastutatvat, paca-[p.75]nam indriyavijanana jaatvapratipadakatvac ca, na-agamad api kalpanapohasyaiva vijanasya pratyakatvam iti na yuktam etat.
461
atively called pratyaka].179 But twin moons and other such [illusions] do not, from the point of view of the cognition of one without cataracts, have the quality of being perceptible (apratyakatvam) although from the point of view of one with cataracts, [such illusions] have precisely the quality of being perceptible180.
179 This, then, is the point Candrakirti has been driving at all along: if it is finally the adjectival sense of pratyaka (perceptible) that is primary, and if this motivates the derivative usage of the word that denotes as well the cognition that has [any perceptible thing] as its object (tadviayea janena saha), then the word cannot be thought to pick out a privileged cognitive instrument (one that is kalpanapoha). In that case, it must be allowed that abstractions (samanyalakaas), too, are perceptible. Rueggs translation misses what I thus take to be the point: Therefore, for ordinary folk in the world, if [as claimed by you, there indeed exists] a lakya, or if [there indeed exists both] a svalakaa and a samanyalakaa, all will in fact be not unamenable to perception, for there will [then] be immediate apprehension. (2002:131) As quickly becomes clear, the confusion here follows from Rueggs not taking this all to represent Candrakirtis own conclusion about the proper understanding of pratyaka. Hence, Ruegg completes this passage thus: And [by you] the pratyaka is accordingly set out systematically along with the (vi)jana having it for its object. 180 75.2-5: Tasmal loke yadi lakya, yadi va svalakaa samanyalakaa va, sarvam eva sakad upalabhyamanatvad aparoka, ata pratyaka vyavasthapyate tadviayea janena saha. Dvicandradina tv ataimirikajanapekaya-apratyakatva, taimirkadyapekaya tu pratyakatvam eva. Ruegg takes the last sentence as stating an unwanted consequence for Dignaga: However, [following your doctrine,] in respect [even] to those who are affected by eye-disease and the like there will indeed be direct perceptibility. (Ruegg 2002:132; cf., his n.256) I take it that this is, rather, part of the account of pratyaka that Candrakirti is commending as contra Dignagas, and that it therefore states a positively desired consequence desired, that is, insofar as it undercuts the privileged status of Dignagas category of perception. Candrakirtis point, then, is that perception is not intrinsically better suited, independent of context, to confer justification; rather, what is perceptible is always relative to perceivers and their contexts. Candrakirtis point about the genuine perceptibility of twin moons represents precisely the sort of claim that is reversed by such later svatantrikas as Janagarbha and Santarakita. Consider, e.g., Ichigos statement of the impetus behind the svatantrika distinction between true and false conventional: Santarakita owes one of his definitions of conventional truth to Janagarbha's basic idea of conventional truth as it appears.' This being the nature of conventional reality, should we then also regard as conventional truth the double moon that appears to those who have defective vision? Partly in response to this issue, Janagarbha distinguishes two types of conventional truth, namely true and false conventional truth. (Ichigo 1989:169) To the extent that Candrakirtis point is to emphasize only how dramatically limited is our perspective relative to the ultimate truth, he is not simply saying that, on the conventional level, anything goes. Indeed, Candrakirti may posit something analogous to the svatantrikas mithyasavti in the form of alokasavti (non-worldly conventional); cf., Prasannapada 493.2-4. With his char-
462
DAN ARNOLD
But cognition whose object is [something] invisible, [when such cognition is] produced by a mark that is invariably concomitant (avyabhicarin) with the thing to be proven, [is known as] inference181. The speech of those who are accomplished, who know directly things that are beyond the senses this is [known as] tradition. Understanding of a thing not [previously] experienced, based on [its] similarity [with something familiar, is known as] comparison for example, [when one learns that,] a cow is like an ox. Just so: everyones understanding of things is established based on this fourfold [scheme of] reliable warrants.182
acteristic stress that conventional means lacking in critical analysis, though, Candrakirti may invite some clarification such as the svatantrikas elaborated. We can nevertheless appreciate that Candrakirtis basic point here is simply to relativize our epistemic instruments: none is intrinsically suited to confer justification, which will always be a matter of context and the context of those of us who are not already Buddhas is such that none of our cognitive instruments can be thought to put us into cognitive contact with anything that is ultimately existent (paramarthasat). 181 Not having seen that the preceding section represented Candrakirtis preferred account of pratyaka, Ruegg now seems not to see the natural segue here to Candrakirtis endorsing (as conventionally valid, at least) the standard Naiyayika list of pramaas. Thus, Ruegg translates: On the other hand, a jana that. (2002:132) But Candrakirti is not, I take it, here offering an alternative to the foregoing; rather, having endorsed a characteristically Naiyayika understanding of perception (i.e., as including abstractions among its objects), he is now proceeding more generally to endorse this alternative to Dignagas spartan epistemology. 182 75.6-9: Parokaviaya tu jana sadhyavyabhicarilingotpannam, anumana. Sakad atindriyarthavidam aptana tad vacana, sa agama. Sadsyad ananubhutarthadhigama upamana, gaur iva gavaya iti yatha. Tad eva pramaacatuayal lokasyaarthadhigamo vyavasthapyate. Cabezn, commenting on a quotation of this passage by the dGe-lugs-pa scholar mKhasgrub-rje, considers it a conundrum why Candrakirti chose to cite four types of valid cognitions (as the Naiyayikas do, for example), and not the standard two of Dignaga and Dharmakirti. (1992:454n) He takes the point of the passage, in the hands of mKhasgrub-rje, to be that of proving that the Madhyamikas do not in general repudiate the notion of a valid cognition (1992: 118) that is, that Madhyamaka can retain the project of Dignaga and Dharmakirti as at least conventionally useful. It should be clear by now, though, that Candrakirtis endorsement of this fourfold schema is meant to be an endorsement of what he takes to be an adequate account of our conventional epistemic practices and that specifically contra the account of Dignaga, which Candrakirti takes to be (not only not ultimately but) not even conventionally valid. The characteristically dGelugs-pa fudging of this point serves the goal of taking Candrakirti as normative, while at the same time retaining precisely the epistemological discourse he so clearly dismisses. It raises interesting historical and philosophical questions that this Tibetan tradition should thus have wedded what Candrakirti, at least, took to be antithetical projects.
463
And these are established in dependence upon one another: given reliable warrants, there are warrantable objects, and given warrantable objects, there are reliable warrants183. But it is emphatically not the case that the establishment of reliable warrants and their objects is essential (svabhaviki)184. Therefore, let the mundane remain just as it is seen185. Enough of this subject (ala prasangena)! We will [now] explain the real matter at hand. The teaching of the dharma of the blessed Buddhas [was given with them] having located themselves precisely in the worldly perspective186. Abbreviations
Abhyankar, K. V. 1977. A Dictionary of Sanskrit Grammar. Second, Revised edition. Baroda: Oriental Institute. Alston, William. 1989. A Doxastic Practice Approach to Epistemology, in Marjorie Clay and Keith Lehrer, eds., Knowledge and Skepticism, pp.1-29. Boulder: Westview Press. Anandasrama Sanskrit Series, editors of. 1928. Madhava, Sarvadarsanasagraha. Trivandrum: Anandasrama Sanskrit Series, vol. 51. Apte, Vaman Shivaram. 1992. The Practical Sanskrit-English Dictionary. Revised and enlarged edition, edited by P. K. Gode and C. G. Karve. Kyoto: Rinsen. (Reprint; revised and enlarged edition first published in Poona, 1957.)
Cf., Oetkes comment, n.87, above. Here, Candrakirti finally makes clear the presupposition that, on his view, guides Dignagas whole project (and the presupposition, therefore, that he finally means to target with all of the foregoing critique) viz., that Dignagas having abstracted privileged pramaas is tantamount to his having posited them as independent (or essential, natural, etc.) epistemic perspectives on what there is. Against this, what Candrakirti has chiefly wanted to stress all along is that pramaas and prameyas are, like everything else, upadaya prajaptaya that is (as he puts it elsewhere), they exist simply as being mutually interdependent (parasparapekamatrataya). 185 That is, let it not be thought (per the interlocutors contention at p.58.14, ff.; n.44, above) that a systematic re-description of our conventional epistemic practices is called for. Ruegg translates: Let there be, therefore, only the worldly [i.e. transactional-pragmatic convention of ordinary folk] that conforms with what is known by experience (2002:134). 186 75.10-13: Tani ca parasparapekaya sidhyanti: satsu pramaeu prameyartha, satsu prameyev artheu pramaani. No tu khalu svabhaviki pramaaprameyayo siddhir iti; tasmal laukikam eva-astu yathadam ity; ala prasangena. Prastutam eva vyakhyasyama. Laukika eva darsane sthitva buddhana bhagavata dharmadesana. This programmatic statement concludes the part of Candrakirtis Prasannapada that is framed specifically as commenting on MMK 1.1 and it is with this passage that Rueggs translation from chapter one ends, as well.
184 183
464
DAN ARNOLD
Arnold, Dan. 2001. How to Do Things with Candrakirti: A Comparative Study in Anti-skepticism. Philosophy East & West 51/2 (2001): 247-279. Arnold, Dan. 2003. Candrakirti on Dignaga on Svalakaas, Journal of the International Association of Buddhist Studies 26/1: 139-174. Arnold, Dan. 2005. Buddhists, Brahmins, and Belief: Epistemology in South Asian Philosophy of Religion. New York: Columbia University Press. Bendall, Cecil, ed. 1970. ikshasamuccaya: A Compendium of Buddhist Teaching compiled by antideva chiefly from earlier Mahayana-Sutras. Bibliotheca Buddhica, Vol. I. Osnabrck: Biblio Verlag. (Reprint; first published 1897-1902.) Bhattacharya, Kamaleswar. 1980. Nagarjunas Arguments against Motion: Their Grammatical Basis, in A. L. Basham, et al, eds., Corpus of Indian Studies: Essays in Honor of Professor Gaurinath Sastri, pp.85-95. Calcutta: Sanskrit Pustak Bhandar. Bhattacharya, Kamaleswar. 1980-81. The Grammatical Basis of Nagarjunas Arguments: Some Further Considerations, Indological Taurenensia 8-9: 35-43. Bhattacharya, Kamaleswar. 1990. The Dialectical Method of Nagarjuna: Vigrahavyavartani. Third edition, revised and enlarged (containing, along with Bhattacharyas English translation, a critical edition of the Sanskrit by E. H. Johnston and Arnold Kunst). Delhi: Motilal Banarsidass, 1990. Cabezn, Jos Ignacio. 1992. A Dose of Emptiness: An Annotated Translation of the sTong thun chen mo of mKhas grub dGe legs dpal bzang. Albany: SUNY Press. De Jong, J. W. 1978. Text-critical Notes on the Prasannapada. Indo-Iranian Journal 20, nos.1/2: pp.25-59; nos.3/4: pp.217-252. De Jong, J. W. 1981. (Review of Mervyn Sprung, Lucid Exposition of the Middle Way), Indo-Iranian Journal 23: 227-230. Dreyfus, Georges. 1997. Recognizing Reality: Dharmakirti's Philosophy and its Tibetan Interpreters. Albany: SUNY. Dreyfus, Georges B. J., and Sara L. McClintock, eds. 2003. The SvatantrikaPrasangika Distinction: What Difference Does a Difference Make? Boston: Wisdom Publications. Eckel, Malcolm D. 1978. Bhavaviveka and the early Madhyamika theories of language. Philosophy East & West 28/3: 323-337. Griffiths, Paul. 1981. Buddhist Hybrid English, Journal of the International Association of Buddhist Studies 4/2: 17-32. Hall, Bruce Cameron. 1983. Vasubandhu on Aggregates, Spheres, and Components: Being Chapter One of the Abhidharmakosa. Ph.D. dissertation, Harvard University. Hattori Masaaki. 1968. Dignaga, On Perception, being the Pratyakapariccheda of Dignagas Pramaasamuccaya from the Sanskrit fragments and the Tibetan versions. Cambridge: Harvard University Press, 1968.
465
Huntington, C. W. 2003. Was Candrakirti a Prasangika?, in Dreyfus and McClintock, pp.67-91. Ichigo Masamichi, ed., trans. 1989. Santarakitas Madhyamakalakara, in Luis Gmez and Jonathan Silk, eds., Studies in the Literature of the Great Vehicle: Three Mahayana Buddhist Texts, pp.141-240. Ann Arbor: Center for South and Southeast Asian Studies, The University of Michigan. Ingalls, Daniel H. H. 1954. The Comparison of Indian and Western Philosophy. Journal of Oriental Research, Madras 22: 1-11. Katre, Sumitra M. 1987. Aadhyayi of Paini. Texas Linguistics Series. Austin: University of Texas Press. La Valle Poussin, Louis de, ed. 1970a. Mulamadhyamakakarikas (Madhyamikasutras) de Nagarjuna, avec la Prasannapada Commentaire de Candrakirti. Bibliotheca Buddhica, Vol. IV. Osnabrck: Biblio Verlag. (Reprint; originally St.-Ptersbourg: Acadmie Impriale des Sciences, 1903-1913.) La Valle Poussin, Louis de, ed. 1970b. Madhyamakavatara par Candrakirti: Traduction tibtaine. Bibliotheca Buddhica, Vol. IX. Osnabrck: Biblio Verlag. (Reprint; originally St.-Ptersbourg: Acadmie Impriale des Sciences, 1907-1912.) Lang, Karen. 1986. Aryadevas Catusataka: On the Bodhisattva's Cultivation of Merit and Knowledge. Copenhagen: Akademisk Forlag. Lang, Karen. 1992. sPa tshab nyi ma grags and the introduction of Prasangika Madhyamaka into Tibet, in G. Kuppuram and K. Kumudamani, eds., Buddhist Heritage in India and Abroad, pp.267-281. Delhi: Sundeep Prakashan. Lang, Karen. 2003. Four Illusions: Candrakirtis Advice to Travelers on the Bodhisattva Path. New York: Oxford University Press. Lindtner, Chr. 1987. Nagarjuniana: Studies in the Writings and Philosophy of Nagarjuna. Delhi: Motilal Banarsidass, 1987. (Reprint; first published in Copenhagen, Institute for indisk filologi, 1982.) Malvania, Paita Dalsukhbhai, ed. 1971. Paita Durveka Misras Dharmottarapradipa. Patna: Kashiprasad Jayaswal Research Institute; 2nd ed. (Tibetan Sanskrit Works Series, vol.11.) Matilal, B. K. 1990. The Word and the World: Indias Contribution to the Study of Language. Delhi: Oxford University Press. May, Jacques, trans. 1959. Candrakirti Prasannapada Madhyamakavtti: Douze chapitres traduits du sanscrit et du tibtain, accompagns d'une introduction, de notes et d'une dition critique de la version tibtaine. Collection Jean Przyluski, Tome II. Paris: Adrien-Maisonneuve. Mookerjee, Satkari. 1957. Candrakirtis Exposition of the Madhyamika, in Satkari Mookerjee, ed., The Nava-Nalanda-Mahavihara Research Publication (Patna), vol. 1: 42-58. Oetke, Claus. 2003a. Some Remarks on Theses and Philosophical Positions in Early Madhyamaka, Journal of Indian Philosophy 31: 449-478.
466
DAN ARNOLD
Oetke, Claus. 2003b. (Review of Jonardon Ganeri, Philosophy in Classical India), Indo-Iranian Journal 46: 135-155. Pandeya, Ram Chandra, ed. 1999. Madhyantavibhagasastram, Acaryamaitreyaktakarikabhi, Acaryavasubandhuktena bhayea, Sthiramatiktaikaya ca savalitam. Delhi: Motilal Banarsidass. Plantinga, Alvin. 1993. Warrant: The Current Debate. New York: Oxford University Press. Pradhan, Prahlad, ed. 1975. Abhidharmakosabhayam of Vasubandhu. Patna: K. P. Jayaswal Research Institute. Rhys Davids, T. W., and William Stede. 1995. The Pali Text Societys PaliEnglish Dictionary. Oxford: Pali Text Society. (Reprint; first published 1921-1925.) Ruegg, David Seyfort. 1981. The Literature of the Madhyamaka School of Philosophy in India. A History of Indian Literature (ed. Jan Gonda), Vol. VII, Fasc. 1. Wiesbaden: Otto Harrassowitz. Ruegg, David Seyfort. 1992. Some Reflections on Translating Buddhist Philosophical Texts from Sanskrit and Tibetan, Etudes Asiatiques / Asiatische Studien 46/1: 367-391. Ruegg, David Seyfort. 2002. Two Prolegomena to Madhyamaka Philosophy: Candrakirtis Prasannapada Madhyamakavtti on Madhyamakakarika I.1, and Tson kha pa blo bzan grags pa / rGyal tshab dar ma rin chens dKa gnad / gnas brgyad kyi zin bris: Annotated Translations. Wiener Studien zur Tibetologie und Buddhismuskunde, Heft 54. Wien: Arbeitskreis fr tibetische und buddhistische Studien, Universitt Wien. Ruegg, David Seyfort. 2004. The Indian and the Indic in Tibetan Cultural History, and Tson kha pas Achievement as a Scholar and Thinker: An Essay on the Concepts of Buddhism and Tibet and Tibetan Buddhism, Journal of Indian Philosophy 32: 321-343. Saito Akira. 1984. A Study of the Buddhapalita-mulamadhyamaka-vtti. Ph.D. dissertation, Australian National University. Schayer, Stanislaw. 1931. Ausgewhlte Kapitel aus der Prasannapada (V, XII, XIII, XIV, XV, XVI): Einleitung, Uebersetzung und Anmerkungen. Krakw: Polska Akademja Umiejetnosci, Prace Komisji Orjentalistycznej. Shastri, Swami Dwarikadas. 1998. Acaryayasomitraktasphuarthavyakhyopetam Acaryavasubandhuviracitam svopajabhayasahita ca Abhidharmakosam. Bauddha Bharati Series, 5-8 (printed in two volumes). Varanasi: Bauddha Bharati. Siderits, Mark. 1981. The Madhyamaka Critique of Epistemology II. Journal of Indian Philosophy 9: 121-160. Sprung, Mervyn, trans. 1979. Lucid Exposition of the Middle Way: The Essential Chapters from the Prasannapada of Candrakirti. In collaboration with T. R. V. Murti and U. S. Vyas London: Routledge & Kegan Paul.
467
Stcherbatsky, Th. 1927. The Conception of Buddhist Nirvaa. Delhi: Motilal Banarsidass, 1989. (Reprint; originally published in Leningrad by the Publishing Office of the Academy of Sciences of the USSR.) Steinkellner, Ernst. 1982. (Review of Mervyn Sprung, Lucid Exposition of the Middle Way), Journal of the American Oriental Society 102: 411-414. Thurman, Robert. 1991. The Central Philosophy of Tibet: A Study and Translation of Jey Tsong Khapas Essence of True Eloquence. Princeton: Princeton University Press. (Paperback reprint of Tsong Khapas Speech of Gold in the Essence of True Eloquence, 1984.) Tillemans, Tom J. F. 1990. Materials for the Study of Aryadeva, Dharmapala and Candrakirti: The Catusataka of Aryadeva, Chapters XII and XIII, with the Commentaries of Dharmapala and Candrakirti: Introduction, Translation, Sanskrit, Tibetan and Chinese Texts, Notes. Wiener Studien zur Tibetologie und Buddhismuskunde, Heft 24, 1-2. Wien: Arbeitskreis fr tibetische und buddhistische Studien. Tillemans, Tom J. F. 2003. Metaphysics for Madhyamikas, in Dreyfus and McClintock, eds., pp.93-123. Vaidya, P. L., ed. 1960. Madhyamakasastra of Nagarjuna, with the Commentary: Prasannapada, by Candrakirti. Buddhist Sanskrit Texts, No. 10. Darbhanga: Mithila Institute of Post-Graduate Studies and Research in Sanskrit Learning. Walleser, Max, ed. 1970. Buddhapalita: Mulamadhyamakavtti: Tibetische bersetzung. Bibliotheca Buddhica, Vol. XVI. Osnabrck: Biblio Verlag. (Reprint; originally St.-Ptersbourg: Acadmie Impriale des Sciences, 19131914.) Yonezawa Yoshiyasu. 1999. *Lakaaika: A Sanskrit Manuscript of an Anonymous Commentary on the Prasannapada. Indogaku Bukkyogaku Kenkyu 47/2: 1024-1022. Yonezawa Yoshiyasu, et al. 2001. Introduction to the Facsimile Edition of a Collection of Sanskrit Palm-leaf Manuscripts in Tibetan dBu med Script. Taisho University: The Institute for Comprehensive Studies of Buddhism (Study Group of Sanskrit Manuscripts in Tibetan dBu med Script). Yonezawa Yoshiyasu. 2004. *Lakaaika: Sanskrit Notes on the Prasannapada (1), Journal of Naritasan Institute for Buddhist Studies, no.27: 115-154. Yoshimizu Chizuko. 1996. Die Erkenntnislehre des Prasangika-Madhyamaka: Nach dem Tshig gsal ston thun gyi tshad mai rnam bsad des Jam Dbyans Bzad Pai Rdo Rje: Einleitung, Textanalyse, bersetzung. Wiener Studien zur Tibetologie und Buddhismuskunde, Heft 37. Wien: Arbeitskreis fr tibetische und buddhistische Studien Universitt Wien. Yotsuya Kodo. 1999. The Critique of Svatantra Reasoning by Candrakirti and Tsong-kha-pa: A Study of Philosophical Proof According to Two Prasangika Madhyamaka Traditions of India and Tibet. Stuttgart: Franz Steiner Verlag.