Marginalia To Dharmakirtis Pramanaviniscaya I II

Author(s): Christian Lindtner
Publisher: Christian Lindtner
Catalog link: https://jainqq.org/explore/269415/1

Page #1

ARCHIV FÜR INDISCHE PHILOSOPHIE
MARGINALIA TO DHARMAKIRTI'S PRAMAŅAVINISCAYA I-II
By Christian Lindtner, Nærum
I
If the general task of historical-philological inquiry - as in contradistinction to purely philosophical inquiry is to acquire a knowledge of what happened and an understanding of what was understood – primarily from autopsy of literary sources - and if it absolves this task roughly through three interrelated phases: 1) linguistic interpretation dealing with textual criticism, lexicography, etc., 2) higher criticism dealing with authenticity, date, etc. of texts, and, 3) hermeneutics, or interpretation, dealing with the background, circumstances, motives, content, form and meaning, etc. of the texts, and if, finally, it is the ultimate task of a historically orientated philologist to present a systematically reconstructive interpretation of his text (and its author)', then, surely, any philologist or philosopher taking an interest in Indian logic and epistemology is in a debt of gratitude to ERNST STEINKELLNER, TILMANN VETTER and the late ERICH FRAUWALLNER for the pains-taking efforts and results achieved by them concerning the works and the thought of the Buddhist philosopher Dharmakīrti (ca. 530-600)2.
In these methodological reflections I am mainly following the Danish classical philologist J. N. MADVIG, for an account of whose views on the method and tasks of philology see PoVL JOHS. JENSEN, Johan Nicolai Madvig. Et Mindeskrift, København 1963, pp. 16-43.
2 FRAUWALLNER'S contributions are collected in GERHARD OBERHAMMER and ERNST STEINKELLNER, Erich Frauwallner. Kleine Schriften, Wiesbaden 1982. The most important monograph on Dharmakirti we owe to TILMANN VETTER, Erkenntnisprobleme bei Dharmakīrti, Wien 1964. Among ERNST STEINKELLNER'S contributions special mention may be made of his Die Entwicklung des Kṣaṇikatvānumānam bei Dharmakirti, WZKS 12-13 (1968) 361-377 and Wirklichkeit und Begriff bei Dharmakīrti, WZKS 15 (1971) 179-221. The celebrated pioneer works of F. TH. STCHERBATSKY are now largely superseded. For the date ca. 530-600 (against FRAUWALLNER's ca. 600-660) and for further references see my paper Apropos Dharmakīrti - Two new Works and a new Date, AO 41 (1980) 27-37.

Page #2

150
Ch. LINDTNER
While the problem of “Die Reihenfolge und Entstehung der Werke Dharmakīrti's” has already been established in rough outlines by FRAUWALLNER in a paper thus entitled", and while the general historical significance of Dharmakīrti seems to boil down to his penetrating contributions to "scientific methodology" (i.e. to logic and epistemology rather than to ethics and ontology), still numerous textual and historical puzzles as well as problems of interpretation remain to be solved before we are in a position to reconstruct a full picture of Dharmakīrti as an individual author and philosopher with the historical setting as contrasting background.
The following marginal notes to the first two chapters of the Pramanaviniscaya (PVin) - Dharmakirti's most systematic major work - as edited and translated by VETTER and STEINKELLNERS are only intended as a modest contribution towards that end. The approach of its author can be characterized as that of a philologist primarily concerned with philosophical issues.
First, however, it will be useful for our purpose briefly to recall the frame-work of his thought as presented in PVin I-II and subsequently summarized by him in the epitome Nyāyabindu (NB). Though his
* Reprinted in his Kleine Schriften, pp. 667–689. - While his views on the formation and sequence of Dharmakīrti's works is basically valid he does, however, leave several of the works ascribed to him out of account. Moreover, the Yogācāra background of Dharmakīrti is entirely left out of consideration, though, in my opinion, we here find a basic motive in the formation of his thought.
+ I have adopted this characterization from a paper read in Copenhagen (May 1983) by Ernst Steinkellner whom I here wish to thank for precious advice and support in preparing this contribution.
5 TILMANN VETTER, Dharmakīrti's Pramāņaviniscayah. 1. Kapitel: Pratyakşam. Einleitung, Text der tibetischen Übersetzung, Sanskritfragmente, deutsche Übersetzung, Wien 1966. - ERNST STEINKELLNER, Dharmakīrti's Pramāņaviniscayah. Zweites Kapitel: Svārthānumanam. Teil I. Tibetischer Text und Sanskrittexte, Wien 1973; id., Dharmakīrti's Pramānaviniscayah. Zweites Kapitel: Svārthānumānam. Teil II. Übersetzung und Anmerkungen. Wien 1979. See also ERNST STEINKELLNER, Verse-Index of Dharmakīrti's Works (Tibetan Versions), Wien 1977.
• Dharmottarapradīpa. Being a subcommentary on Dharmottara's Nyāyabindutīkā, a commentary on Dharmakīrti's Nyāyabindu. Dec. and ed. D. MALVANIA, Patna 1955. - For a general characterization of the relationship of NB to PVin see e.g. ERNST STEINKELLNER, Dharmakīrti's Hetubindu. Teil I. Tibetischer Text und rekonstruierter Sanskrit-Text, Wien 1967, pp. 24–25: "Der Nyāyabinduḥ, welcher im engsten Zusammenhang mit dem Pramāņaviniscayah steht, ist der Form nach eine Art Epitome aus diesem, dem Inhalt nach ein knapper Leitfaden zur Einführung in die Theorie der Erkenntnismittel und der Logik. Die Einteilung in drei Kapitel (pratyaksam, svārthānumānam,

Page #3

Marginalia to Dharmakīrti's Pramāņaviniscaya
151
remaining extant logical and epistemological works - Pramāņavārttika (PV, earlier then PVin and NB), Hetubindu (later than PV and PVin), Vādanyāya (perhaps his latest work) and the two “essays” Sambandhaparīkņā and Samtānāntarasiddhi – as a rule are composed with a more specific purpose and are more rich in details (thus, it may be added, occasionally offering us a fascinating glimpse into his workshop) they all nevertheless presuppose the same over-all structure (and should be interpretated with that context in mind) as presented in PVin and NB - a structure basically inherited from Dignāga's epoch-making Pramāṇasamuccaya?
Dharmakīrti, who sees himself as a commentator to Dignāga, is primarily concerned with valid knowledge (samyagjñāna) as a means of successful action. Only two means of valid knowledge (pramāņa) are admissible: pratyaksa and anumāna. Once an object of interest (artha) has been determined by either of them we are in a position to get what we want and avoid what we do not want. The artha known to us either occurs directly in perception as "evident” (pratyaksa) or indirectly, inferred through anumāna. Any empirical object of pratyaksa has the power (Śakti) - in contradistinction to that of anumāna - to impress itself clearly and directly in cognition (jñāna) in its unique particularity (svalaksana). Nothing could be more obvious - pratyaksa - than a thing in its svalaksaņa, just as it presents itself. The other, or "secondhand”, means of valid knowledge deals with exactly the same thing, but only indirectly through concepts, as a perceptual derivative: svalaksanadar
parārthānumānam) steht über Pramāņaviniscayaḥ und Pramāņavārttikam noch in der Tradition des Pramāṇasamuccayah, bestimmt den Nyāyabinduh also ein Lehrbuch der Erkenntnismittel, nicht eines der Logik”.
For the first chapter see MASAAKI HATTORI, Dignāga: On Perception. Being the Pratyakşapariccheda of Dignāga's Pramāṇasamuccaya from the Sanskrit fragments and the Tibetan versions, Cambridge Mass. 1968. For chapters II, III, IV and VI see H. KITAGAWA, Indo Koten Ronrigaku no Kenkyū, Tokyo 1965. For chapter V we now have M. HATTORI, The Pramāņasamuccayavrtti of Dignāga with Jinendrabuddhi's commentary. Chapter five: Anyāpoha-parīkņā. Tibetan Text with Sanskrit Fragments, Kyoto 1982. An annotated translation of this chapter by OLE HOLTEN-PIND will appear Copenhagen 1985.- Historically speaking it is to be noted that Dignāga seems to be responsible for the following innovations in Buddhist epistemology: i) He is the first to classify yogijñāna as a particular form of pratyakşa. ii) He is the first to regard the visaya of pratyaksa and anumāna as respectively sva- and sāmānyalaksaņa. iii) He does not distinguish pramāna from pramānaphala ( = svasamvitti). These are the three main innovations of interest to us here. In my opinion the main source (among other minor ones) of influence in the formation of these innovations is to be found in Dignāga's Yogācāra background (cf. his Alambanaparīkņā, Prajñāpāramitāpiņdārthasamgraha (see esp. vv. 26-29, quoted,

Page #4

152
CH. LINDTNER
sanahitavāsanākṛtaviplavarupāḥ sarva eva vikalpāḥ8. Direct knowledge is always free from conceptual construction (vikalpa) but, in order to be valid in practice, it must be abhrānta, i.e. unimpaired by timira, etc. Otherwise we clearly err in whatever we do.
So far a higher degree of naive realism in regard to sensa can hardly be conceived. Apart from indriyajñāna Dharmakīrti (following Dignāga) recognizes three kinds of pratyaksa: mental knowledge (mānasa), self-awareness of all thoughts and mental phenomena (svasamvedana) and yogic intuition (yogijñāna). In all cases the visaya of pratyaksa is still svalakṣaṇa, said to be ultimately real (paramarthasat), arthakriyāsāmarthya being the ultimate criterion of the empirical reality of things and "facts". If an object (vastu, bhava, artha) is incapable of arthakriya it cannot serve any sensible purpose.
incidentally, Jñanaśrīmitranibandhavali, pp. 504-505, etc.]), more precisely in the śrutamayi bhumiḥ section (X) of the Yogacarabhūmi. Here (TD, No. 4035, fol. 189b 7 seqq.) we learn, inter alia that pratyakṣa (said of the artha) must be aparoksa, anabhyuhita, anabhyuhya and abhrānta. There are four kinds of pratyakṣapramāna, viz. that of rupindriya-, manasanubhava-, loka- and suddhapratyaksa, etc. I hope to revert to this and the parallel epistemological passages in Yogacara literature at some future occasion, and shall therefore confine myself to these brief remarks for the present.
8 Quoted (from an untraced source) in H. R. KAPADIA (ed.), Anekāntajayapatākā by Haribhadra Sūri. Volume I, Baroda 1940, p. 251. For the idea see e. g. PVin II, 24-28 and the references given in Steinkellner's translation, p. 43. Also PVin I, p. 58, 16-26. For the meaning of upaplava (or viplava) in PVin II, 27-28 see PV III, 212-214. The Yogacara background is unmistakable: upaplava corresponds to the grahyagrahakavikalpa (cf. e. g. note 26), i. e. to the paratantrasvabhava, whereas the status of parikalpitasvabhāva must be assigned to the field of anumana (cf. e. g. PV I, 84: yathapratitikathitaḥ sabdartho sav asann api...). PV III, 215 (q. v.) is also unintelligible without the lakṣaṇaniḥsvabhāvatā doctrine of Samdhinirmocanasūtra VII, 4, etc. A good instance of utpattiniḥsvabhāvatā occurs PV I (ed. R. GNOLI) p. 51, 1.9: na hi vikalpā yathārtham eva jayante... As known, paratantrasvabhava is pratyayodbhava.
See PV I, 166; III, 3; NB I, 13-15. Dharmakirti's use of the terms samvṛtisat and paramarthasat tends to create some confusion. In this regard PV II, 55ab (= PV III, 3ab) is significant: arthakriyasamartham yat tad atra paramarthasat. STEINKELLNER's translation (p. 94): "Was fähig ist, einen Zweck zu erfüllen, das ist hier in Wirklichkeit seiend (paramarthasat)" fails to see the point of atra which here means vyavaharataḥ. Similarly in PVin II, 20 where atra is translated as "in unserem System" whereas Dharmottara and Bu ston quite consistent with the context take it as 'gro ba 'dir "in the empirical world". The objection in PV III, 4a aśaktam sarvam iti ced... (cf. AO 41 [1980] 31) compels Dharmakirti to admit that the śakti experienced in sprouts, etc. is only maintained to be real samvṛtyä. Ultimately nirvyāpārāḥ sarvadharmaḥ (Pramāṇasamuccayavṛtti ad I, 9d, q. v.). While I would not outrule that Dharmakirti is here indebted to Bhavya's distinction between a mithya- and a tathyasamvṛti

Page #5

Marginalia to Dharmakīrti's Pramāņaviniscaya
153
After a thorough discussion with spokesmen of other schools - as is to be expected from a good commentator - concerning the number of pramānas, their nature, their object and the four types of pratyaksa within this frame-work, Dharmakīrti somewhat abruptly (PVin I, p. 78; NB I, 18, etc.) enters upon a discussion concerning the “result" attained by cognition, the pramānaphala. Dharmakīrti introduces this theme as unexpectedly as did his main authority (ācārya), Dignāga (Pramāņasamuccaya I, 8cd; Nyāyamukha, T. 1628; 3b 22–23 = T. 1629; 8c 16-17)*0. Till now the "object" of the pramāṇas has simply - but, no doubt, with a conscious naiveté - been taken for granted without raising questions about its ontological status. His attention has been directed to vyavahāra "usage", the everyday world of things, action, language, etc. where arthakriyāsāmarthya, as said, is the ultimate criterion of reality.
Now, however, we suddenly learn that there is really no external object ([bāhyārtha] PVin I, p. 90, 1. 17) and that the assumed "objects” are merely acts, or manifestations, of mind ([vijñaptimātra) PVin I, p. 9.4, 1. 14). From this point of view (already familiar to us from PV), obviously, there is no real or substantial difference between the agent, instrument and the result of cognition. If everything is "only mind" and if there is nothing "out there the distinction between these three aspects is only apparent (yathānudarsana)". The fact that consciousness from another angle appears bifurcated as subject (grāhaka, svābhāsa) versus object (grāhya, visayābhāsa) is really also an illusion. Consequently all cognition is, in the final analysis, merely cognition of
satya (defined as arthakriyāsamartha!) it does, however, seem more likely that he is once again under the influence of his Yogācāra background. In this milieu numerous speculations concerning the various subdivisions of samurti- and paramārthasatya were thriving. See L. DE LA VALLÉE POUSSIN, Vijñaptimātratāsiddhi. La Siddhi de Hiuan-tsang, Paris 1928-1948, pp. 549-551 (paramārthasamortisatya, etc.). Abhidharmakośa VI, 4 does not come into consideration.
10 Pramāṇasamuccaya, 8cd: savyāpārapratītatvāt pramānam phalam eva sat “The result of cognition is really (not different from] the means of cognition, because [the latter is, in fact, merely] assumed to perform a function (resulting in cognition)". - Cognition, then, does not grasp anything apart from itself, but somehow splits what, consequently, can only be cognition of itself into what seems to be two different parts. Obviously we here again find the underlying Yogācāra doctrine of a real paratantrasvabhāva appearing as an unreal pari kalpitasvabhāva.
"PVin I, 48 = PV III, 356. Here the prefix anu hardly has any particular significance; cf. PVin, p. 90, 1. 18-20: 'di la rnam pa gzag pa 'di ni ji Itar snan ba bzin yin gyi, de kho na ñid ji lta ba bzin du ni ma yin noll. Siddhiviniscayatīkā, p. 468, 1. 21: 469, 1.23; 495, 1. 11 reads: yathādarśanam eveyam mānameyaphalasthitih.

Page #6

154
CH. LINDTNER
cognition itself, or svasamvitti. Needless to add, Dignaga and Dharmakirti here adhere to the celebrated "idealistic" Yogacara doctrine of cittamātra which we know, from other sources, to have flourished in those days. However, they may seem to have done so at the heavy cost of coming into conflict with the presuppositions of their own "realistic" pramana-system. As we have seen, they seem on the one hand to presuppose the existence of external objects, and, on the other, to assume that there are no external objects.
Did our authors (in this respect I see no dissension between Dignaga and Dharmakīrti) like so many other Indian philosphers - simply accept these two truths as theoretically (not practically) irreconcilable by having recourse to agama, or did they, as we would expect from "critical philosophers", seek some way of reconciling these apparent contradictions 12?
-
To answer this question we must briefly look at the approach taken to the doctrine of cittamatra by the most influential Yogacara philosopher before Dignaga, viz. Vasubandhu, the author of the Vimsatikā, the Trimśikā, Karmasiddhi, Abhidharmakośa and Vyakhyāyukti, etc. 13. The latter text (which, incidentally, has been unduly neglected by modern Western scholarship) deals with sutra-exegesis and, inter alia, shows him as an adherent of the Mahāyāna doctrine of two truths: samurti being the visaya of laukikajñāna and paramartha being that of lokottarajñāna1. In the Trimśikā – which gives his vijñānapariņāmasystem and is based on agama - lokottarajñāna is maintained to occur in vijñaptimātratva where there is neither any grahya nor, consequently; any grahaka15. The initial stanza of the Vimsatikā - the yukti-pendant
12 Cf. my paper Atisa's Introduction to the Two Truths, and its Sources JIP 9 (1981) 161-214.
13 Cf. my forthcoming paper on Vasubandhu the Vaitulika.
14 See Vyakhyāyukti (TD, No. 4061, fol. 109b 7 110a 3): las dan rnam par smin pa dag ni kun rdzob tu rdzas su yod / don dam par ni med de / jig rten pa'i ses pa'i yul yin pa'i phyir ro || dam pa ni ye ses jig rten las 'das pa yin te / de 'i don yin pas don dam pa 'o // de gñis kyi ran gi mtshan ñid ni de 'i yul ma yin te / de 'i yul ni brjod du med pa 'i phyi 'i mtshan ñid yin pa 'i phyir ro || 'dir jig rten pa'i ses pa 'am / jig rten las 'das pa tshad ma yin źe na / gcig kho nar ni jig rten las 'das pa yin no // jig rten pa ni dbye ba yod de jig rten las 'das pa 'i rjes las thob pa gan yin pa de ni tshad ma yin no gźan ni tshad ma ma yin no //.
15 Trimśikā, 28-29: yadā tv ālambanam jñānam naivopalambhate tadā / sthitam vijñānamātratve grāhyābhāve tadagrahat || acitto 'nupalambho 'sau jñānam lokottaram ca tat / āśrayasya parāvṛttir dvidhā dauṣṭhulyahānitah ||. - For vijñānamātratve in 28c (rather than vijñaptimātratve) cf. Mahāyānasamgrahabhāṣya (TD, No. 4050, fol. 124a 4): ... ses bya'i mtshan ñid de no bo ñid gsum

Page #7

Marginalia to Dharmakīrti's Pramāņaviniscaya
155
to the Trimsikā - gives the syllogism expressing Vasubandhu's basic argument in favour of cittamātra as revealed in āgama: (p) vijñaptimatram evedam, (h) asadarthāvabhāsanāt; (d) yathā taimirikasyāsatkeśacandrādidarsanam 16.
In proving asadarthāvabhāsa(na) Vasubhandu avails himself of the old and well-known Madhyamika argument later technically known as the ekānekaviyogahetu"7. Vasubandhu could easily have added other stock-arguments known from numerous Mahāyāna texts to support his doctrine of cittamātra in refuting the existence of external objects 18.
Here, however, the important thing to note is that the approach of Dignāga and Dharmakīrti differs decisively from that of Vasubandhu. The latter stressed the "ontological approach arguing in favour of the unreality of external objects, whereas the former - without, of course, being unaware of the "ontological" approach - stressed the "epistemological" one: They start, almost like Descartes, with cognition as the indisputable first and show that cognition is impossible to explain if an absolute distinction is made between the means and the result of cognition. The "object" cannot be isolated from cognition: esse est percipi (... abhedo nīlataddhiyoḥ: PVin I, 55b, etc.). To use a later terminology Vasubandhu is a nirākāravādin assigning the status of parikalpitasvabhāva to the "object" (artha, grāhya) whereas Dignāga and Dharmakīrti are sākāravādins assigning the status of paratantrasvābhava to the "object"19. So, while definitely coming to the same conclusion as Vasubandhu, viz. cittamātra, their approach is certainly more advanced by their being able to account for the "object" as a part (arthāvabhāsa = visayābhāsa = grāhyākāra = grāhyāmša) of an apparently bifurcated but basically univalent "mind” and not just by discarding it as foreign to and disintegrated from mind.
It is, especially to-day one and a half millennium later, not easy to say what personal” motives may have lead Dignāga - and Dharmakīrti - to take an interest in logical and epistemological issues, but to some
mo // ses bya'i mtshan ñid de la ji ltar jug par 'gyur ba 'am /gan gis jug par 'gyur ba ste / rnam par ses pa tsam mo Il.
16 This reading of the second hemistich is supported by the early quotations found in Bhavya's Ratnapradīpa and Tarkajvālā.
17 Cf. my Nagarjuniana. Studies in the Writings and Philosophy of Nāgārjuna, Copenhagen 1982, p. 273 and Catuḥsataka, IX, 12-19.
18 Cf. Mahāyānasamgraha II, 14 with the Bhāsya (which, in my view, is authentic) and Trisvabhāvanirdesa, 35 (the authenticity of which seems less certain).
19 Cf. my forthcoming paper on Bhavya's Critique of Yogācāra in the Madhyamakaratnapradīpa, chapter IV.

Page #8

156
CH. LINDTNER
extent it must have been an outcome of historical circumstances, more specifically under the pressure of the necessity of having to solve some problems prevalent within the Yogācāra tradition. In case of Dharmakīrti we of course know about his controversy with Kumārila and other non-Buddhist philosophers 20, but somehow the problems he takes up are also to be seen in the perspective of the vehement controversy taking place in those years between Bhavya and Dharmapāla, Dharmakīrti's "religious teacher”, if we are to believe tradition on this point21. This, however, is not the place to go deeper into this problem where so much preliminary work still remains to be done 22
On this background, I think, we are enabled to interpret the epistemological works of Dignāga and Dharmakīrti in a more proper perspective than has hitherto been the case. We know, of course, that Dharmakīrti is mostly concerned with "scientific methodology" but the motive for this concern is not just "scientific curiosity" but, as I hope will be clear from what follows, a certain "religious urge”, a desire for moksa. This finally brings us back to the question about Dharmakīrti's attitude to the problem of two "truths", i.e. to the problem of reconciling the rational realism of Sautrāntika forming his startingpoint with the idealism of vijñaptimātratā which, from the stand-point of sākāravāda, denies the existence of external objects.
On the final page of PVin (p. 100, 1. 12 seqq.) an opponent asks how, as an adherent of vijñaptimātra denying that cognition has a real object (visaya), Dharmakīrti can speak of upaplava and the contrary (i.e. how he can distinguish true cognition from false cognition). This puts Dharmakīrti with his back against the wall and his reply is to be found in the Nyāyabhūşaņa (NBhūş) in its Sanskrit original: upaplavavāsanābhisamdhidosād aprabuddhasyāpy anāśvāsikam vyavahāram utpaśyann ekam apramānam ācaksīta, aparam ā samsāram avislistānubandhadrdhavāsanatvād iha vyavahārāvisamvādāpeksayā pramānam, sāmvyavahāri kasya caitat pramāṇasya rūpam uktam; atrāpi pare vimūdhā visamvādayanti lokam iti. - cintāmayım eva prajñām anuśīlayante vibhramavivekanirmalam ana pāyi pāramārthikam pramānam abhimukhīkurvanti 23. - In other words: In samsāra correct knowledge "works", wrong
20 See the indexes to STEINKELLNER's translations of Hetubindu and Pramāņaviniscaya.
21 Cf, my remarks in AO 41 (1980) 32 and the linking together of Dharmapāla and Dharmakīrti in the Anekāntajayapatākā, II, p. 36.
22 At present the most important tasks would be editions and translations, etc. of Bhavya's Yogācāratattvaviniscaya (Tarkajvālā V) and Dharmapāla's commentary on the Catuhsataka (T. 1571).
23 NBhūş, p. 57, 1. 14-19 (in l. 17 yuktam has been emended to uktam; in PVin I, p. 100, 1. 15 med pa'i should be deleted). Further fragments (apart from

Page #9

Marginalia to Dharmakīrti's Pramāņaviniscaya
157
knowledge does not. As long as we live under the law of karman our cognition - whether right or wrong – is empirical or conventional. To realize the ultimate cognition it is necessary to have resort to prajñā in its three phases: the first (not mentioned here) is srutamayī, the second (and the one of greatest interest to a critical philospher) is cintāmayī, while the third (based on the former) is bhāvanāmayī. Dharmakīrti thus admits two "levels" of pramāna – an empirical and a transcendental -- and adds that he has only said a bit about the latter: de'i (i.e. don dam pa'i ses pa'i acc. to Dharmottara) yan cha tsam bstan pa yin no (PVin, loc. cit., 1. 24–25). This remark can only refer to the commentary to PVin I, 28: yoginām api śrutamayena jñānenārthan grhītvā yukticintāmayena vyavasthapya bhāvayatām tannispattau yat spastāvabhāsi jñānam tat pratyaksam (loc. cit., p. 72, 1.30 – 72, 1. 1)24. This becomes somewhat clearer in the light of the following passage from the first chapter of PV: sarvesām viplave 'pi pramāṇatadābhāsavyavasthā ā āśrayaparāvýtter arthakriyāyogyābhimatasamvādanāt; mithyātve 'pi prasamanānu kūlatvān mātssamjñādivat25. From these passages we see that the pāramārthikam pramāņam is a spastāvabhāsi jñānam that presupposes the application of bhāvanāmayī prajñā and only occurs when āśrayaparāvýtti has taken place. This “revulsion of the basis” must take place before the intuition of a yogin becomes free from the
those noted by VETTER, loc. cit.) occur Siddhiviniscayatīkā, p. 489, 1.20–24; Jñānaśrīnibandhāvali, p. 419, 1. 14. See also K. MIMAKI, Blo gsal grub mtha', Kyoto 1982, p. 136. - Dharmottara's commentary to the final paragraph is to be found in the Dravyālamkārasvopajñatīkā (ed. by M. JAMBŪVIJAYA in Studien zum Jainismus und Buddhismus. Gedenkschrift für Ludwig Alsdorf, Wiesbaden 1981, p. 137): pāramārthikam api pramāņam na nirhetukam; na ca bhāvanāvyatirikto hetuh. bhāvanā ca sāmvyavahārikapramānaparicchinnārthavisayā. tataś ca tat sāmvyavahārikam pramāņam samyag nirūpitam pāramārthikajñānahetuḥ sampadyate, tatas tadvisayo yatnaḥ paramārthavisaya eva, mithyajñānena hi visayīkytā bhāvā nityādibhir ākārair bhāvyamānā na pāramārthikajñānanibandhanam bhavanti, anityādibhis tv ākārair bhāvyamānā nibandhanam bhavanty eva, tasmād ato vyāmoham vyāvartya paramārthanaye 'vatārayitavyo janaḥ, sthūlavisayatvād asya vyāmohasya; etad vyāmohanivetti pūrvikā ca paramarthaprāptiḥ. For Tib. see AO 41 (1980) 36. [Professor Vetter (12. 1. 1984) kindly suggests these corrections in the NBhūş-fragment: 'vāsanānabhisamdhi", "ānubandham, vibhramavivekam, and adds that his ,Übersetzung auf p. 101 so zu ändern ist, daß mi mkhas pas kyan (aprabuddhasyāpi) nicht auf das Sehen, sondern auf das Sichverlassenkönnen bezogen ist; ferner muß Versenkung' (p. 101, Z. 4 v. u.) ,Reflexion' werden“.]
24 See ERNST STEINKELLNER, New Sanskrit-Fragments of Pramāņaviniscayah, First Chapter, WZKS 16 (1972) 203. I have changed praśrutao to Srutao
25 R. GNOLI, The Pramāņavārttikam of Dharmakīrti. The first chapter with the autocommentary, Roma 1960, p. 51, 1. 3-5. This passage is translated by VETTER, Erkenntnisprobleme..., p. 37.

Page #10

158
CH. LINDTNER
dichotomy (vikalpa) of grāhyagrāhaka. See for instance Samtānāntarasiddhi (s. 80): gnas ma gyur pa'i phyir rnal byor pa gzun ba dan 'dzin pa'i rnam par rtog pa ma spans pa rnams kyis gźan gyi sems ses pa yan, tha sñad la mi slu ba ñid kyis gzugs la sogs pa mthon ba bźin du tshad ma ñid yin no 26, and PV III, 281: prāguktam yoginām jñānam teṣām tad bhāvanāmayam / vidhutakalpanajālam spaṣṭam evāvabhāsate ||27.
So far there is no reason to suspect that Dharmakirti (and Dignaga) differ seriously from the Vijñaptimātratāsiddhi of Vasubandhu apart from in one respect. This divergence is suggested by the terminology employed by Dharmakīrti in the passage quoted above: Where Vasubandhu speaks of two kinds of jñāna - a laukika and a lokottara - Dharmakirti speaks of two kinds (rupa) of pramāņa - a sāmvyavahārika and a paramarthika. The scheme is, of course, the same but the approach is entirely different. We have already seen how Dignaga and Dharmakirti in their attempt to account for cittamatra by advocating sākāravāda differed from Vasubandhu's nirākāravāda and thus were enabled to render a certain unity to their system quite alien to that of Vasubandhu. The same epistemological approach also accounts for the other important innovation within the development of Yogacāra with which we are here concerned. In this respect PVin I, 28 is significant: bhāvanābalataḥ spaṣṭam bhayadav iva bhāsate | yaj jñānam avisaṇvādi tat pratyakṣam akalpakam || 28
The decisive term is pratyakṣa. By classifying the highest form of cognition on a par with the most simple form of sensation (the term pratyakṣa applies in both cases) Dignaga (followed faithfully by Dharmakīrti) had indeed taken a new and bold step to render the relationship between samvṛtisatya and paramarthasatya more coherent and
26 Samtānāntarasiddhi - "die eine Abhandlung Dharmakirti's im Anschluß an Vimśatika 18 und 21 ist" (VETTER, op. cit., p. 65, n. 49) here in addition presupposes the tradition of Trimśikā, 28-29 (quoted n. 15). For grahyagrahakavikalpa see Madhyāntavibhāgabhāṣya I, 1. For asrayaparāvṛtti cf. LAMBERT SCHMITHAUSEN, Der Nirvana-Abschnitt in der Viniscayasamgrahaṇī der Yogacarabhumiḥ, Wien 1969, pp. 90-103.
27 See the namaskarasloka of PV for vidhutakalpanajala (said of Samantabhadra).
28 The v. 1. given in NBhūs, p. 171, 1. 13 (viz. pramāņam for pratyakṣam) misses the point Dharmakirti wants to stress (namely that yogijñāna is in fact pratyakṣa) entirely and is neither supported by Tib., Dharmottara nor any of the two other quotations of this verse hitherto traced; cf. AO 41 (1980) 36, n. 37 and Tantraloka, vol. VII, p. 170 (cf. G. TUCCI, Minor Buddhist Texts. Part I & II, Kyoto 1978, p. 594). See also ERNST STEINKELLNER, Yogische Erkenntnis als Problem im Buddhismus, in G. OBERHAMMER (ed.). Transzendenzerfahrung, Vollzugshorizont des Heils, Wien 1978, pp. 121-134.

Page #11

159
Marginalia to Dharmakirti's Pramāṇaviniscaya
intelligible than ever before in Buddhist tradition. To be sure, Dignāga was not the first to use the epithet pramāṇabhuta of Bhagavat but when he revived the term in the sense of "pratyaksa in person" the demands of his system gave it an interpretation without precedent 29. This innovation solved some old problems but also created new ones hardly foreseen by its author and his commentator. Here is an interesting field for future research.
The importance of bhāvanābala within this scheme now emerges clearly. It is generally agreed that bhāvanā possesses the power to transform or "digest" anything - be it real or unreal - so that it is clearly intuited without conceptual constructions 30. It is the power of habitual meditation that makes things familiar and obvious. It is the task of critical philosophy (yukti- cinta) to secure that only scientific facts are presented to be digested by bhāvanā. It derives its inspiration from agama. Philosophical critique thus becomes ancillary to religious experience. If we credit Vasubandhu with having created a system of vijñānapariņāma we may as well credit Dignaga (followed by Dharmakīrti) with having created one of *pramāṇapariņāma31, so to speak.
In view of its cardinal importance in his system it may seem surprising that Dharmakīrti does not devote much space to justify the role
29 See HATTORI, op. cit., p. 23 with n. 3 and ERNST STEINKELLNER, The Spiritual place of the Epistemological tradition in Buddhism (= Nanto Bukkyō 49 [1982]). The meaning of the epithet comes out clearly in a fragment to which STEINKELLNER has drawn attention in his Philological Remarks on Śākyamati's Pramāṇavārttikaṭīkā (Studien zum Jainismus und Buddhismus. Gedenkschrift für Ludwig Alsdorf, 1981, p. 290):... bhāvanābalaniṣpannanirmalāvikalpābhrāntajñānātmakatvad bhagavataḥ pratyakṣapramāņasvabhāvatā sākṣād asty eva... In his interpretation of Dignaga's verse and Dharmakirti's commentary (i. e. the pramāņasiddhi chapter of PV) STEINKELLNER (following VETTER) fails to clarify the distinction between Bhagavat as a sāmvyavahārika- and as a pāramarthika-pramāņa.
30 See e. g. PV III, 281-287 and Kambala's Alokamālā (known to Dharmakīrti) 57-60; 117-118, etc. In both authors bhāvanā and abhyasa are used interchangeably.
31 This point is decisive for an adequate interpretation of almost any technical term in the epistemological system. Correct knowledge (pramāņa, samyagjñāna) changes from having an object to having no object, from being conceptual (indirect) to being non-conceptual (direct), from being simple sensation to being pure intuition. The deliberately oscillating vagueness of terms such as arthakriyasamartha, avisamvada, svalakṣana, etc. is easily understood once it is recognized that they have to apply to all four kinds of pratyakṣa, etc. and, consequently, shift their meaning accordingly. Without being too general or too specific NB I, 1 summarizes the system of pramaṇapariņāma with unmatched brevity and comprehensiveness: samyagjñānapūrvikā sarvapuruṣārthasiddhir iti...

Page #12

160
CH. LINDTNER
of bhāvanā in his extant works. There are, however, historical grounds for this apparent omission. As shown elsewhere there are reasons to believe that Dharmakirti composed a work entitled Tattvanişkarṣa in which, inter alia, he discussed the necessity of bhavana for realizing tattvärtha. (I shall revert to this lost work of his below.) Moreover there are reasons to assume that Dharmakirti was familiar with a text in which the significance of bhāvanā is stressed, viz. Kambala's Alokamālā. And, of course, we should not forget that the importance of bhāvanā was a matter of common consent within Mahāyāna. On this background there was no need for Dharmakirti to delve further into the matter.
II
After this attempt of interpretating a part of the background and some of the basic motives in the formation of Dignaga's and Dharmakirti's epistemology we shall now direct our attention to some more specific historical and textual problems posed by the first two chapters of PVin as edited and translated by VETTER and STEINKELLNER.
In PVin I, 19d (!) and 21-27 Dharmakīrti discusses the third form of pratyaksa, viz. svasamvedana. To some extent his discussion is based on the one given in his earlier work PV. This is clear from the fact that some of the verses are taken over from that work, sometimes without change, sometimes with slight or considerable modifications and sometimes they are even rendered into prose. Moreover PVin expressly refers to PV 32. But apart from this well-known fact there is reason to assume that he is also basing this paragraph on an earlier work of his. This may be shown as follows. In the Nyāyāvatāravārtikavṛtti (NAVV) of Śantisūri3 we come across the following quotation dealing with vijñāna as svasamvedana:
-
(1) tadatadrūpino bhāvās tadatadrupahetujāḥ /
tat sukhadi kim ajñānam vijñānābhinnahetujam || (2) aviseṣe 'pi bahyasya viseṣāt pritiṭāpayoḥ/
bhāvanāyā viśesena nārtharūpah sukhādayah || (3) prajñādivad visisyante bhavanabalabhāvataḥ /
nārthena janitākāro buddhau bhogas tadātmanaḥ || (4) niyataviṣayatvena jñānākāro nivartitaḥ /
arthānvayatirekena vyāpto nīlādibhāsavat ||
32 PVin I, p. 74, 1. 5 (read gtan la phab); II, p. 62, 1. 8; II, p. 96, 1. 19 (read 'grel du).
33 Nyāyāvatāravārtika-vṛtti of Śrī Santi Sūri. Critically... edited... by Pandita DALASUKHA MALWANIYA, Bombay 1949, p. 20, 1. 31-32 and p. 21, 1. 1-10.

Page #13

Marginalia to Dharmakīrti's Pramāņaviniscaya
161
(5) bhinnābhaḥ sitaduḥkhādir abhinno buddhivedane /
abhinnābhe vibhinne ced bhedābhedau kimāśrayau // (6) nānātvaikatvalopah syād evam sati jagattraye /
tasmād antarbhavā ete cetanās ceti sādhitam // Now two of these stanzas are already known to us from PV (1 = PV III, 251 and 5 = PV III, 279) and they also occur as PVin I, 22 and I, 27, q.v. But what is the source of the remaining four stanzas? Where does the NAVV quote them from, directly or indirectly? To answer this question we may have a closer look at PVin I, 22-27 (VETTER's edition, pp. 64-72). The second verse of our quotation proves to be identical with PVin I, 23:
phyi rol khyad par med na yan || dga'dan yons su gdun ba dag/ goms las khyad par 'gyur bai phyir /bde sogs don gyi ran bżin min /34
Stanzas 3, 4, 6 of the quotation do not occur as verses in PVin but there are unmistakable traces of them all in the prose; compare 3 and 4 with VETTER's edition, p. 66, 1. 20–22. ... sňon po la sogs pa'i snan ba'i khyad par bżin no de'i khyad par la mi ltos par goms pa'i khyad par dari rjes su 'brel pa de dag ni blo ñid yin te / ses rab la sogs pa bžin no ll, and 6 with op. cit., p. 75, 1. 24: de'i phyir bde ba la sogs pa nan yin zin myon ba yan yin no ll.
Thus we see that all the verses of our quotation occur in some form or other in the prose of PVin, but also that their source is neither PVin nor PV. On the other hand they are certainly by Dharmakīrti. Hence we seem forced to assume that they hail from another of Dharmakīrti's works - now lost. The most natural thing to assume is that the NAVV is here quoting Dharmakīrti's Tattvanişkarşa, a text from which another fragment, four couplets dealing with bhāvanā, are already known from Bhavya's Madhyamakaratnapradīpa 35. This conclusion is important in several respects: First of all it gives us a vague general idea of the nature of his Tattvanişkarşa. It must have discussed svasamvedana and bhāvanā (most probably in connection with yogipratyaksa). In other words one of its main topics was the various forms of pratyakşa and, as we may assume from the title of the work, their relationship to tattva. As shown above this is exactly what we would have expected Dharmakīrti to deal with more extensively than he did in any of his
34 Emend PVin I, p. 66, 1. 14 to gdun ba; 1. 19 to gdun ba'i; 1. 30 to 'brel pa'i; p. 68, 1. 12 to gdun bar; 1. 25 to sin tu dan ba (D); p. 78, 1. 7 to logs sig tu.
35 See AO 41 (1980) 29. - For Dharmakīrti tattva is equivalent to dvayaSūnyatā (cf. especially PV III, 213; 360 and Vimsatikā, 28-29). It is in this sense we have to understand PV II, 253cd: muktis tu sūnyatādrstes tadarthāh sesabhāoanā (v.1.). So for Dharmakirti (as for Kambala, etc.) ũngata is grāhuagrāhakābhāva, i.e. parinispannasvabhāva.

Page #14

162
CH. LINDTNER
extant works. So here Tattvanişkarşa may be held to provide the "missing link”. Secondly this new fragment (even if it does not hail from Tattvanişkarşa, but from some other work of Dharmakīrti's, the title of which is not known), sheds some light on the formation of at least this part of PVin. Now we know that it was composed as a systematic reshuffle not only of PV but also of one or more) of the author's earlier works, among these the Tattvanişkarşa. This is decisive when we have to assess the position of PVin within the literary output of Dharmakīrti as a whole. Thirdly, we have now found a place for depositing some of the quotations from Dharmakīrti occurring in later philosophical works but not to be traced in any of his extant treatises 36
In order to gain a better understanding of the historical position of Dharmakīrti it is obviously incumbent upon us to determine hiş relationship to predecessors and opponents. Here we need not discuss the profound influence of Vasubandhu and Dignāga but will confine ourselves to point out some reminiscences of another and less known Buddhist philosopher, namely Kambala, the author of Alokamālā (ca. 500 A. D.) 87. There are at least two stanzas (57, 84) in the. Alokamālā - which itself betrays unmistakable influence from Vasubandhu and Dignāga - that here come into consideration: 57. bhāvyate yad yad evetaḥ pāramparyena bālišaih /
tat tad eva puraḥ khyāti bhāvanābalanirmitam / 84. iti buddhyā vibhāgo 'sya panditaiḥ parikalpyate /
abhāgasyāpi cittasya lokasamortisatyatah // With these compare PV III, 282 cd (= PVin I, 29 cd), 284 cd and 285 (= PVin I, 31):
abhūtān api paśyanti purato 'vasthitān iva // spastābham nirvikalpam ca bhāvanābalanirmitam // tasmād bhūtam abhūtam vā yad yad evābhibhāvyate / bhāvanā parinispattau tat sphutākalpadhīphalam |
36 For our purpose two stanzas are of special interest: nīlapītādi yaj jñānam bahirvad avabhāsate / tan na satyam ato nāsti vijñeyam tattvato bahih // tada peksā ca samvitter matā yä kartprūpatā / sāpy atattvam ataḥ samvid advayeti vibhāvyate Il. It is quoted Anekāntajayapatākā, Vol. II, p. 82 (with the wrong reading apeksayā for apeksā, and mistakenly ascribed to PV) and in the Syādvādakalpalatā to Šāstravārtāsamuccaya, 393. Cf. also PV III, 220 to which Yasovijayajī (ad 392) incidentally quotes a clause from Devendravyākhyā: citrajñāne hi yo nīlādiņ pratyavabhāsate, jñānopādhir jñānavisesano 'nubhavasvātmabhūta iti... (TD, No. 4217, fol. 196 a 3 (v. 1.]). For Devendrabuddhi, cf. FRAUWALLNER, Kleine Schriften, pp. 842-846. - For the background: Alambanaparikṣā, 6; Alokamālā, 26-27.
37 For this text see WZKS 26 (1982) 191–194. My edition will appear in Indiske Studier 5 (1984).

Page #15

Marginalia to Dharmakīrti's Pramāņaviniscaya
163
and PV III, 353 (= PVin I, 45) and III, 212:
avibhāgo 'pi buddhyātmā viparyāsitadarsanaih / grāhyagrāhakasamvittibhedavān iva laksyate // paricchedo 'ntar anyo 'yam bhāgo bahir iva sthitaḥ /
jñānasyābhedino bheda pratibhāso hy upaplavah // From these parallels we not only see one of Dharmakīrti's sources but also how, in his characteristic manner, he almost rewrites them rendering their meaning more precise and specific so as to suit his own purpose. Kambala, who follows the nirākāravāda of Vasubandhu, does not classify yogijñāna as pratyaksa, nor does he take svasamvitti as pramānaphala, nor does he, consequently, accept the triple division of vijñāna. These features are specific to the sākāravāda of Dignāga and Dharmakīrti, etc. - In the same way it is a task for future research to trace instances where Dharmakīrti has critical allusions to various nonBuddhist authors and thus reconstruct the living dialogue in which he was participating 38. (A tremendous task in itself, needless to add, is to trace and reconstruct the subsequent influence of Dharmakīrti's own contributions to the debate inside and outside the Buddhist fold.)
Moreover there are some problems with regard to the numbering and identification of the stanzas of PVin as edited by VETTER and STEINKELLNER. Some of these may be solved by searching for further fragments than have hitherto been identified in later sources.
First of all, to be sure, there seems to be no problem with regard to PVin I, 19d where abc occur on p. 58 whereas d first occurs on p. 62, an antaraśloka having been inserted as stanza 20. There are other instances of this awkward procedure, as we shall see, occurring in the second chapter of PVin. The Sanskrit of PVin I, 19 is quoted in Kalpalatāviveka, p. 46, 1. 27–28: mānasam cākşavijñānānantarapratyayodbhavam / tadarthāntaragrāhi sukhādinām svasamvedanam // 39.
With regard to stanzas 41 and 42 we face problems. 41 ab corresponds, as Vetter points out, to PV III, 332 cd whereas “Der Anfang
38
38 Thus, for instance, there are several interesting allusions to Vākyapadīya (VP), compare PV II, 2d with VPI, 13 ab; PVin I, 47b with VP I, 37; PV I, 21 with VP I, 32, etc. - On the other hand when Sākyamati ad PV II, 5 (TD, No. 4220, fol. 79 a 5) quotes Siddhasena's Nyāyāvatāra 2 (q.v.), this celebrated Jaina author surely is under the influence of Dharmakīrti, here as elsewhere. This is a new piece of evidence for the discussion about the date of Siddhasena.
39 I owe this reference to Ernst Steinkellner who, for pāda d also refers to NBhūş, p. 101, 1. 7. In a I have emended cāksuo to cāksa'. - For PVin I, 21 ab Steinkellner kindly calls my attention to Távarapratyabhijñavivstivimarsinī, I, p. 116, 1. 25; II, p. 361, 1. 11; Bhāmatī, p. 537, 1. 19; Nyāyakaạikā, p. 190, 1. 9. See Pāda-Index to PVin.

Page #16

164
CH. LINDTNER
des Verses 42 ist wohl genauso wie der Rest von Vers 41 in die Prosa aufgegangen" (op. cit., p. 90, note to line 7). This leaves us either with one verse of two padas and one of three pādas, or with one stanza of five pādas. As there seems to be no precedent to such a metrical irregularity in Dharmakirti the translators (or clerks) are apparently to be held responsible for this unsatisfactory state of affairs. On p. 88, 1. 23, however, we read: 'dod pa dan mi 'dod par snan ba ni rtog pa yin gyi dban po 'i blo ni ma yin no że na /, which is a literal prose version of PV III, 345ab: iṣṭāniṣṭāvabhāsinyaḥ kalpanā nākṣādhir yadi /.
To support this identification we must also account for the missing pada of 42a. On p. 90, 1. 3 we read: de ñid kyi phyir tshad ma dan bras bu dag yul tha dad pa yan ma yin no . Vetter identifies this with a line occurring in the Dharmottarapradīpa (p. 91, 1. 16): na pramanaphalayor visayabhedaḥ, but here the de ñid kyi phyir (*tasmat) as well as the kyan (*api) are missing. VETTER also calls attention to PV III, 350 ab: tasmād viṣayabhedo 'pi na, svasamvedanam phalam /. This reference, I think, solves our difficulty: de ñid kyi phyir (= tasmad) and yul tha dad pa yan ma yin no (= visayabhedo 'pi na) should be printed as stanza 42a (b) whereas tshad ma dan bras bu dag (=pramanaphalayor; cf. Dharmottarapradīpa, loc. cit.) originally was composed as prose by Dharmakīrti. There are, as known, numerous similar instances of this misrakavyakhyāna style in the first chapter of PV40. In this way we may retain Vetter's numbering and read two regular stanzas as we would expect Dharmakirti to have composed (PVin I, 41-42):
tadanyasamvido bhāvāt svasamvit phalam isyate / istāniṣṭāvabhasinyaḥ kalpana nākṣadhir yadi || tasmad viṣayabhedo 'pi na, svasamvedanam phalam uktam svabhāvacintāyām tādātmyad arthasamvidaḥ || Similar problems occur in PVin II. STEINKELLNER (p. 40) prints verse 28 as follows: /mtshan ñid de dan Idan pa yi // gtan tshigs de ni mi dmigs dan bdag dan bras bu źes bya gsum || kho na 'o, and in his translation (p. 44) he notes: "Der Vers ist nicht vollständig, aber die Kommentare enthalten keinen Hinweis auf eine mögliche Ergänzung". However, there is a solution. On p. 30 (!) we read as prose: rjes dpag bya dan de mtshuns la / (DERGE has //!) yod dan med la med par ni || nes pa. In the Nyāyāvatāravivṛtti of Siddharsi (p. 51, 1. 22-23 and elsewhere11) we find the following stanza: anumeye tha tattulye sadbhāvo nāstitāsati/niścitānupalambhātmakāryākhyā hetavas trayaḥ . So what we read on p. 30 is in fact the beginning of a verse - PVin II, 9 - the
40 Cf. GNOLI, op. cit., p. XXXI.
41 Cf. STEINKELLNER's translation of Hetubindu, pp. 207-208.

Page #17

Marginalia to Dharmakīrti's Pramāņaviniscaya
165
rest of which occurs as verse "28" on p. 40. No less than nineteen antaraślokas have been inserted! Verse 9 should then read:
/ rjes dpag bya dan de mtshuns la ll yod dan med la med par mi/
| res pa gtan tshigs mi dmigs dan || bdag dan 'bras bu żes bya gsum / We thus see that mtshan ñid de dan ldan pa yi (p. 40, 1. 2) and kho na'o (p. 40, 1.5) really belong to the prose whereas de ni (p. 40, 1.3) belongs nowhere and must have been inserted by the translators (or revisers) not seeing that the nes pa (p. 30, 1.2) would have saved them from the difficulties of metre. (The numbering of verses should, of course, be changed accordingly.)
According to Steinkellner verse 32, "völlig zerrissen” (p. 61, n. 182 of his translation, q. v.) as it is, is to be found on p. 42, 1.12-13, p. 44, 1.30-31 and p. 52, 1.10-11 giving us:
I med par nes pa'i 'bras bu can || mi dmigs, 'jug pa'i bye brag gis /
rnam bzi, bskal ba rnams la ni || med par nes pa yod ma yin/ Based on various parallels STEINKELLNER reconstructs this verse as follows (p. 117 of his edition): asajjñānaphalā prayogabhedād anupalabdhiḥ / caturvidhā viprakrstesv abhāvaniscayābhāvāt //. This, however, is impossible metrically and otherwise: in a we would thus have a javipulā (-U) and in d the penultimate would be long, etc. A literal reconstruction of pādas a and b must be: abhāvaniscayaphalānu palabdhiś caturvidhā /. Whatever we do there is no place for prayogabhedāt metrically and it can only be explained as belonging to the prose just as it does in the parallel passage in PV noted by STEINKELLNER (p. 45). The literal reconstruction above is also the only metrically satisfactory one. This brings us to the remaining two pādas of verse 32. As noted by STEINKELLNER bskal pa (read: ba) rnams la ni ll med par nes pa yod ma yin | has its paralles in NB II, 27 (later than PVin!): ...viprakrstesu ... abhāvaniscayābhāvāt. But no matter how we turn and twist this prose passage it is impossible to turn it into the two missing pādas so as to get the Sanskrit hemistich we are searching for. We should therefore seek it elsewhere. On p. 56, 1.4-5 we find a verse consisting of two pādas only (!), numbered "35" by the editor. A stanza of two pādas only is without precedent in Dharmakīrti, and this is obviously the missing hemistich of verse 32 which then reads: abhāvaniscayaphalānupalabdhis caturvidhā / isto 'yam arthaḥ sakyeta jñātum so 'tiśayo yadi //^2 (The numbering of the following verses is to be changed accordingly.)
“? In his translation (p. 66, n. 205) STEINKELLNER says: “Dharmakīrti hat... den letzten Pāda geändert, so daß nun nicht die Besonderheit (atisaya), sondern die qualifizierte Person (*visesavat) Object ist". But metrically there is no space for *vićesavat (or rather for *viếisto which would be the normal

Page #18

166
CH. LINDTNER
The Tibetan version of PVin II, 33 reads as follows: | snan ba mthon ba med pa dan // dños po mi dmigs pa la ni /
| rgyu mi srid par gyur pa na || dños po med par rtogs par 'gyur / This, as STEINKELLNER notes, has its parallel in PV I, 203 (= 205): drsyasya darśanābhāvakāraṇāsambhave sati / bhāvasyānupalabdhasya bhāvābhāvaḥ pratiyate ll. The Tibetan version of this PV verse is:
snan run mthon ba med pa'i rgyu // mi srid par ni gyur pa na /
snan ran dros po ma dmigs pa'i || dños po med pa rtogs par 'gyur /
In a note to his translation STEINKELLNER (p. 58) remarks: “Dharmakīrti hat den übernommenen Vers offensichtlich nicht nur umgestellt, sondern auch den Sinn geändert. Im PV hat der Vers die Aufgabe, die Nichtbeobachtung des Wesens (svabhāvānu palabdhi) zu beschreiben. Hier, im PVin, hat er aber die Aufgabe, alle eben besprochenen Arten der Erkenntnis des Nicht vorhandenseins zusammenzufassen. ... Eine Rückübersetzung des problematischen Verses wäre folgend denkbar: *dréyadarśanābhāve bhāvasyānupalabdhasya ca / kāraņāsambhave sati bhāvābhāvaḥ pratīyate //”. This reconstruction is scarcely possible: The reading kāraṇāsambhave sati would give the impermissible ja-vipulā if taken as pāda c. It must be retained as pāda bwhich, as it stands, violates the metre - just as pāda d must be retained unaltered. Moreover if dréyadarsanābhāve was correct we would have to have mthon ba med na for mthon ba med pa in the first pāda of the PVin version. Consequently we are forced to admit that the Sanskrit form of PVin could not have differed from that of PV (given above). Still it is clear that the Tibetan version of the PVin verse is peculiar and requires an explanation. What has happened, in my opinion, is that the translators here - as elsewhere 43 - were under the influence of Dharmottara's exegesis (cf. STEINKELLNER, loc. cit.). The result, as we see from their version of PVin 33, was an unsatisfactory compromise which is neither faithful to the Sanskrit nor to Dharmottara's exegesis (which does, it should be added, not pretend to be verbatim).
Again on p. 100 there seems to be a problem of verse "67" ( = 66 according to the revised numbering) consisting of five pādas. Here ldog pa sgrub pa ma mthon tsam (1.5) obviously - as STEINKELLNER notes - corresponds to vyatirekasādhanasyādarśanamātrasya. It is, however, quite impossible to reduce this passage to a pāda of eight syllables. So again the translators (or scribes) have mislead us by writing mthon for mthon ba, etc. Verse 66 then should read: yasyādarsanamātrena vyatire
rendering of khyad par can). In fact nothing has been changed: atiśayo is simply to be taken as an adjective.
43 Cf. VETTER's observation, PVin I, p. 104, n. 37 and p. 106, n. 65.

Page #19

Marginalia to Dharmakīrti's Pramāņaviniscaya
167
kaḥ pradarsyate / tasya samsayahetutvāc chesavat tad udāhstam Il. Incidentally this and the previous verse (i.e. 65 with the wrong reading na yuktādrstimātrena for na cādarśanamātrena in a) are quoted NBhūş, pp. 142–143.
On p. 106 pāda d of verse "70" (= 69) is problematic. STEINKELLNER reconstructs the pāda (p. 118): asiddhiyojanāvācyā. This is an exact rendering of ma grub sbyar ba... brjod bya min (p. 106, 1.4 and 1.8) but metrically impossible. We should read PVin II, 69d as na vācyāsiddhiyojanā / exactly as PV I, 18d, to which there is no alternative, metrically or otherwise 14
Further problems face us on pp. 112-113. Certainly the first two pādas of "74" (in fact = 71 cd, as we shall see) are correctly identified by STEINKELLNER as PV I, 26 ab: tasmăd vaidharmyadrstānte nesto 'vaśyam ihāśrayaḥ 1. For pādas c and d the Tibetan version has: / yan na rgyu yi dños por te // me med na yan du ba ste /- the exact equivalents of which are found in the Svavṛtti to PV as pointed out by STEINKELLNER: ... hetubhāvo vā... dahanābhāve ca dhūmaḥ. STEINKELLNER consequently (p. 118) reconstructs pāda d as: dahanābhāve ca dhūmaḥ. Now this again is metrically impossible whatever you do and STEINKELLNER wisely abstains from reconstructing pāda c45. Again the Tibetan version is misleading. This passage in PVin was not composed as verse by Dharmakīrti but as prose, i.e. exactly as in the parallel passage in PV. But this fact apparently leaves us with the new problem of a verse consisting of one hemistich only. There are, to be sure, no traces of the missing hemistich in the sequel. To solve our puzzle we most turn back to p. 106 (!) where we find (1.31-32) the following prose passage: de bžin du géan la yan tshad ma gian gyis gnod pa srid de /. This is an exact rendering of PV I, 20 ab where the translators (or revisers) once again failed to recognize the verse: tathānyatrāpi sambhāvyam pramāṇāntarabādhanam /. To be sure, the corresponding Tibetan version of PV runs: | de bzin du ni gzan la yan || tshad ma gźan gyis gnod pa srid /. This then
44 Probably the prose of PVin II, p. 106, 1. 7 is corrupt. See the corresponding passage of the Tib. version of PV (TD, No. 4216 fol. 268 a 3): gan yan ma grub pa'i sbyor ba de bżin du mthun pa'i phyogs la yod pa dan / med pa zes bya ba la sogs pa la yan ci rigs par brjod par bya'o zes bya ba de lta bu la sogs pa/ ma grub sbyar ba de yan brjod bya min ll. Here ma grub sbyar ba brjod bya min renders na vācyāsiddhiyojanā, whereas de yan (= sāpi) belongs to the prose just as the initial ma grub pa'i sbyor ba does. In PVin, I. 7 we should accordingly for de yan read de «lta bu la sogs pa / ma grub sbyar ba de> yan. The lacuna may either be due to the omission of a scribe (haplography) or to an instance of editorial banalization. PVin, 1. 4, then, should be brought back to its original prose form.
45 Like 57a (= 55a) missing in STEINKELLNER'S "Versindex”.

Page #20

168
CH. LINDTNER
is the first hemistich of “71"(= 70). The second (= 70 cd) is to be found on p. 110, 1.25–26 (where STEINKELLNER takes it as 71 ab). The first hemistich of the next verse (71 ab) affords no problems and corresponds to STEINKELLNER'S “71 cd". Then follow two antaraslokas ("72" and "73" = 72 and 73) and finally the two missing pādas of 71 that we were looking for: the first two pādas of "74". This solves our problem and it only remains to be noted that the final verse ("75") should consequently be numbered as 74.
These observations have mainly been concerned with the identification and numbering of the 133 verses of PVin I and II. No doubt the prose still poses several – mostly minor - problems as far as text, translation and interpretation are concerned. I shall have to postpone remarks on these for another occassion. Considering the difficulty of Dharmakīrti's thought and style and the problems connected with the transmission of his works, this is only natural. Dharmakīrti is not easy reading and, as we know, prāyaḥ prākrtasaktir apratibalaprajño janaḥ. Experience from Classical philology, the elder sister, as it were, of Indian and Tibetan philology, shows that abstruse philosophical texts have to go through several editions before we get a text edited to the satisfaction of all - if ever. It would therefore be quite absurd if the above observations should leave the impression that the work done by VETTER and STEINKELLNER done so far is not first rate. At present no one could have done a better and more admirable job than they have.
Much work remains to be done and Dharmakīrti surely deserves it. Speaking of one of Proclus' writings COLERIDGE once remarked that "The most beautiful and orderly development of the philosophy which endeavours to explain all things by an analysis of consciousness, and builds up a world in the mind out of materials furnished by mind itself, is to be found in the Platonic Theology of Proclus”. Had he had the chance to know the works of Dharmakīrti he might well have met hesitations in deciding whom of these congenial almost contemporaries of widely different background to let carry off the palm.
Pāda-Index to PVin 146 akārakam api svayam 37d adarsanāj jagaty asminn akşadhir yady apekseta 5c anapekṣitasādharmyao atadātmani tādātmyao 52c anirdesasya vedakam
53 a
54c
150
46 The numbering of verses follows VETTER's edition from 1966. Subsequent identifications by STEINKELLNER (cf. n. 24) or myself are included and few emendations - all obvious - have tacitly been made. Note that 10, 11ab, 12, 17ab, 24, 30, 33(c)d and 59(c)d are still missing in Sanskrit.

Page #21

Marginalia to Dharmakirti's Pramāṇaviniścaya
anupaplutacakṣuṣām anyataḥ pratipattitaḥ anyathaikasya bhāvasya anyathaivāvabhāsante
anyasyānyatvahāneś ca apratyakṣasya sambandhad apratyakṣopalambhasya abhinnavedanasyaikye
abhinnābhe vibhinne ced abhinno buddhivedane abhutan api pasyanti abhedo nilataddhiyoh abhrāntam abhilāpinī ayogolakavahnivat
arthagrahaḥ katham satyam arthasya dṛṣṭāv iva tad arthasyasambhave bhāvāt arthasyeṣṭā prameyatā artḥādhigamanātmanā arthāpaye 'pi netradhiḥ arthena ghatayaty enam arthopayoge 'pi punaḥ avibhago pi buddhyātmā aviseṣe 'pi bahyasya avedyavedakākārā asakyasamayo hy ātmā astiyam api ya tv antar°
atma meyaḥ phalam svavit asin me kalpanedṛśī
iti ced grahyatām viduḥ iti vetti na purvoktā iti sa yogyatā mānam iyam sarvatra samyojyā iṣṭāniṣṭāvabhāsinyaḥ
ikṣate sākṣajā matiḥ
uktam svabhāvacintāyām ucchinna sarvavastuṣu upaplavasamudbhavā
47b
1d
49 a
46 c
50 a
lc
55c
26a
27c
27b
29 c
55b
4b
25b
44 c
15c
3 a
43d
36d
6d
34 a
5a
45a
23 a
39 a
21 a 53 c
57d
14b
20b
14c
57 c
56 c
41 c
13d
42 c
25d
53d
upayogāviseṣataḥ
ekatra drsto bhedo hi ekasyāpi tadātmanaḥ
kalpana nākṣadhir yadi kāmasokabhayonmada keśādijñānabhedavat kriyate vidyamānāpi kvacin nanyatra dṛśyate
khādīnām svasamvedanam
169
tat pratyakṣam akalpakam tatrātmaviṣaye māne tatrapy anubhavātmatvāt tat sukhadi kim ajñānam tat sphuṭakalpadhiphalam tathakṛtavyavastheyam tatha pratyeti nānyathā tathavabhāsamānasya tathaivadarsanat teṣām tadatadrupahetujāḥ tadatadrūpino bhāvās tadarthantaragrahi su tadanyasamvido bhāvāt tadekatvasya hanitaḥ taddṛṣṭāv eva dṛṣṭesu taddṛṣṭes taddhvanau smṛtiḥ
6b
16 a
53b
41d
29a
40b
48 c
16b
19b
7 c
39d
grhitva samkalayyaitat grahakākāraviplavā grāhakākārasamkhyātā grahyagrahakalakṣaṇā
58d
40d
grahyagrahakavaidhuryāt 38 c
45c
grahyagrahakasamvitti grahyagrahakasamvidām
48d
caurasvapnadyupaplutāḥ
jñanasya hetur artho 'pity jñānākārārpaṇakṣamam
29b
43 c
20d
28d
56a
57 a
22 c
31d
40 a
7d
43 a
47 a
22b
22 a
19c
41 a
49d
18a
11d

Page #22

170
CH. LINDTNER
8c
23c
taddhetutve samam dvayam 3d purato 'vasthitān iva 29d tadrūpam nästi tattvatah 51 b pārvāparaparāmarsao tadrūparahitā api
46d pratibaddhasvabhāvasya tadvaśāt tadvyavasthānādi 37c pratisedhāc ca kasyacit tasmāt prameyādhigateḥ 34cpratītiḥ kalpanārthasya 4c tasmāt prameyādhigateh 35a pratyakşam kalpanāpodham 4a tasmād bhūtam abhūtam vā 31a pratyakşam anumānam ca la tasmād vişayabhedo 'pi 42a pratyaksābhaḥ + + + + 33c tasmād višesavisayā
17c pratyakse 'pi pramānatā 3b tasyā nānubhavo 'parah 38b pramānam meyarūpatā 34d tādātmyād arthasamvidaḥ 420 pramānam svātmavedanam 580 tādęso 'nyādņšo 'pi vā 43b pramāṇāntarasadbhāvaḥ 2c tena nārthāntaram phalam 36b pramāņetarasāmānyao 2a te yogyāḥ svātmasamvidi 57b pramāne sadrśātmanā lb teşām ataḥ svasamvittir
21c
buddhāv apratibhāsanāt 15b dadhānam tacca tām ātmany 36c dūre yathā vā marusu 47cbhayādāv iva bhāsate (v: 1.) 28b odygādis taimirādivat
54d
| bhāvanaparinispattau bhaana
31c drstasamkalanātmakam 8b bhāvanābalataḥ spastam 28a dosodbhavā prakrtyā sā 54a bhāvanāyā višeseną i
bhāvā yena nirūpyante 51a dhir abhedam vyavasyati 50d
bhinnakālam katham grāhyam 20a na ca tat tādrg arthavat 32d
bhinnābhaḥ sitaduh khādir 27a na jāne 'ham apidȚsam 44d
bhedavān iva laksyate 45d na tasmād bhinnam asty anyat 16c
bhedābhedavyavasthaivam 25c. onantarapratyayodbhavam 19b bhedābhedau kimāśrayau' 27d na vikalpānubaddhasya 32a bhrāntir nāmopajāyate 52b na siddham bhedasādhanam 260 na svasamvedanam phalam 42b manträdyupaplutākṣānām 46a na hi muktvärtharūpatām 34 b
mahān alpo 'pi dęśyate 47d nānārūpāvabhāsinah
49b mānasam cāksavijñānā nānyo 'nubhāvyo buddhyāsti 38a
meyamānaphalasthitiḥ 48b nābhijalpānuşangini (v. 1.) 210 meyamānaphalasthitih 56d nābhedo 'rūpadarśanāt 50b nārthadrstih prasidhyati 55d yaḥ prāg ajanako buddher 6a närtharūpaḥ sukhādayaḥ 23d yaj jñānam avisamvādi 28c nārthasamnidhim īkseta 9c yatsannidhāne yo drstas nivartetecchayā matiḥ
yathākathamcit tasyārthao 44 a yathānudarsanam ceyam
48a paricchedātmatātmani 58b yathā bhrāntair nirīksyate 39b punar vikalpayan kimcid 14a yathā mrcchakalādayah 46b
19a
11c
9b

Page #23

Marginalia to Dharmakīrti's Pramāņaviniscaya
171
yathā rāgādivedane yadā tadā na samcodya yad yad evābhibhāvyate yan naivam tad vibhedavat yasmād ekam anekam ca
56b 40c 31b 26b 51c
que
8a 49c 6c 7b 350 170
37 a
rūpam teşām na vidyate o rūpam muktvāvabhūsinaḥ rūpabhedam hi paśyanti
51d 44b 50c
55a 36a 35b 350 52 a
14d 9a 33a 22d
samketasmaranopāyam satyāḥ katham syur ākārās sa paścâd api tena syād sambandham laukikīm sthitim
sambandho na prasidhyati sarvaivendriyajā matiḥ savyāpāram ivābhāti sahopalambhaniyamād sā ca tasyātmabhūtaiva sādhanam meyarūpatā sādhane 'nyatra tatkarma sādharmyadarśanāl loke sāmarthyena samudbhavāt sāmānyam buddhyabhedataḥ sā yogyateti ca proktam siddhiḥ syād vyatirekataḥ sidhyed asādhanatve 'sya sukhādīnām ananyabhak so'rtho vyavahito bhavet stimitenantarātmanā osthiter anyadhiyām gateh sthito 'pi caksuşā rūpam spastārthapratibhāsitā smaraņād abhilāseņa smārtam sabdānuyojanam svapne 'pi smaryate smärtam svayam saiva prakāšate svasamvit phalam isyate
ovasthāyām indriyād gatau vikalpotthāpitā sā ca vikalpo 'vastunirbhāsād vijñānābhinnahetujam vitathapratibhāsini viparyäsitadarsanaiḥ · vibhaktalaksaņagrāhyao
visesaņam visesyam ca višesāt prītitāpayoḥ visamvādād upaplavah °vyavasāyena neha tat vyavahāraḥ pravartate vyāpāreņa svakarmaņi
54b
45 b 39c
7а
160 58c 59d 26c 21 b 5d 13b
2b 13c 32b
23b 33b 52d 180 37b
18c
śabdenāvyāpstākşasya ośūnye tac cāksuse katham
15a 8d
5b 32c 38d 41 b
samvitsāmarthyabhāvinaḥ samsargād avibhāgaś ced samhrtya sarvataś cintām
18b 25a 13a
hetutvam eva yuktijñā
20c
A revised edition of the stanzas of PVin II"
anumānam dvidhā svārtham trirūpāl lingato 'rthadȚk /
atasmims tadgraho bhrāntir api sambandhatah pramā // 2. yo hi bhāvo yathābhūtaḥ sa tādrglingacetasaḥ /
hetus tajjā tathābhūte tasmād vastuni lingidhih //
47 The numbering of the verses follows my remarks above. Spaced words or clauses are reconstructed from Tib. and parallel Sanskrit sources. In 55c I prefer the v. 1. teşām for tasmāt. STEINKELLNER's reconstruction of 52 ab (his 53 ab) violates the metre. A few emendations, mainly of misprints, have tacitly

Page #24

172
CH. LINDTNER
3. lingalingidhiyor evam pāram paryena vastuni /
pratibandhāt tadābhāsasūnyayor apy avañcanam // tayos tadrūpasūnyayos tadrūpādhyavasāyatah / tadrūpāvañcakatve 'pi krtā bhrāntivyavasthitih | manipradīpaprabhayor manibuddhyābhidhāvatoḥ / mithyājñānāvišeşe 'pi việeso 'rthakriyām prati | yathā tathāyathārthatve 'py anumānatadābhayoh / arthakriyānurodhena pramāṇatvam vyavasthitam // prāmāṇyam vastuvisayam dvayor arthabhidām jagau / pratibhāsasya bhinnatvād ekasmims tadayogataḥ // atadrūpaparāvịttavastumātraprasādhanāt /
sāmānyavisayam proktam lingam bhedāpratisthiteh / 9. anumeye 'tha tattulye sadbhāvo nāstitāsati /
niscitānupalambhātmakāryākhyā hetavas trayah / 10. ayogam yogam aparair atyantāyogam eva ca/
vyavacchinatti dharmasya nipāto vyatirecakah // 11. višesanavišesyābhyām kriyayā ca sahoditah /
vivaksāto 'prayoge 'pi tasyārtho 'yam pratīyate | . vyavacchedaphalam vākyam yataś caitro dhanurdharah /
pārtho dhanurdharo nīlam sarojam iti vā yathā // 13. pratiyogivyavacchedas tatrāpy arthesu gamyate /
tathā prasiddheḥ sāmarthyād vivaksānugamād dhvaneh ll 14. tad ayogavyavacchedād dharmī dharmavisesanam/
tadvisistatayā dharmo na niranvayadosabhāk // 15.
nivrttyabhāvas tu vidhir vastubhāvo 'sato 'pi san / vastvabhāvas tu nāstīti paśya bāndhyavijrmbhitam // nivýttir yadi tasmin na hetor vrttiḥ kim isyate / sāpi na pratisedho 'yam nivsttiḥ kim nişidhyate || vidhānam pratişedham ca muktvā sābdo 'sti nāparaḥ /
vyavahāraḥ sa căsatsu neti prāptātra mūkatā // 18. satām ca na nişedho 'sti so 'satsu ca na vidyate /
jagaty anena nyāyena nañarthaḥ pralayam gataḥ // desakālanisedhaś ced yathāsti sa nişidhyate /
na tathā na yathā so 'sti tathāpi na nişidhyate / 20. tasmād āśritya sabdārtham bhāvābhāvasamāśrayam /
abāhyāsrayam atrestam sarvam vidhinişedhanam //
16.
17.
19.
been made. PVin II, 7 occurs Nyāyāvatāravārtikavrtti, p. 70, 1. 21 and Nyāyaviniscayavivarana, II, p. 4, 1. 2. See also NBhūş, p. 289, 1. 24 and Isvarapratyabhijñāvivrtivimarsinī, p. 226, 1. 16. PVin II, 53 cd occurs in Nyāyakanikā, p. 138, 1. 13 and NBhūş, p. 288, 1. 9. These identifications were kindly communicated to me by Ernst Steinkellner (9. 11. 1983).

Page #25

Marginalia to Dharmakīrti's Pramāņaviniscaya
173
21. tābhyām sa dharmi sambaddhaḥ khyāty abhāve 'pi tādęśaḥ /
sabdapravștter astīti so 'pīsto vyavahārabhāk l 22. anyathā syāt padārthānām vidhānapratişedhane /
ekadharmasya sarvātmavidhānapratisedhanam // 23. anānātmatayā bhede nānāvidhinisedhavat /
ekadharminy asamhāro vidhānapratisedhayoh // 24. ekam dharminam uddibya nānādharmasamāśrayam /
vidhāv ekasya tadbhājam ivānyeşām upeksakam / 25. nişedhe tadviviktam ca tadanyesām apeksakam /
vyavahāram asatyārtham prakalpayati dhir yathā || 26. tam tathaivāvikalpyārthabhedāśrayam upāgatāḥ /
anādivāsanodbhūtam bādhante 'rtham na laukikam // tatphalo 'tatphalas cārtho bhinna ekas tatas tataḥ /
tais tair upaplavair nītasañcayāpacayair iva || 28. atadvān api sambandhāt kutaścid upanīyate /
drstim bhedāśrayais te 'pi tasmād ajñātaviplavāḥ || 29. sarve bhāvāh svabhāvena svasvabhāvavyavasthiteh/
svabhāva parabhāvābhyām yasmād vyāvrttibhāginah // 30. tasmād yato yato 'rthānām vyāvrttis tannibandhanāḥ /
jātibhedāḥ prakalpyante tadviseşāvagāhinaḥ 11 31. tasmād yo yena dharmeņa visesaḥ sampratīyate /
na sa sakyas tato 'nyena tena bhinnā vyavasthitiḥ // 32. abhāvaniscayaphalānu palabdhiś caturvidhā /
isto 'yam arthaḥ sakyeta jñātum so 'tiśayo yadi // 33. drsyasya darśanābhāvakāraṇāsambhave sati /
bhāvasyānupalabdhasya bhāvābhāvah pratīyate // 34. istam viruddhakārye 'pi deśakālādyapeksanam/
anyathā vyabhicāri syād bhasmevāsītasādhane // 35. svayam rāgādimān nārtham vetti vedasya nānyataḥ /
na vedayati vedo 'pi vedārthasya kuto gatih // 36. tenägnihotram juhuyāt svargakāma iti srutau /
khādec chvamāmsam ity esa nārtha ity atra kā pramā // 37. prasiddho lokavādaś cet tatra ko 'tīndriyārthadrk /
anekārthesu subdesu yenārtho 'yam vivecitaḥ // 38. svargorvaśyādiśabdaś ca drsto 'rūdhārthavācakah /
śabdāntaresu tādrksu tādréy evāstu kalpanā // 39. 40. prasiddhiś ca nļņām vādaḥ pramānam sa ca nesyate /
tataś ca bhūyo 'rthagatiḥ kim etad dviştakāmitam // 41. atha prasiddhim ullarghya kalpane na nibandhanam /
prasiddher apramānatvāt tadgrahe kim nibandhanam //

Page #26

174
CH. LINDTNER
42. utpäditā prasiddhyaiva sankā sabdārthaniscaye /
yasmān nānārthavrttitvam sabdānām tatra dęśyate //
46
45. anyathāsambhavābhāvān nānāśakteḥ svayam dhvaneḥ /
avasyam sankayā bhāvyam niyāmakam apaśyatām // esa sthānur ayam mārga iti vaktīti kascana /
anyaḥ svayam bravīmīti tayor bhedah parīksyatām // 47. sarvatra yogyasyaikārthadyotane niyamaḥ kutah /
jñātā vātīndriyah kena vivakşāvacanād ste | 48. vivaksā niyame hetuḥ samketas tatprakāśanah /
apauruseye sā nāsti tasya saikārthatā kutaḥ // 49. svabhāvaniyame 'nyatra na yojyeta tayā punah /
samketaś ca nirarthaḥ syād vyaktau ca niyamaḥ kutaḥ // 50. yatra svātantryam icchāyā niyamo nāma tatra kah /
dyotayet tena samketo nestām evāsya yogyatām // 51. vrttibuddhipūrva katvād istā sāpravsttiphalā /
anyā tu pravrttiphalā tasyā nimittadarsanāt ll. 52. tadbhāvamātrā nurodhe svabhāvo hetur ātmani /
upādhyapeksaḥ śuddho vā nāśe kāryatvasattvavat // ahetutvād vināśasya svabhāvād anubandhitā /
sāpekṣāņām hi dharmāņām nāvasyambhāviteksyate // 54. etena vyabhicāritvam uktam kāryāvyavasthiteḥ /
sarvesām nāśahetūnām hetumannāšavādinām // 55. arthakriyāsamartham yat tad atra paramārthasat /
asanto 'ksamikās tesām kramākramavirodhataḥ // asāmarthyāc ca taddhetor bhavaty esa svabhāvatah /
yatra nāma bhavaty asmād anyatrāpi svabhāvataḥ / 57. kāryam svabhāvair yāvadbhir avinābhāvi kārame /
hetus tadvyabhicāre sa hetumattām vilanghayet // 58. nityam sattvam asattvam vähetor anyānapeksaņāt /
apeksāto hi bhāvānām kādācitkatvasambhavaḥ // 59. agnisvabhāvaḥ sakrasya mūrdhā yady agnir eva sah/
athānagnisvabhāvo 'sau dhūmas tatra katham bhavet // 60. dhūmahetusvabhāvo hi vahnis tacchaktibhedavān/
adhūmahetor dhūmasya bhāve sa syād ahetukah // anvayavyatirekād yo yasya dęsto 'nuvartakaḥ /
svabhāvas tasya taddhetur ato bhinnän na sambhavaḥ // 62. kāryakāranabhāvād vā svabhāvād vā niyāmakāt /
avinābhāvaniyamo 'darśanān na na darsanāt // 63. avasyambhāvaniyamaḥ kah parasyānyathā paraih /
arthāntaranimitte vā dharme vāsasi rāgavat //
56.
athan
61

Page #27
Marginalia to Dharmakirti's Pramanaviniscaya 175 64. arthantaranimitto hi dharmah syad anya eva sah / pascad bhavan na hetutvam phale 'py ekantata kutah || 65. na cadarsanamatrena vipakse 'vyabhicarita / sambhavyavyabhicaratvat sthalitandulapakavat || yasyadarsanamatrena vyatirekah pradarsyate/ tasya samsayahetutvac chesavat tad udahstam || hetos trisv api rupesu niscayas tena varnitah | asiddhaviparitarthavyabhicarivipaksatah || na ca nastiti vacanat tan nasty eva yatha yadi / nasti sa khyapyate nyayas tada nastiti gamyate || yady adsstya nivsttih syac chesavad vyabhicari kim vyatireky api hetuh syan navacyasiddhiyojana || tathanyatrapi sambhavyam pramanantarabadhanam / tasmat tammatra8ambamdah 80abhaoo bhaoam eva xa || 71. misartaget karama Da kargam agabhicarata / tasmad vaidharmyadsstante nesto 'vasyam ihasrayah || 72. anyathaikanivittyanyavinivrttih katham bhavet / nasvavan iti martyena na bhavyam gomatapi kim ll samnidhanat tathaikasya katham anyasya samnidhih / goman ity eva martyena bhavyam asvavatapi kim || 74. hetusvabhavabhavo 'tah pratisedhe ca kasyacit / hetur yuktopalambhasya tasya canupalambhanam //