Early Tibetan Ideas On Ascertainment Of Validity

Author(s): Ernst Steinkellner
Publisher: Ernst Steinkellner
Catalog link: https://jainqq.org/explore/269440/1

Page #1

Early Tibetan Ideas on the Ascertainment of Validity
(nges byed kyi tshad ma)
Ernst STEINKELLNER
(Vienna)
The intensive efforts of the Tibetan epistemologists in appropriating the Indian epistemological tradition in continuation particularly of the "seven works of Dharmakirti" resulted in the fact that their explanations must be considered as a qualified testimony for any interpretation of the Indian tradition, a fact that has to date rarely been seriously acknowledged. Of course, the contextual presuppositions, systematic as well as philological, must always be taken serious. But in view of the depth and consequence with which the leading Tibetan scholars treated the inherited themes and problems their interpretational opinions can no longer simply be neglected. On the other hand it is well known and clearly exemplified in Leonard van der Kuijp's pioneer study of the early period of Tibetan epistemology", that the appropriation and understanding of the themes, theories, concepts and problems raised and developed by Dharmakirti and his school was a historical process that was governed by the dialectics of trial and error. A host of scholars from the eleventh century onwards took part in this living and lively process until by the fourteenth and fifteenth century?) authoritative commentaries, above all on the Pramāņavārttika and the Pramāņaviniscaya, were composed which can be considered as a valid statement on the tradition irrespective of their mutual deviations and irrespective of their "methodological deficiencies" in terms of modern critical historical scholarship.
We have no right to denigrate these beautiful fruits of Tibetan intellectual creativity as being of merely interpretational character and not philosophically innovative. For do we not do the same? That is, do we not also try to understand Dharmakirti's works, e.g. within and from their context as a document of reflection on man and his actions? We only distinguish our work from that of the Tibetan exegetes by our concern for the respective context and by our sincere dedication to the exclusion of any creative philosophical activity of our own with regard to the problems confronted. For the Tibetan scholar the interpretation of Dharmakirti and his own epistemological creativity is only one and the same because a living Buddhist tradition provides him with a framework within which he can securely rely on an integral spiritual identity in all his intellectual enterprises.
But if we follow our own obsessive presuppositions that creativity must reach new shores, has to result in new original creations, where and how can we find the original Tibetan mind at work?
No doubt in all those areas where the stream of Indian heritage carried tributaries of heterogeneous and contradictory traditions, and where the need for the establishment of a coherent character resulted in differentiations on all interpretational levels, thereby creating new and comprehensive views of Buddhist theory and praxis, as e.g. in Tibetan Madhyamaka or Tantricism.
However, this mind can also be found where "error, misinterpretation" has contributed to the progress towards a finally accepted interpretation of an Indian complex. The epistemological theories of the famous Phywa pa Chos kyi Seng ge (1109-1169) are a good example. Famed for
1) Leonard W. J. van der Kuijp. Contribution to the Development of Tibetan Buddhist Epistemology from the eleventh to the
thirteenth century. Wiesbaden 1983. 2) A periodization has been proposed recently in van der Kuijp 1989: 5ff.

Page #2

258
Ernst STEINKELLNER
his introduction of the 'thal phyir-praxis in dialectical discussion, hardly any of this innovative man's epistemological ideas escaped the critique of Sa pan and his followers. But his contributions are certainly original and his influence was lasting and was of remarkable import to the theories of the later dGe lugs pa scholars.")
Finally there are many topics in all the theoretical traditions that have not been dealt with in all their details and consequences within the Indian Buddhist documents.
Such topics provide another broad field for the activity of the Tibetan scholars who tried to express the implications and to bring the often rudimentary beginnings in their sources to systematical completion.5)
An example of the latter is the theme of the present paper. Already van der Kuijp indicated in a note that we here face an area that was hardly developed by the Indian scholars but must have been of considerable interest to the Tibetans during the first period of their study of epistemology on the basis of the Indian tradition.
And in fact the problem under discussion in an interesting one, for it is related to no less than the question of how validity of a cognition (tshad ma nyid, prämänya) can be ascertained (nges pa, niścita).
In the Tibetan epistemological literature this theme is usually dealt with under the topic of "the valid cognition which ascertains the validity (of a valid cognition)" (mtshan nyid nges byed kyi tshad ma)." This topic forms a corollary to the definition of valid cognition (tshad ma'i mtshan nyid) and can be found dealt with - more or less extensively - in all relevant treatises.)
The question of whether the validity (pramanya) of a valid cognition (pramana) is certain on account of the cognition itself (svatah), is intrinsic to the cognition, or is in need of another
3) Cf. L. W. J. van der Kuijp. Phya-pa chos-kyi seng-ge's Impact on Tibetan Epistemological Theory. JIPh 5, 1978,
355-369[=Tibetan Studies, ed. M. Brauen and Per Kvaerne, Zürich 1978, 163-177] and van der Kuijp 1983: 59fT. 4) Even if we cannot yet say anything definite regarding his often indicated dependency on Sankaranandana (cf. e.g, van
der Kuijp 1983, note 340). It should also be considered in this connection that he evidently did not know Sanskrit (cf. van der Kuijp 1983: 61). On his impact on Tibetan dialectical practice cf. Jackson 1987: 129 and van der Kuijp 1989:
13ff. 5) Cf. the system of invalid cognitive acts (tshad min) in van der Kuijp 1978: 358ff. 6) Cf. van der Kuijp 1983: 60 and note 207. 7) E.g. Rang 'grel 100a6: mtshan gzhi la mishan nyid nges pa'i tshad ma, or Mun sel 120, 20: mtshan nyid nges byed kyi Ishad ma,
or similar formulations. 8) The following is a survey of major sources and of major paraphrases. It is not comprehensive, but includes examples from the different traditions mainly for the earlier period and adding some from later times.
rNgog Lotsāba Blo ldan shes rab: paraphrases in Pham byed II 366,6–367,5; 350,2f.; Ngog lugs: paraphrases in Rang 'grel 101b2-4; Rol mtsho 618,6-619.1.
Phywa pa Chos kyi seng ge: paraphrases in Pham byed II 367,5-370,3; Rol mtsho 617,2-7; 623,5-6. gTsang nag pa br Tson grus seng ge: bs Dus pa 26a2-27b5. Sa skya pandita: Rigs gter VIII 546-58a; Rang 'grel 100a6-102al. 'U yug pa Rigs pa'i seng ge: Rigs mdzod (A), 197,1-3. Bo dong pan chen Phyogs las rnam rgyal: Rigs snang (Vol. 7) 479,3-488,6. Go ram pa bSod nams seng ge: rNam bshad 127al-134a4; gSal byed 5605-60a5; "Od zer 105a6-b6.
gSer mdog pan chen Shakya mchog Idan: Pham byed II 341,4–375,1; Rol mtsho 615,6-619,2; Kun bzang 300,4-301,3. Mus rab 'byams pa Thugs rje dpal bzang: Rab gsal 166,5-169,6. Glo bo mkhan chen bSod nams Ihun grub: Nyi ma 278,3–280,2. Ngag dbang chos grags: dPal ster 351,5-376,6.
Tsong kha pa: 'Jug sgo 22,10-24.9. mKhas grub rje: Mun sel 120,20-122,2; Rigs pa'i rgya mtsho (Vol. Da) 25b1-26b2.
Gyal tshab rje: Thar lam sal byed 129b2-131a2; dGongs pa rab gsal 12b3-14a6. dGe 'dun grub: Rigs rgyan 118,3-125,5. Pan chen bSod nams grags pa: Gongs gsal 408,1-429,2. "Jam dbyangs bzhad pa'i rdo rje: Blo rigs 339,1-342, 1.
A kya Yongs dzin dByang can dgra ba'i blo gros: Blo rigs 523,1-5 (cf. A Compendium of Ways of Knowing, Dharamsala 1976, 48-52).
Phur Icogs Yongs dzin Do rje chang: Blo rigs 10b1-12b4.
Jam dpal bsam 'phel: Blo rigs 16,6-19,10 (cf. Lati Rinbochay, Mind in Tibetan Buddhism, Valois 1981, 121-128, for a modern commentary).

Page #3

Early Tibetan Ideas on the Ascertainment of Validity (nges byed kyi tshad ma)
259
cognition (paratah), since its ascertainment is extrinsic to it, is a well-known theme of the polemics between the Indian Buddhist epistemologists and the Mimāmsä. It owes its origin to the Mimāmsāka's insistence on the independent authority of the eternal Veda and was promulgated strongly by Kumarila'') who became the main opponent of the Buddhist epistemologists from Dharmakirti onwards.
The difference between these two traditions regarding the derivation of validity is succinctly characterized by Säntarakṣita when he says: "Because some strongly pronounce the restriction, that (a cognition) is (valid) only by itself (svatah), while others hold no (such) restriction, (when they say) that some (cognitions are valid) by themselves, and others because of other (cognitions) (anyatah).")) The first opinion is the Mimāımsaka's, the second the Buddhist's. Sāntarakṣita does not elaborate on the question of which cases of valid cognitions are in fact of the one or the other kind according to the Buddhist epistemological tradition, but Kamalasila provides us with a neat list. He distinguishes seven kinds of valid cognition and says that
(1) the selfawareness-perception (svasamvedanapratyakşam), (2) yogic cognition (yogijñānam), (3) cognition of the fulfillment of purpose (arthakriyajñānam) (4) inference (anumānam), and
(5) trained perception (abhyāsavat pratyakşam) are valid intrinsically (svatah), and that
(6) cognition produced by prescription (codanäjanitam jñanam), and
(7) perception that is not free from causes for error (pratyakşam anapagatabhrāntinimittam) are extrinsically (anyatah) valid.12)
This small list is the only one I know of. A comparable but different list can be gathered frein Dharmottara's commentary on the Pramāņaviniscaya,13) to which I shall refer later. 4) We may further assume from later Tibetan informations15) that Sankaranandana treated this subject too in one of his relevant works, but they are not yet accessible.16) Finally there are some short remarks of relevance by Manorathanandin") and Mokşākaragupta.18)
Not anticipating detailed research on these and further materials -nothing has been done so far - I can only summarize at present a first rough impression on the early Indian development of our theme.
In Dharmakīrti's works only two small remarks are available that can be understood - and in fact were thus understood by some Tibetans19) - as referring to the issue of an ascertainment of validity. Dharmakirti says in PV II 4d-5a: "The cognition of the proper nature (of a cognition as such follows) from itself, (but) that this cognition) is valid is only known) through practice."20)
These words clearly state that the cognition of validity depends on something else, but the later treatment of the problems, e.g. in Santarakṣita's discussions of TS 2956ff., also makes clear that these remarks by Dharmakirti cannot be taken as a comprehensive answer to our question. Dharmakirti simply did not deal with this problem at all. But that it was discussed to some extent
9) Cf. e.g. the long Svatahprāmanyapariksă in Santaraksita's Tattvasangraha (vv. 2810–3122). 10) Cf. Schmithausen 1965: 189ff. and D'Sa 1980: 180ff. 11) TS 2944: yenaikaih svata eveti proccair niyama uyate / kincit svato 'nyatah kiñcit parais caniyamo matah // 12) TSP 938, 19–29. 13) Cf, Steinkellner/Krasser 1989: 13,3ff. 14) Cf. below p. 266 ff. 15) Cf. below p. 262 f. 16) Sankaranandana wrote three Prämánya-treatises, a Suksma-, a Madhya- and a Brhatprämänyakärikä (cf. Much 1988:
16; for a description of the texts from glass negatives at the Bihar Research Institute cf. Bühnemann 1980). The extent texts have only the kärikäs, but it is possible that Sankaranandana also added a commentary in prose (cf. below p. 262). Although I have a xerox-copy of these texts at hand, they are virtually illegible and the editing of them by an
expert Pandit would be an urgent task. 17) Cf. PVV 2,22–3,8. 18) Cf. TBh' 3,5-9 (3,11-15); cf. Kajiyama 1966: 27 and note 19. 19) E.g. by mKhas grub rje, cf. Mun sel 120,21-121,4. 20) PV II 40-5c: svarūpasya svato gatih // pramanyam vyavahärena. For a detailed explanation cf. van Bijlert 1989: 147f.

Page #4

260
Erst STEINKELLNER
in his school is evident from Kamalasila's list and from Dharmottara's treatment. Nevertheless we can say that in the Indian Buddhist epistemological school the problem was - with the exception of Dharmottara - of only marginal interest.
This impression is vividly corroborated by the words of gSer mdog pan chen Shākya mchog ldan (1428-1507) who in his larger commentary on the Tshad ma rigs gter says with respect to the sources of this theorem (lung khungs bsam pa): "With respect to the twofold divisions of intrinsic ascertainment and extrinsic ascertainment in the case of a valid cognition the sources (consisting) of the Pramāṇasamuccaya and the Seven Treatises are not clear (gsal po med). The commentators, too, did not clearly distinguish that this was the way these two have been taught by the sources and (this was) the definition of each (definiendum), and this was) the exemplificate. Particularly the opinion that intrinsic ascertainment and extrinsic ascertainment are respectively explained by the two (!) feet 'PV II 4d' and 'PV II 5c"]?!) is not held by anybody competent (tshad Idan) and is also in general impossible."22)
Shakya mchog Idan's words and, even more so, his elaborate discussions23) are representative for a more final stage of the Tibetan development of our theme. We shall not deal with it as such, although mainly his and other, even later texts will be exploited as sources of information on ideas of the early period.
There are two main sources presently available for the treatment of our theme in this earliest period of Tibetan interest in it: the Pramāņaviniscaya-commentary21) of g Tsang nag pa br Tson 'grus seng ge (t after 1105),25) a pupil of Phywa pa Chos kyi seng ge, and the Tshad ma rigs pa'i gter, together with an autocommentary by the Sa skya Pandita Kun dga' rgyal mtshan (1182-1251).26)
For two reasons I shall not incorporate the treatment of g Tsang nag pa br Tson 'grus seng ge into the first step of this study: g Tsang nag pa's terminology naturally shows the strong impact of Phywa pa Chos kyi seng ge's systematical categories and terms. And those are still hardly known except for the general information from Shakya mchog Idan which was interpreted by van der Kuijp.27) I assume that van der Kuijp will present a study of Phywa pa's theories on the basis of Shakya mchog Idan's summary and the available fragments and paraphrases in the future. For the time being I have to refrain from offering the results of a study of his pupil's statements except for an occasional reference to them, where they supplement other information. The second and more decisive reason is that Sa skya pandita limits his critique of earlier views in connection with the present theme to opinions propounded by rNgog Lotsāba and his school, and does not seem to - deal with Phywa pa Chos kyi seng ge and his pupils at all.
21) The second foot has been lost in the edition (?). 22) Pham byed II 342,1-3: Ishad ma la rang nges dang gzhan nges gnyis su 'byed pa 'di la / mDo dang De bdun gyi khungs gsal po
med cing/'grel pa byed pa dag gis kyang de gnyis gzhung gis ji ltar bstan pa'i shul dang/ so so'i mtshan nyid dang/mtshan gzhi 'di yin zhes bsal bar phye ba med doll khyad par du / rang las rang gingo bo rtogs/zhes dang/tha snyad las ni tshad ma nyid /> zhes Ishig rkang gnyis kyis rang nges dang gzhan nges so sor bshad par 'dod pa ni / Ishad Idan sus kyang bshad pa min zhing/spyir yang
rung ba min no // 23) Pham byed 11 341,4-375,1 ("mtshan gzhi'i steng du mtshan nyid dang mishon byu nges byed kyi Ishad ma bshad pu") which
forms the last part of his study of the definition of a valid cognition (ibid. 281,5ff.) that van der Kuijp rightly called
"extremely thought provoking" (van der Kuijp 1983: note 340). 24) Tshad ma rnam par nges pa'i çika, Legs bshad bsdus pa. A manuscript is in the possession of the Otani University
Library (cf. Otani Tibetan Catalogue No. 13971, Tome 968). It was indicated to me years ago by Gene Smith whom I would like to take the opportunity of thanking here. This manuscript has now been published in the new series of the Ouani Library (Otani University Tibetan Works Series 2, Kyoto 1989) with a foreword by van der Kuijp (cf.
particularly van der Kuijp 1989: 22-26 for analytical remarks). Cf. also van der Kuijp 1983: note 267 25) On g Tsang nag pa cf. van der Kuijp 1983: 69 and 96 (for a collection of fragments), van der Kuijp 1989: 2-5, also
Jackson 1987: 130. Except for Shakya mchog Idan "gTsang nag pa's work seems to have been little studied by the Tibetan authors (van der Kuijp, letter of July 23, 1988). Van der Kuijp changed the date of his death from 1171 to
after 1185 (van der Kujip 1989: 2). 26) On Sa skya Pandita's life, studies and works cf. now the exhaustive introduction in Jackson 1987: 15-189 (for the Rigs
gter cf. 44-46 and 131ff.). 27) Cf. van der Kuijp 1983: 66ff, and 71-84. g Tsang nag pa treats the topic mtshan gahi la mtshan wyid nges byed bsams pa in
26a2-27b6.

Page #5

Early Tibetan Ideas on the Ascertainment of Validity (nges byed kyi tshad ma)
261
With the exception of g Tsang nag pa's treatment,28) fragments and/or paraphrases from the works of Tibetan epistemologists preceding Sa pan's work can be gathered mainly from Sa pan's work itself and the many commentaries written on it, 29) above all from the large commentary of Shakya mchog Idan already referred to.
In the following I shall therefore present only the essentials of Sa skya Pandita's explanations, first of his own position and then of the earlier statements criticized by him. For these I shall supplement the information from other sources, but only as far as it facilitates an understanding of Sa pan's arguments.
Finally I shall look at the three kinds of valid cognitions held to be extrinsically ascertained (gzhan las nges) by Sa pan. Here the alleged dependency of Sa pan on Dharmottara can be examined more closely.
* * *
Our theme is dealt with by Sa pan in the third section of the third part of Chapter VIII.50) This section has the title "the valid cognition which ascertains the definition in case of the exemplificate" (mtshan gzhi la mtshan nyid nges pa'i tshad ma)).
The structure of this section is as follows: III. mtshan gzhi la mtshan nyid nges pa'i tshad ma
100a6-102al 1. dri ba dgod pa (putting the question)
100b1-2 2. lan (answer)
100b2-102al a. gzhan gyi lan dgag (refuting the answer of others)
100b2-5 b. rang gi lan (own answer)
100b5-101b2 c. de la rtsod pa spang ba (rejecting contentions against this [answer]) 10152-102al
The question is put forth in Rigs gter VIII 54bc: "If it be asked), whether an ascertainment of these valid cognitions is known intrinsically (rang) or extrinsically (gzhan)"52) and is explained as raising the problem in the commentary: If its validity was known intrinsically, it would be futile to teach its definition. And if it was known extrinsically, an infinite regression would be unavoidable.33)
Sa pan's own position is expressed in Rigs gter VIII 550-57a.54) Accordingly he distinguishes the following valid cognitions that are either intrinsically or extrinsically ascertained in their validity:35)
28) bs Dus pa 26a 2-27b6 with the title mtshan) gzhi la mtshan nyid nges byed bsams pa. 29) For these cf. David P. Jackson, Commentaries on the Writings of Sa-skya Pandita. A Bibliographical Sketch, in: The
Tibet Journal 8/3, 1983, (3-23) 8-12. Additions: No. (15) 1 has in the meantime been published: Mus chen rab 'byams pa Thurgs rje dpal bzang, Rigs gter 'grel pa rigs lam rab gsal, Dehradun 1985. Also No. (19) by Ngag dbang chos
grags, New Delhi 1988. 30) For an analysis of the Rigs gter cf. Zoltán Horváth, Structure and Content of the chad-ma rigs-pa'i gter, an
Epistemological Treatise of Saskya Pandita, in: Tibetan and Buddhist Studies Commemorating the 200th Anniversary of the
Birth of Alexander Csoma de Körös, ed. by Louis Ligeti, Vol. 1, Budapest 1984, 267-302. 31) Rang 'grel 100a6-102al. 32) Rigs gter VIII 54bc:/ ishad ma de dag nges pa ni /
/ rang gzhan gang las rtogs she na / 33) Rang 'grel 100blf.: nges shes Ishad ma de nyid kyi stobs kyis 'dren nam gzhan gyi stobs kyis 'dren / de nyid kyis 'dren na tshad ma'i
mtshan nyid la mi mthun pa med par 'gyur la gzhan las nges na thug med du 'gyur ro zhes zer ro. Cf. Go ram pa, rNam bshad 127a2f., and Mus rab 'byams pa, Rab gsal 166,6f.
The question and its explanation are given in substantially the same way by gTsang nag pa (bs Dus pa 26a 2-3). It is possible, because of the referential way of presenting it (...zhes zer ro) that Sa paņ is referring to gTsang nag pa here.
But lacking other indications we cannot say so for certain. 34) Rigs gter VIII 550-57a:/ don rig gnyis dang rang rig dang (55)
/ rjes su dpag pa rang las (: la) nges / / dang po pa dang yid ma gtad / / 'khrul rgyu can mnams gzhan las nges / I don byed pa dang goms pa can (56)
Irjes su dpag pas nges par byed / 35) Cf. Rang 'grel 10065-6.

Page #6

262
Ernst STEINKELLNER
I. intrinsically ascertained: 1. don byed par snang ba'i mngon sum ("perception in which the actual] fulfilment of a
purpose appears", *arthakriyānirbhāsam pratyakşam) 2. mngon sum goms pa can ("trained perception", *abhyastapratyakşam) 3. rang rig ("self consciousness", svasamvedanam)
4. rjes dpag ("inference", anumānam) II. extrinsically ascertained:
5. mngon sum dang po pa ("first perception") 6. yid ma gtad pa ("[perception) with distracted mind")
7. 'khrul rgyu can ("[perception) with grounds for error", bhrāntinimitta-) III. ascertaining cognitions for the perceptions of Nos. 5, 6, 7 are the perceptions no. 1, 2 and inference. 56)
In the first (100b2-5) and third part (10152-102al) of his answer Sa paņ indicates the views of different opponents that I would like to identify in the following.
The subsequently refuted answer of others to the above question is condensed in Rigs gter VIII 54c-55a: "Some assume, that a valid cognition) is intrinsically ascertained. Some (others) say, that (it) is extrinsically ascertained."37) And this is supplemented in the commentary by the words: "Some say: 'It is only intrinsically ascertained' and some say: 'It is only 'extrinsically ascertained. 38)
The second of these theorems, that it is extrinsically ascertained (gzhan las) is unanimously attributed to the Bram ze (chen po) by the commentators, 59) i.e. Sankaranandana. Go ram pa presents in rNam bshad 127a5f. a quotation from Sankaranandana that consists of a half verse and an explanation in prose:
"**Although something (don) is proven by a valid cognition, the validity (tshad ma, *prämänya) [of this valid cognition) is not (proven) by this same [valid cognition).' (That means:) Although the respective object of activity (jug yul) is ascertained by the two valid cognitions, an ascertainment of the definitory character (mtshan nyid), (i.e.) reliability (mi slu ba) is not provided through itself. Therefore the general definition and its) contrary concomitance ('gal 'brel) etc. are ascertained by a cognition called 'examination (dpyod pa, *vicāra) which occurred subsequently to the valid cognition. As such the definitory character of a valid cognition is ascertained only by another (valid cognition)."40)
36) Cf. Rang 'grel 101a4-101bl. 37) Rigs gter VIII 540-55a: la la rang las nges par 'dod /
l 'ga' zhig gzhan las nges zhes zer / 38) Rang 'grel 100b3: kha cig rang kho na las nges so zhes zer ba dang/ 'ga' zhig gzhan kho nas nges par byed do zhes zer ba nil... 39) Cf. rNam bshad 127a5f. and gSal byed 57a3; Pham byed II 370,5 (Bram ze chen po Shal ka ril[?]; Rab gsal 167,2; Nyi
ma 278,5; dPal ster 354,5f. 40) Go ram pa, rNam bshad 127a5f.: Bram zes -
/ Ishad ma las don" grub pa yang
/shad ma de nyid las ma yin/ zhes Ishad ma gnyis kyis rang rang gi 'jug yul nges par byed kyang mtshan nyid mi bslu ba la nges par rang stobs kyis 'dren par mi byed pas tshad ma'i rjes su skyes pa'i blodpyod pa zhes bya bas spyi'i mtshan nyid dang/gal 'brella sogs pa nges par byed pas na tshad ma'i mtshan wyid ni gzhan kho na las nges par byed do-zhes bzhed do // "don (cf. Nyi ma 278,6): ston
The half verse quoted is attributed to Sankaranandana's Sambandhapariksa ('Brel pa brtag pa) by Glo bo mkhan chen (cf. Nyi ma 278,5f.), but to be found in neither the Sambandhaparikså nor its commentary, the Anusara (cf. Frauwallner 1934: 263ff. for the text of the verses in the translation of the Anusara).
The above explanation is also referred to in Ngag dbang chos grags' dPal ster 354,58; cf. also Bo dong pan chen's Rigs snang 483,4-6.

Page #7

Early Tibetan Ideas on the Ascertainment of Validity (nges byed kyi tshad ma)
263
At this time we can only assume that this verse comes from one of the three Prāmāṇyaparīkņātexts written by Sankaranandana,4l) but the ascription of the theorem to this Indian scholar need not be doubted.
The ascription of the first theorem, that validity is intrinsically ascertained (rang las nges pa), seems to be more problematic because most of the commentators 2) think that the opponent here is the same that is adduced later in the third part of the answer (Rang'grel 101b2f.), namely Gangs pa she'u.43) However, only the theorem as such is shared by Gangs pa she'u and in the present first part of the answer it is evidently introduced as being propounded by the Mimāmsā.44) Indeed, Glo bo mkhan chen in his 'Grel bshad (278,4f.) rightly identifies this opponent as Kumārila by referring to SV II 47.45)
That Sa paņ here in Rigs gter VIII 54c-55a first refers to the Indian background is more likely for two reasons: firstly the second theorem is unanimously attributed to Sankaranandana. And secondly Sa paņ only starts refuting earlier Tibetan viewpoints, among them also that of Gangs pa she'u, in the third part under the title "refutation of contentions against this answer of our own)."46)
In this third part Sa paņ concerns himself only with objections to his own treatment as coming from the interpretation of rNgog Lotsāba Blo ldan shes rab's school, the so-called rNgog lugs.) The two objections as formulated in Rigs gter VIII 57bc and supported by two quotations 8) in Rang 'grel 101b2f. and 101b3f. are directed against the possibility of an ascertainment of a cognition's validity by means of another cognition, i.e. against the later part of Sa pan's system where he holds that some cognitions are ascertained as valid by others. The two objections differ
41) Cf. Bühnemann 1980: 192. The three Prämānyapariksā-texts are in verse only. It is however quite possible that
Sankaranandana has also composed commentaries in prose on these texts of which nothing so far is known. The prose
part of Go ram pa's quotations might well belong to the fragment from Sankaranandana too. 42) Cf. rNam bshad 127a6-127b2 and gSal byed 57a3f.; Rab gsal 167,2; dPal ster 354,6; 43) Cf. below p. 264. 44) CF. Schmithausen 1965: 193-198. 45) This verse by Kumārila is also reproduced in TS 281 lab and 2812cd, while Kamalasila has only the second part (TSP
904,17f.) although Glo bo mkhan chen refers only to the quotations of "Kālasila". For the ascription of his theorems to
the Mimāmsā cf. also TS 2944ab and TSP 938,18. 46) de la risod pa spang ba, Rang 'grel 101b2f1.
The anachronism in this reference to an earlier scholar as attacking Sa pan's position can be explained by the fact that Sa pah was convinced that his own position was identical with that of Dharmottara. Therefore he felt justified in presenting polemics of earlier Tibetan scholars against Dharmottara as polemics against his own position. But if he deals with Gangs pa she'u only in this last part he cannot be referred to here in the first part, although Kumarila's theorem of intrinsic ascertainment was, no doubt, shared by Gangs pa she'u. This common ground must have thus been the reason for the fact that most commentators took both passages as referring to Gangs pa she'u. On the rNgog Lotsāba and his school cf. Naudou 1968: 171f., 174; van der Kuijp 1983: 29-58; Jackson 1987: 127ff.; van der Kuijp 1989: 11 ff. and 19f. With a view to the great originality of Phywa pa Chos kyi seng ge it is perhaps advisable in this period before Sa pan's activities to speak of a rNgog-school and of a Phywa pa-school irrespective of the fact that Phywa pa as a student of rGya dmar Byang chub grags (van der Kuijp 1983: 60) belongs to the
rNgog-tradition historically. 48) The Tibetan scholars intentionally quote quite literally from Indian sources, and deviations are mostly due to a lapse
of memory. We can therefore easily distinguish real "quotations" which thus transmit textual "fragments" of the quoted original from mere "reference" when they adduce Indian Buddhist sources. But when they refer to Tibetan
s their manner of quoting is much less rigorous. They often present us with only a report of the meaning or general idea of a statement propounded by a Tibetan predecessor or opponent where the literalness of their report seems to be directly related to the authoritativeness associated with the personality quoted. It is therefore particularly difficult to decide in many cases whether, e.g., a certain term occurring belongs to the terminological usage of the text or person quoted or of the author who quotes. Also, in general, the reliability of such quotations has to be carefully judged in each case (cf. van der Kuijp 1989: 22 for a warning in respect of Shakya mchog Idan's "paraphrases" of g Isang nag pa).
With a view to this delicate situation I shall apply for practical reasons and with only tentative claim a qualified terminology in reference to such quotations from Tibetan sources. "Quotation" will be used for a text from another source without defining its quality of literalness between the extremes of presenting a real fragment of the source and of being a general paraphrase. If such a quotation for whatever reasons may be taken as a reliable re-statement I shall use the term "fragment", and if it is clear that the source is only re-stated in a referring way I shall use the term "paraphrase".

Page #8

264
Ernst STEINKELLNER
in that they base their argument on a reference to the object of the cognitions (yul la brtags nas) and on the essential nature of the cognition (ngo bo la brtags nas) respectively.
Sa pan's quotations have been identified by Shākya mchog Idan: both as of (rNgog) Lotsāba, Khyung and others (Rol mtsho 618,6ff.), by Mus rab 'byams pa: both as of Khyung, Gang(s) (pa) she('u) and others (Rab gsal 169,2ff.), and by Ngag dbang chos grags: both as of Gangs pa she'u and others (dPal ster 354,6ff.)
These identifications can only be taken as tracing the quotations back to the direct pupils of rNgog Lotsāba,9) thus representing a "rNgog lugs view-point". The seeming anachronism here can be explained by Sa pan's conviction that his own position is not different to Dharmottara's.50) Since the objections adduced by Sa paņ must be understood as originally having been directed against Dharmottara we have to interpret them accordingly. Due to lack of time and in view of the fact that a manuscript of rNgog Lotsāba's commentary on Dharmottara's Pramāņaviniscayaţikā has now been found in Beijing I shall not discuss rNgog Lotsāba's critique of Dharmottara and his own view in detail.51) This must be reserved for a later date. I shall refer to the information now available from fragments and paraphrases only in order to clarify the objections adduced by Sa pan.
Since it seems clear that Sa pan refers to two opponents, but at the moment the divergent identifications of the commentators cannot be clarified, I shall for the sake of convenience refer to the first objection as that of Khyung (Rin chen grags) and to the second as that of Gangs pa she'u (Blo gros byang chub).
Both objections share the idea that the validity of a valid cognition can only be ascertained intrinsically and their arguments are formulated against the possibility of an extrinsical ascertainment as assumed to be neccessary at least for some cognitions by Dharmottara and Sa pan.
49) On Khyung Rin chen grags and Gangs pa she'u Blo gros byang chub cf. van der Kuijp 1983: 59f. 50) Cf. above note 46. 51) Several relevant quotations from rNgog Lotsāba have been identified by van der Kuijp (van der Kuijp 1983: 58,60 and
note 205), and others can be associated with these. rNgog Lotsāba contributed to not only the translation of the Pramāņaviniscaya and of Dharmottara's commentary, the "Thad Idan, but also wrote a summary (bs Dus don) and a subcommentary (rNam bshad) on the latter (Kuijp 1983: 34).
These works are mentioned in Bu ston's catalogue with the titles Tshad ma rnam nges tik dang bcas pa'i don bsdus (Nishioka 3091) and rNam bshad chen po (Nishioka 3092). Van der Kuijp tentatively identifies the "tik chen" mentioned by Shakya mchog Idan (dGa' byed 12,6) with the latter and "dka' 'grel" (dGa' byed 29,3) with the former. Jackson (1987: 128) rightly assums that the subcommentary "was still available to Shakya-mchog-Idan at the end of the 15th century".
To our great surprise a work with the title Tshad ma rnam nges kyi dka' gnas rnam bshad of rNgog Blo Idan shes rab occurs in a handlist ("Bod yig gi tshad ma'i skor kyi dpe tho") of pramāna-works kept at the Library of the Cultural Palace of National Minorities (Minzu wenhua Gong Tushuguan) in Beijing which was compiled by Prof. Huang Min-xin. According to this list this is a manuscript with three chapters: mNgon sum gyi le'u (46ff.), Rang gi don gyi rjes su dpag pa'i le'u (45ff.) and gZhan gyi don gyi rjes su dpag pa'i le'u (33ff.) The title of this manuscript would allow the assumption that this is in fact the "rNam bshad", but that this rNam
difficult points". We could thus tentatively identify this "dka' gnas rnam bshad" with the two titles "tik chen" and "dka' 'grel" given by Shakya mchog Idan and take the attribute chen to be meaningful only in relation to the bsDus don work, but not in an absolute sense.
Two longer quotations which I consider as fairly reliable because of their specific terminology can give us a fairly broad knowledge of rNgog Lotsāba's ideas on the ascertainment of validity and of the kind of critique he had on Dharmottara's views.
The first quotation (quoted by Shakya mchog Idan in Pham byed II 366,6–367,5 and referred to in Pham byed II 374,3 and by Go ram pa in rNam bshad 127b3f.) contains a definition of intrinsical and extrinsical ascertainment, a list of cases of valid cognitions as examples for the two kinds of ascertainment, a short critique of Dharmottara's division of three kinds of "first perception" and a summary of his own views. The second quotation in Pham byed II 350,28.) states which cognitions can be considered as ascertaining cognitions. Thus these two fragments cover the same ground as Sa pan's verses Rigs gter VIII 550-56c and 560-57a respectively. Both quotations are said to be taken from the rNam par nges pa'i tīkā of rNgog Lotsāba.

Page #9

Early Tibetan Ideas on the Ascertainment of Validity (nges byed kyi tshad ma)
265
Khyung bases his argument on a consideration of the object (yul) of the cognition.52) And the second argument, possibly Gangs pa she'u's, is based on a consideration of the nature (ngo bo) of cognition.53)
In his answer to these two objections Sa paņ makes the point that they both are irrelevant because they confound the functions (byed pa) of perception (mngon sum) and conception (rtog pa)") and thereby do not take into consideration that only perceptual cognitions are possible items where the validity is extrinsically determined.
In his clarification Sa pan first recalls the definition of perception: "A perception is a valid cognition only because of the fact that it is) non-conceptional and non-erroneous"55), and then offers a statement that can be taken as his formula of defining both intrinsical and extrinsical ascertainment.56) He says: "If a certain cognition (nges shes) can be produced immediately by force of the perception only, (this perception) is ascertained [as valid] intrinsically; and (a cognition) for which a certain cognition cannot be produced immediately by force of the perception and (therefore) has to be ascertained [as valid) later by another valid cognition is ascertained extrinsically."57)
To sum up: if Sa pan correctly interprets both objections as being directed against the possibility that one cognition can be ascertained as valid by another cognition, it can be assumed that both Khyung and Gangs pa she'u deviate from rNgog Lotsāba's original opinion as attested by a paraphrase of Skākya mchog Idan.58) Because there rNgog Lotsäba is said to have clearly distinguished between the ascertainment of the nature of a valid cognition which is only intrinsic 9) and the ascertainment of the valid cognition to be defined which is both intrinsic and
52) It is presented by Sa pan in the following form: Rang 'grel 101b2f.: nges bya'i tshad ma dang nges byed kyi tshad ma gnyis yul
gcig na nges byed kyi Ishad ma gryis pa bcad pa'i yul can du 'gyur la yul mi gcig na yul gzhan dngos por grub pas / gzhan ji ltar dngos por grub - ces zer ba dang ...("When both, the valid cognition to be ascertained and the valid cognition ascertaining, have one and the same object, the second valid cognition which ascertains would refer to an object (already) determined (bcad pa). And when they have different objects, another object is established as real [for the later, ascertaining cognition). How then is the first (gzhan) [i.e. the object of the cognition to be ascertained] established as real?"]
This refutation is already attested in gTsang nag pa br Tson 'grus seng ge's Pramanaviniscaya commentary (bsDus pa 26a7f.: tshad ma mams ni rang nyid kyis grub pa yin te / blo mnams yul mi gcig pas gzhan gyis nges pa mi rung ba'i phyir ro zhes zer ba ni /...).
Van der Kuijp in his recent publication changed his earlier calculation of 1171 for the year of g Tsang nag pa's death (Kuijp 1983: 83) to after ca. 1185 (cf. Kuijp 1989: 2).
Several other paraphrases of his objection, mostly in connection with the second objection are extant, e.g. in Shakya
mchog Idan's Rol mtsho 618,6f., Go ram pa's rNam bshad 133a6f., Mus rab 'byams pa's Rab gsal 169,2f. 53) Rang'grel 101b3f.: yang kha cig-gzhan las nges pa thams cad kyis kyang de'i tshe sgro 'dogs ma chod na Ishad ma nyid ma yin pas
gzhan nges su 'ang mi thad la / sgro 'dogs chod na rang las nges su gyur ro // des na guhan las nges kyi tshad ma gtan mi srid do - zhes zer ro // ["(A cognition) is certainly not a valid cognition, if the wrong superimpositions (sgro 'dogs) (on its object)" are not excluded even by all extrinsical ascertainment at its own time. Therefore no extrinsical ascertainment is possible at all. And when the wrong superimpositions are excluded, (the cognition) would be intrinsically ascertained (as valid). Thus a valid cognition extrinsically ascertained (as valid) is absolutely impossible."]
* CF. Rol mtsho 619,1: rang gi 'jug la sgro 'dogs, rNam bshad 133b3: rang yul la sgro 'dogs.
For other paraphrases ct. Rol mtsho 618,71.. rNam bshad 133b3, Rab gsal 169,31. 54) Rigs gter VIII 570-58a: de dag gis ni mngon sum dang/ riog pa'i byed pa 'dres par zad / 55) Rang'grel 101b4f.: mngon sum rtog bral ma 'khrul ba tsam gyis ishad mar 'gro la /... 56) Such definitions of the two ascertainments are later usually given in the introductory sections of the theme (cf. e.g. Go
ram pa, rNam bshad 130a4ff. or gSal byed 57b6ff.) and already in the fragment from rNgog Lotsāba referred to above (cf. note 51) they are found at the beginning (cf. Pham byed II 366,6f.). But Sa pan states them only here
parenthelically in a refutation. 57) Rang 'grel 101b5f.: mngon sum nyid kyi rang stobs kyis de ma thag tu nges shes 'dren nus na rang las nges yin cing/mngon sum
gyi rang stobs kyis de ma thag tu nges shes 'dren ma nus pa / phyis tshad ma gzhan gyis nges dgos la / gzhan las nges yin pa'i... 58) Pham byed II 366,6–367,5. Cf. above note 51. 59) Pham byed II 367,3–4: Ishad ma yin na rang nyid gang la tshad ma yin pa'i don de la rang las nges yin dgos te / tshad ma yin na
rang gi gzhal bya bden par nges dgos la de nges pa na rang nyid mi bslu bar shugs la nges nus pa'i phyir /["If (a cognition) is a valid cognition the ascertainment with regard to that meaning (don) of a valid cognition which is (its) nature must be intrinsical. For, if (a cognition) is a valid cognition, its respective cognitional object must be certain as real, and when this is certain the (cognition's) reliable nature can be ascertained by implication.")

Page #10

266
Ernst STEINKELLNER
extrinsic.60) And Khyung and Gangs pa she'u would deviate from rNgog insofar as they argue against a possible extrinsic ascertainment seemingly also in case of the latter.
*
*
*
We shall now investigate the three kinds of perception said by Sa pan to be extrinsically ascertained (gzhan las nges), because it is with regard to this theme that we can best judge to which extent Sa pan's view can be said to conform with the view of Dharmottara as is assumed e.g. by Shakya mchog Idan.61)
Sa pan does not himself state that he depends on Dharmottara, 62) but the differentiation and examples he gives point clearly in this direction and Shākya mchog Idan's philological efforts to determine the respective sources are well grounded.65)
Sa paņ says: "The first perception (dang po pa) and that with a distracted mind (vid ma gtad pa) and that (perception) which is affected by grounds for error ('khrul ba'i rgyu mtshan gyis zin pa) are extrinsically ascertained (as valid)."64)
These terms and their mutual systematic relation need to be clarified first in their usage by Sa paņ and then we shall try to relate Sa pan's differentiations to their alleged source, Dharmottara's Pramāņaviniscayaţikā. In this Shakya mchog Idan's exegetical ideas will be of considerable help because he seems to have been the first who tried to point out the relevant passages in Dharmottara for this whole system with great care and precision.
1) dang po pa, "the first", is a perception of something for the first time. The term is not attested so far in Indian sources but was already used by rNgog Lotsāba in his description of Dharmottara's classifications.65)
Sa pan does not offer a definition of this first perception, but gives two clear examples (dper brjod na): a perception of this kind is the case when the appearance of the object (snang ba) is intrinsically ascertained, but the share of reality (bden pa'i cha) in this appearance is extrinsically ascertained. E.g., when the form of something shining red from afar (rgyang ring po nas dmar 'bar ba'i rnam pa) is intrinsically ascertained, the doubt (the tshom za ba) whether this is as fire real or merely a wrong superimposition (sgro btags) like an illusion or a magical creation is settled extrinsically. 67)
60) Pham byed II 367,4–5: de lta na yang tshad ma yin pa'i tha snyad rang las nges pa dang/gzhan las nges pa gnyis yin no//{"The
term (tha snyad) of a valid cognition is, nevertheless, both, intrinsically and extrinsically ascertained.")
Cf. also Shakya mchog Idan's short summary of this position in Pham byed II 374,3: mdor na/ mtshoni bya tshad ma'i tha snyad la ltos nas rang nges gzhan nges su 'jog cing/mtshan nyid mi bslu ba la ltos le rang nges kyis khyab par bzhed pa Lo chen gyi.
lugs /. 61) Cf. Pham byed II 354,7; 374,2f. and dGa' byed 41,3 (translated in van der Kuijp 1983: 115). 62) Already noted in van der Kuijp 1983: note 346. 63) It will be also necessary to pursue the indications of Go ram pa, who adduces not only Dharmottara, but also
Devendrabuddhi and Sakyamati in his attempt to pin down the Indian sources of these differentiations (cf. Nam bshad 132b3-133a6). But this will be done on another occasion, when the Indian development can be more fully
elaborated than is possible at this time. 64) Rang 'grel 100b6: mngon sum dang po pa dang yid ma gtad pa dang/ 'khrul ba'i rgyu mtshan gyis zin pa gsum po gzhan las nges
par byed do Il; cf. Rigs gter VIII 56bc (quoted above, note 34). 65) Cf. Pham byed II 366,1f.; Ngog uses the term, however, slightly differently to Sa pan in that it serves as a general
term for the three perceptions doubtful with regard to don yod pa sam, or to spyi yod par nges kyang khyad par or snang ba
nyid, while with Sa pan it would refer only to the first two cases. 66) For a definition from the Sa skya tradition cf. Go ram pa's rNam bshad 131a5f.: dang po pa ni dngos po 'ga' zhig gi rigs
'dra sngar ma mthong ba gsar du mthong ba ste / ba men dang por mthong ba'i mngon sum lta bu'ol/["A new perception of a kind of thing that has not been perceived before, e.g. a perception of somebody) seeing a Gayal for the first time."] Cf. dPal
ster 367,1. 67) Rang'grel 100b6-10lal: dper brjod na snang ba rang las nges bden pa'i cha guhan las nges pa rgyang ring po nas dmar 'bar ba'i
mam pa rang las nges kyang mer bden nam 'on te sgyu ma'am" sprul ba la sogs pa sgro btags yin zhes the tshom za ba gzhan las nges par byed pa'am /
* ma'am SB: mar ram D ba D: bar SB

Page #11

Early Tibetan Ideas on the Ascertainment of Validity (nges byed kyi tshad ma)
267
And the second case of a first perception of this kind is when the general aspect (spyi) of something with branches appearing afar is intrinsically ascertained, but the specific aspect (bye brag), e.g. that it is a Simšapā-tree is ascertained extrinsically. 68)
Both cases are meant as cases of "first perceptions" (dang po pa), for Sa pan connects the examples with the disjunctive particle "or" ('am).
The original source of these two cases of "first perception" has been found by Shakya mchog ldan in Dharmottara's Pramāņaviniscayațīkā69) as a part of Dharmottara's exposition of extrinsically ascertained cognitions.)
Dharmottara explains the first case as a cognition with an object determined as existent in its appearance but doubtful with regard to the necessarily connected thing: is that fire real or not?71) And the second case is more elaborately explained by Dharmottara, because he also considers the ascertainment of general aspects other than the one already intrinsically ascertained.72)
Although it is clear therefore that Sa paņ takes these cases from Dharmottara, the systematic aspect of his subsuming them under the category "first perception" is not extant in Dharmottara's exposition. It is however preformed by rNgog Lotsāba's arrangement of three cases of "first perception", 75) although rNgog's third case is treated differently by Sa pan.
This third case of "first perception" in rNgog Lotsāba's description of Dharmottara's explanation is the case of a perception which is doubtful with regard to the appearance itself
(snang ba nyid), i.e. when it is not clear that something has actually appeared in one's perception. . Sa paņ evidently refers only to the first two cases as cases of "first perception" and explains the
third case in another way, namely by splitting it up into two kinds, which he calls yid ma gtad pa and 'khrul rgyu can respectively.74)
Although already rNgog Lotsāba distinguishes three kinds of "first perception" according to Dharmottara75) and refutes the third kind, 76) and others like Phywa pa Chos kyi seng ge) and Shakya mchog Idan78) also clearly attest to such a division for Dharmottara, I am unable to find a threefold division of that kind in either Dharmottara's Pramāņaviniscayaţikā or his other works, that is particularly the two Prämāṇyapariksä-essays.
Dharmottara seems rather to be strict about the fact that a doubt (samsaya) only relates to the later determining function of a perception) and that the first function of perception, determination of the appearance of an object, cannot be impeded by a doubt at all.80) He even offers a quotation from Dharmakirti to this effect.SI)
68) Rang'grel 101alf.: rgyang ring po na yul daba dang ldan pa snang ba na/ spyi shing Isam laltos nas sgro 'dogs chos nas rang las
nges shing bye brag sha pa la rtog bral ma 'khrul ba yin pas mngon sum Ishad ma yin yang sha pa'i Idog pa la sgro 'dogs ma chod pas gzhan las nges so II
a 'dab SB: 'bad pa D isam la SB: Esal D nas SB: pas D 69) Cf. Pham byed II 349,6f. 70) PVint(a) 14,16-18,5. 71) PVin T(a) 16,2-4: 'ga' zhig yod par zhen par byas pa'i yul can ci me 'di bden nam 'on te ma yin zhes med na mi 'byung bar the Ishom
za ba 'di la yang / 72) Cf. PVin T(a) 16,5-18,5, where Sa pan's second case corresponds only to PVinta) 16,5-17.14. 73) Cf. above note 65. 74) Cf. Pham byed II 349,7f. and 353,2ff. 75) Cf. above note 65. 76) Cf. Pham byed II 367,2f. 77) Cf. his polemics against Dharmottara in the fragment of Pham byed II 367,5-370,3 (368,2 and 368,4f.). 78) In his Rol mtsho 616,6f. he even quotes a passage from Dharmottara: slob dpon Chos mchog ni/gzhan las nges kyi Ishad ma
gsum ste / snang ba vang las nges shing bden pa gzhan las nges pa dang/ spyi rang las nges shing khyad par gzhan las nges pa dang/ snang ba nyid kyang gzhan las nges dgos pa dang gsum moll zhes gsung ba...
This quotation may be only a summarizing paraphrase, but I can find neither a text of this kind in Dharmottara's works nor a similar statement. 79) Cf. PVin T(a) 18,8ff. 80) Cf. PVin?(a) 18,14ff. 81) Cf. PVinȚ(a) 18.12ff.; this quotation, said to be from the Pramāņavārtika, has not been identified so far, but a similar
passage is to be found in PVSV 31,21-23 (cf. Steinkellner/Krasser 1989: 87, note 67).

Page #12

268
Ernst STEINKELLNER
Thus it is highly questionable whether this third kind of "first perception" where the appearance itself (snang ba nyid) is doubtful has been proposed by Dharmottara at all.
We have to wait until rNgog Lotsaba's commentary is available in order to see how he extracts such a threefold division from Dharmottara's work. For the time being we must be content with the assumption that Sa pan follows rNgog Lotsaba in this respect and presents the following two categories of extrinsically ascertained perceptions with the purpose not only of defending Dharmottara but also of giving a complete and reasoned presentation of his alleged opinion.
2) yid ma glad pa, "(perception) with distracted mind", is a case of perception where the appearance itself (snang ba rang nyid) is ascertained extrinsically.
Again Sa pan does not give a definition82) but only an example: "If one doubts when somewhere fire appears 'Did I have a cognition of fire appearing or not?', the sense-cognition which cannot be ascertained by the [following] mental cognition (vid shes) is extrinsically ascertained, although it is nonconceptual (and) nonerroneous (and) therefore a valid cognition."83)
In accordance to what I have said above with regard to the difficulty of tracing the idea of a doubtful appearance (snang ba nyid) in Dharmottara it is not astonishing that this perception under distraction can also not be found in Dharmottara's work. On the contrary, a small remark in Dharmottara's second Pramanyaparīkṣā propounds the very opposite of Sa pan's assumption. For Dharmottara says in an answer that a doubt has the form "is this thing which appears in the perception real or not?" but not the form "does it appear in the perception or not?"84)
3) 'khrul rgyu can, "(perception) with grounds for error", or 'khrul ba'i rgyu mtshan gyis zin pa,85) "(perception) affected by grounds for error."
In this case Sa pan gives neither a definition nor an example. But this perception is sufficiently clear. It is the case of an erroneous perception. Go ram pa's definition may serve as an explanation: "Although (this congnition) has really arisen on the basis of the three conditions, 86) an ascertainment cannot be brought about by its own force, because some external (or) internal ground for error is impeding."87)
In other words Sa pan distinguishes here erroneous sense-perceptions as a second kind of perception where the appearance is doubtful. There is however a clear statement by Dharmottara to the fact that an erroneous sense-perception - and he gives the example of a perception of sunrays as water is not a valid cognition.88)
82) I cannot even find a definition later, but for other examples cf. Go ram pa, rNam bshad 131a6: rtog pa nang du bsdus pa'i gnas skabs kyi gzugs 'dzin dbang mngon lta bu'o // ["E.g. a sense perception which grasps a form, when the mind (rtog=yid) is withdrawn."] Cf. also 132a2 and Pham byed II 353,3ff.; dPal ster 367.1.
83) Rang 'grel 101a3f.: yang snang ba rang nyid kyang gzhan las nges pa dper na 'ga' zhig tu me snang ba na / kho bo la me snang ba'i. blo geig byung ngam ma byung zhes the tshom za ba na dbang shes rtog bral ma 'khrul pas tshad ma yin yang yid shes kyis niges ma" nus pa gzhan las nges par byed pa'o II
a ma om. D.
84) Cf. PPar II 242a6f.: gal te bdag cag mngon sum la the tshom skyes so zhe na bden te / mngon sum la snang ba'i don 'di yang dag pa yin nam 'on te ma yin snyam du the tshom za'o // 'di snang ngam mi snang snyam du ni ma yin te 85) Rang 'grel 100b6.
86) rkyen gsum, i.e. alambana-, adhipati-, samanantarapratyaya.
87) rNam bshad 131a6f.: rkyen gsum las dngos su skyes kyang phyi nang gi 'khrul rgyu 'ga' zhig gis gegs byed pas nges pa rang stobs kyis 'dren mi nus pa ste ; cf. also Pham byed 353,5ff. for further subdivisions and examples dPal ster 367,1ff. 88) Cf. PVinȚ(a) 15,12-16. This evident deviation from Dharmottara in Sa pan's interpretation of cognitions where the appearance itself needs extrinsical ascertainment is corroborated by his polemics against an inclusion of such cognitions among the list of invalid cognitions (tshad min). For in the second chapter of his Rigs gter he refutes the theory of Phywa pa Chos kyi seng ge (cf. the paraphrase in rNam bshad 35b1ff.) who lists a cognition uncertain with regard to its appearance (snang la ma nges pa) as a distinct kind of cognition (cf. Rang 'grel 39a6ff.) and distinguishes three subdivisions: yul rnam pa mi gsal ba, yid nye bar ma gtad pa, 'khrul pa'i rgyu mtshan gyis zin pa.
Sa pan's major argument is: "If (a cognition) uncertain with regard to (its) appearance is (considered to be) an invalid cognition, every perception would be an invalid cognition, because certainty in perception is negated (by us)."
gal te snang la ma nges pa II tshad ma min na mngon sum kun // tshad ma min 'gyur mngon sum la // nges pa nyid ni bkag phyir ro/ Rigs gter II 9c-10b). In other words, Sa pan does not follow Phywa pa in classifying the two cognitions yid ma gtad pa and 'khrul rgyu can as invalid cognitions (cf. rNam bshad 38b1f.).

Page #13

269
Early Tibetan Ideas on the Ascertainment of Validity (nges byed kyi tshad ma)
It seems that Go ram pa tries to cover this curious discrepancy. For after first quoting Devendrabuddhi and Śākyamati he quotes Dharmottara as a source first for the intrinsically ascertained cognitions (PVinŢ(a) 13,3f. and 13,5-7 in rNam bshad 133a3f.) and then for the extrinsically ascertained ones he paraphrases the beginning of PVinȚ(a) 14,16f.: "Some (cognitions) are only extrinsically (ascertained): (perception) with grounds for error and...".89) He then states that Rigs gter VIII 55d-57a, the verses with the list of cognitions, are in accordance with these passages. The second reference is, however, quite problematic, because in this form it cannot be traced to the context in the PVinȚ referred to.
Shakya mchog Idan is much more cautious in this connection. He refers to the same passage as Go ram pa, literally quoting the beginning of Dharmottaras text PVinȚ(a) 14,16f., continues to say that Dharmottara has differentiated the cases of reality (bden pa) and the specific (khyad par) as extrinsically ascertained, and then states that Dharmottara certainly did not mention the appearance itself (snang ba nyid) as extrinsically ascertained in this context (skabs der ma gsungs mod). After that he asks for a careful investigation of whether Dharmottara clearly has such a distinction in mind. And he ends this remark with the note that the two kinds of yid ma gtad pa and 'khrul rgyu can are in this place assumed to be valid cognitions in Sa pan's Rigs gter.90)
Again, when Shakya mchog Idan later explains the example of a perception of sunrays as water, the second kind of his own subdivision of a 'khrul rgyu can-perception of this kind, he notes that this is not in accordance with Dharmottara and gives an identifiable quotation from the 'Thad Idan: "Since the sunrays are not ascertained in (their) nature, although the individual entity (rang gi mtshan nyid) of the sunrays has been grasped, (this cognition) is not a perception of the sunrays. In this way other cases too are to be understood."92)
Supported by Shakya mchog Idan's careful exegetical observations we may therefore safely conclude that Sa pan himself conceived the idea to arrange these last two categories in this place among the valid cognitions in order to give some substantiation to the third kind of extrinsically ascertained perception, which must have been assumed to be held by Dharmottara by Sa pan's predecessors as well as by himself.
The terms and their definitions, however, are not his own invention but are taken from Phywa pa Chos kyi seng ge's list of kinds of cognitions as known from paraphrases by Go ram pa and Shakya mchog Idan.93)
A discussion of the earlier history of these terms in Phywa pa Chos kyi seng ge would exceed the limit of this paper, but their inclusion among the kinds of valid cognitions extrinsically ascertained evidently was Sa pan's way of defending Dharmottara's position against the polemics of rNgog Lotsaba and Phywa pa. For the time being, however, it must also be stated that this seems to be a
89) rNam bshad 133a4f.: la la ni gzhan kho na las gyur te 'khrul rgyu can dang zhes dang !..
90) Pham byed II 349,6-350,1: yang ji skad du / Chos mchog las / la la ni don med na mi 'byung ba gzhan las (: la) nges par byed do // zhes mtshams sbyar nas bden pa gzhan las nges dang/ khyad par gzhan las nges gnyis gsal bar gsungs shing/ snang ba nyid kyang gzhan las nges pa skabs der ma gsungs mod / slob dpon Chos mchog 'di mngon sum du bzhed mi bzhed rnam par dpyad dgos pa yin no II Rigs pa'i gter las ni 'di la yid ma gtad dang 'khrul rgyu can gnyis su phye nas gnyis ka mngon sum gyi tshad mar bzhed payin no ll
91) Pham byed II 354,5ff.
92) Pham byed II 354,6f.: ji skad du / Thad Idan las / smig rgyu'i rang gi mtshan nyid bzung du zin kyang smig rgyu rang bzhin du ma nges pa'i phyir / smig rgyu la mngon sum min te rnam pa de lta bu gzhan dag kyang shes par bya'o zhes gsungs ngo // This is a quotation of PVinȚ 15a5f. (cf. also PPar I 226b6-8; PPar II 242b2; with fragmentary Sanskrit-materials in TBV 471,12 and TSP 946,22f.); for remarks in the same sense cf. PVinȚ(a) 11,19-12,5.
93) Cf. rNam bshad 35b1ff., particularly 36a2f. for a definition of snang la ma nges pa'i blo and its three kinds yul rnam pa mi gsal ba (a perception where the object is of indistinct appearance), yid ma gtad pa (a perception with distracted mind) and 'khrul pa'i rgyu mtshan gyis zin pa (a perception affected by grounds for error). The examples given for the second and third kind as probably conceived by Phywa pa chos kyi seng ge himself are: rtog pa nang du bsdus pa'i gnas skabs kyi gzugs 'dzin gyi mig shes (rNam bshad 35b4), and yid nya phyis la dngul du 'khrul pa'i gnas skabs kyi nya phyis 'dzin pa'i mig shes (rNam bshad 35b4).
The whole paraphrase of Phywa pa's list of cognitions has already been interpreted, however with some gross misunderstandings, by van der Kuijp in his paper of 1978 (cf. note 3 above), 361ff.
For Shakya mchog Idan's presentation cf. Pham byed II 74,3-76,2.

Page #14

270
Erst STEINKELLNER
case of overinterpretation, since Dharmottara does not really hold the category he is supposed to have held.
Although we seem to be forced by these textual facts to "accuse" Sa pan of an overinterpretation of Dharmottara when he followed this curious interpretation of his predecessors, I think we can also suggest a good reason for this overinterpretation that has nothing to do with exegetical correctness in interpreting an acceptable Indian authority. This reason can be found, I think, in an attempt at systematical completeness, i.e. completeness of the list of possible items relevant under the category of "cognitions which are extrinsically ascertained".
I cannot elaborate this point here, but it seems to me that one of the specific characteristics of Tibetans in their theoretical disciplines is a natural desire towards a complete covering of an investigated area or theme. Thus I would dare to say - as an hypothetical afterthought that to Sa pan it seemed natural that rNgog Lotsāba's reference to the case where "appearance itself" is doubtful as having been taught by Dharmottara was not suspicious as such. Since - for completeness' sake - it would have been a well acceptable item on the list, and once recognized, isolated and labelled - for whatever reasons that we do not yet know - it would have been very odd not to find it in Dharmottara's writings at all.
This concludes my brief attempt to look into a corollary of the great theme of the definition of a valid cognition and the early development of its differentiation. In spite of the fragmentary material still available for the early period of Tibetan epistemology, the efforts of the scholars involved to interpret the Indian heritage and at the same time to shape systematical forms with considerable creativity are fascinating and impressive. The study of these developments of the various theoretical themes and problems of the Buddhist epistemological tradition proves to be a direct path to the very heart of Tibetan intellectual culture.

Page #15

Early Tibetan Ideas on the Ascertainment of Validity (nges byed kyi Ishad ma)
271
Abbreviations
1. Indian Sources
TBV TBh
TBh?
TS
TSP PPar II
PV
Tattuabodhavidhāyini. Ed. S. Sanghavi, B. Dośī. 5 vols., Ahamedabad 1924-1931. Tarkabhāṣā of Mokşakara Gupta. Ed. with a Sanskrit Commentary Embar Krishnamacharya. Baroda 1942. Tarkabhāṣā: Tarkabhāșa and Vādasthāna of Mokşākaragupta and Jitāripada. Ed. H. R. Rangaswami Iyengar. Mysore 21952. Tattvasangraha: Tattvasangraha of Acārya Shāntaraksita with the Commentary 'Pañjikā' of Shri Kamalashila. Ed. Dvarikadas Shastri. 2 Vols., Varanasi 1981. Tattvasangrahapanjikā (Kamalasila): cf. TS [Laghu-]Prāmānyapariksa (Dharmottara): P 5747 (Vol. 138, Tshad ma, Ze 236b1-252b4). Pramānavārttika-Kārika (Sanskrit and Tibetan). Ed. Yusho Miyasaka. Acta Indologica 2 (1971/72), 1-206. [The sequence of chapters in Miyasaka for my sequence I, II, III is III, I, II). Pramāņavārttikavștti: Dharmakīrti's Pramanavärttika with a commentary by Manorathanamdin. Ed. Rahula Sünkrityāyana. Patna 1938–40. Pramāņavārttikasvavṛtti: Raniero Gnoli, The Pramānavārttikam of Dharmakīrti, the first chapter with the autocommentary. Text and Critical Notes. Roma 1960. Pramānaviniscayaţikä (Dharmottara): cf. Steinkellner/Krasser 1989. Slokavārttika (Kumārila): Slokavārtikavyākhyā Tātparyatīkā of Umveka Bhatta. Ed. S.K. Ramanatha Sastri, revised K. Kunjunni Raja and R. Thangaswamy. Madras 1971.
PVV
PVSV
PVinT(a) ŚV
2. Tibetan Sources
(CW
"Collected Works") Kun bzanig
gSer mdog Paņ chen Shākya mchog Idan, Gyas pa'i bstan bcos tshad ma rnam
'grel gyi rnam bshad kun bzang chos kyi rol mtsho. CW 18, Thimphu 1975. 189ff. dGa' byed
---, Tshad ma'i mdo dang bstan bcos kyi shing rta'i srol rnams ji ltar 'byung ba'i tshul gtam du bya ba nyin mor byed pa'i snang bas dpyod ldan mtha' dag dga' bar byed pa.
CW 19, Thimphu 1975, 1ff. dGongs pa rab gsal rGyal tshab Dar ma rin chen, bsTan bcos tshad ma rnam nges tik chen dgongs pa
rab gsal. CW 7, New Delhi 1982, 3ff. dGongs gsal Paņ chen bSod nams grags pa, rGyas pa'i bstan bcos tshad ma rnam 'grel dgongs
pa rab gsal. CW 1, Mundgod 1985, 53ff. 'Jug sgo
Tsong kha pa Blo bzang grags pa, sDe bdun la 'jug pa'i sgo don gnyer yid kyi
mun sel. Sarnath 1969. Nyi ma
Glo ba mkhan chen bSod nams lhun grub, sDe bdun mdo dang bcas pa'i dgongs 'grel tshad ma rig«s» pa'i gter gyi 'grel pa'i rnam bshad rig lam gsal ba'i nyi ma. CW
2, Dehradun, 1ff. Thar lam gsal byed rGyal tshab Dar ma rin chen, Tshad ma mnam 'grel gyi Ishig le'ur byas pa'i rnam
bshad thar lam phyin ci ma log par gsal bar byed pa. CW 6, New Delhi 1982, 3ff. bs Dus pa
gTsang nag pa brTson 'grus seng ge, Tshad ma rnam par nges pa'i ți ka legs
bshad bsdus pa. (Otani University Tibetan Works Series 2) Kyoto 1989. rNam bshad
Go ram pa bSod nams seng ge, Tshad ma rigs pa'i gter gyi dka' ba'i gnas mnam
par bshad pa sde bdun rab gsal. CW in: SB 12, Tokyo 1969, No. 41. Nishioka
Nishioka Soshū, "Putun Bukkyo-shi" mokurokubu sakuin. I-III. Tokyo Daigaku Bungakubu Bunkaköryü Kenkyūshisetsu Kenkyü Kiyo 4, 1980, 61-92; 5, 1981, 43-94; 6, 1983, 47-201.

Page #16

272
Ernst STEINKELLNER
dPal ster
Pham byed II
Blo rigs
Blo rigs
Blo rigs
Blo rigs
Mun sel
"Od zer
Rang 'grel
Ngag dbang chos grags, Tshad ma rigs pa'i gter gyi dgongs don gsal bar byed pa'i legs bshad ngag gi dpal ster. New Delhi 1983. gSer mdog Pan chen Shakya mchog Idan, Tshad ma rigs gter gyi dgongs rgyan rigs pa'i 'khor los lugs ngan pham byed ces bya ba'i bstan bcos smad cha. CW 10. Thimphu 1975, Off Jam dpal bsam 'phel, Blo rig gi rnam bzhag nyer mkho kun 'dus blogsal mig 'byed. In: Rigs lam sgo brgya 'byed pa'i 'phrul gyi Ide mig dang po. Mundgod 1979. "Jam dbyangs bzhad pa'i rdo rje, Blo rigs kyi rnam bzhag nyung gsal legs bshad gser gyi phreng mdzes. CW 15, New Delhi 1973, 303ff. Phur Icogs Yongs 'dzin rdo rje chang, Tshad ma'i gzhung don 'byed pa'i bsdus grwa'i mnam bzhag rigs lam 'phrul gyi Ide mig ces bya ba las rigs lam che ba yul yul can dang blo rig gi rnam par bshad pa. In: Textbooks of Se-ra Monastery, ed. Tshulkrim-Kelsang and Shunzo Onoda. Kyoto 1985, 46ff. A kya yongs 'dzin dByang can dgra ba'i blo gros, Blo rigs kyi sdem tshig blang dor gsal ba'i me long. CW 1, New Delhi 1971, 515ff. mKhas grub Ge legs dpal bzang po, Tshad ma sde bdun gyi rgyan yid kyi mun sel. Beijing 1984. Go ram pa bSod nams seng ge, Gyas pa'i bstan bcos tshad ma rnam 'grel gyi rnam par bshad pa kun tu bzang po'i 'od zer. CW in: SB 11, Tokyo 1969, No. 38. Sa skya Pandita Kun dga' rgyal mtshan, Tshad ma rigs pa'i gter gyi rang 'grel. CW in: SB 5, Tokyo 1969, No. 20 (=SB); Tshad ma rigs gter gyi rtsa dang rang grel. Dehradun 1985 (= D). Mus rab 'byams pa Thugs rje dpal bzang, Tshad ma rigs, pa'i gter gyi rnam par bshad pa rigs lam rab tu gsal ba'i nyi ma. In: Kun mkhyen Go bo rab 'byams pa bSod nams seng geli bka' 'bum kha skong, pha. Dehradun 1985. dGe 'dun grub pa, Tshad ma'i bstan bcos chen po rigs pa'i rgyan. CW 4, Gangtok 1979, 97ff. Sa skya Pandita Kun dga' rgyal mtshan, Tshad ma rigs pa'i gter. CW in: SB 5, Tokyo 1969, No. 19. Bo dong Paņ chen Phyogs las rnam rgyal, Tshad ma rigs snang. CW 7, New Delhi 1969. mKhas grub dGe legs dpal bzang po, rGyas pa'i bstan bcos tshad ma rnam 'grel gyi rgya cher bshad pa rigs pa'i rgya mtsho. CW 10, 620ff.; 11, 3ff. New Delhi 1980. 'U yug pa Rigs pa'i seng ge, Tshad ma rnam 'grel gyi 'grel pa rigs pa'i mdzod. 2 Vols., Delhi 1982. gSer mdog Pan chen Shakya mchog Idan, Tshad ma rigs pa'i gter gyi rnam par bshad pa sde bdun ngag gi rol mtsho. CW 19, Thimphu 1975, 447ff. Sa skya pa'i bka' 'bum. 15 Vols., compiled by bSod nams rgya mtsho. Tokyo 1968-1969. Go ram pa bSod nams seng ge, sDe bdun mdo dang bcas pa'i dgongs pa phyin ci ma log par 'grel pa tshad ma rigs pa'i gter gyi don gsal bar byed pa. CW in: SB 11, Tokyo 1969, No.40.
Rab gsal
Rigs rgyan
Rigs gter
Rigs snang
Rigs pa'i rgya
mtsho
Rigs mdzod
Rol mtsho
SB
gSal byed
3. Secondary Sources
Bühnemann 1980
van Bijlert 1989
Gudrun Bühnemann, Der Allwissende Buddha. Ein Beweis und seine Probleme. Ratnakirti's Sarvajñasiddhi. Wien. Vittorio A. van Bijlert, Epistemology and Spiritual Authority. The Development of Epistemology and Logic in the Old Nyāya and the Buddhist School of Epistemology

Page #17
273 Early Tibetan Ideas on the Ascertainment of Validity (nges byed kyi tshad ma) with an Annotated Translation of Dharmakirti's Pramanavarttika II (Pramana siddhi) vv. 1-7. Wien. Frauwallner 1934 E. Frauwallner, Dharmakirtis Sambandhaparikna. WZKM 41, 261-300. D'Sa 1980 Francis X. D'Sa, Sabdapramanyam in Sabara and Kumarila. Towards a Study of the Mimamsa Experience of Language. Vienna. Jackson 1987 David P. Jackson, The Entrance Gate for the Wise (Section III). Sa-skya Pandita on Indian and Tibetan Traditions of Pramana and Philosophical Debate. 2 Vols., Wien. Kajiyama 1966 Yuichi Kajiyama, An Introduction to Buddhist Philosophy. An Annotated Translation of the Tarkabhasa of Moksakaragupta. Memoirs of the Faculty of Letters, Kyoto University No.10. van der Kuijp 1978 Leonard W. J. van der Kuijp, Phya-pa Chos-kyi seng-ge's Impact on Tibetan Epistemological Theory. Journal of Indian Philosophy 5, 355-369. van der Kuijp 1983 ---, Contributions to the Development of Tibetan Buddhist Epistemology. From the eleventh to the thirteenth century. Wiesbaden. van der Kuijp 1989 ---, An Introduction to Gtsang-nag-pa's Tshad-ma rnam-par nges-pa'i ti-ka legs-bshad bsdus-pa. An Ancient Commentary on Dharmakirti's Pramanaviniscaya, Otani University Collection No. 13971. Otani University Tibetan Works Series, Vol. II. Much 1989 Michael Torsten Much, A Visit to Rahula Sankstyayana's Collection of Negatives at the Bihar Research Society: Texts from the Buddhist Epistemological School. Wien. Naudou 1968 Jean Naudou, Les Bouddhistes Kasmiriens au Moyen Age. Paris. Schmithausen 1965 Lambert Schmithausen, Mandanamisra's Vibhramavivekah. Mit einer Studie zur Entwicklung der indischen Irrtumslehre. Wien. Steinkellner/Krasser Pramanaviniscayatika (Dharmottara) P 7b6-13a7: Dharmottaras Exkurs zur 1989 Definition gultiger Erkenntnis im Pramanaviniscaya. Tibetischer Text, Sanskritmaterialien und Ubersetzung von Ernst Steinkellner und Helmut Krasser. Wien.