Once Again Vaisesika Sutra 3 1 13

Author(s): Johannes Bronkhorst
Publisher: Johannes Bronkhorst
Catalog link: https://jainqq.org/explore/269610/1

Page #1
ONCE AGAIN VAISESIKA SUTRA 3.1.13* Johannes BRONKHORST, Lausanne The study of philosophical Sutra works is beset with difficulties. Apart from the condensed style, which makes them sometimes difficult to understand even where no other problems intervene, we often have reason to suspect that these texts may have undergone interpolations and other modifications. In practice this means that, in order to understand a Sutra text, we should know as much as possible of its history, of the vicissitudes it has undergone from its beginning until today. Such detailed knowledge of the history of individual Sutra texts is not normally available. This is the reason why we have to be content, in most cases, with a global understanding of the kind of influences that Sutra texts undergo. Here we will concentrate on one such influence, viz., the one exerted by the commentary or commentaries that accompany them. It is known that Sutra texts are frequently extracted from commentaries that contain them. During this process of extraction mistakes can easily creep into the Sutra text: a sutra may be overlooked; or, more probably, a statement properly belonging to the commentary may be taken to be a sutra. Confusions of this kind were facilitated by the fact that commentaries of around the middle of the first millennium C.E. often fail to contain clear indications as to what is sutra, and what commentary. The use of the socalled Varttika style could not but add to the confusion. The extraction of a Sutra text from a commentary could lead to an incorrect result in other ways, too. There is evidence to show that commentators of around the middle of the first millennium occasionally felt free to comment upon the sutras in an order which deviates slightly from the 'correct' one. In itself this need not be looked upon as an attempt to change the order of the sutras. But whatever the intentions of these commentators, the effect of such a procedure might very well be that the Sutra text which someone else subsequently extracted from such a commentary would have some of the sutras in a modified order. Usually Indian Sutra texts are handed down to us in one single 'line of descent', at least where their early period is concerned. It is only on rare * 1 I thank T. Tillemans for help and advice. See Bronkhorst, 1992, for a brief survey.

Page #2
666 JOHANNES BRONKHORST occasions that we can show with certainty that commentators did actually comment upon the sutras in a changed order. There are, however, some clear cases, which I will now present. Consider first the Samkhya Karika. This work does not consist of sutras, but of karikas. A considerable number of more or less early commentaries on it have been preserved, and the exact chronological relationship between them is not easy to determine. The most elaborate and interesting of these commentaries is the Yuktidipika. This text is not only interesting from the point of view of its contents. Its form, too, is special; it constitutes a perfect example of what I have called the Varttika style. In the present study we are interested neither in its contents nor in its style, but in the manner in which it deals with the text it comments upon. The author of the Yuktidipika is aware of the fact that the Samkhya Karika consists of karikas. This we must conclude from his use of the term saptati 'seventy', hence 'work consisting of seventy karikas', to refer to the Samkhya Karika in his introductory verses. This same term saptati, along with the term arya which refers to the metre of the work, occurs again in the concluding verses of the Samkhya Karika as they are found, and paraphrased, in the Yuktidipika. In spite of this, the Yuktidipika, unlike all other surviving commentaries, treats the Samkhya Karika as if it consisted of sutras, not of karikas.3 It frequently divides the karikas into smaller parts, which it comments upon and refers to as sutras. Indeed, it never gives the slightest hint that these 'sutras' together constitute karikas, so much so that its third Ahnika ends right in the middle of the discussion of what we call karika 15; the remainder of karika 15 is commented upon in Ahnika 4. Sometimes sutra and karika coincide; in such cases a whole karika can actually be referred to as sutra; an example is karika 19, which is called sutra in its discussion in the Yuktidipika (p. 84 1. 7-8).5 Interestingly, on two occasions the 'sutras' of the Yuktidipika do not occur in the order of the karika concerned, as these latter are known from all the other surviving commentaries. 2 Solomon (1974) studies eight of them. In this respect the Yuktidipika has parallels in the Abhidharmakosa Bhasya and Madhyantavibhaga Sastra; see Bronkhorst, 1992. See, e.g., YD p. 91. 10, p. 671. 2, p. 98 1. 3. Note that Sadyojyotis' commentary (before 9th century) on the Svayambhuvasutrasangraha calls the verses of this text 'sutras', as does the title itself. According to Filliozat (1991: xvii), the term sutra here "refere plutot a la parole d'un etre a qui l'on attribue la plus haute autorite".

Page #3
ONCE AGAIN VAISESIKA SUTRA 3.1.13 667 Consider karika 4. This reads, in all the commentaries except the Yuktidipika: drstam anumanam aptavacanam ca sarvapramanasiddhatvat/ trividham pramanam istam prameyasiddhih pramanad dhill. The Yuktidipika (p. 29 f.) comments, in this order, on the following parts: (i) prameyasiddhih pramanad dhi, (ii) trividham pramanam istam, (iii) sarvapramanasiddhatvat, (iv) dystam anumanam aptavacanam ca. These parts constitute the karika, but their order has been reversed. No need to add that in this order nothing remains of the arya metre. . In the case of karika 4 one might think that the author of the Yuktidipika took this karika as a single unit, and commented upon its parts in a different order. No such position can be maintained in connection with karikas 6 and 7. These karikas occur, in almost the same form, in all the surviving commentaries, and must therefore be looked upon as integral parts of the Samkhya Karika, at least at the time of composition of the Yuktidipika. The Yuktidipika has these two karikas, but it has interposed karika 7 between the first and second half of karika 6. That is to say: the normal order is 6a-6b-7a-7b, but the Yuktid pika has 6a-7a-7b-6b. Again, the arya metre is thoroughly disturbed in this manner.6 There can be no doubt that the author of the Yuktidipika consciously changed the order of the 'sutras' of the Samkhya Karika; or perhaps: he consciously decided to comment upon them in an order which differs from the original one. The tradition preserved in all the other commentaries guarantees this sufficiently. This certainty makes the procedure of the 6 K. Preisendanz has kindly sent me a portion of her forthcoming book, in which she mentions the possibility that the author of the Yuktid Ipika did not (yet?) look upon SK 7 as a karika. She makes this suggestion because of the modified order of the karikas, and because karika 7, in the interpretation of the Yuktid Ipika, is invoked by an opponent. I find her suggestion nonetheless problematic, mainly because already the Samkhya Karika as translated into Chinese by Paramartha contains this karika. (For a discussion of the date of the Yuktidspika, see Bronkhorst, 1985: 93-94.) Preisendanz's suggestion further seems to necessitate the assumption that the Yuktidipika is not only older than all the other commentaries, but also that it was looked upon by the authors of the other commentaries as in some way authoritative. This again is hard to harmonize with the doctrinal differences which exist between the Yuktidipika and some of the later commentaries. One such difference concerns the question whether the tanmatras have one quality each, or an increasing number from one to five, depending on which tanmatra one is talking about; and the related question whether the tanmatras produce one element each, or whether they produce the elements jointly; see Bronkhorst, 1994, for details. The problems mentioned by Preisendanz can, of course, equally well be solved by the assumption that the author of the Yuktidipika felt free, not only to comment upon the karikas in a slightly modified order, but also to interpret one karika as representing the opinion of an opponent.

Page #4
668 JOHANNES BRONKHORST Yuktidipika all the more interesting. It shows beyond reasonable doubt that at least some commentators in the first millennium felt free to change the order of the sutras on which they commented. The Samkhya Karika is certainly not the only text the order of whose sutras has been changed. It may however be the only text where there is so little occasion to look for alternative explanations. It is known, for example, that the Brahma sutras occur at some places in a different order in the commentaries of Sankara and Ramanuja.7 Here, too, it is reasonable to assume that someone changed the original order. However, the Brahma sutras as they survive today are written in such a manner that it is virtually impossible to decide what this original order may have been. Let us now turn to Vaisesika sutra (VS) 3.1.13.8 It reads: atmendriyamano'rthasannikarsad yan nispadyate tad anyat A number of authors, among them the oldest whose testimony has been preserved, see in this sutra a definition of perception: (i) Dignaga remarks in his Pramanasamuccaya: "For the Vaisesikas there is a definition, mentioned in the Sutra, of perception in respect to substance (dravya), [which is made meaningful] by a certain relation [to the preceding sutras]. It says: "That [cognition] which is brought about by the contact of the soul (atman), the sense (indriya), the mind (manas), and the object (artha) is [perception as] a separate one [of the pramanas]'." There can be no doubt that the sutra quoted by Dignaga is VS 3.1.13. Dignaga's remark to the extent that the sutra "[is made meaningful] by a certain relation [to the preceding sutras]" is noteworthy and must be kept in mind; we'll return to it later. 7 See Bronkhorst, 1981: 317-18 n. 6. 8 This is its number in the version contained in Candrananda's commentary, edited by Jambuvijaya (C). It is 3.1.20 in the version of Bhatta Vadindra, also contained in the anonymous Vrtti, both edited by A. Thakur (V), 3.1.18 in the version contained in Sankara Misra's Upaskara, reproduced and translated in Sinha, 1911 (U). Where we use only one number, the reference is to C. The present sutra contains the word manas in versions C and V and in a number of quotations of this sutra in other works; manas is lacking in version U and in "one demonstrably wrong translation of the Pramanasamuccaya Vrtti" (Isaacson, 1990: 27). 9 Translation Hattori, 1968: 42; the two Tibetan versions on the basis of which the translation was made are reproduced ibid. p. 198-199.

Page #5
ONCE AGAIN VAISESIKA SUTRA 3.1.13 669 (ii) Simhasuri quotes the following definition of perception in his Nyayagamanusarini (ed. Jambuvijaya, vol. I, p. 110): atmendriyamano'rthasannikarsad yan nispadyate tad anyat, atma manasa mana indriyena indriyam artheneti catustayatrayadvayasannikarsad utpadyamanam pratyaksam This is our Vaisesika sutra along with an explanation. The sutra was apparently quoted - in order to be rejected - in Mallavadin's Dvadasara Nayacakra, which Simhasuri comments. 10 (iii) The Yuktidipika (p. 34, 1. 29-30) cites this sutra besides other definitions of perception. (iv) Jayantabhatta's Nyayamanjari cites the sutra in the following, slightly amplified, form (p. 280, 1. 15-16): yad api kaiscit pratyaksalaksanam uktam 'armendriyamano'rthasannikarsad yad utpadyate jnanam tad anyad anumanadibhyah pratyaksam'iti... Here too there can be no doubt that the sutra is read as a definition of perception. (v) The anonymous Vitti on the Vaisesika Sutra edited by A. Thakur (1957), too, explains the sutra as a definition of perception. So does the commentary by Bhatta vadindra edited by Thakur (1985), of which the former is an abbreviation. (vi) The anonymous Sarvasiddhantapravesaka, in its chapter on Vaisesika, contains the following passage (Jambuvijaya, 1961: 145): aha pratyaksalaksanam kim iti cet, tadaha 'atmendriyamano'rthasannikarsad yan nispadyate tad anyat '/asya vyakhya: atma manasa yujyate mana indriyena indriyam artheneti/ tatas catustayasannikarsad ghatarupadijnanam, trayasannikarsac chabde, dvayasannikarsat sukhadisul evam pratyaksam nirdistam/ 10 Be it noted in passing that Simhasuri appears to quote in matters Vaisesika usually from the katandi, a Vaisesika work probably written before Dignaga's Pramanasamuccaya (see Bronkhorst, 1993). Do we have to conclude that also the present explanation given by Simhasuri derives from the Katandi? It is not possible at this point to address this question.

Page #6
670 JOHANNES BRONKHORST (vii) Akalanka's Tattvartha-Varttika (p. 53, 1. 32.) cites VS 3.1.13 in an enumeration of definitions of perception. This interpretation, though supported by early authorities - among them our earliest, is not accepted by some Sanskrit authors, and by several modern scholars (Hattori, 1966; Oetke, 1988: 303-319; Preisendanz, 1989: 150 f.; Nozawa, 1989: 71).11 They reject this interpretation for various reasons, which all boil down to one single factor: the context. The whole of Ahnika 3.1 is believed to be intended to prove the existence of the soul. 12 It is in this way that the three commentators of the sutras explain this section, and it is clear that a definition of perception in the midst of such an argument would be out of place. Here the question can legitimately be raised whether we are entitled to discard the massive and ancient evidence provided by Dignaga and the other authors mentioned above on the basis of mere considerations of context. We have no surviving commentary on the Vaisesika Sutra that is even approximately as old as Dignaga, nor do we possess certain knowledge of what the context of VS 3.1.13 looked like during his time. (Recall that according to Dignaga the interpretation of 3.1.13 as a definition of perception is supported by its context; in the present situation of the text this can hardly be said to be the case.) Arguments based on context show a marked degree of confidence in the reliability of the Sutra text as it has been handed down to us, and this without supporting evidence. The (present) context of VS 3.1.13, when looked at more closely, presents a number of peculiar features, which justify a certain suspicion with regard to the present order of the sutras. Ahnika 3.1, as stated above, is believed to prove the existence of the soul (atman). But strangely enough, the proof of the existence of the soul is again taken up in VS 3.2.4, without the slightest hint that this is the second time the topic is addressed. 13 Sutras 3.2.1-3, which are situated between the two sections purportedly dealing with the proof of the soul, discuss the proof of the existence of the mind (manas). If 11 Honda (1990: 144 (29)) accepts the sutra as a definition of perception. 12 Hattori is explicit about this (p. 897 (100)): "Although VS is not skilful in its arrange ment of topics, there certainly is an order, which does not allow any arbitrary interpretation to be put on a sutra." In Oetke's interpretation, 3.1.13 is a reply to an objection which finds expression in 3.1.7. The intervening sutras are perhaps interpolated. 13 In version V sutra 3.2.4 concerns only the proof of the existence of the soul in others. This reading of the sutra (pranapananimeson mesaj fvanamanogat indriyantaravikarah paratmani lingam) disagrees however with the evidence of the Padarthadharmasangraha and its commentaries. See below.

Page #7
ONCE AGAIN VAISESIKA SUTRA 3.1.13 671 we follow the (relatively recent) indigenous commentaries and several modern scholars, we are asked to believe that Adhyaya 3 contains two sections dealing with the proof of the soul, which are separated, for no obvious reason, by a section which establishes the existence of the mind. 14 The difficulties do no end here. Sutras 3.1.1-12, which allegedly establish the existence of the soul (atman), do not once mention the word atman, nor indeed any other word for soul. This constitutes a marked contrast with the second section concerned with the proof of the soul: VS 3.2.4 enumerates a number of atmalingas, thus leaving no doubt as to its intentions. The preceding observations show that the context of VS 3.1.13 is by far not as clear as some may maintain. It is instructive at this point to study how the first discussion on the existence of the soul is initiated. The discussion starts with sutra 3.1.1: prasiddha indriyarthah No commentator claims that this sutra by itself introduces the topic of the soul. The sutra is rather presented as an introduction to sutra 3.1.2 (3.1.3 in the version of Bhatta Vadindra) which, it is claimed, presents an inference proving the existence of the soul. 15 Sutra 3.1.2 begins with the word indriyarthaprasiddhi, which obviously refers back to sutra 3.1.1. What strikes us here, is that there is no need in this context of sutra 3.1.1. In presenting "the knowledge / establishment of senses and objects / objects of the senses." as logical ground for something else, presumably the soul (or the difference of the soul from other things), it is superfluous to have this preceded by another sutra which states that "the senses and objects l objects of the senses are known / established". The only reasonable explanation I can think of for the presence of 3.1.1 at this place is that this is a sutra which the author of 3.1.2 (or 3.1.2-3) used as excuse and pretext for the introduction of one or more new sutras. This is of course only possible if 3.1.2 (3.1.2-3 in the case of Bhatta Vadindra) is a later addition to the Sutra text, newly composed when 3.1.1 was already considered to constitute part and parcel of the traditionally accepted Sutra text. Seen in this way, 3.1.1 does not, and never did, constitute part of the proof of the soul, but it could be used as point of departure for such a discussion. 14 Oetke (1988: 304) admits: "Dass das gesamte erste Ahnika des dritten Adhyaya dem Nachweis der Existenz einer Seele gewidmet ist, ist keineswegs so selbstverstandlich, wie es von manchen indischen Kommentaren und europaischen Interpreten angenommen wird." The emphasis here is not, however, on Seele, but on Existenz, as is clear from the following remark on the very next page: "Es sieht so aus, dass die zu beweisende These nicht die Existenz einer Seele ist, sonder der Umstand, dass der Atman ein (von den am Anfang des Werkes angefuhrten Padarthas) verschiedener Gegenstand ist." Or rather, with Oetke, that the soul is different from the padarthas enumerated thus far.

Page #8
672 JOHANNES BRONKHORST It may be possible to explain in this way the use that was made of 3.1.1 by a later commentator. But what was its function before this commentator used it to introduce a discussion on the existence of the soul? An easy explanation can be provided if we are willing to consider that the commentator concerned did not only use 3.1.1 for his purposes, but moved it away from its original context. Explaining 3.1.1 in its present position is difficult, irrespectively of how one wishes to interpret the remainder of Ahnika 3.1. The fact that it deals with indriya and artha, suggests that it originally belonged between 3.1.13 - which deals with indriya, artha, manas and atman - and the discussions of manas and atman in sutras 3.2.1 ff. If we remove 3.1.14 (which is suspect because it seems to presuppose that the preceding sutra dealt with the existence of the soul), 16 we arrive at the following sequence of sutras (the variants in the other two versions are minor and do not affect the argument): 3.1.13: atmendriyamano'rthasannikarsad yan nispadyate tad anyat 3.1.1: prasiddha indriyarthah 3.2.1: atmendriyarthasannikarse jnanasyabhavo bhavas ca manaso lingam 3.2.2: dravyatvanityatve vayuna vyakhyate 3.2.3: prayatnayaugapadyaj jnanayaugapadyac caikam manah 3.2.4: pranapananimesonmesajivanamanogatindriyantaravikarah sukha-1 duhkhe iccha-dvesau prayatnascety atmalingani 3.2.5: dravyatvanityatve vayuna vyakhyate This would then mean: - That which comes about as a result of contact (sannikarsa) between soul (atman), sense organ (indriya), mind (manas) and object (artha), is a different [kind of cognition) (viz., perception) (3.1.13) [From among these four factors) sense organs and objects are well-known [and need no further explanation) (3.1.1) 16 Sutra 3.1.14 appears to deal with the proof of the existence of a soul in others. It cannot therefore have found its present position until after the rest of Ahnika 3.1 had been given an interpretation that concerns the existence of the soul. This does not necessarily imply that 3.1.14 is a late sutra. Nozawa (1989) has argued that it is old, dating from the time when the soul was still thought of as of limited size. (See however Bronkhorst 1993a: 87 f. on the size of the soul in early Vaisesika.) All we can say is that its present position cannot be all that old.

Page #9
ONCE AGAIN VAISESIKA SUTRA 3.1.13 673 - The inferential mark for the existence] of a mind is that there is (some times] cognition and (sometimes] not, even though there is contact between soul, sense organ and object (3.2.1) The fact that (mind) is a substance and eternal is explained by the same arguments as in the case of] wind17 (3.2.2) - There is (only) one mind (in each body] because (several] efforts do not occur simultaneously, nor do (several] cognitions (3.2.3) - The inferential marks (for the existence] of a soul are: breathing in and breathing out, shutting and opening the eyes, life, movement of the mind, the modifications of the other senses, pleasure and pain, desire and aver sion, and volition (3.2.4) - The fact that (soul] is a substance and eternal is explained by (the same arguments as in the case of] wind18 (3.2.5) This sequence makes sense, is coherent, and even clear enough to allow us to interpret the sutras without the help of a commentary. Recall that the reason why someone should comment upon 3.1.1 at its present position, i.e., at the beginning of Adhyaya 3, is quite clear: this allowed him to address the question of the existence of the soul at that place. We will see below why this could be a concern to this commentator. First, however, we must consider the question what originally preceded sutra 3.1.13. If we accept, with the ancient witnesses cited above, that this sutra was a definition of perception, the question can be answered with a fair amount of confidence. The sutra must have been preceded by a related discussion. This is shown by the peculiar form of 3.1.13. Recall that this definition of perception does not mention the word 'perception'; instead it has anyat '[something different / the other one'. It clearly continues a discussion, which distinguished (at least) two kinds of knowledge or cognition (nana(?); or perhaps pramana? see below). The form of 3.1.13 indicates that it was preceded, in all probability, by one or more sutras about inferential knowledge. This is what Jayantabhatta suggests in so many words (see above),19 and indeed, even in its present shape the Vaisesika Sutra has some sutras somehow dealing with inference immediately preceding 3.1.13. 17 This refers to sutras 2.1.11 adravyavattvad dravyam and 2.1.13 adravyavattvena nityatvam uktam, both of which occur in the discussion of wind. 18 See preceding note. 19 Similarly Jinendrabuddhi; see below.

Page #10
674 JOHANNES BRONKHORST It would be sheer temerity to pretend to be able to reconstruct the original form and history of sutras 3.1.3-12. One thing seems however certain: sutra 3.1.13 was once preceded by sutras dealing with inferential knowledge. The word anyat in 3.1.13, moreover, suggests that at one time these preceding sutras contained some such neuter noun as jnanam. No such noun is at present to be found in 3.1.3-12. Nor do any of these sutras seem to introduce, or define, inferential knowledge. Here, however, the following is to be observed. The first part of sutra 3.1.8 reads: samyogi, samavayi, ekarthasamavayi, virodhi ca. The commentators seem to think that these adjectives characterize the word linga 'inferential mark', which is not mentioned in the sutra. But there is another sutra (9.18) which reads: asyedam karyam karanam sambandhi ekarthasamavayi virodhi ceti laingikam. The similarities with 3.1.8 are striking, yet 9.18 does not speak of inferential marks, but of inferential knowledge. It can be translated: "Inferential [knowledge is characterized by the relation:] 'this is the effect of that', 'this is the cause of that', 'this is related to that', 'this inheres in the same object as that', 'this is opposed to that'." It is therefore conceivable that 3.1.8, too, introduced inferential knowledge. The person who changed the order of the sutras may have adjusted sutra 3.1.8 to his purposes by trimming it. Alternatively we may consider the possibility that the scribe who extracted the sutra from its commentary failed to extract the whole sutra. * Whatever the exact original shape of the sutras, it seems probable that Adhyaya 3,20 prior to the changes pointed out above, discussed inferential and perceptual knowledge before turning to the mind (manas) and the soul (atman). Mind and soul being the last two of the nine substances (dravya) enumerated in sutra 1.1.4, Adhyaya 3 completes the discussion of the substances, the earlier ones having been enumerated in Adhyaya 2, as follows: earth (prthivi) 2.1.1; water (ap) 2.1.2; fire (tejas) 2.1.3; wind (vayu) 2.1.4; ether (akasa) 2.1.5; wind 2.1.9 f.; ether 2.1.26 f.; time (kala) 2.2.6 f.; space (dis) 2.2.12 f. Why was the treatment of inferential and perceptual knowledge inserted into the discussion of the substances? The answer is obvious: sutra 3.2.1, which proves the existence of the mind, refers back to the definition of perception. Even when there is contact sannikarsa) between soul (atman), sense organ (indriya) and object (artha), there may or may not be knowledge; this fact indicates the existence of a fourth factor, viz., the mind 20 It is not, of course, claimed here that the original Vaisesika Sutra was already divided into Adhyayas and Ahnikas.

Page #11
ONCE AGAIN VAISESIKA SUTRA 3.1.13 675 (manas).21 The fact that the soul is dealt with after the mind is explained by the fact that sutra 3.2.4, which proves the existence of the soul, presents as one of the arguments the movement of the mind (manogati). What was the purpose of the commentator who changed the order of the sutras? Again it is not difficult to divine the answer. This commentator apparently wanted the discussion of the substances to continue without interruption. After the treatment of space (dis) the next substance mentioned in sutra 1.1.4 was the soul (atman). He introduced this topic in the way we now know, i.e., before the mind. One final question must be addressed: When did the change of order take place? Better perhaps: when was the commentary written which commented upon the sutras in their modified order? Here22 we have to consider the following statement in Prasastapada's Padarthadharmasangraha (Ki p. 97 1. 25-26, Ny p. 219 1. 3, Vy I p. 134 1. 17-18): atmalingadhikare buddhyadayah prayatnantah siddhah "In the section on inferential marks of the soul (the qualities) from consciousness (buddhi) to effort (prayatna) have been established." The early commentators on the Padarthadharmasangraha all agree that this statement refers to the Vaisesika Sutra. Sridhara (Ny p. 219 1. 9) specifies that the reference is to the pranapanadisutra. Udayana provides the following commentary (Kip. 98 1. 10-11): pranadisutre buddhyadayah prayatnantah siddhah yady api buddhis tatra kantharavena nasti tathapi sukhadaya eva svakaranataya tam aksipanti/ "In the pranadisutra the qualities from consciousness to effort have been established. Although consciousness does not figure explicitly in that (sutra), [the qualities) happiness (sukha) etc. suggest it as it is their cause." These remarks show that the three early commentators on the Padarthadharmasangraha, as perhaps Prasastapada himself, knew sutra 3.2.4 more or less in the form which we find in versions C and U (pranapananimesonmesajivanamanogatindriyantaravikarah sukhaduhkhe icchadvesau prayatnas cety atmalingani / sukhaduhkhecchadvesaprayatnas catmano lingani), and not as we find it in V (pranapananimesonmesajivanamanogatindriyantaravikarah paratmani lingam). But there is a problem connected with the identification proposed by the commentators, as pointed out by Udayana. Sutra 3.2.4 does not mention consciousness (buddhi). Udayana's solution to the prob 21 VS 3.2.1: atmendriyarthasannikarse jnanasyabhavo bhavas ca manaso lingam. Hattori (1966: 893 (104)) has already drawn attention to the parallelism that exists between the interpretations of VS 3.1.1-2 offered in the three oldest commentaries, and a passage in the Padarthadharmasangraha (Ki p. 84 & 86, Ny p. 176, 179 & 182, Vy p. 133-34).

Page #12
676 JOHANNES BRONKHORST lem is not convincing. A far more convincing solution presents itself if we assume that Prasastapada already knew the beginning of Adhyaya 3 more or less in the form in which we know it. Sutra 3.1.2 (3), in particular, uses indriyarthaprasiddhi as inferential mark to prove the existence of the soul. Indriyarthaprasiddhi is, of course, a kind of knowledge (jnana) or consciousness.23 Sutra 3.1.13, too, is interpreted by some commentators (C, U) as presenting consciousness as an inferential mark of the soul. There is no evidence, on the other hand, to believe that already the author of the Carakasamhita knew the present order and interpretation of Adhyaya 3 of the Vaisesika Sutra. This text enumerates a number of inferential marks of the highest self (lingani paramatmanah) in Sarirasthana 1.70-72.24 A. Comba (1987: 54 f.) has pointed out that this enumeration draws upon two sources, the one being VS 3.2.4, the other Yajnavalkyasmrti 3.174-175 (Stenzler, 1849: p. 99 of the edition).25 The elements of VS 3.2.4, Comba suggests, were taken as basis, to which the elements of the Yajnavalkyasmrti have been added. The fact that buddhi 'consciousness' figures in the list of the Caraka-samhita does not, therefore indicate that Ahnika 3.1 of the Vaisesika Sutra was read and understood as it is at present, for buddhi occurs in the list of the Yajnavalkyasmrti. The fact that buddhi is added after the elements occurring in VS 3.2.4, agrees with the general procedure of the author of the Carakasamhita, drawn attention to by Comba, to add the elements of the Yajnavalkyasmrti after those taken from the Vaisesika Sutra. It looks, then, as if the commentator who used sutra 3.1.1 as an introduction to a discussion of the proof of the soul, lived before Prasastapada, though perhaps not very long before him. Some facts suggest that his new interpretation of the sutras of Ahnika 3.1 could not impose itself immediately. There is, on the one hand, the ongoing tradition of authors who look upon VS 3.1.13 as a definition of perception. Equally interesting is the fact that both the commentators Sridhara and Udayana fail to understand that 23 This is precisely what Sarkara Misra says in his Upaskara (p. 85 1. 17-19): yady api jnanam eva lingam iha vivaksitam tathap Indriyarthaprasiddhe rupadisaksatkarasya prasiddhatarataya tadrupyenaiva lingatvam uktam. 24 Caraka Sa. 1.70-72: pranapanau nimesadya jIvanam manaso gatih/ indriyantarasamcarah preranam dharanam ca yat// desantaragatih svapne pancatvagrahanam tatha/ drstasya daksinenaksna savyenavagamas tatha// iccha dvesah sukham duhkham prayatnas cetana dhrtih/ buddhih smrtir ahankaro lingani paramatmanah//. 25 Yajnavalkyasmrti 3. 174-75 (ed. Stenzler): ahankarah smrtir medha dveso buddhih sukham dhrtih/ indriyantarasancara iccha dharanaj Ivite// svargah svapnas ca bhavanam preranam manaso gatih/ nimesas cetanayatna adanam pancabhautikam//.

Page #13
ONCE AGAIN VAISESIKA SUTRA 3.1.13 677 Prasastapada's buddhi in the statement cited above refers to sutra 3.1.2 and/ or 13 (Vyomasiva's position cannot be determined with certainty). Do we have to conclude that they still knew the earlier interpretation, perhaps even the earlier order of the sutras in Ahnika 3.1? Did the two interpretations of Ahnika 3.1 exist for a while side by side? In this connection it is interesting to cite Hattori's (1968: 134-35 n. 4.3) paraphrase of some remarks from Jinendrabuddhi's commentary on Dignaga's Pramanasamuccaya: "Jinendrabuddhi remarks that the relation of VS, III, i, 13, to the preceding sutras is variously interpreted by different commentators. He refers to the following two interpretations: (1) The universal apprehension (prasiddhi) is nothing other than knowledge (jnana). It therefore follows that it is an attribute (guna), and is non-eternal (anitya). That which is non-eternal has a cause (karana). Thus the sutra in question indicates the cause of knowledge and also mentions that knowledge as an effect is different from its causes, as a pot as an effect is different from its cause, clay. (2) Since the preceding sutras explain anumana, one might consider anumana as the only pramana. VS, III, i, 13, forestalls this by mentioning pratyaksa as a separate pramana. As Jinendrabuddhi says, VS, III, I, 13, can be understood as providing the definition of pratyaksa according to the second interpretation but not the first. Dignaga's implication when he says "by a certain relation [to the preceding sutras)' (kenacit sambandhena) should be understood as referring to these different interpretation; (Pramanasamuccaya-tika] [Sde-dge ed., Tohoku, No. 4268] 53a.3-536.1 ([Peking ed., Tibetan Tripitaka, No. 5766] 59b.4-60a.3)." Note in particular the remark, in the second interpretation, "since the preceding sutras explain anumana" (rtags las byung ba tshad mar ba rjod la, which Muni Jambuvijaya (1961: 174 1. 5) translates into Sanskrit laingikapramane 'bhihite). This seems to confirm our earlier supposition that a discussion of inferential knowledge (laingikam jnanam; perhaps better laingikam pramanam?) once preceded sutra 3.1.13, not a discussion of the inferential mark (linga) as maintained, for example, by Candrananda. Also Akalanka must have known two interpretations of VS 3.1.13, as has been pointed out by K. Preisendanz (1989: 152).26 Preisendanz (1989: 151 n. 39) also refers to *Vimalaksa's commentary on Mula-Madhyamaka-Karika 14.1, translated by Walleser from the Chinese into German (1912: 90). She observes that here "VS 3.1.13 is obvi 26 Compare Akalarka's remark cited above with Tattvartha-Varttika p. 46 1. 6-8: yasya matam - atmano jnanakhyo gunah, tasmac carthantarabhutah, "atmendriyamanorthasannikarsat yan nispadyate tad anyar" iti vacanad iti ...; also p. 50 1. 9-14.

Page #14
678 JOHANNES BRONKHORST ously used to demonstrate the difference between perception, perceptible object, and perceiver". She thinks that this is "a related line of interpretation" to her own, in which "3.1.13 most probably serves to show that cognition, here specifically perception, is different (anyar) from the soul, the latter being a factor involved in its production" (p. 150). This, if correct, would push the reordering and reinterpretation of Ahnika 3.1 back to a date well before Kumarajiva (344-413), who translated this commentary into Chinese. However, *Vimalaksa's remarks do not force us to draw such a conclusion. It is true that an opponent in his commentary - presumably a Vaisesika - uses VS 3.1.13 to demonstrate the difference between perception, perceptible object, and perceiver.27 But this can very well be done, even if one looks upon this sutra as a definition of perception. Also when interpreted as a definition of perception, this sutra distinguishes between the self (atman), the object of knowledge (artha), and the perceptual knowledge which results from their contact (sannikarsa) with the mind (manas) and the sense organ (indriya). We must therefore conclude that *Vimalaksa's remarks do not constitute evidence that the "new" interpretation of sutra 3.1.13 existed already in the fourth century C.E. Similar remarks should be made with regard to the following passage in Vasu's commentary on the Satasastra:28 The unbeliever says: A disciple of Uluka, who reads the Vaisesika Sutra, says that knowledge and atman are different, and that therefore the atman does not fall into the state of non-eternity, and yet that it is not without knowledge. Why? "Because atman and knowledge are united just like the possessor of an ox." For example, if a man is united with an ox, he is called the possessor of an ox. In the same way, from the union of the atman, the senses, the manas, and the objects, to the atman there is an occurrence of knowledge. Because of the union of the atman with knowledge, the atman is called a possessor of knowledge. 27 See Walleser, 1912: 90: "Frage: Selbst (atman), Vorstellung (manas), Sinn (indriya), Sinnesbereich (gocara): da (diese) vier Dinge vereinigt sind, ist Entstehen des Erkennens. Man kann Krug, Tuch usw., alle Dinge erkennen. Deshalb ist Sehen, zu Sehendes, Seher." The Chinese is to be found T. 1564 (vol. 30) p. 19a 1. 13-15. 28 I thank M. Nozawa, who drew my attention to this passage, and provided me with a translation - different from Tucci's (1929: 23-24) - which I here reproduce (with minor modifications). Nozawa points out that the latter part of the underlined portion (which corresponds to VS 3.1.13) follows Ui's and Hatani's Japanese translation. An alternative translation might be "the atman arises as a possessor of knowledge". The Chinese occurs T. 1569 (vol. 30) p. 1715 l. 7-12.

Page #15
ONCE AGAIN VAISESIKA SUTRA 3.1.13 679 Here VS 3.1.13 is used to prove the difference between the self and knowledge. This can be done, even if 3.1.13 is considered a definition of perception. We have come to the end of this article. It must be admitted that the reconstruction of the context and of the interpretation here presented of VS 3.1.13 cannot be definitely proved to be correct. They do, however, solve a number of problems which other interpretations had failed to solve. The least one can deduce from them is that the arguments which discard the interpretation of 3.1.13 as a definition of perception on the basis of its context are not conclusive. VS 3.1.13 can be interpreted as a definition of perception, as maintained by our earliest witnesses; and other difficulties surrounding Adhyaya 3 can be solved, if only we are willing to consider the possibility that the order of sutras which we find in the surviving versions of the Vaisesika Sutra may in one point deviate from their original order. This possibility in its turn, as we have seen, is supported by the fact that other commentators on Sutra texts are known to have occasionally changed the order of sutras on which they comment. In view of all this, we may conclude with a variant of an observation made by Oetke (1988: 310): The hypothesis that the sutras of Ahnika 3.1 have reached us in their original order may not be less speculative than the opposite assumption. References Akalanka: Tattvartha-Varttika. Edited by Mahendra Kumar Jain. 2nd ed. Delhi: Bharatiya Jnanpith. 1982. (Jnanpith Murtidevi Jaina Granthamala, Sanskrit Grantha no. 10.) Bronkhorst, Johannes (1981): "Yoga and sesvara samkhya." Journal of Indian Philosophy 9, 309-320. Bronkhorst, Johannes (1985): "A possible quotation from the Niruktavarttika known to Durga in the Yuktidepika." In: Proceedings of the Fifth World Sanskrit Conference (Varanasi, October 21-26, 1981). New Delhi: Rashtriya Sanskrit Sansthan. Pp. 90-100. Bronkhorst, Johannes (1992): "Two literary conventions of classical India." Etudes Asiatiques Asiatische Studien 45(2), 1991 (1992), 210-227. Bronkhorst, Johannes (1993): "The Vaisesika vakya and bhasya." Annals of the Bhandarkar Oriental Research Institute 72-73 (1991-92), 145-169. Bronkhorst, Johannes (1993a): "Studies on Bharthari, 5: Bharthari and Vaisesika." Etudes Asiatiques / Asiatische Studien 47 (1), Proceedings of the First International Confer ence on Bharthari, pp. 75-94. Bronkhorst, Johannes (1994): "The qualities of Samkhya." Wiener Zeitschrift fur die Kunde Sudasiens 37 (Festschrift G. Oberhammer), 309-322.

Page #16
680 JOHANNES BRONKHORST Carakasamhita: The Charakasamhita by Agnivesa, revised by Charaka and Dridhabala, with the Ayurveda-Dipika Commentary of Chakrapanidatta. Edited by Vaidya Jadavaji Trikamji Acharya. Third edition. Published by Satyabhamabai Pandurang, for the Nirnaya Sagar Press, Bombay. 1941. Comba, Antonella (1987): "Carakasamhita, Sarirasthana I and Vaisesika philosophy." In: Studies on Indian Medical History. Ed. G. Jan Meulenbeld and Dominik Wujastyk. Groningen: Egbert Forsten. (Groningen Oriental Studies, 2.) Pp. 43-61. Filliozat, Pierre-Sylvain (1991): Le Tantra de Svayambhu, vidyapada, avec le commentaire de Sadyojyoti. Geneve: Droz. Hattori, Masaaki (1966): "Studies of the Vaisesikadarsana (1): On the Vaisesikasutra III, i, 13." Journal of Indian and Buddhist Studies 14, 902-890/(95)-(107). Hattori, Masaaki (1968): Dignaga, On Perception. Being the Pratyaksapariccheda of Dignaga's Pramanasamuccaya, from the Sanskrit fragments and the Tibetan versions translated and annotated. Cambridge, Mass.: Harvard University Press. (Harvard Oriental Series, 47.) Honda, Megumu (1990): "A reading in the Vaisesika Sutra." Dobo Daigaku Kiyo 4, 172-79 (=(1)-(94)). Isaacson, H. (1990): A study of early Vaisesika: the teachings on perception. Thesis University of Groningen. Jambuvijaya, Muni (ed.) (1961): Vaisesikasutra of Kanada, with the Commentary of Candrananda. Baroda: Oriental Institute. (Gaekwad's Oriental Series, 136.) Jambuvijaya, Muni (ed.) (1966): Dvadasaram Nayacakram of Acarya Sri Mallavadi Ksamasramana, with the commentary Nyayagamanusarint of Sn Simhasuri Gani Vadi Ksamasramana. Part I. Bhavnagar: Sri Jain Atmanand Sabha. (Sr Atmanand Jain Granthamala Serial No. 92.) Jayantabhatta: Nyayamanjari. Vol. I. Edited by K.S. Varadacharya. Mysore: Oriental Research Institute. 1969. Nozawa, Masanobu (1989): "On the interpretation of Vaisesikasutra 3.1.14." Hokkaido Journal of Indological and Buddhist Studies 4, 62-74. Oetke, Claus (1988): 'Ich' und das Ich. Analytische Untersuchungen zur buddhistischbrahmanischen Atmankontroverse. Stuttgart: Franz Steiner. (Alt- und Neu-Indische Studien, 33.) Prasastapada: Padarthadharmasangraha. 1) Ki: Prasastapadabhasyam with the commentary Kiranavali of Udayanacarya, edited by Jitendra S. Jetly. Baroda: Oriental Institute. 1971. 2) Ny: Nyayakandali [of Sridhara], being a commentary on Prasastapadabhasya, with three sub-commentaries, edited by J.S. Jetly and Vasant G. Parikh. Vadodara: Oriental Institute. 1991. 3) Vy: Vyomavati of Vyomasivacarya, edited by Gaurinath Sastri. Varanasi: Sampurnanand Sanskrit Vishvavidyalaya. 2 vols. 1983-1984. Preisendanz, K (1989): "On atmendriyamanorthasannikarsa and the Nyaya-Vaisesika theory of vision." Berliner Indologische Studien 4/5, 141-213. Sankara Misra: Upaskara. In: Vaisesikadarsane maharsipravara-Prasastadevacaryaviracitam Prasastapadabhasyam vidvaccudamanis-Sankara-Misravinirmitah Upaskaras ca. Ubhayatra Kasistha-Vedavidyalayadhyapaka-Nyayopadhyaya Pam. Dhundhirajasastrikrtam Vivaranam. Sarvam etat samskrtam uktasastrinaiva. Kasi: Caukhamba. 1923. Sinha, Nandalal (tr.)(1911): The Vaisesika Sutras of Kanada with the Commentary of Sankara Misra and Extracts from the Gloss of Jayanarayana, together with notes from the commentary of Chandrakanta and an introduction by the translator. Reprint. Delhi: S.N. Publications. 1986.

Page #17
ONCE AGAIN VAISESIKA SUTRA 3.1.13 681 Solomon, Esther (1974): The Commentaries of the Samkhya Karika - A Study. Ahmedabad: Gujarat University. Stenzler, Adolf Friedrich (1849): Yajnavalkya's Gesetzbuch. Sanskrit und Deutsch. Neudruck der Ausgabe 1849. Osnabruck: Biblio Verlag. 1970. Thakur, Anantalal (ed.) (1957): Vaisesikadarsana of Kanada, with an anonymous commen tary. Darbhanga: Mithila Institute. Thakur, Anantalal (1966): "Studies in a fragmentary Vaisesikasutravstti." Journal of the Oriental Institute (Baroda) 14 (1965/66), 330-335. Thakur, Anantalal (ed.) (1985): Bhattavadindraracita-Vaisesikavartika-Krsnabhupalaracita Trisutriprakasa'jnatakartrkavrttibhir vilasitam Maharsi-Kanada-prantam Vaisesika darsanam. Darbhanga: Kamesvarasimha-Darabhanga-Samsksta-Visvavidyalaya. Tucci, Giuseppe (1929): Pre-Dinnaga Buddhist Texts on Logic from Chinese Sources. Sec ond edition. Madras: Vesta Publications. 1981. Walleser, Max (1912): Die Mittlere Lehre des Nagarjuna. Nach der chinesischen Version ubertragen. Heidelberg: Carl Winter. Yuktidipika. Edited by Ram Chandra Pandeya. Delhi: Motilal Banarsidass. 1967.