Studies On Bhartrhari 5 Bhartrhari And Vaisesika

Author(s): Johannes Bronkhorst
Publisher: Johannes Bronkhorst
Catalog link: https://jainqq.org/explore/269426/1

Page #1
STUDIES ON BHARTRHARI, 5: BHARTRHARI AND VAISESIKA Johannes Bronkhorst, Lausanne There are reasons to think that Bharthari's writings may shed light on the early history of Vaisesika. One of these is that he obviously knew the Vaisesika system. Almost all of its categories play a role in his work. Separate sections (samuddesa) of the Vakyapadiya are dedicated to the categories jati, dravya, guna and kriya. The relationship called samavaya - a special feature of Vaisesika - is mentioned and used repeatedly. Vaisesika substances appear as 'powers' (sakti), most notably kala (time) and dis (space) A second reason is Bhartrhari's chonological position. I have argued in another publication that Prasastapada's Padarthadharmasarngraha; as well as Dignaga's Pramanasamuccayavrtti before it, were heavily indebted to the katandi, a work written not long before Dignaga. This Katandi, I further argued, exerted a dominating influence on all Vaisesika literature that came after it, including perhaps the versions of the Vaisesika Sutra itself, not to speak of the surviving commentaries on this Sutra work.? Bhartrhari, on the other hand, lived long enough before Dignaga that someone different from Bhartrhari could write a commentary on the first two kandas of his Vakvapadiya still before Dignaga. Bhartrhari, therefore, lived and worked most probably before the katandi! If his work provides information on Vaisesika, it would then be one of the very few sources of information dating form the pre-Katandi period of this system. In what follows we shall consider some possible links between Bhartshari's Vakyapadiya and the Vaisesika of his days. 1 I thank A. Wezler and J. Houben for critical comments. The earlier articles in this series have appeared in the following periodicals: no. 1, Bulletin d'Etudes Indiennes 6 (1988), 105-143, no. 2, Studien zur Indologie und Iranistik 15 (1989), 101-117, no. 3, Asiatische Studien /Etudes Asiatiques 45 (1991), 5-18, no. 4, Asiatische Studien / Etudes Asiatiques 46, 1 (1992), 56-80. See Bronkhorst, forthcoming. 2

Page #2
JOHANNES BRONKHORST 1. The variegated colour (citrarupa) Karl H. Potter explains the variegated colour in his Encyclopedia of Indian Philosophies vol. II, which deals with the tradition of Nyaya-Vaisesika up to Gangesa, in the following terms (1977: 118): "Consider a substance with a mottled surface of more than one shade. Nyaya-Vaisesika insists on treating this substance as a single entity with one color of its own, but surely it is evident that it has several colors. Does this mean that one thing can be both, say, red and green all over at once? Uddyotakara seems to have originated one sort of answer to this, which is that in the list of shades one has to count as one kind of color that called 'variegated color' (citrarupa)." The problem which the variegated colour is meant to solve is clear. An object is, in Vaisesika ontology, different from its parts; it is a completely different entity, which has, necessarily, a colour of its own, different from the colours of its parts. What is the colour of a whole whose parts do not all have the same colour? The problem is inherent in the most fundamental assumptions of Vaisesika, and is likely to be as old as the system itself. Why then do we not find this particular answer until Uddyotakara, a Nyaya author who may have been a contemporary of Prasastapada? The reason why we don't find the variegated colour mentioned in our earliest Vaisesika texts appears to be that we have so few of them. There is reason to believe that the variegated colour played a role in the system already before Prasastapada. Otto Grohma (1975: 151f.) has drawn attention to the passage in Vasubandhu's Abhidharmakosa Bhasya which polemicizes against the existence of a whole cloth as different from its parts. Vasubandhu mentions here the variegated colour in the following lines (p. 189 1. 24-26): "In case the threads have different colours ... the cloth could not have a colour .... If you accept] 'variegated' as its colour ... there would be production of (a colour) belonging to a different universal (from the colours in the threads)." (bhinnarupajatikriyesu tantusu patasya rupadyasarbhayat/ citrarupaditve vijafiyarambho 'pi syat). It must be admitted that the variegated colour in this passage from the Abhidharmakosa Bhasya is rather hypothetical, and does not prove beyond doubt that anyone known to Vasubandhu believed in it. Vyomasiva's commentary on the Padarthadharmasarngraha, called Vyomavati, cites a sutra in its discussion of the variegated colour. The

Page #3
BHARTRHARI AND VAISESIKA 77 sutra reads: "Because there cannot be, in one non-omnipresent substance, (several] specific qualities (visesaguna) that are perceived by the same sense-organ." It is not impossible that this sutra was indeed meant to justify variegated colour as a quality. Unfortunately it is only known through this passage of the Vyomavafi; it does not occur in the different versions of the Vaisesika Sutra that have been preserved, nor does it appear to be cited by any other commentator. For the most certain attestation of the variegated colour in early Vaisesika we have to turn to the following verses of Bhartrhari's Vakyapadiya: Just as the single variegated colour is described by way of different (colours) such as blue etc., which point to divisions in the one and indivisible variegated colour); in the same way the single sentence, which is completely self-sufficient, is described by way of other linguistic units (viz. words) which require one another. It is not necessary to recall here that for Bhartrhari the sentence is the real unit of language, the individual words being the result of an artificial analysis. The comparison with the variegated colour is therefore particularly appropriate, for that colour too cannot be looked upon as a collection of constituent colours. The comparison further reminds us of the fact that Bhartrhari's observations on the sentence as an indivisible unit are of an ontological rather than linguistic or psychological nature. 2. Sound (1) Sound (sabda), in classical Vaisesika, is a quality of ether (akasa). It is already described as such in the Padarthadharmasangraha of Prasastapada, in Candramati's *Dasapadarthi, and in some of the Vaisesika sutras. The Padarthadharmasargraha gives the following description: fabdo 'mbaragunah frotragrahyah ksanikah.../sa dvividho varalaksano dhvani- (v.l. 'varna-) laksanas ca/ tatra akaradir varalaksanah Sarikhadinimitto dhvanilaksanas 3 Vy vol. 1, p. 63 1. 20: avibhuni drave samanendriyagrahyanam visesagunanam asambhavar. VP 2.8-9: citrasyaikasya rupasya yathabhedanidarsanaih/niladibhih samakhyanar kriyate bhinnalaksanaih// tathaivaikasya vakyasya ninakariksasya sarvatah/ Sabdantaraih samakhyanam sakariksair anugamyate // I thank W. Halbfass for some useful observations. Np. 287-88, ki p. 262, Vy vol. 2 p. 237. 5 6

Page #4
78 JOHANNES BRONKHORST ca (v.l. vamalaksanah)/ tatra vamalaksanasyotpattir atmamanasoh samyogat smrtyapeksad vamoccaraneccha/ tadanantaram prayatnah/ tam apeksamanad atmavayusamyogad vayau karma jayate/ sa cordhvam gacchan kanthadin (v.l. urahkanthadin) abhihanti/ tatah sthanavayusamyogapeksamanat (v.l. -apeksat) sthanakasasamyogad varnotpattih/ Sound is a property of ether. It is perceptible by the ear. It is momentary.... It is of two kinds (1) in the form of speech sounds and (2) in the form of noise in general. [Sound] in the form of speech sounds is [the sounds] a etc. [Sound] in the form of noise in general is produced by the blowing of a conch and such things. Sound of the former kind proceeds from the contact of the mind and soul as influenced by remembrance: - First of all there is desire for pronouncing the sound; this is followed by an effort on the part of the speaker; and when this effort brings about the conjunction of the soul with wind, there is produced in this wind a certain motion; this wind moving upwards strikes such places as the throat and the like; this contact of the places of articulation and the wind brings about contact of the places of articulation with akasa; and this contact produces the speech sounds. (tr. Ganganatha Jha, modified) - This passage is quite clear that speech sounds are the sounds of the alphabet (a etc.) and are momentary. Words and phrases, on the other hand, are combinations, or rather sequences, of speech sounds. They present, therefore, a problem which resembles to some extent that of the variegated colour, discussed above. There are, however, important differences. Words and phrases are sequences of speech sounds, and cannot in any way be looked upon as collections of simultaneously existing sounds. Moreover, words and phrases do not, unlike the variegated colour, have a different substrate from their constituent sounds. It is further of some interest to note that the sutra cited in the Vyomavafi and discussed above, which was supposedly meant to justify the existence of the variegated colour, confines itself explicitly to non-omnipresent substances. It may here be recalled that ether, the substrate of sound, is omnipresent. How, then, did the Vaisesikas look upon words and phrases? No statements from early Vaisesika texts are known to me that attribute or deny ontological status to words and phrases. But some passages discuss the link between words and the things they denote. VS 7.2.19/19/19 sabdarthav asambaddhau claims that "words and designated objects have no connection", and VS 7.2.24/20/20 samayikah sabdad arthapratyayah adds that "the understanding of an object from a word is based on convention". It seems clear that the presence of an 'existent' link between words and things is here rejected. This, however, would seem to imply that the 'existence' of words is not in doubt. A similar position appears to be taken in Vyomasiva's Vyomavati. This commentary raises the problem that the definition of samavaya risks to

Page #5
BHARTRHARI AND VAISESIKA 79 cover the relation between a word and its designated object as well, as happens in the case of the word 'ether' and its designated object. The solution to the problem presented by Vyomasiva has no relevance to our question. The fact that he does not point out that the word akasa ('ether') does not exist' creates however the impression that Vyomasiva, too, accepts the existence of whole words. Yet in another passage he points out that we understand the meaning 'cow' when we hear the sequence g-au-h. Candrananda's commentary on Vaisesika Sutra 7.2.23 (which has no parallel in the other versions of this text) discusses the link that exists between sound and ether, and the one between ether and objects. The combined link which thus exists between a word and the object it denotes is not accepted, because it leaves a doubt as to which object is denoted by which word. But Candrananda, too, fails to point out that the designating words do not exist in the first place. He seems to have no difficulty accepting the existence of whole words. Turning now to the Vakyapadiya, we notice that Bhartrhari knows the conception of sound as a quality of ether. This we must conclude from a number of stanzas in the Sambandhasamuddesa, which discuss the relation, in Vaisesika terms, between words and objects. We find here, for example, the following statement (VP 3.3.16ab): svasrayena tu sariyuktaih samyuktar vibhu gamyate What is 'omnipresent' is known, being in contact with (objects that are in contact with its own substrate. We know from the Padarthadharmasangraha (Kip. 148 1. 16; N p. 141 1. 5; Vy vol. 2 p. 72 1. 19) that omnipresent objects have no mutual contact. We may therefore conclude that 'its own substrate' is omnipresent. But it seems certain that 'it' is sound (more precisely, the word vibhu 'omnipresent (object)'); it would be difficult to make sense of the surrounding stanzas without this assumption. This in its turn means that sound has as substrate an omnipresent substance, which can only be ether. From stanza 3.3.13 we learn, moreover, that the relation with the own substrate is 7 Vy vol. 1 p. 26 1. 13-17: tatha hy akasasabdenakasam abhidhiyata ity anayor adharyadharabhave sati vacyavacakabhavah .../ tadvyavacchedartham avadharanam adharyadhanabhutanam eva yah sambandhah sa samavaya iti. Vy vol. 2 p. 241 1. 23-25: yatra yatra gakaraukaravisarjaniyanam itthambhutanupurvim upalabhase, tatra tatra gorvavifisto'nthah pratipattavah pratipadayitavyas ceti sariketagrahe sati tathavidhar sabdam upalabhamanas tam arthar pratipadyate pratipadayati ceti. See also the discussion on p. 184.

Page #6
80 JOHANNES BRONKHORST samavaya (samavayat sva adharah ... prafiyate); that is to say, the relation between the word 'ether' and ether is samavaya. Sound is therefore clearly looked upon, in this passage of the Vakyapadiya, as a quality of ether. But this passage does more than this. Like the passages from Vyomasiva and Candranda discussed above, and like the Vaisesika Sutra itself, it does not appear to find fault with the idea that whole words (such as vibhu 'omnipresent (objects and akasa 'ether') are treated as 'existing entities, about the 'existence' of whose links with the denoted objects one can reasonably discuss. It may be worthwhile to recall at this point that there were thinkers in the age concerned who did not look upon words as entities in their own right. An example is found in the Sabara Bhasya, the classical commentary on the Mimarsa Sutra which is probably earlier than the Padarthadharmasarigraha." According to this text words are nothing but collections of speech sounds, which alone 'exist'. This point of view is introduced in the so-called Vrttikaragrantha on sutra 1.1.5, and attributed to someone called Upavarsa. 3. Sound (2) There is a further problem with sound in early Vaisesika. The Vaisesika sutra that enumerates all the qualities, no. 1.1.5, does not mention sound, nor several of the other qualities that figure in the classical list. Instead of the classical number of 24 qualities, it lists 17 of them. This smaller number is confirmed by the Jaina author Jinabhadra, in his Visesavasyakabhasya. 2 We are entitled to assume that the Vaisesika sutras that do mention or treat sound as a quality are later additions to the text. Their removal offers valuable insights into the earlier construction of the Vaisesika Sutra. 9 10 Bhartrhari does not even hesitate to speak about the universals residing in (whole) words, see Bronkhorst, 1991: 9. There are reasons to think that Bharthari did not yet know the Sabara Bhasya; see Bronkhorst, 1986, 1989. Frauwallner, 1968: 38: atha 'gaurity atra kah sabdah? gakaraukaravisarjaniya iti bhagavan upavarsah/ Compare this passage with the one from the Vyomavali cited above. See Halbfass, 1980: 285 n. 55; Wezler, 1983: 36 n. 5.

Page #7
BHARTRHARI AND VAISESIKA 81 These sutras occur in two groups, in Ahnikas 2.1 and 2.2 respectively. The first group follows sutra 2.1.20/19/20 which presents leaving and entering a place as the inferential mark of ether (niskramanam pravesanam ity akasasya lirigam)." This inferential mark is rejected in the then following group of sutras, and replaced by another inferential mark, sound, which is proved to be a quality of ether. This whole discussion - which covers sutras 21-26 in Jambuvijaya's edition, 20-30 in Thakur's, and 21-27 and 30-31 in Sinha's - is therefore added onto another, older inferential mark meant to prove the existence of ether. Since ether is enumerated as one of the substances in sutra 1.1.4, we may safely assume that in earliest - or at any rate, earlier - Vaisesika the motions of leaving and entering a place were deemed to prove the existence of the substance ether. This conclusion also teaches us to regard with suspicion any discussions that may occur within the body of the Vaisesika Sutra. Sound is again discussed in Ahnika 2.2. The context is, again, peculiar. Sutra 2.2.19/17/17 introduces the topic, which is doubt (sambaya); this topic continues until 2.2.23/21/20. Then the topic sound is introduced, in 2.2.24/22/21, and the following sutra 25 (it occurs only in Candrananda's version) makes clear that this topic is meant to illustrate a particular case of doubt: is sound a substance, an action, or a quality (tasmin dravyam karma guna iti sarsayah)? This illustration now steals the show completely, and is the sole topic of discussion - according to the commentators - until the end of the Ahnika. It seems clear that this long excursus on sound is an intrusion into the text, and that Adhyaya 2 originally ended with a discussion of 'doubt'. Vaisesika, then, underwent a change in its conception of sound. The new conception, according to which sound is a quality of ether, is already known to Nyaya Sutra 1.1.12-14. Caraka Samhita, Sutrasthana 1.49, moreover, enumerates the Vaisesika qualities in such a manner that it is clear that its author knew the expanded list: mention is made of guru etc., and we may conclude that sound, too, was considered a quality. All this suggests that the change took place at a rather early date. However, Jinabhadra's Visesavasyakabhasya states in so many words that the number of Vaisesika qualities is 17, as we have seen. This text 13 Compare this with "giving room" (avakasadana), mentioned as mark of ether in the (according to Ruben, spurious) Nyaya sutra gandhakledapakavyuhavakasadanebhyah paricabhautikam, see Ruben, 1928: 64. Cf. Adachi, 1990: 909 (35); Narain, 1976: 108f.

Page #8
JOHANNES BRONKHORST may have been composed in the year 609 C.E.,'' i.e., much later than Bhartrhari. It appears, therefore, that the earlier conception of sound existed for a long time side by side with the one that came to replace it. It is, for this reason, not impossible that it was still known to Bhartrhari. Before we deal with this question, we must address another one: what conception did early Vaisesika have of sound? Note first that it is not possible to assume that the author of sutra 1.1.5, which enumerates the qualities, simply overlooked sound. Such an argument may be possible in the case of 'heaviness' (gurutva), 'fluidity' (dravatva), viscidity' (sneha), and the other qualities (sarskara, dharma, adharma) that do not figure in the original list. Sound is too obviously a "thing - besides colour, taste, smell, and touch, all of them accepted as qualities in early Vaisesika - not to be given a place in the Vaisesika scheme of what there is. What then was sound? Given the Vaisesika ontological scheme, it must have been a substance, a quality, an action, a universal, a particular, or the special type of relationship which is called samavaya. It seems clear that, out of this list, sound can only belong to the categories substance or quality. Since quality is excluded, sound must then have been looked upon as a kind of substance. Which substance? Vaisesika enumerates nine substances, among them the five elements earth, water, fire, wind and ether. If we are forced to make a choice, wind (vayu) seems most appropriate. It seems therefore a priori not unlikely that for the early Vaisesikas sound was a form of wind. The link between sound and wind is obvious where speech sounds are concerned. The Padarthadharmasagraha explains how the movement of wind plays a crucial role in the production of speech sounds in the passage which we studied in the preceding section. A verse cited in the Vrtti on Bhartrhari's Vakyapadiya describes the same process in almost the same terms, with this difference that here wind itself is stated to become sound:16 15 16 Chatterjee, 1978: 109. Unfortunately I have had no access to the Visesavajyakabhasya. Ed. Iyer, Kanda 1, p. 173; included in Rau's edition as 1.111. Sabara's Bhasya on Mimarnsa sutra 1.1.22 ascribes to the Siksakaras the words: wayur apadyate sabdatam. (D'Sa, 1980: 79 n. 8, surprisingly, ascribes this position to the Vaisesikas, see however further below.) Sabara makes a further remark which may explain how sound could be conceived of as wind: vayaviyas cec chabdo bhaved vayoh sannivesavisesah syar "If sound were made of wind, it would be a special configuration of wind". The idea that sound is wind occurs elsewhere, too; p. Somananda's Sivadrsti 2.36: vaco vayvatmata na kim?

Page #9
BHARTRHARI AND VAISESIKA 83 labdhakriyah prayatnena vaktur icchanuvartina/ sthanesy abhihato vayuh sabdarvari pratipadyate// The wind set in motion by the effort corresponding to the desire of the speaker, strikes at the different places of articulation and is transformed into sounds (tr. Iyer, modified) The parallelism between these two passages, combined with the fact that iccha (desire) and prayatna (effort) are qualities of the soul in the Vaisesika scheme of things, suggests that the quoted stanza in the Vrtti draws upon Vaisesika ontology, and may even express a Vaisesika point of view. This point of view, however, is that wind becomes sound, in other words, that sound is wind. Consider now the following passage of the Padarthadharmasargraha. It reads, in the translation of Ganganatha Jha (p. 129):17 Sound cannot be the property of those substances that can be touched - (1) because, being perceptible, its production is not preceded by any quality in the material cause of the substance (to which it belongs); (2) because it does not pervade over, and is not coeval with, the substance to which it belongs; (3) because it is perceived elsewhere than in the substratum wherein it is produced. It cannot be regarded as belonging to the soul, (1) because it is perceptible by an external senseorgan; (2) because it is perceived by other souls; (3) because it is not found to inhere in the soul; and (4) because it is perceived as apart from all idea of 'I'. It cannot be the quality of space (dis), time and mind, (1) because it is perceptible by the ear, and (2) because it is a visesaguna (a specific quality). And thus the only substance to which it could belong as a quality, and be a distinguishing feature of, is akasa. As the distinguishing feature of sound is common to all akasa, this is regarded as one only. From this unity follows its individual separateness or isolation. akasa being spoken of as vibhu (omnipresent or all-pervading), it points to its dimension being the largest or highest. In as much as akasa is spoken of as the cause of sound, it follows that it has conjunction and disjunction. Every sentence in this passage reflects a Vaisesika sutra;18 this is not 17 Pdhs Ki p. 71-74, N p. 58, Vy vol. 1 p. 108: sabdah pratyaksarve saty akaranagunapurvakatvad ayavaddravyabhavitvad asrayad anyatropalabdhes ca na sparsavadvisesagunah/ bahyendriyapratyaksarvadatmantaragrahyatvadatmany asamavayad aharikarena vibhaktagrahanac ca natmagunah/ (frotragrahyatvad vaibesikagunabhavac ca na dikkalamanasam/ parisesad guno bhutva akatasyadhigame lirigam Sabdalirgavisesad ekatvarn siddham/ tadani vidhanadekaprthaktvam/vibhavavacanat paramamahat parimanam/sabdakaranatvavacanat sarnyogavibhagav iti/. The part in brackets has been omitied in Ki, no doubt by mistake. Compare with the preceding note the following sutras (2.1.24-26) found in Jambuvijaya's edition: karanagunapurvah karye guno drstah, karyantarapradurbhavac ca sabdah sparsavatam agunah/ paratra samavayat pratyaksatvac ca natmaguno na manogunah/ lirigam akasasya). Corresponding sutras are found in the other two versions of the text. 18

Page #10
JOHANNES BRONKHORST however indicated. Yet it is Prasastapada's habit to give an indication to that effect when he refers to a sutra. The final portion of our passage illustrates this. Consider the sentence "akasa being spoken of as vibhu, it points to its dimension being the largest or highest". This refers to sutra 7.1.28/24/22: vibhavan mahan akasah "because of its omnipresence ether (akasa) is large". The sutra contains an argument (if perhaps a bad one), which Prasastapada could have simply repeated. Instead he invokes the authority of the sutra. This only makes sense on the assumption that Prasastapada prefers referring to a sutra to repeating its contents on his own authority. This assumption, if correct, has far-reaching consequences. It implies that all the other sutras whose contents are repeated in this passage, were not yet recognized as such by Prasastapada. In other words, some of the Vaiaesika sutras which describe sound as a quality were not yet considered sutras by Prasastapada. Others, to be sure, were. The Padarthadharmasangraha (Kip. 235 1. 1-3; Np. 239 1. 14-16; Vy vol. 2 p. 200 1. 14-15) cites VS 2.2.26/x/22 from a 'sastra', most probably from the Vaisesika Sutra. Moreover, Prasastapada expresses in no uncertain terms that he looks upon sound as a quality. Our passage refers explicitly to two sutras. The first one has already been discussed. The second one cannot but be 2.2.36/30/31: saryogad vibhagac chabdac ca sabdanispatteh / -nispattih "sound originates from conjunction, from disjunction, and from (other) sound". We may assume that this second sutra was accepted as such by Prasastapada, and may therefore be older than at least some of the sutras which describe sound as a quality. With this in mind we turn to Bhartrhari's Vakyapadiya. Consider Vakyapadiya 1.105: yah sarnyogavibhagabhyan karanair upajanyate/ sa sphotah Sabdajah Sabda dhyanayo 'nyair udahntah// Others declare that the sphota is what is produced by the organs (of speech] by means of contact and separation; the sounds born from (this initial] sound are the dhvanis. Note the similarity of this verse, at least of certain parts of it, with the Vaisesika sutra (2.2.36/30/31) which appears to be old. It seems likely that The version edited by Sinha contains some additional elements in the sutras 2.1.27, 30-31: parisesal lingam akasasya/ Sabdalirgavifesad visesalirgabhavac ca/tadanuvidhanad ekaprthaktvam ceti. Sutra 28 tattvarn bhavena refers back to 1.2.18 sallirgavisesad visesalirgabhavac caiko bhava iti.

Page #11
BHARTRHARI AND VAISESIKA 85 Bhartrhari had this sutra in mind when he wrote the verse. Our next question must be: how is this verse to be understood?" To begin with, note that this verse describes the opinion of 'others'. It offers, by doing so, an alternative to the opinion of 'some', which is presented in the preceding verses. The first of these preceding verses is nr. 1.96, which reads:20 Some consider that the sphota is the universal revealed by the various individual instances, and they consider that the individuals belonging to this universal) are the dhvanis. We see that two alternatives are placed side by side. According to the first alternative, the sphota - that is the real, eternal word - is a universal; according to the second one the sphota is produced by the speech organs. But what is produced by the speech organs? Several factors combine to show that the sphota, on the second alternative, is some form of substance (dravya). Recall, to begin with, that for Bhartrhari the world has two sides: the one real and eternal, the other unreal and non-eternal. Regarding the real, eternal side, Bhartrhari does not care much what we call it. Some consider the eternal aspect of an object to be its universal, others its substance. We are free to choose, as long as we agree that every object has an eternal aspect. The second verse of the Jatisamuddesa (3.2) states therefore:21 In the analysis of objects denoted by words, the eternal objects denoted by all words have been described as 'universal' or as 'substance'. The remainder of the Jatisamuddesa occupies itself with the alternative that the eternal part of all objects is its universal; the then following Dravyasamuddesa takes up the alternative view that substance constitutes their eternal part. What is true for all things', is true for words, too. The real, eternal part of words is either a universal or a substance; both views are acceptable. The conclusion cannot but be that the verse (1.105), which appears 19 20 See in this connection also Bronkhorst, 1991: 14. VP 1.96: anekavyaktyabhivyarigva jatih sphota iti smrta/ kaiscid vyaktaya evasya dhvanitvena prakalpitah// VP 3.1.2: padarthanam apoddhare jatir va dravyam eva va/ padarthau sarvasabdana nityav evopavarnitau// 21

Page #12
86 JOHANNES BRONKHORST to draw its inspiration from the Vaisesika sutra, concerns the sphota as substance. This conclusion is confirmed by verse 1.110, which is one of the verses that elaborate the notions introduced in 1.105. This verse states how different schools of thought conceive of sound:22 Some accept that sound is wind, (others) that it is atoms, (others again) that it is knowledge, for in the presentations the different points of view are endless. The identification sound = knowledge looks puzzling at first. But obviously any idealistic school of thought will maintain that substance derives its reality from, is nothing but, thought or knowledge. In fact, Bharthari himself says so in a passage of his commentary on the Mahabhasya. We may conclude that the three points of view according to which sound is wind, atoms, and knowledge respectively, share in common that sound is substance. We have seen that the view in which sound is knowledge must belong to an idealistic school of thought. The view that sound is atoms is part of the world-view of the Sarvastivadins and the Jainas.24 Remains the view according to which sound is wind. The fact that Bhartrhari appears to refer to a Vaisesika sutra in this very passage, suggests that this view belonged to the early Vaibesikas. It would seem, then, that Bhartrhari knew indeed both the positions of Vaisesika with regard to sound: the more recent one according to which it is a quality of ether, and the older one according to which it is wind. One final observation. Bhartrhari may not yet have known Sabara's Bhasya on the Mimamsa Sutra. He may therefore be earlier than Sabara, or roughly contemporaneous with him. For Sabara, sound is eternal and resides in the omnipresent ether. The fleeting sounds we hear are manifested, and not produced by the speaker who utters them. In this context Sabara adds the following intriguing remark:26 "But for him who 22 VP 1.110: vajor anuna iranasya sabdatvapattir isyate/ kaiscid darsanabhedo hi pravadesv anavasthitah// CE I P. 22. I. 19-20, AL p. 27 1. 4-5, Sw p. 32 1. 11-13: 'dravyan hi nityam'/ nityah prthividhatuh/ prthividhatau kim satyam/ vikalpah/ vikalpe kim satyam/jnanam/... For a description of the way in which, according to the Sarvastivadins, sound joins other atoms in order to form a molecule, see La Vallee Poussin, 1980: I: 144-145. For the position of the Jainas, see Jaini, 1920: 118 (Tattvartha Sutra 5.24). See note 10, above. Sabara on sutra 1.1.13 (p. 93 1. 5-6): yasya punah kurvanti tasya vayaviyah sarnyogavibhaga vayvasritatvad vayusv eva karisyanti, yatha tantavas tantus eva param. 26

Page #13
BHARTRHARI AND VAISESIKA 87 believes that contacts and separations produce [sounds rather than manifest them], contacts and separations, which occur in wind, will produce (sounds) nowhere else than in wind, because they subsist in wind; just as yarns produce a cloth in the yarns themselves." Sabara's ideas are frequently close to those of the Vaisesikas, so that it is possible to believe that he had the Vaisesikas in mind while writing this passage. For the Vaisesikas do indeed believe that sounds are produced, not manifested. It is therefore possible that Sabara, too, still knew of Vaisesikas who believed that sound is wind. 4. The omnipresent soul? VS 5.2.18-20 read, in Candrananda's version: 5.2.18: kayakarmanatmakarma vakhyatam 5.2.19: apasarpanam upasarpanam asitapitasaryogah karyantanasamyogas cety adrstakaritani 5.2.20: tadabhave saryogabhavo 'pradurbhavah sa moksah The first of these sutras has a different form in the Vyakhya edited by Thakur: kayakarmanatmakarmadharmayor anupapattih (5.2.16); as observed by A. Wezler (1982: 659), it is difficult to make satisfactory sense of this reading. Nothing corresponding to this sutra is found in the version known to Sankara Misra. The remaining two sutras, on the other hand, occur in the other versions with only insignificant variations.26 Candrananda's explanation of these sutras contains some suspect features, most notably the following: 1) Candrananda interprets atman in 5.2.18 to mean wind (vayu). 2) In his interpretation 5.2.19 is about the manas. A straightforward interpretation of the sutras would rather suggest that 5.2.18 talks about the acitivity of the soul (atmakarman), and that 5.2.19 continues this topic and therefore talks about the soul too.29 Regarding 5.2.19 we know that already Prasastapada interpreted it like Candrananda: his Padarthadharmasangraha refers to this sutra in the 27 Some of the questions here discussed have also been dealt with in a paper called "Mysticisme et rationalite en Inde: le cas du Vaisesika", to be published in the Proceedings of the Colloque 'Mystique et rationalite: Inde, Chine, Japon' (Geneve, November 29-30, 1990), Asiatische Studien/Etudes Asiatiques 47 (4), 1993. ..pitasaryogah instead of pitasarnyogah; "pradurbhavas ca instead of 'pradurbhavah. 29 See Wezler, 1982: 654f.

Page #14
JOHANNES BRONKHORST context of the description of the manas, in order to show that the manas can have samyoga and vibhaga.30 Note in passing that Prasastapada's remark shows that he looked upon VS 5.2.19 as a sutra. Elsewhere in the Padarthadharmasangraha it is pointed out that the two activities of the manas called apasarpana and upasarpana are the result of contact between the soul and the manas, which depends on adrsta;31 again the sutra is interpreted as referring to the manas. 88 There is evidence to show that the above sutras at one time concerned the atman and its activities. Consider first VS 6.2.19/18/16: atmakarmasu mokso vyakhyatah. Wezler (1982: 654) observed already that this sutra "obviously refers back to VS 5.2.20". It does, however, more than just this: it suggests strongly that 5.2.20, and therefore 5.2.19 as well, concern atmakarman 'the activity of the soul'. With this in mind we turn to Bhartrhari's Vakyapadiya. The Sambandhasamuddesa of this work explores the question what connection exists - to be described in Vaisesika terms, i.e., combinations of samyoga and samavaya - between a word and the object it designates. This leads to no satisfactory results. Indeed, VP 3.3.17 points out that this approach would not limit the designation of a word to its appropriate object. Here VP 3.3.18 counters: adrstavrttilabhena yatha samyoga atmanah/ kvacit svasvamiyogakhyo 'bhede 'nyatrapi sa kramah// Just as the samyoga of the soul is [only] called 'connection of owner and owned' with regard to certain objects, because adrsta operates [in these cases], even though there is no difference [between this special kind of samyoga and samyoga in general], just so is the situation in the case of other [relations], too. This verse cannot but mean that a virtually limitless number of samyogas of the soul is limited by the operation of adrsta to those few which link the soul to 'its' body, etc. This in its turn implies that, in Bhartrhari's opinion, the soul of the Vaisesikas is in contact with far more objects than just its 'own' body etc., and therefore most probably infinitely large, as it is in classical Vaisesika. 30 Ki p. 101 1. 16-17; N p. 89 l. 15-16; Vy vol. 1 p. 156 1. 14-15: apasarpanopasarpanavacanat samyogavibhagau. 31 Ki p. 270 1. 12; N p. 308 1. 22-23; Vy vol. 2 p. 266 1. 6: apasarpanakarmopasarpanakarma catmamanahsamyogad adrstapeksat.

Page #15
BHARTRHARI AND VAISESIKA 89 The explicit mention of adrsta in this verse leaves little doubt that Bhartrhari paraphrases here VS 5.2.19, which he apparently considered to concern the soul. If we now try to translate VS 5.2.18-20 in agreement with the interpretation which Bhartrhari to all appearances accorded them, we get: 5.2.18: The activity of the soul is explained by the activity of the body 5.2.19: Retreating, approaching, contact with what is eaten and drunk, contacts with other effects, (these functions of the soul] are caused by adrsta. 5.2.20: When there is no (activity of the soul], there is no contact (with objects that belong to it), no manifestation (of the soul in a body); that is moksa. In view of VP 3.3.18, the contacts of the soul referred to in VS 5.2.19-20 pertain to the subgroup of saryogas called 'contact between owner and owned'. Contacts in general exist between each soul and every finite object, the soul being omnipresent; but these general contacts are not relevant in the context of 'activity of the soul'. This 'activity of the soul', too, must be interpreted to bring about the special contacts called 'contact between owner and owned'. These special contacts are confined to the body, and so is therefore this 'activity of the soul'. It goes without saying that this limitation of the contacts of an infinitely large soul to a restricted number of objects is hard to explain in terms of the Vaisesika categories. Adrsta is meant to explain, or cover up, this mystery, and would not seem to have much to do with dharma and adharma, which constitute adrsta in the classical system. Indeed, if adrsta in VP 3.3.18 meant dharma and adharma, also the connection between words and their meanings should be determined by dharma and adharma, a point of view which Helaraja rects as impossible. At this point we must pay attention to a passage of the Nyaya Sutra, along with Paksilasvamin's Bhasya (3.2.61-73 (Ananda Asrama ed.)/60-72 (tr. Jha)/ 59-71 (ed. Ruben)). This passage deals with the formation of the body and with the factors that play a role in it. Sutra 61/60/59 gives the opinion of the author: "Its formation is due to the persistence of previous acts" (purvakrtaphalanubandhat tadutpattih). Sutra 67/66/65 explains further: "Just as karman is the cause of the formation of the body, so is it also of the connection (of the body with a particular soul]" (sarirotpattinimittavat samyogotpattinimittam karma; tr. Jha). Sutras 6973/68-72/67-71 now reject an alternative opinion regarding the formation 32 My attention was drawn to this passage by J.E.M. Houben.

Page #16
90 JOHANNES BRONKHORST of the body and its connection with its soul (?). The first two of these sutras are of most interest to us: 69/68/67: tad adrstakaritam iti cet punas tatprasargo 'pavarge If (it be asserted] that the (formation of the body) is due to adrsta, then (our answer is that in that case) even after final release there would be likelihood of (a body being produced). (tr. Jha) 70/71/72: manah (v.l. manasah) karmanimittarvac ca saryoganucchedah There would be no severance of connection - this being due to the action of mind. (tr. Jha) The then following sutras go on to show absurd consequences of the rejected opinion, but the above two are most important, for they allow us to identify the rejected opinion as that of VS 5.2.19-20, studied above. The link with VS 5.2.19 is again emphasized by the Nyaya Bhasya on NS 3.2.70/69/68, which raises questions regarding the apasarpana and upasarpana of the manas, using exactly the terms also found in VS 5.2.19. Two observations must be made here. The first concerns the interpretation of VS 5.2.19 offered in these Nyaya sutras. NS 3.2.70/69/68 speaks of the activity of the mind (manahkarman), and this is apparently how it interpreted the Vaisesika sutra - just like Prasastapada and Candrananda, as we have seen. The second observation pertains to the meaning of adrsta in NS 3.2.69/68/67. Adrsta cannot here be identical with karman, nor even be the fruit of previous acts (purvakrtaphala), because these are presented as the true causes of the formation of the body in sutras 61/60/59 and 67/66/65 (see above). The author of the Nyaya Bhasya understood this very well: he offers two interpretations of adrsta, neither of which appears to have much to do with karman and its effects. It appears, then, that NS 3.2.69-70/68-69/67-68 directly criticise VS 5.2.18-20, which they interpret in a way that deviates from the original interpretation. The meaning assigned to adrsta, on the other hand, is still pre-classical. The criticism centres in a way on VS 5.2.20, which describes moksa, liberation. According to these Nyaya sutras, liberation would not be possible if VS 5.2.19 were correct. It is difficult to avoid the conclusion that the Nyaya sutras just considered must be later than VS 5.2.18-20, so much later that the original interpretation of VS 5.2.19 was no longer known, or used. This is all the more surprising since Bhartrhari, as we have seen, still knew the original interpretation of VS 5.2.19. Is it possible that the section NS 3.2.61-73/ 6072/59-71 was added later?

Page #17
BHARTRHARI AND VAISESIKA 91 This is indeed likely, for this group of sutras constitutes an excursion which interrupts the regular order of topics, as was already noted by Ruben (1928: 209 n. 237). It may here further be observed that tad- in tadutpattih (NS 3.2.61/60/59) supposedly refers to the body; but the body is not mentioned in the preceding sutras! We may safely conclude that the whole group of sutras constitutes a later addition to the text. The results of the above investigation can be presented as follows. A verse of Bhartrhari's Vakyapadiya allows us to catch a glimpse of the earliest interpretation of VS 5.2.18-20. It shows us that then already the Vaisesikas looked upon the soul as infinitely large. This did not prevent them from speaking about the activity or movement of the soul (atmakarman). The soul can be active because besides the general contact (saryoga) which it has with every finite object, it can have a specific contact - described as 'contact of owner and owned' - with a restricted number of objects, primarily the 'own' body and all that is contained in it. The 'activity' of the soul that brings about, or maintains, these special contacts, coincides therefore normally with the movement of the body. In terms of the Vaisesika system there is something very mysterious about these special contacts; this is why they are stated to be occasioned by adrsta, the unseen. These special contacts, as well as the 'activity' ascribed in this way to the soul, can come to an end; the soul does then no longer manifest itself in a body. This state of the soul is called 'liberation' (moksa). The implausibility of this way of speaking about the 'activity' of a none-the-less omnipresent soul is obvious. It does not surprise that the idea was discarded. But discarding an idea proved easier than discarding the sutras which expressed it. This led to a reinterpretation of the sutras concerned. We find the first evidence of this in a set of sutras inserted at an unknown date into the Nyaya Sutra. This set criticizes VS 5.2.18-20, but while doing so it shows that the idea of an activity of the self had been given up. Contact between the soul and 'its' body are now ascribed to the activity of the mind (manas), which corresponds to the later, classical doctrine. But the interpretation of the term adrsta had not yet reached its classical form. Adrsta is not yet short-hand for dharma and adharma, which are the effects of karman; adrsta is, on the contrary, contrasted with karman and its effects. It will be clear that with the interposition of a manas between a soul and its body, the mystery of the special relationship between the soul and its body disappears, and that, consequently, preclassical adrsta has no more role to play in it. No wonder that our set of Nyaya sutras attacks this notion.

Page #18
JOHANNES BRONKHORST With Prasastapada we arrive at the classical exposition of the Vaisesika system, and apparently also at the classical interpretation of VS 5.2.18-20. It is clear that Prasastapada knows at least VS 5.2.19 and considers it a sutra. He believes, furthermore, that it concerns the manas. But also adrsta has with Prasastapada reached its classical meaning: it has become more or less identical with the effect of karman (purvakrtaphala). Prasastapada does no longer have to attack the notion of adrsta; the new interpretation of this term allows him to agree with the author of the above set of Nyaya sutras, while yet accepting the Vaisesika Sutra as authoritative. Abbreviations N Mahabhasyadipika of Bhartrhari, ed. Abhyankar-Limaye "Critical edition of Bhartrhari's Mahabhasyadipika Padarthadharmasarigraha of Prasastapada, ed. Jetly Padarthadharmasarigraha of Prasastapada, ed. Dvivedin Mahabhasyadipika of Bharthari, ed. Swaminathan Vakyapadiya of Bharthari, ed. Rau Padanthadharmasarigraha of Prasastapada, ed. Gaurinath Sastri Sw VP Vy References Adachi Toshihide (1990): "The acceptance of the Vaisesika padartha theory in the Carakasamhita." Journal of Indian and Buddhist Studies 38 (2), 911-907. Bhartrhari: Mahabhasyadipika. 1) Edited by K.V. Abhyankar and V.P. Limaye. Poona: Bhandarkar Oriental Research Institute. 1970.(Post-Graduate and Research Department Series No. 8.) 2) Partly edited by V. Swaminathan under the title Mahabhasya Tika. Varanasi: Banaras Hindu University. 1965. (Hindu Vishvavidyalaya Nepal Rajya Sanskrit Series Vol. 11.) 3) Manuscript reproduced. Poona: Bhandarkar Oriental Research Institute. 1980. 4) Critical edition', Ahnika 1. Edited and translated by Johannes Bronkhorst. Poona: Bhandarkar Oriental Research Institute. 1987. Bharthari: Vakyapadiya. (i) Critical edition by Wilhelm Rau. Wiesbaden: Franz Sieiner. 1977. (Abhandlungen fur die Kunde des Morgenlandes XLII,4) (i) Kanda I, edited, with the Vrtti and Vrsabhadeva's Paddhati, by KA. Subramania Iyer. Poona: Deccan College. 1966. (Deccan College Monograph Series, 32.) Bronkhorst, Johannes (1986): "Tantra and prasanga." Aligarh Journal of Oriental Studies 3 (2), 77-80. Bronkhorst, Johannes (1989): "Studies on Bharthari, 2: Bharthari and Mimarnsa." Studien zur Indologie und Iranistik 15, 101-117.

Page #19
BHARTRHARI AND VAISESIKA 93 Bronkhorst, Johannes (1991): "Studies on Bhartrhari, 3: Bhartrhari on sphota and universals." Asiatische Studien / Etudes Asiatiques 45, 5-18. Bronkhorst, Johannes (forthcoming): "The Vaisesika vakya and bhasya." Annals of the Bhandarkar Oriental Research Institute. Chatterjee, Asim Kumar (1978): A Comprehensive History of Jainism (up to 1000 A.D.). Calcutta: Firma KLM. D'Sa, Francis X. (1980): Sabdapramanyam in Sabana and Kumarila. Towards a study of the Mimarsa experience of language. Vienna. (Publications of the De Nobili Research Library, 7.) Frauwallner, Erich (1968): Materialien zur altesten Erkenntnislehre der Karmamimarsa. Wien: Hermann Bohlaus. (Osterreichische Akademie der Wissenschaften, philosophischhistorische Klasse, Sitzungsberichte, 259. Band, 2. Abhandlung. Veroffentlichungen der Kommission fur Sprachen und Kulturen Sud- und Ostasiens, Heft 6.) Grohma, Otto (1975): "Theorie zur bunten Farbe im Alteren Nyaya und Vaisesika bis Udayana." Wiener Zeitschrift fur die Kunde Sudasiens 19, 147-182. Halbfass, Wilhelm (1980): "Karma, apurva, and 'natural' causes: observations on the growth and limits of the theory of samsara." Karma and Rebirth in Classical Indian Traditions. Edited by Wendy Doniger O'Flaherty. Berkeley - Los Angeles - London: University of California Press. Pp. 268-302. 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.) Hattori, Masaaki (1968a): "Two types of non-qualicative perception." Wiener Zeitschrift fur die Kunde Sudasiens 12-13 (1968/1969, Festschrift fur Erich Frauwallner), 161-169. Jaini, J.L. (1920): Tattvarthadhigama Sutra, edited with introduction, translation, notes and commentary. Reprint: AMS Press, New York. 1974. Jha, Ganganatha (tr.) (1915): Padarthadharmasarigraha of Prasastapada, with the Nyayakandali of Sridhara. Varanasi: Chaukhambha Orientalia. 1982. (Chaukhambha Oriental Studies, 4.) Jha, Ganganatha (tr.)(1939): Gautama's Nyayasutras, with Vatsyayana-Bhasya. Poona: Oriental Book Agency. (Poona Oriental Series, 59.) La Vallee Poussin, Louis de (tr. (1980): L'Abhidharmakosa de Vasubandhu. 5 tomes. Nouvelle edition anastatique presentee par Etienne Lamotte. Bruxelles: Institut Belge des Hautes Etudes Chinoises. (Melanges Chinois et Bouddhisques, XVI.) Narain, Harsh (1976): Evolution of the Nyaya-Vaieesika Categoriology. Vol I. Varanasi: Bharati Prakashan. Paksilasvamin: Nyaya Bhasya. Edited, with the Nyaya sutras and Visvanatha Bhattacarya's Vrtti, by Nagesa Digambara Sastri Josi. Poona: Ananda Asrama. 1922. (Anandasrama sanskrtagranthavali, 91.) Potter, Karl H. (1977): Encyclopedia of Indian Philosophies. II. Indian Metaphysics and Epistemology: The Tradition of Nyaya-Vaisesika up to Gangesa. Delhi: Motilal Banarsidass. Prasastapada: Padarthadharmasargraha (Prasastapadabhasya). 1) Edited, with the com mentary Nyayakandali of Sridhara, by Vindhyesvari Prasad Dvivedin. Second edition. Delhi: Sri Satguru Publications. 1984. 2) Edited, with the commentary Kiranavali of Udayanacarya, by Jitendra S. Jetly. Baroda: Oriental Institute. 1971. 3) Edited, with the

Page #20
94 ; JOHANNES BRONKHORST commentary Vyomavati of Vyomasivacarya, by Gaurinath Sastri. Varanasi: Sampur nanand Sanskrit Vishvavidyalaya. 2 vols. 1983-84. Ruben, Walter (ed., tr.X1928): Die Nyayasutra's. Text, Obersetzung, Erlauterung und Glossar. Leipzig. Genehmigter Nachdruck: Kraus Reprint, Nendeln, Liechtenstein. 1966. (Abhandlungen fur die Kunde des Morgenlandes, XVIII, 2.) Sabara: Mimamsabhasya. 1) Edition in: Mimarnsadarsanam: Tarkapada. Poona: Ananda srama. 1976. (Anandasramasanskrtavali, 97.) 2) Translation: Sabara-Bhasya, translated by Ganganatha Jha. Vol. I. Baroda: Oriental Institute. 1973. (Gaekwad's Oriental Series, 66.) Schmithausen, Lambert (1970): "Zur Lehre von der vorstellungsfreien Wahrnehmung bei Prasastapada." Wiener Zeitschrift fur die Kunde Sudasiens 14, 125-129. Somananda: Sivadrsti. Edited, with the Vrtti by Utpaladeva, by Madhusudan Kaul Shastri. Srinagar. 1934. (Kashmir Series of Texts and Studies, LIV.) Tillemans, Tom J.F. (1990): Materials for the Study of Aryadeva, Dharmapala and Candrakinti. The Catuhsataka of Aryadeva, chapters XII and XIII, with the commentaries of Dharmapala and Candrakinti: introduction, translation, Sanskrit, Tibetan and Chinese texts, notes. 2 vols. Wien: Arbeitskreis fur Tibetische und Buddhistische Studien, Universitat Wien. (Wiener Studien zur Tibetologie und Buddhismuskunde, 24,1 & 2.) Vaisesika Sutra. 1) Edited, with the commentary of Candrananda, by Muni Jambuvijaya. Baroda: Oriental Institute. 1961. (Gackwad's Oriental Series, 136.) 2) Edited, with an anonymous commentary, by Anantalal Thakur. Darbhanga: Mithila Institute. 1957. 3) Edited, with a translation of the sutras, of the commentary of Sankara Misra and of extracts from the gloss of Jayanarayana, by Nandalal Sinha. Delhi: S.N. Publications. 1986. The three numbers that specify a sutra refer to these three editions respectively. Where only one number is given, the reference is to the first of these three. Vasubandhu: Abhidharmakosa Bhasya. Edited by P. Pradhan. Second edition. Patna: K.P. Jayaswal Research Institute. 1975. Vrsabhadeva: Paddhati. See under 'Bhartrhari'. Vyomasiva: Vyomavali. Edited by Gaurinath Sastri. Varanasi: Sampurnanand Sanskrit University. 2 vols. 1983-84. Wezler, A. (1982): "Remarks on the definition of yoga' in the Vaisesikasutra." Indological and Buddhist Studies, Festschrift J.W. de Jong. Edited by L.A. Hercus et al. Canberra: Faculty of Asian Studies. Pp. 643-686. Wezler, A. (1983): "A note on concept adrsta as used in the Vaisesikasutra." Aruna-Bharali. Professor A.N. Jani Felicitation Volume. Edited by B. Datta et al. Baroda: Oriental Institute. Pp. 35-58. Yogabhasya = Patanjalayogadarsanam, edited by Narayana Misra. Varanasi: Bharatiya Vidya * Prakasana. 1971.