Studies On Bhartrhari 3

Author(s): Johannes Bronkhorst
Publisher: Johannes Bronkhorst
Catalog link: https://jainqq.org/explore/269605/1

Page #1
STUDIES ON BHARTRHARI, 3. Bhartrhari on sphota and universals. Johannes Bronkhorst, Universite de Lausanne 1. It will soon be 40 years since John Brough's influential article "Theories of general linguistics in the Sanskrit grammarians" appeared in print.2 Among the topics discussed is the sphota. Brough complains that "this term sphota, which is of prime importance for Indian linguistic theory, has unfortunately been subjected by modem writers to a great deal of unnecessary mystification" (p. 405). Two writers in particular are mentioned, A. Berriedale Keith and S.K. De. Keith had described the sphota as a mysterious entity, a sort of hypostatization of sound", while De had used the expression "a somewhat mystical conception". Brough concludes that "it is hardly to be wondered at if the westem reader, in the face of numerous comparable accounts, should come to the conclusion that the sphota-theory represents a departure from lucidity which, coming as it does from men whose professional task was the clear presentation of linguistic facts, is quite inexplicable" (p. 406). For Brough the sphota is "simply the linguistic sign in its aspect of meaning-bearer (Bedeutungstrager)" (p. 406), or "simply the word considered as a single meaningful symbol" (p. 409). "In this conception of the sphota," Brough continues, "it seems to me that there is nothing 'mysterious': it is merely an abstraction to assist us in the handling of our linguistic material". He concludes on p. 410: "It will thus be seen that the sphota-doctrine, so far from being something 'mysterious', is in fact of central importance for the theory of language-symbolism." It may be that Brough's observations are useful for general linguistics and linguistic philosophy. Indeed, this is what Brough had in mind, for he wrote this article - as he put it - "not merely as a matter of antiquarian curiosity, but because in their extraordinary linguistic and philosophic acumen these ancient authors are still, I believe, worthy of our respect" (p. 402). But whether or not 1 Thanks are due to the Rockefeller Foundation which enabled me, for a period of one month, to direct my undivided attention to Bharthari's Vakyapadiya, in the Villa Serbelloni, Bellagio, Italy. Studies in Bharthari 1 and 2 have appeared in Bulletin d'Etudes Indiennes 6 (1988). pp. 105-143, and Studien zur Indologie und Iranistik 15 (1989), pp. 101-117, respectively. 2 Transactions of the Philological Society, 1951, pp. 27-46. Reprinted in A Reader on the Sanskrit Grammarians, edited by J.F. Staal, MIT Press, Cambridge - Massachusetts and London - England, 1972, pp. 402-414. Page numbers refer to the reprint.

Page #2
JOHANNES BRONKHORST similar ideas are, or should be, present in modern linguistics, this has nothing to do with the question whether Brough's observations help us to understand the ancient Indian grammarians - and among them Bhartrhari in particular. Brough was aware that there may be more to the sphota than is clear from his above-quoted remarks. He mentions in passing the fact that the Indians themselves appear to have given 'ontological status' to this abstraction, and to have considered it as a sort of quasi-Kantian 'Wort-an-sich'' (p. 409). On another page he mentions the fact that on the basis of the sphota-theory there was erected a metaphysical superstructure" (p. 411). Towards the end of the article (p. 412) he even quotes one of the few stanzas of the Vakyapadiya which make a statement about the ontological status of the sphota; VP 1.96 (ed. Rau) says that 'according to some' the sphota is a jati 'universal'.? Brough rejects this view and claims that Bharthari's sphota was rather an individual. Nothing further is however said about the metaphysical superstructure'. We have to face the question whether we really understand Bharthari any better by knowing that one of his concepts corresponds to a modern linguistic one, without knowing how it fits in his 'metaphysical superstructure'. As long as the sphota is not satisfactorily accounted for within the context of Bharthari's theory, are we not correct in describing it as a 'mysterious entity' or as a 'somewhat mystical conception'? The main effect of Brough's article is that it creates in us a sense of familiarity with respect to the sphota, but familiarity is not the same as understanding. For example, Brough's exposition may make us receptive to the idea that the sentence is an undivided entity (cf. p. 412 f.); but this does not help us to understand why, for Bhartphari, also objects like pots are indivisible (VP 3.243). Nor can Brough's arguments explain why the whole of the Rigveda is considered a unity by Bhartrhari (VP 3.553). There can be no doubt that the transcultural assimilation of concepts can remove the feeling of strangeness, but this should not be confused with understanding. It may, on the contrary, in certain cases give rise to confusion. The comparison of Panini's grammar with modern linguistics, for example, - besides contributing greatly to the general appreciation of Panini - has tended to overlook, or even misinterpret, certain aspects of this grammar. Brough's stated aim to demystify the concept of sphota, therefore, appears to be an attractive slogan rather than an achieved goal. 2. The second publication I will consider is Bhartrhari and the Buddhists, An Essay in the Development of Fifth and Sixth Century Indian Thought, by 3 Brough translates 'class', but 'universal' seems more appropriate.

Page #3
BHARTRHARI ON SPHOTA AND UNIVERSALS Radhika Herzberger. Only a part of this book deals with Bharthari's ideas, and only some aspects of this part will here be discussed. Herzberger complains about the absence of an integrated portrait of Bhartrhari's thought, a portrait that would convey the essential links between his grammatical ideas and his metaphysical ones" (p. 10). This shows that she attaches more value to Bharthari's metaphysical ideas than Brough did. Yet she describes Brough's above-discussed article as a first step in the direction of a demystification of Bhartrhari's metaphysical ideas (id.). How does Herzberger do justice to these ideas? The basic question of her approach is: "What is the basis on which names are given to things?" (p. xvii-xviii, xxi). The main ideas which she attributes to Bhartrhari in this context can be briefly described as follows: Bharthari distinguishes two kinds of universals: thing-universals (arthajati) and word-universals (sabdajati). From among these two, word universals are by far the more important; indeed, "the status of the thing universals is ignored" (p. 37). Word-universals - Herzberger calls them sometimes simply 'universals', as in the last line of p. 20-, on the other hand, are "made up of three strands: a phonological strand, a syntactic strand and a semantic strand" (p. 21). The result is clear: "The speaker on the basis of the form of a word has immediate and unerring access to its meaning, its syntactic and phonological features" (p. 21). The semantic aspect of a word-universal makes the next step possible: word-universals participate in a hierarchical structure. "Thus the name simsapa has access through its universal simsaparvam (sic!) to vrksatvam (sic!) (treeness) which is located in the name vrksa" (p. 33). This hierarchy can be extended upward. A simsapa is a tree, and for that reason animate, etc. At the top of this hierarchy we find the Great Being, which is the Supreme Universal, and which is consequently designated by all words (p. 3536). In order to confront this scheme with the text of the Vakyapadiya, I lift out the following points: 1) There are two kinds of universals: thing-universals and word-universals. 2) Word-universals have a semantic aspect. 3) Word-universals participate in a hierarchical structure. 4) The top of this structure is constituted by the Great Being, which is the Supreme Universal. Let us now deal with these points one by one. 4 Dordrecht / Boston / Lancaster / Tokyo: D. Reidel. 1986. (Studies of Classical India, 8.)

Page #4
8 JOHANNES BRONKHORST 1) The first point is easily established, and is obviously correct. VP 3.6 is thus translated by Herzberger (p. 29):5 All words first of all express their own universal; thereafter this universal (lit. it) is thought to be superimposed upon the forms of universals of [external] things. 2) The second point is more problematic. The stanza which supposedly shows that word-universals have a semantic aspect, is VP 3.3: kesamcit sahacaryena jatih saktyupalaksanami khadiradisv asaktesu faktah pratinidhiyate!! Herzberger translates (p. 20):6 6 7 According to some, the universal indicates a capacity by way of accompaniment [of the individual]; when [a post made of] khadira lacks the capacity [to perform the function enjoined by the injunction] something which has the capacity is substituted. She concludes: "There does not seem to be, in view of this stanza, any reason for denying that universals belonging in words lack serhantic features." This conclusion shows and the word 'indicates' in the translation suggested it already - that in Herzberger's opinion this stanza is about word-universals. It isn't, but it is easy to see how Herzberger arrived at this incorrect opinion. It is the result of her incorrect understanding of the preceding stanza VP 3.2. She translates it correctly (p. 71):7 In the artificial analysis of meanings / objects of words, a universal or an individual have been described as the two really eternal objects/meanings of all words. This stanza obviously concerns things. Yet Herzberger concludes from it that "Bhartrhari preferred the two-fold division of words into individuals (dravya) and universals (jati)" (p. 71), as if a division of words rather than of meanings /objects of words were here under consideration. It is true that the preceding stanza VP 3.1 deals with the division of words, but Herzberger is clearly mistaken in thinking that "Bhartrhari had meant to subsume the former classification (of VP 3.1) within the latter, more embracing categories (of VP 3.2)" (p. 20). One does not subsume a classification of words within a classification of meanings/objects of words. 5 The reading accepted by Herzberger is (p. 28): sva jatih prathamam sabdaih sarvair evabhidhiyatel tato 'rthajatirupesu tadadhyaropakalpana!! The translation of the same stanza on p. 75 is slightly different. VP 3.2: padarthanam apoddhare jatir va dravyam eva val padarthau sarvasabdanam nityav evopavarnitau!!

Page #5
BHARTRHARI ON SPHOTA AND UNIVERSALS 9 In the translation of VP 3.3 we may replace the word 'indicates' with 'coimplies', a term which renders the Sanskrit upalaksana at least as well, and makes very good sense. In the injunction "Tie up the beast to a post of khadira," the word khadira refers to the universal of khadira wood and, by coimplication, to the capacity of khadira wood to perform its function. 3) In order to substantiate the hierarchical structure of word universals, Herzberger adduces several stanzas. Consider first her translation of VP 3.7-8 (p. 31, 85; what follows is really an amalgamation based on these two translations):8 Just as the essence (tattva), which is in the quality red, is designated in lacquer (kasaya) and, as a result of contact with the conjoint (samyogisannikarsa), is grasped even in garments; so also the universal, which is fixed in a word, as a result of the relation between word and object, brings about the effect of universals (jatikarya), when universals belonging in things are designated. Herzberger makes much of the phrase "the effect of universals' (jatikarya) in the second of these two stanzas. "The effect of universals'," she observes on p. 33, "derives from the hierarchical structure to which a universal located in a name has access. ... Thus the name simsapa has access through its universal, simsapatvam to vrksatvam (treeness) which is located in the name vrksa. The name, on the basis of its own universal, has negative access to the universal located in the name palasa." Later on the same page she sums up: "Thus the effect of universals' ... represents a theory of the analytic and antonymic content of names." These statements do not, of course, constitute evidence for the correctness of their contents, and Herzberger is aware of it. The evidence, as she indicates on p. 33, follows these statements, and it seems clear that VP 3.10 is adduced to fulfil this role. This stanza has to be read in combination with the one that precedes it, and I reproduce both of them as found in Rau's critical edition: VP 3.9: jatisabdaikasese sa jatinam jatir isyatel sabdajataya ity atra tajjatih sabdajatisull VP 3.10: ya sabdajatisabdesu sabdebhyo bhinnalaksanal jatih sa sabdajatitvam avyatiloramya vartatell The two stanzas deal with certain complications arising in connection with ekasesa - translated by Herzberger as 'Remaindering of One'. An ordinary ex 8 VP 3.7-8: yatha rakte gune tattvam kasaye vyapadisyatel samyogisamnikarsac ca vastradisv api srhyatell tatha sabdarthasambandhac chabde jatir avasthita) vyapadese 'rthajatinam jatikaryaya kalpatell

Page #6
10 JOHANNES BRONKHORST ample of ekasesa is vrksas ca vrksas ca vrksas ca vrksah, which justifies the use of one single word vrksah 'trees' to refer to three or more objects, with the help of a plural ending. The discussion in the Mahabhasya shows that this is possible because words do not only refer to individuals, but also to universals. If now we wish to form the plural jatayah 'universals', we need the universal of the individuals referred to by the word jati, i.e., the universal of universals. To justify the plural sabdajatayah 'word-universals', similarly, we need the universal located in word-universals. However, no universals inhere in universals. How, then, is the formation of the plurals jatayah and Sabdajatayah to be explained? The answer is provided by VP 3.8 (see above), which stipulates that there where thing-universals (in the plural) are designated (vyapadese 'rthajatinam), the corresponding word-universal ([s] abde jatir avasthita) brings about the effect of universals (jatikaryaya kalpate), i.e., justifies the plural. The role of the thing-universal - which in this particular case does not exist - is taken over by the word-universal. This is possible because of the link that unites words and things (sabdarthasambandhalt)) and therefore, indirectly, word-universals and thing-universals. This explanation is confirmed by VP 3.9-10, which can be translated as follows: In the case of ekasesa of the word jati (i.e., in the formation of the plural jatayah), we need that universal of universals (viz., the universal inhering in the word jati). In the case of the plural) sabdajatayah, the universal of that (word sabdajati) resides in the manner indicated in stanzas 7-8) in the word-universals (Sabdajati). The universal which [inheres] in the words fabdajati (and makes the plural sabdajatayah possible] is different from those words. [but) is nothing beyond a word-universal. In order to understand Herzberger's interpretation of these stanzas, we must know that she follows the reading found in Iyer's non-critical edition, which deviates from Rau's in the case of stanza 10. Iyer has here: ya sabdajatih sabdesu sabdebhyo bhinnalaksana jatis sa sabdajatitvam apy atikramya vartalell Herzberger translates the two stanzas as follows (p. 34, 90, cp. p. 93): It i.e. the higher word universal] is held to be the universal of flower, more specific) universals when the operation) Remaindering of One is performed for words which signify universals as: '(these are] word universals here'; the higher word universal (lit. Cf. Mahabhasya on P. 1.2.64 vt. 53 (ed. Kielhom vol. I p. 246 1. 14-15: na hy akrti. padarthikasya dravya na padartho dravyapadarthikasya vakrtir na padarthah ubhayor ubhayam padarthah

Page #7
BHARTRHARI ON SPHOTA AND UNIVERSALS 11 it) is located in word universals. That word universal which is located in (all) words, (but) which is different from the words (in which it is located), resides there having even crossed over its own) word universalness. Note that even this translation should leave no doubt that these stanzas concern the very special case of the plural of words which signify universals. It is hard to see how they could possibly be considered to justify a hierarchy of worduniversals. We must assume that Herzberger drew some inspiration from the obscurity of stanza 10. But we have already seen that this obscurity can be avoided by accepting the reading which is anyway to be preferred on the basis of Rau's critical comparison of the Mss., and which gives a perfectly satisfactory meaning. Note that the reading accepted by Herzberger, and indeed her own translation, are still far removed from the 'hierarchy of word-universals', which can only be read into them with great effort. Indeed, Herzberger seems to be aware of this, for she introduces her explanatory remarks with the words: "I read this stanza in the following manner" (p. 34). She then continues: "A universal has the capacity to cross over both its own substratum as well as the phonetic features associated with it. Thus vrksatva loses its phonetic marks when it lodges in simsapa. Simsapa has the sense of vrksa, but not its phonetic features." A farfetched interpretation indeed! 4) VP 3.33 is quoted in order to show that "the hierarchy (of universals] reaches all the way up to the Supreme Universal, (mahasamanya), the Great Plenum in which all words are properly fixed" (p. 35). Herzberger translates it as follows:10 Divided into cows and so forth through distinctions present in those things which are its relata, [this] Being is called the Supreme) Universal; and all words are fixed in this Universal Again it is difficult to find support for Herzberger's point of view in this stanza. The only hint in that direction which I find in the translation is the word 'Supreme'. But this word is rightly put between hooks, for no word in the Sanskrit text corresponds to it; the addition of Supreme' is clearly an invention of the translator. 11 I shall not here discuss the question in how far Herzberger's interpretation - which does not fit the text of the Vakyapadiya, as we have seen - represents 10 VP 3.33: sambandhibhedat sattaiva bhidyamana gavadisul jatir ity ucyale tasyam sarve sabda vyavasthital 11 Note that Brough, too, made a similar addition while translating this stanza; he has "the Class (par excellence)". See however below.

Page #8
12 JOHANNES BRONKHORST Helaraja's views. Herzberger herself expresses her 'surprise at the discovery that Helaraja deviates from her interpretation at a crucial junction (p. 54). Nor does her interpretation of Bhartrhari find much support in her understanding of Dignaga, which - as she frankly admits on p. xxiii - "is shaped largely by my reading of Bhartrhari". 12 The conclusion is inevitable that Herzberger has not succeeded in her courageous attempt to elucidate Bhartrhari's ideas on the subject of universals. 3. Both Brough and Herzberger worked from below' 'upward' in their attempt to understand Bhartrhari's thought. Brough never reaches the metaphysical 'superstructure', whose existence he none-the-less does not deny. For Herzberger the 'superstructure is the 'top' of a construction built by her 'from below'. For Bhartrhari, however, we can be sure that the metaphysical superstructure did not come at the end, but rather at the beginning. It comes at the beginning literally, for the first stanzas of the Vakyapadiya speak of Brahman. But it must have come at the beginning in another sense as well: Bharthari wrote his work starting from a vision, in which the metaphysical aspects of his thought were already clearly represented. This at any rate seems an extremely reasonable assumption to make. Let us therefore try to understand Bharthari's ideas - at least in as far as they concern the sphota and universals - 'from top to bottom'. We begin with a stanza discussed by both Brough and Herzberger, VP 3.33, which we shall study in its context:13 From among the real and the unreal parts which are present in each thing, the real (part) is the universal, while the individuals are traditionally said to be unreals. (32) Being itself, when divided into cows etc. on account of the different things with which it is connected, is called 'universal'; all words are based on it. (33) They call it the meaning of the nominal stem and the meaning of the verbal rool. It is eternal, it is the great atman; (the abstract suffixes) tva, tal etc. refer to it. (34) 12 See also p. 106: "Dignaga wrote against the assumed background of Bhartihari's thought, and without an awareness of this background Dignaga's laconic statements remain obscure and odd." 13 VP 3.32-39: satyasatyau tu yau bhagau pratibhavan vyavasthitam satyam yat tatra sa jatir asatya vyaktayah smrta // sambandhibhedat sattaiva bhidyamana gavadisul jatir ity ucyate tasyam sarve sabda vyavasthita I tam pratipadikartham ca dhatvartham ca pracaksatel sa nitya sa mahan atma tam ahus tvataladayah/l praptakrama visesesu kriya saivabhidhiyatel kramarupasya samhare tat sattvam iti kathyatell saiva bhavavikaresu sad avasthah prapadyatel kramena saktibhih svabhir evam pratyavabhasatell atmabhutah kramo 'py asya yatredam kaladarsanam/ paurvaparyadirupena pravibhaktam iva sthitam/l tirobhavabhyupagame bhavanam saiva nastita labdhakrame tirobhave nasyatiti pratiyatell purvasma! pracyuta dharmad aprapta cottaram padam/ tadantarale bhedanam asrayaj janma kathy atell

Page #9
BHARTRHARI ON SPHOTA AND UNIVERSALS 13 When it assumes sequence in individual cases, it is called 'activity: when its sequential form is destroyed, it is called 'Being' (sattva). (35) It reaches the six states in the transformations of things, in order; on account of its own powers it appears like that. (36). Also sequence belongs to it. In it there is the experience of the power called] 'time, divided as it were in earlier, later, and so on. (37) It is the (posterior) non-existence of things, when we agree that they have disappeared; when the disappearance is in progress, it is known in the form it is being destroyed'. (38) It is called 'birth' when it has left its earlier characteristic and has not (yet) reached its next position, because in the meantime it is the basis of different [forms of appear ance). (39) These stanzas undoubtedly describe Bharthari's absolute, which he sometimes calls 'Brahman'. In the next article of the present series I intend to argue that this absolute is conceived of as a whole, as the totality of all there is, was, and will be. The present stanzas support this interpretation. Stanza 33, for example, speaks of Being which is divided into cows etc. The stanzas also refer to the "powers' of Brahman, which play a role in producing the unreal world of our experience. Reality, on the other hand, only belongs to Brahman. For further details of Bharthari's vision of the world I must refer to future articles in the present series. Here we must concentrate on universals. Stanza 33 identifies Being - i.e., Brahman - and 'universal'. Does Bhartrhari have here some kind of 'supreme universal' in mind, as Brough and Herzberger maintain? Nothing in the stanza - nor indeed in any other stanza - suggests that. Nor is this interpretation in any way necessary. Consider stanza 32. It states that every object (bhava) has a real and an unreal part. The real part is its universal. We may add that the real part of every object is Brahman. How? Stanza 33 explains: it is Brahman as divided into cows etc. We see that the division of Brahman must be visualized as consisting of two phases. There is the division of Brahman into universals. These universals are essentially identical with Brahman and do not contain any 'unreal' elements. 'Unreal' elements appear when a further division takes place under the influence of the 'powers' of Brahman. These powers introduce spatial and temporal divisions, among other things, and give rise to our 'unreal phenomenal world. Stanza 35 strongly suggests that the introduction of sequence - the effect of time - is an important factor on the way from 'real' to 'unreal'. The universals themselves contribute in the continuous creation of the phenomenal world:14 14 VP 3.25-27: na tad ulpadyate kimcid yasya jatir na vidyalel atmabhivyaktaye jatih karananam prayojikall karanesu padar ketva nityanityesu jataya / kvacit karyesv ab

Page #10
JOHANNES BRONKHORST Nothing originates which has no universal; the universal urges the causes to manifest it. (25) The universals, entering both the eternal and the non-eternal causes, manifest them selves again and again in certain effects. (26) The universal is also effective in producing activity, it urges the activity to manifest the object in which it resides. (27) The picture which thus evolves of universals is hardly that of an abstract entity different from the things in which it manifests itself, like the universals of the Vaisesika philosophy. In an important way Bhartrhari's universal rather is the thing. It is not correct to think that there is a pot, and the universal potness which is different from it. Quite on the contrary, the pot in as far as it really exists is the universal; its not really existing shadow in the phenomenal world is the individual. It is therefore not possible to say that pot and potness are different, even though the former has a spatial and a temporal dimension, which the latter has not. Universals, seen in this way, can most easily be compared with Plato's ideas: they are real and unchanging, while the things that figure in our experience are their unreal reflections. Returning now to Bhartrhari's sphota, if the real pot is the universal, the same must be true of words: the real word, i.e. the sphota, is a universal. This is exactly the opinion attributed to 'some' in VP 1.96, the stanza so easily brushed aside by Brough:15 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 sounds. If we forget for a moment the attribution of this opinion to 'some', we see that we have arrived at a perfect understanding of the sphota in the context of Bhartrhari's theory. To repeat the main points: Like everything else, words too have two aspects, the real word and its phenomenal manifestations, which are not real. The phenomenal manifestation of the word is sound, the real word its universal, which is the essence of the word (sabdatattva), identical with Brahman (VP 1.1). 15 hivyakt im upayanti punah punah// nirvartyamanam yar karma jatis tatrapi sadhanam/ svas. rayasyabhinispattyai sa kriyayah prayojikall VP 1.96: anekavyak yabhivyangya jatih sphora iti smrta kaiscid vyaktaya evasya dhvanitvena prakalpita //

Page #11
BHARTRHARI ON SPHOTA AND UNIVERSALS 15 As in the case of the pot, it is not possible to say that sphota and sound are different, even though the latter, unlike the former, has a spatial and a temporal dimension. This is exactly what is stated in VP 1.99:16 And a connection with space etc. is also seen in the case of corporeal objects (such as pots); [in the same way] there is no difference between sound and word (i.e. sphota). even though we distinguish different locations [in the case of sound].17 The identical nature of sphota and sound is illustrated with the help of the doctrine according to which the sense organ is of the same nature as the objects it perceives:18 Just as sense organs and their objects are suited to each other in a way which is fixed and does not change, in the same way sphota and sound [are suited to each other] as manifested and manifestor. (100) In the world the cause which elucidates odour etc. - [which are objects] whose sense organs have the same nature (as they themselves] is fixed and determined for each substance. (101) According to this doctrine, the organ of smell is constituted of earth, of which smell is the characteristic property; the organ of sight is fire, which has colour as its characteristic property; and so on. The nature of the sense faculty and its object are therefore identical. Why is the view of sphota as universal attributed to 'some'? Does it mean that Bhartrhari himself did not accept this point of view? The situation appears to be somewhat more complicated. In point of fact, Bhartrhari recognizes two possible views as to the thing denoted by words: it is the universal or the substance (dravya). 19 In the Jatisamuddesa (VP 3.1-110) the point of departure is the view that words denote universals; in the following Dravyasamuddesa (VP 3.111-128) words are taken to denote substance. Bhartrhari does not appear to make a choice between these two alternatives. 16 VP 1.99: desadibhis ca sambandho drstah kayavatam apil desabhedavikalpe 'pi na bhedo dhvanisabdayohll 17 This interpretation of the stanza differs from the one offered in the Vrtti; see Appendix. 18 VP 1.100-101: grahanagrahyayoh siddha yogyata niyata yathal vyangyavyanjakabhavena tathaiva sphotanadayohil sadrsagrahananam ca gandhadinam prakasakami nimittam niyatam loke pratidravyam avasthitam// I prefer the reading -bhavena in 100c to -bhave 'pi, which is slightly better supported by the Mss. 19 See VP 3.2, quoted and translated above. Herzberger translates dravya with 'individual'; I prefer 'substance'.

Page #12
JOHANNES BRONKHORST Consider now the first two stanzas of the Dravyasamuddesa:20 'Self' (atman), 'abiding essence' (vastu). 'own nature' (svabhava), body' (sarira) and 'true principle' (tattva), these are synonyms of 'substance (dravya); it is traditionally believed to be eternal. (111) The abiding essence (vastu), which is real, is known through its forms which are un real. The real (abiding essence) is denoted by words which have unreal delimitations. (112) The content of the second of these two stanzas resembles to some extent VP 3.32-33, studied above. Here again we find that objects have a real and an unreal part. But in the case of the present stanza the real part is the substance, not the universal. Substance and universal are not the same thing for Bhartshari. Bharthari rather deals, in these two sections of the third Kanda of his Vakyapadiya, with the two views regarding the denotation of words, and shows that either way, whether one accepts the one or the other, all words denote Brahman. Let us again retum to the sphota. Besides the view that the sphota is a universal, we would, in view of the above, expect some stanzas in the first Kanda which present the opinion of 'others' according to whom the sphota is substance. This is exactly what we find. The discussion of the sphota as universal begins in VP 1.96 and extends up to 1.104. VP 1.105-110 and 120-121 (1.11-119 are really part of the Vrtti)21 then present the altermative view; 1.105 reads:22 Others declare that the sphota is the utterance) produced by the organs (of speech) on account of their contact and separation; the utterances bom from (this initial) utterance are the sounds. The 'substantial' nature of the sphota here described becomes especially clear in the stanzas 110 and 120:23 Some accept that the (real) word is wind, (others) that it is atoms, (others again) that it is knowledge; for in debates the different points of view are endless. (110) The (real) word (whether it be wind, atoms, or knowledge), though ceaselessly active, is not per 20 VP 3.111-112: atma vastu svabhavas ca sariram tattvam ity apil dravyam ity asya paryayas tac ca nityam iti smrtain satyam vastu tadakarair asatyair avadharyatel asatyopadhibhih Sabdaih satyam evabhidhiyatell 21 See "Etudes sur Bharthari, 1. L'auteur et la date de la Vrtti" section 4 (see note 1, above). 22 VP 1.105: yah samyogavibhagabhyam karanair upajanyatel sa sphotah Sabdajah Sabda dhvanayo 'nyair udartan/l VP 1.110, 120: vayor anunam jnanasya sabdatvapattir isyatel kaiscid darsanabhedo hi pravades anavasthitahl ajasravittir yah sabdah suksmatvan nopalabhyatel vyajanad vayur iva sa svanimittat pratiyatell

Page #13
BHARTRHARI ON SPHOTA AND UNIVERSALS 17 ceived because of its subtle nature; it is noticed because of its cause, just as wind (is noticed) on account of a fan (which moves it). (120) We see that Bhartrhari, on the substantialist alternative, pictures sound as a superfine substance which is not noticed until certain causes specific to it have exerted their influence. This substance by itself does not undergo modifications; it is rather its 'power' which does so when words are pronounced:24 The power of the (word), which resides in the breath and in the mind, is differentiated when it manifests itself in the points of articulation. The timelessness of the sphota can thus be maintained. The enumeration of knowledge' (jnana) in VP 1.110 might cause surprise; knowledge is not normally considered a substance. This depends however on one's point of view. For an idealist substance derives its reality from, is nothing but, knowledge. And indeed, Bhartrhari himself, in his commentary on the Mahabhasya, enumerates knowledge among a number of 'substances which are all, ultimately, identical with Brahman:25 "Because substance is eternal' (Mbh I p. 71. 11-12). The element earth is etemal. What is the true (part) in the element earth? The analytic imagination. What is the true (part) in the analytic imagination? Knowledge. What is the true (part) in knowledge? Om. And that is Brahman. Back to VP 1.110. Here, as so often, Bhartrhari declines to choose between the alternatives. It doesn't matter to him which substance constitutes the sphota, as long as it is clear that the view that the sphota is a substance is shown to be tenable. As we have seen, it may also be a universal. Either way the duration of the sphota is not affected by the duration of the perceived sound (cf. VP 1.106). Appendix: the authorship of the Vrtti. There are a number of reasons which have convinced me that the Vrtti was not composed by the author of the stanzas of the Vak yapadiya; these have been presented in another publication.26 Here I propose to deal with one argument - 24 VP 1.121: tasya prane ca ya Saktir ya ca buddhau vyavasthital vivartamana sthanesu saisa bhedam prapadyatell 25 Mahabhasyadipika of Bhart hari, Fasc. IV. Ahnika I (Bhandarkar Oriental Research Insti tute 1987) p. 22 1. 19-21: dravyan hi nityam nityah prthividhatuh prthividhatau kim satyam/ vikalpah/ vikalpe kim satyam jnanam jnane kim satyam/ om atha tad brahmal. 26 "Etudes sur Bharthari, 1. L'auteur et la date de la Vrtti." (above, note 1.)

Page #14
JOHANNES BRONKHORST admittedly neither the most important nor strongest - which is directly related to the interpretation of VP 1.96-110, 120-121 presented in the last part of the present article. This interpretation deviates from the one given in the Vrtti in some important details. I shall contrast the two interpretations, and show that the one given in the Vrtti is more forced and artificial than its competitor. The interpretation of the Vrtti leads to difficulties under VP 1.99, translated above. Its last pada states that there is no difference between sound and word (na bhedo dhvanisabdayoh); this at any rate would be its straightforward interpretation. This interpretation makes good sense in the context of sphota conceived as a universal, for universals and individuals represent the same thing, be it from its real and from its unreal side; see VP 3.32 translated above. According to the Vrtti, on the other hand, there is no denial of difference between sound and sphota in this stanza, but denial of difference of location. And pada c (desabhedavikalpe 'pi) - which we translated 'even though we distinguish different locations (in the case of sound)' - is, of necessity, interpreted in the Vrtti as 'even though we wrongly distinguish different locations (for sound and sphota)'.27 But this makes little sense, for the tendency is to confuse sound and sphota, not to assign different locations to them. Regarding the stanza as a whole, the Vrtti feels obliged to consider it an answer to a rather absurd double objection. The first objection is:28 "The word is not manifested, because there is a difference of location (between it and that which manifests it). For pots etc. are manifested by lamps etc. (only) when they are in the same location. But words are perceived at a location different from the conjunctions and disjunctions of the organs (of speech) which manifest them." This first objection loses its force if one assumes that rather the sounds manifest the word,29 so the Vrtti raises its second objection:30 "How is a word, which is located in one single place, manifested by sounds which are located in several places, far removed from the word]?" The absurdity of this second objection - the only one that remains - follows from the fact, already stated above, that the tendency is to confuse sound and sphota, not to assign different locations to them. 27 Ed. Iyer p. 163 1. 2-3: ... saty api desabhedavikalpabhimane naivasau tayor bhedo vidyata iti. 28 Ed. Iyer p. 162 1. 3-5: desabhedan nabhivyajyate sabdah samanadesastha hi ghaladayah pradipadibhir vyajyantel karanasamyogavibhagabhyam tu vyanjakabhyam anyatra Sab dopalabdhir itil . 29 Id. 1. 5: sa cayam dhvanisu vyanjakesv aprasangah 30 Id. 1. 5-6: katham ekadesasthah Sabdo nanadesair ativiprakrstair dhvanibhir vyajyata itil