Two Literary Conventions Of Classical India

Author(s): Johannes Bronkhorst
Publisher: Johannes Bronkhorst
Catalog link: https://jainqq.org/explore/269609/1

Page #1
TWO LITERARY CONVENTIONS OF CLASSICAL INDIA Johannes Bronkhorst The centuries around the middle of the first millennium of the Common Era are extraordinarily important for the study of Indian culture. These centuries saw, among other things, the rule of the Guptas over large parts of India, and it is very likely that the peace and stability imposed by these rulers, along with their tolerance and encouragement, gave rise to a cultural renaissance. In the realm of literature, a large number of texts which we now consider classical attained their definite forms in this period. It is true that our knowledge of the chronology of Indian literature is very incomplete, yet it is not impossible that, for example, the great epic of India, the Mahabharata, reached in these centuries the form which has been brought to light in the critical edition of this text. It appears that this was a time of collecting and codifying. The Jaina canon of the Svetambaras was collected in this period. The classical texts of several schools of philosophy date from this period, such as the Nyaya Bhasya of the Naiyayikas, and the Padarthadharmasamgraha, or Prasastapadabhasya, of the Vaisesikas. The Samkhya system found its classic exposition in the Samkhya Karika, the Yoga in the Yoga Bhasya. The Mimamsakas codified their system in the Sabara Bhasya, and Sanskrit grammar produced its most important, and perhaps first, commentary on the Mahabhasya, by Bhartshari. But also other kinds of works have been brought in connection with the Gupta period, such as the Kama Sutra, the Artha Sastra, and the Manu Smrti. Also the non-Brahmanical religions were productive. I may mention here only a few of their literary productions: the Tattvarthadhigama Bhasya of the Svetambara Jainas, and the Abhidharmakosa Bhasya of the Sarvastivada and Sautrantika Buddhists have remained classic expositions of these sects. 1 This is the slightly modified text of a lecture given on a few occasions both inside and outside Switzerland, most recently in Poona (India). It briefly discusses some of the issues which have engaged the attention of the author for some time, and are likely to occupy him in the future. Apart from presenting some results of earlier research, it raises a number of questions, not all of which may allow of a definite answer at present. G. Buhler - in Buhler and Kirste, 1892 - has collected evidence in support of the view that the Mahabharata had reached its present form in about the fifth century C.E. 2

Page #2
TWO LITERARY CONVENTIONS OF CLASSICAL INDIA 211 This enumeration is of course not complete. Nor do I wish to give a complete survey of the literature of this period. What I wish to emphasize is that for the study of pre-Gupta India we are often to a large extent dependent upon texts which reached their definite form in the centuries now under consideration. The relationship between these texts and their predecessors can be of various types. In the case of a text like the Mahabharata it is clear that this definite form is really a collection of parts many of which may be considerably older than the collected form. The same is true of the Manu Smrti, if indeed this text reached it classical form in the middle centuries of the first millennium. It seems likely that the Manu Smrti had predecessors, at least one of which was a Dharma text of the Manava school of the Maitrayani Samhita. It is true that there have been different opinions regarding the question whether the predecessor of the Manu Smrti was written in prose or in verse. But both the main exponents of these two views viz., Buhler and P.V. Kane agreed that there was a predecessor of this text, even though they could not adduce positive evidence to support this. In the meantime, however, it has become almost certain that Bhartrhari, who was himself a Manava, was still acquainted with the, or a, Dharma text of that school, and that he identified a verse line as belonging to it.' A collection whose date is rather precisely known is the Jaina Svetambara canon. The Svetambaras themselves believe that its final redaction took place 980 or 993 years after the death of Mahavira, i.e., in 453 or 466 C.E.4 Not all texts from the period under consideration are collections or reeditions of earlier works. Apart from the really original works, which will not be dealt with in this lecture, there are a great many commentaries amongst them. Most commonly these are commentaries on earlier sutras. or verses, in both cases on works which express themselves briefly and concisely. From among the works enumerated above we may mention the Nyaya Bhasya which comments on the Nyaya sutras, the Yoga Bhasya which explains the Yoga sutras, and the Tattvarthadhigama Bhasya which is a commentary on the Tattvartha sutras. The Abhidharmakosa Bhasya comments not on sutras, but on verses. It however treats these verses as sutras by cutting them into pieces; it even refers to these pieces as sutras. The verses of the first two chapters of Bhartrhari's Vakyapadiya are 3 Bronkhorst, 1985. 4 Schubring, 1962: 78; Jaini, 1979: 51-52.

Page #3
212 JOHANNES BRONKHORST similarly commented upon in a commentary, the Vrtti, which unfortunately has been preserved only partially. In this lecture I wish to concentrate on the relationship between the various commentaries and the sutras or verses contained in them. The importance of a correct understanding of this relationship is beyond question. The sutras in particular are not infrequently, the earliest expressions of certain systems of thought which we have, and the Bhasyas are so to say the glasses through which we have to look at them. Bhasyas enclose sutras. Together they form a whole which reads like a single work in prose that contains short nominal phrases, the sutras. This single whole might erroneously be considered the work of one single author. What is remarkable is that some authors of Bhasyas appear to have gone out of their way to create this impression that sutras and Bhasya together are indeed one whole. The following cases illustrate this: (i) The Yoga Bhasya is ascribed by the later tradition to a mythical person called Vyasa, and the sutras to Patanjali. The earlier tradition knows nothing of Vyasa, and the colophon of the Bhasya calls the whole work - sutras and Bhasya - not Yoga Bhasya but Yogasastra, and refers to but one single author, Patanjali. The Bhasya never mentions any variant readings of sutras, and what is more, where it refers to a sutra it uses the first person, as if the sutras were composed by the author of the Bhasya. Yet there can be no doubt that they, or most of them, were not. Some sutras have not been correctly interpreted by the Bhasya, which would be impossible if the Bhasyakara had been their author. This is not the occasion to deal in detail with the sutras which have been misinterpreted in the Yoga Bhasya, the more so since I have dedicated an article to this question.? I find it hard, however, to resist the temptation to briefly mention one example. Yoga sutra 1.25 reads: tatra niratisaya sarvajnabijam 5 6 7 Bronkhorst, 1985a: 203 f. Bronkhorst, 1985b: 170. Bronkhorst, 1985a.

Page #4
TWO LITERARY CONVENTIONS OF CLASSICAL INDIA 213 The preceding sutra deals with God (isvara), which is a special kind of self. The present sutra can therefore be translated: In Him is the unsurpassed germ of the omniscient one. This is not however the way the Yoga Bhasya interprets this sutra. I shall not quote the Sanskrit text, but merely observe that according to this Bhasya the present sutra contains an inference which supposedly shows that there must be an omniscient one. In reality this sutra speaks about Kapila, who is an incarnation of the special self which is God, as can be proved in various ways. (ii) As said already, it is not now possible to go deeper into this and other related questions. Instead we turn to another example of a text which, though commentary, treats itself and the sutras enclosed in it as one indivisible whole. This text is the Tattvarthadhigama Bhasya. As you may know, the Tattvarthadhigama Bhasya is accepted as an authoritative work by the Svetambara Jainas, who hold moreover that its author, Umasvati, was also the author of the Tattvartha sutras contained in it. This view is contested by the Digambara Jainas, who agree with the contents of the sutras but not with those of the Bhasya. * Like the Yoga Bhasya, the Tattvarthadhigama Bhasya never mentions variant readings of sutras; and references to the sutras often use the first person. Yet other indications leave no doubt that the sutras had a different author. Sutras and Bhasya differ on certain points of doctrine, and their choice of words differs; certain sutras, moreover, are incorrectly interpreted in the Bhasya. Again it is not possible to go into details, which have been discussed elsewhere.8 (iii) After discussing a Brahmanical and a Jaina work, our third example should be a Buddhist text. The Madhyantavibhaga Sastra of Vasubandhu is a combination of verses, the karikas, and prose, the Bhasya. Unlike the Yoga Bhasya and the Tattvarthadhigama Bhasya, the Madhyantavibhaga Bhasya refers to the verses contained in it in the third person, so that one is not misled into thinking that both verses and Bhasya have one author. What is more, the initial verse of the Bhasya provides some information about the author of the verse text. It reads: 8 Bronkhorst, 1985b.

Page #5
214 JOHANNES BRONKHORST sastrasyasya pranetaram abhyarhya sugatamajam/ vaktaram casmadadibhyo yatisye 'rthavivecane// Having honoured the author/promulgator of this sastra and him who taught it to/expressed it for me and others, I shall make an effort to explain its meaning. The commentator Sthiramati is of the opinion that the author of the verse text is Bodhisattva Maitreya, its teacher Asanga; but this is not stated in the verse, nor indeed anywhere else in the Madhyantavibhaga Sastra. The verse can be interpreted differently and does not help to determine the author of the verse text. The only information regarding authorship occurs at the end of the Bhasya and says that Vasubandhu is the author." The fact that the verse text came to be ascribed to Maitreya reminds us of the Yoga Sutra, which came to be ascribed to an equally legendary person, Vyasa, probably for the same reason that no indications regarding its true authorship are provided. For our present purposes it is particularly interesting to see that verses and Bhasya occasionally join syntactically. Verse 1.14c, for example, is embedded in a Bhasya sentence, as follows (MAVS p. 36): yas casau tadabhavasvabhavah sa na bhavo napi cabhavah Another instance is verse 1.17cd (p. 40): yadi samala bhutva nirmala bhavati katham vikaradharminitvad anitya na bhavati/ yasmad asya abdhatukanakakasasuddhivac chuddhir isyate agantukamalapagamat na tu tasyah svabhavanyatvari bhavati/ Before we leave this text an observation may be made regarding its name. The colophons call it Madhyantavibhaga-karika-bhasya or Madhyantavibhaga-sastra. The commentator Sthiramati, however, speaks about the Madhyantavibhaga-sutra-bhasya (p. 3). It seems obvious that the karikas and their parts are here referred to as sutras, as we saw was the case in the Abhidharmakosa Bhasya. (iv) Our fourth and final example is the Artha Sastra, supposedly written by Kautilya. This work too consists of verses and prose. Hartmut Scharfe (1968) has shown that at least two persons left their traces in the composition of this work, one of whom, the earlier one, wrote in verse, the other one in prose. Scharfe adduces several arguments in support of this, 9 MAVS p. 192: krtir acaryabhadantavasubandhoh

Page #6
TWO LITERARY CONVENTIONS OF CLASSICAL INDIA 215 among them the fact that the contents of the verses do not always agree with those of the prose. The verse text, moreover, calls its author Kautilya in the very beginning and states that he tore away the land of the Nandas at the very end, while the prose text calls itself a compilation in the first line and its author Visnugupta in the last 10 The exact relationship between the portions of Kautilya and those of Visnugupta is not clear. The concluding lines of the text state that Visnugupta composed both Sutra and Bhasya. What exactly is meant is not clear. It seems likely that here too the verses and parts of verses adopted in the prose are referred to as sutras. This is what happened in the case of the Abhidharmakosa Bhasya, while Sthiramati referred to the verses of the Madhyantavibhaga Sastra as sutras. The concluding lines of Visnugupta are interesting in this context. They form a verse in arya metre and read: drstva vipratipatti bahudha sastresu bhasyakaranam/ svayam eva visnuguptas cakara sutram ca bhasyam ca// The second line means, of course, that Visnugupta himself made Sutra and Bhasya, which does not exclude the possibility that he borrowed extensively from earlier authors, as we shall see. The first line can be interpreted in different ways. Vipratipatti means basically 'opposition' or 'contradiction'. The line may therefore speak of the opposition of the Bhasyakaras against the Sutra, or against each other. In the first case it concerns an incorrect interpretation of the Sutra, in the second a difference of opinion among themselves. Another and at least as important difficulty lies in the word sastresu. Does this word refer to the books, or sciences, on which the Bhasyakaras wrote their Bhasyas? Another interpretation is possible. The whole line may be understood to speak about the opposition of the Bhasyakaras in the Sastras." This would mean that the Bhasyakaras were at the same time the writers of Sastras. This is less peculiar than it seems. Visnugupta describes himself in the same verse as the author of a Bhasya, but he is also the author of a Sastra, the Artha Sastra. A parallel case is constituted by the Yoga Bhasya, which calls itself - including the sutras contained in it - Yoga Sastra. And the names Madhyantavibhaga-karika Bhasya, Madhyantavibhaga-sutra Bhasya and 10 Scharfe, 1968: 80-81. 11 Falk (1986: 59, 58 n. 12) has a third interpretation: "Visnagupta sah haufig einen Widerspruch in den Lehren der Kommentar-Verfasser...".

Page #7
216 JOHANNES BRONKHORST Madhyantavibhaga Sastra are used side by side, as we have seen. A Sastra is in these cases a work which combines sutras (or karikas) and Bhasya, a work which brings a number of elements together and unites them into one. This is exactly what Visnugupta's Artha Sastra says in its first line: ... yavanty arthasastrani parvacaryaih prasthapitani prayasas tani samhrtyaikam idam arthasastram krtam This single (eka) [work called] Artha Sastra has mainly been made by compiling all the Artha Sastras produced by earlier teachers. This is not the place to study how many authors have contributed to the Artha Sastra as we now know it. It is clear that the prose sections may contain parts which derive from various commentators preceding Visnugupta. The statistical investigations of Th.R. Trautmann (1971) do indeed support multiple authorship.12 These four examples - the Yoga Sastra, the Tattvarthadhigama Bhasya, the Madhyantavibhaga Sastra and the Artha Sastra - must suffice to show that there was a tendency in the period which we consider to unite sutras and Bhasya into one indivisible whole, which retained no traces of the original separateness, and authorship, of the enclosed sutras. INT Besides this tendency - perhaps we should say literary convention - there is a second one to which I would like to draw your attention. It finds expression in what I will call the Varttika style. In order to understand this style and its probable origin we must turn to the grammatical literature of ancient India. I do not need to remind you that among the sciences of India grammar is one of the oldest and most important. Its influence on other fields of knowledge was consequently great. It has even been claimed that the grammar of Panini played in India a role similar to that of Euclid's geometry in Europe. Both were, in their respective contexts, methodological guidelines for science and philosophy13 One of the most important texts of Paninian grammar is the Vyakarana-Mahabhasya, or simply Mahabhasya, attributed to Patanjali st and most impat. It has evemilar to that dia a ros even belce on oth h were 12 See also Falk, 1986, esp. p. 69. 13 Staal, 1965.

Page #8
TWO LITERARY CONVENTIONS OF CLASSICAL INDIA 217 (who is not the same as the Patanjali who composed the Yoga Sutra or Yoga Bhasya). The Mahabhasya is an ancient text, and may indeed date back to the second century preceding the Common Era. This Mahabhasya contains within itself nominal phrases which are called 'varttikas'. The researches of Franz Kielhorn in the last century have shown that most of these varttikas derive from an author different from Patanjali, who was called Katyayana.14 Kielhorn was not the first to recognize this fact. To a great extent he followed the Sanskrit commentators on the Mahabhasya, primarily Kaiyata, whose work he completed by trying to identify each and every varttika. The point to which I wish to draw your attention is that there is reason to think that these nominal phrases called varttikas have not always been known to derive from a different author named Katyayana. In works belonging to the centuries which engage our attention the word varttika is used to designate portions of the Mahabhasya which are far more than just the nominal phrases; sometimes the portions called varttika do not even contain such nominal phrases. The word varttika is used in this peculiar way in the Yuktidipika - the most extensive commentary on the Samkhya Karika - and, more frequently and more importantly, in Bhartrhari's commentary on the Mahabhasya. Once again it is not possible, within the time reserved for this lecture, to discuss these points in detail. Those of you who wish to pursue this question may refer to an article which has recently been published in the Wiener Zeitschrift fur die Kunde Sudasiens (Bronkhorst, 1990). The relevant passages strongly suggest that around the middle of the first millennium the nominal phrases which we know by the name varttikas were not recognized as the work of a different author. This observation finds further and unsuspected support in the testimony of the Chinese monk I-ching, who visited India in the seventh century. I-ching's remarks about Sanskrit grammatical literature have always seemed rather problematic. A detailed study by John Brough has led him to conclude that I-ching could not distinguish between varttikas and Bhasya.s What Brough did not know, and could not know, is that I-ching was apparently not the only one who was not aware of this distinction. It seems possible that no one at that time was aware of it. 14 Kielhorn, 1876. 15 Brough, 1973: 257.

Page #9
JOHANNES BRONKHORST Let me make it clear that the nominal phrases which we call varttikas had not escaped the attention of the grammarians of the middle centuries of the first millennium. They even had a separate name for them: vakyas. My point is that they do not seem to have been considered as having an own author in many cases. The evidence is complicated and not completely satisfactory. The one fact which seems to stand out clearly, however, is that the word varttika was used to cover more than just vakyas; they covered vakyas along with the accompanying Bhasya-portion, or even portions of the Bhasya that are without vakyas altogether, Whether or not I have been able to convince you that the varttikas in the Mahabhasya were not looked upon as deriving from a different author, a number of works from the period which we are now studying have the appearance of being imitations of the Mahabhasya considered in this way. Note that the Mahabhasya, once the varttikas are no longer looked upon as the work of someone else, becomes a work characterized by a remarkable style, a style in which ordinary prose passages are frequently interrupted by short nominal phrases vakyas which are subsequently explained. This remarkable style - which we may call 'Varttika style' - was noticed, and more than that, it was imitated as well. Several works of the middle of the first millennium of the Common Era imitate this style, and even call themselves Varttikas. An example is the Tattvartha Varttika of Akalanka, which reads like the Mahabhasya including Katyayana's varttikas. An other example is the Raja Varttika alias Yuktidipika, which I just mentioned. Other works again imitate the Varttika style, but do not call themselves Varttika. Perhaps the best known example is the Nyaya Bhasya, in which this style was already noticed by Ernst Windisch in 1888. 218 IV These, then, are the two literary conventions which I wanted to bring to your notice. The first one is the tendency in commentaries, usually Bhasyas, to swallow up the sutras, or verses, on which they comment, so that together they come to look like one single work: I shall use the expression 'Bhasya style' to refer to it; note however that this Bhasya style does not necessarily occur in all Bhasyas. The second is the tendency to write in what I have called the 'Varttika style': a style in which ordinary prose and short nominal phrases alternate. Again I do not claim that this style is found in all works that call themselves Varttika.

Page #10
TWO LITERARY CONVENTIONS OF CLASSICAL INDIA 219 I have no doubt that many of you will have reservations about the existence of these two conventions, and I cannot blame you for it. It is not possible within the time allotted for this lecture to present all the supporting evidence. This evidence has been published in a few articles, and those of you who are interested may refer to those. During the remainder of this lecture I shall start from the assumption that these two literary conventions are a fact for the period under consideration, and I shall deal with some of the questions which arise in connection with these. Some of these questions may be answerable; others may remain unanswered. (i) I begin with two texts the single or plural authorship of which has been debated for a while. These are the Vakyapadiya and the commentary on its first two books often referred to as the Vrtti. All traditional authors have accepted that both these texts were written by Bhartshari. Doubts about this have not been raised until modern times. Let us look at the arguments which supposedly support the view that Vakyapadiya and Vrtti have one single author. I quote Cardona (1976: 297): The major arguments for concluding that the Vitti was composed by Bharthari himself are as follows. The Vrtti does not record variant readings of verses, but later commentators do. Later authors consider the verses and Vrtti to form a single work. Further, there are striking similarities in thought and expression between the Tripadi (this is the name Cardona uses for Bhartrhari's commentary on the Mahabhasya) and the Vstti. The author who has most vigorously argued that verses and Vrtti have one single author, is Ashok Aklujkar (1972). Aklujkar recognizes that the argument of similarity between the commentary on the Mahabhasya and the Vrtti does not carry much weight. He however emphasizes the fact that Vrtti and verses were intended to be read consecutively, and illustrates this with the help of a number of examples. All this boils down to the following three points: (a) The Vrtti does not record variant readings of verses. (b) Vrtti and verses are meant to be read consecutively, they form one whole. (c) Later authors look upon verses plus Vrtti as one whole. It will be clear that these three points do no more than exemplify the Bhasya style which we discussed in the beginning of this lecture, and which occurs in a number of other works, as we have seen. These three points cannot therefore be used as evidence to show that verses and Vttti had

Page #11
220 JOHANNES BRONKHORST one single author. The fact that the later tradition is unanimous in ascribing the Vrtti to the author of the verses carries as little weight as the tradition among the Svetambara Jainas that Tattvarthadhigama Bhasya and Tattvartha Sutra have one single author. There are, on the other hand, a number of indications which show that verses and Vitti have different authors. I mention the most important ones:16 (a) In a few cases the Vstti gives two alternative explanations of one verse." (b) On two occasions the Vstti quotes a tatrabhavat. Both the views ascribed to this tatrabhavat coincide with views expressed in the Vakyapadiya. 18 (c) The concluding verses of the second kanda of the Vakyapadiya are not commented upon in the Vstti. This is reason to think that they are the concluding verses of the Vrtti. And indeed, they contain the line pranito gurunasmakam ayam agamasamgrahah (2.487 in Rau's critical edition). This means, in Aklujkar's (1978) translation: Our teacher composed this compendium of traditional knowledge. The conclusion is inescapable that the author of the Vrtti is different from the author of the verses. (ii) It is known that the Vaisesika Sutra was once commented upon by a Vakyakara and by a Bhasyakara. This suggests that there was once a commentary in Varttika style on the Vaisesika Sutra, containing both vakyas and Bhasya-portions. This possibility, in its turn, explains some otherwise obscure facts. I shall confine myself to one single example. The Padarthadharmasangraha contains some passages in Varttika style which appear to be borrowings from another text. One of those passages fits so badly into its context that the commentators have great difficulty making sense of it all. This passage begins with the nominal phrase "No, because body, sense-organs and mind are not conscious" (na, sarirendriya 16 17 See Bronkhorst, 1988. These have been discussed by K.A. Subramania Iyer in the Introduction of his English translation of the first chapter of the Vakyapadiya (1965: xxix-xxxi). See Bronkhorst, 1988: 110 f. 18

Page #12
TWO LITERARY CONVENTIONS OF CLASSICAL INDIA 221 manasam ajnatvat).19 However, the preceding lines contain nothing to which this nominal phrase could be a response. The following explanation of this nominal phrase, on the other hand, continues the preceding discussion in a satisfactory manner. It seems clear that Prasastapada, the author of the Padarthadharmasangraha, borrowed here a passage in Varttika style from another work. About the nature of that other work there can be little doubt, for Prasastapada is known to have written a commentary on the Vaisesika Vakya-cum-Bhasya.20 (iii) The Nyaya Bhasya comments on the Nyaya sutras. The first of these sutras gives a brief survey of the topics to be dealt with, and most of the remaining sutras fit well into this scheme. This is a reason to think that the Nyaya sutras as a whole are no loose collection. Some few sutras however, do not fit into the scheme. Sutras 4.1.11-40, for example, look like an insertion, because they do not correspond to anything announced in the initial sutras. But if these sutras were inserted, the question is: who inserted them? We have seen already that the Nyaya Bhasya is an example of a text which uses the Varttika style. This means that the Nyaya Bhasya commented on nominal phrases - the sutras -, and besides this contained nominal phrases the vakyas which characterize the Varttika style. It is clear that in such a situation confusion can easily arise. One possible answer to the question who inserted the additional Nyaya sutras may therefore be: they were inadvertantly taken over from the Bhasya. I do not maintain that this is necessarily the right answer to this question. There are complications, which I have referred to in a published article (1985c). Yet it is clear that our awarenes of the Varttika style can influence the way we approach problems of this kind. (iv) A similar situation presents itself in the commentary on Aryadeva's *Sataka ascribed to Vasu. Karen Lang (1988) has studied this commentary and expressed the view that it examplifies the Varttika style. This, she argues, may have the following consequence. It has long been assumed that the *Sataka cites four of the Nyaya sutras. In reality, according to Lang, these sutras may not be cited by the *Sataka, but by Vasu's commentary. The confusion could arise owing to the Varttika style of that 19 Pdhs p. 69 1. 10-11. 20 See Bronkhorst, forthcoming.

Page #13
222 JOHANNES BRONKHORST commentary. It is clear that Lang's thesis, if true, might have chronological consequences. In that case we cannot take it for granted anymore that Aryadeva knew these Nyaya sutras. (v) Several of the preceding examples dealt with the Varttika style and its possible effects on the texts commented upon in this style. Our last example, like the first one, will deal with the Bhasya style. Tradition states that both the Abhidharmakosa - i.e. the verse text - and the Abhidharmakosa Bhasya were composed by one and the same person, viz., Vasubandhu. I am not going to bore you with a detailed account of the controversy which has arisen regarding the reliability of the tradition of the life and works of Vasubandhu. This controversy mainly concerns the belief that Vasubandhu became a Mahayanist later in life. No one seems to have seriously asked the question whether one and the same person wrote both Kosa and Bhasya. This is remarkable, for verses and commentary represent different points of view: the verses mainly the Vaibhasika, or Sarvastivada, position, the commentary the Sautrantika position. The traditional account gives some kind of explanation for this, but one which on close inspection does not look very plausible. What is more, Kosa and Bhasya do not just represent Vaibhasika and Sautrantika positions, as tradition would have it. If the Bhasya is to be believed, some of the verses express Sautrantika views. And what is even more surprising, the Bhasya differs from the Kosa regarding the correct Vaibhasika position in a few cases. An example is the Bhasya on Abhidh-k 3.2. This verse states that there are 17 places' (sthana) in the Rupadhatu, viz., three 'stages' (bhumi) in the first three Dhyanas, eight in the fourth.21 The Bhasya specifies these stages, enumerating, among others, Brahmapurohitas and Mahabrahmans in the first Dhyana. Then the Bhasya continues: "There are [only] 16 (places) according to the Kashmirians. As is well-known (kila), among the Brahmapurohitas a higher place has been erected for the Mahabrahman, which is like a tower (? parigana), inhabited by [only] one ruler; this is not however another stage (bhumi)."22 Abhidh-k 3.2: urdhvam saptadasasthano rupadhatuh prthak prthak/dhyanam tribhumika tatra caturthai tv astabhumikam// Abhidh-k-bh (P) p. 111 1. 26-27: ... sodaseti kasmirah/brahmapurohitesv eva kila sthanam utkrstataram mahabrahmanah parigana ivabhinirvrttam ekanayakam na tu bhumyantaram

Page #14
TWO LITERARY CONVENTIONS OF CLASSICAL INDIA 223 There can be no doubt that the 'Kashmirians' here referred to are the Vaibhasikas of Kashmir, for their opinion is found in the Mahavibhasa, after which the Vaibhasikas were named.23 Moreover, the 'Kashmirians' are a few times explicitly connected with the Prakarana(-pada), one of the canonical Abhidharma works of the Sarvastivadins.24 And frequently the opinions ascribed to the 'Kashmirians' can be found in the Mahavibhasa.25 A similar case is constituted by the Bhasya on Abhidh-k 1.10c. This quarter verse states that smell is of four kinds (caturvidho gandhah). The Bhasya explains the four kinds of smell: good and bad smell which can be excessive or non-excessive. 26 Then the Bhasya continues: "But (smell is) threefold according to the Sastra, which says) 'Smell is good, bad, or indifferent'.1:27 The quotation is from the Prakaranapada, 28 a canonical text of the Sarvastivadins. Here again, therefore, verses and Bhasya disagree as to what is the orthodox view of the Sarvastivadins. On one occasion the Bhasya points at an insufficiency in a verse and rectifies it. This happens under verse 2.50, which reads: Coexisting (causes) (sahabhu) have one another as effects, such as the elements (bhuta), thought and the accompaniments of thought, the characteristics and what they characterize. This definition is not fully satisfactory, since the secondary characteristics (anulaksana, i.e. jatijati etc.; see 2.46a) have as coexisting cause the dharma which they accompany, but not vice versa. The Bhasya therefore completes the definition: "It must be added (upasamkhyatavyam) that even without mutuality a dharma is coexisting cause of its secondary characteristics, they not of it."30 The references in the Bhasya to the author of the verses do not allow us to draw any conclusions whatsoever. Sometimes these references use the first person. For example, the expression pascad vaksyamah 'we'll 23 See Abhidh-k (VP) II p. 3 n. 1. 24 Abhidh-k-bh (P) p. 84 1. 10-15 (on 2.51), p. 89 1. 7-13 (on 2.54). 25 See Abhidh-k (VP) I p. 76 n. 1, p. 89, p. 205, II p. 13 n. 3. 26 Abhidh-k-bh (P) p. 7 1. 6: sugandhadurgandhayoh samavisamagandharvat. Yasomitra explains: anutkatotkatagandhatvad ity arthah. 27 Abhidh-k-bh (P) p. 71. 6-7: trividhas tu sastre/ sugandho durgandhah samagandha iti/ See Abhidh-k (VP) I p. 18. Abhidh-k 2.50: sahabhur ye mithahphalah/bhutavac cittacittanuvartilaksanalaksyavat// Abhidh-k-bh (P) p. 83 1. 23-24: vinapi canyonyaphalatvena dharmo 'nulaksananam sahabhuhetur na tani tasyety upasankhyatavyam.

Page #15
224 JOHANNES BRONKHORST discuss [this] later' is used in the Bhasya on 1.10 (p. 7 1. 10) to refer to verse 1.12; vyakhyasyamah at p. 89 1. 4 (on 2.54) refers to verse 5.12; the same term at p. 274 1. 24 (on 4.125) refers to verse 6.17; vaksyamah at p. 353 l. 12 introduces verses 6.29 f. After what we have learned from the Yoga Bhasya and Tattvarthadhigama Bhasya we will not be tempted to derive conclusions from this usage, the more not because the references may be to the Bhasya which explains those verses. But nor can we draw conclusions from the references in the third person. On a number of occasions the Bhasya uses vaksyati 'he'll say' in order to refer to a verse. For example, the one but last sentence of the first chapter of the Bhasya (p. 37 1. 14-15) states: "He will explain later (pascad vaksyati) that the female and male (sexual] organ's are part of the dhatu (called) 'body'." This refers to verse 2.2 which explains (at least in the interpretation of the Bhasya) that there are six organs (indriya), and that the female and male sexual organs are merely distinguished from the body, but not different from it, because of their supremacy regarding femininity and masculinity:31 The Bhasya on the first part of Abhidh-k 2.33 indicates with the help of vaksyati that the last word of the verse (cetasah) is to be understood here too (p. 60 1. 25). The Bhasya on Abhidh-k 2.67 uses the same device to show that anantaram is here valid from verse 68 (p. 103 1. 20). The use of vaksyati on Abhidh-k 3.17 (p. 128 1. 28) serves a similar purpose. References to the Bhasya, on the other hand, use the first person: vaksyamah (p. 107 1. 3 and 17, on 2.72) and pravaksyamah (p. 400 1. 15, on 7.13) introduce immediately following portions of the Bhasya; cintayisyamah (p. 93 1. 16-17, on 2.55) refers to the Bhasya on 5.27; pascad vaksyamah (p. 343 1. 19) refers to the Bhasya on 7.13 (p. 400). All these cases do not allow us to draw any conclusions, because cases are known where an author uses the third person to refer to his own verses. An example is Mandana Misra, who - in the Brahmasiddhi, which consists of verses and commentary, both by the same author - uses on several occasions the third person in the commentary part to refer to his verses. 32 31 Abhidh-k 2.2: svarthopalabdhyadhipatyat sarvasya ca sadindriyam/ stritvapumstvadhipatyat tu kayat stripurusendriye // The Bhasya explains (p. 39 1. 14-15): kayendriyad eva stripurusendriye prthak vyavasthapyete/ narthantarabhute/ kascid asau kayendriyabhaga upasthapradeso yah stripurusendriyakhyam pratilabhate/ 32 E.g., p. 75 1. 4: darsayati; p. 23 1. 17: aha.

Page #16
TWO LITERARY CONVENTIONS OF CLASSICAL INDIA 225 We finally consider one more point: the Bhasya refers to the author of the verses as Acarya.93 Verse 1.3, for example, is introduced in the following manner: Why (should there be) teaching of Abhidharma, and by whom has it been taught for the first time, that the Acarya piously applies himself to pronouncing the Abhidharmakosa?34 The author of the verses is again referred to as Acarya in the Bhasya on Abhidh-k 1.11. This verse explains a concept of the Vaibhasikas. The Bhasya points this out, then adds that the word ucyate 'it is said/is called' in the verse shows that this is said by the Acarya.95 The purpose of the last part of this lecture was to raise questions, rather than to solve them. The case of the Abhidharmakosa and Bhasya is particularly complex, and much more research will have to be done before reliable conclusions can be drawn. The same applies to the other examples which have been discussed. My main purpose has been to ask questions. In some cases an answer seems possible, in other cases this may not yet be the case. In spite of this, I hope that these questions constitute a modest contribution to the progress of our field of study. After all, the right question is often half the answer. Added in proofs: Long after this article had been submitted for publication I discovered that the essentials of the 'Varttika style' had already been correctly described by V.G. Paranjpe in his article "The text of the Nyayasutras according to Vacaspatimisra", PAIOC 10, 1941, 296-309. 33 Ruegg (1990: 64) considers this point not decisive and draws attention to Haribhadrasuri's Anekantajayapataka (ed. Kapadia, vol. i, p. 2.12) for a parallel. 34 Abhidh-k-bh (P) p. 2 1. 18-19: kimartham punar abhidharmopadesah kena cayari prathamata upadisto yata acaryo 'bhidharmakosam vaktum adriyat(e). 35 Abhidh-k-bh (P) p. 8 1. 9: ucyata iti acaryavacanam darsayati.

Page #17
226 JOHANNES BRONKHORST References and abbreviations Abhidh-k = Abhidharmakosa Abhidh-k (VP) = L'Abhidharmakosa de Vasubandhu, traduit et annote par Louis de La Vallee Poussin, 6 vols., Paris 1923-1931. Abhidh-k-bh (P) = Abhidharmakosabhasya of Vasubandhu, ed. P. Pradhan, rev. 2nd ed. Aruna Haldar, Patna 1975. Aklujkar, Ashok (1972): "The authorship of the Vakyapadiya-Vrtti." Wiener Zeitschrift fur die Kunde Sudasiens 16, 181-198. Aklujkar, Ashok (1978): "The concluding verses of Bharthari's Vakya-Kanda." Annals of the Bhandarkar Oriental Research Institute 58-59 (1977-78; Diamond Jublilee Volume). 9-26. Bronkhorst, Johannes (1985): "The origin of an Indian dietary rule: evidence for a lost Manava work on Dharma." Aligarh Journal of Oriental Studies 2 (1-2) (Ram Suresh Tripathi Commemoration Volume), 123-132. Bronkhorst, Johannes (1985a): "Patanjali and the Yoga sutras." Studien zur Indologie und Iranistik 10 (1984 (19851), 191-212. Bronkhorst, Johannes (1985b): "On the chronology of the Tattvartha Sutra and some early commentaries." Wiener Zeitschrift fur die Kunde Sudasiens 29. 155-184. Bronkhorst, Johannes (1985c): "Nagarjuna and the Naiyayikas." Journal of Indian Philoso phy 13, 107-132. Bronkhorst, Johannes (1988): "Etudes sur Bhartrhari, 1: L'auteur et la date de la Vrtti." Bulletin d'Etudes Indiennes 6, 105-143. Bronkhorst, Johannes (1990): "Varttika." Wiener Zeitschrift fur die Kunde Sudasiens 34, 123-146. Bronkhorst, Johannes (forthcoming): "The Vaisesika vakya and bhasya." Annals of the Bhandarkar Oriental Research Institute. Brough, John (1973): "I-ching on the Sanskrit grammarians." Bulletin of the School of Oriental and African Studies 36, 248-260. Buhler, G., and Kirste, J. (1892): "Indian studies, No. II: Contributions to the history of the Mahabharata." Sitzungsberichte der Philosophisch-Historischen Classe der Kaiserlichen Akademie der Wissenschaften 127, XII. Abhandlung. Cardona, George (1976): Panini, a Survey of Research. Delhi: Motilal Banarsidass. 1980. Falk, Harry (1986): "Die Prufung der Beambten im Arthasastra." Wiener Zeitschrift fur die Kunde Sudasiens 30, 57-72. Iyer, K.A. Subramania (tr.) (1965): The Vakyapadiya of Bhartshari with the Vrtti. Chapter I. Poona: Deccan College. (Deccan College Building Centenary & Silver Jubilee Series, 26.) Jaini, Padmanabh S. (1979): The Jaina Path of Purification. Delhi: Motilal Banarsidass. Kielhorn, Franz (1876): Katyayana and Patanjali: Their Relation to Each Other and to Panini. Bombay. Reprint: Kleine Schriften I. Wiesbaden: Franz Steiner. 1969. Pp. 1-64. Lang, Karen (1988): "On Aryadeva's citation of Nyaya texts in the *Sataka." Wiener Zeitschrift fur die Kunde Sudasiens 32, 131-140. Mandana Misra: Brahmasiddhi. Edited, with Sankhapani's commentary, by S. Kuppuswami Sastri. Second edition. Delhi: Sri Satguru Publications. 1984. (Sri Garib Das Oriental Series, 16.)

Page #18
TWO LITERARY CONVENTIONS OF CLASSICAL INDIA 227 MAVS = Madhyanta-Vibhaga-Sastra, ed. Ramchandra Pandeya. Delhi: Motilal Banarsidass. 1971. Pdhs = Padarthadharmasangraha: The Prasastapada Bhashya with Commentary Nyayakandali of Sridhara, edited by Vindhyesvari Prasad Dvivedin. Second edition. Delhi: Sri Satguru. 1984. (Sri Garib Dass Oriental Series, 13.) Rau, Wilhelm (ed.) (1977): Bhartrhari's Vakyapadiya. (Abhandlungen fur die Kunde des Morgenlandes, 42.) Wiesbaden: Franz Steiner. Ruegg, D. Seyfort (1990): "On the authorship of some works ascribed to Bhavaviveka/Bhavya." In: Earliest Buddhism and Madhyamaka. Edited by David Seyfort Ruegg and Lambert Schmithausen. (Panels of the VIIth World Sanskrit Conference, 2.) * Scharfe, Hartmut (1968): Untersuchungen zur Staatsrechtslehre des Kausalya. Wiesbaden: Otto Harrassowitz. Schubring, Walther (1962): The Doctrine of the Jainas. Translated from the revised German edition by Wolfgang Beurlen. Delhi: Motilal Banarsidass. 1978. Staal, Frits (1965): Euclid and Panini. In: Universals. Studies in Indian Logic and Linguistics. University of Chicago Press. 1988. Pp. 143-160. Trautmann, Th.R. (1971): Kautilya and the Arthasastra. A statistical investigation of the authorship and evolution of the text. Leiden: E.J. Brill. Windisch, Ernst (1888): Ueber das Nyayabhashya. Leipzig.