valkya by four intermediate teachers, the first of whom (Asuri)^ is repeatedly quoted in the second (and once each in the first, fourth, and fourteenth) ka;/rt'as. Although these indications do not, of course, supply more than a terviinus a quo for the final settlement of this part of the work, they would nevertheless seem to favour the supposition that the combination of the fire-ritual with the sacrificial system cannot have taken place at a time far removed from that of Sa;«^ivi-putra. The custom of forming metronymics by means of 'putra' is of some interest. It first shows itself in the predecessor of Saw^ivi-putra's teacher in the Ya^>7a- valkya line, and continues from thence down to the very end of the vawj'a. Unfortunately, however, we have no means of ascertaining whether this custom had already been commonly practised, in certain localities, before that time, or whether, as seems to me more probable, it was a fashion of recent date. If the latter alternative could be proved, it might help to settle the chronological relations between Ya^;?avalkya and Pa;^ini, since it would appear from Pa;/, IV, i, 159'^ (and VI, i, 13), that the great gram- marian was well acquainted, not only with the practice of forming metronymics of this kind, but also with that of forming patronymics from such metronymics.
The relative date of Pacini and Ya^>7avalkya has been discussed more than once by Sanskrit scholars^; but no agreement has as yet been come to on what Goldstiicker justly called ' one of the most important problems of Sanskrit literature.' The chief difficulty of this problem lies in the ambiguity of Katyayana's well-known varttika to Pa«. IV, 3, 105. According to Pa;/ini's rule the names
' He is also the Ri'M of Va^. S. Ill, 37.
* This rule, which applies to the people of the north, is not explained in the Mahabhashya. The Kasika Vntti gives the patronymics of Gargiputra and Vatsiputra, both of whom occur in our va»?Ja. It is worthy of remark that Kavasha Ailusha, who is mentioned in Ait. Br. II, 19, and to whom the hymns Rig-veda X, 30-34 are ascribed, is called Kavasha Ailushiputra in the A'a/Aaka 25, 7. Cf. Weber, Ind. Stud. Ill, pp. 459, 157, 485.
^ See especially Max Miiller, History of Ancient Sanskrit Literature, p. 360 seq.; Goldstiicker, Pa'/ini, p. 132 seq. ; Webef, Ind. Stud. V, 65 seq. ; XHI, 443 ; Biihler, Sacred Laws of the Arj'as, I, p. xxxix note.
XXXVl ^ATAPATIIA-BRAHMAJVA.
* Pa«ini, p. 138.
INTRODUCTION. XXXVii
Professor Weber in his review of Goldstiicker's ' Pa;nni.' The interpretation of the varttika adopted in the Ka^ika V;-/tti — according to which Pa;/ini's ride does not apply to those works, because Yag-yTavalkya and others are not old authorities in the sense of Pa;nni's rule — is likewise rejected by him, since in that case Katyayana's exception would be no exception at all. On the other hand, Professor Weber thinks that, if we accept Kaiya/a's interpretation, Katya- yana's additional remark ' on account of contemporaneous- ness' would be entirely superfluous. He, therefore, pro- poses, in the passage of the Mahabhashya, to take 'api' in the sense of 'even,' and to interpret the passage thus^: 'Among the Brahmawas and Kalpas proclaimed by the ancients, Pa//ini ought to have made an exception in regard to Ya^yiavalkya &c., because the Brahmawas and Kalpas proclaimed by them, though indeed going back to ancient (sages), are nevertheless contemporaneous (with Pa/nni him- self).' This rather paradoxical argumentation, on the part of Pata;7^ali, would have to be understood to mean, that the Ya^;7avalkani Brahma;/ani and similar works, though ascribed to old authorities, are in reality modern productions ; or — if we may venture to express it in somewhat different words — Pacini ought to have made an exception in regard to works which, in point of fact, are no exception at all. Now, if this be the correct interpretation, I can only say this — that, had Pata;7^ali been anxious to conceal his real meaning, he could scarcely have done so more effectually than by choosing words which, at first sight, look as clear as day.
Professor Biihler^, who has recently touched upon this controversy, sides with Kaiya/a and Goldstucker ; and I, too, can take no other view. But, like him, I sec no necessity for accepting the inferences which Goldstucker has drawn from this varttika, viz. that we have to assume so long an interval between Pa//ini and Katyayana, that authors, whom Katyayana considered as far older than Pa;nni, were in reality his contemporaries. This assumption, surely, would involve a degree of ignorance, on the part of
1 Ind. Stud. V, 6S seq.j XIII, 443.
' Sacred Laws of the Aryas, I, p. xxxix note.
XXXVlll ^'ATAPATHA-BRAHMAiVA.
Katyayana, regarding the age of Pa/nni, such as would seem altogether unaccountable. The weakness of Goldstucker's argument lies in his identification of the Ya^^avalkani Brahma;/ani with the Brahma;/a of the Va^asaneyins. With Professor Weber I believe that Pa/nni was perfectly well acquainted with the term ' Va^asaneyina/^,' but saw no occasion for specially mentioning it in his rules. Surely, if his silence could possibly have been construed into an act of negligence, Katyayana, who was so intimately connected with the White Ya^us that, on Goldstucker's own showing, he composed the Va^'"asaneyi-pratij'akhya before he wrote his varttikas, would have been the first to notice it. The Ya^/Iavalkani Brahma;^ani, in their relation to the sacred canon of the school, seem to me to stand somewhat on a par with the'Tittiri«a prokta/^ i'loka//^' which, in Pata/7^ali^s time, were excluded from the term ' Taittiriya// ' as uncanonical, and which Professor Weber would identify, perhaps rightly, with some portions of the Taittiriyara;/yaka. Both kinds of tracts probably belong to the last floating materials of Advaryu tradition, which had not yet been incorporated with the canon. Whether or not the Ya^«avalkani Brah- mawani form part of the text of the vSatapatha which has come down to us, and what exact portions of that text we have to understand by this designation, must remain un- certain for the present. Most probably, however, we have to look for them to certain portions of the last book (or books) in which Ya^;7avalkya figures so prominently. If we had a complete copy of the Ka«va recension, we might perhaps be in a better position for forming an opinion on this subject ; for if that version should really turn out to consist of 104 adhyayas, four of these adhyayas may have to be considered as a later interpolation ; and the fact might have become obscured in the Madhyandina recension by a different division of the text-. But, however this may
' Mahabhashya on Pa«. IV, 2, 66; 3, 104.
^ Possibly, however, this redundancy may have been caused by the insertion of the thir.l or uddhari-ka«(fa, consisting of 1 24 l<a«(/ikas, to which there seems to be nothing corresponding in the Madhyandina text. We have no MS. of this parti- cular ka/w'a. I may also mention that, while in the first kajida (or second Ka/na"), the Madh)andinas count 9, and the Kii/aas 8 adhyayas, — in the fouith ka;/(;'a (or
Introduction, Xxxix
be, it appears to me quite intelligible why such portions should have been considered as of equal age to the body of the work ; in fact they would probably go back to about the same time as some of the earlier portions ; only that, owing to a longer state of uncertain transmission, they may have been more liable to changes and additions. If these tracts are not mentioned by Pa;/ini, it may be an accidental omission on his part, or he may not have been aware of their existence, for geographical or other reasons : we can hardly expect Paz/ini to have been so intimately acquainted with the Ya^us texts as Katyayana. As regards the dates of Katyayana and Pata/?^ali, I accept with Professor Biihler and others, as by far the most probable, the fourth and the middle of the second century B.C. respectively.
Under the title of Va^asaneyaka, the ^atapatha-brah- ma;/a is quoted once in La/yayana's vSrauta-sutra IV, 12, 12 ; but I have not been able to find the passage either in the Madhyandina text or in that part of the Ka«va text which I have hitherto had at my disposal, viz. kandas I, II, IV-VII (Ka«va). Far more frequently the work is quoted, either as Va^asaneyaka or as Va^asaneyi-brahma;^a, by Apastamba, both in his ^rauta and his Dharma-sutras, On comparing one of these quotations in the Dharma- sutras (1,4,12,3) with the corresponding passage in the Madhyandina recension, Professor Biihler found that ' its wording possessed just sufficient resemblance to allow us to identify the passage which Apastamba meant, but dif- fered from the vSatapatha-brahma/^a in many details ^' From this he naturally inferred that Apastamba probably took his quotations from the Ka;/va recension. Now, although I have not been able to compare this particular passage with the Kawva text-, I have done so regarding a number of other passages quoted from Apastamba in Karka's commentary on the Katiya-5rauta-sutra. The result was that in no single case did Apastamba's quota- tions agree with the corresponding passages in the Ka//va,
fifth Ka«va\ on the other hand, the Ka^vas have 8, instead of 6 adhyayas ; and in the fifth ki/ida. (,or sixth and seventh Ka«vas) they have together 7, instead of 5 adhyayas.
1 Biihler, kic. cit. p. xxv. ^ The oassage occurs in Madhyandina XI, 5, 6, 3.
xl 5ATAPATHA-BRAHMAJVA.
any more than they did with those of the Madhyandina text^. In some cases they came nearer to the one text, in others to the other. To several quotations, again, I could find nothing corresponding in either text. Now, supposing the quotations, as given by Karka, to be on the whole correct, there seem to be only two ways of accounting for these discrepancies, viz. either Apastamba did not mean to quote the passages literally, but only to give the substance of them ; or he had a third recension of the vSatapatha before him. While some passages would seem to be in favour of the former alternative, others would scarcely admit of this expla- nation. This question,however, requires further investigation, before it can be definitely settled. In connection with this question the fact will also have to be taken into account, that Katyayana, in composing his Va^asaneyi-pratij-akhya, seems to have had before him a different recension of the Samhiti, from those of the Ka;/va and Madhyandina schools^.
Professor Buhler appears to be inclined to place Apa- stamba somewhere about the fifth century B. C. ; and though probably he himself does not consider the reasons he adduces as conclusive, they seem at any rate to show that that writer cannot have lived later than the third century B.C. From
' I select a few passages : —
1. Ap. (Katy.VII, i, 36). pragvamsasya madhyamam sthu/zara^am alabhya
^apaliti va^'-asaneyakam. Madhy. Ill, i, i, 11. sa purvardhyaw sthuwarae-am abhipadyaitad yas'ur aha. Ka«va IV, 1,1, 7. sa yasaii varshish/Aa piirvardhe salasthuwa bhavati tam
abhipadya ^apati.
2. Ap. (Katy. V, 3, 6). atrapi meshaw ^ameshiwz ^a karotiti va^asaneyakam. Madhy. II, 5, 2, 15. tatrapi meshaw ^-a meshiw ka. kurvanti.
Kanva I, 5, i, 13. mesham ia. va api meshi/n ka. kurvanti.
3. Ap. (Katy.VII, 2, 34). (vntrasya kaninikasiti traikakudena/T^anenankte) yadi
traikadukaw nadhigaWAed yenaiva kenawo'anenaM^iti va^'asaneyakam. Madhy. Ill, i, 3, 12. (traikadukam bhavati; yatra va indro vr/tram ahaws
tasya yad akshy asit taw girim trikakudam akarot . . . .) yadi traikakudaw
na vinded apy atraikakudam eva syat samani hy eva/lo'anasya bandhuta. Ka«va IV, 1, 3, 10. (tat traikakuda/K syat; yatra va indro vr?tram ahams
tasya ha ya kaninakasa yat ^akshus tam etaw girim ^ak.ira trikakuda/w ;
sa yat traikakudam bhavati iakshushy evaita* Xakshur dadhati) ; yadi
traikakudam na vinded api yad eva kinka syat.
4. Ap. (Katy.VII, 3, 28). ush«ishe«a pradakshi«aw siro vesh/ayata iti vao'asa-
neyakam. Madhy. Ill, 2, i, 16-17. sa pror«ute. Ka«va IV, 2, i, 11-12. athainaw pror«oti, — sa pronjoti. * See Weber, Ind. Stud. IV, p. 69.
INTRODUCTION.
the fact that 5vetaketu, the son of Uddalaka Aru;/i, the reputed teacher (and rivaP) of Ya^wavalkya, is counted by Apastamba among the Avaras or moderns, Dr. Biihler infers that the promulgator of the White Ya^us cannot have preceded Apastamba 'by a longer interval than, at the utmost, two or three hundred years.' That the two authors may not have been separated from each other by a longer interval seems likely enough ; but, on the other hand, Apastamba, by his remark, pays no very great com- pliment to the inspired texts of his own school, since Aruwa Aupave.yi, the grandfather of vSvetaketu Aru;/cya, is twice referred to in the Taittiriya-sa;;zhita^