English
How utterly aitificial and preposterous is the sandhi between two lines of a Vedic stanza has long since been pointed out (JAOS. x. 425). Thus at RV. ix. 97. 38 a b we have sd punand dpa sure nd dkaidhld aprd rddasl vi sd dva^y Pada-patha dhdta \ d \ ubhi | aprdJi |. The a belongs of course to b; and we must pronounce thus, dhaidi 6hM^ &c. The gi'aphical combina¬ tion of dlidta with 6hhA would give dhatddhkS ; instead of which, we have dlidtd first combined with d to dkata^ and then that with uhM ! In the case of vii. 20. 4, uhM cid indra rddasl mahitvd paprdtha tdvinhhis tuvismalj^ it is true that four considerations (the lexical ^ one; the metrical; the accentlessness of paprdtha ; and the Pada-patha) unite to show that b must begin with d paprdtha ; but I know of no parallels for such absurdities in English, save in the humorous verses of Thomas Hood or the like. lUnstrations of the results of non-division. —The non-division of words in Sanskrit books, in connexion with the laws of sandhi, is a fruitful source of misunderstandings. It remains to enforce this thesis (and incidentally the above-cited statement of the Pratiiakhya) by some actual examples. These illustrations arc chosen, a few from very many; are intended to be typical; concern famous texts or interesting matters; and are cumulative in their nature. From the Big-Veda: yuv^rdrava, &c. —By the rules of sandhi, both the combinations, and + give the same result, namely, In ^. + a- and -o/^ + ra- yield -ora-. In combinations of this typo the ambiguity is complete. For convenience, let the texts come first : * Finite forms of prCt without a or tl are of extremcst rarity: correct PW. xl Notes on the Externals of Indian Boohs viii. 39. 2 c d e ny dvatlrdravnam vUvd aryd dratlr ltd yucckantv dmura/t X. 40, 7 yuvdm ha Bhujyum yuvdm Alvina VdAavi 0 0 yuvdm Sihjdram Usdnam updrathuk \ yuvdrdrdvd pdri saJckydm dsate yuvdr ahdm dvasd sumndm a cake 1| vii, 68. 7 utd iydm Bhujy'dm Ah'iTid sdkhdyo mddhye jakur durSvdsa/i samudri \ nir im par-^ad drdvd yd yuvdkuh || At viii. 39.2, &kalya divides ni \ drdtlh \ rdrdvndm |; and Sayana rendei'S c: [And bum] down the enemies of those who offer oblation, raravn^ havik pi^ayacchatdm aratih Sairuiis ca nidaha. At x. 40. 7, &kalya again divides yuvdh | rdrdvd \; and Sayana renders c; The friendship of you two, the giver of oblations, the sacrificer, await = awaiteth, yuvor yuvayoh sakhyam mitratvam rarava lixivi^dm praddtd yajamdno (sic) parydsate] parydste || vaeana-vyatyayah ||. Roth, in 1871, followed Sakalya, and set up (PW. vi. 284) a stem rdrdvan, fieigebig, for viii. 39. 2 and x. 40. 7, without being able to adduce a single other example.* In his oral lectures, however, of April 28, 30, 1875, upon RV^. X. 40, he dissented from ^akalya, and divided yuvdr drdvd and drdtlr drdvndm.^^ It is noteworthy that Ludwig, in his Rigveda, 1875, agrees with Roth’s later view as to the word-division (but not the translation) of both vm. 39. 2 and x. 40. 7; and that Ludwig’s view and Roth’s are doubtless independent each of the other.® In 1884, in the minor PW., v. 171, Bohtlingk assumed, ‘with Delbruck.’ d-rdvan, ‘hostile,’ for viii! 39. 2, and dravan = drvan, ‘ steed,’ for x. 40. 7. In 1894 Roth published, ZDMG. 48. 113, his orally given solution of 1875 for x. 40. Now it matters not how the question is decided. The fact remains that two of the greatest Vedic scholars of India, ancient or mediaeval, agree in > And 80 Gm^mann, Wtb. and Ueb., 1873-77. The assumed stem ,^-ra-ra appears tobe as isolated in respect of formation (Whitney, Gram., § 1169d) as of occurrencl Ibis solution was mentioned by me in the Jenaer Literatur-Zeitunc. for 1875 no. 52. ® ^ ® Ludwigs preface is dated Nov., 1875. He renders d-ravndm by ‘der nicht-f^eber* (vol. 1 , p. 429); ijuvdr dra-va by ‘euer speichen-reicher’ (i. 83); and dra-cd bv ‘der [wagen] mit speichen ’ (i. 61). ^ * The word rdmvan does not happen to be treated in the Vedische Studien- but Geldner ( 111 . 84; against Aufrecht and Roth, KZ. 25. 601-2) suggests in fact another possibility for the seemingly parallel case of vivM)ahsi at i. 69.8 Cf vi .31 .3 • vii 21.4; iv. 19.10. ' xli By the Editor of this Series deciding it in one way; and that two European scholars of unquestioned eminence decide it in just the opposite way. Of the same type is agnirdvena at RV. i. 128. 6. Sakalya, the Pratisakhya (v, 28=:376)jand Sayana, all bid us divide it as agnih \ dvena^ although a stem dva is unquotable elsewhere, and the nati is hardly paralleled save by gdr 6hena, Ludwig declares with emphasis (iv. 276) that rdvena is right. Other examples abound, but—sat prata biberunt. Prom certain editions of an Upanishad,—At RV. i. 114. 8 d, as part of the oft-repeated md nas toke, we have the verse f Sayana renders sddam it tvd by sarvaddiva tvdm; Mahidhara, by saddiva tydin. The verse recurs at Svetasv. Upanishad iv. 22. Roer’s edition* (p. 350) reads ^ (so spaced): Jibananda’s reprint (p. 93) of course follows suit; and even Deussen's version (p. 304) implies sadasL Moreover, strange to say, Roer’s reading of Samkara’s comm, supports Roer’s text by giving sadasi tvd havdmahe asmin setdasi tvdm &c. I In Kesavalal’s ed. of Bombay, 1865, p. 8, I find ^ (so spaced); and, much to my regret, also in the charming little silk-bound ed. of Panasikara,'*^ Twenty- eight Upanishads, p. 251. Since a double mute in combination with any other consonant is written as simple ® by the MSS., it is quite true that itvd may be taken either as itvd or as ittvd ; but I fear that these editors can hardly be excused on this ground. The Ananda Asrama text of Samkara’s comm. (p. 62) reads sadam it tvd havdmahe saddiva ; and the reading which the three other comm’s in that volume explain (p. 34, p. 20, p. 41) is sadam it tvd. If the blunder cannot claim an age so respectable as that of Samkara, it has at least surely misled many readers. From the Pada-patha of the Atharva-Veda.—This work stands on a wholly different footing from that of the Rigveda.^ At xix. 22. 21 the padakura divides spdrddhitum kdk as epdrddhi \ tumkdh I and at 68. 4 ho divides vdrmdsivyadhvdm ... purahkpixLdhvammdyaslr (a muddled version of RV. X. 101.8) as vdrma ( asi | vUadkvdm | ... krnudhvam | md | dyasU, But disregarding the pada-readings of book xix,® it is easy to cite from earlier books downright mistakes such as no one can deny. Thus at xiv. 1.9, ydt ... eavitadaddt is divided as ydt | savitd | adaddt. The staunchest * Bloomfield, Concordance, 1064 a, cites from this ed., and cites it as reading sadam it tvd. * Bombay, Nirnaya Sagara Press, 1906. “ Whitney, Gram., § 232; AV.Pr. iii. 28; TPr. xiv. 1; e.g. dattvd^datvd. * See HOS. vii, p. Ixix, line 13. ^ See HOS. vii, p. Ixix, n. 1; and viii, p. 896, top. f xlil Notes on the Externals of Indian Books defender of the padakara is silenced here, if not by the rules of accent, then by Sakalya himself, who, at RV. x. 85. 9, resolves aright ydt | savitd \ ddadat. At AV. iv. 19. 6, the padakara resolves tddydmeti as tdt \ ydm | etij although the slightest heed of the accent would have shown him that it is impossible ^ for the combination to mean anything but tdd dydm eti. His resolution of sdmdtvdm {=s6mdt | tvdm) as s6md | tvdm, at iv. 10. 6, is quite as bad. At xiii. 2. 35, he resolves aprdddydvdprthivt antdriJcsam as d I aprdt \ dydvd-. Since dy means either ddy or dy (Whitney, § 232 again), and since aprdt for apvds is not without parallel (§ 889 a, § 555 a), we may perhaps absolve him from ignorance of Vedic grammar, but not from amazing unfamiliarity with the ubiquitous recurrences of this verse.^ Prom the Taittinya-Samhita, i. 1. 1.—The celebiated opening phrases of this text are tvd \ urji tvd |. These are addressed by the adhvaryu to the switch which he cuts to drive the calves away from their mothei’s before milking the latter for the oblation of milk. Then follow vdydva stlia I updydva stka^ addressed to the calves. If the fourth phrase means ‘you’ll come back again’ (when we’ve milked them), the third ought to mean ‘you’re going away now’ (while we milk them); and such is fndecd the antithesis which the comm’s first explanation of it implies. He says (Poona ed., p. 12): vdnti gaccliantiti vdyavo gantdrah | upa samipe yaja-‘ mdnagrke punar dydnty dgacchantUy updyava/t \ he vatsds trnahhaha- ndya pratharnam rndtrsakdadd apetya svecchaydivdranye gantdro bkavata | sdyam punar yajamdnagrhe samdgantdro bhavata |. This antithesis is very ill expressed by vdyavah. The PW., at vi. 932, sets up a stem vdyu from m for vdydvaA, and renders it ‘ appetentes, ascher, zudringlich (intruders). The comm., in his alternative explana¬ tion, takes it in its usual sense of ‘ winds but labours hard to give it pertinence.^ Bohtlingk believes ^ that the enigmatical vdydva/i is due simply to misdivision of v/rj4tvavdydvasthxiy and that the true division is urji tvd \ avaydyah sika. His ava^dydvas, in respect of fonnation,^ pairs well with upa-aydvas ; ynd as to the meaning,^ it is a perfect antonym of upa- aydvas. If Bohtlingk is right, then the blunder (often repeated in Yajus texts ) must be, as he says, a very old one."^ ’ See HOS. vii, p. Ixvii. = ZDMG se^m ’ ™ 2- 35- * From ava + i: cf. ai/u from , and catjd from ci. ronL"'iI!Vd.ld. Gespenst/ a, Geldner ® See Bloomfield’s Concordance, 860 a, ’ Cf. ^B. i. 7 p By the Editor of this Scries xliil Trom the Taittiriya-Brahmana. —The original text of the quaint akhyayika at ii. 1.