368

THE INDIAN ANTIQUARY.

[DEcEMBER, 1873.

whose roaring, from their mouth falls the grass.” Again, Mārkal ſleya in the section on the Sāvari, a

Keraka is never used in the Mrichchhakatika or any other play in the sense of a genitive postposi

kind of Śāmdā'i, has the sūtra (fol. 66 b): || kerake

tion; it never determines the case of another

kelake väsyā5 || amhakerakam Dhalam ambakela kam vá l; and Chandrašekhara, the best commenta

noun; it has never been anything else but a real adjective noun. Prof. Hoernle denies having said that the geni tive of santúna was formerly santána keraka. At p. 132, however, he says: “Take, for instance, the genitive of santina, a child ; it would be santána

tor of the Śakuntalá, remarks : kerakasabdah prä krite àtmiye wartate. Thus kera, keraka, kelaka, are found even in the latest and most corrupt dialects.

When should it have been curtailed,

and what particular necessity could induce the Bangălis alone to shorten it, while all the others have either lengthened it or retained it unal tered P According to Vararuchi, III. 18, 19, corresponding to Hemachandra, VIII. 1. 155

keraka.”

What else can this mean but what I

have concluded from it? That the Bangāli adjec tives have dropped all case, number, and gender terminations I knew as well as Prof. Hoernle does :

but exactly because all of them have done it, and

and VIII. 2. 63-64, Trivikrama, I. 4. 59-60, the

because this is the rule, it is difficult to see how

words türya, stºrya, and dhairya may elide the ya and become tºra, stºra, dhira (comp. Lassen,

keraka alone could have been curtailed to such

Inst. prácr. p. 247). After the same principle hdrya becomes kāra ; the word has not been

noticed by the grammarians, because it existed al ready in Sanskrit. This kira is preserved in the Bangāli genitive aſſºſ.TART", i.e. sITTT + 3ſº, and has been curtailed to sIFRTF, Firſt, and in Urdu to hamſīrd, tumhórá. Hemachandra, VIII.4. 434, in the section on the Apabhrafisa has the sūtra :

| yushmadāderiyasya dārah ||apabhramäe yushma dādibhyah parasya iyapratyayasya dāra ity fideso bhavati ||, and among the examples tuhárá, amhird, mahārd are quoted. Trivikrama, III. 3. 23, and inharāja, fol. 73 b, have : || chhasya yushmadāder dārah ||. If we compare these siltras with the siltras mentioned above, nobody, I think, can

an extent. In the language of the gipsies, where, as I have remarked above, kera is very frequently employed, the adjectives are treated in almost the same way as in Bangāli, but still kera had retained its old shape. Whether keraka occurs fourteen or twenty-eight times in the Mrichchha katiká is of no consequence. I should not have mentioned that at all if I had not been struck by the astonishing confidence with which Prof. Hoernle asserted that this word in the determina

tive sense—according to his views—is found in the Mrichchhakatiká only ; a confidence all the more astonishing as he confesses now himself that he has not even examined, to say nothing of read, such plays as the Milaviki and the Mudrárák shasa /

That the word keraka must have been

doubt that dra, which, as the Bangāli shows,

very common in the colloquial speech Prof. Hoernle

originally was kira, and our kera are only modi

need not tell me.

fications of the same word, viz. kārya. Raira could easily be curtailed after a homogeneous vowel, being of frequent occurrence already in Sanskrit; but kārya in the shape of kera is a mere Prākritic word. Originally its use was restricted

why it should have been curtailed; the ques

to the pronouns and the words para and rájan; afterwards it was lengthened and used in connec tion with substantives.

tailed.

It has never been cur

Secondly, the change of r to l forbids us

to accept Prof. Hoernle's theory. There can be no doubt that kelaka is the more modern form; and

that the change of r to l in this word is not arti ficial, but thoroughly organic, is proved by the Ma rābhi keli, keli, kºlem, nd the Low Hindi kail." mentioned by Prof. Hoernle himself. Indeed it would be a strange phenomenon if the same word kera had not only retained its original shape in the vernaculars, but had also been changed into kela and again shortched to er. This is impossi ble, because it is unnatural and against the genius of language. Thirdly, keraka is nowhere a sort of affix. If we style kerakº an affix, we must do the same with innumerable other adjectives.

This, however, is no reason

tion is not how often keraka occurs, but what

changes it may have undergone. If every word of frequent occurrence were curtailed to one syllable, our language would soon resemble the Chinese language. It is due to the uncritical editions of Sanskrit plays by the Indian Parlits that the word is not met with oftener in other plays. In the Śakuntalà I shall restore it in three more pas sages where the best manuscripts have it, though it is not found in any of the present editions of this play. The first instance which I quoted from the Śakuntalá is not a false one; keraka is used there pleonastically ; it could be omit ted very well. The second instance is not in the least doubtful, but as certain as anything can be. Monier Williams is no authority, his edi tion—apart from its being a pons asinorum— being founded upon the worst possible manu scripts. I gladly recognize the superiority of Prof. Hoernle in every otber respect, but as for the Śakuntalà I must lay claim to know a little more about the play than he, having collated,