________________
THE INSCRIPTIONS OF PIYADASI.
MAY, 1892.]
that
I think that this comparison, unless I am much mistaken, goes directly contrary to his conclusion. Modern forms like graha, grika, mraga, mranga, mriga, by the side of which we also find others such as mirga, &c., are in no way direct derivatives of the Sanskrit. mṛiga, but are tateamas; that is, nothing but simple equivalents of the form mṛiga, griha, which itself is also used in the modern languages. They are only instances of such approximate spelling as could be realized with the elements really existing in the popular language, instead of borrowing from the learned language a special sign, corresponding to a special pronunciation which has ceased to exist for more than two thousand years. In both cases, the situation is not only analogous, but is identical. I offer for both, one and the same explanation, which is incontestable for the more recent one: in mruga, graha, dridha, vrachha of the inscriptions, I can see, as in mriga, graha, dradha, vraksha or crackha of existing languages, only tatsumas, loans really taken from the learned language, but represented by an orthography which, by the absence (whether voluntary or not is of little importance at the present stage of the inquiry) of the sign for the vowel ri, was condemned to tentative and approximative devices. These examples in no way argue against my method of treating the groups dhr, pr, &c., in the words which I have quoted. On the contrary, they present certain precedents of a return towards the learned language, operating even at the price of imperfect orthographical expedients. It is exactly in the same light that we must consider the spellings which now occupy us.
149
In the first place, the state of affairs at Kapur di Giri, so far as concerns consonantal compounds including an r, strongly resembles that which we have established for Girnar. We find there pati beside prati (also prati and patri), save, savatra, by the side of sarvé, sarvan, sarvatra, &c. Without attempting to compile exact statistics, the fact is, in a general way, indisputable. It is natural to deduce from it the same conclusions as those to which we have come with regard to Girnar. We must not, therefore, treat the orthographical peculiarities of this language with absolute rigour. If the r in the words which we are discussing, is taken from the learned language by an arbitrary artifice of writing, why should we be astonished that the writers should have allowed themselves some liberty in the manner of representing it, when they have just as often taken the liberty of omitting it altogether? In Hindi the spellings dharama, karama, gandhrava, in no way correspond to any peculiar phonetic phenomena, but are merely equivalent modes of writing the tatsamas dharma, karma, gandharva.
Mr. Beames (Compar. Gram. I, 321) has quoted in the ancient Hindi of Chand, spellings such as érabba (= sarva), dhramma (= dharma), sovranna (suvarga), brana (= varna), brannaná ( varnaná), prabata (parvata), kramma (= karma), krana (: karna), &c. I do not think that these examples can be appealed to against the argument, which I here maintain. It is more than clear that all these spellings were, at the time of Chand, loans taken from the vocabulary of the learned language. The doubling of the consonant in irabla, kramma, &c., sufficiently proves that the true pronunciation of the people was sabba, kamma, &c. Different motives, metrical or otherwise, may have suggested these spellings, but they prove nothing as to the real pronunciation. Far from being contrary to my opinion, they supply, at a distance of some fifteen hundred years, a phenomenon, strictly comparable with that which we have shown to exist at Kapur di Giri. This resemblance of methods is explicable by the resemblance of the conditions which called them into being. In each case we have a language, which, not having as yet a regulated system of spelling, attempts, with groping and uncertainty, to approximate itself, by the simplest means available, to the practice of a language which enjoys a higher degree of reverence.
If we consider the facts by themselves, would this change of dharma to dhrama, of purva to pruva, of karma to krama be likely or probable? I think not. Alongside of pruva, there is at least one passage (VI, 14), in which it seems clear that we must read purva. So also we find that coins wrote varma alongside of dhrama; that by the side of drasana at Kapur di Giri, we have, at Girnar, an example of darsana. The form which all these words have invariably taken in the popular pronunciation, dhashma, puvva, kahma, vassa or vása, &c., depends uniformly on
