________________
176
THE INDIAN ANTIQUARY.
[JUNE, 1892.
We can also assert that the sign for the vowel ri did not yet exist in the time of Asoka. The reason is simple, and is quite independent of any à priori argument. It is clear to every one that the sign J of the vowel ri, in the most ancient form in which it appears, is derived from the sign used to mark r in composition with a preceding consonant, viz., J; and we have just seen that this sign did not develop till after the time of Piyadasi.
Another lacuna is more significant still; it is the absence of three distinct signs corresponding to the three sibilants of the learned orthography. I am now speaking only of the Indian alphabet. Khâlsi allows us to show that this absence was perfectly real, and that it was neither voluntary or merely apparent.
It will be remembered that Khâlsi, in addition to, the ordinary sign for s, also employs another form, . This s has been considered as representing the palatal . It is true that this last letter has an identical or analogous form in the most ancient inscriptions in which it appears, i.e., at Nasik and at Girnar. But we must understand matters. It is not possible to admit that, at Khalsi, the first edicts and the last ones differ between themselves in dialect, and I consider that the conclusions to which I came in the Introduction of this work are unassailable, that, at Khâlsi, is merely an alternative graphic form of . Other facts confirm my opinion. The sign reappears in the Edict of Bairât, and in the two inscriptions of Ramnath, the first presents only a single example, in the word starga, in which the palatal é has no right to exist. The inscriptions of Râmnâth are, unfortunately, either badly defaced or very badly reproduced. Such as we have them, they do not lend themselves to a translation, or even to an approximate interpretation; all that we can remark is that the first uses the sign and that alone, and the other sign d and that alone. This is a very strong reason for considering that the two signs are simple equivalents. The demonstration is completed by facts drawn from the other end of India. Mr. Rhys Davids (Ind. Ant., 1872, p. 130) was the first to point out, in the most ancient inscriptions of Ceylon, the parallel use of two sibilants and A. The second is clearly only a modification of the of Khálsi or of its prototype. Since then, Dr. E. Müller (Ancient Inscript. of Ceylon, No. 1) has published one in which the form alone figures. He has drawn from these facts (p. 16) the only reasonable conclusion, that which Mr. Rhys Davids had already very justly put forward, that the two signs express indifferently one and the same sound. We cannot come to a conclusion for the north different from that to which we have come for the south. The distance between the two localities of occurrence, and the absolute analogy of the facts prohibit us from thinking of a dialectic differentiation between the two sibilants. The sequel of paleographic history shows us that the form came to be subsequently employed to express the palatal é, when a need to express it, that is to say, to write in Sanskrit, was experienced. At the time of Piyadasi, the Indian alphabet did not yet possess the palatals; and it therefore had not yet been applied, in anything like a regular and consequent way, to the learned language.
-
Another strictly parallel fact indirectly confirms this. By the side of c, the inscription of Khâlsi, in its second half, frequently uses a form . Dr. Bühler (p. 26) transliterates it by sh, and approves of my having recognized its relationship with the cerebral sh of the complete alphabet. I fear that there has been a misunderstanding here. I do, it is true, believe that the of Nasik and of Girnar (Rudradaman) is derived from this , but I in no way believe that this last form had the value of a cerebral at Khålsi. In spite of the transliteration sh, I would not venture to assert that such is even the opinion of Dr. Bühler, and in any case I could not agree with him if it is. The sign does not appear till about the 10th Edict, and only becomes common in the 11th, 12th and 13th, although the form is not absolutely unknown to the former ones, as we have it also in the 4th Edict, 1. 11. In the more than 110 instances in which Dr. Bühler reads sh at Khâlsi, there are only thirty in which the cerebral sh could be expected. Under these circumstances, and the transition between the forms du and being easy, the steps being marked out.by several intermediate shapes both at Khâlsi and elsewhere, it is absolutely impossible to consider the sign as anything other than a graphic variant
