________________
150
THE INDIAN ANTIQUARY.
(MAY, 1892.
a former pronunciation, dharma, and not dhrama, rarea and not vrasa &c. If people said arvu, why should they have said pruva ?
We should doubtless be glad to discover with certainty the cause of these inconsistencies; but oor hesitation in this respect proves nothing against corclusions, which appear to me to be Hatisfactorily proved. It is no use counting all the variations in the mode of writing : by the side of sarva, we frequently have sava; mita beside mitra, puta beside putra, &c., &c.; we find written Icirti, and vadhati, vadhita, &c. It is not surprising that, in an orthography which is the arbitrary imitation of a learned pronunciation, a certain approximation should have appeared sufficient. The example of Girnar proves that we must not take the phonetic value of the signs too strictly. It is clear that in d and in the same character 6 signifies at one time, vra, and at another time rra. Reasons of graphic convenience may have had their share of influence. A cursive sign for r following a consonant bad been fixed at this epoch but they had not fixed one for an r preceding one. It is easy to prove this in the more recent inscriptions. Tiney retained the first sign, and invented a new one for the second case (of. the inscription of Suë Vihar.) The direct combination of the characters and 7, 4 and was sufficiently easy and symmetrical, but the combination of with u; 3 &c., being more complicated, gave greater opportunities for confusion. Without doubt such a consideration can only have been a secondary one, but the special conditions under which, as I have pointed out. this spelling was applied with its etymologioal tendency, are precisely such as to make its action admissible. They rendered much less urgent both the invention of a new sign, and the use of compound letters which might be awkward to engrave.
We are thus led to recognise in certain cases a graphic method, which not only does not faithfully represent the real pronunciation, but which in endeavouring to approximate itself to etymological writing, treats it with a certain amount of freedom. This forms a very useful basis from which to judge, what is, in my opinion, a more difficult case. I refer to the groups kod, and at Girnar, regarding which I regret to find that my conclusions did not meet with the concurrence of Dr. Pischel. This difference of opinion renders it necessary for me to completo the observations outlined on pages 20 and 29 of the Introduction to Vol. I. of the original work.
It is quite clear, as Dr. Pischel allows, that the appearance of the group & will not help us to decide between the transcription pta, and the transcription tpa. Every one agrees in reading & 8! and ck st. The exact position of the sign is therefore irrelevant. All the more bas the question embarrassed the various commentators, and they have successively proposed various readings. The arguments invoked in favour of pta are far from convincing me. I cannot admit that the form appá for átman presupposed an intermediate aptá. The group pt regularly gives tt in Prakpit, as in gutta. It is to which gives pp, as in uppala. Now apa is the very form which the most modern inscriptions of the west, near Girnar, regularly give us for átman, and I do not think that any one would suggest a pronunciation aptá as necessarily intermediate between átmá and attá. It is in the same way that chattáró is derived directly from chatcáró, like satta from satva, and atta from atva for átmá. If, under the influence of 1, the v of atvá can have become a p, the same phenomenon is equally explicable in chatpáró for chatváry, and árabhitpå for drabhitvá.
For example rya. We catch, I think, this new notation in course of formation in instances such as the Bigo 4. rkhé (arkheviyasa) of the coins of Archebios (cf. Sullet, Die Nachf. Alexanders, p. 113).
Note by translator. The following extracts from the statistical portion of this chapter, previously published, see ante, pp. 3 and 10. will assist the reader in following the argument.
GIRNAR.--em becomes tp in upas, XII,3,4,5,6.
tv becomes to alichapi, XIV, 6: Grabhitpa, I, 3; chatpiri, XIII. 8: darayitpa, IV, 1; hitalpa, VI, 11 ; pievit fitpd, 3, 4; tad ltpany, x, 1; it becomes t in satiyaput II, 2.
KAPUR DI GIRI.--tm becomes t in ata, XII, passim. te becmest.
