» J 



58 Journal of the Asiatic Society' of Bengal [N.S., XIII, 



Greek transcripts ought to represent something like the Iranian 

 pronunciation of the first century after Christ. And it is re- 

 markable that this pronunciation often approaches much 

 nearer to the modern Persian than to the sound indicated by 

 the written Pahlavi, which latter seems to preserve traces of an 

 older pronunciation more like that of ancient Persian and the 

 Awesta; this is especially seen in the names Mihro, Sahrevaro, 

 and Vado, when compared with the Phlv, Mitro, Satraver, and 

 Yato. The final vowel o, which is very rarely replaced by 1 or 

 e on these coins, is used after every consonant that occurs; but 

 in MS. Pahlavi it occurs only after b, p, t, 8, k, n and g. 



Sir M. Aurel Stein has come, correctly in my opinion, to 

 the conclusion that this legend represents the Zoroastrian deity 

 Ashisvang or Ardishvang, but he hesitates in the absence of 

 more conclusive evidence, as he cannot reconcile her common 

 name with the form of the legend. He also hesitates about the 

 legend At>A6IXt>0 which he would identify, again correctly as 

 I believe, with Ardibahist, provided that any satisfactory evi- 

 dence could be found for the phonetic change of sht into 

 khsh. 



One very strong evidence in favour of the above identifica- 

 tion of Ashishvang is the representation of the goddess on coins 

 with cornucopise in her hands, thus resembling the Greek 

 Tyche, the personification of Wealth and Fortune. The Hindu 

 oddess Lakshmi is an exact prototype of Ashishvang. 



Mr. Thanawalla takes the final o in the legend as an iza* 

 fat: but this simple explanation does not help us, as an izafat 

 necessarily implies possession of something, and there is nothing 

 possessive in the phrase. 



In Pahlavi an optional o is often found following the let- 

 ters b, p, f, t, eh, j, k, and n either in the body or at the end 

 of words. Even the oldest MSS. fail to observe any uniformity 

 as regards this redundant o, but insert or omit it indiscrimin- 

 ately. This seems to show that o though inserted long ago 

 was not pronounced. European savants ive this redundant 

 letter the sound 6. Formerly it was explained as being a- 

 remnant of the Avesta case terminations. This ending in ° 

 may be a dialectic peculiarity of the Indo-Scythian race. 

 Examples are not wanting of Grecised Persian names, M- 

 Tiridates, Oromazis, Pharnabazos, Tiribazos, etc. 



Should this letter be pronounced ? Perhaps not, as we 

 have similar examples of mute letters in other living languages. 

 For example, in English we find words like know, fruit, con- 

 demn, column, calm, psalm, etc., where a letter or letters are 

 quiescent; in French the consonants at the end of words are 

 generally mute; and the Persian language, which owes its origin 

 to Pahlavi, has h mute in many words. 



Both Drs. Haug and West have agreed that this optional 



