________________
128
EPIGRAPHIA INDICA.
[Vol. XIL
for. They are clearly not essential, for they do not occur in the shorter inscriptions? Perhaps they are merely ornamental. D2 has little in common with the other inscriptions and it is engraved on the bottom of its urn. The other urn inscriptions are engraved horizontally round the several urns.
The four inscriptions A, B1, C and D1, are all of one type and I propose to make a detailed comparison of them here. Their resemblances and differences will probably turn out to be matters of importance. For, be it remembered, " Pya" is a language of which as yet only a very small number of words have been identified, and when one is invited to decipher and interpret inscriptions in it which, unlike the Myazedi one, are not accompanied by translations in other langunges, one finds oneself face to face with the difficulty of not knowing how to begin or where to seek for clues. It seems to me that our best chance of interpreting these records is to ascertain what is essential or "common form " in them so as to be able to distinguish it from what is accidental or individual. Every new record of this class that may turn up in the future will help us to draw this important distinction. Then, when we are tolerably certain of the general intent and parport of the essential words, & comparison with the known Tibeto-Barman languages ought to give us clues to their exact meaning. Bat we ought first to be fairly clear as to the sort of meanings that we should look for.
This is particularly necessary in the case of quasi-monosyllabic languages, where there are always a number of words that have several distinct meanings in different contexts, an inherent ambiguity which is only imperfectly met by differentiation of tone. "Pyu " appears to fall into this class. It is not strictly monosyllabic, but largely so, and it apparently rejects final consonants altogether, thus immensely reducing the possible number of its syllabic combinations. I am still of opinion that the dots or little cirolos resembling antsvāra, visarga, and their combinations, used in the “ Pgu" script, represent tonal marks. If that is correct, the" Pyu " tones must have numbered half a dozen or more. In any case it is necessary to reproduce these diacritical marks in our transcription, or we should be mixing up quite a number of distinct words.
There are other difficulties in connexion with these inscriptions. They are, it is true, engraved for the most part in clear and fairly well preserved characters of the same archaic, South Indian type as the " Pya" text of the Myazedi inscription, and most of the letters are easily recognizable. But some of them are only doubtfully identified as yet. There appear to be several that resemble one another rather closely and are difficult to distinguish, particularly those which in the Myazedi inscription I have provisionally read as d, d, and I (and there may possibly be a ţ and l amongst them also). Further the compound aksharas are not always easy to decipher, the subscript forms of the letters being different from the isolated forms and by no means easy to identify. Also there is a strong resemblance, amounting almost to identity, between the lower portions of the letters k, r and subscript 1. Accordingly the transcripts which I now propose to give must be regarded as tentative and subject to such further correction as subsequent enquiry may show to be necessary. To emphasise this point I put into parentheses euch letters as I consider doubtful for want of certainty of identification. Square brackets, on the other hand, will serve to indicate places where the reading is conjectural because the stone has suffered damage. In order to show clearly the points of resemblance and difference amongst the four inscriptions I place the corresponding words directly in the same vertical lines. The actual text of each of these four inscriptions and also of B2) begins with the three paragraph marks which apprar at the beginning of the Myazedi insoription.
1 They seem to occur sporadically in No. 8 and, to a small extent, in the Myazedi inscription,
