________________
THE EASTERN BUDDHIST
However, in his reaction to the practice of editors who have Sanskritized their texts Edgerton has sinned in the opposite direction. Edgerton admits the genuineness of 3rd plural, optative and aorist forms in -itsuC”) and -etsu(b) because they occur very often in the manuscripts of the Mahāvastu. Brough and Regamey are undoubtedly right in rejecting the evidence of the recent Nepalese manuscripts in this case. There is no doubt that Edgerton's Grammar contains many forms for which the manuscript evidence is slight and doubtful. It will be necessary to verify in each case whether the manuscript readings can be accepted as such or whether a different reading must be assumed. Let me quote one example which has been discussed by Břough. In the Lalitavistara one finds anyatra karma sukrtāt (37:7). In 8.9 Edgerton explains karma as an abl. of an a-stem resulting from a shortening of -ä(t) metri causa. In 17.13 he proposed an alternative explanation as a stem-form. Brough prefers this explanation. However, if one takes into account the context: na ca samskste sahāyā na mitrajñātijano ca parivārāḥ/anyatra karma sukstād anubandhati prşthato yāti, it is obvious that anyatra is here not a preposition but an adverb meaning "on the contrary, only" (cf. Edgerton's Dictionary s.v. anyatra). The original reading must have been anyatra karma sukrtam. A misunderstanding of the meaning of anyatra has led to the transformation of sukrtam into sukstād. Edgerton has pointed out that a syllable ending on an anusvāra before a vowel is used metri causa in order to obtain a long syllable. In his critical examination of Edgerton's view Helmer Smith prefers to speak of metrical doublets: for instance -am, -ām or -amm before a vowel instead of -am. Edgerton's assumption of lengthening and shortening of vowels because it is required by the metre has been rejected by Nobel with reference to Smith's article. Edgerton has replied (JAOS, 77, p. 187) by stating that “Smith thought that such changes should be recognized only when there was some historic, phonological or morphological “justification” for them.” I believe this does not reproduce Smith's opinion quite adequately. In Les deux prosodies du vers bouddhique Smith admits lengthening of a short vowel at the end of a pāda, of an initial vowel preceded by a prefix (an-ābhibhūto) and shortening of -e to -i, -ā to -a, -am to-U, -o to -u. The principal point of difference between Edgerton and Smith is that, according to Smith, Middle Indic orthography admits a short vowel before a caesura where metrically a long vowel was pronounced, for instance the fifth syllable of a trişțubh-jagatī, and also in other places where the metre
54