Reviews Of Different Books

Page 34

280 REVIEWS Finally, Rona-Tas' letter again directs my attention to a number of important problems, each too involved to be considered here in anything resembling a complete manner, but concerning which I would like to comment briefly as follows: 1. In connection with my remarks (pp. 154-5) on the possibility of influences by way of non-NE Tibetan loans, Rona-Tas writes as follows: "In Buddhist terminology learned borrowings and in everyday business life the Tibetan koine had surely an important influence - not to speak about pilgrims. I was surprised, nevertheless, (to find in my study) that most of the Monguor forms showed neither literary nor Central Tibetan or Lhasa items as their immediate source." This is an important point, and very well-taken; and the remarks in my review (which still, I think, are of value) should certainly be read with this statement in mind. 2. In connection with my remarks (pp. 156-7) on the problems presented by the labial semivowel x in Monguor loanwords from Tibetan, Rona-Tas expresses himself as remaining unconvinced by my arguments, and suggests that "we have here to do with an areal phenomenon". This seems to me a most important suggestion, and one that should be followed up by future students of the problem, though I still do not see why the development of a labialized velar phoneme from an earlier labial-plus-velar sequence should present any particular difficulties of interpretation. 3. In connection with my remarks (p. 158) on vowel-harmony assimilations in Tibetan forms underlying Monguor loans, Rona-Tas writes as follows: "The possibility of reflecting Tibetan vowel harmony features by Monguor forms is made difficult by the fact that vowel harmony is an essential part of the phonological system of any Mongolian language, so naturally also of Monguor. The so-called 'breaking' can also be taken into account in such cases as araDag, murGuo, suro, etc." The comments in my review should have paid more considered attention to the implications of the Mongol phonological canon than they did, and the point that Rona-Tas makes is of importance in any future consideration of these issues. 4. In connection with my remarks (p. 159) on the importance of the problem of morpheme-boundaries in Tibetan compounds, Rona-Tas points out that his etymology #285, Mgr. lisGa 'work, deed, custom' can only be explained in the light of a morpheme-division different from that of WT las-dka', var. las-ka, i.e., as li-sGa< *la(s)-ska, an important example that had until now escaped my attention. But on my suggestion with regard to the morpheme-segmentation of WT rdo-rje 'vajra-sceptre', Rona-Tas writes, "I do not think that a segmentation rdor-je can be justified." Since the suggestion that the forms in the living Tibetan languages point in the direction of an earlier rdor-je, rather than toward the segmentation of the form in the received orthography, WT rdo-rje, is one that I have tentatively put forth in the literature several times since 1955 (references in my review, p. 159), a few additional lines here in support of such a view may not be out of order. The problem is, to be sure, essentially concerned with the internal morpheme boundaries of the form, in historical terms, but it does not stop there, since the initials of each of the two morphemes, and especially of the first, are equally critical in the problem. Briefly, the most important considerations that enter into the question are the following: (a) If the Tibetan form that historically is responsible for Mgr. Duor Dzi, var. Dor DZi 'vajra-sceptre' is correctly represented by the received orthography of WT rdo-rje, either the initial of the first morpheme, or the initial of the second morpheme, must be anomalous in terms of its historical development. In the Monguor loans, original Tibetan r- before -C- either is retained as such, or otherwise appears in transparent guise as s- or s-: if we center our attention solely on forms with original initial rd-, and -o- vocalism (and hence exactly parallel to the phonological configuration of the form at issue), we find WT rdo/,, Mgr. $D- (Rona-Tas, p. 91), WT rdo-'bum, Mgr. reD- (p. 77), WT rdog, Mgr. ar D- (p. 40), WT rdon-rgan, Mgr. rD- (p. 75), and WT '/dor-mo, Mgr. $D- (p. 91). This is why Rona-Tas writes, "OT r-- [appears as)