Reviews Of Different Books

Page 35

REVIEWS 281 ar- in the first syllable [in a single form]: $- before non-palatal or non-palatalized radicals in every position ...; [and] - before palatal or palatalized radicals ..." (p. 138). But the Monguor form for 'vajra-sceptre' has only initial D(u)s, not *reD-, or any other overt indication of an original r- before -C- in its representation of the initial of this form. (b) Nor do any of the dialect forms that can be cited for 'vajra-sceptre give any hint of an original r-before-C-in this word. Those forms that can be cited, and that usually are given (as for example, in Rona-Tas, p. 45, #93), are not dialect reflexes of the form 'vajra-sceptre'; they are dialect reflexes of the word for 'stone', e.g. Bal. rdoah, Pur. rdoa, Lad. rdoa, all 'stone'. They go, of course, with WT rdo 'id.'. The same is true, it is worth noting, of the fifteenth-century linguistic data represented by Nishida's 'Seibango A', where WT rd- before -o- is regularly and uniformly represented in the pronunciation-transcriptions into Chinese characters in the Paris - Societe asiatique Ms. as [rd], with the single exception of the form 'vajra-sceptre', where we find anomalous [d]: rdo # and page in Nishida WT #53, p. 83 #905, p. 118 rdo(-mthin) #483, p. 100 rdo (rin-po-che) #936, p. 120 rdog(-cig) #283, p. 92 rdo-rje Pronunciation-transcription [rdo] [rdo] [rdo] [rdo] [do-rd3e] (c) On the basis of these data, either of two equally satisfactory conclusions is possible: (a) the etymological identification of the first morpheme in the Tibetan word for vajra-sceptre' suggested by the received Tibetan orthography, by which this first morpheme is taken to be identical with the word for 'stone', is not, in historical fact, to be maintained; or, (B) rdo-rje came into Monguor, and into the 'Seibango A' language, and so also into most (if not all) the modern Tibetan dialects (resp. languages), not as a regular genetic inheritance but rather as a loan from some earlier dialect, now otherwise apparently lost, in which the r-element in the initial rС- had been simplified to zero; but this would necessarily have been a dialect quite different historically from that responsible for the bulk of the Monguor loans. (d) But in the meantime, and regardless of whether (a) or (1) above is elected, the phonological canons of the modern dialects known to me, and particularly of modern Central Tibetan and Lhasa, determine that a historical development from a rdor-je segmentation is regular, while a development from the received rdo-rje segmentation is anomalous. (e) Even apart from all this, that a sequence rdor is something more than an abstraction arrived at solely on the basis of linguistic evidence, may easily be demonstrated by a consideration of the shape of the morpheme WT rdor, which appears as a combiningform for rdo-rje, as for example in phyag-rdor 'vajrapani', the compendious equivalent of Mhv. 649, phyag-na rdo-rje 'id.'. Nor may it be suggested that this is a mere lexicographer's ghost; for in fact, Rona-Tas (pp. 45, 95) twice cites yaraypu 'id' from the Tangutsko-Tibetskaja okraina, II, of G. N. Potanin (where it appears in text J, 'from an unnamed Tangut'). A simplification phyag-na rdor-je > phyag-rdor is regular; phyag-na rdo-rje > phyag-rdor is anomalous; hence, I opt for the former. This, incidentally, remains true whether we think in terms of an original rdor-rje > rdor-je, or an original rdor-je with no prior assimilation and simplification -r-rj-> -r-j-. And so both varieties of evidence, linguistic and philological, point in the direction of final -r in the prior morpheme, rather than an open-syllable in -o with initial r- in the next morpheme. (f) Finally, on this important word, I would in addition be tempted to point out