Phonétique historique du mycénien et du grec ancien By Michel Lejeune (review)

738LANGUAGE, VOLUME 50, NUMBER 4 (1974) complete Greek lexicon? Also, the material is edited with care.3 But one cannot avoid feeling disappointed with the dubious results caused by the neglect of elementary and time-honored principles of historical linguistics. To a large extent, the methods used and their outcome cannot be subjected to any falsification test. So the principal usefulness of this book is as a storehouse of raw data, supplementing the etymological dictionaries. However, I should note that F himself often qualifies the significance of his assertions by expressing doubts, particularly in etymologizing proper names. This is much to his credit in these doubtful realms of substratum linguistics, and distinguishes him from other researchers who seem to believe in nearly everything they assume they have discovered. Here one is better off consulting Furnée's compendium. REFERENCES Chadwick, John. 1970. Greek and Pre-Greek. TPS 1969.80-98. Gindin, L. A. 1967. Jazyk drevnejsego naselenija jugobalkanskogo poluostrova: fragment indoevropejskogo onomastiki. Moscow: Nauka. Hester, D. A. 1968. Recent developments in Mediterranean 'substrate' studies. Minos 9.219-35. Heubeck, Alfred. 1961. Praegraeca: sprachliche Untersuchungen zum vorgriechischindogermanischen Substrat. Erlangen. Kuiper, Franciscus B. J. 1968. Pre-Hellenic labio-velars ? Lingua 21.269-77. Phonétique historique du mycénien et du grec ancien. By Michel Lejeune. (Tradition de l'humanisme, 9.) Pp. xi, 398. Paris: Éditions Klincksieck, 1972. F 56.00. Reviewed by Brian Newton, Simon Fraser University This work is based closely on the author's excellent Traité de phonétique grecque, the first edition of which was reviewed by Roland G. Kent in 1948 (Lg. 24.195-8); the main changes involve the incorporation of Mycenean data. The Linear B syllabary reflects the survival of labio-velar [kw] {i-qo 'horse') and of [w] {we-to 'year'), but the failure of its designers to make the radical adaptations needed for the unambiguous representation of Greek greatly reduces the potential linguistic value of the available material. One need only mention the confusion of r and /; of voiced, voiceless, and aspirated stops (except for d, t); and of long and short vowels—plus the frequent omission ofsyllable-final consonants {ko-wa for [korwa:] 'girl'). Since many of the phonological processes which we would like to date and formulate involve precisely the features unrepresented in the script, shaky deductions often replace direct inspection. Intervocalic [s] was weakened in Greek to [h], and this was metathesized over an initial vowel; thus we would like to know how far the Myceneans had carried the word meaning 'priest', for which initial *[is] is hypothesized : here the variant spelling i-e-re-u for usual i-ye-re-u with normal transitional glide may indicate the [iherêws] stage rather than [hierêws] (p.95)—or it may not; and so it goes. 3 Even copying and typographical errors are rare, e.g. the constant funny replacement of the scholar's name Knobloch by Knoblauch (German for 'garlic'). REVIEWS739 The classical reflex of *ky and *khy in the better-known dialects was [ss] (Attic [U])—and of gy, [zd] ; it is particularly frustrating not to know the value of the corresponding Mycenean consonants (conventionally transliterated as z). Lejeune favors an affricate pronunciation. If [ts] is assumed for *ky, we apparently have a case of later re-ordering (counter-feeding to feeding), since [ts] was presumably an intermediate stage in the development of *ty, which is almost certainly [ss] in Mycenean. Thus in Mycenean the de-affrication of [ts] from *ty preceded the rule which created [ts] from ky. That *su:kia: 'fig tree' was spelt su-za might suggest, though, that ? represents rather a palatal [k'] (for *ky and k"y) or [g'] (for *gy). In any case, the problem remains of accounting for the divergence of the vast majority of items with original *ty, which show [ss], and the small group containing [tósos] 'so much' which undergo degemination (optionally in Homer). We may note that, in the latter group, the determining environment would generally be synchronically intramorphemic, while in the first, [ss] alternates with morpheme-final [t] ([mélitos] 'honey' (gen.), [mélissa] 'bee' [ra:], [re:] > [re:], [rwa:] > [re:], [rea] > [re:], [ewa:] > [ea:], [eea] > [ea...