In lieu of an abstract, here is a brief excerpt of the content:

BOOK NOTICES 479 from the sentence types they tested—the idea is amply refuted in the literature. In general, fuller knowledge of how the limited set of English contours is spread across syntactic domains would have cast some doubt on C&S's hypothesis that a simple boundary strength metric can predict the steepness of F0 curves. One study in this chapter answers affirmatively the question of whether remote syntactic structure influences intonation assignment; as one's ear confirms, there is a steep pitch-fall on the word directly before a gapped constituent. I only wish that C&S had also tested less audible contours; it would be fascinating to find out, e.g., whether gaps from WH-movement also affect F0. Chap. 4 raises an interesting question but answers it only superficially. C&S reconfirm and refine the hypotheses that contrastive stress on one word causes a compensatory lowering of F0 on the next, and that a voiceless consonant causes a fall in Fo during the first 50 msec, of the following vowel. The question is, what syntactic distance between the determinant and the word containing the focus will block these effects ? In compensatory lowering, C&S determine that the effect will not operate between sentences joined by a coordinate conjunction, but they do no follow-up work to find out just how strong the boundary must be to block the effect. The results of the voiceless-consonant study are disappointing to C&S; the small magnitude of the F0 fall is overshadowed by rises in F0 at the beginnings ofphrases. Still, with the cleverness of experimental design which C&S demonstrate in the declination studies, it seems they could have overcome this obstacle; e.g., a comparison of the first 50 msec, of vowels following voiced vs. voiceless consonants at various syntactic boundary types might have yielded significant results. The first chapter surveys the literature on physiological control of F0; and Chap. 5 discusses the relevance of Fo studies to a wide range of topics—not all of which are clearly linked to the main subject matter of the book. [Ellen M. Kaisse, University of Washington.] Harvard studies in phonology, vol. II. Ed. by George N. Clements. Bloomington: Indiana University Linguistics Club, 1981. Pp. ?, 427. $8.75. This collection often papers, largely by Harvard faculty and graduate students, is remarkable for its consistently high quality. Even those few contributions not of immediate interest to the general reader are bound to be useful to specialists in the languages treated; Slovak, Serbo-Croatian, Akan, Tonga, Tubatulabal, Zulu, Mohawk, Turkish, and the Bongo-Bagirmi group all receive detailed attention. The papers display a not surprising emphasis on autosegmental (henceforth a-s) phonology, particularly as it relates to the treatment of African tone languages —this being the forte of the editor, who is also the teacher of several of the contributors. Perhaps the most theoretically interesting of the papers is by Clements himself. In 'The hierarchical representation of tone features', he argues for a highly restrictive yet predictive method of describing tone systems, using only the relative features h and /. He presents a lucid account of what downdrift is, reduces downstep to a kind of downdrift, and explains the universal restrictions on its occurrence. A very different but equally impressive paper is E. Sezer's 'The klH alternation in Turkish'. This is the most intelligent and thorough discussion I have seen of the long-standing abstractness debate on the existence of morpheme -final g in Turkish. S resolves the issue once and for all in favor of the more concrete analysis, with an elegant combination of empirical and formal argumentation. K. Michelson's 'Stress, epenthesis and syllable structure in Mohawk' examines the phenomenon used so controversially by Postal to argue for absolute neutralization, and shows that epenthesis is in fact a set of several rules. It is argued that, in the cases where it fails to separate kw clusters, a syllabification where lew forms a syllable onset is independently required. Clements' second paper, 'Akan vowel harmony : A non-linear analysis', gives the most persuasive argument for an a-s treatment of vowel harmony that I have encountered. Skeptical readers may still ask, however, if the 'universals ' of harmony captured by the a-s treatment , such...

pdf

Share