Phonological segments: Their structure and behaviour By Helga Humbert (review)

422 LANGUAGE, VOLUME 74, NUMBER 2 (1998) changes are much more apparent in the index (itself much extended and improved) than in the table of contents. Despite very similar chapter and section headings, comparing index entries across the two editions for 'sex differences', 'space (multidimensional )', 'solidarity', and 'network(s)' gives a good sense of certain major changes in emphasis in the new edition. The first two of these topics are not represented at all in the index of the first edition (1980), while the second-edition entries for the last two topics go well beyond those of the first, with new subentries as well. H argues in the second edition, as in the first, that language is located in the individual rather than in the community, but his argument is strengthened this time around by proliferating research on speech accommodation and on social networks. The small social network seems on the whole to predict similarities in linguistic systems better than other social constructs do. in particular the indeterminate and unbeatable (in H's view) 'speech community', while accommodation theory can explain otherwise puzzlingly partial and selective uses ofdistance-reducing (or for that matter distance-increasing) linguistic variables. H mounts strong objections to reliance on group scores in the quantitative study of linguistic variation, as has been the practice in the Labovian variationist tradition He points out that the notion of discrete social groupings of the sort that social classes represent is ever more difficult to accept as speech accommodation research and social network research point to more fundamental and more fluid alignments in social interaction and linguistic differentiation —what H, following R. B. Le Page and Andr ée Tabouret-Keller (1985. Acts ofidentity: Creolebased approaches to language and ethnicity, Cambridge : Cambridge University Press.), term acts of identity in multidimensional space. An important and entirely new 'theoretical summary ' chapter concludes the second edition. In it, among other things, H identifies a major consequence arising from different approaches to locating linguistic structures and their social functions: If language is located in the individual, differences between individuals are to be expected, and the similarities require explanation; if language is located in the community, similarity is to be expected, and individual differences require explanation. H's book has a clear point of view and coheres exceptionally well across all topics and chapters, offering students and colleagues alike a consistent and thoughtful exposition Because H is also very clear and very cogent about the impossibility of any artificial separation between grammar and the social aspects of language, his book would make a particularly useful text for departmental foundation courses in sociolinguistics. [Nancy C. Dorian, Bryn Mawr College.] Phonological segments: Their structure and behaviour. By Helga Humbert. (HIL dissertation 12.) The Hague: Holland Academic Graphics, 1995. Pp. xiii, 287. Paper DFL 40. This dissertation proposes a model of segmental representation and a typology of assimilation and dissimilation processes. It is divided into two parts. The first (Chs. 1-3) lays down the theoretical framework of both the representation and the behavior of segments in different phonological phenomena such as assimilation, dissimilation, and nasalization. The second (Chs. 4-7) offers empirical evidence, from different languages, for the analysis of the processes presented in it. Ch. 1, 'Theoretical considerations' (4-28), sets the theoretical basis. Humbert favors a unary feature system over a binary one Based on insights ofdependency phonology, government phonology, and feature geometry, she lays the ground for an alternative proposal of segment representation using a minimum offeatures therefore 'increasing the explanatory burden of the structure' (12). This structure is made out of three modules: stricture, location of stricture, and phonation. Since she uses a minimum of features, the notions of markedness and underspecification play a significant role and are addressed in this chapter. Ch. 2, "The representation of obstruents' (29-81), takes the obstruent-sonorant distinction to be the basic distinction over the consonant-vowel one. There are, according to H, two reasons for this. The first is that vowels and sonorant consonants are sonorants distinguished in their place component; the second is that the manner component is considered relevant in the syllabification processes. An obstruent is a segment which has a C element in...