Atomism and binding Ed. by Hans Bennis, Pierre Pica, and Johan Rooryck (review)

236 LANGUAGE, VOLUME 76, NUMBER 1 (2000) comparison, the author raises and answers the questions why the former follows the definite article and why it appears in the plural neuter form. On a broader scale, he emphasizes the importance of artificial, written languages that can evolve for diachronic studies. This collection is very useful, offering insights into modern investigations of Greek phonology, morphology , and syntax. Nearly every paper deals with a different framework in the given field, thus enriching the reader with different theories and background assumptions yet offering a variety of linguistic phenomena , plenty of data, and stimulating theoretical analyses. [Kleanthes K. Grohmann, University of Maryland.] The phonology of coronals. By T. Alan Hall. (Amsterdam studies in the theory and history of linguistic science 149.) Amsterdam & Philadelphia: John Benjamins , 1997. Pp. x, 176. Hall discusses how the various places of articulation deemed 'coronal' are distinguished from one another in feature terms. There are four main chapters. Ch. 1 (The feature [coronal]') offers a phonetic definition of [coronal] that excludes palatals. Ch. 2 ('Feature specifications for coronals') discusses featural distinctions among the various coronal places of articulation . Ch. 3 ('Coronal consonant inventories') discusses why languages use only certain subsets of the places distinguished in Ch. 2. Ch. 4 ('Rhotics') discusses r-type sounds. In an appendix ("The features and their definitions') H surveys the whole phonological feature inventory. H draws on an impressive variety of languages, and his reference list contains over 500 items. His discussion, however, is at best somewhat pedestrian, and at worst illogical. His account of palatals and palatalization illustrates his strengths and weaknesses nicely. Are palatals coronal or not? H shows that the facts cited in favor involve almost exclusively stops, not fricatives. He also cites evidence that 'palatal' stops (IPA [c, j, Ji]) are articulated further forward than palatal fricatives ([ç, ¡]). He therefore argues that [c, j,Ji] should be reclassified as alveolopalatal and suggests that ' "true" palatal noncontinuants are unattested sounds' (14-15). Nevertheless, palatal noncontinuants do need to be recognized, at least in featural terms because the palatalized velar stops [kJ, g>] differ from the palatalized velar fricatives [xJ, vJ] only in respect of [± continuant], and the latter are said to be featurally identical to the 'true' palatals [ç, j] (70-71, 76). H insists that there should be a single 'palatalization feature' (77) and proceeds to inquire what that feature is. His answer is [ — back], which is shared (he says) by all and only front vowels, palatoalveolars, alveolopalatals, palatals, and palatalized segments (83). But [ — back] means 'fronted tongue body', where 'tongue body' includes 'both the predorsal and laminai regions'. Does not this imply that any plain laminai [t] has to be regarded as 'palatalized' too? How then does H distinguish between [t] and [tJ]? H's answer involves his implicational universal 'If [ — anterior, +distributed], then [ — back]', which states that 'alveolopalatals are inherently palatalized' (51-52). On page 98 he claims that this universal entails that [ + anterior] coronals, such as [t], are not [-back], in contrast to [c], But there is a logical error here. H would be right only if the implication were reversed: 'If [ — back], then [ — anterior, + distributed ]'. This formulation cannot be what H wants, however, because it renders [ ± back] inapplicable to dorsals. H compounds the confusion by adding in a footnote that dentals and alveolare can after all be [ — back] (or [ +back]), in which case they are palatalized (or velarized). The feature [±back] thus emerges with three values: [ +back] for [tv], [0 back] for [t], and [ — back] for [tJ]. At the root ofthis trouble is H's notion of 'inherent palatalization'. What is needed is some way of distinguishing primary from secondary articulation, an issue that H never addresses . Phonologists interested in coronals should find this book useful for its thoroughness but will need to approach H's analyses with caution. [Andrew Carstairs -McCarthy, University of Canterbury, New Zealand.] Atomism and binding. Ed. by Hans Bennis, Pierre Pica, and Johan Rooryck . (Linguistic variations 1.) Dordrecht : Foris Publications, 1997. Pp. xx, 412. This collection of papers deals with explorations in the 'distribution and interpretation of definite descriptions , pronouns and reflexives' (i, italics omitted ), i.e. the syntax and semantics ofbinding. Bennis, Pica, and...