Mastering English: An advanced grammar for non-native and native speakers By Carl Bache and Niels Davidsen-Nielsen, and: Mastering English: A student's workbook and guide By Alex Klinge (review)

232 LANGUAGE, VOLUME 76, NUMBER 1 (2000) mon's 10-volume Encyclopedia of language and linguistics (ed. by Ron E. Asher, 1994, and reviewed by Bernard Comrie in Language 71:146-50, 1995). The series also includes volumes on educational linguistics , language pathology, history oflinguistic sciences , grammatical categories (?), and pragmatics, although the philosophy-of-language volume covers pragmatics quite generously. Each volume has an editorial introduction that surveys the field usefully, as well as indexes of names and terms. The syntax volume also has a 35-page glossary. New in the syntax volume are articles on the minimalist program (by Martin Atkinson), construction grammar (Adele Goldberg), andnew developments in lexical-functional grammar (Louisa Sadler). Of the 50-or-so articles taken from the Encyclopedia, one has been completely replaced, and all but two have been revised and expanded. The philosophy volume presents 152 articles with somewhat less updating. They cover a wide range of subjects, including semantics, pragmatics, and logic. It's all here, from Plato to Grice to Zadeh, from propositional attitudes to generalized quantifiers, and (in the syntax volume) from stratificational grammar to HPSG. Books like this make it obvious just how much linguistic research has been done dunng the past three or four decades. Linguists of all ages—those who remember the developments chronicled and those who don't—will enjoy browsing through these volumes and filling gaps in their knowledge. [Michael A. Covington, University of Georgia.] Discontinuous constituency. Ed. by Harry Bunt and Arthur van Horck. (Natural language processing, 6.) Berlin & New York: Mouton de Gruyter, 1996. Pp. viii, 349. Discontinuity is a fundamental problem in syntax. If close semantic or distributional association between syntactic units always conesponded to adjacency , we could get by with little more than contextfree phrase-structure rules, and the order of rule application would map perfectly onto the observed structure. In real life, closely associated elements are not always together. Discontinuity provides the strongest argument for transformations (such as whmovement ) as well as a steady supply of challenges for parsing algorithms and semantic analysis. Of the fourteen papers in this volume, one particularly stands out. In 'Toward a minimalist theory of syntactic structure' (11-61), David R. Dowty argues thatChomskyanminimalismisnotminimalenough; it still assumes tree structure. A truly minimalist theory postulates only the syntactic groupings that are actually motivated; the default structure is an unordered multiset of words (multiset, rather than set, because the same element can occur more than once). Dowty distinguishes between tectogrammatical structure (the sequence or grouping of rule applications in a generative description) and phenogrammatical structure (the structure of the phrase or sentence itself). He points out thatmany classic arguments for constituent structure are only tectogrammatical . For instance, the structure [Vp could [vp have [Vp been [VP slicing . . . ]]]] is motivated by related alternative sentences—evidence of tectogrammar—rather than the properties of any single sentence. Likewise, compositional semantics is tectogrammar, not phenogrammar. Dowty's tentative system comprises categorial combination rules; linear precedence relations (as in GPSG); bounding categories (which define phrases that must be contiguous, such as S); two additional, marked, syntactic operations (ordering and attaching ); and government and agreement mechanisms like those of categorial grammar. He presents brief analyses of phenomena in Finnish and English. I consider this paper significant because the isomorphism of tectogrammar and phenogrammar is one of the most fundamental (and, nowadays, unquestioned ) assumptions of Chomskyan linguistics. Many phenomena that raise difficulties for generative analysis, such as free word order, are difficult precisely because they challenge this isomorphism. Without toppling the Chomskyan paradigm, Dowty has clearly opened up a new class of alternatives. Several of the other papers demonstrate the close relationship of theoretical syntax to computer science . Harry Bunt, 'Formal tools for describing and processing discontinuous constituent structure' (62-83), builds some pretheoretical infrastructure. Patrick Saint-Dizier, 'Expressing discontinuous constituency in Dislog' (85-98), extends the programming language Prolog to parse discontinuous structures by performing inferences in unspecified order. Several other papers deal with formal semantics . Finally, a descriptive paper by Norbert Corver ('Discontinuity and the wat voor construction', 165-79) analyzes a Dutch idiom that is commonly discontinuous. [Michael A. Covington, University of Georgia.] Mastering English: An advanced grammar for non-native and...