Semantics (review)

206 LANGUAGE, VOLUME 75, NUMBER 1 (1999) sentential switching are all problematic in that data can be found to refute claims that have been made about them. Boeshoten further claims that the role of codeswitching in language change has been neglected . Abdelali Bentahila and Eirlys E. Davies ('Codeswitching: An unequal partnership?' 25-49) continue the examination of claimed universals, mounting a serious critique to Carol Myers-Scotton's matrix model which, they claim, does not provide for situations of asymmetry in the roles of the languages involved. They give examples where, by one definition ofmatrix language (providing the larger number of morphemes), French is the clear matrix language, yet, contra another claim ofthe matrix model, Arabic provides occasional system morphemes. They propose that a variety of mixed discourse types needs to be distinguished. Rudolfo Jacobson ('Conveying a broader message through bilingual discourse: An attempt at contrastive codeswitching research' , 5 1 -76) reviews the attempts that have been made towards finding universals in codeswitching behavior, discusses some of the significant counterexamples, and concludes that postulating universals is premature. In 'Language norms and models and how to describe them' (79-90), Delia Haust and Norbert Dittmar describe twelve different types of codeswitching found in their Gambian data and use these to characterize speakers' performances. The intricacies ofthe matrix language frame model make Carol Myers-Scotton's paper, 'Structural uniformities vs. community differences in codeswitching ' (91-108), most welcome m this collection . Myers-Scotton is careful to provide definitions of important concepts such as matrix language, embedded language, content vs. system morpheme, and a variety of island types, in structural terms and to distinguish these from the sociolinguistic or community factors that influence their actual realization in discourse, making the strong claim that 'all existing performance patterns can be accommodated within the Matrix Language Frame model' (105). However, Shoji Azuma ('Meaning and form in codeswitching ', 109-23) claims that a functional approach is more accurate. Erica McClure ('The relationship between form and function in written national language-English codeswitching', 125-50), following the approach of Susan Gal, shows how codeswitching to English in Spanish and Bulgarian texts reflects the positions of Mexico, Spain, and Bulgaria in the world economy. Li Wei ('Banana split? Variations in language choice and code-switching patterns oftwo groups of Britishborn Chinese in Tyneside', 153-75) takes a more strictly sociolinguistic approach with a detailed analysis of network affiliation and codeswitching patterns . Immigrant bilinguals are also the focus of Jeanine Treffers-Daller' s 'Variability in codeswitching styles: Turkish-German codeswitching patterns' (177-98), in which she argues for relative rather than absolute constraints. Rajeshwari V. Pandharipande ('Is genetic connection relevant in code-switching? Evidence from South Asian languages', 201-20) argues that speakers ' perceptions of the 'distance' between matrix and embedded language affect how codeswitches are structured. The final article, by Rosita Rindler Schjerve (221-47), analyzes Sardinian-Italian data to argue that codeswitching does not necessarily accelerate language shift. While some of these papers are more convincing than others, the volume presents a valuable view of the variety of current approaches to a well-defined problem in sociolinguistics. [Susan Meredith Burt, University of Wisconsin, Oshkosh.] Semantics. By John I. Saeed. Oxford: Blackwell Publishers Ltd., 1997. Pp. xix, 360. Paper $21.95. Several introductions to semantics have been published recently: Robert A. Hipkiss's 1995 Semantics: Defining the discipline (Oxford: Blackwell), John Lyons's 1995 Linguistic semantics: An introduction (New York: Cambndge University Press), and now John I. Saeed's Semantics. Lyons's and Saeed's texts are distinguished by their balanced attention to lexical semantics and to sentential semantics. Semantics is divided into three parts. The first part with two chapters ('Semantics in linguistics' and 'Meaning, thought and reality') is introductory. The second part attends to the language phenomena that have been identified as 'semantic'. And the third introduces some theories of semantics. Part II is the longest and contains six chapters. The first chapter concerns lexical meaning, and the remaining five concentrate on aspects of the semantics ofgrammar. The phenomena treated here include the semantic relations between sentences (Ch. 4), the semantics of events (Ch. 5), the semantics ofparticipant relations (Ch. 6), the relation of utterance to...