Possessor chains in Norwegian By Toril Fiva (review)

BOOK NOTICES 189 equately), and their model distinguishes "rigidly' between formal expression and content and, in order to correct an 'error' in valency theory, further distinguishes between a 'virtual and material level' with a useful set of verb data. Although Murtadha Bakir (Basrah) uses a rather old (1973) work of Chomsky's (in fact, several of the articles in this book seem lacking in current referencing), in his 'Notes on subjacency as a syntactic constraint in Arabic and English' he presents useful data for the continuing argument on the reality of movement rules in grammar. In Stanisíaw Puppel's (Poznan) 'Coarticulatory propensity: The case of English and Polish consonant clusters', a number of facts are nicely laid out and the 'propensity' towards insertion of sonorant segments in both languages in a clear phonotactic pattern (but with different frequencies) appears solid. In 'Contrastive analysis at discourse level and the communicative teaching of languages', Sophia Marmaridou (Athens), from an interlanguage (IL) perspective, describes an EnglishGreek experiment and proposes a functionalist CL model. Her study of crosslinguistic examples should be useful to those interested in such pragmatic concerns as convergent and divergent linguistic strategies of 'making suggestions' across languages by native and nonnative speakers. Also from an IL perspective, in the longest article in the collection (45 pages)— 'Connecting Ll and FL in discourse-level performance analysis'—Lars Sigfred Evensen & Irmgard Lintermann Rygh (Trondheim) explore connectors and their underlying logical relationships in a large-scale descriptive study of these discourse features. Importantly, in these studies the scope of CL is expanded to include comparison of native-language and IL performance . Finally, the traditionally difficult problem of what exactly counts as 'interlingual equivalence ' is addressed in terms of several concerns. First, R. R. K. Hartmann (Exeter) is interested in the creation of bilingual dictionaries in 'Equivalence in bilingual lexicography: From correspondence relation to communicative strategy'. A process-oriented codeswitching perspective, though not put in these terms, adds data to our knowledge base of cognitive domains underlying lexical choice. Though Hartmann uses translation as a device, this is Gunter Weise's (Halle) primary focus, in 'Contrastive studies and the problem of equivalence in translation'. I like his bringing the pragmatic purpose of a text into his consideration of interlingual equivalence, e.g. the briefing of a medical colleague about a patient. What this seems to do is to limit the range of possible translations. But in spite ofthe interesting contrastive facts here, one is left slightly dissatisfied by collections of this type. I keep hoping for a pullingtogether of the CL experience, a reframing of it in more general and more current terms, perhaps as a set of coherence principles that unite various CL studies. The need for such a generalized statement gains urgency when one reflects that, no matter how difficult it may be to conceive of theoretically, the linguistic world is hardly a set of separate and separable monolingual experiences. Even if one's goal is a linguistic theory of mental/internalized language, it is increasingly hard to pretend that interlingual data is not relevant. Much interlingual experience that would be useful for theory building is captured in the vast, but largely buried, CL literature. [Larry Selinker, University of Michigan.] Possessor chains in Norwegian. By Toril Fiva. Oslo: Novus, 1987. Pp. 104. This book gives a unified analysis of possessive constructions in Norwegian within the framework of Government-Binding theory. Norwegian has a larger set of possessive constructions than English currently has, as exemplified by the following four translations for 'Peter's car': (a) Per sin bil (lit. 'Peter his car'), (b) Pers bil ('Peter's car'), (c) bilen til Per (lit. 'car.the to Peter'), and (d) bilen bans Per (lit. 'car.the his Peter'). For all of these constructions , Fiva proposes an analysis in which the possessor NP is base-generated in postnominal position. She also argues that the genitival sin is included in the chain consisting of the prenominal NP and its postnominal trace, and that -5 is a varient form of sin. F notes (11, 84) that when the possessor NP occurs in postnominal possession, as in (c), the head noun must be definite, but that when the...