Language and the brain By Loraine K. Obler and Kris Gjerlow (review)

490 LANGUAGE, VOLUME 76, NUMBER 2 (2000) ture but concludes that all of them are inadequate. Some of these are first mention of an element—but L shows that first-mentioned NPs occur with the same distribution as non-first-mentioned ones; definiteness —which also has no effect on distribution; and the 'theme-first' principle—which is consistently violated in Cheyenne, as thematic information seems to come later in the clause. (If this is true, though, thematic structure does play a role in word order.) In Ch. 4, L concludes that the discourse principle of 'newsworthy first' is what determines word order in Cheyenne. An element of a sentence is newsworthy if it points out a significant contrast, introduces or changes a topic, or represents significant new information . The collected texts, L argues, show the effects of this principle. In Ch. 5, L describes an effective experiment designed to test the hypothesis that word order is determined by newsworthiness. She shows silent videos and asks Cheyenne speakers to narrate them. The results are presented in Ch. 6, where it is argued that the newsworthy first principle does in fact account for the word orders used. While the hypothesis leaves many questions unanswered , some of which L acknowledges (such as the interaction of newsworthiness and demonstratives, adverbs, and obviation, and what determines the relative order of elements that are less newsworthy than the first element), this study does show that many suggested factors are apparently irrelevant to word order in Cheyenne. It also presents useful data for students ofAlgonquian and discourse analysis. [Benjamin Bruening, Massachusetts Institute of Technology .] Mood at the interface. By Josep Quer. The Hague: Holland Academic Graphics , 1998. Pp. 293. Tosep Quer examines mood systems, primarily in Romance languages (with specific focus on Catalan and Spanish), in this detailed study of subordination. His main hypothesis is that mood alternations signal a change in the model where a proposition is evaluated . Specifically, subjunctive mood indicates that the proposition is to be evaluated with respect to an element in the sentence or higher clause, for instance a matrix subject, while indicative mood is either a default or indicates that the proposition should be evaluated with respect to the speaker. In addition, where mood is grammatically specified, i.e. where no alternation is permitted, it is because certain lexical items, such as verbs and complementizers, select for a certain mood, e.g. subjunctive. This selection follows from their semantics; for instance, items that select subjunctive are intensional anchoring elements . Q contrasts this theory of mood with other recent theories throughout the book and shows how they differ in empirical coverage. Ch. 1 briefly describes such previous theories, for instance the theories that subjunctive mood is a defective tense or is akin to indefinite noun phrases, and extracts what insights they offer. It also introduces the dynamic semantics model that Q adopts and the possible worlds approach to modality. It then outlines the basic hypothesis of the study. Ch. 2 explores the distribution ofmood in subordinate clauses, including argument clauses and adjunct clauses. Q shows that lexical selection is what determines mood in intensional domains (both complements of strong intensional predicates and purpose clauses). In epistemic contexts, in contrast, mood is not grammatically determined; either indicative appears as a default or, where there is a clausal operator, the choice between subjunctive and indicative triggers a shift in model evaluation. Ch. 3 investigates the influence of mood on the interpretation of relative clauses. Subjunctive mood is permitted under tightly constrained conditions: A subjunctive relative must occur in the scope of an intensional predicate or other operator (such as negation ). Again, the contrast between subjunctive and indicative in these cases indicates the model according to which the proposition is to be evaluated: subjunctive within the model introduced by the operator and indicative outside that model (either with respect to a higher element or the speaker). The effects of mood choice on free relatives are discussed in Ch. 4, where it is shown that subjunctive free relatives behave like free choice items. Q proposes an analysis of free choice readings and shows how it extends to subjunctive free relatives. Indicative free relatives behave as definite...