Information status and noncanonical word order in English By Betty J. Birner and Gregory Ward (review)

BOOK NOTICES 219 properly understood as complementizers; H analyzes for and whether as prepositions. In the following chapter ('Lexical properties of prepositions', 190-235), H considers the subclass of lexical prepositions , to which she assigns temporal adverbial subordinators , and argues for the necessity of three additional functional projections dominating PP. She ties together her analyses ofcomplementizers, prepositions , and their projections in Ch. 8 ('Subordinating conjunctions straddle the dividing line between the lexical and the functional universe', 236-61) and introduces the functional head Subcon and the constituents which it both dominates and is dominated by. Though clearly organized, this study would have been made much more readable had it been more closely proofread for typos and, especially, English style. [Mark L. Louden, University of Texas at Austin.] Information status and noncanonical word order in English. By Betty J. Birner and Gregory Ward (Studies in language companion series 40.) Amsterdam & Philadelphia: John Benjamins , 1998. Pp. 314. The pragmatics of word order is a popular issue in present-day linguistics: few linguistic phenomena lend themselves that easily to empirical research, and it is an area in which more or less everyone agrees on the general principles involved, i.e. on the overall role played by the information status of the ordered constituents. But the notions 'given' vs. 'new', and similar dichotomies, are usually used in a fairly unrefined way. Studies in this area also often single out particular syntactic constructions without considering their information structural similarities to or differences from other constructions. Birner and Ward do not upset the quasiconsensus on the role of information status in word order ('given' before 'new'), but they claim that it is possible to make generalizations about the functions of syntactically similar constructions on condition that the information structural notions employed are sufficiently fine-grained. Three kinds of construction that deviate from the unmarked canonical word order ofEnglish are distinguished : those that involve 'proposing' (i.e. when some argument of the verb appears to the left of its canonical position, as in For that last bold assertion there are no statistics), those that involve 'postposing ' (some argument of the verb appears to the right of its canonical position, as in There 's a dog running loose somewhere in the neighborhood), and those that involve both (argument-reversing constructions: the logical subject appears in postverbal position and some canonically postverbal argument appears in preverbal position, as in Also a nice woman is our next guest). Each of these is treated in a separate chapter (Chs. 2-4). Preposed constituents (to be distinguished from left-dislocated constituents, which are formally and functionally different) represent discourse -old (rather than simply 'given') links that situate the information presented in the current utterance with respect to the prior context. Postposed constituents (formally and functionally different from rightdislocated constituents) represent information that is unfamiliar in some sense (rather than simply 'new'): existential fnere-sentences need a postverbal NP that represents information that is (presumed to be) new to the hearer, whereas presentational i/iere-sentences only require a postverbal NP that represents information that is new to the discourse. Argument-reversing constructions serve to preserve the normal ordering of information (given before new) in cases where canonical word order would result in relatively new information preceding relatively old information. 'Relatively' is the operative word here: with preposing and postposing the information constraints are 'absolute'; in the case of argument reversal the first constituent must only be at least as familiar within the discourse as the second. Ch. 5 compares the distribution ofpreposing, postposing , and argument reversal in discourse; examines properties shared by preposing and argument reversal ; and looks at correlations between information structure and intonation. In the concluding sixth chapter some crosslinguistic data from Farsi, Italian, and Yiddish are presented to suggest that the proposed generalizations are not limited to English. The book is somewhat repetitive (perhaps because of the fact that it is a reworked compilation of eight previously published articles), but without question B&W have added much-needed explicitness to the study of information packaging in their efforts to arrive at generalizations about the form-function, or syntax-discourse, relation. The publication of their work in book format will help to...