And then there were two. Children and second language learning By Terry Piper (review)

BOOK NOTICES 615 Those readers who come to this collection of articles expecting to encounter profound shifts in our understanding of the pragmatic motivations underlying word order are bound to be disappointed . What one finds instead is a more modest undertaking, a compilation of research on individual languages which provides both a refinement of some traditional discourse concepts and evidence to back past proposals about the influence of pragmatic factors on word order. The factors which are explored include: the (dis)continuity ofreferents in a text, the relative newsworthiness of these referents, themerheme structure, thematic/temporal (discontinuity , and discourse genre. The volume contains an introduction by Payne and twelve articles that generally relate to the issue of which pragmatic considerations motivate the arrangement of constituents in a clause, particularly in those languages that ostensibly do not have a 'basic' word order. Since the majority of the contributors have been affiliated at some point with the Linguistics Department at the University of Oregon, it is not surprising that the collection taken as a whole is representative of the research on discourse which has characterized Oregon for some time. Among other things, there is a crosslinguistic orientation, an emphasis on data drawn from spoken and written texts, and argumentation based largely on statistically significant tendencies which are located in these data. In her introduction (1-13), Payne identifies Marianne Mithun's contribution, 'Is basic word order universal?' (15-61; previously published in 1987), as establishing the thematic complexion ofthe volume. In this paper, Mithun argues that there is no correlation between grammatical relations and the order of clausal elements in Coos, Cayuga, and Ngandi. Thus, the notion of basic word order is not useful for these languages, at least for the description of linear relations in the clause. Mithun then goes on to demonstrate that the order of constituents is due to their relative 'newsworthiness'. Indeed, this article does set the tenor for the remainder of the volume, in that most of the contributions have the same basic form as Mithun 's paper. First, there is a discussion of why the notion of basic word order is problematic or descriptively vacuous for language X. Second, there is a proposal about nonsyntactic factors which can be used to explain various constituent orderings in language X. Although the ubiquity of this pattern of argumentation gets tedious as one reads through the articles, there is enough of interest in the descriptive details of specific analyses to hold one's attention, especially since the languages which are analyzed comprise a genetically diverse set: Ojibwa (Russell S. Tomlin & Richard Rhodes, 117-35; previously published in 1979), Nez Perce (Noel Rude, 193-208), Klamath (Karen Sundberg Meyer, 167-91), Polish (Barbara Jacennik & Matthew S. Dryer, 209-41), Chamorro (Ann Cooreman, 243-63), Biblical Hebrew (John Myhill, 265-78), Agutaynen (J. Stephen Quakenbush, 279-303), and 'O'odham (Doris L. Payne, 137-66). Three of the articles in this collection do not fit comfortably with the rest. Ken Hale's 'Basic word order in two "free word order" languages ' (63-82) argues that, despite surface appearances , a syntactically based basic word order is determinable in Papago. He goes on to suggest that any variation on this order is the result of movement rules. Not only does this conclusion seem to run counter to the general theme of the book, it is in a more direct tension with Payne's own contribution (also on Papago ), 'Nonidentifiable information and pragmatic order rules in 'O'odham', in which she proposes that visible syntax like word order should be separated from less visible syntax like head-dependent relations. The other two articles whose presence in the volume seems awkward are The privilege of primacy: Experimental data and cognitive explanation ' (83-116), by Morton Ann Gernsbacher & David Hargreaves, and 'On interpreting text-distributional correlations: Some methodological issues' (305-20), by T. GivóN. The former is a presentation of experimental evidence for a general cognitive strategy called 'structure building', which can be used to explain why focused or newsworthy information tends to be placed sentence-initially in many languages. The latter is essentially a response to Johanna Nichols, who in an unpublished manuscript questioned the...