Minimalism, scope and VP structure by Thomas Stroik (review)

668 LANGUAGE, VOLUME 74, NUMBER 3 (1998) direct object, both ofwhich N eventually rejects The analysis culminates in Ch. 7, 'Ergativity in Inuktitut' (191-236), in which she concludes that, from the point of view ofargument structure and grammatical relations, Inuktitut must be seen as sui generis, falling outside of all previous treatments ofergativity, since a transitive sentence has two subjects and no object. N's analysis often appears to rely on its own set of unquestioned assumptions such as that a stable relationship should be found between case marking and thematic role or between case marking and subjecthood ; her declaration that only ergative subjects are agents needs more argumentation than she gives it. Why should the ergative noun be agentive in arnaup quimmiq takuvanga 'the woman sees the dog', but the absolutive noun be nonagentive in arnaq qimmimik takujuq 'the woman sees the/a dog'? Similarly , her claim that neither the ergative nor the absolutive 'subject' qualifies as the sole 'subject' (214-30) seems to focus on these case categories without looking at the relationships between them as carefully as might be hoped: N states that incorporation always acts on the absolutive without mentioning the well-known stranding of instrumental casemarked modifiers, and claims that anaphoric coreference gives evidence for the subjecthood only of the ergative NP, which is highly questionable. Likewise, she chooses instrumental case-marked NP in antipassive sentences as the only true direct object, since it 'fulfills an important criterion for direct objects, that of being the obligatory nominal supplement to the predicate' (235). However, she omits mentioning that the absolutive NP in an ergative-absolutive transitive sentence is obligatory for purposes of interpretation, although it can be phonologically null, thanks to verbal cross-referencing. In the final chapter, 'Transforming the images' (237-77), N presents her conclusions and gives a brief grammatical sketch of Inuktitut more in accordance with her theoretical notions. The ramifications ofN's radical theoretical claims are left largely open, hopefully to be pursued in future research. In general, the book is well-written and clear although there is some variation in orthography and citation forms, i.e. kuni- 'kiss' (259) vs. kunik- 'kiss' (273), which may cause some confusion to the non-Eskimologist reader. An unfortunate glitch has caused all footnotes after fn. 1 38 to be numbered incorrectly: the reader must thereafter add 87 to the footnote reference. [David Parkinson, Microsoft Corporation.] Minimalism, scope and VP structure. By Thomas Stroik. Thousand Oaks, London, & New Delhi: Sage Publications , 1996. Pp. vii, 173. $37.00. The main aim ofthis book is to argue for a particular view of the structure and ordering of constituents within the extended VP projection. Stroik is arguing here against various versions of this ordering proposed in the past and using evidence from four main areas (scope, superiority effects, bare NP adverbials, and psych-verbs) to support his claim that the goal argument is generated structurally higher than the theme, that temporal and locative adverbs are sisters to the verb, and that whereas manner adverbs are VPadjoined , reason adverbs are AgrSP-adjoined. The book is well-organized and clear although the data are notoriously tricky, and some readers may find it difficult to agree with certain grammaticality judgments regarding sentences involving multiple WH-insitu phrases or scope effects between multiple quantified phrases. Ch. 1, 'Introduction to syntactic representation' (1-31), serves as a useful introduction to current issues in government and binding/minimalist syntactic theory and allows S to discuss some of the assumptions he will rely on throughout the book. S presents the debate between conditions on economy ofderivation vs. conditions on representation by arguing in favor of a representational approach to the empty category principle (ECP) and the distribution of PRO, as opposed to a more orthodox minimalist derivational approach. Using Aoun & Li's minimal binding requirement as a means for gauging the representational conditions which must hold at LF, S examines four main sources of evidence for the proposed VP structure. Ch. 2, 'Scope and the VP shell hypothesis' (32-72), focuses on the interactions between multiple quantified phrases and WH-words in sentences such as Why did Jill read no one anything (65), using the alternation between...