Reviewed by: The syntax of ditransitives: Evidence from clitics by Elena Anagnostopoulou Marcel den Dikken The syntax of ditransitives: Evidence from cliticsBy Elena Anagnostopoulou. Berlin: Mouton de Gruyter, 2003. Pp. xiv, 379. ISBN 3110170280. $114.40 (Hb). The syntax of ditransitives has long been a densely researched topic—an octopus whose tentacles have touched upon many a major issue in linguistic research, including the representation of argument structure, the syntax of argument-structure alternations, case theory, configurational structure and the binary branching hypothesis, A′-syntax and scope, c-command-based vs. linearity-based approaches to asymmetries, and locality. Elena Anagnostopoulou’s eminently detailed book contributes to this fertile field a study that focuses on the last-mentioned issue: the ways in which ditransitive constructions impinge on locality theory—particularly, the locality of A-movement. And by doing so, it defends a syntax of ditransitives in which double-object and prepositional ditransitives have radically different (and transformationally unrelated) structures, and, along the way, it elucidates the role of case theory in reining in the possibilities of A-movement by showing that case theory plays no major role here at all (thus lending support to Chomsky’s (2000, 2001) ‘demotion’ of case as a factor in triggering and constraining movement). Its results bear obliquely on the many questions surrounding the argument structure of triadic verbs; but by making the locality restrictions on A-movement its focal point, this study sheds much-needed light on major questions that have lingered in footnotes or disclaimers for much too long, issues on which the minimalist approach to syntactic derivation affords us a particularly clear window. A sets herself two major goals in her book: (i) ‘establishing that NP-movement can be used as an analytic tool for probing into the properties of ditransitives’, and (ii) showing that ‘cross-linguistic variation results from the interplay between the morpho-syntactic properties of double object constructions and the syntax of clitics’ (1). This second goal earns the book its subtitle, which, interestingly enough, is not reflected in the blurb text on the back cover. And indeed, evidence from clitics plays a central role only in the last two, relatively short, chapters of the book, with Ch. 4 making a detailed case for the idea that clitics obviate locality effects and Ch. 5 addressing person restrictions. Clitics figure prominently in the context of these person restrictions: in ditransitives with two weak pronominal or clitic objects, the direct object has to be third person. But one of the main contributions of Ch. 5 is precisely its demonstration of the fact that these restrictions by no means cluster specifically around clitics: in Icelandic constructions with a quirky dative subject, the nominative object agreeing with the finite verb has to be third person as well. Overwhelming in empirical detail, drawing directly and heavily on other scholars’ work, and often contenting itself with stipulations instead of explanations, Ch. 5 is a somewhat unfortunate finale. It is nonetheless an important piece of the puzzle because of its impact on the fine-tuning of the definition of locality and the cycle that the book needs. To explain how the indirect object restricts the person specifications of the direct object, A has the indirect object clitic raise to the v-head that checks the clitic’s ϕ-features first, followed by the direct object. On the assumption (supported by Icelandic quirky subject agreement but not derived from anything) that the indirect-object clitic checks person but not number on v, cliticization of the indirect object leaves v with a number feature to check against the direct-object clitic, which it can indeed check if the direct-object [End Page 980] clitic is third person (i.e. nonperson; Benveniste 1966) but not if it is first or second person: a first or second person direct-object clitic, possessing both number and person features, fails to match a v-head that only has a number feature left.1 For this analysis to work, we must make the minimal link condition (MLC) an integral part of the definition of Move/Attract: If we required that the derivation proceed strictly cyclically within phases (Chomsky 2001), the direct-object clitic...