Configurations of Sentential Complementation: Perspectives from Romance Languages (review)

This is largely a collection of some of Rooryck’s previously published articles (namely, Ch. 1 and Chs. 4–7, in not ‘crucially different’ form, xii); Ch. 2 is a modified version of a co-authored paper with João Costa, while the content of Chs. 3 and 8 was previously only presented at conferences. It is my understanding that the ideas presented in this volume basically range from 1989 to 1997, and as such some book-internal mismatches or outdated analyses are to be expected (or as R says, ‘there is some redundancy in places’, xii), especially as they are all supposed to have a bearing on each other in that R aims at ‘provid[ing] some answer to […] classical problems’ (xi); moreover, the eight chapters of the book fall into three natural parts (one discussing infinitival complementation, one treating clitic constructions, and one considering wh-sentences). But especially judging from recent years, this seems to be a growing trend in linguistics. And as in many comparable cases, it takes nothing away from the content. This is a highly stimulating book, containing interesting approaches to and analyses of, indeed, ‘classical problems’.

R begins by discussing infinitival complementation in the first three chapters—‘Raising’ (1–53), ‘Pseudo-raising’ (54–72), and ‘Control’ (73–114). Here raising proper (aka ‘subject-to-subject raising’) and ECM-constructions receive a unified analysis in that both involve movement from the embedded AgrSP-position to the embedded SpecCP, argued to take place for focus reasons. Following related research, R dispenses with seem and believe as separate lexical items but takes them as basically the same, with the latter being the accusative counterpart to the former. The first chapter also compares raising differences between English and French, the second chapter English and Portuguese, especially with respect to ‘raising’ from tensed clauses (cf. Sandy seems like she is happy). Ch. 3 centers around why the choice of the controller of PRO is always determined by the matrix verb (viz. subject vs. object control/promise vs. persuade). The analysis explores the relevance of (arguments present in) [-realized] subevents and ties in nicely with the proposal about raising in that it is always selectional properties of the matrix verb on the embedded C-head that derive the different interpretations and complementation structures.

The clitic part contains two chapters, ‘Enclitic ordering in imperatives and infinitives’ (115–43), which R takes to be the result from (the clitic) being left behind as the verbal complex moves up, and ‘Clitic climbing’ (144–63), which extends the analysis from the previous chapter. Both titles (like all the others, actually) are self-explanatory in terms of content.

Three chapters on wh-structures round the book off: ‘Negative and factive islands’ (164–96), ‘On two types of underspecification: Evidence from agreement in relative clauses’ (197–222), and ‘A unified analysis of French interrogative and complementizer que/qui’ (223–46). Ch. 6 picks up the earlier relevance of matrix verbs and properties of selection (on embedded C-heads); the last two chapters concern wh-elements and properties of C and SpecCP.