178LANGUAGE, VOLUME 74, NUMBER 1 (1998) Verb movement and expletive subjects in the Germanic languages. By Sten Vikner. (Oxford studies in comparative syntax.) New York: Oxford University Press, 1995. Pp. x, 294. Reviewed by Ans van Kemenade, Vrije Universiteit, Amsterdam This is a revised version of Vikner's doctoral dissertation, University of Geneva 1991; in the unpublished form, it has already been influential in the literature, not least for its incisive discussion of the core issues in the extensive literature on verbal position in the Germanic languages. The book aims to give an in-depth description and analysis of V-movement phenomena in the Germanic languages from the perspective of relativized minimality (RM), following Rizzi's (1990) alternative to the locality theory in Chomsky (1986). This yields a clearly chosen approach to a well-defined empirical domain though nothing much seems to hinge on the choice: RM does not appear to have a marked explanatory edge over alternatives when it comes to the treatment of V-movement. Extensive discussion of Vs native language, Danish, is brought to bear on the discussion. Part I consists of two preliminary chapters. Ch. 1 gives an introductory overview of the Germanic languages and a brief explication on language acquisition and the theory of UG as viewed from the principles and parameters perspective. Ch. 2 gives a summary of RM. Part II, on verb movement, consists of three chapters. Ch. 3 is on verb second in its classic sense: the absolute word order constraint found in languages like Dutch and German, which requires that in the root clause the finite verb is always in second constituent position, regardless of its base position, as in the German Vielleicht hat Peter dieses Buch gelesen 'perhaps has Peter this book read'. The first constituent may be one of a range: adverb, subject, object, some prepositional phrase etc. The example illustrates subject-verb inversion when the first constituent is not the subject. The position of the finite verb in German V2 clauses is in complementary distribution with that of the subordinating conjunction in embedded clauses. This is why V2 is analyzed, by V following many others, as movement to the presentential head position C: (1)[a, XP [c V1 [IP subject [: ....... [VP ....t,]]]]] V-movement in embedded clauses is generally blocked by the presence of a lexical complementizer like that. V argues at length that V-movement in root clauses is always to C, remaining neutral on the question what property of C requires such movement. Ch. 4 is about the landing site for V-movement in embedded V2 clauses. The approach in Ch. 3 predicts that V2 is restricted to root clauses because it is incompatible with a complementizer . There are two classes of exceptions to this: In one, V2 is of exactly the same type as in root clauses which are embedded under certain classes of verbs, labelled confusingly 'bridge verbs' in the literature. Here again, languages are oftwo subtypes: One is exemplified by German, where V2 orders are found with the same complementary distribution between Vf and complementizer as in root clauses; the second type is exemplified by Danish, where V2 orders cooccur with a complementizer: (2)Vi ved at denne bog har Bo ikke lest We know that this book has Bo not read 'We know that Bo hasn't read this book' Vs analysis of this phenomenon involves a recursive CP so that this type really reflects an embedded root clause. The second type, found in Icelandic and Yiddish, occurs in a wider range of embedded clauses, and its analysis is more contentious. Contra previous literature, in which Yiddish and Icelandic embedded V2 is analyzed as V-movement to I, with topicalization to Spec-IP, and the nominative subject in Spec-VP, V argues at length that embedded V2 in Yiddish and Icelandic involves a (generalized) recursive CP. The arguments comprise various types of distributional evidence. The distribution of expletive subjects in Icelandic and Yiddish root and nonroot clauses parallels that of expletives in the German root clause, which would seem to favor a CP analysis; the distribution of adverbs and of objects in object-shift constructions shows that the position of the nominative...
Read full abstract