Object scrambling and quantifier float in German

Abstract

The distribution of floated quantifiers in German bears a number of similarities to data discussed by Shlonsky (1991) for Hebrew, showing agreement in some cases between the floated quantifier and the nominal quantified over. Although the pattern of agreement seen in Hebrew appears to be only partially attested in German, I argue that when the covert part of the derivation is taken into account, German too exemplifies a biconditional relation between the presence of agreement on a quantifier head and the position of the DP quantified over. The German data can be clearly accounted for under Sportiche’s (1988) account of floating quantifiers as stranded by movement of the associated nominal; competing analyses of this phenomenon, which treat the floated quantifier as an adverbial, are rejected as having no account of the relevant agreement facts. I show further that the asymmetries in the possible positions of quantifiers floated from subjects vs. those floated from objects follow straightforwardly from recent proposals concerning structure of the clause, specifically that AgrSP and TP dominate AgrOP, taken in conjunction with standard accounts of adjunct extraction islands. Quantifiers associated with subjects can be stranded in Spec-AgrSP or Spec-TP, while those associated with objects strand in Spec-AgrOP (in addition to stranding in base positions, possible for both subjects and objects). This difference can be seen with respect to sentential adverbials, which adjoin to TP (Holmberg 1993). On the bases of these data, I argue with Mahajan 1994 that ‘scrambling’ should be understood as a cover term for two distinct types of movement: A-movement to a specifier of a functional projection (where a quantifier can be stranded), and A'-movement consisting of adjunction to any of the functional projections of the articulated Infl. I begin by briefly reviewing the two major competing analyses, rejecting the adverbial analysis and adopting Sportiche’s (1988) quantifier stranding approach. In section 2 I discuss the German data and argue that the observed positional and

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 58 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.