This paper focuses on differential scope-taking options of various quantifier classes. A critique is presented of a recent influential account based on feature-checking A-bar movements of the various quantifier classes to specialized functional projections, proposed by Beghelli and Stowell (B&S) (1994, 1995, 1997) and Szabolcsi (1997). We reject B&S`s theory on the grounds that it (i) cannot capture the clause-boundedness of the scope of universals, (ii) it creates a conceptual redundancy, (iii) it both undergenerates and overgenerates. It is argued, contra Szabolcsi, that Hungarian with its overt A-bar movements does not in fact furnish overt evidence for such an account: (i) projections proposed by Szabolcsi do not parallel those of B&S, (ii) Hungarian exhibits a radical scopal freedom within the postverbal field, (iii) some predicted overt movements do not actually take place. The model proposed here incorporates Quantifier Raising for GQs, existential closure over choice function variables as in Reinhart (1997) for bare numeral indefinites, and A-reconstruction. We build on Krifka (1999) in taking modified numeral indefinites to involve focus on the numeral, and we argue that A- reconstruction for scope across modified numeral indefinites is generally blocked, falling under a variety of quantificational intervention discussed recently in Boeckx (2001). It is demonstrated that this approach, combining independent recent results, is empirically accurate in treating differing scope interactions of various quantifier classes.

• Balazs Suramyl