One possibility is that non-focused elements are moved from the nuclear scope into the restrictor. A different view is that the nuclear scope contains the entire sentence (without the Q-adverb), and does not lose any material. Its effect on interpretation is to provide a set of alternatives whose union is accommodated into the restrictor.

It is not often realized, but the two views are quite significantly different. Choosing one or the other has empirical as well as theoretical consequences. The first view is unappealing, in that it suffers from a theoretical problem - it implies that focus has two, unrelated roles: introducing alternatives and causing movement. The second view, however, suffers from an empirical problem: it implies that quantificational adverbs are conservative, but it turns out that there are non-conservative readings of some quantificational adverbs.

In this paper I argue for the second view, by demonstrating how these non-conservative readings can be accommodated into its framework. I propose that, while under the usual, conservative reading, the quantificational adverb associates with focus, under the non-conservative reading it also associates with a B-accented element (so called "contrastive topic").