Intervention effects for inverse scope readings: a drt account

Abstract

Negated sentences with a universally quantified subject are usually interpreted with a wide scope of the negation. Consequently, sentences of the form Every N not VP and Not every N VP should behave very similarly. I discuss a contrast between the two types of sentences with respect to their NPI-licensing potential and their possible discourse continuations. I propose a DRT account of the discourse continuation facts that corroborates a representational theory of NPI licensing. Which readings are available for sentences of the form Every N not VP is one of the big puzzles in natural language semantics (see for example Horn 1989). In this paper I do not attempt to solve it, but rather add to the puzzle. I discuss two sets of data with respect to which sentences of the form Every N not VP diverge from the allegedly synonymous sentences of the form Not every N VP.