Frontedquantificationaladverbs.docVIP

Fronted Quantificational Adverbs* Ariel CohenBen-Gurion University of the Negev, Israel arikc@bgumail.bgu.ac.il Abstract Fronting a noun phrase changes the focus structure of a sentence. Therefore, it may affect truth conditions, since some operators, in particular quantificational adverbs, are sensitive to focus. However, the position of the quantificational adverb itself, hence its informational status, is usually assumed not to have any semantic effect. In this paper I discuss a reading of some quantificational adverbs, the relative reading, which disappears if the adverb is fronted. I propose that this reading results from an association not only with focus, but with B-accent (fall-rise intonation) as well. I argue that when an operator is marked with an informational feature, such as B-accent, its associated constituent cannot be marked with the same feature. Since a fronted Q-adverb is usually pronounced with B-accent, it cannot associate with B-accent, and the relative reading is not available. Thus, the effects of fronting range more widely than is usually assumed, and quantificational adverbs are a useful tool with which to investigate these effects. The Problem It is well known that left-dislocation and topicalization can affect logical form, hence meaning. For example, Rooth (1985) points out that sentences like (1) are ambiguous. (1) Ballerinas usually escort officers. Under one reading, (1) means that most ballerinas escort an officer, whereas the other reading is that most officers are escorted by ballerinas. The two readings have different truth conditions. Suppose there are 10 ballerinas and 100 officers. Nine ballerinas accompany an officer each; but the remaining 91 officers are accompanied by ladies who are not ballerinas. In this case, the first reading would be true, but the second reading would be false. As Rooth points out, (1) is disambiguated by focus: depending on which noun phrase is focused, we get one or the other reading. Another