Aron at Georgetown

If Al only sings, why can't he dance?

November 21, Aron is across town at the Speaker Series for the Department of Linguistics at Georgetown University, talking about "Constraining Alternatives." Aron will also be meeting with the department's semantics group, led by Paul Portner, to discuss his paper with Bernhard Schwarz, "Singular which, mention-some, and variable scope uniqueness."

Only applies to a proposition, p, and asserts that alternatives to p are false. Yet, there must be constraints on alternatives. Consider (1). When (1) occurs out of the blue, only excludes (2a), so the sentence conveys that Al did not run. Yet, only does not exclude (2b). If (2b) were excluded, the sentence would convey that Al did run, contrary to fact. Why is (2a) an attested alternative, but not (2b)?

(1) Al only [slept]F. 

(2) a. Al ran. b. Al did not run  

Fox & Katzir (2011) propose that the grammar disallows alternatives which are syntactically more complex than the prejacent. In syntactic terms, (2b) (Al did not run) is more complex than the prejacent in (1) (Al slept), due to the presence of negation. Instead of appealing to syntax, we will pursue a view where alternatives are regulated pragmatically, based on the Question Under Discussion (as in Beaver & Clark 2008, Roberts 1996/2012; cf. Schwarz & Wagner 2024). We will argue that this approach can account for (1), along with a range of further data, which would not be predicted by a structural complexity constraint.