Colloquium: The scope of English indefinites: an experimental investigation

January 15, 3:30-5pm, Cobb 201
Tania Ionin, University of Illinois at Urbana-Champaign

Much theoretical literature has discussed the possible scope readings of indefinites, and their ability to take long-distance scope in particular (Fodor and Sag 1981, Reinhart 1997, Kratzer 1998, among many others): namely, whether indefinites such as "a (certain) professor" in (1) can take widest-scope out of a scope island (the reading in (2a)) or intermediate scope out of the island but beneath a higher quantifier (the reading in (2b)).

(1) Every student read every book that a (certain) professor had recommended.
(2a) paraphrase of the widest scope reading of the indefinite:
There exists a professor x such that every student read every book that x had recommended.
(2b) paraphrase of the intermediate scope reading of the indefinite: For every student y, there exists a professor x such that y read every book that x had recommended.
(2c) paraphrase of the narrow scope reading of the indefinite: Every student read every book that was recommended by at least one professor.

While theories of indefinites often rely on fairly subtle judgments, such as the availability of wide vs. intermediate scope readings, there has been fairly little experimental work on the interpretation of indefinites (but see Marti 2007 and Alonso-Ovalle and Menendez-Benito, submitted, on the interpretation of Spanish indefinites). The goal of the present work is to investigate the scope-taking possibilities of English indefinites experimentally, using controlled interpretation tasks with linguistically naive subjects. This talk reports on three studies testing the interpretation of a-indefinites and a certain-indefinites in a variety of contexts. It is generally accepted that while a-indefinites can take widest, intermediate, or narrow scope (2a-c), a certain-indefinites can take widest or intermediate scope (2a-b) but not narrowest scope (2c). Study 1 tests the proposal of Schwarz (2001) that the long-distance (widest scope and intermediate scope) readings of a-indefinites and a certain-indefinites are derived by different semantic mechanisms. Study 2 explores the scope of a-indefinites and a certain-indefinites in a greater variety of contexts, and shows that speakers have a strong preference for the narrowest-scope readings of a-indefinites. Study 3 examines the role that modification plays in the scope readings of a-indefinites. Taken as a whole, the results of the three studies provide evidence in favor of Schwarz's (2001) proposal, as well as in favor of Schwarzschild's (2002) view that a-indefinites are singleton indefinites. The studies also raise new questions concerning the pragmatic constraints on the availability of long-distance readings of indefinites.