Team: Nicole Gotzner (SiGames, ZAS), Jacopo Romoli (Ulster University), Paolo Santorio (University of California, San Diego)
Disjunctions in the scope of possibility modals give rise to so-called ‘free choice’ inferences (Kamp 1974).
(1) Iris can take Spanish or Calculus. ↝ Iris can take Spanish and she can take Calculus
An influential approach tries to derive free choice as a kind of scalar implicature. Standard algorithms for computing implicatures proceed by negating alternatives to a sentence and adding the information so obtained to the assertion (exclusion). Approaches in this vein have succeeded at capturing free choice and related phenomena (Fox 2007, Chemla 2010, Klinedinst 2007, Santorio and Romoli 2017, Franke 2011 a.o.). But recent experimental findings have challenged exclusion theories. Chemla (2009) shows that free choice occurs both in the scope of universal quantifiers as in (2) and negative existentials as in (3):
(2) Every student can take Spanish or Calculus ↝ every student can take Spanish and every student can take Calculus
(3) No student must take both Spanish and Calculus ↝ every student can avoid Spanish and every student can avoid Calculus
(2) is predicted by exclusion theories, but (3) is not. On these grounds, Bar-Lev and Fox 2017 (BL&F) (see also Santorio 2017) propose a novel theory of free choice, which can account for (3). BL&F adopt a novel algorithm for implicature computation, which directly adds information to the assertion (inclusion), without passing through negation. The functioning of the two algorithms is summarized below.
(4) Exclusion algorithm:
(i) consider all maximal sets of alternatives s.t. negating them is consistent with the assertion;
(ii) negate the alternatives in the intersection of all these sets (the ‘Innocently Excludable’ (IE) alternatives).
(5) Inclusion algorithm:
(i) Consider all maximal sets of alternatives consistent with the assertion and the negation of all IE alternatives;
(ii) conjoin with the assertion all the alternatives that are in all these subsets (the ‘Innocently Includable’ (II) alternatives).
The project aims at producing evidence for deciding between exclusion and inclusion theories. We focus on the interaction between free choice and a number of linguistic items, including nonmonotonic quantifiers, universal quantifiers, and presupposition triggers. We have two goals:
i. We will show that free choice effects occur in a variety of environments where they are unexpected, given exclusion theories.
ii. We will develop a refined inclusion-style approach to free choice that can deal with a variety of data (including the interaction with presupposition projection).