A theory of bouletic reasoning
- Hillary Harner, Navy Center for Applied Research in Artificial Intelligence, Naval Research Laboratory, Washington, District of Columbia, United States
- Sangeet Khemlani, Navy Center for Applied Research in Artificial Intelligence, Naval Research Laboratory, WASHINGTON, District of Columbia, United States
AbstractNo present theory explains or models the inferences people draw about the real world when reasoning about “bouletic” relations, i.e., predicates that express desires, such as 'want' in Lee wants to be in love. Linguistic accounts of such bouletic relations define them in terms of their relation to a desirer’s beliefs, and how its complement is deemed desirable (cf. Heim, 1992; Villalta, 2008; Rubinstein 2012). In contrast, we describe a new model-based theory (cf. Johnson-Laird, 2006; Khemlani, Byrne, & Johnson-Laird, 2018) that posits that such predicates are fundamentally counterfactual in nature. In particular, X wants P should imply that P is not the case, because you cannot want what is already true. The theory makes empirical predictions about how people assess the consistency of bouletic relations as well as how they use such relations to eliminate disjunctive possibilities. Two experiments tested and validated the theory’s central predictions. We assess the theory in light of alternative accounts of human reasoning.