A Comprehensive Theory of Reasoning from Utility Conditionals

SPEAKER

Jean-Francois Bonnefon

ABSTRACT

Many `If p, then q' conditionals have decision-theoretic features, such
as antecedents or consequents that relate to the utility functions of
various agents. These decision-theoretic features leak into reasoning
processes, resulting in various paralogical conclusions. The theory of
utility conditionals offers a unified account of the various forms this
phenomenon can take. The theory is built on two main components: (1) a
representational tool (the utility map), which summarizes in compact
form the decision-theoretic features of a conditional; and (2) a set of
folk decision heuristics, which reflects reasoners' beliefs about the
way most agents make their decisions. Applying the folk heuristics to
the utility map of a conditional allows to systematically predict the
paralogical conclusions invited by its decision-theoretic features. The
theory of utility conditionals significantly extends the scope of
current theories of conditional inference, and moves reasoning research
toward a greater integration with decision-making research.