Causality, Responsibility, and Blame: A Structural-Model Approach

Abstract

This talk will provide an overview of work that I have done with Hana Chockler and Judea Pearl [1,4,5] on defining notions such as causality, explanation, responsibility, and blame. I first review the Halpern-Pearl definition of causality-what it means that A is a cause of B–and show how it handles well some standard problems of causality. This definition is based on what are called structural equations, which are ways of describing the effects of interventions. The definition (like most in the literature) views causality as an all-or-nothing concept. Either A is a cause of B or it is not. I show how the account can be extended to take into account the degree of responsibility of A for B. For example, if someone wins an election 11-0, each person is less responsible for his victory than if he had won 6-5. Finally, I discuss more recent work [2,3] on combining a theory of normality (or defaults) with the structural equations. A slightly revised definition of causality that uses normality deals well with problems that have been pointed out in the original Halpern-Pearl definition, and helps explain different intuitions that people have regarding causality.