When many things contribute to an outcome, people consistently judge certain ones to be the outcome’s “actual” cause. For instance, people believe the lit match, not the surrounding oxygen, was the cause of the fire. Why? Here, we offer a functional account of actual causation: Repeatedly judging whether something (e.g. the match) was the actual cause of an outcome (e.g. the fire) helps compute the probability that introducing it would produce the outcome. In other words, judgments of actual causation accumulate evidence about the effectiveness of potential interventions. We offer a formal account of this process, and show how it explains three basic qualitative features of causal judgment: why actual causes tend (1) to be necessary, (2) to be abnormal (the “abnormal selection” effect), and (3) to lack abnormal counterparts (the “supersession” effect). We show that this approach – which we call the “Sample-based Approximation Method for Predicting the Likelihood of Effectiveness”, or the SAMPLE approach – makes quantitative predictions that closely match participants’ judgments in a novel experiment
<< Back to list of publications