Proceedings of the Annual Meeting of the Cognitive Science Society

Two descriptive accounts of actual causation judgment, the Counterfactual Simulation Model (CSM) and the Counterfactual Effect Size Model (CESM), propose that actual-cause judgments for an observed event are derived from operations over mentally simulated counterfactual alternatives of the event. We argue that counterfactual models face computational and conceptual obstacles and propose the Retrieval of Invariant Causal Knowledge (RICK) model, which utilizes forward mental simulation to map type-level causal knowledge directly to token causal events. We demonstrate two predicted shortcomings of the CSM and compare with our model's performance.