Explanations of Counterfactual Inferences

When engaging in counterfactual thought, people must imagine changes to the actual state of the world. In this study, we investigated how people reason about counterfactual scenarios by asking participants to make counterfactual inferences about a series of causal devices (i.e., answer questions such as If component X had not operated [had failed], would components Y, Z, and W have operated?) and to explain their reasoning. Participants avoided breaking deterministic causal links (i.e., W always causes X), but were willing to break probabilistic causal links (i.e., W sometimes causes X) to keep prior causal events in the same states as in the actual world. Participants’ explanations supported this pattern of inferences. When the causal links were deterministic, participants reasoned diagnostically to infer that the states of prior causal events would have been different in the counterfactual world. In contrast, when the links were probabilistic, participants cited the links’ unreliability as an explanation for why the states of prior causal events would have been the same as in the actual world. Additionally, participants who were told that a component “had failed” (vs. “had not operated”) were more likely to attribute the state of that component to it being “internally broken” and infer that causally upstream components would have operated. Our results suggest that people use their explanation of the antecedent event (the “if” clause) to guide their counterfactual inferences. We discuss the implications of these findings for two rival Bayes-net theories of counterfactual reasoning: Pearl’s (2000) and Hiddleston’s (2005).