Causal judgments approximate the effectiveness of future interventions

Abstract

When many things contributed to an outcome, people consistently judge certain ones to be more causal than others. For instance, people believe that a fire was more caused by the lit match than by the surrounding oxygen that fueled it. Why? Here, we offer a functional account of such patterns in causal judgment: By selecting causes as people naturally do, repeated judgments of whether something (e.g. the match) was the cause of an outcome (e.g. the fire) can be averaged to obtain the probability that introducing those things would produce the outcome (e.g., that lighting a match would start a fire). In other words, token causal judgments accumulate evidence about the general effectiveness of potential future interventions. We offer a formal account of this process, and show how it explains three basic qualitative features of causal judgment: why the causes people select tend (1) to be necessary, (2) to be abnormal, and (3) to lack abnormal counterparts.

Immoral Professors and Malfunctioning Tools: Counterfactual Relevance Accounts Explain the Effect of Norm Violations on Causal Selection

“…A central question in research on causal cognition concerns the role of norms. It is well-known that both statistical and moral norms influence judgments of actual causation (i.e., a judgment that some particular event, e, was the cause of some particular outcome, o) (Alicke, 2000;Gerstenberg, Goodman, Lagnado, & Tenenbaum, 2015;Hitchcock & Knobe, 2009;Kominsky, Phillips, Gerstenberg, Lagnado, & Knobe, 2015;Morris, Phillips, Gerstenberg, & Cushman, 2019;Morris et al, 2018). Specifically, when some outcome o depends on the occurrence of a set of antecedent events, e 1 Àe n , people are more inclined to select a given antecedent event e i as the cause of o if e i was either very unlikely to happen or morally prohibited.…”

When do we think that X caused Y?

“…This study replicated several classic empirical findings about causal judgments, and uncovered several new patterns, thanks to its unprecedented high resolution. [1,9,17,18,19,20]. Their work thus makes two things possible.…”

A counterfactual simulation model of causal judgments for physical events

“…In general, people have a tendency to select abnormal over normal events as the cause of an outcome (Gerstenberg, Halpern, & Tenenbaum, 2015;Hart & Honoré, 1959Hilton & Slugoski, 1986;Hitchcock & Knobe, 2009;Icard et al, 2017;Kominsky et al, 2015). While the preference for abnormal causes has long been noted as an empirical phenomenon (Hart & Honoré, 1959Hilton & Slugoski, 1986;Hitchcock & Knobe, 2009), there are now also a number of accounts that quantify how normality considerations affect causal judgments (Halpern & Hitchcock, 2015;Icard et al, 2017;Kominsky et al, 2015;Morris et al, 2018).…”