Plausibility of Counterfactual Conditional

comparison of Counterfactual Potency and Conditional Probability

Abstract

Over et al. (2007) propose that the plausibility of a counterfactual conditional of the form ”If it had been p, then it would have been q” can be assessed by P(q|p), the conditional probability of q given p. Petrocelli et al. (2011), on the other hand, contend that counterfactual plausibility can be captured by counterfactual potency (CP), which is the product of joint probabilities P(p)*P(q|p), where P(p) is the probability of p occurring and P(q|p) is the conditional probability of q given p. CP has been demonstrated to accurately predict counterfactual effects, such as regret (”I wish I had done p better”). The current study examines which model, P(q|p) or CP, better accounts for counterfactual plausibility, causality between p and q, and emotional responses such as regret. We presented participants with a counterfactual conditional sentence ”If I had drawn the red lottery ticket, I would have won,” along with information on event frequency in the form of ”There are ten red winning tickets.” Participants rated counterfactual plausibility, causal strength and emotional intensity. Based on their ratings, we compared the predictive accuracy of P(q|p) and CP for plausibility, causality and emotions. The results indicate that CP predicts counterfactual plausibility, P(q|p) predicts causality between p and q, and P(p) predicts emotion intensity. We discuss the rationale behind these findings.