A synthesis view of counterfactuals
Abstract: There are two prominent accounts of counterfactuals in the literature—variably strict conditional accounts first developed by Stalnaker and Lewis, and strict conditional accounts defended by von Fintel, Hájek, and others. Unfortunately, both accounts face serious challenges. In the first half of this paper, I argue that existing strict conditional accounts fail to accommodate our probabilistic judgments about counterfactuals. The same goes for several other semantic theories of counterfactuals.
Having presented these challenges, I introduce and defend my own positive account of counterfactuals. The account incorporates a key insight of variably strict accounts–namely, that the antecedent of a counterfactual often influences what worlds are relevant to its truth conditions at a context. But unlike variably strict accounts, my view preserves a strict conditional semantics according to which Antecedent Strengthening is valid, thereby capturing the data that have motivated other strict conditional theorists. By adopting the right account of the pragmatics of counterfactuals, we can endorse a strict conditional semantics and still straightforwardly capture the full range of our intuitive judgments, including probabilistic judgments about counterfactuals.