Fri Apr 14, 2017
234 Moses, 12:30–2 PM
|Meaning Sciences Club
Cian Dorr (New York University)
Conditionals, Closeness, and Probability
Abstract: The modern era in the study of conditionals begins In 1968, with the publication of Robert Stalnaker’s seminal paper ‘A Theory of Conditionals’. In this paper—based on a book I am writing, very slowly at irregular intervals, with John Hawthorne—I will defend a theory closely modeled on Stalnaker’s: ‘If P, Q’ is true just in case either there is no accessible ‘P’-world, or the closest accessible ‘P’-world is a ‘Q’-world. The notion of ‘accessibility’ here is taken to be highly context-dependent, and generally different for indicatives (where accessible worlds are required to be epistemically live for some relevant agent or group) and counterfactuals (where there is no such requirement, and accessibility is typically a matter of match with respect to some body of historical fact). After briefly defending some central features of this view, I will focus on the extent to which it can capture certain plausible generalisations about the probabilities (chances or rational credences) of conditionals, such as the generalisation that if the current chance that P is positive, the current chance that if P it would be that Q is equal to the chance that P-and-Q divided by the chance that P (what is called the ‘conditional chance of Q on P’). Drawing on the work of Bas van Fraassen, I will show how, by divorcing the notion of closeness from that of similarity and thinking of it instead on the model of a random process—as if God had picked a sequence of worlds one by one out of a hat—we can recover quite strong, though not entirely unrestricted, versions of such generalisations.