In Defense of a Probability Based Semantics for Counterfactuals

In a recent paper Lee Walters criticizes a number of philosophers - including Gundersen - for committing a failure in the argumentative strategy' when they attempt to amend the standard Lewis semantics for counterfactuals in order to avoid the so-called principle of Conjunction Conditionalization. In this article we defend a Gundersen-style probability-based semantics against Walter's major misgivings: that it is not logically conservative, that it is committed to the Connection Hypothesis, and that it cannot deal satisfactory with irrelevant semi-factuals.