No CrossRef data available.
This article raises two difficulties that certain approaches to causation have with would-cause counterfactuals. First, there is a problem with David Lewis's semantics of counterfactuals when we ‘suppose in’ some positive event of a certain kind. And, second, there is a problem with embedded counterfactuals. I show that causal-modeling approaches do not have these problems.
I would like to thank Stephen Barker and Jonathan Schaffer for comments on some of this material.