Chancy Causation and the Problem of Aggregate Events


According to David Lewis's account of chancy causation, an actual event c is a cause of an actual event e if and only if (i) e causally depends on on c, or (ii) there is a chain of events d1, d2, ... dn, such that such that d1 causally depends on c, d2 causally depends on d1, ..., and e causally depends on dn. e causally depends on c if and only if, had c not occurred, e’s probability of occurring would have been smaller *by a large factor*. It is a direct corollary of Lewis’s account that, in certain cases, an event c that only slightly raises e’s probability of occurring cannot be a cause of e. This paper argues that this feature of Lewis’s view has implications that are hard to defend. I present three problems – the Problem of Aggregate Effects, the Problem of Aggregate Causes, and the Problem of Causation and Prevention – which together seriously undermine the attractiveness of a causal dependence approach to chancy causation, lending support instead to an account in terms of causal contribution.

Last updated on 11/10/2019