Causing problems: The nature of evidence and the epistemic theory of causality

The epistemic theory of causality maintains that causality is an epistemic relation, so that causality is taken to be a feature of the way an agent represents the world rather than an agent-independent or non-epistemological feature of the world. The objective of this essay is to cause problems for the epistemic theory of causality. This is not because I think that the epistemic theory is incorrect. In fact, I spend some time arguing in favour of the epistemic theory of causality. Instead, this essay should be regarded as something like an exercise in stress testing. The hope is that by causing problems for a particular version of the epistemic theory, the result will be a more robust version of that theory.

My gripe is with a particular version of the epistemic theory of causality, a version that is articulated with the help of objective Bayesianism. At first sight, objective Bayesianism looks like a plausible theory of rational belief. However, I argue that it is committed to a certain theory of evidence, a theory of evidence that recent work in epistemology has shown to be incorrect. In particular, objective Bayesianism maintains that evidence is perfectly accessible in a certain sense. But evidence just is not so perfectly accessible, according to recent developments in epistemology. However, this is not the end of the line for the epistemic theory of causality. Instead, I propose an epistemic theory of causality that dispenses with the assumption that evidence is perfectly accessible in the relevant sense.