Liam Kofi Bright

I discuss the construction of methods of comparing “degrees of incoherence”. These are methods of comparing agents which give a precise sense to the idea that agents can violate norms of probabilistic reasoning more or less severely. I discuss two previous attempts at constructing such a method, demonstrate that they lead to different rankings of incoherent agents in at least one case, and introduce adequacy criteria which would allow us to decide between these efforts. I state a result that tells us one of the methods is only provable adequate by these criteria in a relatively narrow class of circumstances.

(Note: Some people have had trouble opening this in Firefox browsers. It seems to work reliably in Chrome)