Formal tools for confusion

Dave Ripley (University of Melbourne, University of Connecticut)

There are a bunch of familiar problems around Vesper and Lucifer: what happens when we treat one thing as though it were two? There are also a bunch of problems, although perhaps less familiar, around the reverse kind of case, where we treat two things as though they were one. Following Millikan and Camp in particular, I'll call this kind of thing `confusion'.

Here I want to look at some desiderata for formal approaches to confusion. How do we design a formal system to capture important aspects of this phenomenon? I'll point to two key desiderata: a confused system should not treat the "confusanda" differently from each other, and confusion should not invalidate any valid arguments. Building on these desiderata, I'll give a way of generating, for any consequence relation of a certain sort, a confused version of that relation, and explore some of the features of this way of proceeding.