A good year for imprecise probability

The theory of imprecise probability (IP) has been slow to catch on in philosophy, although there are signs that this is beginning to change. Gert de Cooman’s link between IP and logic, presented in a terrific talk at progic 2007, is sure to draw interest from philosophical logicians; a general result in de Cooman and Miranda (2007) showing how Finetti’s exchangeability theorem drops out as a special case should receive attention from philosophers of science; and a proposal by Carl Wagner (2007) to get around weak book arguments by replacing the classical theory of linear previsions with IP’s lower/upper previsions suggests some interest in Bayesian epistemology. Even so, the theory of imprecise probability is loaded with surprises, which may explain the spare use of the theory so far. But 2007 was a watershed year for imprecise probability, owing to a paper by Teddy Seidenfeld, Mark Schervish, and Jay Kadane (2007). I want to call attention to this result, which builds on their work on imprecise probability over the last two decades.