probabilities assumed in Knightian decision theory. I argue that some standard uses of classical inference are Knightian in spirit …