It is truly a great honor to speak on this occasion. In her work, Nancy Schwartz maintained the highest standards of integrity and of the imaginative and skillful use of mathematics. I cannot pretend that the material I will present achieves the same level of elegance. Rather, I will describe a point of view that, when developed and applied, may lead to new economic insights and even to interesting calculations. But this point of view is still very much at a speculative and experimental stage.

My starting point is the distinction made by Frank Knight (1921) between risk and uncertainty. He defined risk as randomness with a known probability distribution and uncertainty as randomness with an unknown distribution. He argued that uncertainty was uninsurable and that the role of entrepreneurs was to undertake investment involving such uninsurable chances of loss. It is not clear from Knight's work what he thought was special about uncertainty - whether it was market failure due to moral hazard and adverse selection or whether it was differences in behavior in the face of risk and uncertainty. In a recent paper, LeRoy and Singell (1987) argue that Knight had in mind moral hazard and adverse selection. I do not wish to discuss what Knight intended, but I rather wish to propose that, in fact, there is a distinction in people's reactions to risk and uncertainty.