An Outcome-Oriented Theory of Choice and Empirical Paradoxes in Expected Utility Theory

I analyze observed choice between lotteries from an outcome-oriented point of view in the framework of choice between random variables. I characterize a choice maker, who faces a choice situation between lotteries, by (1) a surmising process that associates, with a pair of lotteries, a set of well-defined choice situations between random variables, and (2) a choice set that is a collection of well-defined choice situations. I give a partial axiomatic foundation of the theory. The theory is applied to explain the well-known paradoxes in expected utility theory.