Multimode utility theory

• A non-boolean model for decision making is proposed.
• Rationality for context dependence and emotion is discussed.
• Topological event spaces are used for modeling probability.

SEU is expanded to incorporate emotion, bias, and other modes that can influence a decision maker’s judgment of utilities of lotteries. In multimode utility theory (MUT), judgments are based on subjective interpretations that often vary with the mode, resulting in the same event having multiple interpretations and the decision maker making mode dependent judgments. In MUT, an event’s multiple interpretations are modeled in a manner so that they are related semantically. The semantic relationship is characterized through topological and algebraic means. This is done in a way that allows for systems of probabilities and decision making for lotteries that are arguably subjectively coherent in the sense that they can be viewed as satisfying standard rationality assumptions. They are not, however, necessarily behaviorally coherent, e.g., Dutch book criteria for rationality may fail. The conflict between the simultaneous holding of subjective coherence and the failure of behavioral coherence poses interesting rationality questions that are discussed in the article, e.g., Which coherence concepts should “rationality” be based on? Are SEU’s primitive concepts too impoverished to provide a reasonably realistic theory for rational human decision making?MUT uses a new kind of event space to characterize the algebraic structure of the space of subjective interpretations. This space bears a close algebraic resemblance to a generalization of classical propositional logic known as “intuitionistic propositional logic,” which has been much studied in the foundations of mathematics and in philosophical logic.MUT provides alternative methods for modeling various phenomena in the literature. In particular, this article uses it to model decision making for lotteries with catastrophic events and outcomes and to model situations where emotions such as fear and hope influence decisions.