Rationally Selected Utility – A New Theory of Choice

Traditional prescriptive decision theories from Pascal’s to Expected Utility have struggled to explain inconsistent human choices, leading to predictive models like Prospect Theory. This paper introduces a novel approach to understand why these behaviors occur, reconciling these observed "irrationalities" with a simple rational objective. Here we propose that the human nervous system, faced with limited cognitive capacity (a finite representational precision for the representation of reward value), optimally selects a task-appropriate utility function to either maximize long-run average earnings or to minimize errors. We argue that the utility functions we have measured are tools, sculpted by knowledge of the choice environment's probabilistic structure and the decision-maker's precision level, for optimally achieving a rational objective. Our results show that almost every experimentally observed utility function can be described as a near-optimal tool for achieving a maximization of objective earnings (a goal abandoned since Bernoulli) or for minimizing errors. We quantify the specific gains that result from adding cognitive capacity and find that these gains are quite small as long as the utility function is correctly adjusted to cognitive capacity. Our research provides a novel framework that offers a testable and foundational explanation for why human and animal decision-makers evolved the puzzling range of utility functions and risk attitudes that have been widely observed.