Rational Choice and Barriers to Exponential Complexity in Social Optimization Computations under Uncertainty

The main results here are a pair of social-cognitively significant necessary conditions for spin glass complexity that support rational choice in the social optimization problem under uncertain mean social returns and their correlations:(N1) the random mean social returns must be small relative to correlations, and (N2) the social portfolio weights or commitments to separate social activities cannot be continuously adjusted. For certain applications of social optimization computations under uncertainty where these hold, the further demonstration of exponentially large sets of near optimal solutions rules out rational choice.