Environmental economists have traditionally addressed the issue of choice under uncertainty by applying the expected utility hypothesis. When policymakers face a panel of experts with widely divergent beliefs, the problem may more accurately be described as a case of choice under pure uncertainty. Arrow and Hurwicz (1972) demonstrate that rational choice criteria under pure uncertainty will focus on the extremes of the state space, not the midpoint as is typically assumed. Using Nordhaus's DICE model, uncertainty-based choices about global warming policies are evaluated. Linkages to the safe minimum standard and the precautionary principle are also discussed.