Much of mainstream organization theory has been concerned with the implications for organizational design and policy process of high levels of uncertainty or complexity in task environments. Decentralization, disjointed incrementalist decision strategies, and quasi-market coordinative mechanisms have been advanced as rational responses to the complexity of most problems in the socio-political sphere. This article presents and illustrates four conditions which reduce the relative utility of this approach as a means of coping with uncertainty. The propositions are shown to be implicit in the logic of “muddling through,” and are used to help explain/predict the evolution of relatively centralized and planned organizations in certain types of complex task environments.