Abstract
Traditional theory allows for uncertainty in the form of risk or random shocks, without altering the form of the problem of rational choice. Theories of bounded rationality introduce genuine constraints upon the chooser, such as limitations upon the chooser's processing capabilities relative to the complexity of the problem. Three analyses in this newer tradition are examined here (Heiner, Nelson and Winter, and Simon). All three tend to conclude that complexity will engender behaviour governed by rules and routines. This leaves us short of a thoery to account for changes in the routines themselves. The paper shows how a learning procedure must be introduced which renders the process of change endogenous. Rationality in this context must involve search activity which is linked to the recognition of a problem situation and ends up enlarging the set of possible alternative solutions.

This publication has 12 references indexed in Scilit: