Methodological Foundations for Bounded Rationality as a Primary Framework
Experimental observations have shown that economic agents behave in ways different from the maximization of any utility function. Herbert Simon sought to deal with this by positing that individuals do not maximize, but rather "satisfice." This was a radical departure from the traditional economic framework, and one that still has not been adequately formalized. But Simon's suggestion is only the smallest part of what is needed for a theory that reflects the actual behavior. For instance, Simon's framework cannot deal with the observation that the act of choice changes the chooser. This dissertation is further developing Simon's original ideas through embracing John Dewey's transactional thinking to attain an adequate theory of economic choice that accounts for boundedly rational agents.
I clarify that substantive rationality and bounded (procedural) rationality share the same basic utilitarian assumption of predetermined goals. In terms of a Deweyan (transactional) analysis, the idea of utilitarian "optimization" ultimately guides and constrains both theories. But empirical study of choice behavior and the behavior of subjects in experimental laboratories, both indicate that neither substantive nor procedural rationality can effectively account for actual economic choices.
I emphasize the importance of treating bounded rationality without reference to the rational framework. To me, bounded rationality implies a realistic picture of behavior, which is associated with emerging goals and not ones that exist prior to the making of a choice. I consider uncertainty as a normal characteristic of the situation, which in turn allows consideration of acting based on inconsistent information, just as people actually do. The basis of a systematic approach to behavior that can capture inconsistency is developed by Tom Burke. He mathematizes Dewey's logic. He allows for impossible worlds in the set of states. Thus, not only can the initial state space hold inconsistent states, the information set can include mutually inconsistent elements.
So the current neoclassical paradigm resembles the representative realism, but is there any good reason why we should accept this methodology as economists? Whatever one's ultimate metaphysics and epistemology, I want to show that an alternative approach to economic decision-making may prove highly useful in theory and practice.