The Role of Crucial Counterexamples in the Growth of Economic Knowledge: Two Case Studies in the Recent History of Economic Thought

A ‘crucial experiment’ in science is a simple definitive empirical test which either falsifies or corroborates a scientific theory. The ideal crucial experiment excludes all but one of the theories competing as possible explanations. Even a cursory examination reveals that the history of economic science contains few, if any, such crucial experiments. There are of course a variety of well-known reasons (or some would say excuses) for this lack of empirical falsification or attempted falsification within economics. Some of these reasons include the lack of opportunity for laboratory experimentation, the ‘immunizing stratagem’ of ceteris paribus, the general ‘complexity’ of economic phenomena, and the variable nature of the underlying economic system. But regardless of the reasons, the fact remains that if ‘refutations’ or ‘falsifications’ occur at all in economics, they occur as the result of a sustained continuous bombardment of contrary evidence, not as the result of a single strategic crucial experiment. This absence of empirical crucial experiments prompts some rather obvious questions for the history and philosophy of economic science. For instance, what within the development of economics, initiates the type of theory change usually attributed to crucial experiments in natural science? In other words, does economics have a methodological surrogate for crucial experiments? Recently a number of methodologically concerned economists have addressed this question. Goodwin (1980), for example, suggests that theory change is the result of “external crucial experiments,” that is, changes in the social/political/intellectual environment in which the theory functions. Progress is “associated with major external policy challenges or what society takes to be such challenges.”* Remenyi (1979) on the other hand argues that empirical anomalies enter the economic core theory principally through the portholes of subdisciplines (demi-cores) and that change occurs as the result of “core demi-core interaction.” In certain areas of economic science these explanations of progressive theory change are quite appropriate. For instance Goodwin discusses the Keynesian revolution, no doubt a case where such external influences had a significant impact. The problem is, cases like the Keynesian revolution do not exhaust theory change in economics. Much progress occurs in a far more pedestrian way essentially insulated from such external influences. What will be discussed below is an alternative, non-empirical, and nonexternal avenue for the theory change in economics. In some important areas of economic theory, particularly those which have been highly formalized mathematically, theory change is initiated by mathematical ‘counterexamples .’ These counterexamples are not ‘facts’ in any sense, but yet they provide a way of falsifying certain conjectures about the implications of the theory. The effect of these counterexamples on the evolution of the economic theory is very similar to the effect attributed to an anomalous crucial experiment, that is, the abandonment or at least redirection of an area of research. Section I1 below provides two case studies from the recent history of economic theory where such ‘crucial counterexamples’ have significantly affected the development of the theory. Both studies are from Walrasian general-equilibrium theory, but similar cases could be cited for international- trade theory, optimal growth theory, or any other highly formalized area of economic ~c i e n c eW.~h ile these case studies have historical interest in their own right, the primary motivation for this study is methodological. These methodological implications are discussed in Section 111