Formalism is defined here as the use of an abstract language such as mathematics or symbolic logic in place of natural linguistic or literary methods of presentation. Formal systems were originally developed to investigate the rules of deductive inference, and for this purpose the exact meaning to be attached to the primitive terms employed is unimportant. For economics, however, the interpretation and economic meaning of terms is important, and the use of formal methods usually involves the construction of a “model” that both embodies a formal structure and provides the basis for interpreting the symbolism.

Using the terminology of formal and natural languages may give the impression that it should be possible to translate one into the other without much difficulty. It is true that some arguments can be translated from one language into another with relatively little difficulty or alteration of the content, but this is far from always the case. To formalize a theory is not simply to make it more precise; rather, some aspects of the theory are singled out to produce a highly idealized representation of it. In this, the ambiguity of natural language may be overcome, but at the cost of losing much of its richness and suggestiveness.

Although formalism can be traced far back in the history of economics, it is associated in particular with the development of neoclassical economics. Early neoclassical contributors such as Walras, Jevons, and Edgeworth drew on the example of physics and adopted the mathematics of constrained maximization (Mirowski 1989, 1991).