Friedman contends that an hypothesis can be tested only by the conformity of its predictions or implications with observable phenomena and not by comparing its assumptions with reality. He applied the ideas developed by Karl Popper for use in the natural sciences in the realm of economics. Friedman argues that, as in the natural sciences, theories should be accepted provisionally or rejected only on the basis of the degree of correspondence of the predictions of a theory with the factual evidence obtained. Friedman asserted that a single empirical counterexample to a theory's predictions will falsify the theory. His idea of the role of testing thus involves induction via elimination or rejection. Friedman's approach is similar to the pragmatism of John Dewey.

          For Friedman, the significance of a theory is not considered to be a direct result of the descriptive realism of the theory's assumptions. In fact, he extols the virtues of descriptively false assumptions. According to Friedman, "Truly important and significant hypotheses have assumptions that are wildly inaccurate descriptive representations of reality, and in general the more significant the theory, the more unrealistic the assumptions." Friedman explains that a person cannot say that a theory's assumptions are true because any of its conclusions are true but that a person could say that there must be at least one assumption that is false whenever some specific conclusion is false.

          Friedman states that for a theory to be useful it must be an oversimplification of reality and therefore a false picture of reality. He goes on to say that, depending upon the theory and its uses, some of these false pictures, along with their assumptions, may be more appropriate (i.e., more predictively accurate) than others. For Friedman, abstraction involves a theory in which many actual characteristics of reality are designated as absent from the theory. Given this perspective, every advance in knowledge, new discovery, or inclusion of additional attributes would defeat, invalidate, and falsify the previous theory. Friedman views any theory as deficient and descriptively false when it fails to specify all of the characteristics of reality including all extraneous, non-explanatory, and irrelevant characteristics. It follows that, for Friedman, all models or theories are imperfect or incomplete because they do not take into account all aspects of the economy. It is no wonder that he turned his attention to prediction rather than to explanation(1).

          It is difficult to reconcile Friedman's practice with his methodological principles. In practice, he frequently argues on the basis of the plausibility or realism of a model's assumptions. After comparing predictions with outcomes he revises hypotheses and assumptions with the intention of making them more realistic. In the light of test results, economists (including Friedman) modify their hypotheses and assumptions in an ongoing and iterative process of observation and theory. Of course, Friedman should be concerned with the realism of assumptions. How else would he know what to attempt next when his predictions fail to square with the facts? If an economist does not assess the realism of his assumptions, the process of theory modification would be inefficient, futile, and based on speculative guesswork. It is apparent that Friedman is not a great theoretical system builder and is even inconsistent with respect to conforming his theory with his practice.

          Ayn Rand's theory of concept formation transcends both Mises' apriorism and Friedman's empiricism. Rand explains that the acquisition of knowledge requires both induction and deduction. What is known is distinct from, and independent of, the knower. Knowledge is therefore gained via various processes of differentiation and integration from perceptual data. Empirical knowledge is acquired through observational experience of external reality. For example, people observe goal-directed actions from the outside and attain an understanding of causality and other categories of action by observing the actions of others to reach goals. Individuals also learn about causality by means of their own acting and their observation of the outcomes. Introspection is a reliable but ancillary source of evidence and knowledge with respect to what it means to be rational, purposeful, volitional, and acting human being.

          For Rand, the essential characteristics of a concept are epistemological rather than metaphysical. Concepts are epistemologically objective in that they are produced by a man's consciousness in accordance with the facts of reality. Concepts are mental integrations of factual data. Rand explains that concept formation involves the process of measurement omission. A concept is a mental integration of units possessing the same differentiating characteristics with their particular measurements omitted.

          Mises explained that a man's introspective knowledge that he is conscious and acts is a fact of reality and is independent of external experience. Mises deduced the principles of economics and the complete structure of economic theory entirely through the analysis of the introspectively-derived a priori idea of human action. While it is certainly important to understand and acknowledge the useful role of introspection in one's life, it is also necessary to realize that its role is limited, secondary, and adjunct to the empirical observation and logical analysis of empirical reality. It would have been better if Mises had said that external observation and introspection combine to reveal that people act and employ means to achieve ends. Introspection aids or supplements external observation and induction in disclosing to a man the fundamental purposefulness of human action.

          Randian epistemology recognizes that Mises' action axiom could be inductively derived from perceptual data. Human actions would then be viewed as performed by entities who act in accord with their natural attributes of rationality and free will. All the essential principles and laws of economics could then be deduced from the action axiom.

          The Randian or Objectivist view is that the best way for judging a model or a theory is to examine the plausibility of its assumptions. As Ayn Rand herself would put it, "Check your premises." Things which the assumptions abstract from should pertain importantly to the problem under examination. An hypothesis, complete with its assumptions, is both explanatory and predictive if it abstracts the crucial and common elements from the mass of detailed and complex conditions surrounding the phenomenon to be explained and furthers valid predictions of that phenomenon. The Objectivist view is that economic and other theories need to omit a lot of details. Randian realism does not require that all nonexplanatory extraneous particulars be specified.

          The Randian conception of abstraction is to attend to some aspects or attributes of a phenomenon and to exclude others. From this perspective, certain actual characteristics are simply absent from specification or measurement. This is far different from and superior to Friedman's view that some actual characteristics are specified as nonexistent or lacking. From a Randian perspective, new knowledge may expand or refine a theory, concept, or model but it does not necessarily contradict or invalidate it. In other words and contrary to Friedman's view, the theory does not have to be descriptively false(2).

          Ayn Rand's methodological approach overcomes the false alternatives of rationalism and empiricism by explaining how abstract knowledge of reality can be soundly derived from valid perceptual experience. Her method for overcoming these dichotomies includes a number of processes such as differentiation, induction, integration, deduction, reduction, measurement omission, and so on.