Abstract

The 'assumptions controversy' which mainly took place between Friedman, Machlup and Samuelson did not lead to a satisfactory answer to the question whether 'assumptions' in economic theory should or should not be 'abstract'. If 'abstract', as Friedman suggested, means 'false' then falsity of (i) the antecedent clauses entails inapplicability of any theoretical argument. Falsity of (ii) the law statements in the explanans implies that the aim of theoretical progress is abandoned in favour of applicability of economic theory to only a given set of problems (instrumentalism). While 'abstractness' in the sense of 'incompleteness' of any explanation is unavoidable, semantic (use of theoretical terms) and syntactic 'abstractness' (use of strictly universal statements) turned out to be highly desirable characteristics of theories. Finally, axiomatic structuring of a theory, a last kind of 'abstractness', is a heuristic device for the improvement of the testability and applicability of economic theory.