Felix Kaufmann, Do Synthetic Propositions a Priori Exist in Economics? A Reply to Dr. Bernardelli (1937)

Published in Economica in August 1937, Kaufmann’s methodological reply addresses Bernardelli’s criticism of his Methodenlehre der Sozialwissenschaften. Its central contention is that principles cannot establish truths about experience independently of experience. Moving from the status of geometry to economic behaviour and marginal utility, Kaufmann distinguishes definitions, whose validity makes no claim about reality, from empirical propositions, which remain open to refutation.

The opening corrects Bernardelli’s account of Kaufmann’s criticism of Kant. Kaufmann had not argued that the mere consistency of non-Euclidean geometries destroyed Kant’s position: it had shaken the identification of geometry with Euclidean geometry and prompted investigation of its physical applicability. Poincaré’s conventionalism and Einstein’s general relativity supplied the decisive challenge. Non-Euclidean geometry establishes that Euclid’s postulate is not analytic; it does not establish that the postulate necessarily governs physical experience. Kaufmann draws the general conclusion:

The general rule which this line of thoughts leads to, runs as follows: if we understand by the phrase "synthetic propositions a priori", principles which, without being based upon experience, can lay down certain rules about experience, then there are no such propositions.

This rejection leaves an important role for philosophical methodology. Scientists employ general assumptions, often without explicitly examining them; methodology should identify, organise, and justify those assumptions. Kaufmann therefore accepts part of Bernardelli’s programme while rejecting the necessary validity that Bernardelli, following Fries and Nelson, attributes to its results. The methodological danger is precise:

The chief danger of this doctrine is that it leads to a confusion between empirical statements and definitions and thus to begging the question.

The economic examples demonstrate how this confusion arises. If certain properties are deemed indispensable to “economic behaviour,” so that behaviour lacking them is excluded from the category, their necessity belongs to the definition. No discovery about human conduct follows. If economic behaviour is instead identified independently of those properties, the assertion that it possesses them becomes testable. Whether conduct follows the economic principle is thus an empirical question only when conformity to that principle has not already been built into the concept.

Kaufmann next separates three meanings commonly assigned to the marginal utility principle. First, it can articulate what is meant by rational valuation of a unit within a given stock. So understood, it is definitional, not a claim about actual conduct. Second, the applicability of rational economic behaviour and marginal utility to actual phenomena is itself an empirical issue:

The statement that this is possible is an assumption about reality and therefore can be subjected to proof and may be shown to be false.

Third, diminishing marginal utility can mean that, within a given planning period, additional units yield less utility once their number exceeds a certain level. Kaufmann treats this as an experiential proposition supported by the introspectively accessible satiation of needs—Gossen’s first law. He explicitly separates that law from the preceding definitional and applicability questions. Confusing these meanings creates the appearance of a principle that is both independent of experience and informative about reality.

The concluding discussion applies the same distinction to the “least important need.” If this means whichever need an individual actually sacrifices, saying that the individual sacrifices the least important need merely redescribes the choice. Explanation requires identifying the need independently of the behaviour being explained; only then does the prediction acquire empirical content.

But when we make use of a definition of the second type, we cannot claim any absolute validity for the proposition that economic behaviour always involves renouncing the satisfaction of the least important need.

The reply’s relevance lies in this exchange of certainty for explanatory content. Kaufmann does not discard rational valuation or marginal utility as economic concepts. He demands clarity about their logical role: definitions organise inquiry, while assertions about their applicability and about human needs must answer to experience. What appears to be a necessary economic law may otherwise conceal a circular explanation.