Joseph A. Schumpeter, The Meaning of Rationality in the Social Sciences (1984)

Schumpeter’s methodological article distinguishes the rationality of scientific analysis from the rationality attributed to the people it studies. Published in 1984 with an editorial introduction by Wolfgang F. Stolper and Rudolf Richter, it preserves a manuscript marked “Still only a sketch,” probably presented to an informal discussion group in early 1940. The published text comprises sections II–V; its introductory section was omitted and is described by the editors. The argument moves from quantitative economic models through purposive action and conscious calculation to a critical reassessment of economic systems. Its central concern is to defend rational models without confusing their analytical usefulness with universal rationality in human conduct.

Section II separates the observer’s rational procedures from the characteristics of the observed world:

We will emphasize at once that this observer's rationality has in itself nothing whatever to do with the presence or absence of rationality in the human types or human actions observed, or even with the applicability of the concept of rationality to the subject matter under investigation.

Relations among production, prices, interest rates, and demand deposits can yield meaningful economic propositions without specifying individual motives. Although further inquiry would reach acting persons, the model possesses a limited analytical autonomy. Its boundaries cut through real connections, as those of other scientific models do. Economic relations are less stable than physical ones, but this practical difficulty does not establish a different principle of scientific reasoning.

Section III nevertheless rejects restricting social science to measurable relations. Its distinctive resource is the interpretation of action through intelligible ends. Schumpeter’s example is the monopoly model that identifies the output maximizing instantaneous profit. This construction neither reproduces a businessman’s thoughts nor requires him to know marginal revenue and marginal cost:

For a man's behavior may conform to it and be economically described by it, even if its contents are as foreign to his mind as the law of gravitation is foreign to a stone.

Unlike the stone’s motion, however, the firm’s conduct is approached through a norm specifying the conditions for attaining an understood goal. This is “objective rationality”: a rational pattern applicable to the object of inquiry, not necessarily conscious calculation within it. The norm’s logical validity is distinct from its empirical realization; its descriptive usefulness depends on whether the selected end and conditions fit the case.

Schumpeter locates many apparent failures of rational analysis in that selection. Instantaneous profit maximization cannot adequately describe a firm pursuing profits over a longer period. Conflicting ends likewise make conduct appear irrational when tested against a single objective. His example of the abolition of food duties shows that a ruling class might sacrifice economic interests to preserve political power. The analyst must also identify the acting agent correctly: politicians, corporate executives, and union officials need not pursue the interests attributed to their supporters or organizations. Rational individual actions, moreover, can produce collectively destructive results. Economic crises undermine any automatic passage from individual maximization to social welfare. Interpretation across cultural distances introduces further risks when unfamiliar ends are judged through the observer’s own assumptions.

Section IV distinguishes this objective conformity from “subjective rationality,” the actor’s own rational processes. Earlier economists frequently exaggerated conscious calculation and mistook professed motives for genuine explanations rather than rationalizations. Yet eliminating subjective rationality altogether would discard an important causal and diagnostic resource:

But the question itself why it is that reality sometimes conforms to rational schemata is by no means uninteresting.

A firm following a pattern through calculation may react differently to rapid change from one following it through habit or environmental pressure. Knowing the source of conformity therefore matters for prediction, expectations, and adjustment. Conversely, investigating management’s reasoning can explain a model’s failure or indicate a better model. Technological inefficiency may reflect inertia, but it may also be economically sensible when managerial attention earns greater returns elsewhere. A footnote qualifies the terminology: “personal” rationality might be preferable, since practiced rational operations can become automatic without losing their rational character.

This distinction has particular force for democratic theory. The older doctrine presupposes unrealistic clarity of interests, command of means, and responsiveness to rational argument. A reconstructed theory could retain rational interpretive schemata while substantially reducing its assumptions about citizens’ conscious rationality. Schumpeter also recognizes the danger of explaining every action as rational by inventing a suitable end. Inquiry into actual personal reasoning helps resist this tautology.

Section V tests the distinctions against physiocracy, utilitarianism, Marx, and the economics of Marshall, Wicksell, Walras, and Pareto:

His schemata are devised for certain special uses and must be judged by their success in actual application, and not without reference to it.

Physiocratic theology should be distinguished from its achievement in representing economic reproduction. Benthamite economics can partly dispense with conscious rationality, whereas Benthamite politics depends heavily upon it. Marx’s impersonal evolutionary explanation incorporates rational patterns while exposing ideological rationalizations—an exposure that might also rebound upon class struggle itself. Finally, classical economics suffered criticism partly because its language blurred objective and subjective rationality. Pareto removed much hedonistic baggage but retained simplified purposive models. Schumpeter ends by stressing their limitation: separately modeled ends supply no rational rule for combining the whole of life. The article’s enduring contribution is thus a differentiated defense of rational analysis, bounded by empirical inquiry into agents, purposes, mental processes, and unintended consequences.