¹ “Human Society and Scientific Law,” XVIII, no. 2, 05, 1952, 184–94.Google Scholar

² New observations may, of course, invalidate previously accepted propositions and require their revision. However, no matter how many of these previously accepted propositions are falsified, there will be some accepted framework into which the new observations may be fitted. At worst, this framework may consist of basic postulates, in which case we would rebuild the particular science from scratch. Thus the “principle of permanent control” does not affect the point made in the text which is simply an explication of the term “explanation” and as such independent from the rules of procedure to be observed in the process of explanation. See Kaufmann, F., Methodology of the Social Sciences (New York, 1944)Google Scholar; also Feigl, H., “Some Remarks on the Meaning of Scientific Explanation,” Psychological Review, 1948 Google Scholar, reprinted in Readings in Analytical Philosophy (New York, 1949).Google Scholar

³ Geiger argues on p. 185 that the weakening of the concept of strict causality in physics has “broken down” the barrier of method between the natural and the social sciences. This I believe to be a misunderstanding about the nature of this “barrier.” It does not hinge on the specific form in which propositions are formulated; nor does Heisenberg's principle of indeterminacy resolve the difficulties met by the social scientist. The problem is that of the “unique event.” I suspect that Geiger overestimates the change which has occurred in the attitude of the physical sciences and somewhat underestimates the complexities of the problem at hand.

⁴ See, for example, Cramér, Harald, Mathematical Methods of Statistics (New Haven, Conn., 1946), ch. XIII.Google Scholar

⁵ See Weber, Max, Methodology of the Social Sciences (Glencoe, Ill., 1949)Google Scholar, especially “Critical Studies in the Logic of the Cultural Sciences”; also Kaufmann, F., Methodology of the Social Sciences, 178.Google Scholar

⁶ See, for example, Vining, R., “Methodological Issues in Quantitative Economics,” American Economic Review, XL, 1950, especially pp. 272–9.Google Scholar Vining summarizes approvingly Knight's, F. H. position as saying: “The social sciences are unique in that the behavior and change and variation that are to be dealt with cannot be included within fixed classes” (p. 273).Google Scholar In trying to support this claim, he falls into the error of “proving” a postulated assertion about the real world by a purely formal argument.

⁷ No attempt has been made to define the “event” rigorously and fortunately there is no need to do so. Actually the “event” is defined, in each case, by the question asked. Thus we may be interested in the Thirty Years' War, a particular campaign in the war, or the action of a single soldier in a fleeting moment during a batde. Again, we may be interested in a vast galaxy in space or in the brief brilliance of an exploding nova. In each case the “event” will be differently bounded and defined.

⁸ Nothing is said here as to whether these sets are finite or infinite, demimerable or not. In fact, the very concept of “all” properties of an event does not seem to be meaningful. A discussion of these issues, important as they are, transcends the rather modest scope of these notes.

⁹ Henceforth the terms “historical” and “unique” are used interchangeably.

¹⁰ No reference is made here to the psychological processes involved in arriving at a decision. We are only concerned with the methodological nature of the problem, with explication rather than explanation.

¹¹ See, for example, Shackle, G. L. S., Expectations in Economics (Cambridge, 1949)Google Scholar; also Carter, C. F., “Expectations in Economics,” Economic Journal, LX, 1950.Google Scholar

¹² Whether such knowledge is explicitly organized or whether it enters decision-making as the result of non-verbalized processes is here immaterial. This is a psychological question and must be dealt with by the psychologist. Therefore also the speed with which decisions are arrived at is irrelevant in the present context, as is the question of why there are poor and brilliant historians.

¹³ “Human Society and Scientific Law,” 191.

¹⁴ See, for example, zoology and botany.

¹⁵ This holds true for functions of several variables such as occur frequently in the social sciences. For example, consider:

Then

where each partial derivative times its differential in the right-side expression, gives the effect on F of a change in a particular variable, and dF is the result of all these changes.

¹⁶ A particular science or discipline is defined by a bundle of methods shaped to deal with a set of properties of the observable phenomenon. Thus the distinction between disciplines is itself arbitrary and subject to constant change. At any point in time the division of knowledge into various disciplines is a function of the development of knowledge.

¹⁷ Economics has traditionally formulated propositions about human behaviour without explicitly introducing psychology. Classical and neo-classical theory proceeded on the basis of some simple behaviour-postulates such as the “rational principle.” In recent years this procedure has been subjected to increasing criticism as evidenced, for example, by the developments in the theory of the firm. Apparently psychology tends to assume a more important role in the economist's reasoning and research.

¹⁸ Patterns of Culture (Books, Mentor ed., New York, 1948), ch. IV.Google Scholar

¹⁹ F. Kaufmann's position on this point is unfortunately surrounded by some vagueness. On p. 168 and especially on p. 177 his argument appears similar to the one developed in the text. Yet, on p. 178 he asserts: “Usually a physical explanation or prediction is based on a rather small number of facts…. A larger number of factors must be taken into account if we are to explain or predict social facts…. But we must not regard it as a fundamental difference, derivable from the nature of the two kinds of facts. If we do so, we are prone to accept the thesis that historical science is purely descriptive, that there are no historical laws. But this view is untenable.” (Methodology of the Social Sciences.) It is not clear what Kaufmann means here by historical laws but this passage suggests that he thinks of them in a fashion somewhat similar to the one adopted by Geiger.

²⁰ Such regularities may appear in the form: oppressed people always rebel, and the like.

²¹ The designation of science is not a badge of honour and no claim to superiority goes with it. It is a useful methodological concept and little can be gained by diluting it. This may be the place to raise an objection against Geiger's conception of the significance of quantification, measurement, and experiment. He correctly lays stress on the importance of quantitative procedures. Yet, by apparently banning non-quantifying theorizing to the limbo of “humanities,” he seems to confuse the distinctions: empirical (testable)–metaphysical; quantitative-qualitative; operational-non-operational. Most disciplines pass through a pre-quantifying stage. Pyscho-analysis today certainly does not quantify; yet one would hesitate to consider it as non-scientific. General theoretical models at first usually are non-operational. In economics quantification has not, as Geiger claims, been introduced through econometrics, although it may perhaps be said that experimental and measuring procedures have been greatly stimulated by this approach. Economists, on the other hand, have thought in quantitative terms since the sixteenth century as evidenced by the writings, for example, of Copernicus and Jean Bodin. Certainly the neo-classical school quantified and, indeed, used mathematical language. What can be said is that these theories are essentially non-operational. On the other hand, econometrics, in spite of its mathematical sophistication, in parts exhibits tendencies towards that mere description of facts which Geiger himself designates as historiography and non-scientific (p. 185).

²² As pointed out above, the concepts of “all” properties and of “fully” explained are most likely meaningless.

²³ See Hempel, C. G., “The Function of General Laws in History,” Journal of Philosophy, 1942 CrossRefGoogle Scholar, reprinted in Readings in Analytical Philosophy. I would like to take this opportunity to indicate my indebtedness to this brilliant paper. See also Sombart, W., Die Drei Nationalökonomien (München, 1929, 1950), 311.Google Scholar

²⁴ See above, section II.

²⁵ Neglecting here the technical difficulties connected with ascertaining the facts of past events. These difficulties pertain to historiography and are, in the present context, of minor importance.