¹ On the problem of a unit of measure of utility, see Walras, Eléments d'économie politique pure, 4th edition (reprint, 1952), p. 74.

² The theory of chance is protracted by Cournot into a distinction between natural entities and those which are " manifestly artificial": "In the evolution of phenomena, one part follows permanent and regular laws, and one part is left to the influence of prior facts. To suppose that this distinction is not essential is to admit that time is nothing but an illusion" (Essai sur les fonde ments de la connaissance, p. 460). Hence, "a nuvola which the telescope resolves into a mass of stars irregularly grouped is constituted fortuitously, accidentally; whereas the constitution of a sun and planets in a flattened spheroid follows a law or a necessity of nature" (Considérations sur la marche des idées, Boivin, 1934, vol. 1, p. 2). A river is a "system" and cannot be reduced to "a collection of drops of water" (Essai sur les fondements…, p. 242); in contrast, what geographers call a river-basin is carved out artificially by them. One would likewise say that a glacier is a system, a natural object, which has a "reason" and which depends on a glacial system, whereas a mountain, fearlessly carved out by the hazards of erosion, and isolated from the rest of the chain only by the on-looker who regards it as an individual thing, is an artificial object which only has causes (on the difference between reason and cause, cf. Mate rialisme, vitalisme, rationalisme, p. 219). The finest passage is in the Essai sur les fondements de la connaissance, p. 97: "Natural phenomena, linked one to another, form a network all of whose parts adhere to each other, but not in the same manner nor in the same degree. So that one sees the design of a leaf perfectly determined as to its principal ribs, whereas for the last ramifications and for the agglomeration of cells which fill up the intervals and compose the parenchyme of the leaf the fortuitous game of secondary circumstances gives rise to innumerable modifications and to details which are no longer fixed from one individual specimen to another. One strays alike from a faithful interpre tation of nature both in failing to understand systematic coordination in those fundamental traits where it is clearly revealed, and in wrongly conceiving the bonds of coordination and of solidarity in the cases where collateral series, each of which is governed by its own laws after it branches off from the common trunk, no longer share anything but accidental similarities and fortuituous parallels." Cournot here sheds light on a limit of historical nominalism.

³ P. Birot, Les méthodes de la morphologie, P.U.F., 1955, p. 2-17 and 167 (along the same order of studies, cf. R. Brunet, Les phénomènes de discontinuité en géographie, 2^d edition, C.N.R.S., 1970).

⁴ Chance, according to Aristotle, is a determinism which imitates an intention (the "game of nature," the monkeys which, by dint of batting a typewriter, end up by chance with the Iliad; chance according to Poincaré is a determinism whose result could have been reversed by a minimal variation in the initial conditions ("if Cleopatra's nose had been shorter…").

⁵ A sociological treatment of esthetic or political problems, whether it be right-wing or left-wing, dresses up sophistry in scientific clothing. It was Koyré, I believe, who said that Protagoras (of the Protagoras and the Theaetetus) was the ancestor of sociologists. To confront historicism and sociologism, we have seen and today see once again the resurrection of the grand tradition of political philosophy, with Leo Strauss, Natural Law and History, University of Chicago Press, 1953; S. Landshut, Kritik der Soziologie und andere Schriften zur Politik, Berlin, Luchterhand, 1969 (reprint of the studies which have appeared since 1929); W. Hennis, Politik und praktische Philosophie: eine Studie zur Rekonstruktion der politischen Wissenschaft, Luchterhand, 1963; W. Hennis, Politik als Wissenschaft, Aufsätze zur politischen Theorie und Regierungslehre, 1968; C.J. Friedrich, Prolegomena der Politik, politische Erfahrung und ihre Theorie, Duncker und Humblot, 1967; E. Voegelin, The New Science of Politics, University of Chicago Press, 1952. One hardly hesitates any longer to think that sociology is only a denomination that arbitrarily groups diverse activities, and not an element in the system of the sciences, when one remem bers that, until about 1850, another classification reigned over the same pack of goods. According to this classification, one distinguished history, political philo sophy (Aristotle, Hobbes, Spinoza…), the Fürstenspiegel, and lastly "statistics" (one intended by this word the political, geographical, economical, military, psychological, etc. description of contemporary States; Tocqueville would not at that time have been styled a sociologist, but a political philosopher and a "statistician"). In the contemporary atmosphere these old denominations do not appear expert enough; today, a "mirror of princes" would be as old-fashioned as the Henriade or verse-tragedy. We have thus changed words, in place of changing things: sociology is old wine poured into a modern bottle. It remains true, however, that the bottle gives flavour to the wine, and that what one does under the name of sociology has a scientific and sophisticated after-taste.

⁶ The title is significant: Norbert Elias, Die höfische Gesellschaft, Unter suchungen zur Soziologie des Königstums und der höfischen Aristokratie, mit einer Einleitung: Soziologie und Geschichtswissenschaft, (Luchterhand, 1969, 456 pages).

⁷ P. Veyne, Comment on écrit l'histoire, essai d'épistémologie, Editions du Seuil, 1971, chap. XII.

⁸ Gottfried Martin, Leibniz, Logik und Metaphysik. Kölner Universitätsverlag, 1960. (English edition: Leibniz: Logic and Metaphysics, translated by P. G. Lucas and K. J. Northcott. Manchester University Press, 1964).

⁹ La pensée et le mouvant, p. 16: " If there had never been a Rousseau, a Chateaubriand, a Vigny, a Hugo, not only would Romanticism never have been discerned among the Classicals of the past, but it really would not even have existed. For Romanticism only comes into being among the Classicals when we carve out of their work a certain aspect, and our carving, with its particular form, no more existed in Classical literature before the appearance of Romanticism than in a passing cloud exists the amusing design that the artist sees there while organizing the amorphous mass to the liking of his fantasy."

The transformation of the historical perspective according to the issue and sequel of an event is entirely another problem; cf. Raymond Aron, Introduction à la philosophie de l'histoire, essai sur les limites de l'objectivité historique, N.R.F., 1948, p. 133-136. (English edition: Introduction to the Philosophy of History: an Essay on the Limits of Historical Objectivity, translated from the revised Franch edition by George J. Irwin; Boston, Beacon Press, 1961.)