THE AMERICAN JOURNAL OF SOCIOLOGY
in taking too large a unit. To reach inductively true laws of suc- cession, we have only to pass to the little series of transforma- tions that occur repeatedly in the life of a single society. Such are the consecutive changes by which a luxury becomes a conven- tional necessity, a difference in means becomes a difference in status, an elective head becomes a hereditary head, a usurping dynasty becomes legitimate, an innovation becomes orthodoxy, a custom becomes a right, a vice becomes a sin. Such is the cycle that lies between two conquests or two economic crises, or two revivals of religion. Thus from numerous cases it is possible to formulate the normal development of an innovation or a fashion, to declare what is typical in the formation of a myth, the fixa- tion of a tradition, the canonization of a hero, or the assimilation of an immigrant.
In social life there are indeed ricorsi, only they are much more minute and numerous than Vico supposed. It is only the petty phenomenon that is often repeated. The bane of sociology has been the employment of large units, the comparison en bloc instead of the comparison en detail. Parallels have been drawn between the English Revolution and the French Revolution, between Caesar's usurpation and Napoleon's, between classic society and modern society, between England and Carthage, between the Roman empire and the British. We have, further- more, the supposed similarity of all nations with the same form of government, of all civilizations developed in the same climatic zone.
Tarde is perfectly right when he says: "This attempt to con- fine social facts within lines of development, which would compel them to repeat themselves en masse with merely insignificant variations, has hitherto been the chief pitfall of sociology." Advance there will not be until, renouncing the comparison of a few huge and only superficially integrated complexes of phe- nomena such as nations, epochs, and civilizations we conde- scend to compare and group together numbers of small and elementary social facts. Instead of generalizing on the basis of a few gross and fanciful resemblances, we ought to generalize on the basis of numerous minute and exact resemblances. Just
