But who wrote this text, which proposed in such a startnng way coordinate research in the social sciences? It was the work of a Durkheimian sociologist, who therefore represented a new scientific method, aggressive, but still restricted, and for a long time to come, to the fringes of the univer sity and of the social world. It is not without significance that the two basic statements were pronounced from the periphery of the institutions of knowledge. The first statement, on the unity of social science, was soon to be denounced as sociological imperialism. The second statement, on the place of history within it, so upsetting to traditional historians, gave to history a very particular role. While, in principle, nothing was to distinguish the prac tice of the historian from that of the sociologist, or of the economist, or of the geographer, history was nonetheless ussigned the role of the empirical testing ground for theories to be developed by others. The temporal dimension provided the only possibility of experimentation for those scientific areas studying facts which are, by definition, non reproducible, at least as the term is used by the exact sciences. The practice of history came to be assigned a double role: that of one social science among others, responsible for accounting for past phenomena, and the more specific role of an experi mental social science, the adjunct or testing a ground, as you wish, of the other social sciences. At the beginning of the twentieth century, history was given a role comparable to the one played by the ethnological voyage in the foundation and the critique of eighteenth-century anthropology, comparable too to the role that ethnology had once again to play in the redefinition of the social sciencen in the middle of the twentieth century.