The article of Simiand was written against a conception of history which he calls"historisante" and which we are in the habit of calling positivist. Neither of these two labels is very satisfactory, and it is quite possible that the complex of practices covered by this denomination has not been studied carefully enough. Basically, this is the type of history which feels that the essential task is to establish the facts with as much certainty as is possible(according to the rules of erudition). It is assumed that these facts are data whose meaning is previously given, and that all that is necessary is to reconstruct the original reality. Each fact is thought to constitute a seir- sufficient unit, and the facts may be ordered by themselves into a sort of objective narrative, a sort of plot that is, the chronology of evolution in progress. The task of the historian is only to make the story visible and certain. The critique of Simiand is part of the questioning of the notion of progress and of the project of constituting a positive social science. In this perspective, the isolated fact means nothing. The fact has to be inserted into a series which will make it possible to determine patterns and therefore explanatory laws. The temporal dimension no longer constitutes the restric- tive frame of reference of a linear chronology, it makes it possible to study variations and recurrences. It serves as a laboratory for research which declares itself from the outset to be comparative. The classification of social facts leads to the identification of stable systems"If then, the study of human facts tends to offer explanation in the scientific sense of the word its main purpose is to identify stable, welldefined relations which may appear between phenomena"The whole of Annales is virtually contained in this definition: history as problem, the search for models, the long term the convergence of the social sciences, and even the encouragement of collective work, the research project, which was to play such a large role in subsequent historiography.