Like the young philosophers of his generation, Claude Lévi-Strauss first rejected the heritage of Durkheim, brought back to a conservative moralism, before hailing his work on classification systems. However, the Durkheimian method strongly inspired his works through the character of Marcel Granet, author in 1940 of an essay on kinship categories in ancient China, which already establishes the structural typology of the elementary forms of kinship and describes its implications for the choice of spouse and the social cohesion. Magnifying the influence of Roman Jakobson, Lévi-Strauss minimized, even denigrated, the contribution of Granet, and this, for forty years, before admitting it to some extent in 1988. The article explores the reasons of this attitude and attempts to identify how Lévi-Strauss, via Granet, is indeed in the wake of the political philosophy of Durkheim.