The study of religious phenomena was very important for early sociologists. It is to be expected, therefore, that they tried to define religious phenomena precisely. For example, Durkheim, Mauss, and Hubert attached great value to definitions of the social phenomena under research. However, they differed on the meaning of the adjective "social". In this article, I examine the implicit and explicit definitions given by these writers. More precisely, the article focuses on the distinctions they made between religion and magic. It turns out that for these scholars the given definitions and the distinctions between religion and magic are ambiguous. The paper argues that, when it comes to defining religion and magic, there were considerable differences between the central members of the Année sociologique. The paper concludes that differences between Mauss' and Hubert's definitions, on the one hand, and Durkheim's, on the other, can be related both to a difference in their initial motivations for research into religious phenomena and to their differing definitions of the adjective "social".