This article argues that Lockwood's piece on social and system integration was quite insightful at the time, but that it failed in correctly solving the issues it raised. More recent versions of the distinction, such as Giddens's and Archer's, are deemed not satisfactory either. Three steps are then taken in order to provide an alternative solution, which tackles the general problem of the relationship between structure and agency. The critical analysis of the notion of ‘emergent properties’, along with those of ‘structure’ and ‘mechanism’, is the groundwork for a discussion of collective causality that substantiates an alternative view which deÞnes social systems as collective subjectivities, including some methodological considerations. A broad notion of realism is upheld in the article.