A view of corruption as disembedded from society and history is predominant today. In this view, corruption is basically the same thing everywhere and inherently a bad thing because it gets in the way of proper processes. In opposition to this view, we argue for understanding corruption as socially and historically embedded. While there are many viable ways to embed corruption, we advocate a comparative historical sociology of corruption in particular. This approach has in mind a view of corruption as “a moving object,” that is, as subject to variation across social space and transformation over time. It focuses on the processes through which a course of action is worked out in relation to historically specific structural conditions. By tracing these processes and embedding “corrupt” practices in the situations where they were developed and make sense, we gain a deeper understanding of these practices and are in a better position to evaluate them. We are also able to make better comparisons, comparing objects shaped by similar processes rather than objects identified by definition alone. We proceed, first, by situating our intervention in the context of the rise of a disembedded approach to corruption. Second, we review a selection of more or less embedded approaches in anthropology and sociology. Third, we describe what a comparative historical sociology of corruption entails. Finally, we highlight the costs of a disembedded view in terms of ineffective and pathologizing anti-corruption efforts.