An account of the ontological nature of collectives would be useful for several reasons. A successful theory would help to show us a route through the thicket of views known as “methodological individualism”. It would have a bearing on the plausibility of legal positivism. It would be relevant to the question whether collectives are capable of acting. The debate about the ontology of collectives is therefore important for such fields as the theory of action, social and political philosophy, the philosophy (...) of law and the philosophy of social science. I hope to contribute to the debate by proposing an account of collectives, and by showing its theoretical soundness and utility. In its essentials, my proposal is that collectives are “mereological sums” of “stages” of persons linked by a “unity relation”. The influence on this account of some theories of personal identity will perhaps be obvious. But the motivation is different, for it lies at least as much in the ability of the theory to help us deal with the issues already mentioned as in problems concerning the identity of collectives over time. I acknowledge that my proposal is not commonsensical; however, I believe that no account could be both commonsensical and adequate. I will begin by introducing some constraints on a theory of collectives. (shrink)