In his 1991 book, Parts of Classes, David Lewis discusses the idea that composition is identity, alongside the idea that mereological overlap is a form of partial identity. But this notion of partial identity does nothing to help Lewis achieve his goals in that book. So why does he mention it? I explore and resolve this puzzle, by comparing Parts of Classes with Lewis's invocation of partial identity in his 1993 paper ‘Many But Almost One’, where he uses it to address Unger's problem of the many. I raise some concerns about this way of thinking of partial identity, but conclude that, for Lewis, it is an important defence against accusations of ontological profligacy.

What kind of entity is a committee, a book group, or a band? I argue that committees and other such social groups are concrete, composite particulars, having ordinary human beings among their parts. Thus, the committee members are literally parts of the committee. This mereological view of social groups was popular several decades ago but fell out of favor following influential objections from David-Hillel Ruben. Recent years have seen a tidal wave of work in metaphysics, including the metaphysics of parts and wholes. We now have the resources to rehabilitate the mereological view of social groups. I show how this can be done and why we should bother.

Roughly speaking, perdurantism is the view that ordinary objects persist through time by having temporal parts, whilst endurantism is the view that they persist by being wholly present at different times. (Speaking less roughly will be important later.) It is often thought that perdurantists have an advantage over endurantists when dealing with objects which appear to coincide temporarily: lumps, statues, cats, tail-complements, bisected brains, repaired ships, and the like. Some cases – personal fission, for example – seem to involve temporary coincidence between objects of the same kind. Other cases – a cat and its flesh, a statue and its lump – seem to involve objects of different kinds.