David Lewis and his mereological interpretation of Zermelo-Fraenkel set theory. A reconstruction (Q2253829)

A typical trait of nominalistic theories is the rejection of abstract objects. Accordingly it is a problem for nominalism to make sense of set theory, which deals with statements about the existence of sets. A nominalistically acceptable `concrete' alternative to sets are `fusions': One can e.g. take the fusion of all cats, which is the (spatially widely distributed) object consisting of all cats. However, fusions behave differently from sets in various respects: For example, the fusion of all cats is identical with the fusion of all parts of cats, while the set of all cats is not the set of all parts of cats. Hence, the question arises whether we can in some way `reduce' set theory to mereology, the study of the relation between parts and wholes. Given a sufficiently formal account of mereology, this turns into the technical question whether ZFC is interpretable in formal mereology. In his book [Parts of classes. Oxford: Basil Blackwell (1991; Zbl 0900.03061)], \textit{D. Lewis} has sketched such an interpretation. The book under review is a formal reconstruction of Lewis' account, proving formally that ZFC is interpretable with parameters in second order mererology, amended with the statement that there is a `strongly inaccessible partition', which can be seen as a kind of merological substitute for a strongly inaccessible cardinal (or rather the corresponding level of the von Neumann hierarchy). The exposition is concise and focuses strongly on the formal details, with little remarks on background, motivation or explanation. It should maybe best be seen as a supplement to Lewis' work. Also, the notation is occasionally a bit idiosyncratic, though the attentive reader is likely to get used to it. The work is accessible to readers with a basic background in mathematical logic and set theory, though some prior acquaintance with (formal) mereology and Lewis' account is recommendable.