- One prominent answer to the question of what the correct theory of parthood and composition is for material objects, which has been embraced by three-dimensionalists and four-dimensionalists alike, is that these objects are best viewed as ‘mereological sums’, ‘fusions’, or ‘aggregates’, according to a particular, standard, conception of mereology, viz., the family of systems formally analogous to Henry Leonard’s and Nelson Goodman’s ‘Calculus of Individuals’ ("Leonard (Henry S.) & Goodman (Nelson) - The Calculus of Individuals and its Uses", 1940) and referred to in "Simons (Peter) - Parts: A Study in Ontology" (1987) as ‘Classical Extensional Mereology’ or (CEM). (A brief synopsis of the basic principles of standard mereology is given in section 2.) Mereological sums, according to this standard conception, are like sets in that their existence and identity depend on nothing more than the existence and identity of the parts that compose them; no structural requirements are set on the manner in which these parts must be arranged.
- The relative merits of standard mereology in the context of giving an analysis of material objects have received quite a bit of attention from metaphysicians in recent years. A question that has not been pursued to the same degree, however, is what sort of semantic repercussions a commitment to mereological sums in the standard sense might have in particular on the predicted behavior of singular terms and our practices of using such terms to refer to objects. The basic problem in this connection for those who believe that standard mereology has a place in a metaphysical analysis of material objects, it seems to me, is that our practice of using singular terms to refer to objects, at least on the face of it, pretty obviously does not track mereological sums: the objects we are interested in and care about, and hence which are represented in our discourse by means of singular terms, generally do not exhibit the sorts of persistence conditions we would expect them to have if these objects were in fact correctly analyzed as mereological sums in the standard sense. We are much more likely to talk about individual trouts and turkeys than we are to talk about trout-turkeys1, to use David Lewis’s well-known example.
- The apparent mismatch between our actual referential practices and the persistence conditions attributed to material objects by the supporters of standard mereology puts these philosophers, other things being equal, at a disadvantage compared to those whose ontology matches more closely the observed behavior of singular terms, as they are commonly used in ordinary discourse. In particular, as will become apparent below, the proponents of standard mereology suffer from self-imposed handicaps with respect to the sorts of constraints on reference to which they can appeal in the context of offering an account of why our singular terms have the content-determining powers they appear to have. If they believe, as David Lewis does (see especially Lewis 1983, 1984), that some constraints on reference originate from the side of the object referred to (as opposed to the speaker or the speaker’s linguistic community) and that the constraints in question are not exclusively causal in nature, then additional machinery is needed to make up somehow for the fact that ordinary language-users speak as though there are trouts and turkeys but not trout-turkeys. To this end, David Lewis leans heavily on his distinction between natural and non-natural properties: a highly eligible referent, according to Lewis, tends to be one whose boundaries are well-demarcated with respect to the highly natural properties; the distinction between natural and non-natural properties, in his view, ultimately derives from fundamental physics.
- While I am sympathetic to Lewis’s position that a credible account of reference will need to appeal to the presence of well-demarcated boundaries among the objects referred to, I am skeptical that his already heavily burdened natural/non-natural distinction can bear the additional weight of drawing boundaries in the right places around the referents of our singular terms: for conglomerations of undetached rabbit-parts, for example, will in general contrast with their environment just as much as single rabbits do, when it comes to such considerations as, say, the density of matter, the relative abundance of chemical properties, favored loci of causal chains, and the like, in the relevant regions of spacetime. Thus, Lewis’s natural/non-natural distinction among properties, as I will argue below, is not sufficient to avoid Quinean indeterminacy for singular terms. Those who are in the business of giving an analysis of constructions involving full-fledged predication, as opposed to constructions that presuppose merely spatial overlap among denotations (see Evans 1975), will thus want to go in for an ontology that places more stringent structural constraints on the referents of singular terms than are supplied by standard mereology.
See Koslicki - Mereological Sums and Singular Terms.
Text Colour Conventions (see disclaimer)
- Blue: Text by me; © Theo Todman, 2019
- Mauve: Text by correspondent(s) or other author(s); © the author(s)