Two Kinds of Universals and two Kinds of Groups

Philosophers have long debated about two ways of conceiving of universals: as Platonic universals and as Aristotelian universals. Roughly, Aristotelian universals are inherent in the particulars that instantiate them; they can be multiply located (located just where the instances are located), and they exist only if they have at least one instance. Platonic universals, by contrast, are truly abstract objects: their existence is independent of the particulars that instantiate them, and they are not located in space and time at all. This historical distinction between two conceptions of universals, I would like to suggest, can be construed in terms of two ways of assigning properties to entities: Aristotelian universals are assigned properties only on the basis of properties of or relations among their instances (e.g. an Aristotelian universal has a locational property P just in case some instances of that universal has P). By contrast, platonic universals are assigned properties just like ordinary objects (Platonic universals thus can't be located in space and time because there is no particular from which they could inherit a location). If the distinction between Aristotelian and Platonic univerals is conceived in this way, then the view becomes plausible that both sorts of universals exist, distinguished from each other by the way their properties are fixed. In this paper, I will argue that such a distinction in the way of assigning properties to entities plays an important role in the semantics of natural language, namely first in the semantics of nominalizations that refer to universals (wisdom, the property of being wise), and second, in an analogous way, in the semantics of plurals. Natural language, I will argue, allows reference to two kinds of universals and two kinds of groups which are distinguished from each other by the way their properties are fixed: namely either in the ordinary way (as with individuals) or on the basis of properties of or relations among instances or group members, in particular by projecting properties of instances or group members onto the entire kind or group. It is this distinction, I will argue, rather than some formal distinction among two sorts of entities, that helps explain some crucial facts about nominalizations and plurals, such as different readings of certain classes of predicates, the way existential constructions and intensional predicates are understood, and the possibility of distributive interpretation.