The debate between Monism and Pluralism concerns how many particulars there are. The problem of universals is different: it concerns how many kinds of things there are. In particular, are there both particulars and universals, or can we get by with just one or the other?
This page concerns whether we can make do with particulars, and leave universals out of our ontology (nominalism), or whether we need to include universals in our ontology (realism). My companion page on particulars takes up the converse issue of whether we need particulars as a separate part of our ontology, or can define them in terms of universals.
| An Inventory of Ontological Views About Universals |
Let's expand our inventory of possible views a little. Possible views include (more or less in order of increasing ontological commitment):
Nominalism. (Quine links this with formalism in the philosophy of mathematics.) There are only particular things; there are no universals. But wait: if there are no universals, then what are we doing when we group objects by their shared properties? Answers include:
Class Nominalism. Classes are abstract, but they aren't universals. (I think.) Any bunch of particulars can be grouped into a set. To say that an object is red (for example) is only to say that it is an element of a particular set of objects.
Resemblance Nominalism. To say that an object is red is to say that it is at least as similar to certain paradigm objects as they are to each other.
Conceptualism. (Quine links this with intuitionism in the philosophy of mathematics.) There are universals, but they are mind-dependent. They are constructed by us (they are concepts or ideas), and exist only in our minds.
Realism. (Quine links this with logicism in the philosophy of mathematics.) There are universals, and they aren't mind-dependent. No, they really, honest-to-gosh exist. --But how many universals are there? Where are they located? Must they have instances?
Universalia in rebus. There are universals, but only where they are instantiated (that is, only where their instances exist). (So they can be entirely in several places at once. But since universals are only where their instances are, there can be no universals without instances.)
Universalia ante rem. There are universals, and they are not located in space and time at all. "Where are they?" is a bad question, since "where" asks for a location; locations are positions in space and time; and universals don't have such positions. But the idea that they exist, but do not exist anywhere in space and time, leads naturally to the Platonic view that there is a sort of separate world of Forms.
| A Small Sampling of Arguments Concerning Universals |
(Mostly arguments from Price, concerning resemblance nominalism vs. realism.)
| Realism | Nominalism | |
| Definition | There are universals as well as particulars. These universals are objective, mind-independent entities. | There are no universals, only particulars. |
| By virtue of what does a predicate (e.g. "round" or "white") apply to multiple things? | A predicate applies to more than one thing if it expresses a universal, and the things are all instances of the universal. | Resemblance Nominalism (Price): resemblances are a basic characteristic of the universe, and don't need to be explained by universals |
| Problem 1 | How can this view accommodate the fact that some things are only sort of white or round or whatever? | Things aren't white because they resemble each other, period; they are white because they resemble each other in a particular respect. But what is a "respect of resemblance" if not a universal? |
| Response to Problem 1 | We must distinguish determinables from determinates. | For each predicate, there are some paradigm instances. The predicate applies to a non-paradigm thing if and only if that thing resembles the paradigms at least as closely as they resemble one another. |
| critique of the response | Hard to see how this can avoid circularity: how are the paradigms selected in the first place? | |
| Problem 2 | the relation of resemblance itself seems to be a universal of relation. | |
| Response to Problem 2 | no, we can apply the resemblance theory to the relation of resemblance as well. Resemblance is a second-order resemblance between things. |
|
Last update: August 26, 2007 |