Hello everyone. I’m new here, so please call me out if this post is in the wrong place or there are already active threads on this subject.
I’ve been reading about Metaphysics in the Anglo-American tradition recently, and one thing that particularly struck me were the debates surrounding the nature and existence of universals. I guess I’m interested in how much interest there is in this problem among people studying philosophy, and the sort of solutions people are sympathetic towards.
Of course you might think that Metaphysics in the Anglo-American tradition (broadly speaking, specifying the fundamental entities from which reality is structured) is totally redundant, and ought to be replaced by the sort of alternatives one finds in the Continental tradition, e.g. Heideggerean Poetics, Deleuzian Difference or Badiou’s set theory. I don’t know enough about these ontologies or the questions they set out to properly address the question of whether they offer a more legitimate metaphysical project. Perhaps that could be a separate topic of discussion.
With that qualification aside, let me set out the problem of universals as I understand it, and some of the solutions I’ve come across. Incidentally, David Armstrong’s book called ‘Universals: An Opinionated Introduction’, is a pretty good starting place.
Anyway, I think a good way of getting to grips with the problem is through the token/type distinction. For example, how many words are there on this line:
THE THE
There is a sense in which the answers ‘one’ and ‘two’ are both correct. ‘Two’ is correct if you treat each word as an individual token. ‘One’ is correct if you treat both words as tokens of the same type.
The question is, what is going on when two tokens are of the same type? In the case of our two words, it seems as though there is a sense in which they are identical. But how are we to understand ‘identical’? If we mean it in the strict sense, then there must be some property that both tokens share. And what is a property? Well it seems to be a determinate repeatable, something which can be instantiated at more than one place at the same time.
Perhaps we do not want to allow strict identity between are tokens though, but rather suggest that they are identical in some looser sense. This would avoid talk about this funny thing we called a ‘property’, which could exist at different non-overlapping points in space at the same time.
David Armstrong puts it this way: some things are electrons; most things are not. What is it for something to be an electron, of the type electron?
I think the problem can be more clearly understood if we discuss it in terms of ‘sets’. We can think of an infinite number of sets, but some sets seem to be more natural than others. For example, the set of all electrons seems to have a more natural unity than the set [my computer, a dog, an electron, a tree]. Both are legitimate sets, but the members of one seem to have a lot more in common than the members of the other. So we can ask What distinguishes the sets of tokens that mark out a type from the sets that do not?
A realist about universals would say something like the following. All tokens of a particular type have something in common, a property. Indeed they all have the same property. This is possible because properties are not like particulars. They can exist at more than one place at the same time. They are abstract entities which can be multiply instantiated.
A nominalist would reject this entirely. She would say that the only constituents of the world are particulars. These particulars can be grouped together in certain ways. For example some nominalists say that particulars fall into certain natural sets. It is just a basic feature of the world that such sets exist. Other people say that particulars can resemble each other. Resemblance is a primitve relation that holds between objects and which allows them to be grouped into certain sets. Some other philosophers say that particulars do have properties, but that each property is unique to a certain particular. These properties are often called ‘tropes’. When two particulars are grouped into a particular natural set, it is because one of their tropes resembles each other. Doing things this way round is useful because it allows you to divide things up more finely - you do not just have particular obects, you can talk about particular properties too.
I’m not sure how clear all that is. There are problems with all the different accounts, which I’m happy to go into with people. Personally I’m at a loss. Every Nominalist theory seems inadequate to me, but I can’t get my head around the idea that there are abstract properties which can be instantiated at different non-overlapping points at the same time.