Friday, November 15, 2013

Ontology and the necessity for universals

To: Stuart B.
From: Geoffrey Klempner
Subject: Ontology and the necessity for universals
Date: 23rd February 2011 11:37

Dear Stuart,

Thank you for your email of 14 February with your essay for the University of London BA Metaphysics module, in response to the question, 'Are universals necessary elements of any adequate ontology?'

This is an excellent essay, which lays out clearly the strategy which you have adopted in answering this question, and which delivers that answer, with strong arguments to back it up. And yet, somehow, I find that I am still left somewhat confused by this question.

Right at the start you distinguish between two ways of taking the question: Whether universals are necessary, albeit possibly derivative elements of any adequate ontology; OR whether universals are both necessary and fundamental elements of any adequate ontology.

To cut a long story short: Realists believe that universals are necessary and fundamental; nominalists believe that universals are dispensable therefore not necessary; while conceptualists hold the view that universals are necessary but not fundamental.

(It turns out, however, that on the version of conceptualism you wish to defend particulars are not fundamental either. The only fundamental entity is 'the world'! I have already offered comments on this position.)

My difficulty is this. Maybe a bit of personal history would be relevant. In my second year as an undergraduate I read Dummett's book 'Frege: Philosophy of Language'. I didn't just read it, I scoured it from cover to cover filling the margins with copious notes. We'd covered the Presocratics in our first year, looked at the '3rd Man' argument in the Parmenides. The 'problem of universals' so-called was something I might have come across in an old text book but no-one ever talked about it. So when Dummett wrote that the whole questions of universals was old hat, I didn't need much convincing. My outlook (from lectures as well not just from reading) was thoroughly Fregean.

So what IS the fuss about? In the Phaedrus, Plato gives an analogy which is so powerful and appealing that philosophers still use it today: 'Carving the meat at the joints.' We want our language, our concepts, to carve the world at its joints. This implies that the world has 'joints', it is not just an amorphous lump (corned beef from a tin). What is your view? Human beings have needs and interests, which can change from one context to another, and the concepts which we choose to employ are intended to enable us to pursue those interests successfully, given the nature of the world. But, obviously, this implies that the world is some 'nature' rather than another. It has 'joints', an articulation which, admittedly, we only discover indirectly through trial and error, in the course of praxis. The joints can't be seen, they are not recollected by the soul or any such nonsense. But they must 'exist' in some sense, in order to impact on our praxis.

To me, the philosopher whose views come closest to this position is Quine. See his essay 'Epistemology Naturalized' (in 'Ontological Relativity and Other Essays'). This is especially relevant given your interest in evolution. The world impacts on us and the classifications we use because it made us. All there is to say about the way human beings carve the world up in language can be said using the sciences of physics, chemistry and biology and the apparatus of first-order predicate logic. There is no problem of universals. They don't exist, not because nominalists are right but because the whole dispute has been superseded. Nominalism, like conceptualism or realism, is a position with respect to 'universals' and no position needs to be taken because there isn't a meaningful question to answer.

You state that for nominalists, what takes the place of universals are sets. But what is a set? Where do sets exist? Consider the set of all predominantly golden Calico cats which are currently on some floor somewhere (your example). It seems to me absurdly unlikely that such a set 'exists'. Leaving aside the question of simultaneity, consider the aspect of vagueness in 'predominantly', or 'golden', or 'floor'. There is no such set. But why was a set needed in the first place? You suggest the answer: that the naive model of meaning requires a designatum for every designator. Wittgenstein has a nice simile to counter this: in the cab of a locomotive there are lots of levers, but the levers don't all work in the same way. One lever might be an on-off switch, another you have to push and pull continuously and so on.

This variety is regimented (Quine's term) to some extent by the use of first-order predicate logic. Tarski's definition of truth lays down clear and very simple rules for the different elements of the calculus: names, n-adic predicates, variables, propositional connectives, etc.

Before I finish, I would like to mention one thing in particular that you said which is so true: 'While attributing to others the archetypic positions, their own positions appear to blur the distinctions.' In other words, philosophers love to attack straw men. If we were more keenly aware of the extent of this lazy practice, much academic philosophizing would be swept away.

All the best,

Geoffrey