Wednesday, March 11, 2009

Ontological Cheating and Ockham's Razor

I’ve written a brief reply to Jonathan Tallant’s forthcoming Analysis paper, 'Ontological Cheats Might Just Prosper', that argues in favour of being the kind of ‘cheating’ presentist and actualist that simply takes truths concerning the past, or what could have been, to be ungrounded. Tallant argues that Ockham’s razor suggests we should be cheats, because if we can do without past or merely possible ontology, Ockham’s razor says we should. Don’t multiply entities beyond necessity, so since it’s possible not to believe in such things, you shouldn’t believe in them.

I argue that this has to be a bad understanding of Ockham’s razor: were it good, we should be Ontological Nihilists and believe that nothing exists. Since it’s possible to believe in nothing at all, believing in anything multiplies entities beyond what’s necessary, hence we shouldn’t believe in anything! Since we’re not Ontological Nihilists, we can’t be operating with this version of the razor.

Why aren’t we Ontological Nihilists? Because while it’s ontologically parsimonious, it’s ideologically extravagant. (See Jason’s paper.) Ockham’s razor has to allow that ontological parsimony needn’t be purchased if the cost is extravagant! But once we allow this, we’re just back to the old game of weighing up the admitted ontological benefits of cheating against what should be the admitted costs in its ideology and/or in its account of how truth depends on reality. The debate hasn’t progressed any.

I argue that every theory owes us an account of three things: what exists, what is true, and how truth links up to ontology. Ockham’s razor tells us, I suggest, that we shouldn’t accept a theory that postulates the existence of some things that don’t, according to its own view of how truth depends on ontology, do any work in accounting for what it itself says is true. That principle is going to tell us not to say, e.g., both that truths about the past are brute but yet there are nevertheless past entities. And that’s as it should be: that’s a bizarre combination of views to hold. But it’s never going to let us decide between two theories just by looking at their ontologies. And I think that’s as it should be: we have to look at the other two components as well, and see if the ontological advantage is being paid for at an appropriate price. And I can’t see any version of the razor that will mandate accepting the ‘cheating’ theories that won’t also mandate accepting Ontological Nihilism.

(I’ve also written a reply to a forthcoming paper by Stefano Predelli which argues against my view that there are no musical works. It’s here.)

3 comments:

David Gawthorne said...

Do you think that there might also be some use in differentiating a strong form of the razor (against mulitplying kinds of entities) from a weak form (multiplying instances)? After all, the non-presentist could argue that the past and future did and will contain all of the same kinds of things as the present.

Ross Cameron said...

I guess it depends on what we mean by 'kinds'. If there's a past, it contains things of the kind *dinosaur*, whereas if there's no past there are no things of that kind. On the other hand, if the kinds we're interested in are *sets* and *individuals* then, of course, admitting non-present entities doesn't increase the kinds of things we admit. I guess this is exactly parallel to the Melia/Divers debate as to whether Lewis' modal realism is qualitatively parsimonious. Melia says: no, it's not - Lewis is maximally unparsimonious, because he believes there are things of any possible kind. Divers says: he believes in sets and individuals, neither of which you get to disbelieve in just by being an actualist.

I must admit, I find it hard to judge such debates without some independent grip on what 'level' we're meant to count kinds. If I believed in ontological categories, I'd have such a grip. But I don't.

David Gawthorne said...

You make a good point. Thanks.