Sunday, April 06, 2008

Quine and ontology

In his "Things and their place in theories" Quine presents his views on ontology. In particular, he presents three sorts of ontologies one could adopt. The first is the physical object ontology which, unsurprisingly, includes physical objects amongst its things. The second is the space-time ontology which replaces physical objects with the regions of space-time points that the objects occupied. This is broadened then to include empty regions as well. The third sort of ontology replaces space-time points with quadruples of real numbers, and so gets by with just set theory. Sets are all you need to make sense of science on this view.

The first two options seem reasonably attractive. What strikes me as odd about the third option, Quine's favored one, is that it doesn't seem to mesh with what he says about perception. The essay opens by saying, "Our talk of external things, is just a conceptual apparatus that helps us to foresee and control the triggering of our sensory receptors in the light of previous triggering of our sensory receptors." The sensory stimulations constitute observations and perceptions. I'm going to set aside how Quine gets from the former to the latter. It makes sense to say that we observe physical objects and, possibly, that we observe space-time points or regions of them. We don't observe sets or quadruples of reals. Thus, we shouldn't adopt this third option for ontology, since it doesn't help us explain our sensory stimuli and observations.

My uneasiness can probably be alleviated with a more thorough going Quinean response. In the above, I said that "we don't observe sets" understanding "we" in the physical object sense and "observe" in the roughly standard sense. Replacing objects with quadruples of reals (or sets of them) will require likewise adjusting how we understand "observe". Proxy functions could be inserted appropriately. It will have to be some set-theoretic relation containing the observers and the observed quadruples. The observers, us, will likewise have to be understood in terms of sets of quadruples of reals, with certain quadruples appearing in the sets that represent our surface irritations. This more consistently Quinean approach responds to my earlier worry, but there is a lingering one, although it seems a bit lame. The worry is that I am not a set of quadruples of reals, although I could be represented as one. If we are worried about what there is, why should we concern ourselves with representations of things, rather than the things themselves? It happens that in this case the representations are full-fledged objects in their own rights. I'm not sure how moved I am by this last consideration. In writing it up, I started to think I was missing something in Quine's view.

1 comment:

Greg said...

I dunno -- there is a substantive and often-occurring issue here. Philosophers (and mathematicians, too) often want to analyze one sort of talk into another sort of talk (delta-epsilon definitions of continuity, definitions of arithmetical notions in set-theory, e.g.).

I think there is a real question about what these (should be taken to) show. Is it ever OK? If so, when? Last week in my philosophy of logic class, we read Read's chapter on modality. He dismissed the Ersatzers with: 'Yes, perhaps all the interesting/ relevant aspects of possible worlds can be represented by sets of maximal consistent sentences (or whatever your favored replacement is), but we all know that sets of sentences aren't possible worlds.' And he lumped set-theoretic definitions of number into the same bag. I found Read's reply frustrating, since the whole point of introducing some surrogate for possible worlds is that you get rid of possible worlds and all their spookiness.

But the general question is: when is any sort of replacement OK -- and then, does Quine's weird replacement via quadruples meet that criterion? And those seem like legitimate (and important and interesting) questions.