Bacchanalian realism

Tue 18 Apr 2006 11:05 AM

An old line of thought, resurfacing in recent rumination:

In asking whether categories are real, the problem is sometimes posed in this way: Is the world really objectively divided into real kinds of things, or is it just facts about us (our languages, our cultures, our interests, the ways our minds are set up) that make us tend to divide it in one way rather than in any other?*

The question presupposes a dichotomy between two positions:

(1) Facts about us do not influence which categories should appear in our science, because there is some relatively small number of real kinds. One job of science is to figure out what those kinds are. Call this position realist monism.

(2) The world could be divided up in many different ways. Our science should describe kinds that reflect what we want, because there is no privileged list of correct kinds. Call this position anti-realist pluralism.

From the way I labeled these positions, it should be clear that I do not think that these are the only two options. One can readily construct two others:

(3) Anti-realist monism might seem like a weird position, but it is not so far from the idealist tradition. According to Kant, the kinds in the empirical world do depend on how are minds are set up. Yet there is a relatively constrained set of objective kinds. Kant insists that these are empirically real but transcendentally ideal-- which might just be a way of saying that he is an anti-realist monist about kinds.

(4) Realist pluralism has been argued for by a number of recent philosophers, although not by that name. John Dupre has argued that there is no privileged biological taxonomy. The kind 'fish' in the old sense of finned sea life is as real and out in the world as the modern kind 'fish'. Biologists distinguish species in several different ways, all of them marking real distinctions in nature. Cooks and gardeners distinguish species in ways that may not be useful for biological purposes but which are no less real.

Philip Kitcher argues that the world contains too many real kinds to represent in our science. There really are more things on Heaven and Earth! So our interests lead us to select one set of categories rather than one of the infinitely many others that we might select. Yet the truths about these categories are not determined by what we want or by facts about us. Once we adopt one meaning of 'fish' rather than another, the world determines which sentences about fish are true.

Suppose accept some variety of realist pluralism. How pluralist should it be?

To take an example I've used in lecture: Let 'liz' mean the front end of a lizard, and let 'ard' mean the back end of a lizard. Surely there are lizards out in the world, and most of them have front ends and back ends. So there are front ends of lizards and back ends of lizards out in the world. But are the categories liz and ard real?

Dupre coined the phrase promiscuous realism. It suggests a pluralism that embraces all the kinds you could imagine, even stupid ones with no plausible usage-- kinds like liz and ard. In a conversation at the last PSA, however, he suggested that he had nothing quite so extreme in mind. He only meant to include different biological kinds, along with non-scientific kinds used by gardeners, cooks, and so on. All of those kinds are practically useful and can be used to formulate regularities that we care about.

Kitcher seems committed to the reality of liz. He proposes an object, the career of which consists of the last decade of Bertrand Russell and the first decade of his dog Bertie. There is no decisive objection against the reality of this Bert-cum-Bert, but it does not count as a creature in our taxonomy. Moreover, there is no reason to adopt a taxonomy that would distinguish this gerrymandered monster. That is just about the categories we have decided to recognize however, not about whether there really is such a thing in the world.

Kitcher does not give a name to this position, but it is more promiscuous than Dupre's promiscuous realism. This is unfortunate, because there is no natural term for something that is more than promiscuous. Orgiastic realism? Many other jokes suggest themselves, but I leave them as an exercise for the reader.

* This formulation is a direct quote from Robert M. Martin's chatty textbook Scientific Thinking, p. 215.


from: Matt Brown

Sat 29 Apr 2006 09:53 PM

One might be able to make a principled, pragmatist distinction, here, to support Dupre's limits. "Real" kinds are the ones that result from any kind of actual purpose that we have. Or, in the Deweyan terminology, real kinds are the ones that we get as a product of actual inquiries. Useless dreams like Kitcher's may be hypothetical but not real kinds. This still allows for some promiscuity (because biologists, cooks, common sense, etc. may have different purposes/problems and come to different conclusions) without allowing in all the useless stuff.

At least, the line seems plausible, but maybe just because I'm in the midst of reading Dewey's Logic with Paul and some others at the moment.