Dear Jay, more about natural kinds

Wed 14 Jun 2006 03:29 PM

Jay Odenbaugh wrote a provocative reply to my last post on realism. I was going to leave a short reply in the comments, but it ran long.

Reply to Jay

As I understand Jay's reply, LIZ (lizard front ends) might be a legitimate kind. Yet Jay wants to resist accepting any artibrary set as a kind; he gives the example of the members of the set {Jay's nose, Michael Keaton, Eiffel Tower}. Call this would-be-kind jayskind. He suggests two criteria that might be used to bar such a monster.

(1) Non-circularity

jayskind is not a kind, because the feature that distinguishes members of the set is just that something be a member of the given set.

Yet the items in the things might also be described as things featured in Jay's example; so he adds:

(2) Mind-independence

jayskind is not a kind, because the fact that the members figure in an example is just the result of his deciding that they should.

I do not think mind-independence can be settled so easily. Frog food is a natural kind because it is picked out from the environment by frogs. Although it is a frog-dependent kind, it counts as a natural kind in biological study of frogs. Thorough-going naturalism allows for a scientific study of philosophers. Although jayskind isn't an especially important kind, it is only Jay-dependent in the way that frog food is frog-dependent. Now that he has given it as an example, he cannot change what counts as a member of that set just by thinking a different thought about it.

Jay and I began discussing these issues at the last PSA, and we had the good fortune of having Dick Boyd join the conversation; the frog food example is Boyd's. If I recall Boyd's position correctly, he avoids counting collections like jayskind as a natural by only counting something as a real kind if it appears in some actual enquiry. No one has done a sociological study of the Magnus-Odenbaugh correspondence, and no one will. So jayskind is not a real kind. This embraces the first horn of the dilemma that I posed a while back: Kinds that now appear in our science were not real ones prior to the science in which they appeared. This is at least somewhat odd.

(3) Intension

I thought up this further criterion, which is akin to non-circularity. In Jay's example, there are three things that are members of this would-be kind. Because the extension of the set is stipulated, further members of the kind are precluded. Yet scientific kinds are not like this. They are open ended. Although we think that there are eight (or nine) planets in our solar system, one natural moon orbiting Earth, and zero golden mountains, these numbers are contingent. Nothing about the categories themselves determines the numbers to be 8, 1, or 0.

We can bar monsters like jayskind by insisting that there be more to natural kinds than simply a stipulated extension. Natural kinds have intensions, too. This requires (contra Quine) that there be intensions, but that is not asking too much of our philosophy of language.

Alas, this criterion will probably not be enough for Jay. Although jayskind does not count as a natural kind, things featured in Jay's example still does; and the latter is extensionally equivalent to the former.

'Natural Kind' as a relation

A different way of understanding Boyd's suggestion is this: Being a natural kind is not a second-order property that obtains of a predicate. Rather it is a relation between a predicate and a specified enquiry. "Frog food is a natural kind" must be elliptical for something like "Frog food is a natural kind for biology."

Although we might say that jayskind is a natural kind for the micro-sociology of the Magnus-Odenbaugh correspondence, that seems less brazen than saying that it is a natural kind simpliciter. It is certainly not a natural kind for physics, ecology, pharmacy, gastronomy, or any other interesting area of enquiry.

This would also allow us to distinguish between categories that are natural kinds for some actual enquiry and ones that are natural kinds for some possible enquiry. We would just need quantifiers and predicates that distinguish on kind of enquiry from another. We can say that electron was a natural kind for particle physics even before electrons were discovered; it was the existence of physics that changed, rather than the existence of the kind electron.

How about it, Jay?


from: jay

Thu 15 Jun 2006 10:21 PM


Oh dear. You raise very good questions about the criteria I previously suggested. So, let us continue.

(1) Consider things used in my example; Jay's nose, Michael Keaton, Eiffel Tower. Now PD correctly notes that we can devise a category which so describes these things, things featured in Jay's example. However, this suggests x is a thing in Jay's example just in case x is in things featured in Jay's example and that surely is circular or is dangerously close.

Incidentally, I think this also relates to PDs discussion of intensions vs. extensions. Natural kinds are specified via intensions (i.e., properties) which do not make ineliminable references to concrete particulars like Jay. Historical entities like species are concrete particulars and can be so specified by such intentions whereas kinds like gold cannot (maybe).

(2) Your second worry is a good one and I think this is one that realists need to take seriously. Let me put the worry in the following way. Surely, mind is a natural kind. However, it is not mind-independent. Hence, mind-independence is not necessary for something to be a kind. Here we need a more sophisticated account of mind-independence.

Consider a kind like gold. My representing the kind it being the object of some attitude of mine does not affect the category simply in virtue of my so representing it. Similarly, my representing the kind mind does not affect it in virtue of simply representing it. Of course, representations can affect stuff causally though behavior as Ian Hacking often argues but that is a different matter.

Now apply this to the example in (1). If I had chosen different objects for my example, then the category would have been different. This is an example of a mind-dependent category par excellance. I need to extrapolate a general characterization of mind-independence and that is something I haven't done but that is the direction I would go in.

PD have I lost my realist mind?



from: P.D.

Sun 18 Jun 2006 01:41 PM

Thanks for replying, Jay, and I apologize for putting you on the spot. I am not entirely convinced by your reply, although I am not sure if I have any compelling arguments that should convince you either.[blockquote](1) ... Historical entities like species are concrete particulars and can be so specified by such intentions whereas kinds like gold cannot (maybe).[/blockquote]There are a number of things packed into this move. First, you are presuming that species are individuals rather than kinds. I worry that this debate among philosophers of biology might play on a false dilemma. If we agreed that there exist both a set that contains all of the members of the species and a mereological fusion of which all the members are parts, it seems to me that species-as-individual speak could be translated without loss into species-as-kind speak (and vice versa). Second, a natural kind like frog food will contain reference to historical entities; viz. frogs. Yet frog food is not itself a concrete particular. You might deny that frog food is a natural kind, but I think that would be a mistake.[blockquote](2) ... If I had chosen different objects for my example, then the category would have been different.[/blockquote]There is a sense in which the category does not depend on your mental state, since it does not depend on what you now think. You can no more exclude Michael Keaton from the stipulated set by an act of will than you can exclude my wedding band from the set of platinum things. How is your initial stipulation any different than the formation of a specialized vocabulary? Whether the terms of a vocabulary pick out natural kinds has got to be a separate question from whether the meanings of those terms initially depended on some mental states.[blockquote]Have I lost my realist mind?[/blockquote]I suspect not, but I recommend a dose of pragmatism just to be safe.

from: Waldemar

Sat 21 Oct 2006 09:59 AM

zer gut