Here. It's really a first-rate discussion, and brings out aspects of the argument that were not clear (to me at least) in the earlier iterations, namely, that the key claim of Fodor and Piattelli-Palmarini (hereafter FPP) isn't that there isn't a difference between traits that are selected for and those that are mere free-riders (e.g., genetic hitch-hikers), but that the theory of natural selection can state no laws governing the distinction, i.e., no laws about what it is to "select for" traits of a particular type. An excellent place to start is Sober's summary of FPP's argument at roughly 6 to 11 minutes (a characterization Fodor accepts), which is much clearer than Fodor's own introductory remarks I thought. The crucial exchange then occurs betweeen roughly 14:30 and 29:00, in which Fodor, as far as I can see, concedes to Sober that actual theories in evolutionary biology do state counterfactual-supporting generalizations about which traits are selected for--but then he claims (this is 25-28 minutes in), bizarrely, that this is a merely "statistical" explanation not the kind of biological explanation that could support talk of "functions" and "teleology." Sober replies, unsurprisingly, that statistics are used to express probablistic causal relations, so, in essence, he doesn't know what Fodor is talking about. Those who have read Sober's work know that he is a lucid expositor of philosophical and scientific ideas; the same skill is in evidence in this discussion, and it shows up the peculiarity of Fodor's position very clearly.
I should note that it is really very disappointing, indeed a bit shocking, to see Fodor, who wants nothing to do with the intelligent design charlatans, adopt the language of the ID propoagandists, in talking about "Darwinists" and the views of "Darwinists," a point we have remarked on previously. Apart from that reprehensible and irresponsible lapse, it's a high-level and informative discussion, but it does vindicate one of Block and Kitcher's original worries, namely, that the whole argument of FPP suffers from ignorance of the actual theories put forth by evolutionary biologists. Fodor will have to turn in his Quinean union card, alas.
UPDATE: Jerry Fodor takes issue with my characterization of the exchange and the issue in dispute:
My claim is that the theory of natural selection does not, and cannot, provide a mechanism to do what Darwin says that ecological filtering does; i.e. choose from among a creature’s phenotypic traits the ones that are causes of fitness. I don’t propose to defend that claim here (sound though it quite certainly is). But I do want to point out that the brief discussion of statistical models in my blog conversation with Elliott Sober is entirely orthogonal to this issue. I took Elliott to be arguing, correctly in my view, that contemporary biology could perfectly well proceed with model construction even if what I’ve been saying about the theory of natural selection is true. What I took Elliott NOT to be saying is that the construction of a statistical model might provide a mechanism for the selection of causes of fitness. Statistical models don’t specify causal mechanisms; nor do I suppose for a minute that Elliott thinks they do.
Listeners can obviously decide for themselves which characterization is apt.
I invited Elliott Sober to comment on this, and he sent me the following:
Contrary to what Jerry says in the BloggingHeads exchange that he and I did together, his
book with Piattelli-Palmarini DOES assert that “natural selection cannot distinguish between coextensive phenotypic traits." This is a quotation from p. 154. There are other passages that say the same thing.When two traits are coextensive, I think there can be selection-for one of them but no selection-for the other. The book denies this.
The book also asserts that there can't be a theory of selection-for and this is something that Fodor also defends in our BloggingHeads exchange. I disagreed with him about this. I think there are numerous counterfactual supporting generalizations in evolutionary biology about selection-for. I described these in our conversation as causal models. I also gave an example. Jerry chose to call them "statistical." He never explained why they aren't causal. Jerry demands that a theory of adaptation must say what all adaptations have in common, "as such." Theorizing about sex ratio, optimal foraging, and cooperation (to name just three subjects that the theory of natural selection addresses) isn't like this. Fodor and Piattelli-Palmarini's book doesn't discuss theoretical modeling. For him the only thing there is in the "theory" of natural selection is the statement that if two traits differ in fitness, then, all else being equal, the fitter of the two will increase in frequency. This isn't a theory, but is, more or less, a definition.
A long time ago, Creationists argued that the theory of natural selection is a "tautology." Fodor and Piattelli-Palmarini are recycling this old saw, though they are not creationists.
Recent Comments