by Matthew Cobb
In the letters page of this week’s issue of the Times Literary Supplement, Fodor and Piatelli-Palmarini reply to the trashing of their book by Samir Okasha. Surprise, surprise, Okasha didn’t understand what they were saying – “an egregious misreading”. For most of us, if someone doesn’t understand the point we’re making, it’s because we haven’t expressed ourselves in the right way… Some philosophers seem to think that any misunderstandings are inevitably the fault of the reader. Who finds their riposte convincing? I hope Okasha comes out fighting.
Sir, – Samir Okasha’s review of What Darwin Got Wrong (March 26) contains a number of serious criticisms of a line of argument the conclusion of which is that there is something radically wrong with Darwin’s Theory of Natural Selection (TNS). Since we do think that there is something radically wrong with TNS, this would worry us a lot if the argument that Okasha deconstructs were even remotely like the one to which our book is committed. But it’s not. Okasha’s review is a really egregious misreading of the book; we don’t hold (and didn’t publish) the views that Okasha says we do. In fact, we explicitly don’t hold these views, and we devote a lot of the book to explaining why no one should. We know from experience that reviewing is hard work and that it conduces to fast reading. But still.
Here is the most flagrant example: “. . . Fodor and Piattelli-Palmarini insist that . . . there is no basis on which to distinguish the selected-for traits from the free-riders. This distinction could only be drawn, they argue, by invoking an intelligent designer . . .”. If words were the kinds of things that can be false, every one of those would qualify. In particular, we think (as does Okasha) that the selected-for traits are the ones that causally contribute to fitness. We take this to be common ground for everybody involved in the present conversation, Darwin included. For present purposes, it’s OK with us if you take that to be true by definition. (This does not, of course, commit us to accepting that any traits are selected-for; only that, if any are, then they are causes, not just correlates, of fitness.) Nor do we think that the distinction between causes and correlates can only be made by invoking an intelligent designer. To the contrary, we devote a whole chapter (Chapter Seven) to discussing some of the ways in which causes and correlates are routinely deconfounded, both in science and in everyday life. Running control experiments, formal or informal, is the standard technique; and there are many, many others. But (so we argue) only things with minds can run experiments; and since it is also common ground that Natural Selection hasn’t got a mind, distinguishing causes from confounds by running experiments isn’t among its options. Likewise for all the other ways of distinguishing among confounded variables that we could think of. If we’re right about that, then the conclusion isn’t that there is an intelligent designer; it’s that there is something wrong with TNS. So, then, our view is not that it is impossible to deconfound causes of fitness from freeriders. Still less is it that there is no such distinction. What we do think (and what we do think our book shows) is that Darwin’s theory can’t specify a mechanism by which selection could reliably distinguish causes of fitness from correlates of causes of fitness. This is not, to repeat, because there is no such distinction; it’s because TNS recognizes only exogenous variables as selectors, and the only (relevant) fact to which such variables are sensitive, according to TNS, is the strength of the correlations between phenotypic changes and changes of fitness. And, of course, correlation doesn’t imply causation. It patently doesn’t in the kind of cases we discuss, where phenotypic traits are linked, so that the correlation with fitness is identical for both of the candidate causes.
Our difficulty with Darwin is very like our difficulty with our stockbroker. He says the way to succeed on the market is to buy low and sell high, and we believe him. But since he won’t tell us how to buy low and sell high, his advice does us no good. Likewise, Darwin thinks that the traits that are selected-for are the ones that cause fitness; but he doesn’t say how the kinds of variables that his theory envisages as selectors could interact with phenotypes in ways that distinguish causes of fitness from their confounds. This problem can’t be solved by just stipulating that the traits that are selected for are the fitness-enhancing traits; that, as one said in the 1960s, isn’t the solution; it’s the problem.
We have other complaints as well. Three examples: we do not hold a “covering law” view of scientific theories; we are explicit that we don’t. What we hold is that if there were laws of selection, that would solve the problem of reconciling TNS with the intensionality of “select-for”. But we don’t think there are such laws in biology, and Okasha doesn’t either. Also: we disapprove of Okasha’s appealing to the “paradigm” explanatory power of mathematical models of natural selection to rebut our objections to TNS. Models don’t even purport to reveal the mechanisms that underlie the phenomena they’re models of; and our claim is that no mechanism could do what TNS says natural selection does. Also: Okasha summarizes several recent discoveries in biology that our book recounts. He sets them aside saying (correctly) that “they simply concern aspects of biology about which traditional neo-Darwinism didn’t have much to say”. But our point about these discoveries is not that neo-Darwinists ignore them; it’s the marginalization of TNS that they imply; it seems the action is mostly in a different part of town.
There is, however, one place where we admit to a fair cop. Samir Okasha chides us for not telling our readers about the distinction between “random” variation and “undirected” variation. Guilty as charged. But that isn’t because, as Okasha graciously suggests, we don’t understand the difference; it’s because we try not to make readers attend to distinctions which, though perfectly valid, aren’t germane to the topic being discussed. Insisting that readers should is just the sort of thing that gives pedantry a bad name.
Department of Philosophy, Rutgers University, New Jersey 08903.
Cognitive Science Program, University of Arizona, Arizona 85701.