ON FODOR ON DARWINIAN EVOLUTION

shorter commentary on Fodor's Summary of "Against Darwinism"
(longer critique: "On Fodor on Darwin on Evolution")

Stevan Harnad
Canada Research Chair in Cognitive Sciences
Universite du Quebec a Montreal
Montreal, Quebec,  Canada  H3C 3P8
http://www.crsc.uqam.ca/
&
Department of Electronics and Computer Science
University of Southampton
Highfield, Southampton, United Kingdom SO17 1BJ
http://www.ecs.soton.ac.uk/~harnad/


Jerry Fodor (JF):  "A theory of adaptation is supposed to be a theory of the fixation of phenotypic traits." 


Actually, it is a theory of the redistribution of genotypes (not phenotypes) as a consequence of the effect of the environment on the survival/reproduction of the phenotypes of the organisms bearing them.

So it is not phenotype fixation, but just genotype proportion change (following random mutation, recombination, and reproductive success).

JF: "The Darwinian principle is that phenotypic traits are selected for their (presumably causal) connection with fitness."


"Fitness" is also a secondary and somewhat misleading property. The right property is just the distribution of genotypes from generation to generation, and it couldn't be simpler:

The distribution of genotypes from generation N to generation N+1 will tend to change in favor of the genotypes that generate the phenotypes that are more successful in surviving and reproducing (if there is any systematic difference in success).

No need to mention "fitness." No need to mention "fixation." Just a fairly mechanical (and quasi-tautological, once we realize that there are heritable traits at all) regularity relating genotype distribution and reproductive success across generations.

JF: "Prima facie, this principle is falsified by phenotypic traits that go to fixation but are not connected to fitness (drift and the like)"


Why is it falsified, rather than just extended? Some traits are inherited because they have systematically enhanced reproductive success; other traits are inherited because of random variation and because they have not systematically diminished reproductive success.  

JF: "Some such cases can be handled by `friendly amendments' to the Darwinian principle. But there is a serious problem about `free riding': cases where a trait that is NOT connected to fitness goes to fixation because it is (locally or otherwise) linked to a trait that IS connected to fitness. (This is the arch/spandrel situation)."


Neutral correlated traits are not at all a problem for the simple heritability process just described (heritable traits that enhance survival and reproduction are more likely to be transmitted than traits that diminish survival and reproduction; and linked, neutral fellow-travelers can hang in there until and unless they happen to diminish survival and reproduction).  

JF: "The usual way to understand cases where a process distinguishes between coextensive traits is to claim that the process is INTENSIONAL."


In experimental genetics, the way to find out which genetically inherited traits do and don't make a causal contribution to reproductive success and which ones are merely linked fellow-travelers (when it's not obvious) is to try to unlink them and test the effects on reproductive success.

(I am using "reproductive success" as shorthand for "adaptive success," because it is more transparent that we are talking about the success of heritable traits in managing to get inherited!)

The experimental test is much the same as it is with Skinner's pigeons, if they are trained to respond to a green triangle and not to a non-green, non-triangle: Just test to see whether they respond to a green non-triangle or a non-green triangle. (Fellow-travelers are no problem for either a Skinnerian or a Darwinian -- though Skinnerians have plenty of other problems!)

JF: "This is what Darwin does; he introduces the intensional context `select for...'. By stipulation, one but not the other of two coextensive traits can be selected for in an evolutionary process."


I think Darwin doesn't mean anything nearly so complicated with "select for": He just means whatever enhances reproductive success (in a given environment).

Heritable traits can be coupled. A breeder can select deliberately for one of them (and ignore the fellow-traveler). Reproductive success can also crown one of them (and not handicap the fellow-traveler).

JF: "In practice, this comes down to claiming either that there are laws of selection, or that selection involves intervention by a mind (an agent)."


There's no need to speak of anything as fancy as a "law." It is enough to point out the simple (but true, and richly predictive and productive) regularity (eventually discovered to be embodied in the genome) that there is variance in heritable traits and that those (heritable) traits that enhance reproductive success tend to increase their frequency in the next generation, relative to (heritable) traits that diminish it. That's really all there is to it.

And there is no need at all to speak of (or to worry about somehow being committed to) mental intervention; the point of the breeder analogy was precisely to show that mental intervention is not necessary to explain how heritable traits evolve across the generations.

JF: "Both these possibilities appear to be ruled out as explications of `select for...'. The first because there are no laws of selection, the second because there is no Tooth Fairy."


The simple, true regularity, the only one at issue, stands firm: There are heritable traits. (Their mechanism has even been discovered, since Darwin [via Mendel to Watson/Crick]: genes, mutations, and recombinatory DNA.) There is variation (both random mutation and recombination.) And reproductive success in a given environment does influence the relative frequency of genotypes in succeeding generations.

What more can one ask? And what is missing?

JF: "This appears to be a dilemma; one from which, as far as I can tell. Darwin has no exit."


I don't think it's any sort of dilemma at all. It seems to be a consequence of taking a few of Darwin's metaphors ("selection") too literally.

JF: "Historical addendum: I suspect that Darwin got into this mess because he assumed that natural selection can be modeled by artificial selection: Start with breeding, take the breeder away, and you have natural selection. If so, then he was the victim of a fallacy of subtraction. That could happen to anybody."


There may be a general fallacy of subtraction, but certainly Darwin did not fall into it. What he said was perfectly correct: With animal breeding, the breeder decides consciously what traits will have "reproductive success." With evolution, no conscious decision is needed: reproductive success in a given environment determines the genotypic distribution across generations.

(In my critique -- On Fodor on Darwin on Evolution -- I added something that I suspect Jerry found especially wrong-headed: That in fact mindful selection by human breeders is merely a special case of ordinary Darwinian evolution, in which it is human breeders' tastes that determine (say) dogs' survival/reproductive success, rather than, say, predators' tastes...)

JF: "I don't mind people ignoring this argument; that's gone on for some 150 years. Nor do I mind being preached to about strong adaptation, weak adaption, and so on. (One very distinguished Darwinist has suggested to me that the Theory of Evolution isn't a theory at all; it is, he said, a `tool box'.) But enough is enough. Would some body very kindly reply to the argument? Or would everybody very kindly leave me alone?"


<>One reply above. Please don't leave it alone!