Fodor on Natural Selection

I am back from Philly. I had a veggie-cheese steak and saw the Liberty Bell.

As for the SPP, it seemed to be doing quite well. I missed the first day of the conference, which I was bummed about. I wanted to see Devitt’s talk on ‘The “Linguistic Conception” of Grammars’. From what I know of Devitt’s recent work I would guess that the talk was an attack on the assumption that the rules of grammar are explicitly or implicitly represented in the mind of a typical language user. I am sure that would have been interesting to see.

I did catch Jerry Fodor’skeynote speech on Friday. He was there giving a version of a talk that I have heard a couple of other times at various conferences around the New York area. The basic theme of his latest work has been an attack on standard formulations of adaptationism of the Darwinian variety. This theory aims to explain how it comes to be the case that animals come to be adapted to their environment. This is a question that has puzzled biologist for some time. The Darwinian answer is in terms of natural selection. The basic idea is as follows. The traits that are useful for the species (cashed out in terms of reproductive success) are selected for and the traits that are detrimental to reproductive success are selected against.

Fodor’s argument againts this well established view proceeds in a couple of steps. The first step is to look at intentional actions and point out a certain kind of problem. Take the frog. It snaps at flies in order to survive. Now consider the intentional state of the frog itself. Is it intending to snap at flies? Or is it intending to snap at ‘ambient black dots’? These are twp different properties. It happens to be the case that whichever one the frog is doing will help it in the environment it finds itself in, so in a sense natural selection cannot distinguish between these two very different properties. Since it can’t distinguish between them it cannot select between them.

This problem generalizes, according to Fodor. Consider the Polar bear. The standard story that is supposed to explain why the polar bear is white involves selection for white polar bears. But ‘selection for’ as an intensional context. What this means is that we do not get truth preservation with substitution of co-referring terms. So, in the case of belief attribution which are the exemplars of what happens in an intensional context. So, I can believe that 50 cent is a great rapper without believeing that Curtis JAckson is an excellent rapper even though 50 cent and Curtis Jackson are the same person. Now consider ‘selection for’. Say that white is my favorite color. Being white and being my favorite color are different properties and selection for one is not selection for the other. The problem then becomes of explaining how we can say that white polar bears were selected for, as opposed to polar bears that are my favorite color. There are an indefinate number of properties which are distinct from yet coextensive with being white. If you don’t like that of being my favorite color Fodor’s example is ‘being the same color as the environment’. Which of these properties were selected for? To answer that question you would need a selector, but the Darwinist is not allowed to this claim. So then, the traditional adaptionist theory fails to do what it promises to do.

As usual, the scientist in teh room were very upset with Fodor, and it was quite entertaining. Lot’s of people have been upset with Fodor recently. Dennett says that Fodor’s argument is good for two things. Providing a reductio of the Fodor-inspired Representational theory of Mind in cognitive science. And demonstrating how NOT to do philosophy. I worry that this is half right. It seems to me that if one does approach the problem in the way that Fodor does, which whether or not correct is one of the most widely held views about the mind in current cog sci, then you do have this problem.

How much of a problem it turns out to be remains to be seen. Fodor thinks that it is no big deal. He certainly doesn’t think that this should provide comfort to the intelligent design folks.  


15 thoughts on “Fodor on Natural Selection

  1. Not sure what frodor is trying to prove unless its just that people should rephrase either natural selection or the representational theory of mind to more accurately describe reality.

    To me (and dennett I suppose) natural selection is more anchored in society and more proven so it would be RTM that would give.

  2. Hey GNZ,

    He doesn’t think there is anyway to rephrase the theory of natural selection to overcome this problem (or RTM).

    You are right that Dennett rejects basic tennets of RTM….that exactly what I am worried about! If Fodor is right about this, then Dennett’s view looks more appealing…but his view is crazy!

  3. If he doesn’t think it can be rephrased what does he think?
    Are we disputing the facts of what happened in evolution? or the appropriate methodology to investigate it?

    As to intentionality… Not sure if it is relevant but what if we wonder not about the frog but instead about the fly. Something saw the fly used a trait of the fly to determine if it intended to eat it – and then removed it from the gene pool… Seems the majority of natural selection in higher animals is of that sort.

  4. His point is that once you have the very detailed natural histories then there is no more work for the theory of natural selection to do. So he is disputing the very theory itself, though not the data that the theory is trying to explain. Certainly animals that are around today are the result of animals that survived and passed down their genetic material. He doesn’t deny that. What he denies is that saying that a trait was selected for doesn’t explain anything because it doesn’t tell us how it happens to be the trait specified one way as opposed to another.

    You proposed plan to focus on the fly doesn’t help either. You say “something saw the fly and used a trait of the fly” well, which trait did the thing in question see? The problem immediately arises again when you try to specify the trait that is ‘noticed’…

  5. I’m not sure I see the problem. In regards to the fly example, it’s an empirical question. Was natural selection “content” with the rule “snap at black dots”? Or did selection pressure require a more robust “definition” of things that should be snapped at? It would be simple enough to test. Get some frogs and present them with various dots, and see what it takes to get them to snap at the dots. I’m sure it would be an interesting thing to know, but I don’t see how this is in any way a fundamental question.

    The same is true in regards to the polar bear. To hunt seals, you are better off if you remain unnoticed as long as possible. Thus, bears that blended in better with the surroundings tended to be more successful. In the environment of polar bears, this means that you are better off being white. The two “competing” explanations are simply alternative ways of presenting the same idea. “White fur” and “fur that matched the environment” are the same thing in this case. They are both imprecise but useful ways of stating a particular gene-frequency pattern.

  6. > once you have the very detailed natural histories then there is no more work for the theory of natural selection to do

    if you mean once you have all the empirical facts there is no work to be done by a theory of those facts then – isn’t that true about every theory? surely he doesn’t mean that?

    > which trait did the thing in question see?

    I can see the issue there, but I don’t see the better theory – and in he absence of that this one seems to be OK.

  7. Hi Peter Borah,

    I take it that Fodor’s argument is that there is no way to empirically distinguish the two claims in the envirnment that the frog actually finds itself in. The theory of natural selection is, after all, supposed to be a theory for how frogs, in the habitat they find themselves in, come to acquire the trait ‘snapping at flies’. The way that the frog accomplishes this is by having mental states that represent the flies. So, Fodor says, the frog has to token a mental state with a determinent content. So, what is the content of the frog’s mentalstate? Is it ‘snap at flies’ or ‘snap at ambient black dots’? These are distinct contents and having one versus the other will make a differene to the forg’s behavior. So which was selected for? In the actual envirnonment that the frog finds itself in these two contents will have indistinguishable success rates. That is, whether the frog does one or the other it will be equally succesful at surviving and reproducing so in effect natural selection cannot distinuish between these two properties. Since there is no way for natural selection to distinguish between these two mental states there is no way that it could have selected one over the other. Fodor thinks that if there were a designer then the answer would be easy, but since we are naturalists and do not want to appeal to the intentional states of God we are stuck with this problem. The problem is magnifies when you realize that there are an indefinite number of properties that are co-extensive with the flies in the frogs envirnment. Suppose that some frogs really like the color they experience when they see what causes what we would call black. Then frogs who token intententions to snap at things that are their favorite color will have an advantage…etc…etc…this seems like a pretty fundamental problem to me…

    The same is true of the polar bear, according to Fodor…you go on to object that ‘being white’ and ‘being the same color as the environment’ are just different ways of specifying “a particular gene-frequency pattern”; but this is precisely the problem! ‘Selected for x’ is an intensional context and so co-refferring terms cannot be substitued with the preservation of truth. So we cannot say which of the very different properties were selected. All we can say, as you point out, is that there is an animal with a certain genetic ‘spectrum’. We can tell stories about individual animals that survived and reproduced and those that didn’t. But at no point do we need to, or get to according to Fodor, appeal to anything like a theory of natural selection.


    No that’s not true of everything! Do you think it is?

  8. I’m probably missing something important, but to me it seems like your entire first paragraph presents a truism. Assuming that I agree with everything in the first paragraph (which I mostly do), can you explain to me why we should regard that as a “problem”, and not just an explanation of how natural selection works?

    As for your second paragraph, I think you’re taking natural selection to be far more anthropomorphic than it actually is. The basic concept is almost a tautology, something along the lines of “genetic patterns that successfully replicate themselves (and provide for future replication) will end up with more descendant copies”. (For simplicity’s sake, I’ll refer to genes as the unit of selection, though I recognize there’s a debate on that subject.) The reason that this is important is that it gives us a framework in which we can begin to understand– in broad, general terms– why particular genes spread.

    A truly rigorous explanation of polar bear fur would have to detail the individual genes in thousands or millions of bears and which ones were replicated at which time. But it’s far more interesting and useful, as a first approximation, to say “genes that caused their bear to blend in were more likely to replicate before the bear died of starvation”. In the same way, a truly rigorous explanation of why the Mississippi flows south instead of north would have to include calculations about trillions of interactions among billions of molecules. But it’s far more interesting and useful, as a first approximation, to say “rivers flow downhill, because of gravity”.

    It’s possible that I’m entirely missing the point, as my grasp on intentionality is rather weak. (I just graduated high school.) But assuming I understand your point, I don’t think it’s nearly as big a problem as you make it out to be.

  9. Any (good) theory (or natural law if you like) is a generalization designed to approximate reality. If you knew everything then the knowledge of the generalizations would be superfluous (and probably wrong in some sense).
    Of course we don’t and we don’t expect to.

  10. Hi Peter Borah,

    You may not be seeing the problem because you already have an intuitive understanding of natural selection that lines up with Fodor’s account. So, he pretty much agrees with you that all you get from theories of natural selection is something like a tautology. But that isn’t informative. To be told that the reason polor bears are white is because being successful is a good way of being succesful…but to really see what Fodor is up to you would have to deal with his theory of what he calls ‘covering laws’ in special sciences (if you are interested look up his paper entitled ‘The Special Sciences’)…here is a link to a Londer Book Review Fodor wrote where he sets this problem out…read it and let me know what yo uthink…


    If you knew everything you would know the correct theory….but simply knowing all the data wouldn’t determine which theory was correct…or so some people seem to think…

  11. I’m going to try not to make this post incredibly long, but we’ll have to see whether I can pull that off…

    First off, your post. You claim that natural selection, since it’s based on a tautology, can’t be informative. I can’t see how that’s true. We don’t say that polar bears are white because “being successful is a good way of being succesful”. We say that they are white because the genes that coded for white fur resided (obviously) in bears that were white, and thus were more likely to be passed onto the next generation. This is interesting for many reasons. For one, it explains how the amazing, interlocking variety of life could arise by natural processes. For another, it allows us to make predictions about the features of animals we have yet to observe. The most famous example of an evolutionary prediction comes from Darwin himself, who was able to predict the existence of Xanthopan morgani (a moth with an especially large proboscis) decades before it was actually discovered.

    Second, “The Special Sciences”. I can’t see how this is related. His paper offers a pretty convincing account of why strict reductionism is wrong, but I can’t make the connection to his critique of natural selection. I did a google search for “covering laws”, since they aren’t mentioned in the paper, and couldn’t find anything relevant there either.

    Finally, the article you linked. This is the most difficult to respond to, partly because it’s long, but partly because he seems to be arguing against a version of natural selection that’s very nearly a strawman. His argument rests on a lot of assumptions that simply aren’t true. I’m having difficulty classifying them, so I’ll just take some of his own examples as demonstrations of the sorts of places he goes wrong.

    Lets start with spandrels. He claims that Darwinism can’t distinguish between two accounts of a hypothetical cathedral evolution, one in which spandrels are selected for, and one in which arches are selected for. There are in fact two very simple ways out of the problem. First, you could figure out which one positively affects survival rates, by studying how spandrels and arches interact with the environment, or by testing (either physically or as a thought experiment) what the cathedral would be like with arches but not spandrels, or vice versa.

    I admit, however, that this would be somewhat subjective and hard to pin down. I don’t think there are any insurmountable difficulties with that approach, but maybe we want something more concrete. So, an even better way out of the problem is to stop trying to atomize the concept of “traits”. It’s not true that either either spandrels or arches was selected for, and then the other one popped into being when the selection was made. Rather, the mutation that created arches necessarily created spandrels at exactly the same time. So, natural selection had to select between “both spandrels and arches”, or “neither spandrels nor arches”. The trait that was selected for was neither spandrels nor arches, but “spandrels-and-arches”.

    Notice that this approach neatly covers his polar bear example as well. The trait selected for was not “whiteness”, and it was not “being the same color as the environment”. It was “whiteness-when-the-environment-is-also-white”. Saying that natural selection selected for “whiteness” or “being the same color as the environment” is fine, as long as you recognize that this is a simplification. The true story has to include all of the relevant factors.

    After he discusses spandrels and polar bears, he moves on to “the empirical issue”. This appears to be nothing more than a survey of the various factors that affect evolution other than pure adaptionism. He fails, however, to explain why the existence of other factors rules out natural selection. It’s as if an attempted argument against gravity pointed out that gravity couldn’t explain a paperclip rising to meet a magnet. Modern evolutionary biology accepts all of the factors he mentions without jettisoning adaptionism. He also fails to offer any alternative explanation for the apparent “designedness” of nature, the main motivation for the introduction of natural selection in the first place.

    He ends the article by making fun of some particularly silly examples of evolutionary psychology run amok. I think this may be his actual target. If you recast his arguments as against certain abuses of natural selection, they start to make more sense. For instance, the sillier parts of evolutionary psychology do tend to ignore other possible explanations, and they do tend to play fast and loose with what counts as a “trait”. I think his arguments are useful in helping to refine our understanding of natural selection, and in giving some perspective that certain evolutionary psychologists sorely need. But as arguments against natural selection itself, I don’t think they work.

  12. If the data (and here I mean ALL the data) doesn’t distinguish between two theories then are they not both correct and are effectively the same theory?

    OK here are some thoughts directly on Frodors article

    1) He seems to be trying to shed some doubt on the argument that we might be poorly able to deal with situations that our ancestors never faced – but is this really in doubt?

    A specific story might be but only because there are so many possible specific stories. the story could be true and add value without being the total answer (and it cant be the total answer, but if the facts are true it must be part of the answer)

    2) natural selection selects for genes not traits. So any talk of traits is an approximation that biologists are well aware of. So in the short term you can have tameness associated with curly tails IF curly tails are not a major disadvantage. In the long run NS will find other solutions (there is more than one way to be tame in fact there are probably many ways already in the genome and curly tails was one of the best) if there is even a fairly minor advantage so in the wider scheme of things one might be excused for glossing over it.

    “Curly tails aren’t fitness enhancing, they just happen to be linked to tameness, so selection for the second willy-nilly selects the first.This case is much like that of spandrels, but much worse from an adaptationist’s point of view. ”

    What we seem to have here is the fact that if there is a major selection for tameness you can have a selection for curly tails but this also comes with a selection for solutions without curly tails depending on how disadvantageous curly tails are.

    3) “How could a studied decision to breed for one trait or another be ‘the very same thing’ as the adventitious culling of a population”

    To a biologist the sum of the intents of a million breeders (which might primarily to ‘have social prestige’) are no different from the sum of the intent of a million predators. Can you really say there was a fact of the matter regarding what was selected for in regard to wheat or cattle?

    4) “Metaphors are fine things; science probably couldn’t be done without them. But they are supposed to be the sort of things that can, in a pinch, be cashed.”

    yes I guess its a matter of parsimony preciseness and power (explanatory).

    5) “It is, in short, an entirely empirical question to

    what extent exogenous variables are what shape phenotypes; and it’s entirely possible that adaptationism is the wrong answer.”

    but as he stated already no one holds the position that only exogenous variables effect animals or that they are not limited by the possibilities – otherwise we would not just be well adapted – we would be gods.

    I think this leads to a valid point that natural selection actually hinders evolution – leaps in evolution (as in the creation of novelty) occur in situations of LOW competition because intermediary steps are probably unstable. I don’t think any serious biologists doubt that that is sometimes the case (although not everyone would agree with me that it is usually the case). If in a roundabout way he is supporting that position then cool but few of us consider it to make Natural selection invalid.

    6) “Such properties would co-evolve with tameness even if they have little or no systematic effect on fitness”

    but generally an evolutionary biologist might say that a certain trait is effected by natural selection for a certain reason – for that to be wrong and for it to be selected in the way he highlights there would need to be a more significant force pulling in the opposite direction or for them to be wrong about the facts of the matter.

  13. Hey you guys, these comments are long!

    I am going to have to come back and address them when I have more time…I have to go right now. But thanks for the comments, and I will get back to you.

  14. [Here’s]( his 2008 article from _Mind and Language_. (warning: pdf!) The blog post doesn’t mention one of Fodor’s main criticisms, namely that the properties needed to figure into laws of selection will have to be context-sensitive, and bona fide law-like generalizations just are not defined over context-sensitive properties. . . . Even if he’s wrong it helps to get clear on just what sort of laws are operative in selection.

  15. “He certainly doesn’t think that this should provide comfort to the intelligent design folks.”

    No, of course not. That idea is obviously false no matter how much the going theory sucks.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s