Saturday, February 19, 2011

Dan Dennett Replies

This is Daniel Dennett's reply to my earlier post: Daniel Dennett's View of Adaptationism.
Dear Larry,

Your blog was drawn to my attention today, and I decided it was well worth a response. Thanks for leading with your chin.

I’ve been wondering whether anybody would respond vigorously enough, and with enough authority, to my statement about the central importance of adaptation in evolution and the centrality of adaptationist thinking in biology to make it worth my time and energy to expand and explain. You have done so, and your commentators—especially anonymous and z—have already done a good job expressing at least a large part of my response, and I am grateful to them. I particularly endorse anonymous in his comment on the huge space of possible proteins and the fact that “functional proteins” are what inhabit that space.

You already accept the centrality of adaptation, as you say yourself in response to anonymous: “We agree that adaptation is a very important part of evolution and to ignore it completely would be ridiculous.” You need to remember that a great many non-biologists do not agree about this, and many of them have read the Gould & Lewontin essay (reputedly one of the most-cited papers in academia) as showing that (as Jerry Fodor once said to me, years ago) “adaptationism is completely bankrupt.” One of my chief aims in Darwin’s Dangerous Idea was to redress the balance, showing philosophers and other humanists and social scientists that they had to take evolutionary thinking (chiefly adaptationist thinking) seriously. Pluralism is not the lesson Fodor took from Gould & Lewontin’s essay, as his recent book with Piatelli-Palmarini makes clear. The “Spandrels” essay was the chief inspiration for his preposterous claim that “Darwin was wrong”. So I must ruefully admit that my efforts to squelch this widespread misreading of the message of “Spandrels” failed utterly to reach some thinkers.

But what about your biology students and their examination question about my notorious hymn to adaptationist reasoning? I wish you had also given them the passage in the same book, a few pages later, inwhich I quote Niles Eldredge and Michael Ghiselin, who make incautious claims about how we can replace “what is good” (adaptationism) with the more sober question “What has happened?” (pp240-41) I point out there that the very examples they cite depend, tacitly, on adaptationist assumptions—obvious assumptions but so much the better. I have much the same message for you, in response to your paragraph:

“I wonder how adaptationist thinking helps us understand sequence-based phylogenetic trees and the molecular clock. At the other extreme, how crucial a role does adaptationism play in deciding whether birds are dinosaurs or punctuated equilibria are the dominant pattern in the fossil record? I’m thinking that it might be a problem grading the answer to this question. Can a student defend Dennett’s statement and still get a passing grade?”

So let me take your wonders in turn. Nobody can reason about sequence-based phylogenetic trees without some assumptions about what historical processes created the data we now have available in the DNA of living and—in some cases—recently extinct species where DNA can be extracted. Those assumptions include, trivially, assumptions about the relative high fidelity of DNA replication and transmission, the role of DNA expression in (partially) determining phenotypic features, and the tendency of selection to weed out dysfunctional mutations and combinations.

Consider in particular how, for instance, biologists identify and explain gene duplication events. The uncontroversial interpretation of two suspiciously similar sequences in today’s DNA is as evidence—often considered conclusive—of a (roughly datable) duplication event followed by the preservation of one copy for its old (functional) role, freeing up the “extra” copy for exploitation/pruning for some new (functional) role. Duplication events just happen, of course, and not for any reason. The vast majority of them, we may safely suppose, disappear in a few generations or even sooner, but when they persist, it is because they get exploited and preserved for their functional roles. I suppose it is the obvious safety of the adaptationist assumptions here that hides them from view, creating the illusion, apparently, that there is no dependence on adaptationist premises here at all.

As for the molecular clock, it too cannot be relied on without help from adaptationist distinctions For one thing, you can’t distinguish Kimura’s neutral theory from Ohta’s “nearly neutral theory” without taking on board the role of slightly harmful gene differences that are subject to selection (which changes the rate at which such mutations go to fixation). And Ohta’s theory can explain some data that Kimura’s cannot. There are many other complications that arise for the molecular clock, regarding different rates in different taxa, that call for—and receive—clarification from adaptationist reasoning. For instance, molecular evolution in bacteria is faster than molecular evolution in mammals like us. Why? We have elaborate proof-reading systems the bacteria lack, and this raises the high fidelity of our replication processes. Trying explaining that without any appeal to function.

So there’s the answer to your first wonder. IF you want to avail yourself of the standard account of gene duplication events (to take just one uncontroversial example), and the limits on the utility of the molecular clock, you have to give adaptationist reasoning an essential role in your explanation—so central and unchallenged that it need not be mentioned.

Second wonder: “how crucial a role does adaptationism play in deciding whether birds are dinosaurs”? Well, unless you are asking a deliberately “philosophical” (as opposed to scientific) question about “where we draw the line” between (true) dinosaurs (with dinosaur essences) and true birds, a question that does not require or deserve an answer, you are asking for the evidence that birds descended, by a gradual sequence of intermediaries, from dinosaurs, and that is, I think, well established on multifarious grounds. All of those grounds depend, trivially, on assumptions about the absence (or huge unlikelihood) of hopeful monsters, on the necessary viability or fitness of all the intermediate forms, and so forth. Those are adaptationist assumptions. Perhaps because they are so uncontroversial they are not recognized as adaptationist, but there is no other reasoning that supports them. Those who think genetic drift explains a great deal—and of course it does—don’t make the mistake of holding that it can permit gene migrations across deep fitness valleys in adaptive landscapes. This shows the centrality, the non-optionality, of adaptationist reasoning even for the account of the birds-from-dinosaurs history. You cannot answer the “What has happened?” question without adding (sotto voce) “assuming that the descendants obeyed the fundamental constraints of natural selection”.

And lest anyone think that there is no more detailed role for adaptationist reasoning in the dinosaur-to-bird story, consider the question of how wings evolved and why they have the shapes they do. Now you don’t have to consider such questions, and the relevance to them of, say, the independent evolution of (functional) wings in insects and mammals, but it does seem to be an important part of the story.

What, then, of the question of whether “punctuated equilibria are the dominant pattern in the fossil record’? I haven’t encountered any reasoning about this issue that doesn’t involve discussion of whether the equilibria are due to stabilizing selection or other factors, and whether the punctuation episodes are driven by novelties in the environment (in adaptive radiations, for instances) or have some more endogenous trigger. So I guess I am unfamiliar, as a non-biologist, with the alternative settings of these issues that somehow avoid those adaptationist topics. I hope you will enlighten me.

Your final question is whether a student could agree with me and get a passing grade. Well, that is for you to answer: Is there is enough material in the various comments on your blog, and in my response here, to support a passing response on your examination? I have tried to answer your questions. Here is a question for you:

You preface the paragraph I quoted above with this question: “Can anyone figure out why biochemistry would collapse if we stop attributing everything to adaptation?” My question is: can you see that this is an unsympathetic caricature of my claim? I never say we need to “attribute everything to adaptation” or anything close to that. You agree that adaptation is important; and I expect you will agree with me now that (obvious) adaptationist reasoning undergirds all explanation of historical process in biology. That is what I mean by saying that adaptationist reasoning is “not optional, it is the heart and soul of evolutionary biology.”

I look forward to your reply.

Best wishes,

Dan Dennett


10 comments:

  1. The very existence of this whole discussion seems very odd to me. That so much of evolution is of non-adaptive nature had always seems like the much more philosophically interesting result than the role that adaptation plays. In fact, it happens to be one of the most profound insights into many of the questions philosophers have been pondering over the millenia that science has ever provided. So one would expect philosophers to be very interested in it, not to be sweeping it under the rug. Maybe there are such philosophers and I just haven't happened to hear about them (I don't claim to be familiar with the current trends in that field), but on the other side maybe there aren't any and that's actually why this discussions is happening...

    ReplyDelete
  2. This is a fascinating discussion. Thank you Larry and Daniel. The students are lucky to have it presented.

    ReplyDelete
  3. You already accept the centrality of adaptation, as you say yourself in response to anonymous: "We agree that adaptation is a very important part of evolution and to ignore it completely would be ridiculous."

    Some reasoning... So if one admits that ignoring something is ridiculous then one accepts the centrality of that something? Really???

    Nobody can reason about sequence-based phylogenetic trees without some assumptions about what historical processes created the data we now have available in the DNA of living and—in some cases—recently extinct species where DNA can be extracted. Those assumptions include, trivially, assumptions about the relative high fidelity of DNA replication and transmission, the role of DNA expression in (partially) determining phenotypic features, and the tendency of selection to weed out dysfunctional mutations and combinations.

    Wrong. One can (and many have) build and analyze sequence-based phylogenetic trees without *any* of the assumptions mentioned.

    Duplication events just happen, of course, and not for any reason. The vast majority of them, we may safely suppose, disappear in a few generations or even sooner, but when they persist, it is because they get exploited and preserved for their functional roles.

    Wrong again. One of the easiest ways to see why is read Larry's posts on "Onion test".

    At this point I lost an interest in reading any further. I find absolutely nothing fascinating in this discussion. Disappointing, maybe. Dr. Dennett needs to learn some biology before writing books about it. It's as simple as that.

    ReplyDelete
  4. "Dr. Dennett needs to learn some biology before writing books about it."

    DK

    Your comment is so rude I'm surprised it passed moderation.

    ReplyDelete
  5. ... and many of them have read the Gould & Lewontin essay (reputedly one of the most-cited papers in academia) as showing that (as Jerry Fodor once said to me, years ago) “adaptationism is completely bankrupt.”
    One should distinguish between "adaptationism is bankrupt" and "adaptation is bankrupt." For myself, I see adaptationism as bankrupt, yet I see adaptation as important.

    ReplyDelete
  6. This seems to be a defense of adaptationism purely on the grounds that nearly neutral evolution explains more than neutral evolution.

    This might be true, but I don't think it's legit to claim that nearly neutral evolution falls under adaptationism.

    Explaining something in terms of negative vs. positive selection often leads to very different predictions.

    For example, if we include nearly neutral evolution within adaptationism, then Michaeal Lynch-style explanations of genome size and "complexity" (low population sizes lead to large genome sizes and eukaryote-style genomic shambles) must be considered adaptationist.

    I'm pretty sure they aren't. (An adaptationist explanation of larger genome sizes might be that the extra materials are providing a selective benefit to the organisms that have them).

    The specific points raised are also a little funky. None of the assumptions mentioned are needed to build phylogenies. However, many models of evolution do incorporate e.g. different rates of change between different kinds of amino acids, which might be said to be assuming nearly neutral evolution. They generally allow change among chemically similar amino acids with reasonable probabilities. You wouldn't expect that under positive selection; you'd expect the opposite, if anything.

    Actually if positive selection was the most frequent kind of evolution, we wouldn't expect phylogenies to have much historical signal in them at all. Convergent evolution would be the dominant pattern.

    Evolutionary rate differences between human and bacteria don't seem to be because of differences in error repair (and to the extent they are, humans aren't necessarily better at repairing DNA), so not sure where that statement comes from.

    So I don't think it's correct to use Ohta's work to defend adaptationism...

    ReplyDelete
  7. DK: (re: duplication events) Wrong again. One of the easiest ways to see why is read Larry's posts on "Onion test".

    Dennett is referring to duplication of protein-coding genes, so the onion test isn't particularly relevant. DK, the discussion is fascinating... really, you should try giving both perspectives a careful re-reading.

    ReplyDelete
  8. @Anonymous:

    The claim that was made is completely wrong regardless of whether protein coding or non-coding sequence is considered.

    ReplyDelete
  9. Um, is any mainstream modern biologists that is not a pluralist with respect to adaptation/neutral evolution?

    I think that regardless of the actual relative role of adaptation and neutral evolution in explaining the diversity of life, adaptation has a more intuitive story, which is, in part, why it so popular. It's currency is also matter that we humans are familiar: limbs, wings, feathers. Neutral theory deals with codons, amino acids, phenotype spaces, and other unfamiliar oddities.

    ReplyDelete
  10. is any mainstream modern biologists that is not a pluralist with respect to adaptation/neutral evolution?

    Most biochemists that I know don't even know what neutral evolution is.

    ReplyDelete