More on Dick Lewontin and WEIT: what’s the deal with natural selection?

Several days ago I called attention to Richard Lewontin’s review of WEIT and several other books in The New York Review of Books.  In it, Dick (excuse the informality, but he was my Ph.D advisor) praises the book but takes me to task for implying that the evidence for natural selection is as strong as the evidence for evolutionary change per se:

Where he is less successful, as all other commentators have been, is in his insistence that the evidence for natural selection as the driving force of evolution is of the same inferential strength as the evidence that evolution has occurred. So, for example, he gives the game away by writing that when we examine a sequence of changes in the fossil record, we can

“determine whether the sequences of changes at least conform to a step-by-step adaptive process. And in every case, we can find at least a feasible Darwinian explanation.”

But to say that some example is not falsification of a theory because we can always “find” (invent) a feasible explanation says more about the flexibility of the theory and the ingenuity of its supporters than it says about physical nature. Indeed in his later discussion of theories of behavioral evolution he becomes appropriately skeptical when he writes that

“imaginative reconstructions of how things might have evolved are not science; they are stories.”

While this is a perfectly good argument against those who claim that there are things that are so complex that evolutionary biology cannot explain them, it allows evolutionary “theory” to fall back into the category of being reasonable but not an incontrovertible material fact.

There is, of course, nothing that Coyne can do about the situation. There are different modes of “knowing,” and we “know” that evolution has, in fact, occurred in a stronger sense than we “know” that some sequence of evolutionary change has been the result of natural selection. Despite these misgivings, it is the case that Coyne’s book is the best general explication of evolution that I know of and deserves its success as a best seller.

This “critique” has been picked up by several bloggers (see below), and I want to respond in a bit more detail.

First of all, yes, it’s true that the evidence for natural selection as the cause of most evolutionary change in the past is not as strong as the evidence that evolutionary change occurred.  It cannot be otherwise.  We can see evolution happening in the fossil record, but it is infinitely harder to parse out the causes of that change.   We weren’t around when it occurred, so we must rely on inference.  This difficulty is one reason why it took biologists much longer to accept natural selection than to accept evolution.   But to say that the evidence for selection is weaker than for evolution does not mean that the evidence for natural selection is weak, a conclusion I fear that creationists will extract from Lewontin’s comment.

Here is why selection still seems the best hypothesis for the origin of adaptive features of organisms.

1.  It is the only scientific theory, among all of those that have been adumbrated, that currently makes sense.  Failed explanations include teleology, intelligent design, and Lamarckism.  Some of these were once valid scientific alternatives to natural selection, but have failed either because they are untestable or because they were testable and shown to be wrong. If Lewontin and others want to say that some process other than selection is responsible for the limbs of tetrapods, the fins of whales, and the white color of polar bears, they must say what they envision.  Yes, Lewontin and Gould showed that many things for which we can concoct adaptive stories may be “spandrels” — nonadaptive traits hitchhiking on other adaptations — but this does not mean, as Lewontin seems to imply, that selection may not play a major role in creating adaptations.

2.  In cases where we can actually investigate whether selection is responsible for an adaptive change in a species, it is.  I give several examples in WEIT, including coat colors in mice and the famous work of Rosemary and Peter Grant on Darwin’s finches. And of course there are those dozens of cases of antibiotic resistance in bacteria, insecticide resistance in arthropods, and herbicide resistance in weeds.  In bacteria, for instance, we can show that the genetic variation for resistance preexisted in the population and not invoked by the selective agent, precisely as the theory of natural selection posits.

3.   In tests where we envision that selection was responsible for an adaptation, we can do laboratory tests to see if the adaptation at least gives a fitness advantage to those individuals possessing it.  One example of this is the Browers’ work on Batesian mimicry in the viceroy butterfly.  It was shown that exposure to a toxic monarch made naive bluejays sick, and that later these bluejays avoided the nontoxic viceroys, giving a survial advantage to mimics.  This is precisely what has to happen for that mimicry to evolve by natural selection.  Likewise with color in guppies:  brightly colored guppies get eaten more often in Trinidadian streams than do their duller confreres.  This explains why guppies are less colorful in predator-filled streams.

4.  The prerequisites for selection — the heritability of traits, the fact that there is competition between individuals, and that there are fitness differences between individuals with different traits — have all been demonstrated in living organisms in nature.  If few traits showed any heritable genetic variation, we’d be justified in rejecting selection as a major cause of evolution. Guppy coloration is heritable.

5.   Even in ancient species we can test the likelihood that selection caused evolutionary change.  Horses lost their toes right about the time when the forests were disappearing on the Great Plains.  We know that hooves are more effective adaptations for running in open grassland than are multi-toed feet. Likewise, horse teeth become higher and more robust precisely when silicon-rich grasses were replacing the leafy forests.  We know that herbivores need higher and more robust teeth to deal with grass. It is a good inference that the appearance of grassland was the selective factor promoting the loss of horse toes and the change in horse teeth.

6.  As discussed in previous posts, selection as we envision it has certainly been adequate to explain the evolution of complex adaptations like the eye, and in geologically reasonable periods of time.  Therefore it remains a viable hypothesis for adaptive change. This didn’t have to be the way it turned out.

What about my supposed double standard about accepting natural selection for many traits but being skeptical when it comes to evolutionary psychology?  This is a reasonable tactic for one important reason: we have many more alternative theories for the appearance of human behavioral traits than we do for morphological adaptations in other species.  How many alternative theories do we have for the appearance of flippers in proto-whales, or for the movement of their nasal passages to the top of their heads?  In contrast, there are many alternative theories for the appearance of traits like human rape, depression, music, art, religion, etc.  Blowholes aren’t likely to be spandrels; the appearance of music and poetry might well be.  Humans have culture and rationality to a degree possessed by no other animal, and can learn many things not permitted in species having smaller (or no) brains.   That’s why we need to be more cautious about imputing selection to human behaviors than to blowholes.

Now I think Dick did have a point: I should have pointed out (though I might have; I can’t remember!) that it is a lot easier to come up with evidence for evolution than for selection.  But I think Lewontin’s own anti-selectionist biases are intruding here. As I mentioned in an earlier post on ideological grounds neither he nor Gould were ever very strong promoters of selection. I’m not sure what the connection is between selection and politics (it may be the misuse of selectionism that Gould and Lewontin saw among sociobiologists), but neither of these chaps were avid promoters of selection.  They preferred to emphasize other processes, including pleiotropy, spandrels, genetic drift and the like. I think this was a deliberate strategy.

Over on EvolutionBlog, Jason Rosenhouse analyzes Dick’s review and has some good comments:

I don’t understand what it means to say that “natural selection is the driving force of evolution.” Given Lewontin’s past writing (most notably his spandrels paper with Stephen Jay Gould) I would guess that his point is that some biologists are too quick to attribute some anatomical feature of some organism to the prolonged working of natural selection.

That may be true, but when we are talking about adaptations the evidence for natural selection seems to me to be very strong. For one thing, it is the only natural mechanism known that can account for complex structures (like bird wings or vertebrate blood clotting systems). For another, every complex structure studied to date shows clear evidence of being a cobbled together Rube Goldberg machine, which is exactly what we would expect if they were crafted by natural selection.

On top of this, biologists routinely use adaptive reasoning to generate testable hypotheses about the creatures they are studying. Lewontin would know better than I whether biologists engaging in flights of fancy is a genuine problem in the field, but it is undeniable that “the adaptationist program” has yielded great dividends over the years . . .

. . In fairness, I think Stephen Jay Gould was pretty clear on this point [the ubiquity and importance of selection] in several of his essays. I compiled some of his statements on the matter in this essay. But I share Coyne’s frustration. I’ve never really understood what it is exactly that anti-selectionists are complaining about. If they agree that complex adapations arise as the result of gradual accretion mediated by natural selection, then I fail to see how they are really so different from people like Richard Dawkins or Daniel Dennett (two people often described as being beknighted uber-selectionists). If they do not agree then I would like to hear their proposed alternative mechanism.

Now I have great affection for Lewontin (as all of his students say, “I love that man”), but I would like to see him make an explicit statement about what aspects of nature he imputes to natural selection.  We’re not just talking about rape, male domination, and music here, but coat colors, physiology, feathers, gills, flowers, toxins, and the like.  Like Jason, I think the anti-selectionists have gone way, way overboard, and have thrown out the baby with the bathwater.  (These people also include the “structuralists,” and those who attribute adaptations to the self-organizing properties of biological matter.)


  1. Dave
    Posted May 12, 2009 at 10:12 am | Permalink

    I’ve only skimmed through your post so far, so I apologize for a this simple observation. I have always cherished this debate though, and it’s great to see you bring something forward.

    You wrote:

    –“They preferred to emphasize other processes, including pleiotropy, spandrels, genetic drift and the like.”–

    This was an observation Dawkins made some time ago in an interview, he said the real differences between him and Gould were ones of “emphasis”. Of course, Gould being the Bully for Brontosaurus had a way to take this argument to the fairly powerful conclusion that Dawkins and Maynard Smith were akin to Darwinian fundamentalist. The political aspect you mention is also discussed in a fairly good book; Dawkins vs. Gould: Survival of the Fittest. Some of this debate reminds me of the multi-selection vs. selfish gene[as complete] theories. It also appears to be one of possible emphasis. David Sloan Wilson actually makes that similar argument in his, what I consider quite excellent blog post; Truth and Reconciliation. I’m still waiting for anyone to offer the transformation of the obvious. 🙂

    Can’t wait to have more time to digest your essay here, thanks…

  2. Dave
    Posted May 12, 2009 at 10:32 am | Permalink

    I was shaving thinking…

    I’ve always wondered about the emphasis Dawkins has used. He hasn’t been a practicing scientist for about three decades now and some reactions I’ve seen from him to E.O. Wilson seem to be around the emphasis argument, but he appears to rant (to the point of near irrationality). Why? I’ve always thought that Dawkins’ motivations, emphasis and reactions are also possibly influenced by a cultural view. Sometimes at once to argue we can not overreach evolution, but then actually employ largely non-reducible evolutionary mechanisms as explanation (though this point seems limited – it’s focus primarily on religion and religious beliefs).

  3. Dave
    Posted May 12, 2009 at 10:58 am | Permalink

    Just another quickie while I’m thinking about it. I thought about this the first time I read you say “anti-selectionist”. I’ve seen Dawkins say something about multi-level selection like “anti-gene selectionist”. Though, it seems you and others who make the charge recognize it not completely against. If I am anti-slavery, I am not somewhat for it, I don’t argue I’m accepting of it at some level. If I recognize it as a reality, I can still be apposed to it and recognize it, but if I say I am anti-slavery that means I am arguing it is completely wrong – it is beyond the question of emphasis, it is a dismissal of the rational. So, in a way it appears the emphasis may be cutting both ways and perhaps while each claims blinders and lack of pure justification what we are dealing with is some open questions, not “anti” movements. It’s akin to Wilson, both now, saying that Dawkins is missing the point by claiming multi-level selection dismisses kin selection, it does not such thing, in fact it appears possible that its not only incorporated but Dawkins objections simply misses that and instead claims it is an “anti” position. Some of this reminds me of the position you hold on the Pigliucci inspired Altenberg conference.

  4. NewEnglandBob
    Posted May 12, 2009 at 11:14 am | Permalink

    I have seen no evidence for multi-level selection that overrides natural selection.

    I dismiss the ‘spandrels’ paper and silliness.

    Dave, I do not agree with your second post and the third one verges on incomprehensibility. What do you refer to as “transformation of the obvious” in your first post?

    • CharlesInCharge
      Posted May 12, 2009 at 11:48 pm | Permalink

      In my opinion, Spandrels was a crap philosophy of biology paper that spawned a lot of good work on the topics they raised.

      I’ve always found Gould and ESPECIALLY Lewontin were exceptionally good at mischaracterizing others’ arguments.

  5. Dave
    Posted May 12, 2009 at 11:22 am | Permalink

    Ok, read it through in a bit of detail.

    Lewontin states: ~”There is, of course, nothing that Coyne can do about the situation. There are different modes of “knowing,” and we “know” that evolution has, in fact, occurred in a stronger sense than we “know” that some sequence of evolutionary change has been the result of natural selection.”~

    It appears you address this immediately and only illustrate what he is recognizing as fact. However, I fail to recognize that exposing potential weakness means dismissal as the best current theory, especially since no other is offered. Which makes your argument about competing theories irrelevant. I had seen this with Gould, here was Gould defending evolution by natural selection and others would read into his work that he was “anti-selectionist” to the point to which are doing now, which is to paint Lewontin as anti-natural section. Saying the “theory” is not incontrovertible material fact seem reasonable, though it obviously seems to pressure you to ask, well what else is there. But, that misses the point, doesn’t it? It appears you are making a blanket assumption in that your list means to offer Lewontin skepticism means a rejection of this facts a priori. I think his juxtaposition of using evolutionary psychology may be well taken and makes me wonder why you see it as an opportunity to point out competing theory’s. I wonder who here is throwing out the baby with the bathwater, or at least the one who thinks the baby is still in the water when the water is thrown out.

  6. Dave
    Posted May 12, 2009 at 11:28 am | Permalink


    -“I have seen no evidence for multi-level selection that overrides natural selection.”-

    I actually think that may be part of the problem with some of these debates. You are assuming that mult-level selection is meant to do away with natural selection, that even kin selection is dismissed. That is just a mistake, a radical misunderstanding. I like the Spandrels, in fact it reminds me of other by-product theories that have been floated around. I also use that argument to some extent to show how humans were not inevitable to exist.

    The third post is only a thought on labeling and the problems that can stem from it, a cautionary tail more or less.

    When I say “transformation of the obvious” I am simply using the phrase to point out that I have not seen the argument that settles the arguments.

  7. James
    Posted May 12, 2009 at 11:36 am | Permalink

    “I’m not sure what the connection is between selection and politics (it may be the misuse of selectionism that Gould and Lewontin saw among sociobiologists), but neither of these chaps were not avid promoters of selection.”

    Didn’t you mean ” neither … were avid promoters … “?

  8. Dave
    Posted May 12, 2009 at 12:02 pm | Permalink

    In other words, Bob.-

    What I am partly saying is it is a question of modified mechanism, not complete dismissal of the theory. We are dealing then with potential expanding explanations of natural selection. It is not throwing out natural selection, Jerry mentioning Intelligent Design actually makes me wince, but there have been competitors against a strict formulation of Darwin’s theory which have been dismissed, however that does not mean we are dealing with incontrovertible fact for potential takers where weakness is exposed (lately in fact it seems many sides quote Darwin a great deal to show he understood their position). I mean Jerry almost seems to completely dismiss potential importance of any exaptations, again a question of emphasis. The blowhole, ok probably no, cultural by-products, possibly, that’s used in many cultural contexts, and so is the addition of cognitive based arguments, these of course could be incorporated or dismiss a more multi level approach.

    Again, as with what happen to Gould, the potential is there to overstate the case and label for fear of how appointments of evolution by natural selection may react and use. This happened with punctuated equilibrium and it happens anytime there is debate over the mechanisms of evolution. But, to bad, let science do what it does best. I think Jerry is doing a fantastic job with his book and explanation evolution, he doesn’t need to worry needlessly, they have no theories, they are putting out reformulations of way to keep the faith, we can’t give in to that.

  9. Dave
    Posted May 12, 2009 at 12:19 pm | Permalink

    Need to correct at least one of my errors:

    “Again, as with what happen to Gould, the potential is there to overstate the case and label for fear of how opponents [not appointments] of evolution by natural selection may react and use.”

  10. araujo
    Posted May 12, 2009 at 1:56 pm | Permalink

    The criticism of evolutionary psychology are unfair. See the work of John Tooby and Leda Cosmides, in the Center for Evolutionary Psychology, David Buss and many others. See many articles in

  11. Jeremy
    Posted May 12, 2009 at 2:00 pm | Permalink

    Jerry, I think this is a fantastic post. I’ve always been rather irritated with Gould/Lewontin/et al.’s glib dismissal of adaptionism, for they never seem to offer any constructive methodology in its place. (I think this criticism applies even to their “spandrels” paper.)

    They may not see it (or more likely, they don’t want to see it), but adaptionism has long-since been placed upon a sturdy theoretical framework by luminaries such as G.C. Williams (Adaptation and Natural Selection) and Dawkins (The Extended Phenotype). For instance, Williams devotes a large section of the above book to delineating just when it is inappropriate to use the “onerous concept” of natural selection to explain a particular trait. Gould and Lewontin bequeathed us no such boon, but preferred to simply pour scorn upon the entire enterprise.

    • NewEnglandBob
      Posted May 12, 2009 at 2:04 pm | Permalink


      You expressed my thoughts exactly.

  12. araujo
    Posted May 12, 2009 at 2:01 pm | Permalink

    Link for EP articles:

  13. Dave
    Posted May 12, 2009 at 11:40 pm | Permalink

    I am honestly confused by your reaction to this Jerry.

    Take the quote he pulled:

    “determine whether the sequences of changes at least conform to a step-by-step adaptive process. And in every case, we can find at least a feasible Darwinian explanation.”

    Now, it appears he’s simply taking you to task for perhaps stepping a bit to far where it may not be warranted. You are saying in “every case” we can “find”, but to say that gives the appearance you are prepared to give certainty that you could create the explanation before you even have the observation. There is no uncertainty to your statement, you are in essence saying that your explanations are unfalsifiable, because in “every case” you can feasible Darwinian explanation (which I even take you to mean adaptation in every case), no exceptions, which further seems like you could just invent the appropriate explanation to fit the preconceived idea that it is possible to do in every case.

    I think this is why he pulls out EP. You are showing skepticism for the “just so” stories, but how do we know you’re not prepared to offer “just so” stories. Again, it’s not to do with the possibility of other potential theories, or even the weakness in natural selection, only a casual warning of overstepping your claims.

    I think Jason is making the same error. Lewontin is saying natural selection is the driving force, but by overstepping you only show how clever you may be, but that does not add to the theory without proper validation, otherwise it’s unfalsifiable, “just so” stories.

    That is not determining inherent weakness of the theory, he is only offering the point of demarcation between the weight given to the facts of evolution and to the theory to further clarify the point you are actually arguing the theory is as strong without exception (“every case” “find”).

    It really does appear that you both are reading past him, perhaps because of a preconception of what one might expect from Lewontin.

  14. Dave
    Posted May 13, 2009 at 12:05 am | Permalink

    Another thing I find interesting by the positions both you and Jason take on another issue that is in fact related to this.

    Jason wrote: “I wouldn’t care so much but for the fact that the stridency of some anti-selectionist rhetoric (see Lynn Margulies, for example) gives ample fodder to the creationists.”

    Jerry wrote: “But to say that the evidence for selection is weaker than for evolution does not mean that the evidence for natural selection is weak, a conclusion I fear that creationists will extract from Lewontin’s comment.”

    I think this also spells trouble and perhaps explains why there could be a bit of over concern to the point of possibly contributing factors which are not evident in Lewontin’s critique. That, and a level of expectation due to past experience.

    It’s almost hard to believe that this worry about what those scary creationist will do with this is still shows itself so prevalently, possibly to the point of clouding judgment. Of course they will exploit what they can, but heaven for bid this creates an atmosphere of silence due to fear that the enemy is listening in. Perhaps Niles Eldredge and Stephen Jay Gould shouldn’t have snuffed their theory knowing it gave an appearance of a supposed spontaneous creation which those little devils love. Heck, lets go back to the Origin, what a mess that made.

  15. Dave
    Posted May 13, 2009 at 12:08 am | Permalink

    Correction: should have* snuffed….

  16. David Burbridge
    Posted May 13, 2009 at 5:15 am | Permalink

    The hypothesis of selection is not always true, but it is usually a good starting point for research. In the century or so after 1859, valuable discoveries were often missed or delayed because selection was not taken seriously enough. For example, even in Darwin’s time it was known that many fish had organs which generated electric fields, but with a few exceptions (like the electric eel) they were not strong enough to shock or stun prey. But instead of saying ‘Hmm, these organs must be useful for something, let’s find out what it is’, most biologists said ‘Aha, these organs are useless, so they can’t be explained by natural selection’. In consequence, discovery of the electric senses of fish (which are used to detect prey, to navigate through murky water, etc) was delayed by a century or so.

  17. Dave
    Posted May 13, 2009 at 6:47 am | Permalink

    I think David Burbridge offers a decidedly excellent cautionary tail. It illustrates an almost radical form of skepticism – rejectionism without due process. The problem of course is “most” biologist were doing it and so it became self fulfilling, how do you know that it is it incorrect to assume useless when the standard becomes fact. The explanatory mechanism which dissolved the debate eventually won out by force of reasoned argument and further detailed evaluation of the evidences.

    I think to some who are within these debates may fail to recognize how isolated the discussions and evidences are to the rest of us. A fisherman may have deduced the the eel’s useful biological behavior by observation centuries ago and it took more than a 100 years to scientifically validate as useful adaptation.

    I think the question of emphasis is a good one. Unjustified a priori rejections are not the norm in science, that position is reserved mainly for the creationist. However, skepticism is a warranted an useful trait of the scientific process. I think it can also catch when someone is possibly over stating a case, which I think that is what Lewontin is doing here, and not much else. For example, is there any evidence that Lewontin would dismiss the evidences Jerry’s offers in defense of selection? I think Lewontin would accept all 6 (?) quite handily. Which only raises concern when Jerry says something like: -“If few traits showed any heritable genetic variation, we’d be justified in rejecting selection as a major cause of evolution.”- It simply goes back to emphasis. I think a troublesome aspect of what appears to be a very defensive position (perhaps partly unjustified to reading past) is one misses the utility offered by a Lewontin skepticism and further attempts to minimize the contributions of those like Gould and Lewontin (and with other proposed additions to preexisting mechanisms). Suddenly its all about complete rejecting and “anti natural selection”, when in fact no evidences has been forwarded that Lewontin is anti-natural selection, however is skepticism is viewed and where ever emphasis is placed on evidential counter examples which shit, modify or clarify the norm.

  18. Dave
    Posted May 13, 2009 at 6:59 am | Permalink

    Fix last sentence, perhaps I’ll do the longer comments in a word program first, these little boxes are not very conducive to longer comments and its easy for me to miss misspellings and errors in sentence structure.


    “however his skepticism is viewed and wherever emphasis is placed on evidential counter examples which shit, modify or clarify the norm.”

  19. Adam Dimond
    Posted August 3, 2009 at 5:40 am | Permalink

    If we accept A Type of God as a first condition, then the disparity between natural (physics, chemistry and biology)and social(sociology, anthropology, philosophy, and religious belief) is logically resolved.

    • NewEnglandBob
      Posted August 4, 2009 at 2:49 pm | Permalink

      That statements says absolutely nothing. That is a word salad with nonsense dressing.

  20. Marcello Pucciarelli
    Posted August 26, 2009 at 10:18 am | Permalink

    Quote: “Here is why selection still seems the best hypothesis for the origin of adaptive features of organisms…”

    I don’t dispute the reasons you give, but perhaps the main issue is that ‘adaptive features’ are (by definition?) the product of natural selection. Besides, ‘origin’ and ‘fixation’ of adaptive features are not the same thing, and may involve different kinds of explanation.

    I think Lewontin’s ‘anti-selectionism’ has its roots (and justification) in his understanding of the differences between ‘Darwinian selection’ (traits improving the fitness of their bearers will spread in the population) and natural selection as understood by population genetics. He strongly objects to projecting back onto Darwinian selection the explanatory power of population genetics (see the paper coauthored with André Ariew, The Confusions of Fitness). To be sure, Lewontin also stresses the shortcomings of population genetics, but to me this seems a different kind of criticism (more ‘from inside’) than the one reserved to Darwinian selection.

    As for Gould, I would like to recall that in the paper coauthored with Elizabeth Vrba on Exaptation he took a quite traditional stance in defining adaptation (i.e., a feature modeled by natural selection). In his paper on the Irish Elk he criticized the ‘correlation of growth’ hypothesis for explaining the animal’s huge antlers, and proposed an adaptive scenario. I think it might be time to reassess whether Gould really deserves the label of ‘anti-selectionist’, since ‘integrative’ could be more appropriate.

One Trackback/Pingback

  1. […] Follies: 13 May Morning Edition Science Here is a discussion of a review of Jerry Coyne’s book Why Evolution is True. The gist of the review is a discussion of how much evidence we have for […]

%d bloggers like this: