Monday, September 30, 2013

The Problems With The Selfish Gene

Lots of people fail to understand that the "selfish gene" is a metaphor. They criticize Richard Dawkins for promoting the idea that genes can actually take on the characteristics of selfishness.

Andrew Brown and Mary Midgley are prominent examples of people with this kind of misunderstanding and Jerry Coyne has set them straight in Poor Richard’s Almanac: Andrew Brown and the Pope go after The Selfish Gene and “Selection pressures” are metaphors. So are the “laws of physics.”

However, there are two other problem with the metaphor. The first is rather trivial, it refers to the fact that it's actually alleles, or variants, of a gene that are "selfish." Dawkins knows this. He explains it in his book but I don't think he puts enough emphasis on the concept and in most parts of the book he uses "gene" when he should be saying "allele." I grant that The Selfish Allele is not a catchy title.

The second problem is the emphasis on natural selection—a hallmark of Dawkins' writing. He doesn't even consider the success of "lucky" genes (alleles) that aren't "selfish." His readers are left with the impression that only selfish genes succeed. The Lucky Gene is a book that still has to be written.

Jerry Coyne thinks that The Selfish Gene is a very important book. He says, ...
The Selfish Gene, which has by now sold well over a million copies in a gazillion languages, is a seminal work, and has opened the eyes of millions to a gene-centered view of evolution and all that it explains: cooperation, conflict, and, in its brilliant central metaphor, the process of natural selection. I can’t count the number of people who have told me, either in person or on this site, that it changed their lives, opening them up to the wonders of science.
Jerry seems to be thinking of non-scientists. I'm not sure if scientists, especially biologists, think the same. I believe that most biologists still think of individual organisms as the primary unit of selection (evolution). Most of them have not adopted the gene-centric view of evolution expressed by Dawkins in The Selfish Gene. Neither have I.

I understand population genetics and its central role in evolution but that's an entirely different model than the one Dawkins writes about in his book. Dawkins doesn't talk very much about evolving populations or allele frequencies. That's a shame because, if he had, he would have made a much greater impact on people's understanding of evolution.

I do not think The Selfish Gene is a "seminal work." To me it's a collection of just-so stories and it serves as the perfect example of the view that Gould and Lewontin criticized in their Spandrels paper. What do others think?

I know that Richard Dawkins is very proud of the book. It was first published in 1976 and republished in 1989. In the preface to the 1989 edition, Richard Dawkins says ...
In the dozen years since The Selfish Gene was published its central message has become textbook orthodoxy. This is paradoxical, but not in the obvious way. It is not one of those books that was reviled as revolutionary when published, then steadily won converts until it ended up so orthodox that we now wonder what the fuss was about. Quite the contrary. From the outset the reviews were gratifyingly favourable and it was not seen, initially, as a controversial book. Its reputation for contentiousness took years to grow until, by now, it is widely regarded as a work of radical extremism. But over the very same years as the book's reputation for extremism has escalated, its actual content has seemed less and less extreme, more and more the common currency.

The selfish gene theory is Darwin's theory, expressed in a way that Darwin did not choose but whose aptness, I should like to think, he would instantly have recognized and delighted in. It is in fact a logical outgrowth of orthodox neo-Darwinism, but expressed as a novel image. Rather than focus on the individual organism, it takes a gene's-eye view of nature. It is a different way of seeing, not a different theory.
I still see the book as controversial. I do not think that the "orthodox neo-Darwinism" described in the book has become "textbook orthodoxy."


  1. I think Gould and Lewontin are perfect examples poor argumentation on many fronts, to say nothing of their frequent howlers on Evolutionary Psychology. Thankfully, Dawkins is getting the credit he is due in the field of biology. And so are Lewontin and Gould, who both hold the distinction of having entire fallacies named after them. ;)

  2. Dawkins frames his discussion on the units of selection based on the replicator question, not the interactor question as Gould did (as defined by Elisabeth Lloyd: He understands that there could be different interactor levels (individuals, populations, etc), but recognizes the individual organism as being the most likely, and argues against the importance of the population level in the selection process.
    I you read the Selfish Gene then you will understand why he chooses this question and discusses the matter in terms of inheritance and genes.

  3. I have Jerry's book "Speciation" on my desk. It deals, specifically, with the process of speciation, deemed -the- major issue in evolution by Jerry. It has a heavy genetic focus.

    Per the index, there is a single citation to Dawkins. Eight to Gould, seven to Lewontin.

    The citation to Dawkins is rather neutral, as are almost all the citations to Gould. Jerry knocks the founder-effect explanation for punk eek but then points out that not even Gould believed it circa 2002. I didn't check the Lewontin citations, but since he was Jerry's advisor and they mostly agreed, I'm guessing none are overly critical. The one citation to Dawkins is not to the Selfish Gene.

    Michael Payton: Lewontin was a poor arguer? Are you serious? He's one of the most important and influential evolutionary biologists of the past 50 years. Even Gould, most people know only his "flashy" work on punk eek and his popular writings, rather than his laying the foundations for a more quantitative approach to the history of life.

    If you disagree with the spandrels paper, we can certainly argue about that specifically, but to assert that both of those two were poor arguers untenable position, to say the least.

  4. It appears that you have to declare your allegiance either to Dawkins or Gould as champion of evolutionary biology. Larry, are you holding out for Kimura? Remind me why we're doing this.

    Now, I do think there's selection on multiple levels. The gene-centric view is often useful. So is the individual-centric view. Even, though rarely, the population- or species-centric view. Selection is important. So is drift, though not so much when you're trying to explain adaptation. Why can't we all just get along?

    1. Exactly. Neither Dawkins nor Gould were particularly impressive as actual working evolutionary biologists as opposed to popularizers. Personally, I like the Sean Carroll style of popularizer who has actually earned their stripes in the peer reviewed literature before seeking to influence the public.

    2. Jonathan Badger says,

      Personally, I like the Sean Carroll style of popularizer who has actually earned their stripes in the peer reviewed literature before seeking to influence the public.

      Sean is keen on storytelling as a way of teaching science. I disagreed with him when we talked about this last week but I didn't press the issue because there were other things we needed to discuss.

      Here's what Michael Lynch says in his book. Let me know if you still think Sean Carroll is a better popularizer than Richard Dawkins or Stephen Jay Gould.

      ... numerous biologists, particularly in the area of development, have expressed reservations about the entire population genetic enterprise. Consider this quote from Carroll [in "Endless Forms Most Beautiful"]: "Since the Modern Synthesis, most expositions of the evolutionary process have focused on microevolutionary mechanisms. Millions of biology students have been taught the view (from population genetics) that 'evolution is the change in gene frequencies.' Isn't that an inspiring theme? This view forces the explanation toward mathematics and abstract description of genes, and away from butterflies and zebras. ... The evolution of form is the main drama of life's story, both as found in the fossil record and in the diversity of living species. So, let's teach that story. Instead of 'change in gene frequencies' let's try 'evolution of form is change in development.'" Many similar statements could be quoted from other authors.

      Even ignoring the fact that the vast majority of species are unicellular, differentiated by metabolic rather than developmental features, this type of statement paints an inaccurate portrait of the current field of evolutionary biology. Evolution is much more than a storytelling exercise, and the goal of population genetics is not to be inspiring, but to be explanatory.

    3. Obviously, I'm a microbiologist myself interested in evolution at the molecular level, but I hardly see how Carroll is really any worse in this regard than either Dawkins or Gould as neither of whom were particularly interested in non-metazoa or molecular evolution.

      And I hardly think "Endless Forms" is guilty of any more storytelling than is Gould's "Wonderful Life" or Dawkins' "Ancestor's Tale". Making a collection of facts and theories into a "story" is what makes it popularization.

      BTW, you may be amused that circa 1990 when I was an undergrad at the University of Wisconsin, Sean was my biochemistry professor -- bet you didn't think he ever taught that!

    4. @Jonathan,

      Thanks for replying. I guess you don't really see any differences between Sean Carroll, Richard Dawkins, and Stephen Jay Gould when it comes to the popularization of science.

      I guess your earlier statement (above) just means that you prefer Sean Carroll's style for reasons that have nothing to do with how effective it is.

    5. I didn't say that they are not any different, I said in terms of simplifications his work wasn't any worse than theirs. "Endless Forms" is clearly better because rather than the bizarre pseudo-molecular metaphors of "replicators" and so forth that Dawkins prefers, or the emphasis on paleontology (which is not really a hot topic these days) that Gould preferred, Caroll is talking about evo-devo, a field which *is* a major research topic today, and one for which he has been recognized by practically every award short of the Nobel.

      I'm not an evo-devo person myself, and I understand the complaints of people like Lynch (or myself for that matter) that work in less "sexy" fields, but the fact is a high school student reading Carroll would have a far better idea of what topics are current in biology than they would reading either Dawkins or Gould, and in the end, that's why popularizations are relevant.

    6. Paleontology is not a hot topic these days?! Try telling that to the Chinese who are discovering loads of transitional bird and transitional mammal fossils. Did you notice the recent clamour about Entelognathus - the first placoderm fish discovered with a jaw similar to bony fishes & tetrapods, again from China.

      Nothing can capture the imagination of the public quite like stunning fossils. Paleontology will always remain a hot topic

  5. I read "The Selfish Gene" when I was a line cook in my mid-twenties beginning to develop a fascination with evolution. It was a good place to start. Now I'm a grad student in my early-thirties, and although I have a certain nostalgic fondness for this book, I can see that it over-emphasizes the role of natural selection to the exclusion of so many other evolutionary mechanisms.

    I think the adaptationist focus is fairly common in popular books about evolution. During the past few years I've learned a lot about the other more stochastic mechanisms of evolution, sometimes from this very blog. I think it's time for Larry Moran to write "The Lucky Allele". There's a world of line cooks and other non-scientists out there who would read it. Maybe some of them might become scientists too. Then maybe they could leap frog over the adaptationist stage of their own intellectual evolution.

  6. "Allele" is a dreadful term. "Gene" has pretty-much eaten its lunch. You can't tick people off for using "gene" instead of "allele" any more - everyone who's anyone does that.

    1. The problem is that "gene" is stretched too far. You need the other terms. If we just use "gene" I would say that at my Hemoglobin Beta "gene" I have two "genes" (one from my mother and one from my father). However they may or may not be the same "gene". It's a lot clearer if (for the last three uses) I say "locus" for the first one, "gene" or "gene copy" for the second, and "allele" for the third.

    2. I generally use "gene locus" and "meme locus" if I want to refer to a contextual spot and "variant" for a modified copy. "Allele" is irregular terminology. Are we *really* supposed to use 'allelomeme' for an inherited cultural variant? None of the "alllele" terminology proponents appear to have thought this issue through properly :-(

      IMO, we should not be fighting to preserve "allele", but struggling to destroy it. Those who treat use of "allele" as a badge of evo-bio status and understanding are not helping!

    3. We linguists use the prefix allo- to refer to any of a number of different variants of the same abstract entity -- e.g. an allophone is one of many possible phonetic realisations of a phoneme, an an allomorph is a variant shape of a morpheme. Our terminological pattern would yield "allogene" and "allomeme".

    4. "Gene" and "meme" won long ago as terms for these concepts though.

      If you search you'll see that "dominant gene" beats "dominant allele" and "recessive gene" beats "recessive allele". "Gene" and "meme" are the best terms - they should may to the most useful concepts - and these days, they do - perhaps thanks in part to Dawkins, Waddington and other "gene" enthusiasts.

  7. I think proponents of all the different views of evolution have a habit of overstating the importance of their particular pet view. There's pretty far between the truly balanced view.

    In the end, how does one even quantify to which extend the different modes contribute to evolution? It seems an almost impossible task with multiple modes (neutral drift, selection) of evolution operating on multiple levels (allelic, individual, species). Instead of even tryiing to answer which one is "most important", shouldn't we simply teach and emphasize that to properly understand evolution, we need to include all of them, we don't necessarily have to decide which one is "biggest"?

    Scientists and their hobbyhorses XD

    1. It is certainly unjustified to argue that group selection and species selection are invalid, or unscientific, or impossible. Natural selection can occur at all levels. However, it is equally wrong to wave one's hand blithely and invoke group or species selection without checking to see if its conditions are met. The higher up you go the more restrictive are the conditions.

      A skepticism of the power of group selection or kin selection is justified. Individuals die every generation. Local populations and whole species die much less often. So there are good reasons to worry about which of these levels of selection is most important, but no justification for saying that the higher levels are impossible.

  8. I think that Dawkin's selfish gene is a very useful metaphor for introducing basic concepts in evolution. It is far easier to talk of "genes" as fixed entities than to discuss the slippery concepts of phenotypes. However, when using the genic view as a way of introducing evolution the second sentence should be "of course it is so oversimplified as to be completely incorrect."

    I have started to develop a counter argument in favor of a phenotypic view of evolution. Look for a summary to date this week that ties talk of genetic complexity back to why the genic view of evolution is basically a failure.

    1. I would not go as far as Charles, but I would say that TSG is a brilliant piece of popularization. As a scientific treatise it leaves too much out. I've heard web site commenters claim that TSG had a major effect on evolutionary biology. I don't think so, though it did establish Dawkins's reputation as a populariser.

      The "gene-centered" view has its predecessor in JBS Haldane, but was really brought to the attention of evolutionary biologists by Bill Hamilton, and by George Williams and John Maynard Smith in papers and books of 1964-1966. By the late 1960s and early 1970s evolutionary biologists were aware of models of kin selection as an explanation of evolution of social interactions. E.O. Wilson's book Sociobiology: The New Synthesis in 1975 was a major presentation of the implications of this, though it was critcized for its discussion of human behavior.

      All that was before TSG appeared in 1976.

    2. When you argue for the phenotypic view, don't forget Shozo Yokoyama and my 1978 paper in PNAS on Kin selection in an altruistic trait considered as a quantitative character. (I suspect that you are not unaware of it, but just in case, I bring it up here). It's a model showing the direction of response to kin selection for a polygenic trait. It ends up finding conditions that are identical to Hamilton's. There are no individual loci in the model. You could consider it as phenotype-centered.

  9. I think that TSG was (and still is) influential on prospective biologists. For evolutionary biologists already in the field at the time (80's) it did not contain anything new, but for undergraduates, young researchers and even biologists in other fields I think it was very enlightening.

    As for the criticisms of Dawkins:
    -I have never met a biologist that misunderstood the analogy of the selfish gene. So IMHO this argument is exaggerated.
    -Larry criticizes Dawkins in his essay: "Evolution by Accident" that he does not take into account the randomness of evolution. I think he does "The fundamental units of natural selection, the basic things that survive or fail to survive, that form lineages of identical copies with occasional random mutations, are called replicators." - TSG
    or "non-random survival of randomly varying replicators" - RD
    He only emphasizes that after mutations are formed randomly, NS kicks in in a non-random way.
    -As for the gene centric view and the unit of selection question: I may be wrong, but I think the gene centric view is orthodoxy even in MLS theory. The discussion is about the possible levels (organelle, cell, organism, group, population, species) through which the survival of genes is influenced and not about whose survival is influenced.

    Criticisms of the gene centic view often include:
    -epigenetic inheritance - genetic imprinting
    -atomising genes and ignoring their complex interactions, gene expression and regulation pathways
    -interaction of phenotypes and the nuanced effects of multiple genes on them
    I think the gene centric view encompasses all these if we see genes as alleles in the gene pool interacting and cooperating with each other - as I think RD sees them.

    I think there are some valid criticisms though:
    -ignoring genetic drift and overemphasizing NS as the most important factor in evolution - Larry also talks about this in the same essay (altough in later years RD is mostly saying that NS is one important factor in evolution and is the only process leading to adaptation - which is less strict)
    -instead of the tree of life we should talk about the web of life in the non-eukaryotic domain - just to mention two.

    1. hey Gabor, it seems you really understand what The Selfish Gene and the Gene's-eye view of evolution is about (unlike Larry). You clearly recognize that most criticisms directed to the theory are simply wrong. However, you say there are some valid criticisms. I don't think the examples you give are valid:

      1. I think we have to separate the gene centric view of evolution from the emphasis Dawkins puts in natural selection and adaptation in his book. Both ideas are logically independent. We can perfectly envision evolution as the random sorting (drifting) of genes. Just give a value of cero to the selective coefficient of each gene and you end up with the 'Lucky Gene' theory.

      2. The web of life concept does not pose a challenge to the gene centric view. Instead it supports it. First, gene (replicator) replication always occurs the same way, independently of how their survival machines (hosts) evolve. So genes form bifurcating lineages, in contrast to cells. Second, the different mechanisms that have evolved during cell evolution (transduction, transformation, conjugation, GTAs, nanotubes) can be seen as evolved mechanisms for the propagation of genes. Plamids, transposable elements, introns, viruses (almost naked genes), even operons, and the like usually behave as selfish genetic elements (the ultimate genetic parasites). The web of life concept is inherently gene-centric. This concept shows how genes have found different strategies to maximize their survival, i.e., jumping between unrelated hosts. (I've been using Dawkinian metaphorical language)

      The issue of Dawkins' pan-adaptationism is different. I am not defending it here.

    2. I did not mean those as criticisms of the gene's eye view, rather of TSG in general and pan-adaptationism. Yes, I agree that the gene's eye view is perfectly compatible with the web of life concept.

  10. I certainly found it a highly influential work (though hate the term 'seminal' too much to ever actually use it). I was an undergrad in 1976 and it turned biology 'the right way round' as far as I was concerned. And it sparked an interest in evolutionary biology, not a part of my course.

    It's a book about Natural Selection. Criticising it for not being a book about drift seems somewhat unfair. Charges of 'panadaptationism' may be more apposite, though I think speculation on adaptive significance is legitimate, if unprovable without access to historic selection coefficients. But 'Spandrels' makes me gag, reading like some Party memo - "we fault the adaptationism programme ...".

    He has also dealt at length with what he means by 'Selfish' and 'Gene', and acknowledges in the intro his debt to Maynard Smith, Hamilton and Williams.If biologists have not adopted gene-centrism, it's *their* ideas that failed to stick, not his. But I don't think it's the case anyway; gene-centrism is simply one of the available means to view the process; I'm not aware it has been discarded.