Sunday, January 28, 2007

The Richard Dawkins Definition of a Gene Is Seriously Flawed

(This is an updated version of an article that I originally posted to on Sept. 6, 1999)

We are interested in the correct definition of a "gene" (see ...). Part of the confusion is due to popular science writers who don't get it right. For example, Richard Dawkins does some serious handwaving in The Selfish Gene and he compounds it in The Extended Phenotype.

Dawkins knows that his defintion of "gene" in the Selfish Gene is unusual so he returns to the subject in The Extended Phenotype in his discussion of the selfish replicator. Dawkins is forced to concede that his use of the word "gene" is incorrect. That's why he says,
I am happy to replace 'gene' with 'genetic replicator where there is any doubt.
Nevertheless, he tries very hard to defend his point of view by claiming that geneticists and molecular biologists can't come up with a good definition of gene either. This leads him to make some very silly statements about genes and cistrons. He defines his genetic replicators in terms of alleles which means that they don't exist unless there is variation in the genome. He then goes on to restrict his discussion of changes in frequency to the results of natural selection, which means that his "genes" are effectively defined by the mechanism he prefers. This is why he quotes George Williams,
This is the rationale behind Williams's definition: 'In evolutionary theory, a gene could be defined as any hereditary information for which there is a favorable or unfavorable selection bias equal to several or many times its rate of endogenous change.'
                ....The Extended Phenotype p.89
The hand-waving in The Selfish Gene is even more obvious,
In the title of this book the word gene means not a single cistron but something more subtle. My definition will not be to everyone's taste, but there is no universally agreed definition of a gene. Even if there were, there is nothing sacred about definitions. We can define a word how we like for our own purposes, provided we do so clearly and unambiguously. The definition I want to use comes from G.C. Williams. A gene is defined as any portion of chromosomal material that potentially lasts for enough generations to serve as a unit of natural selection.
                ....The Selfish Gene p.28
In the new version of The Selfish Gene (1989) Dawkins adds a footnote where he again addresses his critics, especially Sewall Wright. Dawkins defends his definition of a gene as a unit of selection.

More handwaving,
I am using the word gene to mean a genetic unit that is small enough to last for a number of generations and to be distributed around in many copies. ... The more likely a length of chromosome is to be split by crossing-over, or altered by mutations of various kinds, the less it qualifies to be called a gene in the sense I am using the term.
                ....The Selfish Gene (1989) p.32
I said that I preferred to think of the gene as the fundamental unit of natural selection, and therefore the fundamental unit of self-interest. What I have now done is to *define* [Dawkins' emphasis] the gene in such a way that I cannot really help being right!
                ....The Selfish Gene (1989) p.32
The fact that Dawkins uses the word "gene" in such a non-standard way is not an issue as long as one recognizes that the Dawkins "gene" has nothing to do with the genes that molecular biologists and geneticists talk about. It's not an issue as long as one doesn't try and pretend that Dawkins has avoided handwaving and "clearly" refuted the problems raised by his critics.

The most reasonable definition of gene is that it is a piece of DNA that is transcribed but there are exceptions to everything in biology. Some genes are made of RNA, for example, and sometimes it's better to define a gene in terms of the protein it encodes. In no case is it reasonable to define a gene in terms of its ability to be selected or whether recombination can occur within it.


  1. Those quotes from TSG are about where I stopped reading it. I'm glad to see I'm not the only one irritated by Dawkins' circular definition!

  2. Larry, why is it that you, who know a great deal about biology, can see the flaws in Dawkins' definition of a gene, yet when he gives equally flawed and simplicistic definitions of "faith" or "religion" and is called out by people who know about these subjects, you denounce them as "appeasers"?

  3. Because that's not what Dawkins and I mean by "appeasers." Appeasers are people who agree with us that religion is superstition but who bend over backwards to avoid offending the so-called theistic evolutionists.

    Apparently it's okay to criticize the IDiots for believing silly things but we mustn't ever criticize the likes of Francis Collins or Ken Miller for believing equally silly things. The appeasers think that people like Collins and Miller are really on the same side as the atheists. Isn't that strange?

    Dawkins is attacking belief in the supernatural, especially the supernatural being known as God. I don't have a problem with his definition of theism or religion. Anyone who believes in the existence of supernatural beings is fair game for Dawkins.

    Who are these people that "know about these subjects" and what is their word for believers in supernatural beings?

  4. Is there a better definition for what Dawkins misleadingly refers to as a "gene"?

  5. For some of the problems in defining gene, take a look at Scott Gilbert's site:

  6. Gilbert's site is a good example of why the geneticist's standard definitions are not very useful for an evolutionary biologist using a "gene"-centered framework.

    What any exposition such as that in the Selfish Gene requires is a name for a "chunk" of genetic material - preferably contiguous, although that doesn't absolutely have to be true if we don't mind extra complexity. The function of the "gene" - whether transcribed or not - is absolutely irrelevant, except as a means of conferring competitive (dis)advantage.

    I was initially annoyed with Dawkins on this too, I'd rather he had coined a new, short term for "genetic replicator". But "gene", as a "small enough" piece of genetic material, works just fine - and conveniently makes it simple to translate "dawkinsgene"-based arguments into "othergene"-based arguments, which is more interesting for most readers. (As an example, talking about a gene conferring increased resistance to some disease... a rather conventional sense of the gene, but in Dawkins' case this "gene" doesn't have to be an entire transcribed unit or control region, but a portion - just 15 nucleotides crucial to a binding site for instance.) This makes Dawkins' "gene" more robust for a discussion of evolutionary biology than that used when we are talking about genes in other contexts.

    A couple of other small points:
    "The most reasonable definition of gene is that..."
    Too bad this definition, even in this little blog entry, is laden with exceptions. Dawkins' definition doesn't require any exceptions to be used in discussions of evolutionary biology.

    And: claiming that by using "alleles" Dawkins restricted his definition to cases where there is variation is ... not consistent with what I read, though my memory may be at fault. Even if true, it would be a fine definition since, absent variation, there is no reason whatsoever for an evolutionary biologist to talk about the "gene".

  7. I think it all depends on who you're talking to. That and the fact that the word was invented before we knew what DNA did.

    For palaeontologists and evolutionary biologists a gene is a chunk of genetic information with a phenotypic effect. The internal structure of Hox genes isn't particularly interesting to me, what is interesting is how many of them there are, where they are expressed and how they are shared across phyla. I would lump the regulatory region in with the rest of the gene as a mutation here affects the gene's expression (and I wouldn't describe the regulatory region as a gene in its own right).

    A behavioural ecologist could make do with something approaching Dawkin's definition of a gene. By the time a gene's effcts have made it to the outside world it doesn't matter if the change that distinguishes it and an allele is in the intron or the regulatory region.

    To someone looking more closely this glosses over a whole world of complexity, sections of DNA coding for more than one product, non-transcribed bits affecting transcribed bits, products made form apparantley unrelated sections of DNA.

    For my part, (on Pharyngula) Peter Ellis's definiton of: "A gene is a unit of DNA which is transcribed, has a phenotypic effect when transcribed, and whose phenotypic effect is altered by mutation" is the best one for the interested (or disinterested) layperson to start with. If they then start to ask questions about how exactly these "gene" things work the detail comes in.

  8. "What is a gene?"

    Now this is a very thorny subject.

    I don't think there is a good definition.

    Several facts to consider:

    We talk about gene copy numbers. Are they really the same gene? They encode for the same protein but have slightly different position and regulation. Any biologist will acknowledge that they are different DNA segments, but we seem to perceive them as the same gene. Similarly it is really difficult to describe genes without its cis elements. But in eukaryotes these can be several kb apart and can be shared by multiple genes.Implying there is a many to many relationship between fragments of the chromosome and the genes Different DNA/Same Gene implies a gene is NOT a DNA fragment

    How about transposons? They are often called jumping genes. And in Dawkins terms they are definitely genes, but they don't necessarily encode a protein product ( use other transposon products). This rules out the central dogma definition, a protein product is a very effective and the most common mechanism of survival for a gene, but it is not the only one.

    How about viral genes? Larry already noted that some of them are formed of RNA. But there is more than that: Some of them have a life cycle that encompasses both DNA and RNA forms. You can make similar arguments when talking about genes during the course of evolution. Presence in multiple physical forms across time imply abstract and continuant properties

    There is actually an ontological explanation. What is preserved is the information not the physical entity.
    A gene would be classified as a "schema" by the top level categories:

    I think in that sense Dawkins' definition makes much more sense than "Gene is a DNA fragment...more or less" definition. Maybe we, biologists, should clear our own backyard before criticizing other people for hand waving,

  9. Dawkins definition is more based on function than on physical structure. While the criticism of Dawkins definition seem to be based on that the definition cannot be translated to a given physical structure, there are no rules that say this is more or less flawed than other definitions.

    Dawkins functional definition of a gene fits very well with how the term is used in most media and everyday speech.

  10. You make a pretty interesting point, I am no chemist or biologist, but I am aware that a significant number of knowledgable scientists take issue with Dawkins, I take issue with some of his work on philosophical as well scientific grounds.

    I think that irrespective of ones personal theological beliefs Dawkins needs to be challenged and challenged very seriously.

    As a Prof. of the “Public Understanding of Science” he not only frequently presents unsound arguments, he is actually misrepresenting science to his (often uncritical) audience.

    By applying the term “scientific” to his arguments, he implies a certain credibility that in actuality is often unjustified.

    My own growing irritation over these matters has led me to devote some time to a new blog:

    I will strive to not personally debate too much in this blog, but rather catalog and discuss what I see as serious weaknesses in his arguments on a range of topics.

    I have not placed any restriction on who may comment either.

    I'm thinking I will mention the topic you have raised next, and link back to your blog here.

  11. Larry, thank you for your valuable lesson in how different uses of the word "gene" may or may not be justified by the user, on several levels.

    With regard to your argument on Dawkins and "appeasers", I wish that you would reconsider.

    The reason I do so is that some "appeasers" merely object ot false statements used in argument, due perhaps to Dawkins' refusal to actually study what he talks about.."it's a non-subject"; not worthy of study.

    As an "escaped inculcee", I resent his false statements, as I see them hindering others who are attempting to escape the clutch of religion. If they turn to THE CHAIR for correct information, and find equivocation, oversimplification, and poor logic all round, I think it's NOT helpful at all.
    Those who find error should not be called appeasers, out of hand - merely because some see them as supporting a "side", rather than supporting accuracy in reporting.
    I can point out many errors in logic that Dawkins uses against religion, never mind that he never has shown evidence for Penguins selfishly pushing others as a test for predators. Real scientists in the Antarctic say it doesn't happen - and yet it's part of his premise for the book.

    What i'm getting at is that oppponents of Dawkins anti- religion proselytizing from The Chair can be right sometimes and don't deserve the label applied, and when right, they should be supported

    Also thanks to Dave Godfrey for his comment on definition of "The Gene".

  12. Larry, after consideration of your points made, a layman might wonder if there is any easy way to understand what Dawkins has done:

    Dawkins' "Gene" is actually "gene plus regulators, and any epigentic effects" - in other words, what some people call "the genetics"

    Are the regulators and epigenetic factors included in "Genotype", or in "Genome" ?

    It seems not. But why not ?

    Because Dawkins is fudging ?

    Is this the correct answer (or a correct way to question ? )

  13. Karen Armstrong
    "God is the God beyond God"

    Richard Dawkins
    "The gene is the gene beyond the gene"

  14. I could not agree more. I think the most succinct definition of the word gene is from Wikipedia: “A gene is a sequence of DNA or RNA which codes for a molecule that has a function”. Dawkins has, rather obviously, chosen a definition to fit his preconceptions and created an edifice of circularity. He says “We can define a word how we like for our own purposes.” Well no we can’t, definition is central to scientific thought. Dawkins may mean something when he talks about “genes” but whatever he means it is not a gene in any real sense. He never makes clear in real, molecular, terms what he is wittering on about when he talks about a gene, a non-trivial problem for a book called “The Selfish GENE”.