Scientists…in disagreement!

Yesterday, I reposted an article on homology within the neck and shoulder, which describes an interesting technique of using patterns of gene expression to identify homologous cellular pools; the idea is that we can discern homology more clearly by looking more closely at the molecular mechanisms, rather than focusing on final morphology and tissue derivation. Trust me, if you don’t want to read it all—it’s cool stuff, and one of the interesting points they make is that they’ve traced the fate of a particular bone not found in us mammals, but common in our pre-synapsid ancestors, the cleithrum. They argue from a common cellular origin that this bone has been reshaped into a ridge on our shoulder blade, the scapular spine.

As many readers might know, though, the word “homology,” especially when coupled with a novel technique for its determination, is always good for an argument. This one is no exception.

I won’t get too much into the details of the disagreement, but I just want to point out a very common phenomenon: every scientist is presenting ideas and evidence for hypotheses, and it is expected that other scientists will respond by trying to argue against them. It’s what makes for a healthy science, that there is vigorous debate. Sometimes this goes on in private communications back and forth, sometimes it occurs at meetings, and sometimes it works its way into print, and you can find the argument enshrined in the pages of the science journals.

Here’s a good example: after the publication of the Matsuoka et al. paper that I described, Sánchez-Villagra and Maier wrote a three-page rebuttal and published it in Evolution and Development. They make a strong case that perhaps the comparative descriptions are oversimplified, minimizing the great variation found in musculoskeletal organization, and that fossil and phylogenetic evidence shows an absence of both the cleithrum and the scapular spine, making the homology dubious.

It is possible that Matsuoka et al. (2005) have found ev-
idence that the population of cells that give rise to parts of the
scapular spine of therian mammals is homologous to the
populations of cells that give rise to the cleithrum of basal
gnathostomes. But homology at the cellular level does not
necessarily imply homology at higher levels of the hierarchy. So the cleithrum does not “survive as the scapular spine in living mammals” (Matsuoka
et al. 2005, p. 347)—it has been extinct in synapsids for more
than 250 million years! If no traditional homology of the
cleithrum was challenged by Matsuoka et al. (2005), then this
needs to be explicitly stated, otherwise the readers are misled.

Ahlberg and Koentges have their own reply to the rebuttal. It’s also three pages long, and is rich with detail and ideas—it’s also good science. Here’s a short excerpt:

S.-V. & M.’s point
that the cleithrum has been extinct in synapsids for more
than 250 million years and cannot “survive” as the scapular
spine is moot. We are not claiming that the scapular spine is
a remnant of the cleithrum ossification, but we do infer that
it is made from the same cell populations that once made the
leading edges of that cleithrum ossification. At that inter-
section between two seemingly contradictory homology
statements emerges a mechanistic explanation of evolutionary change: the cleithrum has been lost through the re-specification of cell fate within a stable cell population map. We
provide evidence for genes involved in imparting such skeleton-muscular connectivity and find that these genes (part of
the Hox gene network) have been expressed in neural crest

I’m not going to resolve the argument here, of course. I will say that Sánchez-Villagra and Maier are making a perfectly reasonable argument from structure, while Ahlberg and Koentges are making a perfectly reasonable argument from the developmental process…and my personal bias is to find the logic of how a structure develops to be more persuasive than what it ends up looking like, so I’m inclined to favor the Ahlberg and Koentges explanation. What I look forward to, though, is the continuing exploration of the problem, and the appearance of further data to support one or the other position…or something else altogether. This is how science works.

One other thing: both papers have pleasant acknowledgments.

We thank Per Ahlberg and G. Koentges for their prompt and in-
formative response to our communication with them, leading to this
paper, and for generously engaging on an open discussion about the
issues at hand. We also thank S. Kuratani for discussion of ideas.

We thank Marcelo Sánchez-Villagra and Wolfgang Maier for their open and collegial approach to this discussion.

I just wanted to bring this up to illustrate a few important points. When I summarize a paper here, I’m often simplifying and skipping over colossal amounts of deep background—these explanations are not definitive, final answers! Also, if you want to see some dissenting opinions on that last paper, now you have some sources to trace down. And finally, much as my bias in evolution leans towards the process side, I have the same bias in valuing science: the important part isn’t so much the final answer, but how you get to it.

Sánchez-Villagra MR, Maier W (2006) Homologies of the mammalian shoulder girdle: a response to Matsuoka et al. Evolution and Development 8(2):113-115.

Ahlberg P, Koentges G (2006) Homologies and cell populations: a response to Sánchez-Villagra and Maier. Evolution and Development 8(2):116-118.


  1. Great White Wonder says

    I have the same bias in valuing science: the important part isn’t so much the final answer, but how you get to it.


    Compare the little argument you described above with the debacle of the bogus ivory-billed woodpecker rediscovery.

  2. DragonScholar says

    Definitely a good example of intelligent scientific debate – and civil debate as well. I like seeing people argue with the goal of understanding first – winning doesn’t come into the picture so much as a good conclusion.

  3. Gary says

    I agree with DragonScholar.

    I once attended a talk by a mathematician (sorry, can’t remember his name) who
    made the case that the most important reason for teaching students science is not because it is useful to know this stuff or because it is beautiful or to create an informed citizenry, but because it teaches people how to argue vigorously, but with civility, from evidence, and in pursuit of answers we can agree on.

    For those of us who teach, it is probably worth highlighting examples like this. Thanks for pointing it out.

  4. says

    I have the same bias in valuing science: the important part isn’t so much the final answer, but how you get to it.

    Ouch. A biologist can afford to say that, because you guys do find final answers with some regularity. But in my field, archaeology, the same argument is used to defend speculative activities (I won’t call it research) that do not aim to find any truthful answers: just to “enrich the source material with new interpretations”. And this is an attitude that I simply loathe. Far too many untestable and sesquipedalian hypotheses are published in archaeology.

    So I think the important thing is to find final answers, and that the only way of checking whether they really are final is to use the right methods.

  5. Ian Menzies says

    So I think the important thing is to find final answers, and that the only way of checking whether they really are final is to use the right methods.

    We can never know when we have the final answer. But if we use the right method than we can be confident that our current answers are (on average) less wrong than the answers we had before. Thus the focus must be on the journey (and ensuring the correct path is followed at all times) not on the destination.

  6. Jeff Chamberlain says

    I’d like some more discussion about the business of “getting answers” versus “using the right process” for scientific inquiries. It’s easy to understand that “answers” in science are not “final.” But Mr. Rundkvist’s comment is also important. I’d like to think that we’ve done considerably better than “less wrong,” at least sometimes. Would there really be much point to it if we didn’t reach some answers which, if not “final” are at least “secure” (or, dare I say it, “true”)?

  7. says

    Very strictly speaking, all scientific knowledge is of course provisional to some extent. But in practice, we accept many hypotheses as proven truth once they’ve survived repeated testing from different angles using different methods and datasets. We’re not swimming in some vague post-modern sea of unknowables and subjectivity.

    The problem over at the Arts faculty where I am is that we have neighbours in lit-crit and art-crit whose view of research is “to comment on works of art in an educated way”. And that attitude leaks into archaeology and history as well. Our source materials (texts, artefacts) are in fact works of art. But in my opinion, they primarily provide scientific data allowing us to know something about what happened to people long ago.

  8. says

    Jeff Chamberlain: The solution to your puzzle is, as far as I can tell, to admit that truth comes in degrees. (Short argument to this effect: truth is the dual to error, and any scientist knows error comes in degrees) Subsequently the goal of science would be to long term increase the truth values of our hypotheses. This is not the only goal of science. (Argument: one can write a computer program that prints “n = n” for every n from zero until some really huge value – those statements are all true, but the finding of them is useless.) The others include (at least) systematicity – we want our hypotheses to be organizable into systems (i.e. theories) and depth, which is hard to characterize, but one can start by noting the identities I just mentioned are not very deep.