Richard Buggs: Obsolete Dawkinsian evidence for evolution

Comments invited from @Joe_Felsenstein , @John_Harshman , and other mol phylogeny folks here.

[@moderators: added more informative title, and edited old title into post body.]


It seems to me to consist entirely of a sort of nitpicky rhetorical response to a short colloquial summary of phylogenetic evidence by Dawkins.

Wow. The argument needs to retire because Richard Dawkins (who is not a specialist in the relevant field) said something that can be interpreted as not 100% accurate about a technical topic?

Not a good article. Not untypical of other stuff we see from the DI, this time couched in feigned concern for “thousands of scientists working in the field of phylogenomics”. Yes I’m sure they’re all deeply hurt that Dawkins doesn’t blather on to a lay audience about software designed to handle incongruencies in phylogenetic analysis.

Come on dude.


I think the blog post is quite lacking.

The way it was written, I was expecting Buggs to go on and try to justify his claim that the incongruence that is present is too high to be compatible with Dawkins’ phrase “approximately the same tree”, but Buggs basically just says “there is incongruence” and then the blog post abruptly ends.

Of course there is incongruence, that’s why Dawkins added the caveat “approximately” rather than saying “every gene delivers the same tree of life”. If you want to claim that Dawkins is wrong, you have to get into the weeds and try to quantify “how much incongruence is too much”? I would contend that even the incongruence that is known to be present in many trees is still far more congruent than expected by chance. Often the incongruence represents a difficulty to narrow tree space down to a single tree from, say, 5 trees. The fact that there is enough still enough congruence to narrow it down to 5 trees from the many thousands or millions is already a strong signal.


Thanks for sharing @pnelson , and convey my regards to Richard.

That doesn’t seem to be what Buggs is arguing. It seems he is saying that the argument made by Dawkins, in that forum, was not technically correct.

Is he right?

Much of it depends on what we mean by " every gene delivers approximately the same tree of life". That approximately is important wiggle room for both Dawkins and Buggs. What we mean by this has a big impact.

The bigger question is about whether or not this evidence is evidence for common descent. Independent of what we mean by “approximately” it certainly is strong evidence for common descent. However, I do not think this is the more clear or helpful way to explain the genetic evidence, and frankly it seems pretty out dated.

So that does seem to be what he is arguing.

The argument needs to retire.

Perhaps Buggs is not even arguing that phylogenetic evidence for common descent needs to retire, it’s just Dawkins specific way of characterizing it that does? Do we have to pretend that’s what’s going on here? Buggs is just really concerned that Dawkins characterization can be interpreted as inaccurate and hurt the feelings of phylogeneticsts who want to know that dealing with incongruencies takes time and work?

Look, how stupid do we have to pretend to be here? Buggs is painfully obviously trying to undermine a specific argument for common descent: Consilience of independent phylogenies.

To try to do this, he gives a misleadingly cursory analysis of a sentence uttered by Richard Dawkins*, without even saying anything about the technicalities of the actual argument.

*The universal badguy, Satan, Beelzebub, blood-sworn lieutenant of Baphomet, the horned one, eater of babies, He who shall not be named, etc.


I think that Buggs is exaggerating needlessly and misleadingly. Specifically, at the end of the essay, he states:

“I hardly need to labour my point to the present audience. Dawkins’ statements are simply wrong. Gloriously and utterly wrong.”

Buggs seems to be saying that the claim of Dawkins that is being criticized, that there is general agreement in phylogenetic trees built using different genes, is “(g)loriously and utterly wrong.” Which in turn sounds to me like an assertion that there is little or no agreement at all between trees so built. As in, that such trees give totally different outcomes, with no agreement about the general structures, relationships between groups of organisms, or anything else.

Methinks some editing and explanation is needed.


This. As I recall, this very question came up here some time ago, and sustained a long thread. The problem is, if universal common descent (UCD from LUCA) predicts “signal + noise,” then that represents arguably all the data. “Too much” can never arise as a disconfirming threshold, because incongruence can always be dumped into the “just more noise” bin.

Some biologists are beginning to say that their “too much” threshold to disconfirm UCD from LUCA has been reached, and wonder if the origin of the cell (for instance) occurred multiple times independently:

1 Like

There is such a thing as statistical significance. That’s how signal and noise are handled throughout the sciences.

Did Dawkins give a detailed explanation of statistical methods used to determine if there is a significant phylogenetic signal in the data? No. Most scientists don’t go into statistics when they talk to the lay public about science. Whether that’s a good thing is a discussion worth having.

Here is a random regression analysis I found online:


Only a few of the data points fall on the regression line. There are incongruencies in the data. Does this mean there is not a statistically significant correlation in the data? Of course not.

When Buggs claims that phylogenetic evidence can’t be used because their are incongruencies he is in essence arguing that we shouldn’t use statistics in science. That’s absurd.


Are dogs larger than cats? Many people have claimed this, but I compared my neighbor’s chihuahua with another neighbor’s Maine coon cat, and found to my astonishment that the cat was larger than the dog. This assertion about cats and dogs is gloriously wrong and needs to be retired!


Yes. But I repeat evograd’s comment: “…you have to get into the weeds and quantify ‘how much incongruence is too much?’”

Statistics is a collection of mathematical tools to analyze data, which you can be certain Buggs uses. But statistics does not (indeed, cannot) answer the question posed by evograd.

Evograd’s question is a biological, not a mathematical or statistical, puzzle. Saying “well, of course there will be some noise in the data” is not an answer, but pushes the puzzle down the road for someone else to solve.

I don’t think any professionals in the actual field care about anything Dawkins has to say, much less a blog post about what someone else thinks about what he has to say. And I doubt Dawkins much cares about such a blog post either. So the article itself is pointless. And is the author really writing about something Dawkins said 11 years ago?

Ultimately, I think that this article is at least as misleading as what Dawkins said in the first place while bringing unwarranted attention to something from over a decade ago said by someone irrelevant to the actual field. The author should be embarrassed to have published it.


The only thing I have to add to this accurate description of this misleading blog post is this: those who understand the science here know that the background here is not evolution, not even slightly. The way to understand this piece is to understand the pathetic fixation of numerous Christian scholars—and especially of zillions of Christian laypeople—on Richard Dawkins. We’ve seen it here on PS. It’s rampant at BioLogos. It’s common. And it’s pathetic.


Self-correcting science in action. No one cares, Dawkins – say whatever you please.

Re: the linear plot passing through points. Typical fakery by evolutionists. T aquaticus has made the points in the plot large enough that about 5 of them touch the line. If they were smaller not one of them would be on the line. So this whole plot is faking linearity!


What is? The blog post? No, that’s not science at all.

This is unintelligible. Try again.


This seems confused, because the authors there are not denying common descent that I can see. They are are wondering whether LUCA, the common ancestor of bacteria and archaea, was a cell.

They seem to go the opposite direction you are invoking them to argue, and they suggest instead the possibility of the monophyly of all biological systems. A FUCA, first universal common ancestor from which all the molecules known in all biological systems ultimately descent.

First, we have seen that the presence of ribosomes in cellular structures should not
be used as evidence of a common origin for all cells. On the contrary, it must be seen as
evidence for the monophyly of all biological systems, including progenotes, viruses,
bacteria and archaea. This differentiation is absolutely necessary to better understand the
natural history of cells and differentiate it from the natural history of other biological
systems. Following this reasoning and reinterpreting the classification originally
proposed by Woese et al. (1990), we conclude that these authors were verifying a
common origin for the ribosomal molecule and the translation system; but not for cells.
Thus, it is indeed possible to assume that LUCA may have never existed as an ancestor
for all cells under the form of an already cellular organism (Koga et al. 1998).
Alternatively, the common ancestor of Bacteria and Archaea may have had the form of a
community of open biological systems sharing a translation machinery, reinforcing the
monophyly amongst all biological systems and ascendance from a progenote era. FUCA
was clearly a first ancestor for entities sharing a chemical and informational code, i. e.,
those ones on in which the translation system has emerged and matured through the selforganization of an organic code (Farias et al. 2020).


Hence, when one takes seriously the possibility that cells could be polyphyletic,
it becomes necessary to understand that the natural history of cells is different from the
natural history of biological systems. In that sense, we use the term biological systems to
refer to any system made of chemical molecules with the capacity to form organic codes
(Farias et al. 2020), as the genetic code on which amino acids are encoded by codons in
a nucleic acid. Those biological systems therefore include not only cells and viruses, seen
as different contemporary strategies for maintaining organicity (Farias et al. 2019), but
also the pre-biotic open systems known as progenotes. And even though these life
enduring strategies may have most likely emerged more than once, they have in common
the fact that they are based on the same molecular language, since the organic
phenomenon presents strong evidence that appeared only once on our planet. Thus, we
presented earlier the proposal that biological systems emerged from a chemical symbiosis
(Prosdocimi et al. 2020) that started with an interaction between nucleic acids and amino
acids. This interaction stabilized and increased the chance that these heteromolecular
systems could be maintained and persisted. From this starting point, the system has
probably evolved until the maturation of a ribonucleoprotein machinery that gave rise to
the ribosome, generating the first universal common ancestor (FUCA) (Prosdocimi et al.
2019). FUCA therefore was composed by a population of open-systems, self-replicating
ribonucleoproteins that inaugurated the progenote era and evolved into maturity when
self-organization processes resulted in the creation of a genetic code.

I mean, Paul, did you even read that one?


First off, it oozes smarm and patently false regard for Dawkins, which I find off-putting. Second, while Dawkins’s claims are not true, they are still fairly close to true, much closer than Buggs’s characterization of the field, a field that in my understanding he rejects. Horizontal transfer is rare in eukaryotes. Incomplete lineage sorting is more common, but is not a major factor in any tree with many taxa, affecting a few nodes at most. Molecular convergence other than convergence in base composition is vanishingly rare. The main problem in phylogenetics is that many branches are very short, so not much happens along them, and that many branches are very long, so much that happened earlier is over-written; old short branches followed by long branches are hard to resolve.

Briefly: Dawkins, though wrong, is closer to right than Buggs.


No Paul.

  1. Ten years ago …

  2. … Richard Dawkins said something outside his field of expertise …

  3. … on a Youtube video …

  4. … that might not have been entirely correct.

The question is, does the scientific community have the time to police everything that any scientist has ever said on Youtube? Or would they be better off concentrating on policing things like primary journal articles and textbooks – things that actually have a substantive impact on the way that science is propagated.


A quibble, perhaps, but statistics CAN be used to evaluate incongruence, lack-of-fit (or over-fit), agreement, and disagreement. Building on the linear regression example, looking at the residuals can tell you a lot about what your model may be missing.

image source

And we see that missing pieces have been found (HGT, indels, etc…). I think we could say that Dawkins’ statements are naive in the sense there is much more to the story. What is missing from Buggs’ editorial is any alternative to the claim that the evidence supports evolution.

Evograd’s question is a biological, not a mathematical or statistical, puzzle. Saying “well, of course there will be some noise in the data” is not an answer, but pushes the puzzle down the road for someone else to solve.

This is always true of any analysis, except maybe the simplest of controlled experiments. The interesting question is not whether the answer is right or wrong, but rather “How wrong is it?”


A population of “ribonucleoproteins” is not a cell. Not an ancestor of cells, in other words, at least in any biological sense we can identify.

When you can show me what “open-system, self-replicating ribonucleoproteins” are, in the absence of cells, then I’ll take the notion of a population of them seriously.

The serious part of their paper is their critical analysis of LUCA. Never existed, as far as they can see. The cellular state arose independently multiple times.