Luskin's Hopeless Monster

I’m checking in from the airport on the way back from Evolution 2013. For me, highlights of the meeting included presenting my BioGeoBEARS R package and some Ph.D. results at the Ernst Mayr Symposium, hearing about all the cool things going at NIMBioS, anticipating and thus having a seat in the room while observing the Felsenstein Effect, and meeting Jerry Coyne in person for the first time, and having a friendly conversation rather than an argument. (What will our respective readers think of us? We have reputations to uphold!)

Part of the reason for harmony was that Jerry recently blogged such nice things about my review of the half-baked ID book Darwin’s Doubt: The Explosive Origin of Animal Life and the Case for Intelligent Design. Other folks, however, have not been so positive. Rather than actually defending Meyer’s book from my quite specific criticisms, the Discovery Institute’s Casey Luskin is now pretty much claiming – and the various ID fans out there are blindly, uncritically repeating – that I haven’t read the book, that I wrote most of my review before the book came out, and that I made up quotes of Meyer.

Well, here’s the reality. I did not have an advance copy or pre-write a review or anything. I got the book around lunchtime last Tuesday when it came out. I read it during lunch, then again for snippets of the afternoon (we computational biologists often have bits of downtime while we wait for programs to run), and then most of the rest of it that night and the next morning. The book was not impressive, and I resolved to not bother with a review, and to work on stuff I should be doing. However, when I got into work on Wednesday, I started seeing the fawning, so-innocent-of-the-problems-and-the-science-it-was-almost-cute positive reviews of the book coming out from ID creationists, and I realized that the best way to stop getting distracted would be to bang out a review. I spent most of Wednesday on it and put the review up that night. I felt quite guilty, really, putting even that much time into it, considering everything else I should be doing, but like I said, it was much easier to focus afterwards.

For people who find this all surprising, what can I say? You must be slow readers. More seriously, folks, it’s not like this is my first rodeo. It might help to remember that I spent 3 years at NCSE researching the ID movement and basically crawling inside their heads, and then 6 years in graduate school studying, and TAing, and publishing phylogenetics. Unlike most scientists, I am deeply familiar with the ID arguments, their weird vague question-begging definitions of crucial terms and premises in their argument (“information”, “fundamentally new” whatevers, etc.), and so I don’t have to spend a lot of time mentally unravelling the multiple levels of confusion and misunderstanding and wishful thinking that are going on whenever Meyer rehashes some oft-used, previously refuted ID talking point. I can focus on what little is new and unique to the book in question – in the case of Meyer’s book, this is basically the stuff about the Cambrian and phylogenetics.

(Plus I’ve been keeping up on the Cambrian literature for years – e.g. my “Down with Phyla!” posts are crucial reading if you did not understand what I was talking about in the Meyer review. Heck, I personally know Charles Marshall of Marshall (2006) and Jim Valentine of Valentine (2004) and Erwin and Valentine (2013) – they are professors in my department! – and I read all of these references when they came out.

For what it’s worth, I should say briefly that Erwin and Valentine (2013), while a capable review of the topic of the Cambrian Explosion, has some significant weaknesses in the realm of phylogenetics and taxonomy. The authors work hard to include up-to-date phylogenetic thinking and terminology, and do quite well compared to their previous works, but the book nevertheless still carries a lot of stage #2, Linnaean, ranked-taxonomy thinking within it. This makes sense considering that Valentine was trained in the 1960s and Erwin, I believe, in the 1980s, and that both are in invertebrate paleontology, which for various reasons has hung on to ranks-based analysis longer than most other subfields. However, it causes various internal contradictions in their work. But I digress.)

As for the claim of fake quotes, in all cases, Luskin is just sloppily misreading. It is quite clear from context when I am actually quoting Meyer, and when I am using scare quotes to highlight a term or concept that I think is problematic and/or mistaken, or using a paraphrase marked with quotes (or sometimes dashes, although-this-gets-unwieldy-quite-quickly) to efficiently summarize a difficult-to-describe position. There are a lot of weird and obscure positions in play at the intersection of the Cambrian, systematics, and ID creationism, so sometimes this is necessary, at least when, as now, I don’t have time to spend paragraphs explaining all the basics from scratch.

Other stuff

In addition to trying to discredit my review through well-poisoning based on information-free speculation about my reading and writing practices, Luskin tries a few substantive arguments. These don’t go well, and just further demonstrate just how throughly Luskin and Meyer are misunderstanding the basic terminology and concepts and evidence necessary to even have a meaningful discussion of the Cambrian. (I say Luskin and Meyer, since Luskin says he was Meyer’s research assistant on the book.)

To wit:

Matzke does attempt to address the first problem posed by the Cambrian explosion. He does so by claiming that methods of phylogenetic reconstruction can establish the existence of Precambrian ancestral and intermediate forms – an unfolding of animal complexity that the fossil record does not document.

Well, no. I claimed that phylogenetic methods can establish, and have established, the existence of Cambrian intermediate forms, which are collateral ancestors of various prominent living phyla. The case is clearest with the most common and most-fossilized Cambrian phylum, the arthropods, but there is a fair bit of similar evidence for other major phyla. (Some phyla, primarily soft-bodied worms, have few fossils anyway, and there of course intermediate fossils are scare, although even if we had them they would be difficult-to-identify worms.) All of the leading authorities (Valentine, Erwin, Conway Morris, Briggs, Budd, etc.) would agree with me. More precisely, I agree with them, and they have all said in print what I just said. Furthermore, they would all agree that this is extremely important evidence for understanding the origin of “phyla”, evidence which cannot be ignored. But Meyer/Luskin ignore it, instead occupying themselves with hunting around in the Precambrian.

Similarly, Valentine, Erwin, Conway Morris, Briggs, Budd, etc., would all agree that it is utterly impossible to have a sensible discussion of the Cambrian Explosion while ignoring the 30-million year sequence of surface-crawling worms, then burrowing worms, then armored worms, then small shellies, THEN identifiable relatives of phyla, most of which are (not coincidentally) stem groups rather than members of the crown phyla, and which have characters suites transitional between the major crown phyla. These are fatal, catastrophic omissions Meyer’s book, which is allegedly supposed to be a serious commentary on the Cambrian Explosion. The only way forward for the IDists is to forthrightly admit the error to the books’ readership. From there, they could perhaps try to maintain their argument by arguing that the 30-million-year worms-shellies-stem-groups sequence is irrelevant, and that the stem group fossils with transitional morphologies are irrelevant or have been misinterpreted by the experts or something. But they haven’t got a chance in heck of convincing anyone serious as long as they pretend to their readers that these data don’t exist.

Luskin also says:

Though he accuses Meyer of being ignorant of these phylogenetic methods and studies, he seems unaware that Meyer explains and critiques attempts to reconstruct phylogenetic trees based upon the comparisons of anatomical and genetic characters in his fifth and sixth chapters.

Now who’s not reading? I explicitly devoted a section of my review to Meyer’s discussion of phylogenetic conflict, and made a list of points that any professional, serious discussion of phylogenetic conflict would have to address, which Meyer did not address. (Luskin later contradicts himself and refers to my critique of Meyer’s claims about phylogenetic conflict, but he mostly just asserts Meyer’s book is correct. I suspect Luskin did a lot of the quote-mining for the phylogenetic conflict section. Earth to Luskin: do some statistics to back up your assertions, or you and Meyer aren’t worth listening to on the topic of phylogenetic conflict.)

Things get worse with Luskin’s discussion of “phylum” lobopods and Anomalocaris as an “arthropod”.

In the first quote, from page 53, we see that Meyer called Anomalocaris “either arthropods or creatures closely related to them,” showing his awareness that there is ambiguity and debate over whether Anomalocaris belongs directly within arthropods, or was a close relative. Matzke never quotes Meyer’s statement on this point, which is consistent both with what Matzke says about anomalocaridids, and with the relevant scientific literature. Instead, Matzke seems unfamiliar with what Meyer actually wrote.

In the second quote, from page 60, Meyer suggests that Anomalocaris may in fact be an arthropod. Would it be a “basic error” to make that claim? Not at all, because many leading authorities on the Cambrian explosion have suggested precisely the same thing –that Anomalocaris is an arthropod.

This just further demonstrates the epic-level misunderstandings that Luskin and Meyer have when it comes to phylogenetics, systematics, and the Cambrian. You cannot even discuss this question without specifying what various authorities mean by “arthropod”, which Meyer never does. The most common meaning of “arthropod” today is “crown group arthropod”. This is what is used by e.g. Erwin and Valentine 2013, as well as all the other authorities I cited. On this definition, Anomalocaris is clearly outside of arthropods. Now, some scientists, usually those slightly less hip with phylogenetic systematics, use the term “arthropod” to refer to anything in the crown or on the arthropod stem. On this definition, Anomalocaris is an “arthropod”, but all of these people would also agree that Anomalocaris is not in the arthropod crown group.

This is the crucial point – you cannot just say “Anomalocaris is an arthropod”, flat-out, without specifying what you mean by “arthropod” and what the authorities you are citing mean. It’s clear enough to experts, usually, what various scientists at various times mean (for example, I know Thomas Cavalier-Smith is an old-school evolutionary systematist, but Luskin, who cites him, doesn’t), but in any book for a general audience, this must be specified. Erwin & Valentine do it capably, what the heck is Meyer & Luskin’s problem?

We can see Luskin’s misunderstanding further when he quotes Paterson et al. (2011):

These fossils also provide compelling evidence for the arthropod affinities of anomalocaridids, [and] push the origin of compound eyes deeper down the arthropod stem lineage.

“Arthropod affinities” and “arthropod stem lineage” do not mean “Anomalocaris=arthropod” – they mean Anomalocaris is on the arthropod stem! Which is a common finding, well-understood to everyone in the field.

Luskin comments further:

The paper firmly places anomalocaridids as stem-group arthropods, very close to the crown-group arthropods, and has some weighty co-authors, including John R. Paterson of the University of New England in Australia, Diego C. García-Bellido of the Instituto de Geociencias in Spain, Michael S. Y. Lee of South Australian Museum and the University of Adelaide, Glenn A. Brock of Macquarie University, James B. Jago of the University of South Australia, and Gregory D. Edgecombe of the Natural History Museum in London. In covering this paper, Discover Magazine stated: “Paterson also argues that the eyes confirm that Anomalocaris was an early arthropod, for this is the only group with compound eyes.”

Another way to say “The paper firmly places anomalocaridids as stem-group arthropods” is to say “The paper firmly places anomalocaridids outside of crown-group arthropods, i.e. outside of what most people, and all general readers, are thinking of when you say ‘arthropod’.” What this paper actually does, phylogenetically, is provide some characters (compound eyes) that strengthen the evidence for Anomalocaris being on the arthropod stem, rather than the onychophoran stem, or on the onychophoran-arthropod LCA stem, both of which are somewhat possible placements. The reporter misinterprets this as the simple statement “Anomalocaris was an early arthropod”, which is exactly the mistaken statement that Meyer makes and which Luskin did not correct as “research assistant”.

Likewise Benjamin Waggoner (then of UC Berkeley, now at the University of Central Arkansas) writes in the journal Systematic Biology that “the anomalocarids and their relatives (Anomalopoda) fall out very close to the base of the traditional Arthropoda and should be included within it.” A 2006 paper in the journal Acta Palaeontologica Polonica likewise refers to the “anomalocaridid arthropods.” The leading authorities Charles R. Marshall and James W. Valentine note in a 2010 article in the journal Evolution, titled “The importance of preadapted genomes in the origin of the animal bodyplans and the Cambrian explosion,” that “Anomalocaris most likely lies in the diagnosable stem group of the Euarthropoda (but in the crown group of Panarthropoda).”

Waggoner also says in his paper that Anomalocaris falls outside of “Euarthropoda”, which is another term for crown-group arthropods defined by living taxa. For reasons that are unclear to me, Waggoner defines a larger “Arthropoda” that is supposed to be a crown group, but which is defined to include extinct forms outside of the clade of the living taxa. This is not the usual definition of “crown”, because crowns are supposed to be at the top of the tree, i.e. the present. Waggoner seems to be trying to say Anomalocarids go back to the Ediacaran, and thereby say the arthropods go back to the Ediacaran, and thereby connect the origin of arthropods to Spriggina and other Ediacaran forms. As far as I know none of these suggestions are widely accepted.

Regarding the Marshall and Valentine quote, again, “stem group of the Euarthropoda”, means outside of the crown group, i.e. outside what most people think of when you say “arthropod”. “[I]n the crown group of Panarthropoda”, however, provides no support at all for calling Anomalocaris an arthropod, because Panarthropoda is the crown clade made up of three phyla, namely arthropods plus tardigrades plus onychophorans!

The point of all of this is that you can’t just say “arthropod” when discussing Anomalocaris. You have to specify crown or stem, or some similar qualification, unless it is already clear within the discussion which you mean (which is the case in some expert discussions, but certainly not in the case of Meyer and his readers). All of the experts Luskin cites know Anomalocaris’s probable stem-group status, and they usually specify this qualification in some fashion. The only one that doesn’t is the Discover reporter, which just confirms my point – it’s an amateur mistake, unfit for a serious discussion of the Cambrian.

(An aside: read the next sentence of Marshall & Valentine 2009:

For example, Anomalocaris most likely lies in the diagnosable stem group of the Euarthropoda (but in the crown group of Panarthropoda). In fact, it appears that most fossil taxa in the Cambrian belong to diagnosable stem groups. (bold added)

Why in the world should it be that the animal fossils observed in the Cambrian – the ones furthest back in time – also just happen to tend to be cladistically basal on the cladograms? Evolutionists know why – but ID/creationists don’t even know about this evidence, or at least don’t dare tell their innocent readership about it.)

Luskin digs deeper:

Meyer doesn’t try to enter into the debate over whether Anomalocaris is a “stem group” or “crown group” arthropod, or a member of euarthropoda, or panarthropoda.

Oh god. These are not all either-or questions. “[E]uarthropoda” EQUALS “crown group” arthropod”, and “[E]uarthropoda”/”crown group” arthropod” AND “stem group” arthropods are ALL within panarthropoda. Anomalocaris is a member of panarthropoda no matter how you slice it, and there isn’t actually a “debate” slicing it anyway, since I think there is no analysis that places Anomalocaris clearly within euarthropoda (/crown-group arthropoda as defined by living taxa).

And, anyway, again, one cannot even enter a serious discussion of the origin of Cambrian taxa without having some statement about what taxonomy and relationships are being proposed as the basis for discussion. Pretending to punt on this (actually, Luskin and Meyer think that basically everything is specially created, as far as I can tell) just further discredits the idea that Meyer is engaging in serious scientific scholarship.

Since Meyer states that anomalocaridids are “either arthropods or creatures closely related to them,”

This statement is word salad, because Luskin has been arguing that saying “Anomalocaris is an arthropod” is correct because authorities say it’s on the arthropod stem. On Luskin’s current definition of what Meyer meant by “arthropod”, Luskin is therefore saying “anomalocaridids are either closely related to arthropods or closely related to arthropods.”

But, of course Matzke doesn’t accuse Nature, Budd, Jensen, or the authors of any of these other papers of committing a “basic error” for calling Anomalocaris an “arthropod.”

That’s because they don’t. They usually say “stem arthropod.” Which is correct.

About Lobopodia

Luskin writes,

And what about Matzke’s other accusation of an alleged error – his claim that Lobopodia isn’t a phylum? [italics original]

Um, phylum names don’t get italicized. Only genus/species names. And, anyway, I didn’t claim that Lobopodia isn’t a phylum – I don’t know what the formal, objective definition of a “phylum” is, and neither does anyone else, including those who still rely on the concept; the term only has meaning as a matter of convenience and convention. What I claimed was that you can’t write a responsible book about the origin of bodyplans/phyla without mentioning that lobopods, whether a phylum or not, are a paraphyletic grade containing taxa intermediate between, and ancestral to, crown arthropods, crown onychophorans, and crown tardigrades. Here is what I said:

A related problem is Meyer’s treatment (mostly non-treatment) of “Lobopodia”, which he treats as a distinct phylum and includes in his phylum count. Meyer never spends a word on an actual critical discussion of what “Lobopodia” is supposed to mean - the term appears in a few picture captions, in the titles of some of his references, and in a quote of Simon Conway Morris. Whatever the method of naming the various scientists who use the term “Lobopodia” - Linnaean ranks, rank-free, etc. - as far as I know every authority would agree that lobopods are a paraphyletic grab-bag on the stems of the crown-group phyla Arthropoda and Onychophora (and perhaps also on the stem below their common ancestor). In other words, the arthropod and velvet-worm phyla evolved from lobopods, and lobopods contain a whole series of transitional forms showing the basics of how this happened! How anyone could write a book on the origin of Cambrian animals, without mentioning Cambrian Explosion 101 findings like this, is mystifying.

Erwin and Valentine and everyone else discusses this. Why doesn’t Meyer? Either he doesn’t want readers to know about these transitional fossils, or he doesn’t know about them. Either way, it’s shockingly bad, and invalidates the book as being a competent piece of scholarship.

So, Meyer and Luskin can call lobopods a phylum if they want, but if they do, they have to mention to readers that it is morphologically in-between 3 other phyla (thus all phyla aren’t morphologically disconnected, the lobopod phylum contains 3 other phyla which makes you wonder what “phylum” is supposed to mean, etc. But this would have all kinds of subversive implications for their thesis, which I suspect is why they are either conceptually blind to it, or just left it out so as not to concern their innocent, unskeptical readership.

Interestingly, though, Luskin’s defense of phylum Lobopodia makes things worse for his position anyway. He cites the Supplemental Material of Erwin et al. (2009), which contains a big table of phyla – the same table appears as a supplement to Erwin and Valentine 2013. Luskin screen captures the table showing the listing of lobopodia as a phylum. But Luskin missed the other mention of lobopods in that table. Together, they are (shorn of formatting, sorry):

unranked stem Cambrian lobopods Luolishania Cam 3 e.g. Chen & Zhou 1997 (132)

Lobopodia Cam 3 class stem Microdictyon Cam 3 Hinz 1987; 15995; Kouchinsky et al. 2011 (8) Hadranax Cam 3 Budd and Peel 1998; 546 gilled lobopods Kerygmachela Cam 3 Budd 1993; 30407

So, which is it? Are lobopods an unranked stem, or a phylum? Or two phyla with the same name? (Plus the three nested inside, I suppose?) I suspect what we are seeing here is the older Linnaean taxonomy (my stage #1-2) and the newer, phylogenetic, rank-free taxonomy (stage #3) crashing into each other in the same data table, with the person compiling the table (“Prepared by Sarah Tweedt”, according to p. 343, Erwin & Valentine 2013) either making a mistake, or, more likely, just reflecting the contradictions in scientific literature caused by having phylogenetic and non-phylogenetic taxonomic systems both in play. (This is more evidence for why Erwin & Valentine’s continued reliance on Linnaean taxonomy (although they are somewhat apologetic about it in their text) is problematic, by the way.)

Finally, Luskin shows a screen capture of a chapter heading from a 2004 book, with chapter 14 entitled “Phylum Lobopodia” (http://www.evolutionnews.org/phylumlobopodia.jpg ) But, right there in the first paragraph, we see yet more evidence why it is so problematic to refer to this “phylum” without mentioning its paraphyly:

The Recent species, members of Onychophora…

In other words, phylum Onychophora nests within phylum Lobopodia. This should not happen, if the phylum rank is supposed to be some indicator of morphological distinctness and bodyplan uniqueness.

Random other points

Luskin says,

Page 419 of Darwin’s Doubt has a very nice discussion of stem groups and crown groups

No it doesn’t. First, this is hidden in an endnote, when it has to be front and center in any modern discussion (as it is in e.g. the works by Marshall, Erwin, and Valentine), and, second, Meyer gets the definition of “crown group” wrong, as I pointed out in my original post.

Luskin says,

Nonetheless, Matzke makes bizarre charges like this:

I think that if you plunked those fossils down in front of an ID advocate without any prior knowledge except the general notion of taxonomic ranks, the ID advocate would place most of them in a single family of invertebrates, despite the fact that phylogenetic classification puts some of them inside the arthropod phylum and some of them outside of it.

Luskin doesn’t say why my charge is bizarre, though. Here’s a challenge for Casey: explain why it’s bizarre. Please provide definitions of “family” and “phylum” and then explain why those fossils in the figure oh-so-clearly would fit in distinct phyla if someone didn’t know their phylogenetic relationships.

But Matzke seems unaware that Meyer has a lengthy 450+ word endnote on page 432 where he not only writes about long branch attractions, but addresses why that idea and many other ad hoc explanations fail to account for conflicts among phylogenetic trees.

No one ever says “long branch attractions” – is that some sort of new inter-tree romance or something? Anyway, in that endnote, Meyer only briefly discusses (1) horizontal gene transfer, then admitting it’s basically irrelevant when it comes to animals; (2) long-branch attraction, but incompetently failing to mention that there are several known solutions to long-branch attraction, such as adding more taxa to make branches shorter, and using more accurate sequence substitution models in likelihood and Bayesian approaches; and (3) incomplete lineage sorting (Meyer, strangely, when listing causes of incongruence, writes this item in the list: “coalescent (e.g. incomplete lineage sorting)”. Within the field, scientists only ever write “the coalescent” or “coalescence”; this makes me think Meyer doesn’t know what these are.) Meyer lists a few other sources of incongruence, like contamination, but without any discussion at all. Meyer says that these processes are related to convergence, which is false. Luskin claims that these explanations are ad hoc, which is also false. For example, how is contamination related to convergence, or how is it ad hoc? Sometimes the worm you are studying recently ate a worm from another clade, and your DNA sequencer gets a mix of DNA from both. It is easy to see how this could cause phylogenetic conflict – that is just life, it is regular science. It is perfectly checkable and fixable through methods just as resequencing a number of specimens, and starving the specimens before you sequence them. Similarly, long-branch attraction is not ad hoc, it is a direct mathematical result of using parsimony on branches that are long enough where the parsimony assumption (minimum number of changes) is wrong. The effect can be easily produced with simulation (as shown by one of my advisor’s and grand-advisor’s more famous papers, actually). Incomplete lineage sorting is also not convergence or ad hoc, it is a direct, unavoidable result of population genetic processes (drift) in the context of short speciation times. All of these processes are well-studied, well-understood, can be tested for, and thus it is just silly to ignorantly claim, without any study or due diligence whatsoever, that these explanations are just made up to cover up phylogenetic conflict. This kind of thinking is no better than 9-11 truther conspiracy thinking, sans knowledge of building engineering and similar necessary background.

Alright, my plane has finally arrived. My basic counterarguments against Meyer’s book (and, I guess, Luskin’s research assistance) are, if anything, strengthened by this analysis of Luskin’s rebuttal. In many cases, he still doesn’t understand the mistakes he is making.

Same rules as the other thread, and don’t expect my active participation, I have real science to work on once I’m back in the office.