Tuesday, August 25, 2009

A dubious honour...

It seems that my latest paper has been nominated for a dubious honour. That is, I've been singled out as having committed a cardinal sin of systematics: appeals to the reality or significance of paraphyletic groups.

This post was some time ago, and I have not had time to address it. And, I'll mostly not address it in detail here as it is not terribly worth it. Mostly, it is a kind of juvenile stunt, rather than a serious academic undertaking. However, since the authors Williams & Ebach (with whom I actually agree about much, even with respect to fossils), have ascribed to me ideas I do not actually subscribe to: namely a belief in paraphyletic groups, I'll post a little response here. In fact the point of Brazeau (2009) is to demonstrate that a group that is commonly appealed to in the literature, the "Acanthodii" is, in fact, a non-real group.

Most of Williams & Ebach's gripe with my paper is derived from either a BBC report or a non-specialist, non-technical, non-peer-reviewed interview piece in Nature. I have never used the term "missing link" in my article, nor did I use it in discussions with journalists. In fact, I try as much as possible to disabuse journalists of such popular misconceptions.

No, what is most surprising are the factual errors about my work that Williams and Ebach have made:

What any systemtist should do - re-classify the osteichthyans and chondrichthyans in light of this new evidence. Brazeau is naive to suggest that this discovery will "...not overturn a general consensus about gnathostome interrelationships" If Ptomacanthus is more closely related to chondrichthyans then bang goes the acanthodians. They need to be reclassified along with the chondrichthyans.

This contains several patently wrong statements. The monophyly of the Chondrichthyes and Osteichthyes remains after my analysis, as did their status as each other's extant sister group (which my analysis could hardly have contradicted apart from finding if their respective monophyly is not challenged). That general consensus is not changed by my result, so there is no need to re-classify either osteichthyans or chondrichthyans.

The acanthodians do not all get re-classified with chondrichthyans because, as my results showed, some "acanthodians" are members of the osteichthyan stem. So, we have to reclassify some as chondrichthyans and some as osteichthyans. Something I entirely agree with. Figure 3 of my paper clearly shows where I have placed Ptomacanthus in the group Chondrichthyes and a bunch of other "acanthodians" under Osteichthyes and highlighted in bright colours so that you could see that this is what I already did!

Figure caption: a, Strict consensus trees of the 2,904 shortest trees from the global analysis (left; treelength: 318 steps; consistency index: 0.44; retention index: 0.76; rescaled consistency index: 0.34) and the 30 most parsimonious trees from the endocranial data set (right; treelength: 83 steps; consistency index: 0.64; retention index: 0.85; rescaled consistency index: 0.54). b, Bothriolepis. c, Buchanosteus. d, Tetanopsyrus. e, Ptomacanthus. f, Cladodoides. g, Acanthodes. h, Mimia. Vertical arrow shows position of palatoquadrate-braincase articulation that corresponds to the basipterygoid articulation shown in Fig. 2. Double digits indicate percentage bootstrap support; single digits show Bremer decay indices (when greater than 1). Illustrations are modified from refs 5 and 18 (also see Supplementary Information).


Continuing, Williams & Ebach write:
But rather than saying the obvious, Brazeau descends into evolutionary explanation "... populates the long, naked internal branches, revealing a much richer picture of character evolution in early gnathostomes". No it does not reveal anything other than that Ptomacanthus is a chondrichthyan and that acanthodians are paraphyletic!
I did state the obvious. It's in the figure. Look at it. I did not "descend into evolutionary explanation". The nested series of monophyletic groups that imply acanthodian paraphyly actually do provide sequences of character acquisition along the chondrichthyan and osteichthyan stem segments. As Williams & Ebach know well, each monophyletic group is supported by synapomorphies, and those nested groups synapomorphies are simply synonymous with what we call 'sequences of character acquisition'. This is how we make sense of fossils (or any other newly discovered taxon) and the implications fossils have, if any, on further hypotheses of synapomorphy (homology). If it's not the sequences of nested homologies that define monophyletic groups (the groups that matter) then what does? I'm perplexed as to why Williams & Ebach, of all people, would challenge this, since this seems to be their own view. I thought we had accepted and moved beyond disputing the idea that "evolution", when talking about fossils and the unrepeatable past, was only reducible to our best systematic hypotheses. In the quoted statement, that is all it is to me. It seems, perhaps, I wasn't careful enough and Williams & Ebach saw what they wanted to see in it. If so, then I'll take responsibility for my error, but note that my critics are playing fast and loose ascribing ideas to me which I have not explicitly stated.

Finally, they raise the following gripe:
"The study also suggests that some acanthodians are ancestors to all modern jawed vertebrates" (BBC Online, 19 January 2009).
This is false and misleading - the study shows quite the opposite.
Mostly, Williams & Ebach are just being pedantic and annoying, but this is infuriating bullshit. Those are not my words!

My words in the BBC article were:
"This figures in nicely with the emerging idea that acanthodians don't form a group of fishes that are all closely related to each other. Some of these fossils are primitive sharks while others are primitive bony fishes."
Even in the BBC article I state clearly that some are chondrichthyans (though I used the term "sharks" as a shorthand) and others are osteichthyans.

I believe my primary sin in that paper is to refer to terminal taxa as "basal". As I will cover here in another post, this is a problematic use of the term "basal", and one that is infectiously used amongst people who apply systematic methods. Maybe that could net me a Pewter Leprechaun, but if you nominate me on that basis you have to nominate just about anybody who talks about trees these days.

3 comments:

RBH said...

It's early days, but it'll be interesting to see if Williams & Ebach respond to your post here.

Martin Brazeau said...

Took me months to reply to them. I see no reason why they should feel the need for a speedy reply.

frank said...

Williams and Ebach should have just stuck to the actual paper (which they should have read without a jaundiced eye) and also just concern themselves with direct quotes from the BBC interview. As it is, they've simply embarrassed themselves.