Monday, December 9, 2013

Is punctuated equilibrium an argument against phylogenetic systematics?

I have mentioned before on this blog that one of my interests is the continuing opposition of a minority of systematists against the requirement that formally accepted supraspecific taxa have to be monophyletic.

I was originally trained under the paradigm of phylogenetic systematics and considered the issue to be settled. Sometime around 2010/2011 I grew sufficiently puzzled about the flood of opposing articles being published especially in the botanical journal Taxon to wonder whether I had overlooked something. Maybe the opposition was right? Maybe I had overlooked something, and cladism was really incoherent or at least undesirable as a scientific practice?

So I systematically read through a large number of publications promoting the acceptance of paraphyletic taxa to examine their arguments - only to find very nearly all of them severely wanting, and the one that at least made some sense to be based on a number of assumptions that can certainly be questioned, such as that we can ever positively know that a fossil species is ancestral to extant species instead of a side branch of the tree of life close to the real ancestor.

What is more, except for that last one, which was advanced somewhat more convincingly by the late Richard Brummitt, the papers advocating the acceptance of paraphyletic taxa can mostly be assigned to one of three styles of argumentation:
  1. The Gish Gallop, i.e. the attempt to throw as many PRATTs out there as possible, clearly in the hope that the reader is too overwhelmed by their number to ask if any of them make sense and that a cladist will find it impossible to address all of them in the limited space of a rebuttal. Many of the aforementioned publications in Taxon fall into this category.
  2. Mere assertion, in other words short opinion pieces or letters that simply proclaim organisms should be grouped by overall similarity instead of relatedness and/or that nomenclatural stability is more important than scientific accuracy. They generally do not present any arguments whatsoever with the possible exception of an argumentum ad populum, presumably because the author considers their position to be too self-evident to stoop to actual reasoning.
  3. Utterly incoherent gibberish.
If you find that a position is consistently promoted as badly as this, the likelihood is great that there are no good arguments to be found because if there were they would probably have been used in the first twenty papers you read. This observation has, of course, sometimes been made in the context of the creationism manufacturoversy but it applies equally to the proponents of paraphyly.

Nonetheless, or perhaps precisely because of the wry amusement that can be derived from seeing catastrophic errors of reasoning presented either as deep insights or as common sense, I remain interested in papers that promote paraphyletic taxa. (And who knows, maybe I am wrong, maybe one of them will finally present a new argument that will sway me?)

Google Scholar appears to understand me better than I expected because it recently suggested I might want to read an opinion piece published in the journal Zootaxa: Why Drosophila is not Drosophila any more, why it will be worse and what can be done about it?, by one Jaroslav Flegr. And indeed this paper does not disappoint because it presents a line of argumentation that I have not come across before.

Sadly, that does not mean that it is any more convincing than the others. If you are here for the plant pictures, you may want to look away now because this is not going to be pretty. Also, it is going to be long. I apologize in advance.

The background

Flegr writes his contribution on the occasion of a recent decision by the International Commission on Zoological Nomenclature (ICZN) not to retypify the name Drosophila. Molecular phylogenetic studies have demonstrated that the model organism D. melanogaster would have to be moved to the genus Sophophora to make both genera monophyletic, and changing the name of such an important species will be very disruptive and expensive. It had therefore been suggested that D. melanogaster should be made the type of the genus Drosophila so that it could keep its name, but the ICZN turned that suggestion down.

Now there are several ways Flegr, concerned about the conflict between nomenclatural stability and phylogenetic systematics, could have gone from here. The most obvious one would be to complain about the obstinacy of the ICZN because obviously that body could simply have decided differently. Although I must admit that I am ignorant of the arguments that have been made for and against retypification I strongly suspect that this would have been my approach if I were a zoologist. However, we know that Flegr did not locate the problem there because then you would not be reading this post.

This leaves at least two alternatives: One can blame phylogenetic systematics because without the monophyly criterion the name would not have needed to be changed, or one could blame Linnean taxonomy because it is really its use of the genus name as the first half of the species name that causes all the trouble.

Flegr chooses the former. So far so good, everybody is entitled to their opinion, although one could add that names changed plenty of times even before the advent of cladism. But it is symptomatic of what one might call paraphylists that he never even so much as mentions the possibility that a nomenclatural system devised in complete ignorance of common descent may not be ideal to describe biological reality as understood today, after the development of the theory of evolution.

There are quite a few cladists who would agree that Linnean taxonomy and phylogenetic systematics are incompatible but whose solution is to discard Linnaeus. And really, although as somebody who works only on extant organisms I do not see quite that much incompatibility myself, that makes a lot of sense. If your description of natural history does not fit natural history then it is your description that you should change because changing natural history is not really an option.

The point is that Flegr starts his contribution with a false dilemma: stability or monophyly. There are considerably more options than he is willing to admit, and it should simply not be possible to write so many words on this issue without so much as mentioning the PhyloCode, even if only to reject it.

Some samples from the text

Before we deal with the paper's actual argumentation, it might be instructive to examine a few snippets from the text, just to get a feeling for what we are dealing with.

After an explanation of the Drosophila issue, we are given this gem:
In the past decade, i.e. after the advent of multigene or even whole-genome studies, molecular phylogenetics evolved from an amusing toy for molecular biologists searching for a 'cheap publication' or for a simple project for a pregraduate or postgraduate student to a potent tool that can provide a reliable picture of the phylogeny of particular groups of organisms. With this advance, more and more biologists have been forced to admit that many traditional taxa, including the genera, were wrongly constructed and needed to be redefined or even abandoned.
Although I try to be charitable, it has to be said that this paragraph reads as both breathtakingly arrogant and utterly clueless about the field it comments on. It uncritically regurgitates the myth that molecular data as such are leading to recircumscription of taxa; in reality it is cladistic thinking that made the difference, regardless of the data used. Many of the most striking advances in our understanding - the non-monophyly of traditionally circumscribed dinosaurs, dicotylous flowering plants, and bryophytes, for example - were derived from parsimony analysis of morphological and anatomical characters.

Conversely, all the molecular data in the world would not lead to taxonomic changes without a clear scientific criterion for their circumscription. (The latter is necessarily one of the premises of Flegr's entire chain of reasoning because his ideal situation would be lots of molecular data but no cladism, but perhaps internal consistency was not a priority here.)

And the condescending remark that molecular phylogenies are cheap publications and were until recently mere playthings is downright ridiculous when we consider the pioneers who spent months in the lab painstakingly producing the first datasets and obsessing about the right way to analyze them more than twenty years ago, at a time when much less lab equipment and computer power were available than today, and the progress we have made in our understanding of evolutionary history using these 'playthings'.

This distressing nonsense is swiftly followed by one of the most confusing explanations of phylogenetic systematics that I have ever come across. It also contains this fine sentence:
When the species of clade A were originally assembled into taxon A and species of clade B were originally assembled into taxon B of the same or higher taxonomic rank as taxon A then new results change the status of taxon A from monophyletic (holophyletic in Ashlock’s (1971) terminology) to paraphyletic (polyphyletic in Hennig’s (1966) terminology).
Yes, Ashlock's redefinition of 'monophyletic' to include 'paraphyletic' did not actually help to make communication easier but, and correct me if I am wrong, I am fairly sure that nobody, and least of all Flegr's brothers-in-paraphyly, has ever called 'paraphyletic' what Hennig called 'polyphyletic'. Touches like this one might make one start to wonder whether the author of the piece is well enough informed about the topic he is writing about to be taken seriously. Skipping ahead:
paraphyletic taxa (except the species, whose paraphyletic nature is usually tolerated even by cladistics)
We are seeing here in brackets another PRATT, although there is also a non-zero number of cladists who are confused about the issue - and so, I will gladly admit, was I until a few years ago. Sexually reproducing biological species do not show a phylogenetic structure but instead a net-like structure. In the absence of phylogenetic structure there cannot be any paraphyly or monophyly because there is simply no 'phyly' at all. It is thus not the case that cladists tolerate paraphyletic species because there are no paraphyletic sexually reproducing species to be tolerated. Instead well-informed cladists know that phylogenetic systematics only applies in phylogenetic structures. Or in other words: where there are no clades there can be no cladism.

Let me now cite at length a part that showcases one of the most common problems with paraphylists: circular reasoning.
In contrast, in the punctuational world, the paraphyletic taxa will be relatively common as a species can switch from the frozen to the plastic state in any place of the phylogenetic tree. The origin of a new genus as a result of the switch of some frozen species to the plastic state will be most probably accompanied by the transition of the originally monophyletic (holophyletic) genus to the paraphyletic state (Zander, 2010), but see also Podani (2010b).
The problems with paraphyletic taxa do not only concern the genera but also the taxa of any taxonomic rank. In any rank-based taxonomy, a higher rank taxon is originally composed of species that shared the same unique combination of phenotypic characters. When a certain species of a particular taxon becomes so plastic that even these characters radically change and taxonomists separate this species and its descendants into a new taxon of the same or higher taxonomic rank, the taxon becomes paraphyletic.
Although the last sentence implies at least some form of recognition that taxa are defined by taxonomists, the muddled rest of this section is fairly typical. Flegr starts with the assumption that some degree of morphological change moves a species to a different taxon at the same level and argues from there. In other words, he is simply building his desired conclusion into the premises of his argument. In reality, there are no naturally paraphyletic taxa, there is only a phylogenetic structure in which taxonomists circumscribe taxa to be either monophyletic or non-monophyletic.

Having only recently seen another such case in the newsletter of a botanical society, I am deeply puzzled why it is so hard for paraphylists to notice the glaring flaw in their logic. If I were to tell one of them, "New Zealand should be a part of Australia because it is, after all, a part of Australia", they would surely realize what is wrong with my reasoning. Why are they so consistently unable to spot the exact same logical fallacy when the topic is taxonomy?

The formal argument, such as there is

So what is Flegr's argument for paraphyletic taxa? Working a bit forwards and backwards through the text, we can summarize as follows.
Premise #1: Changing important names is inconvenient.
Premise #2: That we have to change important names is the fault of phylogenetic systematics.
Premise #3: Evolution is punctuational.
Premise #4: Punctuational evolution means that speciation is in most cases accompanied by little change (leading to groups of related species that are very similar) but rarely accompanied by massive morphological and physiological change making a new species and its descendants fundamentally different from the rest of their relatives.
Conclusion #1 (from P3 & P4): Relatedness tells us nothing about 'outer' or even 'inner' similarity of species.
Conclusion #2 (from C1): Strong morphological divergence justifies treating part of a clade as a separate taxon at the same level of the classification as the paraphyletic rest of the same clade.
Conclusion #3 (from C2): Phylogenetic systematics would make sense if evolution were gradualistic because then relatedness would be predictive of similarity. But because of the reality of punctuationalism, phylogenetic systematics does not make sense. A system recognizing paraphyletic taxa is more appropriate to classify the results of evolution than a phylogenetic system.
The first thing we might notice is that the first two premises, in other words the first few paragraphs of the paper, lead absolutely nowhere. One might say that an additional conclusion - we need to get rid of phylogenetic systematics for the sake of nomenclatural stability - is implied here but it is not spelled out. Which is probably wise because premise #2 is flat out wrong (as mentioned before, names were changed plenty before the advent of phylogenetic thinking), and the entire line of thought is profoundly anti-science.

Of course science is anathema to stability. Its entire point is to poke at things and find out if we can understand and describe them more accurately than we currently do. Surely it was also inconvenient for doctors trained in bloodletting when medical science moved away from that practice, and it was inconvenient for textbook publishers when Pluto was downgraded from its status as a planet. So what?

Everything from premise #3 to the end is at least a coherent line of thought that goes somewhere. But it should immediately give us pause when we see punctuated equilibrium used as an argument. This topic is one of evolutionary biology's equivalents to quantum physics: easily misunderstood and consequently a magnet for cranks. I will also gladly admit that I am not an expert either, and that I have myself not read Gould's seminal paper presenting the idea. So considering that premise #4 is central to the logic of the piece, let's compare Flegr's understanding of punctuationalism against other sources.

This is how Flegr describes it:
Some of these models suggest that the species are evolutionarily frozen under normal conditions and can only partly respond to the existing selection pressures (Carson, 1968; Flegr, 1998, 2010; Mayr, 1963; Templeton, 1980). Under some conditions, for example after peripatric speciation, they can turn to the plastic state in which they start to respond to selection; however, after some time, as a result of an accumulation of genetic polymorphism by frequency-dependent selection, the plastic species switches to the normal, frozen, state again and the adaptive evolution of the new species ends. Transient plasticity of a new species can lead to a radical phenotypic switch without any change to the environment and can be accompanied by a rapid radiation and diversification of the new species resembling the adaptive radiation. In the contrast with other (more frequent) types of speciation (e.g. vicariant, sympatric, parapatric), the speciation coupled with the switch from the frozen to plastic state can lead to such a prominent change in the phenotype or may even give origin to several new divergent phenotypes
For comparison, we might want to look at the richly sourced Wikipedia entry on punctuated equilibrium, in particular under the section helpfully labelled "common misconceptions" (accessed 7 Dec 2013). For example the following (note that I would generally be careful with using Wikipedia but this section mostly consists of direct quotes from the guy who came up with punctuationalism and from other evolutionary biologists):
Punctuated equilibrium is often portrayed to oppose the concept of gradualism, when it is actually a form of gradualism.[47] This is because even though evolutionary change appears instantaneous between geological sediments, change is still occurring incrementally, with no great change from one generation to the next. To this end, Gould later commented that "Most of our paleontological colleagues missed this insight because they had not studied evolutionary theory and either did not know about allopatric speciation or had not considered its translation to geological time. Our evolutionary colleagues also failed to grasp the implication(s), primarily because they did not think at geological scales".[12]
The really funny thing here is that Flegr wants to have it both ways. He writes that "the evolution of most of multicellular species is not gradualistic but punctuational in nature", but that would mean that most speciation events would produce about the same degree of divergence, and that that degree cannot be too large anyway because then we would never get groups of species that are similar to each other. On the other hand, for his argument to make sense he also has to claim that the majority of speciation events are gradualistic and a few are ridiculously saltational, which looks pretty much like a severe misrepresentation of the concept of punctuated equilibrium. Could we have some intellectual consistency here?

So unless somebody can come up with a very different interpretation of the above than I did, premise #4 is dead in the water. I may be wrong but it looks to me as if Flegr does not understand punctuated equilibrium, or perhaps he deliberately misrepresents the concept to use it as a cudgel against cladism.

It is also unclear to me whether he can point to any empirical data showing that punctuationalism leads to massive divergence in the majority of characters of an organism (especially 'inner' characters, to adopt his terminology), which would be necessary to justify his conclusion that relatedness is not predictive of similarity. Punctuated equilibrium applies to individual speciation events, but the cases he appears to be most concerned about in his text - e.g. birds nested in dinosaurs, tetrapods nested in fish - show divergence that accumulated over such long time periods and through so many well known fossil intermediates that they are perfectly gradualistic. What he needs for his argument to work is pretty much a chicken hatching hatching from a T. rex egg, and even then the T. rex would still be more similar to the chicken than it would be to most other dinosaur lineages!

That does not mean that there cannot be extant clades that are superficially quite divergent from the extant paraphyletic residues that they are nested in. Of course there are, and over time they would come to exist even in an entirely gradualistic world. But it means that the argument reduces to the same old, same old: the assertion that relatedness should not matter beyond some degree of divergence.

And it also comes with the same old problems: the degree of divergence that is considered sufficient for the acceptance of paraphyletic taxa is a matter of subjective (i.e. unscientific) preference, and because evolution is still gradual the observed divergence is in all cases an illusion, merely an artifact of extinction and of an incomplete fossil record. Worse, the illusion could be broken at any time: we might still discover intermediate extant species or fossils, blurring the line between a paraphyletic taxon and its formerly 'sufficiently divergent descendant'.

Essentially then this piece regurgitates a point refuted a thousand times before and dresses it up with a few big words whose meaning is at best misunderstood, at worst misrepresented.

And now prepare for the scariest part of the paper

Ready?
I would like to thank to all three referees for their constructive suggestions and comments.
Yes, Flegr's contribution to Zootaxa was refereed by three peer reviewers. Apparently none of them pointed out his factual mistakes, none of them explained to him that punctuated equilibrium is not the same as hopeful-monster-saltationism or that there is no evidence for the latter, none of them helped him make his two explanations of taxonomic practice any less confusing, and none of them suggested he at least be honest enough to mention the PhyloCode in his introduction.

This is the state of the peer review process in taxonomy. May the gods have mercy on our souls.

2 comments:

  1. I think this is more widespread now, but I am familiar with an official list of common and scientific names of USA fishes. This list is revised about every 10 years. The common names are quite stable, and the scientific names changes with changes in knowledge of relationships. So why not solve the Drosophila problem by declaring Drosophila a fixed common name for the beloved experimental fruit fly, and let its taxonomy flourish among those who care?

    There is an argument that there should not be monospecific genera. I disagree in that a genus should be morphologically coherent as well as monophyletic. Secondly, if the sister group of a species is a genus, that species should occupy a monophyletic genus.

    The first killifish species I described is now in its fourth, and last, one hopes, genus. Previous generic placements made sense at the time.

    ReplyDelete
  2. Your comment is getting at something that I have been thinking about a bit but have not yet been able to express clearly enough. We have use for a phylogenetic system and we also obviously have use for various non-phylogenetic ones such as common names, trade names, ecological roles, growth form of plants, etc etc, the sky is the limit. So yes, I could not care less about people continuing to call that species Drosophila as long as we also have a science-based system that depicts its relationships accurately. A cladist could live happily with all these systems next to each other as long as there is a phylogenetic one among them.

    The problem is, many proponents of paraphyly want to destroy the only one that has been made increasingly phylogenetic, and they want to do so by turning it into some phylogenetic/phenetic hybrid that is useless for any purpose whatsoever because you never know if something is member of a taxon due to its phylogenetic position or due to its phenetic similarity with other members.

    ReplyDelete