Penguins & Ceratopsids, Lumpers & Splitters.

Adrian Currie writes...

On 14th December 2015 we all awoke to astounding news: the number of Little Blue Penguins (LBPs), otherwise known as kororā, fairy penguins or Eudyptula minor had… doubled.

Little Blue Penguins: yet another reason to visit the antipodes...

Little Blue Penguins: yet another reason to visit the antipodes...

Well, okay, they hadn’t doubled in terms of LBP units: there weren’t more penguins waddling about. Rather, media reports proclaimed they had doubled in terms of LBP species. Gosser et al argued that Eudyptula minor contains two cryptic species: one based in New South Wales (but with colonies in Otago), the other spread throughout New Zealand. So, what prompted this dramatic shift in LBP classification?

The main driver was the discovery of divergence in genetic markers: if you look at the genes of these critters, you find two distinct populations, which likely split around 750,000 years ago. They also have distinctive cries; ‘accents’. One imagines that this is a source of some cruelty, if NZ-Australia relations in Homo sapiens are anything to go by.

Okay, but why is this enough to warrant changes to LBP classification? Let’s consider the language of the paper which drove the flurry of media reports (catchily titled “Coalescent Modelling Suggests Recent Secondary-Contact of Cryptic Penguin Species”). They describe the taxa as “… potentially distinct species”. By the author’s lights, their (super interesting) nuclear DNA sequencing and analysis, is sufficient to claim the “…genus comprises two genetically distinct units”, evidence which is “…further supporting recognition of multiple Eudyptula species”. They finish with a recommendation that the LBPs get divided in two. Clearly, the authors think their evidence is sufficient to claim a new species-level division in Eudyptula, but their careful wording, and their use of a recommendation as opposed to an assertion implies that disagreement is possible.

What would such disagreement look like? Well, most obviously, someone might dispute the empirical results. But more interestingly, everyone might agree with the details: that there are two distinct LBP taxa, both genetically, vocally and geographically seperated by hundreds of thousands of years. And yet, while one side insists this is sufficient to divide the taxa, the other side might equally vehemently deny this. The former are LBP splitters and the latter are LBP lumpers. Splitters like their phylogenetic trees bushy, lumpers like them sparce.

On the left, a lumper's christmas tree. On the right, a splitter's christmas tree.

On the left, a lumper's christmas tree. On the right, a splitter's christmas tree.

Today, I’m interested in challenging a fairly wide-spread idea about the nature of the split/lump divide. Namely, that lumping or splitting is largely a matter of convention, or worse yet—personal taste. I think the debate is much more about how different biologists see the world. In other words, I see it as a largely empirical dispute.

In a recent post Leonard implies just the view I’m worried about. He wonderfully draws together the lump/split divide and biological ‘monsters’: that is, unique critters. He implies – plausibly, I think – that taking splitting to extremes and admitting unique individuals as biological categories is a mistake. What’s more, he does this via the Loch Ness monster. Definite bonus points on that one. Leonard’s way into the debate is not just through Nessie, but also Darwin: he describes a tension in the Big Man’s thought between the individuating nature of natural selection, and the desire for good categories:

In Darwin's theory, then, we have a tension between liberalism about taxonomic differentiation—caused by selection, which is 'daily and hourly scrutinising, throughout the world, every variation, even the slightest' (Darwin 1859 p. 84) and generating new biological taxa through that scrutiny—and conservatism about principles of classification.

On this kind of view, splitters are taxonomic liberals, emphasizing natural selection’s capacity to differentiate between lineages; lumpers are conservatives, preferring nice clean taxonomic systems.

My disagreement with Leonard concerns the latter aspect of the tension. I don’t think that conservatism about taxonomy is primarily driven (or perhaps I should say: it shouldn’t be primarily driven…) by ‘principles of classification’—the thought that lumpers just want a tidy taxonomic house. Rather, I think Darwin’s work is driven by concern for two empirical phenomena. On the one hand, the biological world is wildly diverse: it has rotifers, octopus, and LBPs. On the other hand, it is remarkably conservative: all vertebrates are basically minor variations on the ‘give them a backbone and shove a head on one end and an anus on the other’ strategy of building bodies. Darwin did not only explain the diversity of life in reference to natural selection’s capacity to drive changes in populations over time, he also explained the conservatism of life as being due to shared inheritance. I grew up in a different environment than my father, and this explains many of our differences, such as his indifference, and my love, of hiphop. But we are connected through chains of inheritance, and this explains many of our similarities. We both have red-tinged hair, for instance.

On my view, lumping and splitting are not driven by the conventions of categorization clashing with natural selection’s creative impulses. Rather, taxonomic decisions are made on the basis of how we explain the differences and similarities among critters. Consider the penguins again. Grosser et al provide an empirical argument to suggest that some genetic and behavioural patterns in LBP populations are explained by differing evolutionary histories. In the past, the two populations became isolated and split, and this is why the LBPs in Otago talk so funny. They appeal to extra-lineage causes. Their opponents are hypothetical so far as I know, but if there were to be opponents, they’d possibly argue that the differences were due to some patterns in LBP development: they’d appeal to some intra-lineage causes like differences due to ontogenetic stage, sex, developmental plasticity, and so on. To make this position explicit, let’s take a look at a palaeontological case where the opponents definitely aren’t hypothetical.

If a world with two species of LBP was a shock, that had nothing on awaking in mid-2010 to learn that the famed tri-horned beast of the late Cretaceous, Triceratops, had never existed! John Scannella and Jack Horner argued that what we once thought were two distinct genera, Triceratops and Torosaurus, were in fact two ontogenetic stages of the same species. Specifically, Triceratops were baby Torosaurus.

(Now, equating that with the non-existence of triceratops is very misleading. First off, by the conventions of taxonomic nomenclature, because Triceratops was named first it is Torosaurus who is up for the chop. More importantly, there’s a difference between saying that thing doesn’t exist and saying that thing is a baby version of that other thing.)

Anyway, how did Scannella and Horner make their argument? I’ll be going into more detail in my next post, but the short version is roughly this…

Get an ‘adult’ triceratops. Run its ontogenetic processes forwards (that is, keep it growing). You end up with something suspiciously close to a Torosaurus. Consider the two skulls below:

Either the skulls of two genera of Ceratopsid, or the young and mature versions of a single species... (from Dinosaurs, by Matthew (1915))

Either the skulls of two genera of Ceratopsid, or the young and mature versions of a single species... (from Dinosaurs, by Matthew (1915))

The thought is that if you keep the developmental processes acting on the one on the left going, you’ll end up with a skull quite a lot like the one on the right. This suggests that the differences between the specimens is not that they have different evolutionary histories, but that one has grown for longer than the other.

In the background is the idea that, over their lifespans, dinosaurs had much more morphological variation than we generally thought, and this has led to differences in ontogenetic stages being mistaken for differences in evolutionary histories. Which is to say, we’ve been splitters where we should have been lumpers. And note that this doesn’t turn on some preference of ours, or on conventions about what makes for a good taxonomic system. Rather, it turns on our background beliefs about the evolutionary and developmental features of the relevant critters. Roughly, the question is whether inter-lineage or extra-lineage causes are the best explanation of the patterns of similarity and difference we see in the relevant specimens.

As Longrich and Field (splitters in the context of the ‘Toroceratops’ debate) have said,

Before one can use variation to classify species, it is necessary to understand the nature of that variation. That is, do the differences between two fossils represent variation between different species, which is a result of separate evolutionary histories, or do these differences reflect variation within a single species, which can result from variation within a population, sexual dimorphism, or change in morphology over the course of development.

Decisions to lump or split typically turn on arguments about the sources of variation across the biological world. These are not, primarily, about what good classification looks like, but rather about the nature of the critters in question. It is an empirical dispute. We’ve just seen this play out in debates about living penguins and extinct ceratopsids.

In the LBP case, scientists (convincingly, in my opinion) argue that variation between the penguins in Otago and those in the rest of New Zealand is best explained by their having separate evolutionary histories. Now, this doesn’t quite get rid of the role of taste or convention. One might disagree that the differences which Gosser et al point to are interesting enough to constitute a new species designation (perhaps some kind of ‘subspecies’ category is more appropriate), but at the very least this pushes such disputes back a step, and rightly puts the emphasis firmly on empirical claims about the evolution and development of the relevant critters.

In the ceratopsid case, it’s much less clear whether we should side with the lumpers or the splitters (but again, this is because of empirical uncertainty, not disputes about how best to categorize). In my next post, I’ll delve into the ‘Toroceratops’ controversy in more detail, and suggest that we philosophers have been messing up by focusing so much on what we call ‘concepts’…

(oh, my take on the ‘Toroceratops’ debate, along with some fun metaphysical speculation, is here…)