Tuesday, August 5, 2014

Lumpers, Splitters, and Hobbits

(warning-- anthropology rant ahead.)

Last year, as part of an independent study, I wrote a paper on some of the methodologies and controversies surrounding hominin taxonomy (naming species of human ancestors). Although the paper is fairly broad and not super technical  (it was meant for my own edification), I did come away with the sense that paleoanthropology is naming too damn many species.

This is not my own novel idea, of course.  There is longstanding and lively debate between "lumpers" (those who prefer to acknowledge only a few species in the human fossil record) and "splitters" (those who see evidence for many species).  Ideally I'd prefer not to pitch my tent in either camp because the ever-growing fossil record is continually changing our perspective.  But if pushed I'd have to side with the lumpers.  Given the tiny sampling of actual extinct diversity that the currently-known fossil record displays, it seems egregiously rash to be assigning so many species names.  The chart below is one author's hypothesis for hominin phylogeny.

The purported connections between species (cladogenesis or anagenesis) is based on analysis of shared traits between fossils, and isn't affected much by one's affinity for lumping vs. splitting.  But notice all of the species shown above.  Really? And look at our own genus, Homo, which shows up around 2 million years ago. A splitter may describe 10 or more human species in our genus.  But 2 million years is the blink of an eye: many mammals (including primates!) have been around for 2 million years, and we don't split them up into separate species.  We could make sense of this by suggesting that humans have experienced an unusually high number of speciation events in just 2 million years.  However, this would require populations to have been severely isolated from each other, with genetic drift working swiftly and powerfully.  I don't doubt that both of these things probably did happen, but enough to produce up to 10 species of human?  To me, it seems more likely--more parsimonious, a gold standard for scientific hypothesizing-- to instead conclude that we have conflated our own importance as a genus and in turn recognized far too many species.

All of this turns on which "species concept" you adhere to. In defining nonhuman species, most biologists use the biological species concept*, which basically states that 2 populations can be considered different species if they cannot produce fertile offspring together.  This species concept allows for testable hypotheses about species designations-- see if they can interbreed!  It also has the benefit of being, essentially, a proxy for measuring what really should matter-- degree of genetic change between populations.  That's really what speciation is all about.  Instead, paleoanthropology relies on other species concepts, in part because you can't really test the biological species concept with extinct species.  The result is that two researchers may reach very different conclusions when analyzing degree of similarity between sets of fossils, and we can't even agree on how much dissimilarity is required for separate-species status.  Subjectivity is high. And I think we're again guilty of anthropocentrism-- maybe even something reflecting our racist tendencies?-- in seeing the differences between fossils as evidence of otherness.  The phylogenetic history of most living mammals (including nonhuman primates) isn't marked by multiple, recognized speciation events over the past 2 million years, but this same time span is described as hosting multiple human species. Hmm.  (*The biological species concept is imperfectly applied, even in biology. Grizzlies and polar bears can interbreed, clearly negating their separate-species designation. But it's the closest we've gotten to objectivity.)

All of this rant came to mind when I read the latest on "the Hobbit", aka Homo floresiensis, defined primarily by one cranium (LB1) found in 2003.  It's very small and very, very recent.  If it's a unique species, it challenges our current model of human origins in some big ways.  So assigning this specimen to its own species is, well, the least parsimonious move available to the researchers, but that's what they did.  A paper just published in PNAS (here ) takes aim at this interpretation, and you get the sense that it's personal.  They describe the process of naming this species as a tale of science gone wrong, with ad hoc reasoning and sensationalism trumping the process. (They offer their own explanation of the Hobbit-- time will tell whether it's accurate. But regardless, it's a more measured hypothesis.) Tim White (2013), in reference to the debate over Homo rudolfensis, sums up the broader issue:

“Paleoanthropology’s ecosystem of publishing, access, fundraising, career advancement, media promotion and celebrity seems squarely aligned against the field’s ability to self-regulate, a condition exacerbated by the limited fossil resources available. There is ample and obvious motivation for authors to generate ‘new’ species names in this environment.”

I wonder if the pendulum will swing back?

2016 update on the Hobbit: Well, shit... similar, likely ancestral, specimens were found on Flores and dated to 700,000 years ago, lending support to the interpretation of the Hobbit as a unique species descended from something more primitive (likely erectus).

No comments:

Post a Comment