Blue Collar Scientist

Some interesting research about the evolutionary development of the human skull is being reported by UC Davis. The researchers have concluded that random change - called genetic drift in the parlance - accounts for most of the differences between human and Neandertal skulls.

In their new study, Weaver and his colleagues crunched their fossil data using sophisticated mathematical models — and calculated that Neanderthals and modern humans split about 370,000 years ago. The estimate is very close to estimates derived by other researchers who have dated the split based on clues from ancient Neanderthal and modern-day human DNA sequences.

This is significant, because it tells us that rigorous study of morphological changes in fossils gives us dates in good agreement with that of DNA methods. The picture here is that two separate disciplines offer mutually reinforcing insights into human evolution over this time period.

But then the press release gets a little strange. The principal investigator, Tim Weaver, says:

A take-home message may be that we should reconsider the idea that all morphological (physical) changes are due to natural selection, and instead consider that some of them may be due to genetic drift. This may have interesting implications for our understanding of human evolution.

As a layperson with a pretty solid understanding of evolution, I’ve been suspicious at times of some scientists’ tendency to see natural selection where I felt¹ some things might more simply be explained as neutral features - not selected against, not selected for, just arising randomly without any particular immediately adaptive traits. On the other hand, I’ve always been aware that I’m not the expert, so I’ve been inclined to trust the authorities and quell my own misgivings.

But despite my being predisposed to have sympathy with this conclusion, I have to admit that this is where the press release falls down into a morass of uselessness.

There is, almost certainly, a reason why Weaver thinks that his research indicates genetic drift, rather than selection, has been a major influence on the evolutionary development of the human skull. I’d bet that reason is a really good one, too. But the UC Davis release, as well as the PhysOrg coverage, are completely silent on what that reason might be. The closest they come to giving a reason for their belief is that they used “sophisticated mathematical models.”

I think it is problematic for a press release to pass up an opportunity to explain not only what we know, but how we know it. As has been often repeated, science is not a collection of purported facts. Science is a process for finding things out. A press release that gives such thin treatment of how a discovery was made, in favor of discussing what the discovery was, fails in its fundamental task of informing the general public.

The model behind this press release may be that laypeople are little people who live outside the University and can’t really understand what is being done, but might have a chance at understanding the conclusions. Or it could be that the public relations writer who wrote the press release² couldn’t understand, or didn’t have time to understand, the reasons why the research led to the conclusions it did.

Whatever the reason, this press release crosses the line between “providing an elegant explanation suitable for laypersons,” and “dumbing down science.” This is an extreme example of dumbing down, because the release appears to make the dual assumption that not only will the public not understand the reasons - however well expressed - but they also will find the principles behind the mathematical model too “sophisticated” to understand.

I call this the Moses model - some bearded guy on a mountain conveying the results of his research on tablets of stone to the masses below. What is demanded of all researchers (and their PR collaborators) in this environment of deplorably poor science education is to provide compelling examples of the scientific method in action, and compelling explanations of their research.

Wondering if I could improve upon the press release, I went searching for the paper. I found it here; and for a measly ten bucks, I can purchase the privilege of being able to read the paper for two days. I’m not going to do that, because (a) the subject is outside my field of expertise, so I’m not likely to get as much out of it as someone more familiar with this field of research; and (b) I’m not actually working on an educational program on human-neandertal skull divergence. But I did read the abstract, and noticed the first line was this:

Recent research has shown that genetic drift may have produced many cranial differences between Neandertals and modern humans.

So, it turns out that this paper does not lead to the conclusion that human skull evolution was driven by genetic drift as opposed to natural selection; it’s the other way around: the conclusion led to the paper. The conclusion was raised as a possibility by previous research, and this paper provides a test of the hypothesis. If only I could be cited back to that previous research, perhaps that abstract would further illuminate me. But I don’t know where to look, because (as is largely customary in abstracts), there is no citation.

The abstract again:

Close correspondence between cranial and DNA-sequence results implies that both datasets largely, although not necessarily exclusively, reflect neutral divergence, causing them to track population history or phylogeny rather than the action of diversifying natural selection.

Now this is fine for an abstract - if you want to learn the reasons why this correspondence is evidence for genetic drift, you are supposed to continue on and read the paper.

But this kind of thing is not fine for a press release, which must provide an accessible explanation of why the scientists believe the things they are asserting. Without doing this, the press release is useless as a tool to increase public awareness or education about the subject. It is far more difficult to write a press release than a research paper abstract, and the system that generated this release has had a major malfunction. However obvious the conclusions are to the research team, they are not going to be obvious - nor necessarily even interesting - to a layperson reading the press release.

I have a psychic³ prediction to make: this paper is going to have virtually no penetration into the public awareness. But it could have had widespread penetration, and it could have been an important event in educating the general public about human evolution, if only some kind of explanation of the conclusions had been offered that the average person could not only understand, but embrace as interesting and logical. As it stands, UC Davis gives us only a dry set of assertions.

Fail.

1. For no good reason, admittedly. [↩]
2. Most press releases are not written by anyone on the research team; the research people provide information to the PR department of their university, and then it is largely out of their hands. [↩]
3. Not! [↩]

Andrew Kitchen, Michael M. Miyamoto, and Connie J. Mulligan report in PLoS-ONE on their development of a three-stage model for the colonization of the Americas by Homo sapiens. This issue is of deep interest to anthropology outreach in Alaska, and I’m accordingly very interested in the paper. The attention that these ideas will likely receive in Alaska suggests several major avenues for effective public outreach:

• It provides an opportunity for “what is the nature of science and knowledge” education. The concepts of falsifiability and refinement of knowledge over time are particularly rich opportunities with these new results.
• It provides an opportunity to provide some “cutting edge” science to students. As noted below, many of the interpretive materials in greater Anchorage on these subjects reflect what was “cutting edge” twenty years ago, but which is now largely rejected in paleoanthropology.
• This paper is largely about analysis of genetic populations, and statistics. Therefore, it is an open door to talk about mutation rates and evolution, and some simple statistical exercises could easily be devised to give students an idea of what the authors are doing in their analysis.
• It provides an example of multidisciplinary work in science. The authors present a genetics analysis but subject it to controls imposed from other fields.
• Because some broadly similar studies of the past have not been subject to those controls, it provides an example of why there might be apparent disagreement about knowledge amongst scientists. For example, I’ve heard about genetic data that supports migration into the Americas both much earlier, and significantly later, than well-dated archaeological sites. By not imposing constraints from other fields of study, such findings result in apparent disagreement, without necessarily being valid disagreement. The distinction is worth teaching since organized antiscience uses such cases as a wedge.

The authors propose that the population of Amerind ancestors expanded out of east central Asia between 43,000 and 36,000 years ago, and occupied Beringia, the easternmost portion of Asia and the western part of Alaska, including the sea floor which was exposed at the time. A stable population of 8,000 to 10,000 people remained there from that time until around 16,000 years ago, at which time 1,000 to 5,400 of them rapidly expanded into the Americas. The study conforms to prior hypotheses that this expansion occurred either through an ice-free corridor in eastern Alaska and western Canada, or along the coast.

Consistent with other recent work, this paper proposes a single migration, as opposed to studies of the past that considered Amerinds, Na-Dene, and Eskimo-Aleuts to be the result of different migrations. This hypothesis gained popularity in the mid-1980’s, and is the model adopted by a number of interpretive materials in and around Anchorage. The model has been in disfavor for some time in the professional literature, and it seems likely that this new study would help to change these interpretive aids (assuming that scientific evidence trumps political expediency).

The authors point out that the genetic studies to date have strongly supported a single-migration model, but that they have varied significantly concerning the proposed date of the migration, with dates anywhere from about 13,000 years ago, to 40,000 years ago. As a result, that data has been interpreted by a variety of scenarios involving additional migrations, migrations of various ages, and so on. At least from the layman’s perspective, many of these seemed like clever possibilities that had the unfortunate air of being ad-hoc about them.

The new study accommodates some of the more puzzling aspects of the prior genetic studies, particularly ones that come up with very old dates of 30,000 years or more for the migration. A stable population in Beringia for some thousands of years would explain those results, and also explain why there are no American archaeological sites older than around 15,500 years old, while accommodating nicely the archaeological evidence that Homo sapiens was in northwest Beringia by about 30,000 years ago.

The study incorporates data from both nuclear and mitochondrial DNA of both Native Americans and Asians. Mitochondrial DNA evidence was cited in 2005 (with quite a bit of publicity, at least in Alaska) to support the idea that the population colonizing North America was extremely small, so it appears to me that the re-analysis of the mitochondrial data is of particular interest. Also of interest is that this study, unlike some in the discipline, uses archaeology, geology, and paleoecology as opportunities for imposing controls on the analysis of the genetic data. Some of the genetics studies of the past have given the appearance of being statistical analyses that avoided giving very much consideration to what is known from other disciplines. The study constrains divergence time to 15,000 years ago, and by trying out different migration rates between Asia and Beringia (and back), it is shown that the lower (and “more biologically realistic,” as the authors put it) the migration rate the larger the population of Amerind ancestors:

Our results demonstrate that smaller estimates of Ne depend upon a substantial level of migration from Asia to account for present-day levels of Amerind genetic diversity, e.g. Hey’s estimate of ≈70 founders is associated with a mAsia→NW > 9.0, which is twice the migration rate for contemporary Europe (m = 4.3).

Emphasis mine. I agree with the authors that the high migration rates assumed by other studies are implausible. Intuitively, I have a hard time accepting that the rate of migration on a modern industrial continent serviced by jets and trains is substantially lower than that found in east Asia in the Pleistocene, but I’m not an expert.

The authors also build into the paper a very nice opportunity for those doing outreach to talk about “what is science:”

Our goal is to provide a comprehensive model for the initial settlement of the Americas that generates new testable hypotheses and has high predictive power for the inclusion of new datasets. In light of our results, we propose a three-stage model in which a recent, rapid expansion into the Americas was preceded by a long period of population stability in greater Beringia by the Paleoindian population after divergence and expansion from their ancestral Asian population.

In other words, science produces conclusions that are testable. When you come to a conclusion, you are sticking your neck out a bit - because by definition a scientific finding is subject to being disproved at some point by someone who has better data, or is better at interpreting your data than you are.

One of the most interesting aspects of this paper, from an outreach perspective, is the opportunity to discuss how we know the dates. Here in a single paper are incorporated various methods for dating prehistoric events and materials (carbon dating, stratigraphy, genetic statistics, and surely a few others), and all of the methods agree that this recent event in world geological history still took place thousands of years before some believe the world was even created. The contrivances that are required to refute these vastly different, yet mutually-supporting dating techniques are awesome in their implausibility, and that’s where the teaching opportunity comes from.

–

This blog article is about:

Kitchen, A., Miyamoto, M.M., Mulligan, C.J., Harpending, H. (2008). A Three-Stage Colonization Model for the Peopling of the Americas. PLoS ONE, 3(2), e1596. DOI: 10.1371/journal.pone.0001596