When anthropological and geological facts collide

This passage is the first paragraph of the introduction to Franz Weidenreich's monograph, The Skull of Sinanthropus pekinensis Weidenreich:ZKD:1943.

In my earlier contributions to the study of Early Man I pointed out repeatedly the danger of confusing anthropological facts with geological facts. In determining the character of a given fossil form and its special place in the line of human evolution, only its morphological features should be made the basis of decision; neither the location of the site where it was recovered nor the geological nature of the layer in which it was imbedded [sic] are important. Discrepancies cannot be smoothed out by bringing morphological facts and opposing geological data into closer harmony with artful interpretations or by touching-up reconstructions. It is a generally accepted conception that Man has developed in the course of time by gradual transformation from an ape-like type to the type he presents today. Viewed from this fundamental standpoint, it is logical to assume that the more a form resembles the supposed ancestor the more ancient it will be, or that the more ancient it is the more "primitive" it should be.

I am concerned with this passage today because of a re-emerging mismatch of evidence from the morphology of Middle Pleistocene humans and the genetics of Neandertals. Some paleoanthropologists have asserted that Europeans of the Middle Pleistocene were the exclusive ancestors of Neandertals. I have in the past written that Middle Pleistocene Europeans were among the ancestors of Neandertals, with sustained gene flow from other populations including Africa accretion. The Sima de los Huesos people, maybe 600,000 years old, resembled the (much) later Neandertals in several aspects of their anatomy, as did other Middle Pleistocene Europeans.

The genetic differences between living people and the ancient Neandertal genomes appear consistent with the emergence of distinct African and Neandertal populations only within the last 400,000 years or less Green:draft:2010, Reich:Denisova:2010.

Such a recent date seems a poor match for the morphological evidence of Neandertal ancestry in Europe. I can think of several ways to make these morphological and genetic comparisons concordant with each other, all of which balance some shift in one body of inference against the other. As long as we can't pin down the human mutation rate within a factor of two ("What is the human mutation rate?"), there's a lot of room to make different population models consistent with the genetic data.

This is, in today's language, Weidenreich's point. Morphological data must be interpreted in accordance with evolutionary principles, and if it doesn't fit a temporal scheme, it doesn't fit. Likewise, genetic similarities must be explained in their own evolutionary framework. These two sources of evidence must in the end be consistent with a single history. We will find that consistency not by shoehorning the evidence together, but by interpreting each with the strongest possible skepticism concerning assumptions and models.

Weidenreich's introduction illustrates two cases. The simpler, from our point of view today, was Piltdown. Many establishment anthropologists, particularly in Britain, had maintained that Piltdown was a morphologically advanced ancestor of modern humans, which had lived early in the geological record of human evolution. Weidenreich had been an early and prominent critic of this idea, because he was convinced that the specimen simply did not fit together with its supposed geological context.

I cannot believe, even making very liberal allowances for these uncertainties, that such incongruity between morphology and chronology as is found in the case of Piltdown can be completely brought into accord. The only hope of solution in this case would lie in assuming that the human bones were not contemporaneous with the layer in whih they were found but were deposited there later. Otherwise, modern man must be much more ancient than we ever imagined, or else Western European man did not pass through evolutionary stages as did the hymans of other regions of the earth.

We now know, of course, that Weidenreich was entirely correct. The apparent geological facts were false, and the "advanced" characters of the specimen were simple reflections of the fact that the skull is a modern human skull.

The other problem Weidenreich discussed in some detail was the phylogenetic position of the Steinheim skull. Like Piltdown, this specimen had been placed in a Presapiens context by other workers. Steinheim lacks most of the derived characteristics of later Neandertal specimens. Weidenreich, along with many of his contemporaries, accepted its lack of Neandertal features as evidence for affinity with modern humans. In Weidenreich's view, this similarity with modern humans was "anachronistic". Even so, the case did not challenge an evolutionary interpretation, only the assumption that features could evolve from "primitive" to "modern" along a single line. If we admit that Neandertal features were not in all cases "primitive", even if they may resemble superficially the characteristics of some apes, we can accommodate specimens like Steinheim within a population model where both moderns and Neandertals may have derived (and in some cases, secondarily derived) characters that appeared afterward.

This scenario requires us to straighten out the analysis of the characters themselves, a process for which larger fossil samples are essential. It was to that end that Weidenreich supposed the Sinanthropus sample to be of such great utility. The subtext of the introduction was to illuminate the kinds of evolutionary problems that could be further illuminated by a full description of fossil variation. Finding variation in fossil humans did not repudiate the concept that modern humans had evolved in stages from primitive ancestors, but helps to clarify cases where the evolution has not been a simple linear progression. In many cases, features that are superficially "primitive" may actually have been secondarily derived in recent humans compared to earlier hominins.

Along similar lines, I ran across this old post: "Dobzhansky on Weidenreich's species concept", in which Dobzhansky predicts:

Some modern populations may carry genes that were present in the Neanderthaloids, and other moderns may not carry such genes.