Following on my post from yesterday, I wanted to write a little more specifically about the significance of scale and our session at the AAAs.
Paleoanthropological data are produced and address questions at different scales. Consider the following:
Fossil – Fossils provide data at two ends of the scalar spectrum. First, fossils provide fine-grained data about the life-events of individual specimens. For example, the sub-adult fossils from Dmanisi (D2700+associated mandible and post-cranial elements) provides a lot of information specific to that individual. We can infer patterns of dental eruption, epiphyseal fusion, osteologically-preserved life-history events, dental and skeletal pathologies and possibly behavioral features such as the individuals diet. That is a lot of potential fine-scale data. Fossils also provide coarse-grained data about population/species level patterns of variation. We use fossils as not just individual specimens, but as representatives for fossil populations and/or fossil species. Here, the fine-grained individual data can translate directly into more coarse-grained species/population data, such as individual dental eruption serving as a proxy for population or species-specific patterns of eruption. The coarseness in the data are the result of the uncertainty in translating from the individual to the population or species level. Individuals vary. In lots of ways populations vary, too. In some ways, species vary. How these three levels relate in living populations is something that can be observed and recorded, but in the fossil record is a hypothesis that can (and should) be tested.
Adding additional coarseness to fossil data, at both the individual and population/species level is the temporal uncertainty associated with any given find. Unfortunately we do not have direct dates on a large fraction of the fossil record. Even those that do have directly dated contexts, do so with a methodologically-associated degree of uncertainty. Depending on the questions of interest, this temporal uncertainty is important.
Archaeology – Archaeological data is generally finer in scale than fossil data, spanning a successive range of time-sequenced layers. These layers generally represent a time-averaged accumulation of materials, whose temporal precision is a product of the specific site formation and sedimentological processes. In rare instances, archaeological data might provide fine-scale data approaching an instantaneous view of a horizontal spatial expanse, or so-called “living surface” (though the interpretations of sites as living surfaces likely outnumbers the reality of living surfaces by a large margin). Compared to fossils, however, archaeological data typically provide a view of an evolutionary sequence of time rather than a single point in time.
Genetics – Genetic data are interesting because the scalar perspective they provide depends on their source. Genetic data drawn from living people provide a window on our evolutionary past by revealing a complex pattern of inheritance going back in time. Using coalescent methods, this pattern can be seen as a semi-scaled, probabilistic window into the past, one that becomes fuzzier the deeper into the past you try to look. I say semi-scaled because the data represent genealogical relationships that require some calibration, usually an empirically-derived or estimated mutation rate (and/or recombination rate).
Increasingly, genetic material derived from prehistoric skeletal and fossil remains inform our understanding of human evolution. These data are particularly interesting because they can be used to look backwards (as we do with contemporary DNA) or forwards, providing an alternative evolutionary scaling mechanism by which to assess genetic variation in living humans. Much of the support for the Neandertal Genome Project, for example, was to narrow the range of evolutionary time encompassed by “human-specific” lineage changes to the period of time following a human-Neandertal divergence rather than the much larger window reflecting the divergence of fossil hominins from our (terminal Miocene/Pliocene) Pan-Homo ancestor.
There are many other lines of evidence I could mention (climatic, environmental, behavioral…), but even drawing from just the three above the potential complications associated with integrating these data, and the potential rewards for doing so, can become clear.
Getting back to our session, several of our presenters made scale an explicit focus. Jamie Clark, for example, talked about the differing interpretations of environmental change at Sibudu Cave, South Africa, depending on the scale at which you divide the data. Sibudu encompasses a long Middle to Late Pleistocene archaeological sequence that includes the appearance of a complex Howiesons Poort industry and the subsequent rapid demise and replacement of this industry by a less complex, more MSA-typical assemblage. This loss of complexity has been argued to be the result of an environmental change at the site, and viewed at a coarse scale, this argument is supported by the faunal and paleobotanical remains. If the data are examined at a level by level scale, however, it can be seen that the Howiesons Poort disappearance precedes the most dramatic environmental shift, presumably supporting a very different kind of argument for the archaeological transition.
More broadly, many of the talks (mine included), focused on the benefits from looking at Pleistocene evolutionary questions from a population, rather than paleontological, perspective. Such a view emphasizes the importance of community niche construction (AgustÃn Fuentes’ talk), population expansion and contraction (Luke Premo’s talk), and population structure (John Hawks’ talk).
My head is swirling in a bunch of nascent Thanksgiving Day preparations, so I will cut myself off now. But I would just like to conclude by saying that these issues of scale are a challenge. As Steve Kuhn wrote recently, the scale of our explanations needs to match the scale of the phenomena we are observing in the evolutionary record. But these issues of scale also represent an opportunity. They represent an opportunity to add greater strength to evolutionary arguments by creatively using scale, particularly changing scales of perspective, to circumvent problems of equifinality. In my talk, one of my conclusions was that the pattern of morphological variation in Neandertals and fossil humans provides support for very few speciation events in our recent (i.e. the Pleistocene) evolutionary past. But I also argued that such a view emphasizes population-level processes that cannot be adequately addressed with single lines of evidence but instead demand validation from different sources of data and perspective.
