Quantifying differences in hominin flaking technologies with 3D shape analysis.

Genetic and climate-driven estimates of past population dynamics are increasingly influential in broader models of hominin migration and adaptation, yet the contribution of stone artifact variability remains more contentious. Scientists are increasingly recognizing the potential of unretouched stone flakes ('flakes') in exploring existing models of hominin behavioral evolution. This is because flakes (1) were produced by all stone tool manufacturing groups in the past, (2) are abundant from the inception of the archaeological record up into the ethnographic present, and (3) preserve under most conditions. The statistical tools of 3D geometric morphometrics capture detailed approximations of flake form that are challenging to document with conventional artifact analyses. We analyze a collection of 717 3D scans of experimentally produced flakes from 5 production strategies that were practiced by hominins through large parts of the Pleistocene and that scientists have drawn on also to make demographic arguments about past human behavior (n = 45 reduction sequences, n = 3 knappers naive toward the study objectives). First, as a proof of concept, we demonstrate that we can estimate the strategies used to produce these flakes at a high success rate even when flakes from early stages of core reduction are included. We frame the significance of this finding against archaeological classifications from several key Middle Paleolithic assemblages in France (n = 4 sites, n = 28 layers, n = 16,467 flakes). Second, we show that 3D geometric morphometrics captures subtle differences in these strategies that influence flake formation on a flake-by-flake basis and that reflect decisions made by knappers about platform selection, preparation, and core-surface management. We explore the broader potential of our model with a cross-validation approach, and we describe a means of assessing flake form on a continuum wherein variability among assemblages separated by large expanses of space and time can be meaningfully explored.

Optimal linear estimation models predict 1400-2900 years of overlap between Homo sapiens and Neandertals prior to their disappearance from France and northern Spain.

Recent fossil discoveries suggest that Neandertals and Homo sapiens may have co-existed in Europe for as long as 5 to 6000 years. Yet, evidence for their contemporaneity at any regional scale remains highly elusive. In France and northern Spain, a region which features some of the latest directly-dated Neandertals in Europe, Protoaurignacian assemblages attributed to Homo sapiens appear to 'replace' Neandertal-associated Châtelperronian assemblages. Using the earliest and latest known occurrences as starting points, Bayesian modelling has provided indication that these occupations may in fact have been partly contemporaneous. The reality, however, is that we are unlikely to ever identify the 'first' or 'last' appearance of a species or cultural tradition in the archaeological and fossil record. Here, we use optimal linear estimation modelling to estimate the first appearance date of Homo sapiens and the extinction date of Neandertals in France and northern Spain by statistically inferring these 'missing' portions of the Protoaurignacian and Châtelperronian archaeological records. Additionally, we estimate the extinction date of Neandertals in this region using a dataset of directly-dated Neandertal fossil remains. Our total dataset consists of sixty-six modernly produced radiocarbon determinations which we recalibrated using the newest calibration curve (IntCal20) to produce updated age ranges. The results suggest that the onset of the Homo sapiens occupation of this region likely preceded the extinction of Neandertals and the Châtelperronian by up to 1400-2900 years. This reaffirms the Bayesian-derived duration of co-existence between these groups during the initial Upper Palaeolithic of this region using a novel independent method, and indicates that our understanding of the timing of these occupations may not be suffering from substantial gaps in the record. Whether or not this co-existence featured some form of direct interaction, however, remains to be resolved.

Introducing platform surface interior angle (PSIA) and its role in flake formation, size and shape.

Four ways archaeologists have tried to gain insights into how flintknapping creates lithic variability are fracture mechanics, controlled experimentation, replication and attribute studies of lithic assemblages. Fracture mechanics has the advantage of drawing more directly on first principles derived from physics and material sciences, but its relevance to controlled experimentation, replication and lithic studies more generally has been limited. Controlled experiments have the advantage of being able to isolate and quantify the contribution of individual variables to knapping outcomes, and the results of these experiments have provided models of flake formation that when applied to the archaeological record of flintknapping have provided insights into past behavior. Here we develop a linkage between fracture mechanics and the results of previous controlled experiments to increase their combined explanatory and predictive power. We do this by documenting the influence of Herztian cone formation, a constant in fracture mechanics, on flake platforms. We find that the platform width is a function of the Hertzian cone constant angle and the geometry of the platform edge. This finding strengthens the foundation of one of the more influential models emerging from the controlled experiments. With additional work, this should make it possible to merge more of the experimental results into a more comprehensive model of flake formation.