The Contribution of 2D and 3D Geometric Morphometrics to Lithic Taxonomies: Testing Discrete Categories of Backed Flakes from Recurrent Centripetal Core Reduction

Abstract

AbstractPaleolithic lithic assemblages are usually dominated by flakes and display a high degree of morphological variability. When analyzing Paleolithic lithic assemblages, it is common to classify flakes into categories based on their morphological and technological features, which are linked to the position of the flake in the reduction sequence and how removals are organized in a given production method. For the analysis of Middle Paleolithic lithic assemblages, two categories of flakes are commonly identified: core–edge flakes and pseudo-Levallois points. A third type, core–edge flakes with a limited back, is also commonly found in the archaeological literature, providing an alternative category whose definition does not match the two previous types but shares many of their morphological and technological features. The present study addresses whether these three flakes constitute discrete categories based on their morphological and technological attributes. 2D and 3D geometric morphometrics are employed on an experimental set composed of the three categories of flakes to quantify morphological variation. Machine learning models and principal components biplots are used to test the discreteness of the categories. The results indicate that geometric morphometrics succeed in capturing the morphological and technological features that characterize each type of product. Pseudo-Levallois points have the highest discreteness of the three technological products, and while some degree of mixture exists between core edge flakes and core edge flakes with a limited back, they are also highly distinguishable. We conclude that the three categories are discrete and can be employed in technological lists of products for the analysis of lithic assemblages and that geometric morphometrics is useful for testing for the validity of categories. When testing these technological categories, we stress the need for well-defined and shared lithic analytical units to correctly identify and interpret the technical steps and decisions made by prehistoric knappers and to properly compare similarities and differences between stone tool assemblages. These are key aspects for current research in which open datasets are becoming more and more common and used to build interpretative techno-cultural models on large geographical scales. Now more than ever, lithic specialists are aware of the need to overcome differences in taxonomies between different school traditions.