ABSTRACT
Across Europe the period 45-40 ka (thousands of years ago) is associated with several technological changes, including the emergence of the Châtelperronian technocomplex in France and northern Spain. The Châtelperronian, stratigraphically located between the Mousterian and Aurignacian, is characterized by Upper Palaeolithic features, such as volumetric blade reduction, curved backed blades, end-scrapers, bladelets, bone tools and ornaments. Concurrently, repeated, though debated, associations with Neanderthal remains and Mousterian elements suggest a local technological development. Following recent critiques and cumulating technological studies, this paper provides data-driven contextualisations of the Châtelperronian and late Mousterian archaeological records and a primary comparative assessment of a major linking element, backed knives, to re-assess the origin of the Châtelperronian. The results demonstrate the challenging nature of the 50-35 ka record, with many interpretive problems caused by poorly recorded excavations, resulting in only 25 well-contextualised assemblages from the claimed 143 Châtelperronian find spots. These 25 assemblages facilitate more detailed chronostratigraphic and typo-technological assessments and show that the Châtelperronian has a homogenous set of technologies and tools. A similar evaluation of the late Mousterian indicates a wide-ranging late Neanderthal skill set, commonly including laminar blank production and backing. Further, conceptual similarities were noted both in blank selection and edge modification between Mousterian and Châtelperronian backed knives, alongside their near-absence in other, contemporaneous technocomplexes. A Europe-wide contextualisation shows that while the current coarse-grained record still allows for several potential scenarios, the data throughout this paper point towards a most parsimonious model of a Châtelperronian made by Neanderthals, with roots in the late Middle Palaeolithic technological skill set. However, this change seems triggered by early arrivals of modern humans either indirectly, through stimulus diffusion, or directly, after ca. 42 ka. Fully testing this model requires an ongoing focus on site formation and assemblage integrity, alongside in-depth analyses of recently excavated assemblages and existing collections.