# Unpacking lithic assemblage variability in the Early Upper Palaeolithic: A multivariate approach to the structure of the Iberian Aurignacian

## TL;DR

This paper uses statistical methods to analyze stone tool variability in Iberia during the Early Upper Palaeolithic, revealing patterns linked to geography rather than time.

## Contribution

The study applies multivariate analysis to Iberian Aurignacian assemblages, revealing spatial patterns of variability not explained by time.

## Key findings

- Inter-assemblage differences increase with spatial distance but not with temporal distance.
- The findings challenge the Aquitaine model's applicability to Iberian Aurignacian variability.
- Diagnostic traits are heterogeneously represented across and within temporal classes of assemblages.

## Abstract

The Aurignacian technocomplex of the Early Upper Palaeolithic remains a long-standing focal point for understanding the expansion of modern humans across Europe. Diagnostic assemblages occur across vast swathes of the continent, suggesting the existence of broadly connected groups and traditions around 43–32 ka cal BP. However, while its extensive distribution is often regarded as proxy evidence for the spread of modern human groups, artefact assemblages are known to be synchronically and diachronically variable in ways that reveal an inconsistent representation of diagnostic traits. In the Iberian Peninsula, this variability is exemplified by an idiosyncratic material record in which diverse Aurignacian assemblages occur alongside undiagnostic or ‘culturally indeterminate’ ones, leading many Aurignacian occupations to be disputed. In this paper, we assimilate this regional record through quantitative analyses of techno-typological attributes from all sufficiently published and chronologically relevant assemblages of the Early Upper Palaeolithic. Using two multivariate techniques, we first explore associations between assemblages and thereafter test whether inter-assemblage variability is related to spatial and temporal distances. Our results cast light on the spatial structure of variability by revealing that inter-assemblage differences increase with spatial distance but show no linear relationship to temporal distance. This spatial finding challenges the cross-regional applicability of the Aquitaine model of techno-typological change, whilst the absence of temporally structured variability suggests a heterogeneous representation of diagnostic traits across and within temporal classes of assemblages.