Geochemical data treatment and interpretive uncertainty: a reanalysis of Stonehenge stone fragments (‘Debitage’)

Abstract

This paper centers on the reinterpretation of previously published geochemical data from Stonehenge stone fragments, material often disregarded in geoarchaeological investigations, to highlight the importance of reflexive evaluation of analytical choices and their consequences for archaeological interpretation. Because of the large spreads in many elemental concentrations, earlier interpretations relied on elemental concentration ratios, with Zr as the normalizing element. Here, the same datasets are re-examined using non-normalized elemental concentrations and a visual/graphical, geochemically informed approach that explicitly recognizes analytical variability and interpretive uncertainty. The reanalysis indicates that, while Zr-normalized data may reveal some geochemical similarities, normalization can also obscure compositional differences, generate apparent agreements not supported across multiple elements, and fail to identify other coarse geochemical groupings present in the data. Using an alternative, non-normalized approach, some stone fragment samples can be tentatively grouped with previously characterized sarsen sources, while many remain unsourced or appear to represent geochemistries not yet documented. These observations suggest that the geochemical variability of silcrete materials at Stonehenge may be greater than previously assumed and that normalization-based interpretations should be applied with caution. This variability is in line with recent interpretations that emphasize the use in Stonehenge of materials derived from different parts of Britain. Rather than proposing definitive provenance assignments, this study emphasizes the interpretive consequences of analytical choices and argues for caution in the application of normalization procedures in archaeometric provenance studies.