This paper advocates for a reflexive approach to using radiocarbon datasets for studying prehistoric dispersals in lowland South America. It emphasizes considering data quality, chronometric uncertainty, and hypothesis testing. Using the Arauquinoid and Tupiguarani traditions as examples, the authors develop an analytical solution that incorporates chronometric uncertainty in bivariate regression and tests for statistically significant evidence of dispersal. The analysis reveals issues with resolution and data quality hindering dispersal identification through simple spatial gradients. The findings suggest a need for reflexivity in theoretical frameworks of prehistoric mobility, informing the creation of more critical statistical null models and consideration of alternative cultural expansion models beyond demographic models.

Philip Riris, Fabio Silva