Abstract

The last decade saw a renewed interest in prehistoric demography, a trend stimulated by the growing number of case studies reconstructing past population dynamics through summed-probability distribution of radiocarbon dates (SPD). On the one hand the increasing availability of 14C dates and the possibility to rely on absolute chronology is making SPD a universal toolkit for synthetic research, enabling large scale comparative analysis that was not previously not possible. On the other hand, critiques raise concerns regarding biases and assumptions related to this technique, cautioning its use or even dismissing the entire enterprise. In this paper, I will review the most recent methodological advances on this ‘dates as data approach’, how they solve some of the issues raised, and how those that still need to be tackled are deeper rooted and shared across other proxies of past population change such as floor-size area or site counts. I will conclude by highlighting the upcoming challenges, such as the formal integration of multiple proxies, the necessity of more nuanced middle-range theories, and the possible use simulation- based generative inference.