부산대학교 도서관

The lengths of—and numbers of steps in—stone artifact production sequences are often used as proxies for the emergence of cultural complexity in our lineage. This line of reasoning has been applied to the heat treatment of silcrete rocks by past hominins, where both the lengths and numbers of steps involved are contested. To simplify, one position argues that silcrete heat treatment was an elaborate, multi-level procedure involving the prolonged and indirect exposure of these rocks to fire. Each stage of production (building a fire, burying rocks at a specific depth below the fire, maintaining prolonged insulated heat, managing gradual cooling, and making or shaping blanks) was potentially reliant on the stage that came before. Others argue that silcrete heat treatment was far simpler, involving fewer steps (direct exposure of blanks to embers and/or flames in an open-air burning context), and that this process was managed by hominins no differently to everyday fire-related activities. A more nuanced position draws on the variable fabric texture and structural response to heating of different silcretes to argue that no single approach to heating may be suitable for all types. Determining maximum temperature exposure and/or heating rate would potentially allow one to differentiate these processes as described and thus resolve the implications of heat treatment for cultural complexity. Here, we present the results of experiments wherein multiple specimens of two structurally different South African silcretes were systematically heated to a controlled set of temperatures ranging from 110 to 750 °C, and the infrared responses to each temperature change were quantified. Importantly, measurements were collected non-destructively on stone flakes, providing for the archaeological applicability of the results to the Middle Stone Age record of southern Africa. We then use a combination of multivariate statistics, differential equations, and decision trees to develop a framework for reconstructing the annealing temperatures of silcrete flakes, demonstrating that heating temperature range can be predicted at success rates of 79.5 to > 95% in a large sample of N = 225 observations.