Using the radiocarbon dates of Central Africa for studying long-term demographic trends of the last 50,000 years: potential and pitfalls

This paper presents the first review of biases impacting Pleistocene and Holocene radiocarbon dates from Central Africa. Based on the pooling of the research expertise of the co-authors, twenty-four biases are listed, explained and documented and their impact on any radiocarbon date corpus demonstrated. To achieve this, a new corpus has been created of 1764 radiocarbon and TL assays from 601 archaeological sites published in the literature. Each date has been checked for its context. The irregular dynamics of research in space and time seriously impact the end result of previous analyses aiming to achieve a regional understanding of past demographic fluctuations. While peaks in the number of dates from the late Holocene seem to correspond to a positive demographic trend, it is suggested that the declines identified cannot be of any such use for the time being and that today’s picture does not presently support claims of a population “crash” at a regional or local level for any time period. The numbers are obscured by overall research deficits identifiable throughout the region. The maps of the dated sites presented offer good evidence of this and illustrate the vast expanses where no archaeological research has yet been carried out. The number of radiocarbon dates in Central Africa is more an indicator of the effort archaeologists have put into understanding a settlement than it is of ancient demographics. Successive waves of incoming people since c. 3500–3000 cal. BP, the two most important ones known since the 1990s, have created a cultural mosaic of coexisting technological groups. The last 40 years of research have revealed the inner complexity of these waves, some of which avoided parts of the region for centuries, thereby creating an irregular cultural mosaic of land use that is outlined by patterning in the radiocarbon dates.