Prof Christopher Ramsey

Abstract:

Bayesian statistics provide a methodology for combining information from diverse sources, assuming that such information can be objectively quantified. Critically, the mathematical methods are also well suited to parameters that have non-normal probability distributions. In archaeological and related environmental contexts, there is usually a wide range of information that has bearing on chronology; this includes stratigraphic information, information on deposition processes and other constraints, which can be applied relative to the passage of calendar time. In addition, one of the most widely used dating techniques in archaeology, radiocarbon dating, requires a calibration process that inevitably leads to non-normal probability distributions. Bayesian statistics provide the ideal way to combine the information from these different sources and has been highly successful in improving the precision and accuracy of archaeological chronologies. The same is increasingly true in environmental science, where some of the methods first applied in archaeology are now finding application in different disciplines.

Abstract:

This paper presents formally modelled date estimates,within a Bayesian chronological frame-work, for the sequence of deposits and material at the great Neolithic tell of Vinca-Belo Brdo near Belgrade, Serbia, on the basis of the first excavations, led by Miloje Vasic (1908-1934). This is part of a three-strand approach to dating the occupation and finds from the tell that is being undertaken as part of The Times of Their Lives project. A total of 85 radiocarbon measurements are now available from known depths through the Vasic sequence at Vinca. A Bayesian chronological model is constructed, using a poisson-process depositional model.