A Simulation Approach to the Measurement of Change in Ceramic Frequency Seriation

Abstract

This paper examines the problems encountered by users of ceramic frequency seriation in distinguishing chronologically meaningful trends from the background stochastic “noise” inherent in all ceramic assemblages. Aspects of manufacture, breakage and discard behavior are found to be important in the creation of “noise” in the archaeological record. A computerized model (POTS), which simulates both chronological changes and background “noise” in the ceramic assemblages of hypothetical villages, is used to evaluate the effects of sample size on the chronological accuracy of seriated sherd deposits. It is concluded that in order to maximize the utility of frequency seriation as a chronological ordering device, users should be concerned with the statistical adequacy of their sherd samples and that additional experimentation is needed with artificial data sets generated under ethnographically controlled conditions.