Detecting the effects of selection and stochastic forces in archaeological assemblages

Detecting and diagnosing the causes of change through time in archaeological assemblages is a core enterprise of archaeology. Evolutionary approaches to this problem typically cast the causes of culture change as being either stochastic in origin, or arising from selection. Stochastic sources of change include random innovations, copying errors, drift and founder effects among dispersing groups. Selection is driven by differences in payoffs between cultural variants. Most efforts to identify these evolutionary forces in the archaeological record have relied on assessing how well the predictions from a neutral-stochastic model of cultural transmission fit a data set. Selection is inferred when the neutral-stochastic model fits poorly. A problem with this approach is that it does not test directly for the presence selection. Moreover, it does not account for the fact that both neutral-stochastic and selective forces can act at the same time on the same cultural variants. A different approach based on the Price Equation allows for the simultaneous measurement of selective and stochastic forces. This paper extends use of the Price Equation to the analysis of selective and stochastic forces operating on multiple artifact types within an assemblage. Ceramic data presented by Steele et al. (2010, Vol. 37(6): 1348–1358) from the Late Bronze Age Hittite site of Boğazköy-Hattusa, Turkey, provide an opportunity to evaluate the efficacy of this model. The results suggest that selection is a dominant process driving the frequency evolution of different bowl rim types within the assemblage and that stochastic forces played little or no role. It is also clear, however, that we should be attentive to combinations of direct and indirect selective effects within assemblages consisting of multiple artifact types.