Disentangling faunal skeletal profiles. A new probabilistic framework

Abstract

Faunal skeletal profiles from archaeological assemblages have been long analysed regarding differential transport of carcasses to infer hunting preferences, human mobility, or even dietary stress. However, the existence of several possible accumulating agents, together with the effect of bone attrition, is known to introduce a potential bias, thus hindering the possibilities of meaningful concussions. In order to overcome this problem, several methods were proposed during the late 90s and early 2000’s, although a consensus was not reached, mainly because the different approaches were based on a certain initial hypothesis that significantly affected the output. Building on that previous experience, a new methodological framework is proposed and compared here. Moving from rather deterministic techniques, a Bayesian alternative approach based on a Monte Carlo Markov Chain sampling is presented and applied to several ethnographic and Pleistocene key sites. This new method makes use of the available information to constrain the possible degrees of attrition and carcass processing strategies, leading to easily comparable results.

Keywords:

• Skeletal profiles
• Bayesian inference
• equifinality
• prey transport
• attrition

Acknowledgments

We would like to thank Jean-Philip Brugal and Christiane Denys the invitation to the 4th ICAZ Taphonomy Working Group meeting held in Paris, September 2016. The authors want to acknowledge to Prof. R. Foley and M. Mirazon Lahr for their support. The code developed for this research in MATLAB is available upon request.