University of Southampton OCS (beta), CAA 2012

Font Size: 
Philip Verhagen, Laure Nuninger, François-Pierre Tourneux, Frédérique Bertoncello, Karen Jeneson

Last modified: 2011-12-16


In this paper, we want to present the results of a study into integrating socio-cultural factors into predictive modelling (PM). So far, PM has largely neglected the social and cultural dimensions of past landscapes. Input is commonly derived from correlations between archaeological sites and natural landscape features. The result is a rather static way of modelling, in which the human factor remains unexplored. Furthermore, issues of temporality have been addressed uncritically or insufficiently. To maintain its value for archaeological research, therefore, PM needs new methodologies, concepts and theories.

For this study, we have departed from the methodology developed in the 1990s during the Archaeomedes Project (Favory et al. 1999, van der Leeuw 1998; van der Leeuw et al. 2003; Nuninger et al. 2007). By analyzing the surroundings of Roman settlements in the French Rhône Valley, it proved possible to make cross-regional comparisons of settlement location factors, in particular slope, aspect, solar radiation and soil type. However, at the time this method was not used for PM purposes, and did not include an analysis of non-environmental factors. For the current research, we modified this approach, and analyzed not just the surroundings of the known Roman settlements, but of every pixel in three different study regions, Vaunage (Languedoc, France), Argens-Maures (Provence, France) and Zuid-Limburg (Netherlands). Furthermore, the set of variables analyzed was expanded with a number of ‘socio-cultural’ factors, in particular accessibility, visibility, and the effect of previous occupation (the ‘memory of landscape’). From this, we created predictive models using a combination of principal component analysis and cluster analysis. By making a side-by-side comparison of predictive models based on environmental variables, and models based on a combination of environmental and socio-cultural ones, we could show that, at least for this particular archaeological context, the inclusion of socio-cultural variables will increase the predictive power of the models. More importantly however, we have developed a protocol for PM using both environmental and socio-cultural factors that can easily be implemented for different regions and time periods.


Favory, F., Girardot, J.-J., Nuninger, L., Tourneux F.-P. 1999. ARCHAEOMEDES II : une étude de la dynamique de l’habitat rural en France méridionale, dans la longue durée (800 av. J.-C.-1600 ap. J.-C.). AGER 9, 15-35.

Nuninger, L., Tourneux, F.-P., Favory, F. - From Archaeomedes to Archaedyn. In: Poslushny, A., Lambers, K., Herzog, I. eds. – Layers of perception, proceedings of the 35th Computer Applications and Quantitative Methods in Achaeology Conference, Berlin, Germany, April 2-6, 2007, Bonn (Germany): Habelt, 2008 (Kolloquien zur Vor- und Frühgeschichte, vol. 10) : p. 278 and CD.

Van der Leeuw, S.E. (Ed.) (1998). The Archaeomedes Project - Understanding the natural and anthropogenic causes of land degradation and desertification in the Mediterranean. Luxemburg: Office for Official Publications of the European Union.

Van der Leeuw, S., Favory, F. & Fiches, J.-L. (Eds.) (2003). Archéologie et systèmes socio-environnementaux: études multiscalaires sur la vallée du Rhône dans le programme ARCHAEOMEDES.  Valbonne: Presses du CNRS (Monographies du CRA).


predictive modelling; socio-cultural variables