I spent a good bit of time this past week reading A. Wilson and A. Bowman’s newest edited volume in the Oxford Studies on the Roman Economy series. The book is titled Settlement, Urbanization, and Population and is a follow up to their 2009 volume, Quantifying the Roman Economy (for a quick set of reflections see here). Like the 2009 volume, Settlement, Urbanization, and Population continues a strong commitment to quantitative methods in reconstructing patterns in the ancient economy. The contributors to the Bowman and Wilson volume see the Roman economy as inexorably tied to larger trends in demography and settlement across the empire. As a result, they applied quantitative methods to population and settlement patterns as much as, say, efforts to reconstruct trade or taxation levels across the empire. The very reasonable perspective shared by almost all the contributors is that population size contributes to the structure of settlement which, in turn, informs the degree of economic organization necessary in the Mediterranean.
Efforts to estimate ancient populations are not new and, in fact, date back to 19th century efforts by scholars like Julius Beloch. In recent times, the rapidly growing body of archaeological data – particularly that produced by landscape and region survey projects – would appear to offer a new, more robust body of data upon which to base conclusions. Ironically, this data remains problematic for estimating population as scholars have yet to agree on such basic matters as urban population density.
From a methodological perspective, things are even more muddled, as the nature, size, and chronology of non-urban sites documented through intensive survey procedures remains open to significant dispute. The first group of contributions in the Wilson and Bowman book attempt to address some of these concerns while remaining faithful to idea that the archaeological material on the surface of the ground can produce meaningful information regarding specific types of settlement. They continue to see survey data as producing villas, farmsteads, and villages in the countryside and to discuss “site recovery rates’ as a useful metric for considering the relationship between a particularly group of rural sites and the total number of sites existing at any one time.
It was interesting to see very little attention paid to the work of so-called “siteless survey” which regards the artifact as the basic unit of analysis. Unlike survey projects that look for sites in the landscape, siteless survey sees the surface assemblage as a object of study and accepts the inherent ambiguity in the size, chronology, and formation processes of individual variations within the overall distribution of artifacts on the surface. In other words, siteless survey challenges the existence of so-called “sites” – with all the various aspects of human activities. This means, by extension, that analyses of population grounded in recognizing sites as unproblematic loci of human activity has certain significant theoretical limitations grounded in the very assumptions used to generate certain types of survey data. This perhaps explains the absence of data generated from survey projects in Greece where the growing commitment in siteless survey is most pronounced.