This weekend, I had the pleasure of reading Grace Erny’s very recent article from the Journal of Mediterranean Archaeology, “Statistical Approaches to Small Site Diversity: New Insights from Cretan Legacy Survey Data.” It represents yet another significant contribution to the study of intensive pedestrian survey data and to the potential of reasonably well-published legacy “data.”
Erny’s article looks at the result of seven intensive surveys in Crete and compares the rural sites dating to the Geometric to Hellenistic period documented by these surveys. Her arguments are complex, nuanced, and quantitative. She argues, among other things, that the over 300 small sites documented by these survey on Crete demonstrate considerable variation in artifact diversity even when controlling for variables that might impact site formation and variation in documentation practices. For example, Erny demonstrates that diversity seems to be independent of site size. At the same time, she notes that from larger samples (i.e. total number of artifacts) we should expect greater diversity of artifacts and that this relationship is logarithmic. In other words, diversity should plateau as sample size increases. While Erny admits that she does not know the total number of artifacts recovered from any of the sites that she studies, she plausibly argues that larger sites seem likely to have produced more artifacts (especially since increased artifact density is a factor for defining sites on these projects). This allows her to observe that diversity appears independent of site size.
(As an aside, it is interesting to note that many years ago, David Pettegrew, Dimitri Nakassis, and I made a similar argument, albeit in a less elegant way. We suggested that especially diverse assemblages of artifacts from units with low visibility [and correspondingly low artifact densities and small assemblages] may well reward greater attention [in part to compensate for the small sample size produced by inadequate surface visibility]. Erny notes that by limiting her sample to sites, this controls to some extent for the vagaries of both post-depositional processes [i.e.. the sample of visible sites on the surface represents a subset of all the sites in a region that remain discoverable] and recovery practices [i.e. since the study is based on sites qua sites, we needn’t get lost in the murky realm of site definition].)
This suggests that hierarchical interpretations of rural sites which assume most of these small sites represented primary production in the countryside. Instead, Erny’s argument, which is grounded on the survey data itself, indicates that rural sites likely served a wide range of functions or reductive categories like “primary production” or “habitation” or “farmsteads” are inadequate to capture the diversity of activities in the countryside. This tends to support recent observations made by projects like the Roman Peasant Project which excavated 8 rural sites and discovered that only one them appeared to be a farmstead.
More significantly, it offers a valuable reminder that our view of the countryside is as detailed as it is incomplete. The landscape hinted at by Enry’s analysis may well reflect a significant degree of change over time, functional dynamism, and perhaps nibbling along the edges some of the challenges for systematic documentation of rural sites whose size make them particularly susceptible to various formation processes.
As another aside, it is especially refreshing to see a new wave of analysis of “rurality” built on data collection from intensive survey. At some point in the last 20 years, handwringing over matters of methodology seemed to distract archaeologists from the potential of intensive survey data to contribute to broader historical (and archaeological debates) about the character of the Greek countryside. Erny shows that by using legacy data from the vast number of surveys completed over the last forty years that it is possible to find new ways of thinking about the Greek landscape. This is a very welcome development, indeed!