Over the last three months there has been some discussion of the chronotype system of sampling and identifying pottery in the context of regional intensive survey projects. The most recent critiques have appeared in D. Frankel, Review of B. Knapp and M. Given, Sydney Cyprus Survey Project. AJA 112 (2008), 182-183 and Y. A. Lolos, B. Gourley and D. R. Stewart, "The Sikyon Survey Project: A Blueprint for Urban Survey?" JMA 20 (2007), 271 (for more on this article and this exciting project see their spectacular Sikyon Survey Project web site and my less spectacular previous post).
Tim Gregory developed the chronotype system for the Sydney Cyprus Survey Project and it has subsequently been employed by Australian Paliochora-Kythera Survey Project, the Eastern Korinthia Archaeological Survey, the Troodos Archaeological and Environmental Survey Project, and my own Pyla-Koutsopetria Archaeological Project on Cyprus. In very basic terms, the chronotype system involves both a sampling strategy and a somewhat nested terminology for artifact types designed to facilitate the reading and analysis of survey pottery. The sampling strategy (as alluded to in the film short from Emerging Cypriot released yesterday) involves collecting every unique artifact type from a survey unit to ensure that one of each kind of artifact in the unit is represented in the collected assemblage. It was a response to an older generation of survey projects that tended to collect only "diagnostic artifacts" which often meant rims, handles, bases, and feature sherds. By collecting one of every type of artifact present in the unit, the chronotype system shifted the responsibility for determining whether an artifact was diagnostic from the fieldwalker (who in traditional surveys would have to decide whether a sherd was likely to be significant for establishing the chronology or function of the site) to the ceramicist who would study an assemblage that reflected the material present. All the fieldwalker needed to do was to determine whether a sherd was similar or different from the ones he or she had already collected.
The nested terminology for artifact types involved the ceramicist assigning to each sherd a standardized chronotype which was a combination of date range (e.g. "Late Roman", "Early Bronze Age", or "Ancient"), fabric type (e.g. "medium coarse", "fine ware", "cooking ware" et c.) and description (e.g. "combed", "black glazed", "drip painted"). Standardizing the way in which artifacts were described facilitated the quantitative and qualitative analysis of the survey pottery which demands that artifact types be normalized consistently across the data set. It also provided a fairly well-developed set of artifact identifiers that could be (and was) exported to other survey projects.
Frankel expressed concern that the chronotype system, by eschewing "conventional pottery terminology" would make inter-site and inter-project comparability more difficult. This is certainly a concern. Inter-site comparability (e.g. Side-by-Side) is of increasing significance as more and more of the Mediterranean world is covered by intensive survey and, perhaps more importantly, as survey relies upon excavated contexts for establishing the chronology of surface pottery. That being said, I am not as convinced that "conventional pottery terminology" is so stable that introducing a new set of standardized terminology designed for the vagaries of survey pottery will have any inherent incompatibility with more traditional nomenclature. For the periods where my research focuses (which are generally historical), there is sufficient diversity in the conventional terminology to require some translation between projects(consider, for example, the typologies for Late Roman amphora). As survey and excavation data sets make the slow migration to accessible digital archives which will allow for more direct comparison between projects, there will certainly be a need to create concordances of ceramic terminology that take into account not only the variety of terms employed to describe particular sherds, but also changes in identification of certain types of pottery. Frankel has identified an area that will require the attention of archaeologists in the very near future.
Lolos et al. critiqued the sampling strategy employed by the chronotype system. They questioned whether a fieldwalker could consistently determine whether a sherd was "different" and therefore worthy of collection. They also wondered if by collecting each unique sherd we would lose the ability to talk about relative frequency of particular artifact types within a unit . Their critiques are, indeed, valid, and we have worked to address them in several recent publications (Caraher et al. "Siteless survey and intensive data collection in an artifact-rich environment: case studies from the eastern Corinthia, Greece. JMA 19 (2006), 7-43; T. Tartaron et al. "The Eastern Corinthia archaeological survey: integrated methods for a dynamic landscape," Hesperia 75 (2006), 453-523; Caraher et al. "Pyla-Koutsopetria Archaeological Project: Second Report 2005-2006" RDAC 2007, in press; D. Pettegrew "The Busy Countryside of Late Roman Corinth:Interpreting Ceramic Data Produced by Regional Archaeological Surveys," Hesperia 76 (2007), 743–784; R. S. Moore, "A Decade Later: The Chronotype System Revisited," in Archaeology and History in Medieval and Post-Medieval Greece: Studies on Method and Meaning in Honor of Timothy E. Gregory. W. Caraher, R. S. Moore, L. J. Hall, eds. Forthcoming.) The two salient points here are as follows:
1) We regularly tell fieldwalkers "when in doubt regarding whether a sherd is unique, collect it." Some recent studies (see Moore (Forthcoming) and Caraher et al 2006) have suggested that fieldwalkers tend, if anything, to over collect; that is to say err on the side of caution and collect too many examples of even relatively undiagnostic sherd. While this cannot anticipate whether the walkers have overlooked certain types of artifacts completely (i.e. artifacts that are so indistinguishable from other sherds that they are disregarded consistently as duplicates), it suggest that they did not. In any case, the chronotype sampling method should ensure a more robust sample of the variety of material present on the surface than techniques which involve only collecting highly visible "diagnostic" (e.g. rims, handles, feature sherds, et c.) artifacts.
2) The chronotype sampling strategy will not create assemblages that are as robust as so-called "total collection" strategies which involve the collection of all the material from a unit or a walker's swath. We conducted experiments in 2005 and 2006 and determined that, indeed, total collection from a 5% sample of the unit would produce an assemblage that would allow discussions of proportional representation within the unit. This being said, it is not clear exactly what the significance of the varying proportions present in a surface assemblage means. The formation processes that create the surface assemblage are incredibly complex and varied across periods; thus, for example, it is difficult to understand what a proportionately greater quantity of, say, Classical period artifacts in a unit means. Does it mean more Classical activity in a particular spot? More intensive use of that spot? More people? Or does it suggest that the Classical period material was particularly susceptible to certain site formation processes. From my perspective we need to understand the processes that lead to the creation of the surface assemblage much better than we do today to find any significance in the proportion of pottery represented in a unit. Of course, gross differences -- like the incredible, overwhelming presence of Late Roman material at Pyla-Koutsopetria -- will come out in the chronotype system as in almost all cases the more pottery there is from a particular period the more diversity there is present in the assemblage. Thus, the chronotype system, which excels in documenting diversity, will produce a use and valid indicator of particularly prevalent periods.
The point of this post is not to attack Frankel and Lolos et al. -- far from it, in fact! -- but to expand the dialogue into the blogosphere and, perhaps more importantly, shed some light on an exciting (and surprisingly robust!) little debate in the world of Mediterranean survey archaeology.
Recent Comments