Thinking about Sites in Survey, Part 2

Yesterday I wrote the first half to an epic, two-part series of blog posts loosely related to a very recent article by Nathan Meyer in the Journal of Mediterranean Archaeology 35.2 (2022), “Finding Sites in Mediterranean Survey.” 

Here’s part two! 

4. Modeling and Efficiency. Nathan’s argument in the JMA paper (yes, this is a post about the JMA paper!) is that predictive modeling of landscapes can improve the effectiveness and efficiency of intensive pedestrian survey. Of course, models that will work in any given landscape demand different variables, different testing (and iterative) processes, and different scales. It seems to me that most contemporary intensive pedestrian survey project tend to focus on regions <50 sq km. This is as much a product of the rules governing survey permits in various countries as the area that can be studied intensively over a reasonable number of field seasons (say, 3-5 years). Most survey projects look to identify landscapes where intensive field walking is viable approach to documenting the landscape in a consistent way. In other words, survey projects tend to avoid densely vegetated landscapes, dangerous or rugged hillsides, densely built up or suburban areas or areas inaccessible by roads (or by sea). We tend to prefer valley bottoms and gently sloping (and terraced) hill sides, agricultural regions, and regions that preserve the traces of infrastructure reflecting historical investment over the past century (roads, paths, buildings). There are some exceptions, of course, but even Antikythera is laced with a network of rural roads necessary to deploy field teams.

My point here is not to challenge any of Nathan’s arguments, but to suggest that in many cases intensive survey projects already engage in a kind of modeling of the landscape dictated as much by contemporary concerns (permits, access, and land use) as the kinds of questions that we seek to answer. Of course, additional predictive modeling would complement these approaches and make clear that our contemporary concerns as archaeologists represent a variable present in site identification. Moreover, our understanding of the variable that might make a viable predictive model, from soil type, erosional conditions, vegetation, access to water, and so on depend on some of the same contemporary variables that shape intensive survey from the start. 

Finally, matters of efficiency remain (for me at least!) a hot button issue in intensive pedestrian survey. A survey project that doesn’t have the time or labor to document their entire survey zone using intensive methods has to compromise. These compromises always necessarily temper a project’s interpretative ambitions. Making sure a project’s methods are compatible with a project’s goals seems like it should go without saying, but it is still far to common to see project’s say that owing to time constraints, that they could not test or verify this or that conclusion. This always struck me as odd mismatch of expectations and planning.   

5. Resolution and Scale. This brings us to one of Nathan’s most valuable contributions to how we think about pedestrian survey. He argues that we need to embrace and integrate more extensive practices into intensive survey projects. I have done a good bit of extensive survey over my two decades of pedestrian survey work and generally found this approach to be productive especially for identifying architectural remains, more recent assemblages, and paths and routes through the landscape. Even the most methodologically committed siteless survey advocate (see yesterday’s post) would not recommend ignoring sites identified through extensive practices.

That said, there is real value to considering how extensive survey approaches, which might be as simple as single individual taking grab samples from various points in the landscape for later study, can provide methodologically consistent ways to expand the scale of intensive survey work and inform our understanding of the landscape. The key here, it seems to me, is how we understand the relationship between resolution and the presence of sites. Extensive survey has a fundamentally different purpose, at least presently executed, compared to intensive survey. When they’re combined, such as when an intensive team conducts the survey of a site identified through extensive survey, the goals of the intensive component are different. Intensive survey of a fortification, for example, identified through extensive survey is less to try to identify small sites that might have otherwise been missed, but to produce a more nuanced, diverse, and complex assemblage of material associated with a known site.

6. Assemblages. One of the gaps in Nathan’s thoughtful and thought provoking discussion of survey methods is their role in producing assemblages of artifacts susceptible to interpretation. In fact, one of the most basic criteria that I’ve used for the identification of a site in the field has less to do with density of artifacts and more to do with whether we can plausibly argue that a spatially coherent assemblage of material is capable of sustaining some kind of functional interpretation. Two medium coarse ware body sherds of indeterminate date cannot sustain interpretation. One Classical amphora sherd, a Classical fine ware sherd, and a Classical pithos rim might suggest domestic activity and I would likely classify them as a site. (Here it is sometimes useful to remember that in the US a single projectile point can be officially classified as a “site” depending on the situation.) It goes without saying that these three Classical sherds may not reflect contemporary activities, but considering the vagaries of 2500 years of site formation, it is impossible to know for certain. What we can say is that artifacts reflecting the activities of storage, transport, and dining during the Classical period exist at this site and considering the tendency of pithoi to stay put, it is possible to argue that these represent domestic activities.   

This same kind of logic allows us to argue that chronologically diverse assemblages tend to inform one another. A single Archaic sherd in an assemblage of material dating to the Classical and Hellenistic periods may suggest earlier activity at the site. A handful of Roman period artifacts may well speak to later activity as well. Whether this amounts of uninterpreted continuity of use at a site or episodic activity over hundreds of years is unknowable, but also less relevant for survey projects that operate on the scale of 30-50 square kilometers.

The point of this simplistic interpretative exercise concerning assemblages is that more intensive techniques especially in fields where surface visibility is very limited is not just about artifact densities. After all, finding three functionally or chronologically undiagnostic sherds doesn’t really inform how we understand the landscape more than one. What intensive practices sometimes can produce is more chronological and functional diversity and whatever one’s attitude toward a site is, diversity is just as important as quantity for making functional interpretations. One of the (few) redeeming elements of traditional surveys that ramp up intensity to document sites is that they increase the likelihood of recovering more diverse chronologically diagnostic material. In such cases, intensive practices can open windows into “hidden landscapes” that more methodologically consistent survey practices would have overlooked. This can be a very good thing indeed, but in my experience it rarely does any more than tell us what we should expect anyway: survey is good at showing presence, but not good at producing evidence for absence. More than that, finding traces of difficult to recognize periods in “on-site” assemblages produced by more intensive survey practices is not guaranteed to shed meaningful light on material collected from “off-site” areas where less intensive practices prevail.

~

As should be pretty clear by now, I really enjoyed returning to my roots in intensive pedestrian survey and using Nathan’s article as a kind of jumping off point to reflect on some of the issues facing both survey methods and the interpretation of survey data. I am not so proud to admit that at various points in my career, I privileged methodology and methods in the field over the need to produce the kind of data or to appeal to the kinds of concepts that allow survey to contribute to widest range of interpretative questions.

The concept of the “site,” of course, remains quite fuzzy especially in the field. We’ve all recognized ceramic assemblages that only appear in post-processing because they stretch across adjacent survey unit walked at different times by different teams. We know that there are spare but meaningful collections of sherds that appear undetected by field walkers amid low surface visibility. We also realize that complex diachronic assemblages often consist of overlapping and even unrelated scatters of pottery that only post-analysis can tease out.

Even realizing these relatively common eventualities, there is much to be said for giving into intuition and suspicion in the field. Slowing down, collecting more rigorously at times, and less intensively at others allows us to connect our time in the field with the kinds of assemblages our work produces. This sense of connection, whether established through long solo extensive treks through the Greek landscape or working more intensively with hardened(!) teams of survey veterans should inform our view of the “survey universe.” The kind of recursive, field-based, encounters that Nathan advocates, even if they’re directed by technology laden predictive models, should continue to play a role in archaeological place making.