The First Days of the Western Argolid Regional Project 2016

Today was the first full field day of the final full field season of the Western Argolid Regional Project (WARP). It was immeasurably better than the first full day of the project last year and probably a bit better than our first field day in 2014.

While we still have some open plains around the Inachos river, for this season, our survey area is a striking mix of narrow valleys and steep hill slopes. 

P1140966

Oranges, apricots, olives, peaches, vineyards, and the occasional pomegranate trees, planted in neat rows organize our survey units.

P1140977

P1150007

More than any other year, we’ll have to contend with the early modern and modern landscape.

IMG 4747

P1150021

P1150037

P1150041

So think of us as we stagger to our cars at 6:30 in the morning.

IMG 4740

The Archaeological Institute of America’s Annual Meeting in Review

After a couple of days at the AIA annual meeting in San Francisco, I started to wonder whether archaeologists should only be allowed to go to the meeting once every two or three years. This is not meant as a criticism to those there or the bustle of the conference (or the need for structured and unstructured opportunities to interact), but as a way to observe that changes in the field are thrown into high relief after taking a few years aways from the conference.

Here are a few observations:

1. Late Antiquity. When I first started to attend the AIA, panels on Late Antiquity were shoved unceremoniously into the Saturday or Sunday morning sessions, safely sequestered from the proper business of Classical archaeology. Over the past few years, however, Late Antiquity panels have migrated to the grown-up table, and this year, a panel on Greece in Late Antiquity happened at 8 am on Friday, prime time for the conference. 

The panel was a nice blend of senior scholars and new comers and established projects and new field work. The focus was on ceramics (3 papers) and to a lesser extent architecture (2 papers) and skewed later, into 6th and 7th century (4 papers) abandoning to some extent the 4th and 5th century sweet-spot favored by earlier scholars of this era. Three of the papers featured explicitly quantitative analysis, and two of the paper drew upon recent field work: a survey and an excavation. 

While 5 papers are hardly a representative sample of the work in Late Antiquity, the distribution of papers in this panel offers vague confirmation to my gut feeling that Late Antiquitists are working on later period and more fixated on ceramics than ever before.

2. Survey Archaeology. Like Late Antiquity, it wasn’t very long ago when you’d expect someone to stand up at any panel on survey archaeology and ask whether we could really base any arguments on material found on the surface. Those days have passed, it would seem (whether we have resolved the underlying issues associated with survey archaeology and formation processes or not) and the panel at this year’s AIA drew a standing room only crowd.

The papers were good, and projects appeared sound. None of the paper appear to genuinely embrace an analysis based on siteless survey, and in almost all cases preferred to talk about the landscape as a series of sites with distinct functions. At the same time, none of the paper really talked about any sites smaller than the ambiguous “settlement.” I don’t recall any farmsteads, sanctuaries, or site functions defined by size. There was also very little discussion of method.

3. Abandonment. I enjoyed the twin sessions on abandonment which both problematized abandonment as a symptom of decline, as well as a key stage in the formation of sites in the archaeological landscape. The convergence of concerns about periodization (period are frequently defined by episodes of abandonment) and archaeological formation processes points creates an intriguing and productive space around historical narratives that have become so dependent upon patterns of rise and fall. In fact, the ambiguity surrounding abandonment offered a temporary respite for anyone exhausted by popular narratives of decline that are so prevalent in our media today.

We can’t avoid change.

Efficiency

Since we’re all about efficiency and archaeological Taylorism here on the Western Argolid Regional Project, I decided to run some numbers, out of curiosity more than anything.

The primary productive unit of the survey is the five member field team. It consists of a team leader and four field walkers. They walk an average of slightly over 100 units per day with occasional outings in the mid-100s. We run 5 field teams a day since one team is in the pottery storerooms. It takes field teams about 7 minutes to walk the average unit with some units taking as much as 7 or 8 times that long (and others taking almost no time). Most teams start their first unit a little after 7 am and finish their last unit around 12:45 pm.  So our field day runs for about 6 hours (to simplify). The teams walk for about 2 hours, 15 minutes per day (or about a third of the time their in the field). The rest of the day is devoted to filling out forms and traipsing from one unit to the next. Lest this makes our field walkers sound lazy, I should point out that, over the course of our field season field, walkers walked over 1000 km (that over 600 miles for Americans). There’s no lack of energy and commitment on the part of our field walkers!

Fieldwalking

What I discovered is that the average field team used only 3 walkers for field walking. In fact, the average number of walkers per field was almost exactly 3 (the mode was also 3). This got me thinking that, next year, we should take our 6 field teams of 4 field walkers and divide them into 8 field teams with 3 field walkers. This would have the clear advantage of putting 7 field teams into the field daily (with one team heading to the pottery storerooms each day), and this should increase the number of units walked per day by about 40%. 

When I pitched this to a few team leaders, they responded that the teams often used the fourth field walker to help record information when not walking units. If resulted in an increase in efficiency, we should see that 3 walker units are completed more quickly than 4 walker units. The numbers, however, don’t bear this out. Both 3 and 4 walker units get done in about 9 minutes despite 4 walker units being generally longer (by around 25 meters) than 3 walker units. So, there doesn’t seem – on the face of it – to be any real efficiency gained by 3 walker teams. (I do know that some field teams operated at below full strength, but even when I did some rough work to control for this, it didn’t seem to impact the overall numbers very much).

There is one hitch: Around 65% of our units used fewer than 4 walkers, but about 20% units used 4 walkers exactly. But this, I think, is an artifact of our units being mapped to accommodate 4 walker teams. This might account for why units with more than 4 walkers (but less than 9) average about 11 minutes which is a substantial increase over those with 3 or 4. This is the result of teams having to double walk the unit; that is: walkers having to walk the unit once and then again. Curiously, the 11 minute average is not twice the time taken to walk a unit where every walker walks only once. This is probably because we tended to make larger units from areas where the fields are disturbed and unlikely to produce much pottery. While I haven’t run the numbers recently, historically our ceramic densities decline as unit size increases. So, I suspect one thing that might happen if we shrink our field teams is that we’d shirt our unit size to accommodate the smaller teams. So we’ll do more units, but maybe not survey more ground. 

Of course, to make this all work, we have to find two more excellent team leaders to complement our fine group of six. Moreover, we’d have less margin of error for individual teams. This year we lost a few field walkers each week to ailments ranging from dehydration to sea urchin attacks. Teams dropping to two walkers would struggle to be flexible enough to walk large units and would probably suffer just walking average sized units.

Embiggening the number of teams (by debigulating the number of walkers) might also lead us to increase the number of cars and would almost certainly require us to increase the number of devices assigned to team (cameras, GPS units, Sharpies, et c.). But as a good buddy once quipped, if you can’t afford to do maximum archaeology, perhaps you should just stay in the library. 

Curated versus Automated Revisits

There’s a good bit of buzz lately about Apple Music’s “curated” playlists, and TIDAL, my preference for a music streaming service, offers a range of curated music playlists as well. In general, the term curation, like crafted, artisanal, or any of the other tech-media, marketing buzzwords has come to mean that a human, rather than an algorithm has produced a collection. As many, many have observed, the term curation is annoying and overused.

But I still want to use for a little bit in reference to our work on the Western Argolid Regional Project. This morning, I took some time out of the field to start to analyze some of our finds and field data. We plan to revisit a few units before the season concludes and to collect some more material. Our hope is that these targeted revisits will help us both to refine our survey methods by offering some points to calibrate our sampling strategy, they’ll help us produce more robust assemblages of types of pottery that might only appear in very small quantities using our typical collection approach, and revisits will allow us to document archaeological features a bit more intensively than we would have time and resources to do over the course of intensive survey.

RevisitMap4BlogSM

We target sites for revisit in three ways. First, our field teams can tick a check box and provide a brief explanation for why a particular unit is worth revisiting. Our ceramicists, Scott Gallimore and Sarah James, can also identify units as being interesting, important, or confusing and consequently worth revisiting. Finally, we can analyze data through our GIS and databases that target units with certain characteristics (such as low visibility with either high densities or diverse assemblages). Our revisit lists generated by team leader and ceramicists are not fortified by statistics, but generated through careful observations and total situational awareness. These units represent the slow archaeology approach to landscape and artifact analysis.

So far, it has been heartening to recognize that the lists of revisit units curated by our team leaders and ceramicists are remarkably consistent with the units generated from my analysis of our various databases. In fact, combining the curated list of unit with list of units generated through our analysis of GIS tend to complement each other by expanding the potential target units for revisit. As we nuance the criteria for revisit a bit over the next week, I’m sure that we’ll discover some counterintuitive units that will serve as tests of our archaeological instincts. For now, though, we’ll proceed into the final week of the season with just a bit of confidence that our experiences in the field and at the pottery tables reflects the complexity of our study area.

Speed

One of the more interesting trends emerging so far during the Western Argolid Regional Project season is competition among field teams. At the end of each field day, I typically ask team leaders how many units they have walked. This seemingly benign question helps us measure our progress through the survey area and gauge how much mapping is necessary to keep ahead of the survey teams. A quick tally of the number of units walked lets me begin to plan the next day as soon as the previous field day is over. 

P1140176

Generally our 5 field teams walk between 15 and 20 units and around 90 total. Each unit is around 3000 sq m. so we walk about 1.3 and 1.5 sq. km per week. The number of units we walk depend considerably on the character of the terrain, the size of the units, and the density of artifacts, vegetation, and other distractions to artifact recovery. The size of our field teams is four plus a team leader, but this week we lost a few field walkers to dehydration and bumps and bruises. So a team down a walker will move a bit more slowly than one at full strength especially if the units are slightly larger than average. Historically, field teams walk about 4 units per hour over a 6 hour field day with a couple of breaks for water, znacks (snacks), and transit to and from the field site. 

Teams generally develop a routine where one walker writes tags, one takes a center GPS point, one walker helps with forms, one takes photographs et c. This streamlines the bookkeeping and data recording aspects of intensive pedestrian survey and as the season progresses, small efficiencies occur based on familiarity with the process as much as anything. As the process become more efficient, we usually have to nudge the team leaders to slow things down just a bit to ensure that the teams recognize where they are in the survey area, fill out forms properly, and actually, you know, enjoy the process. Since our project runs as a field school, we see very little benefit to an overly mechanical process that makes our field walkers (and team leaders) into field walking robots (beep, boop, boop, beep, boop).

One thing that I did not anticipate this summer is that teams would start to compete with each other to walk the most units per day. It’s hard not to like the harmless morale boost that comes with walking the most units or besting a team nearby is fun. Moreover, we recognize the field walking – particularly in challenging topography which is difficult to grasp as a coherent space – can be boring and seem pointless. The assembly line was soul crushing in part because of the repetitive character of the work and, in part, because the repetition could obscure the role an individual played in the work’s final result. Unit counts keep the field day interesting.

At the same time, we’ve starting wonder whether there are some less than desirable byproducts of this competition. For example, we don’t want the push to walk more to exhaust field teams more quickly and to contribute to the attrition of team members. We also don’t want to compromise our data collection for some good-natured fun. Finally, we don’t want teams who walk more challenging areas to feel like their contributions are less significant because they didn’t walk enough units. The last thing we want is sad field walkers.  

  

Survey Archaeology and Forms

Anyone who has done archaeology lately knows that we almost spend as much time looking at form (or its digital equivalent) as the trench, survey unit, landscape, or architectural feature. In general, forms are unattractive and at best functional (at worst, they are overwhelming belches of blank lines, boxes, and cryptic instructions. 

Tomorrow the 2015 Western Argolid Regional Project season starts. We had a few little tweaks to make to the database and that led to some tweaking of the form and that led to some modifications in its appearance. 

I’m sure I’m violating several laws of graphic design in my efforts, but I think I’ve improved our forms legibility and added a bit of style. The font is Prime; it’s a free, sans serif, highly geometric font which adds some bling without encroaching too much on the utility of the form. 

I also tried to standardize the placement of boxes. Almost all archaeological forms that I’ve encountered try to do too much in too little space. For WARP, we want to keep the form to a front and back page. So I tried to find ways to negotiate the constrained space of the form so that it was a little bit easier to follow and I tried to play a bit with orientation by extending some things to the right of the margin and some boxes to the left (in an orderly way) to index the form a bit and to give some more room for the free text boxes. 

WARPForm 01

Managing the Modern in Intensive Survey

I’ve made it over to the Argolid and am ensconced in the comfortable accommodations in the village of Myloi for the next two months. My colleagues Dimitri Nakassis and Scott Gallimore have been in the village for a week or so already getting ready for the second field season of the Western Argolid Regional Project. 

I’m excite for this year’s survey area because it encompasses at least two modern settlements which are in states of abandonment. We’re anticipating already a greater amount of modern and early modern (for Greece this is the 19th century) material associated with these settlements. Most recent intensive survey projects make a big deal about being diachronic, but to be fair, the modern period tends to present particular challenges to survey projects. In general, survey archaeologists recognize that we cannot treat the modern period the same way that we treat earlier periods. 

The reasons are both complex and simple. The simple reason is that we simply cannot accommodate the super abundance of most modern material in our survey units. As Richard Rothaus and I discussed a few months ago on our podcast, there is a storage crisis in archaeology, and collecting modern material will only make this worse. In the Eastern Korinthia Archaeological Survey we tried to document modern material without collecting using a “modern sweep” form. This form consisted of a long list of check boxes that tried to take into account the most common form of trash found in the Greek countryside. In practice, however, the survey teams mostly checked the box for “scattered modern trash,” and either failed or refused to distinguish between the various events that created the distribution of modern material through agricultural lands around contemporary villages. 

P1070636

I suspect that the difficulties dealing with the modern landscape also speaks to more complex challenges involving how we understand modern artifact distribution in the countryside where most modern survey projects are based. Modern material represents both very familiar practices – typically those associated with opportunistic discard of unneeded objects – and practices that are rather unfamiliar to archaeologists who are not well versed in modern, sometimes ad hoc, use of modern material in contemporary Mediterranean agricultural practices. For example, last year, I took numerous photographs of modified plastic water bottles hung from trees throughout the Argolid and the ingenious use of beer cans in modified irrigation systems.

P1070677

Our familiarity with the primary use of objects and simple discard practices has perhaps made it easier to overlook creative examples of reuse in the countryside. Modern objects have become so specialized, so disposable, and so common that we have to train our eyes to see them and our archaeological awareness to consider the range of uses possible in the countryside.

Rural Roman Landscapes of Greece

Now that I’m back working in Greece, I’ll have to start paying closer attention to the annually published Archaeological Reports, and a number of my colleagues helped me out by tipping me off to some of the nice contributions to this year’s edition. Generally speaking, Archaeological Reports summarize recent research in particular chronological period, and mostly they have focused on newly discovered and published sites.  

I was especially glad to read Daniel Stewart’s summary treatment of rural Greece during the Roman period. He does a nice job surveying (pun, pun) the work of intensive pedestrian survey projects in Greece, and this is no easy task as many of these projects have not published traditional archaeological volumes, but in scattered articles in edited volumes and journals. Better still, he goes a step further and considers the general direction of intensive survey in Greece with special reference to the challenges of the Roman period. This attention transforms what could have been a parochial survey of newly discovered Roman rural sites into a must read for anyone interested in intensive pedestrian survey.

Stewart identifies four major areas of development in intensive survey: challenges to ceramic typologies, refined collections strategies, studies across landscape zones, and interdisciplinarity. He does a nice job communicating the problems associated with ceramic chronologies for the Roman period and the vexing, but somehow inescapable dependence on the Early, Middle, and Late Roman chronological division. (I blame prehistorians for passing this chronological structure onto us.) David Pettegrew’s landmark Hesperia article on the “busy countryside” of Late Roman Greece was cited with approval (pdf here). 

At the same time, I think any close observer of survey archaeology would agree with these developments broadly speaking, although one could also say that these recent development have characterized the general trajectory of intensive survey since the 1980s. For example, survey archaeologists have always been working to refine their collection strategies to sample more effectively the material on the surface, and Stewart’s attention to re-survey is less a product of recent methodological refinement and more of a particular opportunistic, expression of longstanding interest in how best to sample and document kaleidoscopic surface assemblages. Stewart is right in recognizing that site classification remains a challenge for intensive survey projects and this is tied directly to the intensity of sampling. More rigorous sampling techniques produce a greater range of sites both in terms of size and, in many cases, in terms of functional assemblage. In some conditions, as few as a handful of fine ware sherds can represent activity in the landscape, but they intensity, type, and duration of activities at that particular place must remain undefined. 

The same could be said for recent attention to interdisciplinarity. The earliest efforts at intensive survey in Greece incorporated ethnographic and scientific components to their work embracing the twin influences of processual archaeology and the unstructured perambulations of early modern travelers. By the late 20th century, it was unthinkable to conduct a survey without geologists, a plan for sectioning pottery, biologists to help understanding flora and fauna, and ethnographers to interpret local knowledge. It was odd that Stewart did not mention the influence of geologists as being particularly important to recent trends in intensive survey. 

Finally, efforts to survey different landscape zones has been part of the survey archaeologist’s tool kit from at least the dawn of the Second Wave of survey projects. This is hardly a new trend or one deserving particular mention. In fact, one could argue that recent (21st century) permit limits that impose a 30 sq km maximum study area for intensive survey project have led to a shift from more extensive approaches to the Greek landscape to a more intensive focus on collection and sampling strategies. Intensive survey is committed to saying more with less.  

I also think that Stewart’s emphasis on the fragility of the surface assemblage in light of more intensive agriculture and development in Greece is misplaced or, at least, poorly defined. It seems hard to image even the most intensive collection regimes putting much of a dent in the abundant material present in a surface assemblage. In fact, our work on Cyprus in conditions in every way compatible with those in Greece suggested that typical sampling methods for intensive survey (20% of the surface) collect less than 10% of the material visible and that assemblage of material is only a tiny fraction of the material present. While deep ploughing/plowing does present a risk to archaeological remains (not to mention soil health), from the perspective of intensive survey, the danger is more closely related to movement of artifacts in the landscape than to any significant destruction of the archaeological record. 

I would have liked Stewart to focus more (any?) attention on the reluctance of the significant second wave survey projects (i.e. Pylos Regional Archaeology Project (update: I included PRAP accidentally in this list!), Nemea Valley Archaeological Project, Eastern Korinthia Archaeological Survey, Kythera Island Project, et c.) to make their raw digital data freely accessible. This has had a substantial impact on our ability to comparing and synthesizing the landscapes produced by these projects.

I might have also liked to see some critique of the tendency toward parochialism in Greek archaeology of the Roman period. Of course, this is a generalization that some might see as unfair, but it nevertheless would have been useful to understand how our understanding of rural Greece in the Roman period contributes or responds to similar interest elsewhere in the Mediterranean. For example, scholars invested in intensive survey methods have focused on rural Roman landscape across the Mediterranean basin. The work of these scholars have produced significant data both in terms of material and methodology for any understanding of Roman Greece.  

Despite my critiques (which are mostly saying that I’d write a different article!), Stewart’s article provides a nice summary of recent work and a great point of departure for anyone interested in staying abreast of recent research in the rural world of Roman Greece. 

Check out David Pettegrew’s review of this article here.

Connectivity in Cyprus and Corinth

Over the last few weeks, David Pettegrew and I have been working on an article that compares finds data from the Corinthia and from our site of Pyla-Koutsopetria on Cyprus. We were particularly interested in understanding how the types of ceramics that we can identify in survey assemblages shapes the types of economic relationships we can recognize in the Eastern Mediterranean. As one might expect, our focus has been on the Late Roman world, and we have been particularly interested in the difference between the kind of economic relationships manifest in assemblages comprised of highly visible amphoras and those manifest in highly diagnostic Late Roman red slip wares. The entire project is framed by Horden and Purcell’s notion of connectivity and that’s the unifying theme of the volume to which this paper will contribute.

The paper is exciting because it represents a step beyond the work that David has been doing on his book on the Isthmus of Corinth. I’ve read a draft of the book and it’ll be exciting. It also represents the next step for our work with the Pyla-Koutsopetria data. It is significant that all of our survey data upon which this paper is based, is available on Open Context. Our book should be available in time for the holidays. 

The draft below is 95% of the way there with only a few niggling citations to clean up. Enjoy and, as always, any comments or critiques would be much appreciated!

Some Quick Notes on Intensive Survey Method in the Argolid

This weekend I finally got around to putting together my various notes from database and GIS crunching and field observation on the Western Argolid Regional Project. Since we’re still working to analyze finds from this season, our main body of data derives from artifact densities. That being said, we have been able to spend a little time figuring out what variables had the greatest influence on artifact recovery throughout the survey area.

FEET

Visibility. The overall visibility in the survey area was right around 50%. Surface visibility did not correspond with artifact densities in a linear way, as survey archaeologists have come to expect. The highest artifact densities peaked first in units with 50% visibility and then in units with 70%—90% densities before dropping off in units with 100% artifact densities. In fact, units with 100% visibility produced fewer artifacts per ha then the average for all units. This serves as a useful reminder that visibility and artifact densities are independent variables even if the drop in density at 100% visibility hints that something strange must occur to artifact recovery rates in fields which have been finely plowed and cleared of all vegetation.

Surface Clast Size. We also recorded surface clast size for each field. Most of our fields consisted of 19-75 mm coarse gravel and these fields along with those with cobble sized (>75 mm) surface clast produced the highest densities. The average visibility in these fields falls between 41% and 53% respectively. Cobbles tended to produce more artifacts per ha than average visibility alone might suggest, but not by a vast margin (1040 artifacts per ha rather than the 913 artifact per ha that units with 50% visibility tend to produce). Units with coarse gravel were consistent with visibilities. Interestingly, units with fine gravel or sandy soil produced fewer artifacts than their average visibilities would suggest. Sandy soils, although relatively rare, had 41% visibility but produced only 390 artifacts per ha. It’s tempting to see sandy soils as recently deposited riverine sediments, but they don’t necessarily pattern that way across the survey area.

Background Disturbance. Recently, survey archaeologists have begun to think about background disturbance as a major influence on artifact recovery. This term describes the amount of objects in the soil matrix that distract the eye from the ceramic and man-made lithic objects we are supposed to be identifying.  We recorded background disturbance as either light, moderate, or heavy (or none). Our data showed that units with moderate and light background disturbance performed more or less consistently with their visibility. Units with heavy, background disturbance, however, had much higher than average visibilities (70%) and much lower than predicted artifact densities than this visibility alone would predict. This suggests that high background disturbance might influence recovery rates in a substantial way.

Dominant Vegetation Height. For each unit we recorded the dominant vegetation height. This correlated strongly with surface visibility – as one might expect – with densely overgrown units with vegetation head high or higher (!) having average visibility in the teens (18% and 17% respectively), and waist high vegetation averaged a paltry 33% visibility. Interestingly, head high or higher vegetation produced lower artifact densities than suggested by visibility alone, but we’ve long reckoned that our visibility scale runs to imprecise with very low visibility fields. Units with vegetation at knee height coincided produced densities that coincided with expected visibility, but units with ankle height vegetation produced more artifacts than one might expect from visibility alone.

These short studies demonstrate that artifact recovery rates are influenced by a range of variables present in the landscape. Using visibility and artifact density as a baseline for understanding artifact recovery allowed us to recognize the influence of a range of variables that impacted field walker performance. The highest recovery rates appear to come from units with cobble or coarse gravel, ankle high vegetation, plowed, loose soils, and light or moderate background disturbance producing visibilities of between 70% and 90%.