Survey Archaeology

Since the 1960s, archaeologists in the Old and New Worlds have increasingly complemented subsurface excavation with a whole new variety of multidisciplinary techniques in data recovery. The most significant of these is surface survey, the visual examination of a landscape for variations in ground surface (which might reveal buried walls or buildings) or for the distribution of artifacts on the surface. This new method results, in part, from a recognition that much of our information about the past can be gleaned by examining what lies exposed on the landscape today. The cost effectiveness and ease of "field walking" has made survey archaeology such a valuable method of understanding the archaeological past. A team of surveyors walking across the land, noting artifacts, and recording important observations, might cover several hundred acres a day depending upon the nature of detail recorded and the care taken in the process (Figure 7.1). Moreover, surface survey does not destroy the area under investigation in the same way that excavation does; one can revisit a site numerous times to gather additional facts. Survey is also important in that it provides an entirely different kind of information about an area. While excavation may reveal detailed information about the use of a city or site through time, survey illuminates past utilization of countryside, landscape, and regions. And where excavation in the Mediterranean has traditionally been concerned primarily with only one kind of past activity area (the ancient city), survey reveals a great variety of sites, from overnight camp spots to small-scale settlements to ancient agricultural fields. For all these reasons, surface survey has been elevated in classical archaeology to a viable and important tool in understanding the human past.

Archaeological survey grew out of the sampling techniques carried out by botanists and geologists in their own specific surveys. Researchers in Mesopotamia and the Americas created a foundation on which Greek archaeologists would begin to build in the late 1960s in the Minnesota Messenia Expedition, directed by W. McDonald of the University of Minnesota. The project was initially concerned with the recovery of Bronze Age material, but the discovery of artifacts from periods other than the Bronze Age could not be ignored. The Minnesota project influenced all later surveys in Greece. The ensuing Southern Argolid Exploration Project developed a more systematic kind of survey directed toward all areas and periods in certain areas of the Argolid. Later surveys, such as the Cambridge-Bradford Boeotia Expedition (late 1970s), Pylos Regional Archaeological Project, Melos Project, Nemea Valley Archaeological Project, Berbati-Limnes Archaeological Survey, and the Sydney Cyprus Survey Project would follow and yet also advance the methodology of Greek survey. Basing its method on the achievements of these earlier projects, researchers from OSU are cooperating with others from Florida State University and Wichita State University (Kansas) in the Eastern Korinthia Archaeological Survey (Figure 7.2). Much of the discussion below makes use of the methods of this project (EKAS), but it should be borne in mind that EKAS is only one of many survey projects. Different projects do some of these things differently, but the basic proecesses are similar to those described below.

Landscape and Geomorphology

One of the principal trends in Mediterranean archaeology is the study of human interaction with the landscape through time. While an artifact scatter exposed on the surface may be indicative of past activities, the formation processes that created that scatter are complex and difficult to understand. Recent projects in Greece have recognized that the current state of the landscape has been affected both by human (anthropogenic) and natural factors. In order to understand spheres of past behavior, surveyors must try to determine natural features of the landscape, areas transformed by human activity, and the effect of these processes on artifact distribution through time.

Archaeologists have increasingly turned to geomorphologists to help in this process. Geomorphology is the study of the landscape and the processes that have created and shaped it. In the Korinthia, several kinds of natural processes have continued to shape the form of the land, which consequently influence the human use of that land: 1) Frequent tectonic shifts (earthquakes) and changing sea levels have either submerged or raised ancient shorelines; the analysis of sediment samples along the coast allow archaeologists to determine the location of the ancient shorelines. 2) The processes of alluvium (erosion) and colluvium (gravity) continue to operate in the movement of soil down the slopes of the mountains and deposition in the valleys; the study of the soils allows geomorphologists to identify the effect of these processes on the landscapes which form part of the survey area. Archaeologists can use this data to determine whether the artifacts found on the surface, such as pottery and lithics, originally belonged to soil context in which they were found or whether they were imported through natural or human processes.

Many surveys have incorporated an earth science component, but EKAS has elevated the significance of geomorphology to an indispensable underpinning of the project. Traditionally, survey has been conducted according to a site-based approach. Field walkers spaced at set intervals walk in a line across large agricultural fields, examining the ground for artifacts; find areas are called "sites" and are arbitrarily defined according to horizontal spread and density of artifacts. In EKAS, geomorphological zones rather than arbitrary, modern field boundaries govern the delineation of survey areas. Before the survey actually begins, the landscape is geomorphologically mapped. The basic survey areas, called Discovery Units, never cross geomorphological boundaries, ensuring that artifacts found in those units belong to the same formation context, landscape shaped by the same natural and human processes. Geomorphological units (and subsequently, Discovery Units) are generally small, and are assigned based on soil changes, drainage conditions, slope, or obvious human activities (e.g., bulldozing or terracing).

Artifacts

Over a three-week survey period in the EKAS 1999 season, ten field walkers counted 44,000 artifacts, more than 4,000 pieces per person. While surveyors certainly became better in recognizing artifacts through the season, all participants underwent a brief training in artifact identification before the survey ever began. For those who had little experience with survey archaeology, this training was useful and significantly improved the field walkers' ability to identify material during survey. Ultimately it takes years to develop a truly discerning eye to recognize objects against a variety of soils, grasses, and rocks, and for this reason, additional workshops and lectures are often scheduled throughout the remaining field season to help in this process. Artifactual materials of all kinds are numerous in the eastern Korinthia but are generally divided into several categories.

Ceramics are by far the most ubiquitous type of artifact in Mediterranean survey since well-baked clay breaks down slowly through time and people used vessels in virtually every period (Figure 7.3). The EKAS survey distinguishes types of ceramic by function. Fine Ware may include vessels used for domestic dining, ritual, or mortuary purposes. Thin, well-fired Archaic-Classical Black Slipped pottery or fine Roman red ware are typical for this category. Light Utility ware was used for more daily, utilitarian activities such as dining, cooking, or carrying; heavily-tempered, cookpot sherds are commonly found in EKAS survey. Heavy Utility ware, such as amphoras and micaceous "waterjars," were employed for the more rugged purposes of transport and storage of grain and olive oil. A pithos is a gigantic and heavy vessel that was also used for storage. When found in the countryside, pithoi are indicative of large-scale storage. Architectural ceramics are common in the Korinthia in the form of tiles, bricks, and decorative objects. They signify the presence of domestic and public buildings. Finally, ceramics are occasionally found which functioned for symbolic or personal purposes. Although these are rare, terracotta pipes, loom weights, figurines, and jewelry have survived in the archaeological record.

Lithics are even more durable and permanent than ceramics and are commonly found in many survey projects in Greece. Local outcroppings of chirt and flint, as well as obsidian imported from the Aegean island of Melos, were crafted into weapons, tools, and bladelets. The debitage remains from the tool-making process are most commonly found on the Greek landscape today.

Ceramics and lithics are the most frequent types of artifacts found during EKAS surveys but certainly not the only. Ground stone tools, such as mortars, pestles, and threshing blocks for grinding grain, as well as polished celts and axes, are sometimes noticed in survey. Beautiful marble revetment, which often lined the walls of buildings as a facade covering the stucco and concrete construction beneath, are still found today. Small marble and limestone tessarae represent the ancient presence of mosaic decorations. Bits of concrete used from the classical Greek period are another kind of building material. Finally, a wide variety of modern artifacts, from eighteenth century ceramics to plastic water bottles, are found and recorded during survey. While much of this amounts to scattered refuse, it is indicative of contemporary landscape utilization; more generally, it instructs our knowledge of human settlement patterns and enables conclusions about ancient usage.

Survey Procedures

The actual process of field survey differs from project to project, but there are some overriding principles that govern most processes. Many survey projects incorporate two or more phases of research with different purposes, involvement, and technique. At the basic level, surveys are generally investigative in that they aim to discover the kinds of artifacts and features present (or absent) over an extensive region. Walkers spaced at equal distances walk as a group across a field, counting artifacts and flagging representative pieces for later analysis (Figure 7.4). This phase is efficient, cheap, and painless, allowing a quick assessment of artifact potentials for the survey area. A second, more involved and intensive phase is necessary for areas with high artifact concentrations and will be discussed in the next section.

In the first phase of survey, projects often incorporate a nonsystematic component, which involves scouting out different parts of the landscape to get a general impression of the material there. If the survey is systematic, walkers will be spaced at regular intervals across the field, thereby covering a representative area. The amount of representation depends upon the goals of the project but one walker for every ten to fifteen meters is typical. In EKAS, this first level of research is called the Discovery Phase. Walkers placed every ten meters can visually cover one meter to the left and one to the right, a total coverage of 20% for each Discovery Unit; this means that eight of every ten meters goes unexamined and artifacts in these areas will be unnoticed.

Survey procedure in general is simple and straightforward. When the team leader gives the signal, participants transect the landscape, walking in their "swath" at a set pace and direction. Often where crops or trees are large, it is easiest to walk in the same direction as the vegetation; if the survey area is an open field, the team leader may choose a set bearing to follow with the compasses. The pace varies according to the amount of ground visible, amount of background disturbance (such as rocks, wood, and leaves), and the density of artifacts. It is necessary to take more time in a field that is cluttered with artifacts, stones, and vegetation than a field with light scatter of debris.

As the walkers transect the field, they constantly scan the ground for artifacts (Figure 7.5). When someone sees cultural material on the ground, he / she informs the other walkers by shouting "pottery," "flake," or whatever the artifact is (If the artifact is especially rare, the field walker may do a little dance). This informs the other crew members about artifact scatter areas which may spread into their own swaths and generally encourages participants to keep the eyes to the ground and remain attentive. For site-based surveys, all artifacts found in the same general location belong to the same "site" and are placed in the same site bag. An archaeological "site" in this sense is simply the term used to denote a cluster of artifacts that are spatially definable. The cluster itself is not always culturally significant and may entail nothing more than coincidental association. EKAS uses not a site-based approach but a geomorphological one. Artifacts found in the same Discovery Unit (usually a small field or area considered to be geomorphologically the same) belong to the same context even when they are not found in proximity to one another. Thus, if a Roman amphora sherd and an obsidian bladelet spread twenty meters apart belong to the same Discovery Unit, they are artifacts with similar geomorphological histories.

Artifacts are collected according to the research design and project goals, and a collection strategy is decided before survey begins. What will be done with the archaeological material found in the course of survey? Many projects employ a grab-sample collection strategy which removes a representative sample of the material from the field. There are variants of this such as the timed sample which only allows a limited amount of time to collect at any one place, or a selective sample where only the more diagnostic artifacts are removed. There are advantages to these kinds of strategies such as the opportunity for numerous researchers to analyze and study the material at a laboratory. And conjectures about the use of the landscape can be continually tested by referring to the actual objects themselves. On the other hand, removing artifacts from their primary context is destructive to the archaeological record; material removed cannot be put back. And it is questionable how much additional information is gained by collecting artifacts. Moreover, processing artifacts is time consuming and artifact storage requires a considerable amount of space (some Mediterranean surveys recover hundreds of thousands of artifacts). Finally, some researchers now have the resources to analyze the pottery, lithics, and other artifacts in situ. Digital cameras, for example, which take images that are immediately viewable, provide surveyors with quality photos of artifacts before they even leave the field. For these reasons, some projects have recognized the benefits of using a survey strategy that does not necessitate collecting artifacts.

The EKAS project employs a non-collection strategy based on a "ChronoType" system. A ChronoType is defined as a type of object that has both distinct physical attributes (e.g., decoration, color, fabric, temper, thickness) and a specific temporal association (e.g., Classical Greek, Late Roman). The categorization of the ChronoTypes is both open ended and hierarchical. On the one end are body sherds (e.g., Coarse Red body sherds) with few distinct attributes that are only very loosely connected with a temporal period of production (e.g., the ancient world rather than modern times). On the other end are sherds with very distinctive physical characteristics and a tightly defined period of production (e.g., Late Geometric IIIC amphora). The ChronoType system classifies all artifacts by their most recognizably distinct attributes.

In the EKAS project, artifacts are counted with handheld "clickers." Crew members use these tally counters to facilitate counting for dominant artifacts such as pottery and tiles; participants click once for each artifact. Rarer artifacts, such as lithics and marble revetment, are recorded mentally. Every potsherd representing a distinct ChronoType is flagged. Once a ChronoType has been flagged, no other field walker needs to flag another of the same kind of artifact for that Discovery Unit, although additional artifacts of the same ChronoType will be counted with the tally counter. For example, a common ChronoType found during survey in the Eastern Korinthia is the combed-ware body sherd, dating to the Late Roman period. Once one of these sherds has been seen and flagged in a Discovery Unit, it is no longer necessary to flag another. This system allows walkers to flag distinctive artifacts and avoid duplicates; it also saves time for the processing team, who will come through later and analyze the flagged artifacts in the Discovery Unit. After walkers finish their tracts across the DU, they report tally count totals to the team leader. The number of potsherds, lithics, and other types of artifacts are recorded for each person's tract. If there is still more area to survey, walkers will again pace off ten meters and walk back across the Discovery Unit.

In every survey project, it is necessary to record additional information about survey conditions, such as procedures, field conditions, artifact patterning, weather, and even the morale of the team. Especially important are notes on ground cover (Figure 7.6). The density of stones and vegetation that distract the field walker's visual attention is called background disturbance. The visibility of the ground to the walker is also an important condition to record. The types of debris on the surface and the crop planted that year affect the surface exposed to the field walker. The land may be so covered with vegetation that survey is not possible. Other times, a freshly plowed field with little or no plant coverage renders an ideal survey situation. The percentage of ground visible is usually estimated and varies from field to field. In the Eastern Korinthia project, the team leader keeps a written narrative of the survey process and field walkers complete four pages of forms for every Discovery Unit. This includes such information as the date, the names of the walkers, the location of the survey area, walker direction and bearing, artifact counts and notes, land cover, visibility, soil conditions, the current utilization of the land, features, and evidence for modern activities. While this type of data recording takes additional time, it is nonetheless significant in determining whether the artifacts collected are sufficiently representative of the material present in the field. If vegetation cover totally obscures the surface, it may be necessary to re-survey the field at a later time when the area is more visible.

In survey projects employing a collection strategy, artifacts are taken back to a laboratory and analyzed. Because EKAS is a non-collection project, all processing and analysis takes place in the field. After the survey team has finished a Discovery Unit, the Object Processing Team comes in and analyzes flagged artifacts. Potsherds and lithics are measured, sketched, and described; photographs are taken with a digital camera; artifacts are designated to an appropriate ChronoType. The survey teams and the processing team use walkie-talkies to communicate in the field, and observations of the latter can aid the surveyors in determining what artifacts to flag. The diagnostic artifacts identified by the Object Processing Team can provide specific information about chronology and human activities in ancient times.

Intensive Survey

Many projects undergo a second phase of survey that will explore more intensively an area discovered in the first phase. In EKAS, after a team has finished a Discovery Unit, the team leader may recommend a closer, more intensive examination of the area at a later point on the basis of the high concentrations of artifacts or unusual kinds of artifacts present (Figure 7.7). This survey area, called a LOCA (Localized Cultural Anomaly), requires different strategies and procedures that will delineate the site boundaries and define the chronology and functions of the site. Strategies might include, but are not limited to, further geomorphological study, intensive collection within a grid, excavation, sketching plans of relevant features, or geophysical survey. The particular strategy depends on the nature of the LOCA and the specific research questions. In 1999, an apparently important site (with identified material from prehistoric to Roman times) was gridded into 10x10 m units. In the center of each of these units, participants collected artifacts from an area constituting 10% of the unit. The artifactual material was collected for later description and analysis. The data gathered from the LOCAs can be studied in terms of chronology and space, informing researchers about the use of the site through time.

Computers and GIS

Archaeologists in EKAS use computer technology to record and manage the data collected during survey (Figure 7.8). Several interrelated databases exist and one of the main priorities of the project is to transform all data to digital format. All paper forms filled out during survey (Discovery Unit forms, Geomorphological Unit forms, and artifact analysis forms) are keyed into databases and are made accessible to project participants through the internet. Geographical information is also organized and managed through computer programs called Geographical Information Systems (GIS) which convert spatial data into mathematical and graphical format. In EKAS, researchers "digitize" their topographic maps (1:5000 scale) and aerial photographs before survey actually begins. Once graphical representations of actual maps exist and survey begins for the season, project participants must locate all Discovery Units in real geographical space. While surveying, they use Global Positioning System (GPS) units to determine their location relative to satellites overhead, or get an approximate location by relating visible markers such as roads and streams on 1:5000 topographic maps and aerial photographs to the actual road and streams. In either case, the boundaries of the Discovery Unit are sketched onto maps in the field and then later digitized into GIS programs. Moreover, other types of information collected during survey is also keyed. When these data have been entered, archaeologists can begin to ask complex questions about the relationships of variables such as land slope, Discovery Units, artifact densities, soil, and environment. For example, EKAS researchers can test the relationship between slope and the high density of Late Roman amphora sherds. Were there consistently more Late Roman artifacts at higher elevations? Or, location and function of sites can be related to the distance to the nearest water source, the coast, settlements, or even cities (e.g., ancient Korinth). Like several transparencies stacked one on the other, GIS allows one to simultaneously view and interpret layers of data in light of spatial information.

Experimental Survey

Over the last several decades, archaeologists have increasingly recognized that what is discovered through survey is not always representative of what is actually present. For example, dense clusters of rain-washed artifacts are much more obtrusive and visible than scattered random artifacts which are highly encrusted with a thin limestone patina. Field walkers will recognize an over-proportionate amount of highly obtrusive artifacts while the discovery of artifacts in areas of low density are often matters of chance. Recently, projects have incorporated an experimental component to test the conditions which effect artifact recognition during survey as well as the investigative technique and sampling strategy. How good are humans at recognizing cultural material when 50% of the surface is covered with vegetation, or when there is a high density of pebble and gravel? Further, how does the pace of field walking affect identification? Will a pace twice as fast decrease recognition by half? The questions are important ones as they inform how much field walkers are recognizing as they survey fields.

The EKAS project is currently conducting experiments to improve our understanding of the affect of a variety of conditions on survey results. In seeding experiments, a team of researchers carefully planted potsherds in specific positions along a fifty-meter tract and plotted their positions on a plan. All potsherds were photographed, analyzed, and described before their placement. Two kinds of potential survey conditions were tested. In the first, artifacts placed in a tract with 50% visibility were tested against a tract with 100% visibility. How much would weeds and grain stubble hinder recognition? Secondly, artifacts placed in a tract with high background disturbance were tested against a tract where there were few visual distractions. Participants walked these various kinds of tracts and flagged all artifacts which they saw one meter to either side of their path. After the field walker has completed the tract, the experimental team can analyze which artifacts were recognized during the pass and make broad interpretations about the conditions under which survey was conducted. Additionally, since all of these experiments are timed, the researchers can test efficiency of artifact collection against the precision. How many more artifacts will be found when greater time is allowed in survey? Finally, these experiments test the visual range that surveyors can conceivably cover during a tract. Will increasing visual range from two to four meters double the number of artifacts discovered? The answers to these questions can be applied to future survey work in the EKAS project and Greek survey overall.