[Printout | Contents | Search ]

Context sensitive: Computers in Archaeology at the University Museum

By Randall Couch

Archaeology is a science in which the basic experiment-the excavation of a specific site - can never be repeated. Once the site has been opened and its objects exposed, recorded, and collected, their original context is altered forever. Everything we hope to know about the people who lived in that place depends upon how well the context is recorded.

Context, however, has assumed this degree of importance only in fairly recent times. The earliest "archaeologists" - conquerors, antiquarians, adventurers - were interested mainly in spectacular objects. They did not hesitate to dig or blast indiscriminately in their hunt for treasures. The history of the field has been a gradual development of disciplined practice and improved techniques for recognizing and recording the relationships among objects, and between objects and their surroundings.

Penn researchers were quick to adopt new field methods such as as those pioneered by British archaeologist Mortimer Wheeler in the early part of this century. The later excavation of the royal graves at Ur, directed by Leonard Woolley and cosponsored by the University Museum, set a new standard of accuracy for its time. The work of Penn's William Coe at the Classic Maya site of Tikal represents more recent context-sensitive practice. Today, Penn researchers are breaking new ground in the application of computer technology to excavations.

Dr. Harold Dibble, of the Museum and the Department of Anthropology, explains the challenge: "The central issue of archaeological context is understanding provenience - how objects lie in the ground. That leads naturally to site mapping. The standard practice of site mapping has been in place for 50 years or so. Most field techniques were designed to facilitate mapping: Digging in squares, for instance, facilitates measuring from the squares' edges. But every aspect of field technique is a compromise designed to maximize the amount of information gained per unit cost. The main cost is not capital cost, but the time spent on the excavation."

The computer's role

Electronic data acquisition includes the use of electronic calipers and scales for measuring objects, video imaging systems to measure surface area and form, and other instruments that can log field data directly in machine-readable form for later analysis. A key development in this area, first published by Dibble, was the adaptation of a surveying system called a total station (an electronic theodolite for determining angles coupled with an electronic distance meter or E.D.M.) to record the three-dimensional location of artifacts in a site. The total station takes highly accurate measurements of distance and relative direction by reflecting a laser beam from a small prism placed on the distant point to be measured. Total stations were already in use for large-scale topographic mapping of sites. "Harold's innovation," says McPherron, "was to take something that had been used to measure kilometers and use it to measure centimeters." The resulting benefits are described below.

Data analysis depends on constructing and manipulating complex databases of archaeological information. Data from many levels of an excavation and from many sources must be brought together in order to draw conclusions. For example, information about where in a site objects were found can be combined in a single database with data recorded later in the lab-after the objects have been cleaned-to reveal patterns of distribution of similar objects. The capacity to enter and manage more categories of information permits increasingly sophisticated hypotheses to be formed and tested. And, with the advent of portable computers that can be used to enter data on site, analysis performed each evening can be used to guide researchers in the next day's digging.

Data presentation itself can also provide powerful analytical aid in the form of visualization tools. Graphics programs derive intricate stratigraphic sections directly from electronic measurements taken in the field. Views can be rotated and color-coded to emphasize particular characteristics. Distribution patterns that suggest how artifacts came to be deposited, or how they may have been used, become immediately apparent.

Gleaning in the field

Reaping the benefits

"You now have access within seconds to provenience information in great detail," says Dibble. "A water-movement hypothesis on one site was due entirely to computer graphics. We first noticed anomalies in the data, but seeing the graphic representation instantly revealed the pattern as one of water deposition-bones had floated to a bank while stones had sunk to the hollows." The ability to analyze each day's work immediately, combined with the fixed reference point provided by the total station, also frees the project director to excavate in the areas that appear most promising. "You don't have to dig in squares," says McPherron. "The electronic distance meter permits digs to be circular or irregular, and you never lose your coordinates."

Tying it all together

Dibble also hopes to improve methods to digitize three-dimensional object data for image manipulation and for calculation of volumes and proportions. "Manual measurements reduce the number of parameters you can record," he says, "limiting the analysis you can do on, say, skulls." He is also interested in algorithms to conjecturally rejoin images of broken artifact pieces: piecing together pot sherds, matching flint tool fragments and the matrices they were flaked from, and so on. Currently such matching is done by hand, by trial and error, over months or years. "You learn a lot," says Dibble, "but it's all anecdotal. You're always asking if you're learning enough to justify the time. There's no archaeologist who doesn't work with fragments. If you could find a way to do it quickly, it would truly revolutionize the field."

Sidebar: A Macintosh approach to site mapping

COMPASS, funded by a grant from Apple Computer, is a Macintosh-based system for mapping archaeological sites, architecture, and excavations that is being developed by Andrew Weiss at the Museum Applied Science Center for Archaeology (MASCA). The central software component is MiniCAD+, an computer-assisted drafting and database program that can be customized using a powerful PASCAL-like programming language. Survey coordinates from total stations (see main story) can be automatically plotted, combined with pre-existing maps that are scanned and traced, and then interactively drafted to produce topographic maps, site plans, thematic maps, and 3D reconstructions of buildings and sites. Data can be dynamically linked to objects on the maps and used to modify display attributes and to generate reports and analyses.

At the Classic Maya site of Copan, Honduras, COMPASS is being used to record a 400-year sequence of acropolis construction as part of a project directed by Dr. Robert Sharer, curator of the Museum's American Section. By tunneling into the 40-meter-high acropolis from a river cut, several thousand square meters of buildings, temples, platforms, and plazas have been revealed in three major building levels. Surveying in the dark, narrow, twisting tunnels, Weiss and anthropology graduate student Loa Traxler can quickly generate working maps of the architecture in the field on a Mac Portable, and a wide variety of final maps at the MASCA computer lab. The ultimate goal is to create a 3D animation of the growth of the Copan acropolis over time. COMPASS is also being used at University Museum projects at the Old Kingdom town of Abydos, Egypt; copper production sites in Thailand; and the Bronze Age mound of Tel es Sweyhat, Syria.

Shown in print version are an architectural plan of the Copan acropolis east court at about 700 A.D. and a 3D reconstruction of the platform structure "Ante," both produced using the COMPASS system.