Remains of the worlds first cities are the most noteworthy features of the landscape in southern Iraq, and for nearly two centuries archaeologists have probed them and puzzled over their artifacts. Built up over the course of five and a half millennia of intermittent occupation, these tells--mounds of building rubble and associated artifacts--can be as large as a mile in diameter; some rise more than 100 feet above the plain. Babylon, Ur, Uruk, Nippur and Kish have yielded abundant evidence of the material culture of Mesopotamian society. Thanks to their citizens relatively imperishable writing medium--clay tablets--they have also provided detailed textual testimony on political, intellectual, religious and social institutions.

Nevertheless, the physical and social organization of these most ancient cities is still poorly understood, for a variety of reasons. Paradoxically, the very richness of evidence has led to ignorance. The tells are so massive that even the best-financed field parties can excavate only tiny fractions of each city. More important, the arrangements of buildings that archaeologists uncover generally do not represent a city that actually existed at one particular time. The ancient inhabitants built on earlier structures in some cases and swept them away or modified them in others. One might imagine a similar problem facing archaeologists trying to understand London a few thousand years hence: they would be confronted by the mixed remains of modern skyscrapers, Victorian buildings, Norman castles and even a Roman garrison; reconstructing the city as it looked during any given period would be almost impossible.

Urban sociologists have long known that the plans of contemporary cities reflect patterns of social organization. Our own survey of non-Mesopotamian early cities shows that similar conclusions can be drawn about early urban sites. Where power is highly focused and based on coercion, centers of administration, religion, manufacture and trade cluster together, surrounded by residences of the elite. In contrast, societies in which diverse groups share control and in which decision making takes place at various levels of the social hierarchy show little or no evidence of such concentration. The intimate ties between elites and the rest of the population in these decentralized cities are mirrored by a mixture of rich and poor houses in each of the residential districts.

Where are humankinds first cities to be placed in this spectrum? Archaeologists have tended to emphasize centralization, but a close look at their reasoning, combined with our recent findings at a site called Mashkan-shapir, indicate that this view needs revision. Early excavations in Mesopotamia focused on seats of wealth and power--palaces and temples--and led researchers to take a similarly narrow view in reconstructing the society that built them. Yet concentration on the physical remains of high status obscured the fact that Mesopotamian texts do not identify clearly differentiated social classes. Instead they record the importance of general assemblies in decision making.

There may also be a more subtle bias at work. Historians recognize that industrialism and capitalism have so transformed the world that there are no modern analogues for ancient cities. Rather than considering a wide range of potential urban organizations, however, some scholars have perhaps too readily posited a unified model for a "preindustrial city" based on a few, well-studied (and centralized) examples. In devising this model they have rarely looked further afield than ancient Greece and sometimes no further than medieval Europe.

As a result, researchers have in effect taken for granted that cities in Mesopotamia were shaped by the same forces as were later European ones, among them a stable agricultural base and a fixed value for any given plot of land. In fact, the economic base in this region was anything but geographically stable--as indicated by the importance of nomadic herding. Even cultivated land was impermanent: annual floods, high evaporation rates and rapid poisonous salinization of land under cultivation led to a constantly shifting mosaic of rich irrigated fields and orchards, deserts and marshes, in which wealth or power had little to do with permanent control of a particular parcel. Detailed descriptions of many preindustrial urban civilizations--in West Africa, the Islamic Middle East and the New World at the time of the conquistadors--show considerable variability in organization; they also suggest a link between the permanence of agricultural land and the degree of social and political centralization. There is thus no reason to assume a priori that Mesopotamian cities were especially centralized.

An Untouched Site
THE ARCHAEOLOGICAL project that led to the discovery of Mashkan-shapir came about when we decided to attack the question of urban organization by seeking out a site that had been occupied only during a single period. The ruins of such a short-lived city would provide a snapshot of urban layout; by analyzing that physical organization, we would be able to draw some conclusions about whether it arose through coercion by priests and kings or by consensus among diverse segments of society. We were looking for a site that was of urban scale but otherwise quite different from the great cities that had been investigated before--a site occupied for a short time and left relatively undisturbed since its abandonment.

Looking through data collected by other archaeologists on the overall distribution of ancient settlements in Iraq, we chose a site that Robert McC. Adams, then at the University of Chicago, had found in the mid-1970s. He had given it a number but no name.

Our initial visit in January 1987 took place under less than ideal circumstances--it coincided with the first day of Iran's "final offensive" against Iraq during the Iran-Iraq war. Nevertheless, we could see that the site was littered with traces of occupation. Walls, pottery, graves, even ancient canals were all clearly visible across an area more than half a mile in diameter. Most of the shards dated to the first quarter of the second millennium B.C.E. Both the quantity and quality of the remains made it clear that the only recent visitors to this place, bedouin and their camels, had left it largely undisturbed.

The site stayed nameless for another two years while we made arrangements to return. Then, in 1989, shortly after we began our survey, we chanced on a chunk of baked clay near the remains of a gate in the city wall. The piece bore a cuneiform text. In short order we brushed clean 150 similar fragments, which together turned out to carry multiple copies of an inscription commemorating the walls construction. On the third piece we examined were three clearly legible signs of the four that are used to write Mashkan-shapir: the name of what was once, albeit briefly, one of the most important cities in the world.

Mashkan-shapir first appears in the historical record as a small sheep-rearing village on the fringe of the Mesopotamian heartland in the latter part of the third millennium B.C.E. It probably would have remained obscure had it not been for the political intricacies of the early second millennium. Shortly before 2000 B.C.E., an empire centered at Ur, which had controlled the entire alluvial plain, collapsed. For the next two centuries, several cities vied for hegemony, foremost among them Isin and Larsa. Larsa was probably more powerful, but Isin was farther upstream on the Euphrates River and so could impede its rivals access to essential commodities, such as wood, metal and stone, coming down the river from the northwest. Larsa countered by moving to control the eastern part of the valley and to secure access to the Tigris River. As Larsas northern outpost, Mashkan-shapir quickly grew to urban size, eventually serving as a second capital for the kingdom.

The city retained its role after Isin was defeated, because the rise of Babylon under Hammurabi's rule in the early 18th century B.C.E. created similar competition for control of access to goods. Soon enough, however, Hammurabi's conquests, which reunited most of southern Mesopotamia, made Mashkan-shapir's strategic position irrelevant. The city was abandoned around 1720 B.C.E., and the watercourses that had sustained it fell into disrepair. Mashkan-shapir disappeared into the desert.

There is ample historical testimony as to what kind of a place Mashkan-shapir was during its brief flowering. As a second capital to Larsa, it was a political city where much diplomatic activity was conducted. It was also a gateway to the trade route up the Tigris. Finally, it housed a primary sanctuary to Nergal, god of death, among the most powerful deities in the Mesopotamian pantheon. Because administration, religion and trade were the major activities in other large Mesopotamian cities, the organization of Mashkan-shapir has strong implications for that of other sites.

Difficulties in the Field
OPPORTUNITIES for work at the site have been constrained both by the time required to raise funds for extended fieldwork and by the political vicissitudes of the region. We undertook the initial exploration in two three-week campaigns in 1987 and 1989. From January through May 1990 we were able to work continuously at the site thanks to the support of the National Science Foundation, the National Endowment for the Humanities, the National Geographic Society and the American Schools of Oriental Research. We conducted a thorough surface survey and aerial reconnaissance to map the main features of the site, supplemented by modest excavations to determine the relation between surface traces and subsurface remains. We also made use of satellite imagery to understand local geology.

Shortly after we finished the first phase of the project and returned to the U.S., Iraqs invasion of Kuwait brought an end to archaeological work by foreigners. We had hoped that someday we would be able to use the results of our survey to make excavations at locations that would be most informative about the operation of this ancient city. We have now learned, however, that Mashkan-shapir was one of the victims of the looting of Mesopotamian archaeological sites that began after the 1991 Gulf War and has accelerated since 2003. A photograph taken from a helicopter by John M. Russell of the Massachusetts College of Art in January 2004 shows the complete destruction of much of the central part of the site.

Mapping Mashkan-shapir was not easy. The citys remains hardly form a tell in the traditional sense. The site has been seriously eroded by wind, which has reduced the latest buildings to foundations and left heavier artifacts exposed on the surface. Only in a few places do the contours of the site rise more than two meters above the plain.

As a result, larger architectural patterns, such as the path of the city wall, were difficult to make out from ground level. Indeed, Mashkan-shapir's wall is visible only from the air for much of its circumference, and in some places it has vanished entirely. Aerial reconnaissance was essential to our work. We took advantage of the sites strong winds to loft a camera-bearing kite. The elevation of the kite changed with wind velocity, and so the area included in each image varied considerably. We did our best to make up for these irregularities by taking many pictures and having a great deal of redundancy in our coverage.

The 1,600 aerial photographs we took would have been of little use in mapping were it not for software designed to facilitate analysis of satellite images and to create maps for city planners and geographers. We had marked the corners of the squares in our survey grid, each measuring 50 meters on a side, with crosses that would be visible in the photographs, thus making it possible to orient the images and compensate for geometric distortions caused by the swinging of the camera. Digitized versions of the images were corrected, recomposed at a uniform scale and assembled in a mosaic detailed enough to identify the position of each individual brick on the site's surface.

We combined aerial reconnaissance with a pedestrian survey based on the same 50-meter grid. A member of our team walked over each grid square in a pattern designed to ensure that nothing was missed, marking key features and artifacts with surveyor's flags. We indicated scattered fragments of bricks, potsherds, copper or ceramic slag, kiln fragments and bitumen on our map in terms of concentrations rather than individual pieces because there were far too many to count. From sampling in limited areas we estimated that at least 30 million pieces of pottery larger than a fingernail lay on the surface.

During the course of the survey, we also mapped graves, platforms of baked brick and mud brick, edges of canals and traces of the city wall. And we located more than 1,200 individual artifacts: tools, weapons, jewelry, pieces of statuary, plaques, figurines, small models and whole pots. The variable distribution of artifacts bore testimony to the complexity of this short-lived city.

For a broader view of the geography around Mashkan-shapir, we turned to an image taken by a French SPOT remote-sensing satellite in May 1988. The image reveals an ancient bed of the Tigris near the site, which explains why the city was built more than 20 miles from the nearest modern watercourse. The picture also shows the outlines of a series of canals originating in the river and bisecting the city.

A City of Canals
FROM THESE FINDINGS we were able to discern a great deal about life in Mashkan-shapir. Like all Mesopotamian cities of its era, it was surrounded by a mud-brick wall interrupted by a number of gates. We identified three gates. Two were for road traffic; they were located near major canals presumably because then, as now, roads ran along the waterways that served as vital lines of communication. The pylons of the third gate are on either side of a canal; perhaps it served to regulate either the flow of water or of water traffic.

Surprisingly, the city wall was not always built close to the edge of dense settlement. An area between the settlement and the wall remained empty, except for six isolated buildings that seem to have been storehouses. Many Mesopotamian texts suggest that commodity exchange took place near city gates, and so this space may have served as some kind of market. Another apparently unoccupied area, near a canal, may have been a garden. We know that some Mesopotamian cities incorporated such gardens because a map of Nippur, scratched on a tablet dated a few centuries later, shows a large garden planted on one corner of the city.

The city proper was laid out in five districts, separated by canals. A quad-rilateral region in the middle was surrounded by four larger areas to the north, south, east and west [see box on page 64]. The biggest districts, to the north and east, were themselves divided by canals. Broad harbors occupied two of the canal junctions; they must have been centers for commerce.

Many of the city's buildings were lined up along the waterways, making the canals an integral part of the urban fabric. Mashkan-shapir is not the only Mesopotamian city known to have been arranged in this manner, but the extent of canals elsewhere has been obscured by the great height of the deposits that have filled them in. Streets, which were also covered by debris at other sites, are just barely visible at Mashkan-shapir. Some follow the lines of canals; others cut across districts. On one such street, the baked-brick traces of a bridge (or perhaps two quays) can be seen where the street crosses a canal. Excavations at other Mesopotamian cities have shown that the main streets were supplemented by a network of alleys that gave access to individual houses.

The aerial photographs reveal additional demarcations. One wall surrounded a segment of the western district, and another wall cut across the southern part of the central one. Both are similar to an internal wall at Ur, which marked the sacred space surrounding the city's primary temples.

How were the political, religious, economic and social functions of Mashkan-shapir laid out within the physical structure defined by streets, canals and internal walls? From the ruins, we can tell where various activities took place.

The primary temple at Mashkan-shapir--that of Nergal--would have been the visual focal point of the city. Raised up on a platform or ziggurat, it could be seen for miles, and, much like a medieval cathedral, it was a symbol of power. The remains of the baked-brick and mud-brick platforms that seem to have supported the most important sanctuaries lie in the southern region, which was cut off by a canal. The religious character of these platforms was made clear by the discovery of 70 fragments of life-size terra-cotta statues of humans, lions, dogs and horses. Statues of lions frequently adorned the entrances of even minor temples in Mashkan-shapir's time, but archaeologists have found the more complex human and animal statues only at major cities such as Isin.

Across the canal in the central district was another area with religious overtones, identifiable by another platform--the only one so far discovered outside the religious quarter. This region contains traces of numerous burials and a concentration of such grave goods as jewelry and weapons. The area is also cut off from the rest of the settlement by a wall. Most graves at other Mesopotamian sites (and many at Mashkan-shapir) appear in domestic areas, and so this cemetery suggests that one segment of the society--perhaps that associated with the religious or administrative center--had a separate burial ground, probably with an attached temple.

A third distinctive region within the city was the walled-off enclosure in the west, which we believe was an administrative center. It contained a series of very regularly built structures, quite unlike the more haphazard construction of private houses. Although the structures clearly do not constitute a palace such as the one at Mari, which dates to the same era, they may have performed some of the same administrative functions. Like palaces elsewhere, this enclave at Mashkan-shapir is at the periphery of the city. Furthermore, during the 1990 season we excavated numerous unbaked clay sealings from the buildings in this enclosure. These pieces of clay, bearing the impression of carved stone seals, were attached to ropes that closed doors or were embedded in the material that closed storage jars, much as a wax seal might be placed on a letter. They have no place in domestic contexts.

The enclosure contained an oddity as well--a concentration of model chariots decorated with representations of the major gods of the city, Nergal and Shamash, the sun god. It is difficult to say for certain what function these small, two-wheeled vehicles served. Their prevalence in an official space rules out the possibility that they were toys. One possible interpretation is that they served to represent these gods in such legal issues as oath taking.

No Focus of Power
NO OTHER AREAS of specialized architecture at Mashkan-shapir have been identified by our survey. Everywhere else the artifacts may be characterized as domestic--figurines, small tools and weapons, commonplace jewelry (such as shell rings), and the traces of houses and burials.

The major roads and canals probably divided this large urban space into discrete residential neighborhoods, but if so they were not segregated by wealth or status. Stone bowls and metal objects (made of imported material and representing significant investments of labor) were scattered quite evenly. Had they been more prevalent in one area than another, an argument could be made for elite residential districts. Cylinder seals--ancient badges of office and items of considerable value in themselves--were also evenly dispersed.

Similarly, we found that manufacturing took place throughout the city. There seems to have been some concentration of "smokestack" industries, including pottery making on the southeastern, leeward side of the city, but artisans appear to have undertaken their work in houses surrounded by other residences. But this was not the only manufacturing district. Coppersmiths--their workshops marked by concentrations of copper slag--plied their trade along the main street of the central district, for example, but we found most decorative stones and their associated grinders in the southeast. There were two centers of ceramic production (marked by ceramic slag and kiln fragments), one in the north and one in the east. Both were clustered around the smaller canals that ran through these areas. In short, the spatial arrangement of manufacturing yields the same mostly decentralized picture as the arrangement of houses and artifacts.

Uncentralized System
OUR SURVEY of Mashkan-shapir does not appear to support a highly centralized model for Mesopotamian cities--or for their social organization. It is true that we identified clear foci of religion and administration. But they were sequestered in the southern part of the site. Indeed, they were separated from each other and from the rest of the city by major canals. Moreover, these potential power centers were far from the regions where commerce took place--the harbors and the known city gates. The production of goods seems to have been in the hands of artisans who lived within broader residential neighborhoods that housed both commoners and members of the elite.

The overall organization of Mashkan-shapir suggests that textual sources have not misled us about the broad involvement of Mesopotamian city dwellers in shaping their local power relationships. This conclusion, in turn, may reflect on earlier social structures: if citizens lived in a relatively uncentralized system during the Old Babylonian period, when Mashkan-shapir was thriving, it appears highly unlikely that local authority was more firmly in the hands of a small elite in earlier periods. As a result, the grounds for seeking the origins of civilization in processes of conquest and coercion seem, at the very least, far from compelling.

ELIZABETH C. STONE and PAUL ZIMANSKY have collaborated for more than 30 years on the study of ancient cities in the Middle East. Stone, who received her Ph.D. in 1979 from the University of Chicago, is professor of anthropology at Stony Brook University, New York. Zimansky, who received his Ph.D. in 1980 from the University of Chicago, is professor of archaeology at Boston University.