McGuire Gibson
Professor of Mesopotamian Archaeology
The Oriental Institute
The University of Chicago

(This article originally appeared in Al-Rafidan, Vol. XIV, 1993, and is made available electronically with the permission of the editor.)

The importance of the Mesopotamian holy city, Nippur (Fig. 1), is reflected even today in the great size of the mound, Nuffar (Fig. 2), located between Baghdad and Basra in southern Iraq. Nippur was one of the longest-lived sites, beginning in the prehistoric Ubaid period (c. 5000 B. C. ) and lasting until about A. D. 800, in the Islamic era (Gibson 1992).

From earliest recorded times, Nippur was a sacred city, not a political capital. It was this holy character which allowed Nippur to survive numerous wars and the fall of dynasties that brought destruction to other cities. Although not a capital, the city had an important role to play in politics. Kings, on ascending the throne in cities such as Kish, Ur, and Isin, sought recognition at Ekur, the temple of Enlil, the chief god of the Mesopotamian pantheon (Fig. 3). In exchange for such legitimization the kings lavished gifts of land, precious metals and stones, and other commodities on the temples and on the city as a whole. At the end of successful wars, rulers would present booty, including captives, to Enlil and the other gods at Nippur. Most important, kings carried out for the city elaborate construction and restoration of temples, public administrative buildings, fortification walls, and canals. Even after 1800 B. C., when the Babylonians made Marduk the most important god in southern Mesopotamia, Enlil was still revered, kings continued to seek legitimization at Nippur, and the city remained the recipient of pious donations. The city underwent periodic declines in importance [Gibson 1992) but rose again because its function as a holy center was still needed. The greatest growth of the city (Fig. 2), which occurred under the Ur III kings (c. 2100 B.C), was almost matched in the time of the Kassites (c. 1250 B.C.) and in the period when the Assyrians, from northern Iraq, dominated Babylonia (c. 750-612 B.C.).

The strength of Mesopotamian religious tradition, which gave Nippur its longevity, can be illustrated best by evidence from the excavation of the temple of Inanna, goddess of love and war. Beginning at least as early as the Jemdet Nasr Period (c. 3200 B.C.), the temple continued to flourish as late as the Parthian Period (c. A.D. 100), long after Babylonia had ceased to exist as an independent state and had been incorporated into larger cultures with different religious systems (Persian, Seleucid, and Parthian empires). The choice of Nippur as the seat of one of the few early Christian bishops, lasting until the city's final abandonment around A.D. 800, was probably an echo of its place at the center of Mesopotamian religion. In the Sasanian Period, 4th to 7th Centuries, A.D., most of the major features of Mesopotamian cultural tradition ceased, but certain aspects of Mesopotamian architectural techniques, craft manufacture, iconography, astrology, traditional medicine, and even some oral tradition survived, and can be traced even today not just in modern Iraq but in a much wider area.

The origins of Nippur's sacred character cannot be determined absolutely, but some suggestions can be made. The city's special role was derived, I would suggest, from its geographic position on an ethnic and linguistic frontier. To the south lay Sumer, to the north lay Akkad; the city was open to the people from both areas and probably functioned as an arbiter in disputes between these potential enemies. The existence of the frontier can be demonstrated from texts as early as the Early Dynastic III period (c. 2600 B.C.), when Sumer was the dominant cultural entity. In tablets from Shuruppak, a city 45 kilometers southeast of Nippur, more than 95% of the scribes had Sumerian names, while the rest had Akkadian names. In contrast, at Abu Salabikh, 12 kilometers to the northwest of Nippur, literary and other scholarly texts were written in equal numbers by Sumerian and Akkadian scribes [Biggs 1967]. But, Biggs notes that in the preparation of administrative texts at Abu Salabikh there was a greater representation of Sumerian scribe names, about 80%. This fact may indicate that although Akkadians were deeply involved in all aspects of life in the area just north of Nippur, government affairs may have remained predominantly the preserve of Sumerians in the pre-Sargonic period. For Nippur, we do not know as yet what percentage of scribes had Akkadian names in Early Dynastic III, but Biggs [1988] has suggested that the percentages at Nippur would be more like those of Shuruppak than like those of Abu Salabikh. I would suspect, however, that the percentages for non-governmental texts were closer to those at Abu Salabikh, with a good number of Akkadian scribes in evidence.

As is the case with the world's other holy cities, such as Jerusalem, Mecca, and Rome, Nippur was a vibrant economic center. Besides the economic benefits derived from gifts and on-going maintenance presented by kings and rich individuals, there was probably a continuing income from pilgrims. Nippur was the center of an agricultural district, with much of the land in the possession of temples. The temples produced manufactured goods, predominantly textiles and finished items, some of which were meant for export. But the temples were only part of the economic picture [Maekawa 1987]. Even though it was more dominated by religion than other towns, Nippur, like them, had a mixed economy, with governmental, religious, and private spheres (see, e.g. Westenholz [1987]). Steadily accumulating evidence indicates that the public spheres were closely integrated, with final control in the hands of government officials (see esp. Maekawa [1987]).

The work-force for much of the large-scale manufacture was probably connected with the major institutions, especially the temples. As in most countries until modern times, the temples in Mesopotamia had an important function as social welfare agencies, including the taking in of widows and orphans who had no families or lineages to care for them [Gelb 1972]; temples also were the recipients of war prisoners, especially those from foreign lands, who worked in agricultural settlements belonging to temples or in other temple service [Gelb 1973]. Such dependent people probably worked for generations in the service of one temple as workers and soldiers (gurus/erin), rather than as slaves (sag) [Gelb 1973: 94-95].

All institutions, whether the governor's palace, a government-sponsored industry, or a temple, were not just buildings and not just abstract bureaucratic hierarchies or economic establishments, but were social organizations within a broader social network. As happens in most societies, large institutions in ancient Mesopotamia tended to be dominated by families, lineages, and even larger kinship groups and I would argue that it is this web of kinship that furnishes the long-term, underlying continuity for civilizations, making it possible to reassemble the pieces even after disastrous collapses. For Mesopotamia, the role and power of such kinship organizations is best observed ironically in the Ur III Period, the most centralized, bureaucratized period in Mesopotamian history. The abundance of records of administrative minutiae allows the reconstruction not just of the administrative framework, but of the social network underlying and imbedded within it. The best reconstruction of such a kin-based organization within an institution is Zettler's [1992] work on the Inanna temple. One branch of the Ur-me-me family acted as the administrators of the temple, while another dominated the governorship of Nippur and the administration of the temple of Enlil.

It is important to note that the Ur-me-me family remained as adminstrators of the Inanna temple from some time within the Akkadian period to at least as late as the early years of the Isin dynasty. Thus, while dynasty replaced dynasty and the kingship of Sumer and Akkad shifted from city to city (Akkad to Ur to Isin) the family remained in charge of the Inanna temple.

From the listing of members of two and three generations as minor figures on the temple rolls, it is clear that it was not just the Ur-me-me family that found long-term employment within the temple's economic and social skucture. Through the continued association of families with the institution, not only were generations of people guaranteed a livelihood, but the institution was guaranteed a cadre which would pass on the routines that made the institution function. The temple could add key personnel not only through a kind of birth-right (family or lineage inclusion), but also through recruitment; important individuals within the institution's adrninistration would have acted as patrons not just for nephews, nieces, and more distant relatives but also for unrelated persons. By incorporating clients of its important men and women, an institution could forge linkages with the general population in the city as well as in the supporting countryside and in other cities; these recruits, in taking up posts within a temple, a municipal establishment, the royal bureaucracy, or in a large family business, would ensure that the patron had loyal adherents.

We know from cuneiform texts found at Nippur and elsewhere that the temples, rather than controlling the cities through a "Temple Economy," as was proposed earlier in this century, were under supervision by a king or a royally appointed governor, even in the Early Dynastic III period (c. 2600 B.C.) [Foster 1981; Maekawa 1987]. In the Akkadian period (c. 2300 B.C.), the temples of Inanna and Ninurta seem to have been under very close control of the governor, but the ziggurat complex, dedicated to Enlil, appears to have been more autonomous, reporting directly to the king in Agade [Westenholz 1987: 29]. During the Ur III period (c. 2100 B.C.) at Nippur, the administrator of the Inanna temple had to report to his cousin, the governor, on the financial affairs of the temple, and even had to go to the governor's storehouse to obtain the ritual equipment for specific feasts of the goddess [Zettler 1992]. The situation was much the same in the Isin-Larsa period, with texts from one agency (presumably the governor's office) recording distribution of goods to several temples; it is unfortunate that a recent article [Robertson 1992] revives, again, the notion of "temple economy" to cover these transactions.

The characteristics of administration and support that can be reconstructed from texts for a few temples at Nippur must be assumed to have been operative in the rest of Nippur's temples. The relationship of those temples to governmental institutions and to private entities and individuals is only beginning to be worked out. To reconstruct life in ancient cities one cannot rely on written documents alone, since they do not cover the entire range of ancient activity. Often, crucial insights can be obtained by the correlation of non-inscribed evidence, for instance the repeated co-occurrence of a set of artifacts in one type of find-spot. Especially valuable are correlations that illustrate human adaptations to natural environrnental conditions. When one can bring texts into such correlations, truly innovative syntheses can be made. Whenever possible, documents must be viewed in their archaeological contexts, treating them as an extraordinarily informative class of artifacts to be studied in relationship to all other items. When such relationships are studied, a much more detailed picture emerges. Although that procedure would appear to be self-evidently valuable, it is rare that texts have been treated in this manner. At Nippur, we have made a concerted effort to combine all kinds of information in our interpretations of the site, and we think that we have made some important discoveries by so doing.

Nippur has been the focus of major excavation since 1889 when the University of Pennsylvania opened the first American expedition in the Middle East. Finding the site a rich source for cuneiform tablets, that expedition continued to excavate at Nippur until 1900 [Hilprecht 1903; Peters 1897]. The main achievements of the expedition were to locate the ziggurat and temple of Enlil and to recover more than 30,000 cuneiform tablets of extraordinary literary, historical, grammatical, and economic importance. More than 80% of all known Sumerian literary compositions have been found at Nippur. Included were the earliest recognized versions of the Flood Story, parts of the Gilgamesh Epic, and dozens of other compositions. It was these Sumerian works, plus an invaluable group of lexical texts and bilingual (Sumerian/Akkadian) documents that allowed scholars to make real progress in deciphering and understanding Sumerian. As important in historical terms are royal inscriptions from all periods, especially those of the Kassite Dynasty which ruled Mesopotamia from about 1600 to 1225 B. C. More than 80% of our knowledge of this dynasty has come from Nippur texts. In a special category of Nippur texts are the business archives of the Murashu family, merchant bankers who controlled vast commercial and agricultural interests under the Achaemenid Persian kings (c. 500 B.C.) [Stolper 1985].

For almost a half-century after the University of Pennsylvania left the site, Nippur lay unexcavated. In 1948 the University of Chicago initiated a Joint Expedition to Nippur with the University of Pennsylvania. It was felt at that time that although Nippur had been inundated by a sea of dunes since the 1920's, the information to be gained, especially on Sumerian culture, justified the extraordinary expense and difficulty caused by those dunes. A stated goal of the new excavations was to establish an archaeological context for the extraordinary artifacts, especially the tablets, that the earlier expedition had found. When the University of Pennsylvania withdrew from the expedition in 1952 it was succeeded by the American Schools of Oriental Research until 1962. The University of Chicago has continued its commitment to the site to the present day, and the last season of work in the winter of 1990 constituted the nineteenth campaign since 1948.

For the first three seasons of modern work, 1948-52, excavation was concentrated on the area of the ziggurat and the adjacent mound called Tablet Hill (Fig. 2). The early Pennsylvania excavators gave the name Tablet Hill or The Scribal Quarter to that mound in the belief that all or most of the scribes at Nippur had lived in that one part of the site. Although many important tablets were found in Tablet Hill, a study of all the records of the old Pennsylvania expedition shows that even more texts were found in the southern end of the West Mound. Recent excavations have proven that tablets, including school texts, probably are to be found in every part of the site. Because it had more than a hundred temples [Berhnardt and Kramer 1975] as well as governmental offices and numerous private businesses, it is not surprising that written records are to be found all over Nippur. But, so far, Sumerian literary texts do appear to be more highly concentrated at Tablet Hill.

The Joint Expedition's work in Trenches TA and TB on Tablet Hill yielded a valuable sequence of houses with artifacts in situ. This sequence, especially the pottery, dating from the Akkadian through the Achaemenid period (2300-500 B.C.) became a standard of reference for all of Mesopotamia [McCown and Haines 1967].

While working on Tablet Hill, the expedition began to make exploratory trenches at numerous locations in the eastern half of the site. In one of these trenches, R.C. Haines exposed the North Temple [McCown et al. 1978], dedicated to a god/goddess as yet unidentified (Fig. 2). More important, another trench encountered the temple of Inanna [Zettler 1992], goddess of love and war, one of the most important deities in the Mesopotamian pantheon. For ten years (Seasons 3-8, 1953-62) the expedition concentrated on this one area, and exposed seventeen rebuildings of the temple, one upon another, dating from the Jemdet Nasr Period (c. 3200 B.C.) until the Parthian Period (c. A.D. 100). As with other temples built of unbaked mudbrick, when the Inanna Temple began to age, it was demolished and a new, larger, more elaborate building was constructed upon its ruins. This long sequence of temples, especially the earliest ten (3000-2200 B.C.) with their thousands of artifacts (statues, reliefs, stone bowls, cylinder seals, and pottery), has furnished yet another standard of comparison for all other Mesopotamian sites [ Hansen 1965; Porada et al. 1992].

In 1964, Chicago, by then the sole sponsors of the Nippur expedition, signed a revised agreement with the Iraqi government, promising to continue excavating on a long-term basis. It was decided that the ziggurat area should once more be the focus of research, since that is the most important single structure at Nippur. This focus required the re-excavation of a large Parthian fortress that Pennsylvania had exposed partially in the 1890's. After recording the fortress, the expedition was supposed to demolish it so that the Sumerian levels around the ziggurat could be exposed fully. The 9th and 10th Seasons (1964-67) were expended in excavating the fortress, but when the task was finished, the expedition was not permitted to remove the remains to continue its proposed program because the fortress was judged to have value for tourism.

For five years, the site lay neglected. In 1972, when I became director of Nippur, I instituted a new program, meant to bring to light not just the religious aspects of the city, but its governmental and private sectors as well. I wanted to investigate the city's origins and history, the function of various parts, and the relationship of the city to its region and its environment. I proposed to examine the city walls, put trenches into parts of the site that had never been sampled, and also try to fill gaps in the Mesopotamian sequence (especially the Akkadian and Kassite Periods), and examine the later periods (Sasanian and Islamic) that had rarely been excavated systematically in Mesopotamia. Very important in our work was a commitment to linking archaeological to epigraphical data and an attempt to understand the ecological and social systems of ancient Nippur. We also introduced a new, up-to-date method of excavation, recording, and analysis of material. And we proposed to bring to the archaeology of the historical periods of Mesopotamia some of the techniques and theoretical viewpoints, called the "New Archaeology," that had been developed for prehistoric sites elsewhere. Such an approach was new to Mesopotamia, as it was to the historical ranges of most other parts of the Near East. Now, twenty years later, these methods and viewpoints have become commonplace not just in Iraq, but in the area as a whole.

To carry out our new program, we turned away from the eastern mounds, which were considered to be the more religious side of the city, and began to work on the West Mound, which had not been touched since 1899. Our first operation, WA (=West Mound, Operation A) was located in the bottom of a huge pit left by Pennsylvania (Fig. 2). Here, that expedition had found a large villa of Parthian data (c. 100 A.D.), and, in nearby locations, the Murashu archive and a group of Kassite administrative tablets. We thought we had a chance to expose, eventually, not only buildings that might relate to the Murashu family, but also a Kassite administrative building. Very soon we realized that we had come down upon yet another sequence of temples (Fig. 4), dating from at least the Ur III (c. 2100 B.C.) to the Neo-Babylonian period (c. 600 B.C..). We worked here for three seasons, having great difficulty because of the continual movement of dunes into our excavations, and were able to expose only a part of successive levels of a very large and important temple. We could not identify the deity venerated there. We assumed that this sequence of buildings would be much older than the lowest level we reached at that time (Ur III) and that it would rival the Inanna Temple in importance if conditions made it possible to carry the excavations to conclusion.

In Area WB, toward the south end of the West Mound (Fig. 2) we did, in fact, discover a totally unexpected Kassite administrative building, a badly destroyed palace (c. 1250 B.C.). This building (Fig. 5), half the size of the Kassite royal palace at Dur Kurigalzu near Baghdad, was the governor's palace, according to tablets found there [Gibson 1978a]. We know from other cuneiform documents, found by the old Pennsylvania expedition in the area to the south of WB, that the administrative center of the city and the province was located in this area from at least the Akkadian Period (c. 2300 B.C.) to the 7th Century B.C. The existence of governmental buildings in this part of the city must explain the great number of tablets found in this part of the site by the old Pennsylvania expedition.

Directly below the Kassite palace in Area WB was an Old Babylonian house (c. 1750 B.C.) owned by a family of bakers, who used the front half of the building as an office and shop and the space outside for the baking of bread and meat [Gibson 1978a]. Texts found in the house show that the family baked on contract for the city administration, temples, and individuals. On the floor of the building we found dozens of objects left in place-pottery, a bread oven, grinding tools, cuneiform tablets, and other items. The debris on the last occupation floor gave the impression that the occupants had left suddenly, expecting to return soon, but never did. In time, sand drifted over the artifacts on the floor, and the walls of the house were eroded by rain and finally collapsed. This dramatic instance of sudden abandonment brought into clear focus evidence of similar breaks in stratigraphy in other Old Babylonian contexts on the site. We realized that there had been a crisis in the history of the city that had resulted in a total, or almost total, abandonrnent. The cessation of dated texts at around 1720 B.C., noticed by earlier excavators but not discussed [McCown and Haines 1967: 74-76], had to be correlated with the archaeological evidence. I knew that there was a similar halt in dated texts at other sites in Babylonia (e. g., Ur, Larsa, Isin) during the reign of Samsuiluna, and I knew that only those cities lying along or close to the river's western branches, such as Babylon, Kish, Sippar, Borsippa, and Dilbat, continued to produce dated texts. I began to suggest in lectures, as early as 1973-74, that there may have been a general catastrophe in Babylonia at that time, due to a major environmental crisis, probably the shifting of water away from the main branch of the Euphrates that had passed through Nippur. Elizabeth Stone, in an important restudy of Tablet Hill [Stone 1977; 1987], summarized the available evidence for the crisis and abandonrnent at Nippur. Hermann Gasche [1989: 109-43] subsequently laid out the evidence, in very graphic form, for a general collapse of central and southern Babylonia during the period.

The catastrophic abandonment of the heart of Babylonia, with a subsequent formation of dunes, was not to be reversed until about 1300 B.C., when irrigation water was brought back to the center of the country by the Kassite dynasty. As the Kassites began to revive Nippur and the other cities, they must have done a kind of archaeology to allow them to identify individual buildings. Only such a procedure can explain how, after hundreds of years of abandonment, the Kassites could have placed their versions of the Inanna Temple, the North Temple, the temple in WA, and other buildings, over their Old Babylonian predecessors. The reconstruction by the Kassites of this holiest of cities on so grand a scale and with such care for detail is consistent with that dynasty's deliberate efforts to revive other aspects of ancient Mesopotamian culture, such as a resurrection of the long-dead Sumerian language and literature.

Our appreciation for that effort of reconstruction was heightened by work we carried out on the lowest parts of the site. In our 13th Season of excavation, 1975, we began to investigate Area WC in the southernmost corner of the city (Fig. 2). We had noticed that a ridge, appearing on an air photograph of the site (Fig. 6), seemed to coincide with a corner of the city wall on a Kassite map that had been found at Nippur by the University of Pennsylvania (Fig. 7). This city plan shows the ziggurat complex, Ekur and Ekiur, "the canal in the middle of the city," and a number of city gates, as well as measurements along sections of the city wall. I was already convinced that Samuel Kramer [1956] had been correct in arguing that the Kassite map represented the entire city, not just the eastern half, as other scholars have thought [Fisher 1905]. Miguel Civil, our expedition epigrapher, in conducting a new study of the map, showed me that the measurements along the walls made sense only if the entire city were represented and if the map were oriented as I present it here.

The correct orientation of the map was proven by the cutting of trenches WC-1 and WC-2 (Fig. 2) across the ridge at the southern corner of the site, where we found evidence of a city wall more than 14 meters in thickness [Gibson 1978b: 118-20]. There is difficulty in overlaying the ancient plan on the topographical plan of the site (Fig. 8), however, because of inaccuracies in the angles of the city wall as given by the Kassite scribe; if Ekur and the southern corner of the city (Area WC) are aligned, many of the other features are skewed and if the river Euphrates is laid over the Kassite canal that we excavated to the west of WC, another set of features is then skewed. Even with the difficulty in alignment, however, the similarity of detail in both maps is obvious. By excavation, we also determined that an ancient canal west of WC-1 was Kassite in date and it lay approximately where the Euphrates is located on the ancient map. We even located what must be the Birdu canal, which branches off from the Euphrates at the western corner of the city. In a long trench at the northwest of the mound, we discovered at four meters below the present plain level many thousands of Kassite pottery vessels embedded in greenish clay, laid down in conditions that our soil specialist interpreted as ponded water. This area on the ancient map is marked hirtum, which can be translated "moat, " that is, an area of ponded water. In summary, I can say that we have been able to verify Kramer's interpretation of the map by a combination of archaeological, geomorphological, and philological evidence.

While we worked for three seasons on the southern end of the mound, exposing private houses of several periods just inside the city wall, the dunes that had hampered our excavations on the high mounds began to retreat rapidly towards the east. This phenomenon allowed us to carry out investigations of the city wall east of the ziggurat (Areas EA, EB, EC) and a very important operation, TC, at the end of the TA trench on Tablet Hill (Fig. 2). In Area TC, we were able to prove that not only had there been a crisis and abandonment of Nippur during the Old Babylonian period, but also a second crisis in the period after the Kassite occupation. James A. Armstrong, in an outstanding example of archaeological excavation and reasoning [Armstrong 1989], proved that the original excavations from 1948 to 1952 had involved a misunderstanding of the stratigraphy. When correctly reassembled, the evidence clearly shows sharp breaks in pottery traditions not only in the Old Babylonian period but also in the post-Kassite period. And in both periods of abandonment, dunes invaded the site, just as they have done in the past hundred years. The abandonment at the end of the 2nd Millenrlium meant that there was the necessity for a second revival of Nippur, which seems to have taken place in the 8th Century B.C., reaching its peak under Assurbanipal in the late 7th Century. The breaks in the pottery sequence, which reflected the abandonments, had been somewhat apparent in a table in the original publication of Tablet Hill [McCown and Haines 1967: Table II] but were made indistinct by the confusion of stratigraphy. Armstrong's revision of that table, now nearing completion, will illustrate very graphically the two gaps in occupation of the city. We cannot state, absolutely, that the entire city was abandoned each time; there is a possibility that the ziggurat and the Enlil temple may have survived with a small staff that could derive water from wells and could have been supplied with food from the irrigated areas to the west. In future, we hope to investigate the problem in the ziggurat area.

By 1989, with most of the sand off the site, we decided to return to Area WA to reopen the investigation of the sequence of temples that we had found in the early 1970's. In the years that we had been working on the lower parts of Nippur, we had achieved several of our objectives, such as sampling unexcavated parts of the city through surface collection of sherds and soundings; we have not yet uncovered any industrial areas except the bakery of Area WB and some areas of pottery production of various periods, but we do have a better idea of the history of occupation of the city as a whole; we have also examined the city walls in Areas WC, EA, EB, EC (Fig. 2); and, by the inclusion of environmental specialists on the expedition since 1972, we have made significant strides in understanding the environment both in modern times and in antiquity (e.g., Brandt [1990] ).

Our first step in reopening work on the high mound in 1989 was to make a sizable excavation of Sasanian and Islamic levels in Area WG, just to the southwest of Area WA. With this operation we achieved yet another of our long-range goals, the systematic investigation of the last two periods of occupation at Nippur. The excavation of this area was also meant to give us room to expand Area WA toward the location of the Murashu archive. At the same time, we sank a deep pit (WF) in the southern end of WA, in order to expose levels that would make possible a revision in our understanding of the transition from the Early Dynastic to the Akkadian period.

In the winter of 1990 we resumed excavation on the temple sequence in Area WA. Although we did not expose the entire temple at any level, we were able to gain enough information to hazard an estimate that the latest (Neo-Babylonian, c. 600 B.C.) building was probably about 100 meters by 40 meters in size. In addition, although only the bottoms of the walls of the 7th Century and Kassite (13th Century) levels remain (Fig. 9), we were able to recover enough artifacts in these buildings to identify the deity to whom this temple is dedicated. On floors, and buried in the plaster on walls, we found several figurines of dogs (Fig. 10). We also found fragmentary figurines of human beings in attitudes of pain; for instance one with his hand to his throat, another with one hand to his head and one to his stomach (Fig. 11) and (Fig. 12). Knowing that the dog was the special symbol of Gula, the goddess of medicine, we began to hypothesize that this was her temple, even though there are very few mentions of a Gula Temple in Nippur tablets. The identification was made positive by the finding of a small fragment of a lapis lazuli disc with the incription a-na dGu-la "to Gula." Muhammad Ali Mustafa, an Iraqi scholar, had excavated a small Kassite mound near Dur Kurigalzu, where he had discovered dozens of similar figurines [Mustafa 1947]. On some of his animal figurines there were prayers to Gula, making certain the association of such figurines with the goddess.

We had been assuming since 1973 that the WA temple, being so large, might be dedicated to Ninurta, who is the second most important god at Nippur. It may be proven in future that the temple of Gula lies beside a large temple dedicated to Ninurta, but it is more likely that the part of the WA temple that we have thus far exposed is only the Gula section of the temple of Ninurta, since Gula was the wife of Ninurta from the Old Babylonian period onwards. The situation in WA may, then, be the reverse of what has been found at Isin, where Gula, the chief deity of that city, shared her temple with Ninurta [Hrouda 1981: 200]. At least one other scholar, A. Westenholz [1987: 97-98], has argued that the Ninurta temple is to be located in the West Mound.

Our plan to continue excavation of the Gula Temple in the winter of 1991 was cancelled by the Gulf War. We still hope to spend several years exposing the temple systematically, level by level, until we reach the earliest one. We wish to examine not just the temple but also the area around it, to try to put it in its urban context. And we will be conducting analyses of soil and floral and faunal remains that can expand our knowledge of the environment of ancient Mesopotamia. In the early levels, we know that the temple will not be dedicated to Gula, whose name appeared only at about 2,000 B.C.; the early versions of the temple probably will be dedicated to a Sumerian counterpart, Bau or another of the goddesses of medicine.

If we can carry out our program, we may gain important new information on Mesopotamian medicine, on its practitioners the asu and the asipu, as well as on their relationship to the temple. We know that the asu was something like a modern doctor, making diagnoses, prescribing remedies, and recording the results. We also know that the asipu was a magician, performing rituals and giving potions. We do not know how the two professions related to Gula or to her temple. Perhaps the Mesopotamians dealt with illness as many people do today. They went to the doctor for a cure. If that didn't work, they tried alternative medicine-a faith healer or a folk healer. Maybe at the same time, they went to the temple to leave a figurine or obtain a figurine and say a prayer.

In their attitude toward medicine, as in other things, I would suggest that the ancient people of Nippur and of Mesopotamia in general, rather than having "mythopoeic minds" [Frankfort 1946], were only a little less complex than we are and probably just as sensible. As is the case with most people, the ancient Mesopotamians had contradictory aspects to their personalities, being religious when it was called for but forgetting religion in most situations. In my understanding of written records, the ancient Mesopotamians, even those at a religiously dominated city such as Nippur, were in most aspects of life very pragmatic and extremely rational in working out problems. They were the inventors of many procedures that still underlie modern life, e.g. in commerce and law. Their art objects show an ability to objectify reality, but there are also artifacts, such as figurines of monsters, that express superstition and fear. They could express lofty ideas of justice and mercy, but punish with severity, and even carry out acts of senseless brutality. And besides great art and literature, they could create riddles and jokes and probably pornography.

As a culture, ancient Mesopotamia must be recognized as a tremendously resilient and strong tradition. In a harsh and demanding environment, Mesopotamians created the world's first civilization and sustained it for more than three thousand years. That culture was, in fact, so elaborate, changing, and elastic an adaptation that it could be maintained even when major states collapsed. Nippur, its spiritual center, was probably more intimately involved in that continuation of tradition than most other sites. The city is, then, an extraordinarily important focus for sustained research and deserves continued excavation well into the future, even though there has already been a century of archaeological research on the site.