City Gate: Pottery Analysis

Introduction

The ceramics recovered from Operation CG span the entire occupation of city gate as well as the lower town itself, dating from the end of the Ninevite 5 period (Leilan Period IIId) through the Habur Ware period (Leilan Period I) (ca. 2600-1700 BC). The only notable absence from this mid-third-second millennium span coincides with the abandonment of this settlement from approximately 2200-1900 B.C. during the Habur Hiatus. As a result, this ceramic sequence parallels those retrieved from operations 7 and 8, two soundings in the Lower Town, excavated in 1991, which provide a similar ceramic sequence (Pulhan 2000). The excavation techniques used in Operation CG affected the recovery of ceramics in both quantitative and qualitative terms. The emphasis on dry-sieving and flotation resulted in the recovery of far larger numbers of sherds than would have occurred without the use of these procedures. As a result, despite the extremely small “excavated area” we recovered thousands of sherds. Similarly, the procedures used for the high control area, meant that mqny very small and fine ware sherds were recovered, as there were fewer opportunities to overlook them. This means that our percentages of fine ware sherds may be exaggerated in comparison to those that would have been retrieved using normal excavation techniques. At the same time, of course, this percentage probably more accurately reflect ancient realities, as it seems likely that large number of fine ware sherds are consistently overlooked during the course of excavation. The constraints of working in this area, the excavation techniques used, the nature of the occupation, and the creation of the post-excavation GIS have all affected the ways that we have chosen to analyze this pottery. Along with the C-14 dates and the sealings, the pottery recovered here provides much of the chronology of Op. CG. At the same time, given the non-domestic nature of the occupation at the city gate, the limited number of ceramic types is of interest, as it helps us to more clearly define the nature of the administration. In total, 1031 diagnostic sherds from the high and low control areas were analysed, while 2404 non-diagnostic sherds from the high control area were analysed. The following report begins with a brief consideration of the ceramics by phase, then discusses the variables used to reach the ceramic periodization, before concluding with a discussion of the functional character of this assemblage.

Phases

Phases 1-2: Leilan IIId

Phase 1: 24 diagnostics, 44 non-diagnostics
Phase 2: 132 diagnostics, 172 non-diagnostics

Both of these phases were dominated by fine ware cups, and contained several sherds with Ninevite 5 incising (5, 21% and 10, 8% respectively, see Fig. 1: 2-3). Phase 2 also had several medium ware sherds with lugs (n=4, 3%, see Fig. 3: 19, Fig. 1: 12) Unlike later phases, phase 1 contained fewer fine ware sherds-only 29%, while by phase 2 fine ware cups were overwhelmingly dominant; fine ware sherds form 69% of this assemblage. The same three fine-ware cup rims were the most common diagnostics in these two phases: simple straight wall (phase 1: 3, 12.5%; phase 2: 22, 17%, see Fig. 1: 7), simple curved wall (phase 1: 3, 12.5%; phase 2: 29, 22%, see Fig. 1: 8-11), and chai tea cups (phase 1: 3, 12.5%; phase 2: 19, 14% Fig. 1: 1-2, 4). In phase 1, collared, closed pot rims were also common (3, 12.5%, Fig. 2: 13)), this pot rim only appeared in this phase. In phase 2, a beaded bowl rim became more common and persisted in later phases (Fig. 1: 6). Base types during these phases were similarly conservative, round, mini-flat (Fig. 3: 20; Fig. 1: 1), and pointed bases (Fig. 3: 22-3; Fig. 1: 9) were the most common.

Phases 3-4: Leilan IIa

Phase 3: 132 diagnostics, 245 non-diagnostics
Phase 3a: 102 diagnostics, 1163 non-diagnostics
Phase 4: 78 diagnostics, 170 non-diagnostics

The pottery assemblage retrieved from these phases is almost identical to the one from phases 1-2. The largest difference is the decrease in the number of Ninevite 5 incised sherds. The very high percentage of green and buff fine ware sherds in these three assemblages (phase 3: 160, 65%, phase 3a: 1046, 98.5%, phase 4: 162, 95%) is similar to the pattern already established in phase 2. The three most common rim types: simple straight wall, simple curved wall, and slightly everted, carinated remain the same. These three open forms (small cups or bowls) continue to dominate the assemblage, comprising 73% of all rim types in phase 3 (n=61), 69% in phase 3a (n=69), and 54% in phase 4 (n=35). Otherwise, several examples of a fine, closed collared rim, which is similar to the everted, carinated “chai tea-cup” appeared in phase 3 (5, 4%, see Fig. 1: 5 for an example from phase 5), while a closed, everted rim was common in both this phase and phase 3a (n=4, 3%; n=4, 4% respectively, see Fig. 2: 15). Round, pointed, and mini-flat bases continued to be the most common in all phases, while flat (Fig. 3: 24) and pedestal bases (Fig. 3: 21) were also present. The great similarity between the assemblages from phases 2-4 no doubt is the result both of the short time span represented (see Radiocarbon Dates), as well as continuity in the function of this area.

Phases 5-5a: Leilan IIa

Phase 5: 58 diagnostics, 64 non-diagnostics
Phase 5a: 76 diagnostics, 109 non-diagnostics

During phase 5, the area under consideration seems to have shifted in function. The changing function of this area seems to have meant the use (and primary deposition) of fewer pottery sherds. These lots are more varied than the preceding ones in terms of ware, rim type, and base type. Similarly, few sherds that belonged to reconstructable vesels were found-suggesting that most of the pottery from this phase comes from secondary deposition in phase 5. As phase 5a represents a trash deposit we can assume that all the recovered sherds are from secondary deposition. The biggest difference between this and the preceding phases is the change of percentages in both fine ware and rim sherds. Sherds which could be characterized as fine ware slipped down to 59% of the assemblage in phase 5 (n=73), and 46% in phase 5a (n=86). Similarly, chai tea cup rims decrease to a very small percentage of the assemblage. Simple rims remain common, with straight wall simple rims dominant, 19% in phase 5 (n=11) and 27% in phase 5a (n=12), fine curved simple rims less important 11% in phase 5a (n=5), but only 5% in phase 5 (n=3), the third most common rim type in both phases belonged to another medium rim shape, an open flat rim (n=3, 5% in phase 5, n=5, 11% in phase 5a). Several examples of a medium ware jar with a square, everted rim (5, 9% see Fig. 2: 14) appeared in phase 5, while medium everted rim jars were common in phase 5a. The base sherds also saw a shift, with a much greater proportion of rounded bases (14, 74% of bases), and a shift from steep pointed bases to wide pointed bases (3, 26%) in phase 5, and a greater percentage of flat bases in 5a (n-3, 16%). Mini-flat bases disappear almost completely.

Phases 6-7: Leilan IIb

Phase 6: 113 diagnostics, 104 non-diagnostics
Phase 7: 217 diagnostics, 76 non-diagnostics

This phase, which witnesses the rebuilding of the main northern city wall (W-B) as well as the construction of a new southern wall (W-D) shows a clear shift in common ceramic forms from the preceding period. Although the architecture of this space might suggest that it is analogous to phase 4 in terms of function, the ceramics are nothing like those of the preceding phase. Although fine ware sherds are still in the majority, they only make up 52% of the entire population (113 sherds). A new decorative element is introduced, an appliqué braid motif (2, 1% of assemblage). Similarly, with the exception of straight simple rims (22, 29%) which dominate the assemblage, all of the common rim forms are new. The second-most common rim form is a medium ware, closed, slightly inverted jar rim (n=9, 11.5%). The third most common form are open and slightly cocked medium ware rims (6, 8%, Fig. 4: 26). The fourth most common type the “sila ware rim” is made of a distinctive over-fired, green ware, termed “clinky ware” at Tell Leilan (6, 8%). These same types are also common in phase 7, with the addition of open, exterior beaded medium ware rims (n=10, 7%, see Fig. 4: 25). In general, the proportion of pots and jars, as opposed to cups and bowls increase greatly (jars now 38%, #=30) as opposed to 26% (11) in phase 5a. Base types also shift, the most common base type is now a string-cut flat base (n=9, 29% in phase 6, n=24, 39% in phase 7), with plain flat bases in second, (n=8, 26% in phase 6, n=15, 25% in phase 7, see Fig. 4: 27) with green clinky ware “sila bowl bases” in third (n=7, 23% in phase 6, n=8, 13% in phase 7, see Fig. 4: 28). Round bases still occur (n=3, 10% in phase 6, and n=7, 11% in phase 7), but they make up much less of the assemblage. Phase 7 also witnessed a high percentage of decorated forms, with 15 (6%) of the sherds decorated with horizontal and wavy combing.

Phase 8: Habur Hiatus

Phase 8: 29 diagnostics, no non-diagnostics

Very few sherds were recovered from Phase 8, represented in this excavation by a meter thick deposit of infilled dust. This is due both to the nature of the operation-this deposit was only exposed to limited excavation, and was mostly only visible in section, as well as the nature of the deposit-which was extremely clean. Sila bowl rims predominated, (9, 53%), followed by closed and slightly cocked medium ware rims (3, 18%).

Phase 9: Leilan I

Phase 9: 35 diagnostics, no non-diagnostics

Once again, very few sherds (35) were recovered from this phase-due mostly to the very limited excavation undertaken. A few Habur ware sherds (n=3, see Fig. 5: 32) may provide a link to the second millennium. On the whole, however, these sherds were not retrieved from an obvious living surface. Common rim types were open ledge rims (n=5, 23%, see Fig. 5: 34) and closed ledge rims (4, 18%, see Fig. 5: 32).

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Periodization

We were able to recognize, both from internal characteristics, as well as from comparing this assemblage to others excavated at Leilan, four transitions in this pottery. The assemblage of the first two phases, points to a date at the tail-end of the Ninevite 5 sequence, due to the presence of sherds incised with late Ninevite 5 motifs (21% of phase 1, 5% of phase 2). By phase 3, the percentage of Ninevite 5 sherds declines to 1.4%. Other than the dwindling of Ninevite 5 motifs, the pottery shows few changes between phases 2 and phases 3. This emphasizes the continuity between Ninevite 5 pottery and the late third millennium pottery (period II) that follows it. In addition, the forms for both periods fit nicely into a site-wide as well as a regional wide chronology (see pottery catalogues for a list of comparanda). The stratigraphic sequence illustrates the same thing, namely that “period IIa” which we have termed a state-consolidation phase at Tell Leilan was a natural outgrowth of the state development processes underway during period IIId. The second shift noted from internal characteristics, between phase 4 and phase 5, does not seem to be a general chronological marker from Tell Leilan, but rather a shift in the use of space at the city wall, and in the pottery that accompanied this alteration. In general, this alteration is due to the increasing presence of larger (medium ware) vessels at the city gate-vessels of this type being previously very uncommon. As a result, the emergence of new dominant medium ware forms does not serve as a chronological marker. The second shift observed from internal characteristics occurred between phases 5a and phases 6. In phase 6, the presence of new medium ware jar rims, as well as the emergence of a new ware type (green “clinky” ware), and new dominant bases (flat as opposed to round, mini-flat, or pointed), does signal a shift to period IIb. In fact, the percentages of the dominant forms at the city gate correlates nicely with similar analyses done both at the Acropolis Northwest (Ristvet 1999) and at the Leilan Lower Town (Senior 1998). Unlike the transition between Leilan IIId and Leilan IIa, where a continuous use of space was emphasized, there seems to have been a short break in the use of the city gate between phase 5 and phase 6-corresponding to phase 5a which we have characterized as a decline in the use of this area (after the fire installation went out of use and trash was allowed to pile up in this area). We have no evidence of a shift in ceramics between phases 7 and 8, although such a shift has been noted in other excavated assemblages in Northern Mesopotamia (between Taya 5 and Taya 6, Reade 1982) for example, as well as between Akkadian and Ur III at Brak, (Oates et al. 2001). Instead, the final shift in periodization was recognized between phases 6/7 and phase 9, mostly by the introduction of a new rim type (ledge rims), a new ware type (poorly made, heavily straw tempered red and buff wares), and a new decoration style (painted black and red bands, Habur ware).

Function

The preceding analysis of the pottery of this operation by phase suggests a number of points about the changing function of this area through time, as well as the shifting nature of depositional practices. The ceramics of phases 1-4 are dramatically similar, comprised overwhelmingly of fine-ware drinking vessels. In the beginning of this sequence (phases 1 and 2), several of these vessels are decorated with Ninevite 5 excising. By the end, this decoration has fallen off, but the chosen form used, as well as the ware remains identical. This assemblage is extremely restricted, and includes very low percentages of cooking sherds, coarse ware (storage), or medium ware (common domestic). The ceramic assemblage emphasizes the non-domestic nature of this occupation. It seems likely that many of these sherds may have related to primary deposit, both because the extremely small size of many of them (on average these sherds were only 2X2cm) meant that they could have been swept into corners of the room and missed, or simply become embedded into the floor, and second because many of the larger sherds could be at least be partially reconstructed, suggesting they belonged to vessels which were simply left on the floors during each rebuilding. In phase 5, the changing ceramic assemblage emphasizes a change in the use of this space. During this phase, fine ware sherds decline to 59% of the assemblage, down from a high of more than 90% in the previous phases. At the same time, the rim forms that had dominated the preceding phases were no longer so much in evidence. Instead of a preponderance of cups, the pottery from this phase was generally more diverse, and less easy to characterize generally. On the whole sherds were larger, and there were no joins. This, together with the archaeological evidence, which shows that a large, non-domestic fireplace now dominated this area, along with a baked brick platform (or possibly road), may suggest that the ceramics no longer relate to direct use of this space. The fact that the construction of walls B and D, which made this area look, at least superficially, much the way it did during phase 4, did not lead to a return in preponderant fine ware, cup and bowl types, may imply that the function of this space during period IIb was slightly different than it had been earlier, despite the similarity in the architecture. The discovery of several concatenations of ceramics in phase 7, along with the discovery of an almost complete bowl (plate x), suggests that the pottery recovered from this phase may also relate to occupation surfaces. Unfortunately, bulldozer damage and mixing of deposits prevented us from carefully excavating this phase. The same thing applies to the final phase, 9, where a few sherds were found, possibly mixed in with the final earthwork structures. Outside of our central section, to the west of the limits of our excavation, a few chance sherds also may imply that bulldozer damage destroyed the floors and structures associated with the city gate during this phase.

Catalogue of Drawings