Çatalhöyük 1994 Archive Report

ÇATALHÖYÜK 1994 ARCHIVE REPORT

POTTERY REPORT
Part I: Report on Neolithic pottery from Çatalhöyük excavations, 1961 - 1965

Jonathan Last

The published reports on James Mellaart's excavations at the Neolithic site of Çatalhöyük provide little more than a general discussion of the ceramic material found (Mellaart 1962; 1963; 1964; 1966). The information from the reports has been summarised by Stucki (1984). It may be supposed that the main interests of the excavators lay elsewhere, and indeed Mellaart (1963, 101) states that "if one single category of finds at Çatal Hüyük might be described as relatively rare and unimportant, then it is pottery", although he did recognise the significance of the ceramic material from the lower levels of the site, dating to the later 7th millennium BC, which is still among the earliest known in the Near East.

In comparison to later Neolithic sites like Hacilar, pottery is not abundant at Çatalhöyük. A thin, dark, mineral-tempered pottery appears in small quantities in level VIII and increases through the upper levels. Forms are simple with bowls and holemouth jars predominating. The pottery is generally not fired to a high temperature and often has grey cores but never any organic temper. While many vessels are finely burnished other forms of decoration are rare: virtually no painted sherds were found and only a handful with incised decoration below the rim (Fig. 1:1 [note - illustrations are of pottery found at Çatalhöyük in 1994]). The prevailing surface colour is brownish-black, although red shades are also found. According to Mellaart, buff pottery is common until level VI but above this level the size of vessels increases and there are more light-coloured wares, with cream, orange and pinkish pottery increasing in quantity up to level I. The prevailing lug form also changes, from perforated knobs to ledge handles. In level II more developed forms like disc bases and tubular lugs appear.

Deeper soundings subsequently revealed earlier pottery, continuing down as far as level XII, the lowest reached. However, ceramics are much scarcer below level VIa than in the upper levels. The pottery is also more primitive in character, less well fired with thick black cores and often including vegetable temper. The earliest pottery is a heavy buff, cream or fight grey ware, already burnished and frequently having mottled surfaces. Shapes are not dissimilar to the later forms, but deep bowls are more common than holemouth jars.

Further information, particularly of a quantitative nature, was only available by returning to the excavated material. We considered this to be of value for two main purposes: firstly in order to tie in to Mellaart's chronology the material deriving from our investigation of surface deposits in different areas of the mound; and secondly to see whether there are any consistent differences between assemblages from Mellaart's four categories of structure - shrines, houses, storerooms and courtyards. The sherd material is stored in Konya Museum; complete and restorable vessels were taken to Ankara, although many were subsequently returned to Konya. These proved to be of less use because of 27 Neolithic vessels only 17 were recorded by level, and just two of those were assigned to a particular structure (both in level III). The initial problem with the extant sherd material, on the other hand, was to work out how representative it is of the total excavated assemblage. Mellaart reports that some 300 sherds were found in the 1965 sounding from levels VIb to XII. The sherd bags yielded 312 items from these levels, including a few not from that sounding. Therefore it appears that all sherds from level VIb and below are present in Konya. The upper levels are more problematic: IV to VIa are represented by numerous body sherds as well as diagnostics so may be more or less entire assemblages, minus the complete or restorable vessels; for levels II and III there are no bodies and far fewer sherds overall, so these clearly represent a selection from the excavated assemblage. No finds at all from levels 0 and I were deposited in Konya Museum.

Despite this variation in the quantity of finds many general trends in the ceramic material can be traced through levels XII to II. These may be helpful in terms of the first aim outlined above. Attributes of the rim sherds show varying degrees of change over time (Table 1). The first two columns chart mean sherd thickness, both of the vessel wall and the rim itself. The former shows that walls were thickest at the start of the sequence in level XII and decreased slightly in size through levels XI and X and again, more sharply, in VIII and VII. This no doubt relates to improving ceramic manufacturing and firing technologies. From level VIb through to III which accounts for the majority of the sherds - wall thickness is constant, although it increases again in level II, when there is also an increase in light-coloured wares which probably again relates to changes in firing techniques. Rim thickness follows a similar pattern, with a general decrease through the early levels and stability from level VII, before another decrease occurs in level III. This indicates a change of fashion in level II, with the rim much narrower in comparison to the body thickness, which fits alongside other typological changes at this time. The proportion of flat or squared-off rims reflects this too, with a lower figure in levels III and II compared to that for VIa to IV. Hence narrower, more rounded rims are found at the end of the sequence. The highest proportion of flat rims occurs between levels X and VIb, indicating a clear morphological difference between the earlier and later levels. However, levels XI and XII have lower values again, so changing rim shape is not merely a reflection of decreasing wall thickness.

The size of vessels is less variable, despite Mellaart's statement that vessel size increases from level V. In fact there seems to be a greater range of sizes in the upper levels, including miniature vessels. Many of the diameter measurements are estimates from rather small sherds and should be treated with caution, especially as the repertoire includes oval forms and the complete pots show that mouths are often rather irregular in shape. However, it does appear that, if anything, mean rim diameter decreases between levels XII-IX and VI-III, with levels VII and VIII more variable and a further decline in level II fitting with the other changes at the end of the sequence. The preservation of vessels shows less variation, with the mean vessel equivalent (VE) per sherd failing below 10% only in levels XII and VIb and reaching above 12% only in level VIa. More significant in terms of chronological sequence are changes in vessel form. I recorded only a simple distinction, based on rim angle, between closed forms (mouth diameter less than maximum vessel diameter) and open forms (maximum vessel diameter at mouth). In general terms this distinguishes plain bowls from holemouth jars, although other vessel shapes are also found in smaller numbers. However, there is no necessary correlation between mouth form and vessel shape. Holemouth bowls and open-mouthed jars are also known, particularly from the lower levels. Hence the shorthand used here, distinguishing "bowls" (open) and "holemouths" (closed), should not be taken as an indication that vessel forms fall into just two categories; rather it reflects an attempt to use the information from rim sherds to distinguish broad vessel families which might reflect different sets of practices perhaps relating to storage on the one hand and cooking or serving on the other. In fact rim angles of holemouth forms vary from nearly vertical to about 45 degrees; some have straight profiles, others have a slight neck (Fig. 1:1-4). The open vessels vary from simple hemibowls to more angular profiles in the later levels (Fig. 1:5-6), which prefigure the Early Chalcolithic carinated shapes found at Çatalhöyük West (Mellaart 1965).

Just over 50% of vessels are open forms in levels XII and XI, while more bowls are found in levels X and IX. These early levels have small samples and the figures may relate in part to the focus on a single room in levels XI and XII (see below). Subsequently there is a gradual increase in the number of holemouths from level VIII, peaking in level V at around 70%, before the proportion of open forms increases again and holemouth forms are rare by level II. The open : closed ratio may therefore be a useful chronological indicator for the surface assemblages.

The base sherds also show metrical and morphological variation (Table 2). The proportion of bases to rims is quite variable but generally high until level VIII, low in levels VII and VI, and moderate thereafter. This may reflect the preferential selection of diagnostic rim sherds in the later assemblages, although this was probably not so for levels VII and VIb. It may therefore represent a real difference in the relationship of base diameter and rim diameter, and hence relate to changes in vessel shape from squarer forms in the early levels to smaller, more rounded bases later. The mean angle between wall and base certainly appears steeper in levels XII to IX than thereafter, and the widest bases are found in levels X to VII. However, it is also clear that better defined bases (with an angular rather than a rounded junction) increase proportionally in the later levels. In addition, more pronounced bases and footed vessels (Fig. 1:7-8) are not found before level VI and are only common from level III. Base thickness follows a similar pattern to rim thickness with the chunky early forms (XII to VIII) giving way to thinner vessels later, via a transitional phase perhaps covering levels VII to VIb (although this is a very small sample).

Lugs are more restricted chronologically but still show some variation (Table 3). They are found only very rarely before level VI and no definite lugs are known below level IX. The proportion of lugs to rims is low until VIb, peaks in VIa and V and declines again from level IV. This reflects the changing proportion of holemouth forms on which lugs are mainly but not exclusively found. The complete absence of lugs in level II probably reflects the selection policy, since at least one of the restored vessels from this level is lugged. From levels VI to Ill the proportion of perforated lugs is consistently between 80% and 90%, so the change to ledge handles and unperforated forms (Fig. 2:4) reported by Mellaart must occur in level II. The size of lugs appears to increase somewhat between levels VIa and IV but there is little consistent variation in other attributes like the diameter of the perforation or the distance of the lug below the vessel rim. The prevailing form of lugs and handles also serves as a chronological indicator. The three main lug shapes may be summarised as pointed, flaring and straight (Fig. 2:1-3). The pointed form is found in all levels, but the flaring lugs are generally later than the straight ones, with an overlap in level V. Of the rarer forms, vertically placed lugs (Fig. 1:4) show no consistent pattern while basket handles and double perforations are not found after level V.

The body sherds may also furnish some quantitative information, particularly in terms of colour and wall thickness (Table 4). Colour groups were picked out visually and are based on external surface colour only, so they are to some extent subjective. Nevertheless the proportion of "dark" (dark grey or dark brownish grey, rarely a true black) sherds shows a clear change between levels VIII and VII when they increase greatly in number. This point also marks a change from pottery with at least some organic temper (93% of rims in level VIII) to a purely mineral tempered fabric (75% of rims in level VII; 96% in VIb). The lack of bodies from levels III and II prevent a quantitative analysis of later surface colours. Mean body thickness more or less mirrors the measurements from the rim sherds although the gradual decline in size from levels XII to IX is not apparent in the former, so seemingly indicative of changes in rim form within this early phase. Analysis of sherd size is more problematic since collection and selection policies have not been made explicit. This is of more use when comparing assemblages within a level; nevertheless there is an overall pattern with levels IX-VIa generally having smaller sherds than earlier or later phases. This may be related to changes in wall thickness and firing temperatures which would affect the susceptibility of vessel fragments to further breakage. The very thick walls before level IX and the better fired fabrics after level VI may have been more resistant to processes such as trampling.

The second part of the analysis considers differences between assemblages from different contexts within each level. The majority of sherds from each level, with the exception of VIb, are assigned to a particular room or court. In combination with the plans and descriptions published in the preliminary reports it is possible to locate many of these rooms and divide them into ‘shrines' and ‘houses'. Particularly noticeable is the fact that for most levels only a very few of the rooms have any sherds preserved at all. It is unclear whether this reflects a genuine absence of ceramic material from the majority of the site or merely the excavation techniques, but Mellaart mentions an average of about 6 pots per house from the lower levels and more later, which suggests that the ceramic assemblage discussed here is not fully representative above level VI. In addition the location of finds within the rooms is not recorded: are they from fill or floor contexts? This question is clearly relevant to how far we may interpret an assemblage as reflecting the original function of that room. In only one case (VII.44 and 45) are some sherds marked "fill"; but it appears unlikely that the majority of the finds, which are not so labelled, are all from floors. However, differences between assemblages from different contexts in many levels may reflect genuine differences in the use of these rooms.

From levels XII up to VIII excavation was restricted to a small area, investigated by the sounding dug in 1965. In level XII (Fig. 3) the majority of sherds came from room 29 but there are also some from 26. The latter has only open rims while 29 has closed forms as well; room 29 has rounded base junctions, 25 more angular forms. In addition 29 has more dark sherds than 25, all of which indicates that differences in room assemblages can be traced back to this earliest excavated level. In the succeeding level XI (Fig. 3) the western part of the area is occupied by an open courtyard but all the finds came from room 29. Deposition in the courtyards, of ceramics at least, begins in level X . While there was a single bowl rim from the earliest so-called "shrine", X.1 (see level VIII plan), most of the finds lay to the east. The area of rooms 29, 25 and 28 was ‘one large courtyard" (Mellaart 1966). Here the main distinction is between the west part of the court, contexts 25 and 28, which have mainly open forms with rounded bases and large body sherds (mean size 63.1 mm), and context 29 to the east which has more closed mouth forms, angular bases, and smaller sherds (55.5 mm). Overall the proportion of open vessel forms increases in this level: the relatively large number of holemouths in level XI may be an effect of that sample being restricted to room 29, since this area in level X also has mainly closed forms. Hence there are evident functional or depositional differences within the open area. , In level IX (Fig. 3) the shrines (1 and 8) lack holemouths entirely - these are confined to courtyard 25 and room 29A, reminiscent of the pattern seen further up in level IV. The "large storerooms" 31 had a large part of a single bowl, rather like the stores in levels V and VIb. Levels IX and X show no significant variation in the distribution of light and dark wares.

In level VIII (Fig. 3) finds came from a number of buildings in the area of the 1965 sounding, discussed in some detail by Mellaart. In addition a few body sherds came from shrine 45 on the other side of the site (see level VII plan). These were not noticeably different. In the southern area shrines VIII.1 and 8 had no finds while shrine 14, with textile and net patterns, yielded a single dark holemouth rim. Shrine 25, which had similar paintings, produced base and body sherds only. Shrine 27, also with net or textile motifs as well as a leopard relief, had only holemouth rims, including a number of fineware sherds without organic temper; the "red shrine" 31 also lacked open forms. The shrines as a whole have slightly smaller pots than the other rooms. Of these latter, the majority of the finds came from 28 (probably the room labelled on the plan as 24) and "meeting room" 29, each of which had mainly open-mouthed vessels and more rounded rim forms. In terms of the bodies the houses have thicker (8.8 to 7.7 mm) but slightly smaller (42.8 to 45.8 mm) sherds than the shrines, suggesting a more trampled assemblage. Dark colours are more common in shrines 14 and 27 while room 29 had a number of mottled sherds. Hence there is some indication of patterned differences between "domestic" and "shrine" assemblages.

From level VII. (Fig. 3) a much larger excavated area is available, but there are few sherds. Pottery came from shrines 44 and 45, houses 25, 43 and 71, and court 15. Room 45, the "boar's head shrine", had only two small body sherds from the building fill. Structure 44, the "leopard shrine", had the smashed remains of at least one dark burnished vessel, which agrees with Mellaart's report that two dark cooking pots were found. A majority of the vessels from 43 and 44 are bowls while the three rim sherds from 25 and 71 are probably all from holemouths. Hence the level VIII pattern is reversed. The courtyard yielded only a single body sherd, which seems to show that the use of the courts for refuse, seen in levels X and IX and to a greater degree later on in V and IV, was not practised in levels VII or indeed in VI, for which there were no sherds from courtyards. This may however reflect excavation or collection strategies.

In level VIb (Fig. 4) only 7 rim sherds are located to particular rooms. Storeroom E.32 had a single bowl rim, this is considered to be the storeroom of shrine E.70 which also had one bowl rim. Room E.49, on the other hand, adjacent to shrine E.50 with kilim paintings, had 60% closed forms, close to the level VIb mean, but rather larger vessels than from the level as a whole. Little else can be said about this phase since most finds had no context, although we may reasonably assume that they came from the 1965 sounding (Mellaart 1966). It is noticeable that rooms E.49 and E.50 to the north lack the red-brown wares common among the unlabelled finds.

In level VIa (Fig. 4) finds came from E.71-75, which are classed as houses without special decoration, and F.1, which may be a shrine like the building above it (F.1 may in fact refer to the context written up as E.VI.80 [Mellaart 1966]). All three vessels from the latter building are holemouths; of the others only the largest house, E.75 (4 rims), lacked open forms. One of the pots in F.1 comprises 6 sherds; about half of the rim is preserved. Large portions of individual vessels are also found in 71 and particularly 74, which is the storeroom of 75. These two contexts along with 75 also had larger body sherds on average than the other rooms. Only two bodies from F.1 were preserved but they are also large. This may indicate a difference between "primary" and "secondary" refuse assemblages in different contexts, although nothing is known of the exact provenance of these finds within the rooms. There is also some evidence that dark-coloured vessels were more common in 74-75 than the other rooms.

In level V (Fig. 5) there is a fourfold division of contexts between shrines (E.3, 4, F.1; 28 rims mainly in F.1), houses (E.7,75; F.3-5, 79 8; 36 rims), courts (E, FY F.29 5, 6; 64 rims) and one storeroom (E.3 store; 1 rim). The storeroom contained only a single bowl rim (like E.Vlb.32). Between the other categories there was little difference in vessel form and size, although the houses generally had more open forms than shrine F.1, a so-called "hunting shrine" with a large painting of a red bull, or the courtyards (this assumes that F.V.8, not found on a plan, is a room rather than a court). The few sherds from shrines E.3 and 4 are light-coloured bowls, while the Area E courtyard and room 75 have darker holemouths. Rarer features suggest possible locations of specialised vessel types, with two of the three lugs with double perforations found in shrine F.1, and two of the three basket handles in the nearby room E.7. The body sherds show some differences too, with more dark wares in F.S and 6 than in F.1-4. The thickest sherds (largest vessels?) come from the courtyards (other than F.6) and house F.7. In terms of preservation, the F.S courtyard has the largest sherds (61.7 mm average); shrine F.1, court F.6, and F.8 have relatively well-preserved sherds (54.1-54.8 mm); while the other courts and rooms, especially F.3 and 5 (43.2 and 45.7 mm), have smaller sherds. This may reflect specific patterns of activity or discard in this region but does not reveal general differences between Mellaart's structural categories other than suggesting that, as in the smaller sample from level IV, rooms tend to have fewer and smaller sherds than the courts.

In level IV (Fig. 5) a three way division could be made between a shrine (E.4; 10 rims), other rooms (E.5, 11-12, F.6; 30 rims) and the Area F courtyard (60 rims). The shrine had most bowl forms (80%), along with room E.5 (70%) and F.6 (75%). The latter context is presumed to be a room although it was not located on a plan (however, the underlying F.V.6 is a courtyard). In level IV the courtyard (69%) and rooms E.1 1-1 2 (67%) had most holemouths. Reflecting this, nearly all the lugs came from the court and while the shrine and room E.5 had mainly light coloured and red vessels, E.1 1-1 2 and the courtyard were dominated by dark wares. Rim diameters were similar, but preservation was better in the rooms (mean VE of 13%) than the shrine (8%). Footed bases were absent from the court and from E.1 1-1 2 but more common in the shrine and room E.5. Hence the shrine and an adjacent room have significantly different assemblages from the courtyard and the more peripheral rooms. The Area F finds also include body sherds: the court has a higher proportion of dark coloured bodies than F.6, and the sherds from the court are generally larger (49.8 mm compared to 41.7), reflecting rather different formation processes. Given the relative quantities of material, it probably indicates that dumping of fresh refuse was again concentrated in the courtyard.

From level IV, but particularly in level III (Fig. 5), the excavated area is again more restricted. Here the finds derived from shrine 1, which had wail-paintings of a deer hunt, and the adjacent rooms 2, 7 and 13, as well as from rooms 4 and 12. The reconstruction drawing indicates that room 4 contained a "bull pillar" (Mellaart 1963: 47). The shrine again yielded a very small assemblage (4 rims) which did not deviate from the overall level Ill means in vessel form and size. The highest proportion of holemouths (57%) was in room 13, which from its shape is probably a storeroom adjacent to shrine 1. Both rims in room 2 were also from closed forms. Four of the five preserved lugs from this level came from room 13 and the other lug (an unperforated example) was in room 2. In contrast to 2 and 13, room 7 on the other side of the shrine only had bowl forms, so again a possible functional division of space is reflected in the sherd assemblage. Sherd colours support this difference with 2 and 13 largely having dark wares, 7 mottled colours and shrine 1, as with E.4 in the level below, having light-coloured wares.

Finally in level II (Fig. 5), where wall-paintings were not apparent and the ceramic assemblage is rather different, finds came from shrine A.1 as well as houses B.1 and B.2 and their associated storerooms. The majority of finds are from B.2. Some differences are apparent although there are very few sherds from the shrine (4 rims) and storerooms (9 rims). Holemouth forms are most common in the storerooms (44%) and least in the shrine (none present). All rooms had a variety of surface colours although the shrine again lacked dark wares. Vessel thickness and diameter are lowest in the shrine (due to the presence of a miniature vessel) and largest in the houses, with the storerooms intermediate. Preservation is best in the shrine (mean VE of 18%) and lowest in the houses (12%), which presumably relates to the differences in vessel size. Hence the latest shrine assemblage seems to consist of a few small, open, light-coloured vessels.

While there are sufficient differences between rooms to suggest that variability in room use and formation processes can be discerned in the excavated material, there is no clear distinction between "domestic" and "shrine" assemblages that is valid for all phases. The shrines have more open vessel forms in levels IX, IV and II and more closed forms in levels VIa and VIII. In levels V, VIa and VIII shrines contain relatively large sherds, in level IV they are apparently rather smaller than average. Light-coloured wares are common in shrines from level V. Non-shrine rooms are more variable, although in the upper levels they seem to have smaller sherds than the courtyards. Storerooms generally furnished few finds: in levels II and III these were predominantly holemouths, in the earlier levels open forms. Continuity in a single domestic structure is seen in the earlier phases: from level XII to VIII the focus for ceramic deposition within the excavated area is clearly room 29 and it is also marked by the consistent occurrence of potdiscs in this room in levels XII, XI, X (when part of a court) and VIII. The only other potdisc from these levels was in courtyard X.28.

From the preserved assemblage, it appears that courtyards were used for discard only in levels X, IX, V and IV. They provide a variety of material although the predominance of holemouths in the level IX and IV courts is notable. However, variation between the courtyard areas in level V, and between different areas of the level X court point to significant small-scale patterning of depositional practice. It is these patterns relating to the everyday practices of ceramic use and discard which can be further elucidated only by more detailed contextual recording. It may then be possible to develop a deeper understanding of the degree of routinisation of activities within the context of a domestic space that was clearly structured by ritual practice.