This paper and the research behind it would have not been possible without the unwavering support and kind guidance of my supervisors, Susan Pollock and Reinhard Bernbeck—thanks for all your challenging questions and insightful comments. I am grateful to the workshop organizers for inviting me to present my research, and in particular to Johnny Samuele Baldi who has enthusiastically shared his knowledge of Chalcolithic ceramic technology with me over the last years. I wish to thank the RMO in Leiden for granting me access to the Balikh pottery collections and Prof. Peter Akkermans and Prof. Annelou van Gijn for making my study of the ceramics at Leiden University possible. I am most thankful to the FU Berlin, the DAAD and the BerGSAS for the crucial financial support for this research. Thanks to Matilda Siebrecht for proofreading this article. I am grateful to the reviewers for their challenging comments and questions—thanks for helping me in making this paper and my research better. Finally, my heartfelt thank goes to Olivier Nieuwenhuyse, who was the first to believe in this regional study of ceramic technology since he knew “a collection which would be perfect for it”.
1The Ubaid and the Late Chalcolithic (LC) 1 and 2 periods in northern Mesopotamia (ca. 5300-3900 BCE) have received increasing scholarly attention in the last thirty years (Henrickson and Thuesen 1989; Carter and Philip 2010a; Marro 2012; Iamoni 2016a). A growing body of literature has recognised the importance of the social and economic transformations of these periods, traditionally seen as setting the stage for the complex societies and early states of the mid-4th millennium BCE. These studies have highlighted the need for investigating new elements in material culture, especially in pottery, which we rely upon heavily for studying prehistory. But pots are just things we use as a proxy to get to the people who made and used them. Ceramic morpho-stylistic typologies aim at classifying and chronologically ordering these artefacts, enabling cross-dating over a broad area. However, they cannot explain the underlying social processes. Toward this aim, we should focus instead on practices of ceramic manufacture and use, to trace changes in these elements and their relation to social transformations (Pollock 2010; Baldi 2012a, 2012b, 2015, 2016a).
2In this paper, I examine pottery-making practices of the Northern Ubaid and the LC1 and 2 periods and present some thoughts on their implications in terms of changes in society. I will use the Balikh valley, modern northern Syria as a case study. The 1981-1984 University of Amsterdam’s excavations at Tell Hammam et-Turkman have provided one of the first ceramic and chronological sequences for this period in northern Mesopotamia (Akkermans 1988). Recently, our knowledge has been enriched by new excavations at Tell Zeidan (2008-2010) by the Oriental Institute of Chicago and the Directorate General of Antiquities and Museums (Stein 2009, 2010b, 2011; Fisher 2017). Additionally, surveys of the floodplain were carried out in 1983 by the University of Amsterdam (Akkermans 1993) and between 1992 and 1995 by the Oriental Institute of Chicago (Wilkinson 1998). Based upon these data, Maria Trentin (2010) reconstructed the regional settlement pattern of this period. The ceramic assemblage from the Balikh survey, at present unpublished, is currently being studied in detail by the present author, together with the published ceramics from the excavations at Tell Hammam et-Turkman. In this paper, I focus explicitly on ceramic technology; using a chaîne opératoire approach, I examine how ceramics were made at different sites along the Balikh river in order to trace the development of the ceramic traditions of the region and shed light on the underlying transformations in society. I start with an overview of the recent research on the Ubaid and LC1 and 2 and continue with an introduction to the archaeological context of the Balikh valley. I subsequently introduce the theory and methods underpinning this research. In the main part of the paper, I present the fabrics and techniques identified in the assemblage. I conclude with some preliminary observations on these technological elements.
3Our knowledge of the Ubaid and LC1 and 2 periods in northern Mesopotamia is based upon an increasing number of excavated sites, though difficulties in identifying the presence and extent of usually small, prehistoric occupations remain (Pollock 1999: 45-46; Wilkinson 2003: 37; Akkermans 2013; Bernbeck 2013). Our chronologies for prehistory are built upon type-site terminology (“Halaf”, “Ubaid”), resulting in the construction of chrono-cultural entities perceived as internally homogeneous over broad geographical areas, which is clearly a reflection of a normative conception of culture (Campbell 2007). By focusing on similarities rather than on differences, the complexity and diversity of local ways of life moves to the background, turning “ancient social groups into internally unified and externally sharply bounded entities” (Bernbeck and Nieuwenhuyse 2013: 21). As a result, archaeologists have applied the label of “transitional” to those phases in between, the lengths of which seem hard to define due to gaps in excavated sequences and paucity of radiocarbon data—one can think of the Halaf-Ubaid “Transition” or the Late Ubaid/LC1 “transition”. Thus, transitions often represent gaps in our knowledge; we see gradual transformations in the archaeological evidence, the nature of which remains unclear (Frangipane 2012).
4In the last century, mirroring the changing theoretical perspectives in our discipline, the term “Ubaid” has been used to describe a pottery style (the so-called “black-on-buff” pottery, first excavated at Tell al-Ubaid in southern Mesopotamia; Hall and Woolley 1927), a period (Oates 1960), and a culture (Carter and Philip 2010b). Based on the archaeological evidence (ceramics, clay objects, architecture) retrieved in Greater Mesopotamia, it appears that different regions participated in the Ubaid material-cultural phenomenon whilst maintaining a local character (Henrickson and Thuesen 1989; Akkermans and Schwartz 2003: 154-180; Carter and Philip 2010a). Scholars have therefore introduced new terms and concepts for the Ubaid, such as interaction sphere, horizon, or oikoumene (Nissen 2001; Stein and Özbal 2007; Stein 2010a). This includes those sites, phases of occupation, or assemblages historically classified as Ubaid based upon the presence of black-on-buff pottery, which can be found from the northern Levant and southeastern Anatolia, to Mesopotamia, the Persian Gulf, and southwestern Iran (Carter and Philip 2010b). According to Gil Stein (2010a), the Ubaid interaction sphere emerged from the spread of an ideological system whose material aspects (tripartite house plan, pottery styles, labrets, and communal cemeteries) were adopted in light of local cultural systems and practices, leading to the formation of hybrid identities. Within this sphere, there is neither a centre nor a hierarchical structure. Ideas, thoughts, and representations moved along many axes and directions as communities engaged in multiple interactions—social or economic, through direct or indirect contact, between neighbouring sites, through long-distance trade, etc. (Karsgaard 2010).
5From the Terminal Ubaid (ca. 4800-4600 BCE) onwards, the production and use of black-on-buff pottery decreased, with the emergence of unpainted, organic-tempered, chaff-faced ceramics, often with scraped surfaces and dark cores (Akkermans 1988; Baldi 2016b). This has been interpreted as a change in the significance of ceramics, as new pottery-making practices, modes of production, ways of consuming food, and forms of socio-political organization were introduced (Marro 2012). These general trends are documented across Greater Mesopotamia, yet emerged gradually and differently at various sites against the background of a long, local process of transformations. The shape and (limited) decoration repertoire preserved a local character, despite technological similarities across northern Mesopotamia. Thus, researchers introduced the terms “Chaff-faced” or “Standardized ware” horizon to refer to the ceramic developments following the Ubaid horizon (R. J. Braidwood and L. S. Braidwood 1960; Marro 2010). In northern Mesopotamia, this process of regionalization had its roots in the Ubaid interaction sphere (Frangipane 2012; Baldi 2016b). As Ubaid material culture gradually disappeared, the ceramic provinces of the LC1 (ca. 4600-4200 BCE) became archaeologically visible, each with its local assemblage. In the LC2 (ca. 4200-3900 BCE), with the spread of new elements such as Coba bowls and, in general, unpainted pottery, these provinces merged into two ceramic macroprovinces, with the Balikh valley as a sort of border between them (Baldi 2016b). Clearly, this should not be seen as a sharp boundary, as people and ideas were still circulating. When moving from the material evidence to the underlying socio-economic processes, the Ubaid and the LC1 and 2 appear to be closely linked. The Ubaid system with predominantly small sites/villages (ca. 1-2 ha) and only a few larger sites but few instances of settlement hierarchy (Iamoni 2016b) changed into a more structured territorial system, likely based on new political affiliations with emerging central authorities, thus setting the stage for the centralised societies of the LC3 (Frangipane 2012; Baldi 2016b).
6The Balikh valley is located between modern southern Turkey and northern Syria, the latter including most of the Balikh basin, and stretches for approximately 100 km in north-south direction. The region has an arid climate with low annual rainfall (ca. 200-600 mm; Akkermans 1993), thus falling within the zone of agricultural uncertainty (Wilkinson 2000). The Balikh is a small stream, 6 m wide on average and around 116 km long in Syrian territory. It has multiple springs, some of them located ca. 30 to 50 km to the north in Turkey.
7In comparison to other areas of south-west Asia, archaeological research in the Balikh valley started relatively recently. In 1938, Max Mallowan (1946) directed a brief regional survey and a five-day excavation at Tell Mefesh, in the middle of the valley at the Balikh-Wadi Qaramokh confluence, where he uncovered Ubaid layers. Afterwards, the valley was deserted by archaeologists for the next thirty years. It was only at the end of the 1960s that the construction of the Tabqa dam prompted the beginning of intense archaeological research along the Balikh (Akkermans 1993: 12-14). In 1981, Maurits van Loon (1988) started excavations at Tell Hammam et-Turkman focusing on the Bronze Age occupation. However, layers dating between the late 6th and the early 4th millennium BCE were encountered in a deep sounding and step trench (Meijer 1988). These were attributed to periods IV and V, dated to the northern Ubaid and the LC respectively. Based upon changes in the ceramics, and thus not always coinciding with main stratigraphic divisions, four phases A-D (12 strata) were distinguished within period IV, with only two phases A and B (seven strata) in period V (Meijer 1988: 69-78). The architecture of period IV (deep sounding) included mainly the remains of mudbrick houses. Period V yielded the ruins of tripartite buildings. As these were dug out in a step trench over four squares and whose steps were not joined together, the presence of unrecognized intermediary layers cannot be ruled out (Meijer 1988: 75). Nine samples were taken for radiocarbon dating (Akkermans 1988). A Bayesian analysis of the five dates from Hammam VB gave a combined range of 4220-3990 cal. BCE (Wright and Rupley 2001: 98-100). This dating was used to define the LC2 in the Santa Fe chronology.
- 1 Radiocarbon dates are available only for phases IVA-B and VB (Akkermans 1988: 129-130).
8The Tell Hammam et-Turkman sequence results from the seriation of the ceramics supported by absolute dates and comparisons with neighbouring regions. In his analysis, Peter Akkermans (1988) identified ceramic types defined as a shape with specific technological features such as paste (mineral or organic-tempered), surface finishing, and decoration, and then isolated clusters of types with a limited vertical time span. Yet, as for period IV, some types could be ascribed to a single-phase IVA, IVC or IVD, while others were encountered in both phases IVA-B or IVC-D, and no specific type could be attributed exclusively to phase IVB. Thus, the early phase IVA-B (ca. 5300-4800 BCE) and the later one IVC-D1 were defined. Other criteria for the internal subdivision of period IV were the ratio of painted to unpainted pottery, together with differences in designs and their execution. It is important to remember here that, despite the decrease in painted pottery from phase IVA to IVD, it always constitutes a minority within the assemblage of each phase, which is dominated by unpainted vessels instead. The Balikh regional periodization was built upon this sequence, with its periods IV and V respectively.
9Tell Zeidan lies on the east bank of the Balikh at its confluence with the Euphrates, ca. 5 km east of Raqqa. As is the case with other tells in the Balikh valley, Tell Zeidan is a multi-mounded site composed of three tells for an estimated extent of ca. 10 ha (Stein 2009). The Syrian-American expedition defined four main prehistoric occupation phases, i.e. Halaf, Ubaid, LC1 and 2, with mostly domestic architecture. After a gap encompassing almost the whole 4th millennium, the site was briefly reinhabited at the beginning of the Early Bronze Age before the definitive abandonment (Stein 2011). Absolute dates were obtained through the accelerator mass spectrometry radiocarbon dating of 33 samples (Stein 2009, 2011).
10Thanks to recent research, including the new data from Tell Zeidan, it is possible to date the Balikh IVA-B phase (ca. 5300-4800 BCE) to the Ubaid period, Balikh IVC-D (ca. 4800-4600 BCE) to the Terminal Ubaid, Balikh VA (ca. 4600-4200 BCE) to the LC1, and finally Balikh VB (ca. 4200-3900 BCE) to the LC2 (Baldi 2016b; Fisher 2017).
- 2 The Balikh Survey was conducted in Syrian territory, i.e. between the modern Syro-Turkish border to (...)
- 3 Each site was identified by means of the prefix BS (Balikh Survey) followed by an Arabic number.
11In the autumn of 1983 a survey was conducted in the Balikh flood plain2 within the Balikh Valley Prehistoric Project (Akkermans 1993), and a new survey took place within the Western Jazirah Prehistoric Project between 1992 and 1995 (Wilkinson 1998). Twenty-three sites of the northern Ubaid and LC were recorded (Trentin 2010),3 mostly tells close to water courses. Tells along the Balikh vary from mounds with a complex layout, usually multi-period sites, to small ones hard to recognize. This is due not only to the site history of occupation but also to post-depositional physical and cultural processes erasing or obscuring archaeological remains (Akkermans 1993: 138-145). Although sedimentation processes have affected portions of the valley differently, it can be expected that small prehistoric sites lie beneath the current surface and are thus unknown to us.
12The Balikh valley is characterised by a cyclical or sequential occupation of a site or an area of a site (Akkermans 1993; Trentin 2010), with the large mounds such as Tell-Sawwan (BS147) and Tell Zeidan (BS2) starting as separate mounds that merged over time. For phase IVA-B, 13 sites were recorded, including Tell Sawwan and Tell Zeidan with an estimated size of ca. 10 ha, and possibly Tell Hammam et-Turkman (BS175; fig. 1a). In phase IVC-D, small sites increased in number (24 in total; fig. 1b) especially in the north close to Tell Hammam et-Turkman. However, in the centre of the valley at Tell Sawwan, the occupation started to contract. Only 13 and 10 mounds date to phases VA and VB respectively (fig. 1c-d). Interestingly, the large sites of the previous phase undergo some changes. At Tell Zeidan, LC1 levels are not well represented architecturally, so that the excavators suggested that this site too started to contract or lose population in the mid-late LC1, i.e. Balikh VA and possibly already Balikh IVD (Fisher 2017: 268-269). At Tell Hammam et-Turkman, a monumental building was attributed to phase VB (Meijer 1988: 76-78). At Tell Sawwan, the occupation contracted to ca. 2 ha. At Tell Zeidan, Balikh VB diagnostic sherds were collected on all the three mounds, although in limited numbers in comparison to the previous phase IVC-D, which is well attested at the site in the form of hundreds of beaded-rim bowls. However, excavations showed a change in occupation in the LC2, with the southern mound possibly being abandoned (Fisher 2017). In general, the paucity of sherds of period V in comparison to those of early period IV suggests a gradual abandonment of the Balikh valley, possibly already starting at the end of phase IVC-D at the Ubaid-LC1 transition, at least in the middle portion of the valley, and culminating with the region being virtually deserted in the LC3.
Fig. 1 – The Balikh IV-V periods settlement patterns
After Trentin 2010: 332, fig. 20.1; 333, fig. 20.2; 335, fig. 20.3.
13As archaeologists, we usually look at pottery assemblages so as to identify what did or did not change in time and space, in order to formulate a plausible interpretation of how this change originated in a given context. In this regard, crucial insights are provided by the anthropology of techniques which conceives of technology as “a total human phenomenon” (Mauss 1936) situated within a specific socio-historical context (Sigaut 1994). Social and cultural factors affect technological choices of a community and are expressed not only in what, but also in how a community produces, uses, and discards (Dobres and Hoffman 1994). While creating and transforming materials, individuals are also creating and transforming themselves; to understand the social dimension of technical choices it is essential to examine in detail the main components of a technological activity, i.e. matter, objects and tools, gestures and movements, skills and knowledge (Lemonnier 1986).
14For such a goal, an approach that is both theoretical and methodological is needed. In archaeology, this has required an analysis of the making of ceramics in terms of chaînes opératoires, i.e. the sequence of operations through which a raw material is transformed into a finished product or tool (Cresswell 1976). To master such a knowledge of techniques, tools, and materials, the apprentice learns from a tutor who usually belongs to the apprentice’s social context (Lave and Wenger 1991; Bril 2002; Stark et al. 2008). Regarding this, practice theory offers a useful theoretical foundation to archaeological studies of technology. Sociologist Pierre Bourdieu (1977: 72-87) called habitus the system of dispositions, tendencies, and predispositions necessary to do something, assimilated by imitating other people’s actions. These learned dispositions have a bodily dimension, as they are expressed through characteristic gestures and postures; they are embodied practical knowledge (Ingold 2001). In the case of pottery-making, at the end of the learning process and through repetitive practice the learner has embodied the necessary motor and cognitive skills to make a pot according to the model given in his/her group. The learning process allows the cultural transmission of a chaîne opératoire—a way of making pottery resulting from the technical choices of a group, imbued with memories and cultural perceptions; in short, a technical tradition situated within a specific social, spatial, and chronological context (Dietler and Herbich 1998; Roux 2011). Within a ceramic tradition, the learning of forming techniques usually requires a longer time and necessitates a close interaction between apprentice and tutor because of the difficulty of mastering the necessary motor skills, which makes these techniques more resistant to change. Concerning this, Bourdieu (1977: 159-171) defined doxa as the natural, undisputed way of how things are, as perceived by individuals socialized within a specific context. This allows for the reproduction of social institutions, structures, and behaviours. Yet, this perceived naturalness might be questioned and challenged as new demands and requests emerge within a society; new choices and responses engender new practices and learning networks and ultimately lead to social transformations. This is a crucial aspect when dealing with continuity and change in technological practices and ceramic traditions.
- 4 Wilkinson recorded sites BS117, BS270, BS363, BS400, BS402, BS407, BS412 and BS415; Trentin include (...)
15The Balikh IV-V ceramics discussed here are stored at the Rijksmuseum van Oudheden (RMO) in Leiden, the Netherlands. The assemblage includes 2,390 sherds; 1,954 were collected at 15 sites during the 1983 Balikh Survey4 and 436 come from excavations at Tell Hammam et-Turkman, the latter mostly from phase IVA-B (table 1). This limitation of the dataset should be considered when interpreting the data.
Table 1 – Distribution of Balikh IV-V periods pottery sherds at each site.
Balikh Survey ID |
Site Name |
Diagnostic sherds |
Sherds for techn. analysis |
BS2 |
Tell Zeidan |
870 |
409 |
BS35 |
Merj Abu Sharib |
156 |
0 |
BS82 |
- |
25 |
0 |
BS83 |
Tell Semen |
1 |
0 |
BS110 |
Tell Damer |
30 |
0 |
BS138 |
Tell Mefesh |
308 |
0 |
BS139 |
Tell Khadriya |
4 |
0 |
BS147 |
Tell Sawwan |
449 |
124 |
BS167 |
- |
76 |
0 |
BS175 |
T. Hammam et-Turkman |
436 |
214 |
BS180 |
- |
1 |
0 |
BS183 |
- |
1 |
0 |
BS235 |
Tell Wazgol |
12 |
0 |
BS280 |
- |
13 |
0 |
BS282 |
- |
2 |
0 |
BS330 |
- |
6 |
0 |
|
|
2390 |
747 |
16The present study of the Balikh IV-V ceramic chaînes opératoires is based upon the conceptual and methodological framework proposed by Valentine Roux (2019), whose English terminology I apply. This approach proceeds through a hierarchical classification in three stages, i.e. technical, petrographic, and morpho-stylistic. The reason for considering technical features first is that these result from more specialised gestures, learned and embodied through long apprenticeship (Gosselain 2002: 79). They are therefore less likely to vary in time and space than morpho-stylistic attributes, which tend to follow consumer demands and can be more easily imitated by potters, thus spreading and changing faster than technical elements. Combined with the study of the fabrics, these two stages allow us to define techno-petrographic groups and to reconstruct the manufacturing process—hence the chaînes opératoires at a given site—so as to identify groups of producers. The third stage considers instead the morphological variability within each techno-petrographic group; by focusing on the range of vessels produced in a specific way, it is possible to unveil connections between producers and consumers so as to assess the functional or sociological variability of ceramic assemblages. However, “the analytical procedure must be pragmatic and flexible” (Roux 2019: 249) and consider quantity, fragmentation and conditions of preservation of the assemblage.
- 5 During the data collection, I used the ceramic typology of Tell Hammam et-Turkman (Akkermans 1988) (...)
17The quantity and conditions of preservation of the Balikh IV-V sherds vary strongly from site to site. Moreover, being mostly survey material, I decided to first date the ceramic fragments based upon their compositional and morpho-stylistic features5 and only later did I select samples for the technological analysis. Thus, I adapted Roux’s methodology by changing the order of the steps. At present, I have completed the classification of fabrics and technical features, but I am still working on the morpho-stylistic step. Therefore, in this paper I discuss in detail fabrics and fashioning methods and give only general information on the shapes. Admittedly, this creates an imbalance in the quantity and detail of information provided, still it allows us to follow the manufacturing sequence from raw material preparation to shaping and to put forward some considerations on the technological variability of the Balikh IV-V assemblage.
18To reconstruct clay preparation strategies, I conducted a macroscopic fabric analysis according to the procedure of Orton and Hughes (2013). Using an optical microscope (20×), I examined sherd surfaces and sections on a fresh break so as to define six fabrics groups and a total of 10 single fabrics according to the colour, feel, and hardness of the sherds as well as the type, frequency, sorting, size, and shape of inclusions and voids. I did this analysis on all 2,390 pottery fragments from all Balikh IV-V sites.
- 6 This implies that modelling or percussion techniques, usually used for fashioning the base and lowe (...)
- 7 I am grateful to the ceramic technology expert and potter Loe Jacobs, formerly at the Laboratory fo (...)
- 8 The petrographic analysis is underway.
19To reconstruct the fashioning methods of the vessels, i.e. the series of operations through which potters transform clay into the desired form, I examined sherd surfaces and sections, the latter at different magnifications. My aim was to identify macrotraces and deformations in the paste, respectively, which are indicative of specific forming and finishing techniques. In general, the larger the preserved portion of the original ceramic container, the more technological information one can gather. However, there are no complete vessels in the Balikh IV-V assemblage, while complete and almost complete profiles (always bowls) are present in limited numbers. Thus, I selected sherds larger than ca. 2 cm that could be attributed to a morphological type characteristic of a single phase; small rim, painted, or scraped body fragments and bases not attributable to a vessel type and phase had to be excluded.6 Based on these criteria, I selected 747 well-preserved sherds from three sites, i.e. 214 from Tell Hammam et-Turkman (BS175), 124 from Tell Sawwan (BS147) and 409 from Tell Zeidan (BS2). To identify combinations of traces diagnostic of a specific technique, I first examined each sherd with the naked eye and at low-angled light. For the paste and microtopography inspection, I used an optical microscope Stereo Blue euromex (0.75×-4.5×) and a stereomicroscope Leica M80 (0.75×-6.0×) provided with a camera Leica MC120 HD. The data were interpreted using references from experimental and ethnographic research (Gosselain 2002; Gelbert 2003; Baldi 2015; Roux 2017, 2019; van As and Jacobs, 2004).7 Based on these observations at different scales, I defined five different technical groups describing the manufacture sequence from forming to finishing. Finally, in order to connect technical and fabric groups, I sampled fabrics and different parts of the vessel from each of the five technical groups for thin-section petrography.8
20Six macrogroups for 10 single fabrics were identified. Each group is labelled by means of a capital letter A-F, with the single fabric named by the combination of the group letter and a number (fig. 2; table 2).
Fig. 2 – Fabrics identified in the Balikh IV-V assemblage. A. A1; B. A2; C. A3; D. B1; E. B2; F. C1; G. D1; H. E1; I. E2; J. F1.
Table 2 – Macroscopic classification of fabrics (sample: 2,390 sherds from all Balikh IV-V sites).
Macrogroup |
Predominant colour |
Predominant non-plastic inclusions |
Texture |
Fabric |
A |
Pale yellow to pale brown |
Limestone, quartz, iron, clay pellets |
Very fine |
A1 |
Fine |
A2 (with mica) |
Fine |
A3 |
B |
Pale yellow to pale brown |
Limestone, coarse organic, quartz, iron, clay pellets |
Coarse |
B1 |
Very coarse |
B2 |
C |
Pale yellow to pale brown |
Multicolour fine sand |
Fine |
C1 |
D |
Pale yellow to pale brown |
Multicolour fine sand, fine organic |
Coarse |
D1 |
E |
Brown to dark grey |
Fine organic, calcite |
Coarse |
E1 |
Coarse organic, limestone |
Coarse |
E2 |
F |
Red to brown |
Grey fine sand, coarse organic |
Coarse |
F1 |
21The first group A is characterised by very fine (<0.1 mm) to fine (0.1-0.2 mm) predominant limestone inclusions. Quartz, iron-rich inclusions and clay pellets were also observed. Fine organic components are present up to 10%, likely naturally occurring in the clay (Nieuwenhuyse 2006: 87). The inclusions distribution, low porosity, and homogeneity of the clay matrix indicate that potters carefully cleaned and mixed the paste. Sherds of group A are mostly completely or slightly incompletely oxidised and pale yellow to pale brown in colour, typical of carbonatic clays fired in oxidising atmosphere at 700-800°C (Albero Santacreu 2014: 95). This group includes three fabrics A1-A3 differing in their texture and frequency of inclusions (fig. 2a-c).
22Fabrics of group B have fine, moderate (10%-20%) to abundant (>20%) limestone and moderate coarse (>5 mm) organic temper, the latter leaving characteristic voids visible along the sherds’ wall and section. Other mineral components are quartz, iron, and clay pellets. Based on the sorting and heterogeneities of the clay matrix, it seems that potters did not prepare the paste as carefully as with group A, likely refraining from time-consuming cleaning and mixing operations. The addition of organic temper results into a higher porosity in comparison to group A. Typical sherd colours are pink to reddish yellow, very pale brown, and pale yellow. They can be completely oxidised (B1; fig. 2d) or with a dark core (B2; fig. 2e).
23Group C is characterised by multicolour fine sand, with rare limestone. Sherds are mostly completely oxidised, pale yellow to very pale brown, but also pink and reddish yellow in colour, evidence of firing in oxidising atmosphere. This group has a single fabric C1 (fig. 2f).
24This group is marked by multicolour fine sand and moderate (10-15%) fine organic temper, the latter mostly observed in the sherds’ section. Typical colours are pink to reddish yellow, very pale brown, and pale yellow, suggesting firing in an oxidising atmosphere. As with group C, group D has only one fabric D1 (fig. 2g). It is possible that potters exploited clay sources similar to those of fabric C1, though they opted for less time-consuming paste preparation resulting in heterogeneities in the clay matrix and voids due to incomplete clay mixing or joining of coils.
25Group E is defined by the presence of moderate (10-20%) organic temper and brown to dark grey surfaces and cores as a result of incomplete oxidation or reduction. Two fabrics were distinguished; E1 with crushed calcite and fine (<5 mm; fig. 2h) organic temper, and E2 with coarse organic temper (fig. 2i).
26This final group includes only a single fabric F1 (fig. 2j). This presents a variety of mineral inclusions (quartz, limestone, rock fragments of grey colour) and sparse (5-10%) coarse organic temper. Sherds range in colour from reddish yellow to brown and are mostly incompletely oxidized.
27The macrogroups above are characterised by mineral (A, C) and organic inclusions (B, D, E, F), the latter varying in the size and frequency of their organic components. In general, these groups differ in their chronology (table 3) and morphological repertoire. Very fine and fine mineral fabrics of period Balikh IVA-B (Ubaid), mostly found with black-on-buff pottery, strongly decrease from period Balikh IVC-D (transition with the LC1) onwards. This is accompanied by an increase in organic-tempered fabrics from phase IVC-D. Fabrics B1 and D1 are more frequent in phase IVC-D but decrease in period V, as fabric B2, characterised by a dark core, is found more often. Fabrics of macrogroups E-F already appear in phase IVC-D and are also attested in phases VA and VB, though always in very small numbers.
Table 3 – Distribution of macrogroups of fabrics at Tell Zeidan (BS2), Tell Sawwan (BS147) and Tell Hammam et-Turkman (BS175) in Balikh IV-V periods.
Macrogroup |
Phase |
BS2 |
BS147 |
BS175 |
A |
IVA-B |
110 |
235 |
266 |
IVC-D |
15 |
43 |
14 |
IVA-D |
9 |
22 |
0 |
VA |
9 |
0 |
3 |
VB |
0 |
0 |
0 |
VA-B |
3 |
3 |
0 |
B |
IVA-B |
10 |
25 |
15 |
IVC-D |
52 |
40 |
71 |
IVA-D |
5 |
23 |
0 |
VA |
15 |
18 |
42 |
VB |
14 |
3 |
13 |
VA-B |
22 |
9 |
0 |
C |
IVA-B |
157 |
16 |
3 |
IVC-D |
87 |
1 |
0 |
IVA-D |
12 |
3 |
0 |
VA |
3 |
0 |
0 |
VB |
4 |
1 |
0 |
VA-B |
4 |
2 |
0 |
D |
IVA-B |
19 |
0 |
1 |
IVC-D |
241 |
1 |
0 |
IVA-D |
14 |
0 |
0 |
VA |
37 |
1 |
0 |
VB |
5 |
0 |
0 |
VA-B |
9 |
0 |
0 |
E |
IVA-B |
0 |
0 |
0 |
IVC-D |
0 |
0 |
2 |
IVA-D |
0 |
0 |
0 |
VA |
0 |
0 |
0 |
VB |
0 |
1 |
5 |
VA-B |
6 |
0 |
0 |
F |
IVA-B |
0 |
0 |
0 |
IVC-D |
4 |
0 |
0 |
IVA-D |
0 |
0 |
0 |
VA |
4 |
0 |
1 |
VB |
0 |
1 |
0 |
VA-B |
2 |
0 |
0 |
28Further differences between these groups include the degree of refinement of the clay. Mineral groups A and C have a homogeneous clay matrix, suggesting that potters carefully prepared and mixed the paste. On the contrary, organic-tempered fabrics (B, D, E, F) show varying degrees of sorting and distribution of inclusions and voids, due to incomplete clay mixing or joining of coils. As such, it appears that from phase IVC-D, potters refrained from time-consuming clay preparation practices and possibly exploited new clay sources, in general experimenting more than in phase IVA-B. In period V, fabrics decrease in number and show frequent organic inclusions, perhaps as a tendency towards standardization.
- 9 The petrographic analysis will allow me to ascertain this.
29Turning now to ceramic production, it can be suggested that this took place at each site over periods IV and V, as hinted by the overfired sherds and wasters collected. Looking at fabric distribution, some local tendencies can be seen, especially in period IV. Buff-yellowish fabrics with limestone (groups A, B) are predominant at Tell Hammam et-Turkman and Tell Sawwan. In contrast, at Tell Zeidan, groups C and D with fine sand are the most frequent, and rarely appear north of this site. This might be explained in light of the local geology, with potters exploiting clay sources close to Tell Zeidan at the Balikh-Euphrates junction as the Euphrates transports sediments of volcanic origin downstream from Anatolia (Russo et al. 2018). Overall, as the inclusions observed through the macrofabrics analysis seem compatible with the local geology,9 it can be hypothesised that potters used sedimentary clays from the Balikh region and close to the production site. Taken together, these results confirm what was known of technological practices of this period in the region and neighbouring areas (Akkermans 1988; Marro 2012).
30Within the Balikh IV-V assemblage, two main roughout techniques were identified: coiling and modelling a clay lump. Based on differences in the preforming and finishing operations, I defined five technical groups (TG). These are described below together with their diagnostic elements.
Fig. 3 – Diagnostic elements of TG1.
A. Irregular, wavy profile with oblique fissures and orientation of inclusions and voids (marked with a dashed line); B. Outer surface with elongated vertical marks left by a tool (marked with arrows) and thin striations from smoothing, partly covering those left from scraping. Oblique concentric fissures indicating joining of coils (dashed line) can be seen on the lower body; C. Inner surface with vertical depressions left by the hand supporting the wall during preforming.
31Potters made the rough-out by pinching and spreading coils (1.5-2.5 cm) that were obliquely joined (fig. 3a). They used a support such as their hand but possibly also broken containers and pots. Different elements point to these techniques, such as the vessel’s irregular, sometimes wavy profile, the horizontal rows of rounded depressions and bumps predominantly on the interior where coils were joined together and flattened by discontinuous pressure (Gelbert 2003: 77-78), the small crevices, thin elongated fissures and overthicknesses with a regular concentric arrangement, and the fractures with a bevelled profile. At high magnification, the paste appears as strongly deformed with inclusions, voids, and fissures with an internally oblique or oblique alternated orientation (S/Z configuration; Roux 2019: 160-165).
32Potters then shaped the rim by continuously pressing the wet clay with a wet soft tool (fingers or piece of cloth). This can be inferred from the fluid microtopography of the rim, with horizontal parallel bands of ribbed striations, visible both with the naked eye and at high magnification (Roux 2019: 196-197).
33Afterwards, potters scraped the exterior of the preform. As the clay would dry out, this operation would turn into smoothing, partly obliterating the scraping traces (fig. 3b). On the exterior, thin elongated vertical marks and horizontal bands of thin ribbed striations were observed along the upper body (Roux 2019: 64-66). In the lower body, at high magnification, the microtopography appears to be fluid, with subparallel bands of ribbed striations in a horizontal to oblique direction and from the bottom upwards. At the upper-lower body join, these striations overlap the more regular surface of the upper body. Taken together, these elements point to the use of a wet tool on wet clay (Roux 2017: P0/2), moved in a horizontal to oblique direction to give the vessel’s bottom its final shape.
34As potters performed these operations, they would hold the interior of the preform. On this surface, vertical elongated depressions are visible, left by the hand supporting the inner wall while the outer one was being scraped (fig. 3c; Gelbert 2003: 75-78; Roux 2019: 174-175).
35The manufacture sequence could end here or include optional steps, usually painting the exterior.
Fig. 4 – Diagnostic elements of TG2.
A. Outer surface with rounded depressions (marked with dashed circles) and bumps. An incompletely erased coil join along the rim (dashed line). Along the body there are oblique, intersecting bands of deep striations left by scraping the wet clay with a lithic blade moved from the bottom upwards; B. Inner surface with a combination of compact and fluid microtopography with subparallel bands of thin ribbed striations in a horizontal to oblique direction left by a wet wool on a not-so-wet clay.
36Similarly to TG1, potters started by spreading and obliquely joining coils (1.5 to 2.5 cm) and then made the rim by continuously pressing the wet clay with a wet soft tool. Afterwards, they scraped the exterior with a hard tool, likely a lithic blade, in a horizontal to oblique direction from the bottom upwards (fig. 4a). With the naked eye, the microtopography is irregular to fluid, with overthicknesses and fine to deep striations in multidirectional, at times intersecting bands, indicative of a wet hard tool on wet clay (Baldi 2015: 127). These striations are sometimes overlapped by reticulated threaded ones left by the hand or fingers on wet clay, likely created as potters held the vessel while scraping its exterior or finishing its interior.
37Potters subsequently regularized the interior of the vessel by moving their wet fingers in a horizontal to oblique direction; this operation turned into a smoothing (fig. 4b). With the naked eye, the microtopography appears as a combination of compact and fluid areas, with subparallel bands of thin ribbed striations in multiple directions, sometimes overlapping the horizontal ones of the rim. At high magnification, the grains are embedded and protruding, partly covered by clay, depending on the paste composition and hygrometry. These elements are diagnostic of the use of a wet tool on a not-so-wet clay (Roux 2017: P1/4).
38After finishing the interior, potters usually left the form to dry. At times, they painted the vessel’s surfaces.
Fig. 5 – Diagnostic elements of TG3.
A. Inner surface with fluid microtopography with subparallel bands of ribbed striations associated with thin marks (marked with arrows) left by a hard tool on wet clay; B. Outer surface with horizontal rows of rounded depressions (marked with dashed circles) and bumps along the rim. The latter has a fluid microtopography, which contrasts with the compact one of the lower body; C. Outer surface, lower body with compact microtopography with crevices and rare deep striations from shaving.
39TG3 includes techniques similar to TG1 but in a different order and on different surfaces. After spreading 1.5-2.5 cm coils, potters regularized the surfaces of the rough-out with their hand. They shaped the rim by continuously pressing the wet clay with a wet soft tool (fingers or cloth). The corresponding traces are similar to those of TG1, but are mostly preserved on the exterior.
40As a next step, the interior was scraped with a wet tool in a horizontal to oblique direction; as the clay dried up, the continuous scraping turned into smoothing (fig. 5a; Baldi 2015: 131). On the interior, the microtopography is fluid with horizontal, subparallel ribbed striations, usually associated with thin, elongated tool imprints, pointing to the use of a hard tool (Roux 2019: 174-175).
41After finishing the interior, and likely after a drying phase, potters regularized the exterior by moving a wet soft tool on a not-so-wet surface. In the case of open shapes with a more or less pronounced carination, they possibly turned the vessel with the rim facing down, and then shaved the leather-hard clay from the bottom upwards until the carination (fig. 5b). These operations left different traces. On the upper body, at high magnification a combination of fluid and compact microtopography can be observed, with thin to deep ribbed striations, horizontally oriented. On the lower body, the compact microtopography shows crevices and rare deep striations, horizontal to oblique, characteristic of a leather-hard clay tearing under the action of the shaving tool (fig. 5c; Roux 2019: 177). These striations overlap the ribbed, horizontal ones of the upper body, showing that the lower body was finished at the end of the sequence.
42After this final step, potters could paint or burnish the vessel, mostly its exterior.
Fig. 6 – Diagnostic elements of TG4.
A. Irregular profile of the body and base, with a peripheral coil (marked with a dashed line) added to the exterior of the base-body join and to the rim; B. Outer surface roughly regularised with the wet hand, with rounded depressions left by fingers (marked with dashed circles) during roughout still visible; C. Inner surface smoothed by the means of a wet tool moved in a horizontal direction.
43In contrast to what was seen with TG1-3, in TG4 potters started by modelling a clay lump by pinching and drawing. The base is flattened. Diagnostic elements are the vessels’ irregular profile, with no discontinuities attributable to coils (fig. 6a); the exterior is irregular and uneven with horizontal rows of rounded depressions and bumps left by interdigital pressures on the clay during pinching (fig. 6b). The base has an irregular thickness, as potters could not control its thickness while shaping it, which is characteristic of drawing a clay ball. At high magnification, fissures, voids, and inclusions are elongated and with a subparallel orientation in proximity to the walls, but are irregular in the inner part of the wall. These are diagnostic elements for a clay mass deformed by vertical compression through drawing (Roux 2019: 169-170).
44Through discontinuous pressure, potters added a peripheral coil along the exterior of the base-body join to strengthen it. Similarly, they could add a 2-3 cm coil to make the rim. In section, a curvilinear fissure can be observed at the base-body join (fig. 6a). At high magnification, voids and inclusions of the peripheral coil have a different orientation than those of the base; the same applies to the coil of the rim (Roux 2019: 163-165). In this case, sherds show either a preferential horizontal fracture along the upper body or a distinctive externally bevelled fracture (fig. 6b; Roux 2019: 160-162).
45For preforming the interior, potters mostly used their wet fingers for the base and body, and very occasionally a hard tool moved in a horizontal direction and only for the upper body (fig. 6c). This operation would turn into smoothing as the clay dried up. Along the body, fine ribbed striations in subparallel horizontal bands were observed, sometimes associated with a thin elongated vertical imprint left by a hard tool.
46Potters roughly regularized the exterior by hand when the clay was still wet (fig. 6b). It is possible that while the active hand preformed the interior, the other hand regularized the exterior from time to time, i.e. the two operations happened concomitantly. Therefore, rounded depressions and bumps left by pinching during the rough-out are still visible on the exterior, together with fine reticulated threaded and ribbed striations.
Fig. 7 – Diagnostic elements of TG5.
A. Outer surface with overthickness (marked with a dashed line), depressions (marked with dashed circles), and bumps along the rim, and scraping marks along the body; B. Inner surface with overthickness (marked with a dashed line) at the rim-body join and very fine striations left while preforming by hand.
47Potters started by modelling a clay ball through pinching and drawing; these operations left traces similar to those of TG4. A ca. 2 cm coil could be added to make the rim, formed through continuous pressure with a wet soft tool moved in horizontal direction. On both surfaces, elongated horizontal overthicknesses are visible at the rim-body join, with an oblique fissure in the clay paste visible in section (fig. 7a-7b). The microtopography of the rim is fluid with ribbed striations.
48As a next step, the interior was preformed by moving the wet hand on the wet clay in a horizontal to oblique direction. The inner microtopography combines irregular and fluid areas, with very fine striations in discontinuous, subparallel bands in a horizontal to oblique direction, sometimes intersecting (fig. 7b). Finger prints are preserved together with localized clay accumulations, due to the hand moving the wet clay along the surface. These elements are distinctive of a wet clay worked with wet fingers (Baldi 2015: 127-128).
49Finally, potters turned the form with the rim facing down and scraped the exterior from the bottom upwards with a lithic blade (fig. 7a; van As and Jacobs 2004; Baldi 2015: 126-129). With the naked eye, the surface appears as irregular with intersecting bands of deep striations oriented from the bottom upwards, sometimes overlapped by finger prints possibly left by the hand holding the preform while scraping.
- 10 Also in the Middle Euphrates, at Tell Abr, Stage II (Yamazaki 2010).
50Technical groups 1-5 are present at all three sites considered here (table 4). TG1-3, characterised by joining 1.5-2.5 cm coils, are predominant in period IV (tables 5-7). In contrast, TG4-5, based on modelling a clay lump, are better attested in period V (tables 8-9). In phase IVA-B, it appears that TG1 and TG3 (coiling and smoothing) were associated with both black-on-buff and unpainted shapes, while TG2 (coiling and outside scraping) is attested only in limited numbers for both painted and unpainted productions. However, in the following phase IVC-D TG1 and 3 decrease at all sites together with the general decrease in black-on-buff pottery. Conversely, TG2 becomes more frequent, especially at Tell Zeidan. Yet, the scraping technique for the exterior was already in use in phase IVA-B,10 not only with TG2 but also with TG1, where at times scraping traces were not completely obliterated during smoothing. TG4-5 are found mostly in period V, though the number of sherds of this group (and, in general, period) is rather low in comparison to that of other groups. TG4 includes predominantly Coba bowls; based upon their manufacture, a similarity with Type III of Baldi’s Coba bowl typology (2012b: 397-398) can be suggested here. Similarly, TG5 also mostly includes Coba bowls, comparable to Baldi’s Type II (2012b: 398).
Table 4 – Techno-petrographic groups at Tell Zeidan (BS2), Tell Sawwan (BS147) and Tell Hammam et-Turkman (BS175) in Balikh IV-V periods.
TG |
Fabric |
BS2 |
BS147 |
BS175 |
TG1 1.5-2.5 cm coils, ext. scraping-smoothing, int. scraping |
A1 |
5 |
9 |
30 |
A2 |
1 |
1 |
6 |
A3 |
12 |
20 |
15 |
B1 |
6 |
10 |
10 |
B2 |
1 |
3 |
5 |
C1 |
33 |
2 |
1 |
D1 |
19 |
0 |
0 |
F1 |
1 |
0 |
0 |
TG2 1.5-2.5 cm coils, ext. scraping, int. scraping-smoothing |
A1 |
1 |
2 |
3 |
A2 |
2 |
0 |
0 |
A3 |
20 |
17 |
15 |
B1 |
21 |
8 |
9 |
B2 |
8 |
0 |
5 |
C1 |
56 |
2 |
1 |
D1 |
179 |
0 |
0 |
F1 |
5 |
0 |
0 |
TG3 1.5-2.5 cm coils, ext. scraping, int. scraping-smoothing |
A1 |
5 |
11 |
33 |
A2 |
0 |
0 |
5 |
A3 |
2 |
18 |
24 |
B1 |
6 |
11 |
15 |
B2 |
5 |
1 |
25 |
C1 |
8 |
1 |
0 |
D1 |
3 |
0 |
0 |
E1 |
0 |
1 |
3 |
E2 |
1 |
0 |
3 |
TG4 Modelling clay ball, int. + ext. smoothing |
B1 |
0 |
2 |
0 |
B2 |
8 |
1 |
1 |
F1 |
0 |
1 |
1 |
TG5 Modelling clay ball, ext. scraping |
B1 |
0 |
3 |
0 |
B2 |
0 |
0 |
4 |
F1 |
1 |
0 |
0 |
|
409 |
124 |
214 |
Table 5 – Distribution of techno-petrographic groups (TPG) of technical group 1 at Tell Zeidan (BS2), Tell Sawwan (BS147) and Tell Hammam et-Turkman (BS175) in Balikh IV-V periods.
TPG |
Phase |
BS2 |
BS147 |
BS175 |
TG1.A1 |
IVA-B |
5 |
9 |
28 |
IVC-D |
00 |
0 |
2 |
VA |
0 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG1.A2 |
IVA-B |
1 |
1 |
5 |
IVC-D |
0 |
0 |
1 |
VA |
0 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG1.A3 |
IVA-B |
8 |
11 |
13 |
IVC-D |
3 |
8 |
1 |
VA |
1 |
0 |
1 |
VB |
0 |
0 |
0 |
VA-B |
0 |
1 |
0 |
TG1.B1 |
IVA-B |
1 |
1 |
2 |
IVC-D |
4 |
7 |
7 |
VA |
0 |
1 |
1 |
VB |
0 |
0 |
0 |
VA-B |
1 |
1 |
0 |
TG1.B2 |
IVA-B |
0 |
0 |
0 |
IVC-D |
1 |
0 |
4 |
VA |
0 |
3 |
1 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG1.C1 |
IVA-B |
25 |
1 |
1 |
IVC-D |
8 |
0 |
0 |
VA |
0 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
1 |
0 |
TG1.D1 |
IVA-B |
2 |
0 |
0 |
IVC-D |
14 |
0 |
0 |
VA |
2 |
0 |
0 |
VB |
1 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG1.F1 |
IVA-B |
0 |
0 |
0 |
IVC-D |
0 |
0 |
0 |
VA |
1 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
Table 6 – Distribution of techno-petrographic groups (TPG) of technical group 2 at Tell Zeidan (BS2), Tell Sawwan (BS147) and Tell Hammam et-Turkman (BS175) in Balikh IV-V periods.
TPG |
Phase |
BS2 |
BS147 |
BS175 |
TG2.A1 |
IVA-B |
1 |
2 |
3 |
IVC-D |
0 |
0 |
0 |
VA |
0 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG2.A2 |
IVA-B |
1 |
0 |
0 |
IVC-D |
0 |
0 |
0 |
VA |
0 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG2.A3 |
IVA-B |
17 |
12 |
15 |
IVC-D |
2 |
5 |
0 |
VA |
0 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
1 |
0 |
0 |
TG2.B1 |
IVA-B |
1 |
2 |
1 |
IVC-D |
12 |
6 |
4 |
VA |
2 |
0 |
3 |
VB |
1 |
0 |
1 |
VA-B |
5 |
0 |
0 |
TG2.B2 |
IVA-B |
0 |
0 |
0 |
IVC-D |
1 |
0 |
1 |
VA |
0 |
0 |
2 |
VB |
2 |
0 |
2 |
VA-B |
5 |
0 |
0 |
TG2.C1 |
IVA-B |
23 |
1 |
1 |
IVC-D |
29 |
1 |
0 |
VA |
2 |
0 |
0 |
VB |
2 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG2.D1 |
IVA-B |
5 |
0 |
0 |
IVC-D |
140 |
0 |
0 |
VA |
28 |
0 |
0 |
VB |
2 |
0 |
0 |
VA-B |
4 |
0 |
0 |
TG2.F1 |
IVA-B |
0 |
0 |
0 |
IVC-D |
0 |
0 |
0 |
VA |
2 |
0 |
0 |
VB |
1 |
0 |
0 |
VA-B |
2 |
0 |
0 |
Table 7 – Distribution of techno-petrographic groups (TPG) of technical group 3 at Tell Zeidan (BS2), Tell Sawwan (BS147) and Tell Hammam et-Turkman (BS175) in Balikh IV-V periods.
TPG |
Phase |
BS2 |
BS147 |
BS175 |
TG3.A1 |
IVA-B |
5 |
11 |
32 |
IVC-D |
0 |
0 |
1 |
VA |
0 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG3.A2 |
IVA-B |
0 |
0 |
5 |
IVC-D |
0 |
0 |
0 |
VA |
0 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG3.A3 |
IVA-B |
2 |
9 |
21 |
IVC-D |
0 |
9 |
2 |
VA |
0 |
0 |
1 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG3.B1 |
IVA-B |
0 |
3 |
4 |
IVC-D |
3 |
4 |
9 |
VA |
0 |
1 |
1 |
VB |
2 |
3 |
1 |
VA-B |
1 |
0 |
0 |
TG3.B2 |
IVA-B |
0 |
0 |
0 |
IVC-D |
1 |
0 |
6 |
VA |
1 |
1 |
17 |
VB |
3 |
0 |
2 |
VA-B |
0 |
0 |
0 |
TG3.C1 |
IVA-B |
7 |
1 |
0 |
IVC-D |
1 |
0 |
0 |
VA |
0 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG3.D1 |
IVA-B |
0 |
0 |
0 |
IVC-D |
3 |
0 |
0 |
VA |
0 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG3.E1 |
IVA-B |
0 |
0 |
0 |
IVC-D |
0 |
0 |
0 |
VA |
0 |
0 |
0 |
VB |
0 |
1 |
3 |
VA-B |
0 |
0 |
0 |
TG3.F1 |
IVA-B |
0 |
0 |
0 |
IVC-D |
0 |
0 |
2 |
VA |
0 |
0 |
0 |
VB |
0 |
0 |
1 |
VA-B |
1 |
0 |
0 |
Table 8 – Distribution of techno-petrographic groups (TPG) of technical group 4 at Tell Zeidan (BS2), Tell Sawwan (BS147) and Tell Hammam et-Turkman (BS175) in Balikh IV-V periods.
TPG |
Phase |
BS2 |
BS147 |
BS175 |
TG4.B1 |
IVA-B |
0 |
0 |
0 |
IVC-D |
0 |
0 |
0 |
VA |
0 |
2 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG4.B2 |
IVA-B |
0 |
0 |
0 |
IVC-D |
0 |
0 |
0 |
VA |
8 |
1 |
1 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG4.F1 |
IVA-B |
0 |
0 |
0 |
IVC-D |
0 |
0 |
0 |
VA |
0 |
1 |
1 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
Table 9 – Distribution of techno-petrographic groups (TPG) of technical group 5 at Tell Zeidan (BS2), Tell Sawwan (BS147) and Tell Hammam et-Turkman (BS175) in Balikh IV-V periods.
TPG |
Phase |
BS2 |
BS147 |
BS175 |
TG5.B1 |
IVA-B |
0 |
0 |
0 |
IVC-D |
0 |
0 |
0 |
VA |
0 |
3 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG5.B2 |
IVA-B |
0 |
0 |
0 |
IVC-D |
0 |
0 |
1 |
VA |
1 |
0 |
3 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
TG5.F1 |
IVA-B |
0 |
0 |
0 |
IVC-D |
0 |
0 |
0 |
VA |
1 |
0 |
0 |
VB |
0 |
0 |
0 |
VA-B |
0 |
0 |
0 |
51At this stage of analysis, some degree of variability in terms of fabrics and manufacturing methods appears in the Balikh IV-V assemblage. Research shows that variability can result from functional or sociological factors (Roux 2019: 1-14) as there is no fixed relation between a technique and a specific fabric and vessel size or shape, and potters can use the very same techniques to make a variety of vessels. In other words, potters can choose specific techniques and fabrics in light of the function of the vessel to be made, with cooking pots being a classical example. However, when different procedures are applied to make the same morpho-functional categories, one should rather seek a sociological explanation and thus envisage a scenario where different groups of producers were active in an area at a given time.
- 11 Due to the fragmentation of the assemblage, information on closed shapes is generally limited.
52This study set out with the aim of assessing pottery-making practices in the Balikh valley between the Ubaid and the LC2 by examining fabrics and manufacturing techniques. One interesting finding is the existence of six fabric macrogroups and five different technical groups, whose combinations result in 31 techno-petrographic groups (table 4). The coil-based groups 1-3 appear as the most frequent at Tell Hammam et-Turkman, Tell Sawwan and Tell Zeidan over period IV (tables 5-7). In phase IVA-B, TG1 and 3, characterized by smoothing the exterior and interior respectively, are mostly associated with fine mineral fabrics (macrogroups A, C) and both black-on-buff and unpainted vessels, such as flaring, hemispherical, sharply incurving, s-shaped or carinated bowls with a variety of rims, and jars with a bow-rim or outrolled lip.11 In phase IVC-D, these groups together with fine fabrics and painted decoration decrease in number at all sites, with a parallel increase in coiled vessels with scraped outer surfaces of TG2, particularly well-documented at Tell Zeidan (table 6). TG2 occurs for both open and closed shapes already in phase IVA-B, although in small numbers; in phase IVC-D, it is mostly associated with flaring, hemispherical or slightly carinated, beaded-rim bowls. In general, the increase in exterior scraping coincides with the more frequent use of organic-tempered fabrics (macrogroups B, D). In period V, both coiling and modelling techniques are attested in the assemblage (tables 8-9), although the number of sherds is admittedly small. TG3 appears with organic-tempered, brown to dark grey fabrics (macrogroup E; table 7) used for holemouth pots with beaded rim as well as bowls with grooves on the exterior (Akkermans 1988: 144, fig. 9: 138-139; 145, fig. 10: 151-153). TG4-5, based on modelling a clay lump, are associated only with coarse organic-tempered fabrics (B, F; tables 8-9) used for mass-produced bowls (Baldi 2012b). Overall, these results are in line with those of previous studies (Akkermans 1988; Baldi 2016b) showing changes in technical behaviour as Ubaid fine-mineral painted vessels slowly disappeared and organic-tempered, unpainted vessels, and related serial productions emerged, starting in the transition to the LC1.
- 12 Social boundaries (ethnic, linguistic, kin, etc.) between producers are hard to assess for prehisto (...)
53At the beginning of this paper I stated that a chaîne opératoire and the habitus underlying it are an expression of a learning network and thus a social group. This technological study showed that at least 31 techno-petrographic groups are present in the Balikh valley over periods IV and V. As such, the Balikh IV-V ceramics can be characterized as a complex homogeneous assemblage (Roux 2019: 246-247), where different communities of potters followed different technical practices. On a synchronic axis, this variability is an expression of different groups of producers12 active in the region. The variability in fabrics might result from differences in clay procurement and processing, with potters accessing a variety of clay sources close to the production site or within the radius of the exploited territory (Gosselain 1998). This is especially true for organic-tempered fabrics of phase IVC-D, which show internal differences in the sorting and size of their organic and mineral inclusions, despite coming from a similar geological environment, thus pointing to a low level of standardization in the exploitation of clay sources and paste recipes (Arnold 2000; Fragnoli 2021). On a diachronic axis, the decrease in old traditions (TG1, 3) and their fabrics (A, C) parallel to the increased use of a tradition (TG2) and fabrics (B, D), and the emergence of new traditions (TG4-5) and categories of vessels suggest changes in socio-economic structures, with new forms of production including emerging specialization.
54Looking closely, 25 TPGs are documented in the valley in period IV. This technological variability where potters accessed similar clay sources and made similar functional groups of pottery suggest that pottery making was mainly based in the household. Part-time, likely seasonal, specialists produced mostly for household needs while sharing kilns with the community (Akkermans and Schwartz 2003: 170-171; Özbal 2010; Yamazaki 2010). The Ubaid painted pottery shows a high degree of uniformity with simple decorative motifs. With the transition to the LC1 (Balikh IVC-D), these vessels changed their social value and were replaced by unpainted ones, increasingly standardized in their technology and morphology. In period V, the disappearance of painted vessels, the emergence of new ceramic traditions with limited fabric variability and serial production (Coba bowls) support the hypothesis of increasing specialization, hinting at a change in the cultural framework and related behaviour, i.e. habitus. This change happened slowly. Selection, innovation, borrowing or assimilation change technical traditions over time, provided that people accept these technological and social transformations. The increase in the occurrence of organic-tempered fabrics and coiling-and-scraping as well as the introduction of the modelling technique show the gradual acceptance of specific paste preparation and manufacturing practices deemed more suitable for people’s changing needs. In a society with new social relations of production where competing, large households provided the food consumed in communal settings (Frangipane 2012; Kennedy 2015), was there any need for distinctiveness in ceramics? Over time, pottery became just a utilitarian product expediently produced by a reduced number of part-time specialists in workshops (Baldi 2012a, 2012b).
55Based on the distribution of the techno-petrographic groups and the evidence of on-site ceramic production, it can be hypothesised that production centres were located at each of the three sites investigated, with different groups of potters living together in the same settlement. With the transition to the LC1, large households started competing with each other to control territory with its resources and perhaps sought to extend their social ties by means of allegiances. Given that a chaîne opératoire is the technical expression of a social identity (Baldi 2016a), the diachronic transformations in the techno-petrographic groups point to changing social components. Perhaps some producers stopped making ceramics or joined other groups, thus learning new skills—for example, the modelling-based TG4-5 used for Coba bowls, or TG3.E1 and TG3.E2, both organic-tempered, brown to dark grey fabrics, used only for holemouth cooking pots and grooved bowls. Over time the decrease in the variability of chaînes opératoires and paste recipes might indicate that some potters’ groups abandoned their practice and likely even the region as a result of changes in the socio-economic context of the period, which led to the emergence of the first urban centres of the LC3 in the nearby Khabur plain, a phenomenon for which there is no evidence at present in the Balikh valley.