Thèse : référence bibliographique
Claire Padovani, L’organisation de la production céramique des premières sociétés urbaines.Analyse technique et sociale des pyrotechnologies et des espaces de travail potiers de l’Asie du Sud-Ouest (4e et 3e millénaires av. n. è.), thèse de doctorat en archéologie, Université Paris 1 Panthéon-Sorbonne, soutenue le 20 décembre 2023, deux volumes (930 p).
Directeur de thèse
Pascal Butterlin, professeur d’archéologie du Proche-Orient ancien, Paris 1 Panthéon-Sorbonne
Jury
Eleni Hasaki, présidente : Professeure, Université d’Arizona
Régis Vallet, co-encadrant : Chargé de recherche, UMR 8068 Temps
Luca Peyronel, rapporteur : Professeur, Université de Milan - La Statale
Johnny Samuele Baldi, examinateur : Chargé de recherche, UMR 5133 Archéorient
Augusta McMahon, examinatrice : Professeure, Université de Chicago
1My thesis focuses on the organisation of ceramic production in early urban societies and aims to document the transformations of pyrotechnologies and pottery spaces between the beginning of the 4th and the end of the 3rd millennium BCE, based on fresh excavation data and bibliographic sources, spanning a territory from western Iran to western Syria. The very recent discovery, in northern Iraq, of large-scale pottery workshops characterised by the use of a significant number of kilns, has instigated the reconsideration of mundane ceramic production organisation during the crucial phase of urbanisation. Moreover, it triggered the critically reassessment of a simplistic and linear techno-social evolution, with an innovative approach integrating elements typically overlooked or marginalised in pottery production studies.
2In the first part of my thesis, I delved into the historiography of firing structures and craftsmanship in ancient Southwest Asia, with the aim of deconstructing theoretical concepts such as mode of production, specialisation, innovation, workshop, and industry. I also appraised the various methodologies archaeologists have employed in studying production techniques and pottery kilns. Furthermore, considering that the organisation of production activities is a structuring element of societies (Durkheim 1895), I synthesised an overarching understanding of social and productive protohistoric systems, mainly by reviewing the archaic signs of the potter’s name and the translated cuneiform texts from the late 3rd millennium related to production (Barrelet 1964; Sallaberger 1996). Finally, a whole chapter of the first part was dedicated to the functioning of the firing structures. I approached this topic in a theoretical way, by documenting the general mechanics of firing; and also in practical, anthropological, sensorial ways, by excavating workshops from different periods in France, Jordan, Kurdistan, and southern Iraq, constructing a kiln, conducting three firing experimentation sessions, and engaging with potters to discuss and observe firings, particularly in the villages of La Borne in France, Chemchamal, and Aliawa in Kurdistan. This comprehensive perspective allowed me to develop a method focused on the technical and social analysis of production facilities and spaces.
3The second part of my thesis discusses the morpho-technical profiles of the 479 firing structures, tools, installations and production contexts identified on the 50 sites of the corpus, classified by region and period. Additionally, data were formatted for implementation into an online database, with the ambition to ensuring open access to it1.
4In the third part, I compared and put into perspective data on production structures and spaces by period and by region. I highlighted regional traditions of firing and organisation of ceramic production labour, which overlap with dynamics of innovation diffusion and transformations associated with urbanisation.
5The primary focus of research in pottery production has been on the final product. The spaces and means of production such as firing structures, tools, and other related production facilities (basins, working surfaces, etc.) have been largely overlooked (Brumfiel, Earle 1987; Clark, Parry 1990; Stein, Blackman 1993; Stein 1996). At stake, the low visibility of production remains in the field compared to more monumental vestiges – such as palaces, temples, and ziggurats – which have been the focus of excavations. This taphonomique bias led archaeologists to assume that production took place on the less explored outskirts of inhabited sites, or that it was largely controlled by state or royal authorities (Mumford 2019 (1966); Gelb 1969; Stein 1996). Recently, a paradigm shift helped by the increase of survey campaigns, directed our attention to these areas and their significance in understanding the organisation of societies through the populations’ daily activities (Costin 2020; Pollock 2013).
6Although the pottery kiln played a pivotal role in perfecting firing techniques, artisanal firing structures in Southwest Asia have received relatively little attention. Archaeologists initially attempted to categorise these structures from simple to complex, or from the least efficient to the most efficient, efficiency being interpreted based on the supposed capacity to reach higher temperatures (Delcroix and Huot 1972; Majidzadeh 1975). However, all firing structures, including a simple hearth, can reach the necessary temperature of 600-800°C to transform clay material into ceramic (Gosselain 1992; Livingstone Smith 2001b; Thér 2014).
7Therefore, the concept of efficiency is not a relevant criterion for distinguishing protohistoric structures. Moreover, archaeologists encountered a large diversity of firing structures types, which did not appear to follow any regional or chronological pattern (Boroffka, Becker 2004). In sum, this linear and functionalist perception of the technical evolution, and the tendency to consider the kiln as an object isolated from its context, hindered their ability to identify the nature of pyrotechnological structures and the actual diffusion processes responsible for their morpho-technical variability.
8I used the philosophy of techniques and the work of Gilbert Simondon (1958), in order to understand the technical nature of firing structures and to discern their dynamics of transformation. Kilns must be considered as systems, characterised by an internal arrangement of functional units. The whole structure achieves the ceramic firing, with each unit serving its own technical function: heat production, heat circulation, or heat transfer to the clay material. Depending on the configuration of these units within each structure – for example, in a kiln with superposed chambers or in a single-chamber structure – the method of heat transmission to the clay material differs, thus resulting in different firing techniques.
9The philosophy of techniques highlights the need to differentiate the structural arrangement (i.e., the technical design) from the shape. Therefore, two kilns with the same technical design can have different shape. Furthermore, two kilns with the same technical design can share certain morphological attribute (the same grate), while differing in others (different grate supports). The technical design and the shape are not transmitted as a uniform package but independently of each other, showing that potters can borrow morphological elements from multiple traditions. Therefore, it is unhelpful to establish a strict typo-chronology; instead, it is more valuable to document trends in the use of specific technical design, shape or attribute within particular regions or periods. Through the differentiation of technical design and shape, and the contextualisation of structures, their variability becomes more apparent, facilitating the comprehension of their diffusion over time and space.
10Understanding social and technical transformations on the large scale required an examination of a lengthy period spanning two millennia, and an investigation of a consistent territory comprising three macro-regions linked by protohistoric dynamics:
-
the Mesopotamian basin and Khuzestan.
-
the Tigris region, including Diyala, Hamrin, Qara Dagh, and Upper Tigris.
-
the Jezireh, Khabur, Middle Euphrates, and three sites in the Northern Levant.
11Data from Girdi Qala, Logardan and Aliawa were collected during multiple excavation campaigns by the excavation teams and myself. They constitute the core of the corpus in terms of quality and their collection required to learn the excavation, recognition, and recording of structures related to ceramic production areas (Willems, Favennec 2023).
12The surface area of excavation varies from one site to another, ranging from a few square meters to the entire site. Consequently, the presence and size of production areas on each site depends on the extent of the excavation. This problem is inherent to archaeology, especially in difficult-to-access territories. The second half of the data comes from publications and excavation reports dating from the 19th century to the present day. Firing structures and production spaces have seldom been the focus of dedicated publications or thorough referencing. Bibliographic data were highly scattered and collected through keyword searches in catalog search engines, as well as systematic reading of publications related to the studied period.
13The lack of a systematic excavation protocol, variations in data origin (authorship, time period, scientific approach), and disparities in the preservation of remains contributed to significant variability in the availability and quality of data within the same site and across different sites. Consequently, it was necessary to (re)evaluate both the bibliographic and newly excavated datasets. In the specific case of firing structures, I faced a double challenge: distinguishing between cooking ovens and ceramic kilns within pyrotechnological installations, and identifying technical design despite the poor state of preservation of some structures. Among the 479 firing structures in the corpus, only three were fully preserved. In all other instances, the upper portions of the structures were significantly eroded. Nevertheless, I chose to include all structures in the online database in order to provide as many comparative examples as possible and to tackle challenges in interpretation arising from the lack of documentation. To address identification issues of fragmentary data, I relied on contemporary examples. In addition, to facilitate data reuse, I assigned a confidence index to each structure, ranging from 0 (not a ceramic firing structure) to 3 (well-preserved, reliable identification), which I considered in my final interpretation.
14I established a three-steps method to describe and compare the firing structures and production spaces within the corpus.
15Over the period under study, I identified five technical designs by deconstructing each firing structure into distinct functional units and observing their internal arrangement:
-
technical design V (for Vertical draught kiln), the firing chamber is above the combustion chamber, creating a vertical draught.
-
technical design H (for Horizontal draught kiln), the firing chamber is juxtaposed to the combustion chamber, creating a horizontal draught.
-
technical design S (for Single-chamber structure), the combustion and firing chambers are combined in a single chamber, the direction of the draught is contingent upon the positioning of the air inlet and outlet.
-
technical design P (for Pit), there is no chamber, pots are arranged directly on fuel, placed in a pit.
-
technical design O (for Open hearth), there is no chamber, pots are arranged directly on fuel, above the floor.
16The terminology employed is not new (Delcroix, Huot 1972), but holds technical significance, especially in distinguishing kilns from other firing structures. Kilns are characterised by the separation of the heat production and heat transfer units. They mainly use convection (or draught) for firing clay material, whereas in other structures, firing is predominantly achieved by heat concentration and transmission through conduction.
17Additionally, the shape of each technical design can differ depending on the dimensions of each functional unit and the presence/absence of internal architectural features such as pillars, pilasters, benches or grates. This parameter is not trivial, as the shape choice can impact the firing process without requiring to modify the structure, i.e., the arrangement of functional units.
18The activities of the ceramic chaîne opératoire are interrelated through the manufacture of the final product (Leroi-Gourhan 1943). However, they exhibit a certain degree of technical and spatial independence (Cresswell 1976; Lemonnier 1983). Although they don't require immediate proximity, these tasks must remain reasonably close for practical purposes. For instance, shaping areas cannot be too distant from firing areas due to the difficulty of transporting raw pottery over long distances. A potter's “workshop” is not a homogeneous and continuous space; rather, it consists of multiple working spaces where one or more activities of the chaîne opératoire take place (Hasaki 2011).
19The function of each working space is identified through the presence of tools, in situ structures, or waste associated with specialised activities (Nieuwenhuyse and Akkermans 2019; Pollock 2013; Verhoeven 1999). Usually, the most visible features in production contexts are firing structures and over-fired potsherds. Consequently, these elements are the initial focus of the analysis: once firing areas are identified, it becomes easier to spot indications of other workspaces associated with ceramic production nearby.
20I compare the production contexts to identify recurring spatial organisation forms. I examine the degree of concentration of the working spaces and their location within each site:
-
degree 1, isolated activity (usually indicating that the context of the structures has not been recorded);
-
degree 2, some production installations are gathered in the same vicinity;
-
degree 3, "Workshop", the installations of the pottery chaîne opératoire are concentrated within a well-defined area;
-
degree 4, cluster of identical structures used for one activity of the chaîne opératoire (for example, a concentration of kilns for firing).
21These working spaces can be situated in various locations such as the outskirts of an inhabited site, in courtyards between residential areas, within a domestic space, within a building accommodating multiple productions, in a specific location distant from residential areas, etc. Then, I use ethnographic data to propose hypotheses on the social organisation of labour for each pattern of spatial organisation.
22I delivered an inventory of all excavated production areas in Southwest Asia during the 4th and 3rd millennia, along with detailed descriptions of firing structures and other associated production installations and tools, accessible through an online repository. This endeavor fills a gap in the (pre)history of artisanal pyrotechnology development. Moreover, drawing from archaeological data, I aimed to highlight the entire socio-technical system of ceramic production.
23During the 4th millennium BCE, vertical draught kilns were preferentially used in Mesopotamia, Khuzestan, Hamrin, and Qara Dagh, while single-chamber structures were more commonly employed in the Syrian region and Upper Tigris. In the 3rd millennium BCE, I identified an equal number of vertical draught kilns and single-chamber structures, but the former was widespread across a greater number of sites and tended to become the foremost pottery firing technique. The latter was mainly used in the Syrian region and north-western Iraq or in production contexts associated with an administration or a temple. The horizontal draught kilns were only found in the Tigris region.
24If we observe this distribution in the long term – i.e. back to the 6th millennium BCE, at the beginning of ceramic pyrotechnology – we notice a regional continuity in firing techniques. Single chamber structures were more commonly used by the Hassuna and Halaf groups located in the northern part of the territory under study, while vertical draught kilns were predominantly used by the Obeid communities in Mesopotamia. We suppose that vertical draught kilns spread northward, gradually and in a not-linear way, from the Halaf period. It was only from the 3rd millennium BCE that this technique became a widely shared standard for ceramic firing. Horizontal draught kilns were used in the eastern Tigris region and the Khabur as early as the Obeid 3-4 period (5300-4800 BCE), but were never diffused on a large scale. This technique is much more widespread in Iran. In sum, we infer the long-lasting existence of regional technical memories or technical backgrounds inherited from the beginnings of ceramic pyrotechnology.
25Three major innovations emerged, related to the circulation of heat within and between firing structures. These innovations reflect the expertise of potters in mastering convection, leading to optimising time and energy usage.
-
The firebox. Within a firing structure, the firebox is the constructed feature in which the hearth is located. It extends the heat production unit, allowing more space for combustion gases to burn and increase the temperature. In summary, it optimises fuel consumption. In older models of vertical draught kilns, the firebox is absent; instead, the hearth is located in the combustion chamber where combustion gases burn. However, starting from the 6th millennium BCE, Obeid potters used kilns with an elongated mouth. This element constitutes the beginnings of the firebox and was widespread towards the Diyala by being integrated into kilns of different shapes. By the 5th millennium BCE, this elongation between the mouth and the combustion chamber is constructed and covered, becoming a fully-fledged architectural attribute. It was found on some Iranian sites, and I presume it should also be found in the Mesopotamian basin, although we do not yet have data to confirm this. In the 4th millennium BCE, starting from the Middle Uruk period, this invention was transformed into innovation and spread to Iran, as well as through the Uruk network towards the Diyala and the Syrian region.
-
The airflow control points. They are 10 cm diameters pipes that can be opened and closed, connecting the inside to the outside of the kiln. Operating like air pumps, they facilitate the influx of fresh air into the combustion chamber and enable the combustion of gases. The oldest known examples date back to the mid-4th millennium.
-
The connections between kilns. The earliest known interconnected kilns date back to the inception of the 4th millennium. Their walls are constructed together, and their combustion chambers are linearly connected (from chamber to chamber) as observed at Girdi Qala and Logardan in the Qara Dagh region. At Girdi Qala level 7, three vertical draught kilns are interconnected to create a larger system and facilitate heat circulation. Passages between chambers are redden, and the earthen walls are protected from heat degradation using stones. Usually, a part of the heat produced by the combustion is not transmitted to the clay material and is lost (Valtat 2022). If the kilns of the chain are lit one after the other, a part of the energy produced by the first is used to preheat the second, and then the third. The system enables saving fuel by preheating the combustion chamber and optimising all the energy produced. During the latter half of the 3rd millennium, kiln connections are employed on a large scale and ducts are used to create networks of firing structures. These heat channels facilitate the sharing of hot gases to slowly preheat the structures and their clay material, thereby accelerating the drying process. The construction and operation of these kilns networks require collaborative efforts among several potters.
26During the 4th and 3rd millennia, standardised "workshops" do not exist. Nonetheless, I identified four spatial organisations and used ethnographic examples to interpret social organisations:
-
Domestic workshop. Working spaces are multifunctional and integrated into dwellings, with firing taking place in the household courtyard. Drawing from ethnographic data, only one individual is in charge of the production, yet the entire family help with the difficult tasks such as raw material collection and firing (London 2024). Occasionally, certain tasks may be delegated to individuals out of the family group.
-
Cluster of nucleated workshops. The working spaces of each chaîne opératoire are concentrated into workshop and multiple workshops are clustered in the same area. This suggests that various groups of potters work in close proximity to each other and likely emphasizes the possibility of intergroup assistance for difficult tasks (Golvin, Thiriot, Zakariya 1982).
-
Shared working areas. Multiple production installations are gathered in open areas. This concentration of means provides the opportunity for technical assistance and suggest common space management, though not necessarily cooperation among potters groups sharing these areas (Gosselain 2002).
-
Cooperative working areas. The design and layout of connected spaces and structures suggest that multiple potters use the area and work together for certain tasks such as construction and firing of connected kilns. Drawing from ethnographic insights, these areas are likely planned and overseen by a group with a higher level of involvement. While each group of potters carries out its own production process, they achieve economic or technical benefits from cooperation for difficult tasks (Underhill 2003).
27There is no linear transformation from domestic to cooperative production. Instead, regional patterns indicate different traditions of labour organisation across the territory. Domestic and nucleated workshops are mostly localised in the Djézireh, Middle Euphrates, and upper Tigris, while shared or cooperative working areas are mostly concentrated in Khuzestan, Diyala, and the eastern Tigris region. The former suggests that production was predominantly centered around the family or small units whereas the latter suggests a greater degree of economic diversification and specialisation within each family unit and a more structured production management at the site level.
28Actually, pyrotechnological innovations and the adoption of the vertical draught kiln developed more rapidly and extensively in regions where potters tend to gather to work. This phenomenon highlights the link between production organisation and technological development.
29The continuity of technical and organisational traditions is crucial to comprehend production transformations, as it underscores the influence of regional socio-economic backgrounds over the long term. Nevertheless, between the 4th and 3rd millennia, distinct dynamics emerged, adding complexity to this historical continuum.
30In the 4th millennium, available data suggest that in large centers (as Tell Brak) and on Uruk sites, potters were concentrated on the periphery or near key passages (streets, exchange networks), probably by the influence of the ruling power. In Uruk sites or in Khuzestan, potters shared the same firing areas and sometimes gathered all their means of production into more organised and cooperatives complexes as seen at Girdi Qala, Logardan, Tell el’Oueili and Tell Muqayyar (Ur). However, there is evidence suggesting that large households probably possessed their own production annexe, as indicated by findings at Habuba Kabira.
31In the subsequent millennium, a radical shift occurred. Despite data for cities in the Mesopotamian basin are limited, shared firing areas appeared to have become integrated between residential zones, distributed across different neighborhoods such as in Fara and Lagash. Syrian cities exhibited a plurality of organisations: staple ceramic goods were produced in domestic workshops and cluster of nucleated workshops distributed across the settlement, whereas pottery production for institutional purposes, such as for palaces and temples, involved potters being concentrated in shared working area.
32In the latter half of the 3rd millennium, an externalisation of the production to specialised sites, located outside inhabited centers, is observed in the Tigris region (at Aliawa and Logardan) and possibly in southern Mesopotamia (Umm al-Hafriyat). This process facilitated meeting ceramic demands for a broader region than a neighborhood. This change in production scale was not necessarily driven by an increasing socio-economic complexity but originated from a combination of organisational factors (tradition of sharing working areas and cooperating in work), technological factors (exceptional skill in fire management), and probable political factors (imperial power).
33The collected data and results will serve as a working hypothesis for further research. I aim to delve deeper into the question of production organisation by analysing social networks. This approach involves investigating social interactions through the lens of sociological theories, using a variety of data sources, and employing mathematical formulas. Such analyses can be applied on various scales and across multiple regions. Drawing from the outcomes of my research, it would be particularly interesting to explore the connections between Mesopotamia and Iran, as well as to identify potential shifts in organisation depending on the prevailing political model. This project presents the opportunity to use the vast amount of data already collected spanning from the 7th to the 3rd millennium and will require the acquisition of new archaeological and ethnological data. Moreover, I intend to further develop the open database project on firing structures initiated during this thesis by extending the temporal and regional scope.