Pyrotechnological connections? Re-investigating the link between pottery firing technology and the origins of metallurgy in the Vinča Culture, Serbia

Highlights

• Vinča culture sites known to host the earliest metal production in Eurasia.

• Opportunity to study ceramic technology during the transition into the Metal Age.

• Potential of applying material science to the study of pottery technology.

• The results illustrate the remarkable craftsmanship achieved in the Neolithic/Chalcolithic Balkans.

• The results shed new light on the relation between pottery and metal pyrotechnology.

Abstract

The present paper re-examines the purported relationship between Late Neolithic/Early Chalcolithic pottery firing technology and the world's earliest recorded copper metallurgy at two Serbian Vinča culture sites, Belovode and Pločnik (c. 5350 to 4600 BC). A total of eighty-eight well-dated sherds including dark-burnished and graphite-painted pottery that originate across this period have been analysed using a multi-pronged scientific approach in order to reconstruct the raw materials and firing conditions that were necessary for the production of these decorative styles. This is then compared to the pyrotechnological requirements and chronology of copper smelting in order to shed new light on the assumed, yet rarely investigated, hypothesis that advances in pottery firing technology in the late 6th and early 5th millennia BC Balkans were an important precursor for the emergence of metallurgy in this region at around 5000 BC. The results of this study and the recent literature indicate that the ability to exert sufficiently close control over the redox atmosphere in a two-step firing process necessary to produce graphite-painted pottery could indeed link these two crafts. However, graphite-painted pottery and metallurgy emerge at around the same time, both benefitting from the pre-existing experience with dark-burnished pottery and an increasing focus on aesthetics and exotic minerals. Thus, they appear as related technologies, but not as one being the precursor to the other.

Introduction

Pyrotechnology is defined as the “deliberate process utilising the control and manipulation of fire” (McDonnell, 2001, p. 493), or put simply, the use of fire as a tool (Bentsen, 2014). The term is commonly used in connection with high temperature processes including cooking, heating, illumination and particularly the production of synthetic materials such as plaster, ceramics, metals and glass (Roberts and Radivojević, 2015, p. 300). The emergence of metallurgy in particular is seen as an important advance in the history of humankind and has been the focus of historical narratives explaining the evolution of social complexity, amongst others (e.g. Childe, 1944; Craddock, 1995; Renfrew, 1973; Wertime, 1964). Debate has focused on the questions of when and where ancient humans first learned to use fire to extract metal from naturally occurring ore (Gourdin and Kingery, 1975; Jovanović and Ottaway, 1976; Jordan and Zvelebil, 2009; McDonnell, 2001; Roberts et al., 2009; Wu et al., 2012 and literature therein). One of the most influential studies on this topic is by Childe (1944), who asserted that the Near Eastern prehistoric communities were the sole inventors of extractive metallurgy, which then spread to other parts of the globe. This view was challenged by Renfrew (1969), who argued instead for multiple inventions of metallurgy in different independent centres across Eurasia, basing the argument largely on artefact typology and C14 dates.

Recent excavations at the 7000 year old Vinča culture site of Belovode in eastern Serbia (Fig. 1) and laboratory analysis of unearthed archaeometallurgical artefacts revealed the earliest evidence for copper smelting in Eurasia (Radivojević et al., 2010; Radivojević, 2013). At Pločnik, a Vinča culture settlement in the south of the country, the recovery of a well-contextualised tin bronze foil dated to c. 4650 BC has also suggested the presence of a very early but short-lived tin bronze making tradition in the Balkans, technologically linked to the early copper making by Vinča culture communities (Radivojević et al., 2013). Both discoveries reinforce the theory of multiple independent centres of metallurgy invention in Eurasia, with Iran and possibly the Iberian Peninsula as other likely contemporary metallurgy heartlands alongside the Balkans. Radivojević and Rehren (2016) have suggested that the evolutionary trajectory of copper metallurgy in this part of the world is connected to the knowledge of material properties of black and green manganese-rich copper minerals, which feature as raw materials for both copper and tin bronze making. Such knowledge emerged locally and was subsequently transmitted across the Balkans over the course of c. 2000 years, starting in the late seventh millennium and continuing into the mid to late fifth millennium BC (Radivojević and Rehren, 2016, p. 228).

Advances in pottery firing technology, which predates metallurgy in many parts of the world, have been proposed as precursors to the emergence of metal extraction (Wertime, 1964 and literature therein). According to this theory, metallurgists gained transferrable skills from potters, including the ability to reduce metal oxides (Wertime, 1964, pp. 1264–1266). This appears to be supported by archaeological evidence from the Balkans, where together with other forms of decoration (e.g. cinnabar, calcite, black-topped) dark-burnished and graphite-painted pottery (Fig. 2) are typical productions. The high firing temperatures of about 1000 °C or above and predominantly reducing atmosphere that were assumed to be necessary for the production of dark-burnished and graphite-painted decoration were taken to indicate that potters of the Late Neolithic Balkans already possessed a sophisticated understanding of pyrotechnology and the behaviour of naturally occurring inorganic materials at high temperatures (Gimbutas, 1976, pp. 173–176; Kaiser et al., 1986; Renfrew, 1969) by the time that metallurgy emerged, arguing that this knowledge and skill must have featured as a crucial prerequisite for the development of copper smelting technology.

Whilst this is an attractive and convenient interpretation, it has never been rigorously tested. In particular, an in-depth understanding of the technology involved in the production of dark-burnished and graphite-painted pottery in the Vinča Culture is lacking, especially in the context of most recently reported detailed technological studies on the emergence and evolution of metallurgy in this culture (e.g. Radivojević et al., 2010 and literature therein). The present paper addresses this gap in our knowledge by studying in detail a total of eighty-eight well-dated sherds that include a relevant selection of dark-burnished and graphite-painted sherds from Belovode and Pločnik (Table 1). Using X-ray powder diffraction (XRPD), scanning electron microscopy (SEM), thin section petrography and traditional macroscopic observations, the raw materials, pyrotechnological conditions and procedures required to produce the ceramics' distinctive decoration have been reconstructed. This has then been compared to the contemporary knowledge of copper production technology at the two studied sites and the Vinča culture in general in order to shed more light on the relationship between pottery making and the emergence of metallurgy, as well as the likelihood that the former was a key precursor to the latter in this part of the world. The study represents a significant contribution to the study of the late Neolithic and early Chalcolithic in the Balkans at a time of remarkable craftsmanship and pyrotechnological advancements by the communities in this region.

The Vinča culture is a Neolithic/Chalcolithic phenomenon that covered a vast area comprising parts of the northern and central Balkans, including North Macedonia, Serbia, northeast Bosnia, the Vojvodina, southern Transdanubia, the Banat, Oltenia, west Transylvania, and the lower Tisza valley (Fig. 1). This cultural phenomenon has been the subject of intense research, including key studies by Chapman (1977, 1981), Garašanin (1951, 1979), Jovanović (1971), Marić et al. (2016), Radivojević (2012), Renfrew (1970), Schier (1996) and Tasić et al. (2015). According to published absolute dates, the estimated duration of the Vinča culture ranges from c. 5350 to c. 4600 BC (Breunig, 1987; Ehrich and Bankoff, 1992; Schier, 2000; Whittle et al., 2016) and has been divided into different phases according to the observable stratigraphic sequences and typological developments within the ceramic material culture, with the most widely used divisions based on Garašanin (1951, 1979, 1993) and Milojčić (1949) (supplementary materials).

In this study we focus on the Gradac Phase, which starts at the beginning of the fifth millennium BC (Garašanin, 1994/95; Jovanović, 1993/1994, 2006; Schier, 1996; Whittle et al., 2016) and lasts for around a century in the northern part of the Vinča culture phenomenon (defined as Gradac I-II) or until its end in settlements of the Morava valley and its tributaries in central and south Serbia, termed as Gradac I, II and III (the end varies between c. 4600 and c. 4400 BC, see Radivojević and Grujić, 2018; Radivojević et al., forthcoming). The Gradac Phase has been broadly correlated with the expansion of metallurgy and mining activities in the Vinča culture, particulary at the site of Rudna Glava (Jovanović, 1980, 1993/1994), as well as Belovode and Pločnik (Radivojević and Kuzmanović-Cvetković, 2014).

According to Jovanović (1993/1994) developments corresponding to the appearance of the Gradac phase clearly denote significant social changes at the time that he linked with the beginning of the Chalcolithic period in the Vinča culture and the entire Balkans. Garašanin (1994/95) also noted that this phase in the broader cultural and geographic context of this region belongs to a clearly distinguished and important period. Therefore, the influence which the appearance of metal had played within the Vinča culture is an important matter of debate, as much as its origin. To address themes such as invention, innovation and cultural change, a closer look into the material culture with a technological approach that includes archaeometric analysis seems to be particularly important. The study of the sites of Belovode and Pločnik offers the opportunity to approach these themes by investigating the archaeological records from two sites that gave important evidence for metallurgical activities and accordingly could have played a major role in the invention and adoption of metallurgy in Europe.

The site of Belovode is situated on a plateau located close to the village of Veliko Laole, c. 140 km southeast of Belgrade (Fig. 1). It has yielded the earliest known evidence for copper smelting in the world, dated at around 5000 BC (Radivojević et al., 2010). Pločnik in southern Serbia lies on a fertile floodplain on the left bank of the Toplica River (Fig. 1). It yielded the world's earliest known tin-bronze object, dated to approximately 4650 BC, alongside more than 40 massive copper implements (e.g. Radivojević et al., 2013; Radivojević and Kuzmanović-Cvetković, 2014). In both sites abundant pottery finds were unearthed, including dark-burnished pottery (Fig. 2) coming from different features recognised during the excavations (e.g. dwellings, pits) that belong to building horizons corresponding to different Vinča culture phases (Radivojević et al., forthcoming). In Pločnik graphite-painted sherds emerge in the Gradac phase concurrently with metallurgy, as seen in the case of a copper chisel from Trench 14 that is dated to the context associated with 5040–4860 BC (95% probability, Radivojević and Kuzmanović-Cvetković, 2014, p. 18).

Dark-burnished pottery, also known as black-burnished ware, is a pottery tradition with a widespread distribution during the Late Neolithic across the Balkans (Bonga, 2013, pp. 133–178; Chapman, 2006, 2007, p. 296; Holmberg, 1964). According to Garašanin (1954), this pottery type may have originated in Anatolia due to the finds of aesthetically similar pottery, however their technological link to the Balkan examples has never been thoroughly investigated.

Nevertheless, other scholars argue that it could have evolved independently in the Balkans (Chapman, 2006; Childe, 1936/1937, p. 29) based on a few convincing arguments. Dark-burnished pottery was one of the main features of Vinča material culture and is found from its earliest development (c. 5350 BC). Also, its colour and brightness well match the Neolithic Balkan visual identity based on striking and dark colours (Chapman, 2006, 2007). Besides pottery, other examples for this aesthetical preference include black and green ores used for copper smelting, and obsidian (Radivojević and Rehren, 2016).

The distinctive black or dark grey decoration of the Balkan dark-burnished pottery could have been produced in several different ways including iron reduction, the application of manganese, the deposition of carbon as soot, or painting with graphite pigment. The iron reduction technique fires iron-rich clay above 500 °C under reducing conditions (Cuomo di Caprio, 2007, p. 121; Jones, 1986, p. 762; Maritan, 2004; Noll, 1991, p. 121). A reducing atmosphere is achieved during firing when little or no ‘free oxygen’ is available due to restricted air supply or the addition of excess fuel. In this situation, the iron in the clay is reduced to ‘ferrous’ minerals such as magnetite (Fe₃O₄), and carbonised amorphous organic matter in the clay is not burnt off, giving the pottery a grey or black colour. Manganese black decoration is formed by the presence of manganese-rich mineral phases such as pyrolusite which are applied to pottery as a pigment or within a clay-rich slip, then fired in oxidising conditions (Jones, 1986, p. 762; Noll, 1991, p. 140; Spataro, 2019).

Carbon black decoration is typically produced by adding organic material and firing under reducing conditions, resulting in the formation of a layer of charcoal or soot (Jones, 1986, pp. 763–764; Letsch and Noll, 1978, 1983; Noll, 1991, p. 175). A typical method involves ‘smudging’ (Jones, 1986, pp. 763–764), that is the deposition of carbon on the surface of a vessel and within open pores during the firing process, for example by smothering the pots with fine-textured fuel at the end of the firing. The coating is composed of a very fine crystalline or amorphous carbon (Jones, 1986, p. 763) producing a shiny ‘Glanzkohlenstoff’ (lustrous carbon) finish (Letsch and Noll, 1978, 1983). Significant technological skill is required to produce carbon black as timing is crucial and it is essential to maintain reducing conditions in order for the coating not to be burnt off. Letsch and Noll (1978) argued that the black finish on Neolithic and Bronze Age pottery from the Balkans, Anatolia, the Near East and Egypt is due to the deposition of carbon primarily from the smudging, but also from the organic matter contained within the clay body.

Painting pottery with graphite pigment is another method of achieving a highly reflective black surface finish (Jones, 1986, p. 768). Vessels with geometric patterns painted in graphite are found across the fifth millennium BC Balkans (e.g. Gaul, 1948, pp. 98–99; Leshtakov, 2005; Martinon, 2017; Todorova, 1986, p. 107). The earliest documented use of graphite decoration is considered to come from Promachon-Topolnica (Fig. 1) in the Struma valley and is dated to the beginning of the fifth millennium (Vajsov, 2007). Within the Vinča culture it appears for the first time during the Gradac phase (Perić, 2006, p. 238). Graphite is a crystalline form of carbon that occurs naturally in highly metamorphic rocks such as marble, schist and especially gneiss. It was ground to a fine powder, mixed with water and perhaps clay, then applied, often onto a burnished surface. The reduction during the firing should be well controlled to preserve the graphite layer (Kreiter et al., 2014).

It has been suggested that the use of graphite decoration on pottery was closely related to the emergence of early metal production. Its light-reflective qualities produce a metallic sheen that may have been aesthetically appealing to prehistoric communities (Todorova, 1981). The acquisition of graphite would have required the participation in specialist trade networks comparable to those required for copper exploitation (e.g. Leshtakov, 2005; Radivojević and Grujić, 2018). Another link that has been proposed, as will be discussed below, is that the high temperatures necessary for copper metallurgy (around and exceeding c. 1100 °C) could also have been required to produce graphite-painted pottery (Renfrew, 1969).

Noteworthy, the nature of the relationship between the emergence during the Gradac phase of the Vinča Culture of graphite-painted pottery and extractive metallurgy has never been properly investigated.

The earliest investigation of the pyrotechnological link between pottery and metallurgy in the Balkans was carried out by Frierman (1969), who analysed a late fifth millennium BC dark-burnished sherd decorated with graphite from the site of Karanovo in Bulgaria (Karanovo VI) by determining its fusion point via thermal analysis. He estimated that the sample had been fired to a temperature around 1050 °C in a strongly reducing atmosphere. The latter was beneficial for graphite application, since under oxidising conditions graphite burns off above c. 725 °C. Frierman (1969) therefore suggested that firing took place in a kiln, given the high temperature and prolonged period of reduction required to produce this type of pottery. This finding was taken forward by Renfrew (1969, p. 38) who suggested that “refractory technology in the south-east European Chalcolithic had evolved sufficiently in the firing of pottery to provide the conditions required for smelting and casting of copper”. However, a few years later Kingery and Frierman (1974) re-fired the same sherd at 700, 800, 900 and 1000 °C in reducing conditions and concluded that it had in fact been subjected to a maximum temperature of <800 °C, and possibly as low as 700 °C.

Kaiser et al. (1986) studied the firing temperature of dark-burnished pottery and other pottery types from the Vinča culture sites of Selevac and Gomolava in Serbia via thermal expansion (also Kaiser and Lucius, 1989) and SEM to document the vitrification microstructure. This indicated that the ceramics they studied were variously fired between 850 and 1000 °C under oxygen-poor conditions. Despite this variability, the authors concluded that potters of the western Balkans were routinely capable of achieving temperatures of 1000 °C under reducing atmospheres, and that this pointed to a sophisticated knowledge of the firing process, including managing the required resources of labour, fuel and time. Since the pottery came from different contexts at these two relatively distant sites (c. 100 km), it may be inferred that this knowledge was widely shared between Vinča culture communities at the time and could have been transferred to craftspeople who specialised in other pyrotechnologies, such as the smelting of copper metal.

Other studies on the firing of dark-burnished and graphite-painted pottery from the Balkans and Greece include those by Gardner (1978, 1979, 2003), Goleanu et al. (2005), Maniatis and Tite (1981), Perišić et al. (2016), Spataro (2014, 2017, 2018) and Yiouni (1995, 2000, 2001). Among these, Perišić et al. and Spataro focused especially on Vinča pottery. Perišić et al. (2016) analysed ten samples from Pločnik, but only a few were dark-burnished, and their typology and chronology were not contextually secure. The research of Spataro (2018) includes the materials from the eponymous site of Vinča Belo Brdo, originating from contexts excavated between 1930 and 1936 by Miloje Vasić, which have no direct association with metal artefacts from this site. All these studies applied a wide range of techniques including thin section petrography, SEM, re-firing tests, FTIR, XRPD and thermo-analytical studies. These investigations revealed that firing temperatures were highly variable, and unlike the findings of Frierman (1969) and Kaiser et al. (1986), did not appear to have exceeded 900 °C. Gardner (1978, 2003, p. 289) observed that graphite-painted vessels from Phases III from the site of Sitagroi in Greece (Fig. 1) have a red core, suggesting that the firing process involved two steps. This may have included an initial firing step under oxidising conditions below 700 °C, followed by a second smoky reduction phase.

Thus, considerable uncertainty surrounds the topic of Late Neolithic/Chalcolithic ceramic pyrotechnology in the Balkans, particularly the conditions required to achieve dark-burnished and graphite-painted decorations and their role in the inception of early metallurgy. It appears that too much emphasis has been placed on firing temperature and not enough attention has been given to other pyrotechnological parameters such as the redox conditions and length of firing. The former is of crucial importance to the process of smelting copper (Gardner, 1979, pp. 20–21; Rehren, 1997), as a reducing environment is necessary for the formation of metallic copper, the chemical change starting at temperatures as low as 700 °C, whilst a more oxidising environment and a rise in temperatures up to the melting point of pure copper at 1083 °C are required to initiate the physical change from solid to liquid metal (Pollard et al., 1991; Radivojević et al., 2010).