Introduction of the potter’s wheel as a reflection of social and economic changes during the La Tène period in Central Europe

Abstract

This study focuses on the introduction of pottery-forming methods employing rotational motion in relation to social and economic conditions and their transformations during the La Tène period in Central Europe. It explores the diversity of technological practices on a broader geographical scale in several regions of the Czech Republic with various demographic, social, and environmental conditions during this period. The study is based on the idea that a technological process is a cultural trait whose adoption is the result of a cultural selection. These interactions are facilitated by the performances of the technological process and its products. The technological analysis relies on a recently developed quantitative analytical technique based on calculating the orientation of components of the ceramic body supplemented by qualitative classification of diagnostic features observed on X-ray images and CT reconstructions. By applying the methodology to an extensive collection of pottery samples, we have obtained a robust picture of the adoption and spread of different variants of the application of rotational motion. Based on this evidence, we proposed evolutionary scenarios that show the unique interplay of the performances of the individual variants of this general innovative idea with specific local socio-cultural conditions.

Introduction

Innovations in pottery technology that emerged in the La Tène period in Central Europe reflect the most notable changes in pottery production since its inception in Neolithic societies. These innovations have contributed to the considerable technological diversity observed during this period. The diverse technological practices reflect the development of complex socio-economic relationships within the pottery production system and indirectly indicate the dynamic evolution of the structure of the La Tène society (Arnold and Gibson 1995; Schönfelder 2009; Buchsenschutz et al. 2012). The potter’s wheel or, more precisely, the use of rotational motion in pottery forming was one of the principal innovations. This innovation is particularly important for studying La Tène society due to its specific requirements and performances (Roux and Corbetta 1989; Roux 2019a).

The first wheel-made pottery in Central Europe appeared during the transition between the Hallstatt and La Tène periods. The beginning of the La Tène period (the LT A phase 480/460–390/375 BC) was characterised by high social stratification, accompanied by strong elite manifestations such as Mediterranean imports, fortified settlements, and elaborate burials under mounds (Fig. 1). This particularly applied to the central regions of Northeastern France, Southern Germany, and the southern and western parts of Bohemia. On the other hand, there is no evidence for the concentration of large-scale craft production (e.g. Frankenstein and Rowlands 1978; Kristiansen 1998; Bintliff 1984; Nash 1985; Dietler 1989, 1995; Brun 1995; Pare 1991; Pauli 1984). The transition between the LT A and LT B phases was accompanied by significant social changes, including the discontinuity of burial practices and settlement areas in most known cases (Venclová 2013a; Dreslerová et al. 2022). The entire subsequent period of LT B-C1 (390/375–190/175 BC) was characterised by social decentralisation and the absence of distinctive elite features (Fig. 1). Lowland agricultural settlements predominated. It is difficult to identify them on the basis of the poorly distinguishable settlement pottery, especially in the LT B phase (Rulf and Salač 1995; Salač and Kubálek 2015; Dreslerová et al. 2022). The deceased were buried with more or less standardised grave goods in flat, predominantly inhumation cemeteries. During the LT C phase, in addition to the common agricultural settlements, large central lowland agglomerations with evidence for specialised production and long-distance contacts begin to appear (e.g. Büchsenschütz 1995; Collis 1995; Cumberpatch 1995; Salač 1996; Augstein 2006; Kaenel 2006; Salač 2011a, 2014; Trebsche 2020). Simultaneously, the settlement network became denser, expanding into previously unfavourable peripheral areas (Waldhauser 1985; Danielisová et al. 2019). Further significant social changes are evident in the transition between LT C1 and LT C2 (Fig. 1). These changes are associated, among other things, with changes in burial rituals, which resulted in a disappearance of burial evidence from the beginning of LT C2 onwards. The LT C2-D1 period (190/175–50/30 BC) is also associated with further diversification of settlement forms, indicating an increase in economic interconnectedness within society and a more developed division of labour (Renfrew 1974; Crumley 1987, 1995a, b; Büchsenschütz 1995; Thurston 2009). Fortified centres known as oppida became a significant phenomenon during this period. The oppida partly take over the functions of the open lowland settlement agglomerations, which mainly engaged in long-distance exchange and specialised production (Collis 1984, 1995; Brun 1995; Büchsenschütz 1995; Crumley 1995a; Wells 1995; Salač 1996; Venclová 2002; Augstein 2006; Danielisová 2011).

Fig. 1

Schematic outline of the dominant trends in the main archaeologically documented phenomena during the La Tène period in Central Europe (after Waldhauser 2001, modified and complemented)

This study was carried out in an attempt to understand the process of technological innovation. This relies on the idea that the technological process is an amalgam of cultural traits—a ‘recipe’ which represents a unit of cultural transmission, i.e. a mental phenomenon that one acquires through teaching and learning (Lyman and O’Brien 2003; Mesoudi and O’Brien 2008). The spread of locally produced wheel-made pottery in Central Europe suggests the rapid acquisition of complex and discontinuous technological behaviour. The transmission of such a novelty is facilitated by learning biases, through which learners non-randomly adopt a new cultural variant on the basis of its interactions with people and the environment (content or direct biases), or the adoption arises from the learning context (context or indirect biases) (Boyd and Richerson 1985; Henrich and McElreath 2003, 2007). In such a transmission, the spread of a new cultural variant is usually the result of a selection process.¹ For selection to take effect, it is essential that at least some of the transmitted variations exhibit differential interactions (i.e. result in differences in fitness). These interactions are facilitated by the performance of the technological process as well as the performance of its products (Braun 1983; Schiffer and Skibo 1987; Lemonnier 1992; O’Brien et al. 1994; Fitzhugh 2001; Skibo and Schiffer 2001, 2008; Schiffer 2004). The performance characteristics of a culturally transmitted variant that affect its permanency and transmission relative to other similar phenomena define its ‘cultural fitness’ (Durham 1991; Mesoudi 2010). Performances not only are related to the technical and economic aspects but also facilitate social and symbolic interactions of an object or process (e.g. Schiffer and Skibo 1987; Lemonnier 1992; O’Brien et al. 1994; Fitzhugh 2001; Skibo and Schiffer 2001, 2008; Schiffer 2004).

The concept of performance is particularly interesting when considering the use of rotational motion for pottery forming because this general technological idea can be practiced in various ways. The different ways of applying the concept imply the different performance characteristics of the respective methods (skill, production, and tool requirements) with the consequences for their potential to be transmitted within the population. In this respect, the distinction among methods based on the contribution of rotational motion in the forming sequence is crucial (Henrickson 1991; Courty and Roux 1995; Berg 2007, 2008a; Roux 2010; Thér et al. 2017). We define three levels that simplify the diversity of the use of rotational motion but represent critical differences in the performance of forming methods:

• (a) The application of rotational movement to even the surface and correct the shape of the vessel (wheel finishing). The technique requires specific skills, but the time required to learn these skills is similar to the requirements of the respective hand-building techniques, as minor transformations of the vessel do not require the significant contribution of rotational kinetic energy (RKE) in their forming. The time requirements needed to complete the vessel are, in principle, comparable to hand-building methods without the application of rotational motion (Thér et al. 2015b).

• (b) The use of RKE to shape and thin the walls of the vessel (wheel shaping). The technique requires means to ensure effective utilization of RKE and the development of basic skills related to handling and controlling the transformation. These skills are attainable by apprenticeship lasting substantially longer than the respective hand-building techniques, but the entire forming process has little potential to be significantly more time-efficient compared to the respective hand-building techniques that are employed without the contribution of rotational energy (Foster 1959; Nicklin 1971; Thér et al. 2015b).

• (c) Using RKE throughout the forming process (wheel throwing). This technique is significantly more efficient than wheel finishing and shaping in terms of the time needed to complete the vessel. However, mastering this technique requires a radical change in skills, with greater demands on their acquisition compared to the previous variants (Roux and Corbetta 1989; Roux 2019a).

It can be asserted with a high degree of probability that none of the basic variants was invented in Central Europe. They came as a ready-made concept available in the ancient Mediterranean area and could be adapted to local technical and economic conditions. The question to be addressed is the degree of discontinuity of the individual variants in a given context. In accordance with Roux (2010), by discontinuous change, we understand the introduction of a new physical principle in the technological process. Roux finds two discontinuous changes that gave rise to new technological lines alongside the traditional ones in the evolution of techniques in the Southern Levant in the 5th millennium BC: the use of RKE to transform the walls of a vessel (analogous to the term wheel shaping used in this study) and the use of RKE to transform a clay mass into a vessel (wheel throwing). The discontinuity of the wheel finishing variant is questionable. The use of RKE, rather than the use of rotational motion itself, is considered by some authors to be the principal innovation, representing a more fundamental change than the rotation itself, which is often used to a limited extent also in hand-building methods (Childe 1954; Foster 1959; Roux 2010). However, in the studied context, it brings a significant change and comes hand in hand with a new shape and decorative expression of ceramics. The ancestral trait here is the use of muscular energy; the new trait is the rotation of the formed object around the vertical axis. The rotation is not necessarily fast and stable enough to produce RKE sufficient for effective use in shaping. Although such a technique is still primarily based on muscular energy and does not require the use of a true potter’s wheel, it does represent a significant change in muscular movements.

The performance characteristics of these methods, encompassing skill and tool requirements, efficiency, and visible effects on the finished product, play a pivotal role in replicating and disseminating these practices within society. However, cultural transmission is significantly affected by the nature of the selective environment (O’Brien and Shennan 2010). The value attributed to a particular performance is not universal and depends on the cultural, social, and economic context (the role of the cultural and social context in the innovation process is frequently emphasised in the anthropological literature, e.g. Haudricourt 1987; van der Leeuw and Torrence 1989; Gosselain 1992, 1998, Lemonnier 1992, 1993; Pfaffenberger 1992; Marcia-Anne Dobres and Hoffman 1994; Stark 1998; Maria-Anne Dobres 2000; Sillar and Tite 2000; Skibo and Schiffer 2008).

For instance, wheel throwing combines the time efficiency of production with a high learning cost. Consequently, it has the potential for adoption within a context (a) where production time is significant; i.e. it is advantageous to produce pots more quickly; i.e. the variant is transmitted within a community of potters, who depend on the production of pottery as its main economic activity and operate on an open market (cf. Childe 1954; Foster 1959; Nicklin 1971), and (b) society allows potters to engage intensively in a specialised activity and achieve the necessary skills in it, meaning there is a developed level of division of labour.

Discontinuous technological changes are usually linked to social changes (Creswell 1996; Roux 2010). The nature of the learning process limits the dynamics of change in technical practices and is the reason why they do not undergo jump changes unless the context of their application is changed, and they persist even under conditions where their use is disadvantageous from a cultural fitness perspective. Most of the technical behaviour is on the level of practical consciousness—individuals know how to act in particular situations without knowing how to or needing to articulate this ability. Practical consciousness represents complex and deeply rooted bonds between mind, body, and environment. It is learned without becoming an object of cognisance, thus an object of choice (Bourdieu 1977; Giddens 1984). The learned moves and gestures represent automatic processes that are difficult to change. It requires extended consistent practice to develop alternatives to them (Schneider and Fisk 1983). Consequently, they are considered to be one of the most conservative aspects of human behaviour (Caine and Caine 1994; Minar 2001; Minar and Crown 2001; Roux 2007). Especially, technological stages that do not leave apparent traces on the finished product and rely principally on specialised gestures rather than tools and shared information about clay sources and recipes are resistant to change. Above all, the conservatism of pottery-forming techniques is stressed (e.g. Nicklin 1971; Arnold 1985; Gosselain 2000; Wallaert-Pêtre 2001; Mayor 2011; Roux 2019b, 2020). An interdependence of the individual phases of the technological process is another aspect contributing to the stability. The change in one part of the chain usually influences other parts, and moreover, the dependences go beyond the technological chain itself, encompassing the ways of using the products in a given social context (cf. Skibo and Schiffer 2008). The need for complex change, bringing uncertainty and potential economic risk, often prevents the adoption of a single innovation (Cardew 1958; Quarcoo and Johnson 1968). When a system is reasonably well-adapted to a given environment, the behaviour of households in small-scale societies is motivated to minimise subsistence risk (cf., Winterhalder et al. 1999; Fitzhugh 2001).

Consequently, understanding the discontinuous changes in pottery-forming practices must be based on the perspective of the particular evolution of society. Communities of practice (social groups sharing technological traditions; Lave 1991; Lave and Wenger 1991; Wenger 2000) in different societies and different stages of their evolution may approach the learning of a craft differently, with important implications for the potential for technological change (Crown 2001; Wallaert-Pêtre 2001; Manzo et al. 2018). The universal principles of evolution are blended into the unique scenario of a specific historical situation (cf. Roux 2007, 2010). An analysis that draws on these universal principles and reflects them in unique historical circumstances can contribute to the understanding of processes of technological change.

Study of the use of rotational motion in pottery forming has a fruitful tradition, especially in Near East and Mediterranean archaeology. It has been focused on the development of techniques suitable for the identification of various means of the use of rotational devices and the conceptualisation of this innovation within the social and cultural context (Roux and Courty 1998; Knappett 1999, 2004, 2016; Roux 2003, 2009; Berg 2007, 2009; Roux and de Miroschedji 2009; Jeffra 2011, 2013; Crewe and Knappett 2012; Knappett and Leeuw 2014; Gauss et al. 2015; Gorogianni et al. 2016; Röckl and Jacobs 2016; Baldi and Roux 2016). This kind of approach to studying wheel-made pottery in Central Europe is exceptional. Many authors intuitively use technical characteristics or concepts of organisation of manufacturing process without grounding them in adequate analytical methodology, argumentation, or evidence. Petrographic and mineralogical analyses are usually focused on provenance without a theoretical framework for an appropriate interpretation of the data. However, a few exceptions must be mentioned. Gosden applied the most comprehensive approach to understanding the introduction of wheel-made pottery in the La Tène A period in Bohemia (Gosden 1983, 1985, 1987). Attention was also focused on the aspects of origin, spread, and introduction process of the first wheel-made pottery in the context of the Late Hallstatt and Early La Tène periods in Germany (Balzer 2004, 2009, 2015; Tappert 2015; comprehensively with further references de Groot et al. 2023) and the Early Iron Age Vekerzug culture in Hungary (Czifra et al. 2020). Other publications were focused on the later stages of the La Tène period (Cumberpatch and Pawlikowski 1988; Cumberpatch 1993a, 1993b, 1995). Apart from the provenance and visual characteristics of pottery, only minor attention was paid to the diversity of manufacturing processes and their significance for interpreting changes in pottery production in the context of the social and economic environment.

This study builds on our previous work on the introduction of the potter’s wheel in Central Europe in a small region in Eastern Bohemia (Chrudim region) (Thér et al. 2014, 2015a, b, 2017) and the densely occupied Brno region in Southern Moravia (Thér and Mangel 2021). Analysis of these regions demonstrated the differences in trajectories leading to establishment of the various uses of rotational motion in pottery forming in regions with different socio-economic conditions, suggesting the dependence of the innovation process on these conditions. The Brno region exhibited continuous development with an increasing proportion of wheel-thrown pottery, indicating technology transmission via a wide learning network of producers. In contrast, the Chrudim region provided evidence for disruptive technological changes, characterised by regression of the use of rotational movement in pottery forming and even a change in the direction of rotation (Thér and Mangel 2021). The presented study substantially enlarges the analytical dataset and explores the diversity of technological practices on a broader geographical scale in several regions of the Czech Republic with various demographic, social, and environmental conditions during the relevant period.

The main questions in the analysis are as follows:

• 1. How was rotational movement employed in the pottery-forming sequence?

• 2. Is the appearance of the innovation and its form associated with the attributes of social complexity and inequality?

• 3. How did these aspects change during the La Tène period?

• 4. How has the application of the innovation and its integration into pottery production varied in different regions?

Materials and method

Selected regions and sites

Archaeological sites from four geographical areas with evidence of La Tène settlement were selected to study the use of the potter’s wheel: central part of Central Bohemia, eastern edge of Central Bohemia, Northwest Bohemia, and Western Bohemia (Figs. 2 and 3; Supplementary 1). Settlement in these regions created geographically distinct areas, especially during the Ha D2–LT A period, but the boundaries of these regions became partially blurred in the later phases (Danielisová et al. 2019; Dreslerová et al. 2022). However, during the La Tène period, the regions also differed in certain cultural aspects and dynamics of development. The cultural boundaries between them can best be identified by regional differences in pottery production, which can be observed in Bohemia from the beginning of the LT C. Waldhauser (1996) defined three essential areas in this respect: (1) the Lower Vltava-Elbe area, which coincides with the area of Central and Eastern Bohemia and includes the first two sampled regions; (2) the foothills of the Ore Mountains area, which coincides with the region of NW Bohemia; and (3) the Otava-Middle Vltava area, which coincides with the southern Bohemian region (not included in the sampling). Recently, another area with distinctive pottery style has been defined (Metlička et al. 2022), which covers the sampled region of W Bohemia.

Fig. 2

Map of selected regions and sampled archaeological sites against the background of La Tène settlement in the Czech Republic (according to the materials prepared by B. Danielisová). 1, central part of Central Bohemia; 2, eastern edge of Central Bohemia; 3, Northwest Bohemia; and 4, Western Bohemia

Fig. 3

Characteristics of the sampled regions in the individual phases of the La Tène period

The central area of Central Bohemia can be broadly defined as the northwest and central part of the Central Bohemian region. Its northern parts belong to the flat and fertile terrains of the Central Bohemian Table and the Prague Plateau (Demek and Mackovčin 2006). Dense occupation during the La Tène period can mainly be observed in the basin of the lower Vltava, between its confluences with the Berounka to the south and the Elbe to the north and its smaller tributaries. The settlement density, the continuity of occupation, the presence of central sites, larger cemeteries, and evidence for interregional contacts suggest the central character of the region. The main source for understanding the LT A phase consists of numerous lowland settlements. The central fortified settlement of Závist, known primarily for the presence of an enclosed acropolis with stone podiums influenced by Mediterranean building traditions, held a specific position in the settlement structure of this period (Drda and Rybová 2008). The presence of elites is indicated by a higher concentration of Mediterranean imports or their imitations found at local Late Hallstatt and Early La Tène settlements (Trefný 2011, 2022; Trefný and Polišenský 2013, 2014). The concentration of elite burials in the area of the Vltava and Labe confluence also indicates the existence of a significant settlement agglomeration (Chytráček 2013; Mangel et al. 2023). For the following LT B-C1 period, similar to the rest of Bohemia, there is a reduction in evidence of settlements, partly due to the poorer recognizability of settlement areas from this period. In addition to a smaller number of known settlements, numerous burial areas, including several extensive necropolises, attest to the continued importance of the region (Hlava 2017; Sankot 2022). A significant increase in known lowland settlements occurs again in the LT C period. During the LT C2-D1 period, several oppida also emerged in the southern part of Central Bohemia (Drda and Rybová 1997). The northernmost of these was located on the southern edge of the discussed region, in the location of the older central hillfort at Závist. The existence of Central Bohemian oppida is associated with the control of the so-called Vltava route, which connected Central Bohemia with the Danubian region (Venclová 2013a, 2013b). Another important communication corridor, the Elbe route, also passed through the region connecting the Central Bohemian area with NW Bohemia, Saxony, and Central Germany to the west and with Eastern Bohemia and Moravia to the east. The presence of iron ore deposits located in the Říčany area in the eastern part of the sampled region may have also shaped the settlement of the region during the later stages of the La Tène period (Venclová 2008). Most of the sampled sites belong to common lowland settlements. A specific position among them is held by the Late Hallstatt to Early La Tène settlement of Prague-Pitkovice, from which several artefacts of Mediterranean origin and their imitations associated with the presence of elites have been found (Trefný and Polišenský 2008, 2013, 2014). Another exceptional site is the Závist hillfort, from which samples were taken both from the context of the Early La Tène acropolis and situations related to the existence of the oppidum during LT C2-D1 (Motyková-Šneidrová et al. 1978; Drda and Rybová 2008).

The area located on the eastern edge of Central Bohemia was analysed separately. It encompasses the fertile territory of the Central Elbe Table, with the central course of the Elbe River forming its backbone (Demek and Mackovčin 2006). During the La Tène period, we can observe continuous and uninterrupted settlement in this area, although its density appears to be slightly smaller compared to the central and northwest parts of Central Bohemia. The settlement of the Middle Elbe region seamlessly transitions into this region. Further east and northeast, it transforms into a specific settlement area of Eastern Bohemia, which has a more peripheral character (e.g. Mangel et al. 2013). The occupation from Ha D2 to LT A is represented by a network of lowland settlements and a few burials. The continuous utilisation of the area during the subsequent LT B-C1 period is reflected in several smaller necropolises (e.g. Čižmář and Valentová 1977; Sedláčková and Waldhauser 1987). In this period, no later than during LT B2-C1, a significant lowland settlement in Žehuň was established in the region. It maintained its central functions until the end of the La Tène period (Danielisová et al. 2018). In the LT C2-D1 period, when common lowland settlements again represent the primary source of archaeological knowledge, the role of the region is further strengthened by the emergence of a hilltop (possibly fortified?) settlement—a castellum located in Kolo near Týnec nad Labem, whose significance was undoubtedly supra-regional. The recently excavated settlement at the foot of this centre, located directly on the banks of the Labe River, yielded numerous coin and non-coin imports, e.g. segments of astragal belts and mirror handles from the Middle Danube region, fragments of metal vessels, and a republican ring with a glass gem (Militký and Beneš 2016; Beneš 2020). These indicate the involvement in long-distance contacts directed towards the Amber Road corridor, the Middle Danube, and the areas west of the Bohemian Basin. These two sites and likely the entire region along the middle course of the Elbe River benefited to some extent from the location in the east–west corridor of long-distance communication following the course of the Elbe River. The region is represented in the study by the settlement of Kanín, dated to LT A (Elšíková 2009; Megaw and Megaw 2010), and the settlement at the foot of castellum in Kolo near Týnec nad Labem, in LT D1 (Beneš 2015).

The area of NW Bohemia represents the territory between the Ore Mountains and the Ohře River. The La Tène settlement is primarily associated with the fertile lowland areas of the Most Basin and the Lower Ohře Table (Demek and Mackovčin 2006). Besides favourable agricultural conditions, the region also provided a wide range of mineral deposits (iron ores, ores of non-ferrous metals, and rocks for quern production), which were exploited (Venclová 2013a). The watercourses of the Ohře and Bílina rivers formed the backbone of intensively settled areas. The Elbe River forms the region’s natural northeastern border. A long-distance communication route connecting the Central Bohemian region with the territory of present-day Saxony and further into Central Germany ran parallel to the river (Salač 1998; Venclová 2013a, 2013b). Another essential communication route followed the course of the Ohře River towards the southwest (Salač and von Carnap-Bornheim 1994; Chytráček and Metlička 2004). The settlement of the LT A period is represented by relatively numerous and often continuous lowland settlements covering a broader period from Ha D2 to LT A. Some of these settlements were also inhabited in the subsequent phases until the end of the La Tène period (e.g. Soběsuky, Radovesice: Holodňák 1991; Waldhauser 1993; Salač and Kubálek 2015). Evidence of the presence of elites during the fifth century BC is not as significant as in Central or W Bohemia. However, it is indicated, for example, by the enclosed farmstead (Herrenhof) with Attic pottery finds in Droužkovice (Bouzek and Smrž 1994; Smrž 1996) or the late Hallstatt hillfort at Rubín situated on the southern border of NW Bohemia (Sankot 2009). The occupation continued with slightly reduced intensity in the following LT B-C1 period. The extent of settlement in this period is demonstrated by numerous necropolises, some of which are relatively extensive, in addition to lowland settlements (Zápotocký 1973; Waldhauser 1978, 1987; Holodňák 1988). The increase in the number of known settlements is then associated with LT C2-D1. No oppida or other hilltop centres have been documented in NW Bohemia during this period. However, the role of a central site was fulfilled by the extensive agglomeration in Lovosice (Salač 1991, 2012). Situated on the Elbe River, in a position beneficial for contact with other regions, the site provided evidence of a range of specialised production activities (especially the production of rotary querns and pottery). It was actively involved in trade of supra-regional significance. The sampling of NW Bohemian sites focused on situations from three settlement areas. Pohlody represents a typical lowland settlement with the LT A and C2-D1 phases (Vlčková 1991). The settlement in Soběsuky was continuously inhabited from Ha D to LT D1 (Salač and Kubálek 2015). Of particular significance among the NW Bohemian settlements is the production and distribution centre in Lovosice, from which samples were taken from the complex with one of the excavated pottery kilns (Salač 1991; Mangel and Thér 2018).

The last region subjected to sampling represents W Bohemia. The region is situated in the higher elevated and less agriculturally suitable areas of the Plzeň Uplands (Demek and Mackovčin 2006). The La Tène occupation is mainly associated with the area of the Plzeň Basin and its wider surroundings in the basins of the Mže, Radbuza, Úhlava, Úslava, Střela, and Berounka rivers. In the fifth century BC, W Bohemia belonged to the areas with strong manifestations of elites. The significance of the region, especially during the transition from Ha D3/LT A and in the earlier phase of LT A, is indicated by a series of elite burials equipped with harness components, two-wheeled wagons, and Mediterranean imports (Chytráček 1983, , 2000, 2012; Kozáková et al. 2016; Trefný 2017). Numerous hillforts and highland settlements dating from Ha D2-3 have been documented in the region, but their number significantly decreased during the transition to LT A (Chytráček and Metlička 2004). From the advanced phase of the LT A period, there is only sporadic evidence for settlements in hilltop locations and, at the end of this period, they ceased to exist. On the other hand, relatively numerous lowland settlements from this phase have been documented. The presence of elites and the settlement of the less agriculturally suitable environment in W Bohemia is probably related to the corridor of long-distance communication running parallel to the Berounka and Radbuza rivers, connecting the more northern and eastern regions of Bohemia with the territory of Bavaria and the Bavarian Danube (Chytráček and Metlička 2004). The presence of iron ore and ores of non-ferrous metals (Waldhauser and Klsák 1998; Chytráček and Metlička 2004; Trefný 2017) may have played a role in the occupation of the area. However, there is a significant shift in the transition between the LT A and LT B phases. The entire region transforms into a peripheral transit area. Currently, from the entire W Bohemia region, only three burials from two sites have been found, which can be roughly dated to LT B-C1 (Baštová 1986; Čechura 2013). This transformation is associated with a deterioration in the climate that occurred around 400 BC, and its impact may have been more pronounced in the reduction of stable settlements in less agriculturally suitable areas (Maise 1998; Chytráček and Metlička 2004; cf. Dreslerová et al. 2022). An increase in stable occupation, represented by about 50 lowland settlements, can be observed again during LT C and LT D1 (Waldhauser and Klsák 1998; Řezáč 2004; Metlička et al. 2022). For analysis, ceramics from the LT A period (up to LT A/B1) from settlements in Líně, Mašovice, and Plzeň-Roudná were selected. The remaining samples came from collections associated with settlements of the LT C period, or specifically LT C2-D1.

Sampling

The sampling unit was a sunken hut. Archaeologically, sunken huts are manifested as roughly rectangular pits with evidence of postholes, characterised by a relatively abundant presence of pottery. The goal was to sample three sunken huts from one settlement phase at each site. In some sites, fewer than three sunken huts were available, so sampling had to be limited to fewer units or other types of archaeological features were selected. Sunken huts are considered to represent households/consumer units, i.e. groups of users who obtained ceramics based on specific conditions determined by their social and economic connections. Of course, we cannot consider ceramic assemblages from sunken huts as closed-find units representing the inventory of a given household. The assemblages may have diverse histories and formation dynamics, and it is necessary to consider that they were created over a long period and contain objects from other parts and phases of the settlement. Therefore, we consider the resulting statistics to be orientational, pointing only to notable trends in the composition of ceramic assemblages. The sporadic and selective occurrence of closed-find contexts containing a sufficient quantity of pottery within the entire spectrum of archaeologically excavated types of sites dated to the La Tène period in the given geographic context precludes basing systematic sampling on chronologically more unequivocal assemblages (Supplementary 1).

The development of the use of the potter’s wheel was traced throughout the La Tène period (480/460–50/30 BC). Three primary phases corresponding to three successive chronological periods were compared, reflecting the aforementioned significant changes in the archaeological record, which also indicate major changes in society within the period: the LT A (480/460–390/375 BC), LT B-C1 (390/375–190/175 BC), and LT C2-D1 (190/175–50/30 BC).

The analysis was conducted in two phases:

• (a) The initial evaluation phase was based on macroscopic analysis, wherein we focused solely on the basal characteristics of the phenomena relevant to the objectives of the study as specified in the ‘Macroscopic analysis’ section below. From the dataset, ceramics were selected for more detailed description and further analysis. The number of individual ceramic objects in the selected and recorded collections, representing the given phase at the site, varies between 468 and 2992. A total of 42,664 individual ceramic objects were recorded for the three regions (Supplementary 2).

• (b) The detailed evaluation of the selection of samples consisted of structural analysis based on thin-section microscopy, X-ray imaging, and computed tomography. A total of 578 samples were chosen for the analysis to determine the forming techniques, in order to obtain at least ten samples of fine-grained wheel-made pottery per analytical unit (one chronological phase per archaeological site) for this analysis. Representatives of other ceramic classes were also selectively sampled. The method is described in the ‘Microscopic analysis, X-ray imaging, and computed tomography’ section below.

If we found a site assemblage that was (a) strongly heterogeneous in terms of the presence of chronologically diagnostic artefacts or (b) it was impossible to sample sunken huts or alternative archaeological features due to poor excavation documentation and restricted access to the artefacts, but it was necessary to sample such an assemblage because there was no alternative site available in the region or the site played a specific role in our research plans, we omitted the first phase of the analysis and only took samples for the second phase.

Macroscopic analysis

Two basic variables were observed macroscopically: (a) ceramic pastes and (b) surface topography and morphology.

• (a) The classification of ceramic pastes is purposefully divided into two general categories: fine and coarse ceramics. Fine pottery represents ceramics manufactured from materials without a macroscopically observable aplastic component or with an admixture of fine-grained sands in fractions up to 1 mm. Occasionally, larger grains may appear in the mass (Fig. 4A-D). Coarse pottery represents other pottery that does not meet the criteria for fine pottery (Fig. 4E-H). Within the category of coarse pottery, we further distinguish pottery tempered with graphite-rich rocks—graphite pottery. In terms of the organisation of pottery production, this material is of interest due to its limited sources in the Czech Republic, making the production of graphite pottery dependent on a broader distribution network of raw materials (Waldhauser 1992; Hlava 2008; Mangel and Danielisová 2014). This fact and the higher proportion of wheel-made ware within the graphite pottery production compared to other coarse ware suggest an association of graphite pottery with specialised production (Trebsche 2011; Thér and Mangel 2014 with further references).

  In macroscopic analysis, these categories can be quickly and reliably distinguished, as each possesses distinct physical properties that may require different approaches to shaping and subsequent stages of the technological process. La Tène wheel-made pottery is primarily associated with fine tableware, creating a more pronounced distinction between these material classes compared to other periods. The most notable dichotomy is observed in the LT B2–LT D1 phases. Late Hallstatt and Early La Tène ceramics exhibit less discontinuity between pastes regarding particle size distribution and the proportion of non-plastic components. Consequently, this results in less clear boundaries between fine and coarse ceramics, leading to more significant uncertainty during macroscopic differentiation. Nonetheless, in the context of the La Tène period, we generally consider this distinction significant in terms of the living culture, i.e. reflecting the potters’ perception of the distinctiveness of pottery masses.

  Fig. 4

  

  Variability in fine (A–D) and coarse pottery (E–H) textures

• (b) Surface topography in the context of pottery analysis means the phenomena that characterise the overall shape of the surface, especially surface regularity and discontinuity (Roux 2019a). Morphology perceives hollows and protrusions on the surface as separate objects and studies their shape and distribution. Two phenomena were essential: the evenness of the surface in terms of the presence of local deviations from the ideal shape² of the object and morphological traces (traction lines, grooves) indicating the application of continuous pressure on the vessel wall facilitated by rotational motion. Based on these characteristics, the following three categories were defined (see the overview of the technological categories identified in each phase of the analysis in Fig. 5):

  Fig. 5

  

  Diagram of the links among the technological categories related to pottery forming identified in each phase of the analysis

• Hand-built pottery (HB): There is an absence of features indicating the application of continuous pressure on the vessel wall facilitated by rotational motion. A rotational motion was not used in any forming phase (Fig. 6A).

• Wheel-made pottery (WM): Based on macroscopic observations, rotational motion was used in at least the final forming phase (Fig. 6C). However, based on this observation, we cannot determine the extent to which rotational motion was utilised in most cases. Consequently, all the basic variants—wheel finishing, shaping, and throwing—are possible. The use of rotational motion is indicated by horizontal to diagonal traction striations, grooves, or facets created by lifting the vessel wall during the throwing and evening of the surface, smoothing, and polishing with a potter’s blade (Fig. 7). These traces can occur without rotation, which is why their straightness (consistent direction) and uniformity in the case of multiple traces are important. The term wheel-made implies the use of a potter’s wheel. However, in the given analysis phase, we cannot provide any more specific information about how rotational motion was utilised. We cannot determine what kind of rotational device was used. The more accurate term would be pottery produced using rotational motion. Nevertheless, with this terminological clarification, we choose to use the established term wheel-made for better text readability.

  Fig. 6

  

  Examples of hand-built (HB) (A), unspecified (U) (B), and wheel-made (WM) (C) bowls

  Fig. 7

  

  Diagnostic features for wheel-made (WM) pottery. A, B Facets resulting from smoothing by a hard tool on a fast-rotating device; C drag marks resulting from a frictional action against the clay paste causing the dragging of non-plastic grains over the surface; D striations as a result of evening the surface with the fingers; E stacked rilling; F concentric undulations resulting from the lifting of the clay when shaping a vessel

• Unspecified pottery (U): Based on macroscopic observations, it is impossible to determine whether rotational motion was used during the forming. In these cases, clear evidence of rotational motion (horizontal traction lines and grooves) is absent, but the surface is well-evened (Fig. 6B). Horizontal facets created by smoothing and polishing actions can be observed, but it is not possible to assess their directional consistency to determine the application of the rotational device unambiguously. Theoretically, the observed features can also be achieved without using rotational motion, albeit with a more significant investment of time in the final shaping phase. Therefore, we consider this phenomenon symptomatic of using rotational motion but not directly indicative of this technique. We approach this category more critically than most archaeologists studying La Tène pottery in Central Europe, who consider it evidence for the use of a potter’s wheel. Nonetheless, we still perceive this category as a technological phenomenon that indirectly points to the use of rotational motion in forming. Consequently, this category does not imply a purely negative absence of characteristics that could determine the technology. However, it is essential to note that smaller fragments are also classified in this way, where the size of the preserved fragment surface primarily limits the decision on the application of rotational motion. We can assume that, if a larger surface area were preserved, it would be possible to make a definitive decision regarding the use of rotational motion for a more significant number of smaller fragments.

Further macroscopically observable diagnostic features that would allow for the determination of more specific details regarding the employed techniques were not examined. These features can only be observed in a small proportion of pottery fragments, and the results would not apply to an overall statistical characterisation of the assemblages. Given the detailed structural analysis in the study’s second phase, we do not consider their use beneficial.

Microscopic analysis, X-ray imaging, and computed tomography

The analysis relies on a recently developed quantitative analytical technique based on a single diagnostic feature: the orientation of components within the ceramic body. This feature can be observed and measured on each ceramic fragment, providing a consistent basis for analysis (Thér 2016; Thér and Toms 2016, 2021). This quantitative approach for analysing the orientation enhances the analysis of preferred orientation by delineating intervals of orientation variability for each specific forming technique and their combinations. To validate the theoretical assumptions regarding the relationship between individual forming methods and orientation patterns (Rye 1981; Carr 1990; Courty and Roux 1995; Pierret 1995; Whitbread 1996; Middleton 2005; Livingstone Smith 2007; Berg 2008), we analysed multiple experimental collections.

We applied this methodology to estimate the contribution of rotational movement in pottery forming and to identify the forming techniques used to form the initial segments or roughout of the vessel. For this purpose, tangential sections are the most significant (Thér and Toms 2016). In this study, we rely only on these sections to minimise the destructive impact of the analysis on the archaeological ceramics. The entire area of the thin section was recorded at a magnification of × 200 using a Keyence VHX6000 digital microscope. The resultant images have a resolution of 1.11 μm. The analysis followed a published methodology (Thér, 2016; Thér and Toms 2016). The components of the ceramic materials were extracted using automatic area measurement tools available in the Keyence VHX6000 measurement software. The range of threshold values chosen to separate objects was based primarily on colour saturation, which shows the best results for the thin sections with uneven thickness (resulting in uneven brightness of the captured image).

Two basic measurements were selected to express the orientation of the extracted object: (a) the mean direction—average orientation of objects based on determination of the angle between the major axis of the object’s secondary moment and the horizontal axis (0–180°, clockwise), and (b) the circular standard deviation (CSD)—the dispersion of the values from the average value (Fisher 1993; Mardia and Jupp 2000). The values were plotted in a polar diagram (for the explanation of a polar diagram see Thér and Toms 2016). To make the plots more readable, samples with a CSD of more than 50° (which usually means a random orientation) are plotted on a line defining a CSD of 50°.

After the current refinement of the analysis (using only highly elongated voids and particles, weighing the length of the objects), the delimited zone of the typical values for wheel-thrown pottery is the mean directions deviating from the horizontal axis by 10–45° and a CSD up to 20°. We consider these intervals to be a focal area of the orientation typical for wheel throwing, but this does not mean that every sample located outside this zone was not wheel-thrown and vice versa. The fundamental limitation of this method is that it provides information on the orientation only within a limited surface area of the sample. Particularly in cases of combined forming methods, such a small area may not accurately represent the overall structure of the material. Also, the deviations from optimal process performance (insufficient skills of the manufacturer, complicated shapes requiring further significant transformation of the object beyond the initial lifting of the basic shape) can cause deviations from typical values, as has been shown in the experimental samples (Thér and Toms 2016).

Therefore, we supplemented the analysis of thin sections with X-ray imaging and CT reconstruction of complete fragments (either one or the other imaging technique was employed for each sample with a measured orientation). These imaging techniques allowed us to verify whether the computed orientation corresponded to the overall character of the structure and further supplement the orientation analysis with the identification of additional structural phenomena, especially discontinuities caused by joining of the segments from which the vessel was assembled (Figs. 8 and 9).

Fig. 8

Diagnostic features observed on X-ray images. A, B Discontinuity of the structure in the transition between the base and lower body of the vessel typical for WF and WS; C Omnidirectional structure typical for DPP techniques; D Continuous structures in the transition between the base and lower body of the vessel typical for WT

Fig. 9

Diagnostic features observed on CT reconstructions. Structures are viewed on radial and non-planar sections defined in the plane parallel to the vessel surface. A Alternation of horizontal zones at the junctions between coils and omnidirectional to vertical zones at the cores of the original coils typical for TC; B Multiple coil joints typical for C; C Evidence for coil joints observed on WM pottery; D Discontinuity of the structure in the transition between the base and lower body of the vessel typical for WF and WS; E Continuous diagonal orientation typical for WT; F Random orientation observed on WF fine pottery typical for the use of DPP techniques; G Exceptional evidence for an arc-shaped joint between slabs; H Vertical orientation of the structure reflecting drawing movements during shaping the vessel

X-ray images were acquired using an X-ray generator using a lamp with a focal spot of 0.05 mm at a voltage of 120 kV and a current of 160 µA (obtained on the X-ray device Explorer X test 200–120/400 by Testima). Each CT reconstruction was based on 600 RTG images (acquired with the same settings as single X-ray images). CT reconstructions were created using the LometomArk software provided with the X-ray device mentioned above. The resolution of the resulting CT reconstructions varied depending on the size (ranging between 55 and 120 µm).

In combination with the results of macroscopic analysis, the following categories were differentiated:

• Wheel-finished pottery (WF): Based on macroscopic observation, a rotational motion was most probably used (WM and U categories of macroscopic observation). Nonetheless, the material structure does not exhibit significant transformation as a result of the use of RKE (Fig. 9C, F, H). This combination is characteristic for ceramics where rotational motion was employed solely for surface finishing and shape correction.

• Wheel-shaped pottery (WS): The structure of the clay mass shows significant transformation resulting from using RKE. The orientation is similar to that of WT. Nonetheless, we observed poorer alignment as a result of the transformation from the structure corresponding to the primary forming technique to the structure corresponding to the use of RKE. In addition, structural discontinuities can occur in the ceramic body. A particularly characteristic feature is the discontinuity between the base and the lower part of the vessel wall (Figs. 8A, B and 9D). Such features are typical of pottery, where rotational motion was employed during the shaping phase, leading to wall thinning and substantial changes in the overall vessel shape.

• Wheel-thrown pottery (WT): The alignment of the structure (strongly oriented diagonal structure in tangential section) and the absence of discontinuities in the structure indicate that the entire forming process was executed using RKE (Figs. 8D and 9E).

• Clockwise rotation (…-cw): Most of the WM production at the WS and WT levels was carried out with counterclockwise rotation. However, there are instances where clockwise rotation is observed (Fig. 8B). The ‘-cw’ suffix indicates these cases.

• Coiling (C): A horizontal orientation identifies the technique in the tangential sections and omnidirectional on the radial sections (inspected on CT reconstructions), influenced by how individual segments (coils) are incorporated into the vessel wall. We can expect residues of horizontal joints between the coils (Fig. 9B, C), although these joints may not always be preserved.

• Discontinuous perpendicular pressure techniques (DPP): This designation encompasses a broad category of techniques involving discontinuous perpendicular pressure on the vessel wall or the formed segment from which the vessel is assembled. This category includes techniques like pinching, slab-building, or pressing into a mould. All of these techniques exhibit a similar orientation of the structure, distinct from the coiling technique: omnidirectional in the tangential section and parallel to the wall margins in the radial section (Figs. 8C and 6F, G). The critical evidence lies in the presence or absence of joints between segments to differentiate between specific techniques within this category (Fig. 9G). However, this evidence is rare. Therefore, techniques within this category are generally classified as those employing discontinuous perpendicular pressure without further specification.

• Transformed coils (TC): From a structural perspective, this technique represents a hybrid of the previous two. It is characterised by thick coils that undergo significant transformation through discontinuous perpendicular pressure when incorporated into the vessel wall. This technique is not delineated in classification systems as a distinct forming technique (e.g. Roux 2019a). It appears as a variant of coiling. However, it is distinguished within this context due to its significant implications for the structure of the ceramic material, which is different from the use of thinner coils. The resulting structure typically retains a predominant horizontal orientation in the tangential section but with lower alignment. Alternation of horizontal zones at the junctions between coils and omnidirectional to vertical zones at the cores of the original coils can be observed (Fig. 9A). Consequently, in a thin section that captures only a small portion of the structure, this technique may be manifested in a relatively strong horizontal orientation and a random to vertical orientation. It is essential to observe structural discontinuities over larger surface areas of the samples (using X-ray or CT scanning). As a result of these phenomena, the use of this technique in the assemblages is underestimated.

Detailed descriptions of composite methods (e.g. WS-C, WF-DPP) can be confirmed only where these combinations can be documented. This is particularly characteristic of WF, where the structure in the wall core remains untransformed by rotational motion, fully corresponding to the primary forming technique.

Results

Trends that distinguish Central Bohemia sites³ from NW and W Bohemia in the proportion of fine pottery can be observed. In the former region, we observed an approximately 1/4–1/3 proportion of fine pottery in the assemblages of the first two phases, which decreased significantly in the last phase (Figs. 10 and 19A). This trend corresponds to the disappearance of HB fine pottery dated from the LT B2-C1 phase (Figs. 10 and 19B). In NW and W Bohemia, on the other hand, no clear trend can be observed, and the data obtained from individual sites cannot be averaged due to significant differences in the proportion of fine pottery among them (Fig. 10). These are determined by the presence of specific pottery traditions or by the specificity of the sites. In contrast to Central Bohemia, HB fine pottery did not disappear in the later phases of the La Tène period. However, even in this aspect, the picture in NW and W Bohemia is not uniform. On some sites, the proportion of HB pottery is marginal (SOB, LOV, UHE) while, in others, it forms a significant portion of the fine pottery (POH, NYR, VAL, OPL). In these cases, fine-grained ceramic material was used to produce goods that elsewhere would have been made from coarse-grained materials (e.g. large bowls with in-turned rims, storage jars, and large pots). In Central Bohemia, the fine-grained material is exclusively used for WM pottery during the LT C2-D1 phase and its proportion in production was relatively low compared to other phases and regions (Figs. 10 and 19B).

Fig. 10

Proportion of macroscopically observed categories within fine and coarse pottery at sampled archaeological sites. WM wheel-made pottery, U unspecified pottery, HB hand-built pottery, SUM total number of samples in the given category. Graphs show absolute frequencies

In the LT A phase, when pottery manufactured with the use of rotational motion first appeared, only two samples revealed diagnostic features typical for WT pottery. Both are from Praha-Pitkovice (PIT; Fig. 12B). Given that (a) these are isolated cases and (b) based solely on orientation with no evidence for structural discontinuities, the use of a combined forming methods cannot be conclusively ruled out, and therefore, definitive conclusions cannot be made regarding the appearance of the WT in this period, albeit sporadically. Nevertheless, it is interesting that exceptional evidence for pottery potentially made on a potter’s wheel appears precisely at a site with remarkable evidence for Mediterranean ceramic imports accompanied by other imports exhibiting long-distance contacts (Trefný 2011; Trefný and Polišenský 2014) and also with evidence for specialised production activities associated with metal working, indicating the supra-regional significance and central role of the site (Trefný and Polišenský 2019; Trefný et al. 2020). This leads to the hypothesis that the site represents one of the centres of innovation, benefiting from contact with advanced Mediterranean society and possessing sufficient resources to invest in costly innovations.

Regional differences in the extent of the use of rotational movement and specific characteristics of individual sites were observed for the LT A phase. While Central and NW Bohemia show that approximately half of the WM production was significantly transformed by the use of rotational motion (the proportion of WS in Central Bohemia without ZAV is around 48%, in KAN 42%, and in NW Bohemia 53%; Figs. 11A, 12, 14 and 15), in W Bohemia, the proportion is significantly lower (average 18%; Figs. 11A and 16).

In Central Bohemia, hillfort Závist (ZAV) represents an exception, where out of 18 samples of WM pottery, only one exhibited noticeable structural transformation due to the use of rotational motion (Fig. 12D). The hillfort represents a significant supra-regional central site with long-term development from the Ha D2 phase (Motyková et al. 1984; Drda and Rybová 2008). Its extensive fortified area with evidence for long-distance contacts and stone podiums built in the early LT A phase under the influence of southern building practices constitutes an exceptional phenomenon that can be understood as a Central Bohemian counterpart of the elite environment manifestations in Southwestern Bohemia, where it is manifested in the form of burials with two-wheeled wagons (Chytráček 2012), as well as in numerous hillforts (Chytráček and Metlička 2004; Chytráček et al. 2010). From this point of view, it is interesting to note the similarity of technological phenomena observed on pottery from Závist and W Bohemia. W Bohemian assemblages are technologically more similar to Závist than Central Bohemian sites. We are not able to demonstrate this interesting connection with additional elements of the material culture. Nonetheless, it can be supposed that these phenomena represent the same socio-cultural tradition, distinct from the typical Early La Tène expressions in Central Bohemia. In this case, this suggests that different dynamics of technological evolution in pottery production is associated with the manifestations of elites in W Bohemia during the Late Hallstatt and Early La Tène periods. The extent of the use of rotational movement lags significantly behind that in Central and NW Bohemia.

On some sites in Central Bohemia, we observed the presence of pottery manufactured using clockwise rotation in the LT A phase. In Libčice-Chýnov (LIB; Fig. 12C) and Kanín (KAN; Fig. 14A), this pottery accounted for approximately half of the WS pottery; in Praha-Pitkovice (PIT; Fig. 12B), it accounted for a quarter (Fig. 11B). Clockwise rotation associated with WS is also predominant in the Chrudim region in the LT A phase (Thér et al. 2017). It has not been observed in the other sampled regions and disappears in all the regions in the following period (LT B2-C1; Fig. 11A). The presence of clockwise rotation in the LT C2-D1 stage at oppidum Závist (ZAV) is unique among all the sampled regions (Fig. 13G).

Let us return to the proportion of different levels of the use of rotational motion. In the LT B2-C1 phase, we can already observe a clear presence of WT pottery. Central and W Bohemia show similarly lower proportions of WT pottery (21% in Central Bohemia, 25% in W Bohemia; Figs. 11A and 19D). These regions differ mainly in the proportions of WF and WS pottery. In W Bohemia, 39% of WM pottery does not exhibit any significant structural transformation due to the use of rotational movement (WF) while in Central Bohemia, WF was used only marginally (6%). In this phase, Soběsuky (SOB) exhibited an exceptional proportion of WT pottery (57%; Fig. 11A). Soběsuky is the only site in NW Bohemia from which samples from this chronological phase have been obtained. Therefore, we do not know whether the high proportion of WT pottery represents an anomaly or reflects more general trends in the region. In the final phase of the La Tène period, LT C2-D1, WT pottery predominates in the assemblages in Central Bohemia (73%; Figs. 11A and 13). The lowest proportion is found in W Bohemia (41%; Figs. 11A and 16F, G), with a significantly higher proportion of WF pottery (28%) compared to Central Bohemia (9%). Two high proportions of WT pottery are associated with important (supra)regional centres. Týnec nad Labem (TYN; 89% of fine WM pottery is WT; Fig. 11A, 14B) on the eastern edge of Central Bohemia represents a settlement at the foot of castellum in Kolo near Týnec nad Labem (Militký and Beneš 2016; Beneš 2020). Lovosice (LOV; 85% of fine WM pottery is WT; Figs. 11A and 15F) was part of a settlement agglomeration that played a specific role as a production and distribution centre (Salač 1991, 2012). However, the high proportion of WT pottery could be caused by the presence of a pottery kiln (Mangel and Thér 2018). Higher proportions of WT pottery have been repeatedly documented in relation to pottery kiln complexes in other cases (comprehensively Mangel and Thér 2017). The production waste of a specific production unit may influence the assemblage composition. In other cases where the evidence of production units in spatial proximity to the selected archaeological situations is absent, the assemblages most likely reflect the consumer spectrum of ceramics.

When evaluating the variability of primary forming techniques used in the production of HB and WM pottery formed by combined methods, it is necessary to critically consider the obtained data, as many assemblages focused on WM pottery do not provide representative data (especially in the later phases, where WT often predominates, and there is also a significant proportion of WS pottery without specification of the primary forming technique). However, basic trends and anomalies can be identified. In the LT A phase, Central Bohemian sites exhibit a mixture of primary techniques reflecting the variability of methods used by local potters (Fig. 11C). For example, Praha-Křeslice (KRE) and Praha-Pitkovice (PIT) are settlements about 3 km apart, yet only a quarter of analysed samples from Praha-Křeslice exhibit the use of C, while in Praha-Pitkovice, this is 67%. Once again, Závist represents an exception, where most samples show evidence for DPP techniques (89%). In this regard, Závist is similar to sites in W Bohemia, where DPP techniques significantly predominate at all the sampled LT A sites (with an average proportion of 80% of all samples, 93% of WM pottery samples). The proportion of C reaches its highest proportion in the easternmost sampled LT A assemblage from Kanín (KAN; 79%). The results are consistent with our findings in the Chrudim region, where C was the basic primary HB technique throughout the La Tène period (Thér et al. 2017). A substantial decline in DPP techniques was generally observed in the LT B2-C1 phase (overall proportion of C in all regions: LT A 36%, LT B2-C1 91%). The most contrasting situation is in W Bohemia, where the overall proportion of C (including the TC variant) increases from 20% (only 7% of U and WM pottery with identified primary technique) in the LT A phase to 94% in the LT B2-C1 phase. In contrast, there are no significant changes in the overall proportions observed in NW Bohemia (with approximately half of the pottery manifesting C techniques). However, in the LT C2-D1 phase, there are significant differences among the individual sites (Fig. 11C). In Lovosice (LOV) and Pohlody (POH), C predominates, whereas in Soběsuky (SOB), C is absent (the proportion of C gradually decreases during the La Tène period). Another exception in the LT C2-D1 phase is oppidum Závist (ZAV), where DPP pottery accounts for half of the sampled pottery with a specified primary technique.

Coarse pottery did not receive detailed attention in this study. Nevertheless, some findings interestingly complement the analysis of fine pottery. Apart from the Central Bohemian region, the use of rotational motion for coarse pottery is a marginal phenomenon in all phases of the La Tène period (Fig. 10). In Central Bohemia, however, we observed a significant increase in the proportion of pottery for which we are at least uncertain whether rotational motion was used in its production. For coarse pottery in Central Bohemia, a gradual increase in the regularising of the vessel surfaces can be observed, culminating in the LT C2-D1 phase. This trend primarily affects pots and bowls with in-turned rims, which are often roughened on the outer surface by scraping with a blade held perpendicularly to the surface in the leather-hard stage of drying (referred to as grating or deep roughening, which was undoubtedly performed away from the potter’s wheel). The inner surfaces of the vessels are well-regularised. Local production styles were observed even for this pottery, causing a significantly higher proportion of pottery with the uncertain use of rotational motion in some Central Bohemian sites during the LT B2-D1 phases than in other sites (Fig. 10).⁴ The difference in the technique/style of production at individual sites is also reflected in a significant average proportion of WM coarse pottery in Central Bohemian sites during the LT C2-D1 phase (5%). Additionally, a significant proportion of WT pottery (30%) within the sampled coarse WM pottery from the Central Bohemian was documented in this phase.

Finally, we considered a specific type of coarse pottery—graphite pottery. Due to its minimal representation in the recorded assemblages, we did not discuss graphite pottery at the level of individual sites but only in summary, excluding NW Bohemia, where graphite pottery was recorded only in isolated fragments. In contrast to the rest of the coarse pottery, graphite pottery shows a significant proportion of WM pottery from the LT B2-C1 phase onwards. Due to the size of the recorded graphite pottery assemblages, we did not comment on specific proportions. However, most graphite pottery in the later phases of the La Tène period was made using rotational motion, which contrasts significantly with the rest of the coarse pottery (Fig. 17). This conforms to the previous finding (e.g. Trebsche 2011; Thér and Mangel 2014) (Figs. 11, 12, 13, 14, 15, 16, 17, 18 and 19).

Fig. 11

The results of the second phase of the analysis for the sampled archaeological sites. A Proportion (expressed as a percentage) of the variations in the use of rotational motion within fine wheel-made (WM) pottery; B Proportion (expressed as a percentage) of the pottery manufactured with the use of clockwise (CW) and counterclockwise (CCW) rotation within fine wheel-made (WM) pottery; C Proportion (expressed as a percentage) of hand-building techniques within samples where such a technique has been identified (C coiling, TC transformed coils, DPP discontinuous perpendicular pressure techniques); D Total number of samples analyzed in the second phase of the analysis

Fig. 12

Polar diagrams of the mean direction and circular standard deviation (CSD) of the ceramic body’s components measured in tangential sections of the samples from archaeological sites in central part of Central Bohemia (LT A-C1)

Fig. 13

Polar diagrams of the mean direction and circular standard deviation (CSD) of the ceramic body’s components measured in tangential sections of the samples from archaeological sites in central part of Central Bohemia (LT C2-D1)

Fig. 14

Polar diagrams of the mean direction and circular standard deviation (CSD) of the ceramic body’s components measured in tangential sections of the samples from archaeological sites in eastern edge of Central Bohemia

Fig. 15

Polar diagrams of the mean direction and circular standard deviation (CSD) of the ceramic body’s components measured in tangential sections of the samples from archaeological sites in NW Bohemia

Fig. 16

Polar diagrams of the mean direction and circular standard deviation (CSD) of the ceramic body’s components measured in tangential sections of the samples from archaeological sites in W Bohemia

Fig. 17

Comparison of the development of the proportion of wheel-made (WM) pottery between coarse and graphite pottery

Fig. 18

Map showing the proportion of the variations in the use of rotational motion within fine wheel-made (WM) pottery at sampled archaeological sites. A The LT A period; B The LT B-C1 period. C The LT C2-D1 period

Fig. 19

Comparison of the development of the selected technological phenomena during the La Tène period between Central and W Bohemia. A Proportion of coarse and fine pottery; B Proportion of wheel-made pottery within fine pottery; C Proportion of wheel-made pottery within coarse pottery; D Proportion of the variations in the use of rotational motion; E Proportion of the variants of primary forming techniques within wheel-made pottery

Summary

LT A

• In the LT A phase, the use of rotational motion did not attain the level of WT in any of the regions, except for two samples. RKE was much more extensively used in pottery production in Central and NW Bohemia regions than in W Bohemia.

• There is considerable variability in the methods of pottery production observed at the individual sites in Central Bohemia that combine C and DPP techniques using rotational motion in both the clockwise and counterclockwise direction.

• The results from Závist are more similar to the W Bohemian sites than the Central Bohemia sites (predominantly WF and DPP techniques).

• Pottery from the sites on the boundary between Central and Eastern Bohemia has a strong representation of the phenomena found to be dominant in the earlier analysis of La Tène pottery from Eastern Bohemia (Chrudim region).

LT B2-C1

• Changes compared to the previous phase are substantial and, in some areas, discontinuous.

• In the LT B2-C1 phase, WT pottery regularly appeared in the assemblages of all regions. Its proportion within the WM pottery ranges around one-fourth, except for Soběsuky (SOB), where it corresponds to more than half of the assemblage. Soběsuky is the only site sampled in NW Bohemia, so that it cannot be determined whether this is a regional characteristic or specific to that particular site. In the assemblages from W Bohemia, most WM pottery does not exhibit signs of significant structural transformation by the use of rotational motion.

• Within the primary HB techniques of the LT B2-C1 phase, the dominance of C is evident in Central and W Bohemia. The HB techniques employed in the WM pottery production sequence are generally less variable than in the previous phase.

• Evidence for wheel rotation in the clockwise direction is absent in this phase. From this phase onwards, potter’s wheels or other rotational devices used in pottery production uniformly rotated counterclockwise, corresponding to the findings in previously analysed regions.

• One tradition from the diverse spectrum of methods of the LT A phase prevailed in the LT B2-C1 phase in Central Bohemia. In W Bohemia, the domination of the similar method represents a discontinuous change in primary forming techniques.

• Graphite pottery, in terms of the applied production methods, demonstrates a specific position within the production of coarse pottery, and it corresponds to a large proportion of WM production from the LT B-C1 phase onwards across all the regions (except for NW Bohemia, where it is practically absent).

LT C2-D1

• Changes are gradual compared to the previous phase.

• In the LT C2-D1 phase, most WM pottery from the Central Bohemian region is WT, while the proportion of WT pottery remains significantly lower on average in W Bohemia.

• NW Bohemia exhibits significant variability in the proportions of the observed technological phenomena among sites.

• Závist (ZAV) is again seen as an exceptional site in Central Bohemia. It exhibits a high proportion of DPP techniques and the presence of a unique tradition of combined methods using rotational motion in the clockwise direction.

• The high proportion of WT pottery in Central Bohemia is accompanied by the absence of fine HB pottery, a low proportion of fine tableware in the ceramic assemblages, and the use of rotational motion in the production of coarse pottery, phenomena that are not observed in the W and NW Bohemia.

Discussion

The discussion focuses mainly on the results from Central and W Bohemia. The variable results obtained from fewer sites in NW Bohemia do not provide a clear picture of technological trends. The observed variability in this region may be related to the partial heterogeneity in coarse ware ceramic production observed within the region in LT C2–D1 in terms of morphology and decoration (Salač 1990; Rulf and Salač 1995; Salač and Kubálek 2015).

In Central Bohemia, the region with dense settlements in all the observed periods, we find considerable diversity in technological chains with a high representation of combined forming methods, accompanied by significant structural transformation caused by the use of rotational motion (WS) in the LT A period. On the contrary, the first WM pottery introduced to W Bohemia does not exhibit a noticeable transformation in the internal structure of the vessel walls, indicating minimal use of rotational motion during forming (WF; Figs. 18A and 19D). This suggests a lower level of motor skills required to master the work on the rotational device. Typical new LT A shapes of tableware usually produced using rotational motion were also manufactured without its use. Therefore, a strict distinction between HB and WM production was not observed.

The results contrast to our previous findings in another peripheral region of the La Tène settlement—the Chrudim region in Eastern Bohemia (Thér et al. 2017). The introduction of wheel-made (WM) pottery in the Chrudim region marked a significant technological shift, characterised by an apparent discontinuity from other pottery production practices. The contrast between WM and HB pottery production is evident in the traditions expressed by the exerted skills and materialised in the products. Our conclusion suggested that WM pottery represented an item of high social value crafted by a specific group of skilled potters (for the discussion on the relation between social charged goods and requirements for their production, see Earle 1981; Miller 1982; Brumfiel and Earle 1987; Peregrine 1991; Hayden 1995, 1998). We propose that influential individuals of high social status played a pivotal role in introducing WM pottery during the La Tène A period. These individuals probably acted as opinion leaders (for the concept of a leader of opinion, see Rogers and Shoemaker 1971; Rogers 1983; Bargatzky 1989), promoting the new pottery style. They introduced potters who mastered the potter’s wheel to the region as attached specialists established mutually beneficial relationships with them. Consequently, these potters created a strictly distinctive community of practice, transmitting the technology in a closed learning system (Thér et al. 2017). The observation conforms to the well-documented scenario that, in traditional societies, discontinuous innovations are initiated by individuals having some form of power not for their techno-economic advantages but for their symbolic or social value (Creswell 1996; Roux 2010). A similar scenario for introducing wheel-made pottery was proposed for the Southern Levant (Roux and Courty 2005; Roux 2010) and Mesopotamia (Baldi and Roux 2016). For the La Tène period in Central Europe, it was considered for wheel-made pottery (Gosden 1983, 1987) and for iron production (Bauvais 2008). Elite involvement in the specialised production of metal items or the processing of amber in the Early Iron Age in Moravia has been hypothesised, based on spatial correlation between the evidence for the presence of elites and craft production (Mírová 2019; Golec and Fojtík 2020; Golec and Mírová 2020).

In the Chrudim region, we focused on a small and relatively peripheral region characterised by low population density, presumably simpler inter-group economic relations, and limited social networks, resulting in a lower degree of cultural interconnectedness compared to other regions. Paradoxically, such an environment may have provided favourable conditions for the control of WM pottery production, as there were no stimulating factors for the development of independent craft specialisation and local technologies could supplement the limited availability of luxury imports.

The conditions for the use of pottery as a socially significant item probably did not arise in W Bohemia. Despite relatively unsuitable agricultural conditions, significant manifestations of elites can be observed in the LT A period (Chytráček 1983, 2000, 2012; Bašta et al. 1989; Chytráček and Metlička 2004; Kozáková et al. 2016; Trefný 2017) in contrast to the Chrudim region. The involvement of local elites in interregional exchange played a significant role in the potential for the effective use of imports as a means of social negotiation (e.g. Bouzek et al. 2017; Trefný 2017), and we can hypothesise that there was no need to use local technology for this purpose, at least in the case of pottery.⁶

Similarly, in Central Bohemia, we do not observe features typical for the control over the production of the first WM pottery. However, other factors were at play here. The settlement density and the extent of social networks and connectivity with other regions probably hindered the effective control of WM pottery manufacture and distribution.

The transition from LT A to LT B in Central Bohemia witnessed a reduction in technological diversity and the emergence of the WT method, which already gained significant prominence in LT B-C1 and predominated in LT C2-D1, especially in producing fine tableware (Figs. 18B, C and 19D), but it was also employed for coarse ware (Fig. 13). The predominance of WT can be attributed to changes in the selective environment resulting from increased socio-economic complexity during the studied period. The results suggest the existence of independent specialists sensitive to cost-effective production practices. The estimated time required to produce typical La Tène bowls is approximately four times shorter when utilising WT compared to other WM methods (Thér et al. 2015b). The existence of market primarily based on impersonal considerations of value determined by supply and demand is the prerequisite for this performance to manifest differential interaction, creating a selective environment in which efficiency became a significant component of the cultural fitness of WT. The petrographic and geochemical analyses of the LT C-D1 pottery from Eastern Bohemia demonstrated the operation of a larger number of coexisting workshops with small distribution circles (Thér et al. 2015a). A similar picture emerged from the chemical analysis of bronze wheel amulets in the Czech Republic (Danielisová et al. 2020) and the analysis of iron production in the Paris Basin (Bauvais and Fluzin 2013). A strong indication for an open market is the significant degree of monetization of the society from the 2nd half of the third century BC onwards, manifested in the large number of finds of small silver circulation coins on common settlements (Militký 2018). The conditions for the division of labour in pottery production among households were established in densely settled Central Bohemia, with the occurrence of oppida resulting from population nucleation across various social levels (for the discussion on the role of oppida, see Collis 1984, 1995; Woolf 1993; Brun 1995; Büchsenschütz 1995; Crumley 1995a; Wells 1995; Salač 1996; Venclová 2002; Augstein 2006; Danielisová 2011; Fernández-Götz 2014; Moore 2017). These large Late La Tène fortified sites do not represent a homogeneous category: some functioned as craft and trade centres, others are seen more as socio-political and religious centres, and others are seen as refuges. Nevertheless, in many cases, the concentration of non-agricultural production within oppida is well-documented (e.g. Collis 1984; Venclová 2002; Danielisová 2011), creating an extensive market. The development of the use of the potter’s wheel exhibits fundamental aspects similar to our previous findings in the Brno region in Moravia (Thér and Mangel 2021) but with more pronounced manifestations of central region characteristics, including a greater diversity of techniques during the LT A period and the utilisation of WT for coarse ceramics during LT C2-D1.

If oppida represented the climax of socio-economic complexity in the Late La Tène period, we would expect a higher proportion of WT pottery at the oppida than at common settlements. However, although this can be observed for other types of (supra)regional centres (Týnec nad Labem and Lovosice), this is not the case in at least one sampled oppidum (Závist). Therefore, oppida seems to be more of a symptom of the region’s socio-economic complexity rather than its culmination, indicating that the oppidum environment itself does not necessarily provide better conditions for specialisation in the pottery craft.⁷ The occurrence of two-chamber pottery kilns, considered evidence for specialised production, supports this thought. While they can be found in oppida (e.g. Bratislava, Manching, Staré Hradisko: Čižmář 2002; Leicht 2013; Vrtel 2016), their relative numbers are comparable to those of rural settlements and unfortified centres (Mangel and Thér 2018). Moreover, the Závist oppidum is specific, as the Late La Tène settlement here follows an older tradition associated with the existence of the LT A period central hillfort, with stone podiums on the acropolis interpreted as a ritual area (Drda and Rybová 1997, 2001, 2008). Such continuities with older settlements associated with the presence of shrines are also observed in other oppida, which primarily functioned as public spaces for religious and political purposes and may also have played an important role in the formation of a collective identity (Fernández-Götz 2014; Moore 2017). This could indicate that the primary functions of the Závist oppidum lay on a more political-religious level. Definitive answers to this question can be provided only through research explicitly focused on the differences between oppida and common settlements in the region. The current dataset includes only one oppidum.

In contrast to Central Bohemia, in the later phases of the La Tène period, W Bohemia represents a transitional periphery with sparser settlements. Correspondingly, there is a significantly lower representation of WT pottery in the LT C2-D1 assemblages (Figs. 18C and 19D). The absence of two-chamber pottery kilns would also suggest poorer conditions for specialisation in the pottery craft. So far, not a single example from W Bohemia has been documented while at least seven kilns from the LT B2-D1 have been found in Central Bohemia (Mangel and Thér 2018; Mangel et al. 2021).

Conclusions

The Late Iron Age is the period when the use of rotational motion in pottery forming emerged in Bohemia and also disappeared at the end of this period. This study has complemented the existing knowledge on the application of rotational movement in this context (Thér et al. 2015a, 2017; Thér and Mangel 2021) with evidence for evolutionary scenarios that show the unique interplays of the performances of different variants of this general innovative idea with specific local socio-cultural conditions. We have proposed expectedly different scenarios in central and peripheral regions, but also unique differences that compel a deeper search for specificities, broadening the perspective for understanding the respective societies. In the concluding summary, we will present the discussed scenarios from the transmission context’s perspective.

Some of the sampled regions, namely Central Bohemia (including its eastern edge) and NW Bohemia, are characterised by good conditions for agriculture, high population density, and stable settlement throughout the period. In these regions, we observed the most profound technological changes in the transition between the Hallstatt and the La Tène periods, with a high proportion of pottery produced using methods efficiently utilising RKE (wheel shaping) from the beginning of the occurrence of wheel-made pottery. In Central Bohemia, the high variability of technological solutions with significant representation of techniques efficiently utilising RKE and increased technical requirements and learning costs reflects the dynamics of development and increased receptivity to innovations in the LT A period. Based on the archaeological evidence presented in this work, we are unable to decisively determine whether the observed diversity represents the establishment of distinctive traditions reflecting distinctive communities of practice that have developed different technological recipes transmitted in closed learning systems⁸ (cf. Dumont 1952; Foster 1956; Nicklin 1971; Rye and Evans 1976; Roux 2007) or an open network with numerous connections crossing the borders of communities of practice reflecting the blending of various traditions during the crystallisation of a new cultural pattern. Significant changes in material culture, settlement structure, burial practices, and other cultural aspects at the transitions between LT A and LT B-LT C1 (e.g. Collis 1995; Venclová 2013a, 2013b; Thér and Mangel 2014; Danielisová et al. 2019) were also accompanied by profound changes in pottery-forming methods in Central Bohemia. However, the changes were not entirely discontinuous: one of the diverse range of primary forming techniques of the previous period, coiling became dominant; all the rotational devices rotated in the same direction; and a new variant of the use of rotational motion appeared—wheel throwing. The foundation for its adoption was laid in the region by the popularity of wheel shaping in the previous period. Wheel throwing became the dominant method for the production of fine tableware in the final phase of the La Tène period when social complexity culminated. The development corresponds to a robust and open learning system of independent specialists operating on the open market.

Another sampled region, W Bohemia, is characterised by less fertile soils, with a sparser population and disruptive changes in settlement during the La Tène period. Here, the first adoption of techniques utilising rotational motion was less discontinuous than in other regions. Wheel-finished pottery is dominant in LT A and decreases slowly in the subsequent periods. The region exhibits a more gradual introduction of innovations. The typical La Tène fine tableware shapes were produced using just hand-built techniques and in combination with wheel finishing, suggesting adopting new ideas into the local pottery community without indications of changes in the organisational forms of production. The limited application of rotational movement, not requiring a radical change in the learning of new skills, is consistent with the diffusion of novelty based on strong social ties within a kinship network (Collar et al. 2015; cf. Roux 2020) and the dissemination of knowledge in an open system. In the LT A period, W Bohemia was strongly connected to the exchange with surrounding regions through elites. The extent and intensity of weak ties (infrequently accessed social connections; Granovetter 1973; Collar et al. 2015; Manzo et al. 2018) were prerequisites for the adoption of new ideas into the local pottery community. The demand for new shapes and, in general, the new visual performance of ceramics corresponds to the overall receptivity of the society to new elements of the emerging La Tène culture. It seems that the elites had no interest in using wheel-made pottery as a socially charged item, so they did not engage economically in its manufacture to gain control over the production. Consequently, the technology has been adapted to local conditions, requiring continuity with existing skills or parameters of other parts of the technological chain and reflecting the absence of conditions for applying high-learning-cost technologies. It leads to local innovation—application of wheel finishing instead of methods effectively using RKE. Also, copying errors in transmission under weak ties could play a role in the local innovation (cf., Eerkens and Lipo 2005). This picture contrasts with the results of our previous analysis of the introduction of wheel-made pottery in another peripheral region, the Chrudim region, which exhibited typical features of the presence of attached specialists operating in a fragile and closed learning system (Thér et al. 2017).

The discontinuity at the transition from LT A to LT B-C1 affected W Bohemia more significantly. In contrast to Central and NW Bohemia, in W Bohemia, there is an almost complete absence of evidence for settlements and burials, during the LT B period (Metlička et al. 2022). The situation is similar in the Chrudim region, where the temporary absence of wheel-made pottery and a change in the direction of rotation in newly introduced wheel-made production coincided with the decline of elites of the LT A phase (Thér et al. 2017). The subsequent gradual increase in settlement in both these regions during the LT C period can be linked to a significant population change, which may also underlie the disruptive changes in pottery manufacturing practices. In W Bohemia, there is a striking discontinuity in hand-building techniques. Coiling techniques became supremely predominant, as in Central Bohemia. Considering that a dramatic shift in population accompanied this change and that the discontinuity pertains to the techniques whose performance does not affect the fitness of technology under changing economic conditions (coiling vs. DPP techniques), this change most probably reflects not a shift in stimuli for technological development but rather a discontinuity in the overall settlement of the W Bohemia and its recolonisation from central areas. This is reflected not only in hand-building techniques but also in a comparable proportion of wheel-thrown pottery between Central and W Bohemia in the LT B-C1 period. However, in Central Bohemia, there were more favourable conditions for the development of specialisation in pottery production. Hence, over time, the use of wheel throwing increased, while in W Bohemia, it stagnated, and its use for producing fine tableware remains a minority also in final phase of the La Tène period—LT C2-D1 (for the ethnographic examples of the persistence of not exploiting the full potential of the potter’s wheel in relation to lower craft specialisation, see Nicklin 1971).

The bottom line of the search for links between technological evolution and society is the end of the La Tène period. With the breakdown of social structures and the onset of a new era, wheel-made pottery disappeared entirely for several centuries (Salač 2011b; Beneš 2019).