Definition

Simplifying a complex debate, definitions of ethnicity often stress either strategic situational factors, an individual’s emotional needs for self-identification and belonging or categorizations imposed by one set of groups on others (see below). In all three cases, the fact that an identification is ‘ethnic’ (as opposed to, e.g., class-related) boils down to the importance of shared common origin to make the group appear primordial (often largely mythical and/or unverifiable, e.g., Weber (Reference Weber1976, 237)). In contrast to kinship groups, this ‘common descent’ generally refers to a time long past and often a particularly pertinent generative event, such as a localized creation story or a migration (e.g. Cipolla Reference Cipolla2013; Lorenz Reference Lorenz, Berger and Lorenz2008; Prem Reference Prem1996; Voss Reference Voss2008; Reference Voss2015).

Similarly, ethnicity does not require that an actual daily arena for interaction exists, for example, one based on co-residence. As a result, there is no clear spatial scale (such as a settlement, group of settlements or ‘culture province’) that archaeologists could point to as reflecting ethnicity. However, ethnicity can form an important identifier in deciding on common political action, just as shared political action in turn can create a sense of ethnicity in the first place (e.g. Barth Reference Barth and Barth1969; Blanton Reference Blanton2015; Hu Reference Hu2013). As such, ethnicity is a level of analysis best suited to identifying targeted, group-based political action, rather than necessarily an individual’s self-definition. Once it has come into being, the ethnos then perpetuates its boundaries through, for example, visible aspects of dress or behaviour (Weber Reference Weber1976, 239; Weik Reference Weik2014), or other aspects of material culture. This differs from the concept of ‘race’ as a social difference primarily applied to selected visible biological characteristics, such as skin colour. In its modern guise, this concept is rooted in 15th-century Europe and is strongly focused on negative categorizations of groups by (European) outsiders (Lorenz Reference Lorenz, Berger and Lorenz2008, 37–41). Although it may be interesting to investigate whether biological differences were socially marked in prehistory, this paper maintains a distinction between ethnicity as a self-constructed boundary based on common history and race as an imposed category based on perceived physical distinctions. Similarly, issues of language have been excluded from the current paper, partly because linguistic uniformity can contribute to, but is not necessary for, ethnic group formation (e.g. Metcalf Reference Metcalf2010, 69–73) and partly because including the highly polarized discussion (e.g. Demoule Reference Demoule2023) goes beyond the confines of this paper.

Developing new perspectives

In archaeology, two main strategies have developed for dealing with ethnic groups (pointed out by, e.g., Jones (Reference Jones1997, 15–279) and Fernández-Götz (Reference Fernández-Götz2013)), largely split along sub-disciplinary boundaries. The first strategy, particularly common amongst prehistorians, is denial – either vocally or through lack of interest. Here, the problem is the link of ‘ethnic’ identifications to fundamental analytical units, notably the culture concept. In much continental European prehistoric archaeology, the entire chronological system, classification of material culture and method of registration for finds and sites is based on culture attribution. We fear that, in practice, this is now virtually impossible to change (in spite of continued valiant efforts, e.g., Feinman and Neitzel (Reference Feinman and Neitzel2020)) and sits at the very core of archaeological reasoning. The default setting has therefore been to pay lip service to the idea that cultures are ‘only’ a classification tool (e.g. Lüning Reference Lüning1972). This has at best been partially successful, as the people using the material culture are given equivalent names (‘die Bandkeramiker’, ‘the Yamnaya’ and suchlike) and therefore an identity group is still implied. As a result, changes between ‘cultures’ are still often explained by population shifts, while episodes of violence within ‘cultures’ are seen as ‘internal’ breakdown, and surprise and consternation reign when aspects of two different cultures are found mixed (causing the proliferation of hyphenated groupings such as Pfyn-Altheim and the like). While some researchers (Burmeister Reference Burmeister2017; Clarke Reference Clarke1968; Eggert Reference Eggert1978; Furholt Reference Furholt2009; Reference Furholt2020; Lüning Reference Lüning1972; Müller-Scheeßel Reference Müller-Scheeßel2000; Vander Linden Reference Vander Linden, Fowler, Harding and Hofmann2015) have pushed for a polythetic model of culture, such thinking remains to be widely adopted.

By contrast, a second strategy is found in Early Medieval/Migration Period archaeology, where ethnicity could not be ignored, given the presence of ethnic labels in written sources. The appearance of new burial customs or metal dress accessories was linked to newly incoming groups, with archaeology mainly used to illustrate the migration processes known from historical records. In accordance with discourses in the social sciences (e.g. Hall Reference Hall, Hall and du Gay1996), the material record was not seen as prima facie evidence for ethnicity but rather as a reflection of diverse strategies of social interaction between different groups in the immigrant and host society (Brather Reference Brather2004; Kazanski Reference Kazanski1989; Von Rummel Reference Von Rummel2007).

The central problem which triggers critique and rejection of ethnicity as a concept is the tendency to treat it as primordial and static, an approach Brubaker (Reference Brubaker, Squires, May and Modood2004, 50–51) calls ‘groupism’. Brubaker argues that much of the conceptual confusion stems from the failure to differentiate between groups and categories. He characterizes groups as a ‘mutually interacting, […] mutually recognizing, mutually oriented, effectively communicating, bounded collectivity with a sense of solidarity, corporate identity, and capacity for concerted action’ (Brubaker Reference Brubaker, Squires, May and Modood2004, 55), while categories are imposed upon people and are potentially used for the creation of groups – occupation, place of origin, age or gender, class, physical appearance and so forth (see also Jenkins Reference Jenkins, Stone and Dennis2003). For Brubaker (Reference Brubaker, Squires, May and Modood2004, 52–53), the most interesting question therefore is how and when categories (as ultimately a kind of abstraction) are mobilized in the creation of groups as entities for concrete common action.

Applying Brubaker’s (Reference Brubaker, Squires, May and Modood2004; Reference Brubaker2006) ideas to the prehistoric past requires a change of perspective which does not treat the existence of groups or ethnicity as premises but rather sees them as dynamic and historical outcomes of social interactions – as contingent phenomena which emerge under specific circumstances and have specific consequences. This historical contingency forces archaeologists to acknowledge the possibility of low or altogether missing groupness at times but also helps direct attention to situations of higher social cohesion which are discernible archaeologically. These dynamics of groupness are often masked in the archaeological record owing to the static classificatory systems of cultures, ‘groups’, ‘complexes’ or chronological phases. Yet, conceptually, there is no hindrance to breaking up these artificial entities.

Groupness as a variable can potentially emerge (as well as fade) around different categories, such as kinship, occupation, access to wealth or origin. Ethnicity is a specific manifestation of groupness, which is usually actively created or enhanced by specific interest groups to create boundaries for some people and strengthen solidarity with others, according to an ideology of a common history. In this sense, ethnicity is the active mobilization of latent layers of identity in a larger group, activated to serve particular interests. Its success depends on the willingness or unwillingness of the targeted representatives to contribute their collective power to the group-making process (Brubaker Reference Brubaker, Squires, May and Modood2004, 62–66). Such ethnic identities, variable and fluctuating as they may be, are often powerful factors of real-world interactions, not least among these conflict and war (Brubaker Reference Brubaker, Squires, May and Modood2004, 56), and should therefore not be easily discarded, even if they might be hard to define clearly in prehistoric periods.

Brubaker’s approach is very close to that of Barth (Reference Barth and Barth1969), who charts the ways in which ethnic identity is bound up with economic and political goals and can be situationally redefined to reflect these. Although subsequently critiqued for downplaying both the asymmetrical power relationships inherent in the ethnic classification of ‘others’ and the emotional salience of ethnic attachments (e.g. Stone Reference Stone2003, 32–38; Voss Reference Voss2008, 26–27), this provided the conceptual toolkit for stressing that ethnic identities react flexibly to their wider context. These more instrumentalist or processual definitions avoid much of the often racist political baggage that came with static ethnic designations imposed by outsiders, for example, in a US colonial context (Hu Reference Hu2013; Jenkins Reference Jenkins, Stone and Dennis2003; Lema Reference Lema1993; Smith Reference Smith2008, 29–31; Voss Reference Voss2008, 657; Weik Reference Weik2014, 296).

For European prehistory, the question is how these ‘constructivist’ analytical foci can be combined with the deeply felt sense of commonality that ‘primordial’ approaches to ethnicity stress, and that is just as important for political efficacy (Hu Reference Hu2013, 376). Following Stovel (Reference Stovel2013), one can identify a passive view of ethnicity based on Bourdieu’s concept of the habitus, according to which ethnic identity is most clearly visible in unreflected aspects of life such as ‘the good way’ of living together. This privileges ‘primordial’ understandings. Another strand, based on Barth’s work, sees ethnicity as consciously communicated in a situation of competition, or at least contact, between groups. This must therefore rely on material culture indicators which are visible to outsiders (Stone Reference Stone2003, 44).

Taking our cue from Stovel (Reference Stovel2013), in this paper the term ‘ethnicity’ is used in the latter sense of an actively, strategically communicated group identity. Methodologically, we must thus distinguish, on the one hand, the pool of broad similarities in lifeways across a region from, on the other, the conscious selection of some aspects of these lifeways for boundary marking. These boundaries, and the features they are based on, are temporary and historically specific. The identification of migrant groups (who may, e.g., stand out through specific material culture patterns in contexts that are not visible to society as a whole, see Burmeister (Reference Burmeister, Heyd and Aholaforthcoming)) is hence not enough to identify the presence of different ethnic groups. This means that an archaeological investigation of ethnicity needs to cover the wider temporal cycle of the creation, maintenance and eventually dissolution of specific practices of ethnicity (Kohl’s (Reference Kohl1998, 232) ‘ethnomorphosis’ or Brubaker’s (Reference Brubaker, Squires, May and Modood2004, 61) ‘curves of groupness’).

In what follows, we highlight three situations in which migration events kick off these cycles of increasing and decreasing groupness. Migrations, in particular longer-distance ones, can bring into contact persons and groups who previously had no or only little interaction, leading to situations in which potential conflicts are heightened by an appreciation of mutual difference (Clark et al. Reference Clark, Birch, Hegmon, Mills, Glowacki, Ortman, Dean, Gauthier, Lyons, Peeples, Borck and Ware2019; Hu Reference Hu2013; Voss Reference Voss2015). Boundaries need to be negotiated and references to common origins are likely to occur. Our case studies, all centred on Europe, each foreground another aspect of how ethnic groups can form or change in migration scenarios. The Migration Period, with its more fine-grained data, is used to trace the construction of new ethnic identities from a variety of inspirations at times of social upheaval. The complexities of this process provide a new inspiration also for our prehistoric case studies. The Early Neolithic Linearbandkeramik culture illustrates how the ethnicity-marking qualities of material culture, in this case pottery, change over a longer time frame. The Late Neolithic Corded Ware scenario evokes the regionally diverse responses to new possible modes for ethnic self-definition. In every case, a migration event can be seen as the impetus for ethnogenesis, creating the need for explicit group demarcation, actively used to frame common action.

Ethnicity in the Migration Period

The Migration Period offers a good starting point for further debate. Within a relatively short period of time, we are confronted with a large number of different migration processes that covered much of the European continent. These processes have been richly described and commented on in ancient historiography, and the multiple scales, modalities and directions of movement are increasingly documented bioarchaeologically (see Depaermentier Reference Depaermentier2023). Overall, this creates a picture of highly dynamic societies. Although the migrant collectives themselves did not contribute any significant written records, there is a wealth of iconographic and archaeological sources (primarily grave finds) that complete the picture.

The historian Wenskus (Reference Wenskus1961) laid the foundation for a new perspective on the development of the early medieval gentes. He emphasized the multi-ethnic character of historically recorded migration collectives and developed the concept of the tradition core: many groups of different origins gathered around a military elite. This elite formed the so-called core of tradition and gave the ethnic name to the union of different groups. As a result, the previous ethnic concept of self-contained groups of the same descent dissolved as a construct. It took over 30 years for this realization to find a broader resonance, and it was only via the new ethnicity debate in the social sciences that it found its way back into the historical sciences (Pohl Reference Pohl and Gillet2002).

Historians were uneasy about the concept of the traditional core, as it ultimately focused too much on the social elite and did not take collective group processes into account (Heather Reference Heather1998, 177; Pohl Reference Pohl, Pohl and Reimitz1998, 67). For archaeology, this concept was difficult to handle anyway. Whose material culture manifested itself in the archaeological record, that of the leading, but perhaps numerically inferior, elite or that of one of the larger components of the migrating collective? A direct link between the traditional ethnic group designation and material culture thus dissolved, at least in the sense of a group with common ancestry. This calls for a new way of thinking about the social correlates of material culture.

Lines of tradition between individual regions linked by migration are certainly visible in various areas of the cultural habitus, for example, in burial customs or traditional costume accessories. Such evidence is often used to track the course of a given migration, for example, the migration of the Goths from what is now Poland (Wielbark culture) to the western Black Sea region (Černjachow culture) (Gomolka-Fuchs Reference Gomolka-Fuchs1999; Magomedov Reference Magomedov2004). The migration routes of other Germanic peoples, such as the Anglo-Saxons, Lombards and Vandals, are also reconstructed in this way (e.g. Böhme Reference Böhme1986; Bierbrauer Reference Bierbrauer, Straume and Skar1992; Reference Bierbrauer and Pohl2004; Kokowski Reference Kokowski, Uelsberg and Wemhoff2020). Consequently, the respective tribal history is projected backwards, and an ethnic continuity is constructed. However, later cultural expressions are usually not a straightforward development of earlier ones; despite some remarkable similarities, they also show new influences. Nevertheless, e.g., the people using Černjachow material culture were in lasting contact with people using Wielbark material culture, and the two obviously acted jointly to combat the Roman Empire (Bursche and Myzgin Reference Bursche and Myzgin2020). Even if an ethnic identity cannot be deduced from this in a straightforward way, the observed lines of tradition suggest mobility, contact and cultural transfer.

Historians are now sceptical of such ethnic perspectives. Pohl (Reference Pohl and Gillet2002, 237) doubts that the Germanic tribes imported their identity unchanged into the Roman Empire; according to Heather (Reference Heather1998), for example, Gothic identity only developed in the Roman Empire. The controversy reflected here can essentially be summarized in two positions: An essentialist and static approach to interpretation – which is reflected in traditional central European archaeology – is contrasted with a dynamic and open approach. For the understanding of cultural practice, this can be cast as follows: Populations have a culture, or they create a culture.

The first aDNA analyses of Anglo-Saxon cemeteries have provided concrete insights into the composition of immigrant societies (Gretzinger et al. Reference Gretzinger, Sayer, Justeau, Altena, Pala, Dulias, Edwards, Jodoin, Lacher, Sabin, Vågene, Haak, Ebenesersdóttir, Moore, Radzeviciute, Schmidt, Brace, Abenhus Bager, Patterson and Schiffels2022; Schiffels et al. Reference Schiffels, Haak, Paajanen, Llamas, Popescu, Loe, Clarke, Lyons, Mortimer, Sayer, Tyler-Smith, Cooper and Durbin2016), revealing a differentiated picture. On the one hand, there is clear evidence of biological relationships between natives and immigrants. People of Romano-British descent were in some cases very richly endowed in the Anglo-Saxon style. However, in other cases, groups with different biological ancestries were buried together over many generations but did not form any family ties and were equipped with different grave goods, implying some separation. The relationship between people of Romano-British and those of Anglo-Saxon descent varied greatly from region to region, which warns against lumping together regionally distinct migration processes. A similar pattern is evidenced by a new aDNA study on Avar immigration in Austria (Wang et al. Reference Wang, Tobias, Pany-Kucera, Berner, Eggers, Gnecchi-Ruscone, Zlámalová, Gretzinger, Ingrová, Rohrlach, Tuke, Traverso, Klostermann, Koger, Friedrich, Wiltschke-Schrotta, Kirchengast, Liccardo, Wabnitz and Hofmanová2025).

The demographic upheavals at the end of the Roman Empire led to the reorganization of numerous groups. The debate about the ethnic interpretation of the Franks and Alamanni shows the problems of ethnic interpretation of archaeological data in general particularly well. The two groups appear as actors in Roman written sources from the third century onwards. They later developed into important political players in the early Middle Ages. In his comprehensive statistical analysis, Siegmund arrives at a clear differentiation of the average patterns of grave furnishings (Siegmund Reference Siegmund2000). For the sixth and seventh centuries, he differentiates two regional groups on the basis of their burial customs. As these can be matched regionally with the historically recorded territories of both groups, he sees confirmation for an ethnic interpretation, in accordance with the historical record, especially since the two groups are clearly separated from each other, particularly in the contact area.

Brather and Wotzka (Reference Brather, Wotzka, Burmeister and Müller-Scheeßel2006; with a reply by Siegmund (Reference Siegmund and Müller-Scheeßel2006)) decisively reject Siegmund’s ethnic interpretation on two levels: Firstly, the distribution pattern of the various traits used to distinguish the groups is determined by a variety of factors that do not necessarily have ethnic connotations; secondly, other multivariate approaches do not confirm the clear spatial separation of the two groups observed by Siegmund, especially in the contact area. The criticism hits on an essential point: Based solely on archaeological data, we can never be sure which traits were ethnically significant, i.e., that they were used and understood as ethnic markers by the ancient actors. The spread of certain cultural patterns can have other causes than ethnic identity and demarcation. But even if the mechanisms behind the development of spatial patterns cannot always be reconstructed with certainty, these areas still reflect different intensities of communication and social interaction – and ultimately, in the example of the Franks and Alamanni, also converge with their political formations, at least in broad outline.

During the Migration Period, many diverse population groups clashed with one another (Meier Reference Meier2019). Since they were all in a constant process of negotiating their social situation, it can hardly be assumed that ethnic identities were consistently preserved in this social melting pot. Where social negotiation processes picked up on the attributes of group-specific identities, Von Rummel (Reference Von Rummel2007) has argued that the habitus barbarus could have functioned as a form of social distinction for a new elite. As a consequence, elements of traditional costume lost their ethnic connotation in a conventional sense: Dress became playing with ethnically associated accessories but was not an expression of primordial belonging. The dictum attributed to Theodoric characterizes the situation perfectly: A poor Roman imitates a Goth, and a respected Goth imitates a Roman (quoted from Meier (Reference Meier2019, 89)).

In the course of the disintegration of the Roman social order, both the original inhabitants of the Roman provinces and the immigrant barbarians sought new forms of social distinction. The row-grave fields in the Empire’s former northern provinces were long interpreted as burial sites of immigrating Germanic groups but may instead reflect a cultural reorientation of the population in this ethnic contact zone (Fehr Reference Fehr2010). The social, political and cultural milieu of the Migration Period was a constant process of negotiation, which is not fully appreciated by either Siegmund (Reference Siegmund and Müller-Scheeßel2006) or Brather and Wotzka (Reference Brather, Wotzka, Burmeister and Müller-Scheeßel2006), who more or less explicitly assume that ethnic groups are separated from one another by spatially sharp borders. Numerous ethnographic examples refute this assumption as a general feature of ethnic identity (e.g. Hodder Reference Hodder1982).

With the growing social importance of barbarian groups in the Roman Empire and the increased social prestige of immigrants in Roman society, the cultural habitus of all those involved also changed. Germanic finds or cultural practices on Roman territory cannot be directly identified with Germanic immigrants but could also be the expression of acculturated Romans or a newly hybrid society (see, e.g., Fehr Reference Fehr2010). One could therefore reject the ethnic interpretation of archaeological data. The creative handling of cultural habitus detaches cultural traits from their original bearers so that they lose their ethnic significance. However, the opposite is the case here. New identities emerge in the performance of group membership. Processes of identity formation and group formation are mutually dependent. The ability of different groups to integrate facilitated the emergence of new ethnic groups at the end of the Migration Period (see Koch Reference Koch2012). Even if their names were based on the old tribal names, they were completely newly formed groups. Both the historical and the archaeological record illustrate these processes. They can be understood by moving away from the primordial paradigm of static, self-contained ethnic groups.

These processes can be traced in more detail in the Migration Period owing to the existence of written sources, and not least thanks to a long-standing history of research into migration and ethnicity. While the specifics of the case are historically situated, prehistoric situations were likely just as complex. In the following, we illustrate this with two examples, one from the beginning and one from the end of the European Neolithic.

Linearbandkeramik ethnicity. Of pots and people

Given its large spatial extent, spanning from modern-day Ukraine in the east to the Paris Basin in the west and from Hungary to the northern European Plain (Fig. 2), the Linearbandkeramik (LBK; ca 5500–4900 B.C.) could never have formed a coherent social unit. Ancient DNA studies have established that LBK material culture spread largely by migration (e.g. Haak et al. (Reference Haak, Balanovsky, Sanchez, Koshel, Zaporozhchenko, Adler, Der Sarkissian, Brandt, Schwarz, Nicklisch, Dresely, Fritsch, Balanovska, Villems, Meller, Alt and Cooper2010), and many papers since), initially from a source area in what is now western Hungary, Lower Austria and Slovakia (e.g. Oross and Bánffy Reference Oross and Bánffy2009), and later also from secondary settlement areas. In terms of ethnic identity, this could have created a common pool of social structures and material resources from which to draw when meeting forager groups. In what follows, we trace how these encounters changed over the longer term, using the example of pottery. While pottery is often associated with household consumption, there are public dimensions to its making (which can, for example, include communal installations for firing) and use (in larger gatherings) which qualify it as potentially functioning in ethnic signalling.

Figure 2. Simplified map showing the extent of the LBK (made with Natural Earth, free vector and raster map data @ naturalearthdata.com).

What the interaction between foragers and farmers over much of central Europe looked like is still hard to reconstruct, partly also because of the likely demographic imbalance between the groups that came into contact (Ringbauer et al. Reference Ringbauer, Novembre and Steinrücken2021). A small component of ‘Western Hunter-Gatherer’ DNA has been attested to in most LBK individuals but likely stems from admixture events much earlier on in the expansion of Neolithic groups from the Near East (Childbayeva et al. Reference Childbayeva, Rohrlach, Barquera, Rivollat, Aron, Szolek, Kohlbacher, Nicklisch, Alt, Gronenborn, Meller, Friederich, Prüfer, Deguilloux, Krause and Haak2022, 4). In spite of this demographic swamping, there are arguments for the coexistence of farmers and hunter-gatherers over several generations (summarized, e.g., in Hofmann (Reference Hofmann2015)). In addition, both lithic traditions (e.g. Robinson, Jadin and Bosquet Reference Robinson, Jadin and Bosquet2010; Robinson, Sergant and Crombé Reference Robinson, Sergant and Crombé2013) and pottery-making suggest hunter-gatherer continuity in LBK settings. Archaeologically, this seems clearest on the very western fringes of the LBK, which are thus our focus here. These areas are poor in aDNA data owing to preservation conditions (e.g. the Netherlands) and the absence of large cemeteries (e.g. the Paris Basin). However, a higher contribution of Western Hunter-Gatherer DNA in later populations in these areas (Rivollat et al. Reference Rivollat, Jeong, Schiffels, Küçükkalıpçı, Pemonge, Rohrlach, Alt, Binder, Friederich, Ghesquière, Gronenborn, Laporte, Lefranc, Meller, Réveillas, Rosenstock, Rottier, Scarre, Soler and Haak2020) is consistent with an interaction scenario. Broadly speaking, three stages in the interaction between producers of LBK and non-LBK pottery can be defined.

In the first stage, the earliest LBK communities in areas east of the Rhine came into contact with makers of pottery that was highly dissimilar to Linearbandkeramik ware in shape, decoration and production technique (e.g. temper), generally grouped under the term La Hoguette pottery (Kirschneck Reference Kirschneck2021). Although there is debate as to whether this is the product of forager technologies (e.g. Crombé Reference Crombé, Jordan and Zvelebil2009; Jeunesse and Van Willigen Reference Jeunesse, Van Willigen, Gronenborn and Petrasch2010; Lüning, Kloos and Albert Reference Lüning, Kloos and Albert1989; Manen and Mazurié de Keroualin Reference Manen and Mazurié de Keroualin2003; Schauer Reference Schauer2023), in this article we favour this position on the basis of the fundamental differences between the two types of ware, as well as the appearance of such vessels outside LBK contexts (see summary in Kirschneck (Reference Kirschneck2021)). The dating of this material is vague, but we suggest our first horizon begins just before the earliest LBK groups reached areas east of the Rhine in the 53rd century Cal. B.C. (Jakucs et al. Reference Jakucs, Bánffy, Oross, Voicsek, Bronk Ramsey, Dunbar, Kromer, Bayliss, Hofmann, Marshall and Whittle2016). The earliest LBK and La Hoguette ware then overlapped for a time, but the two traditions were used in very different settings.

Early La Hoguette pottery includes some particularly large and impressive vessels with a geographically wide distribution which may have been employed in the contexts of feasts and other gatherings (Kirschneck Reference Kirschneck, Becker, Bunnefeld, O’Neill, Woltermann, Beier and Einicke2020; also Hofmann Reference Hofmann, Amkreutz, Haack, Hofmann and van Wijk2016). In such a setting, pottery was most likely not intended to signal exclusive identities but rather shared practices and networks of tradition. In contrast, pottery of the so-called earliest LBK horizon is rather uniform over a large area and relatively sparsely decorated. Alongside other overall similarities in material culture, this can be interpreted as related to a migration situation in which internal differences were systematically downplayed to allow for the necessary flexibility in group membership and for the creation of support networks. For both groups, pottery was central to perpetuating identity through shared practice, but for the makers of La Hoguette pottery, this involved inter-group interaction, while LBK pottery encouraged conformity. In a situation of co-residence, when surrounded by a majority of people whose pottery worked as a means of signalling group belonging, novice potters may have been pushed towards learning the dominant LBK pottery traditions if they wanted to fit into their new communities – an active choice we can qualify as ethnic identity building. Indeed, the production of La Hoguette pottery on earliest LBK sites east of the Rhine is short-lived (Hofmann Reference Hofmann, Amkreutz, Haack, Hofmann and van Wijk2016).

This situation changes after around 5200 B.C., when regionalization in almost all aspects of LBK material culture had set in, in spite of continued personal mobility traced through isotopes (e.g. Hedges et al. Reference Hedges, Bentley, Bickle, Cullen, Dale, Fibiger, Hamilton, Hofmann, Nowell, Whittle, Bickle and Whittle2013). In the LBK settlements of the Low Countries, Alsace and the Paris Basin, non-LBK pottery is present throughout all phases and occurs in increasing technological and stylistic diversity. Alongside La Hoguette, there were now Limburg and Begleitkeramik styles, as well as others which do not fit into any predefined categories (Hofmann Reference Hofmann, Amkreutz, Haack, Hofmann and van Wijk2016; Kirschneck Reference Kirschneck2021). This pottery was locally idiosyncratic; it increasingly anchored its producers in a specific place and tradition while distinguishing these wares from LBK-style ones. This was now seemingly tolerated, indicating a change in how pottery functioned as an identity-marking artefact in LBK communities.

Taking the LBK of the Paris Basin (Rubané récent du Bassin parisien (RRBP)) as an example, communities began to differentiate themselves from more eastern areas in several highly visible ways. Pottery now sported garland-style decorations, with vertical line bundles suspended from a horizontal set of lines – a far cry from the meandering spiral bands of earlier times and more eastern areas (e.g. Meunier Reference Meunier2012, 233–34). Other novelties include changes in mortuary tradition, the virtual absence of polished stone axes in graves (Jeunesse Reference Jeunesse1995, 127) and a divergent settlement structure in which houses of different lengths also exhibited different economic orientations. The largest houses have the highest proportion of domesticates in their faunal assemblages, while in the shortest houses only just over half the animal bone spectrum consists of domesticates (Hachem Reference Hachem2000; Reference Hachem2011). It is here that we also see the greatest ceramic diversity (Gomart et al. Reference Gomart, Hachem, Hamon, Giligny and Ilett2015, 239), although this is here defined mainly in vessel shaping traditions). This pattern persisted over several settlement phases and is here interpreted as persons of forager origin settling in LBK sites but maintaining some economic traditions.

In sum, LBK groups newly migrating to the Paris basin made a concerted effort to differentiate their material culture from the more easterly areas where they originated, involving several visible categories of practice. The migration itself could have served as a foundational event that established a new, distinctly ‘RRBP’ identity whereby differences within the migrating group (for example, origin in different settlements) were downplayed. An ethnic split was being created within the phenomenon archaeologists usually lump together as Linearbandkeramik culture. At the same time, these newly created RRBP communities came into contact with local foragers, some of whom began to live in LBK settlements. This required some conformity, for example, living in longhouses and adopting limited agriculture. However, diversity in pottery traditions was tolerated in these likely mixed-origin households, perhaps because difference was acceptable just as long as it also differed from the pottery of LBK communities further east. In either case, acculturated foragers were accepted into RRBP communities, of which they formed a stable part for several generations.

Further mutual influencing then took place, leading to the final phase of our interaction model. Around 4950 B.C. (Dubouloz Reference Dubouloz2003), a new set of material practices was created which archaeologists have termed the Villeneuve-Saint-Germain (VSG) culture (sometimes also Blicquy-VSG) and which fuses LBK and forager-derived elements. Longhouses were still constructed (albeit now trapezoidal ones), but wild animal ornaments were used in burial rites (Jeunesse Reference Jeunesse and Guilaine2002). The pottery shows a general continuity with LBK styles but also has elements characteristic of non-LBK wares, such as bone tempering and herringbone decoration (e.g. Constantin Reference Constantin2013; Gomart Reference Gomart2010; Meunier Reference Meunier2012, 231, 241). This technological fusion took place at a regional, rather than local scale (Meunier Reference Meunier2012, 250–51), whereby Paris Basin communities again suppressed their internal diversity and aligned themselves to some other former LBK settlement areas, for instance, in modern-day Belgium, while increasing their difference to others. At this point, the migration history which plausibly may have served as a springboard for initial ethnic group formation was a long time past, and other triggers likely drove the formation of the VSG.

In the course of the LBK, therefore, migrations helped form the kinds of identities we have here described as ‘ethnic’, and these various stages of interaction are reflected in the social roles of pottery. In the earliest LBK, with its uniform material culture, no coexistence with other pottery traditions was possible. In subsequent LBK migrations, there was a greater tendency to use material culture to actively but selectively stress a difference from more immediate areas of origin. There was a higher level of tolerance for new ways of doing things, so long as they differed from specific ‘others’. This allowed the persistence of ceramic variability over several generations. Out of these admixtures, the successor groups of the LBK were eventually formed. These trajectories were not the same over the whole LBK territory – for example, mixed pottery assemblages suggest more porous boundaries at its eastern margins (e.g. Jakucs Reference Jakucs2020). This stresses the need for local, contextual appreciations of ethnicity, something our next case study also foregrounds.

Third millennium migration and ethnicity

Genomic and genome-wide ancient DNA research has revolutionized our opportunities to detect migration in the third millennium B.C., and in less than 10 years the use of such data to understand migration processes has developed significantly. The large-scale archaeological units of Yamnaya, Corded Ware and Bell Beaker cultures (Fig. 3) are now thought to be connected to a marked uptick in mobility, including migration in a largely east–west direction. Initially, research tended to focus on the appearance of new genetic components (Allentoft and Willerslev Reference Allentoft, Sikora, Sjögren, Rasmussen, Rasmussen, Stenderup, Damgaard, Schroeder, Ahlström, Vinner, Malaspinas, Margaryan, Higham, Chivall, Lynnerup, Harvig, Baron, Della Casa, Dąbrowski and Willerslev2015; Haak et al. Reference Haak, Lazaridis, Patterson, Rohland, Mallick, Llamas, Brandt, Nordenfelt, Harney, Stewardson, Fu, Mittnik, Bánffy, Economou, Francken, Friederich, Garrido Pena, Hallgren, Khartanovich and Reich2015; Kristiansen et al. Reference Kristiansen, Allentoft, Frei, Iversen, Johannsen, Kroonen, Pospieszny, Price, Rasmussen, Sjögren, Sikora and Willerslev2017), potentially overemphasizing initial migration events and downplaying later movements and mixtures that leave less conspicuous genetic signals (Hofmann et al. Reference Hofmann, Frieman, Furholt, Burmeister and Johannsen2024). However, a more multifaceted process, lasting several centuries and with repeated movements in different directions, was emerging. As demonstrated by Papac et al. (Reference Papac, Ernée, Dobeš, Langová, Rohrlach, Aron, Neumann, Spyrou, Rohland, Velemínský, Kuna, Brzobohatá, Culleton, Daněček, Danielisová, Dobisíková, Hložek, Kennett, Klementová and Haak2021) for Bohemia, early Corded Ware burials in that region most likely represent a motley crew of migrants of different origins, only becoming more genetically homogeneous in later centuries of the third millennium. This does not differ fundamentally from the situation discussed above for the Migration Period.

Figure 3. Simplified map showing the extent of main archaeological culture complexes during the third millennium B.C. in Europe (made with Natural Earth, free vector and raster map data @ naturalearthdata.com).

The Corded Ware phenomenon as a whole, despite its transregional display of new material symbols and burial rituals, also shows marked regional diversity from the start (Furholt Reference Furholt2014), which can be best explained by a combination of an initial diversity of migrating groups (see Papac et al. Reference Papac, Ernée, Dobeš, Langová, Rohrlach, Aron, Neumann, Spyrou, Rohland, Velemínský, Kuna, Brzobohatá, Culleton, Daněček, Danielisová, Dobisíková, Hložek, Kennett, Klementová and Haak2021), different levels of incorporation of diverse local traditions (Nielsen and Johannsen Reference Nielsen and Johannsen2023) and subsequently diverse historical trajectories (see below).

Corded Ware burial rituals were not brought as a full package into central Europe from the eastern European steppes but represent a novel complex of practices and material manifestations with some steppe tradition elements, such as the kurgan, individual interments and regular body orientations. To this, novel elements were added. Some of these, such as binary gender differentiation, the prominence of stone shaft-hole axes, the prevalence of drinking and serving vessels (Furholt Reference Furholt2019) or the appearance of circular ‘mortuary palisades’ around graves (Nielsen and Johannsen Reference Nielsen and Johannsen2023), came to prevail supra-regionally. Other elements, such as stone cists (Fischer Reference Fischer1956) or multiple burials (Dresely Reference Dresely2004), remained more regionally restricted. Parts of the characteristic and supra-regionally visible material culture also have roots in specific regions, for example, the defining ‘A-amphorae’, which typologically derive from Polish Globular Amphorae (Beran Reference Beran, Buchvaldek and Strahm1992; Furholt Reference Furholt2008; Reference Furholt2014), or the ‘A-axe’, which is most varied on the Jutland Peninsula (Furholt Reference Furholt2014). While Corded Ware A-axes are typologically novel, the tradition of placing shaft-hole axes in burials already started centuries earlier in different parts of central and western Europe (Schultrich Reference Schultrich, Preda-Bălănică and Ahola2023).

Overall, the newly forming Corded Ware burial ritual is an amalgam of patterns with longer histories in different regions of central Europe and new, steppe-derived elements. In some regions, these new burial rituals were only sporadically implemented, or not at all. In the coastal Netherlands, along the Baltic coasts and in what is now Switzerland, Corded Ware material was produced and used in settlements that were not fundamentally different from previous periods, and the material was gradually (the Netherlands), partially (western Switzerland) or abruptly (eastern Switzerland) introduced.

In conclusion, the archaeological record does indicate a dynamic social situation of multiple encounters between migrants and locally established individuals, with potential tensions and conflicts between different values, beliefs and traditions, social roles, gender concepts, political institutions or over rights to land, resources and privileges. Migration is logistically and socially demanding, requiring some level of group cohesion; however, the cultural encounters that followed further accelerated the potential for group-making, and we can expect ethnicities to form or consolidate to mobilize allegiances and demarcate social boundaries. This could have been connected to – among others – the creation of the new complex of burial customs with their clear reference to steppe traditions, especially in their most durably visible form of the round barrow or kurgan. At the same time pan-European – and thus regionally more familiar – elements such as the battle axe and certain pottery types were incorporated. While the genetic difference between migrants and locals was quickly blurred, overall burials which more closely conform to the new ideal type generally also displayed higher percentages of genetic steppe ancestry (Haak et al. Reference Haak, Furholt, Sikora, Rohrlach, Papac, Sjögren, Heyd, Fischer Mortensen, Nielsen, Müller, Feeser, Kroonen, Kristiansen, Willerslev, Kroonen and Kristiansen2023). It is thus plausible that possible Corded Ware ethnic identities did relate to an eastern European origin and to the ‘founding processes’ of migration, which emerged out of a highly dynamic cultural and demographic situation where groups of steppe, forest steppe and central European origins interacted and mixed (Allentoft et al. Reference Allentoft, Sikora, Refoyo-Martínez, Irving-Pease, Fischer, Barrie, Ingason, Stenderup, Sjögren, Pearson, Sousa da Mota, Schulz Paulsson, Halgren, Macleod, Schjellerup Jørkov, Demeter, Sørensen, Nielsen, Henriksen and Willerslev2024). However, it is also important to emphasize the role of new social values, which more likely originated in the subsequently forming Corded Ware areas themselves. This includes the binary gender concept, drinking/feasting and specific weapon or tool types, as well as possibly the concept of male warriorhood (Vandkilde Reference Vandkilde2007) or indeed alternative male social ideals, such as ‘cultivatorhood’ (see Wentink (Reference Wentink2020), who argues that shaft-hole axes were specialized agricultural tools; also Frieman (Reference Frieman2012)).

As Papac et al. (Reference Papac, Ernée, Dobeš, Langová, Rohrlach, Aron, Neumann, Spyrou, Rohland, Velemínský, Kuna, Brzobohatá, Culleton, Daněček, Danielisová, Dobisíková, Hložek, Kennett, Klementová and Haak2021) have shown for Bohemia, the initial genetic diversity of individuals in Corded Ware burials gives way to a more homogeneous, and apparently isolated set of Corded Ware individuals. Later on, they are found side by side with individuals buried following Bell Beaker and eventually Early Bronze Age traditions, evidently without much genetic admixture. While individuals in Bell Beaker graves genetically derive from a subset of Corded Ware lineages (Y-chromosome haplotype R1b-P312), by around 2500 B.C. there is almost no one with this specific haplotype buried in a Corded Ware grave. In central Europe, where this split between male biological lineages (as represented by Y-chromosome haplotypes) most likely occurred, Corded Ware- and Bell Beaker-associated burial customs are very similar. Yet, in some regions, such as central and southern Germany (Fischer Reference Fischer1956) and Bohemia (Buchvaldek Reference Buchvaldek1967; Großmann Reference Großmann2016), the main orientation in Bell Beaker-associated graves was changed from west–east to north–south, and the gender-specific body position was reversed. Most likely, the initial Bell Beaker package (Shennan Reference Shennan, Lanting and van der Waals1976) formed in connection with a new incident of strategic ethnic marking, this time stressing the difference between distinct, albeit closely biologically related male lineages. Ethnic identity was hence used to unite different migrants into a single group at the start of the third millennium, resulting in the creation of Corded Ware burial customs and material culture. Around 2500 B.C., however, ethnicity was mustered to create or exploit an already existing split within this group in parts of central Europe, resulting in the formation of Bell Beaker burial customs. Outside the direct contact area between Corded Ware- and Bell Beaker-associated customs, such as in western Europe, this opposition did not exist. An interesting case is the Netherlands (e.g. Wentink Reference Wentink2020), where there is no difference in preferred orientations between Corded Ware and Bell Beaker graves. It is possible that we are dealing with a situation in which the instrumental use of ethnicity, which marked the creation of the Beaker package further east, had ceased to be salient, and a marked opposition to users of Corded Ware was no longer deemed important.

A scenario for the third millennium could be envisioned in which certain subgroups, in the course of the complex migration and mixing processes around the turn of the millennia, tried to combine new value systems concerning specific gender roles with a collective experience and, gradually, memory of migration from eastern Europe more broadly, possibly focusing on the kurgan tradition from the south-east. Importantly, this group-making effort was only successful in some regions and appears to have taken roots to a much lesser degree, or not at all, in others.

Concluding discussion

Our three case studies have highlighted how group identities are formed and transformed in the course of migrations. In all our settings, the migration process itself functioned as the likely catalyst for ethnogenesis – it was the common origin point which first brought groups together or demonstrated their connectedness. However, how precisely this happened also varied from case to case.

In our Migration Period study, we have drawn out how belonging to one or another ethnic group could be an active choice. In some places, people with different genetic signatures flexibly opted into newly forming groups by adopting cultural markers, while elsewhere these options were more restricted. This had to do with the historical and personal constellations acting on a particular case, including the initial core of ethnic belonging in mobile male warriors. For the Linearbandkeramik, pottery was more or less strongly involved in creating boundaries between groups, separating a first migration event which stressed group uniformity from later migrations, where people marked their difference from specific origin areas. This probably had more to do with different interactions with local foragers than with the composition of migrating groups, which generally involved both men, women and children from different origin communities. In the Corded Ware case, the extent to which migration played a foundational role in creating new ethnic identities varied regionally. The overall effect was a patchwork of selective adoption, partial rejection or enthusiastic embracing of new customs and material culture forms which were variously used to differentiate oneself from others. In all three case studies, migration brought together people from diverse backgrounds, and it is our current focus on genetic signals that encourages a view of closed, internally homogeneous groups of migrants rather than diverse histories of combining cultural and biological ancestries.

We are well aware that identifying identity groups of any sort archaeologically, particularly in prehistory, is challenging and rarely leads to a definite answer. Therefore, we are not proposing a kind of quick-fix recipe of how to ‘see ethnicity’ or a toolbox to identify ethnic identity, let alone quantify it or separate it from other social arenas. Nevertheless, we argue that explicit engagement with ethnicity in prehistory is necessary for two main reasons.

First, recent research combining archaeological material, aDNA data and historical linguistics (e.g. Kristiansen et al. Reference Kristiansen, Allentoft, Frei, Iversen, Johannsen, Kroonen, Pospieszny, Price, Rasmussen, Sjögren, Sikora and Willerslev2017) has opened up exciting interpretative spaces, but it has also run the danger of resuscitating the unholy trinity of culture, biology and language that formed the basis of 19th- and early-20th-century primordial notions of ethnicity. This danger stems from a failure to explicitly engage with the fundamental politics of groupmaking in the past, in other words, with ethnicity. Although most scholars reject the existence of closed, homogenous identity groupings, in practice it is often easier to simply assume that all three aspects coincided, or focus in on the few situations where they might, rather than laboriously tracing and contextualizing longer-term developments. This has the unfortunate side effect of feeding a simplistic public discourse on the matter, rather than using the opportunity to let archaeology contribute more creatively to current debates (e.g. Frieman and Hofmann Reference Frieman and Hofmann2019).

Second, thinking in terms of ethnicity (rather than, for instance, community) explicitly foregrounds active political choice in the past. For us, ethnicity is an explicit boundary-marking process that happens in conscious opposition to ‘others’. Identifying ethnic groups relies on a perceived sense of common origins, including a common history, and it requires that these ‘origins’ be deployed to frame shared interests. While this can rarely be identified beyond any doubt, we need to broaden our interpretative horizons. Joining the sociological and social anthropological discourse on ethnicity will help us problematize and explicitly discuss group formation processes, to investigate when groups emerge, how long they last, and how they change, and which interest groups drive the instrumentalization of shared practices in this way. This contributes to restoring a necessary sense of political agency to past actors (cf. Graeber and Wengrow Reference Graeber and Wengrow2021).

In prehistoric contexts, this will be challenging. In many areas, we have good background archaeological data on how people made their living or buried their dead. We can map material culture similarity, but the resulting distribution patterns will not directly map ethnic groups (if only for reasons of scale, see Wotzka Reference Wotzka1997). Sometimes, isotopic signatures inform us on personal histories of mobility, while new kinds of aDNA data, such as shared identity by descent (IBD) segments, can tell us whether people chose their reproductive partners broadly or from a restricted group. Yet, how this maps onto other kinds of identity-marking practices is not straightforward. Methodologically speaking, a great deal of contextual data and a long-term view are necessary to put forward the interpretation that a given group formation process can be characterized as ‘ethnogenesis’. We must trace the development of many different material culture indicators over time, check where one or more of them change abruptly and how generally visible this may be. On this basis, we can document instances in which specific things and practices are used for boundary marking and investigate how changing mobility patterns or newly arriving populations could be implicated. Yet, the effect is not spatially coherent or temporally static. The extent to which ethnic mobilization succeeded in the first place, how long it could be sustained or the relative openness of boundaries all varied. Thinking through a case study on ethnic terms thus highlights dynamic aspects, rather than the kind of static classification proposed in the past.

Migrations are one situation in which ruptures and new social constellations can emerge, and we have chosen it here to highlight the starting points of new cycles of group-making. This shows that the ‘shared origins’ on which ethnic groups are based are not necessarily rooted in biology but are active responses to changing situations. Ethnic identities can go from smaller to more inclusive units, as in the Migration Period, or the other way around, as in the emerging Corded Ware–Bell Beaker difference in the Late Neolithic. Who ethnic signalling is aimed at may also change over time, as in the LBK. Realizing that human collectives use ideas of ethnic belonging opportunistically helps explain how apparently strong elements of cultural cohesion can arise with great force within a relatively short time. In archaeology we often struggle to explain the more rapid and decisive episodes of change in social terms. Thinking about ethnicity as a flexible, political concept helps us fill this troublesome gap in our explanatory arsenal. The lens of ethnicity focuses us on how individuals and groups dealt with moments of upheaval and rapid transformation, which have otherwise remained difficult to handle.

In sum, where ethnicity highlights the situational, strategic use of group identities in socio-political discourses, it provides a way to understand important aspects of how people deployed material culture in strategic interaction with others. This starting point does not give us a clear, off-the-shelf methodology or a definite answer. However, it gives us the possibility to characterize one way in which collective identities formed. Ethnicity as a specific and powerful social discourse has to be part of our theoretical understanding and our explanatory models for reasoning about past societies. It should not be discarded out of fear of re-introducing discredited, essentialist notions of culture or ethnicity. Rather, it must be grasped and developed if we are to make the most of new methodologies and interdisciplinary collaborations.