Military and Ethnic Identity Through Pottery: A Study of Batavian Units in Dacia and Pannonia

1 Introduction

Natives of the Low Rhine region (modern Netherlands – Nijmegen area), the Batavi were a community of ethnic soldiers engaged in a treaty with Rome and led by their native aristocracy, overseeing a heavy supply of men to the Roman auxiliary units. Loyal and fierce, they soon became elite soldiers, the emperors’ personal guards, and some of the Empire’s most valued auxiliaries. Regardless of (as well as due to) their extraordinary status, in AD 69, they were the protagonists of one of the fiercest rebellions faced by the Roman Empire during the first century.

The present article, part of a wider research, examines the pottery discovered at and retrieved from three Batavian sites: Războieni, Adony, and Romita. Their status as military sites shaped around the consumption habits of Batavian auxiliary troops makes them comparable. It is a known fact that the Batavians’ material culture is remarkably modest and unspecific; researchers use the term loosely to refer to material finds from the area they consider to belong to Batavian territory, but much of this is widespread in the area of the Lower Rhine. For comparison, the neighbouring tribe, the Cananefates distinguish themselves by retaining their traditional hand-made pottery at least until about 160/200 AD (De Bruin, 2019).

Meanwhile, the Batavians change to imported pottery as early as the first century AD. While somehow elusive, material culture holds the key to our understanding of the Batavians’ integration, acculturation, and becoming during the late second and third centuries. Early on, the unicity of Batavian material culture did not consist of specific pottery types, but rather unusual combinations of universal pottery styles. This became particularly obvious in the rural cemeteries of returning Batavian veterans, who combined the indigenous Gallo-Belgic beakers with the more cosmopolitan and military-related colour-coated beakers. These selections spoke on the multi-faceted identity of these soldiers as part of the process of Batavian ethnogenesis (Pitts, 2019, p. 189; Roymans, 2004).

This article aims to explore the supply and utilisation of vessel fabric, form class, and type at Războieni, Romita, and Adony, spanning Dacia and Pannonia from the early second century through the third century. This comprehensive examination will facilitate an understanding of vessel origins and their integration into regional contexts. The three sites serving as case studies played a role in both local and wider regional supply networks, connecting not only regionally with neighbouring sites but also cross-provincially with one another, as evidenced by imports of sigillata, colour-coated wares, local coarse wares, and Pannonian slipped wares (PSWs), which will be further explored below through the lens of social practice by highlighting pottery-associated behaviours such as supply, demand, and replication to illuminate the expression of identity within the pottery from these three forts associated with Batavian units.

2 Methodology

The methodology presents each site within its specific context, including its historical development, the relevant facts about the troop and their interactions, and the archaeological findings. The latter is especially significant for understanding the pottery under analysis. As the reader will see, the processes of excavation, data collection, and recording vary significantly across the sites, and in no case are we dealing with fully excavated sites. However, it is important to note that in each case, we personally examined the ceramic material and brought it to a comparable standard, regardless of the methodological setbacks. Therefore, the analyses presented here are based on our direct interaction with the artefacts, rather than relying on bibliographic sources or excavation reports when referring to the three case studies.

Methodological aspects have been discussed independently in regards to each site. This section, however, aims to show the more general framework of data collection applied across the three sites. Two types of data have been recorded, qualitative and quantitative. Qualitative data refer to the descriptive information on the qualities or attributes of the pottery, meaning that it is often expressed in a narrative form.

1. Fabric: This field designates the main fabric categories at the sites under analysis. They have been assigned according to:

   • their firing type (oxidised [OXID], reduced [COAR]),

   • level of coarseness (oxidised fine [OXIDF], reduced fine [FINE], and terra sigillata imitations [TS IMIT, PSW]), and

   • those that have traditionally been their own categories (mortaria [MORT] and amphorae [AMPH]).

1. Form class: It designates the main morphological categories of vessels, such as flagons (FL), bowls (BO), dishes (DS), beakers (BE), jars (JA), lids (LID), turibula (TUR), mortaria (MORT), and amphorae (AMPH). They have been defined according to Webster’s “Romano-British coarse pottery: a student’s guide” (1969), except the authors have merged the categories of “dish” and “platter” into the one of “dish,” to declutter the already abundant terminology existent on this matter.

2. Form type: It refers to the sub-categories of each form class, such as flat/round-rimmed bowls, everted rim jars, and storage jars. These categories have been developed according to the Museum of London Archaeology Roman pottery form codes. Where possible the types have been linked to existing typologies, to pinpoint exact analogies.

3. Functional categories: this tends to be a rather loose term that assumes the way vessels would have been used. However, throughout the text, the term “tablewares” has been employed to simply designate one concept when mentioning dishes and bowls at the same time. Similarly, when referring to jars, cooking/storage wares have been employed.

While the first two levels of assessment, the fabrics and form classes, provide wider and sometimes shallow perspectives, more suitable for a general characterisation, the third layer of assessment, the form types, offers a glimpse into the ways identities may have been negotiated via pottery consumption and practices among the Batavian auxilia in this case. This type of information highlights the diversity of styles, decorations, forms, and types of pottery involved in consumption at any site as part of the typological analysis of circulating forms in their wider imperial setting. This step is essential in the methodological approach to the material under analysis, since it collates the data into a comprehensive typology, which not only offers an overview of the selections at any site but is also suitable for further regional, national, and international comparisons.

Quantitative data refer to the information which describes the amount of analysed data, and it is often expressed in a numerical form. Primary data were collected first hand from the three case studies, allowing freedom in selecting the quantification methods. The main criterion in this process was to make the primary data compatible with already published local sites, especially when inter-site comparisons occurred. Therefore, Războieni, Adony, and Romita were quantified both by a minimum number of rims (MNR). The data quantification plays a crucial role in understanding the ratios of the qualitative attributes discussed above. Once obtaining this information, one can use it for intra- and inter-site analyses (e.g. Evans, 1985; Evans et al., 1988; Pitts, 2019; Willis, 2011).

3 Războieni, Adony, and Romita: Motivation and Historical Overview

After the AD 69 revolt, the Batavian troops were reorganised (Alföldy, 1968, pp. 47–48; Spaul, 2000) from nine cohorts that existed before the rebellion to four only. At least two of these new cohorts (certainly the second and third) were cohortes equitatae, meaning that they also had a cavalry intervention group, and all of them became miliariae by the second century. The ala, which was the elite cavalry troop, was a milliaria, theoretically comprising 1,000 men (around 800 realistically) who were highly skilled riders, often from traditional military families and trained for a military career since childhood.

The journey of ala I Batavorum milliaria post-revolt started in Germania Inferior, as suggested by a diploma discovered at Elst (Haalebos, 2000), followed by its transfer to Pannonia Superior, likely in preparation for the Dacian wars, particularly the second one (105–106) (Lörincz, 2001, no. 306, Kat. 510). The location of the ala at the beginning of Hadrian’s reign is unknown, yet probable to be Pannonia until its later deployment to Dacia Superior, at Războieni-Cetate (Figure 1), where it stayed until the abandonment of the province during Aurelian’s reign (271–274). The troop’s earliest firm attestation in Dacia is documented in a diploma from 136/138 (Piso, Benea, 1984, p. 278), but it was probably deployed here earlier, during Hadrian’s fights with the Iazyges (117/118). Numerous stamped tiles indicate the troop’s presence in the Războieni fort (Piso & Varga, 2018). A thriving rural settlement developed around the fort. Its ancient name is unknown, but an inscription from Apamea (Syria) (AE 1993, 1577), dated during the first half of the third century AD, states that Aelius Verecundinus was natus in Dacia ad Vatabos. It is generally believed that “Ad Vatabos” is the settlement of Războieni. However, without additional epigraphic information, this assertion could be challenged, as it may also represent the vicus connected to the Batavian cohort’s fort in Dacia. The fort and its adjacent settlement were “green-field” territories: neither did they overlap a La Tène habitation nor was another troop attested on location before the ala.

Figure 1

Map of Dacia and Pannonia Inferior, with Adony, Războieni, and Romita marked.

Besides the ala, cohors I Batavorum milliaria was also brought on the Dacian limes (Weiss, 2002) and represents one of the three case studies in this article. Its precise location is under discussion, but it is likely to have been Romita in northern Dacia (Dacia Porolissensis) (Figure 1). Along with the cohors III Batavorum, this might be one of the Batavian troops that faced the Caledonians at Mons Graupius (Tacitus, Agricola XXXVI–XXXVIII) in Britannia. During the Dacian wars, it was stationed in Pannonia possibly at Solva (RMD II 86; Spaul, 2000, p. 211). The earliest known record of the troop as part of Dacia’s army is from a military diploma dated to 123 (AE 2011, 179), and it was likely stationed at Romita at that time. The absence of stamped construction material builds up a certain degree of uncertainty regarding the exact location of the troop. Cohors II Britannica milliaria and cohors VI Thracum were also attested at Romita, but we cannot establish an exact succession or system of cohabitation at this time.

From Pannonia Inferior, we are focusing on cohors III Batavorum milliaria equitata, stationed at Vetus Salina (modern-day Adony) (Figure 1). This troop’s itinerary after the Batavian rebellion included Britannia, Raetia, and Pannonia Inferior (Spaul, 2000, p. 213). The earliest attestation of the troop on site, at Adony, is the diploma of 135 (Roxan, 1999, pp. 249–255), but, as in the case of the ala, it was probably deployed here earlier. The first stages of the fort construction here are not fully clear, but apparently, cohors II Batavorum first garrisoned the camp, to be replaced by cohors II Alpinorum, who arrived most probably in 102, the first year of the Dacian wars and until 117–120, when cohors III Batavorum took over and stayed here as a permanent garrison.

These are the short historical outlines of the troops’ and sites’ merged history that have been analysed below from the lens of social practice and pottery consumption to shed light on the extent to which their ceramic goods are confined to the regional supply and whether any vessels may have originated from “home” as a means of preserving a specific facet of the Batavian identity abroad.

4 Războieni

This section analyses the pottery assemblage from the vicus of Războieni, focusing on the methodological considerations and approaches used to identify the pottery consumption patterns. Next, it displays the quantitative and qualitative aspects of the data, discussing their meaning towards a wider understanding of identities associated with Batavian units and the general military environment.

4.3 Qualitative Interpretation

The qualitative interpretation will explore the form types regarding their range at Războieni.

4.3.1 Form Types

Within the reduced greywares fabric category, flat-rim jars emerged as the predominant form, with 25%, followed by everted-rim jars at 15% (Figure 2: 20–23). In contrast, flagons were mostly part of the oxidised category, comprising approximately 32% of this fabric class total quantity. Among these, reeded rim flagons (Figure 2: 6) accounted for 10%, followed by ring flagons (9%) (Figure 2: 4), continuous body flagons (Figure 2: 2) (8%), collared flagons (Figure 2: 1) (4%), and amphora type flagons (Figure 2: 3) (2%). Tablewares, including bowls and dishes, also featured prominently within the oxidised and reduced categories. The bowls made in reduced fabrics were predominantly flanged (Figure 2: 18–19). These types are more suitable for cooking, as they can be easily held by the elongated rims. Conversely, the bowls made in oxidised fabrics and local sigillatas were predominantly imitations of Dr37 and Dr44 (Figure 2: 24–28). These distinctions are less obvious in regards to dishes. In fact, the straight wall dishes were made in oxidised, reduced, and local sigillata fabrics. Not only do these findings offer comprehensive insights into form preferences and their connection to fabrics and related uses but also reveal material culture trends during the second and third centuries AD at Războieni vicus.

Figure 2

The range of form types from the three case studies.

5 Adony

This section analyses the pottery assemblage from the ancient site and fort Vetus Salina (now Adony), focusing on methodological challenges and approaches to identify pottery patterns. The results, presented in quantitative (fabric and form class figures) and qualitative data (form types), highlight their significance in the regional context.

6 Romita: Seeking for Cohors I Batavorum

7 Usage and Consumption: Comparative Remarks

The first mutual element across the three sites was the consumption of terra sigillata imitations. Their presence represents a mutual trait of the second and third centuries in the provinces of Pannonia and Dacia. Jelinćić (2003, p. 85) observed that local production increased locally in Pannonia at Ilok, now including local imitations of sigillata, aside from the original imported products. The black-slipped pottery class (PSW) imitated a wide range of terra sigillata forms, but from AD 130s–140s onwards, the Dr37 bowl became the predominant imitation at several sites in Pannonia, such as Mursa, Cibalae, Aquincum, and Siscia (Leleković, 2018). The pottery assessment from Adony confirms the adherence of its community to these provincial Pannonian trends.

Equally, the high consumption of similar fabric classes in Dacia, particularly the red-coated local sigillata, is most likely an effect of the pottery production development in central and northwestern Dacia in this period (Rusu-Bolindeţ, 2016, p. 379). Războieni, along with other eleven sites from Dacia, have been identified as official producers of local sigillata products (Rusu-Bolindeţ, 2014, p. 159), resulting thus in a high return of such fabrics. Additionally, the arrival of the Roman armies in Dacia at the beginning of the second century AD from the neighbouring province, Pannonia, may have meant the introduction of PSW prototypes (Schindler Kaudelka, 2022, p. 26), which ultimately inspired the local production. The pottery assessment from both Romita and Războieni appears to now confirm indeed the passing of fashions upon military arrivals. In short, the results from Dacia and Pannonia highlight that Adony, Războieni, and Romita were part of a wider supply route that covered neighbouring provinces and served the needs of these sites and their associated communities.

Table 5 focuses further on the vessel form classes from a series of sites in Pannonia and Dacia, dating to early and mid-Roman periods and analysing altogether the three case studies and several of their neighbouring sites for a better local and regional contextualisation. From Pannonia Inferior, assemblages from Baracs (Kovács, 2001) and Ilok (Jelinćić, 2003) have been analysed as they included material from the second and third centuries AD, compatible with Adony. Adony, Baracs, and Ilok are sites of military character and their consumption appears rather similar from the perspective of form preference, opting mainly for jars and bowls. The same pattern seems to follow at both Războieni and Romita, where jars and tablewares (dishes and bowls) are among the most popular forms across the assemblages under analysis. This goes to show that the pottery supply and consumption were interconnected provincially and regionally, resulting in similar trends across Pannonia and Dacia, which could have extended to cooking, eating, and drinking practices. The form class patterns show that pottery consumption in military environments had its own character.

8 Leftover Batavian Traits?

While the pottery assessment has confirmed the existence of a separate generic military identity of the auxiliary units across the three case studies, it is inevitable to ask whether a separate Batavian identity could be recognised at any of these sites. Initially, it was stated that the transfer of Batavian units to the continent starting with the end of the first century AD marked their “denationalisation” (van Rossum, 2004, p. 120) and the general end of ethnic recruitment (Haalebos, 1999, p. 201; Kraft, 1951, p. 50). However, subsequent studies (e.g. Roymans, 2004; van Driel-Murray, 2009) proved that not only is it likely that some form of ongoing ethnic recruitment may have continued after the transfer of the units to the continent, but also that they kept some sort of ethnic identity spirit, often visible through the epigraphic record. Based on such evidence, an additional assemblage from Războieni was also analysed from this perspective to trace Batavian identity in material culture.

9 Drinking and Beakers

The assemblage discussed in this section consists of the pottery recovered from the 2018 excavation campaign at Războieni fort. Table 5 shows the total MNR of 86, along with the percentages of form classes. From these results, the most striking pattern to emerge is the beaker consumption at Războieni fort. It presents a higher quantity than the Potaissa fortress which amounted to a 507 MNR. Considering the differences in the assemblage quantities, it is striking to see Războieni displaying such a high percentage. Additionally, the nearby civilian settlement from Războieni also displayed a significantly lower quantity of beakers. While these patterns could be a result of the different sizes of the assemblages, nonetheless, the gathered data present reliable quantities to achieve initial patterns which could ultimately be assessed further in future excavations.

The high beaker consumption observed in the fort might be an indication of a specific drinking practice linked to the presence of the Batavian ala at Războieni. The importance of drinking within the Batavian society has been observed on multiple occasions (Pitts, 2019, p. 71; Roymans, 2004, p. 240; T. Vindol. III, 628). Thus, the quantities of beakers may represent a preliminary trait of the Batavian character of the unit; a possible continuation of this behaviour stereotypically associated with these units from the beginning and ongoing even in the late second and third centuries. To further scrutinise this possibility, it is important to assess the form types and see whether the practice alone was kept, or whether it was supplemented with foreign vessel types.

The beakers within the fort displayed varied profiles: curve-rimmed, straight, or oblique profile. The straight/oblique type, likely glass imitations prevalent in military contexts, is part of a category termed “legionary ware” (Greene & Dore, 1977; Swan, 2004). Notably, a biconical beaker from Războieni fort (Figure 3: BEC1) stands out morphologically and aesthetically, resembling examples from northern Gaul rather than typical Dacian assemblages (Crizbasan, 2023, 2024). Its fine, reduced firing and polished surface suggest that it may be an imitation of the P54-57/Holwerda 26 biconical beaker, particularly the P56.1 typological form originating from regions such as Gallia Belgica (Deru, 1996, p. 131; Pitts, 2019, pp. 188–189).

Figure 3

Beakers and face pots typology from Războieni.

However, it is worth noting that while the style resembles the terra nigra beaker, there is a significant chronological gap between its actual popularity “at home” and the Războieni example. Additionally, in terms of quantities of BEC1 alone, the low percentages from Războieni may indicate personal ownership rather than the formal supply of these beakers at this stage. Therefore, the overall beaker consumption, along with this specific type, shows possible remaining links of the unit to northern Gaul, inviting further research in the future. The Batavian auxilia may have maintained a certain awareness of the trendy objects at home, and when they arrived in foreign territories, they attempted to reproduce some of them to maintain a familiar consumption facilitated by familiar objects.

10 Conclusion

In conclusion, the comprehensive examination of pottery consumption patterns at Războieni, Adony, and Romita provides valuable insights into the dynamics of material culture in Dacia and Pannonia during the early to mid-Roman periods. Although we are fully aware of the limitations of the datasets – stemming from the archaeological excavation methods used (particularly in the case of Adony, where older research methods are involved) or from the nature of the site itself (such as Războieni, where modern disturbances make it challenging to establish a precise decade-by-decade chronology) – we believe we have succeeded in extracting relevant and interesting data from the ceramic datasets. While future discoveries may influence or alter some of our results, especially in quantitative terms, the validity of most of the identified patterns remains certain.

The methodological approaches employed in assessing these pottery assemblages have facilitated a nuanced understanding of supply networks, consumption preferences, and cultural interactions within these frontier regions. The quantitative analyses of fabric and form class distributions reveal distinctive trends at each site, reflecting variations in local production, regional trade networks, and social practices. Despite differences in specific fabric compositions and form preferences, mutual traits emerge, such as the prevalence of terra sigillata imitations and the prominence of functional vessels like tablewares and cookwares across all sites. These shared consumption patterns suggest high connectivity and cultural exchange within the military communities of Dacia and Pannonia.

The qualitative interpretations further highlight the significance of certain vessel types and their cultural associations. While these findings emphasise the similarities in pottery consumption across the three forts, influenced by military presence and regional trade networks, they also suggest potential indicators of Batavian identity within the auxiliary units stationed there. It is well known that Batavian material culture is not distinct in itself – especially not at the level of common pottery – but is instead integrated into the broader production and consumption patterns of the Lower Rhineland and northern Gaul regions. This makes it extremely hard to pinpoint “Batavian artefacts,” but the presence of certain material traits connected with the aforementioned area, in the context of a Batavian troop stationing on site, suggests the presence of soldiers drafted from the Batavian area. For example, the presence of beakers at Războieni hints at a Batavian presence and the preservation of certain habits within the auxiliary unit, as reflected in the material culture. Meanwhile, imported face pots at the same site provide intriguing insights into cultural connections with the Rhineland.

In summary, the qualitative interpretation of pottery may hint at broader cultural connections, revealing facets of ethnic identity among auxiliary units. Overall, this study emphasises the multifaceted nature of pottery consumption in frontier regions and underscores the importance of material culture in understanding identity formation, cultural exchange, and social practices in auxiliary units. By contextualising pottery within broader historical and archaeological frameworks, this research contributes to a more nuanced understanding of the cultural aspects of auxilia.