This paper considers specific artefact sets and mortuary practices from the Roman world of the south-eastern Alpine region that have generally been interpreted as material expressions of identity and definitions of spatial boundaries. These include the ‘Norican-Pannonian’ costume set, the ‘Norican-Pannonian’ barrow phenomenon, and the so-called ‘House’ urns of the Latobici that appear after the Roman occupation of the region, clearly overlapping in space and time (Figure 1). Given the archaeological evidence present, it is necessary to discuss the issues relating to the concept of Romanization (e.g. Keay and Terrenato 2001; Revell 2009; Gardner 2013: 1–25), which differs from the debates on the subject of ethnogenesis in the Late Roman and Early Medieval periods (e.g. Mirnik Prezelj 1998: 361–381; Brather 2002: 149–176; 2005; Gillett 2002: 1–18; Pohl 2002: 221–239). Although there may be some elements in common, for instance, with the case of the Pannonians, of an identity that was ascribed by outsiders to groups of different ethnic origins in a geographical area with poorly defined, fluid boundaries (Džino and Domić Kunić 2012: 95–101), we believe these interpretations to be too simplistic. Importantly, it is necessary to take into account the possibility that these funerary activities were at least partially related to different social boundaries, and could indicate a range of identities related to gender, status, and ethnicity, in addition to reflecting both pre-Roman and newly formed provincial identities.

Initial Roman involvement in the south-eastern Alpine world came in the aftermath of the Second Punic War. In particular, the foundation of Aquileia as an entrepot for trade with the regnum Noricum between 183 BC and 181 BC and the conquest of Istria in 177 BC brought the Roman state into the Caput Adria region and on to the borders of what later became the provinces of Noricum and Pannonia Superior (Pauli 1982: 29–30). The Roman south-eastern Alpine region can then be divided between the province of Noricum, parts of the province of Pannonia Superior, and Regio X of Italy. While Pannonia and Regio X were conquered in a series of wars over the course of the second and first centuries BC, culminating in the Pannonian Revolt of AD 6–9 (Džino and Domić Kunić 2012: 98–100; Istenić 2014: 9–24), the occupation of the area that later became the provinces of Pannonia Superior and Pannonia Inferior took place during the Augustan period following the conclusion of the Pannonian Revolt in AD 9 and the foundation of the colonia of Emona in AD 14. Interestingly, Noricum exhibits a different trajectory. Although it seems to have originally been part of this wider ‘Pannonian’ region, it first became a client state in the first century BC and was peacefully absorbed into the empire and organised as a province during the reign of Claudius (Pauli 1984: 31–33; Cunliffe 1997: 217–218). It is in the context of the aftermath of these events that we seek to examine the ways in which incorporation into the Roman Empire was negotiated through various phenomena specific to the Roman south-eastern Alpine world. However, a brief consideration of the theoretical approaches to Romanization and ethnogenesis is appropriate before treating the material evidence since a rethinking of the ethnic and social boundaries that existed in the south-eastern Alpine region during the Roman period is first required.

Romanization was formerly seen by German and Italian scholars in the 19^th century as a simple matter of the imposition of Roman imperial organisation and culture on the subject populations of the Empire by the means of military conquest, leading to the creation of a relatively homogenous political and cultural entity; however, this model has long been deemed inadequate (Keay and Terrenato 2001: xi–x; Terrenato 2001a: 1–6; Van Dommelen 2001: 71–72). Keay (2001a: 113) defined his view of Romanization as being ‘a symbiotic but unequal process of cultural exchange [which] is born out of unequal relationships between a dominant imperial power and its subject communities.’ He emphasised the important role of local elites in the adoption of Roman cultural symbols, such as monuments, buildings, and inscriptions, but it would also be appropriate to look at certain forms of material culture since the adoption of these symbols may have taken place as part of the context of elite strategies of self-empowerment and denial to non-elites (Keay 2001a: 113). All of the papers on the western provinces in the Iberian Peninsula, Gaul, and Britain in the influential volume Italy and the West: Comparative Issues in Romanization edited by Terrenato and Keay (2001) (e.g. Alcock 2001: 227–230; Castro Lopez and Gutierrez Soler 2001: 145–160; James 2001: 187–209; Keay 2001a: 113–116; 2001b: 117–137; Woolf 2001: 173–186) discuss the elite negotiation model that was formulated for the incorporation of more complex groups in Italy. This model shows the attraction of incorporation into the Roman state for complex elites in (some) Etruscan and Hellenistic cities, which included stability of land ownership and continuing involvement in municipal government, as opposed to the resistance to incorporation by other groups such as the Samnites and Liguri (Terrenato 2001b: 54–67). This may have also been the case with some of the Cisalpine Gauls (Williams 2001: 89–101).

This does not mean that all groups and elites (or parts of such groups) were incorporated into the Roman state voluntarily or without resistance. In the regions mentioned above elite incorporation often took place after military activity, but the conquest of both Gaul and southern Britain was made possible (in-part) by the active participation of some native elites, which, in turn, helped form the basis for later organisation of the civitates which have their roots in the polities of the pre-Roman Iron Age (James 2001: 193–196). For example, engagement with the Roman state guaranteed landownership and transformed military competition into competition through the municipal and provincial elected bodies and religious collegia, especially the imperial cult (Woolf 2001: 173–183).

The approaches discussed above have also been supplemented by another theoretical school of thought known as the ‘Vienna School’ model of ethnogenesis. This influential model originated in the 1960s with Wenskus’ (1961) study of Germanic antiquity in Germany and Austria and it has been widely applied in the study of Late Antique/Early Medieval history and archaeology (Gillett 2006: 241–246). It suggests that particular dynamics of ethnic identity formation that pre-dated the Roman Empire were muted by it, but re-emerged in the Late Roman and Early Middle Ages to become the basis for the formation and maintenance of ‘peoples’ and ‘states’ (Gillett 2006: 243).

The model of ethnogenesis was initially applied to the formation of northern European barbarian groups such as the Goths (Wolfram 1990), Franks (Murray 2002: 53–54), and Alemanni (Brather 2002: 158–161). It proposes that their central defining characteristic was a political and cultural process of self-identification that was not fixed or hereditary, but generated and reified by elites in order to integrate the diverse members of their followings into a single group loyal to them (Bowlus 2002: 245–246). In addition, Pohl (1991: 39–42) has suggested that there are three important factors to produce ethnogeny and maintain (or sustain) group cohesion: military success of war leaders; origin myths of the royal dynasty and their communication; and the presence of tradition bearers—an inner circle of elite members who subscribe to a group identity and the myths attached to it. These tradition bearers supposedly constitute a more direct descent group and are known as the ‘Traditionskern’ or ‘kernel of tradition’ (Pohl 2002: 224–225; 2005: 255–265; Gillett 2006: 245–246).

The ethnogenesis model in archaeological studies discusses the importance of this ‘kernel of tradition’ in addition to artefacts, which are believed to define group membership in the Roman, Late Roman, and Early Medieval periods. However, this has been attacked as perpetuating the equation of elements of material culture with ethnic groups (Curta 2007: 160–162). Brather (2002: 164–167) has also noted that caution should be exercised in the isolation of individual elements such as brooches and their use to represent particular ethnic groups. Artefact distributions in mortuary contexts may represent different types of identity, such as social rank and gender, as well as indicating directions of trade and exchange (Brather 2002: 169–175).

Studies on the distribution of artefacts recovered along the historically attested Early Medieval frontier between the Franks and the Alemanni have shown notable differences in the combinations of artefact types found in mortuary contexts on either side of that frontier (Curta 2007: 168–169). It has been posited that these combinations probably represent a deliberate choice of material culture made to highlight inter-group differences given the use of ‘emblemic’ and assertive styles which tend to appear at critical junctures in the regional political economy when changing social relations would impel displays of group identity (Curta 2007: 173). Although the artefacts may not have been produced with this in mind, they seem to have been employed in this way.

The Roman frontier on the Rhine and Danube rivers also played an important role in the formation of distinct ethnic groups in the region (Curta 2005: 173–204). The Alemanni and Franks initially appear in the written sources as both raiders and troops employed by the Romans from the late third century AD onwards (Brather 2002: 159). These interactions are marked by the presence of Roman military belt sets in Germanic mortuary contexts in the area to the east of the Rhine and the adoption of Kerbschnitt (chip-carved) decorative techniques to fibula decoration in the fourth and fifth centuries AD (Brather 2005: 150–154). A composite Romano-Germanic frontier culture with Reihengräberfelder (row grave cemeteries) developed in northern Gaul and along the Rhine in the mid fifth century AD, marking the acculturation of the Roman provincial population and the incoming Germanic groups. These are cemeteries of flat inhumation graves organised in rows, which are equipped with grave goods characteristic of Germanic and other ethnic groups in western, central, and eastern Europe in the Early Medieval period (Brather 2005: 162, 167–168).

However, there are also earlier examples of ethnogenesis on Roman frontiers – a recent study of the Batavians by Roymans (2004) highlights the importance of the Batavian royal clan as a ‘Traditionskern’ initially, but in conjunction with military service and command of auxiliary units and the worship of the Batavian Hercules in the first century AD (Roymans 2004: 251–254, 258). Yet after the revolt of Julius Civilis this clan was replaced by a broader and more diffuse group of ‘military’ families in the second and third centuries AD (Roymans 2004: 254–258). Interestingly, the separate ethnic identity of the Batavians was reinforced by military service and deliberately encouraged by the Romans. Roymans (2004: 251–260) also highlights the existence of smaller ‘ethnic’ identities within the macro categories of Germans and Gauls, and it is precisely these nuances of identity that this paper seeks to address in the Roman south-east Alpine region.

Clothing and costume work as a code that enables people to speak to their audience about their place in society and their identity (Carroll 2013: 288). Certain clothes could give the wearer a group identity as well as exclude those who could not or were not allowed to (or did not want to) wear them. The ‘Norican-Pannonian’ costume set is closely, but not exclusively linked to the phenomenon of ‘Norican-Pannonian’ barrows and barrow cemeteries, which occur throughout western Pannonia and eastern Noricum and seem to represent a specific elite burial rite in the early Roman period (Garbsch 1965; Garbsch 1985: 546–577; Palágyi 2003: 257–258) (Figure 1:1; 1:2). The epigraphic evidence suggests that this group of elites incorporated elements of Italic names, but was largely dominated by those with native names (Palágyi 2003: 259). The names of the deceased themselves link both the elite burial rite and the costume elements found in mortuary contexts with those depicted on stone figural representations to a specific provincial elite context, emphasising local elite identities through new mediums of expression.

The ‘Norican-Pannonian’ costume is often depicted as the apparel of elite women on funerary monuments in parts of Noricum and Pannonia Superior in the first and second centuries AD; although it should be noted that the majority of the funerary monuments have been dated on stylistic grounds and were not found in their original contexts (Horvat 1999b: 278–280). Indeed the names of female individuals depicted in ‘Norican-Pannonian’ costume are exclusively native, causing the interpretation of this style of dress to be understood as a native costume (Garbsch 1965: 134). However, the most recent study of the Norican-Pannonian costume shows it to have much greater variability than the earlier studies revealed as Rothe (2012: 222; 2013) interprets it as having ‘a myriad of local styles.’

The Norican-Pannonian costume-set consisted of a number of garments including a long-sleeved bodice and underskirt (Garbsch 1985: 558; Rothe 2012: 180–183). This was supplemented, in the case of adult women, by a sleeveless over-tunic fastened together with a pair of characteristically shaped brooches—one on each shoulder. This costume is illustrated by a damaged tombstone from Ig near Ljubljana that is dedicated to Quarta, who died at fifty years of age, and her daughter Tertia, who died at fifteen years of age (City Museum of Ljubljana, inv. n. 510: LJU; 0036870). The upper part of the tombstone bears the frontal portrait of the deceased woman in a niche with her clothing clasped by a brooch on each shoulder (Figure 2). Brooches of this type are sometimes described as ‘Norican-Pannonian’, or ‘norisch-pannonisch’ (Garbsch 1965: 26–79). This term is also applied to wing brooches and the ‘Doppelknopf brooch’ (Rothe 2012: 137–138) (Figure 3). A distinction can be made between women’s fashion and girl’s fashion, with further regional variants. For example, the ‘traditional’ costume, which is dated to the first through early third centuries AD, features matrons wearing a special headpiece, typically called a bonnet, while girls were often depicted holding a mirror in one hand and a box/jug/towel in another; they also wore a belt of a specific form around their waist (Figure 3). This constituted the typical female costume with a large number of variants in both towns and rural areas in the provinces of Noricum, Pannonia Superior, and the Danube bend in the northern part of Pannonia Inferior (Rothe 2012: 137). In formal terms, the elements of the ‘Norican-Pannonian’ costume lack any direct Late Iron Age precursors. However, the geographical distribution of certain variants of the over-tunic and bonnet do seem to reflect the known locations of some of the civitas in Pannonia and Noricum, although in the case of the bonnet this is also complicated due to chronological differences (Rothe 2012: 183–190, 193, 198–211).

The ‘Norican-Pannonian’ costume-set is a very distinct style of dress, not only in relation to the ‘Roman’, but also in relation to representations of the opposite gender—e.g. ‘manhood’. The latter, although quite indistinct in the contemporary archaeological record, is clearly recognisable on tombstones where a man in a toga sits next to a woman wearing a ‘Norican-Pannonian’ costume (Šašel Kos 1997: 412–413). The toga—a traditional Roman piece of clothing—is often considered an important marker of male status, civil rights, and position in Roman society; it essentially functions as a national costume and a symbol of romanitas (Larsson Lovén 2014: 430–434). Moreover, it is clearly an item relating to males only and, in particular, an item of clothing worn only by elite males. However, there are instances in which this is negated, e.g. infames, such as prostitutes, wearing a toga (Ackerman 2016: 11). As such, the ‘Norican-Pannonian’ style of dress should not just be seen as an expression of the conservatism of female members of the provincial elite, it also might have been part of a demonstration of both status and gender in provincial Roman terms alone, lacking any reference to the pre-Roman tradition. Traditional interpretations of the ‘Norican-Pannonian’ costume rely upon the assumption that those women preferred to think of themselves as of the natives or, in other words, within the postulated veracity of a dichotomy of native vs. Roman (Garbsch 1965: 3–4). However, this may be a construct of our present rather than a reality of the past, as these women may have viewed themselves as respectable females of their space and time whose clothing clearly marked a distinction between the unmarried/marriageable or bride and the married. This was one of the crucial distinctions in Roman society, where there was hardly an unmarried person after a certain age (Lefkowitz and Fant 2005: 102). If this is true, then the ‘Norican-Pannonian’ costume would have aptly served as an external manifestation of correct and proper female identity in that place and time; most probably part of a female’s ceremonial image just like the toga for males. The costume of a married women then represents the power and status enjoyed by a provincial matron. What is involved here is thus not a revolt, but a particular mode of compliance with Roman society, in which costume was a medium for non-verbal communication and a reflection of the social status of the individual. It was also a means of constructing identity (Larsson Lovén 2014: 439); yet the message carried by this costume would be dependant on the individual reading it. A Roman woman from central Italy may have seen a woman from Emona in the local variant of ‘Norican-Pannonian’ costume as being different, foreign, and/or non-Italian. However, other women from Emona would have likely been aware of the numerous shades of meaning with regard to the wealth and status of the women wearing this costume, in addition to, or perhaps even dependant on, the ethnic connotations mentioned above.

It might be more reasonable to rely upon gender-related notions than ethnos-related, since throughout a life-time the human body moves dynamically through numerous identities, which are often regarded as static by archaeologists. Although clothing is not the only means of expressing gender roles, it is one of the most visible ways of expressing identity with regard to physical gender. Being a woman in the Roman Empire constituted an altogether different experience from being a woman in the pre-Roman Celtic world, or rather, the Late Iron Age world of the Middle Danube. In Roman society, a woman was clearly subordinate to a man, i.e. to her father, husband, brother, or son. A girl—unlike a boy—grew up overnight, through a marriage that conferred a new identity upon her – an identity that, in Roman society, was mainly related to reproduction (Carcopino 1967: 89–97). However, this may imply the existence of a female high status that differs in some way to the extremely patriarchal Roman standard, the perceived ‘Roman’ norm. For example, in Late Iron Age ‘Celtic’ societies women are generally thought to have had higher status than they did in Roman society (Cunliffe 1997: 109–110). This is also posited for the Early Iron Age Dolenjska group in south-eastern Slovenia, which partially overlaps with the southern part of the distribution of ‘Norican-Pannonian’ barrows (Teržan 1985: 77–105).

Indeed, there is evidence for female high status costume in the mortuary record in both the Early and Late Iron Age throughout the area of the later distribution of the ‘Norican-Pannonian’ costume set. It should also be noted that much of this also employs distinctive fibula types in combination with arm and ankle rings to form locally distinctive costume groups. It is also depicted in figurative repoussé decoration on bronze vessels, which show evidence of head coverings or shawls (Teržan 1985: 80–83, 88–89, 92–94; Mason 1996: 113–116; Turk 2005). Thus, there is a longstanding tradition of distinctive regional and socially predicated costume in the south-eastern Alpine region.

It has been suggested that female status varied over the region of study in the Early Iron Age and that there is evidence for female sacrifice in the elite mortuary rituals in the Sulmtal Early Iron Age group, which is coterminous, in part, with the distribution of the ‘Norican-Pannonian’ female costume set and the ‘Norican-Pannonian’ barrow phenomenon (Teržan 1990: 124–125). This is somewhat problematic given that the evidence for this is largely derived from poorly documented nineteenth century excavations of the barrow cemeteries surrounding the Burgstall at Klein-Klein in Styria (Dobiat 1980: 47–63) that assigned gender solely on the basis of grave goods and not the cremated remains of the dead. The presumed female sacrificial victims are equated with the typologically female grave sets in the most elaborate barrows (Mason 1996: 67). These graves also contain elaborate weapon sets, armour, and sheet metal vessels that are interpreted as typologically male grave goods. However, similar graves sets in other barrow groups in the same complex are identified as being indicative of high status women. Thus, it is not clear if the individuals represented by the typologically female grave goods in the most elaborate barrows are considered to be of high status, or that their presence is merely indicative of the high status of the typologically male elite burials in these barrows (Teržan 1990: 126–137).

It is interesting that the Norican-Pannonian dress style is associated with female costume and might represent an emphasis of traditional female identity and, thus, a tacit resistance to the reduction of female status by Roman society. This is especially true of the examples in which the female costume depicted on tombstone portraits was not realistic since such examples were probably intended to represent an idealised version of the deceased. The tombstones can be seen to represent a space for (idealised) self-representation that could reflect the social persona of the deceased along with their attendant ethnic identity, wealth, and status.

The complexity of the costume and its variants are known precisely due to the representations of their many perishable elements on the tombstones, as the surviving grave goods could not provide this detail. There is a growing body of epigraphic evidence that also bears witness to geographically circumscribed female costume in other areas of the western provinces (Garbsch 1965; Rothe 2009; Carroll 2013a). Here one might consider the evidence from the Rhineland, the Moselle, or even the northern frontier in Britain, suggesting that certain areas of other western provinces also possessed distinctive female costume only depicted in the iconography (Carroll 2013a: 288–299; 2013b: 7–10). Indeed, it might be appropriate to consider these costumes in terms of ‘emblemic’ styles that are noted on the basis of grave goods in the Early Middle Ages (Curta 2007: 173).

The major difference between the Early Medieval mortuary evidence and that available for the range of ‘Norican-Pannonian’ costume sets is the additional detail available from the iconographic record. Depictions of these female costumes are located along the frontiers and indeed to some extent localised within civitas boundaries in the period from the first century AD to the early third century AD. They seem more area specific than the representations of male costume, comprising elements that are found in other areas of the north-western provinces and the general male identification, which are generally considered as ‘Pannonian’ (Rothe 2012: 143–173). Indeed it might be argued that the elements of the male costume of the Middle Danube supplanted traditional Roman male costume by the third century AD (Rothe 2012: 145–148). Indeed, the term ‘Pannonian’ was employed to describe a ‘male’ frontier identity it was in this period that comes to the fore in the third century AD (Džino and Domić Kunić 2012: 93–115).

The most complete and/or elaborate ‘Norican-Pannonian’ dress sets are found in the so-called ‘Norican-Pannonian’ barrows. These are large purpose-built earthen barrows that contain elite burials in a central rectangular stone chamber accessed in some cases via a stone dromos passage (Palágyi 2003: 258–259). The ‘Norican-Pannonian’ barrows are, as the name implies, found in both Noricum and the western part of Pannonia Superior. They extend into south-eastern Slovenia, north of the Gorjanci hills and into western Croatia, as far as colonia Flavia Siscia (see Figure 1: Symbol 1). The female cremation burials are fully or partially equipped with fibulae, belt sets, and other elements of the ‘Norican-Pannonian’ dress sets, whilst the male cremation burials are sometimes equipped with carts/chariots and hunting weapons (spears and bows), as well as elaborate grave goods including glass cremation urns, glass and fine ware drinking vessels, and dining services of fine ware bowls, dishes, and plates (Palágyi 2003: 257–261). They are found in lowland areas that had a tradition of barrow burial in the Early Iron Age, but not in all areas of Noricum and Pannonia that had an Early Iron Age barrow tradition. In fact, the immediate pre-Roman period was marked by flat cremation burial throughout the region (Božič 1991: 471–477; Božič 1999: 192–195).

It is possible that the barrow phenomenon represents the identification of the ‘Romanized’ or Roman provincial elite with earlier forms of elite expression, which would also associate them with a specific area/territory. Thus, these barrows may be related to the assertion of rights of tenure over a specific landholding/estate. Their origins may have been rooted in the pre-Roman past and legitimated by association with pre-Roman forms of affirmation (barrow cemeteries, the ‘illustrious dead’), but not the flat cemeteries, which were the mortuary practice in the immediate pre-Roman period. Equally, barrow burials could also be understood by the Romans through the medium of Greco-Roman mythology and the Etruscan and Italic past. The barrows created loci or ancestral places in the landscape that were associated with new Roman-style villa estate centres without evidence of previous mortuary activity (Breščak 1990a: 43–44).

Large barrow cemeteries are concentrated in the territories of Poetovio, Flavia Solva, and Savaria, whilst single barrows are more common south of the Dravinja, extending as far as the Dolenjska area (Horvat 1999c: 288–291; Palagyi 2003: 257–261). Within these areas they are at least partly associated with villa estates, for example at Miklavž, Lancova vas, Medvedjek, and Groblje (Breščak 1985: 39–40; Horvat 1999a: 228–231). However, it should not be assumed that all of the provincial elite were buried in ‘Norican-Pannonian’ barrows or used ‘Norican-Pannonian’ costume. Indeed, part of the provincial elite of native origin in Celeia were buried in funerary monuments in overtly classical style and decorated with classical mythological motifs at Šempeter (Kolšek 1997). It should also be noted that the trans-cantonal nature of the Celeia elite is represented in a cemetery at Jezero between Trebnje and Novo mesto on the route of the main Roman road from Emona to Siscia. Here, a recently discovered tombstone commemorates two members of the town council of Celeia, who had Italic names, suggesting that the cemetery was probably connected with a villa estate or a road station (Breščak and Lovenjak 2010: 295–310).

The ‘Norican-Pannonian’ dress set also appears in attenuated form outside the sphere of barrow burials and funerary depictions, as fibulae and other individual elements of the ‘Norican-Pannonian’ costume set are known in Early Roman flat cremation graves in the region, e.g. Beletov vrt—Grave 27 and Grave 28 (Knez 1992: 29–30, Table 8: 17, 18, 30, Table 10: 8, 9); Obrežje—Grave context no. 1819 (Mason 2012: 398); and Ljubljana-Kongresni trg—Grave 1042 (Gaspari et al. 2015: 141–143) (Figure 4). The graves from Novo mesto contain paired wing brooches and are part of a Late Iron Age/Early Roman cemetery, whilst the Obrežje grave contains a single double-knobbed fibula and is part of a small cemetery within the abandoned Roman vexillation fort (Mason 2012: 398) However, the example from Ljubljana was deposited in a military weapons burial—the soldier was buried with his equipment: a sword in its scabbard; a shield, of which the umbo survives; two spearheads; and a knife. In addition, a ‘Norican-Pannonian’ brooch was found together with a set of pottery vessels, comprising a beaker, a plate and two jugs interpreted as being used for the service of food and drink. The deceased was a member of an auxiliary unit (auxilia), who was buried in the first years of the first century AD, either during or before the construction of the colony of Emona. The grave was located close to a group of Early Iron Age barrows, thus certain elements were also deemed suitable for deposition in apparently non-elite mortuary contexts, some of which were also anomalous, given the apparently military male context of supposedly female grave goods. Whatever the case, this metalwork also cuts across status and gender boundaries and appears in both urban and rural contexts, seemingly representing a provincial identity that was not exclusively elite or female in nature.

The early Roman period in the south-eastern Alpine region was also characterised by a third more specific artefact form, exclusively mortuary in character and more restricted in geographic distribution, known as the so-called Latobici ‘House’ urn. This is a specific ceramic closed colour coated vessel form with a modelled conical roof and window or door apertures, which suggests a round house with apertures in the upper body and an attached lid in the form of a conical roof, surmounted by a variety of knobs or even a stylised cockerel (Petru 1966: 361–371) (Figure 5). These vessels are known from Roman cemeteries in south-eastern Slovenia and the adjacent part of Croatia and have been interpreted as being connected geographically with the civitas Latobicorum.

The Latobici were a Celtic group that was probably originally part of the Taurisci macro group in the pre-Roman period, but are not attested to as a separate entity until their mention as a distinct administrative unit in the epigraphic record in the Flavian period (Božič 1991: 471–477; Božič 1999: 192–201). The centre of this civitas was the municipium Neviodunum (modern Drnovo). This administrative unit borders that of the Varciani and either the territory of Colonia Flavia Siscia and/or that of the Colapiani, (Lovenjak 2003: 93; Lolić 2003: 132–133; Nemeth-Ehrlich and Kušan Špalj 2003: 107–111). The Varciani and the Colapiani may have formerly been part of the Late Iron Age Segestiani macro group, centred on Segestica (Siscia) in the pre-Roman period (Džino and Domić Kunić 2012: 104).

The Latobici ‘House’ urns seem to have been made specifically for burial and would only have been visible in rituals leading up to and including burial. They seem to be rooted in the creation of new local cultural identities, based on identification with existing pre-Roman and new contemporary Roman loci created in the landscape in the wake of the establishment of the Civitas Latobicorum. However, it should be noted that this pottery form is without antecedents in the pre-Roman Iron Age, although it is found in Roman period graves in the large Late Iron Age/Roman cemeteries that were associated with major Late Iron Age hill fort centres (Mason 2012: 392–393). The most notable examples are the Beletov vrt cemetery, associated with the Marof hill fort above Novo mesto (Božič 2008: Plate.13: 3–5, 206; Plate.6: 3), the Strmec cemetery below the Vinji vrh hill fort (Dular 1991: 54–59, T. 51–T. 82; 157–188), and the Mihovo cemetery below the Trnišće hill fort (Dular 2008: 112–126, 136–137). These cemeteries continued in use into the first and second centuries AD after the hill forts were abandoned.

House urns are also relatively common in the flat cremation cemeteries associated with the newly founded roadside settlements at Ribnica (Romula) (Breščak 2005: 230–232) and Dolge njive/Draga (the possible site of Crucium) (Križ 2003: 24–27; Mason 2006: 55–57). They also appear in graves in cemeteries associated with some industrial centres, river ports, and villa centres in the Kolpa valley in northern Bela krajina, as well as with some villa centres in the Radulje valley (Breščak 1990c: 103–104; Dular, J. 1974: 353–369; Dular, A. 1976: 191–223; Udovč in press). House urns have also been found in Roman period rural flat cemeteries that have no known associated settlements, including the cemetery with a Late Iron Age founder’s grave at Verdun, near Stopiče in the foothills of the Gorjanci hills (Breščak 1990b: 99–102).

This vessel type was also placed in some graves in cemeteries associated with Early Iron Age barrows, such as those at Dobrnić (Bavec 2006: 130–132), Medvedjek (Breščak 1990a: 43–44; Predan 2005: 194–195), and Mačkovec (Mason 2012: 391–392) (Figure 6). They also occur in the cemeteries of the civitas centre, Neviodunum (Drnovo near Krško) (Petru 1971: 27–38; Petru and Petru 1978: 38; Petru 1990: 90). However, the House urn is only present in some of the graves, and the accompanying material varies. Thus, it seems that only part of the population employed a recognisable and distinct element of material culture to emphasise identity, which was reaffirmed during funerary rituals in different locales, or rather in cemeteries associated with different types of settlement. As such, this vessel seems to have only had a symbolic meaning to the local population within this limited area.

The burial rites and grave structures in all the flat cemeteries within the area are very similar regardless of location. Cremation was the rule from the first to the third century AD, as it was in the pre-Roman Iron Age. The grave structures varied within cemeteries, but the most common forms are simple oval, sub-rectangular or rectangular grave pits, rectangular stone cists lined with upright stone slabs, or rectangular tombs lined with dry stone, mortared stone, or tegulae (Županek and Sivec 2017: 98–100). The graves might be arranged in rows or around stone walled grave plots as they were at Draga (Križ 2003: 24–27; Županek and Sivec 2017: 98–100). However, it must be stressed that these grave types occur together in the same flat cemeteries—cremations may be inurned, or more commonly, placed in the base of the grave.

The locations of rural cemeteries associated with Early Iron Age barrows were clearly carefully selected in this area. They would have been highly visible with the pre-existing barrow emphasising the antiquity of the associated cemetery, and would have also served to emphasise the legitimacy of claims to the surrounding land by the community interred in it. The fact that the well-drained arable land on the ridge tops was removed from cultivation to accommodate these cemeteries also suggests that the communities associated with them wished to emphasise the importance and extent of their rights over the land in the area. Isolated Early Iron Age barrows and barrow groups were known places in the landscape associated with an ancestral past that predated the immediate pre-Roman Iron Age (Mason 2008: 97–106). Indeed, it would seem that some of these communities were seeking to identify and lay claim to specific areas of land through a possibly fictitious ancestral past that was independent of, or pre-dated, the final Late Iron Age, as was the case at Medvedjek, Mačkovec or Zglavnice near Velika Dobrava (Brešćak 1990a: 43–44; Predan 2005, 194–195; Mason 2012: 391–392; Novšak 2006: 228–230). The creation of new ‘Norican-Pannonian’ barrows in association with villa sites might also have had the effect of referencing a mythic Early Iron Age past, but could equally be referring to the Italic past. This expansion in the number of formal mortuary sites in the Early Roman period should perhaps be seen in changing systems of power and means of expressing tenure over land that was inevitable in the wake of the incorporation of the area into the Roman state in the late first century BC and the first century AD.

Thus, it may be suggested that association with increasingly politically defunct Late Iron Age elite centres was supplanted by smaller scale identification with specific local areas in the landscape, probably as a means of exerting or emphasising direct control over the land, even at the expense of sacrificing part of it to mortuary space in the process. It is possible that the burial populations of these new cemeteries are drawn from groups that exercised control over these areas in the Late Iron Age, but were previously buried at the regional centres. It is also likely that the traditional elites were increasingly drawn to the civitas centre as an arena for expressing and maintaining status, or were supplanted by new elites of native and/or non-native origin based on association with the Roman state, some of which employed the ‘Norican-Pannonian’ barrow burial rite found close to villa estate centres, as was the case at Groblje near Novo mesto, or rural settlements as was the case at Medvedjek (Breščak 1990a: 43–44).

Other groups attempted to assert control over land through links to the recent past, but in new locations. The Verdun cemetery combines the Late La Tène warrior burial tradition with burials containing Roman military equipment (Breščak 1990: 99–102). This suggests legitimization through the linking of the immediate warrior past to the professional military present and the status that military service in the Early Roman period could confer on a member of the local population. The burials within the abandoned Augustan fort at Obrežje date to the late Julio-Claudian and Flavian periods. A link between all these cemeteries is created in the similarity of location, visibility from roadways, the eclectic nature of the mortuary structures, and the grave goods found in them—the ‘House’ urn is an element that links all these diverse mortuary sites in the civitas Latobicorum.

The ‘Celtic’ Latobici are also mentioned to in a military diploma as being part of an infantry cohort together with the ‘Pannonian’ Varciani (Lovenjak 2003: 95). There may have been an interest in emphasising a separate identity in death as Latobici, rather than Varciani by at least part of the population. This civitas was squeezed between the colony of Emona and Regio X of Italy to the west, Celeia and the southern boundary of Noricum to the north, and the colony of Siscia to the south-east (Horvat 1999a: 227–231). This boundary was also reinforced towards the civitas of the Colapiani, a Pannonian group in the Kolpa/Kupa valley, which were possibly what remained of the Segestiani after the foundation of the colony of Siscia (Džino and Domić Kunić 2012: 104). It is interesting to note that the southern distribution of the so-called ‘House’ urn closely matches the southern boundary between the Mokronog group and the Vinica group in the Late Iron Age (Božič 1999: 201–202; Božič 2001: 181–198.)

The archaeological record suggests that the use of ‘Norican-Pannonian’ ‘female elite’ costume, and the costume’s elements, is more nuanced than the simple dichotomies of elite vs. non-elite and ‘Roman’ vs. ‘native’. Concepts of female elite status in Noricum and western Pannonia differed from those in Italy. The iconographic record clearly shows that different female costume elements were combined to represent a series of geographically defined costumes, which seem to be related, to a degree, to specific local ethnic, age (unmarried/married) and status identities (Rothe 2012: 193, 212–213, 222). However, it should be noted that individual costume elements may also be incorporated in non-elite/non-military graves in former military sites such as in Grave context no. 1819 in the vexillation fortress at Obrežje (Mason 2005a: 75; 2004b: 209), or cut across gender boundaries to be incorporated into male military graves close to the Augustan fortress at Emona (Gaspari et al. 2015: 125–169).

The creation of ‘Norican-Pannonian’ barrow cemeteries in association with villa estates may represent the desire to associate local elites with visible manifestations of elite burial without being associated with former Late Iron Age power centres. On the other hand, the reuse of Early Iron Age barrows as foci for flat cemeteries subverts this activity. It may be an attempt perhaps to assert peregrine rights over land away from the former hill fort centres, using symbolic, possibly fictional, ancestral burial places in order to emphasise community control over land at a local level, i.e. a control, like that of the ‘Norican-Pannonian’ barrows, symbolically rooted in a time that predated the immediate Pre-Roman Iron Age. Other groups in the region may have sought to achieve this through the continued use of adjacent Late Iron Age cemeteries. This seems to be the case with the continued use of the Late Iron Age/Early Roman Beletov vrt cemetery by the recently discovered Early-Middle Roman settlement in the historic town centre in Novo mesto. This then may relate to association with immediate pre-conquest rights of tenure. Real or imagined connections to the new Roman power structures by association with the immediate past of the conquest itself might be symbolised in the isolated example of mortuary activity within the Obrežje fort, although there is evidence for Roman rural settlement in the vicinity (Mason 2005b: 209). The creation of these specific new pottery forms was instrumental in the creation of non-elite and elite social identities at local, civitas, and provincial levels during the early stages of incorporation into the Roman state (Petru 1972). Such pottery vessels could only have been visible during burial ceremonies and the rituals associated with death, but seem to match a civitas identity (Mason 2012: 398–399). At the same time, the relative homogeneity of the new material culture and structures found within these cemeteries and settlements cuts across their different implicit origins and serves to express a wider provincial identity that supplants that of the preceding Late Iron Age. This identity is not merely rural, but can also be seen in the formal ‘urban’ cemeteries of roadside settlements and indeed those associated with the civitas capital. It is an identity made up of an eclectic mix of old, fictitious, and new elements of the landscape and material culture (Mason 2012: 397–399; see also Curta 2001: 31). Furthermore, it should be noted that these examples relate to practice and custom on a local scale that did not follow provincial borders. It would seem that the Latobici ‘House’ urns relate to an identity connected with the territory of a specific civitas. However, caution must be exercised because the civitas itself is defined in part by the boundaries of the distribution of the Latobici ‘House’ urns as much as by the presence of the two place names that actually include the element ‘Latobicorum’. These are the civitas capital Municipium Flavium Latobicorum Neviodunum (Drnovo near Krško) and the major post/customs station at Praetorium Latobicorum (Trebnje) (Lovenjak 2003: 94). The area under discussion was characterised by the material culture associated with the Mokronog La Tène group in the Late Iron Age. This group is putatively identified with the Taurisci macro group, which was present in the wider eastern and south-eastern Alpine region in the second and first centuries BC (Božič 1991: 447–477; Božič 1999: 192–201). Whatever underlies this appearance of the ‘House’ urn element, it is something that was a product of the Roman present rather than the Iron Age past.

These changing identities endured throughout the early Roman period and into the middle Roman period, finally disappearing with the changing economic and social conditions in the third and fourth centuries AD. The disappearance of many rural cemeteries and the rural identity associated with them may be connected with the increasing importance of villa estates, particularly after the disruption caused by the Marcommanic wars of the late second century AD. The award of citizenship to the free population of the Empire by Caracalla, the impact of civil wars in the third century AD, and the impact of Christianity served to undermine the local identity of the rural population in the area under consideration, subsuming it into a wider ‘Roman’ identity.

In conclusion, it can be seen that current interpretations of Romanization operate on the somewhat simplistic terms of Roman material culture reflecting the ‘big picture’ of Roman conquest and the organization and administration of the provinces. Debates on Romanization often concentrate on dichotomies between pre-Roman ethnic/socio-political groups and ‘Romans’, between civilian and military, elite and non-elite. However, facets of personal/group identity must also be taken into account when looking at the above phenomena in the south-eastern Alpine region in the first and second centuries AD. What takes place with the occupation of the region is not a fixed form, but rather a continuous negotiation, a negotiation that was also taking place prior to Roman occupation. It was not only at the level of ‘us’ vs. the ‘other’—that ‘other’ could be of the opposite sex or a different social class—it relates to all aspects of identity, not just ethnic identity. The politically unified, but strongly culturally diverse melting pot of the Roman Empire sometimes used the old, but frequently created new forms—new regional costumes might appear in the Alpine or middle Danubian provinces, as they did in other provinces, but these costumes did not necessarily follow pre-Roman forms or simply use imported Roman forms from the Mediterranean world.